recent advances and challenges in antibacterial drug development: the editorial doi: http://doi.org/10.5599/admet.1315 89 admet & dmpk 10(2) (2022) 89-90; doi: http://dx.doi.org/10.5599/admet.1315 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index editorial recent advances and challenges in antibacterial drug development: editorial lynn silver ll silver consulting, new jersey, united states; silverly@comcast.net received: march 1, 2022; revised: march 1, 2022; published: march 4, 2022 there is a continuing need for new antibacterial drugs to deal with the burgeoning of bacterial pathogens resistant to the existing armamentarium of antibacterial drugs, compromising their effectiveness; furthermore, emerging pathogens are now considered to be a major microbiologic public health threat. over the years, these issues have been addressed with the modification of existing antibiotic classes or the search for new structural classes. the latter approach has not been particularly successful, as evidenced by the fact that the last novel antibacterial class to be developed was discovered in 1984. today the clinical pipeline is predominantly occupied by derivatives of established classes, although there are some novel entries in development [1]. since the beginning of the century, big pharma began exiting antibacterial drug discovery despite a growing clinical need [2]. apart from financial issues and regulatory hurdles, antibacterial discovery research, especially target-based approaches, has not been productive. the current preclinical pipeline has high diversity, with > 40 % of the projects representing new classes, new mechanisms of action or new targets [3]. however, this has generally been true of the preclinical pipeline all along (even when it was largely the domain of big pharma). now, these projects are most often developed by smes residing in north america or europe. the problem arises in the validation and successful development of these preclinical projects into actual clinical candidates – and that is where most projects fail. clearly focused efforts toward overcoming the obstacles to antibacterial discovery are needed [4]. research and resources are essential to progress novel approaches to clinical practices in order to sustainably fight against antibacterial resistance and emerging bacterial pathogens. this special issue of admet and dmpk is guest-edited by lynn silver, ll silver consulting and ana budimir, university of zagreb. articles in this issue include a brief review of “recent advances and challenges in antibacterial drug development” by valeria gigante et al. from the who [5] and a review of “old and modern antibiotic structures with potential for today’s infections” by david newman [6]. guangshun wang and a.f. mechesso provide a “realistic and critical review of the state of systemic antimicrobial peptides” [7]. in the area of new technologies, there is a review of “recent advances in nanoparticles as antibacterial agents” by murat ozda and sumeyra gurkok [8]. balbina plotkin and monika konaklieva discuss the “impact of host factors on susceptibility to antifungal agents” [9]. matthias fellner presents “newly discovered staphylococcus aureus serine hydrolase probe and drug targets” [10] and andrei bogdanov, et al. [11] describe a novel antibacterial scaffold in “synthesis and diverse biological activity profile of triethylammonium isatin-3-hydrazones.” http://dx.doi.org/10.5599/admet.1315 http://dx.doi.org/10.5599/admet.1315 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:silverly@comcast.net https://pub.iapchem.org/ojs/index.php/admet/article/view/1215 https://pub.iapchem.org/ojs/index.php/admet/article/view/1215 https://pub.iapchem.org/ojs/index.php/admet/article/view/1172 https://pub.iapchem.org/ojs/index.php/admet/article/view/1172 https://pub.iapchem.org/ojs/index.php/admet/article/view/1164 https://pub.iapchem.org/ojs/index.php/admet/article/view/1137 https://pub.iapchem.org/ojs/index.php/admet/article/view/1179 https://pub.iapchem.org/ojs/index.php/admet/article/view/1179 lynn silver admet & dmpk 10(2) (2022) 89-90 90 references [1] u. theuretzbacher, s. gottwalt, p. beyer, m. butler, l. czaplewski, c. lienhardt, l. moja, m. paul, s. paulin, j.h. rex, l.l. silver, m. spigelman, g.e. thwaites, j.-p. paccaud, s. harbarth. analysis of the clinical antibacterial and antituberculosis pipeline. the lancet infectious diseases 2019, 19(2) (2019) e40-e50. https://doi.org/10.1016/s1473-3099(18)30513-9. [2] s. j. projan. why is big pharma getting out of antibacterial drug discovery?. curr. opin. microbiol. 6(5) (2003) 427-430. https://doi.org/10.1016/j.mib.2003.08.003. [3] u. theuretzbacher, k. outterson, a. engel, a. karlen, a. the global preclinical antibacterial pipeline. nat. rev. microbiol. 18(5) (2020) 275-285. https://doi.org/10.1038/s41579-019-0288-0. [4] l.l. silver. challenges of antibacterial discovery. clin. microbiol. rev. 24(1) (2011) 71-109. https://doi.org/10.1128/cmr.00030-10. [5] v. gigante, h. sati, p. beyer. recent advances and challenges in antibacterial drug development. admet and dmpk 10 (2022) 147-151. https://doi.org/10.5599/admet.1271. [6] d.j. newman. old and modern antibiotic structures with potential for today’s infections, admet and dmpk 10 (2022) 131-146. https://doi.org/10.5599/admet.1272. [7] g. wang, a.f. mechesso. realistic and critical review of the state of systemic antimicrobial peptides. admet and dmpk 10 (2022) 91-105. https://doi.org/10.5599/admet.1215. [8] m. ozdal, s. gurkok. recent advances in nanoparticles as antibacterial agent. admet and dmpk 10 (2022) 115-129. https://doi.org/10.5599/admet.1172. [9] b. plotkin, m. konaklieva. impact of host factors on susceptibility to antifungal agents. admet and dmpk 10 (2022) 153-162. https://doi.org/10.5599/admet.1164. [10] m. fellner. newly discovered staphylococcus aureus serine hydrolase probe and drug targets. admet and dmpk 10 (2022) 107-114. https://doi.org/10.5599/admet.1137. [11] a. bogdanov, o. tsivileva, a. voloshina, a. lyubina, s. amerhanova, e. burtceva, s. bukharov, a. samorodov, v. pavlov. synthesis and diverse biological activity profile of triethylammonium isatin-3hydrazones. admet and dmpk 10 (2022) 163-179. https://doi.org/10.5599/admet.1179. ©2022 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/) https://doi.org/10.1016/s1473-3099(18)30513-9 https://doi.org/10.1016/j.mib.2003.08.003 https://doi.org/10.1038/s41579-019-0288-0 https://doi.org/10.1128/cmr.00030-10 https://doi.org/10.5599/admet.1271 https://doi.org/10.5599/admet.1272 https://doi.org/10.5599/admet.1215 https://doi.org/10.5599/admet.1172 https://doi.org/10.5599/admet.1164 https://doi.org/10.5599/admet.1137 https://doi.org/10.5599/admet.1179 http://creativecommons.org/licenses/by/4.0/ ten years for physchem forum-japan (pcf-j) doi: https://doi.org/10.5599/admet.1452 245 admet & dmpk 10(4) (2022) 245-247; doi: https://doi.org/10.5599/admet.1452 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index letter to editor ten years for physchem forum-japan (pcf-j) takashi mano* ono pharmaceutical co.,ltd.,1-1, sakurai 3-chome, shimamoto-cho, mishima-gun, osaka 618-8585, japan, e-mail: mano@ono.co.jp; tel.: +81-75-961-1151; fax: +81-75-962-9314 * on behalf of physchem forum-japan organization committee received: july 09, 2022; revised: july 18, 2022; published: august 16, 2022 abstract the ten years of physchem forum-japan and konstantin tsinman’s great contributions to the forum are briefly described. ©2022 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords physical-chemical property; oral absorption; physchem forum; international association of physical chemists; konstantin tsinman. in 2012 several scientists of pharmaceutical physical-chemical properties working at japanese companies got together and discussed how to improve recognition of the criticality of physicochemical properties in drug discovery and development from pharmaceutical science viewpoints. this was due to concerns that in industry, the assessment and application of physical-chemical properties of active pharmaceutical ingredients (apis) such as solubility, lipophilicity, pka’s, solubility product (ksp) and crystallinity might not be implemented properly in drug discovery and development processes, even though it was well recognised that the properties of apis play important roles in oral drug absorption (i.e., dissolution, permeation and even precipitation). if this is the case, it may lead to poor productivity in drug discovery and development. before starting pcf-j, some scientists had participated in the physchem forum meetings in europe and the international association of physical chemists (iapc) meetings. they acknowledged its values, for example, in cutting-edge and multidisciplinary topics, energetic scientific discussion by diverse participants from many countries and friendly networking [1,2]. considering the practices of physchem forum, they decided to set up a forum for pharmaceutical physical-chemical scientists in japan. the original committee members organised the 1st physchem forum japan (pcf-j) meeting on 25th feb. 2014 (a permission of the use of “physchem forum” was granted by the ‘european’ physchem forum in advance). pcf-j meetings were then held on an annual basis. even during the covid-19 pandemic, the meetings were successfully held online in 2020 and 2021, with more than one hundred participants from industry, academia (faculties and students) and regulatory bodies (table 1). https://doi.org/10.5599/admet.1452 https://doi.org/10.5599/admet.1452 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:mano@ono.co.jp http://creativecommons.org/licenses/by/4.0/ takashi mano admet & dmpk 10(4) (2022) 245-247 246 the lead committee member of each pcf-j meeting upheld each theme, which tended to be dependent on the interests of the leader, and the pcf-j committee members constructed the programs. applying the best practice of the iapc meetings, the committee organized the growth project and invited several oversea lectures for the 4th pcf-j meeting. dr. konstantin tsinman was one of the lecturers that accepted our offer and volunteered to give lectures [3]. konstantin gave an excellent talk on permeability, explaining the basics and applications of the theories of drug permeation in the gastrointestinal tract, including the unstirred water layer over the mucus layer. his talk extended to using the parallel artificial membrane permeability assay (pampa) system and its application to improve the accuracy of permeability assessment. konstantin included quizzes within his talk to arouse the audience’s curiosity, which created a harmonious atmosphere in the growth project session. the contribution of konstantin and other speakers from countries outside of japan created the international environment in pcf-j, and the 6th meeting was held as a joint meeting with iapc as “joint 7th iapc & 6th pcf-j meeting” in 2018 in osaka, japan [4]. this year (2022), the pcf-j 10th meeting is scheduled for december 8th and is planned to be held online. the committee is expecting as many as 100 participants.5 pcf-j has offered opportunities where pharmaceutical physical chemists meet in a multidisciplinary manner, and at the same time, new sciences and technologies meet to create new spaces for advancing theories and developing new methodologies. pcf-j has also helped many scientists grow. in its course, konstantin’s significant contributions should not be forgotten. also, we cannot forget his friendly and warm personality. we are so lucky that konstantin’s paths and ours crossed and that we could share those moments with him. the committee members of physchem forum-japan continue to hold the forum meetings, in one which konstantin’s attendance was of great importance. his overall contributions greatly influenced all pcf-j meetings in the past and will continue in the future. table 1. physchem forum-japan meetings. conflict of interest: the author declares no conflicts of interest associated with this manuscript. references [1] physchem forum for physical chemists by physical chemists <http://www.physchem.org.uk/> (27th june, 2022) no. date theme venue/location 1 25th feb. 2014 supersaturation hoshi uni./tokyo 2 11th dec. 2014 from physical chemical property parameters to dissolution the mishima chamber of commerce and industry/shizuoka 3 25th nov. 2015 api design: api optimisation in drug discovery ono pharmaceutical/osaka 4 9th dec. 2016 log p world: drug discovery from pharmaceutical molecular property perspectives asahi kasei/tokyo 5 1st-2nd dec. 2017 pat practices the mishima chamber of commerce and industry/shizuoka 6 28th-30th aug. 2018 iapc-7/pcfj-6 joint international symposium ritsumei uni./osaka 7 12th-13th nov. 2019 application of mechanical-based simulation in cmc/pharmaceutical sciences eisai/ibaraki 8 4th dec. 2020 application of crystallisation and crystal structure analysis in drug discovery online 9 9th-10th dec. 2021 nano: drug substance or dds, this is the question online 10 8th dec. 2022 ivivc online http://www.physchem.org.uk/ admet & dmpk 10(4) (2022) 245-247 ten years for physchem forum-japan doi: https://doi.org/10.5599/admet.1452 247 [2] international association of physical chemists <https://www.iapchem.org/> (27th june, 2022) [3] itochu techno-solutions corporation <https://ls.ctc-g.co.jp/event/report_20170411.html> (27th june, 2022, japanese only) [4] iapc-7 meeting <https://www.iapchem.org/index.php/iapc-7-home> (27th june, 2022) [5] pjf-j <https://www.c-sqr.net/home> (27th june, 2022, japanese only) ©2022 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) https://doi.org/10.5599/admet.1452 https://www.iapchem.org/ https://ls.ctc-g.co.jp/event/report_20170411.html https://www.iapchem.org/index.php/iapc-7-home https://www.c-sqr.net/home http://creativecommons.org/licenses/by/3.0/) new therapeutic modalities in drug discovery and development: insights & opportunities doi: http://doi.org/10.5599/admet.1209 227 admet & dmpk 9(4) (2021) 227-230; doi: http://dx.doi.org/10.5599/admet.1209 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index editorial new therapeutic modalities in drug discovery and development: insights & opportunities manfred kansy 1 and giulia caron 2 1 independent consultant, 79111, freiburg im breisgau, germany (manfred.kansy1@gmail.com) 2 department of molecular biotechnology and health sciences, university of torino, quarello 15, 10135 torino, italy (giulia.caron@unito.it) received: december 11, 2021; revised: december 13, 2021; published: december 15, 2021 the study of the statistics of new drug registrations over time shows an interesting pattern. at the end of the last century, productivity, measured in terms of market launches, was declining and charts as the one shown in figure 1 have often been used to highlight potential productivity losses in pharmaceutical r&d [1,2]. this trend seems to have reversed in the last decade. certainly, the number of launches alone is not a good measure of productivity and other, partly controversially discussed measures have been applied [3-5]. figure 1. visualization of cder new molecular entity (nme) drug and new biologic approvals 1985-2021 [6]. what might be the reason for this potentially positive trend of increasing numbers of nmes and biologics approvals over the last ten years? it is obvious that massive efforts in strategy adjustments, early admet optimization and the introduction of new assays, methods and computational techniques by pharmaceutical r & d and academia have had an effect. new disease-relevant targets have been identified and potential improvements in regulatory processes could have been beneficial. moreover, the introduction of new therapeutic http://dx.doi.org/10.5599/admet.1209 http://dx.doi.org/10.5599/admet.1209 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:manfred.kansy1@gmail.com mailto:giulia.caron@unito.it m. kansy, g. caron admet & dmpk 9(4) (2021) 227-230 228 modalities, besides the so-called classical rule of 5 (ro5) [7] compliant molecules, has had some impact. antibodies, antibody-drug conjugates, fusion proteins, neurotoxins, peptides, polymers, small interfering rna, vaccines, etc., are reaching the market. furthermore, the introduction of the so-called "beyond rule of five compounds" (bro5) indicates that even for some demanding targets larger compounds (mw >> 500) can be designed and optimized to have adequate admet profiles and resulting bioavailability with desired in vivo effect. although general strategies to design oral bioavailable bro5 drug candidates have not yet been reported, conformational flexibility has been identified as an important parameter to describe the adme profile of large and flexible derivatives. for example, it has been recently shown that flexibility descriptors helped to improve computational solubility prediction of bro5 compounds [8]. moreover, many bro5 compounds seem to have an inherent potential to conformationally adjust to the surrounding media and behave like molecular chameleons. for instance, the formation of intramolecular hydrogen bonds in an environment-dependent way has been analyzed to explain the passive permeability skills of some new therapeutic modalities [9,10]. certainly, small molecules still make up the majority of new drug market introductions, and their urgent need in the treatment of newly discovered diseases and the improvement of current therapies is impressively demonstrated during the severe ongoing covid 19 pandemic. this special issue of admet and dmpk describes new therapeutic modalities, innovative compound characterization tools for drug safety evaluation/optimization and newly developed computational prediction tools for estimating drug brain uptake as examples of how the drug discovery and research fields are continuously innovating and adapting to the changing needs. in a first impressive work, experts in the field of degraders, i.e., bifunctional compounds that promote degradation of target molecules instead of inhibiting them, also called protacs (proteolysis targeting chimeras), have analyzed degrader data in terms of target distribution [11]. moreover, they impressively demonstrate how such large and partially flexible bro5 drug molecules can be engineered in a more rational way to improve admet properties. this excellent work by giuseppe ermondi and his colleagues at the university of turin may be of even greater interest, as preliminary results from a recent phase 1 study show that the degrader principle appears to work in humans [12,13]. colleagues from the russian academy of sciences in saratov describe how selenium nanoparticles (senps) can be used to increase the bioavailability of silymarin [14], a drug with discussed liver-protective and anti-cancer activity [15] but unfavorable physicochemical properties. by combining the anti-cancer activity of senps with the effect of poorly soluble silymarin, the in vitro activity could be increased and the side effects of se could be partially alleviated. this can be considered a good example of how the therapeutic effects and side effects of np can be potentially controlled by combining drugs in nps in the future. in addition to new therapeutic principles, innovative new methods for testing and optimizing drug safety are urgently needed. in a fascinating mini-review [16], a joint team of experts from academia in australia and the pharmaceutical industry in switzerland described how 3d bioprinted heart cells could be strategically used in pharmaceutical research and development and potentially improve and streamline regulatory processes. finally, a team of experts from the universities of naples and edinburgh describes the development of in silico prediction models based on immobilized artificial membrane (iam) chromatography measurements and predicted drug properties to estimate the blood-brain barrier passage of drugs [17]. although the admet & dmpk 9(4) (2021) 227-230 editorial: new therapeutic modalities in ddd doi: http://doi.org/10.5599/admet.1209 229 method is premature, due to the currently limited amount of data, it might be applied in drug discovery and development in selecting appropriate drug candidates with improved brain penetration properties in the future. overall, this special issue of admet and dmpk provides interesting insights into selected innovative areas of drug discovery and development. references [1] i.m. cockburn. is the pharmaceutical industry in a productivity crisis?. innovation policy and the economy 7 (2006) 1–32. https://doi.org/10.1086/ipe.7.25056188. [2] a. schuhmacher, o. grassmann, m. hinder. changing r&d models in research-based pharmaceutical companies. journal of translational medicine 14 (2016)105-116. https://doi.org/10.1186/s12967016-0838-4. [3] f. pamolli, l. magazzini, m. riccaboni. the productivity crisis in pharma r &d. nature reviews in drug discovery 10 (2011) 428-438. https://doi.org/10.1038/nrd3405. [4] y. feyman. is pharmaceutical productivity in decline? maybe not. forbes. april 13, 2015. https://www.forbes.com/sites/theapothecary/2015/04/13/is-pharmaceutical-productivity-indecline-maybe-not/. [5] k.j. hult. economics, 2015. incremental innovation and pharmaceutical productivity. http://terramedica.hsinetwork.com/hult%20%20incremental%20innovation%20and%20pharmaceutical%20productivity.pdf. [6] u.s.food and drug administration website. https://www.fda.gov/drugs/drug-approvals-anddatabases/compilation-cder-new-molecular-entity-nme-drug-and-new-biologic-approvals (status 29/11/2021). [7] c. lipinski, f. lombardo, b. dominy, p. feeney. experimental and computational approaches to estimate solubility and permeability in drug discovery and development settings. advances in drug delivery reviews 23 (1997) 3-25. https://doi.org/10.1016/s0169-409x(00)00129-0. [8] a. avdeef, m. kansy. “flexible-acceptor” general solubility equation for beyond rule of 5 drugs. molecular pharmaceutics 17 (2020) 3930–3940. https://doi.org/10.1021/acs.molpharmaceut.0c00689. [9] m. rossi sebastiano, b.c. doak, m. backlund, v. poongavanam, b. over, g. ermondi, g. caron, p. matsson, j. kihlberg. impact of dynamically exposed polarity on permeability and solubility of chameleonic drugs beyond the rule of 5. journal of medicinal chemistry 61 (2018) 4189–4202. https://doi.org/10.1021/acs.jmedchem.8b00347. [10] g. caron, g. ermondi. molecular descriptors for polarity: the need for going beyond polar surface area. future medicinal chemistry 8 (2016) 2013-2016. https://doi.org/10.4155/fmc-2016-0165. [11] d.g. jimenez, m.r. sebastiano, g. caron, g. ermondi. are we ready to design oral protacs®?. admet and dmpk 9 (2021) 243-254. https://doi.org/10.5599/admet.1037. [12] a. mullard. targeted degraders clear first safety hurdles. nature reviews drug discovery 19 (2020) 435. https://doi.org/10.1038/d41573-020-00109-w. [13] a. mullard. targeted protein degraders crowd into clinic. nature reviews drug discovery 20 (2020) 247-250. https://doi.org/10.1038/d41573-021-00052-4. [14] s. staroverov, s. kozlov, a. fomin, k. gabalov, v. khanadeev, d. soldatov, i. domnitsky, l. dykman, s.v. akchurin, o. guliy. synthesis of silymarin−selenium nanoparticle conjugate and examination of its biological activity in vitro. admet and dmpk 9 (2021) 255-266. https://doi.org/10.5599/admet.1023. [15] d-h. won, l-h. kim, b. jang, i-h. yang, h-j. kwon, b. jin, s. h. oh, j-h. kang, s-d. hong, j-a. shin, s-d. cho. in vitro and in vivo anti-cancer activity of silymarin on oral cancer. tumor biology 40 (2018) 111. https://doi.org/10.1177/1010428318776170. http://dx.doi.org/10.5599/admet.1209 https://doi.org/10.1086/ipe.7.25056188 https://doi.org/10.1186/s12967-016-0838-4 https://doi.org/10.1186/s12967-016-0838-4 https://doi.org/10.1038/nrd3405 https://www.forbes.com/sites/theapothecary/2015/04/13/is-pharmaceutical-productivity-in-decline-maybe-not/ https://www.forbes.com/sites/theapothecary/2015/04/13/is-pharmaceutical-productivity-in-decline-maybe-not/ http://terramedica.hsinetwork.com/hult%20-%20incremental%20innovation%20and%20pharmaceutical%20productivity.pdf http://terramedica.hsinetwork.com/hult%20-%20incremental%20innovation%20and%20pharmaceutical%20productivity.pdf https://www.fda.gov/drugs/drug-approvals-and-databases/compilation-cder-new-molecular-entity-nme-drug-and-new-biologic-approvals https://www.fda.gov/drugs/drug-approvals-and-databases/compilation-cder-new-molecular-entity-nme-drug-and-new-biologic-approvals https://doi.org/10.1016/s0169-409x(00)00129-0 https://doi.org/10.1021/acs.molpharmaceut.0c00689 https://doi.org/10.1021/acs.molpharmaceut.0c00689 https://doi.org/10.1021/acs.jmedchem.8b00347 https://doi.org/10.4155/fmc-2016-0165 https://doi.org/10.5599/admet.1037 https://doi.org/10.1038/d41573-020-00109-w https://doi.org/10.1038/d41573-021-00052-4 https://doi.org/10.5599/admet.1023 https://doi.org/10.5599/admet.1023 https://doi.org/10.1177/1010428318776170 m. kansy, g. caron admet & dmpk 9(4) (2021) 227-230 230 [16] l. polonchuk, c. gentile. current state and future of 3d bioprinted models for cardio-vascular research and drug development. admet and dmpk 9 (2021) 231-242. https://doi.org/10.5599/admet.951. [17] l. grumetto, g. russo. cδlog kwiam: can we afford estimation of small molecules’ blood-brain barrier passage based upon in silico phospholipophilicity?. admet and dmpk 9 (2021) 267-281. https://doi.org/10.5599/admet.1034. ©2021 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/) https://doi.org/10.5599/admet.951 https://doi.org/10.5599/admet.951 https://doi.org/10.5599/admet.1034 http://creativecommons.org/licenses/by/4.0/ manuscript doi: 10.5599/admet.3.1.142 77 admet & dmpk 3(1) (2015) 77-83; doi: 10.5599/admet.3.1.142 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index note differential effects of mitogen-activated protein kinase pathway inhibitors on p-glycoprotein activation hirokazu wakuda 1, *, shino miyauchi 1 , kana maruyama 1 , satomi kagota 1 , kazuki nakamura 1 , keizo umegaki 2 , shizuo yamada 3 and kazumasa shinozuka 1 1 department of pharmacology, school of pharmacy and pharmaceutical sciences, mukogawa women’s university, 1168 koshien, kyuban-cho, nishinomiya 663-8179, japan 2 information center, national institute of health and nutrition, 1-23-1 toyama, shinjuku-ku, tokyo 162-8636, japan 3 center for pharma-food research (cpfr), graduate school of pharmaceutical sciences, university of shizuoka, 52-1 yada, suruga-ku, shizuoka, shizuoka 422-8526, japan *corresponding author: e-mail: wakuda@mukogawa-u.ac.jp; tel.: +81-798-45-9944; fax: +81-798-45-9944 received: november 12, 2014; revised: march 17, 2015; published: march 31, 2015 abstract the aim of this study was to evaluate the effects of the mitogen-activated protein kinase (mapk) pathway inhibitors sb203580, cmpd-1, sb239063, sp600125, and fr180204 on the activity of p-glycoprotein (p-gp) and to assess whether the mapk pathway affects p-gp directly. changes in the fluorescence of residual rhodamine 123, a marker of p-gp activity, in the apical region of caco-2 cells were measured in the presence of mapk pathway inhibitors using time-lapse confocal laser scanning microscopy at 0, 10, 20, 30, and 60 min. significant differences were observed between the fluorescence levels of control cells and cells treated with sb203580 for 20, 30, or 60 min. however, no significant change was observed in the residual rhodamine 123 fluorescence of cells treated with cmpd-1, sb239063, sp600125, or fr180204. among the p38-mapk pathway inhibitors investigated, cmpd-1 and sb239063 showed no detectable effect on the activity of p-gp. further, jnk 1, 2, 3-mapk pathway (sp600125) and erk1/2 pathway (fr180204) inhibitors did not affect p-gp activity. however, sb203580 enhanced the transfer of rhodamine 123 across the apical cell membrane. thus, sb203580 activated p-gp, although not through the p38-mapk pathway. importantly, the mapk pathway did not affect p-gp activity shortly after treatment. keywords confocal laser scanning microscopy; mapk; sb203580 introduction p-glycoprotein (p-gp) is the most thoroughly studied member of the adenosine triphosphate-binding cassette transporter superfamily and is expressed throughout the intestinal epithelium, hepatocytes, renal tubular cells, and the blood-brain barrier [1]. p-gp is encoded by the abcb1/mdr1 gene in humans [2]. because of its ubiquitous expression and broad specificity, changes in p-gp expression or efflux activity induced by drug treatments, diet, environmental factors, or single nucleotide polymorphisms (snps) can greatly affect drug disposition, pharmacokinetics, and clinical response [3]. hence, identification of p-gp http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:wakuda@mukogawa-u.ac.jp wakuda et al. admet & dmpk 3(1) (2015) 77-83 78 figure 1. the mapk pathway substrates is important for therapeutic optimization and to avoid drug-drug interactions. however, little is known of p-gp regulation. mitogen-activated protein kinases (mapks) are a family of ser/thr protein kinases that are widely conserved among eukaryotes and are involved in many cellular functions. the p38-mapk and c-jun nterminal kinase (jnk)-mapk signalling pathways allow cells to interpret a wide range of external signals and respond appropriately by generating a plethora of different biological effects. there are four p38 subtypes (p38α, p38β, p38γ, p38δ) and three jnk subtypes (jnk1, jnk2, jnk3) [4,5]. the extracellular signalregulated kinase (erk1/2)-mapk signalling pathway can be activated in response to a diverse range of extracellular stimuli, including mitogens, growth factors, and cytokines (fig. 1) [6]. several studies have indicated that mapk signalling pathways may regulate membrane trafficking proteins. for example, fujishiro et al. reported that activation of the p38-mapk pathway is not necessary for insulin-induced glucose uptake but that it regulates glucose transporter expression [7]. nagelin et al. reported that murine 12/15-lipoxygenase regulates the expression and function of the atp-binding cassette transporter g1 through p38and jnk2-dependent pathways [8]. these reports suggest that the mapk pathway transmits signals to membrane trafficking proteins to regulate their activity. in the present study, we evaluated the effects of mapk pathway inhibitors (sb203580, competitive p38α and β inhibitor; cmpd-1, non-competitive p38α inhibitor; sb239063, competitive p38α and β specific inhibitor; sp600125, competitive jnk 1, 2, 3 inhibitor; fr180204, competitive erk1/2 inhibitor) on the admet & dmpk 3(1) (2015) 77-83 mitogen-activated protein kinase pathway inhibitors doi: 10.5599/admet.3.1.142 79 activity of p-gp using time-lapse confocal laser scanning microscopy. the confocal laser scanning microscopy-based assay was used to measure residual rhodamine 123 in the apical region of caco-2 cells; this method is more sensitive than conventional methods that rely on microplate readers and fluorescence microscopy [9,10]. experimental materials sb203580 was obtained from invivogen (san diego, ca, u.s.a.). cmpd-1 was obtained from tocris bioscience (ellisville, mo, u.s.a.). sb239063 was obtained from sigma-aldrich (st. louis, mo, u.s.a.). sp600125 was obtained from cayman chemical (ann arbor, mi, u.s.a.). fr180204 was obtained from santa cruz biotechnology (santa cruz, ca, u.s.a.). rhodamine 123 was obtained from molecular probes (eugene, or, u.s.a.). cell culture of caco-2 cells the human colon adenocarcinoma cell line (caco-2) was purchased from ds pharma biomedical (osaka, japan). the cells were used for experiments 7 d after seeding and between passages 47 and 52. figure 2. the protocol for measurement of p-glycoprotein-mediated transport activity using confocal laser-scanning microscopy in the presence or absence of inhibitor agents. quantification of rhodamine 123 by time-lapse confocal laser scanning microscopy transport measurements were performed in a transwell chamber (bd biosciences, san jose, ca, u.s.a.). caco-2 cells were seeded at a density of 3 × 104 cells per filter on polyethylene terephthalate filters in the cell culture inserts. the cell culture inserts were placed on a glass-bottomed dish, and the caco-2 cells were incubated with 5 µm rhodamine 123 for 90 min at 37°c. the cells were rinsed with hanks’ balanced salt solution (hbss). hbss with or without mapk pathway inhibitors was added to the cells (sb203580, 10 µm; cmpd-1, 10 µm; sb239063, 100 nm; sp600125, 100 nm; fr180204, 1 µm), and cells were observed by time-lapse confocal laser scanning microscopy (fig. 2). quantification of the intracellular concentration of rhodamine 123 in the wakuda et al. admet & dmpk 3(1) (2015) 77-83 80 caco-2 cells was performed using fluorescent images at several time points (0, 10, 20, 30, and 60 min) by a chronological measurement program (ez-c1; nikon, tokyo, japan). the fluorescence intensity at 0 min was defined as 100 % [9,10]. statistical analysis all values represent the mean ± standard error. the data were analysed using a two-way anova followed by a bonferroni post-test to determine significance (p < 0.05). statistical analyses were performed using the graphpad prism 4.03 software (graphpad software, la jolla, ca, u.s.a.). results and discussion the level of residual rhodamine 123 fluorescence in control cells and cells treated with mapk pathway inhibitors were measured with time-lapse confocal laser scanning microscopy to assess p-gp activity. treatment with 10 μm sb203580, a competitive inhibitor of the p38α and β atp binding pocket, enhanced the fluorescent decrease significantly at 20, 30, and 60 min (fig. 3a). in contrast, the decrease in fluorescent intensity was not affected by treatment with the jnk 1, 2, 3 inhibitor, sp600125 (100 nm), or the erk1/2 inhibitor fr180204 (1 μm) (fig. 3b, c). further, the decrease in fluorescent intensity was not affected by the non-competitive p38α inhibitor cmpd-1 (10 μm) (fig. 3d) or a second, more specific competitive inhibitor of the p38α and β atp binding pocket, sb239063 (100 nm) (fig. 3e). we evaluated the effects of mapk pathway inhibitors on the activity of p-gp, using time-lapse confocal laser scanning microscopy. sb203580 enhanced the fluorescent decrease significantly after 20 min. it is hypothesized that the enhanced effect of sb203580 is due to changes in p-gp activation, as changes to the p-gp expression level would require at least 24 h of drug treatment. sb203580 is a pyridinyl imidazole that competitively inhibits atp binding and is widely used in studies for elucidating the roles of p38-mapk [11]. barancík et al. reported that treatment of the multidrug resistant mouse leukaemia cell line l1210/vcr with sb203580 (30 µm) for 3 days inhibited p-gp [12]. in contrast, our results suggest that treatment of the human colon adenocarcinoma cell line caco-2 with sb203580 (10 µm) for 60 min activates p-gp. the difference in effect may be due to differences in cell lines, sb203580 concentration, and duration of treatment. furthermore, the decrease in fluorescent intensity was not affected by sp600125 and fr180204. sp600125 is a jnk 1, 2, 3 inhibitor and fr180204 is an erk1/2 inhibitor. our results suggest that the jnkmapk and erk1/2-mapk pathways do not affect p-gp activity directly or over the time course studied. supporting our data, kim et al. have also reported that sp600125 had no effect on p-gp [13]. the enhancement of p-gp activity following sb203580 treatment suggested that activation of the p38mapk pathway might inhibit p-gp. as mentioned earlier, sb203580 competitively inhibits atp binding to p38 α and β. further, p-gp is an atp-dependent transporter. thus, we hypothesized that inhibition of atp binding to p38α and β by sb203580 might enhance available atp levels, thereby enhancing the activation of p-gp. to further assess the contribution of p38-mapk pathways, cells were treated with either cmpd-1 or sb239063. cmpd-1 is a selective, non-competitive inhibitor of p38α, and sb239063 is a competitive inhibitor of p38-mapk [14,15]. importantly, sb239063 is more specific than sb203580 [15]. interestingly, both cmpd-1 and sb239063 had no detectable effect on the activity of p-gp. thus, our results suggest that admet & dmpk 3(1) (2015) 77-83 mitogen-activated protein kinase pathway inhibitors doi: 10.5599/admet.3.1.142 81 the p38-mapk pathway does not affect p-gp directly after a short duration of treatment. however, sb203580 increased the transfer of rhodamine 123 across the cell membrane after 20 min. figure 3. influence of mapk pathway inhibitors on the time-dependent decrease in residual rhodamine 123 fluorescence in the apical region of caco-2 cells, measured by time-lapse confocal laser scanning microscopy. each chamber represents a p38-mapk pathway inhibitor: (a) sb203580, 10 µm: (b) sp600125, 100 nm; (c) fr180204, 1 µm: (d) cmpd-1, 10 µm: (e) sb239063, 100 nm. each point and bar represents the mean ± standard error. n = 12-18. *,*** significant difference between the control group and sb203580 group (p < 0.05, 0.001). there are two possible reasons for the enhanced rhodamine transport. firstly, the effect of sb203580 could be mediated through pathways other than the p38-mapk pathway. the working concentration of sb203580 reported in the literature is 1-20 µm. however, at high concentrations, sb203580 can also affect raf-1, phosphoinositide-dependent protein kinase 1 (pdk1), the transforming growth factor-beta (tgf-β) receptor, cyclooxygenase 1 (cox-1), and cyclooxygenase 2 (cox-2) [16]. additionally, lali et al. reported that low concentrations of sb203580 can inhibit the phosphorylation and activation of protein kinase b wakuda et al. admet & dmpk 3(1) (2015) 77-83 82 (pkb), also known as akt, by inhibiting the pkb kinase, pdk1 (ic50, 3-5 µm) [17]. it is possible that the activation of p-gp by sb203580 may occur via one of these pathways. alternatively, sb203580 could cause a structural change in p-gp. p-gp is one of the best-characterized abc transporters; however, the mechanism of substrate transport is unclear. in particular, it has been hypothesized that there are many possible drug-binding sites on p-gp, including up to seven putative binding sites located in the transmembrane domain of the protein [18-20]. it is generally accepted that transmembrane helices 5, 6, 11, and 12 are involved in p-gp substrate binding. one popular model published by shapiro et al. suggested that two different functional binding sites, the r-site and h-site, interact in a positive cooperative manner.18 rhodamine 123 and anthracylines are substrates of the r-site, whereas colchicine is a substrate of the h-site of p-gp. this two-site hypothesis is the most convenient working model that explains the mutual stimulation of p-gp-mediated transport by several substrates. additionally, there is evidence of a third allosteric binding site that serves a regulatory function and may be the site of progesterone binding [21]. sterz et al. reported that some substances are activators of p-gp [22]. further, kerboeuf et al. reported that macrocyclic lactone anthelmintics activate p-gp in nematodes and suggested that several substituents in the macrocyclic lactone structure are involved in modulating p-gp activation [23]. conclusions the mapk pathway does not affect short-term p-gp activation. rather, activation of p-gp by sb203580 is likely due to either pathways other than the p38-mapk pathway or allosteric structural changes in p-gp. thus, further studies should be performed to understand the mechanism underlying sb203580-mediated p-gp activation. acknowledgements: the authors declare that they have no conflicts of interest to disclose and funding. the authors are grateful to prof. yoshiyuki kagawa, prof. junko sugatani and prof. hiroshi yamada for their suggestions. we also thank our colleagues for their helpful advice and support. references [1] f. thiebaut, t. tsuruo, h. hamada, mm. gottesman, i. pastan, mc. willingham, proc. natl. acad. sci. usa. 84 (1987) 7735-7738. [2] vj. wacher, cy. wu, lz. benet, mol. carcinog. 13 (2000) 129-134. [3] m. hitzl, s. drescher, h. van der kuip, e. schäffeler, j. fischer, m. schwab, m. eichelbaum, mf. fromm, pharmacogenetics 11 (2001) 293-298. [4] a. cuadrado, ar nebreda, biochem. j. 429 (2010) 403-417. [5] ma. bogoyevitch, kr. ngoei, tt. zhao, yy. yeap, dc. ng, biochim. biophys. acta. 1804 (2010) 463475. [6] s. nishimoto, e. nishida, embo rep. 7 (2006) 782-786. [7] m. fujishiro, y. gotoh, h. katagiri, h. sakoda, t. ogihara, m. anai, y. onishi, h. ono, m. funaki, k. inukai, y. fukushima, m. kikuchi, y. oka, t. asano, j. biol. chem. 276 (2001) 19800-19806. [8] mh. nagelin, s. srinivasan, jl. nadler, cc. hedrick, j. biol. chem. 284 (2009) 31303-31314. [9] h. wakuda, n. nejime, y. tada, s. kagota, k. umegaki, s. yamada, k. shinozuka, biol. pharma. bull. 33 (2010) 1238-1241. [10] h. wakuda, n. nejime, y. tada, s. kagota, oa. fahmi, k. umegaki, s. yamada, k. shinozuka. j. pharm. pharmacol. 63 (2011) 1015-1021. admet & dmpk 3(1) (2015) 77-83 mitogen-activated protein kinase pathway inhibitors doi: 10.5599/admet.3.1.142 83 [11] a. cuenda, j. rouse, yn. doza, r. meier, p. cohen, tf. gallagher, pr. young, jc. lee, febs lett. 364 (1995) 229-233. [12] m. barancík, v. bohácová, j. kvackajová, s. hudecová, o. krizanová, a. breier, eur. j. pharm. sci. 14 (2001) 29-36. [13] jh. kim, m. chae, ar. choi, h. sik kim, s. yoon, eur. j. pharmacol. 723 (2014) 141-147. [14] w. davidson, l. frego, gw. peet, rr. kroe, me. labadia, sm. lukas, rj. snow, s. jakes, ca. grygon, c. pargellis, bg. werneburg, biochemistry 43 (2004) 11658-11671. [15] dc. underwood, rr. osborn, cj. kotzer, jl. adams, jc. lee, ef. webb, dc. carpenter, s. bochnowicz, hc. thomas, dw. hay, de. griswold, j. pharmacol. exp. ther. 293 (2000) 281-288. [16] y. zhang y, j. zhou, w. xu, a. li, j. zhou, s. xu, j. toxicol. environ. health. a. 72 (2009) 774-781. [17] fv. lali, ae. hunt, sj. turner, bm. foxwell, j. biol. chem. 275 (2000) 7395-7402. [18] ab. shapiro, v. ling, biochem. pharmacol. 53 (1997) 587-596. [19] c. martin, g. berridge, cf. higgins, p. mistry, p. charlton, r. callaghan, mol. pharmacol. 58 (2000) 624-632. [20] ar. safa, curr. med. chem. anticancer agents. 4 (2004) 1-17. [21] ab. shapiro, k. fox, p. lam, v. ling, eur. j. biochem. 259 (1999) 841-850. [22] k. sterz, l. möllmann, a. jacobs, d. baumert, m. wiese, chemmedchem 4 (2009) 1897-1911. [23] d. kerboeuf, f. guégnard, antimicrob. agents chemother. 55 (2011) 2224-2232. ©2015 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ manuscript doi: 10.5599/admet.3.1.161 45 admet & dmpk 3(1) (2015) 45-50; doi: 10.5599/admet.3.1.161 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper effect of bioenhancers on amoxicillin bioavailability kalyani barve*, kushal ruparel spp school of pharmacy and technology management, svkm’s nmims, vile parle (west), mumbai -400056, india. *corresponding author: e-mail: barve.kalyani@gmail.com; tel.: +91-22-42332000; fax: +91-22-26185422 received: january 29, 2015; revised: february 24, 2015; published: march 31, 2015 abstract amoxicillin, which inspite of being effective is losing its importance due to less bioavailability. bioavailability could be enhanced by combining the antibiotics with bioenhancers like piperine and ginger resin. the present abstract deals with the use of piperine and ginger resin in increasing the bioavailability of amoxicillin using the insitu spip method. piperine, isolated was found to be 99 % pure and ginger resin contained 4.36 % w/w of gingerol. the absorption of amoxicillin increased with addition of piperine (1.2, 1.5, 1.8 % w/w of amoxicillin) in a dose dependant manner reaching saturation after 45 minutes wheras ginger resin (6, 7, 8 % w/w of amoxicillin) failed to show this effect in the invitro studies. piperine enhanced the permeation of amoxicillin but ginger resin failed to show this effect using the insitu spip method. the method used may be appropriate for bioenhancers acting on transporters, metabolizing enzymes or modulators of diffusivity but not for those acting on gut motility. the combination of piperine and amoxicillin trihydarte may have a potential in reducing the dose, shortening the treatment period thus reducing drug-resistance problems of the antibiotic. keywords piperine; gingerol; bcs classification; black pepper; ginger; ayurveda introduction three major factors namely solubility, dissolution, and intestinal permeability, affecting oral drug absorption can be estimated using the biopharmaceutics classification system (bcs). it classifies the drug into four classes: class i (high solubility, high permeability), class ii (low solubility, high permeability), class iii (high solubility, low permeability) and class iv (low solubility, low permeability). some of the very commonly used antibiotics fall into class iii and class iv category according to this system. one such antibiotic is amoxicillin which has a low permeability [1] and has been classified into class iii drugs [2]. amoxicillin is absorbed by passive diffusion and via active transport through oligopeptide carrier localized at the apical enterocyte membrane (pept1). the drug is reported to have low bioavailability in rats [3] as well as in humans [4]. in spite of being an effective antibiotic, it is losing importance in chemotherapy due to less bioavailability. the permeability of such low permeability drugs can be enhanced by co administration of bioenhancers. http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:barve.kalyani@gmail.com barve and ruparel admet & dmpk 3(1) (2015) 45-50 46 bioenhancers are agents, which by themselves are not therapeutic entities but when combined with an active drug lead to the potentiation of the pharmacological effect of the drug. such enhancers have been found to increase the bioavailability of a number of drugs at very low concentrations. in general, they can enhance the bioavailability by enhancing the passive diffusion or by inhibiting the efflux transporters [5] (pgp) and other multidrug resistance proteins (mrps) [6]. piperine and ginger resin are known to be the constituents of trikatu, an ayurvedic formulation, indicating three drugs which enhance the bioavailability [7]. in the present study the effect of piperine and ginger resin on the absorption of amoxicillin has been studied using in vitro and in vivo models. experimental plant material: black pepper and ginger was procured from the local market and subjected to extraction of the active constituents. chemicals: amoxicillin trihydrate, atenolol and propranolol hydrochloride were obtained as gift samples from ipca laboratories, mumbai. urethane was procured from sigma aldrich, india and other chemicals and solvents were of analytical grade and procured from commercial sources. isolation of piperine [8]: 300 g of black pepper powder was extracted with 1.5 liters of 95 % ethanol in a soxhlet extractor at 60-70 °c. the solution was filtered and concentrated in vacuum on a water bath at 60 °c. alcoholic potassium hydroxide (10 % w/v) was added to the filtrate residue and kept overnight in a refrigerator at 2-8 °c. the yellow crystals of piperine were deposited at the bottom of the flask which were then washed and purified with organic solvents. recrystallisation was done with acetone. piperine crystals were stored for further use. analysis of isolated piperine: the isolated piperine was subjected to hplc analysis using microsorb-mv 100-5 c18 column, acetonitrile: 0.1 % orthophosphoric acid (60:40) as mobile phase, uv –visible detector at 340 nm at ambient temperature and a flow rate of 1 ml/min. preparation of ginger resin: 500 g of ginger powder was extracted with 2 l of 95 % ethanol in a soxhlet extractor at 60-70 °c. the extract was evaporated by distillation and collection of the excess filtered solvent was carried out. the extract was further concentrated on a rotary evaporator. ginger resin was stored for further use. quantitative estimation of ginger resin: ginger resin sample was sent to natural remedies pvt. ltd. (bangalore) for estimation of gingerol content by hptlc. in vitro studies using franz diffusion cell: amoxicillin solution (10 mg/ml) was added to the reservoirs with phosphate buffer solution (pbs) ph 7.4 as the medium in the diffusion cell. varying concentrations of piperine (1.2, 1.5 and 1.8 % w/w) [9] and ginger resin (6, 7 and 8 % w/w) [9,10] were added to different reservoirs to check the enhancement in diffusion across the synthetic semi permeable membrane. invitro studies using non everted gut sac method [3]: the pig intestine was procured from deonar slaughter house, mumbai, india. it was cleaned and sacs of about 3 cm in length were made from the ileum. the empty sac was filled with 1 ml of amoxicillin (6 mg/ml) in pbs (ph 7.4) using a blunted syringe. the needle was then slipped off carefully, and the loose ligature on the proximal end was tightened. the serosal compartment contained buffer in the sac. the distended sac was placed inside organ tube of organ bath containing 50 ml pbs. this gut sac bath was surrounded by a water jacket maintained at 37 ± 0.5 °c. the mucosal fluid compartment content was continuously mixed with air bubbles using an aerator. at predetermined intervals, 5 ml of sample was withdrawn from the organ tube and same volume was admet & dmpk 3(1) (2015) 45-50 bioenhancement of amoxicillin by piperine doi: 10.5599/admet.3.1.161 47 replaced with fresh buffer. the concentration of drug that traversed intestinal surface was monitored at 273 nm spectrophotometrically. experiment was repeated with 1.5 % w/w of piperine and 8 % w/w of ginger resin added to the sac. in vivo studies using single pass intestinal perfusion method (spip) [11]: the intestinal transport of amoxicillin trihydrate was investigated using spip technique. male wistar rats (200-250 g) were procured from bharat serum, thane. animals were acclimatized for a week before use in the spip study. all animal experiments were carried out in accordance with guidelines of cpcsea and the study was approved by the institutional animal ethical committee (cpcsea/ iaec/sptm/p-20/2014). rats were fasted for 12 h (water ad libitum) prior to each experiment. anesthesia was induced with urethane (1.25 gm/kg, i.p.). the abdomen was opened with a midline incision and an intestinal segment of approximately 10 cm was measured, isolated and cannulated with plastic tubing. initially, the intestinal segment was rinsed with isotonic saline until the outlet solution was clear. amoxicillin solution (100 µg/ml) was given by a constant perfusion at flow rate of 0.2 ml/min. each perfusion experiment lasted for 60 min and the perfusate was quantitavely collected at 15, 30, 45 and 60 min marks. the collected samples were then analyzed by uv spectroscopy. krebs-ringer buffer solution was used as blank perfusion solution. phenol red solution was added as non-absorbable biomarker to calculate net water flux. propranolol hydrochloride and atenolol were used as high permeability and low permeability markers to validate the procedure. experiment was repeated with 1.5 % w/w of piperine and 8 % w/w of ginger resin added to the perfusate. effective permeability values were calculated from the steady-state concentrations of amoxicillin trihydrate in the perfusate collected from the outlet. in vivo drug intestinal permeability in humans (peff,man) can be predicted from rat peff values according to following formula [12]: peff, man = 3.6*peff,rat + 0.03*10 -4 . results and discussion in vitro studies using franz diffusion cell as shown in table 1, there was a dose dependant and time dependant increase in the diffusion of amoxicillin trihydrate on addition of piperine. piperine at a concentration of 1.2 % w/w did not show a significant increase at 15 and 30 min but the diffusion increased significantly at 45 and 60 minutes. piperine concentrations of 1.5 and 1.8 % w/w significantly enhanced the diffusion of amoxicillin trihydrate at 15, 45 and 60 min (p< 0.0001). however 1.5 and 1.8 % w/w of piperine additions did not show a significant difference in the diffusion of amoxicillin trihydrate when compared with each other. hence 1.5 % w/w of piperine was selected for the rest of the studies. use of ginger resin with amoxcillin trihydrate failed to show a dose dependant increase in the diffusion. 6 and 8 % w/w of ginger resin demonstrated a significant increase (p<0.0001) in the diffusion which was also dose and time dependant. however 7 % w/w of ginger resin failed to show such increase till 45 min, but a significant increase (p<0.0001) was observed at 60 min. hence, we selected 8 % w/w of ginger resin for the rest of the experiments. in vitro studies using non everted gut sac method the results for this study are shown in table 2. 1.5 % w/w piperine showed a significant increase (p<0.0001) in the permeation at all time, however after 45 minutes it reached a plateau and did not show barve and ruparel admet & dmpk 3(1) (2015) 45-50 48 any significant increase in the permeation. 8 % w/w of ginger resin demonstrated a significant reduction in the permeation of amoxicillin trihydrate. table 1. cumulative concentration of amoxicillin trihydrate using semi permeable membrane time amoxicillin (µg/ml) amoxcillin + piperine (µg/ml) amoxcillin + ginger resin (µg/ml) 1.2 % w/w 1.5 % w/w 1.8 %w/w 6 % w/w 7 % w/w 8 % w/w 15 min 6.79±0.06 6.05±0.21 8.34±0.07*** 9.89±0.09*** 7.12±0.10*** 4.4±0.050 8.82±0.14*** 30 min 9.91±0.08 8.83±0.11 10.06±0.12 12.28±0.06* 12.28±0.06 5.55±0.08 27.04±2.86*** 45 min 10.07±0.08 18.14±0.54*** 29.14±0.25*** 32.18±0.63*** 15.75±0.18*** 10.37±0.11 32.76±0.18*** 60 min 14.53±0.14 26.35±0.2*** 54.59±0.18*** 54.91±4.30*** 23.38±0.09** 26.55±0.12*** 42.87±4.17*** n=6, values are written as mean ± sd, ***p<0.0001 , **p<.001, *p<0.01 when compared to amoxicillin trihydrate using one way anova followed by bonferroni multiple comparison. table 2. cumulative concentration of amoxicillin trihydrate using non everted gut sac time amoxicillin (µg/ml) amoxcillin + piperine (µg/ml) amoxcillin + ginger resin (µg/ml) 1.5 % w/w 8 % w/w 15 min 15.76±0.14 38.33±0.07*** 7.44±0.05*** 30 min 27.96±0.07 41.51±0.14*** 15.9±0.08*** 45 min 35.64±0.07 43.41±0.08*** 20.17±0.12*** 60 min 50.02±0.31 44.85±0.08 23.51±0.08*** n=6, values are written as mean ± sd, ***p<0.0001 when compared to amoxicillin trihydrate using one way anova followed by bonferroni multiple comparison in vivo studies using single pass intestinal perfusion method (spip) table 3 shows the effective permeability in rats and humans. the peff of atenolol was found to be 0.106±0.012 x 10 -4 and 0.4116± 0.041 x 10 -4 in rats and humans respectively which is in compliance with the reported values 12. similarly the peff of propranolol hydrochloride was found to be 1.14 ±0.082 x 10 -4 and 4.134±0.236 x 10 -4 in rats and humans respectively, which again complies with the reported values for effective permeability [13]. this indicates the validity of the spip method. for spip techniques drugs with peff,rats > 0.2 x 10 -4 cm/sec and peff,man > 0.7 x 10 -4 cm/sec can be considered as highly permeable [14,15]. the peff of amoxicillin trihydrate in rats and humans was found to be 0.16 x 10 −4 cm/ in rat and 0.617 x 10 −4 cm/s respectively, thus it can be classified as poorly absorbed drug. 1.5 % w/w of piperine increased the permeation by 60 % and 54.13 % in rats and humans respectively (p<0.001). 8 % w/w of ginger resin did not show any significant change in the peff of amoxicillin trihydrate showing no bioenhancement. admet & dmpk 3(1) (2015) 45-50 bioenhancement of amoxicillin by piperine doi: 10.5599/admet.3.1.161 49 table 3. effective permeability values of atenolol, propranolol hydrochloride and amoxicillin trihydrate atenolol propranolol amoxicillin amoxicillin + piperine (1.5 % w/w) amoxicillin + ginger (8 % w/w) peff in rats (cm/s) 0.106±0.012 x 10 -4 1.140±0.082 x 10 -4 0.16±0.03 x 10 -4 0.256±0.021** x 10 -4 0.19±0.025 x 10 -4 peff in humans (cm/s) 0.412±0.041 x 10 -4 4.134±0.236 x 10 -4 0.618±0.056 x 10 -4 0.9516±0.08** x 10 -4 0.714±0.068x 10 -4 n=6, values are written as mean ± sd , **p<0.001 when compared to amoxicillin trihydrate using one way anova followed by dunnetts multiple comparison we have tried to evaluate enhancement of bioavailability using the in situ spip method. the in situ spip technique not only preserves the microclimatic conditions of the intestine but also counts the transporter activity in the same experimental condition. the result obtained with piperine confirms the use of this method for such type of study. amoxicillin trihydrate is absorbed from the intestine via passive diffusion and active transport using the pept1 transporters but at the same time is also a substrate to an energy demanding efflux system distinct from the p-glycoprotein mediated transport system which may contribute to restricted intestinal absorption [4]. piperine is known to induce alterations in membrane dynamics and permeation characteristics and thereby increase the passive diffusion of amoxicillin trihydrate [16,17]. piperine is also reported to be inhibitor of drug efflux pump one of them being pgp efflux transporter [18], thereby reducing the efflux and enhancing the availability of amoxicillin trihydrate. piperine at the same time inhibits a number of enzymes involved in metabolism of drugs and nutrients, thus reducing the metabolism of drugs and prolonging the drug effect [19]. moreover it is also known to reduce the rate of glucuronidation permitting the entry of active drug substance [19]. ginger is known to enhance the bioavailability by regulating the git function [20]. although there are reports which indicate that use of ginger extract enhances the bioavailability of many drugs [21-23], we did not get satisfactory results , which is suggestive that bioenhancers working on the gut motility may not be the right candidates to be studied by this method. it may also be possible that the ginger extract might be forming a coating over the mucosal surface thereby hindering the absorption of amoxicillin trihydrate. ginger is reported to have prokinetic property [24,25], thus it increases the intestinal motility, due to this the drug remains in contact with the mucosal surface for lesser time and hence may not be absorbed to the expected significance. conclusions the results demonstrate that piperine can be effectively used in combination with amoxicillin trihydrate to increase the biovaialbility of the later, however ginger resin did not give anticipated results of bioenhancement. the combination with piperine may help in reducing the dose, shortening the treatment period thus reducing drug-resistance problems. the treatment may be made cost effective, minimizing drug toxicity and adverse reactions. in future, formulation may be developed by combining piperine with amoxicillin trihydrate. such nature derived bioenhancers may be used in combination with other drug classes like, antitubercular, antiviral, antifungal and anticancer drugs to enhance the bioavailability. oral absorption of vitamins, minerals, herbal extracts, amino acids and other nutrients can also be enhanced. barve and ruparel admet & dmpk 3(1) (2015) 45-50 50 acknowledgements: the authors are grateful to spp-school of pharmacy and technology management, svkm’s nmims for funding this project. the authors would also like to acknowledge ipca laboratories, mumbai, india for providing gift samples. references [1] h. lennernasa, l. knutsonb, t. knutsonb, a. hussainc , l. leskoc , t. salmonsond, european journal of pharmaceutical sciences 15 (2002) 271–277. [2] w. chi-yuan, z.b. leslie, pharm. res. 22 (2005) 11-23. [3] b. ali, s. amin, j. ahmad, a. ali, m. ali ,s. r. mir, indian j. med. res. 135 ( 2012) 555-559. [4] i. legen , m. kra cun, m. salobir, j. kerc, international journal of pharmaceutics 308 (2006) 84–89. [5] b.j. aungst, j. pharm. sci. 89 (2000) 429-442. [6] k. bogman, f. erne-brand, j. alsenz, j. drewe, j. pharm. sci. 92 (2003) 1250–1261. [7] w.c. evans, trease and evans pharmacognosy, elsevier limited, china, 2009, p. 56. [8] r. ikan, natural products a laboratory guide, academic press, new delhi, india, 2005, p.233-235. [9] www.hindawi.com, http://dx.doi.org/10.1100/2012/637953 (27/01/2015). [10] c.v. jacob,(synthite industries ltd) us20120058208a1 (2012). [11] j.r. patel, k.h. barve, indian journal of pharmaceutical sciences 74 (2012) 478-481. [12] u. fagerholm, m. johansson, h. lennernas, pharmaceutical research 13 (1996) 1336-1342. [13] h. lennernas, l. knustson, r. knutson, l . lesko, t. salmonson, g.l. amidon, eur. j. pharm. sci. 4 (1996) s69. [14] d.a. volpe, h. moller, l.x. yu, in: lehr cm (ed.) cell culture models of biological barriers: in vitro test systems for drug absorption and delivery. london, new york. taylor & francis publishing group, 130-139. [15] p.z. milani, h. valizadeh, h. tajerzadeh, y. azarmi, z. islambolchilar, s. barxegar, m. jalai, j. pharm. pharmaceut. sci. 10 (2007) 368-379. [16] s. bajad, k.s. bedi, a.k. singla, r.k. johri, planta med. 67 ( 2001) 176–179. [17] s. wadhwa, s. singhal, s. rawat, asian journal of biomedical and pharmaceutical sciences 4 (2014) 1-8. [18] u.k. patil, a. singh, a.k. chakraborty, international journal of recent advances in pharmaceutical research 4 (2011) 16-23. [19] p. gohil , a. mehta, oriental pharmacy and experimental medicine 9 (2009) 269-276. [20] d.v. tatiraju, v.b. bagade, p.j. karambelkar, v.m. jadhav, v. kadam, journal of pharmacognosy and phytochemistry 2(2013)55-60. [21] s. o. nduka, e.o. okonta, l.z. adonu, j.m. okonta, international journal of applied research in natural products 6 (2013) 15-18. [22] b. jhanwar, s. gupta, international journal of pharmtech research 6 (2014) 443-454. [23] n. atal , k.l. bedi, j. ayurveda integr. med. 1 (2010) 96–99. [24] m.n. ghayur , a.h. gilani, digestive diseases and sciences 50 (2005) 1889-1897. [25] a.m. bode, z. dong, in herbal medicine: biomolecular and clinical aspects. eds: i.f.f. benzie, s. wachtel-galor, 2nd edition., crc press, boca raton, florida, united states of america 2011. ©2015 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ manuscript doi: 10.5599/admet.2.3.123 185 admet & dmpk 2(3) (2014) 185-190; doi: 10.5599/admet.2.3.123 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index short communication electrochemistry of redox-active mn porphyrin-based sod mimic mntnbuoe-2-pyp 5+ study of redox species involved in ros/rns scavenging tin weitner 1 *, ines batinic-haberle 2 1 university of zagreb, faculty of pharmacy and biochemistry, ante kovačića 1, zagreb 10000, croatia 2 department of radiation oncology, duke university school of medicine, durham, nc 27710, usa *corresponding author: e-mail: tweitner@pharma.hr; tel.: +385-1-6394-452; fax: +385-1-6394-400 received: september 09, 2014; revised: september 10, 2014; published: september 16, 2014 abstract manganese ortho tetrakis(n-n-butoxyethylpyridinium-2-yl)porphyrin, mntnbuoe-2-pyp 5+ , is a thirdgeneration redox-active compound currently undergoing preclinical exploration. this work is intended to complement the already extensive research of its chemical and biological properties by a simple electrochemical study. the thermodynamic parameters related to the mn(iv) porphyrin species of mntnbuoe-2-pyp 5+ determined in this work support its observed reactivity as an efficient scavenger of peroxynitrite. the corresponding driving forces for the possible single-electron or two-electron reductions of onoo have been estimated as well. keywords metalloporphyrins, oxidative stress, superoxide, peroxynitrite, cyclic voltammetry. introduction therapeutic potential of metalloporphyrin-based mimics of superoxide dismutase (sod) enzyme has been well documented [1-5]. briefly, a true sod mimic should be able to dismutate superoxide (i.e. oxidize or reduce o2 • ) at a catalytic rate, kcat(o2 • ), higher than the rate of o2 • self-dismutation (5 × 10 5 m -1 s -1 at ph = 7). the formal reduction potential of the sod mimic should be optimally midway between the singleelectron formal reduction potentials for oxygen (e 0’ = – 0.16 v vs. nhe) and superoxide (e 0’ = + 0.89 v vs. nhe) [6-7]. over the last 15 years, a rational approach to the design of potential sod mimics led to the development of several manganese porphyrin complexes (mnps) with unique electronic properties and high reactivity towards o2 • , but also other reactive species such as peroxynitrite, onoo [2-5]. indeed, a structure-activity relationship (half-wave reduction potential, e1/2, vs. log kcat(o2 • )) has been developed and found to be valid for any other class of sod mimics as well [1]. recent efforts have been directed towards increasing the bioavailability and reducing the toxicity of mnps. the original structure of a potent first-generation sod mimic, manganese ortho tetrakis(nethylpyridinium-2-yl)porphyrin (mnte-2-pyp 5+ ) was modified via lengthening the alkyl chains of pyridyl http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:tweitner@pharma.hr weitner and batinic-haberle admet & dmpk 2(3) (2014) 185-190 186 substituents, yielding the more lipophilic compound, the manganese ortho tetrakis(n-n-hexylpyridinium-2yl)porphyrin (mntnhex-2-pyp 5+ ) [8]. this second-generation compound showed much higher brain and mitochondrial distribution than mnte-2-pyp 5+ , but was also found to be toxic at higher concentrations and could have a potentially limited use [9-10]. the third-generation compound, manganese ortho tetrakis(n-nbutoxyethylpyridinium-2-yl)porphyrin (mntnbuoe-2-pyp 5+ ), was designed via introduction of oxygen atoms into the alkyl chains on the pyridyl substituents. this compound preserved high redox-activity and lipophilicity but exhibited significantly reduced toxicity relative to mntnhex-2-pyp 5+ [11]. comparison of mnte-2-pyp 5+ and mntnbuoe-2-pyp 5+ with regards to their redox properties (e1/2 for mn iii p/mn ii p redox couple), reactivities towards superoxide, log kcat(o2 • ), and peroxynitrite, log kred(onoo ), as well as lipophilicities (distribution between n-octanol and water, log pow) is given in table 1. table 1. physicochemical properties of redox-active mn porphyrin-based sod mimics mnte-2-pyp 5+ and mntnbuoe-2-pyp 5+ . data are taken from references [3-5,11-16] ; a at 25 °c; b at 37 °c. ar n ethyl n-butoxyethyl e1/2 (mv vs. nhe) + 228 + 277 log kcat (o2 •) a 7.76 7.83 log kred (onoo ) b 7.53 7.54 log pow -7.79 -4.10 comprehensive studies on bioavailability of these compounds have been performed recently as well [1719], in order to gain further insight into the biodistribution of mnps, but also provide guidelines for proper dosing regimens. after assessing preliminary efficacy and toxicity data, the porphyrins were administered to mice intraperitoneally (ip) or intravenously (iv) at following doses: 10 mg/kg of mnte-2-pyp 5+ (ip) and 2 mg/kg of mntnbuoe-2-pyp 5+ (iv). drug levels in plasma and different organs were followed during at least 24 hours after a single dose, thus allowing the calculation of relevant pharmacokinetic (pk) parameters via non-compartmental analysis of pk curves. these studies showed the highest mnp bioavailability in liver (maximum concentration, cmax, values of 8.1 and 13.1 µmol per l of tissue homogenate for mnte-2-pyp 5+ and mntnbuoe-2-pyp 5+ , respectively) and kidney, followed by spleen, lung and heart, whereas lower levels were found in plasma and lowest in the brain. remarkably, mntnbuoe-2-pyp 5+ has 8-fold higher brain bioavailability (expressed as the area under pk curve, auc, divided by dose) than the less lipophilic mnte-2-pyp 5+ (figure 1), which may be highly significant for its potential application as a brain tumor radioand chemo-therapeutic [20-21]. additionally, it has been shown that kcat(o2 • – ) parallels the ability of mnps to reduce onoo [4-5,14], due to the same thermodynamic and kinetic factors facilitating the reaction of electron-deficient cationic mnps with electron-donating anionic species, such as o2 • – and onoo [3-5,22]. specifically, the removal of onoo can occur either through its binding to the mn(iii) site followed by its one-electron reduction to no2 • radical (coupled to the oxidation of mn iii p to mn iv p), or the two-electron reduction of onoo by the mn ii p species (coupled to the oxidation of mn ii p to mn iv p). the latter reaction produces less toxic nitrite, no2 , and might in fact be dominant in vivo due to (i) the abundance of cellular reductants keeping mnps in mn(ii) state [1,13,23-25] and (ii) a higher driving force making it thermodynamically favorable [26]. admet & dmpk 2(3) (2014) 185-190 redox active mn-porphyrin sod mimetic doi: 10.5599/admet.2.3.123 187 figure 1. bioavailability of redox-active mn porphyrin-based sod mimics mnte-2-pyp 5+ and mntnbuoe-2-pyp 5+ , demonstrated as 24-hour dose-adjusted auc values: raw auc values have been divided by dose (10 mg/kg for mnte-2-pyp 5+ and 2 mg/kg for mntnbuoe-2-pyp 5+ ). the values for mnte-2-pyp 5+ determined via iv route are taken as 100 % availability (plasma bioavailability via ip route was found to be approximately 84% of bioavailability via iv route for mnte-2-pyp 5+ ) [18]. mnps’ charges have been omitted from the plot for simplicity. mntnbuoe-2-pyp 5+ represents the latest generation of redox-active compounds and is currently undergoing preclinical exploration. this work is intended to complement the already extensive research of its chemical and biological properties by a simple electrochemical study. experimental the investigated metalloporphyrin, mn iii tnbuoe-2-pypcl5 was synthesized according to the published procedure [11]. stock solution was prepared by dissolution of purified and characterized solid (known molar absorbance coefficient, ε456 (mntnbuoe-2-pyp 5+ ) = 1.78 × 10 5 m -1 cm -1 ) [11]. cyclic voltammetry of buffered and deaerated solutions was performed as reported in a previous study [26]. all potentials are reported vs. the normal hydrogen electrode (nhe), using the known potential of mnte-2-pyp 5+ , e1/2 = + 228 mv vs. nhe at ph 7.8, as a reference [11,27-28]. results and discussion cyclic voltammograms of the aqueous solutions of mn iii ps can show two distinct electron transfers: only one current peak pair attributed to the mn iii p/mn ii p redox couple is observed in a neutral or mildly basic medium (ph = 7.8, figure 2), whereas in the more basic medium an additional current peak pair attributed to the mn iv p/mn iii p redox couple appears at more positive potentials (ph = 11, figure 2). the cathodic and anodic current peak potentials, epc and epa, of both redox couples shift toward more negative values upon the increased ph of the solution. the cathodic-anodic peak separation, δep = |epc epa| for both redox processes is larger than 59 mv even at the scan rate of 0.02 v s -1 , indicating quasi-reversible electron transfer processes. weitner and batinic-haberle admet & dmpk 2(3) (2014) 185-190 188 figure 2. comparison of cyclic voltammograms of 0.5 mm aqueous solutions of mntnbuoe-2-pyp 5+ at ph=7.8 and ph = 11, ν = 0.02 v s -1 , [nacl] = 0.1 m,  = 25 °c. similar behavior has been reported previously for methyl, ethyl and butyl analogues [15,26,29]. redox couples shown relate to the mn iii p/mn ii p (right cycle) and the mn iv p/mn iii p couples (left cycle). these results are in agreement with the previously established scheme of behavior of mnps in aqueous solutions [26,30]. at ph = 7.8 only a single-electron reduction of mn iii p can be observed, shown by eq. (1), whereas at ph = 11 the single-electron reduction of mn iv p is accompanied by a simultaneous dissociation of two protons, according to eq. (2). the latter electron transfer cannot be observed in a neutral medium due to a formal potential more positive than that of water (e 0’ (o2(g),h + (aq.)/h2o(l)) = 816 mv vs. she at ph = 7) [30-31].           iii 5 ii 42 2 22h o mn tnbuoe 2 pyp e h o mn tnbuoe 2 pyp h o (1)             iv 4 iii 52 2 2o h o mn tnbuoe 2 pyp 2h e h o mn tnbuoe 2 pyp (2) table 2. calculated reduction potentials for the mn iii p/mn ii p, mn iv p/mn iii p and mn iv p/mn ii p redox couples of mntnbuoe-2-pyp 5+ . values for mnte-2-pyp 5+ are taken from reference [26] and adjusted to the value e1/2 = + 228 mv vs. nhe for its mn iii p/mn ii p redox couple at ph=7.8. the reduction potential of the mn iv p/mn ii p couple is calculated as: e1/2(mn iv p/mn ii p) = [e1/2(mn iii p/mn ii p) + e1/2(mn iv p/mn iii p)]/2. mnp e1/2 (mv vs. nhe) ph = 7.8 ph = 11 mn iii p/mn ii p mn iii p/mn ii p mn iv p/mn iii p mn iv p/mn ii p mntnbuoe-2-pyp 5+ + 277 + 235 + 624 + 430 mnte-2-pyp 5+ + 228 + 208 + 592 + 400 admet & dmpk 2(3) (2014) 185-190 redox active mn-porphyrin sod mimetic doi: 10.5599/admet.2.3.123 189 as shown previously [11], the e1/2 of mn iii p/mn ii p redox couple of mntnbuoe-2-pyp 5+ is close to the value of the mn sod enzyme itself, accounting for its high reactivity towards o2 •– (even higher than that of mnte-2-pyp 5+ , table 1). thermodynamic parameters related to the mn iv p species of mntnbuoe-2-pyp 5+ determined in this work (table 2) support its observed reactivity as an efficient scavenger of peroxynitrite. estimating the shift of the e1/2 (mn iv p/mn iii p) value for +118 mv per unit decrease of ph from ph = 11, yields the values of e1/2 (mn iv p/mn iii p) = + 1049 mv vs. nhe at the physiological ph, according to eq. (2). the e1/2 (mn iv p/mn ii p) at the physiological ph can then be calculated as e1/2(mn iv p/mn ii p) = [e1/2(mn iii p/mn ii p) + e1/2(mn iv p/mn iii p)]/2 = + 663 mv vs. nhe. finally, the driving forces for the single-electron and two-electron reduction of onoo , emf (1e ) = 551 mv and emf (2e ) = 637 mv, can then be calculated as emf = ec ea, where the values of ec are the formal potentials e 0’ (onoo /no2 • ) = 1.6 v and e 0’ (onoo /no2 ) = 1.3 v [32], respectively, whereas the values of ea are the formal potentials e1/2(mn iv p/mn iii p) and e1/2(mn iv p/mn ii p), respectively. conclusions considerable interest in the impact of oxidative stress on cellular function prompted an intensive search for natural and synthetic antioxidants. high bioavailability of mntnbuoe-2-pyp 5+ , coupled with its exceptional physicochemical properties as a redox-active ros/rns scavenger, makes it a promising candidate in the ongoing cancer and radioprotection preclinical studies. acknowledgements: authors acknowledge financial help from duke university’s ctsa grant 1 ul 1 rr024128-01 from ncrr/nih, the robert preston tisch brain tumor center at duke career developmental award to ibh, and croatian ministry of science (mzoš) grant 006-0061247-0009. ibh is a consultant with biomimetix pharmaceutical, inc. references [1] i. batinic-haberle, a. tovmasyan, e. r. roberts, z. vujaskovic, k. w. leong and i. spasojevic antioxidants & redox signaling 20 (2014) 2372-415. [2] i. batinic-haberle, i. spasojevic, h. m. tse, a. tovmasyan, z. rajic, d. k. st clair, z. vujaskovic, m. w. dewhirst and j. d. piganelli amino acids 42 (2012) 95-113. [3] i. batinic-haberle, j. s. reboucas, l. benov and i. spasojevic in handbook of porphyrin science, vols.11-15; k.m. kadish, k.m. smith, r. guillard, eds.; world scientific: singapore, 2011; pp 291-393. [4] i. batinic-haberle, z. rajic, a. tovmasyan, j. s. reboucas, x. ye, k. w. leong, m. w. dewhirst, z. vujaskovic, l. benov and i. spasojevic free radical biology and medicine 51 (2011) 1035-53. [5] i. batinić-haberle, j. s. rebouças and i. spasojević antioxidants & redox signaling 13 (2010) 877-918. [6] p. m. wood biochemical journal 253 (1988) 287-289. [7] d. t. sawyer and j. s. valentine accounts of chemical research 14 (1981), 393-400. [8] i. batinić-haberle, i. spasojević, r. d. stevens, p. hambright and i. fridovich dalton transactions (2002) 2689-2696. [9] i. spasojevic, s. miriyala, a. tovmasyan, d. salvemini, p. fan, z. vujaskovic, i. batinic-haberle and d. k. s. clair free radical biology and medicine 51 (2011) s98-s99. [10] i. spasojevic, t. weitner, a. tovmasyan, h. sheng, s. miriyala, d. leu, z. rajic, d. s. warner, d. s. clair, t.-t. huang and i. batinic-haberle free radical biology and medicine 65 (2013) s132. weitner and batinic-haberle admet & dmpk 2(3) (2014) 185-190 190 [11] z. rajic, a. tovmasyan, i. spasojevic, h. sheng, m. lu, a. m. li, e. b. gralla, d. s. warner, l. benov and i. batinic-haberle free radical biology and medicine 52 (2012) 1828-1834. [12] s. carballal, v. valez, i. batinic-haberle, g. ferrer-sueta and r. radi free radical biology and medicine 65 (2013) s121-s122. [13] g. ferrer-sueta, d. vitturi, i. batinic-haberle, i. fridovich, s. goldstein, g. czapski and r. radi journal of biological chemistry 278 (2003) 27432-27438. [14] g. ferrer-sueta, c. quijano, b. alvarez and r. radi methods in enzymology 349 (2002) 23-37. [15] g. ferrer-sueta, i. batinic-haberle, i. spasojevic, i. fridovich and r. radi chemical research in toxicology 12 (1999) 442-449. [16] s. carballal and r. radi, unpublished. [17] a. tovmasyan, t. weitner, h. sheng, x. chen, k. ashcraft, p. fan, m. w. dewhirst, d. s. warner, z. vujaskovic, i. batinic-haberle and i. spasojevic, (2014) in preparation. [18] t. weitner, i. kos, h. sheng, a. tovmasyan, j. s. reboucas, p. fan, d. s. warner, z. vujaskovic, i. batinic-haberle and i. spasojevic free radical biology and medicine 58 (2013) 73-80. [19] i. spasojevic, y. chen, t. j. noel, p. fan, l. zhang, j. s. rebouças, d. k. st. clair and i. batinic-haberle free radical biology and medicine 45 (2008) 943-949. [20] d. h. weitzel, a. tovmasyan, k. a. ashcraft, z. rajic, t. weitner, c. liu, w. li, a. f. buckley, m. r. prasad, k. h. young, r. m. rodriguiz, w. c. wetsel, k. b. peters, i. spasojevic, j. i. herndon ii, i. batinic-haberle and m. w. dewhirst molecular cancer therapeutics, (2014) in revision. [21] d. leu, y. zou, t. weitner, a. tovmasyan, i. spasojevic, i. batinic-haberle and t.-t. huang in 60th annual meeting of the radiation research society: las vegas, nevada, usa, 2014. [22] j. s. rebouças, i. spasojević, d. h. tjahjono, a. richaud, f. méndez, l. benov and i. batinić-haberle dalton transactions (2008) 1233-1242. [23] g. ferrer-sueta, l. hannibal, i. batinic-haberle and r. radi free radical biology and medicine 41 (2006) 503-512. [24] v. valez, a. cassina, i. batinic-haberle, b. kalyanaraman, g. ferrer-sueta and r. radi archives of biochemistry and biophysics 529 (2013) 45-54. [25] m. c. jaramillo, m. m. briehl, j. d. crapo, i. batinic-haberle and m. e. tome free radical biology and medicine 52 (2012) 1272-1284. [26] t. weitner, i. kos, z. mandic, i. batinic-haberle and m. birus dalton transactions 42 (2013) 1475714765. [27] a. g. tovmasyan, z. rajic, i. spasojevic, j. s. reboucas, x. chen, d. salvemini, h. sheng, d. s. warner, l. benov and i. batinic-haberle dalton transactions 40 (2011) 4111-4121. [28] i. kos, j. s. rebouças, g. defreitas-silva, d. salvemini, z. vujaskovic, m. w. dewhirst, i. spasojevic and i. batinic-haberle free radical biology and medicine 47 (2009) 72-78. [29] a. budimir, t. šmuc, t. weitner, i. batinic-haberle and m. biruš journal of coordination chemistry 63 (2010) 2750-2765. [30] t. weitner, a. budimir, i. kos, i. batinic-haberle and m. birus dalton transactions 39 (2010) 1156811576. [31] p. w. atkins physical chemistry; 6th ed.; oxford university press: oxford, 1998, 936. [32] w. h. koppenol free radical biology and medicine 25 (1998) 385-391. ©2014 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ manuscript doi: 10.5599/admet.2.4.155 248 admet & dmpk 2(4) (2014) 248-253; doi: 10.5599/admet.2.4.155 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper rapid screening of permeation of rutin through skin using alaptide enantiomers aneta cernikova, radka opatrilova, pavel bobal and josef jampilek department of chemical drugs, faculty of pharmacy, university of veterinary and pharmaceutical sciences, palackeho 1/3, 61242 brno, czech republic corresponding author: e-mail: josef.jampilek@gmail.com; tel.: +420-541-562-926 received: december 21, 2014; revised: december 22, 2014; published: january 09, 2015 abstract the investigation deals with the influence of permeation of rutin from carboxymethyl cellulose gel through full-thickness pig ear skin by (s)and (r)-alaptide as potential excipients. alaptide, 8-methyl-6,9diazaspiro[4.5]decan-7,10-dione, is the original czech compound. by means of this rapid screening it was found out that the permeation of rutin through the skin increased linearly with time and was enhanced by both enantiomers of alaptide: approx. 1.2-fold by (r)-alaptide and approx. 1.5-fold by (s)-alaptide. keywords (s)-alaptide; (r)-alaptide, rutin; carboxymethyl cellulose gel; skin permeation modifiers. introduction permeation modifiers (enhancers and retardants) affect the barrier properties of the skin. chemical permeation enhancers (cpes) are employed to increase the transdermal penetration of therapeutic agents [1−6 and refs. therein]. alaptide, (s)-8-methyl-6,9-diazaspiro[4.5]decan-7,10-dione (see fig. 1) was designed as an analogue of melanocyte-stimulating hormone release-inhibiting factor (mif-1) [7−10] and thus is able to influence the creation and function of keratinocytes [11−13]. based on these facts and structural analogy with other cpes [4], it was suggested to investigate alaptide as a potential cpe of many anti-inflammatory drugs, antimicrobial chemotherapeutics, sex hormones/genital system modulators or drugs of central/vegetative nervous system [14−16]. permeation through the skin of micronized and nanonized (s)-alaptide from semisolid formulations was investigated recently [17], and the transdermal modifying effect of micronized and nanonized (s)-alaptide was also described [14−16,18,19]. it was found that enhancers can become retardants in different formulation environments or when different particle sizes were used and vice versa. since modifiers influence drug delivery, further exploration of these compounds is required to understand their modifying action on the properties of the skin. today more than 50 % of marketed drugs are chiral. it has been well recognized that the stereochemistry of chiral drugs has a major influence on their pharmacological, pharmacokinetic and toxicological actions. studies on enantiomeric differences in the percutaneous permeation of chiral compounds have been actively pursued in recent years. stratum corneum, the rate-limiting barrier in http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:josef.jampilek@gmail.com mailto:josef.jampilek@gmail.com admet & dmpk 2(4) (2014) 248-253 rutin permeation enhancement by alaptide enantiomers doi: 10.5599/admet.2.4.155 249 transdermal permeation, is made up mainly of keratin and ceramides, which could potentially provide chiral environment. transdermal delivery is facilitated by the presence of penetration enhancers, which act primarily by altering the diffusion by disrupting the highly ordered membrane structure or by affecting the partitioning behaviour of the diffusant molecules. enantioselective permeation was observed with some chiral excipients, including, for example, terpene enhancers [20−22]. based on the above mentioned facts and the excellent enhancement effect of (s)-alaptide [14−16], (r)-alaptide was prepared, and both enantiomers were tested by means of rapid screening for their ability to modify the skin barrier properties. bioflavonoid glycoside rutin, 2-(3,4-dihydroxyphenyl)-5,7-dihydroxy-4oxo-4h-chromen-3-yl 6-o-(6-deoxy-α-l-mannopyranosyl)-β-d-glucopyranoside, see fig. 1, was chosen as an model compound, since rutin is one of the most often used drugs for decreasing capillary permeability and for treatment of haemorrhoids, varicosis and microangiopathy [23]. enhanced permeation through the skin and higher availability of the compound would be favourable for the treatment of vascular wall fragility. within this study the permeation of rutin from carboxymethyl cellulose gel and subsequently the permeation of rutin with the addition of the individual enantiomers of alaptide as modifiers of transdermal permeation from the same formulation are investigated. flavonoids (bioflavonoids, vitamin p) are a class of plant secondary phenolic metabolites. today more than 4000 flavonoid-based compounds were described [24,25]. flavonoids are glycosides containing aglycone derived from benzopyran (chromene) with lateral phenyl in the positions c(2), c(3) or c(4) of the benzopyran scaffold. the phenyl rings are marked a and b, and the pyran ring c is ortho-condensed with the ring a. in nature flavonoids are substituted (especially in position 3 of the benzopyran scaffold) by various sugars and generate flavonoid glycoside. aglycone or the glycosyl group can be substituted by a hydroxyacid [25,26]. both aglycone and glycosides have biological effects. bioflavonoids normalize permeability of capillaries and remove their fragility; they have antihaemorrhagic and anti-edematous effects [25−28]. also antibacterial, antioxidant and anti-inflammatory activities are typical for them [25−28]. bioflavonoids are used as venopharmaceuticals, i.e. drugs for the treatment of chronic venous insufficiency, superficial thrombophlebitis and haemorrhoids. they can be applied locally for the treatment of dermatomicroangiopathy, periphlebitis, haematomas and traumatic oedema [28]. figure 1. structures of investigated compounds. experimental preparation of alaptide both alaptide enantiomers were synthesized by the standard process [29] and characterized [30]. the specific optical rotation of 0.2 mol/l methanolic solution is [α]d 20 = +21.2° for the (r)-enantiomer and [α]d 20 = -21.2° for the (s)-enantiomer. the melting point of (s)-alaptide is 232-234 °c, while that of jampilek et al. admet & dmpk 2(4) (2014) 248-253 250 (r)-alaptide is 230-232 °c. rutin was purchased from sigma, all other reagents were purchased from aldrich and lach-ner. investigated semisolid formulations the composition of gel with 10% (w/w) rutin was the following: rutin 10 g and carboxymethyl cellulose gel (carboxymethyl cellulose sodium 5 g, macrogol 300 10 g, propylene glycol 2.5 g, methylparaben 0.2 g, propylparaben 0.2 g and purified water 87.3 g) up to 100 g. the composition of gel with the addition of alaptide 1% (w/w; in relation to rutin amount) was the following: rutin 10 g, alaptide 0.1 g and carboxymethyl cellulose gel up to 100 g. in vitro transdermal permeation experiments skin samples were obtained from porcine ear. full thickness skin was cut in fragments and stored at −20 °c until utilized. skin samples were slowly thawed (at 4 °c overnight and then at ambient temperature) before each experiment. the permeation enhancing effect of both alaptide enantiomers was evaluated in vitro, using a vertical franz diffusion cell (ses – analytical systems, germany) with a donor surface area of 0.6359 cm 2 and a receptor volume of 5.2 ml. the skin was mounted between the donor and receptor compartments of the franz diffusion cell with the epidermal side up. the receptor compartment was filled with phosphate buffered saline (ph 7.4) and maintained at 34.0±0.5 °c, using a circulating water bath. the receptor compartment content was continuously stirred using a magnetic stirring bar. the skin was kept in contact with the receptor phase for 0.5 h prior to the experiment. the investigated gel formulations were applied to the skin surface as a donor sample, and the donor compartment of the cell was covered by parafilm®. samples (0.5 ml) of the receptor phase were withdrawn at five pre-determined time intervals (30, 60, 90, 120 and 180 min), and the cell was refilled with an equivalent amount of fresh buffer solution. a minimum of five determinations was performed using skin fragments from a minimum of 2 animals for each compound, and the data was expressed as means ± sd. the samples were immediately analysed by hplc. the analysis of the samples was performed using an agilent 1200 series hplc system, equipped with a diode array detection (dad) system, a quaternary model pump and an automatic injector (agilent technologies, germany). data acquisition was performed using chemstation chromatography software. a zorbax eclipse xdb c18 5 μm, 4.6×150 mm (agilent, usa) chromatographic column was used. the total flow of the column was 0.5 ml/min; injection was 10 μl; column temperature was 30 °c; and sample temperature was 10 °c. the detection wavelength of 254 nm was chosen; the time of analysis was 10 min. a mixture of meoh (hplc grade, 85.0 %) and h2o (hplc – mili-q grade, 15.0%) was used as a mobile phase. the retention time (tr) of rutin was 3.70±0.05 min; the limit of detection (lod) was 0.99 ng/ml; and the limit of quantification (loq) was 3.3 ng/ml. results and discussion rutin was chosen as a frequently applied venopharmaceutical for the topical treatment of the fragility of vascular wall that can be associated with various injuries of the skin, because alaptide itself shows significant curative effect on the skin [8,31]. the rapid in vitro skin permeation screening experiments were performed using static franz diffusion cells [32] within 3 hours only. full-thickness pig ear skin was selected for in vitro evaluation of permeation. this tissue is a suitable in vitro model of human skin [33], because porcine skin has shown to be histologically and biochemically similar to human skin, and it has been used in numerous studies [4]. the permeation of rutin through full-thickness pig ear skin without and with 1 % admet & dmpk 2(4) (2014) 248-253 rutin permeation enhancement by alaptide enantiomers doi: 10.5599/admet.2.4.155 251 (w/w related to rutin) (s)or (r)-enantiomer of alaptide was tested from carboxymethyl cellulose gel as a donor vehicle. all the results are provided in table 1 and illustrated in figure 2. table 1. cumulative permeated amounts qt per unit area [μg/cm 2 ] of rutin from carboxymethyl cellulose gel without and with (s)-alaptide or (r)-alaptide as modifiers of transdermal permeation achieved in in vitro transdermal permeation experiments using franz diffusion cell. qt are expressed as mean ± sd (n = 5 experiments). time [h] cumulative permeated amounts qt per unit area[µg/cm 2 ] rutin rutin+(s)-alaptide rutin+(r)-alaptide 0.5 3.7±0.6 5.5±0.7 4.6±0.6 1.0 6.5±0.7 9.7±0.9 7.7±0.8 1.5 9.0±0.5 15.4±1.0 11.9±1.0 2.0 13.2±1.1 19.4±1.2 15.1±0.8 3.0 16.7±1.3 25.9±1.5 20.0±1.1 0 5 10 15 20 25 30 0.5 1 1.5 2 3 q t [ g /c m 2 ] time [h] rut rut+sala rut+rala figure 2. in vitro profile of cumulative permeated amounts qt per unit area [μg/cm 2 ] in time of rutin (rut) alone and after addition of 1 % w/w of (s)-alaptide (sala) or (r)-alaptide (rala) from carboxymethyl cellulose gel through skin. qt values are expressed as mean ± sd (n = 5 experiments). as the presented study represents a preliminary rapid in vitro screening, all the experiments were performed only in the time range from 0.5 to 3.0 h. the values obtained from the permeation experiments were expressed as the cumulative permeated amount of the drug (qt [μg]) per unit of skin surface area (0.6359 cm 2 ), see table 1. due to the short time interval of experiments steady-state permeation flux and lag time cannot be calculated, but this was not the aim of the experiments. the dependences of the cumulative permeated amount of the drug per unit of skin surface area in time (qt versus t) are shown in figure 2. based on the results it can be stated that permeation of rutin through the skin increased linearly with time; for rutin alone correlation coefficient was r = 0.9891 and slope 5.95 /cm 2 /h; for rutin with (r)-alaptide r = 0.9896 and slope 7.16 /cm 2 /h; and for rutin with (s)-alaptide r = 0.9915 and slope 9.21 /cm 2 /h. based on the ratios slopes of the linear dependences qt versus t permeation enhancement within the studied time span (up to 3 h) can be evaluated. the permeation of rutin with (r)-alaptide was 1.22-fold higher than without the addition of (r)-alaptide. the addition of (s)-alaptide to carboxymethyl cellulose gel caused 1.54-fold higher permeation of rutin through the skin than from the jampilek et al. admet & dmpk 2(4) (2014) 248-253 252 formulation without (s)-alaptide. on the other hand, rutin permeated from the formulation containing (s)-alaptide only 1.27-fold more compared with (r)-alaptide. as mentioned above the components of stratum corneum form a chiral environment, therefore there is potential for stereoselective interactions of an enhancer with skin components. stereoselective interactions of individual drug enantiomers (e.g., propranolol, metoprolol, timolol, ketorolac, ketoprofen, flurbiprofen, ephedrine, oxybutynin, amlodipine) with the skin components involving different permeation rates of individual enantiomers through the skin were tested. some authors found different permeation rates for enantiomers, while other studies showed no enantiomeric differences in the permeation [20−22,34−36]. similarly, investigations of interactions of individual cpe enantiomers with the skin provided opposite results [20,22,37,38]; therefore additional studies are needed to be performed to answer whether the skin is an enantioselective membrane and given its mechanism of action, enhancer conformation can influence its enhancing properties or the differences in the permeation rates between the individual enantiomers are not related to any stereoselective interactions, but they can be explained only by different physicochemical properties. conclusions the ability of (s)and (r)-alaptide to modify the permeation of rutin through the porcine ear skin was examined using the franz cell in the rapid in vitro screening. both enantiomers of alaptide applied in concentration 1 % w/w related to the amount of rutin influenced the permeation of the drug from the carboxymethyl cellulose gel through the skin. according to the calculated cumulative permeated amount of the drug per unit of skin surface area in time, rutin with addition of the (r)-enantiomer of alaptide permeated 1.22-fold more than from the formulation without the alaptide enantiomers, and with addition of the (s)-enantiomer of alaptide permeated 1.54-fold more than without the alaptide enantiomers. the permeation of rutin through the skin increased approximately linearly with time. it can be concluded that the presence of alaptide facilitated permeation of rutin. acknowledgements: this study was supported by the centralized development project “improving qualification of students of ph.d. educational programs & postdoctorates of agricultural and veterinary branches through engagement in multidisciplinary interuniversity teams”. (studie byla finančně podpořena centralizovaným rozvojovým projektem “zvyšování kvalifikace studentů doktorských studijních programů a post-doktorandů zemědělských a veterinárních oborů zapojením do multidisciplinárních meziuniverzitních týmů”.) references [1] h.y. thong, h. zhai, h.i. maibach, skin pharmacology and physiology 20 (2007) 272-282. [2] i.b. pathan, c.m setty, tropical journal of pharmaceutical research 8 (2009) 173-179. [3] r.k. subedi, s.y. oh, m.k. chun, h.k. choi, archives of pharmacal research 33 (2010) 339-351. [4] j. jampilek, k. brychtova, medicinal research reviews 32 (2012) 907–947. [5] j. jampilek, journal of bioequivalence & bioavailability 5 (2013) 233–235. [6] d. kaushik, p. batheja, b. kilfoyle, v. rai, b. michniak-kohn, expert opinion drug delivery 5 (2008) 517–529. [7] e. kasafirek, j. vanzura, i. krejci, j. krepelka, a. dlabac, m. valchar, (united pharmaceutical works & research institute for pharmacy and biochemistry), belg. 897843 (1984) & cs 231227 (1986). [8] s. radl, e. kasafirek, i. krejci, drugs of the future 15 (1990) 445–447. [9] r. lapka, journal of pharmacy and pharmacology, 43 (1991) 874-876. [10] j. jampilek, r. opatrilova, a. rezacova, z oktabec, j. dohnal, (faculty of pharmacy, university of veterinary and pharmaceutical sciences brno), wo/2014/019556 a1 (2014). admet & dmpk 2(4) (2014) 248-253 rutin permeation enhancement by alaptide enantiomers doi: 10.5599/admet.2.4.155 253 [11] j.a. mcgrath, r.a. eady, f.m. pope, rook's textbook of dermatology, 7th ed. t. burns, s. breathnach, n. cox, c. griffiths, (eds)., blackwell publishing, oxford, 2004, pp. 3–7. [12] w. james, t. berger, d. elston. andrews' diseases of the skin: clinical dermatology, 10th ed., saunders-elsevier, philadelphia, 2006, pp. 5–6. [13] f.m. watt, bioessays 8 (1988) 163–167. [14] j. jampilek, r. opatrilova, l. coufalova, a. cernikova, j. dohnal, (faculty of pharmacy, university of veterinary and pharmaceutical sciences brno), wo/2013/020527 a1 (2013). [15] j. jampilek, a. cernikova, r. opatrilova, j. dohnal, (faculty of pharmacy, university of veterinary and pharmaceutical sciences brno), cz pv 2013-1000 (2013). [16] j. jampilek, r. opatrilova, l. dvorakova, k. brychtova, j. dohnal, (faculty of pharmacy, university of veterinary and pharmaceutical sciences brno), cz 304915 b6 (2014). [17] r. opatrilova, a. cernikova, l. coufalova, j. dohnal, j. jampilek, the scientific world journal, 2013 (2013), article id 787283 (8 pages). [18] r. opatrilova, j. jampilek, admet & dmpk 2 (2014) 56–62. [19] a. cernikova, r. opatrilova, j. jampilek, military medical science letters 83 (2014) 34–39. [20] i.k. reddy, t.r. kommuru, a.a. zaghloul, m.a. khan, critical reviews in therapeutic drug carrier systems 17 (2000) 285–325. [21] i.k. reddy, r. mehvar, chirality in drug design and development, marcel dekker, new york, 2004. [22] k. vavrova, j. zbytovska, a. hrabalek, current medicinal chemistry 12 (2005) 2273–2291. [23] drug bank ver. 4.1 – rutin, http://www.drugbank.ca/drugs/db01698 (december 12, 2014). [24] a. benthsath, s.t. rusznyak, a. szent-gyorgyi, nature 139 (1937) 326–327. [25] q.m. andersen, k.r. markham. flavonoids: chemistry, biochemistry and applications. crc press, boca raton, 2006. [26] s. kumar, a.k. pandey. the scientific world journal 2013 (2013) article id 162750 (16 pages). [27] l.h. cazarolli, l. zanatta, e.h. alberton, m.s. figueiredo, p. folador, r.g. damazio, m.g. pizzolatti, f.r. silva, mini-reviews in medicinal chemistry 8 (2008) 1429–1440. [28] m.s. gohel, a.h. davies, current vascular pharmacology 7 (2009) 303-308. [29] e. kasafirek, m. rybak, i. krejci, a. sturs, e. krepela, a. sedo, life science 50 (1992) 187–193. [30] o. julinek, v. setnicka, a. rezacova, j. dohnal, v. vosatka, m. urbanova, journal of pharmaceutical and biomedical analysis 53 (2010) 958–961. [31] e. kasafirek, l. korbova, j. kohout, m. jiraskova, i. krejci, a. galatik, (united pharmaceutical works & research institute for pharmacy and biochemistry), cs 276270 (1992). [32] t.j. franz, journal of investigative dermatology 64 (1975) 190–195. [33] u. jacobi, m. kaiser, r. toll, s. mangelsdorf, h. audring, n. otberg, w. sterry, j. lademann, skin research and technology 13 (2007) 19–24. [34] c.m. heard, k.r. brain, chirality 7 (1995) 305–309. [35] j. valentova, k. bauerova, l. farah, f. devinsky, archives of dermatological research 302 (2010) 635-638. [36] a. zeng, c. wang, b. yuan, g. yang, q. fu, drug development and industrial pharmacy 36 (2010) 724–734. [37] j. novotny, p. kovarikova, m. novotny, b. janusova, a. hrabalek, k. vavrova, pharmaceutical research 26 (2009) 811–821. [38] y.s.r. krishnaiah, a. nada, pharmaceutical development and technology 17 (2011) 574–582. ©2014 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://www.drugbank.ca/drugs/db01698 http://creativecommons.org/licenses/by/3.0/ manuscript doi: 10.5599/admet.3.1.154 68 admet & dmpk 3(1) (2015) 68-76; doi: 10.5599/admet.3.1.154 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index short communication p-glycoprotein-dependent pharmacokinetics of irinotecan and its active metabolite, sn-38 in rats: effect of verapamil tripta garg* 1 , manu jaggi 2 , roop k. khar 3 and sushama talegaonkar 4 1 full affiliation, address 1department of scientific and industrial research, ministry of science and technology, new mehrauli road, new delhi 110016, india 2 dabur research foundation, 22, site iv, sahibabad, ghaziabad, u.p. 201010, india 3 principal, b. s. anangpuria institute of pharmacy, faridabad, india 4 dept. of pharmaceutics, faculty of pharmacy, jamia hamdard (hamdard university, hamdard nagar, new delhi 110062, india *corresponding author: e-mail: dr.tripta@gmail.com received: december 09, 2014; revised: january 19, 2015; published: march 31, 2015 abstract we have recently demonstrated that the oral bioavailability of irinotecan (80 mg/kg) can be increased at least 7-fold by co-administration of the p-gp blocker verapamil (25 mg/kg, oral). as a result, co-treatment with p-gp inhibitor could be a useful strategy for bioavailability enhancement. however, in view of narrow therapeutic index, the co-administration of irinotecan and verapamil may result in unanticipated toxicities. therefore, dose optimisation studies of irinotecan were performed when it is given in conjunction with a p-gp inhibitor. for dose optimization study, the bioavailability and pharmacokinetic parameters were studied in rats after oral administration of irinotecan at three doses (i.e. 20, 40 and 80 mg/kg) alone and in combination with verapamil (25 mg/kg, oral). the area under the plasmaconcentration time curve (auc) of irinotecan at 20, 40 and 80 mg/kg was 3.51 ± 1.20, 8.81 ± 1.93 and 14.03 ± 2.18 h µg/ml, respectively which after treatment with verapamil, increased dose dependently to 7.84 ± 1.20, 19.94 ± 2.39 and 61.71 ± 15.0 h µg/ml, respectively. in addition to irinotecan, plasma concentrations of sn-38, one of the major active metabolite of irinotecan, were also monitored. the less than proportional increase in sn-38 auc from 20 to 80 mg/kg is consistent with the saturation of carboxylesterase. our results indicate that oral drug treatment of irinotecan in presence of temporary p-gp inhibition could be as equally safe and effective as intravenous administration. neve rtheless, safe p-gp inhibitors need to be identified as alternatives to verapamil for development of efficacious oral irinotecan formulations. keywords irinotecan, bioavailability, verapamil, p-glycoprotein, pharmacokinetics introduction irinotecan is a worldwide approved anti-cancer agent for the treatment of colorectal cancers and other malignancies. it is currently marketed for intravenous use although few reports of oral irinotecan administration exist which demonstrate its low and highly variable oral bioavailability [1-4]. as a result novel formulations of irinotecan exhibiting better oral absorption and bioavailability need to be developed. http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:dr.tripta@gmail.com admet & dmpk 3(1) (2015) 68-76 p-gp pharmacokinetics of irinotecan doi: 10.5599/admet.3.1.154 69 in our previous studies, irinotecan at 1 and 10 µm showed much higher basal-to-apical transport than apical-to-basal transport in caco-2 cells saturable at 100 µm concentrations. this could be due to the active carrier mediated transport of irinotecan by the intestinal drug efflux pump, p-glycoprotein (p-gp). keeping this in view, pharmacokinetic studies of irinotecan were conducted with verapamil in wistar rats. verapamil increased the absolute bioavailability (f) of irinotecan by 4.3 fold and decreased its biliary excretion. it appears that the concomitant and synergistic inhibition of p-gp present in rat intestine and liver is a plausible explanation for prominent increase in oral bioavailability of irinotecan [5]. however, in view of narrow therapeutic index, the co-administration of irinotecan and verapamil may result in unanticipated toxicities. co-administration of p-gp inhibitor would maintain irinotecan plasma concentrations in the same range even at lower doses as compared to when it is administered alone at high dose. we propose that oral drug treatment of irinotecan in presence of temporary p-gp inhibition could be as equally safe and effective as intravenous administration. in the current investigation, the bioavailability and pharmacokinetic parameters were studied in female wistar rats after oral administration of irinotecan alone and in combination with verapamil (25 mg/kg, oral) at three doses (i.e. 20, 40 and 80 mg/kg). the main objective of the study is to evaluate the magnitude of improvement in oral delivery of irinotecan via co-treatment with oral p-gp inhibitor. experimental chemicals and animals irinotecan (>99 %), sn-38 (>96 %), topotecan (>98 %) and irinotecan hydrochloride injection (20 mg ml -1 ) originated from dabur pharma limited (u.p, india). all chemical and reagents were of analytical or hplc grade as appropriate and procured locally. healthy female wistar rats (180-200 g, n=5) used for pharmacokinetic studies were obtained from breeding stock of dabur research foundation (ghaziabad, u.p, india). rats housed in cages were kept in a room under controlled temperature (20-22 °c) and 12 h day-night cycle. animals were used for pharmacokinetic studies after one-week acclimatization with free access to water and feed. all animal procedures were approved by institutional animal ethics committee (dabur research foundation, u.p, india). verapamil (sigma-aldrich, st. louis, mo, usa) solutions were prepared in distilled water containing 5 % dmso. the purity of both irinotecan and verapamil solutions was >96 % and was checked before administration into rats by hplc [5]. in-vivo studies the animals were divided into three different dose levels of 20, 40 and 80 mg/kg. within each dose level, the animals were further sub-divided into a control group and a pre-treated group. rats in control group received irinotecan orally via gavage using a ball-tipped needle. the pretreatment group received oral verapamil (25 mg/kg), 2 h prior to irinotecan administration. in the 20 mg/kg-dose group, three additional rats received the irinotecan only as single dose injection via the lateral tail vein. blood samples were withdrawn prior to dosing and at 0.5, 1, 2, 3, 4, 6, 8, 10, 12 and 24 h post-dosing from retro-orbital plexus into microtubes containing the anticoagulant [5]. plasma was obtained immediately and processed by protein precipitation. samples were stored at -80 °c until hplc analysis. garg et al. admet & dmpk 3(1) (2015) 68-76 70 analytical assay procedures the concentrations of irinotecan and sn-38 in plasma were determined using high performance liquid chromatographic (hplc) method using ultraviolet detection (λmax = 254 nm, 380 nm) as described previously [6]. the sample pre-treatment from plasma involved a single protein precipitation step with cold acetonitrile. topotecan, a structurally related camptothecin, was used as an internal standard. method was found to be selective, linear (r 2 ≈0.999), accurate (recovery ± 15 %) and precise (<5 % c.v.) in the selected concentration ranges for both the analytes. the quantification limit for irinotecan was 40 ng ml -1 and for sn-38 was 25 ng ml -1 [6]. pharmacokinetic data analysis pharmacokinetic parameters were calculated by non-compartment model using winnon-lin 5.0 programme (pharsight, mountain view, ca, usa). the plasma irinotecan and sn-38 concentration-versus-time curves were used to determine maximum plasma concentration (cmax), time to achieve maximum plasma concentration (tmax), mean residence time (mrt), area under the concentration time curve to the respective sampling point (auc0-t), volume of distribution (vd), elimination rate constant (kel), half-life (t1/2) and total body clearance. c0 was the initial plasma concentration of drugs obtained by back-extrapolation to y-axis. the absolute bioavailability (f) of irinotecan after the oral administration (80 mg/kg) compared to the intravenous (i.v) administration (20 mg/kg) was calculated as follows: oral iv auc i.v dose x 100 auc oral dose f  (1) statistical analysis data of five different experiments/animals was reported as mean ± standard error of means (s.e.m) unless otherwise noted. statistical analysis was performed using graphpad prism software version 4.0 (san diego, ca, usa). student’s unpaired t-test was used to test the significance of differences between the controls and treated groups. differences between the concentration time profiles over the entire range tested were analyzed by two-way anova (bonferoni post test). the differences were considered to be significant at p < 0.05. results plasma concentration versus time curves of irinotecan following oral administration at three dose levels (20, 40 and 80 mg/kg) in the absence and presence of a concomitant verapamil oral dose (25 mg/kg) are depicted in figure 1. tables 1, 2 and 3 show the pharmacokinetic parameters for each dose and administration route. the plasma concentrations of irinotecan following oral administration reached peak values at the 2-3 h, indicating that the absorption of irinotecan is from the intestine. irinotecan was absorbed rapidly after oral administration and the observed time to peak irinotecan and sn-38 levels was within 2 h of administration. after an initial absorption phase, plasma concentrations of irinotecan declined admet & dmpk 3(1) (2015) 68-76 p-gp pharmacokinetics of irinotecan doi: 10.5599/admet.3.1.154 71 monoexponentially. verapamil administration was associated with an increase in irinotecan and sn-38 plasma concentrations following both i.v and oral administration. oral irinotecan concentrations increased significantly leading to pronounced alteration in the pharmacokinetics. the overall auc was raised paralleled with a reduction in irinotecan cl and increase in half-life (p < 0.05). the oral bioavailability of irinotecan was 30-40 % without verapamil, which was increased significantly when given with verapamil. figure 1. mean plasma irinotecan (cpt-11) concentration-time plot for a) 20, b) 40 and c) 80 mg/kg after oral administration in control (-●-) and in verapamil (ver) treated (-■-) rats. verapamil at 25 mg/kg was administered orally 2 h prior to irinotecan dosing. each data point represents the mean ± sem of five different rats. **statistical significant difference between control and verapamil treated rats, p<0.01 a) b) c) ** ** ** ** ** ** ** garg et al. admet & dmpk 3(1) (2015) 68-76 72 table 1. pharmacokinetic parameters of irinotecan with and without pre-treatment with oral verapamil (25 mg/kg), after oral and intravenous (20 mg/kg) administration of irinotecan in rats [5]. pk parameters of irinotecan irinotecan dose 20 mg/kg, i.v 20 mg/kg, oral control control treated change auc0-last (h µg/ml) 10.76 ± 2.0 3.51 ± 1.20 7.84 ± 1.20* +220 % c0 (µg/ml) 9.51 ± 1.4 -- cmax (µg/ml) 2.47 ± 0.41 0.63 ± 0.27 1.05 ± 0.23* +66 % tmax (h) -3.21 ± 0.95 2.94 ± 1.17 mrt (h) -5.52 ± 0.20 5.49 ± 0.11 t1/2 (h) 3.1 ± 0.60 3.24 ± 0.22 3.11 ± 0.06 clobs (ml/h/kg) 1206.4 ± 159.7 5664.8 ± 363.9 2573.8 ± 498.4* -54 % vss,obs (ml/kg 4852.2 ± 703.8 27368.4 ± 2440.9 12115.4 ± 3105.5* -56 % kel (h -1 ) 0.232 ± 0.05 0.214 ± 0.0.02 0.214 ± 0.0.02 f (%) 32 72 each value represents the mean ± sem of five rats. *statistical significant difference between control and verapamil treated rats, p<0.05 table 2. pharmacokinetic parameters of irinotecan with and without pre-treatment with oral verapamil (25 mg/kg), after oral (40 mg/kg) administration of irinotecan in rats. pk parameters of irinotecan irinotecan dose 40 mg/kg, oral control treated change auc0-last (h*µg/ml) 8.81 ± 1.93 19.94 ± 2.39* +126% cmax (µg/ml) 1.48 ± 0.57 2.24 ± 0.54 +51% tmax (h) 2.33 ± 0.58 4.67 ± 1.15 mrt (h) 5.21 ± 0.63 6.19 ± 1.10 t1/2 (h) 4.25 ± 2.08 3.59 ± 1.84 clobs (ml/h/kg) 4578.7 ± 1073.02 1981.9 ± 299.7* -56% vss,obs (ml/kg 26088.9 ± 6232.1 9761.5 ± 3385.9* -62% kel (h -1 ) 0.186 ± 0.07 0.223 ± 0.09 f (%) 41 93 each value represents the mean ± sem of five rats. *statistical significant difference between control and verapamil treated rats, p<0.05 table 3. pharmacokinetic parameters of irinotecan with and without pre-treatment with oral verapamil (25 mg/kg), after oral (80 mg/kg) administration of irinotecan in rats [5]. pk parameters of irinotecan irinotecan dose 80 mg/kg, oral control treated change auc0-last (h*µg/ml) 14.03 ± 2.18 61.71 ± 15.0** +440% cmax (µg/ml) 2.93 ± 0.37 10.75 ± 1.0* +266% tmax (h) 2.6 ± 0.89 1.75 ± 1.8 mrt (h) 3.6 ± 0.69 5.13 ± 1.48 t1/2 (h) 2.24 ± 0.51 4.18 ± 1.2* clobs (ml/h/kg) 5613.8 ± 1126.3 897.09 ± 177.9* -84% vss,obs (ml/kg 32197.0 ± 8067.1 7571.1 ± 151.3* -76% kel (h -1 ) 0.293 ± 0.09 0.17 ± 0.05* f (%) 33 143 each value represents the mean ± sem of five rats. *statistical significant difference between control and verapamil treated rats, p<0.05 admet & dmpk 3(1) (2015) 68-76 p-gp pharmacokinetics of irinotecan doi: 10.5599/admet.3.1.154 73 the plasma concentrations versus time curves of sn-38, after oral administration of irinotecan at three doses are shown in figure 2. oral pharmacokinetic parameters of sn-38 were significantly modified by verapamil (p < 0.05) (table 4). sn-38 showed biphasic plasma disposition after oral administration, with a terminal half-life of about 1-2 h. figure 2. plasma concentrations of sn-38 following irinotecan (cpt-11) administration in control and verapamil (ver) treated female wistar rats. the control group was given an oral dose of a) 20, b) 40 and c) 80 mg/kg irinotecan only (-♦-). the treated (-■-) were administered verapamil at 25 mg/kg orally 2 h prior to irinotecan dosing. each data point represents the mean ± sem of five different rats. **statistical significant difference between control and verapamil treated rats, p<0.01 a) b) c) ** ** ** garg et al. admet & dmpk 3(1) (2015) 68-76 74 table 4. pharmacokinetic parameters of sn-38, major metabolite of irinotecan, with and without pre-treatment with oral verapamil (25 mg/kg), after oral (20, 40 and 80 mg/kg) and intravenous (20 mg/kg) administration of irinotecan in rats irinotecan dose groups pk parameter of sn-38 auc0-last (h µg/ml) c0 (µg/ml) cmax (µg/ml) tmax (h) t1/2 (h) 20 mg/kg, i.v control 8.7 ± 0.49 2.10 ± 0.61 --1.1 ± 0.16 20 mg/kg, oral control 0.70 ± 0.026 -0.10 ± 0.01 2 2.18 ± 0.69 treated 0.94 ± 0.08* -0.27 ± 0.03* 2 2.83 ± 1.11 change +34 % +170 % 40 mg/kg, oral control 1.76 ± 0.01 -0.26 ± 0.01 1.0 ± 0.87 2.56 ± 0.19 treated 2.30 ± 0.30* -0.56 ± 0.02* 2 2.35 ± 0.73 change +30 % +115 % 80 mg/kg, oral control 1.81 ± 0.30 -0.48 ± 0.07 2.8 ± 0.45 1.47 ± 0.35 treated 2.99 ± 0.34* -0.62 ± 0.06* 4.2 ± 1.64 1.28 ± 0.26 change +65 % +29 % discussion oral anti-cancer chemotherapy has gained wide acceptance and became standard approach for the treatment of cancer due to various advantages such as greater safety and flexibility, more convenient and cost-effectiveness. unfortunately, the majority of anticancer drugs have a low and highly variable oral bioavailability making intravenous route as the only alternative. data available on irinotecan absorption and disposition showed discouraging results with variable absorption, poor efficacy and toxicity profiles [7]. as a result novel formulations of irinotecan exhibiting better oral absorption and bioavailability need to be developed. verapamil is the most extensively characterized p-gp inhibitor and multi-drug resistance (mdr) reversal agent that has entered clinical trials [8]. therefore, the effect of co-administration of verapamil on the oral bioavailability and pharmacokinetics of irinotecan at various doses was the subject of current investigation. the irinotecan auc increased linearly with dose from 20 to 80 mg/kg indicating linear pharmacokinetics. after treatment with verapamil, irinotecan increased linearly with dose from 20 to 40 mg/kg, but from 40 to 80 mg/kg the irinotecan auc increased non-linearly (i.e. twofold increase in dose, sevenfold increase in auc). overall, the oral combination of irinotecan with verapamil was well tolerated in rat without acute toxicities. a disproportionate increase in oral bioavailability was seen at 80 mg/kg with the same dose of p-gp inhibitor (four fold increase in irinotecan dose resulted in sevenfold increase in auc). the increase in systemic exposure to irinotecan in combination with verapamil is of the same order of magnitude as that with high dose of irinotecan. p-gp inhibition may enable chronic oral therapy with irinotecan. however, safe p-gp inhibitors need to be identified as alternatives to verapamil. this was tested by decreasing the dose in the same proportion as clearance of irinotecan is decreased by verapamil administration. irinotecan has been reported to undergo metabolic saturation (non-linear clearance) at higher concentrations which were achievable in this study, by verapamil pre-treatment [9]. the irinotecan admet & dmpk 3(1) (2015) 68-76 p-gp pharmacokinetics of irinotecan doi: 10.5599/admet.3.1.154 75 clearance in the control groups was unaffected by the increase in dose and remained almost constant. the decrease in irinotecan clearance following verapamil pre-treatment is ~50 % at both 20 and 40 mg/kg. however, ~80 % decrease in irinotecan clearance is observed at 80 mg/kg. hydrolysis of irinotecan by carboxylesterase enzyme has been reported to be responsible for the conversion of irinotecan to sn-38 [7]. the detecachievabletion of sn-38 at early time points may be a consequence of the presence of carboxylesterase in rat plasma, first pass metabolism of irinotecan or metabolism of irinotecan by intestinal carboxylesterase, and subsequent absorption of sn-38. the less than proportional increase in sn-38 auc from 20 to 80 mg/kg is consistent with the saturation of carboxylesterase [9]. figure 3. role of p-gp in determining pharmacokinetics of irinotecan (cpt-11) after oral and intravenous (i.v) administration. p-gp is present in intestinal epithelium affecting absorption; in hepatocytes affecting metabolism and at biliary canalicular surface affecting excretion. in addition, gastro-intestinal toxicity of irinotecan is proposed to be caused by biliary excretion of its major active (100-1000 times) metabolite, sn-38; and its glucuronide (sn-38g). p-gp inhibition at both intestine and bile eliminates first pass effect, resulting in increased oral absorption and systemic exposure. diarrhea could also be ameliorated due to inhibition of biliary excretion of metabolite causing decreased accumulation in conclusion, p-gp in the gastro-intestinal mucosa limits the absorption of orally administered xenobiotics and at least in part, due to its high affinity for p-gp efflux pump. the observed effect may be beneficial in a way to develop oral irinotecan dosage forms using safe p-gp inhibitors to improve its oral bioavailability. clinically, concomitant verapamil and irinotecan treatment would allow dose reduction (about 50-80 %) and still achieve comparable exposure of irinotecan. more importantly, the risk of intestinal toxicity could be substantially reduced because of reduced dose and lowered biliary secretion and accumulation (figure 3). the results of this study could be utilized to evaluate different dosing strategies and methods of administration for irinotecan in humans. oral formulations of irinotecan having better efficacy and less toxicity could be developed using appropriate p-gp inhibitor. however, further garg et al. admet & dmpk 3(1) (2015) 68-76 76 studies are required to compare the efficacy of verapamil with other novel p-gp inhibitors to enable pharmacokinetic modulation of irinotecan and its metabolites. references [1] r.l. drengler, j.g. kuhn, l.j. schaaf, g.i. rodriguez, m.a. villalona-calero,l.a. hammond, j.a. stephenson jr, s. hodge, m.a. kraynak, b.a. statona, g.l. elfring, p.k. locker, l.l. miller, d.d. von hoff, m.l. rothenberg, j. clin. oncol. 17(2) (1999) 685-696. [2] n.e. schoemaker, i.e. kuppens, w.w. huinink, p. lefebvre, j.h. beijnen, s. assadourian, g.j. sanderink, j.h. schellens, cancer chemother. pharmacol. 55(3) (2005) 263-270. [3] o. soepenberg, h. dumez, j. verweij, d. semiond, m.j. de jonge, f.a. eskens, j. ter steeg, j. selleslach, s. assadourian, g.j. sanderink, a. sparreboom, a.t. van oosterom, j. clin. oncol. 23(4) (2005) 889-898. [4] c.f. stewart, w.c. zamboni, w.r. crom, p.j. houghton, cancer chemother. pharmacol. 40(3) (1997) 259-265. [5] t. bansal, g. mishra, m. jaggi, r.k. khar, s. talegaonkar, eur. j. pharm. sci. 36(4-5) (2009) 580-590. [6] t. bansal, a. awasthi, m. jaggi, r.k. khar, s. talegaonkar, talanta 76(5) (2008) 1015-1021. [7] r.h. mathijssen, r.j. van alphen, j. verweij, w.j. loos, k. nooter, g. stoter, a. sparreboom, clin. cancer res. 7(8) (2001) 2182-2194 [8] r. perez-tomas, curr. med. chem. 13(16) (2006) 1859-1876. [9] l.p. rivory, j.r. bowles, j. robert, s.m. pond, biochemical pharmacology, 52(7) (1996) 1103–1111. ©2015 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ investigating antimicrobial features and drug interactions of sedoanalgesics in intensive care unit: an experimental study doi: http://dx.doi.org/10.5599/admet.1042 219 admet & dmpk 9(3) (2021) 219-226; doi: https://doi.org/10.5599/admet.1042 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper investigating antimicrobial features and drug interactions of sedoanalgesics in intensive care unit: an experimental study ozge unlu 1 , emre sertac bingul 2 , sevgi kesici 3 *, mehmet demirci 4 1 istanbul atlas university, faculty of medicine, department of medical microbiology, turkey 2 recep tayyip erdogan university, training and research hospital, department of anesthesiology and reanimation, turkey 3 health sciences university, hqamidiye etfal training and research hospital, department of anesthesiology and reanimation, turkey 4 kirklareli university, faculty of medicine, department of medical microbiology, turkey *sevgi kesici: e-mail: md.kesici@mynet.com ; tel.: +902124426649 received: july 03, 2021; revised: august 25, 2021; available online: september 06, 2021 abstract study objective: aim of this study was to evaluate antimicrobial effects and interaction between analgesic combinations of fentanyl citrate, dexmedetomidine hydrochloride and tramadol hydrochloride on staphylococcus aureus, klebsiella pneumoniae, escherichia coli, pseudomonas aeruginosa and candida albicans which are some of the most common nosocomial infection related microorganisms. design: in vitro prospective study. setting: university clinical microbiology laboratory. measurements: in order to evaluate in vitro antimicrobial effects and interaction between analgesic combinations, tramadol hydrochloride, fentanyl citrate and dexmedetomidin were used against s. aureus atcc 29213, k. pneumoniae, e. coli atcc 25922, p. aeruginosa atcc 27853 and c. albicans atcc 10231 standard strains by microdilution method. main results: according to microdilution assays tramadol has shown the most efficient antimicrobial activity also it has been observed that 10 g/ml concentrated dexmedetomidine has antimicrobial effects on s. aureus, k. pneumoniae and p. aeruginosa. fentanyl has displayed evident inhibitory potency on the pathogens except for klebsiella pneumoniae, nevertheless our predefined minimum concentration inhibited growth by 9.5 %. fentanyl and dexmedetomidine together exhibited more antimicrobial effect on p. aeruginosa and e. coli growth. additionally, when the three drugs examined together, microbial inhibition occurred more than expected on e. coli again and also on c. albicans growth. conclusions: our results revealed the antimicrobial properties and synergy with the different combinations of fentanyl, dexmedetomidine and tramadol against the most common nosocomial infection agents in the icu. this is the first study in the literature looking into the microbial “interactions” of opioids and sedative drugs but more research is needed in order to define clinico-laboratory correlation. ©2021 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords sedoanalgesics; nosocomial agents; antimicrobial effects; drug interactions. http://dx.doi.org/10.5599/admet.1042 https://doi.org/10.5599/admet.1042 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:md.kesici@mynet.com http://creativecommons.org/licenses/by/4.0/ sevgi kesici et al. admet & dmpk 9(3) (2021) 219-226 220 introduction intensive care patients often require sedation procedures for several reasons during their hospitalization. since intensive care units (icu) are prone to nosocomial infections; intravenous (iv) medications gain more importance in the aspect of microbial features. contamination based infections may complicate clinical follow-up and little is known whether the sedative agents have antimicrobial properties. among them, propofol is a well-studied drug that is now known to be inducing bacterial growth [1–4]. however, there are many other drug combinations used in routine practice to maintain sedation. -2 selective blocker dexmedetomidine has become more popular and widely used for its less respiratory depressant effects instead of midazolam which triggers delirium in the elderly [5–7]. many clinicians also add opioids to the treatment just to deepen sedation and fentanyl appears to be one of the most common choices due to its hemodynamic stable properties. there are studies comparing dexmedetomidine and midazolam from the microbiological aspect in the literature [8,9]. but since midazolam has gained a bad reputation for delirium triggering properties another sedative agent fentanyl is yet to be explored. this study aimed to evaluate antimicrobial effects and interaction between analgesic combinations of fentanyl citrate, dexmedetomidine hydrochloride and tramadol hydrochloride on staphylococcus aureus, klebsiella pneumoniae, escherichia coli, pseudomonas aeruginosa and candida albicans which are some of the most common nosocomial infection related microorganisms. these drugs were chosen for being the most frequently used infusion sedoanalgesics in our institute. materials and methods tested sedoanalgesics tramadol hydrochloride (ultramex), fentanyl citrate (talinat) and dexmedetomidin (hipnodex) were tested to reveal their antimicrobial activity and interaction between analgesic combinations. bacterial strains s. aureus atcc 29213, k. pneumoniae, e. coli atcc 25922, p. aeruginosa atcc 27853 and c. albicans atcc 10231 standard strains, which are primary pathogens of bacteremia and sepsis were chosen in order to compare analgesics antimicrobial activity. antibacterial broth test antimicrobial activity of the selected analgesics was measured with microdilution method according to the instructions of clinical laboratory standards institute (clsi). tramadol hydrochloride (abbott, rocky mount, nc, usa, supplied in a 2 ml ampoule containing 50 µg/ml in sodium chloride solution) was diluted with brain heart infusion broth (bhi) to the final concentrations of 500 µg/ml, 250 µg/ml, 125 µg/ml, 62.5 µg/ml, 31.25 µg/ml. fentanyl citrate (abbott, rocky mount, nc, usa, supplied in a 10 ml ampoule containing 0,05 mg/ml in sodium chloride solution) was diluted with brain heart infusion broth (bhi) to the final concentrations of 5 µg/ml, 2.5 µg/ml, 1.25 µg/ml, 0.625 µg/ml, 0.3125 µg/ml. dexmedetomidine hydrochloride (abbott, rocky mount, nc, usa, supplied in a 2 ml ampoule containing 100 µg/ml in sodium chloride solution) was diluted with brain heart infusion broth (bhi) to the final concentrations of 10 µg/ml, 5 µg/ml, 2.5 µg/ml, 1.25 µg/ml, 0.625 µg/ml. overnight broth cultures of s. aureus atcc 29213, k. pneumoniae atcc 13883, e. coli atcc 25922, p. aeruginosa atcc 27853 and c. albicans atcc 10231 were adjusted to the turbidity of a 0.5 mcfarland standard. 10 µl of each strain were inoculated to each well. broths without any analgesic materials were served as controls for comparison. plates incubated overnight and bacterial growth in each well were measured with epoch spectrophotometer (biotek, germany) at 600 nm. admet & dmpk 9(3) (2021) 219-226 are we ready to design oral protacs®? doi: http://dx.doi.org/10.5599/admet.1042 221 interaction between analgesic combinations via broth microdilution method tramadol hydrochloride (th), fentanyl citrate (fc) and dexmedetomidine hydrochloride (dh) combinations were diluted in brain heart infusion broth (bhi) with or without s. aureus atcc 29213, k. pneumoniae atcc 13883, e. coli atcc 25922, p. aeruginosa atcc 27853 and c. albicans atcc 10231 were adjusted to the turbidity of a 0.5 mcfarland standard. 10 µl of each strain were inoculated to each well. broths without any analgesic materials were served as controls for comparison. plates were incubated overnight and bacterial growth in each well were measured with epoch spectrophotometer (biotek, germany) at 600 nm. the preparation of the microdilution in microwells with single agent concentrations of the analgesic to determine the minimum inhibitory concentration (mic), control wells and paired combinations of the analgesics were used in unique concentrations to determine the fractional inhibitory concentration (fic). the value of the ∑fic index is then used to determine whether synergism, indifference or antagonism occurred between the antimicrobial agents and it was used to interpret the nature of the interactions: synergism ≤0.5, indifference >0.5 to ≤4, antagonism >4 according to american society of microbiology [10]. results antimicrobial activity for analgesics antimicrobial susceptibility results revealed that 500 µg/ml tramadol hydrochloride inhibited the growth of k. pneumoniae 100 %. also, inhibition rates for p. aeruginosa, s. aureus, e. coli, and c. albicans were 40, 35, 28.8 and 37.5 %, respectively. moreover, the inhibitory effects of tramadol hydrochloride were observed at descending rates until the lowest concentration tested which was 31,25 µg/ml for k. pneumoniae, e. coli, and c. albicans. in addition, 5 µg/ml fentanyl citrate inhibited the growth of s. aureus, p. aeruginosa, e. coli, and c. albicans at the following rates 38, 13.3, 4.5 and 7.5 %, respectively. also, fentanyl citrate did not show any inhibitory effect on the growth of k. pneumoniae at maximum concentration tested, whereas it inhibited growth 9.5 % in its minimum concentration (3,125 µg/ml). moreover, 10 µg/ml dexmedetomidine had an inhibitory effect on the growth of s. aureus, k. pneumoniae and p. aeruginosa at the rates of 34, 21 and 12.1 %, respectively. also, dexmedetomidine did not show any inhibitory effect on the growth of e.coli at maximum concentration tested, whereas it inhibited growth 5 % in its minimum concentration. moreover, it has been observed that the growth of c. albicans inhibited 8.4 % at the maximum concentration of the analgesic while it has been inhibited 28 % at the minimum concentration tested. all the results of microdilution assays for tramadol hydrochloride, fentanyl citrate and dexmedetomidine were given in table 1, table 2 and table 3, respectively. table 1. antimicrobial effects of tramadol hydrochloride in different concentrations against to tested microorganisms microorganisms 500 µg/ml 250 µg/ml 125 µg/ml 62.5 µg/ml 31.25 µg/ml k. pneumoniae 100.00 % 80.02 % 58.41 % 58.06 % 60.63 % p. aeruginosa 40.03 % 0.00 % 0.00 % 0.00 % 0.00 % e. coli 28.74 % 28.74 % 20.49 % 7.77 % 3.89 % s. aureus 34.62 % 33.46 % 26.25 % 7.21 % 0.00 % c. albicans 35.55 % 26.39 % 27.47 % 27.11 % 27.65 % http://dx.doi.org/10.5599/admet.1042 sevgi kesici et al. admet & dmpk 9(3) (2021) 219-226 222 table 2. antimicrobial effects of fentanyl citrate in different concentrations against to tested microorganisms microorganisms 5 µg/ml 2.5 µg/ml 1.25 µg/ml 0.625 µg/ml 0.3125 µg/ml k. pneumoniae 0.00 % 4.09 % 6.78 % 6.89 % 9.58 % p. aeruginosa 13.34 % 8.67 % 4.85 % 1.04 % 0.69 % e. coli 4.59 % 6.60 % 6.95 % 6.48 % 7.77 % s. aureus 38.22 % 39.64 % 34.75 % 21.24 % 1.93 % c. albicans 7.54 % 26.57 % 36.45 % 38.24 % 30.70 % table 3. antimicrobial effects of dexmedetomidine in different concentrations against to tested microorganisms microorganisms 10 µg/ml 5 µg/ml 2.5 µg/ml 1.25 µg/ml 0.625 µg/ml k. pneumoniae 21.14 % 8.53 % 3.04 % 0.00 % 0.00 % p. aeruginosa 12.13 % 13.00 % 12.11 % 5.37 % 0.69 % e. coli 0.00 % 6.83 % 6.12 % 3.42 % 4.83 % s. aureus 33.98 % 26.00 % 33.20 % 20.21 % 6.31 % c. albicans 8.44 % 28.73 % 33.75 % 24.78 % 20.83 % according to microdilution results, tramadol hydrochloride had the most powerful effect against bacterial growth. moreover, antimicrobial susceptibility results suggest that antimicrobial effect of analgesics vary according to the type of the analgesic, pathogen and the concentration of the material. also, higher concentration does not always result higher antimicrobial effect. interaction between analgesic combinations ∑fic index for tramadol hydrochloride, fentanyl citrate and dexmedetomidine hydrochloride combinations against standard strains were shown in table 4. table 4. ∑fic index for tramadol hydrochloride, fentanyl citrate and dexmedetomidine hydrochloride combinations against standard strains. ∑fic th+fc th+dh fc + dh th+fc+dh c. albicans 1.17 0.58 1.53 0.45 e. coli 1.46 0.64 0.23 0.29 s. aureus 1.17 0.90 1.34 0.89 k. pneumoniae 0.72 0.90 3.23 0.53 p. aeruginosa 1.69 1.60 0.43 0.74 * tramadol hydrochloride (th), fentanyl citrate (fc) and dexmedetomidine hydrochloride (dh) when the interactions between analgesic materials were analyzed, it was found that fc + dh combination had a synergistic antimicrobial effect on the growth of e. coli and p. aeruginosa standard strains. moreover, it has been observed that th + fc + dh combination had a synergistic antimicrobial effect on e. coli and c. albicans standard strains. discussion in this study, we have observed that 10 g/ml concentrated dexmedetomidine has antimicrobial effects on s.aureus, k. pneumoniae and p. aeruginosa. these findings were slightly different from a previous study in which researchers have established the same activity with higher (32 g/ml) concentrations [10]. even though dexmedetomidine is a more preferred sedative drug than midazolam recently; antimicrobial effects were not described as well as it has been studied for midazolam. dexmedetomidine seems to inhibit bacterial growth when prepared in correct concentrations. commonly used infusions are prepared at a concentration of 4 g/ml and according to our results, 10 g/ml infusion fluids might be sufficient enough admet & dmpk 9(3) (2021) 219-226 are we ready to design oral protacs®? doi: http://dx.doi.org/10.5599/admet.1042 223 to prevent bacterial colonization. one should declare that keles et al. [9] have not found any antimicrobial effect of dexmedetomidine. however, this difference may be caused by the chosen microbiological assays. we have performed spectrophotometry after microdilution in bhi which is more sensitive than incubation in blood agar that keles and her colleagues chose. dexmedetomidine hydrochloride does not contain preservatives or stabilizers. therefore the inhibitor effect might be related to its chemical structure. the actual mechanism is unclear and also no microbial growth due to dexmedetomidine was declared in the literature until now [8,9]. in addition to its antimicrobial effect, taniguchi et al. [11] have demonstrated anti-inflammatory features of this drug in rats. according to this study, 2-adrenoceptor agonist has modulated interleukin and tnf- responses reducing excessive neutrophil accumulation on lung tissue. with that knowledge, dexmedetomidine might be a more appropriate drug than midazolam in septic patients who are in need of analgesia and sedation which requires further clinical studies. fentanyl citrate has displayed evident inhibitory potency on the pathogens except for klebsiella pneumoniae, nevertheless our predefined minimum concentration inhibited growth by 9.5 %. to our knowledge, there are few studies focused on fentanyl’s antimicrobial properties and some of them oppose our results [12]. for instance; rota et al. [13] were not able to demonstrate the inhibiting effect of fentanyl. in contrast, graystone et al. [14] have investigated 15 medications that are used in icu and concluded all of them except propofol were bactericidal including fentanyl. based on spectrophotometric calculations our study method was not designed to investigate the bactericidal effects of the drugs. however, as interpreted in the mentioned studies, we may propose specific assumptions. fentanyl citrate ampoules contain 500 g/10 ml which is a large dose for one use only. therefore, multi-dosing with the same syringe is a quite common application practice in the icu settings [12-15]. it could be argued that the drug would degrade and become susceptible to contamination during the time kept at room temperature in the same syringe. however, fentanyl is a physicochemically stable drug and can be stored in the syringe up to 1 week as a mixture with other drugs such as midazolam under 32 °c [16]. refilling of the syringes with this drug does not cause microbial contamination as observed in one prospective longitudinal study [12]. the reliability of multi-dosing with the same syringe is often questioned by the clinicians and considering our results this approach seems adequate. perhaps further clinical studies are required. a fentanyl analog “remifentanil” was investigated by apan et al. [17] in the same manner. they have observed similar microbial inhibition on s. aureus, p. aeruginosa, e. coli and c. albicans depending on the concentration. however, this effect was attributed to “glycine” which is present in remifentanil ampoules as a preservative. it might be possible that glycine has bacteriostatic potency but since both remifentanil and fentanyl are 4-anilidopiperidine structured drugs, further evaluations are needed including also alfentanil and sufentanil [18,19]. fentanyl citrate does not contain preservatives. tramadol is known to be a “good” analgesic that has few respiratory and hemodynamic adverse effects [20]. despite being a weak opioid analgesic, it provides adequate analgesia especially in elder patients who are prone to cognitive dysfunction [21]. therefore, it is mostly used for the postoperative period especially after orthopedic surgery in intensive care units. many clinicians do not reckon tramadol as a sedative agent. but when combined with dexmedetomidine, these drugs show remarkable synergistic effects [20]. as a striking result of our study tramadol has shown the most efficient antimicrobial activity. from this aspect preparing combined infusions of dexmedetomidine and tramadol is rather considerable due to their fewer side-effect properties. it is worth noting that the synergy occurs only on sedation and analgesia but not on http://dx.doi.org/10.5599/admet.1042 sevgi kesici et al. admet & dmpk 9(3) (2021) 219-226 224 bacteriostatic activity. to the best of our knowledge, this is the only study examining antimicrobial features of tramadol. intravenous treatments are generally applied via central venous catheters in the icus. occasionally multiple sedative drugs are needed to be given from the same catheter line and usually infused at lower rates. whether these drug combinations covering the same line attenuate catheter related nosocomial infections is questionable. we have observed evident synergy with the different combinations against particular microorganisms in our study. fentanyl and dexmedetomidine together exhibited more antimicrobial effects on p. aeruginosa and e. coli growth. additionally, when the three drugs were examined together, microbial inhibition occurred more than expected on e. coli again and also on c. albicans growth. this interaction needs to be evaluated more. since this is an in vitro study, in vivo examinations are necessary to adapt clinical practice. as far as we know there is no other study in the literature looking into the microbial “interactions” of opioids and sedative drugs. we believe these interactions are remarkably important which may affect routine iv therapy and shape the practitioner’s approach. as a well-known fact, critical patients receiving total parenteral nutrition via the central venous catheters are at great risk for candidemia. the combination of th, fc and dh infusion is worth investigating from this aspect. it is difficult to come to a conclusion with our laboratory results for sure but further in vitro and in vivo studies may reveal clinically significant interactions. perhaps the most important limitation of our study was the lack of hourly made spectrophotometric calculations. therefore we could not demonstrate microbial inhibition or growth trend during 24 hours. interestingly, graystone et al. [14] have observed an increase in s. aureus number during the first 4 hours which declined later with all opioids except fentanyl. pyrogenic reactions are reported because of extrinsically contaminated fluids even if there is no bacteremia [22]. bacterial load in infusion fluids may be reduced by choosing the appropriate agent and concentration to enhance patient comfort in both sedation and anti-inflammatory meanings. conclusions we have observed definite antimicrobial properties and synergy with the different combinations of fentanyl, dexmedetomidine and tramadol against the most common nosocomial infection agents in the icu. this is the first study in the literature looking into the microbial “interactions” of opioids and sedative drugs but more research is needed in order to define clinico-laboratory correlation. limitations of this work are: (i) the study is in vitro study and (ii) clinical practice is required. conflict of interest: the authors declare no conflict of interest. references [1] m.b. sosis, b. braverman. growth of staphylococcus aureus in four intravenous anesthetics. anesth. analg. 77 (1993) 766-768. [2] c.b. berry, t. gillespie, j. hood, n.b. scott. growth of micro-organisms in solutions of intravenous anaesthetic agents. anaesthesia 48 (1993) 30-32. https://doi.org/10.1111/j.1365-2044.1993.tb06786.x [3] j. crowther, j. hrazdil, d.t. jolly, j.c. galbraith, m. greacen, m. grace. growth of microorganisms in propofol, thiopental, and a 1:1 mixture of propofol and thiopental. anesth. analg. 82 (1996) 475-478. https://dx.doi.org/10.1097/00000539-199603000-00008. https://doi.org/10.1111/j.1365-2044.1993.tb06786.x https://doi.org/10.1111/j.1365-2044.1993.tb06786.x https://dx.doi.org/10.1097/00000539-199603000-00008 admet & dmpk 9(3) (2021) 219-226 are we ready to design oral protacs®? doi: http://dx.doi.org/10.5599/admet.1042 225 [4] d.v. thomas. propofol supports bacterial growth. br. j. anaesth. 66 (1991) 274. https://dx.doi.org/ 10.1093/bja/66.2.274-a. [5] schroeder me. delirium in the elderly. in: pain [internet]. cham: springer international publishing; 2019 [cited 2019 jul 10]. p. 1117–20. available from: http://link.springer.com/10.1007/978-3-31999124-5_238. [6] f. santana santos, l.o. wahlund, f. varli, i. tadeu velasco, m. eriksdotter jonhagen. incidence, clinical features and subtypes of delirium in elderly patients treated for hip fractures. dement. geriatr. cogn. disord. 20 (2005) 231-237. https://dx.doi.org/ 10.1159/000087311. [7] n. bhana, k.l. goa, k.j. mcclellan. dexmedetomidine. drugs 59 (2000) 263-268; discussion 269-270. https://dx.doi.org/ 10.2165/00003495-200059020-00012. [8] h. ayoglu, c. kulah, i. turan. antimicrobial effects of two anaesthetic agents: dexmedetomidine and midazolam. anaesth. intensive care 36 (2008) 681-684. https://dx.doi.org/ 10.1177/0310057x0803600508. [9] g.t. keleş, s. kurutepe, d. tok, h. gazi, g. dinç. comparison of antimicrobial effects of dexmedetomidine and etomidate-lipuro with those of propofol and midazolam. eur. j. anaesthesiol. 23 (2006) 1037-1040. https://dx.doi.org/10.1017/s0265021506000949. [10] m.g. botelho. fractional inhibitory concentration index of combinations of antibacterial agents against cariogenic organisms. j. dent. 28 (2000) 565-570. https://dx.doi.org/10.1016/s0300-5712(00)00039-7. [11] t. taniguchi, y. kidani, h. kanakura, y. takemoto, k. yamamoto. effects of dexmedetomidine on mortality rate and inflammatory responses to endotoxin-induced shock in rats. crit. care med. [internet] 32 (2000) 1322–1326. https://doi.org/10.1097/01.ccm.0000128579.84228.2a. [12] g. morgan. microbial growth in a mixture of hyperbaric bupivacaine and fentanyl prepared in a multi-dose syringe in the operating theatre environment. south african j. anaesth. analg. 16 (2010) 17–21. https://doi.org/10.1080/22201173.2010.10872661. [13] s. rota, k. kaya, o. timliothlu, o. karaca, s. yzdep, e. ocal. do the opioids have an antibacterial effect. can. j. anaesth. 44 (1997) 679-680. https://dx.doi.org/ 10.1007/bf03015460. [14] s. graystone, m.f. wells, d.j. farrell. do intensive care drug infusions support microbial growth. anaesth. intensive care 25 (1997) 640-642. https://dx.doi.org/ 10.1177/0310057x9702500608. [15] s. segal, a. gunawan, d.h. mclaughlin, e. palavecino. microbial stability of syringes of anesthetic drugs prepared in the operating room. j. clin. anesth. 55 (2019) 20-23. https://dx.doi.org/ 10.1016/j.jclinane.2018.12.029. [16] g.m. peterson, k.a. miller, j.g. galloway, p.f. dunne. compatibility and stability of fentanyl admixtures in polypropylene syringes. j. clin. pharm. ther. 23 (1998) 67-72. https://dx.doi.org/ 10.1046/j.1365-2710.1998.00141.x. [17] t.z. apan, a. apan, s. sahin, m. cakirca. antibacterial activity of remifentanil and mixtures of remifentanil and propofol. j. clin. anesth. 19 (2007) 346-350. https://dx.doi.org/10.1016/j.jclinane.2007.02.005. [18] a. obayashi, s. oie, a. kamiya. microbial viability in preparations packaged for single use. biol. pharm. bull. 26 (2003) 667-670. https://dx.doi.org/10.1248/bpb.26.667. [19] r.s. vardanyan, v.j. hruby. fentanyl-related compounds and derivatives: current status and future prospects for pharmaceutical applications. future med. chem. 6 (2014) 385-412. https://dx.doi.org/10.4155/fmc.13.215. [20] c.r. lee, d. mctavish, e.m. sorkin. tramadol. a preliminary review of its pharmacodynamic and pharmacokinetic properties, and therapeutic potential in acute and chronic pain states. drugs 46 (1993) 313-340. https://dx.doi.org/10.2165/00003495-199346020-00008. [21] e. freye, j.v. levy. the effects of tramadol on pain relief, fast eeg-power spectrum and cognitive function in elderly patients with chronic osteoarthritis (oa). acute pain 8 (2006) 55–61. https://doi.org/10.1016/j.acpain.2006.03.001. http://dx.doi.org/10.5599/admet.1042 https://dx.doi.org/%2010.1093/bja/66.2.274-a https://dx.doi.org/%2010.1093/bja/66.2.274-a http://link.springer.com/10.1007/978-3-319-99124-5_238 http://link.springer.com/10.1007/978-3-319-99124-5_238 https://dx.doi.org/%2010.1159/000087311 https://dx.doi.org/%2010.2165/00003495-200059020-00012 https://dx.doi.org/%2010.1177/0310057x0803600508 https://dx.doi.org/%2010.1177/0310057x0803600508 https://dx.doi.org/10.1017/s0265021506000949 https://dx.doi.org/10.1016/s0300-5712(00)00039-7 https://dx.doi.org/10.1016/s0300-5712(00)00039-7 https://doi.org/10.1097/01.ccm.0000128579.84228.2a https://doi.org/10.1080/22201173.2010.10872661 https://dx.doi.org/%2010.1007/bf03015460 https://dx.doi.org/%2010.1177/0310057x9702500608 https://dx.doi.org/%2010.1016/j.jclinane.2018.12.029 https://dx.doi.org/%2010.1016/j.jclinane.2018.12.029 https://dx.doi.org/10.1046/j.1365-2710.1998.00141.x https://dx.doi.org/10.1046/j.1365-2710.1998.00141.x https://dx.doi.org/10.1016/j.jclinane.2007.02.005 https://dx.doi.org/10.1016/j.jclinane.2007.02.005 https://dx.doi.org/10.1248/bpb.26.667 https://dx.doi.org/10.4155/fmc.13.215 https://dx.doi.org/10.4155/fmc.13.215 https://dx.doi.org/10.2165/00003495-199346020-00008 https://doi.org/10.1016/j.acpain.2006.03.001 https://doi.org/10.1016/j.acpain.2006.03.001 sevgi kesici et al. admet & dmpk 9(3) (2021) 219-226 226 [22] s.n. bennett, m.m. mcneil, l.a. bland, m.j. arduino, m.e. villarino, d.m. perrotta, d.r. burwen, s.f. welbel, d.a. pegues, l. stroud. postoperative infections traced to contamination of an intravenous anesthetic, propofol. n. engl. j. med. 333 (1995) 147-154. https://dx.doi.org/10.1056/nejm199507203330303. ©2021 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) https://dx.doi.org/10.1056/nejm199507203330303 https://dx.doi.org/10.1056/nejm199507203330303 http://creativecommons.org/licenses/by/3.0/ manuscript doi: 10.5599/admet.3.1.163 34 admet & dmpk 3(1) (2015) 34-44; doi: 10.5599/admet.3.1.163 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper characterization of slc transporters in human skin marion alriquet, karine sevin, alexandre gaborit, pierre comby, bernard ruty and hanan osman-ponchet * galderma r&d, les templiers , 2400 route des colles, bp 87, f-06902 sophia-antipolis *corresponding author: e-mail: hanan.osman-ponchet@galderma.com; tel.: +33-492-386-786; fax: +33-493-957-071 received: january 30, 2015; revised: march 19, 2015; published: march 31, 2015 abstract most identified drug transporters belong to the atp-binding cassette (abc) and solute carrier (slc) families. recent research indicates that some of these transporters play an important role in the absorption, distribution and excretion of drugs, and are involved in clinically relevant drug-drug interactions for systemic drugs. however, very little is known about the role of drug transporters in human skin in the disposition of topically applied drugs and their involvement in drug-drug interactions. the aim of this work was to compare the expression in human skin (vs human hepatocytes and kidney) of slc transporters included in the ema guidance as the most likely clinical sources of drug interactions. the expression of slc transporters in human tissues was analyzed by quantitative rt-pcr. modulation of slc47a1 and slc47a2 (mate1 and mate2) expression was analyzed after treatment of human skin in organ-culture with rifampicin and uv irradiation. the expression of slco2b1 (oatpb), slco3a1 (oatpd), slco4a1 (oatpe), slc47a1 and slc47a2 (mate1 and mate2) was detected in human skin, oatpe and mate1 being the most expressed. oatpe is about 70 times more expressed in human skin than in human hepatocytes. moreover, the expression of slc47a1 and slc47a2 was down-regulated after treatment with rifampicin or after exposure to uv light. the present findings demonstrate that slco4a1 (oatpe) and slc47a1 (mate1) are highly expressed in human skin and suggest the involvement of slc transporters in the disposition of topically applied drugs. keywords solute-carrier transporter; mate1; uv light; skin absorption introduction the human skin, one of the largest organ in the organism with about 6% of total body weight (5-6 kg, 2 m 2 in area), acts as a physical and biochemical barrier to protect the body from environmental factors and also plays a pivotal role in homeostasis by preventing the uncontrolled loss of water [1]. although constitutive expression of xenobiotic-metabolizing enzymes has been detected in normal keratinocytes [2,3] the levels of drug-metabolizing enzymes are generally much lower in the skin than those in the liver and intestine [4-7]. thus, skin permeability rather than drug metabolism appears to be the major barrier to topical bioavailability [8,9]. http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:hanan.osman-ponchet@galderma.com admet & dmpk 3(1) (2015) 34-44 expression of slc transporters in human skin doi: 10.5599/admet.3.1.163 35 most identified drug transporters belong to the atp-binding cassette (abc) and solute carrier (slc) families. xenobiotic transporters (abcs and slcs) have broad specificity and are involved in both uptake (influx) and secretion (efflux) of their substrates, thereby affecting their cellular disposition. constitutive expression of multidrug-resistance associated proteins (mrp, abcc) or solute carrier organic anion transporting polypeptides (oatps) has been detected in human keratinocytes [2,3,10,11]. the expression of abcb1 (multidrug resistance protein 1, p-glycoprotein or p-gp) has only been found after keratinocyte treatment with dexamethasone [2]. current knowledge about the function of xenobiotic transporters in the skin is limited. in addition to their role as a contributor to multidrug resistance, it has been demonstrated that p-gp as well as mrp1 help maintain a healthy immune response in the skin as they play an important role in the migration of langerhans cells out of the skin by way of the lymphatic vessels [12]. in addition, these transporters also play a role in the active efflux of contact allergens and endogenous compounds such as steroid hormones from normal human epidermal keratinocytes (nhek) [13]. further studies demonstrated p-gp involvement in migration and invasion of resistant melanoma cells [14], whereas very low or no expression of p-gp was detected in primary human melanocytes [15]. recently the expression pattern of human atp-binding cassette transporters was reported in human skin [16]. in addition the characterization of abc transporters in human skin was reported recently by our group [17]. the present study was not intended to be an exhaustive search for slc transporter expression in human skin. this has been reported in human keratinocytes by schiffer et al. [3] and in human skin by li et al. [18] and more recently by fujiwara et al. [19]. it should be noted that mate1 and mate2 transporters were not included in those studies. the aim of our study was to compare the expression in human skin (vs human hepatocytes and kidney) of slc transporters included in the ema guidance [20] as the most likely clinical sources of drug interactions (slco1b1, slco1b3, slc22a1, slc22a2, slc22a6, slc22a8, slc47a1 and slc47a2). this list of slc transporters is essentially the same as the one reported in the recent review by maeda and sugiyama [21]: list which is based on the fda guidance [22] to which these authors added mate1 and mate2 transporters. three others transporters slco2b1 (oatpb), slco3a1 (oatpd) and slco4a1 (oatpe) were included as a control as they are known to be expressed in human skin [3,18,19]. furthermore, the modulation of expression of slc47a1 and slc47a2 (mate family) transporters by rifampicin and uv light exposure was also evaluated in human skin in organ-culture. experimental chemicals and reagents rifampicin was from sigma-aldrich (st louis, mo, usa). total rna from human kidneys came from life technologies (carlsbad, ca, usa). skin organ-culture fresh healthy human skin surgical waste samples excised during cosmetic surgery were collected and used with the consent of the patients. skin tissues were handled under aseptic condition and processed with sterile dissection tools. after removing adipose tissue and hypodermis with surgical scissors, skin biopsies of 6 mm diameter were taken using sterile biopsy punch tools, and transferred to a 6-well plate alriquet m admet & dmpk 3(1) (2015) 34-44 36 filled with long term skin culture medium (biopredic, france). this proprietary long-term skin culture medium has been developed specifically to maintain the survival of human skin explants. the specific composition of this medium (fetal calf serum, antibiotics, pituitary bovine extract, sodium pyruvate, growth factors and hormones such as insulin and epidermal growth factor, hydrocortisone and selenium) allows the preservation of dermal and epidermal cell populations in normal metabolic conditions for at least 3 days. four biopsies per well were used (figure 1). the culture plate was maintained in a humidified incubator at 37 °c with 5 % co2. figure 1. skin organ-culture. skin biopsies were taken using biopsy punch and transferred to a 6-well plate filled with culture medium. treatment with rifampicin skin biopsies were treated during 72 hours with 20 µm rifampicin, refreshing culture medium every 24 hours. control skin biopsies were cultured in the same conditions but rifampicin was omitted in the vehicle (i.e. 0.1 % dmso). during incubation, culture plate was placed in a humidified incubator at 37 °c with 5 % co2. skins from at least three different individuals were used in each experiment unless mentioned. treatment with sunlight simulator skin biopsies were exposed during one hour every day during 3 days in a sunlight simulator (model 91293; oriel instruments) equipped with a 1000-w xenon arc lamp power supply set at 850 w. the sunlight simulator dispensed a mix of uva at 110 w/m² and uvb at 20 w/m² which corresponds to approximately 40 j/cm² uva and 3.6 j/cm² uvb for a 1 hour exposure period. control skin biopsies were cultured in the same conditions but without exposure in the sunlight simulator. human hepatocytes in primary culture fresh human hepatocytes in primary culture were purchased from biopredic (france). hepatocytes (identification number hh0711b) were seeded at 1.8 x 10 6 cells/well in 6-well plate. each well was filled with 2 ml spfm culture medium (gibco, france). hepatocyte culture was maintained during 3 days in a humidified incubator set at 37 °c with 5 % co2. culture medium was refreshed every day. in some experiments, hepatocytes were treated with 20 µm rifampicin during 72 hours with refreshing culture medium every 24 hours. cryopreserved human hepatocytes cryopreserved human hepatocytes were used to evaluate constitutive expression of slc transporters. cryopreserved human hepatocytes, batch number hue26c from life technologies, consisted of a pool of 26 different donors. admet & dmpk 3(1) (2015) 34-44 expression of slc transporters in human skin doi: 10.5599/admet.3.1.163 37 total rna isolation after homogenization of skin samples or hepatocytes in lysis buffer (promega, france), total rna was isolated using sv total rna isolation system (promega, france) according to the instructions provided by the supplier. rna concentrations were quantified at 260 nm using biotek synergy 2 multi-mode microplate. human kidney total rna human kidney total rna samples were purchased from life technologies. two different samples of total rna from two women were provided at final concentration 1 mg/ml in 1 mm sodium citrate, ph 6.4. rna samples were stored at –80 °c. real time pcr expression of eleven human slc transporters (slco1b1, slco1b3, slco2b1, slco3a1, slco4a1, slc22a1, slc22a2, slc22a6, slc22a8, slc47a1 and slc47a2) and 4 nuclear receptors (ahr, nr113, nr112 and rxra) was measured by real-time pcr techniques. experiments were carried out on a 7500 real-time pcr system (applied biosystems) using taqman assay-on-demand gene expression products. for this, 500 ng of total rna were reverse-transcribed using the high capacity rna to cdna master mix kit (applied biosystems). table 1. assay-on-demand used in the gene expression profile experiments. gene/protein assay-on-demand reference slco1b1/oatp1b1 hs00272374_m1 slco1b3/oatp1b3 hs00251986_m1 slco2b1 (slc21a9)/oatpb hs00939778_m1 slco3a1 (slc21a11)/oatpd hs00983988_m1 slco4a1 (slc21a12)/oatpe hs01030343_m1 slc22a1/oct1 hs00427552_m1 slc22a2/oct2 hs01010723_m1 slc22a6/oat1 hs00537914_m1 slc22a8/oat3 hs00188599_m1 slc47a1/mate1 hs00217320_m1 slc47a2/mate2 hs00945650_m1 ahr/ahr hs00169233_m1 nr1i3/car hs00231959_m1 nr1i2/pxr hs01114267_m1 rxra/rxrα hs01067636_m1 gapdh hs99999905_m1 alriquet m admet & dmpk 3(1) (2015) 34-44 38 pcr amplifications were performed in a total volume of 25 µl using the taqman universal master mix (applied biosystems). denaturation was performed at 95 °c for 10 min, followed by 40 pcr cycles with the following specifications: 95 °c 15 s and 60 °c 60 s. glyceraldehyde-3-phosphate dehydrogenase (gapdh) was used as a reference gene for normalization in each sample. the references of primer and taqman probe sequences used in the expression profiling experiments are indicated in table 1. all real-time pcr measurements were performed in triplicate. the comparative threshold cycle (ct) method, also called delta ct method was used to analyse the real time pcr data. statistical analysis all experiments were performed at least three times. data are expressed as mean ± sem. student’s t-test was used to determine the statistical significance of differences between two groups. a p value of less than 0.05 was considered as statistically significant. results gene expression of slc transporters in human skin the expression of 11 slc transporters (8 of clinically importance in drug-drug interactions and three controls) was measured in human skin (three different donors) by real time pcr (figure 2a). among the 11 tested transporters, the expression of six slc transporters was not detected in human skin (oatp1b1, oatp1b3, oct1, oct2, oat1 and oat3) while the expression of 5 slc transporters was observed. oatpb, oatpd and mate2 were found to be poorly expressed while two slc transporters oatpe and mate1 were found to be highly expressed. data from our laboratory showed that expression level of mate1 transporters gene was not markedly modified in cultured skin samples (ct 26.6) compared to tissues without culture (ct 25.1) (data not shown). these results show that some of slc transporter involved in clinically relevant drug-drug interactions for systemic drugs are expressed in human skin and that human skin in organ-culture is a useful tool to study the involvement of slc transporters in the absorption and distribution of topically applied drugs. comparison of expression of slc transporters in human skin, hepatocytes and kidney liver and kidney tissues were used as a positive control because strong expression of slc transporters has been described previously in these organs [23]. the results show that the expression profiles of slc transporters are specific to the organ considered. in human cryopreserved hepatocytes oct1 and oatp1b1 transporters are the most strongly expressed (figure 2b; middle) while in human kidney (figure 2c; bottom), it is oat1 and oat3. moreover, oatpe expression in human skin is 70 times higher than in human hepatocytes and as high as in human kidney. effect of rifampicin on gene expression of slc transporters in human skin previous studies showed that treatment with rifampicin decreased the expression level of abcc1 (mrp1) in human skin [17] and in hepg2 cells [24]. the effect of rifampicin on the expression of slc transporters was thus investigated. after 72 hours treatment of skin samples with rifampicin 20 µm, gene expression of slc transporters was measured by real time pcr. admet & dmpk 3(1) (2015) 34-44 expression of slc transporters in human skin doi: 10.5599/admet.3.1.163 39 a: human skin (data from three donors) b: human hepatocytes (pool of 26 donors) c: human kidney (data from 2 donors) figure 2. constitutive expression of some slc transporters in human skin (a), human hepatocytes (b) and human kidney (c). quantitative real-time reverse transcription-pcr was used to investigate the constitutive expression of slc transporters in human tissues. gene expression was normalized to gapdh to ensure equality of loading. all analyses were performed in triplicate. results are means and standard error of 3 individual skin samples from 3 different donors, one hepatocyte pool of 26 donors, or 2 kidney mrna from 2 donors. all measures were done in triplicate. alriquet m admet & dmpk 3(1) (2015) 34-44 40 the results presented in figure 3 show that treatment of skin samples with rifampicin strongly decreased the expression of mate1 and mate2 transporters in human skin (figure 3a; top). indeed, rifampicin triggers a 44 % decrease in mate1 expression (p < 0.05) and a 30 % decrease in mate2 expression. however, treatment of skin samples with rifampicin had no effect on the expression level of oatpb, oatpd and oatpe (data not shown). the same experiment was performed with fresh human hepatocytes in primary culture treated with rifampicin 20 µm for 72 hours. the results presented in figure 3b bottom, show that the expression level of mate1 in human hepatocytes decreased by 48 % after being treated with rifampicin (data performed in triplicate on one single donor). overall, it means that the expression of mate1 and mate2 transporters in human skin is regulated by rifampicin and that the same phenomenon occurs in fresh human hepatocytes with mate1. a: human skin b: human hepatocytes figure 3. effect of rifampicin on gene expression of slc47a1 (mate1) and slc47a2 (mate2) in human skin (a) and human hepatocytes (b). human skin biopsies in organ-culture (three donors) and human hepatocytes in primary culture (one donor) were treated with rifampicin (20 µm) during 72 hours. expression of slc47a1 and slc47a2 was analysed by quantitative real-time reverse transcription-pcr. expression of mrna was normalized to gapdh. all analyses were performed in triplicate. data are mean and sd of three donors (human skin). all measures were done in triplicate. * statistically significant (p < 0.05). rifampicin regulates ahr and rxra expression in human skin it has been shown that the expression of abc transporter mdr1 and mrp2 is induced by rifampicin via pxr and car nuclear receptors [25]. as a consequence, we investigated the involvement of nuclear receptors such as pxr, car, ahr and rxrα in regulation of slc transporters by rifampicin. the expression admet & dmpk 3(1) (2015) 34-44 expression of slc transporters in human skin doi: 10.5599/admet.3.1.163 41 of these nuclear receptors was measured by real time pcr in skin samples treated with 20 µm rifampicin for 72 hours and compared to control untreated skin samples. the constitutive expression of the four nuclear receptors in human skin, hepatocytes and kidney (figure 4a, top) shows that unlike in hepatocytes, the expression of car and pxr was not detected in human skin (ct above 35) and very low in kidney. however, the expression of ahr and rxrα was detected in human skin but at lower levels compared to hepatocytes and kidney. moreover, expression of ahr and rxrα in human skin was down regulated by rifampicin, just like mate1 and mate2 transporters (figure 4b; bottom). indeed, the expression of ahr and rxrα decreased by 32% and 33% when skin biopsies were treated with rifampicin compared to untreated controls. this indicates that ahr and rxrα could be involved in mate1 and mate2 down regulation by rifampicin. a: constitutive expression of nuclear receptors b: effect of rifampicin on nuclear receptors figure 4. constitutive expression of nuclear receptor in human hepatocytes, human kidney, and human skin (a), and effect of rifampicin on the expression of rxrα and ahr in human skin. (a) results are means and standard error of 2 individual skin samples from 2 different donors, one hepatocyte pool of 26 donors, or 2 kidney mrna from 2 donors. (b) results from one single donor. all measures were done in triplicate. effect of sunlight simulator on the expression of slc transporters in human skin the effect of exposure to uv light on the expression of slc transporters was investigated in two different donors. human skin biopsies were maintained in organ-culture for three days as explained in the experimental section and exposed to uv light for 1 hour every 24 hours using a solar simulator. control alriquet m admet & dmpk 3(1) (2015) 34-44 42 samples were not exposed to uv lights. the expression of the slc transporters found to be expressed in human skin was measured by real time pcr. the results presented in figure 5 show that following the exposure of skin biopsies to uv light, the expression level of mate1 and mate2 decreased by 43% (p < 0.05) and 60%, respectively. the expression of slco3a1 and slco4a1 was not modified by exposure of skin biopsies to uv light (data not shown). figure5. effect of uv light exposure on the expression of slc47a1 and slc47a2 in human skin. results are means and standard error of 2 individual skin samples from 2 different donors. all measures were done in triplicate. * statistically significant (p < 0.05). discussion transporters are recognized as important actors of drug absorption, distribution and excretion. it has also been demonstrated that they are involved in drug-drug interactions, turning them into elements to be studied in drug development. abc and slc transporters have been well studied in liver and kidney which are responsible for most drug excretion. however, their expression and role in human skin are poorly understood. in the present study, we have observed the expression of oatpb, oatpd, oatpe, mate1 and mate2 transporters in human skin: oatpe and mate1 transporters being the most highly expressed. a recent study has shown that slc transporters were expressed in the human skin to a moderate and variable extent, but mate transporter expression was not investigated [19]. to our knowledge it is first time that there is evidence for the expression of mate transporter in human skin. mate transporters are found in many living organisms and are involved in multidrug resistance. in human, mate1 is expressed in both liver and kidney while mate2 is kidney-specific [26]. those transporters play a role in renal and biliary excretion of endogenous and exogenous compounds including xenobiotics and drugs [26,27]. although mate1 transporter belongs to slc family, it is an efflux transporter and its activity depends on physiological ph. mate1 is thought to be responsible for the last step of excretion of organic cations involving proton exchange [27]. by showing that mate1 is strongly expressed in human skin, these data show its possible involvement in absorption, distribution and excretion of topically applied drugs. moreover, this transporter is down regulated by rifampicin which suggests that it might play a role in drug-drug interactions. further investigations are needed to characterize the functional role of mate transporters in human skin. admet & dmpk 3(1) (2015) 34-44 expression of slc transporters in human skin doi: 10.5599/admet.3.1.163 43 furthermore, oatpe is poorly known but is highly expressed in human skin. it would be very interesting to better characterize this transporter and to investigate its role in drug disposition and to understand its function. the results presented in this work showed that some slc transporters have a specific expression profile in human skin. mate1 transporter is highly expressed in human skin compared to other slc transporters, suggesting that mate1 may play a role in topical drug exposure and drug-drug interactions. the extensive expression of mate1 transporter in human skin and the down-regulation by solar simulator exposure suggest that this transporter might serve as a biomarker for skin cancers, particularly those induced by solar exposure. furthermore, the expression of mate1 transporter in human skin may play an important role in chemosensitivity of cutaneous cancer cells and thus may be the basis for the discovery of novel agents for the treatment of skin cancers [28]. conclusions our present findings suggest that oatpe and mate1 transporters are highly expressed in human skin and that mate1 is down-regulated by exposure to uv light, but investigations are needed to identify the potential role of oatpe and mate1 transporters in the skin disposition of topically applied drugs. references [1] p.m. elias, current allergy asthma reports, 8 (2008) 299-305. [2] j.m. baron, d. holler, r. schiffer, s. frankenberg, m. neis, h.f. merck, f.k. jugert, journal of investigative dermatology, 116 (2001) 541-548. [3] r. schiffer, m. neis, d. höller, f. rodriguez, a. geier, c. gartung, f. lammert, a. dreuw, g. zwadloklarwasser, h. merck, f. jugert, j.m. baron journal of investigative dermatology, 120 (2003) 285291. [4] h. schaeffer, t. redelmeier, skin barrier: principles of percutaneous absorption, s. karger, basel, ch, 1996 . [5] h. swanson, chemico-biological interactions, 149, (2004) 69-79. [6] a. williams, transdermal and topical drug delivery, pharmaceutical press, london, 2003. [7] m. pham, j. magdalou, m. totis, s. fournel-gigleux, g. siest, b. hammock, biochemical pharmacology, 38 (1989) 2187-2194. [8] k. madison, journal of investigative dermatology, 121 (2003) 231-241. [9] m. prausnitz, s. mitragotri, r. langer, nature reviews drug discovery, 3 (2004) 115-124. [10] m.a. sleeman, j.d. watson, j.g. murison, journal of investigative dermatology, 115 (2000) 19-23. [11] a. dreuw, h.m. hermanns, r. heise, s. joussen, f. rodriguez, y. marquardt, f. jugert, h.f. merck, p.c. hinrich, j.m. baron, journal of investigative dermatology, 124 (2005) 28-37. [12] g.j. randolph, s. beaulieu, m. pope, i. sugarawa, l. hoffman, r.m. steinman, w.a. muller, proceedings of the national academy of sciences usa, 95 (1998) 6924-6929. [13] r. heise, c. skazik, f. rodriguez, s. tanzel, y. marquardt, s. joussen, a.f. wendel, m. wosnitza, h.f. merck, j.m. baron, journal of investigative dermatology, 130 (2010) 305-308. [14] m. colone, a. calcabrini, l. toccacieli, g. bozzuto, a. stringaro, m. gentile, m. cianfriglia, a. ciervo, m. caraglia, a. budillon, g. meo, g. arancia, a. molinari, journal of investigative dermatology, 128 (2008) 957-971. [15] s. heimerl, a.k. bosserhoff, t. langmann, j. ecker, g. schmitz, melanoma research, 17 (2007) 265273. [16] s. takenaka, t itoh, r. fujiwara, pharmacology research & perspectives 1 (2013) 1-8. alriquet m admet & dmpk 3(1) (2015) 34-44 44 [17] h. osman-ponchet, a. boulai, m. kouidhi, k. sevin, m. alriquet, a. gaborit, b. bertino, p. comby, b. ruty, drug metabolism and drug interactions, 29 (2014) 91-100. [18] q. li, h. tsuji, y. kato, y. sai, y. kubo, a. tsuji, journal of controlled release 110 (2006) 542-56. [19] r. fujiwara, s. takenada, m. hashimoto, t. narawa, t, itoh, scientific reports, 4 (2014) 1-8. [20] european medicines agency (ema), 2012, guideline on the investigation of drug interactions. [21] k. maeda, y. sugiyama, molecular aspects of medicine 34 (2013) 711-718. [22] us food and drug administration (fda), 2012. guidance for industry drug interaction studies— study design, data analysis, implications for dosing, and labeling recommendations. [23] f.g.m. russel, in enzyme and transporter-based drug-drug interactions. progress and future challenges, k.s. pang, a.d. rodrigues, r.m. peter, springer, 2010, p. 27-49. [24] s. haenisch, s. laechelt, h. bruckmueller, a. werk, a. noack, o. bruhn, c. remmler, i. cascorbi, molecular pharmacology, 80 (2011) 314-320. [25] m. sun, w. cui, s.k. woody, j.l. staudinger, drug metabolism and disposition, 43 (2015) 335-43. [26] h. omote, m. hiasa, t. matsumoto, m. otsuka, y. moriyama, trends in pharmacological sciences, 27 (2006) 587—593. [27] m. otsuka, t. matsumoto, r. morimoto, s. arioka, h. omote, y. moriyama, proceedings of the national academy of sciences usa, 102 (2005) 17923-17928. [28] h. osman-ponchet, (galderma research & development), patent wo-2014/184265 a. ©2015 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ proall-d: protein allergen detection using long short term memory a deep learning approach doi: https://doi.org/10.5599/admet.1335 231 admet & dmpk 10(3) (2022) 231-240; doi: https://doi.org/10.5599/admet.1335 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper proall-d: protein allergen detection using long short term memory a deep learning approach pallavi m. shanthappa*, rakshitha kumar* department of computer science, amrita school of arts and sciences, mysuru campus, amrita vishwa vidyapeetham, india *corresponding authors: e-mail: palls.ms@gmail.com; rakshitha.k.k1999@gmail.com. received: april 03, 2022; revised: july 15, 2022; available online: august 21, 2022 abstract background: an allergic reaction is the immune system's overreacting to a previously encountered, typically benign molecule, frequently a protein. allergy reactions can result in rashes, itching, mucous membrane swelling, asthma, coughing, and other bizarre symptoms. to anticipate allergies, a wide range of principles and methods have been applied in bioinformatics. the sequence similarity approach's positive predictive value is very low and ineffective for methods based on fao/who criteria, making it difficult to predict possible allergens. method: this work advocated the use of a deep learning model lstm (long short-term memory) to overcome the limitations of traditional approaches and machine learning lower performance models in predicting the allergenicity of dietary proteins. a total of 2,427 allergens and 2,427 non-allergens, from a variety of sources, including the central science laboratory and the ncbi are used. the data was divided 80:20 for training and testing purposes. these techniques have all been implemented in python. to describe the protein sequences of allergens and non-allergens, five e-descriptors were used. e1 (hydrophilic character of peptides), e2 (length), e3(propensity to form helices), e4(abundance and dispersion), and e5 (propensity of beta strands) are used to make the variable-length protein sequence to uniform length using acc transformation. a total of eight machine learning techniques have been taken into consideration. results: the gaussian naive bayes as accuracy of 64.14 %, radius neighbour's classifier with 49.2 %, bagging classifier was 85.8 %, ada boost was 76.9 %, linear discriminant analysis has 76.13 %, quadratic discriminant analysis was 84.2 %, extra tree classifier was 90%, and lstm is 91.5 %. conclusion: as the lstm, has an auc value of 91.5 % is regarded best in predicting allergens. a web server called proall-d has been created that successfully identifies novel allergens using the lstm approach. users can use the link https://doi.org/10.17632/tjmt97xpjf.1 to access the proall-d server and data. ©2022 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords allergen prediction; acc transformation; lstm model; gaussian naive bayes; classifier; extra tree classifier; bagging classifier; ada boost; linear discriminant analysis; quadratic discriminant analysis introduction allergy, often described as an autoimmune disorder, is a clinical condition characterized by the immune system’s sensitivity to normally innocuous elements. the substance that causes allergy is known as an allergen. allergens can be dust, pollen, cosmetics, and food. in food, allergy is usually caused by proteins. proteins are an essential part of our diet, but some proteins can also be harmful to some individuals. one of https://doi.org/10.5599/admet.1335 https://doi.org/10.5599/admet.1335 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:palls.ms@gmail.com mailto:rakshitha.k.k1999@gmail.com https://doi.org/10.17632/tjmt97xpjf.1 http://creativecommons.org/licenses/by/4.0/ pallavi and rakshitha admet & dmpk 10(3) (2022) 231-240 232 the reasons behind this is that nowadays, the use of genetically modified crops that are transgenic food crops is increasing rapidly. thus, it is necessary to assess them before they are introduced into the food chain. allergy can be innate, acquired, predictable, and at times rapid. allergic reactions are caused by an antibody called immunoglobulin e (ige), which causes hyperactivity in white blood cells such as mast cells and basophils, resulting in the production of inflammatory chemicals like histamine. apart from symptoms such as uneasiness, sneezing, wheezing, and swelling, allergic reactions can also lead to life-threatening situations. as a result, assessing them is critical to protect society's wellbeing. according to the food and agriculture organization, a protein is a potential allergen if it has a homology of six successive amino acids or a sequence identity of more than 35 percent [1]. poms et al. developed pcr (polymerase chain reaction) in conjunction with elisa (enzyme-linked immunosorbent assay) to find potential allergens in foods or an indicator to detect the existence of the offending foods [2]. algpred uses meme/mast motif search to predict allergens and svm for classification based on single and dipeptide composition [3]. allerhunter uses svm as the classifying method and an incremental pairwise sequence comparison indexing approach to identify probable allergens and allergic cross-reactivity in proteins. the paired vectorization system models the essential elements of allergens that are involved in cross-reactivity [4]. using pseudo-amino acid composition (pseaac) and svm, a new technique for identifying and predicting allergenic proteins was developed. it looked at sequence vector representations derived from sequence attributes. the minimum reliability and maximal significance feature selection approach were used to assess the impact and efficiency of each feature [5]. vijaykumar et al. developed an innovative fuzzy rule-based approach to investigate protein allergenicity when the similarity between known allergens and non-allergens is low for characterizing allergens. the results of five different modules were combined: computational classifier, pattern analysis, global comparison with allergens, fao management framework, and prototype approach [6]. allertop was developed as an alignment-free allergen prediction method. protein properties were defined using z-descriptors. the acc transformation was used to convert the variable-length strings to uniformlength strings. the knn (k-nearest neighbor) algorithm was applied for classification and consistently outperformed other algorithms [7]. allergenfpwas designed for distinguishing allergens and non-allergens, and a sequence descriptor-based fingerprint technology was presented. the strings of varying lengths were transformed into arrays of similar lengths using the acc transformation. the results were compared using tanimoto coefficients followed by the transformation of vectors to binary fingerprints [8]. for allergenicity prediction, dimitrov et al. developed artificial neural network-based algorithms. as a final step before the ann modeling, the vectors were transformed into binary fingerprints [9]. allertop v2 is a highly accurate allergen prediction model based on amino acid characteristics. the acc procedure was used to transform variable-length strings into uniform-length vectors. in comparison to other classification approaches, the knn algorithm produced a stable output graph [10]. allerdictor is a pattern-based allergy prediction software that interprets sequence data as a textual information and detects allergens through text classification using support vector machines [11]. cross-react was a computational framework approach for predicting allergenic protein’s cross-reactivity. it is based on the hypothesis that surface regions with peptide compositions similar to an antigen in a known allergen can be detected on three-dimensional structures of probable allergens [12]. a study of ge (genetically engineered) crops found they are as safe as conventional food crops. screening the recombinant protein for predicting potential allergens is one of the assessment procedures for ge crops. it implies that there is currently no clear parameter that can be used to anticipate the pathogenicity of proteins [13]. allercatpro analyses potential allergenic protein based on the three-dimensional structural admet & dmpk 10(3) (2022) 231-240 proall-d: protein allergen detection doi: https://doi.org/10.5599/admet.1335 233 similarity. shifting between sequential frame similarities to b-cell epitope-like 3d surface similarity with anticipated architectures was investigated, and so an entropy-adjusted hexamer hit method was also investigated [14]. pallavi et al. used computational analysis to compare three medications and identified the impediment to malignant cells that cause skin cancer. they also used homology modeling to create the 3d framework of a braf(v600e) protein genotype, which they confirmed using the ramachandran plot [15]. aller screener predicts protein allergenicity by analyzing hla binders derived from recognized allergens. by creating binders to hla class ii proteins, it could predict whether a given substance is safe to eat or drink [16]. algpred v 2.0 provides various options, including searching for motifs in proteins found by meme/mast and merci features such as blast-based similarity searches and ige epitope mapping [17]. wang et al. demonstrated the superiority of their proposed technique by using numerous supervised algorithms as baseline classifiers. the greatest auc value was 0.9578 for the deep learning model, which was superior to the ensemble learning and baseline approaches, according to the results of 5-fold cross-validation [18]. this paper comprises the report on the development methods of a set of novel allergen prediction models that use the knowledge gained via a publicly available server: allertop v.2 [10]. we propose the long shortterm memory (lstm) as a rapid model-based recurrent neural network for protein allergen detection. lstm is primarily used to handle long-term dependence problems and is best suited for time series or sequential data. because our dataset, the protein dataset, is similarly in sequential form, lstm would be a better method for classification with enhanced performance. also, a few machine learning and ensemble learning algorithms have been evaluated for comparison and classification purposes. methods protein data sets we gathered a total of 2,427 allergens and 2,427 non-allergens from a variety of sources, including the central science laboratory and the national center for biotechnology information (ncbi). the redundancies were eliminated. a sample protein sequence is shown in figure 1. >gi│83715928│dbj│bae54429.1│ tropomyosin [sepia esculenta] mdaikkkmlamkmekevatdkaeqteqslrdledaknkteedlstlqkkysnlendfdna neqltaantnleasekrvaeceseiqglnrriqlleedlerseerltsaqskledaskaa desergrkvlenrsqgdeeridllekqleeakwiaedadrkfdeaarklaitevdlerae figure 1. sample dataset in fasta format of a protein sequence e-descriptors five e-descriptors were used to characterize the protein sequences of allergens and non-allergens [10]. venkatarajan et al. used principal component analysis to calculate the quantitative descriptor values based on 237 physical-chemical properties of amino acids. pca of amino acid properties extracted 5 orthogonal edescriptors, which are the eigenvectors of the covariance matrix [19]. we have considered the same five edescriptors that were used to define the characteristics of amino acids [10]. e1 denotes the hydrophilic nature of peptides, e2 their length, e3 their tendency for helical formation, e4 their abundance and distribution, and e5 their tendency for β strand formation. auto cross-covariance transformation proteins are composed of amino acid sequences, each distinct and varies in length. so, acc transformation was employed to convert the variable-length sequence to uniform length so that the classification algorithms could be applied to it. here, the 5 e-descriptors have been considered, which were derived from 237 physiochemical properties of amino acids as used by dimitrov et al. [10]. https://doi.org/10.5599/admet.1335 pallavi and rakshitha admet & dmpk 10(3) (2022) 231-240 234 auto cross covariance includes auto covariance and cross covariance. the equation to calculate auto covariance is as follows: j,i j,i+lag jj i ( ) n lag e e acc lag n lag      (1) the equation to calculate cross covariance is as follows: j,i k,i+lag jk i ( ) j k n lag e e acc lag n lag       (2) index j was used for the e-descriptors (j = 1, 2, 3,4,5), n is the number of amino acids in a sequence, the index i is the amino acid position (i = 1, 2, ...n) and l is the lag, length of the minimum sequence (l = 1, 2, ...l). in order to investigate the influence of close amino acid proximity on protein allergenicity, a short range of lags (l = 1, 2, 3, 4, 5) was used [20]. the classification methods adopted in this research for classification, a few machine learning, ensemble learning, and deep learning algorithms have been considered, which include: the gaussian naive bayes, radius neighbour's classifier, bagging classifier, ada boost, linear discriminant analysis, quadratic discriminant analysis, extra tree classifier, and lstm, which have been implemented in python. hochreiter and schmidhuber proposed the term lstm (long short-term memory) in 1997. lstms are a kind of rnn (recurrent neural networks) that aid in the resolution of the long-term dependence problem. a conventional rnn encounters the difficulty of vanishing gradient problems, which makes learning extended sequences difficult. this is where lstm comes in to address the challenge described above. the lstm model is built iteratively. to construct the model, one must add different sorts of layers with varied parameters and experiment with dropout layers. the network constructed here consists of four layers with three relu activation functions and one softmax function. "categorical cross-entropy," was considered as the loss function with "rmsprop" as the optimizer. evaluation of performance for training and testing purposes, the data was split in the ratio 80:20. the accuracy results of the model based on training and testing data has been represented in table 2. true positives (tp) and true negatives (tn) were assigned to the allergens and non-allergens that were accurately predicted. false negatives (fn) and false positives (fp) were assigned to the allergen and non-allergen identified inaccurately. precision [(tp) /(tp + fp)] is the fraction of correctly predicted samples among the retrieved instances. the recall [(tp)/(tp + fn)] is the ratio of accurately identified true positives to total actual true positives. and the f1 score is computed as [(2 *(accuracy * recall)/(precision + recall)]. web server for allergenicity prediction a web server, namely proall-d, has been developed to predict the potential allergens using the lstm algorithm. it is developed using the python django framework, which is fast and user-friendly. the detailed functioning of the webserver has been described in the supplementary section. results and discussion acc includes both autocovariance and cross-covariance. auto covariance is calculated between the same e descriptors, that is between e1 and e1, along with the lag value. ac111 represents the autocovariance admet & dmpk 10(3) (2022) 231-240 proall-d: protein allergen detection doi: https://doi.org/10.5599/admet.1335 235 between e1 and e1 along with the lag value as shown in figure 2. cross covariance is calculated between the different e descriptor values, like between e1 and e2, along with the lag value. the cross-covariance values will be represented as ac121, ac131, ac145, and ac431. the details of acc implementation have been specified in the supplementary data. figure 2. output of acc transformation the classification methods considered in this research are: the gaussian naive bayes, radius neighbour's classifier, bagging classifier, ada boost, linear discriminant analysis, quadratic discriminant analysis, extra tree classifier, and lstm, which have been implemented in python. gaussian naive bayes is a statistical predictive model for classification that is based on the naive bayes algorithm. for our dataset, this algorithm produced an accuracy of 64.14 percent (table 1). the radius neighbours classifier is an extended version of the knn algorithm that produces results using all instances within a range of a new instance rather than the k clusters, which would be beneficial for our dataset, but the model failed to provide the expected results, resulting in an accuracy of 49.2 percent. adaboost, also known as adaptive boosting, is an ensemble method. the weights are reallocated to each instance, with larger weights applied to inaccurately identified instances. boosting is used in supervised learning to minimize bias as well as variation. the accuracy of this model was 76.9 percent. linear and quadratic discriminant analysis resulted in an accuracy of 76.13 and 84.2 percent. a bagging classifier is an ensemble classifier that works on random samples of the data and then combines various instances to obtain the final output. this method resulted in an accuracy of 85.8 percent. extra tree classifier, an extended version of the random forest algorithm, resulted in an accuracy of 90 percent. lstm model resulted in an accuracy of 91.5 percent. table 1. analysis of the effectiveness of different classifiers method accuracy precision recall f1-score gaussian naive bayes 64.14 0.74 0.46 0.56 radius neighbours classifier 49.2 0.49 1.00 0.66 ada boost 76.9 0.79 0.75 0.77 linear discriminant analysis 76.13 0.80 0.71 0.75 quadratic discriminant analysis 84.2 0.93 0.74 0.83 bagging classifier 85.8 0.89 0.86 0.88 extra tree classifier 90 0.95 0.85 0.90 lstm (long short-term memory) 91.5 0.91 0.91 0.91 table 1 represents the results of performance evaluation metrics and accuracy of all the algorithms implemented. it is evident that the lstm approach is superior and has consistent performance across all measures. so, lstm has been considered for protein allergen prediction. https://doi.org/10.5599/admet.1335 pallavi and rakshitha admet & dmpk 10(3) (2022) 231-240 236 table 2. analysis of the classification results of different classifiers based on training and testing data. method training data testing data gaussian naive bayes 63.8 59.7 radius neighbours classifier 50.5 47.6 ada boost 84.6 78.1 linear discriminant analysis 78.6 74.9 quadratic discriminant analysis 88.7 81.6 bagging classifier 88.4 86.3 extra tree classifier 93.4 89.8 lstm (long short-term memory) 94.1 91.5 table 2 represents the accuracy of the algorithms for training and testing dataset. the lstm method was implemented to a well-known benchmark data set for protein allergen identification, in which a protein has to be classified as allergen or non-allergen. lstm delivers highly defined classification performance that is substantially quicker than other algorithms with comparable classification performance. lstm is five times faster than marginal classification algorithms (methods based on distance) and two times faster than the quickest svm-based methods (which have lower classification performance than lstm). all the implemented methods were tested and compared using performance evaluation measures. the top-performing model was lstm, which had an accuracy of 91.51 percent. lstm is a more sophisticated version of the rnn (recurrent neural network). the lstm has been considered for our problem because of its robustness against long-term dependency problems. since the protein sequences are also correlated with each other, lstm would be a likely method for solving long-term dependencies and would overcome the drawbacks of the alignment method. table 3. assessment of web servers for allergenicity prediction in table 3 the performance of the lstm model was compared to nine freely available servers. the lstm resulted in an accuracy of 91.5 percent. a roc curve (short for receiver operating characteristic) plots the rate of true positives vs. false positives to assess the effectiveness of a classification model. auc measures how well a model distinguishes between positive and negative classes. auc is not affected by the classification threshold value. modifying the threshold value does not affect auc because it is an aggregate measure of roc. figure 3 indicates that when the true positive rate increases, so does the false positive rate and after a certain point that is near 0.91 the graph is constant. the area under the roc curve between (0,0) and (1,1) is defined as the area under the curve (auc). auc essentially aggregates the model's performance overall threshold values. the lstm model has the highest auc, indicating that it has the largest area under the curve and is the best model for correctly classifying observations. the software for the server proall-d has been developed to predict potential allergens using the lstm algorithm. it is developed using the python django framework, which is fast and user-friendly. server accuracy allerhunter 0.871 algpred (svm_single_aa) 0.775 algpred (svm_dipeptide) 0.796 algpred(arp) 0.842 appel 0.783 proap(motif) 0.505 proap(svm) 0.843 allertop v.1 0.828 allergenfp 0.879 allertop v.2 0.887 lstm model 0.915 admet & dmpk 10(3) (2022) 231-240 proall-d: protein allergen detection doi: https://doi.org/10.5599/admet.1335 237 figure 3. the lstm model’s roc curve there are three different sections, namely home, datasets, and method description as shown in figure 4. in the home section, the user enters the protein sequence in a one-letter code, and the models predict whether the entered sequence is allergenic or non-allergenic as shown in figure 5. in the dataset part, we have uploaded the data considered in our research in the fasta file format as represented in figure 6. the method description provides the user with a brief description of the methodologies we have considered. figure 4. interface of proall-d for protein allergen detection. https://doi.org/10.5599/admet.1335 pallavi and rakshitha admet & dmpk 10(3) (2022) 231-240 238 figure 5. working of proall-d. figure 6. data-set section the supplementary section describes the detailed functioning of the web application. conclusions this study builds on and expands earlier studies to develop the analysis of potential protein allergens. our first aim has been to update the technique while keeping the previous factors in mind. we evaluated deep learning, ensemble learning, and machine learning models such as the gaussian naive bayes, radius neighbour’s classifier, bagging classifier, ada boost, linear discriminant analysis, quadratic discriminant analysis, and lstm to predict the allergenicity of proteins. extensive testing produced excellent results. they admet & dmpk 10(3) (2022) 231-240 proall-d: protein allergen detection doi: https://doi.org/10.5599/admet.1335 239 were superior and corroborated earlier research. furthermore, the auc value of lstm (the best performance) was 0.9152. so far, this is the only study to apply the aforementioned methods to evaluate protein allergenicity, and it will serve as a paradigm for future protein allergen prediction. conflict of interest: all the authors declare no conflict of interest. references [1] m.b. stadler, b.m. stadler . allergenicity prediction by protein sequence. faseb j. 17 (2003) 1141. https://doi:10.1096/fj.02-1052fje. [2] r.e. poms, e. anklam, m. akuhn . polymerase chain reaction techniques for food allergen detection. journal of aoac international 87 (2004) 1391-1397. https://doi.org/10.1093/jaoac/87.6.1391. [3] s. saha, g.s. raghava. algpred: prediction of allergenic proteins and mapping of ige epitopes. nucleic acids research 34 (2006) w202-w209. https://doi.org/10.1093/nar/gkl343. [4] h.c. muh, j.c. tong. allerhunter: a svm-pairwise system for assessment of allergenicity and allergic cross-reactivity in proteins. plos one 4 (2009) e5861. https://doi.org/10.1371/journal.pone.0005861. [5] h. mohabatkar, b.m. mohammad, k. abdolahi, s. mohsenzadeh. prediction of allergenic proteins by means of the concept of chou's pseudo amino acid composition and a machine learning approach. medicinal chemistry 9 (2013) 133-137. https://doi.org/10.2174/157340613804488341. [6] s. vijayakumar, p.t.v. lakshmi. ieee international conference on bioinformatics and biomedicine, a fuzzy inference system for predicting allergenicity and allergic cross-reactivity in proteins, tongji university, china. (2013) 49-52. https://doi.org/10.1109/bibm.2013.6732458. [7] i. dimitrov, d.r. flower, i. doytchinova. bmc bioinformatics, allertop – a server for in silico prediction of allergens, cambridge, uk, 2013, 1-9. https://doi.org/10.1186/1471-2105-14-s6-s4. [8] i. dimitrov, l. naneva, i. doytchinova, i. bangov. allergenfp: allergenicity prediction by descriptor fingerprints. bioinformatics 30 (2014) 846-851. https://doi.org/10.1093/bioinformatics/btt619. [9] i. dimitrov, l. naneva, i. bangov, i. doytchinova. allergenicity prediction by artificial neural networks. journal of chemometrics 28 (2014) 282-286. https://doi.org/10.1002/cem.2597. [10] i. dimitrov, i. bangov, d.r. flower, i. doytchinova. allertop v.2 – a server for in silico prediction of allergens. journal of molecular modelling 20 (2014) 1-6. https://doi.org/10.1007/s00894-014-2278-5. [11] ha. x. dang, c.b. lawrence. allerdictor: fast allergen prediction using text classification techniques. bioinformatics 30 (2014) 1120-1128. https://doi.org/10.1093/bioinformatics/btu004. [12] s.s. negi, w. braun. cross-react: a new structural bioinformatics method for predicting allergen crossreactivity. bioinformatics 33 (2017) 1014-1020. https://doi.org/10.1093/bioinformatics/btw767. [13] g.s. ladics. assessment of the potential allergenicity of genetically-engineered food crops. journal of immunotoxicology 16 (2019) 43-53. https://doi.org/10.1080/1547691x.2018.1533904. [14] s. maurer-stroh, n.l. krutz, p.s. kern, v. gunalan, m.n. nguyen, v. limviphuvadh, f. eisenhaber, g.f. gerberick. allercatpro-prediction of protein allergenicity potential from the protein sequence. bioinformatics 35 (2019) 3020-3027. https://doi.org/10.1093/bioinformatics/btz029. [15] m.s. pallavi, h.s. pramod kumar. in-silico analysis to determine the efficient drug for malignant melanoma using molecular dynamics. biomedical and pharmacology journal 13 (2020) 1463-1470. https://dx.doi.org/10.13005/bpj/2018. [16] i. dimitrov, m. atanasova. allerscreener – a server for allergenicity and cross-reactivity prediction. cybernetics and information technologies 20 (2020) 175184. https://doi.org/10.2478/cait-2020-0071. [17] n. sharma, s. patiyal, a. dhall, a. pande, c. arora, g.p.s raghava. algpred 2.0: an improved method for predicting allergenic proteins and mapping of ige epitopes. briefings in bioinformatics 22 (2021) bbaa294. https://doi.org/10.1093/bib/bbaa294. https://doi.org/10.5599/admet.1335 https://doi:10.1096/fj.02-1052fje https://doi.org/10.1093/jaoac/87.6.1391 https://doi.org/10.1093/nar/gkl343 https://doi.org/10.1371/journal.pone.0005861 https://doi.org/10.2174/157340613804488341 https://doi.org/10.1109/bibm.2013.6732458 https://doi.org/10.1186/1471-2105-14-s6-s4 https://doi.org/10.1093/bioinformatics/btt619 https://doi.org/10.1002/cem.2597 https://doi.org/10.1007/s00894-014-2278-5 https://doi.org/10.1093/bioinformatics/btu004 https://doi.org/10.1093/bioinformatics/btw767 https://doi.org/10.1080/1547691x.2018.1533904 https://doi.org/10.1093/bioinformatics/btz029 https://dx.doi.org/10.13005/bpj/2018 https://doi.org/10.2478/cait-2020-0071 https://doi.org/10.1093/bib/bbaa294 pallavi and rakshitha admet & dmpk 10(3) (2022) 231-240 240 [18] l. wang, d. niu, x. zhao, x. wang, m. hao, h. che. a comparative analysis of novel deep learning and ensemble learning models to predict the allergenicity of food proteins. foods 10 (2021) 809. https://doi.org/10.3390/foods10040809. [19] m.s. venkatarajan, w. braun. new quantitative descriptors of amino acids based on multidimensional scaling of a large number of physical–chemical properties. molecular modeling annual 7 (2001) 445453. https://doi.org/10.1007/s00894-001-0058-5. [20] i.a. doytchinova, d.r. flower. vaxijen: a server for prediction of protective antigens, tumour antigens and subunit vaccines. bmc bioinformatics 8 (2007) 1-7. https://dx.doi.org/10.1186%2f1471-2105-8-4. ©2022 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) https://doi.org/10.3390/foods10040809 https://doi.org/10.1007/s00894-001-0058-5 https://dx.doi.org/10.1186%2f1471-2105-8-4 http://creativecommons.org/licenses/by/3.0/ newly discovered staphylococcus aureus serine hydrolase probe and drug targets doi: https://dx.doi.org/10.5599/admet.1137 107 admet & dmpk 10(2) (2022) 107-114; doi: https://doi.org/10.5599/admet.1137 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index review newly discovered staphylococcus aureus serine hydrolase probe and drug targets matthias fellner department of biochemistry, school of biomedical sciences, university of otago, dunedin, new zealand. matthias.fellner@otago.ac.nz; tel.: +64 34797897; received: october 10, 2021; revised: october 25, 201; available online: october 28, 2021 abstract there is an urgent need for new diagnosis and treatment options for the bacterial pathogen staphylococcus aureus. this review will summarize data on ten recently discovered biofilm-associated serine hydrolases called fluorophosphonate-binding hydrolases (fpha-j). based on the summarized findings, many of these proteins represent intriguing new targets for probe and drug development. ©2021 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords inhibitor; biofilm; fluorophosphonate introduction antimicrobial resistance presents a major challenge to public health, with the past year accelerating the problem [1, 2]. this is especially true for infections by the bacterial pathogen staphylococcus aureus, which are a major cause of mortality, often linked to community-acquired drug-resistant strains (mrsa) [3]. this calls for an urgent need to find new solutions for effective diagnosis and treatment options to overcome resistance to avoid depletion of our arsenal of antibiotics. new protein targets within s. aureus are required for the development of efficient diagnostic probes both for imaging applications and as a treatment strategy to block productive infection by the bacteria without pressuring the organism to select resistance mutants. a fluorophosphonate-based activity-based probe identified ten previously uncharacterized active serine hydrolases in s. aureus during biofilm-promoting growth conditions, which could fill this need. these were named fluorophoshonate-binding hydrolases (fphs), with sequential letters for each enzyme based on their predicted size (52 kd fpha – 22 kd fphj) [4]. all are members of the α/β hydrolase superfamily, characterized by a core of eight β-strands connected by several α helices with an active site triad of serinehistidineaspartate or glutamate. the nucleophilic serine is used for the hydrolysis of substrates and can easily and specifically be targeted by small molecules [5, 6]. in general, these proteins play important roles in processing of metabolites, peptides and lipids as a means of controlling cell signalling and metabolism; however, to date the biological function remains unknown for all fphs and only the structure of fphf has been determined [6, 7]. their active state during biofilm-forming conditions makes them accessible to modification by chemical inhibitors that can be developed into probes and drugs. such new compounds https://dx.doi.org/10.5599/admet.1137 https://doi.org/10.5599/admet.1137 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:matthias.fellner@otago.ac.nz http://creativecommons.org/licenses/by/4.0/ matthias fellner admet & dmpk 10(2) (2022) 107-114 108 have the potential for diagnosis, monitoring or treatment of s. aureus infections. this review will summarize our current knowledge about each fph gene and fph protein with the gene name of the reference s. aureus strain nctc 8325, also listed in the headline. fph proteins fpha saouhsc_02751 fpha is significantly larger than the other fph proteins at 52kd (fig. 1a; secondary structure figure made with topdraw [8]). the size increase stems from a possible additional β-strand and especially from large extensions of the connecting loops and helices to the core β sheet. fpha is predicted to be a carboxylesterase (fig. 1b, pfam [9]) and structure prediction suggests a narrow substrate profile due to the range of hydrophobic and hydrophilic residues in a small but surface-exposed active site pocket [6]. a library screen of 1,2,3-triazole urea molecules failed to identify a highly specific inhibitor for fpha with fpha showing a broad range of low level of inhibition compared to a more defined profile of specific hits for other investigated fph proteins [10]. fphb saouhsc_02844 fphb stands out with predicted additional n-terminal extensions that contain three large α-helices preceding the core hydrolase fold (fig. 1a and 1b). function and structure predictions do not give a confident suggestion of the role of these helices, with possible roles being membrane association or activesite capping [4, 6]. fphb may have a role in stress responses as transcriptomic datasets showed a moderate upregulation of fphb gene transcription (~1.8–2.9-fold) upon exposure to various cell-wall-acting antibiotics and antimicrobial peptides (locus tags sa2323 [11-13] and sa2549 [14], respectively); to neutrophilassociated azurophilic granules, hydrogen peroxide, hypochlorus acid (locus tag mw2456) [15]; to acid shock (locus tag sa2549) [16]; and antibacterial skin fatty acids (temporary upregulation, locus tag sa2323) [17]. another study showed fphb upregulation comparing s. aureus exponentially and stationary phases in media, in a s9 human bronchial epithelial cell line infection model and a murine pneumonia model [18]. using commercially available fluorogenic substrates revealed that fphb has cleavage-specificity towards esterase substrates with a preference for c4 > c7 > c8 fatty acid esters and failing to cleave c2 or c10 and longer substrates. for fphb, small molecule inhibitor screening identified a chloroisocoumarin-based covalent inhibitor jcp251 [4] and a triazole-based inhibitor aa395 [10], both with specificity over the other fph proteins. jcp251 effectively reduced infectivity in a mouse model, suggesting that it may be a viable therapeutic target for the treatment or management of staphylococcus infections. this inhibitor also had a fast inactivation rate when the esterase activity was tested in vitro. jcp251 was subsequential modified into a fluorescent probe jcp251-bt, which only reduced the inactivation rate 3-fold. in live s. aureus cells, this probe only labelled fphb; it also did not label any targets in several other bacterial pathogens and predominantly labelled fphb in live bacteria cocultured with peripheral mononuclear cells and mouse raw264.7 macrophage cells. this allows for visualization of probe-labelled bacteria over weak background fluorescence caused by nonspecific probe uptake and/or weak labelling of host cell proteins. various bacteria-host cell coculture experiments suggested that s. aureus (strains atcc35556 and newman) regulates fphb activity in response to host-cell-derived signals and that fphb is not secreted into the culture medium. compared with a cytosolic gfp signal, fphb labelling is concentrated in specific regions of the cell envelope, especially in the septal cross-wall, suggesting that fphb might act on substrates located in the bacterial cell wall. however, the fphb-deficient strain had no growth defect in liquid culture and showed a similar doubling time to the wild-type, ruling out a general role in cell division. however, a ~10to admet & dmpk 10(2) (2022) 107-114 new staphylococcus aureus serine hydrolase probe doi: https://dx.doi.org/10.5599/admet.1137 109 100-fold decrease in bacterial loads in liver and heart tissues of infected mice from wild-type to the fphbdeficient strain suggests a critical role in the colonization of these specific organs as no phenotype was observed in kidneys. pre-treatment of wild type with jcp251 showed a similar reduction of bacterial burden in the liver to levels comparable to those observed upon infection with the fphb-deficient strain. overall, fphb is an important virulence factor in the early stages of infections in the heart and liver but not the kidneys, potentially due to different environments during initial colonization. the inhibitory effect of jcp251 demonstrates that fphb is an ideal target for the development of small molecules that could stop the spread of primary infection to other sites that cause increased morbidity and mortality [4]. studies on the 58 % sequence identical fphb homolog from the commensal s. epidermidis showed similar substrate preferences, with an additional ability to process an acetate ester, with inhibition of the protein also not influencing the growth of the bacteria. however, in contrast to fphb in s. aureus, the homolog in s. epidermidis appears to not have a role in colonization in vivo based on preliminary studies in skin-infected mice [19]. fphc saouhsc_01279 fphc and fphg were most difficult to label among the fph proteins with the biotinylated and fluorescent probes used to initially characterize the fph proteins, possibly due to comparable lower abundance and beyond their initial discovery, not much is known about these two hydrolases [4]. interestingly fphc, fphg and fphh are the most conserved fph proteins in the genus staphylococcus [19]. structure prediction suggests that fphc may lack the first core β-strand (fig. 1a) but otherwise shows a typical α/β hydrolase fold with a helix capping the active site [6]. fphc was upregulated comparing s. aureus exponentially and stationary phases in media [18]. fphd saouhsc_01279 fphd is the only fph protein with a significant structural homology hit in the protein data bank, with the 63 % fphd homolog from s. epidermidis (pdb id 3fle, unpublished). this structure shows a large surface exposed active site with mostly hydrophobic residues, however, the n-terminal is missing in the structure, and fphd is predicted to contain several helices here (fig. 1a and 1b). similar to fphb these helices could serve various functions in fphd, however, capping the active site is more likely here due to the large surface exposed area [6]. beyond this, not much is known about fphd, its expression may be linked to fphh based on gene knockout studies [4] and fphd was downregulated in an s9 bronchial epithelial cell infection model [18]. fphe saouhsc_02900 analysis of 50 different species in the genus staphylococcus showed that fphe is the least conserved fph protein, with less than half containing a sequence relevant homolog [19]. studies suggest that fphe could be an s. aureus specific driver of virulence, with a fphe transposon mutant showing a small, though statistically significant, reduction of bacterial loads in the mice livers compared to wild type, with no reduction in the hearts or kidneys [20]. fphe and fphg were the only fph proteins identified to bind βlactam probes [20, 21]. a screening of 1,2,3-triazole urea compounds identified inhibitors for fphe, one was modified into a fluorescent probe, resulting in a highly specific cellular probe to study fphe activity or to sort bacterial populations based on probe-labelling status [10]. in both s. aureus usa300 and newman strains, fphe-specific labelling was highest in cells growing under biofilm-forming conditions compared to exponential growth and in the late stationary phase. throughout all culture stages, fphe labelling was higher in usa300 compared to newman. overall, studies with the fluorescence probe suggest that fphe https://dx.doi.org/10.5599/admet.1137 matthias fellner admet & dmpk 10(2) (2022) 107-114 110 levels are dynamically regulated and are subject to the bacterial growth state and environment in a strainspecific manner [10]. fphe was also shown to be upregulated comparing s. aureus exponentially and stationary phases in media in an s9 human bronchial epithelial cell line infection model while being downregulated in a murine pneumonia model [18]. fphf saouhsc_02962 whatsgnu analysis on all publicly available s. aureus genomes showed an extremely conserved fphf sequence within individual s. aureus clonal complexes but varied between each complex (gnu scores of 2218 or 3370 of 10,350). across bacterial species, putative fphf sequences are highly sequence divergent, while clustering to the expected tree of life [7]. interestingly, some species like the commensal s. epidermidis reference strain rp62a contain two fphf homologs, something only observed for fphf among the fph proteins [19]. fphf is the only fph protein for which the atomic structure has been determined. the crystal structure of fphf shows a tetramer, however, in solution fphf forms a dimer. the overall structure represents a typical α/β hydrolase superfamily fold with the core eight β-strands connected by several α helices (fig. 1c, figure made with ucsf chimera [22]). the active site is formed by the traditional ser-his-asp triad, with apo crystals containing an electron density at the active site interpreted as a sodium ion chelated by two water molecules (pdb id 6vh9). the active site is exposed to the surface, with one side containing a large hydrophobic pocket and the other side having a polar pocket [6]. a later study also independently determined the apo structure of fphf (pdb id 7l0a). magnesium ions were modelled at the active site, similar to the positions of the mentioned sodium. otherwise, both structures closely resemble each other [7]. independent studies have characterized fphf as a carboxylesterase. the first study identified, using fluorogenic substrates, that fphf cleaved lipid ester substrates but was unable to process phosphate, phosphonate, or glycosidic substrates. towards hydrophobic saturated lipid substrates, a promiscuous specificity profile was observed with activity for 4-methylumbelliferyl-acetate to decanoate with the highest activity towards heptanoate [6]. another study showed fphf activity against 4-nitrophenyl-acetate and butyrate with no activity against 4-nitrophenyl-trimethylacetate, and a lack of glyoxalase activity using the substrate s-lactoylglutathione. in addition, fphf was identified to remove one pivaloyloxymethyl moiety from the carboxy ester prodrug pom-hex, with a second moiety removed in concert by the hydroxyacylglutathione hydrolase glob in s. aureus, as well as in vitro, to activate the prodrug. prodrug activation was uniform across the population in all observed bacterial cells, demonstrating ubiquitous expression of fphf. screening of a 32-compound ester substrate library identified that fphf had the highest activity toward oxygen ethers with a preference for unbranched substrates with little regard for chain length or the end-of-chain bulk. branching at the position following the ester carbonyl was deleterious to fphf activity and when oxygen is included in the chain, positioning at the β-position to the carbonyl is strongly preferred over the γ-position [7]. the substrate preference was further confirmed via an atomic substrate-bound structure (pdb id 6wcx). here the heptyl acyl moiety of the preferred model substrate 4-methylumbelliferyl-heptanoate is bound to the active site serine inside the now fully occupied hydrophobic pocket. in conjunction with docking studies, the preference for intermediate sized acyl groups is evident as shorter chains never correctly positioned over the active site and longer chains do not fit [6]. the biological function of fphf is unknown. in various databases, it is currently annotated as an sadmet & dmpk 10(2) (2022) 107-114 new staphylococcus aureus serine hydrolase probe doi: https://dx.doi.org/10.5599/admet.1137 111 formylglutathione hydrolase, which mediates detoxification of cellular formaldehyde, based on studies on the e. coli enzyme frmb [23]. however, the two proteins only share 25 % sequence identity. major differences are evident when comparing the structure of fphf to other s-formylglutathione hydrolases, like the 24 % sequence homolog to fphf and 60 % sequence homolog to frmb from neisseria meningitides [24]. the narrow active site environment of s-formylglutathione hydrolases with c2 substrate preference is clearly distinguishing from fphf. the structural observations in combination with the promiscuous esterase activity profile suggest that fphf forms a new esterase subfamily [6]. for fphf a highly specific covalent sulfonyl fluoride-based inhibitor jcp678 was discovered [4] and inhibition was also confirmed in vitro [6]. a triazole-urea based inhibitor screen also resulted in specific hits with the regio-isomers kt129 and kt130 (fig. 1d) [10]. atomic structures (pdb id 6vhd kt129-bound and 6vhe kt130-bound) show that the 2-phenylpiperidine-1-carbonyl moiety of these inhibitors binds similar to the hydrophobic lipids. the phenylpiperidine fits well into the hydrophobic pocket of fphf. these inhibitorbound structures offer a blueprint of compound design targeting fphf [6]. fphg saouhsc_01912 as mentioned fphg characterizations with biotinylated and fluorescent probes were more difficult compared to the other fph proteins [4]. it is known that fphg is able to bind β-lactam probes [20, 21], but not much else is known about the protein, except that it is highly conserved within staphylococcal species [19]. structure prediction indicates that fphg has the greatest diversity of residues predicted to form a narrow acyl pocket which appears to be designed for a specific substrate [6]. fphh saouhsc_00802 the fphh transposon mutant was the only fph s. aureus single knockout mutant that strongly suggested a functional link to the other fph proteins, as a gel-based labelling assay suggested upregulation of several other fph proteins, for example, fphe and fphd [4]. interestingly, this response to a lack of fphh was not observed in all cells in the population, and the increase of fphe was more significant in the s. aureus newman strain compared to the usa300 strain [10]. these phenomena were observed in the late stationary phase, suggesting a particular relevance of fphh activity during this phase [10]. apart from the mentioned high conservation of fphh in staphylococcal species, [19] not much else is known about fphh. structure prediction indicates a lack of a well-defined acyl binding pocket compared to the larger fph proteins [6]. fphi saouhsc_00417 fphi might be a distant homolog of fphh, as these two are the only fph proteins that share more than 25 % sequence identity at 28 %, and they are predicted to fall into the same pfam family (fig. 1b). similar to fphh, not much is known about fphi. it showed moderate upregulation in an s9 human bronchial epithelial cell line infection model and was detected in a murine pneumonia model [18]. fphj saouhsc_02824 fphj with 22 kd is significantly smaller than the other fph proteins but appears to still form a typical α/β hydrolase domain (fig. 1b). the active site is predicted to be the least defined of all fph proteins due to the reduced size of active site helices [6]. https://dx.doi.org/10.5599/admet.1137 matthias fellner admet & dmpk 10(2) (2022) 107-114 112 figure 1. (a) overview of the α/β hydrolase fold shared by all fph proteins. shared and unique features color coding illustrated. (b) molecular size and hydrolase domain pfam family association for fpha-j. grey box represents the predicted size of the hydrolase domain with unique helices in cyan. (c) fphf structure as an example of the hydrolase domain (pdb id 6vh9). active site triad residues in black with hydrogen bonds as red dotted lines. α helices in purple, active site helices marked black. core β-strands in red. (d) fphf kt130 inhibitor bound to the hydrophobic active site pocket. residues of this pocked in purple, with active site triad and covalent bound kt130 in black. conclusions for the development of targeted probes for localization and risk stratification of staphylococcus spp. infections fphb is the most promising newly discovered active serine hydrolase in s. aureus. this is due to its presence on the bacterial surface during infections and the preliminary results of several probes and inhibitors for fphb. fphf is the best target for intracellular probes as it appears fphf is expressed at high levels during the bacterial life cycle and can easily be targeted with inhibitors. in addition, library screening hits and inhibitor-bound structures are already available to serve as a blueprint for further development. but fphf also has the potential to be used in prodrug development. fphb and fphe are new therapeutic targets for the treatment or management of staphylococcus infections. for these and all other fph proteins, further characterizations are needed to determine their exact biological roles. it is highly likely that phenotypes relevant to the clinics are hidden by redundancy mechanisms within this family of serine hydrolases. admet & dmpk 10(2) (2022) 107-114 new staphylococcus aureus serine hydrolase probe doi: https://dx.doi.org/10.5599/admet.1137 113 conflict of interest: the author declares to have no conflict of interest. references [1] j. hsu. how covid-19 is accelerating the threat of antimicrobial resistance. bmj-british medical journal 369 (2020) m1983. https://doi.org/10.1136/bmj.m1983. [2] antimicrobial resistance in the age of covid-19. nat. microbiol. 5 (2020) 779. https://doi.org/10.1038/s41564-020-0739-4. [3] j.d. turnidge, d. kotsanas, w. munckhof, s. roberts, c.m. bennett, g.r. nimmo, g.w. coombs, r.j. murray, b. howden, p.d.r. johnson, k. dowling, a.n.z. cooperative. staphylococcus aureus bacteraemia: a major cause of mortality in australia and new zealand. medical journal of australia 191 (2009) 368-373. https://doi.org/10.5694/j.1326-5377.2009.tb02841.x. [4] c.s. lentz, j.r. sheldon, l.a. crawford, r. cooper, m. garland, m.r. amieva, e. weerapana, e.p. skaar, m. bogyo. identification of a s. aureus virulence factor by activity-based protein profiling (abpp). nat. chem. biol. 14 (2018) 609-617. https://doi.org/10.1038/s41589-018-0060-1. [5] d.l. ollis, e. cheah, m. cygler, b. dijkstra, f. frolow, s.m. franken, m. harel, s.j. remington, i. silman, j. schrag, j.l. sussman, k.h.g. verschueren, a. goldman. the alpha/beta hydrolase fold. protein eng. 5 (1992) 197-211. https://doi.org/10.1093/protein/5.3.197. [6] m. fellner, c.s. lentz, s.a. jamieson, j.l. brewster, l. chen, m. bogyo, p.d. mace. structural basis for the inhibitor and substrate specificity of the unique fph serine hydrolases of staphylococcus aureus. acs infect. dis. 6 (2020) 2771-2782. https://doi.org/10.1021/acsinfecdis.0c00503. [7] j.j. miller, i.t. shah, j. hatten, y. barekatain, e.a. mueller, a.m. moustafa, r.l. edwards, c.s. dowd, p.j. planet, f.l. muller, j.m. jez, a.r. odom john. structure-guided microbial targeting of antistaphylococcal prodrugs. elife 10 (2021) e66657. https://doi.org/10.7554/elife.66657. [8] c.s. bond. topdraw: a sketchpad for protein structure topology cartoons. bioinformatics 19 (2003) 311-312. https://doi.org/10.1093/bioinformatics/19.2.311. [9] s. el-gebali, j. mistry, a. bateman, s.r. eddy, a. luciani, s.c. potter, m. qureshi, l.j. richardson, g.a. salazar, a. smart, e.l.l. sonnhammer, l. hirsh, l. paladin, d. piovesan, s.c.e. tosatto, r.d. finn. the pfam protein families database in 2019. nucleic acids res 47 (2019) d427-d432. https://doi.org/10.1093/nar/gky995. [10] l. chen, l.j. keller, e. cordasco, m. bogyo, c.s. lentz. fluorescent triazole urea activity-based probes for the single-cell phenotypic characterization of staphylococcus aureus. angew. chem. int. ed. engl. 58 (2019) 5643-5647. https://doi.org/10.1002/anie.201900511. [11] s. utaida, p.m. dunman, d. macapagal, e. murphy, s.j. projan, v.k. singh, r.k. jayaswal, b.j. wilkinson. genome-wide transcriptional profiling of the response of staphylococcus aureus to cellwall-active antibiotics reveals a cell-wall-stress stimulon. microbiology (reading) 149 (2003) 27192732. https://doi.org/10.1099/mic.0.26426-0. [12] m. pietiainen, p. francois, h.l. hyyrylainen, m. tangomo, v. sass, h.g. sahl, j. schrenzel, v.p. kontinen. transcriptome analysis of the responses of staphylococcus aureus to antimicrobial peptides and characterization of the roles of vrade and vrasr in antimicrobial resistance. bmc genomics 10 (2009) 429. https://doi.org/10.1186/1471-2164-10-429. [13] p. sass, a. jansen, c. szekat, v. sass, h.g. sahl, g. bierbaum. the lantibiotic mersacidin is a strong inducer of the cell wall stress response of staphylococcus aureus. bmc microbiol. 8 (2008) 186. htpps://doi.org/10.1186/1471-2180-8-186. [14] a. muthaiyan, e.m. martin, s. natesan, p.g. crandall, b.j. wilkinson, s.c. ricke. antimicrobial effect and mode of action of terpeneless cold-pressed valencia orange essential oil on methicillinresistant staphylococcus aureus. j. appl. microbiol. 112 (2012) 1020-1033. https://doi.org/10.1111/j.1365-2672.2012.05270.x. [15] a.m. palazzolo-ballance, m.l. reniere, k.r. braughton, d.e. sturdevant, m. otto, b.n. kreiswirth, e.p. skaar, f.r. deleo. neutrophil microbicides induce a pathogen survival response in communityhttps://dx.doi.org/10.5599/admet.1137 https://doi.org/10.1136/bmj.m1983 https://doi.org/10.1038/s41564-020-0739-4 https://doi.org/10.1038/s41564-020-0739-4 https://doi.org/10.5694/j.1326-5377.2009.tb02841.x https://doi.org/10.1038/s41589-018-0060-1 https://doi.org/10.1093/protein/5.3.197 https://doi.org/10.1021/acsinfecdis.0c00503 https://doi.org/10.7554/elife.66657 https://doi.org/10.1093/bioinformatics/19.2.311 https://doi.org/10.1093/nar/gky995 https://doi.org/10.1002/anie.201900511 https://doi.org/10.1099/mic.0.26426-0 https://doi.org/10.1186/1471-2164-10-429 https://doi.org/10.1186/1471-2180-8-186 https://doi.org/10.1111/j.1365-2672.2012.05270.x https://doi.org/10.1111/j.1365-2672.2012.05270.x matthias fellner admet & dmpk 10(2) (2022) 107-114 114 associated methicillin-resistant staphylococcus aureus. j. immunol. 180 (2008) 500-509. https://doi.org/10.4049/jimmunol.180.1.500. [16] e. bore, s. langsrud, o. langsrud, t.m. rode, a. holck. acid-shock responses in staphylococcus aureus investigated by global gene expression analysis. microbiology (reading) 153 (2007) 22892303. https://doi.org/10.1099/mic.0.2007/005942-0. [17] y. neumann, k. ohlsen, s. donat, s. engelmann, h. kusch, d. albrecht, m. cartron, a. hurd, s.j. foster. the effect of skin fatty acids on staphylococcus aureus. arch. microbiol. 197 (2015) 245-267. https://doi.org/10.1007/s00203-014-1048-1. [18] s. michalik, m. depke, a. murr, m. gesell salazar, u. kusebauch, et al. a global staphylococcus aureus proteome resource applied to the in vivo characterization of host-pathogen interactions. scientific reports 7 (2017) 9718. https://doi.org/10.1038/s41598-017-10059-w. [19] l.j. keller, c.s. lentz, y.e. chen, r.j. metivier, e. weerapana, m.a. fischbach, m. bogyo. characterization of serine hydrolases across clinical isolates of commensal skin bacteria staphylococcus epidermidis using activity-based protein profiling. acs infect. dis. 6 (2020) 930-938. https://doi.org/10.1021/acsinfecdis.0c00095. [20] i. staub, s.a. sieber. beta-lactam probes as selective chemical-proteomic tools for the identification and functional characterization of resistance associated enzymes in mrsa. j. am. chem. soc. 131 (2009) 6271-6276. https://doi.org/10.1021/ja901304n. [21] i. staub, s.a. sieber. beta-lactams as selective chemical probes for the in vivo labeling of bacterial enzymes involved in cell wall biosynthesis, antibiotic resistance, and virulence. j. am. chem. soc. 130 (2008) 13400-13409. https://doi.org/10.1021/ja803349j. [22] e.f. pettersen, t.d. goddard, c.c. huang, g.s. couch, d.m. greenblatt, e.c. meng, t.e. ferrin. ucsf chimera--a visualization system for exploratory research and analysis. j. comput. chem. 25 (2004) 1605-1612. https://doi.org/10.1002/jcc.20084. [23] c.f. gonzalez, m. proudfoot, g. brown, y. korniyenko, h. mori, a.v. savchenko, a.f. yakunin. molecular basis of formaldehyde detoxification. characterization of two s-formylglutathione hydrolases from escherichia coli, frmb and yeig. j. biol. chem. 281 (2006) 14514-14522. https://doi.org/10.1074/jbc.m600996200. [24] n.h. chen, r.m. counago, k.y. djoko, m.p. jennings, m.a. apicella, b. kobe, a.g. mcewan. a glutathione-dependent detoxification system is required for formaldehyde resistance and optimal survival of neisseria meningitidis in biofilms. antioxid. redox signal. 18 (2013) 743-755. https://doi.org/10.1089/ars.2012.4749. ©2021 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) https://doi.org/10.1099/mic.0.2007/005942-0 https://dx.doi.org/10.1007%2fs00203-014-1048-1 https://doi.org/10.1038/s41598-017-10059-w https://doi.org/10.1021/acsinfecdis.0c00095 https://doi.org/10.1021/ja901304n https://doi.org/10.1021/ja803349j https://doi.org/10.1002/jcc.20084 https://doi.org/10.1074/jbc.m600996200 https://doi.org/10.1089/ars.2012.4749 http://creativecommons.org/licenses/by/3.0/ determination of salicylic acid content in pharmaceuticals using chitosan@fe3o4/cpe electrode detected by swv technique doi: https://doi.org/10.5599/admet.1682 175 admet & dmpk 11(2) (2023) 175-184; doi: https://doi.org/10.5599/admet.1682 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper determination of salicylic acid content in pharmaceuticals using chitosan@fe3o4/cpe electrode detected by swv technique sudarut pitakrut, phetlada sanchayanukun and sasithorn muncharoen* department of chemistry, faculty of science, burapha university, chonburi, 20130, thailand corresponding author: e-mail: muncharoen@go.buu.ac.th (s. muncharoen); tel.: 66-81-949-3166 received: january 30, 2023; revised: february 25, 2023; published: march 1, 2023 abstract chitosan-coated magnetite nanoparticles (chitosan@fe3o4) were used to modify the carbon paste electrode (chitosan@fe3o4/cpe) to enhance sensitivity for salicylic acid (sa) analysis using square wave voltammetry (swv). the performance and behaviour of the purposed electrodes were investigated using cyclic voltammetry (cv). the results showed that the mixed behaviour process was observed. furthermore, parameters affecting swv were also studied. it was discovered that the optimum conditions were a twolinearity range of sa determination, 1-100 and 100-400 m. the limit of detection (lod) and the limit of quantitation (loq) for sa are 0.57 m and 0.90 m, respectively. the proposed electrodes were successfully used to determine sa in applications employing pharmaceutical samples. ©2023 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords salicylic acid; magnetite nanoparticles; wart treatment drugs. introduction 2-hydroxybenzoic acid, also known as salicylic acid (sa) [1], has played a very important role in daily human life from the past to the present. due to its antibacterial properties, sa has been used in many fields, such as in the food industries and agriculture. sa is used to control plant diseases and prolong the spoilage of agricultural products. it is also used as a preservative in fermented food [2]. additionally, in cosmetic and pharmaceutical industries, sa is an ingredient in creams or serums for treating acne or wart, including various skin medicines because of its antiseptic and anti-inflammatory properties [3,4]. although sa is enormously useful, it is also dangerous for humans. given high doses of sa, fever, difficulty breathing, cancer, tinnitus and possibly death are detected [2]. consequently, the analysis of sa content has attracted great interest from scientists. the literature has shown that there are several popular techniques for sa analysis, such as high-performance liquid chromatography, uv-visible spectrophotometry and fluorescence spectroscopy [5-7]. nonetheless, the above-mentioned techniques, even with high accuracy and precision, have limitations on the size of the tool, high cost and require experts to analyse. one of the most popular techniques used for sa analysis is various modes of electrochemical techniques. they are quick techniques with the potential to fabricate a small-size https://doi.org/10.5599/admet.1682 https://doi.org/10.5599/admet.1682 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:muncharoen@go.buu.ac.th http://creativecommons.org/licenses/by/4.0/ s. pitakrut et al. admet & dmpk 11(2) (2023) 175-184 176 device. the improvement of these analytical electrochemical techniques is mostly a modification of the working electrodes to improve sensitivity and selectivity. currently, nanoparticles have received a lot of attention for their good catalytic properties. these can conduct electricity and improve the surface of the working electrode, causing to increase the analytical efficiency even higher [8,9]. one of the nanoparticles that are very popularly used in analytical electrochemical techniques is magnetite nanoparticles (fe3o4) [10,13]. particularly, the magnetite nanoparticles were combined with materials such as gold nanoparticles, carbon nanotubes and reduced graphene oxide, including chitosan [14-16]. the surface of the magnetite nanoparticles was enhanced with chitosan to enhance the analytical efficiency due to adsorption between positively charged chitosan and negatively charged species [17], like sa in this work. hence, the notable aim of this work is to use the magnetite nanoparticles coated with chitosan (chitosan@fe3o4/cpe) as the working electrode to enhance the analytical efficiency [19] for sa determination in wart treatment pharmaceuticals. experimental reagents and materials distilled water (di water) was used throughout all experiments from a milli-q plus system (millipore). all chemicals used in this work were analytical-grade reagents (ar-grade). salicylic acid (c7h6o3) was purchased from rankem (india). potassium dihydrogen phosphate (kh2po4), acetone (c3h6o) and 99 % ethanol (c2h5oh) from qrëctm (new zealand) were obtained. graphite powder (≤20 m powder) and whatmantm qualitative filter paper, grade 1 from sigma-aldrich (island) and paraffin oil from fisher scientific (usa) were used for the preparation of the proposed electrode. fabrication of the electrode the fabrication procedure of the carbon paste electrode modified with chitosan@fe3o4 in this work was adopted from [18]. briefly, it was prepared by weighing 0.0800 g of the graphite powder and 0.0100 g of chitosan@fe3o4. then 1.0 ml acetone was added into the mixture and sonicated using an ultrasonic homogenizer (zealway/china) for 30 min. it became the carbon paste modified with chitosan@fe3o4. after that, 50 l paraffin oil was added to the paste and mixed up. the paste was compressed into an electrode mole (bas/japan) until tight. the carbon paste electrode modified with chitosan@fe3o4 was polished using whatmantm qualitative filter paper grade 1 to smooth the electrode surface. apparatus the potentiostat (autolab pgstat204, metrohm, netherlands) were used for all electrochemical experiment. a conventional three-electrode system: chitosan@fe3o4 modified cpe (chitosan@fe3o4/cpe), ag/agcl (in 3 m kcl), platinum wire as a working electrode, a reference electrode and an auxiliary electrode were utilized. for the standard method, a specord 210 plus uv-visible spectrophotometer (analytica jana/ germany) was used. electrochemical methods the square wave voltammetry (swv) was carried out in a 0.1 m phosphate buffer solution containing ph 5. scan potential 0.2 to 1.4 v, frequency 50 hz, amplitude 50 mv/s, step potential 10 mv/s, and preconcentration step at potential 0.2 v for 30 s were used as voltammetric parameters. the results of the study of the optimum conditions for swv parameters are shown in figure s1 in the supplementary document. the electrochemical impedance spectra (eis) were created using a frequency range of 0.1 hz to 10 khz and an amplitude of 10 mv/s at a voltage of 0.2 v. admet & dmpk 11(2) (2023) 175-184 determination of salicylic acid in pharmaceuticals doi: https://doi.org/10.5599/admet.1682 177 sample analysis generally, sa is one of the important ingredients in medicine for treating warts. then, wart medicine was selected as a sample in this study. these samples are usually paste-like. thus, all these experiments were treated the same way as in sa solutions. in this work, the standard addition was selected to determine sa in all samples. briefly, the samples were diluted 200 times with 70 %v/v ethanol to determine the amount of sa using the square wave voltammetry technique. results and discussion electrochemical response of sa on chitosan@fe3o4/cpe the responses of sa on various electrodes: bare cpe, fe3o4 modified carbon paste electrode (fe3o4/cpe) and chitosan-coated on fe3o4 modified cpe (chitosan@fe3o4/cpe) were studied. the results demonstrate that the oxidation current of sa detected by chitosan@fe3o4/cpe (i = 3.0 a) gave higher currents than on the bare cpe (i = 2.4 a) and fe3o4/cpe (i = 2.7 a) as shown in figure 1. it was observed that the nanoparticles-modified cpe showed higher currents compared to bared cpe due to the high surface area from the nanoparticles [10]. the surface area of the modified (chitosan@fe3o4/cpe) and unmodified electrodes (bare cpe) was calculated according to the randles-ševčík equation [20]. it was found that the electroactive surface area of 30.11 mm2 and 24.09 mm2 for chitosan@fe3o4/cpe and bared cpe, respectively. additionally, another reason for the higher observed current from chitosan@fe3o4/cpe may be the attraction between the protonated amine group of chitosan coating on the fe3o4 and sa, which leads to the enhancement of response currents according to our previous work [18]. the proposed mechanism between sa and electrode surface is shown in figure s2 (supplementary). figure 1. cyclic voltammograms of 50 m salicylic acid (sa) in 0.1 m phosphate buffer ph 5, scan rate 100 mv/s on bared cpe (orange line), modified carbon paste electrode (fe3o4/cpe) (green line) and chitosan coated on fe3o4 modified cpe (chitosan@fe3o4/cpe) (blue line) additionally, the electrochemical impedance spectroscopy (eis) was used for the investigation of the proposed electrode surface. eis data can be represented by a nyquist diagram from which some conclusions about the interface properties of the electrode can be drawn. generally, the frequency dependence of eis can be divided into two regions, semi-circular and straight-line. the semi-circular sections show the resistances of the electrodes. in the case of a narrow semicircle, the resistance is small or there is good electrical conductivity. https://doi.org/10.5599/admet.1682 s. pitakrut et al. admet & dmpk 11(2) (2023) 175-184 178 on the other hand, a wide semicircle indicates a large electrical resistance or lower conductivity. the straightline section depicted the diffusion process of the electrodes [21,22]. eis data of 0.5 mm fe(ii) in 0.1 m kno3 were obtained for three various types of electrodes: bare cpe (orange line), fe3o4/cpe (green line) and chitosan@fe3o4/cpe (blue line) (figure 2). from the results, it can be observed that the smallest semicircle was obtained for fe3o4/cpe electrode. this is because of the good electrical conductivity and fast electron transfer rates at fe3o4 nanoparticles. although the chitosan@fe3o4/cpe similarly contains fe3o4 nanoparticles, the surface of these particles is encapsulated with chitosan. therefore, the chitosan@fe3o4/cpe shows the wider semicircle part in nyquist diagrams compared to the fe3o4/cpe. however, the oxidation current of sa detected with chitosan@fe3o4/cpe illustrated the highest response (figure 1), which may help confirm the attraction between the protonated amine group of chitosan favours the electrochemical reaction. figure 2. nyquist diagrams of bared cpe (orange line), fe3o4/cpe (green line) and chitosan@ fe3o4/cpe (blue line), and equivalent circuit model of the chitosan@fe3o4/cpe (inset). electrochemical behaviour the behaviour of the chitosan@fe3o4/cpe as the working electrode was studied using cyclic voltammetry at various scan rates (50-160 mv/s) (figure 3(a)). the results shown in figure 3 were plotted as follows: figure 3(b) is the plot between the logarithm of peak current (y-axis) and logarithm of scan rate (x-axis) to study the mixed behaviour (adsorption and diffusion), figure 3(c) is the plot between the peak current (y-axis) and the scan rate (x-axis) to study adsorption process and figure 3(d) is the plot between the peak current (y-axis) and square root of scan rate (x-axis) to study diffusion process [23]. as considering the slope value of the plot in figure 3(b), this value (0.7238) was greater than 0.5 but less than 1.0, indicating that charge transfer is under mixed control according to wyantuti, hartati, panatarani & tjokronegoro [24]. additionally, to confirm the behaviour of the electrode, the relative coefficients of figure 3(c) and figure 3(d) were compared. it was found that the relative coefficients of both figure 3(c) and figure 3(d) were close to 1, indicating that the electron transfer at the developed electrodes is under mixed control, which is consistent with the above [24]. effect of ph in this work, the study of the electrochemical response of sa in dependence on the solution ph was done in a phosphate buffer in the ph range between 2 to 8 because of pka1 of sa is 2.9 [25]. the results show that the oxidation current signals of sa detected by chitosan@fe3o4/cpe in the ph range between 2-6 exhibit no differrence, while the signals decreased at ph values above 7, as shown in figure 4(a). due to the deprotonated amine admet & dmpk 11(2) (2023) 175-184 determination of salicylic acid in pharmaceuticals doi: https://doi.org/10.5599/admet.1682 179 group of chitosan (ph 7-8), the attraction between chitosan and sa was diminished. however, in a highly acidic medium, losing of the cross-linkage between chitosan and glutaraldehyde occurred, as referred to in the report of freire et al. [11]. additionally, it was found that sa responses at ph 2-4 were slightly lower than the sa response at ph 5 as shown in figure 4(b). therefore, ph 5 was selected as the optimum medium for further work. figure 3. (a) cyclic voltammogram of 50 m sa in 0.1 m phosphate buffer ph 5 at different scan rates (50-160 mv/s), (b) plot between of logarithm of peak current and the logarithm of scan rate, (c) plot between of the peak current and the scan rate and (d) plot between of peak current and square root of scan rate for 50 m salicylic acid in 0.1 m phosphate buffer ph 5 figure 4. (a) a plot between of peak current and various ph (b) the square wave voltammograms of salicylic acid 50 m in 0.1 m phosphate buffer. the measurement conditions: deposition time 30 s, deposition potential 0.2 v, frequency 50 hz, amplitude 50 mv/s, step potential 10 mv/s. interference study electrochemical techniques can detect the following sample interferences: resorcinol, phenol, lactic acid, and fluorouracil, each to a different extent. in this study, the measurements were conducted by adding https://doi.org/10.5599/admet.1682 s. pitakrut et al. admet & dmpk 11(2) (2023) 175-184 180 interference to 50 m sa under optimum conditions. as shown in table 1, the tolerance limit is set to have a relative error of not more than 10 % [26]. in addition, for sample analysis, the standard addition method was used, the determination should not also be affected by the interfering agents. table 1. interferences study on chitosan@fe3o4/cpe for determination of sa in drug to treat warts sample. interference agent concentration ratio of sa: if tolerance limit (%rsd) resorcinol 1:0.5 5.5 phenol 1:0.1 6.6 lactic acid 1:1000 3.8 fluorouracil 1:2 9.2 *if : interference agent, sa: salicylic acid analytical performance linear range the various concentrations of sa (1.0-400.0 m) were measured by the proposed electrode under optimal conditions using square wave stripping voltammetry (swv). the calibration plots between the various concentrations of sa and oxidation peak current are indicated in figure 5. it was found that the linearity was in the range of 1.0-100.0 m (r2=0.9994) and 100.0-400.0 m (r2=0.9983). figure 5. the calibration plots of sa detected by chitosan@fe3o4/cpe, insert the square wave voltammograms of salicylic acid 1.0 to 100 m and 100.0 to 400.0 m in 0.1 m phosphate buffer ph 5. the measurement conditions: deposition time 30 s, deposition potential 0.2 v, frequency 50 hz, amplitude 50 mv/s, step potential 10 mv/s. limit of detection (lod) and limit of quantitation (loq) the detection limit can be calculated from 3s.d./slope and the limit of quantitation is calculated from 10sd/slope in which s.d. is the standard deviation and slope from the calibration plot. the measured detection and the quantitation limits were 0.57 and 0.90 m, respectively, as shown in table 2. repeatability the stability of the proposed electrode was investigated by measurement of 50 m sa in 0.1 m phosphate buffer ph 5 using the same electrode. a relative standard deviation (%rsd) was 1.17 % (n=35), according to the aoac standard stating that %rsd has not to exceed 7.3 % (table 2) [27]. admet & dmpk 11(2) (2023) 175-184 determination of salicylic acid in pharmaceuticals doi: https://doi.org/10.5599/admet.1682 181 table 2. analytical performance for the determination of sa under optimal conditions parameter result linear range (m) 1.0-100.0 and 100.0-400.0 linear equation 1.0-100.0 m y = 0.1375x 0.0584 (r2 = 0.9994) 100.0-400.0 m y = 0.0882x + 5.4656 (r2 = 0.9983) limit of detection (lod) (m) 0.57 limit of quantification (loq) (m) 0.90 % rsd (n = 35) 1.17 application for real samples as mentioned above, the standard addition was used for sa analysis in all samples. the obtained results were compared with those of uv-visible spectrophotometry as the standard method, as shown in table 3. it was found that the two methods were not significantly different at the 95% confidence level (tstat = 0.69, tcrit = 2.78). in addition, the recovery percentages of samples adding with 3 concentrations of 10.00-30.00 mm sa were in the range of 94.40-113.51 that remained within the acceptable range (85-115 %) in response to the aoac standard (table 4) [27]. table 3. comparison of sa content analysis in pharmaceutical samples using between the standard method (uv-visible spectrophotometric method) and the proposed method. sample proposed method (mm) standard method (mm) (n=3) (n=3) s1 1.39±0.04 1.39±0.00 s2 0.84±0.08 0.86±0.00 s3 0.60±0.03 0.62±0.00 s4 0.93±0.04 0.91±0.00 s5 0.06±0.02 0.06±0.00 table 4. recovery percentage for determination of sa in samples using the developed method sample concentration of sa, m recovery, % add found (mean ± sd; n = 3) (n = 3) s1 0.00 13.94 ± 0.38 10.00 23.47 ± 0.79 95.32 20.00 34.22 ± 1.36 101.39 30.00 43.77 ± 2.00 99.42 s2 0.00 10.08 ± 0.93 10.00 21.12 ± 1.53 110.61 20.00 31.32 ± 3.23 106.33 30.00 43.77 ± 2.00 112.36 s3 0.00 7.07 + 0.49 10.00 17.19 ± 1.19 98.64 20.00 28.79 ± 1.70 107.15 30.00 41.41 ± 2.89 113.51 s4 0.00 7.07 ± 0.49 10.00 18.06 ± 1.20 110.06 20.00 28.89 ± 2.13 106.95 30.00 39.81 ± 3.91 107.67 s5 0.00 9.91 ± 0.12 10.00 19.53 ± 0.07 94.40 20.00 30.54 ± 0.56 103.14 30.00 41.13 ± 3.21 104.08 comparison with previously published research the comparison of the proposed method using chitosan@fe3o4/cpe with some studies in the literature is shown in table 5. the procedure of all reports shown in this table consisted of various voltammetric modes https://doi.org/10.5599/admet.1682 s. pitakrut et al. admet & dmpk 11(2) (2023) 175-184 182 and different working electrode types. it was discovered that most of the electrodes modified with nanoparticles showed high sensitivity, including this work [2,4,28-31]. although techniques for electrode preparation using nanoparticles require complicated electrode fabrication methods, a suitable wide working range for determining sa in the pharmaceutical samples is found especially in this work. table 5. a comparison of the performance of the proposed electrodes with previous studies to determine sa. electrode technique sample linear range, m lod, m ref. nitio3/cpe dpv aloe vera 3-40 and 40-1000 0.068 [28] mip/gce swv 60-100 20 [29] spce dpv pickled vegetables and fruits juice 1-200 1.6 [2] spce swv urine 16-300 5.6 [30] ce/zro2/cpe swv human serum, milk and pharmaceuticals 5-1000 1.1 [4] gce dpv pharmaceuticals 7-434 0.65 [31] chitosan@fe3o4/cpe swv pharmaceuticals 1-100 and 100-400 0.57 this work conclusions in this work, the carbon paste electrodes modified with chitosan-coated magnetite nanoparticles (chitosan@fe3o4/cpe) were developed for sa analysis. at acidic conditions, the amine proton of chitosan at magnetite nanoparticle as the positively charged surface attracted the negatively charged sa increasing the sensitivity and selectivity of sa analysis. under the optimum conditions, the linearity ranges of 1.0-100.0 and 100.0-400.0 m with r2 of 0.9994 and 0.9983) were observed. the proposed method gave high sensitivity (lod = 0.57 m). the steadiness of the proposed electrodes as precision showed actual stability with 1.17 %rsd (n = 35). furthermore, this method was also successfully applied for the determination of sa in wart treatment pharmaceutical samples. conflict of interest: none acknowledgements: this work was supported as research funding from faculty of science, burapha university (sc07/2565) and undergraduate students (b.sc. program) in 2020 academic year by department of chemistry, faculty of science, burapha university. references [1] p.k. keszycka, m. szkop, d. gajewska. overall content of salicylic acid and salicylates in food available on the european market. j. agric. food chem. 65 (2017) 11085-11091. https://doi.org/10.1021/acs.jafc.7b04313. [2] w. detpisuttitham, c. phanthong, s. ngamchana, p. rijiravanich, w. surareungchai. electrochemical detection of salicylic acid in pickled fruit/vegetable and juice. journal of analysis and testing 4 (2020) 291-297. https://doi.org/10.1007/s41664-020-00127-y. [3] n. gissawong, s. srijaranai, s. sansuk. a simple capture-release strategy based on an instantly formed mixed metal hydroxide sorbent for determination of salicylic acid in cosmetics. sustainable chemistry and pharmacy 13 (2019). https://doi.org/10.1016/j.scp.2019.100154. [4] t. alizadeh, s. nayeri. electrocatalytic oxidation of salicylic acid at a carbon paste electrode impregnated with cerium-doped zirconium oxide nanoparticles as a new sensing approach for salicylic acid determination. journal of solid state electrochemistry 22 (2018) 2039-2048. https://doi.org/10.1007/s10008-018-3907-1. [5] m. ramos payan, m.a. bello lopez, r. fernandez-torres, j.l. perez bernal, m. callejon mochon. hplc determination of ibuprofen, diclofenac and salicylic acid using hollow fiber-based liquid phase https://doi.org/10.1021/acs.jafc.7b04313 https://doi.org/10.1021/acs.jafc.7b04313 https://doi.org/10.1007/s41664-020-00127-y https://doi.org/10.1016/j.scp.2019.100154 https://doi.org/10.1007/s10008-018-3907-1 https://doi.org/10.1007/s10008-018-3907-1 admet & dmpk 11(2) (2023) 175-184 determination of salicylic acid in pharmaceuticals doi: https://doi.org/10.5599/admet.1682 183 microextraction (hf-lpme). anal chim acta 653 (2009) 184-190. https://doi.org/10.1016/j.aca.2009.09.018. [6] u. saha, k. baksi. spectrophotometric determination of salicylic acid in pharmaceutical formulations using copper(ii) acetate as a colour developer. analyst 110 (1985) 739-741. https://doi.org/10.1039/an9851000739. [7] m.m. karim, h.s. lee, y.s. kim, h.s. bae, s.h. lee. analysis of salicylic acid based on the fluorescence enhancement of the as(iii)-salicylic acid system. anal. chim. acta 576 (2006) 136-139. https://doi.org/10.1016/j.aca.2006.02.004. [8] m. reza ganjali. determination of salicylic acid by differential pulse voltammetry using zno/al2o3 nanocomposite modified graphite screen printed electrode. international journal of electrochemical science (2017) 9972-9982. https://doi.org/10.20964/2017.11.49. [9] h.-y. liu, j.-j. wen, z.-h. huang, h. ma, h.-x. xu, y.-b. qiu, w.-j. zhao, c.-c. gu. prussian blue analogue of copper-cobalt decorated with multi-walled carbon nanotubes based electrochemical sensor for sensitive determination of nitrite in food samples. chinese journal of analytical chemistry 47 (2019) e19066-e19072. https://doi.org/10.1016/s1872-2040(19)61168-0. [10] f. moreira, t. de andrade maranhão, a. spinelli. carbon paste electrode modified with fe3o4 nanoparticles and bmi.pf6 ionic liquid for determination of estrone by square-wave voltammetry. journal of solid state electrochemistry 22 (2017) 1303-1313. https://doi.org/10.1007/s10008-0173678-0. [11] t.m. freire, l.m.u. dutra, d.c. queiroz, n. ricardo, k. barreto, j.c. denardin, f.r. wurm, c.p. sousa, a.n. correia, p. de lima-neto, p.b.a. fechine. fast ultrasound assisted synthesis of chitosan-based magnetite nanocomposites as a modified electrode sensor. carbohydr. polym. 151 (2016) 760-769. https://doi.org/10.1016/j.carbpol.2016.05.095. [12] f. beigmoradi, h. beitollahi. fe3o4/go nanocomposite modified glassy carbon electrode as a novel voltammetric sensor for determination of bisphenol a. journal of electrochemical science and engineering (2022). https://doi.org/10.5599/jese.1482 [13] h. tashakkorian, b. aflatoonian, p.m. jahani, m.r. aflatoonian. electrochemical sensor for determination of hydroxylamine using functionalized fe3o4 nanoparticles and graphene oxide modified screen-printed electrode. journal of electrochemical science and engineering (2021). https://doi.org/10.5599/jese.1145 [14] m. sedki, g. zhao, s. ma, d. jassby, a. mulchandani. linker-free magnetite-decorated gold nanoparticles (fe(3)o(4)-au): synthesis, characterization, and application for electrochemical detection of arsenic (iii). sensors (basel) 21 (2021). https://doi.org/10.3390/s21030883. [15] s. qu, j. wang, j. kong, p. yang, g. chen. magnetic loading of carbon nanotube/nano-fe(3)o(4) composite for electrochemical sensing. talanta 71 (2007) 1096-1102. https://doi.org/10.1016/j.talanta.2006.06.003. [16] m. tabrizi, s.-a. shahidi, f. chekin, a. ghorbani-hasansaraei, s.n. raeisi. reduce graphene oxide/fe3o4 nanocomposite biosynthesized by sour lemon peel; using as electro-catalyst for fabrication of vanillin electrochemical sensor in food products analysis and anticancer activity. topics in catalysis 65 (2022) 726-732. https://doi.org/10.1007/s11244-021-01541-x. [17] i.a. mattioli, p. cervini, e.t.g. cavalheiro. screen-printed disposable electrodes using graphitepolyurethane composites modified with magnetite and chitosan-coated magnetite nanoparticles for voltammetric epinephrine sensing: a comparative study. mikrochim acta 187 (2020) 318. https://doi.org/10.1007/s00604-020-04259-x. [18] p. sanchayanukun, s. muncharoen. chitosan coated magnetite nanoparticle as a working electrode for determination of cr(vi) using square wave adsorptive cathodic stripping voltammetry. talanta 217 (2020) 121027. https://doi.org/10.1016/j.talanta.2020.121027. [19] s. dehdashtian, m.b. gholivand, m. shamsipur, s. kariminia. construction of a sensitive and selective sensor for morphine using chitosan coated fe3o4 magnetic nanoparticle as a modifier. mater. sci. eng. c mater. biol. appl. 58 (2016) 53-59. https://doi.org/10.1016/j.msec.2015.07.049. https://doi.org/10.5599/admet.1682 https://doi.org/10.1016/j.aca.2009.09.018 https://doi.org/10.1016/j.aca.2009.09.018 https://doi.org/10.1039/an9851000739 https://doi.org/10.1039/an9851000739 https://doi.org/10.1016/j.aca.2006.02.004 https://doi.org/10.1016/j.aca.2006.02.004 https://doi.org/10.20964/2017.11.49 https://doi.org/10.1016/s1872-2040(19)61168-0 https://doi.org/10.1007/s10008-017-3678-0 https://doi.org/10.1007/s10008-017-3678-0 https://doi.org/10.1016/j.carbpol.2016.05.095 https://doi.org/10.5599/jese.1482 https://doi.org/10.5599/jese.1145 https://doi.org/10.3390/s21030883 https://doi.org/10.1016/j.talanta.2006.06.003 https://doi.org/10.1016/j.talanta.2006.06.003 https://doi.org/10.1007/s11244-021-01541-x https://doi.org/10.1007/s00604-020-04259-x https://doi.org/10.1016/j.talanta.2020.121027 https://doi.org/10.1016/j.msec.2015.07.049 s. pitakrut et al. admet & dmpk 11(2) (2023) 175-184 184 [20] a. garcia-miranda ferrari, c.w. foster, p.j. kelly, d.a.c. brownson, c.e. banks. determination of the electrochemical area of screen-printed electrochemical sensing platforms. biosensors (basel) 8 (2018). https://doi.org/10.3390/bios8020053. [21] y.-h. wang, c.-m. yu, z.-q. pan, y.-f. wang, j.-w. guo, h.-y. gu. a gold electrode modified with hemoglobin and the chitosan@fe3o4 nanocomposite particles for direct electrochemistry of hydrogen peroxide. microchimica acta 180 (2013) 659-667. https://doi.org/10.1007/s00604-0130977-8. [22] h. derikvand, a. azadbakht. an impedimetric sensor comprising magnetic nanoparticles–graphene oxide and carbon nanotube for the electrocatalytic oxidation of salicylic acid. journal of inorganic and organometallic polymers and materials 27 (2017) 901-911. https://doi.org/10.1007/s10904-0170535-7. [23] d. asangil, i. hudai tasdemir, e. kilic. adsorptive stripping voltammetric methods for determination of aripiprazole. j pharm anal 2 (2012) 193-199. https://doi.org/10.1016/j.jpha.2012.01.009. [24] s. wyantuti, y.w. hartati, c. panatarani, r. tjokronegoro. cyclic voltammetric study of chromium (vi) and chromium (iii) on the gold nanoparticles-modified glassy carbon electrode. procedia chemistry 17 (2015) 170-176. https://doi.org/10.1016/j.proche.2015.12.109. [25] r. kort, a. nocker, a. de kat angelino-bart, s. van veen, h. verheij, f. schuren, r. montijn. real-time detection of viable microorganisms by intracellular phototautomerism. bmc biotechnol 10 (2010) 45. https://doi.org/10.1186/1472-6750-10-45. [26] d. eskiköy bayraktepe, z. yazan. application of single‐use electrode based on nano‐clay and mwcnt for simultaneous determination of acetaminophen, ascorbic acid and acetylsalicylic acid in pharmaceutical dosage. electroanalysis 32 (2020) 1263-1272. https://doi.org/10.1002/elan.201900601. [27] aoac. official methods of analysis of the association of official analytical chemists: official methods of analysis of aoac international. (2019). [28] s.m. ghoreishi, f.z. kashani, a. khoobi, m. enhessari. fabrication of a nickel titanate nanoceramic modified electrode for electrochemical studies and detection of salicylic acid. journal of molecular liquids 211 (2015) 970-980. https://doi.org/10.1016/j.molliq.2015.08.035. [29] j. kang, h. zhang, z. wang, g. wu, x. lu. a novel amperometric sensor for salicylic acid based on molecularly imprinted polymer-modified electrodes. polymer-plastics technology and engineering 48 (2009) 639-645. https://doi.org/10.1080/03602550902824499. [30] s. rawlinson, a. mclister, p. kanyong, j. davis. rapid determination of salicylic acid at screen printed electrodes. microchemical journal 137 (2018) 71-77. https://doi.org/10.1016/j.microc.2017.09.019. [31] a.a. torriero, j.m. luco, l. sereno, j. raba. voltammetric determination of salicylic acid in pharmaceuticals formulations of acetylsalicylic acid. talanta 62 (2004) 247-254. https://doi.org/10.1016/j.talanta.2003.07.005. ©2023 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) https://doi.org/10.3390/bios8020053 https://doi.org/10.1007/s00604-013-0977-8 https://doi.org/10.1007/s00604-013-0977-8 https://doi.org/10.1007/s10904-017-0535-7 https://doi.org/10.1007/s10904-017-0535-7 https://doi.org/10.1016/j.jpha.2012.01.009 https://doi.org/10.1016/j.proche.2015.12.109 https://doi.org/10.1186/1472-6750-10-45 https://doi.org/10.1002/elan.201900601 https://doi.org/10.1002/elan.201900601 https://doi.org/10.1016/j.molliq.2015.08.035 https://doi.org/10.1080/03602550902824499 https://doi.org/10.1016/j.microc.2017.09.019 https://doi.org/10.1016/j.talanta.2003.07.005 https://doi.org/10.1016/j.talanta.2003.07.005 http://creativecommons.org/licenses/by/3.0/ voltammetric determination of hydrochlorothiazide at a modified carbon paste electrode with polypyrrole nanotubes doi: https://doi.org/10.5599/admet.1706 293 admet & dmpk 11(2) (2023) 293-302; doi: https://doi.org/10.5599/admet.1706 open access : issn : 1848-7718 http://pub.iapchem.org/ojs/index.php/admet/index original scientific paper voltammetric determination of hydrochlorothiazide at a modified carbon paste electrode with polypyrrole nanotubes arefeh mohammadnavaz1 and fariba garkani-nejad2,* 1department of chemistry, graduate university of advanced technology, kerman, iran 2environment department, institute of science and high technology and environmental sciences, graduate university of advanced technology, kerman, iran *corresponding author: e-mail: f.garkani95@gmail.com received: february 09, 2023; revised: february 28, 2023; published: march 15, 2023 abstract in this paper, the electrochemical behavior of hydrochlorothiazide (hctz) is described using carbon paste electrodes modified with polypyrrole nanotubes (ppy-nts/cpes) at ph value 7. experiments revealed that the presence of hctz greatly impacts the electrochemical behavior of modified cpes. the synthes ized ppy-nts were utilized as a sensing material for the electrochemical detection of hctz and were investigated by cyclic voltammetry (cv), differential pulse voltammetry (dpv) and chronoamperometry. the key experiment conditions, including supporting electrolyte and electrolyte ph, were studied and optimized. under optimized conditions, the prepared sensor displayed the linear relationships for the concentrations of hctz from 5.0 to 400.0 μm (r2 = 0.9984). the detection limit of the ppy-nts/cpes sensor was found to be 1.5 μm using the dpv method. the ppy-nts is highly selective, stable and sensitive for the determination of hct. therefore, we believe the newly prepared ppy -nts material can be useful for different electrochemical applications . ©2023 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords carbon paste electrodes; polypyrrole nanotubes; modified electrode; voltammetry; hydrochlorothiazide introduction hypertension (high blood pressure) adds to the workload of the heart and arteries. if it continues long, the heart and arteries may not function properly. this can damage the blood vessels of the brain, heart, and kidneys, resulting in a stroke, heart attack or kidney failure. these problems may be less likely to occur if blood pressure is controlled. hydrochlorothiazide (hctz) combination is used to treat hypertension. hydrochlorothiazide is a thiazide-type diuretic that has been used clinically for more than half a century. the drug has been widely used to treat hypertension globally and is relatively safe. hydrochlorothiazide acts on the distal convoluted tubules and inhibits the sodium chloride co-transporter system. this action leads to a diuretic action that lowers blood pressure, but there is also a potassium loss in the urine. hydrochlorothiazide is also helpful in removing the excess water from the body; however, calcium retains in the body. moreover, as a therapeutic option for congestive heart failure, diabetes insipidus, renal tubular acidosis, and symptomatic edema. in addition, it https://doi.org/10.5599/admet.1706 https://doi.org/10.5599/admet.1706 http://pub.iapchem.org/ojs/index.php/admet/index mailto:f.garkani95@gmail.com http://creativecommons.org/licenses/by/4.0/ a. mohammadnavaz and f. garkani-nejad admet & dmpk 11(2) (2023) 293-302 294 prevents kidney stones [1-5]. all these aspects encourage investigating the electrochemical features of hctz on novel materials, and it has been studied with an increasing trend in recent research activities around the world. the various techniques available for the selective and sensitive detection of hctz include hplc, capillary zone electrophoretic, spectrophotometric/hplc and electrochemical methods [6-11]. electrochemistrybased methods can be employed considering their lower cost, speediness, portability, reasonable selectivity, simple preparation process, suitable accuracy, and precision for the analysis of biological compounds, gas pollutants, drugs, food and water pollutants compounds [12-22]. the chemical modification of inert substrate electrodes offers significant advantages in the design and development of electrochemical sensors. the properties of chemically modified electrodes that have driven their development include increased selectivity and sensitivity, chemical and electrochemical stability, larger usable potential windows, and resistance to fouling [23-35]. the incorporation of nanomaterials has had a great impact on the development of electrochemical sensors [36-40]. significant progress has been made toward synthesizing nanomaterials with controllable morphologies, dimensions, surface charges, and physicochemical properties [41-62]. conducting polymers (cps) derivatives of polypyrrole nanotubes (ppy-nts) are 1-d nanostructured materials and are considered as one of the ideal candidates due to their large usage in various industrial applications and tremendous physico-chemical properties, such as high electrical conductivity, good mechanical stability, nanoscale particle size, lightweight, and wide surface area. ppy-nts has an excellent material for device fabrication in electrochemical sensors or biochemical fields due to their small diameter with an average length, which may allow more scope for proper chemical interaction with a doping material and enhance the surface area and sensitivity compared to the bulk material. ppy-nts show a high conductivity due to extended π conjugation and a tunable doping process that promotes the electrochemical oxidation activity [63,64]. based on the above-mentioned information and in light of the information presented, the current study was based on the development and application of ppy-nts/cpes for the voltammetric determination of hctz. inspired by the aforementioned discussions, we modified the bare electrodes with ppy-nts, which are applied to improve the selectivity and sensitivity of the electrochemical sensor. hence, the ppy-nts were synthesized and developed as a highly sensitive and selective platform for detecting hctz. the constructed sensor has good performance characteristics, simplicity of preparation, high selectivity, stability, wide linear range and a small limit of detection. it was successfully applied for the voltammetry determination of hctz in biological samples. experimental equipment and materials in order to do electrochemical tests at ambient temperature, we utilized the auto-lab potentiostat /galvanostat (pgstat 302n, eco chemie, the netherlands) with gpes (general purpose electrochemical system-version 4.9) software to control the system. electrochemical measurements were performed at room temperature in a conventional electrochemical cell with a ppy-nts/cpe as the working electrode, 3.0 m ag/ agcl/kcl as a reference electrode (azar electrode, urmia, iran) and platinum wire as a counter electrode (azar electrode, urmia, iran). moreover, ph was measured using the metrohm 713 ph-meter with a glass electrode (switzerland). hydrochlorothiazide and all other solutions used during the procedure were prepared by reagent-grade chemicals from merck and sigma-aldrich and deionized water was supply from millipore, germany. orthophosphoric acid was utilized to prepare the phosphate buffer solutions (pbss), and sodium hydroxide was used to adjust the desired ph values (ph range between 2.0 and 9.0). admet & dmpk 11(2) (2023) 293-302 hydrochlorothiazide determination on polypyrrole modified cpe doi: https://doi.org/10.5599/admet.1706 295 preparation of ppy nanotubes polypyrrole nanotubes (ppy-nts) were prepared by the oxidation of pyrrole monomer with iron(iii) chloride in the presence of a structure-guiding agent, methyl orange. in a typical synthesis, 0.784 g (2.3 mmol) methyl orange and 3.888 g (23 mmol) fecl3 were dissolved into 480 ml of deionized water. then 0.84 ml (12.1 mmol) of pyrrole was added to the solution and stirred for 24 h at room temperature. the formed ppy precipitate was washed with deionized water/ethanol several times until the filtrate was colorless and neutral and finally dried under a vacuum atmosphere at 65 °c for 20 h. figure 1 shows the fe-sem image of ppy nanotubes. figure 1. fe-sem image of ppy nanotubes. preparation and surface modification of electrode to prepare ppy-nts/cpe, 0.95 g graphite powder and 0.05 g ppy-nts were mixed. next, a suitable amount of paraffin oil was poured into the resulting mixture, followed by mixing well for 30 min to obtain a uniformly wetted paste. an appropriate amount of the paste was tightly packed into a glass tube and a copper wire was positioned over the carbon paste to make electrical contact. results and discussion electrochemical behavior of hctz on polypyrrole nanotubes according to our knowledge, the electrooxidation of hctz is closely related to the ph value of the solution. so, the effect of ph was investigated using the differential pulse voltammetry (dpv) method. the results show that the oxidation peak current increased slowly from ph 2.0 to 7.0, and then the current conversely decreased when the ph value increased from 7.0 to 9.0. according to obtained results, ph 7.0 was chosen as the optimal experimental condition for other experiments. the electrochemical reaction of hctz involves two electrons and two protons, according to scheme 1. scheme 1. mechanism for the oxidation of hctz at the surface of ppy-nts/cpe. https://doi.org/10.5599/admet.1706 a. mohammadnavaz and f. garkani-nejad admet & dmpk 11(2) (2023) 293-302 296 the electrochemical behavior of the cpe, ppy-nts/cpe was studied by the cyclic voltammetry (cv) technique in the 0.1 m phosphate buffer (ph=7.0) as the supporting electrolyte at a scan rate of 50 mv s−1 (figure 2). as shown in figure 2, in comparison to the bare cpe (a), ppy-nts/cpe (b) presents a well-defined irreversible oxide peak with a higher current signal (hctz concentration equal to 200.0 µm). figure 2. cyclic voltammograms of a) cpe and b) ppynts/cpe in the presence of 200.0 μm hctz at a ph 7.0 of 0.1 m pbs, respectively. role of variable scan rates the effect of the potential scan rates (5-100 mv s-1) on the electrochemical oxidation of hctz was studied by linear sweep voltammograms (lsv). figure 3 shows the lsv of 200.0 µm of hctz in the 0.1 m phosphate buffer solution at the ppy-nts/cpe. these results show that the anodic current increases with increasing scan rate. the oxidation current of hctz increased linearly with the square root of the scan rate (figure 3, inset), demonstrating a diffusion-controlled electrochemical process. figure 3. linear sweep voltammograms of hctz (200 μm) at ppy-nts/cpe at different scan rates of a) 5, b) 10, c) 20, d)30, e) 40, f) 50, g) 60, h) 70, i) 80, j) 90 and k) 100 mv/s in 0.1 m pbs (ph 7.0). insert: plot of ip versus ν 1/2 for the oxidation of hctz at ppy-nts/cpe. chronoamperometric analysis the chronoamperometric measurements of htcz at the ppy-nts/cpe surface were done to estimate the apparent diffusion coefficient of htcz. figure 4 shows the current-time profiles obtained by setting the working electrode potential at 950 mv for different concentrations of htcz. at long enough experimental times (t=0.33s), where the electron transfer reaction rate of htcz is more than its diffusion rate toward the working electrode surface, the current is diffusion controlled. figure 4, inset a, shows the experimental plots of i versus admet & dmpk 11(2) (2023) 293-302 hydrochlorothiazide determination on polypyrrole modified cpe doi: https://doi.org/10.5599/admet.1706 297 t-1/2 with the best fit for different concentrations of htcz employed. the slopes of the resulting straight lines were then plotted versus the htcz concentration (figure 4, inset b). based on the cottrell equation [65], the slope of this plot (figure 4 inset b) can be used to estimate the apparent diffusion coefficient of htcz. from the slope of this plot (11.644 a s1/2 mm-1), the value of diffusion coefficient was found to be 1.710-6 cm s1. figure 4. chronoamperograms obtained at the ppy-nts/cpe in the presence of a) 0.1, b) 0.7, c) 1.3 and d) 2.0 mm hctz in the 0.1 m buffer solution (ph 7.0). a) plot of i versus t-1/2 for electrooxidation of hctz obtained from chronoamperoms a–d. b) plot of slope from straight lines versus hctz level. dpv analysis of hctz dpv was used for the determination of hctz at ppy-nts/cpe due to its high sensitivity. the dpv responses for different concentrations of hctz are illustrated in figure 5. the linear range was found to be 5.0 μm to 400.0 μm. the linear equation was ip (μa)=0.8477-0.0511 chctz (μm) with a correlation coefficient of 0.9984. the detection limit was 1.5 μm (s/n=3). figure 5. dpv curves of ppy-nts/cpe in the 0.1 m buffer solution (ph 7.0) containing different concentrations of hctz. a-i corresponds to 5.0, 20.0, 40.0, 60.0, 80.0, 100.0, 200.0, 300.0 and 400.0 μm hctz. inset: plots of oxidation peak. conclusion a novel electrochemical protocol using ppy-nts/cpe was fabricated for the sensitive determination of hctz. the modified electrode electrocatalytically oxidizes the hctz at a less positive potential with an https://doi.org/10.5599/admet.1706 a. mohammadnavaz and f. garkani-nejad admet & dmpk 11(2) (2023) 293-302 298 increased oxidation current. the electrocatalytic oxidation current of hctz was linearly increased with the increased concentration of hctz. the sensor under the optimized circumstances possessed a fast current response to hctz, with a linear dynamic range between 5.0-400.0 μm, a thin limit of detection of 1.5 μm, and an appreciable sensitivity of 0.0511 µa/µm. according to the analyses, the modified electrode demonstrated acceptable electrocatalytic activities and sensitivity. also, excellent features, like a wide linear range, low detection limit, high reproducibility and repeatability and longtime stability, proved the successful application of this sensor for the determinations of hctz. conflict of interest: none references [1] a. nezhadali, m. mojarrab. computational study and multivariate optimization of hydrochlorothiazide analysis using molecularly imprinted polymer electrochemical sensor based on carbon nanotube/polypyrrole film. sensors and actuators b: chemical 190 (2014) 829-837. https://doi.org/10.1016/j.snb.2013.08.086 [2] r. lal, a. tahira, a.a. khand, i.n. qureshi, j. mangi, s.a. lakho, u. aftab, b. lal, s. basha, a.m. karami, s.l. al-saeedi. flower-like cuo/polyaniline composite for electrochemical determination of hydrochlorothiazide. bulletin of materials science 44 (2021) 1-8. https://doi.org/10.1007/s12034-02102528-w [3] m.e. ernst, m.a. fravel. thiazide and the thiazide-like diuretics: review of hydrochlorothiazide, chlorthalidone, and indapamide. american journal of hypertension 35(7) (2022) 573-586. https://doi.org/10.1093/ajh/hpac048 [4] z. shahabi, v. zare-shahabadi, s. sayyahi, j. burromand-piroz. novel cuo/polymethylenedisulfide nanocomposite for high performance electrocatalytic determination of hydrochlorothiazide in real samples. journal of porous materials 29(4) (2022) 1123-1135. https://doi.org/10.1007/s10934-02201236-y [5] b. núñez-acevedo, j. domínguez-ortega, b. rodríguez-jiménez, c. kindelan-recarte, m.a. pérezfernández. severe and rare adverse reaction to hydrochlorothiazide. revista alergia méxico 65(4) (2018) 442-445. https://doi.org/10.29262/ram.v65i4.363 [6] a. hemdan, n.f. al‐tannak, e.h. mohamed. development of a multivariate model with desirability‐ based optimization for determination of atenolol and hydrochlorothiazide by eco‐friendly hplc method with fluorescence detection. journal of separation science 45(4) (2022) 824-831. https://doi.org/10.1002/jssc.202100711 [7] s.m. haque. box–behnken experimental design for optimizing the hplc method to determine hydrochlorothiazide in pharmaceutical formulations and biological fluid. journal of molecular liquids 352 (2022) 118708. https://doi.org/10.1016/j.molliq.2022.118708 [8] h.m. ahmed, t.s. belal, r.a. shaalan, f.a. el yazbi, s.m. elonsy. validated capillary zone electrophoretic method for simultaneous analysis of benazepril in combination with amlodipine besylate and hydrochlorothiazide. acta chromatographica 32(4) 2020 219-227. https://doi.org/10.1556/1326.2019.00686 [9] g. tiris, m. mehmandoust, h.m. lotfy, n. erk, s.w. joo, e.n. dragoi, y. vasseghian. simultaneous determination of hydrochlorothiazide, amlodipine, and telmisartan with spectrophotometric and hplc green chemistry applications. chemosphere 303 (2022) 135074. https://doi.org/10.1016/j.chemosphere.2022.135074 [10] e.f. silva, a.a. tanaka, r.n. fernandes, r.a.a. munoz, i.s. da silva. batch injection analysis with electrochemical detection for the simultaneous determination of the diuretics furosemide and hydrochlorothiazide in synthetic urine and pharmaceutical samples. microchemical journal 157 (2020) 105027. https://doi.org/10.1016/j.microc.2020.105027 https://doi.org/10.1016/j.snb.2013.08.086 https://doi.org/10.1016/j.snb.2013.08.086 https://doi.org/10.1007/s12034-021-02528-w https://doi.org/10.1007/s12034-021-02528-w https://doi.org/10.1093/ajh/hpac048 https://doi.org/10.1007/s10934-022-01236-y https://doi.org/10.1007/s10934-022-01236-y https://doi.org/10.29262/ram.v65i4.363 https://doi.org/10.1002/jssc.202100711 https://doi.org/10.1016/j.molliq.2022.118708 https://doi.org/10.1556/1326.2019.00686 https://doi.org/10.1556/1326.2019.00686 https://doi.org/10.1016/j.chemosphere.2022.135074 https://doi.org/10.1016/j.chemosphere.2022.135074 https://doi.org/10.1016/j.microc.2020.105027 admet & dmpk 11(2) (2023) 293-302 hydrochlorothiazide determination on polypyrrole modified cpe doi: https://doi.org/10.5599/admet.1706 299 [11] m.f. khanfar, e.s. abu-nameh, m.m. saket, l.t. al khateeb, a.al ahmad, z. asaad, z. salem, n. alnuman. detection of hydrochlorothiazide, sulfamethoxazole, and trimethoprim at metal oxide modified glassy carbon electrodes. international journal of electrochemical science 15 (2020) 17711787. https://doi.org/10.3390/ma13112521 [12] h. karimi-maleh, f. karimi, y. orooji, g. mansouri, a. razmjou, a. aygun, f. sen. a new nickel-based co-crystal complex electrocatalyst amplified by nio dope pt nanostructure hybrid; a highly sensitive approach for determination of cysteamine in the presence of serotonin. scientific reports 10(1) (2020) 1-13. https://doi.org/10.1038/s41598-020-68663-2 [13] a. lohrasbi-nejad. electrochemical strategies for detection of diazinon. journal of electrochemical science and engineering 12(6) (2022) 1041-1059. https://doi.org/10.5599/jese.1379 [14] j. mohanraj, d. durgalakshmi, r.a. rakkesh, s. balakumar, s. rajendran, h. karimi-maleh. facile synthesis of paper based graphene electrodes for point of care devices: a double stranded dna (dsdna) biosensor. journal of colloid and interface science 566 (2020) 463-472. https://doi.org/10.1016/j.jcis.2020.01.089 [15] m. mazloum-ardakani, h. beitollahi, z. taleat, h. naeimi, n. taghavinia. selective voltammetric determination of d-penicillamine in the presence of tryptophan at a modified carbon paste electrode incorporating tio2 nanoparticles and quinizarine. journal of electroanalytical chemistry 644(1) (2010) 1-6. https://doi.org/10.1016/j.jelechem.2010.02.034 [16] y.f. mustafa, g. chehardoli, s. habibzadeh, z. arzehgar. electrochemical detection of sulfite in food samples. journal of electrochemical science and engineering 12(6) (2022) 1061-1079. https://doi.org/10.5599/jese.1555 [17] m. miraki, h. karimi-maleh, m. a. taher, s. cheraghi, f. karimi, s. agarwal, v.k. gupta. voltammetric amplified platform based on ionic liquid/nio nanocomposite for determination of benserazide and levodopa. journal of molecular liquids 278 (2019) 672-676. https://doi.org/10.1016/j.molliq.2019.01.081 [18] m. velicky, a.n. rodgers, r.a. dryfe, k. tam. use of voltammetry for in vitro equilibrium and transport studies of ionisable drugs. admet and dmpk 2(3) (2014) 143-156. https://doi.org/10.5599/admet.2.3.22 [19] m. kazemipour, m. ansari, a. mohammadi, h. beitollahi, r. ahmadi. use of adsorptive square-wave anodic stripping voltammetry at carbon paste electrode for the determination of amlodipine besylate in pharmaceutical preparations. journal of analytical chemistry 64 (2009) 65-70. https://doi.org/10.1134/s1061934809010134 [20] r.m. mohabis, f. fazeli, i. amini, v. azizkhani. an overview of recent advances in the detection of ascorbic acid by electrochemical techniques. journal of electrochemical science and engineering 12(6) (2022) 1081-1098. https://doi.org/10.5599/jese.1561 [21] h. karimi-maleh, m. sheikhshoaie, i. sheikhshoaie, m. ranjbar, j. alizadeh, n.w. maxakato, a. abbaspourrad. a novel electrochemical epinine sensor using amplified cuo nanoparticles and anhexyl-3-methylimidazolium hexafluorophosphate electrode. new journal of chemistry 43(5) (2019) 2362-2367. https://doi.org/10.1039/c8nj05581e [22] a. meoipun, k. kaewjua, o. chailapakul, w. siangproh. a simple and fast flow injection amperometry for the determination of methimazole in pharmaceutical preparations using an unmodified borondoped diamond electrode. admet and dmpk (2023). https://doi.org/10.5599/admet.1584 [23] z. taleat, m. m. ardakani, h. naeimi, h. beitollahi, m. nejati, h. r. zare. electrochemical behavior of ascorbic acid at a 2, 2'-[3, 6-dioxa-1, 8-octanediylbis (nitriloethylidyne)]-bis-hydroquinone carbon paste electrode. analytical sciences 24(8) (2008) 1039-1044. https://doi.org/10.2116/analsci.24.1039 [24] p.m. jahani. flower-like mos2 screen-printed electrode based sensor for the sensitive detection of sunset yellow fcf in food samples. journal of electrochemical science and engineering 12(6) (2022) 1099-1109. https://doi.org/10.5599/jese.1413 [25] s.a. alavi-tabari, m.a. khalilzadeh, h. karimi-maleh. simultaneous determination of doxorubicin and dasatinib as two breast anticancer drugs uses an amplified sensor with ionic liquid and zno https://doi.org/10.5599/admet.1706 https://doi.org/10.3390/ma13112521 https://doi.org/10.1038/s41598-020-68663-2 https://doi.org/10.5599/jese.1379 https://doi.org/10.1016/j.jcis.2020.01.089 https://doi.org/10.1016/j.jcis.2020.01.089 https://doi.org/10.1016/j.jelechem.2010.02.034 https://doi.org/10.5599/jese.1555 https://doi.org/10.5599/jese.1555 https://doi.org/10.1016/j.molliq.2019.01.081 https://doi.org/10.1016/j.molliq.2019.01.081 https://doi.org/10.5599/admet.2.3.22 https://doi.org/10.5599/admet.2.3.22 https://doi.org/10.1134/s1061934809010134 https://doi.org/10.5599/jese.1561 https://doi.org/10.1039/c8nj05581e https://doi.org/10.5599/admet.1584 https://doi.org/10.2116/analsci.24.1039 https://doi.org/10.5599/jese.1413 a. mohammadnavaz and f. garkani-nejad admet & dmpk 11(2) (2023) 293-302 300 nanoparticle. journal of electroanalytical chemistry 811 (2018) 84-88. https://doi.org/10.1016/j.jelechem.2018.01.034 [26] c.m. brett. electrochemical impedance spectroscopy in the characterisation and application of modified electrodes for electrochemical sensors and biosensors. molecules 27(5) (2022) 1497. https://doi.org/10.3390/molecules27051497 [27] s.z. mohammadi, f. mousazadeh, m. mohammadhasani-pour. electrochemical detection of folic acid using a modified screen printed electrode. journal of electrochemical science and engineering 12(6) (2022) 1111-1120. https://doi.org/10.5599/jese.1360 [28] j.b. raoof, r. ojani, h. beitollahi. electrocatalytic determination of ascorbic acid at chemically modified carbon paste electrode with 2, 7-bis (ferrocenyl ethynyl) fluoren-9-one. international journal of electrochemical science 2(7) (2007) 534-548. [29] v. sanko, a. şenocak, s. o. tümay, y. orooji, e. demirbas, a. khataee. an electrochemical sensor for detection of trace-level endocrine disruptor bisphenol a using mo2ti2alc3 max phase/mwcnt composite modified electrode. environmental research 212 (2022) 113071. https://doi.org/10.1016/j.envres.2022.113071 [30] h. karimi-maleh, a.f. shojaei, k. tabatabaeian, f. karimi, s. shakeri, r. moradi. simultaneous determination of 6-mercaptopruine, 6-thioguanine and dasatinib as three important anticancer drugs using nanostructure voltammetric sensor employing pt/mwcnts and 1-butyl-3-methylimidazolium hexafluoro phosphate. biosensors and bioelectronics 86 (2016) 879-884. https://doi.org/10.1016/j.bios.2016.07.086 [31] s. saghiri, m. ebrahimi, m. bozorgmehr. electrochemical amplified sensor with mgo nanoparticle and ionic liquid: a powerful strategy for methyldopa analysis. chemical methodologies 5(3) (1999) 234239. https://doi.org/10.22034/chemm.2021.128530 [32] t. eren, n. atar, m.l. yola, h. karimi-maleh. a sensitive molecularly imprinted polymer based quartz crystal microbalance nanosensor for selective determination of lovastatin in red yeast rice. food chemistry 185 (2015) 430-436. https://doi.org/10.1016/j.foodchem.2015.03.153 [33] s. saghiri, m. ebrahimi, r. bozorgmehr. nio nanoparticle/1-hexyl-3-methylimidazolium hexafluorophosphate composite for amplification of epinephrine electrochemical sensor. asian journal of nanosciences and materials 4(1) (2021) 46-52. https://doi.org/10.26655/ajnanomat.2021.1.4 [34] h. beitollahi, s.z. mohammadi, m. safaei, s. tajik. applications of electrochemical sensors and biosensors based on modified screen-printed electrodes: a review. analytical methods 12(12) (2020) 1547-1560. https://doi.org/10.1039/c9ay02598g [35] h. karimi-maleh, r. darabi, m. shabani-nooshabadi, m. baghayeri, f. karimi, j. rouhi, m. alizadeh, o. karaman, y. vasseghian, c. karaman. determination of d&c red 33 and patent blue v azo dyes using an impressive electrochemical sensor based on carbon paste electrode modified with zif-8/g-c3n4/co and ionic liquid in mouthwash and toothpaste as real samples. food and chemical toxicology 162 (2022) 112907. https://doi.org/10.1016/j.fct.2022.112907 [36] s. azimi, m. amiri, h. imanzadeh, a. bezaatpour. fe3o4@sio2-nh2/cosb modified carbon paste electrode for simultaneous detection of acetaminophen and chlorpheniramine. advanced journal of chemistry-section a 4(2) (2021) 152-164. https://doi.org/10.22034/ajca.2021.275901.1246 [37] s. mahari, s. gandhi. electrochemical immunosensor for detection of avian salmonellosis based on electroactive reduced graphene oxide (rgo) modified electrode. bioelectrochemistry 144 (2022) 108036. https://doi.org/10.1016/j.bioelechem.2021.108036 [38] h. karimi-maleh, c. karaman, o. karaman, f. karimi, y. vasseghian, l. fu, m. baghayeri, j. rouhi, p. senthil kumar, p.l. show, s. rajendran. nanochemistry approach for the fabrication of fe and n codecorated biomass-derived activated carbon frameworks: a promising oxygen reduction reaction electrocatalyst in neutral media. journal of nanostructure in chemistry 12(3) (2022) 429-439. https://doi.org/10.1007/s40097-022-00492-3 [39] a. hosseini fakhrabad, r. sanavi khoshnood, m.r. abedi, m. ebrahimi. fabrication a composite carbon paste electrodes (cpes) modified with multi-wall carbon nanotubes (mwcnts/n, n-bis (salicyliden)https://doi.org/10.1016/j.jelechem.2018.01.034 https://doi.org/10.1016/j.jelechem.2018.01.034 https://doi.org/10.3390/molecules27051497 https://doi.org/10.5599/jese.1360 https://doi.org/10.1016/j.envres.2022.113071 https://doi.org/10.1016/j.envres.2022.113071 https://doi.org/10.1016/j.bios.2016.07.086 https://doi.org/10.1016/j.bios.2016.07.086 https://doi.org/10.22034/chemm.2021.128530 https://doi.org/10.1016/j.foodchem.2015.03.153 https://doi.org/10.26655/ajnanomat.2021.1.4 https://doi.org/10.1039/c9ay02598g https://doi.org/10.1016/j.fct.2022.112907 https://doi.org/10.22034/ajca.2021.275901.1246 https://doi.org/10.1016/j.bioelechem.2021.108036 https://doi.org/10.1007/s40097-022-00492-3 admet & dmpk 11(2) (2023) 293-302 hydrochlorothiazide determination on polypyrrole modified cpe doi: https://doi.org/10.5599/admet.1706 301 1,3-propandiamine) for determination of lanthanum (iii). eurasian chemical communications 3(9) (2021) 627-634. doi: http://dx.doi.org/10.22034/ecc.2021.288271.1182 [40] m. yang, z. sun, h. jin, r. gui. sulfur nanoparticle-encapsulated mof and boron nanosheet-ferrocene complex modified electrode platform for ratiometric electrochemical sensing of adriamycin and realtime monitoring of drug release. microchemical journal 177 (2022) 107319. https://doi.org/10.1016/j.microc.2022.107319 [41] m. farahmandjou, p. khalili. zno nanoparticles synthesized by co-precipitation method; morphology and optoelectronic study. asian journal of green chemistry 5(2) (2021) 219-226. https://doi.org/10.22034/ajgc.2021.261206.1287 [42] p.s. kumar, b.s. sreeja, k.k. kumar, g. padmalaya. static and dynamic analysis of sulfamethoxazole using go/zno modified glassy carbon electrode by differential pulse voltammetry and amperometry techniques. chemosphere 302 (2022) 134926. https://doi.org/10.1016/j.chemosphere.2022.134926 [43] s.a. ismaeel; y.k. al-bayati. determination of trace metformin in pharmaceutical preparation using molecularly imprinted polymer based pvc-membrane. eurasian chemical communications 3(11) (2021) 812-830. http://dx.doi.org/10.22034/ecc.2021.300477.1224 [44] y. dessie; s. tadesse. a review on advancements of nanocomposites as efficient anode modifier catalyst for microbial fuel cell performance improvement. journal of chemical reviews 3(4) (2021) 320-344. http://dx.doi.org/10.22034/jcr.2021.314327.1128 [45] e.c. martins, e.r. santana, a. spinelli. nitrogen and sulfur co-doped graphene quantum dot-modified electrode for monitoring of multivitamins in energy drinks. talanta 252 (2023) 123836. https://doi.org/ 10.1016/j.talanta.2022.123836 [46] f. zare kazemabadi, a. heydarinasab, a. akbarzadehkhiyavi, m. ardjmand. development, optimization and in vitro evaluation of etoposide loaded lipid polymer hybrid nanoparticles for controlled drug delivery on lung cancer. chemical methodologies 5(2) (2021) 135-152. https://doi.org/10.22034/chemm.2021.121495 [47] s. wang, h. wang, s. liu, h. guo, j. meng, m. chang, s. wu. highly sensitive detection of fluoride based on poly (3-aminophenylboronic acid)-reduced graphene oxide multilayer modified electrode. food chemistry 400 (2023) 134042. https://doi.org/10.1016/j.foodchem.2022.134042 [48] r. kavade, r. khanapure, u. gawali, a. patil, s. patil. degradation of methyl orange under visible light by zno-polyaniline nanocomposites. journal of applied organometallic chemistry 2(2) (2022) 101-112. http://dx.doi.org/10.22034/jaoc.2022.349558.1056 [49] h. shayegan, v. safarifard, h. taherkhani, m.a. rezvani. efficient removal of cobalt(ii) ion from aqueous solution using amide-functionalized metal-organic framework. journal of applied organometallic chemistry 2(3) (2022) 109-118. doi: http://dx.doi.org/10.22034/jaoc.2022.154718 [50] m. mazloum-ardakani, h. beitollahi, b. ganjipour, h. naeimi. novel carbon nanotube paste electrode for simultaneous determination of norepinephrine, uric acid and d-penicillamine. international journal electrochemical science 5 (2010) 531-546. [51] h. duan, d. wang, y. li. green chemistry for nanoparticle synthesis. chemical society reviews 44(16) (2015) 5778-5792. https://doi.org/10.1039/c4cs00363b [52] m. bijad, a. hojjati-najafabadi, h. asari-bami, s. habibzadeh, i. amini, f. fazeli. an overview of modified sensors with focus on electrochemical sensing of sulfite in food samples. eurasian chemical communications 3(2) (2021) 116-138. doi: http://dx.doi.org/10.22034/ecc.2021.268819.1122 [53] m.m. abdul hassan, s. hassan, k.a. hassan. green and chemical synthesis of bimetallic nanoparticles (fe/ni) supported by zeolite 5a as a heterogeneous fenton-like catalyst and study of kinetic and thermodynamic reaction for decolorization of reactive red 120 dye from aqueous pollution. eurasian chemical communications 4 (2022) 1062-1086. https://doi.org/10.22034/ecc.2022.342067.1466 [54] sh. ariavand, m. ebrahimi, e. foladi. design and construction of a novel and an efficient potentiometric sensor for determination of sodium ion in urban water samples. chemical methodologies 6 (2022) 886-904. https://doi.org/10.22034/chemm.2022.348712.1567 https://doi.org/10.5599/admet.1706 http://dx.doi.org/10.22034/ecc.2021.288271.1182 https://doi.org/10.1016/j.microc.2022.107319 https://doi.org/10.1016/j.microc.2022.107319 https://doi.org/10.22034/ajgc.2021.261206.1287 https://doi.org/10.22034/ajgc.2021.261206.1287 https://doi.org/10.1016/j.chemosphere.2022.134926 http://dx.doi.org/10.22034/ecc.2021.300477.1224 http://www.jchemrev.com/article_140627.html http://www.jchemrev.com/article_140627.html http://dx.doi.org/10.22034/jcr.2021.314327.1128 https://doi.org/10.1016/j.talanta.2022.123836 https://doi.org/10.1016/j.talanta.2022.123836 https://doi.org/10.22034/chemm.2021.121495 https://doi.org/10.22034/chemm.2021.121495 https://doi.org/10.1016/j.foodchem.2022.134042 http://dx.doi.org/10.22034/jaoc.2022.349558.1056 http://dx.doi.org/10.22034/jaoc.2022.154718 https://doi.org/10.1039/c4cs00363b http://dx.doi.org/10.22034/ecc.2021.268819.1122 https://doi.org/10.22034/ecc.2022.342067.1466 http://www.chemmethod.com/article_155298.html http://www.chemmethod.com/article_155298.html https://doi.org/10.22034/chemm.2022.348712.1567 a. mohammadnavaz and f. garkani-nejad admet & dmpk 11(2) (2023) 293-302 302 [55] b. paulchamy, g. arthi, b.d. lignesh. a simple approach to stepwise synthesis of graphene oxide nanomaterial. journal of nanomedicine & nanotechnology 6(1) (2015) 1. https://doi.org/10.4172/2157-7439.1000253 [56] g. nabi bidhendi, n. mehrdadi, m. firouzbakhsh. removal of lead from wastewater by iron– benzenetricarboxylate metal-organic frameworks. chemical methodologies 5 (2021) 271-284. https://doi.org/10.22034/chemm.2021.130208 [57] a. dehno khalaji, n. mohammadi, m. emami. nio nanoparticles: synthesis, characterization, and methyl green removal study. progress in chemical and biochemical research 4(4) (2021) 372-378. https://doi.org/10.22034/pcbr.2021.294420.1194 [58] a. obaid, s. al-ghabban, r. al-hussain. appraising antioxidant and antibacterial activities of zinc oxide nanoparticles synthesized biologically by iraqi propolis, chemical methodologies 6(5) (2022) 366-371. https://doi.org/10.22034/chemm.2022.332390.1448 [59] m. pirozmand, a. nezhadali, m. payehghadr, l. saghatforoush. ultratrace determination of cadmium ion in petro-chemical sample by a new modified carbon paste electrode as voltammetric sensor. eurasian chemical communications 2 (2020) 1021-1032. https://doi.org/10.22034/ecc.2020.241560.1063 [60] s. e. mousavi ghahfarokhi, k. helfi, m. zargar shoushtari. synthesis of the single-phase bismuth ferrite (bifeo3) nanoparticle and investigation of their structural, magnetic, optical and photocatalytic properties. advanced journal of chemistry-section a 5(1) (2022) 45-58. https://doi.org/10.22034/ajca.2021.309069.1284 [61] l. qian, s. durairaj, s. prins, a. chen. nanomaterial-based electrochemical sensors and biosensors for the detection of pharmaceutical compounds. biosensors and bioelectronics 175 (2021) 112836. https://doi.org/10.1016/j.bios.2020.112836 [62] v. tallapaneni, l. mude, d. pamu, v.v.s.r. karri. formulation, characterization and in vitro evaluation of dual-drug loaded biomimetic chitosan-collagen hybrid nanocomposite scaffolds. journal of medicinal and chemical sciences 5 (2022) 1059-1074. https://doi.org/10.26655/jmchemsci.2022.6.19 [63] a. kannan, s. radhakrishnan. fabrication of an electrochemical sensor based on gold nanoparticles functionalized polypyrrole nanotubes for the highly sensitive detection of l-dopa. materials today communications 25 (2020) 101330. https://doi.org/10.1016/j.mtcomm.2020.101330 [64] h. ganesha, s. veeresh, y.s. nagaraju, d.s. suresh, h. devendrappa. micelles self-degraded polypyrrole nanotube-cobalt oxide nanocomposite based electrochemical sensor for detection of ascorbic acid. inorganic chemistry communications 145 (2022) 109975. https://doi.org/10.1016/j.inoche.2022.109975 [65] a. j. bard, l. r. faulkner, electrochemical methods: fundamentals and applications, john wiley & sons, new york, 2nd edition, 2001. ©2023 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) https://doi.org/10.4172/2157-7439.1000253 https://doi.org/10.4172/2157-7439.1000253 https://doi.org/10.22034/chemm.2021.130208 https://doi.org/10.22034/pcbr.2021.294420.1194 https://doi.org/10.22034/chemm.2022.332390.1448 https://doi.org/10.22034/ecc.2020.241560.1063 https://doi.org/10.22034/ajca.2021.309069.1284 https://doi.org/10.22034/ajca.2021.309069.1284 https://doi.org/10.1016/j.bios.2020.112836 https://doi.org/10.26655/jmchemsci.2022.6.19 https://doi.org/10.1016/j.mtcomm.2020.101330 https://doi.org/10.1016/j.inoche.2022.109975 http://creativecommons.org/licenses/by/3.0/ encapsulated polycaprolactone with triazole derivatives and selenium nanoparticles as promising antiproliferative and anticancer agents doi: https://doi.org/10.5599/admet.1789 1 admet & dmpk 0(0) (2023) 000-000; doi: https://doi.org/10.5599/admet.1789 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper encapsulated polycaprolactone with triazole derivatives and selenium nanoparticles as promising antiproliferative and anticancer agents ahmed e. abdelhamid1, ahmed a. el-sayed2,*, samira a. swelam2, abdelmohsen m. soliman3 and ahmed m. khalil2 1polymers & pigments department, national research centre, el-bohouth st., dokki 12622, giza, egypt 2photochemistry department, national research centre, 33 el-bohouth st., dokki 12622, giza, egypt 3therapeutic chemistry department, national research centre, 33 el-bohouth st., dokki 12622, giza, egypt *corresponding author: e-mail: ahmedcheme4@yahoo.com; tel.: +2-0100-865-3440 received: march 27, 2023; revised: june 20, 2023; published: june 28,2023 abstract background and purpose: polycaprolactone nanocapsules incorporated with triazole derivatives in the presence and absence of selenium nanoparticles were prepared and evaluated as antiproliferative and anticancer agents. polycaprolactone nanoparticles were prepared using the emulsion technique. experimental approach: the prepared capsules were characterized using ft-ir, tem and dls measurements. the synthesized triazolopyrimidine derivative in the presence and absence of selenium nanoparticles encapsulated in polycaprolactone was tested for its in vitro antiproliferative efficiency towards human breast cancer cell line (mcf7) and murine fibroblast normal cell line (balb/3t3) in comparison to doxorubicin as a standard anticancer drug. key results: the results indicated that encapsulated polycaprolactone with selenium nanoparticles (senps) and triazolesenps were the most potent samples against the tested breast cancer cell line (mcf7). on the other hand, all compounds showed weak or moderate activities towards the tested murine fibroblast normal cell line (balb/3t3). conclusion: as the safety index (si) was higher than 1.0, it expanded the way for newly synthesized compounds to express antiproliferative efficacy against tumour cells. hence, these compounds may be considered promising ones. however, they should be examined through further in-vivo and pharmacokinetic studies. ©2023 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords carbohydrate polymers, metal nanoparticles, breast cancer cell line (mcf7), murine fibroblast normal cell line (balb/3t3) introduction with advances in drug discovery technology, more than 40 % of newly discovered drug candidates have low aqueous solubility that limits their administration roots [1]. the encapsulation technique is considered an efficient alternative to overcome the solubility difficulty and enable the delivery of these compounds into purposed sites in the body. the capsules (micro or nanocapsules) are made up of two parts: the core and the shell material. in general, the core material contains an active ingredient, whereas the shell material protects https://doi.org/10.5599/admet.1789 https://doi.org/10.5599/admet.1789 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:ahmedcheme4@yahoo.com http://creativecommons.org/licenses/by/4.0/ a. e. abdelhamid et al. admet & dmpk 00(0) (2023) 000-000 2 the core material from the outside environment and allows for good release characteristics. micro or nanocapsules are very useful carriers due to their physiological stability and ability to reach local areas by targeting intended tissues or organs. natural and some synthetic polymers are used as encapsulating materials due to their biodegradability and excellent biocompatibility. polycaprolactone (pcl) is widely used for drug delivery and other medical applications with stable properties over time. it is unaffected by ph in aqueous solution [2-4]. polycaprolactone (pcl), either combined with or without poly(ethylene glycol) nanoparticles (np), was synthesized and assessed as potential drug nanocarriers[5]. the investigated nps did neither affect the cells' viability nor elicit an immune response in the raw 264.7 mouse murine macrophage cell line. preparation of paclitaxel-loaded polycaprolactone nanoparticles by nanoprecipitation was investigated for lung cancer treatment [6]. the prepared nanoparticles were coated with chitosan or poly-l-lysine to obtain a cationic surface charge and to increase the cellular interaction. the in vitro release studies showed a prolonged release for up to 96 hours. however, less than 50 % of the therapeutic load was released in the first 24 hours. the medication's core material can be used to treat several diseases such as angina, hypertension, raynaud's phenomenon, and cancer, etc. [7–9]. heterocyclic compounds are well known for their high biological activity, including anticancer capability [10,11]. triazole derivatives isolated from natural products or synthesized in laboratory conditions possess diverse pharmacological properties such as antibacterial, antitubercular, anticancer, and antimalarial activities [12-16]. the most important tools in organic transformations and pharmaceuticals are multicomponent reactions. triazole derivatives based on reacting hydrazinylthienopyrimidine with a number of monosaccharides for antiviral activity against influenza virus h5n1 were tested [15]. moreover, selenium (se) is a substantial nutrient for humans. in food, it can be obtained from various sources, including meats and fish. selenium nanoparticles (senps) (the zero-valent selenium) have recently drawn more scientific attention. this characteristic may be correlated to their obvious biological activities and biosafety properties.senps have a variety of biomedical applications, including drug and targeted gene delivery, anticancer, antibacterial and anti-inflammatory activity, and biosensors [17,18]. many reports describe the synthesis of senps using various methods such as laser ablation, microwave-assisted synthesis, chemical reduction, electrodeposition, solvothermal synthesis, and green synthesis. however, harsh chemical conditions such as acidic ph and high temperatures limit their use in biomedical applications [19-21]. cancer is one of the most important modern health issues because it knows no boundaries and can affect any organ. the world health organization (who) estimates that cancer is among the major causes of roughly many million deaths worldwide [22]. the majority of anticancer medications created to treat cancer have been demonstrated to cause cancer cells to undergo apoptosis [23]. although there are various treatments available, persistent resistance to these treatments focuses on the need for innovative cancer control alternatives [24]. we aim in this study to evaluate the in vitro antiproliferative potency of polycaprolactone nanoparticles encapsulated with triazole derivative and conjugated with senps. there are four formulations prepared in this study; blank polycaprolactone 1, polycaprolactone encapsulated with triazole derivative 2, polycaprolactone encapsulated with triazole-senps 3, polycaprolactone encapsulated with senps 4. the bioactivity assessment of these formulations towards human breast cancer (mcf7) and murine fibroblast normal (balb/3t3) cell lines will be investigated. experimental chemistry stuart melting point apparatus (smp 30) was used to measure melting points. the reactions were monitored using pre-coated (0.25 mm) silica gel plates (merck 60 f254, germany). the spots were visualized using a uv lamp (254 nm). nmr spectra were recorded in (dmso) at 1h nmr (400 mhz) and 13c nmr admet & dmpk 00(0) (2023) 000-000 encapsulated polycaprolactone as antiproliferative and anticancer agents doi: https://doi.org/10.5599/admet.1789 3 (100 mhz) using tms as an internal standard on a bruker nmr spectrometer. mass spectra were performed on the direct inlet part of the mass analyzer in a thermo scientific gcms model isq. the uv-vis (shimadzu spectrophotometer) was used to monitor the formation of selenium nanoparticles in the range of 400 and 700 nm. the shape and size of senps were practically obtained using high-resolution transmission electron microscopy (hrtem) jeol (jem-2100 tem). their diameters were assessed by using imagej software. specimens for tem measurements were prepared by placing a drop of colloidal solution on a 400 mesh carbon-coated copper grid with evaporating the solvent in the air at 25 °c. dynamic light scattering was used to determine the average diameter and size distribution of the encapsulated samples using zeta sizer (nano-zs, malvern instruments ltd., zeta sizer ver, 704, uk). the samples were sonicated for 30-60 minutes before the assessment to obtain a good suspension of the particles in the solution. ethyl-7-acetamido-5-(5-methylfuran-2-yl)-[1,2,4]triazolo[1,5-a]pyrimidine-6-carboxylate (2) a mixture of compound 1 (2.87g, 10 mmol) and acetic anhydride (10 ml) was heated under reflux for 4 h. the reaction mixture was allowed to cool to room temperature and poured into water (100 ml). the solid formed was collected by filtration, dried, and crystallized from ethanol. brown crystals; yield 60 %; mp 120-122 °c, ir (kbr, ν /cm-1); 3436 (nh), 1711, 1659 (c=o), 1054 (c=c); 1h-nmr (dmso-d6, δ / ppm), 1.20-1.28 (t, 3h, j=5.4hz, ch3), 2.33 (s, 3h, ch3), 2.47 (s, 3h,ch3), 4.22-4.28 (q, 2h, j = 5.4hz, ch2), 6.50-6.55 (d, 1h, j = 2.5 hz, -ch, furan ring), 7.43-7.45 (d, 1h, j = 2.5 hz, -ch, furan ring), 7.99 (s, 1h,ch, triazole proton), 10.75 (broad singlet (br s), 1h, nh, d2o exchangeable); 13c-nmr (dmsod6, δ / ppm), 21.46, 22.62, 23.88, 62.40, 95.04, 112.03, 116.05, 138.96, 144.79, 147.51, 157.84, 161.21, 163.05, 168.01, 172.49; ms ((m/z) / %) (329, 21); analyatical calculated for c15h15n5o4 (329.32), c, 54.71; h, 4.59; n, 21.27; found: c, 54.60; h, 4.66; n, 21.19 %. 7-amino-8-methyl-5-(5-methylfuran-2-yl)pyrimido[5,4-e][1,2,4]triazolo[1,5-a]pyrimidin-6 (7h)-one (3). the equimolar amount of compound 2 (3.29, 10 mmol) and hydrazine hydrate (10 mmol) in ethanol (20 ml) was refluxed for 6 h. the precipitate formed was filtered off, washed with cold ethanol, dried and recrystallized from ethanol. brown crystals, yield 60 %; mp 220-222 °c, ir (kbr, ν / cm-1); 3430 (nh2), 1669 (c=o), 1054 (c=c); 1h-nmr (dmso-d6, δ / ppm): 2.12 (s, 3h, -ch3), 2.33 (s, 3h, -ch3), 3.63 (br s, 2h, -nh2 exchangeable with d2o), 6.31-6.37 (d, 1h, j = 2.5hz, -ch, furan ring), 7.38-7.43 (d, 1h, j = 2.5hz, -ch, furan ring), 8.35 (s, 1h, -ch, triazole ring); ms ((m/z)/ %) (297, 58); 13c-nmr (dmso-d6, δ / ppm), 21.17, 21.73, 111.53, 116.89, 136.91, 139.78, 142.89, 143.20, 152.96, 153.57, 155.03, 155.72, 168.29; analytical calculated for c13h11n7o2 (297.28): c, 52.52; h, 3.73; n, 32.98; found: c, 52.40; h, 3.90; n, 32.90 %. 8-methyl-5-(5-methylfuran-2-yl)-7-((2,3,4-trimethoxybenzylidene)amino)-pyrimido[5,4-e][1,2,4]triazolo[1,5a]pyrimidin-6-one (4) an equimolar mixture of compound 3 (10 mmol) and 2,3,4-trimethoxybenzaldehyde (10 mmol) dissolved in ethanol (10 ml) and a catalytic amount of acetic acid was refluxed for eight hours. the reaction mixture was allowed to cool to room temperature. the precipitate so-formed was filtered off, washed with cold ethanol (1 ml) and recrystallized from ethanol. pale yellow crystals yield 74 %; mp 201-203 °c. ir (kbr, ν / cm-1); 1701 (c=o), 1064 (c=c); 1h-nmr (dmso-d6, δ / ppm), 2.36 (s, 3h,ch3), 2.38 (s, 3h, ch3), 3.39 (s, 3h, -och3), 3.41 (s, 3h, -och3), 3.44 (s, 3h, -och3), 6.58-6.59 (d, 1h, j = 2.5hz, -ch, furan ring), 7.57-7.58 (d, 1h, j = 2.5hz, -ch, furan ring), 7.73-7.78(d, 1h, j = 6.7hz , aromatic proton), 8.12-8.15(d, 1h, j = 6.7hz, aromatic proton), 8.08 (s, 1h, ch, triazole ring), 9.02 (s, 1h, -ch=n); ms ((m/z) / %) (475, 68); analyatical calculatedfor c23h21n7o5 (475.47): c, 58.10; h, 4.45; n, 20.62; found: c, 58.22; h, 4.30; n, 20.88 %. https://doi.org/10.5599/admet.1789 a. e. abdelhamid et al. admet & dmpk 00(0) (2023) 000-000 4 synthesis of in-situ selenium nanoparticles (senps) using the synthesized heterocyclic compound 4 selenious acid (h2seo3, 0.013 g, 0.01 mmol.) was dissolved in 80 ml of deionized water. compound 4 (0.01 mmol.) in dmso (10 ml) was added to the solution of selenious acid, then the solution was heated to 60 °c under continuous stirring for 1 h. afterwards, 200 µl of 40 mm ascorbic acid was added as a catalyst to reduce selenium ions into selenium nanoparticles. the formation of selenium nanoparticles was confirmed and characterized by uv-spectrophotometer, particle size, sem and tem. preparation of polycaprolactone nanocapsule the polycaprolactone nanocapsules were prepared using the emulsion technique, as reported in the literature [25]. briefly, polycaprolactone (0.25 g) was dissolved in 10 ml water-immiscible organic solvent with a low boiling point as methylene chloride. the triazole derivative was also dissolved in the same solvent in 5 ml and added dropwise to the pcl solution keeping the ratio of triazole to pcl was 10 %. the produced solution was dropped into 30 ml water containing 0.05 % polyvinyl alcohol (pva) as stabilizing agent under vigorous stirring using a homogenizer at 18.000 rpm for about 30 min. the organic solvent was left to evaporate at room temperature (35 °c) for 24 h under mild stirring to obtain the nanocapsule of pcl containing the triazole derivative. in the case of conjugation of selenium nanoparticles, the selenium nanoparticles were synthesized firstly by reduction of (1 mm) selenium oxide using ascorbic acid (half equimolar ratio) in 0.05 % pva aqueous solution. the obtained suspension solution was centrifuged at 6000 rpm for about 30 min and the precipitate was freeze-dried. anticancer activity evaluation cells 1-human breast cancer cell line (mcf7) was obtained from american type culture collection (rockville, maryland, usa) and is being maintained in the ludwikhirszfeld institute of immunology and experimental therapy (wrocław, poland). cells were cultured in eagle medium (iiet, wroclaw, poland) supplemented with 2 mm l-glutamine, 10 % fetal bovine serum, 8 g/ml of insulin and 1 % minimum essential medium (mem) with non-essential amino acid solution 100(all from sigma–aldrich chemie gmbh, steinheim, germany). 2-murine fibroblast normal cell line (balb/3t3) was cultured in dmem (gibco, uk). it was supplemented with 2 mm l-glutamine, 10 % fetal bovine serum (ge healthcare, logan, ut, usa). all culture media were also supplemented with antibiotics. 100 μg/ml streptomycin (sigma–aldrich chemie gmbh, steinheim, germany) and 100 units/ml penicillin (polfatarchomin sa, warsaw, poland). all cell lines were grown at 37 °c with 5 % co2 humidified atmosphere. an in vitro antiproliferative assay 24 hours before adding the tested compounds, the cells were plated in 96-well plates (sarstedt, germany) at a density of 104 cells per well. the assay was performed after 72 hours of exposure to varying concentrations of the tested compounds. the in vitro cytotoxic effect of all compounds was examined using the srb assay. cytotoxic test srb the details of this technique were described by skehanet al. [26]. the cells were attached to the bottom of plastic wells by fixing them with cold 50 % tca (trichloroacetic acid, sigma-aldrich chemie gmbh, steinheim, germany) on the top of the culture medium in each well. the plates were incubated at 4 °c for 1 hour and then washed five times with tap water. the cellular material fixed with tca was stained with 0.4 % sulphorhodamine b (srb, sigma-aldrich chemie gmbh, steinheim, germany) dissolved in 1 % acetic acid (poch, gliwice, poland) for 30 minutes. unbound dye was removed by rinsing (5 times) in 1 % acetic acid. the admet & dmpk 00(0) (2023) 000-000 encapsulated polycaprolactone as antiproliferative and anticancer agents doi: https://doi.org/10.5599/admet.1789 5 protein-bound dye was extracted with 10 mm buffered tris base (poch, gliwice, poland) for determination of the optical density (λ = 540 nm) in synergy h4 multi-mode microplate reader (biotek instruments usa). the relation between surviving fraction and drug concentration is plotted to get the survival curve for each cell line after the specified time. the concentration required for 50 % inhibition of cell viability (ic50) was calculated and the results are given in table 1. the results were compared to the antiproliferative effects of the reference control doxorubicin. we used the traditional drug doxorubicin as an independent control in the experiments with in vitro antiproliferative activity. table 1. in vitro antiproliferative activity (ic50) of the encapsulated polycaprolactone towards human breast cancer (mcf7) and murine fibroblast normal (balb/3t3) cell lines and the calculated values of the selectivity index (si) of 1 tested encapsulated polycaprolactone, 2 encapsulated polycaprolactone with triazole, 3 encapsulated polycaprolactone with triazole-senps and 4 encapsulated polycaprolactone with senps. compounds ic50 ± sd / µg ml-1 selectivity index human breast cancer (mcf7) murine fibroblast normal (balb/3t3) 1 35.8 ± 6.9 83.5 ± 12.3 2.4 2 24.6 ± 3.4 76.8 ± 10.4 3.1 3 7.2 ± 0.4 38.1 ± 9.2 5.3 4 9.1 ± 0.3 45.3 ± 10.2 4.9 doxorubicin 5.03 ± 0.7 3.8 ± 0.2 0.75 dmso n.a. n.a. n.a. data were expressed as mean ± sd of three independent experiments. ic50: 1 to 10 µg ml-1 very strong, 11 to 25 µg ml-1 strong, 26 to 50 µg ml-1 -moderate; 51 to 100 µg ml-1 – weak; dox: doxorubicin is the drug reference; n.a.: no activity statistical analysis the results are reported as mean ± standard error (se). the experiments investigating the in vitro antiproliferative activity were set in three biological replicates per sample for analysis. results and discussion the synthesis of triazolopyrimidine(s) (1) was reported by a simple addition of an equimolecular mixture of amino triazole, carbonyl compounds and active methylene compounds in an aqueous medium [27]. the triazolopyrimidine (1) was refluxed in acetic anhydride for 4 hours to produce compound (2). the latter showed a new ir peak at 1659 cm-1 (c=o), denoting the resulting amide carbonyl with recent signals for 1h-nmr (dmso-d6, δ/ ppm) and 2.33 (s, 3h, ch3). triazolopyrimidine (2) treated with hydrazine hydrate under reflux formed compound (3) possessing a new ir peak at 1669 cm-1 (c=o), with the disappearance of the acetate carbonyl group. the synthesis of triazolopyrimidine derivative (4) was carried out by the reaction of amino triazole (3) with trimethoxybenzaldehyde in the presence of acetic acid to have new signals for the new three methoxy groups and signals for the new aromatic protons 1h-nmr (dmso-d6, δ / ppm), 3.39 (s, 3h, -och3), 3.41 (s, 3h, -och3), 3.44 (s, 3h, -och3), 7.73-7.78 (d, 1h, j = 6.7 hz , aromatic proton), 8.12-8.15 (d, 1h, j = 6.7 hz, aromatic proton) as shown in scheme 1. synthesis of in-situ selenium nanoparticles using heterocyclic compounds (4) (het-senps) the synthesis of selenium nanoparticles (senps) conjugated with the synthesized compound (4) was performed. the triazole derivative has an adequate balance of low reducing and stabilizing properties of nanoparticles. ascorbic acid was utilized as a catalyst. it is employed to generate senpsby stabilizing their nanostructure upon reducing the se+ cation into se [28-31]. some organic compounds with reductive groups, such as -oh, -sh, -nh, can reduce selenium cation. the pathway to synthesize senps is depicted in figure 1a. uv-vis spectroscopy was used to monitor the formation of senps, as illustrated in figure 1b. according to the surface plasmon of selenium colloidal solution, a resonance peak appears in the absorption spectrum at 465 nm, referring to senps. https://doi.org/10.5599/admet.1789 a. e. abdelhamid et al. admet & dmpk 00(0) (2023) 000-000 6 scheme 1. the proposed steps for the synthesis of triazolopyrimidine derivatives. wavelenght, nm figure 1. (a) the pathway for the synthesis of senps, (b) uv-vis spectrum of senps. the formation of selenium nanoparticles was confirmed by tem analysis [32]. in figure 2a, senpsare displayed as spheres with few agglomerates in a global micrograph. diameter, nm figure 2. tem of senps and their diameter distribution histogram from image analysis. the diameter distribution of these nanoparticles ranges between 25-34 nm, as displayed in figure 2b. the diameter distribution of these nanoparticles was assessed through imagej software in the experimental section. the nanoparticles seem to be apart in an acceptable allocation [33,34]. admet & dmpk 00(0) (2023) 000-000 encapsulated polycaprolactone as antiproliferative and anticancer agents doi: https://doi.org/10.5599/admet.1789 7 the reactive functional groups of pcl and its encapsulated nanoparticles are shown in figure 3. polycaprolactone shows a strong peak at 1736 cm-1 of carbonyl c=o groups. the peak at 1185 corresponds to c-o stretching band. ch symmetrical and asymmetrical vibrations appear at 2868 and 2947 cm−1, respectively. for triazole-encapsulated pcl, there is a small peak at 1640 cm−1 related to c=n of heterocyclic structure. in addition, the peak at 3445 cm−1 indicates o-h or n-h stretching bands. wavenumber, cm-1 figure 3. ftir of pcl, pcl-triazole, pcl-triazole-se and pcl-se nanoparticles. the size and shape of the prepared polycaprolacone nanoparticles were evaluated using tem, as shown in figure 4. one can see in figure 4a that the formed blank pcl was in semi-spherical shape with a major average size of 50 to 90 nm. figure 4b displayed pcl encapsulated triazole derivative with two distinct phases in black and grey color, indicating encapsulated particles and associated with some aggregates. figure 4c indicates the encapsulated triazole with selenium nanoparticles emerged as semi-spherical particles demonstrating black spots in the middle of each particle. their sizes varied from 14 to 144 nm, with the major size around 25 nm. figure 4d showed pcl encapsulated selenium nanoparticles with obvious selenium nanoparticles embedded in spherical polymer particles in the range of 38 to 72 nm. the particle size distribution of the prepared pcl nanoparticles was explored using dynamic light scattering measurement. the measurements of dls based on the scattered beam resulted from the movement of the nanoparticle in the suspension solution. the particle size of the pcl and its encapsulated compounds are sown in figure 5. the size of the blank pcl was large (48 m), indicated as aggregated particles. the encapsulated pcl was around 980 nm and selenium conjugated triazole was 370 nm for selenium pcl, showing two peaks; the major at 110 nm and the minor at 600 nm. larger particle sizes were obtained from dls than in tem. this feature may be attributed to the fact that, in tem, the image revealed a small selected part of the dried sample of the nanoparticles, while in dls, it gives a complete overview of the whole sample, which may be concentrated or aggregated and swelled particles. https://doi.org/10.5599/admet.1789 a. e. abdelhamid et al. admet & dmpk 00(0) (2023) 000-000 8 figure 4. tem of (a) pcl, (b) pcl-triazole, (c) pcl-triazolese and (d) pcl-se nanoparticles. figure 5. particle size distribution from dls for (a) pcl, (b) pcl-triazole, (c) pcl-triazole-se and (d) pcl-se nanoparticles. anticancer performance (antiproliferative activity) the in vitro antiproliferative activity of the encapsulated polycaprolactone samples was investigated towards human breast cancer (mcf7) and murine fibroblast normal (balb/3t3) cell lines via standard established assays. the ic50 value is defined as the concentration of a compound at which 50 % growth inhibition is observed. the selectivity index (si) was calculated for each compound using the equation (1): admet & dmpk 00(0) (2023) 000-000 encapsulated polycaprolactone as antiproliferative and anticancer agents doi: https://doi.org/10.5599/admet.1789 9 si = ic50 for normal cell line (balb/3t3) / ic50 for breast cancer cell line (mcf7) (1) a beneficial si>1.0 indicates that the drug with efficacy against tumour cells is greater than the toxicity against normal cells. both the tumour and normal cell lines showed normal growth in our culture system. dmso did not seem to have any noticeable effect on cellular growth. a gradual decrease in the viability of cancer cells was observed with increasing concentration of the tested compounds, in a dose-dependent inhibitory effect. the data in table 1, in addition to figures 6 and 7, indicate that two of the obtained encapsulated capsules (3 and 4) were shown to be active against the investigated human breast cancer cell line (mcf-7) with strong potency at lower concentrations and a higher relative affinity to tumour cells over normal cells. moreover, the results indicated that the salectivity index (si) was higher than 1.0. hence, we have newly synthesized compounds with higher antiproliferative efficacy against tumour cells. this potency is greater than the toxicity against normal cells. so, it can be concluded that newly synthesized compounds showed a significant potency towards tested cancer cell lines. however, they displayed weak or moderate activities towards normal cells (balb/3t3). figure 6. antiproliferative activity of the encapsulated polycaprolactone towards human breast cancer cell line (mcf7): (a) untreated, (b) polycaprolactone 1, (c) encapsulated polycaprolactone + triazole 2, (d) encapsulated polycaprolactone + triazole-senps 3 and (e) encapsulated polycaprolactone + senps 4. figure 7. antiproliferative activity of the encapsulated polycaprolactone towards murine fibroblast normal cell line (balb/3t3): (a) untreated, (b) polycaprolactone1, (c) encapsulated polycaprolactone + triazole 2, (d) encapsulated polycaprolactone + triazole-senps 3 and (e) encapsulated polycaprolactone + senps 4. https://doi.org/10.5599/admet.1789 a. e. abdelhamid et al. admet & dmpk 00(0) (2023) 000-000 10 regression equation to determine the regression between the ic50 values for the tested two types of cells, we will estimate the correlation coefficient (r), which is used to measure the strength of the relationship between the two ic50 values using equation (2): ( ) ( ) 2 2 2 2 n yx x y r n x x n y y − =    − −               (2) where: r = correlation coefficient, n= number of samples, x = total of all ic50 values for breast cancer cells, y = total of all ic50 values for murine fibroblast normal cells, xy= sum of the product of first and second values, x2=sum of squares of the first value, y2=sum of squares of the second value. the calculation of r is presented in equation (3):    4×320.4-76.7×243.7 -0.3 4×4882.89-5882.89 4×59389.69-59389.69 r = = (3) the present result is the product-moment correlation coefficient (or pearsoncorrelation coefficient). it is known that the value of r always lies between –1 and +1. a value of the correlation coefficient close to +1 indicates a strong positive linear relationship (i.e., one variable increases with the other). a value close to -1 indicates a strong negative linear relationship (i.e., one variable decreases as the other increases). a value close to 0 indicates no linear relationship; however, there could be a nonlinear relationship between the variables. for the regression between the ic50 values of the tested two types of cells presented in table 1, the correlation coefficient = -0.3, indicates a moderate negative linear relationship between the two variables. as shown in figure 7, the antiproliferative activity of the encapsulated polycaprolactone compounds (1, 2, 3 and 4) towards the murine fibroblast normal cell line (balb/3t3) can be compared with the activities of these compounds towards tested cancer cell line along with doxorubicin as the traditional anticancer drug [35]. conclusion selenium nanoparticles were prepared and conjugated with triazole derivatives, then encapsulated into polycaprolactone-forming nanocapsules. these synthesized nanocapsules were investigated to explore their antiproliferative and anticancer activities. the tested nanocapsules (3 and 4) samples exert significant antiproliferative potency towards the human breast cancer cell line (mcf7) by reducing cell proliferation, resulting in a reasonable significant growth inhibitory effect. on the other hand, all tested compounds gave weak or moderate activities towards normal cells (balb/3t3). so, the present study reveals that human breast cancer cells (mcf7) are more sensitive to the tested compounds than normal cells (balb/3t3). these findings indicate a specific antiproliferative potency of these newly synthesized compounds towards cancer cell lines. further pharmacological investigations are needed to study the efficacy of these compounds. conflict of interest: the co-authors declare that there is not any conflict of interest. acknowledgements: the co-authors acknowledge the national research centre (nrc) egypt for funding this work. references [1] j. z. setschenow. über die konstitution der salzlösungen auf grund ihres verhaltens zu kohlensäure. z. physik. chem. 4 (1889) 117-125. https://doi.org/10.1515/zpch-1889-0409. (in german) https://doi.org/10.1515/zpch-1889-0409 admet & dmpk 00(0) (2023) 000-000 encapsulated polycaprolactone as antiproliferative and anticancer agents doi: https://doi.org/10.5599/admet.1789 11 [2] f.a. long, w.f. mcdevit. activity coefficients of nonelectrolyte solutes in aqueous salt solutions. chem. rev. 51 (1952) 119-169. https://doi.org/10.1021/cr60158a004. [3] n. ni, m.m. el-sayed, t.sanghvi, s.h. yalkowsky. estimation of the effect of nacl on the solubility of organic compounds in aqueous solutions. j. pharm. sci. 89 (2000) 1620-1625. https://doi.org/ 10.1002/1520-6017(200012)89:12. [4] s. endo, a. pfennigsdorff, k.-u. goss. salting-out effect in aqueous nacl solutions: trends with size & polarity of solute molecules. environ. sci. technol. 46 (2012) 1496-1503. https://doi.org/10.1021/ es203183z. [5] s.h. lubbad, b.k.a. al-roos, f.s. kodeh. adsorptive–removal of bromothymol blue as acidic–dye probe from water solution using latvian sphagnum peat moss: thermodynamic assessment, kinetic & isotherm modeling. curr. green chem.6 (2019) 53-61. https://doi.org/10.2174/2452273203666 190114144546. [6] m. pudipeddi, a.t.m.serajuddin. trends in solubility of polymorphs. j. pharm. sci. 94 (2005) 929-939. https://doi.org/10.1002/jps.20302. [7] p. friberger, g. åberg. some physicochemical properties of the racemates and optically active isomers of two local anaesthetic compounds.acta pharm. suec. 8 (1971) 361-364. [8] s.k. el-arini, d. giron, h. leuenberger. solubility properties of racemic praziquantel and its enantiomers. pharm. dev. tech. 3 (1998) 557-564. https://doi.org/10.3109/10837459809028638. [9] a. avdeef,e. fuguet, a.llinas, c.ràfols, e. bosch, g. völgyi,t. verbić, e. boldyreva, k. takács-novák. equilibrium solubilitymeasurement of ionizable drugs − consensus recommendations forimproving data quality. admet dmpk 4 (2016) 117−178. https://doi.org/10.5599/admet.4.2.292. [10] w.h.xie, w.y. shiu, d. mackay. a review of the effect of saltson the solubility of organic compounds in seawater. mar. environ. res. 44 (1997) 429-444. [11] a. al-maaieh, d.r. flanagan. salt effects on caffeine solubility, distribution, & self-association. j. pharm. sci. 91 (2002) 1000-1008. https://doi.org/10.1002/jps.10046. [12] e. furia, a. beneduci, l. malacaria, a. fazio, c. la torre, p. plastina.modeling the solubility of phenolic acids in aqueous media at 37oc. molecules 26 (2021) 6500. https://doi.org/10.3390/molecules262 16500. [13] s.h. lubbad, b.k.a. al–roos, f.s. kodeh. adsorptive–removal of bromothymol blue as acidic–dye probe from water solution using latvian sphagnum peat moss: thermodynamic assessment, kinetic & isotherm modeling. curr. green chem. 6 (2019) 53-61. https://doi.org/10.2174/24522732036661 90114144546. [14] g. schill. photometric determination of amines and quaternary ammonium compounds with bromothymol blue. part 2. association of bromothymol blue in aqueous solutions. acta pharm. suec. 1 (1964) 101-122. [15] v.d. gupta, d.e. cadwallader. determination of first pka’ value &partition coefficients of bromothymol blue. j. pharm. sci.57 (1968) 2140-2142. https://doi.org/10.1002/jps.2600571224. [16] g. völgyi, a. marosi, k. takács-novák, a. avdeef. salt solubility products of diprenorphinehydrochloride, codeine and lidocaine hydrochlorides and phosphates – novel method of data analysis not dependent on explicit solubility equations. admet dmpk 1 (2013) 48-62. https://doi.org/10.5599/ admet.1.4.24. [17] a. avdeef. absorption and drug development, 2nd ed., john wiley & sons, inc., hoboken, nj, 2012. isbn 978-1-118-05745-2. [18] a. avdeef. anomalous solubility behavior of several acidic drugs. admet dmpk 2 (2014) 33-42. https://doi.org/10.5599/admet.2.1.30. [19] a.avdeef. phosphate precipitates and water-soluble aggregates in re-examined solubility-phdata of twenty-five basic drugs. admet dmpk 2 (2014) 43-55. https://doi.org/10.5599/admet.2.1.31. [20] a. avdeef. suggested improvements for measurement of equilibrium solubility-ph of ionizable drugs. admet dmpk 3 (2015) 84-109. https://doi.org/10.5599/admet.3.2.193. https://doi.org/10.5599/admet.1789 https://doi.org/10.1021/cr60158a004 https://doi.org/‌10.1002/1520-6017(200012)89:12 https://doi.org/‌10.1002/1520-6017(200012)89:12 https://doi.org/10.1021/es203183z https://doi.org/10.1021/es203183z https://doi.org/10.2174/2452273203666190114144546 https://doi.org/10.2174/2452273203666190114144546 https://doi.org/10.1002/jps.20302 https://doi.org/10.3109/10837459809028638 https://doi.org/10.5599/admet.4.2.292 https://doi.org/10.1002/jps.10046 https://doi.org/10.3390/molecules26216500 https://doi.org/10.3390/molecules26216500 https://doi.org/10.2174/2452273203666190114144546 https://doi.org/10.2174/2452273203666190114144546 https://doi.org/10.1002/jps.2600571224 https://doi.org/10.5599/admet.1.4.24 https://doi.org/10.5599/admet.1.4.24 https://doi.org/10.5599/admet.2.1.30 https://doi.org/10.5599/admet.2.1.31 https://doi.org/10.5599/admet.3.2.193 a. e. abdelhamid et al. admet & dmpk 00(0) (2023) 000-000 12 [21] g. butcher, j. comer, a. avdeef. pka-critical interpretations of solubility–ph profiles: pg-300995 and nsc-639829 case studies. admet dmpk 3 (2015) 131-140. https://doi.org/10.5599/admet.3.2.182. [22] a. pobudkowska, c. ràfols, x. subirats, e. bosch, a. avdeef. phenothiazines solution complexity – determination of pka and solubility-ph profiles exhibiting sub-micellar aggregation at 25 and 37 oc.eur. j. pharm. sci. 93 (2016) 163-176.https://doi.org/10.1016/j.ejps.2016.07.013. [23] c.a.s. bergström, a. avdeef. perspectives in solubility measurement and interpretation. admet dmpk 7 (2019) 88-105. http://dx.doi.org/10.5599/admet.686. [24] o.s. marković, m.p. pešić, a.v. shah, a.t.m. serajuddin, t.ž. verbić, a. avdeef. solubility-ph profile of desipramine hydrochloride in saline phosphate buffer: enhanced solubility due to drug-buffer aggregates. eur. j. pharm. sci. 133 (2019) 264–274. https://doi:10.1016/j.ejps.2019.03.014. [25] e. fuguet, x. subirats, c. ràfols, e. bosch, a. avdeef. ionizable drug self-associations and the solubility dependence on ph: detection of aggregates in saturated solutions using mass spectrometry (esi-qtof-ms/ms). mol. pharmaceutics 18 (2021) 2311-2321. https://doi.org/10.1021/acs.molpharma ceut.1c00131. [26] o.s. marković, n.g. patel, a.t.m. serajuddin, a. avdeef, t.ž. verbić. nortriptyline hydrochloride solubility-phprofiles in a saline phosphate buffer: drug-phosphate complexes and multiple phmaxdomains with a gibbs phase rule"soft" constraints. mol. pharmaceutics 19 (2022) 710-719. https://doi.org/10.1021/acs.molpharmaceut.1c00919. [27] a. avdeef, j.j. bucher. accurate measurements of the concentration of hydrogen ions with a glass electrode: calibrations using the prideaux and other universal buffer solutions and a computer-controlled automatic titrator.anal. chem. 50 (1978) 2137-2142. https://doi.org/10.1021/ac50036a045. [28] f.h. sweeton, r.e. mesmer, c.f. baes, jr. acidity measurements at elevated temperatures. 7. dissociation of water.j. solut. chem. 3 (1974) 191-214. https://doi.org/10.1007/bf00645633. [29] m.h. abraham, j. le. the correlation and prediction of the solubility of compounds in water using an amended solvation energy relationship.j. pharm. sci. 88(1999) 868-880.https://doi.org/10.1021/ js9901007. [30] j.a. platts, d.butina, m.h.abraham,a. hersey. estimation of molecular linear free energy relation descriptors using a group contribution approach. j. chem. inf. comput. sci. 39 (1999) 835-845. https://doi.org/10.1021/ci980339t. [31] a. avdeef, m. kansy. predicting solubility of newly-approved drugs (2016–2020) with a simple absolv and gse(flexible-acceptor) consensus model outperforming random forest regression. j. solution chem. 51 (2022) 1020-1055. https://doi.org/10.1007/s10953-022-01141-7. [32] t. shimada, k. tochinai, t. hasegawa. determination of ph dependent structure of thymol blue revealed by cooperative analytical method of quantum chemistry and multivariate analysis of electronic absorption spectra. bull. chem. soc. jpn. 92 (2019) 1759-1766. https://doi.org/10.1246/ bcsj.20190118. [33] k. yamaguchi, z. tamura, m. maeda. (1997). molecular structure of bromophenol blue having a γsultone ring. anal. sci. 13 (1997) 1057-1058. https://doi.org/10.2116/analsci.13.1057. [34] p. balderas-hernández, m.t. ramírez, a. rojas-hernández, a. gutiérez. determination of pka’s for thymol blue in aqueous medium: evidence of dimer formation. talanta 46 (1998) 1439-1452. https://doi.org/10.1016/s0039-9140(98)00015-0. [35] t.ž. verbić, k.y. tam, d.ž. veljković, a.t.m. serajuddin, a. avdeef. clofazimine pkadetermination by potentiometry and spectrophotometry – reverse cosolvent dependence as an indicator of presence of dimers in aqueous solutions. mol. pharmaceutics (2023). https://doi.org/10.1021/acs.molphar maceut.3c00172. ©2023 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) https://doi.org/10.5599/admet.3.2.182 https://doi.org/10.1016/j.ejps.2016.07.013 http://dx.doi.org/10.5599/admet.686 https://doi:10.1016/j.ejps.2019.03.014 https://doi.org/10.1021/acs.molpharmaceut.1c00131 https://doi.org/10.1021/acs.molpharmaceut.1c00131 https://doi.org/10.1021/acs.molpharmaceut.1c00919 https://doi.org/10.1021/ac50036a045 https://doi.org/10.1007/bf00645633 https://doi.org/10.1021/js9901007 https://doi.org/10.1021/js9901007 https://doi.org/10.1021/ci980339t https://doi.org/10.1007/s10953-022-01141-7 https://doi.org/10.1246/bcsj.20190118 https://doi.org/10.1246/bcsj.20190118 https://doi.org/10.2116/analsci.13.1057 https://doi.org/10.1016/s0039-9140(98)00015-0 https://doi.org/10.1021/acs.molpharmaceut.3c00172 https://doi.org/10.1021/acs.molpharmaceut.3c00172 http://creativecommons.org/licenses/by/3.0/ impact of host factors on susceptibility to antifungal agents doi: https://dx.doi.org/10.5599/admet.1164 153 admet & dmpk 10(2) (2022) 153-162; doi: https://doi.org/10.5599/admet.1164 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper impact of host factors on susceptibility to antifungal agents balbina plotkin 1 *, monika konaklieva 2 1 department of microbiology and immunology, college of graduate studies, midwestern university, downers grove, il 60515 2 department of chemistry, american university, college of arts and sciences, 4400 massachusetts ave, nw, washington, dc 20016 *corresponding author: e-mail: bplotk@midwestern.edu; tel.: +1-630-515-6163; fax: +1-630-515-7249 received: november 08, 2021; accepted: december 30, 2021; published: january 07, 2022 abstract an obstacle to drug development, particularly in this era of multiple drug resistance, is the underappreciation for the role the host environment plays in microbial response to drugs. with the rise in fungal infection and drug resistance, particularly in individuals with co-morbidities, the influence serum and its components have on antimicrobial susceptibility requires assessment. this study examined the impact of physiologically relevant glucose and insulin levels in the presence and absence of 50 % human plasma on mics for clinical isolates of candida lusitaniae, candida parapsilosis, candida albicans, candida tropicalis, candida glabrata, candida krusei and cryptococcus neoformans. the addition of insulin or glucose at physiologic levels in rpmi medium alone altered the mic in either a positive or negative fashion, depending on the organisms and drug tested, with c. glabrata most significantly altered with a 40, >32 and 46-fold increase in mic for amphotericin b, itraconazole and miconazole, respectively. the addition of candida-antibody negative plasma also affected mic, with the addition of glucose and insulin having a tandem effect on mic. these findings indicate that phenotypic resistance of candida and cryptococcus can vary depending on the presence of insulin with glucose and plasma. this modulation of resistance may help explain treatment failures in the diabetic population and facilitate the development of stable drugresistant strains. furthermore, these findings indicate the need for a precision approach in the choice of drug treatment and drug development. ©2022 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords insulin; glucose; human serum; candida; cryptococcus. introduction phenotypic response to environmental shifts is essential for microbial survival and is of particular importance with respect to changes that occur in vivo, which affect response to antimicrobial agents. the potential changes in response to drugs upon growth within the host could both be responsible for treatment failures and provide the potential for novel drug development. thus, it is essential to define these phenotypic responses to host factors, and metabolic conditions, to predict effective drug design and utility. a common metabolic condition associated with increased risks for infections is type ii diabetes. people https://dx.doi.org/10.5599/admet.1164 https://doi.org/10.5599/admet.1164 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:bplotk@midwestern.edu http://creativecommons.org/licenses/by/4.0/ plotkin and konaklieva admet & dmpk 10(2) (2022) 153-162 154 with type ii diabetes can exhibit hyperglycemia, hyperinsulinemia, and/or combined hyperglycemiahyperinsulinemia, as a result of insulin intolerance [1,2]. a metabolic immunosuppressive condition, type ii diabetes, places individuals at increased risk for fungal diseases [3-6]. thus, during the infectious process, a pathogen can be exposed to elevated levels of insulin and/or glucose as compared to the metabolically normal host. a critical disconnect regarding predicted sensitivity via in vitro testing and efficacy is the lack of congruency with the metabolic disorders most commonly at risk for fungal infections, e.g., type ii diabetes, since antimicrobial activity testing is predicated on the host being in a normal metabolic state. most studies examining serum effects are focused on the impact normal serum/plasma has on antimicrobial activity, or on the effect of high glucose (2 mg/dl) concentrations, which are reported to affect the ability of antifungal agents to bind to their targets [7-9]. to date, studies of glucose-fungal-antifungal agent interactions fail to take into consideration the concomitant presence of insulin with glucose. this becomes important since organisms across the taxonomic kingdoms are reported to produce and/or react to insulin [10-16]. in candida albicans, insulin has been associated with increased expression of virulence-associated morphological transition, from blastospore to hyphal production [17]. the mechanism via which insulin promotes yeast to hyphal transition is by enhancing uptake of proline by c. albicans. conversely, glucose concentrations can affect morphogenic expression [18,19]. in the present study, the effect on in vitro antifungal activity in the presence and absence of human plasma by insulin and/or glucose, at levels analogous to those reported in type ii diabetes, was determined. experimental fungal species and culture candida lusitaniae, candida parapsilosis, candida albicans, candida tropicalis, cryptococcus neoformans, candida glabrata, and candida krusei clinical isolates (a generous gift from the lab of paul schreckenberger, loyola university medical school) were maintained at -70 °c and passed at least twice on sabouraud dextrose agar (sda) prior to testing. drug testing antifungal activity determinations were done using the standard clsi (nccls m27-t) microbroth dilution method with modifications [20,21]. susceptibility to amphotericin b (sigma-aldrich), fluconazole (pfizer), itraconazole (janssen pharmaceuticals), and miconazole (sigma-aldrich) was determined using antifungal agents constituted and serially diluted. microtiter plates containing the serial dilutions of the antifungal agents were made as recommended and stored frozen (-70 °c; < 7 days) or refrigerated (itraconazole; < 3 days). antifungal activity in the presence or absence of normal human plasma (type a, rh +), was performed using 2x rpmi medium (50 % v/v). plasma was screened for specific fungal antibodies, as described below, by indirect immunofluorescent microscopy (fitc-protein a) to detect the presence of yeast-bound antibodies. minimum lethal concentrations (mlc) were determined by plating 10 µl of all wells showing no growth onto sda. to simulate glucose and insulin levels reported for type ii diabetes, glucose (285 mg/dl) and/or insulin (200 µu/ml), final concentrations were tested in rpmi with and without 50 % plasma medium. endogenous glucose levels were determined using a kit according to the manufacturer’s instructions (sigma-aldrich). the reported mid-range value for normal human insulin levels (18 µu/ml) was used as the basis for calculating the amount of insulin to add. glucose and/or insulin in rpmi medium had no effect (data not admet & dmpk 10(2) (2022) 153-162 host factors on susceptibility to antifungal agents doi: https://dx.doi.org/10.5599/admet.1164 155 shown) on fungal growth kinetics as determined by both turbidimetric measurements (absorbance 560 nm) and a direct microscopic count of yeast and hyphal forms. the growth rates and the relative production of yeast and hyphal forms during growth in rpmi medium with and without glucose and/or insulin were measured to determine if changes in the mic and mlc of the antifungal agents tested in the presence of glucose and/or insulin could be correlated to alterations in fungal growth. no significant differences in generation time or proportion of yeast and hyphal forms were measured. the differences in the rate of growth between organisms also appear to be unrelated to the alterations in mic and mlc measured in the presence of glucose and/or insulin and to the relative susceptibility of the organisms. the antifungal agents were not exposed to the media supplements (glucose and insulin) until organisms were added to the agent. sterility controls and growth controls for all test conditions were run simultaneously with the tests. to determine antifungal activity in the presence of normal human plasma (type a, rh +), a modified protocol was developed. plasma was screened for specific fungal antibodies by indirect immunofluorescent microscopy using fitc-protein a to detect the presence of bound antibodies. equal volumes of normal human plasma were added to 2x nccls rpmi medium to yield a medium that contained 1x rpmi with 50 % v/v of plasma. for glucose in a medium containing 50 % plasma (plasma medium), glucose and/or insulin were added to ensure equivalent amounts were used for the medium alone. antifungal agents were prepared, as described above, in this plasma medium. antifungal activity in the presence and absence of plasma was determined on the same day using the same drug lot and microbial suspension. minimum inhibitory concentrations (mic) were determined by visual inspection of the wells at 24, 36 and 72 hrs after initiation of the incubation period. minimum lethal concentrations (mlc) were determined by plating 10 μl of all wells showing no growth onto sabouraud dextrose agar (difco) at 36 hrs post-initiation of the incubation period. results and discussion amphotericin b the overall effect of insulin and glucose alone or together were drug and fungal species-specific (table 1). amphotericin b (amb) activity was relatively unaffected by insulin and glucose combined with the exception of c. glabrata (40-fold greater mic than rpmi alone; table 1). for amb, the presence of plasma with and without glucose and/or insulin decreased the mic < 4 fold of that measured for medium alone for all organisms, with the exception of c. glabrata and c. neoformans. for amphotericin b in plasma supplemented with glucose, its mic increased from four to a maximum of 32-fold for c. krusei compared to medium with plasma alone (table 2). like with itraconazole, with amphotericin b, there was a difference between the mic and mlc. the mlc of c. albicans was 32-fold higher than the mic in the presence of plasma without supplements. the plasma effect on the candidal mlc was overcome by the addition of glucose and/or insulin. the mlc was within 2-fold of the amphotericin b mic for all other isolates. comparison of medium with plasma to medium alone (table 3) indicated that the addition of 50 % normal human plasma to medium decreased the mic of amphotericin b > 16-fold for all isolates, except c. lusitaniae and c. krusei, which were unaffected. 5-fluorocytosine the mic of 5-fluorocytosine in media supplemented with glucose alone either was unaffected (c. lusitaniae, c. neoformans, and c. krusei), or decreased (table 1). the effect of insulin on the mic in media was also isolate-dependent. insulin alone had the greatest negative impact on 5-fluorocytosine activity for c. tropicalis (~25-fold), with c. parasilosis and c. glabrata exhibiting a ~6-fold increase in mic. in general, https://dx.doi.org/10.5599/admet.1164 plotkin and konaklieva admet & dmpk 10(2) (2022) 153-162 156 the mic of all isolates in medium supplemented with glucose and insulin was decreased compared to medium alone, although the degree of decrease was isolate-dependent. the presence of plasma (50 % v/v rpmi medium) resulted in an increase in 5-fluorocytosine mic maximally for c. parasilosis, (32-fold) and c. tropicalis (16-fold), as compared to rpmi medium alone for all isolates, with the exception of c. glabrata, whose mic was markedly decreased (table 2). the mlc in media, and in the presence of plasma with or without supplements correlated within two-fold of the mic, with the exception of c. glabrata, which in the presence of plasma was four-fold higher than that measured for media alone. the addition of insulin to both medium alone, and with plasma, resulted in an increase in the mic to a maximum increase measured for c. glabrata when exposed to 5-fluorocytosine (40-fold increase; 25 μg/ml in media with plasma and insulin vs. 0.625 μg/ml media with insulin alone). comparison of medium with plasma, to medium alone (table 3), indicated that the addition of 50 % normal human plasma to medium decreased the mic 25-fold for c. tropicalis only (0.25 μg/ml media with 50 % plasma vs. 6.25 μg/ml rpmi medium alone). fluconazole supplementation of rpmi medium containing plasma with insulin, resulted in a 1024-fold increase in mic of c. albicans to fluconazole with the resultant reclassification from sensitive to resistant. however, addition of glucose to insulin restored the sensitivity of c. albicans to fluconazole. the only change in classification with regards to fluconazole was a shift from sensitive to sensitive-dose dependent for c. lusitaniae when both glucose and insulin were present. in contrast, the presence of plasma resulted in a shift from resistant to sensitive for c. neoformans with respect to fluconazole. itraconazole for itraconazole, the addition of glucose and/or insulin to rpmi medium did not significantly alter the mic or mlc (< 2-fold of mic in medium alone) for all organisms with the exception of c. lusitaniae and c. glabrata (table 1). the addition of glucose and/or insulin to rpmi medium resulted in a mic for c. glabrata that exceeded the drug solubility level (see table 1). the effect of glucose on the c. lusitaniae response to itraconazole was opposite of its effect when combined with insulin (8-fold higher mic for glucose alone vs. 8-fold lower mic for glucose and insulin), indicating that there may be an interactive effect of glucose and insulin. the presence of plasma alone had a modest (< 4-fold) effect on the mic for all isolates with the exception of c. glabrata, where the mic was decreased nearly 103-fold as compared to that measured for medium alone (table 2 and 3). the addition of glucose and/or insulin to plasma did not significantly alter the mic or mlc of itraconazole for all organisms, except for c. neoformans, where supplementing plasma with insulin resulted in a 20-fold decrease in the mic. interestingly, the mlc for all conditions tested differed < 2-fold from the mic for all organisms, except for c. neoformans (8-fold higher than mic in plasma supplemented with glucose and insulin) and c. glabrata (128-fold higher than mic in plasma supplemented with glucose). the mlc for c. neoformans in rpmi alone was less affected, exhibiting a fivefold decrease in mlc compared to an mlc measured in plasma alone, or plasma supplemented with glucose. this differential effect of the presence of plasma, as compared to medium alone, may explain the treatment failure reported despite laboratory-confirmed sensitivity [22-27]. miconazole the combination of glucose and insulin resulted in a significant alteration in the miconazole mic for all isolates except c. parapsilosis. resistance to miconazole was increased by the presence of insulin and glucose to a maximum of 46-fold and 23-fold over that of rpmi alone for c. glabrata and c. krusei, admet & dmpk 10(2) (2022) 153-162 host factors on susceptibility to antifungal agents doi: https://dx.doi.org/10.5599/admet.1164 157 respectively (table 1). the mic in media supplemented with glucose was decreased from four-fold to a maximum 16-fold measured for c. albicans in miconazole. c. neoformans was the only organism whose mic was affected by the presence of insulin. interestingly, while plasma alone significantly increased drug activity for all drugs tested against c. glabrata, the addition of insulin and/or glucose restored this activity to the levels measured for rpmi with plasma alone (tables 2 and 3). in contrast, the presence of plasma negatively impacted miconazole (c. lusitaniae, > 185-fold; c. tropicalis, > 23-fold; c. krusei, > 92-fold) and with the addition of glucose and insulin, mic returned to levels within 4-fold that of rpmi with plasma. the addition of glucose to plasma caused either a decrease in the mic or did not alter the mic. in all test situations, the mlc correlated with the mic. table 1. effect of insulin and glucose on fungal mic (g/ml) organism drug rpmi rpmi + gluc c ratio rpmi + gluc rpmi rpmi + insu d ratio rpmi + insu rpmi rpmi + glucose & insulin ratio rpmi + gluc & insu rpmi c. lusitaniae amb a 0.039 (s) b 0.0195 (s) 0.5 0.0195 (s) 0.5 0.078 (s) 2 5-fc a 0.15625 (s) 0.15625 (s) 1 0.15625 (s) 1 0.00936 (s) 0.0599 flucon a 4.0 (s) 8.0 (s) 2 2.0 (s) 0.5 16.0 (s-dd) 4 itra a 0.5 (s-dd) b 4.0 (r) b 8 0.5 (s-dd) 1 0.0625 (s) 0.125 micon a 0.195 (s) 0.195 (s) 1 0.195 (s) 1 2.2425 (s) 11.5 c. parapsilosis amb 0.078 (s) 0.0195 (s) 0.25 0.039 (s) 0.5 0.078 (s) 1 5-fc 0.78 (s) 0.39 (s) 0.5 4.99 (i) b 6.4 0.312 (s) 0.4 flucon 2.0 (s) 8.0 (s) 4 2.0 (s) 1 8.0 (s) 4 itra 0.25 (s-dd) 0.25 (s-dd) 1 0.25 (s-dd) 1 0.25 (s-dd) 1 micon 0.39 (s) 0.78 (s) 2 0.39 (s) 1 0.546 (s) 0.71 c. albicans amb 0.039 (s) 0.0195 (s) 0.5 0.0195 (s) 0.5 0.039 (s) 1 5-fc 0.78 (s) 0.1 (s) 0.1299 0.39 (s) 0.5 0.078 (s) 0.1 flucon 0.125 (s) 0.125 (s) 1 0.125 (s) 1 0.5 (s) 4 itra 0.03125 (s) 0.0156 (s) (2) 0.03125 (s) 1 0.0625 (s) 2 micon 0.39 (s) 0.0245 (s) 0.0629 0.78 (s) 2 0.01757 (s) 0.045 c. tropicalis amb 0.039 (s) 0.0195 (s) 0.5 0.039 (s) 1 0.156 (s) 4 5-fc 0.195 (s) 0.0975 (s) 0.5 4.99 (i) 25.6 0.156 (s) 0.8 flucon >128.0 (r) >128.0 (r) >1 >128 (r) >1 >128 (r) >1 itra >64.0 (r) >64.0 (r) >1 >64 (r) >1 >64 (r) >1 micon 1.56 (s) 0.78 (s) 0.5 1.56 (s) 1 9.048 (r) 5.8 c. neoformans amb 0.039 (s) 0.0195 (s) 0.5 0.039 (s) 1 0.039 (s) 1 5-fc 6.25 (i) b 6.25 (i) 1 0.78 (s) 0.125 0.78 (s) 0.125 flucon 8.0 (r) 8.0 (r) 1 8.0 (r) 1 8.0 (r) 1 itra 0.5 (s-dd) 0.5 (s-dd) 1 0.5 (s-dd) 1 0.5 (s-dd) 1 micon 0.9766 (s) 0.1953 (s) 0.2 0.09766 (s) 0.1 4.492 (s) 4.6 c. glabrata amb 0.078 (s) 0.0195 (s) 0.25 0.039 (s) 0.5 3.12 (r) 40 5-fc 8.0 (i) 4.0 (i) 0.5 51.2 (r) 6.4 1.6 (i) 0.2 flucon 0.09766 (s) 0.3906 (s) 4 0.195 (s) 2 0.195 (s) 2 itra 2.0 (r) >64 (r) >32 >64 (r) >32 >64 (r) >32 micon 0.39 (s) 0.78 (s) 2 0.39 1 17.94 (r) 46 c. krusei amb 0.078 (s) 0.039 (s) 0.5 0.078 (s) 1 0.078 (s) 1 5-fc 12.5 (i) 12.5 (i) 1 12.5 (i) 1 5.0 (i) 0.4 flucon 64.0 (r) 64.0 (r) 1 64.0 (r) 1 64.0 (r) 1 itra 1.0 (r) 1.0 (r) 1 1.0 (r) 1 1.0 (r) 1 micon 0.39 (s) 0.78 (s) 2 1.56 (s) 4 8.97 (r) 23 a amphotericin b (amb), flucytosine (5-fc) fluconazole (flucon), itraconazole (itra) , and miconazole (micon) b s=sensitive; r=resistant; i=intermediate resistance; s-dd= sensitive based on reported disc diffusion concentration c gluc=glucose; d insu=insulin https://dx.doi.org/10.5599/admet.1164 plotkin and konaklieva admet & dmpk 10(2) (2022) 153-162 158 table 2. effect of insulin and glucose in rpmi medium with human plasma (50 % v/v; rpmi-p) on the mic (g/ml) organism drug rpmi-p c ratio rpmi-p rpmi rpmi-p & glu d ratio rpmi-p & gluc rpmi-p rpmi-p & insu e ratio rpmi-p & insu rpmi-p rpmi-p, gluc & insu ratio rpmi-p, gluc & insu rpmi-p c. lusitaniae amb a 0.078 (s) b 2 9.75x10 -3 (s) 0.25 0.0195 (s) 0.25 0.0195 (s) 0.25 5-fc a 0.78 (s) 5 0.78 (s) 1 0.78 (s) 1 0.78 (s) 1 fluco n a 1.0 (s) 0.25 1.0 (s) 1 4.0 (s) 4 4.0 (s) 1 itra a 2.0 (r) b 4 2.0 (r) 1 2.0 (r) 1 2.0 (r) 1 micon a >36.075 (r) >185 >36.075 (r) >1 0.3608 (s) 0.01 >36.08 (r) >1 c. parapsilosis amb 0.0195 (s) 0.25 0.078 (s) 4 0.039 (s) 2 0.0195 (s) 1 5-fc 24.96 (i) b 32 24.96 (i) 1 24.96 (i) 1 24.96 (i) 1 fluco n 1.0 (s) 0.5 0.5 (s) 0.5 8.0 (s) 8 0.5 (s) 0.5 itra 0.125 (s) 0.5 0.125 (s) 1 0.125 (s) 1 0.5 (s-dd) 2 micon >39.88 (r) >92 >39.88 (r) 1 >39.88 (r) >1 >39.88 (r) >1 c. albicans amb 0.00975 (s) 0.25 0.0195 (s) 2 0.0195 (s) 2 0.0195 (s) 2 5-fc 3.12 (s) 4 3.12 (s) 1 3.12 (s) 1 3.12 (s) 1 fluco n 0.075 (s) 0.5 0.075 (s) 1 >6.8 (s-r) 1024 0.3 (s) 4 itra 0.03125 (s) 1 0.03125 (s) 1 0.03125 (s) 1 0.03125 (s) 1 micon 0.39 (s) 1 0.065 (s) 0.166 0.0325 (s) 0.083 0.78 (s) 2 c. tropicalis amb 0.0195 (s) 0.5 0.0195 (s) 1 0.0195 (s) 1 0.0195 (s) 1 5-fc 3.12 (s) 16 3.12 (s) 1 3.12 (s) 1 3.12 (s) 1 fluco n >128 (r) >1 >128 (r) 1 >128.0 (r) >1 >128.0 (r) 1 itra >64 (r) 1 >64 (r) 1 >64.0 (r) 1 >64.0 (r) 1 micon >35.88 (r) >23 >35.88 (r) 1 >35.88 (r) >1 >35.88 (r) >1 c. neoformans amb 0.00488 (s) 0.125 0.00975 (s) 2 0.00975 (s) 1 0.00975 (s) 2 5-fc 25.0 (i-r) 4 12.5 (i) 0.5 12.5 (i) 0.5 12.5 (i) 0.5 fluco n 1.0 (s) 0.125 0.25 (s) 0.25 1.0 (s) 1 2.0 (s) 2 itra 0.25 (sdd) b 0.5 0.25 (s-dd) 1 0.0125 (s) 0.05 0.25 (s-dd) 1 micon 1.953 (s) 2 0.3906 (s) 0.2 0.078 (s) 0.04 7.812 (r) 4 c. glabrata amb 3.71x10 -5 (s) 4.76x10 -4 3.7 x 10 -5 (s) 1 3.71x10 -5 (s) 1 3.71x10 -5 (s) 1 5-fc 0.0615 (s) 7.69x10 -3 <0.0615 (s) <1 <0.0615 (s) <1 0.246 (s) 4 fluco n <2.4x10 5 (s) <2.4x10 -4 <2.4x10 -5 (s) <1 <2.4x10 -5 (s) 1 <2.4x10 -5 (s) <1 itra 0.0194 (s) 9.7x10 -3 0.0194 (s) 1 0.0194 (s) 1 0.0194 (s) 1 micon 1.08x10 -3 (s) 2.8x10 -3 5.41 x 10 -4 (s) 0.5 0.0964 (s) 89 0.001083 (s) 1 c. krusei amb 0.156 (s) 2 0.624 (s) 4 0.156 (s) 1 0.156 (s) 1 5-fc >25.0 (i-r) >2 >25.0 (i-r) 1 >25.0 (i-r) >1 >25.0 (i-r) >1 fluco n 64.0 (r) 1 64.0 (r) 1 64.0 (r) 1 64.0 (r) 1 itra 4.0 (r) 4 4.0 (r) 1 4.0 (r) 1 4.0 (r) 1 micon 35.88 (r) >92 8.97 (r) 0.25 574.08 (r) 16 >35.88 (r) >1 a amphotericin b (amb), flucytosine (5-fc) fluconazole (flucon), itraconazole (itra), and miconazole (micon) b s=sensitive; r=resistant; i=intermediate resistance; s-dd= sensitive based on reported disc diffusion concentration c p=plasma d gluc=glucose e insu=insulin thus, the addition of plasma to the medium resulted in an overall decrease in the concentration of miconazole, itraconazole, 5-fluorocytosine, and amphotericin b needed to inhibit the growth of c. glabrata (table 3). in contrast, plasma overall caused an increase in the concentration of miconazole, itraconazole, 5-fluorocytosine, and amphotericin b needed to inhibit the growth of c. lusitaniae, c. tropicalis and c. krusei. the presence of the medium supplements insulin and/or glucose does not affect fungal mic or mlc in a predictable manner regarding specific antifungal agents or organisms tested. the growth rate and relative production of yeast and/or hyphal forms in medium with and without glucose and/or insulin do not appear to be related to changes in the organism's mic or mlc. admet & dmpk 10(2) (2022) 153-162 host factors on susceptibility to antifungal agents doi: https://dx.doi.org/10.5599/admet.1164 159 table 3. ratio of effects of insulin and glucose on antifungal activity of amphotericin b, flucytosine, fluconazole, itraconazole, and miconazole in the presence of rpmi with 50 % human plasma to homologous rpmi medium alone organism drug ratio rpmi-plasma rpmi ratio rpmi-p c & gluc d rpmi-gluc ratio rpmi-p & insu e rpmi-insu ratio rpmi-p, gluc & insu rpmi-gluc & insu c. lusitaniae amb 2.0 0.5 1.0 0.25 5-fc 5.0 5.0 5.0 83.3 flucon 0.25 0.125 2.0 0.25 itra 4.0 0.5 4.0 32.0 micon >185.0 >185.0 1.85 16.1 c. parapsilosis amb 0.25 4.0 1.0 0.25 5-fc 32.0 64.0 5.0 80.0 flucon 0.5 0.0625 4.0 0.0625 itra 0.5 0.48 0.5 2.0 micon >92.0 >51.13 >102.3 >73 c. albicans amb 0.25 1.0 1.0 0.5 5-fc 4.0 31.2 8.0 40 flucon 0.5 0.6 54.4 0.6 itra 1.0 2.0 1.0 0.5 micon 1.0 2.65 0.042 44.4 c. tropicalis amb 0.5 1.0 0.5 0.125 5-fc 16.0 32.0 0.625 20.0 flucon >1.0 1.0 1.0 1.0 itra 1.0 1.0 1.0 1.0 micon >23.0 46.0 23.0 3.97 c. neoformans amb 0.125 0.5 0.25 0.25 5-fc 4.0 2.0 16.0 16.0 flucon 0.125 0.03 0.125 0.25 itra 0.5 0.5 0.025 0.5 micon 2.0 2.0 0.7987 1.74 c. glabrata amb 4.76x10 -4 1.897x10 -3 9.5x10 --4 1.2x10 -5 5-fc 7.69x10 -3 1.5x10 -2 1.2x10 -3 1.54x10 -1 flucon <2.4x10 -4 6.1x10 -5 1.23x10-4 1.23x10 -4 itra 9.7x10 -3 3x10 -4 3x10 -4 3x10 -4 micon 2.8x10 -3 6.9x10 -4 2.47x10 -1 6.0x10 -5 c. krusei amb 2.0 16.0 2.0 2.0 5-fc >2.0 2.0 2.0 5.0 flucon 1.0 1.0 1.0 1.0 itra 4.0 4.0 4.0 4.0 micon >92.0 11.5 368.0 4.0 a amphotericin b (amb), flucytosine (5-fc) fluconazole (flucon), itraconazole (itra), and miconazole (micon) b s=sensitive; r=resistant; i=intermediate resistance; s-dd= sensitive based on reported disc diffusion concentration c p=plasma; d gluc=glucose; e insu=insulin conclusions the testing of fungal susceptibility in the presence of plasma with and without the supplements glucose and/or insulin to simulate type ii diabetes appears warranted in light of their sometimes adverse effect on the mic of the antifungal agents. these results taken together indicate that alterations in the levels of serum components associated with type ii diabetes, and simulated herein, cause an alteration in the in https://dx.doi.org/10.5599/admet.1164 plotkin and konaklieva admet & dmpk 10(2) (2022) 153-162 160 vitro activity of certain antifungal agents that is also dependent on the presence of plasma. since treatment failure in candidal infections has been an issue and is increasing, it is likely that a portion of treatment efficacy failure is due to phenotypic resistance [28,29]. while multiple in vitro studies show that glucose induces phenotypic resistance, the organism is always exposed to glucose in the presence of insulin in vivo [27,30,31]. most commonly, these studies focused on the effect glucose has on end-point determinations for amphotericin b, fluconazole and itraconazole [32]. this study demonstrated that testing of individual serum components, e.g., insulin and glucose in vitro showed that not only does insulin or glucose affect susceptibility in a manner that was species and drug-specific, but that the interaction of the two components can negate the effect, either positive or negative, of the other additive. this impact of insulin and glucose is further compounded by testing in the presence of human plasma vs. rpmi medium alone. these unpredictable phenotypic changes in susceptibility to antifungal agents were demonstrated for amphotericin b and c. glabrata, which were sensitive to the drug in medium alone or supplemented with either glucose or insulin. however, when combined, the mic was > 40-fold higher than medium alone or with either glucose or insulin. this effect was eliminated with the addition of anti-candidal antibodynegative plasma. in summary, these findings present evidence that a precision approach is needed for the determination of fungal drug susceptibility, particularly for individuals with type ii diabetes. acknowledgements: the authors thank the midwestern university office of research and sponsored programs and the midwestern college of graduate studies for their support of this work. conflict of interest: the authors declare that no conflict of interest exists. references [1] a.m. willis, w.a. coulter, c.r. fulton, j.r. hayes, p.m. bell, p.-j. lamey. oral candidal carriage and infection in insulin-treated diabetic patients. diabetic medicine 16 (1999) 675-679. https://doi.org/10.1046/j.1464-5491.1999.00134.x. [2] h.m. calvet, t.t. yoshikawa. infections in diabetes. infectious disease clinics of north america 15 (2001) 407-421. https://doi.org/10.1016/s0891-5520(05)70153-7. [3] m.a. de resende, l.v.n.f. de sousa, r.c.b.w. de oliveira, c.y. koga-ito, j.p. lyon. prevalence and antifungal susceptibility of yeasts obtained from the oral cavity of elderly individuals. mycopathologia 162 (2006) 39-44. . https://doi.org/10.1007/s11046-006-0029-6 [4] r.h.p. goncalves, e.t. miranda, j.e. zaia, m. giannini. species diversity of yeast in oral colonization of insulin-treated diabetes mellitus patients. mycopathologia 162 (2006) 83-89. https://doi.org/10.1007/s11046-006-0038-5. [5] m.f. gonzalez-lara, l. ostrosky-zeichner. invasive candidiasis. seminars in respiratory and critical care medicine 41 (2020) 3-12. https://doi.org/10.1055/s-0040-1701215. [6] o. leroy, j.p. gangneux, p. montravers, j.p. mira, f. gouin, j.p. sollet, j. carlet, j. reynes, m. rosenheim, b. regnier, o. lortholary, g. amarcand study group. epidemiology, management, and risk factors for death of invasive candida infections in critical care: a multicenter, prospective, observational study in france (2005-2006). critical care medicine 37 (2009) 1612-1618. https://doi.org/10.1097/ccm.0b013e31819efac0. [7] t.s. ng, m.n.m. desa, d. sandai, p.p. chong, l.t.l. than. growth, biofilm formation, antifungal susceptibility and oxidative stress resistance of candida glabrata are affected by different glucose concentrations. infection, genetics and evolution 40 (2016) 331-338. https://doi.org/10.1016/j.meegid.2015.09.004. [8] m.h. nguyen, c.j. clancy, v.l. yu, y.c. yu, a.j. morris, d.r. snydman, d.a. sutton, m.g. rinaldi. do in vitro susceptibility data predict the microbiologic response to amphotericin b? results of a https://doi.org/10.1046/j.1464-5491.1999.00134.x https://doi.org/10.1046/j.1464-5491.1999.00134.x https://doi.org/10.1016/s0891-5520(05)70153-7 https://doi.org/10.1007/s11046-006-0029-6 https://doi.org/10.1007/s11046-006-0038-5 https://doi.org/10.1007/s11046-006-0038-5 https://doi.org/10.1055/s-0040-1701215 https://doi.org/10.1097/ccm.0b013e31819efac0 https://doi.org/10.1016/j.meegid.2015.09.004 https://doi.org/10.1016/j.meegid.2015.09.004 admet & dmpk 10(2) (2022) 153-162 host factors on susceptibility to antifungal agents doi: https://dx.doi.org/10.5599/admet.1164 161 prospective study of patients with candida fungemia. journal of infectious diseases 177 (1998) 425430. https://doi.org/10.1086/514193. [9] a. rodaki, i.m. bohovych, b. enjalbert, t. young, f.c. odds, n.a. gow, a.j. brown. glucose promotes stress resistance in the fungal pathogen candida albicans. molecular biology of the cell 20 (2009) 4845-4855. https://doi.org/10.1091/mbc.e09-01-0002. [10] a. becker, r. roth. insulysin and pitrilysin: insulin-degrading enzymes of mammals and bacteria. methods enzymol. 248 (1995) 693-703. https://doi.org/10.1016/0076-6879(95)48046-3. [11] s. christensen. insulin rescues the unicellular eucaryote tetrahymena from dying in a complete, synthetic nutrient medium. cell biol. internat. 17 (1993) 833-837. https://doi.org/10.1006/cbir.1993.1145. [12] s.e. fawell, j. lenard. a specific insulin receptor and tyrosine kinase activity in the membranes of neurospora crassa. biochemical and biophysical research communications 155 (1988) 59-65. https://doi.org/10.1016/s0006-291x(88)81049-0. [13] d. leroith, m. lesniak, j. roth. insulin in insects and annelids. diabetes 30 (1981) 70-76. https://doi.org/10.2337/diab.30.1.70. [14] d. leroith, j. shiloach, r. heffron, c. rubinovitz, r. tanenbaum, j. roth. insulin-related material in microbes: similarities and differences from mammalian insulins. can. j. biochem. cell biol. 63 (1985) 839-849. https://doi.org/10.1139/o85-106. [15] s. nolting, h. hagemeier, k. fegeler. effect of insulin on germ tube and mycelial formation of candida albicans. mykosen 25 (1982). [16] b.j. plotkin, s.m. viselli. effect of insulin on microbial growth. curr. microbiol. 41 (2000) 60-64. [17] n. dabrowa, s. taxer, d.h. howard. germination of candida albicans induced by proline. infection and immunity 13 (1976) 830-835. https://doi.org/10.1007/s002840010092. [18] e. evans, f. odds, m. richardson, k. holland. optimum conditions for initiation of filamentation in candida albicans. canadian journal of microbiology 21 (1975) 338-342. https://doi.org/10.1139/m75-048. [19] a. man, c.n. ciurea, d. pasaroiu, a.-i. savin, f. toma, f. sular, l. santacroce, a. mare. new perspectives on the nutritional factors influencing growth rate of candida albicans in diabetics. an in vitro study. memórias do instituto oswaldo cruz 112 (2017) 587-592. https://doi.org/10.1590/007402760170098. [20] n.c.f.c.l. standards., reference method for broth dilution antifungal susceptibility testing of yeasts; tentative standard, national committee for clinical laboratory standards, wayne, pa, 1995. [21] j.l. rodriguez-tudela, j. berenguer, j.v. martinez-suarez, r. sanchez. comparison of a spectrophotometric microdilution method with rpmi-2% glucose with the national committee for clinical laboratory standards reference macrodilution method m27-p for in vitro susceptibility testing of amphotericin b, flucytosine, and fluconazole against candida albicans. antimicrobial agents and chemotherapy 40 (1996) 1998-2003. https://doi.org/10.1128/aac.40.9.1998. [22] o.t. malazy, m. shariat, r. heshmat, f. majlesi, m. alimohammadian, n.k. tabari, b. larijani. vulvovaginal candidiasis and its related factors in diabetic women. taiwanese journal of obstetrics and gynecology 46 (2007) 399-404. https://doi.org/10.1016/s1028-4559(08)60010-8. [23] c. amar, j. ashish, v. hajare, b. yogesh. study of prevalence and antifungal susceptibility of candida. international journal of pharma and bio sciences 4 (2013). https://doi.org/10.1186/s12879-0194736-2. [24] k.-h. lin, c.-m. chen, t.-l. chen, s.-c. kuo, c.-c. kao, y.-c. jeng, m.-w. ho. diabetes mellitus is associated with acquisition and increased mortality in hiv-uninfected patients with cryptococcosis: a population-based study. journal of infection 72 (2016) 608-614. https://doi.org/10.1016/j.jinf.2016.01.016. https://dx.doi.org/10.5599/admet.1164 https://doi.org/10.1086/514193 https://doi.org/10.1091/mbc.e09-01-0002 https://doi.org/10.1016/0076-6879(95)48046-3 https://doi.org/10.1006/cbir.1993.1145 https://doi.org/10.1006/cbir.1993.1145 https://doi.org/10.1016/s0006-291x(88)81049-0 https://doi.org/10.2337/diab.30.1.70 https://doi.org/10.1139/o85-106 https://doi.org/10.1007/s002840010092 https://doi.org/10.1139/m75-048 https://doi.org/10.1139/m75-048 https://doi.org/10.1590/0074-02760170098 https://doi.org/10.1590/0074-02760170098 https://doi.org/10.1128/aac.40.9.1998 https://doi.org/10.1016/s1028-4559(08)60010-8 https://doi.org/10.1186/s12879-019-4736-2 https://doi.org/10.1186/s12879-019-4736-2 https://doi.org/10.1016/j.jinf.2016.01.016 https://doi.org/10.1016/j.jinf.2016.01.016 plotkin and konaklieva admet & dmpk 10(2) (2022) 153-162 162 [25] r. acharya, k. khanal, p. upadhyaya, s. kafle, v. savaliya. diabetes mellitus as a risk factor for cryptococcal meningitis in immunocompetent. idcases 22 (2020) e00988. https://doi.org/10.1016/j.idcr.2020.e00988. [26] r. diamond, j. bennett. prognostic factors in cryptococcal meningitis: a study in 111 cases. annals of internal medicine 80 (1974) 176-181. https://doi.org/10.7326/0003-4819-80-2-176. [27] i.v. ene, a.k. adya, s. wehmeier, a.c. brand, d.m. maccallum, n.a. gow, a.j. brown. host carbon sources modulate cell wall architecture, drug resistance and virulence in a fungal pathogen. cellular microbiology 14 (2012) 1319-1335. https://doi.org/10.1111/j.1462-5822.2012.01813.x. [28] j.d. sobel. treatment of vaginal candida infections. expert opinion on pharmacotherapy 3 (2002) 1059-1065. https://doi.org/10.1517/14656566.3.8.1059. [29] n.b. mcclenny, h. fei, e.j. baron, a.c. gales, a. houston, r.j. hollis, m.a. pfaller. change in colony morphology of candida lusitaniae in association with development of amphotericin b resistance. antimicrobial agents and chemotherapy 46 (2002) 1325-1328. https://doi.org/10.1128/aac.46.5.1325-1328.2002. [30] s.m. mandal, d. mahata, l. migliolo, a. parekh, p.s. addy, m. mandal, a. basak. glucose directly promotes antifungal resistance in the fungal pathogen, candida spp. journal of biological chemistry 289 (2014) 25468-25473. https://doi.org/10.1074/jbc.c114.571778. [31] l. knight, j. fletcher. growth of candida albicans in saliva: stimulation by glucose associated with antibiotics, corticosteroids, and diabetes mellitus. the journal of infectious diseases 123 (1971) 371377. https://doi.org/10.1093/infdis/123.4.371. [32] m.h. nguyen, y.y. christine. influence of incubation time, inoculum size, and glucose concentrations on spectrophotometric endpoint determinations for amphotericin b, fluconazole, and itraconazole. journal of clinical microbiology 37 (1999) 141-145. https://doi.org/10.1128/jcm.37.1.141-145.1999. ©2022 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) https://doi.org/10.1016/j.idcr.2020.e00988 https://doi.org/10.1016/j.idcr.2020.e00988 https://doi.org/10.7326/0003-4819-80-2-176 https://doi.org/10.1111/j.1462-5822.2012.01813.x https://doi.org/10.1517/14656566.3.8.1059 https://doi.org/10.1128/aac.46.5.1325-1328.2002 https://doi.org/10.1128/aac.46.5.1325-1328.2002 https://doi.org/10.1074/jbc.c114.571778 https://doi.org/10.1093/infdis/123.4.371 https://doi.org/10.1128/jcm.37.1.141-145.1999 http://creativecommons.org/licenses/by/3.0/ a simple and fast flow injection amperometry for the determination of methimazole in pharmaceutical preparations using an unmodified boron-doped diamond electrode doi: https://doi.org/10.5599/admet.1584 303 admet & dmpk 11(2) (2023) 303-315; doi: https://doi.org/10.5599/admet.1584 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper a simple and fast flow injection amperometry for the determination of methimazole in pharmaceutical preparations using an unmodified boron-doped diamond electrode adison meoipun1, kantima kaewjua1, orawon chailapakul2 and weena siangproh1* 1department of chemistry, faculty of science, srinakharinwirot university, sukhumvit 23 rd. wattana, bangkok, 10110, thailand 2electrochemistry and optical spectroscopy center of excellence, department of chemistry, faculty of science, chulalongkorn university, pathumwan, bangkok 10330, thailand *corresponding author: weenasi@hotmail.com, weena@g.swu.ac.th received: november 01, 2022; revised: december 29, 2022; published: january 01, 2023 abstract in this work, an automated flow injection analysis (fia) connected to a boron-doped diamond electrode (bdde) was originally developed for the analysis of methimazole in pharmaceutical preparations. at a modification-free bdde, methimazole was easilly oxidized. for the analysis of the mechanisms occurring at the electrode surface, cyclic voltammetry was employed to evaluate the impact of fundamental experimental parameters, such as ph and scan rate, on the bdde response. for the quantitative detection, the fia amperometric approach was constructed and used as a fast and sensitive method. the suggested approach provided a broad linear range of 0.5–50 µmol/l and a low detection limit of 10 nmol/l (signal-tonoise ratio = 3). furthermore, the bdde was successfully utilized to quantify methimazole in genuine samples from a variety of medicines, and its performance remained steady after more than 50 tests. the findings of amperometric measurements exhibit excellent repeatability, with relative standard deviations of less than 3.9 and 4.7 % for intra-day and inter-day, respectively. the findings indicated that, compared with traditional approaches, the suggested method has the following advantages: quick analysis time, simplicity, highly sensitive output, and no need for complicated operational processes. ©2023 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords methimazole; boron-doped diamond thin film electrode; flow injection analysis; drug formulations; amperometry introduction methimazole (2-mercapto-1-methyl imidazole, mmi) is used to treat hyperthyroidism or an overactive thyroid. it is also prescribed before undergoing thyroid surgery or radioactive iodine treatment. the most prevalent cause of hyperthyroidism is graves’ disease, an autoimmune condition caused by antibodies that bind to receptors on thyroid hormone-producing cells in the thyroid gland, inducing excessive thyroid hormone synthesis. by mixing iodine with a protein called thyroglobulin, an enzyme called peroxidase generates thyroid hormones, such as thyroxine (t4) and triiodothyronine (t3). mmi stops iodine and https://doi.org/10.5599/admet.1584 https://doi.org/10.5599/admet.1584 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:weenasi@hotmail.com mailto:weena@g.swu.ac.th http://creativecommons.org/licenses/by/4.0/ weena siangproh et al. admet & dmpk 11(2) (2023) 303-315 304 peroxidase from producing t4 and t3 from their usual interactions with thyroglobulin. therefore, thyroid hormone synthesis is reduced as a result of this activity. the conversion of t4 to t3 is also hampered by mmi. because t3 is more effective than t4, it decreases thyroid hormone activity [1,2]. mmi was authorized by the fda in march 1999 [3]. however, mmi can induce nephritis, liver cirrhosis, skin irritation, allergies, pharyngitis with fever, and other adverse effects [4]. in pharmaceutical, nutrition, and clinical chemistry, accurate drug determination using simple and rapid procedures is a prerequisite. the analytical methods used for mmi determination are described, including thin-layer chromatography [5], high-performance liquid chromatography with tandem mass spectrometry [6], ultraviolet detection [7,8], spectroscopy [9-11], potentiometry [12], liquid chromatography with amperometric detection [13], and capillary zone electrophoresis [14]. among these, liquid chromatography is one of the most commonly employed in pharmaceutical research. while it is often utilized in laboratories, there are instances when complicated procedures must be developed and validated. they could also take a lot of time and call for qualified employees. on the other hand, because they are quick, simple, accurate, and exceedingly sensitive, electroanalytical methods have proved effective in detecting a wide range of analytes, including those present in biological and pharmaceutical samples. however, electrochemistry has a limitation in that it is not an automated method causing it not to support routine work. the discovery of a flow-based system led to widespread application in quantitative pharmaceutical analysis. flow-based technology has been useful and can be an option for routine analysis with a fast response and a substantially lower cost. moreover, flow injection analysis (fia) offers the possibility to work in continuous rather than in batch mode, with higher sample throughput and various detectors. thus, electrochemical techniques have been widely used as a detector in flow injection techniques for pharmaceutical applications [15,16]. one of the most frequent working electrodes used in electrochemical research is the glassy carbon electrode (gce). unfortunately, mmi oxidation at the gce generally has a weak electrochemical response. therefore, chemically modified electrodes have been frequently documented for the sensitive and selective determination of mmi. acetylene black/chitosan film [17], mwcnts [18], mwcnts/electro-copolymerized cobalt nanoparticle–poly (pivalic acid) composite film [19], and carbon paste electrodes modified with schiff base complexes of vanadium and cobalt [20,21], nanocomposite of cds np–rgo/il [22], and mwcnt–titanium dioxide nanoparticles [23] are among the modifiers. in several applications mentioned, time was wasted in the modification and pretreatment steps. furthermore, there is still a significant issue with electrode preparation repeatability. based on our expertise, the chemical and physical properties of a boron-doped diamond electrode (bdde) (low background currents, wide potential window, good resistance to fouling, chemical and mechanical stability, lack of a surface oxide film, and controllable surface termination) have been successfully used in a variety of applications [24-26]. in this regard, bdde can be realized for the quantification analysis of mmi [27]. however, the difference of this study from the previous work related to the bdde is that the modification-free or unpretreated bdde was used to explore the electrochemical property of mmi via cyclic voltammetry as a continuation of our previous work on the usage of the bdde in applications. as previously mentioned, thus, this is the first time that reported the use of an unmodified bdde as a detector in an fia system to determine mmi. the online fia and the promising characteristics of the bdde enable the detection of extremely low mmi levels. good analytical detection performances were observed with detection limits in the nanomolar range, strong sensitivity, outstanding response accuracy, and stability. therefore, the ideal protocol to work with bdde for fia enables the electrochemical pretreatment step and sequential detection to simplify regular applications and routine measurements. admet & dmpk 11(2) (2023) 303-315 fia determination of methimazole doi: https://doi.org/10.5599/admet.1584 305 experimental chemicals and reagents without additional purification, all compounds were of analytical grade or above. all solutions and subsequent dilutions were prepared in deionized water. phosphate solutions (0.1 mol/l, ph 5–8) were made by combining 0.1 mol/l potassium dihydrogen phosphate (merck, darmstadt, germany) and 0.1 mol/l sodium hydrogen phosphate (fluka, germany). phosphate solution (ph 2.5) was prepared from 0.1 mol/l potassium dihydrogen phosphate, and the ph was adjusted with orthophosphoric acid (85 %, carlo erba). phosphate solution (ph 9), 0.1 mol/l, was prepared from 0.1 mol/l potassium dihydrogen phosphate, and the ph was adjusted with 0.1 mol/l sodium hydroxide. prior to use, the standard methimazole (sigma-aldrich, darmstadt, germany) solutions were prepared daily using the same solution. electrodes the equipment utilized for the diamond film generation and the experimental setup have been extensively documented elsewhere [28]. the films were created by microwave plasma-assisted chemical vapor deposition of bdd thin films on highly conductive n-si (111) substrates. a layer thickness of around 30 mm was attained after 10 hours of deposition. the normal degree of boron doping in the film was around 1021 cm3, and the nominal b/c atomic ratio in the gas phase was 1:100. before usage, the bdde was cleansed with ultrapure water. for the glassy carbon electrode (gce, (0.07 cm2, bioanalytical system, inc.), the pretreatment involved sequential polishing with 1 and 0.05 µm of alumina/water slurries on felt pads, followed by rinsing with ultrapure water prior to use. voltammetric study an autolab potentiostat 100 (pg100, metrohm, switzerland) with a typical three-electrode setup was used to record the electrochemical signals. the planar working bdde was pressed against a smooth ground joint at the bottom of the cell, isolated by an o-ring. the geometric area of the working electrode was 0.07 cm2. a platinum wire served as the auxiliary electrode, and ag/agcl was used as the reference. the placement of the back side of the si substrate on a brass plate led to an ohmic contact. in the comparative study with the bdde, a gce (0.07 cm2, bioanalytical system, inc.) was also utilized as a working electrode. cyclic voltammetry was employed to probe the electrochemical reaction. the electrochemical measurements were performed in a faradaic cage to reduce electronic noise. all experiments were conducted at room temperature. flow injection analysis with amperometric detection the fia system included an electrochemical detector (pg100), a peristaltic pump (ismatec), a thin-layer flow cell (all from bioanalytical system, inc.), and an injection port with a 20-l injection loop (rheodyne7725). a reagent delivery module regulated the carrier stream, which was 0.1 mol/l phosphate solution (ph 9) at a flow rate of 1 ml/min. to minimize the pulsation caused by the peristaltic pump’s roller alternation, a pulse dampener was used in series. a silicone gasket served as a spacer, and the thin-layer flow cell had an ag/agcl reference electrode, a stainless-steel tube as an auxiliary electrode, and an outlet. to decrease electrical noise, the experiments were conducted in a copper faradaic cage. before performing the amperometric measurement, a hydrodynamic voltammogram was acquired to obtain the suitable detection potential. after each injection, the peak current and the matching background current were recorded. hydrodynamic voltammograms were created by plotting this data as a function of applied voltage. the hydrodynamic voltammograms provided a maximum signal-to-background (s/b) ratio since the amperometric measurements were performed at the potential of +0.3 v. https://doi.org/10.5599/admet.1584 weena siangproh et al. admet & dmpk 11(2) (2023) 303-315 306 sample preparations tablets containing mmi were acquired from local pharmacies in bangkok, thailand. the medication tablet was homogenized in an agar mortar prior to measuring. the accurate amount of the powdered mass analyte was dissolved in 100 ml of 0.1 mol/l phosphate solution (ph 9) by stirring for 10 min. the macromolecule excipient was separated by filtration with whatman filter paper (no. 1), and the residue was washed three times with phosphate solution (ph 9). to determine the mmi amount of the tablets, the final test solution was diluted with phosphate solution (ph 9) to keep the mmi concentration within the linear dynamic range (0.5–50 mol/l). results and discussion voltammetric study the electrochemical responses of two electrodes (the bdde and gce) to mmi were first examined and compared via cyclic voltammetry. figure 1a presents the voltammetric response of the bdde in a 0.1 mol/l phosphate solution (ph 9) containing 1 mmol/l mmi, whereas figure 1b shows the voltammetric response of the gce in the same circumstances. the response at each electrode exhibited no detectable oxidative signals across the investigated voltage range, according to the cyclic voltammograms obtained before the addition of mmi (0.0–1.2 v). furthermore, the bdde consistently exhibited lower capacitance than the gce when background currents at each electrode were evaluated. the response at the bdde changed when mmi was added to the solution, with a new well-defined electrochemical signal developing at +0.45 v. the mmi response at the gce, on the other hand, had two pairs of ill-defined peaks observed as an increase in oxidation current at +0.35 v and +0.98 v. for comparison, the cv response of mmi at the bdde is more significant and has a stronger anodic peak current than bare gce. this may be concerned with the fact that a bdde provided a very lower background current than a gce, suggesting that the bdde has shown an efficient catalytic activity toward the electrooxidation of mmi. in addition, no additional reductive processes were identified in the voltage range investigated at each electrode when the scan direction was reversed. these irreversible behaviors provide consistent results with a chemically modified electrode reported previously. effect of ph value on the electrochemical behavior of mmi investigating the effect of solution ph on the electrochemical responses of mmi is essential for achieving the best results in high responsiveness. thus, the electrochemical characteristics of mmi were investigated using a phosphate solution with a concentration of 0.1 mol/l as the supporting electrolyte. the tests were conducted at ph levels of 2.5, 5.0, 7.0, 8.0, and 9.0. according to the findings, mmi may be oxidized in both neutral and alkaline conditions. when the ph value increased, the peak potential shifts toward less positive potentials (figure 2). additionally, the oxidation of sulfur-containing compounds at the bdde can explain this electrooxidation. the oxidation process involved the dissociation of the proton on the thiol group, followed by the electrochemical oxidation of the mmi anion. as a result, mmi easily lost the proton at higher ph levels and formed a more stable reduced form. the use of a phosphate solution at ph 9 yielded the largest oxidation current peak. hence, this ph was selected as the starting point for all the following studies. scan rate and concentration dependence to confirm that this oxidative process is diffusion-controlled, an investigation into the effects of scan rate on the electrochemical signal was carried out. as presented in figure 3, the electrooxidation responses of mmi varied with the scan rate, with the current plateau rising linearly with the square root of the scan rate (r2 = 0.9988). such a relationship indicates that mmi oxidation at a bdde surface is really controlled by the diffusion process. admet & dmpk 11(2) (2023) 303-315 fia determination of methimazole doi: https://doi.org/10.5599/admet.1584 307 figure 1. cyclic voltammograms for a) bdde and b) gce versus ag/agcl in 1.0 mmol/l mmi in 0.1 mol/l phosphate solution ph 9 (solid lines) and 0.1 mol/l phosphate solution ph 9 (dashed lines). sweep rate, 0.05 v/s; area of electrode, 0.07 cm2 figure 2. cyclic voltammogram for 1-mmol/l mmi in 0.1 mol/l phosphate solution at different ph solutions at a bdde (potential scan rates: 0.05 v/s and area of electrode: 0.07 cm2). the calibration curve of the relationship between potential (v) and ph is also presented in the inset of this figure. https://doi.org/10.5599/admet.1584 weena siangproh et al. admet & dmpk 11(2) (2023) 303-315 308 figure 3. bar graph for 1 mmol/l mmi in 0.1 mol/l phosphate solution (ph 9) at bdde for a series of potential scan rate; area of electrode, 0.07 cm2. the calibration curve of the relationship between current (a) and (scan rate)1/2 is also presented in the inset of this figure. the measurement obtained based on three replicates (n=3) in order to perform a preliminary study on the analytical effectiveness of mmi at the bdde, cyclic voltammetry was used to measure the change in voltammetric current as a function of various mmi concentrations. it was discovered that the response of the bdde was linear over the concentration range of 0.025-3.0 mmol/l, with a limit of detection of 50 µmol/l at signal-to-background (s/b) = 3. the data is depicted in figure 4. however, the obtained linear range and lod are insufficient for the quantitative detection of mmi in applications such as biological samples or pharmaceutical preparations. therefore, the use of a highly sensitive electrochemical technique is required. electrochemical behavior of mmi at the bdd electrode to examine the mechanism of electrooxidation of mmi at the bdde interface, the electrooxidation behavior of mmi was measured via cyclic voltammetry at various potential scan rates. the oxidation current linearly increases with the square root of the scan rate, as presented in figure 2. in addition, the obtained linear relationship between ipa and the ν1/2 (y = 75.513x −1.0473; r2 = 0.9988) was used to calculate the diffusion coefficient of mmi (d0) using andrieux and savèant’s theoretical model (eq. 1): 1/2 1/2 1/2 pa 00.496 f i facd rt    =     (1) where f is faraday’s constant = 96,485 c/mol, r = 8.31447 j/mol⋅k, t = 298 k, a = 0.07 cm2; cs is the mmi concentration = 1 × 10−6 mol/cm3, ip is the oxidative current of mmi, and ν is the scan rate. the calculated diffusion coefficient (d0) of mmi was found to be 1.30 × 10−5 cm2/s. the randles–ševčík relationship for the irreversible process (eq. 2) was employed to calculate the number of electron transfers at the electrode interface: ( )5 1/2 1/2 1/2pa s 0[( ) ]2.99 10 1 i n n ac d = − (2) 0 s pa ln ln (1 ) (1 ) (1 ) rtkrt rt e e nf nf nf     = + + − − − (3) admet & dmpk 11(2) (2023) 303-315 fia determination of methimazole doi: https://doi.org/10.5599/admet.1584 309 figure 4. cyclic voltammogram for mmi in 0.1 mol/l phosphate solution (ph 9) at bdde for a series of methimazole concentrations. the potential scan rate was 0.05 v/s; the area of electrode, was 0.07cm2. the calibration curve is presented in the inset. the measurement obtained based on three replicates (n=3). the factor (1-)n was firstly calculated from the linear regression (eq. 3) between epa and ln ν (y = 0.0223x + + 0.7576; r2 = 0.9963) and was found to be 1.15. after calculating the important factors, the number of electrons involved in the electrooxidation of mmi was calculated with eq. 2 and was found to be n = 0.92 (n1). moreover, it was found that the peak potential of mmi shifted to the negative potential by increasing ph. this is predictable due to the participation of proton in the oxidation process of mmi [18,19]. the nernst equation (eq. 4) was used to describe the relationship between epa and ph (inset in figure 2) the linear relationship between the ph values and oxidation peak potentials was plotted (y = −0.0572x + 1.0074; r2 = 0.9652). a 0 pa b 0.0591 ox log 0.0591 ph r m e e n n      = + −          (4) where m and n are the number of protons and electrons at the electrode interface, respectively. the slope was found to be 57.2 mv, which is in agreement with the theoretical slope (2.303 mrt/2f) of 59 mv/ph. based on the calculation, m = n = 1, since the results show that the quantities of electrons and protons exchanged at the electrode surface are equal. consequently, the possible mechanism of mmi at the bdde involves one electron and one proton. hydrodynamic voltammetry hydrodynamic voltammetry is a helpful technique to figure out what potential to use for fia/amperemetric detection. a standard solution of mmi was repeatedly injected, while the fia/amperemetry working potential was adjusted at +0.20, +0.30, +0.35, +0.40 and +0.45 v, respectively. figure 5a shows the hydrodynamic voltammogram of the standard solution, which contains 100 mol/l of mmi. the current response of mmi increased as the applied voltage increased; however, it did not reach its maximum value. to reach the maximum potential point, the s/b ratio was computed (figure 5a) at each point to construct a graph as a function of s/b ratios and applied potential (figure 5b). in this work, +0.30 v was selected as the optimum detection potential for subsequence work owing to high sensitivity and low noise. remarkably, the reference electrode used for cyclic voltammetry and flow system is different. thus, it may lead to a little different detection potential being obtained between batch and flow systems. https://doi.org/10.5599/admet.1584 weena siangproh et al. admet & dmpk 11(2) (2023) 303-315 310 figure 5. (a) hydrodynamic voltammogram of (-●-) 0.1mol/l phosphate solution (ph 9, background current) and (-▲-) 100 mol/l of mmi in 0.1 mol/l phosphate solution (ph 9) with repeated four injections of analysis using 0.1 mol/l phosphate solution (ph 9) as a carrier solution. (b) the hydrodynamic of the signalto-background (s/b) ratio. the flow rate was 1 ml/min. fia coupled with amperometric detection after setting the suitable detection potential for the fia technique, amperometric measurements were carried out in the phosphate solution (ph 9) containing various mmi concentrations in order to get the analytical curve. the fia current-time response for different mmi concentrations (figure 6) demonstrated a linear relationship between the current values (at +0.30 v) obtained and the mmi concentrations exhibited from 0.5 to 50 µmol/l with a correlation coefficient of 0.9969. figure 6. the flow injection analysis results of increasing concentrations of methimazole at a bdde in 0.1 mmol/l pbs at ph 9. methimazole concentrations: 0.5, 5, 15, 40, and 50 μmol/l. inset: a plot of anodic peak currents versus increasing concentration of methimazole. the linearity is disrupted at concentrations greater than 50 mol/l. the experiment revealed that the detection limit was found to be 10 nmol/l (signal-to-noise (s/n) ratio = 3). as demonstrated in table 1, the proposed method gave the lowest detection limit compared to previously published studies, particularly those admet & dmpk 11(2) (2023) 303-315 fia determination of methimazole doi: https://doi.org/10.5599/admet.1584 311 using chemically modified electrodes. this means that a bdde may detect mmi with extreme sensitivity while maintaining its speed and ease of use. furthermore, because it allows for rapid real-time measurements and eliminates the need for electrode cleaning, this system is ideal for pharmaceutical analysis. table 1. comparison of analytical performance for the determination of mmi using different analytical methods techniques and detectors linearity range, mol/l detection limit, mol/l ref. flow-injection: chemiluminescence 1.7510-5 – 8.7510-4 8.7510-6 [29] capillary electrophoresis: a carbon fiber micro disk electrode 10-7 – 2.010-4 5.0010-8 [30] amperometry: screen-printed enzymatic biosensor modified with carbon nanotube 7.410-8 – 6.3510-5 5.6010-8 [31] differential pulse voltammetry: carbon-paste electrode modified with a chiff base complex of cobalt 10-6 – 1.010-4 5.0010-7 [21] hplc: spectrophotometer detector 5.010-8 – 5.510-7 7.4010-5 [32] differential pulse voltammetry: acetylene black/chitosan/gce 10-7 – 2.010-5 2.0010-8 [17] square wave voltammetry: fluorine-doped tin oxide electrodes 6.010-6 – 24010-6 1.9810-6 [33] square wave voltammetry: gce 710-6 – 1.310-5 3.7010-6 [34] square wave voltammetry: pretreated anodically bdde 4.38×10−6– 2.19×10−4 7.4510−7 [27] amperometry: flow-injection analysis/unmodified bdde 0.510-6 – 5010-6 10−8 this work determination of mmi in commercially available pharmaceuticals the use of the fia amperometric approach for the assessment of mmi in commercially available drug samples was performed to test the analytical validity of the suggested method. however, the active ingredients of mmi tablets contain lactose monohydrate, magnesium stearate, starch, corn, and talc [35]. these substances are non-electroactive species that cannot be oxidized at the same oxidation potential as mmi at a bdde. to ensure that the effects of matrices that may be present in real samples can be reduced, the mmi content was calculated using a standard addition method. the recovery of this method was determined by calculating the percentage of recovery after adding known amounts of standard mmi to the sample solutions. table 2 presents the recovery for intra-day and inter-day from the determination of mmi using the proposed method. on the basis of intra-assay, the method's precision was established. three concentrations of the added solution (0.33, 0.65, and 1.31 g/ml) were selected for the investigation to check the results obtained from the low, medium, and high concentrations with respect to the probable range of interest in the samples. the results obtained from 10 injections were within 3.9 % of the relative standard deviation (rsd, %). the rsd values for the day-to-day assays of mmi were also investigated. it was discovered that during a week in the same laboratory, the %rsd measurements did not vary by more than 4.7 %. these findings indicate that bdde detection was very reproducible and accurate. in addition, the bdde were tested for their long-term stability in continuous operation (50 injections). they generated highly repeatable responses from day to day, and the responses were also repeatable after several days in the laboratory environment. the intra-day recovery from the proposed technique was between 95.2 and 101.7 %, whereas the inter-day recovery of mmi was between 93.5 and 103.8 %. as presented in table 3, for all the five supplements tested, the experimentally determined values were quite close to the manufacturer’s claims. these results were compared with those obtained using the hplcuv technique, and statistical analysis was conducted using a paired t-test with a 95 % confidence interval. the null hypothesis was supported by the paired two-tailed test, which yielded a computed t-value of https://doi.org/10.5599/admet.1584 weena siangproh et al. admet & dmpk 11(2) (2023) 303-315 312 0.906, less than the crucial t-value (2.776). a summary of the findings is presented in table 3, the results of the fia combined with bdde amperometry are comparable and not significantly different from those obtained using the traditional hplc-uv method. table 2. the intraand inter-precisions and recoveries of the fia-bdde method (n = 10) samples spiked level, g/ml intra-day inter-day mean of recovery, % (± sd) rsd, % mean of recovery, % (± sd) rsd, % sample 1 sample 2 sample 3 sample 4 sample 5 0.33 0.65 1.31 0.33 0.65 1.31 0.33 0.65 1.31 0.33 0.65 1.31 0.33 0.65 1.31 95.3 ± 1.5 96.5 ± 2.6 101.4 ± 1.1 98.4 ± 1.2 99.1 ± 1.6 101.2 ± 1.2 96.4 ± 2.2 101.6 ± 1.9 98.8 ± 1.4 95.4 ± 3.3 101.2 ± 2.1 99.3 ± 1.3 95.2 ± 2.5 100.8 ± 2.0 101.7 ± 1.4 3.9 3.4 2.6 3.8 2.6 1.5 3.8 2.6 1.6 3.5 2.1 2.5 3.3 2.9 1.2 93.5 ± 2.1 97.4 ± 2.3 100.8 ± 1.8 99.2 ± 0.7 101.2 ± 2.2 100.7 ± 1.8 93.5 ± 2.4 103.7 ± 2.2 100.1 ± 1.5 94.8 ± 2.7 103.8 ± 1.9 102.3 ± 1.7 93.5 ± 1.8 99.8 ± 1.9 103.8 ± 1.2 4.5 3.6 2.8 4.7 3.1 2.0 4.1 3.1 1.9 3.9 2.9 2.2 4.4 2.7 1.3 table 3. determination of the mmi levels in different drug samples (n = 5) using the traditional hplc-uv method and the developed fia-bdde method reported herein. sample amount drug label, mg/tablet amount found, mg/capsule (± sd) recovery, % fia-bdde a hplc-uv b fia-bdde a hplc-uv b sample 1 sample 2 sample 3 sample 4 sample 5 5.00 5.00 5.00 5.00 5.00 4.93 ± 0.2 4.90 ± 0.8 4.99 ± 1.7 4.96 ± 2.5 5.03 ± 1.4 4.88 ± 0.5 4.93 ± 1.7 5.22 ± 1.5 4.95 ± 2.4 5.05 ± 1.3 98.6 ± 1.3 98.0 ± 1.7 99.8 ± 3.5 99.2 ± 2.4 100.6 ± 2.7 97.6 ± 1.5 98.6 ± 1.3 104.4 ± 2.3 99.0 ± 1.9 101.0 ± 2.3 a = this developed method, b = traditional method conclusions a simple and fast electrochemical method for the determination of mmi was successfully developed using an automatic fia connected to an unmodified bdde. the results indicate that bdde is a promising material for further research in terms of repeatability, sensitivity, and low background current. among other interesting features of the approach, it was described as rapid, simple, precise, and extremely sensitive, making it useful for the mmi analysis and quality control of pharmaceutical products. furthermore, this method may be successfully used to determine mmi in commercially available medications, with results that match the label and standard procedure. this evidence will help understand specific drug pathways and metabolism, leading to their development in the various phases of pharmaceutical research as well as clinical investigations of the mmi levels. acknowledgments: the authors would like to acknowledge the partial financial support from srinakharinwirot university and the national research council of thailand (nrct) through grant number n41a640073. the authors would like to thank enago (www.enago.com) for the english-language review. conflicts of interest: all the authors declare no conflict of interest. references [1] b. marchant, j.f. lees, w.d. alexander. antithyroid drugs. pharmacology & therapeutics, part b: general and systematic pharmacology 3 (1978) 305-348. https://doi.org/10.1016/s0306039x(78)80001-4. http://www.enago.com/ https://doi.org/10.1016/s0306-039x(78)80001-4 https://doi.org/10.1016/s0306-039x(78)80001-4 admet & dmpk 11(2) (2023) 303-315 fia determination of methimazole doi: https://doi.org/10.5599/admet.1584 313 [2] h. nakamura, j.y. noh, k. itoh, s. fukata, a. miyauchi, n. hamada. comparison of methimazole and propylthiouracil in patients with hyperthyroidism caused by graves’ disease. the journal of clinical endocrinology & metabolism 92 (2007) 2157-2162. https://doi.org/10.1210/jc.2006-2135. [3] methimazole, https://www.medicinenet.com/methimazole/article.htm) (accessed august 17, 2022). [4] m. baretić, s. balić, g. gudelj. miastenična kriza kao nuspojava liječenja metimazolom: prikaz slučaja. acta clinica croatica 49 (2010) 67-71. https://hrcak.srce.hr/56826. [5] m. aletrari, p. kanari, d. partassides, e. loizou. study of the british pharmacopeia method on methimazole (thiamazole) content in carbimazole tablets. journal of pharmaceutical and biomedical analysis 16 (1998) 785-792. https://doi.org/10.1016/s0731-7085(97)00119-2. [6] y. ito, h. sakai, h. nagao, r. suzuki, m. odawara, k. minato. development of a high-performance liquid chromatography–tandem mass spectrometric method for the determination of methimazole in human blood matrices. journal of chromatography b 1144 (2020) 122083. https://doi.org/10.1016/j.jchromb.2020.122083. [7] g. moretti, p. betto, p. cammarata, f. fracassi, m. giambenedetti, a. borghese. determination of thyreostatic residues in cattle plasma by high-performance liquid chromatography with ultraviolet detection. journal of chromatography b: biomedical sciences and applications 616 (1993) 291-296. https://doi.org/10.1016/0378-4347(93)80397-m. [8] r. zakrzewski. determination of methimazole in urine with the iodine-azide detection system following its separation by reversed-phase high-performance liquid chromatography. journal of chromatography b 869 (2008) 67-74. https://doi.org/10.1016/j.jchromb.2008.05.021. [9] m. el-bardicy, y. el-saharty, m. tawakkol. determination of carbimazole and methimazole by first and third derivative spectrophotometry. spectroscopy letters 24 (1991) 1079-1095. https://doi.org/10.1080/00387019108018174. [10] c. sanchez-pedreno, m. albero, m. garcia, v. rodenas. flow-injection spectrophotometric determination of carbimazole and methimazole. analytica chimica acta 308 (1995) 457-461. https://doi.org/10.1016/0003-2670(94)00606-m. [11] m.s. garcía, m.i. albero, c. sánchez-pedreño, l. tobal. kinetic determination of carbimazole, methimazole and propylthiouracil in pharmaceuticals, animal feed and animal livers. analyst 120 (1995) 129-133. https://doi.org/10.1039/an9952000129. [12] s. pinzauti, g. papeschi, e. la porta. potentiometric titration of thiols, cationic surfactants and halides using a solid-state silver-silver sulphide electrode. journal of pharmaceutical and biomedical analysis 1 (1983) 47-53. https://doi.org/10.1016/0731-7085(83)80007-7. [13] s. zhang, w.-l. sun, w. zhang, w.-y. qi, l.-t. jin, k. yamamoto, s. tao, j. jin. determination of thiocompounds by liquid chromatography with amperometric detection at a nafion/indium hexacyanoferrate film modified electrode. analytica chimica acta 386 (1999) 21-30. https://doi.org/10.1016/s0003-2670(98)00799-5. [14] a. wang, l. zhang, s. zhang, y. fang. determination of thiols following their separation by cze with amperometric detection at a carbon electrode. journal of pharmaceutical and biomedical analysis 23 (2000) 429-436. https://doi.org/10.1016/s0731-7085(00)00326-5. [15] b. uslu, s.a. ozkan. electroanalytical methods for the determination of pharmaceuticals: a review of recent trends and developments. analytical letters 44 (2011) 2644-2702. https://doi.org/10.1080/00032719.2011.553010. [16] s.a. özkan, b. uslu, h.y. aboul-enein. analysis of pharmaceuticals and biological fluids using modern electroanalytical techniques. critical reviews in analytical chemistry 33 (2003) 155-181. https://doi.org/10.1080/713609162. [17] w. yazhen. electrochemical determination of methimazole based on the acetylene black/chitosan film electrode and its application to rat serum samples. bioelectrochemistry 81 (2011) 86-90. https://doi.org/10.1016/j.bioelechem.2011.04.001. https://doi.org/10.5599/admet.1584 https://doi.org/10.1210/jc.2006-2135 https://www.medicinenet.com/methimazole/article.htm https://hrcak.srce.hr/56826 https://doi.org/10.1016/s0731-7085(97)00119-2 https://doi.org/10.1016/j.jchromb.2020.122083 https://doi.org/10.1016/0378-4347(93)80397-m https://doi.org/10.1016/j.jchromb.2008.05.021 https://doi.org/10.1080/00387019108018174 https://doi.org/10.1080/00387019108018174 https://doi.org/10.1016/0003-2670(94)00606-m https://doi.org/10.1039/an9952000129 https://doi.org/10.1016/0731-7085(83)80007-7 https://doi.org/10.1016/s0003-2670(98)00799-5 https://doi.org/10.1016/s0731-7085(00)00326-5 https://doi.org/10.1080/00032719.2011.553010 https://doi.org/10.1080/00032719.2011.553010 https://doi.org/10.1080/713609162 https://doi.org/10.1016/j.bioelechem.2011.04.001 weena siangproh et al. admet & dmpk 11(2) (2023) 303-315 314 [18] x. xi, l. ming, j. liu. electrochemical determination of thiamazole at a multi-wall carbon nanotube modified glassy carbon electrode. journal of applied electrochemistry 40 (2010) 1449-1454. https://doi.org/10.1007/s10800-010-0122-x. [19] a. kutluay, m. aslanoglu. multi-walled carbon nanotubes/electro-copolymerized cobalt nanoparticlespoly (pivalic acid) composite film coated glassy carbon electrode for the determination of methimazole. sensors and actuators b: chemical 171 (2012) 1216-1221. https://doi.org/10.1016/j.snb.2012.06.083. [20] j. fahimeh, m. loghman, r. mahmoud. electrocatalytic determination of anti-hyperthyroid drug, methimazole, using a modified carbon-paste electrode. african journal of pharmacy and pharmacology 7 (2013) 269-274. https://doi.org/10.5897/ajpp12.717. [21] s. shahrokhian, m. ghalkhani. voltammetric determination of methimazole using a carbon paste electrode modified with a schiff base complex of cobalt. electroanalysis: an international journal devoted to fundamental and practical aspects of electroanalysis 20 (2008) 1061-1066. https://doi.org/10.1002/elan.200704149 [22] p. norouzi, v.k. gupta, b. larijani, m.r. ganjali, f. faridbod. a new methimazole sensor based on nanocomposite of cds nps–rgo/il–carbon paste electrode using differential fft continuous linear sweep voltammetry. talanta 127 (2014) 94-99. https://doi.org/10.1016/j.talanta.2014.03.061. [23] m. fouladgar, s. mohammadzadeh. determination of methimazole on a multiwall carbon nanotube titanium dioxide nanoparticle paste electrode. analytical letters 47 (2014) 763-777. https://doi.org/10.1080/00032719.2013.855782. [24] j. thangphatthanarungruang, a. lomae, o. chailapakul, s. chaiyo, w. siangproh. a low‐cost paper‐ based diamond electrode for trace copper analysis at on‐site environmental area. electroanalysis 33 (2021) 226-232. https://doi.org/10.1002/elan.202060305. [25] a. lomae, s. nantaphol, t. kondo, o. chailapakul, w. siangproh, j. panchompoo. simultaneous determination of β-agonists by uhplc coupled with electrochemical detection based on palladium nanoparticles modified bdd electrode. journal of electroanalytical chemistry 840 (2019) 439-448. https://doi.org/10.1016/j.jelechem.2019.04.003. [26] k. pungjunun, s. chaiyo, i. jantrahong, s. nantaphol, w. siangproh, o. chailapakul. anodic stripping voltammetric determination of total arsenic using a gold nanoparticle-modified boron-doped diamond electrode on a paper-based device. microchimica acta 185 (2018) 1-8. https://doi.org/10.1007/s00604-018-2821-7. [27] s. seyidahmet, f. dönmez, y. yardım, z. şentürk. simple, rapid, and sensitive electrochemical determination of antithyroid drug methimazole using a boron-doped diamond electrode. journal of the iranian chemical society 16 (2019) 913-920. https://doi.org/10.1007/s13738-018-1562-1. [28] t. yano, d. tryk, k. hashimoto, a. fujishima. electrochemical behavior of highly conductive boron‐ doped diamond electrodes for oxygen reduction in alkaline solution. journal of the electrochemical society 145 (1998) 1870. https://doi.org/10.1149/1.1838569 [29] a. economou, p.d. tzanavaras, m. notou, d.g. themelis. determination of methimazole and carbimazole by flow-injection with chemiluminescence detection based on the inhibition of the cu (ii)catalysed luminol–hydrogen peroxide reaction. analytica chimica acta 505 (2004) 129-133. https://doi.org/10.1016/s0003-2670(03)00176-4. [30] j. sun, c. zheng, x. xiao, l. niu, t. you, e. wang. electrochemical detection of methimazole by capillary electrophoresis at a carbon fiber microdisk electrode. electroanalysis: an international journal devoted to fundamental and practical aspects of electroanalysis 17 (2005) 1675-1680. https://doi.org/10.1002/elan.200403272. [31] n.a. martinez, g.a. messina, f.a. bertolino, e. salinas, j. raba. screen-printed enzymatic biosensor modified with carbon nanotube for the methimazole determination in pharmaceuticals formulations. sensors and actuators b: chemical 133 (2008) 256-262. https://doi.org/10.1016/j.snb.2008.02.025. [32] r. zakrzewski. determination of methimazole in pharmaceutical preparations using an hplc method coupled with an iodine-azide post-column reaction. journal of liquid chromatography & related technologies® 32 (2008) 383-398. https://doi.org/10.1080/10826070802631451. https://doi.org/10.1007/s10800-010-0122-x https://doi.org/10.1016/j.snb.2012.06.083 https://doi.org/10.1016/j.snb.2012.06.083 https://doi.org/10.5897/ajpp12.717 https://doi.org/10.1002/elan.200704149 https://doi.org/10.1016/j.talanta.2014.03.061 https://doi.org/10.1080/00032719.2013.855782 https://doi.org/10.1002/elan.202060305 https://doi.org/10.1016/j.jelechem.2019.04.003 https://doi.org/10.1007/s00604-018-2821-7 https://doi.org/10.1007/s00604-018-2821-7 https://doi.org/10.1007/s13738-018-1562-1 https://doi.org/10.1149/1.1838569 https://doi.org/10.1016/s0003-2670(03)00176-4 https://doi.org/10.1002/elan.200403272 https://doi.org/10.1002/elan.200403272 https://doi.org/10.1016/j.snb.2008.02.025 https://doi.org/10.1080/10826070802631451 admet & dmpk 11(2) (2023) 303-315 fia determination of methimazole doi: https://doi.org/10.5599/admet.1584 315 [33] l. molero, m. faundez, m.a. del valle, r. del río, f. armijo. electrochemistry of methimazole on fluorine-doped tin oxide electrodes and its square-wave voltammetric determination in pharmaceutical formulations. electrochimica acta 88 (2013) 871-876. https://doi.org/10.1016/j.elect acta.2012.10.142. [34] f. jalali, z. hatami. fast electrocatalytic determination of methimazole at an activated glassy carbon electrode. iranian journal of pharmaceutical research: ijpr 15 (2016) 735. https://www.ncbi.nlm.ni h.gov/pmc/articles/pmc5316251/. [35] methimazole tablet https:/dailymed.nlm.nih.gov/dailymed/druginfo.cfm, (accessed december 14, 2022). ©2023 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) https://doi.org/10.5599/admet.1584 https://doi.org/10.1016/j.electacta.2012.10.142 https://doi.org/10.1016/j.electacta.2012.10.142 https://www.ncbi.nlm.nih.gov/pmc/articles/pmc5316251/ https://www.ncbi.nlm.nih.gov/pmc/articles/pmc5316251/ https://dailymed.nlm.nih.gov/dailymed/druginfo.cfm http://creativecommons.org/licenses/by/3.0/ drug-like properties and fraction lipophilicity index as a combined metric doi: https://doi.org/10.5599/admet.1022 177 admet & dmpk 9(3) (2021) 177-190; doi: https://dx.doi.org/10.5599/admet.1022 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper drug-like properties and fraction lipophilicity index as a combined metric anna tsantili-kakoulidou 1 * and vassilis j. demopoulos 2 1 department of pharmaceutical chemistry, school of pharmacy, national and kapodistrian university of athens, panepistimiopolis, zografou, 157 71 athens, greece. e-mail: tsantili@pharm.uoa.gr 2 department of pharmaceutical chemistry, school of pharmacy, aristotle university of thessaloniki, 54124 thessaloniki, greece. e-mail: vdem@pharm.auth.gr *corresponding author received: june 13, 2021; revised: september 30, 2021; published online: october 10, 2021 abstract fraction lipophicity index (fli) has been developed as a composite drug-like metric combining log p and log d in a weighted manner. in the present study, an extended data set confirmed the previously established drug-like fli range 0-8 using two calculation systems for log p/log d assessment, the freeware medchem designer and clogp. the dataset was split into two classes according to the percentage of fraction absorbed (%fa) class 1 including drugs with high to medium absorption levels and class 2 including poorly absorbed drugs. the fli and fli-c (clogp based fli) drug-like range covers 92 % and 91 % of class 1 drugs, respectively. using mlogp, a narrower drug-like fli-m range 0-7 was established, covering 91 % of class 1 drugs. the dependence of the degree of ionization to intrinsic lipophilicity within the fli (fli-c, fli-m) drug-like range as well as the inter-relation between the other ro5 properties (mw, hd, ha) was explored to define drug-like / non-drug-like combinations as a safer alternative to single properties for drug candidates’ prioritization. in this sense, we propose a combined metric of mw and the number of polar atoms (mw/no) to account for both size and polarity. setting the value 50 as cutoff, a distinct differentiation between class 1 and class 2 drugs was obtained with mw/no>50 for more than 70 % of class 1 drugs, while the opposite was observed for class 2 drugs. ©2021 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords drug-likeness; oral drugs; fraction lipophilicity index; rule of 5; molecular weight; polar atoms introduction the role of physicochemical properties in controlling the fate of drug molecules within the organism and their binding to macromolecules has been well documented. emphasis has been given to lipophilicity, molecular size and polarity as the most crucial properties, which have been incorporated in the majority of the metrics used to set boundaries in the drug-like chemical space [1,2]. there is strong evidence that compounds with higher lipophilicity and higher molecular weight, e.g., with high molecular ‘obesity’ are more likely to be discontinued during clinical development [3-6]. they are associated with difficulties in oral absorption, while they are considered to exhibit increased ‘promiscuity’ towards biomacromolecules. as a result, they may lack selectivity, leading to side effects while they accumulate in the organism, increasing https://doi.org/10.5599/admet.1022 https://dx.doi.org/10.5599/admet.1022 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:tsantili@pharm.uoa.gr mailto:vdem@pharm.auth.gr http://creativecommons.org/licenses/by/4.0/ tsantili-kakoulidou and demopoulos admet & dmpk 9(3) (2021) 177-190 178 non-selective toxicity. the minimal hydrophobicity concept was formulated in 1987 by hansch et al. [7], who suggested that molecules exert their action rather by specific binding and not by non-specific hydrophobic interactions. ten years later, the well-known rule of five (ro5) formulated by lipinski for compounds intended for oral administration [8] suggested upper limits for lipophilicity (clogp≤5), molecular weight (mw≤500), counts of hydrogen bond donor (hd≤5) and acceptor sites (ha≤10), the latter two expressing polarity. compounds violating more than two of the ro5 conditions are prone to gastrointestinal absorption problems. ro5 as well as limits or ranges for additional drug-like properties such as a number of rotatable bonds, polar surface area, or a number of aromatic rings [2,9-11] may be used as virtual screening filters in early drug discovery. however, this strategy has been recently disputed, as there is an increasing interest in exploring beyond the rule of five (bro5) chemical space in drug discovery [12,13]. the concept of drug-likeness was further advanced to address drug safety by normalizing affinity to size or/and lipophilicity, as reflected in metrics like ligand efficiency (le), lipophilic ligand efficiency (lle) and ligand efficiency dependent lipophilicity (lelp) [1,14-16]. on the other hand, ghose et al. [17] suggest qualifying ranges of -0.4 to 5.6 for log p and 160 to 480 for molecular weight as filters for drug-likeness, thus setting lower limits lipophilicity and size [17]. in the above metrics, lipophilicity is expressed by log p, which corresponds to intrinsic lipophilicity of the neutral species, ignoring the effect of ionization in the case of drugs containing ionizable groups. intrinsic lipophilicity governs hydrophobic binding to biomacromolecules. however, permeability is affected rather by apparent lipophilicity, log d, as dictated by the partition-ph hypothesis, although intrinsic lipophilicity remains the driving force. in a retrospective analysis of human bioavailability data, compounds with log d values at ph 6.5 in the range of -2 to 3 were found to display increased bioavailability [18]. for physiological ph, an optimum log d range between 1-3 has been proposed [19], while waring established lower limits for log d7.4 in a molecular weight-dependent manner to achieve > 50 % chance for high permeability [20]. according to the author, the log d thresholds occur at higher values for higher molecular weight ranges, with large molecules (mw>500) necessitating log d> 4.5. analogous results on the dependence of optimal log d ranges on molecular weight were also reported by other authors [21]. the above suggestions underestimate the role of intrinsic lipophilicity, although the dependence of log d to molecular weight in waring’s approach indirectly introduces the bulk component, which is related to the hydrophobicity of the molecule, according to the dual nature of lipophilicity [22]. moreover, since log d values depend on log p, their magnitude alone does not provide information of the degree of ionization. a specific log d value may correspond to log p values spanning within 3 log units, considering that this is the difference between log p of the neutral and fully ionized species. recently, we have developed a new metric for assessing oral drug-likeness of ionizable chemical entities as a weighted combination of both log p and log d, considering that the two measures have a distinct role to play [23,24]. we refer to this metric as fraction lipophilicity index (fli). fli is defined as the logarithm of the quotient of partition coefficients (p) and the fraction of the neutral form (f n ) at a given ph (i.e., log(p/f n )), which upon analyzing log f n leads to equation 1: fli = 2log p-log d (1) eq.1 was further modified to eq. 2 by considering the absolute value of log d. this modification was assumed necessary since negative log d values, although unfavorable for absorption, contribute positively to fli calculation shifting the values to the center of the fli range. fli = 2log p │log d│ (2) admet & dmpk 9(3) (2021) 177-190 drug-likeness and fraction lipophilicity index doi: https://doi.org/10.5599/admet.1022 179 the drug-like fli range 0-8, based on calculations with the medchem designer freeware, was found to accommodate 89 % of drugs classified as highly or moderately absorbed, with the range 5 to 6 being the most populated. there is a slight shift of fli distribution to the upper limits of the above range for acidic compounds, which is restored if fli@ph5.5 is considered. this is justified assuming that absorption of acids is favored in the upper part of the intestine. the same drug-like fli range was found to be covered by 73 % of drugs with low absorption, while fli distribution is shifted towards negative values. the present study is a continuation of our previous investigation [23,24] in the aim to further elaborate the fli metric, extending the data set to include drugs approved till december 2020, as well as a larger number of older drugs, while using two additional log p calculation systems, namely clogp for direct comparison with lipinski’s ro5 and mlogp for which an upper limit 4.15 is suggested. the dependence of the degree of ionization to intrinsic lipophilicity within the drug-like space and the inter-relation between the ro5 properties are explored to define drug-like / non-drug-like combinations as a safer alternative to single properties for drug candidates’ prioritization, integrating fli in this perspective. materials and methods the entire dataset comprises 643 drugs, presented in table 1s, in supplementary material along with their smiles. it includes all drugs analyzed in our previous publication [24], drugs compiled by newby et al. [25], and later generation drugs approved in 2017-2020. all drugs included in the data set possess ionizable centers and display a degree of ionization at ph 7.4 for bases and ph 5.5 for acids higher than 30 %, corresponding to a difference δ(log p-log d) > 0.14. depending on their isoelectric point, zwitterions were considered if log dmax > 0.14 than log d at ph 7.4 or ph 5.5. the dataset is split into two classes according to their fraction absorbed (%fa) taken from newby’s compilation [25]. fraction absorbed was considered more appropriate than bioavailability, used in our previous publications, since the latter is a composite parameter involving several factors, such as gastrointestinal absorption, chemical stability, and the first-pass effect. for classification to class 1 (highly and medium absorbed drugs) and class 2 (poorly absorbed drugs), the cutoff value was set at 45 % fa. drugs in class 1 are designated with ‘m’ if %fa is in the range of 46-60 and with ‘h’ if %fa is higher than 60. drugs in class 2 are designated with ‘l’. for drugs with not available %fa in ref [25], information on absorption was taken from drugbank or other sources, as presented in table 1s, supplementary material. moreover, latest generation drugs with not available %fa data have been classified into class 1, if they are administered orally or to class 2 if other route of administration is reported. class 1 comprises 527 drugs (336 bases, 178 acids and 13 zwitterions) with high or moderate absorption, while class 2 comprises 116 drugs (54 bases, 54 acids and 8 zwitterions). calculation of fli partition coefficients (logps + ) and distribution coefficients at physiological ph for bases and ph 5.5 for acids (logds + ) were calculated using the freeware medchem designer(tm) version 3.0.0.28 (https://www.simulations-plus.com/). the same program provides log p calculations according to moriguchi (mlogp) [26]. for direct comparison with ro5, the software clogp for windows v.4.0 [http://www.biobyte.com/] was also used (clogp). mlogd and clogd values were assessed considering the difference δ(log p – log d) generated by medchem designer. δ(log p – log d) corresponds to [log(10 ±ph-pka) +1] and is independent of the log p calculation system. in the expression δ(log p – log d), log d stands for log d7.4 and log d5.5 for bases and acids, respectively. the program did not provide calculation of clogp for five compounds with high mw (>1000). for these cases, clogp values were predicted using eq. 3 which reflects the correlation between https://doi.org/10.5599/admet.1022 https://www.simulations-plus.com/ https://www.simulations-plus.com/ http://www.biobyte.com/ tsantili-kakoulidou and demopoulos admet & dmpk 9(3) (2021) 177-190 180 clogp and logps + values. clogp=1.021(±0.014)logps + + 0.014(±0.045) (3) n=626 r=0.948 s=0.678 equation 3 practically corresponds to 1:1 correlation. the mean difference δ(clogp log ps + ) is 0.07, although larger deviations are observed in individual cases (table 1s, supplementary material). eq. 3 was generated upon exclusion of 16 drugs with a high difference, │δ(clogp logps + )│ > 2. all lipophilicity data are presented in table 1s, supplementary material. in the same table, molecular weight (mw), number of hydrogen bond donor (hd) and acceptor sites (ha), designated as [n+o], as well as ro5 violation scores, are provided. mw, hd and [n+o] were calculated by medchem designer. ro5 are labeled according to log p calculation systems as scorero5/logps + , scorero5/clogp and scorero5/mlogp, respectively. calculations of ro5 scores have been performed manually for the first two, while medchem designer has provided scorero5/mlogp. fli values were calculated according to eq. 2 for ph=7.4 for bases and ph=5.5 for acids. fli values of zwitterions were calculated either at ph 7.4 or at ph 5.5 according to their isoelectric point. fli-c and fli-m were calculated using the corresponding clogp/clogd and mlogp/mlogd values. fli, fli-c and fli-m are included in table 1s, supplementary material. statistical analysis the program statistica 7.1, copyright© statsoft inc. 1984-2006, 2300 east 14th street, tulsa, ok 74104, usa was used for statistical analysis and histograms’ construction. results and discussion data overview the property overview of the drugs considered in the present study is shown in table 1 (supplementary tables). mean logps + , clogp and mlogp values are 2.63, 2.68 and 2.02, respectively, while mean mw is 372.5. clear differentiation in the mean values is observed between class 1 and 2 drugs. class 1 drugs display higher mean lipophilicity and lower mean mw, while the opposite is found for class 2. moreover, class 2 drugs possess a considerably higher number of hydrogen bond donor (hd) and acceptor sites ([n+o]). for class 1 drugs, high mw is associated with polar atoms [n + o] up to 15, while in class 2, large molecules possess up to 33 polar atoms and lower log p values. it should be noted that mean log p and mw values for class 1 are considerably higher than those reported for 1791 oral drugs of the chembl database -including both neutral and ionizable molecules which are close to 2.5 and 333, respectively [27,28]. as also commented in our previous publication [24], there is a shift towards higher log p and mw for drugs approved after 2000, which display mean logps + /clogp and mw values 3.34/3.42 and 430.9 respectively. in particular, for drugs approved between 2000-2009 mean logps + /clogp/mlogp and mw are 3.39/3.38/2.71 and 365.1, respectively, while a further increase in mean mw to 465.1 is observed for drugs approved in the last decade (2010-2020). for drugs before 2000, descriptive statistics show mean logps + clogp and mlogp 2.40, 2.44 and 1.88 respectively and a mean mw equal to 353.5. more to the point, a tendency to increase the number of polar atoms is evident after 2000 (table 1, supplementary tables). analogous findings are reported by hann and keserü [29] and shulz [30], who outlined property differences between target classes and companies, suggesting apparently different requirements of the expanding target space. this is, for instance, the case of pparγ agonists [31] or protease inhibitors [32], indicating that drug-like properties should be adapted over time. admet & dmpk 9(3) (2021) 177-190 drug-likeness and fraction lipophilicity index doi: https://doi.org/10.5599/admet.1022 181 exploring drug-like properties inter-relation as aforementioned, log d values per se do not give information about ionization if log p is not considered. to explore ionization throughout the logps + range, the difference δ(log p – log d) was compared to two logps + levels, above and below the drug-like limit of five. it was found that maximum δ(log p – log d), e.g. maximum ionization, is differentiated between the two logps + levels in both class 1 and class 2, with reduced ionization for highly lipophilic drugs (table 2, supplementary tables). the ionization pattern according to logps + boundaries is provided in figure 1. the log p dependent ionization will be further discussed in relation to the fli concept. figure 1. ionization pattern for drugs with logps + ≤5 (left) and logps + >5 (right) for class 1 drugs, ionization is observed to be also size-dependent. setting a discriminant at mw=600 reduced ionization is observed for drugs with mw>600 with maximum δ(log p – log d) equal to 1.75, while for drugs with mw<600, maximum δ(log p – log d) is 3.49. the analogous situation is observed for class 2 drugs, although higher ionization is generally observed with maximum δ(log p – log d) values 2.31 and 4.89 for mw>600 and mw<600, respectively (table 2, supplementary tables). figure 2 shows the different ionization levels according to mw for class 1 and class 2 separately. figure 2. mw dependent ionization levels for class 1 and class 2. (xaxis is in different scale to avoid compressing of the histograms in class 1, which does not contain drugs with very high molecular weight, as it is the case for class 2 drugs). https://doi.org/10.5599/admet.1022 tsantili-kakoulidou and demopoulos admet & dmpk 9(3) (2021) 177-190 182 regarding log d, there is a tendency to increase with mw. however, this trend is not followed for drugs with mw>600 and very hydrophilic drugs (figure 3a). for large molecules, this is due to a higher number of polar atoms [n + o], which contribute negatively to log p and, thereupon to log d values (figure 3b). in particular, for class 2 drugs, there is a correlation with r=0.860 between mw and the number of polar atoms [n+o] (figure 3c). figure 3. a: plot of mw versus logps + and logds + for the entire set. b,c: plot of mw versus the number of polar atoms [n+o] for the entire set and class 2 drugs respectively. drugs beyond the drug-like limits recently, there has been an increased interest in the development of drugs beyond the rule of five (bro5), especially in the area of oncology and direct-acting antivirals (daa) [13]. in this aspect we explored violations of ro5 based on logps + (scorero5/logps + ), clogp (scorero5/clogp) and mlogp (scorero5/mlogp) classifying drugs to those approved after or before 2000. as shown in figure 4, there is a considerable increase in the percentage of drugs with more than two violations after 2000 with 20, 18 and 14 % according to logps + , clogp and mlogp systems, respectively, compared to the corresponding 8, 8 and 7 % before 2000 (figure 4). if we focus on class 1 drugs approved after 2000, the bro5 cases are 16, 15 and 10 %, respectively (figure not shown). nevertheless, there is still a considerable differentiation in the distribution of ro5 violations between class 1 and class 2 drugs. for class 1, 82, 79 and 84 % of the drugs show 0 violations, respectively, 11,15 and 11 % display one violation and only 6, 5 and 4 % have higher than two violations. for class 2 drugs, zero violation is displayed for 48, 47 and 48 %, respectively, while 33, 31 and 29 % showed higher than 2. in figure 5 the distribution of violations in classes 1 and 2 are displayed as bars (left), and in the case of scorero5/logps + as pies (right). admet & dmpk 9(3) (2021) 177-190 drug-likeness and fraction lipophilicity index doi: https://doi.org/10.5599/admet.1022 183 figure 4. number of polar atoms [n+o] according to mw levels for class 1 and class 2 drugs. figure 5. distribution of ro5 violations in class 1 and class 2 drugs based on logps + and clogp. violations are presented in detail and compared to the corresponding %fa levels in table 3 (suppl. tables), also considering low lipophilicity limits, namely log p ≤ -0.4, the minimum value suggested by ghose et al. [17] and log p ≤ -1, set in this paper as a more determinative limit for oral absorption. 47 drugs display logps + <-0.4, 27 out of them with additional violations of the ro5 drug-like limits. eleven drugs (55 %) out of the 20 with no other ro5 violations and 20 drugs with additional ro5 violations (74 %) have low absorption. if the cutoff value is set to logps + < -1, the percentage of low %fa increases to 66.7 % (8 out of 12 drugs) and 82 % (14 out of 17) for drugs without additional ro5 violations. looking at the upper extreme, 67 drugs have logps + >5. among them, only eight drugs (12 %) have low %fa. considering the total population of class 2, 33 drugs (28 %) are more hydrophilic (logps + <-0.4), while only 8 (6.9 %) exceed the ro5 log p limit, indicating that high lipophilicity is rather overrated as an issue for oral absorption. regarding molecular weight, 80 drugs exceed the limit value of 500, 39 out of them (48.7 %) showing low %fa. among 20 drugs with twofold logps + /mw violations, four (20 %) have low absorption. on the other hand, among the 13 drugs with logps + <-0.4 and mw>500, 12 (92 %) and all ten drugs with logps + <-1 and mw >500 display low absorption. https://doi.org/10.5599/admet.1022 tsantili-kakoulidou and demopoulos admet & dmpk 9(3) (2021) 177-190 184 an analogous pattern is observed if clogp or mlogp is used instead of logps + (table 3, supplementary tables), although the cases are not always overlapping. these findings indicate that high mw in combination with polarity is more crucial than high lipophilicity for absorption difficulties. more to the point, considering hydrogen bonding potential, among the 22 drugs with violations of both hydrogen bond donor and acceptor sites (hd+[n+o]) 20 (91 %) show low absorption. fraction lipophilicity index (fli) in our previous investigation [24] the drug-like fli range 0-8 has been suggested, covering 89 % of the drugs with satisfactory absorption. this range was found to cover 73 % of the drugs with lower absorption, with the distribution shifted towards negative fli values. the above findings were further supported by the extended data set in the present study, with 92 % of class 1 drugs lying within the fli range 0-8. 6 % have negative fli values and 1 % fli values > 8. most populated is the fli range 2-6, covering 66 % of the drugs. for class 2 drugs the fli range 0-8 is covered by 56 % of the drugs and 34 % are within the range 2-6. 1 % have fli values >8, while 42 % have negative fli values (figure 6). the substantially lower coverage of the drug-like fli range by class 2 drugs, compared to our previous report [24], should be attributed to the use of %fa in this study instead of the more complex bioavailability. analogous results are obtained using fli-c, which is based on clogp. the fli-c drug-like range 0-8 is covered by 91 % of class 1 drugs, 61 % being within the range 2-6. for class 2 the range 0-8 is covered by 51 % of the drugs, while 46 % have negative fli-c values (figure 7). based on the relatively lower mlogp values, fli-m shapes a narrower drug-like range between 0-7, covering 91 % of class 1 drugs with most populated the region 3-5 (46 %). for class 2 drugs, 53 % are within the range 0-7 and 45 % have negative values (figure 8). it is evident that fli, fli-c and fli-m based on different calculation systems show the same performance, although in fli-m the drug-like range is compressed, in agreement with the lower mlogp upper limit. thus, in the next section, fli and fli-m are considered for further discussion. figure 6. fli distribution for class 1 and class 2 drugs as histograms (left) and pie charts (right). admet & dmpk 9(3) (2021) 177-190 drug-likeness and fraction lipophilicity index doi: https://doi.org/10.5599/admet.1022 185 figure 7. fli-c distribution for class 1 and class 2 drugs as histograms (left) and pie charts (right). figure 8. fli-m distribution for class 1 and class 2 drugs as histograms (left) and pie charts (right). fli as a combined metric comparison with ro5 considering ionizable drugs, the drug-like fli range 0-8 expands the lipophilicity cutoff values beyond 5 (or 4.15 for mlogp). in fact, log p values (logps + or clogp) expand to 7.7, provided that the compounds do not exhibit higher ionization than 50 % (considering that at 50 % ionization, lipophilicity decreases by 0.3 log units) or to slightly higher (7.87) if 30 % ionization is considered. for mlogp the upper limits expand to 6.7 or 6.87. such hard limit values are not realistic for the design of new drugs. our proposition is to think in terms of combinations of properties rather than single properties. in this aspect, fli represents a weighted combination of log p with log d, suggesting reduced ionization for drugs at the upper lipophilicity extremes. on the other hand, low log p and log d values leading to negative fli values are more discriminating for low absorption levels. in figure 9, the fli values inside the drug-like range along with the associated δ(log p – log d) levels for class 1 drugs are presented. for drugs with logps + >5 δ(log p – log d) values do not exceed 2.7, with most drugs showing δ(log p – log d)<2. although the majority for class 1 drugs with logps + ≤5 has https://doi.org/10.5599/admet.1022 tsantili-kakoulidou and demopoulos admet & dmpk 9(3) (2021) 177-190 186 also δ(log p – log d)<2, 67 drugs show higher ionization. similarly, no class 1 drugs with mlogp > 4.15 have δ(log p – log d)>2.5, while for 14 drugs with mlogp≤4.15 δ(log p – log d) exceeds 2.5. figure 9. left: ionization pattern for class 1 drugs, according to fli levels within the drug-like range, for subsets including drugs with logps + >5 and logps + ≤5 respectively. right: ionization pattern for class 1 drugs, according to fli-m levels within the drug-like range, for subsets including drugs with logps + >4.15 and logps + ≤4.15 respectively. in figure 10, the combination of fli and rule of five (scorero5/logps + ) is shown for class 1 and the subset, including fli values inside the drug-like range for class 2 and the subset including drugs outside the drug-like range. it is shown that 25 drugs (5.2 %) in class 1 drug-like subset (figure 10b) display more than twofold ro5 violation, 12 out of them involving logps + and mw. considering the fli metric, these 12 drugs would not receive a second alert, and thus they would not be considered bro5 drugs. on the other hand, 33 drugs (28 %) of class 2 (figure 10d) show none or one ro5 violation. however, in terms of their fli values, a warning should be set for these drugs according to this metric. considering fli-m, 16 class 1 drugs within the drug-like range (figure 11b) show the twofold violation of the corresponding ro5 score, 5 of them involving mlogp >4.15. on the other hand, there should be a warning for 37 drugs of class 2 outside the drug-like range, although they show 0 or 1 ro5 violation (figure 11d). considering the merits of fli as a composite metric and in the light of the combination mentioned above of high mw with high polarity as a crucial issue for low absorption, we go a step further to propose the normalization of mw to the number of polar atoms [n+o]. the term mw/no is calculated according to the expression mw/(1+[n+o]), where 1 stands to avoid division by 0 in the case of compounds that lack n and o atoms. considering the distribution of mw/no in classes 1 and 2 (figure not shown), a discriminant value of 50 is set to differentiate the two classes. for class 1, the majority (72 %) of drugs display mw/no>50, while the opposite occurs for class 2 drugs, with 75 % showing mw/no≤50 (figure 12). admet & dmpk 9(3) (2021) 177-190 drug-likeness and fraction lipophilicity index doi: https://doi.org/10.5599/admet.1022 187 figure 10. rule of 5 violations according to fli levels. a: class 1, b: class 1 subsetdrugs inside the drug-like fli range, c: class 2, d: class 2, subsetdrugs outside the drug-like fli range. figure 11. rule of 5 violations according to fli-m levels. a: class 1, b: class 1 subsetdrugs inside the drug-like fli-m range, c: class 2, d: class 2, subsetdrugs outside the drug-like fli-m range. https://doi.org/10.5599/admet.1022 tsantili-kakoulidou and demopoulos admet & dmpk 9(3) (2021) 177-190 188 figure 12. pie charts of the distribution of mw/no for class 1 and class 2 drugs. conclusions the previously established drug-like fli range 0-8 is confirmed using an extended dataset, covering 93 % of highly to moderately absorbed drugs. application of three different methods, medchem designer, clogp and mlogp for lipophilicity assessment in fli generation, supported the robustness of the results while showing that the metric is overall independent of log p calculation system. however, the fli-m values generated by mlogp, which is associated with a lower drug-like upper limit, shape a narrower drug-like range between 0-7, covering 91 % of class 1 drugs. fli tolerates higher log p values for drugs, provided that ionization is limited. it offers the option to explore a wider chemical space for drug discovery, notwithstanding the disadvantages of high lipophilicity regarding other issues, like promiscuity, extensive metabolism or toxicity. on the other hand, negative fli values minimize the chances for oral absorption. the merit of fli lies in the fact that it is a combined metric of lipophilicity and ionization and provides more information on drug characteristics while ‘softening’ the hard limits. we, therefore, suggest that a combination of metrics may be preferable to single properties. in this sense and considering the crucial two-fold violation including mw and [n+o], we propose a combined metric of mw and the number of polar atoms [mw/no] to account for both size and polarity. exploration of mw/no will follow in future investigations. conflict of interest: authors claim no conflict of interest. references [1] p.d. leeson, b. springthorpe. the influence of drug-like concepts on decision-making in medicinal chemistry. nature reviews drug discovery 11 (2007) 881-90. https://doi.org/10.1038/nrd2445. [2] i. yusof, m.d. segall. considering the impact drug-like properties have on the chance of success. drug discovery today 18 (2013) 659-666. https://doi.org/10.1016/j.drudis.2013.02.008. [3] m. wenlock, r.p. austin, p. barton, a.m. davis, p.d. leeson. a comparison of physiochemical property profiles of development and marketed oral drugs. journal of medicinal chemistry 46 (2003) 1250–1256. https://doi.org/10.1021/jm021053p. [4] m.m. hann. molecular obesity, potency and other addictions in drug discovery. medicinal chemistry communications 2 (2011) 349-355. https://doi.org/10.1039/c1md00017a. [5] s. mignani, s. huber, h. tomás j. rodrigues, j.p. majoral. compound high-quality criteria: a new vision to guide the development of drugs, current situation. drug discovery today 21 (2016) 573-584. https://doi.org/10.1016/j.drudis.2016.01.005. [6] s. wang, g. dong, c. sheng. structural simplification: an efficient strategy in lead optimization. acta pharmaceutica sinica b 9 (2019) 880-901 https://doi.org/10.1016/j.apsb.2019.05.004. https://doi.org/10.1038/nrd2445 https://doi.org/10.1016/j.drudis.2013.02.008 https://doi.org/10.1021/jm021053p https://doi.org/10.1039/c1md00017a https://doi.org/10.1016/j.drudis.2016.01.005 https://doi.org/10.1016/j.apsb.2019.05.004 admet & dmpk 9(3) (2021) 177-190 drug-likeness and fraction lipophilicity index doi: https://doi.org/10.5599/admet.1022 189 [7] c. hansch, j. p. bjorkroth, a. leo. hydrophobicity and central nervous system agents: on the principle of minimal hydrophobicity in drug design. journal of pharmaceutical sciences 76 (1987) 66-687. https://doi.org/10.1002/jps.2600760902. [8] c.a. lipinski, f. lombardo, b.w. dominy, p.j. feeney. experimental and computational approaches to estimate solubility and permeability in drug discovery and development settings. advanced drug delivery reviews 46 (1997) 3-26. https://doi.org/10.1016/s0169-409x(00)00129-0. [9] d.f. veber, s.r. johnson, h.-y. cheng, b.r. smith, k.w.ward, k.d. kopple. molecular properties that influence the oral bioavailability of drug candidates. journal of medicinal chemistry 45 (2002) 26152623. https://doi.org/10.1021/jm020017n. [10] t.i. oprea. current trends in lead discovery: are we looking for the appropriate properties? journal of computer-aided molecular design 16 (2002) 325-334. https://doi.org/10.1023/a:1020877402759. [11] t.j. ritchie, s. j. f. macdonald. how drug-like are 'ugly' drugs: do drug-likeness metrics predict adme behaviour in humans? drug discovery today 19 (2014) 489-495. https://doi.org/10.1016/j.drudis.2014.01.007. [12] s. pathania, p.k. singh. analyzing fda-approved drugs for compliance of pharmacokinetic principles: should there be a critical screening parameter in drug designing protocols?, expert opinion on drug metabolism & toxicology 17 (2021) 4, 351-354. https://doi.org/10.1080/17425255.2021.1865309. [13] d.a. degoey, h-j chen, p. b. cox, m. d. wendt. beyond the rule of 5: lessons learned from abbvie’s drugs and compound collection journal of medicinal chemistry 61 (2018) 2636−2651. https://doi.org/10.1021/acs.jmedchem.7b00717. [14] a.l. hopkins, c.r. groom, a. alex. ligand efficiency: a useful metric for lead selection. drug discovery today 9 (2004) 430-431. https://doi.org/10.1016/s1359-6446(04)03069-7. [15] a.l. hopkins, g.m. keserü, p.d. leeson, d.c. rees, c.m. reynolds. the role of ligand efficiency metrics in drug discovery. nature reviews drug discovery 13 (2014) 105-121. https://doi.org/10.1038/nrd4163. [16] g.m.keserű, g.m. makara. the influence of lead discovery strategies on the properties of drug candidates. nature reviews drug discovery 8 (2009) 203-12. https://doi.org/10.1038/nrd2796. [17] a.k. ghose, v.n. viswanadhan, j.j. wendoloski. a knowledge-based approach in designing combinatorial or medicinal chemistry libraries for drug discovery. 1. a qualitative and quantitative characterization of known drug databases. journal of combinatorial chemistry 1 (1999) 55-68. https://doi.org/10.1021/cc9800071. [18] f.yoshida, j.g. topliss. qsar model for drug human oral bioavailability. journal of medicinal chemistry 43 (2000) 2575-85. https://doi.org/10.1021/jm0000564. [19] j.e.a, comer. drug bioavailability: estimation of solubility, permeability, absorption and bioavailability methods and principles in medicinal chemistry, h.van de waterbeemd, h. lennernäs, p. artursson eds, wiley vch, weinheim, 2003. p. 21-45. https://doi.org/10.1002/3527601473.ch2. [20] m.j. waring. defining optimum lipophilicity and molecular weight ranges for drug candidatesmolecular weight dependent lower logd limits based on permeability. biorganic medicinal chemistry letters 19 (2009) 2844-2851. https://doi.org/10.1016/j.bmcl.2009.03.109. [21] g. camenisch, j. alsenz, h. van de waterbeemd, g. folkers. estimation of permeability by passive diffusion through caco-2 cell monolayers using the drugs’ lipophilicity and molecular weight. european journal of pharmaceutical sciences 6 (1998) 313–319. https://doi.org/10.1016/s09280987(97)10019-7. [22] b.testa, p.-a. carrupt, p. gaillard, and r.-s. tsai, r.-s. intramolecular interactions encoded in lipophilicity: their nature and significance. v. pliska, b.testa, and h. van de waterbeemd, h. eds. lipophilicity in drug action and toxicology. weinheim: vch, 1995 pp. 49-71. [23] a. tsantili-kakoulidou, m. chatzopoulou, and v. j. demopoulos. fraction lipophilicity index (fli): a metric for assessing oral drug-likeness of ionizable chemical entities. international journal of quantitative structure-property relationships. 4 (2019) 41-66. https://doi.org/10.4018/ijqspr.2019010104. https://doi.org/10.5599/admet.1022 https://doi.org/10.1002/jps.2600760902 https://doi.org/10.1016/s0169-409x(00)00129-0 https://doi.org/10.1021/jm020017n https://www.ncbi.nlm.nih.gov/pubmed/12489682 https://link.springer.com/journal/10822 https://link.springer.com/journal/10822 https://doi.org/10.1023/a:1020877402759 https://doi.org/10.1016/j.drudis.2014.01.007 https://doi.org/10.1016/j.drudis.2014.01.007 https://doi.org/10.1080/17425255.2021.1865309 https://doi.org/10.1021/acs.jmedchem.7b00717 https://doi.org/10.1016/s1359-6446(04)03069-7 https://doi.org/10.1038/nrd4163 https://doi.org/10.1038/nrd4163 https://doi.org/10.1038/nrd2796 https://doi.org/10.1021/cc9800071 https://doi.org/10.1021/jm0000564 https://onlinelibrary.wiley.com/action/dosearch?contribauthorstored=lennern%c3%a4s%2c+hans https://onlinelibrary.wiley.com/action/dosearch?contribauthorstored=artursson%2c+per https://doi.org/10.1002/3527601473.ch2 https://doi.org/10.1016/j.bmcl.2009.03.109 https://doi.org/10.1016/s0928-0987(97)10019-7 https://doi.org/10.1016/s0928-0987(97)10019-7 https://doi.org/10.4018/ijqspr.2019010104 https://doi.org/10.4018/ijqspr.2019010104 tsantili-kakoulidou and demopoulos admet & dmpk 9(3) (2021) 177-190 190 [24] a. tsantili-kakoulidou, v.j. demopoulos. fraction lipophilicity index (fli). a drug-like metric for orally administered ionizable drugs. sar and qsar in environmental research. 30 (2019) 643-653. https://doi.org/10.1080/1062936x.2019.1653363. [25] d. newby, a.a. freitas, t. ghafourian. decision trees to characterise the roles of permeability and solubility on the prediction of oral absorption european journal of medicinal chemistry. 90 (2015) 751-65. https://doi.org/10.1016/j.ejmech.2014.12.006. [26] i. moriguchi, s.hirono, q. liu, i. nakagome, y.matsushita. simple method of calculating octanol/water partition coefficient. chemical & pharmaceutical bulletin 40 (1992) 127-130. https://doi.org/10.1248/cpb.40.127. [27] m.p.gleeson p, a. hersey, d. montanari, j. overington. probing the links between in vitro potency, admet and physicochemical parameters. nature reviews drug discovery. 3 (2011) 197-208. https://doi.org/10.1038/nrd3367. [28] h.a. zhong, v. mashinson, t.a. woolman, m. zha. understanding the molecular properties and metabolism of top prescribed drugs. current topics in medicinal chemistry. 13 (2013) 1290-1307. https://doi.org/10.2174/15680266113139990034. [29] m. hann, g. keserü. finding the sweet spot: the role of nature and nurture in medicinal chemistry. nature reviews drug discovery 11 (2012) 355–365. https://doi.org/10.1038/nrd3701. [30] m. shulz. two decades under the influence of the rule of five and the changing properties of approved oral drugs. journal of medicinal chemistry 62 (2019) 1701–1714. https://doi.org/10.1021/acs.jmedchem.8b00686. [31] t. vallianatou, c. giaginis, a. tsantili-kakoulidou. the impact of physicochemical and molecular properties in drug design: navigation in the "drug-like" chemical space. advances in experimental medicine and biology 822 (2015) 187-94. https://doi.org/10.1007/978-3-319-08927-0_21. [32] m.vieth, j.j.sutherland. dependence of molecular properties on proteomic family for marketed oral drugs. journal of medicinal chemistry 49 (2006) 3451−3453. https://doi.org/10.1021/jm0603825. ©2021 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/) https://www.tandfonline.com/toc/gsar20/current https://doi.org/10.1080/1062936x.2019.1653363 https://doi.org/10.1016/j.ejmech.2014.12.006 https://doi.org/10.1248/cpb.40.127 https://www.ncbi.nlm.nih.gov/pubmed/?term=montanari%20d%5bauthor%5d&cauthor=true&cauthor_uid=21358739 https://www.ncbi.nlm.nih.gov/pubmed/?term=overington%20j%5bauthor%5d&cauthor=true&cauthor_uid=21358739 https://doi.org/10.1038/nrd3367 http://www.ncbi.nlm.nih.gov/pubmed/?term=zhong%20ha%5bauthor%5d&cauthor=true&cauthor_uid=23675936 https://doi.org/10.2174/15680266113139990034 https://doi.org/10.1038/nrd3701 https://doi.org/10.1021/acs.jmedchem.8b00686 https://doi.org/10.1021/acs.jmedchem.8b00686 https://pubmed.ncbi.nlm.nih.gov/25416989/ https://pubmed.ncbi.nlm.nih.gov/25416989/ https://doi.org/10.1007/978-3-319-08927-0_21 https://doi.org/10.1021/jm0603825 http://creativecommons.org/licenses/by/4.0/ therapeutic potential of interferon-gamma in tuberculosis doi: http://dx.doi.org/10.5599/admet.1078 63 admet & dmpk 10(1) (2022) 63-73; doi: https://doi.org/10.5599/admet.1078 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index review therapeutic potential of interferon-gamma in tuberculosis svetlana a. berns 1 , julia a. isakova 2 and polina i. pekhtereva 3, * 1 federal state institution national medical research center for therapy and preventive medicine of the ministry of healthcare of the russian federation, svberns@yandex.ru 2 spp “pharmaclon” ltd., moscow, russia, julia.is.alex@gmail.com 3 spp “pharmaclon” ltd., moscow, russia, pekhterevapolina@gmail.com *corresponding author: polina i. pekhtereva; e-mail: pekhterevapolina@gmail.com; tel.: +7-909-675-9643 received: august 11, 2021; revised: november 27, 2021; published: january 14, 2021 abstract tuberculosis is one of the critical health problems worldwide. the search for ways to improve the results of tuberculosis treatment and overcome drug resistance lies in understanding the pathogenesis of the development of the infectious process. the interferon system, particularly the role of interferon-gamma, has been identified as the main link in the immune response in tuberculosis. the clinical efficacy of interferon-gamma has been studied and evaluated in clinical trials since the end of the last century. there was obtained evidence of the clinical efficacy of interferon-gamma as part of complex therapy. recent experimental data make it possible to consider interferon-gamma as a promising therapeutic option for the treatment of multidrug-resistant tuberculosis as part of complex therapy worthy of further studies. ©2021 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords interferon gamma; mrd-tb; immunotherapy; ifn-γ; complex treatment. introduction the problem of resistance of microorganisms to antimicrobial drugs and, consequently, issues related to the ineffectiveness of treatment remain one of the most urgent in medicine. the world health organization (who) developed and published in 2001 the who global strategy for containing antimicrobial resistance, which recommends considering this problem as one of the priorities of national health systems [1]. the significance of the problem of antibiotic resistance is determined by the widespread prevalence of resistant microorganisms in inpatient departments and the wide access to drugs used for the treatment of multi-resistant strains [2]. a decrease in the effectiveness of therapy ultimately requires the use of more expensive drugs and an increase in the duration of treatment, which leads to hospitalization time increase, economic losses, and mortality risk [3]. tuberculosis (tb) is a socially significant disease and the deadliest infection, continuing to be the most pressing challenge for the healthcare system around the world. according to who, tb is the ninth leading cause of death in the world, having been the leading cause of death from any single infectious agent, ahead of human immunodeficiency virus (hiv) / acquired immunodeficiency syndrome (aids), till 2020, when http://dx.doi.org/10.5599/admet.1078 https://doi.org/10.5599/admet.1078 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:svberns@yandex.ru mailto:julia.is.alex@gmail.com mailto:pekhterevapolina@gmail.com mailto:pekhterevapolina@gmail.com http://creativecommons.org/licenses/by/4.0/ svetlana a. berns et al. admet & dmpk 10(1) (2022) 63-73 64 covid-19 occupied this place [4]. in 2019, who estimates that 1.21 million hiv-negative people died from tb (less than 1.7 million deaths in 2000), and 208,000 deaths were reported among hiv-positive people [5]. thus, global tb deaths have been reduced by 35 % from 2015 to 2020. the incidence of tb has decreased by 20 % over the same period. in this way, there are notable successes in the fight against this infectious disease, but so far, not enough to achieve the global goals set by the who [5]. most tb deaths can be prevented with early diagnosis and appropriate treatment. millions of people are diagnosed and successfully treated with tb each year, but there are still significant gaps in the detection and therapy of the disease. the situation may be aggravated by the epidemic threat of covid-19, in particular, due to the redistribution of human, financial, and other resources not in favour of diagnosis and treatment of tuberculosis. the number of medical institutions providing inpatient and outpatient care for tb patients has decreased, and there are overlaps in data collection and reporting [5]. in conditions of unstable drug supply, underdeveloped infrastructure, and limited resources [6], the threat of drug-resistant tb remains. according to who estimates, in 2019, in the world, 3.3 % (95 % confidence interval (ci): 2.3–4.3 %) of newly diagnosed tb cases and 18 % (95 % ci: 9.7–27 %) of previously treated patients have been identified as resistant to rifampicin (rr-tb), the most effective first-line drug, or multidrug-resistant tb (mdr-tb). among them, 20.1 % (95 % ci: 15.5-25.0 %) were also resistant to fluoroquinolones, the drugs of choice in establishing resistance to rifampicin or isoniazid. it is alarming that about half of these cases are registered in india, china and the russian federation [5]. the problem of overcoming drug resistance to tb is also complicated because chemotherapy itself has an antimicrobial effect and suppresses the immune system's activity and can aggravate the patient's immunological reactivity. impaired immune response in tb patients can be corrected by immunotherapy. the correction of the immune status in mdr-tb patients is of particular importance because these patients initially have a worse prognosis and a more severe course of the disease [7]. in such conditions, the possibility of adjuvant immunotherapy, which can change the course of the disease and alleviate the patient's condition, gains a special role. one of the ways of such therapy is to regulate the body's immune response. features of the formation of the immune response in bacterial infection returning to the pathophysiological background of the immune response in bacterial infection, it is known that when the pathogen enters the human body for the first time, the innate and adaptive immune response mechanisms get activated [8]. in case of repeated exposure to the same pathogen, i. e. upon reinfection, the removal of the pathogen can occur at a higher rate due to protective immunity and immunological memory [9]. the important role of interferons (ifns) in the implementation of the immune response process was proven back long ago [10,11]. the search for new ways to overcome drug resistance in patients with mdrtb aimed at the role of ifn as an immunomodulator in the complex therapy of tb. the ifn system includes the ifn cytokines themselves, transcriptional activators and repressors, cellular receptors and enzyme systems activated by them. ifns increase the resistance of cells to infectious agents and propagate signals to other cells that have not previously had contact with either the antigen or ifn. all ifns are divided into 3 types: ifn of type i (-α, -β, -κ, -ε, -ω, -τ) [12], type ii (-γ) [13] and type iii (-λ1-4, -£1-3) [14]. at an early stage of infection (several hours), ifns of types i and ii are produced as the first line of immune defense in order to attract the largest number of dendritic cells and macrophages to the site of injury, which trigger active phagocytosis and inactivate the pathogen. however, if there is a defect in the admet & dmpk 10(1) (2022) 63-73 therapeutic potential of interferon-gamma in tuberculosis doi: http://dx.doi.org/10.5599/admet.1078 65 immune mechanism, attracted macrophages can promote infection, providing the microorganism with a target for intracellular growth and spread [15,16]. in this case, type i ifns can contribute to the development of tb by inducing il-10 and deactivating macrophages [17,18]. the immune defense of the first level will be broken and the cellular response to the antigen blocked. as a result, a latent form of tb may develop [19]. the optimal immune response is formed in a few days after infection. cd4+ and cd8+ effector t cells are directed to the affected area and start inducing type ii ifn, strongly shifting the balance towards this class of cytokines [15] and reducing the risk of developing an active tb process [20]. it was found that cd4+ lymphocyte count below 100 cells/mm3 before treatment significantly increases the likelihood of an unfavourable outcome of generalized tuberculosis in hiv patients [21]. at least 10 genes are known to be associated with increased susceptibility to mycobacterial infection [22]. first of all, these are genes encoding ifn-γ and its receptors, as well as cellular factors stat1 [23] and nfκb [24], directly related to the mechanism of ifn-γ action on the cell: antigen recognition, the realization of an innate immune response, and oxidative properties of macrophages [25], and necessary for protection against tb [26] and other infectious diseases, since they determine nonspecific pathways of action on the pathogen (figure 1). figure 1. mechanism of ifn-γ action via the immune cells http://dx.doi.org/10.5599/admet.1078 svetlana a. berns et al. admet & dmpk 10(1) (2022) 63-73 66 disseminated forms of infections in children often arise due to congenital defects in the structure of the ifn-γ protein [27]. the deficiency of genes encoding protein chains of receptors is one of the most common mutations that cause increased susceptibility to mycobacterial infection. the ifnγr1 gene accounts for 39 % of mutations [28]. for reasons still unclear, ifnγr2 deficiency is much less common [29]. in patients with complete deficiencies of any of these genes, there is an absence or incomplete formation of granulomas and diffusely loaded macrophages [30]. features of the formation of bacterial resistance the main mechanisms of the resistance development include genetic changes that ensure the "removal of the effect" of the antibiotic (change in the target, metabolism, method of removing the antibiotic from the cell), as well as the isolation of the microorganism from the drug (intracellular localization, biofilm formation) [31]. in the case of m. tuberculosis, the microorganism forms resistance due to the hydrophobic cell wall, β-lactamases, and mutations in the bacterial genes [7]. the immune system ensures the elimination of the pathogen primarily due to the strength of the cellular immunity by activating the phagocytosis. it was shown that ifn-γ could affect each of the main aspects of the phagocytes activation, thereby playing an important role in the immune response [32-34]:  activates macrophages, attracts them to the site of infection;  induces the release of nitric oxide;  serves as the only inducer in the cell of the synthesis of proteins of the second class histocompatibility complex (mhc class ii) an antigen-presenting complex for extracellular pathogens. the primary feature of the functioning of phagocytes, which ensures the relationship of nonspecific and specific immunity, is the performance of the role of antigen-presenting cells. antigen-presenting cells are located on the main routes of entry of antigens into the body (in the skin and the mucous membranes), from where, after capturing antigens, they migrate to the peripheral organs of the immune system, where they present antigens to lymphocytes. phagocytes are the cells that directly kill bacteria through phagocytosis ("eating" them) and play one of the leading roles in the body's first immune response to bacterial infection. hereditary and acquired disorders of phagocytosis can lead to severe pathologies of the functioning of the immune system [35]. mechanisms of action of phagocytes:  a phagocyte cell carries receptors on its surface that recognize an infectious agent (for example, the cell wall of a bacterium). when a phagocyte comes into contact with a bacterium, receptors on its surface bind to it, leading to the bacterium's absorption by the phagocyte.  release of reactive oxygen species and nitric oxide by phagocytes cause oxidative stress that destroys bacteria;  stimulation of other cells of the immune system through "antigen presentation" is a process in which phagocytes place fragments of pathogen proteins on the surface of their cells, thereby presenting these fragments to other cells of the immune system, developing an antigen-specific immune response [8,35-37]. it is important to note that at an early stage of infection, ifns of types i and ii are produced in equal concentrations, and ifns of type i are able to reduce the expression of the major histocompatibility admet & dmpk 10(1) (2022) 63-73 therapeutic potential of interferon-gamma in tuberculosis doi: http://dx.doi.org/10.5599/admet.1078 67 complex (mhc or mhc) of class ii on antigen-presenting cells, induced by ifn of type ii, preventing it action [38]. thus, during the development of a bacterial infection, the effects of ifn-γ provide a relationship between the two most important links of the immune response of a macroorganism, enhancing the antigen-dependent immune response and stimulating the work of phagocytes. an additional factor affecting the course of a bacterial infection is the ability of bacteria to inhibit the synthesis of pro-inflammatory cytokines. it has been shown that some microorganisms can specifically repress the synthesis of endogenous ifn-γ [39]. in a study by fortes et al., it was demonstrated that patients infected with an antibiotic-resistant strain of mycobacterium tuberculosis have a reduced level of endogenous ifn-γ compared to normal patients [40]. it has also been shown that the use of beta-lactam antibiotics can decrease the level of endogenous ifn-γ [41]. thus, the ifn system and directly ifn-γ are key regulators of the body's immune response to the infection caused by tb. therefore, additional exposure to exogenous ifn-γ can play the role of an inducer of the immune system under conditions of a reduced response to antibiotics and correct the endogenous cytokine deficiency. despite all the attractiveness of the idea, unfortunately, interest in the development has not yet been represented by widely deployed large-scale clinical trials (except for diagnostics use research). this is probably due to the limited number of medical products based on interferon-gamma presented on the world market. nevertheless, in our view, the diagnostic and clinical value of the molecule will undoubtedly gain momentum in the nearest future. clinical efficacy of interferon-gamma in the treatment of tuberculosis on the open resource https://clinicaltrials.gov/ 99 clinical trials with keywords “interferon-gamma” and “tuberculosis” were found, 15 of which are in active status, 64 are completed, 20 are cancelled or have an unknown status. out of 64 completed studies, 16 studies have published results. current research is mainly concerned with the diagnostic use of the interferon-gamma secretion test to assess the immune status of the body [42]. the therapeutic activity of ifn-γ against tb and mdr-tb has been intensively studied since the end of the last century; a number of clinical studies have demonstrated the effectiveness of ifn-γ against tb, including mdr-tb [43-49]. a meta-analysis that pooled data from 9 clinical trials showed a statistically significant effect for ifn-γ as an adjunct drug for tb [43]. the studies of nebulized ifn-γ showed statistically significant benefits as a result of negative sputum conversion and chest radiographs: the pooled relative risk (rr) for conversion was 1.97 (95 % confidence interval (ci) 1.20-3.24, p = 0.008) after 1 month of treatment, 1.74 (95 % ci 1.302.34, p = 0.0002) after 2 months of treatment, 1.53 (95 % ci 1.16-2.01, p = 0.003) after 3 months of treatment, 1.57 (95 % ci 1.20-2.06, p = 0.001) after 6 months of treatment, and 1.55 (95 % ci 1.17-2.05, p = 0.002) at the end of treatment; the pooled rr for chest radiographs was 1.38 (95 % ci 1.10–1.17, p = 0.006) at the end of treatment [44]. a randomized controlled trial with aerosol and subcutaneous injected ifn-γ showed a significant reduction in fever symptoms, wheezing and night sweats compared with the control group after one month of treatment [46]. for intramuscularly administered ifn-γ, the meta-analysis included three studies that showed significant improvement in negative sputum conversion after two months of treatment [43]. intramuscular injection of ifn-γ also showed a beneficial effect in treating pulmonary mdr-tb [45]. the authors concluded that adjuvant therapy with ifn-γ, especially in an aerosol form, may be beneficial for tb patients. http://dx.doi.org/10.5599/admet.1078 https://clinicaltrials.gov/ svetlana a. berns et al. admet & dmpk 10(1) (2022) 63-73 68 the results of another randomized clinical trial published in 2016 also confirmed the high clinical efficacy of ifn-γ as part of the complex therapy for tb. this study evaluated the clinical efficacy of ifn-γ in the intensive phase of complex treatment of newly diagnosed patients with destructive pulmonary tb with bacterial excretion. the dynamics of laboratory and immunological parameters were also assessed. the study included 60 patients (the study group included 30 newly diagnosed patients who received 500,000 iu of ifn-γ intramuscularly every other day, for a month in addition to standard therapy; control group 30 patients who received only standard therapy). the predominant clinical form was infiltrative, diagnosed in 80 % of patients in both groups (p> 0.05). clinical efficacy was assessed in patients who received ifn-γ after two months of the intensive phase of treatment. according to the results of a control x-ray examination, cavities in the lungs closed in 23 % of cases, and by four months, this indicator was 50 %. while in patients in the control group, the closure of cavities in the lungs was 10 %, in the study group, this indicator was three times higher (37 %) (p <0.05). abacillation occurred in 90 % of patients receiving ifn-γ. a month later, the normalization of indicators of the clinical analysis of blood occurred in 70 % of patients. in the group of patients who additionally received ifn-γ, the number of cd4+ lymphocytes increased to 41 %, exceeding the initial level by 1.7 times (p <0.001). the authors of the study concluded that the addition of immunomodulatory therapy is a significant addition to anti-tb drugs in terms of pathogenesis. the use of ifn-γ in the complex treatment of patients with destructive pulmonary tb with bacterial excretion has shown its high efficiency in the intensive phase of chemotherapy, accompanied by an increase in patients' quality of life. the results of the study indicate a significant improvement in the quality of life based on the assessment according to the methodology recommended by the who. after complex treatment, when studying the quality of life of patients in the main group, there was a positive trend in aspects characterizing social relationships and the psychological sphere, as well as reflecting the physical and mental health of the patient. it should be noted that the addition of ifn-γ to anti-tuberculosis therapy was not accompanied by a significant increase in the number of adverse reactions (20 % vs. 17 % in the control group) (p> 0.05). all adverse reactions were removable and stopped by pathogenetic methods of concomitant treatment [48]. one more clinical example was published in the scientific literature about the experience of ifn-γ medical use in mdr-tb. the treatment of mdr-tb of the brain in a patient with acute lymphocytic leukemia, after five months of adjuvant therapy using subcutaneously injected ifn-γ, showed a positive trend with the achievement of complete remission after 12 months of therapy [47]. in another study evaluating the effects of ifn-γ in mdr-tb, it was found for resistant forms of tb that there is an increase in the basal and il12/il27-induced secretion of ifn-γ by blood lymphocytes in vitro, with the maximum severity of changes in the disseminated form of the disease. hypersecretion of ifn-γ in vitro is associated with a low content of cd3-ifn-γ+ lymphocytes and an increase in the proportion of cd3ifn-γ+ against the background of general t-lymphocytopenia [50]. a recent study to assess the effects of ifn-γ on macrophage activity in overcoming mdr-tb drug resistance demonstrated the high efficacy of ifn-γ as an adjuvant therapy for mdr-tb. the authors of the study draw encouraging conclusions about the prospect of using ifn-γ as a part of adjuvant therapy for mdr-tb. it increases the activity of macrophages against mycobacterium tuberculosis and improves the activity of macrophages treatment outcomes in patients with poor response to therapy [7]. the use of exogenous ifn-γ in adjuvant therapy, especially in the case of diseases caused by microorganisms resistant to antibiotics, can have a significant effect on the course of the disease and accelerate the patient's recovery, contributing to earlier abacillation of sputum. admet & dmpk 10(1) (2022) 63-73 therapeutic potential of interferon-gamma in tuberculosis doi: http://dx.doi.org/10.5599/admet.1078 69 assessment of the economic feasibility of using interferon-gamma in the treatment of tuberculosis (experience of the russian federation) as a socially significant disease, tb poses a significant social and economic burden on the healthcare system around the world [51]. the economic burden of tb is conditioned by several components: direct economic losses associated with medical costs and indirect economic losses of the country's gross domestic product due to the loss of “human capital”. the costs of drug therapy account for more than half of the direct medical costs of tb treatment [52]. in this regard, when adding new drugs to tb treatment regimens, not only clinical indicators of the effectiveness of therapy should be taken into account, but also their impact on the health economics. assessment of health technologies when deciding on the inclusion of medicines in the drug supply system is a mandatory component of health systems in all countries of the world, regardless of economic development, morbidity and mortality. in the russian study published in 2018 [53] on the assessment of the cost-effectiveness for ifn-γ, a positive impact on the budget from the use of ifn-γ as a part of complex tb therapy was demonstrated due to reducing the duration of hospitalization. as a result of the behavioural pharmacoeconomic analysis, it was shown that ifn-γ contributed to an increase of the duration of quality life by 2.1 qaly, and its use will be accompanied by savings in the budget of the russian federation in the amount of up to 27 % of funds for the treatment of this group of patients [53]. results and discussion analysis of data from preclinical and clinical studies shows that the inflammatory cytokine ifn-γ occupies an important place in the pathogenesis of bacterial infections, including tb. the significance and potential of the use of interferons are discussed in all areas of phthisiology: diagnosis, therapy, and prevention of tb [54]. for instance, the main alternative to tuberculin skin test widely used for diagnostic purposes is ifn-γ release assays (igra) test. it measures the ex vivo production of ifn-γ from t-cells stimulated with the tb-specific antigens. the test doesn’t have cross-reactivity to bcg and nontuberculous mycobacteria showing higher specificity keeping the same level of sensitivity [55-57]. preclinical data confirm the enhancement of the cellular component of immunity when using ifn-γ in combination with the recombinant bcg vaccine (bacillus calmette guérin (bcg)) [58]. the use of ifn-γ as a means of preventing infection, as well as the transition of the latent phase of tb to the active form, requires the study of this agent in clinical trials of vaccines and other therapeutic drugs [20]. the greatest amount of evidence and data today is obtained in the direction of the use of ifn-γ as a therapeutic option in the complex therapy of pulmonary tb. it has been shown that ifn-γ plays a significant role in the pathogenesis of tb [15,59], is a marker of type 1 th-helper lymphocytes [55], and is considered a key immunomodulatory agent for the maintenance adjuvant therapy of tb patients. many factors play a role in the infectious process and outcomes of tb, but the most important of them is the activity of the body's immune system and its ability to resist infection. the human immune system is able to control it. weakness of the immune response is the main condition for the development of the disease and subsequently for the prolonged course of the infectious process against the background of antibiotic therapy. in patients with immunodeficiency, the treatment of the disease is always difficult, even with maximum doses of drugs as part of complex antibiotic therapy [6]. in conditions of bacterial resistance to antibiotics, the role of the immune response in the fight against infection and the stress on the immune system increases dramatically. s.h. kaufman et al., in their review published in 2014 in the lancet [60], focused on host-directed therapies defining such strategy “especially beneficial for patients with multidrughttp://dx.doi.org/10.5599/admet.1078 svetlana a. berns et al. admet & dmpk 10(1) (2022) 63-73 70 resistant tuberculosis”. the use of traditional anti-tuberculosis drugs only is no more a sufficient tactic to achieve the target [60]. conclusions understanding the pathogenesis of the development of the infectious process and the role of ifn-γ in the activation of the immune system in tb, made it possible to identify ifn-γ as a promising therapeutic option for improving treatment results and overcoming the development of bacterial resistance. the clinical efficacy of ifn-γ against tb has been studied in clinical trials for over 20 years. the clinical experience and medical application of ifn-γ accumulated in the course of long research allow us to recommend it as a safe and effective therapeutic agent in a complex, comprehensive tb treatment, not only from the standpoint of evidence-based medicine but also from the point of view of health economics. for the full disclosure of the therapeutic potential of ifn-γ in overcoming multidrug resistance of tb, further study of its effectiveness in controlled clinical trials is necessary. conflict of interest: the authors declare that they have no competing interests. references [1] world health organization. who global strategy for containment of antimicrobial resistance. geneva, 2001. who/cds/csr/drs/2001.2. https://www.who.int/drugresistance/who%20global%20strategy%20-%20executive%20summary%20-%20english%20version.pdf?ua=1. [2] j.p. cegielski, t. dalton, m. yagui m, et al. extensive drug resistance acquired during treatment of multidrug-resistant tuberculosis. clinical infectious diseases 59(8) (2014) 1049-1063. https://doi.org/10.1093/cid/ciu572. [3] a.v. dehnich. antibiotic resistance: problems and solutions. materials of the ix national congress on cystic fibrosis "cystic fibrosis in children and adults 2009», cjsc « reklamno-izdatel'skaja gruppa megapro», 2009, стр. 34-35. [4] johns hopkins university. corona virus resource center. 2020. available from: https://coronavirus.jhu.edu/map.html. [5] who, global tuberculosis report, 2020, https://apps.who.int/iris/bitstream/handle/10665/336069/9789240013131-eng.pdf. [6] a.k. azad, w. sadee, and l. s. schlesinger. innate immune gene polymorphisms in tuberculosis. infection and immunity 80(10) (2012) 3343–3359. https://doi.org/10.1128/iai.00443-12. [7] t.a. khan, h. mazhar, sh. saleha, h.n. tipu, n. muhammad, m.n. abbas. interferon-gamma improves macrophages function against m. tuberculosis in multidrug-resistant tuberculosis patients. chemotherapy research and practice (2016) 7295390. https://doi.org/10.1155/2016/7295390. [8] c.a. jr. janeway, p. travers, m. walport, et al. immunobiology: the immune system in health and disease. 5th edition. new york: garland science (2001). [9] r.j. meshkova. immunization guidelines for physicians. smolensk state medical academy (1998). [10] s. sharma, m. bose. role of cytokines in immune response to pulmonary tuberculosis. asian pac j allergy immunol. 19(3) (2001) 213-219. [11] j. parkin, b. cohen. an overview of the immune system. lancet 357(9270) (2001) 1777–1789. https://doi.org/10.1016/s0140-6736(00)04904-7. [12] j. piehler, c. thomas, k.c. garcia, g. schreiber. structural and dynamic determinants of type i interferon receptor assembly and their functional interpretation. immunological reviews 250 (2012) 317–334. https://doi.org/10.1111/imr.12001. [13] j. ye, j.r. ortaldo, k. conlon, r. winkler-pickett, h.a. young. cellular and molecular mechanisms of ifn-gamma production induced by il-2 and il-12 in a human nk cell line. journal of leukocyte biology 58(2) (1995) 225–233. https://doi.org/10.1002/jlb.58.2.225. https://www.who.int/drugresistance/who%20global%20strategy%20-%20executive%20summary%20-%20english%20version.pdf?ua=1 https://www.who.int/drugresistance/who%20global%20strategy%20-%20executive%20summary%20-%20english%20version.pdf?ua=1 https://doi.org/10.1093/cid/ciu572 https://doi.org/10.1093/cid/ciu572 https://coronavirus.jhu.edu/map.html https://coronavirus.jhu.edu/map.html https://apps.who.int/iris/bitstream/handle/10665/336069/9789240013131-eng.pdf https://apps.who.int/iris/bitstream/handle/10665/336069/9789240013131-eng.pdf https://doi.org/10.1128/iai.00443-12 https://doi.org/10.1155/2016/7295390 https://www.ncbi.nlm.nih.gov/pubmed/?term=sharma%20s%5bauthor%5d&cauthor=true&cauthor_uid=11826917 https://www.ncbi.nlm.nih.gov/pubmed/?term=bose%20m%5bauthor%5d&cauthor=true&cauthor_uid=11826917 https://www.ncbi.nlm.nih.gov/pubmed/11826917 https://www.ncbi.nlm.nih.gov/pubmed/11826917 https://doi.org/10.1016/s0140-6736(00)04904-7 https://doi.org/10.1111/imr.12001 https://doi.org/10.1002/jlb.58.2.225 admet & dmpk 10(1) (2022) 63-73 therapeutic potential of interferon-gamma in tuberculosis doi: http://dx.doi.org/10.5599/admet.1078 71 [14] h. bierne, l. travier, t. mahlakõiv, et al. activation of type iii interferon genes by pathogenic bacteria in infected epithelial cells and mouse placenta. plos one 7(6) (2012) e39080. https://doi.org/10.1371/journal.pone.0039080. [15] l. desvignes, a.j. wolf, j.d. ernst. dynamic roles of type i and type ii interferons in early infection with mycobacterium tuberculosis. journal of immunology. 188(12) (2012) 6205–6215. https://doi.org/10.4049/jimmunol.1200255. [16] s.v. jamwal, p. mehrotra, a. singh, z. siddiqui, a. basu, k.v. rao. mycobacterial escape from macrophage phagosomes to the cytoplasm represents an alternate adaptation mechanism. scientific reports 6 (2016) 23089. https://doi.org/10.1038/srep23089. [17] t.n. ottenhoff, r.h. dass, n. yang, et al. genome-wide expression profiling identifies type 1 interferon response pathways in active tuberculosis. plos one 7(9) (2012) e45839. https://doi.org/10.1371/journal.pone.0045839 [18] m. travar, m. petkovic, a. verhaz. type i, ii, and iii interferons: regulating immunity to mycobacterium tuberculosis infection. archivum immunologiae et therapiae experimentalis (warsz) 64(1) (2016) 19–31. doi: https://doi.org/10.1007/s00005-015-0365-7. [19] m.a. huaman, d. henson, p.l. rondan, et al. latent tuberculosis infection is associated with increased unstimulated levels of interferon-gamma in lima, peru. plos one 13(9) (2018) e0202191. https://doi.org/10.1371/journal.pone.0202191. [20] e. petruccioli, t.j. scriba, l. petrone, et al. correlates of tuberculosis risk: predictive biomarkers for progression to active tuberculosis. the european respiratory journal 48(6) (2016) 1751-1763. https://doi.org/10.1183/13993003.01012-2016. [21] v.n. zimina, i.a. vasilieva, a.v. kravchenko, f.a. batyrov. generalized tuberculosis in hiv-infected patients with aids. journal of the international aids society 13(4) (2010) p195. https://doi.org/10.1186/1758-2652-13-s4-p195. [22] e. van de vosse, m.a. hoeve, t.n. ottenhoff. human genetics of intracellular infectious diseases: molecular and cellular immunity against mycobacteria and salmonellae. lancet infect. dis. 4 (2004) 739–749. https://doi.org/10.1016/s1473-3099(04)01203-4. [23] l.c. platanias. mechanisms of type-iand type-ii-interferon-mediated signalling. nature reviews. immunology 5(5) (2005) 375–386. https://doi.org/10.1038/nri1604. [24] b.h. salt, j.e. niemela, r. pandey, e.p. hanson, r.p. deering, r. quinones, a. jain, j.s. orange, e.w. gelfand. ikbkg (nuclear factor-kappa b essential modulator) mutation can be associated with opportunistic infection without impairing toll-like receptor function. the journal of allergy and clinical immunology 121 (2008) 976–982. https://doi.org/10.1016/j.jaci.2007.11.014. [25] j. bustamante, a.a. arias, g. vogt, c. picard, l.b. galicia, et al. germline cybb mutations that selectively affect macrophages in kindreds with x-linked predisposition to tuberculous mycobacterial disease. nature immunology 12 (2011) 213–221. https://doi.org/10.1038/ni.1992. [26] j.d. ernst. the immunological life cycle of tuberculosis. nature reviews immunology 12 (2012) 581– 591. https://doi.org/10.1038/nri3259. [27] o. filipe-santos, j. bustamante, a. chapgier, g. vogt, l. de beaucoudrey, j. feinberg, e. jouanguy, s. boisson-dupuis, c. fieschi, c. picard, j.l. casanova. inborn errors of il-12/23and ifn-γ-mediated immunity: molecular, cellular, and clinical features. seminars in immunology 18 (2006) 347–361. https://doi.org/10.1016/j.smim.2006.07.010. [28] s. al-muhsen, j.l. casanova. the genetic heterogeneity of mendelian susceptibility to mycobacterial diseases. journal of allergy and clinical immunology 122 (2008) 1043–1051. https://doi.org/10.1016/j.jaci.2008.10.037. [29] s.d. rosenzweig, s.e. dorman, g. uzel, s. shaw, a. scurlock, m.r. brown, r.h. buckley, s.m. holland. a novel mutation in ifn-γ receptor 2 with dominant negative activity: biological consequences of homozygous and heterozygous states. journal of immunology 173 (2004) 4000–4008. https://doi.org/10.4049/jimmunol.173.6.4000. http://dx.doi.org/10.5599/admet.1078 https://doi.org/10.1371/journal.pone.0039080 https://doi.org/10.1371/journal.pone.0039080 https://doi.org/10.4049/jimmunol.1200255 https://dx.doi.org/10.1038%2fsrep23089 https://doi.org/10.1371/journal.pone.0045839 https://doi.org/10.1007/s00005-015-0365-7 https://doi.org/10.1371/journal.pone.0202191 https://doi.org/10.1183/13993003.01012-2016 https://doi.org/10.1186/1758-2652-13-s4-p195 https://doi.org/10.1016/s1473-3099(04)01203-4 https://doi.org/10.1038/nri1604 https://doi.org/10.1016/j.jaci.2007.11.014 https://doi.org/10.1038/ni.1992 https://doi.org/10.1038/ni.1992 https://doi.org/10.1038/nri3259 https://doi.org/10.1016/j.smim.2006.07.010 https://doi.org/10.1016/j.jaci.2008.10.037 https://doi.org/10.1016/j.jaci.2008.10.037 https://doi.org/10.4049/jimmunol.173.6.4000 https://doi.org/10.4049/jimmunol.173.6.4000 svetlana a. berns et al. admet & dmpk 10(1) (2022) 63-73 72 [30] j.f. emile, n. patey, f. altare, s. lamhamedi, e. jouanguy, f. boman, j. quillard, m. lecomte-houcke, o. verola, j.f. mousnier, f. dijoud, s. blanche, a. fischer, n. brousse, j.l. casanova. correlation of granuloma structure with clinical outcome defines two types of idiopathic disseminated bcg infection. j pathol. 181 (1997) 25–30. https://doi.org/10.1002/(sici)1096-9896(199701)181:1<25::aid-path747>3.0.co;2-z . [31] s.n. kozlov, r.s. kozlov. modern antimicrobial chemotherapy. a guide for doctors. 3rd edition revised and enlarged. medicinskoe informacionnoe agentstvo (2017) 400 с. [32] t. eddens, j.k. kolls. host defenses against bacterial lower respiratory tract infection. curr. opin. immunol. 24(4) (2012) 424–430. https://doi.org/10.1016/j.coi.2012.07.005. [33] c.v. harding, w.h. boom. regulation of antigen presentation by mycobacterium tuberculosis: a role for toll-like receptors. nat. rev. microbiol. 8 (2010) 296–307. https://doi.org/10.1038/nrmicro2321. [34] a. razaghi, l. owens, k. heimann. review of the recombinant human interferon gamma as an immunotherapeutic: impacts of production platforms and glycosylation. journal of biotechnology 240 (2016) 48–60. https://doi.org/10.1016/j.jbiotec.2016.10.022. [35] f.c. fang. antimicrobial reactive oxygen and nitrogen species: concepts and controversies. nat. rev. microbiol. 2(10) (2004) 820-32. https://doi.org/10.1038/nrmicro1004. [36] e. pashkov, a. karaulov, s. bykov, m. korn, a. bykov, a. vorob'ev, v. zverev. atlas of medical microbiology, virology and immunology. m.: medicinskoe informacionnoe agentstvo (2008) 272 с. [37] r.s. orlov, a.d. nozdrachev. normal physiology: textbook. second edition, revised and enlarged m. gjeotar-media (2010) 832 p. [38] e.l. lousberg, c.k. fraser, m.g. tovey, k.r. diener, j.d. hayball. type i interferons mediate the innate cytokine response to recombinant fowlpox virus but not the induction of plasmacytoid dendritic celldependent adaptive immunity. journal of virology 84(13) (2010) 6549–6563. https://doi.org/10.1128/jvi.02618-09. [39] p.o. pinheiro, e.b. oliveira, g. santos, g.m. sperandio da silva, b.j. andrade silva, r.m.b. teles, a. milagres, e.n. sarno, m.p. dalcolmo, e.p. sampaio. different immunosuppressive mechanisms in multi-drug-resistant tuberculosis and non-tuberculous mycobacteria patients. clinical and experimental immunology 171(2) (2013) 210–219. https://doi.org/10.1111/cei.12007. [40] a. fortes, k. pereira, p.r.z. antas, c.l.m.c. franken, m. dalcolmo, m.m. ribeiro-carvalho, k.s. cunha, a. geluk, a. kritski, a. kolk, p. klatser, e.n. sarno, t.h.m. ottenhoff, e.p. sampaio. detection of in vitro interferon-γ and serum tumour necrosis factor-α in multidrug-resistant tuberculosis patients. clinical and experimental immunology 141(3) (2005) 541–548. https://doi.org/10.1111/j.13652249.2005.02872.x. [41] b.m. brooks, c.a. hart, j.w. coleman. 2005. differential effects of beta-lactams on human ifngamma activity. the journal of antimicrobial chemotherapy 56(6) (2005) 1122-1125. https://doi.org/10.1093/jac/dki373. [42] https://clinicaltrials.gov (date accessed 10.06.2021). [43] x.f. gao, z.w. yang, j. li. adjunctive therapy with interferon-gamma for the treatment of pulmonary tuberculosis: a systematic review. international journal of infectious diseases 15 (2011) e594–e600. https://doi.org/10.1016/j.ijid.2011.05.002. [44] r. condos, w. n. rom, and n. w. schluger. treatment of multidrug-resistant pulmonary tuberculosis with interferon-gamma via aerosol. lancet 349(9064) (1997) 1513–1515. https://doi.org/10.1016/s0140-6736(96)12273-x [45] r. suarez-mendez, i. garcía-garcía, n. fernández-olivera et al. adjuvant interferon gamma in patients with drug resistant pulmonary tuberculosis: a pilot study. bmc infectious diseases 4 (2004) 44. https://doi.org/10.1186/1471-2334-4-44. [46] s. k. park, s. cho, i. h. lee et al. subcutaneously administered interferon-gamma for the treatment of multidrugresistant pulmonary tuberculosis. international journal of infectious diseases 11(5) (2007) 434–440. https://doi.org/10.1016/j.ijid.2006.12.004. https://doi.org/10.1002/(sici)1096-9896(199701)181:1%3c25::aid-path747%3e3.0.co;2-z https://doi.org/10.1002/(sici)1096-9896(199701)181:1%3c25::aid-path747%3e3.0.co;2-z https://doi.org/10.1016/j.coi.2012.07.005 https://doi.org/10.1038/nrmicro2321 https://doi.org/10.1016/j.jbiotec.2016.10.022 https://doi.org/10.1038/nrmicro1004 https://doi.org/10.1128/jvi.02618-09 https://doi.org/10.1128/jvi.02618-09 https://doi.org/10.1111/cei.12007 https://doi.org/10.1111/j.1365-2249.2005.02872.x https://doi.org/10.1111/j.1365-2249.2005.02872.x https://doi.org/10.1093/jac/dki373 https://clinicaltrials.gov/ https://doi.org/10.1016/j.ijid.2011.05.002 https://doi.org/10.1016/s0140-6736(96)12273-x https://doi.org/10.1016/s0140-6736(96)12273-x https://doi.org/10.1186/1471-2334-4-44 https://doi.org/10.1016/j.ijid.2006.12.004 admet & dmpk 10(1) (2022) 63-73 therapeutic potential of interferon-gamma in tuberculosis doi: http://dx.doi.org/10.5599/admet.1078 73 [47] i. raad, r. hachem, n. leeds, r. sawaya, z. salem, and s. atweh. use of adjunctive treatment with interferon-gamma in an immunocompromised patient who had refractory multidrug-resistant tuberculosis of the brain. clinical infectious diseases 22(3) 1996 572–574. https://doi.org/10.1093/clinids/22.3.572. [48] a.a. maslennikov, n.i. obolonkova. efficiency of ingaron in the treatment of patients with destructive pulmonary bacteriologicaly proven tuberculosis. research result. medicine and pharmacy series 1(7) 2016. https://doi.org/10.18413/2313-8955-2016-2-1-10-16. [49] r. dawson, r. condos, d. tse, et al. immunomodulation with recombinant interferon-gamma1b in pulmonary tuberculosis. plos one 4(9) (2009) e6984. https://doi.org/10.1371/journal.pone.0006984. [50] m.s. pisarenko. pecularities of ifn-g secretion in drug-resistant pulmonary tuberculosis. fundamental research 9 (2013) 444-447. [51] i.a. vasilyeva, e.m. belilovsky, s.e. borisov, s.a. sterlikov. incidence, mortality and prevalence as indicators of tuberculosis burden in who regions, countries of the world and the russian federation. part 2. tuberculosis mortality. tuberculosis and lung diseases 95(7) (2017) 8-16. https://doi.org/10.21292/2075-1230-2017-95-6-9-21. [52] t. tanimura, e. jaramillo, d. weil, m. raviglione, k. lönnroth. financial burden for tuberculosis patients in lowand middle-income countries: a systematic review. the european respiratory journal 43(6) (2014) 1763-1775. https://doi.org/10.1183/09031936.00193413. [53] s.l. plavinskiy, p.i. shabalkin. evaluation of the clinical and economic effectiveness of interferongamma in the treatment of tuberculosis. epidemiology and infectious diseases. russian journal 22(6) (2017) 276-280. https://doi.org/10.17816/eid40981. [54] k.l. chin, f.z. anis, m.e. sarmiento, m.n. norazmi, a. acosta. role of interferons in the development of diagnostics, vaccines, and therapy for tuberculosis. journal of immunology research 2017 (2017) 5212910. https://doi.org/10.1155/2017/5212910. [55] s. anderson, v. ansdell, p.m. arguin, et al.. the travel and tropical medicine manual (fifth edition). elsevier (2017) 646p. https://doi.org/10.1016/b978-0-323-37506-1.00061-1. [56] p. c. hopewell, m. kato-maeda, j. d. ernst. murray and nadel's textbook of respiratory medicine (sixth edition). elsevier (2016). https://doi.org/10.1016/b978-1-4557-3383-5.00035-x. [57] t. togun, b. kampmann, m. pai. diagnosis of childhood tuberculosis, reference module in biomedical sciences. elsevier (2017). https://doi.org/10.1016/b978-0-12-801238-3.64157-0. [58] y. xu, b. zhu, q. wang, et al. recombinant bcg coexpressing ag85b, esat-6 and mouse-ifn-gamma confers effective protection against mycobacterium tuberculosis in c57bl/6 mice. fems immunology and medical microbiology 51(3) (2007) 480–487. https://doi.org/10.1111/j.1574-695x.2007.00322.x. [59] g. c. k. w. koh, m.f. schreiber, r. bautista, et al. host responses to melioidosis and tuberculosis are both dominated by interferon-mediated signaling. plos one 8(1) (2013) e54961. https://doi.org/10.1371/journal.pone.0054961. [60] s.h. kaufmann, c. lange, m. rao, et al. progress in tuberculosis vaccine development and hostdirected therapies – a state of the art review. the lancet. respiratory medicine 2(4) (2014) 301-320. https://doi.org/10.1016/s2213-2600(14)70033-5. ©2022 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://dx.doi.org/10.5599/admet.1078 https://doi.org/10.1093/clinids/22.3.572 https://doi.org/10.1093/clinids/22.3.572 https://doi.org/10.18413/2313-8955-2016-2-1-10-16 https://doi.org/10.1371/journal.pone.0006984 https://doi.org/10.21292/2075-1230-2017-95-6-9-21 https://doi.org/10.1183/09031936.00193413 http://www.medlit.ru/journalsview/infections/main/ http://www.medlit.ru/journalsview/infections/view/journal/en/2017/issue-6/ http://www.medlit.ru/journalsview/infections/view/journal/en/2017/issue-6/ https://doi.org/10.17816/eid40981 https://doi.org/10.1155/2017/5212910 https://www.sciencedirect.com/book/9780323375061/the-travel-and-tropical-medicine-manual https://doi.org/10.1016/b978-0-323-37506-1.00061-1 https://doi.org/10.1016/b978-1-4557-3383-5.00035-x https://doi.org/10.1016/b978-0-12-801238-3.64157-0 https://doi.org/10.1111/j.1574-695x.2007.00322.x https://doi.org/10.1371/journal.pone.0054961 https://doi.org/10.1371/journal.pone.0054961 https://doi.org/10.1016/s2213-2600(14)70033-5 http://creativecommons.org/licenses/by/3.0/ cδlog kwiam: can we afford estimation of small molecules’ blood-brain barrier passage based upon in silico phospholipophilicity? doi: http://dx.doi.org/10.5599/admet.1034 267 admet & dmpk 9(4) (2021) 267-281; doi: https://doi.org/10.5599/admet.1034 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper cδlog kw iam : can we afford estimation of small molecules’ bloodbrain barrier passage based upon in silico phospholipophilicity? lucia grumetto 1 , giacomo russo 2 * 1 pharm-analysis & bio-pharm laboratory, department of pharmacy, school of medicine and surgery, university of naples federico ii, via d. montesano, 49, i-80131, naples, italy. 2 school of applied sciences, sighthill campus, edinburgh napier university, 9 sighthill ct, eh11 4bn edinburgh, united kingdom. *corresponding author: e-mail: g.russo@napier.ac.uk; tel.: +44 (0) 131 455 3464; fax: +44 (0) 131 455 3555 received: june 27, 2021; revised: november 27, 2021; published: december 15, 2021 abstract 56 compounds, whose log bb values were known from the scientific literature, were considered and their phospholipophilicity values were calculated in silico. these values, along with either experimentally determined or calculated lipophilicity values, were used to extract cδ/δ’log kw iam parameters. cδ/δ’log kw iam values were found inversely related to data of blood-brain barrier passage, especially in the < -0.20 log bb range and on the iam.pc.dd2 phase (r 2 = 0.79). in multiple linear regression, satisfactory statistic models (r 2 (n-1) = 0.76), based on c/’log kw iam.mg along with other in silico calculated descriptors, were achieved. this method brings the potential to be applied, along with other methodologies, to filter out solutes whose bbb permeation is foreseen to be substandard, thus allowing pharmaceutical companies/research institutes to focus on candidates that are more likely to concentrate in the brain. ©2021 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords immobilized artificial membrane; biochromatography; blood brain barrier; brain targeting; phospholipophilicity. introduction combinatorial chemistry involves the generation of a large array of structurally diverse compounds, i.e., a chemical library, through systematic, repetitive and covalent linkage of various “building blocks” [1]. this technique can be exploited in parallel, delivering hundreds, if not thousands, of molecules of pharmaceutical interest in a handful of hours. while the organic synthesis throughput has expanded so noticeably in recent years, screening methodologies are still lagging behind, instead [2]. indeed, most of the testing still requires animal models that have the undeniable advantage of mirroring more closely the complexity of human beings than cells. however, animal models are facing criticism from the public since they often require the sacrifice of vertebrates [3] and heavily impact the environment due to the huge number of carcasses to dispose of. the assessment of the ability of a drug to cross the biological membranes in the early stages of its http://dx.doi.org/10.5599/admet.1034 https://doi.org/10.5599/admet.1034 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:g.russo@napier.ac.uk http://creativecommons.org/licenses/by/4.0/ grumetto and russo admet & dmpk 9(4) (2021) 267-281 268 development plays a pivotal role in pharmaceutical industrial research. notably, the development of drugs acting toward the central nervous system (cns) has poorer success rates and requires longer times than non-cns drugs [4]. this occurs due to the complexity of the blood-brain barrier (bbb). in fact, in a healthy brain, the bbb plays a crucial role in protecting normal brain functions from potentially harmful compounds occurring in the bloodstream [5]. strategies for brain drug delivery have developed in the last decades, and various techniques are available to study the bbb's role in drug uptake. these include in vivo, in vitro [6] and in situ techniques [7]. separation science offers valuable alternatives to animal testing that can provide effectiveness in the drug discovery/drug development pipeline as biomimetic liquid chromatography [8-11], performed employing stationary phases emulating biological components or using mobile phase ingredients simulating physiological environments. a consistent branch of this is represented by liquid chromatography (lc) conducted on stationary phases based on immobilized artificial membranes (iam). iam phases are based on membrane phospholipid analogs covalently bound to aminopropyl silica [7,12,13]. some of these phases are available commercially as iam.pc.mg and iam.pc.dd2. both these support phosphatidylcholine analogs (pc), but they differ from each other in the end-capping of the free aminopropyl moieties, which is performed with methyl glycolate (mg) or with c3 or c10 anhydrides (dd2). in recent years, we parameterized the excess of the polar/electrostatic interactions occurring between drugs and biological membranes as δlog kw iam [14-19]. δlog kw iam is obtained by combining n-octanol/water lipophilicity with phospholipophilicity, i.e., the affinity of the compound for the iam phases measured as a retention factor extrapolated at 100 % of aqueous phase (kw iam ) [20]. this represents the difference between the logarithm of the chromatographic retention factor (log kw iam ) measured for each analyte, i.e., the experimentally determined phospholipophilicity, and the value expected for a neutral isolipophilic molecule that is estimated by correlative equations. δlog kw iam values were inversely related to the drug passage of complex biological barriers, such as the bbb and the intestinal wall [13,17,18]. the increasing need for high-throughput drug discovery methods has provided several in silico models of bbb permeation based on in vivo log bb values [21,22]. log bb is generally measured on murine models and is still nowadays considered as a solid indication for bbb delivery [23]. log bb is defined as (eq. 1): log bb = log 𝐶brain 𝐶blood (1) in which cbrain is the concentration that the analyte realizes in the brain tissues, and cblood is the concentration that it achieves in the blood. the in silico models bring the advantages of being much faster to perform and applicable to molecules that are not yet synthesized and/or not easily detectable. back in 2017, we developed some statistical models to predict the phospholipophilicity of small molecules based on more than 200 individual measurements performed in our laboratories. this also materialized in an online service, namely log kw iam.mg/dd2 calculator, offering the opportunity to predict the phospholipophilicity of all compounds included in pubchem collection as log kw iam on both mg and dd2 chromatographic columns [24]. in the present study, we aim at applying these statistical models to calculate the phospholipophilicity of a dataset of compounds whose log bb is known from the scientific literature and from there to estimate δlog kw iam , based either on experimentally determined or calculated lipophilicity values. our goal is to evaluate whether these parameters calculated in silico, therefore called from here on cδlog kw iam , can offer effectiveness in screening libraries of compounds for their potential to reach the brain. if so, we will look at ways to implement these procedures in the drug discovery/development industrial programs. admet & dmpk 9(4) (2021) 267-281 bbb passage based on in silico phospholipophilicity doi: http://dx.doi.org/10.5599/admet.1034 269 experimental data collection experimental lipophilicity values were collected from the scientific literature. specifically, all the log p values were taken from pubchem but those of nevirapine and thioxolone, which were taken from the literature [25]. calculated log p values were obtained by either alogps [26] or by marvinsketch [27]. for acidic compounds, whose ’log kw iam but not their log kw iam were previously found related to log bb, log d 7.4 values were again calculated by marvinsketch software [27]. log bb values were taken from the literature [28]. c/’log kw iam values calculation c/’log kw iam values were calculated from phospholipophilicity values estimated in silico according to a procedure we developed in 2017 [24]. in brief, the best relationships were found to be: log kw iam.mg = -0.1405 (0.1282) + 0.4401(0.0297)milogp + 0.0536 (0.0057)heavy atoms 0.0833 (0.0201)hlbm 0.0435(0.0144)rotatable bonds (2) n=204 r 2 = 0.81 q 2 = 0.80 se = 0.438 f4,199 = 213.92 p<1.0 10 -8 pc = 39.403 and log kw iam.dd2 = -2.3989 (0.2812) + 0.4936(0.0379)milogp + 0.4354 (0.0470)volume diameter 0.0640 (0.0226)hlbpsa 0.0497(0.0173)rotatable bonds (3) n=160 r 2 = 0.85 q 2 = 0.84 se = 0.459 f4,155 = 212.94 p<1.0 10 -8 pc = 33.974 a detailed explanation of the main descriptors, along with relevant references, is reported in supporting information (table s1). in these equations, n is the number of data considered to derive the regression equation, r 2 is the square of the correlation coefficient, se is the standard error of the estimate, f (the subscripts are the degrees of freedom and the number of variables) is the fisher statistic of the regression, p is the observed significance level, i.e., the probability of obtaining a result equal to or “more extreme” than what was observed, and pc is the amemiya predictive criterion of the regression. the hydrophilic-lipophilic balance (hlb) can be taken into account by the methods by griffin [29] (hlbg), davies [30] (hlbd), and taking into account the steric effects (hlbpsa), not considered by the two approaches. hlbpsa is defined as follows: hlbpsa=20⋅psa/surface where psa is the polar surface area and surface is the total surface. hlb (hlbm) is the mean resulting from the values by all three methods. milogp is the octanol-water partition coefficient predicted by the online program for the calculation of molecular properties and bioactivity prediction [31]. the calculations were made completely automated and easily accessible to anyone via a user-friend tool to predict log kw iam.mg and log kw iam.dd2 , a web service and a set of scripts for vega zz program [24]. this is available at https://www.ddl.unimi.it/vegaol/logkwiam.htm and offers a calculation of log k iam.mg/dd2 of any molecule included in the pubchem collection as implemented in the script version. log kw iam values were calculated as the difference between the log kw iam computed from equations (2) and (3) and the log kw iam expected for neutral isolipophilic molecules. indeed, as reported in our previous studies [14,15], iam retention data on both iam phases relate unambiguously with log p values of http://dx.doi.org/10.5599/admet.1034 https://www.ddl.unimi.it/vegaol/logkwiam.htm grumetto and russo admet & dmpk 9(4) (2021) 267-281 270 structurally non-related neutral compounds, in the log p range 1.0–4.8. these relationships are expressed by the following equations: log kw iam.mg = 0.792 (±0.038) log p 0.732(±0.105) (4) n = 36 r 2 = 0.926 s = 0.248 f1, 34 = 422.40 f1, 34 α,0.001 = 12.90 log kw iam.dd2 = 0.934 (±0.038) log p 0.883(±0.104) (5) n = 36 r 2 = 0.946 s = 0.246 f1, 34 = 595.74 f1, 34 α,0.001 = 12.90 for acidic compounds, analogously to what was reported in our previous study [15], log d 7.4 rather than log p n was used for the computation of delta values in equations (4) and (5). their values were therefore named ’log kw iam values to avoid any ambiguity. molecular modeling an ample array (> 1,600) of physico-chemical descriptors, subdivided into 20 logical blocks (atom type, functional group, fragment counts, topological and geometrical descriptors), were calculated by the web service e-dragon 1.0 [32]. in brief, the molecules were input as smiles code in a text document and converted by the integrated applet corina in 3d before all the indices were computed. the quantitative structure-property relationship (qspr) models were obtained by the automatic stepwise approach implemented in the “automatic linear regression” script of vega zz software [33], calculating regression models, including from one to five independent variables. the predictive strength of the best equation was evaluated by leave-one-out (loo) cross-validation. the regression coefficients were calculated to evaluate the set in terms of the standard deviation of errors (se), regression coefficients (r 2 is the square of the correlation coefficient, q 2 is the square of the correlation coefficient after cross-validation), intercept, fisher statistic for the regression (f), probability (p) and amemiya prediction criterion (pc). descriptors with too low regression coefficient (r 2 < 0.1) were excluded, and collinear descriptors were detected by evaluating the variance inflation factor (vif) whose threshold value was set to 5. data handling data were input in a spreadsheet and data points were plotted from microsoft excel, part of the microsoft office 365 suite of programs. results and discussion c/’log kw iam : simple linear regression in our previous studies [14-19, 34], /’log kw iam values were found inversely related to the passage of complex biological barriers, such as the bbb and the intestinal wall. the calculation of /’log kw iam parameters are based on two physico-chemical properties, i.e., n-octanol/water lipophilicity either of the neutral species (giving log kw iam ) or of the mixture of the species at the physiological ph, i.e., 7.4 (giving ’log kw iam ) and the affinity of the compound for iam phases. indeed, in our previous studies [14-16], we verified that for acidic compounds, significant relationships vs. log bb data could only be obtained when delta parameters were calculated by using the lipophilicity of the mixture of the species present in solution at the experimental ph, i.e., log d 7.4 , rather than that of the neutral species, i.e., log p n . for neutrals, bases and ampholytes, delta parameters were estimated by using log p n values instead. for consistency, we extended the same approach to delta values surrogated in silico (c/’log kw iam ). however, while there are plenty of tools available to surrogate log p values [35], to the best of our knowledge, the in silico platform we developed is the only service that predicts phospholipophilicity. table admet & dmpk 9(4) (2021) 267-281 bbb passage based on in silico phospholipophilicity doi: http://dx.doi.org/10.5599/admet.1034 271 1 lists names, chemical nature (a= acid, b=basic, bb= bibasic, n= neutral), calculated log kw iam.mg and log kw iam.dd2 , exp log p n , clog p n and calculated log d 7.4 (for acids only) values for the dataset considered. c/’log kw iam values are reported in table 2 along with the experimental log bb values. table 1. names, chemical nature (a= acid, b=basic, bb= bibasic, n= neutral), calculated log kw iam.mg and log kw iam.dd2 , exp log p n , clog p n values for the dataset considered. molecule nature clog kw iam.mg clog kw iam.dd2 exp log p n [36] clog p n (1) [26] clog p n (2) [37] clog d 7.4 [37] 1,1,1-trichloroethane n 1.063 1.247 2.49 2.45 2.04 1,2-dimethylbenzene n 1.400 1.635 3.12 3.16 2.98 1,4-dimethylbenzene n 1.421 1.599 3.15 3.15 2.98 1,7-dimethylxanthine a -0.073 -0.001 -0.22 -0.63 0.09 0.09 1-chloro-2,2,2trifluoroethane n 0.811 0.774 1.82 1.86 1-hydroxymidazolam n 1.839 2.043 3.09 2.9 2,2-dimethylbutane n 1.283 1.578 3.82 3.74 2.85 2-methylpentane n 1.388 1.706 3.21 3.66 2.82 3-methylhexane n 1.561 1.949 4.18 3.21 3-methylpentane n 1.292 1.599 3.60 3.98 2.82 4-hydroxymidazolam n 1.950 2.191 3.05 3.35 acetaminophen n 0.184 0.302 0.91 0.51 1.09 acetone n -0.247 -0.359 -0.24 -0.29 0.38 aminopyrine n 1.045 1.349 1.00 0.94 1.60 amobarbital a 0.899 1.181 2.07 1.87 1.86 1.60 antipyrine n 0.901 1.139 0.56 1.18 1.61 bretazenil n 2.103 2.447 3.05 2.29 cyclohexane n 1.476 1.671 3.44 2.38 cyclopropane n 0.284 0.105 1.72 1.56 1.19 desmonomethylpromazine b 2.287 2.703 4.28 3.68 didanosine a -0.404 -0.294 -1.24 -1.26 -0.50 -1.06 diethylene glycol divinyl ether n -0.127 0.200 0.87 1.26 0.87 enflurane n 1.075 1.203 2.10 2.24 2.42 ethanol n -0.534 -0.683 -0.31 -0.40 -0.22 ethyl ether n 0.162 0.308 0.89 1.12 0.70 ethylbenzene n 1.398 1.616 3.15 3.27 2.91 flunitrazepam n 1.621 1.739 2.06 2.20 2.58 fluroxene n 0.570 0.637 1.70 1.58 halothane n 1.165 1.300 2.30 2.50 1.97 indinavir bb 2.864 2.745 2.90 3.26 2.39 isobutyl alcohol n 0.045 0.169 0.76 0.60 0.65 isoflurane n 1.074 1.207 2.30 2.48 http://dx.doi.org/10.5599/admet.1034 grumetto and russo admet & dmpk 9(4) (2021) 267-281 272 table 1. continued… molecule nature clog kw iam.mg clog kw iam.dd2 exp log p n [36] clog p n (1) [26] clog p n (2) [37] clog d 7.4 [37] isopropyl alcohol n -0.241 -0.243 0.05 0.04 0.19 mesoridazine b 2.640 3.027 3.90 3.83 3.41 methoxyflurane n 0.864 1.070 2.21 2.01 1.91 methyl cyclopentane n 1.140 1.347 3.37 3.15 2.31 methyl ethyl ketone n 0.047 0.057 0.29 0.41 1.01 mirtazapine b 1.969 2.287 2.90 3.38 m-xylene n 1.410 1.641 3.20 3.15 2.98 nevirapine n 1.152 1.332 2.50 [25] 1.75 2.19 n-heptane n 1.791 2.189 3.28 n-hexane n 1.545 1.861 2.88 nordazepam n 1.838 2.055 2.79 3.24 northioridazine b 3.120 3.607 5.29 5.1 n-pentane n 1.299 1.529 2.49 quinidine b 2.016 2.394 3.44 2.82 2.32 sulforidazine b 2.684 3.057 4.45 4.32 3.6 teflurane n 1.029 1.066 2.07 1.63 thioridazine b 3.318 3.816 5.90 5.93 5.48 thioxolone n 2.414 2.834 3.90 2.69 2.93 tiotidine b 0.186 0.375 0.68 0.59 1.18 triazolam n 2.102 2.365 2.42 2.94 3.31 trichloroethylene n 0.837 0.944 2.45 2.17 trifluoperazine bb 3.305 3.651 5.03 4.87 4.72 valproic acid a 1.135 1.542 2.75 2.54 2.61 0.37 zidovudine a -0.063 0.094 0.05 -0.1 -0.22 -0.28 figure 1 illustrates the relationships between log bb and the c/’log kw iam values on the mg (a) and dd2 (b) stationary phases and a clear descending trend is visible in both plots. these values, calculated by considering exp log d 7.4 values for acids and log p n values for all the other molecules, are reported in table 2 along with the experimental log bb values. the dataset was divided according to the molecules’ ionization in neutrals (n), bases supporting one (b) or two (bb) basic groups and acidic (a) compounds. this was set to evaluate whether any specific trend was visible in each subgroup. three of the assayed molecules markedly deviate from the pattern identified by the main distribution of points are triazolam, trifluoperazine and valproic acid. the chromatographic behaviorur of small molecules on iam phases has been characterized by many research groups for more than three decades [38,39]. trifluoperazine is a highly lipophilic base, and it is well ascertained [14] that these interact with phospholipids weaker than isolipophilic neutral compounds, especially on the iam.pc.dd2 phase. as to triazolam and valproic acid, the reasons for these deviating from the main distribution of points do not seem that straightforward to spot. triazolam is a benzodiazepine derivative featuring a structure of three condensed rings covalently bound to one chlorobenzene moiety sharing the same plane. it has been again already characterized [39] that those planar structures tend to interact with iam phases stronger than the extent expected based on their lipophilicity, but it is hard to assess whether this played a role in this instance. for its being an acid, the calculation of cδ’log kw iam of valproic acid was based on log d 7.4 rather than log p n . however, since we could not retrieve the admet & dmpk 9(4) (2021) 267-281 bbb passage based on in silico phospholipophilicity doi: http://dx.doi.org/10.5599/admet.1034 273 experimental value from literature sources, we had to rely on the calculated value, whose closeness to the actual value cannot be reasonably taken for granted. interestingly, a descending trend is visible for neutral compounds between the c/’log kw iam and log bb values ranging from +1 to 0 but the distribution flattens for log bb < 0. figure 1. relationships between log bb values and c/’log kw iam.mg (a) and c/’log kw iam.dd2 values (b). table 2. c/’log kw iam.mg , c/’log kw iam.dd2 values and experimental log bb values for the dataset considered. molecule cδ/δ’log kw iam.mg cδ/δ’log kw iam.dd2 log bb 1,1,1-trichloroethane -0.177 -0.196 0.40 1,2-dimethylbenzene -0.339 -0.396 0.37 1,4-dimethylbenzene -0.342 -0.460 0.31 1,7-dimethylxanthine 0.588 0.798 0.06 1-chloro-2,2,2-trifluoroethane 0.070 -0.080 0.08 1-hydroxymidazolam 0.274 0.217 -0.07 2,2-dimethylbutane -1.010 -1.107 1.04 2-methylpentane -0.422 -0.409 0.97 3-methylhexane -0.249 -0.166 0.90 3-methylpentane -0.827 -0.880 1.01 4-hydroxymidazolam 0.029 -0.055 -0.30 acetaminophen 0.195 0.335 -0.31 acetone 0.675 0.748 -0.15 aminopyrine 0.985 1.298 0.00 amobarbital 0.364 0.570 0.04 antipyrine 1.189 1.499 -0.10 bretazenil 1.021 1.191 -0.09 cyclohexane -0.516 -0.659 0.92 cyclopropane -0.346 -0.618 0.00 desmonomethylpromazine 0.104 0.149 0.59 didanosine 1.168 1.579 -1.30 diethylene glycol divinyl ether -0.084 0.270 0.11 http://dx.doi.org/10.5599/admet.1034 grumetto and russo admet & dmpk 9(4) (2021) 267-281 274 table 2. continued… molecule cδ/δ’log kw iam.mg cδ/δ’log kw iam.dd2 log bb enflurane 0.144 0.125 0.24 ethanol 0.444 0.490 -0.16 ethyl ether 0.189 0.360 0.00 ethylbenzene -0.365 -0.443 0.20 flunitrazepam 0.721 0.698 0.06 fluroxene 0.051 0.044 0.13 halothane 0.075 0.035 0.35 indinavir 1.299 0.919 -0.74 isobutyl alcohol 0.175 0.342 -0.17 isoflurane 0.174 0.166 0.42 isopropyl alcohol 0.451 0.593 -0.15 mesoridazine 0.283 0.267 -0.36 methoxyflurane -0.154 -0.111 0.25 methyl cyclopentane -0.797 -0.918 0.93 methyl ethyl ketone 0.549 0.669 -0.08 mirtazapine 0.404 0.461 0.53 m-xylene -0.392 -0.465 0.29 nevirapine -0.096 -0.120 0.00 n-heptane -0.075 0.008 0.81 n-hexane -0.004 0.054 0.80 nordazepam 0.004 -0.088 0.50 northioridazine -0.187 -0.273 0.75 n-pentane 0.059 0.086 0.76 quinidine 0.024 0.064 -0.46 sulforidazine -0.108 -0.216 0.18 teflurane 0.470 0.427 0.27 thioridazine -0.623 -0.812 0.24 thioxolone 0.057 0.074 0.40 tiotidine 0.379 0.623 -0.82 triazolam 0.917 0.988 0.74 trichloroethylene -0.150 -0.200 0.34 trifluoperazine 0.053 -0.164 1.44 valproic acid 1.574 2.079 -0.22 zidovudine 0.891 1.239 -0.72 figure 2 instead displays the relationship occurring between the data of permeation through the bbb and experimental n-octanol/water lipophilicity values. the experimental log p values for 1-chloro-2,2,2trifluoroethane, 1-hydroxymidazolam, 3-methylhexane, 4-hydroxymidazolam, bretazenil, desmonomethylpromazine, fluroxene, isoflurane, n-heptane, n-hexane, nordiazepam, northioridazine, n-pentane, teflurane and trichloroethylene were not available and, therefore, calculated values were used instead. clearly, log p represents an index of paramount importance in pharmaceutical discovery and development [40]. the assumption is that lead compounds should lie in a specific range of lipophilicity to be considered for further admet & dmpk 9(4) (2021) 267-281 bbb passage based on in silico phospholipophilicity doi: http://dx.doi.org/10.5599/admet.1034 275 implementations. the expectation is that lipophilicity should be positively related with data of drugs’ passage through complex barriers, including the bbb [40]. however, the extremely scattered data points of figure 2 evidence that no relationship between log p and log bb values can be observed. likewise, no trend is visible between the two considered variables if all the compounds are considered. however, an ascending trend is visible for acidic compounds, albeit their number is limited. figure 2. relationships between log bb values and log p values. conversely, the situation changes noticeably when considering only the lowest range of log bb (< -0.20). indeed, as figure 3 displays, the relationship between log bb and cδ/δ’log kw iam becomes inverse linear for this subset with a rather solid r 2 value, i.e., > 0.59, with a superior accuracy afforded by delta values on the dd2 phase. this is analogous to what we achieved using delta values obtained from experimentally determined log kw iam values [14-19,34] instead of the calculated ones. we subsequently compared the performance in predicting log bb values of delta descriptors again vs. experimentally determined log p n values (detailed in figure 4). although a direct linear relationship is observable between log bb (< -0.20) and log p n values, its accuracy as assessed from r 2 is inferior to that of the relationship developed from cδ/δ’log kw iam.dd2 values. if cδ/δ’log kw iam are calculated from in silico rather than experimental log p data, the relationship between log bb (< -0.20) and cδ/δ’log kw iam values weaken, although not much, especially on the dd2 phase (r 2 = 0.68 by using clogp values calculated by marvinsketch, data not shown). although the size of our dataset is relatively limited (n = 56), we can extract some interesting information from the results achieved. specifically, the method for predicting cδ/δ’log kw iam cannot yet be used alone in the discovery phase. however, this can be run as complementary along with other assays for profiling the adme potential of drug candidates as it can provide additional information that is not afforded by other early assessments, e.g., lipophilicity. moreover, the method hereby reported seems to be more selective in the identification of the candidates with the slimmest chances to gain access to the brain. this is advantageous, especially if the potency of the candidates that are screened is high enough to be effective, even if the amounts that are successfully delivered to the brain are low. http://dx.doi.org/10.5599/admet.1034 grumetto and russo admet & dmpk 9(4) (2021) 267-281 276 figure 3. relationships beween log bb values (<0.20) and c/’log kw iam.mg (a) and c/’log kw iam.dd2 values (b). figure 4. relationships beween log bb (<0.20) and log p values. these considerations would support the implementation of this method as a filter in the discovery phase to filter out the compounds intended to act toward the brain, with substandard potential to partition in the cns. the method seems to work better if the estimation of cδ/δ’log kw iam relies on experimental lipophilicity data rather than calculated ones. this is not an obstacle since many high throughput platforms for log p assessments are now available on the market [41] or being described [42] and for sure log kw iam measurements are more demanding since they require samples to be run over (at least) three organic modifier concentrations. a further consideration concerns the models used to calculate phospholipophilicity. these are already rather good but could be improved by analysing more and more chemically diverse solutes to broaden their applicability space. c/’log kw iam : multiple linear regression the passage of therapeutics through the bbb is unanimously recognized as an extremely complex admet & dmpk 9(4) (2021) 267-281 bbb passage based on in silico phospholipophilicity doi: http://dx.doi.org/10.5599/admet.1034 277 phenomenon, which results from an interplay of various passage patterns, including transcellular passive, transcellular active and paracellular passage pathways [43]. therefore, it is rather unlikely that a sole descriptor can encode all the interactions taking place in bbb uptake. for this reason, we calculated an ample array (> 1,600) of physico-chemical descriptors by the software e-dragon 1.0 and studied them in (i) simple linear regression and (ii) multiple linear regression vs. log bb values. task (i) was done to establish how c/’log kw iam indexes compared to other physico-chemical descriptors in terms of predictive strength, while task (ii) was accomplished to study whether using multiple variables to model the bbb passage of the dataset could yield some useful statistic models. the results of the simple linear regression analysis are listed in table 3, along with the relevant statistics. an analysis of the data suggests that all regression coefficients are significant at the 99 % level. among all the e-dragon descriptors, c/’log kw iam.dd2 and c/’log kw iam.mg ranked fourth and fifth, respectively and their r 2 values were exceeded only by parameters referring to polarity (tpsa(no)), molecular lipophilicity (alogps), and the number of oxygen atoms (no). a detailed explanation of these and relevant references is reported in supporting information (table s1). the aspect that both the topological surface area and the number of oxygen atoms relate to a fair extent with the bbb passage of the molecules in the dataset may suggest that h-bonding may act by preventing the uptake of these chemicals through the bbb. this agrees well with the observations made by diamond and co-workers [44] and other research groups [45]. subsequently, the incorporation of c/’log kw iam descriptors was attempted in multiple linear regression reported below: log bb = 0.0668 + 0.1548 alogps_logp 0.0779 cδ/δ’log kw iam.mg 0.0046 tpsa(tot) 0.3464 nroh (6) r 2 = 0.73 q 2 = 0.67 se= 0.280 f = 34.45 p=6.26e-14 pc=4.450 and after loo: log bb = 0.0425 + 0.1657 alogps_logp 0.0609 cδ/δ’log kw iam.mg 0.0043 tpsa(tot) 0.3709 nroh (7) r 2 = 0.76 se= 0.266 f = 38.79 p=9.35e-15 pc=3.93 exc: mesoridazine table 3. variable considered and r 2 vs log bb values. a detailed description of the descriptors is offered in supporting information (table s1). the statistics of the each regressor is reported in 2.3. the statistics of the equations has been detailed in 2.3, while exc identifies the compound that was excluded from the regression. according to both equations (6) and (7), the bbb diffusion of chemicals seems to be promoted by molecular lipophilicity and hindered for molecules featuring high c/’log kw iam.mg , which is an index accounting for the excess of the polar/electrostatic interaction forces realized in the interplay between the chemical species and membrane phospholipids. again, according to the models presented above, the bbb uptake of molecules is less efficient for those supporting many polar atoms (and specifically many hydroxyl groups). an r 2 (n-1) value equal to 0.76 suggests that the statistic model is robust and reliable. a plot experimental vs. predicted (according to eq. (7)) log bb values is presented in figure 5. the value of each descriptor is reported in supporting information (table s2). variable r 2 q 2 se f p pc tpsa(no) 0.46 0.42 0.385 46.08 9.08e-09 7.985 alogps_logp 0.42 0.37 0.398 39.36 6.19e-08 8.561 no 0.38 0.33 0.411 33.46 3.77e-07 9.138 cδ/δ’log kw iam.dd2 0.38 0.32 0.413 32.82 5.36e-07 9.205 cδ/δ’log kw iam.mg 0.37 0.32 0.414 32.36 5.36e07 9.255 http://dx.doi.org/10.5599/admet.1034 grumetto and russo admet & dmpk 9(4) (2021) 267-281 278 figure 5. experimental vs predicted log bb values calculated according to eq. (7). conclusions the present study proposes a method to streamline the drug discovery/development process and filter out solutes whose bbb permeation is envisaged to be substandard. even if the dataset is limited in size and the method is not mature enough for broad implementation alone, it may be applied, along with other methodologies by pharmaceutical companies and research institutes, to focus only on candidates that tend to concentrate on the brain. this way, all the others can be neglected, thus saving time and money resources. conflict of interest: the authors lg and gr do not have any competing financial interests to disclose. references [1] j.p. kennedy, l. williams, t.m. bridges, r.n. daniels, d. weaver, c.w. lindsley. application of combinatorial chemistry science on modern drug discovery. j. comb. chem. 10 (2008) 345-354. https://dx.doi.org/10.1021/cc700187t. [2] j.p. renaud, c.w. chung, u.h. danielson, u. egner, m. hennig, r.e. hubbard, h. nar. biophysics in drug discovery: impact, challenges and opportunities. nat. rev. drug discov. 15 (2016) 679-698, https://dx.doi.org/10.1038/nrd.2016.123. [3] s. saeidnia, a. manayi, m. abdollahi. from in vitro experiments to in vivo and clinical studies; pros and cons. curr. drug discov. technol. 12 (2015) 218-224. https://dx.doi.org/10.2174/1570163813666160114093140. [4] i. zeiadeh, a. najjar, r. karaman. strategies for enhancing the permeation of cns-active drugs through the blood-brain barrier: a review. molecules 23 (2018) 1289. https://dx.doi.org/10.3390/molecules23061289. [5] m.m. patel, b.m. patel. crossing the blood-brain barrier: recent advances in drug delivery to the brain. cns drugs 31 (2017) 109-133. https://dx.doi.org/10.1007/s40263-016-0405-9. https://dx.doi.org/10.1021/cc700187t https://dx.doi.org/10.1038/nrd.2016.123 https://dx.doi.org/10.2174/1570163813666160114093140 https://dx.doi.org/10.2174/1570163813666160114093140 https://dx.doi.org/10.3390/molecules23061289 https://dx.doi.org/10.3390/molecules23061289 https://dx.doi.org/10.1007/s40263-016-0405-9 admet & dmpk 9(4) (2021) 267-281 bbb passage based on in silico phospholipophilicity doi: http://dx.doi.org/10.5599/admet.1034 279 [6] x. dong. current strategies for brain drug delivery. theranostics 8 (2018) 1481-1493. https://dx.doi.org/10.7150/thno.21254. [7] m. heymans, e. sevin, f. gosselet, s. lundquist, m. culot. mimicking brain tissue binding in an in vitro model of the blood-brain barrier illustrates differences between in vitro and in vivo methods for assessing the rate of brain penetration. eur. j. pharm. biopharm. 127 (2018) 453-461. https://dx.doi.org/10.1016/j.ejpb.2018.03.007. [8] m. chrysanthakopoulos, f. tsopelas, a. tsantili-kakoulidou. biomimetic chromatography: a useful tool in the drug discovery process. adv. chromatogr. 51 (2014) 91-125. [9] g. russo, g. ermondi, g. caron, d. verzele, f. lynen. into the first biomimetic sphingomyelin stationary phase: suitability in drugs' biopharmaceutic profiling and block relevance analysis of selectivity. eur. j. pharm. sci. 156 (2021) 105585. https://dx.doi.org/10.1016/j.ejps.2020.105585. [10] c. stergiopoulos, f. tsopelas, k. valko. prediction of herg inhibition of drug discovery compounds using biomimetic hplc measurements. admet and dmpk 9 (2021) 191-207. https://doi.org/10.5599/admet.995. [11] k.l. valko, t. zhang. biomimetic properties and estimated in vivo distribution of chloroquine and hydroxy-chloroquine enantiomers. admet and dmpk 9 (2021) 151-165. https://doi.org/10.5599/admet.929. [12] f. tsopelas, t. vallianatou, a. tsantili-kakoulidou. advances in immobilized artificial membrane (iam) chromatography for novel drug discovery. expert opin. drug discov. 11 (2016) 473-488. https://dx.doi.org/10.1517/17460441.2016.1160886. [13] g. russo, l. grumetto, r. szucs, f. barbato, f. lynen. screening therapeutics according to their uptake across the blood-brain barrier: a high throughput method based on immobilized artificial membrane liquid chromatography-diode-array-detection coupled to electrospray-time-of-flight mass spectrometry. eur. j. pharm. biopharm. 127 (2018) 72-84. https://dx.doi.org/10.1016/j.ejpb.2018.02.004. [14] l. grumetto, c. carpentiero, f. barbato. lipophilic and electrostatic forces encoded in iam-hplc indexes of basic drugs: their role in membrane partition and their relationships with bbb passage data. eur. j. pharm. sci. 45 (2012) 685-692. https://dx.doi.org/10.1016/j.ejps.2012.01.008. [15] l. grumetto, c. carpentiero, p. di vaio, f. frecentese, f. barbato. lipophilic and polar interaction forces between acidic drugs and membrane phospholipids encoded in iam-hplc indexes: their role in membrane partition and relationships with bbb permeation data. j. pharm. biomed. anal. 75 (2013) 165-172, https://dx.doi.org/10.1016/j.jpba.2012.11.034. [16] l. grumetto, g. russo, f. barbato. indexes of polar interactions between ionizable drugs and membrane phospholipids measured by iam-hplc: their relationships with data of blood-brain barrier passage. eur j pharm sci 65 (2014) 139-146. https://dx.doi.org/10.1016/j.ejps.2014.09.015. [17] l. grumetto, g. russo, f. barbato. relationships between human intestinal absorption and polar interactions drug/phospholipids estimated by iam-hplc. int. j. pharm. 489 (2015) 186-194. https://dx.doi.org/10.1016/j.ijpharm.2015.04.062. [18] l. grumetto, g. russo, f. barbato. immobilized artificial membrane hplc derived parameters vs pampa-bbb data in estimating in situ measured blood-brain barrier permeation of drugs. mol. pharm. 13 (2016) 2808-2816. https://dx.doi.org/10.1021/acs.molpharmaceut.6b00397. [19] l. grumetto, g. russo, f. barbato. polar interactions drug/phospholipids estimated by iam-hplc vs cultured cell line passage data: their relationships and comparison of their effectiveness in predicting drug human intestinal absorption. int. j. pharm. 500 (2016) 275-290. https://dx.doi.org/10.1016/j.ijpharm.2016.01.019. [20] t. braumann, g. weber, l.h. grimme. quantitative structure—activity relationships for herbicides: reversed-phase liquid chromatographic retention parameter, log kw, versus liquid-liquid partition coefficient as a model of the hydrophobicity of phenylureas, s-triazines and phenoxycarbonic acid derivatives. journal of chromatography a 261 (1983) 329-343. https://doi.org/10.1016/s00219673(01)87961-9. http://dx.doi.org/10.5599/admet.1034 https://dx.doi.org/10.7150/thno.21254 https://dx.doi.org/10.7150/thno.21254 https://dx.doi.org/10.1016/j.ejpb.2018.03.007 https://dx.doi.org/10.1016/j.ejpb.2018.03.007 https://dx.doi.org/10.1016/j.ejps.2020.105585 https://doi.org/10.5599/admet.995 https://doi.org/10.5599/admet.995 https://doi.org/10.5599/admet.929 https://doi.org/10.5599/admet.929 https://dx.doi.org/10.1517/17460441.2016.1160886 https://dx.doi.org/10.1016/j.ejpb.2018.02.004 https://dx.doi.org/10.1016/j.ejpb.2018.02.004 https://dx.doi.org/10.1016/j.ejps.2012.01.008 https://dx.doi.org/10.1016/j.jpba.2012.11.034 https://dx.doi.org/10.1016/j.ejps.2014.09.015 https://dx.doi.org/10.1016/j.ijpharm.2015.04.062 https://dx.doi.org/10.1021/acs.molpharmaceut.6b00397 https://dx.doi.org/10.1016/j.ijpharm.2016.01.019 https://dx.doi.org/10.1016/j.ijpharm.2016.01.019 https://doi.org/10.1016/s0021-9673(01)87961-9 https://doi.org/10.1016/s0021-9673(01)87961-9 grumetto and russo admet & dmpk 9(4) (2021) 267-281 280 [21] t.s. carpenter, d.a. kirshner, e.y. lau, s.e. wong, j.p. nilmeier, f.c. lightstone. a method to predict blood-brain barrier permeability of drug-like compounds using molecular dynamics simulations. biophys. j. 107 (2014) 630-641. https://dx.doi.org/10.1016/j.bpj.2014.06.024. [22] g. russo, f. barbato, l. grumetto, l. philippe, f. lynen, g.h. goetz. entry of therapeutics into the brain: influence of exposed polarity calculated in silico and measured in vitro by supercritical fluid chromatography. int. j. pharm. 560 (2019) 294-305. https://dx.doi.org/10.1016/j.ijpharm.2019.02.008. [23] p.k. pandey, a.k. sharma, u. gupta. blood brain barrier: an overview on strategies in drug delivery, realistic in vitro modeling and in vivo live tracking. tissue barriers 4 (2016) e1129476. https://dx.doi.org/10.1080/21688370.2015.1129476. [24] g. russo, l. grumetto, f. barbato, g. vistoli, a. pedretti. prediction and mechanism elucidation of analyte retention on phospholipid stationary phases (iam-hplc) by in silico calculated physicochemical descriptors. eur. j. pharm. sci. 99 (2017) 173-184. https://dx.doi.org/10.1016/j.ejps.2016.11.026. [25] d.s. wishart, y.d. feunang, a.c. guo, e.j. lo, a. marcu, j.r. grant, t. sajed, d. johnson, c. li, z. sayeeda, n. assempour, i. iynkkaran, y. liu, a. maciejewski, n. gale, a. wilson, l. chin, r. cummings, d. le, a. pon, c. knox, m. wilson. drugbank 5.0: a major update to the drugbank database for 2018. nucleic acids res. 46 (2018) d1074-d1082. https://dx.doi.org/10.1093/nar/gkx1037. [26] i.v. tetko, v.y. tanchuk. application of associative neural networks for prediction of lipophilicity in alogps 2.1 program. j. chem. inf. comput. sci. 42 (2002) 1136-1145. https://dx.doi.org/10.1021/ci025515j. [27] marvinsketch version 17.1.16.0, chemaxon ltd, budapest, hungary http://www.chemaxon.com (2017). [28] j.a. platts, m.h. abraham, y.h. zhao, a. hersey, l. ijaz, d. butina. correlation and prediction of a large blood-brain distribution data set--an lfer study. eur. j. med. chem. 36 (2001) 719-730. https://dx.doi.org/10.1016/s0223-5234(01)01269-7. [29] w.c. griffin. calculation of hlb values of non-ionic surfactants. j. soc. cosmet. chem. 5 (1954) 249256. [30] j. davies. a quantitative kinetic theory of emulsion type . i . physical chemistry of the emulsifying. 2003. [31] molinspiration: log p n-octanol-water partition coefficient. http://www.molinspiration.com/services/logp.html. 2021. [32] a. mauri, v. consonni, r. todeschini, molecular descriptors, in: j. leszczynski, a. kaczmarek-kedziera, t. puzyn, m. g. papadopoulos, h. reis, m. k. shukla (eds.) handbook of computational chemistry, springer international publishing, cham, 2017, pp. 2065-2093. https://dx.doi.org/10.1007/978-3-31927282-5_51. [33] a. pedretti, l. villa, g. vistoli. vega-an open platform to develop chemo-bio-informatics applications, using plug-in architecture and script programming. j. comput. aided mol. des. 18 (2004) 167-173. https://dx.doi.org/ 10.1023/b:jcam.0000035186.90683.f2. [34] g. russo, l. grumetto, r. szucs, f. barbato, f. lynen. determination of in vitro and in silico indexes for the modeling of blood-brain barrier partitioning of drugs via micellar and immobilized artificial membrane liquid chromatography. j. med. chem. 60 (2017) 3739-3754. https://dx.doi.org/10.1021/acs.jmedchem.6b01811. [35] j. verma, v.m. khedkar, e.c. coutinho. 3d-qsar in drug design--a review. curr. top. med. chem. 10 (2010) 95-115. https://dx.doi.org/10.2174/156802610790232260. [36] s. kim, j. chen, t. cheng, a. gindulyte, j. he, s. he, q. li, b.a. shoemaker, p.a. thiessen, b. yu, l. zaslavsky, j. zhang, e.e. bolton. pubchem in 2021: new data content and improved web interfaces. nucleic acids research 49 (2020) d1388-d1395. https://dx.doi.org/10.1093/nar/gkaa971. [37] m.s.f.w.-b.p.v. http://www.chemaxon.com/products/marvin/marvinsketch/. calculation module designed by chemaxon. https://dx.doi.org/10.1016/j.bpj.2014.06.024 https://dx.doi.org/10.1016/j.ijpharm.2019.02.008 https://dx.doi.org/10.1080/21688370.2015.1129476 https://dx.doi.org/10.1080/21688370.2015.1129476 https://dx.doi.org/10.1016/j.ejps.2016.11.026 https://dx.doi.org/10.1016/j.ejps.2016.11.026 https://dx.doi.org/10.1093/nar/gkx1037 https://dx.doi.org/10.1021/ci025515j https://dx.doi.org/10.1021/ci025515j http://www.chemaxon.com/ https://dx.doi.org/10.1016/s0223-5234(01)01269-7 http://www.molinspiration.com/services/logp.html http://www.molinspiration.com/services/logp.html https://dx.doi.org/10.1007/978-3-319-27282-5_51 https://dx.doi.org/10.1007/978-3-319-27282-5_51 https://dx.doi.org/%2010.1023/b:jcam.0000035186.90683.f2 https://dx.doi.org/10.1021/acs.jmedchem.6b01811 https://dx.doi.org/10.1021/acs.jmedchem.6b01811 https://dx.doi.org/10.2174/156802610790232260 https://dx.doi.org/10.1093/nar/gkaa971 http://www.chemaxon.com/products/marvin/marvinsketch/ admet & dmpk 9(4) (2021) 267-281 bbb passage based on in silico phospholipophilicity doi: http://dx.doi.org/10.5599/admet.1034 281 [38] f. tsopelas, t. vallianatou, a. tsantili-kakoulidou. advances in immobilized artificial membrane (iam) chromatography for novel drug discovery. expert opinion on drug discovery 11 (2016) 473-488. https://dx.doi.org/10.1517/17460441.2016.1160886. [39] a. taillardat-bertschinger, f. barbato, m.t. quercia, p.-a. carrupt, m. reist, m.i. la rotonda, b. testa. structural properties governing retention mechanisms on immobilized artificial membrane (iam) hplc columns. helvetica chimica acta 85 (2002) 519-532. https://dx.doi.org/10.1002/15222675(200202)85:2<519::aid-hlca519>3.0.co;2-q. [40] j.a. arnott, s.l. planey. the influence of lipophilicity in drug discovery and design. expert opin. drug discov. 7 (2012) 863-875. https://dx.doi.org/10.1517/17460441.2012.714363. [41] h. wan, a.g. holmen. high throughput screening of physicochemical properties and in vitro adme profiling in drug discovery. comb. chem. high throughput screen. 12 (2009) 315-329. https://dx.doi.org/10.2174/138620709787581701. [42] z. chen, s.g. weber. high-throughput method for lipophilicity measurement. anal. chem. 79 (2007) 1043-1049. https://dx.doi.org/10.1021/ac061649a. [43] p. jeffrey, s.g. summerfield. challenges for blood-brain barrier (bbb) screening. xenobiotica 37 (2007) 1135-1151. https://dx.doi.org/10.1080/00498250701570285. [44] j.m. diamond, e.m. wright, d.h. smyth. molecular forces governing non-electrolyte permeation through cell membranes. proceedings of the royal society of london. series b. biological sciences 172 (1969) 273-316. https://dx.doi.org/10.1098/rspb.1969.0022. [45] e.g. chikhale, k.-y. ng, p.s. burton, r.t. borchardt. hydrogen bonding potential as a determinant of the in vitro and in situ blood–brain barrier permeability of peptides. pharmaceutical research 11 (1994) 412-419. https://dx.doi.org/10.1023/a:1018969222130. ©2021 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://dx.doi.org/10.5599/admet.1034 https://dx.doi.org/10.1517/17460441.2016.1160886 https://dx.doi.org/10.1002/1522-2675(200202)85:2%3c519::aid-hlca519%3e3.0.co;2-q https://dx.doi.org/10.1002/1522-2675(200202)85:2%3c519::aid-hlca519%3e3.0.co;2-q https://dx.doi.org/10.1517/17460441.2012.714363 https://dx.doi.org/10.2174/138620709787581701 https://dx.doi.org/10.1021/ac061649a https://dx.doi.org/10.1080/00498250701570285 https://dx.doi.org/10.1098/rspb.1969.0022 https://dx.doi.org/10.1023/a:1018969222130 http://creativecommons.org/licenses/by/3.0/ challenges, current status and emerging strategies in the development of rapidly dissolving fdm 3d-printed tablets: an overview and commentary doi: https://doi.org/10.5599/admet.1661 277 admet & dmpk 11(2) (2023) 277-291; doi: https://doi.org/10.5599/admet.1661 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper use of azospirillum baldaniorum cells in quercetin detection matvey v. kanevskiy1, irina s. kosheleva1, vladislav o. menukhov1, elizaveta s. zhdanova1, svetlana v. borisova1, gennady l. burygin1,2, svetlana a. konnova1,2, victor d. bunin3 and olga i. guliy2* 1chernyshevsky saratov state university, saratov 410012, russia 2institute of biochemistry and physiology of plants and microorganisms – subdivision of the federal state budgetary research institution saratov federal scientific center of the russian academy of sciences (ibppm ras), saratov 410049, russia 3elosystems gbr, germany *corresponding author e-mail: guliy_olga@mail.ru received: january 7, 2023; revised: march 6, 2023; published: march 21, 2023 abstract the possibility of detection and determination of flavonoids by using microbial cells was shown for the first time using the quercetin azospirillum baldaniorum sp245 model system. the activity of the flavonoids quercetin, rutin and naringenin toward a. baldaniorum sp245 was evaluated. it was found that when the quercetin concentration ranged from 50 to 100 µm, the number of bacterial cells decreased. rutin and naringenin did not affect bacterial numbers. quercetin at 100 μm increased bacterial impedance by 60 %. under the effect of quercetin, the magnitude of the electro-optical signal from cells decreased by 75 %, as compared with the no-quercetin control. our data show the possibility of developing sensor-based systems for the detection and determination of flavonoids. ©2023 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords flavonoids; quercetin; azospirillum; electro-optical analysis; impedance introduction quercetin is one of the most widespread and studied flavonoids plant secondary metabolites that are actively used by humans owing to their high biological activity. the beneficial effect of flavonoids has been repeatedly proven in the treatment of a number of diseases, including cancer, alzheimer’s disease, and atherosclerosis [1]. flavonoids are indispensable in various pharmaceutical, medicinal, and cosmetic products, and they are used as antioxidants, angioand cardioprotective drugs, antiinflammatory agents, antidiabetic and anticancer drugs, and so on [2-7]. the total flavonoid content is a very important indicator of the quality of plant materials, which is associated with the overall antioxidant activity of flavonoids. because of the growing need for natural antioxidants, there is an increasing need to devise new flavonoid-detection methods and prepare standards for flavonoid determination [8]. https://doi.org/10.5599/admet.1661 https://doi.org/10.5599/admet.1661 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:guliy_olga@mail.ru http://creativecommons.org/licenses/by/4.0/ m. v. kanevskiy et al. admet & dmpk 11(2) (2023) 277-291 278 the simplest way to detect and determine flavonoids is to use a spectrophotometric method based on the color reaction of flavonoids with aluminum chloride [9]. this method is recommended by the european pharmacopoeia [8] because it is highly sensitive (up to 10 ng/ml), widely available, and does not require complex sample preparation. however, it is more suitable for determining the total content of phenolic compounds because individual flavonoids in a mixture are difficult to identify owing to the similarity of their spectra [10-12]. the most common methods for fractionating and detecting flavonoids in plant extracts are thin-layer chromatography (tlc) and high-performance liquid chromatography (hplc). tlc is mainly used for the primary rapid analysis of the composition of isolated substances and is ineffective at separating mixtures of flavonoids [13]. hplc quickly and effectively separates mixtures of flavonoids and their derivatives (omethylated and glycosylated forms) and is highly sensitive (up to 1 ng/ml), but it requires expensive equipment and a wide variety of highly purified standards [14,15]. the use of solid-liquid extraction, solidphase extraction, and ultrahigh-performance liquid chromatography coupled with tandem mass spectrometry strongly expands the possibilities of hplc and makes it the most informative method for the characterization of the flavonoid structures in a mixture. the weak point, however, is the need for complex and expensive equipment and for highly trained operators [16,17]. for the analysis of flavonoids, capillary electrophoresis is also used, which is highly effective at separating substances [18,19] and is highly sensitive (up to 0.3 µg/ml). however, depending on the expected flavonoid composition in the extract, it becomes necessary to vary buffer composition, which limits the applicability of this method. few data exist on the use of bacteria-based biosensors for the detection and determination of flavonoids. in particular, siedler et al. [20] showed that the introduction of special gene constructs makes it possible to use e. coli cells to detect flavonoids from 100 µm onward. but for the operation of the sensor to be successful, it is necessary to first transform bacterial cells and then carry out the detection with fluorescence, which also makes the method more expensive and labor-intensive. despite the many existing methods for the detection and determination of flavonoids, it is very important to search for new alternative and universal methods for their determination in extracts. a promising approach to detecting various compounds can be one that is based on the analysis of microbial cell responses. for example, velichko et al. [21] showed that a decrease in the polarizability tensor γ of herbaspirillum lusitanum cells is directly proportional to the content of sodium dodecyl sulfate in the medium in the range 0.01–2 mg/l. microorganisms can be used as a simple and fast-sensing element in the determination of the flavonoid of interest. in particular, nonpathogenic rhizobacteria of the genus azospirillum can be used in biosensors because flavonoids inhibit the growth of azospirilla and change the physicochemical properties and composition of their surface glycopolymers [22,23]. here, the possibility of detecting and determining quercetin when acting on azospirillum baldaniorum sp245 was evaluated by the changes in the physicochemical characteristics of bacteria. experimental bacteria and culture conditions a. baldaniorum sp245 (ibppm 219) [24,25] was from the collection of rhizosphere microorganisms, institute of biochemistry and physiology of plants and microorganisms, russian academy of sciences (ibppm ras) (http://collection.ibppm.ru). bacteria were grown in an iron–free liquid malate medium [26] with constant stirring on a shaker at 30 °c for 24 h. http://collection.ibppm.ru/ admet & dmpk 11(2) (2023) 277-291 quercetin detection with azospirillum cells doi: https://doi.org/10.5599/admet.1661 279 flavonoids quercetin (≥98 %; diaem, russia), rutin (≥94 %; sigma–aldrich, usa), naringenin (≥95 %; sigma–aldrich; usa) were added to the medium as a solution in dimethyl sulfoxide (dmso) to final concentrations 50, 100, and 200 μm. the content of dmso in the medium was 1 vol.%. the inoculum was added to the medium to a600 = 0.09–0.11. phase-contrast microscopy both control and flavonoid-incubated samples were subjected to phase-contrast microscopy (magnification, 40×) on a leica lmd 7000 laser dissector (leica microsystems, germany). the microscopy was done at the simbioz center for the collective use of research equipment (ibppm ras, saratov). samples were stained with aqueous 1 % crystal violet, as described in [27]. microbial cells were placed on a coverslip, moistened with 1 % crystal violet solution, and incubated at room temperature for 30 min. the stain was then removed, and the samples were carefully washed with a phosphate buffer until the washing solution was completely clear. the stained biofilms were dried at 50 °c and subjected to phase-contrast microscopy. measurement of orientational spectra of bacterial suspensions measurements were made with an elus eo analyzer (elosystems gbr, germany), as described in [28]. the measuring conditions were: electric field strength, 89.4 v/cm; light wavelength, 670 nm (relative to vacuum); field application time, 4.5 s. before study, the bacteria were washed by double centrifugation at 2800 g for 5 min in distilled water (conductivity, 1.8 μs/cm). to remove cellular aggregates, we centrifuged the cell suspension again at 110 × g for 1 min and used the suspension that remained in the supernatant liquid for further work. the absorbance (a670) of the suspension was adjusted from 0.4 to 0.42. the orientingfield frequencies were 50, 100, 200, 300, 400, 500, 700, 1000, 2000, and 3000 khz. orientational spectra were presented as the frequency dependence of the difference (δa) between the absorbance values of cell suspensions. this difference was measured during the preparation of a beam of unpolarized light along and across the orienting-field direction and was normalized to the absorbance value measured for randomly oriented cells [29,30]. optical sensor systems are based on the effects of an electric field on cells suspended in an aqueous medium; i.e., they are based on the use of particle polarizability in an electric field and the measurement of the optical manifestation of the polarizability results [kerr effect/electro-optical effect (eo)]. when an electric field is applied to a cell suspension, the cellular structures become polarized and, as a result, the cells acquire an induced cellular dipole moment. the physical manifestation of this effect is the electrical orientation of cells, which manifests itself as a transition of cells to an oriented state. the phenomenon of electric-field orientation is due to the effect of an external electric field on the dipole moment of cells. polarization arises when a field is applied and decreases when it is removed not instantly but after some finite time, called the relaxation time [29,30]. figure 1 shows the general scheme for the eo analyzer. impedance dispersion analysis figure 2 shows the general scheme and outside appearance of the installation used for impedance dispersion analysis. the container for impedancemetry was a 2-ml plastic container in which platinum electrodes were mounted at a distance of 1 cm from each other. before the experiment, a control was done in which the resistance of the nacl solution (9 g/l) was measured in the entire frequency range. the obtained value was subtracted from the experimental one to level the effect of the washing solution on impedance. https://doi.org/10.5599/admet.1661 m. v. kanevskiy et al. admet & dmpk 11(2) (2023) 277-291 280 figure 1. general block diagram of the eo analyzer. the photometric scheme is made according to a two-beam scheme. two orthogonal light beams from led1 and led2 alternately pass through the eo cell. then, by using a system of mirrors, the beams are focused on a photodetector. the eo cell contains nine electrodes, which are connected to an electric field generator. the electric field generator contains a frequency synthesizer with two antiphase outputs, a two-channel variable gain amplifier, a switch, and a power amplifier. the outputs of the four-channel power amplifier are connected to the groups of electrodes of the eo cell a b figure 2. a) general scheme and b) appearance of the installation for impedance dispersion analysis bacteria were spun down by centrifugation (10,000 g, 5 min) and washed free of the medium twice with nacl (9 g/l), after which the culture impedance was measured. impedance dispersion was examined by using an installation consisting of a g6-27 signal generator, a v7-27 a/1 digital voltmeter, a k762 power supply unit, and a measuring cuvette (russia). a sample was placed in the measuring cell, and the incoming voltage (uin) (voltage in the circuit before the electric current passes through the sample) was measured. next, the voltage in the circuit was measured after the current had passed through the sample (uout). for the measuring cell, the exemplary resistance r0 was calculated and found to be 4.99 kohm (constant). in accordance with ohm’s law, a formula was derived led1 led2 photodetector frequency generator mirror mirror keys amplifier controller computer usb interface admet & dmpk 11(2) (2023) 277-291 quercetin detection with azospirillum cells doi: https://doi.org/10.5599/admet.1661 281 for calculating the impedance (1) because the voltage of the alternating current is proportional to its resistance. the uin and uout were measured at 101, 102, 103, 104, 105, and 106 hz, after which the resistance value was found from the formula: 0 in out 1 r z u u = − (1) where uin is the input voltage, uout is the output voltage, and r0 is 4.99 kohm. additionally, the polarization coefficient (pc) was calculated as an indicator of the ability of charged particles to move in the dielectric under the influence of an electric field. the pc was calculated by evaluating the steepness of the impedance dispersion curve and the cell viability. intact cells have high resistance at low electric-current frequencies because the membrane acts as a capacitor. membrane damage leads to a decrease in the capacitive component of impedance; accordingly, cell viability decreases and the impedance dispersion curve becomes flatter. the flavonoid-affected pc of the azospirillum cells was also calculated. the pc [31] is the ratio between the impedance indicators at two frequencies, 10 khz and 1 mhz. eps extraction for isolating extracellular polysaccharides (epss), the cells were pelleted by centrifugation at 10,000 g for 30 min. three volumes of chilled ethanol were added to cold cell-free culture liquid, and the mixture was left to stand at 4 °c for 24 h [32]. the precipitated epss were separated by centrifugation (10,000g , 30 min), suspended, dialyzed against distilled water for 24 h, concentrated on a rotary evaporator, and lyophilized. colony‑forming‑unit (cfu) counting colonies formed from individual viable cells after growth with the flavonoids were counted by the standard method of plating on solid nutrient media [33]. the control was the colonies grown without the flavonoids. gas-liquid chromatography (glc) samples were hydrolyzed with 2 m cf3cooh (120 °c, 2 h), reduced with nabh4, and acetylated [34]. eps monosaccharide composition was examined by glc of polyol acetates on a hewlett-packard 5890 chromatograph. antibodies polyclonal rabbit antibodies were prepared to glutaraldehyde-treated a. baldaniorum sp245 cells, as described [35]. statistics for each series of experiments, at least five repetitions were done. data were analyzed with microsoft excel 2010 and with standard statistical processing methods. results and discussion cfu counting flavonoids can be divided into subclasses, including chalcones, flavones, flavonols, flavanones, and isoflavones. quercetin is one of the first isolated and best-studied flavonoids; therefore, we investigated its https://doi.org/10.5599/admet.1661 m. v. kanevskiy et al. admet & dmpk 11(2) (2023) 277-291 282 effect on the rhizospheric bacterium a. baldaniorum sp245. for this purpose, bacteria were grown with quercetin (concentrations 50, 100, and 200 μm) and postincubation changes in cfu were analyzed. the quercetin concentrations had been chosen on the basis of previous studies [23,30]. figure 3 shows that quercetin at 50 and 100 μm inhibited culture growth, decreasing the cell number by 10 times (9.7×107 and 4.8×107 cfu/ml, respectively), as compared with the control (2.4 × 108 cfu/ml). at 200 μm, quercetin precipitated, precluding further studies with this concentration. figure 3. cfu of a. baldaniorum sp245 grown with flavonoids. *significant differences for p<0.01 to evaluate the specificity of the quercetin effect on a. baldaniorum sp245, we also made studies with rutin and naringenin, whose structural formulas are similar to quercetin's (figure 4). these flavonoids belong to different subclasses and, therefore, differ in hydrophobicity and degree of glycosylation [36,37]. the experimental procedure was the same as that used with quercetin. cfu counts showed neither rutin nor naringenin affected culture growth (figure 3). figure 4. structures of the flavonoids used. quercetin has high biological activity, including bacteriostatic activity [38-40]. importantly, quercetin and naringenin are aglycone forms of flavonoids; however, the addition of naringenin did not inhibit culture growth. rutin, a glycoside of quercetin, did not affect the viability of a. baldaniorum sp245 either. these results can be explained in terms of the chemical structure of the flavonoids used. according to the literature data [2], the antimicrobial activity of flavonoids is associated with the presence and number of hydroxyl groups in their structure. unlike quercetin, naringenin does not have hydroxyl groups at positions 3 and 3', whereas in rutin, the hydroxyl group at position 3 is involved in the formation of a glycosidic bond with the rutinose residue (figure 4). admet & dmpk 11(2) (2023) 277-291 quercetin detection with azospirillum cells doi: https://doi.org/10.5599/admet.1661 283 phase-contrast microscopy to confirm the inhibition of a. baldaniorum cell growth by quercetin at 50 and 100 μm, we used phasecontrast light microscopy. this kind of microscopy was chosen because the leica lmd 7000 system makes it possible to identify relevant cells and ensures noncontact and contaminant-free isolation of individual cells. the high numerical aperture of the lens objectives and the short laser wavelength provide high-resolution images along the optical and transverse directions. this microscopy was used previously to evaluate the effect of antibacterial inhibitors on bacteria [41]. figure 5 shows the images for cell suspensions unexposed (figure 5 a) and exposed to flavonoids (concentrations 50, 100, and 200 μm) (figure 5 b-i). neither naringenin nor rutin (figure 5 d-i) affected the change in the cell number, as compared with the control (figure 3 a). however, with increasing quercetin concentration (figure 5 b-c), the number of cells in the field of view decreased compared to the control. this finding is consistent with the postexposure cfu counts. figure 5. light microscopy of a. baldaniorum sp245. a) control culture. cultures grown with flavonoids: b) quercetin, 50 μm; c) quercetin, 100 μm; d) rutin, 50 μm; e) rutin, 100 μm; f) rutin, 200 μm; g) naringenin, 50 µm; h) naringenin, 100 µm; i) naringenin, 200 µm. the ruler size is 50 µm impedancemetry of azospirillum cells because flavonoids can interact with the bacterial surface and change the barrier properties of the cell wall, impedancemetry was used to evaluate the effect of the flavonoids on a. baldaniorum. by impedancemetry, one https://doi.org/10.5599/admet.1661 m. v. kanevskiy et al. admet & dmpk 11(2) (2023) 277-291 284 can clearly show cell-resistance changes that occur under the influence of external conditions. the greatest influence on the impedance of a. baldaniorum sp245 was exerted by quercetin at 100 µm (figure 6 a). at high frequencies, the impedance value was greater than the control one by 2.1 times. the explanation for the sharp increase in bacterial ohmic part of resistance could be that high concentrations of quercetin in the growth medium stimulate the cell to stabilize the membrane. however, the pc (figure 7 a) decreased with increasing flavonoid concentration and reached its minimum (60 %) at 100 μm, as compared with the control. figure 6. impedance of a. baldaniorum sp245 grown with a) quercetin, b) rutin, and c) naringenin the change in the impedance of the cells incubated with rutin and naringenin (figure 6 b and 6 c) was significantly less (by 43 and 51 %, respectively). the pc also decreased with increasing concentrations of flavonoids, but even at the maximum (200 μm) rutin and naringenin concentration, it did not decrease as much as it did with quercetin, equaling 73 and 71 %, respectively (figure 7 b and 7 c). figure 7. polarization coefficients for a. baldaniorum sp245 grown with a) quercetin, b) rutin, and c) naringenin in electrical engineering, impedance is the opposition to an alternating current presented by the combined effect of resistance and reactance in a circuit. for living objects, impedance is the sum of the resistance of the membranes (capacitive component) and that of the internal environment⸻the cytoplasm (ohmic component). the impedance of the internal contents depends only on the chemical composition of the cytoplasm and does not change depending on the frequency of the electric current. the impedance of the membranes will decrease with increasing frequency; at high frequencies, the impedance will be represented only by the ohmic component. depending on the state of the membrane (intact or damaged), the value of the capacitive component will change, and hence the impedance as a whole will change, too. in our work, the influence of the external environment on impedance (ohmic part of resistance+reactance) was excluded because we examined the impedance of the cells washed from the nutrient medium. figure 6 shows that the presence of quercetin in the growth medium causes changes in both capacitive and ohmic components of impedance. because it is assumed that the effect of flavonoids on bacteria results in changes in the cell membranes, we considered only the change in the capacitive component of impedance. admet & dmpk 11(2) (2023) 277-291 quercetin detection with azospirillum cells doi: https://doi.org/10.5599/admet.1661 285 eo analysis of azospirillum cell suspensions because flavonoids may damage the cell surface of bacteria, we used eo analysis to confirm the results obtained. previous studies have used eo analysis to evaluate changes in microbial cells' electrophysical and morphometric variables subjected to an electric field. the evaluations were done in real time and did not involve special sample preparation. during measurements, the effect of the electric field on the cells was minimal and did not lead to cell death, and the effect of the supporting medium on the measuring accuracy was insignificant [29,42]. the presence of quercetin in the growth medium lowered the eo signal up to 75 % across the range of frequencies used, as compared with the control (figure 8 a). according to the literature, the cell surface charge depends on the integrity of cell membranes [43]. quercetin-caused damage to bacterial membranes may decrease the cellular potential. according to the goldman equation, damage to cell membranes is associated with an increase in their permeability and, consequently, passive transport between intraand extracellular spaces decreases the transmembrane potential. in addition to the decrease in the cell charge as a result of quercetin-caused damage to the cell membranes, the membrane potentials of the cells decrease and the cells become dehydrated; this, ultimately, changes the eo signal (figure 8 a). figure 8. eo spectra of suspensions of a. baldaniorum sp245 grown with a) quercetin, b) rutin, and c) naringenin. d) the magnitude of the eo signal at 500 khz the addition of naringenin led to a slight (no more than 3 %) change in the magnitude of the eo signal across the range of frequencies used (figure 8 b), as compared with the control. the presence of rutin in the culture medium mediated an increase in the eo signal magnitude in the low-frequency region by 13 % (figure 8 c). this indirectly indicates that the changes affected the polymers present on the cell surface. in the high-frequency range, on the contrary, the magnitude of the eo signal decreased by 20 %, as compared with the control, only with 200 μm rutin. these changes in the anisotropy of cell polarizability at 1000, 2000, and 3000 khz, reflecting the state of the cytoplasm, are the earliest response to changes in the vital parameters of a bacterial culture under the effect of rutin and indicate a change in the physicochemical properties of the cell cytoplasm. for the convenience of presentation, figure 8 d shows the change in the eo https://doi.org/10.5599/admet.1661 m. v. kanevskiy et al. admet & dmpk 11(2) (2023) 277-291 286 signal magnitude at an orienting-field frequency of 500 khz. at different orienting-field frequencies, different cell structures take part in the formation of an induced dipole, including cell surface structures, membrane biomolecules, and cytoplasm components [29]. at a low orienting-field frequency (below 200 khz), an induced dipole is formed owing to the structures protruding into the extracellular space. the changes in the eo signal between 200 and 1000 khz indicate a change in the membrane composition. the signal differences above 1000 khz may indicate changes in the cytoplasm composition. because we showed by impedancemetry and eo analysis that quercetin caused changes in the cell membrane surface, we made additional studies to confirm the preservation of the antigenic determinants characteristic of the a. baldaniorum sp245 cell surface. the main component of the bacterial outer membrane is lipopolysaccharide (lps); therefore, we used antibodies specific for the lps of strain sp245. it is known that changes in the eo signal in the low-frequency region (50–200 khz) point to changes that have occurred on the surface of microbial cells [29,44]. therefore, eo analysis was used to record interactions of the antibodies with the cell surface. antibodies were added to the cell suspension to a final concentration of 6 μg/ml, and measurements were made 5 min after the start of cell incubation with the antibodies. the antibody concentration and exposure time were chosen on the basis of previous studies evaluating the effect of antibodies on bacteria [43]. the specific interaction of the antibodies with cells grown in the presence of quercetin (figure 9 b, c), naringenin (figure 9 d, e, f), and rutin (figure 9 g, h, i) indicated that the antigenic determinants characteristic of sp245 strain were preserved on the cell surface. however, the obtained results do not exclude the appearance of new structures on the cell surface. figure 9. eo analysis of cell suspensions of a. baldaniorum sp245. a) control culture (antibodies to strain sp245). cultures grown with flavonoids: b) quercetin, 50 μm; c) quercetin, 100 μm; d) naringenin, 50 μm; e) naringenin, 100 μm; f) naringenin, 200 μm; g) rutin, 50 µm; h) rutin, 100 µm; i) rutin, 200 µm admet & dmpk 11(2) (2023) 277-291 quercetin detection with azospirillum cells doi: https://doi.org/10.5599/admet.1661 287 eps extraction and monosaccharide composition analysis under the effect of flavonoids, the composition of azospirillum polysaccharides undergoes changes [22,23]. therefore, we isolated the epss of the cultures incubated with quercetin, naringenin, and rutin (each used at 100 μm) and examined their monosaccharide composition. glc analysis showed the preservation of the qualitative composition of the eps-constituting monosaccharides, in agreement with the data by eo analysis. however, the epss of the bacteria grown with rutin and quercetin showed an increased relative proportion of glucose (table 1), possibly indicating the appearance of a new-composition polymer. a similar effect for lps had previously been shown with this strain [45] as a result of the replacement of the carbon source in the nutrient medium. for other azospirillum strains, modifications of the lps composition and structure had also been found in bacteria grown with flavonoids and an increased proportion of rhamnose had been found [23]. such changes may signal adaptation by bacteria transitioning from a free-living to a symbiotic lifestyle. table 1. eps monosaccharides composition in a. baldaniorum sp245 monosaccharide sample control quercetin (100 μm) rutin (100 μm) naringenin (100 μm) content, % of the sum of the polyol acetate peaks rhamnose 84.0±4,2 73±3.6 75±3.5 86±4.3 glucose 16.0±0.8 27±1.5 25±1.4 14±0.6 today, much attention is paid to the search and use of plant materials in various industries, in particular as food, cosmetic, and nutritional supplements and as livestock feed and biomass sources. phenolic compounds, ubiquitous in plants, are indispensable in the human diet and animal feed, mainly owing to their antioxidant properties. in the past 20 years, much attention has been paid to flavonoids, with most of the work being addressed to their potential positive effects and only a few aiming to further improve the existing methods of analysis. as recommended by the european pharmacopoeia, flavonoids are isolated by extraction of plant material, usually with ethyl alcohol, methyl alcohol, or aqueous alcohols (most often, 70 % alcohol as an optimal extractant). this procedure is a lengthy one [8]. as can be seen from table 2, the main methods used to determine flavonoids, including quercetin, take much time to complete. table 2. brief description of the methods used to determine flavonoids method flavonoid limit of detection analysis time, h reference physicochemical methods tlc quercetin rutin gallicacid 1 μg/ml 0.5 μg/ml 0.3 μg/ml 1–2 [13,46] hplc quercetin morin rutin 0.2ng/ml 0.4 μg/ml 0.2 μg/ml 1–2 h [10,47-49] hplc–ms quercetin rutin apigenin hesperedin 1 ng/ml 10 ng/ml 1.1 ng/ml 0.8 ng/ml 3–5 h [16,50] capillary electrophoresis naringin naringenin quercetin rutin apigenin 0.4 µg/ml 0.3 µg/ml 0.3 µg/ml 0.5 µg/ml 0.5 µg/ml 1–2 h [18,19] spectrophotometry total content of flavonoids 10 ng/ml 10–30 min [10] biosensor-based methods fluorimetric analysis of transformed bacterial cells quercetin 10 µg/ml 2–3 h (26-27 h*) [20] eo analysis, impedancemetry quercetin 10 µg/ml 1–2 h (25-26 h*) this work *allowance for culture growth https://doi.org/10.5599/admet.1661 m. v. kanevskiy et al. admet & dmpk 11(2) (2023) 277-291 288 biosensors, which consist of a sensing element and a sensor for signal recording, hold promise for the detection and determination of flavonoids. an important point in the development of a bacteria-based sensor system is the selection of the sensing element. the use of a. baldaniorum sp245 as the sensing element in the detection and determination of quercetin is very promising because, as shown in this work, quercetin specifically affects this bacterium and its response may indicate the presence of the flavonoid in the sample being tested. the obtained data may be useful in the development of a rapid method for the detection and determination of quercetin by using a. baldaniorum sp245. the available literature contains almost no data on the use of bacteria as a sensing element in the detection and determination of flavonoids, including quercetin. the results of this work show the promise of eo analysis and impedancemetry, with a. baldaniorum sp245 as the sensing element, for use in the detection of quercetin. the proposed approach can be used for the rapid detection of flavonoids and evaluation of the antioxidant activity of samples. conclusions quercetin inhibits the viability of a. baldaniorum sp245 from 50 μm onward. by contrast, rutin and naringenin do not affect a. baldaniorum viability. quercetin at 100 μm has the greatest effect on the impedance of a. baldaniorum sp245, as compared with rutin and naringenin. under the effect of quercetin, the eo signal decreases by 75 %, as compared with the control. the quercetin-induced changes affect the polymers present on the cell surface. the interaction of specific antibodies with cells (after incubation with quercetin, naringenin, and rutin) indicates the preservation of characteristic antigenic determinants but does not exclude the appearance of new structures on the bacterial surface. we conclude that a. baldaniorum sp245 is promising for use as a sensing element in the detection and determination of quercetin. conflict of interest: the authors declare no conflict of interest. declaration of competing interest: the authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. acknowledgements: this work was supported by the russian science foundation project no. 22-29-00587. the equipment for the studies was provided by the simbioz center for the collective use of research equipment (ibppm ras, saratov). we thank members of the simbioz center for the collective use of research equipment (ibppm ras, saratov) for their support with phase-contrast microscopy. author contribution statement: m. v., o. i., s. a. conceived and designed research. m. v., i. s., v. o., e. s., s. v., g. l. conducted experiments. m. v., i. s., o. i., v. b., s. a. analyzed data. m. v., o. i., s. a., v. b. wrote the manuscript. all authors read and approved the manuscript. references [1] a.n. panche, a.d. diwan, s.r. chandra. flavonoids. journal of nutritional science 5 (2016), e47. https://doi.org/10.1017/jns.2016.41 [2] l.e. alcaraz, s.e. blanco, o.n. puig, f. tomas, f.h. ferretti. antibacterial activity of flavonoids against methicillin-resistant staphylococcus aureus strains. journal of theoretical biology 205 (2000) 231-240. https://doi.org/10.1006/jtbi.2000.2062 [3] k.c. murthy, j. kim, a. vikram, b.s. patil. differential inhibition of human colon cancer cells by structurally similar flavonoids of citrus. food chemistry 132 (2012) 27-34. https://doi.org/10.1016/ j.foodchem.2011.10.014 [4] a.k. verma, h. singh, m. satyanarayana, s.p. srivastava, p. tiwari, a.b. singh, a.k. dwivedi, s.k. singh, m. srivastava, c. nath, r. raghubir, a.k. srivastava, r. pratap. flavone-based novel antidiabetic and antidyslipidemic agents. journal of medicinal chemistry 55 (2012) 4551-4567. https://doi.org/ 10.1021/jm201107g https://doi.org/10.1017/jns.2016.41 https://doi.org/10.1006/jtbi.2000.2062 https://doi.org/10.1016/j.foodchem.2011.10.014 https://doi.org/10.1016/j.foodchem.2011.10.014 https://doi.org/10.1021/jm201107g https://doi.org/10.1021/jm201107g admet & dmpk 11(2) (2023) 277-291 quercetin detection with azospirillum cells doi: https://doi.org/10.5599/admet.1661 289 [5] m. chen, t. wang, z.-z.jiang, c. shan, h. wang, m.-j.wu, s. zhang, y. zhang, l.-y. zhang. antiinflammatory and hepatoprotective effects of total flavonoid c-glycosides from abrusmollis extracts. chinese journal of natural medicines 12 (2014) 590-598. https://doi.org/10.1016/s18755364(14)60090-x [6] j. zhang, y. wu, x. zhao, f. luo, x. li, h. zhu, c. sun, k. chen. chemopreventive effect of flavonoids from ougan (citrus reticulata cv. suavissima) fruit against cancer cell proliferation and migration. journal of functional foods 10 (2014) 511-519. https://doi.org/10.1016/j.jff.2014.08.006 [7] t.y. wang, q. li, k.s. bi. bioactive flavonoids in medicinal plants: structure, activity and biological fate. asian journal of pharmaceutical sciences 13 (2018) 12-23. https://doi.org/10.1016/j.ajps.2017.08.004 [8] european pharmacopoeia. – 10th ed. – sup. 10. – strasbourg: european department for the quality of medicines, 2020 (isbn: 978-3-7692-7453-0) [9] s.a. baba, s.a. malik. determination of total phenolic and flavonoid content, antimicrobial and antioxidant activity of a root extract of arisaema jacquemontii blume. journal of taibah university for science 9 (2015) 449-454. https://doi.org/10.1016/j.jtusci.2014.11.001 [10] s.g. dmitrienko, v.a. kudrinskaya, v.v. apyari. methods of extraction, preconcentration, and determination of quercetin. journal of analytical chemistry 67 (2012) 299-311. https://doi.org/10.11 34/s106193481204003x [11] k. csepregi, m. kocsis, é. hideg. on the spectrophotometric determination of total phenolic and flavonoid contents. acta biologica hungarica 64 (2013) 500-509. https://doi.org/10.1556/abi ol.64.2013.4.10 [12] l.a.l. da silva, b.r. pezzini, l. soares. spectrophotometric determination of the total flavonoid content in ocimumbasilicum l. (lamiaceae) leaves. pharmacognosy magazine 11 (2015) 96. https://doi.org/10.4103/0973-1296.149721 [13] o.v. trineeva. development of theoretical approaches to the determination of the main groups of biologically active substances of medicinal plant materials by tlc. development and registration of medicines 10 (2021) 69-79. (in russian). https://doi.org/10.333 80/2305-2066-2021-10-2-69-79 [14] v.s. saraswathi, d. saravanan, k. santhakumar. isolation of quercetin from the methanolic extract of lagerstroemia speciosa by hplc technique, its cytotoxicity against mcf-7 cells and photocatalytic activity. journal of photochemistry and photobiology b: biology 171 (2017) 20-26. https://doi.org/10.1016/j.jphotobiol.2017.04.031 [15] m. tzanova, p. peeva. rapid hplc method for simultaneous quantification of trans-resveratrol and quercetin in the skin of red grapes. food analytical methods 11 (2018) 514-521. https://doi.org/ 10.1007/s12161-017-1022-z [16] s. bajkacz, i. baranowska, b. buszewski, b. kowalski, m. ligor. determination of flavonoids and phenolic acids in plant materials using sle-spe-uhplc-ms/ms method. food analytical methods 11 (2018) 3563-3575. https://doi.org/10.1007/s12161-018-1332-9 [17] h. yin, j. ma, j. han, m. li, j. shang. pharmacokinetic comparison of quercetin, isoquercitrin, and quercetin-3-o-β-d-glucuronide in rats by hplc-ms. peerj 7 (2019) e6665. https://doi.org/10.7717/ peerj.6665 [18] a.f. memon, a.r. solangi, s.q. memon, a. mallah, n. memon, a.a. memon. simultaneous determination of quercetin, rutin, naringin, and naringenin in different fruits by capillary zone electrophoresis. food analytical methods 10 (2017) 83-91. https://doi.org/10.1007/s12161-016-0552-0 [19] s. zhang, s. dong, l. chi, p. he, q. wang, y. fang. simultaneous determination of flavonoids in chrysanthemum by capillary zone electrophoresis with running buffer modifiers. talanta 76 (2008) 780-784. https://doi.org/10.1016/j.talanta.2008.04.025 [20] s. siedler, s.g. stahlhut, s. malla, j. maury, a.r. neves. novel biosensors based on flavonoid-responsive transcriptional regulators introduced into escherichia coli. metabolic engineering 21 (2014) 2-8. https://doi.org/10.1016/j.ymben.2013.10.011 https://doi.org/10.5599/admet.1661 https://doi.org/10.1016/s1875-5364(14)60090-x https://doi.org/10.1016/s1875-5364(14)60090-x https://doi.org/10.1016/j.jff.2014.08.006 https://doi.org/10.1016/j.ajps.2017.08.004 https://doi.org/10.1016/j.jtusci.2014.11.001 https://doi.org/‌10.11‌34/s106193481204003x https://doi.org/‌10.11‌34/s106193481204003x https://doi.org/10.1556/abiol.64.2013.4.10 https://doi.org/10.1556/abiol.64.2013.4.10 https://doi.org/10.4103/0973-1296.149721 https://doi.org/10.333%2080/2305-2066-2021-10-2-69-79 https://doi.org/10.1016/j.jphotobiol.2017.04.031 https://doi.org/10.1007/s12161-017-1022-z https://doi.org/10.1007/s12161-017-1022-z https://doi.org/10.1007/s12161-018-1332-9 https://doi.org/10.7717/‌peerj.6665 https://doi.org/10.7717/‌peerj.6665 https://doi.org/10.1007/s12161-016-0552-0 https://doi.org/10.1016/j.talanta.2008.04.025 https://doi.org/10.1016/j.ymben.2013.10.011 m. v. kanevskiy et al. admet & dmpk 11(2) (2023) 277-291 290 [21] n.s. velichko, o.i. guliy, m.v. kanevsky, m.a. kupryashina, y.p. fedonenko. whole-cell electric sensor for determination of sodium dodecyl sulfate. world journal of microbiology and biotechnology 38 (2022) 1-13. https://doi.org/10.1007/s11274-022-03309-1 [22] s.e. fischer, m.j. miguel, g.b. morri. effect of root exudates on the polysaccharide composition and the lipopolysaccharide profile of azospirillum brasilense cd under saline stress. fems microbiology letters 219 (2003) 53-62. https://doi.org/10.1016/s0378-1097(02)01194-1 [23] m.v. kanevskiy, s.a. konnova, a.s. boyko, y.p. fedonenko, e.n. sigida, v.v. ignatov. effect of flavonoids on the composition of surface glycopolymers of azospirillum lipoferum sp59b. microbiology 83 (2014) 15-22. https://doi.org/10.1134/s0026261714020106 [24] v.l. baldani, j.i. baldani, j. döbereiner. effects of azospirillum inoculation on root infection and nitrogen incorporation in wheat. canadian journal of microbiology 29 (1983) 924-929. https://doi.org/10.1139/m83-148 [25] n.dos s. ferreira, f.h. sant’anna, v.m.reis, a.ambrosini, c.g. volpiano, m. rothballer, s. schwab, v.a. baura, e. balsanelli, f.de o. pedrosa, l.m.p. passaglia, e.m.de souza, a. hartmann, f. cassan, j.e. zill. genome-based reclassification of azospirillum brasilense sp245 as the type strain of azospirillum baldaniorum sp. nov. international journal of systematic and evolutionary microbiology 70 (2020) 6203-6212. https://doi.org/10.1099/ijsem.0.004517 [26] s.a. konnova, o.e. makarov, i.m. skvortsov, v.v. ignatov. isolation, fractionation and some properties of polysaccharides produced in a bound form by azospirillum brasilense and their possible involvement in azospirillum-wheat root interactions. fems microbiology letters 118 (1994) 93-99. https://doi.org/ 10.1111/j.1574-6968.1994.tb06809.x [27] g.a. o'toole, r. kolter. initiation of biofilm formation in pseudomonas fluorescens wcs365 proceeds via multiple, convergent signalling pathways: a genetic analysis. molecular microbiology 28 (1998) 449461. https://doi.org/10.1046/j.1365-2958.1998.00797.x [28] o.i. guliy, n.s. velichko, y.p. fedonenko, v.d. bunin. use of an electro-optical sensor and phage antibodies for immunodetection of herbaspirillum. talanta. 202 (2019) 362-368. https://doi.org/ 10.1016/j.talanta.2019.04.086 [29] o.i. guliy, v.d. bunin. biointerface engineering: prospects in medical diagnostics and drug delivery. springer, singapore, 2020 https://doi.org/10.1007/978-981-15-4790-4_11 [30] m.v. kanevskiy, v.v. shardin, v.d. bunin, o.i. guliy. electrophysical sensor systems for in vitro monitoring of bacterial metabolic activity. biosensors and bioelectronics: x. 11 (2022) 100179. https://doi.org/10.1016/j.biosx.2022.100179 [31] m. kanevskiy, e. borisova, i. mironova, s. konnova, a. galitskaya, a. khorovodov, i. agranovich, p. pavlova, o. semyachkina-glushkovskaya. stress-induced gastric adenocarcinoma: fluorescent and electrical measurements. journal of optoelectronics and advanced materials 22 (2020) 316-322. https://joam.inoe.ro/articles/stress-induced-gastric-adenocarcinoma-fluorescent-and-electrical-measurements/ [32] m. del gallo, a. haegi. characterization and quantification of exocellular polysaccharides in azospirillum brasilense and azospirillum lipoferum. symbiosis 9 (1990) 155-161. (https://dalspace.libra ry.dal.ca/bitstream/handle/10222/77099/volume%209-numbers%201-3-1990-page%20155.pdf ?sequence=1) [33] p. gerhardt. manual of methods for general bacteriology, american society for microbiology 1981. p. 524. (isbn 0914826298) [34] j.s. sawardeker, j.h. sloneker, a. jeanes. quantitative determination of monosaccharides as their alditol acetates by gas liquid chromatography. analytical chemistry 37 (1965) 1602-1604. https:// doi.org/10.1021/ac60231a048 [35] o.i. gulii, l.y. matora, g.l. burygin, l.a. dykman, v.v. ignatov, o.v. ignatov. electro-optical properties of the microbial suspensions during a cell’s interaction with the antibodies of a different specificity. applied biochemistry and microbiology 46 (2010) 61-64. https://doi.org/10.1134/s0003683810010102 https://doi.org/10.1007/s11274-022-03309-1 https://doi.org/10.1016/s0378-1097(02)01194-1 https://doi.org/10.1134/s0026261714020106 https://doi.org/10.1139/m83-148 https://doi.org/10.1099/ijsem.0.004517 https://doi.org/10.1111/j.1574-6968.1994.tb06809.x https://doi.org/10.1111/j.1574-6968.1994.tb06809.x https://doi.org/10.1046/j.1365-2958.1998.00797.x https://doi.org/‌10.1016/j.talanta.2019.04.086 https://doi.org/‌10.1016/j.talanta.2019.04.086 https://doi.org/10.1007/978-981-15-4790-4_11 https://doi.org/10.1016/j.biosx.2022.100179 https://joam.inoe.ro/articles/stress-induced-gastric-adenocarcinoma-fluorescent-and-electrical-measurements/ https://joam.inoe.ro/articles/stress-induced-gastric-adenocarcinoma-fluorescent-and-electrical-measurements/ https://dalspace.library.dal.ca/bitstream/handle/10222/77099/volume%209-numbers%201-3-1990-page%20155.pdf?sequence=1 https://dalspace.library.dal.ca/bitstream/handle/10222/77099/volume%209-numbers%201-3-1990-page%20155.pdf?sequence=1 https://dalspace.library.dal.ca/bitstream/handle/10222/77099/volume%209-numbers%201-3-1990-page%20155.pdf?sequence=1 https://doi.org/‌10.1021/ac60231a048 https://doi.org/‌10.1021/ac60231a048 https://doi.org/10.1134/s0003683810010102 admet & dmpk 11(2) (2023) 277-291 quercetin detection with azospirillum cells doi: https://doi.org/10.5599/admet.1661 291 [36] m.a. matus-cádiz, t.e. daskalchuk, b. verma, d. puttick, r.n. chibbar, g.r. gray, c.e. perron, r.t. tyler, p. hucl. phenolic compounds contribute to dark bran pigmentation in hard white wheat. journal of agricultural and food chemistry 56 (2008) 1644-1653.https://doi.org/10.1021/jf072970c [37] s. cesco, t. mimmo, g. tonon, n. tomasi, r. pinton, r. terzano, g. neumann, l. weisskopf, g. renella, l. landi, p. nannipiery. plant-borne flavonoids released into the rhizosphere: impact on soil bioactivities related to plant nutrition. biology and fertility of soils 48 (2012) 123-149. https://doi.org/ 10.1007/s00374-011-0653-2 [38] p.h. bais, t.l. weir, l.g. perry, s. gilroy, j.m. vivanco. the role of root exudates in rhizosphere interactions with plants and other organisms. annual review of plant biology 57 (2006) 233-66. https://doi.org/10.1146/annurev.arplant.57.032905.105159 [39] c.h. kong, h. zhao, x.h. xu, p. wang, y. gu. activityandallelopathyofsoilof flavones o-glycosides from rice. journal of agricultural and food chemistry 55 (2007) 6007-6012. https://doi.org/10.10 21/jf0703912 [40] s. hassan, u. mathesius. the role of flavonoids in root–rhizospheresignalling: opportunities and challenges for improving plant–microbe interactions. journal of experimental botany 63 (2012) 34293444. https://doi.org/10.1093/jxb/err430 [41] w.l. johnson, d.c. france, n.s. rentz, w.t. cordell, f.l. walls. sensing bacterial vibrations and early response to antibiotics with phase noise of a resonant crystal. scientific reports 7 (2017) 1-12. https://doi.org/10.1038/s41598-017-12063-6 [42] o.i. guliy, v.d. bunin, v.i. korzhenevich, o.v. ignatov. electro-optical assays for immunoindication of microbial cells. current immunology reviews 13 (2017) 153-162. https://doi.org/10.2174/1573395513 666171010142039 [43] a.m. james. the electrical properties and topochemistry of bacterial cells. advances in colloid and interface science 15 (1982) 171-221. https://doi.org/10.1016/0001-8686(82)80001-8 [44] m.v. kanevsky, v.o. menukhov, i.s. kosheleva, a.yu. kostritsky, i.v. kanevskaya, s.a. konnova. changes in the physicochemical and cultural properties of the bacteria azospirillum baldaniorum sp245 under the influence of some synthetic coumarins. izvestiya of saratov university. chemistry. biology. ecology 22 (2022) 215-225. (in russian). https://doi.org/10.18500/1816-9775-2022-22-2-215-225 [45] s.s. yevstigneyeva, e.n. sigida, y.p. fedonenko, s.a. konnova, v.v. ignatov. structural properties of capsular and o-specific polysaccharides of azospirillum brasilense sp245 under varying cultivation conditions. microbiology 85 (2016) 664-671. https://doi.org/10.1134/s0026261716060096 [46] a.a. maltseva, o. v. trineeva, a. s. chistyakova, t. a. brezhneva, a. i. slivkin, a. a. sorokina. thin-layer chromatography in the analysis of flavonoids of plant objects. pharmacy 1 (2013) 13-16. (in russian) (https://www.elibrary.ru/item.asp?id=18891084) [47] s.l. hsiu, c.w. tsao, y.c. tsai, h.j. ho, p.d.l. chao. determinations of morin, quercetin and their conjugate metabolites in serum. biological and pharmaceutical bulletin 24 (2001) 967-969. https://doi.org/10.1248/bpb.24.967 [48] k. ishii, t. uruta, y. kasuya. high-performance liquid chromatographic determination of quercetin in human plasma and urine utilizing solid-phase extraction and ultraviolet detection. journal of chromatography b. 794 (2003) 49-56. https://doi.org/10.1016/s1570-0232(03)00398-2 [49] e.j. lee, k.s. yoo, b.s. patil. development of a rapid hplc‐uv method for simultaneous quantification of protodioscin and rutin in white and green asparagus spears. journal of food science 75 (2010) 703709. https://doi.org/10.1111/j.1750-3841.2010.01824.x [50] n.b. sarikahya, a.c. goren, s. kirmizigul. simultaneous determination of several flavonoids and phenolic compounds in nineteen different cephalaria species by hplc-ms/ms. journal of pharmaceutical and biomedical analysis 173 (2019) 120-125. https://doi.org/10.1016/j.jpba.20 19.05.019 ©2023 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) https://doi.org/10.5599/admet.1661 https://doi.org/10.1021/jf072970c https://doi.org/10.1007/s00374-011-0653-2 https://doi.org/10.1007/s00374-011-0653-2 https://doi.org/10.1146/annurev.arplant.57.032905.105159 https://doi.org/10.1021/jf0703912 https://doi.org/10.1021/jf0703912 https://doi.org/10.1093/jxb/err430 https://doi.org/10.1038/s41598-017-12063-6 https://doi.org/10.2174/1573‌395513‌666171010142039 https://doi.org/10.2174/1573‌395513‌666171010142039 https://doi.org/10.1016/0001-8686(82)80001-8 https://doi.org/10.18500/1816-9775-2022-22-2-215-225 https://doi.org/10.1134/s0026261716060096 https://www.elibrary.ru/item.asp?id=18891084 https://doi.org/10.1248/bpb.24.967 https://doi.org/10.1016/s1570-0232(03)00398-2 https://doi.org/10.1111/j.1750-3841.2010.01824.x https://doi.org/10.1016/j.jpba.2019.05.019 https://doi.org/10.1016/j.jpba.2019.05.019 http://creativecommons.org/licenses/by/3.0/ electrochemical sensor for determination of butylated hydroxyanisole in real samples using glassy carbon electrode modified by [co(hl)2cl2] nano-complex doi: http://dx.doi.org/10.5599/admet.1703 185 admet & dmpk 11(2) (2023) 185-199; doi: https://doi.org/10.5599/admet.1703 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper electrochemical sensor for determination of butylated hydroxyanisole in real samples using glassy carbon electrode modified by [co(hl)2cl2] nano-complex mahbubeh fazli and niloufar akbarzadeh-t* department of chemistry, university of sistan and baluchestan, p.o. box 98135-674, zahedan, iran *corresponding author: e-mail: n.akbarzadeh@chem.usb.ac.ir received: february 07, 2023; revised: february 22, 2023; published: march 2, 2023 abstract a new mononuclear co(ii) complex with the formula [co(hl)2cl2] (1) (hl= n-(2-hydroxy-1-naphthylidene)2-methyl aniline) has been synthesized and characterized by fourier transform infrared spectroscopy, uv– vis, elemental analysis and single crystal x-ray structure analysis. single crystals of the complex [co(hl)2cl2] (1) were obtained through slow evaporation of an acetonitrile solution at room temperature. the crystal structure analysis revealed that the two schiff base ligands create a tetrahedral geometry via oxygen atoms and two chloride atoms. the nano-size of [co(hl)2cl2] (2) have been synthesized by the sonochemical process. characterization of nanoparticles (2) was carried out via x-ray powder diffraction (xrd), scanning electron microscopy (sem), uv-vis, and ft-ir spectroscopy. the average sample size synthesized via the sonochemical method was approximately 56 nm. in this work, a simple sensor based on a glassy carbon electrode modified with [co(hl)2cl2] nano-complex was developed ([co(hl)2cl2] nano-complex/gce) for convenient and fast electrochemical detection of butylated hydroxyanisole (bha). the modified electrode offers considerably improved voltammetric sensitivity toward bha compared to the bare electrode. applying linear differential pulse voltammetry, a good linear relationship of the oxidation peak current with respect to concentrations of bha across the range of 0.5–150 µm and a detection limit of 0.12 µm was achieved. the [co(hl)2cl2] nano-complex/gce sensor was applied to the determination of bha in real samples successfully. ©2023 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords schiff base ligand; cobalt(ii) nano-complex; butylated hydroxyanisole; voltammetry introduction butylated hydroxyanisole, a substituted phenolic compound, is one of the potent chemical antioxidants used as a preservative in food, food packaging, animal feed, cosmetics, pharmaceutical preparation, rubber and petroleum products. it is also used as a vitamin a stabilizer. bha is added to food products to improve their stability, especially to prevent rancidity in products containing lipids or fats. bha is exclusively used in food products because of its ability to remain active even at high temperatures while cooking or baking [1-3]. although they are powerful in protecting product quality in food distribution, excess antioxidants added to food might produce toxicities or mutagenicities and, thus, endanger people's health [4,5]. thus, it is necessary to reliably determine the amounts of bha in food or other products. http://dx.doi.org/10.5599/admet.1703 https://doi.org/10.5599/admet.1703 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:n.akbarzadeh@chem.usb.ac.ir http://creativecommons.org/licenses/by/4.0/ m. fazli and n. akbarzadeh-t. admet & dmpk 11(2) (2023) 185-199 186 many methods, including spectrophotometry [6], gas chromatography [7], high-performance liquid chromatography [8], micellar electrokinetic capillary chromatography [9] and electrochemical methods [10,11] have been developed for the determination of bha. electrochemical sensors with excellent ability for the determination of electroactive substances, especially food samples, have been suggested as a powerful analytical tool in recent years [12-18]. electrochemical sensors are fast, portable and affordable. however, in many cases, the detection limit was insufficient when traces of these compounds had to be determined. the results might also be irreproducible due to surface fouling, requiring high oxidation potentials and being time-consuming [19-26]. chemically modified electrodes (cmes) improve the sensitivity and selectivity of electrochemical analysis by improving the kinetics of the reaction via an electrocatalytic process at the cme surface [24-35]. recently, more attention has been focused on the synthesis and application of nanoparticles (nps), and on properties of nps such as high electrical conductivity, high surface area, and chemical stability [36-40]. the nps can be used to promote electron transfer reactions when it is used as electrode material in an electrochemical device. therefore, the modification of the electrochemical interface by nanostructures is one of the recent approaches used extensively in the development of sensing platforms [41,43]. schiff base ligands are among the most common coordinating ligands in coordination chemistry [44], through a direct forward condensation reaction between an amine and a carbonyl compound [45]. during the past few decades, considerable attention has been paid to the chemistry of metal complexes of schiff bases containing nitrogen and other donor atoms [46]. this may be attributed to their stability, biological activity and potential applications in many fields, such as oxidation catalysis, electrochemistry [47], and biological properties, such as antibacterial, antifungal and antioxidant activities [48,49]. in general, the antibacterial and antifungal activities of schiff bases are attributed to the presence of lone pair of electrons on the nitrogen atom and the electron-donating character of the double bond of the azomethine group [50]. facile and inexpensive synthesis, allied with the wide range of structural and electronic features of schiff bases and their coordination complexes, has increased interest in chemically modifying electrodes with these compounds. therefore, the resulting electrodes have been used as sensors and/or probes in a variety of fields. all this interest has culminated in the development of low-cost sensors. in particular, the many structural and electronic properties of schiff bases and their transition metal nano-complexes allow them and the analyte to establish different interactions, which improves sensor selectivity and sensitivity [51-56]. in this work, schiff base ligand [hl] and its cobalt(ii) complex (1) with (n-(2-hydoxynaphtalen)2-methyl aniline) schiff base ligand were synthesized. the complex was characterized using spectroscopic techniques such as ft-ir, uv-vis and elemental analysis. the crystal structure of title complex was determined by x-ray crystallography. the schiff base cobalt nano-complex (2) was prepared by the sonochemical method. this synthesis was performed without the use of any surfactants. the nano-sized complex (2) was identified using xrd and sem techniques. also, we developed a voltammetric sensor based on [co(hl)2cl2] nano-complex modified glassy carbon electrode for the detection of bha. under optimal conditions, the sensor displayed excellent sensitivity, low detection limit and rapid response toward the analyte. moreover, this developed sensor was applied in the determination of bha in real samples successfully experimental materials and instrumentation 1h-nmr spectrum was accomplished with bruker avance3 3-300mhz spectrometer using dmso-d6 as solvent and at room temperature (298 k). fourier-transform infrared spectra were performed on a perkinelmer ft-ir spectrophotometer model spectrum two with kbr discs in the range of 4000-400 cm-1. for https://en.wikipedia.org/wiki/fourier-transform_infrared_spectroscopy admet & dmpk 11(2) (2023) 185-199 butylated hydorsyanisole determination using modified gce doi: http://dx.doi.org/10.5599/admet.1703 187 elemental analyses of new compounds, a costech ecs 4010 chns elemental analyzer was utilized. uv-vis spectra were recorded in the range of 200–800 nm on optizen view 2120uvplus spectrophotometer ver1-2. single crystal measurements were done with rigaku od supernova, equipped with atlas s2 ccd detector and mo-kα radiation from a micro-focused sealed x-ray tube. the data reduction, scaling and absorption correction were made in crysalis pro. mce was used for the visualization of electron density maps. all the electrochemical measurements were carried out on a pgstat302n potentiostat/galvanostat autolab consisting of a traditional three-electrode system: a bare or modified glassy carbon electrode (gce) as the working electrode, an ag/agcl as the auxiliary electrode and a pt wire as the counter electrode. solution ph values were determined using a 713 ph meter combined with a glass electrode (metrohm, switzerland). bha and other chemicals used were analytical grade and were purchased from sigma aldrich and merck. orthophosphoric acid was used to prepare the phosphate buffer solutions (pbss), and sodium hydroxide was used to adjust the desired ph values (ph range between 2.0 and 9.0). synthesis of schiff base ligand [hl] the ligand (hl: n-(2-hydroxy-1-naphthylidene)-2-methyl aniline) was synthesized by adding a solution of 2-hydroxy-1-naphthaldehyde (0.17 g, 1 mmol) in 10 ml of ethanol to a solution of 2-mehyl aniline (0.11 ml, 1 mmol). then, the mixture was refluxed for 6 h. the resulting yellow solution was placed at room temperature. after 24 h, yellow crystals of [hl] were formed. the schiff base ligand was characterized by 1hnmr, ft-ir, elemental analysis and single-crystal x-ray diffraction: mp:(154 °c);(yield: 88 %);(mw: 261.31); anal. calc. for [c18h15no]: c82.73;h6.12; n 5.79; found: c82.70; h6.09; n5.76. ftir (kbr, cm-1), 3421ν (n-h), 1620 ν(c=n), 1161 ν(c-o), 1479 ν(c=c); 1h-nmr (300 mhz, dmso d6, ppm) 9.61 (1h, s, ch=n), 16.05 (1h, oh), 2.40 (3h,ch3) [37]. synthesis of complex (1) the schiff base ligand (hl: n-(2-hydroxy-1-naphthylidene)-2-methyl aniline) (0.26 g,1 mmol) was dissolved in 10 ml of methanol and to this yellow solution, a solution of [cocl2.6h2o] (0.12 g, 0.5mmol) in 10 ml of methanol was added dropwise. the reaction mixture was refluxed for 6 h. the complex was precipitated, and the precipitate was filtered and dried at room temperature. green crystals of [co(hl)2cl2] complex suitable for single crystal x-ray diffraction were obtained by the crystallization from the acetonitrile solution of the precipitate after two weeks. the complex was characterized by uv-vis, ftir, elemental analysis and singlecrystal x-ray diffraction (scheme 1). mp:250 °c; (yield: 85 %); (mw: 652.5); anal. calc. for [c36h30cl2con2o2]: c66.27; h4.63; n4.29; found: c66.32; h4.69; n4.27. ft-ir (kbr, cm-1), 1619 ν(c=n); 1146 ν(c-o), 535 ν(co-o). scheme1. synthetic route for complex (1). synthesis of nanocomplex (2) a solution of cocl2.6h2o (0.12 g, 0.5 mmol) in methanol (10 ml) was positioned in a high-density ultrasonic probe for 10 min. then to this solution, 10 ml of methanolic solution of schiff base ligand (0.26 g, 1 mmol) was added dropwise. the resultant solution was then irradiated for 60 min at 60 °c with a power of 100 w. http://dx.doi.org/10.5599/admet.1703 https://en.wikipedia.org/wiki/fourier-transform_infrared_spectroscopy https://en.wikipedia.org/wiki/fourier-transform_infrared_spectroscopy https://en.wikipedia.org/wiki/fourier-transform_infrared_spectroscopy https://en.wikipedia.org/wiki/fourier-transform_infrared_spectroscopy https://en.wikipedia.org/wiki/fourier-transform_infrared_spectroscopy https://en.wikipedia.org/wiki/fourier-transform_infrared_spectroscopy m. fazli and n. akbarzadeh-t. admet & dmpk 11(2) (2023) 185-199 188 the obtained brown precipitate was filtered and dried in air. mp: 245 °c; (yield: 87 %); (mw: 652.5); anal. calc. for [c36h30cocl2n2o2]: c 66.27; h 4.63; n 4.29; found: c 66.34; h, 4.69; n,4.27. ft-ir (kbr, cm-1): 1621 ν(c=n); 1147 ν(c-o), 534ν (co-o). preparation of [co(hl)2cl2] nano-complex/gce the bare glassy carbon electrode was coated with [co(hl)2cl2] nano-complex according to the following simple procedure. 1 mg [co(hl)2cl2] nano-complex was dispersed in 1 ml aqueous solution within 50 min ultrasonication. then, 4 µl of the prepared suspension was dropped on the surface of carbon working electrodes. it remains at room temperature until it becomes dry. the surface areas of the [co(hl)2cl2] nanocomplex/gce and the un-modified gce were obtained by cv using 1 mm k3fe(cn)6 at various scan rates. using the randles–ševčik equation for [co(hl)2cl2] nano-complex/gce, the electrode surface was found to be 0.109 cm2 which was about 3.5 times greater than un-modified gce. results and discussion single crystal x-ray diffraction green crystals of [co(hl)2cl2] (1) were obtained by slow evaporation of an acetonitrile solution. its structure was determined by single-crystal x-ray diffraction. the complex crystallized in the monoclinic system in space group c2/c with an asymmetric unit consisting of a metal cation in a special position, one chloride anion and one disordered hl ligand. table 1 shows the crystallographic data for complex (1). all hydrogen atoms of the strongly occupied part of ligands were discernible in difference fourier maps and could be refined to reasonable geometry. according to common practice, h atoms bonded to c were kept in ideal positions with c–h = 0.96 å, while positions bonded to n were refined with restrained bond length. in both cases, uiso(h) was set to 1.2ueq(c, n). all non-hydrogen atoms were refined using harmonic refinement. a disordered ligand was observed in both structures. given the scale of disorder and numerous overlaps, we decided to use molecular refinement to model the disorders. two models were used to allow for free rotation of the tolyl group. the occupancies of two positions were refined with the sum constrained to 1. resulting occupancy ratios were 832(2):168(2). for further information on data collection and refinement, see table 1. a view of compound (1) is shown in figure 1. the occupied position is shown in figure 2. figure 1. the labeled diagram of compound (1). thermal ellipsoids are at 50% probability level symmetry code: (i) 1-x, y, 0.5-z. admet & dmpk 11(2) (2023) 185-199 butylated hydorsyanisole determination using modified gce doi: http://dx.doi.org/10.5599/admet.1703 189 table1. crystallographic data and refinement parameters for complex (1) compound (1) figure 2. disorder in complex (1), weakly occupied position depicted in blue with dashed bonds. formula c36h30cl2con2o2 formula weight (g mol-1) 652.5 λ (å) 0.71073 crystal system monoclinic space group c 2/c hall group -c 2yc a (å) 24.3845(7) b (å) 7.5848(3) c (å) 17.9596(9) β (°) 114.763(3) t (k) 95 v (å3) 3016.2(2) z 4 dcalc (g cm-3) 1.437 μ (mm-1) 0.784 f(000) 1348 h,k,lmax 33,10,15 θmax (°) 29.34 measured refl. 7862 indep. refl. (rint) 3594 (0.018) obs. refl. (i>3σ(i)) 3110 r1(obs) 0.0343 wr2(all) 0.0998 s 1.510 parameters 211 δρmax, δρmin (e å-3) 0.35, -0.36 ccdc 1995453 the cation in complex (1) is four-coordinate with two chloride anions and two phenolic oxygen atoms forming a slightly distorted tetrahedron. only the oxygen atom of the neutral zwitterionic hl ligand is coordinated with the metal cation, while the presence of an azomethine proton denies the possibility of coordination via the nitrogen atom. a list of the most important bond distances and angles is reported in table 2. table 2. selected bond lengths and angles in structures (1) and (2) x–y (1) x–y (å) x–y–z (1) x–y–z (°) co–cl1 2.2505(5) cl1–co–cl1i 115.11(2) co–cl1i 2.2505(5) cl1–co–o1a 104.82(6) co–o1a 1.9618(16) cl1–co–o1ai 110.14(5) co–o1ai 1.9618(16) cl1i–co–o1a 110.14(5) co–o1b 2.122(16) cl1i–co–o1ai 104.82(6) co–o1bi 2.122(16) o1a–co–o1ai 111.97(7) cl1–co–o1b 114.6(2) cl1–co–o1bi 99.06(19) cl1i–co–o1b 99.06(19) cl1i–co–o1bi 114.6(2) o1b–co–o1bi 115.4(3) symmetry code: (i) 1-x, y, 0.5-z. the coordination tetrahedron formed by chloride anions and the ligand with higher occupancy has only minor distortions from the ideal values with co–cl1 bond lengths at 2.2505(5) å and co–o1a bond lengths at http://dx.doi.org/10.5599/admet.1703 m. fazli and n. akbarzadeh-t. admet & dmpk 11(2) (2023) 185-199 190 1.9618(16) å. the bond angles are in the range from 104.82(6)° to 115.11(2)°. the coordination of weakly occupied oxygen atoms appears weaker compared to the oxygen atoms with higher occupancy, with co–o1b bonding longer by 0.1602 å compared to co–o1a. this elongation brings the bond length closer to the length of co–cl1 bond, which should result in lower distortions from the ideal tetrahedron. regardless of the similarities in the bond lengths, the distortions in the coordination tetrahedron formed with the weakly occupied ligand were even more pronounced, with the angles ranging from 99.06(19)° to 115.4(3)°. ft-ir spectra the ft-ir analysis data is listed in table 3. the ft-ir spectrum of the free schiff base ligand [hl] exhibits a band in 3421 due to ν(nh+) and a band in 1620 cm-1 due to ν(c=n) azomethine. a band at 1161 cm-1 is assigned to the ν(c-o) phenolic group. this band has been shifted to lower frequencies for both complex (1) and nano-complex (2), in the region of 1146-1147 cm-1, which indicates that both compounds are formed by the coordination of the oxygen atom of [hl] to the metal ion (figure 3)s. ft-ir spectra of both complex (1) and nanocomplex (2) show weak bands at 535 cm-1 and 534 cm-1 that assign to (co-o), and don't show any shift due to ν(c=n) and ν(nh+) that this display the azomethine group of [hl] doesn't participate in complex formation [57]. figure 3. the ft-ir spectra of the complex (1) and its nanoparticles (a2) table 3. selected ir frequencies (cm−1) of [hl], (1) and (2). compound ν(n-h) ν(c=n) ν(c-o) ν(co-o) hl 3421 1620 1161 co(hl)2cl2 (1) 3425 1619 1146 535 co(hl)2cl2 (2) 3425 1621 1147 534 hl:n-(2-hydroxy-1-naphthylidene)-2-methyl aniline uv-vis spectra uv-vis spectra of [hl], (1) and (2) in methanol solution contain different peaks related to the transition bands data. these absorption bands (nm) are listed in table 2. the schiff base ligand display two bands at 230 and 250 nm attributed to π →π* and a band at 330 nm that is assigned to n→π* transitions. in the electronic spectra of both (1) and (2), intra-ligand transitions (n→π* and π →π*) were shifted to another wavelength that this is due to the coordination of the metal to the ligands. both of electronic spectra of complex (1) and nano-complex (2) showed three d-d absorption bands at 520-635 nm are assigned to the transition 4a2→4t2(f),4a2→4t1(f) and 4a2→4t1(p), respectively [58] (figures 4 and 5). admet & dmpk 11(2) (2023) 185-199 butylated hydorsyanisole determination using modified gce doi: http://dx.doi.org/10.5599/admet.1703 191 figure 4. the spectrum of uv-vis of (1) in methanol. figure 5. the spectrum of uv-vis of (2) in methanol. xrd and sem the xrd patterns of nano-sized co(ii) complex (2) and standard powder co(ii) complex (1) were obtained from single-crystal x-ray diffraction and are exhibited in figure 6. the xrd pattern shows the crystalline phase and the nature of the complex. by investigating the location and intensity of the peaks of both patterns, it can be deduced that the diffraction angle in both complexes obtained by different methods is the same [59]. this indicates that the nano-sized complex (2) has a single crystalline phase that this phase is similar to that obtained by single-crystal x-ray diffraction [60]. the width of the diffraction peaks shows the nanocrystal complex (2) particles are of nanometer scales [61]. the particle means the size of the nanocrystal complex (2) was calculated using the debye scherrer equation. sem images of the nanocrystal complex (2) are shown in figure 7. the sem photos also show the shape of the nanoparticles and the surface morphology of the nanocrystal complex. the average size diameter obtained from the debye-scherrer equation of nanocrystal complex (2) was approximately 56 nm. figure 6. the xrd patterns of a: standard powder co(ii) complex (1) b: nano-sized co(ii) complex (2) figure 7. sem images of the nanocrystal complex (2) http://dx.doi.org/10.5599/admet.1703 m. fazli and n. akbarzadeh-t. admet & dmpk 11(2) (2023) 185-199 192 electrochemical behavior of bha at the surface of various electrodes according to our knowledge, the electrooxidation of bha is closely related to the ph value of the solution (scheme 2). so, the effect of ph was investigated using the differential pulse voltammetry (dpv) method. the results show that the oxidation peak current increased from ph 2.0 to 7.0, and then the current conversely decreased when the ph value increased from 7.0 to 9.0. according to obtained results, ph 7.0 was chosen as the best optimal experimental condition for other experiments. scheme 2. the proposed mechanism for the oxidation of bha at the [co(hl)2cl2] nano-complex/gce. the electrochemical behavior of bha was investigated by linear sweep voltammetry (lsv). the linear sweep voltammograms obtained using the bare gce and [co(hl)2cl2] nano-complex/gce in 0.1 m pbs (ph 7.0) in the presence of 50.0 μm bha are shown in figure 8. at the bare cpe, a weak oxidation peak current (ipa = 3.6 μa) could be seen at 0.57 v. in contrast, [co(hl)2cl2] nano-complex/gce exhibited an enhanced sharp anodic peak current (ipa =9.4 μa) at much lower overpotential ep = 0.43 v. these results confirmed that the [co(hl)2cl2] nano-complex improved the sensitivity of the modified electrode by enhancing peak current and decreasing the overpotential of the oxidation of bha. effect of scan rate on the determination of bha at [co(hl)2cl2] nano-complex/gce the influence of the scan rate (ʋ) on the peak currents (ipa) of bha at [co(hl)2cl2] nano-complex/gce was investigated by lsv. figure 9 shows the voltammetric response of 50.0 μm bha at [co(hl)2cl2] nanocomplex/gce at different scan rates in the range of 10 to 400 mv/s. the oxidation peak current of bha increases linearly with increasing scan rate. a linear regression equation was obtained from the plot ipa and vs. ʋ1/2 (square root of scan rate) as follows; ipa (μa) = 1.3684 ʋ1/2 (mv/s)2 – 0.5184 (r2 = 0.9995) for the oxidation process, which indicates that the reaction of bha at [co(hl)2cl2] nano-complex/gce is diffusion controlled. in order to obtain some information on the rate-determining step, we drew a tafel plot (figure 10) using the data from the rising part of the current-voltage curve recorded at a low scan rate of 10 mv s-1 for 50.0 μm bha. the linearity of the e versus log i plot implies the intervention of the kinetics of the electrode process. the slope of this plot can be used to estimate the number of electrons transferred in the ratedetermining step. according to figure 10 inset, the tafel slope for the linear part of the plot was estimated to be equal to 0.1316 v. the value of the tafel slope indicates that the one-electron transfer process is the rate-limiting step, assuming a transfer coefficient (α) of about 0.55. admet & dmpk 11(2) (2023) 185-199 butylated hydorsyanisole determination using modified gce doi: http://dx.doi.org/10.5599/admet.1703 193 figure 8. the linear sweep voltammograms of (a) bare gce and (b) [co(hl)2cl2] nano-complex/gce in 0.1 m pbs (ph 7.0) in the presence of 50.0 μm bha at the scan rate 50 mvs-1. figure 9. linear sweep voltammograms of [co(hl)2cl2] nano-complex/gce in 0.1 m pbs (ph 7.0) containing 50.0 μm bha at various scan rates; 1-6 correspond to 10, 50, 100, 200, 300, and 400 mv s-1, respectively. inset: variation of anodic peak current vs. ν1/2. figure 10. linear sweep voltammograms response for 50.0 μm bha with 10 mvs-1 scan rate. inset: the tafel plot derived from the rising part or the corresponding voltammogram. http://dx.doi.org/10.5599/admet.1703 m. fazli and n. akbarzadeh-t. admet & dmpk 11(2) (2023) 185-199 194 chronoamperometric analysis the analysis of chronoamperometry for bha samples was performed by use of [co(hl)2cl2] nanocomplex/gce vs. ag/agcl/kcl (3.0 m) at 0.48 v. the chronoamperometric results of different concentrations of bha in pbs (ph 7.0) are demonstrated in figure 11. the cottrell equation for the chronoamperometric analysis of electroactive moieties under mass transfer limited conditions is as follow: i = nfad1/2cbπ-1/2t-1/2 where d represents the diffusion coefficient (cm2 s-1), and cb is the applied bulk concentration (mol cm-3). experimental results of i vs. t-1/2 were plotted in figure 11a, with the best fits for different concentrations of bha. the resulting slopes, corresponding to straight lines in figure 11a, were then plotted against the concentration of bha (figure 11b). the mean value of d was determined to be 8.2 × 10-5 cm2/s according to the resulting slope and cottrell equation. figure 11. chronoamperograms obtained at [co(hl)2cl2] nano-complex/gce in 0.1 m pbs (ph 7.0) for different concentration of bha. the 1-4 correspond to 0.1, 0.5, 0.9, and 2.0 mm of bha. insets: (a) plots of i vs. t-1/2 obtained from chronoamperograms 1-4. (b) plot of the slope of the straight lines against bha concentration. calibration curve because dpv commonly has a higher sensitivity than cv, the dpv technique was applied for the quantitative detection of bha. figure 12 shows the differential pulse voltammograms of bha at various concentrations using [co(hl)2cl2] nano-complex/gce. as seen, the oxidation peak currents of bha enhance gradually by increasing its concentration. the oxidation peak currents (ipa) show a good linear relationship with the concentrations of bha ranging from 0.5 m to 150.0 μm. the linear equation is ipa/μa = 0.171cbha/μm + 0.4624 (r2 = 0.9994) (figure 12 (inset)). also, the limit of detection, cm, of bha was calculated using the following equation: cm=3sb/m where, m is the slope of the calibration plot (0.171 μa/ μm) and sb is the standard deviation of the blank response obtained from 15 replicate measurements of the blank solution. the detection limit for determination of bha using this method of 0.12 μm was obtained. admet & dmpk 11(2) (2023) 185-199 butylated hydorsyanisole determination using modified gce doi: http://dx.doi.org/10.5599/admet.1703 195 figure 12. dpvs of [co(hl)2cl2] nano-complex/gce in 0.1 m (ph 7.0) containing different concentrations of bha. numbers 1–7 correspond to 0.5, 5.0, 15.0, 45.0, 75.0, 100.0, and 150.0 µm of bha. inset: plot of the electrocatalytic peak current as a function of bha concentration in the range of 0.5-150.0 µm. conclusions we have reported the synthesis of a schiff base ligand and a new co(ii) complex (1) . the [hl]and its cobalt(ii) complex (1) were characterized by spectroscopic techniques and elemental analysis. the single crystal x-ray diffraction analysis of the complex showed that metal ions reacted with the ligand in a 1:2 molar ratio. in the formation of the complex, oxygen atoms of two ligands are coordinated to the metal ion, and tetrahedral geometry is formed around the metal ion with two chloride ions attached to it. also, the nanocomplex (2) have been synthesized by the sonochemical process. the nanocomplex (2) was characterized via x-ray powder diffraction (xrd) and sem. the xrd patterns indicated both (1) and (2) compounds prepared by different synthesis methods have the same crystal structure. a sensitive and fast voltammetric method to detect the bha based on co(hl)2cl2] nano-complex (2) modified glassy carbon electrode was established. the voltammetric investigation demonstrates that electrooxidation of bha at the surface of [co(hl)2cl2] nano-complex/gce showed very distinct characteristics due to the presence of nanoparticles of [co(hl)2cl2] complex on the surface of electrode. [co(hl)2cl2] nano-complex/gce exhibited good catalytic activity towards the oxidation of bha over a linear range from 0.5 to 150.0 μm with an enhanced sensitivity of 0.171 μa/μm of bha and favourable characteristics for the determination of bha with a detection limit of 0.12 μm. finally, the proposed method was successfully applied in the determination of bha in real samples with satisfactory results. acknowledgment: the authors sincerely thank the university of sistan and baluchestan for providing financial support for this work. conflict of interest: the authors declare no conflict of interest. references [1] k.h.g. freitas, o. fatibello-filho. simultaneous determination of butylated hydroxyanisole (bha) and butylated hydroxytoluene (bht) in food samples using a carbon composite electrode modified with cu3(po4)2 immobilized in polyester resin. talanta 81 (2010) 1102-1108. https://doi.org/10.1016/j.talanta.2010.02.004 [2] d. thomas, z. rasheed, j.s. jagan, k.g. kumar. study of kinetic parameters and development of a voltammetric sensor for the determination of butylated hydroxyanisole (bha) in oil samples. journal of food science and technology 52 (2015) 6719-6726. https://doi.org/10.1007/s13197-015-1796-1 http://dx.doi.org/10.5599/admet.1703 https://doi.org/10.1016/j.talanta.2010.02.004 https://doi.org/10.1016/j.talanta.2010.02.004 https://doi.org/10.1007/s13197-015-1796-1 m. fazli and n. akbarzadeh-t. admet & dmpk 11(2) (2023) 185-199 196 [3] s.r. prabakar, s.s. narayanan. surface modification of amine-functionalised graphite for preparation of cobalt hexacyanoferrate (cohcf)-modified electrode: an amperometric sensor for determination of butylated hydroxyanisole (bha). analytical and bioanalytical chemistry 386 (2006) 2107-2115. https://doi.org/10.1007/s00216-006-0854-y [4] m. cui, s. liu, w. lian, j. li, w. xu, j. huang. a molecularly-imprinted electrochemical sensor based on a graphene–prussian blue composite-modified glassy carbon electrode for the detection of butylated hydroxyanisole in foodstuffs. analyst 138 (2013) 5949-5955. https://doi.org/10.1039/c3an01190a [5] z. rasheed, a.e. vikraman, d. thomas, j.s. jagan, k.g. kumar. carbon-nanotube-based sensor for the determination of butylated hydroxyanisole in food samples. food analytical methods 8 (2015) 213221. https://doi.org/10.1007/s12161-014-9894-7 [6] l.f. capitán‐vallvey, m.c. valencia, e.a. nicolás. simple resolution of butylated hydroxyanisole and n‐ propyl gallate in fatty foods and cosmetics samples by flow‐injection solid‐phase spectrophotometry. journal of food science 68(5) (2003) 1595-1599. https://doi.org/10.1111/j.1365-2621.2003.tb12297.x [7] n.b. tombesi, h. freije, application of solid-phase microextraction combined with gas chromatography–mass spectrometry to the determination of butylated hydroxytoluene in bottled drinking water. journal of chromatography a 963 (2002) 179-183. https://doi.org/10.1016/s00219673(02)00217-0 [8] a. sanches-silva, j.m. cruz, r. sendón-garcĺa, p. paseiro-losada. determination of butylated hydroxytoluene in food samples by high-performance liquid chromatography with ultraviolet detection and gas chromatography/mass spectrometry. journal of aoac international 90 (2007) 277283. https://doi.org/10.1093/jaoac/90.1.277 [9] y. guan, q. chu, l. fu, t. wu, j. ye. determination of phenolic antioxidants by micellar electrokinetic capillary chromatography with electrochemical detection. food chemistry 94 (2006) 157-162. https://doi.org/10.1016/j.foodchem.2005.01.015 [10] s. han, y. ding, f. teng, a. yao, q. leng. molecularly imprinted electrochemical sensor based on 3dflower-like mos2 decorated with silver nanoparticles for highly selective detection of butylated hydroxyanisole. food chemistry 387 (2022) 132899. https://doi.org/10.1016/j.foodchem.2022.132899 [11] n.s. sangeetha, s.s. narayanan. effective electrochemical detection of riboflavin and butylated hydroxy anisole based on azure a and nickel hexacyanoferrate framework on graphite electrode. chemical data collections 30 (2020) 100544. https://doi.org/10.1016/j.cdc.2020.100544 [12] h. karimi-maleh, r. darabi, m. shabani-nooshabadi, m. baghayeri, f. karimi, j. rouhi, c. karaman. determination of d&c red 33 and patent blue v azo dyes using an impressive electrochemical sensor based on carbon paste electrode modified with zif-8/g-c3n4/co and ionic liquid in mouthwash and toothpaste as real samples. food and chemical toxicology (2022) 112907. https://doi.org/10.1016/j.fct.2022.112907 [13] h. peyman, h. roshanfekr, a. babakhanian, h. jafari. pvc membrane electrode modified by lawson as synthetic derivative ionophore for determination of cadmium in alloy and wastewater. chemical methodologie 5(5) (2020) 446-453. http://www.chemmethod.com/article_135266.html [14] h. beitollahi, s.z. mohammadi, m. safaei, s. tajik. applications of electrochemical sensors and biosensors based on modified screen-printed electrodes: a review. analytical methods 12 (2020) 15471560. https://doi.org/10.1039/c9ay02598g [15] s. saghiri, m. ebrahimi, m.r. bozorgmehr. nio nanoparticle/1-hexyl-3-methylimidazolium hexafluorophosphate composite for amplification of epinephrine electrochemical sensor. asian journal of nanosciences and materials 4 (2021) 46-52. https://doi.org/10.26655/ajnanomat.2021.1.4 [16] y.f. mustafa, g. chehardoli, s. habibzadeh, z. arzehgar. electrochemical detection of sulfite in food samples. journal of electrochemical science and engineering 12 (2022) 1061-1079. https://doi.org/10.5599/jese.1555 [17] t. eren, n. atar, m. l. yola, h. karimi-maleh. a sensitive molecularly imprinted polymer based quartz crystal microbalance nanosensor for selective determination of lovastatin in red yeast rice. food chemistry 185 (2015) 430-436. https://doi.org/10.1016/j.foodchem.2015.03.153 https://doi.org/10.1007/s00216-006-0854-y https://doi.org/10.1039/c3an01190a https://doi.org/10.1007/s12161-014-9894-7 https://doi.org/10.1111/j.1365-2621.2003.tb12297.x https://doi.org/10.1111/j.1365-2621.2003.tb12297.x https://doi.org/10.1016/s0021-9673(02)00217-0 https://doi.org/10.1016/s0021-9673(02)00217-0 https://doi.org/10.1093/jaoac/90.1.277 https://doi.org/10.1016/j.foodchem.2005.01.015 https://doi.org/10.1016/j.foodchem.2022.132899 https://doi.org/10.1016/j.cdc.2020.100544 https://doi.org/10.1016/j.fct.2022.112907 https://doi.org/10.1016/j.fct.2022.112907 http://www.chemmethod.com/article_135266.html https://doi.org/10.1039/c9ay02598g https://doi.org/10.26655/ajnanomat.2021.1.4 https://doi.org/10.5599/jese.1555 https://doi.org/10.5599/jese.1555 https://doi.org/10.1016/j.foodchem.2015.03.153 admet & dmpk 11(2) (2023) 185-199 butylated hydorsyanisole determination using modified gce doi: http://dx.doi.org/10.5599/admet.1703 197 [18] m.s. sengar, s. saxena, s.p. satsangee, r. jain. silver nanoparticles decorated functionalized multiwalled carbon nanotubes modified screen printed sensor for voltammetric determination of butorphanol. journal of applied organometallic chemistry 1 (2021) 95-108. https://doi.org/10.22034/jaoc.2021.289344.1023 [19] m. mazloum-ardakani, h. beitollahi, b. ganjipour, h. naeimi. novel carbon nanotube paste electrode for simultaneous determination of norepinephrine, uric acid and d-penicillamine. international journal of electrochemical science 5 (2010) 531-546. [20] a. lohrasbi-nejad. electrochemical strategies for detection of diazinon. journal of electrochemical science and engineering 12 (2022) 1041-1059. https://doi.org/10.5599/jese.1379 [21] h. karimi-maleh, a.f. shojaei, k. tabatabaeian, f. karimi, s. shakeri, r. moradi. simultaneous determination of 6-mercaptopruine, 6-thioguanine and dasatinib as three important anticancer drugs using nanostructure voltammetric sensor employing pt/mwcnts and 1-butyl-3-methylimidazolium hexafluoro phosphate. biosensors and bioelectronics 86 (2016) 879-884. https://doi.org/10.1016/j.bios.2016.07.086 [22] z. mehdizadeh, s. shahidi, a. ghorbani-hasansaraei, m. limooei, m. bijad. monitoring of amaranth in drinking samples using voltammetric amplified electroanalytical sensor. chemical methodologies 6 (2022) 246-252. https://doi.org/10.22034/chemm.2022.324073.1423 [23] s. tajik, h. beitollahi, f.g. nejad, i.s. shoaie, m.a. khalilzadeh, m.s. asl, m. shokouhimehr. recent developments in conducting polymers: applications for electrochemistry. rsc advances 10(62) (2020) 37834-37856. https://doi.org/10.1039/d0ra06160c [24] a. shamsi, f. ahour. electrochemical sensing of thioridazine in human serum samples using modified glassy carbon electrode. advanced journal of chemistry-section a 4 (2021) 22-31. https://doi.org/10.22034/ajca.2020.252025.1215 [25] p.m. jahani. flower-like mos2 screen-printed electrode based sensor for the sensitive detection of sunset yellow fcf in food samples. journal of electrochemical science and engineering 12 (2022) 10991109. https://doi.org/10.5599/jese.1413 [26] s.a. alavi-tabari, m. a. khalilzadeh, h. karimi-maleh. simultaneous determination of doxorubicin and dasatinib as two breast anticancer drugs uses an amplified sensor with ionic liquid and zno nanoparticle. journal of electroanalytical chemistry 811 (2018) 84-88. https://doi.org/10.1016/j.jelechem.2018.01.034 [27] s. azimi, m. amiri, h. imanzadeh, a. bezaatpour. fe3o4@sio2-nh2/cosb modified carbon paste electrode for simultaneous detection of acetaminophen and chlorpheniramine. advanced journal of chemistry-section a 4 (2021) 152-164. https://doi.org/10.22034/ajca.2021.275901.1246 [28] s. tajik, m.a. taher, h. beitollahi. first report for simultaneous determination of methyldopa and hydrochlorothiazide using a nanostructured based electrochemical sensor. journal of electroanalytical chemistry 704 (2013) 137-144. https://doi.org/10.1016/j.jelechem.2013.07.008 [29] m. vardini, n. abbasi, a. kaviani, m. ahmadi, e. karimi. graphite electrode potentiometric sensor modified by surface imprinted silica gel to measure valproic acid. chemical methodologies 6 (2022) 398-408. http://www.chemmethod.com/article_147528.html [30] s. ariavand, m. ebrahimi, e. foladi. design and construction of a novel and an efficient potentiometric sensor for determination of sodium ion in urban water samples. chemical methodologies 6(11) (2022) 886-904. https://doi.org/10.22034/chemm.2022.348712.1567 [31] h. karimi-maleh, m. sheikhshoaie, i. sheikhshoaie, m. ranjbar, j. alizadeh, n.w. maxakato, a. abbaspourrad. a novel electrochemical epinine sensor using amplified cuo nanoparticles and anhexyl-3-methylimidazolium hexafluorophosphate electrode. new journal of chemistry 43 (2019), 2362-2367. https://doi.org/10.1039/c8nj05581e [32] r.m. mohabis, f. fazeli, i. amini, v. azizkhani. an overview of recent advances in the detection of ascorbic acid by electrochemical techniques. journal of electrochemical science and engineering 12 (2022) 1081-1098. https://doi.org/10.5599/jese.1561 http://dx.doi.org/10.5599/admet.1703 https://doi.org/10.22034/jaoc.2021.289344.1023 https://doi.org/10.22034/jaoc.2021.289344.1023 https://doi.org/10.5599/jese.1379 https://doi.org/10.1016/j.bios.2016.07.086 https://doi.org/10.1016/j.bios.2016.07.086 https://doi.org/10.22034/chemm.2022.324073.1423 https://doi.org/10.1039/d0ra06160c https://doi.org/10.22034/ajca.2020.252025.1215 https://doi.org/10.22034/ajca.2020.252025.1215 https://doi.org/10.5599/jese.1413 https://doi.org/10.1016/j.jelechem.2018.01.034 https://doi.org/10.1016/j.jelechem.2018.01.034 https://doi.org/10.22034/ajca.2021.275901.1246 https://doi.org/10.1016/j.jelechem.2013.07.008 http://www.chemmethod.com/article_147528.html https://doi.org/10.22034/chemm.2022.348712.1567 https://doi.org/10.1039/c8nj05581e https://doi.org/10.5599/jese.1561 m. fazli and n. akbarzadeh-t. admet & dmpk 11(2) (2023) 185-199 198 [33] s. tajik, m.a. taher, h. beitollahi. simultaneous determination of droxidopa and carbidopa using a carbon nanotubes paste electrode. sensors and actuators b: chemical 188 (2013) 923-930. https://doi.org/10.1016/j.snb.2013.07.085 [34] a. hosseini fakhrabad, r. sanavi khoshnood, m.r. abedi, m. ebrahimi. fabrication a composite carbon paste electrodes (cpes) modified with multi-wall carbon nano-tubes (mwcnts/n, n-bis (salicyliden)1,3-propandiamine) for determination of lanthanum (iii). eurasian chemical communications 3 (2021) 627-634. https://doi.org/10.22034/ecc.2021.288271.1182 [35] k. harismah, m. mirzaei, m. dai, z. roshandel, e. salarrezaei. in silico investigation of nanocarbon biosensors for diagnosis of covid-19. eurasian chemical communications 3(2) (2021) 95-102. https://doi.org/10.22034/ecc.2021.267226.1120 [36] h. karimi-maleh, c. karaman, o. karaman, f. karimi, y. vasseghian, l. fu, a. mirabi. nanochemistry approach for the fabrication of fe and n co-decorated biomass-derived activated carbon frameworks: a promising oxygen reduction reaction electrocatalyst in neutral media. journal of nanostructure in chemistry (2022) 1-11. https://doi.org/10.1007/s40097-022-00492-3 [37] a. dehno khalaji, p. machek, m. jarosova, α-fe2o3 nanoparticles: synthesis, characterization, magnetic properties and photocatalytic degradation of methyl orange. advanced journal of chemistry-section a 4(4) (2021) 317-326. https://doi.org/10.22034/ajca.2021.292396.1268 [38] r. hajinasir, m. esmaeili jadidi. synthesis of zno nanoparticles via flaxseed aqueous extract. journal of applied organometallic chemistry 2(2) (2022) 101-108. https://doi.org/10.22034/jaoc.2022.154805 [39] m. alidadykhoh, h. pyman, h. roshanfekr. application of a new polymer agcl nanoparticles coated polyethylene terephetalat [pet] as adsorbent for removal and electrochemical determination of methylene blue dye. chemical methodologie 5(2) (2021) 96-106. https://doi.org/10.22034/chemm.2021.119677 [40] n. koohzadi, z. rezayati zad. a highly sensitive colorimetric determination of paraquat by silver nanoparticles. advanced journal of chemistry-section b: natural products and medical chemistry 3(4) (2021) 311-322. https://doi.org/10.22034/ajcb.2021.302596.1092 [41] m. miraki, h. karimi-maleh, m.a. taher, s. cheraghi, f. karimi, s. agarwal, v.k. gupta. voltammetric amplified platform based on ionic liquid/nio nanocomposite for determination of benserazide and levodopa. journal of molecular liquids 278 (2019) 672-676. https://doi.org/10.1016/j.molliq.2019.01.081 [42] h. beitollahi, i. sheikhshoaie. novel nanostructure-based electrochemical sensor for simultaneous determination of dopamine and acetaminophen. materials science and engineering: c 32(2) (2012) 375-380. https://doi.org/10.1016/j.msec.2011.11.009 [43] m. bijad, a. hojjati-najafabadi, h. asari-bami, s. habibzadeh, i. amini, f. fazeli. an overview of modified sensors with focus on electrochemical sensing of sulfite in food samples. eurasian chemical communications 3(2) (2021) 116-138. https://doi.org/10.22034/ecc.2021.268819.1122 [44] a. sahraei, h. kargar, m. hakimi, m.n. tahir. distorted square-antiprism geometry of new zirconium (iv) schiff base complexes: synthesis, spectral characterization, crystal structure and investigation of biological properties. journal of molecular structure 1149 (2017) 576-584. https://doi.org/10.1016/j.molstruc.2017.08.022 [45] a. trzesowska-kruszynska. copper complex of glycine schiff base: in situ ligand synthesis, structure, spectral, and thermal properties. journal of molecular structure 1017 (2012) 72-78. https://doi.org/10.1016/j.molstruc.2012.03.003 [46] j. joseph, g.a.b. rani. metal based sod mimetic therapeutic agents: synthesis, characterization and biochemical studies of metal complexes. arabian journal of chemistry 10 (2017) s1963-s1972. https://doi.org/10.1016/j.arabjc.2013.07.024 [47] x. liu, j.r. hamon. recent developments in penta-, hexa-and heptadentate schiff base ligands and their metal complexes. coordination chemistry reviews 389 (2019) 94-118. https://doi.org/10.1016/j.ccr.2019.03.010 https://doi.org/10.1016/j.snb.2013.07.085 https://doi.org/10.22034/ecc.2021.288271.1182 https://doi.org/10.22034/ecc.2021.267226.1120 https://doi.org/10.22034/ajca.2021.292396.1268 https://doi.org/10.22034/jaoc.2022.154805 https://doi.org/10.22034/chemm.2021.119677 https://doi.org/10.22034/chemm.2021.119677 https://doi.org/10.22034/ajcb.2021.302596.1092 https://doi.org/10.1016/j.molliq.2019.01.081 https://doi.org/10.1016/j.molliq.2019.01.081 https://doi.org/10.1016/j.msec.2011.11.009 https://doi.org/10.22034/ecc.2021.268819.1122 https://doi.org/10.1016/j.molstruc.2017.08.022 https://doi.org/10.1016/j.molstruc.2017.08.022 https://doi.org/10.1016/j.molstruc.2012.03.003 https://doi.org/10.1016/j.molstruc.2012.03.003 https://doi.org/10.1016/j.arabjc.2013.07.024 https://doi.org/10.1016/j.ccr.2019.03.010 https://doi.org/10.1016/j.ccr.2019.03.010 admet & dmpk 11(2) (2023) 185-199 butylated hydorsyanisole determination using modified gce doi: http://dx.doi.org/10.5599/admet.1703 199 [48] s. kaya, s. erkan, d. karakaş. computational investigation of molecular structures, spectroscopic properties and antitumor-antibacterial activities of some schiff bases. spectrochimica acta part a: molecular and biomolecular spectroscopy 244 (2021) 118829. https://doi.org/10.1016/j.saa.2020.118829 [49] a.m. abu-dief, i.m. mohamed. a review on versatile applications of transition metal complexes incorporating schiff bases. beni-suef university journal of basic and applied sciences 4(2) (2015) 119133. https://doi.org/10.1016/j.bjbas.2015.05.004 [50] s.a. dailami, s.a. onakpa, m.a. funtua. fe (iii) and cr (iii) complexes with phenolic schiff base: synthesis, physico-chemical characterisation, antimicrobial and antioxidant studies. iosr journal of applied chemistry 9(9) (2016) 18-23. [51] e. tozzo, s. romera, m.p. dos santos, m. muraro, h.d.a. regina, l.m. liao, e.r. dockal. synthesis, spectral studies and x-ray crystal structure of n, n′-(±)-trans-1, 2-cyclohexylenebis (3ethoxysalicylideneamine) h2 (t-3-etosalchxn). journal of molecular structure 876 (2008) 110-120. https://doi.org/10.1016/j.molstruc.2007.05.043 [52] m. masoudi, m. behzad, a. arab, a. tarahhomi, h.a. rudbari, g. bruno. crystal structures, dft calculations and hirshfeld surface analyses of three new cobalt (iii) schiff base complexes derived from meso-1, 2-diphenyl-1, 2-ethylenediamine. journal of molecular structure 1122 (2016) 123-133. https://doi.org/10.1016/j.molstruc.2016.05.092 [53] p.g. cozzi. metal–salen schiff base complexes in catalysis: practical aspects. chemical society reviews 33 (2004) 410-421. https://doi.org/10.1039/b307853c [54] o. fatibello‐filho, e.r. dockal, l.h. marcolino‐junior, m.f. teixeira. electrochemical modified electrodes based on metal‐salen complexes. analytical letters 40 (2007) 1825-1852. https://doi.org/10.1080/00032710701487122 [55] w. al zoubi, n. al mohanna. membrane sensors based on schiff bases as chelating ionophores–a review. spectrochimica acta part a: molecular and biomolecular spectroscopy 132 (2014) 854-870. https://doi.org/10.1016/j.saa.2014.04.176 [56] é.n. oiye, m.f.m. ribeiro, j.m.t. katayama, m.c. tadini, m.a. balbino, i.c. eleotério, m.f. de oliveira. electrochemical sensors containing schiff bases and their transition metal complexes to detect analytes of forensic, pharmaceutical and environmental interest. a review. critical reviews in analytical chemistry 49(6) (2019) 488-509. https://doi.org/10.1080/10408347.2018.1561242 [57] r.h. taha, z.a. el-shafiey, a.a. salman, e.m. el-fakharany, m.m. mansour. synthesis and characterrization of newly synthesized schiff base ligand and its metal complexes as potent anticancer. journal of molecular structure 1181 (2019) 536-545. https://doi.org/10.1016/j.molstruc.2018.12.055 [58] e. yousif, a. majeed, k. al-sammarrae, n. salih, j. salimon, b. abdullah. metal complexes of schiff base: preparation, characterization and antibacterial activity. arabian journal of chemistry 10 (2017) s1639s1644. https://doi.org/10.1016/j.arabjc.2013.06.006 [59] m. biabani, h. saravani, v. eigner, m. dusek. a novel coordination polymer of cd (ii) based on 2mercaptopyrimidine: sonochemical synthesis, characterization, and antibacterial properties. journal of molecular structure 1166 (2018) 470-478. https://doi.org/10.1016/j.molstruc.2018.04.054 [60] h. sadeghzadeh, a. morsali. sonochemical synthesis and characterization of nano-belt lead (ii) coordination polymer: new precursor to produce pure phase nano-sized lead (ii) oxide. ultrasonics sonochemistry 18(1) (2011) 80-84. https://doi.org/10.1016/j.ultsonch.2010.01.011 [61] m. saif, h.f. el-shafiy, m.m. mashaly, m.f. eid, a.i. nabeel, r. fouad. synthesis, characterization, and antioxidant/cytotoxic activity of new chromone schiff base nano-complexes of zn (ii), cu (ii), ni (ii) and co (ii). journal of molecular structure 1118 (2016) 75-82. https://doi.org/10.1016/j.molstruc.2016.03.060 ©2023 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://dx.doi.org/10.5599/admet.1703 https://doi.org/10.1016/j.saa.2020.118829 https://doi.org/10.1016/j.bjbas.2015.05.004 https://doi.org/10.1016/j.molstruc.2007.05.043 https://doi.org/10.1016/j.molstruc.2016.05.092 https://doi.org/10.1039/b307853c https://doi.org/10.1080/00032710701487122 https://doi.org/10.1080/00032710701487122 https://doi.org/10.1016/j.saa.2014.04.176 https://doi.org/10.1080/10408347.2018.1561242 https://doi.org/10.1016/j.molstruc.2018.12.055 https://doi.org/10.1016/j.arabjc.2013.06.006 https://doi.org/10.1016/j.molstruc.2018.04.054 https://doi.org/10.1016/j.ultsonch.2010.01.011 https://doi.org/10.1016/j.molstruc.2016.03.060 https://doi.org/10.1016/j.molstruc.2016.03.060 http://creativecommons.org/licenses/by/3.0/ in memoriam: konstantin tsinman (1968 2020) doi: https://doi.org/10.5599/admet.1556 241 admet & dmpk 10(4) (2022) 241-244; doi: https://doi.org/10.5599/admet.1556 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index in memoriam konstantin tsinman (1968 2020) alex avdeef, phd, faaps, cofounder of pion inc. in-adme research, 1732 first avenue #102, new york, ny 10128 usa; alex@in-adme.com dr. konstantin tsinman, chief scientific officer at pion inc., succumbed after a three-week struggle, to the ravages of covid-19 infection on friday, 24 april 2020. those of us who knew konstantin were stunned to hear the horrible news. it is tragic to lose such a gifted scientist in the prime of his career, with so many valuable contributions to his credit and with so many yet to come. and to his wife oksana and their daughter tonia, what can anyone say to soften the deep pain and sorrow of loss? i admired konstantin very much. he had a calm demeanor, was very smart, unpretentious, very easily approachable. i still see his smiling face in my mind. konstantin joined pion in 1998 (earliest photo of him above), just as the seedling company was starting to expand, poised to develop innovative scientific equipment for pharmaceutical research. the start-up secured its first external funds from a prominent pharmaceutical company in new jersey, in exchange for the promise of delivery of the first “pampa” instruments (parallel artificial membrane permeability assay). we had been working behind the scenes with dr. manfred kansy of roche (basel), the inventor of pampa, for the preceding two years, before his famous seminal paper was published in 1998. so, that year we took the investment we secured and spent the whole lot on (i) a lease to a modest space on the second floor of a twostory industrial building, above mike’s gym, on the outskirts of cambridge, massachusetts, (ii) a tecan robot workstation (pictured above – which several of us carried up the rickety wooden staircase in the elevatorless building), (iii) hiring a laboratory scientist (who later went on to medical school to become a physician), and (iv) hiring a scientific programmer – konstantin. he was a phd physicist interested in robotics programming, with a ‘can do just about anything’ attitude. he received his advanced degrees from the institute of metal physics and the dnepropetrovsk state university in ukraine. he and his family had just then emigrated to america. his wife, oksana, also a scientist, eventually joined the pion team, and took over the running of the laboratory – a truly remarkable husband-wife duo at work. the first few years were extremely exciting, albeit somewhat stressful. at the same time, many good ideas kept coming our way, which eventually led to new and innovative instrument products. we even were https://doi.org/10.5599/admet.1556 https://doi.org/10.5599/admet.1556 http://www.pub.iapchem.org/ojs/index.php/admet/index alex avdeef admet & dmpk 10(4) (2022) 241-244 242 able to convince a group of scientists at the fda to join us in a collaborative study to validate a newly-invented potentiometric method for measuring intrinsic solubility of drugs. the resultant publication in 2000 spurred a lot of interest in pion’s products for pharmaceutical research, laying the groundwork for its future. relaxing with friends in the back bay of boston on the fourth of july 2005. konstantin is the second person from the left. oksana is to his left. his life-long friend, dima, is in the blue and white shirt near the center. dima worked at pion for many years. konstantin was the principal developer of the high throughput permeability analyzer and subsequently the high-throughput method for measuring solubility-ph profiles. he continued investigating physicochemical factors influencing intestinal absorption and penetration across the blood-brain-barrier of pharma research molecules, expanding the scope of applications for the uv fiber-optic technique. a longstanding challenge in drug dissolution measurement was the assessment of multicomponent active pharmaceutical ingredients in exploratory formulations. he developed the so-called “zim” technique elegantly to address the problem. its later applications to studying nanoparticle suspensions were groundbreaking. it was such a clever idea that allowed the implementation of uv-vis spectroscopy for real time concentration monitoring of complex phenomena. konstantin had been involved in multiple collaborative research projects with scientists from pharmaceutical companies and academic institutions, expanding the scope of applications for pion's in situ uv fiber-optic dissolution technology, eventually coupling it with permeability measurement. he co-authored many articles in primary scientific journals and held several patents. konstantin was more than a laboratory researcher. he jumped right in to participate and lead in user training programs. he frequently traveled with the sales force to demonstrate products. he was an excellent speaker at scientific symposia all around the world. one could find him at the pion booths at tradeshows, big and small. he always had a welcoming smile. admet & dmpk 10(4) (2022) 241-244 in memory of konstantin tsinman doi: https://doi.org/10.5599/admet.1556 243 konstantin with manfred kansy at the first international pampa conference in san francisco, california in 2002. the last photo i have of konstantin – the scientist amid scientists – in a quaint restaurant in the ancient adriatic port of split, croatia, at the iapc-8 meeting, september 2019. there were many other tables around the corner and there were musicians serenading and electrifying the spirits. the man standing is professor zoran mandić (university of zagreb), the seemingly tireless organizer of the iapc series of conferences on topics related to drug discovery and development. https://doi.org/10.5599/admet.1556 alex avdeef admet & dmpk 10(4) (2022) 241-244 244 as news of konstantin’s passing spread, i received many emails from those who knew him. a sampling of thoughts: “i was shocked to hear the bad news, about the tragic death of konstantin. it is incredible, such a young person as a victim of this virus pandemic. … i know … how bright a scientist he was. i met him first last year in split and found a very nice guy and excellent speaker. i am very sorry… i am sitting and thinking on the life. it is so fragile.” “very shocked … such a good person.” “i am really shocked and feel so sorry for the family. i remember the 4th of july on the roof in boston so well. a happier time. every day is precious…” “i'm shocked … i met him in boston last october. he has been very kind to me since i first met him.” “… terribly sad news regarding konstantin. i can’t believe it, he will be very sadly missed… i travelled with him on many occasions visiting customers and at trade shows. he was great company and taught me a lot about the science behind the pion products. please pass on my deepest sympathy.” “it is very sad news. although i didn't know him well, i met him on several occasions and always had pleasant and sensible discussions.” “my dear dear friend konstantin, i´m out of words…konstantin was one of the reasons for going to the aaps, having chats and laughs, friendly but spot-on scientific discussions. the one who took care of me when arriving to pion and boston in 1998 as a visiting phd student …just starting my research career. i am devastated. so so sorry. what a big loss to this world he will be so missed, for his sharp mind and friendliness.” konstantin will be deeply missed by his many friends. his soft-spoken, upbeat and cheerful manner endeared him to all with whom he had contact. he was an explorer, he was a teacher, and he was an inspiration to all. искра жизни драгоценна и может быть мимолетной. покойся с миром, друг мой. new york, november 2022. ©2022 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/) http://creativecommons.org/licenses/by/4.0/ synthetic routes to theranostic applications of carbon-based quantum dots doi: https://doi.org/10.5599/admet.1740 1 admet & dmpk 0(0) (2023) 000-000; doi: https://doi.org/10.5599/admet.1740 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper polyvinyl alcohol nanoparticles loaded with propolis extract: fabrication, characterization and antimicrobial activity benazir subaşı-zarbaliyev, gozde kutlu and fatih tornuk* yildiz technical university, faculty of chemical and metallurgical engineering, department of food engineering, davutpasa campus, istanbul, turkey *corresponding author: e-mail: ftornuk@yildiz.edu.tr; tel.: +90-212-383-4380; fax: +90-212-383-4572 §equal contribution received: march 14, 2023; revised: april 24, 2023; published: may xx, 2023 abstract background and purpose: propolis has high potential beneficial bioactive properties such as anti-oxidative, antimicrobial, and anti-tumour activities. however, the bitter taste and the insolubility nature of propolis in water lead to some limitations in their usage in functional food applications. experimental approach: herein, we evaluated the effects of nanoencapsulation of propolis at the different concentration levels (0, 0.4, 0.8, 1.0, and 1.2 %) into the polyvinyl alcohol (pva) nanoparticles using the electrospraying method, on the structural, physical, antioxidant, antimicrobial and thermal properties. key results: the results revealed that the fabricated nanocapsules (pva-nps) obtained under optimal conditions had uniform size distribution and unstable particles with small particle size between 104-258 nm, a polydispersity index <0.317, and a zeta potential between -5 and +5 mv. the maximum encapsulation efficiency of pva-nps was about 25.32 % for 1 % of the initial propolis loading level. dsc thermal experiments showed an increase in the thermal stability of the propolis loaded pva nanoparticles as compared to the neat pva nanoparticles. the percent inhibition of dpph radical scavenging activity of the nanocapsules was between 80 and 89 %. sem analysis revealed that pva-nps had a spherical shape with a rough surface and were composed of long and thin fibres at nanometric diameters. ft-ir analysis showed that no indications of any chemical reactions were found between the constituents of the core and wall material due to their physical mixing. antibacterial efficacy was evaluated by the broth dilution method and pva-nps exhibited good inhibitory activity against s. aureus at low concentration ratios, whereas it had no inhibitory activity against e. coli o157:h7. conclusion: pva-nps fabricated using the electrospraying technique can be used for the development of a new promising natural and bioactive agent in the food and pharmaceutical industry. ©2023 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords encapsulation, electrospraying, thermal property, antimicrobial activity, morphological properties introduction materials with reduced size up to nanoscale can display quite different properties as compared to micro or macro-scale equivalents. therefore, nanotechnology has become one of the most efficient and popular topics with a wide range of application fields such as medical, agricultural and food science. one of the most common research and development areas in nanotechnology applications in the food and pharmaceutical industry is the encapsulation of bioactive compounds. encapsulation can be defined as a process that is the https://doi.org/10.5599/admet.1740 https://doi.org/10.5599/admet.1740 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:ftornuk@yildiz.edu.tr http://creativecommons.org/licenses/by/4.0/ benazir subaşı-zarbaliyev et al. admet & dmpk 00(0) (2023) 000-000 2 entrapment or coating of a pure material or a mixture (core material, actives, fill, internal phase, or payload) into a coating material (the capsule, wall material, membrane, carrier or shell) [1,2]. the electrospraying (electro-hydrodynamic atomization), a nanoencapsulation technique directed by the electrostatic forces leading to liquid atomization, has found application on laboratory and industrial scales for the fabrication of the nanoparticles [3]. the main factors affecting the electrospraying process include solution (e.g., viscosity, conductivity, molecular weight, and surface tension), process (e.g., applied electric field, tip to collector distance and feeding or flow rate), and environmental conditions (e.g., the humidity and temperature of the surroundings) [4]. this technique can produce stable nanocapsules with enhanced bioactive properties. moreover, this technique is simple, promising, one-step, highly reproducible and usable for bulk applications. it is also possible to set process parameters and produce more uniform particles with lower particle sizes and higher encapsulation efficiency [5]. polyvinyl alcohol (pva) is a water-soluble, odourless, tasteless, non-toxic, non-carcinogenic and fully biodegradable synthetic polymer obtained by partial or complete hydrolysis of polyvinyl acetate. it forms an excellent particle and fibre structure either alone or when blended with other natural polymers and offers a great promise for better biocompatibility, particularly due to its good chemical resistance, stability and flexibility. also, it shows stability in a wide range of temperature and ph [6,7]. propolis (bee glue) is a natural honeybee product with a naturally heterogeneous complex matrix used for the construction, maintenance, insulation, and protection of hives by the honeybees from various plant sources [8]. it is a sticky, balsamic and resinous material that can vary in colour from light yellow to dark brown and has a sweet but also bitter taste [9]. its composition can vary depending on its botanical and geographical origin [8]. additionally, it has a great potential for health benefits in terms of its antiseptic, antibacterial (particularly against gram-positive bacteria), anti-mycotic, anti-viral, anti-protozoal, antioxidative, anti-inflammatory, and anti-tumour properties [10]. although it has unique beneficial health effects, its use in its original form is limited due to its low solubility in water and organic solvents and its strong flavour [11]. to date, several fabrication techniques such as spray-drying [12,13], electrospinning [11], complex-coacervation [14], electrospraying [15], and pressurized carbon dioxide anti-solvent crystallization [16] have been used for propolis encapsulation to overcome solubility problems. to the best of our knowledge, no published papers have focused on the encapsulation of propolis into the pva by the electrospraying method. therefore, the purpose of the current work was (i) to optimize the nanoencapsulation conditions of propolis using pva at the highest encapsulation efficiency and to make their characterization by their antioxidant (dpph scavenging), antimicrobial, morphological (sem), thermal (dsc), and structural (ft-ir) properties. experimental materials and reagents raw propolis used in this study was obtained from arifoğlu spice& food industry trade company (i̇stanbul, turkey). polyvinyl alcohol (sigma-aldrich company, germany), nutrient broth (merck, germany), maximum recovery diluent (mrd, merck, germany) and 2,2-diphenyl-1-picrylhydracyl (dpph) (sigmaaldrich company, germany) were purchased. methanol and ethanol were analytical grades. preparation of propolis extract propolis extract was obtained by using 80 % ethanol as solvent. firstly, 8 % of ethanolic propolis solution was prepared. the obtained solution was stirred with a magnetic stirrer (ika c-mag hs 7, germany) for 24 h at room temperature and then kept at -20 °c for 3 days to precipitate the impurities. after that, the https://www.sciencedirect.com/topics/materials-science/atomizer-nozzle admet & dmpk 00(0) (2023) 000-000 polyvinyl alcohol nanoparticles loaded with propolis extract doi: https://doi.org/10.5599/admet.1740 3 supernatant was taken and filtered through filter paper. following the filtration, the solution was first centrifuged (centrifuge multifuge x3 fr, thermo scientific) twice at -5 °c for 5 min at 4.200 rpm. afterwards, the supernatant was filtered by passing through a 0.22-μm filter. removal of the solvent from ethanol extract was done by vacuum evaporation (heidolph rotary evaporator, laborota 4000, schwabach, germany) at 50°c. the resulting propolis extract was kept in the refrigerator at 4 °c until further use. optimization of electrospraying conditions based on pva concentration in this study, prior to propolis encapsulation, optimum pva concentration was determined to obtain homogeneous and low-diameter pva nanoparticles using the electrospraying technique. for this purpose, different pva solutions were prepared at various concentrations to determine the most appropriate pva concentration and to optimize the electrospraying conditions and then the nanoparticles were fabricated depending on the different processing conditions using the electrospraying method. in light of the obtained results of the size and zeta potential measurements, optimization studies were performed. preparation of the coating material different aqueous pva solutions 5 to 10 % (w:v) were prepared and each solution was shaken in a water bath (memmert model sv 1422, schwabach, germany) at 100 rpm for 20 min at 50 °c to form a homogeneous solution. the nanoparticles were fabricated using an electrospraying device (nanodev company, ankara, turkey). for this purpose, 5 ml plastic syringes (steriject, with a 0.8038 mm diameter needle), each filled with neat pva feed solutions, were inserted into the syringe pump (new erapump systems inc., ne-300, hauppauge, ny) to spray the feed solutions. during electrospraying, different process parameters, such as voltage (1020 kv), solution flow rate (0.3-0.4-0.5 ml h-1), and the tip-to-collector plate distance (10-20 cm), were tested. the most suitable processing parameters were determined by measuring the particle size and assessing the microscope images of resulting nanoparticles applied to the different process parameters mentioned above. particle size measurements of neat pva nanoparticles a malvern zetasizer nano zs instrument (malvern instruments co. ltd., worcestershire, uk) was used to determine the variations in the particle size and polydispersity index (pdi) values depending on the pva concentration in the feed solutions (5, 6, and 10 %; w/v). for this purpose, the nanoparticles were diluted with sterile distilled water, homogenized using an ultra-turrax (daihan, hg-15d, gang-won-do, south korea) for 30 min, and then passed through 0.45 µm pore size disposable filter. measurements were carried out in triplicate. fabrication and characterization of pva nanoparticles loaded with propolis extract (pva-nps) a slightly modified procedure reported by durán et al. (2007) [17] was followed to fabricate pva-nps with optimization of the fabrication parameters outlined below. firstly, pva solutions containing propolis extracts at different concentrations (0, 0.4, 0.8, 1.0, and 1.2 %) were prepared. for this purpose, 10 ml of 6 % pva solution was prepared and placed in a water bath (memmert model sv 1422, schwabach, germany) with constant shaking at 80 °c for 60 min. then, the propolis extract was added and stirred for 24 h at ambient temperature with a magnetic stirrer (ika c-mag hs 7, germany) to obtain a completely homogeneous solution. afterwards, the solutions were filled into the 5 ml plastic syringe fitted with a steel needle (steriject, with a 0.8038 mm (diameterlength) needle) and fed into the electrospraying device applying the following process parameters: voltage, 15 kv; the tip to collector plate distance, 10 cm; and solution flow rate, 0.5 ml/h. eventually, va-nps were collected from the aluminium foil-coated plate. the samples were coded as pva-np0, pva-np1, pva-np2, pva-np3 and pva-np4 based on propolis concentrations of 0, 0.4, 0.8, 1.0, and 1.2 %, respectively. https://doi.org/10.5599/admet.1740 benazir subaşı-zarbaliyev et al. admet & dmpk 00(0) (2023) 000-000 4 surface charge (ζ-potential) and particle size measurements particle size and zeta potential measurements of pva-nps were measured using a malvern zetasizer nano zs analyser (malvern instruments co. ltd., worcestershire, uk) at 25 °c. for this purpose, pva-nps were diluted with distilled water and the measurements were conducted. each result was calculated as the average of 10 (one set of 10) measurements. the results of particle size and ζ-potential were presented as nanometre and millivolt, respectively. determination of encapsulation efficiency the encapsulation efficiency of pva-nps was determined according to a method described by durán et al. [17], with some modifications. pva-nps containing 10 ml of deionized water were continuously stirred for 60 min using a magnetic stirrer and centrifuged (centrifuge multifuge x3 fr, thermoscientific) twice for 5 min at 5.000 rpm. then each supernatant was filtered through the filter papers, the absorbance measurements were done using a uv visible spectrometer (uv-1800, shimadzu, kyoto, japan). the percent encapsulation efficiency was determined using the following equation. encapsulation efficiency = ((absorbancetotal-absorbancefiltrate)/absorbancetotal)100 (1) ft-ir molecular characterization of the propolis extract and pva-nps were performed using a bruker ft-ir spectrometer model tensor 27 (bremen, germany) coupled to a diamond platinum atr unit (spectral range of 400-4000 cm-1, resolution: 4 cm−1, and scans: 32). dsc thermal properties of pva-nps were determined using a dsc instrument (dsc, q100, ta instruments inc., new castle, de, usa) in the temperature range of 40-400 °c at a heating rate of 10 °c min−1 under a nitrogen atmosphere. about 4-5 mg of pva-nps sample was placed on aluminium-coated pans and loaded into the instrument chamber at ambient temperature. an empty pan was used as a reference. dpph radical scavenging activity firstly, 1.5 ml of ethanolic extract of pva-nps was well-mixed with 1.5 ml of dpph solution, vigorously vortexed for 1 min and left in a dark environment for 30 min at room temperature. then the absorbance values were measured at 517 nm using a uv-visible spectrometer (uv-1800, shimadzu, kyoto, japan). the dpph scavenging activity was calculated using the following equation: dpph scavenging activity = ((absorbancecontrol-absorbancesample)/absorbancecontrol)100 (2) antimicrobial activity the broth dilution method described by feyzioglu et al. [18] was used for the determination of the antimicrobial activity of pva-nps against two different bacterial strains (s. aureus (atcc 25923) and escherichia coli o157:h7 (atcc 33150)). initially, after activating bacterial strains twice in nutrient broth at 37 °c for 24 h, the test tubes, including 7 ml of nutrient broth, were seeded with the bacterial strains (a final inoculum of approximately 108 colony-forming units (cfu ml-1) and then incorporated with2 % of the pvanps samples. following incubation of the test tubes for 5 h in the shaking incubator at 37 °c at 140 rpm, serial dilutions were prepared, and colony counts were analysed using a spread plating method. colony forming units were counted after 24 h of incubation of petri plates at 37 °c. the results were presented as logarithmic values. admet & dmpk 00(0) (2023) 000-000 polyvinyl alcohol nanoparticles loaded with propolis extract doi: https://doi.org/10.5599/admet.1740 5 sem the surface morphology of pva-nps was analysed by scanning electron microscopy (sem, fei quanta feg 250, usa) under vacuum at 20 kv and at a 7.6 mm of working distance. prior to analysis, the samples were lyophilized and coated with gold. statistical analysis a minimum of three independent experiments with two replicates were performed to achieve efficient results and the findings were tabulated as mean ± standard deviation. a one-way anova (analysis of variance) test for the determination of the statistical differences between the data was applied using a t-test at the p<0.05 level and the statistical analysis of data was performed by the jmp software (v6; sas institute inc., cary, nc, usa). results and discussion optimization of electrospraying processing conditions in this study, firstly, different concentrations (5, 6 and 10 %) of pva solutions were fed to the electrospraying system. the particle size of the nanoparticles was monitored. the highest particle size value was measured in the solution containing 10 % pva (212-255 nm), followed by the solution containing 5 % pva (144-180 nm), and the solution including 6 % pva (86-106 nm). microscope images showed the formation of nanofibers at higher concentrations than 7 %, while dark (black) beads were obtained at lower concentrations (5 %). moreover, laboratory trials showed that an increase in the ultrasonication time before the electrospraying led to the formation of more homogeneous particles. overall, ultrasonication frequency and coating solution concentration were determined as 50 hz and 6 %, respectively, in this study. moreover, optimum electrospraying conditions were an applied voltage of 15 kv, a tip-to-collector plate distance of 10 cm and a solution flow rate of 0.5 ml / h. characterization of pva-nps following the optimization of electrospraying conditions, pva-nps were prepared with the incorporation of propolis extract. then the resulting nanoparticles were characterized. ζ-potential, pdi and particle size the summary of the ζ-potential, pdi and mean particle size findings for the pva-np0, pva-np1, pva-np2, pva-np3 and pva-np were given in table 1. the different wall materials and types of encapsulating agents influence the various capsule characteristics (e.g., particle size, encapsulation efficiency, oxidative stability, and morphology) [19]. table 1. polydispersity index, mean particle size, stability, encapsulation efficiency and dpph scavenging activity values of pva-nps. sample name polydispersity index mean particle size, nm ζ-potential, mv encapsulation efficiency, % dpph scavenging activity, % pva-n0 0.23 ± 0.16bc 106.37 ± 14.84c 3.09±0.50a 0.00 ± 0.00d pva-n1 0.32 ± 0.16ab 137.17 ± 61.82b 2.77±0.29a 5.09 ± 0.37d 78.13 ± 0.43c pva-n2 0.22 ± 0.03c 186.75 ± 8.27b -0.24±0.06b 19.43 ± 1.16b 85.09 ± 2.82b pva-n3 0.42 ± 0.12a 258.51 ± 90.51a -2.44±1.27b 25.32 ± 0.32a 87.32 ± 0.21a pva-n4 0.48 ± 0.08a 249.33 ± 41.40b -0.11±0.10b 18.15 ± 0.26c 89.25 ± 1.22a ζ-potential: surface charge, pva: polyvinyl alcohol, pva-n0: neat pva nanoparticles, pva-n1: pva nanoparticles loaded with propolis in the ratio of 0.4 %, pva-n2: pva nanoparticles loaded with propolis in the ratio of 0.8 %, and pva-n3: pva nanoparticles loaded with propolis in the ratio of 1.0 %, and pva-n4: pva nanoparticles loaded with propolis in the ratio of 1.2 %. a-d: means followed by the same lowercase letters in the same column showed no statistical difference at the 5 % significance level applying-test between the tested samples. https://doi.org/10.5599/admet.1740 benazir subaşı-zarbaliyev et al. admet & dmpk 00(0) (2023) 000-000 6 pva-np0 had lower average particle size values (106.37 nm) compared to the other formulations. the average particle size of pva-nps loaded with the aforementioned five concentrations of propolis was found to be between 106.37 and 258.51 nm. moreover, an increase in the loaded propolis ratio led to an increase in the average particle size except for pva-np4. however, no significant differences were found between pva-np1, pva-np2 and pva-np4 (p>0.05). the ζ-potential is a very good index for the specification of the surface electrical status (the electrostatic or charge repulsion/attraction) of colloidal systems. this parameter was used for the evaluation of the encapsulation of the core material in the nanoparticle and the stability or aggregation of the nanoparticle in the suspension. all the ζ-potential values were in the -0.11 to 3.09 mv range with no statistically significant influence of propolis ratio between pva-np0/pva-np1 or pva-np2/pva-np3/pva-np4 (p>0.05). in the current study, pva-nps had zeta potential values near zero with slight differences depending on the propolis concentration. increasing propolis caused a more negative surface potential, while the highest ζ-potential was measured in pva-np0 (3.09 mv). the ζ-potential values between -30 and +30 mv are considered as the formation of stable water suspensions [13], showing that the fabricated pva-nps formed unstable colloidal dispersions due to the low electrostatic repulsion. ζ-potential was lower (more negative) as the propolis ratio increased, but the variations were not very high. similar findings were reported by baysan et al. [20], who found the ζ-potential values of the propolis-loaded capsules prepared with different coating material combinations (whey protein isolate, gelatin, na-caseinate, maltodextrin, lactose, and gum arabic) in the range of −6.16 to 0.42 mv. the reduction in diameter from pva-np4 to pva-np0 was probably due to the changing surface charge of the pva-nps and the agglomeration of nanoparticles in an aqueous medium. pdi is one of the best indicators to evaluate the particle size distribution in suspensions and estimate the physical and rheological features of the colloidal systems (e.g., stability, the release of core material and encapsulation efficiency) [21]. lower pdi means more homogeneous particle size distribution. pdi values of the samples varied from 0.22 to 0.48 (table 1), indicating that the fabricated nanoparticles were relatively homogeneous samples with intermediate pdi. we can interpret the particle size distribution of pva-nps as extremely polydisperse nor broad, as previously noted by wrona et al. [22]. the use of different concentrations of propolis had a significant effect on pdi (p<0.05), but pva-np3 and pva-np4 exhibited statistically no significant differences (p>0.05). additionally, pva-np0 and pva-np2 had a more homogenous particle size distribution as compared to the other pva-nps. in a previous study, particle size (200-280 nm) on a nanometric scale, the ζ-potential values (−12.7 to -33.5 mv) and pdi (0.089 to 0.169) for the polymeric nanoparticles loaded with brazilian red propolis extract were determined by do nascimento et al. [23]. one recent research also reported the average particle (208.1-362.2 nm), pdi (0.23-0.26) and ζ-potential (-26.9 to -31.7 mv) of propolis-loaded zein/caseinate/alginate nanoparticles [24]. our results showed that the fabricated pva-nps could be used for food and pharmaceutical purposes due to their small average particle size and moderate pdi values (<0.4), as also supported by gonçalves and co-workers [25]. encapsulation efficiency the encapsulation efficiency gives information corresponding to the retention degree of the core material within the hollow structure of the coating material(s) compared to the reference core amount incorporated into the encapsulation system. pva-np3 and pva-np1 had the highest (25.32 %) (p<0.05) and the lowest (5.09 %) encapsulation efficiency values, respectively (table 1). increasing propolis concentration resulted in higher encapsulation efficiency values up to 1 %, then decreased. this indicates that the excessive use of core material does not always give a better encapsulation yield [26]. similar findings were reported by durán et al. [17], who found encapsulation efficiency levels around 25 % for poly(e-caprolactone) microparticles admet & dmpk 00(0) (2023) 000-000 polyvinyl alcohol nanoparticles loaded with propolis extract doi: https://doi.org/10.5599/admet.1740 7 loaded with propolis extract. in another study conducted by baysan et al. [27], the encapsulation efficiency values of maltodextrin and whey protein isolate powder particles containing propolis extract were between 29.79 and 99.73%. additionally, zhang et al. [24] used zein/dissociated sodium caseinate micelles/alginate nanoparticles for the encapsulation of propolis and noted the encapsulation efficiency as 76.2-86.5 %, depending on the applied method. ina recent study conducted by shakoury et al. [28], the encapsulation efficiency of whey protein nanoparticles loaded with propolis extract was in the range of 65.49 to 84.03 %. structural characterization by atr-ft-ir the ftir spectra of pva-np0, pva-np1, pva-np2, pva-np3 and pva-np4 obtained in the range 4000 to 400 cm−1 were given in figure 1. the wavelengths near 3336 and 1347 cm-1 could be related to the phenolic hydroxyl groups [29]. figure 1. ftir spectra of the pva-nps; pva: polyvinyl alcohol, pex: propolis extract, pva-n0: neat pva nanoparticles, pva-n1: pva nanoparticles loaded with propolis in the ratio of 0.4 %, pva-n2: pva nanoparticles loaded with propolis in the ratio of 0.8%, and pva-n3: pva nanoparticles loaded with propolis in the ratio of 1.0 %, and pva-n4: pva nanoparticles loaded with propolis in the ratio of 1.2 %. the wavelengths found at 1725-1705 cm-1 can be ascribed to the c=o stretch in aliphatic ketones [23]. the band at 1670 to 1600 cm−1 could be related to the c=c stretch of conjugated aromatic rings [30]. the observed bands near 1400 and 1340 cm−1 correspond to ch2 bending, oh in-plane bending, and the c-h deformation (ch3 or o-h in-plane bending) [31]. additionally, the weak peak of the stretch of aromatic ether c-o bond (for flavonoids) appeared around 1045 cm−1. the propolis extract spectra also exhibited the band's disappearance at 1043 cm−1 from pva-nps. furthermore, corresponding to the pva-nps, no new bands were determined. in the ftir spectra of the propolis extract, the wavelength around 1045 cm−1 shifted to the higher band (1083 cm−1) with the fabrication of pva-nps. moreover, molecules containing the plane of aromatic c-h regarding the angular deformation were attributed to near 877 cm−1 [32]. similar ft-ir spectra were also given for the polymeric nanoparticles loaded with brazilian red propolis extract by do nascimento et al. [23]. overall, it can 0 400 800 1200 1600 2000 2400 2800 3200 3600 4000 0.00 0.04 0.08 0.12 wavenumber, cm-1 tr a n sm it ta n ce , % pp0.4 pp0.8 pp1 pp1.2 pp0 pex https://doi.org/10.5599/admet.1740 https://www.sciencedirect.com/topics/biochemistry-genetics-and-molecular-biology/maltodextrin benazir subaşı-zarbaliyev et al. admet & dmpk 00(0) (2023) 000-000 8 be inferred from the ftir spectra that the encapsulation was successfully performed and a good interaction between the constituents of the system was achieved. thermal characterization by dsc the thermographs of pva-np0, pva-np1, pva-np2, pva-np3, and pva-np4 were illustrated in figure 2 and glass transition temperature (tg), onset temperature (to), peak temperature (tp) and ending temperature (te) values were determined. as can be seen, while the dsc thermographs of pva-np1, pva-np2, pva-np3, and pva-np4 were similar, whereas pva-np0 possessed a broader endothermic peak from 242 °c to 390 °c with a maximum temperature of 305 °c. it can be due to the changes in thermal characteristics of pva with the addition of propolis. the pva-np0 depicted the three endotherm events with to at 44, 150 and 242 °c, tp at 105, 225 and 305 °c and te at 150, 237.5 and 390 °c and its corresponding tg was determined at 50 °c. figure 2. dsc thermograms of the pva-nps; pva: polyvinyl alcohol, pva-n0: neat pva nanoparticles, pva-n1: pva nanoparticles loaded with propolis in the ratio of 0.4 %, pva-n2: pva nanoparticles loaded with propolis in the ratio of 0.8 %, and pva-n3: pva nanoparticles loaded with propolis in the ratio of 1.0 %, and pva-n4: pva nanoparticles loaded with propolis in the ratio of 1.2 %. encapsulation of propolis extract into the pva polymer matrix resulted in four endothermic peaks. our findings supported the results obtained by do nascimento et al. [23], who also observed four endotherm events when brazilian red propolis extract was encapsulated into the polymeric nanoparticles. additionally, the formation of extra small endothermic peaks at around 40-45 °c could be due to the little crystalline nature of the particles. it was also determined that to of the third endothermic peak for pva-np0 shifted to higher temperatures after encapsulation of propolis, which implied that encapsulation supplied thermal protection of propolis. variations of peak positions indicated the interactions between pva and propolis extract that led to the formation of new structural organizations. among the tested pva-nps, in general, pva-np1 and pva-np4 exhibited similar endothermic peaks with lower temperature values, while pva-np3 showed endothermic peaks at greater degrees (tp of 45, 137, 223, and 325 °c). moreover, the dsc thermogram of the pva-np2 illustrated four endothermic peaks at tp of 40, 128, 223, and 329 °c. nascimento et al. [23] reported that the observed endothermic peak around 135 °c for the polymeric nanoparticles loaded with brazilian red propolis extract was related to the fusion processes of low molecular weight compounds such as flavonoids in the mixture and other phenolic compounds found in the propolis extract. the findings revealed that propolis-loaded -11.0 -9.0 -7.0 -5.0 -3.0 -1.0 0.0 80.0 160.0 240.0 320.0 400.0 h e a t fl o w , w g -1 temperature, °c pp0.4 pp0.8 pp1 pp1.2 pp0 admet & dmpk 00(0) (2023) 000-000 polyvinyl alcohol nanoparticles loaded with propolis extract doi: https://doi.org/10.5599/admet.1740 9 nanoparticles were decomposed at a higher temperature as compared to their neat forms, showing that the thermal stability of pva-nps was enhanced by the incorporation of propolis into the pva polymer matrix. dpph radical scavenging activity the dpph scavenging activity (table 1) of the pva-nps was found in the following order: pva-np4 (89.25 %) > pva-np3 (87.32 %) > pva-np2 (85.09 %) > pva-np1 (78.13 %) >pva-np0 (0 %), showing that the dpph-radical scavenging abilities of the samples were correlated with the loaded propolis level. the findings showed that the neat pva nanoparticle formulation (pva-np0) possessed no antioxidant activity. additionally, no significant differences were determined in the radical scavenging activities of pva-np3 and pvanp4 (p>0.05). our findings were supported by the results obtained by aytekin et al. [33], who reported that dpph scavenging activities of various propolis liposomal formulations varied from 76.21 to 100.3 %. in another study conducted by andrade and co-workers [34], dpph inhibition levels of the microcapsules containing spray-dried brazilian brown, green and red propolis were found between 82.00 and 89.00 %. antibacterial activity the antibacterial activity of pva-np0, pva-np1, pva-np2, pva-np3 and pva-np4 were tested against two pathogenic bacterial strains (e. coli o157:h7 and s. aureus) and the findings were given in table 2. the type of pva-nps and the bacterium strain affected the bacterium growth inhibition. pva-np0, which does not contain propolis did not show any inhibitory activity against any of the pathogens, which shows the ineffectiveness of pva itself. among the studied pathogenic strains, propolis-loaded pva-nps exhibited stronger antibacterial activity against the gram-positive s. aureus, where a complete elimination was achieved by the effect of pva-np2, pva-np3 and pva-np4. a significant reduction in e. coli numbers was also provided by the application of propolis-loaded pva-nps. in a previous study, przybyłek and karpiński [35] reported that the propolis extract had more bacterial inhibition capability on gram-positive strains as compared to the gram-negative bacteria. this could be due to variations in the structural composition of cells and the cell wall of gram-positive and gram-negative bacteria [33,36]. table 2. the effects of the pva-nps on the growth inhibition of e. coli o157:h7 and s.aureus. name of strain log (antibacterial activity / cfu ml-1) sample control pva-n0 pva-n1 pva-n2 pva-n3 pva-n4 e. coli 8.20 ± 0.10a 8.52 ± 0.09a 8.15 ± 0.17b 8.25±0.11b 8.03 ± 0.19c 8.12 ± 0.09b s. aureus 8.24 ± 0.08a 8.06 ± 0.07a 4.37 ± 0.33b 0.00 ± 0.00c 0.00 ± 0.00c 0.00 ± 0.00c cfu: colony-forming units, pva: polyvinyl alcohol, pva-n0: neat pva nanoparticles, pva-n1: pva nanoparticles loaded with propolis in the ratio of 0.4 %, pva-n2: pva nanoparticles loaded with propolis in the ratio of 0.8 %, and pva-n3: pva nanoparticles loaded with propolis in the ratio of 1.0 %, and pva-n4: pva nanoparticles loaded with propolis in the ratio of 1.2 %.a-d: means followed by the same lowercase letters in the same line indicates that there was no statistical difference at the 5 % significance level applyingtest between the tested samples. the results also demonstrated that propolis extract kept its antibacterial activity after nanoencapsulation. the findings regarding s. aureus agreed with the observations reported by previous researchers [37,38]. similarly, almuhayawi [39] reported that the antibacterial effects of propolis were dependent on several factors, such as the origin of propolis and the type of bacterial strains. the results were also in accordance with those of lu et al. [40], who reported that s. aureus was almost susceptible to the ethanolic extract of propolis. the antimicrobial activity of propolis could be due to the presence of hydrophilic and hydrophobic phenolic compounds (e.g., such as flavonoids, aromatic esters and acids) that may directly influence the cell wall of the bacteria [33]. https://doi.org/10.5599/admet.1740 https://www.sciencedirect.com/science/article/pii/s1319562x21010287#b0470 https://www.sciencedirect.com/science/article/pii/s1319562x21010287#b0025 https://www.sciencedirect.com/science/article/pii/s0963996922007098#b0230 benazir subaşı-zarbaliyev et al. admet & dmpk 00(0) (2023) 000-000 10 morphological characterization by sem sem images of the pva-nps are shown in figure 3 (a-e). corresponding to the five different sem images (magnification 40000), various particle shapes and mean particle size diameters were seen depending on the five formulations. sem images indicated no cracks or holes on the surface of pva-nps. sem results clearly revealed that pva-nps had a spherical shape with a rough surface and were composed of long and thin fibres at nanometre thickness. the first reason for the formation of nanoparticles and nanofibers in the images could be the applied high voltage to the solutions so-called as taylor-cone based spraying effect [41]. according to this theory, the droplets were charged by the high voltage and the surface tension of the polymer solution was overcome by the arising repulsive coulombic forces. then, the taylor cone was formed and when the voltage exceeded the limits, cone-jet formed based on polymer concentration resulting in thin fibre formation [42]. figure 3. morphology and size distribution of pva nanoparticles loaded with propolis in the ratio of 0 % (a), 0.4 % (b), 0.8 % (c), 1.0 % (d) and 1.2 % (e). additionally, another possible reason could be due to the high viscosity of the coating material. earlier studies reported that the increase in the viscosity of the solution resulted in spindle-like droplets with pure fibres [43]. sem examination showed that the mean particle size diameter of the pva-nps was measured in (a) (b) (d) (c) (e) admet & dmpk 00(0) (2023) 000-000 polyvinyl alcohol nanoparticles loaded with propolis extract doi: https://doi.org/10.5599/admet.1740 11 the range of 100 to 250 nm, confirming the particle size determined by the dynamic light scattering analysis. although we determined that the dispersions of pva-nps were uniform and homogeneous based on the pdi measurement results, sem photographs showed that the pva-nps were non-homogenous. this can be linked with the quickly absorbed moisture from humid air by the pva-nps and this phenomenon led to the particle agglomeration. conclusions propolis has been popular in recent years for its health-related and pharmaceutical properties. however, it cannot be consumed in its original form due to its bitter and unpleasant taste. its insolubility nature in water and the easy loss of its antioxidant activity and phenolic compounds are other limitations of propolis in the food industry. in the current work, propolis at different concentrations (0, 0.4, 0.8, 1.0, and 1.2 %) were encapsulated into pva polymer matrix using the electrospraying method. the dynamic light scattering analysis showed that the nanoparticles had nanoscale particle size with moderate pdi values and formed unstable colloidal dispersions. encapsulation efficiency of up to 25 % and high dpph scavenging activity of up to 89.25 % were found for propolis-loaded pva nanoparticles. ft-ir analysis showed the successful propolis encapsulation and interaction between the constituents. dsc analysis revealed that the thermal stability of propolis-loaded pva nanoparticles was higher in comparison to their neat nanoparticle forms. regarding antibacterial activity, the loaded nanoparticles showed stronger antibacterial activity against s. aureus than e. coli. sem images demonstrated that the nanoparticles had a spherical shape with a rough surface and were composed of long and thin fibres with nanometre thickness. overall, the current work confirmed that nanoencapsulation of propolis using electrospraying technique could offer a new and effective approach to overcome its limitations for utilization in the food and pharmaceutical industries. acknowledgment: this research has been supported by yildiz technical university scientific research projects coordination department. project number: 2016-07-05-yl01. conflict of interest: the author declares no conflict of interest. references [1] b.f. gibbs, s. kermasha, b. inteaz al. encapsulation in the food industry. international journal of food sciences and nutrition 50(3) (1999) 213-224. https://doi.org/10.1080/096374899101256 [2] r. biswas, m. alam, a. sarkar, i. haque, m. hasan, m. hoque. application of nanotechnology in food: processing, preservation, packaging and safety assessment. heliyon 8(11) (2022) e11795. https://doi.org/10.1016/j.heliyon.2022.e11795 [3] o. salata, o. tools of nanotechnology: electrospray. current nanoscience 1(1) (2005) 25-33. https://doi.org/10.2174/1573413052953192 [4] n. bhardwaj, s. c. kundu. electrospinning: a fascinating fiber fabrication technique. biotechnology advances 28 (3) (2010) 325-347. https://doi.org/10.1016/j.biotechadv.2010.01.004 [5] m.t. yilmaz, a. yilmaz, p.k. akman, f. bozkurt, e. dertli, a. basahel, b. al-sasi, o. taylan, o. sagdic. electrospraying method for fabrication of essential oil loaded-chitosan nanoparticle delivery systems characterized by molecular, thermal, morphological and antifungal properties. innovative food science & emerging technologies 52 (2019) 166-178. https://doi.org/10.1016/j.ifset.2018.12.005 [6] p. supaphol, s. chuangchote. on the electrospinning of poly(vinyl alcohol) nanofiber mats: a revisit. journal of applied polymer science 108(2) (2008) 969-978. https://doi.org/10.1002/app.27664 [7] v. sedlařík, n. saha, i. kuřitka, p. sáha. preparation and characterization of poly (vinyl alcohol)/lactic acid compounded polymeric films. international journal of polymer analysis and characterization 11(4) (2006) 253-270. https://doi.org/10.1080/10236660600750190 https://doi.org/10.5599/admet.1740 https://doi.org/10.1080/096374899101256 https://doi.org/10.1016/j.heliyon.2022.e11795 https://doi.org/10.2174/1573413052953192 https://doi.org/10.1016/j.biotechadv.2010.01.004 https://doi.org/10.1016/j.ifset.2018.12.005 https://doi.org/10.1002/app.27664 https://doi.org/10.1080/10236660600750190 benazir subaşı-zarbaliyev et al. admet & dmpk 00(0) (2023) 000-000 12 [8] g. a. burdock. review of the biological properties and toxicity of bee propolis (propolis). food and chemical toxicology 36(4) (1998) 347-363. https://doi.org/10.1016/s0278-6915(97)00145-2 [9] a.f. wali, a. mushtaq, m.u. rehman, s. akbar, m.h. masoodi. bee propolis (bee’s glue): a phytochemistry review. journal of critical reviews 4(4) (2017) 9-13. https://www.jcreview.com/ admin/uploads/files/61c7167c4e7c21.89487709.pdf [10] p.p. wieczorek, n. hudz, o. yezerska, v. horčinová-sedláčková, m. shanaida, o. korytniuk, i. jasickamisiak. chemical variability and pharmacological potential of propolis as a source for the development of new pharmaceutical products. molecules 27(5) (2022) 1600. https://doi.org/10.3390/molecules27051600 [11] r. bilginer, a. arslan yildiz. a facile method to fabricate propolis enriched biomimetic pva architectures by co-electrospinning. materials letters (2020) 128191. https://doi.org/10.1016/ j.matlet.2020.128191 [12] v. m. busch, a. pereyra-gonzalez, n. šegatin, p.r. santagapita, n. poklarulrih, m.p. buera. propolis encapsulation by spray drying: characterization and stability. lwt 75 (2017) 227-235. https://doi.org/10.1016/j.lwt.2016.08.055 [13] m.a. çakır, n. c. icyer, f. tornuk. optimization of production parameters for fabrication of thymolloaded chitosan nanoparticles. international journal of biological macromolecules 151 (2020) 230238. https://doi.org/10.1016/j.ijbiomac.2020.02.096 [14] m.p. nori, c.s. favaro-trindade, s. matias de alencar, m. thomazini, j.c. de camargo balieiro, c.j. contreras castillo. microencapsulation of propolis extract by complex coacervation. lwt food science and technology 44(2) (2011) 429-435. https://doi.org/10.1016/j.lwt.2010.09.010 [15] d. sözeri-atik, e. bölük, f. bildik, f. altay, e. torlak, a.a. kaplan, b. kopuk, i̇. palabıyık. particle morphology and antimicrobial properties of electrosprayed propolis, food packaging and shelf life 33 (2022) 100881. https://doi.org/10.1016/j.fpsl.2022.100881 [16] s.j. yang, j.j. wu, y.c. wang, c.f. huang, t.m. wu, c.j. shieh, c.m.j. chang. encapsulation of propolis flavonoids in a water soluble polymer using pressurized carbon dioxide anti-solvent crystallization. the journal of supercritical fluids 94 (2014) 138-146. https://doi.org/10.1016/j.supflu.2014.07.009 [17] n. durán, p.d. marcato, c.m. buffo, m.m. de azevedo, e. espósito. poly (epsiloncaprolactone)/propolis extract: microencapsulation and antibacterial activity evaluation, pharmazie 62 (2007) 287-290. https://doi.org/10.1691/ph.2007.4.6058 [18] g. c. feyzioglu, f. tornuk. development of chitosan nanoparticles loaded with summer savory (satureja hortensis l.) essential oil for antimicrobial and antioxidant delivery applications. lwt food science and technology 70 (2016) 104-110. https://doi.org/10.1016/j.lwt.2016.02.037 [19] s.m. jafari, e. assadpoor, y. he, b. bhandari. encapsulation efficiency of food flavours and oils during spray drying. drying technology 26(7) (2008) 816-835. https://doi.org/10.1080/07373930 802135972 [20] u. baysan, a. zungurbastıoğlu, n. ö. coşkun, d. konuktakma, e. ülkeryıldızbalçık, h. sahin-nadeem, m. koç. the effect of coating material combination and encapsulation method on propolis powder properties. powder technology 384 (2021) 332-341. https://doi.org/10.1016/j.powtec.2021.02.018 [21] s. hasani, s.m. ojagh, m. ghorbani. nanoencapsulation of lemon essential oil in chitosan-hicap system. part 1: study on its physical and structural characteristics. international journal of biological macromolecules 115 (2018) 143-151. https://doi.org/10.1016/j.ijbiomac.2018.04.038 [22] m. wrona, m.j. cran, c. nerín, s.w. bigger. development and characterisation of hpmc films containing pla nanoparticles loaded with green tea extract for food packaging applications. carbohydrate polymers 156 (2017) 108-117. https://doi.org/10.1016/j.carbpol.2016.08.094 [23] t.g. do nascimento, p.f. da silva, l. f. azevedo, l. g. da rocha, i.c. de moraes porto, t.f.a. lima e moura, i.d. basílio-júnior, l.a. grillo, c.b. dornelas, e.j. fonseca, e. de jesus oliveira, a.t. zhang, d.g. watson. polymeric nanoparticles of brazilian red propolis extract: preparation, characterization, antioxidant and leishmanicidal activity. nanoscale research letters 11(1) (2016). http://dx.doi.org/ 10.1186/s11671-016-1517-3 https://doi.org/10.1016/s0278-6915(97)00145-2 https://www.jcreview.com/admin/uploads/files/61c7167c4e7c21.89487709.pdf https://www.jcreview.com/admin/uploads/files/61c7167c4e7c21.89487709.pdf https://doi.org/10.3390/molecules27051600 https://doi.org/10.1016/j.matlet.2020.128191 https://doi.org/10.1016/j.matlet.2020.128191 https://doi.org/10.1016/j.lwt.2016.08.055 https://doi.org/10.1016/j.ijbiomac.2020.02.096 https://doi.org/10.1016/j.lwt.2010.09.010 https://doi.org/10.1016/j.fpsl.2022.100881 https://doi.org/10.1016/j.supflu.2014.07.009 https://doi.org/10.1691/ph.2007.4.6058 https://doi.org/10.1016/j.lwt.2016.02.037 https://doi.org/10.1080/07373930802135972 https://doi.org/10.1080/07373930802135972 https://doi.org/10.1016/j.powtec.2021.02.018 https://doi.org/10.1016/j.ijbiomac.2018.04.038 https://doi.org/10.1016/j.carbpol.2016.08.094 http://dx.doi.org/10.1186/s11671-016-1517-3 http://dx.doi.org/10.1186/s11671-016-1517-3 admet & dmpk 00(0) (2023) 000-000 polyvinyl alcohol nanoparticles loaded with propolis extract doi: https://doi.org/10.5599/admet.1740 13 [24] h. zhang, y. fu, y. xu, f. niu, z. li, c. ba, b. jin, g. chen, x. li. one-step assembly of zein/caseinate/alginate nanoparticles for encapsulation and improved bioaccessibility of propolis. food & function 10 (2018) 635-645. http://dx.doi.org/10.1039/c8fo01614c [25] m. c. f. gonçalves, o. mertins, a.r. pohlmann, n.p. silveira, s.s. guterres. chitosan coated liposomes as an innovative nanocarrier for drugs. journal of biomedical nanotechnology 8(2) (2012) 240-250. https://doi.org/10.1166/jbn.2012.1375 [26] m. soleimanifar, s.m. jafari, e. assadpour. encapsulation of olive leaf phenolics within electrosprayed whey protein nanoparticles; production and characterization. food hydrocolloids (2019) 105572. https://doi.org/10.1016/j.foodhyd.2019.105572 [27] u.baysan, f. elmas f, m. koç.the effect of spray drying conditions on physicochemical properties of encapsulated propolis powder. journal of food processing engineering 42(4) (2019) e13024. https://doi.org/10.1111/jfpe.13024 [28] n. shakoury, m.a. aliyari, m. salami, z. emam-djomeh, b. vardhanabhuti, a.a. moosavi-movahedi. encapsulation of propolis extract in whey protein nanoparticles, lebensmittel-wissenschaft und technologie 158 (2022) 113138. https://doi.org/10.1016/j.lwt.2022.113138 [29] z. liu, j. chen, g. knothe, x. nie, j. jiang. synthesis of epoxidized cardanol and its antioxidative properties for vegetable oils and biodiesel. acs sustainable chemistry & engineering 4(3) (2016) 901906. https://doi.org/10.1021/acssuschemeng.5b00991 [30] b.c. smith (2011). fundamentals of fourier transform infrared spectroscopy, second edition, crc press, taylor & francis group. https://doi.org/10.1201/b10777 [31] h. barud, a.m. de araújo júnior, s. saska, l.b. mestieri, j. a. d. b. campos, r. m. de freitas, n.u. ferreira, a.p. nascimento, f.g. miguel, vaz m. m. de o.l.l., e.a. varizon, f. marquele-oliveira, a.m.m. gaspar, s.j.l. ribeiro, a. a. berretta. antimicrobial brazilian propolis (epp-af) containing biocellulose membranes as promising biomaterial for skin wound healing. evidence-based complementary and alternative medicine (2013) 703024. https://doi.org/10.1155/2013/703024 [32] s.m.a. el-sheikh, a.a.f.a. el-alim, h.a.f. ibrahim, e.a. mobarez, d.m.a. el-masry, w.a. el-sayed. preparation, characterization and antibacterial activity of chitosan nanoparticle and chitosanpropolis nanocomposite. advences in animal veterinary sciences 7(s2) (2019) 183-190. https://nexusacademicpublishers.com/uploads/files/aavs_7_s2_183-190.pdf [33] a. a. aytekin, s. t. tanrıverdi, f. aydınköse, d. kart, i̇. eroğlu, ö. özer. propolis loaded liposomes: evaluation of antimicrobial and antioxidant activities. journal of liposome research 30(2) (2020) 107-116. https://doi.org/10.1080/08982104.2019.1599012 [34] j. k. s. andrade, m. denadai, g. r. s. andrade, c. da cunha nascimento, p. f. barbosa, m. s. jesus, n. narain. development and characterization of microencapsules containing spray dried powder obtained from brazilian brown, green and red propolis. food research international 109 (2018) 278287. https://doi.org/10.1016/j.foodres.2018.04.048 [35] i. przybyłek, t.m. karpiński. antibacterial properties of propolis. molecules 24(11) (2019) 2047. https://doi.org/10.3390/molecules24112047 [36] s. bakirdere, m.t. yilmaz, m. f. tornuk, s. keyf, a. yilmaz, o. sagdic, b. kocabas. molecular characterization of silver-stearate nanoparticles (agstnps): a hydrophobic and antimicrobial material against foodborne pathogens. food research international 76 (2015) 439-448. https://doi.org/ 10.1016/j.foodres.2015.08.005 [37] r. béji-srairi, i. younes, m. snoussi, k. yahyaoui, g. borchard, r. ksouri, v. frachet, m.k. wided. ethanolic extract of tunisian propolis: chemical composition, antioxidant, antimicrobial and antiproliferative properties. journal of apicultural research 59(5) (2020) 917-927. https://doi.org/ 10.1080/00218839.2020.1732572 [38] s. bouchelaghem. propolis characterization and antimicrobial activities against staphylococcus aureus and candida albicans, saudi journal of biological sciences 29(4) (2022) 1936-1946. https://doi.org/10.1016/j.sjbs.2021.11.063 https://doi.org/10.5599/admet.1740 http://dx.doi.org/10.1039/c8fo01614c https://doi.org/10.1166/jbn.2012.1375 https://doi.org/10.1016/j.foodhyd.2019.105572 https://doi.org/10.1111/jfpe.13024 https://doi.org/10.1016/j.lwt.2022.113138 https://doi.org/10.1021/acssuschemeng.5b00991 https://doi.org/10.1201/b10777 https://doi.org/10.1155/2013/703024 https://nexusacademicpublishers.com/uploads/files/aavs_7_s2_183-190.pdf https://doi.org/10.1080/08982104.2019.1599012 https://doi.org/10.1016/j.foodres.2018.04.048 https://doi.org/10.3390/molecules24112047 https://doi.org/10.1016/j.foodres.2015.08.005 https://doi.org/10.1016/j.foodres.2015.08.005 https://doi.org/10.1080/00218839.2020.1732572 https://doi.org/10.1080/00218839.2020.1732572 https://doi.org/10.1016/j.sjbs.2021.11.063 benazir subaşı-zarbaliyev et al. admet & dmpk 00(0) (2023) 000-000 14 [39] m.s. almuhayawi. propolis as a novel antibacterial agent. saudi journal of biological sciences 27(11) (2020) 3079-3086. https://doi.org/10.1016/j.sjbs.2020.09.016 [40] l.c. lu, y.w. chen, c.c. chou. antibacterial activity of propolis against staphylococcus aureus. international journal of food microbiology 102(2) (2005) 213-220. https://doi.org/10.1016/j.ijfood micro.2004.12.017 [41] m. wilm. principles of electrospray ionization. molecular & cellular proteomics 10(7) (2011) m111.009407. https://doi.org/10.1074/mcp.m111.009407 [42] a.k. jain, v. sood, m. bora, r. vasita, d.s. katti. electrosprayed inulin microparticles for microbiota triggered targeting of colon. carbohydrate polymers 112 (2014) 225-234. https://doi.org/10.1016/ j.carbpol.2014.05.087 [43] c. mit-uppatham, m. nithitanakul, p. supaphol. effects of solution concentration, emitting electrode polarity, solvent type, and salt addition on electrospun polyamide-6 fibers: a preliminary report. macromolecular symposia 216(1) (2004) 293-300. https://doi.org/10.1002/masy.200451227 ©2023 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) https://doi.org/10.1016/j.sjbs.2020.09.016 https://doi.org/10.1016/j.ijfoodmicro.2004.12.017 https://doi.org/10.1016/j.ijfoodmicro.2004.12.017 https://doi.org/10.1074/mcp.m111.009407 https://doi.org/10.1016/j.carbpol.2014.05.087 https://doi.org/10.1016/j.carbpol.2014.05.087 https://doi.org/10.1002/masy.200451227 http://creativecommons.org/licenses/by/3.0/ old and modern antibiotic structures with potential for today’s infections doi: http://dx.doi.org/10.5599/admet.1272 131 admet & dmpk 10(2) (2022) 131-146; doi: https://doi.org/10.5599/admet.1272 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index review old and modern antibiotic structures with potential for today’s infections david j. newman nih special volunteer, wayne, pa 19087 us. email: djnewman664@verizon.net. received: january 31, 2022; revised: february 13, 2022; available march 04, 2022 abstract due to the lack of new antibiotics with efficacy against the eskape and other resistant microbes, coupled to the demise of major pharmaceutical company antibiotic discovery programs, due to a number of factors but mainly roi calculations and the lack of efficacy of combinatorial chemistry as a substitute, the search for novel antibiotics may well have moved to the utilization of older structures with significant synthetic chemistry input. this short review demonstrates how modern synthetic chemistry, when applied to either modification of current resistant antibiotics such as glycopeptides, or production of novel peptidic agents based on natural product sourced antimicrobial peptides (amps) and other potential initial peptide-based agents from genomic searches and baiting techniques, have produced active agents of significant utility. in addition, synthetic chemistry practitioners have now shown that they can produce bioactive molecules of greater than 800 daltons in kilogram quantities under cgmp conditions. ©2022 by the author. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords peptides; synthesis; pathogens introduction with the advent of what might be called the “dangerous sextet” of current bacterial infections viz enterococcus faecium, staphylococcus aureus, klebsiella pneumoniae, acinetobacter baumannii, pseudomonas aeruginosa, and enterobacter spp. (aka the eskape pathogens), came to the frequently belated realization that the antibiotics in current use were not effective as single agents against most, if not all, of these organisms. in a few advanced countries, expensive antibiotic drugs with suitable activities might be available, but these infections ignore economic boundaries. thus, it behooves researchers to reinvestigate and/or modify agents that, in earlier days, were ignored due to their toxicities, difficulty in dosing, chemical stability or side effects due to their unknown interactions. the world health organization included the organisms above in their listing of dangerous microbes, as shown below. on the 17 th of february 2017, they published a list of twelve microbes that were of concern as they were all resistant to a significant number of antibiotics then in current use. this list was meant to be a “wake-up call” to the pharmaceutical industry/scientists involved in the discovery of new antibiotics, be they natural products, semisynthetic or totally synthetic entities, by showing how widespread resistance is. http://dx.doi.org/10.5599/admet.1272 https://doi.org/10.5599/admet.1272 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:djnewman664@verizon.net http://creativecommons.org/licenses/by/4.0/ david j. newman admet & dmpk 10(2) (2022) 131-146 132 on the list (www.who.int/mediacentre/news/releases/2017/bacteria-antibiotics-needed/en/),  the first three (priority 1 listed as critical) were the gram-negative organisms acinetobacter baumannii (carbapenem-resistant); pseudomonas aeruginosa (carbapenem-resistant); enterobacteriaceae (carbapenem-resistant, esbl-producing).  the next 6 (priority 2 listed as high) contained two very well-known gram-positive pathogens, enterococcus faecium (resistant to vancomycin) and staphylococcus aureus (mrsa), and four gramnegative pathogens. helicobacter pylori (clarithromycin-resistant); campylobacter spp. (fluoroquinolone-resistant); salmonellae (fluoroquinolone-resistant) and neisseria gonorrhoeae (cephalosporin-resistant, fluoroquinolone-resistant).  the last three (priority 3 listed as medium) were the gram-positive streptococcus pneumoniae (penicillin-non-susceptible) and the two gram-negative haemophilus influenzae ampicillinresistant) and shigella spp. (fluoroquinolone-resistant). what will be described in this article are examples culled from the recent literature where elegant chemistry and modernized microbiological techniques have led to modifications of old agents and/or the administration of “adjuncts” that has enabled the renewed usage of agents that in the “halcyon days of yore”, which covered from the early 1960s until the effective elimination of large-scale antibiotic discovery programs in “big pharma”, which with a few exceptions would be pre-1990. the demise of combinatorial chemistry as a source of antiinfective lead structures it would be worth commenting here that possibly the major “player” in the elimination of searching for microbial-sourced agents by “big pharma” was the concept of large-scale to very large scale (>10 6 ) syntheses of small compounds via combinatorial chemistry. the idea was that these agents would be patentable and “easily synthesized”, rather than have the large-scale systems necessary for natural product sourced agents. in addition, a major problem with microbial sourcing (the major source used in all earlier cases) was the task(s) associated with what became known as “dereplication”, or the ability to rapidly decide whether an isolated compound was new or old and often known under a variety of names. it should be noted that the advent of >100 mhz nmr, portable computers, accessible compound databases and automated hplcs were way in the future, even in the early 1980s. the irony in the early days of combichem was that the original “successes quoted for increased bioactivity” used peptide-based molecules or small compounds based on bioactive natural products as their starting points. an excellent example of what happened when combichem structures were substituted for products from natural sources and/or their derivatives was the review in 2007, giving the total lack of success when massive numbers of combichem structures met up with ~70 isolated targets at gsk [1]. i should point out at this stage that i was part of the earlier sk&f antibiotics program based on microbial sources and modification of beta-lactam based molecules from 1976 to its sudden demise in april of 1985. that program was certainly more successful than its subsequent descendants. a 2015 paper by blaskovich et al. gives more of the subsequent history of the demise of promising antibiotic programs by large pharma in the subsequent years to 2015 [2]. in the following sections, i will illustrate how, by building upon “old or at times new structures” by competent scientists (chemists and microbiologists), and/or by utilizing modern techniques based on earlier systems, old structures suddenly became significant candidates for clinical trials, and nowadays utilization of the enormous capabilities of genetic manipulation can lead to novel agents with the desired attributes against the eskape pathogens and their subsequent descendants. http://www.who.int/mediacentre/news/releases/2017/bacteria-antibiotics-needed/en/ admet & dmpk 10(2) (2022) 131-146 chemical syntheses and resistant antibiotics doi: http://dx.doi.org/10.5599/admet.1272 133 vancomycin and other glycopeptides; initial and later history for a significant number of years, vancomycin (1) and then later its close chemical relatives were often defined as the “antibiotic of last resort”. vancomycin was originally introduced into clinical medicine by lilly in the middle to late 1950s, though there is a debate in the literature as to the actual date. the integrity r database (now clarivate analytics) uses 1955, but butler et al. in a review in 2014 [3] used 1958. as the fda approval date. however, the actual structure was not fully defined until 1982 when the presence of asparagine was confirmed [4]. resistance to vancomycin occurred relatively early, but before the advent of advanced structural techniques, including amino acid analyses of gram-positive cell walls, the reason for the resistance was not known. with the advancement of knowledge, it became apparent that vancomycin (and its later naturally occurring or semisynthetic “chemical cousins”) inhibited bacterial growth by binding to the l-lys-d-ala-dala-co2h terminals in the cross-links in the gram-positive cell wall. subsequent work demonstrated that the van r phenotype was as a result of the simple change to d-lactate in place of the terminal d-ala residue in the cases of vana, vanb, vand and to d-ser for vanc, vane, vang clinical phenotypes. in addition to vancomycin resistance, s. aureus cultures that were classified as methicillin resistant (mrsa), were also becoming vancomycin-resistant as well. the vancomycin resistance level due to the d-lac substitute was approximately 1000-fold and for the d-ser modification approximately 140-fold. interestingly, the idea that resistance was due to the use of glycopeptide antibiotics in animal feeds that then led to the resistance seen was shown to be inaccurate by two reports. one in 2011 by d’costa et al., [5] and the other by wright et al. [6] the following year, both papers demonstrating that microbes over 10,000 years old isolated from yukon ice fields had similar resistance phenotypes. in the last few years, three semisynthetic glycopeptides entered clinical use in the usa and other countries. in 2009 telavancin (a close relative of vancomycin) was approved by the fda and then in 2014, dalbavancin, which was derived from part of the a40926 complex, and oritavancin derived from chloroeremomycin were approved. the drug complex teicoplanin, which is a mixture of closely related compounds, is used in europe and interestingly, there are some vanr phenotypes that are not resistant to this mixture and the reverse also occurs, though in general, most strains are resistant to all. synthetic modifications of vancomycin the laboratory that has had the greatest effect on “modifying” the basic structure of vancomycin (1), with the aims of reducing resistance and potentially expanding the range of microbes that it can “inhibit”, is, without doubt, the boger laboratory at the scripps research institute in la jolla, california. over the last ten plus years, that laboratory has published papers showing how by making internal modifications to the peptide backbone in this molecule, that were simple in concept but required very clever synthetic chemical processes to achieve, have led to a series of related vancomycin molecules that demonstrated very significant antibiotic activities against mrsa and e. faecalis (both vana & vanb phenotypes) depending upon very simple changes in one position. then the boger group extended the synthetic chemistry to other parts of the base molecule by adding small parts from other glycopeptides in clinical use. structure (2), which was redrawn from the proceedings of the national academy of science paper by okano et al. [7] shows the substitutions used. the compilation of vancomycin structural changes and the tables in that report demonstrated how these various modifications “converted” total resistance to e. faecalis and e. faecium (mics of vancomycin > 250 µg ml -1 ) to molecules with mics from 5 to 0.005 µg ml -1 for these van a/e resistant microbes. http://dx.doi.org/10.5599/admet.1272 david j. newman admet & dmpk 10(2) (2022) 131-146 134 then in 2020, the same group reported the results of further guanidino modifications on the c-terminus of vancomycin that improved antimicrobial activity and appeared to provide a synergistic mechanism of action independent of d-ala-d-ala [8]. by using structure (2) as the base, then coupling the 4chlorobiphenyl)-methyl (cbp) modification, which they had previously shown to give significant increases in activity, with x = o and r = a variety of guanidino substituents, they produced a series of relatively simple modified vancomycins that displayed sub-microgram activities (mic levels) against significant vancomycinresistant clinical specimens. the following is a direct quote from that paper. “a prototypical member of the series, g3-cbpvancomycin (15) exhibits no hemolytic activity, displays no mammalian cell growth inhibition, possesses improved and especially attractive in vivo pharmacokinetic (pk) properties, and displays excellent in vivo efficacy and potency against an especially challenging multidrug-resistant (mrsa) and vana vancomycinresistant (vrsa) staphylococcus aureus bacterial strain.” the structure (15) mentioned in the direct quote above from wu et al. [8] is shown as structure (3) in figure 1. figure 1. vancomycin and boger lab modifications; structures 1 to 3 admet & dmpk 10(2) (2022) 131-146 chemical syntheses and resistant antibiotics doi: http://dx.doi.org/10.5599/admet.1272 135 in addition to the papers referred to above, another recent paper from the boger lab [9] gives an excellent precis of the modified vancomycin derivatives mentioned above, plus others from the boger lab. they comment that these agents are now known in that laboratory as “maxamycins”. this paper is well worth reading to gain further insight into these chemical modifications of a microbial product that first saw “light of day” in the middle to late 1950s, in order to overcome microbial resistance. a comment on modern synthetic methods as a route(s) to cgmp product(s) it should be emphasized at this juncture that synthetic organic chemists have succeeded in the last few years in producing large quantities of cgmp-quality, partial or complete natural products that have either become drugs, eribulin (4) mw 730 being the prime example utilizing data from the kishi synthesis of halichondrin b (5) mw 1111. or nowadays as potential leads to drugs, with the total synthesis of a derivative of halichondrin b e7130 (6); mw 1111 at the 10-gram level, reported by the same eisai group that synthesized eribulin [10], which is currently in phase i clinical trials. these structures are shown in figure 2. figure 2. halichondrin b and synthetic derivatives; structures 4 to 6 tetracycline-like molecules rather than cover the modified tetracyclines omadacycline, eravacycline and saracycline (structures not given), that came from the lederle (then wyeth and now pfizer) laboratories in the 1990s with these variations being approved as drugs in 2018, there were two compounds reported that had the basic nucleus of the tetracyclines but with extensions on the base four-ring scaffold. these compounds were the viridicatumtoxins, now known as a (7) and b (8). they were first reported by hutchinson et al. in 1973 from the fungus penicillium viridicatum [11] as viridicatumtoxin. subsequently the b variant was reported by zheng et al. in 2008 [12], and the synthesis of the b (8) variant was originally reported by the nicolaou group in 2013, [13] with a subsequent structure revision and syntheses of related “microbiologically active compounds” in 2014 [14]. this molecule (as the mixture of enantiomers) was significantly active against resistant strains of e. faecalis, e. faecium and mrsa, and the natural a isomer was comparably active against the same strains. recently the biosynthtetic gene cluster (bgc) for viridicatumtoxin was identified in a. nidulans, a fungus not previously reported to produce this agent. that report also identified the same cluster in p. brasilianum and in three other aspergillus spp [15]. interestingly, the capon group at the university of queensland reported five years earlier in 2015, that the paecilomyces sp., known as cmb-mf010, which they had isolated from the inner tissues of the intertidal pulmonated mollusk siphonaria sp., yielded viridicatumtoxins a (7) and b (8) and the new viridicatumtoxins cf (9-12), and spirohexaline (structure not http://dx.doi.org/10.5599/admet.1272 david j. newman admet & dmpk 10(2) (2022) 131-146 136 given). of these compounds, viridicatumtoxins b (8) and c (9) demonstrated the largest separation of antimicrobial activity versus cell-line cytotoxicity with between 15-to-40-fold differences in favor of the antimicrobial activity. of the six viridicatumtoxins, a (7) and b (8) were the most active against mrsa and vancomycin-resistant e. faecalis. what is also of interest is that the capon group demonstrated that “regular tetracyclines” could be transformed by ring-opening by the paecilomyces sp., mentioned above, perhaps demonstrating a resistance methodology by fungi against these molecules, since none of the modified agents demonstrated any significant bioactivity [16]. in 2020, the antibacterial mechanism of these agents was proven by li et al. in a report in acs infectious diseases [17]. this report confirmed that, as reported earlier, this series of molecules bind directly to the undecaprenyl pyrophosphate synthases (upps) of e. faecalis, s. aureus and e. coli in a direct and highaffinity manner. this paper is worth consulting as the models can be used for further optimization of these molecules and analogues as upps is an essential component of cell wall biosynthesis. since the nicolaou group have successfully synthesized these molecules, viridicatumtoxin a (7) and b (8) and the capon group have shown the simple difference in c (9), these are excellent candidates for large scale production, synthetically or genomically via modification of the bgc plus fermentation, to provide another candidate for use against members of the gram-positive eskape series. the relevant structures are shown in figure 3. figure 3. viridicatumtoxins; structures 7 to 12 peptidic candidates (other than vancomycin) this section will include both naturally occurring and partially synthesized molecules. in some cases, the molecules will be well-known and not in others. colistins and polymyxins these peptidic antibiotics were frequently the “bane” of microbiologists and chemists in the 1960s to late 1970s when pharmaceutical companies with active antibiotic search programs based on microbial fermentation were in vogue. this was because these compounds and their close chemical relatives, had “for then” significant toxicities when compared to the tetracyclines, aminoglycosides etc. they were usually just “tossed aside” and no further work was done with them, nor with the microbe(s) producing them. admet & dmpk 10(2) (2022) 131-146 chemical syntheses and resistant antibiotics doi: http://dx.doi.org/10.5599/admet.1272 137 with the current advent of multiple resistant microbes, combination therapy including these agents has been utilized, using both polymyxins and colistin (aka polymyxin e). however, in 2016, a report was published [18] giving details of a transmissible plasmid expressing the mcr gene. this gene encoded the phosphoethanolamine transferases (petn) that catalyzed the addition of a phosphoethanolamine moiety to lipid a in the outer membrane of gram-negative bacteria, thus generating resistance to the antibiotics in current use, including these peptidic agents. a major problem with the bacteria that contain this plasmid (or close variants) is that they are easily transmissible as “water-borne” microbes, and thus have spread from their original source in mainland china, to become significant problems in multiple areas, as shown by the 2021 paper by cherak et al [19]. in an effort to produce new variations on the old and well used/known polymyxin b and colistin, a group of researchers from china, south australia and germany recently (2021) published the results of a largescale program to discover potential agents against resistant a. baumannii [20]. at the time of writing of that paper over 2000 polymyxin analogues had been reported in the literature, with some early and some more recent examples also reported by vaara in 2019 [21]. of the 2000 plus variations referred to earlier, only 3 demonstrated improved activities against gram-negative pathogens. using clever chemistry and sequential alanine substitution, jiang et al [20] derived a potent and novel polymyxin analogue, faddi-287 (13) that folded similarly to pmb3 (14) and demonstrated mics of 0.125-0.5 µg ml -1 against a. baumannii compared to polymyxin b. faddi-287 is covered by the patent wo2015149131 which has also been granted in the usa and china. very recently, the brady group presented details on a semisynthetic, optimized amp based upon a genomic analysis of sequences in bacteria that might well produce agents that are analogues of colistin (15). they reported the modified colistin known as biphenyl macolacin (16) in nature in early 2022 [22]. interestingly a major substitution was the biphenyl group, one that boger used in his successful synthesis of the highly bioactive maxamycins. vancomyxins although this discussion could fit under the discussion of vancomycin modifications above, because it now extends the activity of vancomycin analogues into gram-negative space, and does not come from the boger group, it is probably a better fit in this section. although vancomycin by itself (as described earlier) is effectively inactive against gram-negative bacteria, it will bind to the lipid ii component in gram-negative bacteria if the outer membrane is “made permeable.” it has been known since 1989 that coadministration of suitable “penetrating” agents with vancomycin will synergize the effects of vancomycin against the gram-negative bacterium e. coli [23]. then in 2010, gordon et al demonstrated significant activity of a vancomycin-colistin physical combination (i.e.co-administration of the two antibiotics) [24]. moving to 2021, van groesen et al came up with what they are calling “vancomyxins,” where a short spacer was inserted between vancomycin and what they called the “polymyxin e nonapeptide pmen (17).” depending upon the specific combination of these molecules, they could be linked via the c-terminus of vancomycin, or in other examples, linked to the vancosamine amino group via a slightly more complex method involving a triazole linker, giving rise to a series of compounds [25]. their structures are not given due to their complexity/sizes, but the combinations were evaluated against four gram-negative and five gram-positive bacteria with vancomycin and pmen separately, and then vancomycin plus a constant 8 µg ml -1 of pmen versus six of the vancomyxins. in most cases, one or http://dx.doi.org/10.5599/admet.1272 david j. newman admet & dmpk 10(2) (2022) 131-146 138 more of the vancomyxins demonstrated mics of between 8 and 16 µg ml -1 , against microbes that were resistant to the unpaired molecules with no activities at or below 128 µg ml -1 for the gram-negative microbes except for one outlier, the p. aeruginosa strain (atcc 27853) that gave the only positive result when the unlinked combination was used. as would be expected, against the gram-positive microbes significant activities were seen, even as low as less than 0.008 µg ml -1 for an s. simulans strain that had an mic of 0.125 µg ml -1 against vancomycin alone. thus, in contrast to the very interesting boger compounds referred to earlier, simple combined “old” antibiotics did provide some excellent results. their hemolytic and nephrotoxic activities were assessed by means of a viability assay using conditionally immortalized proximal tubule epithelial cells (ciptecs), with relative mitochondrial activity after 24 h as the end point. in those assays, toxicity arose at concentrations well above the mic values reported. however, in vivo activities and toxicities were not measured and would have to be performed in the future. figure 4. colistin and polymyxin derivatives; structures 13 to 17 admet & dmpk 10(2) (2022) 131-146 chemical syntheses and resistant antibiotics doi: http://dx.doi.org/10.5599/admet.1272 139 antibacterial-active cyclic peptides (excluding polymyxins and colistins) what is not often realized is that the first natural product antibiotic to go into clinical use but not under terms equivalent to formal approval as a drug entity in the modern sense, was not penicillin as most people think, but the cyclic peptide tyrocidine and the probable second, a derivative of “gramicidin” that was initially mentioned in a 1940 report [26], with the structure of tyrocidine a (18), reported later in 1952 [27]. these two molecules are usually not mentioned in any discussion of early antibiotics. however, contrary to what most people in the west may think, the aminoglycoside streptomycin (19) was not the next antibiotic to go into use for battlefield wounds at around the same time as penicillin in world war ii. the next antibiotic to go into use for these conditions was gramicidin s (20), first used by the ussr for battlefield injuries in 1943 [28,29]. further reports in the same basic time frame from the then ussr, covered more clinical aspects [30,31]. then in 1946, gause discussed the large-scale production and chemistry of this agent [32]. even today, variations on this cyclic peptide are still being investigated. very recently there have been three reviews relating to antimicrobial peptides from naturally occurring and synthetic sources. in 2020, an excellent review by lazzaro et al [33] discussed the evolution of antimicrobial peptides. in this paper they suggested that there was evidence showing adaptive maintenance of polymorphism in amp genes, and perhaps adaptative loss of amp activity. they also suggested as shown earlier with the polymyxins/colistins, that synergy might well occur with “conventional antibiotics”. what is interesting, is that they show synergy between magainin 2 and the amp pgla [34] both of which are amps found in the skin of the african frog xenopus laevis and eukaryotes can rapidly deploy multiple distinct classes of amps simultaneously. although it was always thought that amps have broad spectrum activities, genetic disruption of amp genes result in different responses to infections by s. aureus in the beetle tenbrio molitor. similar separation of activities/genes in other insects have caused an alteration in the idea that “broad spectrum responses” may not be the case every time, but specificity in responses appears to be one response across many eukaryotic phyla. in 2021 a group from malaysia [35] reported the results of a literature review from 2011 to 2021 using a variety of available databases. table 1 in that review (which is an open access paper) covers analogies between naturally occurring and synthetic agents, with the rest of the review covering methodologies and some history of both natural and synthetic amps, with an up-to-date series of references [35]. closely following the malaysian paper was one from an italian group covering semisynthetic and synthetic amps, with an extension into some with antifungal and antiviral activities [36]. this review finishes with a comment on the semisynthetic amp murepavadin (21), a cyclic beta hairpin peptidomimetic based on the antimicrobial peptide, protegrin (pg-1). it is nonlytic and targeted the lptd protein transporter in p. aeruginosa. it reached phase iii clinical trials as an iv-preparation against pneumonia, but in july 2019, the two trials were terminated due to high levels of unexpected kidney injury. in december 2020, the uk medicines agency permitted a phase i trial to begin targeting oral inhalation route in people with cystic fibrosis who are infected with p. aeruginosa. an excellent and very recent paper that covers some of the above examples, and other potential agents that can be targeted against the outer membrane(s) of gram-negative microbes, is the review by klobucar and brown in 2022 that should be consulted as they cover other agents as well [37]. http://dx.doi.org/10.5599/admet.1272 david j. newman admet & dmpk 10(2) (2022) 131-146 140 figure 5. early (1943) and late (2010 plus) antibiotics; structures 18 to 24 admet & dmpk 10(2) (2022) 131-146 chemical syntheses and resistant antibiotics doi: http://dx.doi.org/10.5599/admet.1272 141 other peptide-based antibiotic classes with potential due to the ability today to synthesize natural-product derived agents, particularly peptidic molecules under glp and when necessary, cgmp conditions in bulk, once a suitable compound has been identified by using up to date versions of the classical searching methods from microbes, they can be synthesized and/or modified as needed in order to extend testing schedules. two examples and three recent reviews will be discussed under this rubric. the first will be the material found by the lewis’ group at northeastern university, and the second will be what i call the “grasslands dna program” at rockefeller university. teixobactin in 2015, the lewis group at northeastern university reported the isolation of a previously unknown microbe that they discovered by use of what can be best described as a “baiting technique”, though it was similar in concept to a technique used years earlier by a company named “one-drop” that was eventually purchased by diversa, though details were lost when diversa went under financially. the compound that was isolated was active against multiply resistant gram-positive bacteria, and it was found to target pyrophosphate and sugar moieties present in lipid ii and iii. lipid ii is linked to peptidoglycan (target of the glycopeptide antibiotics) and the lipid iii binding is linked to down-regulation of cell wall teichoic acid biosynthesis. over the last few years, there have been three reports on the total synthesis of this agent, plus several dozen reports covering analogues and sar studies as shown by references in gao et al. [38], all using solid-phase techniques. in 2019, gao et al. published the first solution-phase synthesis, thus adding to the synthetic repertoire for this agent and the potential for many analogues [38]. thus there are at least two different routes to the compound and derivatives. an example of modifying the agent and adding activity against p. aeruginosa can be seen in the report by ng et al. in 2018 [39]. grasslands dna in 2014, the brady group at rockefeller published an article in current opinion in microbiology describing the varied then current methods that used metagenomic studies to obtain potential antibiotic information without isolating the microbe(s) [40]. in 2015 charlop-powers and brady published an informatic package that linked geographic analysis and related microbiome data [41]. following on, in 2016, the group published their initial data on “urban park microbiomes” as sources of biosynthetic diversity (hence the title for this section) [42]. in this work, they reported a gene cluster that could be linked to a novel calcium-dependent antibiotic (cda). in 2018 they reported on the culture-independent synthetic-bioinformatic natural product (syn-bnp) discovery approach covering the survey of analogues of paenimucillin a (23) that led to the discovery of paenimucillin c (24) which inhibited the growth of multidrug-resistant a. baumannii isolates, including a rat cutaneous wound model, with mechanistic studies implying a membrane-associated mode of action [43]. in 2020 this report was expanded by the same group giving details of the synthesis of 157 cyclic peptides from 96 nrps clusters. these compounds were then tested against the eskape pathogens yielding nine antibiotics with activities against some of these microbes as well as m. tuberculosis [44]. that same year, the group reported on the culture-independent discovery of the cda compound referred to in their earlier general paper in 2016, demonstrating excellent activities (mic values between 0.1 to 0.8 µg ml -1 ) for malacidin a (25) depending upon the particular resistant organism tested. these http://dx.doi.org/10.5599/admet.1272 david j. newman admet & dmpk 10(2) (2022) 131-146 142 included microbes resistant to β-lactams, tetracyclines, vancomycin and aminoglycosides. the other cda that they reported, malacidin b (26), differs by one methylene group from malacidin a [45]. this paper has over 180 citations in scopus as of the middle of january 2022. then in 2020, the complete synthesis of malacidin a was published by a joint group from the university of hong kong and rockefeller [46]. as a result of the malacidin findings, a second total synthesis of malacidin a and analogues was recently published by the brimble group in new zealand [47]. three recent reviews on other peptidic antibiotics in 2021, two interesting reviews on peptidic antibiotics were published, with a third in 2022. the first in 2021 discussed emerging peptides with therapeutic potential, with one author from the biotech company polyphor in switzerland, which has been involved in clinical trials with such agents, demonstrating the variation in linear and cyclic peptides with significant bioactivities against resistant microbes [48]. the second in 2021, was from a group in china and an interesting part of that review is a long table listing natural, semisynthetic and synthetic agents reported in the last five years. that table makes interesting reading, as it aptly demonstrates what can be done with current synthetic methods [49]. the third, accepted at the very end of december 2021 and currently online, is full of up to date listings of agents, sources, and clinical trial results, plus discussions as to probable and actual mechanisms of action [50]. conclusion synthetic chemists have now shown that they can modify natural product-based antibiotics that have ceased to be viable due to the build-up of multiple resistance profiles. as an example of what can be done today by competent synthetic chemists, the recent review by nicolaou and rigel shows some of the capabilities of experts in this field [51]. to this review can be added the work referred to earlier on the total synthesis of a derivative of halichondrin b e7130 (mw 1111, 6) producing 10 grams of a cgmp product [10], together with the production of 1 gram of cgmp bryostatin 1 by the wender group [52]. the compounds referred to in the previous paragraph and those alluded to in the body of the text are just some examples of the many totally synthetic methods described in the literature for bioactive agents from natural product sources in the relatively recent past, covering compounds in many biological areas. it should also be noted that semi-syntheses starting from a natural product precursor have been a routine technique practiced by many chemists, particularly with marine-sourced bioactive agents and have been covered in detail by many authors. as a result, for “antibiotic discoverers/developers” of today, a close relationship between microbiologists, natural products and synthetic chemists may well be the current optimal linkage to find and subsequently develop novel agents with clinical relevance. in addition, the advent of searchable databases that cover compounds, activities and genomic information should also be part of the mixture required to discover and develop agents against the ever-increasing number of resistant microbes. admet & dmpk 10(2) (2022) 131-146 chemical syntheses and resistant antibiotics doi: http://dx.doi.org/10.5599/admet.1272 143 figure 6. malacidin a and b; structures 25 to 26 conflict of interest. there are no coi’s to report. all work was performed by me including the structures from public sources. references [1] d.j. payne, m.n. gwynn, d.j. holmes, d.l. pompliano. drugs for bad bugs: confronting the challenges of antibacterial discovery. nat. rev. drug discov. 6 (2007) 29-40. https://doi.org/10.1038/nrd2201. [2] m.a.t. blaskovich, j. zuegg, a.g. elliott, m.a. cooper. helping chemists discover new antibiotics. acs infect. dis. 1 (2015) 285-287. https://doi.org/10.1021/acsinfecdis.5b00044. [3] m.s. butler, k.a. hansford, m.a.t. blaskovich, r. halai, m.a. cooper. glycopeptide antibiotics: back to the future. j. antibiot. 67 (2014) 631-644. https://doi.org/10.1038/ja.2014.111. [4] c.m. harris, t.m. harris. structure of the glycopeptide antibiotic vancomycin. evidence for an asparagine residue in the peptide. j. am. chem. soc. 104 (1982) 4293-4295. https://doi.org/10.1021/ja00379a062. [5] v.m. d'costa, c.e. king, l. kalan, m. morar, w.w.l. sung, c. schwarz, d. froese, g. zazula, f. calmels, r. debruyne, g.b. goldiing, h.n. poinar, g.d. wright. antibiotic resistance is ancient. nature biotechnol. 477 (2011) 457-461. https://doi.org/10.1038/nature10388. [6] g.d. wright, h. poinar. antibiotic resistance is ancient: implications for drug discovery. trends microbiol. 20 (2012) 157-159. https://doi.org/10.1016/j.tim.2012.01.002. [7] a. okano, n.a. isley, d.l. boger. peripheral modifications of [ψ[ch2nh]tpg4]vancomycin with added synergistic mechanisms of action provide durable and potent antibiotics. proc. natl. acad. sci., u.s.a. (2017) e5052-e5061. https://doi.org/10.1073/pnas.1704125114. [8] z.-c. wu, m.d. cameron, d.l. boger. vancomycin c‑terminus guanidine modifications and further insights into an added mechanism of action imparted by a peripheral structural modification. acs infect. dis. 6 (2020) 2169-2180. https://doi.org/10.1021/acsinfecdis.0c00258. [9] z.-c. wu, d.l. boger. maxamycins: durable antibiotics derived by rational redesign of vancomycin. acc. chem. res. 53 (2020) 2587-2599. https://doi.org/10.1021/acs.accounts.0c00569. [10] s. kawano, k. ito, k. yahata, k. kira, t. abe, a. t., m. asano, k. iso, y. sato, f. matsuura, i. ohashi, y. matsumoto, m. isomura, t. sasaki, f. t., y. miyashita, y. kaburagi, a. yokoi, o. asano, t. owa, y. kishi. a landmark in drug discovery based on complex natural product synthesis. sci. rep. 9 (2019) 8656. https://doi.org/10.1038/s41598-019-45001-9. [11] r.d. hutchison, p.s. steyn, s.j. van rensburg. viridicatumtoxin, a new mycotoxin from penicillium viridicatum westling. toxicol. appl. pharmacol. 24 (1973) 507−509. https://doi.org/10.1016/0041008x(73)90057-4. http://dx.doi.org/10.5599/admet.1272 https://doi.org/10.1038/nrd2201 https://doi.org/10.1021/acsinfecdis.5b00044 https://doi.org/10.1038/ja.2014.111 https://doi.org/10.1021/ja00379a062 https://doi.org/10.1021/ja00379a062 https://doi.org/10.1038/nature10388 https://doi.org/10.1016/j.tim.2012.01.002 https://doi.org/10.1073/pnas.1704125114 https://doi.org/10.1021/acsinfecdis.0c00258 https://doi.org/10.1021/acs.accounts.0c00569 https://doi.org/10.1038/s41598-019-45001-9 https://doi.org/10.1016/0041-008x(73)90057-4 https://doi.org/10.1016/0041-008x(73)90057-4 david j. newman admet & dmpk 10(2) (2022) 131-146 144 [12] c.-j. zheng, h.-e. yu, e.-h. kim, w.-g. kim. viridicatumtoxin b, a new anti-mrsa agent from penicillium sp. fr11. j. antibiot. 61 (2008) 633−637. https://doi.org/10.1038/ja.2008.84. [13] k.c. nicolaou, c. nilewski, c.r.h. hale, h.a. ioannidou, a. el marrouni, l.g. koch. total synthesis and structural revision of viridicatumtoxin b. angew. chem. int. ed. 52 (2013) 8736-8741. https://doi.org/10.1002/anie.201304691. [14] k.c. nicolaou, c.r.h. hale, c. nilewski, h.a. ioannidou, a. el marrouni, l.g. nilewski, k. beabout, t.t. wang, y. shamoo. total synthesis of viridicatumtoxin b and analogues thereof: strategy evolution, structural revision, and biological evaluation. j. am. chem. soc. 136 (2014) 12137-12160. https://doi.org/10.1021/ja506472u. [15] m.t. drott, r.w. bastos, a. rokas, l.n.a. ries, t. gabaldón, g.h. goldman, n.p. keller, c. greco. diversity of secondary metabolism in aspergillus nidulans clinical isolates. msphere 5 (2020) e0015600120. https://doi.org/10.1128/msphere.00156-20. [16] z. shang, a.a. salim, z. khalil, p.v. bernhardt, r.j. capon. fungal biotransformation of tetracycline antibiotics. j. org. chem. 81 (2016) 6186−6194. https://doi.org/10.1021/acs.joc.6b01272. [17] w. li, l. l., c. zhang, y. cai, q. gao, f. wang, y. cao, j. jlin, j. li, z. shang, w. lin. investigations into the antibacterial mechanism of action of viridicatumtoxins. acs infect. dis. 6 (2020) 1759−1769. https://doi.org/10.1021/acsinfecdis.0c00031. [18] y.-y. liu, y. wang, t.r. walsh, l.-x. yi, r. zhang, j. spencer, y. doi, g. tian, b. dong, x. huang, l.-f. yu, d. gu, h. ren, x. chen, l. lv, d. he, h. zhou, z. liang, j.-h. liu, j. shen. emergence of plasmidmediated colistin resistance mechanism mcr-1 in animals and human beings in china: a microbiological and molecular biological study. lancet infect. dis. 16 (2016) 161-168. https://doi.org/10.1016/s1473-3099(15)00424-7. [19] z. cherak, l. loucif, a. moussi, j.-m. rolain. epidemiology of mobile colistin resistance ( mcr ) genes in aquatic environments. j. global antimicro. resist. 27 (2021) 51-62. https://doi.org/10.1016/j.jgar.2021.07.021. [20] x. jiang, n.a. patil, m.a.k. azad, wickremasinghe, h., h. yu, j. zhao, x. zhang, m. li, b. gong, l. wan, w. ma, p.e. thompson, k. yang, b. yuan, f. schreiber, l. wang, t. velkov, k.d. roberts, j. li. a novel chemical biology and computational approach to expedite the discovery of new generation polymyxins against life-threatening acinetobacter baumannii. chem. sci. 12 (2021) 12211. https://doi.org/10.1039/d1sc03460j. [21] m. vaara. polymyxins and their potential next generation as therapeutic antibiotics. front. microbiol. 10 (2019) 1689. https://doi.org/10.3389/fmicb.2019.01689. [22] z. wang, b. koirala, y. hernandez, m. zimmerman, s. park, d.s. perlin, s.f. brady. a naturally inspired antibiotic to target multidrug-resistant pathogens. nature 601 (2022) 606-611. https://doi.org/10.1038/s41586-021-04264-x. [23] d.m. shlaes, j.h. shlaes, j. davies, r. williamson. escherichia coli susceptible to glycopeptide antibiotics. antimicro. ag. chemother. 33 (1989) 192-197. https://doi.org/10.1128/aac.33.2.192. [24] n.c. gordon, k. png, d.w. wareham. potent synergy and sustained bactericidal activity of a vancomycin-colistin combination versus multidrug-resistant strains of acinetobacter baumannii. antimicro. ag. chemother. 54 (2010) 5316-5322. https://doi.org/10.1128/aac.00922-10. [25] e. van groesen, c.j. slingerland, p. innocenti, m. mihajlovic, r. masereeuw, n.i. martin. vancomyxins: vancomycin-polymyxin nonapeptide conjugates that retain anti-gram-positive activity with enhanced potency against gram-negative strains. acs infect. dis. 7 (2021) 2746-2754. https://doi.org/10.1021/acsinfecdis.1c00318. [26] r.d. hotchkiss, r.j. dubos. fractionation of the bacteriocidal agent from cultures of a soil bacillus. j. biol. chem. 132 (1940) 791-792. [27] a.r. battersby, l.c. craig. the chemistry of tyrocidine. i. isolation and characterization of a single peptide. j. chem. soc. 74 (1952) 4019-4023. https://doi.org/10.1021/ja01136a014. [28] g.f. gause. gramicidin s and its use in the treatment of infected wounds. nature 154 (1944) 703. https://doi.org/10.1038/154703a0. https://doi.org/10.1038/ja.2008.84 https://doi.org/10.1002/anie.201304691 https://doi.org/10.1021/ja506472u https://doi.org/10.1128/msphere.00156-20 https://doi.org/10.1021/acs.joc.6b01272 https://doi.org/10.1021/acsinfecdis.0c00031 https://doi.org/10.1016/s1473-3099(15)00424-7 https://doi.org/10.1016/j.jgar.2021.07.021 https://doi.org/10.1016/j.jgar.2021.07.021 https://doi.org/10.1039/d1sc03460j https://doi.org/10.3389/fmicb.2019.01689 https://doi.org/10.1038/s41586-021-04264-x https://doi.org/10.1038/s41586-021-04264-x https://doi.org/10.1128/aac.33.2.192 https://doi.org/10.1128/aac.00922-10 https://doi.org/10.1021/acsinfecdis.1c00318 https://doi.org/10.1021/ja01136a014 https://doi.org/10.1038/154703a0 admet & dmpk 10(2) (2022) 131-146 chemical syntheses and resistant antibiotics doi: http://dx.doi.org/10.5599/admet.1272 145 [29] g.f. gause, m.g. brazhnikova. gramicidn s origin and mode of action. lancet 244 (1944) 715-716. https://doi.org/10.1016/s0140-6736(00)88377-4. [30] a.n. belozersky, t.s. passhina. chemistry of gramacidin s. lancet 244 (1944) 716-717. https://doi.org/10.1016/s0140-6736(00)88378-6. [31] p.g. sergiev. clinical use of gramicidin s. lancet 244 (1944) 717-718. https://doi.org/10.1016/s01406736(00)88379-8. [32] g.f. gause. gramacidin s review of recent work. lancet 248 (1946) 46-47. https://doi.org/10.1016/s0140-6736(46)90004-9. [33] b.p. lazzaro, m. zasloff, j. rolff. antimicrobial peptides: application informed by evolution. science 368 (2020) eaau5480. https://doi.org/10.1126/science.aau5480. [34] j. zerweck, e. strandberg, o. kukharenko, r. j., j. bürck, p. wadhwani, a.s. ulrich. molecular mechanism of synergy between the antimicrobial peptides pgla and magainin 2. sci. rep. 7 (2017) 13153. https://doi.org/10.1038/s41598-017-12599-7. [35] n.h.z. baharin, n.f.k. mokhtar, m. nasir, m. desa, b. gopalsamy, n.n.m. zaki, m.h. yuswan, a.r. muthanna, n.d. dzaraly, s. abbasiliasi, a.m. hashim, m.s.a. sani, s. mustafa. the characteristics and roles of antimicrobial peptides as potential treatment for antibiotic-resistant pathogens: a review. peerj 9 (2021) e12193. https://doi.org/10.7717/peerj.12193. [36] t. vanzolini, m. bruschi, a.c. rinaldi, m. magnani, a. fraternale. multitalented synthetic antimicrobial peptides and their antibacterial, antifungal and antiviral mechanisms. int. j. mol. sci. 23 (2022) 545. https://doi.org/10.3390/ijms23010545. [37] k. klobucar, e.d. brown. new potentiators of ineffective antibiotics: targeting the gram-negative outer membrane to overcome intrinsic resistance. curr. opin. chem. biol. 66 (2022) 102099. https://doi.org/10.1016/j.cbpa.2021.102099. [38] b. gao, s. chen, y.n. hou, y.j. zhao, t. ye, z. xu. solution-phase total synthesis of teixobactin. org. biomol. chem. 17 (2019) 1141-1153. https://doi.org/10.1039/c8ob02803f. [39] v. ng, s.a. kuehne, w.c. chan. rational design and synthesis of modified teixobactin analogues: in vitro antibacterial activity against staphylococcus aureus, propionibacterium acnes and pseudomonas aeruginosa. chem. eur. j. 24 (2018) 9136-9147. https://doi.org/10.1002/chem.201801423. [40] z. charlop-powers, a. milshteyn, s.f. brady. metagenomic small molecule discovery methods. curr. opin. microbiol. 19 (2014) 70-75. https://doi.org/10.1016/j.mib.2014.05.021. [41] z. charlop-powers, s.f. brady. phylogeo: an r package for geographic analysis and visualization of microbiome data. bioinformatics 31 (2015) 2909–2911. https://doi.org/10.1093/bioinformatics/btv269. [42] z. charlop-powers, c.c. pregitzer, c. lemetre, m.a. ternei, j. maniko, b.m. hover, p.y. calle, k.l. mcguire, j. garbarino, h.m. forgione, s. charlop-powers, s.f. brady. urban park soil microbiomes are a rich reservoir of natural product biosynthetic diversity. proc. natl. acad. sci., u.s.a. 113 (2016) 14811-14816. https://doi.org/10.1073/pnas.1615581113. [43] x. vila-farres, j. chu, m.a. ternei, c. lemetre, s. park, d.s. perlin, s.f. brady. an optimized syntheticbioinformatic natural product antibiotic sterilizes multidrug-resistant acinetobacter baumanniiinfected wounds. msphere 3 (2018) 3:e00528-00517. https://doi.org/10.1128/msphere.00528-17. [44] j. chu, b. koirala, n. forelli, x. vila-farres, m.a. ternei, t. ali, d.a. colosimo, s.f. brady. syntheticbioinformatic natural product antibiotics with diverse modes of action. j. am. chem. soc. 142 (2020) 14158-14168. https://doi.org/10.1021/jacs.0c04376. [45] b.m. hover, s.-h. kim, m. katz, z. charlop-powers, j.g. owen, m.a. ternei, j. maniko, a. estrela, b., h. molina, s. park, d.s. perlin, s.f. brady. culture-independent discovery of the malacidins as calciumdependent antibiotics with activity against multidrug-resistant gram-positive pathogens. nat. microbiol. 3 (2018) 415-422. https://doi.org/10.1038/s41564-018-0110-1. [46] z. sun, z. shang, n. forelli, k.h.l. po, s. chen, s.f. brady, x. li. total synthesis of malacidin a by betahydroxyaspartic acid ligation-mediated cyclization and absolute structure establishment. angew. chem. int. ed. 59 (2020) 19868–19872. https://doi.org/10.1002/anie.202009092. http://dx.doi.org/10.5599/admet.1272 https://doi.org/10.1016/s0140-6736(00)88377-4 https://doi.org/10.1016/s0140-6736(00)88378-6 https://doi.org/10.1016/s0140-6736(00)88379-8 https://doi.org/10.1016/s0140-6736(00)88379-8 https://doi.org/10.1016/s0140-6736(46)90004-9 https://doi.org/10.1016/s0140-6736(46)90004-9 https://doi.org/10.1126/science.aau5480 https://doi.org/10.1038/s41598-017-12599-7 https://doi.org/10.7717/peerj.12193 https://doi.org/10.3390/ijms23010545 https://doi.org/10.1016/j.cbpa.2021.102099 https://doi.org/10.1039/c8ob02803f https://doi.org/10.1002/chem.201801423 https://doi.org/10.1016/j.mib.2014.05.021 https://doi.org/10.1093/bioinformatics/btv269 https://doi.org/10.1093/bioinformatics/btv269 https://doi.org/10.1073/pnas.1615581113 https://doi.org/10.1128/msphere.00528-17 https://doi.org/10.1021/jacs.0c04376 https://doi.org/10.1038/s41564-018-0110-1 https://doi.org/10.1002/anie.202009092 david j. newman admet & dmpk 10(2) (2022) 131-146 146 [47] n. kovalenko, g.k. howard, j.a. swain, y. hermant, a.j. cameron, g.m. cook, s.a. ferguson, l.a. stubbing, p.w.a. harris, m.a. brimble. a concise synthetic strategy towards the novel calciumdependent lipopeptide antibiotic, malacidin a and analogues. front. chem. 9 (2021) 687875. https://doi.org/10.3389/fchem.2021.687875. [48] g. upert, a. luther, d. obrecht, p. ermert. emerging peptide antibiotics with therapeutic potential. medicine drug discov. 9 (2021) 100078. https://doi.org/10.1016/j.medidd.2020.100078. [49] y. liu, j. shi, z. tong, y. jia, b. yang, z. wang. the revitalization of antimicrobial peptides in the resistance era. pharmacol. res. 163 (2021) 105276. https://doi.org/10.1016/j.phrs.2020.105276. [50] y. zhu, w. hao, x. wang, j. ouyang, x. deng, h. yu, y. wang. antimicrobial peptides, conventional antibiotics, and their synergistic utility for the treatment of drug‐resistant infections. med. res. rev. (2022). https://doi.org/10.1002/med.21879. [51] k.c. nicolaou, s. rigol. perspectives from nearly five decades of total synthesis of natural products and their analogues for biology and medicine. nat. prod. rep. 37 (2020) 1404-1435. https://doi.org/10.1039/d0np00003e. [52] p.a. wender, c.t. hardman, s. ho, m.s. jeffreys, j.k. maclaren, r.v. quiroz, s.m. ryckbosch, a.j. shimizu, j.l. sloane, m.c. stevens. scalable synthesis of bryostatin 1 and analogs, adjuvant leads against latent hiv. science 358 (2017) 218-223. https://doi.org/10.1126/science.aan7969. ©2022 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) https://doi.org/10.3389/fchem.2021.687875 https://doi.org/10.1016/j.medidd.2020.100078 https://doi.org/10.1016/j.phrs.2020.105276 https://doi.org/10.1002/med.21879 https://doi.org/10.1039/d0np00003e https://doi.org/10.1126/science.aan7969 http://creativecommons.org/licenses/by/3.0/ manuscript doi: 10.5599/admet.3.1.169 51 admet & dmpk 3(1) (2015) 51-67; doi: 10.5599/admet.3.1.169 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper adaptation of a human gut epithelial model in relation to the assessment of clinical pharmacokinetic parameters for selected tyrosine kinase inhibitors richard j honeywell, christien fatmawati, marita buddha; sarina hitzerd, ietje kathman and godefridus j. peters* dept. of medical oncology, vu university medical center, po box 7057, 1007 mb amsterdam, the netherlands * corresponding author, e-mail: gj.peters@vumc.nl, tel: +3120 444 2633, fax: +3120 444 3844, received: february 19, 2015; revised: march 24, 2015; published: march 31, 2015 abstract the absorption, efflux and transport properties of two of the most commonly used tyrosine kinase inhibitors (tkis), erlotinib (e) and gefitinib (g) were investigated using an adapted workable methodology of a 3-day caco-2 cell monolayer transwell system, a standard model to test drug permeability and uptake of orally administered compounds. monolayer integrity was tested using trans-epithelial electrical resistance (teer) measurements, while drug concentrations were determined with a validated lc-ms/ ms technique. addition of 5 % bovine serum albumin (bsa) maintained drug concentrations at ± 20 µm through the avoidance of chelate formation, (nevertheless, a reduced accumulative mass transport of the protein bound drug was observed). investigation with ko143 (a specific blocker of abcg2) or nan3 (a metabolic inhibitor) indicated an interplay between active transport and to a less degree passive diffusion for gefitinib. however, for erlotinib results indicate a more dominant passive diffusion supported by one or more active transport mechanisms. the use of ko143 suggests that abcg2 is partially involved with accumulation of both erlotinib and gefitinib in the intestinal cell. this adapted methodology is well suited for absorption, efflux and transport studies and may be extended to investigate the dominant mechanism involved in the transport of tkis. keywords tyrosine kinase inhibitors; caco-2, transwell; gut epithelial; gut model; drug absorption. introduction in recent years, approaches to treatment of various different oncological disorders have been directed towards the application of a series of small molecules referred to as tyrosine kinase inhibitors (tkis). this loosely related family of molecules specifically target the tyrosine kinase domain of growth factor receptors in the cellular membrane or cytosol which play an important role in the control of cell growth and replication [1-4]. two of the most commonly used tkis are erlotinib (non-small cell lung cancer (nsclc), pancreatic cancer (pc)) and gefitinib (nsclc) [5]. erlotinib and gefitinib both show similarity in their basic structure with a central pyrimidine core, an adenosine triphosphate (atp) http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:gj.peters@vumc.nl peters et al. admet & dmpk 3(1) (2015) 51-67 52 binding side chain and a basic side chain (figure 1). both gefitinib and erlotinib specifically target epidermal growth factor receptor (egfr) [6]. figure 1. the physical and chemical properties for gefitinib (ge) and erlotinib (er). log p is estimated by viswanadhan’s fragmentation [7]. tkis represent an orally self-administered treatment that can be handled on an outpatient basis. oral administration has a lot of beneficial aspects such as ease of use and suitability for long term treatment. however, resistance to these molecules can develop over time. this may be mediated by one of the multidrug resistance (mdr) proteins which form a group of active efflux pumps of the atp binding-cassette (abc) drug transporters; the pumps may catalyze the efflux of tkis out of the enterocyte cells [8-10]. there are seven subfamilies of abc efflux transporter (abca – abcg) based on their domain structure [11]. drug interaction with the efflux transporter as a substrate or an inhibitor depends on the concentration of drugs, for example abcg2 may transport substrates like erlotinib and gefitinib at low concentrations (0.1 – 1 μm) [11], but shows inhibitory reaction with these drugs at higher concentrations [12-15]. the abcg2 might transport substrates in the same way as abcb1 that expels the drugs out of the cell or only translocate them inside of the cell [15], for which both conditions are lessening the effectiveness of treatment. in addition to mdr, barriers to oral chemotherapy are the intestinal drug permeability uptake through epithelial cells of the inner intestinal wall, metabolism in the liver and enzymatic degradation in the intestinal tract [16]. the overall prediction of efficacy of an oral based treatment is possible by using an appropriate model system in vitro [17]. an ideal model system for intestinal absorption is based on the growth of a cellular monolayer and the careful comparison of the drug transport across this barrier. this system is referred to as a transwell system and consists of a donor, receiver and cellular compartment. figure 2 illustrates the schematic diagram of a compartmental model and working model of a transwell system. transport processes can occur with both a unidirectional and/or a bi-directional mechanism of either active transport by a receptor based protein system or by passive diffusion along a concentration gradient [18]. active transport requires energy (atp dependent) and can be against a concentration gradient, it plays a major role in the uptake of hydrophilic compounds into the cell and in the efflux of drug out of the cell. passive diffusion from higher to lower concentration is atp independent and occurs through lateral diffusion, transcellular or paracellular pathway [18]. admet & dmpk 3(1) (2015) 51-67 adaptation of human gut epithelial model doi: 10.5599/admet.3.1.169 53 figure 2. schematic diagram of caco-2 cell monolayer and drug transport processes in the transwell system. apical side represents the luminal side of the enterocyte and basolateral side represents the blood circulation. the dashed arrow describes drug uptake from the apical to the basolateral side where the apical side serves as donor compartment and basolateral side serves as receiver compartment. single headed arrows in the transport processes between compartments describe unidirectional processes while the double-headed arrows describe bidirectional processes. drug transport between cellular compartments involves bidirectional processes. a widely used model system consisting of a caco-2 cell monolayer, which is a polarized cell line derived from human epithelial colorectal adenocarcinoma cells [19]. when cultured as a monolayer (in vitro), this cell line demonstrates the same phenotype, morphology and functions as enterocytes of the small intestine [20-23]. since this cell line expresses the same enzymes and transport proteins that mediate drug uptake or efflux in the intestine, it is possible to determine significant correlation between permeability of compounds under in vivo conditions [23]. the transwell system has become the standard model to test the drug permeability and uptake of orally administered compounds [17]. caco-2 cell monolayers enable investigation into the absorption of drugs, the passive and the active transport systems of its transport processes across a barrier directly comparable to the human system. a disadvantage of this model system is the standard 21 day protocol for the formation of the caco2 monolayer [23]. an alternative 3-day culture system offers a more convenient and productive alternative that has been demonstrated to provide comparable results to the standard system [24]. we optimized this model to investigate the absorption, efflux and transport properties of two tkis, namely erlotinib and gefitinib, utilizing lc-msms sensitivity to reliably measure the individual compound levels [25]. experimental materials erlotinib and gefitinib were purchased from lc laboratories (woburn, ma, usa). all general reagents were purchased from sigma (zwijndrecht, the netherlands). ko143 was a kind gift of professor gj koomen, university of amsterdam, the netherlands. fetal bovine serum (fbs) was purchased from paa laboratories gmbh (pasching, austria) while the bovine serum albumin (bsa) fraction v was purchased from roche diagnostics (mannheim, germany). roswell park memorial institute (rpmi) 640 medium, trypsin-­edta, penicillin, streptomycin (10000 u/ml) and 1 m hepes buffer (in 0.85 % nacl) were purchased from lonza benelux bv (breda, nl). hank’s balanced salt peters et al. admet & dmpk 3(1) (2015) 51-67 54 solution (hbss) containing cacl2 and mgcl2 as the transport medium for drug transporting cross cell monolayer was purchased from invitrogen (breda, the netherlands). the growth medium (dulbecco’s modified eagle medium (dmem) based), differentiation medium (a serum free medium containing butyric acid) to induce fully differentiation to enterocytes and supplemental medium to maintain the cells under serum free condition specific for caco-2 cell monolayer were purchased from bd biocoat™ (breda, the netherlands). high purity erlotinib and gefitinib was obtained from lc-laboratories (massachusetts, usa). analytical grade solvents like acetonitrile, methanol and isopropanol were purchased from biosolve bv (valkenwaard, nl). bio­rad protein assay was purchased from bio-­rad laboratories gmbh (münchen, de). milliq water was supplied via a milliq water purification system (millipore, nl). all other reagents were of an analytical grade unless stated and sourced locally. equipment the biocoat® hts caco­2 assay system and bd falcontm 24-well multiwell plates were purchased from becton dickinson bv (breda, nl). breathe-easier microplate sealing film was purchased from diversified biotech bv (ulvenhout, nl). the microplate reader was provided by tecan benelux bvba (giessen, nl) and spectra fluor software (xfluor4 version v 4.50) was used. the trans epithelial electrical resistance (teer) meter (millicell® – ers) was provided by millipore (amsterdam, nl). the liquid chromatography coupled to mass spectrometry (lc-ms/ms) analyses were performed using a dionex ultimate 3000 system coupled with an api 3000 mass spectrometer. for this system the following software was used; analyst version 1.5.2 from applied biosciences, in combination with dionex, chromeleon lc modules version 6.8, controlled by dionex mass link (dms) version 2.10 . cell culture the transwell system was developed using the wild type caco-2 cell line (passage 15 – 25) after defrost. cells were cultured routinely in dmem containing 4.5 g/ liter glucose and l-glutamine supplemented with 10 % fbs and 20 mm of hepes at standard conditions of 37 °c, 5 % co2 and 100 % humidity. confluent cells were detached using trypsin edta and past twice weekly. cells were seeded at a density 6.6 x 10 5 cells/ cm 2 on a biocoat 24 well transwell plate (1 µm pore size, 0.31 cm 2 surface area) pre-wetted with 50 µl of growth medium for 5 – 10 minutes prior to seeding. plates were incubated for 20 – 24 hours with growth medium and then for 44 – 48 hours with differentiation medium at standard conditions. both media were enriched with the supplemental medium (1:1000) and with 1 % penicillin/ streptomycin while maintaining conditions for growth and differentiation as specified by the supplier’s protocol. in addition, the plate was covered with breathe-easier cell culture foil during incubation period to maintain identical environmental conditions for each well. monolayer integrity – transepithelial electrical resistance (teer) the development of a good integrity monolayer on semi-permeable filters in the transwell system is the initial crucial step prior to any drug uptake and/or transport investigation. transepithelial electrical resistance (teer) measures the ion permeability through the paracellular pathways and is used to determine the “intactness” of each grown monolayer. the benefit of teer is the speed of measurements and the accuracy with which the integrity can be measured. to this end transport medium was prepared on the day of treatment by buffering hbss; ph 7.4 with 25 mm hepes and 0.35 g/ litre nahco3 [23] then adjusted to ph 7.4 with naoh. within the transwell plate the differentiation medium was replaced with prepared transport medium (apical 300 µl and basolateral 1 ml) and incubated for 15 minutes with gentle agitation (100 rpm) under admet & dmpk 3(1) (2015) 51-67 adaptation of human gut epithelial model doi: 10.5599/admet.3.1.169 55 standard conditions. teer was determined using the potential difference between two electrodes suspended across the monolayer. measurement between each well was performed after a short washing period of the electrodes in ethanol then transport medium. the resistance measured for each monolayer before and after the experiment was adjusted by a blank resistance determined from the wells without a monolayer and multiplied by 0.31 cm 2 (the area of effective membrane diameter). a cut off value of 600 – 1600 ω cm 2 was used for determining monolayer integrity according to specifications obtained from bd biocoat™. transport studies transport studies of 20 µm erlotinib and gefitinib were performed in the direction apical to basolateral (a b) and in the direction basolateral to apical (b a). inhibition of cellular pump function was investigated with either 200 nm ko143 or 1 mm nan3. wells were pre-incubated under standard conditions for 20 minutes with 400 nm ko143 or 1 hour with 3 mm nan3; gentle agitation (100 rpm) was applied during the incubation. all drugs were dissolved in transport medium either containing 5 % bsa or containing 5 % bsa + 1 mm nan3 and added to the donor compartment either apical (300 µl) or basolateral (1 ml). the initial concentration of the drug was verified from a 20 µl sample taken immediately from the donor compartment after drug administration (t = 0). subsequently, samples (50 µl) were taken from each receiver compartment at the time points 15, 30, 60, 90, 120 and 180 minutes after drug administration. following each sampling, the volume of sample taken was replaced by the same amount of pre-warmed transport medium. the plate was incubated with gentle agitation (100 rpm) at standard conditions between each time point. a final sample of 20 µl was taken from the donor compartment after 180 minutes. all samples were placed in pre-labelled tubes, stored on ice temporarily during the experiment, snap frozen in liquid nitrogen and stored at -80°c until required for analysis. after the final sample time point, cells were washed, trypsinized and harvested; the pellets were stored at -80 °c until analysis. liquid chromatography analysis extractions for chromatographic analysis of standard and samples were all performed on ice as detailed. analysis of samples taken from the donor and receiver compartments as well as the prepared cell pellet was performed by lc-ms/ ms techniques. a simple extraction procedure of protein precipitation with acetonitrile was performed for each sample and standard preparation as reported previously [25]. lc-ms/ ms analysis was performed with a mobile phase consisting of acetonitrile, ammonium acetate (20 mm, ph 7.8) and methanol in the ratio of 66.1:24.5:8.3 % (v/v) with 1 % isopropyl alcohol added as a chromatographic modifier. chromatographic separation was obtained with a phenomenex prodigy ods3 column, 3 µm particle sizes, 100 x 2.00 mm (phenomenex, the netherlands) at a flow rate of 0.2 ml/ minute. all mobile phases were filtered through a 0.2 µm sartorius membrane filter and degassed for 5 minutes under vacuum with sonication. calculation and statistics the permeability coefficient (papp) represents a measure for the efficiency of transport and was calculated using the total drug concentration per sample well with the following equation: ac v x t q p app . 0    peters et al. admet & dmpk 3(1) (2015) 51-67 56 where ∆q/ ∆t = the rate of increase in drug concentration (accumulative mass transport) in the receiver compartment over time (µm/ second), v = volume in the receiver compartment (ml), c0 = the initial concentration of drug in the donor compartment (µm), and a = the membrane surface area (cm 2 ). the efflux ratio was calculated according to the equation: efflux ratio bap abp app app    efflux ratios > 1 indicate that drug efflux occurred during the drug transport experiment. the papp ratios are shown as a mean value of three or more measurements ± sem. statistical significance was determined using a simple student’s t-test where a p value of < 0.05 was considered significant. results and discussion methodology development drug concentration to establish a baseline for the proposed transwell transport studies a plate with no prepared monolayer was used as a control system. either erlotinib or gefitinib was placed in the donor compartment dissolved in medium. however, control samples taken from the donor compartment medium revealed a curious decrease in concentration over time (table 1). a drug concentration of 20 µm had been added to the donor compartment, but the actual concentration measured after the 3 hr experiment was repeatedly lower than expected, gefitinib by 1.5 fold and erlotinib by 2 to 3 fold. the observed decrease was not related to the added stock solution since analysis of this gave the correct concentration and the amount transported to the receiver compartment was negligible compared to the change observed. tkis of this type are extremely insoluble in aqueous solutions (hbss buffer in this instance) and are known systemically to be 90 – 100 % bound to plasma protein [6]. therefore it was investigated whether the drug solubility was the issue behind the decreasing drug concentrations. analysis of drug concentration in the prepared medium immediately after dilution showed concentrations in the 20-21 µm range. table 1. apical concentrations of erlotinib and gefitinib (without bsa) after 0 and 3 h’s in medium under standard conditions of 37 °c, 5 % co2 and 100 % humidity (n=4). donor concentration (µm) apical 0 hour (µm ± sem) apical 3 hour (µm ± sem) erlotinib 20 10.1 ± 0.2 7.1 ± 0.03 gefitinib 20 19.2 ± 0.3 14.8 ± 0.7 however, the same solution demonstrated a similar decrease in concentration after 24 hours as compared to the 3 hours post experimental sample. the lowering of the tkis’ concentration was investigated by lc-ms/ ms and complexes were observed that could be linked to the chelation of the tkis to the metal ions in the hbss buffer, mg 2+ and ca 2+ (data not shown). it was determined that admet & dmpk 3(1) (2015) 51-67 adaptation of human gut epithelial model doi: 10.5599/admet.3.1.169 57 these compounds will form chelation complexes in the ratios 2:3 for gefitinib and 1:2 for erlotinib, these complexes were not broken up during sample preparation and hence the relative concentration of drug appears to decrease over time. to prevent chelation, bovine serum albumin (bsa) was added to the transport medium at a concentration of 5 % v/v, subsequently measured concentrations after 3 hours at 37 °c matched the expected concentration of 20 µm for both erlotinib and gefitinib (figure 3). in addition, it was also determined that the addition of nan3 or ko143 to the transport medium did not significantly affect the initial and post donor concentrations of either erlotinib or gefitinib when 5 % bsa was added. c o n c e n tr a ti o n ( µ m ) w it h o u t b s a w it h b s a w it h b s a +k o 14 3 w it h b s a +n an 3 0 10 20 30 erlotinib gefitinib figure 3. comparison of the concentration determined for prepared 20 µm solutions of erlotinib or gefitinib in hbss buffer after 3 hours at 37 °c with and without 5 % bovine serum album – (bsa). addition of bsa prevented chelation of both erlotinib and gefitinib during the course of the 3 hour experimental procedure which was not affected by the addition of ko143 or sodium azide. (µm ± sem of n=16). using these conditions the transport over the blank transwell system (no-monolayer) was determined. samples were taken from the receiver compartment on the same time schedule as the proposed experimental procedure. figure 4 shows the graphical representation of the transport characteristics of gefitinib under these conditions. linear transport characteristics are observed with both conditions. post experiment each donor well was drained and washed three times with fresh medium. using a 1 % dmso solution of ethanol each used compartment was washed and the effluent collected. analysis of the effluent revealed no evidence gefitinib remaining on the compartment surfaces (data not shown). peters et al. admet & dmpk 3(1) (2015) 51-67 58 transwell monolayer reproducibility cells were prepared as a monolayer using a 3-day culture protocol by yamashita et al [24] supplied as a commercial kit by becton dickinson bv (the netherlands). monolayer integrity was determined by teer measurements prior to and after 3 hours post addition of the drugs. values such as 260 ± 65 ω cm 2 [23] and 300 – 600 ω cm 2 have been quoted in literature as being a specification cut-off. [26] however, under 300 ω cm 2 we observed transport characteristics that were similar to a well with no monolayer while above 1600 ω cm 2 transport characteristics were highly different in regards to mass transported. hence, the specifications for monolayer integrity was set in the range 600 – 1600 ω cm 2 , this was in agreement with the technical information as supplied by becton dickinson. initial teer values revealed a problem to the formation of consistent monolayers linked to the well position within each 24 well plate. wells closer to the sides of the plate often had teer values indicating an incomplete monolayer, whereas more centrally placed wells were within the approved specification, indicating a complete monolayer. it was observed that caco-2 monolayers were highly sensitive to minor fluctuations in temperature, humidity and co2 content; these fluctuations affected individual wells depending on the plate orientation and position within the incubator. to regulate the environment within each well a breathable membrane seal was applied, this was sufficient for reproducible monolayers across the entire plate to be formed within the 3-day growth period. an additional problem was observed with the teer measurements post experiment, here teer values of 300 ω cm 2 or lower were observed indicating loss of monolayer integrity at an unknown stage during the experimental procedure. post experiment each well was carefully washed free of drug containing medium with the intention of recovering as many cells as possible for accumulation analysis. additionally this washing step was included since it was unknown whether the added drugs to the medium would affect the teer measurements. however, this washing step disrupted the monolayer significantly giving the “out of specification” teer values. avoiding the final washing stage teer values demonstrated good integrity over the course of the 3 hour experiment for all the wells tested both with and without the addition of the drugs under investigation (table 2) recovered cells were subsequently washed during the recovery process. table 2. good monolayer integrity was indicated by teer value in the range of 600 – 1600 ω cm 2 . each box represents one well in the transwell plate. the grey boxes indicate the wells without monolayer (blank). the teer measurements were performed before and after the experiment of transport studies. 1 2 3 4 5 6 a 1016.8 1674.0 1246.2 737.8 62.0 62.0 b 1109.8 905.2 1271.0 713.0 651.0 688.2 c 883.5 1153.2 1023.0 1147.0 744.0 775.0 d 1240.0 1550.0 1209.0 1581.0 1612.0 781.2 admet & dmpk 3(1) (2015) 51-67 adaptation of human gut epithelial model doi: 10.5599/admet.3.1.169 59 referenced literature recommended 500 µl as the sampling volume from the receiver compartment, replacing volume taken at each time point with fresh drug free medium at the correct temperature [23]. in initial plates post experiment teer values revealed that in many cases monolayer integrity was compromised with values both lower and higher than specification. subsequent analysis of the receiver compartment samples showed abrupt changes in mass transport, increasing for teer values under specification but decreasing for teer values above specification. these abrupt changes were not observed in wells with consistent teer values prior to and post drug addition. the problem was associated with the mixing of the media in the receiver compartment after sampling. it was concluded that hydrostatic pressure from pipette aspiration could either induce stress on the monolayer causing a break in integrity or introduce air bubbles from poor technique into the pores of the filter under the monolayer. this isolated the monolayer from the receiver compartment, thereby decreasing mass transport noticeably. a broken monolayer exhibited a decrease in its teer value, but air bubbles increase the electrical resistance across the monolayer artificially elevating the subsequent teer values. the 500 µl of sample was initially taken to provide sufficient material for extraction and analysis of each drug but caused serious problems with the monolayer. therefore, analytical procedures were developed to reduce the amount of sample required for analysis down to 20 µl while maintaining analytical sensitivity. hence it became possible to reduce the sample volume taken on each time point to 50 µl, this provided sufficient sample for analysis and additional volume for unforeseen analytical problems. the reduced sample volume decreased the chances that air could get under the well insert and decreased the hydrostatic pressure seen with the 50 % liquid replacement technique. verification of transport with gefitinib and erlotinib the transport of gefitinib across the transwell system without a monolayer established a baseline for passive diffusion across the experimental setup. the subsequent sampling procedure dictated the calculation methodology; each 15 minute sample dilutes the well concentration by a factor of 0.02. therefore, absolute mass is determined at each time point, adjusted for sample dilution and used as the baseline value for each subsequent increase in mass to the next time point. addition of absolute mass transport for each following time point gave the total mass transported. in this way absolute mass transported can be determined for all the time points and the papp calculated. transport decreased significantly when an intact monolayer was used consistent with expectations (figure 4). this could be explained when taking into account the chelation complex being formed in both the donor and receiver compartments, reducing the total drug being measured by the highly specific lcmsms technique. comparison of the accumulated mass transport for gefitinib with and without the addition of bsa also demonstrated a clear trend (figure 4). with the addition of bsa a decrease in the relative amounts transported across the monolayer was observed (693.8 pmol vs 473.9 pmol, t=3 hr). this could be explained by the high protein binding properties of gefitinib which would decrease the availability of the drug for transport across the membrane. all remaining experiments were performed with medium containing 5 % bsa in both the apical and basolateral compartments. peters et al. admet & dmpk 3(1) (2015) 51-67 60 experiment time (min) t o ta l m a s s ( p m o l) t ra n s p o rt e d 0 50 100 150 200 0 500 1000 1500 2000 without monolayer with monolayer with monolayer + 5% bsa figure 4. graphical representation of transport characteristics of gefitinib (n=4) with and without a monolayer across the transwell setup. where the medium (hbss buffer) did not include bsa the total mass transported would have been affected by chelation of gefitinib with either calcium or magnesium reducing the measurable drug concentrations significantly. addition of a monolayer to the system slows transport hence total accumulated mass while further addition of 5 % bsa also would slow transport lowering the total accumulated mass as well. model pharmacokinetic absorption apical to basolateral (a b) accumulative mass transport of gefitinib in the apical to basolateral direction (figure 5) showed a clear linear increase over the time period measured (papp of 0.38 ± 0.05 µm/ s). when sodium azide (nan3) was added to the system all atp dependent transport processes would have been blocked. in this situation gefitinib showed decrease in the observed mass transported (papp of 0.32 ± 0.018 µm/s) while an even greater decrease was observed when using the specific abg2 blocker ko143 (papp of 0.26 ± 0.013). passive transport mechanisms would not be affected by nan3 but all active mechanisms would, similarly ko143 would not inhibit passive diffusion but all mechanisms involving the abcg2 transporter would be inhibited. therefore it can be concluded from this evidence that gefitinib demonstrated evidence of a partial role for active transport mechanisms that occur alongside more passive mechanisms. admet & dmpk 3(1) (2015) 51-67 adaptation of human gut epithelial model doi: 10.5599/admet.3.1.169 61 0 50 100 150 200 0 200 400 600 800 apical to basolateral apical to basolateral + ko143 apical to basolateral + nan 3 t o ta l m a s s ( p m o l) t ra n s p o rt e d time (minutes) 0 50 100 150 200 0 200 400 600 800 basolateral to apical + nan 3 basolateral to apical + ko143 basolateral to apical figure 5. graphical representation of accumulated drug transport in the apical to basolateral (a) and the basolateral to apical (b) directions for gefitinib using hbss medium containing 5 % bsa. interference in total accumulated transport is demonstrated using ko143 (abcg2 inhibitor) and nan3 (inhibitor of atp processes). erlotinib (figure 6) demonstrated a similar linear increase over the time period to gefitinib but at 4 4.5 fold higher concentrations (papp of 1.72 ± 0.08 µm/ s). with the addition of nan3 erlotinib had an increase in mass transported (papp of 2.06 ± 0.15 µm/ s), however, this was not significantly different to the transport flow without nan3. the limited effect of nan3 on the transport of erlotinib across the membrane suggested that erlotinib transport is predominantly a passive system. however, with the addition of ko143 a decrease in mass transported was observed that was significantly different to the control condition (papp of 1.23 ± 0.06, p> 0.05). this suggested that an active mechanism was involved but was an elimination mechanism on the apical membrane only. table 3. summary of the papp values in both a-b and b-a directions for erlotinib and gefitinib, with and without the inclusion of the inhibitors ko143 and nan3.statistical comparisons were made using paired students t-test (#) comparing the papp a-b to the papp b-a. additional comparisons were made (*) between the control and either the addition of ko143 or nan3 gefitinib (n=9) papp (a-b) papp (b-a) efflux ratio control 0.38 ± 0.05 µm/s 0.39 ± 0.038 µm/s 1.03 # +ko143 0.26 ± 0.013 µm/s** 0.26 ± 0.012 µm/s** 1.00 # +nan3 0.32 ± 0.018 µm/s* 0.57 ± 0.14 µm/s** 1.78 ## erlotinib (n=12) papp (a-b) papp (b-a) efflux ratio control 1.72 ± 0.08 µm/s 4.51 ± 0.16 µm/s 2.62 ### +ko143 1.23 ± 0.06 µm/s** +nan3 2.06 ± 0.15 µm/s* where * or # not significant paired students t-test ** or ## significantly different (p>0.05) paired students t-test *** or ### highly significantly different (p>0.01) paired students t-test (a) (b) peters et al. admet & dmpk 3(1) (2015) 51-67 62 time (min) a c c u m u la ti v e m a s s t ra n s p o rt ( p m o l) 0 50 100 150 200 0 2000 4000 6000 apical to basolateral basolateral to apical apical to basolateral + ko143 apical to basolateral + nan3 figure 6. graphical representation of accumulated drug transport in the apical to basolateral and the basolateral to apical direction for erlotinib using hbss medium containing 5 % bsa. interference of the atp based active transport mechanisms is demonstrated using nan3 and ko143. basolateral to apical (b a; +5 % bsa) to correctly interpret the pharmacokinetic properties of these drugs as determined using the model in the apical to basolateral direction, transport from the blood or cellular situation back to the epithelial side of the membrane was also investigated. gefitinib demonstrated very similar flow characteristics in the b a (papp 0.39 ± 0.038 µm/ s) direction compared to a – b (figure 5). however, with the addition of nan3 the flow of gefitinib increased significantly (papp – 0.57 ± 0.14 µm/ s, p > 0.05), whereas with the addition of ko143 the flow decreased significantly (papp – 0.26 ± 0.012 µm/s, p>0.05). the net efflux ratio of gefitinib was determined to be less than 2 (1.0) which indicates a predominate flow in the basolateral to apical direction. with the introduction of nan3 the efflux ratio decreased (0.46) indicating a significant role of several active atp transport mechanisms on both the apical and basolateral membranes in the uptake of gefitinib. for erlotinib the b a transport was significantly greater in comparison to a b (papp – 4.51 ± 0.16 µm/ s vs papp – 1.72 ± 0.08 µm/ s, p > 0.009) indicating a very strong apical efflux flow as suggested by the data for the apical to basolateral flow. cellular accumulation of erlotinib and gefitinib analysis of cellular accumulation for each drug was performed in the cell pellet after completion of the transport studies experiment (table 4). the accumulation of gefitinib was consistent for both a-b and b-a transport mechanisms observed during the transwell investigation. for a-b transport gefitinib accumulation was 792.3 pmol/ mg protein which was lower than the accumulation seen for b-a transport (2059.0 pmol/ mg protein) indicating either a decreased uptake via the basolateral layer admet & dmpk 3(1) (2015) 51-67 adaptation of human gut epithelial model doi: 10.5599/admet.3.1.169 63 compared to the apical or a decreased efflux via the apical layer compared to the basolateral. addition of nan3 to the a-b donor compartment during transport slightly reduced cellular accumulation (22.9 %); while the addition of ko143 had a more significant lowering effect (45.5 %). for b-a transport addition of nan3 led to a similar decrease in accumulation (28 %) but ko143 had a significantly lower difference in accumulation (17.8 %). these results indicate that the measured papp for gefitinib transport is due to a combination of both passive and multiple active processes located on differing membranes. table 4. the cellular accumulation of gefitinib and erlotinib. concentration of drug was determined in the recovered cells of the monolayer at the end of the transwell experiment (180 min). gefitinib (pmol / mg protein) erlotinib (pmol / mg protein) a – b b – a a – b b a control (20 µm) 792.3 ± 178.8 2059 ± 256.5 741.1 ± 88.8 674.6 ± 71.6 20 µm + ko143 431.7 ± 165.7 1692 ± 242.6 451.6 ± 81.7 no data 20 µm + nan3 611.3 ± 106.2 1475 ± 156.9 835.8 ± 44.2 erlotinib demonstrated a similar accumulation in both a-b and b-a directions without addition of any inhibitors. with the addition of nan3 to the apical compartment during a-b transport an increase in cellular accumulation is seen (12.8 %), in contrast the addition of ko143 reduced the cellular accumulation similar to that of gefitinib (39.1 %). discussion in this paper we described the optimization and validation of a caco-2 gut epithelial model system in order to simulate uptake characteristics of the family of compounds classified under the name of tyrosine kinase inhibitors. this series of small molecule inhibitors have demonstrated strong in vitro chemotherapy potential but many exhibit limited to no clinical effect [27]. the phase 2 and phase 3 trial failure rates for these compounds is very high with regard to solid tumours. tkis are referred to as “targeted chemotherapy” drugs but have been shown to actually inhibiting a broad range of targets which can lead to toxicity and tumour resistance [27]. however, lack of target specificity does not completely explain the lack of clinical efficacy. hence we developed an adaptable model system to investigate the pharmacokinetic uptake of these molecules; we used the registered compounds gefitinib and erlotinib to validate the models applicability [28]. the first major point of the caco-2 model system is the time in which the monolayers need to be prepared, traditionally this has been a time consuming 21 days. to shorten this we utilized a commercial system from bd biosciences that required only a 3 day growth period. monolayers using this system have been shown to have characteristics identical to those grown over 21 days [24]. initial testing demonstrated several problems using this system for the investigation of gefitinib or erlotinib. it was observed that the donor compartment concentrations were not stable during the course of the experiment showing decreases of 60 % or more of the total drug added to the culture medium. peters et al. admet & dmpk 3(1) (2015) 51-67 64 investigation into this phenomenon indicated that this was due to the buffer matrix used (hbss),however, this buffer contained only ca 2+ and mg 2+ salts and no degradation of the compounds could be observed. mass spectral analysis of gefitinib and erlotinib incubated at 37°c in hbss buffer suggested that instead a chelation effect was the cause of the loss of concentration. yamashita et al reported that the relative importance of considering the “physiological conditions of the in vivo drug absorption when optimizing the in vitro experimental conditions”. the reference reports that the addition of bsa to the medium improved the transport of high lipophilic drugs with poor medium solubility across the caco-2 monolayer [29]. both gefitinib and erlotinib are poorly soluble in aqueous solutions and are very highly bound to proteins when circulating in the human system, hence by the simple addition of 5 % bsa we could more closely mimic the “real situation”. more importantly 5 % bsa in the medium stabilized the drugs while in solution by preventing “chelation” with the magnesium or calcium ions. this, in turn, yielded more reproducible and accurate permeability characteristics. it should be noted that the addition of transport inhibitors to this protocol did not affect the concentrations of compounds in the initial matrix or after the 3 hour experimental time. the next issue we observed was due to the sampling technique at individual time points whereby the physical removal of too much of the receiver compartments medium during the experiment led to problems with the monolayer where either a rapid increase or a sharp decline in accumulated transport could be observed on a random basis. the monolayer could not be observed physically but it was reasoned that the observed results were due to either stress breaks or “bubbling effects”. the bubbling effect was where air bubbles from sampling procedures had become trapped under the monolayer, isolating the donor compartment from the receiver compartment. this had the consequence of having very high variation between experimental duplicates and inconsistent results on an intra-day basis. lowering the volume sampled and very careful mixing of a compartment contents reduced this issue down to acceptable parameters. however, sample volume was reduced from 500 µl to 50 µl, requiring the adaptation of the lcms analytical techniques to be able to utilize only 10-20 µl of sample volume. for gefitinib and erlotinib this proved to be possible. with experimental parameters optimized a further problem was identified in that the monolayers did not grow consistently across the 24 well plates. wells in the centre portion of the plate were observed to reach starting experimental conditions much slower than those located on the outer rim of the plate. this created the condition where centrally placed monolayers were just within specifications while outer monolayers were at the maximum limit. this created significant variation in the observed transport characteristics. this issue was solved by isolating each well with a breathable membrane during the entire growth and experimental periods. the resulting protocol was subsequently used to assess transport characteristics of both gefitinib and erlotinib. initial investigations demonstrated significant differences in the transport characteristics between gefitinib and erlotinib. gefitinib demonstrated a 4 fold lower uptake compared to erlotinib when considering the transport in the apical to basolateral direction. however, in the reverse direction erlotinib demonstrated a significantly higher flow whereas for gefitinib it was similar to the apical to basolateral flow. the high flow of erlotinib indicated a potentially large negative flow from the system which would lower its overall bioavailability. to test the applicability of the model system further two active transport inhibitors were used in combination with gefitinib and erlotinib. active transport systems dependent on atp would have been blocked and any difference in transport behaviour would indicate an active transport mechanism for these compounds. the admet & dmpk 3(1) (2015) 51-67 adaptation of human gut epithelial model doi: 10.5599/admet.3.1.169 65 compound sodium azide inhibits all atp dependent processes [30]; therefore, if an atp dependent transporter is involved the total amount accumulated would alter depending on which membrane it was located on and the direction of flow. this model system demonstrated consistent differences between control and inhibited situations indicating that both multiple active and passive transport mechanisms are involved in the uptake of both these compounds. these results also indicate that the polarized membranes of caco-2 cells had different mechanisms for both uptake and efflux which were different for both compounds. the other inhibitor used was ko143, a more specific inhibitor of abcg2 only [31]. this inhibitor also showed differences for both drugs and between apical or basolateral membranes. for gefitinib, the presence of nan3 decreased the observed mass transport in the apical to basolateral direction but not completely. this indicated that gefitinib is apparently partially transported by both an active transport mechanism and partly by passive diffusion [7]. inhibition of abcg2 seemed to block the drug from accumulating within the cell suggesting that abcg2 might be one of several transporters involved. however, nan3 inhibition in the basolateral to apical direction increased gefitinib transported whereas ko143 decreased the total suggesting the role of another uncharacterized transporter in addition to abcg2. investigation into characteristics of erlotinib demonstrated a different apical to basolateral pattern. here nan3 increased slightly the amount transported while ko143 decreased the total similar to gefitinib in the basolateral to apical directions. however, the total erlotinib transported was significantly higher than gefitinib despite the same starting concentration being used. for erlotinib significantly lower drug accumulation in the cell was detected suggesting that equilibrium between donor and receiver compartments was the driving force of this mechanism with the cellular membrane having little resistance to passive diffusion; it should be considered whether the paracellular route is possibly the predominant mechanisn involved for this compound. theoretically passive diffusion results in the equilibrium of the compounds between donor, cellular and receiver compartments [32]. however, for erlotinib the most significant of the observed effects is that this concentration gradient is higher in the basolateral (blood) to apical (gut) direction. this could have significant consequences when considered clinically; these results suggest that high single dose schedules would have a better bioavailability compared to regular but lower dosing. in conclusion we validated a gut epithelial model system to study the potential gut uptake mechanisms for two widely used tkis, erlotinib and gefitinib. the caco-2 cell monolayer transwell system with a 3-day culture system proved to be very adaptable to study drug transport and will be very useful to investigating the role of drug transporters using a blocker like probenecid (blocker of abcc) and verapamil (blocker of abcb1), that would contribute important overview of drug transporters involved in transport of tkis. conclusions in conclusion we validated a gut epithelial model system to study the potential gut uptake mechanisms for two widely used tkis, erlotinib and gefitinib. the caco-2 cell monolayer transwell system with a 3-day culture system proved to be very adaptable to study drug transport and will be very useful to investigating the role of drug transporters using a blocker like probenecid (blocker of peters et al. admet & dmpk 3(1) (2015) 51-67 66 abcc) and verapamil (blocker of abcb1), that would contribute important overview of drug transporters involved in transport of tkis. references [1] e. k. rowinsky, annu. rev. med. 55 (2004) 433-457. [2] f. duffaud and c. a. le cesne, target oncol. 4 (2009) 45-56 [3] f. roszkiewicz, r. garidi, i. vaida, b. royer, a. parcelier, j. p. marolleau, and g. damaj, pharmacology 84 (2009) 38-41. [4] m. tolomeo, f. dieli, n. gebbia, and d. simoni, anticancer agents med. chem. 9 (2009) 853-863. [5] r. l. comis, oncologist. 10 (2005) 467-470. [6] n. p. van erp, h. gelderblom, and h. j. guchelaar, cancer treat. rev. 35 (2009) 692706. [7] m. galetti, r. r. alfieri, a. cavazzoni, m. s. la, m. bonelli, c. fumarola, p. mozzoni, p. g. de, r. andreoli, a. mutti, m. mor, m. tiseo, a. ardizzoni, and p. g. petronini, biochem. pharmacol. 80 (2010) 179-187. [8] p. borst and r. o. elferink, annu. rev. biochem. 71 (2002) 537-592. [9] c. lemos, g. jansen, and g. j. peters, br. j. cancer 98 (2008) 857-862. [10] m. m. gottesman, t. fojo, and s. e. bates, nat. rev. cancer 2 (2002) 48-58. [11] m. dean, a. rzhetsky, and r. allikmets, genome res. 11 (2001) 1156-1166. [12] j. li, g. cusatis, j. brahmer, a. sparreboom, r. w. robey, s. e. bates, m. hidalgo, and s. d. baker, cancer biol. ther. 6 (2007) 432-438. [13] z. shi, x. x. peng, i. w. kim, s. shukla, q. s. si, r. w. robey, s. e. bates, t. shen, c. r. ashby, jr., l. w. fu, s. v. ambudkar, and z. s. chen, cancer res. 67 (2007) 1101211020. [14] z. shi, s. parmar, x. x. peng, t. shen, r. w. robey, s. e. bates, l. w. fu, y. shao, y. m. chen, f. zang, and z. s. chen, oncol. rep. 21 (2009) 483-489. [15] c. f. higgins and m. m. gottesman, trends biochem. sci. 17 (1992) 18-21. [16] p. stenberg, u. norinder, k. luthman, and p. artursson, j. med. chem. 44 (2001) 1927-1937. [17] r. i. mahato, a. s. narang, l. thoma, and d. d. miller, crit rev. ther. drug carrier syst. 20 (2003) 153-214. [18] a. t. heikkinen, t. korjamo, and j. monkkonen, basic clin. pharmacol. toxicol. 106 (2010) 180-188. [19] j. fogh and g. trempe, new human tumor cell lines. in human tumor cells in vitro. j.fogh, editor., plenum press, new york, (1975) 115-159. [20] p. h. vachon and j. f. beaulieu, gastroenterology 103 (1992) 414-423. [21] p. shah, v. jogani, t. bagchi, and a. misra, biotechnol. prog. 22 (2006) 186-198. [22] i. j. hidalgo, t. j. raub, and r. t. borchardt, gastroenterology 96 (1989) 736-749. [23] i. hubatsch, e. g. ragnarsson, and p. artursson, nat. protoc. 2 (2007) 2111-2119. [24] s. yamashita, k. konishi, y. yamazaki, y. taki, t. sakane, h. sezaki, and y. furuyama, j. pharm. sci. 91 (2002) 669-679. admet & dmpk 3(1) (2015) 51-67 adaptation of human gut epithelial model doi: 10.5599/admet.3.1.169 67 [25] r. honeywell, k. yarzadah, e. giovannetti, n. losekoot, e. f. smit, m. walraven, j. s. lind, c. tibaldi, h. m. verheul, and g. j. peters, j. chromatogr. b analyt. technol. biomed. life sci. 878 (2010) 1059-1068. [26] p. r. wielinga, e. de waal, h. v. westerhoff, and j. lankelma, j. pharm. sci. 88 (1999) 1340-1347. [27] f. broekman, e. giovannetti, and g. j. peters, world j. clin. oncol. 2 (2011) 80-93. [28] e. galvani, r. alfieri, e. giovannetti, a. cavazzoni, m. s. la, m. galetti, c. fumarola, m. bonelli, m. mor, m. tiseo, g. j. peters, p. g. petronini, and a. ardizzoni, curr. pharm. des. 19 (2013) 818-832. [29] s. yamashita, t. furubayashi, m. kataoka, t. sakane, h. sezaki, and h. tokuda, eur. j. pharm. sci. 10 (2000) 195-204. [30] j. h. hooijberg, h. m. pinedo, c. vrasdonk, w. priebe, j. lankelma, and h. j. broxterman, febs lett. 469 (2000) 47-51. [31] c. lemos, i. kathmann, e. giovannetti, h. dekker, g. l. scheffer, c. calhau, g. jansen, and g. j. peters, int. j cancer 123 (2008) 1712-1720. [32] c. g. da silva, r. j. honeywell, h. dekker, and g. j. peters, expert. opin. drug metab toxicol. 29 (2015) 1-15. ©2015 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ thickness of the aqueous boundary layer in stirred microtitre plate permeability assays (pampa and caco-2), based on the levich equation doi: https://doi.org/10.5599/admet.1568 249 admet & dmpk 10(4) (2022) 249-252; doi: https://doi.org/10.5599/admet.1568 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index short communication thickness of the aqueous boundary layer in stirred microtitre plate permeability assays (pampa and caco-2), based on the levich equationǂ alex avdeef in-adme research, 1732 first avenue #102, new york, ny 10128 usa. e-mail: alex@in-adme.com; tel.: +1-646-678-5713 orcid id: 0000-0002-3139-5442 received: october 24, 2022; revised: october 27, 2022; published: december 06, 2022 in permeability assays using microtitre plates, either based on cellular models (e.g., caco-2, mdck) or pampa (parallel artificial membrane permeability assay) [1-3], the thickness of the aqueous boundary layer (abl) has been approximated by: aq abl d h k f    (1) where daq is the diffusivity, f is the stirring frequency (rpm), and k and  are fitted constants [4-9]. based on testosterone caco-2 measurements, karlsson and artursson [5] reported  = 1 and implied k = 0.57 x 10-6 cm/s. adson et al. [6] reported k = 4.1 x 10-6 cm/s and  = 0.8, also for testosterone. orbital shakers were used to agitate the microtitre plates during the assay, as is the common practice in cellular assays. both groups noted that the k parameter is a function of aqueous diffusivity, kinematic viscosity, and geometrical factors. adson et al. [6] pondered on the  factor being greater than the theoretically expected value of 0.5 and reasoned that the asymmetric hydrodynamic conditions of the transwell plates may have led to the elevated values. in a pampa study of 53 ionizable molecules, avdeef et al. [9] determined the abl permeability, pabl, using the pkaflux method at four different stirring speeds (49, 118, 186, 622 rpm). efficient individual-well magnetic stirring (using the gut-box device) was used in their study. since pabl = daq / habl, the constants k and  can be determined for each molecule by linear regression based on log pabl = log k +  log f. the pkaflux method uniquely made such an analysis possible. the least-squares refined parameters (based on several molecules) were reported as k = 23.1 x 10-6 cm/s and  = 0.709. the implicit assumption in the analysis was that for a given rate of stirring, there is a unique abl thickness for all molecules. ǂ this contribution is dedicated to the memory of dr. konstantin tsinman. the work here is based on discussions with konstantin before his tragic passing in 2020 from covid-19. during the discussions, he kindly shared the data reported here at 21 and 313 rpm. https://doi.org/10.5599/admet.1568 https://doi.org/10.5599/admet.1568 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:alex@in-adme.com alex avdeef admet & dmpk 10(4) (2022) 249-252 250 in the above three studies, different values of  were reported, all greater than the theoretical value of 0.5 expected from the solution to the convective diffusion model partial differential equation, based on the rotating disk geometry, according to levich [10]. in the theoretical model, the thickness of the abl may be calculated from: levich 1/6 1/3 1/2 abl aq4.98h d f   (2) where  is the kinematic viscosity (cm2/s). if the levich equation were applicable to microtitre plate permeability assay geometries, then eq. (2) suggests that k = 0.201 1/6 daq2/3, provided  were 0.5 in eq. (1). hence, each molecule in the permeability assay would be expected to have its own habl value, depending on its diffusivity. according to pohl et al. [11], such "theoretical predictions ...[are]... widely ignored." moreover, using ion-selective microelectrodes, pohl and coworkers unequivocally showed that habl varied with ionic substances at a given level of stirring. in this communication, it is hypothesized that the theoretical  = 0.5 was obscured in prior caco-2 and pampa microtitre plate permeability studies [5-9], either because (i) k was evaluated without explicit consideration of the daq term from the levich equation, and/or (ii) the data were not of sufficient sensitivity to reveal the theoretical values. we proceeded to test the hypothesis with pampa data by re-arranging the levich equation into a parametric form. combining pabl = daq / habl with eq. (2) and converting into the logarithmic form: abl aq 2 log log log 3 p d a b f   (3) with the theoretical constants a = log (0.201 -1/6) = -0.356 (25 °c) and b = 0.5. we applied eq. (3) to the (pabl, f) data of avdeef et al. [9], augmented with additional measurements at 21 and 313 rpm (table 1), and found a = -0.731 and b = 0.505 (r2 = 0.93, sd = 0.09, f = 50, n = 6). the plot of the data used in the re-analysis is shown in figure 1. the slope factor, 0.505, is so close to the theoretical value that we propose to simply use the theoretical value henceforth. substituting the new parameters into eq. (3) and converting the resulting equation to the form of eq. (2) results in: pampa 1/6 1/3 1/2 abl aq11.8h d f   (4) table 1. aqueous boundary layer permeability data from eqs. (2) and (4), hablpampa/habllevich = 2.4. the geometry of the rotating disk apparatus allows the convective flow to reach closer to the rotating surface (thus diminishing the thickness of the pure diffusion layer) compared to the geometry of magnetically stirred pampa wells. the stirring of caco-2 plates by orbital shakers produces even a greater ratio, hablcaco-2/habllevich, indicating less "efficient" stirring [9]. table 2 shows sample calculations using eqs. (2) and (4) for three drugs, widely ranging in size. pohl et al. [11] suggested that if a single reference compound is used to calibrate the geometrical factor, then calculations of subsequent habl should be according to the diffusivity dependence in the levich equation: f (rpm) log pabl ⅔ log daq a sd n b 21 0.037 c 0.119 15 49 0.004 d 0.154 5 118 0.294 d 0.104 6 186 0.435 d 0.340 51 313 0.495 c 0.268 49 622 0.737 d 0.222 22 a pabl is aqueous boundary layer permeability determined by the pkaflux method [2,8,9]. b number of measurements averaged. c this work. d values averaged from ref [9]. admet & dmpk 10(4) (2022) 249-252 thickness of the aqueous boundary layer in pampa doi: https://doi.org/10.5599/admet.1568 251 1/3 aqref abl abl ref aq d h h d          (5) eq. (5) was experimentally verified with several combinations of ions and buffers by pohl et al. [11], using ph and other ion-selective microelectrodes to directly measure the change in concentrations in the aqueous boundary layer adjacent to black lipid membranes. figure 1. the averaged log pabl 2/3logdaq vs. log f (rpm) plot of ionizable molecules, with pampa measurements done at six different stirring speeds. the data are from avdeef et al.[9], augmented with previously unpublished measurements at 21 and 313 rpm. the values in parentheses refer to the rpm values. table 2. aqueous boundary layer (abl) thickness at 300 rpm compound t (oc)  (cm2/s) a daq (cm2/s) b habllevich (m) c hablpampa (m) d vincristine 25 0.00893 3.49e-06 20 33 37 0.00697 4.67e-06 21 35 testosterone 25 0.00893 5.58e-06 23 39 37 0.00697 7.47e-06 25 41 benzoic acid 25 0.00893 8.21e-06 26 44 37 0.00697 1.10e-05 28 47 a values of kinematic viscosity, , were taken from riddick and bunger [12]. b diffusivity, daq, calculated by the procedure described elsewhere [8]. c values of the abl thicknesses, habllevich, were calculated by eq. (2). d hablpampa calculated by eq. (4). in conclusion, the stirring frequency exponent of -1/2 in the theoretical levich expression appears to apply to pampa assays, where efficient individual-well magnetic stirring ( > 20 rpm) is used. the same may be true for caco-2 assays, although additional measurements at varied stirring speeds would make this a more confident assertion. if a single molecule is used as a stirring calibrant, then it seems reasonable to use the scaling suggested by eq. (5) with microtitre plate data. as table 2 suggests, the error in calculating habl based on unscaled hablref can be as high as 30 %. hence, it is prudent to incorporate eq. (5) in the calibration procedure. this is especially important to bear in mind for nonionizable molecules since the pkaflux method cannot be directly applied to them. this is of practical importance in pampa and perhaps cellular assays as well. https://doi.org/10.5599/admet.1568 alex avdeef admet & dmpk 10(4) (2022) 249-252 252 references [1] m. kansy, f. senner, k. gubernator. physicochemical high throughput screening: parallel artificial membrane permeability assay in the description of passive absorption processes. j. med. chem. 41 (1998) 1007-1010. https://doi.org/10.1021/jm970530e. [2] a. avdeef. absorption and drug development. 2nd edn. wiley-interscience, hoboken, nj, 2012. [3] m. kansy, a. avdeef, h. fischer. advances in screening for membrane permeability: high-resolution pampa for medicinal chemists. drug discovery today: technologies. 1 (2005) 349-355. https://doi.org/10.1016/j.ddtec.2004.11.013. [4] i. komiya, j.y. park, n.f.h. ho, w.i. higuchi. quantitative mechanistic studies in simultaneous fluid flow and intestinal absorption using steroids as model solutes. int. j. pharm. 4 (1980) 249-262. https://doi.org/10.1016/0378-5173(80)90140-4. [5] j. karlsson, p. artursson. a method for the determination of cellular permeability coefficients and aqueous boundary layer thickness in monolayers of intestinal epithelial (caco-2) cells grown in permeable filter chambers. int. j. pharm. 71 (1991) 55-64. https://doi.org/10.1016/03785173(91)90067-x. [6] a. adson, p.s. burton, t.j. raub, c.l. barsuhn, k.l. audus, n.f.h. ho. passive diffusion of weak organic electrolytes across caco-2 cell monolayers: uncoupling the contributions of hydrodynamic, transcellular, and paracellular barriers. j. pharm. sci. 84 (1995) 1197-1204. https://doi.org/10.1002/jps.2600841011. [7] n.f.h. ho, t.j. raub, p.s. burton, c.l. barsuhn, a. adson, k.l. audus, r. borchardt. quantitative approaches to delineate passive transport mechanisms in cell culture monolayers. in: amidon, g.l., lee, p.i., topp, e.m. (eds.). transport processes in pharmaceutical systems. marcel dekker, new york, 2000, pp. 219-316. [8] a. avdeef, p. artursson, s. neuhoff, l. lazarova, j. gräsjö, s. tavelin. caco-2 permeability of weakly basic drugs predicted with the double-sink pampa pkaflux method. eur. j. pharm. sci. 24 (2005) 333349. https://doi.org/10.1016/j.ejps.2004.11.011. [9] a. avdeef, p.e. nielsen, o. tsinman. pampa a drug absorption in vitro model. 11. matching the in vivo unstirred water layer thickness by individual-well stirring in microtitre plates. eur. j. pharm. sci. 22 (2004) 365-374. https://doi.org/10.1016/j.ejps.2004.04.009. [10] v.g. levich. physicochemical hydrodynamics. prentice-hall, englewood cliff, nj, 1962. [11] p. pohl, s.m. saparov, y.n. antonenko. the size of the unstirred layer as a function of the solute diffusion coefficient. biophys. j. 75 (1998) 1403-1409. https://doi.org/10.1016/s0006-3495(98)74058-5. [12] j.a. riddick, w.b. bunger. organic solvents physical properties and methods of purification. wileyinterscience, new york, 1970, pp. 28-34. ©2022 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/) https://doi.org/10.1021/jm970530e https://doi.org/10.1016/j.ddtec.2004.11.013 https://doi.org/10.1016/0378-5173(80)90140-4 https://doi.org/10.1016/0378-5173(91)90067-x https://doi.org/10.1016/0378-5173(91)90067-x https://doi.org/10.1002/jps.2600841011 https://doi.org/10.1002/jps.2600841011 https://doi.org/10.1016/j.ejps.2004.11.011 https://doi.org/10.1016/j.ejps.2004.04.009 https://doi.org/10.1016/s0006-3495(98)74058-5 https://doi.org/10.1016/s0006-3495(98)74058-5 http://creativecommons.org/licenses/by/4.0/ pharmacokinetics of ceftriaxone-tazobactam (8:1) combination in healthy and escherichia coli induced diarrhoeic birds doi: http://dx.doi.org/10.5599/admet.1170 180 admet & dmpk 10(3) (2022) 180-196; doi: https://doi.org/10.5599/admet.1170 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper pharmacokinetics of ceftriaxone-tazobactam (8:1) combination in healthy and escherichia coli induced diarrhoeic birds mithin u.c.1, rinku buragohain1, pradip k das2, tapan k mandal1, rabindra n hansda3, siddhartha n joardar4, indranil samanta4 and tapas k sar1,* 1 department of veterinary pharmacology and toxicology, west bengal university of animal and fishery sciences, 37 k. b. sarani, kolkata-700037, west bengal, india 2 department of veterinary physiology, west bengal university of animal and fishery sciences, 37 k. b. sarani, kolkata700037, west bengal, india 3 department of veterinary pathology, west bengal university of animal and fishery sciences, 37 k. b. sarani, kolkata700037, west bengal, india 4 department of veterinary microbiology, west bengal university of animal and fishery sciences, 37 k. b. sarani, kolkata-700037, west bengal, india *corresponding author: e-mail: tapas.sar@rediffmail.com; tel.: +919474821915 received: november 13, 2022; revised: march 03, 2022; available online: march 12, 2022 abstract antibiotic-resistant escherichia coli infection of poultry causes significant economic losses. extended spectrum β lactamases (esbl) producing e. coli was inoculated in a broiler, rhode island red and haringhata black birds orally at 56×108 c.f.u. ml-1 for induction of diarrhoea. pharmacokinetics of ceftriaxone-tazobactam combination (8:1) was studied following a single intramuscular injection at 28.125 mg kg-1 and the combination was administered twice daily to treat such infection. plasma concentration of both ceftriaxone persisted up to 8 h in experimental birds and maintained an approximate ratio of 8:1 with tazobactam for a period of 2 h, 0.25 h and 0.75 h, respectively in a broiler, rhode island red and haringhata black birds. the kel was significantly lower in all experimental birds compared to healthy birds. efficacy study was conducted in diarrhoeic birds by administration of ceftriaxone-tazobactam combination at 28.125 mg kg-1 body weight twice daily intramuscularly for three days which caused an increase in specific antibody titre in the broiler on 5th day and in rhode island red birds 10th day. however, haringhata black birds were inherently showed more resistance towards the infection. the combination of ceftriaxone and tazobactam in the ratio of 8:1 can be an effective treatment to combat esbl producing e. coli infections. ©2022 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords disposition; efficacy study; antibiotic-β lactamase inhibitor combination; poultry; intramuscular injection introduction the poultry sector continues to grow rapidly in many parts of the world. the poultry sector is broadly divided into two sub-sectors, one is the highly organized commercial sector with about 80 % of the total market share and the other being the unorganized sector with about 20 % of the total market share in india [1]. the unorganized sector is also referred to as backyard poultry which plays a key role in supplementary http://dx.doi.org/10.5599/admet.1170 https://doi.org/10.5599/admet.1170 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:tapas.sar@rediffmail.com http://creativecommons.org/licenses/by/4.0/ admet & dmpk 10(3) (2022) 180-196 ceftraxone-tazobactam combination in poultry diarrhoea doi: http://dx.doi.org/10.5599/admet.1170 181 income generation and family nutrition to economically weaker sections (ews) in india. the rhode island red, an american breed of chicken (gallus gallus domesticus) is a backyard dual-purpose poultry reared by the ews section of people for its egg-laying ability and hardiness. whereas haringhata black is an indigenous poultry breed found in the northern part of north 24 parganas and southern part of nadia districts of west bengal, india. escherichia coli infections have various disease expressions in domestic birds, including salpingitis, synovitis, omphalitis, and/or chronic respiratory disease. colibacillosis is one of the principal causes of morbidity and mortality in the poultry industry and it is responsible for significant worldwide economic losses [2]. the experimental use of vaccines against e. coli has had limited success due to the antigen used and the methods of inactivation and administration. vaccination of broiler breeding hens with homologous e. coli had demonstrated that maternally derived antibodies protected against colibacillosis for only two weeks post-hatching [3,4]. recommended treatment protocols for e. coli infections include antibiotics with a broad spectrum of activity and sulpha drugs. however, isolates of e. coli from poultry are sometimes found to be resistant to commonly used antibacterial drugs. dhillon and jack reported failure of oxytetracycline treated feed to reduce losses during outbreaks of colibacillosis in commercial caged layers [5]. in another study performed in spain, blanco et al. isolated a septicaemic strain of e. coli resistant to fluoroquinolones [6]. extended spectrum β lactamases (esbls) are plasmid-mediated β lactamases that have the ability to hydrolyze β lactam antibiotics containing an oxyimino group (e.g. ceftazidime, ceftriaxone, cefotaxime or aztreonam). these esbls were most commonly found in klebsiella pneumoniae, but are being increasingly found also in e. coli, proteus mirabilis and other members of the enterobacteriaceae. the vast majority of esbls are derivatives of tem-1 (the common plasmid-mediated β lactamase of organisms such as e. coli) or shv-1 (the common chromosomally mediated β lactamase of k. pneumoniae). tem-1 and shv-1 can inactivate ampicillin but not the third-generation cephalosporins [7,8]. such esbl containing e. coli has already been isolated from backyard poultry in india [9,10]. the infections caused by esbl strains of e. coli were often associated with nosocomial epidemics, and no clear association with food consumption or contact with animals was noticed. but, sometimes poultry meat is also consumed without proper boiling, which may be a source of resistant e. coli infections in consumers. hussain et al. reported that prevalence rates of esbl producing e. coli was 46 % among broiler chicken meat and was 15 % among free-range/backyard chicken meat. e. coli isolated from broiler and free-range chicken meat exhibited 68 % and 8 % prevalence rates, respectively, for multi-drug resistant e. coli [11]. ceftriaxone is a broad-spectrum cephalosporin with potent activity against gram-positive and gram-negative bacteria, including enterobacteriaceae, heamophilus influenza, streptococcus pneumonia and other nonenterococcal streptococci [12]. li et al. studied the pharmacokinetics of ceftriaxone in broiler poultry following single intravenous administration at 50 mg kg-1 and determined tissue residue level of ceftriaxone in plasma, liver, kidney, heart, lungs and muscle tissue [13]. pharmacokinetics study of ceftriaxone in layer poultry following single-dose intravenous and intramuscular administration at 50 mg kg-1 showed favourable pharmacokinetic characteristics for its use as an effective antibiotic against septicemic diseases of poultry [14]. queenan et al. reported a minimum inhibitory concentration (mic) of ≤ 0.12 µg ml-1 for ceftriaxone against inoculums of 105 and 106 c.f.u. ml-1 of tem-1 containing e. coli [15]. tazobactam is an inhibitor of a variety of plasmid-mediated β lactamases elaborated by some bacteria. the ceftriaxone and sulbactam (another β lactamase inhibitor) combination was reported to be more effective than ceftriaxone alone for the prevention of mutation in esbl producing organisms in vitro with mpc (mutation prevention concentration) of > 256 µg ml-1 for ceftriaxone-sulbactam combination and >512 µg ml-1 for ceftriaxone alone [16]. whereas payne et al. reported that tazobactam has greater β lactamases inhibitory activity than sulbactam [17]. ceftriaxone-tazobactam combination (8:1) therapy for esbl producing e. coli infection could be an attractive option. in our previous study, we reported the pharmacokinetics of ceftriaxone and http://dx.doi.org/10.5599/admet.1170 t.k. sar et al. admet & dmpk 10(3) (2022) 180-196 182 tazobactam following intramuscular administration at 25 mg kg-1 and 3.125 mg kg-1, respectively, in the broiler haringhata black and rhode island red poultry [18]. moreover, ceftriaxone-tazobactam combination (8:1) at 28.1 mg kg-1 intramuscularly twice daily for three days in poultry did not alter aspartate transaminase and alanine transaminase activities significantly [19]. therefore, the present research work was undertaken to study the pharmacokinetic profile of ceftriaxone-tazobactam combination (8:1) in healthy and esbl e. coli infected broiler, rhode island red and haringhata black birds following single intramuscular dosing and to evaluate the efficacy of ceftriaxone-tazobactam combination (8:1) against esbl producing e. coli infection. experimental drugs and chemicals analytical grade ceftriaxone sodium (purity ≥ 95 %) was obtained from alembic limited, mumbai, india and analytical grade tazobactam was obtained from sigma aldrich. all other chemicals used in the study were obtained from e. merck (india) and sigma chemicals co., usa. experimental birds a total of eighteen clinically healthy adult poultry (six poultry each of broiler, rhode island red and haringhata black) were collected from instructional poultry farm, west bengal university of animal and fishery sciences, kolkata, india. the broiler birds were approximately 5 weeks of age and the rhode island red and haringhata black birds were approximately 12 weeks of age. all birds were quarantined for 14 days period prior to the start of the experiment. six birds of each breed were caged in three separate metabolic cages made of stainless steel and provided with commercial grower feed with ad libitum potable drinking water. all the experimental procedures were approved by the institutional animal ethics committee (iaec), west bengal university of animal and fishery sciences, india and conducted as per the ethical guidelines of the committee for the purpose of control and supervision of experiments on animals (cpcsea), india. pharmacokinetic study the pharmacokinetic profile of ceftriaxone and tazobactam was studied in six healthy broiler birds (gr bct-h), six rhode island red birds (gr rct-h) and six haringhata black birds (gr hct-h) following single intramuscular administration of ceftriaxone-tazobactam combination (8:1) at 28.125 mg kg-1. the same six healthy broiler, rhode island red and haringhata black birds were used for induction of diarrhoea by oral inoculation of esbl producing e. coli culture after allowing a washout period of 15 days. the pharmacokinetic profile of ceftriaxone and tazobactam was also studied in these diarrhoeic broiler birds (gr bct-d), rhode island red birds (gr rct-d) and haringhata black birds (gr hct-d) following single intramuscular administration of ceftriaxone-tazobactam combination (8:1) at 28.125 mg kg-1. blood samples were collected from the collateral wing vein at predetermined time intervals for estimation of each individual drug concentration. efficacy study for efficacy study, six apparently healthy birds each of broiler (gr be), rhode island red (gr re) and haringhata black (gr he) breed were used for induction of diarrhoea following oral inoculation of esbl producing e. coli. following induction of diarrhoea, ceftriaxone-tazobactam combination (8:1) was administered at 28.125 mg kg-1 two times daily (at 12 h interval) intramuscularly for three days. the dosage regimen was calculated on the basis of maintenance of plasma ceftriaxone concentration above mic level in a pharmacokinetic study. treatment with a prescheduled dosage regimen of the ceftriaxone-tazobactam admet & dmpk 10(3) (2022) 180-196 ceftraxone-tazobactam combination in poultry diarrhoea doi: http://dx.doi.org/10.5599/admet.1170 183 combination was employed on 7th-day post-inoculation in broiler and rhode island red birds and on 8th-day post-inoculation of the second challenge in haringhata black birds. all the birds were closely observed during and after treatment up to a period of 1 month. induction of infection with pathogenic e. coli possessing esbl genes the pathogenic e. coli was isolated from broiler birds in a local poultry farm. the bird was 29 days old and was suffering from diarrhoea followed by death. the strain belonged to o62 serogroup, pathogenic to experimental chickens and possessed the genes for tem (bla tem). however, the strain was negative for other esbl genes such as bla ctx-m and bla shv. the isolate was maintained at the department of veterinary microbiology, west bengal university of animal & fishery sciences, kolkata. for induction of diarrhoea in the experimental broiler, rhode island red and haringhata black birds, 56 × 108 c.f.u. ml-1 of the bacterial culture was inoculated by oral route. however, haringhata black birds did not show any clinical sign following the initial challenge of esbl producing e. coli (tem-1) sub-culture (56 × 108 c.f.u. ml-1). due to failure of the initial challenge, these birds were again inoculated orally with a higher second dose (112 × 108 c.f.u. ml-1 subculture) after 21 days of oral inoculation with 56 × 108 c.f.u. ml-1 subculture of esbl producing e. coli antibiotic sensitivity test the esbl producing e. coli isolates were tested for their sensitivity and resistance to ceftriaxone and ceftriaxone-tazobactam combination by disc diffusion method [20]. the result was interpreted as per the clsi guidelines or the standard information provided by the manufacturer. collection of samples blood samples (2 ml) were collected from the wing vein in heparinized test tubes at 0 (pre-dosing), 0.04, 0.08, 0.25, 0.5, 0.45, 1, 2, 4, 6, 8, and 12 h post-dosing. plasma was then separated by centrifugation at 3000 rpm for 20 min and stored at 4 °c for pharmacokinetics analysis. blood samples (1 ml) without anticoagulant were collected from the experimental birds at a pre-inoculation time (0 day) and 3, 6, 9, 12, 15, 18 and 21 days post-inoculation of esbl producing e. coli. the collected blood samples were allowed to clot at 25 °c and the serum was collected in sterile vials. the serum samples were further centrifuged at 2500 rpm for 15 min, to remove residual rbc and it was stored at -20 °c until further analysis. faecal samples were collected following oral inoculation of esbl producing e. coli and on 3rd day of ceftriaxone-tazocactam treatment in the induced diarrhoeic birds for the efficacy study. estimation of ceftriaxone in plasma ceftriaxone concentration in plasma was estimated with the method of sar et al. [21]. the mobile phase was prepared with the method mentioned in united states pharmacopoeia (usp). the mobile phase consisted of 44 ml of potassium phosphate buffer (ph 7.0), 4 ml sodium citrate buffer (ph 5.0), 400 ml of hplc grade acetonitrile, 3.2 g of tetraheptyl ammonium bromide and 552 ml of hplc grade milipore water. to 1 ml of plasma, 1 ml acetonitrile (hplc grade) was added for deproteinisation and shaken vigorously for 1 min followed by centrifugation at 5000 rpm for 20 min. the supernatant was collected and filtered through 0.2 µm membrane filter. the sample filtrate volume was adjusted to 1 ml with acetonitrile and 20 µl of the sample was injected into high-performance liquid chromatography (hplc). measurement was performed at a wavelength of 254 nm. estimation of tazobactam in plasma a binary mobile phase consisting of 25 mm potassium dihydrogen phosphate buffer, ph 6.2 and http://dx.doi.org/10.5599/admet.1170 t.k. sar et al. admet & dmpk 10(3) (2022) 180-196 184 acetonitrile (94:6, v/v) was used for the estimation of tazobactam. extraction of tazobactam was done with the method of ocampo et al. to 0.5 ml of plasma sample, 0.5 ml of mobile phase and 3 ml of acetonitrile was added and vortexed for 3 min. the whole mixture was centrifuge at 2000 rpm for 10 min and the supernatant was decanted into a test tube containing 3 ml of dichloromethane. the whole mixture was vortexed for 1 min followed by centrifugation at 3000 rpm for 10 min. 20 µl of the aqueous layer was injected into hplc for the estimation of tazobactam concentration [22]. measurement was performed at a wavelength of 210 nm [23]. instrumental condition shimadzu lc-20 at liquid chromatograph coupled with photo diode-array (pda) detector attached with computer spd-mxa 10 software was used for the analysis of the drugs. 5µ luna c18 (2); 250 × 4.6 mm (rp) column was used and 1.5 ml min-1 flow rate was maintained. pharmacokinetic parameters pharmacokinetic parameters of ceftriaxone and tazobactam were determined from the computerized curve fitting programme “pharmkit” supplied by the department of pharmacology, jipmer, puducherry, india. the data obtained from this programme in healthy and diseased birds were analysed for deriving some of the pharmacokinetic parameters as per standard formulae of baggot [24]. the fitted pharmacokinetic parameters were mrt, β, t½ β, auc, vdarea, clb, cmax and tmax while the extrapolated pharmacokinetic parameters were a, b, ka, t½ ka, vdc, vdss, k12, k21, kel, fc and t~p. bacterial colony count bacterial colony count of faecal samples was conducted with the standard protocol by quinn et al. [25]. the cloacal swabs were collected from the studied birds using sterile cotton swabs (himedia, india) and placed into a sterile transport medium (hi-media, india). the cloacal swabs were stored at 4 °c, and the content was processed for bacteriological enumeration on the same day of collection. the swab content was serially diluted to 10 folds with sterile phosphate buffer saline solution (pbs). from 10−2 and 10 −4 dilution, 10 μl was placed on eosin methylene blue (emb) agar (himedia, mumbai, india) for e. coli. the plates were incubated aerobically at 37 °c for 24 hours. the typical colonies were enumerated in a colony counter (digital colony counter, la663, himedia, india) and the numbers were expressed as colony-forming units (c.f.u.) per gram. i-elisa to detect anti e. coli antibody in serum samples of diarrhoeic broiler, rhode island red and haringhata black bird, plate elisa was performed based on principles of i-elisa as per mockett et al. [26] with some modifications. the original protocol detected the titer at 405 nm. however, the present study detected the titer at 492 nm wavelength. preparation of somatic antigen the laboratory maintained strain of esbl producing e. coli was revived on a trypticase soy broth (tsa) slant. it was then transferred to tsb and incubated at 30 °c for 24 h. the cells were harvested by centrifugation at 7500 rpm for 25 min at 4 °c. the cell pellets were washed with normal saline solution (nss) and finally re-suspended in 10 ml of the nss. sonication was performed after adding 25 mm of pmsf and 24 mm of edta. the bacterial cell suspension in nss was sonicated on the ice at the amplitude of 50 for 0.5 to 1 min, for each cycle giving 1 min interval in between. the process was repeated by 6-8 cycles. soluble sonicated extract was centrifuged at 7500 rpm for 25 min at 4 °c. the supernatant was collected and the soluble protein (somatic antigen) was concentrated by sucrose [27]. admet & dmpk 10(3) (2022) 180-196 ceftraxone-tazobactam combination in poultry diarrhoea doi: http://dx.doi.org/10.5599/admet.1170 185 statistical analysis the data were expressed as mean ± standard error (s.e.). the data were analyzed statistically using glm (general linear model) method, comparison through lsd (least significant difference) and t-test of ibm spss statistic, version 21, 2012. results and discussion antibiotic sensitivity test the esbl producing e. coli isolates showed intermediate sensitivity for ceftriaxone and high sensitivity to ceftriaxone-tazobactam combination (figure 1). but, the esbl producing e. coli isolates were resistant to ampicillin, amoxicillin, streptomycin, cephalexin and cefotaxime. figure 1. plate showing sensitivity pattern of only ceftriaxone and ceftriaxone-tazobactam (8:1) combination against esbl producing e. coli during antibiotic sensitivity test. standardization of the analytical techniques of ceftriaxone and tazobactam the recovery percentages of ceftriaxone and tazobactam from plasma were 81.14 ± 4.95 % and 83.01 ± 3.6 %, respectively. since the recovery percentages of both the drugs were more than 80 % in plasma, the recoveries were found to be satisfactory. the limit of detection for both the drugs in plasma was 0.5 ppm and sensitivity was 0.25 ppm. the linearity of calibration curves was checked for both the drugs and linearity was found to be maintained in the range of 0.5 to 25 ppm for ceftriaxone and 0.5 to 10 ppm for tazobactam in plasma. however, the retention time of ceftriaxone and tazobactam showed inter-day variation. pharmacokinetics of ceftriaxone and tazobactam in healthy and esbl e. coli infected broiler birds following single intramuscular administration of ceftriaxone-tazobactam combination (8:1) ceftriaxone persisted for a longer duration of 8 h in bct-d birds compared to bct-h birds (figure 2). the elimination rate (β) of ceftriaxone is slower in the gr bct-d birds that corresponded with significantly higher plasma concentrations at 2, 4 and 6 h and a 1.5-fold higher auc0-inf (table 1). tazobactam persisted up to 6 h in bct-d and bct-h birds and maintained an approximate ratio of 1:8 with ceftriaxone up to 2 h. http://dx.doi.org/10.5599/admet.1170 t.k. sar et al. admet & dmpk 10(3) (2022) 180-196 186 the elimination rate of tazobactam from the central compartment (kel) was significantly slower in bct-d birds (table 2), which also corresponds with the higher plasma concentration of tazobactam at 2, 4 and 6 h. figure 2. semilogarithmic plot of plasma concentration of ceftriaxone and tazobactam in healthy (gr bct-h) and esbl e. coli infected broiler birds (gr bct-d) following single intramuscular administration of ceftriaxonetazobactam combination (8:1) at 28.125 mg kg-1. table 1. pharmacokinetic parameters of ceftriaxone in healthy (gr bct-h) and esbl e. coli infected broiler birds (gr bct-d) following single intramuscular administration of ceftriaxone-tazobactam combination (8:1) at 28.125 mg kg-1. kinetic parameters gr bct-h gr bct-d a (μg ml-1) 23.47* ± 1.27 20.23 ± 0.86 b (μg ml-1) 23.40* ± 1.25 20.15 ± 0.85 ka (h-1) 42.16 ± 8.51 44.77 ± 7.29 t½ ka (h) 0.02 ± 0.01 0.02 ± 0.00 β (h-1) 0.69* ± 0.05 0.43 ± 0.02 t½ β (h) 1.03 ± 0.07 1.62* ± 0.07 auc0-inf (μg h ml-1) 34.86 ± 1.35 47.39* ± 2.63 vdarea (l kg-1) 1.09 ± 0.07 1.23 ± 0.06 clb (l kg-1 h-1) 12.34* ± 0.58 8.89 ± 0.49 mrt (h) 1.45 ± 0.08 2.35* ± 0.10 vdc (l kg-1) 0.54 ± 0.03 0.62* ± 0.03 vdss (l kg-1) 1.14 ± 0.08 1.28 ± 0.06 k12 (h-1) 20.11 ± 4.23 21.79 ± 3.65 k21 (h-1) 21.38 ± 4.25 22.55 ± 3.63 kel (h-1) 1.36* ± 0.10 0.86 ± 0.04 fc 0.51 ± 0.00 0.50 ± 0.00 t~p 0.96 ± 0.01 0.98 ± 0.00 cmax_calc (μg ml-1) 21.18 ± 1.47 19.00 ± 0.89 tmax_calc (h) 0.16 ± 0.03 0.15 ± 0.01 t > mic 11.08 h (92 %) f 0.14 0.12 mean values in a raw bearing superscript * vary significantly (p < 0.05); [n=6] admet & dmpk 10(3) (2022) 180-196 ceftraxone-tazobactam combination in poultry diarrhoea doi: http://dx.doi.org/10.5599/admet.1170 187 table 2. pharmacokinetic parameters of tazobactam in healthy (gr bct-h) and esbl e. coli infected broiler birds (gr bct-d) following single intramuscular administration of ceftriaxone-tazobactam combination (8:1) at 28.125 mg kg-1. kinetic parameters gr bct-h gr bct-d a (μg ml-1) 3.37 ± 0.22 3.06 ± 0.11 b (μg ml-1) 3.42 ± 0.21 3.10 ± 0.13 ka (h-1) 14.53 ± 3.12 11.91 ± 1.39 t½ ka (h) 0.06 ± 0.01 0.06 ± 0.01 β (h-1) 0.24* ± 0.01 0.19 ± 0.01 t½ β (h) 2.99 ± 0.18 3.72* ± 0.24 auc0-inf (μg h ml-1) 15.02 ± 1.30 16.72 ± 0.74 vdarea (l kg-1) 0.96 ± 0.06 1.04 ± 0.05 clb (l kg-1 h-1) 3.76 ± 0.34 3.32 ± 0.16 mrt (h) 4.38 ± 0.26 5.41* ± 0.38 vdc (l kg-1) 3.76 ± 0.22 4.09 ± 0.17 vdss (l kg-1) 7.89 ± 0.45 8.51 ± 0.28 k12 (h-1) 6.90 ± 1.56 5.66 ± 0.71 k21 (h-1) 7.40 ± 1.55 6.06 ± 0.68 kel (h-1) 0.46* ± 0.03 0.37 ± 0.02 fc 0.51 ± 0.01 0.51 ± 0.00 t~p 0.95 ± 0.02 0.96 ± 0.01 cmax_calc (μg ml-1) 3.09 ± 0.17 2.83 ± 0.09 tmax_calc (h) 0.36 ± 0.05 0.39 ± 0.03 f 0.16 0.15 mean values in a raw bearing superscript * vary significantly (p < 0.05); [n=6] pharmacokinetics of ceftriaxone and tazobactam in healthy and esbl e. coli infected rode island red birds following single intramuscular administration of ceftriaxone-tazobactam combination (8:1) ceftriaxone persisted up to 8 h in rct-d birds (fig. 3) with a cmax value of 16.99 μg ml-1. the cmax value was significantly lower in rct-d birds which corresponds with significantly higher vdarea. the slower elimination rate (β) of ceftriaxone in rch-d birds corresponds with the significantly higher auc0-inf (table 3). tazobactam persisted up to 6 h in both rct-d and rct-h birds and it maintained 1:8 plasma concentration ratio with ceftriaxone for a much shorter time of 0.25 h. the mean value of kel was significantly higher in gr rct-d birds (0.46 ± 0.02 h-1) compared to gr rct-h birds (0.39 ± 0.02 h-1) (table 4). pharmacokinetics of ceftriaxone and tazobactam in healthy and esbl e. coli infected haringhata black birds following single intramuscular administration of ceftriaxone-tazobactam combination (8:1) ceftriaxone persisted for a longer duration in the plasma of hct-d group (up to 8 h) compared hct-h group (up to 6 h) (fig. 4). the elimination rate of ceftriaxone (β) is slower in gr hct-d group that corresponded with the higher auc0-inf. the cmax was significantly lower in hct-d birds, which corresponds with significantly higher vdarea values (table 5). tazobactam was detected up to 8 h in both hct-h and hctd birds and maintained a plasma concentration of 1:8 with ceftriaxone up to 0.75 h (figure 4). mean values of ka, k12 and k21 were significantly higher in gr hct-d birds compared to gr hct-h birds indicating rapid absorption as well as distribution of tazobactam in diseased haringhta black birds (table 6). http://dx.doi.org/10.5599/admet.1170 t.k. sar et al. admet & dmpk 10(3) (2022) 180-196 188 figure 3. semilogarithmic plot of plasma concentration of ceftriaxone and tazobactam in healthy (gr rct-h) and esbl e. coli infected rhode island red birds (gr rct-d) following single intramuscular administration of ceftriaxone-tazobactam combination (8:1) at 28.125 mg kg-1. table 3. pharmacokinetic parameters of ceftriaxone in healthy (gr rct-h) and esbl e. coli infected rhode island red birds (gr rct-d) following single intramuscular administration of ceftriaxone-tazobactam combination (8:1) at 28.125 mg kg-1. kinetic parameters gr rct-h gr rct-d a (μg ml-1) 22.73* ± 0.31 18.34 ± 1.00 b (μg ml-1) 22.65* ± 0.29 18.28 ± 1.00 ka (h-1) 32.46 ± 4.48 33.78 ± 4.24 t½ ka (h) 0.02 ± 0.00 0.02 ± 0.00 β (h-1) 0.70* ± 0.05 0.44 ± 0.01 t½ β (h) 1.01 ± 0.06 1.57* ± 0.05 auc0-inf (μg h ml-1) 33.66 ± 1.93 41.70* ± 1.67 vdarea (l kg-1) 1.10 ± 0.02 1.43* ± 0.11 clb (l kg-1 h-1) 12.86* ± 0.86 10.45 ± 0.59 mrt (h) 1.43 ± 0.08 2.25* ± 0.07 vdc (l kg-1) 0.55 ± 0.01 0.69* ± 0.04 vdss (l kg-1) 1.16 ± 0.02 1.43* ± 0.10 k12 (h-1) 15.24 ± 2.21 16.27 ± 2.12 k21 (h-1) 16.55 ± 2.25 17.08 ± 2.12 kel (h-1) 1.37* ± 0.10 0.88 ± 0.03 fc 0.5110* ± 0.00 0.5063 ± 0.00 t~p 0.96 ± 0.01 0.98* ± 0.00 cmax_calc (μg ml-1) 20.21* ± 0.38 16.99 ± 1.02 tmax_calc (h) 0.17 ± 0.02 0.17 ± 0.02 t > mic 11.16 h (92.67 %) f 0.14 0.11 mean values in a raw bearing superscript * vary significantly (p < 0.05) admet & dmpk 10(3) (2022) 180-196 ceftraxone-tazobactam combination in poultry diarrhoea doi: http://dx.doi.org/10.5599/admet.1170 189 table 4. pharmacokinetic parameters of tazobactam in healthy (gr rct-h) and esbl e. coli infected rhode island red birds (gr rct-d) following single intramuscular administration of ceftriaxone-tazobactam combination (8:1) at 28.125 mg kg-1. kinetic parameters gr rct-h gr rct-d a (μg ml-1) 2.96 ± 0.17 3.48* ± 0.19 b (μg ml-1) 3.07 ± 0.15 3.47 ± 0.19 ka (h-1) 13.28 ± 2.59 19.32 ± 3.19 t½ ka (h) 0.07 ± 0.02 0.04 ± 0.01 β (h-1) 0.20 ± 0.01 0.23 ± 0.01 t½ β (h) 3.45 ± 0.19 3.03 ± 0.18 auc0-inf (μg h ml-1) 15.55 ± 1.00 15.37 ± 1.30 vdarea (l kg-1) 1.05 ± 0.05 0.80 ± 0.15 clb (l kg-1 h-1) 3.56 ± 0.27 3.54 ± 0.26 mrt (h) 4.96* ± 0.28 4.28 ± 0.23 vdc (l kg-1) 4.19* ± 0.20 3.65 ± 0.18 vdss (l kg-1) 8.96* ± 0.57 7.49 ± 0.35 k12 (h-1) 6.31 ± 1.33 9.33 ± 1.61 k21 (h-1) 6.78 ± 1.26 9.76 ± 1.59 kel (h-1) 0.39 ± 0.02 0.46* ± 0.02 fc 0.52 ± 0.01 0.51 ± 0.00 t~p 0.93 ± 0.04 0.98 ± 0.01 cmax_calc (μg ml-1) 2.79 ± 0.12 3.21* ± 0.15 tmax_calc (h) 0.38 ± 0.05 0.30 ± 0.03 f 0.15 0.17 mean values in a raw bearing superscript * vary significantly (p < 0.05); [n=6] figure 4. semilogarithmic plot of plasma concentration of ceftriaxone and tazobactam in healthy (gr hct-h) and esbl e. coli infected haringhata black birds (gr hct-d) following single intramuscular administration of ceftriaxone-tazobactam combination (8:1) at 28.125 mg kg-1. http://dx.doi.org/10.5599/admet.1170 t.k. sar et al. admet & dmpk 10(3) (2022) 180-196 190 table 5. pharmacokinetic parameters of ceftriaxone in healthy (gr hct-h) and esbl e. coli infected haringhata black birds (gr hct-d) following single intramuscular administration of ceftriaxone-tazobactam combination (8:1) at 28.125 mg kg-1. kinetic parameters gr hct-h gr hct-d a (μg ml-1) 23.47* ± 1.27 18.41 ± 0.94 b (μg ml-1) 23.40* ± 1.25 18.34 ± 0.92 ka (h-1) 42.16 ± 8.51 37.00 ± 4.44 t½ ka (h) 0.02 ± 0.01 0.02 ± 0.00 β (h-1) 0.69* ± 0.05 0.44 ± 0.02 t½ β (h) 1.03 ± 0.07 1.59* ± 0.07 auc0-inf (μg h ml-1) 34.86 ± 1.35 42.32* ± 2.13 vdarea (l kg-1) 1.09 ± 0.07 1.39* ± 0.08 clb (l kg-1 h-1) 12.34* ± 0.58 10.19 ± 0.57 mrt (h) 1.45 ± 0.08 2.29* ± 0.10 vdc (l kg-1) 0.54 ± 0.03 0.69* ± 0.03 vdss (l kg-1) 1.14 ± 0.08 1.41* ± 0.07 k12 (h-1) 20.11 ± 4.23 17.88 ± 2.23 k21 (h-1) 21.38 ± 4.25 18.68 ± 2.21 kel (h-1) 1.36* ± 0.10 0.87 ± 0.04 fc 0.51 ± 0.00 0.51 ± 0.00 t~p 0.96 ± 0.01 0.98 ± 0.00 cmax_calc (μg ml-1) 21.18* ± 1.47 17.12 ± 0.94 tmax_calc (h) 0.16 ± 0.03 0.17 ± 0.02 t > mic 11.24 h (93.3 %) f 0.13 0.10 mean values in a raw bearing superscript * vary significantly (p < 0.05); [n=6] efficacy study the efficacy of the treatment schedule was evaluated on the basis of feacal e. coli count though the particular counting of esbl producing e. coli (tem-1) could not be performed (table 7). the feacal e. coli count reduced considerably in all the three groups following treatment with the ceftriaxone-tazobactam combination (8:1) at 28.125 mg kg-1 twice daily for three days. however, a significant number of e. coli were still present even after the recovery of the birds from diarrhoea because a large number of beneficial e. coli remain as a commensal organism in the intestine of birds and specific esbl producing e. coli could not be counted by the present method. antibody titre broiler (fig. 5a) and rhode island red (fig. 5b) birds showed lower antibody response during the 5th day of post-inoculation, but haringhata black birds showed higher antibody response during the 15th day of first inoculation (fig. 5c). the clinical signs of diarrhoea were observed in broiler and rhode island red birds but antibody titre did not increase significantly on day 5 (during the infection period) compared to 0 day. the antibody titre was increased significantly after 5 days of treatment with ceftriaxone-tazobactam (8:1) combination in broiler and 10 days of treatment in rhode island red birds. whereas in the case of haringhata black birds no diarrhoea was observed after initial inoculation of esbl e. coli, but antibody titre increased significantly on 15 days compared to 0 day due to natural resistance. however, following readmet & dmpk 10(3) (2022) 180-196 ceftraxone-tazobactam combination in poultry diarrhoea doi: http://dx.doi.org/10.5599/admet.1170 191 inoculation of these birds with esbl e. coli on the 21st day of first inoculation diarrhoea was observed, but treatment with ceftriaxone did not cause a significant increase in antibody titre on the 5th and 10th day of treatment. table 6. pharmacokinetic parameters of tazobactam in healthy (gr hct-h) and esbl e. coli infected haringhata black birds (gr hct-d) following single intramuscular administration of ceftriaxone-tazobactam combination (8:1) at 28.125 mg kg-1. kinetic parameters gr hct-h gr hct-d a (μg ml-1) 3.06 ± 0.11 3.22 ± 0.10 b (μg ml-1) 3.10 ± 0.13 3.20 ± 0.11 ka (h-1) 11.91 ± 1.39 18.77* ± 1.73 t½ ka (h) 0.06* ± 0.01 0.04 ± 0.00 β (h-1) 0.19 ± 0.01 0.21 ± 0.01 t½ β (h) 3.72 ± 0.24 3.30 ± 0.17 auc0-inf (μg h ml-1) 16.72 ± 0.74 15.40 ± 0.80 vdarea (l kg-1) 1.04 ± 0.05 0.99 ± 0.03 clb (l kg-1 h-1) 3.32 ± 0.16 3.50 ± 0.17 mrt (h) 5.41 ± 0.38 4.82 ± 0.21 vdc (l kg-1) 4.09 ± 0.17 3.91 ± 0.13 vdss (l kg-1) 8.51 ± 0.28 8.00 ± 0.25 k12 (h-1) 5.66 ± 0.71 9.09* ± 0.87 k21 (h-1) 6.06 ± 0.68 9.46* ± 0.86 kel (h-1) 0.37 ± 0.02 0.42 ± 0.02 fc 0.51 ± 0.00 0.50 ± 0.00 t~p 0.96 ± 0.01 0.98 ± 0.00 cmax_calc (μg ml-1) 2.83 ± 0.09 3.00 ± 0.09 tmax_calc (h) 0.39* ± 0.03 0.30 ± 0.02 f 0.14 0.15 mean values in a raw bearing superscript * vary significantly (p < 0.05); [n=6] table 7. mean values of faecal e. coli colony count in broilers, rhode island red and haringhata black birds during diarrhoea (before the start of treatment) and after three days of treatment with ceftriaxone-tazobactam combination at 28.125 mg kg-1 bid intramuscularly for 3 days. [n=6] groups during diarrhoea (c.f.u. gm-1) after treatment (c.f.u. gm-1) gr be 25 × 109 19 × 107 gr re 31 × 109 25 × 106 gr he 45 × 109 23 × 105 http://dx.doi.org/10.5599/admet.1170 t.k. sar et al. admet & dmpk 10(3) (2022) 180-196 192 (a) (b) (c) figure 5. assessment of anti-e. coli antibody of sensitized (a) broiler birds, (b) rhode island red birds and (c) haringhata black birds. the kinetic behaviour of both ceftriaxone and tazobactam followed “two compartments open model” in both healthy and esbl e. coli infected broiler, rhode island red and haringhata black birds, following single intramuscular administration of ceftriaxone-tazobactam (8:1) combination. ceftriaxone achieved peak plasma concentration at 0.08 h and persisted up to 6 h in a healthy broiler, rhode island red and haringhata black birds. significantly higher values of pharmacokinetic parameters like t½ β and kel corresponded with longer persistence of ceftriaxone in diarrhoeic birds up to 8 h compared to healthy birds. the plasma concentrations of ceftriaxone and tazobactam maintained an approximate ratio of 8:1 for an appreciable period (2 h) in broiler birds but for a much shorter period in rhode island red (0.25 h) and haringhata black (0.75 h) birds post-dosing. disposition study showed that ceftriaxone achieved a plasma concentration of 1.51 ± 0.15 µg ml-1, 1.84 ± 0.14 µg ml-1 and 1.64 ± 0.18 µg ml-1 in esbl producing e. coli (tem-1) induced diarrhoeic broiler, rhode island red and haringhata black birds, respectively at 0.04 h which suggested that ceftriaxone concentration reached above the mic value (0.12 µg ml-1) within 0.04 h. moreover, the mean plasma concentrations of ceftriaxone were above the mic value for more than 11 h (t > mic) in all the diarrhoeic birds. gavin et al. reported that the treatment was successful in more than 90 % of patients with non-urinary isolates when % t > mic exceeded 40 % against esbl producing e. coli [28] which was in corroboration with our findings though higher percentages of t > mic (> 90 %) were observed. considering these results, it was decided to administer the ceftriaxone-tazobactam combination (8:1) at 28.125 mg kg-1 at 12 h interval for the efficacy study. in the efficacy study, we selected b.i.d. dosing to maintain the plasma concentration of ceftriaxone within the range of 5 – 12 µg/ml (mic 0.12 µg ml-1). in the ceftriaxone-tazobactam (8:1) combination, ceftriaxone is the antibacterial component and therefore, the mic of only ceftriaxone was considered for the efficacy study. the tazobactam component of the used combination is a beta-lactamase inhibitor which is usually incorporated to prevent the destruction of ceftriaxone by beta-lactamases produced by microorganisms like esbl e. coli and thereby maintaining the mic of ceftriaxone above the target mic. the most suitable ratio of ceftriaxone and tazobactam for effective inhibition of beta-lactamases and to increase the efficacy of ceftriaxone was reported to be 8:1 [29] and the ceftriaxone-tazobactam 8:1 combination is available commercially for effective treatment of susceptible infections. during experimental induction of infection with oral inoculation of esbl producing e. coli sub-culture (56 × 108 c.f.u. ml-1), broilers birds showed severe diarrhea with higher frequency; the major clinical sign of esbl producing e. coli infection on 7th day whereas moderate diarrhoea was manifested in rhode island red birds from the same day. these results were correlated with the lower antibody response of broilers and rhode island red during the 5th day (during the infection period) of postinoculation. interestingly, ceftriaxone-tazobactam combination treatment at 28.125 mg kg-1 body weight intramuscularly twice daily for three days exhibited increased specific antibody titre on the 5th and 10th day of treatment in broiler and rhode island red birds, respectively. ceftriaxone, being a beta-lactam antibiotic, admet & dmpk 10(3) (2022) 180-196 ceftraxone-tazobactam combination in poultry diarrhoea doi: http://dx.doi.org/10.5599/admet.1170 193 might cause increased specific antibody titre in the present study in broiler, rhode island red birds. the role of ceftriaxone in immune stimulation by increasing antibody titre against esbl e. coli was not reported earlier and is warranted for further study. previous studies suggested that the presence of antipneumococcal antibodies led to therapeutic efficacy with sub-inhibitory concentrations of β-lactam antibiotics [30] and, phagocytosis mediated by human and mouse neutrophils was increased when antibiotic-resistant pneumococcal strains were incubated with serum containing specific antibodies and sub-mic concentrations of β-lactams [31]. interestingly, haringhata black birds did not show any clinical sign following the initial challenge of esbl producing e. coli (tem-1) subculture. this result was also correlated with the higher antibody response of haringhata black birds during the 15th day of the first inoculation. this finding strongly supports the higher resistance of haringhata black birds, which is an indigenous poultry breed of west bengal available only at the northern part of north 24 parganas and the southern part of the nadia districts of west bengal, india. diarrhoea began to subside on 2nd day of treatment, and a complete recovery was noticed on 3rd day of treatment with ceftriaxone-tazobactam (8:1) combination in all the birds. the treatment was discontinued after 3 days of treatment due to complete recovery from diarrhoea in all the birds of three groups and no other clinical signs, including diarrhoea reappeared during the observation period. the feacal e. coli count reduced significantly in all the birds following treatment with the ceftriaxone-tazobactam combination (8:1). the feacal e. coli count results were correlated with the lower antibody response of broilers and rhode island red during the 5th day of post-inoculation. conclusions the ceftriaxone-tazobactam (8:1) combination showed favourable pharmacokinetics to use this combination as an effective treatment to combat particularly esbl producing e. coli infection in poultry birds by administering at 28.125 mg kg-1 body weight intramuscularly twice daily for three days. abbreviations a: zero time plasma concentration intercept (distribution phase); auc0-inf: total area under the plasma concentration versus time curve; b: zero time plasma concentration intercept (elimination phase); c.f.u.: colony forming units; clb: total body clearance of the drug; clsi: clinical and laboratory standards institute; cmax_calc: maximum plasma concentration calculated from the fitted data; cpcsea: committee for the purpose of control and supervision of experiments on animals; edta: ethylenediaminetetraacetic acid; esbl: extended spectrum β lactamase; f: bioavailability; fc: fraction of drug in the body that is contained in the central compartment; hplc: high performance liquid chromatography; iaec: institutional animal ethical committee; k12: first order rate constant for transfer of drug from central compartment to peripheral compartment; k21: first order rate constant for transfer of drug from peripheral compartment to central compartment; ka: rate of distribution; kel: first order rate constant for drug elimination from central compartment; mic: minimum inhibitory concentration; mpc: mutation prevention concentration; mrt: mean residence time; pd: post dosing; pda: photo diode array; ppm: parts per million; pmsf: phenylmethylsulfonyl fluoride; rpm: revolutions per minute; t > mic: time above mic; t~p: tissue to plasma ratio; tmax_calc: maximum plasma concentration time calculated from the fitted data t½ ka: biological half-life (distribution phase); t½ β: biological half-life (elimination phase); tsa: trypticase soy agar; tsb: trypticase soy broth; usp: united states pharmacopoeia; vdarea: apparent volume of distribution (area method); vdc: apparent volume of distribution in central compartment; vdss: steady state volume of distribution; β: zero time plasma concentration intercept (elimination phase). http://dx.doi.org/10.5599/admet.1170 t.k. sar et al. admet & dmpk 10(3) (2022) 180-196 194 acknowledgements: we are grateful to professor anjan bhattacharyya, exprofessor, department of agricultural chemicals, bidhan chandra krishi vishwavidyalaya, mohanpur, nadia, west bengal, india for his kind help to standardize the hplc analytical methods. conflict of interest: all the authors declare no conflict of interest. references [1] department of animal husbandry, dairying & fisheries, ministry of agriculture & farmers welfare, government of india. national action plan for egg & poultry2022 for doubling farmers’ income by 2022. https://www.dahd.nic.in/sites/default/filess/seeking%20comments%20on%20national%20action%2 0plan-%20poultry-%202022%20by%2012-12-2017.pdf. [2] h.j. barnes, w.b. gross. colibacillosis in diseases of poultry, b.k. calnek, h.j. barnes, c.w. beard, eds., iowa state university press, mosby-wolfe, (1997), p. 138–144. [3] j.k. rosenberger, p.a. fries, s.s. cloud. in vitro and in vivo characterization of avian escherichia coli. iii. immunization. avian dis. 29 (1985) 1108–1117. [4] d. melamed, g. letner, e.d. heller. a vaccine against avian colibacillosis based on ultrasonic inactivation of escherichia coli. avian dis. 35 (1991) 17–22. [5] a.s. dhillon, o.k. jack. two outbreaks of colibacillosis in commercial caged layers. avian dis. 40 (1996) 742–746. [6] m. blanco, j.e. blanco, a. mora, j. blanco. escherichia coli septicemicos aviares: serotipos, factores de virulencia, resistencia a antibiotics y desarrollo de vacunas. medicina veterinaria 13 (1996) 525–537. [7] j.m. casellas. south america: a different continent, different esbls. j. antimicrob. chemother. 44(a) (1999) 16. [8] l. poirel, i. lethomas, t. naas, a. karim, p. nordmann. biochemical sequence analyses of ges-1, a novel class a extended-spectrum betalactamase, and the class 1 integron in52 from klebsiella pneumoniae. antimicrob. agents chemother. 44 (2000) 622–632. https://doi.org/10.1128/aac.44.3.622-632.2000. [9] i. samanta, s.n. joardar, p.h. das, p. das, t.k. sar, t.k. dutta, s. bandyopadhyay, s. batabyal, d.p. isore. virulence repertoire, characterization, and antibiotic resistance pattern analysis of escherichia coli isolated from backyard layers and their environment in india. avian dis. 58(1) (2014) 39-45. https://doi.org/10.1637/10586-052913-reg.1. [10] m. chowdhury, r. bardhan, s. pal, a. banerjee, k. batabyal, s.n. joardar, g.p. mandal, s. bandyopadhyay, t.k. dutta, t.k. sar, i. samanta. comparative occurrence of esbl/ampc betalactamase-producing escherichia coli and salmonella in contract farm and backyard broilers. lett. appl. microbiol. 74 (2021) 53-62. https://doi.org/10.1111/lam.13581. [11] a. hussain, s. shaik, a. ranjan, n. nandanwar, s.k. tiwari, m. majid, r. baddam, i.a. qureshi, t. semmler, l.h. wieler, m.a. islam, d. chakravortty, n. ahmed. risk of transmission of antimicrobial resistant escherichia coli from commercial broiler and free-range retail chicken in india. front. microbiol. 8 (2017) 2120. https://doi.org/10.3389/fmicb.2017.02120. [12] h.c. neu, n.j. meropolol, k.p. fu. antibacterial activity of ceftriaxone (ro 13-9904) and a β-lactamase stable cephalosporin. antimicrob. agents chemother. 19 (1981) 414–423. https://doi.org/10.1128/aac.19.3.414. [13] t. li, g.l. qiao, g.z. hu, f.d. meng, y.s. qiu, x.y. zhang, w.x. guo, h.l. yie, s.f. li, s.y. li. comparative plasma and tissue pharmacokinetics and drug residue profiles of different chemotherapeutants in fowls and rabbits. j. vet. pharmacol. ther. 18(4) (1995) 260-273. https://doi.org/10.1111/j.13652885.1995.tb00590.x. [14] p. kumar, k.p. singh, v. ahuja, a.h. ahmed. pharmacokinetics of ceftriaxone following single dose i.v. and i.m. administration in layer birds. journal of veterinary pharmacology and toxicology 8(1-2) (2010) 10-12. https://www.dahd.nic.in/sites/default/filess/seeking%20comments%20on%20national%20action%20plan-%20poultry-%202022%20by%2012-12-2017.pdf https://www.dahd.nic.in/sites/default/filess/seeking%20comments%20on%20national%20action%20plan-%20poultry-%202022%20by%2012-12-2017.pdf https://doi.org/10.1128/aac.44.3.622-632.2000 https://doi.org/10.1128/aac.44.3.622-632.2000 https://doi.org/10.1637/10586-052913-reg.1 https://doi.org/10.1111/lam.13581 https://doi.org/10.3389/fmicb.2017.02120 https://doi.org/10.1128/aac.19.3.414 https://doi.org/10.1111/j.1365-2885.1995.tb00590.x https://doi.org/10.1111/j.1365-2885.1995.tb00590.x admet & dmpk 10(3) (2022) 180-196 ceftraxone-tazobactam combination in poultry diarrhoea doi: http://dx.doi.org/10.5599/admet.1170 195 [15] a.m. queenan, b. foleno, c. gownley. effects of inoculum and β-lactamase activity in ampcand extended-spectrum β-lactamase (esbl)-producing escherichia coli and klebsiella pneumoniae clinical isolates tested by using nccls esbl methodology. j. clin. microbiol. 42 (2004) 269–275. https://doi.org/10.1128/jcm.42.1.269-275.2004. [16] a. payasi, s. kumar, m.a. chaudhary. comparative study of sulbactomax versus ceftriaxone and betalactamase inhibitor and their effect on mutant prevention in esbl producing organisms. international journal of drug development and research 3(3) (2011) 366-371. [17] d.j. payne, r. cramp, d.j. winstanley, d.j knowles. comparative activities of clavulanic acid, sulbactam, and tazobactam against clinically important β-lactamases. antimicrob. agents chemother. 38(4) (1994) 767–772. https://doi.org/10.1128/aac.38.4.767. [18] u.c. mithin, r. buragohain, t.k. sar, t.k. mandal. disposition kinetics of ceftriaxone and tazobactam in broiler, rhode island red and haringhata black poultry following single dose intramuscular administration. international journal of modern pharmaceutical research 4(6) (2020). [19] u.c. mithin, r. buragohain, p.k. das, t.k. mandal, r.n. hansda, s.n. joardar, i. samanta, t.k. sar. monitoring of liver markers in poultry during post inoculation of extended spectrum β lactamases producing escherichia coli and after treatment by ceftriaxone-tazobactam combination. indian j. anim. hlth. 59(2) (2020) 237-242. https://doi.org/10.36062/ijah.59.2spl.2020.237-242. [20] clinical and laboratory standards institute. performance standards for antimicrobial susceptibility testing: twenty-third informational supplement m100-s23. clsi, wayne, pa, usa, 2013. [21] t.k. sar, p.h. patra, j.r. dash, t.k. mandal. pharmacokinetic interaction of intramammary ceftriaxone and oral polyherbal drug (fibrosin) in goats. drug metabol. drug interact. 24(4) (2011) 191-196. https://doi.org/10.1515/dmdi.2011.019. [22] a.p. ocampo, k.d. hoyt, n. wadgaonkar, a.h. carver, c.v. puglisi. determination of tazobactam and piperacillin in human plasma, serum, bile and urine by gradient elution reversed-phase hplc. j. chromatogr. 496 (1089) 167-179. https://doi.org/10.1016/s0378-4347(00)82563-3. [23] s.r. tamboli, d.d. patil. rp-hplc method for simultaneous estimation of cefepime hydrochloride and tazobactam sodium in bulk and pharmaceuticals. j. chem. 208057 (2013) 6 pages. https://doi.org/10.1155/2013/208057. [24] j.d. baggot. the basis of veterinary clinical pharmacology, w.b. saunders co., philadelphia, london, 1977. [25] p.j. quinn, m.e. carter, b.k. markey, g.r. carter. clinical veterinary microbiology, wolf publishing, london, uk, 1994. [26] a.p.a. mockett, j.k.a. cook, m.b. huggins. maternally derived antibody to infectious bronchitis virus: its detection in chick trachea and serum and its role in protection. avian pathol. 16 (1987) 407–416. https://doi.org/10.1080/03079458708436391. [27] k.h. choi, s.k. maheswaran, l.j. felice. characterization of outer membrane protein enriched extracts from pasteurella multocida isolated from turkeys. am. j. vet. res. 50 (1989) 676-683. [28] p.j. gavin, m.t. suseno, r.b. thomson, j.m. gaydos, c.l. pierson, d.c. halstead, j. aslanzadeh, s. brecher, c. rotstein, s.e. brossette, l.r. peterson. clinical correlation of the clsi susceptibility breakpoint for piperacillin-tazobactam against extended-spectrum-β-lactamase-producing escherichia coli and klebsiella species. antimicrob. agents chemother. 50(6) (2006) 2244–2247. https://doi.org/10.1128/aac.00381-05. [29] a. georgopoulos, a. buxbaum, w. graninger. efficacy of β-lactam and inhibitor combinations in a diffusion chamber model in rabbits. j. antimicrob. chemother. 43(4) (1999) 497–501. https://doi.org/10.1093/jac/43.4.497. [30] j. casal, l. aguilar, i. jado, j. yuste, m.j. giménez, j. prieto, a. fenoll. effects of specific antibodies against streptococcus pneumoniae on pharmacodynamic parameters of beta-lactams in a mouse sepsis model. antimicrob. agents chemother. 46 (2002) 1340–1344. https://doi.org/10.1128/aac.46.5.1340-1344.2002. http://dx.doi.org/10.5599/admet.1170 https://doi.org/10.1128/jcm.42.1.269-275.2004 https://doi.org/10.1128/aac.38.4.767 https://doi.org/10.36062/ijah.59.2spl.2020.237-242 https://doi.org/10.1515/dmdi.2011.019 https://doi.org/10.1016/s0378-4347(00)82563-3 https://doi.org/10.1155/2013/208057 https://doi.org/10.1155/2013/208057 https://doi.org/10.1080/03079458708436391 https://doi.org/10.1128/aac.00381-05 https://doi.org/10.1093/jac/43.4.497 https://doi.org/10.1128/aac.46.5.1340-1344.2002 https://doi.org/10.1128/aac.46.5.1340-1344.2002 t.k. sar et al. admet & dmpk 10(3) (2022) 180-196 196 [31] f. cafini, j. yuste, m.j. giménez, d. sevillano, l. aguilar, l. alou, e. ramos-sevillano, m. torrico, n. gonzález, e. garcía, p. coronel, j. prieto. enhanced in vivo activity of cefditoren in pre-immunized mice against penicillin-resistant s. pneumoniae (serotypes 6b, 19f and 23f) in a sepsis model. plos one. (2010) 5:e12041. https://doi.org/10.1371/journal.pone.0012041. ©2022 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) https://doi.org/10.1371/journal.pone.0012041 http://creativecommons.org/licenses/by/3.0/ the optimization of electrochemical immunosensors to detect epithelial sodium channel as a biomarker of hypertension doi: http://dx.doi.org/10.5599/admet.1629 211 admet & dmpk 11(2) (2023) 211-226; doi: https://doi.org/10.5599/admet.1629 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper the optimization of electrochemical immunosensors to detect epithelial sodium channel as a biomarker of hypertension tias f.h. lestari1, riyanto setiyono1, nina tristina2, yulia sofiatin3 and yeni wahyuni hartati1* 1department of chemistry, faculty of mathematics and natural sciences, universitas padjadjaran, indonesia 2department of clinical pathology, faculty of medicine, universitas padjadjaran, indonesia 3department of public health, faculty of medicine, universitas padjadjaran, indonesia *corresponding author: e-mail: yeni.w.hartati@unpad.ac.id; tel.: +628122132349 received: december 04, 2022; revised: february 13, 2023; published: february 17, 2023 abstract the epithelial sodium channel (enac) is a transmembrane protein that regulates the balance of sodium salt levels in the body through its expression in various tissues. the increase in sodium salt in the body is related to the expression of enac, thereby increasing blood pressure. therefore, overexpression of the enac protein can be used as a biomarker for hypertension. the detection of enac protein using anti-enac in the biosensor system has been optimized with the box-behnken experimental design. the steps carried out in this research are screen-printed carbon electrode modification with gold nanoparticles, then anti-enac was immobilized using cysteamine and glutaraldehyde. optimum conditions of the experiment, such as anti-enac concentration, glutaraldehyde incubation time, and anti-enac incubation time, were optimized using the box-behnken experimental design to determine the factors that influence the increase in immunosensor current response and the optimum conditions obtained were then applied to variations in enac protein concentrations. the optimum experimental conditions for anti-enac concentration were 2.5 µg/ml, the glutaraldehyde incubation time was 30 minutes, and the anti-enac incubation time was 90 minutes. the developed electrochemical immunosensor has a detection limit of 0.0372 ng/ml and a quantification limit of 0.124 ng/ml for the enac protein concentration range of 0.09375 to 1.0 ng/ml. thus, the immunosensor generated from this study can be used to measure the concentration of normal urine samples and those of patients with hypertension. ©2023 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords box-behnken design; screen-printed carbon electrode; immunosensor; hypertension introduction hypertension is the number one cause of death in the world and in southeast asia, where indonesia is one of the countries in the region [1]. the world health organization [2] explains that as many as 1.13 billion people worldwide suffer from hypertension. the number of people suffering from hypertension continues to increase every year. based on diagnostic data in the basic health research [3] conducted by doctors, it was found that the prevalence of hypertension in indonesia in adolescents aged 18 years was 34.1%, while in the 31-44 year age group, it was 31.6 %, in the age group 45-54 years 45.3 %, and 55-64 years 55.2 %. in addition, there are other tools used for measuring blood pressure, namely the sphygmomanometer, but in its use, it http://dx.doi.org/10.5599/admet.1629 https://doi.org/10.5599/admet.1629 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:yeni.w.hartati@unpad.ac.id http://creativecommons.org/licenses/by/4.0/ t. f.h. lestari et al. admet & dmpk 11(2) (2023) 211-226 212 requires supporting tests [4]. hypertension can be detected through biomarkers (biological markers) in the human body. salt-sensitive hypertension is a type of hypertension that can occur due to increased blood pressure driven by excess salt intake or salt load and due to abnormal sodium excretion function in the kidneys [5,6]. the epithelial sodium channel (enac) is widely expressed in tissues such as the kidney, colon, lung, intestine, sweat and salivary glands [7]. the kidneys play a role in balancing the sodium content in the body. in the kidney, precisely in the distal nephron, the total sodium (na+) balance in the body is regulated through enac. unbalanced sodium levels in the distal nephron can increase extracellular fluid volume, which causes an increase in blood pressure [8,9]. enac is located in apically polarized cell membranes that function as mediators of na+ transport across the renal epithelium [9]. urine is a substance excreted by the kidneys, so enac protein can be detected using a urine sample. sofiatin and rusli [10] conducted a study to detect enac protein in human urine samples using the enzymelinked immunosorbent assay (elisa) method. the results of the study showed enac protein levels in the urine were normal, hypertension without a history, and with a history of 1.12, 4.0 and 2.7 ng/ml. however, the elisa method has several drawbacks, including requiring trained medical personnel, a long processing time, and complicated procedures, so other methods are needed for enac protein in human urine. there is another method that can be used to detect enac protein, namely electrochemical immunosensor [11], including the research of hartati et al. [12] electrochemical immunosensor study based on gold nanoparticles-anti-enac bioconjugate, voltammetric immunosensor using a screen-printed carbon electrode modified with reduced graphene oxide [13], and electrochemical immunosensor study based on a modified screen-printed carbon electrode with gold nanoparticles-mpa-edc-nhs [14]. in addition, the enac protein detection method can also be carried out with an electrochemical aptasensor [14], who used ceria electrodeposition and fazrin et al. [15] study, who used a screenprinted carbon electrode modified with ceo2/aunp-streptavidin-biotinylated aptamer. there is also research using an optical-based biosensor for the detection of enac protein by garcia rubio et al. [16]. biosensor is an analytical tool that has been widely developed, which is a tool that combines the use of biological elements such as enzymes, cells, and antibodies with a transducer to convert biological interactions into a readable output signal [17–21]. biosensors are useful in the fields of medicine, environmental diagnostics and the food and beverage industry [20]. electrochemical immunosensor is a type of biosensor that can detect an analyte based on the interaction between antibody-antigen through electrochemical transduction [22]. screen-printed carbon electrode (spce) is an electrochemical device that is widely used in sensor applications [23-25] because it has several advantages, including low background current and a wide potential window, and does not require expensive costs [26]. the spce consists of three electrodes, namely a working electrode, a counter electrode and a reference electrode located on the same platform and can be made simultaneously with the screen printing technique [27]. increasing the signal sensitivity of an immunosensor device can be done by modification of nanoparticles, such as gold nanoparticles [22]. gold nanoparticle-modified spce is widely used because of its great biocompatibility [28] so biomolecules can be immobilized stably. in addition, gold nanoparticles can also increase the electroactive area of the electrode, thereby providing many active areas for protein binding and electron transfer, and are better at maintaining protein bioactivity [29]. one of the biomolecular immobilization techniques that can be done is a self-assembled monolayer [30]. a self-assembled monolayer is an ordered molecular assembly formed spontaneously by chemical adsorption and selforganization of long chains of molecules on a suitable substrate surface [31]. to identify the factors that influence the experimental results, to minimize uncontrollable factors and evaluate their influence, experiments were carried out using an experimental design [32]. response surface methodology (rsm) is a combination of statistical and mathematical techniques used in modeling and problem analysis to determine the variables in order to optimize the response [33]. one type of experimental admet & dmpk 11(2) (2023) 211-226 electrochemical immunosensor for enac doi: https://doi.org/10.5599/admet.1629 213 design from the response surface methodology is the box-behnken experimental design, a three-level design that matches the response surface [34]. the principle of this experimental design is based on the number of experiments carried out, the number of parameters, and the number of central points, as well as at three levels/levels for each factor, namely -1 (low), 0 (medium), +1 (high). the box-behnken design is more efficient because it only performs 15 experiments. the box-behnken experimental design is a three-level design method used to estimate the interactions that occur, carried out by maximizing and minimizing the response. the resulting designs are generally very efficient in terms of the number of rounds required [35]. to determine the optimal conditions in the box-behnken experimental design, a second-order polynomial model was used to fit the parameter and response relationship. the polynomial equations obtained are [36]: y = β0 + x1β1 + x2β2 + x3β3 + x1x2β1.2 + x1x3β1.3 + x2x3β2.3 + x1.2β1.2 (1) where y = response, xi = parameter that affects response (i = 1, 2, …, n) and βi = regression coefficient (i = 1, 2, …, n). in this study, an electrochemical immunosensor method was developed for the detection of enac protein using au-modified spce, cysteamine, glutaraldehyde, and anti-enac. on the surface of spce/au, cysteamine is added to form a self-assembled monolayer via a thiol group (-sh). glutaraldehyde has two terminal aldehyde groups (-cho) which covalently bind to the free amine group (-nh2) of cysteamine. to increase the selectivity of the immunosensor, anti-enac was added to spce modified by gold nanoparticles, cysteamine, and glutaraldehyde. then the box-behnken experimental design was used to optimize several factors that affect this electrochemical immunosensor experiment, including the concentration of anti-enac, glutaraldehyde incubation time, and anti-enac incubation time. optimum conditions of these factors are required to obtain as much attachment of the bioreceptors to the modified electrode as possible. the detection and quantification limits were determined. experimental tools the equipment used in this study were autoclave sterilizer (hirayama autoclave hve-50), hot plate (ika c-mag hs 7), magnetic stirrer (eppendorf), micro pipette (eppendorf), potentiostat zimmer and peacock connected to a computer using pstrace 5.8 software, centrifugator (eppendorf), uv-vis spectrophotometer (thermo scientific), scanning electron microscope (hitachi tm 3000, japan), homemade screen-printed carbon electrode (spce) produced at the department of chemistry, padjadjaran university, micro tube (eppendorf), and micro pipette tips, as well as all glass equipment in the laboratory of the department of chemistry, padjadjaran university. materials the materials used in this study were demineralized water (pt ikapharmindo putramas indonesia), antienac antibody (santa cruz biotechnology), hydrochloric acid (merck; p.a), chloroauric acid (sigma aldrich), ethanolamine (merck.pa), glutaraldehyde (sigma aldrich), potassium ferricyanide (sigma aldrich), potassium chloride (merck.pa), sodium hydroxide (merck.pa), phosphate buffered saline ph 7.4 (merck p.a.), enac protein (santa cruz biotechnology), cysteamine (sigma aldrich). preparation of 1 % trisodium citrate dihydrate solution: 0,1 g of trisodium citrate dihydrate (na2c6h5o7.2h2o) was dissolved with 10 ml of demineralized water, then homogenized. preparation of 0.75 mm 30 ml gold nanoparticle (aunp) solution from chloroauric acid: 268.53 µl of haucl4×3h2o solution was added with 19.731.4 µl of demineralized water, then stirred using a magnetic https://doi.org/10.5599/admet.1629 t. f.h. lestari et al. admet & dmpk 11(2) (2023) 211-226 214 stirrer and heated using a hot plate until boiling. then 1730 µl of 1 % na2c6h5o7.2h2o was added and stirred until it changed color (yellow-colorless-purple-burgundy). preparation of phosphate buffer saline (pbs) solution ph 7.4: 1 pbs tablet was dissolved in 100 ml of demineralized water and stirred until homogeneous. then the ph of the solution was adjusted by adding 0.1 m naoh or hcl solution and measured using a ph meter to show a ph of 7.4. preparation of 0.01 m k3[fe(cn)6] solution in 0.1 m kcl: k3[fe(cn)6] was weighed as much as 0.8233 grams, then kcl was weighed as much as 1.8639 grams. the solid potassium ferricyanide and potassium chloride were dissolved in 250 ml of demineralized water. preparation of 0.1 m cysteamine solution: the cysteamine solid was weighed as much as 0.7715 g. then put into a 100 ml volumetric flask and dissolved with demineralized water up to the mark and shaken until homogeneous. preparation of 2.5 % glutaraldehyde solution: 5 ml of 50 % glutaraldehyde dissolved in phosphate buffer saline ph 7.4 to 100 ml. preparation of 50 ml 1 m ethanolamine solution: 3.08 ml of 98 % ethanolamine solution was diluted with demineralized water in a 50 ml volumetric flask, and then the solution was shaken until homogeneous. screen-printed carbon electrode (spce) modification with gold particles (spce-au) and spce-au characterization spce electrode modification is done through the spray coating technique. prior to the spray coating technique, the spce electrode was activated by irradiating it with ultraviolet light for 15 minutes. after the spce electrode was activated, a spray coating technique was applied using a solution of gold nanoparticles (aunp) 10 times sprayed and dried. then rinsed using 40 µl of demineralized water and dried at room temperature. then, spce that had been modified by gold nanoparticles was characterized by differential pulse voltammetry using a 0.01 m k3[fe(cn)6] redox system in 40 µl 0.1 m kcl over a potential range of -1.0 to +1.0 at a scan rate of 0.008 v/s for 3-4 minutes. spce surfaces before and after modified gold nanoparticles were observed by sem (scanning electron microscope) and characterized using electrochemical impedance spectroscopy (eis). anti-enac immobilization on electrode surface and characterization the spce-au surface was rinsed with demineralized water and dried at room temperature before adding 40 µl of 0.1 m cysteamine solution and incubating at room temperature for 2 hours in the dark, followed by rinsing with demineralized water [13]. next, the sensor was covered with 2.5 % glutaraldehyde solution for 30 minutes to form a cross-linking monolayer on the sam-modified au electrode and washed thoroughly with demineralized water to remove unreacted glutaraldehyde [37]. next, the electrodes were dripped with 25 l of 2.5 µg/ml anti-enac and incubated for 90 minutes at 25 °c to obtain spce-au/cysteamine/glutaraldehyde/anti-enac. the resulting immunosensor was characterized by differential pulse voltammetry using a 0.01 m k3[fe(cn)6] redox system in 40 µl 0.1 m kcl in the potential range of -1.0 to +1.0 at the scanning rate. 0.008 v/s for 3-4 minutes to determine whether cysteamine, glutaraldehyde, and anti-enac were immobilized at the electrode. then it was characterized using sem and eis [13]. determination of immunosensory response to enac electrodes immobilized with anti-enac were incubated in 30 µl enac with a certain concentration for 30 minutes at 25 °c. furthermore, at each concentration of enac, the current response was measured using differential pulse voltammetry to obtain spce-au/cysteamine/glutaraldehyde/anti-enac/enac electrodes admet & dmpk 11(2) (2023) 211-226 electrochemical immunosensor for enac doi: https://doi.org/10.5599/admet.1629 215 with a redox system of 0.01 m k3[fe(cn)6] solution in kcl 0 .1 m over a potential range of -1.0 to +1.0 v at a scan rate of 0.008 v/s for 3-4 minutes [14]. optimization of parameters affecting experiments parameter optimization was carried out using the box-behnken method with the concentration of antienac (x1), incubation time of glutaraldehyde (x2), and incubation time of anti-enac (x3), optimized through three different levels as shown in table 1. lowest (-1), medium (0), and highest (+1) levels. each measurement was processed and the determination of the optimal value for each factor was carried out using the boxbehnken experimental design on the minitab19 program. table 1. optimization of the factors that affect the experimental conditions with the box-behnken design factor level -1 0 +1 anti-enac concentration, µg/ml 0,5 1,5 2,5 glutaraldehyde incubation time, min 30 60 90 anti-enac incubation time, min 30 60 90 calibration curve variations in the concentration of the enac solution (0, 0.09375, 0.1875, 0.375, 0.75, 1 and 1.5 ng/ml) were measured using an electrochemical immunosensor (differential pulse voltammetry with a redox system of 0.01 m k3[fe(cn)6] in 0.1 m kcl in the potential range from -1.0 to +1.0 v at a scan rate of 0.008 v/s for 34 min) using the optimal conditions of the box-behnken method. next, a curve is made between the concentration and the difference in the average peak current (δi) for each resulting measurement as x and y, so that the equation y = a + bx is obtained [32]. results and discussion preparation and characterization of gold nanoparticles (aunp) gold nanoparticles (aunp) are currently widely used in immunosensor fabrication because of their high conductivity, good compatibility with biomolecules, and can increase the electroactive area of the electrode [29]. in addition, gold nanoparticles function as an electron-conducting pathway between the prosthetic group and the electrode surface and act as a catalyst for electrochemical reactions [38]. colloidal gold nanoparticles (aunp) were prepared with 0.75 mm chloroauric acid (haucl4.3h2o) solution, which was heated to boiling, then mixed with 1 % sodium citrate (na3c6h5o7.2h2o) solution and vigorous stirring until the solution color changed from colorless to burgundy, as in figure 1a. figure 1. (a) burgundy gold nanoparticle colloid. (b) the results of aunp characterization using visible light spectrophotometer. the maximum absorption wavelength is at 521 nm. chlorouric acid solution acts as the main precursor in the synthesis of gold nanoparticles. chlorouric acid solution has gold in au3+ oxidation state, while gold nanoparticles have au0 oxidation state [39]. trisodium citrate solution acts as a capping agent to maintain the stability of colloidal gold nanoparticles so that they have excellent https://doi.org/10.5599/admet.1629 t. f.h. lestari et al. admet & dmpk 11(2) (2023) 211-226 216 stability. citrate can prevent aggregation of the formed gold nanoparticles [40,41]. the citrate ion has a hydroxyl (-oh) and a carboxylate (coo-) functional groups. the hydroxyl functional group on the citrate ion will interact with the reduced gold so that the gold particles will be surrounded by citrate ions (negatively charged) and prevent aggregation of the nanoparticles through the repulsion between the negative charges on the surface. the colloidal nanoparticles formed were characterized using a visible light spectrophotometer. the characterization results in figure 1b show that the maximum absorption of gold nanoparticles is at a wavelength of 521 nm. gold nanoparticles generally have a strong plasmon resonance absorption band at a wavelength of 510-550 nm with a spherical shape [42,43]. when gold particles are nano-sized, they can appear red, purple or blue. the color difference of the resulting gold nanoparticles comes from the surface plasmon resonance (spr) phenomenon. when light shines on a metal surface, a surface plasmon occurs, a group of electrons that move back and forth simultaneously across the metal surface. when electrons travel at the same frequency as light, plasmons are in resonance. when plasmons are in resonance, electrons absorb and scatter light, producing the colors seen (complementary colors). gold nanoparticles resonate at frequencies in the visible light spectrum. the smaller gold nanoparticles absorb and resonate with the blue-green light wavelengths of the spectrum (~450 nm), resulting in a red color, while the larger gold nanoparticles absorb and resonate with green-red light wavelengths, resulting in blue color. with increasing particle size, the spr absorption wavelength shifts to the infrared region of the spectrum and the color changes according to the wavelength [41]. based on the spr theory, small nanoparticles have a large band gap so that the energy required to excite electrons out of the band gap is very large. the energy is inversely proportional to the wavelength, so that a small absorption of visible light wavelengths indicates a small aunp size. gold nanoparticles have also been characterized using a particle size analyzer (psa) and dynamic light scattering (dls) techniques reported in [44] with an average gold nanoparticle diameter of 38.3 nm. spce-au modification screen-printed carbon electrode (spce) is an electrode widely used because of its simplicity in the process, low background current, requires a small number of samples and is easily modified with gold nanoparticles. spce modification with gold nanoparticles aims to increase the electroactive area of the electrode. the high conductivity of gold can increase the electroactive properties of the electrode, thus providing many active regions that play a role in protein binding and electron transfer. figure 2 shows the stages of spce modification using gold nanoparticles and biological elements with electrochemical characterization. in this study, the surface of the spce electrode was modified with gold nanoparticles (spce-au) using a spray coating technique involving the formation of a fine aerosol through a nozzle [45]. the aerosol droplets will then hit the electrode surface and stick evenly to the electrode surface after being modified. prior to spray coating, spce was irradiated with ultraviolet light for 15 minutes to oxidize the carbon particles so that the carbon functional groups on the spce surface were active. activation of carbon shows that carbon from the relaxation energy level is excited to a higher energy level, then carbon becomes activated carbon. so that when gold nanoparticles are added, they will settle on the carbon surface. this pretreatment is also carried out to clean the electrode surface, change the microstructure and surface chemistry of the electrode, and generate new active sites that affect the sensitivity of the electrode to the target analyte [46]. spce before and after modification with gold nanoparticles was characterized using differential pulse voltammetry to determine whether spce modification with aunp was successful. this electrochemical measurement was carried out by observing the redox activity of the electroactive indicator, namely k3[fe(cn)6] 10 mm in 0.1 m kcl on the electrode surface. the k3[fe(cn)6] system is used because it has a sensitive electrochemical response to carbon-based surfaces [47]. the redox reaction that occurs is the reduction of ferricyanide ions to ferrocyanide, as shown in the following reaction [48]: fe(cn)64 ⇌ fe(cn)63-+ e (2) admet & dmpk 11(2) (2023) 211-226 electrochemical immunosensor for enac doi: https://doi.org/10.5599/admet.1629 217 figure 2. illustration of electrochemical immunosensor for detection of enac protein. in figure 3 it is shown that the modified spce electrode is more electroactive than the unmodified electrode. this is indicated by the increase in peak current to 77.298 µa after the electrode was modified with aunp (spce-au). the peak current before modification (spce bare) was 31.355 µa. the increasing peak current after modification indicates that the addition of aunp to spce by spray coating technique and irradiation with uv light can increase the electron transfer process between the electrode and the analyte so that the electrode becomes more sensitive [12,49]. figure 3. immunosensor response using differential pulse voltammetry: (a) spce bare; (b) spce-au; (c) spceau/cysteamine; (d) spce-au/cysteamine/glutraladehyde; (e) spce-au/cysteamine/glutaraldehyde/antienac; and (f) spce-au/cysteamine/glutaraldehyde/anti-enac/enac with a redox system of 10 mm k3[fe(cn)6] solution in 0.1 m kcl over a potential range of -1.0 to +1.0 v and scan rate 0.008 v/s. anti-enac immobilization on electrode surface spce-au was modified first using a self-assembled monolayer (sam). sam has the potential to create immobile protein layers in which some control over orientation is exerted. the formation of sam is carried out by forming a covalent coupling of cysteamine and glutaraldehyde as a functional amino cross-linking structure. sam can basically be used in two ways, namely, chemical modification of a protein with a linking structure that can react with a substrate to form sam and carrying out reactions on sam to link biomolecules [50]. spce-au was incubated in cysteamine solution for 120 minutes. cysteamine is an organic molecule that belongs to https://doi.org/10.5599/admet.1629 t. f.h. lestari et al. admet & dmpk 11(2) (2023) 211-226 218 alkanethiol, can oscillate spontaneously and is stable on a gold surface. in its structure, alkanethiol consists of three parts, namely the main functional group (amine), which binds with other functional groups to produce a chemical interaction, the sulfur (s) group, which interacts with metal (au), and the methylene group [12]. when a thiol-based solution is added to the spce-au electrode, the thiol group will be chemically adsorbed onto the gold surface through the formation of a thiolate bond. the attractive van der waals forces between the alkyl chains increase the stability and regularity of the sam so that a strong and stable s-au reaction will be formed. figure 4 shows van der waals dispersion forces mechanism between s-au [51]: figure 4. mechanism of van der waals dispersion forces between s-au [51]. the s-au reaction is a lewis acid reaction, which is a reaction based on the binding of free electrons owned by the sulfur atom (s) of cysteamine with au. the formation of s-au occurs through the coordination of covalent bonds. in the s-au reaction, as many as six free electrons owned by sulfur bind to au at the electrode surface. this reaction can be written as follows [52]: rs – h + aun0 → rs – au + aun0 + ½h2 (3) furthermore, spce-au/cysteamine, glutaraldehyde solution was added and incubated for 30 minutes. glutaraldehyde consists of two terminal aldehyde groups (-coh), which is a bifunctional reagent because it can react with two primary amine groups, namely from biomolecules and solid surfaces [53]. bifunctional reagents in binding biomolecules to solid media, is widely used because it facilitates the formation of a stable single layer. anti-enac was immobilized on spce-au/cysteamine/glutaraldehyde and incubated for 90 min. glutaraldehyde binds to the amino acid residue lysine (positively charged), then reacts covalently with the carbonyl group of glutaraldehyde to create a more stable structure. anti-enac plays an important role as a capture molecule to recognize and detect the presence of enac antigen in a sample. after anti-enac is immobilized on the electrode surface, ethanolamine is added as a blocking agent whose presence will replace the -coh groups on the electrode surface that are not bound to the terminal amine group of anti-enac. this is done to prevent non-specific binding to the surface of the spce electrode, which can interfere with the immunoassay process. good antibody orientation can increase the antigen-binding capacity to reduce the detection limit and increase the sensitivity to antigen. spce blocked with ethanolamine was then incubated with enac protein for 30 minutes at room temperature. after the enac protein was immobilized on the electrode surface, the electrode was rinsed using a pbs solution of ph 7.4 to remove unreacted or attached species to the electrode surface so that it did not interfere with the analysis process and the immunoreaction process between antibodies and antigens and minimized the measurement error of the immunosensor response. then, characterization was carried out using dpv. the presence of the enac protein bound to anti-enac causes the ferricyanide current peak to be lower. this is because the enac protein is a large biomolecule that can cause the electron transfer process to be more hindered. the measured current response will be proportional to the amount of enac protein involved. the higher the enac protein concentration, the lower the current generated, while the lower the enac protein concentration, the higher the current generated. the higher the concentration of enac protein, the more enac protein is attached to the electrode surface. admet & dmpk 11(2) (2023) 211-226 electrochemical immunosensor for enac doi: https://doi.org/10.5599/admet.1629 219 characterization of spce-au/cysteamine/glutaraldehyde/anti-enac/enac with differential pulse voltammetry and electrochemical impedance spectroscopy figure 4 shows the voltammogram of the modified spce characterization using differential pulse voltammetry. characterization was carried out to determine whether anti-enac had been successfully modified on the spce surface. after the addition of cysteamine on the electrode surface to form spce-au/cysteamine, the current decreased to 67.363 µa due to the formation of a strong and stable s-au bond on the electrode surface so that the electron transfer process between ferricyanide species and the electrode was hindered. furthermore, there was a decrease in current to 47.375 µa after the addition of glutaraldehyde, which indicated that glutaraldehyde had reacted with cysteamine on the electrode surface and blocked the electron transfer process. after the addition of anti-enac again, the current decreased to 31.886 µa, this is because the enac protein is a large biomolecule that can hinder the electron transfer process. spce-au characterization was also carried out by electrochemical impedance spectroscopy (eis). eis is a measurement based on impedance, so eis is inversely proportional to dpv whose measurement is based on the generated current. impedance data are usually represented by a nyquist plot with the dependence of real impedance (z') on imaginary impedance (z'') [54]. characterization of spce-au can be followed by eis the changes in charge transfer resistance (rct), resulting from the interaction of the analyte with the receptor [55]. rct can be evaluated from the semicircular portion of the nyquist plot at higher frequencies. the increasing current due to electron transfer, the smaller the impedance and the smaller the rct diameter [56]. in the results of spce characterization with dpv it is known that there is an increase in current due to the process of electron transfer between the electrode and the analyte. meanwhile, the results of the spce characterization using eis (figure 5) show that the larger rct diameter indicates a lower current resulting from electron transfer so that the resistance obtained is large. table 2. shows a comparison of the measurement results of each stage of immunosensor modification using dpv and eis. figure 5. spce characterization results using electrochemical impedance spectroscopy: (a) spce bare; (b) spce-au; (c) spce-au/cysteamine; (d) spce-au/cysteamine/glutaraldehyde; and (e) spceau/cysteamine/glutaraldehyde/anti-enac with a redox system of 10 mm k3[fe(cn)6] solution in 0.1 m kcl with a frequency of 0.1 hz to 1000000 hz at an anodic peak current potential of 0.01 v. table 2. comparison of the measurement results of the enac immunosensor using dpv and eis. spce modification stage current, μa resistance, ω spce bare 31.355 0.767 spce-au 77.298 0.290 spce-au/cysteamine 67.363 0.445 spce-au/cysteamine/glutaraldehyde 47.375 1.027 spce-au/cysteamine/glutaraldehyde/anti-enac 31.866 1.150 spce bare 31.355 0.767 https://doi.org/10.5599/admet.1629 t. f.h. lestari et al. admet & dmpk 11(2) (2023) 211-226 220 morphological characterization of enac immunosensors with a scanning electron microscope spce surfaces before and after being modified with gold nanoparticles were characterized using sem. figure 6 shows the sem images of (a) spce bare, (b) spce-au, and (c) spce-au/cysteamine/glutaraldehye/anti-enac. figure 6. the modified spce characterization using sem: (a) spce bare, (b) spce-au, and (c) spceau/cysteamine/glutaraldehyde/anti-enac. the results of spce-au characterization showed that after being modified with gold nanoparticles (spce-au), the surface of the spce electrode became more closed due to the attached gold nanoparticles. this indicates that the aunp has been evenly distributed on the entire surface of the electrode. meanwhile, the spceau/cysteamine/glutaraldehyde/anti-enac characterization showed that the anti-enac immobilization had been successfully carried out, where antibody molecules were seen covering the electrode surface. determination of optimum conditions with box-behnken experimental design the optimum conditions of the immunosensor were determined by carrying out a box-behnken experimental design on the minitab19 application with three factors optimized in the experiment, namely antienac concentration (x1), glutaraldehyde incubation time (x2) and anti-enac incubation time (x3). the effect of anti-enac concentration, glutaraldehyde incubation time, and anti-enac incubation time was tested using differential pulse voltammetry with a redox system of 10 mm k3[fe(cn)6] solution in 0.1 m kcl at a potential range of -1.0 to 1 0.0 v at a scan rate of 0.008 v/s, estep 0.004 v with an epulse of 0.025 v and a tpulse of 0.05 s. each factor is designed at three levels, namely the lowest level (-1), the middle level (0), and the highest level (+1), as shown in table 3. the relationship between the response and the factors is determined through a sequence of experiments to obtain the optimum response results. experiment with three factors and three different levels so as to give the number of trials 15 times. from the experimental current response obtained, the regression equation (2) is obtained as follows: y = 60.13 + 7.82 x1 0.534 x2 0.279 x34.126 x1.x1 + 0.004356 x2.x2 + 0.000550 x3.x3 0.0617 x1.x2 0.0465 x1.x3 + 0.003596 x2.x3 (4) from equation (4), can be seen that a factor with a negative value reduces the response, while a factor with a positive value increases the response of the experimental stream. based on the current response analysis, the anova results are shown in table 3 through the p-value, which explains the variability of the data. a p-value less than 0.05 indicates that a single variable exhibits a linear effect that is consistent with the linear model. the p-value is also used to determine the significance of each variable and the interaction effect shown by the combination of the two variables [44]. based on the results of the anova analysis shown in table 3, it is known that all the optimized factors, namely the concentration of anti-enac, incubation time of glutaraldehyde, incubation time of anti-enac significantly affected the decrease in the current response, because it had a p-value <0.05. in table 3, can be seen that the p-value of each factor is x1 = 0.000, x2 = 0.001, and x3 = 0.0012. (a) (b) (c) admet & dmpk 11(2) (2023) 211-226 electrochemical immunosensor for enac doi: https://doi.org/10.5599/admet.1629 221 table 3. analysis of variance. source df adj ss adj ms f-value p-value model 9 1291.72 143.524 64.99 0.000 linear 3 1097.84 365.945 165.71 0.000 x1 1 975.76 975.759 441.84 0.000 x2 1 89.82 89.820 40.67 0.001 x3 1 32.26 32.257 14.61 0.012 square 3 130.53 43.511 19.70 0.003 x1x1 1 62.85 62.854 28.46 0.003 x2x2 1 56.76 56.756 25.70 0.004 x3x3 1 0.90 0.903 0.41 0.551 2-way interaction 3 63.35 21.117 9.56 0.016 x1x2 1 13.69 13.686 6.20 0.055 x1x3 1 7.77 7.770 3.52 0.120 x2x3 1 41.89 41.893 18.97 0.007 error 5 11.04 2.208 lack-of-fit 3 9.63 3.211 4.56 0.185 pure error 2 1.41 0.704 total 14 1302.76 figure 7. response optimization of the optimal experimental conditional. response optimization needs to be done to identify setting variables and to optimize one or more responses. it is necessary to evaluate the number of responses to a variable [44]. based on the results in figure 7, the best optimization is close to the optimal state for the overall operating conditions of anti-enac concentration at 2.5 ng/ml, glutaraldehyde incubation time of 30 minutes, and anti-enac incubation time of 90 minutes. figure 8. contour plot of the experimental conditional. https://doi.org/10.5599/admet.1629 t. f.h. lestari et al. admet & dmpk 11(2) (2023) 211-226 222 determination of analytical parameters after the optimum conditions of the parameters are known, a calibration curve is made and the detection and quantification limits of the developed enac immunosensor are calculated. the calibration curve for the enac immunosensor was made using various concentrations of the enac antigen tested on the immunosensor. the variation of the enac antigen concentration was made to 0.09375; 0.1875; 0.375; 0.75; 1; and 1.5 ng/ml. the response of the enac immunosensor current obtained was plotted against the response in the form of a reduction current difference measured three times for each antigen concentration. the enac immunosensor was characterized using differential pulse voltammetry (dpv) with a k3[fe(cn)6] solution redox system at a potential range of -1.0 to +1.0 and a scan rate of 0.008 v/s. the voltammogram of the characterization results for each concentration of enac antigen is shown in figure 9(a). based on the results of the characterization, it can be seen that the higher the concentration of the enac antigen, the lower the current generated. the enac antigen is a large biomolecule that is not electroactive so when more enac antigen binds to the antibody on the electrode surface, the electron transfer process between the analyte and the electrode will be increasingly hindered. figure 9. (a) the results of differential pulse voltammetry measurements for various concentrations of enac (0.09375; 0.1875; 0.375; 0.75; 1; 1.5) ng/ml with a redox system of 10 mm k3[fe(cn)6] solution in kcl 0 .1 m over a potential range of -1.0 to +1.0 v, estep 0.004 v, epulse 0.025 v, tpulse0.05 s, and scan rate 0.008 v/s. (b) enac immunosensor calibration curve with variations in concentration (0.09375; 0.1875; 0.375; 0.75; 1) ng/ml. furthermore, the results of the characterization of the variation of the enac antigen concentration were plotted to create a calibration curve, as shown in figure 9 (b), so that the linear regression equation was obtained, y = 5.6786x + 2.1157 with r2 = 0.996. the detection limit is 0.0372 ng/ml and the quantification limit is 0.124 ng/ml. in this study, the lod obtained is lower than similar immunosensor studies as a method of detecting hypertension biomarkers, which can be seen in table 4. this shows that the electrochemical immunosensor method using screen-printed carbon electrode modified with gold nanoparticles and a selfassembled monolayer can be used for the detection of enac protein as a biomarker of hypertension. table 4. electrochemical biosensors studies to detect epithelial sodium channel (enac). methods sensing method linear range, ng/ml lod, ng/ml reference spce-au/mpa/edc-nhs dpv and eis 0,1 – 1,5 0,037 [14] spce/au/cysteamine/ bioconjugate dpv and eis 0,09375 – 1,0 0,084 [12] spce/ceo2/streptavidin/aptamer dpv and eis 0,05 – 3,0 0,012 [53] spce/rgo dpv and eis 0,09375 – 1,5 0,198 [13] spce/ceo2/bioconjugate dpv 0.047 – 3,0 0,110 [15] spce-au/cysteamine/glutaraldehyde dpv and eis 0,09375 – 1,0 0,0372 this work conclusions based on the results and discussion, it can be concluded that anti-enac was successfully mobilized on the spce-au surface. the optimum experimental conditions, obtained through the box-behnken experimental (a) admet & dmpk 11(2) (2023) 211-226 electrochemical immunosensor for enac doi: https://doi.org/10.5599/admet.1629 223 design, were anti-enac concentration of 2.5 g/ml, the incubation time of glutaraldehyde for 30 minutes, and the incubation time of anti-enac for 90 minutes. the performance of the immunosensor developed in this study resulted in a detection limit and a quantification limit of 0.0372 ng/ml and 0.124 ng/ml. the lod obtained shows that this electrochemical immunosensor can differentiate enac protein levels in urine samples of non-hypertensive patients, without hypertension, or with a family history of hypertension, between 1.12 to 4.0 ng/ml or higher. this proves that the developed electrochemical immunosensor has the potential to be a point of care testing compared to the elisa method in the detection of enac protein as a salt-sensitive biomarker of hypertension. acknowledgements: we would like to give an acknowledgment for this article to the indonesian ministry of education and culture for penelitian tesis magister no. 094/e5/pg.02.00.pt/2022 and academic leadership grant of universitas padjadjaran, indonesia no .2203/un6.3.1/pt.00/2022. conflict of interest: the authors declare that they have no conflict of interest. references [1] r. castilo. prevalence and management of hypertension in southeast asia. special satellite symposium 03 17 (2016) 1. [2] who. world health statistics 2015. (2015) 1–164. https://www.who.int/docs/default-source/ghodocuments/world-health-statistic-reports/world-health-statistics-2015.pdf (accessed july 20, 2021). [3] b.p. dan p.k. kementerian kesehatan ri. hasil utama riskesdas 2018. https://kesmas.kemkes.go.id/ assets/upload/dir_519d41d8cd98f00/files/hasil-riskesdas-2018_1274.pdf (accessed july 17, 2021). [4] d. irawan, i. muhimmah, t. yuwono. prototype smart instrument untuk klasifikasi penyakit hipertensi berdasarkan jnc-7. jurnal teknologi informasi dan terapan 4 (2019) 111–118. https://doi.org/10.25047/jtit.v4i2.68. [5] y. sun, j.n. zhang, d. zhao, q.s. wang, y.c. gu, h.p. ma, z.r. zhang. role of the epithelial sodium channel in salt-sensitive hypertension. acta pharmacologica sinica 32 (2011) 789–797. https://doi.org/ 10.1038/aps.2011.72. [6] k. ando, t. fujita. pathophysiology of salt sensitivity hypertension. annals of medicine 44 (2012) 119– 126. https://doi.org/10.3109/07853890.2012.671538. [7] h. garty, l.g. palmer. epithelial sodium channels: function, structure and regulation. physiological reviews 77 (1997) 359–396. https://doi.org/10.1152/physrev.1997.77.2.359. [8] i. hanukoglu, a. hanukoglu. epithelial sodium channel (enac) family: phylogeny, structure-function, tissue distribution, and associated inherited diseases. gene 579 (2016) 95–132. https://doi.org/10.1016/j.gene.2015.12.061. [9] s. kellenberger, l. schild. epithelial sodium channel/degenerin family of ion channels: a variety of functions for a shared structure. physiological reviews 82 (2002) 735–767. https://doi.org/10.1152/physrev.00007.2002. [10] y. sofiatin, r. ma roesli. detection of urinary epithelial sodium channel (enac) protein. american journal of clinical medicine research 6 (2018) 20–23. https://doi.org/10.12691/ajcmr-6-2-1. [11] a. khoirunnisa, y.w. hartati. detecting epithelial sodium channel (enac) protein levels as a biomarker of hypertension. journal of medical pharmaceutical and allied sciences 11 (2022) 5004–5011. https://doi.org/10.55522/jmpas.v11i4.2840. [12] y.w. hartati, s. gaffar, d. alfiani, u. pratomo, y. sofiatin, t. subroto. a voltammetric immunosensor based on gold nanoparticle anti-enac bioconjugate for the detection of epithelial sodium channel (enac) protein as a biomarker of hypertension. sensing and bio-sensing research 29 (2020) 100343. https://doi.org/10.1016/j.sbsr.2020.100343. [13] y.w. hartati, s.f. yusup, fitrilawati, s. wyantuti, y. sofiatin, s. gaffar. a voltammetric epithelial sodium channels immunosensor using screen-printed carbon electrode modified with reduced graphene oxide. current chemistry letters 9 (2020) 151–160. https://doi.org/10.1016/j.sbsr.2020.100343. https://doi.org/10.5599/admet.1629 https://kesmas.kemkes.go.id/assets/upload/dir_519d41d8cd98f00/files/hasil-riskesdas-2018_1274.pdf https://kesmas.kemkes.go.id/assets/upload/dir_519d41d8cd98f00/files/hasil-riskesdas-2018_1274.pdf https://doi.org/10.25047/jtit.v4i2.68 https://doi.org/10.25047/jtit.v4i2.68 https://doi.org/10.3109/07853890.2012.671538 https://doi.org/10.1152/physrev.1997.77.2.359 https://doi.org/10.1016/j.gene.2015.12.061 https://doi.org/10.1016/j.gene.2015.12.061 https://doi.org/10.1152/physrev.00007.2002 https://doi.org/10.1152/physrev.00007.2002 https://doi.org/10.12691/ajcmr-6-2-1 https://doi.org/10.55522/jmpas.v11i4.2840 https://doi.org/10.1016/j.sbsr.2020.100343 https://doi.org/10.1016/j.sbsr.2020.100343 t. f.h. lestari et al. admet & dmpk 11(2) (2023) 211-226 224 [14] y.w. hartati, n. satriana, s. gaffar, j. mulyana, s. wyantuti, y. sofiatin. electrochemical label-free immunosensor for the detection of epithelial sodium channels using gold modified screen-printed carbon electrode. iconistech (2019). https://doi.org/10.4108/eai.11-7-2019.2298070. [15] e.i. fazrin, a.k. sari, r. setiyono, s. gaffar, y. sofiatin, h.h. bahti, y.w. hartati. the selectivity and stability of epithelial sodium channel (enac) aptamer as an electrochemical aptasensor. anal. bioanal. electrochem. 14 (2022) 715–729. [16] d.l. garcía‐rubio, m.b. de la mora, d. cerecedo, j.m. saniger blesa, m. villagrán‐muniz. an opticalbased biosensor of the epithelial sodium channel as a tool for diagnosing hypertension. biosensors and bioelectronics 157 (2020) 112151. https://doi.org/10.1016/j.bios.2020.112151. [17] a.f. collings, f. caruso. biosensors: recent advances. progress in phyiscs 60 (1997) 1397–1445. https://doi.org/10.1088/0034-4885/60/11/005 [18] d. grieshaber, r. mackenzie, j. vörös, e. reimhult. electrochemical biosensors sensor principles and architectures. sensors 8 (2008) 1400–1458. https://doi.org/10.3390/s8031400. [19] g. shruthi, c. amitha, b.b. mathew. biosensors: a modern day achievement. journal of instrumentation technology 2 (2014) 26–39. https://doi.org/10.12691/jit-2-1-5. [20] a.p.f. turner. biosensors. current opinion in biotechnology 5 (1994) 49–53. [21] c. moina, g. ybarra. fundamentals and applications of immunosensors, 2012. https://doi.org/10.1021/ac60368a765. [22] g.k. ahirwal, c.k. mitra. gold nanoparticles based sandwich electrochemical immunosensor. biosensors and bioelectronics 25 (2010) 2016–2020. https://doi.org/10.1016/j.bios.2010.01.029. [23] s. kubendhiran, s. sakthinathan, s.m. chen, c.m. lee, b.s. lou, p. sireesha, c. su. electrochemically activated screen printed carbon electrode decorated with nickel nano particles for the detection of glucose in human serum and human urine sample. international journal of electrochemical science 11 (2016) 7934–7946. https://doi.org/10.20964/2016.09.11. [24] h. beitollahi. advanced electrochemical sensors based on the functional carbon materials. journal of electrochemical science and engineering 12 (2022) 1–2. https://doi.org/10.5599/jese.1302. [25] p.m. jahani. flower-like mos2 screen-printed electrode based sensor for the sensitive detection of sunset yellow fcf in food samples. journal of electrochemical science and engineering 12 (2022) 1099– 1109. https://doi.org/10.5599/jese.1413. [26] m.i. gonzález-sánchez, b. gómez-monedero, j. agrisuelas, j. iniesta, e. valero. electrochemical performance of activated screen printed carbon electrodes for hydrogen peroxide and phenol derivatives sensing. journal of electroanalytical chemistry 839 (2019) 75–82. https://doi.org/10.1016/j.jelechem.2019.03.026. [27] n. rongwaree, t. watcharawittayakul, c. sooksamphanwong, p. sakdarat, c. thanachayanont, s. prichanont, p. pungetmongkol. low cost cuo nanorod growth on screen-printed carbon electrode for pesticide analysis. proceedings of the ieee conference on nanotechnology 2020 (2020) 189–192. https://doi.org/10.1109/nano47656.2020.9183705. [28] j.m. jian, l. fu, j. ji, l. lin, x. guo, t.l. ren. electrochemically reduced graphene oxide/gold nanoparticles composite modified screen-printed carbon electrode for effective electrocatalytic analysis of nitrite in foods. sensors and actuators, b: chemical 262 (2018) 125–136. https://doi.org/10.1016/j.snb.2018.01.164. [29] j. zhou, l. du, l. zou, y. zou, n. hu, p. wang. an ultrasensitive electrochemical immunosensor for carcinoembryonic antigen detection based on staphylococcal protein a au nanoparticle modified gold electrode. sensors and actuators, b: chemical 197 (2014) 220–227. https://doi.org/10.1016/j.snb.2014.02.009. [30] m.m.s. silva, i.t. cavalcanti, m.f. barroso, m.g.f. sales, r.f. dutra. gold electrode modified by selfassembled monolayers of thiols to determine dna sequences hybridization. journal of chemical sciences 122 (2010) 911–917. https://doi.org/10.1007/s12039-010-0079-7. https://doi.org/10.4108/eai.11-7-2019.2298070 https://doi.org/10.1016/j.bios.2020.112151 https://doi.org/10.1088/0034-4885/60/11/005 https://doi.org/10.3390/s8031400 https://doi.org/10.12691/jit-2-1-5 https://doi.org/10.1021/ac60368a765 https://doi.org/10.1021/ac60368a765 https://doi.org/10.1016/j.bios.2010.01.029 https://doi.org/10.20964/2016.09.11 https://doi.org/10.5599/jese.1302 https://doi.org/10.5599/jese.1413 https://doi.org/10.1016/j.jelechem.2019.03.026 https://doi.org/10.1016/j.jelechem.2019.03.026 https://doi.org/10.1109/nano47656.2020.9183705 https://doi.org/10.1016/j.snb.2018.01.164 https://doi.org/10.1016/j.snb.2014.02.009 https://doi.org/10.1016/j.snb.2014.02.009 https://doi.org/10.1007/s12039-010-0079-7 admet & dmpk 11(2) (2023) 211-226 electrochemical immunosensor for enac doi: https://doi.org/10.5599/admet.1629 225 [31] m.ç. canbaz, ç.s. şimşek, m.k. sezgintürk. electrochemical biosensor based on self-assembled monolayers modified with gold nanoparticles for detection of her-3. analytica chimica acta 814 (2014) 31–38. https://doi.org/10.1016/j.aca.2014.01.041. [32] j.n. miller, j.c. miller. statistics and chemometrics for analytical chemistry, sixth edit, prentice hall england, edinburgh gate, 2010. https://doi.org/10.7861/clinmedicine.14-6-677. [33] d.c. montgomery. design and analysis of experiments eighth edition. arizona state university, 2013. [34] g.e.p. box, d.w. behnken. some new three level desing for study of quantitative variables. technometrics 2 (1960) 455–475. https://www.tandfonline.com/doi/abs/10.1080/00401706.1960.10489912. [35] d.c.a.s.u. montgomery. design and analysis of experiments ninth edition, 2017. www.wiley.com/go/permissions.%0ahttps://lccn.loc.gov/2017002355. [36] p.d. patel, a. lakdawala, r.n. patel. box–behnken response surface methodology for optimization of operational parameters of compression ignition engine fuelled with a blend of diesel, biodiesel and diethyl ether. biofuels 7 (2016) 87–95. https://doi.org/10.1080/17597269.2015.1118784. [37] f. arduini, s. guidone, a. amine, g. palleschi, d. moscone. acetylcholinesterase biosensor based on self-assembled monolayer-modified gold-screen printed electrodes for organophosphorus insecticide detection. sensors and actuators, b: chemical 179 (2013) 201–208. https://doi.org/10.1016/j.snb.2012.10.016. [38] j.m. pingarrón, p. yáñez-sedeño, a. gonzález-cortés. gold nanoparticle-based electrochemical biosensors. electrochimica acta 53 (2008) 5848–5866. https://doi.org/10.1016/j.electacta.2008.03.005. [39] d.r. baer, m.h. engelhard, g.e. johnson, j. laskin, j. lai, k. mueller, p. munusamy, s. thevuthasan, h. wang, n. washton, a. elder, b.l. baisch, a. karakoti, s.v.n.t. kuchibhatla, d. moon. surface characterization of nanomaterials and nanoparticles: important needs and challenging opportunities. journal of vacuum science & technology a: vacuum, surfaces, and films 31 (2013) 050820. https://doi.org/10.1116/1.4818423. [40] s. wyantuti, m. permadi, d. hendrati, y.w. hartati. modifikasi elektrode glassy carbon dengan nanopartikel emas dan aplikasinya untuk mendeteksi kromium(vi) secara voltammetri pulsa differensial. al-kimia 5 (2017) 12–20. https://doi.org/10.24252/al-kimia.v5i1.2844. [41] e.i. fazrin, a.i. naviardianti, s. wyantuti, s. gaffar, y.w. hartati. review: sintesis dan karakterisasi nanopartikel emas (aunp) serta konjugasi aunp dengan dna dalam aplikasi biosensor elektrokimia. pendipa journal of science education 4 (2020) 21–39. https://doi.org/10.33369/pendipa.4.2.21-39. [42] l. nie, f. liu, p. ma, x. xiao. applications of gold nanoparticles in optical biosensors. journal of biomedical nanotechnology 10 (2014) 2700–2721. https://doi.org/10.1166/jbn.2014.1987. [43] v. amendola, r. pilot, m. frasconi, o.m. maragò, m.a. iatì. surface plasmon resonance in gold nanoparticles: a review. journal of physics condensed matter 29 (2017). https://doi.org/10.1088/1361-648x/aa60f3. [44] a.k. sari, y.w. hartati, s. gaffar, i. anshori, d. hidayat, h.l. wiraswati. the optimization of an electrochemical aptasensor to detect rbd protein s sars-cov-2 as a biomarker of covid-19 using screen-printed carbon electrode/aunp. journal of electrochemical science and engineering 12 (2022) 219–235. https://doi.org/10.5599/jese.1206. [45] f.c. krebs. fabrication and processing of polymer solar cells: a review of printing and coating techniques. solar energy materials and solar cells 93 (2009) 394–412. https://doi.org/10.1016/j.solmat.2008.10.004. [46] a. rana, n. baig, t.a. saleh. electrochemically pretreated carbon electrodes and their electroanalytical applications a review. journal of electroanalytical chemistry 833 (2019) 313–332. https://doi.org/10.1016/j.jelechem.2018.12.019. [47] m. zaib, m.m. athar. electrochemical evaluation of phanerocheaete chrysosporium based carbon paste electrode with potassium ferricyanide redox system. international journal of electrochemical science 10 (2015) 6690–6702. https://doi.org/10.5599/admet.1629 https://doi.org/10.1016/j.aca.2014.01.041 https://doi.org/10.7861/clinmedicine.14-6-677 https://www.tandfonline.com/doi/abs/10.1080/00401706.1960.10489912 https://www.tandfonline.com/doi/abs/10.1080/00401706.1960.10489912 https://doi.org/10.1080/17597269.2015.1118784 https://doi.org/10.1016/j.snb.2012.10.016 https://doi.org/10.1016/j.snb.2012.10.016 https://doi.org/10.1016/j.electacta.2008.03.005 https://doi.org/10.1116/1.4818423 https://doi.org/10.24252/al-kimia.v5i1.2844 https://doi.org/10.33369/pendipa.4.2.21-39 https://doi.org/10.1166/jbn.2014.1987 https://doi.org/10.1088/1361-648x/aa60f3 https://doi.org/10.1088/1361-648x/aa60f3 https://doi.org/10.5599/jese.1206 https://doi.org/10.1016/j.solmat.2008.10.004 https://doi.org/10.1016/j.solmat.2008.10.004 https://doi.org/10.1016/j.jelechem.2018.12.019 https://doi.org/10.1016/j.jelechem.2018.12.019 t. f.h. lestari et al. admet & dmpk 11(2) (2023) 211-226 226 [48] s. petrovic. cyclic voltammetry of hexachloroiridate(iv): an alternative to the electrochemical study of the ferricyanide ion. the chemical educator 5 (2000) 231–235. https://doi.org/10.1007/s00897000416a. [49] y.w. hartati, d.r. komala, d. hendrati, s. gaffar, a. hardianto, y. sofiatin, h.h. bahti. an aptasensor using ceria electrodeposited-screen-printed carbon electrode for detection of epithelial sodium channel protein as a hypertension biomarker. royal society open science 8 (2021) 1-11. https://doi.org/ 10.1098/rsos.202040. [50] r.k. mendes, r.f. carvalhal, l.t. kubota. effects of different self-assembled monolayers on enzyme immobilization procedures in peroxidase-based biosensor development. journal of electroanalytical chemistry 612 (2008) 164–172. https://doi.org/10.1016/j.jelechem.2007.09.033. [51] j.r. reimers, m.j. ford, s.m. marcuccio, j. ulstrup, n.s. hush. competition of van der waals and chemical forces on gold–sulfur surfaces and nanoparticles. nature reviews chemistry 1 (2017) 1-2. https://doi.org/10.1038/s41570-0017. [52] m. tachibana, k. yoshizawa, a. ogawa, h. fujimoto, r. hoffmann. sulfur-gold orbital interactions which determine the structure of alkanethiolate/au(111) self-assembled monolayer systems. journal of physical chemistry b 106 (2002) 12727–12736. https://doi.org/10.1021/jp020993i. [53] f. lópez-gallego, j.m. guisan, l. bentacor. glutaraldehyde-mediated protein immobilization. in: immobil. enzym. cells third ed. methods mol. biol., 2013: pp. 33–41. https://doi.org/10.1007/978-162703-550-7. [54] e.p. randviir, c.e. banks. electrochemical impedance spectroscopy: an overview of bioanalytical applications. analytical methods 5 (2013) 1098–1115. https://doi.org/10.1039/c3ay26476a. [55] m. el-azazy. electrochemical impedance spectroscopy (eis) in food, water, and drug analyses: recent advances and applications. electrochemical impedance spectroscopy (2020) 1–17. https://doi.org/ 10.5772/intechopen.92333. [56] m. emami, m. shamsipur, r. saber, r. irajirad. an electrochemical immunosensor for detection of a breast cancer biomarker based on antiher2–iron oxide nanoparticle bioconjugates. analyst 139 (2014) 2858–2866. https://doi.org/10.1039/c4an00183d. ©2023 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) https://doi.org/10.1007/s00897000416a https://doi.org/10.1007/s00897000416a https://doi.org/10.1098/rsos.202040 https://doi.org/10.1098/rsos.202040 https://doi.org/10.1016/j.jelechem.2007.09.033 https://doi.org/10.1038/s41570-0017 https://doi.org/10.1021/jp020993i https://doi.org/10.1007/978-1-62703-550-7 https://doi.org/10.1007/978-1-62703-550-7 https://doi.org/10.1039/c3ay26476a https://doi.org/10.5772/intechopen.92333 https://doi.org/10.5772/intechopen.92333 https://doi.org/10.1039/c4an00183d http://creativecommons.org/licenses/by/3.0/ recent advances and challenges in antibacterial drug development doi: http://dx.doi.org/10.5599/admet.1271 147 admet & dmpk 10(2) (2022) 147-151; doi: https://doi.org/10.5599/admet.1271 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index short communication recent advances and challenges in antibacterial drug development valeria gigante*, hatim sati, and peter beyer 1 impact initiatives and research coordination unit, global coordination department, amr division, who, geneva, switzerland *corresponding author: e-mail: gigantev@who.int received: january 31, 2022; revised: february 13, 2022; published: march 04, 2022 keywords: antimicrobial resistance; who pipeline; research and development; public health. antimicrobial resistance (amr) is one of the leading causes of death globally. recent estimates reveal that 1.2 million deaths were due to resistant bacterial infections in 2019 and 4.95 million deaths were associated with resistant bacterial infections [1]. there are some evidences suggesting that the current global crisis due to covid-19 pandemic might have exacerbated antibiotic use. in addition, the pandemic has marginalized many public health priorities and programs in countries, resulting in services and programs interruption such as disease surveillance. this all together threatens the gains made in recent years against amr. this includes nourishing the pipeline, investing in the development of new therapies knowing that research and development (r&d) of safe and effective products takes time, and adopting policy at the government level such as advanced remuneration frameworks to create a viable environment for new products [2]. since 2017 world health organization (who) analyses the antibacterial development pipeline annually. the analysis covers traditional direct-acting small molecules and nontraditional antibacterial agents in clinical and preclinical development worldwide. the aim is to evaluate to what extent does the present pipeline address the who bacterial priority pathogens, mycobacterium tuberculosis and clostridium difficile [3]. the end goal is to understand the state of the pipeline and to direct r&d towards public health unmet needs. today who pipeline review is the only global analysis that regularly covers antibacterial agents in development by large pharmaceutical companies as well as by medium and small size enterprises [4]. findings from this analysis are featured in an annual who publication and the data are also made public through who r&d observatory [5] as well as the global amr r&d hub. overall, this information allows for secondary analyses such as the estimation of the source and number of investments across the different development phases as shown in the global amr r&d hub dashboard [6,7]. the long road to approval between july 2017 and november 2021, twelve new antibacterial agents have been approved by either the us food and drug administration (fda), the european medicines agency (ema), or both. only one of 1 the authors are staff members of the world health organization. the authors alone are responsible for the views expressed in this article and they do not necessarily represent the decisions, policy or views of the world health organization http://dx.doi.org/10.5599/admet.1271 https://doi.org/10.5599/admet.1271 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:gigantev@who.int%20n gigante, sati and beyer admet & dmpk 10(2) (2022) 147-151 148 these was approved in the last two years. this translates to an average duration of approximately 10.1 years for a new antimicrobial agent to advance from the discovery and preclinical stage until the market authorization [8]. looking at the progression rate, a who financial model estimates that an antibacterial product in preclinical development has approximately a 12.5 % chance of successfully moving past the registration phase [8]. these estimations reflect the scientific complexity of developing new drug. several factors make the development of innovative antibacterials specifically challenging. these include the limited funding, and the limited return on investment in the current volume-driven revenue models. this is exacerbated by the current epidemic that has slowed down and even halted many clinical and preclinical programmes. however, covid-19 could be looked at as an opportunity to push for ad-hoc regulatory pathways, sustainable public financial models, and other financial incentives that prioritize antibacterial agents as a finite public resource. this push could also include investments in new emerging technologies and nontraditional approaches such as synthetic mrnas/antimicrobial peptides (amps) and phage therapies. surviving the market and proving the value evaluating how the newly approved products address global public health priorities, only one agent, cefiderocol, is said to be active against all three “critical” gram-negative bacteria. almost 50 % of the newly approved agents (n = 5) target one priority pathogen in the critical category, carbapenem-resistant enterobacterales (cre). this leaves two critical targets; carbapenem-resistant acinetobacter baumannii (crab) and carbapenem-resistant pseudomonas aeruginosa (crpa), largely unaddressed at present. with some exceptions, all newly approved agents present limited clinical benefit over existing treatment. over 80 % of the newly approved antibacterials belong to existing classes (mostly ßlactamases) where resistance mechanisms are well established, and the development of resistance is expected. the evident challenge to design new antibacterial classes calls for prudent use of the existing agents to ensure their longevity. in fact, the majority (7 out of 12) of the newly approved antibiotics are classified as “reserve” according to the most recent who aware classification, while three are in the “watch” group [9,10]. one agent, contezolid has not been classified at the time of this publication, but like other oxazolidinones, is expected to be included in the “reserve” group in the next iteration of the aware classification [10]. this also indicates that most new agents presently in development, when approved, will likely be in the reserve group of antibiotics to be used only when other treatment lines have failed. this need to preserve new agents is another challenge to antibacterial development that could further limit investments. it also highlights the need for universal decoupled public financial model(s) through existing push or new pull mechanisms for antibacterial drug development. however, to justify public funding, antimicrobial programs need to demonstrate potential clinical utility of candidates and ideally novelty, by moving away from the development of worn derivatives with limited potential for clinical differentiation. innovation in traditional antibacterial agents traditional antibacterial agents are defined as small-molecule agents that directly target bacteria to either halt their growth (bacteriostatic effect) or to kill them (bactericidal effect) [11]. the current clinical antibacterial pipeline counts approximately 78 antibacterial agents containing either new therapeutic entity or combinations of existing molecules with at least one new therapeutic entity. among the 78 products, 45 are traditional antibacterial agents and 33 are nontraditional agents. of the 45 traditional antibiotics, 27 (60 admet & dmpk 10(2) (2022) 147-151 challenges in antibacterial drug development doi: http://dx.doi.org/10.5599/admet.1271 149 %) are reported to be active against the who bacterial priority pathogens, 13 (31.1 %) against mycobacterium tuberculosis (tb) and five (11.1 %) against clostridium difficile [5]. innovation is a key aspect evaluated when reviewing antibacterial agents in clinical development. there is a lack of consensus on what qualifies as innovative antibacterial agent. who looks for evidence of innovation in: the proposed class; the mechanism of action; the molecular target; and the absence of crossresistance. [5]. considering the chemical class and the molecular target, among the 27 agents identified against the who critical priority pathogens, most of the antimicrobial agents in the clinical pipeline are derivatives of existing classes and β-lactams-β-lactamase inhibitors (blis) accounts for over 40 % (12/27) of these agents. most of these combinations are active against class a, and c and some are active against class d enzymes. very few are reported to target class b enzymes (mbls). both pseudomonas aeruginosa and, to a certain extent, acinetobacter baumannii, have developed resistance mechanisms beyond the production of βlactamases, including decreased permeability of the outer membrane and upregulation of efflux pumps and modified penicillin-binding protein (pbp). this may explain the copious β-lactams-blis combinations approach found in the pipeline, which attempts to overcome these resistance mechanisms challenges. however, this approach also reflects a lack of innovation in terms of new chemical classes and may result in many “me too” agents with limited potential for clinical differentiation. in fact, examining all the 27 traditional agents under development against who bacterial priority pathogens, only a few compounds in clinical development satisfy at least one of the innovation criteria described and address at the same time critical gram-negative bacteria (crab, crpa or cre). the anti-tb clinical antibacterial pipeline, on the other hand, looks more innovative, with seven (53 %) agents showing new chemical structures. it is important to point out here that for any given agent in development, meeting one or more of the innovation criteria described must be interpreted in the context of clinical and public health unmet needs. nontraditional agents, new approaches, same developmental pathway “nontraditional antimicrobials” refers to therapies that are not small molecules and/or aim at a nontraditional target and include a broad range of agents [12]. since 2019, the who antibacterial pipeline analysis has examined these new strategies as complementary and synergistic, or even future alternative to traditional antibacterial agents (i.e., small molecules) [13]. among the 78 products identified in the clinical pipeline, 34 are nontraditional agents in a different stage of development: two are in the marketing, authorization, application/new drug application (maa/nda) stage, five products in phase 3, eleven in phase 2, and five in phase 1/2. nontraditional agents encompass different categories from phage-based therapies to monoclonal antibodies, antimicrobial peptides and antibacterial enhancers. due to the wide range of approaches pursued in the nontraditional pipeline, these agents hold a great potential to curb amr. although very distinct from each other in their nature, for the most part, traditional and nontraditional agents should be considered comparable when it comes to the need to adhere to predictable regulatory requirements for their development and potentially their approvals [14]. during the who technical advisory group meeting on antibacterial research and development held on 29 and 30 november 2021, the scope of the nontraditional section was discussed, and it was decided that bacteriophage treatments will be included in the who pipeline analysis only if they were associated to a http://dx.doi.org/10.5599/admet.1271 gigante, sati and beyer admet & dmpk 10(2) (2022) 147-151 150 randomized controlled clinical trial while expanded access/compassionate use bacteriophage programs won’t be listed in the pipeline analysis. in addition, agents against biothreat pathogens will likely be considered in the 2023 update of the who antibacterial pipeline analysis and will also be evaluated in the context of the next update of the who bacterial priority pathogen list (bppl). in conclusion, in the present scenario where traditional products have a limited lifespan before resistance emerges, non-conventional approaches offer opportunities to preserve new and old traditional agents, and to tackle amr from different angles. preclinical, traveling the valley of death in the preclinical phase, there is a high turnover with almost one out of three developers lost every year. half of the products in preclinical development target a single species, with 56 % targeting staphylococcus aureus, pseudomonas aeruginosa and mycobacterium tuberculosis. this trend, which is also observed in the nontraditional space, requires availability of diagnostics at the point of care for optimal use, often unavailable outside of specialized healthcare facilities and poses a challenge in low-resource settings. overall, our analyses of the preclinical pipeline indicate a volatile landscape, with longer timeframes and challenging milestones to meet before potentially hitting the market. thus, public, and private incentives, including new funding models becomes crucial in this space to continuously feed the clinical development with promising, safe, effective, on target and innovative new compounds [5,12]. final thoughts the findings from 2021 who antibacterial pipeline analysis confirm that r&d is driven by private pharmaceutical companies, but the progress made since last year are extremely modest with only one product obtaining marketing authorization in china, one new product entering phase 3 and five products entering phase 1 and 7 products terminating their development. smaller biotechs are facing major challenges, and the number of overall programs are still largely insufficient to properly address the burden of amr. antibacterial agents in clinical development do not conclusively address the problem of extensive or pandrug-resistant gram-negative bacteria. novel antibiotics targeting the critical who priority pathogens, are still lacking, in particular, carbapenem-resistant a. baumannii and p. aeruginosa. the pipeline also lacks oral antibiotic treatment options for extended spectrum β-lactamases (esbls) and cre, which could allow for treatments outside the hospital-setting, and shorter treatment duration. in addition, optimized pediatric formulations of existing antibiotics are also needed, and there are barely any active pediatric development programs. overall, the recently approved antibacterial agents combined with the late-stage traditional agents under development are insufficient to address the enormous threat posed by amr. more novel compounds that address unmet clinical needs are needed. these compounds must elude known mechanisms of drug resistance, and fit into modern reimbursement mechanisms. global public health efforts continue to focus on the relentless covid-19 pandemic; resources are further diverted from other public health priorities, including amr. this has also impacted clinical and preclinical developments, which have been substantially halted by the covid-19 outbreak, with most programmes incurring considerable costs and delays likely for years to come. a global response is needed to regain the time lost in the race against the silent amr pandemic lurking in the shadows, waiting for its turn. admet & dmpk 10(2) (2022) 147-151 challenges in antibacterial drug development doi: http://dx.doi.org/10.5599/admet.1271 151 conflict of interest: no conflict of interest is declared by the authors. references [1] c.j. murray, k.s. ikuta, f. sharara, l. swetschinski, g.r. aguilar, a. gray, c. han, c. bisignano, p. rao, e. wool, s.c. johnson. global burden of bacterial antimicrobial resistance in 2019: a systematic analysis. the lancet 399 (2022) p629-655. https://doi.org/10.1016/s0140-6736(21)02724-0. [2] m.s. butler, v. gigante, h. sati, s. paulin, l. al-sulaiman, j.h. rex, p. fernandes, c.a. arias, m. paul, g.e. thwaites, l. czaplewski. analysis of the clinical pipeline of treatments for drug resistant bacterial infections: despite progress, more action is needed. antimicrobial agents and chemotherapy (2022) https://doi.org/10.1128/aac.01991-21 . [3] prioritization of pathogens to guide discovery, research, and development of new antibiotics for drug-resistant bacterial infections, including tuberculosis. world health organization; 2017. https://apps.who.int/iris/handle/10665/311820. [4] antibiotics currently in global clinical development, pew’s antibiotic resistance project communication, updated: march 9, 2021. available via: antibiotics currently in global clinical development | the pew charitable trusts (pewtrusts.org) (last accessed 01/27/2022). [5] 2020 antibacterial agents in clinical and preclinical development: an overview and analysis. geneva: world health organization; 2021. licence: cc by-nc-sa 3.0 igo. [6] who global observatory on health research and development. available via: https://www.who.int/observatories/global-observatory-on-health-research-anddevelopment/monitoring/antibacterial-products-in-clinical-development-for-priority-pathogens. (last accessed 01/27/2022). [7] the global amr r&d hub’s dynamic dashboard. pipeline gallery. available via: https://globalamrhub.org/dynamic-dashboard/pipeline-gallery/ (last accessed 01/27/2022). [8] a financial model for an impact investment fund for the development of antibacterial treatments and diagnostics: a user guide. geneva: world health organization; 2020. licence: cc by-nc-sa 3.0 igo. [9] m. sharland, c. pulcini, s. harbarth, m. zeng, s. gandra, s. mathur, n. magrini. classifying antibiotics in the who essential medicines list for optimal use—be aware. the lancet infectious diseases 18 (2018) 18-20. https://doi.org/10.1016/s1473-3099(17)30724-7. [10] 2021 aware classification . who access, watch, reserve, classification of antibiotics for evaluation and monitoring of use. who reference number: who/hmp/hps/eml/2021.04 copyright cc bync-sa 3.0 igo. avaliable via: https://www.who.int/publications/i/item/2021-aware-classification (last accessed on 27/01/2022). [11] b.n. tse, a.a: adalja, c. houchens, j. larsen, t.v. inglesby, r. hatchett. challenges and opportunities of nontraditional approaches to treating bacterial infections. clinical infectious diseases 65 (2017) 495-500. https://doi.org/10.1093/cid/cix320. [12] u. theuretzbacher, l.j. piddock. nontraditional antibacterial therapeutic options and challenges. cell host & microbe 26 (2019) 61-72. https://doi.org/10.1016/j.chom.2019.06.004. [13] antibacterial agents in preclinical development: an open access database. geneva: world, health organization; 2019 (who/emp/iau/2019.12). licence: cc by-nc-sa 3.0 igo. [14] j.h. rex, h.f. lynch, i.g. cohen, j.j. darrow, k. outterson. designing development programs for nontraditional antibacterial agents. nature communications 10 (2019) p61-72. https://doi.org/10.1016/j.chom.2019.06.004. ©2022 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://dx.doi.org/10.5599/admet.1271 https://doi.org/10.1016/s0140-6736(21)02724-0 https://doi.org/10.1128/aac.01991-21 https://apps.who.int/iris/handle/10665/311820 https://www.pewtrusts.org/en/research-and-analysis/data-visualizations/2014/antibiotics-currently-in-clinical-development https://www.pewtrusts.org/en/research-and-analysis/data-visualizations/2014/antibiotics-currently-in-clinical-development https://www.who.int/observatories/global-observatory-on-health-research-and-development/monitoring/antibacterial-products-in-clinical-development-for-priority-pathogens https://www.who.int/observatories/global-observatory-on-health-research-and-development/monitoring/antibacterial-products-in-clinical-development-for-priority-pathogens https://globalamrhub.org/dynamic-dashboard/pipeline-gallery/ https://doi.org/10.1016/s1473-3099(17)30724-7 https://www.who.int/publications/i/item/2021-aware-classification https://doi.org/10.1093/cid/cix320 https://doi.org/10.1016/j.chom.2019.06.004 https://doi.org/10.1016/j.chom.2019.06.004 https://doi.org/10.1016/j.chom.2019.06.004 http://creativecommons.org/licenses/by/3.0/ synthesis and diverse biological activity profile of triethyl­ammonium isatin-3-hydrazones doi: http://dx.doi.org/10.5599/admet.1179 163 admet & dmpk 10(2) (2022) 163-179; doi: https://doi.org/10.5599/admet.1179 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper synthesis and diverse biological activity profile of triethylammonium isatin-3-hydrazones andrei bogdanov* 1 , olga tsivileva* 2 , alexandra voloshina 1 , anna lyubina 1 , syumbelya amerhanova 1 , ekaterina burtceva 1 , sergey bukharov 3 , alexander samorodov 4 and valentin pavlov 4 1 arbuzov institute of organic and physical chemistry, frc kazan scientific center of ras, kazan, 420088, russian federation 2 institute of biochemistry and physiology of plants and microorganisms, saratov scientific centre of the russian academy of sciences, saratov 410049, russian federation 3 kazan national research technological university, kazan 420015, russian federation 4 bashkir state medical university, ufa 450000, russian federation *corresponding authors: andrei bogdanov e-mail: abogdanov@inbox.ru; tel.: +7-962-562-44-18; olga tsivileva email: tsivileva@ibppm.ru; tel.: +7-960-346-25-02 received: november 15, 2021; revised: january 07, 2022; published: january 12, 2022 abstract a series of biorelevant triethylammonium isatin hydrazones containing various substituents in the aromatic fragment have been synthesized. their structure and composition were confirmed by nmr and ir-spectroscopies, mass-spectrometry and elemental analysis. it was found that some representatives show activity against staphylococcus aureus and bacillus cereus higher or at the level of norfloxacin, including methicillin-resistant staphylococcus aureus strains. the study also showed low hemoand cytotoxicity (chang liver) and high antiaggregatory and anticoagulant activity of these compounds. the high potential of new ammonium isatin-3-acylhydrazones in the search for antimicrobial activity against phytopathogens of bacterial and fungal nature has been shown for the first time. ©2021 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords isatin; hemostasis; phytopathogens, antibacterial and antifungal activities introduction as a representative of the class of privileged structures, isatin and its derivatives are widely used in medicinal chemistry [1,2]. the progressing number of works on synthetic procedures [3] and the studies of this heterocycle push researchers in this area to publish generalized data on one or another type of biological activity [4-7]. it is due to the manifestation of isatin and its derivatives of a wide spectrum of activity, such as anti-cancer [8-10], anti-tubercular [11,12], antibacterial [13], anti-covid [14,15], fungicidal [16,17], etc (fig. 1). http://dx.doi.org/10.5599/admet.1179 https://doi.org/10.5599/admet.1179 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:abogdanov@inbox.ru mailto:tsivileva@ibppm.ru http://creativecommons.org/licenses/by/4.0/ bogdanov, tsivileva et al. admet & dmpk 10(2) (2022) 163-179 164 figure 1. biologically active isatins and isatin-3-acylhydrazones one of the modern trends in the design and targeted synthesis of bioactive isatin derivatives is the concept of molecular hybridization [18-21]. in this regard, the combination of a pharmacophore fragment and a quaternized nitrogen atom in one molecule seems to be promising in terms of the search for effective and non-toxic biorelevant drugs [22,23]. in recent years, our research group has been working in this direction, namely in the synthesis and study of the antimicrobial activity of water-soluble isatin hydrazones containing a positively charged nitrogen atom (fig. 2) [24-30]. those studies showed the dependence of antimicrobial activity on many structural factors. thus, in the series of trimethylammonium isatin acylhydrazones, sterically hindered analogs showed the best activity against some gram-positive bacteria [24-26]. we also found that both an increase in the lipophilicity of the benzyl substituent [30] and the presence of an alkyl chain of medium length (c10-c12) [27] in position 1 leads to an improvement in antimicrobial activity. it is important that all previously obtained compounds have low hemoand cytotoxic effects. figure 2. antimicrobial ammonium isatin-3-acylhydrazones experimental materials and methods starting isatins 1a-g were synthesized accordingly to our previously reported procedure [24]. ir spectra were measured with a bruker vector-22 instrument for the samples in kbr pellets. 1 h and 13 c nmr spectra were recorded on a bruker avance-400 or bruker avance-600 bruker spectrometers at 400, 600 and 100.6, 150 mhz, respectively. chemical shifts were reported in ppm relative to residual signals of deuterated solvents. сdcl3, dmso-d6 or a mixture of cdcl3/dmso-d6 were used as the nmr solvents. maldi mass spectra were recorded on an ultraflex iii tof/tof mass spectrometer in linear mode with a recording of positive ions, metal target, and p-nitroaniline matrix. elemental analysis was performed on a euro vector 2000 chns-3 instrument; halogen content was determined by pyrolysis in the oxygen stream. melting points were determined using an smp10 stuart instrument and uncorrected. admet & dmpk 10(2) (2022) 163-179 isatin-3-hydrazones: synthesis and biological activity doi: http://dx.doi.org/10.5599/admet.1179 165 n,n,n-triethyl-2-hydrazinyl-2-oxoethanammonium bromide (1). white crystalline powder. yield: 73 %; m.p. 135 °c. ir (kbr, cm −1 ): 3298 (n-h), 3191 (n-h), 3010 (c-h), 2946 (c-h), 1680 (c=o), 1646 (с=о), 1530 (n-h); 1449 (с-n). 1 h nmr (600 mhz, dmso-d6, δ, ppm): 1.23 t (9h, ch3, et, j 7.2 hz); 3.43 q (6н, ch2, et, j 7.2 hz); 3.98 s (2н, ch2co); 4.51 br. s [2н, nh2]; 9.89 s (1н, nh). 13 c nmr (150 mhz, dmso-d6, δ, ppm): 160.90 (c=o); 55.19 (ch2); 54.24 (ch2); 8.45 (ch3). ms (maldi): 213 [m+k-br] + . found, %: c, 37.67; h, 7.84; br, 31.32; n, 16.46. c8h20brn3o. calculated, %: c, 37.80; h, 7.93; br, 31.44; n, 16.53. scheme 1. synthetic route to the ammonium hydrazones 3a-g general procedure for the synthesis of compounds 3a-g. a mixture of substituted isatin 1a-g (1 mmol) and hydrazide 2 (1 mmol) was magnetically stirred in absolute ethanol (7 ml) for 10 min, followed by the addition of trifluoroacetic acid (20 mol%). then the reaction mixture was heated at reflux for 3 h. after cooling the solution to room temperature, the solvent was rotary evaporated. the formed precipitate was washed with anhydrous diethyl ether, filtered off and dried in vacuum (12 mmhg). 2-(2-(1-(3,5-di-tert-butyl-4-hydroxybenzyl)-2-oxoindolin-3-ylidene)hydrazinyl)-n,n,n-triethyl-2oxoethanammonium bromide (3a). yellow powder. yield: 92 %; m.p. 212 °c. ir (kbr, cm −1 ): 3631 (o-h), 3232 (n-h), 2950 (c-h), 1713 (c=o), 1673 (с=о), 1617 (с=с), 1469 (с=n). 1 h nmr (600 mhz, cdcl3, δ, ppm): 1.36 s (18н, сн3, t-bu); 1.44 t (9h, ch3, et, j 6.9 hz); 3.88 q (6н, ch2, et, j 7.0 hz); 4.75 s [2н, nch2]; 5.02 s (2н, ch2co); 5.17 s (1н, оh); 6.84 d (1н, 7-h, ar, j 7.8 hz); 8.02-8.03 m (3н, 5-h, ar, 2h, benzyl); 7.29 dd (1н, 6-h, ar, j 7.6 hz, j 7.6 hz); 8.03 d (1h, 4-h, ar, j 7.2 hz); 12.83 s (1н, nh). 13 c nmr (150 mhz, cdcl3, δ, ppm): 165.51 (c=o); 160.90 (c=o); 153.50 (c-oh); 143.49; 136.51; 136.38; 132.07 (ch); 125.25; 124.49 (ch); 123.83 (ch); 123.30 (ch); 118.77; 109.77 (ch); 55.19 (ch2); 54.24 (ch2); 43.84 (ch2); 34.15 (c, t-bu); 30.06 (ch3, t-bu); 8.45 (ch3). ms (maldi): 521 [m-br] + . found, %: c, 61.79; h, 7.45; br, 13.16; n, 9.25. c31h45brn4o3. calculated, %: c, 61.89; h, 7.54; br, 13.28; n, 9.31. 2-(2-(1-(3,5-di-tert-butyl-4-hydroxybenzyl)-7-methyl-2-oxoindolin-3-ylidene)hydrazinyl)-n,n,n-triethyl-2oxoethanammonium bromide (3b). yellow powder. yield: 97 %; m.p. 154 °c. ir (kbr, cm −1 ): 3634 (o-h), 3204 (n-h), 2959 (c-h), 1714 (c=o),1683 (с=о), 1606 (с=с), 1446 (с=n). 1 h nmr (400 mhz, cdcl3, δ, ppm): 1.36 s (18н, сн3, t-bu); 1.48 t (9h, ch3, et, j 6.0 hz); 2.35 s (3н, ar-сн3); 3.91 q (6н, ch2, et, j 6.2 hz); 5.06-5.07 m [4н, nch2, ch2co); 5.14 br. s (1н, оh); 6.97-7.00 m (3н, 6-h, ar, 2h, benzyl); 7.08-7.11 m (1н, 5-h, ar); 8.03 d (1h, 4-h, ar, j 7.0 hz); http://dx.doi.org/10.5599/admet.1179 bogdanov, tsivileva et al. admet & dmpk 10(2) (2022) 163-179 166 12.91 s (1н, nh). ms (maldi): 535 [m-br] + . found, %: c, 62.30; h, 7.56; br, 12.81; n, 9.01. c32h47brn4o3. calculated, %: c, 62.43; h, 7.69; br, 12.98; n, 9.10. 2-(2-(1-(3,5-di-tert-butyl-4-hydroxybenzyl)-7-ethyl-2-oxoindolin-3-ylidene)hydrazinyl)-n,n,n-triethyl-2oxoethanammonium bromide (3c). yellow powder. yield: 89 %; m.p. 163 °c. ir (kbr, cm −1 ): 3609 (o-h), 3392 (n-h), 3208 (n-h), 2957 (c-h), 1682 (с=о), 1606 (с=с), 1460 (с=n), 1436 (с=n). 1 h nmr (600 mhz, dmso-d6, δ, ppm): 0.93 t (3н, ch3, et, j 7.3 hz); 1.27-1.29 m (27н, сн3, t-bu, ch3, et); 2.63 q (2н, ch2, et, j 7.3 hz); 3.57-3.62 m (6н, ch2, et); 4.58 s (1н, оh); 4.80 s (2н, ch2co); 5.06 s [2н, nch2]; 6.93 s (2н, 2h, benzyl); 7.13-7.17 m (1н, 5-h, ar); 7.27 d (1н, 7-h, ar, j 7.1 hz); 7.65-7.69 m (1h, 4-h, ar); 12.64 s (1н, nh). 13 c nmr (150 mhz, dmso-d6, δ, ppm): 166.21 (c=o); 161.45 (c=o); 152.96 (c-oh); 140.48; 139.70; 134.41 (ch); 127.73; 127.19; 123.63 (ch); 122.01 (2ch); 119.64; 118.99; 54.18 (ch2); 53.15 (ch2); 44.63 (ch2); 34.41 (c, t-bu); 30.15 (ch3, t-bu); 23.35 (ch2); 15.56 (ch3); 7.49 (ch3). ms (maldi): 549 [m-br] + . found, %: c, 62.80; h, 7.73; br, 12.59; n, 8.78. c33h49brn4o3. calculated, %: c, 62.95; h, 7.84; br, 12.69; n, 8.90. 2-(2-(1-(3,5-di-tert-butyl-4-hydroxybenzyl)-6,7-dimethyl-2-oxoindolin-3-ylidene)hydrazinyl)-n,n,n-triethyl2-oxoethanammonium bromide (3d). yellow powder. yield: 94 %; m.p. 187 °c. ir (kbr, cm −1 ): 3415 (o-h), 3199 (n-h), 2952 (c-h), 1707 (c=o), 1678 (с=о), 1611 (с=с), 1463 (с=n), 1435 (с=n). 1 h nmr (600 mhz, cdcl3, δ, ppm): 1.34 s (18н, сн3, t-bu); 1.46 t (9h, ch3, et, j 7.3 hz); 2.18 s (3н, ar-сн3); 2.24 s (3н, ar-сн3); 3.89 q (6н, ch2, et, j 7.3 hz); 4.97 s [2н, ch2co); 5.05 s [2н, nch2); 5.13 br. s (1н, оh); 6.93-6.95 m (3н, 5-h, ar, 2h, benzyl); 7.77 d (1h, 4-h, ar, j 7.1 hz); 12.83 s (1н, nh). 13 c nmr (150 mhz, cdcl3, δ, ppm): 165.31 (c=o); 162.66 (c=o); 153.10 (c-oh); 143.55; 142.00; 136.57; 136.41; 126.67; 125.90 (ch); 122.61 (ch); 120.54; 120.15; 117.67; 55.26 (ch2); 54.22 (ch2); 45.53 (ch2); 34.23 (c, t-bu); 30.12 (ch3, t-bu); 21.36 (ch3); 13.85 (ch3); 8.50 (ch3). ms (maldi): 535 [m-br] + . found, %: c, 62.35; h, 7.50; br, 12.83; n, 8.99. c32h47brn4o3. calculated, %: c, 62.43; h, 7.69; br, 12.98; n, 9.10. 2-(2-(1-(3,5-di-tert-butyl-4-hydroxybenzyl)-4-bromo-2-oxoindolin-3-ylidene)hydrazinyl)-n,n,n-triethyl-2oxoethanammonium bromide (3e). yellow powder. yield: 89 %; m.p. 194 °c. ir (kbr, cm −1 ): 3592 (o-h), 3367 (n-h), 3199 (n-h), 2952 (c-h), 1686 (с=о), 1607 (с=с), 1447 (с=n), 1433 (с=n). 1 h nmr (400 mhz, cdcl3, δ, ppm): 1.37 s (18н, сн3, t-bu); 1.48 t (9h, ch3, et, j 6.9 hz); 3.90 q (6н, ch2, et, j 7.0 hz); 4.78 s [2н, nch2]; 4.87 s (2н, ch2co); 5.22 s (1н, оh); 6.88 d (1н, 7-h, ar, j 7.8 hz); 7.09 s (2н, 2h, benzyl); 7.18 dd (1н, 6-h, ar, j 7.8 hz, j 7.8 hz); 7.26 d (1н, 5-h, ar, j 7.8 hz); 12.93 s (1н, nh). 13 c nmr (100.6 mhz, cdcl3, δ, ppm): 165.37 (c=o); 160.07 (c=o); 153.42 (c-oh); 144.64; 136.33; 134.60; 132.27 (ch); 128.00 (c-h); 124.67; 124.34 (ch); 117.79; 117.09; 108.87 (ch); 65.40 (ch2); 55.64 (ch2); 43.87 (ch2); 33.94 (c, t-bu); 29.87 (ch3, t-bu); 8.39 (ch3). ms (maldi): 601 [m-br] + . found, %: c, 54.60; h, 6.40; br, 23.33; n, 8.11. c31h44br2n4o3. calculated, %: c, 54.71; h, 6.52; br, 23.48; n, 8.23. 2-(2-(1-(3,5-di-tert-butyl-4-hydroxybenzyl)-5-nitro-2-oxoindolin-3-ylidene)hydrazinyl)-n,n,n-triethyl-2oxoethanammonium bromide (3f). yellow powder. yield: 89 %; m.p. 176 °c. ir (kbr, cm −1 ): 3621 (o-h), 3392 (n-h), 3215 (n-h), 2954 (c-h), 1693 (с=о), 1617 (с=с), 1523 (n=o), 1486 (с=n), 1434 (с=n). 1 h nmr (400 mhz, cdcl3, δ, ppm): 1.38 s (18н, сн3, t-bu); 1.49 t (9h, ch3, et, j 6.3 hz); 3.86-3.90 m (6н, ch2, et); 4.84 s [2н, nch2]; 5.22 s (1н, оh); 5.28 s (2н, ch2co); 6.94 d (1н, 7-h, ar, j 8.5 hz); 7.09 s (2н, 2h, benzyl); 8.18 d (1н, 6-h, ar, j 8.5 hz); 8.90 br. s (1н, 4-h, ar); 12.67 s (1н, nh). 13 c nmr (150 mhz, dmso-d6, δ, ppm): 167.15 (c=o); 161.62 (c=o); admet & dmpk 10(2) (2022) 163-179 isatin-3-hydrazones: synthesis and biological activity doi: http://dx.doi.org/10.5599/admet.1179 167 153.52 (c-oh); 148.20; 143.26; 139.51; 127.92 (c-h); 125.94; 124.28 (2ch); 119.59; 116.21; 110.96 (ch); 54.09 (ch2); 53.90 (ch2); 43.54 (ch2); 34.41 (c, t-bu); 30.20 (ch3, t-bu); 7.49 (ch3). ms (maldi): 566 [m-br] + . found, %: c, 57.42; h, 6.70; br, 12.21; n, 10.70. c31h44brn5o5. calculated, %: c, 57.58; h, 6.86; br, 12.36; n, 10.83. 2-(2-(1-(3,5-di-tert-butyl-4-hydroxybenzyl)-5-chloro-7-bromo-2-oxoindolin-3-ylidene)hydrazinyl)-n,n,ntriethyl-2-oxoethanammonium bromide (3g). yellow powder. yield: 89 %; m.p. 177 °c. ir (kbr, cm −1 ): 3629 (o-h), 3400 (n-h), 3209 (n-h), 2956 (c-h), 1689 (с=о), 1606 (с=с), 1447 (с=n). 1 h nmr (400 mhz, cdcl3/dmso-d6, δ, ppm): 1.29 s (18н, сн3, t-bu); 1.33 t (3н, ch3, et, j 7.1 hz); 3.66-3.70 m (6н, ch2, et); 4.87 s (2н, ch2co); 5.22 s [2н, nch2]; 7.04 s (2н, 2h, benzyl); 7.51 s (1н, оh); 7.68-7.69 m (1н, 6-h, ar); 7.88 br. s (1h, 4-h, ar); 12.58 s (1н, nh). 13 c nmr (100.6 mhz, cdcl3/dmso-d6, δ, ppm): 166.14 (c=o); 161.24 (c=o); 153.59 (c-oh); 139.41; 138.11; 136.48 (c-h); 129.29; 127.30; 126.76; 123.98 (c-h); 123.43; 121.22 (c-h); 103.72; 54.90 (ch2); 54.13 (ch2); 44.49 (ch2); 34.60 (c, t-bu); 30.39 (ch3, t-bu); 8.15 (ch3). ms (maldi): 635 [m-br] + . found, %: c, 51.90; h, 5.91; br, 22.23; cl, 4.72; n, 7.71. c31h43br2cln4o3. calculated, %: c, 52.08; h, 6.06; br, 22.35; cl, 4.96; n, 7.84. antimicrobial activity study the antimicrobial activity of the test compounds was determined by the serial dilution technique in muller hinton broth 2 and in sabouradu broth for fungi. the cultures used for testing included grampositive bacteria: staphylococcus aureus атсс 6538р fda 209p, bacillus cereus атсс 10702 nctc 8035, enterococcus faecalis atcc 29212; gram-negative bacteria: escherichia coli atcc 25922, pseudomonas aeruginosa atcc 9027, and fungi: trichophyton mentagrophytes var. gypseum 1773, and candida albiсans atcc 10231; methicillin-resistant strains of s. aureus (mrsa) was obtained from hospital patients with chronic tonsillitis in the republican clinical hospital (kazan, russia). the bacterial load was 3.0 × 10 5 cfu/ml. the fungi load was 2.0 × 10 3 cfu/ml. the results were recorded every 24 h for 5-7 days. cultures were incubated at 37 °c. the experiment was repeated three times. the dilutions of the compounds were prepared immediately in nutrient media; 5 % dmso was added for better solubility and the test strains were not inhibited at this concentration. the minimum inhibitory concentration (mic) was defined as the minimum concentration of a compound that inhibits the growth of the corresponding test microorganism. the growth of bacteria, as well as the absence of the growth due to the bacteriostatic action of compounds, were recorded. to determine minimal bactericidal concentration (mbc), an aliquot of the bacterial culture was transferred onto mueller-hinton agar or sabouradu agar in a 10-cm petri dish and incubated for 24 h at 37 °с. mbc was the minimal concentration at which bacterial colonies were not detected, indicating that the bacteria were killed with an efficiency of >99.9 % [31]. hemolytic activity hemolytic activity of test compounds was estimated by comparing the optical density of a solution containing the test compound with that of blood at 100 % hemolysis. the experiments were carried out as described earlier [32]. cytotoxicity assay cytotoxic effects of the test compounds on human normal cells were estimated by means of the multifunctional cytell cell imaging system (ge health care life science, sweden) using the cell viability bio app which precisely counts the number of cells and evaluates their viability from fluorescence intensity [32]. two fluorescent dyes that selectively penetrate the cell membranes and fluoresce at different wavelengths were used in the experiments. dapi is able to penetrate intact membranes of living cells and http://dx.doi.org/10.5599/admet.1179 bogdanov, tsivileva et al. admet & dmpk 10(2) (2022) 163-179 168 colors nuclei in blue and propidium iodide dye penetrates only dead cells with damaged membranes, staining them in yellow. dapi and propidium iodide were purchased from sigma. ic50 was calculated using an online tool: mla-“quest graph™ ic50 calculator.” aat bioquest, inc, 15 july, 2021, https://www.aatbio.com/tools/ic50-calculator. chang liver cell line (human liver cells) from n. f. gamaleya research center of epidemiology and microbiology was used in the experiments. the cells were cultured in a standard eagle’s nutrient medium manufactured at the chumakov institute of poliomyelitis and virus encephalitis (paneco company) and supplemented with 10 % fetal calf serum and 1 % nonessential amino acids. the cells were plated into a 96-well plate (nunc) at a concentration of 1×10 5 cells/ml, 150 μl of medium per well, and cultured in a co2 incubator at 37 °c. twenty-four hours after seeding the cells into wells, the compound under study was added at a preset dilution, 150 μl to each well. the dilutions of the compounds were prepared immediately in nutrient media; 5 % dmso that does not induce inhibition of cells at this concentration was added for better solubility. the experiments were repeated three times. intact cells cultured in parallel with experimental cells were used as a control. anticoagulant and anti-aggregation activities study the in vitro experiments were performed using the blood of healthy male donors aged 18-24 years (total 54 donors). the study was approved by the ethics committee of federal state budgetary educational institution of higher education at the bashkir state medical university of the ministry of health of russian federation (no.2 dated 17.10.2012). informed consent was obtained from all participants before blood sampling. the blood was collected from the cubital vein using the system of vacuum blood collection bd vacutainer ® (becton, dickinson and company, usa). a 3.8 % sodium citrate solution in a 9:1 ratio was used as a venous blood stabiliser. the study of the effect on platelet aggregation was performed using the born method [33] using the aggregometer «ат-02» (spc medtech, russia). the assessment of antiplatelet activity of the studied compounds and reference preparations was started with the final concentration of 2×10 −3 mol/l. adenosine diphosphate (adp; 20 µg/ml) and collagen (5 mg/ml) manufactured by tehnologia-standart company, russia, were used as inducers of aggregation. the study on the anticoagulant activity was performed by standard recognised clotting tests using the optical two-channel automatic analyser of blood coagulation solar cgl 2110 (cjsc solar, belarus). the following parameters were studied: activated partial thromboplastin time (aptt), prothrombin time (pt) and fibrinogen concentrations according to the clauss method. the determination of anticoagulant activity of the studied compounds and reference preparation was performed in a concentration of 5×10 −4 g/ml using the reagents manufactured by tehnologia-standart company (barnaul, russia). the results of the study were processed using the statistical package statistica 10.0 (statsoft inc, usa). the shapiro–wilk’s test was used to check the normality of actual data distribution. the form of distribution of the data obtained differed from the normal one; therefore, non-parametric methods were used for further analysis. the data were presented as medians and 25 and 75 percentiles. analysis of variance was conducted using kruskal–wallis test. a p value of 0.05 was considered statistically significant. antiphytopathogenic activity study plant pathogenic bacterial strains micrococcus luteus b-109, pectobacterium atrosepticum 1043, pectobacterium carotovorum subsp. сarotovorum mi, pseudomonas fluorescens el-2.1, and xanthomonas campestris b-610 and fungal strain fusarium oxysporum ibppm 543 were obtained from the specialized scientific culture collection of ibppm ras (wfcc no. 975, wdcm no. 1021) (см ibppm). pathogenic fungus phytophthora cactorum vkm f-985 provided by the all-russian collection of microorganisms (vkm) and deposited at a.e. favorsky irkutsk institute of chemistry, sb ras, was also used as test organism. bacteria https://www.aatbio.com/tools/ic50-calculator https://www.aatbio.com/tools/ic50-calculator admet & dmpk 10(2) (2022) 163-179 isatin-3-hydrazones: synthesis and biological activity doi: http://dx.doi.org/10.5599/admet.1179 169 m. luteus, p. carotovorum subsp. сarotovorum, p. atrosepticum, and ps. fluorescens were grown in meatpeptone medium (bp), and x. campestris was grown in the medium with glucose, yeast extract, and calcium carbonate (gyca). solid media contained bacto agar (18 g/l); ph was adjusted to 7.2-7.4. all bacterial cultures were grown at 28°c. the mycelial cultures of f. oxysporum and p. cactorum were grown on a glucose-peptone-yeast (gpy) nutrient medium at 27 °c. for inoculum preparation, both fungal strains were initially grown on the agar gpy in petri dishes and then transferred into the seed medium by punching out 5 mm of the agar plate culture with a self-designed cutter. antibacterial and antifungal activities of the compounds were explored using the agar diffusion method and the technique of a phytopathogen radial growth inhibition. method of diffusion in agar (measuring the diameter of growth inhibition zones) was used for determining the bactericidal activity. the 6-mm wells were made in ager medium (gyca for xanthomonas campestris or bp for other bacteria). bacterial suspensions were distributed over the agar surface, and the tested compound's solution (150 μl) was added to each well. the width of growth inhibition zones around the wells was determined after incubation for 36-40 h. for the fungicidal activity analysis, radial growth (colony diameters) of the fungi on a solid medium in the absence and in the presence of the compounds 3a,c-g solutions at various concentrations were compared. the method consisted of the following: sterile, melted and then cooled to about 60°c gpy agar medium (20.0 ml) was mixed with the precisely measured volumes of the solutions under question, and poured into a sterile petri dish (90 mm i.d.). after solidification of gpy agar, the media were inoculated by the fungus using 10-day-old cultures of f. oxysporum or p. cactorum. the inoculation was done by transferring a 5-mm (i.d.) gpy-agar block covered with mycelium to the center of the petri dish followed by incubation in a thermostat at 27°c. the fungicidal effect was scored by the size of mycelium colony on a petri dish compared to the control without fungicidal admixtures to gpy agar. each treatment was performed in at least four replicates in two independent experiments. the observation period ended when the control petri dish was filled with mycelium (usually after 12 days). the inhibition of the phytopathogen colony growth by the compounds 3a,c-g or fludioxonil solutions was calculated as a percentage by which the mycelium radial propagation was decreased compared to the unaffected control, the latter was taken as 100 % of growth (or zero percent of inhibition). the ec50 value was calculated as the compound concentration at which the radial growth of the fungus colony was decreased by 50 % relative to the nonfungicidal control, according to the formula ec50 = 50c/i based on the approach developed e.g., in [34,35]. the solutions of tested compounds were prepared with a concentration of 2 mmol/l (stock solution). for comparison, test compounds were also commonly applied disinfecting agents and antibiotics. results and discussion chemistry the synthesis of target compounds is based on our previously developed approach [24-30] and is carried out in two stages (fig. 1). at the first stage, hydrazide 1 (an analog of the girard’s reagent t) was obtained for the first time, containing three ethyl groups at the quaternized nitrogen atom. further, this hydrazide was involved in the condensation with isatin derivatives bearing a phenolic fragment in position 1 and substituents of different nature in the benzo fragment of the heterocycle. the reaction proceeds in ethanol at reflux temperature for 3 hours in the presence of trifluoroacetic acid as a catalyst. after easy workup, the desired reaction products were isolated in pure form with high yields (89-97 %). the structure of all obtained hydrazones 3a-g was proved using nmr and ir spectroscopy. thus, the ir spectra of the new http://dx.doi.org/10.5599/admet.1179 bogdanov, tsivileva et al. admet & dmpk 10(2) (2022) 163-179 170 compounds contain narrow, intense absorption bands in the region of 3600-3630 cm -1 and broad bands of medium intensity at 3210-3380 cm -1 , corresponding to the stretching vibrations of o-h and n-h bonds, respectively. the presence of these functional groups is also confirmed by the data of 1 h nmr spectra, in which the signal of the proton of the hydroxyl group appears in the region of ~ 5.2 ppm, and the signal of the hydrazone proton in the lowest fields at ~ 12.8 ppm. antimicrobial activity evaluation synthesized compounds were further evaluated for antimicrobial activity against test microorganisms: staphylococcus aureus atcc 209p (sa), bacillus cereus atcc 8035 (bc), escherichia coli cdc f-50 (ec), enterococcus faecalis (ef), pseudomonas aeruginosa atcc 9027, aspergillus niger bkmf-1119, trichophyton mentagrophytes var. gypseum 1773, and candida albicans 855-653 including methicillin-resistant strains of s. aureus (mrsa 1 and mrsa 2). the minimum inhibitory (mic) and minimum bactericidal (mbc) concentrations of ammonium salts were determined against these pathogens (table 1). due to the lack of activity of compounds 3a-g against gram-negative bacterial strains and some fungi, the corresponding data are not shown. table 1. antimicrobial activity of compounds 3a-g * . compound mic/mbc, µm sa bc ef mrsa-1 mrsa-2 ca 3a 6.5±0.5/ 13.0±1.2 52±4.7/nd* nd -** 3b 3.2±0.2/ 25.4±2.1 25.4±1.9/ 203±16 50.8±4.1/ 203.2±16.4 25.4±2.3/ 50.8±4.1 3c 3.1±0.2 /6.2±0.5 24.8±1.9/ 99.3±8.2 49.7±3.9/ 99.3±8.3 24.8±2.2/ 99.3±8.3 3d 3.1±0.2/ 12.4±1.1 49.6±3.9/ 198.5±15.7 49.6±3.9/ 49.6±3.9 6.2±0.5/ 12.4±1.2 6.2±0.5/ 49.6±3.8 198.5±15.5/ 3e 8.9±0.7/ 17.7±1.5 71±6.2/ 283.7±22.7 nd nd 3f 48.3±3.9/ 96.7±7.6 397±31.7/nd nd 3g 11±0.9/ 21.9±1.8 43.7±3.5/ 87.4±7.2 nd nd norfloxacin 7.5±0.5/ 7.5±0.6 24.4±2.1/ 24.4±2.1 7.5±0.5/ 7.5±0.6 391.4±30/na§ 30.0±2.6/na * were not determined (compounds possess low activity); § no activity; ** means >500 the initial study of antimicrobial activity showed that triethylammonium derivatives 3a-g, like their trimethylammonium analogs [24-27], selectively act on gram-positive pathogenic bacterial strains with high activity. compared with norfloxacin against sa, compounds 3a-e (mic 8.9-3.1 µm) turned out to be the best in bacteriostatic effect, and hydrazones 3b,c (mic ~25 µm) were the best against bc. according to the bactericidal action derivative 3c was determined as the leader compound, containing an ethyl radical in position 7 with an mbc of 6.2 µm, better than that of the reference drug. it should be noted that only compound 3d showed antimicrobial activity against enterococcus faecalis that often causes a wide range of nosocomial human infections. it is especially important to note that three representatives of this series compounds 3b-d showed high activity against both types of mrsa used. in this case, the most effective was the most lipophilic hydrazone 3d, containing two methyl groups in the aromatic fragment of the heterocycle. this compound also showed moderate activity against the yeast-like fungus candida albicans. admet & dmpk 10(2) (2022) 163-179 isatin-3-hydrazones: synthesis and biological activity doi: http://dx.doi.org/10.5599/admet.1179 171 thus, the data obtained showed that acylhydrazones bearing electron-donor alkyl groups in the benzo fragment possess the best activity against both museum and methicillin-resistant bacterial strains. an important step in the determination of the biological activity of new chemical compounds is the assessment of their cytotoxic action in relation to mammalian cells. the ability of the investigated compound to cause the destruction of human erythrocytes illustrates its toxic effect on the internal environment of the body. the hemolysis assay is a simple screening test, the results of which can help in the study of cytotoxicity in more complex models. such experimental models can be the cell lines obtained from various organs and tissues of a person and allowing to adequately assessing the effect of new potential drugs on cell metabolism. in this regard, the studied compounds were tested for cytotoxicity against blood erythrocytes and the human hepatocytes (chang liver cell lines (fig. 3). hemolytic and cytotoxic activity data are represented by hc50 and ic50 values. it can be seen that all compounds 3a-3g in the range of tested concentrations did not have high hemolytic and cytotoxic activity. hc50 concentrations were 72.2-243.8 µm; ic50 113-168 µm. from the point of view of the effect on red blood cells, the least toxic compound is compound 3b containing methyl group at position 7 of the aromatic fragment. reference drugs gramicidin s and doxorubicin turned out to be much more toxic to red blood cells and liver cells. figure 3. hemolytic and cytotoxic activity of 3a-g the selectivity of compounds for microbial cells is an important criterion for assessing the cytotoxic effect. this indicator is characterized by the value of the selectivity index (si), which for the leading compounds 3b, 3c, and 3d was calculated as the ratio between the hc50 value for erythrocytes (ic50 for eukaryotic cells) and the mic value for bacterial cells (fig. 4). it can be seen that with respect to the s. aureus 209 p test strain, all tested compounds exhibit a sufficiently high selectivity. compound 3d bearing two methyl groups demonstrated the most significant selectivity against resistant strains mrsa-1 and mrsa-2 compared to erythrocytes and hepatocytes. its selectivity index was 3–8 times higher than that of compounds 3b and 3c. http://dx.doi.org/10.5599/admet.1179 bogdanov, tsivileva et al. admet & dmpk 10(2) (2022) 163-179 172 figure 4. selectivity of 3b, 3c and 3d for bacteria (s. aureus 209, mrsa-1and mrsa-2) compared to red blood cells and chang liver cells. anticoagulant and anti-aggregation activities evaluation the development of new drug candidates is a multi-stage process that requires consideration of the influence of many factors, including the level of toxicity and the presence of side effects. there are many examples of drug recalls from the market in world practice, which gradually reveals serious side effects associated, for example, with the risk of developing cardiovascular diseases [36,37]. in this regard, it seemed appropriate and important to assess the effect of the newly synthesized isatin-3-acylhydrazones on the hemostatic system. in this work, we carried out a primary study of the anti-aggregation and anticoagulant activity of compounds 3a-g (table 2). table 2. anticoagulant and anti-aggregation activity of compounds 3a-g compound platelets aggregation, % of control aptt $$ , % of control adp $ collagen 3a -10.5 (8.7-13.9)* -9.4 (8.5-12.3)* +5.2 (4.1-7.6)*, † 3b -9.7 (7.4-12.3)* -10.3 (9.6-12.5)* +7.3 (6.5-9.6)*, † 3c -13.7 (11.5-15.1)** -12.5 (9.8-14.3)* +11.2 (7.9-13.1)*, † 3d -9.0 (7.5-10.6)*, # -8.5 (7.2-12.3)*, # +4.8 (3.7-6.9)†† 3e -0.5 (0.2-0.9)## -2.4 (1.7-3.5)## +3.2 (2.8-5.4)†† 3f -4.6 (3.7-5.3)*, # -4.3 (3.4-5.6)*, # +10.3 (8.5-12.7)*,† 3g -10.2 (8.6-12.2)* -8.9 (7.5-10.3)*, # +8.6 (6.1-10.9)*,† acetylsalicylic acid -13.7 (10.8-16.4)* -14.7 (12.1-16.3)** heparin sodium +20.3 (19.7-21.4)** *р ≤ 0.05, **р ≤ 0.001 compared to control; #р ≤ 0.05, ##р ≤ 0.001 compared to acetylsalicylic acid; †р ≤ 0.05, ††р ≤ 0.001 compared to heparin sodium; $ adp – adenosine diphosphate, $$ aptt activated partial thromboplastin time. the compounds demonstrated the varying extent of the effect on the plasmatic component of the haemostasis system that manifested only by a change in the parameter aptt of the intrinsic coagulation pathway. compounds 3c,f demonstrated anticoagulant activity ≥10 % (p<0.05). regarding the impact on platelet aggregation, the compounds demonstrated similar activity on both aggregation inductors. compounds 3a-c,g demonstrated antiaggregatory activity in vitro at the level of acetylsalicylic acid. the admet & dmpk 10(2) (2022) 163-179 isatin-3-hydrazones: synthesis and biological activity doi: http://dx.doi.org/10.5599/admet.1179 173 most promising active compound is 7-ethyl substituted analog 3c. among the obtained series of new compounds, this hydrazone is the most active. antiphytopathogenic activity evaluation to expand the range of possible practical applications of isatin derivatives, synthesized compounds were examined for their antibacterial and antifungal effects against test organisms. bacterial pathogens isolated from the plants microenvironment and harvested vegetables frequently include micrococcus luteus, pseudomonas fluorescens, pectobacterium carotovorum (erwinia carotovora), xanthomonas campestris, and pectobacterium atrosepticum (erwinia carotovora subsp. atrosepticum) [38-40]. fungal pathogens cause 70–80 % of all plant diseases [41], possessing a potentiality of causing large-scale disease outbreaks in a very limited period of time. among the diverse mycelial pathogenic fungi, those from the genera fusarium and phytophthora are among the most destructive plant pathogens known, have broad host ranges, and are capable of causing crop losses and eventual collapse of whole infected plants [42]. f. oxysporum is referred to as the aggressive pathogen among fusarium species, and, in particular, different branches of the food industry are extremely conscious of fusarium infections of cereals [43]. phytophthora cactorum is a causative agent of phytophthora blight of ginseng (panax ginseng), a plant that is very useful for conventional medicine to treat various diseases, including cancer [44]. along with the solutions of 3a,c-g, test compounds were sodium hypochlorite (1000 µg/ml), chlorohexidin (500 µg/ml), and norfloxacin (500 µg/ml), a synthetic fluoroquinolone with broad-spectrum antibacterial activity against most bacteria. a commercial widely used fungicide fludioxonil capable of inhibiting the pathogenic fungi mycelium propagation served for the purpose of comparison in the course of the fungicidal effect assays. under laboratory conditions, the assessment of the phytopathogenic fungi resistance to this fungicide is referred to be conducted at its concentrations from 0.1 to 10 μg/ml [45]. the results of determining the bactericidal activity against m. luteus b-109, p. atrosepticum 1043, p. carotovorum subsp. carotovorum mi, ps. fluorescens el-2.1, x. campestris b-610 showed non-zero activity against bacterial test systems of all the compounds tested (table 3). table 3. bactericidal activity of hydrazones (2 mm/l) studied*. compound** bacterial phytopathogen m. luteus p. atrosepticum p. carotovorum subsp. carotovorum ps. fluorescens x. campestris 3a 9 8 10 9 10 3c 10.5 9 9 8 11 3d 8 8 8 9 8 3e 13 15 10 9 11 3f 10 10 5 10 7.5 3g 7 9 9 8.5 8 norfloxacin, 500 μg/ml 7 7 8 8 7 sodium hypochlorite, 1000 μg/ml 2.5 4 3 4 4 chlorohexidin, 500 μg/ml 5 4 5 4 5 * shows the values of the width of the inhibition zone (mm), averaged over the results of 3 experiments; ** compound 3b was not tested due to the formation of precipitate at the solution preparation compounds 3a, c-g showed high antibacterial activity in this experiment. only in a few cases was the inhibition zone width of 5 and 7.5 mm (3f versus p. carotovorum and x. campestris, respectively), 7 mm (3g versus m. luteus) found. other drugs with an even more pronounced bactericidal effect formed the growth http://dx.doi.org/10.5599/admet.1179 bogdanov, tsivileva et al. admet & dmpk 10(2) (2022) 163-179 174 of phytopathogens inhibition zone from 8 to 15 mm (table 4). table 4. comparative bactericidal activity of compounds 3a, c-g (2 mm/l). bacterial test system inhibition zone width (mm), not less 15 13 11 10 9 8 micrococcus luteus b-109 – 3e – 3c>3f 3a 3d pectobacterium atrosepticum 1043 3e – 3f 3c,48 3a, 3d pectobacterium carotovorum subsp. carotovorum mi – – 3e, 3a 3c, 3g 3d pseudomonas fluorescens el-2.1 – – – 3f 3e, 3a, 3d 3g >3c xanthomonas campestris b-610 – – 3e, 3c 3a 3d, 3g taking into account the data in table 4, it is quite conditionally possible to arrange the studied 2 mm solutions of compounds 3a,c-g in decreasing order of activity as follows: 3e> 3c> 3a> 3d> 3g> 3f. thus, with respect to the phytopathogens used, the best activity was shown by acylhydrazone 3e containing a bromine atom at position 4 of the heterocycle. screening for the manifestation of antifungal properties of compounds 3a,c-g, introduced into the agar medium to cultivate the fungus f. oxysporum, was carried out in the concentration range of 6-43 μg/ml (table 5). table 5. fungicidal activity of compounds 3a,c-g against fusarium oxysporum ibppm 543. compound с, μg/ml inhibition value, i (%)*, at the age of the fungus (days) ec50, μg/ml at day 7 3 5 7 9 3a 6.02 11 4 9 0 33.44 12.04 14 12 18 3 24.08 24 16 27 4 36.12 28 24 30 12 3c 6.30 2 11 30 17 10.50 12.60 24 26 33 23 25.20 31 33 47 27 37.80 44 42 59 38 3d 6.30 0 5 5 0 48.46 12.60 8 12 13 2 25.20 17 15 20 8 37.80 22 18 24 9 3e 6.81 11 13 19 6 17.92 13.62 14 18 22 9 27.24 19 19 25 20 40.86 40 37 43 29 3f 6.47 13 11 10 9 30.81 12.94 26 23 21 10 25.88 28 27 32 12 38.82 29 28 34 18 3g 7.15 0 4 5 4 55.00 14.30 9 4 11 5 28.60 11 7 26 10 42.90 31 27 32 13 fludioxonil 10 38 23 6 4 83.33 * the average values are given; the standard deviation did not exceed 0.03 from the given value. admet & dmpk 10(2) (2022) 163-179 isatin-3-hydrazones: synthesis and biological activity doi: http://dx.doi.org/10.5599/admet.1179 175 the antifungal properties of compounds 3a,c-g, introduced into the agar medium to cultivate the fungus p. cactorum, were studied in the concentration range 1-15 μg/ml. the amount of inhibition of fungal growth was expressed as a percentage, taking into account that the quantitative characteristic of the absence of inhibition is expressed as 0 % (table 6). table 6. fungicidal activity of compounds 3a,c-g against phytophthora cactorum vkm f-985. compound с, μg/ml inhibition value, i (%)*, at the age of the fungus (days) ec50, μg/ml at day 7 3 5 7 9 12 3a 3.01 4 9 15 1 0 10.03 6.02 13 21 22 4 5 9.03 17 22 25 4 5 12.04 40 47 49 41 26 3c 1.26 17 28 30 12 3 2.10 3.15 29 31 38 20 16 6.30 33 44 45 28 20 9.45 39 45 48 41 36 12.60 50 56 61 58 48 3d 1.26 18 17 5 0 0 11.25 3.15 23 18 14 6 5 6.30 29 19 15 8 7 9.45 31 20 17 9 8 12.60 50 50 51 48 35 3e 1.36 8 14 22 10 3 3.10 3.41 13 21 34 12 5 6.81 17 33 38 16 7 10.22 39 40 53 28 18 13.62 45 62 72 70 66 3f 1.29 17 11 2 0 0 8.99 3.24 33 19 18 1 0 6.47 35 23 20 8 1 9.71 41 36 32 20 16 12.94 59 55 53 48 30 3g 1.43 30 29 15 1 0 4.77 3.58 38 29 22 5 0 7.15 40 33 26 14 11 10.73 48 43 41 31 26 14.30 51 48 46 41 28 fludioxonil 10 46 38 18 12 9 27.78 * the average values are given; the standard deviation did not exceed 0.03 from the given value. all compounds 3a,c-g showed an antagonistic effect against p. cactorum, superior to fludioxonil from 2.5 (3d) to 13.2 (3c) times by seven days of the fungus growth. at the same culture age of f. oxysporum, almost all compounds of this series outperformed fludioxonil in antifungal effect by up to 7.9 times (hydrazone 3c). it is important to note that fludioxonil showed the greatest antifungal effect at the age of both cultures three days, and then the fungicidal effect of the commercial drug was significantly reduced (tables 5, 6). at the same time, the fungicidal properties of compounds 3e, 3a, 3c increased during the development of the mycelium. the inhibition value i by these drugs after 5-9 days of p. cactorum growth increased by a maximum of 27 % (3e), in f. oxysporum by 28 % (3c). moreover, in the case of using the fusarium test system, an increase in antifungal ability in the dynamics of the experiment was revealed in the entire group 3a,c-g (table 6). in decreasing order of activity against f. oxysporum, the following series is obtained: 3c> 3e> 3f> 3a> 3d> 3g. for p. cactorum, the series is as follows: 3c > 3e > 3g > 3f > 3a > 3d. thus, newly synthesized isatin derivatives of 3a,c-g, possessing significant antibacterial activity, were able to http://dx.doi.org/10.5599/admet.1179 bogdanov, tsivileva et al. admet & dmpk 10(2) (2022) 163-179 176 significantly inhibit the growth of pathogenic fungi f. oxysporum and p. cactorum on a dense medium. first of all, these are 4-bromo(3e) and 7-ethyl derivatives (3c) "leaders" in all sequences of the location of compounds in terms of antiphytopathogenic activity: against pathogenic bacteria, against fungi f. oxysporum and p. cactorum. conclusions a number of new isatin-3-acylhydrazones containing a triethylammonium fragment have been synthesized, and their multiple biological profiles were revealed. the study of antimicrobial properties showed that the compounds obtained have selective activity against staphylococcus aureus and bacillus cereus. derivatives containing donor alkyl substituents showed bacteriostatic activity two times better than norfloxacin, and the 7-ethyl analog was slightly better than the reference drug in terms of its bactericidal effect. the results obtained showed low toxicity of new compounds towards healthy human cells (red blood cells, hepatocytes) and the absence of a negative effect on some factors of hemostasis. the study of antiphytopathogenic activity showed that 4-bromine (3e) and 7-ethyl (3c) analogs have high activity against a number of dangerous bacterial (micrococcus luteus and pectobacterium atrosepticum) and fungal (f. oxysporum and p. cactorum) pathogens that exceeds those for reference drugs norfloxacin, chlorohexidine and fludioxonil respectively. thus, the data obtained indicate a high potential for the development of effective biocompatible drugs for both pharmaceutical and agricultural purposes. acknowledgements: the authors thank the spectral and analytical joint center (kazan scientific center, russian academy of sciences) for technical support. part of the work on the synthesis and study of antimicrobial, hemolytic and cytotoxic activities (a. bogdanov, a. voloshina, a. lyubina, s. amerkhanova) was carried out within the framework of the state assignment of the federal research center ksc ras. research on antiphytopathogenic activity was completed in the framework of theme no. 121031100266-3 of the program of fundamental research of the russian academy of sciences. conflict of interest: the authors declare no conflict of interest references [1] m.e. welsch, s.a. snyder, b.r. stockwell. privileged scaffolds for library design and drug discovery. current opinion in chemical biology 14 (2010) 347-361. https://doi.org/10.1016/j.cbpa.2010.02.018. [2] c. melis, r. meleddu, a. angeli, s. distinto, g. bianco, c. capasso, f. cottiglia, r. angius, c.t. supuran, e. maccioni. isatin: a privileged scaffold for the design of carbonic anhydrase inhibitors. journal of enzyme inhibition and medicinal chemistry 32 (2017) 68-73. https://doi.org/10.1080/14756366.2016.1235042. [3] a.v. bogdanov, v.f. mironov. advances in the synthesis of isatins: a survey of the last decade. synthesis 50 (2018) 1601-1609. https://doi.org/10.1055/s-0036-1591946. [4] s.z. drmanić, p. petrović, d.r. brkić, a.d. marinković, j.b. nikolić. a survey on the characterization and biological activity of isatin derivatives. journal of the serbian chemical society 85 (2020) 9791000. https://doi.org/10.2298/jsc200320020d. [5] p. brandao, c. marques, a.j. burke, m. pineiro. the application of isatin-based multicomponentreactions in the quest for new bioactive and druglike molecules. european journal of medicinal chemistry 211 (2021) 113102. https://doi.org/10.1016/j.ejmech.2020.113102. [6] y.‐zh. zhang, h.‐zh. du, h.‐l. liu, q.‐s. he, zh. xu. isatin dimers and their biological activities. archiv der pharmazie (weinheim) 353 (2020) e1900299. https://doi.org/10.1002/ardp.201900299. [7] p. brandao, c. marques, e.p.carreiro, m. pineiro, a.j. burke. engaging isatins in multicomponent reactions (mcrs) – easy access to structural diversity. the chemical record 21 (2021) 1-115. https://doi.org/10.1002/tcr.202000167. https://doi.org/10.1016/j.cbpa.2010.02.018 https://doi.org/10.1080/14756366.2016.1235042 https://doi.org/10.1080/14756366.2016.1235042 https://doi.org/10.1055/s-0036-1591946 https://doi.org/10.2298/jsc200320020d https://doi.org/10.1016/j.ejmech.2020.113102 https://doi.org/10.1002/ardp.201900299 https://doi.org/10.1002/tcr.202000167 admet & dmpk 10(2) (2022) 163-179 isatin-3-hydrazones: synthesis and biological activity doi: http://dx.doi.org/10.5599/admet.1179 177 [8] zh. ding, m. zhou, ch. zeng. recent advances in isatin hybrids as potential anticancer agents. archiv der pharmazie (weinheim) 353 (2020) e1900367. https://doi.org/10.1002/ardp.201900367. [9] r.e.f. de paiva, e. guimaraes vieira, d.r. da silva, c.a. wegermann, a.m. costa ferreira. anticancer compounds based on isatin-derivatives: strategies to ameliorate selectivity and efficiency. frontiers in molecular biosciences 7 (2021) 627272. https://doi.org/10.3389/fmolb.2020.627272. [10] d.-j. fu, y.-f. zhang, a.-q. chang, j. li. β-lactams as promising anticancer agents: molecular hybrids, structure activity relationships and potential targets. european journal of medicinal chemistry 201 (2020) 112510. https://doi.org/10.1016/j.ejmech.2020.112510. [11] r. kumar, p. takkar. repositioning of isatin hybrids as novel anti-tubercular agents overcoming preexisting antibiotics resistance. medicinal chemistry research 30 (2021) 847-876. https://doi.org/10.1007/s00044-021-02699-5. [12] zh. xu, sh. zhang, ch. gao, j. fan, f. zhao, z.-sh. lv, l.-sh. feng. isatin hybrids and their antituberculosis activity, chinese chemical letters 28 (2017) 159-167. https://doi.org/10.1016/j.cclet.2016.07.032. [13] f. song, zh. li, yu. bian, x. huo, ju. fang, l. shao, m. zhou. indole/isatin‐containing hybrids as potential antibacterial agents. archiv der pharmazie 353 (2020) e2000143. https://doi.org/10.1002/ardp.202000143. [14] shagufta, i. ahmad. the race to treat covid-19: potential therapeutic agents for the prevention and treatment of sars-cov-2. european journal of medicinal chemistry 213 (2021) 113157. https://doi.org/10.1016/j.ejmech.2021.113157. [15] m. konwar, d. sarma. advances in developing small molecule sars 3clpro inhibitors as potential remedy for corona virus infection. tetrahedron 77 (2021) 131761. https://doi.org/10.1016/j.tet.2020.131761. [16] v.k. pandey, a. dwivedi, o.p. pandey, s.k. sengupta. organophosphorus derivatives containing isatin3-hydrazones as chemotherapeutants against fungal pathogens of sugarcane. journal of agricultural and food chemistry 56 (2008) 10779. https://doi.org/10.1021/jf801975z. [17] l.s. zampirolli, m.j. de lemos, v.t. goncalves, m.a.a. de souza, s.r. de souza, v.m. rumjanek, j.b.n. dacosta. synthesis, characterization, and biological activity of a new class of dialkylphosphorylhydrazone derivatives of isatin. quimica nova 37 (2014) 989-995. https://doi.org/10.5935/0100-4042.20140162. [18] c. viegas-junior, a. danuello, v. da silva bolzani, e.j. barreiro, c.a.m. fraga. molecular hybridization: a useful tool in the design of new drug prototypes. current medicinal chemistry 14 (2007) 1829-1852. https://doi.org/10.2174/092986707781058805. [19] g. berube. an overview of molecular hybrids in drug discovery. expert opinion on drug discovery 11 (2016) 281-305. https://doi.org/10.1517/17460441.2016.1135125. [20] p.l. bosquesi, th.r. ferreira melo, e.o. vizioli, j.l. dos santos, m.ch. chung. anti-inflammatory drug design using a molecular hybridization approach. pharmaceuticals 4 (2011) 1450-1474. https://doi.org/10.3390/ph4111450. [21] r. meleddu, s. distinto, a. corona, e. tramontano, g. bianco, c. melis, f. cottiglia, e. maccioni., isatin thiazoline hybrids as dual inhibitors of hiv-1 reverse transcriptase. journal of enzyme inhibition and medicinal chemistry 130 (2017) 130-136. https://doi.org/10.1080/14756366.2016.1238366. [22] m.c. jennings, k.p.c. minbiole, w.m. wuest. quaternary ammonium compounds: an antimicrobial mainstay and platform for innovation to address bacterial resistance. acs infection diseases 1 (2015) 288-303. https://doi.org/10.1021/acsinfecdis.5b00047. [23] a.n. vereshchagin, n.a. frolov, k.s. egorova, m.m. seitkalieva, v.p. ananikov. quaternary ammonium compounds (qacs) and ionic liquids (ils) as biocides: from simple antiseptics to tunable antimicrobials. international. journal of molecular sciences 22 (2021) 6793. https://doi.org/10.3390/ijms22136793. [24] a.v. bogdanov, i.f. zaripova, a.d. voloshina, a.s. strobykina, n.v. kulik, s.v. bukharov, v.f. mironov. isatin derivatives containing sterically hindered phenolic fragment and water-soluble acyl hydrazones http://dx.doi.org/10.5599/admet.1179 https://doi.org/10.1002/ardp.201900367 https://doi.org/10.3389/fmolb.2020.627272 https://doi.org/10.1016/j.ejmech.2020.112510 https://doi.org/10.1007/s00044-021-02699-5 https://doi.org/10.1007/s00044-021-02699-5 https://doi.org/10.1016/j.cclet.2016.07.032 https://doi.org/10.1016/j.cclet.2016.07.032 https://doi.org/10.1002/ardp.202000143 https://doi.org/10.1002/ardp.202000143 https://doi.org/10.1016/j.ejmech.2021.113157 https://doi.org/10.1016/j.tet.2020.131761 https://doi.org/10.1016/j.tet.2020.131761 https://doi.org/10.1021/jf801975z https://doi.org/10.5935/0100-4042.20140162 https://doi.org/10.2174/092986707781058805 https://doi.org/10.1517/17460441.2016.1135125 https://doi.org/10.3390/ph4111450 https://doi.org/10.1080/14756366.2016.1238366 https://doi.org/10.1021/acsinfecdis.5b00047 https://doi.org/10.3390/ijms22136793 bogdanov, tsivileva et al. admet & dmpk 10(2) (2022) 163-179 178 on their basis: synthesis and antimicrobial activity. russian journal of general chemistry 88 (2018) 57-67. https://doi.org/10.1134/s1070363218010097. [25] a.v. bogdanov, i.f. zaripova, a.d. voloshina, a.s. strobykina, n.v. kulik, s.v. bukharov, ju.k. voronina, a.r. khamatgalimov, v.f. mironov. synthesis and antimicrobial activity evaluation of some novel water-soluble isatin-3-acylhydrazones. monatshefte chemie 149 (2018) 111-117. https://doi.org/10.1007/s00706-017-2049-y. [26] a.v. bogdanov, i.f. zaripova, a.d. voloshina, a.s. sapunova, n.v. kulik, s.v. bukharov, ju.k. voronina, a.e. vandyukov, v.f. mironov. synthesis and biological evaluation of new isatin-based qacs with high antimicrobial potency. chemistryselect 4 (2019) 6162-6166. https://doi.org/10.1002/slct.201901708. [27] t.n. pashirova, a.v. bogdanov, i.f. zaripova, e.a. burilova, a.e. vandyukov, a.s. sapunova, i.i. vandyukova, a.d. voloshina, v.f. mironov, l.ya. zakharova. tunable amphiphilic π-systems based on isatin derivatives containing a quaternary ammonium moiety: the role of alkyl chain length in biological activity. journal of molecular liquids 290 (2019) 111220. https://doi.org/10.1016/j.molliq.2019.111220. [28] a.v. bogdanov, m.e. kadomtseva, s.v. bukharov, a.d. voloshina, v.f. mironov. effect of the cationic moiety on the antimicrobial activity of sterically hindered isatin 3-hydrazone derivatives. russian journal of organic chemistry 56 (2020) 555-558. https://doi.org/10.1134/s107042802003032x. [29] a.v. bogdanov, o.v. andreeva, m.g. belenok, a.d. voloshina, k.i.enikeeva, a.v.samorodov, v.f. mironov. synthesis of triazolylisatins glycoconjugates and some ammonium hydrazones on their basis. russian journal of general chemistry 91 (2021) 1282-1291. https://doi.org/10.1134/s1070363221070045. [30] a.v. bogdanov, a.d. voloshina, a.r. khamatgalimov, n.v. terekhova, v.f. mironov. on the effect of the nature of substituents on the antimicrobial activity of water-soluble acylhydrazones on the isatin scaffold. doklady chemistry 494 (2020) 136-140. https://doi.org/10.1134/s0012500820090013. [31] a.d. voloshina, s.k. gumerova, a.s. sapunova, n.v. kulik, a.b. mirgorodskaya, a.a. kotenko, t.m. prokopyeva, v.a. mikhailov, l.ya. zakharova, o.g. sinyashin. the structure-activity correlation in the family of dicationic imidazolium surfactants: antimicrobial properties and cytotoxic effect. biochimica et biophysica acta (bba)-general subjects 1864 (2020) 129728. https://doi.org/10.1016/j.bbagen.2020.129728. [32] a.d. voloshina, a.s. sapunova, n.v. kulik, m.g. belenok, i.yu. strobykina, a.p. lyubina, s.k. gumerova, v.e. kataev. antimicrobial and cytotoxic effects of ammonium derivatives of diterpenoids steviol and isosteviol. bioorganic and medicinal chemistry 32 (2021) 115974. https://doi.org/10.1016/j.bmc.2020.115974. [33] g. born. aggregation of blood platelets by adenosine diphosphate and its reversal. nature 194 (1962) 927-929. https://doi.org/10.1038/194927b0. [34] c. pane, m. caputo, g. francese, g. manganiello, r. lo scalzo, g. mennella, m. zaccardelli. managing rhizoctonia damping-off of rocket (eruca sativa) seedlings by drench application of bioactive potato leaf phytochemical extracts. biology 9 (2020) 270. https://doi.org/10.3390/biology9090270. [35] a. moreno-gavira, f. dianez, b. sanchez-montesinos, m. santos. biocontrol effects of paecilomyces variotii against fungal plant diseases. journal of fungi 7 (2021) 415. https://doi.org/10.3390/jof7060415. [36] m. greener, drug safety on trial. embo reports 6 (2005) 202-204. https://doi.org/10.1038/sj.embor.7400353. [37] a.e. schultze, d.b. walker, j.r. turk, j.m. tarrant, m.b. brooks, s.d. pettit. current practices in preclinical drug development: gaps in hemostasis testing to assess risk of thromboembolic injury. toxicologic pathology 41 (2013) 445-453. https://doi.org/10.1177/0192623312460924. [38] h. abd-el-khair, t.g. abdel-gaied, m.s. mikhail, a.i. abdel-alim, h.i s. el-nasr. biological control of pectobacterium carotovorum subsp. carotovorum, the causal agent of bacterial soft rot in vegetables, in vitro and in vivo tests. bulletin of the national research centre 45 (2021) 37. https://doi.org/10.1186/s42269-021-00491-4. https://doi.org/10.1134/s1070363218010097 https://doi.org/10.1007/s00706-017-2049-y https://doi.org/10.1002/slct.201901708 https://doi.org/10.1016/j.molliq.2019.111220 https://doi.org/10.1016/j.molliq.2019.111220 https://doi.org/10.1134/s107042802003032x https://doi.org/10.1134/s1070363221070045 https://doi.org/10.1134/s1070363221070045 https://doi.org/10.1134/s0012500820090013 https://doi.org/10.1016/j.bbagen.2020.129728 https://doi.org/10.1016/j.bbagen.2020.129728 https://doi.org/10.1016/j.bmc.2020.115974 https://doi.org/10.1016/j.bmc.2020.115974 https://doi.org/10.1038/194927b0 https://doi.org/10.3390/biology9090270 https://doi.org/10.3390/jof7060415 https://doi.org/10.3390/jof7060415 https://doi.org/10.1038/sj.embor.7400353 https://doi.org/10.1038/sj.embor.7400353 https://doi.org/10.1177/0192623312460924 https://doi.org/10.1186/s42269-021-00491-4 admet & dmpk 10(2) (2022) 163-179 isatin-3-hydrazones: synthesis and biological activity doi: http://dx.doi.org/10.5599/admet.1179 179 [39] x. huang, j. ren, p. li, s. feng, p. dong, m. ren. potential of microbial endophytes to enhance the resistance to postharvest diseases of fruit and vegetables. journal of the science of food and agriculture 101 (2021) 1744-1757. https://doi.org/10.1002/jsfa.10829. [40] l. ragasova, e. penazova, f. gazdik, j. pecenka, j. cechova, r. pokluda, m. baranek, d. grzebelus, a. eichmeier. the change of bacterial spectrum after storage of x. campestris pv. campestris inoculated cabbage heads (brassica oleracea var. capitata l.). agronomy 10 (2020) 443. https://doi.org/10.3390/agronomy10030443. [41] m. ray, a. ray, s. dash, a. mishra, k.g. achary, s. nayak, s. singh. fungal disease detection in plants: traditional assays, novel diagnostic techniques and biosensors. biosensors and bioelectronics 87 (2017) 708-723. https://doi.org/10.1016/j.bios.2016.09.032. [42] t. wang, c. gao, y. cheng, z. li, j. chen, l. guo, j. xu. molecular diagnostics and detection of oomycetes on fiber crops. plants 9 (2020) 769. https://doi.org/10.3390/plants9060769. [43] k. mastanjevic, v. krstanovic, k. mastanjevic, b. sarkanj. malting and brewing industries encounter fusarium spp. related problems. fermentation 4 (2018) 3. https://doi.org/10.3390/fermentation4010003. [44] j.-h. han, g.-c. park, k.s. kim. antagonistic evaluation of chromobacterium sp. jh7 for biological control of ginseng root rot caused by cylindrocarpon destructans. mycobiology 45 (2017) 370-378. https://doi.org/10.5941/myco.2017.45.4.370. [45] e.m. chudinova, v.a. platonov, a.v. alexandrova, s.n. elansky. biology and resistance of phytopathogenic fungus ilyonectria crassa to fungicides. plant protection news 103 (2020) 192-196. https://doi.org/10.31993/2308-6459-2020-103-3-13431. ©2022 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://dx.doi.org/10.5599/admet.1179 https://doi.org/10.1002/jsfa.10829 https://doi.org/10.3390/agronomy10030443 https://doi.org/10.1016/j.bios.2016.09.032 https://doi.org/10.3390/plants9060769 https://doi.org/10.3390/fermentation4010003 https://doi.org/10.3390/fermentation4010003 https://doi.org/10.5941/myco.2017.45.4.370 https://doi.org/10.31993/2308-6459-2020-103-3-13431 http://creativecommons.org/licenses/by/3.0/ recent advances in electrochemical determination of anticancer drug 5-fluorouracil doi: https://doi.org/10.5599/admet.1711 135 admet & dmpk 11(2) (2023) 135-150; doi: https://doi.org/10.5599/admet.1711 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index review recent advances in electrochemical determination of anticancer drug 5-fluorouracil totka dodevska*, dobrin hadzhiev and ivan shterev department of organic chemistry and inorganic chemistry, university of food technologies, plovdiv, bulgaria *corresponding author: e-mail: dodevska@mail.bg received: february 15, 2023; revised: april 16, 2023; published: april 19, 2023 abstract reliable, rapid, highly selective and sensitive analytical methods for the determination of antineoplastic agent 5-fluorouracil (5-fu) in human body fluids (blood serum/plasma and urine) are required to improve the chemotherapy regimen to reduce its toxicity and improve efficacy. nowadays, electrochemical techniques provide a powerful analytical tool for 5-fu detection systems. this comprehensive review covers the advances in the development of electrochemical sensors for the quantitative determination of 5-fu, mainly focused on original studies reported from 2015 to date. we have summarized recent trends in the electrochemical sensor systems applied for the analysis of 5-fu in pharmaceutical formulations and biological samples, and critically evaluated the key performance metrics of these sensors (limit of detection, linear range, stability and recovery). challenges and future outlooks in this field have also been discussed. ©2023 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords electrochemical detection; electroanalysis; pharmaceutical electrochemistry; sensor; analysis; cancer introduction 5-fluorouracil (5-fu) is a gold-standard antineoplastic agent used in the chemotherapy of solid tumors (colorectal, breast, head, neck, stomach, pancreatic, and cervical cancer). antimetabolite drugs work by inhibiting essential biosynthetic processes or being incorporated into macromolecules, such as dna and rna, and inhibiting their normal function. the fluoropyrimidine 5-fu does both [1]. after intravenous bolus administration, approximately 80 % of 5-fu is catabolized in the liver into pharmacologically inactive metabolite 5-fluoro-5,6-dihydrouracil (5-fuh2) via dihydropyrimidine dehydrogenase (dpd), 5–20 % of 5-fu are excreted in the urine, and only 1–3 % contribute to the anabolism pathway responsible for clinical effects [2]. the mechanism of action of 5-fu is based on anabolic conversion into cytotoxic nucleotides using several pyrimidine metabolic pathway enzymes [3]. anabolic way produces 5-fluorouridine-5′-monophosphate (fump), 5-fluorouridine (5-furd), 5-fluoro-2′-deoxyuridine (5-fdurd) and their derivatives, responsible for 5-fu cytotoxicity in host and tumor cells [4]. the antitumor activity results from the inhibition of thymidylate synthase – an enzyme required for de novo pyrimidine synthesis. 5-fluoro-2′-deoxyuridine 5′-monophosphate (5-fdump) inhibits thymidylate synthase, resulting in the depletion of thymidine triphosphatase (ttp), one of the four nucleotide https://doi.org/10.5599/admet.1711 https://doi.org/10.5599/admet.1711 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:dodevska@mail.bg http://creativecommons.org/licenses/by/4.0/ t. dodevska et al. admet & dmpk 11(2) (2023) 135-150 136 triphosphates used in the in vivo synthesis of dna. other 5-fu metabolites incorporate into both rna and dna; incorporation into rna results in major effects on both rna processing and functions. the 5-fu-associated adverse effects include diarrhea, vomiting, nausea, pyrexia, neutropenia, pulmonary embolism, mucosal inflammation, asthenia, and a decrease of haemoglobin level [5]. leukopenia, leading to an immunocompromised state in these patients, can result in secondary pneumonia or sepsis. topical 5-fu has been used in the treatment of several premalignant and malignant skin diseases. it is used for treating multiple actinic or solar keratoses to prevent progression to squamous cell carcinoma in high-risk individuals. 5-fu is usually administered by application directly to the skin lesions as a 0.5 to 5 % cream or solution (common brand names: carac, efudex, fluoroplex, tolak) twice daily for 14 to 42 days. due to limited absorption, the systemic side effect profile of topically applied 5-fu is minimal. the most common adverse effects are localized skin irritation, erythema, crusting, and eczematous skin reactions. there is a large interindividual variability in the clinical outcome (efficacy and toxicity) of chemotherapy based on 5-fu. the effects of 5-fu are strongly dependent on the balance between anabolism and catabolism and, therefore, on the various factors involved in these two phases, such as the substrate concentration, the level of enzymes in the various tissues and their enzymatic activities [6]. therapeutic drug monitoring (tdm) is the measurement of drug concentrations in biological samples to individualize the drug dosage to improve drug efficacy and reduce related toxicities. tdm is especially important for patients with dpd deficiency, in which 5-fu treatment is the mainstay of treatment for their condition with no clinically efficacious alternatives. patients may have a partial or complete deficiency of dpd and be more likely to experience toxicity with 5-fu. conversely, some patients may also have increased dpd activity, which may result in reduced exposure to 5-fu [7]. over the past years, increasing efforts have been placed on optimizing 5-fu dosing to increase antitumor efficacy while minimizing drug-associated toxicity. there is growing evidence to show that 5-fu dosing based on plasma 5-fu drug level is feasible and that 5-fu tdm can improve clinical outcomes by improving the efficacy of 5-fu-based combination regimens and reducing toxicities [8]. casale et al. 2004 confirmed the relevance of the pharmacokinetic analysis of 5-fu main metabolites, especially 5-fuh2, to better understand the metabolism and improve the therapeutic efficacy [9]. the results from the commented study provided evidence that the 5-fuh2/5-fu ratio calculation should be used as reference values to decide and monitor optimal dose scheduling for the decrease of toxicity and the improvement of therapeutic efficacy. reliable analytical techniques for monitoring 5-fu concentrations in biological fluids after intravenous administration effectively improve the chemical agent's therapeutic index. here it should be noted that blood samples containing 5-fu require more complex handling than samples for most other drug analyses since 5fu continues to degrade in blood once the sample is taken. two approaches are used to stop enzymatic degradation: placing the sample immediately in ice and storing it on ice until further processed in the laboratory or adding a small amount of a dpd enzyme inhibitor immediately after collection [7]. this was observed to remove the need for immediate centrifugation and separation of plasma. in physiological samples of clinical patients, the mean concentration of 5-fu varies over a wide concentration range of 0.106 to 3.77 μg ml–1 (0.82 to 29.2 μm) in serum samples and as high as 10 to 60 μg ml–1 (77.4 to 464.2 μm) in urine samples [10]. the analytical methods developed for the determination of 5-fu in biological matrices have been reviewed extensively by breda & barattè [3]. in 2020 semail et al. presented an updated review on the analytical method development and validation of 5-fu and its metabolites in biological matrices [11]. the article primarily addressed the development of sample preparation methods, admet & dmpk 11(2) (2023) 135-150 electrochemical determination of 5-fluorouracil doi: https://doi.org/10.5599/admet.1711 137 analytical methodologies and the summary of sensitivity and recovery of the compounds from 1980 to february 2020. authors presented a review on chromatographic and electrophoretic systems, which are the most commonly used strategies for the analytical separation of 5-fu and its metabolites from samples. monitoring of 5-fu can be performed using high-performance liquid chromatography (hplc), gas chromatography-mass spectrometry (gc–ms) or liquid chromatography-mass spectrometry (lc–ms). hplc–uv detection is the most frequently reported technique for the determination of 5-fu in plasma, serum, whole blood, tissue and urine. comparing the operational parameters of different techniques for the determination of 5-fu in plasma, breda & barattè conclude that lc–ms is the most sensitive technique; gc-ms and gc are almost similar in sensitivity to lc-ms, but these approaches require a time-consuming, tedious extraction and derivatization steps [3]. generally, the described techniques needed a relatively large plasma volume (2.0 ml), which may not always be available, required sophisticated equipment and are not amenable to rapid and routine clinical assay. therefore, simple, rapid, cost-effective, easily miniaturized, highly selective, sensitive and properly validated analytical methods for the determination of 5-fu in human body fluids (blood serum/plasma and urine) are required to improve the chemotherapy. further, controlling the amount of 5-fu in commercial formulas is vital in the pharmaceutical industry quality control. the received data testify that the electrochemical techniques can offer the advantages mentioned above and provide a feasible path toward the next generation of sensing devices. to the best of our knowledge, there is no review article related to the electrochemical determination of 5-fu. the next parts of the manuscript present the most recent and innovative works on the fabrication of electrochemical sensor platforms and their application for 5-fu determination in biological and pharmaceutical samples. we have comprehensively reviewed the most promising sensor designs and successfully applied them for the electroanalysis of 5-fu, mainly focused on original studies reported from 2015 to date. basic detection principles of the most common electrochemical sensing techniques various electrochemical techniques – chronoamperometry, cyclic voltammetry (cv), linear sweep voltammetry (lsv), staircase voltammetry (scv), differential pulse voltammetry (dpv), square wave voltammetry (swv), and square wave adsorptive stripping voltammetry (swadsv) can be used as an efficient alternative, providing an affordable approach for an accurate, highly selective, sensitive, and fast quantitative determination of 5-fu. the basic detection principles of the most common electrochemical sensing techniques are summarized in table 1. table 1. basic detection principles of the most common electrochemical sensing techniques. electrochemical technique detection principle amperometry amperometry is based on the application of a constant potential to the working electrode, and the subsequent measurement of the current generated by the oxidation/reduction of the electroactive analyte. the resulting steady-state current is proportional to the bulk concentration of the analyte. cv the technique is based on varying the applied potential of the working electrode in both the forward (anodic sweep) and the reverse (cathodic sweep) directions while measuring the corresponding current. cv is often used to investigate the reversibility, reaction kinetics, electron stoichiometry of a system, the diffusion coefficient of an analyte, and the formal redox potential. lsv the current at a working electrode is measured while the potential between the working electrode and a reference electrode is swept linearly in time. scv the potential sweep is a series of stair steps; the current is measured at the end of each potential change, right before the next. thus the contribution to the current signal from the capacitive charging current is reduced. https://doi.org/10.5599/admet.1711 https://en.wikipedia.org/wiki/reference_electrode t. dodevska et al. admet & dmpk 11(2) (2023) 135-150 138 electrochemical technique detection principle dpv short pulses (in the range of 10 – 100 ms) with limited amplitude (up to 100 mv) are superimposed on a linear ramp. the current value is immediately measured before the pulse application and at the end of the pulse. the difference between the currents is recorded and plotted versus the potential. the procedure effectively reduces the capacitive current due to the direct current ramp. swv a symmetrical square-wave pulse is superimposed on a staircase wave. the duration of the pulse is equal to the length of the staircase, and the superposition is obtained in such a way that the forward pulse of the square wave coincides with the first half of that staircase. the first current is measured at the end of the forward square-wave pulse, and the second one is measured at the end of the return square-wave pulse. the signal is obtained as an intensity of the resulting differential current. the change in current between potential steps is plotted versus the potential. electrochemical sensors for 5-fu determination the general principle of electrochemical detection of 5-fu is illustrated in fig. 1. figure 1. illustrative representation of electrochemical detection of 5-fu. earlier electrochemical methods used for the quantitative determination of 5-fu include the use of mercury electrode. mirčeski et al. 2000 studied the redox reaction of 5-fu at a hanging mercury drop electrode (hmde) by means of swv [12]. the effect of the cu(ii) ions on the adsorptive swv response of 5-fu was also discussed from an analytical point of view. authors demonstrated that in the presence of cu(ii) ions, swv provides the determination of 5-fu below nano-molar concentration levels. 5-fu reacts with cu(ii) yielding a stable complex. the complex formed is adsorbed more effectively at the electrode surface, increasing the adsorptive stripping response of 5-fu. a linear calibration plot was established at 1-9×10−11 m with a correlation coefficient of r2 = 0.992 and an extremely low detection limit of 7.7×10−12 m was achieved. the development of new, cost-effective, efficient electrocatalysts plays a key role in the design of highperformance electrochemical sensing devices. the great potential of carbon-based electrodes as sensing platforms is exciting due to their unique electrical and chemical properties, such as a high surface-to-volume ratio, high electrical conductivity, chemical stability, biocompatibility, and robust mechanical strength [13,14]. carbon nanomaterials (carbon nanotubes, graphene, carbon quantum dots, carbon nanofibres, etc.) have exhibited such inherent features that can be easily exploited in the development of novel advanced technology for sensing applications [15,16]. carbon nanotubes (cnts) offer excellent properties that enable a wide range of applications in electrocatalysis. cnts may be comprised of a single graphitic layer or multiple coaxial layers, resulting in the formation of single-walled carbon nanotubes (swcnts) and multi-walled admet & dmpk 11(2) (2023) 135-150 electrochemical determination of 5-fluorouracil doi: https://doi.org/10.5599/admet.1711 139 carbon nanotubes (mwcnts). the variable surface morphology of cnts permits a great variety of surface functionalities for the development of reliable and stable electrochemical sensor platforms. the electrochemical sensing performance (sensitivity and selectivity) of the cnts can be modified by the covalent or non-covalent functionalization of cnts [15]. graphene is a two-dimensional (2-d) sheet of hexagonally arranged carbon atoms with sp2-hybridization. the combination of attributes such as an extremely large surface area (at 2630 m2 g-1, it is double that of swcnts), very large 2-d electrical conductivity, high mechanical strength, and low cost makes graphene an ideal platform for the anchoring of metal nanoparticles for electrochemical sensing applications [14,17]. there are excellent examples of applications of graphene decorated with catalyst nps for electrochemical sensing [17,18]. graphene oxide (go) contains oxygen functional groups, including hydroxyl, carboxylic, and epoxy groups, which demonstrate good catalytic activities, particularly for the oxidation of organic molecules. however, the presence of oxygen-containing groups disrupts the sp2-bonds of carbon rings, thereby lowering conductivity. this issue may be resolved by reducing graphene (electrochemically, thermally, or chemically) to generate reduced graphene oxide (rgo). this form of graphene exhibits higher conductivity due to the cleavage of hydroxyl and epoxy functional groups while retaining most of the carbonyl or carboxylic acid groups at its edges [19]. since the carbon nanomaterials-based sensors are unable to differentiate between the responses of organic compounds with the same functional group in their structures, the accurate quantification of the target analyte becomes difficult. in order to enhance the electrochemical sensing performance of carbon nanomaterials, a mixture or hybrid or composite of these nanomaterials with organic or inorganic compounds has to be used [15]. the addition of some other materials, such as nanoparticles of metals or metal oxides, functionalized nanostructures, as well as organometallic compounds, is an appropriate approach to further enhance the sensitivity and lower the limits of detection for the target analyte through synergistic effects. a high-quality review article [20] surveys state-of-the-art nanomaterials-based electrochemical sensors and biosensors for the detection and quantification of six classes of important pharmaceutical compounds, including anti-inflammatory, anti-depressant, anti-bacterial, anti-viral, anti-fungal, and anti-cancer drugs. a critical review focused on the recent advances in metal nanocomposite-based electrochemical sensors for pharmaceutical analysis was published in 2022 [21]. the authors have presented a comprehensive overview of representative metal nanocomposites with synergistic properties and their recent (2017–2022) application in the context of electrochemical sensing as a means of detecting specific antibiotic, tuberculostatic, analgesic, antineoplastic, anti-psychotic, and anti-hypertensive drugs. hua et al., 2013 reported the electrochemical behavior of 5-fu on a glassy carbon electrode (gce) modified with bromothymol blue (btb) and mwcnts [22]. quantitative analysis of 5-fu was performed in 0.2 m pbs (ph 6.8) using cyclic voltammetry (cv) and a linear relationship between the oxidation peak current and logarithmic concentration of 5-fu in the concentration range of 8×10−7 to 5×10−3 m was obtained (lod = 2.67×10−7 m). the results of the rsd (2.8 %) and recovery (98.6 %) showed that this proposed method can be used efficiently for the determination of 5-fu in the pharmaceutical formulation. the authors point out that the electrode surface was renewed electrochemically after each experiment to minimize electrode fouling and improve the electrochemical response. two years later, a fast and facile electrochemical sensor for the detection of ppb levels of 5-fu with an analysis time as low as 25 s was developed by satyanarayana et al. [10]. the sensing platform was fabricated using a gold nanoparticles-decorated multiwall carbon nanotubes composite modified gce. the electrochemical capability of the fabricated aunps-mwcnts-chit/gce for the detection of 5-fu is examined by cv, electrochemical impedance analysis and dpv. peak current of the dpvs exhibited a linear relationship https://doi.org/10.5599/admet.1711 t. dodevska et al. admet & dmpk 11(2) (2023) 135-150 140 over a wide concentration range of 0.03–10 μm with a low detection limit of 20 nm (2.6 ng ml–1). the recovery values obtained in these experiments varied from 98.3 to 102.6 % and are quite satisfactory. тhe authors have also provided data about the storage stability and reusability of the developed electrode. dpvs of 20 μm 5-fu were recorded at a single aunps-mwcnts-chit/gce electrode over a period of 7 days while storing the modified electrode at ambient laboratory conditions. the peak currents of the recorded dpvs decreased merely by 3.8 % in about 25 measurements. practical utility of the fabricated sensor has been demonstrated for the detection of 5-fu directly from artificial urine and pharmaceutical formulations with good recovery limits and clearly confirms that the developed sensor would be of biomedical interest. a glassy carbon electrode modified with a chemically reduced graphene oxide and chitosan (crgo/cs) composite film was constructed and applied to determine 5-fu using scv and swv techniques [23]. the sensor crgo/cs/gce showed a lower detection limit (1.24 nm for scv, and 4.93 nm for swv) and it was successfully applied to determine 5-fu in real samples. the proposed method was validated for the quantitative determination of 5-fu in pharmaceutical preparations. it was used to detect 5-fu in florac injection (50 mg ml–1) by applying scv using the standard addition method. additionaly, the applicability of the crgo/cs/gce was checked by analyzing 5-fu in spiked urine samples. in these studies, authors report high recovery values (97.3 to 100.7 %) and low rsd values (0.79–1.15 %), revealing the accuracy and precision of the proposed method. the electrochemical behavior of 5-fu was investigated at a self-made mwcnts-paraffin oil paste electrode using cv and dpv [24]. the oxidation was irreversible over the ph range studied (ph = 3.0 to 11.2) and exhibited a diffusion-controlled process. in the range of 10–7 to 5×10–6 m, the current measured by dpv presents a good linear property as a function of the concentration of 5-fu with a detection limit of 3.94×10–8 m. the proposed method was validated for the determination of 5-fu in pharmaceutical preparations in “aduracil” tablets as a real sample by applying dpv using the standard addition method and the results were in good agreement with the content marked in the label. the recovery test of 5-fu ranging from 4.0 μm to 1.0 μm was performed using dpv. the recoveries in different samples were found to lie in the range from 97.15 to 101.82 %, with rsd = 2.09 %. here it should also be noted that during the successive cyclic voltammetric sweeps, a significant decrease in the oxidation peak current was noticed. this phenomenon may be attributed to the fouling of the electrode surface due to the adsorption of the oxidation products and severely affects the analytical performance in terms of sensitivity and linearity of the electrode signal. authors noted that the current response of the prepared electrode would decrease after successive use and the electrode should be prepared again. the established feature would limit the use of this sensor for 5-fu analysis. therefore, future efforts should be focused on appropriate reverse electrode passivation materials to reduce the effect of fouling. carbon paste electrodes (cpes) are widely applicable in electrochemical studies. cpes have multiple beneficial features and practical advantages including excellent electric properties, low cost, easy preparation, simple regeneration of their active surface and possibilities of miniaturization. feasibility of incorporating different substances during the paste preparation makes cpe a strong tool to evaluate different modifiers, building a useful platform for many electrochemical applications. methylene blue (mb) is a cationic dye whose electrochemical properties are well known. mb-modified cpe as a sensor for the voltammetric investigation of 5-fu was reported by bukkitgar & shetti [25]. due to the excellent property of mb as a redox mediator, improved and enhanced results were obtained compared to the unmodified electrode. the peak currents registered with dpv, were linear with the concentration of 5-fu in the range of 10−7 to 4×10−5 m and the detection limit was calculated to be 2.04 nm. simple procedure for preparation and regeneration of surface with good reproducibility, high sensitivity, and low detection limit implied the applicability of the proposed electrode for 5-fu determination. under the optimum experimental admet & dmpk 11(2) (2023) 135-150 electrochemical determination of 5-fluorouracil doi: https://doi.org/10.5599/admet.1711 141 conditions, the authors evaluated the effects of potential interferents on the voltammetric response of 0.1 mm 5-fu. the experimental results showed that 100-folds of citric acid, gum acacia, oxalic acid, sucrose, xanthine, uric acid and urea did not interfere with the voltammetric signal of 5-fu. however, when caffeine and ascorbic acid were used, peak potential showed remarkable variation, indicating that these substances will have an interfering effect on the quantitative detection of 5-fu. nanostructures of lanthanides (or f-block metallic elements) have interesting electrocatalytic properties and can be used as modifiers in sensors. cpe modified with praseodymium erbium tungstate nanoparticles (pr:er nps) was also investigated for the voltammetric detection of 5-fu [26]. an irreversible 5-fu oxidation peak was registered at 1.0 v (vs. ag/agcl) in 0.01 m pbs (ph 7.0). a linear calibration was achieved in the range of 0.01 to 50 μm (lod = 0.98 nm) using swv. the authors stated that the sensor provided good precision and accuracy. the practical applicability of developed electrode material pr:er/cpe to detect selectively 5-fu in spiked urine and plasma samples was tested. the results obtained from the developed sensor satisfactorily agreed with hplc, indicating the reliability of the electrochemical method. researchers have used various types of ionic liquids (ils) in electrode fabrication with the intention of developing novel electrochemical sensor platforms or as unique electrocatalytic substances, as they possess many attractive properties: sophisticated conductivity, low volatility, high chemical stability, and wide electrochemical windows [27]. therefore, ils can be used as modifiers to make a new kind of modified carbon paste electrodes (il-cpe), which have exhibited advantages including improved electrocatalytic ability, antifouling effect and good stability. for the first time, the electrochemical oxidation of 5-fu on an il-cpe was investigated by zhan et al. [28] il-cpe was fabricated using 1-butylpyridinium hexafluorophosphate (bppf6) as the modifier. quantitative analysis of 5-fu was performed in b-r buffer solution (ph 7.0) using differential pulse voltammetry and two linear ranges (5×10−7 to 2×10−6 m and 2×10−6 to 8×10−4 m) were obtained; lod was found to be 1.3×10−8 m. the electrode was stored at room temperature when not in use, it could remain 95.9 % of its initial responses after storage for 15 days, indicating that il-cpe had good storage stability. furthermore, to demonstrate the feasibility of the il-cpe, the amounts of 5-fu in commercial injection samples were tested by dpv, which showed good recovery (in the range from 97.2 to 102.1 %) and practical applicability. combined nanomaterials and ils have synergistic effects and result in improved conductivity, active sites, and accelerated electron transfer rate. in these cases, multiple-amplifying voltammetric response could be determined and increased sensitive analyte detection would be possible. a novel sensitive electrochemical approach was developed by incorporating graphene quantum dots (gqd) and ionic liquid 1-butylpyridinium bromide (bpbr) in the fabrication of a carbon paste electrode (gqd/bpbr/cpe) [29]. the applicability of the gqd/bpbr/cpe in the voltammetric analysis of 5-fu was evaluated. in square wave voltammetry analysis, the gqd/bpbr/cpe showed high selectivity and sensitivity for 5-fu over a wide linear range of 0.001 to 400 μμ. owing to the synergic effect of gqd and bpbr, an extremely low detection limit of 0.5 nμ (s/n=3) was achieved. the oxidation current of 5-fu at gqd/bpbr/cpe showed 92.3 % of its original signal after 14 days, confirming the good stability of the modified electrode. the gqd/bpbr/cpe was successfully applied for the determination of 5-fu in pharmaceutical samples and acceptable results were obtained. the composition of carbon nanotubes, copper(ii) oxide nanoparticles and ionic liquid loaded on cpe, which acted as part of highly sensitive and selective electrode for the analysis of two anticancer drugs, was described by fouladgar [30,31]. the electrode was fabricated using 1-ethyl-3-methylimidazolium tetrafluoroborate (emitfb). doxorubicin and 5-fu can be oxidized at potentials of approximately 0.68 and 0.97 v (vs. ag/agcl) on emitfb/cuo-swcnts/cpe surface that is sufficient for the simultaneous determination of https://doi.org/10.5599/admet.1711 t. dodevska et al. admet & dmpk 11(2) (2023) 135-150 142 drugs in a mixture via swv [30]. the relationship between current and concentration was linear within the range 160 to 760 μm (lod = 6.0 nm) for doxorubicin and 20–180 μm (lod = 0.4 μm) for 5-fu. the sensitivities of the modified electrode toward doxorubicin in the absence and presence of 5-fu were virtually similar, thus indicating that the oxidation of doxorubicin and 5-fu at the modified electrode are independent and simultaneous or independent measurement of the two compounds are possible. the application of emitfb/cuo-swcnts/cpe for the quantitative analysis of doxorubicin and 5-fu in real samples (serum and commercial pharmaceutical samples), produced favourable outcomes. one year later, roushani et al. reported cunps/mwcnts/il/chit/gce, fabricated using 1-methyl-3-octylimidazolium tetrafluoroborate ionic liquid [32]. the sensor has similar characteristics – the peak current of the dpv exhibited a linear relationship against 5-fu over a concentration range of 1–110 μm with a detection limit of 0.15 μm. the modified electrode demonstrated satisfactory long-term stability – it retains 95 % of the initial response after two weeks storage and 93 % after 30 days, respectively. using the standard addition method the concentration of 5-fu in an artificially prepared specimen, by adding known amounts of 5-fu to serum samples, was measured. the obtained recovery (89.05–104.9 %) reveals the potential capability of the method for the determination of 5-fu in human blood serum samples. recently, green nanotechnology has remained at the forefront of scientific research due to its outstanding approaches and applications. green nanotechnology involves the application of green chemistry principles to the design of valuable and sustainable nano-sized materials in a more environmentally benign approach. the so-called “green synthesis”, using mild reaction conditions and natural resources as plant extracts, has received more attention as a cost-effective and valuable alternative for environmentally safe and energyefficient production of metal nanoparticles [33]. there is convincing evidence that bio-inspired synthesis of metal nanoparticles has the potential to provide a new direction in the fabrication of novel, cheap and effective electrocatalytic materials applicable to electroanalysis. lima et al., 2018 reported for the first time utilization of porphyran (pfr) (a sulfated polysaccharide extracted from red seaweed) as a capping, stabilizing, and reducing agent for the cost-effective and environmentally safe synthesis of agnps and their applicability for the development of an electrochemical sensor for 5-fu determination [34]. the fabricated electrochemical platform based on aunps-pfr/cpe, combined with the dpv technique, exhibits a linear current response in the concentration range from 29.9 to 234.4 µm, lod of 0.66 µm, and loq of 2.22 µm, respectively. the good analytical performance of the sensor was confirmed for determining 5-fu in pharmaceutical formulation (commercial injection solution), with good recoveries (96.6 to 101.4%) and an acceptable relative standard deviation (rsd=2.80 %). unfortunately, the authors do not provide data on the long-term storage stability of the modified electrode. in conclusion, although these synthetic methods restrict the use of toxic chemicals, energy and sophisticated instruments, the experimental data show that the electrode surfaces modified with bio-synthesized metal nanoparticles still remain challenging as they are not as stable and reproducible as one would hope. further research needs to be done to improve the electrode performance in terms of operational and storage stability of the electrode-catalysts modified with nanoparticles synthesized in a green way. a novel class of chitosan-based hybrid semi-interpenetrating nanogel catalyst by the reductant-free synthesis of aunps within the chitosan-poly(methacrylic acid) (au@cs-pmaa) polymer network was reported [35]. the covalently cross-linked au@cs-pmaa nanogel with cs chains semi-ipn in the pmaa crosslinked network shows superior colloidal stability, increased catalytic activity and proved to be an effective electrochemical sensing platform for 5-fu. under optimal conditions, the constructed sensor au@cspmaa/gce offered good analytical performance in the determination of 5-fu with a wide linear range from admet & dmpk 11(2) (2023) 135-150 electrochemical determination of 5-fluorouracil doi: https://doi.org/10.5599/admet.1711 143 0.1–497 μm and a detection limit of 0.03 μm. in addition, the peak current response of the au@cspmaa/gce retains 92.5 % of the initial peak current value after storage of 30 days, revealing good long-term stability of the modified electrode. dpv measurements in diluted human blood serum samples prove the practical applicability of the sensor. here it should be pointed out that some drawbacks, such as low homogeneity and reproducibility of the resulting modified surface, could characterize the drop-casting approach used in [10,22,35]. the influence of the “coffee ring” effect can alter the distribution of the nanoparticles drop-casted on an electrode. however, voltammetry requires the formation of uniformly modified surfaces for which the coffee ring and its related effects present a significant limitation on the reproducibility of the drop-casted surfaces [36]. moreover, the modifier that has been physically adsorbed onto the electrode surface may be gradually stripped off in longterm measurements. thermo-sensitive conductive polymer microgels are unique materials applicable in the fields of electronics, medical electrodes, soft robotics, and electrochemical energy storage devices. among all stimulating responsive polymer microgels, thermo-sensitive poly(n-isopropylacrylamide) (pnipam) attracts an extensive range of scientists interested in the field of sensors due to its lower critical solution temperature (around 32 °c), close to the human body temperature. various multifunctional pnipam systems have revolutionized biomedical fields, such as controlled drug delivery, tissue engineering, and self-healing materials [37]. in the field of electroanalysis, thermo-sensitive conductive polymer forming a charged smart membrane by cycling “on-off” over the surface of the electrode allows for a “command interface” that selectively enhances or inhibits the electrochemical reaction [38]. a new type of switch-like temperature-controlled 5-fu electrochemical sensor based on the adaptable thermosensitive conductive microgel consisting of pnipam and poly(3,4-ethylenedioxythiophene) (pedot) was developed [39]. chen et al. constructed a smart electrochemical 5-fu sensor with temperature control using the adaptable conductive polymer microgel pnipam-pedot containing electrically conductive and thermoresponsive properties. the resulting conductive microgel exhibits a thermo-reversible conversion by controlling the internal temperature in the range of 20–40 °c, owing to the contracted and expanded configuration of the pnipam. the sensor was successfully used to achieve thermo-switch-control for detection of 5-fu for the first time. an effectively switched the electrochemical process by tuning the solution temperature and achieved sensitive quantitative determination of 5-fu. the pnipam-pedot/gce has good selectivity and reproducibility for 5-fu detection. furthermore, the sensor demonstrated satisfactory results toward 5-fu detection in real human blood serum samples. authors concluded that the presented novel switchlike electrochemical sensor offers an innovative notion for the application of thermo-responsive polymers. ratiometric electrochemical sensors (recss) have attracted extensive attention because of recss significantly reduce the interference of the inherent background electrical signals, as well as possess great advantages in terms of enhanced accuracy, repeatability and reliability. recss are capable of overcoming the system errors of the conventional electrochemical sensors derived from the alteration of the environment and operating personnel [40]. the recs possesses dual electrochemical signals and the quantitative detection of the target analyte is based on the ratio of these two signals. by applying the ratiometric method, a built-in correction coefficient is created to reduce the influence of general interfering species. hatamluyi et al., 2021 reported a ratiometric electrochemical sensor based on gce modified with pdau/mwcnts-rgo nanocomposite for simultaneous quantification of irinotecan (iri) and 5-fu in both real samples and aqueous solutions [41]. introduction of pd-au/mwcnt-rgo significantly improved the speed of electron transport, specific surface area, and electrical catalytic ability of the sensing system due to the https://doi.org/10.5599/admet.1711 t. dodevska et al. admet & dmpk 11(2) (2023) 135-150 144 synergistic effect of pd-au bimetallic nanoparticles and mwcnt-rgo hybrid structure. the assay strategy was based on the use of a fixed concentration of ferrocene (fc) added to the electrolyte solution (robinson buffer solution). fc is a built-in effective internal reference to achieve ratiometric detection of iri and 5-fu. dpv, registered after the addition of iri and 5-fu, clearly shows three separated and independent characteristic peaks appeared at +0.2, +0.58, and +1.17 v (vs. ag/agcl, sat. kcl), which were ascribed to the oxidation of fc, iri, and 5-fu, respectively. under optimal conditions, a linear relationship between the iri and 5-fu concentrations and ratiometric peak current intensities (iiri/ifc) and (i5-fu/ifc) was obtained. by measuring the (iiri/ifc) and (i5-fu/ifc) as the response signals, simultaneous detection of iri and 5-fu were realized with excellent selectivity and reproducibility in contrast with the single-signal method. based on the developed ratiometric method, sensitive and reliable detection of 5-fu has been realized in the dynamic range 0.05 to 75 μm (lod = 9.4 nm). the authors have provided data on the long-term stability of pd–au/mwcntrgo/gce. six electrodes from the same batch were stored at 4 °c for 6 weeks and applied every week. after 42 days, the original value of i5-fu/ifc exhibited negligible changes – the rsd was estimated <2.1 %, representing acceptable stability. in the analysis of real biological samples (injection solution forms of iri and 5-fu, human urine and serum samples), the sensor possesses high detection recoveries and low rsds. the advanced electrochemical sensor platforms for the detection of 5-fu and their applications in the clinical and pharmaceutical fields are listed in table 2. considering the serious impacts of 5-fu on human health, the here presented novel electrochemical sensors is expected to hold great potential for fast, simple, and on-site 5-fu detection in blood serum and pharmaceutical formulations. table 2. comparison of linear range and lod values for 5-fu determination using electrochemical sensor platforms. electrode method linear range limit of detection real sample ref. aunps-mwcnts-cs/gce dpv 0.03 10−6 m 210−8 m artificial urine, pharmaceutical formulations [10] hmde swv 1 9×10−11 m 7.710−12 m – [12] mwcnts/btb/gce cv 810−7 510−3 m 2.6710−7 m pharmaceutical formulation [22] crgo/cs/gce scv 0.01 0.15 μm 1.24 nm urine [23] mb/cpe dpv 1×10−7 4×10−5 m 2.04 nm urine [25] pr:er/cpe swv 0.01 50 μm 0.98 nm blood serum, urine [26] il-cpe dpv 510−7 810−4 m 1.310−8 m pharmaceutical formulation [28] gqd/bpbr/cpe swv 0.001 400 μμ 0.5 nμ pharmaceutical formulation [29] emitfb/cuo-swcnts/cpe swv 20 180 μm 0.4 µm pharmaceutical formulation, blood serum [30] cunps/mwcnts/il/cs/gce dpv 1 110 μm 0.15 μm blood serum [32] aunps-pfr/cpe dpv 29.9 234.4 µm 0.66 µm pharmaceutical formulation [34] au@cs-pmaa/gce dpv 0.1 497 μm 0.03 μm blood serum [35] pnipam-pedot/gce dpv 0.03 182 μm 15 nm blood serum [39] pd–au/mwcnt-rgo/gce dpv 0.05 75 μm 9.4 nm pharmaceutical formulation, urine, blood serum [41] aunp-spe swv 0.2 50 µg ml−1 0.1 µg ml−1 pharmaceutical formulation [42] p(bcp)/dsdna/gce dpv 1 50 mg l−1 0.31 mg l−1 pharmaceutical formulation [43] znfe2o4/mnps/il/cpe swv 0.1 1400 μm 0.07 µm pharmaceutical formulation, urine [44] admet & dmpk 11(2) (2023) 135-150 electrochemical determination of 5-fluorouracil doi: https://doi.org/10.5599/admet.1711 145 electrode method linear range limit of detection real sample ref. agnps@panints/pge dpv 1.0–300 µm 0.06 µm blood serum [45] gqds-pani/zno-ncs/gce dpv 0.1 50 µm 0.023 µm pharmaceutical formulation, blood serum, urine [46] cufe2o4 nps/spge dpv 0.1 270 µm 0.03 µm pharmaceutical formulation, urine [47] btb – bromothymol blue; hmde – hanging mercury drop electrode; spe – screen-printed electrode; gqd – graphene quantum dots; bpbr – 1-butylpyridinium bromide; pnipam – poly(n-isopropylacrylamide); pedot – poly(3,4-ethylenedioxythiophene); pmaa – poly(methacrylic acid); crgo – chemically reduced graphene oxide; cs – chitosan; pfr – porphyran; mb – methylene blue; p(bcp) – poly(bromocresol purple); dsdna – fish sperm double strand dna; mnps – magnetic nanoparticles; panints – polyaniline nanotubes; pge – pencil graphite electrode; ncs – nanocomposites; emitfb – 1-ethyl-3-methylimidazolium tetrafluoroborate. electrochemical sensors – advantages and limitations electrochemical sensor platforms are ideally suited for rapid and reliable point-of-care sensing of 5-fu. the methods for 5-fu detection have advantages over the spectrophotometric and chromatographic methods mostly because of the ease of conducting experimental work and the cost-effectiveness of the instrumentation. chromatography-based techniques (gc, hplc) and hyphenated approaches (hplc-uv, hplc-ms, gc-ms), the most recent techniques used for detection of 5-fu, have require relatively great investment and expertise, as well as complicated working operation including multistage sample preparation – tedious extraction and derivatization procedures. the high hydrophilicity of 5-fu and its active metabolites hinders their isolation from biological matrices. the high aqueous solubility and the low solubility to organic solvents result in extraction difficulties [48,49]. the most applied techniques to remove interfering endogenous substrates, namely liquid-liquid extraction (lle), solid-phase extraction (spe) and protein precipitation (pp), have been thoroughly reviewed [49]. the main drawback of such techniques is high consumption of organic solvents. extraction usually employs the acidification of serum sample as well as pp (using trichloroacetic acid, mixtures of methanol and water, or mixtures containing acetonitrile) followed by lle (using different solvents such as propanol, methanol, ethyl acetate, diethyl ether, or chloromethane). currently, more attention is being paid to the development of efficient and environmentally friendly microextraction methods that consume a much lower amount of organic solvents. moreover, ms detection is incompatible with microdialysis samples in terms of signal stability because the dialysis salts can induce high background noise. the upkeep of accessory instruments is expensive and the high cost makes the ms detector, coupled with hplc and gc, less widely used. here it must be emphasized that the differences in sample pre-treatment, experimental conditions and instrumental related parameters can cause variations in 5-fu concentrations detected. in contrast, the electroanalytical techniques are quick, simple and offer the possibility of direct analysis. electrochemical sensing is the only method that allows the determination in an aqueous medium without derivatization of the analyte. it should be considered that electrochemical sensors have an enormous linear dynamic range of more than six orders of magnitude and exhibit a high degree of selectivity. the possibility of regenerating the catalytic active electrode surface allows numerous analyses in succession. unlike their spectroscopic and chromatographic counterparts, electrochemical sensors may be incorporated into robust, portable or miniaturized devices while remaining relatively inexpensive and simple compared to spectrophotometric detectors or mass spectrometry. table 3 highlights the advantages and disadvantages of electrochemical sensor devices. although there are still some drawbacks to using this type of equipment, advantages exceed shortcomings. https://doi.org/10.5599/admet.1711 t. dodevska et al. admet & dmpk 11(2) (2023) 135-150 146 table 3. summary of advantages and disadvantages of electrochemical sensors. advantages selectivity and sensitivity electroanalytical techniques are powerful, highly sensitive and selective. the use of specific electrode-catalyst and precisely defined value of working potentials allow matrix interference to be avoided. therefore, the electrochemical sensor platforms offer a remarkable sensitivity for trace levels determination. repeatability and accuracy electrochemical sensors possess an excellent repeatability and accuracy. once calibrated to a known concentration, the sensor device will provide an accurate measurement of a target analyte, that is repeatable. low sample volume and minimum pretreatment low sample volume requirement with no prior treatment of the samples is another big advantage. the principle of operation of the electrochemical sensors allows procedures of the sample preparation to be kept to a minimum. the electrochemical method only requires appropriate sample dilution followed by immediate analysis. fast response important aspect is that electroanalysis is much simpler and requires less time in comparison with the more commonly used analytical methods. the electrochemical sensors provide a rapid response within a few seconds. inherent miniaturization electrochemical sensor is an ideal, attractive candidate for miniaturized analytical system, as it has high compatibility with modern microand nanofabrication technologies. due to the ability to easily integrate electrodes in microchips, electrochemical sensor is actually easy to be miniaturized. moreover, the analytical sensitivity is not compromised by miniaturization. low cost and power requirements ultra-low power consumption, easy-to-carry instruments and cost-effectiveness are among the most important features of electrochemical sensors. real time and on-site detection the electrochemical sensor is a system that simultaneously meets all the key requirements for an on-site sensing. wireless network most of the advanced electrochemical sensors are equipped with the technology allowing them to be used as a part of a wireless network, i.e. they can be connected to laptops, tablets or smartphones. disadvantages non-specific adsorption and biofouling the non-specific adsorption of proteins on the electrode surface is persistent problem that negatively affects electrochemical sensors. this phenomenon occurs because of physisorption and can decrease the sensor’s performance – it results in higher background signal and affects the reliability of sensors. on the other hand, electrochemical reaction products tend to accumulate at the electrode surface leading to loss of catalytic activity and hence the sensor response decreases with time. ph-dependence of electrode signal since the proton (h+) participates directly in the electrochemical process, ph is a key factor affecting the signal intensity. therefore the sample analysis should be performed in a buffer solution at an optimal ph-value. device-to-device reproducibility the sensing performance is highly dependent on the electrode surface morphology and may differ from device-to-device even though the catalytic-active electrode material originates from the same fabrication protocol. electrochemical sensors for 5-fu demonstrate irreplaceable advantages in biomedical and clinical applications, including easy-to-operate, rapid response, excellent detection limits and ability to handle small sample volumes. thus, the electrochemical sensors for 5-fu have a high potential to outcompete traditional methods in the near future. conclusion remarks and outlook the outstanding features of electrochemical sensor devices have brought many prospects in clinical and pharmaceutical applications. high sensitivity, selectivity, portability and affordable investment provide reliable alternatives to the conventional analytical methods used in clinical labs. moreover, electrochemical sensors hold great promise for enabling clinical analysis of 5-fu at the point-of-care. transforming an electrochemical sensor system into a simple, easy-to-use, and portable device from labscale research is the driving force for the commercialisation of sensor systems. although most of the electrochemical sensors discussed here were tested in real samples analysis, there is a considerable gap admet & dmpk 11(2) (2023) 135-150 electrochemical determination of 5-fluorouracil doi: https://doi.org/10.5599/admet.1711 147 between the laboratory tests and the fabrication of commercial sensing devices for practical applications. there are still shortcomings and further research is required for their overall improvement. some issues, such as unsatisfactory long-term stability, electrode fouling and the non-specific adsorption of other species, make the proposed sensors more difficult to apply commercially. therefore, novel strategies for electrode signal amplification are needed to be developed. here we summarize the most promising approaches for improving the electrochemical determination of 5-fu: • to achieve reliable measurements, the effects of fouling can be sufficiently minimized by incorporating the electrochemical activation of the electrode surface at regular intervals as a part of the analysis procedure. • measuring in flowing systems. the use of flow injection analysis (or batch injection analysis) systems combined with modified screen-printed electrodes will substantially reduce the contact between the electrode and fouling agents. additionally, in all these systems, products/intermediates of the electrochemical reaction are removed (washed away) from the working electrode, thus minimizing their deposition on the electrode surface. • application of protecting films. the working electrodes can be modified with protein, polymeric or nonpolymeric films in order to provide a physical barrier between the fouling agents and the electrode surface. • development of novel electrode materials more resistant toward passivation. • the use of boron-doped diamond (bdd) electrodes is still unexplored in this area of research. bdd is a conductive material effective for real-time detection because the sp3-hybridized structure of bdd is resistant to biofouling and biocompatible with organisms. in comparison with other electrode materials (gc, graphite, pt and au), bdd possesses a higher resistance toward biofouling, the largest electrochemical potential window for aqueous media, as well as a low and stable background current, which is attributed to the low capacitance of the bdd. all these properties make bdd suitable for in vivo real-time determination of 5-fu. additionally, it is essential to develop and validate electrochemical sensors for the simultaneous quantification of 5-fu and its metabolites in human plasma. for clinical practice, the method could successfully improve the effectiveness and security of the chemotherapy regimen. we are convinced that the joint work of interdisciplinary research groups will contribute to more recent commercialization of high-performance electroanalytical devices for 5-fu determination. conflict of interest: the author declares no conflict of interest. references [1] d. longley, d. harkin, p. johnston. 5-fluorouracil: mechanisms of action and clinical strategies. nature reviews cancer 3 (2003) 330-338. https://doi.org/10.1038/nrc1074 [2] g.d. heggie, j.p. sommadossi, d.s. cross, w.j. huster, r.b. diasio. clinical pharmacokinetics of 5fluorouracil and its metabolites in plasma, urine, and bile. cancer research 47 (1987) 2203-2206. https://aacrjournals.org/cancerres/article-pdf/47/8/2203/2430813/cr0470082203.pdf [3] m. breda, s. barattè. a review of analytical methods for the determination of 5-fluorouracil in biological matrices. analytical and bioanalytical chemistry 397(3) (2010) 1191-1201. https://doi.org/ 10.1007/s00216-010-3633-8 [4] r.b. diasio, b.e. harris. clinical pharmacology of 5-fluorouracil. clinical pharmacokinetics 16(4) (1989) 215-237. https://doi.org/10.2165/00003088-198916040-00002 [5] k. kadoyama, i. miki, t. tamura, j.b. brown, t. sakaeda, y. okuno. adverse event profiles of 5fluorouracil and capecitabine: data mining of the public version of the fda adverse event reporting https://doi.org/10.5599/admet.1711 https://doi.org/10.1038/nrc1074 https://aacrjournals.org/cancerres/article-pdf/47/8/2203/2430813/cr0470082203.pdf https://doi.org/10.1007/s00216-010-3633-8 https://doi.org/10.1007/s00216-010-3633-8 https://doi.org/10.2165/00003088-198916040-00002 t. dodevska et al. admet & dmpk 11(2) (2023) 135-150 148 system, aers, and reproducibility of clinical observations. international journal of medical sciences 9(1) (2012) 33-39. https://doi.org/10.7150/ijms.9.33 [6] m. chalabi-dchar, t. fenouil, c. machon, a. vincent, f. catez, v. marcel, h.c. mertani, j.-c. saurin, p. bouvet, j. guitton, n. d. venezia, j.-j. diaz. a novel view on an old drug, 5-fluorouracil: an unexpected rna modifier with intriguing impact on cancer cell fate. nar cancer 3(3) (2021) zcab032. https://doi.org/10.1093/narcan/zcab032 [7] j.j. schneider, p. galettis, j.h. martin. overcoming barriers to implementing precision dosing with 5fluorouracil and capecitabine. british journal of clinical pharmacology 87 (2021) 317-325. https:// doi.org/10.1111/bcp.14723 [8] j.j. lee, j.h. beumer, e. chu. therapeutic drug monitoring of 5-fluorouracil. cancer chemotherapy and pharmacology 78(3) (2016) 447-464. https://doi.org/10.1007/s00280-016-3054-2 [9] f. casale, r. canaparo, l. serpe, e. muntoni, c.d. pepa, m. costa, l. mairone, g.p. zara, g. fornari, m. eandi. plasma concentrations of 5-fluorouracil and its metabolites in colon cancer patients. pharmacological research 50(2) (2004) 173-179. https://doi.org/10.1016/j.phrs.2004.01.006 [10] m. satyanarayana, k.y. goud, k.k. reddy, k.v. gobi. biopolymer stabilized nanogold particles on carbon nanotube support as sensing platform for electrochemical detection of 5-fluorouracil invitro. electrochimica acta 178 (2015) 608-616. http://dx.doi.org/10.1016/j.electacta.2015.08.036 [11] n.-f. semail, a.s.a. keyon, b. saad, s.s. noordin, n.n.s.n.m. kamal, n.n.m. zain, j. azizi, s. kamaruzaman, n. yahaya. analytical method development and validation of anticancer agent, 5fluorouracil, and its metabolites in biological matrices: an updated review. journal of liquid chromatography & related technologies 43 (2020) 562-579. https://doi.org/10.1080/10826076.20 20.1781654 [12] v. mirčeski, r. gulaboski, b. jordanoski, š. komorsky-lovrić. square-wave voltammetry of 5fluorouracil. journal of electroanalytical chemistry 490(1–2) (2000) 37-47. https://doi.org/10.1016/ s0022-0728(00)00203-5 [13] p.a. pushpanjali, j.g. manjunatha, n. hareesha. an overview of recent developments of carbonbased sensors for the analysis of drug molecules. journal of electrochemical science and engineering 11(3) (2021) 161-177. https://doi.org/10.5599/jese.999 [14] b.r. adhikari, m. govindhan, a. chen. carbon nanomaterials based electrochemical sensors/biosensors for the sensitive detection of pharmaceutical and biological compounds. sensors (basel) 15(9) (2015) 22490-508. https://doi.org/10.3390/s150922490 [15] r. kour, s. arya, s.-j. young, v. gupta, p. bandhoria, a. khosla. review—recent advances in carbon nanomaterials as electrochemical biosensors. journal of the electrochemical society 167(3) (2020) 037555. https://doi.org/10.1149/1945-7111/ab6bc4 [16] a.c. power, b. gorey, s. chandra, j. chapman. carbon nanomaterials and their application to electrochemical sensors: a review. nanotechnology reviews 7(1) (2018) 19-41. https://doi.org/ 10.1515/ntrev-2017-0160 [17] m. pumera, a. ambrosi, a. bonanni, e.l.k. chng, h.l. poh. graphene for electrochemical sensing and biosensing. trac trends in analytical chemistry 29(9) (2010) 954-965. https://doi.org/10.1016/ j.trac.2010.05.011 [18] e.b. bahadır, m.k sezgintürk. applications of graphene in electrochemical sensing and biosensing. trac trends in analytical chemistry 76 (2016) 1-14. https://doi.org/10.1016/j.trac.2015.07.008 [19] l. qian, a.r. thiruppathi, r. elmahdy, j. van der zalm, a. chen. graphene-oxide-based electrochemical sensors for the sensitive detection of pharmaceutical drug naproxen. sensors 20 (2020) 1252. https://doi.org/10.3390/s20051252 [20] l. qian, s. durairaj, s. prins, a. chen. nanomaterial-based electrochemical sensors and biosensors for the detection of pharmaceutical compounds. biosensors and bioelectronics 175 (2021) 112836. https://doi.org/10.1016/j.bios.2020.112836 https://doi.org/10.7150/ijms.9.33 https://doi.org/10.1093/narcan/zcab032 https://doi.org/‌10.1111/bcp.14723 https://doi.org/‌10.1111/bcp.14723 https://doi.org/10.1007/s00280-016-3054-2 https://doi.org/10.1016/j.phrs.2004.01.006 http://dx.doi.org/10.1016/j.electacta.2015.08.036 https://doi.org/10.1080/‌10826076.20‌20.1781654 https://doi.org/10.1080/‌10826076.20‌20.1781654 https://doi.org/10.1016/s0022-0728(00)00203-5 https://doi.org/10.1016/s0022-0728(00)00203-5 https://doi.org/10.5599/jese.999 https://doi.org/10.3390/s150922490 https://doi.org/10.1149/1945-7111/ab6bc4 https://doi.org/10.1515/ntrev-2017-0160 https://doi.org/10.1515/ntrev-2017-0160 https://doi.org/10.1016/j.trac.2010.05.011 https://doi.org/10.1016/j.trac.2010.05.011 https://doi.org/10.1016/j.trac.2015.07.008 https://doi.org/10.3390/s20051252 https://doi.org/10.1016/j.bios.2020.112836 admet & dmpk 11(2) (2023) 135-150 electrochemical determination of 5-fluorouracil doi: https://doi.org/10.5599/admet.1711 149 [21] y. fang, h. chang, j. li, z. li, d. zhang. recent advances in metal nanocomposite-based electrochemical (bio)sensors for pharmaceutical analysis. critical reviews in analytical chemistry (2022) 1-27. https://doi.org/10.1080/10408347.2022.2128633 [22] x. hua, x. hou, x. gong, g. shen. electrochemical behavior of 5-fluorouracil on a glassy carbon electrode modified with bromothymol blue and multi-walled carbon nanotubes. analytical methods 5 (2013) 2470. http://dx.doi.org/10.1039/c3ay40149a [23] v.p. pattar, s.t. nandibewoor. electroanalytical method for the determination of 5-fluorouracil using a reduced graphene oxide/chitosan modified sensor. rsc advances 5 (2015) 34292-34301. https:// doi.org/10.1039/c5ra04396d [24] ]. c. a. jyothi, n. p. shetti, s. t. nandibewoor. development of voltammetric method for the determination of an anticancer drug, 5-flurouracil at a multi-walled carbon nanotubes paste electrode. synthesis and reactivity in inorganic, metal-organic, and nano-metal chemistry 46 (2016) 814-820. https://doi.org/10.1080/15533174.2014.989586 [25] s.d. bukkitgar, n.p. shetti. electrochemical behavior of an anticancer drug 5-fluorouracil at methylene blue modified carbon paste electrode. materials science and engineering c 65 (2016) 262268. http://dx.doi.org/10.1016/j.msec.2016.04.045 [26] m. rahimi-nasrabadi, f. ahmadi, h. beigizadeh, m. s. karimi, a. sobhani-nasab, y. joseph, h. ehrlich, m. r. ganjali. a modified sensitive carbon paste electrode for 5-fluorouracil based using a composite of praseodymium erbium tungstate. microchemical journal 154 (2020) 104654. https://doi.org/ 10.1016/j.microc.2020.104654 [27] r. chokkareddy, t. niranjan, g.g. redhi. chapter 13 ionic liquid based electrochemical sensors and their applications, editor(s): a.m.a. inamuddin, s. kanchi. green sustainable process for chemical and environmental engineering and science, elsevier (2020) 367-387. isbn 9780128173862. https://doi.org/10.1016/b978-0-12-817386-2.00013-5 [28] t. zhan, l. cao, w. sun, w. hou. electrochemical behavior of 5-fluoro-1h-pyrimidine-2 on an ionic liquid modified carbon paste electrode. analytical methods 3 (2011) 2651. http://dx.doi.org/ 10.1039/c1ay05454f [29] r. emamian, m. ebrahimi, h. karimi-maleh. a sensitive sensor for nano-molar detection of 5fluorouracil by modifying a paste sensor with graphene quantum dots and an ionic liquid. journal of nanostructures 10(2) (2020) 230-238. https://doi.org/10.22052/jns.2020.02.004 [30] m. fouladgar. cuo-cnt nanocomposite/ionic liquid modified sensor as new breast anticancer approach for determination of doxorubicin and 5-fluorouracil drugs. journal of the electrochemical society 165(13) (2018) b559-b564. https://doi.org/10.1149/2.1001811jes [31] m. fouladgar. a novel electrochemical cuo-nanostructure platform for simultaneous determination of 6-thioguanine and 5-fluorouracil anticancer drugs. acta chimica slovenica 67 (2020) 701-709. https://doi.org/10.17344/acsi.2019.4986 [32] m. roushani, z. saeidi, s. hemati, m. hosseini. highly sensitive electrochemical determination of 5fluorouracil using cunps/mwcnt/il/chit composite modified glassy carbon electrode. advances in nanochemistry 1 (2019) 73-77. https://doi.org/10.22126/anc.2019.11282 [33] t. dodevska, d. hadzhiev, i. shterev, y. lazarova. application of biosynthesized metal nanoparticles in electrochemical sensors: review. journal of the serbian chemical society 87 (2022) 401-435. https://doi.org/10.2298/jsc200521077d [34] d. lima, g.n. calac, a.g. viana, c.a. pessôa. porphyran-capped gold nanoparticles modified carbon paste electrode: a simple and efficient electrochemical sensor for the sensitive determination of 5-fluorouracil. applied surface science 427 (2018) 742-753. http://dx.doi.org/10.1016/j.apsusc.2017.08.228 [35] s.k.d. vishnu, p. ranganathan, s.-p. rwei, c. pattamapromc, t. kavitha, p. sarojini. new reductantfree synthesis of gold nanoparticles-doped chitosanbased semi-ipn nanogel: a robust nanoreactor for exclusively sensitive 5-fluorouracil sensor. international journal of biological macromolecules 148 (2020) 79-88. https://doi.org/10.1016/j.ijbiomac.2020.01.108 https://doi.org/10.5599/admet.1711 https://doi.org/10.1080/10408347.2022.2128633 http://dx.doi.org/10.1039/c3ay40149a https://doi.org/10.1039/c5ra04396d https://doi.org/10.1039/c5ra04396d https://doi.org/10.1080/15533174.2014.989586 http://dx.doi.org/10.1016/j.msec.2016.04.045 https://doi.org/10.1016/j.microc.2020.104654 https://doi.org/10.1016/j.microc.2020.104654 https://doi.org/10.1016/b978-0-12-817386-2.00013-5 http://dx.doi.org/10.1039/c1ay05454f http://dx.doi.org/10.1039/c1ay05454f https://doi.org/10.22052/jns.2020.02.004 https://doi.org/10.1149/2.1001811jes https://doi.org/10.17344/acsi.2019.4986 https://doi.org/10.22126/anc.2019.11282 https://doi.org/10.2298/jsc200521077d http://dx.doi.org/10.1016/j.apsusc.2017.08.228 https://doi.org/10.1016/j.ijbiomac.2020.01.108 t. dodevska et al. admet & dmpk 11(2) (2023) 135-150 150 [36] a. k. s. kumar, y. zhang, d. li, r. g. compton. a mini-review: how reliable is the drop casting technique? electrochemistry communications 121 (2020) 106867. https://doi.org/10.1016/j.ele com.2020.106867 [37] e.m. frazar, r.a. shah, t.d. dziubla, j.z. hilt. multifunctional temperature-responsive polymers as advanced biomaterials and beyond. journal of applied polymer science 137 (2019) 48770. https:// doi.org/10.1002/app.48770 [38] c. chen, p. zhao, c. li, y. xie, j. fei. highly sensitive temperature-responsive sensor based on pspdea-ps/c60-mwcnts for reversible switch detection of catechol. electroanalysis 31 (2019) 913. https://doi.org/10.1002/elan.201800769 [39] b. mutharani, p. ranganathan, s.-m. chen. temperature-reversible switched antineoplastic drug 5fluorouracil electrochemical sensor based on adaptable thermo-sensitive microgel encapsulated pedot. sensors & actuators: b. chemical 304 (2020) 127361. https://doi.org/10.1016/j.snb.20 19.127361 [40] t. dodevska, d. hadzhiev, i. shterev. a review on electrochemical microsensors for ascorbic acid detection: clinical, pharmaceutical, and food safety applications. micromachines 14 (2023) 41. https://doi.org/10.3390/mi14010041 [41] b. hatamluyi, r. sadeghian, t.s.s. belin, i. alipourfard, m. rezayi. dual-signaling electrochemical ratiometric strategy for simultaneous quantification of anticancer drugs. talanta 234 (2021) 122662. https://doi.org/10.1016/j.talanta.2021.122662 [42] s. wang, s. fu, h. ding. determination of 5-fluorouracil using disposable gold nanoparticles modified screen-printed electrode. sensor letters 10 (2012) 974-978. http://dx.doi.org/10.1166/ sl.2012.2341 [43] d.k. zeybek, b. demir, b. zeybek, ş. pekyardımcı. a sensitive electrochemical dna biosensor for antineoplastic drug 5-fluorouracil based on glassy carbon electrode modified with poly(bromocresolpurple). talanta 144 (2015) 793-800. http://dx.doi.org/10.1016/j.talanta.2015.06.077 [44] a.f. shojaeia, k. tabatabaeiana, s. shakerib, f. karim. a novel 5-fluorouracile anticancer drug sensor based on znfe2o4 magnetic nanoparticles ionic liquids carbon paste electrode. sensors and actuators b 230 (2016) 607-614. http://dx.doi.org/10.1016/j.snb.2016.02.082 [45] f.m. zahed, b. hatamluyi, f. lorestani, z. es’haghi. silver nanoparticles decorated polyaniline nanocomposite basedelectrochemical sensor for the determination of anticancer drug 5-fluorouracil. journal of pharmaceutical and biomedical analysis 161 (2018) 12-19. https://doi.org/10.1016/j.jp ba.2018.08.004 [46] b. hatamluyi, z. es'haghi, f. m. zahed, m. darroudi. a novel electrochemical sensor based on gqdspani/zno-ncs modified glassy carbon electrode for simultaneous determination of irinotecan and 5-fluorouracil in biological samples. sensors & actuators: b. chemical 286 (2019) 540-549. https://doi.org/10.1016/j.snb.2019.02.017 [47] p. m. jahani, m. jafari, f. n. ravari. cufe2o4 nanoparticles-based electrochemical sensor for sensitive determination of the anticancer drug 5-fluorouracil. admet and dmpk (2023) https://doi.org/ 10.5599/admet.1691 [48] p. mavromatis, k. stampouli, a. vliora, a. mayilyan, v. samanidou, m. touraki. development of an hplc-dad method for the extraction and quantification of 5-fluorouracil, uracil, and 5 fluorodeoxyuridin monophosphate in cells and culture media of lactococcus lactis. separations 9 (2022) 376. https://doi.org/10.3390/separations9110376 [49] j. e. knikman, h. rosing, h-j guchelaar, a. cats, j. h. beijnen. a review of the bioanalytical methods for the quantitative determination of capecitabine and its metabolites in biological matrices. biomedical chromatography 34 (2020) e4732. https://doi.org/10.1002/bmc.4732 ©2023 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) https://doi.org/10.1016/j.elecom.2020.106867 https://doi.org/10.1016/j.elecom.2020.106867 https://doi.org/10.1002/app.48770 https://doi.org/10.1002/app.48770 https://doi.org/10.1002/elan.201800769 https://doi.org/10.1016/j.snb.2019.127361 https://doi.org/10.1016/j.snb.2019.127361 https://doi.org/10.3390/mi14010041 https://doi.org/10.1016/j.talanta.2021.122662 http://dx.doi.org/10.1166/sl.2012.2341 http://dx.doi.org/10.1166/sl.2012.2341 http://dx.doi.org/10.1016/j.talanta.2015.06.077 http://dx.doi.org/10.1016/j.snb.2016.02.082 https://doi.org/10.1016/j.jpba.2018.08.004 https://doi.org/10.1016/j.jpba.2018.08.004 https://doi.org/10.1016/j.snb.2019.02.017 https://doi.org/10.5599/admet.1691 https://doi.org/10.5599/admet.1691 https://doi.org/10.3390/separations9110376 https://doi.org/10.1002/bmc.4732 http://creativecommons.org/licenses/by/3.0/ recent advances in nanomaterials-based electrochemical sensors for tramadol analysis doi: https://doi.org/10.5599/admet.1593 117 admet & dmpk 11(2) (2023) 117-134; doi: https://doi.org/10.5599/admet.1593 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index review recent advances in nanomaterials-based electrochemical sensors for tramadol analysis farideh mousazadeh1, yar-mohammad baghelani2* and shamsi rahimi3 1 department of chemistry, payame noor university, tehran, iran 2 department of chemistry, faculty of sciences, ilam university, ilam, iran 3 department of chemistry, payame noor university, tehran, iran *corresponding author: e-mail: baghelani.2020@gmail.com received: november 09, 2022; revised: april 07, 2023; published: april 16, 2023 abstract tramadol is a centrally-acting analgesic used for treating moderate to severe acute and chronic pain. pain is an unpleasant sensation that occurs most commonly as a result of tissue injury. tramadol possesses agonist actions at the μ-opioid receptor and effects reuptake at the noradrenergic and serotonergic systems. in the last years, several analytical procedures have been published in the literature for the determination of tramadol from pharmaceutical formulations and biological matrices. electrochemical methods have attracted tremendous attention for the quantification of this drug owing to their demonstrated potential for quick response, real-time measurements, elevated selectivity and sensitivity. in this review, we highlighted the recent advances and applications of nanomaterials-based electrochemical sensors for the analysis and detection of tramadol, which is extremely important for the indication of effective diagnoses and for quality control analyses in order to protect human health. also, the main challenges in developing nanomaterials based electrochemical sensors for the determination of tramadol will be discussed. at last, this review offers prospects for the future research and development needed for modified electrode sensing technology for the detection of tramadol. ©2023 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords tramadol; analytical procedures; nanomaterials; electrochemical sensors introduction nowadays, the rapid improvements in the medical and pharmaceutical fields increase the diversity and use of drugs. however, problems such as the use of multiple or combined drugs in the treatment of diseases and the insensible use of over-the-counter drugs have caused concerns about the side-effect profiles and therapeutic ranges of drugs and environmental contamination and pollution problems due to pharmaceutical waste. therefore, the analysis of drugs in various media, such as biological, pharmaceutical, and environmental samples, is an important topic of discussion [1]. pain is an unpleasant sensation occurring majorly due to tissue damage, influenced by thinking, outlook, behaviour, and community factors. it is responsible for causing emotional and psychological discomfort, too [2]. long-term use and misuse of opioids causes serious health problems [3,4]. https://doi.org/10.5599/admet.1593 https://doi.org/10.5599/admet.1593 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:baghelani.2020@gmail.com http://creativecommons.org/licenses/by/4.0/ y. m. baghelani et al. admet & dmpk 11(2) (2023) 117-134 118 opioids bind to and activate opioid receptors on cells located in many areas of the brain, spinal cord, and other organs in the body, especially those involved in feelings of pain and pleasure. when opioids attach to these receptors, they block pain signals sent from the brain to the body and reduce the pain sensation. although a majority of opioids, i.e., morphine, oxycodone, methadone, ketobemidone, tapentadol, tramadol, fentanyl, hydromorphone, sufentanil, buprenorphine and codeine differ in chemical structure, physicochemical properties and in pharmacokinetics, they have one common feature, which is their interaction with the μ opioid receptor as the primary target. despite this similarity, large differences in the clinical responses (efficacy, effectiveness, toxicity and safety) are seen between the classes of opioids [5]. opium addicts suffering from opioid withdrawal symptoms are usually managed with tramadol in the initial stage [6]. tramadol is a 2-(dimethylamino)-methyl-1-(3-methoxyphenyl) cyclohexanol hydrochloride and a 4phenyl-piperidine analog of opioid drug codeine, which was discovered and synthesized in 1962 by a german company for pain treatment. it was introduced by the name tramadol into the market in 1977 and became available in the us market in 1995 [7]. opioid-category drugs such as morphine are very potent analgesic but are used less due to their adverse effects of respiratory depression, addiction, dependency, and constipation. the analgesic potency of tramadol is found to be ten times lesser than morphine but is preferred, being safe than the latter. tramadol is considered safe, as it does not cause respiratory depression and addiction compared to other opioid analgesics [8]. additionally, tramadol, when administered by the parenteral route, has less abuse potential. tramadol acts in two different mechanisms. it binds to the μopioid receptor, its affinity is weak, and the drug is able to inhibit the reuptake of serotonin and norepinephrine [9]. this drug can be taken by injection, suppository and orally. most of the reported side effects of tramadol have happened with intramuscular injection and in recent years, its use has increased in society [10]. the tramadol overdose can cause dizziness, vomiting, nausea, diarrhea, chills, hallucinations and respiratory problems since it is considered a toxic material in nature [11]. therefore, a sensitive method for the measurement of samples containing tramadol is of absolute importance for therapeutic and pharmaceutical reasons. analytical chemistry methods refer to techniques used for the detection, identification, characterization and quantification of chemical compounds [12-15]. these methods are commonly used in biology for research, development and quality control of pharmaceutical products. quantitative determination of tramadol in pharmaceutical and biological samples has been investigated by some instrumental methods such as gas chromatography (gc) [16] gas chromatography-mass spectrometry (gc-ms) [17,18], high-performance liquid chromatography (hplc) [19], capillary electrophoresis [20,21], spectrophotometry [22,23], chemiluminescence [24] and electrochemical methods [25]. however, most of these methods require highly trained technicians and complex sample preparation that hinders their wide application. among these methods, researchers considerably focused on electrochemical methods due to the respective benefits such as fast analysis time, higher selectivity, low cost, simplicity, and ability to be miniaturized [26-32]. it is well-known that conventional electrodes have some serious issues due to their sluggish surface kinetics, which severely affects the sensitivity and selectivity of the electrodes. the analytes on the conventional electrodes usually display a broad peak and generally, no peak appears at lower concentrations. the slow electrode reaction of analytes on the bare conventional electrode surfaces requires the high potential to proceed with the reaction at high rates, which greatly exceeds their formal redox potential. the kinetically hindered electrode reactions require a suitable electrocatalyst that can fasten the electrochemical reaction and lower its redox potential. it is obvious that electrode material can play a crucial role in the https://www.sciencedirect.com/topics/medicine-and-dentistry/nociception https://www.sciencedirect.com/topics/medicine-and-dentistry/opioid-dependence https://www.sciencedirect.com/topics/medicine-and-dentistry/anodyne https://www.sciencedirect.com/topics/medicine-and-dentistry/respiration-depression https://www.sciencedirect.com/science/article/pii/s0753332218373694#bib0055 admet & dmpk 11(2) (2023) 117-134 electrochemical analysis of tramadol doi: https://doi.org/10.5599/admet.1593 119 construction of high-performance electrochemical sensors [33-38]. the researchers are continuously putting their efforts into improving the sensitivity and selectivity of electrochemical sensors. nanotechnology refers to the branch of science and engineering devoted to designing, producing, and using structures, devices, and systems by manipulating atoms and molecules at the nanoscale, i.e., having one or more dimensions of the order of 100 nanometres or less. nanotechnology is used to revolutionize many technology and industry sectors: electronics, energy, material science, medicine, transportation, transportation, and environmental science, to name a few [39-44]. the reasons for their growing applications relate to their unique physical and chemical properties, which are entirely different from their bulk counterparts [45]. recently more attention has been paid to the synthesis of various nanomaterials to achieve not only ultrasensitive but selective electrochemical sensors [46]. electrochemical sensors and biosensors integrated with nanostructured materials were employed as powerful analytical devices owing to numerous advantages such as rapid response, high performance, cost-effective, high sensitivity and selectivity by signal amplification not only via the catalytic activity and high conductivity but also by facilitating the immobilization of chemical and biological reagents on the sensor surface [47-49]. the development of nanomaterials has proven fundamental for the development of electrochemical sensors to be used in different application fields such as biomedical, environmental, and food analysis [50-52]. in this review, we mainly focused on the current progress and application of nanomaterial-enabled electrochemical sensors for tramadol detection. here, we classify the current research based on the types of nanomaterials such as carbon nanomaterials including graphene (gr) and its derivatives, carbon nanotubes (cnts), and graphitic carbon nitride), metal nanoparticles (nps) and metal oxide nps) used in electrochemical tramadol sensors. also, the challenges and prospects were discussed in this review. electrochemical sensors based on nanomaterials for detection of tramadol electrochemical sensors based on carbon nanomaterials carbon-based nanomaterials exhibit distinct physical and chemical properties besides displaying great potential for applications in material preparation, energy storage, environmental science, pharmaceutical analyses, and medical science [53,54]. also, carbon nanomaterials as the most widely studied nanomaterials due to outstanding properties, such as unique chemical/physical stability, high heat resistance and corrosion resistance, wide electrochemical windows, large specific surface area, and ultra-high electrical conductivity, playing an extremely important role in the construction of electrochemical sensors for electroanalysis [55]. electrochemical sensors based on different functions and structures of carbon nanomaterials for electrochemical detection have been established over recent years, including cnts, gr and its derivatives, graphitic carbon nitride, etc. electrochemical sensors based on (gr) and its derivatives since the experimental discovery of gr in 2004 by novoselov et al. [56], it has been extensively employed in various fields [57-59]. gr is composed of single-layer carbon atoms forming a two-dimensional (2d) crystal structure. the inert electrochemistry, wide potential window, and electrocatalytic properties of this 2d material make it a promising candidate as a sensing element in electrochemical sensors to detect various target analytes. besides, the high surface area to volume ratio of gr improves the sensitivity of such electrochemical sensors [60]. https://doi.org/10.5599/admet.1593 https://ec.europa.eu/health/scientific_committees/opinions_layman/glossary/mno/nanotechnology.htm https://ec.europa.eu/health/scientific_committees/opinions_layman/glossary/abc/atom.htm https://ec.europa.eu/health/scientific_committees/opinions_layman/glossary/mno/molecule.htm https://ec.europa.eu/health/scientific_committees/opinions_layman/glossary/mno/nanoscale.htm https://ec.europa.eu/health/scientific_committees/opinions_layman/glossary/mno/nanometre.htm y. m. baghelani et al. admet & dmpk 11(2) (2023) 117-134 120 afkhami et al. [61] reported the preparation of nife2o4/gr nanocomposite and its application as a modifier for the fabrication of an electrochemical sensor for the simultaneous determination of tramadol and acetaminophen. the decoration of gr with nife2o4 nps can provide excellent electrochemical platforms for tramadol and acetaminophen analysis due to the combination of the enlarged active surface area, strong adsorptive capability of the nanomaterial and their specific interactions ability. by using square wave voltammetry (swv), the peak currents of tramadol and acetaminophen increased linearly with their concentration in the range of 0.01 to 9 µm. the detection limit for their determination was found to be 0.0036 and 0.0030 µm, respectively. the capability of the proposed sensor nife2o4-gr modified carbon paste electrode (cpe) was investigated by direct analysis of tramadol and acetaminophen in the commercial pharmaceutical samples (acetaminophen tablet, tramadol tablet, and ultracet tablet) and also in the biological fluids (human serum, and urine). good results for real sample analysis were obtained. manokaran et al. [62] functionalized nitrogen-doped gr using poly(diallyldimethyl ammonium chloride) (pdda-ngr). then, they anchored pt-pd bimetallic nps on pdda-ngr to form pt-pd/pdda-ngr nanocomposite. the simultaneous determination of tramadol and acetaminophen was carried out using ptpd/pdda-ngr modified glassy carbon electrode (gce). two well-defined voltammetric peaks were obtained in swv measurements. the modified electrode detected acetaminophen over a wide linear concentration range from 5.0 to 100.0 μm and tramadol from 12.0 to 240.0 μm. the detection limits were found to be 0.18 and 5.7 μm for acetaminophen and tramadol, respectively. finally, the amounts of acetaminophen and tramadol present in the human serum sample were analyzed by standard addition method using the ptpd/pdda-ngr modified gce. mohamed et al. [63] constructed a voltammetric sensor based on graphene oxide (go) and multiwalled carbon nanotubes (mwcnts) composites (go-mwcnts) for the sensitive determination of tramadol [64]. the electroanalytical sensitivity for the determination of tramadol using the go-mwcnts/cpe sensor was significantly improved over that of an unmodified cpe. the linear response obtained for tramadol using the go-mwcnts/cpe was found to be over the range of 2.0×10-9 to 1.1×10-3 m with good linearity and high correlation (0.9996). also, the limits of detection and quantification were found to be 1.50×10-10 m and 4.99×10-10 m, respectively. in addition, the go-mwcnts/cpe sensor enabled the sensing of tramadol in the presence of the frequently co-formulated drug ketorolac tromethamine and acetaminophen without any interference. rokhsefid et al. [64] prepared a selective and sensitive electrochemical sensor by modifying cpe with au nps-gr nanosheets nanocomposite and 4-hydroxyl-2-(triphenylphosphonio)phenolate (htp). the modified electrode (au nps-gr-htp/cpe) showed a successful application for determining tramadol and simultaneous detection of tramadol and acetaminophene. by modification of cpe with au nps-gr-htp, the oxidation of tramadol was significantly improved. according to differential pulse voltammetry (dpv) measurements, au nps-gr-htp/cpe showed a wide linear dynamic range (1.0–16.0 µm, and 16.0–100.0 μm) for the determination of tramadol. the detection limit of 0.82 µm for tramadol was obtained. in addition, the analytical performance of the prepared sensor was assessed to detect tramadol and acetaminophen in human urine and pharmaceutical samples with acceptable outputs. the simultaneous determination of tramadol, codeine and caffeine by using ceo2-sno2/rgo nanocomposite-modified gce was reported by hosseini et al. [65]. the ceo2-sno2/rgo/gce could shift the oxidation potential of tramadol, codeine and caffeine toward a less positivepotential; however, a small amount, but also boosted their oxidation peak current remarkablywhen compared to rgo/gce, sno2-rgo/gce, and bare gce. the simultaneous determination of three analytes was performed on ceo2-sno2/rgo/gce. in optimum admet & dmpk 11(2) (2023) 117-134 electrochemical analysis of tramadol doi: https://doi.org/10.5599/admet.1593 121 conditions, a dynamic range of 0.008-10 μm and 10-270 μm for tramadol, 0.01-12 μm and 12-260 μm for codeine, 0.01-14 μm and 14-260 μm for caffeine with the detection limit of 0.0056, 0.0053, and 0.0055 μm for tramadol, codeine and caffeine, respectively, were obtained. finally, the presented electrode was successfully applied for the measurement of tramadol, codeine and caffeine in urine and human plasma spiked samples. saichanapan et al. [66] reported the voltammetric determination of tramadol using a hierarchical graphene oxide nanoplatelets modified gce (h-gonps/gce). the h-gonps/gce showed a faster charge transfer rate and larger active surface area. the anodic current response of tramadol was three times higher at the h-gonps/gce than at the gonps/gce. in the optimal condition, by using adsorptive anodic stripping voltammetric (adasv), the calibration curve of tramadol demonstrated good linearity in two tramadol concentration ranges (0.05–5.0 μm, and 5.0–20 μm). a higher detection sensitivity (20.7 μa μm-1 cm-2) and a lower limit of detection (0.015 μm) were obtained. finally, the developed sensor was applied to detect tramadol in pharmaceutical samples and spiked beverage, saliva, and urine samples. electrochemical sensors based on cnts cnts are nanomaterials formed from one or more gr sheets rolled cylindrically from their axis, forming tubular structures with a diameter in the nanometer and a length range ranging from micrometers to centimeters [67]. conceptually, cnts can be divided into two groups: single-walled carbon nanotubes (swcnts) and mwcnts. swcnts are made up of only one sheet of gr and have a diameter ranging from 1 to 5 nm. however, the synthesis methods currently employed to produce only a small fraction of swcnt increase its cost and make it difficult to be applied on a large scale. mwcnts are formed by a set of two or more concentrically coiled gr sheets, which may have diameters of 10–50 nm [68,69]. recently, cnts have been applied in various fields [70-73]. while they have many of the same properties as other types of carbon, cnts offer unique advantages, including enhanced electronic properties, a large edge plane/basal plane ratio, and rapid electrode kinetics. therefore, cnt-based sensors generally have higher sensitivities, lower limits of detection, and faster electron transfer kinetics than traditional carbon electrodes [74]. babaei et al. [75] constructed a chemically modified electrode based on mwcnts-modified gce. the prepared modified electrode was used for the simultaneous determination of acetaminophen and tramadol. they showed that the application of mwcnts increases anodic peak currents for both acetaminophen and tramadol on the electrode surface. the results indicated that the use of mwcnts/gce allows the simultaneous determination of acetaminophen and tramadol with good sensitivity and selectivity. there was a linear relationship between the oxidation peak current and the concentration of acetaminophen over the range of 0.5 to 210 μm and a linear relationship between the oxidation peak current and the concentration of tramadol over the range of 2 to 300 μm. also, the mwcnts/gce exhibited a low limit of detection for acetaminophen (0.085 μm) and tramadol (0.361 μm). the analytical performance of this sensor has been evaluated for the detection of acetaminophen and tramadol in human serum, human urine and some pharmaceutical preparations with satisfactory results. s. branch [76] introduced a voltammetric sensor based on mwcnts-modified gce for the determination of tramadol. the mwcnts/gce facilitated the determination of tramadol with good sensitivity and selectivity. the dpv experiments of various concentrations of tramadol showed two linear dynamic ranges. the first linear dynamic range was from 4 µm to 35 µm, and the second linear dynamic range was between 60 µm to 550 µm. a detection limit of 0.38 µm was obtained. finally, the proposed sensor was used in the determination of tramadol in some real samples like a human serum, urine and some drugs, without the necessity of sample pretreatments or time-consuming extraction, with satisfactory results. https://doi.org/10.5599/admet.1593 y. m. baghelani et al. admet & dmpk 11(2) (2023) 117-134 122 atta et al. [77] fabricated the modified gce by electrodeposition of mono-dispersed au nps onto cnts. the electroanalysis of tramadol was investigated at the surface of the modified electrode by different electrochemical techniques. the increase of the electroactive surface area and the synergistic electrocatalytic activity were achieved by combining aunps with cnts responsible for the modified gce's improved performance. the modified electrode showed a low detection limit (68 nm) and wide linear concentration range (0.1–1000 μm) for the detection of tramadol concentration. tramadol was successfully determined in pharmaceutical dosage forms without any pretreatment of the samples. an amplified sensor based on improved cpe with 1,3-dipropylimidazolium bromide and mgo/swcnts nanocomposite (1,3-di-br/mgo/swcnts/cpe) for tradamol determination was prepared by hosseini et al. [78]. the 1,3-dibr/mgo/swcnts/cpe showed good catalytic ability for the determination of tramadol. swv investigation showed a linear relationship between the tramadol current and concentration within the range of 0.05-280 μm with a detection limit of 8.0 nm. finally, the modified electrode was successfully used for determination of tramadol in the injection and urine samples. foroughi et al. [79] developed an electrochemical sensor based on la3+ doped fern-like cuo nanoleaves/mwcnts modified gce for voltammetric determination of tramadol. fern-like la3+-cuo nanoleaves and mwcnts significantly improved the sensitivity of the sensor. a limit detection of 0.014 μm within a linear range of 0.5-900.0 μm was determined for obtaining the quantitative tramadol detection. finally, the developed sensor was utilized to determine tramadol and acetaminophen in pharmaceutical formulations and urine samples, with successful results. atta et al. [80] modified a cpe with co3o4 nps, ionic liquid (il) and cnts in the presence of sodium dodecyl sulfate (sds). the modified electrode (cntsilco3o4-sds/cpe) was applied for the determination of nalbuphine and tramadol narcotic analgesic drugs [42]. they evaluated the effect of the type of metal oxide on the catalytic activity of the sensor. nano-cobalt oxide showed the highest electrocatalytic activity, excellent conductivity, antifouling ability and charge transfer enhancement compared to other studied metal oxides. under the optimized conditions, simultaneous determination of nalbuphine and tramadol in human urine using the proposed sensor was successfully achieved with sub-nano detection limits of 0.58 nm and 0.62 nm, respectively. finally, the practical analytical performance of the sensor was studied for the determination of nalbuphine and tramadol in their pharmaceutical samples with satisfied recovery results. a voltammetric platform based on nps of antimony oxide (sb2o3 nps) and mwcnts was reported by çidem et al. [81]. the prepared nanocomposite was used for the modification of a gce to the electrooxidation of tramadol. high catalytic activity was observed towards the oxidation of tramadol using the proposed voltammetric platform. the peak current exhibited a linear dependence on the concentration of tramadol in a dynamic range of 4×10-8–3.0×10-5 m with a detection limit of 9.5×10-9 m. finally, the modified electrode was used for the sensitive determination of tramadol in pharmaceuticals. tavana et al. [82] synthesized pt-pd-doped nio nps decorated at the surface of swcnts (pt-pd/nio nps/swcnts) using a simple chemical precipitation method and characterized by various methods. moreover, a highly sensitive electroanalytical sensor was fabricated by incorporating synthesized pt-pd/nio nps/swcnt nanocomposites into a cpe in the presence of 1-ethyl-3-methylimidazolium methane sulfonate (emich3so3-) as a binder. the pt-pd/nio nps/swcnts/emich3so3-/cpe showed a powerful electrocatalytic activity for the electrooxidation of nalbuphine and tramadol. the pt-pd/nio nps/swcnts/emich3so3-/cpe also showed good catalytic activity for the determination of nalbuphine in the presence of tramadol and the oxidation potential of these drugs separated with δe = 460 mv. the pt-pd/nio nps/swcnts/emich3so3-/cpe was used to determine nalbuphine with a detection limit of 0.9 nm and tramadol with a detection limit of 50.0 nm in drug samples. admet & dmpk 11(2) (2023) 117-134 electrochemical analysis of tramadol doi: https://doi.org/10.5599/admet.1593 123 the development of a new voltammetric sensor based on molecularly imprinted poly(acrylic acid)mwcnt nanocomposite (mip-mwcnt) drop coated onto gce was reported by ricardo teixeira tarley and co-workers [83]. the prepared sensor was applied to tramadol determination in pharmaceutical samples. the voltammetric sensor prepared by suspension of mip-mwcnt at 1:1 (w/w) ratio shows an improved performance compared to unmodified gce. the peak currents were proportional to the tramadol concentration in the range of 9.0 to 30.0 μm. thus, the calculated limit of detection and limit of quantification to the proposed sensor were 1.4 μm and 4.8 μm, respectively. the modification of a pencil graphite electrode (pge) with mwcnts capped au nps for electrochemical determination of tramadol was proposed by kolahi ahari and co-workers [84]. the combination of mwcnts and au nps improved the electrocatalytic activity of the sensor toward the oxidation of tramadol in solutions. a linear relationship was observed between dpv response of the modified electrode and the concentration of tramadol in the range of 0.012-0.1 and 0.1-3.0 μm. the detection limit of the method was 0.005 μm. in addition, the proposed sensor (mwcnts-au nps/pge) showed high sensitivity and selectivity for the determination of the concentration of tramadol in tablets and biological fluids. li and wang [85] worked on the development of a stable, sensitive, and selective electrochemical sensor based on coo nps and functionalized carbon nanotubes (coo@f-cnts) for the detection of tramadol. the coo@f-cnts nanocomposite-modified gce was made using an electrodeposition approach. the electrochemical studies using dpv and amperometry revealed that f-cnts and coo nps had a synergistic electrocatalytic effect in promoting charge transfer in the oxidation of tramadol as a sensitive and selective sensor with a linear range of 1 to 300 µm. the detection limit and sensitivity were calculated to be 6 nm and 0.44971 µa/µm, respectively. the usefulness and precision of coo@f-cnts/gce for determining tramadol in prepared real samples from urine samples of athlete volunteers were explored. the results showed that the elisa and amperometric analyses had a high level of agreement. electrochemical sensors based on graphitic carbon nitride graphitic carbon nitride (g-c3n4) emphasized the analog skeleton to graphite, found to be promising and fascinating material containing its sturdy c=n covalent bonds instead of c=c in graphite and the layers that are linked by van der waals forces [86]. the incorporation of heteroatoms such as nitrogen atoms into carbon-based materials can enhance the properties of existing material whereby the nitrogen atoms act as the strong electron donor sites for catalytic conductivity due to the chemical nature of the nitrogen atom [87,88]. therefore, g-c3n4 is an excellent material for use in the surface modification of electrodes [89]. for example, hassannezhad et al. [90] reported an electrochemical sensor based on graphitic carbon nitride-fe3o4 (g-c3n4-fe3o4) nanocomposite modified cpe for voltammetric determination of tramadol. the developed g-c3n4-fe3o4/cpe sensor demonstrated better sensitivity for the analysis of tramadol compared to that of the unmodified cpe. the modified electrode showed a rise in peak current and a decrease in the overpotential for the oxidation reaction of tramadol. under optimized conditions, the proposed electrode showed great detection efficiency for tramadol using dpv over the concentration range of 0.2–14.0 μm and 14.0–120.0 µm and a detection limit of 0.1 μm. the g-c3n4-fe3o4/cpe sensor demonstrated suitable potential to be used for the quantification of tramadol in biological samples (serum, plasma and urine). electrochemical sensors based on metal nps noble metal nanoparticles (mainly au nps, ag nps, pt nps, pd nps, ru nps and their alloys au-ag, au-pt, ag-pt, pt-pd, etc.) possess exceeding advantages over other nanomaterials, including stability, conductivity, biocompatibility, low cytotoxicity and size-related electronic, magnetic and optical properties [91,92]. these https://doi.org/10.5599/admet.1593 y. m. baghelani et al. admet & dmpk 11(2) (2023) 117-134 124 metal nanoparticles are helpful for decreasing the overpotentials of the electroanalytical reactions while maintaining the reversibility of the redox reactions. further, the noble nanoparticles can be easily electrodeposited on any kind of working electrode surface to enhance the overall surface properties. the efficacy of the nafion/blended cetyltrimethylammonium bromide (ctab) protected au nps-modified gce for the electrochemical detection of tramadol in wastewater was studied by amin and co-workers [93]. compared with unmodified gce, the nafion/ctab-au/gce, due to the synergistic effect, exhibited excellent electrocatalytic activity toward the oxidation of tramadol. the square wave anodic stripping voltammetric (swasv) analysis was employed to quantify the amount of tramadol. the swasv response of nafion/ctabau/gce was linear over the 0.5-1 µg/ml and 2-12 µg/ml ranges with a detection limit of 3×10-4 µg/ml. the nafion/ctab-au/gce sensor showed good accuracy and reliability. hassanvand and jalali [94] electrodeposited au nps and l-cysteine on gce to prepare a modified electrode for the simultaneous determination of tramadol and acetaminophen. the enlarged surface area of the electrode and high electrocatalytic activity brought about by the nanocomposite were useful in the sensitive and selective determination of tramadol and acetaminophen. the linear dynamic ranges of concentration obtained by swv, were 0.1-10.7 µm and 0.5-63.5 µm, for acetaminophen and tramadol, respectively. the limit of detection was calculated as 0.03 µm for acetaminophen and 0.17 µm for tramadol. the proposed electrode was used successfully in the simultaneous determination of the drugs in spiked human plasma samples. hojjati-najafabadi et al. [95] synthesized au nps using a biosynthesized strategy by mentha aquatic extract and characterized by uv-vis spectroscopic method. the synthesized au nps were used as a conductive mediator for the modification of the tramadol electrochemical sensor. the modified paste electrode with au nps and 1butyl-3-methylimidazolium tetrachloroborate (bmtcb) showed high catalytic activity for the determination of tramadol in an aqueous solution. with a detection limit of 6.0 nm, the au nps/bmtcb/cpe revealed a linear relationship between tramadol oxidation current and concentration in the range of 0.01–400.0 m. the real sample tests confirmed the good ability of au nps/bmtcb/cpe for the determination of tramadol. electrochemical sensors based on metal oxide nps recently, there has been a growing interest in studying the application of metal oxide nps in various fields due to their attractive physical and chemical properties. metal oxide nps with different morphologies have been made through versatile methods. the main functions of metal oxide nps in electroanalysis involve the large surface-to-volume ratio, high surface reaction activity, and high catalytic efficiency [96]. the use of metal oxide nps was reported to improve the response time, linear range, detection limit, reproducibility and long-term stability of the biosensors [97]. the strong affinity of metal oxide nps to the surface of the working electrode can be achieved by various techniques, including physical adsorption, electrodeposition, chemical covalent bonding and electropolymerization [98]. madrakian et al. [99] fabricated an electrochemical sensor for rapid and sensitive determination of tramadol based on ni-al layered double hydroxide (ni-al ldh)-fe3o4 nps modified gce. the modified electrode showed good performance for tramadol determination. under the optimized conditions, the anodic peak current was linear for the concentration of tramadol in the range 1.0–200.0 μm with a detection limit of 3.0×10-1 m. in addition, the method electrode was successfully used to detect the concentration of tramadol in human serum and urine samples. memon et al. [100] reported the development of an effective and sensitive modified electrode for the quantitative determination of tramadol. this modified electrode was fabricated by modification of gce with co3o4 nps. the co3o4 nps over the gce surface resulted in the production of catalytic current, which helped admet & dmpk 11(2) (2023) 117-134 electrochemical analysis of tramadol doi: https://doi.org/10.5599/admet.1593 125 in the development of a sensitive and effective detection tool. using the modified gce, under optimized conditions, the determination of tramadol was obtained with a linear range of 0.5–45 μm and a detection limit of 0.001 μm. the modified electrode was successfully applied for the quantitative analysis of tramadol in commercial pharmaceutical samples. arabali et al. [101] modified a pge using cuo nps/ polypyrrole (ppy) nanocomposite as a powerful electrochemical strategy for the determination of tramadol. the cuo nps-ppy/pge showed excellent electrocatalytic activity for oxidation determination of tramadol with improving oxidation current about 4.55 times. using swv investigation, the over-potential oxidation of tramadol was decreased by about 40 mv compared to the oxidation potential of a drug at unmodified pge. in the analytical investigation using the swv method, the cuo nps-ppy/pge showed more advantage for the determination of tramadol in the concentration range 5.0 nm–380 µm with a detection limit of 1.0 nm. in addition, the pge/cuo-nps/ppy showed acceptable data for the determination of tramadol in drug samples. khairy and banks [102] demonstrated the applicability of screen-printed electrodes (spe) modified with yb2o3 nps for individual and simultaneous determinations of acetaminophen and tramadol drugs. the acetaminophen and tramadol exhibited non-overlapping voltammetric signals at voltages of +0.30 and +0.67 v (vs. ag/agcl; ph = 9) using yb2o3-spes. pharmaceutical dosage forms and spiked human fluids were analyzed in wide linear concentration ranges of 0.25–654 and 0.50– 115 μm with limits of detection of 55 and 87 nm for acetaminophen and tramadol, respectively. vazirirad et al. [103] constructed a chemically modified cpe based on sno2/α-fe2o3 hierarchical nanorods as a simple electrochemical sensor for a sensitive simultaneous determination of dopamine and tramadol. application of sno2/α-fe2o3 hierarchical nanorods as the modifier of the cpe showed high oxidation peak currents for the determination of dopamine and tramadol due to its catalytic effect and high surface area. using the dpv technique, the sensor showed linear ranges of 0.1-70 µm and 0.5-65 µm and low detection limits of 40 nm and 65 nm toward dopamine and tramadol, respectively. finally, the analytical performance of the sensor was evaluated for the analysis of dopamine and tramadol in human blood serum and urine with satisfactory results. reported electrochemical sensors for the determination of tramadol based on nanomaterials are summarized in table 1. conclusions this review gives an overview of the recent advances in the application of nanomaterials in the electrochemical detection of tramadol. nanomaterials can provide a reliable electrochemical platform for the detection of tramadol. the high surface-to-volume ratio, high conductivity, enhanced electron transfer rate, and simple functionalization process are the main reasons that nanomaterials have gained a lot of attention to construct sensors with high sensitivity. although nanomaterials-based electrochemical sensors have obvious advantages, they still need to be further studied to increase their sensitivity and selectivity in order to provide powerful and effective tools in pharmaceutical and biological samples. the mechanism of the electrochemical reaction of tramadol on the modified electrodes and the mechanism of action in which the modified nanomaterials play the role of enhancing the sensing sensitivity and selectivity were not elucidated clearly in most present reports. the specific interaction between the target and the modifier material, and interfaces interactions between the different modifiers are always speculated or suggested due to the direct proof. the changes in the composite of the nanostructures-based modified electrodes are seldom investigated in most research, though these changes may contain important https://doi.org/10.5599/admet.1593 y. m. baghelani et al. admet & dmpk 11(2) (2023) 117-134 126 information on reaction mechanisms. table 1. the summary of the reported electrochemical sensors for determination of tramadol based on nanomaterials. electrochemical sensor method detection limit linear range ref. carbon nanomaterials gr and its derivatives nife2o4-gr/cpe swv 0.0036 µm 0.01-9 µm [64] pt-pd/pdda-ngr/gce swv 5.7 μm 12.0-240.0 μm [65] go-mwcnts/cpe dpv 1.50×10-10 m 2.0×10-9-1.1×10-3 m [66] au nps-gr-htp/cpe dpv 0.82 µm 1.0–16.0 µm and 16.0– 100.0 μm [67] ceo2-sno2/rgo/gce dpv 0.0056 μm 0.008-10 μm and 10-270 μm [68] h-gonps/gce adasv 0.015 μm 0.05–5.0 μm, and 5.0–20 μm [69] cnts mwcnts/gce dpv 0.361 μm 2-300 μm [78] mwcnts/gce dpv 0.38 µm 4-35 µm and 60-550 µm [79] au nps/cnts/gce dpv 68 nm 0.1–1000 μm [80] 1,3-di-br/mgo/swcnts/cpe swv 8.0 nm 0.05-280 μm [81] la3+-cuo/mwcnts/gce dpv 0.014 μm 0.5-900.0 μm [82] cntsilco3o4-sds/cpe dpv 0.62 nm [83] sb2o3 nps-mwcnts/gce dpv 9.5×10-9 m 4×10-8–3.0×10-5 m [84] pt-pd/nio nps/swcnts/emich3so3/cpe swv 0.9 nm 0.1-750.0 µm [85] mip-mwcnts/gce dpv 1.4 μm 9.0 to 30.0 μm [86] mwcnts-au nps/pge dpv 0.005 μm 0.012-0.1 and 0.1-3.0 μm [87] coo@f-cnts/gce amperometry 6 nm 1 to 300 µm [88] graphitic carbon nitride g-c3n4/fe3o4/cpe dpv 0.1 μm 0.2–14.0 μm and 14.0– 120.0 µm [93] metal nps nafion/ctab-au/gce swasv 3×10-4 µg/ml 0.5-1 µg/ml and 2-12 µg/ml [95] au/cysteic acid/gce swv 0.17 µm 0.5-63.5 µm [96] au nps/bmtcb/cpe swv 6.0 nm 0.01–400.0 m [97] metal oxide nps ni-al ldh-fe3o4 nps/gce dpv 3.0×10-1 m 1.0–200.0 μm [101] co3o4 nps/gce dpv 0.001 μm 0.5–45 μm [102] cuo nps-ppy/pge swv 1.0 nm 5.0 nm–380 µm [103] yb2o3-spe dpv 0.087 μm 0.5–5400 µm [104] sno2/α-fe2o3/cpe dpv 65 nm 0.5-65 µm [105] the selectivity of the target analyte of the modified electrodes is an issue that deserves special concern. the selectivity and sensitivity of tramadol sensing can be improved by the use of bio-specific recognition admet & dmpk 11(2) (2023) 117-134 electrochemical analysis of tramadol doi: https://doi.org/10.5599/admet.1593 127 molecules, e.g., antibodies, biological enzymes, nucleic acid aptamers, and etc. however, the performance of such biological molecules is greatly affected by operating conditions, such as temperature, the buffer solution, the ph value of the solution, etc., that will have a fatal effect on the sensitivity, selectivity, and stability of the sensor. artificial bio-recognition molecules, such as molecularly imprinted polymers, can overcome the abovementioned shortcomings of biological recognition molecules. although conductive monomer molecules are selected as the functional monomers, the conductivity of the resulting polymers synthesized by the functional monomers is not as good as the conductive materials. combining these with high-conductive materials (novel nanostructures with high surface area and conductivity) is necessary for these molecularly imprinted polymers in order to acquire the desired sensitivity and responsiveness of sensors. despite the mentioned limitations and problems, electrochemical sensors could offer a relatively cheap, fast, sensitive, and selective method for online detection or working in complicated matrices. they show great potential for further growth in the near future. conflict of interest: the authors claim no conflict of interest. abbreviations graphene gr carbon nanotubes cnts nanoparticles nps square wave voltammetry swv carbon paste electrode cpe nitrogen doped gr ngr poly(diallyldimethyl ammonium chloride) pdda glassy carbon electrode gce graphene oxide go multiwalled carbon nanotubes mwcnts 4-hydroxyl-2-(triphenylphosphonio)phenolate htp hierarchical graphene oxide nanoplatelets h-gonps adsorptive anodic stripping voltammetric adasv single-walled carbon nanotubes swcnts 1,3-dipropylimidazolium bromide 1,3-di-br sodium dodecyl sulfate sds 1-ethyl-3-methylimidazolium methane sulfonate emich3so3 molecularly imprinted polymer mip pencil graphite electrode pge functionalized carbon nanotubes f-cnts cetyltrimethylammonium bromide ctab square wave anodic stripping voltammetric swasv 1-butyl-3-methylimidazolium tetrachloroborate bmtcb layered double hydroxide ldh polypyrrole ppy references [1] a. cetinkaya, s.i. kaya, g. ozcelikay, f. budak, s.a. ozkan. carbon nanomaterials-based novel hybrid platforms for electrochemical sensor applications in drug analysis. critical reviews in analytical chemistry (2022). https://doi.org/10.1080/10408347.2022.2109125. [2] m. subedi, s. bajaj, m.s. kumar, y.c. mayur. an overview of tramadol and its usage in pain management and future perspective. biomedicine & pharmacotherapy 111 (2019) 443-451. https://doi.org/10.1016/j.biopha.2018.12.085. [3] m. karbakhsh, z.n. salehian. acute opiate overdose in tehran: the forgotten role of opium. addictive behaviors 32 (2007) 1835-1842. https://doi.org/10.1016/j.addbeh.2006.12.014. https://doi.org/10.5599/admet.1593 https://doi.org/10.1080/10408347.2022.2109125 https://doi.org/10.1016/j.biopha.2018.12.085 https://doi.org/10.1016/j.addbeh.2006.12.014 y. m. baghelani et al. admet & dmpk 11(2) (2023) 117-134 128 [4] e.r. zaaijer, j. bruijel, p. blanken, v. hendriks, m.w. j. koeter, m.j. kreek. personality as a risk factor for illicit opioid use and a protective factor for illicit opioid dependence. drug and alcohol dependence 145 (2014) 101-105. https://doi.org/10.1016/j.drugalcdep.2014.09.783. [5] a.m. drewes, r.d. jensen, l.m. nielsen, j. droney, l.l. christrup, l. arendt‐nielsen, a. dahan. differences between opioids: pharmacological, experimental, clinical and economical perspectives. british journal of clinical pharmacology 75 (2013) 60-78. https://doi.org/10.1111/j.13652125.2012.04317.x. [6] t. threlkeld, t.v. parran jr, c.a. adelman, s.f. grey, y. yu. tramadol versus buprenorphine for the management of acute heroin withdrawal: a retrospective matched cohort controlled study. the american journal on addictions 15 (2006) 186-191. https://doi.org/10.1080/10550490500528712. [7] k. shah, b. stout, h. caskey. tramadol for the management of opioid withdrawal: a systematic review of randomized clinical trials. cureus (2020). https://doi.org/10.7759/cureus.9128 [8] l.p. kenzie, r.j. donald, t. margaret. abuse potential and pharmacological comparison of tramadol and morphine. drug and alcohol dependence 27 (1991) 7-17. https://doi.org/10.1016/0376-8716(91) 90081-9. [9] s. koncz, n. papp, n. menczelesz, d. pothorszki, g. bagdy. eeg and sleep effects of tramadol suggest potential antidepressant effects with different mechanisms of action. pharmaceuticals 14 (2021) 431. https://doi.org/10.3390/ph14050431. [10] j. kozak, k. tyszczuk-rotko, m. wójciak, i. sowa. electrochemically activated screen-printed carbon sensor modified with anionic surfactant (aspce/sds) for simultaneous determination of paracetamol, diclofenac and tramadol. materials 14 (2021) 3581. https://doi.org/10.3390/ma14133581. [11] e.s. williams, b.w. brooks. human health risk assessment for pharmaceuticals in the environment: existing practice, uncertainty, and future directions. human pharmaceuticals in the environment 4 (2012) 167-224. https://doi.org/10.1007/978-1-4614-3473-3_8. [12] y. davoudabadi farahani, v. safari fard. an amine/imine functionalized microporous mof as a new fluorescent probe exhibiting selective sensing of fe3+ and al3+ over mixed metal ions. journal of applied organometallic chemistry 2 (2022) 180-194. https://doi.org/10.22034/jaoc.2022.154981. [13] r.t. hussain, a.s. islam, m. khairuddean, f.b.m. suah. a polypyrrole/go/zno nanocomposite modified pencil graphite electrode for the determination of andrographolide in aqueous samples. alexandria engineering journal 61 (2022) 4209-4218. https://doi.org/10.1016/j.aej.2021.09.040. [14] j.f. rontani. use of gas chromatography-mass spectrometry techniques (gc-ms, gc-ms/ms and gcqtof) for the characterization of photooxidation and autoxidation products of lipids of autotrophic organisms in environmental samples. molecules 27 (2022) 1629. https://doi.org/10.3390/mole cules27051629. [15] m. kazemipour, m. ansari, a. mohammadi, h. beitollahi, r. ahmadi, use of adsorptive square-wave anodic stripping voltammetry at carbon paste electrode for the determination of amlodipine besylate in pharmaceutical preparations. journal of analytical chemistry 64 (2009) 65-70. https://doi.org/ 10.1134/s1061934809010134. [16] h. ahmar, m. nejati-yazdinejad, m. najafi, k.s. hasheminasab. switchable hydrophilicity solventbased homogenous liquid–liquid microextraction (shs-hllme) combined with gc-fid for the quantification of methadone and tramadol. chromatographia 81 (2018) 1063-1070. https://doi.org/10.1007/s10337-018-3528-y. [17] f. xu, q. li, w. wei, l. liu, h. li. development of a liquid–liquid microextraction method based on a switchable hydrophilicity solvent for the simultaneous determination of 11 drugs in urine by gc–ms. chromatographia 81 (2018) 1695-1703. https://doi.org/10.1007/s10337-018-3643-9. [18] h. yu, m. choi, j.h. jang, b. park, y.h. seo, c.h. jeong, j.w. bae, s. lee. development of a columnswitching lc-ms/ms method of tramadol and its metabolites in hair and application to a pharmacogenetic study. archives of pharmacal research 41 (2018) 554-563. https://doi.org/10.1007/ s12272-018-1013-7. https://doi.org/10.1016/j.drugalcdep.2014.09.783 https://doi.org/10.1111/j.1365-2125.2012.04317.x https://doi.org/10.1111/j.1365-2125.2012.04317.x https://doi.org/10.1080/10550490500528712 https://doi.org/10.7759/cureus.9128 https://doi.org/10.1016/0376-8716(91)90081-9 https://doi.org/10.1016/0376-8716(91)90081-9 https://doi.org/10.3390/ph14050431 https://doi.org/10.3390/ma14133581 https://doi.org/10.1007/978-1-4614-3473-3_8 https://doi.org/10.22034/jaoc.2022.154981 https://doi.org/10.1016/j.aej.2021.09.040 https://doi.org/10.3390/‌molecules27051629 https://doi.org/10.3390/‌molecules27051629 https://doi.org/10.1007/s10337-018-3528-y https://doi.org/10.1007/s10337-018-3643-9 https://doi.org/10.1007/‌s12272-018-1013-7 https://doi.org/10.1007/‌s12272-018-1013-7 admet & dmpk 11(2) (2023) 117-134 electrochemical analysis of tramadol doi: https://doi.org/10.5599/admet.1593 129 [19] h. cornthwaite, l. labine, j. watterson. semi‐ quantitative analysis of tramadol, dextromethorphan, and metabolites in decomposed skeletal tissues by ultra performance liquid chromatography quadrupole time of flight mass spectrometry. drug testing and analysis 10 (2018) 961-967. https://doi.org/10.1002/dta.2327. [20] s. chalavi, a.r. fakhari, s. nojavan. development of a modified partial filling method in capillary electrophoresis using two chiral plugs for the simultaneous enantioseparation of chiral drugs: comparison with mixed chiral selector capillary electrophoresis. journal of chromatography a 1567 (2018) 211-218. https://doi.org/10.1016/j.chroma.2018.06.052. [21] e. naghdi, a.r. fakhari. simultaneous chiral separation of tramadol and methadone in tablets, human urine, and plasma by capillary electrophoresis using maltodextrin as the chiral selector. chirality 30 (2018) 1161-1168. https://doi.org/10.1002/chir.23008. [22] y. hamid, m.r. fat’hi. a simple vortex-assisted graphene oxide nanosheets dispersive micro-solid phase extraction combined with high-performance liquid chromatography for uvvis detection of tramadol in biological samples. separation science and technology 53 (2018) 1689-1697. https://doi.org/10.1080/01496395.2018.1439960. [23] p. saar-reismaa, e. erme, m. vaher, m. kulp, m. kaljurand, j. mazina-šinkar. in situ determination of illegal drugs in oral fluid by portable capillary electrophoresis with deep uv excited fluorescence detection. analytical chemistry 90 (2018) 6253-6258. https://doi.org/10.1021/acs.analchem.8b00911. [24] a. yousefzadeh, j. hassanzadeh, s.m.j. mousavi, m. yousefzadeh. surface molecular imprinting and powerfully enhanced chemiluminescence emission by cu nanoclusters/mof composite for detection of tramadol. sensors and actuators b: chemical 286 (2019) 154-162. https://doi.org/10.1016/j.snb.20 19.01.155. [25] b. deiminiat, g.h. rounaghi, m.h. arbab-zavar. development of a new electrochemical imprinted sensor based on poly-pyrrole, sol–gel and multiwall carbon nanotubes for determination of tramadol. sensors and actuators b: chemical 238 (2017) 651-659. https://doi.org/10.1016/j.snb.2016.07.110. [26] m. miraki, h. karimi-maleh, m.a. taher, s. cheraghi, f. karimi, s. agarwal, v.k. gupta. voltammetric amplified platform based on ionic liquid/nio nanocomposite for determination of benserazide and levodopa. journal of molecular liquids 278 (2019) 672-676. https://doi.org/10.1016/j.molliq.2019.01.081. [27] h. beitollahi, s.z. mohammadi, m. safaei, s. tajik. applications of electrochemical sensors and biosensors based on modified screen-printed electrodes: a review. analytical methods 12 (2020) 15471560. https://doi.org/10.1039/c9ay02598g. [28] n.h. khand, i.m. palabiyik, j.a. buledi, s. ameen, a.f. memon, t. ghumro, a.r. solangi. functional co3o4 nanostructure-based electrochemical sensor for direct determination of ascorbic acid in pharmaceutical samples. journal of nanostructure in chemistry 11 (2021) 455-468. https://doi.org/ 10.1007/s40097-020-00380-8. [29] s.a. alavi-tabari, m.a. khalilzadeh, h. karimi-maleh. simultaneous determination of doxorubicin and dasatinib as two breast anticancer drugs uses an amplified sensor with ionic liquid and zno nanoparticle. journal of electroanalytical chemistry 811 (2018) 84-88. https://doi.org/10.1016/ j.jelechem.2018.01.034. [30] m. vardini, n. abbasi, a. kaviani, m. ahmadi, e. karimi. graphite electrode potentiometric sensor modified by surface imprinted silica gel to measure valproic acid. chemical methodologies 6 (2022) 398-408. https://doi.org/10.22034/chemm.2022.328620.1437. [31] e.g. arafa gomaa, m. diab, a. elsonbati, h.m. abulenader, a. helmy. interaction parameters for cucl2 plus orange g (og) at 19.1oc using carbon classy electrode (cge) in kcl aqueous solutions. asian journal of nanoscience and materials 5 (2022) 294-302. https://doi.org/10.26655/ajnano mat.2022.4.3. [32] m. mazloum-ardakani, h. beitollahi, m.k. amini, f. mirkhalaf, b.f. mirjalili, a. akbari. application of 2(3,4-dihydroxyphenyl)-1,3-dithialone self-assembled monolayer on gold electrode as a nanosensor for electrocatalytic determination of dopamine and uric acid. analyst 136 (2011) 1965-1970. https://doi.org/10.1039/c0an00823k. https://doi.org/10.5599/admet.1593 https://doi.org/10.1002/dta.2327 https://doi.org/10.1016/j.chroma.2018.06.052 https://doi.org/10.1002/chir.23008 https://doi.org/10.1080/01496395.2018.1439960 https://doi.org/10.1021/acs.analchem.8b00911 https://doi.org/10.1016/j.snb.2019.01.155 https://doi.org/10.1016/j.snb.2019.01.155 https://doi.org/10.1016/j.snb.2016.07.110 https://doi.org/10.1016/j.molliq.2019.01.081 https://doi.org/10.1039/c9ay02598g https://doi.org/10.1007/s40097-020-00380-8 https://doi.org/10.1007/s40097-020-00380-8 https://doi.org/10.1016/j.jelechem.2018.01.034 https://doi.org/10.1016/j.jelechem.2018.01.034 https://doi.org/10.22034/chemm.2022.328620.1437 http://www.ajnanomat.com/article_69758.html http://www.ajnanomat.com/article_69758.html https://doi.org/10.26655/ajnano‌mat.2022.4.3 https://doi.org/10.26655/ajnano‌mat.2022.4.3 y. m. baghelani et al. admet & dmpk 11(2) (2023) 117-134 130 [33] n. baig, m. sajid, t.a. saleh. recent trends in nanomaterial-modified electrodes for electroanalytical applications. trac trends in analytical chemistry 111 (2019) 47-61. https://doi.org/10.1016/j.trac.20 18.11.044. [34] h. karimi-maleh, m. sheikhshoaie, i. sheikhshoaie, m. ranjbar, j. alizadeh, n.w. maxakato, a. abbaspourrad. a novel electrochemical epinine sensor using amplified cuo nanoparticles and anhexyl-3-methylimidazolium hexafluorophosphate electrode. new journal of chemistry 43 (2019) 23622367. https://doi.org/10.1039/c8nj05581e. [35] p. joshi, s. mehtab, m.g.h. zaidi, t. tyagi, a. bisht. development of polyindole/tungsten carbide nanocomposite-modified electrodes for electrochemical quantification of chlorpyrifos. journal of nanostructure in chemistry 10 (2020) 33-45. https://doi.org/10.1007/s40097-019-00326-9. [36] h. karimi-maleh, f. karimi, y. orooji, g. mansouri, a. razmjou, a. aygun, f. sen. a new nickel-based co-crystal complex electrocatalyst amplified by nio dope pt nanostructure hybrid; a highly sensitive approach for determination of cysteamine in the presence of serotonin. scientific reports 10 (2020) 11699. https://doi.org/10.1038/s41598-020-68663-2. [37] e. yarali, e. eksin, h. torul, a. ganguly, u. tamer, p. papakonstantinou, a. erdem. impedimetric detection of mirna biomarkers using paper-based electrodes modified with bulk crystals or nanosheets of molybdenum disulfide. talanta 241 (2022) 123233. https://doi.org/10.1016/j.ta lanta.2022.123233. [38] y. wang, x. wu, j. sun, c. wang, g. zhu, l. p. bai, w. zhang. stripping voltammetric determination of cadmium and lead ions based on a bismuth oxide surface-decorated nanoporous bismuth electrode. electrochemistry communications 136 (2022) 107233. https://doi.org/10.1016/j.elecom.2022.107233. [39] v. tallapaneni, l. mude, d. pamu, v.v.s.r. karri. formulation, characterization and in vitro evaluation of dual-drug loaded biomimetic chitosan-collagen hybrid nanocomposite scaffolds. journal of medicinal and chemical sciences 5 (2022) 1059-1074. https://doi.org/10.26655/jmchemsci.20 22.6.19. [40] m. hasanzadeh, z. hasanzadeh, s. alizadeh, m. sayadi, m. nasiri nezhad, r.e. sabzi, s. ahmadi. coppernickel oxide nanofilm modified electrode for non-enzymatic determination of glucose. journal of electrochemical science and engineering 10(3) (2020) 245-255. https://doi.org/10.5599/jese.699. [41] m.m. ardakani, z. taleat, h. beitollahi, m. salavati-niasari, b.b.f. mirjalili, n. taghavinia. electrocatalytic oxidation and nanomolar determination of guanine at the surface of a molybdenum (vi) complex–tio2 nanoparticle modified carbon paste electrode. journal of electroanalytical chemistry 624 (2008) 73-78. https://doi.org/10.1016/j.jelechem.2008.07.027. [42] k. sudhakara, a. praveen kumar, b. prem kumar, a. raghavender, s. ravi, d. negussa keniec, y. lee. synthesis of γ-fe2o3 nanoparticles and catalytic activity of azide-alkyne cycloaddition reactions. asian journal of nanoscience and materials 5 (2022) 243-251. https://doi.org/10.26655/ajnano mat.2022.4.1. [43] g.k. jayaprakash, r. flores-moreno, b.e.k. swamy, k. mohanty, p. dhiman. pre/post electron transfer regioselectivity at glycine modified graphene electrode interface for voltammetric sensing applications. journal of electrochemical science and engineering 12 (2022) 1001-1008. https://doi.org/10.5599/jese.1438. [44] h. karimi-maleh, c. karaman, o. karaman, f. karimi, y. vasseghian, l. fu, m. baghayeri, j. rouhi, p. senthil kumar, p.l. show, s. rajendran, a.l. sanati, a. mirabi. nanochemistry approach for the fabrication of fe and n co-decorated biomass-derived activated carbon frameworks: a promising oxygen reduction reaction electrocatalyst in neutral media. journal of nanostructure in chemistry 12 (2022) 429-439. https://doi.org/10.1007/s40097-022-00492-3. [45] k. gajanan, s.n. tijare. applications of nanomaterials. materials today: proceedings 5 (2018) 10931096. https://doi.org/10.1016/j.matpr.2017.11.187. [46] p.k. kalambate, j. noiphung, n. rodthongkum, n. larpant, p. thirabowonkitphithan, t. rojanarata, w. laiwattanapaisal. nanomaterials-based electrochemical sensors and biosensors for the detection of non-steroidal anti-inflammatory drugs. trac trends in analytical chemistry 143 (2021) 116403. https://doi.org/10.1016/j.snb.2020.128879. https://doi.org/10.1016/j.trac.2018.11.044 https://doi.org/10.1016/j.trac.2018.11.044 https://doi.org/10.1039/c8nj05581e https://doi.org/10.1007/s40097-019-00326-9 https://doi.org/10.1016/j.ta‌lanta.2022.123233 https://doi.org/10.1016/j.ta‌lanta.2022.123233 https://doi.org/10.1016/j.elecom.2022.107233 https://doi.org/10.26655/jmchemsci.20‌22.6.19 https://doi.org/10.26655/jmchemsci.20‌22.6.19 https://doi.org/10.5599/jese.699 https://doi.org/10.1016/j.jelechem.2008.07.027 http://www.ajnanomat.com/article_63662.html https://doi.org/10.26655/ajnano‌mat.2022.4.1 https://doi.org/10.26655/ajnano‌mat.2022.4.1 https://doi.org/10.5599/jese.1438 https://link.springer.com/article/10.1007/s40097-022-00492-3#auth-mehdi-baghayeri https://link.springer.com/article/10.1007/s40097-022-00492-3#auth-jalal-rouhi https://link.springer.com/article/10.1007/s40097-022-00492-3#auth-p_-senthil_kumar https://link.springer.com/article/10.1007/s40097-022-00492-3#auth-pau_loke-show https://link.springer.com/article/10.1007/s40097-022-00492-3#auth-saravanan-rajendran https://link.springer.com/article/10.1007/s40097-022-00492-3#auth-afsaneh_l_-sanati https://doi.org/10.1016/j.matpr.2017.11.187 https://doi.org/10.1016/j.snb.2020.128879 admet & dmpk 11(2) (2023) 117-134 electrochemical analysis of tramadol doi: https://doi.org/10.5599/admet.1593 131 [47] a. lohrasbi-nejad. electrochemical strategies for detection of diazinon. journal of electrochemical science and engineering 12(6) (2022) 1041-1059. https://doi.org/10.5599/jese.1379. [48] n.f. mahmoud, o.a. fouad, a.e. ali, g.g. mohamed. potentiometric determination of the al (iii) ion in polluted water and pharmaceutical samples by a novel mesoporous copper metal–organic frameworkmodified carbon paste electrode. industrial & engineering chemistry research 60 (2021) 2374-2387. [49] r.m. mohabis, f. fazeli, i. amini, v. azizkhani. an overview of recent advances in the detection of ascorbic acid by electrochemical techniques. journal of electrochemical science and engineering 12(6) (2022) 1081-1098. https://doi.org/10.5599/jese.1561. [50] a.m. attaran, s. abdol-manafi, m. javanbakht, m. enhessari. voltammetric sensor based on co3o4/sno2 nanopowders for determination of diltiazem in tablets and biological fluids. journal of nanostructure in chemistry 6 (2016) 121-128. https://doi.org/10.1007/s40097-015-0186-6. [51] z. mehdizadeh, s. shahidi, a. ghorbani-hasan saraei, m. limooei, m. bijad. monitoring of amaranth in drinking samples using voltammetric amplified electroanalytical sensor. chemical methodologies 6 (2022) 246-252. https://doi.org/10.22034/chemm.2022.324073.1423. [52] s. tajik, m.r. aflatoonian, h. beitollahi, i. sheikhshoaie, z. dourandish, f. garkani-nejad, b. aflatoonian, m. bamorovat. electrocatalytic oxidation and selective voltammetric detection of methyldopa in the presence of hydrochlorothiazide in real samples. microchemical journal 158 (2020) 105182. https://doi.org/10.1016/j.microc.2020.105182. [53] k. harismah, m. mirzaei, m. dai, z. roshandel, e. salarrezaei. in silico investigation of nanocarbon biosensors for diagnosis of covid-19. eurasian chemical communications 3 (2021) 95-102. https://doi.org/10.22034/ecc.2021.267226.1120. [54] s. parlayici, v. eskizeybek, a. avcı, e. pehlivan. removal of chromium (vi) using activated carbonsupported-functionalized carbon nanotubes. journal of nanostructure in chemistry 5 (2015) 255-263. https://doi.org/10.1007/s40097-015-0156-z. [55] a.c. power, b. gorey, s. chandra, j. chapman. carbon nanomaterials and their application to electrochemical sensors: a review. nanotechnology reviews 7 (2018) 19-41. https://doi.org/10.1515/ntrev2017-0160. [56] k.s. novoselov, a.k. geim, s.v. morozov, d. jiang, y. zhang, s.v. dubonos, i.v. grigorieva, a.a. firsov. electric field effect in atomically thin carbon films. science 306 (2004) 666-669. https://doi.org/10.11 26/science.1102896. [57] b.a. aragaw. reduced graphene oxide-intercalated graphene oxide nano-hybrid for enhanced photoelectrochemical water reduction. journal of nanostructure in chemistry 10 (2020) 9-18. https://doi.org/10.1007/s40097-019-00324-x. [58] j. mohanraj, d. durgalakshmi, r.a. rakkesh, s. balakumar, s. rajendran, h. karimi-maleh. facile synthesis of paper based graphene electrodes for point of care devices: a double stranded dna (dsdna) biosensor. journal of colloid and interface science 566 (2020) 463-472. https://doi.org/ 10.1016/j.jcis.2020.01.089. [59] f. ahmad, k.e. kainat, u. farooq. a comprehensive review on materials having high oxygen reduction reaction (orr) activity. journal of chemical reviews. 4 (2022) 374-422. https://doi.org/10.22034/jcr. 2022.357844.1185. [60] m.j. allen, v.c. tung, r.b. kaner. honeycomb carbon: a review of graphene. chemical reviews 110 (2009) 132–145. https://doi.org/10.1021/cr900070d. [61] a. afkhami, h. khoshsafar, h. bagheri, t. madrakian. preparation of nife2o4/graphene nanocomposite and its application as a modifier for the fabrication of an electrochemical sensor for the simultaneous determination of tramadol and acetaminophen. analytica chimica acta 831 (2014) 50-59. https://doi.org/10.1016/j.aca.2014.04.061. [62] j. manokaran, j. narendranath, r. muruganantham, n. balasubramanian. nitrogen doped graphene supported pt-pd nanoparticle modified gc electrode for electrochemical determination of tramadol and paracetamol. indian journal of chemistry 56a (2017) 63-68. https://doi.org/10.5599/admet.1593 https://doi.org/10.1007/s40097-015-0186-6 https://doi.org/10.22034/chemm.2022.324073.1423 https://doi.org/10.1016/j.microc.2020.105182 https://doi.org/10.22034/ecc.2021.267226.1120 https://doi.org/10.1007/s40097-015-0156-z https://doi.org/10.1515/ntrev-2017-0160 https://doi.org/10.1515/ntrev-2017-0160 https://doi.org/10.1126/science.1102896 https://doi.org/10.1126/science.1102896 https://doi.org/‌10.1016/j.jcis.2020.01.089 https://doi.org/‌10.1016/j.jcis.2020.01.089 http://www.jchemrev.com/article_156379.html http://www.jchemrev.com/article_156379.html https://doi.org/10.22034/jcr.%20‌2022.357844.1185 https://doi.org/10.22034/jcr.%20‌2022.357844.1185 https://doi.org/10.1021/cr900070d https://doi.org/10.1016/j.aca.2014.04.061 y. m. baghelani et al. admet & dmpk 11(2) (2023) 117-134 132 [63] m.a. mohamed, s.a. atty, n.n. salama, c.e. banks. highly selective sensing platform utilizing graphene oxide and multiwalled carbon nanotubes for the sensitive determination of tramadol in the presence of co‐formulated drugs. electroanalysis 29 (2017) 1038-1048. https://doi.org/10.1002/elan.2016 00668. [64] n. rokhsefid, m.r. shishehbore. synthesis and characterization of an au nanoparticles/graphene nanosheet nanocomposite and its application for the simultaneous determination of tramadol and acetaminophen. analytical methods 11 (2019) 5150-5159. https://doi.org/10.1039/c9ay01497g. [65] f. hosseini, m. bahmaei, m. davallo. a sensitive method for the electrochemical determination of tramadol, codeine and caffeine by a ceo2-sno2/rgo nanocomposite-modified glassy carbon electrode. analytical and bioanalytical electrochemistry 13 (2021) 264-282. [66] j. saichanapan, k. promsuwan, k. saisahas, a. soleh, k.h. chang, a.f.l. abdullah, w. limbut. voltammetric determination of tramadol using a hierarchical graphene oxide nanoplatelets modified electrode. journal of the electrochemical society 168 (2021) 117512. https://doi.org/10.1149/1945 7111/ac3529. [67] l.s. porto, d.n. silva, a.e.f. de oliveira, a.c. pereira, k.b. borges. carbon nanomaterials: synthesis and applications to development of electrochemical sensors in determination of drugs and compounds of clinical interest. reviews in analytical chemistry 38 (2019). https://doi.org/10.1515/revac-2019-0017. [68] v. schroeder, s. savagatrup, m. he, s. lin, t. m. swager. carbon nanotube chemical sensors. chemical reviews 119 (2019) 599–663. https://doi.org/10.1021/acs.chemrev.8b00340. [69] h. song, x. zhang, y. liu, z. su. developing graphene-based nanohybrids for electrochemical sensing. chemical record 19 (2019) 534–549. https://doi.org/10.1002/tcr.201800084. [70] h. karimi-maleh, a.f. shojaei, k. tabatabaeian, f. karimi, s. shakeri, r. moradi. simultaneous determination of 6-mercaptopruine, 6-thioguanine and dasatinib as three important anticancer drugs using nanostructure voltammetric sensor employing pt/mwcnts and 1-butyl-3-methylimidazolium hexafluoro phosphate. biosensors and bioelectronics 86 (2016) 879-884. https://doi.org/10.1016/j.bi os.2016.07.086. [71] i. raeisi, p. derakhshi, p. aberoomsand azar, m. saber tehrani. desulfurization of gas condensate under visible light using synthesized photocatalysts of mn/tio2/mwcnts and ni/tio2/mwcnts. journal of nanostructure in chemistry 11 (2021) 165-185. https://doi.org/10.1007/s40097-020-00349-7. [72] a. hosseini fakhrabad, r. sanavi khoshnood, m.r. abedi, m. ebrahimi. fabrication a composite carbon paste electrodes (cpes) modified with multi-wall carbon nanotubes (mwcnts/n, n-bis (salicyliden)1,3-propandiamine) for determination of lanthanum (iii). eurasian chemical communications 3 (2021) 627-634. http://doi.org/10.22034/ecc.2021.288271.1182. [73] s. tajik, m.a. taher, h. beitollahi. simultaneous determination of droxidopa and carbidopa using a carbon nanotubes paste electrode. sensors and actuators b: chemical 188 (2013) 923-930. https://doi.org/10.1016/j.snb.2013.07.085. [74] s. wang, m. pan, k. liu, x. xie, j. yang, l. hong, s. wang. a sio2@ mip electrochemical sensor based on mwcnts and aunps for highly sensitive and selective recognition and detection of dibutyl phthalate. food chemistry 381 (2022) 132225. https://doi.org/10.1016/j.foodchem.2022.132225. [75] a. babaei, a.r. taheri, m. afrasiabi. a multi-walled carbon nanotube-modified glassy carbon electrode as a new sensor for the sensitive simultaneous determination of paracetamol and tramadol in pharmaceutical preparations and biological fluids. journal of the brazilian chemical society 22 (2011) 1549-1558. https://doi.org/10.1590/s0103-50532011000800020. [76] s. branch. a sensor for determination of tramadol in pharmaceutical preparations and biological fluids based on multi-walled carbon nanotubes-modified glassy carbon electrode. journal of the chemical society of pakistan 35 (2013) 1106. [77] n.f. atta, r.a. ahmed, h.m. amin, a. galal. monodispersed gold nanoparticles decorated carbon nanotubes as an enhanced sensing platform for nanomolar detection of tramadol. electroanalysis 24 (2012) 2135-2146. https://doi.org/10.1002/elan.201200344. https://doi.org/10.1002/elan.2016‌00668 https://doi.org/10.1002/elan.2016‌00668 https://doi.org/10.1039/c9ay01497g https://doi.org/10.1149/1945-7111/ac3529 https://doi.org/10.1149/1945-7111/ac3529 https://doi.org/10.1515/revac-2019-0017 https://doi.org/10.1021/acs.chemrev.8b00340 https://doi.org/10.1002/tcr.201800084 https://doi.org/10.1016/j.bi‌os.2016.07.086 https://doi.org/10.1016/j.bi‌os.2016.07.086 https://doi.org/10.1007/s40097-020-00349-7 http://doi.org/10.22034/ecc.2021.288271.1182 https://doi.org/10.1016/j.snb.2013.07.085 https://doi.org/10.1016/j.foodchem.2022.132225 https://doi.org/10.1590/s0103-50532011000800020 https://doi.org/10.1002/elan.201200344 admet & dmpk 11(2) (2023) 117-134 electrochemical analysis of tramadol doi: https://doi.org/10.5599/admet.1593 133 [78] f. hosseini, m. ebrahimi, h. karimi-maleh. an amplified sensor based on improved carbon paste electrode with 1, 3-dipropylimidazolium bromide and mgo/swcnts nanocomposite for tradamol determination. international journal of electrochemical science 13 (2018) 4923-4932. [79] m.m. foroughi, s. jahani, h.h. nadiki. lanthanum doped fern-like cuo nanoleaves/mwcnts modified glassy carbon electrode for simultaneous determination of tramadol and acetaminophen. sensors and actuators b: chemical 285 (2019) 562-570. https://doi.org/10.1016/j.snb.2019.01.069. [80] n.f. atta, a. galal, s.h. hassan. ultrasensitive determination of nalbuphine and tramadol narcotic analgesic drugs for postoperative pain relief using nano-cobalt oxide/ionic liquid crystal/carbon nanotubes-based electrochemical sensor. journal of electroanalytical chemistry 839 (2019) 48-58. https://doi.org/10.1016/j.jelechem.2019.03.002. [81] e. çidem, t. teker, m. aslanoglu. a sensitive determination of tramadol using a voltammetric platform based on antimony oxide nanoparticles. microchemical journal 147 (2019) 879-885. https://doi.org/ 10.1016/j.microc.2019.04.018. [82] t. tavana, a.r. rezvani, h. karimi-maleh. pt-pd-doped nio nanoparticle decorated at single-wall carbon nanotubes: an excellent, powerful electrocatalyst for the fabrication of electrochemical sensor to determine nalbuphine in the presence of tramadol as two opioid analgesic drugs. journal of pharmaceutical and biomedical analysis 189 (2020) 113397. https://doi.org/10.1016/j.jpba. 2020.113397. [83] c. ricardo teixeira tarley, j. de cássia mendonça, l. rianne da rocha, t. boareto capelari, m. carolyne prete, m. cecílio fonseca, m. gava segatelli. development of a molecularly imprinted poly (acrylic acid)‐mwcnt nanocomposite electrochemical sensor for tramadol determination in pharmaceutical samples. electroanalysis 32 (2020) 1130-1137. https://doi.org/10.1002/elan.201900148. [84] s. kolahi-ahari, b. deiminiat, g.h. rounaghi. modification of a pencil graphite electrode with multiwalled carbon nanotubes capped gold nanoparticles for electrochemical determination of tramadol. journal of electroanalytical chemistry 862 (2020) 113996. https://doi.org/10.1016/j.jelechem. 2020.113996. [85] q. li, h. wang. fabrication of an electrochemical sensor based on a nanocomposite of coo@f-cnts for determination of tramadol narcotic drug in urine of athlete volunteers. international journal of electrochemical science 17 (2022) 2. [86] c.h. choi, s.h. park, s.i. woo. green chemistry heteroatom doped carbons prepared by the pyrolysis of bio-derived amino acids as highly active catalysts for oxygen electro-reduction reactions. green chemistry 13 (2011) 406e412. https://doi.org/10.1039/c0gc00384k. [87] j. liu, h. wang, m. antonietti. graphitic carbon nitride “reloaded”: emerging applications beyond (photo)catalysis. chemical society reviews 45 (2016) 2308-2326. https://doi.org/10.1039/c5cs00767d. [88] s.c. yan, z.s. li, z. g. zou. photodegradation performance of g-c3n4 fabricated by directly heating melamine. langmuir 25 (2009) 10397-10401. https:// doi.org/10.1021/la900923z. [89] h. karimi-maleh, r. darabi, m. shabani-nooshabadi, m. baghayeri, f. karimi, j. rouhi, c. karaman. determination of d&c red 33 and patent blue v azo dyes using an impressive electrochemical sensor based on carbon paste electrode modified with zif-8/g-c3n4/co and ionic liquid in mouthwash and toothpaste as real samples. food and chemical toxicology 162 (2022) 112907. https://doi.org/ 10.1016/j.fct.2022.112907. [90] m. hassannezhad, m. hosseini, m.r. ganjali, m. arvand. a graphitic carbon nitride (gc3n4/fe3o4) nanocomposite: an efficient electrode material for the electrochemical determination of tramadol in human biological fluids. analytical methods 11 (2019) 2064-2071. https://doi.org/10.1039/ c9ay00146h. [91] a. john, l. benny, a.r. cherian, s.y. narahari, a. varghese, g. hegde. electrochemical sensors using conducting polymer/noble metal nanoparticle nanocomposites for the detection of various analytes. journal of nanostructure in chemistry 11 (2021) 1-31. https://doi.org/10.1007/s40097-02000372-8. [92] i.f.a. karm, a.s. dwaish, o.a.a. dakhil. algae extracts as reduction agents for biosynthesis of silver nanoparticles for alternative medicinal compounds. eurasian chemical communications 4 (2022) 910920. https://doi.org/10.22034/ecc.2022.337634.1409 https://doi.org/10.5599/admet.1593 https://doi.org/10.1016/j.snb.2019.01.069 https://doi.org/10.1016/j.jelechem.2019.03.002 https://doi.org/‌10.1016/j.microc.2019.04.018 https://doi.org/‌10.1016/j.microc.2019.04.018 https://doi.org/10.1016/j.jpba.%20‌2020.113397 https://doi.org/10.1016/j.jpba.%20‌2020.113397 https://doi.org/10.1002/elan.201900148 https://doi.org/10.1016/j.jelechem.%20‌2020.113996 https://doi.org/10.1016/j.jelechem.%20‌2020.113996 https://doi.org/10.1039/c0gc00384k https://doi.org/‌10.1016/j.fct.2022.112907 https://doi.org/‌10.1016/j.fct.2022.112907 https://doi.org/10.1039/‌c9ay00146h https://doi.org/10.1039/‌c9ay00146h https://doi.org/10.1007/s40097-020‌00372-8 https://doi.org/10.22034/ecc.2022.337634.1409 y. m. baghelani et al. admet & dmpk 11(2) (2023) 117-134 134 [93] s. amin, a. hameed, n. memon, a.r. solangi, m. aslam, m.t. soomro. the efficacy of the nafion® blended ctab protected au nanoparticles for the electrochemical detection of tramadol in wastewater: a parametric investigation. journal of environmental chemical engineering 4 (2016) 38253834. https://doi.org/10.1016/j.jece.2016.08.010. [94] z. hassanvand, f. jalali. gold nanoparticles/cysteic acid modified electrode for simultaneous electrochemical determination of tramadol and paracetamol. analytical and bioanalytical chemistry research 6 (2019) 393-404. https://doi.org/10.22036/abcr.2019.170247.1305. [95] a. hojjati-najafabadi, s. salmanpour, f. sen, p. n. asrami, m. mahdavian, m.a. khalilzadeh. a tramadol drug electrochemical sensor amplified by biosynthesized au nanoparticle using mentha aquatic extract and ionic liquid. topics in catalysis 65 (2022) 587-594. https://doi.org/10.1007/s11244-021-01498-x. [96] f.s. yardımcı, s. mehmet, a. baykal. amperometric hydrogen peroxide biosensor based on cobalt ferrite–chitosan nanocomposite. materials science and engineering: c 32 (2012) 269–275. https://doi.org/10.1016/j.msec.2011.10.028. [97] r. devi, s. yadav, r. nehra, s. yadav, c.s. pundir. electrochemical biosensor based on gold coated iron nanoparticles/chitosan composite bound xanthine oxidase for detection of xanthine in fish meat. journal of food engineering 115 (2013) 207–214. https://doi.org/10.1016/j.jfoodeng.2012.10.014. [98] f. wang, s. hu. electrochemical sensors based on metal and semiconductor nanoparticles. microchimica acta 165 (2009) 1-22. https://doi.org/10.1007/s00604-009-0136-4. [99] t. madrakian, s. alizadeh, m. bahram, a. afkhami. a novel electrochemical sensor based on magneto ldh/fe3o4 nanoparticles@ glassy carbon electrode for voltammetric determination of tramadol in real samples. ionics 23 (2017) 1005-1015. https://doi.org/10.1007/s11581-016-1871-2. [100] s.a. memon, d. hassan, j.a. buledi, a.r. solangi, s.q. memon, i.m. palabiyik. plant material protected cobalt oxide nanoparticles: sensitive electrocatalyst for tramadol detection. microchemical journal 159 (2020) 105480. https://doi.org/10.1016/j.microc.2020.105480. [101] v. arabali, s. malekmohammadi, f. karimi. surface amplification of pencil graphite electrode using cuo nanoparticle/polypyrrole nanocomposite; a powerful electrochemical strategy for determination of tramadol. microchemical journal 158 (2020) 105179. https://doi.org/10.1016/j.microc.2020.105179. [102] m. khairy, c.e. banks. a screen-printed electrochemical sensing platform surface modified with nanostructured ytterbium oxide nanoplates facilitating the electroanalytical sensing of the analgesic drugs acetaminophen and tramadol. microchimica acta 187 (2020) 126. https://doi.org/10.1007/s00604-020-4118-x. [103] a. vazirirad, a. babaei, m. afrasiabi. sno2/α-fe2o3 hierarchical nanorods modified carbon paste electrode as the novel sensor for sensitive simultaneous determination of dopamine and tramadol. analytical and bioanalytical electrochemistry 13 (2021) 393-407. ©2023 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) https://doi.org/10.1016/j.jece.2016.08.010 https://doi.org/10.22036/abcr.2019.170247.1305 https://doi.org/10.1007/s11244-021-01498-x https://doi.org/10.1016/j.msec.2011.10.028 https://doi.org/10.1016/j.jfoodeng.2012.10.014 https://doi.org/10.1007/s00604-009-0136-4 https://doi.org/10.1007/s11581-016-1871-2 https://doi.org/10.1016/j.microc.2020.105480 https://doi.org/10.1016/j.microc.2020.105179 https://doi.org/10.1007/s00604-020-4118-x https://doi.org/10.1007/s00604-020-4118-x http://creativecommons.org/licenses/by/3.0/ manuscript doi: 10.5599/admet.3.1.144 1 admet & dmpk 3(1) (2015) 1-14; doi: 10.5599/admet.3.1.144 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index review the extended clearance model and its use for the interpretation of hepatobiliary elimination data gian camenisch* ,1 , julia riede 1,2 , annett kunze 1 , jörg huwyler 2 , birk poller 1 and kenichi umehara 1 1 division of drug metabolism and pharmacokinetics, integrated drug disposition section, novartis institutes for biomedical research, ch-4056 basel, switzerland 2 department of pharmaceutical sciences, division of pharmaceutical technology, university of basel, ch-4056 basel, switzerland *corresponding author: e-mail: gian.camenisch@novartis.com; tel.: +41-79-278-58-17; fax: +41-61-696-85-83 received: november 14, 2014; revised: december 18, 2015; published: march 31, 2015 abstract hepatic elimination is a function of the interplay between different processes such as sinusoidal uptake, intracellular metabolism, canalicular (biliary) secretion, and sinusoidal efflux. in this review, we outline how drugs can be classified according to their in vitro determined clearance mechanisms using the extended clearance model as a reference. the approach enables the determination of the ratedetermining hepatic clearance step. some successful applications will be highlighted, together with a discussion on the major consequences for the pharmacokinetics and the drug-drug interaction potential of drugs. special emphasize is put on the role of passive permeability and active transport processes in hepatic elimination. keywords extended clearance concept classification system; hepatic elimination; passive permeability; transporters extended clearance model historically, hepatic clearance models assumed that (i) the unbound drug concentration in blood is determining the hepatic clearance (metabolism and/or biliary excretion) and that (ii) there is no membrane transport barrier limiting access to the enzymes or transporters in the hepatocyte. improved models, however, reflect the physiological reality more precisely. in the liver drugs first have to overcome the membrane barrier separating the blood in the sinusoid from the cytosol of the hepatocytes. permeation across this barrier might occur by passive diffusion and/or active carrier-mediated transport. once in the cytosol drugs are subject to metabolism, efflux transporter-mediated canalicular (biliary) secretion and/or back-flux (active or passive) into the sinusoid. consequently, generally referred as the extended clearance model (ecm), the overall hepatic intrinsic drug clearance (clh,int) can be described as the interplay between all these processes as follows [1,2]: 𝐶𝐿ℎ,𝑖𝑛𝑡 = 𝑃𝑆𝑖𝑛𝑓 ∙ 𝐶𝐿𝑖𝑛𝑡 𝑃𝑆𝑒𝑓𝑓 + 𝐶𝐿𝑖𝑛𝑡 = (𝑃𝑆𝑖𝑛𝑓 ,𝑎𝑐𝑡 + 𝑃𝑆𝑖𝑛𝑓 ,𝑝𝑎𝑠 ) ∙ (𝐶𝐿𝑖𝑛𝑡 ,𝑚𝑒𝑡 + 𝐶𝐿𝑖𝑛𝑡 ,𝑠𝑒𝑐) 𝑃𝑆𝑒𝑓𝑓 ,𝑎𝑐𝑡 + 𝑃𝑆𝑒𝑓𝑓 ,𝑝𝑎𝑠 + 𝐶𝐿𝑖𝑛𝑡 ,𝑚𝑒𝑡 + 𝐶𝐿𝑖𝑛𝑡 ,𝑠𝑒𝑐 (1) http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:gian.camenisch@novartis.com camenisch et al. admet & dmpk 3(1) (2015) 1-14 2 where, psinf,act and psinf,pas are the active and passive hepatic influx clearances from the blood, respectively, clint,sec is the intrinsic biliary secretion clearance, and clint,met is the intrinsic metabolic clearance. pseff,act and pseff,pas describe the active and passive sinusoidal efflux from the hepatocytes back into the blood, respectively. psinf is the sum of psinf,act and psinf,pas, pseff is the sum of pseff,act and pseff,pas, clint is the sum of clint,sec and clint,met and clint,h is the overall intrinsic hepatic clearance. the extended clearance model allows the identification of the rate-determining hepatic clearance step for a given drug molecule [1]. depending on the relative contributions of the individual processes in eq. (1), four different cases can be distinguished (fig. 1). while (passive) hepatic uptake is the rate-determining step for ecm class 1 compounds, the sum of metabolism and efflux transporter-mediated biliary elimination is predicted to be the rate-limiting step for ecm class 2 compounds. the overall hepatic uptake (sum of active and passive) is projected to be rate-determining for ecm class 3 compounds, whereas the overall intrinsic clearance of ecm class 4 compounds is dependent on the interplay of all processes involved in hepatic elimination (namely metabolism, uptake, and efflux). figure 1. rate-determining hepatic clearance processes (inner panels) derived from the extended clearance model (ecm) based on different pre-requisites (outer panels) and the assumption thyat pseff,pas is equal to psinf,pas and that pseff,act equals zero (eq. (1)). viewing the liver as a single compartment and assuming that drug molecules are distributing instantaneously and homogeneously within the liver upon entering (well-stirred liver model) the hepatic blood clearance (clh) can be calculated with: 𝐶𝐿ℎ = 𝑄ℎ ∙ 𝑓𝑢,𝑏 ∙ (𝑃𝑆𝑖𝑛𝑓 ,𝑎𝑐𝑡 + 𝑃𝑆𝑖𝑛𝑓 ,𝑝𝑎𝑠 ) ∙ (𝐶𝐿𝑖𝑛𝑡 ,𝑚𝑒𝑡 + 𝐶𝐿𝑖𝑛𝑡 ,𝑠𝑒𝑐) 𝑄ℎ ∙ (𝑃𝑆𝑒𝑓𝑓 ,𝑎𝑐𝑡 + 𝑃𝑆𝑒𝑓𝑓 ,𝑝𝑎𝑠 + 𝐶𝐿𝑖𝑛𝑡 ,𝑚𝑒𝑡 + 𝐶𝐿𝑖𝑛𝑡 ,𝑠𝑒𝑐) + 𝑓𝑢 ,𝑏 ∙ (𝑃𝑆𝑖𝑛𝑓 ,𝑎𝑐𝑡 + 𝑃𝑆𝑖𝑛𝑓 ,𝑝𝑎𝑠 ) ∙ (𝐶𝐿𝑖𝑛𝑡 ,𝑚𝑒𝑡 + 𝐶𝐿𝑖𝑛𝑡 ,𝑠𝑒𝑐) (2) admet & dmpk 3(1) (2015) 1-14 extended clearance concept classification system (ecccs) doi: 10.5599/admet.3.1.144 3 where qh is the hepatic blood flow (20.7 ml/(min·kg)) and fu,b is the unbound fraction in blood. the clearance parameters of the extended clearance model can be determined experimentally as discussed in full detail elsewhere [1,3]. in brief, psinf,act and psinf,pas can be assessed by uptake experiments in suspended human hepatocytes (shh), clint,sec can be determined in human sandwich-cultured hepatocyte incubations (generally assuming that metabolism in this system is negligible) whereas clint,met, assuming absence of non-oxidative metabolism, is usually experimentally evaluated in human liver microsomal (hlm) incubations. efflux over the sinusoidal membrane from hepatocytes back into the blood is frequently assumed to occur via passive diffusion only (i.e. pseff,act = 0). the experimental determination of this parameter is difficult. in the absence of measured data, the passive sinusoidal efflux is usually assumed to be equal to the passive influx (i.e. pseff,pas = psinf,pas). the individual clearances can finally be fed into eq. (1) and eq. (2) to predict clh,int and clh of a drug. table 1 provides a compilation of experimentally determined intrinsic clearances for a dataset of 19 physicochemically diverse compounds from the literature together with corresponding in vivo reference data and the ecm classification according to fig. 1 [1, 4-7]. table 1. experimental human hepatic process clearances from in vitro assays and corresponding in vivo reference data. experimental in vitro data in vivo reference data compounds ecm/ecc psinf,act pseff = psinf,pas clint,met,hlm clint,met,hh clint,sec fu,b clh,obs fnren fnmet ml/(min·kg) ml/(min·kg) ml/(min·kg) ml/(min·kg) ml/(min·kg) ml/(min·kg) lovastatin acid 3/1 165.1 145.5 459.0 119.6 0.0 0.08 11.4 0.10 0.90 simvastatin acid 3/1 116.1 297.9 769.2 nd 1.7 0.11 25.2 0.13 0.87 propanolol 4/2 300.7 276.3 110.8 29.2 6.8 0.11 12.8 0.01 0.99 quinidine 4/2 229.4 109.3 28.4 18.0 5.1 0.27 8.7 0.19 0.81 verapamil 2/2 0.0 258.2 127.7 33.4 8.1 0.13 13.7 0.03 0.97 ketoconazole 2/2 0.0 1568.5 97.4 nd 29.6 0.02 3.9 0.03 0.97 cerivastatin 4/2 221.5 243.8 46.9 nd 0.0 0.02 3.4 0.00 1.00 fluvastatin 4/2 218.7 325.5 146.8 nd 0.0 0.04 7.0 0.00 1.00 pitavastatin 4/2 364.3 258.7 17.7 nd 0.0 0.07 3.5 0.00 nd aliskiren 3/3 32.3 25.4 89.2 nd 31.2 0.70 11.3 0.25 0.10 cimetidine 3/3 3.0 3.6 528.7 3.2 0.2 0.84 2.7 0.84 0.14 digoxin 3/3 20.0 6.9 24.2 nd 18.4 0.82 4.6 0.66 0.04 cyclospsorine a 3/4 113.2 41.9 77.6 13.5 9.1 0.03 3.1 0.01 0.96 atorvastatin 4/4 140.4 57.7 64.6 254.4 11.8 0.08 5.9 0.02 0.69 furosemide 4/4 11.1 23.9 19.0 0.9 1.2 0.03 0.4 0.66 0.01 ciprofloxacin 4/4 7.0 22.9 22.0 nd 0.0 0.69 4.5 0.60 0.12 valsartan 4/4 16.0 18.5 4.1 nd 21.5 0.09 0.6 0.29 0.11 pravastatin 4/4 57.9 36.0 0.9 5.3 2.2 0.97 10.4 0.47 0.30 rosuvastatin 4/4 27.2 24.8 1.5 nd 5.7 0.17 nd 0.30 0.10 nd: not determined or experimental data subject to a high degree of uncertainty as discussed elsewhere [4]. with exception of the hepatocyte turn-over data [5,6] all hepatic in vitro and in vivo data were taken from previous in-house manuscripts [1,4,7]. ecm and ecc class assignment was performed according to fig. 1 and fig. 6, respectively (underlying working principle: class 1 or 3 if 2·pseff < (clint,met,hlm + clint,sec), otherwise class 2 or 4). metabolic clearance data from human liver microsomes and human hepatocytes are labeled with the subscripts hlm and hh, respectively. fnmet (= clmet,obs/cltot,obs) and fnren (= clren,obs/cltot,obs) values were calculated from the observed total (cltot,obs), renal (clren,obs) and metabolic (clmet,obs) clearances as derived from peroral human mass balance studies taking into consideration the estimated absolute oral bioavailability (f). thereof, fnh (=1fnren) and fnsec (= fnhfnmet) values discussed in this manuscript can be calculated. camenisch et al. admet & dmpk 3(1) (2015) 1-14 4 role of membrane permeability in hepatic elimination the likelihood that drugs will be subject to enzyme and/or canalicular efflux transporter activities in the liver depends on the extent of their sinusoidal membrane permeation. for the compounds in our dataset (table 1) this is illustrated in fig. 2 depicting the relationship between (in vitro) sinusoidal passive uptake (psinf,pas) and the fractional contribution of metabolism (fnmet, panel a) or biliary secretion (fnsec, panel b) to overall in vivo elimination. figure 2. correlation of sinusoidal hepatic uptake clearances (psinf,pas or psinf) with the fractional contributions of metabolic (panel a), biliary (panel b), hepatic (panel c) and renal (panel d) clearance to overall in vivo elimination. an increase in lipophilicity is well known to promote the passive permeability potential of drugs as well as their affinity to drug metabolizing enzymes [8]. above observed inter-relationship, as depicted in fig. 2 (panel a), is therefore not very surprising. the association of permeability and enzyme activity allows to define an approximate sinusoidal permeability threshold above which drug elimination seems predominantly driven by metabolism (> 200 ml/(min·kg)). the corresponding plot of psinf,pas vs fnsec reveals that compound recognition by canalicular efflux transporters, in contrast to fnmet, is by some means inversely correlated with rising sinusoidal permeability (fig. 2, panel b). fig. 3 shows the contribution of the measured in vivo metabolic clearance to the in vivo hepatic clearance (= fnmet/fnh) for the compounds in our dataset. not astonishingly, the compounds with a high (total) sinusoidal permeability are predominantly cleared via hepatic metabolism, while the contribution of metabolism generally decreases for lower permeability compounds apparently accompanied by an increasing contribution of active canalicular secretion. admet & dmpk 3(1) (2015) 1-14 extended clearance concept classification system (ecccs) doi: 10.5599/admet.3.1.144 5 figure 3. contribution of in vivo metabolism to overall hepatic clearance for the dataset in table 1. the dotted line separates the high permeability compounds (green bars) from the low permeability drugs (orange bars) as determined by uptake experiments in suspended hepatocytes (psinf threshold: 200 ml/(min·kg)). hepatic versus non-hepatic elimination figure 2 (panel c) illustrates the relationship between passive sinusoidal uptake (psinf,pas) and the observed (in vivo) fractional contribution of hepatic clearance to overall drug elimination fnh. from this illustration it appears that a sinusoidal permeability of above 40-60 ml/(min·kg) will result in predominant hepatic clearance (> 80 %). numerous data demonstrates that permeability also exerts a key role in renal elimination similar to the one discussed here for liver [4]. it could be demonstrated that for the highly permeable compounds, despite significant tubular secretion, reabsorption from the tubule back into the blood is so extensive that overall renal clearance for these compounds will be low. for the low permeable compounds on the other hand reabsorption, in agreement with their reduced permeability potential, will be much less. as illustrated in fig. 4, for all compounds in our dataset with a (passive) sinusoidal permeability beyond 40 ml/(min·kg) renal clearance is a minor route of elimination (fnren < 20 %). for the lower permeable compounds the fractional contribution of renal clearance to total body clearance exceeds 20% or even represents the predominant route of elimination suggesting that hepatic and renal elimination are somehow complementary with regards to the role of sinusoidal permeability in drug elimination. figure 4. contributions of the observed renal (red bars) and hepatic (blue bars) clearance pathways to the total body clearance (cltot,obs = clh,obs + clren,obs) for the compounds in table 1. the dotted line separates the compounds with a psinf,pas value greater (left part) or lower (right part) than 40 ml/(min·kg)). camenisch et al. admet & dmpk 3(1) (2015) 1-14 6 fig. 2 (panel d) illustrates the relationship between fnren and total sinusoidal uptake (psinf) for the present dataset. as discussed above, a low permeability is hindering hepatic elimination pushing drugs towards alternative elimination routes (with decreasing lipophilicity plasma protein binding usually decreases which would result in increased glomerular filtration process, possibly increased active tubular secretion and less reabsorption) [4]. beyond a psinf value of 100 ml/(min·kg)) renal elimination becomes very unlikely. consequently, following integration of the information from this and the previous chapter, it becomes apparent that total sinusoidal uptake is the predominant gatekeeper between almost exclusive metabolic hepatic elimination (> 200 ml/(min·kg)), mixed hepatic (metabolism plus biliary secretion, 100-200 ml/(min·kg)) and mixed hepatic/renal elimination (metabolism plus biliary and urinary secretion, < 100 ml/(min·kg)) (fig. 5). exclusive renal clearance is expected only if psinf is approaching a very low value or if substantial entero-hepatic circulation is taking place. yet, renal excretion might become the predominant elimination pathway for low permeable compounds with about psinf < 30 ml/(min·kg) (fig. 2, panel d). for compounds with sinusoidal uptake between 30-60 ml/(min·kg), mixed urinary/biliary elimination seems to be the most probable clearance pathway. consequently, the relationship between psinf,pas and fnsec as shown in figure 2 (panel b) likely needs to be interpreted as a bell-shaped curve with a psinf,pas maximum at around 30 ml/(min·kg) (maximal fnsec ≈ 0.6-0.7). figure 5. schematic illustration of the change in elimination route as a function of sinusoidal uptake (psinf). fnmet, fnren and fnsec denote the fractional contributions of (hepatic) metabolic, urinary and biliary elimination to overall clearance, respectively. above a psinf threshold of about 60 ml/(min·kg) liver is the expected primary clearance organ whereas below 30 ml/(min·kg) renal elimination is likely predominant. extended clearance concept classification system the extended clearance concept (ecc) is the intangible attempt to complete the principles of the extended clearance model (fig. 1) with the gatekeeping role of sinusoidal influx as discussed above (fig. 5) to allow prediction of in vivo pharmacokinetic performance of drugs from in vitro permeability and metabolism data as illustrated in fig. 6. drugs with a high apparent hepatic intrinsic clearance are removed from the blood essentially as fast as they can be delivered to the liver, i.e. independent of protein binding and intrinsic enzyme and canalicular efflux transporter activities. therefore, the elimination of such drugs is highly dependent upon liver blood flow and the inherent ability to cross the sinusoidal membrane rapidly. consequently, according to the well-stirred liver model, a compound can be ranked as highly permeable if admet & dmpk 3(1) (2015) 1-14 extended clearance concept classification system (ecccs) doi: 10.5599/admet.3.1.144 7 clh,int = psinf,pas >> qh (eq. (1) and eq. (2)). from above discussion it is evident that a (passive) sinusoidal permeability above 60 ml/(min·kg) results in an almost complete hepatic clearance (fig. 2, panel c and fig. 5). thus, a compound can be ranked as highly permeable if psinf,pas ≥ 3-fold qh, a threshold substantiating the conceptual approximation psinf ≈ pseff ≈ psinf,pas pertinent to all compounds identified to be hepatic uptake transporter substrates besides demonstrating a high intrinsic permeation potential across the sinusoidal membrane of hepatocytes. hence, it becomes evident that uptake transporter effects for highly permeable compounds are minimally contributing to overall hepatic clearance while they are expected to become important/predominant for the hepatic clearance of lower permeable compounds. applying this fundamental relationship, the 19 compounds in our dataset were assigned to the four different ecc classes as summarized in table 1. figure 6. predominant routes of drug elimination and potential transporter effects on (hepatic) drug disposition according to the extended clearance concept classification system (ecccs). thresholds as defined by the biopharmaceutics drug disposition classification system (bddcs) are provided for comparative reasons also [1,8]. taking all above principles into account it is appropriate to conclude that the major route of elimination for ecc class 3 and 4 compounds is renal and biliary excretion of unchanged drug whereas metabolism is the predominant elimination pathway for ecc class 1 and 2 compounds as previously concluded by others also [8]. however, ecc allows a compound classification of drug molecules based on their in vitro determined (hepatic) clearance parameters in contrast to the biopharmaceutics drug disposition classification system (bddcs) which allocates drugs into four classes according to their in vivo metabolism and solubility potential. both concepts are closely related and have in common that they recognize that the fundamental parameter controlling (hepatic) drug disposition is the compound class-dependent interplay between transporters, enzymes and membrane permeability. both systems provide a similar rational on the predominant routes of drug elimination and the potential effect of transporters on (hepatic) drug disposition as illustrated in fig. 6. yet, assignment into bddcs relies on clinical elimination information and is therefore not really applicable for early drug development. camenisch et al. admet & dmpk 3(1) (2015) 1-14 8 ecc and ivive the correlation between the in vitro predicted (clh,pred) and in vivo observed (clh,obs) hepatic clearances for our dataset based on the mechanistic in vitro-in vivo extrapolation (ivive) method represented by eq. (1) and eq. (2) is illustrated in fig. 7 (panel a). the approach reveals an excellent correlation with 11 out of 18 drugs predicted within two-fold deviation from the clinically observed value. the prediction accuracy in terms of average fold error (afe) and geometric mean fold error (gmfe) was 0.92 and 1.55, respectively. present results demonstrate that the extended clearance concept model is by far exceeding the accuracy and performance of other ivive prediction methods for hepatic clearance based only on in vitro metabolism (fig. 7, panels b and c for microsomes (afe = 0.76, gmfe = 2.11)) and hepatocytes (afe = 0.32, gmfe = 3.56), respectively) or sinusoidal uptake (fig. 7, panel d, afe = 1.71, gmfe = 1.84) data. figure 7. comparison of the reported (clh,obs) and predicted (clh,pred) hepatic clearances using different ivive methods. panel a represents the relationship according to the extended clearance model (ecm). the panels b and c show the predictions based on in vitro metabolism data from human microsomes (hlm) and hepatocytes (hh), respectively, whereas panel d is the representation based on sinusoidal uptake data from suspended human hepatocytes (shh). red diamonds, green squares, blue triangles and orange circles show the class 1, 2, 3 and 4 assignments according to ecc (table 1). the blue line represents the line of unity for each panel. in accordance with previous research, the correlation analysis in fig. 7 reveals a systematic underprediction of in vivo hepatic clearance when using metabolic turn-over data from human hepatocyte incubations whereas sinusoidal uptake data from suspended human hepatocytes tend to over-predict the in vivo situation [9,10]. metabolic clearance data from human liver microsomes on the other hand are highly scattered along the line of unity providing underand over-estimations of the observed human hepatic clearance. the potential rate-determining role of the hepatocyte membrane (missing in the microsomal system) is extensively discussed in literature. in line with this theory more in-depth data examination reveals that metabolism information from microsomes and hepatocytes generally provide reasonable results for the ecc class 1 and 2 compounds while hepatocyte uptake data seem to be highly predictive for the in vivo hepatic clearance of ecc class 3 compounds. ivive based on the extended mechanistic model surely works best for ecc class 4 compounds though. imposing these observations on admet & dmpk 3(1) (2015) 1-14 extended clearance concept classification system (ecccs) doi: 10.5599/admet.3.1.144 9 the principles of ecc the expected performance of the different ivive approaches can be summarized as depicted in fig. 8. it is self-explanatory that with increasing contribution of renal clearance the predictability of all hepatobiliary ivive tools for overall (total) clearance is decreasing (tendency for underestimation). based on above discussion, this is mainly true for ecc class 3 and 4 compounds while the overall prediction performance for ecc class 1 and 2 compounds likely remains good. in-house research intending to improve bottom-up predictions for overall human clearance using different in vitro and in vivo approaches is currently ongoing. figure 8. anticipated hepatobiliary ivive accuracy of different in vitro tools: metabolism assessment in human liver microsomes (hlm) or human hepatocytes (hh), uptake data from suspended human hepatocytes (shh) and integrated approach with the extended clearance model (ecm). for the assignment the following assumptions were made: (i) absence of significant biliary secretion (i.e. clint,sec << clint,met), (ii) absence of phase ii metabolism and, (iii) no or down-regulated transporter activity in hh. static ddi predictions a perpetrator drug may inhibit any active clearance pathway contributing to the total hepatic elimination of a substrate. accordingly, based on eq. (1) (assuming pseff,act = 0 and pseff,pas = psinf,pas), the overall hepatic intrinsic clearance in the presence of a perpetrator (clh,int,i) can be expressed as follows [1,7]: 𝐶𝐿ℎ,𝑖𝑛𝑡 ,𝑖 = � 1 − 𝑓𝑖,𝑖𝑛𝑓 ∙ 𝑃𝑆𝑖𝑛𝑓,𝑎𝑐𝑡 + 𝑃𝑆𝑖𝑛𝑓 ,𝑝𝑎𝑠 ∙ [ 1 − 𝑓𝑖,𝑚𝑒𝑡 ∙ 𝐶𝐿𝑖𝑛𝑡 ,𝑚𝑒𝑡 + (1 − 𝑓𝑖,𝑠𝑒𝑐) ∙ 𝐶𝐿𝑖𝑛𝑡 ,𝑠𝑒𝑐] [𝑃𝑆𝑖𝑛𝑓 ,𝑝𝑎𝑠 + (1 − 𝑓𝑖,𝑚𝑒𝑡 ) ∙ 𝐶𝐿𝑖𝑛𝑡 ,𝑚𝑒𝑡 + 1 − 𝑓𝑖,𝑠𝑒𝑐 ∙ 𝐶𝐿𝑖𝑛𝑡 ,𝑠𝑒𝑐] where fi,inf, fi,sec and fi,met denote the inhibited fractions of active influx, canalicular secretion, and metabolism, respectively. a fi value of zero thereby indicates no inhibition whereas a value of one refers to (3) camenisch et al. admet & dmpk 3(1) (2015) 1-14 10 complete inhibition. based on this relationship the hepatic clearance in the presence of any perpetrator compound (clh,i) can be anticipated in accordance with eq. (2). following oral (po) administration of a drug and its perpetrator, assuming the presence of hepatic and a non-hepatic (e.g. renal) elimination pathways and that the perpetrator drug only affects active processes in the liver, the exposure (auc) fold-change (expressed as aucpo,i/aucpo) can be described as follows: 𝐴𝑈𝐶𝑝𝑜 ,𝑖 𝐴𝑈𝐶𝑝𝑜 = 𝐹ℎ,𝑖 𝐹ℎ ∙ 1 𝑓𝑛ℎ ∙ 𝐶𝐿ℎ,𝑖 𝐶𝐿ℎ� + 1 − 𝑓𝑛ℎ where, fh (= 1-clh/qh) and fh,i (= 1-clh,i/qh) are the fractions of the oral dose escaping hepatic first-pass in the absence and presence of a perpetrator, respectively. under the additional assumption that the liver is the only clearance organ (i.e. fnh = 1) eq. (4) simplifies to: 𝐴𝑈𝐶𝑝𝑜 ,𝑖 𝐴𝑈𝐶𝑝𝑜 = 𝐶𝐿ℎ,𝑖𝑛𝑡 𝐶𝐿ℎ,𝑖𝑛𝑡 ,𝑖 in an abbreviated manner, based on eq. 5 and previous discussion, the drug-drug interaction (ddi) potential for the four ecm cases in fig. 1 can therefore be represented as shown in fig. 9. figure 9. compound class-dependent ddi prediction (inner panel) according to eq. 5 for the four different ecm cases based on different pre-requisites (outer panel). i refers to a change in clearance in the presence of any perpetrator compound. fig. 10 depicts the predicted static ddi risk assessment of 3 representative compounds from our dataset (namely the ecm class 2 compound verapamil, digoxin as a typical ecm class 3 compound and pravastatin representing the ecm class 4) according to eq. (4) and eq. (5), assuming 90% inhibition of the respective (4) (5) admet & dmpk 3(1) (2015) 1-14 extended clearance concept classification system (ecccs) doi: 10.5599/admet.3.1.144 11 processes (i.e. fi,inf = fi,met = fi,sec = 0.9, not taking into consideration fractional contributions of different enzymes or transporters to overall metabolic or transport clearance). in the absence of a renal clearance contribution (fig. 10, panel a), in accordance with the extended clearance model, for the ecm class 2 compound in our selection, metabolism is identified as the major clearance mechanism primarily responsible for causing ddi’s. selective inhibition of active hepatic influx or biliary secretion has no significant effect on exposure though. a similarly distinct overall ddi behavior can be derived for the ecm class 3 compound digoxin with sinusoidal transporter inhibition being the major mechanism causing the interaction. for the ecm class 4 compound pravastatin, assuming absence of a renal clearance contribution, we predict a substantial auc change upon concomitant inhibition of hepatic uptake, biliary secretion and metabolism whereas inhibition of the single clearance pathways results in comparatively moderate auc ratios. taking the clinically observed fnh values according to eq. (4) into account, the individual as well as the overall ddi risk assessment of verapamil is only marginally effected whereas the impact on the projected exposure changes for digoxin and pravastatin is significant (e.g. the overall aucpo,i/aucpo ratio for pravastatin decreases about 8-fold) (fig. 10, panel b). this concept is well reflected by ecm and can be rationalized by the process inter-dependencies as discussed above. it is noteworthy to mention at this point, that the extended clearance model represents the four extremes of hepatobiliary elimination and that most drugs settle somewhere in-between the ultimate limits given in fig. 1. to eventually predict the overall ddi potential of drug molecules it is therefore essential to (quantitatively) assess all the individual hepatic process contributions as defined by eq. (1). figure 10. static victim drug ddi predictions for verapamil (cyp2c8, cyp3a and p-gp), digoxin (p-gp, cyp3a and unknown sodium-dependent sinusoidal uptake transporter) and pravastatin (bcrp, mrp2, oatp1b1, oatp1b3, oatp2b1 and cyp3a4) according to eq. (5) (panel a) and eq. (4) (panel b). the projected auc fold-changes with fi,inf = fi,met = fi,sec = 0.9 are represented as follows: metabolism inhibition only (blue bars), exclusive sinusoidal active uptake inhibition (red bars), inhibition of canalicular efflux (light green bars) and, simultaneous inhibition of all active processes together (yellow bars). neglecting time-dependent concentration changes, static ddi predictions are expected to provide conservative (worst-case) estimates for the effective in vivo situation [1]. nevertheless, ddi assessments according to eq. (4) are generally in good agreement with clinical ddi data for ecm class 1 and 2 compounds as discussed elsewhere [7]. however, with increasing fnren (i.e. decreasing fnh), the predictions become less reliable for ecm class 3 and 4 compounds, often resulting in significant under-estimations of the effective ddi risk observed in clinics (e.g. pravastatin exposure in the presence of cyclosporine a (inhibitor of oatp’s, oat’s, ntcp, p-gp, bcrp, mrp2, cyp3a4 and ugt’s) was reported to increase about 20-fold, significantly exceeding the theoretically possible hepatic ddi potential of 2.6-fold (fig. 10)) [7]. the main reason for this observation is that a significant renal secretion process for these low permeable camenisch et al. admet & dmpk 3(1) (2015) 1-14 12 compounds is typically governed by active transporter processes which often are concomitantly inhibited by perpetrator drugs also interfering with the sinusoidal uptake carrier system in the liver. consequently, for conservative bottom-up ddi estimates the physiologically less appropriate eq. (5) should be preferred as it accommodates for possible cross-reactivity on hepatic and renal (transporters) systems (e.g. the clinical observation for pravastatin in the presence of cyclosporine a is well predicted with eq. (5) (fig. 10)). nevertheless, in combination with clinical ddi data and applying a top-down approach, comparative static ddi assessments according to eq. (4) and eq. (5) can be extremely helpful in revealing alternative (extrahepatic) active elimination processes as discussed in full detail elsewhere [7]. prediction of unbound intracellular hepatic drug concentrations figure 11. ecm class-dependent effect on the unbound liver-to-capillary blood concentration ratio (kpuu) according to eq. 6. according to the so-called free drug hypothesis, at steady-state and in the absence of a membrane transport barrier limiting access to the hepatocyte, the unbound (free) intracellular concentration ch,u is equal to the unbound blood concentration cb,u (i.e. ch,u/cb,u = 1). however, according to the extended clearance model, ch,u is not only governed by uptake and protein binding but also by all intracellular elimination processes. therefore, based on eq. (1), for the liver this relationship needs to be rewritten as follows [11]: 𝐶ℎ,𝑢 𝐶𝑏,𝑢 = 𝐾𝑝𝑢𝑢 = 𝑃𝑆𝑖𝑛𝑓 𝑃𝑆𝑒𝑓𝑓 + 𝐶𝐿𝑖𝑛𝑡 (6) admet & dmpk 3(1) (2015) 1-14 extended clearance concept classification system (ecccs) doi: 10.5599/admet.3.1.144 13 where, kpuu is the unbound liver-to-capillary blood concentration ratio. in line with ecm, the impact of the relative process contributions on kpuu can be summarized as shown in fig. 11. the knowledge of the relevant (effective) intrahepatic concentrations is important when performing ddi risk assessments (compound acting as inhibitor) on active processes such as metabolism and/or canalicular efflux. however, due to experimental inaccessibility of ch,u, static perpetrator risk calculations for the liver generally refer to cb,u (usually expressed as cb,u /ki, where ki denotes the inhibition constant on a particular process). based on the extended clearance model it becomes evident that the use of this substitute might lead to a considerable misjudgment of the ddi potential of perpetrator drugs though. depending on the major ratedetermining process driving hepatic elimination significant under(e.g. kpuu equals 2.4 for ecm class 4 compound pravastatin) or over(e.g. ch,u is about 80-fold less than cb,u for ecm class 3 compound cimetidine) estimations of the real interaction risk might result. the use of systemic unbound drug concentrations (or unbound drug concentrations in the portal vein) is thus likely insufficient to properly assess the actual drug-drug interaction potential on intracellular enzymes/transporters for class 1, 3 and 4 compounds as illustrated in table 2. table 2. effect of kpuu on ch,u and the ddi risk outcome for perpetrator drugs. ecm class effect of kpuu on ch,u (eq. (6)) risk assessment with cb,u /ki 1 ch,u << cb,u overestimation 2 ch,u = cb,u predictive 3 ch,u < or > cb,u underor overestimation (generally overestimation as usually psinf < clint) 4 ch,u < or > cb,u underor overestimation (generally underestimation as usually psinf > pseff) although currently not requested by health authorities present ecm-based principles might change the way we are doing perpetrator ddi risk assessments for intracellular hepatic processes in the future, all the more the mechanistic ivive method presented in this manuscript was demonstrated to provide quantitative reliable estimates of clh,obs and therewith of kpuu (see above). further in-house research on this this topic is currently ongoing. conclusions in this review, we have illustrated the usefulness of the extended clearance model for classifying drug compounds depending on the relative contributions of the individual hepatic elimination processes. the ecm classification system based upon easily accessible in vitro data for metabolism, sinusoidal transport and canalicular efflux allows the quantitative bottom-up assessment of a series of (pharmacokinetic) parameters such as the hepatic clearance, the unbound liver-to-capillary blood concentration ratio and the static prediction of the ddi potential of drug molecules. the method reveals the major role of sinusoidal uptake as a gatekeeper for drug elimination, provides insight into the performance of different ivive approaches for hepatic clearance prediction and highlights potential limitations of currently applied drugdrug interaction risk assessment approaches. consequently, following a thoughtful implementation, the approach may facilitate the compound selection process in pharmaceutical research and improve some of the compound profiling methodologies (e.g. ddi risk assessment) applied in drug development. camenisch et al. admet & dmpk 3(1) (2015) 1-14 14 acknowledgements: the author wishes to acknowledge the many novartis drug metabolism and pharmacokinetic scientists of basel, switzerland, who have supported this work. special thanks go to dr markus zollinger for his critical evaluation of this manuscript. references [1] g. camenisch and k. umehara, biopharmaceutics & drug disposition 33 (2012) 179-1940. [2] y. shitara, h. sato and y. sugiyama, annual review of pharmacology and toxicology 45 (2005) 689723. [3] k. umehara and g. camenisch, pharmaceutical research 29 (2012) 603-617. [4] a. kunze, b. poller, j. huwyler and g. camenisch, drug metabolism and drug interactions in press. [5] d. mcginnity, m. soars, r. urbanowicz and r. riley, drug metabolism and disposition 32 (2004) 1247-1253. [6] d. hallifax, e. turlizzi, u. zanelli and j. houston, european journal of pharmaceutical sciences 45 (2012) 570-574. [7] a. kunze, j. huwyler, b. poller, h. gutmann and g. camenisch, journal of pharmaceutical sciences 103 (2014) 994-1001. [8] l. benet, g. amidon, d. barends, h. lennernas, j. polli, v. shah, s. stavchansky and l. yu, pharmaceutical research 25 (2008) 483-48. [9] m. chiba, y. ishii and y. sugiyama, aaps journal 11 (2009) 262-276. [10] d. hallifax, h. rawden, n. hakooz and j. houston, drug metabolism and disposition 33 (2005) 1852-1858. [11] y. shitara, k. maeda, k. ikejiri, k. yoshida, t. horie and y. sugiyama, biopharmaceutics & drug disposition 34 (2013) 45-78. ©2015 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ development and application of a simple pharmacokinetic model that quantitatively describes the distribution and elimination of the commonly measured proteins doi: https://doi.org/10.5599/admet.1570 57 admet & dmpk 11(1) (2023) 57-80; doi: https://doi.org/10.5599/admet.1570 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper development and application of a simple pharmacokinetic model that quantitatively describes the distribution and elimination of the commonly measured proteins david g. levitt,1* michael d. levitt2 1 department of integrative biology and physiology, university of minnesota, minneapolis, mn 55455, usa; levit001@umn.edu 2 medicine service, veterans affairs medical center, minneapolis, mn. 55417, usa; michael.levitt@va.gov *corresponding author: e-mail: levit001@umn.edu received: october 25, 2022; revised: january 22, 2023; published: january 26, 2023 abstract increased plasma concentrations of a variety of cellular enzymes (alanine transaminase, aspartate aminotransferase, alkaline phosphatase, amylase, etc.) are commonly used as routine screening tests for a range of conditions. an increased concentration usually is assumed to result from an increased rate of delivery to the plasma. factors such as decreased metabolism or excretion or altered extravascular distribution usually are ignored. as a prelude to a detailed analysis of all the factors producing altered plasma enzyme levels, we have reviewed the relevant literature describing the pharmacokinetics (pk) of 13 of the commonly measured plasma proteins and developed a pk model that provides a simple physiological description of all the data. our model starts with the general 3-compartment, 6-parameter system previously developed for albumin and interprets the fluxes in terms of unidirectional sieved protein convectional volume flows from the plasma to the two tissue compartments and equal lymph flows returning to the plasma. this greatly constrains the model such that each protein is characterized by only two adjustable parameters (plasma clearance and sieving factor). in addition to accurately fitting the plasma kinetics, the model can accurately describe the tissue and lymph protein pk. for example, it can describe the thoracic duct lymph protein concentration following an intravenous infusion or the plasma concentration following a subcutaneous tissue injection. this simple model provides a satisfactory framework for the pk of 12 of the 13 proteins investigated. the glycoprotein intestinal alkaline phosphatase is the exception, requiring the addition of a liver recycling compartment involving the asialoglycoprotein receptor. ©2023 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords protein; pharmacokinetics; alkaline phosphatase; alanine transaminase; aspartate aminotransferase; amylase; albumin; liver; pancreas introduction the standard approach to characterizing protein pharmacokinetics (pk) is via the classical 2-compartment (4 parameters) [1,2] or 3-compartment (6 parameters) [3,4] modeling. for most clinical purposes, 2 compartments are sufficient [1,2]. however, in order to accurately fit the details of the plasma concentration https://doi.org/10.5599/admet.1570 https://doi.org/10.5599/admet.1570 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:levit001@umn.edu mailto:michael.levitt@va.gov mailto:levit001@umn.edu http://creativecommons.org/licenses/by/4.0/ d.g. levitt and m.d. levitt admet & dmpk 11(1) (2023) 57-80 58 curve, usually 3 compartments are required [3,4]. each protein is characterized by a six-parameter set determined by fitting the model to the plasma protein concentration following bolus injection. the recent proliferation of therapeutic proteins, especially monoclonal antibodies, has led to an interest in developing physiologically based pharmacokinetic modeling (pbpk) of these proteins [5-8]. in this approach, each of the major organs is described in terms of their physiological parameters (blood flow, capillary permeability, lymph flow, interstitial volume, protein catabolism, etc.) and the pk is related to these parameters. the major theoretical advantage of this approach is that once these parameters are determined for, e.g., the human, they allow one to predict the human protein pharmacokinetics of a novel protein with a minimum of adjustable parameters. in addition, since each organ is described individually, one can predict their individual tissue protein pk, which is useful if trying to direct therapy to specific organs. although there has been great interest in the pk modeling of therapeutic proteins, there is another class whose pk has been almost completely ignored: the diagnostic enzymes such as alanine transaminase, aspartate aminotransferase, alkaline phosphatase, amylase, etc. pathological disruption of normal cellular structure may result in the release of these cellular proteins to the interstitial space, with subsequent transport to the plasma. because low concentrations of cellular enzymes are easily quantified via activity measurements, plasma enzyme values are widely used as routine screening tests for a variety of conditions as well as for evaluating the severity of specific disease states. these diagnostic enzymes are typically characterized by estimates of their “half-life”, based primarily on the time course of the fall in the plasma concentration following the resolution of the problem causing the enzyme elevation. lindena et al. [9] have presented a remarkably thorough compilation of this older data, comparing the reported half-lives for man, dog and rat. these estimates are only approximate because of the lack of quantitative information about the ongoing release of the enzymes and/or recycling from the tissue space. physicians tend to interpret plasma protein concentrations primarily as an indicator of a disordered rate of delivery to the plasma. details of the pk: rate of release into the blood, time course and details of the extravascular distribution, and site and rate of removal usually are ignored. a quantitative description of the pk of these commonly measured plasma proteins, with an emphasis on the roles of distribution and elimination rates as determinants of plasma concentrations, should aid in the clinical interpretation of the time course of the plasma concentration. in this review, we will first discuss the compartmental and pbpk approach to protein pk. we will show that, although the full pbpk approach is probably not feasible for most proteins, the use of a simpler physiological model, in which some organs are lumped together, is useful. we describe a new model that provides an accurate description of the plasma pk requiring, for a given species, only 2 adjustable parameters to characterize each protein. plasma protein pk are quite simple and can be described by a variety of kinetic models. an important constraint on the model validity is its ability to describe the simultaneous tissue or lymph concentration, an additional constraint that is rarely discussed. we show that, in addition to describing the plasma concentration, this model is surprisingly successful at predicting the lymph and tissue concentration, a result that provides additional support for the validity of the model. our model is much simpler and, therefore, more limited than some of the detailed pbpk models that have been developed for the monoclonals [8]. we believe that this simplicity and its ability to describe a wide range of proteins compensate for these limitations. we will first discuss and illustrate compartmental and physiological pk modeling by applying it to albumin, the most extensively studied plasma protein (section ii). then, in the following sections, this physiological model will be applied to the monoclonal antibodies (section iii) and the diagnostic enzymes (sections iv – viii). the resultant pk parameters for the 13 proteins investigated are summarized in table 1 in terms of the admet & dmpk 11(1) (2023) 57-80 protein pharmacokinetic model doi: https://doi.org/10.5599/admet.1570 59 compartmental parameters and in table 2 in terms of the physiological parameters. the range of proteins analyzed is limited by the available pk data. the ideal would be human data for the plasma time course of an intravenous (iv) bolus of the purified human protein. for the proteins where these data do not exist, we assumed that data from studies in large mammals (e.g., dog, lamb, baboon) can be extrapolated to the human. rat data is only sparingly employed because the extrapolation of rat pk data to humans is uncertain. table 1. summary of protein pharmacokinetics expressed in terms of the 5 compartmental time constants corresponding to the steady state parameters of figure 1a. all results are scaled to a 70 kg human. “kupffer” indicate liver kupffer cells and “sinusoidal” are liver parenchymal cells. (**a pk model with an additional liver compartment is used for recap. *tm for recap corresponds to the liver metabolic time constant.) protein species mw catabolic organ tm min tp1 min tp2 min tt1 min tt2 min albumin human 66,500 endothelium 16,750 4385 1929 4182 557 monoclonal antibody human mepolizumab 149,000 endothelium 21202 8770 3859 4182 557 infliximab 149,000 endothelium 11964 8770 3895 4182 557 alkaline phosphatase human recap** 140,000 liver 1100* 8770 3859 4182 557 placental 140,000 liver 7537 8770 3859 4182 557 amylase baboon 57,086 kidney 100 4385 1929 4182 557 lipase baboon 48000 kidney 100 4385 1929 4182 557 alanine transaminase dog 54,600 sinusoidal 2949 4385 1929 2323 310 aspartate aminotransferase dog mitochondrial 47,517 kupffer 49.1 4385 1929 2323 310 cytoplasmic 46,247 sinusoidal 689 4385 1929 2323 310 creatine kinase dog 81,000 kupffer 88.1 6265 2756 2323 310 lactate dehydrogenase lamb ld1 heart 138,000 sinusoidal 2871 1370 3859 726 619 ld5 skeletal muscle 138,000 kupffer 440 1370 3859 726 619 albumin pharmacokinetics: compartmental and physiological modeling the distribution and catabolism of albumin has been extensively studied and will serve as our basis for developing the compartmental and physiological models applied to all the other proteins. albumin capillary permeability must be very low in order to preserve the high plasma concentrations that maintain the colloidal osmotic pressure gradient between plasma and extravascular fluid. the permeability is not zero, however, and lymph drainage of the tissue space is required to maintain this concentration gradient. controversy exists concerning whether this slow albumin leak is via caveolar vesicle transport versus through fixed pores or tight junctions [10]. recent results showing that mice lacking endothelial caveolae have normal capillary albumin permeability [11] provide direct support for the pore mechanism. in addition, the proteins leak at rates are roughly inversely proportional to their molecular radius [12], a characteristic of molecules passing through pores rather than via transcytotic vesicle transport. an important question is whether this albumin leak represents a bidirectional diffusional permeability versus a unidirectional convective transport from the plasma to the tissue and return via lymph [13,14]. although some quantitative measurements of simultaneous plasma and lymph protein pk are supportive of the convection mechanism [12,13,15], these results do not rule out a small diffusional component. it is assumed in the most recent protein modeling that https://doi.org/10.5599/admet.1570 d.g. levitt and m.d. levitt admet & dmpk 11(1) (2023) 57-80 60 the leak is entirely convective with no significant diffusional component [5,6,16], and we will make the same assumption. table 2. summary of protein pharmacokinetics expressed in terms of the lumped physiological parameters of figure 1b: plasma metabolic clearance clp, filtration fractions f1, f2, and lymph flows l1 l2. all results are scaled to a 70 kg human using an assumed plasma volume vp = 2800 ml. for each species, the 4 parameters l1, l2, v1 and v2 are adjusted as described and assumed constant for all proteins studied in that species. only 2 parameters are adjusted for each protein: clp and f ≈f1≈f2. for human and baboon: v1/vp =1.3, v2/vp= 0.4. for dog and lamb: v1/vp =1.0, v2/vp= 0.3. “kupffer” indicate liver kupffer cells and “sinusoidal” are liver parenchymal cells.(**a pk model with an additional liver compartment is used for recap. *clp for recap corresponds to the liver metabolic clearance.) protein species mw catabolic organ clp ml/m f1 f2 l1 ml/m l2 ml/m albumin human 66,500 endothelium 0.167 0.734 0.722 0.87 2.00 monoclonal antibody human mepolizumab 149,000 endothelium 0.132 0.367 0.361 0.87 2.00 infliximab 149,000 endothelium 0.23 0.367 0.361 0.87 2.00 alkaline phosphatase human recap** 140,000 liver 66.2* 0.367 0.361 0.87 2.00 placental 140,000 liver 0.371 0.367 0.361 0.87 2.00 amylase baboon 57,086 kidney 27.9 0.734 0.722 0.87 2.00 lipase baboon 48000 kidney 27.9 0.734 0.722 0.87 2.00 alanine transaminase dog 54,600 sinusoidal 0.949 0.530 0.535 1.2 2.71 aspartate aminotransferase dog mitochondrial 47,517 kupffer 56.9 0.530 0.535 1.2 2.71 cytoplasmic 46,247 sinusoidal 4.06 0.530 0.535 1.2 2,71 creatine kinase dog 81,000 kupffer 31.7 0.370 0.374 1.2 2.71 lactate dehydrogenase lamb ld1 heart 138,000 sinusoidal 0.975 0.53 0.53 3.85 1.35 ld5 skeletal muscle 138,000 kupffer 6.35 0.53 0.53 3.85 1.35 figure 1. a) beeken et al. steady-state kinetic model. b) interpretation of the model in terms of onedirectional convective filtration from plasma to the tissue and return via lymph flows to the plasma. beeken et al. [3] and takeda and reeve [4] have published similar compartmental pk human albumin analyses using the standard 3-compartment modeling approach. in the model of beeken et al. diagramed in figure 1a, the plasma albumin exchanges with two tissue compartments and the kinetics are characterized admet & dmpk 11(1) (2023) 57-80 protein pharmacokinetic model doi: https://doi.org/10.5599/admet.1570 61 by 6 steady state constants: m, ap, a1, a2, r1, and r2. in the steady state, m equals the rate of synthesis which equals rate of catabolism, ap, a1, and a2 are the amounts in each compartment and r1 and r2 are the exchange rates. the pharmacokinetics of proteins has the simplifying characteristic that their distribution is limited primarily to the extracellular space. thus, ap is the amount in the plasma volume vp and the plasma albumin concentration is cp = ap/vp. similarly, a1 and a2 are the amounts in the compartmental interstitial volumes [17]. these 6 parameters can be reduced to 5 by scaling them to the amount in plasma (ap) and expressing them in terms of the following five time constants (in units of minutes): p p p 1 2 m p1 p2 t1 t2 1 2 1 2 a a a a a t t t t t m r r r r (1) in the steady state case, when albumin is being synthesized at the same rate that it is removed by catabolism/excretion, tm would be the time required to catabolize/excrete the initial amount in the plasma (=ap). in the non-steady state in which there is no new synthesis of albumin, tm is the time for catabolism/excretion to lower the blood albumin concentration to 0.368 (=1/e) of its original concentration, assuming no tissue exchange. similarly, tp1 and tp2 are the time constants for the fall in plasma concentration produced by filtration to the tissue, and tt1 and tt2 are time constants for the washout of the tissue compartments. all pk results are scaled to the human with an assumed plasma volume of 2800 ml. figure 2. comparison of pk compartmental protein model (line) for bolus iv human i131 albumin versus experimental data (solid circles) of takeda and reeve. we have incorporated this compartmental model into a general computer routine using maple (maplesoft) with the 6 parameters determined by simple trial and error adjustments. as a test, figure 2 shows the good agreement of the theoretical results (solid line) versus the independent experimental data of takeda and reeve [4] for the time course of plasma i131-albumin following a bolus input to the plasma, using the rate constants listed in table 1, differing by only a few percent from those of beeken et al. [3]. the https://doi.org/10.5599/admet.1570 d.g. levitt and m.d. levitt admet & dmpk 11(1) (2023) 57-80 62 metabolic time constant (tm) is 16,750 minutes (11.6 days). this is the hypothetical time constant for the exponential fall in the plasma concentration where the plasma albumin does not exchange with the tissue. the actual plasma pk, including tissue exchange, is multi-exponential and, therefore, cannot by described by a single time constant. however, because tm is much longer than the other time constants, at long times when the plasma and tissue are approximately equilibrated the plasma pk has a time constant equal to (ap + a1 +a2)/m =tm(1+tt1/tp1+tt2/tp2) (figure 1a) or 26 days. the “half-time” required for the plasma concentration to fall to half its initial concentration is only about 2 days, reflecting primarily the time constants for distribution to the tissue (tp1, tp2). for pbpk modeling, it is necessary to specify the physiological parameters that determine the protein pk (flow, convective transport, lymph flow, tissue volumes, etc.) for each organ. the obvious difficulty with this approach is the huge number of model parameters, most of which cannot be directly determined. for example, in the rat model of shah and betts [6], there are 16 organs, each described by 5 or more parameters, for a total of 90 parameters. pbpk modeling has been primarily useful for some small molecules for which the important parameters are simply the organ blood flow and volume, along with some other parameters that can be predetermined [18,19]. in marked contrast, protein pk depends on organ parameters such as capillary convective transport, lymph flow and the convective sieving factor for the protein, factors that, at present, cannot be accurately predetermined. in an animal model such as the rat, one can use the individual organ tissue concentrations to put some constraints on these parameters [6], but most remain just educated guesses. for the human, tissue measurements are not possible, severely limiting the value of a complete pbpk model. instead, we have developed the “lumped physiological” model shown schematically in figure 1b. it is assumed that the “tissue 1” compartment (figure 1a) with the slow turnover time (tt1, table 1) corresponds to the lumped tissues with relatively low lymph flow (e.g., skeletal muscle, skin, etc.) and the high turnover compartment (tissue 2, tt2) represents the high lymph flow tissues (e.g., gastrointestinal tract). there are unidirectional convectional volume flows from plasma to each tissue space of l1 and l2 and equal lymph flows l1 and l2 returning to the plasma. in the convective flow from the plasma to tissue, the protein is sieved by factors of f1 and f2, so that the rate of transport from blood to tissue is equal to fi li cp. there is no sieving of protein in the returning lymph with the lymph concentration equal to that in the tissue space so the rate of return to the blood is lici. d is the dose into the blood compartment and clp is the plasma metabolic clearance (in units of ml/min). the plasma protein pharmacokinetics is a function of just the 5 compartmental time constants (eq. 1) and does not depend on how they are interpreted physiologically. however, to also model the lymph flow or the lymph protein concentration using the lumped model (figure 1b). it is necessary to assume two additional variables that assign physical volumes to the tissue compartments (v1, v2). using these volumes, the metabolic plasma clearance (clp), lymph flows (l1, l2) and sieving factors (f1, f2) in figure 1b can be expressed as functions of the five compartmental time constants: p 1 2 p 1 2 m 1 2 1 1 2 2 1 2 p p t t p p v v vv v cl l l f f t t t l t l t (2) alternatively, if the two lymph flows are known, then eq. 2 can be used to calculate the volumes v1 and v2, given tt1 and tt2. the time dependence of the concentration in the plasma (cp(t)) and the two tissues (c1(t), c2(t)) is described by the 3 coupled differential equations: admet & dmpk 11(1) (2023) 57-80 protein pharmacokinetic model doi: https://doi.org/10.5599/admet.1570 63 1 1 2 2 1 1 2 2 1 1 1 1 1 1 2 2 2 2 2 2 d ( ) ( ) ( ) ( ) ( ) ( ) d ( ) ( ) ( ) ( ) ( ) ( ) p p p p p p p c t v cl c t f l f l c t l c t l c t t dc t v l c t f l c t dt dc t v l c t f l c t dt (3) the advantage of the lumped physiological model (figure 1b) is that it places major physiological limitations on the model parameters, i.e., there is not a near-infinite number of manipulations that can be employed to make the model fit the data. in our fitting of the data, we have made the following assumptions: 1) for a given species (e.g., human, dog, etc.), the value of the lymph flows l1 and l2 and tissue volumes v1 and v2 must be the same for all proteins studied in that species. 2) the capillary filtration “pore” must be roughly the same for the two tissue regions, implying the sieving factor f1 ≈ f2. 3) the values of f must be physiologically reasonable, i.e., f must be <=1 and, for a given species, proteins with similar molecular weights should have similar f, and f must decrease as the protein molecular weight increases. these assumptions dramatically simplify the model. assuming that l1, l2, v1 and v2 have been previously determined for this species, in place of the 5 adjustable parameters in the general model, there are only two parameters that can be adjusted to fit different proteins for the same species: the plasma metabolic clearance (clp) and the sieving factor (≈f1 ≈ f2). in addition, for a given species, this sieving factor must decrease as the protein molecular weight increases. as shown in table 2, we have been able to successfully fit the pk data for the 13 investigated proteins using these constraints on the model parameters. the most questionable of these assumptions is the assumption that f1 ≈ f2. zhao et al. [16] have used a similar two-compartment convective model to describe the plasma pk of a variety of monoclonal antibodies in 8 different animal species. they assumed that compartment 1 is “leaky” with a sieving coefficient f near 1, and compartment 2 is “tight” with an f near 0. the primary reason for this choice is to account for the liver, whose sinusoids are highly permeable (“leaky”) for most proteins. we disagree with this approach. the liver sinusoids are so highly permeable that the exchange of albumin between the plasma and liver interstitial space is flow-limited [20], that is, the sinusoidal exchange is so fast it is as if the sinusoidal barrier is not present. thus, given the high liver blood flow, the liver interstitial space should equilibrate with the plasma within a few minutes. in the typical pk measurement, one waits for about 5 to 10 minutes after the bolus iv protein injection for plasma equilibration before taking the first plasma sample to follow the rate of decrease in concentration. because of the flow-limited liver protein exchange, the liver will have completely equilibrated before the first sample is taken, and it will not contribute to the subsequent fall in plasma concentration. thus, in our model, the liver is not included in the two tissue compartments in figure 1b and the liver interstitial space is regarded as a component of the plasma volume vp. in all the other non-liver tissues, albumin is highly impermeable, as indicated by the fact that it exerts its theoretical colloid osmotic pressure across the capillary membrane, which is essential for fluid balance in all these tissues. although the mechanism of the slow convective leak is poorly understood, whatever the mechanism, we think it is likely that, for the same species, the value of f should be roughly the same in all these non-liver tissues. it should be noted that plasma protein pk is quite simple, and any model with a few adjustable parameters will be able to fit the plasma concentration data. the ability of the model to also fit the experimental measurements of the lymph concentration places important additional constraints on the model and is a factor that we have emphasized in this review. the liver does complicate the interpretation of the lymph measurements. since the liver lymph flow is relatively large, about 20 % of the thoracic duct flow [21], liver lymphatics, which are neglected in our model, should make a significant contribution. however, as shown below, lymph https://doi.org/10.5599/admet.1570 d.g. levitt and m.d. levitt admet & dmpk 11(1) (2023) 57-80 64 concentration measurements are highly variable and deviations of 20 % are not considered significant. the lumped tissue volumes v1 and v2 and the corresponding lymph flows l1 and l2 (eq. 2) cannot be directly measured experimentally. they are simply model parameters that are empirically adjusted to fit, at least, the following three experimental observations. the total lymph flow relation: 1 2total lymph flow = l l (4) the steady-state tissue concentrations are described by (c1/cp)ss = f1 and (c2/cp)ss = f2, so that the experimentally measurable steady-state total lymph protein concentration relative to the plasma concentration (clss/cp) is: lss 1 1 2 2 ss p 1 2 c f l f l c l l (5) finally, the steady-state extravascular volume of distribution (vecf), which can be determined experimentally from the pk following a bolus injection [18], is described by: 1 2 ecf 1 1 2 2ss p a a v f v f v c (6) where a1 and a2 are the steady-state amounts in the tissue compartments (figure 1a). for each experimental animal model (e.g., human, dog, etc.), v1 and v2 were adjusted to satisfy these three experimental observations (eqs. 4-6). for the human, we estimated that v1/vp = 1.3 and v2/vp = 0.4, which produces albumin sieving factors (eq.2) f1 =0.73 and f2=0.725. these values satisfy the 3 conditions (eqs. 46) for albumin. the corresponding lymph flows (eq. 2) are l1 = 0.87 ml/min and l2=2 ml/min for a total lymph flow (eq. 4) of 2.8 ml/min. although there is no measurement of normal human total lymph flow (thoracic duct plus right lymph duct), this result is similar to the few available normal human thoracic duct measurements [22,23]. the corresponding theoretical steady state albumin lymph/plasma ratio (eq. 5) is 0.725, identical to that reported by witte et al. [24] for human thoracic duct lymph. finally, vecf (eq. 6) is 3479 ml, similar to the value of 3250 determined from the pk analysis following the bolus iv injection (figure 2). these physiological parameters for albumin and all the other proteins studied are summarized in table 2. tissue region 1, which has a larger volume (v1/vp = 1.3) and slower turnover time (t1= 4182 minutes), presumably corresponds to the lumped tissues with relatively low lymph flows such as the skin, subcutaneous and skeletal muscle tissues, while tissue region 2, with the smaller volume (v2/vp = 0.4) and faster turnover time (t2=557 minutes), presumably corresponds to the gi tract which is known to have a relatively high lymph flow rate [13]. the physiological model fit to the albumin data is identical to the compartment fit (figure 2) because we have defined the albumin physiological parameters (table 2) directly in terms of the previously determined compartmental parameters (table 1, eq. 2). the focus of this review is on the pk, not on the detailed metabolic mechanisms. the main assumption of both the compartmental and physiological model is that the catabolism or clearance occurs from the plasma space, i.e., it is proportional to the plasma concentration. both albumin and immunoglobulins are known to be metabolized by endosomal metabolism that is moderated by fcrn binding [5,7,25]. since this endosomal space is usually assumed to be located in the vascular endothelium [5,7] which is in contact with the blood, this is consistent with the metabolic assumption. similarly, the diagnostic enzymes are either metabolized by the liver or cleared by the kidney (table 1), both of which are also proportional to the plasma concentration. admet & dmpk 11(1) (2023) 57-80 protein pharmacokinetic model doi: https://doi.org/10.5599/admet.1570 65 monoclonal antibodies. igg and albumin are the classical markers of capillary protein permeability. with a molecular weight about twice that of albumin, igg has a lower filtration coefficient and a corresponding lower steady-state lymph concentration, about 60 % that of albumin [24,26,27]. the recent development of monoclonal antibodies for therapeutic purposes has resulted in an extensive investigation of the human pk of these proteins and has been the subject of several recent reviews [28-30]. figure 3 (left panel) shows the physiological model fit to the data of smith et al. [31] for the plasma disappearance curve of an iv bolus infusion of the monoclonal antibody mepolizumab. as discussed, since the lymph flows (l1, l2) and tissue volumes should be equal to those determined above for albumin in the human, there are only two parameters that can be adjusted to fit this data, the plasma clearance (clp) and the filtration constant f (≈ f1 ≈ f2), which should be less than then the albumin value of f=0.73 since mepolizumab has a molecular weight 2.2 times that of albumin. the physiological model fit in figure 3 is in accord with these restrictions, having an f1 = 0.367 and f2 = 0.361, half that for albumin, and the same lymph flows and tissue volumes used for albumin (summarized in table 2). as predicted by the physiological model, the pk of another monoclonal, infliximab [32], was accurately described by adjusting only clp, with all the other parameters identical to those for mepolizumab since it has a nearly identical structure and, therefore, should have identical filtration constants (data in table 2, figure 3, right panel). figure 3. comparison of pk physiological protein model prediction (line) following iv bolus monoclonal antibody injection in the human versus experimental data (solid circles). left panel: mepolizumab; right panel: infliximab. only one parameter (plasma clearance) differs for the two model fits. in addition to the iv pharmacokinetics for mepolizumab, smith et al. [31] also described the plasma concentration following a bolus subcutaneous injection. it is of interest to determine if this data is consistent with the predictions of the physiological model (figure 1b) using parameters identical to those determined from the iv input data (figure 3, table 2). as discussed above, the slower lymph flow, longer turnover time region v1 presumably corresponds roughly to that of the subcutaneous tissue. thus, the subcutaneous dose of mepolizumab was presumably injected into a tissue with kinetics similar to that of v1. figure 4 compares the experimental plasma concentration following the subcutaneous injection versus the physiological model prediction for a bolus dose into tissue v1 using the identical model parameters obtained from the iv input (figure 3), with no adjustable parameters. the agreement between model prediction and experimental data is remarkably good, given the simple two-tissue compartment model assumption. this result provides additional support for the physiological model (figure 1b) with its corresponding quantitative implications for the lymph flow rates (eq. 2). https://doi.org/10.5599/admet.1570 d.g. levitt and m.d. levitt admet & dmpk 11(1) (2023) 57-80 66 figure 4. comparison of physiological protein model prediction (line) of the plasma concentration following a subcutaneous bolus mepolizumab injection in the human versus experimental data (solid circles) using the identical model parameters used for iv input in figure 3. amylase and lipase lipase and amylase, two enzymes used to assess the pancreatic injury, are unique among the proteins listed in table 1 in that they are cleared primarily by the kidney. junge and colleagues [33,34] have elucidated their pharmacokinetics in the rat. both are rapidly cleared from the blood with a single compartment, oneexponential metabolic time constants tm of 31 min and 26 min for amylase and lipase, respectively. about 70 % of this clearance is renal, with both enzymes having surprisingly high glomerular filtration rates, roughly 12 % of inulin. about 84 % of the filtered amylase appears in the urine as an intact, enzymatically active protein, while nearly all the filtered lipase is metabolized by the kidney with negligible amounts of active enzyme appearing in the urine. figure 5. comparison of pk physiological protein model prediction (line) following iv bolus of baboon purified pancreatic amylase versus the experimental data in baboon (solid circles). the activity is plotted with the basal background subtracted. levitt and colleagues [35,36] have reported a detailed analysis of the pharmacokinetics of amylase in the baboon. they determined the plasma kinetics of enzyme activity following a bolus iv injection of either salivary or pancreatic amylase, both of which were rapidly cleared with nearly identical kinetics and could be admet & dmpk 11(1) (2023) 57-80 protein pharmacokinetic model doi: https://doi.org/10.5599/admet.1570 67 described by a single exponential with a tm of about 100 min. complicated pk modeling is superfluous for proteins with tm much shorter than the plasma-tissue exchange time constants (tp1, tp2, tt1, tt2) because these proteins, effectively, are limited to the plasma space and should have one compartment, single exponential, kinetics with tm equal to 1/slope of the straight line in the log plot. since tm is uniquely determined by the plasma volume and the clearance (eq. 2), the clearance is the only adjustable parameter. however, it is of interest to check whether a single exponential is consistent with our protein model. figure 5 shows the model fit assuming amylase f1 and f2 equal to those of albumin, which has a similar molecular weight and l1 and l2 for the baboon assumed identical to the human (table 2). only the last experimental data point deviated from this model result. since the experimental data is for the “excess” amylase above the initial basal value, just a small drift (≈ 2 units/ml) in basal amylase could account for this small deviation. or, if the data are correct, it would suggest that the tissue-plasma exchange of amylase is significantly slower than albumin. note, the values of the parameters f1, f2, l1, and l2 listed in table 2 for amylase and lipase only indicate that they are compatible with the plasma pk, but they were not directly determined because of the short metabolic time constant. although there are no equivalent large animal kinetic studies of lipase, we have assumed identical kinetics for lipase in table 1 because of the nearly identical lipase and amylase kinetics in the rat [33]. alanine transaminase (alt) alanine transaminase, also known as alanine aminotransferase or glutamic pyruvate transaminase, is released from a variety of tissues under pathological conditions. its primary site of elimination is probably the liver. at least three different liver systems are involved in protein catabolism: the sinusoidal cells, the kupffer cells and the asialoglycoprotein receptor (see below). kupffer cells are macrophages that line the surface of the sinusoids [37] and their metabolic activity for different enzymes can be quantitated by competition experiments [38]. because the kupffer cells usually have higher rates of metabolism than the sinusoidal cells, the proteins they catabolize usually have shorter time constants. alt is not metabolized by kupffer cells, consistent with its relatively long lifetime [38]. figure 6. comparison of pk physiological protein model (line) for alt bolus iv injection versus experimental data (solid circles) in dog. https://doi.org/10.5599/admet.1570 d.g. levitt and m.d. levitt admet & dmpk 11(1) (2023) 57-80 68 fleisher and wakim [39] determined the dog disappearance kinetics of iv injections of alt purified from either dog liver or heart (no significant differences). importantly, they also reported simultaneous measurements of thoracic duct lymph alt concentrations, which, as discussed above, can be used to estimate the tissue volumes. using these data, we estimated the dog values of tissue volumes of v1/vp = 1.0 and v2/vp = 0.3, differing slightly from the values determined above for the human. these values then allow us to relate the compartmental model time constants to the filtration coefficients and lymph flows using eq. 2. figure 6 shows the good fit of the protein model to the plasma concentration following the iv bolus injection, using the pk data listed in tables 1 and 2. the metabolic time constant (tm) is 2949 min (2 days), 5.5 fold faster than albumin. the corresponding f1=0.53, f2=0.535. the estimated lymph flows are l1 =1.2 and l2 =2.7, for a total lymph flow = 3.9 ml/min, similar to the 3.5 ml/min value reported in resting conscious dogs [40]. (note, all values reported in this review have been extrapolated to the 70 kg human). figure 7. comparison of pk physiological protein model lymph thoracic duct alt concentration following iv bolus injection in dog (line) versus experimental data (solid circles) of fleisher and wakim. the simultaneously measured thoracic duct lymph alt concentration had relatively large variations between dogs [39]. the model total lymph concentration is given in terms of the total lymph drainage of the two tissue compartments (figure 1b): 1 1 2 2 1 2 lymph concentration = l c l c l l (7) figure 7 shows the relatively good model fit obtained for the lymph data for one dog. these tissue volumes and lymph flows should be regarded as only rough approximations because of the large variation between dogs and the assumption that there are only two tissue compartments. however, the finding that the thoracic duct data is in general agreement with the model predictions with reasonable physiological parameters is encouraging and provides additional support for the physiological model (figure 1b), with the parameters summarized in table 2. aspartate aminotransferase (ast) and creatine kinase (ck) aspartate aminotransferase, also known as glutamic-oxalacetic transaminase, is a clinical marker for hepato-cellular injury secondary to a variety of inflammatory conditions. there are two forms of the enzyme found in plasma, the “cytoplasmic” and “mitochondrial”, with similar structures but markedly different pharmacokinetics [41,42]. fleisher and wakim [41] investigated the dog pk for both forms of purified dog admet & dmpk 11(1) (2023) 57-80 protein pharmacokinetic model doi: https://doi.org/10.5599/admet.1570 69 liver alp. since ast has a molecular weight similar to alt, it should have the same f1 and f2 and lymph flows l1 and l2 as those determined above in the dog for alt. thus, there is only one adjustable parameter: clp. figure 8 and table 2 show that, indeed, both cytoplasmic and mitochondrial plasma ast concentration following an iv bolus could be successfully fit by the protein model using the previously determined dog alt f1, f2, l1 and l2. the ability to fit this novel data using only one adjustable parameter dramatically illustrates the advantage of using this physiological model. mitochondrial ast is metabolized by the liver kupffer cells and, as is characteristic of this mechanism, has a very high metabolic rate, with a tm of only 50 minutes. cytoplasmic ast is metabolized primarily by the liver sinusoidal cells and has a tm of 670 minutes, 13-fold slower than mitochondrial ast. figure 8. comparison of pk physiological protein model prediction (line) following iv bolus dose of mitochondrial (left panel) or cytoplasmic (middle panel) ast and creatine kinase (right panel) versus experimental data (solid circles) in dog. creatine kinase (ck) is another kupffer-metabolized protein [38]. aktas et al. [43] have determined the plasma kinetics in the dog following iv injections of ck purified from dog muscle homogenates. ck should have the same dog l1 and l2 as alt, but since it has about twice the molecular weight of alt, it is expected to have smaller filtration constants and, therefore, there are two adjustable parameters: clp and f. consistent with this, the plasma pk data are well fit by the model (figure 8, right panel) with f1 = f2 =0.37, 30 % smaller than alt and ast, and the same dog l1 and l2 as alt and ast (table 2). as expected for liver kupffer metabolism, it has a relatively short metabolic time constant (tm) of 88 min (table 1). lactate dehydrogenase (ld) lactate dehydrogenase is a tetramer constructed from different numbers of the two monomers called h (heart) and m (muscle), resulting in 5 different isoenzymes: ld1 (h4), ld2 (h3m), ld3 (h2m2), ld4 (hm3) and ld5 (m4) [44]. the predominant diagnostic isoenzymes are ld1 which increases in acute myocardial infarction and ld5, which is increased in skeletal muscle (e.g., muscular dystrophy) and liver pathologies [44]. ld5 is metabolized predominantly by liver kupffer cells and, accordingly, should have a relatively short metabolic time constant. in contrast, ld1 is not metabolized by kupffer cells and should have a corresponding slower metabolic rate [38]. https://doi.org/10.5599/admet.1570 d.g. levitt and m.d. levitt admet & dmpk 11(1) (2023) 57-80 70 figure 9. comparison of pk physiological protein model prediction (line) following iv bolus injections in the lamb of the sheep purified protein lactate dehydrogenase isoenzymes ld1 (heart) (left) and ld5 (skeletal muscle) (right) versus experimental data (solid circles). boyd has reported the plasma kinetics in lambs following iv bolus injections of ld1 (purified from sheep heart) and ld5 (purified from sheep skeletal muscle). we assume the tissue space volumes are identical to those determined for the dog (v1/vp = 1.0 and v2/vp = 0.3). as shown in figure 9, both forms were adequately fit using this convection model with the corresponding f1 = f2 =0.53, and lymph flows l1=3.85 and l2 = 1.35 (table 2). these results are more uncertain than the dog or human because there are no lamb lymph protein data to constrain them. these lymph flows (total flow = 5.2 ml/min/70 kg), which are significantly larger than those of dog or human are consistent with the high values reported for lamb thoracic duct [45]. as expected for these two molecules with similar structures, they were fit with identical values of f1, f2, l1 and l2. the tm for ld5 (kupffer cell metabolism) is 465 minutes, more than 7 times shorter than the ld1 tm of 3418 minutes (table 1). alkaline phosphatase (alp) peters et al. [46] have developed a human recombinant alkaline phosphatase (recap) as a potential therapeutic agent. this human chimera is constructed from 414 amino acids from the intestinal plus 70 amino acids from the “crown” domain of placental alkaline phosphatase [47]. since the recap pk is nearly identical to purified bovine intestinal alkaline phosphatase [48], it can be assumed that the recap pk is representative of human intestinal alkaline phosphatase (iap). studies of recap provide some of the best available human protein enzyme pk studies because the protein is readily distinguished from the background endogenous enzyme. since recap has a molecular weight nearly identical to that of igg, it should have the same f1, f2, l1 and l2 that were previously determined for igg in humans, leaving only one adjustable parameter, clp. figure 10 shows an attempt to fit the recap kinetics (solid circles) following a 60 min constant iv infusion of 1000 u/kg in humans with the convective model (figure 1b) using the igg f1 and f2, and the human lymph flows (l1 and l2), adjusting only clp to fit the early time data. it can be seen that recap has marked biphasic kinetics that is very poorly fit by the model. no reasonable variation of the model parameters can fit the anomalously rapid, initial, distribution phase of the enzyme. clearly, a major modification of the protein model in figure 1b is required for recap. admet & dmpk 11(1) (2023) 57-80 protein pharmacokinetic model doi: https://doi.org/10.5599/admet.1570 71 figure 10. comparison of pk physiological protein model prediction (line) following a 60 min iv infusion of recombinant alkaline phosphatase (recap) in the human versus experimental data (solid circles). the human plasma-tissue exchange parameters (f1, f2, l1, l2, figure 1b) were assumed to be identical to those for monoclonals and only the metabolic clearance (clp) was adjusted to try to fit the early time data. studies in the rat, both in vivo well as in isolated liver perfusion studies, have provided a mechanistic explanation of this striking biphasic kinetic behavior [49-52]. alkaline phosphatase is unique among the proteins under study (table 1) in that it is a glycoprotein whose elimination involves binding to the mannose and galactose (also known as asialoglycoprotein) liver receptor. the rates of both the early (distribution) and late (elimination) phases of the alkaline phosphatase curve are dramatically reduced by the addition of asialofetuin, an inhibitor of the asialoglycoprotein receptor. the rat results suggest that the early, extremely rapid phase of the kinetics results from alkaline phosphatase attachment to the asialoglycoprotein receptor, most of which is then taken up by the liver cell and recycled back to the plasma where it is released intact, with little metabolism. figure 11. modified physiological protein model with the addition of a recycling liver compartment with metabolic clearance cll from the liver compartment. https://doi.org/10.5599/admet.1570 d.g. levitt and m.d. levitt admet & dmpk 11(1) (2023) 57-80 72 figure 11 diagrams a modification of the protein physiological model with the addition of another compartment (vl, cl) in an attempt to account for this additional recycling liver component. the rate constant k characterizes the recycling rate between the plasma and liver compartment, and vl is the equivalent volume (i.e., binding capacity) of this compartment. note that the metabolism now occurs from the liver compartment and is proportional to cl and is characterized by metabolic liver clearance (cll). although there are 7 pk parameters in figure 11, only 3 of them are adjustable (k, vl, cll) because f1, f2, l1, and l2 should have the same values as the human monoclonal antibody pk. as shown in figure 12, this modified model, with these limiting conditions, accurately simulates the human biphasic kinetics of recap (using model parameters listed in table 2). the liver compartment vl has a very large volume (e.g., binding capacity), 26 times that of the plasma volume, and rapidly exchanges (k) with the plasma (50 times the rate of plasma to tissue transport). the fit is not perfect, which is not surprising given the complexity of liver recycling and metabolic processes. the red line shows that when the liver exchange rate k is set to zero, there is a dramatic reduction in both the clearance and tissue distribution, very similar to what is observed in the rat when the asialoglycoprotein receptor is competitively inhibited by asialofetuin [49,51]. the green dashed line (figure 12) shows that setting the liver metabolic clearance (cll) to zero has only a minor effect on the early phase kinetics, which is dominated by the distribution into the liver compartment. figure 12. comparison of physiological liver recycling model (figure 11) prediction (black line) following a 60 min iv infusion of recombinant alkaline phosphatase (recap) in the human versus experimental data (solid circles). the red line is for the case where the liver exchange rate (k, figure 11) is set to zero. the green dashed line is for the case where the liver metabolic clearance (cll) is set to zero. an interesting test of this pk model is to see if it can mimic the well-established, transient clinical elevations in plasma intestinal alkaline phosphatase (iap) that follow the ingestion of a high-fat (but not protein or carbohydrate) meal [53]. there is a large individual variation in these fluctuations, with those who are red blood cell b or o antigen secretors having much greater iap release. domar et al. [53] have quantitated these fluctuations following a standard high-fat meal. interestingly, they have shown that the baseline (fasting) iap increases proportionally to the peak increase and that secretors with high fatstimulated values have fasting iap about 10 times greater than non-secretors with small fluctuations, strongly suggesting that the intestinal iap release following a fatty meal is the sole supplier and determinant of plasma iap. admet & dmpk 11(1) (2023) 57-80 protein pharmacokinetic model doi: https://doi.org/10.5599/admet.1570 73 we tried to fit the data of domar et al. [53] with the modified pk model (figure 11) using the model parameters determined above for recap, which is representative of iap. we assumed that following a highfat meal, iap was released into the tissue volume v2 (figure 1b), which, as discussed above, presumably corresponds to the gi tract, at a constant rate for 120 minutes, simulating the release of iap into the intestinal interstitial space following a high-fat meal. in order to compare the results with that of domar et al. for subjects that have established steady-state iap levels, this 120-minute input was repeated at 24-hour intervals until a steady state was established. figure 13 shows the comparison of our pk model predictions to the clinical data of domar et al. [53] for one of their subjects (there was a large subject-to-subject variation). given the crudeness of our model, the agreement with the data is relatively good. of special significance is the fact that both the size of the iap fluctuations and the corresponding basal fasting value (minimums in the plot) are in good agreement with the clinically observed values. this strongly suggests that the fasting value is entirely the result of the amount of iap released during fatty meals. for the subject in figure 13, a total of 360 iu was released with each meal, quantitating the daily human iap release. figure 13. comparison of physiological liver recycling model (figure 11) simulation of the intestinal release of intestinal alkaline phosphatase following a high-fat meal every 24 hours (line) versus the experimental data of domar et al. (solid circles). there are at least 4 gene loci coding 4 different tissue-specific alkaline phosphatase isoenzymes: 1) intestine, 2) liver/bone/kidney, 3) placenta, and 4) germ cell [49,54]. clinical observations indicate that the intestinal form has the shortest half-live (most rapid clearance) placental the longest, with the liver/bone/kidney isoenzyme intermediate, a finding consistent with pk measurements in the rat [49]. based on the pk model developed for recap (figure 11), one would predict that both the early phase (recycling) and late phase (metabolism) of the biphasic kinetics should be proportional to the affinity for the asialoglycoprotein receptor, which depends on the number of glycan chains and their degree of sialyation. this is consistent with the low clearance of the placental alkaline phosphatase, which has only one glycan chain, of which more than 50 % are sialyated [55]. https://doi.org/10.5599/admet.1570 d.g. levitt and m.d. levitt admet & dmpk 11(1) (2023) 57-80 74 figure 14. comparison of standard physiological protein model (figure 1b) prediction (line) following iv bolus injection of human purified placental alkaline phosphatase versus experimental data (solid circles). in an older study, clubb et al. [56] determined the pk of partially purified human placental alkaline phosphatase (pap) following a bolus iv dose in humans, using the heat stability of pap to distinguish it from the endogenous background plasma alkaline phosphatases. because of the very low glycosylation of pap, the liver recycling should be negligible, and we attempted to fit the pap pk with the standard protein model (figure 1b) using the monoclonal antibody f1 and f2, and the human lymph flows l1 and l2, with clp as the only adjustable parameter. as can be seen in figure 14, this fit is excellent, indicating that pap is in the same class as all the other proteins in table 2 and that iap is the exception. clinical relevance the primary aim of this paper was to demonstrate that the distribution and elimination rate of an infused bolus of a wide variety of plasma proteins can be described by a relatively simple physiological pk model. since all studies were carried out in healthy humans and animals, the data provided are normal values. from the medical care standpoint, the important question is to what extent can alterations in these normal distribution and elimination kinetics produce clinically relevant but frequently unrecognized changes in plasma protein concentrations. the answer to this question, which is complicated and different for each protein, is the topic of a subsequent paper. however, some generalities include the following. depending on the protein, either increases or decreases in elimination rate could cause alterations in the plasma value that could be of clinical importance. a clear-cut example of such an elimination defect is the formation of globulin-bound enzymes (“macro enzymes”), an anomaly reported for most commonly measured diagnostic plasma enzymes. the increased molecular size of these macroenzymes dramatically slows their catabolic rate in the liver or glomerular filtration rate (lipase, amylase) such that serum concentration rises roughly in proportion to the fraction of enzyme that exists in the bound form. a disease of the eliminating organ also may cause serum enzyme elevations. the most obvious example is kidney disease; up to three-fold elevations of the two enzymes cleared by the kidney, lipase and amylase, can be observed in renal insufficiency [57,58]. similarly, hepatic injury, e.g., carbon tetrachloride-induced hepatic necrosis in rats, dramatically reduced the clearance of infused ast [42]. thus, the elevated plasma level of admet & dmpk 11(1) (2023) 57-80 protein pharmacokinetic model doi: https://doi.org/10.5599/admet.1570 75 the aminotransferases characteristic of hepatic injury reflects, in part, decreased elimination in addition to the increased release. another example of the effect of decreased liver clearance is the striking increases in alkaline phosphatase produced by circulating inhibitors of the galactose (asialoglycoprotein) receptor, as postulated in a recent paper [59]. as discussed above, the degree of sialyation of glycoproteins has a dramatic effect on their elimination rates. since the desialylation of circulating proteins is enhanced in various disease states, e.g., hyperthyroidism [60], this activity would be expected to have a major effect on the serum glycoprotein protein concentration. hypoalbuminemia is a common finding in a wide variety of disease states, and this finding is associated with a poor prognosis [25]. while clinicians commonly attribute these low serum albumin values to decreased synthesis (“malnutrition”, etc.), multiple studies have shown that, with the exception of cirrhosis and kwashiorkor, it is actually increased elimination that is the major cause of hypoalbuminemia [25]. this enhanced elimination may be due to increased activity of the normally catabolizing organs or an unusual rate of loss of albumin into the gut lumen or urine [61]. the development of igg monoclonal antibody therapies has stimulated interest in the elimination rate of these compounds [28-30]. the frequency of igg dosing required to maintain a therapeutic plasma concentration as a function of elimination rate and trough concentration measurements have become routine clinical assays for some monoclonal antibodies. alteration of the molecular structure of selected monoclonal antibodies has been used to reduce the elimination rate and hence prolong effectiveness [62]. much of the emphasis in this pk modeling discussion was focused on the kinetics of exchange and distribution with the extravascular space. however, at homeostasis, the rate of delivery to and return from the extravascular pools are equal, hence the plasma concentration is independent of the rate of distribution and depends only on the rates of synthesis and elimination. the rates of exchange with the extravascular pools as well as the size of the pool, do influence non-steady state plasma concentrations, such as the timing and magnitude of peak plasma concentrations following periodically administered monoclonal antibodies or the rate of rise and fall of liver enzymes after a brief insult. the appreciable extravascular distribution of proteins (e.g., 1.24 times the plasma volume for albumin at homeostasis) buffers the rate of change in plasma protein in the non-steady state, hence, influencing clinically important values such as the apparent half-time of various proteins. summary the three-compartment model shown in figure 1a is the standard 3-exponential, 6-parameter description routinely used to describe plasma pk. we have interpreted this model in terms of a lumped physiological model (figure 1b) in which the exchange between the blood and tissue is purely convective, dependent on the lymph flows l1 and l2 and filtration coefficients f1 and f2. for a given species, we first estimate the four adjustable model parameters l1, l2 and tissue volumes v1 and v2. assuming that these parameters are the same for any protein in that species implies that each protein is uniquely characterized by only 3 parameters: the clearance clp, and f1 and f2. imposing the additional constraint (satisfied by the results in table 2) that f1 ≈ f2, effectively reduces it to two adjustable parameters. interpreting these parameters in terms of the physiological convection model (figure 1b) has two important advantages. first, as discussed above, it puts strong bounds on the values of the parameters. for example, the human pk of an arbitrary protein should have the same values of v1, v2, l1 and l2 determined in table 2 and since f1 and f2 can be estimated from the molecular weight, there is actually only one primary adjustable parameter, the plasma clearance (clp), in place of the 5 parameters of the compartmental model (figure 1a, eq. 1). secondly, it allows one to use the plasma pk to predict the tissue and lymph protein https://doi.org/10.5599/admet.1570 d.g. levitt and m.d. levitt admet & dmpk 11(1) (2023) 57-80 76 kinetics: 1) the total lymph flow (eq. 4), 2) the steady-state lymph/plasma protein ratio (eq. 5), 3) the simultaneous plasma and lymph protein concentration following a bolus alt iv injection (figure 7), 4) the plasma concentration following a subcutaneous monoclonal antibody injection (fig 4), and 5) the plasma concentration of iap following a high-fat meal (figure 13). the validity of these predictions provides the strongest available evidence in support of this simple physiological model. as discussed, one of the assumptions of the model is that, for a given species, the sieving coefficient f should decrease as its molecular size increases. the results in the human are consistent with this assumption with an f = 0.73 for albumin (mol wt= 66,500) and a smaller f = 0.367 for the larger monoclonal antibodies (mol wt = 149,000) and similar size alkaline phosphatase (mol wt = 140,000). the dog results are also consistent: f= 0.53 for alanine transaminase (mw = 54,600) and similar-sized mitochondrial and cytoplasmic aspartate aminotransferase (mw 47,000) and f = 0.37 for the larger creatine kinase (mw = 81,000). although one could use these results to try to estimate an “equivalent capillary pore radius” [63], the data are too limited to put much confidence in these results. note that these results also suggest that the equivalent pore radius in the human is larger than that in the dog. this review is the first that we are aware of that uses a simple physiological model to describe a comprehensive framework for cataloguing the results of protein pk studies (table 2). the parameters are consistent with the constraints placed on them by the model, i.e., the same v1, v2, l1 and l2 for all proteins in the same species, and f decreases as molecular weight increases. even considering species variations, the plasma clearance (clp) is the main pk determinant, varying by more than 400-fold, with the other parameters varying by about 2-fold. the one protein that did not fit this simple convection model was the intestinal isoform (recap) of the glycoprotein alp. the recap pk is dramatically biphasic. in the first (distribution) phase, the plasma concentration falls to about 5 % of its initial value in 4 hours (see figure 10). since an impossibly large space of about 24 liters would be required to account for such a distribution, it appears that its initial disappearance phase is attributable almost entirely to receptor binding with rapid recycling in the liver. metabolism is negligible during this distribution phase (figure 12). the second slow metabolic phase only becomes apparent at long time periods, and the 24-hour data used in figure 12 may not be long enough for an accurate estimate of the metabolism. this result illustrates one of the main advantages of using a physiological model. although one can always fit the data with a compartment model with its 5 adjustable parameters, using the much more constrained physiological model (2 adjustable parameters) can provide unequivocal evidence that the model is wrong and that the compound under study must belong to a significantly different pk class. in going to the simplified 2 tissue compartment lumped physiological model, one gives up the potential of predicting the detailed individual tissue pk in the 13 or more organs of the full pbpk model. however, given the limited knowledge, at least in the human, of each organ’s lymph flow, volume, and sieving coefficient, this is more of a theoretical than a practical limitation. the two lumped tissue compartments (v1 and v2) in our simplified model are surprisingly good at simulating tissue input. for example, the plasma concentration following a subcutaneous injection is well simulated by injection into the slow turnover v1 compartment (figure 4) and the plasma concentration following gastrointestinal tissue input can be predicted by input into the fast turnover v2 compartment (figure 13). we believe that the physiological model’s simplicity and ability to describe the plasma and lymph concentration for a wide range of proteins compensates for its theoretical limitations. admet & dmpk 11(1) (2023) 57-80 protein pharmacokinetic model doi: https://doi.org/10.5599/admet.1570 77 conflict of interest: the authors report no conflicts of interest in this work. references [1] j.f. brandse, d. mould, o. smeekes, y. ashruf, s. kuin, a. strik, g.r. van den brink, g.r. d'haens. a reallife population pharmacokinetic study reveals factors associated with clearance and immunogenicity of infliximab in inflammatory bowel disease. inflamm bowel dis 23 (2017) 650-660. https://doi.org/10.1097/mib.0000000000001043. [2] a.a. fasanmade, o.j. adedokun, j. ford, d. hernandez, j. johanns, c. hu, h.m. davis, h. zhou. population pharmacokinetic analysis of infliximab in patients with ulcerative colitis. eur j clin pharmacol 65 (2009) 1211-1228. https://doi.org/10.1007/s00228-009-0718-4. [3] w.l. beeken, w. volwiler, p.d. goldsworthy, l.e. garby, w.e. reynolds, r. stogsdill, r.s. stemler. studies of i-131-albumin catabolism and distribution in normal young male adults. j clin invest 41 (1962) 1312-1333. https://doi.org/10.1172/jci104594. [4] y. takeda, e.b. reeve. studies of the metabolism and distribution of albumin with autologous i131albumin in healthy men. j lab clin med 61 (1963) 183-202. https://www.ncbi.nlm.nih.gov/pubmed/13984638. [5] r.h. rose, a. sepp, f. stader, k.l. gill, c. liu, i. gardner. application of physiologically-based pharmacokinetic models for therapeutic proteins and other novel modalities. xenobiotica (2022) 1-27. https://doi.org/10.1080/00498254.2022.2133649. [6] d.k. shah, a.m. betts. towards a platform pbpk model to characterize the plasma and tissue disposition of monoclonal antibodies in preclinical species and human. j pharmacokinet pharmacodyn 39 (2012) 67-86. https://doi.org/10.1007/s10928-011-9232-2. [7] h. wong, t.w. chow. physiologically based pharmacokinetic modeling of therapeutic proteins. j pharm sci 106 (2017) 2270-2275. https://doi.org/10.1016/j.xphs.2017.03.038. [8] p.m. glassman, j.p. balthasar. physiologically-based modeling of monoclonal antibody pharmacokinetics in drug discovery and development. drug metab pharmacokinet 34 (2019) 3-13. https://doi.org/10.1016/j.dmpk.2018.11.002. [9] j. lindena, f. diederichs, h. wittenberg, i. trautschold. kinetic of adjustment of enzyme catalytic concentrations in the extracellular space of the man, the dog and the rat. approach to a quantitative diagnostic enzymology, v. communication. j clin chem clin biochem 24 (1986) 61-71. https://doi.org/10.1515/cclm.1986.24.1.61. [10] b. rippe, b.i. rosengren, o. carlsson, d. venturoli. transendothelial transport: the vesicle controversy. j vasc res 39 (2002) 375-390. https://doi.org/10.1159/000064521. [11] b.i. rosengren, a. rippe, c. rippe, d. venturoli, k. sward, b. rippe. transvascular protein transport in mice lacking endothelial caveolae. am j physiol heart circ physiol 291 (2006) h1371-1377. https://doi.org/10.1152/ajpheart.01364.2005. [12] b. rippe, b. haraldsson. fluid and protein fluxes across small and large pores in the microvasculature. application of two-pore equations. acta physiol scand 131 (1987) 411-428. https://doi.org/10.1111/j.1748-1716.1987.tb08257.x. [13] a. bill. plasma protein dynamics: albumin and igg capillary permeability, extravascular movement and regional blood flow in unanesthetized rabbits. acta physiol scand 101 (1977) 28-42. https://doi.org/10.1111/j.1748-1716.1977.tb05980.x. [14] m. perry, d. garlick. transcapillary efflux of gamma globulin in rabbit skeletal muscle. microvasc res 9 (1975) 119-126. https://doi.org/10.1016/0026-2862(75)90056-4. [15] n.a. lassen, h.h. parving, n. rossing. editorial: filtration as the main mechanism of overall transcapillary protein escape from the plasma. microvascular research 7 (1974) i-iv. [16] j. zhao, y. cao, w.j. jusko. across-species scaling of monoclonal antibody pharmacokinetics using a minimal pbpk model. pharm res 32 (2015) 3269-3281. https://doi.org/10.1007/s11095-015-1703-5. https://doi.org/10.5599/admet.1570 https://doi.org/10.1097/mib.0000000000001043 https://doi.org/10.1097/mib.0000000000001043 https://doi.org/10.1007/s00228-009-0718-4 https://doi.org/10.1172/jci104594 https://www.ncbi.nlm.nih.gov/pubmed/13984638 https://www.ncbi.nlm.nih.gov/pubmed/13984638 https://doi.org/10.1080/00498254.2022.2133649 https://doi.org/10.1007/s10928-011-9232-2 https://doi.org/10.1016/j.xphs.2017.03.038 https://doi.org/10.1016/j.dmpk.2018.11.002 https://doi.org/10.1515/cclm.1986.24.1.61 https://doi.org/10.1515/cclm.1986.24.1.61 https://doi.org/10.1159/000064521 https://doi.org/10.1152/ajpheart.01364.2005 https://doi.org/10.1111/j.1748-1716.1987.tb08257.x https://doi.org/10.1111/j.1748-1716.1987.tb08257.x https://doi.org/10.1111/j.1748-1716.1977.tb05980.x https://doi.org/10.1111/j.1748-1716.1977.tb05980.x https://doi.org/10.1016/0026-2862(75)90056-4 https://doi.org/10.1007/s11095-015-1703-5 d.g. levitt and m.d. levitt admet & dmpk 11(1) (2023) 57-80 78 [17] d.g. levitt. the pharmacokinetics of the interstitial space in humans. bmc clin pharmacol 3 (2003) 3. https://doi.org/10.1186/1472-6904-3-3. [18] d.g. levitt, computer assisted human pharmacokinetics: non-compartmental, deconvolution, physiologically based, intestinal absorption, non-linear, www.pkquest.com, 2017. [19] d.g. levitt. pkquest: pbpk modeling of highly lipid soluble and extracellular solutes. admet dmpk 7 (2019) 60-75. https://doi.org/10.5599/admet.579. [20] c.a. goresky. the processes of cellular uptake and exchange in the liver. fed proc 41 (1982) 3033-3039. https://www.ncbi.nlm.nih.gov/pubmed/6754456. [21] j.c. cain, j.h. grindlay, et al. lymph from liver and thoracic duct; an experimental study. surg gynecol obstet 85 (1947) 558-562. https://www.ncbi.nlm.nih.gov/pubmed/20266768. [22] r.e. girardet, d.l. benninghoff. thoracic duct lymph and lymphocyte studies in man using a thoracic duct "side-fistula". cancer 29 (1972) 666-669. https://doi.org/10.1002/1097-0142(197203)29:3 [23] h.s. mayerson. on lymph and lymphatics. circulation 28 (1963) 839-842. https://doi.org/10.1161/01.cir.28.5.839. [24] c.l. witte, m.h. witte, a.e. dumont, w.r. cole, j.r. smith. protein content in lymph and edema fluid in congestive heart failure. circulation 40 (1969) 623-630. https://doi.org/10.1161/01.cir.40.5.623. [25] d.g. levitt, m.d. levitt. human serum albumin homeostasis: a new look at the roles of synthesis, catabolism, renal and gastrointestinal excretion, and the clinical value of serum albumin measurements. int j gen med 9 (2016) 229-255. https://doi.org/10.2147/ijgm.s102819. [26] m.d. el-zawahry, o.a. el-roubi, a.s. ibrahim. protein content of thoracic duct lymph in patients with bilharzial hepatic fibrosis. gastroenterology 72 (1977) 617-620. https://www.ncbi.nlm.nih.gov/pubmed/838214. [27] b.g. shafiroff, q.y. kau. cannulation of the human thoracic lymph duct. surgery 45 (1959) 814-819. https://www.ncbi.nlm.nih.gov/pubmed/13659321. [28] r.j. keizer, a.d. huitema, j.h. schellens, j.h. beijnen. clinical pharmacokinetics of therapeutic monoclonal antibodies. clin pharmacokinet 49 (2010) 493-507. https://doi.org/10.2165/11531280000000000-00000. [29] j.h. lin. pharmacokinetics of biotech drugs: peptides, proteins and monoclonal antibodies. curr drug metab 10 (2009) 661-691. https://doi.org/10.2174/138920009789895499. [30] e.d. lobo, r.j. hansen, j.p. balthasar. antibody pharmacokinetics and pharmacodynamics. journal of pharmaceutical sciences 93 (2004) 2645-2668. https://doi.org/10.1002/jps.20178 [31] d.a. smith, e.a. minthorn, m. beerahee. pharmacokinetics and pharmacodynamics of mepolizumab, an anti-interleukin-5 monoclonal antibody. clin pharmacokinet 50 (2011) 215-227. https://doi.org/10.2165/11584340-000000000-00000. [32] u. klotz, a. teml, m. schwab. clinical pharmacokinetics and use of infliximab. clin pharmacokinet 46 (2007) 645-660. https://doi.org/10.2165/00003088-200746080-00002. [33] w. junge, m. malyusz, h.j. ehrens. the role of the kidney in the elimination of pancreatic lipase and amylase from blood. j clin chem clin biochem 23 (1985) 387-392. https://doi.org/10.1515/cclm.1985.23.7.387. [34] m. malyusz, p. wrigge, d. caliebe, w. junge. renal handling of 125i-labelled homologous pancreatic lipase and amylase in the rat. j clin chem clin biochem 26 (1988) 611-615. https://doi.org/10.1515/cclm.1988.26.10.611. [35] w.c. duane, r. frerichs, m.d. levitt. distribution, turnover, and mechanism of renal excretion of amylase in the baboon. j clin invest 50 (1971) 156-165. https://doi.org/10.1172/jci106469. [36] w.c. duane, r. frerichs, m.d. levitt. simultaneous study of the metabolic turnover and renal excretion of salivary amylase125 i and pancreatic amylase131 i in the baboon. j clin invest 51 (1972) 15041513. https://doi.org/10.1172/jci106947. [37] l.j. dixon, m. barnes, h. tang, m.t. pritchard, l.e. nagy. kupffer cells in the liver. compr physiol 3 (2013) 785-797. https://doi.org/10.1002/cphy.c120026. https://doi.org/10.1186/1472-6904-3-3 file:///d:/zoranm/downloads/www.pkquest.com https://doi.org/10.5599/admet.579 https://www.ncbi.nlm.nih.gov/pubmed/6754456 https://www.ncbi.nlm.nih.gov/pubmed/20266768 https://doi.org/10.1002/1097-0142(197203)29:3 https://doi.org/10.1161/01.cir.28.5.839 https://doi.org/10.1161/01.cir.28.5.839 https://doi.org/10.1161/01.cir.40.5.623 https://doi.org/10.2147/ijgm.s102819 https://www.ncbi.nlm.nih.gov/pubmed/838214 https://www.ncbi.nlm.nih.gov/pubmed/838214 https://www.ncbi.nlm.nih.gov/pubmed/13659321 https://doi.org/10.2165/11531280-000000000-00000 https://doi.org/10.2165/11531280-000000000-00000 https://doi.org/10.2174/138920009789895499 https://doi.org/10.1002/jps.20178 https://doi.org/10.2165/11584340-000000000-00000 https://doi.org/10.2165/11584340-000000000-00000 https://doi.org/10.2165/00003088-200746080-00002 https://doi.org/10.1515/cclm.1985.23.7.387 https://doi.org/10.1515/cclm.1985.23.7.387 https://doi.org/10.1515/cclm.1988.26.10.611 https://doi.org/10.1515/cclm.1988.26.10.611 https://doi.org/10.1172/jci106469 https://doi.org/10.1172/jci106947 https://doi.org/10.1002/cphy.c120026 admet & dmpk 11(1) (2023) 57-80 protein pharmacokinetic model doi: https://doi.org/10.5599/admet.1570 79 [38] m.j. smit, a.m. duursma, j.m. bouma, m. gruber. receptor-mediated endocytosis of lactate dehydrogenase m4 by liver macrophages: a mechanism for elimination of enzymes from plasma. evidence for competition by creatine kinase mm, adenylate kinase, malate, and alcohol dehydrogenase. j biol chem 262 (1987) 13020-13026. https://www.ncbi.nlm.nih.gov/pubmed/2820961. [39] g.a. fleisher, k.g. wakim. the fate of enzymes in body fluids-an experimental study. i. disappearance rates of glutamic-pyruvic transaminase under various conditions. j lab clin med 61 (1963) 76-85. https://www.ncbi.nlm.nih.gov/pubmed/13945478. [40] h. schad, h. brechtelsbauer. thoracic duct lymph in conscious dogs at rest and during changes of physical activity. pflugers arch 367 (1977) 235-240. https://doi.org/10.1007/bf00581360. [41] g.a. fleisher, k.g. wakim. presence of two glutamic-oxalacetic transaminases in serum of dogs following acute injury of the liver. proc soc exp biol med 106 (1961) 283-286. https://doi.org/10.3181/00379727-106-26310. [42] y. kamimoto, s. horiuchi, s. tanase, y. morino. plasma clearance of intravenously injected aspartate aminotransferase isozymes: evidence for preferential uptake by sinusoidal liver cells. hepatology 5 (1985) 367-375. https://doi.org/10.1002/hep.1840050305. [43] m. aktas, h.p. lefebvre, p.l. toutain, j.p. braun. disposition of creatine kinase activity in dog plasma following intravenous and intramuscular injection of skeletal muscle homogenates. j vet pharmacol ther 18 (1995) 1-6. https://doi.org/10.1111/j.1365-2885.1995.tb00542.x. [44] m. maekawa. lactate dehydrogenase isoenzymes. j chromatogr 429 (1988) 373-398. https://doi.org/10.1016/s0378-4347(00)83879-7. [45] s.a. johnson, m.c. vander straten, j.a. parellada, w. schnakenberg, a.l. gest. thoracic duct function in fetal, newborn, and adult sheep. lymphology 29 (1996) 50-56. https://www.ncbi.nlm.nih.gov/pubmed/8823726. [46] e. peters, j. heuberger, r. tiessen, a. van elsas, r. masereeuw, j. arend, j. stevens, p. pickkers. pharmacokinetic modeling and dose selection in a randomized, double-blind, placebo-controlled trial of a human recombinant alkaline phosphatase in healthy volunteers. clin pharmacokinet 55 (2016) 1227-1237. https://doi.org/10.1007/s40262-016-0399-y. [47] t. kiffer-moreira, c.r. sheen, k.c. gasque, m. bolean, p. ciancaglini, a. van elsas, m.f. hoylaerts, j.l. millan. catalytic signature of a heat-stable, chimeric human alkaline phosphatase with therapeutic potential. plos one 9 (2014) e89374. https://doi.org/10.1371/journal.pone.0089374. [48] p. pickkers, f. snellen, p. rogiers, j. bakker, p. jorens, j. meulenbelt, h. spapen, j.e. tulleken, r. lins, s. ramael, m. bulitta, j.g. van der hoeven. clinical pharmacology of exogenously administered alkaline phosphatase. eur j clin pharmacol 65 (2009) 393-402. https://doi.org/10.1007/s00228-008-0591-6. [49] e. blom, m.m. ali, b. mortensen, n.e. huseby. elimination of alkaline phosphatases from circulation by the galactose receptor. different isoforms are cleared at various rates. clin chim acta 270 (1998) 125137. https://doi.org/10.1016/s0009-8981(97)00217-9. [50] f.g. russel, j.g. weitering, r. oosting, g.m. groothuis, m.j. hardonk, d.k. meijer. influence of taurocholate on hepatic clearance and biliary excretion of asialo intestinal alkaline phosphatase in the rat in vivo and in isolated perfused rat liver. gastroenterology 85 (1983) 225-234. https://www.ncbi.nlm.nih.gov/pubmed/6862151. [51] h.b. scholtens, m.j. hardonk, d.k. meijer. a kinetic study of hepatic uptake of canine intestinal alkaline phosphatase in the rat. liver 2 (1982) 1-13. https://doi.org/10.1111/j.1600-0676.1982.tb00172.x. [52] g.p. young, i.s. rose, s. cropper, s. seetharam, d.h. alpers. hepatic clearance of rat plasma intestinal alkaline phosphatase. am j physiol 247 (1984) g419-426. https://doi.org/10.1152/ajpgi.1984.247.4.g419. [53] u. domar, f. karpe, a. hamsten, t. stigbrand, t. olivecrona. human intestinal alkaline phosphatase-release to the blood is linked to lipid absorption, but removal from the blood is not linked to lipoprotein clearance. eur j clin invest 23 (1993) 753-760. https://doi.org/10.1111/j.1365-2362.1993.tb01296.x. https://doi.org/10.5599/admet.1570 https://www.ncbi.nlm.nih.gov/pubmed/2820961 https://www.ncbi.nlm.nih.gov/pubmed/2820961 https://www.ncbi.nlm.nih.gov/pubmed/13945478 https://doi.org/10.1007/bf00581360 https://doi.org/10.3181/00379727-106-26310 https://doi.org/10.3181/00379727-106-26310 https://doi.org/10.1002/hep.1840050305 https://doi.org/10.1111/j.1365-2885.1995.tb00542.x https://doi.org/10.1016/s0378-4347(00)83879-7 https://doi.org/10.1016/s0378-4347(00)83879-7 https://www.ncbi.nlm.nih.gov/pubmed/8823726 https://www.ncbi.nlm.nih.gov/pubmed/8823726 https://doi.org/10.1007/s40262-016-0399-y https://doi.org/10.1371/journal.pone.0089374 https://doi.org/10.1007/s00228-008-0591-6 https://doi.org/10.1016/s0009-8981(97)00217-9 https://www.ncbi.nlm.nih.gov/pubmed/6862151 https://www.ncbi.nlm.nih.gov/pubmed/6862151 https://doi.org/10.1111/j.1600-0676.1982.tb00172.x https://doi.org/10.1152/ajpgi.1984.247.4.g419 https://doi.org/10.1152/ajpgi.1984.247.4.g419 https://doi.org/10.1111/j.1365-2362.1993.tb01296.x d.g. levitt and m.d. levitt admet & dmpk 11(1) (2023) 57-80 80 [54] u. sharma, d. pal, r. prasad. alkaline phosphatase: an overview. indian j clin biochem 29 (2014) 269278. https://doi.org/10.1007/s12291-013-0408-y. [55] t. endo, h. ohbayashi, y. hayashi, y. ikehara, n. kochibe, a. kobata. structural study on the carbohydrate moiety of human placental alkaline phosphatase. j biochem 103 (1988) 182-187. https://doi.org/10.1093/oxfordjournals.jbchem.a122228. [56] j.s. clubb, f.c. neale, s. posen. the behavior of infused human placental alkaline phosphatase in human subjects. j lab clin med 66 (1965) 493-507. https://www.ncbi.nlm.nih.gov/pubmed/5835974. [57] m.d. levitt, m. rapoport, s.r. cooperband. the renal clearance of amylase in renal insufficiency, acute pancreatitis, and macroamylasemia. ann intern med 71 (1969) 919-925. https://doi.org/10.7326/0003-4819-71-5-919. [58] v.l. royse, d.m. jensen, h.l. corwin. pancreatic enzymes in chronic renal failure. arch intern med 147 (1987) 537-539. https://www.ncbi.nlm.nih.gov/pubmed/2435254. [59] m.d. levitt, s.m. hapak, d.g. levitt. alkaline phosphatase pathophysiology with emphasis on the seldom-discussed role of defective elimination in unexplained elevations of serum alp a case report and literature review. clin exp gastroenterol 15 (2022) 41-49. https://doi.org/10.2147/ceg.s345531. [60] k. goswami, d.n. nandakumar, b.c. koner, z. bobby, s.k. sen. oxidative changes and desialylation of serum proteins in hyperthyroidism. clin chim acta 337 (2003) 163-168. https://doi.org/10.1016/j.cccn.2003.08.009. [61] d.g. levitt, m.d. levitt. protein losing enteropathy: comprehensive review of the mechanistic association with clinical and subclinical disease states. clin exp gastroenterol 10 (2017) 147-168. https://doi.org/10.2147/ceg.s136803. [62] c.g. starr, p.m. tessier. selecting and engineering monoclonal antibodies with drug-like specificity. curr opin biotechnol 60 (2019) 119-127. https://doi.org/10.1016/j.copbio.2019.01.008. [63] d.g. levitt. general continuum analysis of transport through pores. i. proof of onsager's reciprocity postulate for uniform pore. biophys j 15 (1975) 533-551. https://doi.org/10.1016/s00063495(75)85836-x. ©2023 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) https://doi.org/10.1007/s12291-013-0408-y https://doi.org/10.1093/oxfordjournals.jbchem.a122228 https://www.ncbi.nlm.nih.gov/pubmed/5835974 https://doi.org/10.7326/0003-4819-71-5-919 https://doi.org/10.7326/0003-4819-71-5-919 https://www.ncbi.nlm.nih.gov/pubmed/2435254 https://doi.org/10.2147/ceg.s345531 https://doi.org/10.2147/ceg.s345531 https://doi.org/10.1016/j.cccn.2003.08.009 https://doi.org/10.1016/j.cccn.2003.08.009 https://doi.org/10.2147/ceg.s136803 https://doi.org/10.1016/j.copbio.2019.01.008 https://doi.org/10.1016/s0006-3495(75)85836-x https://doi.org/10.1016/s0006-3495(75)85836-x http://creativecommons.org/licenses/by/3.0/ effect of divalent and trivalent metal ions on artificial membrane permeation of fluoroquinolones doi: http://dx.doi.org/10.5599/admet.1427 289 admet & dmpk 10(4) (2022) 289-297; doi: https://doi.org/10.5599/admet.1427 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper effect of divalent and trivalent metal ions on artificial membrane permeation of fluoroquinolones1 nanami nakatani, kiyohiko sugano* molecular pharmaceutics lab., college of pharmaceutical sciences, ritsumeikan university, 1-1-1, noji-higashi, kusatsu, shiga 525-8577, japan *corresponding author: e-mail: suganok@fc.ritsumei.ac.jp; tel.: +81-77-561-2773 received: june 27, 2022; revised: september 02, 2022; published: september 07, 2022 abstract the purpose of the present study was to evaluate the predictability of pampa for the effect of metal ions on the bioavailability of fluoroquinolones (fq). eleven fqs and seven metal ions were employed in this study. the pampa membrane consisted of a 10 % soybean lecithin (sl) – decane solution. a drug solution in mes buffer with or without a metal ion (added as a chloride salt) was added to the donor compartment. in the absence of metal ions, fq showed relatively high permeability (> 5 × 10-6 cm/sec) in sl-pampa despite their hydrophilic and zwitterionic properties. as the pampa permeability ratio with/without metal ions became smaller, the urinary excretion and auc ratios tended to be smaller, suggesting that sl-pampa is a suitable in vitro model to evaluate the potential effect of metal ions on the bioavailability of fq. however, the reduction in auc and urinary excretion was overestimated for low solubility metal ion formulations (dried aluminum hydroxide gel and la2(co3)3・8h2o). in such cases, the dissolution of the metal ion formulations and the permeation of fqs should be simultaneously evaluated. ©2022 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords artificial membrane; permeability; phospholipid; fluoroquinolone; metal; cation introduction multivalent metal cations are present in many pharmaceutical and nutraceutical products, such as antacids, phosphate binders, and iron supplements. the concomitant administration of these multivalent metal salt formulations can reduce the bioavailability of various drugs, such as fluoroquinolones (fq) [1,2], tetracyclines [3], and hiv-integrase inhibitors [4]. in an acidic environment of the stomach, a metal cation exists as a dissociated ion. however, it could form a complex with a drug molecule in a neutral ph environment of the small intestine. it is well known that multivalent metal cations reduce the intestinal membrane permeation of some fqs by chelate formation [2]. to avoid fq multivalent cation interactions in the gastrointestinal tract, it is recommended to separate the drug administration times [5]. however, this may decrease patient medication adherence. currently, more than a dozen fqs and several multivalent metal 1 this paper is dedicated to the memory of dr. konstantin tsinman http://dx.doi.org/10.5599/admet.1427 https://doi.org/10.5599/admet.1427 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:suganok@fc.ritsumei.ac.jp http://creativecommons.org/licenses/by/4.0/ nakatani and sugano admet & dmpk 10(4) (2022) 289-297 290 salt formulations are prescribed in clinical practice. however, it is practically impossible to examine the effect of metal salt formulations on the bioavailability of fqs for all combinations by clinical studies. the parallel artificial membrane permeation assay (pampa) has been widely used to assess the passive membrane permeation of drugs [6–9]. recently, we reported that the membrane permeation of tetracyclines and the effect of metal ions can be evaluated using a phospholipid-based pampa [10,11]. however, it has been unknown whether pampa can predict the effect of metal salt formulations on the bioavailability of fqs. the purpose of the present study was to evaluate the predictability of pampa for the effect of metal ions on the bioavailability of fqs. eleven fqs and seven metal ions are employed in this study (figure 1, table 1). n o n nh f o oh ch3 o n o n n ch3 f o oh ciprofloxacin (cpfx) ofloxacin (oflx) ch3 n o n nh f o oh norfloxacin (nflx) ch3 n o n n nh f o oh enoxacin (enx) ch3 n n f n oo oh f f fleroxacin (flrx) ch3 n nho ch3 n oo oh f gatifloxacin (gflx) ch3 o n nh f o n o oh h h moxifloxacin (mflx) ch3 n o s o oh f n n o o o ch3prulifloxacin (pufx) ch3 n nh ch3 f n oo oh nh2 f sparfloxacin (spfx) n nh2 o n n f f o oh f tosufloxacin (tflx) ch3 o n o n n ch3 f o oh levofloxacin (lvfx) figure 1. chemical structures of fluoroquinolones used in this study experimental materials norfloxacin (nflx), ofloxacin (oflx), tosufloxacin tosylate monohydrate (tflx), calcium dichloride, magnesium dichloride, iron(ii) dichloride tetrahydrate, iron(iii) chloride hexahydrate, aluminum(iii) chloride hexahydrate, zinc chloride, lanthanum chloride heptahydrate, 1-octanol, sodium dihydrogen phosphate dihydrate, sodium chloride, sodium hydroxide, decane, 8 m naoh, and 0.1 vol% trifluoroacetic acidadmet & dmpk 10(4) (2022) 289-297 artificial membrane permeation of fluoroquinolones doi: http://dx.doi.org/10.5599/admet.1427 291 acetonitrile were purchased from wako pure chemical industries, ltd (osaka, japan). ciprofloxacin hydrochloride monohydrate (cpfx), levofloxacin (lvfx), enoxacin sesquihydrate (enx), fleroxacin (flrx), and prulifloxacin (pufx) were purchased from tci (tokyo, japan). gatifloxacin (gflx) and sparfloxacin (spfx) were purchased from funakoshi co., ltd (tokyo, japan). moxifloxacin (mflx) was purchased from combiblocks inc. (san diego, usa). 2-morpholinoethanesulfonic acid (mes) was purchased from dojindo laboratories (tokyo, japan). 0.1 vol% trifluoroacetic acid-distilled water was purchased from kanto chemical co., inc (tokyo, japan). soybean lecithin was provided by tsuji oil mills co., ltd (mie, japan) (slp-white, phosphatidylcholine (24-32 %), phosphatidylethanolamine (20-28 %), phosphatidylinositol (12-20 %), phosphatidic acid (8-15 %), and lysophosphatidylcholines (1-5 %) (based on the product information provided by the manufacturer)). table 1. physicochemical properties of fluoroquinolones used in this study drug mw pka log d (ph 6.5)a uv (nm) pka ref. ciprofloxacin (cpfx) 331 6.1, 8.7 -1.38 ± 0.15 335 [12] enoxacin (enx) 320 6.3, 8.7 -1.20 ± 0.39 330 [12] fleroxacin (flrx) 369 5.5, 8.1 -0.55 ± 0.10 283 [12] gatifloxacin (gflx) 375 5.7, 8.7 -0.97 ± 0.04 330 [13] levofloxacin (lvfx) 361 6.1, 8.2 -0.56 ± 0.11 335 [14] moxifloxacin (mflx) 401 6.4, 9.5 -0.72 ± 0.38 350 [12] norfloxacin (nflx) 319 6.3, 8.4 -1.35 ± 0.17 280 [12] ofloxacin (oflx) 361 6.1, 8.2 -0.55 ± 0.12 330 [12] prulifloxacin (pufx) 461 5.6, 6.3 0.86 ± 0.02 283 [15] sparfloxacin (spfx) 392 6.3, 8.8 -0.26 ± 0.03 300 [12] tosufloxacin (tflx) 404 5.8, 8.7 -0.14 ± 0.10 270 [16] a measured in this study. mean ± sd (n = 3) methods pampa assay the pampa sandwich consisted of a 96-well filter plate (hydrophobic pvdf, 0.45 μm) and a pampa acceptor plate (merck millipore, ma, usa). the acceptor plate was filled with 300 μl of a 50 mm mes buffer solution (ph 6.5). the filter bottom of the donor well was coated with 5 μl of a 10 % soybean lecithin (sl) – decane solution prepared in our laboratory [11]. the soybean lecithin–decane membrane (sl–pampa) was used because it most likely mimics the intestinal membrane [17]. a drug solution (0.5 mm in ph 6.5 50 mm mes buffer, 200 μl) with or without a metal ion was added to the donor compartment. the pampa sandwich was then incubated at 37 °c for 3 h without stirring. after incubation, 150 μl of both the donor and acceptor solutions were transferred to a uv plate. the concentrations of fqs were measured by uv absorption except for tflx (table 1). the concentration of tflx was measured by hplc (lc-20ad, shimazu corporation, kyoto, japan) (column: zorbax eclipse plus c18 narrow bore rr 2.1×50 mm column with 3.5 µm particles (agilent technologies, ca); flow rate: 0.60 ml/min; mobile phase: 20 % 0.1 vol% trifluoroacetic acid-acetonitrile, 80 % 0.1 vol% trifluoroacetic acid-distilled water (isocratic elution); detection wavelength: 270 nm; column temperature: 40 °c ; injection volume: 10 µl). the pampa permeability (pe) was measured in triplicate. the pe value was calculated by the following equations [17]. http://dx.doi.org/10.5599/admet.1427 nakatani and sugano admet & dmpk 10(4) (2022) 289-297 292 𝑃e = − 2.303𝑉d 𝐴∙𝑡 ( 1 1+𝑟𝑣 ) 𝑙𝑜𝑔 (−𝑟v + ( 1+𝑟v 1−𝑅 ) 𝐶d (𝑡) 𝐶d (0) ) (1) 𝑅 = 1 − 𝐶d(𝑡) 𝐶d (0) − 1 𝑟v 𝐶a (𝑡) 𝐶d (0) (2) 𝑟v = 𝑉d 𝑉a (3) where pe is the effective permeation coefficient (cm/s), a is the filter surface area (0.266 cm2), vd and va are the volumes (ml) of the donor and acceptor phases, respectively, t is the incubation time, cd(t) is the concentration of a drug in the donor phase at time t, r is the membrane retention factor, and rv is the volume ratio. these equations are generally used to calculate pe for pampa in the literature. the details for the derivation of these equations have been reported elsewhere [17]. in equations 1 to 3, the mass balance of a drug in the donor fluid, the acceptor fluid, and the membrane are considered. the r values ranged from < 0.1 (cpfx, enx, nflx, oflx, lvfx) to 0.36 (tflx) in this study. because it is difficult to collect samples multiple times from a 96-well plate, pampa pe is usually calculated from one-time point data. in typical pampa experiments, the steady state is achieved within 0 to 45 min [17]. therefore, pampa pe is usually calculated from one-time point data at > 2 h [6–9]. log d measurement of the fluoroquinolones the octanol-buffer distribution coefficient (log d) was determined at ph 6.5 by a shake-flask method. the octanol and buffer phases were mutually pre-saturated before use. a buffer solution of a model drug (1.0 mm, 0.5 ml, 50 mm sodium – phosphate buffer) and octanol (2.5 ml) was added to a 15 ml tube. the sample was vigorously shaken for 90 min at 25 °c. because log d is a physicochemical property of a drug, it was measured at a standard condition of 25 °c, rather than 37 °c. the concentration of fqs in the aqueous phase was determined as described above. log d was measured in triplicate. results effect of physicochemical properties of fluoroquinolones on pampa permeability all pampa permeability data are summarized in the supplemental material (table s1). the effect of the physicochemical properties of fqs on pampa pe was first investigated (figure 2). the pe value tended to increase as log d increased. above log d > 0, the pe value reached a plateau value of about 30 × 10-6 cm/s. there was no correlation between pe and the pka values. figure 2. correlation between pampa permeability, log d, and mw admet & dmpk 10(4) (2022) 289-297 artificial membrane permeation of fluoroquinolones doi: http://dx.doi.org/10.5599/admet.1427 293 in vitro – in vivo correlation of fq – metal ion interaction since metal ions can interact with phosphate and citrate ions, mes buffer was used in this study. the concentration of metal ions was set to reflect the clinical dose and the gastrointestinal fluid volume. in the presence of metal ions, the pe values were reduced in all fqs (supplemental material table s1). trivalent ions (al3+, fe3+, and la3+) tended to show a stronger inhibition effect than divalent ions. the correlations between the urinary excretion (ur) and auc ratios in humans and the pampa pe ratio are shown in figure 3 (categorized by metal ions) and also in supplemental material figure s1 (categorized by fq). clinical and in vivo canine data are summarized in supplemental material table s2 with the references. there is a general trend that as the pe ratio (with/without metal ions) becomes smaller, the reduction in auc and ur becomes more significant. however, pampa tended to overpredict the reduction in auc and ur, especially for al3+ and la3+ with high permeability fqs (pe > 10-5 cm/s). figure 3. correlation between pampa pe ratio with/without metal ions and (a) urinary excretion ratio and (b) auc ratio in humans (annotated by metal ions). the interactions between norfloxacin (nflx) and various metal ions have been investigated in dogs [18]. there is a good correlation between the pampa pe ratio and the in vivo cmax and auc ratios (figure 4). figure 4. correlation between the pampa pe ratio and (a) cmax and (b) auc ratios in dogs. http://dx.doi.org/10.5599/admet.1427 nakatani and sugano admet & dmpk 10(4) (2022) 289-297 294 discussion at ph 6.5, fq molecules exist as a zwitterion (table 1). according to the ph-partition hypothesis, zwitterionic drugs are often assumed to show poor membrane permeability [19]. however, we previously reported that zwitterionic tetracyclines can permeate the pampa membrane in the presence of phospholipids (soybean lecithin) (sl-pampa), despite their hydrophilic and zwitterionic properties [11]. in addition, sl-pampa has been used to investigate the effect of metal ions on the permeation of tetracyclines [10]. therefore, sl-pampa was used in this study. in the absence of metal ions, fqs showed relatively high permeability (> 5 × 10-6 cm/sec) in sl-pampa despite their hydrophilic and zwitterionic properties. the high pe values of fqs in sl-pampa are in good agreement with that these fqs show good oral bioavailability. even in the case of most hydrophilic fqs, nflx (log d = 1.35) and cpfx (log d = 1.38), the bioavailability is moderate and high (> 38 % and 83 %, respectively) [20,21]. these results suggested that sl-pampa is suitable to assess the effect of metal ions on the intestinal membrane permeation of zwitterionic drugs. the pe values of nflx and cpfx are about 10 times higher than previously reported values in pion’s pampa, which uses a 20 % dodecane solution of a lecithin mixture [19,22]. as log d was increased, the pe value was increased until reaching the plateau value (about 30 × 10-6 cm/s at log d > 0) (figure 2a). this plateau value would be attributed to the unstirred water layer adjacent to the pampa membrane [23]. as shown in figure 2b, pampa pe was also positively correlated with mw. the observed correlation between log pe and mw may be a mere coincidence due to the choice of fqs, but it may also be explained as follows. according to the solubility-diffusion theory [24,25], partition and diffusion processes determine the lipid membrane permeation of a drug. as mw increases, the diffusion coefficient decreases. on the other hand, the partition process depends on the lipophilicity of a drug. the lipophilicity of a drug is determined by the balance of drug-solvent interactions (e.g., hydrogen bonding) and the cavity effect, the latter of which is proportional to mw. therefore, when drug-solvent interaction is similar (like among fq series), log d increase with mw (figure 2c). in most cases of drug membrane permeation, the influence of a partition process is greater than that of a diffusion process. it is well known that the membrane permeability of drugs correlates with their lipophilicity [9]. therefore, even though an increase in mw decreases the diffusion process, pe can show a positive correlation with mw. more detailed analysis is required to further investigate this point, such as using abraham’s solute descriptors [26]. in this study, metal ions are added as chloride salts because they are soluble in aqueous media. the pe ratio of nflx was in good agreement with the cmax and auc ratios observed in dogs (figure 4) [18]. in that study, nflx was administered as a 0.01 n hcl solution with the metal chloride salts. therefore, the dissolution process of fqs and multivalent metal salt formulations were neglected. the effect of mg2+ in dogs was slightly overestimated by pampa. the reason for this was not clear. however, in clinical cases where these formulations are administered as solid dosage forms, pampa tended to overpredict the reduction of auc and cmax, especially for al3+ and la3+ formulations (figure 3). in clinical formulations, al3+ is contained as dried aluminum hydroxide gel and la3+ is contained as la2(co3)3・8h2o. ca2+, mg2+, fe2+, and zn2+ are also formulated as the other solid forms (caco3, mgo, etc.). some of these solid forms would dissolve slowly in the intestinal fluid so fqs can be absorbed before metal ions are released from the formulations, especially for high permeability fqs. for example, the dissolution of la2(co3)3・8h2o is about 50 % at 15 min in the gastric environment of ph 1.2 and there is little dissolution in the intestinal environment of ph 6.8 [27]. for more quantitative prediction, the dissolution of a metal ion formulation and the permeation of an fq should be simultaneously evaluated. admet & dmpk 10(4) (2022) 289-297 artificial membrane permeation of fluoroquinolones doi: http://dx.doi.org/10.5599/admet.1427 295 the detailed analysis of equilibria between fqs and metal ions is out of the scope of this study, as it has been reported for some of the fqs [28]. previously, wallis et al. reported that the formation constant of the nflx – metal ion complex correlates with in vivo bioavailability in dogs [18]. the chelate formation constant (log unit) and the percentage of nflx complexed at ph 6.5 are reported to be 2.2 (8.5 %), 3.0 (35 %), 3.8 (81 %), 4.0 (78 %), and 7.0 (99 %) for ca2+, mg2+, zn2+, fe2+, and al3+, respectively. in the present study, the trivalent metal ions (al3+, fe3+, la3+) reduced pe more than the divalent metal ions. among the divalent metal ions, no general trend was observed, except that mg2+ reduced pe more than ca2+ in all fqs. this trend was not observed in tetracyclines [10]. conclusion in conclusion, the pe values of fqs are relatively high in sl-pampa. as the pe ratio (with/without metal ions) becomes smaller, the reduction in auc and ur tended to be more significant. however, the reduction in auc and ur is overestimated for low solubility metal ion formulations (especially al3+ and la3+ formulations). in such cases, the dissolution of a metal ion formulation and the permeation of an fq should be simultaneously evaluated for more quantitative prediction by an in vitro experiment. sl-pampa is suitable to evaluate the potential effect of metal salt formulations on the bioavailability of fqs. conflict of interest: authors declare no conflict of interest. references [1] a. stojković, l. tajber, k.j. paluch, z. djurić, j. parojčić, o.i. corrigan. biopharmaceutical characterisation of ciprofloxacin-metallic ion interactions: comparative study into the effect of aluminium, calcium, zinc and iron on drug solubility and dissolution. acta pharm. 64 (2014) 77–88. https://doi.org/10.2478/acph-2014-0007. [2] ž. simon, b. katja, u. darko, v. marjan, k. albin. metal cation-fluoroquinolone complexes do not permeate through the intestinal absorption barrier. j. pharm. biomed. anal. 53 (2010) 655–659. https://doi.org/10.1016/j.jpba.2010.05.021. [3] j.t. doluisio, a.n. martin. metal complexation of the tetracycline hydrochlorides. j. med. chem. 6 (1963) 16–20. https://doi.org/10.1021/jm00337a003. [4] r. krishna, l. east, p. larson, c. valiathan, k. butterfield, y. teng, m. hernandez-illas. effect of metalcation antacids on the pharmacokinetics of 1200 mg raltegravir. j. pharm. pharmacol. 68 (2016) 1359–1365. https://doi.org/10.1111/jphp.12632. [5] s.k. pitman, u.t.p. hoang, c.h. wi, m. alsheikh, d.a. hiner, k.m. percival. revisiting oral fluoroquinolone and multivalent cation drug-drug interactions: are they still relevant? antibiotics. 8 (2019). https://doi.org/10.3390/antibiotics8030108. [6] m. kansy, f. senner, k. gubernator. physicochemical high throughput screening: parallel artificial membrane permeation assay in the description of passive absorption processes. j. med. chem., 41 (1998), 1007-1010. https://doi.org/10.1021/jm970530e. [7] a. avdeef, s. bendels, l. di, b. faller, m. kansy, k. sugano, y. yamauchi. pampa critical factors for better predictions of absorption. j. pharm. sci. 96 (2007). https://doi.org/10.1002/jps.21068. [8] l. di, p. artursson, a. avdeef, l.z. benet, j.b. houston, m. kansy, e.h. kerns, h. lennernäs, d.a. smith, k. sugano. the critical role of passive permeability in designing successful drugs. chemmedchem. (2020) 1862-1874. https://doi.org/10.1002/cmdc.202000419. [9] k. sugano, m. kansy, p. artursson, a. avdeef, s. bendels, l. di, g.f. ecker, b. faller, h. fischer, g. gerebtzoff, h. lennernaes, f. senner. coexistence of passive and carrier-mediated processes in drug transport. nat. rev. drug discov. 9 (2010) 597-614. https://doi.org/10.1038/nrd3187. http://dx.doi.org/10.5599/admet.1427 https://doi.org/10.2478/acph-2014-0007 https://doi.org/10.1016/j.jpba.2010.05.021 https://doi.org/10.1021/jm00337a003 https://doi.org/10.1111/jphp.12632 https://doi.org/10.3390/antibiotics8030108 https://doi.org/10.1021/jm970530e https://doi.org/10.1002/jps.21068 https://doi.org/10.1002/cmdc.202000419 https://doi.org/10.1038/nrd3187 nakatani and sugano admet & dmpk 10(4) (2022) 289-297 296 [10] s. yamauchi, d. inoue, k. sugano. permeation characteristics of tetracyclines in parallel artificial membrane permeation assay ii: effect of divalent metal ions and mucin. admet dmpk 8 (2020) 129138. https://doi.org/10.5599/admet.797. [11] s. yamauchi, k. sugano. permeation characteristics of tetracyclines in parallel artificial membrane permeation assay. admet dmpk 7 (2019) 151–160. https://doi.org/10.5599/admet.657. [12] f. sörgel, j. bulitta, m. kinzig-schippers. pharmakokinetik der chinolone. pharmaziein unserer zeit. (2001) 418–427. https://doi.org/10.1002/1615-1003(200109)30:5<418::aid-pauz418>3.0.co;2%23 (accessed august 24, 2022). [13] m. caianelo, c. rodrigues-silva, m.g. maniero, j.r. guimarães. antimicrobial activity against grampositive and gram-negative bacteria during gatifloxacin degradation by hydroxyl radicals. environ. sci. pol. res. 24 (2017) 6288-6298. https://doi.org/10.1007/s11356-016-6972-y. [14] levofloxacin hydrate drug product information (cravit tablets), daiichi sankyo company, https://www.info.pmda.go.jp/go/pack/6241013c2024_1_21 (accessed august 24, 2022). [15] p. ravisankar, ch.v.p. rao, n.a. qureshi, h.a. alsubaie, m.k. alharbi, g.i.m. ali, s.m. alsanad, s.a. ali, s. mahanand, s.w.` khan, k.t. magalhães-guedes, j.i. druzian, r.f. schwan, i.l. nunes, j.a.v. costa, t. attard, e. attard, l.m. shengo, t.h. mundongo, i.u. umoh, a.u. ekanem, s.u. rehman, h.m. rafique, f.a. qureshi, s. shahid, h.-k. kang, h.-j. kim, j.-s. kang, i. dey, t.k. das, c. roy, m.k. pillai, s. magama, l.i. santi, n. katsuto, s. jin-ichi, m.s.t. sheikuduman, b. mangla, a. jain, v.r. salunkhe. trends in pharmaceutical research and development vol. 1, book publisher international (a part of sciencedomain international), 2020. https://doi.org/10.9734/bpi/tprd/v1. [16] tosufloxacin tosylate drug product information (ozex tablets), fujifilm toyama chemical co. ltd. https://www.info.pmda.go.jp/go/pack/6241010f1020_2_06/ (accessed august 24, 2022). [17] a. avdeef. absorption and drug development: solubility, permeability, and charge state, second edition, john wiley & sons, inc., 2012. https://doi.org/10.1002/9781118286067. [18] s.c. wallis, b.g. charles, l.r. gahan, l.j. filippich, m.g. bredhauer, p.a. duckworth. interaction of norfloxacin with divalent and trivalent pharmaceutical cations. in vitro complexation and in vivo pharmacokinetic studies in the dog. j. pharm. sci. 85 (1996) 803–809. https://doi.org/10.1021/js960087f. [19] k.y. tam, a. avdeef, o. tsinman, n. sun. the permeation of amphoteric drugs through artificial membranes an in combo absorption model based on paracellular and transmembrane permeability. j. med. chem. 53 (2010) 392–401. https://doi.org/10.1021/jm901421c. [20] n. campbell, m. kara, b. hasinoff, w. haddara, d. mckay. norfloxacin interaction with antacids and minerals. br. j. clin. pharmacol. 33 (1992) 115–116. https://doi.org/10.1111/j.1365-2125.1992.tb04010.x. [21] ciprofloxacin hcl drug product information (ciproxan). bayer. https://www.info.pmda.go.jp/go/pack/6241008f1023_1_23/ (accessed august 24, 2022). [22] m. bermejo, a. avdeef, a. ruiz, r. nalda, j.a. ruell, o. tsinman, i. gonzález, c. fernández, g. sánchez, t.m. garrigues, v. merino. pampa a drug absorption in vitro model: 7. comparing rat in situ, caco2, and pampa permeability of fluoroquinolones. eur. j. pharm. sci. 21 (2004) 429–441. https://doi.org/10.1016/j.ejps.2003.10.009. [23] p.e. nielsen, a. avdeef. pampa a drug absorption in vitro model: 8. apparent filter porosity and the unstirred water layer. eur. j. pharm. sci. 22 (2004) 33–41. https://doi.org/10.1016/j.ejps.2004.02.003. [24] t.x. xiang, y.h. xu, b.d. anderson. the barrier domain for solute permeation varies with lipid bilayer phase structure. j. membr. biol. 165 (1998) 77-90. https://doi.org/10.1007/s002329900422. [25] j.m. diamond, y. katz. interpretation of nonelectrolyte partition coefficients between dimyristoyl lecithin and water. j. membr. biol. 17 (1974) 121-154. https://doi.org/10.1007/bf01870176. [26] m.h. abraham. scales of solute hydrogen-bonding: their construction and application to physicochemical and biochemical processes. chem. soc. rev. 22 (1993) 73-83. https://doi.org/10.1039/cs9932200073. https://doi.org/10.5599/admet.797 https://doi.org/10.5599/admet.657 https://doi.org/10.1002/1615-1003(200109)30:5%3c418::aid-pauz418%3e3.0.co;2-%23 https://doi.org/10.1002/1615-1003(200109)30:5%3c418::aid-pauz418%3e3.0.co;2-%23 https://doi.org/10.1007/s11356-016-6972-y https://www.info.pmda.go.jp/go/pack/6241013c2024_1_21 https://doi.org/10.9734/bpi/tprd/v1 https://www.info.pmda.go.jp/go/pack/6241010f1020_2_06/ https://doi.org/10.1002/9781118286067 https://doi.org/10.1021/js960087f https://doi.org/10.1021/js960087f https://doi.org/10.1021/jm901421c https://doi.org/10.1111/j.1365-2125.1992.tb04010.x https://doi.org/10.1111/j.1365-2125.1992.tb04010.x https://www.info.pmda.go.jp/go/pack/6241008f1023_1_23/ https://www.info.pmda.go.jp/go/pack/6241008f1023_1_23/ https://doi.org/10.1016/j.ejps.2003.10.009 https://doi.org/10.1016/j.ejps.2004.02.003 https://doi.org/10.1016/j.ejps.2004.02.003 https://doi.org/10.1007/s002329900422 https://doi.org/10.1007/bf01870176 https://doi.org/10.1039/cs9932200073 https://doi.org/10.1039/cs9932200073 admet & dmpk 10(4) (2022) 289-297 artificial membrane permeation of fluoroquinolones doi: http://dx.doi.org/10.5599/admet.1427 297 [27] lanthanum carbonate granules drug product information. nipro corporation. https://med.nipro.co.jp/servlet/servlet.filedownload?file=00p2x00000bcdptean (accessed august 24, 2022). [28] b. urbaniak, z.j. kokot. analysis of the factors that significantly influence the stability of fluoroquinolone-metal complexes. anal. chim. acta 647 (2009) 54–59. https://doi.org/10.1016/j.aca.2009.05.039. ©2022 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://dx.doi.org/10.5599/admet.1427 https://med.nipro.co.jp/servlet/servlet.filedownload?file=00p2x00000bcdptean https://doi.org/10.1016/j.aca.2009.05.039 https://doi.org/10.1016/j.aca.2009.05.039 http://creativecommons.org/licenses/by/3.0/) manuscript doi: 10.5599/admet.2.4.147 199 admet & dmpk 2(4) (2014) 199-220; doi: 10.5599/admet.2.4.147 open access: issn: 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper linear combination methods for prediction of drug skin permeation stefan scheler 1, , alfred fahr 2 , xiangli liu 3 1 sandoz gmbh, biopharmaceuticals, pharmaceutical & device development, biochemiestrasse 10, 6336 langkampfen, austria 2 friedrich schiller university of jena, department of pharmaceutical technology, lessingstr. 8, 07743 jena, germany 3 bradford school of pharmacy, university of bradford, bd7 1dp, uk corresponding author: e-mail: stefan.scheler@sandoz.com; tel.: +43 5372 69962793; fax: +43 5372 69965672 received: november 15, 2014; revised: december 22, 2014; published: january 09, 2015 abstract many in-vitro methods for prediction of skin permeability have been reported in literature. cerasome electrokinetic chromatography is one of the most sophisticated approaches representing a maximum level of similarity to the lipid phase of the stratum corneum. one goal of this study was to investigate the affinity pattern of cerasome and to compare it with the permeability profile of human skin. another purpose was to study the applicability of hansen solubility parameters for modelling skin permeation and to investigate the predictive and explanatory potential of this method. visualisation in hansen diagrams revealed very similar profiles of cerasome electrokinetic chromatography retention factors and skin permeability coefficients. in both cases, the characteristic pattern with two clusters of highly retained or highly permeable substances could be shown to be mainly caused by two groups of compounds, one of them with high affinity to ceramides, fatty acids and lecithin and the other being more affine to cholesterol. if based on a sufficiently comprehensive experimental dataset, model-independent predictions of skin permeability data using three-component hansen solubility parameters are able to achieve similar accuracy as calculations made with an abraham linear free energy relationship model in which the compounds are characterized by seven physicochemical descriptors. keywords: cerasome 9005; liposome electrokinetic chromatography; linear free energy relationship analysis; solubility parameter; stratum corneum introduction the transdermal administration route is one of oldest, safest, and most convenient for the patient and can be used for both, local and systemic therapy. however, the primary function of the skin is that of a barrier rather than of an absorption organ. an important role is played by the stratum corneum, the outermost layer of the epidermis. it consists of 15-25 layers of non-viable cells, each of which is about 0.2-0.5 mm thick. as they had become cornified in their terminal stage of differentiation, they are mainly composed of keratin. the interstices between these corneocytes are tightly sealed with various lipids, accounting for about 5-15 % of the stratum corneum [1]. though the stratum corneum is the thinnest layer of the skin at the preferred application sites it is the rate limiting structure for transepidermal drug transport. disregarding the less pronounced drug penetration via pores (transglandular, transfollicular), two types of transepidermal transport can be distinguished: the transcellular and the intercellular route [2]. in case of the first mentioned type the drug has to pass both, the intercellular lipid layers and the hydrophilic http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:stefan.scheler@sandoz.com scheler et al. admet & dmpk 2(4) (2014) 199-220 200 cytoplasma and keratin structures, which requires a very well balanced hydrophilicity/lipophilicity ratio of the molecules [3]. hence, for many drugs and other substances permeation through the skin is mainly accomplished by intercellular transport where the drug diffuses through the lipid filled interstices between the corneocytes [4]. the steady-state flux, js,v of a drug from a vehicle through the skin can be derived, according to fick’s laws, from the concentration of the drug in the vehicle (cv) and the permeability coefficient, kp [5]: vpvs ckj  , (1) kp is defined as the product of the partition coefficient (ps,v) and the diffusion coefficient (ds) divided by the thickness of the skin (hs): s svs p h dp k   , . (2) thus js,v depends on the partition of the drug between the vehicle and the skin. in case of intercellular transport this means partition between the vehicle and the intercellular lipids. it is known that ds correlates only slightly with the chemical structure of the substance, however ps,v is strongly related to the structural properties of the molecules. hence, the permeability coefficient of lipid soluble substances is essentially determined by the lipid/water distribution, which is consequently a main factor governing the flux through the skin. the determination of solubilities in stratum corneum lipids poses experimental difficulties, e.g. with respect to a standardized and well defined composition of the mixture. for this reason, the solubility in other lipophilic solvents, most commonly octanol, is used as a surrogate [6]. however, in contrast to stratum corneum lipids, those substances, like aliphatic alcohols or hydrocarbons, are uncharged and do not show similar interactions with electrolytes as they have to be assumed, for example, with stratum corneum fatty acids. in addition, the ordered structure of lipids arranged in membranes might account for a modified affinity to different solutes as compared with a physical mixture of the same lipids. for this reason, many in-vitro models, mimicking the composition and the structure of human stratum corneum, have been investigated so far in order to mathematically describe the process of skin permeation [7-11]. likewise, different methods have been studied to analyse the raw data obtained in such experimental setups. many of these methods aim to predict the skin penetration from chemical, physicochemical or structural data of the compounds without subjecting them to any experimental test. one of these methods is linear free energy relationship analysis (lfer). the study presented here investigates another approach. based on data obtained from electrokinetic chromatography (lekc) and human skin permeation, hansen solubility parameters (hsp) were applied for data analysis, and advantages and disadvantages of this method were studied in comparison to lfer. as mentioned above, partitioning of the drug into the stratum corneum lipids is only one crucial step in drug skin permeation. another step is the diffusion through the intercellular space. with in-vitro skin models it is not possible to analyse these processes individually. quantitative structure-property relationship approaches modelling the drug skin permeation process as a whole do not facilitate a differentiated analysis of individual factors of influence. for this reason lekc was applied as a model to obtain data reflecting exclusively partitioning into stratum corneum lipids. by separate analysis of both steps, partitioning and diffusion, much deeper insight can be gained regarding the significance and role of involved structures and molecules. admet & dmpk 2(4) (2014) 199-220 prediction of drug skin permeation doi: 10.5599/admet.2.4.147 201 theory liposome electrokinetic chromatography (lekc) chromatographic methods based on immobilized artificial membranes (iams, monolayers of phospholipid analogs covalently bonded to the surface of silica particles) and micelles (bmc = biopartitioning micellar chromatography) have been applied to better mimic the lipid structures in the stratum corneum and to improve the prediction of lipid distribution and skin penetration [7,8]. one of the most sophisticated techniques in this field is liposome electrokinetic chromatography (lekc) [9,10]. as liposomes possess lipid bilayer structures they resemble biological biomembranes much more than the aforementioned systems. most studies applying lekc for studying skin penetration have used conventional phosphatidylcholine (pc)/phosphatidylserine (ps) liposomes which do not reflect the composition of the lipid layers in the stratum corneum. zhang et al. performed lekc measurements using cerasome 9005, a 6.6 % (w/w) vesicular dispersion composed of hydrogenated lecithin, ceramides (np and ns), cholesterol and fatty acids (palmitic acid and oleic acid) [11]. the mixture resembles the composition of the stratum corneum lipids which consist of 40-50 % ceramides, about 25 % sterols, mainly cholesterol, and 10-20 % free fatty acids. however, in contrast to cerasome, the intercellular stratum corneum lipids are free of phospholipids. anyway, they form multilamellar sheets as a result of the steric interaction of amphiphilic ceramides with cholesterol and free fatty acids [12-15]. originally developed and marketed as a prefabricated liposome dispersion for cosmetic and drug formulation, the cerasome dispersion was used by the cited authors as a running solution in lekc, in order to model drug partitioning in stratum corneum lipids and consequently skin permeation. with these data, collected from 71 different compounds, they established a quantitative relationship between the structurally determined physicochemical descriptors and the retention factors in lekc using lfer. linear free energy relationship analysis (lfer) lfer (and also the very similar lser = linear solvation energy relationship [16]) is a type of a quantitative structure-property relationship (qspr). it represents a mathematical model able to predict physicochemical properties (e.g. permeation, absorption, partitioning, solubility or, in case of this work, lekc retention factors) from chemical molecular descriptors of a substance. in lfer the logarithm of the physicochemical property depends linearly on a free energy change as also the various descriptors are related to gibbs free energy. a widely used model, developed by abraham is based on the following equation,   jjjjvvbbaasseecsplog (3) where sp is a free energy-related physicochemical property of a compound in a given solvent system. the lower case letters (e, s, a, b, v, j + , j ) are system constants describing the contribution of the solvent phase to the sorption process [11]. the capital letters are solute descriptors representing the properties of the compounds. in detail, these are the excess molar refraction (e), the solute dipolarity/polarizability (s), the overall solute hydrogen bond acidity (a) and hydrogen bond basicity (b), the mcgowan characteristic volume (v), a descriptor for cations (j + ), and another for anions (j ). v is calculated by summing up all atom volumes of the molecule and subtracting 6.56 cm 3 mol -1 for each bond. zhang et al. demonstrated that eq. (3) is able to describe both the retention factors in lekc and the distribution coefficient at ph 7.4 (d7.4) between water and different solvents for all tested compounds including non-ionic and ionic substances [11]. scheler et al. admet & dmpk 2(4) (2014) 199-220 202 solubility parameters another, quite different approach for computation of solubility related properties is the application of solubility parameters. a main purpose of this study is the comparison of both systems, lfer and solubility parameters, with respect to their ability to predict skin penetration behaviour and the properties of different substances in in-vitro models. in order to work out and to understand analogies and differences between both systems, some general considerations on solubility parameters have to be made first. the concept of solubility parameters is based on the cohesive energy density (ced) which is the energy per unit volume necessary to separate the molecules of a solute or solvent to a distance where all cohesive interactions have disappeared. the simplest system of solubility parameters () was introduced by hildebrand and scott, who defined them as the square root of the ced:   5.0 mv veced  (4) where ev is the energy of vaporization and vm is the molar volume [17]. according to eq. (5), materials having similar solubility parameters have a low non-combinatorial free energy of mixing, g m noncomb, and therefore a high affinity for each other.  2 2121   m m noncomb vg (5) 1 and 2 are the volume fractions of the components and vm is the volume of the mixture. the noncombinatorial free energy of mixing includes all free energy effects other than the combinatorial entropy of solution, occurring because of simple mixing the components [18]. as the free energy of mixing (gm) is the difference between the non-combinatorial energy change and the combinatorial entropy change (ts m comb), m comb m noncomb m stgg  (6) the maximum difference of solubility parameters which is just small enough for solution to occur (i.e. g m = 0), can be calculated from the condition: m comb m noncomb stg  (7) it follows from the above that g m noncomb includes both terms by which g m of a real solution differs from that of an ideal solution, the excess enthalpy and the excess entropy. when calculating the mole fraction solubility, x2, of a substance in a solvent forming a real solution these corrections are accounted for by the rational activity coefficient, 2.     22 1/2 1/20 0 0 2 1 2 2 2 11 22 0 ln ln ln ln f pi h ct t t t t v x x w w rt t r t t rt                             (8) where x2 i is the ideal mole fraction solubility, hf is the heat of fusion of the solute, t is the temperature and t0 is the melting point of the solute, v2 is the molar volume of the solute, 1 is the volume fraction of the solvent and r is the gas constant. cp is the difference in heat capacity between the solid form and the hypothetical super-cooled liquid form of the solute. w11 and w22 are the energy required to generate a void between adjacent solvent molecules and the energy required to overcome the attractive forces between the solute molecules, respectively. the square roots of these energies represent the hildebrand solubility parameters of the solvent (1) and the solute (2). inserting these parameters, eq. (8) can be written as admet & dmpk 2(4) (2014) 199-220 prediction of drug skin permeation doi: 10.5599/admet.2.4.147 203   rt v t t t tt r c t tt rt h x pf 2 122 21 00 0 0 2 lnln                           . (9) most of the compounds investigated in this study are in a solid state at room temperature, however, solubility parameters are based on the enthalpy changes occurring when liquids are mixed. this fact is taken into account by the subtractive term in eq. (9) which is related to the energy required to convert the solids into the hypothetical super-cooled liquids. since the heat capacity of the super-cooled liquid normally cannot be determined, it is a common assumption, valid for nonpolar and for polar aprotic solvents, that cp is small and can be assumed to be zero. this simplifies eq. (9) to eq. (10) [19]:   rt v t tt rt h x f 2 122 21 0 0 2 ln             . (10) though still widely used, the concept of hildebrand takes insufficient account of the fact that different types of forces are acting between the molecules in a mixture, all of them contributing to the miscibility and other interaction properties. many approaches have been made to introduce corrections or to subdivide the hildebrand parameters into partial solubility parameters dedicated to different types of intermolecular interactions. the most widely used method is a three-component parameter model proposed by c.m. hansen [18]. he introduced a system of three parameters, d, p, and h, contributing to dispersive (van der waals), polar and hydrogen-bond interactions, respectively. hpdt   (11) t is the total solubility parameter which is numerically equal to the hildebrand parameter. using only three components, hansen solubility parameters (hsps) have the advantage to be visualizable in a threedimensional coordinate system. another benefit is that hsps for a huge number of various substances are listed in literature and that group contribution methods allow for estimation of hsp values on the basis of a given structure formula. due to its high versatility, the hsp concept has gained widespread use, in many fields of application, like solvent selection, polymer synthesis and engineering, surface characterization, chemical and biological compatibility studies, and even in pharmaceutical development, e.g. in the areas of drug formulation [20-30] or permeability across biological barriers [32,33]. also the use of solubility parameters (hansen or hildebrand) for prediction of skin transport was investigated [5, 34-37]. as the first summand in eq. (10) is the ideal mole fraction solubility x2 i , the logarithm of the rational activity coefficient ln 2 can be written as:   rt v x x i 2 122 21 2 2 2 lnln    . (12) consequently the squared difference of the solubility parameters can be expressed as:   2 212 12 22 )( )/( /ln    rtv xx i . (13) in terms of hansen parameters, the simple univariate quadratic polynomial of eq. (13) can be extended to a trivariate polynomial, scheler et al. admet & dmpk 2(4) (2014) 199-220 204   2 213 2 212 2 21102 12 22 )()()( )/( /ln hhppdd i dddd rtv xx    (14) which can also be written as [38] 2 213 2 212 2 21102 )()()(ln hhppdd ddddx   (15) with d0, d1, d2, d3and d’0, d’1, d’2, d’3 being constants. as is evident from eq. (14) and eq. (15), also the solubility parameter approach can be considered as a linear combination of terms attributed to different types of intermolecular interactions, similar to the abraham lfer model. however, in this case the descriptors (1x-2x) do not represent the solute properties as such but the differences between solvent and solute properties. whilst in the abraham lfer approach the solvent specific constants include both, the solvent properties and a scaling factor to fit the predicted quantity, the constants d and d’ in eq. (14) and eq. (15) are mere scaling factors, correlating each pair (squared difference) of solvent and solute parameter to the physicochemical target property (ln x2). as the solubility parameter approach separates the scaling factors from the solvent parameters, it may better allow for comparing the response for different solvents. as the description of the predicted property is not centred to a particular solvent, it facilitates the graphical depiction of the relationships between solvents and solutes in a diagram with an absolute origin which is not determined by the applied solvent. irrespective of these advantages, the classical hansen model has one major drawback. the systems of hildebrand and hansen solubility parameters are based on the “similarity matching” rule “like dissolves like”. however, it should be considered that particularly in case of acidic and basic components (or also in case of lewis acids and bases) “complementary matching” can be the dominating principle. in order to quantify electron-donor and acceptor properties, beerbower et al. developed a four parameter system in which h is substituted by an acidic (a) and a basic (b) solubility parameter (h 2 =2ab) [39]. this subdivision of the hydrogen-bonding parameter into an acidity and a basicity component has its analogue in the overall solute hydrogen bond acidity and basicity descriptors in lfer. unfortunately, in contrast to hsbs, only very limited sources of a and b values are available in literature and a group contribution method does not exist. for this reason, this study was based on three-component hansen parameters in their non-expanded form. a main goal of this work was to study the applicability of the hsp method to analyze the behaviour of different compounds in cerasome lekc. it is scrutinized whether the hansen approach is able to elucidate further relationships remaining latent with the abraham model and whether it can thus support or complement the abraham approach. furthermore, the applicability of hsps to predict the distribution of compounds in other model systems and the permeation through human skin are addressed. materials and methods data sets used in this study all following considerations regarding the interpretation of lekc retention factors are based on a comprehensive dataset created by zhang, k. et al. [11], which comprises experimentally determined log k7.4 values of 70 compounds (tab. 1, see appendix). for comparison with skin penetration data, another set of 168 substances was compiled from 6 different sources (tab. 2, see appendix). hansen solubility parameters were taken from literature if available. in all other cases, where experimentally determined values could admet & dmpk 2(4) (2014) 199-220 prediction of drug skin permeation doi: 10.5599/admet.2.4.147 205 not be found, hsps were calculated using the group contribution method of hoftyzer and van krevelen [40]. octanol-water distribution coefficients (log d7.4) were compiled from six sources. if not available from literature they were calculated either with acd/physchem suite (acd/labs) or with the property viewer of chemaxon chemicalize.org, both accessible via the chemical database chemspider (royal society of chemistry, london, uk) (www.chemspider.com). intepolation method for prediction of lekc retention factors and of human skin permeability coefficients in this study the k-nearest neighbor method was used for prediction of lekc retention factors and of human skin permeability data. the k-nearest neighbor method is a non-parametric, local, and deterministic spatial interpolation technique which predicts the unknown value of a data point as the average value of its k nearest neighbor points. the calculation of the algorithm was performed using microsoft excel. first the euclidean distance between the unknown and each known data point is calculated. then for the two (k=2 in case of lek retention factors) or three (k=3 in case of skin permeability coefficients) least distant known data points the retention factors or the skin permeability coefficients were looked up and averaged. results and discussions all calculations and considerations made in this study are based on the two sets of data mentioned in the materials and methods section. one of them comprises in-vitro data measured with cerasome lekc, the other encompasses human skin penetration data compiled from different sources of literature. to every compound of each dataset a three-component hsp was assigned, describing its interaction properties with respect to dispersive, polar, and hydrogen bonding forces. if available from literature, experimentally determined hsp values were used. in all other cases hsps were calculated by a group contribution method. hsps are always subject to a certain level of uncertainty, which is lower in case of experimentally created data. hsps obtained by group contribution are described to have an accuracy of approximately 10% [41]. visualization of lekc retention factors or stratum corneum permeabilities in dependence of hsps requires plotting a dependent variable versus three independent variables (d, p, h). as these four-dimensional datasets cannot be visualized in plane diagrams, different methods have been developed in the past to solve this problem by means of data reduction. one approach is to merge two partial solubility parameters by pythagorian addition, for example d and p, as realized in bagley diagrams [35]. another approach, known as teas plot, is the presentation of fractional solubility parameters (e.g. d d + p + h)) in a three component diagram. all these methods have in common a loss of information thus distorting the results [42]. for this reason, three dimensional diagrams (often called hansen space) were chosen here to depict the retention or permeation properties of the tested substances in correlation to hsps. description of cerasome lekc retention factors in terms of hsps fig. 1 shows the data points of all compounds tested in lekc experiments within the hansen space. the retention factors are represented by different shapes of the markers. in contrast to hansen’s basic concept, where all good solvents are clustered within one single solubility sphere, no such single-centered area covering the data points of all compounds with large retention factors can be found in this diagram. instead not only one but two focal areas are formed by strongly retained substances which are separated by a broad region with data points of better eluting compounds. one of these clusters can be found in the region of d = 15 to 18 mpa 1/2 , p = 0 to 4 mpa 1/2 and a second cluster is seen between d = 19 to 21 mpa 1/2 , http://www.chemspider.com/ scheler et al. admet & dmpk 2(4) (2014) 199-220 206 p = 3 to 8 mpa 1/2 . such a pattern of more than one area of high solubility is described to be typical for surfactants and copolymers [43,44]. it is characteristic for molecules in which sites with different interactive properties are spatially separated from one another or where such sites of a number of molecules cluster together. by contrast, mixtures of different types of small molecules without any ordered supramolecular structure do not show multi-centered pattern of solubility in hsp diagrams. instead, they reveal a single solubility maximum whose coordinates can be calculated as the volume-fraction-weighted average of the components’ hsps. hence, the complex pattern of lekc retention factors within the hansen space indicates interactions either with amphiphilic molecules or with ordered structures. looking at the components of cerasome 9005, this product represents a mixture of a variety of amphiphilic compounds. in view of its liposomal nature it must be assumed that most of the components are regularly arranged, thus forming extended hydrophilic and hydrophobic regions. it is difficult to derive experimentally determined hsp values for lecithin, ceramides, cholesterol and fatty acids from literature. only for the last two mentioned substances tabular values could be found. hsps for dipalmitoylphosphatidylcholine (dppc) and n-stearoyl sphingosine (as typical members of the lecithinand ceramide group) were calculated using the method of hoftyzer and van krevelen [40]. fig. 1 shows the hsp coordinate positions of the cerasome components together with the test samples of the lekc experiments. interestingly, compounds with high k7.4 values are mainly clustered around the data points of the cerasome lipids, indicating mutual interactions. the solubility parameters of lecithin, ceramides and fatty acids are relatively close to each other with d being in the range of 16.3 to 16.8 mpa 1/2 , p ranging from 1.3 to 2.8 mpa 1/2 , and h between 5.6 and 8.5 mpa 1/2 . as can be seen by comparison with hexadecane (d = 16.3 mpa 1/2 , p = 0 mpa 1/2 , h = 0 mpa 1/2 ) and octadecane (d = 16.4 mpa 1/2 , p = 0 mpa 1/2 , h = 0 mpa 1/2 ), d is mainly determined by the aliphatic chains of the lipids. from homologue series of hydrocarbons it can be deduced that longer chains result in higher d values. however, values of 20 mpa 1/2 and above cannot be explained by aliphatic carbon chains. as solubility parameters represent the cohesion energy per molar volume, such high dispersion parameters are characteristic for compactly structured polycyclic hydrocarbons like cholestane. for this reason, d of cholesterol (20.4 mpa 1/2 ) [45] is incomparably higher than the value of the other, non-cyclic lipids. due to the amphiphilic character of cholesterol, enabling hydrogen bonding as well as van der waals binding to low polar compounds, it can be assumed that p and h parameters of substances with affinity to cholesterol are ranging over a wide scale. indeed, the cluster of strongly retained compounds around the data point of cholesterol shows a large extension both in p and h direction. in summary, strongly retained compounds form two clusters in the hansen space, one caused by the affinity to cholesterol and the other by binding to lecithin, ceramides and fatty acids. both clusters are divided from each other by a region comprising the data points of substances with low k7.4 values. as discussed before, the occurrence of two separate solubility regions, which is typically not observed with mixtures of miscible substances, raises the suspicion that in the cerasome liposomes cholesterol is not molecularly dispersed within the other lipids but in some way separated. indeed, it is reported that ternary mixtures of 1-palmitoyl-2-oleoyl-sn-glycero3-phosphocholine, ceramide (18:0) and cholesterol with fractions of 50% ceramide and 25% cholesterol, resembling the ratio in stratum corneum lipids, reveal a separate phase of cholesterol monohydrate crystals besides a bilayer phase containing all three components [46]. also in lipid membranes separated sphingolipid-rich and cholesterol-rich ordered lipid domains are believed to exist [47]. admet & dmpk 2(4) (2014) 199-220 prediction of drug skin permeation doi: 10.5599/admet.2.4.147 207 figure 1. lekc retention factors of the compounds tested for retention in cerasome lekc: data locations of 70 tested substances in the hansen space (perspective view and d/p-view). solid squares (with drop lines): log k7.4 > 0, solid triangles (with drop lines): log k7.4: 0 to -0.75, open squares: log k7.4 < -0.75, open circles: components of cerasome 9005. as was shown in a former work of the author, the contours of such irregularly shaped solubility regions are determined by the plurality of different molecular fragments of a large polymer molecule, which applies also to a supramolecular structure [42]. by contrast, small molecular compounds and also their mixtures are characterized by a single centred solubility sphere. in the latter case, according to hansen, the solubility of a substance in another can be estimated from the relative energy difference (red). the red is calculated from the euclidean distance in the three-dimensional hansen space between the data point of one substance and the centre of the other substance’s solubility sphere and is defined as the ratio of this distance and the sphere’s interaction radius r.        rred c h p h c p p p c d p d 5.0222 4   (16) a red lower than 1 indicates high solubility (coordinate position within the sphere), a red higher than 1 characterizes substances with low solubility (coordinate position outside the sphere). however, in case of cerasome the complex-shaped and even divided solubility region makes the classical hansen approach inapplicable. as it is nearly impossible to describe such irregular shapes by a single formula, it likewise renders difficult to find a simple mathematical correlation which is able to predict the solubility, or in this case the lekc retention factor, from hsps. for this reason, it is more expedient to use the full data set of known log k7.4 values for prediction instead of reducing it to an approximation formula with loss of valuable information. however, because of the irregular shape of the solubility region in the hansen space there is no predictable influence of more distant data points, so that only a local approximation is reasonable. the necessary basis for this method is a training set of a sufficient number of homogenously scattered data points without major gaps in between. the simplest way of prediction is by using a k-nearest neighbour algorithm, assigning the value of an object to the average value of its k nearest neighbors. as the dataset in tab. 1 unfortunately exhibits poorly covered regions with large distances to the closest data points, the k-nearest neighbor method can only be applied in a limited way, i.e. excluding the prediction of points which are too far from any point of the training set. in order to test the method without separate sets of training and sample data being available, log k7.4 of each compound in the list was predicted by using the data of all other compounds except the respective one as a trainings set. with k = 2 it could be shown that the method works with high accuracy for all data for which the distance to the nearest point of the training scheler et al. admet & dmpk 2(4) (2014) 199-220 208 set is smaller than 1 mpa 1/2 (average deviation of log k7.4: 0.20) (fig. 2). however, as can be seen by some outliers, in rare cases, where a data point has to be extrapolated in a region with steeply increasing response values, mispredictions cannot entirely be prevented. nevertheless, if a sufficiently large, dense and evenly distributed training dataset is available it can be expected that this model-independent method allows even more accurate predictions. as hsps can be easily calculated by simple group contribution methods, the presented approach allows the prediction of interactions of any organic substances with stratum corneum lipids directly from the molecular structure without the need of sophisticated computer programs. surprisingly, fig. 2 shows that neutral, anionic and cationic compounds are fitted equally well by the same regression line. hence, this example does not identify any necessity to consider charged compounds in a special way. figure 2. lekc retention factors (log k7.4) predicted from hsps (nearest neighbourhood method) vs. measured log k7.4values (compounds with hsp data points more distant than 1 mpa 1/2 to the nearest point of the model’s trainings set were excluded from prediction for lack of accuracy). squares: acids, diamonds: neutral compounds, triangles: bases. lfer analysis as described by zhang et al., however, considers the charge state of the tested compounds [11]. by regressing eq. (3) on the dataset of table 1, they developed the following formula for k7.4:   jjvbsk 912.1164.0782.1427.1587.0844.1log 4.7 (17) fig. 3. shows the values predicted by this linear relationship equation for the same compounds as in fig. 2. comparison of both diagrams reveals very similar results of both methods (r 2 hsp = 0.86, r 2 lfer = 0.90; slope of the linear trendline: hsp: 0.84, lfer: 0.90). however it should not be overlooked that the hsp method needs a more evenly distributed training dataset (or a better interpolation algorithm) in order to predict the whole range of sample compounds with sufficient accuracy. admet & dmpk 2(4) (2014) 199-220 prediction of drug skin permeation doi: 10.5599/admet.2.4.147 209 figure 3. lekc retention factors (log k7.4) of the same subset of compounds as in fig. 2 predicted with lfer. squares: acids, diamonds: neutral compounds, triangles: bases. description of octanol-water distribution coefficients in terms of hsps another more commonly used in-vitro-parameter which was also often applied for the prediction of skin permeability is the octanol-water partition coefficient p or, if all dissociation species at ph 7.4 are considered, the octanol-water distribution coefficient d7.4. however, compared to log k7.4, a different picture emerges if log d7.4 is displayed in the hansen diagram (fig. 4). data points representing high log d7.4 values form a widespread cloud. the hsp of octanol does not represent the center of this cluster, as one might expect from hansen theory. it has to be considered that, as log d7.4 does not describe the solubility in octanol but the distribution between octanol and water, this parameter is also a function of the compounds’ aqueous solubility. the center of the solubility sphere of water is reported d = 15.1 mpa 1/2 , p = 20.4 mpa 1/2 , h = 16.5 mpa 1/2 and the sphere’s interaction radius is given as 18.1 mpa 1/2 (based on the assumption of “good” solutes being soluble to more than 1 % in water) [48]. if the relative energy difference red of the neutral components is plotted against the octanol-water partition coefficient, a red value of 1 can be correlated with a log d7.4 of about 2 (data not shown), which means that compounds with hsps located outside the 1 %-water-solubility sphere are distributed less than 1 % within the waterand more than 99 % within the octanol-phase. the same is also valid if the red of acidic and basic compounds is plotted against log p, indicating that the neutral species of acids and bases partition in the same manner as nondissociating compounds. thus, it can be stated that log d7.4 and log p are substantially governed by the aqueous solubility of the compounds. this is in agreement to the known fact that for similar classes of compounds, excellent linear relationships exist between aqueous solubility cw and log p, as described by b c ap w          1 loglog (18) with a and b being constants [49,50]. it follows that cerasome retention factors provide a more differentiated model for partitioning of drugs into stratum corneum lipids than the octanol-water distribution coefficient. scheler et al. admet & dmpk 2(4) (2014) 199-220 210 figure 4. octanol-water distribution coefficients (log d7.4) of compounds tested for human stratum corneum penetration: data locations of 168 tested substances in the hansen space (perspective view and d/p-view). solid squares (with drop lines in the perspective view): log d7.4 > 2.5, solid triangles: log d7.4: 2.5 to 2, open squares: log d7.4 < 2, open circle: octanol. the perspective view shows also the solubility sphere of water. description of human skin permeability coefficients in terms of hsps though cerasome was developed to largely mimic the lipid composition in the stratum corneum, conditions in real skin are somewhat different from those in liposome dispersions. the differences do not only concern the structural arrangement of the lipid layers but also their composition. as mentioned above, cerasome contains lecithin which is not a component of the stratum corneum lipids. nevertheless, compared to other in-vitro test systems for skin permeation, the cerasome dispersion exhibits one of the closest analogies to human skin. using a database with human skin permeability coefficients (kp) of 168 different compounds, the correlation with lekc retention factors was analyzed. the permeability coefficients were compiled from different literature sources and averaged (tab. 2). fig. 5 shows the hsps of the analyzed compounds as a graphical representation in the hansen space with the permeability coefficients classified as different markers. figure 5. human skin permeability coefficients (log kp) of compounds tested for human stratum corneum penetration: data locations of 141 tested substances in the hansen space (perspective view and d/p -view). solid squares (with drop lines): log kp > -1.2, solid triangles (with drop lines): log kp: -1.2 to 1.32, open squares: log kp < -1.32, open circles: components of stratum corneum lipids and the lecithin component dppc. compared to the lekc data set, the compounds chosen for skin permeability analysis are spread over a more extended hsp range particularly towards lower d and higher p values. however, the distribution pattern of kp is largely the same as that of k7.4 in case of lekc data. again, there is one cluster of high kp admet & dmpk 2(4) (2014) 199-220 prediction of drug skin permeation doi: 10.5599/admet.2.4.147 211 values in the region of ceramides and fatty acids and another around the hsp of cholesterol. also in this case, the k-nearest neighbor method is able to predict the kp values from the hsps of the compounds. fig. 6 shows the correlation between predicted and measured kp values. as done before with k7.4 values, each value of the data set in tab. 2 was predicted on the basis of all other data, which were used as the training set. the predicted value was assigned to the average value of the three nearest neighbor points of the training set. because of the large size of the data set on which the method is based, the number of predictable data points did not have to be restricted. the average deviation of log kp was 0.51. though most of the data points are located along the 45° line through the origin, the slope of the regression line (0.6) deviates from 1 (r 2 = 0.62). this is mainly due to an underestimation of log kp values smaller than -4. the majority of these substances are high molecular weight compounds. as detailed below, for those compounds the low diffusion coefficient can be a biasing factor in prediction of kp. however, for most substances of a medium range molecular weight, this example demonstrates the performance of the method. the accuracy might be additionally increased by the application of a more sophisticated interpolation algorithm, like the modified shepards method described by renka [42,51,52], which, however, was not part of this work. figure 6: human skin permeability coefficients (log kp) predicted from hsps (nearest neighbourhood method) vs. measured log kp -values. squares: acids, diamonds: neutral compounds, triangles: bases. this data analyses demonstrate that lekc retention factors as well as human skin permeability coefficients are predictable from hansen solubility parameters. they show that compounds which are highly retained in lekc and those which are well permeable through human skin have similar hsps and thus comparable distribution patterns in the hansen space. this congruence is somewhat surprising as it is known, that skin permeability is not only governed by the water/stratum corneum partitioning but also by diffusion within the stratum corneum. hence, kp depends on both the water skin partition coefficient ksc and the diffusion coefficient in the stratum corneum dsc, as described by the following equation, h d kk sc scp logloglog  (19) in which h is the thickness of the stratum corneum [53]. as k7.4 quantifies a partitioning process into a lipid system similarly composed as the stratum corneum it is supposed to be a good estimate for ksc [53]. dsc was reported to be a function of the molecular volume vm,   mscsc vbdd  exp 0 (20) scheler et al. admet & dmpk 2(4) (2014) 199-220 212 with d 0 sc being the diffusivity of a molecule with a vanishing small vm and b a constant which is inversely proportional to the average free volume available for diffusion [6]. from eq. (20) a proportionality between log dsc and -vm can be deduced. it may therefore be concluded that log kp should be linearly related not only to log k7.4 but also to -vm. however, as follows from the very similar distribution pattern of lekc retention and skin permeation in the hansen space, the latter seems to be mainly governed by stratum corneum lipid partitioning and, for a wide range of substances, to be determined only to a minor degree by diffusion properties. conclusions summarizing these data, the distribution of substances into cerasome liposomes is primarily governed by specific affinities to the four main components of this system with the different solubility behaviour of cholesterol playing a special role. by contrast, the octanol-water distribution was found to be largely determined by a more unspecific displacement of hydrophobic compounds from the aqueous phase. the affinity of different compounds to cerasome can be portrayed using hansen solubility parameters. as cerasome resembles the stratum corneum lipids not only with respect to their composition but also concerning their lamellar arrangement, it constitutes an excellent in-vitro test system for skin partition. the high similarity to stratum corneum lipids is also expressed in a largely similar pattern of k7.4 and kp intensity distributions in the hansen space. in both systems a main reason for the complex relationship between skin permeability or lekc retention on the one hand and the different kinds of interactive forces on the other are the different solubility effects of cholesterol and the other lipids. references [1] n. dayan, cosmetics & toiletries 120 (2005) 67-76. [2] m. nino, g. calabrò, p. santoianni, dermatology online journal 16 (2010). [3] j.j. berti, j.j. lipsky, mayo clinic proceedings 70 (1995) 581-586. [4] y.g. anissimov, o.g. jepps, y. dancik, m.s. roberts, advanced drug delivery reviews 65 (2013) 169190. [5] k.b. sloan, s.a.m. koch, k.g. siver, f.p. flowers, journal of investigative dermatology 87 (1986) 244252. [6] r.o. potts, r.h. guy, pharmaceutical research 9 (1992) 663-669. [7] j.j. martínez-pla, y. martín-biosca, s. sagrado, r.m. villanueva-camañas, m.j. medina-hernández, biomedical chromatography 17 (2003) 530-537. [8] c.y. yang, s.j. cai, h. liu, c. pidgeon, advanced drug delivery reviews 23 (1997) 229-256. [9] y.j. wang, j. sun, h.z. liu, j.f. liu, l.q. zhang, k. liu, z.g. he, analyst 134 (2009) 267-272. [10] d.l. xian, k.l. huang, s.q. liu, j.y. xiao, chinese journal of chemistry 26 (2008) 671-676. [11] k. zhang, m. chen, g.k.e. scriba, m.h. abraham, a. fahr, x. liu, journal of pharmaceutical sciences 100 (2011) 3105-3113. [12] n.y. schürer, in handbook of atopic eczema, j. ring, b. przybilla, t. ruzicka, eds., springer verlag, berlin, germany, 2006, p. 157. [13] n.y. schurer, p.m. elias, advances in lipid research 24 (1991) 27-56. [14] n.y. schürer, g. plewig, p.m. elias, dermatologica 183 (1991) 77-94. [15] l. coderch, o. lópez, a. de la maza, j.l. parra, american journal of clinical dermatology 4 (2003) 107-129. [16] r.e. baynes, x.r. xia, v. vijay, j.e. riviere, sar and qsar in environmental research 19 (2008) 615630. [17] j. hildebrand, r.l. scott, the solubility of nonelectrolytes, third ed., reinhold, new york, ny, 1950. admet & dmpk 2(4) (2014) 199-220 prediction of drug skin permeation doi: 10.5599/admet.2.4.147 213 [18] c.m. hansen, hansen solubility parameters: a user’s handbook, crc, boca raton, fl, 1999. [19] s.h. neau, g.l. flynn, s.h. yalkowsky, international journal of pharmaceutics 49 (1989) 223-229 [20] a. amin, m. dare, a. sangamwar, a.k. bansal, pharmaceutical development and technology 17 (2012) 614-624. [21] c. bordes, v. fréville, e. ruffin, p. marote, j.y. gauvrit, s. briançon, p. lantéri, international journal of pharmaceutics 383 (2010) 236-243. [22] a. forster, j. hempenstall, i. tucker, t. rades, international journal of pharmaceutics 226 (2001) 147-161. [23] a.n. ghebremeskel, c. vemavarapu, m. lodaya, international journal of pharmaceutics 328 (2007) 119-129. [24] d.j. greenhalgh, a.c. williams, p. timmins, p. york, journal of pharmaceutical sciences 88 (1999) 1182-1190. [25] b.c. hancock, p. york, r.c. rowe, international journal of pharmaceutics 148 (1997) 1-21. [26] d. jain, k. pathak, aaps pharmscitech 11 (2010) 133-142. [27] k. kuksal, k. pathak, indian journal of pharmaceutical sciences 70 (2008) 609-613. [28] m.a. mohammad, a. alhalaweh, s.p. velaga, international journal of pharmaceutics 407 (2011) 6371. [29] m. shah, y. agrawal, pharmaceutical development and technology 18 (2013) 582-590 [30] k. vay, s. scheler, w. friess, international journal of pharmaceutics 416 (2011) 202-209. [31] p. bustamante, j. navarro-lupión, m.a. peña, b. escalera, international journal of pharmaceutics 414 (2011) 125-130. [32] j. breitkreutz, pharmaceutical research 15 (1998) 1370-1375. [33] l.g. martini, p. avontuur, a. george, r.j. willson, p.j. crowley, european journal of pharmaceutics and biopharmaceutics 48 (1999) 259-263. [34] s. abbott, international journal of cosmetic science 34 (2012) 217-222. [35] r. gröning, f.j. braun, pharmazie 51 (1996) 337-341. [36] t. hashiguchi, t. yasutake, t. manako, m. otagiri, international journal of pharmaceutics 158 (1997) 11-18. [37] e. squillante, t. needham, h. zia, international journal of pharmaceutics 159 (1997) 171-180. [38] p. bustamante, b. escalera, a. martin, e. selles, journal of pharmacy and pharmacology 45 (1993) 253-257. [39] a. beerbower, p.l. wu, a. martin, journal of pharmaceutical sciences 73 (1984) 179-188. [40] d.w. van krevelen, properties of polymers, elsevier, amsterdam, netherlands, p. 189. [41] a. jonquières, d., roizard, j. cuny, p. lochon, journal of membrane science 121 (1996) 117-133. [42] s. scheler, journal of applied polymer science 105 (2007) 3121-3131. [43] c.m. hansen, in hansen solubility parameters: a user’s handbook, c.m. hansen, ed., crc, boca raton, fl, 2007, p. 332. [44] a.f.m. barton, crc handbook of solubility parameters and other cohesion parameters, crc, boca raton, fl, 1991, p. 426. [45] c.m. hansen, t.s. poulsen, in hansen solubility parameters: a user’s handbook, c.m. hansen, ed., crc, boca raton, fl., 2007, p. 275. [46] m.r. ali, k.h. cheng, j. huang, biochemistry 45 (2006) 12629-12638. [47] m. london, e. london, journal of biological chemistry 279 (2004) 9997-10004. [48] c.m. hansen, in hansen solubility parameters: a user’s handbook, c.m. hansen, ed., crc, boca raton, fl, 2007, p. 21. [49] m. coates, d.w. connell, d.m. barron, environmental science & technology 19 (1985) 628-632. [50] s.c. valvani, s.h. yalkowsky, t.j. roseman, journal of pharmaceutical sciences 70 (1981) 502-507. scheler et al. admet & dmpk 2(4) (2014) 199-220 214 [51] d. douroumis, s. scheler, a. fahr, journal of pharmaceutical sciences 97 (2008) 907-918. [52] r.j. renka, acm transactions on mathematical software 14 (1988) 139-148. [53] k. zhang, m. chen, g.k.e. scriba, m.h. abraham, a. fahr, x. liu, journal of pharmaceutical sciences 101 (2012) 2034-2044. [54] o.a. santos-filho, a.j. hopfinger, t. zheng, molecular pharmaceutics 1 (2004) 466-476. [55] a. wilschut, w.f. ten berge, p.j. robinson, t.e. mckone, chemosphere 30 (1995) 1275-1296. [56] c. hansch, a. leo, substituent constants for correlation analysis in chemistry and biology, wiley, new york, ny, 1979. [57] g.l. flynn, in principles of route-to-route extrapolation for risk assessment, t.r. gerrity, c.j. henry, eds., elsevier, new york, 1990, p. 93. [58] m.d. barratt, toxicology in vitro 9 (1995) 27-37. [59] m.e. johnson, d. blankschtein, r. langer, journal of pharmaceutical sciences 86 (1997) 1162-1172. [60] w.j. pugh, i.t. degim, j. hadgraft, international journal of pharmaceutics 197 (2000) 203-211. [61] c. hansch, a. leo, d. hoekman, exploring qsar: hydrophobic, electronic, and steric constants, american chemical society, washington dc, 1995, usa, p. 183. [62] j.l. dwan`isa, m. dinguizli, v. preat, a. arien, m. brewster, journal of controlled release 101 (2005) 366-368. [63] p. bustamante, m.a. peña, j. barra, journal of pharmacy and pharmacology 50 (1998) 975-982. [64] m.a. peña, y. daali, j. barra, p. bustamante, chemical & pharmaceutical bulletin 48 (2000) 179-183. admet & dmpk 2(4) (2014) 199-220 prediction of drug skin permeation doi: 10.5599/admet.2.4.147 215 appendix table 1. lekc retention factors (log k7.4), octanol-water partition and distribution coefficients at ph 7.4 (log p and log d7.4), lfer solute descriptors, and hansen solubility parameters d, p, h) of 70 substances tested with cerasome electrokinetic chromatography [11] (zhang et al., 2011). one substance of the original dataset was omitted, because the hsp of iodinated compounds cannot be calculated by current group contribution methods. compound d mpa 1/2 p mpa 1/2 h mpa 1/2 charge state log poct log d7.4 log k7.4 e s a b v j + j cortexolone 18.8 4.6 12.5 neutral 2.52 2.52 -1.11 1.91 3.45 0.36 1.60 2.738 9 0 0 cortexone 18.4 4.2 9.2 neutral 2.88 2.88 -0.82 1.74 3.50 0.14 1.31 2.680 2 0 0 corticosterone 18.6 4.6 12.4 neutral 1.94 1.94 -1.27 1.86 3.43 0.40 1.63 2.738 9 0 0 cortisone 19.1 5.4 12.8 neutral 1.47 1.47 -1.57 1.96 3.50 0.36 1.87 2.754 6 0 0 dexamethasone 18.6 4.7 14.7 neutral 1.83 1.83 -1.40 2.04 3.51 0.71 1.92 2.913 2 0 0 digitoxin 18.7 3.0 14.6 neutral 1.86 1.86 -1.12 3.46 5.56 1.67 4.35 5.693 8 0 0 estriol 18.7 3.8 16.2 neutral 2.54 2.54 -1.37 1.97 1.74 1.06 1.63 2.257 5 0 0 hydrocortisone 19.0 4.9 15.1 neutral 1.55 1.55 -1.47 2.03 3.49 0.71 1.90 2.797 6 0 0 hydrocortisone-21-acetate 18.6 4.3 12.6 neutral 2.19 2.19 -1.41 1.89 2.88 0.46 2.16 3.095 1 0 0 17-hydroxyprogesterone 18.4 4.2 9.2 neutral 3.17 3.17 -0.90 1.64 3.35 0.25 1.31 2.680 2 0 0 prednisolone 19.0 5.1 15.3 neutral 1.62 1.62 -1.50 2.21 3.10 0.71 1.92 2.754 6 0 0 testosterone 18.3 3.6 9.3 neutral 3.29 3.29 -0.85 1.54 2.59 0.32 1.19 2.382 7 0 0 bibenzyl 18.6 0.6 0.0 neutral 4.80 4.80 0.72 1.22 1.04 0.00 0.33 1.606 0 0 0 4-chloro-2-methylphenol 20.3 6.5 13.3 neutral 2.78 2.78 -0.74 0.89 0.91 0.63 0.22 1.038 4 0 0 4-chloro-3,5-dimethylphenol 20.9 5.7 12.4 neutral 3.27 3.27 -0.51 0.93 0.96 0.64 0.21 1.179 3 0 0 3,4-dimethylphenol 17.0 4.2 12.9 neutral 2.23 2.23 -1.34 0.83 0.90 0.55 0.38 1.056 9 0 0 1-fluoro-2,4-dinitrobenzene 21.2 12.9 5.1 neutral 1.47 1.47 -1.63 1.01 1.69 0.00 0.45 1.082 5 0 0 2-naphthol 19.7 6.3 12.3 neutral 2.70 2.70 -0.70 1.52 1.08 0.61 0.40 1.144 1 0 0 resorcinol 18.0 8.4 21.0 neutral 0.80 0.79 -1.49 0.98 1.11 1.09 0.52 0.833 8 0 0 styrene 17.6 0.6 0.0 neutral 2.95 2.95 -0.62 0.85 0.65 0.00 0.16 0.955 2 0 0 toluene 17.5 1.0 0.0 neutral 2.73 2.73 -0.83 0.60 0.52 0.00 0.14 0.857 3 0 0 4-brc6h4oh 21.3 4.9 13.9 neutral 2.59 2.58 -0.87 1.08 1.17 0.67 0.20 0.950 1 0 0 3-ch3c6h4oh 18.0 5.1 12.9 neutral 1.96 1.96 -1.48 0.82 0.88 0.57 0.34 0.916 0 0 0 4-ch3c6h4oh 18.3 4.9 13.9 neutral 1.95 1.95 -1.43 0.82 0.87 0.57 0.31 0.916 0 0 0 c6h5coch3 19.6 8.6 3.7 neutral 1.58 1.58 -1.64 0.82 1.01 0.00 0.48 1.013 9 0 0 2-clc6h4nh2 19.3 5.4 9.2 neutral 1.91 1.91 -1.50 1.03 0.92 0.25 0.31 0.939 0 0 0 2-clc6h4no2 19.6 10.7 4.1 neutral 2.52 2.52 -1.27 1.02 1.24 0.00 0.24 1.013 0 0 0 3-clc6h4oh 19.3 7.5 14.3 neutral 2.50 2.49 -1.05 0.91 1.06 0.69 0.15 0.897 5 0 0 4-clc6h4oh 19.3 7.5 14.3 neutral 2.39 2.39 -1.01 0.92 1.08 0.67 0.20 0.897 5 0 0 4-clc6h4ch2oh 18.7 6.4 13.2 neutral 1.96 1.96 -1.36 0.91 0.96 0.40 0.50 1.038 4 0 0 2-h2nc6h4ph 19.0 1.0 7.3 neutral 2.84 2.84 -1.06 1.60 1.48 0.26 0.41 1.424 0 0 0 3-o2nc6h4oh 19.9 11.9 14.7 neutral 2.00 1.96 -0.83 1.05 1.57 0.79 0.23 0.949 3 0 0 phch2cn 19.5 12.3 3.8 neutral 1.56 1.56 -1.61 0.75 1.03 0.00 0.50 1.012 0 0 0 phch2oh 18.4 6.3 13.7 neutral 1.10 1.10 -1.66 0.80 0.87 0.39 0.56 0.916 0 0 0 phnh2 18.6 1.2 9.6 neutral 0.90 0.90 -1.73 0.96 0.96 0.26 0.41 0.816 2 0 0 phnhet 18.2 1.9 5.0 neutral 2.16 2.16 -1.52 0.95 0.85 0.17 0.43 1.098 0 0 0 phno2 20.0 8.6 4.1 neutral 1.85 1.85 -1.49 0.87 1.11 0.00 0.28 0.890 6 0 0 phoh 18.0 5.9 14.9 neutral 1.47 1.47 -1.56 0.81 0.89 0.60 0.30 0.775 1 0 0 acridine 19.6 5.4 5.8 neutral 3.40 3.39 -0.51 2.36 1.32 0.00 0.58 1.413 3 0 0 aspirin 18.7 4.7 11.0 anion 1.13 -2.79 -2.20 0.93 3.91 0.04 3.03 1.266 4 0 2.122 7 flurbiprofen 19.4 2.2 7.0 anion 3.81 0.32 -1.21 1.59 4.56 0.07 3.36 1.817 4 0 2.538 3 ibuprofen 16.4 7.7 7.2 anion 3.87 0.90 -1.19 0.88 3.50 0.08 3.31 1.755 6 0 2.418 8 ketoprofen 19.5 5.5 7.8 anion 2.77 -0.34 -1.32 1.80 5.49 0.01 3.39 1.956 4 0 2.485 1 mefenamic acid 20.3 2.5 8.1 anion 5.12 2.05 -1.17 1.80 4.71 0.09 3.14 1.899 6 0 2.642 7 scheler et al. admet & dmpk 2(4) (2014) 199-220 216 compound d mpa 1/2 p mpa 1/2 h mpa 1/2 charge state log poct log d7.4 log k7.4 e s a b v j + j naproxen 18.1 6.2 9.2 anion 3.06 -0.19 -1.43 1.66 5.07 0.02 3.11 1.760 6 0 2.426 0 4-brc6h4cooh 19.9 3.7 9.2 anion 2.86 -0.57 -1.16 1.15 3.47 0.04 2.61 1.085 2 0 2.250 4 1-c10h7cooh 20.5 3.2 8.6 anion 3.10 -0.61 -1.61 1.61 4.13 0.05 2.87 1.279 2 0 2.404 1 3-clc6h4cooh 18.9 5.9 9.3 anion 2.71 -0.86 -1.58 0.99 3.25 0.04 2.68 1.032 6 0 2.201 0 4-clc6h4cooh 19.0 5.9 9.4 anion 2.65 -0.77 -1.39 0.99 3.31 0.04 2.60 1.032 6 0 2.187 3 c6h5cooh 19.2 4.3 9.9 anion 1.96 -1.24 -1.16 0.88 3.05 0.02 2.75 0.910 2 0 2.138 5 c6h5(ch2)2cooh 18.8 3.3 8.7 anion 1.89 -1.26 -1.37 0.90 3.43 0.03 3.02 1.192 0 0 2.187 9 c6h5(ch2)3cooh 18.5 2.9 8.2 anion 2.42 -0.26 -1.67 0.91 3.59 0.04 3.01 1.332 9 0 2.218 4 c6h5(ch2)4cooh 18.3 2.6 7.8 anion 2.85 0.00 -1.63 0.92 3.63 0.04 3.10 1.471 8 0 2.279 4 c6h5(ch2)7cooh 17.8 2.0 6.8 anion 4.09 1.72 -1.31 0.94 3.87 0.07 3.26 1.896 5 0 2.425 6 4-mec6h4ch2nhme 17.2 1.6 4.6 cation 1.96 -0.57 -0.63 0.63 2.64 1.47 0.00 1.260 4 1.262 2 0 4-mec6h4ch2nhet 17.2 1.4 4.4 cation 2.38 -0.26 -0.63 0.61 2.69 1.48 0.00 1.401 3 1.264 7 0 4-mec6h4ch2nhpr 17.1 1.3 4.1 cation 2.96 0.38 -0.56 0.59 2.68 1.45 0.00 1.542 2 1.260 5 0 4-mec6h4ch2nhbu 17.0 1.2 4.0 cation 3.49 0.91 -0.44 0.57 2.68 1.46 0.00 1.683 1 1.240 5 0 4-mec6h4ch2nh(ch2)4me 17.0 1.1 3.8 cation 4.26 1.58 -0.08 0.55 2.66 1.41 0.00 1.824 0 1.252 2 0 4-mec6h4ch2nh(ch2)5me 16.9 1.0 3.7 cation 4.96 2.19 0.26 0.54 2.45 1.29 0.00 1.964 9 1.226 8 0 4-mec6h4ch2nh(ch2)6me 16.9 1.0 3.5 cation 5.12 2.50 0.95 0.53 2.51 1.47 0.00 2.105 8 1.121 5 0 acebutolol 18.0 4.3 10.5 cation 2.02 -0.10 -1.01 1.45 6.69 3.62 0.00 2.777 1 2.296 5 0 alprenolol 17.1 4.3 9.9 cation 3.10 0.91 0.06 1.10 4.46 1.78 0.00 2.180 2 2.257 4 0 metoprolol 17.7 7.0 10.1 cation 1.95 -0.28 -0.79 1.02 5.35 2.16 0.00 2.281 9 2.347 6 0 oxprenolol 17.1 4.8 10.8 cation 2.51 0.34 -0.50 1.16 5.09 2.35 0.00 2.238 9 2.202 9 0 penbutolol 17.3 2.4 9.6 cation 4.62 2.10 0.74 0.78 4.66 1.98 0.00 2.619 5 1.963 0 0 pindolol 18.7 3.3 11.6 cation 1.75 -0.39 -0.83 1.55 4.60 2.36 0.00 2.030 5 2.266 1 0 propafenone 17.9 3.3 9.5 cation 3.64 1.42 0.45 1.55 5.67 2.97 0.00 2.846 7 2.346 7 0 propranolol 19.2 7.9 10.0 cation 3.48 1.35 0.47 1.69 4.31 2.07 0.00 2.169 5 2.431 9 0 timolol 17.3 6.2 13.1 cation 1.83 0.01 -0.82 1.32 5.67 2.83 0.00 2.397 4 2.269 2 0 admet & dmpk 2(4) (2014) 199-220 prediction of drug skin permeation doi: 10.5599/admet.2.4.147 217 table 2. skin permeability coefficients (log kp), octanol-water distribution coefficients at ph 7.4 (log d7.4) and hansen solubility parameters d, p, h) of 168 substances. compounds which are undissociated to an extent of more than 80 % at ph 7.4 are referred to as neutral. those which are undissociated less than 20 % at ph 7.4 are referred to as anions or cations. all other compounds are partially dissociated (20-80 % undissociated) and labelled with “a/n” (anionic/neutral) or “c/n” (cationic/neutral). compound d p h ref. log kp (cm/h) ref. charge state log d7.4 ref. 1,1,1-trichloroethane 17.7 9.9 3.5 j -2.34 a,b neutral 2.49 b,c 1,3-dichloropropene 16.6 8.2 2.9 j -2.00 a neutral 1.82 o 17-hydroxyprogesterone 18.4 4.2 9.2 j -3.22 d,e,a neutral 2.74 d,e 2,3-butanediol 15.7 7.8 21.0 j -4.17 d,e,a,f,p neutral -0.92 d,e,c,f 2,4,6-trichlorophenol 19.6 8.7 13.1 j -1.23 d,e,a,f,b anion 2.70 o 2,4-dichlorophenol 21.3 8.3 13.6 j -1.22 d,e,a,f,b neutral 3.10 d,e,c,f,b 2-butoxyethanol 15.8 4.9 13.2 j -2.85 a neutral 0.80 i 2-chlorophenol 20.3 5.5 13.9 k -1.36 d,e,a,f,b,p neutral 2.16 d,e,c,f,b 2-cresol 18.2 4.9 13.9 j -1.68 d,e,a,f,b,p neutral 1.97 d,e,c,f,b 2-ethoxyethanol 16.1 9.2 14.3 k -3.48 d,e,a,f,b,p neutral -0.47 d,e,c,f,b 2-heptanone 16.2 5.7 4.1 k -2.00 a neutral 1.97 i 2-hexanone 15.3 6.1 4.1 k -2.35 a neutral 1.44 i 2-hydroxypropyl nicotinate 16.2 8.5 14.1 j -3.99 p neutral 0.27 i 2-methoxyethanol 16.2 9.2 16.4 k -3.73 a neutral -0.80 i 2-naphthol 19.7 6.3 12.3 j -1.47 d,e,a,f,b,p neutral 2.78 d,e,c,f,b 2-pentanone 15.4 7.1 4.3 j -2.60 a neutral 0.91 i 2-phenylethanol 18.2 4.3 12.9 j -1.71 a,b,f,p neutral 1.50 o 2-toluidine 19.4 5.8 9.4 k -1.44 a neutral 1.40 i 3,4-xylenol 17.0 4.2 12.9 j -1.32 d,e,a,f,b,p neutral 2.23 q 3-cresol 18.0 5.1 12.9 k -1.70 d,e,a,f,b,p neutral 1.96 q 3-nitrophenol 19.9 11.9 14.7 j -2.25 d,e,a,f,b neutral 1.96 q 3-xylene 17.3 0.9 0.0 j -1.10 a neutral 3.14 i 4-bromophenol 21.3 4.9 13.9 j -1.32 d,e,a,f,b,p neutral 2.59 d,e,c,f,b 4-chloro-3,5-xylenol 20.9 4.0 12.4 j -1.13 d,e,a,f,b,p neutral 3.39 d,e,c,f,b 4-chloro-m-cresol 20.3 6.5 13.3 j -0.78 p neutral 2.88 i 4-chloro-m-phenylenediamine 19.0 5.2 12.7 j -2.98 p neutral 0.84 i,o 4-chloro-o-cresol 20.3 6.5 13.3 j -1.25 d,e,a,f,b neutral 3.10 d,e,c,f,b 4-chlorophenol 19.3 7.5 14.3 j -1.32 d,e,a,f,b,p neutral 2.39 d,e,c,f,b 4-cresol 18.3 4.9 13.9 j -1.63 d,e,a,f,b,p neutral 1.94 d,e,c,f,b 4-ethylphenol 18.0 4.2 12.9 j -1.34 d,e,a,f,b,p neutral 2.46 d,e,c,f,b 4-(hydroxymethyl)phenol 19.3 7.0 19.8 j -2.70 p neutral 0.40 i 4-hydroxy-methylphenylacetate 20.5 5.0 15.1 j -1.70 p neutral -1.77 i 4-hydroxyphenylacetamide 19.9 5.8 17.1 j -3.33 p neutral 0.18 i 4-mec6h4ch2nh(ch2)4me 17.0 1.1 3.8 j -2.44 p cation 1.58 q 4-mec6h4ch2nh(ch2)5me 16.9 1.0 3.7 j -2.31 p cation 2.19 q 4-mec6h4ch2nh(ch2)6me 16.9 1.0 3.5 j -2.14 p cation 2.50 q 4-mec6h4ch2nhbu 17.0 1.2 4.0 j -3.16 p cation 0.91 q 4-mec6h4ch2nhet 17.2 1.4 4.4 j -3.41 p cation -0.26 q 4-mec6h4ch2nhme 17.2 1.6 4.6 j -4.15 p cation -0.57 q 4-mec6h4ch2nhpr 17.1 1.3 4.1 j -3.41 p cation 0.38 q 4-methyl-2-pentanol 15.1 4.0 12.6 j -2.33 a neutral 1.57 i 4-nitrophenol 19.9 11.9 14.7 j -2.25 d,e,a,f,b a/n 1.94 d,e,c,f,b 5,5-diethylbarbituric acid (barbital) 16.7 8.5 8.7 j -3.90 d,e,a,f,b,p a/n 0.69 i,o 5-ethyl-5-(3-methylbutyl) barbituric acid (amobarbital) 16.4 6.5 7.6 j -2.59 d,e,a,f,b,p a/n 2.06 i,o 5-ethyl-5-butylbarbituric acid (butobarbital) 16.6 7.0 7.9 j -3.65 d,e,a,f,p a/n 1.71 i,o 5-ethyl-5-phenylbarbituric acid (phenobarbital) 18.2 4.3 6.7 j -3.29 d,e,a,f,b,p a/n -3.11 i,o scheler et al. admet & dmpk 2(4) (2014) 199-220 218 compound d p h ref. log kp (cm/h) ref. charge state log d7.4 ref. 5-fluorouracil 19.6 13.3 11.0 j -3.26 p a/n -1.64 i 8-methoxypsoralen 21.9 6.6 8.3 j -1.56 p neutral 1.52 i acetic acid 14.5 8.0 13.5 k -3.21 a anion -2.90 o acrylic acid 16.7 6.2 12.1 j -3.05 a anion -2.90 o acrylonitrile 16.0 12.8 6.8 k -2.87 a neutral 0.19 i aldosterone 19.0 5.4 13.1 j -4.69 d,e,a,f,b,p neutral 1.08 d,e,c,f,b allyl alcohol 15.3 7.1 16.8 j -2.95 a neutral 0.17 i aniline 18.6 1.2 9.6 j -1.83 a,b,p neutral 0.91 b anisole 17.2 3.7 5.1 j -1.25 a,f,b,p neutral 2.11 c,f,b benzaldehyde 18.8 8.0 6.7 j -0.91 a,f,b,p neutral 1.48 c,f,b benzene 15.5 0.0 0.0 j -0.86 a,f,b,p neutral 2.12 c,f,b benzyl alcohol 18.4 6.3 13.7 k -1.88 d,e,a,f,b,p neutral 1.10 d,e,c,f,b,q benzyl nicotinate 18.3 3.0 7.5 j -1.31 p neutral 2.48 i butanoic acid 16.7 4.7 10.6 j -3.00 d,e,a anion -1.91 o butanone 15.3 8.4 4.7 j -2.18 d,e,a,f,p neutral 0.29 d,e,c,f butyl acrylate 15.6 3.4 7.0 j -2.00 a neutral 2.39 i butyl nicotinate 16.7 3.1 7.7 j -1.30 p neutral 2.32 i c6h5(ch2)2cooh 18.8 3.3 8.7 j -1.57 p neutral -1.26 q c6h5(ch2)3cooh 18.5 2.9 8.2 j -1.49 p neutral -0.26 q c6h5(ch2)4cooh 18.3 2.6 7.8 j -0.94 p neutral 0.00 q c6h5(ch2)7cooh 17.8 2.0 6.8 j -0.50 p neutral 1.72 q c6h5cooh 19.2 4.3 9.9 j -1.55 p anion -1.24 q caffeine 20.8 13.3 12.9 j -3.87 a,f,p neutral -0.05 c,f catechol 20.0 11.3 21.8 k -2.77 a neutral 0.88 i chlorpheniramine 17.3 5.1 6.5 j -2.66 d,e,a,b cation 1.04 o cortexolone 18.8 4.6 12.5 j -4.13 d,e,a,b neutral 2.52 d,e,b cortexone 18.4 4.2 9.2 j -3.35 d,e,a,b neutral 2.88 d,e,b corticosterone 18.6 4.6 12.4 j -3.75 d,e,a,f,b,p neutral 1.94 d,e,c,f,b cyclohexanone 15.9 7.5 4.4 j -2.74 a neutral 0.76 i dexamethasone 18.6 4.7 14.7 j -4.03 a,f,p neutral 1.91 c,f diclofenac 21.3 4.5 8.2 j -2.31 a,f,b anion 1.44 o diethyl ether 14.7 2.8 5.5 j -2.05 d,e,a,f,b,p neutral 0.83 d,e,f,b diethylamine 14.9 1.4 5.5 j -2.75 a cation -2.25 o diethylcarbamazine 16.4 8.1 9.3 j -3.46 d,e,a,b,p c/n 0.08 o digitoxin 18.7 3.0 14.6 j -4.79 d,e,a,b,p neutral 2.44 i dimethyl acetamide 16.8 11.5 10.2 k -2.80 a neutral 0.21 o dimethyl formamide 17.4 13.7 11.3 k -1.98 a neutral -0.83 o dimethyl sulfoxide 18.4 16.4 10.2 k -1.80 a neutral -1.35 i ephedrine 17.2 3.4 11.9 j -2.22 d,e,a,f,b cation -0.93 o estradiol 18.4 3.0 13.1 j -2.45 d,e,a,f,b,p neutral 2.91 d,e,f,b estriol 18.7 3.8 16.2 j -4.40 d,e,a,b neutral 2.47 d,e,b estrone 19.6 4.0 9.7 j -2.44 d,e,a,b neutral 2.76 d,e,b ethanol 15.8 8.8 19.4 k -2.96 d,e,a,f,b,p neutral -0.31 d,e,c,f,b ethyl acrylate 15.5 7.1 5.5 k -2.39 a neutral 1.32 i ethyl benzene 17.3 0.9 0.0 j 0.62 d,e,a,f,p neutral 3.15 d,e,c,f ethyl formate 13.2 10.6 7.8 j -3.01 a neutral 0.30 i ethyl nicotinate 16.8 3.9 8.6 j -1.72 p neutral 1.30 i ethylene dichloride 17.1 6.5 3.1 j -2.00 a neutral 1.41 i ethylene glycol 17.0 11.0 26.0 k -4.07 a neutral -1.69 i etorphine 19.8 3.5 9.9 j -2.44 d,e,a,b cation 1.92 o fentanyl 18.7 4.5 6.2 j -2.26 d,e,a,f,b cation 2.61 o fluocinonide 18.1 4.3 9.8 j -2.77 d,e,a,f,b,p neutral 3.19 d,e,c,f,b formaldehyde 10.7 18.3 10.1 j -2.65 a neutral 0.35 o glycerol trinitrate 17.8 14.6 10.0 j -1.65 p neutral 2.15 i heptanoic acid 16.6 3.0 8.5 j -1.70 d,e,a anion -0.36 o admet & dmpk 2(4) (2014) 199-220 prediction of drug skin permeation doi: 10.5599/admet.2.4.147 219 compound d p h ref. log kp (cm/h) ref. charge state log d7.4 ref. hexachloroethane 19.7 7.3 4.3 j -1.40 a neutral 4.47 i hexanoic acid 16.6 3.4 9.0 j -1.85 d,e,a anion -0.87 o hexyl nicotinate 16.6 2.6 7.1 j -1.27 p neutral 3.34 i hydrocortisone 19.0 4.9 15.1 j,l -4.43 d,e,a,f,b,p neutral 1.56 d,e,c,f,b hydrocortisone 21-(6hydroxy)hexanoate 18.1 3.6 11.6 j -2.92 d,e,a,f,b,p neutral 2.79 d,e,c,f,b hydrocortisone 21-(n,ndimethyl)succinamate 18.1 4.4 9.9 j -4.17 d,e,a,f,b,p neutral 2.03 d,e,c,f,b hydrocortisone 21-hemipimelate 18.3 3.5 10.2 j -2.73 d,e,a,f,b,p anion 0.25 o hydrocortisone 21hemisuccinate 18.5 3.9 10.9 j -3.18 d,e,a,f,b,p anion -0.47 o hydrocortisone 21-hexanoate 17.8 3.4 9.0 j -1.74 d,e,a,f,b,p neutral 4.04 d,e,c,f,b hydrocortisone 21methylpimelate 15.9 3.3 9.5 j -2.27 d,e,a,f,b,p neutral 3.50 d,e,c,f,b hydrocortisone 21methylsuccinate 18.0 3.7 10.1 j -3.68 d,e,a,f,b,p neutral 3.58 d,e,c,f,b hydrocortisone 21-octanoate 17.7 3.1 8.6 j -1.21 d,e,a,f,b,p neutral 5.49 d,e,c,f,b hydrocortisone 21-pimelamate 18.1 3.8 10.2 j -2.93 d,e,a,f,b,p neutral 2.31 d,e,c,f,b hydrocortisone 21-propionate 17.9 3.9 9.6 j -2.47 d,e,a,f,b,p neutral 3.00 d,e,c,f,b hydrocortisone 21-succinamate 18.3 4.3 10.8 j -4.59 d,e,a,f,b,p neutral 1.43 d,e,c,f,b hydroxypregnenolone 18.4 3.6 12.1 j -3.22 d,e,a,b neutral 3.00 d,e,b indometacin 19.9 5.1 8.7 j -2.44 a,f,b anion 0.98 o isoamyl alcohol 15.3 4.6 13.5 j -2.00 a neutral 1.22 i isobutanol 15.1 5.7 15.9 k -2.65 a neutral 0.68 o isoquinoline 18.9 6.6 6.5 j -1.72 d,e,a,f,b,p neutral 2.07 d,e,c,f,b lidocaine 18.5 5.0 6.7 j -2.09 a,f,b c/n 1.67 o mannitol 18.6 10.7 32.4 j -4.86 p neutral -3.26 i meperidine 16.9 4.0 7.2 j -2.43 d,e,a,b cation 1.62 o methacrylic acid 16.9 5.0 10.9 j -2.58 a anion -2.20 o methanol 15.1 12.3 22.3 k -3.14 d,e,a,f,b,p neutral -0.74 d,e methyl 4-hydroxybenzoate 18.2 5.6 14.7 j -1.92 d,e,a,f,b,p neutral 1.96 d,e,c,f,b methyl acrylate 15.3 6.7 9.4 k -2.68 a neutral 0.79 i methyl nicotinate 16.8 4.4 9.2 j -2.21 p neutral 0.88 i methylene chloride 17.8 6.4 6.1 k -2.74 a neutral 1.53 i methyltriglycol nicotinate 16.6 3.7 8.9 j -3.27 p neutral monomethylhydrazine 16.2 8.7 14.8 k -3.75 a cation -2.82 o n,n-dimethylaniline 18.0 6.3 6.3 j -1.70 a neutral 2.33 i naproxen 18.1 6.2 9.2 j,l,m -2.84 d,e,a,f,b anion 0.35 o n-butanol 16.0 5.7 15.8 k -2.52 d,e,a,f,b,p neutral 0.88 d,e,c,f,b n-decanol 16.0 2.6 10.2 j -0.93 d,e,a,f,b,p neutral 4.20 d,e,c,f n-heptanol 15.9 3.5 11.9 j -1.41 d,e,a,f,b,p neutral 2.62 d,e,c,f,b n-hexanol 15.8 4.0 12.6 j -1.71 d,e,a,f,b,p neutral 2.03 d,e,c,f,b nicotine 20.5 6.6 10.7 j -1.72 d,e,a,f,b cation -0.62 o nitroglycerine 16.2 17.8 5.9 k -1.96 d,e,a,b neutral 1.87 d,e,c,b n-methyl-2-pyrrolidone 18.0 11.5 8.5 j -1.80 a neutral -0.40 i n-nonanol 16.0 2.9 10.7 j -1.10 d,e,a,f,b,p neutral 3.80 d,e,f,b n-octanol 15.9 3.2 11.2 j -1.11 d,e,a,f,b,p neutral 3.07 d,e,c,f,b n-pentanol 15.7 4.6 13.6 j -2.10 d,e,a,f,b,p neutral 1.51 d,e,c,f,b n-propanol 16.0 6.8 17.4 k -2.77 d,e,a,f,b,p neutral 0.28 d,e,c,f,b octanoic acid 15.1 3.3 8.2 k -1.60 a anion 0.15 o o-phenylenediamine 19.5 0.8 13.4 j -3.14 p neutral 0.05 i ouabain 18.9 4.2 21.2 j -6.11 a,b,p neutral -2.85 c,b pentanoic acid 16.7 4.0 9.7 j -2.70 a anion -1.39 o phenol 18.0 5.9 14.9 k -1.93 d,e,a,f,b,p neutral 1.46 d,e,c,f,b phenylglycidyl ether 18.2 4.3 6.7 j -2.84 a neutral 1.74 i scheler et al. admet & dmpk 2(4) (2014) 199-220 220 compound d p h ref. log kp (cm/h) ref. charge state log d7.4 ref. piroxicamc 18.9 7.1 13.6 j -2.46 p neutral 1.68 i p-phenylenediamine 19.5 0.8 13.4 j -3.42 p neutral -0.71 i pregnenolone 18.0 3.2 8.7 j -2.82 d,e,a neutral 3.13 d,e,b progesterone 18.0 3.8 3.7 j -2.08 d,e,a,f,b,p neutral 3.84 d,e,c,f,b propionic acid 14.7 5.3 12.4 k -2.94 a anion -2.39 o propylene oxide 15.2 8.6 6.7 k -3.05 a neutral 0.13 i resorcinol 18.0 8.4 21.0 k -3.50 d,e,a,f,b,p neutral 0.80 d,e,c,f,b salicylic acid 19.4 10.1 17.4 k -2.32 d,e,a,f,b anion -1.14 o scopolamine 19.2 4.9 12.3 j -4.30 d,e,a,f,b c/n 0.05 o styrene 17.6 0.6 0.0 j -0.19 d,e,a,f neutral 2.95 d,e,c,f sucrose 17.7 7.5 27.8 j -5.28 d,e,a,b neutral -2.74 d,e,c,b testosterone 18.3 3.6 9.3 j -2.74 d,e,a,f,b,p neutral 3.31 d,e,c,f,b thymol 16.5 3.2 11.4 j -1.15 d,e,a,f,p neutral 3.33 d,e,c,f,b toluene 17.5 1.0 0.0 j -0.04 d,e,a,p neutral 2.74 d,e,c triglycol nicotinate 16.4 5.6 13.3 j -4.52 p neutral urea 20.9 18.7 26.4 k -4.37 p neutral -1.66 i vinyl acetate 16.0 7.2 5.9 k -2.73 a neutral 0.73 i a) reference [54] (santos-filho et al., 2004), b) reference [55] (wilschut et al., 1995), c) reference [56] (hansch and leo, 1979), d) reference [57] (flynn, 1990), e) reference [58] (barratt, 1995), f) reference [59] (johnson et al., 1997), g) reference [60] (pugh et al., 2000), h) reference [61] (hansch et al., 1995), i) calculated with the software acd/chemsketch freeware (version 10.00, advanced chemistry development, inc., toronto, on, canada, www.acdlabs.com, 2006), j) calculated by the group contribution method accordingt to hoftyzer, van krevelen [40] (van krevelen, 1990), k) reference [18] (hansen, 2007a), l) reference [62] (dwan`isa, 2005), m) reference [63] (bustamante et al., 1998), n) reference [64] (peña et al., 2000), o) chemspider (royal society of chemistry, london, uk) (www.chemspider.com), p) reference [53] zhang et al., 2012), q) reference [11] (zhang et al., 2011) ©2014 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ bacteriocin-mediated inhibition of some common pathogens by wild and mutant lactobacillus species and in vitro amplification of bacteriocin encoding genes doi: http://dx.doi.org/10.5599/admet.1053 75 admet & dmpk 10(1) (2021) 75-87; doi: https://doi.org/10.5599/admet.1053 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper bacteriocin-mediated inhibition of some common pathogens by wild and mutant lactobacillus species and in vitro amplification of bacteriocin encoding genes arooj ahsan 1 , bushra mazhar 1 , muhammad kamran khan 1 , madiha mustafa 2 , muhammad hammad 3 , nazish mazhar ali 1 * 1 department of zoology, gcu lahore, pakistan 2 qarshi university lahore, pakistan 3 university of education, lahore, pakistan *corresponding author: e-mail: nazipak@hotmail.com. received: july 08, 2021; revised: october 14, 2021; available online: november 14, 2021 abstract lactobacilli are the most common probiotics used in food and other industries because of their capability of producing bacteriocins. bacteriocins are compounds that are used to kill pathogenic microorganisms. as most bacteria have become resistant to synthetic antibacterial tools, the importance of using probiotics as antibacterial agents has increased. this work was done to check the bacteriocin effect on some common pathogens and the influence of mutation on the bacteriocin activity of lactobacilli was also investigated. four strains were isolated, identified from meat and pickles samples via culturing methods, staining, biochemical tests, and ribotyping. preliminary tests, including gram staining and catalase test, were done for the confirmation of lactobacillus species. all strains were gram-positive and catalase-negative. antibacterial activity was checked against pseudomonas aeruginosa, staphylococcus aureus, bacillus thuringiensis, escherichia coli, and salmonella enteritis via agar well diffusion method. the mutations were done using ethidium bromide and the influence of wild and mutants were also checked. interestingly, mutants developed more virulence than wild ones. it was also observed that they all were sensitive to pepsin. protein estimation was done via bradford method. ribotyping of gcu-w-ps1 revealed 99 % homology with lactobacillus plantarum and gcu-w-ms1 to lactobacillus curvatus (99 % homology). curvacin a, sakacin p, and plantaricin a genes were also amplified using specific primers. gene sequence showed the presence of curvacin a gene in gcu-w-ms1. it was concluded that lactic acid bacteria could be used as antibacterial tools against common pathogens. ©2021 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords bacteriocin; antibacterial activity; lactobacilli; bacteriocin encoding genes. introduction bacteriocins are compounds produced by different bacterial strains and are ribosomally synthesized antimicrobial proteins. their function is to inhibit the growth of similar or closely related bacterial strains [1]. many lactic acid bacteria are found infrequently fermented or non-fermented foods, producing a high http://dx.doi.org/10.5599/admet.1053 https://doi.org/10.5599/admet.1053 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:nazipak@hotmail.com http://creativecommons.org/licenses/by/4.0/ nazish mazhar ali et al. admet & dmpk 10(1) (2022) 75-87 76 variety of bacteriocins. some of the bacteriocins even have the potential to preserve food, which helps foods preserve nature and not only preserve but also restore the nutritional properties of food. in industries, we preserve food by using chemicals, though it is successful, the quality of food becomes doubtful. so for this purpose, the alternative technique that could be used is applied bacteriocin and act as bio preservatives [2]. nisin is the only bacteriocin that is broadly used for food preservation now a day. it is the most extensively implemented bacteriocin. it is vigorous against highly pathogenic and food spoilage microorganisms, including s. aureus and listeria monocytogenes [3]. but scientists are doing work on many bacteriocins to characterize them biochemically, chemically, and genetically so that we will be able to understand their basic mode of action. as a result of which we will revolutionize the food industry [4]. by inducing mutation in lactobacillus we can improve the potency at elevated temperatures, also broadening the antibacterial spectrum [5]. mutation is a method of changing organisms genetically. either this mutation will be beneficial for us or may make the organism more vulnerable for mankind. mutagenesis can be done with ethidium bromide [6] and ultraviolet (uv) light [7]. ethidium bromide is a frame-shift mutagen genotoxic and teratogen which usually become the cause of changes like frame-shift mutation, chromosomal recombination, arrested cell division and developmental problem. however, it has been noted that such types of mutations have positive effects and usually increase the efficiency of mutants in contrast to wild ones. mutagenic activities can be optimized by knowing the type of mutagens and dosage. mutagens specificity effects may also help in the optimization. mutagenesis can be improved or directed to achieve the maximum frequency of enviable mutant products [8]. the objectives of the current work were to isolate, identify, and characterize the lactobacillus species from meat and pickle samples, to check the virulent effect of wild and mutant species of lactobacillus against common pathogenic bacteria, and to detect bacteriocin encoding genes in different wild and mutant species of lactobacilli species. importance of work bacteria develop resistance against many antibacterial agents currently used, so it is a dire need to develop new strategies to control these pathogenic bacteria. lactobacilli are used as probiotics and are now extensively studied for their antibacterial activity against common pathogens. these bacteria produce bacteriocin, which has an impact on pathogens. we studied and detected the bacteriocin encoding genes in our native lactobacilli (wild and mutant), did ribotyping, and obtained their accession number from ncbi. this is original research work and our study with native bacterial isolates. materials and methods sample collection poultry meat samples and a variety of pickles samples (homemade and companies made) were collected in sterilized bags from different shops of sant nagar and krishan nagar, lahore, pakistan. preparation of culture media for the isolation of lactobacillus species de man, rogosa and sharpe (mrs) broth (sigma) (1.5 g of peptone, 1.5 g of beef extract, 0.75 g of yeast extract, 3 g of glucose, 0.15 g of tween-80, 0.3 g of ammonium citrate, 0.75 g of sodium acetate, 0.03 g of magnesium sulfate, 0.0075 g of manganese sulfate and 0.3 g of dipotassium hydrogen and 150 ml of distilled water) and mrs agar medium (sigma)(1.5 g of peptone, 0.75 g of yeast extract, 1.5 g of beef extract, 3 g of glucose, 0.15 g of tween-80, 0.75 g of sodium acetate, 0.3 g of ammonium citrate, 0.03 g of magnesium sulfate, 0.0075 g of manganese sulfate, 0.3 g of dipotassium hydrogen phosphate, 2.25 g of agar, and 150 ml of distilled water) were prepared with neutral ph. admet & dmpk 10(1) (2022) 75-87 bacteriocin-mediated inhibition of pathogens doi: http://dx.doi.org/10.5599/admet.1053 77 isolation of bacteria via aseptic techniques one gram of each sample was taken and mixed with 9 ml of buffered peptone water in the flask. the content was homogenized. the samples were serially diluted and two methods, i.e., spreading and streaking, were used to isolate bacteria on prepared mrs agar medium. morphological and biochemical characterization pure cultures were characterized by cell morphology via gram staining and biochemical characterization via catalase reaction. catalase-negative isolates were further analyzed for bacteriocin production. agar-well diffusion assay the antagonism of bacteriocin-producing strains against responsive strains (gram-positive and gramnegative bacteria) is usually determined using the agar well diffusion method [9]. potential bacteriocin producers were grown in mrs broth at 37 °c for 24 h. the size of disposable petri plates was 60 mm x 15 mm. responsive strains were streaked on the agar plates, dried for a few minutes, five wells of 8 mm were made, and poured with 50 µl of an overnight culture of isolated strains. the agar plates were incubated for 30 min at 4 o c to allow the diffusion of bacteriocin and incubated overnight at 37 °c. zones of inhibitions were checked after 24 h of incubation. the diameter of inhibited zones was measured [10]. induction of mutagenesis mutagenesis with ethidium bromide was done [11]. strains of lactobacilli (i.e., gcu-w-ps1, gcu-w-ps2, gcu-w-ms1, and gcu-w-ms2) were grown up to the late logarithmic phase of growth in trypticase soy broth (tsb). the cells were yielded and washed with 0.9 % nacl solution twice and ethidium bromide with a concentration of 0.5, 1.0, 1.5 g/l were added to 2ml of each cell suspension. the mixture was aerated on a shaking incubator at 30 o c and 42 o c, respectively. the treated cells were incubated in 10ml tsb washed twice and re-suspended in 0.9 % nacl solution and after serial dilution, it was spread on mrs agar plates and incubated at 30 o c and 42 o c for 48 h. then isolation of mutant was done. production of bacteriocins 10 ml of mrs broth was prepared in falcon tubes for each isolate and ph was adjusted to 6.0 to 7.0. falcon’s tubes were autoclaved at 121 °c and 15 lbs. then 100 µl of a freshly prepared broth culture of lactic acid bacteria was inoculated with a micropipette and incubated for 24 hours at 37 °c. after incubation, falcon tubes were placed in a centrifuge (biorad) for 30 minutes at 8000 rpm at 4 °c. with the help of filter paper, the supernatant was filtered. this filtered supernatant was crude extract. protein estimation by bradford method bradford method [12] was used to determine protein concentrations using a spectrophotometer (biorad). bovine serum albumin (bsa) was taken as standard. bsa's stock solution may be prepared if we dissolve 1mg of bsa in 1ml of distilled water. different concentration of bovine serum albumin was prepared for the preparation of the standard curve of bsa. 100 µl of each dilution of bsa was taken in test tubes and labeled according to concentration. 100 µl of distilled water was taken in the test tube as blank. then 5 ml of bradford reagent was added to each test tube mixed and placed at room temperature for 15 minutes. after a few minutes, the sample color changed to blue then the absorbance of the tubes was noted at 595 nm. graph for the standard curve was plotted, taking concentration along the x-axis and absorbance along the y-axis. determination of protein concentration http://dx.doi.org/10.5599/admet.1053 nazish mazhar ali et al. admet & dmpk 10(1) (2022) 75-87 78 for preparing the bradford reagent, we dissolved 100 mg coomassie brilliant blue g-250 in 50 ml 95 % ethanol; added 100 ml 85 % (w/v) with phosphoric acid. then diluted it to 1 liter. when the dye had completely dissolved, it was filtered through whatman #1 paper just before use and stored at 4 °c. to determine the protein concentration of samples, test tubes were labeled according to the sample name and 100 µl of protein sample in the form of cfs was taken while the control tube contained only a growth medium (mrs broth). 5 ml of bradford reagent was added and mixed in each test tube and then incubated at room temperature until it turned blue. the absorbance of the tubes was taken at 595nm by spectrophotometer. the absorbance was compared with standard. characterization of bacteriocin to confirm the proteinaceous nature of bacteriocin, its sensitivity against proteolytic enzymes was tested. proteolytic enzyme with final concentration of 2mg/ml was treated with crude bacteriocin in 0.01 m phosphate buffer at ph 7.0, 0.15 ml (150 µl), phosphate buffer (0.5m, ph 7.0), 0.15 ml (150µl) of bacteriocin and 0.15 ml (150 µl) of pepsin (0.25mg/ml) were taken in a test tube. a control was run by taking only 0.15 ml (150 µl) of bacteriocin and 0.15 ml (150 µl) of distilled water. these samples were kept in an incubator for 2 hours at 37 °c and then boiled at 100 °c for 3 to 5 minutes. no zone of inhibition indicates a positive result. molecular characterization lactobacillus cells from 24 h mrs broth cultures were centrifuged at 6000 rpm for 5 min to pellet out cells for dna isolation using phenol: chloroform extraction method. selected bacteriocin-producing strains were finally identified along the partial length of the 16s rrna gene sequence. amplification with 16s rrna gene is used to identify the microbes to specie level. 50 µl reaction mixtures were used for pcr amplification using a thermocycler (biorad1000). the polymerase chain reaction was performed for 35 reaction cycles in a thermocycler. all amplification reactions were performed in a pcr cycling system using the conditions given below: the first cycle was led by initial denaturation at 94 °c for 5 min, followed by 35 cycles, denaturation at 94 °c for 0.50 sec, annealing gradient ranging from 49.3 °c to 60.2 °c for 0.30 sec and polymerization at 72 °c for 50 sec. the reactions were terminated with 10 min of elongation at 72 °c and then chilled to 4 °c. sequencing and sequence analysis pcr product was run on 1 % agarose gel. bands in the gel were observed in uv (biorad). the products of pcr were sent to 1 st base laboratory, malaysia, for sequencing. the sequences acquired were sent to the national centre for biotechnology information (ncbi) to analyze the nucleotide-nucleotide blast data. characterization of the bacteriocin structural gene various structurally different bacteriocins have been described to date. we performed pcr amplification experiments to determine whether the selected strains carry structural genes of known species that were the same as those of other species. in this experiment, we used specific primers for the detection and identification of specific bacteriocin genes. for pcr amplification of the specific gene, specific primers were designed based on gene sequences available in the genbank database. amplification of bacteriocin encoding genes by performing the polymerase chain reaction, bacteriocin encoding genes of the chosen isolates were identified using different specific primer pairs, the primers for sakacin p, and bacterial genomic dna. after the pcr, amplified products were analyzed by electrophoresis on 1 % agarose gel (1x tae buffer ph 8.3) admet & dmpk 10(1) (2022) 75-87 bacteriocin-mediated inhibition of pathogens doi: http://dx.doi.org/10.5599/admet.1053 79 using a 100 kb dna ladder (enzynomics) as a molecular weight standard weight. electrophoresis was done at 80 v for 60 min. analysis of bacteriocin gene sequence the products of pcr were sent for sequencing to 1 st base laboratory, malaysia. nucleotide sequencing of the pcr amplicon was compared with the genbank database. sequences were blasted against the ncbi genbank database. sequence alignment was performed. the gene sequences related to bacteriocin production were deposited in the ncbi, genbank. results isolation of lactic acid bacteria fourteen bacterial strains were isolated, eight strains of lactobacillus were isolated from four samples of minced meat (p1, p2, p3, p4, p5, p6, gcu-w-ms1, gcu-w-ms2), and six strains of lactobacillus isolated from different sources of pickles (h1, gcu-w-ps1, h3, n1, gcu-w-ps2, n3) (table 1). table 1. isolated lactobacilli strains from poultry meat and pickle sample nature of sample isolated strains minced poultry meat (sant nagar) p1, p2 minced poultry meat (sant nagar) p3, p4 minced poultry meat (krishan nagar) p5, gcu-w-ms1, minced poultry meat (krishan nagar) p6, gcu-w_ms2 pickle (mixed) home made h1 ,gcu-w-ps1 pickle (mixed) shan company h3, pickle (mixed) shangrilla company gcu-w-ps2 pickle (mixed) shezan company gcu-w-ps2, n3 morphological and biochemical characteristics grams staining result showed that all the isolated strains were purple-stained and were gram-positive and were rod or cocci in shape (table 2 and figure 1). the results revealed that all isolated bacterial strains could not degrade hydrogen peroxide and were catalase-negative (figure 2). figure 1. gram staining test showing results of selected strains as a gram-positive http://dx.doi.org/10.5599/admet.1053 nazish mazhar ali et al. admet & dmpk 10(1) (2022) 75-87 80 table 2. morphological and biochemical characterization of bacterial isolates strains colony color colony size colony texture gram staining catalase test p1 white medium wet +ive -ive p2 white medium wet +ive -ive p3 offwhite small wet +ive -ive p4 offwhite medium wet +ive -ive p5 white small wet +ive -ive p6 white small wet +ive -ive gcu-w-ms1 offwhite medium dry +ive -ive gcu-w-ms2 offwhite small wet +ive -ive h1 white small wet +ive -ive gcu-w-ps1 white small wet +ive -ive h3 offwhite medium wet +ive -ive n1 offwhite medium wet +ive -ive gcu-w-ps2 offwhite medium wet +ive -ive n3 white small wet +ive -ive figure 2. catalase test showing negative results for selected strains agar well diffusion assay (inhibitory assay) the agar well diffusion method revealed that gcu-w-ms1, ms2, gcu-w-ps1, and gcu-w-ps2 were found an active producer of bacteriocins (table 3). gcu-w-ms1 showed maximum inhibition of escherichia coli (9 mm). on the other hand, the least or even minimum inhibition of salmonella enteritidis, s. aureus, p. aeruginosa, and b. thuringiensis was recorded. table 3. inhibitory activity of bacteriocin of wild strains of lactobacillus isolated from poultry meat and pickles against some pathogenic bacteria sample no. zone of inhibition (mm) salmonella enteritidis p. aeruginosa escherichia coli s. aureus b. thuringiensis m ± s.e gcu-w-ps2 5.1 ± 00.10 2.0 ± 00.10 3.0 ± 00.10 7.9 ± 00.10 2.0 ± 00.20 gcu-w-ps1 6.0 ± 00.10 resistant 2.7 ± 00.10 7.5 ± 00.20 3.5 ± 00.10 gcu-w-ms1 8.2 ± 00.20 2.1 ± 00.13 8 ± 00.15 8.0 ± 00.10 2.7 ± 00.13 gcu-w-ms2 7.8 ± 00.10 3.3 ± 00.10 5.4 ± 00.10 9.2 ± 00.14 2.4 ± 00.10 gcu-w-ms2 showed maximum inhibition of s. aureus and salmonella enteritidis with 8 mm zone of inhibition, while escherichia coli, p. aeruginosa, b. thuringiensis were resistant to l. johnsonii. gcu-w-ps1 showed maximum inhibition of s. aureus and salmonella enteritidis with 7 mm zone of inhibition, while resistance to p. aeruginosa, escherichia coli and b. thuringiensis showed less degree of inhibition with a zone of 3 mm. gcu-w-ps2 showed a high degree of inhibition of s. aureus (8 mm). p. aeruginosa showed the least bacteriocin activity. admet & dmpk 10(1) (2022) 75-87 bacteriocin-mediated inhibition of pathogens doi: http://dx.doi.org/10.5599/admet.1053 81 mutagenesis in lactobacillus spp. gcu-w-ps1, gcu-w-ps2, gcu-w-ms1 and gcu-w-ms2 strains were the species that showed maximum bacteriocin activities against pathogenic bacteria i.e., s. aureus, p. aeruginosa, b. thuringiensis, escherichia coli, and salmonella enteritis. so, the mutagenesis induction technique was being used on these strains to check either mutation makes them more reactive against a pathogen or not. ethidium bromide is a teratogen chemical and can change the genetic makeup was being used for inducing mutation. interestingly, it was observed morphologically and biochemically (bacteriocin assay) that mutation enhances the activity of l. curvatus, l. plantarum, and lactobacillus sakei. morphological characteristics of mutants’ v/s wild it can be seen that after mutation, the growth of different lactobacillus strains has been increased at a high rate. the increased growth rate in mutants is shown in figure 3, which is a comparison between wild and mutant lactobacillus. figure 3. l. curvatus (wild and mutant) bacteriocin assay bacteriocin assay (agar well diffusion assay) was done again on mutant species of lactobacillus i.e. l. curvatus, l. plantarum, lactobacillus sakei, and lactobacillus johnsonii (figure 4). they showed more inhibition zone compared to wild species of lactobacillus (table 4). figure 4. comparison of antibacterial activity between wild and mutant species on pathogen s. aureus http://dx.doi.org/10.5599/admet.1053 nazish mazhar ali et al. admet & dmpk 10(1) (2022) 75-87 82 table 4. inhibitory activity of bacteriocin of mutant strains of lactobacillus isolated from poultry meat and pickles against some pathogenic bacteria sample no. zone of inhibition (mm) salmonella enteritidis p. aeruginosa escherichia coli s. aureus b. thuringiensis m ± s.e gcu-w-ps2 5.5 ± 00.16 2.0 ± 00.10 4.0 ± 00.10 7.0 ± 00.10 2.0 ± 00.20 gcu-w-ps1 9.0 ± 00.18 4.0 ± 00.10 5.0 ± 00.10 7.0 ± 00.30 3.0 ± 00.10 gcu-w-ms1 7.0 ± 00.25 2.1 ± 00.39 5.0 ± 00.15 7.8 ± 00.14 3.0 ± 00.13 gcu-w-ms2 8.0 ± 00.10 resistant 6.0 ± 00.10 9.6 ± 00.18 2.1 ± 00.10 characterization of bacteriocin isolated strains cps were being treated with pepsin which is a proteolytic enzyme. when pepsin was being added, it was seen that all isolated strains showed sensitivity against it (table 5). table 5. sensitivity of bacteriocin of bacterial isolates isolated from poultry meat and pickle towards pepsin serial no. isolated bacterial strains sensitivity to pepsin (proteolytic enzyme) 1. gcu-w-ps1 (l. plantarum) sensitive 2. gcu-w-ps2 (lactobacillus sakei) sensitive 3. gcu-w-ms2 ( lactobacillus johnsonii) sensitive 4. gcu-w-ms1 ( l. curvatus) sensitive 5. gcu-m-ps1 (l. plantarum) sensitive 6. gcu-m-ps2 (lactobacillus sakei) sensitive 7. gcu-m-ms2 ( lactobacillus johnsonii) sensitive 8. gcu-m-ms1 ( l. curvatus) sensitive protein estimation by bradford method the standard curve of bsa was plotted (figure 5) and protein concentration produced by lactobacillus spp. was calculated. figure 5. standard curve of bsa molecular characterizations strains with the best inhibitory activity were selected for molecular characterization (table 6). 1 % gel was prepared and about 4l dna sample was loaded for gel electrophoresis sharp bands of genomic dna of gcu-w-ps1, gcu-m-ps1, gcu-w-ps2, gcu-m-ps2, gcu-w-ms1 and gcu-m-ms1 were visualized (figure admet & dmpk 10(1) (2022) 75-87 bacteriocin-mediated inhibition of pathogens doi: http://dx.doi.org/10.5599/admet.1053 83 6). universal forward and reverse primers were used to amplify genomic dna and 1 % agarose gel was being used to check pcr product (figure 7). table 6. accession number given to identified strains sr. no. strain name identified species name accession no. 01 gcu-w-ps1 lactobacillus plantarum ky048156 02 gcu-m-ps1 lactobacillus plantarum ky048157 03 gcu-w-ms1 lactobacillus curvatus ky048158 04 gcu-m-ms1 lactobacillus curvatus ky048159 figure 6. 1 % agarose gel gcu-w-ps1, gcu-w-ps2, gcu-w-ms1, gcu-m-ps1, gcu-m-ps, gcu-m-ms1 indicates the genomic dna isolated from poultry and pickle samples and their mutants figure 7. agarose gel 1% gcu-w-ps1, gcu-w-ps2, gcu-w-ms1, gcu-m-ps1, gcu-m-ps2, gcu-m-ms1 indicates the pcr products of genomic dna of isolated wild lactobacilli from poultry and pickle samples and their mutants. sequencing and sequence analysis for the sequencing of pcr products, the samples were sent to 1 st base laboratory, malaysia. it was found that gcu-w-ps1 (wild) and gcu-m-ps1 (mutant) have 97 % and 99 % resemblance with l. plantarum respectively. gcu-w-ms1 (wild) and gcu-m-ms1 (mutant) strains showed a resemblance 98 % with l. curvatus (table 6). http://dx.doi.org/10.5599/admet.1053 nazish mazhar ali et al. admet & dmpk 10(1) (2022) 75-87 84 characterization of bacteriocin encoding genes using specific primers designed for curvacin, sakacin, and plantaricin genes, strains that showed maximum antibacterial activity were amplified. pcr results had shown gcu-w-ms1 (wild) and gcu-m-ms1 (mutant) (l. curvatus) strain amplified for curvacin gene and showed 100 % homology (figure 8). figure 8. agarose gel electrophoresis of pcr fragments generated with curvacin a specific primer from dna of selected gcu-w-ms1 and gcu-w-ms1m strain (l. curvatus) discussion food safety is becoming a big challenge now a day. in the food industry, the use of chemical preservatives has a positive effect; however, they have some negative impacts on the health of human beings [13]. modern society is more concerned about food safety as chemicals and artificial additives are eliciting poisonous concerns. so era demands to use natural sources for bio-safety and health. bacteriocins are the natural compounds found in probiotics, especially lab, i.e., mostly lactobacillus species [14]. these compounds can kill pathogens found in food, preserve it, and avoid the degradation of food. and they were generally recognized as safe (gras). usually, lactobacillus bacteria can be isolated from yogurt, pickles, and meat. bacteriocin produced by lactobacillus promises safe to use as a food preservative found in vegetables, cheese, dairy products, and meat, as they inhibit pathogenic contamination during the processing [15]. they show high antimicrobial activity against disease-causing bacteria [16]. the inhibition zone can be counted as positive if the zone is 2 mm or greater than 2mm. and the present study shows that the strains i.e. gcu-w-ps1, gcu-w-ps2, gcu-w-ms1, and gcu-w-ms2, exhibit a greater inhibition zone of more than 2 mm. the highest degree of inhibition that was measured is more than 9 mm. gcu-w-ms2 was the strain isolated from meat that had shown a high degree of inhibition on the culture plate of s. aureus 9.2 mm. however, a strain isolated from pickles named gcu-w-ps1 showed resistance for p. aeruginosa. [17] used well diffusion method to ensure the antibacterial activity of lactic acid bacteria of raw milk of cattle against s. aureus, bacillus mycoides, proteus valgaris, and klebsiella pneumonia, these pathogens were found to be sensitive for bacteriocin produced by lactobacillus species [18]. as these strains showed more bacteriocin active results, they were selected for the induction of mutagenesis to check whether a mutation will increase the efficiency against a pathogen. the results have shown that mutagenesis increases the efficiency in some bacteria i.e., gcu-w-ps1 which had no bacterial activity and exhibited no admet & dmpk 10(1) (2022) 75-87 bacteriocin-mediated inhibition of pathogens doi: http://dx.doi.org/10.5599/admet.1053 85 inhibition zone. however, present work showed that after mutation, the pathogenicity of gcu-m-ps1 increased, showing an inhibition zone of 4mm against p. aeruginosa. by site-directed mutagenesis in bacterial structural genes and with the help of genomics and proteomics, this could be made possible to construct a new family with peptides to increase the antibacterial activity [19]. biochemical testing (i.e., sensitivity to proteolytic enzymes and protein estimation by bradford method) and molecular characterization have been proved that mutant strains of lactobacillus have more virulence against pathogens. in [20], a large number of gram-positive bacteria, including food-borne pathogen listeria monocytogenes, bacillus cereus, and s. aureus, were inhibited by bacteriocin produced by l. plantarum. the analysis of lab bacteria from poultry meat and pickle in this study indicated several bacterial genera with the ability of bacteriocin production [21]. in conclusion, the p11 strain (l. curvatus) has bacteriocin encoding gene exhibited inhibitory activity against pathogenic bacteria. time-saving methods must be used for characterization. one of them is pcr which is used for rapid identification. through the pcr bacteriocin encoding, genes were detected. the strains p7, p11 were amplified with already reported primers. generally, one species of bacteria may produce more than one type of bacteriocin and one bacteriocin may not produce by single species of bacteria. two subspecies of l. plantarum isolated from different sources, such as fermented sausages and cucumber fermentation were shown to have the same plantaricin encoding gene plna [22,23]. curvacin a encoding gene was identified in both the l. curvatus and lactobacillus sakeii [24]. in the present study, the gcu-w-ps1 (l. curvatus) strain was amplified with curvacin a gene. conclusion the ability to produce bacteriocins provides the antibacterial potential to lactic acid bacteria against various common pathogens like pseudomonas aeruginosa, staphylococcus aureus, bacillus thuringiensis, escherichia coli, and salmonella enteritis. they show antibacterial activity against both their wild and mutant strains. bacteriocins are safe to use with a minimum negative impact on human health because they are natural products and should be considered safe for food and preservation in the food and dairy industries. conflict of interest: the authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. references [1] e. pasolli, f. de filippis, i.e. mauriello, f. cumbo, a.m. walsh, j. leech, p.d. cotter, n. segata, d. ercolini. large-scale genome-wide analysis links lactic acid bacteria from food with the gut microbiome. nat. commun. 11 (2020) 2610. https:/doi.org/10.1038/s41467-020-16438-8. [2] d. dordević, s. jančíková, m. vítězová, i. kushkevych. hydrogen sulfide toxicity in the gut environment: meta-analysis of sulfate-reducing and lactic acid bacteria in inflammatory processes. j. adv. res. 27 (2021) 55-69. https://doi.org/10.1016/j.jare.2020.03.003. [3] w. churklam, s. chaturongaku, b, ngamwongsatit, b. aunpad. the mechanisms of action of carvacrol and its synergism with nisin against listeria monocytogenes on sliced bologna sausage. food control 108 (2020) 106864. https://doi.org/10.1016/j.foodcont.2019.106864. [4] a.b. snyder, r.w. worobo. chemical and genetic characterization of bacteriocins: antimicrobial peptides for food safety. j. sci. food agric. 94 (2014) 28-44. https://doi.org/10.1002/jsfa.6293. [5] d. drider, g. fimland, y. héchard, l.m. mcmullen, h. prévost. the continuing story of class iia bacteriocins. microbiol. mol. biol. rev. 70 (2006) 564-582. https://doi.org/10.1128/mmbr.00016-05. http://dx.doi.org/10.5599/admet.1053 https://dx.doi.org/10.1038/s41467-020-16438-8 https://doi.org/10.1016/j.jare.2020.03.003 https://doi.org/10.1016/j.foodcont.2019.106864 http://dx.doi.org/10.1002/jsfa.6293 https://doi.org/10.1128/mmbr.00016-05 nazish mazhar ali et al. admet & dmpk 10(1) (2022) 75-87 86 [6] d. gawel, m. maliszewska-tkaczyk, p. jonczyk, r.m. schaaper, i.j. fijalkowska. lack of strand bias in uv-induced mutagenesis in escherichia coli. j. bacteriol. 184 (2002) 4449-4454. https://doi.org/10.1128/jb.184.16.4449-4454.2002. [7] f. kamal, n. samadi, n. moazami, m. r. fazeli. mutagenesis of leuconostoc mesenteroides and selection of dextransucrase hyperproducing strains. daru j. pharm. sci. 9 (2001) 18-23. https://www.sid.ir/en/journal/viewpaper.aspx?id=34248. [8] y. sobrun, a. bhaw-luximon, d. jhurry, d. puchooa. isolation of lactic acid bacteria from sugar cane juice and production of lactic acid from selected improved strains. adv. biosci. biotech. 3 (2012). https://doi.org/10.4236/abb.2012.34057. [9] ünlü, b. nielsen, c. ionita. production of antilisterial bacteriocins from lactic acid bacteria in dairybased media: a comparative study. probiotics antimicrob. proteins 7 (2015) 259-274. https://doi.org/10.1007/s12602-015-9200-z. [10] d. lynch, p.m. o'connor, p.d. cotter, c. hill, d. field, m. begley. identification and characterisation of capidermicin, a novel bacteriocin produced by staphylococcus capitis. plos one 14 (2019) e0223541. https://doi.org/10.1371/journal.pone.0223541. [11] h.f. tai, h.l. foo, r. abdul rahim, t.c. loh, m.p. abdullah, k. yoshinobu. molecular characterisation of new organisation of plnef and plw loci of bacteriocin genes harbour concomitantly in lactobacillus plantarum i-ul4. microb. cell fact. 14 (2015) 89. https://doi.org/10.1186/s12934-015-0280-y. [12] m.m. bradford. a rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. anal. biochem. 72 (1976) 248-254. https://doi.org/10.1006/abio.1976.9999. [13] a. endo, s. maeno, y. tanizawa, w. kneifel, m. arita, l. dicks, s. salminen. fructophilic lactic acid bacteria, a unique group of fructose-fermenting microbes. appl. environ. microbiol. 84 (2018) e01290-18. https://doi.org/10.1128/aem.01290-18. [14] r. hatti-kaul, l. chen, t. dishisha, h.e. enshasy. lactic acid bacteria: from starter cultures to producers of chemicals. fems microbiol. lett. 365 (2018). https://doi.org/10.1093/femsle/fny213. [15] l.h. deegan, p.d. cotter, h. colin. bacteriocins: biological tools for bio-preservation and shelf-life extension. international dairy journal 16 (2006) 1058-1071. https://doi.org/10.1016/j.idairyj.2005.10.026. [16] s.f. henriques, d.b. dhakan, l. serra, a.p. francisco, z. carvalho-santos, c. baltazar, a.p. elias, m. anjos, t. zhang, o.d.k. maddocks, c. ribeiro. metabolic cross-feeding in imbalanced diets allows gut microbes to improve reproduction and alter host behaviour. nat. commun. 11 (2020) 4236. https://doi.org/10.1038/s41467-020-18049-9. [17] a. mohankumar, n. murugalatha. characterization and antibacterial activity of bacteriocin producing lactobacillus isolated from raw cattle milk sample. international journal of biology 3 (2011) 128. https://doi.org/10.5539/ijb.v3n3p128. [18] l.g. ruiz rodríguez, f. mohamed, j. bleckwedel, r. medina, l. de vuyst, e.m. hebert, f. mozzi. diversity and functional properties of lactic acid bacteria isolated from wild fruits and flowers present in northern argentina. front. microbiol. 10 (2019) 1091. https://doi.org/10.3389/fmicb.2019.01091. [19] w.s. kim, r. j. hall, and n. w. dunn. improving nisin production by increasing nisin immunity/resistance genes in the producer organism lactococcus lactis." applied microbiology and biotechnology 50 (1998) 429-433. https://doi.org/10.1007/s002530051316. [20] i. askoul, s.a. gorrah, l. al-amir. isolation and characterization of bacteriocin producing lactic acid bacteria from some syrian fermented foods. international journal of chemtech research 6 (2014), 2507-2520. https://sphinxsai.com/2014/vol6pt4/4/(2507-2520)jul-aug14.pdf. [21] a. sharma, s. lee, y.s. park. molecular typing tools for identifying and characterizing lactic acid bacteria: a review. food sci. biotechnol. 29 (2020) 1301-1318. https://doi.org/10.1007/s10068-02000802-x. https://doi.org/10.1128/jb.184.16.4449-4454.2002 https://doi.org/10.1128/jb.184.16.4449-4454.2002 https://www.sid.ir/en/journal/viewpaper.aspx?id=34248 http://dx.doi.org/10.4236/abb.2012.34057 https://doi.org/10.1007/s12602-015-9200-z https://doi.org/10.1007/s12602-015-9200-z https://dx.doi.org/10.1371/journal.pone.0223541 https://doi.org/10.1186/s12934-015-0280-y https://doi.org/10.1006/abio.1976.9999 https://doi.org/10.1006/abio.1976.9999 https://dx.doi.org/10.1128/aem.01290-18 https://doi.org/10.1093/femsle/fny213 https://doi.org/10.1016/j.idairyj.2005.10.026 https://doi.org/10.1016/j.idairyj.2005.10.026 https://doi.org/10.1038/s41467-020-18049-9 https://doi.org/10.5539/ijb.v3n3p128 https://doi.org/10.3389/fmicb.2019.01091 https://doi.org/10.3389/fmicb.2019.01091 https://doi.org/10.1007/s002530051316 https://sphinxsai.com/2014/vol6pt4/4/(2507-2520)jul-aug14.pdf https://doi.org/10.1007/s10068-020-00802-x https://doi.org/10.1007/s10068-020-00802-x admet & dmpk 10(1) (2022) 75-87 bacteriocin-mediated inhibition of pathogens doi: http://dx.doi.org/10.5599/admet.1053 87 [22] r. shylaja, d. selvakumar, j.h. jagannath. isolation and identification of lactic acid bacteria from raw and fermented products and their antibacterial activity. recent research in science and technology 2 (2010) 42-46. [23] s. saltaji, o. rué, v. sopena, s. sablé, f. tambadou, s. didelot, r. chevrot. lactococcus lactis diversity revealed by targeted amplicon sequencing of purr gene, metabolic comparisons and antimicrobial properties in an undefined mixed starter culture used for soft-cheese manufacture. foods 9 (2020) e622. https://doi.org/10.3390/foods9050622. [24] s.d. todorov, c. rachman, a. fourrier, l.m. dicks, c.a. van reenen, h. prévost, x. dousset. characterization of a bacteriocin produced by lactobacillus sakei r1333 isolated from smoked salmon. anaerobe 17 (2011) 23-31. https://doi.org/10.1016/j.anaerobe.2010.01.004. ©2021 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://dx.doi.org/10.5599/admet.1053 http://dx.doi.org/10.3390/foods9050622 https://doi.org/10.1016/j.anaerobe.2010.01.004 http://creativecommons.org/licenses/by/3.0/ cufe2o4 nanoparticles-based electrochemical sensor for sensitive determination of the anticancer drug 5-fluorouracil doi: http://dx.doi.org/10.5599/admet.1691 201 admet & dmpk 11(2) (2023) 201-210; doi: https://doi.org/10.5599/admet.1691 open access : issn : 1848-7718 http://pub.iapchem.org/ojs/index.php/admet/index original scientific paper cufe2o4 nanoparticles-based electrochemical sensor for sensitive determination of the anticancer drug 5-fluorouracil peyman mohammadzadeh jahani1*, maedeh jafari2 and farhad nazari ravari3 1school of medicine, bam university of medical sciences, bam, iran 2 department of pediatrics, school of medicine, kerman university of medical sciences, kerman, iran 3student research committee, school of medicine, bam university of medical sciences, bam, iran *corresponding author: e-mail: peymanjahani1234@gmail.com received: january 31, 2023; revised: february 17, 2023; published: march 15, 2023 abstract a fast and facile electrochemical sensor for the detection of an important anticancer drug, 5-fluorouracil, is fabricated using cufe2o4 nanoparticles modified screen printed graphite electrode (cufe2o4 nps/spge). the electrochemical activity of the modified electrode was characterized by chronoamperometry, cyclic voltammetry (cv) and differential pulse voltammetry (dpv) and linear sweep voltammetry (lsv) experiments. the cufe2o4 nps improved the electrochemical properties of the electrodes and enhanced their electroanalytical performance. electrochemical measurements using differential pulse voltammetry showed a wide linear relationship between 5-fluorouracil concentration and peak height within the range 0.1 to 270.0 µm with a low detection limit (0.03 µm). further, the sensor was testified with a urine sample and 5-fluorouracil injection sample, and the observed remarkable recovery results replicate its practical applicability. ©2023 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords modified electrode; 5-fluorouracil; cufe2o4 nanoparticles; voltammetriy; electrochemical sensor; spe introduction cancer is a class of diseases characterized by out-of-control cell growth. due to ineffective drugs, cancer is the cause of most deaths worldwide. therefore it's essential to develop cancer research in order to identify causes and develop strategies for prevention, diagnosis, treatments and cure [1,2]. 5-fluorouracil is an antimetabolite fluoropyrimidine analog prescribed as a chemotherapy drug. 5-fluorouracil, a derivative of uracil in which the 5th positioned hydrogen is replaced by fluorine, is used as an excellent antineoplastic agent in treatment of cancer such as colorectal, breast, stomach, pancreatic, and cervical. it acts on cancer cells by directly incorporating into nucleic acids and inhibiting the thymidylate synthase enzyme, which is involved in nucleotide synthesis [3-6] . however, an overdose of 5-fluorouracil may result in very toxic side effects such as mucositis, leukopenia, nausea, diarrhea, alopecia, neurotoxicity, ocular toxicity, and cardiac toxicity by accumulation. consequently, controlling the dose of this drug in biological samples and studying the purity of its pharmaceutical forms can help manage its side effects [7-9]. http://dx.doi.org/10.5599/admet.1691 https://doi.org/10.5599/admet.1691 http://pub.iapchem.org/ojs/index.php/admet/index mailto:peymanjahani1234@gmail.com http://creativecommons.org/licenses/by/4.0/ p.m. jahani et al. admet & dmpk 11(2) (2023) 201-210 202 currently, there are several analytical techniques for the determination of 5-fluorouracil, such as capillary electrophoresis [10], high-performance liquid chromatography [11], raman spectroscopy [12] and mass spectrometry [13]. in spite of their good performance, these techniques are generally expensive, timeconsuming and require relatively complex procedures for sample preparation. the use of electrochemical sensors presents an alternative to the aforementioned analyses for the determination of 5-fluorouracil. electrochemical methods are highlighted and deserve special attention since they present high sensitivity, selectivity, reproducibility, low cost, and rapid response generation, and related to dyes, they present functional groups that can undergo redox reactions, enabling their determination. among the electrochemical techniques, voltammetry deserves special attention once different sensors (modified or not) can be applied [14-27]. magnification of the above properties can be achieved by the application of chemically modified electrode. electrode modification in analytical chemistry has always been an interesting field. modifiers effectively transport electrons between the electrodes to an analyte. it works by minimizing the over-potential required for the electrode reactions and increasing the electrode's sensitivity and selectivity [28-39]. with the advent of nanoscience, various kinds of nanoparticles are utilized for electrochemical sensors in several analytical methods due to their unique character. owing to their small size (normally in the range of 1–100 nm), nanoparticles exhibit unique chemical, physical and electronic properties that are different from those of respective bulk materials and can be used to construct novel and improved sensing devices. compared with the traditional macroelectrodes, nanostructured electrodes show an increased masstransport rate, a decreased influence of the solution resistance and a higher signal-to-noise ratio [40-54]. in order to develop a miniaturized sensor, screen-printed electrodes (spes) are a valuable choice. spes produced by printing different inks on plastic or ceramic supports are gaining widespread applicability for the electrochemical monitoring field. recently, spes have emerged as a simple, disposable, nontoxic and lowcost alternative to mercury-based and conventional solid electrodes for the voltammetric determination of many substances. moreover, they have been shown to be a convenient substrate for nanoparticle modification [55-58]. in the present work, we synthesized cufe2o4 nanoparticles (cufe2o4 nps) and screen-printed graphite electrodes were modified with cufe2o4 nps using the drop-casting method. the resulting modified electrode is applied to the determination of 5-fluorouracil by differential pulse voltammetry. experimental apparatus and chemicals all the electrochemical measurements were carried out on a pgstat302n potentiostat/galvanostat autolab. the measurement cell consisted of spge (dropsens; drp-110: spain) containing a graphite counter electrode, a graphite working electrode, and a silver pseudo-reference electrode. solution ph values were determined using a 713 ph meter combined with a glass electrode (metrohm, switzerland). 5-fluorouracil and all chemicals used were of analytical grade and were used as received without any further purification and were obtained from merck and sigma aldrich. orthophosphoric acid was utilized to prepare the phosphate buffer solutions (pbss), and sodium hydroxide was responsible for adjusting the desired ph values (ph range between 2 and 9). preparation of modified electrode cufe2o4 nps/spges were prepared by modifying the bare working electrode of an spge using the dropcasting method. briefly, 4 µl of the solution of cufe2o4 nps (1 mg/ml) were dropped onto the working admet & dmpk 11(2) (2023) 201-210 cufe2o4 nanoparticles for determination of 5-fluorouracil doi: http://dx.doi.org/10.5599/admet.1691 203 electrode surface and dried at room temperature. the obtained electrode was noted as cufe2o4 nps/spge. the surface areas of the cufe2o4 nps/spges and the un-modified spgewere obtained by cv using 1 mm k3fe(cn)6 at various scan rates. using the randles–sevcik equation for cufe2o4 nps/spges, the electrode surface was found to be 0.113 cm2 which was about 3.6 times greater than un-modified spe. results and discussion electrochemical behavior of 5-fluorouracil at the surface of various electrodes the effect of the electrolyte ph on the oxidation of 50.0 μm 5-fluorouracil was investigated at cufe2o4 nps/spge using differential pulse voltammetry (dpv) measurements in the pbs in the ph range from 2.0 to 9.0. according to the results, the oxidation peak current of 5-fluorouracil depends on the ph value and increases with increasing ph until it reaches the maximum at ph 7.0 and then decreases at higher ph values. the optimized ph corresponding to the higher peak current was 7.0, indicating that protons are involved in the reaction of 5-fluorouracil oxidation. figure 1 displays cv responses from the electrochemical oxidation of 100.0 μm 5-fluorouracil at the surface of cufe2o4 nps/spge (curve b) and bare spge (curve a). the results showed that the oxidation of 5fluorouracil is very weak on the surface of the bare spge, but the presence of cufe2o4 nps in spge could enhance the peak current and decrease the oxidation potential (decreasing the overpotential). a substantial negative shift of the currents starting from oxidation potential for 5-fluorouracil and a dramatic increase of the current indicates the catalytic ability of cufe2o4 nps/spge to 5-fluorouracil oxidation. the results showed that the using of cufe2o4 nanoparticles (curve b) definitely improved the characteristics of 5-fluorouracil oxidation, which was partly due to excellent characteristics of cufe2o4 nps, such as good electrical conductivity and high chemical stability. figure 1. cyclic voltammograms of a) bare spge, b) cufe2o4 nps/spge in the presence of 100.0 μm 5-fluorouracil in 0.1 m phosphate buffer solution, ph 7.0. effect of scan rate on the determination of 5-fluorouracil at cufe2o4 nps/spge the influence of potential scan rate (ν) on ip of 100.0 μm 5-fluorouracil at the cufe2o4 nps/spge was studied by linear sweep voltammetry (lsv) at various sweep rates (figure 2). as shown in figure 2, the peak http://dx.doi.org/10.5599/admet.1691 p.m. jahani et al. admet & dmpk 11(2) (2023) 201-210 204 currents of 5-fluorouracil grow with the increasing scan rates and there are good linear relationships between the peak currents and ν1/2 (square root of scan rate) (figure 2 inset). the regression equation is ipa = 1.0767 ν1/2 +1.4301 (ipa/ μa, ν / mv s−1, r2= 0.9985), indicating the oxidation process of 100.0 μm 5-fluorouracil at the cufe2o4 nps/spge was diffusion-controlled. figure 2. linear sweep voltammograms of 5-fluorouracil at cufe2o4 nps/spge at different scan rates, 1-7 correspond to 10, 30, 70, 100, 200, 300 and 400 mv s−1 in 0.1 m phosphate buffer solution, ph 7.0. inset shows a plot of ipa versus ν1/2 for the oxidation of 5-fluorouracil at cufe2o4 nps/spge. to obtain further information on the rate-determining step, the tafel plot for oxidation of 100.0 µm 5-fluorouracil at the surface of cufe2o4 nps/spge using the data derived from the raising part of the current–voltage curve has been recorded in figure 3. using the slope of the tafel plot at a scan rate of 10 mv s-1, the value of electron transfers coefficient (α) was determined as 0.6, confirming an irreversible process for the oxidation of 5-fluorouracil on the surface of cufe2o4 nps/spge. figure 3. linear sweep voltammograms response for 100.0 μm 5-fluorouracil with 10 mv s-1 scan rate. inset: the tafel plot derived from the rising part of the corresponding voltammogram admet & dmpk 11(2) (2023) 201-210 cufe2o4 nanoparticles for determination of 5-fluorouracil doi: http://dx.doi.org/10.5599/admet.1691 205 chronoamperometric studies the electrochemical oxidation of 5-fluorouracil by a cufe2o4 nps/spge was also studied by chronoamperometry. chronoamperometric measurements of different concentrations of 5-fluorouracil at cufe2o4 nps/spge were done by setting the working electrode potential at 1000 mv (figure 4). in chronoamperometric studies, we have determined the diffusion coefficient, d, of 5-fluorouracil. the experimental plots of i versus t−1/2 with the best fits for different concentrations of 5-fluorouracil were employed (figure 4 a). the slopes of the resulting straight lines were then plotted versus the 5-fluorouracil -concentrations (figure 4 b), from whose slope and using the cottrell equation (1): i =nfad1/2cbπ-1/2t-1/2 (1) we calculated a diffusion coefficient of 8.3×10-5 cm2 s−1 for 5-fluorouracil. figure 4. chronoamperograms obtained at the cufe2o4 nps/spge in 0.1 m phosphate buffer solution, ph 7.0, for different concentrations of 5-fluorouracil. the 1-4 correspond to 0.1, 0.5, 0.9 and 1.7 mm of 5-fluorouracil. (a) plots of i vs. t-1/2 for electrooxidation of 5-fluorouracil obtained from chronoamperometry. (b) plot of the slope of the straight lines against 5-fluorouracil concentration. calibration curve and limit of detection since dpv has a much higher current sensitivity than cyclic voltammetry, we used the dpv method for the determination of 5-fluorouracil (step potential=0.01 v and pulse amplitude=0.025 v). figure 5 shows dpvs of different concentrations of 5-fluorouracil and the obtained calibration curve. the results showed a linear segment for 5-fluorouracil concentration from 0.1 to 270.0 μm 5-fluorouracil (figure 5), with a regression equation of ip = 0.0793c5-fluorouracil + 0.6877 (r2= 0.9994, n=9). the detection limit, lod, was obtained by using the equation (2): lod= 3sb/m (2) where sb is the standard deviation of the blank response (n=15) and m is the slope of the calibration plot. the limit of detection was determined to be 0.03 μm for 5-fluorouracil. http://dx.doi.org/10.5599/admet.1691 p.m. jahani et al. admet & dmpk 11(2) (2023) 201-210 206 figure 5. differential pulse voltammograms of the cufe2o4 nps/spge in 0.1 m phosphate buffer solution (ph7.0) containing different concentrations of 5-fluorouracil, numbers 1–9 correspond to 0.1, 5.0, 15.0,30.0, 70.0, 100.0, 150.0, 200.0 and 270.0 µm of 5-fluorouracil. (b) the plot of the voltammetric peak current as a function of 5-fluorouracil concentration. real sample analysis to investigate the applicability of the proposed sensor for the voltammetric determination of 5-fluorouracil in real samples, we selected urine and 5-fluorouracil injection samples for the analysis of 5-fluorouracil contents. the 5-fluorouracil contents were measured after sample preparation using the standard addition method. the results are given in table 1. according to the table, the recovery values within 98.0-103.6 % confirm the powerful ability of cufe2o4 nps/spge for the determination of 5-fluorouracil in real samples. table 1. the application of cufe2o4 nps/spge for determination of 5-fluorouracil in real samples (n=3) sample c / µm recovery, % rsd, % spiked found urine 0 5.0 4.9 98.0 1.9 7.0 7.2 102.9 3.0 5-fluorouracil injection 0 3.0 3.2 2.5 5.7 103.6 1.6 4.5 7.4 98.7 2.8 conclusions the fabrication of sensors for the measurement of 5-fluorouracil was achieved using screen-printed graphite electrodes modified with cufe2o4 nps. the cufe2o4 nps remarkably decreased overvoltage and improved the electrochemical response of 5-fluorouracil in terms of specificity, sensitivity and current response. under optimized conditions, differential pulse voltammetry exhibited linear dynamic ranges from 0.1-270.0 µm with a detection limit of 0.03 µm. also, the cufe2o4 nps/spge was used to detect 5-fluorouracil in real samples and produced satisfactory results. conflict of interest: authors declare no conflict of interest. https://www.sciencedirect.com/topics/chemistry/overvoltage admet & dmpk 11(2) (2023) 201-210 cufe2o4 nanoparticles for determination of 5-fluorouracil doi: http://dx.doi.org/10.5599/admet.1691 207 references [1] a. hatami, z. azizi haghighat. evaluation of application of drug modeling in treatment of liver and intestinal cancer. progress in chemical and biochemical research 4(2) (2021) 220-233. https://doi.org/10.22034/pcbr.2021.277514.1181 [2] o. oyeneyin, t. abayomi, n. ipinloju, e. agbaffa, d. akerele, o. arobadade. investigation of amino chalcone derivatives as anti-proliferative agents against mcf-7 breast cancer cell lines-dft, molecular docking and pharmacokinetics studies. advanced journal of chemistry-section a 4(4) (2021) 288299. https://doi.org/10.22034/ajca.2021.285869.1261 [3] s.d. bukkitgar, n.p. shetti. electro-oxidation of nimesulide at 5% barium-doped zinc oxide nanoparticle modified glassy carbon electrode. chemistry select 1(4) (2016) 771-777. https://doi.org/10.1002/slct.201600197 [4] b. hatamluyi, z. es' haghi, f.m. zahed, m. darroudi. a novel electrochemical sensor based on gqdspani/zno-ncs modified glassy carbon electrode for simultaneous determination of irinotecan and 5fluorouracil in biological samples. sensors and actuators b: chemical 286 (2019) 540-549. https://doi.org/10.1016/j.snb.2019.02.017 [5] d. lima, g.n. calaça, a.g. viana, c.a. pessôa. porphyran-capped gold nanoparticles modified carbon paste electrode: a simple and efficient electrochemical sensor for the sensitive determination of 5fluorouracil. applied surface science 427 (2018) 742-753. https://doi.org/10.1016/j.apsusc.2017.08.228 [6] d. k. zeybek, b. demir, b. zeybek, ş. pekyardımcı. a sensitive electrochemical dna biosensor for antineoplastic drug 5-fluorouracil based on glassy carbon electrode modified with poly (bromocresol purple). talanta 144 (2015) 793-800. https://doi.org/10.1016/j.talanta.2015.06.077 [7] v. selvaraj, m. alagar. analytical detection and biological assay of antileukemic drug 5-fluorouracil using gold nanoparticles as probe. international journal of pharmaceutics 337 (2007) 275-281. https://doi.org/10.1016/j.ijpharm.2006.12.027 [8] m. ganesan, k.d. ramadhass, h. c. chuang, g. gopalakrishnan. synthesis of nitrogen-doped carbon quantum dots@ fe2o3/multiwall carbon nanotubes ternary nanocomposite for the simultaneous electrochemical detection of 5-fluorouracil, uric acid, and xanthine. journal of molecular liquids 331 (2021) 115768. https://doi.org/10.1016/j.molliq.2021.115768 [9] p. r. s. teixeira, a. s. d. n. m. teixeira, e. a. d. o. farias, e. c. da silva filho, h. n. da cunha, j. r. dos santos júnior, c. eiras. development of a low-cost electrochemical sensor based on babassu mesocarp (orbignya phalerata) immobilized on a flexible gold electrode for applications in sensors for 5fluorouracil chemotherapeutics. analytical and bioanalytical chemistry 411(3) (2019) 659-667. https://doi.org/10.1007/s00216-018-1480-1 [10] h. j. lu, y. l. guo, h. zhang, q. y. ou. rapid determination of 5-fluorouracil in plasma using capillary electrophoresis. journal of chromatography b 788(2) (2003) 291-296. https://doi.org/10.1016/s15700232(03)00033-3 [11] j. ciccolini, c. mercier, m.f. blachon, r. favre, a. durand, b. lacarelle. a simple and rapid high‐performance liquid chromatographic (hplc) method for 5‐fluorouracil (5‐fu) assay in plasma and possible detection of patients with impaired dihydropyrimidine dehydrogenase (dpd) activity. journal of clinical pharmacy and therapeutics 29(4) (2004) 307-315. https://doi.org/10.1111/j.1365-2710.2004.00569.x [12] s. farquharson, c. shende, f.e. inscore, p. maksymiuk, a. gift. analysis of 5‐fluorouracil in saliva using surface‐enhanced raman spectroscopy. journal of raman spectroscopy 36(3) (2005) 208-212. https://doi.org/10.1002/jrs.1277 [13] d. ouyang, q. zheng, h. huang, z. cai, z. lin. covalent organic framework nanofilm-based laser desorption/ionization mass spectrometry for 5-fluorouracil analysis and tissue imaging. analytical chemistry 93(47) (2021) 15573-15578. https://doi.org/10.1021/acs.analchem.1c01743 [14] p.m. jahani. flower-like mos2 screen-printed electrode based sensor for the sensitive detection of sunset yellow fcf in food samples. journal of electrochemical science and engineering 12(6) (2022) 1099-1109. https://doi.org/10.5599/jese.1413 http://dx.doi.org/10.5599/admet.1691 https://doi.org/10.22034/pcbr.2021.277514.1181 https://doi.org/10.22034/ajca.2021.285869.1261 https://doi.org/10.1002/slct.201600197 https://doi.org/10.1002/slct.201600197 https://doi.org/10.1016/j.snb.2019.02.017 https://doi.org/10.1016/j.apsusc.2017.08.228 https://doi.org/10.1016/j.ijpharm.2006.12.027 https://doi.org/10.1016/j.molliq.2021.115768 https://doi.org/10.1016/s1570-0232(03)00033-3 https://doi.org/10.1016/s1570-0232(03)00033-3 https://doi.org/10.1111/j.1365-2710.2004.00569.x https://doi.org/10.1002/jrs.1277 https://doi.org/10.1021/acs.analchem.1c01743 https://doi.org/10.5599/jese.1413 p.m. jahani et al. admet & dmpk 11(2) (2023) 201-210 208 [15] s. azimi, m. amiri, h. imanzadeh, a. bezaatpour. fe3o4@ sio2-nh2/cosb modified carbon paste electrode for simultaneous detection of acetaminophen and chlorpheniramine. advanced journal of chemistry-section a 4(2) (2021) 152-164. https://doi.org/10.22034/ajca.2021.275901.1246 [16] b. dogan-topal, b. bozal-palabıyık, b. uslu, s. a. ozkan. multi-walled carbon nanotube modified glassy carbon electrode as a voltammetric nanosensor for the sensitive determination of anti-viral drug valganciclovir in pharmaceuticals. sensors and actuators b: chemical 177 (2013) 841-847. https://doi.org/10.1016/j.snb.2012.11.111 [17] m. bijad, a. hojjati-najafabadi, h. asari-bami, s. habibzadeh, i. amini, f. fazeli. an overview of modified sensors with focus on electrochemical sensing of sulfite in food samples. eurasian chemical communications 3(2) (2021) 116-138. https://doi.org/10.22034/ecc.2021.268819.1122 [18] o.j. d'souza, r.j. mascarenhas, a.k. satpati, v. mane, z. mekhalif. application of a nanosensor based on mwcnt‐sodium dodecyl sulphate modified electrode for the analysis of a novel drug, alpha‐ hydrazinonitroalkene in human blood serum. electroanalysis 29(7) (2017) 1794-1804. https://doi.org/10.1002/elan.201700114 [19] a. hosseini fakhrabad, r. sanavi khoshnood, m.r. abedi, m. ebrahimi. fabrication a composite carbon paste electrodes (cpes) modified with multi-wall carbon nano-tubes (mwcnts/n, n-bis (salicyliden)1, 3-propandiamine) for determination of lanthanum (iii). eurasian chemical communications 3(9) (2021) 627-634. https://doi.org/10.22034/ecc.2021.288271.1182 [20] a. hosseini fakhrabad, r. sanavi khoshnood, m.r. abedi, m. ebrahimi. fabrication a composite carbon paste electrodes (cpes) modified with multi-wall carbon nano-tubes (mwcnts/n, n-bis (salicyliden)1, 3-propandiamine) for determination of lanthanum (iii). eurasian chemical communications 3(9) (2021) 627-634. https://doi.org/10.22034/ecc.2021.288271.1182 [21] m.s. sengar, s. saxena, s.p. satsangee, r. jain. silver nanoparticles decorated functionalized multiwalled carbon nanotubes modified screen printed sensor for voltammetric determination of butorphanol. journal of applied organometallic chemistry 1(2) (2021) 95-108. http://dx.doi.org/10.22034/jaoc.2021.289344.1023 [22] h. karimi-maleh, f. karimi, y. orooji, g. mansouri, a. razmjou, a. aygun, f. sen. a new nickel-based co-crystal complex electrocatalyst amplified by nio dope pt nanostructure hybrid; a highly sensitive approach for determination of cysteamine in the presence of serotonin. scientific reports 10(1) (2020) 1-13. https://doi.org/10.1038/s41598-020-68663-2 [23] s. saghiri, m. ebrahimi, m.r. bozorgmehr. nio nanoparticle/1-hexyl-3-methylimidazolium hexafluorophosphate composite for amplification of epinephrine electrochemical sensor. asian journal of nanosciences and materials 4(1) (2021) 46-52. https://doi.org/10.26655/ajnanomat.2021.1.4 [24] z. li, c. sheng. nanosensors for food safety. journal of nanoscience and nanotechnology 14(1) (2014) 905-912. https://doi.org/10.1166/jnn.2014.8743 [25] e. sharifi pour, h. beitollai. novel electrochemical sensing platform for caffeine using three dimensional nio nanowrinkles modified glassy carbon electrode. eurasian chemical communications 3(8) (2021) 551-558. https://doi.org/10.22034/ecc.2021.287723.1181 [26] a. shamsi, f. ahour. electrochemical sensing of thioridazine in human serum samples using modified glassy carbon electrode. advanced journal of chemistry-section a 4(1) (2021) 22-31. https://doi.org/ 10.22034/ajca.2020.252025.1215 [27] j. mohanraj, d. durgalakshmi, r.a. rakkesh, s. balakumar, s. rajendran, h. karimi-maleh. facile synthesis of paper based graphene electrodes for point of care devices: a double stranded dna (dsdna) biosensor. journal of colloid and interface science 566 (2020) 463-472. https://doi.org/10.10 16/j.jcis.2020.01.089 [28] s. bilge, b. dogan-topal, e.b. atici, a. sınağ, s.a. ozkan. rod-like cuo nanoparticles/waste masks carbon modified glassy carbon electrode as a voltammetric nanosensor for the sensitive determination of anticancer drug pazopanib in biological and pharmaceutical samples. sensors and actuators b: chemical 343 (2021) 130109. https://doi.org/10.1016/j.snb.2021.130109 https://doi.org/10.22034/ajca.2021.275901.1246 https://doi.org/10.1016/j.snb.2012.11.111 https://doi.org/10.22034/ecc.2021.268819.1122 https://doi.org/10.1002/elan.201700114 https://doi.org/10.22034/ecc.2021.288271.1182 https://doi.org/10.22034/ecc.2021.288271.1182 http://dx.doi.org/10.22034/jaoc.2021.289344.1023 http://dx.doi.org/10.22034/jaoc.2021.289344.1023 https://doi.org/10.1038/s41598-020-68663-2 https://doi.org/10.26655/ajnanomat.2021.1.4 https://doi.org/10.1166/jnn.2014.8743 https://doi.org/10.22034/ecc.2021.287723.1181 https://doi.org/10.22034/ajca.2020.252025.1215 https://doi.org/10.22034/ajca.2020.252025.1215 https://doi.org/10.10‌16/j.jcis.2020.01.089 https://doi.org/10.10‌16/j.jcis.2020.01.089 https://doi.org/10.1016/j.snb.2021.130109 admet & dmpk 11(2) (2023) 201-210 cufe2o4 nanoparticles for determination of 5-fluorouracil doi: http://dx.doi.org/10.5599/admet.1691 209 [29] m. miraki, h. karimi-maleh, m.a. taher, s. cheraghi, f. karimi, s. agarwal, v.k. gupta. voltammetric amplified platform based on ionic liquid/nio nanocomposite for determination of benserazide and levodopa. journal of molecular liquids 278 (2019) 672-676. https://doi.org/10.1016/j.molliq.2019.01.081 [30] r. abdel-karim, y. reda, a. abdel-fattah. nanostructured materials-based nanosensors. journal of the electrochemical society 167(3) (2020) 037554. https://doi.org/10.1149/1945-7111/ab67aa [31] m. vardini, n. abbasi, a. kaviani, m. ahmadi, e. karimi, graphite electrode potentiometric sensor modified by surface imprinted silica gel to measure valproic acid. chemical methodologies 6(5) (2022) 398-408. https://doi.org/10.22034/chemm.2022.328620.1437 [32] s. saghiri, m. ebrahimi, m. bozorgmehr. electrochemical amplified sensor with mgo nanoparticle and ionic liquid: a powerful strategy for methyldopa analysis. chemical methodologies 5(3) (2021) 234239. https://doi.org/10.22034/chemm.2021.128530 [33] d.m. fernandes, n. silva, c. pereira, c. moura, j.m. magalhães, b. bachiller-baeza, c. freire. mnfe2o4@ cnt-n as novel electrochemical nanosensor for determination of caffeine, acetaminophen and ascorbic acid. sensors and actuators b: chemical 218 (2015) 128-136. https://doi.org/10.1016/j.snb.2015.05.003 [34] h. beitollahi, s. salari. sensitive detection of hydrochlorothiazide using ce3+/nio hexagonal nanoparticles modified glassy carbon electrode. eurasian chemical communications 3(1) (2021) 26-34. http://dx.doi.org/10.22034/ecc.2021.120302 [35] h. karimi-maleh, m. sheikhshoaie, i. sheikhshoaie, m. ranjbar, j. alizadeh, n.w. maxakato, a. abbaspourrad. a novel electrochemical epinine sensor using amplified cuo nanoparticles and anhexyl-3-methylimidazolium hexafluorophosphate electrode. new journal of chemistry 43(5) (2019) 2362-2367. https://doi.org/10.1039/c8nj05581e [36] t.n. pham, n. van cuong, n.x. dinh, h. van tuan, v.n. phan, n.t. lan, a.t. le. manganese ferrite nanoparticles (mnfe2o4): size dependence for hyperthermia and negative/positive contrast enhancement in mri. journal of the electrochemical society 168(2) (2021) 026506. https://doi.org/10.1149/1945-7111/abde80 [37] d. wang, x. xiao, s. xu, y. liu, y. li. electrochemical aptamer-based nanosensor fabricated on single au nanowire electrodes for adenosine triphosphate assay. biosensors and bioelectronics 99 (2018) 431437. https://doi.org/10.1016/j.bios.2017.08.020 [38] a. lohrasbi-nejad. electrochemical strategies for detection of diazinon. journal of electrochemical science and engineering 12(6) (2022) 1041-1059. https://doi.org/10.5599/jese.1379 [39] s. a. alavi-tabari, m. a. khalilzadeh, h. karimi-maleh. simultaneous determination of doxorubicin and dasatinib as two breast anticancer drugs uses an amplified sensor with ionic liquid and zno nanoparticle. journal of electroanalytical chemistry 811 (2018) 84-88. https://doi.org/10.1016/j.jelechem.2018.01.034 [40] d.r. kulkarni, s.j. malode, k.k. prabhu, n.h. ayachit, r.m. kulkarni, n.p. shetti. development of a novel nanosensor using ca-doped zno for antihistamine drug. materials chemistry and physics 246 (2020) 122791. https://doi.org/10.1016/j.matchemphys.2020.122791 [41] h. karimi-maleh, a. f. shojaei, k. tabatabaeian, f. karimi, s. shakeri, r. moradi. simultaneous determination of 6-mercaptopruine, 6-thioguanine and dasatinib as three important anticancer drugs using nanostructure voltammetric sensor employing pt/mwcnts and 1-butyl-3-methylimidazolium hexafluoro phosphate. biosensors and bioelectronics 86 (2016) 879-884. https://doi.org/10.1016/j.bios.2016.07.086 [42] m. mirzaei, o. gulseren, m. rafienia, a. zare. nanocarbon-assisted biosensor for diagnosis of exhaled biomarkers of lung cancer: dft approach. eurasian chemical communications 3(3) (2021) 154-161. https://doi.org/10.22034/ecc.2021.269256.1126 [43] g. ozcelikay, s. kurbanoglu, a. yarman, f. w. scheller, s. a. ozkan. au-pt nanoparticles based molecularly imprinted nanosensor for electrochemical detection of the lipopeptide antibiotic drug daptomycin. sensors and actuators b: chemical 320 (2020) 128285. https://doi.org/10.1016/j.snb.2020.128285 [44] r.m. mohabis, f. fazeli, i. amini, v. azizkhani. an overview of recent advances in the detection of ascorbic acid by electrochemical techniques. journal of electrochemical science and engineering 12(6) (2022) 1081-1098. https://doi.org/10.5599/jese.1561 http://dx.doi.org/10.5599/admet.1691 https://doi.org/10.1016/j.molliq.2019.01.081 https://doi.org/10.1149/1945-7111/ab67aa https://doi.org/10.22034/chemm.2022.328620.1437 https://doi.org/10.22034/chemm.2021.128530 https://doi.org/10.1016/j.snb.2015.05.003 http://dx.doi.org/10.22034/ecc.2021.120302 https://doi.org/10.1149/1945-7111/abde80 https://doi.org/10.1016/j.bios.2017.08.020 https://doi.org/10.5599/jese.1379 https://doi.org/10.1016/j.jelechem.2018.01.034 https://doi.org/10.1016/j.matchemphys.2020.122791 https://doi.org/10.1016/j.bios.2016.07.086 https://doi.org/10.22034/ecc.2021.269256.1126 https://doi.org/10.1016/j.snb.2020.128285 https://doi.org/10.5599/jese.1561 p.m. jahani et al. admet & dmpk 11(2) (2023) 201-210 210 [45] j. bastos-arrieta, a. florido, c. pérez-ràfols, n. serrano, n. fiol, j. poch, i. villaescusa. green synthesis of ag nanoparticles using grape stalk waste extract for the modification of screen-printed electrodes. nanomaterials 8(11) (2018) 946. https://doi.org/10.3390/nano8110946 [46] n. koohzadi, z. rezayati zad. voltammetric folic acid sensor based on nickel ferrite nanoparticles modified-screen printed graphite electrode. advanced journal of chemistry-section b 3(4) (2021) 311322. https://doi.org/10.22034/ajcb.2021.302596.1092 [47] t. eren, n. atar, m. l. yola, h. karimi-maleh. a sensitive molecularly imprinted polymer based quartz crystal microbalance nanosensor for selective determination of lovastatin in red yeast rice. food chemistry 185 (2015) 430-436. https://doi.org/10.1016/j.foodchem.2015.03.153 [48] k. harismah, m. mirzaei, m. dai, z. roshandel, e. salarrezaei. in silico investigation of nanocarbon biosensors for diagnosis of covid-19. eurasian chemical communications 3(2) (2021) 95-102. https://doi.org/10.22034/ecc.2021.267226.1120 [49] d. talarico, s. cinti, f. arduini, a. amine, d. moscone, g. palleschi. phosphate detection through a costeffective carbon black nanoparticle-modified screen-printed electrode embedded in a continuous flow system. environmental science & technology 49(13) (2015) 7934-7939. https://doi.org/10.1021/acs.est.5b00218 [50] h. karimi-maleh, r. darabi, m. shabani-nooshabadi, m. baghayeri, f. karimi, j. rouhi, c. karaman. determination of d&c red 33 and patent blue v azo dyes using an impressive electrochemical sensor based on carbon paste electrode modified with zif-8/g-c3n4/co and ionic liquid in mouthwash and toothpaste as real samples. food and chemical toxicology 162 (2022) 112907. https://doi.org/10.1016/j.fct.2022.112907 [51] y.f. mustafa, g. chehardoli, s. habibzadeh, z. arzehgar. electrochemical detection of sulfite in food samples. journal of electrochemical science and engineering 12(6) (2022) 1061-1079. https://doi.org/10.5599/jese.1555 [52] s. cinti, s. politi, d. moscone, g. palleschi, f. arduini. stripping analysis of as(iii) by means of screen‐ printed electrodes modified with gold nanoparticles and carbon black nanocomposite. electroanalysis 26(5) (2014) 931-939. https://doi.org/10.1002/elan.201400041 [53] a. dehno khalaji, p. machek, m. jarosova. α-fe2o3 nanoparticles: synthesis, characterization, magnetic properties and photocatalytic degradation of methyl orange. advanced journal of chemistry-section a 4(4) (2021) 317-326. https://doi.org/10.22034/ajca.2021.292396.1268 [54] h. karimi-maleh, c. karaman, o. karaman, f. karimi, y. vasseghian, l. fu, a. mirabi. nanochemistry approach for the fabrication of fe and n co-decorated biomass-derived activated carbon frameworks: a promising oxygen reduction reaction electrocatalyst in neutral media. journal of nanostructure in chemistry 12 (2022) 429-439. https://doi.org/10.1007/s40097-022-00492-3 [55] s. cinti, f. arduini, d. moscone, g. palleschi, a. j. killard. development of a hydrogen peroxide sensor based on screen-printed electrodes modified with inkjet-printed prussian blue nanoparticles. sensors 14(8) (2014) 14222-14234. https://doi.org/10.3390/s140814222 [56] d. talarico, f. arduini, a. amine, d. moscone, g. palleschi. screen-printed electrode modified with carbon black nanoparticles for phosphate detection by measuring the electroactive phosphormolybdate complex. talanta 141 (2015) 267-272. https://doi.org/10.1016/j.talanta.2015.04.006 [57] m. khairy, b.g. mahmoud, c.e. banks. simultaneous determination of codeine and its co-formulated drugs acetaminophen and caffeine by utilising cerium oxide nanoparticles modified screen-printed electrodes. sensors and actuators b: chemical 259 (2018) 142-154. https://doi.org/10.1016/j.snb.2017.12.054 [58] m.á.g. rico, m. olivares-marín, e.p. gil. modification of carbon screen-printed electrodes by adsorption of chemically synthesized bi nanoparticles for the voltammetric stripping detection of zn (ii), cd (ii) and pb (ii). talanta 80(2) (2009) 631-635. https://doi.org/10.1016/j.talanta.2009.07.039 ©2023 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) https://doi.org/10.3390/nano8110946 https://doi.org/10.22034/ajcb.2021.302596.1092 https://doi.org/10.22034/ecc.2021.267226.1120 https://doi.org/10.1021/acs.est.5b00218 https://doi.org/10.1021/acs.est.5b00218 https://doi.org/10.1016/j.fct.2022.112907 https://doi.org/10.1016/j.fct.2022.112907 https://doi.org/10.5599/jese.1555 https://doi.org/10.5599/jese.1555 https://doi.org/10.1002/elan.201400041 https://doi.org/10.22034/ajca.2021.292396.1268 https://doi.org/10.3390/s140814222 https://doi.org/10.1016/j.talanta.2015.04.006 https://doi.org/10.1016/j.snb.2017.12.054 https://doi.org/10.1016/j.snb.2017.12.054 https://doi.org/10.1016/j.talanta.2009.07.039 http://creativecommons.org/licenses/by/3.0/ development of fiber optic in vitro release method for dexamethasone oil solutions doi: https://doi.org/10.5599/admet.1465 315 admet & dmpk 10(4) (2022) 315-329; doi: https://doi.org/10.5599/admet.1465 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper development of fiber optic in vitro release testing method for dexamethasone release from the oil solutions filip kozlina1, ivan meštrović1, viktor novak1, nikola marjanović2 and biserka cetinačižmek*2 1 university of zagreb, faculty of pharmacy and biochemistry, a. kovačića 1, 10 000 zagreb, croatia 2 pliva croatia ltd, teva group member, prilaz baruna filipovića 25, 10 000 zagreb, croatia *corresponding author: e-mail: biserka.cetina-cizmek@pliva.com; tel.: +385-1-372-2648 received: july 19, 2022; revised: september 14, 2022; published: october 11, 2022 abstract for many parenteral drugs, there is still no standardized method for in vitro release (ivr) testing available. this article presents the development of a new ivr method for oil solutions using a dialysis membrane and usp ii apparatus coupled to a fiber optic uv-vis spectrometer. experiments were performed using dexamethasone formulations containing castor oil as a solvent with the addition of cosolvents, 20 % (v/v) of isopropanol or capryol® 90. based on solubility testing results, castor oil was chosen as the best solvent amongst other vegetable oils, while a significant increase in solubility was obtained by adding either of the two cosolvents. partitioning experiments were performed to ensure these formulations could achieve prolonged drug release. ivr testing was performed with model formulations and critical test parameters were varied in order to examine the method’s sensitivity. the developed method was sensitive to temperature and stirring rate, while coupling the usp ii apparatus with a fiber optic uv-vis spectrometer enabled complete automation. moreover, due to the interference of excipients on fiber optic detection of dexamethasone during the release testing, derivative spectroscopy was successfully introduced for the elimination of the interference. the developed ivr method described herein could be useful in preformulation investigations and the early development of novel formulations. ©2022 by the authors.this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords dissolution; prolonged release injection; parenterals; derivative spectroscopy introduction the release of an active pharmaceutical ingredient (api) from a drug product is a prerequisite for all pharmacokinetic processes as well as the therapeutic effect of an administered drug. in vitro release (ivr) testing is often conducted early in the development of drug products as it can potentially give insight into the viability of the drug product and, preferably, predict its in vivo behaviour. furthermore, ivr test plays an important role in quality control and regulatory processes for some drug products [1-4]. several standardized methods of conducting ivr tests are described in the united states pharmacopoeia (usp), european pharmacopoeia (ph. eur.) and japanese pharmacopoeia (jp), but their application is mostly limited to orally administered drug products and transdermal patches. to date, no standardised ivr test for the parenteral https://doi.org/10.5599/admet.1465 https://doi.org/10.5599/admet.1465 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:biserka.cetina-cizmek@pliva.com http://creativecommons.org/licenses/by/4.0/ filip kozlina et al. admet & dmpk 10(4) (2022) 315-329 316 products is available, which makes the development of such methods an enduring challenge [5,6]. oil solutions are a type of prolonged-release dosage form that allows parenteral administration of poorly soluble drugs or unstable in aqueous systems [7]. vegetable oils (such as castor, sesame, cottonseed, etc.), medium chain triglycerides and fatty acid monoesters, or their mixtures normally play the role of a solvent in these products [8]. following their application (usually intramuscular), the oil forms a depot, from which a prolonged drug release (and some excipients) occurs. the prolonged drug release allows for a significantly less frequent application, making oil solutions especially suitable for the treatment of chronic diseases (such as hypogonadism), as well as diseases characterised by variable patient adherence (e.g. schizophrenia) [9,10]. when developing an oil solution, the key points to think of are the solubility of the drug in the lipophilic vehicle, its viscosity, and the partitioning of the drug between the oil and the aqueous phase. by increasing the solubility of the drug, a higher concentration of the drug in the solution is allowed, and thereby administration of a lower volume of the such drug product. this resolves some of the possible major issues regarding this type of formulation, including pain and irritation at the injection site. solubility of the drug is most often increased using cosolvents, usually, alcohols or simple esters, whose addition simultaneously decreases the viscosity of the solution, making it easier to flow through a hypodermic needle (raising the syringeability of the solution). however, these cosolvents allow higher partitioning of the drug into the aqueous phase, which might shorten the release time of the drug, thus decreasing its value as a sustained release formulation. therefore, the amount of cosolvent added to the solution must be carefully determined in order to balance out its positive effects on drug solubility and solution viscosity while maintaining the partitioning of the drug adequate for prolonged drug release [8]. dexamethasone (dex) is a highly potent glucocorticoid used in the treatment of multiple conditions characterised by severe inflammation, e.g., rheumatoid arthritis. dex is a moderately lipophilic molecule with a log p of 1.68, practically insoluble in water and sparingly soluble in ethanol. its most acidic group has a pka value of 12.42, meaning it is a neutral molecule within the physiological ph range [11]. oil solution-based drug products containing dex have been investigated in several studies so far [12,13], as such setup might prove useful in the treatment of inflammatory diseases due to its potential prolonged release. furthermore, such a solution, applied locally, should pose less of a risk due to drug systemic exposure. considering the fact that dex contains a chromophore, its content can be determined via uv-vis spectrometry. analysis on the standard usp ii dissolution system requires sampling the medium at desired time points, followed by filtration and further analysis. sample analysis can be a significant drawback in time and resources if performed by hplc analysis. on the other hand, an online uv-vis spectrometer coupled with a dissolution system significantly decreases the time needed for obtaining drug release profile curves. the use of a fiber optic automated dissolution system has many advantages compared to both of these options. fiber optic probes connected to a uv-vis spectrometer are immersed into the release medium throughout the entire experiment, thus not requiring additional medium sampling at desired time points and concurrent replacement of the fresh medium. this results in significantly lower consumption of time and materials, such as filters, pipettes, or tubing, compared even to an online spectrometer detection, and especially to manual sampling followed by hplc analysis. simplified detection also decreases the intervals between time points, making it possible to obtain readouts even in less than a minute intervals if desired. finally, in-situ spectroscopic detection through optic probes enables real-time monitoring of drug release during the dissolution process [14]. even though the probe is constantly exposed to room light (unlike conventional cuvettes), fiber optic spectrometry has so far proven to be mostly unaffected by this potential interference. due to the lack of admet & dmpk 10(4) (2022) 315-329 dexamethasone release from the oil solutions doi: https://doi.org/10.5599/admet.1465 317 release medium filtration, detection can be compromised by the interference of unsolved particles or excipients, which can cause light scattering on the optic probes [14]. these interferences can be resolved with the use of mathematical filters, such as baseline correction or derivative spectroscopy [15], as demonstrated in this article. in addition, with the use of an appropriate dialysis bag as a sample container, possible interferences in detection can be minimized or totally diminished. some previous attempts at developing an ivr method for oil solutions utilised a dialysis cell model [16,17] and a simple magnetic stirrer setup with the oil formulation dispersed in a release medium [18]. in general, the ivr methods for oil solutions can be divided into (a) those with the oil solution floating atop the release medium, (b) dialysis methods and (c) continuous flow methods. all these approaches share some of the same possible issues. oil solution tends to float atop the release medium and stick to parts of the apparatus. this problem is completely avoided using dialysis bags, and furthermore, the oil packed in a dialysis bag imitates the depot that is formed in vivo after administration in muscle tissue [8]. this article describes the development of the ivr method using a dialysis bag placed in a paddle apparatus (commonly known as the usp ii apparatus) coupled to a fiber optic spectrometer. the aim of such a setup is to provide an automated ivr method utilising a widely available apparatus, avoiding potential sampling and sink condition-related problems by design, for the screening of various formulations in the early development phase. experimental refined oils used in the solubility and ivr experiments were purchased from croda (united kingdom) and include corn oil, peanut oil, cottonseed oil, sesame oil and castor oil. dex powder was purchased from pfizer (usa). cosolvents propylene glycol monocaprylate type ii (capryol® 90) and isopropanol were purchased from gattefosse (france) and merck (germany), respectively. tetrahydrofuran and acetonitrile were obtained from merck (germany) and used for sample preparation for hplc analysis. regenerated cellulose (rc, mwco (molecular weight cut-off) 50 kda) and cellulose ester (ce, mwco 300 kda) dialysis membranes, used as formulation carriers, were obtained from spectrum labs (usa). regenerated cellulose filter membranes used for medium degassing were obtained from ge healthcare (usa). demineralized water was used throughout the study. all chemicals and solvents were of analytical grade. for the determination of partition coefficient and medium in ivr studies, phosphate buffer ph 7.4, made from sodium hydroxide and potassium dihydrogen phosphate, obtained from kemika (croatia), was used. ethanol was obtained from kemika (croatia). solubility of dex solubility of dex was determined in five different vegetable oils (castor, peanut, cottonseed, sesame and corn), pure or mixed with 10 and 20 % (v/v) of one of the cosolvents (capryol® 90 or isopropanol) by dissolving an excess amount of dex in 3 ml of the solvent/mixture using a shaker at 37 °c and 100 rpm for 24 hours. 1 ml of each supernatant was sampled after 2 hours and 24 hours and then centrifuged at 10000 rpm for 15 minutes to remove the undissolved dex. after centrifugation, 200 µl of supernatant was diluted with 600 µl of tetrahydrofuran and 800 µl of acetonitrile in vials for the hplc analysis. all samples were prepared in triplicates. castor oil combined with 20 % (v/v) of isopropanol and castor oil with 20 % (v/v) of capryol® 90 were chosen as the best formulations for the following experiments. determination of partition coefficients papp (apparent partition coefficient) was used to quantify the partitioning of dex in test solutions. equation (1) describes the calculation of papp: https://doi.org/10.5599/admet.1465 filip kozlina et al. admet & dmpk 10(4) (2022) 315-329 318 𝑃app = ci−𝑐v 𝑐v ∙ 𝑉v 𝑉u (1) where ci represents the concentration of prepared oil solution for determination of papp, cv represents the concentration of drug in the aqueous phase (phosphate buffer ph 7.4) at equilibrium, while vv and vu are volumes of aqueous and oil phase at the end of the experiment, respectively [8]. the partition coefficient between the formulation (castor oil, cosolvent and dex, representing the oil phase) and the phosphate buffer ph 7.4 (representing the aqueous phase) was determined by mixing them at 1:1 volume ratio and shaking at 37 °c and 100 rpm for 24 hours. the concentration of added dex was determined previously in the solubility study (1 mg/ml in capryol® 90 mixture and 2 mg/ml in isopropanol mixture). after 2 and 24 hours, oil and aqueous phases were sampled (1 ml) and centrifuged for 15 minutes at 10000 rpm. the obtained supernatant was diluted as already described in the solubility experiments. all samples were prepared in triplicates. hplc analysis samples from the solubility and partitioning studies were analysed on hplc system agilent 1100/1200 series connected to diode array detector adjusted to 241 nm. kinetex c18 column (4.6 x 50 mm, particles 2.6 μm and pore size 100 å) by phenomenex (usa), heated to 40 °c, was used. the chromatograms were obtained with a gradient elution, where acetonitrile and buffer ph 2.0 were components of the mobile phase. injector temperature was adjusted at 35 °c, the flow rate at 1.8 ml/min and sample injection volume at 10 μl. the stock solution used in solubility and partitioning studies was prepared by dissolving 15 mg of dex in 25 ml of tetrahydrofuran. the two working standard solutions were obtained by further diluting the stock solution. the first one was obtained by diluting 5 ml of stock solution in 25 ml of acetonitrile and was used for assay of dex in castor oil solutions. the second standard solution was prepared by diluting 1 ml of stock solution in 100 ml acetonitrile. it was used for dex assay in solutions composed of four other oils. in vitro release ivr experiments were performed using a usp ii apparatus coupled to the fiber optic uv-vis spectrophotometer (agilent, germany) for real-time absorbance measurement. the 20 mm optical path probes were used for precise online absorbance determination. single point detection at 241 nm was used for the calculation of released dex, while a spectrum scan from 200 to 300 nm at each time point was used later on for derivative spectroscopy. phosphate buffer ph 7.4 was used as a release medium, degassed before every experiment following the usp procedure [19], since the non-degassed medium showed unexpected results due to light scattering caused by bubbles of entrapped air in the fiber optic probes. before each experiment, dialysis membranes were soaked in the dissolution medium for thirty minutes. the membranes were cut in sizes compatible with the intended amount of test solutions. prior to every experiment, for precise absorbance determination, the spectrophotometer baseline was determined by measuring the dissolution medium absorbance. subsequently, the absorbance of prepared standard solutions was also determined for each of the six used probes individually to adequately calibrate the optic fibres. standard solutions were prepared by dissolving 10 mg of dex in 25 ml of tetrahydrofuran and further diluted with the dissolution medium to 100 ml or 200 ml final volume, resulting in 0.004 mg/ml or 0.002 mg/ml final concentration of dex in the standard solutions for isopropanol or capryol® 90 test solutions, respectively. this was equal to the concentration of dex in the release medium in the case of 100 % dex release. admet & dmpk 10(4) (2022) 315-329 dexamethasone release from the oil solutions doi: https://doi.org/10.5599/admet.1465 319 samples for ivr analysis were prepared by dissolving dex in each of the two oil formulations to the concentrations obtained within the solubility study. concentration of dex in castor oil containing 20 % (v/v) isopropanol was 2 mg/ml, while in the castor oil containing 20 % (v/v) capryol® 90, it was 1 mg/ml. samples were tested in each experiment in triplicates at least. dialysis bags were sealed at one end with a weighted pin and 1 ml of the formulation was added. they were later carefully pinned on the other side without trapping air bubbles inside the bags. absorbance measurements were performed at 5, 10, 15, 20, 30, 45, 60, 90, 120, 150, 180, 240, 300, 360, 720 and 1440 minutes to determine the drug dissolution profile at given parameters during the method development. for further studies, additional sampling was performed every 12 hours up to a maximum of 72 hours. as a part of the development of the optimal dissolution method, sensitivity to several parameters was tested: type of dialysis membrane, temperature, release medium, stirring rate, the composition of the release medium, and test solution volume. results and discussion solubility solubility of dex was determined in five oils (solvents) and mixtures with 10 or 20 % (v/v) of one of the cosolvents (isopropanol or capryol® 90). evaluated oils are commonly used as carriers for parenteral oil depots, while selected cosolvents serve as model type of excipients that can increase the solubility of the drug in selected oil in order to achieve model drug therapeutic concentrations and obtain satisfactory partitioning for prolonged release. no significant differences in solubility were observed between 2 and 24 hours of incubation, indicating that the equilibrium occurs quickly after the addition of dex to the solvent mixtures. therefore, only the results for solubility after 24 hours are shown (figure 1). figure 1. solubility of dex after 24 hours incubation in various vegetable oils alone and with the addition of cosolvents (capryol® 90 (blue bars) or isopropanol (orange bars) at 10 % or 20 % (v/v)). error bars represent standard deviation, n=3. dex did not show high solubility in pure vegetable oils, except for castor oil. the composition of triglycerides in corn, cottonseed, sesame and peanut oil is very similar and the fatty acid moieties mostly come from oleic and linoleic acid. castor oil is much more hydrophilic due to the high content (about 90 %) of ricinoleic acid residues amongst the triglycerides. ricinoleic acid contains one hydroxyl group, which enables hydrogen bonding with dex's h-bond acceptor groups [20]. https://doi.org/10.5599/admet.1465 filip kozlina et al. admet & dmpk 10(4) (2022) 315-329 320 cosolvent addition clearly made a significant impact on the solubility of dex. solubility greatly increased with the addition of 10 % (v/v) of both cosolvents and increased with the 20 % (v/v) concentrations. dex was more soluble in isopropanol than capryol® 90 solutions, which was expected due to the hydrophilic nature of isopropanol. the highest solubility of dex was achieved in castor oil solution with 20 % (v/v) of isopropanol (around 2.3 mg/ml) and 20 % (v/v) capryol® 90 (around 1.2 mg/ml). based on the results of the solubility study, test solutions with castor oil as a solvent with the addition of 20 % (v/v) of one of the cosolvents were used for partitioning and ivr experiments. partitioning determination of the partition coefficient is very helpful for the prediction of the behaviour of prolongedrelease formulations in vivo. partition coefficients determined for the chosen test solutions with the phosphate buffer ph 7.4 (later used as a release medium) as the aqueous phase are shown in table 1. log p value higher than 1 was obtained (1.25 for the isopropanol formulation and 1.53 for capryol® 90 formulation), indicating that the formulations could achieve prolonged release. furthermore, it took 24 hours of stirring for the isopropanol test solution to achieve equilibrium, while the capryol® 90 test solution achieved equilibrium much earlier. table 1. results of partitioning study for dex test solutions (sd – standard deviation, n=3). dex conc. ratio (oil /aqueous phase) cosolvent t/h mean value sd log p isopropanol 2 21.1 2.1 1.25 24 10.7 0.1 capryol® 90 2 38.9 0.2 1.53 24 34.7 1.8 the appropriate choice of the type and concentration of cosolvent used in prolonged release parenteral product is reflected in both solubility and partition coefficient of the drug. since better solubility and partitioning of dex were observed using the formulation with 20 % (v/v) cosolvent, such test solutions were used for further ivr experiments. higher concentrations of cosolvents might lead to a significantly lower partition coefficient, which could, in turn, diminish the prolonged release properties of the formulation. in vitro release method development following preliminary tests, several test parameters were varied in further method development in order to determine the method’s sensitivity and optimal test conditions. dialysis membranes during the method development, two types of dialysis membranes were inspected, ce membranes with mwco 300 kda and rc membranes with mwco 50 kda. when comparing dexamethasone release profiles from isopropanol and capryol® 90 containing formulations, the great similarity was observed when ce membranes were used, while a certain difference between the profiles was noticeable with rc membranes which have a smaller pore size and contact area (figure 2). since the concentration of dex is twofold higher in isopropanol compared to capryol® 90 containing formulation, the concentration gradient is consequently higher, which results in faster dex release from the isopropanol-containing formulation in the case where rc membranes are a rate-limiting factor. admet & dmpk 10(4) (2022) 315-329 dexamethasone release from the oil solutions doi: https://doi.org/10.5599/admet.1465 321 figure 2. ivr profiles using two different types of dialysis membranes (ce with mwco 300 kda (solid line) and rc with mwco 50 kda (dashed line)) for isopropanol (orange) and capryol® 90 (blue) test solutions. other test parameters: 75 rpm, 37 °c, 500 ml of phosphate buffer ph 7.4 release medium. error bars represent standard deviation, n=3. temperature the sensitivity of the method to temperature conditions was examined using both test solutions at 25 °c and 37 °c. following basic physical principles, a slower release is expected at a lower temperature due to slower rates of every process involved in dex release into the medium. ivr profiles presented in figure 3 confirm this hypothesis and show that the method is sensitive to temperature. controlled temperature during the test is a critical parameter for this method. figure 3. ivr profiles at two different temperatures (25 °c (dashed line) and 37 °c (solid line)) for isopropanol (orange) and capryol® 90 (blue) test solutions. other test parameters: ce membranes, 75 rpm, 500 ml of phosphate buffer ph 7.4 release medium. error bars represent standard deviation, n=3. https://doi.org/10.5599/admet.1465 filip kozlina et al. admet & dmpk 10(4) (2022) 315-329 322 release medium stirring rate the stirring rate is considered an essential parameter for ivr methods and must be carefully chosen and evaluated. unexpected results, which might arise using an insufficient stirring rate, include thickening of the diffusion layer, leading to a lowered concentration gradient and diffusion potential and finally resulting in a lower release rate. on the other hand, if the stirring rate is too high, the method loses its discriminatory power. using the usp ii apparatus, the fda recommends a stirring rate of 50 to 75 rpm [1]. therefore, the release was tested using these two stirring rates. the method is not very sensitive to different stirring rates. however, slightly higher drug release rates were observed when higher stirring rates were applied (figure 4). figure 4. ivr profiles at two different stirring rates (50 rpm (solid line) and 75 rpm (dashed line)) for isopropanol (orange) and capryol® 90 (blue) test solutions. other test parameters: rc membranes, 37 °c, 500 ml of phosphate buffer ph 7.4 release medium. error bars represent standard deviation, n=3. composition of the release medium aiming to reduce the method duration, dex release from test solutions was analysed using a release medium containing 10 % (v/v) ethanol in phosphate buffer ph 7.4. sink conditions were ensured during these analyses as well. the solubility of dex in phosphate buffer ph 7.4 with 10 % (v/v) ethanol is 0.17 mg/ml [21]. according to the usp criteria, the volume of the release medium to ensure sink conditions for dex should be 238 ml, while to ph. eur. criteria in the range from 238 ml to 473 ml. considering that dex is more soluble in ethanol than in water solutions, it was presumed that the release rate is going to be higher in the medium with the addition of ethanol than in the phosphate buffer alone. however, the results were entirely different (figure 5). the overall percentage of dex released after 24 hours was lower for both test solutions using the release medium with 10 % ethanol and even throughout the entire test for the isopropanol test solution. this is probably due to the fact that ethanol itself can diffuse from the release medium into the test solution through the dialysis membrane pores, thus increasing the solubility of both dex and cosolvent and lowering dex partitioning into the release medium, as well as its release rate. this effect is more pronounced with the isopropanol compared to the capryol® 90 test solution, presumably due to its alcoholic nature. admet & dmpk 10(4) (2022) 315-329 dexamethasone release from the oil solutions doi: https://doi.org/10.5599/admet.1465 323 figure 5. ivr profiles using phosphate buffer (ph 7.4) medium alone (solid line) and modified medium with 10 % (v/v) ethanol (dashed line) for isopropanol (orange) and capryol® 90 (blue) test solutions. other test parameters: rc membranes, 37 °c, 75 rpm, 500 ml of release medium. error bars represent standard deviation, n=3. volume of test solution development and evaluation of ivr methods often include evaluation of the sample volume on release, defining whether the method is sensitive to this parameter. for this purpose, a twice lower volume of the test solution (500 μl) when compared to other experiments was also analysed. when using a smaller sample volume of test solutions with rc membranes, great similarity between the ivr profiles of both tested formulations was observed (figure 6). in these conditions, rc membranes are no longer a rate-limiting factor. on the other hand, ce membranes with larger pore sizes are not a rate-limiting factor even when 1 ml of test solution is used (figure 2). figure 6. ivr profiles using 1.0 ml (solid line) and 0.5 ml (dashed line) of test solutions for isopropanol (orange) and capryol® 90 (blue) test solutions. other test parameters: rc membranes, 37 °c, 75 rpm, 500 ml of phosphate buffer ph 7.4 release medium. error bars represent standard deviation, n=3. https://doi.org/10.5599/admet.1465 filip kozlina et al. admet & dmpk 10(4) (2022) 315-329 324 complete ivr profiles using the optimized test parameters method development results indicate that the method is sensitive to temperature and moderately to stirring rate but not to other evaluated parameters. additional analyses were performed with both test solutions to determine the method reproducibility and overall length of the test required to observe the complete release of the drug. since the main purpose for using dialysis membrane in our method was avoiding interference from formulation excipients and mimicking of oil depot that forms after administration in vivo, we didn’t want the choice of dialysis membrane type to be a rate-limiting factor. based on the results observed between the profiles when comparing the two types of membranes, and due to their easier handling, ce membranes were chosen as a better option for the ivr testing method. the test parameters used were as follows: ce membranes, 37 °c (physiological temperature, higher release rate compared to 25 °c), phosphate buffer (ph 7.4) as release medium (biorelevant regarding ph, ensuring faster release compared to ethanol-containing medium), and 75 rpm stirring rate (ensuring moderately higher release rates compared to 50 rpm). 1 ml of each test solution was added to a dialysis bag. dex release up to 48 hours for both test solutions is shown in figure 7. figure 7. complete dex ivr profiles for capryol® 90 (blue) and isopropanol (orange) containing test solutions in castor oil. test parameters: ce membranes, 37 °c, 75 rpm, 1 ml sample volume, 500 ml of phosphate buffer ph 7.4 release medium. error bars represent standard deviation, n=6. during the first few hours of testing, dex is released at a higher rate than in the later ones. in the period between 6 and 12 hours, a change in the curve slope can be observed, indicating a point at which dex release becomes nearly linear. these two phases in the ivr profiles are a consequence of the physicochemical properties of the cosolvent, which is well soluble both in the oil and the aqueous phase (in the case of isopropanol, the solubility of the cosolvent is even higher in the release medium). this leads to a release of the cosolvent into the release medium (alongside the drug) with kinetics different from those observed for the release of dex alone. since the solubility of isopropanol in the release medium is significantly higher than the solubility of dex, its release is faster than the release of dex in the first hours, during which virtually all isopropanol becomes released to the medium. consequently, the partitioning coefficient changes over time. when the rate of isopropanol release becomes constant or all isopropanol is released, dex's solubility and partitioning admet & dmpk 10(4) (2022) 315-329 dexamethasone release from the oil solutions doi: https://doi.org/10.5599/admet.1465 325 coefficient become constant, leading to a linear terminal phase of dex release. furthermore, this linearity is maintained by a constant concentration gradient, which is guaranteed by design in this method, where sink conditions are ensured throughout the experiment. release kinetics from the oil phase to the oil/water interface in this type of delivery system is affected by drug diffusivity in the oil as well as convection induced by stirring in the aqueous phase [22]. ivr setup does have an influence on the convection process. however, the literature predominantly reports first-order release for oil solutions [22,23], which is also valid for profiles obtained in this study. two phases of release from oil formulations with a hydrophilic cosolvent were also observed in vivo in human blood, using a solution of nandrolone decanoate in sesame oil with benzyl alcohol as a cosolvent [24]. reported biphasic ivr profiles nicely fit observed in vivo pk profiles. overall reproducibility (n = 6, from two different experiments) was achieved, quite acceptable for ivr method for modified release products (standard deviation below 10 % up to 48 or 24 hours for isopropanol and capryol® 90 test solution, respectively). in addition, dex release profiles from both test solutions were quite similar up to 36 hours, indicating that the type of cosolvent does not affect the release rate in these conditions. on the other hand, the obtained dex release profiles (figure 7) were elevated and did not show a characteristic plateau pattern around 100 % at the final time points. this was even more pronounced at time points up to 72 hours, especially for capryol® 90 test solution (data not shown). since that was an indicator of interference during the absorbance reading with optic probes, this effect was further studied using the isopropanol test solution. placebo interference an additional dissolution experiment was performed with a placebo solution containing only castor oil and isopropanol. the obtained absorbance profile up to 72 hours, corresponding to the percent of released dex, was compared to the dex release profile from the test solution (figure 8). increased absorbance of placebo was observed from 48-hour time point, indicating a significant effect of excipients from formulation on total absorbance readout in consecutive time points. the interference from the placebo is probably caused by components of castor oil that diffuse through the dialysis membrane at later time points. the absorbance intensity is more pronounced in placebo alone compared to the extent of interference in the test solution. that could be attributed to the slower diffusion of excipients from the test solution due to competition with faster diffusing dex. since the dialysis membrane of choice for performing the experiments was the ce membrane with mwco 300 kda, these interferences could probably have been reduced to a certain extent by using the rc membrane with mwco 50 kda. instead, the elimination of interference by use of derivative spectroscopy is described in the following lines. derivative spectroscopy in order to eliminate the interference caused by excipient's absorbance at 241 nm, derivative spectroscopy was introduced. dex uv spectrum is characterized by a local maximum of absorbance around 240 nm, while the placebo spectrum is predominantly flat around that wavelength (figure 9). the first derivative curve of the dex spectrum should therefore have two extremes at its peak inflection points, while the placebo would be predominantly flat with values around zero. comparing the first derivative values (at the wavelength corresponding to the curve’s extreme) of standard solution spectra (i.e. 100 % dissolved dex) with respective first derivative values of samples spectra at each time point, the dex release curve without placebo interference should be obtained. https://doi.org/10.5599/admet.1465 filip kozlina et al. admet & dmpk 10(4) (2022) 315-329 326 figure 8. dex release profile from isopropanol test solution (orange) compared to absorbance profile of respective placebo solution (green). error bars represent standard deviation, n=3 (for placebo) and n=6 (for test solution). figure 9. comparison of dex (orange) and placebo (green) uv spectra. in line with that, an additional ivr experiment was performed with the isopropanol test solution, according to the already described procedure. in addition to single point readout at 241 nm, spectrum scans from 200 to 300 nm were obtained for standard solution and samples at each time point in 5 vessels during the 48 hours of experiment duration. by applying the first derivative on spectrum scans, 252 nm was identified as the extreme value for obtained curves. the samples to standards ratio of first derivative values at 252 nm resulted in the dex release profile corrected to excipients interference. the plateau with over 90 % released dex was obtained at two finishing time points. when comparing the corrected average release profile to the one obtained by fiber optic detection, reasonable matching was obtained up to 36-hour time point (figure 10). admet & dmpk 10(4) (2022) 315-329 dexamethasone release from the oil solutions doi: https://doi.org/10.5599/admet.1465 327 although the second derivative may be more suitable for extracting drug-specific information, in this case, more reliable results were obtained with the first derivative. since dexamethasone spectrum has a local maximum around 240 nm, while placebo spectrum is predominantly flat in a wide range around that wavelength, the spectra first derivative resulted with local extremes at inflection points for dexamethasone while for placebo it was flat with values around zero. the local extreme at 252 nm was chosen for calculating dexamethasone release percentages by comparing the first derivative values of samples in each vessel with the respective standard first derivative values. although a relatively narrow range window was obtained for the absolute values of first derivative curves at 252 nm, the resulting release curves were similar when comparing different vessels and reasonable matching was obtained between the average release curve and the one with no correction. moreover, the interference from excipients at later time points was eliminated with the first derivative. the second derivative curves were also calculated by using the same data. since the window range is narrowed with every consecutive derivation step, the second derivative resulted in a range that was too narrow for reliable calculation. therefore, the results obtained from the first derivative curves were used for interference elimination. figure 10. dex average release profiles obtained by fiber optic detection at 241 nm (orange) and corrected to excipients interference by applying the first derivative of obtained spectra (dark blue). error bars represent standard deviation, n=5. conclusions the lack of a standardized ivr method for almost all types of parenteral drug products, including oil parenteral solutions, motivated us to develop one such method using a routinely used usp ii apparatus coupled to a fiber optic uv-vis spectrometer. dex was used as a model drug. test solutions containing castor oil as a solvent with the addition of 20 % (v/v) isopropanol or capryol® 90 as a cosolvent were selected based on the results of solubility testing with several vegetable oils, as well as a partitioning study. the solubility of dex was much higher in castor oil than in other tested vegetable oils. both isopropanol and capryol® 90 proved to be efficient cosolvents, helping to achieve higher concentrations of dex in test solutions (2 mg/ml and 1 mg/ml for isopropanol and capryol® 90, respectively). partitioning of dex in these https://doi.org/10.5599/admet.1465 filip kozlina et al. admet & dmpk 10(4) (2022) 315-329 328 test solutions was adequate to ensure the prolonged release of dex, and therefore they were chosen as model formulations. the developed dialysis method has proven to be sensitive to temperature and moderately to stirring rate and has satisfying reproducibility with respect to usp standards for ivr methods. furthermore, an automated system enables online detection and monitoring of the release through the fiber optic probes immersed in the dissolution medium for the whole duration of the experiment without the need for additional sampling and analysis. packing the samples in dialysis bags helped avoid floating and sticking the oil solution to the apparatus, minimized the chance of interference with optic probe reading and acted as a “sample holder” mimicking in vivo distribution of the oil depot. however, most probably due to diffusion of excipients through the membrane, significant interference in reading was observed in later time points, more pronounced with a formulation containing capryol® 90. excipients interference was successfully eliminated by introducing derivative spectroscopy. both test solutions showed similar, prolonged release profiles. taking into account that the concentration of dex achieved in the solution containing isopropanol is twice the one achieved with capryol® 90, it was concluded that the use of both cosolvents can lead to the development of potential prolonged release oil formulation in the form of solution. the choice of the type and amount of cosolvent determines the potential strength of the developed drug product. the developed method described in this article may be used in oil-based drug product development for fast and direct early preformulation and formulation screening. moreover, it serves as a basis for further research, leading to a standardised ivr method with significance in quality control and regulatory affairs purposes. acknowledgements: the authors would like to thank marieta duvnjak romić, ph.d. for her contribution in designing of the experiments and preparing of the manuscript. conflict of interest: the authors declare no conflict of interest. references [1] b. abrahamson, a.l. ungell. biopharmaceutical support in formulation development. in: pharmaceutical preformulation and formulation; gibson m, ed., crc press llc, boca raton, usa, 2004, p. 239-291. [2] m. jug, a. hafner, j. lovrić, m. lusina kregar, i. pepić, ž. vanić, b. cetina-čižmek, j. filipović-grčić. in vitro dissolution/release methods for mucosal delivery systems. admet and dmpk 5(3) (2017) 173182. http://dx.doi.org/10.1016/j.jpba.2017.06.072 [3] m. jug, a. hafner, j. lovrić, m. lusina kregar, i. pepić, ž. vanić, b. cetina-čižmek, j. filipović-grčić. an overview of in vitro dissolution/release methods for novel mucosal drug delivery systems. journal of pharmaceutical and biomedical analysis 147 (2018) 350-366. https://doi.org/10.1016/j.jpba.2017.06.072 [4] m. probst, m. schmidt, k. tietz, s. klein, w. weitschies, a. seidlitz. in vitro dissolution testing of parenteral aqueous solutions and oily suspensions of paracetamol and prednisolone. international journal of pharmaceutics 532(1) (2017) 519–527. https://doi.org/10.1016/j.ijpharm.2017.09.052 [5] d. j. burgess, d. j. crommelin, a.s. hussain, m. l. chen. assuring quality and performance of sustained and controlled release parenterals. european journal of pharmaceutical sciences 21(5) (2004) 679– 690. https://doi.org/10.1208/ps060111 [6] v. gray, s. cady, d. curran, j. demuth, o. eradiri, m. hussain, j. krämer, j. shabushnig, e. stippler. in vitro release test methods for drug formulations for parenteral applications. dissolution technologies 25(4) (2018) 8-13. https://doi.org/10.14227/dt250418p8 http://dx.doi.org/10.1016/j.jpba.2017.06.072 https://doi.org/10.1016/j.jpba.2017.06.072 https://doi.org/10.1016/j.jpba.2017.06.072 https://doi.org/10.1016/j.ijpharm.2017.09.052 https://doi.org/10.1208/ps060111 https://doi.org/10.14227/dt250418p8 admet & dmpk 10(4) (2022) 315-329 dexamethasone release from the oil solutions doi: https://doi.org/10.5599/admet.1465 329 [7] a. owen, s. rannard. strengths, weaknesses, opportunities and challenges for long acting injectable therdruges: insights for applications in hiv therapy. advanced drug delivery reviews 103 (2016) 114156. https://doi.org/10.1016/j.addr.2016.02.003 [8] s. weng larsen, m.a. thing, c. larsen c. oily (lipophilic) solutions and suspensions, in: long acting injections and implants. j. c. wright, d. j. burgess, eds., springer, new york, usa, 2012, p. 113-136. [9] j. zuidema, f. a. j. m. pieters, g. s. m. j. e. duchateau. release and absorption rate aspects of intramuscularly injected pharmaceuticals. international journal of pharmaceutics 47 (1988) 1-12. https://doi.org/10.1016/0378-5173(88)90209-8 [10] s. spanarello, t. la ferla. the pharmacokinetics of long-acting antipsychotic medications. current clinical pharmacology 9(3) (2014) 310-317. https://doi.org/10.2174/15748847113089990051 [11] pubchem, https://pubchem.ncbi.nlm.nih.gov/compound/dexamethasone (accesed 12th july 2022) [12] o. suitthimeathegorn, j. a. turton, h. mizuuchi, a. t. florence. intramuscular absorption and biodistribution of dexamethasone from non-aqueous emulsions in the rat. international journal of pharmaceutics 331(2) (2007) 204-210. https://doi.org/10.1016/j.ijpharm.2006.11.062 [13] r. macoon, t. guerriero, a. chauhan. extended release of dexamethasone from oleogel based rods. journal of colloid and interface science 555 (2019) 331-341. https://doi.org/10.1016/j.jcis.2019.07.082 [14] x. lu, r. lozano, p. shah. in-situ dissolution testing using different uv fiber optic probes and instruments. dissolution technologies 10 (2003) 6-16. https://doi.org/10.14227/dt100403p6 [15] i. nir, x. lu. in situ uv fiber optics for dissolution testing – what, why, and where we are after 30 years. dissolution technologies 25(3) (2018) 70-77. https://doi.org/10.14227/dt250318p70 [16] k. schultz, b. møllgaard, s. frokjaer, c. larsen. rotating dialysis cell as in vitro release method for oily parenteral depot solutions. international journal of pharmaceutics 157(2) (1997) 163-169. https://doi.org/10.1016/s0378-5173(97)00229-9 [17] d. h. larsen, k. fredholt, c. larsen. assessment of rate of drug release from oil vehicle using a rotating dialysis cell. european journal of pharmaceutical sciences 11(3) (2000) 223-229. https://doi.org/10.1016/s0928-0987(00)00105-6 [18] c.-j. li, m.-y. ku, c.-y. lu, y.-e. tien, w. h. chern, j. huang. in vitro and in vivo release of dinalbuphine sebacate extended release formulation: effect of the oil ratio on drug release. international journal of pharmaceutics 531(1) (2017) 306–312. https://doi.org/10.1016/j.ijpharm.2017.08.083 [19] usp <711> in: united states pharmacopoeia usp 40/nf 35. the united states pharmacopeial convention, rockville, usa, 2015, p. 588-598. [20] v. r. patel, g. g. dumancas, l. c. kasi viswanath, r. maples, b. j. j. subong. castor oil: properties, uses, and optimization of processing parameters in commercial production. lipid insights 9 (2016) 1-12. https://doi.org/10.4137/lpi.s40233 [21] a. šutić, m. duvnjak romić, s. miočić, b. cetina-čižmek. development of analytical method for in vitro release testing of dexamethasone nanosuspensions. dissolution technologies 26(2) (2019) 40-46. https://doi.org/doi.org/10.14227/dt260219p40 [22] m. thing, c. larsen, j. østergaard, h. jensen, s. weng larsen. in vitro release from oil injectables for intra-articular administration: importance of interfacial area, diffusivity and partitioning. european journal of pharmaceutical sciences 45(3) (2012) 351-357. https://doi.org/10.1016/j.ejps.2011.12.006 [23] b. t. pedersen, s. weng larsen, j. østergaard, c. larsen. in vitro assessment of lidocaine release from aqueous and oil solutions and from preformed and in situ formed aqueous and oil suspensions. parenteral depots for intra-articular administration. drug delivery 15(1) (2008) 23-30. https://doi.org/10.1080/10717540701828657 [24] r. w. kalicharan, p. schot, h. vromans. fundamental understanding of drug absorption from a parenteral oil depot. european journal of pharmaceutical sciences 83 (2016) 19-27. https://doi.org/10.1016/j.ejps.2015.12.011 ©2022 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) https://doi.org/10.5599/admet.1465 https://doi.org/10.1016/j.addr.2016.02.003 https://doi.org/10.1016/0378-5173(88)90209-8 https://doi.org/10.2174/15748847113089990051 https://pubchem.ncbi.nlm.nih.gov/compound/dexamethasone https://doi.org/10.1016/j.ijpharm.2006.11.062 https://doi.org/10.1016/j.jcis.2019.07.082 https://doi.org/10.14227/dt100403p6 https://doi.org/10.14227/dt250318p70 https://doi.org/10.1016/s0378-5173(97)00229-9 https://doi.org/10.1016/s0928-0987(00)00105-6 https://doi.org/10.1016/j.ijpharm.2017.08.083 https://doi.org/10.4137/lpi.s40233 https://doi.org/doi.org/10.14227/dt260219p40 https://doi.org/10.1016/j.ejps.2011.12.006 https://doi.org/10.1080/10717540701828657 https://doi.org/10.1016/j.ejps.2015.12.011 http://creativecommons.org/licenses/by/3.0/ electronic tongue for determining the limit of detection of human pathogenic bacteria doi: https://doi.org/10.5599/admet.1650 237 admet & dmpk 11(2) (2023) 237-250; doi: https://doi.org/10.5599/admet.1650 open access : issn : 1848-7718 https://pub.iapchem.org/ojs/index.php/admet/index original scientific paper electronic tongue for determining the limit of detection of human pathogenic bacteria aya abu rumaila1, basima abu rumaila1, wafa masoud1, antonio ruiz-canales2 and nawaf abu-khalaf1* 1department of agricultural biotechnology, faculty of agricultural sciences and technology, palestine technical university-kadoorie (ptuk), p.o. box 7, jaffa street, tulkarm, palestine 2department of engineering, school of engineering of orihuela (epso), miguel hernández university (umh), carretera de beniel, km 3.2, 03312 orihuela, alicante, spain *corresponding author: e-mail: n.abukhalaf@ptuk.edu.ps received: december 23, 2022; revised: february 13, 2023; published: february 17, 2023 abstract the electronic tongue (et) has been used as a diagnostic technique in the medical sector. it is composed of a multisensor array set with high cross-sensitivity and low selectivity characteristics. the research investigated using astree ii alpha mos et to determine the limit of early detection and diagnosis of foodborne human pathogenic bacteria and to recognize unknown bacterial samples relying on pre-stored models. staphylococcus aureus (atcc 25923) and escherichia coli (atcc25922) were proliferated in nutrient broth (nb) medium with original inoculum (approximately 107*105 cfu/ml). they were diluted up to 10-14 and the dilutions ranging from 10-14 to 10-4 were measured using et. the partial least square (pls) regression model detected the limit of detection (lod) of the concentration that was monitored to grow the bacteria with different incubation periods (from 4 to 24 h). the measured data were analysed by principal component analysis (pca) and followed by projecting unknown bacterial samples (at specific concentrations and time of incubation) to examine the recognition ability of the et. astree ii et was able to track bacterial proliferation and metabolic changes in the media at very low concentrations (between the dilutions 10 -11 and 10-10 for both bacteria). s.aureus was detected after 6 h incubation period and between 6 and 8 h for e.coli. after creating the strains’ models, et was also able to classify unknown samples according to their foot-printing characteristics in the media (s.aureus, e.coli or neither of them). the results considered et a powerful potentiometric tool for the early identification of food-borne microorganisms in their native state within a complex system to save patients’ lives. ©2023 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords electronic tongue; food-borne pathogens; multivariate data analysis; principal component analysis; partial least squares. introduction worldwide, food-borne diseases influence public health and cause dangerous diseases. a few pathogenic bacteria are adequate to initiate infection and cause potential damage to the human host system. patients can be treated for dangerous bacterial diseases after an accurate and early diagnosis of the infection, which https://doi.org/10.5599/admet.1650 https://doi.org/10.5599/admet.1650 https://pub.iapchem.org/ojs/index.php/admet/index mailto:n.abukhalaf@ptuk.edu.ps http://creativecommons.org/licenses/by/4.0/ a. abu rumaila et al. admet & dmpk 11(2) (2023) 237-250 238 requires combining signs and symptoms with precise diagnostic tests to give suitable treatment and avoid unnecessary antibiotics [ 2,1 ]. therefore, finding suitable detecting approaches and developing new and fast methods is important for health and safety. colony count, enzyme-linked immunosorbent assay (elisa), electrophoresis, polymerase chain reaction (pcr), biosensors and others have all been employed for the detection of these pathogens [ 4,3 ]. the bacterial normal diagnostic process includes culturing, colony counting and phenotypic characteristics. this usually requires 24 to 48 h to grow the pathogen and obtain a pure culture for further antibiotics testing. moreover, the sensitive and available diagnostic methods (elisa, pcr nucleic acid detection, antigen testing and surface recognition) are expensive, time-consuming and require a high sophistication level and complex sample preparation [ 6,5 ]. consequently, ultrasensitive, advanced, new methods are required to improve the detecting capability of a few or even a single pathogenic bacterial species in the target samples (such as water, food, blood or biological tissues) [7]. chemical and biological sensor technologies have recently become popular analytical tools for complex liquid analysis [8-10]. human smell and taste sensing have been mimicked by the electronic nose (en) and the electronic tongue (et) devices (gas and liquid sensors, respectively) and their communication with the human brain [8,11-17]. liquids and complex solutions can be analyzed using et systems. they are based on an array of multisensor schemes having pronounced cross-sensitivity and low selectivity characteristics [18 -20]. signals obtained from sensors and liquids are processed with multivariate data analysis (mvda) techniques, such as principle component analysis (pca), partial least square (pls), soft independent model class analogy (simca) and discrimination function analysis (dfa,) allowing for obtaining qualitative and quantitative information on the analyzed samples and creating models from the gathered data [15,18,21-25]. using et in the medical analysis is promising to have rapid bacterial detection and shortening the detecting period as much as possible for many physicians, medical laboratories, and even patients as it is an alternative, rapid, reliable and highly sensitive system [26-28]. the limit of detection (lod) defines the lowest concentration of a variable in a sample that can be constantly detected by a particular measurement process at a specified level of confidence without the necessity of being quantitated as an exact value [29,30]. this research aims to evaluate and/or determine the limit for early detection (lod) for both the number of colony-forming units (cfu) and incubation or growing periods of food-borne human pathogenic bacteria using et and multivariate data analysis. also, to identify unknown bacterial samples relying on a pre-stored bacterial model. experimental two bacterial isolates (escherichia coli (atcc25922) and staphylococcus aureus (atcc 25923)) obtained from the american type culture collection (atcc) were cultivated on nutrient agar (na) medium. na medium was prepared by dissolving 23 g of na powder in 1 l distilled water (dw) completely with heating, sterilized at 121 °c and 15 psi for 15 min autoclaving program. the purified medium was cooled and poured into 9 cm petri dishes under aseptic conditions on a microbiological safety cabinet (mn 120). it was then used for culturing the bacteria. the plate count was applied for the viable bacterial count. three fresh well-isolated colonies from na culture medium were suspended in 1 ml sterile nutrient broth (nb) medium, homogenized using a vortex, and 0.1 ml of stock was serially diluted in 0.9 ml nb tenfolds. this was followed by culturing 0.1 ml of each admet & dmpk 11(2) (2023) 237-250 electronic tongue for pathogenic bacteria doi: https://doi.org/10.5599/admet.1650 239 dilution on na medium spread with glass hockey sticks and incubating at 37 °c for 24 h. well-isolated colonies were counted and those within the average of 25-250 cfu were recorded for applying the following equation: cfu/ml = number of colonies  dilution factor / volume of the culture plate the process was repeated three times for the average count. bacterial dna isolation was applied using trizol reagent manual (tri reagent) (cat. # t942) (invitrogen, thermo fisher scientific, us). three fresh well-isolated colonies from fresh na culture media were homogenized using vortex in 1 ml of tri reagent in 1.5 ml microfuge tubes. after that, 200 µl of absolute cold chloroform was added to the suspension, shaken vigorously for 15 sec, and left to stand for 15 min at room temperature. the resulting mixture was centrifuged for 10 min at 11573 rpm at 4 °c to give three phases: colourless upper phase (rna), interphase (dna), and red organic phase (protein lower phase). at this point, 300 µl of cold 100 % ethanol was added after removing and discarding the aqueous overlying phase. tubes were inverted a few times to be mixed and let to stand for 3 min at room temperature, then centrifuged at 4730 rpm for 5 min at 4 °c. the resulting supernatant was removed to be discarded and 1 ml of cold 0.1 m trisodium-citrate in 10 % ethanol solution was used for washing the remaining dna pellets (twice). tubes were allowed to stand for 30 min with occasional mixing, centrifuged at 4730 rpm for 5 min at 4 °c, and the resulting pellets were suspended with 1.5 ml of 75 % cold ethanol and allowed to stand for 20 min at room temperature. later, tubes were centrifuged at 4730 rpm for 5 min at 4 °c discarding the resulting supernatant. in the end, under the vacuum hood, pellets were dried for 10 min, dissolved in 50 µl of te buffer (add 10.8 g tris and 5.5 g boric acid in 900 ml distilled water, then add 4 ml 0.5 m na2edta (ph 8.0), then adjust the volume to 1 l), and stored at -20 °c for further use. pcr amplification for the templates was done using a universal 16s bacterial primer set (forward 27f (agatttgatctggctcag)) and reverse primers 1492r (tacggttaccttgttacgactt)). the primers were dissolved in sterilized distilled dnase-free water to have a final concentration of 100 µm and stored at -20 °c. pcr amplification mixture was done using go taq green 2x pcr master mix with 3 mm mgcl2 (cat. # af9pim7120418m712). 25 µl pcr reaction mixture contained 12.5 µl of 2x ready mix pcr master mix (75 mm tris-hcl, 20 mm (nh4)2so4, 0.625 u thermo prime taq dna polymerase, 0.2 mm of each dntps, 1.5 mm mgcl2), 0.5 µl of 50 mm mgcl2, 0.125 µl of 100 µm forward primer, 0.125 µl of 100 µm reverse primer, 10.75 µl of free dnase water and 1 µl of dna template. vertitm 96 well thermal cycler (cat. #: 4375786) (applied biosystems company, california, usa) was used to perform a pcr amplification program. the program started with an initial 94 °c cycle for 3 min, followed by 35 cycles of 45-sec denaturation cycle at 94 °c, 50 sec of 51 °c, and 1 min at 72 °c, and then 7 min of the final cycle at 72 °c. the pcr procedure was duplicated for each isolate to guarantee the reproducibility of the amplified dna fragments. a blank negative control sample was also run. to separate the total extracted bacterial dna, a 0.8 % agarose electrophoresis gel was used. meanwhile, 2 % agarose electrophoresis gel was prepared to separate pcr products. the gel was prepared by dissolving 2 g of agarose powder completely in 100 ml of 1x tbe buffer with heating using the microwave. the mixture was cooled to 60 °c. after that, 4 µl of 1000x gel red dna stain (cat. #41003) (med chem express, usa) was added and stirred. the suspension was then powered and allowed to solidify in a (10 x 10) tray with 13 wells comp. after submerging the gel in 1 x tbe buffer, 5 µl of pcr products were loaded and the device was run for 2 h at 70 volts. a 10000x gel red dna stain and uv-illuminator were used to visualize dna fragments and syngene gene tool system (synoptics ltd., cambridge c, uk) was used to document it using image acquisition and https://doi.org/10.5599/admet.1650 a. abu rumaila et al. admet & dmpk 11(2) (2023) 237-250 240 documentation. for estimating dna fragments size, a dna ready-to-use (rtu) ladder (cat. # dm001-r500) of 100 bp was used as a molecular marker. finally, pcr products were stored and sent for sequencing. the obtained bacterial sequences were aligned using the universal blast program (national center for biotechnology information, maryland, usa). meanwhile, for et measurements, a liquid taste analyzer astree ii et (alpha mos company, toulouse, france) was used. that is composed of seven sensor arrays (ca, jb, ha, zz, bb, je and ga) with an ag/agcl reference electrode. five testing rounds of bacterial samples were measured on et. the first round was for determining the limit of detection (lod) (limited cfu) for e.coli samples that et can detect after 24 h incubation period. the second was for determining the least incubation time for e.coli that et can detect after cultivating the detected least cfu (the same two rounds were applied for s.aureus). the final fifth round was done to test et capability to recognize unknown bacterial samples of e.coli, s.aureus, and others (s.agalactiae and p.aeruginosa) that were grown at the least incubation time and cfu. in each round, 11 bacterial samples with a nutrient broth (nb) media sample (control) were tested in triplicate. nb was prepared by dissolving a complete weight of 13 g nb powder in 1 l dw by heating, suspended in 250 ml erlenmeyer flasks, each containing 100 ml of the suspension that was labelled and sealed with aluminium foil for autoclaving at 121 °c and 15 psi for 15 min, and left to cool. the overall action was also done at aseptic conditions. bacterial proliferation was done by cultivating three fresh colonies (approx. 107105 cfu/ml) of pure cultured bacteria in 100 ml nb media. the dilution test was applied by serially diluting 1 ml of stock in 99 ml of sterilized nb media up to 14 folds. flasks were then incubated at 37 °c with shaking at 150 rpm for 24 h (the samples with dilutions 10-14 to 10-4 were analyzed using et). meanwhile, the growth period test was applied by inoculating the media with the determined least concentration cfu (approx. 8810-9 cfu/ml) of each bacterial type that was then incubated at 37 °c with shaking at 150 rpm for different periods (4, 6, 8, 10, 12, 14, 16, 18, 20, 22 and 24 h). in the fifth final round, bacterial samples of e.coli, s.aureus, s.agalactiae and p.aeruginosa with nb as a control sample were tested. those samples were measured at 10-9 cfu concentration after 10, 12 and 14 h of inoculation to identify unknown bacterial samples relying on prestored bacterial data and if it can recognize them from other types of bacteria. to create the sequence on et a two parts labelling was applied, where the first part has the bacterial name (i.e. e.coli, staph, unec, unsa, unps, unsr and unnb), the other for the concentration (i.e., -04 to -14 or nb) and/or incubation period (i.e., 04h to 24h or nb) (table 1). before et testing, bacterial samples were filtered using a white cheesecloth to obtain approximately 80 ml of each broth to be placed on et’s 16-position autosampler, with an automatic stirrer, after creating the sequence. samples were separated by four water-cleaning samples for cleaning et sensors after each test. after each measurement, the data from each sensor was collected in a folder categorized by bacterial sequence for each round after creating a library of the experiment. the collected raw data from analyzed sensors were exported to unscrambler x (version 10.3, camo software as, oslo, norway), where the signals of each sensor were numerically analyzed and normalized to values to be categorized using pls and pca. admet & dmpk 11(2) (2023) 237-250 electronic tongue for pathogenic bacteria doi: https://doi.org/10.5599/admet.1650 241 table 1. the et five experiments rounds and the labelling for each tested sample round no. sample no. bacterial type dilution factor the incubation period, h et code goal of the experiment round 1 1 e.coli 10-4 24 e.coli_-04 e.coli concentration lod test 2 e.coli 10-5 24 e.coli_-05 3 e.coli 10-6 24 e.coli_-06 4 e.coli 10-7 24 e.coli_-07 5 e.coli 10-8 24 e.coli_-08 6 e.coli 10-9 24 e.coli_-09 7 e.coli 10-10 24 e.coli_-10 8 e.coli 10-11 24 e.coli_-11 9 e.coli 10-12 24 e.coli_-12 10 e.coli 10-13 24 e.coli_-13 11 e.coli 10-14 24 e.coli_-14 12 -------------24 e.coli_nb round 2 1 e.coli 10-9 4 e.coli_04h e.coli lod for incubation periods test 2 e.coli 10-9 6 e.coli_06h 3 e.coli 10-9 8 e.coli_08h 4 e.coli 10-9 10 e.coli_10h 5 e.coli 10-9 12 e.coli_12h 6 e.coli 10-9 14 e.coli_14h 7 e.coli 10-9 16 e.coli_16h 8 e.coli 10-9 18 e.coli_18h 9 e.coli 10-9 20 e.coli_20h 10 e.coli 10-9 22 e.coli_22h 11 e.coli 10-9 24 e.coli_24h 12 -------------24 e.coli_nb round 3 1 s.aureus 10-4 24 staph_-04 s.aureus concentration lod test 2 s.aureus 10-5 24 staph_-05 3 s.aureus 10-6 24 staph_-06 4 s.aureus 10-7 24 staph_-07 5 s.aureus 10-8 24 staph_-08 6 s.aureus 10-9 24 staph_-09 7 s.aureus 10-10 24 staph_-10 8 s.aureus 10-11 24 staph_-11 9 s.aureus 10-12 24 staph_-12 10 s.aureus 10-13 24 staph_-13 11 s.aureus 10-14 24 staph_-14 12 -------------24 staph_nb round 4 1 s.aureus 10-9 4 staph_04h s.aureus lod for incubation periods test 2 s.aureus 10-9 6 staph_06h 3 s.aureus 10-9 8 staph_08h 4 s.aureus 10-9 10 staph_10h 5 s.aureus 10-9 12 staph_12h 6 s.aureus 10-9 14 staph_14h 7 s.aureus 10-9 16 staph_16h 8 s.aureus 10-9 18 staph_18h 9 s.aureus 10-9 20 staph_20h 10 s.aureus 10-9 22 staph_22h 11 s.aureus 10-9 24 staph_24h 12 -------------24 staph_nb round 5 1 e.coli 10-9 10 unec_10 identify unknown bacterial samples relaying on pre-stored bacterial model 2 e.coli 10-9 12 unec_12 3 e.coli 10-9 14 unec_14 4 s.aureus 10-9 10 unsa_10 5 s.aureus 10-9 12 unsa_12 6 s.aureus 10-9 14 unsa_14 8 p.aeruginosa 10-9 14 unps_14 10 s.agalactiae 10-9 14 unsr_14 11 --------------12 unnb_01 12 --------------14 unnb_01 https://doi.org/10.5599/admet.1650 a. abu rumaila et al. admet & dmpk 11(2) (2023) 237-250 242 results and discussion bacterial experiment bacterial colony forming unit (cfu) counting figure 1 represents the plated bacterial dilution (approximately 88*10-9 cfu/ml) with well-separated and countable colonies of 25-250 cfu, considered for the et lod testing procedure. figure 1. plated bacterial dilution of 88*10-9 cfu/ml with well-separated and countable colonies. a: plate with e.coli, b: plate with s.aureus. bacterial dna isolation and pcr the total dna extracted from four bacterial samples using tri reagent method is shown in figure 2. pcr amplification of dna templates using a universal 16s bacterial primer set (27f and 1492r) resulted in 1500 pb bands used for the sequencing process (figure 3). figure 2. gel electrophoresis documentation of bacterial total dna isolation using tri reagent method. where lanes 1 and 2 represent e.coli samples, 3 and 4 represent s.aureus samples, 5 is a negative control. m=100 bp ladder as a molecular size marker. admet & dmpk 11(2) (2023) 237-250 electronic tongue for pathogenic bacteria doi: https://doi.org/10.5599/admet.1650 243 figure 3. gel electrophoresis documentation of bacterial 16s rrna amplification in eight bacterial isolates using primer 27f and 1492r. 1-4 represents e.coli samples, 4-8 represents s.aureus samples and 9 is a negative control m=100 bp ladder as a molecular size marker. sequence identification blastn alignment of the 16s rrna gene sequences of e.coli and s.aureus bacterial samples are shown in figure 4 and figure 5, respectively. it shows obtained sequence homology of 99 % for e.coli to strain nbrc 102203 and 100 % for s.aureus to strain atcc 12600. figure 4. blastn alignment for e.coli sequenced 16s ribosomal rna with 99 % identity to escherichia coli strain nbrc 102203. https://doi.org/10.5599/admet.1650 a. abu rumaila et al. admet & dmpk 11(2) (2023) 237-250 244 figure 5. blastn alignment of s.aureus sequenced 16s ribosomal rna with 100 % identity to staphylococcus aureus atcc 12600. et data analysis lod test of bacterial concentration the calibration curve of the pls recognition model, for determining the limit of detection (lod) test of bacterial concentration, has identified the presence of bacteria between the dilutions 10-11 and 10-10 for both bacteria e.coli (figure 6) and s.aureus (figure 7). figure 6. pls recognition model for e.coli lod of different dilutions ranged from 10-14 to 10-4. et can sense the presence of bacteria, in nb media, between dilutions 10-11 and 10-10. admet & dmpk 11(2) (2023) 237-250 electronic tongue for pathogenic bacteria doi: https://doi.org/10.5599/admet.1650 245 figure 7. pls recognition model for s.aureus lod of different dilutions ranged from 10-14 to 10-4. et can sense the presence of bacteria, in nb media, between dilutions 10-11 and 10-10. lod test of bacterial earliest incubation period the calibration curve of the pls recognition model for determining the lod test of bacterial earliest incubation period after determining the concentration lod (10-9) identified that et can sense the presence of e.coli, in nb media, between 6 and 8 h of incubation (figure 8) and s.aureus after 6 h of incubation (figure 9). the results are summarized in table 2. figure 8. pls recognition model for e.coli lod of different incubation periods ranged from 4 to 24 h. et can sense the presence of bacteria, in nb media, between incubation periods 6 and 8 h. https://doi.org/10.5599/admet.1650 a. abu rumaila et al. admet & dmpk 11(2) (2023) 237-250 246 figure 9. pls recognition model for s.aureus lod of different incubation periods ranged from 4 to 24 h. et can sense the presence of bacteria, in nb media, at an incubation period of 6 h. table 2. limit of detection (lod) results for s.aureus and e.coli. bacterial type lod of concentration lod of the incubation period s.aureus between 10-11 and 10-10 after 6 h e.coli between 10-11 and 10-10 between 6 and 8 h et classification test et was able to identify two well-separated groups of e.coli and s.aureus in the same pca scores plot, after joining the data for both recognized lod tests (dilution greater than 10-10 and growth time greater than 8 h) in the same pca score plot (figure 10). the scoring model had 99 % pc-1 recognition power. figure 10. pca scores plot for both bacterial data at the recognized lod tests (dilution greater than 10-10 and growth time greater than 8 h). e: e.coli, s: s.aureus. et projection model a pca model for e.coli and s.aureus were created using the resulting data for the projection test, where unknown samples of e.coli and s.aureus were incubated with a dilution of 10-9 for 10, 12 and 14 h, and s.agalactiae and p.aeruginosa as gram-positive and gram-negative bacteria were also incubated with a admet & dmpk 11(2) (2023) 237-250 electronic tongue for pathogenic bacteria doi: https://doi.org/10.5599/admet.1650 247 dilution of 10-9 and for 14 h in order to test the created models and to prove et’s ability to recognise between different bacterial samples. e.coli pca projection model projected unknown e.coli samples close enough to the created models’ data. meanwhile, the unknown p.aeruginosa was out of the group, as well as the projected unknown s.aureus and s.agalactiae, which were far away from the model group (figure 11). s.aureus pca projection model projected samples of unknown s.aureus inside the model’s created group. meanwhile, the unknown s.agalactiae was out of the group (at a distance), as well as the projected unknown e.coli and p.aeruginosa, which were far away from the model group (figure 12). figure 11. e.coli pca projection model with projected unknown samples. a: a group of e.coli’s created data with projected unknown e.colis samples (incubated with a dilution of 10-9 and periods at 10, 12 and 14 h), b: projected unknown p.aeruginosa (incubated with a dilution of 10-9 and 14 h), c: projected unknown s.aureus and s.agalactiae that incubated with a dilution of 10-9 and periods at 10, 12 and 14 h. figure 12. s.aureus pca projection model with projected unknown samples. a: projected unknown e.coli samples (incubated with a dilution of 10-9 and periods at 10, 12 and 14 h), b: projected unknown p.aeruginosa (incubated with a dilution of 10-9 and 14 h), c: a group of all s.aureus’s created data with projected unknown s.aureus samples (incubated with a dilution of 10-9 and periods at 10, 12 and 14 h), d: projected s.agalactiae (incubated with a dilution of 10-9 and 14 h). the 99 % homology for e. coli may be due to mutations throughout the subsequent culturing or the sequencing process. it can also be attributed that e. coli used in this study is a different strain from strain nbrc 102203. https://doi.org/10.5599/admet.1650 a. abu rumaila et al. admet & dmpk 11(2) (2023) 237-250 248 the et was able to classify the two types of bacteria according to their gram-negative and gram-positive strains (i.e., e.coli and s.aureus). moreover, et could sense the difference between the same strains (i.e., e.coli and p.aeruginosa as gram-negative, and s.aureus and s.agalactiae as gram-positive). this can be due to bacteria’s different characteristics. conclusions astree ii et was an efficient technique for tracking bacterial growth and following their metabolic changes in nb media. it was able to create a categorization model that is specific for some strains of microorganisms. moreover, et was able to detect these food-borne bacterial strains just a few hours after inoculation up to only 8 h and even 6 h in some strains such as s.aureus. et’s sensitivity was also confirmed for identifying microorganisms’ proliferation even with a very low concentration of an original inoculum (such as a dilution factor up to 10-10). according to these statements, et can be considered a powerful tool for early identification and fast classification of harmful food-borne microorganisms by creating other subsequent steps to create microorganisms’ models and save patients’ lives. in the long term, this will open a wide door for using these sensors as an alternative assessment and fast monitoring technique in industrial, categorizing, fermentable and other applications. et ease of use in tracking microorganism footprints coupled with distinguishing these microorganisms in the native state (in vitro assessment) and being contained in a complex system is important. however, combining et with other technologies can provide a powerful combination in a wide range of applications. further studies should be carried out to monitor sensors' temperature dependence and charge transfer affected by the adsorption of solution components. also, enlarging the specified foot-printing databases of microorganisms that needs the first step of full work. conflict of interest: the authors declare no conflict of interest. acknowledgements: the authors acknowledge the infrastructure and support of palestine technical university—kadoorie (ptuk) through the master’s program in “agricultural biotechnology”. we acknowledge the collaboration of telenatura ebt, sl enterprise, spain. references [1] n.m. aljamali. review on food poisoning (types, causes, symptoms, diagnosis, treatment). global academic journal of pharmacy and drug research 3 (2021) 54-61. https://doi.org/10.36348/gajpdr.2021.v03i04.001 [2] s.m. parry, s.r. palmer, j. slader, t. humphrey. risk factors for salmonella food poisoning in the domestic kitchena case control study. epidemiology & infection 129 (2001) 277-285. https://doi.org/10.1017/s0950268802007331 [3] n. nordin, n.a. yusof, j. abdullah, s. radu, r. hushiarian. sensitive detection of multiple pathogens using a single dna probe. biosensors and bioelectronics 86 (2016) 398-405. https://doi.org/10.1016/j.bios.2016.06.077. [4] q. yu, l. zhai, x. bie, z. lu, c. zhang, t. tao, j. li, f. lv, h. zhao. survey of five food-borne pathogens in commercial cold food dishes and their detection by multiplex pcr. food control 59 (2016) 862-869. https://doi.org/10.1016/j.foodcont.2015.06.027. [5] j. qin, y. cui, x. zhao, h. rohde, t. liang, m. wolters, d. li, c. belmar campos, m. christner, y. song. identification of the shiga toxin-producing escherichia coli o104: h4 strain responsible for a food poisoning outbreak in germany by pcr. journal of clinical microbiology 49 (2011) 3439-3440. https://journals.asm.org/doi/10.1128/jcm.01312-11. https://doi.org/10.36348/gajpdr.2021.v03i04.001 https://doi.org/10.36348/gajpdr.2021.v03i04.001 https://doi.org/10.1017/s0950268802007331 https://doi.org/10.1016/j.bios.2016.06.077 https://doi.org/10.1016/j.bios.2016.06.077 https://doi.org/10.1016/j.foodcont.2015.06.027 https://journals.asm.org/doi/10.1128/jcm.01312-11 admet & dmpk 11(2) (2023) 237-250 electronic tongue for pathogenic bacteria doi: https://doi.org/10.5599/admet.1650 249 [6] x. zhao, m. li, z. xu. detection of foodborne pathogens by surface enhanced raman spectroscopy. frontiers in microbiology 9 (2018) 1236-1249. https://doi.org/10.3389/fmicb.2018.01236. [7] a. rohde, j. a. hammerl, i. boone, w. jansen, s. fohler, g. klein, r. dieckmann, s. al dahouk. overview of validated alternative methods for the detection of foodborne bacterial pathogens. trends in food science & technology 62 (2017) 113-118. https://doi.org/10.1016/j.tifs.2017.02.006. [8] a. fernandez-lopez, m. ferrández-villena, m. oates, c. cabrera, c. conesa, r. j. agustín, n. abukhalaf, a. ruiz-canales. use of lowcost electronic nose, tongue and eye for monitoring agri-food processes. proceeding of the ii university congress on food innovation and sustainability (cuisa), at the higher polytechnic school of orihuela (epso) of the miguel hernández university of elche, spain, 2021 (abstract in english, paper in spanish). [9] s. kumar, v. kumar, p. suthar, r. saini, t. negi. crop improvement, crc press., florida, usa, 2021, p. 225-236. [10] k. r. srivastava, s. awasthi, p. k. mishra, p. k. srivastava. waterborne pathogens, elsevier, amsterdam, netherland, 2020, 237-277. [11] n. abu-khalaf. identification and quantification of olive oil quality parameters using an electronic nose. agriculture 11 (2021) 674-685. https://doi.org/10.3390/agriculture11070674. [12] n. abu-khalaf, w. masoud. electronic nose for differentiation and quantification of yeast species in white fresh soft cheese. applied bionics and biomechanics, (2022) article id 8472661, 1-5. https://doi.org/10.1155/2022/8472661. [13] n. christodoulides, m.p. mcrae, g.w. simmons, s.s. modak, j. t. mcdevitt. sensors that learn: the evolution from taste fingerprints to patterns of early disease detection. micromachines 10 (2019) 251264. https://doi.org/10.3390/mi10040251. [14] d. ha, q. sun, k. su, h. wan, h. li, n. xu, f. sun, p. wang. recent achievements in electronic tongue and bioelectronic tongue as taste sensors. sensors and actuators b: chemical 207 (2015) 1136-1146. https://doi.org/10.1016/j.snb.2014.09.077. [15] w. masoud, a. al-qaisi, n. abu-khalaf. growth prediction of the food spoilage yeast debaryomyces hansenii using multivariate data analysis. palestine technical university research journal 9 (2021) 2232. https://doi.org/10.53671/ptukrj.v9i1.160. [16] s. mudalal, n. abu-khalaf. electronic nose to differentiate between several drying techniques for origanum syriacum leaves. food research 5 (2021) 260-265. https://doi.org/10.26656/fr.2017.5(6).125. [17] m. peris, l. escuder-gilabert. electronic noses and tongues to assess food authenticity and adulteration. trends in food science & technology 58 (2016) 40-54. https://doi.org/10.1016/j.tifs.2016.10.014. [18] e. legin, o. zadorozhnaya, m. khaydukova, d. kirsanov, v. rybakin, a. zagrebin, n. ignatyeva, j. ashina, s. sarkar, s. mukherjee. rapid evaluation of integral quality and safety of surface and waste waters by a multisensor system (electronic tongue). sensors 19 (2019) 1-15. https://doi.org/10.1016/j.bios.2016.06.077. [19] j.k. lorenz, j.p. reo, o. hendl, j.h. worthington, v.d. petrossian. evaluation of a taste sensor instrument (electronic tongue) for use in formulation development. international journal of pharmaceutics 367 (2009) 65-72. https://doi.org/10.1016/j.ijpharm.2008.09.042. [20] w. wang, y. liu. evaluation technologies for food quality. elsevier, amsterdam, netherland, 2019, p. 23-36. [21] m. kumar, s. kumar, a. gupta, a. ghosh. development of electronic interface for sensing applications with voltammetric electronic tongue. ieee sensors (2018) 1-4. https://doi.org/10.1109/icsens.20 18.8589506. [22] w. masoud, n. abu-khalaf. food and agricultural engineering. proceedings of ires international conference, putrajaya, 2022, p. 16-18. [23] a. veloso, m. sousa, l. estevinho, l. dias, a. peres. honey evaluation using electronic tongues: an overview. chemosensors 6 (2018) 28-53. https://doi.org/10.3390/chemosensors6030028. https://doi.org/10.5599/admet.1650 https://doi.org/10.3389/fmicb.2018.01236 https://doi.org/10.1016/j.tifs.2017.02.006 https://doi.org/10.3390/agriculture11070674 https://doi.org/10.1155/2022/8472661 https://doi.org/10.3390/mi10040251 https://doi.org/10.1016/j.snb.2014.09.077 https://doi.org/10.53671/ptukrj.v9i1.160 https://doi.org/10.26656/fr.2017.5(6).125 https://doi.org/10.1016/j.tifs.2016.10.014 https://doi.org/10.1016/j.tifs.2016.10.014 https://doi.org/10.1016/j.bios.2016.06.077 https://doi.org/10.1016/j.bios.2016.06.077 https://doi.org/10.1016/j.ijpharm.2008.09.042 https://doi.org/10.1109/icsens.20‌18.8589506. https://doi.org/10.1109/icsens.20‌18.8589506. https://doi.org/10.3390/chemosensors6030028 a. abu rumaila et al. admet & dmpk 11(2) (2023) 237-250 250 [24] m. wesoly, p. ciosek. comparison of various data analysis techniques applied for the classification of pharmaceutical samples by electronic tongue. sensors and actuators b: chemical 267 (2018) 570-580. https://doi.org/10.1016/j.snb.2018.04.050. [25] a.n. zaid, r. al ramahi, a. alkilany, n. abu-khalaf, m. el kharouf, d.a. dayeh, l. al-omari, m. yaqoup. following drug degradation and consequent taste deterioration of an oral reconstituted paediatric suspension during dosing interval via electronic tongue. saudi pharmaceutical journal 30 (2022) 555561. https://doi.org/10.1016/j.jsps.2022.02.016. [26] r. al-ramahi, a.n. zaid, n. abu-khalaf. evaluating the potential use of electronic tongue in early identification and diagnosis of bacterial infections. infection and drug resistance 12 (2019) 2445-2457. https://doi.org/10.2147/idr.s213938. [27] e.i. mohamed, s.m. abdel-mageed. the electronic tongue basic principles and medical applications. journal of biophysics and biomedical sciences 3 (2010) 290-295. [28] t. wasilewski, d. migon, j. gebicki, w. kamysz. critical review of electronic nose and tongue instruments prospects in pharmaceutical analysis. analytica chimica acta 1077 (2019) 14-29. https://doi.org/10.1016/j.aca.2019.05.024. [29] e. oleneva, m. khaydukova, j. ashina, i. yaroshenko, i. jahatspanian, a. legin, d. kirsanov. a simple procedure to assess limit of detection for multisensor systems. sensors 19 (2019) 1359. https://doi.org/10.3390/s19061359. [30] n. abu-khalaf, k. f. haselmann. characterization of odorants in an air wet scrubber using direct aqueous injection-gas chromatography-mass spectrometry (dai-gc-ms) and solid phase extraction (spe-gc). american journal of environmental engineering 2 (2012) 58-68. https://doi.org/10.5923/j.ajee.20120203.04. ©2023 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) https://doi.org/10.1016/j.snb.2018.04.050 https://doi.org/10.1016/j.jsps.2022.02.016 https://doi.org/10.2147/idr.s213938 https://doi.org/10.1016/j.aca.2019.05.024 https://doi.org/10.3390/s19061359 https://doi.org/10.5923/j.ajee.20120203.04 https://doi.org/10.5923/j.ajee.20120203.04 http://creativecommons.org/licenses/by/3.0/ levothyroxine sodium loaded dissolving microneedle arrays for transdermal delivery doi: http://dx.doi.org/10.5599/admet.1317 213 admet & dmpk 10(3) (2022) 213-230; doi: https://doi.org/10.5599/admet.1317 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper levothyroxine sodium loaded dissolving microneedle arrays for transdermal delivery riyam f. ghazi, mohammed h. al-mayahy* department of pharmaceutics, college of pharmacy, mustansiriyah university, baghdad-iraq *corresponding author: e-mail: mohammedhussain@uomustansiriyah.edu.iq received: march 06, 2022; revised: april 26, 2022; available online: may 21, 2022 abstract levothyroxine (lt-4) sodium has shown variable bioavailability following oral administration. this can be assigned to the significant influence of gastrointestinal conditions, food and drugs administered concomitantly on the rate and extent of absorption from the gastrointestinal tract. thus, the aim of this research study was to establish an efficient transdermal delivery system of lt-4 sodium via the application of hyaluronic acid dissolving microneedles. microneedles-based drug delivery system consists of sharp-tip needles that puncture the top layers of the skin in a minimally invasive manner to create physical channels through which therapeutic molecules can easily diffuse into/across the skin. hyaluronic acid polymer at different ratios (5-60 %) was used to prepare microneedle arrays (100 needles per array) using a micromoulding technique. characterisation tests were carried out to select the optimum formulation. f11 formula containing 50% w/v hyaluronic acid and 1% v/v tween 80 formula showed an appropriate needle shape with dimensions of 432 ± 6.4 μm in height and a tip diameter of 9.8 ± 1.3 μm. the microneedle arrays demonstrated a suitable mechanical strength after applying a force of 32 n per array and an excellent insertion ability both in parafilm m® and human skin. the in vivo dissolution of microneedles was started rapidly within 5 minutes following the insertion in the skin and completed at 1 hour. ex vivo permeation study using human skin has shown a significant improvement in lt-4 sodium delivery across the skin compared to control preparations (drug solution and microneedle free film). the microneedle array f11 has significantly (p ≤ 0.05) increased lt-4 sodium permeation through the skin (cumulative permeated amount of 32 ± 2 µg/cm²) in comparison to the control solution (cumulative permeated amount of 0.7 ± 0.07 µg/cm²) and the microneedle free film (cumulative permeated amount of 0.1 ± 0.02 µg/cm²) over 7 hours. the findings from the irritation test revealed that mild erythema was produced from the application of microneedle arrays which disappeared within 24 hours. accordingly, dissolving hyaluronic acid microneedles could be a feasible and effective approach to delivering lt-4 sodium transdermally without causing significant skin damage. ©2022 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords hyaluronic acid; levothyroxine sodium; dissolving microneedles; micromoulding; transdermal delivery introduction levothyroxine (lt-4) sodium is predominantly used as replacement therapy for the treatment of hypothyroidism, chronic lymphocytic thyroiditis and simple non-endemic goiters. it is administered in small doses (micrograms), which provides a greater opportunity for clinically significant interactions through the http://dx.doi.org/10.5599/admet.1317 https://doi.org/10.5599/admet.1317 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:mohammedhussain@uomustansiriyah.edu.iq http://creativecommons.org/licenses/by/4.0/ r.f. ghazi and m.h. al-mayahy admet & dmpk 10(3) (2022) 213-230 214 absorption phase to occur [1]. several gastrointestinal conditions such as celiac disease, atrophic gastritis, lactose intolerance and pyloric infection in helicobacter could limit the oral absorption of lt-4 sodium [1,2]. simultaneous administration of oral drugs and food intake can significantly affect lt-4 bioavailability, including the onset, rate and extent of drug absorption. thyroxine malabsorption has been noticed following jejunoileal bypass procedures, in short-bowel syndrome, in association with vitamin b12 malabsorption, in severe hepatic cirrhosis, as well as in congestive heart failure [3]. therefore, it is required to develop an alternative delivery system for a more efficient administration of lt-4 sodium, for instance, via the transdermal route. however, lt-4 is regarded as a poor candidate for transdermal drug delivery due to its inappropriate permeability characteristics. thus, an active skin permeation enhancement strategy is needed for its delivery across the skin, such as microneedle arrays. microneedles (mns) represent a unique technology for enhancing the drug permeability through the stratum corneum. mns based drug delivery consists of sharp-tip needles that puncture the top layers of the skin in a minimally invasive manner to create physical channels through which therapeutic molecules can easily diffuse into/across the skin [4,5]. typically, this approach uses 1 to 400 needles (100 μm to 1 mm apart) which may range from 150 μm to 1000 μm in length with a diameter from 50 μm to 80 μm. the mn array is placed on the skin surface such that mns penetrate the stratum corneum layer, bypassing it and delivering therapeutic agents into the viable epidermis directly. since the epidermis has no nociceptors (pain receptors) and the mns are frequently do not reach the dermis, pain sensation is not triggered and therefore, they are regarded as a minimally invasive technique with pain-free delivery. morphologically microneedles can be classified into five different types, including solid microneedles [6], coated microneedles [7], dissolving microneedles [8], hollow microneedles [9] and hydrogel-forming microneedles [10]. they can be manufactured of a variety of materials, such as metals [11], inorganic materials [12] and polymeric materials [13]. drug-loaded dissolving mns containing a drug within the needle matrix can overcome the limitations associated with other mn types. they solve the problem of the two-step application of solid mns. as a result, a one-step application technique will be available [14]. in comparison to coated mns, a higher and more precise drug loading can be achieved by dissolving mns [15]. additionally, they eliminate the need for specialized techniques used in hollow mns to control the flow rate and pressure for driving the flow of liquid into the skin. however, some of the biodegradable polymers, including plga, polylactic acid, polyglycolic acid and polycarbonate used in the manufacturing of dissolving mns are unsuitable for the fabrication of lt-4 sodium loaded mns, particularly during the manufacturing process. this can be attributed to the need for a heating step that may cause a degradation of heat-sensitive drugs such as lt-4 sodium. other materials used in the fabrication of dissolving mns, such as carbohydrates, undergo hydrolysis in a moist environment at a humidity level exceeding 43 %, resulting in needle distortion or disappearance, as well as insufficient insertion ability into the skin [16,17]. hyaluronic acid (ha), commonly utilized in skincare products, was identified to generate mns with good biocompatibility and deformation resistance. the obtained ha mns were easily prepared without the need of a heating step. in addition, they have sufficient strength to reliably pierce the skin, dissolve and rapidly release the contained drug within the skin [18,19]. therefore, in the present study, novel lt-4 sodium-loaded ha mns were developed for enhancing transdermal delivery. to the best of our knowledge, this is the first study conducted to investigate lt-4 delivery across the skin using ha mns. admet & dmpk 10(3) (2022) 213-230 transdermal delivery of levothyroxine via dissolving microneedles doi: http://dx.doi.org/10.5599/admet.1317 215 experimental materials levothyroxine sodium and hyaluronic acid 20 kd were purchased from (look chemical, china). tween 80 was obtained from (chemical point, united kingdom). blue silica gel powder was purchased from (om chemicals, india) and formalin was obtained from (sigma-aldrich, germany). all chemicals and reagents used were of analytical grade. preparation of lt-4 sodium loaded hyaluronic acid microneedle arrays mn arrays were manufactured by micromoulding technique with ha as the structural material. mns manufacturing can be considered as a transcription process from the micromould [20]. initially, lt-4 sodium was dissolved in deionised water (dw) with the aid of tween 80. following this, ha was added to this solution at specific percentages and the volume was completed up to 100 % with dw to prepare different polymeric formulation mixtures as illustrated in table 1. the mixtures were then stirred for 15 minutes using a magnetic stirrer at 600 rpm and left to settle overnight to remove any bubbles formed. afterward, 100 µl of each polymeric formulation mixture containing 50 µg of lt-4 as an active pharmaceutical ingredient was poured into mn moulds (pdms mould consisting of 10 × 10 array on a 0.64 cm² area with 500 µm needle height, 300 µm needle base and 500 µm interspacing). the mn moulds were subjected to centrifugation for 40 minutes for two rounds (2 х 40 min.) to ensure the whole volume of polymeric mixture has entered microcavities of the mould. during the centrifugation process, the mn moulds were covered with silicon caps and adhesive tape to prevent the solution spill out from the mould. the moulded mixture was kept to dry for 24 hours in a desiccator at ambient conditions protected from light. the microneedle arrays were then removed from the mould and examined using a digital microscope. mns were stored wrapped with aluminium foil in a desiccator until use. table 1. composition of polymeric dissolving microneedle formulations. formulation code lt-4 (% w/v) ha (% w/v) tween 80 (% v/v) dw (% v/v) up to f1 0.05 5 1 100 f2 0.05 5 2 100 f3 0.05 10 1 100 f4 0.05 15 1 100 f5 0.05 20 1 100 f6 0.05 25 1 100 f7 0.05 30 1 100 f8 0.05 35 1 100 f9 0.05 40 1 100 f10 0.05 45 1 100 f11 0.05 50 1 100 f12 0.05 55 1 100 f13 0.05 60 1 100 http://dx.doi.org/10.5599/admet.1317 r.f. ghazi and m.h. al-mayahy admet & dmpk 10(3) (2022) 213-230 216 characterisation of hyaluronic acid microneedle arrays assessment of the mechanical properties of ha microneedle arrays. a ta-xt2 texture analyser (stable microsystems, copley, uk) was used to test mn arrays for their mechanical strength. in brief, mn arrays were mounted to the texture analyser's movable probe (0.5 cm in length). the probe was set to move downward at a pre-test speed of 1 mm/s, test speed of 0.5 mm/s and post-test speed of 10 mm/s. a force of 32 n per array was applied for 30 seconds to press the mn arrays down towards a flat metal block of 9.2 × 5.2 cm dimensions [18]. a digital microscope (rohs, china) was utilised to observe the mn arrays prior and following the appliance of the compression load. individual mns’ heights were measured before and after testing using the ruler feature of image j® software (us national institutes of health, bethesda, maryland, usa) in order to determine the percent height reduction using equation number (1) [21]: % compression = hbc hac / hbc*100 . (1) hbc is the height prior to compression and hac is the height following compression. in vitro drug release study. in vitro drug release of lt-4 sodium was performed for mns formulations (f8, f9, f10 and f11) using franz diffusion cells with an exposed surface area of 0.785 cm². mn arrays were fixed on a small sieve and mounted on franz cells with the needles facing downwards. the receptor chamber was filled with 10 ml of the solvent mixture containing ethanol (25 % v/v) and dw (75 % v/v) as a receptor fluid. franz cells were then placed in diffusion cells apparatus (stirring water bath, orchid scientific, india) at 37 °c. the receptor fluid was stirred continuously by a small teflon-coated magnetic bar at 400 rpm. sink condition was maintained; since the samples of 1 ml were removed from the receptor, fluid was recovered with the same volume of the fresh pre-heated solvent mixture. samples withdrawn were analyzed for lt-4 sodium content by uv-spectrophotometer to determine the percentage of drug released [22]. detection of microneedles insertion ability using parafilm m®. in their latest work (larraneta et al., 2014), they demonstrated a great resemblance between parafilm m® (pf) layers and porcine skin, suggesting that simple insertion tests on pf layers may be performed reliably. the film is folded into eight layers to achieve a thickness of approximately 1 mm, with one layer being 127 µm. the mn array was inserted into the eight membrane layers using a force of (32 n) generated by the texture analyser instrument. following insertion, the mn array was withdrawn from the workbench and the pf layers were unfolded, allowing the number of holes remaining in each layer to be determined (using a digital microscope). the thickness of one layer being 127 µm ± 7 µm, hence the insertion depth can be deduced [23]. insertion of microneedle arrays into excised human skin. to investigate the skin insertion ability of mn arrays prepared with different proportions of ha to penetrate the skin and form holes, an insertion and staining test was performed. human skin samples were obtained from 32 years old female after liposuction surgery. mn arrays were inserted into human skin by pressing them against the skin surface for 1 minute and then removed. subsequently, the skin surface was subjected to staining by application of methylene blue dye for 10 minutes to identify the insertion sites. after that, the excess of the dye was removed by tissue paper and the mn treated skin area was observed using a digital microscope. the insertion ratio for each mn array was calculated as the number of blue spots on the skin surface divided by the number of needles in the array [16]. histological examination of microneedles treated skin. to demonstrate the penetration depth of mns, a histological examination was performed. skin cross-sectioning and staining techniques were used to assess the formation of microchannels within skin layers. briefly, human skin samples treated with mns were embedded in paraffin wax blocks and subjected to cross-sectioning into 5 μm slices using an electrical admet & dmpk 10(3) (2022) 213-230 transdermal delivery of levothyroxine via dissolving microneedles doi: http://dx.doi.org/10.5599/admet.1317 217 microtome. these cross-sectioned skin samples were then stained with hematoxylin and eosin (h&e) dyes to be examined under a light microscope [24]. in vivo dissolution study. in vivo dissolution study of the f11 mn array was performed using 2 month old rats weighing approximately 147-150 g (n = 3). the mns arrays were inserted in the rat skin for different time intervals of 0, 5, 10, 20, 30 and 60 minutes. after that, the mn arrays were removed at each specific time point to be examined under a light microscope [25]. ethical approval for all animal studies in this research was obtained from the animal ethics committee (aec) of the college of pharmacy/ mustansiriyah university/ iraq-baghdad under license no. 9. visualization of microneedles by scanning electron microscopy (sem). the morphological characteristics and dimensions of the mns were determined using a tescan mira 3 sem instrument. in this test, a mn array was fixed on an aluminum plate with adhesive tape and covered with gold/palladium of 20 μm thickness. the accelerating voltage was 15 kv and the time of the test was 2 minutes. a mn array was tested by sem at different magnifications; 50, 100 and 500 x, to show the needles’ length, width, needle tip diameter and distance between needles [17]. the sem images were captured using tescan’s essence software. ex vivo permeation study. franz diffusion cells with an exposed surface area of 0.785 cm² were used to investigate the permeation of lt-4 sodium from the optimum mn array formula f11 across human skin. skin samples were obtained from the abdomen of 32 years old woman who had undergone a plastic surgery operation [19]. a consent form was obtained from the patient to collect and use her excised skin in ex vivo experiments. the skin was washed with normal saline and dried using tissue paper. after that, skin samples were wrapped in an aluminum foil and stored at -20 °c until required. skin samples were used within a month of storage. prior to starting the experiment, skin samples were soaked for 1 hour in normal saline to hydrate them and then they were placed on a flat desk with the stratum corneum facing upwards. subsequently, ha mn arrays were physically pressed onto the skin for 30 seconds and secured using adhesive tape clamped between the donor and receptor chambers. the receptor chamber was filled with 10 ml of solvent mixture (ethanol 25 % v/v and dw 75 % v/v) as a receptor fluid. franz cells (n = 6) were then placed in diffusion cells apparatus at 37 °c. the receptor fluid was stirred continuously by a small teflon-coated magnetic bar at 400 rpm. samples of 1 ml were withdrawn from the receptor fluid and replaced with a pre-heated fresh medium following each sampling time (10, 30, 60, 90, 180, 300 and 420 minutes) to maintain the sink condition. the withdrawn samples were then assayed by a uv spectrophotometer to determine the concentration of lt-4 sodium. for comparison, control formulas containing the same ingredients of the optimum formula, one in a liquid form and another one as a film without needles (needle-free), were prepared to highlight the influence of needles on lt-4 sodium permeation. the control formulas were tested for skin permeation under the same experimental conditions as the ha mn arrays. the cumulative amount of lt-4 sodium was plotted against time to calculate the flux and percentage of the dose delivered across the skin [26]. skin irritation test. draize test was chosen to notice erythema and edema on the rat skin following the application of mn arrays [27]. three healthy rats were selected and their skin was determined to be normal in the absence of a scratch or wound. twenty-four hours before the test, the rats’ back skin was shaved. after that, the mn arrays were fixed to the skin for two hours. following the removal of the mn arrays, draize dermal scoring criteria were used to evaluate the erythema and edema at 1, 24 and 72 hours [27]. the following equation was used to estimate the primary irritation index (p.i.i.): http://dx.doi.org/10.5599/admet.1317 r.f. ghazi and m.h. al-mayahy admet & dmpk 10(3) (2022) 213-230 218 p. 𝐼. 𝐼. = ∑ erythema grade at 1 h,24 h and 72 h + ∑ edema grade at 1 h,24 h and 72 h number of rats × number of application skin sites × time of reading (2) draize dermal scoring criteria were used to evaluate the irritancy potential of the mn arrays, as described in table 2. additionally, a traditional hypodermic syringe of gauge 23 g was also used for the comparison with mns for producing erythema and edema. these irritation signs were evaluated at 1, 24 and 72 hours using a digital microscope. table 2. draize dermal scoring criteria [27]: statistical analysis was carried out using one-way anova followed by the tukey test to determine the presence of any significant difference among the data. all data are presented as the mean ± sd with p values of ≤ 0.05 being regarded as statistically significant [5]. results and discussion preparation, visual inspection and microscopic examination of levothyroxine sodium loaded hyaluronic acid microneedle arrays lt-4 sodium-loaded ha mns were prepared using micromoulding technique. ha was used as a matrix polymer in this study for reasons other than its biocompatibility and biodegradability; the manufacturing of ha mns does not necessitate extreme conditions such as high temperatures or photolithography. moreover, the application of ha dissolving mns can improve the transport of active agents into and through the skin while also delivering ha into the skin itself as an added advantage [22]. as a result, the chosen micromoulding method with ha as a matrix polymer is ideally suited for incorporating heat and photosensitive molecules such as lt-4 sodium into mns [28]. not only does the mns matrix material affect the effectiveness of the mns, but it also does the geometry of mns. in case the needle arrays were too close together, a "bed of nails" situation occurs, resulting in penetration failure [29]. furthermore, numerous publications in the literature have linked the shape of needles ranging from conical, octagonal, rectangular, cylindrical and pyramidal to the required penetration force [30]. typically, a pyramidal shape exerted the greatest puncture strength, making application easier [5]. the pyramidal design, according to laue et al., enabled increased drug release by expanding the contact area with skin [31]. in this study, ha was used in different ratios ranging from 5 % to 60 % w/v to investigate the appropriate percentage for producing mns with sufficient mechanical strength required for effective penetration into the skin. the drying time is also a crucial consideration in mns manufacturing affecting the needle shape and the mechanical strength [32]. the drying time varies relatively depending on the concentration of ha. in general, 24 hours were required to obtain completely dry mn arrays in this study which is a suitable duration of time to be fitted with the large-scale production. total dryness of the mns was required to guarantee adequate mechanical strength [23]. on the other hand, inappropriate drying conditions led to wet mn arrays, which bent when removed from the mould resulting in unsatisfactory skin penetration capabilities. tween 80 was added to enhance the solubility of lt4 sodium. in addition to enhancing solubility, tween 80 plays a role in p.i.i. classification 0.0 0.4 no irritation 0.5 1.9 slight irritation 2.0 4.9 moderate irritation 5.0 8.0 sever irritation admet & dmpk 10(3) (2022) 213-230 transdermal delivery of levothyroxine via dissolving microneedles doi: http://dx.doi.org/10.5599/admet.1317 219 providing the formulation with the required elasticity necessary for the easy removal of mns from the mould without any fracture. formulations of f1 and f2 with a smaller concentration 5 % w/v of ha containing 1 % v/v and 2 % v/v of tween 80, respectively, resulted in soft, elastic formulations with a gummy texture that are unsuitable for the mns formation. formulas from f3 to f11 have resulted in the formation of elegant shaped mn arrays that are fully formed. a digital microscope was used for the visualization of the fabricated mns as illustrated in figure 1. f12 and f13 formulas containing 55 % and 60 % w/v of ha, respectively, did not result in fully formed mns. this can be attributed to the polymeric solution used, which was too viscous to fill the micronsized cavities in the mould and therefore a limited number of needles were eventually formed. figure 1. digital microscope images of mn arrays prepared using different ratios of ha, where (a) and (b) represent f1, (c) f2, (d) f3, (e) f4, (f) f5, (g) f6, (h) f7, (i) f8, (j) f9, (k) f10, (l) f11, (m) f12 and (n) f13. characterisation of ha microneedle arrays containing lt-4 sodium measurement of the weight and thickness of the microneedle arrays. the mn arrays demonstrated a weight range of (4 55 mg) depending largely on the amount of ha added. the thickness values were in the range of (1.01 – 1.04 mm). thus, the mn arrays possess a lightweight and delicate thickness suitable for convenient use by the patients. detailed results regarding the weight and thickness of each formula are shown in the supplementary materials. assessment of the mechanical properties of microneedle arrays. to allow consistent application and drug delivery, mns must be inserted into the skin without failure [33]. the mechanical strength of the prepared http://dx.doi.org/10.5599/admet.1317 r.f. ghazi and m.h. al-mayahy admet & dmpk 10(3) (2022) 213-230 220 dissolving mns was determined by evaluating their resistance to compression when an axial force was applied. this is a commonly used method to evaluate the mechanical properties of mns. the applied axial force was 32n, which refers to the maximum force applied by a human during the manual insertion of mns [18]. the results revealed a reduction in post-compression height, bending in certain needles, and needle fracture in some formulas. f1 and f2 resulted in the formation of delicate mns, that were easily broken when moved before subjection to compression. f3-f7 showed needle fracture upon exposure to compression. f8, f9, f13 and f11 revealed mns bending rather than fracturing. this refers to their ability to withstand the applied force better than the rest of the formulas. a higher concentration of ha polymer has resulted in a relatively stronger mn array. f11 was the most successful formulation yielding completely formed and strong mn arrays with the lowest height reduction percentage. the digital microscope images of mn arrays after exposure to compression using a texture analyser instrument were provided in the supplementary materials. the degree of height reduction of mns for different formulations was varied considerably since it is largely reliant on its unique components. the percentage of height reduction after compression for each mn array formula is shown in table 3. the percentage of height reduction was observed to be inversely proportional to the concentration of the ha used. the mechanical strength increased by adding more polymer and an enhancement of elasticity was achieved by further addition of tween 80. in addition to the role of tween 80 in enhancing the solubility of lt-4 sodium, plasticizers like tween 80 are also commonly used to minimize the intramolecular bonding between polymer chains. this provides the desired mechanical film qualities by boosting flexibility and preventing cracking, which is frequently employed in the transdermal drug delivery field [34]. the viscosity of the polymeric solution is proportional to the amount of polymer utilized. the higher the concentration of polymer, the more viscous the polymeric solution is leading to the formation of stronger mns. it is critical to choose the right polymer concentration for microneedle shaft manufacturing; the thickness of the baseplate that supports the microneedles has been diminished by decreasing the polymer content. as a result, the needles were subjected to an extra-axial applied force which explains why mechanical strength decreased by reducing polymer concentration [35]. the mechanical strength evaluation and insertion study are virtually linked. since they are related, the crucial aspect of efficient insertion is based on the ability of the mns to endure the applied pressure [36]. the standard mn array height was found to be 432 ± 6.4 µm and it was used for the comparison of height reduction with other arrays subjected to compression. in vitro drug release study. in vitro cumulative drug release percentages for the chosen formulas (f8, f9, f10 and f11) over 1 hour are illustrated in figure 2. the graph demonstrates that the release profile from the different formulations is almost similar, with a non-significant difference between them (p > 0.05). the release of lt-4 sodium was initiated within the first 5 minutes as the mns’ tips became hydrated and began to dissolve by contact with water. after that, the dissolution of mns increased with a simultaneous increase in lt-4 sodium release over an hour. during this hour, the mn arrays were diminished gradually until they eventually vanished after one hour. although there is a non-significant difference in the release profiles between different formulations, it can be observed that the release of lt-4 sodium from f8 and f9 was relatively higher than f10 and f11. this may be due to the presence of a higher ratio of ha in f10 and f11. after 1 hour, the release of lt-4 sodium from different mn arrays has reached 96-98 %. admet & dmpk 10(3) (2022) 213-230 transdermal delivery of levothyroxine via dissolving microneedles doi: http://dx.doi.org/10.5599/admet.1317 221 table 3. mean height, height percentage and height reduction percentage after compression for each mn array formula. data is presented as the mean ± sd formula code mean height (µm) after compression height % after compression height reduction % f3 277.65 ± 4 67.80 ± 4 32.24 ± 2 f4 292.93 ± 2 68.70 ± 8 31.30 ± 4 f5 326.17 ± 5 75.50 ± 4 24.51 ± 6 f6 349.56 ± 6 80.91 ± 5 19.09 ± 8 f7 367.44 ± 4 85.05 ± 4 14.95 ± 3 f8 382.44 ± 8 88.52 ± 2 11.48 ± 2 f9 399.72 ± 7 92.52 ± 3 7.48 ± 5 f10 402.76 ± 8 93.23 ± 5 6.77 ± 6 f11 411.22 ± 4 95.18 ± 2 4.82 ± 4 the amount of the drug released from the four mn arrays has reached 20-25 % five minutes following the start of the test. this might be attributed to the drug adsorption on the surface of the polymeric mns, besides to the hydrophilic property of ha polymer that will quickly dissolve in water and release drugs [37]. following the complete dissolution of the needles’ tip within 5 minutes, a slower rhythm of drug release has been noticed. this is because the backing layer may act as a reservoir for drug release. figure 2. lt-4 sodium release profile from f8, f9, f10 and f11 mn arrays. detection of microneedles insertion ability using parafilm m®. eight layers of parafilm m® were employed to mimic the skin layers in order to assess the depth of penetration of mns in the skin. as reported by larrañeta et al., penetration is regarded as efficient if greater than 20 % of micro conduits are created in each layer [18]. thus, the insertion test findings indicate that mns are located between the third and fourth layers, as shown in figure 3. moreover, the thickness of one parafilm m® layer is approximately 126 ± 7 µm, implying that the mns were penetrated up to 378 µm depth (approximately 87 % of needles height). as demonstrated in figures 3 http://dx.doi.org/10.5599/admet.1317 r.f. ghazi and m.h. al-mayahy admet & dmpk 10(3) (2022) 213-230 222 and 4, f11 mn arrays showed 100 % penetration in the first and second layers and 50 % in the third layer. this indicates their adequate mechanical strength with the sharpness of the tips required for efficient penetration across the skin tissue barrier. f8, f9 and f10 have exhibited a lower number of micro holes, especially in the third layer as illustrated in figure 4, proposing lower mechanical properties probably due to the lower concentration of ha polymer contained in their mn arrays. figure 3. insertion test using parafilm m® layers by f11 mn array that clearly shows the successful penetration through the 3rd layer. figure 4. illustration of the insertion ratio of the mn arrays (f8, f9, f10 and f11) created in parafilm m® sheets. insertion of microneedle arrays into excised human skin. to assess the mns’ effectiveness in the penetration of the uppermost layer of the skin (stratum corneum), an insertion study using human skin was performed. figure 5 shows the human skin stained with methylene blue dye after treatment with ha mn arrays. the pores created by the mns were stained with the dye, whereas the intact skin remained unstained. the 10 × 10 mn arrays with varying concentrations of ha (f8, f9, f10 and f11) resulted in a 75 to 100 % insertion ratio after 1 minute of application of mn arrays. these findings highlight the capability of the mn arrays to penetrate the skin effectively [32]. the insertion ratio for each sample was estimated by dividing the number of blue spots on the skin after insertion by the number of needles in the arrays [16]. the insertion ratio of mn arrays was increased with an increase in their ha content. this can be related to the higher mechanical strength achieved by increasing the concentration of ha in each formula. consequently, mn formula f11 demonstrated a 100 % insertion ratio due to their higher ha concentration (50 % w/v), resulting in the greatest mechanical strength required for efficient skin penetration. mn arrays were inserted into the skin using gentle thumb pressure. different types of applicators have been designed to minimize variation in insertion forces for the same mn array [38,39]. when appropriately engineered, the consistency of the applied force can be better with an applicator than with a thumb. yet, inserting mns into skin using a thumb force has been a typical technique of administration and multiple groups have clearly confirmed uniform and full mn array insertion while using this approach [40-42]. 0 50 100 150 layer 1 layer 2 layer 3 in se rt io n r a ti o % f8 f9 f10 f11 admet & dmpk 10(3) (2022) 213-230 transdermal delivery of levothyroxine via dissolving microneedles doi: http://dx.doi.org/10.5599/admet.1317 223 figure 5. digital microscope images of human skin following the application of ha mn arrays and staining with methylene blue dye showing the pores formed by different formulas, where (a) represents f11, (b) f10, (c) f9 and (d) f8. characterisation of the optimum formula ha mn array formula f11 was identified as the optimum formula because it showed the required mechanical strength, high drug content (data not shown) and superior penetration into the skin surface. (i.e. highest insertion capability), in addition to its ease of formulation during the fabrication process and elegant appearance. therefore, the f11 formula was subjected to further investigations to ensure its eligibility as a transdermal drug delivery system for lt-4 sodium. histological examination of mn treated skin to demonstrate the penetration depth of the microchannels within skin layers achieved by the application of mn arrays, histological examination of the cross-sectioned and stained human skin samples was performed as shown in figures 6 (a) and (b). it can be observed that f11 mn arrays following the insertion into the skin have developed drug microchannels (drug permeation routes) through the different skin layers, including the stratum corneum, viable epidermis and dermis. these microchannels formed in the skin showed approximately similar shapes to the inserted mns arrays indicating their uniform insertion. it is believed that these microchannels play a significant role in the permeation of lt-4 sodium into the systemic circulation by bypassing the stratum corneum barrier. in addition, figure 6 (a) highlights the difference in size between pores created by the mn arrays and by the traditional hypodermic syringe. the pores formed by hypodermic needles are obviously larger, deeper reaching the deep dermis and having a greater opportunity to stimulate the nerve endings than those formed by mns. ultimately, these observations suggested that f11 mn arrays have the potential for successful transdermal delivery of lt-4 sodium across the skin with a little tendency to trigger the nerve endings located at the deep dermis layer [43]. in vivo dissolution of mn arrays. the images of the in vivo dissolution study of f11 mn arrays within rat skin are illustrated in figure 7. as depicted in the images, the dissolution of the mn arrays was very fast, starting 5 minutes after insertion into the skin, with complete dissolution achieved by 1 hour. the fast dissolution of f11 mns suggests rapid delivery of lt-4 sodium across the skin layers. http://dx.doi.org/10.5599/admet.1317 r.f. ghazi and m.h. al-mayahy admet & dmpk 10(3) (2022) 213-230 224 it has been observed that ha mns possess self-dissolving properties and are easily dissolved upon application to the skin [20]. this rapid dissolution is caused by the high hydrophilicity of ha [44]. in their research study, zhuangzhi zhu et al., have shown that ha is a suitable candidate for dissolving mns, as it starts dissolving rapidly upon application into the skin, proposing that it required 60 120 minutes for their needles to be completely dissolved in the skin [44]. figure 6. digital microscope images of cross-sectioned human skin samples stained with h&e for mn treated skin (a) at power 4x and (b) at power 10x; demonstrating the location of microchannels within the skin in comparison to the hypodermic needle. where: (sc) represents stratum corneum, (ve) viable epidermis and (d) dermis. figure 7. time-course images of f11 mn arrays after being applied to the skin of a live rat and withdrawn at the indicated time points. admet & dmpk 10(3) (2022) 213-230 transdermal delivery of levothyroxine via dissolving microneedles doi: http://dx.doi.org/10.5599/admet.1317 225 visualization of microneedles by scanning electron microscopy (sem) the sem images of f11 mn array are demonstrated in figure 8. the images highlight the proper formation of sharp, pyramidal-shaped and equally interspaced mns. additionally, the images elucidate the dimensions of mns. for example, an individual mn from the array showed a mean length of 432.0 ± 6.4 μm. the needle tip diameter of 9.8 ± 1.3 μm, base diameter of 288 ± 2.2 μm and interspacing distance (the distance between two needle tips) of 500.0 ± 5.5 μm. based on these dimensions, the mns possess sharp tips with a strong base required for successful penetration into the skin [32]. according to the sem images, the prepared ha mn by micromoulding technique were shorter than the master structure (pdms mould), which is 500 μm needle length. this can occur due to the difference in polymer filling capability into the mould, which is based on the polymer viscosity and the shrinkage of polymeric solution during drying [45]. it is anticipated that this reduction in the length of mns does not greatly influence the transdermal delivery of lt-4 sodium since they showed the excellent formation of microchannels through the different skin layers, including the dermis where the drug can be systemically absorbed. figure 8. sem images of ha mn array f11, where: (a) top view of mn array at a magnification of 200x, (b) a single array of pyramidal-shaped mns shows the distance between two needles at a magnification of 100x, (c) top view of a single needle in the array showing the base width at a magnification of 300x, and (d) lateral view of mns showing the needle length at a magnification of 200x. ex vivo permeation study the transdermal delivery of lt-4 sodium is limited due to its low permeability characteristics through the stratum corneum layer. mns may enable lt-4 sodium delivery through the skin by physically disrupting the stratum corneum barrier and transferring molecules through the micro-channels into the dermis microcirculation. therefore, to investigate the influence of mns in enhancing transdermal delivery of lt-4 sodium through the skin, an ex vivo permeation study using human skin was carried out. http://dx.doi.org/10.5599/admet.1317 r.f. ghazi and m.h. al-mayahy admet & dmpk 10(3) (2022) 213-230 226 ha mn array formula f11 was used in this permeation study since it is regarded as the optimum formula, as shown previously. in addition, to the use of f11 mn arrays in the permeation study, control formulas of lt-4 sodium as an aqueous solution and as a polymeric film without needles (needle-free) were utilized for comparison purposes. the cumulative amount of lt-4 sodium permeated from mn array f11, drug solution and the polymeric film is demonstrated in figure 9. as can be observed from the graph, the microneedle array f11 has significantly (p ≤ 0.05) increased lt-4 sodium permeation through the skin (cumulative permeated amount of 32 ± 2 µg/cm²) in comparison to solution control (cumulative permeated amount of 0.7 ± 0.07 µg/cm²) and the polymeric film (0.1 ± 0.02 µg/cm²) over the period of 7 hours. the mns arrays have delivered 64 % ± 4 of their loading dose over 7 hours period, whereas the control preparations of aqueous solution and polymeric film were only able to deliver 1.4 % ± 0.14 and 0.2 % ± 0.04 of total loading dose, respectively, during the same period. this represents an approximately 46-320 fold increase in delivery of lt-4 sodium than that was achieved with the control preparations (aqueous solution and polymeric film, respectively). moreover, as the drug was encapsulated within mn arrays, the lag time was less since the drug began to diffuse quickly upon application of mns. in this study, the ha mns were started to dissolve within 5 minutes of application into the skin, supporting the shortening in lag time. the lag time for the mn arrays was only 10 minutes, while the drug solution (control) lag time was 44 minutes and the polymeric film lag time was 120 minutes. this enhancement of lt-4 sodium delivery is related to the ability of mns to bypass the stratum corneum barrier by creating microchannels through the skin to facilitate drug permeation. in addition, the hydration of the drug reservoir in the polymer matrix is enhanced by the fluid withdrawn from the skin via the microchannels leading to an improvement in drug diffusion [40]. the results obtained from this permeation study suggest the successful application of ha mn arrays in the delivery of lt-4 sodium. this can be considered a promising approach when used in humans since it may improve lt-4 sodium bioavailability. figure 9. the cumulative amount of lt-4 sodium permeated across human skin in µg/cm² from ha mn arrays f11 in comparison to control preparations (aqueous drug solution and polymeric film). data is presented as the mean ± sd (n = 6). admet & dmpk 10(3) (2022) 213-230 transdermal delivery of levothyroxine via dissolving microneedles doi: http://dx.doi.org/10.5599/admet.1317 227 skin irritation test. in the skin irritation test, the scores for erythema and edema from the skin sites treated with mn arrays were evaluated for the rats at 1, 24 and 72 hours [46]. as shown in figure 10, slight erythema but no edema appeared at the application sites at one hour and disappeared within 24 hours in comparison with a hypodermic needle. the p.i.i. of the mn arrays was calculated to be 1.7. the slight redness that was seen on the skin site one hour following the removal of the mns, suggests that the physical compression during the application of mns may cause this skin redness. however, this redness disappeared 24 hours and 72 hours after the application of mns. therefore, these findings and according to draize dermal scoring criteria indicated that the irritation and skin damage caused by ha mn arrays were slight [27]. whereas the hypodermic syringe has resulted in bleeding and a dramatic scar which was obvious even 72 hours following the application of the hypodermic needle. figure 10. images of skin irritation following the application of (a) ha mn arrays f11 and (b) conventional hypodermic syringe on rat skin at 1 h, 24 h and 72 h. conclusions the current study revealed for the first time the potential of ha dissolving mn arrays for efficient transdermal delivery of lt-4 sodium by overcoming the stratum corneum barrier. additionally, this work clearly demonstrates the benefits of ha mn arrays in enhancing skin permeation of drugs with poor permeability characteristics. the lt-4 sodium loaded ha mn arrays represent a minimally invasive system that showed excellent skin insertion with rapid dissolution and release of lt-4 sodium and superior transdermal delivery in comparison to control preparations. thus, they may provide a feasible and convenient alternative approach for oral administration, which is accompanied by variable rate and extent of lt-4 sodium absorption. however, further in vivo studies are required to confirm the mn arrays’ efficacy and safety for clinical use. acknowledgements: the authors would like to thank mustansiriyah university (www.uomustansiriyah.edu.iq) baghdad-iraq for its support in the present work. funding: this research did not receive any specific grant from funding agencies in the public, commercial, or http://dx.doi.org/10.5599/admet.1317 http://www.uomustansiriyah.edu.iq/ r.f. ghazi and m.h. al-mayahy admet & dmpk 10(3) (2022) 213-230 228 not-for-profit sectors. conflict of interest: the authors declare no conflict of interest. references [1] m. skelin, t. lucijanic, d. amidzic klaric, a. resic, m. bakula, a.m. liberati-cizmek, h. gharib, d. rahelic. factors affecting gastrointestinal absorption of levothyroxine: a review. clin. ther. 39 (2017) 378-403. https://doi.org/10.1016/j.clinthera.2017.01.005. [2] j. jonklaas, a.c. bianco, a.j. bauer, k.d. burman, a.r. cappola, f.s. celi, d.s. cooper, b.w. kim, r.p. peeters, m.s. rosenthal, a.m. sawka, r. american thyroid association task force on thyroid hormone. guidelines for the treatment of hypothyroidism: prepared by the american thyroid association task force on thyroid hormone replacement. thyroid 24 (2014) 1670-1751. https://doi.org/10.1089/thy.2014.0028. [3] k.b. ain, s. refetoff, h.g. fein, b.d. weintraub. pseudomalabsorption of levothyroxine. jama 266 (1991) 2118-2120. https://doi.org/10.1001/jama.1991.03470150090036. [4] m.r. prausnitz. microneedles for transdermal drug delivery. adv. drug deliv. rev. 56 (2004) 581-587. https://doi.org/10.1016/j.addr.2003.10.023. [5] d.v. mcallister, p.m. wang, s.p. davis, j.h. park, p.j. canatella, m.g. allen, m.r. prausnitz. microfabricated needles for transdermal delivery of macromolecules and nanoparticles: fabrication methods and transport studies. proc. natl. acad. sci. usa 100 (2003) 13755-13760. https://doi.org/10.1073/pnas.2331316100. [6] k. ita. transdermal delivery of drugs with microneedles—potential and challenges. pharmaceutics 7 (2015) 90-105. https://doi.org/10.3390/pharmaceutics7030090. [7] y. chen, b.z. chen, q.l. wang, x. jin, x.d. guo. fabrication of coated polymer microneedles for transdermal drug delivery. j. control. release 265 (2017) 14-21. https://doi.org/10.1016/j.jconrel.2017.03.383. [8] k. ita. dissolving microneedles for transdermal drug delivery: advances and challenges. biomed. pharmacother. 93 (2017) 1116-1127. https://doi.org/10.1016/j.biopha.2017.07.019. [9] m.s. lhernould. optimizing hollow microneedles arrays aimed at transdermal drug delivery. microsystem technologies 19 (2013) 1-8. [10] a.a. sabeeh, a..m. hussain. the growing role of hydrogel microneedles in transdermal drug delivery. international journal of drug delivery technology 11 (2021) 611-616. https://doi.org/10.25258/ijddt.11.2.67. [11] s. dhivya, j. ajita, n. selvamurugan. metallic nanomaterials for bone tissue engineering. j. biomed. nanotechnol. 11 (2015) 1675-1700. https://doi.org/10.1166/jbn.2015.2115. [12] s. chen, x. hao, x. liang, q. zhang, c. zhang, g. zhou, s. shen, g. jia, j. zhang. inorganic nanomaterials as carriers for drug delivery. journal of biomedical nanotechnology 12 (2016) 1-27. https://doi.org/10.1166/jbn.2016.2122. [13] w. yu, g. jiang, y. zhang, d. liu, b. xu, j. zhou. polymer microneedles fabricated from alginate and hyaluronate for transdermal delivery of insulin. mater. sci. eng. c mater. biol. appl. 80 (2017) 187196. https://doi.org/10.1016/j.msec.2017.05.143. [14] x. jin, d.d. zhu, b.z. chen, m. ashfaq, x.d. guo. insulin delivery systems combined with microneedle technology. adv. drug deliv. rev. 127 (2018) 119-137. https://doi.org/10.1016/j.addr.2018.03.011. [15] n.b. pajić, s. vucen, t. ilić, c. o'mahony, v. dobričić, s. savić. comparative efficacy evaluation of different penetration enhancement strategies for dermal delivery of poorly soluble drugs–a case with sertaconazole nitrate. european journal of pharmaceutical sciences 164 (2021) 105895. https://doi.org/10.1016/j.ejps.2021.105895. [16] i.c. lee, j.s. he, m.t. tsai, k.c. lin. fabrication of a novel partially dissolving polymer microneedle patch for transdermal drug delivery. j. mater. chem. b 3 (2015) 276-285. https://doi.org/10.1039/c4tb01555j. https://doi.org/10.1016/j.clinthera.2017.01.005 https://doi.org/10.1089/thy.2014.0028 https://doi.org/10.1001/jama.1991.03470150090036 https://doi.org/10.1016/j.addr.2003.10.023 https://doi.org/10.1073/pnas.2331316100 https://doi.org/10.3390/pharmaceutics7030090 https://doi.org/10.1016/j.jconrel.2017.03.383 https://doi.org/10.1016/j.jconrel.2017.03.383 https://doi.org/10.1016/j.biopha.2017.07.019 https://doi.org/10.25258/ijddt.11.2.67 https://doi.org/10.1166/jbn.2015.2115 https://doi.org/10.1166/jbn.2016.2122 https://doi.org/10.1016/j.msec.2017.05.143 https://doi.org/10.1016/j.addr.2018.03.011 https://doi.org/10.1016/j.ejps.2021.105895 https://doi.org/10.1039/c4tb01555j https://doi.org/10.1039/c4tb01555j admet & dmpk 10(3) (2022) 213-230 transdermal delivery of levothyroxine via dissolving microneedles doi: http://dx.doi.org/10.5599/admet.1317 229 [17] s. khan, m.u. minhas, i.a. tekko, r.f. donnelly, r.r.s. thakur. evaluation of microneedles-assisted in situ depot forming poloxamer gels for sustained transdermal drug delivery. drug deliv. transl. res. 9 (2019) 764-782. https://doi.org/10.1007/s13346-019-00617-2. [18] e. larrañeta, j. moore, e.m. vicente-pérez, p. gonzález-vázquez, r. lutton, a.d. woolfson, r.f. donnelly. a proposed model membrane and test method for microneedle insertion studies. international journal of pharmaceutics 472 (2014) 65-73. https://doi.org/10.1016/j.ijpharm.2014.05.042. [19] c.h. chen, v.b. shyu, c.t. chen. dissolving microneedle patches for transdermal insulin delivery in diabetic mice: potential for clinical applications. materials (basel) 11 (2018) 1625. https://doi.org/10.3390/ma11091625. [20] s. liu, m.-n. jin, y.-s. quan, f. kamiyama, h. katsumi, t. sakane, a. yamamoto. the development and characteristics of novel microneedle arrays fabricated from hyaluronic acid, and their application in the transdermal delivery of insulin. journal of controlled release 161 (2012) 933-941. https://doi.org/10.1016/j.jconrel.2012.05.030. [21] m. mir, a.d. permana, n. ahmed, g.m. khan, a. ur rehman, r.f. donnelly. enhancement in sitespecific delivery of carvacrol for potential treatment of infected wounds using infection responsive nanoparticles loaded into dissolving microneedles: a proof of concept study. european journal of pharmaceutics and biopharmaceutics 147 (2020) 57-68. https://doi.org/10.1016/j.ejpb.2019.12.008. [22] h.x. nguyen, b.d. bozorg, y. kim, a. wieber, g. birk, d. lubda, a.k. banga. poly (vinyl alcohol) microneedles: fabrication, characterization, and application for transdermal drug delivery of doxorubicin. eur. j. pharm biopharm 129 (2018) 88-103. https://doi.org/10.1016/j.ejpb.2018.05.017. [23] m.s. lhernould, m. deleers, a. delchambre. hollow polymer microneedles array resistance and insertion tests. int. j. pharm. 480 (2015) 152-157. https://doi.org/10.1016/j.ijpharm.2015.01.019. [24] m.j. kim, s.c. park, s.-o. choi. dual-nozzle spray deposition process for improving the stability of proteins in polymer microneedles. rsc advances 7 (2017) 55350-55359. https://doi.org/10.1039/c7ra10928h. [25] h. du, p. liu, j. zhu, j. lan, y. li, l. zhang, j. zhu, j. tao. hyaluronic acid-based dissolving microneedle patch loaded with methotrexate for improved treatment of psoriasis. acs appl. mater. interfaces 11 (2019) 43588-43598. https://doi.org/10.1021/acsami.9b15668. [26] g. yao, g. quan, s. lin, t. peng, q. wang, h. ran, h. chen, q. zhang, l. wang, x. pan, c. wu. novel dissolving microneedles for enhanced transdermal delivery of levonorgestrel: in vitro and in vivo characterization. int. j. pharm. 534 (2017) 378-386. https://doi.org/10.1016/j.ijpharm.2017.10.035. [27] s. liu, m.-n. jin, y.-s. quan, f. kamiyama, k. kusamori, h. katsumi, t. sakane, a. yamamoto. transdermal delivery of relatively high molecular weight drugs using novel self-dissolving microneedle arrays fabricated from hyaluronic acid and their characteristics and safety after application to the skin. european journal of pharmaceutics and biopharmaceutics 86 (2014) 267-276. https://doi.org/10.1016/j.ejpb.2013.10.001. [28] b. saha, w.q. toh, e. liu, s.b. tor, d.e. hardt, j. lee. a review on the importance of surface coating of micro/nano-mold in micro/nano-molding processes. journal of micromechanics and microengineering 26 (2015) 013002. https://doi.org/10.1088/0960-1317/26/1/013002. [29] s.d. gittard, a. ovsianikov, b.n. chichkov, a. doraiswamy, r.j. narayan. two-photon polymerization of microneedles for transdermal drug delivery. expert opin. drug deliv. 7 (2010) 513-533. https://doi.org/10.1517/17425241003628171. [30] s.m. bal, a.c. kruithof, r. zwier, e. dietz, j.a. bouwstra, j. lademann, m.c. meinke. influence of microneedle shape on the transport of a fluorescent dye into human skin in vivo. journal of controlled release 147 (2010) 218-224. https://doi.org/10.1016/j.jconrel.2010.07.104. [31] s. lau, j. fei, h. liu, w. chen, r. liu. multilayered pyramidal dissolving microneedle patches with flexible pedestals for improving effective drug delivery. j. control. release 265 (2017) 113-119. https://doi.org/10.1016/j.jconrel.2016.08.031. http://dx.doi.org/10.5599/admet.1317 https://doi.org/10.1007/s13346-019-00617-2 https://doi.org/10.1016/j.ijpharm.2014.05.042 https://doi.org/10.1016/j.ijpharm.2014.05.042 https://doi.org/10.3390/ma11091625 https://doi.org/10.3390/ma11091625 https://doi.org/10.1016/j.jconrel.2012.05.030 https://doi.org/10.1016/j.ejpb.2019.12.008 https://doi.org/10.1016/j.ejpb.2018.05.017 https://doi.org/10.1016/j.ijpharm.2015.01.019 https://doi.org/10.1039/c7ra10928h https://doi.org/10.1039/c7ra10928h https://doi.org/10.1021/acsami.9b15668 https://doi.org/10.1016/j.ijpharm.2017.10.035 https://doi.org/10.1016/j.ejpb.2013.10.001 https://doi.org/10.1088/0960-1317/26/1/013002 https://doi.org/10.1517/17425241003628171 https://doi.org/10.1016/j.jconrel.2010.07.104 https://doi.org/10.1016/j.jconrel.2016.08.031 r.f. ghazi and m.h. al-mayahy admet & dmpk 10(3) (2022) 213-230 230 [32] y. kim, s.a. bhattaccharjee, m. beck-broichsitter, a.k. banga. fabrication and characterization of hyaluronic acid microneedles to enhance delivery of magnesium ascorbyl phosphate into skin. biomedical microdevices 21 (2019) 1-9. https://doi.org/10.1007/s10544-019-0455-0. [33] s.p. davis, b.j. landis, z.h. adams, m.g. allen, m.r. prausnitz. insertion of microneedles into skin: measurement and prediction of insertion force and needle fracture force. journal of biomechanics 37 (2004) 1155-1163. https://doi.org/10.1016/j.jbiomech.2003.12.010. [34] m.g.a. vieira, m.a. da silva, l.o. dos santos, m.m. beppu. natural-based plasticizers and biopolymer films: a review. european polymer journal 47 (2011) 254-263. https://doi.org/10.1016/j.eurpolymj.2010.12.011. [35] q.l. wang, d.d. zhu, x.b. liu, b.z. chen, x.d. guo. microneedles with controlled bubble sizes and drug distributions for efficient transdermal drug delivery. sci. rep. 6 (2016) 38755. https://doi.org/10.1038/srep38755. [36] d.f.s. fonseca, c. vilela, a.j.d. silvestre, c.s.r. freire. a compendium of current developments on polysaccharide and protein-based microneedles. int. j. biol. macromol. 136 (2019) 704-728. https://doi.org/10.1016/j.ijbiomac.2019.04.163. [37] k. saraswathy, g. agarwal, a. srivastava. hyaluronic acid microneedles‐laden collagen cryogel plugs for ocular drug delivery. journal of applied polymer science 137 (2020) 49285. https://doi.org/10.1002/app.49285. [38] m.i. haq, e. smith, d.n. john, m. kalavala, c. edwards, a. anstey, a. morrissey, j.c. birchall. clinical administration of microneedles: skin puncture, pain and sensation. biomed. microdevices 11 (2009) 35-47. https://doi.org/10.1007/s10544-008-9208-1. [39] m. yang, j.d. zahn. microneedle insertion force reduction using vibratory actuation. biomedical microdevices 6 (2004) 177-182. https://doi.org/10.1023/b:bmmd.0000042046.07678.2e. [40] j.w. lee, j.h. park, m.r. prausnitz. dissolving microneedles for transdermal drug delivery. biomaterials 29 (2008) 2113-2124. https://doi.org/10.1016/j.biomaterials.2007.12.048. [41] h.x. nguyen, a.k. banga. enhanced skin delivery of vismodegib by microneedle treatment. drug deliv transl. res. 5 (2015) 407-423. https://doi.org/10.1007/s13346-015-0241-3. [42] h.x. nguyen, a.k. banga. delivery of methotrexate and characterization of skin treated by fabricated plga microneedles and fractional ablative laser. pharm. res. 35 (2018) 68. https://doi.org/10.1007/s11095-018-2369-6. [43] b.m. lee, c. lee, s.f. lahiji, u.w. jung, g. chung, h. jung. dissolving microneedles for rapid and painless local anesthesia. pharmaceutics 12 (2020) 366. https://doi.org/10.3390/pharmaceutics12040366. [44] z. zhu, h. luo, w. lu, h. luan, y. wu, j. luo, y. wang, j. pi, c.y. lim, h. wang. rapidly dissolvable microneedle patches for transdermal delivery of exenatide. pharm. res. 31 (2014) 3348-3360. https://doi.org/10.1007/s11095-014-1424-1. [45] y. park, j. park, g.s. chu, k.s. kim, j.h. sung, b. kim. transdermal delivery of cosmetic ingredients using dissolving polymer microneedle arrays. biotechnology and bioprocess engineering 20 (2015) 543-549. https://doi.org/10.1007/s12257-014-0775-0. [46] v. kumar, s. mahant, r. rao, s. nanda. emulgel based topical delivery system for loratadine. admet and dmpk 2 (2014) 254-271. https://doi.org/10.5599/admet.2.4.64. ©2022 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) https://doi.org/10.1007/s10544-019-0455-0 https://doi.org/10.1016/j.jbiomech.2003.12.010 https://doi.org/10.1016/j.eurpolymj.2010.12.011 https://doi.org/10.1016/j.eurpolymj.2010.12.011 https://doi.org/10.1038/srep38755 https://doi.org/10.1016/j.ijbiomac.2019.04.163 https://doi.org/10.1002/app.49285 https://doi.org/10.1007/s10544-008-9208-1 https://doi.org/10.1023/b:bmmd.0000042046.07678.2e https://doi.org/10.1016/j.biomaterials.2007.12.048 https://doi.org/10.1007/s13346-015-0241-3 https://doi.org/10.1007/s11095-018-2369-6 https://doi.org/10.1007/s11095-018-2369-6 https://doi.org/10.3390/pharmaceutics12040366 https://doi.org/10.3390/pharmaceutics12040366 https://doi.org/10.1007/s11095-014-1424-1 https://doi.org/10.1007/s12257-014-0775-0 https://doi.org/10.5599/admet.2.4.64 http://creativecommons.org/licenses/by/3.0/ anomalous salting-out, self-association and pka effects in the practically-insoluble bromothymol blue doi: https://doi.org/10.5599/admet.1793 1 admet & dmpk 0(0) (2023) 000-000; doi: https://doi.org/10.5599/admet.1793 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper green synthesis, characterization and biological activities of silver nanoparticles synthesized from neolamarkia cadamba juluri maheswari1, mohammed reshma anjum1, mohan sankari1, golla narasimha2, suresh babu naidu krishna3* and battini kishori1 1department of biotechnology, sri padmavati mahila visvavidyalayam (women’s university), tirupati517 502, a.p. india 2department of virology, sri venkateswara university, tirupati517 502, a.p. india 3institute for water and wastewater technology, durban university of technology, durban – 4000, south africa *corresponding authors: *kktinku@redifmail.com; sureshk@dut.ac.za; tel.: +27-31 3736015 received: march 28, 2023; revised: june 14, 2023; published: june xx, 2023 abstract background and purpose: metal nanoparticles are essential due to their unique catalytic, electrical, magnetic, and optical characteristics, as well as their prospective use in sensing, catalysis, and biological research. in recent years, researchers have focused on developing cost-effective and eco-friendly biogenic practices using the green synthesis of metal nanoparticles (agnp). experimental approach: in the present study, the aqueous extracts prepared from the leaf, stem, bark, and flower of neolamarkia cadamba were used for the synthesis of silver nanoparticles. synthesized silver nanoparticles were characterized using uvvisible spectroscopy, zeta potential, dynamic light scattering, scanning electron microscope (sem), and edax. key results: the current study showed absorption of synthesized agnps at 425, 423, 410, and 400 nm. dynamic light scattering of agnps showed size distribution of agnps synthesized from leaf, stem, and flower aqueous extracts ranges from 80-200 nm and agnps prepared from bark extract ranges from 100700 nm. zeta-potential of the biosynthesized agnps was found as a sharp peak at -23.1 mv for the leaf, 27.0 mv for the stem, -34.1 mv for the bark, and -20.2 mv for the flower. silver nanoparticles and crude extracts of neolamarkia cadamba showed effective antibacterial, antifungal, and antioxidant activities. conclusion: silver nanoparticles have substantial antibacterial activity against gram-positive bacteria and also exhibit the utmost antifungal activity against aspergillus niger. the study concludes that the green synthesis of silver nanoparticles from n. cadamba leaf, stem, bark, and flower extract is a reliable and ecofriendly technique. ©2023 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords antibacterial activity; antifungal activity; antioxidant activity; electron microscopy introduction metal nanoparticle synthesis has emerged in material science in recent years due to its applications in medicine, agriculture, drug delivery, environmental bioremediation, and information storage [1-5] due to its unique optical, catalytic, magnetic, electronic, and antimicrobial properties [6-10]. silver is preferred for the synthesis of nanoparticles due to its antibacterial and catalytic characteristics and is non-toxic to humans in comparison with other metals [11]. nanocrystalline silver particles are often used in high-sensitivity https://doi.org/10.5599/admet.1793 https://doi.org/10.5599/admet.1793 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:kktinku@redifmail.com mailto:sureshk@dut.ac.za http://creativecommons.org/licenses/by/4.0/ juluri maheswari et al. admet & dmpk 00(0) (2023) 000-000 2 biomolecular detection, therapeutics, catalysis, antimicrobials, and microelectronics [12-15]. silver nanoparticles (agnps) can be synthesized by physical, chemical, and biological methods. physical as well as chemical methods are exorbitant, time-consuming, and non-eco-friendly. they require a large quantity of energy, toxic solvents, and hazardous chemicals [16]. the biological method is inexpensive, energy-efficient, non-toxic, and eco-friendly [17]. fungi, bacteria, yeast, and plant sources are used in the synthesis of agnps, which is termed green synthesis [18,19]. green synthesis utilizing plants is common; it is safe, economical, and harmless to the environment [20]. neolamarkia cadamba is a member of the ruiaceae family. this plant is commonly referred to as 'kadam'. many indian groups utilize its bark, leaf, flower, stem, and root to treat a variety of diseases, such as fever, sore throat, cough, inflammation, and infections [21,22]. the pharmacological properties of neolamarkia cadamba involve anticancer [23], antiprotozoal [24], antidiabetic [25], antibacterial [26], antifungal [27], antioxidant [28], anti-inflammatory [29], wound healing [30] and anti-malarial activity [22]. in the present study, agnps were synthesized by eco-friendly green synthesis using aqueous leaf, stem, bark, and flower extracts of neolamarkia cadamba at ambient temperature. characterization of biosynthesized agnps was done by using uv-visible spectroscopy, zeta-potential, dynamic light scattering (dls), scanning electron microscopy (sem), and energy dispersive x-ray analysis (edax). antibacterial activity was tested on selective bacterial strains like staphylococcus aureus and bacillus subtilis, gram-positive, and pseudomonas aeruginosa and escherichia coli, gram-negative bacteria. the dpph method was used to determine antioxidant and antifungal activity against aspergillus niger. experimental collection of samples neolamarkia cadamba leaves, stems, bark, and flowers were collected in tirupati, chittoor district, andhra pradesh, india. the fresh plant parts were gathered in polyethylene zipper bags to eliminate filth, cleaned with tap water, and then distilled water. after that, the samples were shadow dried at room temperature. the samples were then crushed with an electric mixer and put in airtight polyethylene bottles until they could be looked at more closely (figure 1). figure 1. n. cadamba (a) leaf, (b) stem, (c) bark and (d) flower chemicals all the chemicals used were analytical grade. plant extract preparation aqueous extracts of the leaves, stem, bark, and flowers were prepared using distilled water (150 ml) in four different erlenmeyer flasks of 500 ml each and positive control in a separate flask, followed by heating at 70–80 °c in a water bath for 2-3 hours and later cooling at room temperature. the obtained extracts were centrifuged for 5 minutes at 3000 rpm. the recovered supernatant was filtered through whatman no. 1 filter paper using a buchner funnel. the filtrate was stored at 40 °c and used for further analysis. admet & dmpk 00(0) (2023) 000-000 silver nanoparticles synthesized from neolamarkia cadamba doi: https://doi.org/10.5599/admet.1793 3 silver nitrate stock solution to prepare 1 mm silver nitrate solution, 17 mg of silver nitrate was dissolved in 100 ml of distilled water and stored in an amber glass bottle. characterization of silver nanoparticles ten milliliters of agno3 (1 mm) solution was added to 1 ml of plant extract and made up to 15 ml using distilled water. the color shift of the solution from colorless to brown verified the reduction of ag+ to ag0. to eliminate errors caused by the high optical density of the solution, distilled water was added ten times to the particle suspension. uv-visible analysis the bio-reduction of pure ag+ ions was evaluated by sampling aliquots (0.5 ml) and diluting the samples with 5 ml of de-mineralized water, and the samples were further analyzed using uv–vis spectroscopy (perkinelmer lambda 25 spectrophotometer) [31]. absorption peak in the range of 350 to 500 nm revealed the existence and reduction of silver ions. dynamic light scattering (dls) dls (malvern instruments ltd., uk) was used to assess particle size distributions by monitoring dynamic variations in light scattering intensity caused by the brownian motion of the particles. the agnps were diluted and either filtered through a 0.22 μm syringe-driven filter or left unfiltered [17]. zeta potential measurements the physical attribute that determines the net surface charge of nanoparticles was defined as zeta potential. when the zeta potential values ranged from higher than +30 mv to less than -30 mv, the requirements for nanoparticle stability were evaluated [32]. the ph of the samples was then adjusted to the desired amount and vortexed for 30 minutes. after the vortex, the zeta potential was measured and the equilibrium ph was determined. an average of three distinct measures was reported in each case [33]. scanning electron microscopy (sem) analysis scanning electron microscopy characterizes the morphology and size of synthesized agnps. a mercury lamp was used to dry the films on the carbon-coated copper grid for 5 minutes. sem micrographs were taken on a zeiss evo 40 electron microscope [34]. edax measurements edax analysis was performed using a hitachi su6600 fe-sem fitted with edax attachment after drying the agnps on a carbon-coated copper grid at 25 °c. the x-rays were detected using a semiconductor material combined with circuits to analyze the spectrum [35]. antibacterial activity the crude extracts and agnps synthesized from aqueous extracts of n. cadamba were tested for antibacterial activity using the agar well diffusion technique against gram-positive bacteria s. aureus and b. subtilis, as well as gram-negative bacteria e. coli and p. aeruginosa. to do this, 100 μl of active culture inoculums were added to the nutrient agar plates in a clean way and spread out over the medium. a sterile borer was used to create 5 wells of 8 mm diameter on 4 mm thick nutrient agar plates. various concentrations of crude and agnps samples (20, 40, 60, and 80 μg/μl) were placed in four wells, with streptomycin (himedia) (10 μg/μl) serving as a positive control in one well. incubation was done on these agar plates at 37 °c for 24 https://doi.org/10.5599/admet.1793 juluri maheswari et al. admet & dmpk 00(0) (2023) 000-000 4 to 48 hours. the experiment was performed in duplicate to minimize errors. after incubation, each plate was examined, and zones of inhibition were measured using a ruler [36,37]. antifungal activity the crude extracts and agnps synthesized from aqueous extracts of various plant components, i.e., leaf, stem, flower, and bark, were investigated using the agar well diffusion method for antifungal activity. aspergillus niger (atcc 9029) was obtained from the microbial type culture collection in chandigarh, india. a. niger stock cultures were developed and placed at 4 °c. plates of potato dextrose agar was prepared and solidified. to inoculate 100 μl (0.1 ml) of fungal culture, the spread plate method was used. borers were used to drill wells, which were then filled with varying concentrations of crude extracts and synthesized silver nanoparticles, with one well serving as a control. these plates were incubated for 7 days at 26 ± 4 °c. after incubation, a clear zone was formed around the well, which was examined and measured with a ruler. antifungal activity was indicated by the formation of a clear zone [38,39]. antioxidant assay (dpph method) with minor modifications, the procedure was conducted by following the protocol of bhakya et al. [40]. using the stable radical dpph, the scavenging activity of free radicals from crude extracts and agnps, along with standard vitamin c, were measured. 1 ml of crude extracts and agnps at various concentrations (40, 80, 120, 160, 200, and 400 μg/μl) were mixed with freshly prepared 1 ml of dpph solution (0.1 mm in methanol) and vortexed vigorously. the solution was then incubated in the dark for 30 minutes at room temperature. the absorbance was measured at 517 nm using a uv-vis spectrophotometer. dpph (all reagents except the sample) was employed as a control, while methanol served as a blank solution. the scavenging activity of free radicals of agnps generated from different portions of plant extracts was represented as a percentage of inhibition, calculated using equation (1). % dpph radical scavenging activity = ac as/ac×100 (1) where ac is the control absorbance of dpph radical + methanol and as is the sample absorbance of dpph radical + sample agnps / vitamin c. results color change and uv-vis spectroscopy nanoparticles began to form after the extract was mixed with silver nitrate solution. the formation of agnps was confirmed by the observable changes in the solution's color, as shown in figure. 2, i.e., from pale yellow to brown. figure 2. colour change indicates the formation of agnps synthesized from aqueous extracts of: (a) leaf, (b) stem, (c) bark and (d) flower of n. cadamba admet & dmpk 00(0) (2023) 000-000 silver nanoparticles synthesized from neolamarkia cadamba doi: https://doi.org/10.5599/admet.1793 5 spectral analysis was another method of verifying developed agnps. the absorption of synthesized agnps from leaf, stem, bark, and flower aqueous extracts of n. cadamba (figure 3) was observed at 425, 423, 410, and 400 nm due to surface plasmon resonance (spr). wavelength in nm figure 3. uv–vis absorption spectra of agnps synthesized from aqueous extracts of leaf, stem, bark and flower of n. cadamba dynamic light scattering of silver nanoparticles current observation shows that the size distribution of agnps synthesized from leaf, stem and flower aqueous extracts ranges from 80–200 nm and that agnps prepared from bark extract range from 100 to 700 nm. the calculated z-average particle size distribution of synthesized agnps from leaf extract was 77.5 nm, stem extract was 80.9 nm, bark extract was 757.7 nm, and flower extract was 11559.6 nm, as shown in below figure 4. figure 4. dls showing average size of agnps synthesized from extracts of (a) leaf, (b) stem (c) bark (d) flower of n. cadamba zeta potential of silver nanoparticles the zeta potential of the biosynthesized agnps was found as sharp peak at -23.1 mv for the leaf, -27.0 mv for the stem, -34.1 mv for the bark and -20.2 mv for the flower, as shown in figure 5. https://doi.org/10.5599/admet.1793 juluri maheswari et al. admet & dmpk 00(0) (2023) 000-000 6 figure 5. zeta potential of agnps synthesized from extracts of (a) leaf (b) stem (c) bark and (d) flower of n. cadamba scanning electron microscopy (sem) figure 6 illustrates the sem images of synthesized silver nanoparticles from leaf, stem, bark, and flower aqueous extracts of n. cadamba. the results of sem concluded that agnps size and spherical form were not defined due to agglomeration. figure 6. sem micrographs of agnps synthesised from extracts of (a) leaf (b) stem (c) bark and (d) flower of n. cadamba admet & dmpk 00(0) (2023) 000-000 silver nanoparticles synthesized from neolamarkia cadamba doi: https://doi.org/10.5599/admet.1793 7 edx of silver nanoparticles the observations from energy dispersive spectroscopy show very high silver signals and weak chloride and carbon signals, indicating that the conversion of silver ions to silver elements may come from molecules linked on the surface of agnps. silver nitrate was reduced to silver nanoparticles, as evidenced by the dense peak of silver. the silver peak is significantly thicker than the others in the spectrum, as shown in figure 7. figure 7. edax spectra of agnps synthesized from extracts of (a) leaf (b) stem (c) bark (d) flower of n. cadamba antibacterial activity the aqueous crude and agnps extracts from leaf, stem, bark and flower of n. cadamba have potential biological activity towards antibacterial on four different bacterial strains like b. subtilis, s. aureus, e. coli and p. aeruginosa. both crude and agnps of n. cadamba showed different zone of inhibitions at concentrations of 20, 40, 60 and 80 μg/μl along with streptomycin as a positive control, as shown in figure 8. antifungal activity aqueous crude and agnps extracts of leaf, stem, bark and flower of n. cadamba showed antifungal activity against a. niger (figure 9) compared to crude extracts. antioxidant activity aqueous crude and silver nanoparticle extracts of leaf, stem, bark and flower of n. cadamba tested positive for antioxidant properties. it is confirmed that both crude and silver nanoparticle extracts synthesized from different parts of n. cadamba show antioxidant activity depending on the relative dosage. agnps prepared from different parts of n. cadamba showed effective antioxidant activity, as shown in figure 10 (a), compared to crude extracts, as shown in figure 10 (b). https://doi.org/10.5599/admet.1793 juluri maheswari et al. admet & dmpk 00(0) (2023) 000-000 8 figure 8. antibacterial activity of agnps synthesized from extracts of (a) leaf (b) stem (c) bark and (d) flower of n. cadamba figure 9. antifungal activity of agnps synthesized from (a) leaf (b) stem (c) bark and (d) flower of n. cadamba figure 10. antioxidant activity of (a) n. cadamba and (b) crude extracts. admet & dmpk 00(0) (2023) 000-000 silver nanoparticles synthesized from neolamarkia cadamba doi: https://doi.org/10.5599/admet.1793 9 discussion the present work describes the green synthesis of agnps with the help of aqueous extracts of the leaf, stem, bark, and flower of n. cadamba. green synthesis of agnps was popular due to the absence of harmful ingredients, low cost, eco-friendly, and suitable for biomedical and pharmaceutical applications [41]. the color of the silver nitrate solution was changed from pale yellow to brown after the aqueous extracts were added to it. the reduction of silver nitrate into agnps was confirmed by the development of a brown-colored solution [42,43]. the reduction of ag+ to ag0 is directly proportional to the concentration of crude extract, similar results were observed during the synthesis of agnps using olive leaf extract [44]. colloidal agnps were excited by absorbing light in the range of 400 to 450 nm due to surface plasmon resonance [45]. the excitation peak observed using a uv-vis spectrophotometer indicated the presence of agnps. compared to the strong spr peak obtained in the uv-vis spectra, the broad spectrum of the dls analyzer reveals that the particle size was smaller [46]. the surface of the nanoparticles was thought to be negatively charged and was disseminated in the medium. the negative values in zeta potential confirmed the repulsion and stability among the particles [47]. the observations from energy dispersive spectroscopy show very high silver signals and weak chloride and carbon signals [48], indicating the reduction of silver ions to elemental silver, which may have come from molecules linked to the agnps surface. the silver peak is significantly thicker than the other peaks [49]. recent studies revealed that most bacteria gained resistance to recurrently used antibiotics, necessitating the use of alternate medicines [50]. agnps showed effective antibacterial activity compared to crude extract against gram-positive and gramnegative bacteria, similar studies on cucumis prophetarum revealed that nanoparticles showed enhanced activity against bacterial pathogens compared to crude extracts [51,52]. the studies show less zone of inhibition against aspergillus niger at performed concentrations [53]. the color changes were observed upon mixing agnps with dpph solution, indicating the scavenging activity of dpph due to the donation of hydrogen atoms, which revealed that agnps synthesized from n. cadamba displayed effective antioxidant properties compared to crude, studies on agnps synthesized from elephantopus scaber leaf extract exhibited similar results [54]. conclusion the development of a sustainable and environmentally friendly approach for the synthesis of metallic nanoparticles are a key necessity in the field of nanotechnology. nanoparticles are regarded as essential building components in nanotechnology. because of their appealing physiochemical characteristics, silver nanoparticles play a significant role in biological research. the current study explains the development of silver nanoparticles using aqueous extracts of n. cadamba leaf, stem, bark, and flower. bioactive compounds present in the leaf, stem, bark, and flower of n. cadamba show a color change when silver nitrate is reduced to silver nanoparticles. it proves to be an environment-friendly, quick green approach to synthesis, as well as a cost-effective and efficient way to make silver nanoparticles. characterization was done by using methods like uv-vis absorption spectrophotometer, dls, zeta-potential, sem, and edax to the extracted silver nanoparticles. they prove that the capping agent provides stability to the agnps and also the formation of agnps. the phytochemical screening confirms the occurrence of tannins, steroids, cardiac glycosides, saponins, terpenoids, flavonoids, and alkaloids. the green synthesized silver nanoparticles show effective antibacterial activity towards b. subtilis, s. aureus, e. coli, and p. aeruginosa, antifungal activity towards a. niger, and antioxidant activities when compared with crude extracts. the agnps synthesized from n. cadamba were less effective against a. niger and c. albicans at performed concentrations. https://doi.org/10.5599/admet.1793 juluri maheswari et al. admet & dmpk 00(0) (2023) 000-000 10 agnps from leaf extract of n. cadamba showed effective antibacterial, antifungal, and antioxidant activities compared to agnps synthesized from stem, bark, and flower extracts of n. cadamba. the silver nanoparticles synthesized from plant extracts could be used in medicine, food, drug delivery, and in the preparation of pharmaceuticals such as antibiotics because of low cost, non-toxic, organic, and most efficacious against bacteria and fungi, which cause diseases in humans and plants. in the future, we are also planning to do the anticancer and antidiabetic activity of agnps synthesized from the n. cadamba plant. author contributions: bk and gn contributed to conception and design of the study. jm, mra and ms contributed to data acquisition, analysis and interpretation. drafting the initial manuscript by mra and jm. revising the manuscript critically for important intellectual content by bk, gn and ksbn. conflicts of interest: the authors declared no conflicts of interest. data availability: the datasets generated during and/or analyzed during the current study are not publicly available due to confidentiality but are available from the corresponding author on reasonable request. acknowledgements: we are thankful to dst-curie, central instrumentation facility, sri padmavati mahila visvavidyalayam (women’s university), tirupati, for providing the instrumentation to carry out this research work. ksbn would like to thank durban university of technology for research fellowship. references [1] l. xu, y.y. wang, j. huang, c.y. chen, z.x. wang, h. xie. silver nanoparticles: synthesis, medical applications and biosafety. theranostics 10 (2020) 8996-9031. https://doi.org/10.7150/thno.45413 [2] r. prasad, a. bhattacharyya, q.d. nguyen. nanotechnology in sustainable agriculture: recent developments, challenges, and perspectives. front. microbiol. 8 (2017) 1014. https://doi.org/ 10.3389/fmicb.2017.01014 [3] j.k. patra, g. das, l.f. fraceto, e.v.r. campos, m.d.p. rodriguez-torres, l.s. acosta-torres, l.a. diaztorres, r. grillo, m.k. swamy, s. sharma, s. habtemariam, h.s. shin. nano based drug delivery systems: recent developments and future prospects. j. nanobiotechnology 16 (2018) 71. https://doi.org/10.1186/s12951-018-0392-8 [4] f.d. guerra, m.f. attia, d.c. whitehead, f. alexis. nanotechnology for environmental remediation: materials and applications. molecules 23 (2018) 1760. https://doi.org/10.3390/molecules23071760 [5] m. mansuripur. joint international symposium on optical memory and optical data storage, the role of nanotechnology in data storage devices and systems. kauai, hawaii united states (2011) https://doi.org/10.1364/isom_ods.2011.oma2 [6] r. dastjerdi, m. montazer, s. shahsavan. a novel technique for producing durable multifunctional textiles using nanocomposite coating. colloids surf. biointerfaces 81 (2010) 32-41. https://doi.org/ 10.1016/j.colsurfb.2010.06.023 [7] m. catauro, m.g. raucci, f. de gaetano, a. marotta. antibacterial and bioactive silver-containing na2o x cao x 2sio2 glass prepared by sol-gel method. j. mater. sci. mater. med. 15 (2004) 831-837. https://doi.org/10.1023/b:jmsm.0000032825.51052.00 [8] v. vaishnavi, k.k. sadasivuni, d. ponnamma, n. golla. green synthesis of silver nanoparticles from pterocarpus santalinus leaf broth and their antibacterial and antioxidant activities, macromolecular symposia 392 (2020) 2000079 http://dx.doi.org/10.1002/masy.202000079 [9] f. joy, j. devasia, a. nizam, v.v. lakshmaiah, s.b.n. krishna. fungi-templated silver nanoparticle composite: synthesis, characterization, and its applications. applied sciences 13 (2023) 2158. https://doi.org/10.3390/app13042158 [10] m. ozdal, s. gurkok. recent advances in nanoparticles as antibacterial agent. admet dmpk 10 (2022) 115-129. https://doi.org/10.5599/admet.1172 https://doi.org/10.7150/thno.45413 https://doi.org/10.3389/fmicb.2017.01014 https://doi.org/10.3389/fmicb.2017.01014 https://doi.org/10.1186/s12951-018-0392-8 https://doi.org/10.3390/molecules23071760 https://doi.org/10.1364/isom_ods.2011.oma2 https://doi.org/10.1016/j.colsurfb.2010.06.023 https://doi.org/10.1016/j.colsurfb.2010.06.023 https://doi.org/10.1023/b:jmsm.0000032825.51052.00 http://dx.doi.org/10.1002/masy.202000079 https://doi.org/10.3390/app13042158 https://doi.org/10.5599/admet.1172 admet & dmpk 00(0) (2023) 000-000 silver nanoparticles synthesized from neolamarkia cadamba doi: https://doi.org/10.5599/admet.1793 11 [11] m. rai, a. yadav, a. gade. silver nanoparticles as a new generation of antimicrobials. biotechnol. adv. 27 (2009) 76-83. https://doi.org/10.1016/j.biotechadv.2008.09.002 [12] s. schultz, d. smith, j. mock, d. schultz. applications of silver nanoparticles in biosensor. proc. natl. acad. sci. usa 97 (2000) 996–1001. https://doi.org/10.1073/pnas.97.3.996 [13] n. durán, p.d. marcato, o.l. alves, g.i. souza, e. esposito. mechanistic aspects of biosynthesis of silver nanoparticles by several fusarium oxysporum strains. j. nanobiotechnology 3 (2005) 8. https://doi.org/10.1186/1477-3155-3-8 [14] r.m. crooks, m. zhao, l. sun, v. chechik, l.k. yeung. dendrimer-encapsulated metal nanoparticles: synthesis, characterization, and applications to catalysis. acc. chem. res. 34 (2001) 181-190. https://doi.org/10.1021/ar000110a [15] d.i. gittins, d. bethell, r.j. nichols, d.j. schiffrin. diode-like electron transfer across nanostructured films containing a redox ligand. journal of materials chemistry 10 (2000) 79-83. https://doi.org/ 10.1039/a902960e [16] j.s. moodley, s.b.n. krishna, k. pillay, p. govender. green synthesis of metal nanoparticles for antimicrobial activity. novel nanomaterials (2020) 253-278. http://doi.org/10.5772/intechopen.94 348 [17] s. palithya, s.a. gaddam, v.s. kotakadi, j. penchalaneni, n. golla, s.b.n. krishna, c.v. naidu. green synthesis of silver nanoparticles using flower extracts of aerva lanata and their biomedical applications. particulate science and technology 40 (2022) 84-96. https://doi.org/10.1080/02726 351.2021.1919259 [18] a. plucinski, z. lyu, b.v.k.j. schmidt. polysaccharide nanoparticles: from fabrication to applications. j. mater. chem. b 9 (2021) 7030-7062. https://doi.org/10.1039/d1tb00628b [19] j. avilala, n. golla. antibacterial and antiviral properties of silver nanoparticles synthesized by marine actinomycetes. int. j. pharm. sci. res 10 (2019) 1223-1228. http://doi:10.13040/ijpsr.09758232.10(3).1223-28 [20] v.v. makarov, a.j. love, o.v. sinitsyna, s.s. makarova, i.v. yaminsky, m.e. taliansky, n.o. kalinina. "green" nanotechnologies: synthesis of metal nanoparticles using plants. acta naturae 6 (2014) 3544. pmid: 24772325; pmcid: pmc3999464. [21] a. dwevedi, k. sharma, y.k. sharma. cadamba: a miraculous tree having enormous pharmacological implications. pharmacogn. rev. 9 (2015) 107-113. https://doi.org/10.4103/0973-7847.162110 [22] r. verma, f. chaudhary, a. singh. neolamarckia cadamba: a comprehensive pharmacological. global journal of pharmacy & pharmaceutical sciences 6 (2018) 73-78. https://doi.org/10.19080/gjpps.20 18.06.555691 [23] r. puatanachokchai, h. kishida, a. denda, n. murata, y. konishi, u. vinitketkumnuen, d. nakae. inhibitory effects of lemon grass (cymbopogon citratus, stapf) extract on the early phase of hepatocarcinogenesis after initiation with diethylnitrosamine in male fischer 344 rats. cancer lett. 183 (2002) 9-15. https://doi.org/10.1016/s0304-3835(02)00111-8 [24] a. pandey, p.s. negi. traditional uses, phytochemistry and pharmacological properties of neolamarckia cadamba: a review. j. ethnopharmacol. 181 (2016) 118-135. https://doi.org/10.1016/j.jep.20 16.01.036 [25] h.a. mansour, a.s. newairy, m.i. yousef, s.a. sheweita. biochemical study on the effects of some egyptian herbs in alloxan-induced diabetic rats. toxicology 170 (2002) 221-228. https://doi.org/ 10.1016/s0300-483x(01)00555-8 [26] a. pandey, p.s. negi. phytochemical composition, in vitro antioxidant activity and antibacterial mechanisms of neolamarckia cadamba fruits extracts. nat. prod. res. 32 (2018) 1189-1192. https://doi.org/10.1080/14786419.2017.1323209 [27] a. pandey, p.s. negi. traditional uses, phytochemistry and pharmacological properties of neolamarckia cadamba: a review. j. ethnopharmacol. 181 (2016) 118-135. https://doi.org/10.1016/j.jep.20 16.01.036 https://doi.org/10.5599/admet.1793 https://doi.org/10.1016/j.biotechadv.2008.09.002 https://doi.org/10.1073/pnas.97.3.996 https://doi.org/10.1186/1477-3155-3-8 https://doi.org/10.1021/ar000110a https://doi.org/10.1039/a902960e https://doi.org/10.1039/a902960e http://doi.org/10.5772/intechopen.94348 http://doi.org/10.5772/intechopen.94348 https://doi.org/10.1080/02726351.2021.1919259 https://doi.org/10.1080/02726351.2021.1919259 https://doi.org/10.1039/d1tb00628b https://doi.org/10.13040/ijpsr.0975-8232.10(3).1223-28 https://doi.org/10.13040/ijpsr.0975-8232.10(3).1223-28 https://doi.org/10.4103/0973-7847.162110 https://doi.org/10.19080/gjpps.2018.06.555691 https://doi.org/10.19080/gjpps.2018.06.555691 https://doi.org/10.1016/s0304-3835(02)00111-8 https://doi.org/10.1016/j.jep.2016.01.036 https://doi.org/10.1016/j.jep.2016.01.036 https://doi.org/10.1016/s0300-483x(01)00555-8 https://doi.org/10.1016/s0300-483x(01)00555-8 https://doi.org/10.1080/14786419.2017.1323209 https://doi.org/10.1016/j.jep.2016.01.036 https://doi.org/10.1016/j.jep.2016.01.036 juluri maheswari et al. admet & dmpk 00(0) (2023) 000-000 12 [28] m.z. zayed, f.b. ahmad, w.-s. ho, s.-l. pang. gc-ms analysis of phytochemical constituents in leaf extracts of neolamarckia cadamba (rubiaceae) from malaysia. int. j. pharm. pharm. sci. 6 (2014) 123127. https://journals.innovareacademics.in/index.php/ijpps/article/view/1824. [29] s. li, x. liu, x. chen, l. bi. research progress on anti-inflammatory effects and mechanisms of alkaloids from chinese medical herbs. evid based complement alternat med 2020 (2020) 1303524. https://doi.org/10.1155/2020/1303524 [30] s.p. umachigi, g.s. kumar, k. jayaveera, k.d. kishore, k.c. ashok, r. dhanapal. antimicrobial, wound healing and antioxidant activities of anthocephalus cadamba. afr. j. tradit. complement. altern. med. 4 (2007) 481-487. pmid: 20161916; pmcid: pmc2816507. [31] k. mallikarjuna, n.j. sushma, g. narasimha, l. manoj, b.d.p. raju. phytochemical fabrication and characterization of silver nanoparticles by using pepper leaf broth. arabian journal of chemistry 7 (2014) 1099-1103. https://doi.org/10.1016/j.arabjc.2012.04.001 [32] y. zhang, m. yang, n.g. portney, d. cui, g. budak, e. ozbay, m. ozkan, c.s. ozkan. zeta potential: a surface electrical characteristic to probe the interaction of nanoparticles with normal and cancer human breast epithelial cells. biomed. microdevices 10 (2008) 321-328. https://doi.org/10.1007/ s10544-007-9139-2 [33] e. akman, b.g. oztoprak, m. gunes, e. kacar, a. demir. effect of femtosecond ti: sapphire laser wavelengths on plasmonic behaviour and size evolution of silver nanoparticles. photonics and nanostructures-fundamentals and applications 9 (2011) 276-286. https://doi.org/10.1016/j.photo nics.2011.05.004 [34] m.a. ali, k.a. mosa, a. el-keblawy, h. alawadhi. exogenous production of silver nanoparticles by tephrosia apollinea living plants under drought stress and their antimicrobial activities. nanomaterials (basel) 9 (2019) 1716. https://doi.org/10.3390/nano9121716 [35] j.s. devi, b.v. bhimba, d.m. peter. production of biogenic silver nanoparticles using sargassum longifolium and its applications. indian journal of geo-marine sciences 42 (2013) 125-130. http://nopr.niscpr.res.in/handle/123456789/15995 [36] k. seku, s.s. hussaini, b. pejjai, m.m.s. al balushi, r. dasari, n. golla, g.b. reddy. a rapid microwaveassisted synthesis of silver nanoparticles using ziziphus jujuba mill fruit extract and their catalytic and antimicrobial properties. chemical papers 75 (2021) 1341-1354. https://doi.org/10.1007/s11696020-01386-w [37] k. seku, b.r. gangapuram, b. pejjai, k.k. kadimpati, n. golla. microwave-assisted synthesis of silver nanoparticles and their application in catalytic, antibacterial and antioxidant activities. journal of nanostructure in chemistry 8 (2018) 179-188. https://doi.org/10.1007/s40097-018-0264-7 [38] h. khadri, m. alzohairy, a. janardhan, a.p. kumar, g. narasimha. green synthesis of silver nanoparticles with high fungicidal activity from olive seed extract. advances in nanoparticles 2 (2013) 241-246. http://dx.doi.org/10.4236/anp.2013.23034 [39] s. palithya, s. a. gaddam, v. s. kotakadi, j. penchalaneni, n. golla, s.b.n. krishna, c. v. naidu. green synthesis of silver nanoparticles using flower extracts of aerva lanata and their biomedical applications, particulate science and technology 40(2021) 84-96. https://doi.org/10.1080/027263 51.2021.1919259 [40] s. bhakya, s. muthukrishnan, m. sukumaran, m. muthukumar, s.t. kumar, m. rao. catalytic degradation of organic dyes using synthesized silver nanoparticles: a green approach. j. bioremed. biodeg. 6 (2015). 1000312. https://doi.org/10.4172/2155-6199.1000312 [41] k. mohan kumar, m. sinha, b.k. mandal, a.r. ghosh, k. siva kumar, p. sreedhara reddy. green synthesis of silver nanoparticles using terminalia chebula extract at room temperature and their antimicrobial studies. spectrochim. acta a mol. biomol. spectrosc. 91 (2012) 228-233. https://doi.org/10.1016/j.saa.2012.02.001 [42] c. tanase, l. berta, n.a. coman, i. roșca, a. man, f. toma, a. mocan, a. nicolescu, l. jakab-farkas, d. biró, a. mare. antibacterial and antioxidant potential of silver nanoparticles biosynthesized using the spruce bark extract. nanomaterials (basel) 9 (2019) 1541. https://doi.org/10.3390/nano9111541 https://journals.innovareacademics.in/index.php/ijpps/article/view/1824 https://doi.org/10.1155/2020/1303524 https://doi.org/10.1016/j.arabjc.2012.04.001 https://doi.org/10.1007/s10544-007-9139-2 https://doi.org/10.1007/s10544-007-9139-2 https://doi.org/10.1016/j.photonics.2011.05.004 https://doi.org/10.1016/j.photonics.2011.05.004 https://doi.org/10.3390/nano9121716 https://www.researchgate.net/journal/indian-journal-of-geo-marine-sciences-0379-5136 http://nopr.niscpr.res.in/handle/123456789/15995 https://doi.org/10.1007/s11696-020-01386-w https://doi.org/10.1007/s11696-020-01386-w https://doi.org/10.1007/s40097-018-0264-7 http://dx.doi.org/10.4236/anp.2013.23034 https://doi.org/10.1080/02726351.2021.1919259 https://doi.org/10.1080/02726351.2021.1919259 https://doi.org/10.4172/2155-6199.1000312 https://doi.org/10.1016/j.saa.2012.02.001 https://doi.org/10.3390/nano9111541 admet & dmpk 00(0) (2023) 000-000 silver nanoparticles synthesized from neolamarkia cadamba doi: https://doi.org/10.5599/admet.1793 13 [43] a.k. keshari, r. srivastava, p. singh, v.b. yadav, g. nath. antioxidant and antibacterial activity of silver nanoparticles synthesized by cestrum nocturnum. j. ayurveda. integr. med. 11 (2020) 37-44. https://doi.org/10.1016/j.jaim.2017.11.003 [44] m.m.h. khalil, e.h. ismail, k.z. el-baghdady, d. mohamed. green synthesis of silver nanoparticles using olive leaf extract and its antibacterial activity. arabian journal of chemistry 7 (2014) 1131-1139. https://doi.org/10.1016/j.arabjc.2013.04.007 [45] f.e. meva, m.l. segnou, c.o. ebongue, a.a. ntoumba, p.b.e. kedi, v. deli, m.-a. etoh, e.m. mpondo. spectroscopic synthetic optimizations monitoring of silver nanoparticles formation from megaphrynium macrostachyum leaf extract. revista brasileira de farmacognosia 26 (2016) 640-646. https://doi.org/10.1016/j.bjp.2016.06.002 [46] s.c.g.k. daniel, b.n. banu, m. harshiny, k. nehru, p.s. ganesh, s. kumaran, m. sivakumar. ipomea carnea-based silver nanoparticle synthesis for antibacterial activity against selected human pathogens. journal of experimental nanoscience 9 (2014) 197-209. https://doi.org/10.1080/174580 80.2011.654274 [47] s. raja, v. ramesh, v. thivaharan. green biosynthesis of silver nanoparticles using calliandra haematocephala leaf extract, their antibacterial activity and hydrogen peroxide sensing capability. arabian journal of chemistry. 10 (2017) 253-261. https://doi.org/10.1016/j.arabjc.2015.06.023 [48] p. devraj, k. prachi, a. chirom, r. arun. synthesis and characterization of silver nanoparticles using cannonball leaves and their cytotoxic activity against mcf-7 cell line. j. nanotechnol. 6 (2013) 598328. https://doi.org/10.1155/2013/598328 [49] s. phongtongpasuk, s. poadang, n. yongvanich. environmental-friendly method for synthesis of silver nanoparticles from dragon fruit peel extract and their antibacterial activities. energy procedia 89 (2016) 239-247. https://doi.org/10.1016/j.egypro.2016.05.031 [50] m. krychowiak, m. grinholc, r. banasiuk, m. krauze-baranowska, d. głód, a. kawiak, a. królicka. combination of silver nanoparticles and drosera binata extract as a possible alternative for antibiotic treatment of burn wound infections caused by resistant staphylococcus aureus. plos one 9 (2014) e115727. https://doi.org/10.1371/journal.pone.0115727 [51] hemlata, p.r. meena, a.p. singh, k.k. tejavath. biosynthesis of silver nanoparticles using cucumis prophetarum aqueous leaf extract and their antibacterial and antiproliferative activity against cancer cell lines. acs omega 5 (2020) 5520-5528. https://doi.org/10.1021/acsomega.0c00155 [52] k. nesović, a. janković, t. radetić, a. perić-grujić, m. vukasinović-sekulić, v. kojić, k. y. rhee, v. mišković-stanković. poly (vinyl alcohol)/chitosan hydrogels with electrochemically synthesized silver nanoparticles for wound dressing applications. j. electrochem. sci. eng. 10 (2020) 185-198. https://doi.org/10.5599/jese.732 [53] j.s, devi, b.v, bhimba. antibacterial and antifungal activity of silver nanoparticles synthesized using hypnea muciformis. biosci. biotechnol. res. asia 11 (2014) 235-238. http://dx.doi.org/10.13005/ bbra/1260 [54] s.n. kharat, v.d. mendhulkar. "synthesis, characterization and studies on antioxidant activity of silver nanoparticles using elephantopus scaber leaf extract". mater. sci. eng. c mater. biol. appl. 62 (2016) 719-724. https://doi.org/10.1016/j.msec.2016.02.024 ©2023 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) https://doi.org/10.5599/admet.1793 https://doi.org/10.1016/j.jaim.2017.11.003 https://doi.org/10.1016/j.arabjc.2013.04.007 https://doi.org/10.1016/j.bjp.2016.06.002 https://doi.org/10.1080/17458080.2011.654274 https://doi.org/10.1080/17458080.2011.654274 https://doi.org/10.1016/j.arabjc.2015.06.023 https://doi.org/10.1155/2013/598328 https://doi.org/10.1016/j.egypro.2016.05.031 https://doi.org/10.1371/journal.pone.0115727 https://doi.org/10.1021/acsomega.0c00155 https://doi.org/10.5599/jese.732 http://dx.doi.org/10.13005/bbra/1260 http://dx.doi.org/10.13005/bbra/1260 https://doi.org/10.1016/j.msec.2016.02.024 http://creativecommons.org/licenses/by/3.0/ manuscript doi: 10.5599/admet.3.2.167 141 admet & dmpk 3(2) (2015) 141-153; doi: 10.5599/admet.3.2.167 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper formulation and characterization of ketoprofen embedded polycaprolactone microspheres using solvent evaporation method pankaj wagh, jitendra naik* department of pharmaceutical technology, university institute of chemical technology, north maharashtra university, jalgaon, 425 001, maharashtra, india *corresponding author: e-mail: jbnaik@nmu.ac.in; tel: 91257 2258441; fax: +91257 2258403 received: february 13, 2015; revised: april 09, 2015; published: july 01, 2015 abstract the purpose of this study was to prepare polymeric microspheres containing ketoprofen (kfn) by single emulsion [oil-in-water (o/w)] solvent evaporation method. polycaprolactone (pcl), biocompatible polymer, was used for the preparation of sustained released microspheres of kfn. a plackett–burman design was employed by using the design-expert® software (version9.0.3.1, stat-ease inc., minneapolis, mn). eleven factors out of six processing factors were investigated in order to enhance the encapsulation efficiency (ee) of the microspheres. the resultant microspheres were characterized for their size, morphology, ee, and drug release. imaging of particles was performed by field emission scanning electron microscopy. interaction between the drug and polymers were investigated by fourier transform infrared (ftir) spectroscopy, x-ray powder diffractometry (xrpd) and differential scanning calorimetry (dsc). graphical and mathematical analyses of the design showed that concentration of factor pcl (b) and varying speed (f, revolution per minute, rpm) were significant negative effect on the ee and identified as the significant factor determining the ee of the microspheres. the microspheres showed high % ee (31.18 % to 96.81 %). the microspheres were found to be discrete, oval with porous surface. the ftir analysis confirmed no interaction of kfn with the polymer. the xrpd revealed the dispersion of drug within microspheres formulation. sustained drug release profile over 12 h was achieved by pcl polymer. in conclusion, polymeric microspheres containing kfn can be successfully prepared using the technique of experimental design, and these results helped in finding the optimum formulation variables for ee of microspheres. keywords plackett–burman design, encapsulation efficiency, sustained release introduction microspheres are well accepted system to control the drug release from the dosage form to improve bioavailability; it reduces absorption difference in patients, and dosing frequency as well as adverse effects during prolonged treatment [1]. ketoprofen (kfn) is an nsaid having prominent anti-inflammatory, analgesic and antipyretic properties. kfn is an arylpropionic acid derivative and it is one of the most powerful inhibitors of cyclooxygenase at concentrations well within the range of therapeutic plasma concentration (ec50 2μg/l), which is resulting in a reduction in the tissue production of prostaglandins such as pge2 and pgf2a [2]. in addition to its effect on cyclooxygenase, kfn inhibits the lipoxygenase pathway http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:jbnaik@nmu.ac.in wagh and naik admet & dmpk 3(2) (2015) 141-153 142 of the arachidonic acid cascade. kfn is also a strongest inhibitor of bradykinin, an important chemical mediator of pain and inflammation. it also stabilizes lysosomal membranes against osmotic change and prevents the release of lysosomal enzymes that mediate tissue destruction in inflammatory react ions [3]. the half-life of kfn in plasma is about 2–2.5 hours. due to its short half-life, kfn is a very good candidate for the formulation of sustained release dosage forms [2-3]. it is bcs ii drug [4], kfn has low water solubility and exposure to the stomach at high levels can cause severe gastric damage such as ulceration or bleeding. to overcome these disadvantages, sustained release dosage forms have been developed, resulting in less frequent dosing and less gi disturbances [5]. the solvent evaporation technique which was used in this study is a simple process that is also inexpensive enough for scaling up to a commercial level. the purpose of the present study was to prepare kfn microspheres by using solvent evaporation method and to study the effect of drug-polymer concentration on drug release. the main objective of this work was to investigate the possibility of getting sustained release kfn microspheres by the solvent evaporation method using polycaprolactone (pcl). investigation of the effect of various processing and formulation factors such as drug to polymer ratio, stirring speed, surfactant concentration and others on the, yield, encapsulation efficiency (ee), and in-vitro release rate of drug from the microspheres were performed. materials and methods materials ketoprofen (kfn) was procured from swapnroop drugs and pharmaceuticals (aurangabad, india). polycaprolactone (pclmw 70,000) was purchased from sigma – aldrich chemicals india pvt. ltd. (mumbai, india). n-hexane and dichloromethane (dcm) was procured from merck specialities pvt. ltd. (mumbai, india). polyvinyl alcohol (pvamw 140, 000) was procured from himedia laboratories pvt. ltd. (mumbai, india). all other solvents and reagents were of analytical grade and used as provided. methods plackett–burman design (pbd) the pbd factorial design of experiment was performed using design-expert® (doe) software (version 9.0.3.1, stat-ease inc., minneapolis, mn) [6-10]. in pbd, eleven numbers of lowest factors and forty seven numbers of highest factors can be used. in lowest eleven factors, five formulation factors and one process factor was investigated in order to enhance the ee of the microspheres and remaining five factors was dummied. pbd plotted twelve runs containing variation value of factors. six independent variables such as amount of kfn drug (a, mg), pcl (b, mg), dcm (c, ml), water (d, ml), surfactant (e, % w/v) and speed (f, revolution per minute, rpm). the level of screening variables evaluated in this study is listed in table 1. doe analyses the input data, presents the rank ordering of the variables, and designates the sign to the effect to indicate whether an increase in factor value was advantageous or not [11-13]. doe shows, formulation equations, anova for ee and drug loading (dl), p-value, f-value, mean square and r 2 value. it also shows pareto chart, experimental, predicted, and % prediction error values. the significance of the design was determined by the comparisons of statistical parameters, and on the basis of higher values of r 2 . twodimensional (2d) contour plots and three-dimensional (3d) response plots resulting from the equations were constructed using designexpert® software. admet & dmpk 3(2) (2015) 141-153 ketoprofen embedded polycaprolactone microspheres doi: 10.5599/admet.3.2.167 143 table 1. screening variables and their levels in the plackett-burman design factor name unit low actual high actual a ketoprofen mg 100 200 b pcl mg 100 400 c dcm ml 20 30 d water ml 70 100 e pva mg 0.1 0.3 f speed rpm 700 1000 g g -1 1 h h -1 1 j j -1 1 k k -1 1 l l -1 1 preparation of encapsulated microspheres kfn loaded microspheres were prepared using oil-in-water (o/w) single emulsion solvent evaporation method. an accurate weighed quantity of a (kfn), b (pcl) was mixed with a c (dcm) organic solvent as per the design. the resulting oil phase was injected drop wise using a syringe with a 21.5 g needle into a factor d (distilled water) as an external phase with stirring speed 700/1000 rpm using lab stirrer (remi electrotechnik limited, thane, india) in a room temperature containing varying concentrations of pva (factor e) as an emulsifier to produce an o/w single emulsion. after 60 min, n-hexane was added into the solution for hardening of the microspheres. fig. 1 shows the schematic representation of formulation of kfn microspheres. stirring was continued further at a varying speed 700/1000 rpm until the evaporation of organic solvent from solid microspheres. the microsphere were collected by filtration using vacuum filtration (tid -15, temp. 55 °c, speed 1440 rpm) washed with n-hexane, air dried for 48 h in room temperature and was used for further studies [6,14]. characterization of the microspheres encapsulation efficiency (% ee) the amount of kfn encapsulated into microsphere was determined by uv–vis spectrophotometer (hitachi u-2900, tokyo, japan). an accurately weighed 10 mg of microspheres were stirred with dichloromethane (5 ml) to dissolve the polymer as well as drug and extracted in phosphate buffer solution ph 6.8 (up to 100 ml). stirring was continued for 30 min at room temperature to facilitate the evaporation of organic solvent. the dispersion was filtered, and the residue was washed with phosphate buffer solution. the drug content was determined in the filtrate after appropriate dilution with a phosphate buffer solution at 260 nm using uv-vis spectrophotometer. the ee was expressed as the percentage of drug incorporated in the formulation relative to the total amount of drug (theoretical quantity) used in the formulation [6]. the % ee of kfn was calculated using the following equations: drug entrapment efficiency (% ) = (actual drug content/theoretical drug content) x 100 (1) wagh and naik admet & dmpk 3(2) (2015) 141-153 144 figure 1. schematic representation of formulation of kfn microspheres among all twelve run formulations of microspheres having higher encapsulation efficiency were chosen for further characterization study. scanning electron microscopy the morphology of microspheres was investigated by scanning electron microscopy (fesem-s 4800, hitachi, japan) at a working distance of 8.0–8.9 mm and accelerating voltage of 15.0 kv. the particles were examined for shape, size and surface characteristics. for the morphology study, microsphere were mounted on metal stub using double sided adhesive tape and coated with gold for 80 second under vacuum. fourier transform infrared (ftir) spectroscopy study the chemical structure of the kfn, pcl and drug loaded microspheres were analysed using ftir spectrophotometer (ftir-8400, shimadzu, asia pacific pvt. ltd. singapore) by kbr pellet method. sample (1 mg) was mixed with kbr (40 mg) and formed into a disk by applying force in a manual press. spectra were recorded in the scan range of 4000–400 cm −1 . x-ray diffraction (xrd) x-ray diffraction patterns of kfn, pcl and drug loaded microspheres were obtained in x-ray diffractometer (bruker, d8 advanced, germany) with cu kα radiation (λ = 1.5406 ǻ). the samples were analysed over the angle range (2θ) 10°–80°. differential scanning calorimetriy(dsc) thermal behaviour of the sample kfn, pcl and drug loaded microsphere was determined by differential scanning calorimetric (dsc-60, shimadzu). accurately weighed samples (5-10 mg) were sealed in an admet & dmpk 3(2) (2015) 141-153 ketoprofen embedded polycaprolactone microspheres doi: 10.5599/admet.3.2.167 145 aluminium pan and scanned at a temperature range of 30 °c to 400 °c at the rate of 10 °c/min under dry nitrogen atmosphere purge of 50 ml/min. in vitro drug dissolution studies drug release from microspheres was performed in vitro using phosphate buffer (ph 6.8) for 12 h in dissolution test app. (model fc 6x12r electrolab tdt — 08 l, india) uspxxviii, type-i with speed 100 rpm and temperature 37 ± 0.5 °c. the dissolution medium of phosphate buffer (ph 6.8) was prepared according to indian pharmacopoeia 2007 [15]. from 900 ml phosphate buffer i.e. dissolution medium 10 ml were withdrawn at predetermined time intervals and replenished with same volume of fresh dissolution media to maintain the sink condition. the samples were filtered through a whatman filter paper no. 41. the kfn content of each sample after suitable dilution was assayed by uv spectroscopy at λ max of 260 nm using a 1 cm cell. result and discussion statistical design and analysis the pbd is an efficient approach to evaluate the results which are shown in table 2. design shows the formulation equation of % ee can be represented by the following equation: ee = +84.08-0.051a-3.02b+5.15c+7.34d+6.01e–2.85f-8.57g-7.71h+1.64j-6.21l (2) table 2. the plackett-burman experimental design matrix (in coded level) and experimental results runs variables response a b c d e f g h j k l ee, % 1 1 -1 1 -1 -1 -1 1 1 1 -1 1 65.31± 0.56 2 -1 1 1 1 -1 1 1 -1 1 -1 -1 96.81± 1.59 3 -1 -1 1 -1 1 -1 1 1 -1 1 -1 91.86± 2.74 4 1 -1 -1 -1 1 -1 1 -1 1 1 -1 95.04± 2.12 5 -1 1 -1 -1 -1 1 1 1 -1 1 1 96.20± 1.25 6 -1 -1 -1 -1 1 1 -1 1 1 -1 1 87.79± 2.55 7 1 1 1 -1 1 1 -1 1 -1 -1 -1 94.66± 0.57 8 1 1 -1 -1 1 -1 1 -1 -1 -1 1 80.27± 2.43 9 1 1 -1 -1 -1 -1 -1 1 1 1 -1 93.24± 1.86 10 -1 1 1 -1 1 -1 -1 -1 1 1 1 31.18± 0.94 11 -1 -1 -1 -1 -1 -1 -1 -1 -1 -1 -1 92.34± 0.83 12 1 -1 1 -1 -1 1 -1 1 -1 1 1 84.31± 2.24 from the above equation it is observed that the ee increased by the water (d), pva (e), and followed by the dummy factors; (g), (h), (l). dummy factors (g) and (h) have higher and significant effects on ee in the microspheres preparation process. table 3 shows the design of anova for encapsulation efficiency, p-value, f-value, mean square and r 2 of the model. from the wagh and naik admet & dmpk 3(2) (2015) 141-153 146 anova of pbd for drug loading, p-value, f-value, mean square and r 2 of the model are given in the table 4. standardized pareto chart of ee is shown in fig. 2a and dl in fig. 2b. pareto charts showed that the parametrical graph containing the t-value of the student's t-test represents bars with a length proportional to the absolute value of the estimated effects divided by the standard error. there are eleven factors in which water (factor d), pva (factor e) and dummy factor g (unknown in the process) has a significant influence on the ee. table 3. analysis of variance for encapsulation efficiency (ee) source sum of squares d.f. mean squares f value p-value prob>f model 3695.00 101 369.50 1.55 0.5589(ns) a 0.031 1 0.031 1.30 0.9927 b 109.51 1 109.51 0.46 0.6204 c 318.37 1 318.37 1.34 0.4537 d 646.36 1 646.36 2.72 0.3471 e 432.84 1 432.84 1.82 0.4060 f 97.41 1 97.41 0.41 0.6375 g 881.17 1 881.17 3.71 0.3050 h 713.48 1 713.48 3.00 0.3333 j 32.44 1 32.44 0.14 0.7747 l 463.39 1 463.39 1.95 0.3957 residual 237.72 1 237.72 cor total 3932.72 11 d.f. = degree of freedom, s = significant, ns = non-significant. standard deviation of the residual = 15.42. correlation coefficient = 0.9396. table 4. analysis of variance for drug loading (dl). source sum of squares d.f. mean squares f value p-value prob>f model 5554.80 10 555.48 152.10 0.0630(ns) a 978.85 1 978.85 268.03 0.03889(s) b 872.11 1 872.11 238.80 0.0411(s) c 84.48 1 84.48 23.13 0.1305 d 760.98 1 760.98 208.37 0.0440(s) e 152.08 1 152.08 41.64 0.0979 f 438.26 1 438.26 120.01 0.0580 g 673.50 1 673.50 184.42 0.0468(s) h 24.25 1 24.25 6.64 0.2356 j 1282.16 1 1282.16 351.08 0.0339(s) l 288.12 1 288.12 78.89 0.0714 residual 3.65 1 3.65 cor total 5558.45 11 d.f. = degree of freedom, s = significant, ns = non-significant. standard deviation of the residual = 1.91. correlation coefficient = 0.9993. admet & dmpk 3(2) (2015) 141-153 ketoprofen embedded polycaprolactone microspheres doi: 10.5599/admet.3.2.167 147 figure 2. (a) pareto chart of % ee. (b) pareto chart of % dl. (c) encapsulation efficiency between actual and predicted values plot. (d) encapsulation efficiency between residuals and run plot. (e) drug loading, between actual and predicted values plot. (f) drug loading between residuals and run plot. it was found that the kfn (a), pcl (b) and speed (f) had a significant negative effect on the % ee. the experimental, predicted, and % prediction error values for all the experimental runs are listed in table 5 for ee and table 6 for the dl. therefore, it can be concluded that the model is best suitable because of the difference between experimental and predicted value, and % prediction error value is very low in fig. 2 (c and d) for ee and (e and f) for dl. the desirability plot indicating desirable regression ranges for optimal process variable and the overlay plot indicating the region of optimal process variables are presented in fig. 3 (a and b) [6-7]. pbd also showed the equation of % dl which is represented by the following equation: dl=+77.00-9.03a+8.52b+2.65c+7.96d-3.56e-6.04f-7.49g+1.42h-10.34j+4.90l (3) the above equation 3 shows that out of eleven factors, dl increased by the factors; (a), (b), (d) and followed by the dummy factors; (g), (j). dummy factor (j) and (a) has higher and significant effect on drug loading in the microspheres preparation process. it was found that the factors; (a), (e), (f), (g) and (j) had a significant negative effect on the % drug loading. characterization of microspheres encapsulation efficiency (% e.e) the ee of different experimental runs of the microspheres is reported in table 2. the 3d response surface plots are useful in understanding about the main and interaction effects of the independent variables, whereas 2d contour plot gives a visual representation of values of the response [16]. to visualize the effect of independent variables on each response, 3d response surface plots & 2d contour plots (fig. 4) wagh and naik admet & dmpk 3(2) (2015) 141-153 148 were constructed. fig. 4b shows the effects of factor g and h on the response because both factors had more significant effect as compared to others. contour plots and surface response plots (fig. 4a & b) represented the increase in % ee while (fig. 4c & d) represents contour plots and surface response plots of % dl respectively. fig. 4d shows the effect of factor j and a on the response because both factors had more significant effect as compared to others. it was found that the kfn (a), pcl (b) and speed (f) had a significant negative effect on the % ee. the % ee of the microspheres was ranged from 31.18 % to 96.81 %. table 5. experimental and predicted value of the response ee. run experimental value predicted value residual % prediction eroor 1 65.31 60.86 4.45 6.81 2 96.81 92.36 4.45 4.60 3 91.86 87.41 4.45 4.84 4 95.04 99.49 -4.45 -4.68 5 96.20 91.75 4.45 4.62 6 87.79 83.34 4.45 5.06 7 94.66 99.11 -4.45 -4.70 8 80.27 84.72 -4.45 -5.54 9 93.24 88.79 4.45 4.77 10 31.18 35.63 -4.45 -14.2 11 92.34 96.79 -4.45 -4.81 12 84.31 88.76 -4.45 -5.27 percent prediction error was calculated using the formula (experimental value − predicted value)/experimental value × 100. table 6. experimental and predicted value of the response dl. run experimental value predicted value residual % prediction error 1 52.78 52.33 0.55 0.85 2 88.55 88.00 0.55 0.62 3 76.48 75.93 0.55 0.71 4 91.81 92.36 -0.55 -0.59 5 79.09 78.54 0.55 0.69 6 91.16 90.61 0.55 0.60 7 88.33 88.88 -0.55 -0.62 8 14.52 15.07 -0.55 -3.78 9 77.24 76.69 0.55 0.71 10 90.07 90.62 -0.55 -0.61 11 89.42 89.97 -0.55 -0.615 12 84.53 85.08 -0.55 -0.65 percent prediction error was calculated using the formula (experimental value − predicted value)/experimental value × 100. admet & dmpk 3(2) (2015) 141-153 ketoprofen embedded polycaprolactone microspheres doi: 10.5599/admet.3.2.167 149 figure 3. (a) desirability plot. (b) overlay plot. figure 4. (a) % ee contour plot, (b) % ee response surface, (c) % dl contour plot, (d) % dl response surface scanning electron microscopy (fe-sem) shape, size and surface morphology of the kfn loaded microspheres were examined by fe sem. scanning electron microphotographs of microspheres prepared in this study are shown in fig. 5. the size of microspheres was found to be in the range of 20 μm to 293 μm (fig. 5d) and the microspheres are oval in shape with porosity and rough surface (fig. 5c and d). wagh and naik admet & dmpk 3(2) (2015) 141-153 150 figure 5. sem micrographs and surface of kfn microspheres fourier transform infrared (ftir) spectroscopy study ftir spectra were obtained in order to analyse the prepared microspheres. fig. 6 shows typical spectra of pure kfn, polymers and encapsulated kfn in the microspheres. in the spectrum of kfn (a), principal peaks were found at 3288 cm −1 of n-h stretching, at 1654 cm −1 of c=c stretching, at 3055 cm−1 attributed to c-h stretching of the aromatic ring, at 2937 cm −1 attributed to aliphatic c-h stretching vibrations, at 1076 cm −1 because of c-o c stretching, at 2978 cm −1 of ethoxy group stretching. in the spectra of pcl (b), principal peak at 3758 cm −1 of o-h stretching. the small peak at 2969 cm −1 corresponds to a symmetric structure vibration of the ethoxy groups, at 757 cm −1 of c-cl stretching, at 1070 cm −1 of c-o-c stretching, at 1387 cm −1 of o-h stretching, at 1763 cm −1 of carbonyl group stretching, these values remained very close in the ftir spectra of optimized kfn loaded microspheres (c), indicating no existence of the interaction between the kfn and polymer. x-ray diffraction (xrd) the presence of numerous distinct peaks in the xrd patterns shown in fig. 7 of kfn (a), at 2θ; 5.41°, 12.28°, 13.52°, 15.34°, 16.08°, 16.43°,17.54° and 18.23° was present as crystalline material 88.2 % and 11.8 % amorphous. there was no clear peak in the xrd patterns (b) of pcl indicate the amorphous nature of the polymer, crystalline 44.2 % and amorphous 55.8 %. the xrd patterns of drug loaded optimize formulation (c) crystalline 58.9 % and amorphous 41.1 % were characterized by the absence of distinct diffraction peaks of kfn, signifying a drug amorphization. thus, xrd analysis suggested that the degree of crystallinity of kfn was reduced in the microspheres formulation, and most of the drug distributed in the polymer at molecular level. admet & dmpk 3(2) (2015) 141-153 ketoprofen embedded polycaprolactone microspheres doi: 10.5599/admet.3.2.167 151 figure 6. (a) ftir spectra of kfn, (b) pcl, (c) microspheres differential scanning calorimetry (dsc) dsc was carried out to find possible interaction between the drug and polymer. this was performed for pure drug (a), pure polymer (b) and drug loaded microspheres (c) using a differential scanning calorimeter (dsc60, shimadzu). (fig. 8) confirms that there was no interaction between drug and polymer. figure 7. (a) xrd spectra of kfn, (b) pcl, (c) microspheres wagh and naik admet & dmpk 3(2) (2015) 141-153 152 figure 8. (a) dsc spectra of kfn, (b) pcl, (c) microspheres in vitro drug dissolution studies pcl is water insoluble degradable polymer and able to control drug release from microspheres. fig. 9 shows the cumulative % drug release as a function of the dissolution time from the kfn loaded microspheres. the release of kfn was evaluated using phosphate buffer (ph 6.8) as the release media. sustained drug release was observed in the range of 14.52 – 91.81 % in 12 h study of all the experimental runs. figure 9. cumulative % drug release admet & dmpk 3(2) (2015) 141-153 ketoprofen embedded polycaprolactone microspheres doi: 10.5599/admet.3.2.167 153 conclusions this work shows kfn loaded pcl polymeric microspheres were successfully prepared. effects of various independent formulation or process variables and dependent variables had been screened out by employing pbd screening design. the mathematical model for the response developed using statistical analysis quantitatively describes the influence of the selected variables on the responses under study. the % ee of the microspheres was ranged from 31.18% to 96.81%. the size of microspheres was found to be in the range of 20 μm to 293 μm was confirmed by sem analysis. the absence of drug polymer interactions was confirmed by ftir spectroscopy and dsc. the xrpd showed the distribution of drug in the polymer at molecular level. drug release from microspheres showed a sustained release i.e. 14.52 % – 91.81 % over a period of 12 h in a phosphate buffer (ph 6.8). microspheres prepared by such methods may represent a promising approach for efficient encapsulation and sustained drug release of kfn. acknowledgments authors are very much thankful to technical education quality improvement program (teqip-ii), mhrd, new delhi for providing financial assistance to carry out this research work. conflict of interest: ‘all authors are hereby declare that they have no conflict of interest’ animal studies: 'no animal or human studies were carried out by the authors for this article' references [1] m. sandile, khamanga, p. natalie, dissolut. technol. 2 (2009) 15-22. [2] b. deore, h. mahajan, u. deore, int. j. chem. tech. res. 1(3) (2009) 634-42. [3] j. jeevana, g. sunit, j. young pharma. 1 (2009) 24-27. [4] g.e. granero, c. ramachandran, g.l. amidon, pharmazie 61 (2006) 673-676. [5] y. tomofumi, o. hiracu, m. yoswharu, yakugaku zashi 121(3) (2001) 239-245. [6] r.k. deshmukh, j.b. naik, mat. sci. eng. c 36(c) (2014)320-328. [7] r.k. deshmukh, j.b. naik, j. pharm. innov. 8 (2013) 276-287. [8] p. wang, z.w. wang, z.c. wu, chem. eng. j. 193 (2012) 50–58. [9] r.s. liu, y.j. tang, bioresour. technol. 101 (2010) 3139–3146. [10] s. el-malah nazzal, int. j. pharm. 309 (2006) 163–170. [11] s.b. imandi, v.v.r. bandaru, s.r. somalanka, s.r. bandaru, h.r. garapati, bioresour. technol. 99 (2008) 4445–4450. [12] r.l. plackett, j.p. burman, biometrika 33 (1946) 305–325. [13] j. murray. statistical experimental design, data analysis and nonlinear optimisation, john wiley and sons, new york, 1994. [14] j.b. naik, a.b. lokhande, s. mishra, r.d. kulkarni, int. j.pharma biosci. 3(4) (2012) 573-590. [15] indian pharmacopoeia commission ghaziabad, indian pharmacopoeia (2007) pp-480. [16] a.k. nayak, d. pal, j. pradhan, m.s. hasnain, int. j. biol. micromol. 54 (2013) 144–154. ©2015 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ admet & dmpk x(y) (20xx) pp-pp; doi: 10 doi: 10.5599/admet.1.4.18 83 admet & dmpk 1(4) (2013) 83-83; doi: 10.5599/admet.1.4.18 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index editorial board member: brief biography christopher chan dr. christopher chan is currently senior engineer of the office of medical device evaluation (omde), center for measurement standards, industrial technology research institute (itri), taiwan and office-in-charge for the medical device human factors/usability engineering laboratory of omde. prior to joining omde in 2010, he was a senior consultant of the hong kong productivity council and had been the secretariat to the hong kong medical and healthcare device manufacturers association. before moving to hong kong in 2005, christopher was postdoctoral researcher in leiden institute of chemistry working on bionanotechnology and blue copper protein; and at the department of molecular biology and biotechnology working on cell cycle check point control. christopher received his d.phil in 1996 from the department of biochemistry, university of oxford under professor stuart ferguson on molecular bioenergetics related to the process of bacterial denitrification. the industrial technology research institute is a not-for-profit organization established by the taiwan government to assist the industrial development of taiwan, and the office of medical device evaluation is a unit under the center for measurement standards that currently can perform medical device good manufacturing practice audit, and class ii ivd product technical review under the authorization of the food and drug administration of taiwan. omde is also the only us fda 510k approved third party reviewer registered in asia. in addition to medical device evaluation omde also undertake medical device policy study projects to support the development and harmonization medical device regulation. christopher’s current interests include medical device evaluation technology in particularly human factors and usability and regulatory science. ©2013 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution licence (http://creativecommons.org/licenses/by/3.0/) http://www.pub.iapchem.org/ojs/index.php/admet/index http://creativecommons.org/licenses/by/3.0/ synthesis of silymarin−selenium nanoparticle conjugate and examination of its biological activity in vitro doi: http://dx.doi.org/10.5599/admet.1023 255 admet & dmpk 9(4) (2021) 255-266; doi: https://doi.org/10.5599/admet.1023 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper synthesis of silymarin−selenium nanoparticle conjugate and examination of its biological activity in vitro sergey a. staroverov 1,2 , sergey v. kozlov 1 , alexander s. fomin 2 , konstantin p. gabalov 2 , vitaliy a. khanadeev 2 , dmitry a. soldatov 1 , ivan yu. domnitsky 1 , lev a. dykman 2 , sergey v. akchurin 3 , olga i. guliy 1,2* 1 saratov state agrarian university, saratov, russian federation 2 institute of biochemistry and physiology of plants and microorganisms, russian academy of sciences, saratov, russian federation 3 russian state agrarian university moscow timiryazev agricultural academy, moscow, russian federation *corresponding author: e-mail: guliy_olga@mail.ru. received: june 14, 2021; revised: october 21, 2021; available online: november 14, 2021 abstract silymarin (sil) was conjugated to selenium nanoparticles (senps) to increase sil bioavailability. the conjugates were monodisperse; the average diameter of the native senps was ~ 20-50 ± 1.5 nm, whereas that of the conjugates was 30-50 ± 0.5 nm. the use of senps to increase the bioavailability of sil was examined with the mh-22a, epnt-5, hela, hep-2, and spev-2 cell lines. the epnt-5 (glioblastoma) cells were the most sensitive to the conjugates compared to the conjugate-free control. the conjugates increased the activity of cellular dehydrogenases and promoted the penetration of sil into the intracellular space. possibly, senps play the main part in sil penetration of cells and sil penetration is not associated with phagocytosis. thus, senps are promising for use as a sil carrier and as protective antigens. ©2021 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords selenium nanoparticles; silymarin, bioavailability; cell lines. introduction nanoparticles (nps) are natural or chemically synthesized ultradispersed materials sized between 1 and 200 nm [1-2]. various nanostructures, including polymers, dendrimers, liposomes, metal nanoparticles (ag, au, ce, cu, eu, fe, se, ti, y, etc.), and siliconand carbon-based nanomaterials have been successfully used as therapeutic agents and drug carriers [3–10]. the properties of nps, such as their small size, large surface area, surface charge, chemical composition, and multifunctionality, make them unique drug carriers. much research attention has been paid to synthesized selenium nanoparticles (senps). senps have anticancer activity and are less toxic than se salts [11-12]. senps have been used as antimicrobial agents [13] and in the treatment of various diseases, including cancer, diabetes, inflammations, liver fibrosis, and drug-induced poisoning [14]. the main obstacle to the widespread use of se is its low therapeutic index [15]. http://dx.doi.org/10.5599/admet.1023 https://doi.org/10.5599/admet.1023 http://www.pub.iapchem.org/ojs/index.php/admet/index https://www.scopus.com/affil/profile.uri?afid=60069979 mailto:guliy_olga@mail.ru http://creativecommons.org/licenses/by/4.0/ olga guliy et al. admet & dmpk 9(4) (2021) 255-266 256 for example, huang et al. [16] showed that small (5–15 nm) senps are well able to scavenge free radicals. at < 0.5 mm, senps had an excellent antioxidant effect. compared to inorganic and organic se compounds, senps are more active biologically. however, their main disadvantage is their poor cellular penetration. attempts have been made to solve this problem by conjugating the nanoparticles to various bioactive substances. this approach is a good foundation for cancer therapy. for example, senps synthesized with quercetin and gallic acid have antioxidant, antimicrobial, and antitumor activities [17]. surface ligands control the size and stability of senps and improve their cancer selectivity, cellular uptake, bioavailability, and biological activity. the use of amphoteric ligands (polyethylene glycol, peg) to make senps has been described [18]. in another study, senps were conjugated to a synthesized cyclic peptide that showed improved penetration into sk-ov-3 ovarian adenocarcinoma cells [19]. consequently, senps can be used as nanoscale delivery vehicles for differentially charged biomolecules and anticancer drugs. specifically, senp coated with 5-fluorouracil (5-fu) exhibited increased anticancer activity in a375 cells [20]. senps in combination with irinotecan increased the antitumor activity both in vitro and in vivo [11]. the combination of adriamycin and senps proved a powerful approach to cancer chemotherapy. low concentrations of this combination had a synergistic anticancer activity in bel7402 liver cancer cells [21]. antihepatocarcinoma effects were observed in hepg2 cells after the use of anisomycin-functionalized senps, with nps delaying the cell cycle in g0/g1 [22]. senps (25 μg/ml) in combination with doxorubicin (2.5 μg/ml) showed a superior apoptotic effect in mcf-7 human breast cancer cells compared to the drug used alone. senps also exhibit antitumor activity in in vivo-induced mcf-7 cells [23]. biogenic senps synthesized from an l. plantarum strain were immunostimulatory in balb/c mice with breast cancer cells. oral senp treatment significantly increased the production of proinflammatory cytokines such as ifn-γ, il2, il-12, and tnf-α, and it also enhanced the delayed hypersensitivity response. senps reduced tumor volume and increased survival in mice treated with senps due to increased immune response [19]. of particular interest is the use of senps to increase the availability of drugs such as silymarin (sil). sil has a liver-protecting effect, reducing the concentration of free radicals and the degree of damage to the cell membranes, but its bioavailability is limited due to its poor water solubility [24-25]. the aim of this work was to conjugate senps with sil and analyze the biological activity of the resultant conjugate in vitro. experimental the conjugate was prepared by the following scheme: 0.39 g of sil was dissolved in 30 ml of 0.1 m sodium hydroxide, and the solution was thoroughly mixed. then, 2 ml of selenous acid from a 0.1 m solution and 1 ml of 0.03 m l-cysteine were added with vigorous stirring. the mixture was stirred for 1 h and lyophilized. characterization of selenium nanoparticles senps size measurements were monitored by dynamic light scattering (dls) on the malvern zetasizer nanoparticle characterization system (malvern instruments, great britain) with the he-ne laser (wavelength = 633 nm, power = 4 mw). the measurements were carried out at a fixed angle of 173° at 25 °c. the values of the hydrodynamic radius obtained by registering dynamic light scattering amounted to a peak of 25.2±1.5 nm for native gold particles. senps formations were monitored by spectrophotometry as described in [26]. spectrophotometric admet & dmpk 9(4) (2021) 255-266 selenium nanoparticle conjugates and their activity in vitro doi: http://dx.doi.org/10.5599/admet.1023 257 evaluation of nanoparticles was carried out using a uv-vis specord bs 250 spectrophotometer (analytik jena, germany) at a wavelength of 520 nm (a520). it is known that the maximum absorption of native nanoparticles was 526 nm, which is typical for particles with the property of surface plasmon resonance. the average size, morphology and uniformity of the synthesized senps also were studied by transmission electron microscopy (tem) imaging as described in [26]. tem images were recorded on a libra 120 electron microscope (carl zeiss, germany). then these images were analyzed by imagej software to determine the morphology and size distribution of the senps with a minimum of 100 nanoparticles measured. the sil concentration of the conjugates was measured by high-performance liquid chromatography at 288 nm, by using a stayer chromatograph (akvilon, russia) fitted with a uv detector. measurements were made with a 2.0 × 60-mm luna c-18 column, with liquid-chromatography-grade acetonitrile: 1 % acetic acid solution (2:3, v/v) as the eluent. biological activity of the conjugate the biological activity of the conjugate was tested with the mh-22a, epnt-5, hela, hep-2, and spev-2 cell lines. the cells were cultured in dulbecco’s modified eagle’s medium (dmem; biolot) supplemented with 10 % fetal bovine serum (biolot), penicillin (100 u/ml; gibco), streptomycin (100 μg/ml; gibco), and lglutamine (292 μg/ml; gibco). respiratory activity was measured conventionally [27] by the ability of cells to reduce nitrotetrazolium blue [3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide, mtt (sigma–aldrich)] to formazan (mtt test). for each biological sample 10 replicates were analyzed. affinity selection of miniantibodies from a phage library miniantibodies were selected from a phage library of sheep antibodies (griffin.1), kindly provided by professor w.j. harris (aberdeen university, uk) [28]. for the selection of phage carrying anti-si l antibodies, a western s membrane (size, 1 × 1 cm; sigmaaldrich, usa) was used as a solid phase. the membrane was incubated overnight in a solution of the antigen (concentration, 1 mg/ml) at 4 °c. the antigen-coated membrane was then incubated in a 2 % solution of fat-free powdered milk for 30 min and placed in 1 ml of 10 mm tbs-t buffer (ph 7.2) containing the library phage at 10 12 phage particles/ml. after the membrane was incubated overnight at 4 °c, it was washed five times with tbs-t buffer for 10 min. the bound phage was eluted with 7.18 m triethylamine (1 ml), and the ph was adjusted to 7.2 with 1 m trishcl. the eluted phage particles were used to infect e. coli xl-1. the infected e. coli cells were grown overnight at 37 °c in 10 ml of 2yt liquid medium containing 100 μg/ml of ampicillin and 1 % glucose. one liter of the 2yt medium contained 16 g of tryptone, 10 g of yeast extract, and 5 g of nacl. a 1/100 portion (100 μl) of the resultant culture was inoculated into 10 ml of the 2yt medium, and the culture was grown in a thermostated shaker for 6 h until an absorbance (а600) of 0.3 (∼10 12 cells/ml) was achieved. this was followed by the addition of helper phage м13к07 and by incubation at 37 °с for 1 h. after incubation, the cells were sedimented by centrifugation at 2000 g for 10 min. the cell sediment was resuspended in 50 ml of the 2ty medium containing 100 μg/ml of ampicillin, 50 μg/ml of kanamycin, and 100 μg/ml of isopropyl-β-d-thiogalactoside and was grown overnight at 37 °с in a thermostated shaker. the overnight cell culture was again centrifuged at 3000 g for 40 min. to the supernatant liquid containing phage particles, a 1/5 volume of 20 % peg 6000/2.5m nacl was added, and the mixture was incubated on ice for 1.5 h. phage particles were sedimented by centrifugation at 8000 g for 10 min, and the sediment was resuspended in 5 ml of te buffer (1/10 the original culture volume; ph 7.5). the resultant preparation was http://dx.doi.org/10.5599/admet.1023 olga guliy et al. admet & dmpk 9(4) (2021) 255-266 258 clarified by centrifugation under the same conditions, and the phage particles were again precipitated by adding a 1/5 volume of peg 6000/nacl (1 ml) and were centrifuged. the sediment was suspended in 1 ml of te buffer. the phage particle concentration was calculated spectrophotometrically using the ratio a269 = 30 ~ 2 × 10 14 phage particles/ml. the resulting phage particles were used for the next two rounds of selection. these were done under the same conditions but with a shorter incubation time and with fewer particles at the stage of their interaction with the immobilized antigen (1.5 h and 1 h of incubation at room temperature and 10 11 and 10 10 phage particles for the 2 nd and 3 rd rounds, respectively). dot immunoassay the specificity of the obtained phage antibodies was tested by dot immunoassay. a standard sil solution was applied to the western s membrane, and the membrane was blocked for 1 h with 2 % fat-free powdered milk in phosphate buffer. the membrane was dipped into a solution of specific phage, diluted to 10 13 particles/ml of 10 mm phosphate buffer, and incubated at room temperature for 1 h. after that, the membrane was washed free from nonspecifically bound miniantibodies and dipped into a solution of colloid gold conjugated to rabbit antiphage antibodies (a520 = 0.5) [29]. phage labeling the sil-specific miniantibodies (1 × 10 12 pfu) were resuspended in 100 ml of 0.3 m nahco3 solution (ph 8.6) containing 1 mg/ml of tetramethylrhodamine isothiocyanate. the reaction was carried out in the dark at room temperature for 1 h. after incubation, the volume of labeled phage was adjusted to 1 ml with pbs, and the phage was purified by dialysis. finally, the fluorochrome-labeled phage was resuspended in 200 ml of phosphate buffered saline (pbs) and was frozen [30]. isolation of peritoneal macrophages for peritoneal macrophages, the animals were killed and then fixed on their backs. an incision was made along the midline of the anterior abdominal wall, and the skin flap was carefully separated, with care taken to keep the peritoneum intact. after a puncture had been made with a needle connected to a syringe, 50 ml of pbs, ph 7.2, was injected into the peritoneal cavity. the anterior abdominal wall was then gently massaged, and after 5–7 min, peritoneal fluid was collected with a pasteur pipet through a cut made in the peritoneum and filtered into a test tube through a nylon filter. the cells were washed three times by centrifugation in pbs at 750 g, after which they were redissolved in 1 ml of pbs and counted in a goryaev chamber. peritoneal macrophages were cultured by standard procedures [31], as described in table 1. for splenic lymphocytes, an incision was made along the white line of the peritoneum after peritoneal macrophages had been isolated, and the spleen was removed. the spleen was minced with scissors, and the tissue pieces were mashed through a fine sieve into a petri plate containing sterile pbs. the resulting suspension was subjected to ficoll–urografin density-gradient centrifugation. the lymphocyte ring was collected into a new test tube. the lymphocytes were washed three times by centrifugation in pbs, ph 7.4, at 750 ×g for 10 min, and the cell pellet was redissolved in 1 ml of pbs. the lymphocytic cells were counted with a haemascreenvet hematology analyzer (hospitex diagnostics, italy) [31], as presented in table 1. the cellnanoparticle interaction was visualized on a dmlb fluorescence microscope (leica, germany; excitation at 544 nm, emission at 570 nm). the animals were cared for and handled in compliance with the requirements of the ministry of health of the russian federation (work of experimental biology clinics) and with the european convention for the admet & dmpk 9(4) (2021) 255-266 selenium nanoparticle conjugates and their activity in vitro doi: http://dx.doi.org/10.5599/admet.1023 259 protection of vertebrate animals used for experimental and other scientific purposes. table 1. scheme for the growth of macrophages and lymphocytes with se/sil and the method of their staining with silspecific miniantibodies for subsequent microscopy group control sil solution sil/se composition 10 6 cells/ml in dmem + hepes with 10 % embryo serum 10 6 cells/ml in dmem + hepes with 10 % embryo serum 10 6 cells/ml in dmem + hepes with 10 % embryo serum step 1 addition of a sil solution (sil concn, 1 μg/ml) addition of the se/sil conjugate (sil concn, 1 μg/ml) step2 incubate at 37 °c for 2 h, spin down the cells, resuspend the cells in a fresh medium step 3 fix aliquots of the cells on glass with acetone for 2 min step 4 apply phage (1012 per ml of pbs with 2 % bsa) for 1 h on both smears of each group; rinse twice with pbs step 5 rinse for 10 min with trisglycine buffer (ph 2.5); rinse twice with pbs; subject to microscopy (excitation at 544 nm, emission at 570 nm) statistics data were processed by the standard procedures integrated in excel 2007 software (microsoft corp., usa). results the use of nano-se enables researchers to make materials with improved physicochemical characteristics. in this context, many methods for preparing nanosized se have been advanced [32,33]. these include green synthesis methods. green synthesis is reliable, sustainable, and eco-friendly, and it avoids the production of unwanted or harmful byproducts [34]. green synthesis of nanoparticles aims at minimizing waste and implementing sustainable processes. in recent years, green processes using mild reaction conditions and nontoxic precursors have been used in nanoresearch to promote environmental sustainability [34]. сontributions of researchers from different countries have led to remarkable progress in green synthesis. the versatility of green chemistry allows the preparation of a wide range of organic and inorganic nanomaterials with many promising applications [35]. the green analytical procedure index (gapi) evaluates the green character of an entire analytical methodology, from sample collection to final determination. it was created with a tool such as the national environmental methods index (nemi) to provide general and qualitative information [36]. the existing methods are rated on the basis of four criteria that refer to the properties of reagents or wastes used in this method [37]. in this work, senps were conjugated to sil, and because the particles were reduced by using sil (a milk thistle extract), we believe that according to the nemi criteria, this method falls within green synthesis. after synthesis, it is very important to characterize the resulting nanomaterials carefully. the diameter of the synthesized senps was measured by tem and dynamic light scattering. fig. 1 shows that the conjugates were monodisperse and that the average diameter of the native senps was ~ 20-50 ± 1.5 nm with a peak ~ 25.2 ± 1.5 nm. the tem results agree with the hydrodynamic diameter data. according to the literature, the most promising is the use of 2070-nm senps, the main advantage is their low toxicity, which permits them to be used in doses much greater than the daily requirement [33]. therefore, the synthesized senps were used for conjugation with sil. the conjugate diameter was 30-50 ± 0.5 nm. highperformance liquid chromatography showed that the sil concentration in the conjugate was 2 mg/ml. on this basis, the concentration of the conjugate was calculated for its further use in biological research. http://dx.doi.org/10.5599/admet.1023 olga guliy et al. admet & dmpk 9(4) (2021) 255-266 260 figure 1. size distribution of senps, as found by dls (a) and tem (b). we next examined the cytotoxic effect of se/sil (figure 2) on several cancer cell lines. for this purpose, the mh-22a, epnt-5, hela, hep-2, and spev-2 lines were treated with se/sil (sil concentration, 0.8957 μg/ml) for 24 h, and the change in absorbance was measured. there were two controls. in one of these, the cell lines were exposed to the above concentrations of sil. in the other, the cell lines were unexposed to senps or sil. as presented in fig. 2 (f), the most sensitive to the conjugate was the epnt-5 glioblastoma line. its activity decreased by 89 % compared to the control (cells grown without the conjugate). the sil concentration, in this case, was 7.13 μg/ml. the other lines were sensitive to the higher 14.25 μg/ml sil concentration. at 28.5 μg/ml of sil, 85 % of the cells of the mh-22a line died. we next investigated the effect of se/sil on the cells of the reticuloendothelial, or macrophage system because this cell system is responsible for the barrier, phagocytic, and metabolic functions. fig. 3 (а) shows that the conjugate increased cell dehydrogenase activity by 3.4 times (p = 0.0003), as compared to se alone, which increased cellular respiratory activity by 2.5 times (p = 0.0002). fig. 3 (b) shows the effect of se/sil on mouse splenocytes. the conjugate increased cell dehydrogenase activity by 2.5 times (p = 0.0008) compared to se alone, which increased cellular respiratory activity by 1.9 times (p = 0.0003). fluorescence microscopy was used to assess the interaction of se/sil with the cells of the reticuloendothelial system (macrophages and lymphocytes) of laboratory rats. the macrophages and lymphocytes were cultured as described in materials and methods. for visualization, we used antibodies prepared beforehand by phage display technology. the specificity of the phage antibodies was examined by dot immunoassay. fig. 4 shows that the binding of the selected sil-specific antiphage antibodies in the dot immunoassay was detected up to a sil concentration of 0.1 μg/ml. admet & dmpk 9(4) (2021) 255-266 selenium nanoparticle conjugates and their activity in vitro doi: http://dx.doi.org/10.5599/admet.1023 261 figure 2. changes in the cytotoxic effect of senps/sil on the cancer cell lines: epnt-5 (a) hela (b); hep-2 (c); mh22a (d); spev-2 (e): 1 – control (cancer cell lines were unexposed to senps/sil); 2 – cancer cell lines with sil; 3 – cancer cell lines with senps/sil. (f) total changes in the cytotoxic effect of the senps + sil preparation on various tumor cell lines without controls: 1 – epnt-5; 2 – hela; 3 – hep-2; 4 – mh22a; 5 – spev-2. http://dx.doi.org/10.5599/admet.1023 olga guliy et al. admet & dmpk 9(4) (2021) 255-266 262 figure 3. changes in the respiratory activity of mouse peritoneal macrophages (а) and splenocytes (b) grown with se/sil: 1 – control (grown without senps or sil); 2 – grown with sil; 3 – grown with se; 4 – grown with se/sil. figure 4. dot immunoassay with selected sil-specific antiphage antibodies (pvdf membrane, development with rabbit antibodies against the entire library, staining with protein a/colloidal gold). anti-s mab, titrated sil (a); library, application of phage library dilutions (b); helper phage, application of helper phage (10 12 ). the fluorescence microscopy results for the interaction of se/sil with the cells of the reticuloendothelial system are shown in fig. 5a for macrophages and in fig. 5b for lymphocytes. the red fluorescence indicates the presence of sil in the cells. because there was no fluorescence when sil acted alone, one can speculate that senps are crucial for the sil penetration of the cells. probably, senps facilitate the penetration of sil into the intracellular space. because fluorescence was observed when senps were grown with lymphoid cells, the penetration of senps is possibly unrelated to phagocytosis. thus, senps/sil is a promising anticancer conjugate that affects cellular immunity by causing the stimulation of both macrophages and splenocytes. admet & dmpk 9(4) (2021) 255-266 selenium nanoparticle conjugates and their activity in vitro doi: http://dx.doi.org/10.5599/admet.1023 263 figure 5. visualization of the se/sil interaction with cells of the reticuloendothelial system [the macrophages (a) and lymphocytes (b) of laboratory rats]: 1 – control (cells cultured without se/sil); 2 – cells cultured with sil; 3 – cells cultured with se/sil. discussion nanomaterials, including senps, can increase drug bioavailability. in this work, the bioavailability of sil, used as an example, was increased by conjugation with senps. nanoparticles can penetrate cells by bypassing protective barriers (including the bloodbrain and placental barriers) and can selectively accumulate in different parts of a living organism [2]. they can lead to inflammation, fibrosis, and cell and organ dysfunction, and they can also cause pathological disorders [38,39]. this situation imposes requirements for the development of adequate methodological approaches to the study of risks arising from the contact of biological systems with nanomaterials [40]. therefore, any negative consequences should be analyzed before nanoparticle use, and risks should be weighed against any possible benefits. although se is a known antioxidant, it is a toxic element [41]. when choosing the chemical form of se, one should pay attention to its effectiveness and safety. the most promising is the use of senps, the size of which is 2070 nm, because they are less toxic than other se forms. therefore, nanoparticles of this size can be used in doses that largely exceed the daily requirement [33]. in this work, 2050-nm senps were used to increase the bioavailability of sil. the present study has shown promising prospects for the synthesis of senps using silymarin. the epnt-5 line was the most sensitive to the conjugate: the cell activity decreased by 89 % compared to the control. the conjugate increased cell dehydrogenase activity and promoted the penetration of sil into the intracellular space. senps are crucial to sil penetration, and the process of penetration is unrelated to phagocytosis. in size range used, senps are a promising platform for protective antigens and immunomodulators. the use of 2050-nm senps avoids the toxicity of se while increasing the bioavailability of sil and possibly contributes to the anticancer treatment of the liver. the promise of using se in cancer treatment has been repeatedly http://dx.doi.org/10.5599/admet.1023 olga guliy et al. admet & dmpk 9(4) (2021) 255-266 264 demonstrated [41]. the use of senps in combination with sil opens up the possibility of utilizing se for liver cancer treatment. the obtained results are pioneering and will enable the range of anticancer drugs for the liver to be expanded. acknowledgements: this work was supported in part by the russian science foundation (grant no. 19-1400077). conflict of interest: the authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. abbreviations: dls – dynamic light scattering gapi – green analytical procedure index nps – nanoparticles nemi – national environmental methods index pbs – phosphate buffered saline peg – polyethylene glycol pfu – plaque-forming units senps – selenium nanoparticles sil – silymarin se/sil – sil-linked senps tem – transmission electron microscopy uv– ultraviolet references [1] d. peer, j.m. karp, s. hong, o.c. farokhzad, r. margalit, r. langer. nanocarriers as an emerging platform for cancer therapy. nature nanotechnology 2(12) (2007) 751–760. https://doi.org/10.1038/nnano.2007.387. [2] s. bakand, a. hayes. toxicological considerations, toxicity assessment, and risk management of inhaled nanoparticles. international journal of molecular sciences 17(6) (2016) 929. https://doi.org/10.3390/ijms17060929. [3] m. liong, j. lu, m. kovochich, t. xia, s.g. ruehm, a.e. nel, f. tamanoi, j.i zink. multifunctional inorganic nanoparticles for imaging, targeting, and drug delivery. acs nano 2(5) (2008) 889–896. https://doi.org/10.1021/nn800072t. [4] i. brigger, c. dubernet, p. couvreur. nanoparticles in cancer therapy and diagnosis. advanced drug delivery reviews. advanced drug delivery reviews 54(5) (2002) 631–651. https://doi.org/10.1016/s0169-409x(02)00044-3. [5] j. das, j.w. han, y.j. choi, h. song, s.-g. cho, c. park, h.g. seo, j.h. kim. cationic lipid-nanoceria hybrids, a novel nonviral vector-mediated gene delivery into mammalian cells: investigation of the cellular uptake mechanism. scientific reports 6(6) (2016) 29197. https://doi.org/10.1038/srep29197. [6] y. sun, y. xia. shape-controlled synthesis of gold and silver nanoparticles. science 298(5601) (2002) 2176–2179. https://doi.org/10.1126/science.1077229. [7] r.h. baughman, a.a. zakhidov, w.a. de heer. carbon nanotubes--the route toward applications. science 297(5582) (2002) 787–792. https://doi.org/10.1126/science.1060928. [8] s. sangomla, m.a. saifi, a. khurana, c. godugu. nanoceria ameliorates doxorubicin induced cardiotoxicity: possible mitigation via reduction of oxidative stress and inflammation. journal of trace elements in medicine and biology 47 (2018) 53–62. https://doi.org/10.1016/j.jtemb.2018.01.016. [9] p. kumari, m.a. saifi, a. khurana, c. godugu. cardioprotective effects of nanoceria in a murine model of cardiac remodeling. journal of trace elements in medicine and biology 50 (2018) 198–208. https://en.wikipedia.org/wiki/nature_nanotechnology https://doi.org/10.1038/nnano.2007.387 https://doi.org/10.1038/nnano.2007.387 https://doi.org/10.3390/ijms17060929 https://doi.org/10.3390/ijms17060929 https://doi.org/10.1021/nn800072t https://doi.org/10.1016/s0169-409x(02)00044-3 https://doi.org/10.1016/s0169-409x(02)00044-3 https://nature.com/srep https://doi.org/10.1038/srep29197 https://doi.org/10.1126/science.1077229 https://doi.org/10.1126/science.1060928 https://doi.org/10.1016/j.jtemb.2018.01.016 admet & dmpk 9(4) (2021) 255-266 selenium nanoparticle conjugates and their activity in vitro doi: http://dx.doi.org/10.5599/admet.1023 265 https://doi.org/10.1016/j.jtemb.2018.07.011. [10] e. cremonini, e. zonaro, m. donini, s. lampis, m. boaretti, s. dusi, p. melotti, m. m. lleo, g. vallini. biogenic selenium nanoparticles: characterization, antimicrobial activity and effects on human dendritic cells and fibroblasts. microbial biotechnology 9(6) (2016) 758–771. https://doi.org/10.1111/1751-7915.12374. [11] w. liu, x. li, y.-s. wong, w. zheng, y. zhang, w. cao, t. chen. selenium nanoparticles as a carrier of 5-fluorouracil to achieve anticancer synergism. acs nano 6(8) (2012) 6578–6591. https://doi.org/10.1021/nn202452c. [12] y. huang, l. he, w. liu, c. fan, w. zheng, y.-s. wong, t. chen. selective cellular uptake and induction of apoptosis of cancer-targeted selenium nanoparticles. biomaterials 34(29) (2013) 7106–7116. https://doi.org/10.1016/j.biomaterials.2013.04.067. [13] x. huang, x. chen, q.chen, q. yu, d. sun, j. liu. investigation of functional selenium nanoparticles as potent antimicrobial agents against superbugs. acta biomaterialia 30 (2016) 397–407. http://doi.org/10.1016/j.actbio.2015.10.041. [14] a. khurana, s. tekula, m.a. saifi, p. venkatesh, c. godugu. therapeutic applications of selenium nanoparticles. biomedicine & pharmacotherapy 111 (2019) 802–812. https://doi.org/10.1016/j.biopha.2018.12.146. [15] a. robby, j.m. petros, de simone. strategies in the design of nanoparticles for therapeutic application. nature reviews drug discovery 9 (2010) 615–627. https://doi.org/10.1038/nrd2591. [16] b. huang, j. zhang, j. hou, c. che. free radical scavenging efficiency of nano-se in vitro. free radical biology and medicine 35(7) (2003) 805–813. https://doi.org/10.1016/s0891-5849(03)00428-3. [17] y. xia, p. you, f. xu, j. liu, f. xing. novel functionalized selenium nanoparticles for enhanced antihepatocarcinoma activity in vitro. nanoscale research letters. 10(1) (2015) 349. https://doi.org/10.1186/s11671-015-1051-8. [18] s. zheng, x. li, y. zhang, q. xie, y.s. wong, w. zheng, t. chen. peg-nanolized ultrasmall selenium nanoparticles overcome drug resistance in hepatocellular carcinoma hepg2 cells through induction of mitochondria dysfunction. international journal of nanomedicine 7 (2012) 3939–3949. https://dx.doi.org/10.2147/ijn.s30940. [19] a.n. shirazi, r.k. tiwari, d. oh, b. sullivan, a. kumar, y.a. beni, k. parang. cyclic peptide−selenium nanoparticles as drug transporters. molecular pharmaceutics 11(10) (2014) 3631–3641. https://doi.org/10.1021/mp500364a. [20] m.h. yazdi, b. varastehmoradi, e. faghfuri, f. mavandadnejad, m. mahdavi, a.r. shahverdi. adjuvant effect of biogenic selenium nanoparticles improves the immune responses and survival of mice receiving 4t1 cell antigens as vaccine in breast cancer murine model. journal of nanoscience and nanotechnology 15(12) (2015) 10165–10172. https://doi.org/10.1166/jnn.2015.11692. [21] f. gao, q. yuan, l. gao, p. cai, h. zhu, r. liu, y. wang, y. wei, g. huang, j. liang, x. gao. cytotoxicity and therapeutic effect of irinotecan combined with selenium nanoparticles. biomaterials 35(31) (2014) 8854–8866. https://doi.org/10.1016/j.biomaterials.2014.07.004. [22] l. tan, x. jiang, y. zhang, h. tang, s. yao, q. xie. in vitro study on the individual and synergistic cytotoxicity of adriamycin and selenium nanoparticles against bel7402 cells with a quartz crystal microbalance. biosensors and bioelectronics 24(7) (2009) 2268–2272. https://doi.org/10.1016/j.bios.2008.10.030. [23] a.k. mittal, s. kumar, u.c. banerjee. quercetin and gallic acid mediated synthesis of bimetallic (silver and selenium) nanoparticles and their antitumor and antimicrobial potential. journal of colloid and interface science 431 (2014) 194–199. https://doi.org/10.1016/j.jcis.2014.06.030. [24] n. vargas-mendoza, e. madrigal-santillán, á.morales-gonzález, j. esquivel-soto, c. esquivel-chirino, m. garcía-luna y gonzález-rubio, j.a. gayosso-de-lucio, j.a. morales-gonzález. hepatoprotective effect of silymarin. world journal of hepatology 6(3) (2014) 144–149. https://doi.org/10.4254/wjh.v6.i3.144. http://dx.doi.org/10.5599/admet.1023 https://doi.org/10.1016/j.jtemb.2018.07.011 https://doi.org/10.1111/1751-7915.12374 https://doi.org/10.1111/1751-7915.12374 https://doi.org/10.1021/nn202452c https://doi.org/10.1021/nn202452c https://doi.org/10.1016/j.biomaterials.2013.04.067 http://dx.doi.org/10.1016/j.actbio.2015.10.041 http://dx.doi.org/10.1016/j.actbio.2015.10.041 https://doi.org/10.1016/j.biopha.2018.12.146 https://doi.org/10.1016/j.biopha.2018.12.146 https://www.nature.com/nrd https://doi.org/10.1038/nrd2591 https://doi.org/10.1016/s0891-5849(03)00428-3 https://doi.org/10.1186/s11671-015-1051-8 https://doi.org/10.1186/s11671-015-1051-8 https://dx.doi.org/10.2147/ijn.s30940 https://doi.org/10.1021/mp500364a https://doi.org/10.1021/mp500364a http://www.aspbs.com/jnn/ http://www.aspbs.com/jnn/ https://doi.org/10.1166/jnn.2015.11692 https://doi.org/10.1016/j.biomaterials.2014.07.004 https://www.journals.elsevier.com/biosensors-and-bioelectronics/ https://doi.org/10.1016/j.bios.2008.10.030 https://doi.org/10.1016/j.bios.2008.10.030 https://doi.org/10.1016/j.jcis.2014.06.030 https://doi.org/10.4254/wjh.v6.i3.144 https://doi.org/10.4254/wjh.v6.i3.144 olga guliy et al. admet & dmpk 9(4) (2021) 255-266 266 [25] c. loguercio, d. festi. silybin and the liver: from basic research to clinical practice. world journal of gastroenterology 17(18) (2011) 2288–2301. https://doi.org/10.3748/wjg.v17.i18.2288. [26] n.g. khlebtsov, l.a. dykman. optical properties and biomedical applications of plasmonic nanoparticles. journal of quantitative spectroscopy and radiative transfer 111 (2010) 1–35. https://doi.org/10.1016/j.jqsrt.2009.07.012. [27] t. bernas, j.w. dobrucki. the role of plasma membrane in bioreduction of two tetrazolium salts, mtt, and ctc. archives of biochemistry and biophysics 380(1) 2000 108–116. https://doi.org/10.1006/abbi.2000.1907. [28] k.a. charlton, s. moyle, a.j.r. porter, w.j. harris. analysis of the diversity of a sheep antibody repertoire as revealed from a bacteriophage display library. journal of immunology 164 (2000) 6221– 6229. https://doi.org/10.4049/jimmunol.164.12.6221. [29] s.a. staroverov, a.a. volkov, a.s. fomin, v.n. laskavuy, p.v. mezhennyy, s.v. kozlov, s.v. larionov, m.v. fedorov, l.a. dykman, o.i. guliy. the usage of phage mini-antibodies as a means of detecting ferritin concentration in animal blood serum. journal of immunoassay and immunochemistry 36(1) (2015) 100–110. https://doi.org/10.1080/15321819.2014.899257. [30] k.a. kelly, p. waterman, r. weissleder. in vivo imaging of molecularly targeted phage. neoplasia 8(12) (2006) 1011–1018. http://doi.org/10.1593/neo.06610. [31] e.h. leiter. the nod mouse: a model for insulin dependent diabetes mellitus. current protocols in immunology (2001) chapter 15, unit 15.9. https://doi.org/10.1002/0471142735.im1509s24. [32] j.-s. zhang, x.y. gao, l.d. zhang, y.p. bao. biological effects of a nano red elemental selenium. biofactors 15(1) (2001) 27–38. https://doi.org/10.1002/biof.5520150103. [33] h. wang, j. zhang, h. yu. elemental selenium at nano size possesses lower toxicity without compromising the fundamental effect on selenoenzymes: comparison with selenomethionine in mice. free radical biology and medicine 42(10) (2007) 1524–1533. https://doi.org/10.1016/j.freeradbiomed.2007.02.013. [34] j. singh, t. dutta, k.-h. kim, m. rawat, p. samddar, p. kumar. ‘green’ synthesis of metals and their oxide nanoparticles: applications for environmental remediation. journal of nanobiotechnol. 16 (2018) 84. https://doi.org/10.1186/s12951-018-0408-4. [35] p.n. catalano, r.g. chaudhary, m.f. desimone, p.l. santo-orihuela. a survey on analytical methods for the characterization of green synthesized nanomaterials. current pharmaceutical biotechnology 22(6) (2021) 823–847. https://doi.org/10.2174/1389201022666210104122349. [36] j. płotka-wasylka. a new tool for the evaluation of the analytical procedure: green analytical procedure index. talanta 181 (2018) 204–209. https://doi.org/10.1016/j.talanta.2018.01.013. [37] a. gałuszka, p. konieczka, z.m. migaszewski, j. namies´nik. analytical eco-scale for assessing the greenness of analytical procedures. trends in analytical chemistry 37 (2012) 61-72. http://dx.doi.org/10.1016/j.trac.2012.03.013. [38] n. prajitha, s.s. athira, p.v. mohanan. bio-interactions and risks of engineered nanoparticles. environmental research 172 (2019) 98–108. https://doi.org/10.1016/j.envres.2019.02.003. [39] s. bakand, a. hayes, f. dechsakulthorn. nanoparticles: a review of particle toxicology following inhalation exposure. inhalation toxicology 24(2) (2012) 125–135. https://doi.org/10.3109/08958378.2010.642021. [40] m. van der zande, r. junker, x.f. walboomers, j.a. jansen. carbon nanotubes in animal models: a systematic review on toxic potential. tissue engineering reviews (part b) 17(1) (2011) 57–69. https://doi.org/10.1089/ten.teb.2010.0472. [41] s. misra, m. boylan, a. selvam, j.e. spallholz and m. björnstedt. redox-active selenium compounds— from toxicity and cell death to cancer treatment. nutrients 7 (2015) 3536–3556. https://doi.org/10.3390/nu7053536. ©2021 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) https://doi.org/10.3748/wjg.v17.i18.2288 https://doi.org/10.1016/j.jqsrt.2009.07.012 https://doi.org/10.1006/abbi.2000.1907 https://doi.org/10.1006/abbi.2000.1907 https://doi.org/10.4049/jimmunol.164.12.6221 https://doi.org/10.1080/15321819.2014.899257 http://doi.org/10.1593/neo.06610 https://doi.org/10.1002/0471142735.im1509s24 https://doi.org/10.1002/biof.5520150103 https://doi.org/10.1016/j.freeradbiomed.2007.02.013 https://doi.org/10.1016/j.freeradbiomed.2007.02.013 https://jnanobiotechnology.biomedcentral.com/ https://doi.org/10.1186/s12951-018-0408-4 https://doi.org/10.2174/1389201022666210104122349 https://doi.org/10.1016/j.talanta.2018.01.013 http://dx.doi.org/10.1016/j.trac.2012.03.013 http://dx.doi.org/10.1016/j.trac.2012.03.013 https://doi.org/10.1016/j.envres.2019.02.003 https://doi.org/10.3109/08958378.2010.642021 https://doi.org/10.3109/08958378.2010.642021 https://www.liebertpub.com/doi/full/10.1089/ten.teb.2010.0472 https://www.liebertpub.com/doi/full/10.1089/ten.teb.2010.0472 https://doi.org/10.1089/ten.teb.2010.0472 https://doi.org/10.3390/nu7053536 https://doi.org/10.3390/nu7053536 http://creativecommons.org/licenses/by/3.0/ manuscript doi: 10.5599/admet.1.4.10 63 admet & dmpk 1(4) (2013) 63-75; doi: 10.5599/admet.1.4.10 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index mini review estimating the “first in human” dose – a revisit with particular emphasis on oncology drugs kin tam faculty of health science, university of macau, macau, china e-mail: kintam@umac.mo; tel.: +853-8397-8444; fax: +853-8397-8500 received: june 30 th , 2013; revised: october 21 st , 2013; published: december 16, 2013 abstract the initial dose selection is one of the important steps for any investigative new drug (ind) entering the first clinical study in humans. in this mini review, we will discuss the no observed adverse effect level (noael) and the minimum anticipated biological effect level (mabel) approaches for the estimation of the first in human (fih) dose. particular attention will be placed on the development of the fih dose for oncology drugs. keywords initial dose; starting dose; clinical trial; noael; mabel; cytotoxic; anticancer; combination therapy study 1. introduction estimating the first in human (fih) dose is one of the initial steps in the clinical development of any molecule that has successfully gone through all of the hurdles in preclinical evaluations. it is an important parameter in the fih clinical trials, since a high starting dose may cause serious toxicity in volunteers, while a low starting dose could prolong the dose escalation/optimisation, leading to unnecessary delay in the clinical programs. adverse drug reactions do occur in clinical trials [1]. it is unethical to expose human subjects to poorly executed trials [2]. clearly, protecting the volunteers is paramount, especially in the first human trial using healthy individuals. designing the appropriate fih dose requires close collaboration among the pharmacokineticists, toxicologists, and preclinical scientists. data from preclinical experiments such as animal, modelling, and pharmacokinetic data help to determine the initial dose and dosing interval in early clinical development. in 2005, the us food and drug administration (fda) issued guidance on estimating the maximum safe starting dose in initial clinical trials for therapeutics in adult healthy volunteers, which provided a framework to carry out the estimation [3]. typically this involves the use of the no observed adverse effect level (noael) obtained from preclinical toxicology studies, which is then converted into a human equivalent dose (hed) from which a safety factor is applied to derive the maximum recommended start dose (mrsd). http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:kintam@umac.mo tam admet & dmpk 1(4) (2013) 63-75 64 the noael method is based on selecting a dose with minimal risk of toxicity, rather than selecting one with minimal pharmacology activity in humans. the approach appears to work well with new molecules that act on established targets and/or have pharmacology that is more or less understood. however, the life-threatening adverse events that occurred in 2006 after the fih administration of tng1412 (a novel cd28 super-agonist antibody) was a wake-up call to the industry, and led to immediate reactions from different regulators [4]. subsequently, the european medicines agency (emea) has issued new guidance [5], which underlines the need to better mitigate the risk regarding fih dosing, and proposes an improved approach to dose selection through the combined analysis of all of the pharmacology, safety and efficacy preclinical data. one of the main objectives of the new process is to define a starting dose that is expected to result in a minimum anticipated biological effect level (mabel), particularly for investigational medicinal products where risk factors were identified in (1) mode of action, (2) nature of the target and (3) relevance of animal species and models. with regard to oncology drugs, they are somewhat different from other therapeutic agents in terms of clinical evaluations and estimating the fih dose. for example, the fih clinical studies for traditional cytotoxic agents are usually carried out in patients with treatment-refractory cancer instead of healthy volunteers. in these instances, there is always a hope for a therapeutic benefit and a desire to minimise patient exposure to sub-therapeutic doses. the initial dose selection is usually based on the lethal dose in 10 % of animals (ld10) [6]. there are additional considerations in the design of fih dose for molecular targeted anticancer compounds. in this article, we will first discuss the general considerations on fih dose in the context of clinical trials. then, we will briefly outline the noael and the mabel approaches for the estimations of fih dose. particular attention will be placed on the development of fih dose for oncology drugs. 2. entering the clinics the main goal of clinical trials is to evaluate whether the treatment is effective and safe in humans. before an investigational new drug (ind) can be given to humans, it must be tested thoroughly in animals to find out the following [7]: (1) the effects in the body systems (pharmacodynamics (pd)), (2) the blood levels and how it is absorbed, distributed, metabolised and eliminated after dosing (pharmacokinetics (pk)), (3) a range of doses of the ind, up to a toxic level, the organ toxicity, and safety margin in terms of the noael. after the preclinical studies, there are four phases of trials in humans, which sometimes overlap in practice. phases 1 to 3 are performed before a licence is granted, while phase 4 is launched after authorisation [8]. table 1 summarises the four clinical phases of the development of a typical drug. admet & dmpk 1(4) (2013) 63-75 estimating first in human dose doi: 10.5599/admet.1.4.10 65 table 1. summary of the four clinical phases of development of a typical drug * phase typical number and type of subjects objectives 1 20 – 200 healthy subjects (usually) or patients who are not expected to benefit from the ind determine how well the ind can be tolerated in humans, study the clinical pharmacology and toxicity and find the maximum-tolerated dose to aid the selection of dose and/or schedule for phase 2 2 100-400 patients with the target disease examine dose–response curves in patients, and evaluate the clinical benefits that might be seen in a small group of patients 3 1000-5000 patients with the target disease test in a controlled fashion in a large patient population against a placebo or standard therapy with respect to safety and efficacy 4 many thousands of patients with the target disease post-marketing study to gather additional safety information from a larger group of patients in order to understand the long-term safety and monitor drug-drug interactions * the number of subjects listed in the table were adopted from the abpi guidelines [9]. it should be noted that these values could vary, depending on the therapeutic areas, the trial design and the clinical questions to be addressed. for any clinical trial, the number of subjects has to be planned on a justifiable and rational basis (see reference 10 for further detail). before an ind enters a phase 1 trial, one of the most important tasks is to estimate a starting dose that is low enough to be safe in humans, but not so conservative that excessively costly and time-consuming dose escalations are needed. generally, the estimation is largely based on the pharmacology and toxicity data, while pharmacokinetics is a bridge of pharmacology and toxicity. however, the design of a dose escalation strategy in clinical studies is beyond the scope of this article. in the next section (section 3), we will discuss the approach endorsed by fda for estimating the maximum safe starting dose in initial clinical trials. 3. the noael approach for dose selection there were different approaches that could be used to calculate the starting dose in fih studies; to name a few: noael from preclinical toxicology studies multiplied by a safety factor, the similar drug approach and the pharmacokinetically guided approach [11,12]. estimating the optimal starting dose is complicated, especially when extrapolation of doses from animals to humans is based on multiple assumptions. indeed, different methods may yield widely varying results [11]. there was no consensus regarding the best approach for estimating the starting dose. it was not until 2005, when the fda issued their guidance [3], that the approach based on noael and safety factor became the most common method for estimating the starting dose. in the following discussion, we will focus on the noael approach. readers who are interested in other approaches are directed to the relevant review papers for methodological details [11,12]. tam admet & dmpk 1(4) (2013) 63-75 66 the fda guidance outlines a standardised process for deriving the maximum recommended starting dose (mrsd) for adult healthy volunteers in phase 1 studies [3]. the main steps of this process are: (1) the determination of the noaels in animal species, (2) the conversion of noaels to the human equivalent dose (hed) using the allometric scaling factors listed in table 2, (3) selection of the most appropriate animal species, and (4) the application of a safety factor to derive the mrsd. table 2. conversion factors of animal doses to human equivalent doses [3] species to convert animal dose in mg/kg to hed mg/m 2 , multiply by km + to convert animal dose in mg/kg to hed * in mg/kg, either: divide animal dose by multiply animal dose by human 37 mouse 3 12.3 0.08 hamster 5 7.4 0.13 rat 6 6.2 0.16 ferret 7 5.3 0.19 guinea pig 8 4.6 0.22 rabbit 12 3.1 0.32 dog 20 1.8 0.54 monkeys 12 3.1 0.32 marmoset 6 6.2 0.16 squirrel monkey 7 5.3 0.19 baboon 20 1.8 0.54 * assumes 60 kg human. for species not listed or for weights outside the standard ranges, hed can be calculated from the following formula: hed = animal dose (mg/kg) × (animal wt. (kg)/human wt. (kg)) 0.33 + factor for converting mg/kg dose to mg/m 2 dose step 1: noael determination this is the first step in the process, which involves the evaluation of the available animal data to determine the noaels from toxicology studies. here, the noael refers to the highest dose level that does not produce significant adverse effects compared with the control group. biologically significant or pd effects are acceptable provided that they do not raise a safety concern. the adverse effects used to define the noael should be based on an effect that would be unacceptable if produced by the initial dose in a phase i clinical trial. the noael for each toxicology study is usually reported in mg/kg. step 2: hed calculation the noael for each tested species should be determined, and then converted to the hed using appropriate scaling factors, as shown in table 2. for systemically administered agents, the conversion should be based on the normalisation of doses to body surface area (assumed doses scaled 1:1 when admet & dmpk 1(4) (2013) 63-75 estimating first in human dose doi: 10.5599/admet.1.4.10 67 normalized to body surface area). there are exceptions where the mg/m 2 scaling between species is not expected to work well. these include: therapeutics administered by alternative routes (e.g., topical, intranasal, subcutaneous, intramuscular, intrathecal, intravesical, intraocular, or intrapleural) and proteins administered intravascularly. in such cases, alternative means of normalisation should be utilised [3]. conversion of the animal noael to the hed should be normalised to mg/kg. the species that generates the lowest hed is referred to as the most sensitive species. step 3: selection of the most appropriate species in the absence of any information to aid the selection of species to assess the risk in humans, the most sensitive species is then regarded as the most appropriate, since the use of the lowest hed would generate the most conservative starting dose. however, if there is information suggesting a particular species is more relevant for assessing the risk in humans, the hed of that species could be used. step 4: application of safety factor a safety factor should then be applied to the hed to provide a margin of safety for protection of human subjects receiving the initial clinical dose. the mrsd is obtained by dividing the hed by the safety factor; the default safety factor is 10. this is a historically accepted value, but should be evaluated based on available information. for instance, the safety factor should be increased in situations in which steep dose– response curve, severe/irreversible toxicities, unexplained mortality, variability in doses or drug levels eliciting effects, variable bioavailability, or nonlinear pharmacokinetics were seen in preclinical studies. it can be seen that the mrsd was derived based on the adverse effects seen in preclinical toxicology studies. near the end of the fda guidance document [3], there is brief mention of the merits of comparing the mrsd with pharmacological active dose (pad) derived from appropriate pd models, and adjusting/reducing the fih dose accordingly. for certain classes of drug or biologics, the toxicity may arise from exaggerated pharmacologic effects, so the pad could be a more sensitive indicator of potential toxicity. however, little details were given in the guidance document regarding the selection of pad and the strategy for reducing the clinical starting dose. 4. the drug trial that went horribly wrong in 2006, a phase i clinical trial was conducted at northwick park hospital in london to evaluate a humanised monoclonal antibody – a cd28 superagonist, also known as tgn1412 [4]. tgn1412 was in development for the treatment of b cell chronic lymphocytic leukaemia, with the expectation that it would reverse the t cell deficiency in this disease, by activating the t cell compartment [13]. soon after the administration of tgn1412 to 6 healthy volunteers, the trial turned into a catastrophe because the treatment unexpectedly induced a rapid and massive cytokine storm that caused severe and lifethreatening adverse effects in all six volunteers. six volunteers were hospitalised and suffered from multiple organ dysfunction [4]. the drug was immediately withdrawn from development. extensive studies have been carried out to investigate the cause of this adverse side effect in humans [14-16]. prior to the clinical trial, tgn1412 was tested in preclinical species for toxicology studies and did not show any obvious adverse effects [14]. there is still lack of evidence to explain why tgn1412 did not stimulate the profound cytokine release in vitro or in vivo in preclinical species used in safety testing. the underlying molecular mechanisms leading to the adverse reaction in humans are not yet fully understood. it is plausible that a species difference in the organisation and regulation of t cell responses is responsible for this disparate immunopharmacology [17,18]. nevertheless, the tgn1412 study was the first trial of this tam admet & dmpk 1(4) (2013) 63-75 68 type of compound in human at the time and a small amount of published data were available before the clinical investigation. similar clinical trials of antibodies against t cell antigens had already shown toxic effects [14], suggesting that the tgn1412 trial would be a high-risk study. in the tgn1412 trial, the calculation of the starting dose was based on the noael approach (section 3). a starting dose of 0.1 mg/kg was selected, which corresponded to 160 times lower than the hed derived from the cynomolgus monkey studies (assuming this species was most predictive for human). subsequent analysis by the expert scientific group reported that the trial was conducted according to the medicines and healthcare products regulatory agency (mhra) approved policy and there were no errors involved in the manufacture of tgn1412 or in its formulation, dilution or administration to trial participants [4]. it was then concluded that an unpredicted biological action of the drug in humans was the most likely cause of the adverse reactions. however, it has also been suggested that the tragic incident would have been an avoidable rather than an unforeseeable event if the clinical trial had been designed to emphasise the safety of the trial participants [19]. instead of using the noael approach, the report also highlighted the use of mabel to estimate the fih dose, which would result in a much lower value for tgn1412 – about 20-fold lower than the dose used in the trial. it may be possible that the use of a lower initial dose could avoid the adverse effect seen in the tgn1412 trial, but this is a hypothesis that one would never have the opportunity to testify. nevertheless, after the publication of the expert scientific group report, the emae has issued a guidance document that endorsed the use of the mabel approach for estimating the fih dose, especially for high-risk studies [5]. in the next session, we will briefly discuss the emae guideline that endorses the mabel approach for dose selection. 5. the mabel approach for dose selection the emae guideline identifies factors influencing risk for new investigational medicinal products, which covers the transition from nonclinical to early clinical development [5]. strategies for mitigating and managing risk are given. these include: the calculation of the first dose in humans, the subsequent dose escalation, and the conduct of the clinical trial, all of which are thought to have been triggered by the tragic incident of the tgn1412 trial. the guideline is supposed to use in conjunction with several other nonclinical and clinical eu guidelines. the non-clinical safety aspect is particularly emphasised with reference to the ich m3 (non-clinical safety studies for clinical trials) [7] and the ich m6 (preclinical safety evaluation of biotechnology-derived pharmaceuticals) [20], and four other related guidelines [5]. here, we will direct our attention to the estimation of the first dose in humans. as highlighted above, risk identification and risk mitigation are the focuses of this emae guideline. in particular, factors of risk may be derived from the lack of knowledge in the following areas: (1) the mode of action this refers to the knowledge on the nature and intensity of the effect of the medicinal product on the specific target and non-targets and subsequent mechanisms. certain modes of action have been identified to require special attention, which include targets that have pleiotropic effects or are ubiquitously expressed (e.g., as often occur in the immune system) and those that have a biological cascade or cytokine release, including those leading to an amplification of an effect that might not be sufficiently controlled by a physiological feedback mechanisms (e.g., in the immune system or blood coagulation system). monoclonal antibodies against the t cell targets are named as examples of products that have this latter type of mode of action, presumably because of the tgn1412 trial (see section 4). admet & dmpk 1(4) (2013) 63-75 estimating first in human dose doi: 10.5599/admet.1.4.10 69 (2) the nature of the target this may include information on the structure, tissue distribution, cell specificity, disease specificity, regulation, polymorphisms, level of expression, and biological function of the human target, including downstream effects, and how it might vary between individuals in different populations of healthy subjects and patients. (3) the relevance of animal models. the available animal species should be compared to humans taking into account the target, its structural homology, distribution, signal transduction pathways and the nature of pharmacological effects. where the available animal species/models or surrogates are perceived to be of questionable relevance for thorough investigation of the pharmacological and toxicological effects of the medicinal product, this should be considered as adding to the risk. the emea guideline has pointed out that the risk assessment should be performed on a ‘‘case-by-case’’ basis and that a ‘‘weight-of-evidence’’ approach should be used. despite the scope of the guideline covering both chemical and biological investigational medicinal products, the above factors of risk are mostly derived from the intended pharmacological mechanism (rather than from chemical mechanisms), which are more relevant to the exaggerated “on-target” pharmacology as the mechanism of toxicity for many biologics. in the calculation of the fih dose, it has been suggested that all of the information has to be taken into consideration for the dose selection. in general, the noael approach is still regarded as the method that gives the most important information. however, for investigational medicinal products where the above risk factors have been identified, the use of the mabel approach is recommended. the calculation of mabel should use all in vitro and in vivo information available from pk/pd, which may include the following: (1) target binding and receptor occupancy studies in vitro in target cells from human and the relevant animal species (2) concentration-response curves in vitro in target cells from human and the relevant animal species and dose/exposure-response in vivo in the relevant animal species (3) exposures at pharmacological doses in the relevant animal species to further limit the potential for adverse reactions in humans, a safety factor may be applied in the calculation of the first dose in humans from mabel. this should take into account the criteria of risks such as the novelty of the active substance, its biological potency and its mode of action, the degree of species specificity, the shape of the dose-response curve and the degree of uncertainty in the calculation of the mabel. in cases where the methods of calculation (e.g. noael, mabel) give different estimations of the first dose in man, the lowest value should be used. there are challenges related to the calculation of the mabel. for example, the guideline does not provide much detail on the criteria to define the minimal biological effect, the recommended value for the safety factor (if applicable), or any example of how mabel could be calculated. fortunately, the expert scientific group report has given examples of mabel calculations for tgn1412 and two other products, as well as the calculation of mabel for tgn1412 based on receptor occupancy [4]. following the publication of the expert scientific group report, other papers on the use of the mabel approach to estimate the initial dose for biologics have started to appear in open literature [21-26]. these would undoubtedly serve as examples for clinical investigators to select the fih dose with the mabel approach. tam admet & dmpk 1(4) (2013) 63-75 70 6. initial dose selection for oncology agents we now turn to the initial dose selection for anticancer drugs. in 2013, emae issued an updated guidance document on the evaluation of anticancer medicinal products in humans [27]. the document was adopted from early guidelines dating back to 1996, where the focus was mainly on cytotoxic anticancer compounds. in contrast to the previous version published in 2005, the guideline highlighted the importance of the exploratory clinical (phase 1/2) studies to identify the most appropriate target population for treatment with the selected biomarkers, expanded the coverage on confirmatory trial (phase 3 study), and emphasised the importance of understanding the pharmacokinetics, particularly in the potential of drug-drug interactions, and for patients with impaired organ function. the use of pk/pd analysis supported by biomarkers and clinical markers data in phase 2/3 studies has been particularly encouraged. the emea guideline loosely classified anticancer drugs as cytotoxic and non-cytotoxic [27]. cytotoxic compounds elicit the anticancer effects by means of killing or preventing the division of cells. the drug action is nonspecific in targeting abnormal and normal cells, which is usually linked with toxicity. classical cytotoxic chemotherapeutic agents include: the nitrogen mustard derivatives (e.g. cyclophosphamide, melphalan), antimetabolites (e.g. fluorouracil, methotrexate), platins (e.g. cisplatin, oxaliplatin), antimicrotubule agents (e.g. taxol, vinblastine) and antitumor antibiotics (e.g. doxorubicin, mitomycin c). as for non-cytotoxic anticancer compounds, this generally refers to molecular targeted agents where the drug activity may connect with novel mechanisms such as hormones, nuclear-targeting, and signaltransduction targeting. molecular targeted anticancer compounds have the capacity to modify target or receptor function in an ongoing manner, and are usually less toxic. this kind of behaviour can be exemplified by the hormone receptor-directed agents (e.g. tamoxifen, anastrozole) and many onocogenedirected kinase inhibitors such as gefitinib, erlotinib, imatinib, dasatinib, and sorafenib. immune modulators/monoclonal antibodies (e.g. bevacizumab, cetuximab) and combination therapy studies are regarded as two separated categories of anticancer products in the guideline. owing to the differences in drug actions among the cytotoxics and non-cytotoxics (molecular targeted compounds), and immune modulator/monoclonal antibody and combination therapy studies, the approaches for initial dose selection are varied considerably. in the following discussion, we will discuss the fih dose selection approaches for these four classes of treatments. 6.1. cytotoxics the cytotoxic compounds induce irreversible lethal cellular lesions following short-term exposure, which are relatively nonselective in terms of their effects on tumours against normal cells. for these compounds, toxicity and tumour response can be regarded as suitable indicators of drug activity [27]. because of their inherent toxicity, the clinical trials of cytotoxic agents are not conducted in healthy volunteers. instead, the fih trial of these agents usually involves patients with advanced incurable malignancy, which differs considerably from those in most other therapeutic areas. selection of the starting dose for the initial phase i clinical trial of cytotoxic oncology drug is based on animal toxicology studies [6,27]. these studies are typically performed in rodents (e.g. mice/rats) or other non-rodent large species (e.g. dogs). it has been shown that rodent-only toxicity study (with mice and rats) was sufficiently predictive in identifying the phase i trial starting dose and commonly encountered dose-limited toxicities [28]. this suggests that the routine use of a non-rodent large species in preclinical toxicology studies prior to initial clinical trials with cancer therapeutics is not always necessary. from the mice studies, the dose at which 10% of the mice die (the murine ld10) is defined. one-tenth of this value, 0.1 mld10, expressed in admet & dmpk 1(4) (2013) 63-75 estimating first in human dose doi: 10.5599/admet.1.4.10 71 mg/m 2 , has been historically regarded as a safe starting dose in humans provided that the toxicology studies in a second species (e.g. dog) did not show serious, irreversible toxicity and substantial differences in the dose–toxicity relationship [29]. if irreversible toxicities are produced at the proposed starting dose in non-rodents or if the non-rodent is known to be a more appropriate model, then the starting dose could be based on one-third of the lowest dose in non-rodents that produces drug-induced pathological alterations in haematological, clinical, or morphological parameters and which, when doubled, produces no lethality [11]. readers are directed to a literature working example of designing the safe starting dose for a cytotoxic agent for further methodological details [30]. however, ethical concerns exist on using excessive numbers of patients at sub-therapeutic doses of a new cytotoxic agent to reach the maximum-tolerated dose (mtd) during dose escalation [29]. increasing the starting dose and/or pursuing a more aggressive dose escalation schedule could potentially reduce the number of patients treated at sub-therapeutic doses. on the other hand, it is important to protect patients from unacceptable risk in terms of toxicity. there is a fine balance between the starting dose selection, escalation strategy and patient safety. 6.2. non-cytotoxics (molecular targeted compounds) unlike cytotoxic drugs, the pharmacological effects of molecular targeted compounds rarely induce acute cellular damage. most often, these drugs are cytostatic, and are likely to be more effective when administered continuously. usually, molecular targeted compounds will interact with proteins that are specific to tumour cells or that are up-regulated during malignant transformation. thus, they tend to be more selective and less toxic to normal tissue. increasing the dose in a traditional phase 1 trial to the mtd may be irrelevant for molecular targeted compounds as the maximum therapeutic effect may be achieved at doses that are well below the mtd. in some situations, it may be difficult to determine the mtd [31]. as such, the early stages of clinical drug development are likely to be more complex and have to be tailored according to the assumed pharmacology of the compound as defined in preclinical studies. an early assessment of anti-tumour activity using biomarkers and pd (e.g. receptor binding, enzyme inhibition, downstream events, or sensitive imaging techniques) may be required to help to define the dose and schedule. for molecular targeted compounds, there does not appear to be a standardised approach for selecting the fih dose. this is not even mentioned in the emea guideline [27]. indeed, le tourneau et al. reported that there is no consensus on which preclinical models and parameters should define the starting dose for molecular targeted compounds [31]. the authors have analysed 81 fih phase i trials that evaluated 60 different molecular targeted agents; the data were taken from open literature from 1998 to 2009. in the 81 trials, 57 of which have clearly specified the animal models used to derive the starting dose. figure 1 shows the preclinical models and toxicological parameters used to derive the starting dose of these 57 phase i clinical trials of molecular targeted compounds. it can be seen that the selections of fih dose were based on diverse practices using a variety of preclinical toxicological parameters. nevertheless, most of these trials appeared to be safely conducted, which is consistent with the fact that toxicity may not be a critical issue in molecular targeted compounds, as compared with cytotoxic anticancer compounds. tam admet & dmpk 1(4) (2013) 63-75 72 figure 1. the preclinical models and toxicological parameters used to derive the starting dose of the 57 phase 1 cancer clinical trials of molecular targeted compounds (adapted from reference 32). mtd, maximum-tolerated dose; ld10, lethal dose for 10% of animals; tdl, toxic dose low; noael, no observed adverse event level; std10, severely toxic dose for 10% of animals; med, minimum effective dose. 6.3. immune modulators and monoclonal antibodies for these kinds of compounds, it has been suggested that the selection of starting dose may be supported by the mabel approach, as discussed in section 5 [27]. 6.4. combination therapy studies cytotoxic compounds have been used in combination to increase the anti-tumour activity at acceptable levels of toxicity. this may be accomplished by combining compounds with at least partly non-overlapping toxicity, activity or resistance. however, it may be difficult to predict whether this kind of combination therapy would lead to a better treatment outcome than the individual agents that are consecutively used in isolation. in combining two conventional cytotoxic compounds, it may be possible to predict the toxicity based on the toxicities of individual components. assuming that the pk interactions of the two compounds admet & dmpk 1(4) (2013) 63-75 estimating first in human dose doi: 10.5599/admet.1.4.10 73 are negligible, the starting dose could be about one half of the recommended mono-therapy dose for each compound [27]. if the hypothesised interaction would lead to an anticipated increased drug exposure, the first dose-level should be cautiously low. alternatively, it may be reasonable to start at the recommended dose for one compound and a reduced dose (<50%) of the other. in cases where at least of one the two components is a non-cytotoxic compound, it is desirable to have compelling preclinical rationale to support the combination study [33]. this kind of combination phase 1 study is very complex. the methodology for the development of innovative cancer therapies (mdict) task force has recently published a paper concerning recommendations on combination therapy studies [34]. with regard to the schedule and dose, the paper simply remarked that they “should be formally explored for e.g. with a randomised or adaptive design”. as far as we are aware, little, if any, information is available in open literature describing the estimation of starting dose and clinical trial design in this kind of study. given the high uncertainty to predict the add-on activity in non-clinical models, the emea has recommended the incorporation of randomised phase 2 studies to compare the experimental regimen with the chemotherapy-alone regimen [27]. 7. conclusions estimation of the fih dose is one of the important steps in phase 1 clinical trials. the initial human dose has to be low enough to ensure the safety of the human participants, but not so conservative that excessively costly and time-consuming dose escalations are needed. the noeal approach endorsed by the fda [3] is a commonly used method to perform the initial dose prediction. this is based on the data obtained from preclinical toxicology studies. following the tragic incident of the tgn1412 phase 1 trial, it has been recognised that the noeal approach might not be sufficient to provide a safe initial dose estimate, especially for high-risk investigative medicinal products that the mechanisms of the pharmacological actions in human are not fully understood. since then, the emea has endorsed the mabel approaches [5] for estimating the fih dose of high-risk investigative medicinal products. this comprehensive approach made use of all preclinical in vitro and in vivo information available from pk/pd, which appears to provide a more conservative estimate of the initial dose on many occasions. this is particularly useful for biologics and/or high-risk investigative medicinal products where the toxicity is closely related to the exaggerated “on-target” pharmacology. with regard to oncology drugs, they are somewhat different from other therapeutic agents in terms of clinical evaluations and estimating the fih dose. the recently published emea guideline [27] provided upto-date recommendations in the clinical study of oncology products, which include: cytotoxics, noncytotoxics, immune modulator/monoclonal antibody and combination therapy studies. it is noted that the fih dose estimation for cytotoxics is usually based on a fraction of the lethal dose obtained from preclinical toxicology studies. for immune modulator/monoclonal antibodies, the mabel approach is recommended. as for non-cytotoxics, the selection of fih dose was based on diverse practices. as far as we are aware, there is currently no established protocol or consensus regarding the fih dose selection in combination therapy studies involving molecular targeted compound(s). this reflects on the complex nature of the clinical trial design for this kind of combination study, which would have to consider a “case-by-case” basis. references [1] s. derry, y.k. loke, j.k aronson, bmc medical research methodology 1 (2001) 7. [2] d.l. streiner, canadian journal of neurological sciences 34: s1 (2007) s37-s41. tam admet & dmpk 1(4) (2013) 63-75 74 [3] fda guidance for industry estimating the maximum safe dose in initial clinical trials for therapeutics in adult healthy volunteers 2005, http://www.fda.gov/downloads/drugs/guidances/ucm078932.pdf (accessed 19 june, 2013) [4] expert scientific group on phase one clinical trials final report 2006, http://www.tfscro.com/business/doc/final_report_of_the_expert_scientific_group_(esg).pdf (accessed 19 june, 2013) [5] emea guideline on strategies to identify and mitigate risks for first-in-human clinical trials with investigational medicinal products 2007, http://www.emea.europa.eu/docs/en_gb/document_library/scientific_guideline/2009/09/wc500 002988.pdf (accessed 19 june, 2013) [6] j.j. degeorge, c.h. ahn, p.a. andrews, m.e. brower, d.w. giorgio, m.a. goheer, d.y. lee-ham, w.d. mcguinn, w. schmidt, c.j. sun, s.c. tripathi, cancer chemotherapy and pharmacology 41 (1998) 173-185. [7] ich guidelines m3 (r2) on non-clinical safety studies for clinical trials 2009, http://www.emea.europa.eu/docs/en_gb/document_library/scientific_guideline/2009/09/wc500 002720.pdf (accessed 20 june, 2013) [8] d. wang, a. bakhai, clinical trialsa practical guide to design, analysis and reporting, remedica, london, uk, 2006. [9] guidelines for phase i clinical trials 2007, http://www.abpi.org.uk/ourwork/library/guidelines/documents/phase1-trial-guidelines.pdf (asccessed 12 october, 2013) [10] b. röhrig, j.b. du prel, d. wachtlin, r. kwiecien, m. blettner, deutsches ärzteblatt international 107 (2010) 552-556. [11] b.g. reigner, k.s. blesch, european journal of clinical pharmacology 57 (2002) 835–845. [12] i. mahmood, m.d. green, j.e. fisher, journal of clinical pharmacology 43 (2003) 692-697. [13] t. hünig, advances in immunology 95 (2007) 111–148. [14] m.j.h. kenter, a.f. cohen, lancet 368 (2006) 1387–1391. [15] z. waibler, l.y. sender, c. merten, r. hartig, s. kliche, m. gunzer, p. reichardt, u. kalinke, b. schraven, plos one 3(3) (2008) e1708 [16] d. eastwood, l. findlay, s. poole, c. bird, m. wadhwa, m. moore, c. burns, r. thorpe, r. stebbings, british journal of pharmacology 161 (2010) 512–526. [17] b. schraven, u. kalinke, immunity 28 (2008) 591-595. [18] n. müller, j. van den brandt, f. odoardi, d. tischner, j. herath, a. flügel, h.m. reichardt, journal of clinical investigation 118 (2008) 1405-1416. [19] c.j. horvath, m.n. milton, toxicologic pathology 37 (2009) 372-383. [20] ich guidelines s6 (r1) preclinical safety evaluation of biotechnology-derived pharmaceuticals 2011, http://www.emea.europa.eu/docs/en_gb/document_library/scientific_guideline/2009/09/wc500 002828.pdf (accessed 24 june, 2013). [21] p.j. lowe, y. hijazi y, o. luttringer, h. yin, r. sarangapani, d. howard, xenobiotica 37 (2007) 13311354. [22] b.m. agoram, british journal of clinical pharmacology 67 (2008) 153–160. [23] p.y. muller, m. milton, p. lloyd, j. sims, f.r. brennan, current opinion in biotechnology 20 (2009) 722-729. [24] p.j. lowe, s. tannenbaum, k. wu, p. lloyd, j. sims, basic & clinical pharmacology & toxicology 106 (2009) 195–209. [25] j.p. gibbs, the aaps journal 12 (2010) 750-758. [26] j. yu, h. karcher, a.l. feire, p.j. lowe, the aaps journal 13 (2011) 160-178. http://www.fda.gov/downloads/drugs/guidances/ucm078932.pdf http://www.tfscro.com/business/doc/final_report_of_the_expert_scientific_group_(esg).pdf http://www.emea.europa.eu/docs/en_gb/document_library/scientific_guideline/2009/09/wc500002988.pdf http://www.emea.europa.eu/docs/en_gb/document_library/scientific_guideline/2009/09/wc500002988.pdf http://www.emea.europa.eu/docs/en_gb/document_library/scientific_guideline/2009/09/wc500002720.pdf http://www.emea.europa.eu/docs/en_gb/document_library/scientific_guideline/2009/09/wc500002720.pdf http://www.abpi.org.uk/our-work/library/guidelines/documents/phase1-trial-guidelines.pdf http://www.abpi.org.uk/our-work/library/guidelines/documents/phase1-trial-guidelines.pdf http://www.emea.europa.eu/docs/en_gb/document_library/scientific_guideline/2009/09/wc500002828.pdf http://www.emea.europa.eu/docs/en_gb/document_library/scientific_guideline/2009/09/wc500002828.pdf admet & dmpk 1(4) (2013) 63-75 estimating first in human dose doi: 10.5599/admet.1.4.10 75 [27] emea guideline on the evaluation of anticancer medicinal products in man 2012, http://www.ema.europa.eu/docs/en_gb/document_library/scientific_guideline/2013/01/wc5001 37128.pdf (accessed 26 june 2013). [28] d.r newell, s.s. burtles, b.w. fox, d.i. jodrell, t.a. connors, british journal of cancer 81 (1999) 760–768. [29] e.a. eisenhauer, p.j. o’dwyer, m. christian, j.s. humphrey, journal of clinical oncology 18 (2000) 684-692. [30] a.m. senderowicz, clinical cancer research 16 (2010) 1719-1725. [31] n. saigo, t. tamura, k. nishio, cancer chemotherapy and pharmacology 46 (2000) s43-s45. [32] c. le tourneau, a. stathis, l. vidal, m.j. moore, l.l. siu, journal of clinical oncology 28 (2010) 14011407. [33] j. verweij, m.l. disis, s.a. cannistra, journal of clinical oncology 28 (2010) 4545-4546. [34] l.k. seymour, a.h. calvert, m.w. lobbezoo, e.a. eisenhauer, g. giaccone, european journal of cancer 49 (2013) 1808-1814. ©2013 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://www.ema.europa.eu/docs/en_gb/document_library/scientific_guideline/2013/01/wc500137128.pdf http://www.ema.europa.eu/docs/en_gb/document_library/scientific_guideline/2013/01/wc500137128.pdf http://creativecommons.org/licenses/by/3.0/ an electrochemical sensing platform based on a modified carbon paste electrode with graphene/co3o4 nanocomposite for sensitive propranolol determination doi: http://dx.doi.org/10.5599/admet.1705 1 admet & dmpk 11(2) (2023) 227-236; doi: https://doi.org/10.5599/admet.1705 open access : issn : 1848-7718 http://pub.iapchem.org/ojs/index.php/admet/index original scientific paper an electrochemical sensing platform based on a modified carbon paste electrode with graphene/co3o4 nanocomposite for sensitive propranolol determination parisa karami-kolmoti1 and reza zaimbashi2* 1department of chemistry, graduate university of advanced technology, kerman, iran 2 environment department, institute of science and high technology and environmental sciences, graduate university of advanced technology, kerman, iran *corresponding author: e-mail: r.zaim2017@gmail.com received: february 10, 2023; revised: february 23, 2023; published: march 15, 2023 abstract a simple and sensitive method for the determination of propranolol using a modified carbon paste electrode with graphene/co3o4 nanocomposite was presented. the electrochemical measurements of propranolol are studied using differential pulse voltammetry, cyclic voltammetry and chronoamperometry. the graphene/co3o4 nanocomposite exhibits excellent catalytic activity towards the electrochemical oxidation of propranolol in phosphate buffer solution of ph 7.0. the graphene/co3o4 nanocomposite facilitates the determination of propranolol in the concentration range 1.0–300.0 μm and a detection limit and sensitivity of 0.3 μm. and 0.1275 μa/μm were achieved. ©2023 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords carbon paste electrodes; graphene/co3o4 nanocomposite; differential pulse voltammetry; propranolol introduction propranolol (i) [1-(isopropylamino)-3-(1-napthyloxy)-2propanol] is a highly effective and non-selective adrenergic receptor (-blocker) that is widely used in clinical practices to treat cardiovascular disorders, such as angina pectoris, arrhythmia, myocardial infarction, dysfunctional labor and anxiety. it possesses no autonomous nervous system activity and is completely absorbed from the gastrointestinal tract in healthy volunteers after oral administration. it affects the regulation of blood circulation by reducing the cardiac frequency, myocardial contractility, contraction force, coronary flow and secretion of rennin, resulting in a fall in the level of angiotensin ii, which contributes significantly to the hypertensive action of this drug. it also lowers blood pressure by altering the transmission of nerve impulses from the brain to a certain part of the body and causes relaxation (and possibly dilation) of blood vessels, decreasing heart rate and pulse. propranolol has been listed as a banned substance by world anti-doping agency (wada) and international olympic committee in competitive games. therefore, monitoring the level of propranolol in biological fluids is important for clinical practices and doping control. to ensure the reliability of results obtained from http://dx.doi.org/10.5599/admet.1705 https://doi.org/10.5599/admet.1705 http://pub.iapchem.org/ojs/index.php/admet/index mailto:r.zaim2017@gmail.com http://creativecommons.org/licenses/by/4.0/ p. karami-kolmoti and r. zaimbashi admet & dmpk 11(2) (2023) 227-236 228 analytical methods, wada has established the minimum required performance limit of this drug in urine as 500 g/l [1-5]. there are various methods, such as rp-hplc-dad, chromatography, 13c nmr spectroscopic and electrochemical method [5-9]. among all these quantification techniques, electrochemical techniques offered more sensitivity and reliability in sample consumption and miniaturization of the processes [10-17]. nanotechnology has enormous potential for providing innovative solutions to a wide range of applications [18-23]. the development of nanoscience and nanotechnology has allowed trials to apply different nanomaterials for the fabrication of chemically modified electrodes [24-36]. in recent years, various nanomaterials have been used singly or in composite form to modify electrodes [37-47]. the chemical modification of inert substrate electrodes offers significant advantages in the design and development of electrochemical sensors [48-52]. in operations, the redox-active sites often shuttle electrons between a solution of the analyte and the substrate electrodes, with a significant reduction of the activation overpotential [53-55]. a further advantage of chemically modified electrodes is that they are less prone to surface fouling and oxide formation compared to inert substrate electrodes [56-60]. the carbon paste electrode (cpe) has attracted great attention in electrochemistry due to its non-toxic nature, eco-friendly, low-cost, easy preparation, broad operational potential window, easy chemical and mechanical modification, and renewable surface. carbon paste is a widely used electrode due to its electrochemical characteristics, such as very low background current, low ohmic resistance, affordability, easy modification and simple renewal of the electrode surface [61-63]. the objective of the present research is the fabrication of a new sensor by modification of a carbon paste electrode with graphene/co3o4 nanocomposite (gr-co3o4 nc/cpe) for determination of the propranolol. finally, this technique experienced a successful application for detecting propranolol in the real sample with encouraging outputs. experimental equipment and materials in order to do electrochemical tests at ambient temperature, we utilized the auto-lab potentiostat/galvanostat (pgstat 302n, eco chemie, the netherlands) with gpes (general purpose electrochemical systemversion 4.9) software to control the system. electrochemical measurements were performed at room temperature in a conventional electrochemical cell with a gr-co3o4 nc/cpe as the working electrode, 3.0 m ag/ agcl/kcl as a reference electrode (azar electrode, urmia, iran) and platinum wire as a counter electrode (azar electrode, urmia, iran). moreover, ph was measured using the metrohm 713 ph meter with a glass electrode (switzerland). propranolol and all other solutions used during the procedure were prepared by reagent-grade chemicals from merck and sigma-aldrich and deionized water was supplied from millipore, germany. preparation of graphene/co3o4 nanocomposite firstly, graphene oxide (go) (20 mg) was dispersed in 20 ml ethanol and ultrasonicated for 30 min. then, co(no3)2.6h2o (0.001 mol) was dissolved in 20 ml ethanol solution and stirred for 30 min at ambient temperature. subsequently, the prepared two solutions were mixed under stirring and 3.6 ml of ammonia solution (nh3.h2o (wt. 25%)) was dropwise added. the mixture was transferred into a teflon-lined stainless steel autoclave and maintained at 180 °c for 24 h. after completion of the reaction, the product was collected by centrifugation, and washed with deionized water and ethanol. finally, the graphene/co3o4 nanocomposite was dried at 70 °c overnight in an oven. figure 1 shows the fe-sem image of graphene/co3o4 nanocomposite. admet & dmpk 11(2) (2023) 227-236 propranolol determination on graphene/co3o4 nanocomposite doi: https://doi.org/10.5599/admet.1705 229 figure 1. fe-sem image of graphene/co3o4 nanocomposite. preparation and surface modification of electrode to prepare gr-co3o4 nc/cpe, 0.95 g graphite powder and 0.05 g graphene/co3o4 nanocomposite were mixed. next, a suitable amount of paraffin oil was poured into the resulting mixture, followed by mixing well for 30 min to obtain a uniformly wetted paste. an appropriate amount of the paste was tightly packed into a glass tube and a copper wire was positioned over the carbon paste to make electrical contact (scheme 1). the surface areas of the gr-co3o4 nc/cpe and the un-modified cpe were obtained by cv using 1 mm k3fe(cn)6 at various scan rates. using the randles–ševčik equation for gr-co3o4 nc/cpe, the electrode surface was 0.342 cm2, about 3.8 times greater than un-modified cpe. scheme 1. scheme of gr-co3o4 nc/cpe preparation and voltammetric detection of propranolol. results and discussion electrochemical behavior of propranolol on graphene/co3o4 nanocomposite the electrochemical behavior of the cpe, gr-co3o4 nc/cpe was studied by the cyclic voltammetry (cv) technique in the 0.1 m phosphate buffer (ph 7.0) as the supporting electrolyte at a scan rate of 50 mv s−1 (figure 2). as shown in figure 2, in comparison to the bare cpe (a), gr-co3o4 nc/cpe (b) presents a welldefined irreversible oxide peak with a higher current signal (propranolol concentration equal to 100.0 µm). https://doi.org/10.5599/admet.1705 p. karami-kolmoti and r. zaimbashi admet & dmpk 11(2) (2023) 227-236 230 figure 2. cyclic voltammograms of a) cpe and b) gr-co3o4 nc/cpe in the presence of 100.0 μm propranolol at a ph 7.0 of 0.1 m pbs, respectively. role of variable scan rates the effect of the potential scan rates (5-300 mv s-1) on the electrochemical oxidation of propranolol was studied by cyclic voltammetry (cv). figure 3 shows the cv of 70.0 µm of propranolol in the 0.1 m phosphate buffer solution at the gr-co3o4 nc/cpe. these results show that the anodic current increases with an increasing scan rate. the oxidation current of propranolol increased linearly with the square root of the scan rate (figure 3, inset), which demonstrate a diffusion controlled electrochemical process. figure 3. cyclic voltammograms of propranolol (70.0 μm) at gr-co3o4 nc/cpe at different scan rates of a) 5, b) 10, c) 20, d)30, e) 40, f) 50, g) 60, h) 70, i) 80, j) 90 k) 100 l) 200 and m) 300 mv/s in 0.1 m pbs (ph 7.0). insert: plot of ip versus ν 1/2 for the oxidation of propranolol at gr-co3o4 nc/cpe. chronoamperometric analysis the chronoamperometric measurements of propranolol at the gr-co3o4 nc/cpe surface were done to estimate the apparent diffusion coefficient of propranolol under used experimental conditions. figure 4 admet & dmpk 11(2) (2023) 227-236 propranolol determination on graphene/co3o4 nanocomposite doi: https://doi.org/10.5599/admet.1705 231 shows the current-time profiles obtained by setting the working electrode potential at 1000 mv for different concentrations of propranolol. at long enough experimental times (t=0.3-3s), where the electron transfer reaction rate of propranolol is more than its diffusion rate toward the working electrode surface, the current is diffusion controlled. figure 4, inset a, shows the experimental plots of i versus t-1/2 with the best fit for different concentrations of propranolol employed. the slopes of the resulting straight lines were then plotted versus the propranolol concentration (figure 4, inset b). based on the cottrell equation [64], the slope of this plot (figure 4 inset b) can be used to estimate the apparent diffusion coefficient of propranolol. from the slope of this plot (29.458 a s1/2 mm-1), the value of propranolol was found to be 7.4x10-5 cm s1 for propranolol. figure 4. chronoamperograms obtained at the gr-co3o4 nc/cpe in the presence of a) 0.05, b) 0.5, c) 1.5 and d) 2.0 mm propranolol in the 0.1 m buffer solution (ph 7.0). a) plot of i versus t-1/2 for electro-oxidation of hydrochlorothiazide obtained from chronoamperograms a–d. b) plot of slope from straight lines versus propranolol level. differential pulse voltammetry analysis of propranolol differential pulse voltammetry (dpv) was used for the determination of propranolol at gr-co3o4 nc/cpe due to its high sensitivity. the dpv responses for different concentrations of propranolol were illustrated in figure 5 (step potential=0.01 v and pulse amplitude=0.025 v). the linear range was found to be from 1.0 μm to 300.0 μm. the linear equation was ip (μa)=1.6252+0.1275 cpropranolol (μm) with a correlation coefficient of 0.9979. also, the limit of detection, cm, of propranolol was calculated using the following equation: cm=3sb/m where, m is the slope of the calibration plot (0.1275 μa/ μm) and sb is the standard deviation of the blank response obtained from 10 replicate measurements of the blank solution. the detection limit for the determination of propranolol using this method is 0.3 μm. https://doi.org/10.5599/admet.1705 p. karami-kolmoti and r. zaimbashi admet & dmpk 11(2) (2023) 227-236 232 figure 5. dpv curves of gr-co3o4 nc/cpe in the0.1 m buffer solution (ph 7.0) containing different concentrations of propranolol. a-n corresponds to 1.0, 5.0, 10.0, 20.0, 30.0, 40.0, 50.0, 60.0, 70.0, 80.0, 90.0, 100.0, 200.0 and 300.0 μm propranolol. inset: plots of electrocatalytic peak current as a function of propranolol concentration. conclusion we arranged a sensor using gr-co3o4 nanocomposite produced for the detection of propranolol and showed that the results showed the improved behavior of the propranolol sensor with respect to the nanocomposite by s-created. extensive direct access, limited detection limit (0.3 μm), selectivity and incredible robustness were found for the proposed detection framework as a result of improved properties such as large surface area, fast mass vehicle, remarkable biocompatibility and reactive synergism of gr-co3o4 nanoparticles. likewise, the findings recommended gr-co3o4 nc/cpe nanoparticles as a magnificent sensor for the determination of propranolol. conflict of interest: the authors declare no conflict of interest references [1] p. gupta, s.k. yadav, b. agrawal, r.n. goyal. a novel graphene and conductive polymer modified pyrolytic graphite sensor for determination of propranolol in biological fluids. sensors and actuators b: chemical 204 (2014) 791-798. https://doi.org/10.1016/j.snb.2014.08.040 [2] a. santhy, s. beena. a study on the recent developments in voltammetric sensors for the β-blocker propranolol hydrochloride. biomedical engineering applications for people with disabilities and the elderly in the covid-19 pandemic and beyond, (2022) 23-31. https://doi.org/10.1016/b978-0-32385174-9.00004-2 [3] h. b. ladmakhi, s. fathi, f. chekin, j.b. raoof. determination of propranolol at a carbon paste electrode modified with magnetite–graphene oxide in combination with presence of sodium dodecyl sulfate. russian journal of electrochemistry 58(3) (2022) 184-191. https://doi.org/10.1134/s1023193522030065 [4] m. yaqub, s. park, e. alzahrani, a.e. farouk, w. lee. application of data-driven machine learning to predict propranolol and trimethoprim removal using a managed aquifer recharge system. journal of environmental chemical engineering 10(1) (2022) 106847. https://doi.org/10.1016/j.jece.2021.106847 [5] m. khalid, s. ahmad. box–behnken design used to optimize the simultaneous quantification of amitriptyline and propranolol in tablet dosages by rp-hplc-dad method and their stability tests. separations 9(12) (2022) 421. https://doi.org/10.3390/separations9120421 https://doi.org/10.1016/j.snb.2014.08.040 https://doi.org/10.1016/b978-0-323-85174-9.00004-2 https://doi.org/10.1016/b978-0-323-85174-9.00004-2 https://doi.org/10.1134/s1023193522030065 https://doi.org/10.1016/j.jece.2021.106847 https://doi.org/10.3390/separations9120421 admet & dmpk 11(2) (2023) 227-236 propranolol determination on graphene/co3o4 nanocomposite doi: https://doi.org/10.5599/admet.1705 233 [6] n.s. alsultani, a.a. alkarimi, i.a. aljazaery. preparation of monolithic chromatographic column for propranolol hydrochloride determination. journal of pharmaceutical negative results 13(4) (2022) 1629-1636. https://doi.org/10.47750/pnr.2022.13.04.226 [7] s. lee, s. jung. 13c nmr spectroscopic analysis on the chiral discrimination of n-acetylphenylalanine, catechin and propranolol induced by cyclic-(1→ 2)-β-d-glucans (cyclosophoraoses). carbohydrate research 337(19) (2002) 1785-1789. https://doi.org/10.1016/s0008-6215(02)00286-0 [8] a.m. santos, a. wong, o. fatibello-filho. simultaneous determination of salbutamol and propranolol in biological fluid samples using an electrochemical sensor based on functionalized-graphene, ionic liquid and silver nanoparticles. journal of electroanalytical chemistry 824 (2018) 1-8. https://doi.org/ 10.1016/j.jelechem.2018.07.018 [9] m. khairy, a.a. khorshed. simultaneous voltammetric determination of two binary mixtures containing propranolol in pharmaceutical tablets and urine samples. microchemical journal 159 (2020) 105484. https://doi.org/10.1016/j.microc.2020.105484 [10] h. karimi-maleh, f. karimi, y. orooji, g. mansouri, a. razmjou, a. aygun, f. sen. a new nickel-based co-crystal complex electrocatalyst amplified by nio dope pt nanostructure hybrid; a highly sensitive approach for determination of cysteamine in the presence of serotonin. scientific reports 10(1) (2020) 1-13. https://doi.org/10.1038/s41598-020-68663-2 [11] a. lohrasbi-nejad. electrochemical strategies for detection of diazinon. journal of electrochemical science and engineering 12(6) (2022) 1041-1059. https://doi.org/10.5599/jese.1379 [12] h. beitollahi, s.z. mohammadi, m. safaei, s. tajik. applications of electrochemical sensors and biosensors based on modified screen-printed electrodes: a review. analytical methods 12 (2020) 15471560. https://doi.org/10.1039/c9ay02598g [13] j. mohanraj, d. durgalakshmi, r.a. rakkesh, s. balakumar, s. rajendran, h. karimi-maleh. facile synthesis of paper based graphene electrodes for point of care devices: a double stranded dna (dsdna) biosensor. journal of colloid and interface science 566 (2020) 463-472. https://doi.org/10.1016/j.jcis.2020.01.089 [14] y.f. mustafa, g. chehardoli, s. habibzadeh, z. arzehgar. electrochemical detection of sulfite in food samples. journal of electrochemical science and engineering 12(6) (2022) 1061-1079. https://doi.org/10.5599/jese.1555 [15] m. mazloum-ardakani, h. beitollahi, b. ganjipour, h. naeimi. novel carbon nanotube paste electrode for simultaneous determination of norepinephrine, uric acid and d-penicillamine. international journal of electrochemical science 5 (2010) 531-546. [16] m. miraki, h. karimi-maleh, m.a. taher, s. cheraghi, f. karimi, s. agarwal, v.k. gupta. voltammetric amplified platform based on ionic liquid/nio nanocomposite for determination of benserazide and levodopa. journal of molecular liquids 278 (2019) 672-676. https://doi.org/10.1016/j.molliq.2019.01.081 [17] m. velicky, a. n. rodgers, r.a. dryfe, k. tam. use of voltammetry for in vitro equilibrium and transport studies of ionisable drugs. admet and dmpk 2(3) (2014) 143-156. https://doi.org/10.5599/admet.2.3.22 [18] s. azimi, m. amiri, h. imanzadeh, a. bezaatpour. fe3o4@sio2-nh2/cosb modified carbon paste electrode for simultaneous detection of acetaminophen and chlorpheniramine. advanced journal of chemistry-section a 4(2) (2021) 152-164. https://doi.org/10.22034/ajca.2021.275901.1246 [19] s. mahari, s. gandhi. electrochemical immunosensor for detection of avian salmonellosis based on electroactive reduced graphene oxide (rgo) modified electrode. bioelectrochemistry 144 (2022) 108036. https://doi.org/10.1016/j.bioelechem.2021.108036 [20] h. karimi-maleh, c. karaman, o. karaman, f. karimi, y. vasseghian, l. fu, m. baghayeri, j. rouhi, p. senthil kumar, p.l. show, s. rajendran. nanochemistry approach for the fabrication of fe and n co-decorated biomass-derived activated carbon frameworks: a promising oxygen reduction reaction electrocatalyst in neutral media. journal of nanostructure in chemistry 12(3) (2022) 429-439. https://doi.org/10.1007/ s40097-022-00492-3 https://doi.org/10.5599/admet.1705 https://doi.org/10.47750/pnr.2022.13.04.226 https://doi.org/10.1016/s0008-6215(02)00286-0 https://doi.org/10.1016/j.jelechem.2018.07.018 https://doi.org/10.1016/j.jelechem.2018.07.018 https://doi.org/10.1016/j.microc.2020.105484 https://doi.org/10.1038/s41598-020-68663-2 https://doi.org/10.5599/jese.1379 https://doi.org/10.1039/c9ay02598g https://doi.org/10.1016/j.jcis.2020.01.089 https://doi.org/10.1016/j.jcis.2020.01.089 https://doi.org/10.5599/jese.1555 https://doi.org/10.1016/j.molliq.2019.01.081 https://doi.org/10.5599/admet.2.3.22 https://doi.org/10.22034/ajca.2021.275901.1246 https://doi.org/10.1016/j.bioelechem.2021.108036 https://doi.org/10.1007/s40097-022-00492-3 https://doi.org/10.1007/s40097-022-00492-3 p. karami-kolmoti and r. zaimbashi admet & dmpk 11(2) (2023) 227-236 234 [21] s. tajik, m.a. taher, h. beitollahi. simultaneous determination of droxidopa and carbidopa using a carbon nanotubes paste electrode. sensors and actuators b: chemical 188 (2013) 923-930. https://doi.org/10.1016/j.snb.2013.07.085 [22] a. hosseini fakhrabad; r. sanavi khoshnood; m.r. abedi; m. ebrahimi. fabrication a composite carbon paste electrodes (cpes) modified with multi-wall carbon nano-tubes (mwcnts/n, n-bis (salicyliden)1,3-propandiamine) for determination of lanthanum (iii). eurasian chemical communications 3(9) (2021) 627-634. http://dx.doi.org/10.22034/ecc.2021.288271.1182 [23] m. yang, z. sun, h. jin, r. gui. sulfur nanoparticle-encapsulated mof and boron nanosheet-ferrocene complex modified electrode platform for ratiometric electrochemical sensing of adriamycin and realtime monitoring of drug release. microchemical journal 177 (2022) 107319. https://doi.org/ 10.1016/j.microc.2022.107319 [24] m. farahmandjou, p. khalili. zno nanoparticles synthesized by co-precipitation method; morphology and optoelectronic study. asian journal of green chemistry 5(2) (2021) 219-226. https://doi.org/10.22034/ajgc.2021.261206.1287 [25] h. beitollahi, s.g. ivari, m. torkzadeh-mahani. voltammetric determination of 6-thioguanine and folic acid using a carbon paste electrode modified with zno-cuo nanoplates and modifier. materials science and engineering: c 69 (2016) 128-133. https://doi.org/10.1016/j.msec.2016.06.064 [26] p.s. kumar, b.s. sreeja, k.k. kumar, g. padmalaya. static and dynamic analysis of sulfamethoxazole using go/zno modified glassy carbon electrode by differential pulse voltammetry and amperometry techniques. chemosphere 302 (2022) 134926. https://doi.org/10.1016/j.chemosphere.2022.134926 [27] s.a. ismaeel; y.k. al-bayati. determination of trace metformin in pharmaceutical preparation using molecularly imprinted polymer based pvc-membrane. eurasian chemical communications 3(11) (2021) 812-830. http://dx.doi.org/10.22034/ecc.2021.300477.1224 [28] y. dessie, s. tadesse. a review on advancements of nanocomposites as efficient anode modifier catalyst for microbial fuel cell performance improvement, journal of chemical reviews 3(4) (2021) 320-344. http://dx.doi.org/10.22034/jcr.2021.314327.1128 [29] e.c. martins, e.r. santana, a. spinelli. nitrogen and sulfur co-doped graphene quantum dot-modified electrode for monitoring of multivitamins in energy drinks. talanta 252 (2023) 123836. https://doi.org/10.1016/j.talanta.2022.123836 [30] m. mazloum-ardakani, h. beitollahi, m.k. amini, f. mirkhalaf, b.f. mirjalili, a. akbari. application of 2(3, 4-dihydroxyphenyl)-1, 3-dithialone self-assembled monolayer on gold electrode as a nanosensor for electrocatalytic determination of dopamine and uric acid. analyst 136(9) (2011) 1965-1970. [31] f. zare kazemabadi, a. heydarinasab, a. akbarzadehkhiyavi, m. ardjmand. development, optimization and in vitro evaluation of etoposide loaded lipid polymer hybrid nanoparticles for controlled drug delivery on lung cancer. chemical methodologies 5(2) (2021) 135-152. https://doi.org/10.22034/chemm.2021.121495 [32] s. wang, h. wang, s. liu, h. guo, j. meng, m. chang, s. wu. highly sensitive detection of fluoride based on poly (3-aminophenylboronic acid)-reduced graphene oxide multilayer modified electrode. food chemistry 400 (2023) 134042. https://doi.org/10.1016/j.foodchem.2022.134042 [33] r. kavade, r. khanapure, u. gawali, a. patil, s. patil. degradation of methyl orange under visible light by zno-polyaniline nanocomposites. journal of applied organometallic chemistry 2(2) (2022) 101-112. http://dx.doi.org/10.22034/jaoc.2022.349558.1056, [34] h. shayegan, v. safarifard, h. taherkhani, m.a. rezvani. efficient removal of cobalt(ii) ion from aqueous solution using amide-functionalized metal-organic framework. journal of applied organometallic chemistry 2(3) (2022) 109-118. doi: http://dx.doi.org/10.22034/jaoc.2022.154718 [35] h. duan, d. wang, y. li. green chemistry for nanoparticle synthesis. chemical society reviews 44(16) (2015) 5778-5792. https://doi.org/10.1039/c4cs00363b [36] m. bijad, a. hojjati-najafabadi, h. asari-bami, s. habibzadeh, i. amini, f. fazeli. an overview of modified sensors with focus on electrochemical sensing of sulfite in food samples. eurasian chemical communications 3(2) (2021) 116-138. doi: http://dx.doi.org/10.22034/ecc.2021.268819.1122 https://doi.org/10.1016/j.snb.2013.07.085 http://dx.doi.org/10.22034/ecc.2021.288271.1182 https://doi.org/10.1016/j.microc.2022.107319 https://doi.org/10.1016/j.microc.2022.107319 https://doi.org/10.22034/ajgc.2021.261206.1287 https://doi.org/10.22034/ajgc.2021.261206.1287 https://doi.org/10.1016/j.msec.2016.06.064 https://doi.org/10.1016/j.chemosphere.2022.134926 http://dx.doi.org/10.22034/ecc.2021.300477.1224 http://www.jchemrev.com/article_140627.html http://www.jchemrev.com/article_140627.html http://dx.doi.org/10.22034/jcr.2021.314327.1128 https://doi.org/10.1016/j.talanta.2022.123836 https://doi.org/10.22034/chemm.2021.121495 https://doi.org/10.22034/chemm.2021.121495 https://doi.org/10.1016/j.foodchem.2022.134042 http://dx.doi.org/10.22034/jaoc.2022.349558.1056 http://dx.doi.org/10.22034/jaoc.2022.154718 https://doi.org/10.1039/c4cs00363b http://dx.doi.org/10.22034/ecc.2021.268819.1122 admet & dmpk 11(2) (2023) 227-236 propranolol determination on graphene/co3o4 nanocomposite doi: https://doi.org/10.5599/admet.1705 235 [37] m.m. abdul hassan, s. hassan, k.a. hassan. green and chemical synthesis of bimetallic nanoparticles (fe/ni) supported by zeolite 5a as a heterogeneous fenton-like catalyst and study of kinetic and thermodynamic reaction for decolorization of reactive red 120 dye from aqueous pollution. eurasian chemical communications 4 (2022) 1062-1086. https://doi.org/10.22034/ecc.2022.342067.1466 [38] s. ariavand, m. ebrahimi, e. foladi. design and construction of a novel and an efficient potentiometric sensor for determination of sodium ion in urban water samples. chemical methodologies 6 (2022) 886-904. https://doi.org/10.22034/chemm.2022.348712.1567 [39] b. paulchamy, g. arthi, b.d. lignesh. a simple approach to stepwise synthesis of graphene oxide nanomaterial. journal of nanomedicine & nanotechnology 6(1) (2015) 1. https://doi.org/10.4172/2157-7439.1000253 [40] g. nabi bidhendi, n. mehrdadi, m. firouzbakhsh. removal of lead from wastewater by iron– benzenetricarboxylate metal-organic frameworks. chemical methodologies 5 (2021) 271-284. https://doi.org/10.22034/chemm.2021.130208 [41] a. dehno khalaji, n. mohammadi, m. emami. nio nanoparticles: synthesis, characterization, and methyl green removal study. progress in chemical and biochemical research 4(4) (2021) 372-378. https://doi.org/10.22034/pcbr.2021.294420.1194 [42] a. obaid, s. al-ghabban, r. al-hussain. appraising antioxidant and antibacterial activities of zinc oxide nanoparticles synthesized biologically by iraqi propolis. chemical methodologies 6(5) (2022) 366-371. https://doi.org/10.22034/chemm.2022.332390.1448 [43] m. pirozmand, a. nezhadali, m. payehghadr, l. saghatforoush. ultratrace determination of cadmium ion in petro-chemical sample by a new modified carbon paste electrode as voltammetric sensor. eurasian chemical communications 2 (2020) 1021-1032. https://doi.org/10.22034/ecc.2020.241560.1063 [44] s.e. mousavi ghahfarokhi, k. helfi, m. zargar shoushtari. synthesis of the single-phase bismuth ferrite (bifeo3) nanoparticle and investigation of their structural, magnetic, optical and photocatalytic properties. advanced journal of chemistry-section a 5(1) (2022) 45-58. https://doi.org/10.22034/ajca.2021.309069.1284 [45] l. qian, s. durairaj, s. prins, a. chen. nanomaterial-based electrochemical sensors and biosensors for the detection of pharmaceutical compounds. biosensors and bioelectronics 175 (2021) 112836. https://doi.org/10.1016/j.bios.2020.112836 [46] v. tallapaneni, l. mude, d. pamu, v.v.s.r. karri. formulation, characterization and in vitro evaluation of dual-drug loaded biomimetic chitosan-collagen hybrid nanocomposite scaffolds. journal of medicinal and chemical sciences 5 (2022) 1059-1074. https://doi.org/10.26655/jmchemsci.2022.6.19 [47] a. kannan, s. radhakrishnan. fabrication of an electrochemical sensor based on gold nanoparticles functionalized polypyrrole nanotubes for the highly sensitive detection of l-dopa. materials today communications 25 (2020) 101330. https://doi.org/10.1016/j.mtcomm.2020.101330 [48] r.m. mohabis, f. fazeli, i. amini, v. azizkhani. an overview of recent advances in the detection of ascorbic acid by electrochemical techniques. journal of electrochemical science and engineering 12(6) (2022) 1081-1098. https://doi.org/10.5599/jese.1561 [49] h. karimi-maleh, m. sheikhshoaie, i. sheikhshoaie, m. ranjbar, j. alizadeh, n.w. maxakato, a. abbaspourrad. a novel electrochemical epinine sensor using amplified cuo nanoparticles and anhexyl-3-methylimidazolium hexafluorophosphate electrode. new journal of chemistry 43(5) (2019) 2362-2367. https://doi.org/10.1039/c8nj05581e [50] a. meoipun, k. kaewjua, o. chailapakul, w. siangproh. a simple and fast flow injection amperometry for the determination of methimazole in pharmaceutical preparations using an unmodified borondoped diamond electrode. admet and dmpk (2023). https://doi.org/10.5599/admet.1584 [51] p.m. jahani. flower-like mos2 screen-printed electrode based sensor for the sensitive detection of sunset yellow fcf in food samples. journal of electrochemical science and engineering 12(6) (2022) 1099-1109. https://doi.org/10.5599/jese.1413 [52] s.a. alavi-tabari, m.a. khalilzadeh, h. karimi-maleh. simultaneous determination of doxorubicin and dasatinib as two breast anticancer drugs uses an amplified sensor with ionic liquid and zno https://doi.org/10.5599/admet.1705 https://doi.org/10.22034/ecc.2022.342067.1466 http://www.chemmethod.com/article_155298.html http://www.chemmethod.com/article_155298.html https://doi.org/10.22034/chemm.2022.348712.1567 https://doi.org/10.4172/2157-7439.1000253 https://doi.org/10.4172/2157-7439.1000253 https://doi.org/10.22034/chemm.2021.130208 https://doi.org/10.22034/pcbr.2021.294420.1194 https://doi.org/10.22034/chemm.2022.332390.1448 https://doi.org/10.22034/ecc.2020.241560.1063 https://doi.org/10.22034/ajca.2021.309069.1284 https://doi.org/10.22034/ajca.2021.309069.1284 https://doi.org/10.1016/j.bios.2020.112836 https://doi.org/10.26655/jmchemsci.2022.6.19 https://doi.org/10.1016/j.mtcomm.2020.101330 https://doi.org/10.5599/jese.1561 https://doi.org/10.1039/c8nj05581e https://doi.org/10.5599/admet.1584 https://doi.org/10.5599/jese.1413 p. karami-kolmoti and r. zaimbashi admet & dmpk 11(2) (2023) 227-236 236 nanoparticle. journal of electroanalytical chemistry 811 (2018) 84-88. https://doi.org/10.1016/j.jelechem.2018.01.034 [53] c.m. brett. electrochemical impedance spectroscopy in the characterisation and application of modified electrodes for electrochemical sensors and biosensors. molecules 27(5) (2022) 1497. https://doi.org/10.3390/molecules27051497 [54] s.z. mohammadi, f. mousazadeh, m. mohammadhasani-pour. electrochemical detection of folic acid using a modified screen printed electrode. journal of electrochemical science and engineering 12(6) (2022) 1111-1120. https://doi.org/10.5599/jese.1360 [55] v. sanko, a. şenocak, s. o. tümay, y. orooji, e. demirbas, a. khataee. an electrochemical sensor for detection of trace-level endocrine disruptor bisphenol a using mo2ti2alc3 max phase/mwcnt composite modified electrode. environmental research 212 (2022) 113071. https://doi.org/10.1016/j.envres.2022.113071 [56] h. karimi-maleh, a.f. shojaei, k. tabatabaeian, f. karimi, s. shakeri, r. moradi. simultaneous determination of 6-mercaptopruine, 6-thioguanine and dasatinib as three important anticancer drugs using nanostructure voltammetric sensor employing pt/mwcnts and 1-butyl-3-methylimidazolium hexafluoro phosphate. biosensors and bioelectronics 86 (2016) 879-884. https://doi.org/10.1016/j.bios.2016.07.086 [57] s. saghiri, m. ebrahimi, m. bozorgmehr. electrochemical amplified sensor with mgo nanoparticle and ionic liquid: a powerful strategy for methyldopa analysis. chemical methodologies 5(3) (1999) 234239. https://doi.org/10.22034/chemm.2021.128530 [58] t. eren, n. atar, m. l. yola, h. karimi-maleh. a sensitive molecularly imprinted polymer based quartz crystal microbalance nanosensor for selective determination of lovastatin in red yeast rice. food chemistry 185 (2015) 430-436. https://doi.org/10.1016/j.foodchem.2015.03.153 [59] s. saghiri, m. ebrahimi, m.r. bozorgmehr. nio nanoparticle/1-hexyl-3-methylimidazolium hexafluorophosphate composite for amplification of epinephrine electrochemical sensor. asian journal of nanosciences and materials 4(1) (2021) 46-52. https://doi.org/10.26655/ajnanomat.2021.1.4 [60] h. karimi-maleh, r. darabi, m. shabani-nooshabadi, m. baghayeri, f. karimi, j. rouhi, m. alizadeh, o. karaman, y. vasseghian, c. karaman. determination of d&c red 33 and patent blue v azo dyes using an impressive electrochemical sensor based on carbon paste electrode modified with zif-8/g-c3n4/co and ionic liquid in mouthwash and toothpaste as real samples. food and chemical toxicology 162 (2022) 112907. https://doi.org/10.1016/j.fct.2022.112907 [61] h. setiyanto, d. r. purwaningsih, v. saraswaty, n. mufti, m. a. zulfikar. highly selective electrochemical sensing based on electropolymerized ion imprinted polyaniline (iipani) on a bismuth modified carbon paste electrode (cpe-bi) for monitoring nickel (ii) in river water. rsc advances 12(45) (2022) 2955429561. https://doi.org/10.1039/d2ra05196f [62] f. hosseini, m. bahmaei, m. davallo. electrochemical determination of propranolol, acetaminophen and folic acid in urine, and human plasma using cu2o–cuo/rgo/cpe. russian journal of electrochemistry 57(4) (2021) 357-374. https://doi.org/10.1134/s1023193521040054 [63] j. zoubir, y. elkhotfi, c. radaa, n. bougdour, a. idlahcen, i. bakas, a. assabbane . electrocatalytic detection of dimetridazole using an electrochemical sensor ag@ cpe. analytical application. milk, tomato juice and human urine. sensors international 2 (2021) 100105. https://doi.org/10.1016/j.sintl.2021.100105 [64] a. j. bard, l. r. faulkner, electrochemical methods: fundamentals and applications, john wiley & sons, new york, 2nd edition, 2001. ©2023 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) https://doi.org/10.1016/j.jelechem.2018.01.034 https://doi.org/10.1016/j.jelechem.2018.01.034 https://doi.org/10.3390/molecules27051497 https://doi.org/10.5599/jese.1360 https://doi.org/10.1016/j.envres.2022.113071 https://doi.org/10.1016/j.envres.2022.113071 https://doi.org/10.1016/j.bios.2016.07.086 https://doi.org/10.1016/j.bios.2016.07.086 https://doi.org/10.22034/chemm.2021.128530 https://doi.org/10.1016/j.foodchem.2015.03.153 https://doi.org/10.26655/ajnanomat.2021.1.4 https://doi.org/10.1016/j.fct.2022.112907 https://doi.org/10.1039/d2ra05196f https://doi.org/10.1134/s1023193521040054 https://doi.org/10.1016/j.sintl.2021.100105 https://doi.org/10.1016/j.sintl.2021.100105 http://creativecommons.org/licenses/by/3.0/ microsoft word 130-588-1-ed_varadi doi: 10.5599/admet.2.4.130 272 admet & dmpk 2(4) (2014) 272-281; doi: 10.5599/admet.2.4.130 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper efficient transdermal delivery of benfotiamine in an animal model gyula varadi�, zhen zhu and stephen g. carter biochemics, inc., 99 rosewood drive, suite 270, danvers, ma 01923-4537, usa � corresponding author: gyula varadi, e-mail: gvaradi@biochemics.com; tel.: +1-978-750-0090; fax: +1-978-7500080 received: november 13, 2014; revised: december 13, 2014; published: january 09, 2015 abstract we designed a transdermal system to serve as a delivery platform for benfotiamine utilizing the attributes of passive penetration enhancing molecules to penetrate through the outer layers of skin combined with the advance of incorporating various peripherally-acting vasodilators to enhance drug uptake. benfotiamine, incorporated into this transdermal formulation, was applied to skin in an animal model in order to determine the ability to deliver this thiamine pro-drug effectively to the sub-epithelial layers. in this proof of concept study in guinea pigs, we found that a single topical application of either a solubilized form of benfotiamine (15 mg) or a microcrystalline suspension form (25 mg) resulted in considerable increases of the dephosphorylated benfotiamine (s-benzoylthiamine) in the skin tissue as well as in significant increases in the thiamine and thiamine phosphate pools compared to control animals. the presence of a ~8000x increase in thiamine and increases in its phosphorylated derivatives in the epidermis and dermis tissue of the test animals gives a strong indication that the topical treatment with benfotiamine works very well for the desired outcome of producing an intracellular increase of the activating cofactor pool for transketolase enzyme, which is implicated in the pathophysiology of diabetic neuropathy. keywords transdermal drug delivery; benfotiamine; solubilized benfotiamine; microcrystalline benfotiamine cream; thiamine, pentose phosphate cycle, pharmacokinetics; diabetic peripheral neuropathy introduction in diabetic conditions, hyperglycemia induces an altered cellular metabolism, and through a series of events results in the creation of damaging cellular metabolites by influencing a number of glucose metabolizing pathways [1,2]. elevated glucose level increases the metabolism through the polyol pathway; increases fructose-6-phosphate and glyceraldehyde-3-phosphate levels, which latter in turn increases the use of the diacylglycerol pathway leading to an increase in protein kinase c pathway; increases the production of superoxide molecules and results in increased production of advanced glycation endproducts (ages) [2], a biochemical problem known to be affected in neuropathy. ages inhibit key cellular proteins from functioning properly [2-4]. in addition, ages have been implicated in the induction of admet & dmpk 2(4) (2014) 272-281 transderamal delivery of benfotiamine in an animal model doi: 10.5599/admet.2.4.130 273 apoptosis, including peripheral nerve cells resulting eventually in the loss of sensation to the affected area and in an altered neuronal signaling process, which may present as pain [1,2,5]. there are many candidate molecular targets for the treatment of diabetic neuropathy [6-8]. several have been addressed through clinical trial evaluations [8] and most have failed [9]. one of the most considered endpoints for the development of a therapeutic for this diabetic complication is the treatment of pain associated with the neuronal deterioration. as a result, there have been several candidate therapies using oral and topically applied pain relievers ranging from pregabalin [10,11] to capsaicin [12]. apparently, in both forms (type 1 and type 2) of diabetic condition there is about 70 % thiamine deficiency as measured in the plasma of diabetic patients [13-15] but the physiological mechanism associated with this deficiency has not been clarified yet. however, a number of studies found an altered thiamine handling in the kidneys of diabetic patients [16-18]. benfotiamine, (s-benzoylthiamine-omonophosphate), a derivative of the vitamin b1 or thiamine, has a much higher bioavailability than thiamine and, as a result, accumulates in the target tissues much more than thiamine itself [15,19,20]. thiamine absorption from oral benfotiamine is approximately five times greater than from conventional thiamine supplements [21-23]. once benfotiamine has been transported to the target tissue, it is converted through a series of metabolic and enzymatic steps to s-benzoylthiamine, which after internalization into the cell forms thiamine. this latter then serves as a substrate for enzymatic conversion to form thiamine mono-, di-, and triphosphates. it is however the thiamine diphosphate (tdp) that is responsible for binding to and activating the enzyme transketolase, a critical enzyme for the metabolism of the hyperglycemic by-products via the pentose phosphate pathway to generate metabolic products that are not damaging to the cells. benfotiamine was employed to treat alcoholic neuropathy [24,25] sciatica and other painful nerve conditions [26,27] known to be affected with low thiamine content [28-30]. the application of benfotiamine and the resulting activation of the enzyme transketolase in the diabetic, hyperglycemic system can also produce beneficial effects on general nerve health. benfotiamine has been shown in many preclinical [21,31-43] as well as clinical studies [44-57] to prevent the formation of ages [21,53] and eventually with continued treatment to reverse the symptoms associated with the neuropathy [46,47,52,57]. by increasing intracellular levels of thiamine, benfotiamine indirectly induces enzymatic and biochemical pathways [58-60] through elevated concentrations of thiamine diphosphate resulting in a reduced level of damaging glucose-derived chemical protein species [21,22,32,37-39,60-62]. benfotiamine is currently offered as a prescriptive drug for the treatment of diabetic neuropathy in europe in an oral form. it is also available for similar indications as a homeopathic medicine and as a dietary supplement. the typical oral dose recommended is 40-450 mg/day [63,64]. the topical application of benfotiamine may allow more prodrug to be introduced into the target tissues and, as a result, potentially generate a greater therapeutic effect. experimental we prepared two topical formulations for these studies. the topical formulation termed as 3 % benfotiamine was prepared according to a vale formulation [65] and contained 3 % benfotiamine in soluble form, neutralized with arginine. the 5 % topical formulation contained 5 % benfotiamine, also in a vale formulation, but in microcrystalline form. the stability of these products was tested after one year storage at 4 °c and we found that 89 % of benfotiamine was present in the active chemical form. varadi et.al. admet & dmpk 2(4) (2014) 272-281 274 we utilized male hartley guinea pigs (ca. 250-300 g) for pharmacokinetic studies with two topical benfotiamine formulations. the guinea pigs fur was clipped from their back, and an area of 4x4 cm depilated with nair. after ca. 24 hours, 0.5 g of test formulation was applied to a 2x4 cm area. the test formulations tested either contained a solubilized topical benfotiamine (3 %) yielding a dose of 15 mg of benfotiamine or a microcrystalline suspension form of benfotiamine (5 %) yielding a dose of 25 mg benfotiamine. at 0.5, 1, 2, 4, 8 and 24 hours after application, groups of 5 animals were anesthetized with isoflurane and blood removed by cardiac stick. one group of 5 untreated animals was anesthetized and blood removed as vehicle control. blood samples were directly processed to analyze benfotiamine, s-benzoylthiamine, thiamine and its derivatives by hplc [66-70]. the skin below the area of topical application on each animal in the 24 hour exposure group was cleaned, rubbed with alcohol, with the intent to remove the stratum corneum, excised, weighed, homogenized in phosphate buffered saline, centrifuged, and ca. 2 ml extract processed for pharmacokinetic analyses [66-70]. the blood and the homogenized skin tissue were evaluated at various time points following the single topical application for the presence of the formulated active pharmacological ingredient (bc-dn-01: benfotiamine); the dephosphorylated drug intermediate (sbt: s-benzoylthiamine); as well as the cellular metabolites of thiamine (t); thiamine diphosphate; (tdp) and thiamine monophosphate (tmp). results and discussion the biochemistry and required chemical conversions for benfotiamine from a phosphorylated openringed structure to a de-phosphorylated closed ring and then subsequently forming thiamine and the phosphorylated species of thiamine to serve as cofactors in the transketolase reactions are found in many cell types including the skin tissue epidermal and dermal cells, as well as presumably peripheral nerve endings in the skin [22,23,59-62]. we believe that the delivery of benfotiamine to a patient via a topical route, as opposed to an oral administration, offers significantly greater efficacy as a result of the greater quantity of the drug directly deposited into the skin tissue where the neuronal endings are located. to examine this concept, we initiated a set of experiments to first demonstrate the viability and stability of preparing a topical formulation of benfotiamine. while the substance benfotiamine has been available for use as a solid and orally delivered therapeutic for several years the development of a novel topical formulation of benfotiamine has been completed for these studies. we formulated benfotiamine into topical preparation to maximize local drug accumulation and to minimize systemic distribution [65]. the goal of these formulations and administration route was to deliver significantly greater quantities of the drug to the afflicted peripheral dermal neuronal tissue. next, we have performed a series of animal pharmacokinetic and tissue deposition experiments to determine the time course of appearance of thiamine derivatives in the plasma after treatment with benfotiamine formulations. in addition, we also analyzed the amounts of thiamine and the phosphorylated cellular metabolites in the skin tissue at a 24 h time point after topical treatment with benfotiamine. we measured the thiamine (t), thiamine diphosphate (tdp), thiamine monophosphate (tmp) and calculated the level of total thiamine derivatives (total) in the blood after the application of 0.5 g of 3 % solubilized benfotiamine topical formulation. upon treatment, we observed a sharp rise in the blood concentration of each of the thiamine derivatives within 30 minutes of the drug application (figure 1). there was an additional increase until one hour, followed by a modest decline until two hours, then the admet & dmpk 2(4) (2014) 272-281 transderamal delivery of benfotiamine in an animal model doi: 10.5599/admet.2.4.130 275 blood concentration of the individual derivatives and the total remained at an elevated level until 24 h after the application (figure 1). figure 1. plasma concentration of cellular metabolites after topical application of 15 mg bc-dn-01 (solubilized). the pharmacokinetic profile of plasma concentrations of thiamine (t), thiamine diphosphate (tdp) and thiamine monophosphate (tmp) were analyzed following the single application 15 mg of topical bc-dn-01. the data shows a rapid penetration through the skin into the plasma for all three thiamine derivatives with a peak for the cumulative group at 1 hour, however, holding an approximate 50 % increase over normal level up to 24 hours. n=5 for plasma data. the application of 0.5 g of 5 % microcrystalline benfotiamine topical formulation resulted in an appreciable enhancement of thiamine and its phosphorylated derivatives in the blood. again, a sharp rise was observed in the blood concentration of each of the thiamine derivatives within 30 minutes of the drug application (figure 2). this enhancement continued to further increase until one hour, followed by a sharp decline in the blood concentration of the individual derivatives and the total remained at an elevated level until 24 hrs after the application. interestingly, while the microcrystalline formulation contained almost twice as high concentration of the active pharmacological ingredient than the solubilized benfotiamine formulation, the 24 hours steady state concentrations in the blood resembled to those values that were observed with the solubilized formulation containing a lower (3 %) benfotiamine concentration. this is an indication that in case of each applied formulation the benfotiamine level was at saturating concentration. moreover, the more pronounced rise at one hour and the drastic drop thereafter in the concentration of thiamine derivatives is consistent with a robust overload (figure 2). varadi et.al. admet & dmpk 2(4) (2014) 272-281 276 figure 2. plasma concentration of cellular metabolites after topical application of 25 mg bc-dn-01 (suspension). the pharmacokinetic profile of plasma concentrations of thiamine (t), thiamine diphosphate (tdp) and thiamine monophosphate (tmp) were analyzed following the single application 25 mg of topical suspended bc-dn-01. the data shows a rapid penetration through the skin into the plasma for all three thiamine derivatives with a peak for the cumulative group at 1 hour with an approximate 70 % increase in the group, however, dropping off to a more normal level by 4 hours. n=5 for plasma data at 0.5, 1, 2, 4 and 8 hours and n=1 at t=24 hours time point. the measurement of benfotiamine, s-benzoyl-benfotiamine, thiamine and thiamine related compounds in skin tissues attested that there was no obvious change of tdp and tmp levels compared to controls over the 24 hour period when treated with solubilized benfotiamine cream (figure 3a). the level of benfotiamine appeared low, however, there was a drastic rise in s-benzoyl thiamine and thiamine levels. as shown in figure 3a, bft and s-benzoyl thiamine skin level (stratum corneum removed) was 1.64 nmol/g and 1,492.33 nmol/g respectively. concomitant with these changes, the thiamine level increased from 0.32 nmol/g to 809.25 nmol/g, resulting in a 2,500-fold enhancement in treated animals. the results of the accumulation of thiamine and thiamine derivatives in the skin tissue located under the application site after 24 h treatment with the 5 % microcrystalline formulation are summarized in figure 3b. similarly to the treatment with solubilized 3 % benfotiamine formulation, we observed no obvious change of tdp and tmp levels compared to controls over the 24 hour period. the level of benfotiamine appeared low, however, there was a drastic rise in s-benzoyl thiamine and thiamine levels. concomitantly, the thiamine level increased from 0.32 nmol/g to 2,538.29 nmol/g at t = 24 hours that is a 7,880-fold increase in treated animals. we could not determine the significance of the increase in skin tissue thiamine level after 24 hrs with treatment of 3 % solubilized benfotiamine formulation due to the limited number of animal subjects involved in this experiment (n=2), however, the treatment with the 5% microcrystalline benfotiamine cream formulation resulted in more robust enhancement in skin tissue thiamine level and this value appeared significant when compared to non-treated control (p=0.03) (figure 3c). admet & dmpk 2(4) (2014) 272-281 transderamal delivery of benfotiamine in an animal model doi: 10.5599/admet.2.4.130 277 figure 3. a; concentration of bc-dn-01 and metabolites in skin tissue 24 h after 15 mg topical application of bc-dn-01 (solubilized). analysis of the content of bc-dn-01 in skin tissue following a single application of 15 mg of solubilized bc-dn-01 at t=0 for the presence of bc-dn-01, its dephosphorylated intermediate metabolite (sbt), and thiamine (t), thiamine diphosphate (tdp) and thiamine monophosphate (tmp). in addition, control levels of thiamine from untreated skin were analyzed and graphed. n=3 for control, bc-dn-01, tdp and tmp in skin tissue samples. n=2 for determination of sbt and t in skin tissue samples. b; concentration of bc-dn-01 and metabolites in skin tissue 24 h after 25 mg topical application of bc-dn-01 (suspension). analysis of the content of bc-dn01 in skin tissue following a single application of 25 mg of suspended form of bc-dn-01 at t=0 for the presence of bc-dn-01, its dephosphorylated intermediate metabolite (sbt), and the thiamine (t), thiamine diphosphate (tdp) and thiamine monophosphate (tmp). in addition, control levels of thiamine from untreated skin were analyzed and graphed; n=3 for skin. c; thiamine levels in the skin after 24 h with treatment of 3 % bc-dn-01 and 5% bc-dn-01. numbers in red are mean values from independent determinations as indicated in a and b. bc-dn-01 sbt t tdp tmp c o n c e n tr a ti o n i n s k in ( n m o l/ g ) 0 1000 2000 3000 4000 6.51 1875.19 2538.29 0.06 1.03 bc-dn-01 sbt t tdp tmp c o n c e n tr a ti o n i n s k in ( n m o l/ g ) 0 500 1000 1500 2000 2500 1.64 1492.33 809.25 0.04 0.56 a b c 3% bc-dn-01 5% bc-dn-01 control c o n c e n tr a ti o n i n s k in ( n m o l/ g ) 0 1000 2000 3000 809.25 2538.29 0.32 b c varadi et.al. admet & dmpk 2(4) (2014) 272-281 278 these pharmacokinetic studies demonstrated that transdermally-delivered benfotiamine is capable of serving as a prodrug for blood and intradermal thiamine production. by making use of alternative drug forms of the benfotiamine as well as modifying the formulation cofactors and solvents we were able to achieve a significant bioavailability in the target tissue of the skin tissue while also protecting the integrity of the molecule for stability purposes. our experimental data from these pharmacokinetic studies with topical benfotiamine formulation has demonstrated a significant, ~8000-fold, increase in the level of thiamine chemical species in the skin tissue of animals receiving a single 25 mg dose of topical benfotiamine at the 24 hour time point following the application. the administration of benfotiamine as a topical prodrug offers several advantages over an oral delivery. firstly, benfotiamine contains a unique open-ringed structure giving the molecule a significantly higher bioavailability index compared to thiamine [71-73]. thus, the first pass liver metabolism as well as metabolic conversions of the drug crossing the intestinal wall will be avoided with a topical administration. secondly, the oral delivery of benfotiamine has been shown to significantly accumulate in the erythrocytes [74] and then in the liver as opposed to being delivered in significant doses to the peripheral nerves in the skin of the legs and hands. topical delivery of benfotiamine will concentrate the drug in the target tissues immediately without dilution. as a result, the impact on the damaged cells and tissues should respond better than the oral applications. in contrast to using thiamine as the drug to treat diabetic peripheral neuropathy, benfotiamine is a chemical analog of thiamine that is considerably more bioavailable at the systemic as well as the cellular level than thiamine. therefore, if administered to patients with diabetic neuropathy, topical benfotiamine should have a favorable efficacy profile through the introduction of greater quantities of thiamine into the cell. thiamine cellular absorption derived from orally administered benfotiamine is about five times greater than those levels obtained from conventional thiamine supplements [19,71-73]. this differential bioavailability is even greater in the brain and muscle tissues that yield fiveto twenty-five-fold more thiamine [22] after a benfotiamine application compared to equivalent doses of thiamine. for benfotiamine given transdermally, a similar cascade of biochemistry occurs with the dephosphorylation of the native benfotiamine catalyzed by phosphatase enzymes positioned on the outside of epidermal cells producing s-benzoylthiamine. once in this form, the s-benzoylthiamine prodrug follows the same passive transport across the membrane of the cells of the epidermis and dermis including those of the schwann cells and neurons, delivering a high dose of the prodrug and subsequently the thiamine to those cells. the thiamine diphosphate produced from the elevated thiamine activates the transketolase and the pentose phosphate pathway, altering the glucose metabolism to reduce the formation of ages. reduction of ages in the schwann and neuronal cells reduces the expression of neuropathy symptoms and also addresses the base etiology of the disease state. conclusions conclusions from the proof of concept study in guinea pigs is primarily that after single topical application of either a solution form of benfotiamine (15 mg) or a microcrystalline suspension form (25 mg) we found considerable increases of the dephosphorylated benfotiamine (s-benzoylthiamine) in the skin tissue as well as significant increases in the thiamine and thiamine phosphate pools compared to control animals. the presence of a ~8000x increase in thiamine and increase in its phosphorylated derivatives in the epidermis and dermis tissue of the test animals gives a strong indication that the topical treatment with benfotiamine works very well for the desired outcome of producing an intracellular increase of the activating cofactor pool for transketolase. subsequently, it is reasonable to conclude that this metabolite admet & dmpk 2(4) (2014) 272-281 transderamal delivery of benfotiamine in an animal model doi: 10.5599/admet.2.4.130 279 level increase, including tdp, as a function of cellular compartmentalization of metabolites and pools of the thiamine derivatives, will well serve the diabetic neuropathic disease condition by activating the transketolase and activating the pentose phosphate pathway as an alternative mechanism to metabolize glucose into harmless end-products as opposed to the continued production of advanced glycation endproducts. acknowledgements: this work was supported by funds from biochemics, inc. references [1] m. brownlee, nature 414 (2001) 813-820. [2] h.p. hammes, x. du, d. edelstein, t. taguchi, t. matsumura, q, ju, j. lin, a. bierhaus, p. nawroth,d. hannak, m. neumaier, r. bergfeld, i. giardino, m. brownlee, nature med. 9 (2003) 294-299. [3] k.a. head, altern. med. rev. 11 (2006) 294-329. [4] t. varkonyi, p. kempler, diabetes obes. metab. 10 (2008) 99-108. [5] a. bierhaus, t. fleming, s. stoyano, a. leffler, a. babes, c. neacsu, s.k. sauer, m. eberhard, m. schnölze, f. lasischka, w.l. neuhuber, t.i. kichko, i. konrade, r. elvert, w. mier, v. pirags, i.k. lukić, m. morcos, t. dehmer, n. rabbani, p.j. thornalley, d. edelstein, c. nau, j. forbes, p.m. humpert, m. schwaninger, d. ziegler, d.m. stern, m.e. cooper, u. haberkorn, m. brownlee, p.w. reeh, p.p. nawroth, nature med. 18 (2012) 926-933. [6] the diabetes control and complications trial research group. new engl. j. med. 329 (1993) 977986. [7] the effect of intensive diabetes therapy on the development and progression of neuropathy. ann. int. med. 122 (1995) 561-568. [8] m.c. wong, j.w. chung, t.k. wong, bmj (2007) 335:387. [9] r.h. dworkin, d.c. turk, k.w. wyrwich, d. beaton, c.s. cleeland, j.t. farrar, j.a. haythornthwaite, m.p. jensen, r.d. kerns, d.n. ader, n. brandenburg, l.b. burke, d. cella, j. chandler, p. cowan, r. dimitrova, r. dionne, s. hertz, a.r. jadad, n.p. katz, h. kehlet, l.d. kramer, d.c. manning, c. mccormick, m.p. mcdermott, h.j. mcquay, s. patel, l. porter, s. quessy, b.a. rappaport, c. rauschkolb, d.a. revicki, m. rothman, k.e. schmader, b.r. stacey, j.w. stauffer, t. von stein, r.e. white, j. witter, s. zavisi, pain 9 (2008) 105-121. [10] r.a. moore, s. straube, p.j. wiffen, s. derry, h.j. mcquay, the cochrane database of systematic reviews (2009) cd007076. [11] s. quilici, j. chancellor, m. lothgren, d. simon, g. said, t.k. le, a. garcia-cebrian, b. monz, bmc neurology 9 (2009) 6. [12] s. derry, a. sven-rice, p. cole, t. tan, r.a. moore, the cochrane database of systematic reviews 2 (2013) cd007393. [13] g.l. page, d. laight, m.h. cummings, int. j. clin. pract. 65 (2011) 684-690. [14] n. rabbani, p.j. thornalley, diabetes obes. metab. 13 (2011) 577-583. [15] m. gangolf, j. czerniecki, m. radermecker, o. detry, m. nisolle, c. jouan, d. martin, f. chantraine, b. lakaye, p. wins, t. grisar, l. bettendorff, plos one 5 (2010) e13616. [16] p.j. thornalley, r. babaei-jadidi, h. al ali, n. rabbani, a. antonysunil, j. larkin, a. ahmed, g. rayman, c.w. bodmer, diabetologia 50 (2007) 2164-2170. [17] j. r. larkin, f. zhang, l. godfrey, g. molostvov, d. zehnder, n. rabbani, p. j. thornalley, plos one 7 (2012) e53175. [18] r. babaei-jadidi, n. karachalias, n. ahmed, s. battah, p.j. thornalley, diabetes 52 (2003) 2110-2120. [19] r. masson, c. quincy, p. girard, r. later, b. bochu, m.f. cotte, lyon med. 216 (1966) 685-704. [20] r. hilbig h. rahmann, arzneimittelforschung 48 (1998) 461-468. varadi et.al. admet & dmpk 2(4) (2014) 272-281 280 [21] h. stracke, h.p. hammes, d. werkmann, k. mavrakis, i. bitsch, m. netzel, j. geyer, w. köpcke, c. sauerland, r.g. bretzel, k.f. federlin, exp. clin. endocrinol. diabetes 109 (2001) 330-336. [22] m. l. volvert, s. seyen, m. piette, b. evrard, m. gangolf, j. c. plumier, l. bettendorf, bmc pharmacol. 8 (2008) 10. [23] e. beltramo, e. berrone, s. tarallo, m. porta, acta diabetol. 45 (2008), 131-141. [24] h. woelk, s. lehrl, r. bitsch, w. kopcke, alcohol 33 (1998) 631-638. [25] e.i. anisimova, a.b. danilov, zh. nevrol. psikhiatr. im. s. s korsakova 101 (2001) 32-36. [26] e. haupt, h. ledermann, w. kopcke, int. j. clin. pharmacol. ther. 43 (2005) 71-77. [27] g. winkler, b. pal, e. nagybeganyi, i ory, m. porochnavec, p. kempler, arzneimittelforschung 49 (1999) 220-224. [28] h. stracke, w. gauss, u. achenbach, k. federlin, r.g. bretzel, exp. clin. endocrinol. diabetes 116 (2008) 600-605. [29] x. du, d. edelstein, m. brownlee, diabetologia 51 (2008) 1930-1932. [30] c.d. ang, m.j. alviar, a.l. dans, g.g. bautista-velez, v.v. villaruz-sulit, j.j. tan, h.u. co, m.r. bautista, cochrane database systematic reviews 16 (2008) cd004573. [31] n. karachalias, r. babaei-jadidi, n. ahmed, p.j. thornalley, biochem. soc. trans. 31 (2003) 14231425. [32] r. babaei-jadidi, n. karachalias, c. kupich, n. ahmed, p.j. thornalley, diabetologia 47 (2004) 22352246. [33] n. karachalias, r. babaei-jadidi, c. kupich, n. ahmed, p.j. thornalley, ann. n. y. acad. sci. 1043 (2005) 777-783. [34] n. karachalias, r. babaei-jadidi, n. rabbani, p.j. thornalley, diabetologia 53 (2010) 1506-1516. [35] a.f. ceylan-isik, s. wu, q. li, s. y. li, j. ren, j. appl. physiol. 100 (2006) 150-156. [36] y. kohda, h. shirakawa, k. yamane, k. otsuka, t. kono, f. terasaki, t. tanaka, j. toxicol. sci. 33 (2008) 459-472. [37] s. wu, j. ren, neurosci. lett. 394 (2006) 158-162. [38] s. gadau, c. emanueli, s. van linthout, g. graiani, m. todaro, m. meloni, i. campesi, g. invernici, f. spillmann, k. ward, p. madeddu, diabetologia 49 (2006) 405-420. [39] g.m. sanchez-ramirez, n.l. caram-salas, h.i. rocha-gonzalez, g.c. vidal-cantu, r. medinasantillan, g. reyes-garcia, eur. j. pharmacol. 530 (2006) 48-53. [40] r.g. katare, a. caporali, a. oikawa, m. meloni, c. emanueli, p. madeddu, circ. heart fail. 3 (2010) 294-305. [41] r. katare, a. caporali, c. emanueli, p. madeddu, j. mol. cell cardiol. 49 (2010) 625-638. [42] y. kohda, m. umeki, t. kono, f. terasaki, h. matsumura, t. tanaka, j. pharmacol. sci. 113 (2010) 343-352. [43] r. katare, a. oikawa, d. cesselli, a.p. beltrami, e. avolio, d. muthukrishnan, p.e. munasinghe, g. angelini, c. emanueli, p. madeddu, cardiovasc. res. 97 (2013) 55-65. [44] h. stracke, a. lindemann, k. federlin, exp. clin. endocrinol. diabetes 104 (1996) 311-316. [45] e. haupt, h. ledermann, w. köpcke, int. j. clin. pharmacol. ther. 43 (2005) 71-77. [46] s. arora, a. lidor, c.j. abularrage, j.m. weiswasser, e. nylen, d. kellicut, a.n. sidawy, ann. vasc. surg. 20 (2006) 653-658. [47] x. du, d. edelstein, m. brownlee, diabetologia 51 (2008) 1930-1932. [48] n. schupp, e.m. dette, u. schmid, u. bahner, m. winkler, a. heidland, h. stopper, naunyn schmiedebergs arch. pharmacol. 378 (2008) 283-291. [49] s. riaz, v. skinner, s.k. srai, j. pharm. biomed. anal. 54 (2011) 817-825. [50] n. rabbani, s.s. alam, s. riaz, j.r. larkin, m.w. akhtar, t. shafi, p.j. thornalley, diabetologia 52 (2009) 208-212. admet & dmpk 2(4) (2014) 272-281 transderamal delivery of benfotiamine in an animal model doi: 10.5599/admet.2.4.130 281 [51] a. alkhalaf, a. klooster, w. van oeveren, u. achenbach, n. kleefstra, r.j. slingerland, g.s. mijnhout, h.j. bilo, r.o. gans, g.j. navis, s.j. bakker, diabetes care, 33 (2010) 1598-1601. [52] m. gonzález-ortiz, e. martínez-abundis, j.a. robles-cervantes, v. ramírez-ramírez, m.g. ramos-zavala, eur. j. nutr. 50 (2011) 145-149. [53] a. alkhalaf, n. kleefstra, k.h. groenier, h.j. bilo, r.o. gans, p. heeringa, j.l. scheijen, c.g. schalkwijk, g.j. navis, s.j. bakker, plos one 7 (2012) e40427. [54] d.a. fraser, l.m. diep, i.a. hovden, k.b. nilsen, k.a. sveen, i. seljeflot, k.f. hanssen, diabetes care 35 (2012) 1095-1097. [55] f.c. polizzi, g. andican, e. çetin, s. civelek, v. yumuk, g. burçak, exp. clin. endocrinol. diabetes 120 (2012) 329-334. [56] a. stirban, a. pop, a. fischer, s. heckermann, d. tschoepe, diabetes technol. ther. 15 (2013) 733737. [57] a. stirban, a. pop, d. tschoepe, diabet. med. 30 (2013) 1204-1208. [58] e. berrone, e. beltramo, c. solimine, a.u. ape, m. porta, j. biol. chem. 281 (2006) 9307-9313. [59] v. marchetti, r. menghini, s. rizza, a. vivanti, t. feccia, d. lauro, a. fumikazu, r. lauro, m. federici, diabetes 55 (2006) 2231-2237. [60] f. pomero, a. min molinar, m. la selva, a allione, g.m. molinatti, m. porta, acta diabetol. 38 (2001) 135-138. [61] e. beltramo, e. berrone, s. buttiglieri, m. porta, diabetes metab. res. rev. 20 (2004) 330-336. [62] s.h. oh r.p. witek, s.h. bae, h.a. darwiche, y. jung, l. pi, a. brown, b.e. petersen, stem cells dev. 18 (2009) 37-46. [63] t. hanawa, a. watanabe, t. tsuchiya, r. ilkoma, m. hidaka, m. sugihara, chem. pharm. bull. (tokyo) 43 (1995) 872-876. [64] s. simeonov, m. pavlova, m. mitkov, l. mincheva, d. troev, folia med. (plovdiv) 39 (1997) 5-10. [65] s.g. carter, z. zhu, g. varadi, a. veves, j.e. riviere, j. pharmaceut. sci. 102 (2013) 4072-4078. [66] c. herve, p. beyne, e. delacoux, j. chromatogr. b, 653 (1994) 217-220. [67] m. ziems, m. netzel, s. peter, i. bitsch, vitamine und zusatzstoffe in der ernahrung von mensch und tier, 5. symposium jena (1995) 320-323 [68] p.l.m. lynch, e.r. trimble, i.s. young, j. chromatogr. b, 701 (1997) 120-123. [69] d. talwar, h. davidson, j. cooney, d.st. jo’reilly, clinical chemistry 46 (2000) 704-710. [70] b.a.j. van landeghem, j. puts, h.a. claessens, j. chromatogr. b, 822 (2005) 316-321. [71] d. loew, int. j. clin. pharmacol. ther. 34 (1996) 47-50. [72] a. greb, r. bitsch, int. j. clin. pharmacol. 36 (1998) 216-221. [73] t. frank, r. bitsch, j maiwald, g. stein, eur. j. clin. pharmacol. 56 (2000) 251-257. [74] r. bitsch, m. wolf, j. moller, l. heuzeroth, d. gruneklee, ann. nutr. metab. 35 (1991) 292-296. ©2014 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) manuscript doi: 10.5599/admet.3.2.182 131 admet & dmpk 3(2) (2015) 131-140; doi: 10.5599/admet.3.2.182 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific article pka-critical interpretations of solubility–ph profiles: pg-300995 and nsc-639829 case studies george butcher, 1 john comer, 1 alex avdeef 2, * 1 sirius analytical ltd., forest row, west sussex rh18 5dw, uk; 2 in-adme research, 1732 first avenue, #102, new york, ny 10128, usa. *corresponding author: e-mail: alex@in-adme.com; tel.: +1 647 678 5713 received: april 07, 2015; revised: june 19, 2015; published: july 01, 2015 abstract two weak bases, pg-300995 (anti-hiv agent) and nsc-639829 (anti-tumor agent), whose log s – ph profiles had been previously published, but whose pka values had not been reported, were analyzed using a method which can determine pka values from log s – ph data. this “sph-pka” technique, although often practiced, can result in inaccurate pka values, for a variety of reasons. the operational sph-pka values were compared to those predicted by marvinsketch (chemaxon), admet predictor (simulation plus), and acd/percepta (acd/labs). the agreement for the sparingly-soluble pg-300995 was reasonably good. however, a substantial difference was found for the practically-insoluble nsc639829. to probe this further, the pka of nsc-639829 was measured by an independent spectrophotometric cosolvent technique. the log s ph profile of nsc-639829 was then re-analyzed with the independently-measured pka. it was found that the equilibrium model which best fit the solubility data is consistent with the presence of a monocationic nsc-639829 dimeric species below ph 4. this illustrates that an independently-determined accurate pka is critical to mechanistic interpretations of solubility-ph data. apparently, the henderson-hasselbalch equation holds for pg300995, but not nsc-639829. keywords sparingly-soluble; solubility as a function of ph; solubility equations; shake-flask solubility; aggregation introduction it is possible to determine the ionization constant (pka) of an ionizable molecule from its solubilityph profile (log s ph), provided the curve is accurately predicted by the well-known hendersonhasselbalch equation [1] (cf., appendix). this simple pka-from-solubility (sph-pka) method has been popularized by zimmermann [2], and has been practiced often, as noted in the critical pka compilations by prankerd [3]. however, the sph-pka method can be significantly inaccurate when sample molecules in saturated solutions react with buffer components to form water-soluble complexes, or with each other to form water-soluble aggregates/oligomers (dimers, trimers, … [4]) or micelles with large aggregation numbers [5]. ionizable molecules that are surface-active (e.g., long-chain acylcarnitines [5], prostaglandins [6]) or that have strong acid-base hydrogen potentials (e.g., oxytetracycline [7], cefadroxyl [8]) have a tendency to form self-aggregates and/or micelles in saturated aqueous solutions http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:alex@in-adme.com butcher, comer, avdeef admet & dmpk 3(2) (2015) 131-140 132 [9,10], especially under the high concentrations used to characterize drug-salt precipitates [11,12]. these complications are not taken into account in sph-pka determinations. reliable and widely-available pka methods have been developed over the last two decades to measure the pka of practically-insoluble molecules, under conditions where saturation is avoided and self-aggregation is minimized [13,14]. potentiometric and uv methods are particularly well-suited for measuring pka values. the yasuda-shedlovsky [15] and origin-shifted yasuda-shedlovsky [9] methods employing cosolvents have been amply demonstrated to determine accurate pkas of molecules, some with solubility as low as 2 pg/ml (amiodarone[11,12]). in this study we considered two weak bases (fig. 1), pg-300995 (anti-hiv agent) and nsc-639829 (anti-tumor agent), whose log s – ph profiles have been published by ran et al. [16] and jain et al. [17], respectively. figure 1. structures of the molecules considered. in these cases, the pka values of the molecules had not been reported. the undistorted shapes [9] of the two log s-ph profiles suggested that the simple weak-base henderson-hasselbalch equation could be used to describe the solubility-ph relationship. the data were first analyzed by the sph-pka method, using the computer program pdisol-x (in-adme research) [10,12,18]. these operational pka values were compared to those predicted using marvinsketch (chemaxon), admet predictor (simulation plus), and acd/percepta (acd/labs). with pg-300995 (intrinsic solubility, s0 = 51 μg/ml [16]), the agreement was reasonably good with two of the three prediction programs. however, with the practically-insoluble nsc-639829 (s0 ≈ 30 ng/ml [17]), there was a substantial differences between the prediction and the value determined by the sph-pka method. to address the apparent inconsistency, it was decided to measure the pka of nsc-639829 using dedicated instrumentation which can determine pka values to high precision. using the measured pka, the log s ph profile of nsc-639829 was rationalized with a monocationic dimeric species, predominating below ph 4. (we could not obtain samples of pg-300995 to do the experimental pka determination.) experimental materials nsc-639829 (>99 % purity) was synthesized by the national cancer institute and was used as received in the jain et al. [17] study. in the ran et al. [16] study, pg-300995 was a gift from proctor and gamble. methanol was purchased from sigma-aldrich, st. louis, mo, usa. solutions and solvent mixtures were made with distilled water purchased from emd millipore, billerica, ma, usa. analytical grade potassium hydrogen phthalate and kcl were purchased from acros organics, new jersey, usa. the ionic strength of water was adjusted to 0.15 m with kcl. the admet & dmpk 3(2) (2015) 131-140 pka-critical interpretations of solubility-ph profiles doi: 10.5599/admet.3.2.182 133 base titrant was prepared by diluting co2-free koh concentrate (fluka analytical, st. louis, mo, usa) to 0.5 m. it was standardized by titration against potassium hydrogen phthalate. the acid titrant, standardized 0.5 m hcl, was purchased from sigma-aldrich st. louis, mo, usa. legacy solubility data the ran et al. [16] and jain et al. [17] shake-flask solubility protocols were very similar. briefly, excess solid was added to 1-2 ml of buffer solution. enough nacl had been added to the buffers to produce a total ionic strength, i = 0.2 m. the suspensions were agitated mildly for 5 d at 23 °c for nsc639829 and 7-10 d at 25 °c for pg-300995. at the end of the stirring period, ph was read, after which the solutions were filtered and the concentrations were determined by hplc. mostly 10 mm buffers were used: glycine/hcl for ph < 3, citrate for ph 3-5, phosphate for ph 5-8, and glycine/naoh for ph > 8. the data used in this study were digitized from the log s – ph plots in the original publications (11 and 13 ph points for pg-300995 and nsc-639829, respectively). spectroscopic measurement of the pka of nsc-639829 the pka was determined at 25 ± 0.1 °c using the siriust3 uv spectroscopic (uv-metric) method. the sample was added to a siriust3 vial as 5 microliters of 10 mm dmso stock solution and was initially titrated in a fast-uv pka screening assay. the sample was titrated from ph 2 12 at concentrations of 32 16 µm under methanol-water co-solvent conditions. only one pka, with a value of 3.7, was observed. consequently, 5 microliters of a 10 mm dmso stock solution of nsc-639829 were dispensed into a siriust3 vial and the sample was titrated in a triple titration from ph 1.5 5.0 at concentrations of 29 14 µm in three different ratios of methanol-water cosolvent. the average methanol concentration was 40.9, 30.2 and 21.6 % w/w in the first, second and third titrations, respectively. the presence of cosolvent caused the pka to shift from its aqueous value. the apparent pka in the presence of cosolvent is referred to as the pska. multi-wavelength uv spectra (200 750 nm) and ph data were collected every 0.2 ph units throughout each titration. the siriust3 refine software used target factor analysis (tfa) to rationalize the 3d matrix of absorbance vs wavelength vs ph. the pska values for each titration were determined as the point at which the rate of uv absorbance change was greatest across the selected wavelength range as a function of ph. the aqueous pka was determined by yasuda-shedlovsky extrapolation of the pska values from each titration. there was no evidence of significant sample impurities in the pka determination. predicted pka values of pka were predicted using marvinsketch tm 5.3.7 (chemaxon ltd., budapest, hungary; www.chemaxon.com), and corroborated with admet predictor tm 7.0 (simulation plus, inc., lancaster, ca, usa; www.simulations-plus.com), and acd/percepta tm 14 (acd/labs, toronto, canada; www.acdlabs.com). the predictions were used as a guide for some of the sph-pka analysis. refinement of intrinsic solubility, aggregation constants, and sph-pka detail of the mathematical approach in the pdisol-x (in-adme research) computer program has been described by völgyi et al. [18]. briefly, the data analysis method uses measured log s ph, along with standard deviations, sd (log s), as input into the pdisol-x program. an algorithm was developed which considers the contribution of all species proposed to be present in solution, including all buffer http://www.chemaxon.com/ http://www.simulations-plus.com/ http://www.acdlabs.com/ butcher, comer, avdeef admet & dmpk 3(2) (2015) 131-140 134 components (e.g., citrate, phosphate, glycine). the approach does not depend on any explicitly derived extensions of the henderson-hasselbalch equations. the computational algorithm derives its own implicit equations internally, given any practical number of equilibria and estimated constants, which are subsequently refined by weighted nonlinear least-squares regression [9,18]. therefore, in principal, drug-salt precipitates, -aggregates, -complexes, bile salts, -surfactant can be accommodated [4,9,18]. presence of specific buffer-drug species can be tested. the program assumes an initial condition for the suspension of the solid drug in the buffer solution, ideally with the compound remaining saturated over a wide range of ph. first, the program calculates the volume of acid titrant that would lower the ph of the suspension to ~0. from there, a sequence of perturbations with standardized naoh is simulated, and solubility calculated at each point (in ph steps of 0.005-0.2), until ph ~ 13 is reached. the ionic strength is rigorously calculated at each step, and pka values (as well as solubility products, aggregation and complexation constants) are accordingly adjusted [9]. also, the ph electrode parameters are adjusted for the changing ionic strength [9]. at the end of the ph-speciation simulation, the calculated log s vs. ph curve is compared to measured log s vs. ph. a log s-weighted nonlinear least squares refinement commences to refine the proposed equilibrium model, using analytical expressions for the differential equations. the process is repeated until the differences between calculated and measured log s values reach a stable minimum, as described elsewhere in detail [9,18]. results and discussion the aqueous pka value was determined by yasuda-shedlovsky extrapolation to be 3.76 ± 0.03. figure 2a shows the yasuda-shedlovsky plot of pska + log [h2o] vs. 1000/dielectric constant. the spectra measured during titration in 40.9 % methanol are shown in figure 2b. each line represents a spectrum measured at a particular ph between 1.5 and 5.2. figure 2. (a) yasuda-shedlovsky extrapolation of pska (+ log [h2o]) at three ratios of methanol to water. aqueous pka determined from intercept with vertical red line, equivalent to 1000/dielectric constant for water, minus the log [h2o]. (b) spectra measured during titration in 40.9% methanol. each line represents spectrum measured at a different ph between 1.5 and 5.2. the pka results and those of the re-analysis the ran et al. [16] and jain et al. [17] solubility-ph data are summarized in table 1. (a) nsc-637829 y-s extrapolation 14 16 18 20 1000 / dielectric constant 3 4 5 p s k a + l o g [ h 2 o ] (b) nsc-637829 spectra 300 400 wavelength (nm) 0.0 0.2 0.4 a b s o rb a n c ey = -0.386x + 10.42 admet & dmpk 3(2) (2015) 131-140 pka-critical interpretations of solubility-ph profiles doi: 10.5599/admet.3.2.182 135 table 1. summary of the results of the re-analysis of the pg-300995 [16] and nsc-639829 [17] data. a pka values predicted by marvinsketch (chemaxon). underlined values refers to acidic group ionization; otherwise, the value is that of basic group ionization. b refined pka by the sph-pka method. c measured pka by uv cosolvent extrapolation from 21.6-40.9 % w/w ch3oh-h2o (this work). d s0 = intrinsic solubility (this work). e gof = goodness-of-fit [9,18]. f not determined. g acidic pka not found experimentally below ph 12 in fig.3b. also, preliminary uv pka assay from ph 2 – 12 indicated no additional pkas within the range. pg-300995 pka and s0 the two marvinsketch-predicted pkas for pg-300995 are 3.52 (benzimidazole –n= basic functionality) and 10.58 (–nhacidic group). the corresponding amine pka from admet predictor (average over 2 tautomers) was 3.60, in close agreement with that of marvinsketch. however, the value from acd/percepta, 5.30, was less concordant. the refined sph-pka is 3.61 ± 0.06, indicated by the ph in the bend in the log s ph curve in figure 3a, agrees very well with the predictions from two of the commercial programs. there is no indication of the acidic ionization in the solubility profile, suggesting that the second pka must be greater than 9, which is consistent with the predicted value from all three pka prediction programs. since the molecule is relatively simple and not too insoluble, and since the predicted basic pka agrees with the measured value, the refined sph-pka tentatively was taken to be a measure of the true pka. thus, the henderson-hasselbalch equation is thought to be a valid description of the solubility-ph curve. this is, of course, a tentative assignment, since the independently measured value of the pka of pg-300995 is not available. the refined intrinsic solubility, s0 = 51.3 ± 0.7 μg/ml, agrees well with the reported value [16]. each point in figure 3a was assigned the standard deviation of 0.1 (not reported in the original publication). the overall goodness-of-fit, gof = 0.77, suggests that the points are scattered by 0.077 log units about the best-fit curve. nsc-639829 pka the two marvinsketch-predicted pkas for nsc-639829 were far from what was indicated in figure 3b. it is quite clear that there is no acidic ionization corresponding to the urea nh group below ph 12. this was also the finding of the spectrophotometric pka determination. the dimethylaniline amine sphpka and the predicted value (table 1) were different by 2.4 log units. with the adage “prediction guides but experiment decides,” it was decided to measure the pka(s) of nsc-639829 by an independent method. one pka was evident by the spectrophotometric method, and its value was different from both the predicted and the refined sph-pka values (table 1). in depth analyses of many log s – ph profiles [10, 12, 18] show that pka values determined by modern purpose-built pka instrumentation [13, 14] can differ substantially from sph-pka values, suggesting that the simple henderson-hasselbalch equation may not always be an accurate predictor of the ph dependence of solubility. this is particularly evident in examples such as prostaglandin f2α [6]. we consequently pursued possible explanations for the difference between sph-pka = 4.70 ± 0.12 and the uv-measured pka = 3.76 ± 0.03. butcher, comer, avdeef admet & dmpk 3(2) (2015) 131-140 136 figure 3. (a) solubility profiles of (a) pg-300995 and (b) nsc-639829. the diagonal region has slope of -1. the ph in the bend, between the slope -1 and slope 0 is the apparent pka, which may or may not be the true pka. a step-by-step solubility model construction was described recently by avdeef [10]. knowing the accurate pka starts the process. the shape of the log s – ph curve is compared against a series of templates [4,9,10]. for the profile in figure 3b, the key characteristics are: pka app > pka and slope = -1 in the diagonal region, suggesting that a mixed-charge dimer (or higher order oligomer) needs to be considered (case 3b [4,9,10,12]). the log s0 can be estimated as the solubility in the ph >> pka region of the curve. three equilibria are needed to describe such a model: h + +b ↔ bh + (pka), b ↔ b(s) (1/s0), and bh + + b ↔ bhb + (k2). having initial estimates of constants corresponding to proposed equilibrium reactions, it is possible to refine the model by weighted nonlinear regression, a procedure described elsewhere [18]. the iterative refinement process continued to convergence. figure 4 shows the refined results for nsc-639829. the solid (red) line is the best fit to the log s data at various ph values. the dashed line is calculated by the henderson-hasselbalch equation, incorporating the spectrophotometrically-measured pka. figure 4. solubility profile of nsc-639829 incorporating the uv-determined pka in the equilibrium model. the dashed line is calculated using the henderson-hasselbalch equation. admet & dmpk 3(2) (2015) 131-140 pka-critical interpretations of solubility-ph profiles doi: 10.5599/admet.3.2.182 137 figure 5 shows the distribution of various species as a function of ph. at ph = sph-pka (4.70) in a suspension containing 1 mg/ml added drug, the concentration of the free base, [b] = 5.79 x 10 -8 m (= s0, the intrinsic solubility), accounts for 66 % of the total aqueous concentration of nsc-639829. at this ph, [bh + ] accounts for 8 % of the total aqueous concentration of the drug, while [bhb + ] accounts for the remaining 26 %. the molecule is 34 % ionized at this ph. at ph = pka (3.76) in a 1 mg/ml suspension, the concentration of the free base remains the same, but it accounts for only 18 % of the total aqueous concentration of nsc-639829. at this ph, [bh + ] also accounts for 18 % of the total ([b] = [bh + ] when ph = pka), while [bhb + ] accounts for the remaining 64 %. the molecule is 82 % ionized at this ph. according to the equilibrium model, the ratio [bhb + ]/[bh + ] = 3.4 is expected to remain unchanged as long as the solutions remain saturated (as suggested by the two thick diagonal lines of identical slope -1 in figure 5), provided there are no other unaccounted equilibria when the positively charged species become less concentrated than the uncharged free base (ph > 4.7). figure 5. speciation profile of nsc-639829. the thick solid lines correspond to the concentrations of the drug species. the thin lines correspond to the concentrations of the three buffer components used to simulate the data: glycine, phosphate, and citrate. the 1 mg/ml used in the simulation suggests that all of the solid dissolves when ph < 0.5. conclusions we re-analyzed the previously published solubility-ph data of two weak base drugs. it was tentatively proposed, based on the agreement between the apparent pka (sph-pka) and that predicted by marvinsketch and admet predictor (but not acd/percepta), that the solubility profile of the more soluble drug, pg-300995, could be adequately predicted by the henderson-hasselbalch (hh) equation. however, the practically-insoluble nsc-639829 drug could not be predicted by the simple hh equation. its pka was determined here. the inclusion of the independently-measured pka in the equilibrium model butcher, comer, avdeef admet & dmpk 3(2) (2015) 131-140 138 suggested that the “anomalous” profile of nsc-639829 can be explained by presence of a mixed-charge cationic dimer (bhb + ). this illustrates that an independently-determined accurate pka is critical to interpreting solubility-ph data of ionizable compounds. however, predicted pka values can be very helpful guides in the initial stages of such investigations. apparently, the simple henderson-hasselbalch equation holds for pg-300995, but not nsc-639829. acknowledgements the authors wish to thank professor samuel yalkowsky (univ. of arizona) for kindly providing a sample of nsc-639829. we are also grateful for very stimulating discussions with drs. robert fraczkiewicz (simulations plus), andreas klamt (cosmologic), and jozsef szegezdi (chemaxon), regarding the possible effect of internally-stabilized hydrogen bonding in nsc-639829 on the accurate prediction of the pka(s). also, robert fraczkiewicz was kind to share the prediction of the pka of pg300995, using admet predictor. references [1] e. baka, j.e.a. comer, k. takács-novák, j. pharm. biomed. anal. 46 (2008) 335-341. [2] i. zimmermann, int. j. pharm. 13 (1983) 57-65. [3] r. j. prankerd. critical compilation of pka values for pharmaceutical substances. in: h. brittain (ed.). profiles of drug substances, excipients, and related methodology. vol. 33. elsevier, new york, 2007. [4] a. avdeef, adv. drug deliv. rev. 59 (2007) 568-590. [5] s.h. yalkowsky, g. zografi, j. colloid inter. sci. 34 (1970) 525-533. [6] t.j. roseman, s.h. yalkowsky, j. pharm. sci. 62 (1973) 1680-1685. [7] t. higuchi, m. gupta, l.w. busse, j. am. pharm. assoc. 42 (1953) 157-161. [8] e. shoghi,e. fuguet, e. bosch, c. rafols, eur. j. pharm. sci. 48 (2012) 290-300. [9] a. avdeef. absorption and drug development second edition, wiley-interscience, hoboken, nj, 2012, pp. 251-318. [10] a. avdeef, admet & dmpk 2(1) (2014) 33-42. [11] c.a.s. bergström, k. luthman, p. artursson, eur. j. pharm. sci. 22 (2004) 387-398. [12] a. avdeef, admet & dmpk 2(1) (2014) 43-55. [13] j. comer, k. box, j. assoc. lab. automat. 8 (2003) 55-59. [14] n. sun, a. avdeef, j. pharm. biomed. anal. 56 (2011) 173-182. [15] k.j. box, g. völgyi, r. ruiz, j.e. comer, k. takács-novák, e. bosch, c. ràfols, m. rosés, helv. chim. acta 90 (2007) 1538-1553. [16] y. ran, a. jain, s.h. yalkowsky, j. pharm. sci. 94 (2005) 297-303. [17] n. jain, g. yang, s.e. tabibi, s.h. yalkowsky, int. j. pharm. 225 (2001) 41-47. [18] g. völgyi, a. marosi, k. takács-novák, a. avdeef, admet & dmpk 1(4) (2013) 48-62. admet & dmpk 3(2) (2015) 131-140 pka-critical interpretations of solubility-ph profiles doi: 10.5599/admet.3.2.182 139 appendix – derivation of the solubility ph equations in the case of a monoprotic weak base, a saturated solution can be defined by the equations and the corresponding constants bh +  h + + b ka = [h + ][b] / [bh + ] (a1) b(s)  b s0 = [b] (a2) solubility, s, at a particular ph is defined as the mass balance sum of the concentrations of all of the species dissolved in the aqueous phase: s = [b] + [bh + ] (a3) where the square brackets denote molar concentration of species. the above equation can be transformed into an expression containing only constants and [h + ] (as the only variable), by substituting the ionization and solubility eqs. (a1) and (a2) into eq. (a3). log s = log ( [b] + [h + ][b] /ka ) = log [b] + log ( 1 + [h + ] / ka ) = log s0 + log ( 1 + 10 +pka – ph ) (a4) eq. (a4) is usually called the henderson-hasselbalch equation for a monoprotic weak base, and describes a hyperbolic-shaped log s – ph curve. at the bend in the log s – ph curve, the ph equals the pka. it can be hypothesized that the dimeric mixed-charge weak base species, (b.bh + ), also forms, which contains a 1:1 ratio of b and bh + . an additional equilibrium equation needs to be added to the mass balance. b + bh +  (b.bh + ) k2 = [b.bh + ] / [b][bh + ] (a5) eq. (a3) needs to be expanded accordingly. s = [b] + [bh + ] + 2 [b.bh + ] (a6) in logarithmic form, log s = log ( [b] + [h + ][b] /ka + 2 k2[b][bh + ]) = log ( [b] + [h + ][b] /ka + 2 k2[b] 2 [h + ] /ka ) = log s0 + log ( 1 + { 1+2 k2 s0 } 10 +pka – ph ) (a7) for nsc-639829, the factor in the braces in eq. (a7) equals 8.60 (=1 + 2 x 10 +7.80 x 10 -7.22 ). when the logarithm of the factor (0.94) is taken into the exponent, the resulting equation appears to equal eq. (a4), except that the pka is replaced with the sph-pka value (3.76+0.94 = 4.70). the hypothesized presence of the mixed-charge dimer appears to shift the original henderson-hasselbalch equation in the positive ph direction by nearly a log unit. the presence of self-aggregated species of other stoichiometries can distort the shape of the simple henderson-hasselbalch equation in a number of different ways [4,9,10,12,18]. the example in the present study is one of the simpler types of distortion. with increasing ph, it is possible for the protonated dimer, b.bh + , to lose the ionizable proton to become the water-soluble neutral-species dimer, b2. the pka for such a process would not be expected to be exactly the same as the measured pka. for example, ketoprofen bound to sodium taurocholate butcher, comer, avdeef admet & dmpk 3(2) (2015) 131-140 140 micelles shows such a secondary pka [9]. there was no hint of such a process in the case of nsc-639829, but a more thorough investigation would require additional log s-ph measurements with total concentration of the sample varied over a wide range, since aggregates would be expected to have concentration dependence. ©2015 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ realistic and critical review of the state of systemic anti­microbial peptides doi: http://dx.doi.org/10.5599/admet.1215 91 admet & dmpk 10(2) (2022) 91-105; doi: https://doi.org/10.5599/admet.1215 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index review realistic and critical review of the state of systemic antimicrobial peptides guangshun wang*, abraham fikru mechesso department of pathology and microbiology, college of medicine, university of nebraska medical center, 985900 nebraska medical center, omaha, ne 68198-5900, usa *corresponding author email: gwang@unmc.edu received: december 17, 2021; revised: january 14, 2022; january 20, 2022 abstract antimicrobial peptide research remains active not only because of the growing antibiotic resistance problem but also our desire to understand the role of innate immune peptides in host defense. while numerous peptides are currently under active development for topical use, this article highlights peptides with systemic efficacy. the scaffolds of these peptides range from linear to cyclic forms. the neutropenic mouse model is well established to illustrate antimicrobial efficacy from direct killing. the majority of tests, however, are conducted using normal mice so that both direct antimicrobial and immune regulatory effects can be characterized. these systemic examples underscore the possibility of adding new candidates to the list of the existing peptide antibiotics to more effectively combat antibiotic-resistant bacteria, fungi, and parasites. ©2021 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords antibiotics; antimicrobial peptides; peptide design; mic; pk; pd; systemic efficacy; toxicity; intravenous injection 1. introduction the current sars-cov-2 pandemic is a reminder to human beings that pathogenic microbes can have a tremendous impact on every aspect of our society. in addition to viruses, antibiotic-resistant bacteria, fungi and parasites constitute a new potential threat. it is projected that 10 million people could die from untreatable infections by 2050 [1]. consequently, it becomes urgent to search for novel antimicrobials. innate immune antimicrobial peptides (amps) play a critical role in protecting the hosts from infection [2,3]. their lasting efficacy rejuvenates our interest in developing them into novel antibiotics. indeed, some peptide antibiotics such as gramicidin, vancomycin, daptomycin and colistin are already in clinical use [4]. according to the updated antimicrobial peptide database (https://aps.unmc.edu), over 3000 such peptides have been discovered and characterized from natural sources, ranging from bacteria to animals [5,6]. many amps are cationic and possess broad-spectrum antimicrobial activity against both gram-positive and gramnegative bacteria. some peptides, however, eliminate only gram-positive or gram-negative bacteria. frequently, amps consist of both basic and hydrophobic amino acids. while basic amino acids initiate http://dx.doi.org/10.5599/admet.1215 https://doi.org/10.5599/admet.1215 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:gwang@unmc.edu http://creativecommons.org/licenses/by/4.0/ https://aps.unmc.edu/ g. wang and a.f. mechesso admet & dmpk 10(2) (2022) 91-105 92 anionic bacterial recognition, hydrophobic amino acids are important for subsequent peptide anchoring into membranes. this is accompanied by a structural transition from random coils to an amphipathic αhelical structure [2,3,7]. in addition to the α family, there are three other types of structural families: β, αβ, and non-αβ [8]. representative structures for these four families can be viewed on the face page of the amp database website above. some amps have a folded structure even in an aqueous solution. one molecular mechanism results from the forced folding of disulfide bonds. as a consequence, α-defensins are known to adopt β-sheet structures. beta-defensins, longer than α-defensins, can incorporate an α-helical element to generate a mixed αβ fold. like helical peptides, defensins have an effect on membranes as well. some defensins, however, work by inhibiting cell wall synthesis [9,10]. there are also amps that form neither a regular α-helix nor a β-sheet structure. proline-rich peptides are such examples that inhibit bacterial growth by binding to ribosomes [11]. the structural diversity of amps offers numerous r&d opportunities for developing novel peptide antibiotics. this minireview highlights amps with the potential to treat systemic infections. we discuss peptide toxicity, in vivo efficacy, and pharmacokinetic (pk) and pharmacodynamic (pd) studies. these examples support the potential of engineered amps as novel systemic antibiotics. 2. peptide design: antimicrobial robustness, selectivity and stability the basic goal of peptide design is to identify potent peptides with minimal or no toxic effects on normal cells. it is evident that basic and hydrophobic amino acids are the two key elements in designing amphipathic antimicrobial peptides [7,12,13]. our current knowledge implies that basic amino acids are key to the inhibition and killing of gram-negative bacteria, whereas hydrophobic amino acids are critical to eliminating gram-positive bacteria such as methicillin-resistant staphylococcus aureus (mrsa) [14]. although reducing peptide hydrophobicity is a general strategy to improve peptide cell selectivity [15], basic amino acids also play a role [16]. hence, the finally designed peptide achieves an optimal combination of hydrophobic and basic amino acids. selected peptides discussed in this review are listed in table 1. these peptides were obtained by modifying natural peptides or designed based on the key elements of amps. a3apo, onc72, onc112, and api137 are derived from proline-rich peptides originally discovered in insects [1719]. gomesin is a short β-hairpin peptide isolated from scorpions [20]. ampr-11 is a peptide derived from mitochondrial non-selective channel romo-1 [21]. c10-kr8d is a lipopeptide designed based on the minimal antibacterial peptide kr-12 of human cathelicidin ll-37 [22], whereas oh-cath30 is a truncated fragment of the king cobra cathelicidin oh-cath [23]. some peptides are de novo designed. wlbu2 is a synthetic peptide designed using valine, tryptophan, and arginine [24]. dft503 and dft561 [25] are derivatives of the dftamp1, which is the first database filtering technology designed peptide obtained by extracting the key parameters (peptide length, charge, hydrophobicity, and structure) from the antimicrobial peptide database [26]. horine and verine were designed by combining database filtering with structural knowledge [27]. another important goal of peptide design is to improve peptide stability to proteases. several strategies are utilized to enhance peptide stability: terminal capping, incorporation of non-standard amino acids, and peptide cyclization [28]. for example, the incorporation of d-amino acids increased the half-life (t1/2) of horine from ~10 min (l-form) to 60 min (d-form) in vivo [27]. in the case of c10-kr8, the d-form is preferred over the l-form since the d-from did not bind as many serum components as the l-form based on mass spectrometric studies [22]. there are also other methods to improve peptide stability. while peptide mimicries may be made to eliminate susceptible peptide bonds entirely [29-31], formulations can protect peptides from protease degradation [32,33]. admet & dmpk 10(2) (2022) 91-105 antimicrobial peptides with systemic efficacy doi: http://dx.doi.org/10.5599/admet.1215 93 subsequently, after peptide optimization, it is necessary to test peptide antimicrobial robustness. this is because amps such as ll-37 are known to lose activity in the presence of physiological salts, serum, and under acidic conditions [22,24,34]. it is anticipated that the peptides are more likely to display in vivo efficacy if they can retain antimicrobial activity under such conditions. to identify candidates with such properties, we assembled a pipeline of in vitro filters by measuring peptide mic values in rich media supplemented with salts, serum, and at an acidic ph. we found that very few peptides retain antimicrobial activity under all these conditions. it was encouraging that our database-designed peptides that crossed these barriers did show systemic efficacy in vivo [22,27]. 3. in vivo toxicity of designed antimicrobial peptides nephrotoxicity of colistin is the major side effect of this antibiotic [35,36]. therefore, there is a desire to develop the next generation of peptide antibiotics with minimized toxicity. to evaluate the peptide toxicity in vivo, the commonly used routes for peptide administration are intraperitoneal (i.p.), intravenous (i.v.), intramuscular (i.m.), and subcutaneous (s.c.) injections (table 1). of significant interest is the peptide toxicity after intravenous injection. for a3-apo, mice survived with severe transient side effects at 25 mg/kg (i.p.), but a single dose of 50 mg/kg was lethal, indicating a maximum tolerated dose (mtd) was 20 mg/kg. it is equal to the no observed adverse effect limit (noael) in this case. in the same model, the 50 % lethal dose (ld50) of colistin was found to be less than 40 mg/kg [17]. the ld50 of oh-cath30 after i.p. injection was 120 mg/kg and mice survived when subcutaneously injected at 160 mg/kg [23]. the wlbu2 peptide was found to have an mtd of 12 mg/kg (i.v.) [24]. this is fourfold higher than the treatment dose at 3 mg/kg. four alternating d,l-cyclic peptides (6752, 6756, 6853 and 7251) displayed mtds in the range of 20-25 mg/kg or above depending on the amino acid sequence [37]. these concentrations allowed a dosedependent treatment at 16 mg/kg or below. likewise, teixobactin showed no sign of toxicity when administered intravenously at a single dose of 20 mg/kg [38]. a single i.v. injection of ampr-11 (100 mg/kg), tenfold higher than the treatment dose into mice, did not show severe clinical signs of toxicity for 15 days [21]. after one week of daily i.p. injection at the treatment dose (10 mg/kg), horine showed no side effects on body weight, animal behavior, or kidney histology [27]. although a slight body weight loss (12.1 %) was noted after daily i.v. injection (10 mg/kg) for a week, we did not observe any other side effects. these results are encouraging and will lead to the discovery of novel peptides with further reduced toxicity and enhanced potency. 4. systemic efficacy of antimicrobial peptides while most candidates are developed for topical use, some laboratories have observed systemic efficacy for designed peptides (table 1). both neutropenic and normal mice are utilized for systemic efficacy evaluation. neutropenic mice are a classic model to establish direct antimicrobial efficacy and for studying pk/pd of antibiotics since the immune effects are suppressed [39,40]. it is important to establish that bacteria had spread to different organs prior to antimicrobial treatment. after i.p. injection, szabo and colleagues [17] found a 30-minute delay was sufficient for e. coli to spread all over the body of mice. mor and colleagues observed the spread of escherichia coli in neutropenic mice one-hour post i.p. infection [31]. we also detected mrsa or klebsiella pneumoniae in different murine organs two hours post-infection [25,27]. selsted and colleagues also found the dissemination of k. pneumoniae to various organs one-hour post i.p. infection [41]. the peptide efficacy is often reflected by a dose-dependent colony-forming unit (cfu) drop as well as the survival advantage of mice in the treated group relative to the untreated infected group. using the neutropenic thigh model, dartois et al. [37] made alternating d,l-cyclic peptides and observed a decrease of the s. aureus load by 2.1-3.0 log units after i.v. treatment (8 mg/kg) two hours post http://dx.doi.org/10.5599/admet.1215 g. wang and a.f. mechesso admet & dmpk 10(2) (2022) 91-105 94 i.m. infection. although effective in vitro, two such peptides (1316 and 1150) failed to show in vivo activity, probably due to low bioavailability. using the same model, li et al. [23] tested the efficacy of oh-cath30 immediately after thigh infection at 1, 10, and 20 mg/kg via a single s.c. treatment. in the protection experiment, the survival rate of mice decreased from 100 % to 57-70 % when the treatment time was delayed from one hour to four hours post i.p. infection (e. coli at 10 8 cfu). using cd-1 neutropenic mice infected with mrsa intraperitoneally, teixobactin (mic 0.25 µg/ml against mrsa) shows dose-dependent efficacy (i.e., cfu decrease) after a single dose i.v. treatment two hours post-infection [38]. mor observed systemic efficacy of the oak peptide mimics in neutropenic mice but not two amphibian peptide derivatives [31]. likewise, we did not observe systemic efficacy in this model for a human cathelicidin ll-37 derived peptide [25]. one possibility is that the peptide associates with other molecules in mice and is not available to attack bacteria. our reasoning is in line with the molecular promiscuity of ll37 (binding to numerous molecular partners such as human serum proteins and nucleic acids) [22,24,34]. similarly, indolicidin did not show a protective effect when treated in the free form. after formulation, however, it protected mice by 30 % [32]. these results indicate that not all amps are able to display systemic efficacy in mice in the free form. taking this into consideration, we then aimed at identifying more robust peptides that retain activity under various conditions by expanding the database filtering technology from in silico to in vitro. our database-designed peptides are particularly effective against gram-positive pathogens such as mrsa and vancomycin-resistant enterococci (vre). in addition, we discovered that an increase in cationic amino acids led to an increase in toxicity and a loss of in vivo activity of the dftdesigned peptides in neutropenic mice [25]. in keeping with this low cationicity idea, we have designed even shorter peptides such as horine and verine to reduce production costs [27]. these two peptides share a similar amino acid composition but possess different amino acid sequences. they also have different 3d structures and activity spectra: horine mainly against gram-positive mrsa and verine against both grampositive and negative pathogens. during a five-day observation, a single dose injection of horine at 10 mg/kg (i.p.) protects 87.5 % mice infected with mrsa identical to vancomycin, whereas 81.8 % mice infected with antibiotic-resistant k. pneumoniae e406-17 survived after a single dose of verine treatment (15 mg/kg) 2 h post infection. this protective effect of verine was better than doripenem (50 % protection) at the same dose. likewise, we observed a significant decrease in bacterial burden (1-4 logs) in a variety of murine organs, including the spleen, lung, kidney, and liver, when treated either intraperitoneally or intravenously [27]. these results document direct antimicrobial efficacy for amps. other laboratories have primarily used normal (non-neutropenic) mice to test peptide efficacy. in such a model, the peptide may achieve efficacy by direct killing and/or immune boosting. usually, a high bacterial inoculum is required for infection (e.g., ~10 8 cfu for mrsa and ~10 7 cfu for p. aeruginosa). deslouches et al. [24] observed a systemic effect for wlbu2 treated i.v. at 3 mg/kg 30 to 45 min after p. aeruginosa i.p. infection. no bacteria were recovered after 7 to 10 days from tissues or blood. several peptides derived from proline-rich amps show systemic efficacy in mice infected intraperitoneally. examples include a3-apo, onc72, onc112, and api137 (table 1). the designer peptide, a3-apo, is effective in multiple models to protect mice from e. coli neumann infection in a dose-dependent manner. at 10 mg/kg, it was as efficacious as imipenem in the long term. however, s.c. treatment was found to be ineffective against e. coli. the i.p. efficacy is proposed to result from both direct killing and immune-stimulatory effects [17]. onc72 and onc112 (2.5 or 5 mg/kg injected i.p. at 1, 4, and 8 h post-infection) were able to protect mice in a septicaemia model due to e. coli atcc 25922 infection [18]. api137 (an apidaecin derived peptide) showed a dose-dependent protection of the cd-1 mice (67 % survival over 5 days) from e. coli atcc 25922 infection when treated subcutaneously but not intravenously [42]. although the minimal bactericidal admet & dmpk 10(2) (2022) 91-105 antimicrobial peptides with systemic efficacy doi: http://dx.doi.org/10.5599/admet.1215 95 concentrations (mbc) of ampr-11 are moderate (~50 µm), it increases mouse survival rate by 60 % after a single dose (10 mg/kg, i.v.) treatment one-hour post-infection with s. aureus, pseudomonas aeruginosa, k. pneumoniae, or acinetobacter baumannii [21]. this efficacy is comparable to that achieved by imipenem injected four times (every 12 h, i.p.) 24 h post-infection. it is proposed that the observed efficacy results from multiple factors, including rapid killing, a lack of binding to serum/lipoproteins, and lipopolysaccharides (lps) neutralization. although not discussed, peptide-induced immune stimulation as evidenced by lps neutralization could also be an important factor. zhang et al. [43] observed a 2-3 log cfu decrease (relative to pbs treated) in mice injected with dp7 (2 mg/kg, i.v.) one-hour post i.v. infection similar to vancomycin at 10 mg/kg. the rtd-1 peptide, a theta-defensin, shows systemic efficacy against intravenously infected candida albicans when treated i.p. daily [44]. interestingly, the treatment efficacy of rtd-1 at 5 mg/kg does not depend on administration routes. however, a further delay of treatment from 1, 3, to 6 h after infection reduces peptide efficacy. the fact that the measured peptide concentrations in mice are 5-25 fold lower than mics (>100 µg/ml) against c. albicans in the presence of 50 % serum led to the conclusion that peptide efficacy did not result from direct antifungal effect. the same conclusion has been arrived in a recent study using mtd12813, an improved version of rtd-1, to treat mice (i.p.) infected with gram-negative e. coli or k. pneumoniae (i.p.) [41]. likewise, rossi et al. [20] demonstrated the impact of gomesin (a scorpion peptide) on the host immune response against candida infection in mice. in addition to bacterial killing [23], li et al. [45] showed that oh-cath30, a snake antimicrobial peptide, also induces an innate immune response to help the host combat bacterial infection. these examples indicate that amps can selectively boost the host immune response by regulating the release of proinflammatory cytokines and chemokines and promoting the chemotaxis of immune cells [45,46]. in summary, these examples underscore the feasibility of identifying systemic amps. the systemic efficacy of these peptides could result from direct antimicrobial killing, immune stimulation, or a combination of the two mechanisms. in addition, in vivo efficacy could depend on experimental conditions. for instance, cyclic peptides were effective when administered intravenously but not orally or subcutaneously [37]. if we define the selectivity index as the ratio between the mtd and peptide treatment dose, there is a therapeutic window of up to 10 for these peptides to treat bloodstream infections (table 1) and up to 100 to treat animals intraperitoneally [41]. 5. peptide pharmacokinetics and pharmacodynamics the pk and pd data for antimicrobial peptides are very limited. dartois and colleagues [37] followed the serum concentrations of peptides 6752 and 6853 after treatment at different doses. maximum concentration (cmax) and the area under the concentration-time curve (auc) increased steadily after injection from 2 to 8 mg/kg. in contrast, peptides 1316 and 1150, which did not display in vivo efficacy, showed significantly lower cmax values, indicating the need for a sufficient amount of peptide (e.g., above mic) for in vivo bacterial elimination. the t1/2 values for these cyclic peptides ranged from 1.2 to 3.9 h. consistent with in vitro results, the in vivo anti-mrsa effect of the peptide did not depend on the susceptibility of bacteria to conventional antibiotics. the faster killing by peptide 6752 (membrane disruption) than vancomycin and oxacillin (cell wall inhibition) in vivo is consistent with their mechanisms of action. interestingly, when treated at 16 mg/kg in the thigh model, the prolonged antibacterial activity (pae) was greater than 6 h for vancomycin and >7 h for 6752. a significant amount of the peptide was cleared renally. ling et al. [38] showed that intravenous injection of a single dose of teixobactin (20 mg/kg) resulted in serum concentrations above the mic for 4 h, explaining the peptide efficacy in vivo. schmidt et al. [19] also showed that when api88 or api137 was injected i.v. or i.p. at doses of 5 and 20 mg/kg, their http://dx.doi.org/10.5599/admet.1215 g. wang and a.f. mechesso admet & dmpk 10(2) (2022) 91-105 96 plasma levels were similarly low (<3 μg/ml). comparable levels of these peptides accompanied by the same major metabolites were detected in blood, urine, kidney, and liver homogenates. api137 was rapidly degraded at the c-terminus, while api88 was rather stable. the high efficacy in murine infection models and the rapid clearance (60-90 min) of these peptides indicate that their in vivo efficacy is also related to the cmax. rossi et al. [20] found that gomesin mainly accumulated in the liver and rapidly eliminated from the circulation. however, the excretion route remained unexplored. in our recent study [27], we observed different half lives for the land d-forms of horine (t1/2 ~10 min and ~1 h) in murine blood after i.v. injection and a decrease in the amount of the peptide in blood is accompanied by an increase in other organs, including the spleen, lung, kidney, and liver. while the plasma level of d-horine remains above the mic (4 µm) for about one hour, the l-form of horine is rapidly degraded. the fact that both peptides demonstrate in vivo efficacy indicates rapid bacterial killing by horine, consistent with in vitro findings [27]. 6. concluding remarks and future directions there is a clear need for systemic antimicrobials to treat bloodstream infections caused by antibioticresistant pathogens, such as bacteria, fungi, and parasites. our establishment of a pipeline of peptide filters from in vitro to in vivo, inspired by our in silico filtering technology, presents a useful strategy for identifying additional systemic candidates to treat bloodstream infections [25]. in several cases, evidence has been obtained for the establishment of systemic bacterial infections before treatment. based on limited systemic examples summarized here, we conclude that not all amps are able to display systemic efficacy in vivo. some notable features for these systemic peptides in table 1 are (1) low hydrophobic prorich peptides, (2) low cationic database-derived peptides, and (3) cyclic peptides. also, a small size is preferred to reduce production costs. it is proposed that a compact structure can reduce potential binding to other molecules. such features might have improved the bioavailability of these peptides in vivo. in addition, the systemic effect of a peptide may depend on the type of animal models, bacteria species, bacterial inoculum, treatment time post-infection, peptide dose, and treatment frequency. it is clear in each case that a sufficient amount of antimicrobials is required to achieve in vivo efficacy. therefore, further studies are required: 1. to identify the in vitro conditions useful for identifying systemic peptides; 2. to evaluate peptide toxicity more completely by using a variety of mammalian cell lines and animal models; and 3. to characterize the pk/pd properties of the identified amp candidates since only limited data are available at present. finally, one can also learn from the peptide antibiotics already in clinical use (table 2). these include vancomycin, daptomycin, colistin, and their derivatives recently approved by fda. these peptide-based antibiotics are characterized by high efficacy and selectivity, low toxicity, low metabolic stability, and rapid renal clearance [42]. notably, most known peptide antibiotics are cyclic and differ from the classic amphipathic helical peptides. hence, we anticipate more attention to cyclic peptides in developing a new generation of future antibiotics. we also anticipate a continued search of novel peptide structures in nature by combining genetic, bioinformatic, proteomic, and structural biology approaches. it is proposed that a combination of amps with existing antibiotics can bring forth additional treatment benefits otherwise unachievable [4,47]. an improved understanding of amps in the host as well as in association with microbiota will contribute to future personal and precise medicine. admet & dmpk 10(2) (2022) 91-105 antimicrobial peptides with systemic efficacy doi: http://dx.doi.org/10.5599/admet.1215 97 table 1. antimicrobial peptides with systemic efficacy r e f (y e a r) [3 2 ] (1 9 9 5 ) [4 8 ] (2 0 0 2 ) [4 7 ] (2 0 0 4 ) [3 7 ] (2 0 0 5 ) [2 4 ] (2 0 0 5 ) [4 9 ] (2 0 1 0 ) [5 0 ] (2 0 1 0 ) p k /p d m t d 4 0 m g /k g 2 5 m g /k g n d -t h e d e cl in e i n c o n ce n tr a ti o n w a s b e st d e sc ri b e d b y a t w o -c o m p a rt m e n t m o d e l. -m o st o f th e i n je ct e d c o m p o u n d s a re cl e a re d r e n a ll y . -t h e c l, t 1 /2 β a n d v ss w e re 0 .1 9 -0 .2 6 l/ k g /h r, 1 .2 2 .2 h r, a n d 0 .3 3 -0 .4 9 l it /k g , re sp e ct iv e ly . m t d = 1 2 m g /k g n o a ct iv it y i n s e ru m a ft e r s c a d m in is tr a ti o n , re m a in e d s ta b le in t h e c ir cu la ti o n f o r > 1 h s u rv iv a l ra te 3 0 % o f in fe ct e d m ic e s u rv iv e d a ft e r tr e a tm e n t a t 4 0 m g /k g a m o d e st p ro lo n g a ti o n in su rv iv a l ti m e o f in fe ct e d m ic e a ft e r tr e a tm e n t w it h 5 m g /k g o f th e p e p ti d e 7 5 % o f in fe ct e d m ic e s u rv iv e d a ft e r b m a p -2 8 t re a tm e n t n d n d n d 3 0 % s u rv iv a l a t 2 m g /k g a n d 9 0 % a t 5 m g /k g . b a ct e ri a l c f u d ro p ~ 3 l o g 2 .9 -3 .6 l o g n o b a ct e ri a e te ct e d a ft e r 7 t o 1 0 d a y s p o st t re a tm e n t 1 l o g ~ 3 l o g t re a tm e n t ro u te iv iv iv iv iv iv ip /i m / iv in fe ct io n ro u te iv iv iv im ip iv ip /i m m ic ro o rg a n is m a sp e rg il lu s fu m ig a tu s sp o re c a n d id a a lb ic a n s s ta p h yl o co cc u s a u re u s s . a u re u s (m s s a /m r s a ) p se u d o m o n a s a e ru g in o sa p a o 1 m r s a s . a u re u s m ic e b a lb /c m ic e b a lb /c m ic e s w is s w e b st e r m ic e m ic e p e p ti d e in d o li ci d in li p o so m e p 1 9 (8 ) b m a p -2 8 c y cl ic 6 7 5 2 w lb u 2 r a n a le xi n a n d re co m b in a n t ly so st a p h in o a k s c 1 2 (ω 7 )k -β 1 2 http://dx.doi.org/10.5599/admet.1215 g. wang and a.f. mechesso admet & dmpk 10(2) (2022) 91-105 98 table 1. continued… r e f [1 7 ] (2 0 1 0 ) [5 1 ] (2 0 1 0 ) [5 2 ] (2 0 1 2 ) [2 0 ] (2 0 1 2 ) [2 3 ] (2 0 1 2 ) [3 8 ] (2 0 1 5 ) [1 8 ] (2 0 1 6 ) [5 3 ] (2 0 1 6 ) p k /p d m t d 2 5 -5 0 m g /k g t h e i m t h e ra p e u ti c in d e x o f th e p e p ti d e (> 1 2 ) a p p e a re d t o b e h ig h e r th a n t h a t o f co li st in ( a ro u n d 5 – 6 ); m t d > 6 0 m g /k g . r e ta in e d a ct iv it y a g a in st t h e p a ra si te i n b lo o d m a in ly d is tr ib u te d i n l iv e r (6 0 % ) fo r u p t o 2 4 h p o st -i n je ct io n . t 1 /2 < 6 0 m in . c o n ce n tr a ti o n a b o v e t h e m ic f o r 4 h a ft e r a s in g le d o se o f 4 0 m g /k g ; t 1 /2 , 4 .7 h ; a u c to l a st , 5 7 .8 µ g -h /m l; v o lu m e o f d is tr ib u ti o n , 4 7 m l; m t d > 2 0 m g /k g . o n c7 2 : t 1 /2 , 1 2 m in ; v o lu m e o f d is tr ib u ti o n , 3 2 m l; o n c1 1 2 : t 1 /2 , 1 9 .3 m in ; v o lu m e o f d is tr ib u ti o n , 1 3 m l. n d d e te ct e d i n b lo o d , u ri n e , a n d k id n e y , a n d li v e r h o m o g e n a te s a t si m il a r le v e ls , w it h ra p id cl e a ra n ce w it h in ∼ 9 0 m in . s u rv iv a l ra te 8 0 -1 0 0 % s u rv iv a l in ci sp la st in n o n -t re a te d m ic e , 4 0 -1 0 0 % i n c is p la st in tr e a te d m ic e . 1 0 0 % s u rv iv e d a ft e r 1 2 h a n d o n ly 1 4 .3 % s u rv iv e d a ft e r 2 4 h . n d a ll d ie d o n t h e 1 8 th d a y o f tr e a tm e n t, t h re e d a y s la te r th a n t h e p b s c o n tr o l t h e s u rv iv a l ra te : 6 0 t o 9 0 % in a s e p ti ce m ia p ro te ct io n m o d e l, 1 0 0 % s u rv iv a l d u ri n g 4 8 h w h e n t re a te d a b o v e 0 .5 m g /k g . n d n d 1 0 0 % b a ct e ri a l c f u d ro p 3 -4 l o g i n b o th ci sp la st in p re tr e a te d a n d n o n -– tr e a te d m ic e . 0 .5 -2 l o g s ig n if ic a n t p a ra si te lo a d d ro p i n l iv e r 0 .5 -1 l o g 4 -5 l o g ~ 4 l o g /t h ig h 4 -5 l o g 1 0 u n it ~ 1 l o g 1 0 u n it n d t re a tm e n t ro u te ip im /i v iv ip ip iv ip iv ip in fe ct io n ro u te ip ip iv iv ip ip ip ip ip m ic ro o rg a n is m e sc h e rs ic h ia co li a ci n e to b a ct e r b a u m a n n ii l. i n fa n tu m ch a g a si c a n d id a a lb ic a n s e . co li p . a e ru g in o sa s . a u re u s m r s a e . co li m r s a e . co li m ic e c f w -1 c d -1 b a lb /c b a lb /c m ic e c d -1 m ic e m ic e m ic e p e p ti d e a 3 -a p o d im e r a 3 -a p o r p -1 g o m e si n o h -c a t h 3 0 t e ix o b a ct in o n c1 1 2 o n c7 2 d p 7 li p o so m e a p i1 3 7 admet & dmpk 10(2) (2022) 91-105 antimicrobial peptides with systemic efficacy doi: http://dx.doi.org/10.5599/admet.1215 99 table 1. continued… r e f (y e a r) [5 4 ] (2 0 1 8 ) [4 4 ] (2 0 1 9 ) [4 2 ] (2 0 1 9 ) [2 5 ] (2 0 1 9 ) [2 5 ] (2 0 1 9 ) [5 5 ] (2 0 1 9 ) [4 3 ] (2 0 1 9 ) [2 1 ] (2 0 2 0 ) p k /p d m t d : 1 0 m g /k g f o r h y d ra zi d e a n a lo g a n d 2 5 m g /k g f o r a m id e a n a lo g . m t d > 5 0 m g /k g ; a t w o -c o m p a rt m e n t m o d e l b e st d e sc ri b e d t h e i v d a ta . s lo w ly a b so rb e d i p w it h 6 3 .4 % b io a v a il a b il it y . c m a x 2 3 m g /l ( si n g le i v 2 0 m g /k g ); lo w p la sm a l e v e ls , h ig h v o lu m e s o f d is tr ib u ti o n a n d l o w s e ru m s ta b il it y . r a p id ly e li m in a te d f ro m t h e b lo o d st re a m . n d n d n d m t d 1 0 0 m g /k g ; a ct iv it y d e cr e a se d i n p la sm a /s e ru m . t 1 /2 , ~ 5 0 m in i n p la sm a . s u rv iv a l ra te m a jo ri ty o f th e a m id e a n a lo g t re a te d m ic e su rv iv e d a s in g le d o se o f r t d -1 si g n if ic a n tl y e n h a n ce d su rv iv a l r e g a rd le ss o f th e ro u te s o f a d m in is tr a ti o n s u rv iv a l 6 7 % /i p a n d 3 3 % /i v a n d s c a d m in is tr a ti o n n d n d s u rv iv a l 6 0 % 7 0 -9 0 % r e d u ct io n s in th e le th a li ty o f in fe ct io n s u rv iv a l > 6 0 % . b a ct e ri a l c f u d ro p ~ 0 .5 -1 .5 l o g 1 0 u n it ~ 1 -3 l o g f o ld n d 1 .5 -1 .8 l o g 0 .8 1 .8 l o g ~ 1 .5 -2 .5 l o g ~ 0 .5 − 1 l o g t re a tm e n t ro u te im iv /s c /i p iv /s c ip ip iv iv iv in fe ct io n ro u te ip iv ip ip ip in iv iv m ic ro o rg a n is m a . b a u m a n n ii c . a lb ic a n s e . co li m r s a u s a 3 0 0 m r s a u s a 3 0 0 c o -i n fe ct io n p . a e ru g in o sa a n d a . b a u m a n n ii m r s a s . a u re u s p . a e ru g in o sa , k . p n e u m o n ia e , a n d a . b a u m a n n ii m ic e m ic e c d -1 n m r i c 5 7 b l/ 6 c 5 7 b l/ 6 b a lb /c c 5 7 b l/ 6 c 5 7 b l/ 6 p e p ti d e c h e x1 a rg 2 0 a m id e c h e x1 a rg 2 0 h y d ra zi d e r t d -1 a p i1 3 7 d f t 5 0 3 d f t 5 6 1 t a ch y p le si n ii i d p 7 a m p r -1 1 http://dx.doi.org/10.5599/admet.1215 g. wang and a.f. mechesso admet & dmpk 10(2) (2022) 91-105 100 table 1. continued… r e f (y e a r) [2 7 ] (2 0 2 0 ) [2 7 ] (2 0 2 0 ) [2 2 ] (2 0 2 1 ) a b b re v ia ti o n s: m t d , m a xi m u m t o le ra b le d o se ; n d , n o t d e te rm in e d ; ip , in tr a p e ri to n e a l; iv , in tr a v e n o u s; s c , s u b cu ta n e o u s; i n , in tr a n a sa l p k /p d t 1 /2 f o r h o ri n e < 1 5 m in a n d f o r d -h o ri n e ~ 1 h . d -f o rm d e te ct e d i n l u n g , sp le e n , li v e r, a n d k id n e y i n 2 4 h . n d m t d > 2 0 m g /k g ; d -f o rm b in d s le ss t o se ru m t h a n l -f o rm . s u rv iv a l ra te 8 7 .5 % s u rv iv e d /i p ( cf . 8 7 .5 % v a n co m y ci n ) 8 1 .8 % /i p ( cf . 5 0 % d o ri p e n e m ) n d b a ct e ri a l c f u d ro p u p t o 4 .5 l o g /i p ; ~ 2 -2 .5 l o g /i v 1 -3 l o g /i p ; u p t o 7 l o g /i v ~ 0 .5 − 1 l o g t re a tm e n t ro u te ip / iv ip / iv ip in fe ct io n ro u te ip ip ip m ic ro o rg a n is m m r s a u s a 3 0 0 k . p n e u m o n ia e m r s a u s a 3 0 0 m ic e c 5 7 b l/ 6 b a lb /c c 5 7 b l/ 6 b a lb /c c 5 7 b l/ 6 b a lb /c p e p ti d e h o ri n e v e ri n e c 1 0 -k r 8 d admet & dmpk 10(2) (2022) 91-105 antimicrobial peptides with systemic efficacy doi: http://dx.doi.org/10.5599/admet.1215 101 table 2. clinically used peptide antibiotics for systemic and skin infections 1 a p p ro v e d 1 9 7 0 1 9 9 4 1 9 5 8 1 9 8 0 e u ro p e 2 0 0 9 2 0 1 4 2 0 1 4 ; 2 0 2 1 2 0 0 3 1 g + , g ra m -p o si ti v e b a ct e ri a ; g -, g ra m -n e g a ti v e b a ct e ri a ; iv , in tr a v e n o u s; i m , in tr a m u sc u la r; c s s s is , co m p li ca te d s k in a n d s k in s tr u ct u re i n fe ct io n s r o u te iv iv iv im ,i v iv iv iv iv t re a tm e n t m u lt i d ru g re si st a n t g in fe ct io n s u ri n a ry t ra ct a n d b lo o d st re a m in fe ct io n s g + i n fe ct io n s g + i n fe ct io n s s k in i n fe ct io n s (c s s s i) s k in i n fe ct io n s s k in i n fe ct io n s (b e st d ru g to le ra b il it y ) s k in i n fe ct io n s m o a m e m b ra n e l y si s m e m b ra n e l y si s in h ib it s ce ll w a ll sy n th e si s in h ib it s ce ll w a ll sy n th e si s m e m b ra n e l y si s a n d i n h ib it s ce ll w a ll s y n th e si s m e m b ra n e l y si s a n d i n h ib it s ce ll w a ll s y n th e si s in h ib it s ce ll w a ll sy n th e si s m e m b ra n e l y si s a ct iv it y sp e ct ru m g g g + g + g + g + g + ( m r s a ) g + c la ss li p o p e p ti d e li p o p e p ti d e g ly co p e p ti d e g ly co p e p ti d e li p o g ly co p e p ti d e (v a n co m y ci n d e ri v e d ) (v a n co m y ci n -d e ri v e d ) li p o g ly co p e p ti d e (t e ic o p la n in a n a lo g ) li p o p e p ti d e t ra d e n a m e c o li st im e th a te s o d i u m , c o ly m y ci n m p o ly -r x v a n co ci n , v a n co le d t a rg o ci d v ib a ti v o rb a ct iv e d a lv a n ce ; x y d a lb a c u b ic in p e p ti d e a n ti b io ti c c o li st in ( p o ly m y x in e ) p o ly m y x in b v a n co m y ci n t e ic o p la n in t e la v a n ci n ( t d -6 4 2 4 ) o ri ta v a n ci n ( ly 3 3 3 3 2 8 ) d a lb a v a n ci n ( b i 3 9 7 ) d a p to m y ci n http://dx.doi.org/10.5599/admet.1215 g. wang and a.f. mechesso admet & dmpk 10(2) (2022) 91-105 102 conflict of interest: the authors declare no conflict of interest acknowledgement: this study is supported by the grants from the nih (gm138552) and the state of nebraska references [1] j. o’neil, the review on antimicrobial resistance: tracking drug-resistant infections globally. uk (2014). https://amr-review.org/sites/default/files/160518_final%20paper_with%20cover.pdf. [2] h.g. boman. antibacterial peptides: basic facts and emerging concepts. journal of internal medicine 254 (2003) 197–215. https://doi.org/10.1046/j.1365-2796.2003.01228.x. [3] m. zasloff. antimicrobial peptides of multicellular organisms. nature 415 (2002) 389–395. https://doi.org/10.1038/415389a. [4] b. mishra, s. reiling, d. zarena, g. wang. host defense antimicrobial peptides as antibiotics: design and application strategies. current opinion in chemical biology 38 (2017) 87-96. https://doi.org/10.1016/j.cbpa.2017.03.014. [5] z. wang, g. wang. apd: the antimicrobial peptide database. nucleic acids research 32 (2004) d590d592. https://doi.org/10.1093/nar/gkh025. [6] g. wang, c.m. zietz, a. mudgapalli, s. wang, z. wang. the evolution of the antimicrobial peptide database over 18 years: milestones and new features. protein science 31 (2022) 92-106. https://doi.org/10.1002/pro.4185. [7] g. wang. structures of human host defense cathelicidin ll-37 and its smallest antimicrobial peptide kr-12 in lipid structures of human host defense cathelicidin ll-37 and its smallest antimicrobial peptide kr-12 in lipid micelles. the journal of biological chemistry 283 (2008) 32637–32643. https://doi.org/10.1074/jbc.m805533200. [8] g. wang. unifying the classification of antimicrobial peptides in the antimicrobial peptide database. methods in enzymology (2021). in press. https://doi.org/10.1016/bs.mie.2021.09.006. [9] r.i. lehrer, w. lu. α-defensins in human innate immunity. immunological reviews 245 (2012) 84112. https://doi.org/10.1097/moh.0000000000000005. [10] g. diamond, l. ryan. beta-defensins: what are they really doing in the oral cavity? oral diseases 17 (2011) 628-635. https://doi.org/10.1111/j.1601-0825.2011.01799.x. [11] m. graf, m. mardirossian, f. nguyen, a.c. seefeldt, g. guichard, m. scocchi, c.a. innis, d.n. wilson. proline-rich antimicrobial peptides targeting protein synthesis. natural product reports 34 (2017) 702-711. https://doi.org/10.1039/c7np00020k. [12] r.m. epand, h.j. vogel. diversity of antimicrobial peptides and their mechanisms of action. biochimica et biophysica acta 1462 (1999)11-28. https://doi.org/10.1016/s0005-2736(99)00198-4. [13] a. makovitzki, d. avrahami, y. shai. ultrashort antibacterial and antifungal lipopeptides. proceedings of the national academy of sciences 103 (2006) 15997-60002. https://doi.org/10.1073/pnas.0606129103. [14] g. wang, j.l. narayana, b. mishra, y. zhang, f. wang, c. wang, d. zarena, t. lushnikova, x. wang. design of antimicrobial peptides: progress made with human cathelicidin ll -37. in: matsuzaki k. (eds) antimicrobial peptides. advances in experimental medicine and biology 1117 (2019) 215-240. https://doi.org/10.1007/978-981-13-3588-4_12. [15] x. li, y. li, h. han, d.w. miller, g. wang. solution structures of human ll-37 fragments and nmrbased identification of a minimal membrane-targeting antimicrobial and anticancer region. journal of the american chemical society 128 (2006) 5776-5785. https://doi.org/10.1021/ja0584875. [16] qin, h. fang, x. chen, l. wang, c. ma, x. xi, t. chen, c. shaw, m. zhou. exploration of the structure– function relationships of a novel frog skin secretion-derived bioactive peptide, t-dph1, through use of rational design, cationicity enhancement and in vitro studies. antibiotics 10 (2021) 1529. https://doi.org/10.3390/antibiotics10121529. https://amr-review.org/sites/default/files/160518_final%20paper_with%20cover.pdf https://onlinelibrary.wiley.com/action/dosearch?contribauthorraw=boman%2c+h+g https://doi.org/10.1046/j.1365-2796.2003.01228.x https://doi.org/10.1038/415389a https://doi.org/10.1038/415389a https://doi.org/10.1016/j.cbpa.2017.03.014 https://doi.org/10.1016/j.cbpa.2017.03.014 https://doi.org/10.1093/nar/gkh025 https://doi.org/10.1002/pro.4185 https://doi.org/10.1002/pro.4185 https://doi.org/10.1074/jbc.m805533200 https://doi.org/10.1016/bs.mie.2021.09.006 https://doi.org/10.1097/moh.0000000000000005 https://doi.org/10.1111/j.1601-0825.2011.01799.x https://doi.org/10.1039/c7np00020k https://doi.org/10.1016/s0005-2736(99)00198-4 https://doi.org/10.1073/pnas.0606129103 https://doi.org/10.1073/pnas.0606129103 https://doi.org/10.1007/978-981-13-3588-4_12 https://doi.org/10.1021/ja0584875 https://doi.org/10.3390/antibiotics10121529 admet & dmpk 10(2) (2022) 91-105 antimicrobial peptides with systemic efficacy doi: http://dx.doi.org/10.5599/admet.1215 103 [17] d. szabo , e. ostorhazi, a. binas, f. rozgonyi, b. kocsis, m. cassone, j.d. wade, o. nolte, l. otvos . . the designer proline-rich antibacterial peptide a3-apo is effective against systemic escherichia coli infections in different mouse models. international journal of antimicrobial agents 35 (2010) 357– 361. https://doi.org/10.1016/j.ijantimicag.2009.10.015. [18] r. schmidt, e. ostorházi, e. wende, d. knappe, r. hoffmann. pharmacokinetics and in vivo efficacy of optimized oncocin derivatives. journal of antimicrobial chemotherapy 71 (2016) 1003–1011. https://doi.org/10.1093/jac/dkv454. [19] r. schmidt, d. knappe, e. wende, e. ostorházi, r. hoffmann. in vivo efficacy and pharmacokinetics of optimized apidaecin analogs. frontiers in chemistry 5 (2017) 15. https://doi.org/10.3389/fchem.2017.00015. [20] d.c. rossi, j.e. muñoz, d.d. carvalho, r. belmonte, b. faintuch, p. borelli, a. miranda, c.p. taborda, s. daffre . therapeutic use of a cationic antimicrobial peptide from the spider acanthoscurria gomesiana in the control of experimental candidiasis. bmc microbiology 12 (2012) 28. https://doi.org/10.1186/1471-2180-12-28. [21] h.r. lee, d.g. you, h.k. kim, j.w. sohn, m.j. kim, j.k. park, g.y. lee, y.d. yoo. romo1-derived antimicrobial peptide is a new antimicrobial agent against multidrug-resistant bacteria in a murine model of sepsis. mbio 11 (2020) e03258–19. https://doi.org/10.1128/mbio.03258-19. [22] j. l. narayana , r. golla, b. mishra, x. wang, t. lushnikova, y. zhang, a.verma, v. kumar, j. xie, g. wang. short and robust anti-infective lipopeptides engineered based on the minimal antimicrobial peptide kr12 of human ll-37. acs infectious disease 7 (2021) 1795-1808. https://doi.org/10.1021/acsinfecdis.1c00101. [23] s.a. li, w.h. lee, y. zhang. efficacy of oh-cath30 and its analogs against drug-resistant bacteria in vitro and in mouse models. antimicrobial agents and chemotherapy 56 (2012) 33093317. https://doi.org/10.1128/aac.06304-11. [24] b. deslouches, k. islam, j.k. craigo, s.m. paranjape, r.c. montelaro, t.a. mietzner. activity of the de novo engineered antimicrobial peptide wlbu2 against pseudomonas aeruginosa in human serum and whole blood: implications for systemic applications. antimicrobial agents and chemotherapy 49 (2005) 3208–3216. https://doi.org/10.1128/aac.49.8.3208-3216.2005. [25] b. mishra, j.l. narayana, t. lushnikova, x. wang, g. wang. low cationicity is important for systemic in vivo efficacy of database-derived peptides against drug-resistant gram-positive pathogens. proceedings of the national academy of sciences 116 (2019) 13517–13522. https://doi.org/10.1073/pnas.1821410116. [26] b. mishra, g. wang. ab initio design of potent anti-mrsa peptides based on database filtering technology. journal of the american chemical society 134 (2012) 12426-12429. https://doi.org/10.1021/ja305644e. [27] j.l. narayana, b. mishra, t. lushnikova, q. wu, y.s. chhonkerb, y.zhang, d. zarena, e.s. salnikov, x. dang, f. wang, c. murphy, k.w. foster, s. gorantla, b. bechinger, d.j. murry, g. wang. two distinct amphipathic peptide antibiotics with systemic efficacy. proceedings of the national academy of sciences 117 (2020) 19446–19454. https://doi.org/10.1073/pnas.2005540117. [28] g. wang. post-translational modifications of natural antimicrobial peptides and strategies for peptide engineering. current biotechnology 1 (2012) 72-79. https://doi.org/10.2174/2211550111201010072. [29] b.p. mowery, s.e. lee, d.a. kissounko, r.f. epand, r.m. epand, b. weisblum, s.s. stahl, s.h. gellman. mimicry of antimicrobial host-defense peptides by random copolymers. journal of the american chemical society 129 (2007) 15474-15476. https://doi.org/10.1021/ja077288d. [30] r.w. scott, w.f. degrado, g.n. tew. de novo designed synthetic mimics of antimicrobial peptides. current opinion in biotechnology 19 (2008) 620-627. https://doi.org/10.1016/j.copbio.2008.10.013. [31] i.s. radzishevsky, s. rotem, d. bourdetsky, s. navon-venezia, y. carmeli, a. mor. improved antimicrobial peptides based on acyl-lysine oligomers. nature biotechnology 25 (2007) 657659. https://doi.org/10.1038/nbt1309. http://dx.doi.org/10.5599/admet.1215 https://www.sciencedirect.com/science/article/abs/pii/s0924857909004907?via%3dihub#! https://www.sciencedirect.com/science/article/abs/pii/s0924857909004907?via%3dihub#! https://www.sciencedirect.com/science/article/abs/pii/s0924857909004907?via%3dihub#! https://www.sciencedirect.com/science/article/abs/pii/s0924857909004907?via%3dihub#! https://www.sciencedirect.com/science/article/abs/pii/s0924857909004907?via%3dihub#! https://www.sciencedirect.com/science/article/abs/pii/s0924857909004907?via%3dihub#! https://www.sciencedirect.com/science/article/abs/pii/s0924857909004907?via%3dihub#! https://www.sciencedirect.com/science/article/abs/pii/s0924857909004907?via%3dihub#! https://www.sciencedirect.com/science/article/abs/pii/s0924857909004907?via%3dihub#! https://doi.org/10.1016/j.ijantimicag.2009.10.015 https://doi.org/10.1093/jac/dkv454 https://doi.org/10.3389/fchem.2017.00015 https://doi.org/10.3389/fchem.2017.00015 https://bmcmicrobiol.biomedcentral.com/articles/10.1186/1471-2180-12-28#auth-rodrigo-belmonte https://bmcmicrobiol.biomedcentral.com/articles/10.1186/1471-2180-12-28#auth-bluma-faintuch https://bmcmicrobiol.biomedcentral.com/articles/10.1186/1471-2180-12-28#auth-primavera-borelli https://bmcmicrobiol.biomedcentral.com/articles/10.1186/1471-2180-12-28#auth-antonio-miranda https://bmcmicrobiol.biomedcentral.com/articles/10.1186/1471-2180-12-28#auth-carlos_p-taborda https://bmcmicrobiol.biomedcentral.com/articles/10.1186/1471-2180-12-28#auth-sirlei-daffre https://bmcmicrobiol.biomedcentral.com/articles/10.1186/1471-2180-12-28#auth-sirlei-daffre https://doi.org/10.1186/1471-2180-12-28 https://doi.org/10.1128/mbio.03258-19 https://pubmed.ncbi.nlm.nih.gov/?term=lakshmaiah+narayana+j&cauthor_id=33890759 https://pubmed.ncbi.nlm.nih.gov/?term=golla+r&cauthor_id=33890759 https://pubmed.ncbi.nlm.nih.gov/?term=mishra+b&cauthor_id=33890759 https://pubmed.ncbi.nlm.nih.gov/?term=wang+x&cauthor_id=33890759 https://pubmed.ncbi.nlm.nih.gov/?term=lushnikova+t&cauthor_id=33890759 https://pubmed.ncbi.nlm.nih.gov/?term=zhang+y&cauthor_id=33890759 https://pubmed.ncbi.nlm.nih.gov/?term=verma+a&cauthor_id=33890759 https://pubmed.ncbi.nlm.nih.gov/?term=kumar+v&cauthor_id=33890759 https://pubmed.ncbi.nlm.nih.gov/?term=xie+j&cauthor_id=33890759 https://pubmed.ncbi.nlm.nih.gov/?term=wang+g&cauthor_id=33890759 https://pubmed.ncbi.nlm.nih.gov/?term=wang+g&cauthor_id=33890759 https://doi.org/10.1021/acsinfecdis.1c00101 https://doi.org/10.1021/acsinfecdis.1c00101 https://doi.org/10.1128/aac.06304-11 https://doi.org/10.1128/aac.49.8.3208-3216.2005 https://doi.org/10.1073/pnas.1821410116 https://doi.org/10.1073/pnas.1821410116 https://doi.org/10.1021/ja305644e https://doi.org/10.1021/ja305644e https://doi.org/10.1073/pnas.2005540117 https://doi.org/10.2174/2211550111201010072 https://doi.org/10.1021/ja077288d https://www.sciencedirect.com/journal/current-opinion-in-biotechnology https://doi.org/10.1016/j.copbio.2008.10.013 https://doi.org/10.1038/nbt1309 g. wang and a.f. mechesso admet & dmpk 10(2) (2022) 91-105 104 [32] i. ahmad, w.r. perkins, d.m. lupan, m.e. selsted, a.s. janoff. liposomal entrapment of the neutrophil-derived peptide indolicidin endows it with in vivo antifungal activity. biochimica et biophysica acta (bba) biomembranes 1237 (1995) 109–114. https://doi.org/10.1016/00052736(95)00087-j. [33] x. liu, z. li, x. wang, y. chen, f. wu, k. men, t. xu, y. luo, l.yang. novel antimicrobial peptidemodified azithromycin-loaded liposomes against methicillin-resistant staphylococcus aureus. international journal of nanomedicine 11 (2016) 6781–6794. https://doi.org/10.2147/ijn.s107107. [34] j. johansson, g.h. gudmundsson, m.e. rottenberg, k.d. berndt, b. agerberth. conformationdependent antibacterial activity of the naturally occurring human peptide ll-37. journal of biological chemistry 273 (1998) 3718-3724. https://doi.org/10.1074/jbc.273.6.3718. [35] f. jafari, s. elyasi. prevention of colistin induced nephrotoxicity: a review of preclinical and clinical data. expert review of clinical pharmacology. 14 (2021) 1113-1131. https://doi.org/10.1080/17512433.2021.1933436. [36] b.k. prasannan, f.c. mukthar, v.n. unni, s. mohan, k. vinodkumar. colistin nephrotoxicity-age and baseline kidney functions hold the key. indian journal of nephrology. 31 (2021) 449-453. https://doi.org/10.4103/ijn.ijn_130_20. [37] v. dartois, j. sanchez-quesada, e. cabezas, e. chi, c. dubbelde, c. dunn, j. granja, c. gritzen, d. weinberge, m. r. ghadiri, t.r. parr. systemic antibacterial activity of novel synthetic cyclic peptides antimicrobial agents and chemotherapy 49 (2005) 3302–3310. https://doi.org/10.1128/aac.49.8.3302-3310.2005. [38] l.l. ling, t. schneider, a.j. peoples, a.l. spoering, i. engels, b.p. conlon, a. mueller, t.f. schäberle, d.e. hughes, s. epstein, m. jones, l. lazarides, v.a. steadman, d.r. cohen, c.r. felix, k.a. fetterman, w.p. millett, a.g. nitti, a.m. zullo, c. chen, k. lewis. a new antibiotic kills pathogens without detectable resistance. nature 517 (2015) 455-459. https://doi.org/10.1038/nature14098. [39] j.w. van t wout, h. mattie, r. van furth. comparison of the efficacies of amphotericin b, fluconazole, and itraconazole against a systemic candida albicans infection in normal and neutropenic mice. antimicrobial agents and chemotherapy 33 (1989) 147-151. https://doi.org/10.1128/aac.33.2.147. [40] c.b. landersdorfer, j. wang, v. wirth, k. chen, k.s. kaye, b.t. tsuji, j. li, r.l. nation. pharmacokinetics/pharmacodynamics of systemically administered polymyxin b against klebsiella pneumoniae in mouse thigh and lung infection models. journal of antimicrobial chemotherapy 73 (2018) 462-468. https://doi.org/10.1093/jac/dkx409. [41] j.b. schaal, y. eriguchi, d.q. tran, p.a. tran, c. hawes, a.e. cabebe, k. pike, k. trinh, a.j. ouellette, m.e. selsted. a host-directed macrocyclic peptide therapeutic for mdr gram negative bacterial infections. scientific reports 11 (2021) 23447. https://doi.org/10.1038/s41598-021-02619-y. [42] d. knappe, r. schmidt, k. adermann, r. hoffmann. continuous subcutaneous delivery of proline-rich antimicrobial peptide api137 provides superior efficacy to intravenous administration in a mouse infection model. frontiers in microbiology 10 (2019) 2283. https://doi.org/10.3389/fmicb.2019.02283. [43] r. zhang, z. wang, y. tian, q. yin, x. cheng, m. lian, b. zhou, x. zhang, l. yang. efficacy of antimicrobial peptide dp7, designed by machine-learning method, against methicillin-resistant staphylococcus aureus. frontiers in microbiology 10 (2019) 1175. https://doi.org/10.3389/fmicb.2019.01175. [44] v. basso, d.q. tran, j.b. schaal, p. tran, y. eriguchi, d. ngole, a.e. cabebe, a.y. ppark, p.m. beringer, a.j. ouellette, m.e. selsted. rhesus theta defensin 1 promotes long term survival in systemic candidiasis by host directed mechanisms. scientific reports 9 (2019) 16905. https://doi.org/10.1038/s41598-019-53402-z. [45] s.a. li, y. xiang, y.j. wang, j. liu, w.h. lee, y. zhang. naturally occurring antimicrobial peptide ohcath30 selectively regulates the innate immune response to protect against sepsis. journal of medicinal chemistry 56 (2013) 9136–9145. https://doi.org/10.1021/jm401134n. https://www.infona.pl/contributor/0@bwmeta1.element.elsevier-b57cf62a-4dc7-38e6-9d67-a703605403d0/tab/publications https://www.infona.pl/contributor/1@bwmeta1.element.elsevier-b57cf62a-4dc7-38e6-9d67-a703605403d0/tab/publications https://www.infona.pl/contributor/2@bwmeta1.element.elsevier-b57cf62a-4dc7-38e6-9d67-a703605403d0/tab/publications https://www.infona.pl/contributor/3@bwmeta1.element.elsevier-b57cf62a-4dc7-38e6-9d67-a703605403d0/tab/publications https://www.infona.pl/contributor/4@bwmeta1.element.elsevier-b57cf62a-4dc7-38e6-9d67-a703605403d0/tab/publications https://www.sciencedirect.com/science/journal/00052736 https://www.sciencedirect.com/science/journal/00052736 https://doi.org/10.1016/0005-2736(95)00087-j https://doi.org/10.1016/0005-2736(95)00087-j https://doi.org/10.2147/ijn.s107107 https://doi.org/10.2147/ijn.s107107 https://doi.org/10.1074/jbc.273.6.3718 https://doi.org/10.1080/17512433.2021.1933436 https://doi.org/10.1080/17512433.2021.1933436 https://doi.org/10.4103/ijn.ijn_130_20 https://doi.org/10.1128/aac.49.8.3302-3310.2005 https://doi.org/10.1128/aac.49.8.3302-3310.2005 https://doi.org/10.1038/nature14098 https://doi.org/10.1128/aac.33.2.147 https://doi.org/10.1093/jac/dkx409 https://doi.org/10.1038/s41598-021-02619-y https://doi.org/10.3389/fmicb.2019.02283 https://doi.org/10.3389/fmicb.2019.02283 https://doi.org/10.3389/fmicb.2019.01175 https://doi.org/10.3389/fmicb.2019.01175 https://doi.org/10.1038/s41598-019-53402-z https://doi.org/10.1038/s41598-019-53402-z https://pubs.acs.org/action/dosearch?field1=contrib&text1=sheng-an++li https://pubs.acs.org/action/dosearch?field1=contrib&text1=yang++xiang https://pubs.acs.org/action/dosearch?field1=contrib&text1=yan-jie++wang https://pubs.acs.org/action/dosearch?field1=contrib&text1=jie++liu https://pubs.acs.org/action/dosearch?field1=contrib&text1=wen-hui++lee https://pubs.acs.org/action/dosearch?field1=contrib&text1=yun++zhang https://doi.org/10.1021/jm401134n admet & dmpk 10(2) (2022) 91-105 antimicrobial peptides with systemic efficacy doi: http://dx.doi.org/10.5599/admet.1215 105 [46] n. mookherjee, r.e. hancock. cationic host defence peptides: innate immune regulatory peptides as a novel approach for treating infections. cellular and molecular life sciences 64 (2007) 922-933. https://doi.org/10.1007/s00018-007-6475-6. [47] a. giacometti, o. cirioni, r. ghiselli, c. bergnach, f. orlando, g. d’amato, f. mocchegiani, c. silvestri, m.s. del prete, b. skerlavaj, v. saba, m. zanetti, g. scalise. the antimicrobial peptide bmap-28 reduces lethality in mouse models of staphylococcal sepsis. critical care medicine 32 (2004) 2485– 2490. https://doi.org/10.1097/01.ccm.0000148221.09704.22. [48] s. pacor, a. giangaspero, m. bacac, g. sava, a. tossi. analysis of the cytotoxicity of synthetic antimicrobial peptides on mouse leucocytes: implications for systemic use. journal of antimicrobial chemotherapy 50 (2002) 339–348. https://doi.org/10.1093/jac/dkf141. [49] a.p. desbois, c.g. gemmell, p.j. coote. in vivo efficacy of the antimicrobial peptide ranalexin in combination with the endopeptidase lysostaphin against wound and systemic meticillin-resistant staphylococcus aureus (mrsa) infections. international journal of antimicrobial agents 35 (2010) 559–565. https://doi.org/10.1016/j.ijantimicag.2010.01.016. [50] h. sarig, l. livne, v. held-kuznetsov, f. zaknoon, a. ivankin, d. gidalevitz, a. mor a miniature mimic of host defense peptides with systemic antibacterial efficacy. the faseb journal 24 (2010) 19041913. https://doi.org/10.1096/fj.09-149427. [51] e. ostorhazi, f. rozgonyi, d. szabo, a. binas, m. cassone,j.d. wade, o. nolte, c.r. bethel, r.a. bonomo, l,otvos. intramuscularly administered peptide a3-apo is effective against carbapenemresistant acinetobacter baumannii in mouse models of systemic infections. peptide science 96 (2010) 126–129. https://dx.doi.org/10.1002/bip.21443. [52] m.c. erfe, c.v. david, c. huang, v. lu, a.c. maretti-mira, j. haskell, k.w. bruhn, m.r. yeaman, n. craft. efficacy of synthetic peptides rp-1 and aa-rp-1 against leishmania species in vitro and in vivo. antimicrob agents and chemotherapy. 56 (2012) 658-665. https://doi.org/10.1128/aac.05349-11. [53] x. liu, z. li, x. wang, y. chen, f. wu, k. men, t. xu, y. luo, l. yang. novel antimicrobial peptidemodified azithromycin-loaded liposomes against methicillin-resistant staphylococcus aureus. international journal of nanomedicine. 11 (2016) 6781-6794. https://doi.org/10.2147/ijn.s107107. [54] e. ostorhazi, r. hoffmann, n. herth, j.d. wade, c.n. kraus, l. otvos. advantage of a narrow-spectrum host defense (antimicrobial) peptide over a broad spectrum analog in preclinical drug development. frontiers in chemistry 6 (2018) 359. https://doi.org/10.3389/fchem.2018.00359. [55] j. qi, r. gao, c. liu, b. shan, f. gao, j. he, m. yuan, h. xie, s. jin, y. ma. potential role of the antimicrobial peptide tachyplesin iii against multidrug-resistant p. aeruginosa and a. baumannii coinfection in an animal model. infection and drug resistance 12 (2019) 2865–2874. https://doi.org/10.2147/idr.s217020. ©2022 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://dx.doi.org/10.5599/admet.1215 https://doi.org/10.1007/s00018-007-6475-6 https://doi.org/10.1097/01.ccm.0000148221.09704.22 https://doi.org/10.1093/jac/dkf141 https://doi.org/10.1016/j.ijantimicag.2010.01.016 https://doi.org/10.1096/fj.09-149427 https://dx.doi.org/10.1002/bip.21443 https://doi.org/10.1128/aac.05349-11 https://doi.org/10.2147/ijn.s107107 https://doi.org/10.3389/fchem.2018.00359 https://doi.org/10.2147/idr.s217020 https://doi.org/10.2147/idr.s217020 http://creativecommons.org/licenses/by/3.0/ dissolution-permeation of hot-melt extruded amorphous solid dispersion comprising an experimental grade of hpmcas doi: https://doi.org/10.5599/admet.1586 1 admet & dmpk 0(0) (2023) 000-000; doi: https://doi.org/10.5599/admet.1586 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper dissolution-permeation of hot-melt extruded amorphous solid dispersion comprising an experimental grade of hpmcas hironori tanaka1, tetsuya miyano2 and hiroshi ueda3,* 1formulation r&d laboratory, shionogi & co., ltd., hyogo 660-0813, japan 2laboratory for medicinal chemistry research, shionogi & co., ltd., osaka, 561-0825, japan 3analysis and evaluation laboratory, shionogi & co., ltd., osaka, 561-0825, japan *corresponding author: e-mail:: hiroshi.ueda@shionogi.co.jp; tel.: +81-6-6331-5747; fax: +81-6-6332-6385 received: november 4, 2022; revised: july 20, 2023; published: july 22,2023 abstract background and purpose: physicochemical properties of an amorphous solid dispersion (asd) comprising an experimental grade of hydroxypropyl methylcellulose acetate succinate (hpmcas-mx) with lower glass transition temperature have been previously investigated. this study aimed to evaluate applicability of hpmcas-mx to hot-melt extrusion (hme) and dissolution-permeation performance of prepared asds using microflux. review approach: a physical mixture of indomethacin (imc) and hpmcas-mx or -mg (a commercial grade with higher transition temperature) at 20:80 weight ratio was hot-melt extruded to prepare an asd (imc-mx and imc-mg, respectively). the dissolution-permeation performance and the stability of the asds were measured. key results: a torque reduction at 120 °c implied that imc-mx transformed into an amorphous state at this temperature, but imc-mg required around 170 °c. this result was supported by raman mapping of the the hme samples. imc-mg and imc-mx remained in an amorphous state at 40 °c for three months. the initial dissolution rate and solubility of the asds were higher than that of crystalline imc. the apparent permeability of imc from imc-mx and imc-mg was comparable but was approximately two-fold higher than that from crystalline imc. conclusion: hpmcas-mx enabled hme process at a lower temperature and improved the dissolution-permeation performance of indomethacin. ©2023 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords amorphous; microflux™; physical stability; raman mapping; x-ray powder diffraction; extrusion introduction oral administration is an important advantage for small molecular medicine because medium and large molecules show lower intestinal membrane permeability, followed by poor oral absorption [1-2]. however, the number of poorly water-soluble compounds has increased and become a key issue for drug discovery. the development of small molecules due to a decrease in aqueous medium solubility could result in reduced absorption per the oral route [3-4]. solubility issues have been reported in approximately 75 % of new drug candidates. therefore, physicochemical and pharmaceutical studies to improve solubility are required [3]. conversion of the drug form to salt and cocrystal and a reduction of the particle size between the micrometer to nanometer range are commonly attempted to modify the physicochemical properties of compounds [4]. modification of drug properties, and combination with co-solvent, surfactant, ph-modifier, and cyclodextrin, have been considered as potential methods [4]. in addition, amorphization with the disruption of the crystal https://doi.org/10.5599/admet.1586 https://doi.org/10.5599/admet.1586 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:hiroshi.ueda@shionogi.co.jp http://creativecommons.org/licenses/by/4.0/ h. tanaka et al. admet & dmpk 00(0) (2023) 000-000 2 lattice of the compound has been reported to enhance solubility [4-6]. since the amorphous form is thermodynamically unstable and might cause crystallization during the formulation process and/or storage, stabilization of amorphous drugs by combination with a hydrophilic polymer as a carrier was designed as polymeric solid dispersion [5,6]. miscible dispersion of amorphous drugs into a polymeric carrier with the formation of intermolecular interaction could decrease the molecular mobility, which inhibits crystallization of the amorphous drug [7,8]. however, the formulation process is a major concern for the manufacturing of solid dispersion, where solvent-mediated and fusion methods have been utilized [9]. the solvent-mediated technique involves spraydrying drugs and dissolving polymeric carriers in organic and/or aqueous solvents. rapid evaporation of solvent results in drug amorphization with miscible dispersion into the polymeric carrier at the molecular level. in contrast, hot-melt extrusion is a process that does not use solvent as a representative fusion technique. high-shared mixing of drug and polymeric carrier using a twin rotating screw over the glass transition temperature (tg) of a polymeric carrier can transform the polymeric carrier to a rubber state. subsequently, crystalline drug co-melted with rubbery polymeric carrier under high-shared mixing results in a miscible amorphous blend as a solid dispersion. although both processes are effective, recent environmental concerns have led to hot-melt extrusion being considered more attractive as it is a solventfree process [8]. commercial products of solid dispersion prepared by hot-melt extrusion can comprise various polymeric carriers [8]. polyvinylpyrrolidone, copovidone, and cellulosic polymers are widely used in commercially and in research [8-12]. studies on polymeric carriers revealed that hydroxypropyl methylcellulose acetate succinate (hpmcas) has better physical stability and oral absorption than other carriers [13-15] and is used in a commercial product, “noxafil“. although an advantage for solvent-free processes, elevation of the hot-melt extrusion temperature is limited because thermal degradation of the drug and/or polymeric carrier may occur [16]. to overcome this problem, a study on tadalafil, a known heat-sensitive drug with a high melting temperature (tm), was performed [16]. the addition of surfactant as a plasticizer to a polymeric carrier could cause a reduction in tg, enabling the hot-melt extrusion process to be performed at a lower temperature [17,18]. hence, designing and developing a novel solid dispersion carrier with lower tg is desirable. in a previous study, we focused on an experimental grade of hpmcas originally designed with lower tg than commercial grade [19]. physicochemical investigation of its solid dispersion revealed the formation of intermolecular interaction with indomethacin (imc) inhibited crystallization. even though the tg of an experimental grade hpmcas was lower than that of commercial grade, their performances were comparable. this study aimed to investigate the applicability of an experimental grade hpmcas to the hot-melt extrusion process and the dissolutionpermeation performance of the extruded solid dispersion. indomethacin was employed as the model drug, and a commercial grade of hpmcas was used for comparison. experimental materials indomethacin was purchased from kongo chemical co., ltd. (toyama, japan). a commercial grade of hpmcas “shin-etsu aqoat® type mg” and an experimental grade “type mx” originally developed were kindly gifted by shin-etsu chemical co., ltd. (tokyo, japan) [19]. git-0 lipid (20 wt.% phospholipid dissolved into dodecane), prisma ht buffer (ph 6.5), and acceptor sink buffer (asb) were purchased from pion inc. (ma, usa). admet & dmpk 00(0) (2023) 000-000 dissolution-permeation of hot-melt extruded asd doi: https://doi.org/10.5599/admet.1586 3 preparation of solid dispersion solid dispersion formulations were prepared by hot-melt extrusion (hme) process using a micro-conical twin-screw compounder “haake™ mini ctw” (thermo fisher scientific inc., tokyo, japan). physical mixtures comprising imc and hpmcas-mg or hpmcas-mx at a 20:80 ratio were prepared using a mortar and pestle and 10 g of each physical mixture was subjected into a feeder of the hot-melt extrusion apparatus followed by mixing with twin-screw rotation at 20 rpm. both carriers showed no thermal degradation up to 250 °c [19], therefore, the preparation temperature was set to be under 250 °c. the hme experiment was performed to confirm the appropriate temperature for the process. subsequently, the temperature was set at 100 °c, and the mixture was processed for 10 min. the temperature was raised in increments of 10° c from 100 to 170 °c; the hme process was conducted for 10 min at each temperature. the solid dispersion samples prepared at 170 °c were used for the physiochemical and dissolution-permeation studies because the formulations showed amorphization at this temperature. extruder torque readings were recorded for each temperature after each 10-min process. the hot-melt extruded was milled using a tablet mill jf-001j1 (takazono corporation., tokyo, japan) at around 15,000 rpm for 1 min. the resulting powders were treated as solid dispersion samples. the sample comprising 20 % imc and 80 % hpmcas-mg (or -mx) is termed imcmg (or -mx) (20:80) in the current study. scanning electron microscope (sem) the hot-melt extruded samples prepared at 170 °c were glued on 12 mm diameter aluminum sample holders using a round carbon seal (nisshin em co., ltd., tokyo, japan). the samples were sputter-coated with an au alloy using a magnetron sputter (msp-1s, vacuum device, ibaraki, japan). the shapes and surface morphologies of samples were investigated using an sem (ve-8800, keyence corporation, osaka, japan). raman mapping to measure the amorphization and distribution of imc into the carrier non-destructively, raman imaging was performed for the hot-melt extruded samples prepared at 120 and 170 °c. the raman spectra of crystalline and amorphous imc were measured using a ramantouch laser raman microscope (nanophoton corporation, osaka, japan). the samples were placed onto a hole (diameter: 3 mm and depth: 0.2 mm) on an aluminum plate, followed by smoothing of the surface using a spatula. raman spectroscopic measurement was performed in the carbonyl region between 1800-1500 cm−1. the procedure and the conditions of the apparatus were set according to a previous study [19]: excitation wavelength 785.16 nm, excitation power 113.04 mw, nd filter 100 %, spectrograph center wavelength 1400 cm−1, grating 600 gr mm-1, slit width 50 µm, exposure time 3 s, averaging 1, gain high, readout port low noise, readout speed 2 mhz, ccd temperature −70 °c, and objective lens 10×/na 0.3. the wavenumber was calibrated using a spectrum of silicon provided by the equipment. an area of the solid dispersions measuring 200×200 μm was evaluated at a resolution of 2000 nm per spot using raman imager 2 (nanophoton corporation, osaka, japan). the raman image was drawn by the peak ratio of the crystalline and the amorphous imc; the ranges of 1699.8 to 1700.5 cm−1 and 1685.1 to 1695.8 cm−1 were used as the peak areas of the crystalline and the amorphous imc, respectively. x-ray powder diffraction amorphization of the hot-melt extruded samples was measured by x-ray powder diffraction (xrpd) . the smooth samples were placed on an aluminum plate with holes (diameter = 2 mm, depth = 0.1 mm) and analyzed with a smartlab (rigaku corporation, tokyo, japan) instrument equipped with a 9 kw rotating anode using cu kα radiation (λ = 0.154186 nm) and a hypix-3000 detector. the distance between the sample and detector was 331 mm; the diffractometer was equipped with a cross-beam optic, providing a parallel https://doi.org/10.5599/admet.1586 h. tanaka et al. admet & dmpk 00(0) (2023) 000-000 4 beam. a psc with 2.5° and a slit of 0.05 mm height and 0.5 mm width was used; thus, the beam footprint for all measurement configurations was smaller than that of the sample. no slit was used on the receiving side. the cu kα radiation point source was operated at 40 kv and 200 ma. the scan was performed from 2 = 3 to 32°, with  axis rotation (20 rpm) during data collection; the sampling step was 2 = 0.02°, and the count time was 40 s. the data were analyzed using smart lab studio ii x64 version 4.2.111.0 software (rigaku corporation, tokyo, japan). physical stability the crystallization tendency of the solid dispersion samples was evaluated. the samples held on a hole (diameter: 3 mm and depth: 0.2 mm) of an aluminum plate were placed into a desiccator with silica gel. the samples in the desiccator were stored at 40 °c for three months, followed by xrpd measurement. dissolution-permeation test dissolution and permeation of the crystalline imc and the solid dispersion samples were simultaneously measured using microflux™ (pion inc., ma, usa). the apparatus contains donor and acceptor chambers connected through an artificial membrane (ϕ 0.45 μm of polyvinylidenfluoride). the measurement conditions were set by modifying those of previous studies [20]. first, 25 μl of git-0 lipid solution was dripped onto the membrane surface from the donor side. thereafter, 20 ml of prisma ht buffer and acceptor sink buffer (asb)were added to the donor and acceptor chambers. prisma ht, a ph 6.5 buffer, acts as the donor medium and does not contain phosphate, boric acid, citric acid, or benzoate. asb is a hepes-based, ph 7.4 buffer, and acts as the acceptor buffer, which contains chemical scavengers to maintain sink conditions [21,22]. the solid dispersion samples corresponding to 10 mg of imc were added to the donor chamber. the dissolution media in both chambers were stirred at 150 rpm at 37 °c. the uv probes of 2 and 5 mm were inserted into the donor, and acceptor chambers, respectively, and the concentration of imc was measured using a uv spectrometer at 370 nm every 5 s (0 10 min), 20 s (10.5 60 min), 1 min (61 120 min), and 2 min (121 – 480 min). the concentration of imc was corrected with a change in baseline values using the difference between 370 and 380 nm. a standard curve for imc between 20.0-558.1 μg ml-1 in both chambers was plotted. the flux values through the membrane were obtained for crystalline imc and the solid dispersions of imc-mg (20 : 80) and imc-mx (20:80) in accordance with the methods previously described in the literature [23,24]. the flux value was calculated as follows: flux value, (µg min-1) cm-2 = the permeated slope, µg ml-1) min-1  the volume of the acceptor chamber /the surface area of the membrane. solubility test the solubility of crystalline imc in the media used in both donor and acceptor chambers of the microflux™ apparatus was measured. approximately 50 mg of crystalline imc was added to 5 ml of prisma ht (ph 6.5) and asb (ph 7.4) buffers. they were shaken at 70 cycles/min in an mm-10 water bath shaker (taitec, koshigaya, saitama, japan) at 37 °c for 24 h. the test media were filtered through a 0.45-mm syringe filter (gl sciences inc., tokyo, japan). approximately 1 ml of the filtrate was analyzed using a waters acquity uplc h-class system (waters corporation, massachusetts, usa). the measurement conditions were as follows: mobile phase, 0.1 % trifluoroacetic acid/acetonitrile; flow rate 0.4 ml min-1; injection volume 2 µl; uv absorbance 270 nm, and column acquity uplc beh c18 2.1 × 100 mm, 1.7μm at 35 °c. measurement of the drug-rich colloid prior to the experiment, the prisma ht buffer was filtered through a 0.45-mm syringe filter (gl sciences inc., tokyo, japan). a sample of the solid dispersions (5 mg) corresponding to 1 mg of imc was added to 2 ml of the filtered prisma ht buffer. they were shaken at 500 rpm at 37 °c for 4 h using a well plate shaker, m admet & dmpk 00(0) (2023) 000-000 dissolution-permeation of hot-melt extruded asd doi: https://doi.org/10.5599/admet.1586 5 br-022up (taitec corporation, saitama, japan), followed by filtration through a 0.45-μm syringe filter (gl sciences inc., tokyo, japan). approximately 100 μl of the filtrate sample was added to a cuvette to measure dynamic light scattering (dls) using zetasizer nano-zs (malvern panalytical ltd., malvern, uk). the mean diameter was calculated using photon correlation spectroscopy, and the attenuation and measurement settings were optimized automatically by zetasizer software version 8.01 (malvern panalytical ltd., malvern, uk). results and discussion hot-melt extrusion the physical mixture of imc-mg (20:80) and imc-mx (20:80) was subjected to the hot-melt extrusion process. figure 1 shows the torque values as a function of temperature for imc-mg (20:80) and imc-mx (20:80). the imc-mg (20:80) sample was mixed and extruded at 100 °c, but higher torque over 2.5 nm caused lower processability; the powder could not be fed into the twin-screw barrel. the temperature was elevated to 120 °c, and the process was re-performed. the torque measured 0.76 nm and subsequently decreased with each temperature elevation, reaching 0.22 nm at 170 °c. whereas when the imc-mx sample was processed at 100 °c, the torque measured approximately 0.55 nm. the torque was reduced to almost 0.20 nm at 120 °c. previously, an hme experiment using a mixture of ibuprofen and kollidon sr observed a reduction of torque with processing temperature, and this phenomenon could be correlated with the tg of the mixture [25]. the tg of hpmc-mg and hpmc-mx were 121.3±1.2 °c and 107.9±0.4 °c, respectively [19]. the observed rapid reduction of torque at temperatures over 120 °c was caused by the mixtures reaching the glass transition temperature of hpmcas followed by melting of imc. these results suggest that hpmcasmx can be effectively used in the hot-melt extrusion process for the preparation of solid dispersion at a relatively lower temperature than hpmcas-mg. temperature, °c figure 1. torque monitoring during hot-melt extrusion of imc-mg (20:80) and imc-mx (20:80) at 100-170 °c. figure 2 shows the appearance of the hot-melt extrudate imc-mg (20:80) and imc-mx (20:80) processed at 120 °c. the imc-mg (20:80) extrudate was characterized by a visible dispersion of white powder in the extrudate. in contrast, imc-mx (20:80) produced a translucent yellow extrudate. the glass transition of imc results in a change in appearance and a yellow-glassy appearance is formed by the disruption of the crystal lattice due to a change in the form of the intermolecular interaction between functional groups [26-27]. therefore, it suggests a possibility of partial or complete amorphization with the carrier after hot-melt extrusion. these results were consistent with the change in the torque during the hot-melt extrusion process shown in figure 1. imc-mg (20:80) required over 150 °c to reach a plateau in torque reduction, while imcmx (20:80) only required 120 °c. t o rq u e ( n m ) temperature (c) 90 180120 150 ● imc-mg (20:80) ● imc-mx (20:80) 0.0 0.2 0.4 0.6 0.8 1.0 t o rq u e , n m https://doi.org/10.5599/admet.1586 h. tanaka et al. admet & dmpk 00(0) (2023) 000-000 6 figure 2. appearance of hot-melt extruded imc-mg (20:80) and imc-mx (20:80) processed at 120 °c. to confirm if the dispersed white powder was crystalline imc, raman mapping was performed as a nondestructive method, as the milling process for the preparation of powder to be measured by xrpd might cause amorphization [28]. figure 3a shows the raman images of the cutting plane of the hot-melt extrude processed at 120 °c. (a) (b) (c) raman shift, cm-1 figure 3. raman images of imc-mg (20:80) and imc-mx (20:80) processed at (a) 120 °c and (b) 170 °c, and raman spectra of (c) points a-d in the raman images. imc-mg (20:80) at 120c imc-mx (20:80) at 120c imc-mg (20:80) at 120c imc-mx (20:80) at 120c 2.5 0.2 3.4 -1.5 50 µm point a point b point c point d 50 µm imc-mg (20:80) at 170c imc-mx (20:80) at 170c 2.5 -1.1 1.7 -0.1 50 µm 50 µm in te n s it y raman shift (cm-1) 1500 18001700 point a point b point c point d 1600 1700 cm-1 admet & dmpk 00(0) (2023) 000-000 dissolution-permeation of hot-melt extruded asd doi: https://doi.org/10.5599/admet.1586 7 the images were drawn using the peak ratio of crystalline and amorphous imc; the lower intensity is represented as dark gray and reflects crystalline imc. the distribution image of imc-mg (20:80) revealed that crystalline imc remained and was heterogeneously dispersed in the extrude. however, crystalline imc did not appear in imc-mx (20:80), and homogenous dispersion of amorphous imc into the extrude was shown. the results showed that imc-mx (20:80) could be transformed to solid dispersion by processing at 120 °c, but imcmg (20:80) could not. this temperature was lower than the reported values; the study of hot-melt extrusion using hpmcas set the process temperature above the tg at over 130 °c [29,30]. figure 3b shows the raman images of the cutting plane of the hot-melt extrude processed at 170 °c. both the hot-melt extrude of imc-mg (20:80) and imc-mx (20:80) processed at 170 °c showed homogeneous distribution without crystalline imc, suggesting that imc-mg (20:80) also formed amorphous dispersion. figure 3c shows the raman spectra of points a, b, c, and d, highlighted in figure 3a. points b, c, and d highlight the amorphous pattern similar to the previous study [19]. however, the crystalline peak of imc at 1700 cm-1 [31] appeared on point b, corresponding to a dark gray particle in imc-mg (20:80) at 120 °c. it appears that the raman image reflects the distribution of crystalline imc. to compare the performance of the solid dispersion comprising hpmc-mg or hpmc-mx, the solid dispersion prepared at 170 °c under the same conditions were subjected to further study. preparation of the powder and physical stability test the solid dispersion powder was prepared by milling the hot-melt extrude for further experiments. figure 4a and b show the sem images of imc-mg (20:80) and imc-mx (20:80) prepared by hot-melt extrusion at 170 °c followed by milling. a similar particle shape and size with consolidated and smooth glass-like material can be seen in both images. this appearance was almost consistent with a previous study [32]. (a) (b) figure 4. sem images of hot-melt extruded at 170 °c after milling for (a) imc-mg (20:80) and (b) imc-mx (20:80). figure 5 shows the xrpd patterns of imc-mg (20:80) and imc-mx (20:80) prepared by hot-melt extrusion at 170 °c followed by milling. all samples showed no x-ray diffraction peak derived from imc, suggesting amorphization was achieved following the process. furthermore, they were physically stable; no crystallization occurred, and the amorphous form was stored for three months at 40 °c. it is commonly known that tg is a key parameter used to determine the crystallization tendency of amorphous compounds because the molecular mobility is significantly elevated over tg with glass transition [33,34]. in addition, the formation of intermolecular interaction between the drug and polymeric carrier has been reported to play an important role in maintaining an amorphous state [12,15]. solid dispersion comprising of > 60 % of hpmcas-mg or hpmcas-mx showed a tg over 70 °c; the formation of intermolecular interaction between imc and hpmcasmg/hpmcas-mx was characterized by a change in infrared spectra [19]. the difference between the tg and imc-mg (20:80) imc-mx (20:80) https://doi.org/10.5599/admet.1586 h. tanaka et al. admet & dmpk 00(0) (2023) 000-000 8 storage temperature and the formation of intermolecular interaction with hpmcas-mg or hpmcas-mx should result in higher physical stability. 2 / ° figure 5. xrpd patterns of hot-melt extruded at 170 °c after milling for imc-mg (20:80) and imc-mx (20:80) before and three months after storage at 40 °c. dissolution-permeation test dissolution and the subsequent permeation profiles of crystalline imc and the solid dispersions were studied with simultaneous measurement for donor-acceptor chambers using a microflux™. figure 6a shows the dissolution profiles of crystalline imc and the solid dispersions of imc-mg (20:80) and imc-mx (20:80) prepared by hot-melt extrusion at 170 °c. the concentration of crystalline imc in the dissolution medium increased time-dependently and almost achieved a plateau at 185.9  0.9 µg ml-1 between 420 and 480 min. this result was similar to that of the solubility test; the equilibrium solubility of crystalline imc for the acceptor medium was 177.22.7 µg ml-1 , which was also similar to the previously reported value (approximately 200 µg ml-1) [35]. the dissolution profile of imc was significantly improved by formulation of the solid dispersion. the initial dissolution rate and dissolved amount of imc elevated to around 3080.4 and 3160.5 µg ml-1 between 420-480 min for both imc-mg (20:80) and imc-mx (20:80), respectively. furthermore, no precipitation tendency was observed for both the solid dispersions, keeping supersaturation during the test time. these dissolved amounts of imc were similar to a previously reported value of a solid dispersion comprising imc and hpmcas-lf [36]. figure 6b shows the magnified dissolution profiles within 2 min. the slope of the concentration as a function of experimental time was calculated between 0-1 min for the solid dispersions, but crystalline imc was calculated until 2 min because the dissolution was significantly low until 1 min. the initial dissolution rate represented as the slope of crystalline imc was 1.6 (µg ml-1) min-1, while the solid dispersion of imc-mg (20:80) and imc-mx (20:80) were 46.0 and 44.3 (µg ml-1) min-1, corresponding to about a 29and 28-fold bigger improvement, respectively. improvement of the dissolution rate by the formation of solid dispersion using hpmcas as a polymeric carrier has been investigated for various drugs [37-39]. the formation of a drug-rich phase of danazol under a supersaturated state was confirmed prior to crystallization, which was affected by the addition of polymeric carriers [37]. in addition, the effect of hpmcas on the drug-rich phase of atazanavir was studied. the distribution of hpmcas into a drug-rich phase did not agree between different grades, resulting in dissolution performance [38]. structural conformation of hpmcas with random-coil or aggregation affected the supersaturation state of the drug; change in the conformation between grades of hpmcas impacted the supersaturation of celecoxib [39]. these findings suggest that both grades of hpmcas in te n s it y 2-theta () 5 2510 15 20 before storage 3 months at 40c before storage 3 months at 40c imc-mg (20:80) imc-mx (20:80) admet & dmpk 00(0) (2023) 000-000 dissolution-permeation of hot-melt extruded asd doi: https://doi.org/10.5599/admet.1586 9 used in this study formed similar interactions with imc in a supersaturated solution, resulting in comparable results for the initial dissolution rate and concentration at a plateau between 420 and 480 min. (a) (b) time, min time, min figure 6. dissolution profiles of crystalline imc and solid dispersions of imc-mg (20:80) and imc-mx (20:80) in the range of (a) 0-480 min and (b) 0-2.0 min. the error bars represent a standard deviation of n = 3. figure 7a shows the permeation profiles of crystalline imc and the solid dispersions of imc-mg (20:80) and imc-mx (20:80). the concentrations of imc linearly increased depending on experimental time for all samples; both the solid dispersions showed a reduction in lag-time from around 100 to 60-65 min and a higher concentration than that of crystalline imc. the concentrations of all of the samples were within 100 mg ml-1. the solubility test result for crystalline imc in the acceptor medium was 1.840.39 mg ml-1. imc is an acidic drug, and its solubility increase is ph-dependent, especially at ph values greater than 4 [40]. this result suggests that in all samples, imc was not saturated in the acceptor medium during the dissolutionpermeation test. the slope of the samples with concentration and time were calculated in the range of 420 to 480 min, in which the dissolution almost attained a plateau, as described in figure 7a. figure 7b shows the permeation profiles of the samples in the range of 420 to 480 min. the slope of crystalline imc was 0.11 (µg ml-1) min-1, while those of the solid dispersion of imc-mg (20:80) and imc-mx (20:80) were 0.21 and 0.20 (µg ml-1) min-1, respectively. the flux values through the membrane were measured for crystalline imc, and the solid dispersions of imc-mg (20:80) and imc-mx (20:80) were measured according to the previously described process [23-24]. the calculated flux values of crystalline imc and the solid dispersions of imc-mg (20:80) and imc-mx (20:80) were 1.38, 2.71, and 2.60 (µg min-1) cm-2, respectively. formulation of the solid dispersions resulted in an approximately two-fold increase in the flux value of imc; no significant differences in the flux values were observed between the two solid dispersions. it is known that the membrane permeability is affected by the formation of drug-rich colloids in the supersaturation state; polymers used as solid dispersion carriers were distributed into drug-rich colloids, which stabilized the drug [36-38]. resistance of diffusion across the unstirred water layer was reduced by the drugrich colloids, resulting in enhanced membrane penetration [24]. the sizes of the drug-rich colloids of the solid dispersions in the acceptor buffer were measured; the average sizes were 140.2 and 148.5 nm for the solid dispersion imc-mg (20:80) and imc-mx (20:80), respectively, showing no significant difference. hence, it is suggested that the effects of hpmcas-mg and hpmcas-mx on drug-rich colloids are comparable, inducing similar permeation profiles and the flux value of imc. the results of the dissolution and permeation tests determined that while hpmcas-mx has a comparable ability to hpmcas-mg as a solid dispersion carrier, it can undergo the hot-melt extrusion process at a lower temperature. d is s o lv e d ( µ g /m l ) time (min) 0 480240 000 100 300 500 200 400 120 360 ● imc-mg (20:80) ● imc-mx (20:80) ● crystalline imc y = 45.959x + 7.0967 r² = 0.9889 y = 44.343x + 0.7864 r² = 0.9879 y = 1.5997x + 0.2152 r² = 0.8211 d is s o lv e d ( µ g /m l ) time (min) 0.0 2.01.0 000 020 080 040 060 0.5 1.5 ● imc-mg (20:80) ● imc-mx (20:80) ● crystalline imc d e si lv e d a m o u n t, m g m l-1 d e si lv e d a m o u n t, m g m l-1 https://doi.org/10.5599/admet.1586 h. tanaka et al. admet & dmpk 00(0) (2023) 000-000 10 (a) (b) time, min time, min figure 7. permeation profiles of crystalline imc and solid dispersions of imc-mg (20:80) and imc-mx (20:80) in the range of (a) 0-480 min and (b) 420-480 min. the error bars represent the standard deviation of n = 3. conclusions this study aimed to evaluate the potential of an experimental grade of hpmcas-mx for the processability of hot-melt extrusion and improvement of dissolution-permeation of imc as a solid dispersion carrier compared to a commercial grade of hpmcas-mg. monitoring during hot-melt extrusion revealed that torque with high-share mixing was reduced depending on the processing time and temperature. the torque of imcmx (20:80) almost attained a plateau at 120 °c, while hpmcas-mg (20:80) did not. raman mapping characterized that elevation of the processing temperature up to 170 °c induced transformation to the solid dispersions for both formulations. these results show that the use of hpmcas-mx, with its lower tg, enables the hot-melt extrusion process to be conducted at a relatively lower temperature. the solid dispersions prepared by the hot-melt extrusion process showed no crystallization for three months at 40 °c. the dissolution test, followed by the permeation test by simultaneous measurement of the donor and acceptor chambers, showed significant enhancement of the dissolution rate and dissolved amount of imc for both the solid dispersions. moreover, the flux value of imc in the solid dispersion was higher than that of crystalline imc, reflecting the improved membrane penetration of the solid dispersions. therefore, it is concluded that hpmcas-mx has a comparable ability to the commercial grade product as a solid dispersion carrier for both solid and liquid phases. this study revealed the applicability of hpmcas-mx in the hot-melt extrusion process. further studies using hpmcas-mx should be conducted to evaluate the effect of the processing temperature on the physicochemical properties of the solid dispersion. in addition, they should investigate the application of our findings to a drug with a high melting temperature. acknowledgments: we would like to thank yoshiichiro aoyama (shin-etsu chemical co., ltd.) and yasuyuki hirama (shin-etsu chemical co., ltd.) for kindly supplying hpmcas-mg and hpmcas-mx and providing information about them. conflict of interest: all authors are employees of shionogi & co., ltd., and there is no competing interest. references [1] j.d. bucheit, l.g. pamulapati, n. carter, k. malloy, d.l. dixon, e.m. sisson. oral semaglutide: a review of the first oral glucagon-like peptide 1 receptor agonist. diabetes technology and therapeutics 22 (2020) 10-18. http://doi.org/10.1089/dia.2019.0185. [2] p.g. dougherty, a. sahni, d. pei. understanding cell penetration of cyclic peptides. chemical reviews 119 (2019) 10241-10287. https://doi.org/10.1021/acs.chemrev.9b00008. p e rm e a te d (µ g /m l ) time (min) 0 480120 360 000 020 060 100 040 080 240 ● imc-mg (20:80) ● imc-mx (20:80) ● crystalline imc y = 0.2001x 16.405 r² = 0.9992 y = 0.2083x 16.88 r² = 0.9989 y = 0.1062x 9.3842 r² = 0.9957 p e rm e a te d ( µ g /m l ) time (min) 420 480435 465 000 020 060 100 040 080 450 ● imc-mg (20:80) ● imc-mx (20:80) ● crystalline imcp e rm e a te d a m o u n t, m g m l-1 p e rm e a te d a m o u n t, m g m l-1 http://doi.org/10.1089/dia.2019.0185 https://doi.org/10.1021/acs.chemrev.9b00008 admet & dmpk 00(0) (2023) 000-000 dissolution-permeation of hot-melt extruded asd doi: https://doi.org/10.5599/admet.1586 11 [3] j. han, y. wei, y. lu, r. wang, j. zhang, y. gao, s. qian. co-amorphous systems for the delivery of poorly water-soluble drugs: recent advances and an update. expert opinion on drug delivery 17 (2020) 1411-1435. https://doi.org/10.1080/17425247.2020.1796631. [4] h.d. williams, n.l. trevaskis, s.a. charman, r.m. shanker, w.n. charman, c.w. pouton, c.j. porter. strategies to address low drug solubility in discovery and development. pharmacological reviews 65 (2013) 315-499. https://doi.org/10.1124/pr.112.005660. [5] p. pandi, r. bulusu, n. kommineni, w. khan, m. singh. amorphous solid dispersions: an update for preparation, characterization, mechanism on bioavailability, stability, regulatory considerations and marketed products. international journal of pharmaceutics 586 (2020) 119560. https://doi.org/ 10.1016/j.ijpharm.2020.119560. [6] s. alshehri, s.s. imam, a. hussain, m.a. altamimi, n.k. alruwaili, f. alotaibi, a. alanazi, f. shakeel. potential of solid dispersions to enhance solubility, bioavailability, and therapeutic efficacy of poorly water-soluble drugs: newer formulation techniques, current marketed scenario and patents. drug delivery 27 (2020) 1625-1643. https://doi.org/10.1080/10717544.2020.1846638. [7] a.b. anane-adjei, e. jacobs, s.c. nash, s. askin, r. soundararajan, m. kyobula, j. booth, a. campbell. amorphous solid dispersions: utilization and challenges in preclinical drug development within astrazeneca. international journal of pharmaceutics 614 (2022) 121387. https://doi.org/10.1016/ j.ijpharm.2021.121387. [8] m.f. simões, r.m.a. pinto, s. simões. hot-melt extrusion in the pharmaceutical industry: toward filing a new drug application. drug discovery today 24 (2019) 1749-1768. https://doi.org/10.1016/ j.drudis.2019.05.013. [9] n. mendonsa, b. almutairy, v.r. kallakunta, s. sarabu, p. thipsay, s. bandari, m.a. repka. manufacturing strategies to develop amorphous solid dispersions: an overview. journal of drug delivery science and technology 55 (2020) 101459. https://doi.org/10.1016/j.jddst.2019.101459. [10] a.t.m. serajuddin. challenges, current status and emerging strategies in the development of rapidly dissolving fdm 3d-printed tablets: an overview and commentary. admet dmpk. 11 (2023) 33-55. https://doi.org/10.5599/admet.1622 [11] k. deboyace, m. bookwala, i.s. buckner, d. zhou, p.l.d. wildfong. interpreting the physicochemical meaning of a molecular descriptor which is predictive of amorphous solid dispersion formation in polyvinylpyrrolidone vinyl acetate. molecular pharmaceutics 19 (2022) 303-317. https://doi.org/10.1021/acs.molpharmaceut.1c00783. [12] b. tian, x. tang, l.s. taylor. investigating the correlation between miscibility and physical stability of amorphous solid dispersions using fluorescence-based techniques. molecular pharmaceutics 13 (2016) 3988-4000. https://doi.org/10.1021/acs.molpharmaceut.6b00803. [13] f. qian, j. wang, r. hartley, j. tao, r. haddadin, n. mathias, m. hussain. solution behavior of pvp-va and hpmc-as-based amorphous solid dispersions and their bioavailability implications. pharmaceutical research 29 (2012) 2765-2776. https://doi.org/10.1007/s11095-012-0695-7. [14] a. mitra, l. li, p. marsac, b. marks, z. liu, c. brown. impact of polymer type on bioperformance and physical stability of hot melt extruded formulations of a poorly water soluble drug. international journal of pharmaceutics 505 (2016) 107-114. https://doi.org/10.1016/j.ijpharm.2016.03.036. [15] h. konno, l.s. taylor. influence of different polymers on the crystallization tendency of molecularly dispersed amorphous felodipine. journal of pharmaceutical sciences 95 (2006) 2692-2705. https://doi.org/10.1002/jps.20697. [16] d. huang, z. xie, q. rao, e. liamas, p. pan, s. guan, z.j. zhang, m. lu, q. li. hot melt extrusion of heatsensitive and high melting point drug: inhibit the recrystallization of the prepared amorphous drug during extrusion to improve the bioavailability. international journal of pharmaceutics 565 (2019) 316-324. https://doi.org/10.1016/j.ijpharm.2019.04.064. [17] a. kapourani, t. tzakri, v. valkanioti, k.n. kontogiannopoulos, p. barmpalexis. drug crystal growth in ternary amorphous solid dispersions: effect of surfactants and polymeric matrix-carriers. international journal of pharmaceutics: x 3 (2021) 100086. https://doi.org/10.1016/j.ijpx.20 21.100086. https://doi.org/10.5599/admet.1586 https://doi.org/10.1080/17425247.2020.1796631 https://doi.org/10.1124/pr.112.005660 https://doi.org/‌10.1016/j.ijpharm.2020.119560 https://doi.org/‌10.1016/j.ijpharm.2020.119560 https://doi.org/10.1080/10717544.2020.1846638 https://doi.org/10.1016/‌j.ijpharm.2021.121387 https://doi.org/10.1016/‌j.ijpharm.2021.121387 https://doi.org/10.1016/‌j.drudis.2019.05.013 https://doi.org/10.1016/‌j.drudis.2019.05.013 https://doi.org/10.1016/j.jddst.2019.101459 https://doi.org/10.5599/admet.1622 https://doi.org/10.1021/acs.molpharmaceut.1c00783 https://doi.org/10.1021/acs.molpharmaceut.6b00803 https://doi.org/10.1007/s11095-012-0695-7 https://doi.org/10.1016/j.ijpharm.2016.03.036 https://doi.org/10.1002/jps.20697 https://doi.org/10.1016/j.ijpharm.2019.04.064 https://doi.org/10.1016/j.ijpx.20‌21.100086 https://doi.org/10.1016/j.ijpx.20‌21.100086 h. tanaka et al. admet & dmpk 00(0) (2023) 000-000 12 [18] n. siriwannakij, t. heimbach, a.t.m. serajuddin. aqueous dissolution and dispersion behavior of polyvinylpyrrolidone vinyl acetate-based amorphous solid dispersion of ritonavir prepared by hotmelt extrusion with and without added surfactants. journal of pharmaceutical sciences 110 (2021) 1480-1494. https://doi.org/10.1016/j.xphs.2020.08.007. [19] h. ueda, y. hirakawa, h. tanaka, t. miyano, k. sugita. applicability of an experimental grade of hydroxypropyl methylcellulose acetate succinate as a carrier for formation of solid dispersion with indomethacin. pharmaceutics 13 (2021) 353. https://doi.org/10.3390/pharmaceutics13030353. [20] m. imono, h. uchiyama, h. ueda, k. kadota, y. tozuka. in-situ dissolution and permeation studies of nanocrystal formulations with second-derivative uv spectroscopy. international journal of pharmaceutics 558 (2019) 242-249. https://doi.org/10.1016/j.ijpharm.2018.12.086. [21] e. borbás, z.k. nagy, b. nagy, a. balogh, b. farkas, o. tsinman, k. tsinman, b. sinkó. the effect of formulation additives on in vitro dissolution-absorption profile and in vivo bioavailability of telmisartan from brand and generic formulations. european journal of pharmaceutical sciences 114 (2018) 310-317. https://doi.org/10.1016/j.ejps.2017.12.029. [22] j. li, k. tsinman, o. tsinman, l. wigman. using ph gradient dissolution with in-situ flux measurement to evaluate bioavailability and ddi for formulated poorly soluble drug products. aaps pharmscitech. 19 (2018) 2898-2907. https://doi.org/10.1208/s12249-018-1164-3. [23] m. imono, h. uchiyama, s. yoshida, s. miyazaki, n. tamura, h. tsutsumimoto, k. kadota, y. tozuka. the elucidation of key factors for oral absorption enhancement of nanocrystal formulations: in vitroin vivo correlation of nanocrystals. european journal of pharmaceutics and biopharmaceutics 146 (2020) 84-92. https://doi.org/10.1016/j.ejpb.2019.12.002. [24] a.m. stewart, m.e. grass, t.j. brodeur, a.k. goodwin, m.m. morgen, d.t. friesen, d.t. vodak. impact of drug-rich colloids of itraconazole and hpmcas on membrane flux in vitro and oral bioavailability in rats. molecular pharmaceutics 14 (2017) 2437-2449. https://doi.org/10.1021/acs.molpharma ceut.7b00338. [25] i. ozgüney, d. shuwisitkul, r. bodmeier. development and characterization of extended release kollidon sr mini-matrices prepared by hot-melt extrusion. european journal of pharmaceutics and biopharmaceutics 73 (2009) 140-145. https://doi.org/10.1016/j.ejpb.2009.04.006. [26] p.a. priemel, r. laitinen, h. grohganz, t. rades, c.j. strachan. in situ amorphisation of indomethacin with eudragit® e during dissolution. european journal of pharmaceutics and biopharmaceutics 85 (2013) 1259-1265. https://doi.org/10.1016/j.ejpb.2013.09.010. [27] h. wickström, m. palo, k. rijckaert, r. kolakovic, j.o. nyman, a. määttänen, p. ihalainen, j. peltonen, n. genina, t. de beer, k. löbmann, t. rades, n. sandler. improvement of dissolution rate of indomethacin by inkjet printing. european journal of pharmaceutical sciences 75 (2015) 91-100. https://doi.org/10.1016/j.ejps.2015.03.009. [28] k. holzapfel, j. liu, t. rades, c.s. leopold. (co-)amorphization of enantiomers investigation of the amorphization process, the physical stability and the dissolution behavior. international journal of pharmaceutics 616 (2022) 121552. https://doi.org/10.1016/j.ijpharm.2022.121552. [29] y. li, a.k.p. mann, d. zhang, z. yang. processing impact on in vitro and in vivo performance of solid dispersionsa comparison between hot-melt extrusion and spray drying. pharmaceutics 13 (2021) 1307. https://doi.org/10.3390/pharmaceutics13081307. [30] v.r. kallakunta, s. sarabu, s. bandari, a. batra, v. bi, t. durig, m.a. repka. stable amorphous solid dispersions of fenofibrate using hot melt extrusion technology: effect of formulation and process parameters for a low glass transition temperature drug. journal of drug delivery science and technology 58 (2020) 101395. https://doi.org/10.1016/j.jddst.2019.101395. [31] h. ueda, y. ida, k. kadota, y. tozuka. raman mapping for kinetic analysis of crystallization of amorphous drug based on distributional images. international journal of pharmaceutics 462 (2014) 115-22. https://doi.org/10.1016/j.ijpharm.2013.12.025. [32] j. lu, s. obara, f. liu, w. fu, w. zhang, s. kikuchi. melt extrusion for a high melting point compound with improved solubility and sustained release. aaps pharmscitech 19 (2018) 358-370. https://doi.org/10.1208/s12249-017-0846-6. https://doi.org/10.1016/j.xphs.2020.08.007 https://doi.org/10.3390/pharmaceutics13030353 https://doi.org/10.1016/j.ijpharm.2018.12.086 https://doi.org/10.1016/j.ejps.2017.12.029 https://doi.org/10.1208/s12249-018-1164-3 https://doi.org/10.1016/j.ejpb.2019.12.002 https://doi.org/10.1021/acs.molpharma‌ceut.7b00338 https://doi.org/10.1021/acs.molpharma‌ceut.7b00338 https://doi.org/10.1016/j.ejpb.2009.04.006 https://doi.org/10.1016/j.ejpb.2013.09.010 https://doi.org/10.1016/j.ejps.2015.03.009 https://doi.org/10.1016/j.ijpharm.2022.121552 https://doi.org/10.3390/pharmaceutics13081307 https://doi.org/10.1016/j.jddst.2019.101395 https://doi.org/10.1016/j.ijpharm.2013.12.025 https://doi.org/10.1208/s12249-017-0846-6 admet & dmpk 00(0) (2023) 000-000 dissolution-permeation of hot-melt extruded asd doi: https://doi.org/10.5599/admet.1586 13 [33] k. kawakami. crystallization tendency of pharmaceutical glasses: relevance to compound properties, impact of formulation process, and implications for design of amorphous solid dispersions. pharmaceutics 11 (2019) e202. https://doi.org/10.3390/pharmaceutics11050202. [34] k. kawakami, t. harada, k. miura, y. yoshihashi, e. yonemochi, k. terada, h. moriyama. relationship between crystallization tendencies during cooling from melt and isothermal storage: toward a general understanding of physical stability of pharmaceutical glasses. molecular pharmaceutics 11 (2014) 1835-1843. https://doi.org/10.1021/mp400679m. [35] j.h. fagerberg, o. tsinman, n. sun, k. tsinman, a. avdeef, c.a. bergström. dissolution rate and apparent solubility of poorly soluble drugs in biorelevant dissolution media. molecular pharmaceutics 7 (2010) 1419-1430. https://doi.org/10.1021/mp100049m. [36] a.l. sarode, h. sandhu, n. shah, w. malick, h. zia. hot melt extrusion (hme) for amorphous solid dispersions: predictive tools for processing and impact of drug-polymer interactions on supersaturation. european journal of pharmaceutical sciences 48 (2013) 371-384. https://doi.org/ 10.1016/j.ejps.2012.12.012. [37] m.j. jackson, u.s. kestur, m.a. hussain, l.s. taylor. dissolution of danazol amorphous solid dispersions: supersaturation and phase behavior as a function of drug loading and polymer type. molecular pharmaceutics 13 (2016) 223-231. https://doi.org/10.1021/acs.molpharmaceut.5b00652. [38] k. ueda, s.s. hate, l.s. taylor. impact of hypromellose acetate succinate grade on drug amorphous solubility and in vitro membrane transport. journal of pharmaceutical sciences 109 (2020) 24642473. https://doi.org/10.1016/j.xphs.2020.04.014. [39] a.n. bristol, m.s. lamm, y. li. impact of hydroxypropyl methylcellulose acetate succinate critical aggregation concentration on celecoxib supersaturation. molecular pharmaceutics 18 (2021) 42994309. https://doi.org/10.1021/acs.molpharmaceut.1c00372. [40] a. avdeef. cocrystal solubility product analysis dual concentration-ph mass action model not dependent on explicit solubility equations. european journal of pharmaceutical sciences 15 (2017) 218. https://doi.org/10.1016/j.ejps.2017.03.049. ©2023 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) https://doi.org/10.5599/admet.1586 https://doi.org/10.3390/pharmaceutics11050202 https://doi.org/10.1021/mp400679m https://doi.org/10.1021/mp100049m https://doi.org/‌10.1016/j.ejps.2012.12.012 https://doi.org/‌10.1016/j.ejps.2012.12.012 https://doi.org/10.1021/acs.molpharmaceut.5b00652 https://doi.org/10.1016/j.xphs.2020.04.014 https://doi.org/10.1021/acs.molpharmaceut.1c00372 https://doi.org/10.1016/j.ejps.2017.03.049 http://creativecommons.org/licenses/by/3.0/ preparation of self-assembly silica redox nanoparticles to improve drug encapsulation and suppress the adverse effect of doxorubicin doi: https://doi.org/10.5599/admet.1845 1 admet & dmpk 0(0) (2023) 000-000; doi: https://doi.org/10.5599/admet.1845 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper preparation of self-assembly silica redox nanoparticles to improve drug encapsulation and suppress the adverse effect of doxorubicin minh-dat quoc tang1,2,#, nhu-thuy trinh1,2,#, dung vu1,2, thu-ha thi nguyen1,2, hung thanh dong1,2, toi van vo1,2, long binh vong1,2,* 1school of biomedical engineering, international university, ho chi minh 700000, vietnam 2vietnam national university ho chi minh city (vnu-hcm), ho chi minh 700000, vietnam #equal contribution *corresponding author: e-mail: vblong@hcmiu.edu.vn; tel.: +81 29-853-5749 fax: +81 29-853-5749 received: april 27, 2023; revised: june 29, 2023; published: june xx,2023 abstract background and purpose: the utilization of doxorubicin (dox) in clinal trials is also challenging owing to its adverse effects, including low oral bioavailability, generation of reactive oxygen species (ros), cardiotoxicity, and epithelial barrier damage. recently, scavenging of ros reduced the cytotoxicity of dox, suggesting a new approach for using dox as an anticancer treatment. thus, in this study, non-silica and silica redox nanoparticles (denoted as rnpn and sirnp, respectively) with ros scavenging features have been designed to encapsulate dox and reduce its cytotoxicity. experimental approach: dox-loaded rnpn (dox@rnpn) and dox-loaded sirnp (dox@sirnp) were prepared by co-dissolving dox with rnpn and sirnp, respectively. the size and stability of nanoparticles were characterized by the dynamic light scattering system. additionally, encapsulation efficiency, loading capacity, and release profile of dox@rnpn and dox@sirnp were identified by measuring the absorbance of dox. finally, the cytotoxicity of dox@rnp n and dox@sirnp against normal murine fibroblast cells (l929), human hepatocellular carcinoma cells (hepg2), and human breast cancer cells (mcf-7) were also investigated. key results: the obtained result showed that rnpn exhibited a ph-sensitive character while silanol moieties improved the stability of sirnp in physiological conditions. dox@rnpn and dox@sirnp were formed at several tens of nanometers in diameter with narrow distribution. moreover, dox@sirnp stabilized under different ph buffers, especially gastric ph, and improved encapsulation of dox owing to the addition of silanol groups. dox@rnpn and dox@sirnp maintained anticancer activity of dox against hepg2, and mcf-7 cells, while their cytotoxicity on l929 cells was significantly reduced compared to free dox treatment. conclusion: dox@rnpn and dox@sirnp could effectively suppress the adverse effect of dox, suggesting the potential to become promising nanomedicines for cancer treatments. ©2023 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords chemotherapy; reactive oxygen species; ros scavengers; micelles nanoparticles, nanomedicine introduction doxorubicin (dox, also known as adriamycin) has been widely used to treat numerous cancer types since its approval by the fda in 1974 [1]. the most well-known anticancer mechanism of dox is the interaction with the topoisomerases ii, leading to interference with the dna replications. additionally, dox also induces https://doi.org/10.5599/admet.1845 https://doi.org/10.5599/admet.1845 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:vblong@hcmiu.edu.vn http://creativecommons.org/licenses/by/4.0/ m.-d. quoc tang et al. admet & dmpk 00(0) (2023) 000-000 2 reactive oxygen species (ros) via nonenzymatic and enzymatic pathways [2,3]. consequently, these free radicals cause intracellular oxidative stress, and damage cellular membranes and dna, although this mechanism is not considered to play a major role in killing cancer cells [4]. unfortunately, dox is non-specific distribution due to its low molecular weight drug nature, resulting in myocyte, myelocyte, and hematocyte toxicities [5-7]. dox-induced damage on intestinal epithelium has also been reported [8]. besides, the low oral bioavailability of dox was reported at 1 % due to its absorption through the paracellular pathway [9]. lowering the cytotoxicity of dox and enhancement of its oral bioavailability have gained attention to extend the clinical utility of dox. recently, scavengers of free radicals showed prevention of dox-induced cardiotoxicity [10], implying the important role of ros scavengers in reducing the adverse effects of dox. nanoparticles (nps) have been widely developed as carriers for drug delivery systems (dds) to decrease unwanted diffusion, increase drug bioavailability, and control pharmacokinetics [11-13]. oral administration is one of the most popular pathways to deliver nps owing to its unique advantages like non-invasion, painlessness, easy administration, and patient compliance. the various nps, such as liposomes, metal, and polymeric nps, have been designed for oral drug delivery. however, most techniques still face numerous challenges, such as gastric juice influencing the aggregation state of liposomes [14], decreasing migration, retention in the intestinal mucus layer, and low blood uptake of the therapeutic agents. moreover, the activation of various enzymes by oral administration also causes instability and leakage of payload, suggesting an efficient reduction of drug carriers and toxicity owing to the exposure of the drug on the gastrointestinal (gi) tract. therefore, the requirements for oral dds of cancer treatment are stable dispersion, drug encapsulation in the harsh gi tract, and low toxicity. recently, polymeric nps have been applied to improve the oral bioavailability of chemotherapy. for example, dox-encapsulated enoxaparin sodium-plga hybrid nps significantly increased retention time in the pharmacokinetics study [15]. moreover, polymeric micelles, a type of polymeric nps with a core-shell structure, can increase the stability, and permeability in the intestinal epithelium and reduce the degradation of anti-cancer drugs. for instance, dox-loaded polymeric micelles significantly improved the intestinal absorption rate and systemic circulation time as compared to free dox [16]. thus, micelles nps can be considered potential anti-cancer dds in oral administration. recently, we have designed two types of core-shell antioxidant nps, ph non-sensitive (rnpo) and phsensitive (rnpn), as ideal oral dds to solve these problems. rnpo and rnpn were prepared through selfassembly amphiphilic block copolymers containing ros scavengers (nitroxide radicals, tempo) at the side chains of the hydrophobic segment. the ros-scavenging ability of rnps is an important factor in decreasing the dox-induced oxidation stress. both rnps indicated highly dispersible and biocompatible properties with a long half-life in circulation compared to free nitroxide radicals. in addition, rnps have been researched as possible treatments for solid tumours [17,18]. although rnpo prolonged blood circulation as compared to rnpn [19,20], rnpo was not appropriated in encapsulating drugs inside the core [21]. in contrast, gastric ph is also challenging for rnpn due to its ph sensitivity caused by the protonation of amino groups on the hydrophobic core. therefore, we have developed silica-installed redox nps (sirnp) through sol-gel chemistry by the hydrolysis and condensation of tetraethyl orthosilicate (teos) with the presence of ammonia during the preparation of rnpn. crosslinking of silanol groups is expected to improve the stability of nps and the drug encapsulation efficacy via adsorption on the silica surface and electrostatic interactions between dox and silica. in this work, we studied the impacts of rnpn and sirnp loading dox (denoted as dox@rnpn and dox@sirnp, respectively) to reduce the side effects of dox on normal cells and to maintain the cytotoxicity properties of dox on cancer cells. the obtained result indicated that nps were nano-size distribution and high dispersion in the physiological conditions. moreover, the addition of silanol groups significantly improved the encapsulation of dox and the stability of rnpn in gastric ph. compared to free dox treatment, the dox@rnpn and dox@sirnp showed admet & dmpk 00(0) (2023) 000-000 drug encapsulation and suppress the adverse effect of doxorubicin doi: https://doi.org/10.5599/admet.1845 3 lower cytotoxicity against murine fibroblast cells (l929), while they maintained the cytotoxicity on human breast cancer cells (mcf-7) and human hepatocellular carcinoma cells (hepg2). experimental chemicals dimethylformamide (dmf, sigma-aldrich, usa), permeable membrane tube (mwco 3.5 kda, spectrum laboratories inc., japan), tetraethyl orthosilicate (teos, sigma-aldrich, usa), ammonia (nh3, china), doxorubicin hydrochloride (dox, wako, japan), 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (mtt, roche diagnostics, japan), dimethyl sulfoxide (dmso, china), dulbecco’s modified eagle’s medium (dmem, gibco, usa), fetal bovine serum (fbs, sigma-aldrich, usa), and antibiotics (a mixture of penicillin, streptomycin, and neomycin, sigma-aldrich, usa) were purchased. preparation and characterization of rnpn and sirnp rnpn was prepared by self-assembly amphiphilic copolymers peg-b-pmnt, as reported in previous research [18]. briefly, 15 mg of peg-b-pmnt were dissolved in 1 ml of dmf and then stirred on a magnetic stirrer. the mixture was then put into a semi-permeable membrane tube before being dialyzed against distilled water for 24 h. to prepare sirnp, a similar method was conducted, except that 50 µl of teos and 50 µl of nh3 were added during the stirring process (figure 1). the average size, polydispersity index (pdi), and size distribution of rnpn and sirnp were then characterized by the dynamic light scattering (dls, malvern zetasizer, uk) system. figure 1. illustration of dox@rnpn and dox@sirnp preparation. dox@rnpn was prepared by co-dissolving amphiphilic copolymers peg-b-pmnt and dox while dox@sirnp was similarly prepared, except teos was added to form silica nps under the catalyst of nh3 during the stirring process. the stability of rnpn and sirnp phosphate-buffered saline (pbs, 10 mm) was prepared at different phs, including 3.0, 6.5, and 7.4, to mimic the ph levels of the gastric, tumour extracellular, and physiological environments, respectively. rnpn https://doi.org/10.5599/admet.1845 m.-d. quoc tang et al. admet & dmpk 00(0) (2023) 000-000 4 and sirnp were then diluted in the prepared buffers. the stability of rnpn and sirnp was evaluated by measuring the change in light scattering intensity using the dls system at 25 °c for 2 h. preparation and characterization of dox@rnpn and dox@sirnp dox@rnpn and dox@sirnp were prepared by co-dissolving 1 mg of dox and 15 mg of peg-b-pmnt in 1 ml of dmf. similar steps were followed in the preparation of rnpn and sirnp, respectively (figure 1). additionally, various ratios of teos:nh3 (200 μl teos: 200 μl nh3, 100 μl teos: 100 μl nh3, 50 μl teos: 50 μl nh3, 100 μl teos: 50 μl nh3, 200 μl teos: 50 μl nh3) were examined to optimize the size of dox@sirnp. the dls system was used to characterize the average size, pdi, and size distribution of dox@rnpn and dox@sirnp. drug encapsulated efficiency and loading capacity of dox@rnpn and dox@sirnp the encapsulation efficiency (ee / %) and loading capacity (lc / %) of dox@rnpn and dox@sirnp were investigated based on the amount of loaded dox evaluated by a microplate reader (thermo fisher scientific, usa) with a calibration curve of dox measured absorbance at 480 nm. the ee and lc were calculated by using equation (1) and (2): encaplulated theoretical dox total mass 100 total mass ee = (1) encaplulated dox theoretical polymers total mass 100 total mass lc = (2) release profile of dox@rnpn and dox@sirnp 2 ml of dox@rnpn or dox@sirnp were loaded into separate dialysis bags and placed in different beakers under stirring conditions, each with distinct ph including 3, 6.5, and 7.4. at predetermined time points (0, 1, 3, 6, 9, 20, and 24 h), samples were collected and the absorbance at 480 nm was measured to evaluate the amount of drug released from the nps. cytotoxic assay the murine fibroblast cells (l929), human breast cancer cells (mcf-7), and human hepatocellular carcinoma cells (hepg2) were obtained from the american type culture collection (atcc, usa). the cells were cultured in dmem containing 10 % fbs and 1 % antibiotics (a mixture of penicillin, streptomycin, and neomycin) at 37 °c and 5 % co2. the cells were then seeded on 96-well plates (104 cells/well) and incubated for 24 h at 37 °c and 5 % co2. the tested samples, including rnpn, sirnp, dox, dox@rnpn, and dox@sirnp, were added to each well to obtain final concentrations of dox at 0.25, 0.5, 1, 2.5, and 5 µg/ml, and incubated for 24 h. after that, the medium was removed, and mtt solution was added to each well and continuously incubated for 4 h. finally, dmso was added to dissolve crystals of formazan inside the cells, and the plates were measured absorbance at 540 nm by a microplate reader (thermo fisher scientific, usa). the cell viability (%) was calculated by using equation (3): sample control od cell viability 100 od = (3) where odsample is the absorbance of the sample and odcontrol is the absorbance of the solvent. statistical analysis statistical analysis was performed using one-way analysis of variance (anova) to determine significant differences among various groups. a p-value <0.05 was considered statistically significant. data were presented as the mean ± standard deviation (sd). admet & dmpk 00(0) (2023) 000-000 drug encapsulation and suppress the adverse effect of doxorubicin doi: https://doi.org/10.5599/admet.1845 5 results and discussion characterization of rnpn and sirnp rnpn and sirnp were prepared by self-assembly amphiphilic copolymers peg-b-pmnt in dmf, followed by dialysis against distilled water for 24 h to remove the organic solvent. the results showed that both rnpn and sirnp were transparent (figure 2a), with average sizes of 45.5 and 85.5 nm, respectively (figure 2b). furthermore, the pdi of both nps was less than 0.3, indicating a narrow size distribution. the increase in the size of sirnp as compared to rnpn might be explained by a sol-gel reaction of teos in the rnpn core [22]. the size of sirnp was approximately 85 nm after the condensation reaction, suggesting no significant aggregation (figure 2c), even though silica particles formed by sol-gel can reach up to 2000 nm. we next examined the influence of different ph buffers on the stability of nps [23]. as shown in figure 2d, rnpn disintegrated under acidic ph (3.0 and 6.5) and stabilized at ph 7.4. in contrast, sirnp was stable regardless of ph change. this might be explained that the amino groups of the pmnt segment disintegrated due to their protonation at low ph. however, adding silica would maintain the structure of nps at acidic ph, resulting in enhanced stability. figure 2. characterization of rnpn and sirnp. the micelles were obtained by self-assembly amphiphilic copolymers peg-b-pmnt and the addition of silanol groups through the dialysis method for rnpn and sirnp, respectively, and then characterized by the dls system. (a) the photographs after dialysis for 24 h. (b) average size. (c) size distribution. (d) the stability of rnpn and sirnp in pbs with different phs, including 3, 6.5, and 7.4 simulated for the gastric, tumor extracellular, and physiological environments, respectively. data were presented as the mean ± standard deviation (sd); n =3; p < 0.05 (*); ns = non-significant. characterization of dox@rnpn and dox@sirnp dox@rnpn and dox@sirnp were similarly prepared as rnpn and sirnp, respectively, except for the addition of dox during the preparation process. different concentrations of nh3 might result in a change in the size of silica nps despite its catalyzation role in sol-gel reactions [24]. thus, different ratios of teos:nh3 were examined to prepare dox@sirnp with the nano-size distribution. as shown in figure 3a, dox@rnpn, dox@sirnp with ratios of 50 μl teos: 50 μl nh3 and 100 μl teos: 50 μl nh3 were transparent, resulting in https://doi.org/10.5599/admet.1845 m.-d. quoc tang et al. admet & dmpk 00(0) (2023) 000-000 6 the nano-size distribution as compared to other samples with turbidity in solution and micro-size distribution (data not shown). hence, dox@sirnp with a ratio of 50 μl teos: 50 μl nh3 was chosen for further experiments because of its suitable size. as compared to nps, the size of both dox@rnpn and dox@sirnp significantly increased to 56 and 90 nm, respectively, implying the encapsulation of dox in nps (figure 3b). the critical cut-off point for endocytosis has been estimated from 20 to 500 nm for m cells and from approximately 50 to around 100 nm for enterocyte cells [25]. therefore, dox@rnpn and dox@sirnp showed the potential to increase the oral bioavailability of dox. moreover, the pdi of both nps was around 0.3, indicating the narrow size distribution even after drug encapsulation (figure 3c). conventionally, polymeric nps exhibit a low ee by physical entrapment of the drug (less than 10 %) [26-28]. for example, the lc of dox-loaded silica nps was reported at 4.8 % [29]. as shown in figure 3d, the ee was 20.3 and 74.9 %, and the lc was 1.5 and 5 % for dox@rnpn and dox@sirnp, respectively. this result indicates the appropriate design of the dox encapsulation because the silanol groups in the core of sirnp significantly improved the ee and lc of dox. this result could be explained by the pka of the amino group in dox being around 9.9, thus most dox molecules are in the positively charged form at physiological ph [30]. consequently, silanol moieties possessing a negatively charged surface could interact with dox by electrostatics, suggesting the enhancement in ee and lc of dox@sirnp. figure. 3. characterization of dox@rnpn and dox@sirnp. the method for preparation of dox@rnpn and dox@sirnp was similar to the method of rnpn and sirnp except for the incorporation of dox during the stirring process. (a) the photographs of dox@rnpn and dox@sirnp with different ratios of teos:nh3 after dialysis for 24 h. (b) average size. (c) size distribution. (d) the ee and lc of dox@rnpn and dox@sirnp were obtained by measuring the absorbance at 480 nm. data were presented as the mean ± standard deviation (sd); n =3; p < 0.05 (*); p < 0.01 (**). release profile of dox@rnpn and dox@sirnp the release kinetics of dox@rnpn and dox@sirnp were evaluated under different ph buffers, including ph 3.0, 6.5, and 7.4, which mimic the gastric, tumour extracellular, and physiological ph, respectively. gastric ph is one of the challenging barriers for oral drug delivery systems. as shown in figure 4a, dox@sirnp was released significantly slower than dox@rnpn (17.8 % compared to 35.9 %) after 24 h, even though both admet & dmpk 00(0) (2023) 000-000 drug encapsulation and suppress the adverse effect of doxorubicin doi: https://doi.org/10.5599/admet.1845 7 experienced protonations of amino groups at the hydrophobic core. however, more than 80 % of dox@rnpn and 90 % of dox@sirnp were retained after 6 h, suggesting their potential for in vivo application as they can withstand the maximal transit time of the stomach in the human body. the extracellular ph of tumors, which commonly drops around 6.5-7.0, is mainly caused by anaerobic glycolysis in hypoxia [31]. as shown in figure 4b, the release profiles of dox@sirnp and dox@rnpn at ph 6.5 were 16.4 and 33.0 %, respectively. these results are consistent with the previous stability study. it is evident that dox@sirnp was less impacted by acidic ph than dox@rnpn due to the ability of nano-silica to stabilize the structure. after 24 h at physiological ph, dox@rnpn and dox@sirnp leaked at 14 and 15 %, respectively (figure 4c). similarly, several studies have reported that the release profile of micelle nps with ph-sensitive groups is higher at acidic ph than at neutral ph [32,33]. figure 4. the drug release of dox@rnpn and dox@sirnp. the release of dox in pbs with various ph including 3.0, 6.5, and 7.4 was identified by measuring the absorbance at 480 nm. (a) ph 3.0. (b). ph 6.5. (c) ph 7.4. data were presented as the mean ± standard deviation (sd); n =3; p < 0.01 (**); ns = non-significant. in vitro cytotoxicity the cytotoxicity of nps was assessed through mtt assay on normal murine fibroblast cells (l929) and cancer cells, including human breast cancer cells (mcf-7) and human hepatocellular carcinoma cells (hepg2). in addition to the well-known anticancer mechanism, which involves interaction with the topoisomerases ii to interfere with dna replication, it also induces the over-expression of free radicals to kill cells. therefore, rnpn with tempol, a ros scavenger in the hydrophobic core, was used to encapsulate dox to reduce its side effects in an in vitro model. as shown in figure 5a, dox indicated dose-dependence, with approximately 50 % of l929 cell death at a concentration of 5 µg/ml dox, while the nps significantly enhanced cell viability, almost equal to the control sample. despite dose-dependent toxicity observed in dox@rnpn and dox@sirnp, there was more than 90 % of cell viability after 24 h of incubation due to the scavenging ability of tempo groups. in addition, silica nps have been reported to cause cytotoxicity due to the overproduction of ros [34], leading to lower cell viability of sirnp compared to rnpn. however, sirnp treatment exhibited more than 90 % of cell viability, suggesting the biosafe profile of nps. similarly, anticancer effects of dox were observed on human breast cancer cells (mcf-7) and human hepatocellular carcinoma cells (hepg2), with dose dependence (figure 5b and figure 5c). remarkably, dox@sirnp caused fewer cell deaths (around 20 and 25 %) compared to (27 and 32 %) at 5 µg/ml of free dox treatment against mcf-7 and hepg2, respectively. this result may be due to dox being kept inside the core of sirnp, resulting in less exposure to cells, although encapsulation of dox in nps was expected to be more effective at entering cells. in contrast, the anticancer effect caused by dox@rnpn (5 µg/ml of dox) was 36 and 37 % for mcf-7 and hepg2, respectively. it is obvious that the cumulative release of dox@rnpn was faster as compared to dox@sirnp, resulting in higher cytotoxicity when entering cells. https://doi.org/10.5599/admet.1845 m.-d. quoc tang et al. admet & dmpk 00(0) (2023) 000-000 8 figure 5. the cytotoxicity on various cells. the cell viability was measured by mtt assay. (a) murine fibroblast cells (l929). (b) human breast cancer cells (mcf-7). (c) human hepatocellular carcinoma cells (hepg2). data were presented as the mean ± standard deviation (sd); n =3; p < 0.05 (*). conclusions in this study, non-silica and silica redox nps (rnpn and sirnp, respectively) with ros scavenging moieties on hydrophobic side chains were prepared to encapsulate dox to reduce its adverse effects. the dox@rnpn and dox@sirnp were successfully prepared with several tens nm distributions and stable dispersion. the obtained results indicated that the addition of silanol groups significantly improved the stability of rnpn in acidic ph and the encapsulation efficacy of dox into sirnp. due to its ph-sensitive character, dox@rnpn exhibited a faster release of dox under acidic ph environments as compared to dox@sirnp. both dox@rnpn and dox@sirnp suppressed the cytotoxicity of dox against normal fibroblast cells while they exhibited the anticancer activity to induce cell death in liver cancer and breast cancer cell lines. therefore, dox@rnpn and dox@sirnp could be considered potential nanomedicines for cancer treatments. conflict of interest: the authors have no conflicts of interest to declare. acknowledgements: this work was supported by vietnam national university ho chi minh city (vnu-hcm) under grant numbers ncm2020-28-01. authors would like to thank prof. nagasaki yukio (university of tsukuba, japan) for kindly sharing the polymers used in this study. references [1] j. sun, q. wei, y. zhou, j. wang, q. liu, h. xu. a systematic analysis of fda-approved anticancer drugs. bmc systems biology 11 (2017) 1-17. https://doi.org/10.1186/s12918-017-0464-7 [2] l. gianni, j.l. zweier, a. levy, c. myers. characterization of the cycle of iron-mediated electron transfer from adriamycin to molecular oxygen. journal of biological chemistry 260 (1985) 6820-6826. https://doi.org/10.1016/s0021-9258(18)88854-8 [3] r.d. olson, p.s. mushlin. doxorubicin cardiotoxicity: analysis of prevailing hypotheses. the faseb journal 4 (1990) 3076-3086. https://doi.org/10.1096/fasebj.4.13.2210154 [4] h. keizer, h. pinedo, g. schuurhuis, h. joenje. doxorubicin (adriamycin): a critical review of free radicaldependent mechanisms of cytotoxicity. pharmacology & therapeutics 47 (1990) 219-231. https://doi.org/ 10.1016/0163-7258(90)90088-j [5] m.s. horenstein, r.s. vander heide, t.j. l'ecuyer. molecular basis of anthracycline-induced cardio toxicity and its prevention. molecular genetics and metabolism 71 (2000) 436-444. https://doi.org/ 10.1006/mgme.2000.3043 [6] a.g. patel, s.h. kaufmann. how does doxorubicin work? elife 1 (2012) e00387. https://doi.org/ 10.7554/elife.00387 [7] c.f. thorn, c. oshiro, s. marsh, t. hernandez-boussard, h. mcleod, t.e. klein, r.b. altman. doxorubicin pathways: pharmacodynamics and adverse effects. pharmacogenetics and genomics 21 (2011) 440. https://doi.org/10.1097/fpc.0b013e32833ffb56 https://doi.org/10.1186/s12918-017-0464-7 https://doi.org/10.1016/s0021-9258(18)88854-8 https://doi.org/10.1096/fasebj.4.13.2210154 https://doi.org/‌10.1016/0163-7258(90)90088-j https://doi.org/‌10.1016/0163-7258(90)90088-j https://doi.org/‌10.1006/mgme.2000.3043 https://doi.org/‌10.1006/mgme.2000.3043 https://doi.org/‌10.7554/elife.00387 https://doi.org/‌10.7554/elife.00387 https://doi.org/10.1097/fpc.0b013e32833ffb56 admet & dmpk 00(0) (2023) 000-000 drug encapsulation and suppress the adverse effect of doxorubicin doi: https://doi.org/10.5599/admet.1845 9 [8] r.j. rigby, j. carr, k. orgel, s.l. king, p.k. lund, c.m. dekaney. intestinal bacteria are necessary for doxorubicin-induced intestinal damage but not for doxorubicin-induced apoptosis. gut microbes 7 (2016) 414-423. https://doi.org/10.1080/19490976.2016.1215806 [9] j.-e. kim, h.-j. cho, j.s. kim, c.-k. shim, s.-j. chung, m.-h. oak, i.-s. yoon, d.-d. kim. the limited intestinal absorption via paracellular pathway is responsible for the low oral bioavailability of doxorubicin. xenobiotica 43 (2013) 579-591. https://doi.org/10.3109/00498254.2012.751140 [10] s. kotamraju, e.a. konorev, j. joseph, b. kalyanaraman. doxorubicin-induced apoptosis in endothelial cells and cardiomyocytes is ameliorated by nitrone spin traps and ebselen: role of reactive oxygen and nitrogen species. journal of biological chemistry 275 (2000) 33585-33592. https://doi.org/10.1074/ jbc.m003890200 [11] m. elsabahy, k.l. wooley. design of polymeric nanoparticles for biomedical delivery applications. chemical society reviews 41 (2012) 2545-2561. https://doi.org/10.1039/c2cs15327k [12] n. kamaly, b. yameen, j. wu, o.c. farokhzad. degradable controlled-release polymers and polymeric nanoparticles: mechanisms of controlling drug release. chemical reviews 116 (2016) 2602-2663. https://doi.org/10.1021/acs.chemrev.5b00346 [13] h. otsuka, y. nagasaki, k. kataoka. pegylated nanoparticles for biological and pharmaceutical applications. advanced drug delivery reviews 55 (2012) 403-419. https://doi.org/10.1016/j.addr.20 12.09.022 [14] t.x. nguyen, l. huang, m. gauthier, g. yang, q. wang. recent advances in liposome surface modification for oral drug delivery. nanomedicine 11 (2016) 1169-1185. https://doi.org/10.2217/ nnm.16.9 [15] j. wang, l. li, l. wu, b. sun, y. du, j. sun, y. wang, q. fu, p. zhang, z. he. development of novel selfassembled es-plga hybrid nanoparticles for improving oral absorption of doxorubicin hydrochloride by p-gp inhibition: in vitro and in vivo evaluation. european journal of pharmaceutical sciences 99 (2017) 185-192. https://doi.org/10.1016/j.ejps.2016.12.014. [16] j. wang, l. li, y. du, j. sun, x. han, c. luo, x. ai, q. zhang, y. wang, q. fu. improved oral absorption of doxorubicin by amphiphilic copolymer of lysine-linked ditocopherol polyethylene glycol 2000 succinate: in vitro characterization and in vivo evaluation. molecular pharmaceutics 12 (2015) 463-473. https://doi.org/10.1021/mp500833m [17] l.b. vong, t. yoshitomi, h. matsui, y. nagasaki. development of an oral nanotherapeutics using redox nanoparticles for treatment of colitis-associated colon cancer. biomaterials 55 (2015) 54-63. https://doi.org/10.1016/j.biomaterials.2015.03.037 [18] t. yoshitomi, y. ozaki, s. thangavel, y. nagasaki. redox nanoparticle therapeutics to cancer—increase in therapeutic effect of doxorubicin, suppressing its adverse effect. journal of controlled release 172 (2013) 137-143. https://doi.org/10.1016/j.jconrel.2013.08.011 [19] p. chonpathompikunlert, c.-h. fan, y. ozaki, t. yoshitomi, c.-k. yeh, y. nagasaki. redox nanoparticle treatment protects against neurological deficit in focused ultrasound-induced intracerebral hemorrhage. nanomedicine 7 (2012) 1029-1043. https://doi.org/10.2217/nnm.12.2 [20] t. yoshitomi, a. hirayama, y. nagasaki. the ros scavenging and renal protective effects of ph-responsive nitroxide radical-containing nanoparticles. biomaterials 32 (2011) 8021-8028. https://doi.org/0.1016/ j.biomaterials.2011.07.014 [21] t.-h.t. nguyen, n.-t. trinh, h.n. tran, h.t. tran, p.q. le, d.-n. ngo, h. tran-van, t. van vo, l.b. vong, y. nagasaki. improving silymarin oral bioavailability using silica-installed redox nanoparticle to suppress inflammatory bowel disease. journal of controlled release 331 (2021) 515-524. https://doi.org/ 10.1016/j.jconrel.2020.10.042 [22] t. yoshitomi, y. nagasaki. development of silica-containing redox nanoparticles for medical applications. biomaterials science 3 (2015) 810-815. https://doi.org/10.1039/c5bm00057b [23] h.t. tran, l.b. vong, y. nishikawa, y. nagasaki. sorafenib-loaded silica-containing redox nanoparticles for oral anti-liver fibrosis therapy. journal of controlled release 345 (2022) 880-891. https://doi.org/ 10.1016/j.jconrel.2022.04.002 https://doi.org/10.5599/admet.1845 https://doi.org/10.1080/19490976.2016.1215806 https://doi.org/10.3109/00498254.2012.751140 https://doi.org/10.1074/‌jbc.m003890200 https://doi.org/10.1074/‌jbc.m003890200 https://doi.org/10.1039/c2cs15327k https://doi.org/10.1021/acs.chemrev.5b00346 https://doi.org/10.1016/j.addr.20‌12.09.022 https://doi.org/10.1016/j.addr.20‌12.09.022 https://doi.org/10.2217/‌nnm.16.9 https://doi.org/10.2217/‌nnm.16.9 https://doi.org/10.1016/j.ejps.2016.12.014 https://doi.org/10.1021/mp500833m https://doi.org/10.1016/j.biomaterials.2015.03.037 https://doi.org/10.1016/j.jconrel.2013.08.011 https://doi.org/10.2217/nnm.12.2 https://doi.org/0.1016/‌j.biomaterials.2011.07.014 https://doi.org/0.1016/‌j.biomaterials.2011.07.014 https://doi.org/‌10.1016/j.jconrel.2020.10.042 https://doi.org/‌10.1016/j.jconrel.2020.10.042 https://doi.org/10.1039/c5bm00057b https://doi.org/‌10.1016/j.jconrel.2022.04.002 https://doi.org/‌10.1016/j.jconrel.2022.04.002 m.-d. quoc tang et al. admet & dmpk 00(0) (2023) 000-000 10 [24] m. su, h. su, b. du, x. li, g. ren, s. wang. mesoporous silica with monodispersed pores synthesized from the controlled self-assembly of silica nanoparticles. korean journal of chemical engineering 32 (2015) 852-859. https://doi.org/10.1007/s11814-014-0270-5 [25] d.j. mcclements. edible lipid nanoparticles: digestion, absorption, and potential toxicity. progress in lipid research 52 (2013) 409-423. https://doi.org/10.1016/j.plipres.2013.04.008 [26] s. abbad, c. wang, a.y. waddad, h. lv, j. zhou. preparation, in vitro and in vivo evaluation of polymeric nanoparticles based on hyaluronic acid-poly (butyl cyanoacrylate) and d-alpha-tocopheryl polyethylene glycol 1000 succinate for tumor-targeted delivery of morin hydrate. international journal of nanomedicine 10 (2015) 305. https://doi.org/10.2147/ijn.s73971 [27] j. li, z. li, t. zhou, j. zhang, h. xia, h. li, j. he, s. he, l. wang. positively charged micelles based on a triblock copolymer demonstrate enhanced corneal penetration. international journal of nanomedicine 10 (2015) 6027. https://doi.org/10.2147/ijn.s90347 [28] y. wang, l. li, j. li, b. yang, c. wang, w. fang, f. ji, y. wen, f. yao. stable and ph-responsive polyamidoamine based unimolecular micelles capped with a zwitterionic polymer shell for anticancer drug delivery. rsc advances 6 (2016) 17728-17739. https://doi.org/10.1039/c5ra25505h [29] q. zhang, f. liu, k.t. nguyen, x. ma, x. wang, b. xing, y. zhao. multifunctional mesoporous silica nanoparticles for cancer‐targeted and controlled drug delivery. advanced functional materials 22 (2012) 5144-5156. https://doi.org/10.1002/adfm.201201316 [30] a.c. alves, a. magarkar, m. horta, j.l. lima, a. bunker, c. nunes, s. reis. influence of doxorubicin on model cell membrane properties: insights from in vitro and in silico studies. scientific reports 7 (2017) 6343. https://doi.org/10.1038/s41598-017-06445-z [31] y. kato, s. ozawa, c. miyamoto, y. maehata, a. suzuki, t. maeda, y. baba. acidic extracellular microenvironment and cancer. cancer cell international 13 (2013) 89. https://doi.org/10.1186/14752867-13-89 [32] e.s. lee, k. na, y.h. bae. doxorubicin loaded ph-sensitive polymeric micelles for reversal of resistant mcf-7 tumor. journal of controlled release 103 (2005) 405-418. https://doi.org/10.1016/j.jconrel.20 04.12.018 [33] q.n. nguyen-trinh, k.x.t. trinh, n.-t. trinh, n. li, y. nagasaki, l.b. vong. a silica-based antioxidant nanoparticle for oral delivery of camptothecin which reduces intestinal side effects while improving drug efficacy for colon cancer treatment. acta biomaterialia 143 (2022) 459-470. https://doi.org/10.10 16/j.actbio.2022.02.036 [34] m. li, f. cheng, c. xue, h. wang, c. chen, q. du, d. ge, b. sun. surface modification of stöber silica nanoparticles with controlled moiety densities determines their cytotoxicity profiles in macrophages. langmuir 35 (2019) 14688-14695. https://doi.org/10.1021/acs.langmuir.9b02578 ©2023 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) https://doi.org/10.1007/s11814-014-0270-5 https://doi.org/10.1016/j.plipres.2013.04.008 https://doi.org/10.2147/ijn.s73971 https://doi.org/10.2147/ijn.s90347 https://doi.org/10.1039/c5ra25505h https://doi.org/10.1002/adfm.201201316 https://doi.org/10.1038/s41598-017-06445-z https://doi.org/10.1186/1475-2867-13-89 https://doi.org/10.1186/1475-2867-13-89 https://doi.org/10.1016/j.jconrel.20‌04.12.018 https://doi.org/10.1016/j.jconrel.20‌04.12.018 https://doi.org/10.10‌16/j.actbio.2022.02.036 https://doi.org/10.10‌16/j.actbio.2022.02.036 https://doi.org/10.1021/acs.langmuir.9b02578 http://creativecommons.org/licenses/by/3.0/ are we ready to design oral protacs®? doi: http://dx.doi.org/10.5599/admet.1037 243 admet & dmpk 9(4) (2021) 243-254; doi: https://doi.org/10.5599/admet.1037 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper are we ready to design oral protacs®? diego garcia jimenez, matteo rossi sebastiano, giulia caron, giuseppe ermondi* university of torino, molecular biotechnology and health sciences dept., cassmedchem, via quarello 15, 10135 torino, italy. *giuseppe ermondi: e-mail: giuseppe.ermondi@unito.it; tel.: +39 011 6708337 received: june 29, 2021; revised: august 22, 2021; available online: august 31, 2021 abstract protacs® are expected to strongly impact the future of drug discovery. therefore, in this work we firstly performed a statistical study to highlight the distribution of e3 ligases and pois collected in protac-db, the main online database focused on degraders. moreover, since the emerging technology of protein degradation deals with large and complex chemical structures, the second part of the paper focuses on how to set up a property-based design strategy to obtain oral degraders. for this purpose, we calculated a pool of seven previously ad hoc selected 2d descriptors for the 2258 publicly available degraders in protac-db (average values: mw= 972.9 da, nc= 49.5, nar= 4.5, phi= 17.3, nhdon= 4.5, nhacc= 17.7 and tpsa= 240 å 2 ) and compared them to a dataset of 50 bro5 orally approved drugs. then, a chemical space based on nc, phi and tpsa was built and subregions with optimal permeability and bioavailability were identified. bioavailable degraders (arv-110 and arv-471) tend to be closer to the ro5 region, using mainly semi-rigid linkers. permeable degraders, on the other hand, are placed in an average central region of the chemical space but chameleonicity could allow them to be located closer to the two arvinas compounds. ©2021 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords degrader; bro5; bioavailability; permeability; chemical space; protac-db; chameleonicity. introduction traditionally, the search for novel treatments has been undoubtedly driven by small molecule discovery. however, since the release of muromonab in 1986, biologics have gained importance. even though the market impact for this class of molecules was predicted to rise faster, 71 % of 2020s approved drugs (38) correspond to small molecules which still assume, 35 years later, the burden of drug discovery research (www.fda.gov). moreover, the guidelines for small molecule development have evolved from the traditional agonism-antagonism concept to innovative mechanisms of action. thus, small molecule research has enlarged its scope to nucleic acid-based modalities, protein-protein interaction (ppi) modulators, peptide/peptidomimetics and protacs® or degraders [1]. in particular, protacs® have been revealed to be a promising therapeutic area, which has expanded from 13 scientific articles in 2016 to 265 in 2020 (www.pubmed.com). structurally, protacs® are http://dx.doi.org/10.5599/admet.1037 https://doi.org/10.5599/admet.1037 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:giuseppe.ermondi@unito.it http://creativecommons.org/licenses/by/4.0/ http://www.fda.gov/ http://www.pubmed.com/ g. ermondi et al. admet & dmpk 9(4) (2021) 243-254 244 heterobifunctional degraders made of a warhead binding a protein of interest (poi), a ligand recruiting an e3 ligase and a linker coupling both regions (fig.1) [2]. biochemically speaking, their mechanism of action involves the formation of a ternary complex (e3 ligase:protac®:poi) in which the e3 ligase triggers the ubiquitination and subsequent degradation of the poi by the ubiquitin-proteasome system (ups) [2]. figure 1. structure of degraders and their building blocks (arv-825 is shown as an example). unlike traditional small molecules, based on occupancy pharmacological effects, degraders make use of this innovative event-driven mechanism of action, presenting several advantages over conventional strategies. this is considered a catalytic and sub-stoichiometric event, which allows the targeting of several units of the poi with just one protac® molecule [3]. in addition, due to its event-driven pharmacology, the degradative activity of the ups does not require the highest affinity of the warhead for the poi [4]. therefore, unlike traditional inhibitors requiring higher affinities, degraders are able to act on proteins thought to be ´undruggable`, which expands the therapeutic scope from protein-dependent diseases to viral infections, cancer and immune and neurodegenerative disorders [5]. nonetheless, protac technology is still in its infancy and is not yet completely understood [6]. degraders, first introduced by crews and deshaies in 2001 [7], have now a few candidates in clinical trials. in fact, arvinas announced in the first half of 2020 the first two oral degraders to reach phase 2 clinical trials (https://www.arvinas.com). moreover, the number of protacs® in the clinic should rise to 15 at the end of 2021, which reveals a higher interest in this field and a better understanding of their functioning [8]. however, despite their promising pharmacodynamic potential, the cautious introduction of protacs® into the market may also be explained by their poor bioavailability. moreover, this limitation is often related to drug metabolism and pharmacokinetics (dmpk) challenges, such as water solubility and permeability issues, derived from their complex chemical structure. chemically speaking, degraders abandon the classical “drug-like” context of lipinski’s rule-of-five (ro5) to be considered beyond the ro5 (bro5) molecules [9]. more recently, newer guidelines have tried to classify molecules attending to their molecular properties. starting from lipinski´s [9] and weber´s [10] guidelines, kihlberg et al. [11] suggested several physicochemical guidelines for oral bro5 drugs including abbvie´s multiparametric scoring function (ab-mps) [12]. moreover, these criteria were later on applied to protacs® by edmondson et al. in 2019 [13], who evaluated several physicochemical descriptors for 40 model protacs® and discussed their molecular properties based on their e3 ligases. additionally, maple et al. in 2019 [14] developed a study to examine the molecular properties of 422 degraders, establishing a predictive degrader score (deg_s). overall, these studies highlight the necessity for property-based resources that enable the identification of degraders with a reasonable potential to become oral drugs [15]. the first attempt in this direction was recently performed by our research group using a pool of 7 rationally selected 2d molecular descriptors (mw, nc, nar, tpsa, nhacc, nhdon and phi) for a dataset of https://www.arvinas.com/ admet & dmpk 9(4) (2021) 243-254 are we ready to design oral protacs®? doi: http://dx.doi.org/10.5599/admet.1037 245 protacs® downloaded from protac-db (http://cadd.zju.edu.cn/protacdb/) [16], an online database exclusively devoted to degraders [17]. recently protac-db has been significantly updated and thus the first aim of the study is to provide an overview of this resource. then, after recalling the concept of chemical space and tailoring to bro5 molecules [18], we built a protac® chemical space based on the aforementioned descriptors and investigated their distribution in relation to a set of oral available bro5 drugs. finally, we identified in this space the position of a small set of protacs® for which either bioavailability or permeability has been experimentally determined. experimental protac®, warhead, linker and e3 ligand csv files were separately downloaded from protac-db (last download in june 2021) and converted to xlsx files using microsoft excel (v. 16.0). data entries were submitted to osiris datawarrior version 5.2.1 (http://www.openmolecules.org/datawarrior/) and alvadesc 2.0.0 (https://www.alvascience.com/alvadesc/). mw, nc, phi, nhdon, nhacc and tpsa(tot) were calculated with alvadesc and nar with datawarrior. calculated values were then exported to excel and plotted into 3d graphs with datawarrior. results and discussion protacs® available structures: protac-db analysis common databases, chembl for instance, are expected to enhance drug discovery performance. however, the complexity of protacs® (and of most bro5 compounds) may require bespoke approaches [17]. for instance, the chemical probes portal is starting to include a preliminary dataset of degraders along with their targets to help protac® drug discovery, but it is still very limited (www.chemicalprobes.org). probes & drugs (www.probes-drugs.org/home/) have recently updated their content in protacs® with data obtained directly from the protac-db which is an online database fully focused on protacs® (http://cadd.zju.edu.cn/protacdb/). this latter introduces for the first time a search engine based on any of the three components of the degrader. consequently, the search for protacs® becomes more user-friendly, in which every protac entry is now considered as an addition of chemical moieties rather than just a fixed chemical structure with a database id. the protac-db provides first-hand information about chemical structures, biological activities, and physicochemical properties, retrieved from pubchem, chembl [19], bindingdb [20], and literature. in particular, physicochemical properties are calculated by rdkit (www.rdkit.org) and alogps software [21]. up to date, 2258 published protacs®, 275 warheads, 68 e3 ligands and 1099 linkers are contained in the protac-db (last download june 2021). due to the extensive information provided, regarding public and available information, the protac-db is the best online source to perform a preliminary analysis of the state of the art in protac® technology. thus, all the entries including protacs®, warheads and e3 ligands were downloaded from the protac-db. it is necessary to highlight that the number of available entries for each structural group is slightly higher than the number of reported molecular ids. therefore, 3939 protac® entries were reported for 2258 protacs®, 74 e3 ligand entries for 68 e3 ligand structures and 973 warhead entries for 275 warheads. this fact points out the versatility of some structures within every group, which can indeed target more than one protein (i.e., protac® id 2199 can be used to target clk1 or cdk9 proteins). hence, to have a clear picture of the current trend, protac-db entries were classified through their targeted protein and their corresponding protein families. moreover, a count was performed, and the corresponding protein/family clusters were sorted by size. http://dx.doi.org/10.5599/admet.1037 http://cadd.zju.edu.cn/protacdb/ http://www.openmolecules.org/datawarrior/ https://www.alvascience.com/alvadesc/ http://www.probes-drugs.org/home/ http://cadd.zju.edu.cn/protacdb/ http://www.rdkit.org/ g. ermondi et al. admet & dmpk 9(4) (2021) 243-254 246 the main e3 ligase clusters are represented in frequency histograms (fig. 2). from the 74 e3 ligand entries, 27 targeted vhl, 23 iaps and 10 crbn, revealing that e3 ligands focus mainly on these 3 families of targets (36, 31 and 14 %, respectively) (fig. 2a). when considering specific e3 ligase proteins (fig. 2b), vhl remains the most widely used e3 ligase (36 %). moreover, the iap family is subdivided into xiap (14 %), ciap1/birc2 (9 %), ciap2/birc3 (3 %) and unspecified iaps (5 %), revealing the heterogenicity of this e3 ligase family. consequently, snipers (specific and nongenetic iap-based protein erasers) are rising as degraders using a high variability of e3 ligase proteins [22]. finally, crbn (14 %) and mdm2 (7 %) gain the podium of e3 ligase proteins (fig. 3b). this result reveals the reduced variety of e3 ligases and the limited choice of e3 ligands when designing new protacs®. moreover, despite the higher number of known iap ligands compared to crbn ligands, just 1 crbn e3 ligand, pomalidomide, is used in 1366 out of the 3939 protac® entries (35 %). this fact confirms that crbn is one of the most used e3 ligases for protacs, despite crbn´s reduced e3 ligand choice range. consequently, due to the heterogenicity of the iap family and the great variability of iap ligands, it is expected that iap recruiting protacs® provide alternative tools in the near future. figure 2. histogram representation of a) e3 ligase protein families and b) e3 ligases. frequencies and percentages are calculated. values below 5 % for e3 ligase families and e3 ligases are not represented. warheads, on the other hand, have a more heterogeneous distribution (fig. 3) in which 121 warhead entries target the cdc2/cdkx protein kinases subfamily (12 %), 39 mapk (4 %), 38 serine/threonineprotein kinases (stk) (4 %), 33 nuclear hormone receptor (nhr) (3 %) and 28 bromodomain and extraterminal (bet) domain family (3 %) (fig. 3a). additionally, the androgen receptor (ar) (2 %), aurka (2 %), cdk4 (2 %), cdk9 (2 %) and brd4 (2 %), represent the most targeted proteins, although the distribution is almost equal for all the proteins targeted by warheads. since the protein of interest will ultimately determine the potential therapeutic use of a degrader, the discovery of new “undruggable” targets justifies this heterogeneous pattern. figure 3. histogram representation of a) poi families and b) poi targeted by warheads. frequencies and percentages are calculated. values below 3 % for poi families and 1 % for poi are not represented. regarding the 3939 protac® entries (fig. 4), 507 target cdc2/cdkx family proteins (23 %), 336 nuclear hormone receptors (15 %) 325 target bromodomain and extra-terminal (bet) domains (15 %), 307 map admet & dmpk 9(4) (2021) 243-254 are we ready to design oral protacs®? doi: http://dx.doi.org/10.5599/admet.1037 247 kinases (mapk) (14 %), 144 epithelial growth factor receptors (egfr) (7 %) and 139 the insulin receptor family (6 %), (figure 4a). in addition, the study reflects that 5 % of published protacs® target the estrogen receptor (er), 4 % brd4, 4 % cdk4, 4 % the ar and 4 % cdk6. moreover, the wide variety of poi targeted by warheads justifies the heterogeneous and scattered poi profile targeted by protacs. notably, protacdb does not include most proprietary compounds and thus we expect that warheads may cover more poi families. figure 4. histogram representation of a) poi families and b) poi targeted by protacs®. frequencies and percentages are calculated. values below 6 % for poi families and 2 % for poi are not represented. one protac® chemical space the most current definition of chemical space is the ensemble of all possible synthesizable molecules. when defining as bioactive any chemical entity able to specifically interfere withand alter biological systems, navigating the chemical space to find bioactive compounds sets itself as the primary goal of medicinal chemists. unfortunately, this is not a trivial task, since the chemical space potentially includes more than 10 180 molecules, [23]. in order to give a magnitude-order comparison, the biggest structure repository from the american chemical society “just” contains 10 8 structures and only a thousand are bioactive. moreover, it should be also recalled that bioactivity alone is not sufficient to allow a compound to become a drug and that oral drug delivery is the most preferred administration route. therefore, most bioactive molecules (often known as chemical probes) do not become oral drug candidates because of the lack of an adequate adme profile. in general terms, protacs® occupy a subregion of the whole chemical space that can be called the protac® chemical space. this region includes both protacs® with no future as drugs (either inactive molecules or chemical probes) and protacs® with potential as oral drug candidates. therefore, one main aim of pharmaceutical research consists in individuating in the early drug discovery the protac® chemical space subregions enriched in potential drug candidates (fig. 5). one reasonable way to do this is to use molecular properties. molecular properties are quantified by molecular descriptors, which define an n-dimensional chemical space, where n is the number of descriptors considered. molecular descriptors are in this fashion the orienting coordinates within the chemical space and retrospective analysis of already known compounds individuates regions of interest. in very early drug discovery stages, the chemical matter is not available and therefore in silico descriptors are used at the beginning of any research program. since all molecular properties fundamentally derive from the atom quality, quantity and connectivity, the first and most fundamental descriptors can be obtained from the molecular formula, such as mass weight, number of atoms, number of specific elements, etc. one level of complexity above this involves the use of atom connectivity-related descriptors. to date, the definition of the classically intended ”drug-like” chemical space (defined by lipinski’s rule of five, ro5) is purely bordered by molecular descriptors calculated from the structure formula, referred as 2d descriptors. even http://dx.doi.org/10.5599/admet.1037 g. ermondi et al. admet & dmpk 9(4) (2021) 243-254 248 though the ro5 still represents the golden region of the chemical space for small molecules, more and more active compounds discovered violate these guidelines. this occurs especially in the case of larger, highly flexible structures, such as protacs® [15]. figure 5. subregions of the chemical space: schematic representation. one key aspect of bro5 compounds is that the larger and more complex a structure is, the more conformation-dependent molecular properties become. many reports in recent years highlighted the importance of considering the conformational landscape of larger molecules when it comes to defining their molecular properties [24]. unfortunately, this is not a minor task: with the increase of rotatable bonds number and, consequently, the degrees of freedom, the number of potential conformers increases with factorial function. at present, computational conformational sampling procedures lose accuracy and consistency for bro5 compounds, making 3d descriptors calculation still challenging. therefore, due to the inherent conformational limitations of protacs®, we selected a pool of 2d descriptors that provide an unequivocal starting point in the definition of protac® chemical space [17]: mw as a descriptor of molecular size; a set of count descriptors related to both polar (nhacc (also hba), nhdon (also hbd)) and nonpolar (the number of carbon atoms, nc and the number of aromatic rings, nar) molecular moieties; a flexibility descriptor (phi) [25] and tpsa as a polarity index. in practice, we set up a pool of seven descriptors, three of nonpolar nature, three of polar nature and one flexibility descriptor. figure 6 shows violin plots representing the seven molecular property distribution of the 2258 degraders in protac-db (average values are: mw= 972.9 da, nc= 49.5, nar= 4.5, phi= 17.3, nhdon= 4.5, nhacc= 17.7 and tpsa= 240 å 2 ) (table s1). the average of the 7 calculated descriptors for degraders were compared to those previously calculated for 50 orally approved bro5 drugs [17] (mw= 765 da, nc= 39, nar= 2, phi= 14, tpsa= 177 å 2 , nhdon= 3 and nhacc= 13), introduced by kihlberg et al. [18]. results, presented in a radar plot, confirm that protacs® cover a wider property range, being the degrader´s nar more than twice the value for orally approved bro5 drugs (fig. 7a). nevertheless, it must be mentioned that some of the represented protacs® were designed as scientific tools, unintended to be pharmaceutical candidates with optimal dmpk properties. moreover, to better visualize the property distribution we defined a chemical space by including three representative descriptors: nc (nonpolar), tpsa (polar) and phi (flexibility) (fig. 7b). admet & dmpk 9(4) (2021) 243-254 are we ready to design oral protacs®? doi: http://dx.doi.org/10.5599/admet.1037 249 mw 0 500 1000 1500 2000 nc 0 50 100 nar 0 2 4 6 8 10 phi 0 10 20 30 40 50 tpsa 0 100 200 300 400 500 nhdon nhacc 0 10 20 30 40 figure 6. violin plots representing the molecular property distribution of the 2258 degraders in protac-db; molecular weight (mw, da) and number of carbons (nc) (green), number of aromatic rings (nar)(yellow), kier's flexibility index (phi) (purple), topological polar surface area (tpsa, å 2 ) (light blue) and number of hydrogen bond acceptors and donors (nhdon and nhacc) (dark red and blue, respectively). the 50th percentiles or medians are represented as solid horizontal bars and the 25th and 75 th percentiles as dashed lines. the picture emerging is that the protac® chemical space is adjacent and only partially stackable to the region occupied by oral bro5 drugs: the majority of protacs® have a higher number of carbon atoms than the analysed bro5 compounds. this is a bona fide indicator of larger structures. topological polar surface area is also higher. this aspect is of greater interest since a greater polar surface area is often associated with lower permeability. nevertheless, here we are considering fragment-calculated (topological) 2d surface area which does not include 3d information. thus, this aspect must be carefully considered. indeed, protacs® present in protac-db show also a higher kier’s flexibility index (phi), suggesting that the average increase in polar fragments, might be compensated with higher flexibility and polar moieties shielding. this fact has already been extensively reported in the bro5 chemical space [26,27]. finally, it should be recalled that protac-db does not include most of the proprietary compounds which might occupy further regions in the chemical space. subregions in the protacs® chemical space populated with potential oral drugs. a protac® chemical space represents an extraordinary tool to identify groups of degraders or clusters sharing similar chemical or in vitro adme properties. in particular, the potential of degraders to become oral drugs could be defined by bioavailability and permeability, among others. bioavailability (f%) is defined as the fraction of drug that reaches systemic circulation and thus needs an in vivo experimental determination. consequently, the recently disclosed orally bioavailable degraders arv-110 and arv-471 (fig. 8), were introduced into the chemical space defined above. both structures are located in the boundaries of the chemical space. regarding their molecular size, their structure contains 41 and 45 carbons respectively, being placed on the first quartile with respect to the protac-db average (nc= 49.5) http://dx.doi.org/10.5599/admet.1037 g. ermondi et al. admet & dmpk 9(4) (2021) 243-254 250 (fig. 6). moreover, their significantly low phi index (12 and 10 respectively) compared to the protac-db average (phi= 17.3) places both degraders in the first quartile (fig. 6). both structures incorporate a semirigid linker 1-[(piperidin-4-yl)methyl]piperazine (chemical structures in fig. s1), which compared to traditional pegylated or alkyl linkers, are less flexible. considering the importance of flexibility to the formation of intramolecular hydrogen bonds (imhbs), these two degraders occupy a subregion where, at least theoretically, dynamic imhbs are not favored [28] (a dynamic imhb is formed in nonpolar but not in polar media). consequently, this structural property profile reduces arvinas degrader´s capacity to change conformation in an environment-dependent manner, that is, to behave as molecular chameleons. in addition, arv-110 and arv-471 have a tpsa value of 182.86 å 2 and 96.43 å 2 , respectively, which places them in the first quartile (tpsa average= 240 å 2 ). overall, after analysing their properties it could be stated that both degraders occupy a superimposable region to that of oral bro5 drugs: (nc: 40-50, phi: 10-15 and tpsa: 50-200 å 2 ) (fig. 7). thus, this drug design strategy seems to have prioritized candidates staying as close to the ro5 chemical space as possible. figure 7. a) radar plot comparing the seven 2d descriptors for degraders and orally approved bro5 drugs b) chemical space occupied by protacs® and bro5 approved drugs. permeability is one of the major determinants of bioavailability. it is physiologically defined as the ability of a compound to pass across biological membranes and is determined in vitro with different methods, pampa and caco-2 being the most common ones (a discussion about the differences between the two methods and their reliability when applied to bro5 compounds is beyond the aim of the paper [28,29]). however, the limitations of obtaining reliable permeability data for large molecules [27] suggest that degraders with an apparent low passive permeability could surprisingly display decent bioavailability values. for this reason, a set of degraders showing consistent permeability values were identified: protac1 described by kihlberg et al. [30], a small set provided for free by opnme (www.opnme.com) [25], protac-14 by skidmore et al. [31] and 11 degraders reported by lokey et al. [32] (figure s1 and table s2). these degraders were then located into the chemical space (fig. 9a) and classified into permeable (papp > 1 x 10 -6 cm/s) or low permeable (papp < 1 x 10 -6 cm/s) (fig. 9b) (table s2). notably, permeable degraders (protac-1, protac-14, acbi1 and bi-3663) occupy a central region of the chemical space (figure 9b). among them, bi-3663 and protac-14, are the closest to arvinas’s compounds. in particular, protac-14 is the degrader showing the best “drug-like” properties (table s3). it has the structure containing less nc and the lowest phi and tpsa values. moreover, bi-3663 is slightly more polar than acbi1 but has 5 carbons less and thus is closer to arv-110 and arv-471. finally, protac-1 has a higher nc (50), tpsa (265 å 2 ) and phi (19) which places it above the other degraders. protac-1 is permeable although large dimensions and high tpsa are associated with low permeability. however, protac-1 is highly flexible and thus capacity to form http://www.opnme.com/ admet & dmpk 9(4) (2021) 243-254 are we ready to design oral protacs®? doi: http://dx.doi.org/10.5599/admet.1037 251 dynamic intramolecular interactions that confer it the capacity to behave as a molecular chameleon [30]. chameleonicity therefore could explain why protac-1 is permeable despite its position in a less “druglike” region of the protac® chemical space. figure 8. chemical space occupied by arv-110 and arv-471. overall, our analysis suggests that permeable protacs® could be located in three regions: 1) the subregion having the “best drug” like properties (smaller size, lower polarity and flexibility) (protac-14), 2) an intermediate region with acceptable molecular properties (bi-3663 and acbi1) and 3) a region occupied by molecular chameleons (protac-1). however, since chameleonicity cannot be only explained on the basis of 2d descriptors (the determinants of chameleonicity are beyond the scope of this paper) it is possible that molecular chameleons with optimal permeability could occupy further regions of this 2d property-based chemical space. figure 9. chemical space of a a) subset of permeable degraders b) graphical distribution of highly permeable and bioavailable degraders. a b http://dx.doi.org/10.5599/admet.1037 g. ermondi et al. admet & dmpk 9(4) (2021) 243-254 252 conclusions protacs® are an additional modality to the drug discovery toolbox and ad hoc strategies to fully exploit their potential as oral drugs are still under study. in this paper, we first explored the protac-db which represents a useful tool for degrader design. in particular, we identified the most popular targets in degrader technology and calculated statistically their distribution. e3 ligands have been revealed to target mainly three e3 ligase families (vhl, iaps and crbn) that represent up to 81 % of the total. warheads, on the contrary, show a heterogeneous distribution in which only 2 % target the same poi. this high warhead variability confers protacs® the capacity to degrade multiple targets, being the er the most popular target (5 %). the updated version of protac-db allowed us to update and expand the protac® chemical space previously described by some of us [17] and based on a set of seven calculated 2d descriptors (mw, nc, nar, phi, nhdon, nhacc and tpsa). then we compared the protac® with an orally approved bro5 drugs chemical space and provided for the first-time initial insights on the location of bioavailable (arv-110 and arv-471) and a small dataset of permeable protacs®. results show that bioavailable degraders are close to the oral bro5 region whereas permeable protacs® seem to occupy an extended and central region of the chemical space. in addition, the fact that chameleonicity can improve permeability could locate molecular chameleons towards the less “drug-like” region of the chemical space. overall, we proved that the chemical space using nc, tpsa and phi helps to identify regions of protacs® with future as oral drugs if permeability/bioavailability data are available. since limitations associated to the measurement of bro5 permeability are known, the determination of experimental physicochemical descriptors able to predict permeability could accelerate the identification of new oral protacs®. studies along these lines are in due course in our laboratories. acknowledgements: the authors thank the university of torino (ricerca locale ex-60%, bando2020) for financial support. conflict of interest: the authors declare no conflict of interest. references [1] w. yang, p. gadgil, v.r. krishnamurthy, m. landis, p. mallick, d. patel, p.j. patel, d.l. reid, m. sanchez-felix. the evolving druggability and developability space: chemically modified new modalities and emerging small molecules. the aaps journal 22 (2020) 21. https://doi.org/10.1208/s12248-019-0402-2. [2] s. an, l. fu. small-molecule protacs: an emerging and promising approach for the development of targeted therapy drugs. ebiomedicine 36 (2018) 553–562. https://doi.org/10.1016/j.ebiom.2018.09.005. [3] d.p. bondeson, a. mares, i.e.d. smith, e. ko, s. campos, a.h. miah, k.e. mulholland, n. routly, d.l. buckley, j.l. gustafson, n. zinn, p. grandi, s. shimamura, g. bergamini, m. faelth-savitski, m. bantscheff, c. cox, d.a. gordon, r.r. willard, j.j. flanagan, l.n. casillas, b.j. votta, w. den besten, k. famm, l. kruidenier, p.s. carter, j.d. harling, i. churcher, c.m. crews. catalytic in vivo protein knockdown by small-molecule protacs. nature chemical biology 11 (2015) 611–617. https://doi.org/10.1038/nchembio.1858. [4] c.m. crews. targeting the undruggable proteome: the small molecules of my dreams. chemistry & biology 17 (2010) 551–555. https://doi.org/10.1016/j.chembiol.2010.05.011. https://doi.org/10.1208/s12248-019-0402-2 https://doi.org/10.1208/s12248-019-0402-2 https://doi.org/10.1016/j.ebiom.2018.09.005 https://doi.org/10.1016/j.ebiom.2018.09.005 https://doi.org/10.1038/nchembio.1858 https://doi.org/10.1016/j.chembiol.2010.05.011 admet & dmpk 9(4) (2021) 243-254 are we ready to design oral protacs®? doi: http://dx.doi.org/10.5599/admet.1037 253 [5] x. liu, x. zhang, d. lv, y. yuan, g. zheng, d. zhou. assays and technologies for developing proteolysis targeting chimera degraders. future medicinal chemistry 12 (2020) 1155-1179. https://doi.org/10.4155/fmc-2020-0073. [6] s. ramachandran, a. ciulli. building ubiquitination machineries: e3 ligase multi-subunit assembly and substrate targeting by protacs and molecular glues. current opinion in structural biology 67 (2021) 110–119. https://doi.org/10.1016/j.sbi.2020.10.009. [7] k.m. sakamoto, k.b. kim, a. kumagai, f. mercurio, c.m. crews, r.j. deshaies. protacs: chimeric molecules that target proteins to the skp1-cullin-f box complex for ubiquitination and degradation. proceedings of the national academy of sciences 98 (2001) 8554–8559. https://doi.org/10.1073/pnas.141230798. [8] a. mullard. targeted protein degraders crowd into the clinic. nature reviews drug discovery 20 (2021) 247-250. https://doi.org/10.1038/d41573-021-00052-4. [9] c.a. lipinski, f. lombardo, b.w. dominy, p.j. feeney. experimental and computational approaches to estimate solubility and permeability in drug discovery and development settings. advanced drug delivery reviews 23 (1997) 3–25. https://doi.org/10.1016/s0169-409x(96)00423-1. [10] d.f. veber, s.r. johnson, h.-y. cheng, b.r. smith, k.w. ward, k.d. kopple. molecular properties that influence the oral bioavailability of drug candidates. journal of medicinal chemistry 45 (2002) 2615– 2623. https://doi.org/10.1021/jm020017n. [11] v. poongavanam, b.c. doak, j. kihlberg. opportunities and guidelines for discovery of orally absorbed drugs in beyond rule of 5 space. current opinion in chemical biology 44 (2018) 23–29. https://doi.org/10.1016/j.cbpa.2018.05.010. [12] d.a. degoey, h.-j. chen, p.b. cox, m.d. wendt. beyond the rule of 5: lessons learned from abbvie’s drugs and compound collection. journal of medicinal chemistry 61 (2018) 2636–2651. https://doi.org/10.1021/acs.jmedchem.7b00717. [13] s.d. edmondson, b. yang, c. fallan. proteolysis targeting chimeras (protacs) in ‘beyond rule-of-five’ chemical space: recent progress and future challenges. bioorganic and medicinal chemistry letters 29 (2019) 1555–1564. https://doi.org/10.1016/j.bmcl.2019.04.030. [14] h.j. maple, n. clayden, a. baron, c. stacey, r. felix. developing degraders: principles and perspectives on design and chemical space. medchemcomm 10 (2019) 1659–1846. https://doi.org/10.1039/c9md00272c. [15] g. ermondi, d. garcia jimenez, m. rossi sebastiano, g. caron. rational control of molecular properties is mandatory to exploit the potential of protacs as oral drugs. acs medicinal chemistry letters 12 (2021) 1056–1060. https://doi.org/10.1021/acsmedchemlett.1c00298. [16] g. weng, c. shen, d. cao, j. gao, x. dong, q. he, b. yang, d. li, j. wu, t. hou. protac-db: an online database of protacs. nucleic acids research 49 (2020) d1381–d1387. https://doi.org/10.1093/nar/gkaa807. [17] g. ermondi, d. garcia-jimenez, g. caron. protacs and building blocks: the 2d chemical space in very early drug discovery. molecules 26 (2021) 672. https://doi.org/10.3390/molecules26030672. [18] b.c. doak, b. over, f. giordanetto, j. kihlberg. oral druggable space beyond the rule of 5: insights from drugs and clinical candidates. chemistry & biology 21 (2014) 1115–42. https://doi.org/10.1016/j.chembiol.2014.08.013. [19] d. mendez, a. gaulton, a.p. bento, j. chambers, m. de veij, e. félix, m.p. magariños, j.f. mosquera, p. mutowo, m. nowotka, m. gordillo-marañón, f. hunter, l. junco, g. mugumbate, m. rodriguezlopez, f. atkinson, n. bosc, c.j. radoux, a. segura-cabrera, a. hersey, a.r. leach. chembl: towards direct deposition of bioassay data. nucleic acids research 47 (2019) d930–d940. https://doi.org/10.1093/nar/gky1075. [20] m.k. gilson, t. liu, m. baitaluk, g. nicola, l. hwang, j. chong. bindingdb in 2015: a public database for medicinal chemistry, computational chemistry and systems pharmacology. nucleic acids research 44 (2016) d1045–d1053. https://doi.org/10.1093/nar/gkv1072. http://dx.doi.org/10.5599/admet.1037 https://doi.org/10.4155/fmc-2020-0073 https://doi.org/10.4155/fmc-2020-0073 https://doi.org/10.1016/j.sbi.2020.10.009 https://doi.org/10.1073/pnas.141230798 https://doi.org/10.1073/pnas.141230798 https://doi.org/10.1038/d41573-021-00052-4 https://doi.org/10.1016/s0169-409x(96)00423-1 https://doi.org/10.1021/jm020017n https://doi.org/10.1016/j.cbpa.2018.05.010 https://doi.org/10.1021/acs.jmedchem.7b00717 https://doi.org/10.1016/j.bmcl.2019.04.030 https://doi.org/10.1039/c9md00272c https://doi.org/10.1039/c9md00272c https://doi.org/10.1021/acsmedchemlett.1c00298 https://doi.org/10.1093/nar/gkaa807 https://doi.org/10.1093/nar/gkaa807 https://doi.org/10.3390/molecules26030672 https://doi.org/10.1016/j.chembiol.2014.08.013 https://doi.org/10.1016/j.chembiol.2014.08.013 https://doi.org/10.1093/nar/gky1075 https://doi.org/10.1093/nar/gky1075 https://doi.org/10.1093/nar/gkv1072 g. ermondi et al. admet & dmpk 9(4) (2021) 243-254 254 [21] i. v. tetko, v.y. tanchuk. application of associative neural networks for prediction of lipophilicity in alogps 2.1 program. journal of chemical information and computer sciences 42 (2002) 1136–1145. https://doi.org/10.1021/ci025515j. [22] z. ma, y. ji, y. yu, d. liang. specific non-genetic iap-based protein erasers (snipers) as a potential therapeutic strategy. european journal of medicinal chemistry 216 (2021) 113247. https://doi.org/10.1016/j.ejmech.2021.113247. [23] a.-d. gorse. diversity in medicinal chemistry space. current topics in medicinal chemistry 6 (2006) 3–18. https://doi.org/10.2174/156802606775193310. [24] v. poongavanam, e. danelius, s. peintner, l. alcaraz, g. caron, m.d. cummings, s. wlodek, m. erdelyi, p.c.d. hawkins, g. ermondi, j. kihlberg. conformational sampling of macrocyclic drugs in different environments: can we find the relevant conformations? acs omega 3 (2018) 11742– 11757. https://doi.org/10.1021/acsomega.8b01379. [25] g. ermondi, m. vallaro, g. caron. degraders early developability assessment: face-to-face with molecular properties. drug discovery today 25 (2020) 1585–1591. https://doi.org/10.1016/j.drudis.2020.06.015. [26] a. whitty, m. zhong, l. viarengo, d. beglov, d.r. hall, s. vajda. quantifying the chameleonic properties of macrocycles and other high-molecular-weight drugs. drug discovery today 21 (2016) 712–717. https://doi.org/10.1016/j.drudis.2016.02.005. [27] m. rossi sebastiano, b.c. doak, m. backlund, v. poongavanam, b. over, g. ermondi, g. caron, p. matsson, j. kihlberg. impact of dynamically exposed polarity on permeability and solubility of chameleonic drugs beyond the rule of 5. journal of medicinal chemistry 61 (2018) 4189–4202. https://doi.org/10.1021/acs.jmedchem.8b00347. [28] g. caron, j. kihlberg, g. ermondi. intramolecular hydrogen bonding: an opportunity for improved design in medicinal chemistry. medicinal research reviews 39 (2019) 1707–1729. https://doi.org/10.1002/med.21562. [29] p. matsson, j. kihlberg. how big is too big for cell permeability? journal of medicinal chemistry 60 (2017) 1662–1664. https://doi.org/10.1021/acs.jmedchem.7b00237. [30] y. atilaw, v. poongavanam, c. svensson nilsson, d. nguyen, a. giese, d. meibom, m. erdelyi, j. kihlberg. solution conformations shed light on protac cell permeability. acs medicinal chemistry letters 12 (2021) 107–114. https://doi.org/10.1021/acsmedchemlett.0c00556. [31] d.e. scott, t.p.c. rooney, e.d. bayle, t. mirza, h.m.g. willems, j.h. clarke, s.p. andrews, j. skidmore. systematic investigation of the permeability of androgen receptor protacs. acs medicinal chemistry letters 11 (2020) 1539–1547. https://doi.org/10.1021/acsmedchemlett.0c00194. [32] v.g. klein, c.e. townsend, a. testa, m. zengerle, c. maniaci, s.j. hughes, k.h. chan, a. ciulli, r.s. lokey. understanding and improving the membrane permeability of vh032-based protacs. acs medicinal chemistry letters 11 (2020) 1732–1738. https://doi.org/10.1021/acsmedchemlett.0c00265. ©2021 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) https://doi.org/10.1021/ci025515j https://doi.org/10.1016/j.ejmech.2021.113247 https://doi.org/10.1016/j.ejmech.2021.113247 https://doi.org/10.2174/156802606775193310 https://doi.org/10.1021/acsomega.8b01379 https://doi.org/10.1016/j.drudis.2020.06.015 https://doi.org/10.1016/j.drudis.2020.06.015 https://doi.org/10.1016/j.drudis.2016.02.005 https://doi.org/10.1021/acs.jmedchem.8b00347 https://doi.org/10.1021/acs.jmedchem.7b00237 https://doi.org/10.1021/acsmedchemlett.0c00556 https://doi.org/10.1021/acsmedchemlett.0c00194 https://doi.org/10.1021/acsmedchemlett.0c00265 https://doi.org/10.1021/acsmedchemlett.0c00265 http://creativecommons.org/licenses/by/3.0/ voltammetric determination of vitamin b6 in the presence of vitamin c based on zinc ferrite nanoparticles modified screen-printed graphite electrode doi: https://doi.org/10.5599/admet.1702 251 admet & dmpk 11(2) (2023) 251-261; doi: https://doi.org/10.5599/admet.1702 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper voltammetric determination of vitamin b6 in the presence of vitamin c based on zinc ferrite nano-particles modified screenprinted graphite electrode peyman mohammadzadeh jahani1, maedeh jafari2 and sayed ali ahmadi*3 1 school of medicine, bam university of medical sciences, bam, iran 2 department of pediatrics, school of medicine, kerman university of medical sciences, kerman, iran 3 department of chemistry, kerman branch, islamic azad university, kerman, iran *corresponding author: e-mail: ahmadi.iauk59@gmail.com received: february 02, 2023; revised: february 23, 2023; published: march 15, 2023 abstract the zinc ferrite nano-particles (znfe2o4) modified screen-printed graphite electrode (znfe2o4/spge) was used for the voltammetric determination of vitamin b6 in real samples, using differential pulse voltammetry (dpv). it has been found that the oxidation of vitamin b6 at the surface of such an electrode occurs at a potential about 150 mv less positive compared to an unmodified screen-printed graphite electrode. after optimization, a vitamin b6 sensor with a linear range from 0.8 to 585.0 µm and a detection limit of 0.17 µm. the znfe2o4/spge sensor exhibits good resolution between the voltammetric peaks of vitamin b 6 and vitamin c, making it suitable for detecting vitamin b6 in the presence of vitamin c in real samples. ©2023 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords zinc ferrite nanoparticles; voltammetry; vitamin b6; vitamin c; modified electrode introduction vitamins are small organic molecules whose lack or excess may result in several diseases to the organisms that need them. they are classified into two groups by their solubilities, namely water-soluble vitamins (vitamin b and vitamin c) and fat-soluble vitamins (vitamin a, vitamin d, vitamin e, and vitamin k) [1 2]. vitamin b6 belongs to the water-soluble b complex vitamins group, commonly called pyridoxine. it is essential in the diet for the metabolism of amino acids and the maintenance of body cells. the nervous and immune systems need vitamin b6 for efficient functioning; it also plays a major role in the conversion of tryptophan to niacin [3,4]. also, it is found in different chemical forms (pyridoxamine, pyridoxine or 5-phosphate derivatives), but the most stable form is pyridoxine which is used in drug formulations such as multivitamin supplements or in enriched foods [5]. vitamin c is one of the most important water-soluble vitamins and it refers to all compounds exhibiting equivalent biological activity to l-ascorbic acid (aa), including dehydroascorbic acid (dhaa), the oxidation product of aa, its isomers and esters. vitamin c is an antioxidant necessary for the growth, development, and repair of all tissues [6,7]. low levels of vitamin c can result in a condition called scurvy. scurvy may cause https://doi.org/10.5599/admet.1702 https://doi.org/10.5599/admet.1702 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:ahmadi.iauk59@gmail.com http://creativecommons.org/licenses/by/4.0/ p. m. jahani et al. admet & dmpk 11(2) (2023) 251-261 252 symptoms such as rash, muscle weakness, joint pain, tiredness, or tooth loss. vitamin c is an important antioxidant, along with vitamin e, beta-carotene, and many other plant-based nutrients. antioxidants block some of the damage caused by free radicals, substances that damage dna [8,9]. studies show that vitamins b6 and c are essential for the natural synthesis of dopamine in the human body. on the other hand, large doses of vitamins b6 and c may reduce the risk of kidney stone formation in women. so, the simultaneous determination of these compounds is very important for pharmaceutical and biological investigation [10,11]. several methods have been developed to determine vitamins b6 and c, including high-performance liquid chromatography [12], spectrophotometry [13], and flow injection [14,15]. however, these methods require not only advanced technical expertise but also time-consuming, expensive and often need the pretreatment step. electrochemical detection is an attractive alternative approach to these technologies because of the inherent advantages of simplicity, ease of miniaturization, cost-effectiveness, dependability, high sensitivity, and relatively low cost [16-21]. so far, electrochemical sensors have found widespread use in a variety of disciplines, including pharmaceutical, food, and clinical analyses [22-29]. currently, the advancement of screen-printing technology for the fabrication of screen-printed electrodes (spes) is attracting enormous attention due to the advantageous properties of spes compared to conventional electrodes, such as cost-effectiveness, disposability, simplicity, versatility, availability of materials and patterns, elimination of electrode maintenance, the requirement for low volumes of solution, and appropriateness for outside laboratory measurement [30-32]. chemically modified electrodes are best suited for the electrochemical determination of pharmaceutical, environmental, or biological samples. chemically modified electrodes reduce the over-potential required for either the oxidation or reduction of the electro-active compounds [33-39]. also, modification of electrodes is a powerful strategy for overcoming such limitations of un-modified electrodes as low selectivity, poor sensitivity, low stability, and the blockage of the electron transfer [40-45]. nanostructured metal oxides crystallizing in the spinel structure type have been investigated intensively over the years and present a permanent interest due to their wide technological applications such as magnetic and optical materials, semiconductors, pigments, catalysts, or material for biomedical applications [46-51]. ferrites are a well-known class of complex oxides of considerable technological importance. on the other side, nano znfe2o4 as spinel ferrites is found to be one of the most interesting spinel systems because of its unique properties, photochemical stability, good visible-light response and favourable magnetism [52,53]. a characteristic of znfe2o4 is that it has two different metal cations, zn and fe, with o as an anion. the cations occupy two different positions in a spinel structure: tetrahedral (zn) and octahedral (fe) sites along the facecentered cubic lattice formed by o2− cations. the use of bimetallic oxides as electrode materials could enhance both electrical conductivity by two orders of magnitude and electrochemical activity versus materials prepared with unitary metal oxides [54,55]. in the present work, the preparation and application of a screen-printed graphite electrode, modified with zinc ferrite nano-particles (znfe2o4), for the determination of vitamin b6 in the presence of vitamin c is described. the electrochemical behavior of vitamin b6 at znfe2o4/spge was investigated. the results showed the superiority of znfe2o4/spge to the bare electrode in terms of better sensitivity. we have also evaluated the analytical performance of the znfe2o4/spge for the quantification of vitamin b6 in the presence of vitamin c in some real samples. admet & dmpk 11(2) (2023) 251-261 determination of vitamin b6 in the presence of vitamin c doi: https://doi.org/10.5599/admet.1702 253 experimental chemicals and instrumentation all chemicals used were of analytical reagent grade purchased from sigma-aldrich and were used as received without any further purification. double-distilled water was used throughout all experiments. orthophosphoric acid was utilized to prepare the phosphate buffer solutions (pbss), and sodium hydroxide was used to adjust the desired ph values (ph range between 2.0 and 9.0). cyclic voltammetry (cv), linear sweep voltammetry (lsv), chronoamperometry, and differential pulse voltammetry (dpv) investigations were performed in an electroanalytical system autolab pgstat302n, potentiostat/galvanostat connected to an electrode cell, the spge (dropsens; drp-110: spain), containing graphite counter electrode, a graphite working electrode, and a silver pseudo-reference electrode. the system was run on a pc using general purpose electrochemical system (gpes) software. solution ph values were determined using a 713 ph meter combined with a glass electrode (metrohm, switzerland). preparation of modified electrode znfe2o4 nano-particles were used to coat the bare screen printed graphite electrode. a stock solution of znfe2o4 nano-particles in 1 ml of the aqueous solution was prepared by distributing 1 mg of znfe2o4 nanoparticles via ultra-sonication for 50 min, whereas 4 µl of aliquots of the znfe2o4 nano-particles suspension solution was cast on carbon working electrodes and evaporated at room temperature. results and discussion electrochemical behavior of vitamin b6 on the znfe2o4/spge according to our knowledge, the electrooxidation of vitamin b6 depends on the ph value of the solution (scheme 1). so, the effect of ph was investigated using the dpv method. results show that the oxidation peak current increased slowly from ph 2.0 to 7.0, and then the current conversely decreased when the ph value increased from 7.0 to 9.0. consequently, ph 7.0 was chosen as the optimal experimental condition for other experiments. scheme 1. the proposed mechanism for the oxidation of vitamin b6 at the znfe2o4/spge. to investigate the vitamin b6 behavior and the as-produced electrode response to vitamin b6, the performance of znfe2o4/spge was compared to that of unmodified spge. figure 1 shows the cv curve obtained for znfe2o4/spge (curve a) and unmodified spge (curve b) in the presence of 200.0 μm vitamin b6containing pbs at the scan rate of 50 mv/s. the results showed that the oxidation of vitamin b6 is very weak on the surface of the bare spge, but the presence of znfe2o4 nano-particles could enhance the peak current and decrease the oxidation potential (decreasing the overpotential). a substantial negative shift of the currents starting from oxidation potential for vitamin b6 and a dramatic increase of the current indicates the catalytic ability of znfe2o4/spge to vitamin b6 oxidation. the results showed that the use of znfe2o4 nanoparticle improved the characteristics of vitamin b6 oxidation, which was partly due to excellent characteristics of znfe2o4 nano-particles such as excellent electrical conductivity and good chemical stability. https://doi.org/10.5599/admet.1702 p. m. jahani et al. admet & dmpk 11(2) (2023) 251-261 254 figure 1. cyclic voltammograms of a) znfe2o4/spge and b) spge in the presence of 200.0 μm vitamin b6 at a ph 7.0 of 0.1 m pbs, respectively. effect of scan rate the linear sweep voltammograms measurements were carried out to evaluate the association of peak current with scan rate at varied scan rates (10-400 mv/s) in the 100.0 μm vitamin b6-containing 0.1 m pbs (ph = 7.0) on the znfe2o4/spge (figure 2). as shown in figure 2, the peak currents of vitamin b6 grow with increasing scan rates and there are good linear relationships between the peak currents (ip) and the square root of the scan rate (ν 1/2). the results also showed that the action is mass transfer of vitamin b6 controlled at diffusion process. to obtain further information on the rate-determining step, a tafel plot was developed for the vitamin b6 at the surface of znfe2o4/spge using the data derived from the rising part of the current–voltage curve (figure 3). the slope of the tafel plot is equal to 2.3rt/n(1 − α)f, which comes up to 0.1803 v decade−1. we obtained the charge transfer coefficient (α) as 0.67. chronoamperometric measurements the electrooxidation of vitamin b6 by a znfe2o4/spge was also studied by chronoamperometry (figure 4). chronoamperometric measurements of different concentrations of vitamin b6 at the znfe2o4/spge sensor were accomplished by setting the working electrode potential at 760 mv as the first step potential. using chronoamperometric studies, we determined the diffusion coefficient, d, of vitamin b6 in a buffer solution. the experimental plots of ip versus t-1/2 were employed with the best fits for different concentrations of vitamin b6 (figure 4a). the slopes of the resulting straight lines were then plotted versus vitamin b6 concentrations (figure 4b). using these slopes and the cottrell equation, we obtained d= 9.1×10-6 cm2 s-1. admet & dmpk 11(2) (2023) 251-261 determination of vitamin b6 in the presence of vitamin c doi: https://doi.org/10.5599/admet.1702 255 figure 2. linear sweep voltammograms of vitamin b6 (100.0μm) at znfe2o4/spge at different scan rates of a) 10, b) 50, c) 100, d)200, e) 300, and f) 400 mv/s in 0.1 m pbs (ph 7.0). insert: plot of ip versus ν 1/2 for the oxidation of vitamin b6 at znfe2o4/spge. figure 3. linear sweep voltammogram for znfe2o4/spge in the presence of 0.1 m pbs (ph 7.0) with 100.0 µm of vitamin b6 at the scan rate of 10 mv/s; points: outputs used in tafel plot; inset: tafel plot of lsv. figure 4. chronoamperograms obtained at the znfe2o4/spge in the presence of a) 0.1, b) 0.3, c) 0.5, d) 1.0, e) 1.5, f) 2.0, g) 2.7, h) 3.2, and i) 3.5 μm vitamin b6 in the0.1 m buffer solution (ph 7.0). a) plot of i versus t-1/2 for electrooxidation of vitamin b6 obtained from chronoamperograms a–i. b) plot of slope from straight lines versus vitamin b6 level. calibration plot and limit of detection since dpv has a much higher current sensitivity and better resolution than cv and lsv, dpv was used for the determination of vitamin b6. figure 5 shows the dpv curves of znfe2o4/spge in the pbs buffer with variable vitamin b6 levels (step potential=0.01 v and pulse amplitude=0.025 v). it was found that the https://doi.org/10.5599/admet.1702 p. m. jahani et al. admet & dmpk 11(2) (2023) 251-261 256 electrocatalytic peak currents of vitamin b6 oxidation at znfe2o4/spge surface linearly depended on vitamin b6 concentrations above the range of 0.8-585.0 µm (with a correlation coefficient of 0.9997), while determination limit was achieved to be 0.17 µm. figure 5. dpv curves of znfe2o4/spge in the 0.1 m buffer solution (ph 7.0) containing different concentrations of vitamin b6. a-m corresponds to 0.8, 5.0, 10.0, 20.0, 40.0, 60.0, 80.0, 100.0, 200.0, 300.0, 400.0, 500.0, and 585.0 μm vitamin b6. inset: plots of electrocatalytic peak current as a function of vitamin b6 concentration. determination of vitamin b6 in the presence of vitamin c the simultaneous determination of vitamin b6 and vitamin c is one of the most important applications of the proposed modified electrode. this study investigated a simultaneous change in the concentrations of vitamins b6 and c by recording the dpv curves. the result showed two well-defined oxidation peaks with a 530 mv separation of the peaks (figure 6). insets a and b in figure 6 show the dependence of dpv peak currents on the concentration of vitamin b6 and vitamin c, respectively. the sensitivities towards vitamin b6 in the absence and presence of vitamin c were found to be 0.0501 µa/µm (in the absence of vitamin c) and 0.0503 µa/µm (in the presence of vitamin c). these results demonstrated that the znfe2o4/spge successfully detected vitamin b6 and vitamin c simultaneously, both sensitively and selectively. stability of modified electrode for checking znfe2o4/spge sensor stability, we kept the recommended sensor within the ph equal to 7.0 in the pbs for two weeks to test znfe2o4/spge stability and, consequently, we recorded the dpv of the solution consisting of 50.0 μm vitamin b6 to be compared to the dpv observed prior to immersion. the oxidation peak of vitamin b6 did not change and, in comparison to earlier responses to the current, showed a less than 4.5 % reduction in signal, reflecting acceptable stability of znfe2o4/spge. admet & dmpk 11(2) (2023) 251-261 determination of vitamin b6 in the presence of vitamin c doi: https://doi.org/10.5599/admet.1702 257 figure 6. differential pulse voltammograms of znfe2o4/spge in 0.1 m pbs (ph 7.0) containing different concentrations of vitamin c and vitamin b6 mixed solutions of: a)10.0+15.0 , b) 50.0+50.0, c) 125.0+100.0, d) 300.0+150.0, e) 400.0+200.0, f) 600.0+300.0, g) 800.0+400.0,and h) 1000.0+500.0 µm vitamin c and vitamin b6, respectively. insets: (a) plot of the peak currents as a function of vitamin c concentration and (b) plot of the peak currents as a function of vitamin b6 concentration. conclusion a sensor for voltammetric determination of traces of vitamin b6 in real samples, based on the zn-ferrite modified screen printed graphite electrode, was developed. the sensor exhibited a good linear response over the concentration range 0.8-585.0 µm with a detection limit of 0.17 µm for vitamin b6. also, the modified electrode successfully resolves the overlapped voltammetric peaks of vitamin b6 and vitamin c by approximately 530 mv so that the modified electrode displays high selectivity in the dpv measurement of vitamin b6 and vitamin c of in their mixture solutions. as well as, the proposed method could be applied to the determination of vitamin b6 and vitamin c in real samples. conflict of interest: the authors declare no conflict of interest. references [1] s. abdulmlik, p. saifullah, m.a. al-a’adhami. evaluation of vitamin b12 and methylmalonic acid levels as markers with neuropathy in patients of type 2 diabetes mellitus. eurasian chemical communications 4(10) (2022) 956-965. https://doi.org/10.22034/ecc.2022.339922.1452 [2] b. brunetti. recent advances in electroanalysis of vitamins. electroanalysis 28(9) (2016) 1930-1942. https://doi.org/10.1002/elan.201600097 [3] b. habibi, h. phezhhan, m.h. pournaghi-azar. voltammetric determination of vitamin b6 (pyridoxine) using multi wall carbon nanotube modified carbon-ceramic electrode. journal of the iranian chemical society 7 (2010) s103-s112. https://doi.org/10.1007/bf03246189 [4] s. rison, a.t. mathew, l. george, t. maiyalagan, g. hegde, a. varghese, pt nanospheres decorated graphene-β-cd modified pencil graphite electrode for the electrochemical determination of vitamin b6. topics in catalysis (2022) 1-11. https://doi.org/10.1007/s11244-021-01559-1 https://doi.org/10.5599/admet.1702 https://doi.org/10.22034/ecc.2022.339922.1452 https://doi.org/10.1002/elan.201600097 https://doi.org/10.1007/bf03246189 https://doi.org/10.1007/s11244-021-01559-1 p. m. jahani et al. admet & dmpk 11(2) (2023) 251-261 258 [5] s.m. cottica, j. nozaki, h.s. nakatani, c.c. oliveira, n.e.d. souza, j.v. visentainer. voltammetric determination of pyridoxine (vitamin b6) in drugs using a glassy carbon electrode modified with chromium (iii) hexacyanoferrate (ii). journal of the brazilian chemical society 20 (2009) 496-501. https://doi.org/10.1590/s0103-50532009000300014 [6] s.j. devaki, r.l. raveendran. vitamin c: sources, functions, sensing and analysis. in vitamin c. intechopen (2017). https://doi.org/10.5772/intechopen.70162 [7] j.m. zen, d.m. tsai, h.h. yang, direct determination of vitamin c in fruit juices using a polyviologen‐ modified electrode. electroanalysis: an international journal devoted to fundamental and practical aspects of electroanalysis 14(22) (2002) 1597-1600. https://doi.org/10.1002/1521-4109(200211)14:22 %3c1597::aid-elan1597%3e3.0.co;2-u. [8] i. škugor rončević, d. skroza, i. vrca, a.m. kondža, n. vladislavić. development and optimization of electrochemical method for determination of vitamin c. chemosensors 10(7) (2022) 283. https://doi.org/ 10.3390/chemosensors10070283 [9] f. shayanfar, h. sarhadi. determination of vitamin c at modified screen printed electrode: application for sensing of vitamin c in real samples. surface engineering and applied electrochemistry 57 (2021) 487494. https://doi.org/10.3103/s1068375521040141 [10] a. baghizadeh, h. karimi-maleh, z. khoshnama, a. hassankhani, m. abbasghorbani. a voltammetric sensor for simultaneous determination of vitamin c and vitamin b 6 in food samples using zro 2 nanoparticle/ionic liquids carbon paste electrode. food analytical methods 8 (2015) 549-557. https://doi.org/ 10.1007/s12161-014-9926-3 [11] h. bakhsh, i.m. palabiyik, r.k. oad, n. qambrani, j.a. buledi, a.r. solangi, s.t.h. sherazi. sno2 nanostructure based electroanalytical approach for simultaneous monitoring of vitamin c and vitamin b6 in pharmaceuticals. journal of electroanalytical chemistry 910 (2022) 116181. https://doi.org/10.10 16/j.jelechem.2022.116181 [12] t.k. patle, k. shrivas, a. patle, s. patel, n. harmukh, a. kumar. simultaneous determination of b1, b3, b6 and c vitamins in green leafy vegetables using reverse phase-high performance liquid chromatography. microchemical journal 176 (2022) 107249. https://doi.org/10.1016/j.microc.2022.107249 [13] y.b. monakhova, s.p. mushtakova, s.s. kolesnikova. determination of vitamins in mixtures of various composition by spectrophotometry with self-modeling curve resolution. journal of analytical chemistry 65 (2010) 588-595. https://doi.org/10.1134/s1061934810060079 [14] l.g. shaidarova, l.n. davletshina, g.k. budnikov. flow-injection determination of water-soluble vitamins b 1, b 2, and b 6 from the electrocatalytic response of a graphite electrode modified with a ruthenium (iii) hexacyanoruthenate (ii) film. journal of analytical chemistry 61 (2006) 502-509. https://doi.org/ 10.1134/s1061934806050133 [15] m. kolar, d. dobcnik, n. radić. potentiometric flow-injection determination of vitamin c and glutathione with a chemically prepared tubular silver electrode. die pharmazie 55(12) (2000) 913-916. [16] h. karimi-maleh, f. karimi, y. orooji, g. mansouri, a. razmjou, a. aygun, f. sen. a new nickel-based cocrystal complex electrocatalyst amplified by nio dope pt nanostructure hybrid; a highly sensitive approach for determination of cysteamine in the presence of serotonin. scientific reports 10(1) (2020) 1-13. https://doi.org/10.1038/s41598-020-68663-2 [17] m. bijad, a. hojjati-najafabadi, h. asari-bami, s. habibzadeh, i. amini, f. fazeli. an overview of modified sensors with focus on electrochemical sensing of sulfite in food samples. eurasian chemical communications 3(2) (2021) 116-138. https://doi.org/10.22034/ecc.2021.268819.1122 [18] h. beitollahi, s.z. mohammadi, m. safaei, s. tajik. applications of electrochemical sensors and biosensors based on modified screen-printed electrodes: a review. analytical methods 12 (2020) 15471560. https://doi.org/10.1039/c9ay02598g [19] s. saghiri, m. ebrahimi, m. bozorgmehr. electrochemical amplified sensor with mgo nanoparticle and ionic liquid: a powerful strategy for methyldopa analysis. chemical methodologies 5(3) (2021) 234-239. https://doi.org/10.22034/chemm.2021.128530 https://doi.org/10.1590/s0103-50532009000300014 https://doi.org/10.5772/intechopen.70162 https://doi.org/10.1002/1521-4109(200211)14:22%3c1597::aid-elan1597%3e3.0.co;2-u https://doi.org/10.1002/1521-4109(200211)14:22%3c1597::aid-elan1597%3e3.0.co;2-u https://doi.org/10.3390/chemosensors10070283 https://doi.org/10.3390/chemosensors10070283 https://doi.org/10.3103/s1068375521040141 https://doi.org/10.1007/s12161-014-9926-3 https://doi.org/10.1007/s12161-014-9926-3 https://doi.org/10.1016/j.jelechem.2022.116181 https://doi.org/10.1016/j.jelechem.2022.116181 https://doi.org/10.1016/j.microc.2022.107249 https://doi.org/10.1134/s1061934810060079 https://doi.org/10.1134/s1061934806050133 https://doi.org/10.1134/s1061934806050133 https://doi.org/10.1038/s41598-020-68663-2 https://doi.org/10.22034/ecc.2021.268819.1122 https://doi.org/10.1039/c9ay02598g https://doi.org/10.22034/chemm.2021.128530 admet & dmpk 11(2) (2023) 251-261 determination of vitamin b6 in the presence of vitamin c doi: https://doi.org/10.5599/admet.1702 259 [20] a. shamsi, f. ahour. electrochemical sensing of thioridazine in human serum samples using modified glassy carbon electrode. advanced journal of chemistry-section a 4(1) (2021) 22-31. https://doi.org/ 10.22034/ajca.2020.252025.1215 [21] j. mohanraj, d. durgalakshmi, r.a. rakkesh, s. balakumar, s. rajendran, h. karimi-maleh. facile synthesis of paper based graphene electrodes for point of care devices: a double stranded dna (dsdna) biosensor. journal of colloid and interface science 566 (2020) 463-472. https://doi.org/10.1016/j.jcis.2020.01.089 [22] m. mazloum-ardakani, h. beitollahi, b. ganjipour, h. naeimi. novel carbon nanotube paste electrode for simultaneous determination of norepinephrine, uric acid and d-penicillamine. international journal of electrochemical science 5 (2010) 531-546. [23] a. lohrasbi-nejad. electrochemical strategies for detection of diazinon. journal of electrochemical science and engineering 12(6) (2022) 1041-1059. https://doi.org/10.5599/jese.1379 [24] m.s. sengar, s. saxena, s.p. satsangee, r. jain. silver nano-particles decorated functionalized multiwalled carbon nanotubes modified screen printed sensor for voltammetric determination of butorphanol. journal of applied organometallic chemistry 1(2) (2021) 95-108. https://doi.org/10.22034/jaoc.2021.289344.1023 [25] y. f. mustafa, g. chehardoli, s. habibzadeh, z. arzehgar. electrochemical detection of sulfite in food samples. journal of electrochemical science and engineering 12(6) (2022) 1061-1079. https://doi.org/ 10.5599/jese.1555 [26] m.m. ardakani, z. taleat, h. beitollahi, m. salavati-niasari, b.b.f. mirjalili, n. taghavinia. electrocatalytic oxidation and nanomolar determination of guanine at the surface of a molybdenum (vi) complex–tio2 nano-particle modified carbon paste electrode. journal of electroanalytical chemistry 624 (2008) 73-78. https://doi.org/10.1016/j.jelechem.2008.07.027 [27] m. miraki, h. karimi-maleh, m. a. taher, s. cheraghi, f. karimi, s. agarwal, v. k. gupta. voltammetric amplified platform based on ionic liquid/nio nanocomposite for determination of benserazide and levodopa. journal of molecular liquids 278 (2019) 672-676. https://doi.org/10.1016/j.molliq.20 19.01.081 [28] a. hosseini fakhrabad, r. sanavi khoshnood, m.r. abedi, m. ebrahimi. fabrication a composite carbon paste electrodes (cpes) modified with multi-wall carbon nano-tubes (mwcnts/n, n-bis (salicyliden)-1,3-propandiamine) for determination of lanthanum (iii). eurasian chemical communications 3(9) (2021) 627-634. https://doi.org/10.22034/ecc.2021.288271.1182 [29] s. mohammadi, h. beitollahi, a. mohadesi. electrochemical behaviour of a modified carbon nanotube paste electrode and its application for simultaneous determination of epinephrine, uric acid and folic acid. sensor letters 11(2) (2013) 388-394. https://doi.org/10.1166/sl.2013.2723 [30] h. beitollahi, n. arbabi. nanolayeredti3c2 modified screen printed electrode as high‐performance electrode for electrochemical detection of tyrosine. chemical methodologies 6(4) (2022) 293-300. https://doi.org/10.22034/chemm.2022.328263.1436 [31] o.m. istrate, l. rotariu, c. bala. amperometric l-lactate biosensor based upon a gold nanoparticles/reduced graphene oxide/polyallylamine hydrochloride modified screen-printed graphite electrode. chemosensors 9(4) (2021) 74. https://doi.org/10.3390/chemosensors9040074 [32] e.p. medyantseva, d.v. brusnitsyn, r.m. varlamova, m.a. baibatarova, g.k. budnikov, a.n. fattakhova. determination of antidepressants using monoamine oxidase amperometric biosensors based on screenprinted graphite electrodes modified with multi-walled carbon nanotubes. pharmaceutical chemistry journal 48 (2014) 478-482. https://doi.org/10.1007/s11094-014-1135-2 [33] z. mehdizadeh, s. shahidi, a. ghorbani-hasansaraei, m. limooei, m. bijad. monitoring of amaranth in drinking samples using voltammetric amplified electroanalytical sensor. chemical methodologies 6(3) (2022) 246-252. https://doi.org/10.22034/chemm.2022.324073.1423 [34] h. karimi-maleh, m. sheikhshoaie, i. sheikhshoaie, m. ranjbar, j. alizadeh, n.w. maxakato, a. abbaspourrad. a novel electrochemical epinine sensor using amplified cuo nano-particles and an-hexylhttps://doi.org/10.5599/admet.1702 https://doi.org/10.22034/ajca.2020.252025.1215 https://doi.org/10.22034/ajca.2020.252025.1215 https://doi.org/10.1016/j.jcis.2020.01.089 https://doi.org/10.1016/j.jcis.2020.01.089 https://doi.org/10.5599/jese.1379 https://doi.org/10.22034/jaoc.2021.289344.1023 https://doi.org/10.5599/jese.1555 https://doi.org/10.5599/jese.1555 https://doi.org/10.1016/j.jelechem.2008.07.027 https://doi.org/10.1016/j.molliq.20‌19.01.081 https://doi.org/10.1016/j.molliq.20‌19.01.081 https://doi.org/10.22034/ecc.2021.288271.1182 https://doi.org/10.1166/sl.2013.2723 https://doi.org/10.22034/chemm.2022.328263.1436 https://doi.org/10.3390/chemosensors9040074 https://doi.org/10.1007/s11094-014-1135-2 https://doi.org/10.22034/chemm.2022.324073.1423 p. m. jahani et al. admet & dmpk 11(2) (2023) 251-261 260 3-methylimidazolium hexafluorophosphate electrode. new journal of chemistry 43(5) (2019) 23622367. https://doi.org/10.1039/c8nj05581e [35] r.m. mohabis, f. fazeli, i. amini, v. azizkhani. an overview of recent advances in the detection of ascorbic acid by electrochemical techniques. journal of electrochemical science and engineering 12(6) (2022) 1081-1098. https://doi.org/10.5599/jese.1561 [36] s. azimi, m. amiri, h. imanzadeh, a. bezaatpour. fe3o4@sio2-nh2/cosb modified carbon paste electrode for simultaneous detection of acetaminophen and chlorpheniramine. advanced journal of chemistry-section a 4(2) (2021) 152-164. https://doi.org/10.22034/ajca.2021.275901.1246 [37] s.a. alavi-tabari, m.a. khalilzadeh, h. karimi-maleh. simultaneous determination of doxorubicin and dasatinib as two breast anticancer drugs uses an amplified sensor with ionic liquid and zno nanoparticle. journal of electroanalytical chemistry 811 (2018) 84-88. https://doi.org/10.1016/j.jelechem.20 18.01.034 [38] m. mirzaei, o. gulseren, m. rafienia, a. zare. nanocarbon-assisted biosensor for diagnosis of exhaled biomarkers of lung cancer: dft approach. eurasian chemical communications 3(3) (2021) 154161. http://dx.doi.org/10.22034/ecc.2021.269256.1126, [39] p.m. jahani. flower-like mos2 screen-printed electrode based sensor for the sensitive detection of sunset yellow fcf in food samples. journal of electrochemical science and engineering 12(6) (2022) 1099-1109. https://doi.org/10.5599/jese.1413 [40] h. karimi-maleh, a.f. shojaei, k. tabatabaeian, f. karimi, s. shakeri, r. moradi. simultaneous determination of 6-mercaptopruine, 6-thioguanine and dasatinib as three important anticancer drugs using nanostructure voltammetric sensor employing pt/mwcnts and 1-butyl-3-methylimidazolium hexafluoro phosphate. biosensors and bioelectronics 86 (2016) 879-884. https://doi.org/10.1016/j.bi os.2016.07.086 [41] m. vardini, n. abbasi, a. kaviani, m. ahmadi, e. karimi. graphite electrode potentiometric sensor modified by surface imprinted silica gel to measure valproic acid. chemical methodologies 6(5) (2022) 398-408. https://doi.org/10.22034/chemm.2022.328620.1437 [42] s. ariavand, m. ebrahimi, e. foladi. design and construction of a novel and an efficient potentiometric sensor for determination of sodium ion in urban water samples. chemical methodologies 6(11) (2022) 886-904. https://doi.org/10.22034/chemm.2022.348712.1567 [43] j.b. raoof, r. ojani, h. beitollahi, r. hosseinzadeh. electrocatalytic oxidation and highly selective voltammetric determination of l-cysteine at the surface of a 1-[4-(ferrocenyl ethynyl) phenyl]-1ethanone modified carbon paste electrode. analytical sciences 22(9) (2006) 1213-1220. https://doi.org/ 10.2116/analsci.22.1213 [44] t. eren, n. atar, m. l. yola, h. karimi-maleh. a sensitive molecularly imprinted polymer based quartz crystal microbalance nanosensor for selective determination of lovastatin in red yeast rice. food chemistry 185 (2015) 430-436. https://doi.org/10.1016/j.foodchem.2015.03.153 [45] s. saghiri, m. ebrahimi, m.r. bozorgmehr. nio nanoparticle/1-hexyl-3-methylimidazolium hexafluorophosphate composite for amplification of epinephrine electrochemical sensor. asian journal of nanosciences and materials 4(1) (2021) 46-52. https://doi.org/10.26655/ajnanomat.2021.1.4 [46] y. dessie, s. tadesse. a review on advancements of nanocomposites as efficient anode modifier catalyst for microbial fuel cell performance improvement. journal of chemical reviews 3(4) (2021) 320344. https://doi.org/10.22034/jcr.2021.314327.1128 [47] z.r. al-bahadili, a.a.s. al-hamdani, l.a. al-zubaidi, f.a. rashid, s.m. ibrahim. an evaluation of the activity of prepared zinc nano-particles with extract alfalfa plant in the treatments of peptidase and ions in water. chemical methodologies 6(7) (2022) 522-533. https://doi.org/10.22034/chemm.2022.336588.1470 [48] e. ezzatzadeh. chemoselective oxidation of sulfides to sulfoxides using a novel zn-dabco functionalized fe3o4 mnps as highly effective nanomagnetic catalyst. asian journal of nanosciences and materials 4(2) (2021) 125-136. https://doi.org/10.26655/ajnanomat.2021.2.3 https://doi.org/10.1039/c8nj05581e https://doi.org/10.5599/jese.1561 https://doi.org/10.22034/ajca.2021.275901.1246 https://doi.org/10.1016/j.jele‌chem.2018.01.034 https://doi.org/10.1016/j.jele‌chem.2018.01.034 http://dx.doi.org/10.22034/ecc.2021.269256.1126 https://doi.org/10.5599/jese.1413 https://doi.org/10.1016/j.bios.2016.07.086 https://doi.org/10.1016/j.bios.2016.07.086 https://doi.org/10.22034/chemm.2022.328620.1437 https://doi.org/10.22034/chemm.2022.348712.1567 https://doi.org/10.2116/analsci.22.1213 https://doi.org/10.2116/analsci.22.1213 https://doi.org/10.1016/j.foodchem.2015.03.153 https://doi.org/10.26655/ajnanomat.2021.1.4 http://www.jchemrev.com/article_140627.html http://www.jchemrev.com/article_140627.html https://doi.org/10.22034/jcr.2021.314327.1128 https://doi.org/10.22034/chemm.2022.336588.1470 https://doi.org/10.22034/chemm.2022.336588.1470 https://doi.org/10.26655/ajnanomat.2021.2.3 admet & dmpk 11(2) (2023) 251-261 determination of vitamin b6 in the presence of vitamin c doi: https://doi.org/10.5599/admet.1702 261 [49] h. karimi-maleh, c. karaman, o. karaman, f. karimi, y. vasseghian, l. fu, a. mirabi. nanochemistry approach for the fabrication of fe and n co-decorated biomass-derived activated carbon frameworks: a promising oxygen reduction reaction electrocatalyst in neutral media. journal of nanostructure in chemistry (2022)1-11. https://doi.org/10.1007/s40097-022-00492-3 [50] f. zare kazemabadi, a. heydarinasab, a. akbarzadehkhiyavi, m. ardjmand. development, optimization and in vitro evaluation of etoposide loaded lipid polymer hybrid nano-particles for controlled drug delivery on lung cancer. chemical methodologies 5(2) (2021) 135-152. https://doi.org/10.22034/chemm.2021.121495 [51] y.y. muhi-alden, k.a. saleh. removing methylene blue dye from industrial wastewater using polyacrylonitrile/iron oxide nanocomposite. eurasian chemical communications 3(10) (2021) 755-762. http://dx.doi.org/10.22034/ecc.2021.300767.1225 [52] g. osamong, p.k. kamweru, j.m. gichumbi, f.g. ndiritu. surface potential, fermi level and band gap energy of copper doped magnesium nickel ferrite nano-particles. asian journal of nanosciences and materials 4(1) (2021) 1-14. https://doi.org/10.26655/ajnanomat.2021.1.1 [53] m. ognjanović, d.m. stanković, y. ming, h. zhang, b. jančar, b. dojčinović, b. antić. bifunctional (zn, fe) 3o4 nano-particles: tuning their efficiency for potential application in reagentless glucose biosensors and magnetic hyperthermia. journal of alloys and compounds 777 (2019) 454-462. https://doi.org/10.1016/j.jallcom.2018.10.369 [54] d.m. stanković, s. škrivanj, n. savić, a. s. nikolić, p. vulić, d.d. manojlović. application of novel zn‐ ferrite modified glassy carbon paste electrode as a sensor for determination of cd (ii) in waste water. electroanalysis 26(7) (2014) 1536-1543. https://doi.org/10.1002/elan.201400095 [55] a.v. shinde, s.j. patil, s.k. hwang, g.s.r. raju, y.s. huh, y.k. han, n.r. chodankar. surface modified zinc ferrite as a carbon-alternative negative electrode for high-energy hybrid supercapacitor. ceramics international 47(11) (2021) 16333-16341. https://doi.org/10.1016/j.ceramint.2021.02.213 ©2023 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) https://doi.org/10.5599/admet.1702 https://doi.org/10.1007/s40097-022-00492-3 https://doi.org/10.22034/chemm.2021.121495 http://dx.doi.org/10.22034/ecc.2021.300767.1225 https://doi.org/10.26655/ajnanomat.2021.1.1 https://doi.org/10.1016/j.jallcom.2018.10.369 https://doi.org/10.1002/elan.201400095 https://doi.org/10.1016/j.ceramint.2021.02.213 http://creativecommons.org/licenses/by/3.0/ manuscript doi: 10.5599/admet.3.2.162 110 admet & dmpk 3(2) (2015) 110-121; doi: 10.5599/admet.3.2.162 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index review the complexity of roles of p-glycoprotein in refractory epilepsy: pharmacoresistance, epileptogenesis, sudep and relapsing marker after surgical treatment alberto lazarowski 1,2,£ *, liliana czornyj 2,3,£ , luisa rocha 4,£ 1 infibioc, ffyb-university of buenos aires (uba), junin 956 (1113), caba, argentina 2 fundacion investigar, riobamba 426 p-15 (1025), caba, argentina 3 servicio de neurologia, hospital juan p. garraham, combate de los pozos 1881 (1245), caba, argentina. 4 pharmacobiology department, center for research and advanced studies, df, mexico. £ members of geniar (cyted 610rt0405) *corresponding author: e-mail: alazarowski@gmail.com; tel +54-11-4304-2611 received: january 29, 2015; revised: may 04, 2015; published: july 01, 2015 abstract as described initially from clinical and experimental studies, p-glycoprotein (p-gp) plays a central role in the pharmacoresistance of epilepsy, acting by efflux of aeds mainly at blood brain barrier (bbb) level. however, repetitive seizures can produce both brain and heart p-gp overexpression. because p-gp activity induces membrane depolarization, its neuronal expression could be acting in the intrinsic mechanism of epileptogenesis, and its heart expression, can be a high risk factor of death, after severe -continuo convulsive stresses as in fatal status epilepticus or in sudep. additionally, because p-gp is also a stem cell marker, we suggests that its constitutive overexpression in dysplastic neurons from brain epileptogenic areas observed in patients with refractory epilepsies, should be addressed as a risk factor of seizures relapse after surgical treatment. here we discuss these concepts, based on our own clinical and experimental experiences, and reviewing the current literature on these subjects. keywords p-glycoprotein; membrane depolarization; drug resistance; sudep, biomarker, long-term relapse “what is a seizure and what is epilepsy?” seizures and epilepsy are not the same. it is defined that seizure as “the abnormal excessive or synchronous neuronal activity in the brain” that can be produced secondary to a wide spectrum of injuries. however, “epileptic seizures” are produced spontaneously, so, a seizure is an event and epilepsy is a disorder involving recurrent unprovoked seizures. after several years of deliberations on this issue have now been published the result from international league against epilepsy (ilae) commissioned second task force, to develop a practical (operational) definition of epilepsy, designed for use by doctors and patients [1], and adopted as a position of the ilae. some comments were published to better understanding of this definition, however, it is limited to a observational clinical level [2,3], and not http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:alazarowski@gmail.com admet & dmpk 3(2) (2015) 110-121 the roles of p-gp in refractory epilepsy doi: 10.5599/admet.3.2.162 111 describe the cellular or sub-cellular mechanistic process that involve alterations on functional properties of the neuronal membrane. epilepsy is the second most common neurological disorder after cerebrovascular accident (cva) or stroke, and it is estimated that approximately ~1.0-2.0 % of the population is affected by different forms of epilepsy. interestingly, near 30 % of epilepsies, have identifiable injuries to the brain that triggered the development of the named symptomatic epilepsies [4]. so, a non-specific damaging factor producing a brain injury can develop symptomatic epilepsy after a very complex subclinical process, which links this mentioned first insult with the clinically expressed first spontaneous epileptic seizure. and this particular and multifactorial process is recognized as latency phase or “epileptogenesis”. this cellular/molecular phenomenon is started after the brain injury and can be facilitated by some conditions as age, co-morbidity, medications, etc. blood brain barrier (bbb) damage, angiogenesis, gliosis, inflammation, axonal sprouting or injury, acquired channelopathies, and changes of dendritic plasticity, are some of long list of modifications observed during this latency phase. because a wide spectrum of different forms of brain injuries can lead in a secondary epileptic syndrome, it is important to add that the presence of background genes, developmentally regulated genes and/or epigenetic factors, can play a role in these process. perhaps, stroke, brain hypoxia and/or status epilepticus are the more clear conditions that can develop secondary epilepsy as well as to disclose a refractory phenotype [5]. neuronal excitability can arise from structural/functional changes at the postsynaptic membrane level, alterations in the type, number, and distribution of voltageand ligand-gated ion channels or biochemical modification of receptors that increase permeability to ca 2+ , favoring development of the prolonged depolarization that precedes seizures. excessive excitation is the key factor underlying ictogenesis, which may originate from individual neurons, the neuronal environment, or neuronal networks [6]. the imbalance between the excitatory (high) and the inhibitory (reduced) stimuli are the main mechanism of neuronal membrane depolarization and the onset of seizures. neuronal axons have a resting membrane potential of about -70 mv inside vs outside. action potentials occur due to net positive inward ion fluxes, resulting in local changes in the membrane potential. membrane potentials vary with the activation of either ligandor voltage-gated ion channels, which are affected by changes in either the membrane potential or intracellular ion concentration. when a neuron is at -70 mv it is said to be at resting potential. here, a neuron is not being excited by anything and is basically asleep, awaiting input from other neurons. resting potential is maintained by the sodiumpotassium pump, which constantly pushes sodium out of the neuron. the pump operates via a bunch of sodium-potassium transporters located along the membrane. these transporters push three sodium molecules out for every two potassium molecules it takes in. resting potential exists when sodium is on the outside of the cell and potassium is on the inside. when a neuron is excited the membrane potential shoots from -70 mv to about 40 mv and the membrane is depolarized. the excitability of synaptic terminals depends on the amount of excitatory neurotransmitter released (e.g., glutamic acid or activation of glutamergic receptors) as well as of an insufficient amount of inhibitor neurotransmitter released (e.g., γ-aminobutyric acid, or gaba), and consequently of the final extent of membrane depolarization produced. interestingly, through the axis glutamate/nmda receptor/cyclooxygenase-2 (cox-2) signaling, glutamate can induces overexpression of p-glycoprotein (p-gp) in brain [7,8]. the efflux of drugs by the multidrug transporters, particularly p-gp lazarowski et al. admet & dmpk 3(2) (2015) 110-121 112 (abcb1-transporter) is one of the principal means by which several patients with pharmacoresistant epilepsy develops refractoriness to a wide spectrum of antiepileptic drugs (aeds) [9,10]. the questions are how so different factors and cellular/molecular process underlying during brain injury, can induce an enduring phenomenon with chronically recurrent seizures and why persistent (without control) convulsive stress is a main risk factor of develop pharmacoresistant epilepsy? so, is the acquired refractoriness an inducible process? abc-transporters: p-glycoprotein the movement of different compounds as solutes, nutrients, ions, toxics or drugs across cell membranes is a critical biological process where the atp-binding cassette (abc) transporter´s family plays a central role and represents one of the largest families of transport proteins [11-13]. several functional properties of these transporters have been described and their importance is indicated not only by the large number of encoding genes, or their ubiquitous expression, but also by their ancestral origin and because 22 of 48 human abc proteins are implicated in human diseases [14]. the presence of these pumps in the liver, bowel, and kidney and bbb suggests that these transport systems have evolved to protect the organism against, endogenous/circulating chemicals, xenobiotics, organic ions, and drugs [15]. particularly the p-gp was the first eukaryotic abc-transporter identified conferring multidrug resistance to cancer cells [16,17]. p-gp has a molecular architecture with a core structure of two pairs of domains of nucleotide binding domain (nbd) where atp is hydrolyzed and the transmembrane domain (tmd) that consists of six transmembrane helices and where is the drugbinding site. the mechanistic way of drugs transport is based on conformational changes at the nbds level, transmitted to the tmds, and producing the drug export that is energy atp-dependent [18,19]. an ‘‘alternating catalysis model’’ have been described indicating that after atp hydrolysis, adp could be trapped at one of two nbds. normally, in the absence of nucleotide (state i), the two opposing domains of the atp sites are in the ‘‘open’’ conformation, and the binding of two molecules of atp results in the formation of the nucleotide sandwich (state ii) that draws together the opposing faces of the atp sites. this is followed by a conformational change in which one of the two atps is ‘‘occluded’’ in a non-exchangeable form (state iii) (figure 1)[20]. abc-transporters in the brain the central nervous system (cns) effects of many therapeutic drugs are blunted because of restricted entry into the brain. the basis for this poor permeability is the brain capillary endothelium, which comprises the bbb. this tissue exhibits very low paracellular (tight junction, or tj) permeability and expresses a potent system of drug export pumps. the bbb is a highly specialized structural and biochemical barrier that regulates the entry of blood-borne molecules and cells into the brain and preserves ionic homeostasis within the brain microenvironment. admet & dmpk 3(2) (2015) 110-121 the roles of p-gp in refractory epilepsy doi: 10.5599/admet.3.2.162 113 figure 1: drug efflux cycle: the drug binds first to the tmd producing conformational changes of abc-transporter. a) state i: the inactive structure of p-gp. b) state ii: the inward-facing conformation. c) state iii: the outward-facing conformation transporters are abundantly expressed in the brain, mainly in the vascular endothelial cells (vec) of the bbb, the epithelial cells of the brain-cerebrospinal fluid barrier (bcsfb), and also some parenchyma cells in brain, particularly in the food-ending process of astrocytes-touching vec of vessels. drug uptake into the brain highly depends on the efflux transporters expressed at both bbb and bcsfb biological barrier systems. a polarized location of the drug efflux transporters is mainly present on the luminal (apical or blood) side of the endothelial cells in the brain capillary where p-gp, bcrp, and mrp2 are unequivocally expressed at the blood side. this natural distribution of the main drug-transporters at bbb, indicates that the normal presence of some abc-transporters (including p-gp) at bbb level, is not enough to produce the impairment of aeds entrance in the brain for the responders patients. the “transporter`s hypothesis” to explain the refractoriness in epilepsy need not only a functional up-regulation of these transporters at bbb level to clean/export aeds, but also they should be expressed on previously non-expressing cell as neurons. the inducible properties of these transporters as well as the the pharmacoresistant phenotype [9] links pharmacoresistance with epileptogenesis, as discussed below. p-glycoprotein and drug resistant phenotype in epilepsy one-half of newly diagnosed patients with epilepsy obtain full seizure control with the first aed trial, and 13 % more enter remission with the addition of a second drug. however, the 30-40 % of the remaining patients are not to obtain satisfactory seizure control in spite of different drug combinations. it is clear the wrong choice of aeds may be the cause of treatment failure. however, with the correct therapeutic stratagem, sanjay m. sisodiya (2006) suggests that patients, who were considered drug resistant, may not remain so as newer aeds are developed or designed to new target previously unappreciated underlying pathophysiological mechanisms. additionally, he suggests that for individuals who we could define as being drug resistant, perhaps their epilepsy is considered “drug resistant” simply because we do not yet have the appropriate drugs for the treatment [21]. however, in spite of the increased amount of new aeds developed during the last 30 years, the same percentage (30-40 %) of pharmacoresistant patients remains without modifications (fig. 2). lazarowski et al. admet & dmpk 3(2) (2015) 110-121 114 figure 2: the increase of new aeds not decrease the percentage of drug-resistant cases (modified from czornyj & lazarowski: genetics of epilepsy and refractory epilepsy. ed. michael dean. colloquium series on the genetic basis of human disease, morgan & claypool life sciences, 2013) after the initial descriptions of the potential association between p-gp overexpression in the brain and refractory epilepsy [22,23], several other groups have reported high levels of p-gp expression in epileptogenic brain specimens from patients with refractory epilepsy. in these studies, p-gp was highly expressed not only in endothelial cells of bbb but also in brain parenchymal cells [9,10,24-26]. similar results were observed by independent experimental studies of different epileptic models, such as a single dose of intra-cerebroventricular kainic acid, chronic epilepsy, repetitive ptz-induced seizures and pilocarpine-induced status epilepticus. in these studies, seizure-induced overexpression of p-gp at bbb and brain parenchyma cells as astrocytes and neurons were described [25,27-29]. furthermore, after repetitive induced seizures, the simultaneous high p-gp expressions in different brain regions and liver or heart have been reported [30,31]. p-glycoprotein and pharmacokinetics alterations refractory epilepsy is described in patients receiving recommended aeds doses, have an adequate therapeutic levels of aeds in plasma, but remains without control of seizures. it was demonstrated that therapeutic levels of phenytoin (pht) in blood and csf can be achieved independently of the route of administration, as long as conventional loading doses are used, and in the case of pht i.v. administration, 2 h are enough to reach steady state concentrations within the therapeutic range [32]. as previously described in two cases reports [23,33], patients can have persistent low plasmatic levels of at least one of the aeds administered, despite their scrupulous compliance with the prescribed drugs regimen. often, laboratory professionals and physicians have no rational explanations, and non-detectable errors in aed measurement procedures and methods are assumed. most aeds are substrates of p-gp, and the inducible nature of this abc-transporter gene suggests that over-expression of this protein can be observed in all admet & dmpk 3(2) (2015) 110-121 the roles of p-gp in refractory epilepsy doi: 10.5599/admet.3.2.162 115 excretory organs including bbb, playing a critical role in the modification of both systemic and local pharmacokinetics of aeds. experimentally, it was also demonstrated that brain p-gp overexpression, correlated with the lower hippocampal phenytoin concentration in a refractory epileptic model, and whereas the administration of an antagonist of p-gp such as nimodipine (2 mg/kg) restored the normal central pht pharmacokinetics [34]. clinically and experimentally, the uses of some calcium channel blockers have been described to increase plasmatic levels of aeds simultaneously with the recovery of seizures control [35,36]. it is known that abc-transporters are located at the apical membrane of both acinar and ductal cells [37]. those proteins are important determinants of interindividual differences in intestinal absorption or hepatic clearance, and hence plasma levels of some if not all aeds. these variations can also be detected by measurement of saliva drug concentration, and it is sensitive enough for detecting systemic clearance changes. furthermore, some aeds would modify their bioavailability and clearance by inducing efflux transporter throughout chronic treatments, from the first dose to multiple dose administration and it can be documented by saliva drug concentrations. it is important to notice the usefulness of salivary drug concentrations to determine systemic clearance in patients not only from a pharmacokinetic point of view but also as new index to phenotype epileptic patients as refractory to the pharmacological treatment [38]. iannetti et al. [39] recently described 11-years-old boy who developed status epilepticus, after a prolonged right-side simple partial motor seizure, which was unresponsive to long term aggressive treatment with several aeds. control of convulsive seizures was achieved at valproic acid plasma level of 108 μg/ml, but electrical status epilepticus persisted, and the child remained comatose. on day 37, a treatment with verapamil (blocker of calcium l-channel and p-gp) was started and 1.5 hour after the initiation the infusion, the patient regained consciousness, was able to breathe spontaneously, and the status epilepticus, promptly disappeared. the authors suggested that verapamil, a known p-gp inhibitor acted by facilitating the brain penetration of aeds simultaneously administrated to the patient. role of p-glycoprotein in epileptogenesis taking in account that in a normal brain, neurons do not express p-gp, which could be the functional changes in neurons expressing p-gp,? approximately one third of patients with epilepsy do not have satisfactory control of seizures with antiepileptic drugs (aeds) and it is suggested that drug resistant phenotype could be present at the early stage of the disease [40]. at present, antiepileptic drug refractoriness has been explained by the “reduced drug-target sensitivity” and the “drug transporter overexpression” hypotheses. mdr-1 gene upregulation could be acquired as a consequence of repetitive non-controlled seizures [28,29] as well as secondary to pathological conditions such as inflammation, tumors or hypoxia [41-44], supporting the idea that silent mechanisms could overexpress p-gp in brain before epilepsy and refractoriness becomes clinically evident. repetitive seizures, status epilepticus as well as brain hypoxia induces p-gp expression in neurons. this p-gp induced expression depends on the frequency and intensity of seizures and is related to a progressive increase of the pharmacoresistant phenotype [45]. furthermore, the induced “de novo” expression of p-gp in previously non-expressive cells as observed particularly in lazarowski et al. admet & dmpk 3(2) (2015) 110-121 116 neurons from epileptogenic brain areas, suggests a differential role of this transporter, perhaps related to the intrinsic convulsive mechanism [29,33,46]. in this regard, a group of evidence indicates that p-gp can also decrease the plasma membrane potential of several cell types [47,48] and modify swelling-activated cl − currents [49], both physiologic disturbances observed during brain hypoxia [50] and convulsive stress [51]. according to these observation, recently was reported the first study demonstrating that daily ptz-induced seizures produces simultaneously the progressive acquisition of pharmacoresistant phenotype, increased brain p-gp expression and high membrane depolarization in brain areas such as hippocampus and cortex [52]. in this study, the recovery of potential membrane of depolarized slides in hippocampus and neocotex (in vitro experiments) was induced when pht was combined with nimodipine, a calcium channel blocker that also inhibit p-gp function. in addition to the much known drug pump model, some studies indicate that p-gp may mediate drug resistance through channels function regulation [53]. supporting this idea, it has been described that mdr1 protein behaves as some type of cl transporter, stimulates cl channel activity and may alter cell volume in several cell types [54]. its overexpression leads to decreased plasma membrane electrical potential by increasing cl permeability and thus decreasing the dominance of membrane potential by k + conductance, and leads to altered phi regulation [47,48]. so, brain p-gp overexpression could contribute to crucial factors detected in several types of epilepsy such as the intracellular acidosis associated with altered na + /h + exchange [55], enhanced extracellular concentration of potassium [56], and particularly in the increased membrane depolarization [57]. considering all the mentioned evidences, the popular comments saying that “seizures induce seizures” and “seizures without control induce refractoriness” could have a mechanistic and molecular explanation, where progressive and induced p-gp expression in neurons reduces the membrane potential and the seizure thresholds, and increasing the risk of pharmacoressitance. consequently, how much time p-gp can be expressed in brain cells after an initial inducer insult? figure 3: neuronal expression of p-gp (white arrow) can be induced by seizures, hypoxia, inflammation, drugs, cytokines, tumors, metabolic disorders, toxics, trauma, or infections. irrespective which was the cause, in these conditions neuronal membranes will be depolarized. interestingly, all these mentioned factors are the more common conditions that develop acquired epilepsy. admet & dmpk 3(2) (2015) 110-121 the roles of p-gp in refractory epilepsy doi: 10.5599/admet.3.2.162 117 the time course for the expression of p-gp in astrocytes after status epilepticus following intracerebroventricular kainate injection, was 10 consecutive weeks [25], and after the cortical devascularization inducing focal hypoxia, p-gp expression in neurons was maximal at day 7, and remained lightly positive at day 28 [58]. p-gp could be expressed longer times depending on the severity and also the frequency of the repetitive inducer stimulus. so, we can speculate that time during neuronal p-gp remains expressed, results in a pharmacoresistant window, or the time during the patients have higher risk to get a new seizures, or both conditions. if it is the case, pharmacoresistance and epileptogenesis are sharing the same mechanistic risk factor: “p-gp expressed in neurons”. p-glycoprotein and sudden unexpected death in epilepsy (sudep) patients with refractory epilepsy carry an increased mortality risk of that is more elevated than in patients with well-controlled seizures. the sudden, unexpected, witnessed or unwitnessed, no traumatic and no drowning deaths in patients with epilepsy, is usually a clinical phenomenon in which post-mortem examination does not reveal a toxicological or anatomical cause of death [59]. the clinical characteristics of patients with epilepsy at particular risk for sudep are related with a long history (more than 15 years) of pharmacoresistant epilepsy or high seizure frequency (more than 15 seizures/moth). all these patients currently receive antiepileptic drug (aed) polypharmacy, and commonly changes between different therapeutic combinations and/or doses. persistent low levels of aeds in plasma as well as poor adherence to pharmacotherapy are also observed in these patients, and several of all of these features are associated to pharmacoresistant epilepsy [60]. different studies have suggested that seizure activity as inducer of cardiovascular alterations, changes in the respiratory rate including tachypnea, hypopnea or apnea, related also with heart ischemia, suggesting that respiratory and mainly cardiovascular abnormalities during and between seizures, could play a central role in sudep [61-63]. it was also experimentally demonstrated that after both acute or chronic heart hypoxia, p-glycoprotein results to be overexpressed in cardiomyocytes and associated with heart stunning [64,65]. according with these observations it is important notice that continuous convulsive stress could mimic a repetitive apneas or hypoxia and produce heart ischemia. in bases to these considerations, the progressive and simultaneous p-gp overexpression in brain and heart, related with the development of fatal status epilepticus after experimental repetitive seizures was recently demonstrated by our group [66]. furthermore, we can speculate that high p-gp expression in brain, particularly in neurons (as previously described above), could play a role on the risk for se development, and high p-gp expression in heart could play a role in the risk for develop an acute and fatal heart failure. p-gp as stem cell marker of brain malformation/tumor refractory epilepsies and/or relapsing marker after surgical treatment several malformations of cortical development as well as brain tumors are associated with refractory epileptic syndromes and/or developmental delay. the main developmental brain abnormalities identified are hemimegalencephaly, polymicrogyria, agyria-pachygyria, schizencephaly, heterotopic gray matter, transmantle dysplasia, focal cortical dysplasia, or cortical tubers. p-gp and others abc-transporters have been reported to be over-expressed in these malformations of cortical development and brain cortical tubers explaining why all of them are clinically characterized by phamacoresistant phenotype [22,23,67,68]. epilepsy is also common disorder in patients with brain tumors and can substantially affect lazarowski et al. admet & dmpk 3(2) (2015) 110-121 118 daily life, even if the tumor is under control. in these cases, multidrug-abc-transporters expressed at blood brain barrier, brain tumor cells and tumor-surrounding cells, prevent the access of antiepileptic drugs into brain parenchyma, and patients develop a refractory phenotype. in subependimmal giant astrocytoma (sega), abc-transporters (p-gp and bcrp) and cd34 stem cell markers were highly expressed in non-vascularized tumor cells. heterogeneous distributions for these markers were detected with differential immunostaining pattern, showing high immunoreactivity in sega cells far of vessels, and low or negative expression in sega cells near the vessels. this particular expression pattern of both abc-transporters, and cd34 antigen could identify different stem-cell subset in sega [69]. similarly, co-expression of cd34 and neurofilament were described in epilepsy-associated glioneuronal tumors [62,70]. in normal brain, cd34 occurs only transiently during neurulation [71] and recently cd34 immunoreactivity was also detected as a subpopulation of balloon cells confined to the white matter but not observed in neocortical layers from brain specimens of taylor's focal cortical dysplasia [72]. tissues from therapeutic resections of several refractory epilepsies, as dysembryoplastic neuroepithelial tumors, focal cortical dysplasia and hippocampal sclerosis, as well as several brain tumors as glial and glioneuronal tumors, are associated with pharmacoresistant epilepsy, having a much poorer pharmacological outcome compared to other symptomatic epilepsies. in all of them, overexpression of abc-transporters as p-glycoprotein, mrps and bcrp were described [10,24,45,69,73]. abc-transporters, particularly p-gp and bcrp, are transiently expressed at high levels in human neural stem/progenitor cells (hnspcs) but are downregulated in differentiated hnspcs, and in normal conditions, they are expressed in a highly regulated manner, with the highest expression in primitive cells and subsequent down-regulation following commitment to differentiation [74]. furthermore, the constitutive deficiency of abc-transporters leads to distinct impairments in neural stem/progenitor cells maturation and adult neurogenesis in vivo, indicating a functional role of abc-transporters in stem cell maintenance and differentiation [75]. all these data clearly suggests that presence of abc-transporters, in abnormal and immature brain cells as well as in immortalized tumor cells, are clearly a hallmark of a non-fully differentiated progenitor cell. if it is the case, abc-transporters together other stem cell markers, could build a new biomarkers profile to give predictive information for long-term follow-up of surgically treated refractory epilepsies and/or related with the risk of relapse of seizures, even before that a new brain structural anomaly can be detected. furthermore, at date, the available evidence demonstrates that longterm seizure freedom is achieved in about 60% of patients who undergo temporal lobe surgery, and in 30– 40 % of those who undergo extratemporal surgery [76]. conclusions it is important to distinguish between the truly drug-transporter properties of these proteins leading to the pharmacoresistant phenotype, from those related with plasmatic membrane depolarization inducing pro-epileptic effects, directly related with epileptogenesis as described for p-gp. in this sense, abctransporters, as p-glycoprotein and bcrp, could be interpreted as stem-cell markers presents in several brain cortical malformations, being constitutive components of immature not-fully differentiated cells, as observed in dysplastic neurons and ballooned cells or brain tumor cells. interestingly, all these abnormal cells that also play a role in the epileptogenesis, have high expression of abc-transporters and are also refractory to aeds. under these conditions, repetitive seizures can induce simultaneous p-gp overexpression in both brain and heart. it could be a combined risk factor to develop an acute heart failure under severe stress triggered by status epilepticus resulting in a fatal end. admet & dmpk 3(2) (2015) 110-121 the roles of p-gp in refractory epilepsy doi: 10.5599/admet.3.2.162 119 finally, the condition of abc-transporters as stem cell markers, if they are presents in those mentioned abnormal cells, could contribute to build a risk score or predictive profile for long-time seizures relapse after surgical treatment. references [1] r.s. fisher, c. acevedo, a. arzimanoglou, a. bogacz, j.h. cross, c.e. elger, j. jr. engel, l. forsgren, j.a. french, m. glynn, d.c. hesdorffer, b.i. lee, g.w. mathern, s.l. moshé, e. perucca, i.e. scheffer, t. tomson, m. watanabe, s. wiebe, epilepsia 55 (2014) 475-482. [2] r.s. fisher, epilepsia 55 (2014) 492–493. [3] e. perucca, epilepsia 55 (2014) 473–474. [4] w.a. hauser, incidence and prevalence. in: j. engel jr, t.a. pedley, editors. epilepsy: a comprehensive textbook. philadelphia: lippincott–raven;1997. p.47–57 [5] a. pitkänen and k. lukasiuk, epilepsy & behavior 14 (2009) 16–25. [6] s. engelborghs, r. d’hooge, p.p. de deyn, acta neurol. belg. 100 (2000) 201-13. [7] b. bauer, a.m. hartz, a. pekcec, k. toellner, d.s. miller, h. potschka, mol. pharmacol. 73 (2008) 1444-53. [8] r.w. robey, a. lazarowski, s.e. bates, mol. pharmacol. 73 (2008) 1343-6. [9] a. lazarowski, l. czornyj, f. lubienieki, e. girardi, s. vazquez, c. d´giano, epilepsia 48 (2007) s5,140–149. [10] s.m. sisodiya, w.r. lin, b.n. harding, m.v. squier, m. thom, brain 125 (2002) 22–31. [11] m. dean, y. hamon, g. chimini, j. lipid res. 42 (2001) 1007–17. [12] m. dean, t. annilo, annu. rev. genomics hum. genet. 6 (2005) 123–42. [13] m.m. gottesman, s.v. ambudkar, j. bioenerg. biomembr. 33 (2001) 453–8. [14] v. ling, j. nat. cancer inst. 81 (1989) 84–85. [15] c. cordon-cardo, j.p. o’brien, j. boccia, d. casals, j.r. bertino, m.r.melamed, j. histochem. cytochem. 138 (1990) 1277–1287. [16] m.m. gottesman, t. fojo, s.e. bates, nat. rev. cancer 2 (2002) 48–58. [17] c.f. higgins, i.d. hiles, g.p.c. salmond, d.r. gill, j.a. downie, i.j. evans, i.b. holland, l. gray, s.d. buckel, a.w. bell, m.a. hermodson, nature 323 (1986) 448–50. [18] a.l. davidson, j. bacteriol. 184 (2002) 1225–33. [19] c.f. higgins, k.j. linton, nat. struct. mol. biol. 11 (2004) 918–26. [20] z.e. sauna, s.v. ambudkar, mol. cancer ther. 6 (2007) 13-23. [21] s.m. sisodiya, epilepsia 46 (2005) (suppl 10) 33–8. [22] d. tishler, k. weinberg, d. hinton, n. barbaro, a. geralyn, r. corey, epilepsia 36 (1995) 1-6. [23] a. lazarowski, g. sevlever, a. taratuto, m. massaro, a. rabinowicz, ped. neurol. 21 (1999) 731-734. [24] a.j. lazarowski, f.j. lubieniecki, s.a. camarero, h.h. pomata, m.a. bartuluchi, g. sevlever, a.l. taratuto. pediatr. neurol. 34 (2006) 20-4. [25] l. zhang, w. ong, t. lee, exp. brain res. 126 (1999) 509-516. [26] u. seegers, h. potschka, w. loscher, epilepsia 43 (2002) 675-684. [27] m. rizzi, s. caccia, g. guiso, c. richichi, j.a. gorter, e. aronica, m. aliprandi, r. bagnati, r. fanelli, m. d'incalci, r. samanin, a. vezzani, j. neurosc. 22 (2002) 5833-5839. [28] p. kwan, g. still, e. butler, t. gant, b. meldrum, m. brodie, epilepsy 43 (2002) 1318-1323. [29] a. lazarowski, a.j. ramos, h. garcía-rivello, e. girardi, cell mol. neurobiol .24 (2004) 77-85. [30] e.a. van vliet, r. van schaik, p.m. edelbroek, r.a. voskuyl, s. redeker, e. aronica, w.j. wadman, j.a. gorter, jpet 322 (2007) 141–147. lazarowski et al. admet & dmpk 3(2) (2015) 110-121 120 [31] j. auzmendi, a. merelli, e. girardi, s. orozco-suarez, l rocha, a. lazarowski, mol. cell epilepsy 1 (2014) 3:1-9. [32] a.l. rabinowicz, j.m. salvat, r.c. leiguarda, f. demonty, f. salvat, a. cervio, f. manes, a. lazarowski. clin. neuropharmacol. 20 (1997) 438–441. [33] a. lazarowski, m. massaro, a. schteinschnaider, s. intruvini, g. sevlever, a. rabinowicz, ther. drug monit. 26 (2004) 44–46. [34] c. höcht, a. lazarowski, n.n. gonzalez, j. auzmendi, j.a. opezzo, g.f. bramuglia, c.a. taira, e. girardi, neurosci. lett. 413 (2007) 168-72. [35] j.r. hughes, neurol. res. 30 (2008) 920-5. [36] e.a. van vliet, r. van schaik, p.m. edelbroek, s. redeker, e. aronica, w.j. wadman, n. marchi, a. vezzani, j.a. gorter, epilepsia 47 (2006) 672-80. [37] t. uematsu, m. yamaoka, r. doto, h. tanaka, t. matsuura, k. furusawa, arch. oral biol. 48 (2003) 87-90. [38] p. fagiolino, m. vázquez, c. maldonado, m. esperanza ruiz, m.g. volonté, s. orozco-suárez, a. lazarowski, curr. pharm. des. 19 (2013) 6767-6774. [39] p. iannetti, a. spalice, p. parisi, epilepsia 46 (2005) 967-9. [40] ch. eger, epilepsia 44 (2004) suppl. 5: 9-15. [41] d.j. roberts, k.b.goralski, expert opin. drug metab. toxicol. 4 (2008) 1245-64. [42] k. kunishio, m. okada, y. matsumoto, s. nagao, y. nishiyama, j. med. invest. 53 (2006) 285-91. [43] k.m. comerford, t.j. wallace, j. karhausen, n.a. louis, m.c. montalto, s.p. colgan, cancer res. 62 (2002) 3387-94. [44] a. lazarowski, l. caltana, a. merelli, m. rubio, a. ramos, a. brusco, j. neurol. sci. 258 (2007) 84-92. [45] e. aronica, s.m. sisodiya, j.a. gorter, adv. drug deliv. rev. 64 (2012) 919–29. [46] e. aronica, j.a. gorter, g.h. jansen, c.w. van veelen, p.c. van rijen, s. leenstra, m. ramkema, g.l. scheffer, r.j. scheper, d. troost, neuroscience 118 (2003) 417–29. [47] r.m. wadkins, p.d. roepe. int. rev. cytol. 171 (1997) 121–65. [48] p.d. roepe, curr. pharm. des. 6 (2000) 241–60. [49] c. vanoye, a. castro, t. pourcher, l. reuss, g. altenberg, am. j. physiol. 276 (1999) c370–8. [50] m. müller, neuroscience 97 (2000) 33-45. [51] c. le duigou, v. bouilleret, r. miles, j. physiol. 586 (2008) (pt 20): 4891-904. [52] j.a. auzmendi, s. orozco-suárez, i. bañuelos-cabrera, m.e. gonzález-trujano, e. c. gonzález, l. rocha, a. lazarowski, curr. pharm. design 19 (2013) 6732-6738. [53] t. mizutani, m. masuda, e. nakai, k. furumiya, h. togawa, y. nakamura, y. kawai, k. nakahira, s. shinkai, k. takahashi, curr. drug metab. 9 (2008) 167-74. [54] m.a. valverde, m. diaz, f.v. sepulveda, d.r. gill, s.c. hyde, c.f. higgins, nature 355 (1992) 830-3. [55] b.k. siesjö, r. von hanwehr, g. nergelius, g. nevander, m. ingvar, j. cereb. blood flow metab. 5 (1985) 47-57. [56] d.m. durand, e.h. park, a.l. jensen, philos. trans. r. soc. lond. b biol. sci. 365 (2010):2347-62. [57] d.a. prince, b.w. connors, adv. neurol. 44 (1986) 275-99. [58] a.j. ramos, a. lazarowski, m. villar, a. brusco, cell. mol. neurobiol.24 (2004) 101-7. [59] l. nashef, epilepsia 38 (1997) 6-8. [60] s. kirby, r.m, epilepsia 36 (1995) 25-28. [61] l. nashef, f. walker, p. allen, j.w. sander, s. shorvon, d. fish, j. neurol. neurosurg. psychiatry 60 (1996)297–300. [62] s. tigaran, h. molgaard, r. mcclelland, m. dam, a.s. jaffe, neurology 60 (2003) 492–495. admet & dmpk 3(2) (2015) 110-121 the roles of p-gp in refractory epilepsy doi: 10.5599/admet.3.2.162 121 [63] s.u. schuele, p. widdess-walsh, a. bermeo, h. lüders, cleve. clin. j. med. 74 (2007) suppl 1:s121127. [64] a.j. lazarowski, h.j. garcía rivello, g.l. vera janavel, l.a. cuniberti, p.m. cabeza meckert, g.g. yannarelli, a. mele, a.j. crottogini, r.p. laguens. j. histochem. cytochem. 53 (2005) 845–850. [65] r.p. laguens, a.j. lazarowski, l.a. cuniberti, g.l. vera-janavel, p.m. cabeza meckert, g.g. yannarelli, h.f. del valle, e.c. lascano, j.a. negroni, a.j. crottogini, j. histochem. cytochem. 55 (2007) 191–197. [66] j.auzmendi, a. merelli, e. girardi, s. orozco-suarez, l. rocha, a. lazarowski, mol. cell. epilepsy 1 (2014) e66. doi: 10.14800/mce.66. [67] e. brodtkorb, g. nilsen, o. smevik, p.a. rinck, acta neurol. scand. 86 (1992) 24–32. [68] s.m. sisodiya, j. heffernan, m.v. squier, neuroreport 10 ((1999) 3437-41. [69] a. lazarowski, f. lubieniecki, s. camarero, v. cuccia, a.l.taratuto. receptor clin. invest. 1 (2014) e53. doi: 10.14800/rci.53 [70] j. fassunke, m. majores, c. ullmann, ch.e. elger, j. schramm, o.d. wiestler, a.j. becker, lab. invest. 84 (2004)1520–1525. [71] c. reali, f. scintu, r. pillai, s. cabras, f. argiolu, m.s. ristaldi, m.a. sanna, m. badiali, v. sogos. exp. neurol. 197(2006) 399-406. [72] s. fauser, a. becker, a. schulze-bonhage, m. hildebrandt, i. tuxhorn, h.w. pannek, r. lahl, j. schramm, i. blumcke, acta neuropathol. 108 (2004) 272-8. [73] m.s. van breemen, e.b. wilms, c.j. vecht. lancet neurol. 6 (2007) 421-30. [74] t. lin, o. islam, k. heese, cell res. 16 (2006) 857-871. [75] t. schumacher, m. krohn, j. hofrichter, c. lange, j. stenzel, j. steffen, t. dunkelmann, k. paarmann, c. fröhlich, a. uecker, a.s. plath, a. sommer, t. brüning, h.j. heinze, j. pahnke. plos one 7 (2012) e35613 [76] l. czornyj, a. lazarowski. front. biosci. 19 (2014) 1425-1435. ©2015 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ jese manuscript doi: 10.5599/admet.1.1.1 1 admet & dmpk 1(1) (2013) 1-2; doi: 10.5599/admet.1.1.1 open access : issn: 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index editorial close relationships between in vitro admet and dmpk research in pre-clinical drug discovery kin tam editor: admet & dmpk e-mail: kin_tam@iapchem.org received: december 20, 2012; revised: january 10, 2013; published: january 14, 2013 i have recently been invited to become an editor of a new open-access journal, admet & dmpk. i am honored to take up this exciting and challenging position. at first glance, the journal name is just a combination of two acronyms/disciplines in the pharmaceutical industry: admet (absorption, distribution, metabolism, excretion and toxicology) and dmpk (drug metabolism and pharmacokinetics), which sounds odd. practically speaking, admet refers to a suite of in vitro assays that address various aspects of the pharmacokinetic performance of a drug, while dmpk more often refers to a drug’s in vivo pharmacokinetic performance. on thinking about it in more depth, it makes perfect sense to consider these two disciplines together, as a good dmpk profile of a drug candidate would be difficult to achieve without optimal admet properties, and both play an indispensable role in pre-clinical drug discovery. here, i will share my views on the synergistic roles of these two disciplines, and how our new journal might help to foster the development of admet and dmpk. the discovery and development of a new candidate drug is a resource intensive and challenging process in the pharmaceutical industry. it involves evaluating the parameters affecting the likely success of a drug candidate in the preclinical, clinical and commercial phases of drug development. to maximize the chance of success in the clinic, it is crucial to generate and optimize quality lead compounds in the discovery phase. ideally, the discovery project should include: (1) a sound clinical hypothesis, (2) the right drug target and (3) the right chemical agents/series, with demonstrated exposure through the preferred route of administration in a pre-clinical species. to effectively guide drug discovery efforts during lead optimization, it is essential to develop a pharmacodynamic model, which provides a means with which to evaluate whether the exposure level of the tested compound is sufficient to elicit a desirable pharmacological response. clearly, pharmacokinetic studies provide a measure of the exposure level of the tested compound, which directly addresses item (3). the prevalence of high throughput and combinational methods in drug discovery has led to an increase in the number of new chemical entities to be evaluated in the early discovery phase. standard pharmacokinetic studies are no longer able to cope with the number of tests required. to mitigate the risk of failure due to poor pharmacokinetics, the pharmaceutical industry has developed medium/high throughput in vitro admet studies to evaluate potential lead compounds before carrying out http://www.pub.iapchem.org/ojs/index.php/admet/index� mailto:name@address.domain� kin tam admet & dmpk 1(1) (2013) 1-2 2 pharmacokinetic studies in pre-clinical species. table 1 lists representative examples of some in vitro admet assays. table 1. some common in vitro admet assays assay issue(s) to be addressed caco-2 permeability oral absorption transporter study oral absorption, drug–drug interactions protein/serum binding distribution hepatocyte stability metabolism gut stability metabolism cyp450 inhibition drug–drug interactions metabolite identification metabolic mechanism herg /cardiac ion channels qt liability, cardiotoxicity physicochemical properties such as aqueous solubility, dissolution, lipophilicity and chemical stability are often regarded as the key parameters governing the admet properties. for instance, a poorly soluble research compound may suffer from incomplete oral absorption, which could lead to low systematic exposure, and highly lipophilic compounds are likely to show high metabolic clearance. with the advance in numerical/statistical techniques, it is now possible to use the knowledge base from existing research compounds/drugs to build predictive models for in silico assessment of admet and physicochemical properties. the in silico approach is very useful in the drug design process and for prioritizing the synthetic and/or testing resources for those compounds that are most likely to show good pharmacokinetic profiles. it is not uncommon to encounter situations in which the potential lead compound shows good physiochemical properties and reasonable admet properties, but exhibits poor systematic exposure in a pharmacokinetic study. bespoke admet studies could then be used as problem solving tools to identify such issues. typical examples could be the differentiation between gut and hepatic metabolism, and the effects of uptake transporters and/or efflux transporters in absorption or clearance. in such cases, admet and dmpk studies should be conducted together as necessary with the aim of identifying the root cause and guiding the design of new compounds/approaches to address the problem. as the chemical series progresses, all available data from pharmacokinetic studies and in vitro admet studies could be used for scaling purpose. generally, a simple allometric approach may be used to extrapolate the animal data to human data and thus predict pharmacokinetic parameters and, together with pharmacokinetic/pharmacodynamic data, the therapeutic dose levels in humans. i hope this journal will provide a platform for all scientists working in admet and dmpk areas to publish their results/research findings in a timely manner. the journal welcomes original scientific contributions from all areas of absorption, distribution, metabolism, excretion, toxicology and pharmacokinetics of drugs. i firmly believe in the importance of exchange of knowledge and shared learning, which will help drug discovery scientists to come up with better quality lead compounds, and hopefully to minimize costly clinical failures. ©2013 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/� manuscript doi: 10.5599/admet.3.3.173 182 admet & dmpk 3(3) (2015) 182-189; doi: 10.5599/admet.3.3.173 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index review assessing blood brain barrier permeability in traumatic brain injury research george p. liao, benjamin m. aertker, daniel j. kota, karthik s. prabhakara, philippa smith, robert a. hetz, hasen xue, supinder bedi, scott d. olson and charles s. cox jr* department of pediatric surgery, university of texas health science center at houston, 6431 fannin street, msb 5.230, houston, tx 77030, usa *corresponding author: e-mail: charles.s.cox@uth.tmc.edu; tel.: +1-713-500-7307; fax: +1-713-500-7296 received: march 06, 2015; revised: july 09, 2015y; published: september 05, 2015 abstract the blood brain barrier plays an important role in traumatic brain injury, serving at the crossroads of secondary injury and potential therapies. in regards to trauma, this barrier contains an array of cellular and molecular components that protect the central nervous system from derangements in water homeostasis and inflammation. preclinical and clinical assays have been developed to describe and quantify blood brain barrier permeability in relation to the integrity of these blood brain barrier components and the handling of edema. this review will discuss both preclinical and clinical molecular and imaging techniques that are used to assess blood brain barrier function and recovery following traumatic brain injury. keywords extravasation; mri; alexa fluor; evans blue; edema introduction blood brain barrier (bbb) permeability and cerebral edema play crucial roles in the progression of secondary injury following traumatic brain injury (tbi), and are clinically associated with increased intracranial pressure and compromised cerebral perfusion [1]. the pathological consequence of bbb compromise is the exposure of the normally immunologically privileged central nervous system to an influx of neuroinflammatory cells and proinflammatory cytokines [2,3]. neuroinflammation promotes vasogenic edema and co-exists with cytotoxic edema from cellular dysfunction and dysregulation of intracellular volume [4,5]. the injury penumbra is the at-risk area adjacent to the injury site where the bbb is likely in transition between healthy and dysfunctional cerebral microvasculature, and can be evaluated in the preclinical setting via excess tissue water, bbb component protein up and downregulation and imaging the extravasation of various markers. clinically, bbb compromise can be estimated via microdialysis catheters or external ventricular drains as ratios between brain and serum total protein or albumin levels [6]. clinical visualization of penumbral bbb permeability in vivo has been begun to be possible using imaging modalities such as diffusion tensor magnetic resonance imaging and emerging in vivo fluorescence technologies [7]. this review will discuss bbb identification and use as a therapeutic target to protect and restore brain tissue following tbi. http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:charles.s.cox@uth.tmc.edu admet & dmpk 3(3) (2015) 182-189 blood-brain-barrier permeability in trauma research doi: 10.5599/admet.3.3.173 183 physiology the bbb is lined with capillary endothelial cells that lack fenestrations and are attached by tight junctions (comprised of claudin-5, occludin, zona ocludens-1 and ve-cadherin) that together restrict transport [8,9]. adenosine triphosphate-binding cassette transporters prevent drug penetration across the bbb. in comparison with peripheral endothelial cells, the bbb endothelial cells are much more energy dependent, containing five to ten times the amount of mitochondria [3]. these endothelial cells are further covered by astrocyte end processes and have basal lamina shared with luminal pericytes. astrocytes express high levels of aquaporin-4, a water channel protein that is known to be involved in the clearance of edema [10,11]. the mechanical forces of the primary injury transmitted to bbb disrupts blood flow and energy supply, compromising the barrier to the central nervous system, allowing for the passage of neuroinflammatory cells, potent osmotic proteins such as albumin, and neuro excitatory molecules such as glutamate [12]. subsequently, inflammation, edema and excitotoxicity contribute to clinically relevant secondary brain injury that may even manifest in epileptic events [13]. the resident astrocytes and microglia within the central nervous system are activated and also contribute to secondary injury, leading to further neuronal death [14]. preclinical measurements of bbb permeability a number of invasive strategies have been employed by preclinical investigators to assess the status of the bbb. the strategies include measuring brain water content, bbb components and extravasation. the first strategy is using brain water content, the simplest method of indirectly assessing the degree of bbb dysfunction and offers global differences between treatments, but is subject to variations in animal sacrifice technique and post mortem handling of brain tissues [15]. the second strategy is to measure the components involved in forming the bbb, such as perivascular astrocytes, tight junction proteins such as claudin-5, occludin and zona occludens-1, as well as the components responsible for maintaining and restoring water hemostasis, such as aquaporin-4 [10,16,17]. while this strategy can describe how individual components of the bbb change following injury and treatment, the effect of other components and the function of the bbb cannot be directly assessed. the third strategy is to use imaging with various molecular probes to detect the degree of extravasation present due to bbb permeability. evans blue dye extravasation is a common assay used in tbi research. evans blue has affinity to serum albumin (66 kda). following injury, the bbb, with pore size normally allowing for molecules of general 0.4 kda mass to pass, becomes disrupted, allowing for evans blue to leave the vasculature and extravasate into the interstitial tissue. the amount of extravasation can be quantified by extraction [18,19]. despite wide use, the evans blue assay is limited by the relatively large molecular size of albumin, the reliance on detection by absorbance via a wide fluorescence excitation spectrum, and harsh extraction process that destroys the brain tissue and precludes further histologic analysis on the same tissue sample. alexa fluor 680 (lifetechnologies, carlsbad, ca) is a far-red dye that can be bioconjugated to various molecular weight dextrans as well as other molecules and has been used in investigations for drug transit across the bbb [20]. high resolution (approaching 20 μm) infrared laser scanners can detect subtle changes in signal intensities of alexa fluor 680 between adjacent anatomical structures in brain tissue samples. we previously demonstrated that alexa fluor 680 reduces the non-specific signal between sham and injury to 7 % (p<0.001) when compared to evans blue, an eightfold reduction (figure 1). infrared image scanning g.p. liao et al. admet & dmpk 3(3) (2015) 182-189 184 precludes the need of detaching contralateral hemispheres for normalization and decreases signal to noise. maintaining intact brain architecture allows for the use of additional fluorescent markers to evaluate cell integrity and junctional proteins comprising the bbb. the tissue can be further processed for sectioning and staining. figure 1. non-specific signals are reduced with alexa fluor 680 conjugated to 10 kda dextran compared to evans blue dye as evidenced by the high signal from sham animals receiving evans blue. evans blue control cortical injury (cci) (n=11), evans blue sham (n=6), alexa fluor cci (n=9), alexa fluor sham (n=10). the y axis od/wt refers to optical density (absorbance) per brain weight. the x axis integrated density refers to the cumulated signal detected by the far-red scanner across the brain slices. figure 2. four ranges of intensity thresholds applied to the same coronal brain slice for a rat using imagej software. white areas demonstrate tissue edema, red areas represent the alexa fluor 680 conjugated to 10 kda dextran signal captured at 700nm using a li-cor odyssey clx infrared laser scanner. (a) no intensity threshold applied. (b) the low + narrow intensity range of 4000-5000, represents areas of low alexa fluor extravasation and suggests areas with a lower degree of bbb compromise within the at-risk injury penumbra. (c) the low + less narrow intensity range of 5000-7500 represents areas of higher alexa fluor extravasation and suggests areas with a higher degree of bbb compromise within the at-risk injury penumbra. (d) intensity range of 5000-maximum representing all areas of alexa fluor extravasation. we demonstrated that alexa fluor 680 dye conjugated to 10 kda dextran can be used in preclinical studies of tbi to suggest the location and degree of bbb permeability associated with at-risk penumbral admet & dmpk 3(3) (2015) 182-189 blood-brain-barrier permeability in trauma research doi: 10.5599/admet.3.3.173 185 regions. rats undergoing controlled cortical injury were designated to vehicle versus intravenous cell therapy groups and were sacrificed at 72 hours. prior to sacrifice the brains were perfused with alexa fluor 680 conjugated to 10 kda dextran, then following extraction were sectioned coronally into 1mm slices. these slices were then imaged using a li-cor odyssey clx infrared laser scanner (li-cor, lincoln, nebraska) at 700 and 800nm. raw images were then stacked, processed and analyzed in batch using fiji [21], the fully open source version of imagej 1.48p (http://imagej.nih.gov/ij). a low and narrow intensity threshold range identified a rim of alexa fluor signal in the area of the penumbral tissue that may be the subtle transition zone of bbb permeability [22] (figure 2). this threshold range demonstrated significant signal differences between the treated and control tbi rats. wider threshold ranges, with greater alexa fluor signal identified tissue associated with intraparenchymal contusions or microvascular hemorrhage. other infra-red imaging and immunohistochemical stragegies have also been used for extravasation studies. infra-red imaging contrast agents of various sizes have also been used by preclinical investigators in direct and indirect bbb permeability assays. an example of a small agent is the p-glycoprotein efflux transporter contrast molecule, rhodamine 800 (0.5 kda) [23,24]. larger molecular weight permeability markers include irdye 800cw which can vary in size from 15 kda and upwards, depending on conjugation [25]. fluorescein isothiocyanate conjugated to dextran or albumin has been used, as well as probing for igg penetration into the central nervous system after injury [26-30]. two-photon microscopy with tetramethylrhodamine-dextran is less invasive, and has been used in rat and mouse models of stroke and neoplasia for in vivo cerebral blood flow and edema information [31]. translating this technique to the bbb in severe tbi animal models may be challenging due to the condition of injured tissue and the required imaging window. some investigators have used a radiolabeling strategy to quantify movement of material across the bbb. briefly, the isotope alpha-[ 14 c]-aminioisobutyric acid is a tracer introduced into the circulation along with a red blood cell labeling agent, technetium-99 m . the amount of tracer that has moved out of the vasculature across the bbb can be described by a transfer constant, ki by differentiating the parenchymal tracer concentration from the intravascular tracer, using co-localization with red blood cells labeled with technetium-99 m [32,33]. clinical measurement of bbb permeability bbb permeability can be clinically measured through central nervous system (cns) and peripheral fluid sampling or via imaging. bbb disruption has been defined as a total cerebrospinal fluid (csf) protein concentration to total plasma protein concentration ratio greater than 0.007 [6,34]. proteins such as s100b are released by astrocytes in areas of tissue injury and can be measured via jugular vein or peripherally [35]. the principles of identifying areas of bbb compromise and edema can be translated into neuroimaging. in 2006, lescot et al. used computed tomography to measure volume, weight and specific gravity of contused and non-contused areas in patients with severe tbi [36]. positron emission tomography (pet) can identify the injury penumbra using a metabolic strategy. anaerobic metabolism can be identified as hypodense gray matter and also localized by measuring oxygen extraction fraction, cerebral metabolic rates of oxygen and glucose consumption [37]. magnetic resonance imaging (mri) has emerged as an imaging technique that can be used in both preclinical and clinical studies of bbb permeability following tbi. dynamic contrast enhanced mri uses extravasation of low-molecular weight contrast agents along with repeated t1-weighted imaging to http://imagej.nih.gov/ij g.p. liao et al. admet & dmpk 3(3) (2015) 182-189 186 evaluate the degree of bbb disruption and has been used in human and animal models in a variety of applications from tumors to stroke [38-41]. investigators using mri with diffusion weighted imaging (dwi) technology have been able to identify temporal and regional differences in edema following tbi in rabbits. the authors were able to differentiate between vasogenic and cytotoxic edema based on higher or lower apparent diffusion coefficients respectively [42]. newcombe et al. applied diffusion tensor mri to at-risk contusions in acute post tbi patients and found a pattern of concentric regions of varying diffusion similar to our alexa fluor findings. the authors suggested that the outermost rim of hypodensity may represent a ‘traumatic penumbra’ which may be rescued by effective therapy [43]. bbb studies involving mild tbi due to repeat concussions have been studied using a combination of serum s100, auto-antibodies to s100b and diffusion tensor mri technology focused on white matter damage [44,45]. blast shock wave models of mild tbi using animal models suggest that the primary injury not only causes damage from free radical induced oxidative stress resulting tight junction, pericytes and astrocyte end-feet disruption, but also causes vascular lesions. both of these mechanisms may progress into long term neuroinflammatory damage and chronic traumatic encephalopathy (cte) [46-48]. mri imaging technology has emerged as an excellent, non-invasive tool to follow the evolution of both mild and severe tbi in both pathogenesis and potential therapy. blood pool contrast agents such as the large gadofosveset trisodium has been used in mri as a method to examine bbb permeability noninvasively [49]. originally developed for vascular surgery, this particular contrast agent has strong binding to albumin, long half-life, thus much like evans blue can be used in humans using mri with the advantage of repeated measures and localization of leakage [50]. this particular contrast agent has been used both pre-clinically and clinically to examine bbb breakdown in brain tumors, multiple sclerosis, utilization for traumatic brain injury is emerging [51,52]. smaller agents have been developed, such as gadolinium diethylenetriaminepentaacetate (0.5 kda). within the mri imaging, cellular contrast agents are now being used to show specific cellular infiltration through the bbb. inflammatory cells entering the human brain through the bbb can be labeled with superparamagnetic iron oxide particles (spio) and perfluorocarbons [53]. this technology can help investigators describe the cellular neuroinflammatory response across the bbb, but also help track the response to potential therapies. while mri imaging has been the most common clinical imaging modality for bbb permeability, other techniques are emerging. recently, fluorescence-based time resolved optical detection method for assessment of the bbb has found success in humans. this technique is based on monitoring of the fluorescence following the excitement of indocyanine green dye. following the washout of indocyanine green dye circulating in the brain and may offer a useful adjunct or alternative to mri in describing bbb permeability [54,55]. targets and role of cellular therapy strategies to protect and stabilize the bbb include the blockade of molecules or receptors to inflammatory cytokines as well as the blockade of glutamate, transforming growth factor β, fibrinogen, thrombin, vascular endothelial growth factor, bradykinin, histamine and metalloproteinases. glucose transporter modulation has also been investigated in reducing the co-transport of water into injured tissue [56]. as tbi progresses from the primary to secondary injury, the number of deranged components of the bbb and cns increases along with the number of infiltrating cells and molecules. therefore, no single molecular target or pharmacological agent is likely to adequately rescue the bbb and protect the cns from admet & dmpk 3(3) (2015) 182-189 blood-brain-barrier permeability in trauma research doi: 10.5599/admet.3.3.173 187 further injury. cellular therapy offers a treatment strategy that can potentially sense and respond to multiple signals with multiple neuroprotective and neuroregenerative responses. bbb studies utilizing evans blue as well as immunohistochemistry have previously demonstrated that cell based therapies can modulate the neuroinflammatory response following tbi resulting in bbb preservation [57-59]. of note, pati et al. demonstrated in mouse models of tbi that mesenchymal stem cells exert perivascular protective effects to the penumbral bbb via increased junctional protein (ve-cadherin and occludin) expression [60]. walker et al showed increased localization/organization of occludins to the microvasculature [61]. the cox lab is currently using diffusion tensor mri imaging to evaluate the response to cell therapy in acute setting for severe tbi in pediatric and adult clinical trials (nct01851083 and nct 01575470 respectively). conclusions the bbb continues to be an important therapeutic target in tbi research. preclinical and clinical bbb assays continue to evolve. brain water content and bbb component assays can now be complemented by an array of imaging techniques that quantify bbb related extravasation, function and even infiltration of inflammatory cells. non-invasive mri and fluorescence based optical detection show promise for translational research by providing a platform for conducting preclinical studies that can be unified with clinical outcome measures. references [1] j.l. alves, j. neurosci. res. 92(2) (2014) 141-147. [2] p. barzo, a. marmarou, p. fatouros, k. hayasaki, f. corwin, j. neurosurg. 87(6) (1997) 900-907. [3] s.c.thal, w. neuhaus, arch. med. res. 45(8) (2014) 698-710. [4] a. pasco, l. lemaire, f. franconi, y. lefur, f. noury, j.-p. saint-andre, j.-p. benoit, p.j. cozzone, j.-j. le jeune, j. neurotrauma 24(8) (2007) 1321-1330. [5] a. marmarou, neurosurg. focus 22(5) (2007) e1. [6] m.m. saw, j. chamberlain, m. barr, m.p.g. morgan, j.r. burnett, k.m. ho, neurocrit. care 20(2) (2014) 209-216. [7] k. wang, b. liu, j. ma, j.chin. med. (engl) 127(10) (2014) 1964-1968. [8] p. grammas, j. martinez, b. miller, expert. rev. mol. med. 13 (2011) e19. [9] d. vajtr, o. benanda, j. kukačka, r. pruša, l. houstava, p. toupalik, r. kizek, physiol. res. 58(2) (2009) 263-268. [10] j.a. stokum, d.b. kurland, v. gerzanich, j. marc simard, neurochem. res. 40(2) (2015) 317-328. [11] k. taya, c.r. marmarou, k. okuno, r. prieto, a. marmarou, j. neurotrauma 27(1) (2010) 229-239. [12] a.r. frattalone, g.s. ling, neurosurg. clin. n. am. 24(3) (2013) 309-319. [13] d. lozano, g.s. gonzales-portillo, s. acosta, i. de la pena, n. tajiri, y. kaneko, c.v. borlongan, neuropsychiatr. dis. treat. 11 (2015) 97-106. [14] j. lu, s.j. goh, p.y.l. tng, y.y. deng, e.-a. ling, s. moochhala, front. biosci. (landmark ed) 14 (2009) 3795-3813. [15] w.w. van den brink, a. marmarou, c.j. avezaat, acta neurochir. suppl. (wien) 51 (1990) 261-262. [16] s. khosrow tayebati, f. amenta, d. tomassoni, cns neuro . disord. drug targets ( 2015). [17] j. wen, s. qian, q. yang, l. deng, y. mo, y. yu, exp. ther. med. 8(3) (2014) 881-886. [18] m. imer, b. omay, a. uzunkol, t. erdem, p.a. sabanci, a. karasu, s.b. albayrak, a. sencer, k. hepgul, m. kaya, neurol. res. 31(9) (2009) 977-981. g.p. liao et al. admet & dmpk 3(3) (2015) 182-189 188 [19] m. khan, y.b. im, a. shunmugavel, a.g. gilg, r.k. dhindsa, a.k. singh, i. singh, j. neuroinflammation 6:32 (2009). [20] m. nuriya, t. shinotsuka, m. yasui, cereb. cortex 23(9) (2013) 2118-2126. [21] j. schindelin, i. arganda-carreras, e. frise, v. kaynig, m. longair, t. pietzsch, s. preibisch, c. rueden, s. saalfeld, b. schmid, j.-y. tinevez, d.j. white, v. hartenstein, k. eliceiri, p. tomancak, a. cardona, nat. methods 9(7) (2012) 676-682. [22] g.p. liao, s.d. olson, d.j. kota, r.a. hetz, p. smith, s. bedi, c.s. cox jr, j. surg. res. 190(2) (2014) 628633. [23] n.h. on, p. kiptoo, t.j. siahaan, d.w. miller, mol. pharm. 11(3) (2014) 974-981. [24] n.h. on, s. savant, m. toews, d.w. miller, j. cereb. blood flow metab. 33(12) (2013) 1944-1954. [25] n.h. on, f. chen, m. hinton, d.w. miller, pharm. res. 28(10) (2011) 2505-2515. [26] m. krueger, i. bechmann, k. immig, a. reichenbach, w. härtig, d. michalski, j. cereb. blood flow metab. 35(2) (2015) 292-303. [27] b.t. hawkins, y.h. gu, y. izawa, g.j. del zoppo, j. cereb. blood flow metab. 35 (2015) 985-992. [28] f.y. jalal, y. yang, j.f. thompson, t. roitbak, g.a. rosenberg, j. cereb. blood flow metab. 35 (2015) 1145-1153. [29] j.h. kim, d.w. lee, b.y. choi, m. sohn, s.h. lee, h.c. choi, h.k. song, s.w. suh, brain res. 1595 (2015) 156-165. [30] j. badaut, d.o. ajao, d.w. sorensen, a.m. fukuda, l. pellerin, neuroscience 285 (2015) 215-226. [31] s.m. schwarzmaier, m. gallozzi, n. plesnila, j. neurotrauma 32(13) (2015) 990-1000. [32] j. zhang, g.b. sadowska, x. chen, s.y. park, j.-e. kim, c.a. bodge, e. cummings, y.-p. lim, o. makeyev, w.g. besio, j. gaitanis, w.a. banks, b.s. stonestreet, faseb j 29 (2015) 1739-1753. [33] x. chen, g.b. sadowska, j. zhang, j.-e. kim, e.e. cummings, c.a. bodge, y.-p. lim, o. makeyev, w.g. besio, j. gaitanis, s.w. threlkeld, w.a. banks, b.s. stonestreet, neurobiol. dis. 73 (2015) 118-129. [34] k.m. ho, s. honeybul, c.b. yip, b.i. silbert, j. neurosurg. 121(3) (2014) 674-679. [35] b. arfvidsson, t.k. nilsson, l. norgren, clin. chem. lab. med., 53(1) (2015) 111-117. [36] t. lescot, v. degos, a. zouaoui, crit. care med. 34(12) (2006) 3029-3033. [37] h.m. wu, s.c. huang, p. vespa, d.a. hovda, m. bergsneider, j. neurotrauma 30(5) (2013) 352-360. [38] a. montagne, s.r. barnes, m.d. sweeney, m.r. halliday, a.p. sagare, z.zhao, a.w. toga, r.e. jacobs, c.y. liu, l. amezcua, m.g. harrington, h.c. chui, m. law, b.v. zlokovic, neuron 85(2) (2015) 296-302. [39] w. li, j.a. long, l.t. watts, z. jiang, q. shen, y. li, t.q. duong, plos one 9(12) (2014) e114173. [40] n.j. heyer, j.h. derzon, l. winges, c. shaw, d. mass, s.r. snyder, p. epner, j.h. nichols, j.a. gayken, d. ernst, e.b. liebow, clin. biochem. 45(13-14) (2012) 1012-1032. [41] z. merali, j. leung, d. mikulis, f. silver, a. kassner , transl. stroke. res. 6(1) (2015) 39-49. [42] x.e. wei, y.z. zhang, y.h. li, m.h. li, w.b. li, j. neurotrauma. 29(14) (2012) 2413-2420. [43] v.f. newcombe, g.b. williams, j.g. outtrim, d. chatfield, m. gulia abate, t. geeraerts, a. manktelow, h. room, l. mariappen, p.j. hutchinson, j.p. coles, d.k. menon, j. cereb. blood flow metab. 33(6) (2013) 855-862. [44] n. marchi, j.j. bazarian, v. puvenna, m. janigro, c. ghosh, j. zhong, t. zhu, e. blackman, d. stewart, j. ellis, r. butler, d. janigro, plos one 8(3) (2013) e56805. [45] e. bargerstock, v. puvenna, p. iffland, t. falcone, m. hossain, s. vetter, s. man, l. dickstein, n. marchi, c. ghosh, j. carvalho-tavares, d. janigro, plos one 9(7) (2014) e101477. [46] j.e. bailes, a.l. petraglia, b.i. omalu, e. nauman, t. talavage, j. neurosurg. 119(5) (2013) 1235-1245. [47] a.k. shetty, v. mishra, m. kodali, b. hattiangady, front. cell neurosci., 8 (2014) 232. [48] h. ling, j. hardy, h. zetterberg, mol. cell neurosci. 66(pt b) (2015) 114-122. [49] f.l. giesel, a. mehndiratta, m. essig, eur. radiol. 20(10) (2010) 2461-2474. [50] m. essig, k. nikolaou, j.f. meaney, eur. radiol. 17 suppl 2 (2007) b30-37. admet & dmpk 3(3) (2015) 182-189 blood-brain-barrier permeability in trauma research doi: 10.5599/admet.3.3.173 189 [51] j. puig, g. blasco, m. essig, j. daunis-i-estadella, g. laguillo, a.m. quiles, s. remollo, k. bergmann, c. joly, l. bernado, j. sánchez-gonzález, s. pedraza, eur. radiol. 23(4) (2013) 1093-1101. [52] e. fisher, d.s. reich, neurology 81(3) (2013) 202-203. [53] g. weise, g. stoll, front. neurol. 3 (2012) 178. [54] a. liebert, et al., conf. proc. ieee eng. med. biol. soc. 2013 (2013) 3040-3042. [55] n. horie, y. fukuda, t. izumo, k. hayashi, k. suyama, i. nagata, acta neurochir. (wien) 156(5) (2014) 919-926. [56] d. shlosberg, m. benifla, d. kaufer, a. friedman, nat. rev. neurol. 6(7) (2010) 393-403. [57] p. a. walker, s.k. shah, f. jimenez, k.r. aroom, m.t. harting, c.s. cox jr., surgery 152(5) (2012) 790793. [58] p.a. walker, s.s. bedi, s.k. shah, f. jimenez, h. xue, j.a. hamilton, p. smith, c.p. thomas, r.w. mays, s. pati, c.s. cox, j. neuroinflammation 9: 228 (2012). [59] s. pati, a.y. khakoo, j. zhao, f. jimenez, m.h. gerber, m. harting, j.b. redell, r. grill, y. matsuo, s. guha, c.s. cox jr., m.s. reitz jr., j.b. holcomb, p.k. dash, stem cells dev. 20(1) (2011) 89-101. [60] t. menge, y. zhao, j. zhao, k. wataha, m. gerber, j. zhang, p. letourneau, j. redell, l. shen, j. wang, z. peng, h. xue, r. kozar, c.s. cox jr., a.y. khakoo, j.b. holcomb, p.k. dash, s. pati, sci. transl. med. 4(161) (2012) 161ra150. [61] p.a. walker, s.k. shah, f. jimenez, m.h. gerber, h. xue, r. cutrone, j.a. hamilton, r.w. mays, r. deans, s. pati, p.k. dash, c.s. cox jr., exp. neurol. 225(2) (2010) 341-52. ©2015 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ admet & dmpk x(y) (20xx) pp-pp; doi: 10 doi: 10.5599/admet.1.4.20 82 admet & dmpk 1(4) (2013) 82-82; doi: 10.5599/admet.1.4.20 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index editorial board member: brief biography kin tam dr. kin tam holds a dphil in physical chemistry from the university of oxford, uk. he is currently associate professor in biomedical sciences of the faculty of health sciences, university of macau, china. before moving back to china in 2013, he worked for astrazeneca, alderley park, uk, for 12 years, progressing from senior scientist, to associate principal scientist, and working as a dmpk design leader from 2011. he contributed to a number of drug discovery projects in the areas of oncology and led the development of high-throughput physicochemical assays. prior to joining astrazeneca, he developed new instruments with sirius analytical instruments, uk, for 5 years and commercialised uv-based devices for pka determination. his current research interests include permeability and physiochemical studies of drug molecules, drug metabolism and pharmacokinetics as well as drug development. he has authored or co-authored over 70 articles in peer review journals. dr. tam is a fellow of the royal society of chemistry. he is currently the editor-in-chief for admet & dmpk. ©2013 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution licence (http://creativecommons.org/licenses/by/3.0/) http://www.pub.iapchem.org/ojs/index.php/admet/index http://creativecommons.org/licenses/by/3.0/ manuscript doi: 10.5599/admet.2.4.150 191 admet & dmpk 2(4) (2014) 191-198; doi: 10.5599/admet.2.4.150 open access: issn: 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index review skin pampa: application in practice bálint sinkó 1, gábor vizserálek 2 and krisztina takács-novák 2 1 sinkolab scientific, 21 nagyszőlős street, h-1113 budapest, hungary 2 department of pharmaceutical chemistry, semmelweis university, 9. hőgyes e. street, h-1092 budapest, hungary  corresponding author: e-mail: balint@sinkolab.com; tel.: +36-1789-2896; received: december 01, 2014; revised: december 17, 2014; published: january 09, 2015 abstract transdermal drug delivery has been growing extensively in the past decades, therefore new, reliable and cost-effective in vitro models were demanded to support the research and development on this field. model membrane of pampa, which can mimic skin penetration, was first described in 2006, but the need for a more bio-mimetic system has been arisen by new industrial tendencies and a bio-relevant system was published in 2012. since its first publication, skin pampa has already been applied by several universities and industrial groups successfully, and the first articles, podium and poster presentations have been appeared. the original skin pampa model has been further developed in order to extend its application for formulations. examples of liquid or semi-solid formulation projects and transdermal patch studies are available beside standard solution applications. the present review demonstrates the different approaches needed for various type of samples, provides examples of applications and practical conclusions for further improvement. keywords skin penetration; in vitro; artificial skin model, formulation study introduction estimation of skin penetration of compounds has become crucial in the past decades, as the advantages of dermal usage and transdermal absorption route have been recognized by many pharmaceutical and cosmetic companies [1]. it is common that the best estimation is provided by in vivo human skin studies, however its availability and the ethical regulation make it impossible to apply routinely in product development. in vitro human skin experiments could be the most useful replacement to predict human in vivo environment, but accessibility is difficult, especially for cosmetic industry. therefore researchers started to work on development of artificial membrane models that can mimic the features of skin, while having much better availability and being more cost effective. depending on the need, silicon based membranes, cellulose acetate or cellulose nitrate filters are available for franz diffusion cell method; however these models have limited quality of prediction, suffer from poor standardization and do not even meet the high throughput criteria, that would be useful in early stages of discovery and development [2]. the current review introduces the pampa method and the skin pampa model, that provides a good estimation of skin penetration, does not suffer from pure standardization and reproducibility, while having a high throughput and low cost option [3]. http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:balint@sinkolab.com sinko et al. admet & dmpk 2(4) (2014) 191-198 192 with the introduction of the biopharmaceutics classification system [4], fast and reliable permeability measurement strategies were needed to classify the molecules at the earliest stages of discovery. in 1998, kansy and co-workers have introduced the parallel artificial membrane permeability assay (pampa) as a tool for rapid determination of passive membrane permeability of drugs [5]. the interest of the industry was growing extensively in the first decade of the method, mostly because of the method’s low cost and high throughput performance. since the first publication by kansy and co-workers, several companies have developed their own variants and many articles were published, indicating that pampa is a good research tool for physico-chemical property screening in early discovery stage [6]. pampa models have been published for the prediction of gastrointestinal absorption (git) [7,8] and for modelling the blood brain barrier (bbb) [9]. there are two models available for the estimation of skin penetration as well. the first pampa model for skin penetration estimation was published by ottaviani et al. in 2006 [10]. this approach incorporates silicone oil and isopropyl myristate as membrane components, which are not the natural elements of the in vivo barrier. also, that model was developed based on the flynn database that contains human skin permeability results collected from many research groups using diverse experimental conditions. as it was shown in several studies [11-13], the difference between permeability data measured in various laboratories can exceed by an order of magnitude, therefore the prediction potential of this model is not proved yet. following the tendencies of model development in pharmaceutical industry, skin pampa was designed to be bio-mimetic, i.e. the same or similar components have been applied as present in the most important barrier of human skin [3]. as the main barrier of human skin, it is known to be the outermost layer of epidermis, the stratum corneum (sc), skin pampa has been designed to mimic the features of this layer. sc composes of corneocytes embedded into a multilamellar lipid layer. lipid layer consists of a mixture of ceramides, cholesterol and free fatty acids as major components, and provides the main route of penetration, the paracellular pathway [1]. skin pampa membrane was created by using cholesterol, free fatty acid and a ceramide-analog compound that mimics the features of ceramides in the lipid matrix [14]. the ceramide-analogue compound (certramide) has been studied extensively and its properties have been found to be suitable for pampa membrane [15]. during the 15 years history of pampa, researchers have used the system to predict the permeability of compounds through the git, bbb or later through skin. in most cases, the main goal was to determine permeability as kinetic parameter of pure compounds in early stage of drug discovery without any formulation effect. most of the protocols was developed to satisfy this need, therefore buffer solutions at various ph with 0.5-1(-5) % dmso was studied as donor solution and pure ph = 7.4 buffer as acceptor solution [16]. donor concentration was selected according to the physico-chemical (solubility, pka, logp) parameters of the compounds and permeability was calculated as the main parameter for comparison. upon the expanding application of skin pampa, an urgent need has been arisen for the investigation of drug formulations, including semi-solid dosage forms and transdermal patches. the following paragraphs describe the application possibilities of skin pampa, provide some general considerations of assay design and data processing, and demonstrate some recent examples. admet & dmpk 2(4) (2014) 191-198 skin pampa doi: 10.5599/admet.2.4.150 193 applications permeability studies of compounds in solution samples of compounds dissolved in pure solvent or solvent mixture are typically applied in skin pampa, when the permeation properties of active pharmaceutical ingredients (api) are investigated. in these studies, the performance of the api itself is important, without any penetration enhancing or controlling effect. simple solvents, like water based buffers, peg 400, ethanol or propylene-glycol are applied in these projects that have a well-defined and usually predictable effect on permeability. the protocol of these studies is the same or similar to general pampa protocols, i.e. the concentration and the applied dose are selected based on the physico-chemical parameters of the api and not based on the relevant in vivo dose. the donor volume/membrane area ratio is also significantly higher compared to in vivo environment. permeability is determined and used for assay evaluation, that is normalized for donor concentration, for surface area and for time, therefore it is capable to rank the compounds own permeation property. the performance of skin pampa for api characterisation was first investigated by the inventors of the model. two correlations were described and both were established as a result of careful human permeability data selection. selection of appropriate, reliable reference dataset has fundamental importance in model validation, since several published databases contain results of different laboratories that can reduce the significance of the correlation. the first selected dataset was published by lee and co-workers in 2010 [17]. this dataset provided homogenous and normalized permeability data for more than 40 compounds. the results were measured on dermatomed human female back skin using franz diffusion cell method. 45 % peg 400 was added to donor solution to mimic the effect of a basic formulation and ph was selected to be 6.4. the skin pampa experiment for 15 compounds was performed with similar parameters to franz cell experiment. permeability was calculated and used for the comparison as the performance of the apis. high correlation (r 2 = 0.84) was found between human skin data and skin pampa data as shown on figure 1. the good correlation also indicates good ranking order of compounds, so the model was able to characterize the permeability and rank the apis properly. the authors have also compared skin pampa results measured with pure buffer solutions in the donor phase to the same dataset, and found a significantly weaker correlation that indicates the responsiveness of the system for the simple formulation effects of peg 400. the second database was the full validated dataset published by hadgraft and guy [18]. this database contains results measured in different laboratories and despite the careful validation criteria of selection the experimental conditions are very different. therefore data were divided into small homogenous groups. data were sorted based on the incubation temperature and the skin derivative used for the test. skin pampa experiment was performed with the standard solution protocol, pure buffers solutions with 1 % dmso were studied as donor phase, pure buffer at ph = 7.4 was used as acceptor, and permeability was calculated for the correlation study. four comparisons were published and good correlation was found, however statistical evaluation was not possible for the reason of low number of compounds. a united correlation study performed on the published 4 small groups is shown on figure 2. the obtained r 2 = 0.70 value is also an acceptable correlation regardless of the uncertainty caused by the human data. sinko et al. admet & dmpk 2(4) (2014) 191-198 194 figure 1. comparison of human skin permeability data measured by franz cell vs skin pampa data. figure 2. correlation between human skin permeability data of full validated dataset and skin pampa data. karadzovska and riviere published a study of 96 well based skin penetration models in 2013 that included skin pampa as well, beside the other, previously mentioned pampa skin model by ottaviani and strat-m membrane of merck [19]. six compounds dissolved in three different solvents (water, ethanol and propylene-glycol) were studied. compounds were applied in both saturated and unsaturated solutions, but both applications can be considered as infinite for the reason of the total applied dose. permeability was calculated and used for the correlation study. the assays with saturated solutions were not successful, but the results of unsaturated solutions were in good agreement with data measured on human skin. they investigated the membrane retention of 96 well plate based models, and found good correlation to skin retention as well, therefore, all three models were found to be useful for transdermal penetration prediction. skin pampa has been acknowledged as the most bio-mimetic membrane model, where specific skin mimetic interactions are more probable. in a paper published in 2013, vizserálek and co-workers investigated the effect of incubation temperature on permeability using skin pampa model among others [20]. the study concluded that skin pampa results are more sensitive for incubation temperature compared to other, mostly phospholipidbased pampa models. figure 3 demonstrates the temperature dependence of 7 compounds. as the slope of the lines indicates, as high as 0.35 log unit difference can occur by performing the assay either at 25 or 32 °c. therefore, the authors concluded that proper incubation temperature is important for skin pampa admet & dmpk 2(4) (2014) 191-198 skin pampa doi: 10.5599/admet.2.4.150 195 assay, especially when formulations are investigated where the sample itself can have temperature dependent behavior. figure 3. temperature dependence of permeability in skin pampa method. permeability studies with liquid or semi-solid formulations formulations require different approach compared to standard pampa application, as the aim of the study is to investigate the performance of the formulation, and not just the pure api. the modification affects the applied dose, the way of application and the calculated parameters as well. as the goal is to investigate the performance of the formulation, keeping the applied dose close to in vivo environment would provide more reliable information. the previously described general pampa protocol or skin pampa protocol for solutions apply “infinite dosing”, that means significantly higher dose compared to in vivo case. therefore, the dose should be reduced as much as possible to approach the “finite dosing”, i.e. the in vivo relevant amount [21]. the currently available protocols and plates make it possible to apply 10 times finite dose, that is much closer to finite situation than standard pampa assay, and the development of lower dose applications are still in process. new bottom plate (formulation plate) has also been developed by the manufacturer to make the ‘10 times finite dose’ application possible. as the performance of the formulation as a complex matrix is investigated, parameters like flux, concentration in the receiver solution or permeated amount of drug are calculated, that are not normalized for the applied dose. these parameters are suitable for direct comparison of the formulations. it is advantageous if the permeation is studied in wider time period, and the permeated amount vs time profile is used for evaluation. this graph includes all the main information like permeated amount at certain time point, flux and lag time. as these considerations are the result of the latest developments, most of the cited studies below have not been designed according to all of these recommendations. on a poster of clough and co-workers published at aaps in 2013, the permeability enhancing effect of a series of solvents, lipidic fluids and emulsifiers were investigated [22]. infinite dosing approach was followed by the authors and standard pampa sandwiches were used for the study. all samples contained 5 % ibuprofen-na as api and 8 % of one of the investigated component beside water. the samples of lipidic fluids contained 45 % peg 400 as well. figure 4 demonstrates the receiver concentration after one hour of incubation where the authors have found significant differences in the permeability of ibuprofen depending on the penetration enhancer. unfortunately, correlation with human skin data is not available, but the sinko et al. admet & dmpk 2(4) (2014) 191-198 196 results itself suggest that the model is able to differentiate the samples, and therefore can be used for ranking the formulations. figure 4. permeability enhancing effect of different solvents, lipidic fluids and emulsifiers in case of ibuprofen measured by skin pampa method [22]. tsinman and co-workers presented their work with semi-solid formulation of ibuprofen at aaps in 2012 [23]. three formulations and one slurry of ibuprofen were investigated using both skin pampa model and real human skin mounted in franz cell system. they have applied 30 µg formulation/well that is equal to ’10 times finite dose application’ as mentioned before, and they used the receiver concentration for the comparison. good agreement was found between franz cell and skin pampa results in terms of ranking order the samples, however skin pampa model provided higher permeation results compared to human skin data. in 2014 at aaps conference, luo and co-workers presented a comparison of skin pampa results to porcine skin results measured with franz cell method [24]. the study investigated the permeation of 3 apis (ibuprofen, caffeine, uvinul a plus) using saturated solutions. they presented both infinite and ‘10 time finite’ dosing studies and used the cumulative permeated amount vs time profiles for data interpretation and analysis. the comparison shows, that skin pampa can provide an excellent ranking order of samples, while having the advantage of much lower cost and ease of standardization. permeability studies with transdermal patches similarly to formulation studies, the aim of transdermal patch testing is to evaluate the performance of the patch as transdermal therapeutic system and not just the api, therefore similar approach is needed. the dosing is easier in this case as the patch is applied directly on the membrane without any need for normal or modified bottom plate, so the applied dose is predefined by the patch. as a significant amount of available patches must not be cut into pieces, it is recommended to apply them as they are. this application involves the occurrence of “edge effect” [25], which means a possible lateral diffusion of api within the adhesive layer that can increase the flux. it is recommended to determine the cumulative permeated amount vs time profile that can provide all the necessary information for comparison as detailed before. as most of the transdermal patches are designed to provide a zero order kinetic release of api, it is suggested to continue the skin pampa study till the end of the proposed application period of the patch. the first and only study of transdermal therapeutic systems has been presented by vizserálek and coworkers at aaps 2014 [26]. transdermal patches of four apis (nicotine, fentanyl, ketoprofen, rivastigmine) were applied on skin pampa model. figure 5 demonstrates the permeated amount vs time profile for admet & dmpk 2(4) (2014) 191-198 skin pampa doi: 10.5599/admet.2.4.150 197 fentanyl as an example. the blue dashed line shows the in vivo delivery rate calculated from human blood level, while the red dots and black solid line demonstrate the skin pampa results. excellent agreement can be seen between skin pampa results and in vivo human flux data, therefore the study concluded that skin pampa is a good tool for ranking order of transdermal patches, and it can even estimate the human absorption rate with an acceptable precision. the detailed results of this study on transdermal patches will be published soon elsewhere. figure 5. the permeated amount vs time profile of a marketed transdermal patch containing fentanyl. conclusions in the past 2.5 years, skin pampa as a bio-relevant artificial membrane based ht permeability model has been widely applied for the estimation of skin penetration. this paper has collected the most important publications, and grouped them based on the type of sample applied on the membrane. the driving forces of the studies, the ways of application and the criteria of calculated parameter selection have been considered and suggestions for assay planning have been provided. it can be concluded, that the careful definition of study expectation, the proper design of the assay and the calculation of suitable parameters are crucial for satisfying results and for good prediction potential. references [1] j.a. bouwstra, p.l. honeywell-nguyen, g.s. gooris, m. ponecet, progress in lipid research 42(1) (2003) 1-36. [2] d. howes, r. guy, j. hadgraft, j. heylings, u. hoeck, f. kemper, h. maibach, j.p. marty, h. merk, j. parra, d. rekkas, i. rondelli, h. schaefer, u. täuber, n. verbiese, atla: alternatives to laboratory animals 24(1) (1996) 81-106. [3] b. sinko, t.m. garrigues, g.t. balogh, z.k. nagy, o. tsinman, a. avdeef, k. takács-novák, european journal of pharmaceutical sciences 45(5) (2012) 698-707. [4] g.l. amidon, h. lennernäs, v.p. shah, j.r. crison, pharmaceutical research 12(3) (1995) 413-420. [5] m. kansy, f. senner, k. gubernator, journal of medicinal chemistry 41(7) (1998) 1007-1010. [6] a. avdeef, expert opinion on drug metabolism and toxicology 1(2) (2005) 325-342. [7] a. avdeef, absorption and drug development: solubility, permeability, and charge state, 2nd edition. wiley, hoboken, new jersey, united states, 2012. [8] a. avdeef, o. tsinman, european journal of pharmaceutical sciences 28(1-2) (2006) 43-50. [9] o. tsinman, k. tsinman, n. sun, a. avdeef, pharmaceutical research 28(2) (2011) 337-363. [10] g. ottaviani, s. martel, p.a. carrupt, journal of medicinal chemistry 49(13) (2006) 3948-3954. sinko et al. admet & dmpk 2(4) (2014) 191-198 198 [11] g.m. khan, y. frum, o. sarheed, g.m. eccleston, v.m. meidan, international journal of pharmaceutics 303(1-2) (2005) 81-87. [12] r.p. chilcott, n. barai, a.e. beezer, s.i. brain, m.b. brown, a.l. bunge, s.e. burgess, s. cross, c.h. dalton, m. dias, a. farinha, b.c. finnin, s.j. gallagher, d.m. green, h. gunt, r.l. gwyther, c.m. heard, c.a. jarvis, f. kamiyama, g.b. kasting, e.e. ley, s.t. lim, g.s. mcnaughton, a. morris, m.h. nazemi, m.a. pellett, j. du plessis, y.s. quan, s.l. raghavan, m. roberts, w. romonchuk, c.s. roper, d. schenk, l. simonsen, a. simpson, b.d. traversa, l. trottet, a. watkinson, s.c. wilkinson, f.m. williams, a. yamamoto, j. hadgraft, journal of pharmaceutical sciences 94(3) (2005) 632-638. [13] y. frum, g.m. eccleston, v.m. meidan, european journal of pharmaceutics and biopharmaceutics 67(2) (2007) 434-439. [14] b. sinko, m. pálfi, s. béni, j. kökösi, k. takács-novák, molecules 15(2) (2010) 824-833. [15] b. sinko, j. kökösi, a. avdeef, k. takács-novák, chemistry & biodiversity 6(11) (2009) 1867-1874. [16] j.m. reis, b. sinko, c.h.r. serra, mini-reviews in medicinal chemistry 10(11) (2010) 1071-1076. [17] p.h. lee, r. conradi, v. shanmugasundaram, bioorganic & medicinal chemistry letters 20(1) (2010) 69-73. [18] r. guy, j. hadgraft, transdermal drug delivery, 2nd edition. dekker, new york, united states 2003. [19] d. karadzovska, j.e. riviere, european journal of pharmaceutical sciences 50(5) (2013) 569-576. [20] g. vizseralek, t. balogh, k. takács-novák, b. sinkó, european journal of pharmaceutical sciences 53 (2014) 45-49. [21] d. selzer, m.m.a. abdel-mottaleb, t. hahn, u.f. schaefer, d. neumann, advanced drug delivery reviews 65(2) (2013) 278-294. [22] m. clough, n. richardson, n. langley, k. tsinman, o. tsinman, assessment of transdermal penetration enhancement by topical pharmaceutical excipients using skin pampa method . (t2267) in aaps annual meeting and exposition. 2013. san antonio. [23] k. tsinman, o. tsinman, g. schalau, h. aliyar, r. huber, g. loubert, application of skin pampa to differentiate between topical pharmaceutical formulations of ibuprofen. (r6058) in aaps annual meeting and exposition. 2012. chicago. [24] l. luo, b. sinkó, k. tsinman, h. abdalghafor, j.hadgraft, m. lane, a comparison of drug permeation in the skin pampa model and the franz cell model. (w5104) in aaps annual meeting and exposition. 2014. san diego. [25] j. hadgraft, d. lewis, d. beutner, h.m. wolff, international journal of pharmaceutics 73(2) (1991) 125-130. [26] g. vizseralek, b. sinkó, k. tsinman, k. takács-novák, developing a method for skin pampa to test transdermal patches. (m1237) in aaps annual meeting and exposition. 2014. san diego. ©2014 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ manuscript doi: 10.5599/admet.3.3.191 287 admet & dmpk 3(3) (2015) 287-295; doi: 10.5599/admet.3.3.191 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper prion protein and its interactions with metal ions (cu 2+ , zn 2+ , and cd 2+ ) and metallothionein 3 branislav ruttkay-nedecky 1,2 , eliska sedlackova 1 , dagmar chudobova 1,2 , kristyna cihalova 1,2 , ana maria jimenez jimenez 1,2 , sona krizkova 1,2 , lukas richtera 1,2 , vojtech adam 1,2 and rene kizek 1,2* 1 department of chemistry and biochemistry, mendel university in brno, zemedelska 1, cz-613 00 brno, czech republic, european union 2 central european institute of technology, brno university of technology, technicka 3058/10, cz-616 00 brno, czech republic, european union *corresponding author: e-mail: kizek@sci.miuni.cz; tel.: +420-5-4513-3350; fax: +420-5-4521-2044 received: june 10, 2015; revised: august 02, 2015; published: september 05, 2015 abstract the effects of heavy metals (zn 2+ , cu 2+ , and/or cd 2+ ) on escherichia coli expressing either prion (hprp c ) or metallothionein 3 (mt-3) brain proteins capable of binding these metals were investigated. the expression of hprp c or mt-3 in e.coli was confirmed using western-blot and dot-blot methods. after analyzing growth curves, we found that bacteria expressing prion protein better tolerated the presence of zn 2+ in comparison with wild-type bacteria and bacteria expressing mt-3. the addition of cd 2+ and cu 2+ was well tolerated by bacteria expressing mt-3, whereas the bacteria expressing prion protein displayed slower growth when compared to the wild-type. we subsequently determined total content of the mt in bacteria using differential pulsed voltammetry (dpv), and depending on the treatment of the individual metals. mt expression in mt3 transformed cells as well as in control e.coli cells increased at the lowest metal concentration (25 µm), followed by a decrease at higher metal concentrations (50, 75, and 150 µm). the highest increase by cd 2+ were observed. mt expression pattern in hprp c transformed cells was different. after application of cu 2+ an increase in mt expression continued also at higher metal concentrations; and after application of cd 2+ and zn 2+ no decrease in mt expression at higher metal concentrations was observed. keywords prion diseases; prp c ; prpsc; mt; mt-3; copper; zinc; cadmium. introduction numerous studies suggest that prion disease is caused by conformational conversion of prp c (the normal cellular prion protein) to prp sc (its abnormal isoform), which becomes infectious [1]. in 1982, stanley b. prusiner formulated prion theory, recognizing the substance of which was the definition of the prion protein infectious particle that lacks nucleic acid [2]. prp is encoded by a single copy of chromosomal gene [3]. the gene encoding prp (prnp) is a copy of the gene that is located on the 20th chromosome in humans [4]. prion protein (prp c ) is a protein naturally present in all mammalian cells. cellular prion protein (prp c , fig. 1a) is a glycoprotein that is located primarily in the central nervous system (cns), but also occurs http://www.pub.iapchem.org/ojs/index.php/admet/index ruttkay-nedecky et al. admet & dmpk 3(3) (2015) 287-295 288 in smaller quantities in other tissues. the physiological function of this protein is not yet fully understood, but it is probably involved in cell differentiation and synaptic transmission [5]. posttranslational modifications of prp c and conformational changes to pathological isoform prp sc are the molecular basis of prion diseases [6]. in comparison with the knowledge about prp c , very little is known about prp sc . the pathological isoform (prp sc , fig. 1b) is highly resistant to degradation processes in the cells and, upon binding to prp c proteins these isoforms will change the original conformation to the conformation of prp sc . the presence of hardly degradable protein and lack of physiological prp c therefore will lead to the emergence and development of progressive neurodegenerative diseases known as transmissible spongiform encephalopathy (tse), which are clinically heterogeneous and currently incurable [7-9]. these illnesses in humans and animals include the following diseases: creutzfeldt-jakob disease (cjd) or gertsman-straussler-schenker syndrome in humans, scrapie in both sheep and goats, and bovine spongiform encephalopathy (bse) in cows. [10-12]. experiments with tissue culture of cerebellum showed that cells lacking prp c are more susceptible to oxidative stress and readily undergo cell death [13]. some of advances in prion protein detection can be found in [14-19]. figure 1. (a) scheme of the structure of cellular form of prion protein (prpc). (b) the structure of pathological (abnormal) prion protein (prpsc). adapted from [2]. because the prp c is a metal binding protein, there is a hypothesis that the metal ions (cu 2+ , zn 2+ and mn 2+ ) can play a role in the transformation of prp c to prp sc . equally important link between prion diseases and metals is a protein called metallothionein. the brain specific form of metallothionein-3 (mt-3) ensures the maintenance of optimal concentrations of metals in the brain. also the involvement of mt-3 in the formation of neurodegenerative diseases was observed as the decrease in the mt-3 levels led to the formation of neurofibrillary clusters characteristic for neurodegenerative diseases [20-22].the aim of this study was to monitor heavy metal ions influence on the growth and electrochemical properties of e. coli strains transformed with prp c or mt-3 protein. experimental chemicals, preparation of deionized water and ph measurement chemicals used in this study were purchased from sigma-aldrich (st. louis, mo, usa) in american chemical society (acs) purity unless noted otherwise. the deionized water was prepared using reverse a b admet & dmpk 3(3) (2015) 287-295 prion interactions with metals and metallothionein doi: 10.5599/admet.3.3.191 289 osmosis equipment aqual 25 (aqual s.r.o., brno, czech republic). the deionized water was further purified at 18 mω resistance using miliq direct quv equipped with uv lamp. the ph was measured using ph meter wtw inolab (weilheim, germany). chemical transformation, growth conditions and isolation of protein fraction the human mt-3 protein (hmt-3) and human prion protein (hprp c ) were cloned in the prset-b vector (invitrogen, usa). high levels of expression of dna sequences cloned into the prset vectors were made possible by the presence of the t7 promoter. the chemical transformation protocol was performed following the instructions of new england biolabs, using as host bl21 (de3) plyss chemically competent e. coli strain. bacteria transformed with prset-hmt-3 or prset-hprp c plasmid were selected by ampicillin resistance. the positive transformants were confirmed by pcr screening. the positive transformants were grown in lb (luria bertani) medium (10 g tryptone, 5 g yeast extract, 5 g nacl) with 50 µg ml -1 ampicillin and 35 µg ml -1 chloramphenicol shaking at 37 °c overnight. the next day, each culture was grown to 0.1 of od600nm. when the culture reached the exponential phase (0.4 0.6), iptg (isopropylthio-b-d-galactopyranoside) was added to a final concentration of 1 mm and the cells were allowed to grow for 4 to 6 hours. after centrifugation at 4000 rpm for 10 min, the pellet was resuspended, with 20 mm phosphate buffer at neutral ph and frozen in liquid nitrogen. the frozen lysate was thawed at 42 °c (this freeze-thaw action was repeated three times). the protein fraction was harvested by centrifugation at 4000 rpm for 10 minutes at 4 °c. measurement of growth curves after addition of selected heavy metal ions (cu 2+ , zn 2+ and cd 2+ ) to the bacterial cultures procedure for the evaluation of the antimicrobial effect of tested compounds and their combinations consisted of measuring the absorbance using the apparatus multiscan ex (thermo fisher scientific, germany) and subsequent analysis in the form of growth curves. control bacterial cultures of e. coli and e. coli with hmt-3 or hprp c were cultivated in lb medium with 50 µg ml -1 ampicillin, 35 µg ml -1 chloramphenicol and 1 mm iptg for 24 hours with shaking. in the microplate these cultures were mixed with cadmium, zinc and copper ions (10, 50, 150 and 300 µm concentrations) or strains without addition of heavy metal ions alone as a control for measurements. total volume in the microplate wells was always 300 µl. measurements were carried out at time 0, then each half-hourly for 24 hours at 37 °c and a wavelength of 620 nm. preparation of samples each sample (control strains – e. coli, e. coli – mt-3, e. coli hprp c or strains with addition of 10, 25, 50, 75, 150 and 300 µm concentrations of cu 2+ , zn 2+ and cd 2+ ions) was centrifuged at 8000 rpm for 10 minutes. liquid nitrogen was added to the pellet. after evaporation, 1 ml of phosphate buffer (ph 7) was added and samples were mixed for 30 minutes. two minutes of ultrasound were used for the lysis of cells. after centrifugation at 8000 rpm for 10 minutes, the supernatant was used in the following experiments. for electrochemical measurements 25, 50, 75, and 150 µm of metal ions were used. electrochemical measurement of metallothionein by differential pulse voltammetry differential pulse voltammetric measurements were performed with 747 va stand instrument connected to 693 va processor and 695 autosampler (metrohm, switzerland), using a standard cell with three electrodes, a cooled sample holder, and measurement cell cooled to 4 °c (julabo f25, germany). a ruttkay-nedecky et al. admet & dmpk 3(3) (2015) 287-295 290 hanging mercury drop electrode (hmde) with a drop area of 0.4 mm 2 was the working electrode. an ag/agcl/3m kcl electrode was the reference and platinum electrode was auxiliary. for data processing va database 2.2 by metrohm ch was employed. the analysed samples were deoxygenated prior to measurements by purging with argon (99.999 %) saturated with water for 120 s. brdička supporting electrolyte containing 1mm co(nh3)6cl3 and 1m ammonia buffer (nh3(aq) + nh4cl, ph = 9.6) was used. the supporting electrolyte was exchanged after each analysis. the parameters of the measurement were as follows: initial potential of -0.700 v, end potential of -1.750 v, modulation time 0.057 s, time interval 0.2 s, step potential 0.002 v, modulation amplitude -0.250 v, eads = 0 v, volume of injected sample: 20 µl, volume of measurement cell 2 ml (20 μl of sample and 1980 µl brdicka solution) for calibration curves. for calibration curve measurements mt2a standard (ikzus proteomics, italy) was used. the volume for the measurement of bacterial culture of e. coli and e. coli with mt-3 or hprp c with metals was 100 µl of bacterial culture and 1900 µl of brdicka solution. statistical analyses software statistica (data analysis software system) version 10.0 (tulsa, oklahoma, usa) was used for data processing. half-maximal concentrations (ic50) were calculated from logarithmic regression of sigmoidal dose-response curve. general regression model was used to analyse differences between the combinations of compounds. to reveal the differences between cell lines, tukey’s post hoc test within homogenous groups and also f-test was employed. unless noted otherwise, p < 0.05 was considered significant. data from 3 measurements were analysed. results and discussion the aim of this study was to determine the effect of copper, zinc and cadmium ions on metallothionein level and on the growth of two different e. coli bacterial cultures transformed with the plasmid containing hprp c or mt-3 gene, and control e. coli (bl21(de3)) strain. proteins prp c and mt-3 are able to bind heavy metals, thus protecting cells from their toxicity and we attempted to test this hypothesis by the mentioned experiments. verification of the expression of recombinant human prion protein (hprp c ) and metallothionein 3 (mt-3) using western-blot the presence of the hprp c expression was verified using western-blot (fig. 2). to achieve better production and stimulation of proteins in bacterial cells iptg (isopropyl β-d-1-thiogalactopyranoside) was added to culture. bacterial cells, which were expressing hprp c protein, were divided into pellet and supernatant after lysis and subsequent centrifugation. after electrophoretic separation by sds-page (fig. 2a), the proteins were transferred to a nylon membrane where they were detected with antibodies against hprp c protein. positive band of hprp c protein with size of about 20 kda was detected in the pellet of bacterial cells in a sample with the addition of iptg only. this finding can be explained by possible aggregation and crystallization of hprp c in pellet (fig. 2b). these results were also confirmed by dot-blot (not shown). admet & dmpk 3(3) (2015) 287-295 prion interactions with metals and metallothionein doi: 10.5599/admet.3.3.191 291 figure 2. verification of the prion protein (hprpc) presence in e. coli bl21(de3) cells transformed with a plasmid containing hprpc gene. red ellipse indicates occurrence of prion protein (size 20-23 kda). (a) gel stained with coomassie brilliant blue, standard: precision plus protein dual xtra standards (bio-rad); (b) western blot, standard: precision plus protein dual xtra standards (bio-rad) mr-molecular mass standard. for both (a and b), lane 1: hprpc supernatant, lane 2: hprpc pellet, lane 3: hprpc /+iptg supernatant, lane 4: hprpc /+iptg pellet. in the same way, the presence of mt-3 protein expression was verified (fig. 3a). using western-blot a positive band with a size of about 10 kda in both cell fractions (supernatant and pellet) supplemented with iptg was verified (fig. 3b). the same results were achieved also by dot-blot method (not shown). figure 3. verification of the metallothionein (mt-3) presence in e. coli bl21(de3) cells transformed with a plasmid containing mt-3 gene. red ellipses indicate occurrence of metallothionein 3 protein (size 7-10 kda). (a) gel stained with coomassie brilliant blue, standard: precision plus protein dual xtra standards (bio-rad); (b) western blot, standard: precision plus protein dual xtra standards (bio-rad), mr-molecular mass standard. for both (a and b), lane 1 e. coli bl21(de3) supernatant, lane 2: e. coli bl21(de3) pellet, lane 3: mt-3 supernatant, lane 4: mt-3 pellet, lane 5: mt-3/+iptg supernatant, lane 6: mt-3/+iptg pellet. ruttkay-nedecky et al. admet & dmpk 3(3) (2015) 287-295 292 the effect of metal ions (cu 2+ , zn 2+ and cd 2+ ) on growth of bacterial cultures expressing hprp c or mt-3 proteins an effect of copper, zinc and cadmium ions on growth of bacterial cultures expressing hprp c or mt-3 proteins was investigated by growth curve method. the procedure for evaluating antimicrobial activity of three chosen heavy metals (cu 2+ , zn 2+ and cd 2+ ) was based on measuring the absorbance of transformed bacterial cultures of e. coli expressing the hprp c and mt-3 proteins compared to control bacterial e. coli bl21(de3) culture, which was also exposed to the action of mentioned heavy metals. figure 4 shows the effect of metal ions (cu 2+ , zn 2+ and cd 2+ ) on the proliferation of transformed bacterial cultures expressing hprp c or mt-3 genes. figure 4. spectrophotometric determination of growth curves obtained after treatment of e. coli bl21 cells transformed with an empty plasmid (a), plasmid containing hprp c gene (b) or mt-3 gene (c) with 0, 10, 50, 150 and 300 μm concentration of metal ions cd (a), cu (b) and zn (c). average values ic 50 (50 % inhibitory concentration) calculated from all growth curves are given. data represent mean calculated from 3 measurements. cells, which expressed mt-3 protein and were exposed to the increasing concentration of cadmium ions, exhibited more effective protection against toxic effects induced by cadmium(ii) (ic 50 = 90.1 μm) than cells expressing hprp c protein (ic 50 =11.2 μm) (figs. 4a-a,b, c). on the other hand, the application of admet & dmpk 3(3) (2015) 287-295 prion interactions with metals and metallothionein doi: 10.5599/admet.3.3.191 293 copper(ii) ions caused on the cells expressing hprp c higher inhibition of growth (ic 50 = 2.7 μm) and growth was decreased by 46 % in presence of 300 µm concentration of copper (figs. 4b-a;b). the concentration required for growth inhibition of 50 % of cells was measured on 5.1 µm and the highest used concentration of copper (ii) caused 56 % growth inhibition of cells (fig. 4b-a). the best protection against copper (ii) toxicity showed cells expressing mt-3 (ic 50 = 10.0 μm) with 37 % inhibition of bacterial growth (fig. 4b-c). similar findings as in the presence of cadmium (ii) ions were also observed in the case of zinc (ii) ions. the cells expressing hprp c protein were exposed to the increasing concentration of zinc(ii) ions. the best protection against toxicity of metals was observed after addition of zinc (ii) ions (ic 50 = 520.8 μm), as it is shown in figs. 4c-a,b. these bacterial cultures were inhibited by treatment with zinc (ii) ions. the cells expressing mt-3 (ic 50 = 46.9 μm, figs. 4c-c) were less protected (25 % inhibition of bacterial growth in presence of 300 μm concentration of zinc (ii) and the least protected were the control cells (ic 50 = 5.0 μm, fig. 4c-a). the control cells were inhibited by up of 69 % after addition of the same concentration of zinc (ii) ions. electrochemical determination of the total mt level after addition of metal ions (cu 2+ , zn 2+ and cd 2+ ) to the bacterial cultures expressing hprp c or mt-3 proteins before the measurement of total mt level, a calibration curve (fig. 5a) was determined by measuring different concentrations of mt 2a standard (0.39 – 25.0 μg/ml), whereas dp voltammograms are shown in fig. 5b. catalytic signal cat2 measured at the potential of -1.56 v linearly corresponded to the concentration of metallothionein. mt values obtained by reading from the calibration curve were recalculated to mg of protein determined in the sample. figure 5. (a) calibration curve of total mt determined using dpv method, as electrolyte brdicka solution was used (1 mm co(nh3)6cl3 and 1 m ammonium buffer (nh3(aq) and nh4cl, ph = 9.6)). experimental parameters were chosen as follows: initial potential -0.4 v, end potential -1.75 v, modulation time 0.057 s, time interval 0.2 s, step potential 2 mv, and modulation amplitude 250 mv. (b) dp voltammograms of mt standards (0.39 25 µg/ml) obtained from the calibration curve. ruttkay-nedecky et al. admet & dmpk 3(3) (2015) 287-295 294 further, dpv measurements of the total mt level in the control e. coli cells and in e. coli cells expressing hprp c or mt-3 proteins were performed. these cultures were exposed to the increasing concentration of two essential metal ions (cu 2+ and zn 2+ ) and one toxic metal ion (cd 2+ ). bacteria that expressed mt-3 protein, had slightly higher mt levels after application of increasing metal concentrations (25, 50, 75 and 150 µm) in comparison with control cells with similar pattern (fig. 6). mt expression increased at lowest metal concentration (25 µm), followed by a decrease at higher metal concentrations (50, 75, and 150 µm). the highest increase in mt levels in bacterial cells after application of cd 2+ were observed followed by cells with addition of zn 2+ . the smallest increases in mt levels were observed in cells with addition of cu 2+ . mt expression pattern in hprp c transformed cells was different. after application of cu 2+ an increase in mt expression continued also at higher metal concentrations; and after application of cd 2+ and zn 2+ no decrease in mt expression at higher metal concentrations was observed. figure 6. metallothionein content in e. coli cells bl21 transformed with an empty plasmid, the plasmid containing hprp c or mt-3 gene after incubation with different concentrations of (a) cu, (b) zn and (c) cd in concentrations of 25, 50, 75, and 150 μm compared to control cells without addition of metal. data represent mean ± s.d. calculated from 3 measurements, ns, not significant, *p < 0.05. conclusions clarifying the precise function of mt-3 and the conversion of the prion protein into its pathological form would contribute to a better understanding of the emergence of transmissible spongiform encephalopathies and the development of effective drugs to cure them. the first step for research of brain protein functions is the use of a bacterial model, where it is possible to observe the effect of metals on growth of bacterial cells, in which the brain proteins are produced. another interesting topic is the monitoring of interconnection between functions of prion protein and mt-3. in this work the presence of both hprp c and mt-3 protein expression in bacterial cells was verified using western-blot and dot-blot. furthermore, by the growth curve measurement and electrochemical admet & dmpk 3(3) (2015) 287-295 prion interactions with metals and metallothionein doi: 10.5599/admet.3.3.191 295 determination of mt protein levels, it was found that after addition of zn 2+ and cd 2+ mt-3 protein was more effective in protection of bacterial cells against toxicity of metals (zn 2+ and cd 2+ ) than hprp c . the different situation occurred in the case of addition of cu 2+ , when the mt-3 was equally effective as hprp c , whose expression even increased the inhibitory effect of copper ions on the growth of bacterial cells. acknowledgements: financial support by ceitec cz.1.05/1.1.00/02.0068 is highly acknowledged. the authors express their thanks to vladimir pekarik for the preparation of plasmids and to yazan haddad for a correction of the english language style. references [1] s. ortega-cubero, m.r. luquin, i. dominguez, j. arbizu, i. pagola, m.m. carmona-abellan, m. riverol, neurologia 28 (2013) 299-308. [2] s.b. prusiner, m.r. scott, s.j. dearmond, f.e. cohen, cell 93 (1998) 337-348. [3] k.m. pan, n. stahl, s.b. prusiner, protein sci. 1 (1992) 1343-1352. [4] m. glatzel, k. stoeck, h. seeger, t. luhrs, a. aguzzi, arch. neurol. 62 (2005) 545-552. [5] a. gaeta, r.c. hider, br. j. pharmacol. 146 (2005) 1041-1059. [6] k. hur, j.i. kim, s.i. choi, e.k. choi, r.i. carp, y.s. kim, mech. ageing dev. 123 (2002) 1637-1647. [7] h.f. ji, h.y. zhang, trends biochem. sci. 35 (2010) 129-134. [8] p. tiraboschi, f. tagliavini, nat. rev. neurol. 9 (2013) 366-367. [9] p.e. fraser, j. biol. chem. 289 (2014) 19839-19840. [10] v.c. dodelet, n. r. cashman, blood 91 (1998) 1556-1561. [11] s.b. prusiner, science 252 (1991) 1515-1522. [12] h. kretzschmar, j. tatzelt, brain pathol. 23 (2013) 321-332. [13] d.r. brown, w.j. schulzschaeffer, b. schmidt, h.a. kretzschmar, exp. neurol. 146 (1997) 104-112. [14] p. sobrova, m. ryvolova, v. adam, r. kizek, electrophoresis 33 (2012) 3644-3652. [15] p. sobrova, m. ryvolova, d. hynek, v. adam, j. hubalek, r. kizek, int. j. electrochem. sci. 7 (2012) 928-942. [16] p. sobrova, i. blazkova, j. chomoucka, j. drbohlavova, m. vaculovicova, p. kopel, j. hubalek, r. kizek, v. adam, prion 7 (2013) 349-358. [17] p. sobrova, m. ryvolova, v. pekarik, j. hubalek, v. adam, r. kizek, int. j. electrochem. sci. 8 (2013) 12466-12475. [18] a. cardova, p. sobrova, r. bujdoso, v. adam, r. kizek, j. metallomics nanotech. 1 (2014) 27-29. [19] e. sedlackova, a. cardova, v. adam, r. kizek, j. metallomics nanotech. 1 (2014) 23-26. [20] c. cirelli, g. tononi, brain res. 885 (2000) 303-321. [21] g. meloni, v. sonois, t. delaine, l. guilloreau, a. gillet, j. teissie, p. faller, m. vasak, nat. chem. biol. 4 (2008) 366-372. [22] m. vasak, d.w. hasler, curr. opin. chem. biol. 4 (2000) 177-183. ©2015 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ silver(i) complexes with phenolic schiff bases: synthesis, anti­bacterial evaluation and interaction with biomolecules doi: http://dx.doi.org/10.5599/admet.1167 197 admet & dmpk 10(3) (2022) 197-212; doi: https://doi.org/10.5599/admet.1167 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper silver(i) complexes with phenolic schiff bases: synthesis, antibacterial evaluation and interaction with biomolecules natalia loginova*1, maxim gvozdev1, nikolai osipovich2, alina khodosovskaya3, tatiana koval’chuk-rabchinskaya1, galina ksendzova2, dzmitry kotsikau1 and anatoly evtushenkov3 1faculty of chemistry, belarusian state university, leningradskaya str. 14, 220030 minsk, belarus 2research institute for physico-chemical problems of the belarusian state university, leningradskaya str. 14, 220030 minsk, belarus 3faculty of biology, belarusian state university, 4 independence avenue, 220030 minsk, belarus *corresponding author: e-mail: loginonv@gmail.com; tel.: +375-172-095-516; fax: +375-172-424-998 received: november 11, 2021; revised: march 09, 2022; available online: march 18, 2022 abstract novel ag(i) complexes (2a–2c) with phenolic schiff bases were synthesized using 4,6-di-tert-butyl-3-(((5mercapto-1,3,4-thiadiazol-2-yl)imino)methyl)benzene-1,2-diol (1a), 4,6-di-tert-butyl-3-(((4-mercaptophenyl)imino)methyl)benzene-1,2-diol (1b), and 4,6-di-tert-butyl-3-(((3-mercaptophenyl)imino)methyl)benzene-1,2-diol (1c). they were examined by elemental analysis, ft-ir, uv-vis, 1h-nmr spectroscopy, xrd, cyclic voltammetry, conductivity measurements, and biological methods. the complexes are characterized by distorted geometry of the coordination cores agn2s2 (2c), agns (2b) and ags2 (2a). these stable complexes were not typified by the intramolecular redox reaction in organic solvents resulting in the formation of silver nanoparticles (agnps). antibacterial activity of 1a–1c and 2a–2c was evaluated in comparison with agnps and commonly used antibiotics. all the complexes were more active than the ligands against the bacteria tested (14), but they were less active than agnps and commonly used antibiotics. both 1a–1c and their complexes 2a–2c exhibited the capability for the bovine heart fe(iii)-cyt c reduction. the ligands 1b and 1c were characterized by the highest reduction rate among the compounds under study, and they showed a higher reducing ability (determined by cyclic voltammetry) as compared with that of their ag(i) complexes 2b and 2c. ©2022 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords silver; phenolic compounds; nanoparticles; antimicrobial activity; bsa binding; cytochrome c reduction introduction the development of novel effective antibacterial agents assumes prime importance in the context of the ever-growing frequency of infections caused by the strains of bacteria resistant to widely used pharmaceuticals [1]. silver is biocidal in its ionic form and, unlike many antibiotics, has several various mechanisms of antimicrobial action. in particular, sulphhydryl groups both in bacterial cell walls and enzymes were shown to be vulnerable to denaturation; nucleoproteins and nucleic acids are additional http://dx.doi.org/10.5599/admet.1167 https://doi.org/10.5599/admet.1167 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:loginonv@gmail.com http://creativecommons.org/licenses/by/4.0/ loginova, gvozdev, osipovich et al. admet & dmpk 10(3) (2022) 197-212 198 target sites [2]. these multiple targets render the appropriate mutations required for microorganisms to become resistant unlikely. moreover, silver-containing pharmaceuticals have the following advantages: a low induction of resistance compared to antibiotics; a broad spectrum of activities (bacteria, yeasts and moulds); and safety in a proper dose [2]. in this connection, silver-containing pharmaceuticals turned out to be promising for medicine again. pharmacological screening of the ag(i) complexes with biologically active substances for antimicrobial activity represents a strategy for the development of a novel class of antimicrobials that have a different mode of action compared to the commonly used antibiotics. in the last twenty years, screening has been carried out intensively to reveal the antimicrobial activity of the ag(i) complexes with a variety of organic substances [3-10]. though in rare cases, microorganisms were found to be resistant to silver and its compounds. this resistance may be overcome by using silver complexes due to changes in their hydrophilic/lipophilic characteristics, solubility, structure and the ability of the complex to take part in exchange reactions with bioligands [2,11]. thus, the more loosely bound is ag(i) ion in coordination core of the complex, the more probable is its interaction with soft bases of target biomolecules (proteins, dna, etc.), and primarily with sulphur-containing functional groups of proteins [12]. this concept allows one to predict the antimicrobial properties of ag(i) complexes synthesized rather effectively. however, it doesn’t take into account the role of redox activity of ag(i) compounds in realizing their bioactivity and doesn’t consider a possibility of correlation between these properties. different antibacterial agents are known to act in vivo as electron transfer agents in the production of radical species or disruption of normal electron transport [13]. electron transfer can occur either to the metal center or, if in the biologically accessible range, to redox-active ligands, resulting in reactive species capable of attacking biologically relevant target molecules respectively either by ligand displacement at the metal center or by radicals formed at the ligand entity [14,15]. in our view, the redox-active ag(i) complexes with sterically hindered phenolic derivatives are a particularly rich source of effective antibacterial agents. nowadays phenolic derivatives as antibacterial agents are of limited utility because of their toxicity and irritating action. introduction of substituents into benzene ring and ag(i) ion complexation, which change the hydrophilic-lipophilic balance of a compound, will allow one not only to achieve an optimal biocidal effect but also to decrease their toxicity. previously we have carried out the synthesis and characterization of ag(i) complexes with some orthodiphenol derivatives (hydrazone, thiosemicarbazone derivatives, as well as derivatives of thiocarboxylic acids) [16-18]. they displayed bioactivity against microorganisms tested comparable with or even higher than that of some commonly used antibiotics and silver-containing drugs. unfortunately, the abovementioned ag(i) complexes were found to be unstable in solvents with a high solvating power. they were shown to be complexes with partial charge transfer (pct), which belong to a special class of transition metal complexes with redox-active ligands (ortho-dioxolenes). the intramolecular electron transfer between the catechol-containing ligand and the metal ion is observed therein [19]. thus the ag(i) complexes acquire partly semiquinonate character and are typified by an intramolecular redox reaction in some organic solvents resulting in the formation of silver nanoparticles (agnps) [20-23]. for this reason, we undertook a design of the ag(i) complexes with phenolic schiff bases for the purpose of inhibiting bacterial growth and improving the stability of the redox-active ag(i) complexes in solution. we chose schiff bases because they have important implications for drug discovery and development owing to the structural diversity and ease of synthesis; they were found to be potent against fungi, bacteria, protozoa [24]. on the other hand, the generation of agnps by decomposition of the ag(i) admet & dmpk 10(3) (2022) 197-212 silver(i) complexes with schiff bases doi: http://dx.doi.org/10.5599/admet.1167 199 complexes in solution deserves particular attention. we described the peculiarities of agnps formation on using the ag(i) complexes with 2-[4,6-di-(tert-butyl)-2,3-dihydroxyphenylsulfanyl]acetic acid and 4,6-di-tertbutyl-2,3-dihydroxybenzaldehyde isonicotinoyl hydrazone as precursors [20-23]. according to our previous results, a positive feature of this method is the fact that the ag(i) complex with pct (as a precursor) includes an oxidizer, a reductant and an agnps stabilizer. moreover, to understand the mechanism of activity of silver formulations, we must consider the possibility of agnps acting as the ultimate biocidal bullets that will contribute to the efficiency of the redox-active ag(i) complexes [25]. therefore, the second key task of the present work is to compare the antibacterial activity of agnps prepared by the abovementioned method with that of their silver precursors. in the earlier investigations, using cyclic voltammetry, we have shown some sterically hindered phenolic derivatives as well as their ag(i) complexes to be also of a pronounced reducing ability correlating with antibacterial activity and the rate of the reduction of bovine heart cytochrome c (fe(iii)-cyt с) [26]. those results allowed us to suggest that redox processes could play an important part in the biotransformation and pharmacological activity of these compounds, and one of the possible types of their biological macromolecular targets can be components of electron transport chains such as cyt c-like ones. in particular, the most sensitive enzymes’ sites for the action of silver lay between cytochrome b and a3 [2]. owing to subcellular localization, bacterial cytochromes are among the first target enzymes for antiinfective agents on their way into the cell. account should also be taken of the known relation of the abovementioned enzyme system to generation and detoxification of reactive oxygen species [27]. the results obtained are discussed in the view of a supposed relationship between their antibacterial activity, the capability of the compounds under study for reducing fe(iii)-cyt с, redox properties determined by cyclic voltammetry, lipophilicity, and their ability to interact with bsa. experimental materials and methods elemental analyses were carried out with a vario el instrument (chns mode). silver was determined using an atomic emission spectrometer with an inductively coupled plasma excitation source (spectroflame modula). 1н nmr spectra (dmso-d6 was used as a solvent) were recorded using a bruker avanсe-500 spectrometer operating at 500 mhz. mass spectra were registered with a shimadzu qp-5000 spectrometer using direct injection of the specimens into an ion source with a temperature of the source 200 °с and ionization energy of 70 ev. uv-vis absorption spectra of the compounds were recorded in acetonitrile (hplc grade). spectrophotometric experiments were performed with a solar pb 2201 spectrophotometer using a quartz cuvette with 1 cm optical path. the ir spectra were registered with an avatar ft-ir-330 (thermo nicolet) spectrometer in the wavelength range of 400–4000 cm−1 using a smart diffuse reflectance accessory. esr spectra of solid samples were measured with an ers-220 x-band spectrometer (9.45 ghz) at room temperature, using 100-khz field modulation; g factors were quoted relative to the standard marker 2,2-diphenyl-1-picrylhydrazyl (dpph). the molar conductance of 10–4 m solutions of the silver complexes in dmso was measured at room temperature using a tesla bms91 conductometer (cell constant 1.0). fluorometric experiments were performed with a solar cm 2203 spectrofluorometer using a quartz cuvette with 1 cm optical path. the lipophilicity test was made by determining the n-octanol/water partition coefficient (logp) according to the method reported in the literature [28]. cyclic voltammetry measurements were performed under dry nitrogen in a three-electrode two-compartment electrochemical cell using a potentiostat with glassy-carbon (gc) working electrode, pt auxiliary electrode and ag|agcl|0.1 mol·l–1 (c2h5)4ncl reference electrode. the supporting electrolyte was 0.1 mol·l–1 (c2h5)4nclo4. the http://dx.doi.org/10.5599/admet.1167 loginova, gvozdev, osipovich et al. admet & dmpk 10(3) (2022) 197-212 200 ag|agcl|0.1 mol·l–1 (c2h5)4ncl reference electrode was calibrated with the ferrocenium | ferrocene redox couple (this couple has potential +0.54v vs. our reference electrode). anhydrous acetonitrile was used as a solvent. (c2h5)4nclo4 and (c2h5)4ncl used to prepare solutions were dried respectively at 80 and 100 °с under vacuum for 3 h. preparation of solutions and filling the electrochemical cell were carried out in a glove box in dry nitrogen. synthesis of the ligands the aldehyde 1 was obtained by duff reaction using 3,5-di-tert-butylbenzene-1,2-diol and hexamethylenetetramine in glacial acetic acid as starting materials [17]. compounds 1b–1c were prepared by the reaction of aromatic amines with 4,6-di-tert-butyl-2,3-dihydroxybenzaldehyde 1 in the molar ratio 1:1 using anhydrous methanol as a solvent. the ligand 1a was obtained upon refluxing aldehyde 1 with 5amino-1,3,4-thiadiazole-2-thiol in the mixture of acetic acid and methanol (1:2) for 12 h (fig. 1). figure 1. scheme of synthesizing the ligands 1a–1c. compound 1a: orange crystals from ethanol, yield 57 %; m.p. 111–113 ºс. 1h nmr (500 mhz, dmso-d6) δ, ppm: 1.36 s (9h, 3ch3), 1.44 s (9h, 3ch3), 6.82 s (1н, char), 7.09 s (1н, oh), 8.75 s (1h, ch=n), 13.17 s (1н, sh), 14.60 br. s (1h, oh). ft-ir (ν, cm-1): 3342s, ν(o–h); 1616s ν(c=n); 1166m, 1111m ν(c–o), 1271m ν(carom–n), 2576w ν(s–h), 688w ν(c–s), 634w, 601w ν(cr–s–). uv-vis: λ, nm (logε, m-1 cm-1): 211 (4.25), 289 (4.28), 370 (3.62). mass spectrum, m/z (irel, %): 365 (100) [м]+. [с17h23n3o2s2]+. compound 1b: red solid from ethanol, yield 64 %; m.p. 189–191 ºс. 1h nmr (500 mhz, dmso-d6) δ, ppm: 1.37 s (9h, 3ch3), 1.44 s (9h, 3ch3), 5.65 br. s (1h, sh), 6.77 s (1н, снar), 7.27–7.30 m (2h, char), 7.40– 7.42 m (2h, char), 8.25 s (1h, oh), 9.36 s (1н, ch=n), 15.26 s (1h, oh). ft-ir (ν, cm-1): 3394s, ν(o–h); 1610s ν(c=n); 1162m, 1097m ν(c–o), 1247m ν(carom–n), 2560w ν(s–h), 705w, 669w ν(c–s). uv-vis: λ, nm (logε, m1 cm-1): 204 (4.57), 227sh, 302sh, 344 (4.45). mass spectrum, m/z (irel, %): 357 (100) [м]+: [с21h27no2s]+. compound 1c: red solid from ethanol, yield 68 %. m.p. 138–140 ºс. 1h nmr (500 mhz, dmso-d6) δ, ppm: 1.38 s (9h, 3ch3), 1.45 s (9h, 3ch3), 5.75 br. s (1h, sh), 6.78 s (1н, снar), 7.10 d (1h, char, j = 7.8 hz), 7.24 d (1h, char, j = 7.8 hz), 7.29–7.31 m (1h, char), 7.37 t (1h, char, j = 7.8 hz), 8.39 s (1h, oh), 9.35 s (1н, ch=n), 15.12 s (1h, oh). ft-ir (ν, cm-1): 3346s, ν(o–h); 1627s ν(c=n); 1162m ν(c–o), 1263m ν(carom–n), 2577w ν(s–h), 676w, 646w, 613w ν(c–s). uv-vis: λ, nm (logε, m-1 cm-1): 206 (4.43), 225 (4.36), 328 (4.27). mass spectrum, m/z (irel, %): 357 (92) [м]+: [с21h27no2s]+. synthesis of the complexes a novel procedure was developed for synthesizing ag(i) complexes with the ligands 1a–1c making it possible to prevent redox interaction with ag(i) ions and to control the ligand environment of ions and thus to produce complexes with specified characteristics. the ag(i) complexes 2a and 2b were isolated from the admet & dmpk 10(3) (2022) 197-212 silver(i) complexes with schiff bases doi: http://dx.doi.org/10.5599/admet.1167 201 mixture of acetonitrile/acetone (1:5 v/v) upon the reaction respectively of the parent ligands 1a and 1b with ag(i) nitrate in the molar ratio metal/ligand 1:1. the compound 2c was obtained with the molar ratio metal/ligand 1:2 (fig. 2). compound 2a: orange powder, yield 84 %. anal. calcd. for c17h22agn3o2s2: c, 43.23; h, 4.69; ag, 22.84; n, 8.90; s, 13.57; found: c, 43.14; h, 4.53; ag, 22.91; n, 8.74; s, 13.49. 1h nmr (500 mhz, dmso-d6) δ, ppm: 1.36 s (9h, 3ch3), 1.44 s (9h, 3ch3), 6.82 s (1н, char), 7.57 br. s (1н, oh), 8.75 s (1h, ch=n). ft-ir (ν, cm-1): 3520s, 3340s, ν(o–h); 1612s ν(c=n); 1168m ν(c–o), 1270m ν(carom–n), 673w ν(c–s), 588w ν(cr–s–). uv-vis: λ, nm: 212, 288, 367. molar conductivity λ: 20.0 ω-1cm2mol-1. compound 2b: dark red solid, yield 95 %. anal. calcd. for c21h26agno2s: c, 54.32; h, 5.64; ag, 23.23; n, 3.02; s, 6.90; found: c, 54.26; h, 5.47; ag, 23.19; n, 3.11; s, 6.81. ft-ir (ν, cm-1): 3410s, ν(o–h); 1651s ν(c=n); 1168m, 1087m ν(c–o), 1224m ν(carom–n), 690w, 642w ν(c–s), 441m ν(n–ag). uv-vis: λ, nm: 201, 288, 385. molar conductivity λ: 12.5 ω-1cm2mol-1. compound 2c: orange powder, yield 92 %. anal. calcd. for c42h53agn2o4s2: c, 61.38; h, 6.50; ag, 13.12; n, 3.41; s, 7.80; found: c, 61.21; h, 6.01; ag, 13.34; n, 3.32; s, 7.74. 1h nmr (500 mhz, dmso-d6) 1.25 s (18h, 3ch3), 1.32 s (18h, 3ch3), 6.65 s (2h, char), 6.78 br. s (5н, снar), 7.27 br. s (2h, char), 7.43 br. s (2h, char), 8.11 s (2h, oh), 9.16 s (2н, ch=n), 15.11 br. s (2h, oh). ft-ir (ν, cm-1): 3504s, ν(o–h); 1610s ν(c=n); 1162m ν(c–o), 1245m ν(carom–n), 632w, 572w ν(c–s); 1072s ν(c=s), 443m, 418m ν(n–ag). uv-vis: λ, nm: 209, 291, 385. molar conductivity λ: 3.7 ω-1cm2mol-1. synthesis of silver nanoparticles (agnps) silver organosols were produced by chemical decomposition of the ag(i) complexes with 2-[4,6-di-(tertbutyl)-2,3-dihydroxyphenylsulfanyl]acetic acid (3a) and 4,6-di-tert-butyl-2,3-dihydroxybenzaldehyde isonicotinoyl hydrazone (3b) (fig. 3) when the latter was dissolved in media with high donor numbers dn > 19 (ethanol, propanol-2, butanol-1, dimethyl formamide, dimethyl sulphoxide [29]) under permanent stirring. synthesis and characterization of the ag(i) complexes 3a and 3b used in this work were described elsewhere [16-18]. these sols were produced from such an amount of the complex dissolved in an organic solvent as to provide the silver concentration in the sol equal to 10-4 m. the sols were stored in closed, light-tight vessels. if required, argon was bubbled through the solutions to ensure the absence of oxygen. agnps in solid-state were prepared by evaporating the solvent using a rotary evaporator rv 8 v (ika). measuring the mass of the dry residue, the silver concentration in the sol was found to be 88.9–90.0 µg·ml– 1 [20,22,23]. the silver content in the solid phase was determined using atomic emission spectrometry and was equal to 12.1 % [20,22,23]. antibacterial activity the following test microorganisms (collection of department of molecular biology, belarusian state university) were used: gram-negative bacteria (escherichia coli, pseudomonas putida, proteus vulgaris, pantoea agglomerans, serratia marcescens, pectobacterium carotovorum, salmonella typhimurium, pseudomonas aeruginosa), gram-positive bacteria (bacillus subtilis, staphylococcus saprophyticus, bacillus pumilus p10, bacillus pumilus b9, staphylococcus aureus, sarcina lutea). the microorganisms were subcultured for testing in muller-hinton broth (merck) (ph 7.4±0.2). the cells were suspended, according to the mcfarland protocol, in saline solution to produce a suspension of about 106 cfu/ml (colony-forming units per ml). because all the compounds tested are almost insoluble in water, they were dissolved, according to standard recommendations [30,31], in a small volume of an organic solvent (dimethyl sulfoxide), which is well miscible with the aqueous nutrient medium on dilution and doesn’t react with the http://dx.doi.org/10.5599/admet.1167 loginova, gvozdev, osipovich et al. admet & dmpk 10(3) (2022) 197-212 202 compounds tested. if organic solvent was diluted in mueller-hinton broth at 1:10 to 1:100, there was no detectable inhibition or antagonism associated with the residual solvent concentration when tested against bacteria [30,31]. this solvent is used for making stock solutions (1.6 mg·ml–1) of the test compounds. stock solutions can be further diluted with the broth. serial dilutions of stock solutions to certain concentrations were carried out in test tubes. concentrations of the compounds under study evaluated in mueller-hinton broth ranged from 400 µg·ml–1 to 1.6 µg·ml–1. a specified volume of a 24 h old inoculum was added to each tube. the organic solvent (dimethyl sulfoxide) at the final concentration (1–2 %) in the nutrient medium did not affect the growth of the test microorganisms. the incubation was carried out at the optimal growth temperature for 24 h, and optical density (od600) was determined for the bacterial cultures with and without the compounds tested. the contents of the test tubes in which the concentration of these compounds was sufficient to suppress the microbial growth remained clear, while their turbidity was evidence for the presence of bacteria. the mic, defined as the lowest concentration of the test compound that inhibits the visible microbial growth, was determined after incubation. mic values are given in µmol·ml–1 to reveal a correlation between the antibacterial activity and reducing ability of the compounds. tests using dimethyl sulfoxide as negative control were carried out in parallel. results were always verified in three separate experiments. interaction with bsa in the fluorescence quenching experiment, quenching of the tryptophan residues of bsa was done by keeping the constant concentration of bsa (2 µm) in 5 mm tris-hcl buffer (ph 7.4) with 50 mm nacl while varying that of the complexes (0–10 µm). the fluorescence spectra were recorded in the range of 310–450 nm upon excitation at 295 nm and an emission wavelength of tryptophan residues of bsa at 339 nm after each addition of the quencher. solutions of the complexes were prepared in dmso. all experiments were carried out at room temperature. the fluorescence quenching is described by the stern-volmer equation [32]: f0/f = 1 + ksv[q] (1) f and f0 are fluorescence intensities respectively with and without the quencher (the complex), ksv is a stern–volmer quenching constant, and [q] is the quencher concentration. the quenching constant (ksv) was calculated from the slope of the plot of f0/f versus [q]. the hill coefficient (n) and the apparent binding constant (kb) values were calculated from the plot of log [(f0 − f)/f] versus log[q] according to the following equation [32]: log [(f0−f)/f] = log kb + nlog [q] (2) interaction with cytochrome c bovine heart fe(iii)-cyt с (sigma) was used. fe(iii)-cyt с concentration was determined on its interaction with excess sodium dithionite using the absorption coefficient ε550 = 21000 m-1cm-1 [33,34]. argonsaturated dmso solutions of the ligands and complexes and fe(iii)-cyt с (7 µm) were used. experiments were performed in 10 mm tris-hcl buffer (ph 7.6) at 20 °c. aliquots of the compounds under study were added to fe(iii)-cyt с solution up to the final concentration 35.0 μm. the initial rate of fe(iii)-cyt с reduction (ν0) was calculated by the slope of the linear portion of kinetic curve a550 versus time. the results were confirmed in three independent experiments. admet & dmpk 10(3) (2022) 197-212 silver(i) complexes with schiff bases doi: http://dx.doi.org/10.5599/admet.1167 203 results and discussion physicochemical characterization the catechol-based ligands 1a–1c were obtained via condensation of a primary amine with aldehyde 1 (fig. 1). the ag(i) complexes 2a–2c were isolated in amorphous or poorly crystalline state from the mixture of acetone/acetonitrile. they were practically insoluble in non-polar organic solvents, slightly soluble in acetonitrile, but soluble in dmso and dmf. the values of the molar conductivity (λ ≤ 20 ω-1cm2m-1) in dmso characteristic of the metal complexes under study indicate that they are non-electrolytes [35]. 1h nmr spectra of the ligands and the ag(i) complexes were recorded in dmso-d6. the spectra of schiff bases 1a–1c display two singlets at 7.09–8.39 and 14.60–15.26 ppm corresponding to the phenolic protons of a catechol moiety. the aromatic protons of the catechol moiety were observed as singlets at ~ 6.8 ppm. the formation of the ligands 1a–1c was also evidenced by the appearance of singlets at 8.75–9.36 ppm corresponding to the azomethine protons. the signals of sh protons resonated at 5.65, 5.75 and 13.17 ppm as singlets. the disappearance of these signals in the spectra of ag(i) complexes 2a and 2c indicates the participation of the mercapto group in the coordination via deprotonation. the signal of ch=n proton shifted and was observed at 9.16 ppm for 2c, indicating the involvement of the azomethine group in ag(i) coordination. however, in the spectrum of 2a this signal virtually did not change its position in the spectrum. the minor shift of the hydroxyl proton signals in the spectra of the complexes may result from the participation of hydroxyl groups in different processes: coordination of these groups to the metal ion as well as disruption of the strong hydrogen bonding between the free ligand molecules and the formation of hydrogen bonding involving two uncoordinated hydroxyl groups of one of the ligands. analysis of ir spectra of 1a1c and their ag(i) complexes made it possible to identify donor sites of binding of the ligands to ag(i). the general characteristic of ir spectra of these complexes is the fixed position of the bands corresponding to vibrations of the structural fragments not bonded to ag(i) (benzene ring, tert-butyl groups), which are present in the spectra of the free ligands [36,37]. it should be emphasized that the band characteristic of o-benzoquinones [38] is missing from the spectra of the complexes, thus suggesting the lack of any oxidized ligands in the coordination sphere of the latter. in the spectra of the ligands 1a1c in the region 3394–3342 cm–1 there is a broad band assigned to the stretching vibrations of phenolic о–н groups, and in the spectra of the ag(i) complexes, it is blue-shifted, suggesting that there is no coordination of oxygen atoms to ag(i). this fact is also supported by the lack of changes in the range of 1200–1100 cm–1 corresponding to the stretching с–о bond vibrations. in the spectra of the complexes 2b and 2c, there is a shift of the bands in the region 1651–1610 cm–1 corresponding to the stretching c=n bond vibrations. besides, new bands appearing in the region 500–400 cm–1 suggest nitrogen atom participating in coordination to ag(i) ion. by contrast, new bands missing from the range of 500–400 cm–1 and no changes in the position and intensity of the bands in the spectrum of the complex 2а assigned to the stretching c=n bond vibrations (1616–1610 cm–1) suggest that nitrogen atoms take no part in the formation of coordination core of the complex. in the spectra of the ligands 1a1c in the region 2600–2500 cm–1 there is a weak band assigned to the stretching s–h bond vibrations, absent in the spectra of all the complexes, which is evidence for s atom coordinating in thiolate form. in the spectra of the complexes 2a and 2b there are no bands that could be assigned to the stretching с=s bond vibrations (1050–1200 cm–1). but there is a very strong band at 1072 cm–1 in the spectrum of the complex 2с, suggesting that s atom is coordinated in thion-form along with thiolate one. in the complex 2a it is the sulphur atom of the heterocycle which takes part in coordination to ag(i), as is evidenced by the red shift of the band belonging to the stretching –cr–s– bond vibrations (the latter is present at 630–600 cm–1 for the ligand 1a). http://dx.doi.org/10.5599/admet.1167 loginova, gvozdev, osipovich et al. admet & dmpk 10(3) (2022) 197-212 204 the uv-vis absorption spectra of the ligands exhibit bands in the range of 204–227 nm, which may be assigned to π-π* transitions of the aromatic rings [39]. the intense bands at 289–370 nm are due to π-π* and n-π* transitions of the azomethine group [39]. in the spectra of the complexes 2b and 2c, these bands are shifted to the region 388–385 nm, indicating the participation of azomethine group in the coordination. the bands in the spectrum of 2a virtually do not change their position. as the compounds 1a1c are potential ag(i) reductants, their complexes 2a–2c were investigated by esr method in order to elucidate the possibility of the formation of paramagnetic particles by virtue of redox interaction between ag(i) ions and phenolic ligands. esr method is known to be widely applied to detect and identify paramagnetic silver species (ag(0) and ag(ii)) [40]. however, it was found that they were absent in the ag(i) complexes 2a–2c, as no their characteristics (g-factors) were registered in the esr spectra. for interpretation of the esr spectra of these ag(i) complexes, a concept can be used of complexes with pct as a special class of compounds of transition metal ions with redox ligands [19]. their peculiarity is the existence of rapidly establishing thermal equilibrium between the ground and electron-excited states. owing to this, a pct complex can interact with any “third” particle, being simultaneously in two (ground and electron-excited) states, and can demonstrate properties both of initial reacting particles and of electron-transfer products. “ambiguity” of properties of pct complexes is clearly interpreted in terms of the mulliken's theory [41], according to which a pct complex may be conceived as a superposition of states without charge transfer and with a complete charge transfer. the pct complex decomposes from its ground state to yield the initial reagents, while on decomposing from an electron-excited one, it gives electron-transfer products. thus, according to our previous results, the redox-active ag(i) complexes 3a and 3b with sterically hindered o-diphenol derivatives (fig. 3) were found to acquire partly semiquinonate character (giso=2.003÷2.004), and on dissolving in solvents with high donor numbers (dn ˃ 20) that actively solvate metal cations the pct can initiate a redox process resulting in, as noted above, colloidal silver formation [20-23]. in contrast, the behaviour of the complexes 2a–2c is different from that previously reported for other redox-active ag(i) complexes [16,18,21]. it should be emphasized that no singlet signal with giso close to ge is registered in the esr spectra of the complexes 2a–2c. besides, on dissolving these novel ag(i) complexes in strongly solvating solvents, no agnps are formed. the results obtained support the absence of the pct in complexes 2a–2c. thus, the results of the spectral studies suggest an asymmetric mode of binding to functional groups involved in complexation, and the general mode of ligating atoms in the ag(i) complexes 2a–2c can be represented as shown in fig. 2. figure 2. coordination modes of the ligands 1a–1c in their ag(i) complexes 2a–2c. admet & dmpk 10(3) (2022) 197-212 silver(i) complexes with schiff bases doi: http://dx.doi.org/10.5599/admet.1167 205 figure 3. plausible structures of the ag(i) complexes 3a and 3b. redox characterization: electrochemical study and interaction with cytochrome c the redox properties of the newly synthesized compounds 1a–1c and their ag(i) complexes 2a–2c were determined electrochemically using cyclic voltammetry method. the latter is the most widely used technique for acquiring qualitative information about electrochemical reactions; it offers a rapid location of redox potentials of the electroactive species [42]. cyclic voltammetry allows a successful evaluation of the total antioxidant capacity because the redox potentials of phenolic compounds determined by this method are considered good measures of their reducing ability [43,44]. there are essential reasons for carrying out an investigation of the redox properties of phenolic ligands and their ag(i) complexes. first, the phenolic ligands in metal complexes can be in different redox states depending on conditions (diamagnetic singleor double-charged anions, neutral ortho-benzoquinones or paramagnetic ortho-benzosemiquinone anion-radicals [45]. second, according to our previous data [21,26], it is safe to assume that the redox properties of the phenolic ligands and their ag(i) complexes can affect their bioactivity. it was found that the redox processes involving the compounds under study are reversible or irreversible. that is why the potential value for the first oxidation peak (еpa1, v) was taken as a criterion to compare the reducing ability of these compounds of the same type, and the more cathodic this value is, the more active is the compound as a reductant [42]. on the basis of the electrochemical findings, the compounds investigated can be graded in their reducing ability as follows (table 1): 1a>1c≈1b; 2a>2b>2c. since the ag(i) ion in the complexes does not undergo any electrochemical transformations, the anodic processes can be assigned to ligand-centered ones. the oxidation potential ranges of the ligands are close. furthermore, it was found that the complexes rank below the ligands in reducing ability. table 1. voltammetry data and rates of reduction of fe(iii)-cyt c for the compounds 1a–1c and their ag(i) complexes 2a–2c (anodic polarization). we have carried out a spectrophotometrical investigation of bovine heart fe(iii)-cyt c reduction with 1a–1c and their ag(i) complexes 2a–2c. it was established that among the compounds under study it is the ligands 1b and 1c that are characterized by the highest rate of fe(iii)-cyt c reduction (table 1), and they show a higher reducing ability (determined electrochemically) as compared with that of their ag(i) complexes 2b and 2c. among the ag(i) complexes, it is 2a that demonstrates the highest level of fe(iii)-cyt c reduction rate, and according to electrochemical data, this complex is the most active reducing agent. on the contrary, the complex 2a is a weaker reductant than the respective ligand 1a, but the latter reduces the enzyme slower than 2a. thus, the reduction of fe(iii)-cyt c with the sterically hindered phenolic derivatives and their ag(i) compounds е1pa, v ν0 (nmol min-1) 1a 0.90 0.7 1b 1.04 4.2 1c 1.02 3.9 2a 0.96 1.1 2b 1.26 0.4 2c 1.34 0.6 http://dx.doi.org/10.5599/admet.1167 loginova, gvozdev, osipovich et al. admet & dmpk 10(3) (2022) 197-212 206 complexes cannot be due solely to their capacity for oxidation, depending on other physicochemical properties (ionization, solubility etc.) in a more intricate way. antibacterial activity increased activity of ag(i) complexes compared to silver nitrate is known [2]. this effect can be associated with an increase in compound lipophilicity owing to ag(i) complexation with organic ligands. a microbiological investigation of the newly synthesized compounds 1a–1c and their ag(i) complexes 2a–2c was carried out in vitro for test cultures of gram-positive and gram-negative bacteria using the procedures described in section “antibacterial activity”; the data obtained are given in tables 2 and 3. microbiological tests show that the ag(i) complexes are more active than the parent ligands. the ag(i) complexes suppress the growth of test cultures at moderate concentrations, and the activity of the ag(i) complex 2c exceeds not only the inhibiting action of the ligands but also that of other ag(i) complexes, 2b and 2a (table 2). we have chosen thiol-containing schiff bases 1a–1c as the ligands for improving the stability of the redox-active ag(i) complexes in solution due to the high affinity of ag(i) ions for the thiolate sulphur and the high stability of the ag(i) complexes with organosulphur compounds (k ~ 1013) [46]. however, when evaluating the complexation effect of these ligands with ag(i) ions on their antibacterial properties, one is forced to accept the fact that the newly synthesized ag(i) complexes 2a–2c are characterized by mic values higher than those of the ag(i) complexes 3a and 3b with sterically hindered phenolic derivatives previously reported by us [17,21] (table 3). table 2. antibacterial activity of the compounds 1a–1c and their ag(i) complexes 2a–2c. bacterial strain minimal inhibitory concentration (µmol·ml–1) 1a 1b 1c 2a* 2b* 2c bacillus subtilis >0.274 0.280 0.280 0.106 0.108 0.061 staphylococcus saprophyticus >0.274 0.280 >0.280 0.106 0.108 0.061 proteus vulgaris >0.274 >0.280 >0.280 0.106 0.108 0.061 pantoea agglomerans >0.274 >0.280 >0.280 0.106 0.108 0.061 pseudomonas putida >0.274 >0.280 >0.280 0.106 0.108 0.061 escherichia coli >0.274 0.280 0.280 0.106 0.108 0.061 pectobacterium carotovorum >0.274 >0.280 >0.280 >0.106 >0.108 >0.061 serratia marcescens 0.274 >0.280 >0.280 0.106 0.108 0.061 bacillus pumilus p10 >0.274 >0.280 >0.280 >0.106 >0.108 >0.061 bacillus pumilus b9 >0.274 >0.280 >0.280 0.106 0.108 0.061 *mic was calculated per one monomeric unit based on the literature and our experimental data, several explanations for the effect observed can be offered. first, the complex-forming ag(i) ion is bonded to sulphur atom in coordination cores of the ag(i) complexes and thus, according to [2,12], is not able to take part in exchange reactions with bioligands, interacting with soft bases (nitrogen or sulfur atoms) in the composition of target biomolecules of a microbial cell. second, the ag(i) complexes 3a and 3b are effective antioxidants (unlike 2a–2c) owing to a special state of silver in their molecules resulting from pct complex being formed, and for redox-active admet & dmpk 10(3) (2022) 197-212 silver(i) complexes with schiff bases doi: http://dx.doi.org/10.5599/admet.1167 207 complexes, which are active reductants, the antimicrobial activity can be related to their action on electrontransport systems of a microbial cell [21,26,47]. table 3. antibacterial activity of agnps, the ag(i) complexes 3a, 3b. bacterial strain minimal inhibitory concentration (µmol·ml–1) 3a 3b agnps3a agnps3b stm* tcn* chl*** bacillus subtilis <0.004 0.004 0.007 0.007 0.011 0.014 0.009 staphylococcus saprophyticus <0.004 <0.004 0.007 0.007 0.011 0.014 0.019 serratia marcescens <0.004 <0.004 0.007 0.007 0.011 na 0.019 sarcina lutea <0.004 <0.004 0.007 0.007 0.021 0.014 na pseudomonas aeruginosa <0.004 <0.004 0.007 0.007 0.172 0.056 0.039 escherichia coli <0.004 <0.004 0.007 0.007 0.005 0.007 0.019 salmonella typhimurium <0.004 <0.004 0.007 0.007 0.021 0.014 0.019 staphylococcus aureus <0.004 0.004 0.007 0.007 0.011 0.007 0.019 *streptomycin; **tetracycline; ***chloramphenicol the toxic effect is believed to result generally from the catalytic production of reactive radical species that arise via electron transfer and destroy the cell. third, on carrying out microbiological assays in a liquid nutrient medium, it is necessary to take into account the possibility of silver nanoparticles being formed as a result of the pct complex decomposition [20-23]. the mean size of these particles is 5 nm, and it is known that silver particles of this very size demonstrate the highest antimicrobial activity [25,48,49]. in our earlier investigations, we have shown that the ag(i) complexes 3a and 3b are the agnps precursors. in this paper, we investigated the antibacterial activity of agnps produced by decomposition of the ag(i) complexes 3a and 3b (respectively agnps3a and agnps3b, see section “synthesis of silver nanoparticles”). the data obtained are given in table 3. agnps3a and agnps3b as well as the ag(i) complexes 3a and 3b are characterized by very low mic values, their antimicrobial activity exceeds that of the commonly used antibiotics (table 3). interaction with bsa serum albumins are the most abundant proteins in the circulatory system of mammals, responsible for the maintenance of oncotic pressure and blood ph. they also serve as depot and transport proteins for a variety of endogenous and exogenous substances such as drugs, hormones, fatty acids and metal ions [50]. it is known that the affinity of a drug for plasma proteins influences its distribution, efficacy, free concentration, and metabolism. therefore, investigation of protein-drug interactions is important. recently it was found that bsa interacted with ag(i) ions, forming a protein-silver adduct. quenching the fluorescence of bsa with agnps was also studied [50]. in this work, the distinguishing features of the interaction between ag(i) complexes 2a–2c and bsa were determined using fluorescence quenching experiments. bsa was chosen as a model protein in biochemical studies due to its structural homology with human serum albumin [51]. bsa is composed of three domains (i, ii and iii) divided into two subdomains (a and b), and contains two tryptophan residues, trp-134 and trp-213, which can be used as fluorophores. upon addition of the ag(i) complexes to bsa solution, there was a decrease in fluorescence (fig. 4, a). fluorescence data were analyzed to estimate the binding constants of the ag(i) complexes to bsa according http://dx.doi.org/10.5599/admet.1167 loginova, gvozdev, osipovich et al. admet & dmpk 10(3) (2022) 197-212 208 to a 1:1 model. stern-volmer constant (ksv) was calculated from the slope of the plot of f0/f versus [q] (equation (1)). the apparent binding constant (kb) and hill coefficient (n) were calculated according to equation (2). the data obtained are presented in table 4. in the case of complex 2c there is a positive deviation from the linearity of the stern-volmer plot (fig. 4, b), which occurs primarily either when the extent of quenching is large or due to the combination of the static and dynamic mechanisms of quenching. figure 4. a) fluorescence spectra of bsa without the complex 2c and with it at different concentrations. b) stern-volmer plot for fluorescence quenching of bsa by the complexes 2a–2c. c) double logarithmic plot log[(f0-f)/f] vs log[q]. to evaluate ksv, we considered the range within which the stern-volmer plot was linear. according to the results obtained, complex 2c demonstrated a high affinity for the bsa molecule with logkb = 6.5 m–1, while log kb values of the polymeric complexes 2a and 2b were much lower. the hill coefficient for 2c was greater than unity, indicating positive cooperativity. the n values calculated for 2a and 2b were lower than unity, suggesting negative cooperativity. binding constant decreased in the order: 2c<2b<2a. table 4. stern-volmer constants, apparent binding constants, hill coefficients and logp of the ag(i) complexes-bsa interactions. compound ksv 10-4 (m-1) log kb (m-1) n log p 2a 6.48±0.03 4.08±0.23 0.85 2.5 2b 7.99±0.03 4.59±0.46 0.94 2.8 2c 20.5±0.5 6.53±0.28 1.23 2.9 it also should be noted that the direct correlation between the binding constant and log p was observed in the series of ag(i) complexes. the log kb for these complexes was significantly lower than agnps reported previously [50]. conclusions the novel ag(i) complexes 2a–2c with sterically hindered phenolic schiff bases were synthesized and isolated in the amorphous state. according to the data obtained, they are characterized by distorted geometry of the coordination cores [agn2s2], [agns] and [ags2]. the results of our investigation show that structural modification of the phenolic ligands and complexation with ag(i) ions provides a way of obtaining ag(i) complexes which are stable in organic solvents with high solvating power, being the complexes without pct. the complexes 2a–2c are not typified by the intramolecular redox reaction in organic solvents resulting in the formation of agnps. this property makes them different from the ag(i) complexes of 2-[4,6di(tert-butyl)-2,3-dihydroxyphenylsulfanyl]acetic acid and 4,6-di-tert-butyl-2,3-dihydroxybenzaldehyde isonicotinoyl hydrazone with pct (3a and 3b). the novel ag(i) complexes with schiff bases as well as agnps3a and agnps3b have been screened for their admet & dmpk 10(3) (2022) 197-212 silver(i) complexes with schiff bases doi: http://dx.doi.org/10.5599/admet.1167 209 activity against different species of bacteria. all the complexes are more active than the respective ligands 1a–1c, but significantly less active than the complexes 3a, 3b, agnps3a and agnps3b. this antibacterial effect is believed to result generally from the possibility of the pct complex decomposition on carrying out microbiological assays in liquid nutrients to give agnps. compared with ag(i) complexes, the mechanism for the antimicrobial action of agnps may be similar, but agnps may have much better efficiency owing to their surface area to volume ratio is higher [48,52]. in particular, agnps or ag(i) ions can attack the respiratory chain in bacterial mitochondria and lead to cell death [53]. it was found that the ligands 1b, 1c and the complex 2a with a high reducing ability (determined electrochemically) are characterized by the highest rates of fe(iii)-cyt c reduction among the compounds synthesized. but the level of their antibacterial activity was shown to be independent of their reducing ability (determined electrochemically). thus, further studies are required to elucidate the mechanism of the effects observed because no “one-toone” correspondence between their bioactivity and reducing ability was established. in this paper, we presented the results characterizing the peculiarities of the interaction of the ag(i) complexes with bsa. these results may be useful in understanding how ag(i) compounds convert to toxic forms and will provide a pharmacological basis for novel antimicrobial agents. acknowledgements: the work was carried out within the framework of the task 2.2.01.05 srp "chemical processes, reagents and technologies, bioregulators and bioorgchemistry". the authors are grateful to dr. pavel nikishau for his help in obtaining the esr spectral data. conflict of interest: the authors declare no conflict of interest. references [1] m leeb. antibiotics: a shot in the arm. nature 431 (2004) 892-893. https://doi.org/10.1038/431892a. [2] a.b.g. lansdown, silver in health care: antimicrobial effects and safety in use. royal society of chemistry, cambridge, uk, 2010, p. 262. [3] k. nomiya, h. yokoyama. syntheses, crystal structures and antimicrobial activities of polymeric silver(i) complexes with three amino-acids [aspartic acid (h2asp), glycine (hgly) and asparagine (hasn)]. journal of the chemical society, dalton transactions 12 (2002) 2483-2490. https://doi.org/10.1039/b200684g. [4] k. nomiya, r. noguchi, m. oda. synthesis and crystal structure of coinage metal(i) complexes with tetrazole (htetz) and triphenylphosphine ligands, and their antimicrobial activities. a helical polymer of silver(i) complex [ag(tetz)(pph3)2]n and a monomeric gold(i) complex [au(tetz)(pph3)]. inorganica chimica acta 298 (2000) 24-32. https://doi.org/10.1016/s0020-1693(99)00401-6. [5] s. abuskhuna, j. briody, m. mccann, m. devereux, k. kavanagh. synthesis, structure and anti-fungal activity of dimeric ag(i) complexes containing bis-imidazole ligands. polyhedron 23 (2004) 12491255. https://doi.org/10.1016/j.poly.2004.02.006. [6] b. coyle, p. kinsella, m. mccann, m. devereux, r. o'connor. synthesis, x-ray structure, anti-fungal and anti-cancer activity of [ag(nh3)2(salh)2] (salh2=salicylic acid). journal of inorganic biochemistry 98 (2004) 1361-1366. https://doi.org/10.1016/j.jinorgbio.2004.04.016. [7] k. nomiya, k. tsuda, n.c. kasuga. synthesis and crystal structure of hexanuclear silver(i) cluster [ag(hmna)]6·4h2o(h2mna=2-mercaptonicotinic acid) and a supramolecular gold(i) complex h[(au(hmna)2] in solid state, and their antimicrobial activities. journal of the chemical society, dalton transactions 13 (2000) 2091-2097. https://doi.org/10.1039/b001664k. [8] a. melaiye, k.m. hindi, s. durmus, m.j. panzner, l.a. hogue. formation of water-soluble pincer silver(i)–carbene complexes: a novel antimicrobial agent. journal of medicinal. chemistry 47 (2004) 973-977. https://doi.org/10.1021/jm030262m. http://dx.doi.org/10.5599/admet.1167 https://doi.org/10.1038/431892a https://doi.org/10.1038/431892a https://doi.org/10.1039/b200684g https://doi.org/10.1039/b200684g https://doi.org/10.1016/s0020-1693(99)00401-6 https://doi.org/10.1016/j.poly.2004.02.006 https://doi.org/10.1016/j.jinorgbio.2004.04.016 https://doi.org/10.1039/b001664k https://doi.org/10.1021/jm030262m loginova, gvozdev, osipovich et al. admet & dmpk 10(3) (2022) 197-212 210 [9] k. nomiya, s. takahashi, r. noguchi. synthesis and crystal structure of silver(i) complexes with (s)(+)-5-oxo-2-tetrahydrofurancarboxylic acid (s-hothf) and its isomeric forms showing wide spectra of effective antibacterial and antifungal activities. chiral helical polymers in the solid state formed by self-assembly of the dimeric [ag(othf)]2 cores. journal of the chemical society, dalton transactions 8 (2000) 1343-1348. https://doi.org/10.1039/a910251p. [10] k. nomiya, s. takahashi, r. noguchi, s. nemoto, t. takayama, m. oda. synthesis and characterization of water–soluble silver(i) complexes with l-histidine and (s)-(–)2-pyrrolidone-5carboxylic acid showing a wide spectrum of effective antibacterial and antifungal activities. inorganic chemistry 39 (2000) 3301-3311. https://doi.org/10.1021/ic990526o. [11] s. silver. bacterial silver resistance: molecular biology and uses and misuses of silver compounds. fems microbiology reviews 27 (2003) 341-353. https://doi.org/10.1016/s0168-6445(03)00047-0. [12] k. nomiya, a. yoshizawa, k. tsukagoshi, n.c. kasuga, s. hirakawa, j. watanabe. synthesis and structural characterization of silver(i), aluminium(iii) and cobalt(ii) complexes with 4isopropyltropolone (hinokitiol) showing noteworthy biological activities. action of silver(i)-oxygen bonding complexes on the antimicrobial activities. journal of inorganic biochemistry 98 (2004) 4660. https://doi.org/10.1016/j.jinorgbio.2003.07.002. [13] j.r. ames, m.d. ryan, p. kovacic. mechanism of antibacterial action: electron transfer and oxy radicals. journal of free radicals in biology & medicine 2 (1986) 371-391. https://doi.org/10.1016/s0748-5514(86)80040-x. [14] e. reisner, v.b. arion, b.k. keppler, a.j.l. pombeiro. electron-transfer activated metal-based anticancer drugs. inorganica chimica acta 361 (2008) 1569-1583. https://doi.org/10.1016/j.ica.2006.12.005. [15] c.r. kowol, e. reisner, i. chiorescu, v.b. arion, m. galanski, d.v. deubel, b.k. keppler. an electrochemical study of antineoplastic gallium, iron and ruthenium complexes with redox noninnocent α-n-heterocyclic chalcogensemicarbazones. inorganic chemistry 47 (2008) 1103211047. https://doi.org/10.1021/ic8013249. [16] a.a. chernyavskaya, n.v. loginova, g.i. polozov, o.i. shadyro, a.a. sheryakov, e.v. bondarenko. synthesis and antimicrobial activity of silver(i) and copper(ii) complexes with 2-(4,6-di-tert-butyl2,3-dihydroxyphenylsulfanyl) acetic acid. pharmaceutical chemistry journal 40 (2006) 413-415. https://doi.org/10.1007/s11094-006-0141-4. [17] n.v. loginova, t.v. koval’chuk, a.t. gres, n.p. osipovich, g.i. polozov, y.s. halauko, y.v. faletrov, h.i. harbatsevich, a.v. hlushko, i.i. azarko, y.v. bokshits. redox-active metal complexes of sterically hindered phenolic ligands: antibacterial activity and reduction of cytochrome c. part iv. silver(i) complexes with hydrazone and thiosemicarbazone derivatives of 4,6-di-tert-butyl-2,3dihydroxybenzaldehyde. polyhedron 88 (2015) 125-137. https://doi.org/10.1016/j.poly.2014.12.014. [18] n.v. loginova, a.a. chernyavskaya, g.i. polozov, t.v. koval’chuk, e.v. bondarenko, n.p. osipovich, a.a. sheryakov, o.i. shadyro. silver(i) interaction and complexation with sterically hindered sulfurcontaining diphenol derivatives. polyhedron 24 (2005) 611-618. https://doi.org/10.1016/j.poly.2005.01.007. [19] n.b. kozlova, yu.i. skurlatov. activation of oxygen by metalloenzymes and their models. russian chemical reviews 58 (1989) 138-147. http://dx.doi.org/10.1070/rc1989v058n02abeh003431. [20] n.v. loginova, a.a. chernyavskaya, m.s. parfenova, n.p. osipovich, g.i. polozov, y.a. fedutik, t.v. koval’chuk, g.p. shevchenko. complex of silver (i) with 2-[4,6-di(tert-butyl)-2,3-dihydroxyphenylsulfanyl]acetic acid as a precursor of silver nanoparticles. polyhedron 25 (2006) 1723-1738. https://doi.org/10.1016/j.poly.2005.11.015. [21] n. loginova, a. chernyavskaya, g. polozov, n. osipovich, t. koval’chuk, a. gres, y. halauko, a. halauko, r. zheldakova, y. faletrov, i. azarko. bioactive silver(i) complexes with phenolic derivatives of thioglycolic and thiopropionic acids. mini-reviews in organic chemistry 10 (2013) 227-240. https://doi.org/10.2174/1570193x11310030003. https://doi.org/10.1039/a910251p https://doi.org/10.1021/ic990526o https://doi.org/10.1016/s0168-6445(03)00047-0 https://doi.org/10.1016/j.jinorgbio.2003.07.002 https://doi.org/10.1016/s0748-5514(86)80040-x https://doi.org/10.1016/s0748-5514(86)80040-x https://doi.org/10.1016/j.ica.2006.12.005 https://doi.org/10.1016/j.ica.2006.12.005 https://doi.org/10.1021/ic8013249 https://doi.org/10.1007/s11094-006-0141-4 https://doi.org/10.1016/j.poly.2014.12.014 https://doi.org/10.1016/j.poly.2014.12.014 https://doi.org/10.1016/j.poly.2005.01.007 https://doi.org/10.1016/j.poly.2005.01.007 http://dx.doi.org/10.1070/rc1989v058n02abeh003431 https://doi.org/10.1016/j.poly.2005.11.015 https://doi.org/ https://doi.org/ admet & dmpk 10(3) (2022) 197-212 silver(i) complexes with schiff bases doi: http://dx.doi.org/10.5599/admet.1167 211 [22] a.a. chernyavskaya, n.v. loginova, n.p. osipovich, g.p. shevchenko, m.s. parfenova, t.v. koval’chuk, g.i. polozov. formation of silver nanoparticles from a complex of ag(i) with 2-[4,6di(tert-butyl)-2,3-dihydroxyphenylsulfanyl]acetic acid in organic solvents. colloid journal 68 (2006) 345-349. https://doi.org/10.1134/s1061933x06030136. [23] h.i. harbatsevich, n.v. loginova, t.v. kovalchuk, n.p. osipovich, a.t. gres. synthesis of silver nanoparticles by decomposition of a silver complex with the effect of valence tautomerism and their properties. journal of applied spectroscopy 84 (2017) 13-18. https://doi.org/10.1007/s10812017-0419-1. [24] c.m. da silva, d.l. da silva, l.v. modolo, r.b. alves, m.a. de resende, c.v.b. martins, â. de fátima. schiff bases: a short review of their antimicrobial activities. journal advanced research 2 (2011) 18. https://doi.org/10.1016/j.jare.2010.05.004. [25] j.a. lemire, j.j. harrison, r.j. turner. antimicrobial activity of metals: mechanisms, molecular targets and applications. nature reviews microbiology 11 (2013) 371-384. https://doi.org/10.1038/nrmicro3028. [26] n.v. loginova, t.v. koval’chuk, n.p. osipovich, y.v. faletrov, yu.s. halauko, g.i. polozov, a.t. gres’, h.i. harbatsevich, a.v. hlushko, r.a. zheldakova, v.m. shkumatov in: cytochromes b and c: biochemical properties, biological functions, and electrochemical analysis, r. thom (ed)., nova science publishers, new york, usa, 2014, p 268. [27] w.h. koppenol, k.j. van buuren, j. butler, r. braams. the kinetics of the reduction of cytochrome c by the superoxide anion radical. biochimica et biophysica acta 449 (1976) 157-168. https://doi.org/10.1016/0005-2728(76)90130-4. [28] a. leo, c. hansch, d. elkins. partition coefficients and their uses. chemical reviews 71 (1971) 525616. https://doi.org/10.1021/cr60274a001. [29] k. burger, solvation, ionic and complex formation reactions in non-aqueous solvents: experimental methods for their investigation. elsevier, amsterdam, netherlands, 1983, p. 268. [30] v. lorian (ed.), antibiotics in laboratory medicine, 5th ed. lippincott williams & wilkins, philadelphia, usa, 2005, p.889. [31] j.h. jorgensen m.j. ferraro. antimicrobial susceptibility testing: a review of general principles and contemporary practices. clinical infectious diseases 49 (2009) 1749-17455. https://doi.org/10.1086/647952. [32] j.r. lakowicz. principles of fluorescence spectroscopy, 3rd ed. springer, berlin, 2006. [33] h. t. nagasawa, h.r. gutmann, m.a. morgan. the oxidation of o-aminophenols by cytochrome c and cytochrome oxidase. ii. synthesis and identification of oxidation products. journal of biological chemistry 234 (1959) 1600-1604. https://doi.org/10.1016/s0021-9258(18)70057-4. [34] m.m.m. saleem, m.t. wilson. kinetic studies on the reduction of cytochrome c. reaction with dihydroxy conjugated compounds (catechols and quinols). biochemical journal 201 (1982) 433-444. https://doi.org/10.1042/bj2010433. [35] w.j. geary. the use of conductivity measurements in organic solvents for the characterisation of coordination compounds. coordination chemistry reviews 7 (1971) 81-122. http://dx.doi.org/10.1016/s0010-8545(00)80009-0. [36] k. nakamoto, infrared and raman spectra of inorganic and coordination compounds, part b, applications in coordination, organometallic, and bioinorganic chemistry, 6th ed. john wiley & sons inc., hoboken, new jersey, usa, 2009, p. 416. [37] l.j. bellamy, the infra-red spectra of complex molecules, 3d ed. chapman and hall ltd., london, united kingdom, 1958, p. 433. [38] c.w. lange, c.g. pierpont. nickel complexes containing catecholate, benzoquinone and semiquinone radical ligands. inorganica chimica acta 263 (1997) 219-224. https://doi.org/10.1016/s0020-1693(97)05649-1. http://dx.doi.org/10.5599/admet.1167 https://doi.org/10.1134/s1061933x06030136 https://doi.org/10.1007/s10812-017-0419-1 https://doi.org/10.1007/s10812-017-0419-1 https://doi.org/10.1016/j.jare.2010.05.004 https://doi.org/10.1038/nrmicro3028 https://doi.org/10.1038/nrmicro3028 https://doi.org/10.1016/0005-2728(76)90130-4 https://doi.org/ https://doi.org/ https://doi.org/10.1086/647952 https://doi.org/10.1086/647952 https://doi.org/10.1016/s0021-9258(18)70057-4 https://doi.org/10.1042/bj2010433 http://dx.doi.org/10.1016/s0010-8545(00)80009-0 http://dx.doi.org/10.1016/s0010-8545(00)80009-0 https://doi.org/10.1016/s0020-1693(97)05649-1 https://doi.org/10.1016/s0020-1693(97)05649-1 loginova, gvozdev, osipovich et al. admet & dmpk 10(3) (2022) 197-212 212 [39] r. herzfeld, p. nagy. studies of the solvent effect observed in the absorption spectra of certain types of schiff bases. current organic chemistry 5 (2001) 373-394. https://doi.org/10.2174/1385272013375599. [40] a.v. dmitryuk, s.e. paramzina, a.s. perminov, n.d. solov'eva, n.t. timofeev. the influence of glass composition on the properties of silver-doped radiophotoluminescent phosphate glasses. journal of non-crystalline solids 202 (1996) 173-177. https://doi.org/10.1016/0022-3093(96)00175-5. [41] r.s. mulliken. molecular compounds and their spectra. journal of the american chemical society 74 (1952) 811-824. https://doi.org/10.1021/ja01123a067. [42] f. scholz, electroanalytical methods: guide to experiments and applications, 1st ed. springerverlag, berlin, germany, 2002, p. 331 [43] a.m. pisoschi, c. cimpeanu, g. predoi. electrochemical methods for total antioxidant capacity and its main contributors determination: a review. open chemistry 13 (2015) 824-856. https://doi.org/10.1515/chem-2015-0099. [44] a. kedadra, t. lanez, e. lanez, h. hemmami, m. henni. synthesis and antioxidant activity of six novel n-ferrocenyl-methyl-n-(nitrophenyl)and -n-(cyanophenyl)-acetamides: cyclic voltammetry and molecular docking studies. journal of electrochemical science and engineering 12(2) (2022) 293-304. https://dx.doi.org/10.5599/jese.1162. [45] k.p. butin, e.k. beloglazkina, n.v. zyk. metal complexes with non-innocent ligands. russian chemical reviews 74 (2005) 531-554. https://dx.doi.org/10.1070/rc2005v074n06abeh000977. [46] r.a. bell, j.r. kramer. structural chemistry and geochemistry of silver-sulfur compounds: critical review. environmental toxicology and chemistry 18 (1999) 9-22. https://doi.org/10.1002/etc.5620180103. [47] j.r. ames, m.d. ryan, p. kovacic. mechanism of antibacterial action: electron transfer and oxy radicals. journal of free radicals in biology & medicine 2 (1986) 377-391. https://doi.org/10.1016/s0748-5514(86)80040-x. [48] s. sarkar, a.d. jana, s.k. samanta, g. mostafa. facile synthesis of silver nano particles with highly efficient antimicrobial property. polyhedron 26 (2007) 4419-4428. https://doi.org/10.1016/j.poly.2007.05.056. [49] k.k.y. wong, x. liu. silver nanoparticles – the real ‘‘silver bullet’’ in clinical medicine? medicinal chemistry communications 1 (2010) 125-131. https://doi.org/10.1039/c0md00069h. [50] j. mariam, p.m. dongre, d.c. kothari. study of interaction of silver nanoparticles with bovine serum albumin using fluorescence spectroscopy. journal of fluorescence 21 (2011) 21932199. https://doi.org/10.1007/s10895-011-0922-3. [51] d. carter, x. he, s. munson, p. twigg, k. gernert, m. broom, t. miller. three-dimention structure of human serum albumine. science 244 (1989) 1195-1198. https://doi.org/10.1126/science.2727704. [52] c.n. lok, c.m. ho, r. chen, q.y. he, w.y. yu, h. sun, p.k. tam, j.f. chiu, c.m. che. proteomic analysis of the mode of antibacterial action of silver nanoparticles. journal of proteome research 5 (2006) 916-24. https://doi.org/10.1021/pr0504079. [53] i. sondi, b. salopek-sondi. silver nanoparticles as antimicrobial agent: a case study on e. coli as a model for gram-negative bacteria. journal of colloid and interface science 275 (2004) 177-182. https://doi.org/10.1016/j.jcis.2004.02.012. ©2022 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) https://doi.org/10.2174/1385272013375599 https://doi.org/10.2174/1385272013375599 https://doi.org/10.1016/0022-3093(96)00175-5 https://doi.org/10.1021/ja01123a067 https://doi.org/10.1515/chem-2015-0099 https://dx.doi.org/10.5599/jese.1162 https://dx.doi.org/10.1070/rc2005v074n06abeh000977 https://doi.org/10.1002/etc.5620180103 https://doi.org/10.1002/etc.5620180103 https://doi.org/10.1016/s0748-5514(86)80040-x https://doi.org/10.1016/s0748-5514(86)80040-x https://doi.org/10.1016/j.poly.2007.05.056 https://doi.org/10.1016/j.poly.2007.05.056 https://doi.org/10.1039/c0md00069h https://doi.org/10.1007/s10895-011-0922-3 https://doi.org/10.1126/science.2727704 https://doi.org/10.1021/pr0504079 https://doi.org/10.1016/j.jcis.2004.02.012 http://creativecommons.org/licenses/by/3.0/ manuscript doi: 10.5599/admet.3.3.217 154 admet & dmpk 3(3) (2015) 154; doi: 10.5599/admet.3.3.217 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index editorial research in neurodegenerative diseases: challenges and solutions kin tam editor: admet & dmpk e-mail: kin_tam@iapchem.org the need of effective medicines for the treatment of neurodegenerative diseases, such as alzheimer’s disease and parkinson’s disease, is expected to increase strongly in the coming decades. though great efforts have been paid on research, neurodegenerative diseases remain as urgent unresolved problems in modern day medicine. at present, there is no cure for these diseases, resulting in death of neuron cells. generally, the disease mechanisms are not fully understood and are thought to be related to genetics, protein misfolding/degradation or mitochondrial dysfunction etc. this introduces further complexity in the search for therapeutic approaches for these diseases. in most neurodegenerative diseases, the blood–brain barrier (bbb) becomes a hurdle in effective drug delivery. small molecules or biologics need to be engineered to cross the bbb and reach their target in sufficient concentrations to elicit a pharmacodynamics response. the drug concentration in the brain is tricky to determine, while behavioral assessment methods are used as surrogates for pharmacodynamics effects. in this context, it could be challenging to establish an unambiguous pharmacokinetics/pharmacodynamics relationship. in this special issue, some papers reviewed the mechanisms, treatment approaches, animal models used for alzheimer’s disease. a paper reported a trend model for alzheimer’s disease mortality. moreover, two review papers discussed bbb permeability and delivery in traumatic brain injury and in insulin delivery for alzheimer’s disease, respectively. other papers dealt with cognitive deficits. finally, two other papers discussed the use of nanotechnology platform in parkinson’s disease and the interactions of prion protein with metal ions. for the first time, admet & dmpk devotes a special issue on medical research/applications in a particular therapeutic area, with the aims to capture some of the current views and new achievements in neurodegenerative diseases. it is nice to see some papers discussing cross-disciplinary research and new thoughts to tackle these diseases. hopefully, we will call for more special issues aligned to other therapeutic areas in due course. http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:kin_tam@iapchem.org anomalous salting-out, self-association and pka effects in the practically-insoluble bromothymol blue doi: https://doi.org/10.5599/admet.1822 1 admet & dmpk 0(0) (2023) 000-000; doi: https://doi.org/10.5599/admet.1822 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index communication anomalous salting-out, self-association and pka effects in the practically-insoluble bromothymol blue alex avdeef in-adme research, new york, ny 10128 usa e-mail: alex@in-adme.com; tel.: +1-646-678-5713 received: april 17, 2023; revised: may 17, 2023; published: may 23, 2023 abstract background and purpose: the widely-used and practically insoluble diprotic acidic dye, bromothymol blue (btb), is a neutral molecule in strongly acidic aqueous solutions. the schill (1964) extensive solubility-ph measurement of bromothymol blue in 0.1 and 1.0 m nacl solutions, with ph adjusted with hcl from 0.0 to 5.4, featured several unusual findings. the data suggest that the difference in solubility of the neutral-form molecule in 1m nacl is more than 0.7 log unit lower than the solubility in pure water. this could be considered as uncharacteristically high for a salting-out effect. also, the study reported two apparent values of pka1, 1.48 and 1.00, in 0.1 m and 1.0 m nacl solutions, respectively. the only other measured value found for pka1 in the literature is -0.66 (gupta and cadwallader, 1968). experimental approach: it was reasoned that the there can be only a single pka1 for btb. also, it was hypothesized that salting-out alone might not account for such a large difference in solubility observed at the two levels of salt. a generalized mass action approach incorporating activity corrections for charged species using the stokes-robinson hydration equation and for neutral species using the setschenow equation, was selected to analyze the schill solubilityph data to seek a rationalization of these unusual results. key results: btb reveals complex speciation chemistry in saturated aqueous solutions which had been poorly understood for many years. the appear ance of two different values of pka1 at different levels of nacl and the anomalously high value of the empirical salting-out constant could be rationalized to normal values by invoking the formation of a very stable neutral dimer (log k2 = 10.0 ± 0.1 m-1). a ‘normal’ salting-out constant, 0.25 m-1 was then derived. it was also possible to estimate the ‘self-interaction’ constant. the data analysis in the present study critically depended on the pka1 = -0.66 reported by gupta and cadwallader. conclusion: a more reasonable saltingout constant and a consistent single value for pka1 have been determined by considering a self-interacting (aggregation) model involving an uncharged form of the molecule, which is likely a zwitterion, as suggested by literature spectrophotometric studies. ©2023 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords salting-out effect; self-interaction; aggregation; solubility-ph; intrinsic solubility; ionization constant; dimerization constant introduction the intrinsic (i.e., neutral form) solubility of an ionizable substance in water generally decreases when a ‘water-structure maker’ salt (e.g., nacl) is added to the solution. the phenomenon is known as ‘salting-out’ [1-5]. the decrease in solubility seldom exceeds 0.3 log unit per molar salt. such changes in magnitude are https://doi.org/10.5599/admet.1822 https://doi.org/10.5599/admet.1822 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:alex@in-adme.com http://creativecommons.org/licenses/by/4.0/ alex avdeef admet & dmpk 00(0) (2023) 000-000 2 comparable to those seen between different polymorphs of a drug substance [6] and in solubility differences between racemates and individual isomers of an optically-active drug substance [7-9]. in a simple view of the salting-out phenomenon, ions introduced from the salt (e.g., na+, cl–) become strongly hydrated, thus tying up a significant portion of the available water. the added neutral organic solute, effectively deprived of the water molecules associated with the salt ions, dissolves to a lesser extent in terms of the total solution volume, thus appearing to be less soluble. many studies of the salting effect had focused on relatively simple organic nonelectrolytes, including gases and water-immiscible organic solvents [2-4,10-12]. the more soluble the nonelectrolyte, the larger is the salting effect [11]. however, there appear to be very few reported salting studies of practically-insoluble (e.g., intrinsic solubility, s0 < 10-5 m) nonelectrolytes, particularly of drug-like molecules, or substances of similar complexity. in this communication, we re-examine the saturation solution-ph behavior of bromothymol blue (btb, fig. 1), which is practically insoluble in its zero net-charge form in strongly acidic solutions (left-most structure in fig. 1). the molecule may be used as an indicator (yellow for ph <6, blue for ph >7.6, and green in between) to follow ph-dependent cellular processes. it can function as a complexing agent, and has been used in textiles, paints, cleaning products, detergents, and photovoltaic cells [13]. in cases of prolonged general exposure or if absorbed through the skin, btb can be harmful. because of its low solubility and high stability, btb can be a persistent environmental pollutant [13]. btb is a diprotic acid (h2a), with ionization reactions shown in figure 1. in strongly acidic solution (ph <1), the molecule is plausibly a zwitterion forming a red solution. in mildly acidic/neutral and alkaline solutions, btb exists as the anionic species ha– and a2–. the second ionization constant (pka2) has the reported value of 7.12 at ionic strength, i = 0.1 m [14]. in contrast, the value of pka1 has been unresolved for many years, with values reported as low as -0.66 [15] and as high as 1.48 [14]. there has also been a long-standing controversy about the structure of the uncharged species in aqueous solution (ph <1). figure 1. ionization equilibria for the diprotic weak acid bromothymol blue in water. although the value of pka2 is confidently established, the value of pka1 has been elusive. schill [14] extensively studied the solubility-ph behavior of btb in 0.1 m and 1.0 m nacl solutions at 20 °c. in a series of publications, schill applied btb as a sensitive ion-pair extraction reagent to determine the concentrations of amines and quaternary ammonium compounds by spectrophotometry. the reported btb solubility-ph data, along with the originally-determined constants are shown in figure 2. we found three points of interest in schill’s study, which appeared in need of additional examination. • the study indicated surprising departures from the expected salting behavior of neutral electrolytes. the difference between the apparent intrinsic solubility in pure water and in solutions containing 1.0 m nacl was reported to be greater than 0.7 log unit for btb, which is inexplicably high [1-5]. admet & dmpk 00(0) (2023) 000-000 polyvinyl alcohol nanoparticles loaded with propolis extract doi: https://doi.org/10.5599/admet.1822 3 • also, two significantly different values of the first ionization constant were reported by schill in saturated solutions: pka1 = 1.48 in 0.1 m nacl and pka1 = 1.00 in 1.0 m nacl (cf. fig. 2). the stokes-robinson hydration equation [16] (a) includes the debye-hückel expression (ion-ion electrostatic interactions), (b) incorporates decrease in activity of water (work done in immobilizing some of the bulk water to hydrate ions), and (c) accounts for the free energy change of ions (as their concentrations increase when the volume of bulk water decreases upon hydration of ions). from the stokes-robinson hydration equation, the difference between the apparent pka1 values is expected to be near 0.02, far less than the above 0.48 implied value [14]. • schill’s examination of the uv spectra of btb in sub-saturated solutions suggested that (ah–)2 and (ah–)4 aggregates might be forming in the ph region where the btb anion prevails. they reported association constants from analysis of the spectroscopic data in the absence of visible signs of precipitation. in saturated solutions in the same region, micelles appeared to form. the reported limiting slope in figure 2 in the ph 1-3 region is +1, which is not compatible with the aggregation models proposed by schill. if aggregation comprised solely of the ha– species, the slopes in the acid region would have been +2 (dimer) or +4 (tetramer), rather than +1 [17]. if mixed-charge aggregates (h2a.ha–)n were to form, the slope would be +1 [17], but such models cannot resolve the unusual salting-out magnitude reported by schill. so, the stoichiometry of the observed aggregation may need to be further examined. ph figure 2. bromothymol blue solubility-ph profiles reported by schill [14], with data collected at 20 oc, after 24 h equilibration time. the upper saturated solutions profile contained 0.1 m nacl, while the lower profile contained 1.0 m nacl. the ph electrode was calibrated by schill for the concentration scale, using solutions of known [h+] concentrations. the various listed equilibrium constants are those reported by schill. in the present communication, an attempt is made to rationalize the apparently excessive salting behavior, the lack of consistency regarding the value of pka1, and the formation of aggregates by applying a mass action model, using the program pdisol-xtm (in-adme research) [16-26], to analyze the saturation solubility-ph data published by schill [14]. it was hypothesized that the critical aggregation, which could explain some of the above-mentioned anomalies in acidic solutions (ph <3), is due to the self-association of the zero net-charge species, h2a. method mass action equilibrium model consider a diprotic acid, h2a, with the dissociation constants pka1 and pka2. let’s assume it is prone to form zero net-charge dimers, (h2a)2, in aqueous solutions. the equilibrium reactions of relevance in saturated solutions (ph <6) are h2a  ha– + h+ ka1 = [ha–][h+]/[h2a] (1a) lo g ( s / m ) https://doi.org/10.5599/admet.1822 alex avdeef admet & dmpk 00(0) (2023) 000-000 4 ha–  a2– + h+ ka2 = [a2–][h+]/[ha–] (1b) 2h2a  h4a2 k2 = [h4a2]/[h2a]2 (1c) h2a(s)  h2a s0 = [h2a] (1d) na.ha(s)  na+ + ha– ksp = [na+][ha–] (1e) where s0 is the intrinsic solubility of the uncharged (monomeric free acid) form of the molecule. based on the above relations, the total concentration of the weak acid in the saturated aqueous phase, atotaq, is the total solubility, st, which can be expressed solely in terms of equilibrium constants and ph: atotaq = st = [h2a] + [ha–] + [a2–] + 2 [h4a2] (2a) = [h2a] + [h2a] ka1 / [h+] + [h2a] ka1 ka2 / [h+]2 + 2 k2 [h2a]2 (2b) = s0 (1 + ka1 / [h+] + ka1 ka2 / [h+]2 + 2 s0 k2) (2c) in logarithmic form, base 10, log st = log s0 + log (1 + 10–pka1 + ph + 10–pka1 –pka2 + 2ph + 210–ps0 +log k2 ) (3) in the thermodynamically valid constant ionic medium (cim) activity scale [17], the reference ionic strength may be set to iref > 0. in such a framework, all equilibrium constants are expressed in terms of concentrations. in the multiple-ph solubility data analysis, all constants are adjusted for local deviations from iref, using the stokes-robinson hydration equation [16,17]. in the case of uncharged species, similar adjustments for local deviations from iref are made using the salting out equation approach described below. in the cim framework, the operational pah scale (ph meter readings) is standardized to the concentration scale, pch, using the four-parameter equation [27] pah = α + k pch + jh [h+] + joh kw / [h+] (4) where kw is the ionization constant of water [28]. the jh term corrects pah readings for the nonlinear ph electrode response due to liquid junction and asymmetry potentials in highly acidic solutions (ph <2), while the joh term corrects for high-ph nonlinear effects [17]. since schill standardized the electrode to read on the concentration scale and since all saturation data were for ph <6, it was assumed here that α = joh = 0 and k = 1. the jh was adjusted during the regression analysis since many ph values were <1 in the 1 m nacl set (fig. 2). in the analysis of the schill data, pka2 was selected as 7.12. (since most of the data are in the acidic range, pka2 is of minor role here.) the rest of the constants in eq. (3) were determined by nonlinear regression analysis (i.e. mass action model), the details of which are described elsewhere [16,17]. the data were first evaluated as schill had done, and essentially the same constants were determined as those listed in figure 2. for those constants, self-interaction (aggregation) was not included (as per schill’s assumption). salting activity model for uncharged species the setschenow equation [1,2,10-15] describes the salting effect on nonelectrolyte (i.e. h2a) total aqueous solubility, st, as log (st(0) / st(s)) = kcs (5) where k is the empirical setschenow coefficient (m−1) and cs is the concentration (m) of the added salt. st(0) and st(s) are the neutral solute total solubility values in pure water and in water containing salt, respectively. k is positive in the case of salting-out and negative in the case of salting-in processes. generally, the equilibrium constants in eq. (5) can represent a solute partitioning process between different phases. if self-interaction (aggregation) is hypothesized, then a modified form of the setschenow equation may need to be invoked [2,4,11]: admet & dmpk 00(0) (2023) 000-000 polyvinyl alcohol nanoparticles loaded with propolis extract doi: https://doi.org/10.5599/admet.1822 5 log (st(0) / st(s) ) = kscs + ki (st(s) – st(0)) (6) where ks and ki are nonelectrolyte salting and self-interaction parameters, respectively. only for low st or in the absence of nonelectrolyte self-interaction will the empirical setschenow constant, k, be equal to the salting parameter, ks. usually ks is determined from oil-water partition data using small volumes of water-immiscible oil, where aggregates would not be expected to enter the oil phase [11]. eq. (5), cast in the form of partition coefficients, may be used to determine ks. the concentration of the monomer is measured in the oil phase and the corresponding value in the water phase is calculated from mass balance to determine the partition coefficient. once ks is known, then ki can be calculated form eq. (6). this is the approach used by al-maaieh and flanagan [11] to determine the self-interaction parameter for caffeine, theophylline, and theobromine (e.g., ki = -2.06 m-1 for caffeine in 1 m na2so4 solution, indicating less aggregation with increasing salt concentration [11]). a different approach was used presently: ks was determined by substituting into eq. (5) the intrinsic (monomer) solubility, s0 (cf. eq. (1d)), determined by regression analysis. in eq. (5), k was taken to be ks; st(0) and st(s) were substituted with s0(0) *and s0(s), respectively. once ks was so determined, then ki was calculated from eq. (6). abraham linear free energy descriptors used to predict the salting-out constant of btb abraham’s five linear free energy solvation descriptors (a, b, sπ, e, v) have been used to describe the distribution of nonelectrolytes between two phases [29,30]. a is the h-bond acidity and b is the h-bond basicity of the solute. sπ is the dipolarity/polarizability, e is an excess molar refractivity in units of (cm3/mol)/10, and v is the mcgowan characteristic molar volume in units of (cm3/mol)/100. endo et al. [4] used the abraham model to predict salting-out constants (m-1): ks = a0 + a1a + a2b + a3sπ + a4e + a5v (7) the a0-a5 coefficients in eq. (7) were determined by multi-linear regression (mlr): 0.112, -0.047, -0.060, -0.042, -0.020, and 0.171, respectively [4]. evidently, large molecules increase the salting-out behavior, and polar molecules act in the opposite direction. presently, the a-coefficients were re-determined by partial least squares (pls) regression (open-source package from https://cran.r-project.org/web/packages/pls), using 142 ks values as the training set (136 values compiled/measured by endo et al. [4] and six derived here from the data of furia et al. [12]) to predict the ks value of bromothymol blue. values of the abraham descriptors for bromothymol blue were calculated from its 2d structure using the absolv algorithm [30] (cf. www.acdlabs.com). results and discussion of data re-analysis the nature of the structure of the btb free acid species in saturated solutions is still controversial and has been discussed for many decades. several different monomeric structures for the h2a species may exist/coexist (two zwitterion tautomers and a closed sultone ring uncharged form) in saturated solutions below ph 3, each with its own associated ‘micro-constant’ for pka1. an electrostatically-bound dimer based on two zwitterions oriented with opposite-charge groups at contact distances may be feasible. other structures are possible for a btb dimer. a further complication is that the structural form of the btb free acid in the crystal lattice may undergo tautomeric re-arrangement upon dissolution, posing a challenge to the canonical definition of intrinsic solubility, since the form of the substance would be different between the solid state and the solution phases. definitive support for structures of btb species in saturated aqueous solutions is sketchy and is mostly inferred from spectrophotometric measurements in sub-saturated solutions. https://doi.org/10.5599/admet.1822 https://cran.r-project.org/web/packages/pls http://www.acdlabs.com/ alex avdeef admet & dmpk 00(0) (2023) 000-000 6 the speciation analysis of solubility-ph data is based on thermodynamic principles, formulated in terms of ‘macro-constants.’ structural questions can only be addressed indirectly. in this communication, three different model strategies were hypothesized and tested against schill’s solubility-ph data. model a. dimer-free, single pka1 model, based on an unusually-high salting-out factor our first strategy was to be open to the possibility of an unusually high k, to assume that (i) dimers did not form in the ph <3 region, and (ii) one of schill’s pka values was more reliable (but both values could not be simultaneously ‘correct’). we selected (and refined) pka1 = 1.43 ±0.09 (0.1m nacl) since its value would be least affected by the salting-out effect of bromothymol blue. a fit of the two apparent intrinsic solubility values to cs (eq. (5)), yielded intercept salt-free value of ps0 = 5.10 and the slope factor k = 0.778. the refinement of the model in the case of 0.1 m nacl (keeping pka1 fixed at 1.43) yielded at iref = 0.1 m: ps0 = 5.17±0.02 (using the high k to correct for the activity of the net zero charge h2a species) and pksp = 4.41±0.03 (goodness-of-fit, gof = 0.50). however, when the above model was applied to the 1.0 m nacl case, it was not feasible to adhere to the above assumption (i) at the high salt level. it was not possible to fit the schill data without invoking the mixedcharge dimer, h2a.ha–, since such a species would account for the apparently lower pka1 in 1 m nacl solutions reported by schill, while maintaining the slope in the 1-3 ph region at +1 [17]. also, it was necessary to invoke two salt species, with the new addition being na.h2a.ha(s), needed to explain the data in the ph 2.5-3.5 region. in the refinement of the data, pka1 and ps0 were kept fixed at 1.43 and 5.17 (since the 0.1 m nacl data defined these values). the refined constants, also with reference to iref = 0.1 m, were log kh3a2 = 5.87±0.03, pkps (nah3a2) = 4.93 ±0.02, and pksp(naha) = 4.33 ±0.02 (gof = 0.35) for the case of 1.0 m nacl. model b. fitting a single pka1 to schill’s data, assuming a neutral dimer forms in our next strategy, it was hypothesized that there can only be a single pka1 for bromothymol blue and its value would not likely be either 1.48 or 1.00, as schill had reported. we hypothesized that a neutral dimer formed. it was reasoned that this caused the st near ph 0 to be different in the 0.1 and 1.0 m nacl solutions. this was hypothesized to be the explanation for the two values of the reported pka1. since there appeared to be less self-association in the 1 m nacl solutions, the new value for the pka1 was sought in that medium, with iref was set to 1.0 m. starting with the schill pka1 = 1.00 as a fixed constant, the ps0, log k2, and pksp constants were determined iteratively by nonlinear regression. next, these three refined values were fixed, and pka1 was subjected to regression. its value decreased steadily. it was then fixed, and the process was repeated with the first three constants. after several cycles of this ‘block-diagonalization’ refinement, the minimum errors in all four constants were reached, and the overall goodness-of-fit (gof) [17] settled at its minimum value. the excessive correlations between the pka1 and the other three constants did not permit their simultaneous determination, so the ‘block-diagonalization’ approach was used. the pka1 so determined (+0.217) was then applied to the 0.1 m nacl set. the same iterative process led to convergence. it soon became evident that a range of reasonable pka1 values could be proposed and the fitting of the schill data (fig. 2) would be equally good (as indicated by the minimum gof reached). since the correlation between pka1 and the other constants was extreme, the ‘block-diagonalization’ minimum lay over a very shallow well, and normal experimental errors in solubility measurement caused havoc to pin down the ‘best’ pka1 value. model c. applying an independently-measured single pka1 to schill’s data, assuming a neutral dimer forms attention was then directed to finding an independently determined value of pka1 for the further analysis of the schill solubility data. the only value found was reported by gupta and cadwallader [15]: pka1 = -0.66. the authors collected uv-visible spectroscopic data in the ph interval from 0.92 to -1.08 (concentration admet & dmpk 00(0) (2023) 000-000 polyvinyl alcohol nanoparticles loaded with propolis extract doi: https://doi.org/10.5599/admet.1822 7 scale), where the ph of distilled water was adjusted using 12 m hcl. at -log [h+] = -0.662, the concentration of the anion was found to be equal to the concentration of the uncharged species. this defines the value of the pka1 at ionic strength 10+0.662 = 4.59 m. the stokes-robinson hydration equation was then used to harmonize this value to the cim activity scale (iref = 1.0 m), to obtain pka1 = -1.18, the value ultimately to be used as a fixed contribution in the regression analysis of the schiff solubility data. model selection model a described above, with the single pka1 = 1.43 (although it could not support a dimer-free case for both 0.1 and 1.0 m nacl media), is consistent with the ‘red’ species in figure 1 as shedding a proton from a protonated carbonyl group in the cyclohexadienone zwitterion. also, model a supports a more soluble ‘red’ zwitterion with ps0 = 5.17, compared to 8.02 in model c (table 1). however, pka1, derived from schill’s data, depends on two assumptions that could not be fully supported since it was necessary to invoke the presence of a mixed dimer, h2a.ha–, as well as a salt based on the dimer, in 1.0 m nacl solutions. the latter two species could not be fit to the data in 0.1 m solutions. model a was also dependent on an unusually high value of the empirical constant, k = 0.778. model b can be dismissed since a specific value of pka1 could not be determined definitively from the solubility data by the ‘block diagonalization’ refinement procedure. model c is attractive since the pka1 value was derived from an independent study using uv/vis data taken in sub-saturated solutions [15]. the so-determined value, adjusted to iref = 1.0 m is -1.18, a much lower value than either of the two apparent values reported by schill. it may represent the shedding of the proton from a protonated sulfonic acid group in an open sultone ring form of btb, possibly present in the sub-saturated solution used by gupta and cadwallader [15]. however, structural issues cannot be directly addressed by the thermodynamic speciation analysis presented here. it was decided in this communication to focus further discussion on model c. also, the setschenow constant based on model c is much more in line with values taken from the literature for many molecules. mass action equilibrium model the results of the final re-analysis of schill’s two sets of saturation solubility-ph data (in 0.1 and 1.0 m nacl media), employing the gupta-cadwallader [15] pka1 (transformed to the cim activity scale), are listed in table 1 and depicted in figure 3. table 1. refined constants bromothymol blue, 20 °c, 0.1 and 1.0 m nacl media (iref = 1.0 m) [nacl] / m pka1 ps0 log k2 pksp iavg / ma nb gofc 0.10 -1.18 d 8.03 ±0.03 10.11 ±0.10 4.41 ±0.02 0.11 10 0.48 1.00 -1.18 d 8.02 ±0.01 9.94 ±0.02 4.31 ±0.01 1.07 15 0.18 aaverage ionic strength. to minimize dilution effects, 12 m hcl titrant was used in the refinement model. bnumber of ph points. cgof = goodness-of-fit [16], based on assigned individual log s errors of 0.1 log unit. dpka1 = -0.66 at i = 4.59 m from gupta and cadwallader [15] was transformed to iref = 1.0 m here, using the stokes-robinson hydration equation. dependence of equilibrium constants on ıonic strength (stokes-robinson hydration equation corrections) when the ionic strength for a given point in a log s-ph titration is calculated to stray from the designated iref value (1.0 m here), it is beneficial to adjust the equilibrium constants in eqs. (1a), (1b), and (1e). the debyehückel equation is often used for this adjustment. however, the latter equation becomes less accurate as ionic strength exceeds about 0.3 m [17]. since the gupta-cadwallader [15] pka1 was measured at i = 4.59 m, it was not expected that the simple debye-hückel equation would be suitable in correcting for activity changes. https://doi.org/10.5599/admet.1822 alex avdeef admet & dmpk 00(0) (2023) 000-000 8 (a) (b) ph ph figure 3. summary of the re-analysis of bromothymol blue solubility-ph data reported by schill [14], incorporating the pka1 reported by gupta and cadwallader [15] (adjusted to iref used here). the points in the green solid curves were calculated after correcting for salting-out effects of the uncharged saturated species. the curves decrease with ph <0 since salt concentrations are elevated due to addition of titrant. ionic strength, m ionic strength, m figure 4. (a) dependence of pka1 and pka2 on ionic strength, according to the stokes-robinson hydration model. (b) dependence of the negative logarithm of the solubility product, ksp = [na+][ha–] on ionic strength. a more comprehensive correction scheme, which may still be useful for i < 5 m, involves the application of the stokes-robinson hydration theory [16,17]. the scheme is coded into pdisol-x. figure 4 shows how the ionization constants and the solubility product depend on ionic strength in the stokes-robinson model. the gupta-cadwallader reported pka1 = -0.66 (unfilled square in fig. 4a) was adjusted to -1.18 at iref = 1.0 m (solid red circle in the upper curve in fig. 4a). the solubility product, ksp (cf., eq. (1e) and fig. 4b), has a less steep ionic strength dependence, compared to those of pka1 and pka2. since ionic strength is not defined for uncharged species, the stokes-robinson scheme is not directly applicable to the intrinsic solubility (eq. (1c)) and dimerization of uncharged species (eq. (1d)) equilibrium constants. however, the next section addresses how salt levels can affect the activity of uncharged species. salting model the application of eq. (5) to the total solubility values in figure 2 resulted in the linear relationship, pst(s) = 5.10 + 0.778 cs, where the slope factor is k = 0.778. figure 5a shows the equivalent of eq. (5) based on intrinsic (monomer) solubility values. the slope factor listed in the figure corresponds to the true saltingout parameter, ks = 0.250. when this salting-out value is substituted into eq. (6) for the case of cs = 1.0 m, lo g ( s / m ) lo g ( s / m ) p k a 2 p k sp p k a 1 admet & dmpk 00(0) (2023) 000-000 polyvinyl alcohol nanoparticles loaded with propolis extract doi: https://doi.org/10.5599/admet.1822 9 ki = (pst(s) – pst(0) – kscs) / (st(s) – st(0)) = (5.865 – 5.094 – 0.250)/(1.364·10-6 – 8.056·10-6) = -7.787·104 (7) its magnitude exceeds that of any other reported ki values, as far as we could find. once ks and ki were determined, eq. (6) can be used to calculate the intrinsic and dimerization equilibrium constants of btb over a range of salt concentrations. this is evident in the green solid curves in figure 3 for ph values below zero, where titrant additions to achieve very low ph increase the salt content in the saturated solutions. cs / m (of added nacl) cs / m (of added nacl) figure 5. (a) salting-out and (b) dimerization effects in bromothymol blue saturated solutions, as a function of the amount of added salt, nacl. these are the apparent refined constants when ks was set to zero. two salt points preclude the possibility to test nonlinear relationships, as seen in cases of substituted phenolic acids covering a wide range of salt concentrations [12]. distribution of aqueous-phase species in saturated solutions of bromothymol blue the equilibrium model developed using schill’s solubility-ph data and gupta-cadwallader pka1 (table 1) allows for the calculation of the distribution of species at the two levels of salts considered, as illustrated in figure 6. it is evident that the monomer concentration (green dash-dot curve in fig. 6) is dwarfed by that of the dimer (red dash-dot-dot curve in fig. 6). pch pch figure 6. distribution of species in (a) 0.1 m nacl and (b) 1.0 m nacl containing solutions, based on the constants in table 1; pka1 = -1.18 (iref = 1.0 m). p s 0 (s ) lo g k 2 (s ) https://doi.org/10.5599/admet.1822 alex avdeef admet & dmpk 00(0) (2023) 000-000 10 in the 0.1 m nacl medium at ph 1.28 (fig. 6a), 50% of the total bromothymol blue is in the monoanionic form and half is essentially in the dimeric form. so, in figure 2, the schill-proposed ionization constant, 1.48, appears to be indicating the equilibrium between anionic ha– and the dimeric (h2a)2 species (and not the monomer). similarly, in the 1.0 m nacl medium at ph 0.60 (fig. 6b), 50 % of btb is in the monoanionic form and half is essentially in the dimeric form. the schill-proposed ionization constant, 1.00, again appears to be indicating the equilibrium between anionic ha– and the dimeric (h2a)2 species. at the half-point ph1/2, [ha–] ≈ [h4a2]. from eq. (2c), the estimate of pka1 ≈ -log 2 + pch + ps0 – log k2. for the 0.1 m and 1.0 m nacl cases, the resulting approximate pka1 values are -1.3 and -1.4, respectively, compared to the -1.18 used (fig. 3). the slight differences may indicate the shortcomings in attempts to standardize the ph electrode in such low ph regions, as per eq. (4), given the uncertainty over junction potentials and the uncertainty in adjusting the pka1 = -0.66 (at 4.59 m ionic strength) from gupta and cadwallader [15] to the selected reference ionic strength of 1.0 m. comparison of bromothymol blue (btb) to thymol blue (tb) equilibrium speciation the ‘flexible‑acceptor’ gse consensus model [31] predicts the intrinsic solubility values of thymol blue (tb) and btb as 6.67 and 7.68, respectively. given the similarity of structures, it might be anticipated that tb and btb would possess similar equilibrium reactions. shimada et al. [32] analyzed the spectral changes (300 to 650 nm) of tb solutions in the ph -0.02 to 4.35 interval. solutions of tb gradually change from red to yellow as ph is raised in the interval studied. one prominent and two minor isosbestic points were evident in the spectra. using principal components analysis (pca) and alternative least-square (als) regression, the authors concluded that two major species accounted for the tb equilibria in the strong acid solutions. the apparent pka1 of tb was determined to be 1.54 in 0.1 m and 1.45 at 1.0 m ionic strength solutions [32]. the red species was most plausibly the zwitterionic form of tb, possibly a carbocation or its resonance equivalent, like the left structure in figure 1 of btb. the colorless uncharged sultone closed ring form was not expected to be stable in the polar medium of water (although it is present in the crystal structure reported by yamaguchi et al. [33]). since the bromine atoms in btb are electron withdrawing substituents, the pka1 value of btb was suggested (without reference) to be -1.3 (compared to +1.6 of bt), and the red form was anticipated to be dominant in highly acidic aqueous medium [32]. from the rigorous analysis of spectrophotometric, potentiometric, and conductimetric data at 25 oc of tb in 1.1 m ionic strength (nacl) aqueous solutions, balderas-hernández et al. [34] reported pka2 = 8.90 and the constants for the reactions h2a + ha–  h3a2– kh3a2 = [h3a2 –]/[h2a][ha–] = 10+11.48 (8) 2h2a  h4a2 kh4a2 = [h4a2]/[h2a] 2 = 10+11.41 (9) the above two constants were originally reported on the cumulative ‘stability’ basis. the conversions to the step-wise basis [17] in eqs. (8) and (9) employed pka1 = 1.45 for tb from shimada et al. [32]. the latter constant taken with those from eqs. (8) and (9) suggest that the proton dissociation constant of the h4a2 dimeric species has a dimeric ionization constant of 1.38, as indicated in eq. (10). h4a2  h3a2– + h+ ka” = [h3a2–][h+]/[h4a2] = 10-1.38 (10) hence, the anionic dimer in tb is expected to be dominant above ph 1.38, while the zero net-charge dimer is expected to prevail below ph 1.38. attempts to incorporate the above model into the schill data for btb were not successful. given the btb pka1 = -1.18, schill’s solubility-ph data did not support the inclusion of the h3a2– species in both the 0.1 and 1.0 m nacl solutions, as discussed earlier in the context of model a. it is quite remarkable that the uncharged dimeric species from the balderas-hernández et al. [34] tb study is so close in value to that reported here for the btb. the predicted tb intrinsic solubility is an order of admet & dmpk 00(0) (2023) 000-000 polyvinyl alcohol nanoparticles loaded with propolis extract doi: https://doi.org/10.5599/admet.1822 11 magnitude higher than that of btb. it is reasonable to view the tb and btb system as being similar, with ionizations shifted to lower values of ph in the case of btb because of the bromine substituents. prediction of ks using the abraham model and literature values of salting-out constants the 142 ks values mentioned earlier (including the provided abraham descriptors) [4,12] were used to re-determine the abraham coefficients in eq. (7) by pls regression. the a0-a5 coefficients were determined as 0.090, -0.073, -0.064, -0.039, -0.002, +0.188, respectively. the regression coefficients are comparable to those reported by endo et al. [4]. highly-polar molecules are predicted to have weak salting-out effects (a1a4 factors). on the other hand, large molecules are expected to be highly sensitive to salting-out. the absolvcalculated values of the abraham descriptors for bromothymol blue (left-most structure in fig. 1) are a = 0.33, b = 1.28, sπ = 2.48, e = 2.94, and v = 3.99. the polar terms contribute 21 % (negative) to the predicted effect and the volume term contributes 75 % (positive). using the coefficients reported by endo et al. [4], predicted ks = 0.54 for btb. using the pls-derived coefficients here, the predicted ks = 0.63. the two predictions are substantially higher than the ks = 0.25 value determined from the data of schill [14] using model c. it may be that the training set of 142 ks values did not cover the chemical space of molecules like bromothymol blue or molecules that form very stable water-soluble oligomers. limitation of the bromothymol blue equilibrium model as discussed elsewhere [9, 17-23], it is not possible to determine the exact degree of aggregation of neutral species solely from the solubility-ph data (case 1a in fig. 6.6 in ref. [17]). it is necessary to use other stoichiometry-sensitive methods, such as esi-q-tof-ms/ms [25]. the simplest stoichiometry (dimer) was modeled here, but the actual species may be a tetramer, a higher-order/mixture of (h2a)n oligomers [14,15]. any of such species can fit the solubility data equally well for low sample concentrations. the ph1/2 values in figure 6 would remain unchanged for low sample concentrations. the pka1 value is not easy to measure independently since its value is negative, and aggregation is a confounding complication. standardization of the ph electrode is likely to be a challenge. the measurement of the ionization constant in cosolvent mixtures (e.g., methanol-water) is expected to increase the pka1 value, perhaps making the determination more reliable. potentiometric titrations, which can simultaneously determine both the solubility and the pka can be effective [17, 35]. additional measurements of the saturation solubility of bromothymol blue for ph >6 might possibly support some of the suggestions proposed by schill regarding the nature of self-association he studied by uv-vis spectrophotometry in sub-saturated solutions. conclusion bromothymol blue reveals complex speciation chemistry in saturated aqueous solutions, which is influenced strongly by the presence of high concentrations of nacl. schill’s [14] extensive saturation solubility-ph measurements (ph from 0 to 5 in 0.1 and 1.0 m nacl solutions) of the practically-insoluble diprotic weak acid, bromothymol blue (6 ng/ml monomer solubility), was re-analyzed and harmonized using a general mass action approach. a ‘normal’ salting-out constant, ks = +0.25 m-1 was derived. the formation of self-aggregates (here treated as dimers, with zero-salt log k2 = 10.0 ± 0.1 m-1) in strongly acidic solutions raised the total solubility of bromothymol blue nearly 500-fold to zero-salt log st = -5.10. the data analysis in the present study critically depended on the pka1 = -0.66 (at 4.59 m ionic strength) reported by gupta and cadwallader [15]. it was possible to estimate the self-interaction constant, ki = -7.787·10+4 m-1, an unusually high value, reflecting the tendency of bromothymol blue to form oligomers in strongly acidic solutions. https://doi.org/10.5599/admet.1822 alex avdeef admet & dmpk 00(0) (2023) 000-000 12 acknowledgements: we are grateful to drs. abu serajuddin (st. john’s univ., nyc) and tatjana verbić, miloš pešić, olivera marković, dušan veljković (university of belgrade, serbia), and kin tam (university of macau) for the critical discussions regarding determinations of ionization constants of practically-insoluble molecules which can self-aggregate (e.g., clofazimine) in saturated solutions [35]. also, i thank the reviewer for a stimulating discussion of the structural nature of the btb free acid, and for suggesting a new reference. references [1] j.z. setschenow. über die konstitution der salzlosungen auf grund ihres verhaltens zu kohlensaure. z. physik. chem. 4 (1889) 117-125. https://doi.org/10.1515/zpch-1889-0409. [2] f.a. long, w.f. mcdevit. activity coefficients of nonelectrolyte solutes in aqueous salt solutions. chem. rev. 51 (1952) 119-169. https://doi.org/10.1021/cr60158a004. [3] n. ni, m.m. el-sayed, t. sanghvi, s.h. yalkowsky. estimation of the effect of nacl on the solubility of organic compounds in aqueous solutions. j. pharm. sci. 89 (2000) 1620-1625. https://doi.org/10.1002/1520-6017(200012)89:12. [4] s. endo, a. pfennigsdorff, k.-u. goss. salting-out effect in aqueous nacl solutions: trends with size & polarity of solute molecules. environ. sci. technol. 46 (2012) 1496-1503. https://doi.org/10.1021/ es203183z. [5] s.h. lubbad, b.k.a. al-roos, f.s. kodeh. adsorptive–removal of bromothymol blue as acidic–dye probe from water solution using latvian sphagnum peat moss: thermodynamic assessment, kinetic & isotherm modeling. curr. green chem. 6 (2019) 53-61. https://doi.org/10.2174/2452273203666 190114144546. [6] m. pudipeddi, a.t.m. serajuddin. trends in solubility of polymorphs. j. pharm. sci. 94 (2005) 929-939. https://doi.org/10.1002/jps.20302. [7] p. friberger, g. åberg. some physicochemical properties of the racemates and optically active isomers of two local anaesthetic compounds. acta pharm. suec. 8 (1971) 361-364. [8] s.k. el-arini, d. giron, h. leuenberger. solubility properties of racemic praziquantel and its enantiomers. pharm. dev. tech. 3 (1998) 557-564. https://doi.org/10.3109/10837459809028638. [9] a. avdeef, e. fuguet, a. llinas, c. ràfols, e. bosch, g. völgyi, t. verbić, e. boldyreva, k. takács-novák. equilibrium solubility measurement of ionizable drugs − consensus recommendations for improving data quality. admet dmpk 4 (2016) 117−178. https://doi.org/10.5599/admet.4.2.292. [10] w.h. xie, w.y. shiu, d. mackay. a review of the effect of salts on the solubility of organic compounds in seawater. mar. environ. res. 44 (1997) 429-444. [11] a. al-maaieh, d.r. flanagan. salt effects on caffeine solubility, distribution, & self-association. j. pharm. sci. 91 (2002) 1000-1008. https://doi.org/10.1002/jps.10046. [12] e. furia, a. beneduci, l. malacaria, a. fazio, c. la torre, p. plastina. modeling the solubility of phenolic acids in aqueous media at 37 oc. molecules 26 (2021) 6500. https://doi.org/10.3390/molecules262 16500. [13] s.h. lubbad, b.k.a. al–roos, f.s. kodeh. adsorptive–removal of bromothymol blue as acidic–dye probe from water solution using latvian sphagnum peat moss: thermodynamic assessment, kinetic & isotherm modeling. curr. green chem. 6 (2019) 53-61. https://doi.org/10.2174/24522732036661 90114144546. [14] g. schill. photometric determination of amines and quaternary ammonium compounds with bromothymol blue. part 2. association of bromothymol blue in aqueous solutions. acta pharm. suec. 1 (1964) 101-122. [15] v.d. gupta, d.e. cadwallader. determination of first pka’ value & partition coefficients of bromothymol blue. j. pharm. sci. 57 (1968) 2140-2142. https://doi.org/10.1002/jps.2600571224. [16] g. völgyi, a. marosi, k. takács-novák, a. avdeef. salt solubility products of diprenorphine hydrochloride, codeine and lidocaine hydrochlorides and phosphates – novel method of data analysis https://doi.org/10.1515/zpch-1889-0409 https://doi.org/10.1021/cr60158a004 https://doi.org/10.1002/1520-6017(200012)89:12 https://doi.org/10.1021/es203183z https://doi.org/10.1021/es203183z https://doi.org/10.2174/2452273203666190114144546 https://doi.org/10.2174/2452273203666190114144546 https://doi.org/10.1002/jps.20302 https://doi.org/10.3109/10837459809028638 https://doi.org/10.5599/admet.4.2.292 https://doi.org/10.1002/jps.10046 https://doi.org/10.3390/molecules26216500 https://doi.org/10.3390/molecules26216500 https://doi.org/10.2174/2452273203666190114144546 https://doi.org/10.2174/2452273203666190114144546 https://doi.org/10.1002/jps.2600571224 admet & dmpk 00(0) (2023) 000-000 polyvinyl alcohol nanoparticles loaded with propolis extract doi: https://doi.org/10.5599/admet.1822 13 not dependent on explicit solubility equations. admet dmpk 1 (2013) 48-62. https://doi.org/ 10.5599/admet.1.4.24. [17] a. avdeef. absorption and drug development, 2nd ed., john wiley & sons, inc., hoboken, nj, 2012. isbn 978-1-118-05745-2. [18] a. avdeef. anomalous solubility behavior of several acidic drugs. admet dmpk 2 (2014) 33-42. https://doi.org/10.5599/admet.2.1.30. [19] a. avdeef. phosphate precipitates and water-soluble aggregates in re-examined solubility-ph data of twenty-five basic drugs. admet dmpk 2 (2014) 43-55. https://doi.org/10.5599/admet.2.1.31. [20] a. avdeef. suggested improvements for measurement of equilibrium solubility-ph of ionizable drugs. admet dmpk 3 (2015) 84-109. https://doi.org/10.5599/admet.3.2.193. [21] g. butcher, j. comer, a. avdeef. pka-critical interpretations of solubility–ph profiles: pg-300995 and nsc-639829 case studies. admet dmpk 3 (2015) 131-140. https://doi.org/10.5599/admet.3.2.182. [22] a. pobudkowska, c. ràfols, x. subirats, e. bosch, a. avdeef. phenothiazines solution complexity – determination of pka and solubility-ph profiles exhibiting sub-micellar aggregation at 25 and 37 oc. eur. j. pharm. sci. 93 (2016) 163-176. https://doi.org/10.1016/j.ejps.2016.07.013. [23] c.a.s. bergström, a. avdeef. perspectives in solubility measurement and interpretation. admet dmpk 7 (2019) 88-105. http://dx.doi.org/10.5599/admet.686. [24] o.s. marković, m.p. pešić, a.v. shah, a.t.m. serajuddin, t.ž. verbić, a. avdeef. solubility-ph profile of desipramine hydrochloride in saline phosphate buffer: enhanced solubility due to drug-buffer aggregates. eur. j. pharm. sci. 133 (2019) 264–274. https://doi:10.1016/j.ejps.2019.03.014. [25] e. fuguet, x. subirats, c. ràfols, e. bosch, a. avdeef. ionizable drug self-associations and the solubility dependence on ph: detection of aggregates in saturated solutions using mass spectrometry (esi-qtof-ms/ms). mol. pharmaceutics 18 (2021) 2311-2321. https://doi.org/10.1021/acs.molpharma ceut.1c00131. [26] o.s. marković, n.g. patel, a.t.m. serajuddin, a. avdeef, t. ž. verbić. nortriptyline hydrochloride solubility-ph profiles in a saline phosphate buffer: drug-phosphate complexes and multiple phmax domains with a gibbs phase rule "soft" constraints. mol. pharmaceutics 19 (2022) 710-719. https://doi.org/10.1021/acs.molpharmaceut.1c00919. [27] a. avdeef, j.j. bucher. accurate measurements of the concentration of hydrogen ions with a glass electrode: calibrations using the prideaux and other universal buffer solutions and a computer-controlled automatic titrator. anal. chem. 50 (1978) 2137-2142. https://doi.org/10.1021/ac50036a045. [28] f.h. sweeton, r.e. mesmer, c.f. baes, jr. acidity measurements at elevated temperatures. 7. dissociation of water. j. solut. chem. 3 (1974) 191-214. https://doi.org/10.1007/bf00645633. [29] m.h. abraham, j. le. the correlation and prediction of the solubility of compounds in water using an amended solvation energy relationship. j. pharm. sci. 88 (1999) 868-880. https://doi.org/10.1021/ js9901007. [30] j.a. platts, d. butina, m.h. abraham, a. hersey. estimation of molecular linear free energy relation descriptors using a group contribution approach. j. chem. inf. comput. sci. 39 (1999) 835-845. https://doi.org/10.1021/ci980339t. [31] a. avdeef, m. kansy. predicting solubility of newly-approved drugs (2016–2020) with a simple absolv and gse (flexible-acceptor) consensus model outperforming random forest regression. j. solution chem. 51 (2022) 1020-1055. https://doi.org/10.1007/s10953-022-01141-7. [32] t. shimada, k. tochinai, t. hasegawa. determination of ph dependent structure of thymol blue revealed by cooperative analytical method of quantum chemistry and multivariate analysis of electronic absorption spectra. bull. chem. soc. jpn. 92 (2019) 1759-1766. https://doi.org/10.1246/ bcsj.20190118. [33] k. yamaguchi, z. tamura, m. maeda. (1997). molecular structure of bromophenol blue having a γsultone ring. anal. sci. 13 (1997) 1057-1058. https://doi.org/10.2116/analsci.13.1057. https://doi.org/10.5599/admet.1822 https://doi.org/10.5599/admet.1.4.24 https://doi.org/10.5599/admet.1.4.24 https://doi.org/10.5599/admet.2.1.30 https://doi.org/10.5599/admet.2.1.31 https://doi.org/10.5599/admet.3.2.193 https://doi.org/10.5599/admet.3.2.182 https://doi.org/10.1016/j.ejps.2016.07.013 http://dx.doi.org/10.5599/admet.686 https://doi:10.1016/j.ejps.2019.03.014 https://doi.org/10.1021/acs.molpharmaceut.1c00131 https://doi.org/10.1021/acs.molpharmaceut.1c00131 https://doi.org/10.1021/acs.molpharmaceut.1c00919 https://doi.org/10.1021/ac50036a045 https://doi.org/10.1007/bf00645633 https://doi.org/10.1021/js9901007 https://doi.org/10.1021/js9901007 https://doi.org/10.1021/ci980339t https://doi.org/10.1007/s10953-022-01141-7 https://doi.org/10.1246/bcsj.20190118 https://doi.org/10.1246/bcsj.20190118 https://doi.org/10.2116/analsci.13.1057 alex avdeef admet & dmpk 00(0) (2023) 000-000 14 [34] p. balderas-hernández, m.t. ramírez, a. rojas-hernández, a. gutiérez. determination of pka’s for thymol blue in aqueous medium: evidence of dimer formation. talanta 46 (1998) 1439-1452. https://doi.org/10.1016/s0039-9140(98)00015-0. [35] t.ž. verbić, k.y. tam, d.ž. veljković, a.t.m. serajuddin, a. avdeef. clofazimine pka determination by potentiometry and spectrophotometry – reverse cosolvent dependence as an indicator of presence of dimers in aqueous solutions. mol. pharmaceutics (2023) https://doi.org/10.1021/acs.molphar maceut.3c00172. ©2023 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) https://doi.org/10.1016/s0039-9140(98)00015-0 https://doi.org/10.1021/acs.molpharmaceut.3c00172 https://doi.org/10.1021/acs.molpharmaceut.3c00172 http://creativecommons.org/licenses/by/3.0/ challenges, current status and emerging strategies in the development of rapidly dissolving fdm 3d-printed tablets: an overview and commentary doi: https://doi.org/10.5599/admet.1622 33 admet & dmpk 11(1) (2023) 33-55; doi: https://doi.org/10.5599/admet.1622 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index review challenges, current status and emerging strategies in the development of rapidly dissolving fdm 3d-printed tablets: an overview and commentary abu t.m. serajuddin department of pharmaceutical sciences, college of pharmacy and health sciences, st. john’s university, 8000 utopia parkway, queens, ny 11439, usa; serajuda@stjohns.edu received: november 27, 2022; revised: december 13, 2022; published: january 01, 2023 abstract since the approval of a 3d-printed tablet by the fda in 2015 for marketing, there has been a great interest in 3d printing in the pharmaceutical field for the development of personalized and on-demand medications. among various 3d printing methods explored for the development of oral solid dosage form like tablet, the fused deposition modeling (fdm) 3d-printing, where the drug-polymer mixtures are first converted into filaments by hot melt extrusion (hme) and then the filaments are printed into tablets using 3d printers by applying computer-aided design principles, has emerged as the most attractive option. however, no fdm 3d-printed tablets have yet been marketed as the technology faces many challenges, such as limited availability of pharmaceutical-grade polymers that can be printed into tablets, low drug-polymer miscibility, the need for high temperature for hme and 3d-printing, and slow drug release rates from tablets. these challenges are discussed in this article with a special focus on drug release rates since fdm 3d-printing usually leads to the preparation of slow-release tablets while the rapid release from dosage forms is often desired for optimal therapeutic outcomes of new drug candidates. pros and cons of various strategies for the development of rapidly dissolving fdm 3d-printed tablets reported in the literature are reviewed. finally, two case studies on emerging strategies for the development of rapidly dissolving fdm 3d-printed tablets are presented, where one outlines a systematic approach for formulating rapidly dissolving tablets, and the other describes a novel strategy to increase dissolution rates of drugs from fdm 3d-printed tablets, which at the same time can also increase drug-polymer miscibility and printability of tablets and lower processing temperatures. thus, this overview and commentary discusses various issues involving the formulation of rapidly dissolving fdm 3d-printed tablets and provides guidance for the development of commercially viable products. ©2023 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords 3d printing; fused deposition modeling; hot melt extrusion; 3d-printed tablet; drug release; rapid dissolution introduction three-dimensional (3d) printing, where materials are deposited layer by layer to form solid objects, has gained much interest in the pharmaceutical field as it can deliver personalized and on-demand medications, such as tablets, to patients for better therapy [1]. 3d-printing is essentially a form of additive manufacturing (am) that is extensively used in many different industries [2]. the interest in 3d printing in the pharmaceutical https://doi.org/10.5599/admet.1622 https://doi.org/10.5599/admet.1622 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:serajuda@stjohns.edu http://creativecommons.org/licenses/by/4.0/ abu t.m. serajuddin admet & dmpk 11(1) (2023) 33-55 34 field has been most intense since the approval of a 3d-printed tablet, spiritam®, by the us food and drug administration in 2015 [3], which is evident from the exponential increase in the number of publications on 3d printing of pharmaceutical products in recent years [4, 5]. various opportunities provided by 3d printing technology in the pharmaceutical field and different challenges currently facing its application in the development of drug products have been extensively reviewed in the literature [6-12]. several different 3d-printing techniques, such as binder jet printing, fused deposition modeling (fdm), semisolid extrusion (sse), selective laser sintering (sls), and stereolithography (sla), have been employed for the manufacture of tablets [1,11]. among these techniques, fused deposition modeling (fdm) 3d-printing is the most popular and widely used for the development of solid dosage forms like tablets and pellets. in this method, drug-loaded filaments of thermoplastic polymers are produced by hot melt extrusion (hme) and the filaments are then printed into tablets using 3d printers by applying computer-aided design (cad) principles [13,14]. hot melt extruders used for the preparation of filaments are now commonly available in academia and the pharmaceutical industry for many different applications and 3d printers are relatively cheaply available in the market for the printing of filaments into tablets, both of which contributed to the rapid growth of the fdm 3d-printing technology. it is anticipated that, in the future, the filaments can be mass-produced in the pharmaceutical industry or at various pharmaceutical distribution sites and then sent to clinics, retail pharmacies, hospitals, nursing homes, etc., for printing into individualized medications as per the needs of the patients. such personalized therapy can mitigate some of the variability in patients’ responses to medications observed due to differences in their genetic compositions, gender, age, body weight, disease state, and so forth. the fdm 3d printing will be especially suitable for clinical testing of new drug candidates where the dose is often unknown at the outset of clinical studies and may need to be changed or modified depending on clinical observations on patients. under such circumstances, clinical studies may be accelerated as new doses may be printed using the same filaments without having to wait for the development of new formulations. additionally, fdm 3d printing has the potential to provide multiple drugs in the same tablet by printing different drugs layer by layer, side by side or in different compartments [15,16]. the doses may also be individualized in such polypills. objectives of this report despite its early promises and relative simplicity, no products based on the fdm 3d-printing technology have yet been approved by regulatory agencies and marketed by pharmaceutical companies. this is because fdm 3d-printing is a relatively new technology that faces many developmental and technical challenges. the objectives of the present overview and commentary are to critically analyze these challenges, discuss what progress has been made towards resolving them, and present emerging strategies that may lead to the resolution of most of these problems leading to the development of clinically and commercially viable fdm 3d-printed tablets. one special focus will be on the development of rapidly dissolving tablets since fdm 3dprinting mostly produces tablets with slow and incomplete drug release, while tablets with rapid drug release rates are often needed for optimal clinical efficacy in humans. the importance of rapid drug release for the successful therapeutic outcome of new drug candidates will be discussed based on the biopharmaceutical risk assessment roadmap (bioram) developed by an international panel of experts [17,18], and the pros and cons of several strategies for the development of rapidly dissolving 3d-printed tablets reported in the literature will be reviewed. finally, two case studies will be presented where one of them will provide a systematic approach for the development of fdm 3d-printed tablets for rapid drug release based on the work by solanki et al. [19], and the other will provide, based on two recent publications by patel and serajuddin [20,21], a multi-pronged strategy to not only increase drug release rates from printed tablets but admet & dmpk 11(1) (2023) 33-55 development of rapidly dissolving fdm 3d-printed tablets doi: https://doi.org/10.5599/admet.1622 35 at the same time to also enhance drug-polymer miscibility, increase printability of polymers, and reduce processing temperatures of filaments and tablets. challenges in the development of fdm 3d-printed tablets as outlined by solanki et al. [19] and patel and serajuddin [20,21], the fdm 3d-printing technology faces several major challenges that must be addressed before it can progress from the current scientific exploratory phase to a mature technology for the development and manufacture of viable and commercially successful drug products. some of these challenges are: a) limited availability of pharmaceutical-grade polymers that are suitable for fdm 3d-printing, b) low drug drug-polymer miscibility, c) the need for high temperatures for melt extrusion and fdm 3d-printing where drug, polymer or both may degrade, and d) slow and incomplete drug release from tablets. the above challenges in the development of fdm 3d-printed tablets are discussed below using fig. 1, where different steps involved in the preparation of fdm 3d-printed tablets are given in fig. 1a, while fig. 1b shows the printing process on an expanded scale: figure 1. (a) illustration of melt extrusion of filaments and 3d-printing of tablets during the fdm 3d-printing process. (b) printing gear, nozzle, and printing plate are shown with an expanded scale. adapted from ilyes et al. [22] with permission. limited availability of pharmaceutical-grade polymers as shown in fig. 1a, the extrusion of drug-polymer mixtures through a hot melt extruder into filaments is the first major step in the fdm 3d printing process. for this purpose, it is essential that the polymers used are extrudable. however, melt extrusion is a relatively new application of polymers in the pharmaceutical industry, and most polymers currently used for the formulation of conventional tablets were originally developed and introduced by their manufacturers for other pharmaceutical applications like diluent, binder, https://doi.org/10.5599/admet.1622 abu t.m. serajuddin admet & dmpk 11(1) (2023) 33-55 36 disintegrant, drug release modifier, coating agent, etc. as such, they may not have suitable attributes, such as glass transition temperature (tg), melt viscosity, thermal stability, etc., for hme. even when certain polymers and drug-polymer mixtures are extrudable, the filaments produced may not be printable. various issues with the feeding of filaments into 3d printers and then printing as tablets have been reported in the literature [23,24]. as shown in fig. 1b, the filaments are pushed through the printing nozzles by using gears, where they may break if they are too brittle or may not pass through the nozzle at all if they are too soft. therefore, the optimal flexibility of a filament is essential for its printing into a tablet. additionally, once the filaments are fed into the printer using the gear, they are further heated to attain certain optimum melt viscosities such that the molten materials are streamed through the printer head and deposited on the building plate layer by layer to form tablets [25]. the layers must coalesce with each other, and they must also harden rapidly enough that the surface, shape, and internal structures remain intact as the tablets are built by depositing one layer over the other until the printing is completed [26]. all these attributes of filaments and tablets often depend on the properties of the polymers used. for this reason, henry et al. [27] cautioned that despite much academic research and many proof-of-concept studies, a more thorough understanding of polymer properties and the interplay between polymers and processing parameters are needed before the fdm 3d-printing technique can be fully implemented in the development of personalized medicine. currently, the availability of pharmaceutical-grade polymers that are extrudable and printable remains a major limitation in the progress of fdm 3d printing technology. lack of drug-polymer miscibility ensuring drug-polymer miscibility is another major challenge in the development of fdm 3d-printed tablets. although it may not be a major issue for water-soluble drugs since they may dissolve rapidly even if they remain crystalline and phase-separated from polymers, one must consider that approximately 3 out of 4 chemical entities under development as drug candidates are poorly water-soluble and may fall into class ii or iv of the biopharmaceutical classification system (bcs) [28]. this is reflected in the types of drugs used for 3d printing. in a survey of literature for 37 drugs used for fdm 3d printing, pereira et al. (29) observed that 28 (74 %) belonged to bcs class ii or iv, indicating possible solubility and dissolution issues. for the rapid dissolution of such drugs, it is essential that the polymers used for fdm 3d-printing are water-soluble and the drugs are molecularly or amorphously dispersed in polymers during hme and remain such in the filaments as well as in the tablets [30]. the tablets should also be physically stable during their shelf-lives such that no crystallization of drugs occurs. therefore, drug-polymer miscibility is critically important for the development of fdm 3d-printed tablets, especially for drugs with poor aqueous solubility. the drug-polymer miscibility will also dictate how much drug can be loaded in the filaments and then in the tablets. however, in most reported studies on fdm 3d-printing in the literature, not much consideration of drug-polymer miscibility is made, and it has been reported that the poorly water-soluble drugs may remain either partially or fully crystalline in the tablets [31]. in one of the few studies where drug-polymer miscibility was determined for the development of fdm 3d-printed tablets, solanki et al. [19] tested the miscibility of haloperidol with polyvinyl pyrrolidone-vinyl acetate copolymer (kollidon® va64), hydroxypropyl methylcellulose (affinisol® 15cp) and the 1:1 mixture of the two polymers by applying the film casting technique previously developed by parikh et al. [32]. they observed that haloperidol was miscible with kollidon® va64, affinisol® 15cp and the 1:1 kollidon-affinisol mixture to the extents of approximately 20, 10 and 15 percent, respectively. in another study, wei et al. [33] observed that the miscibility of haloperidol with another polymer, polyvinyl alcohol (pva), was <10 % w/w, while a different drug, carvedilol, was miscible with pva up to 20 % w/w. these reports demonstrated that the miscibility of drugs with polymers could be rather limited, and the extent of drug-polymer miscibility may depend on the properties of both drugs and polymers used. it may admet & dmpk 11(1) (2023) 33-55 development of rapidly dissolving fdm 3d-printed tablets doi: https://doi.org/10.5599/admet.1622 37 also be noted that if the drug-polymer miscibility and, thereby, the drug loading is low, the size of printed tablets may become too large for certain doses. need for high processing temperature the hot melt extrusion (hme) temperature to prepare filaments for fdm 3d-printing should be sufficiently high that the polymers must exist in the molten state above their glass transition temperatures, and their melt viscosities should be within a certain range such that they can be extruded into filaments [19,20,24,34]. although the presence of the drug may reduce the melt viscosity and thus the extrusion temperature by the plasticization of polymers [34,35], the melt extrusion temperatures in most of the reported studies in the literature are still high. when the filaments are passed through the printer, as mentioned earlier, the temperature within the printer nozzle must be even higher to further soften the filaments such that they can be deposited layer by layer into tablets. by surveying 58 fdm 3d-printed tablets reported in the literature, cailleaux et al. [26] reported that filaments for 50 of them were melt extruded at temperatures ranging from 100 to 210 °c, where temperatures of 140-180 °c were common. among these tablets, 47 were printed in the temperature range of 150 to 250 °c, where temperatures around 200°c were the most used. thus, high temperatures required for the hme of filaments and even higher temperatures needed for 3d printing could be major issues for the thermal stability of drugs and polymers. at such high melt extrusion and printing temperatures, drugs, polymers, or both may be thermally unstable. it is essential that the processing temperatures are lowered to make fdm 3d-printing more widely applicable to drug product development. drug release rate lastly, the most difficult challenge in the development of fdm 3d-printed tablets is to ensure rapid drug release from the tablets. the drug-loaded filaments produced by hme are essentially nonporous masses, which are then printed into compact and hard tablets. therefore, the tablets dissolve in aqueous media by erosion from the surface without any disintegration into particles, which often leads to slow drug release rate. indeed, cailleaux et al. [26] reported that out of 56 fdm 3d-printed drug products for which drug release rates were reported in the literature, 43 were slow-release dosage forms, 4 were both sustained and immediate release dosage forms, and 9 were for immediate release. however, as will be discussed in the next section, the rapid drug release rate is critically important for most drugs for optimal therapy [17,18]. therefore, as mentioned earlier, the primary focus of the present commentary is to assess the status of the development of rapidly dissolving fdm 3d-printed tablets. the significance of rapid drug release in drug therapy will also be discussed, which will then be followed by the discussion of various strategies employed to increase drug release rates from fdm 3d-printed tablets. why rapid drug release? the therapeutic outcome of drug products and optimal patient benefits depend greatly on how drugs are released from dosage forms and reach the target sites after oral administration and how the drug levels are maintained in the body. therefore, it is essential that the optimal drug release attributes must be built into dosage forms. in 2014, an international panel of experts representing the food and drug administration (fda), academia, and the pharmaceutical industry from the us and europe, headed by dr. arzu selen from the fda, developed a biopharmaceutical risk assessment roadmap (bioram) to delineate optimal drug delivery rates and patterns for achieving the desired therapeutic outcomes [18]. the team also described different biopharmaceutics tools that may be applied to identify and address potential challenges to optimize drug products for patient benefits [17]. according to bioram, blood levels of different drugs in the body may be https://doi.org/10.5599/admet.1622 abu t.m. serajuddin admet & dmpk 11(1) (2023) 33-55 38 classified into four scenarios, as shown in figure 2, to achieve the desired therapeutic outcomes: figure 2. four scenarios of drug product development according to the biopharmaceutical roadmap (bioram). adapted from selen et al. [17] with permission. a) scenario 1 represents drugs where very rapid drug absorption and therapeutic action are needed for such therapies as breakthrough pain, analgesia, acute angina pectoris, insomnia, etc. b) scenario 2 represents multiphasic delivery of a single drug or multiple drugs from the same dosage form that has both rapid action and sustained action, and this scenario may include drug products for attention deficit hyperactivity disorder (adhd), combination histamines, antimigraine drugs, drugs for sleep disorders where the rapid onset must be combined with sustained sleep, and so forth. c) scenario 3 represents drug products where the drugs must be rapidly released and absorbed at a predetermined time and then the exposures are maintained, and this scenario may include drug products for the maintenance of blood glucose levels to match the physiological effects of food intake, certain oncology drugs targeted for tumor growth, antihypertensive agents targeted to address circadian relief of blood pressure, etc. d) scenario 4 includes drug products where the targeted exposure in the body is maintained by repeated dosing, and the drug must be absorbed rapidly enough that its concentration after each dose reaches the therapeutic level. many oral antibiotics, antihypertensives, antiepileptics, anti-alzheimer’s disease drugs, chronic pain medications, antispasmodics, etc., belong to this scenario. it is evident from the above-mentioned scenarios that drugs must be rapidly released from products that fall under scenarios 1 and 2, and drugs must also be released rapidly under scenario 3 after a certain initial lag time. even when a drug product is developed for prolonged or sustained release, as depicted by scenario 4, there is usually a rapid-release aspect of a drug product so that the therapeutic blood levels are attained rapidly after initial or repeated dosing. thus, whatever is the scenario of drug product development, a rapidrelease attribute must be built into the product. since most fdm 3d-printed tablets result in slow-release products, further development of the technology to achieve rapid drug release is needed before it can be applied for the development of therapeutically effective drug products according to different bioram scenarios. admet & dmpk 11(1) (2023) 33-55 development of rapidly dissolving fdm 3d-printed tablets doi: https://doi.org/10.5599/admet.1622 39 current strategies for the development of rapid-release fdm 3d-printed tablets several strategies for the development of rapidly dissolving fdm 3d-printed tablets have been reported in the literature. they include selection of rapidly dissolving polymers with optimum attributes for 3d printing, use of polymer-polymer mixtures, incorporation of additives, modification of internal patterns of tablets and tablet geometry, and supersolubilization of a drug by acid-base interaction. selection of rapidly dissolving polymers the most convenient approach to increase the dissolution rates of drugs from the fdm 3d-printed tablets could be the use of highly water-soluble and rapidly dissolving polymers. however, as mentioned earlier, there is only a limited number of polymers available that are water-soluble and printable. polyvinylpyrrolidone (pvp) and related polymers, which are highly water-soluble, may not be used for fdm 3d printing as they produce brittle filaments that are not printable [19-21]. funk et al. [36] reviewed the literature to determine the effects of polymers on drug release rates from 3d-printed tablets and observed that among various polymers, only eudragit® e, which is a polymethacrylate polymer, could possibly provide the most rapid drug release rate. the authors observed that, among other polymer types, celluloses and polyvinyl alcohols (pva) usually provide slow drug release rates from the fdm 3d-printed tablets, and alternative 3d-printing techniques like semisolid extrusion (sse), selective laser sintering (sls), digital light processing (dlp), etc., are necessary to increase dissolution rates of drugs from such polymers. the reported high dissolution rate of drugs from tablets containing eudragit® e and its impact on drug absorption after oral administration, however, need further scrutiny before it is selected for fdm 3d-printing. eudragit® e (orally administered grade known as eudragit® epo) is a cationic polymer comprised of dimethyl aminoethyl methacrylate, butyl methacrylate and methyl methacrylate at a ratio of 2:1:1 in the polymeric chain [37]. because of the protonation of the basic amino group in the structure, it is soluble at low acidic ph and insoluble at ph>5. several investigators reported relatively rapid drug release (<1 h) when the dissolution testing of the fdm 3d-printed tablets containing eudragit® e was conducted in 0.1 m hcl (ph~1.1), and they did not study the dissolution of the tablets at higher intestinal ph conditions (ph>5), where the polymer is insoluble [38-40]. the dissolution testing was often conducted in 0.1 m hcl, even for the amorphous solid dispersion formulations of drugs in eudragit® e reported in the literature [41,42]. in a few cases where the dissolution was studied at both lower gastric and higher intestinal ph conditions, dissolution rates and extents of drug release at the higher ph of 6.8 were much lower than that at the lower ph [43,44]. considering that the average ph of the gastric fluid is around 2 in the fasted state, which may increase to >5 in the fed state [45], the ph-dependent solubility of eudragit® e may have a major impact on drug release from the fdm 3d-printed tablets after oral administration. the shorter gastric residence time in the fasted state (~30 min) could also be a factor in drug dissolution and absorption since the tablet may pass rapidly through the stomach before all the drugs could be released, and then it may not dissolve in the intestine since the polymer is not soluble at the higher ph. careful consideration must also be given to the printability of filaments containing eudragit® e. since the polymer has a low glass transition temperature (52 °c), the filaments could be very soft at the relatively high printing temperatures used. in most of the reported studies, high amounts of solid drugs and solid additives were added to the polymer to reduce its fluidity after melting. for example, yang et al. [46] added 37.5-50 % talc to eudragit® e-based formulations to make the filaments harder and printable. use of polymer mixtures another approach to increase the drug release rate from fdm 3d-printed tablets that has been introduced https://doi.org/10.5599/admet.1622 abu t.m. serajuddin admet & dmpk 11(1) (2023) 33-55 40 by solanki et al. [19] is using mixtures of polymers. in this way, the polymers that are water-soluble but not normally printable could be printed by mixing with other polymers. the authors mixed polyvinylpyrrolidonevinyl acetate copolymer (kollidon® va64) with hydroxypropyl methylcellulose (affinisol® hme 15 cp) at 1:1 w/w ratio to increase the dissolution rate of a model drug haloperidol as compared to that with cellulose alone. they obtained complete drug release in <1 h at ph 2, and it took ~3 h to completely dissolve the drug at ph 6.8 because of the ph-dependent solubility of the drug used. while kollidon® va64 is not printable due to the brittleness of filaments, mixing with affinisol® hme 15 cp made the formulations printable. this study will be further elaborated in a later section of this article as a case study for a systematic approach to the development of fdm 3d-printed tablets. in another study, saydam and takka [47] used 20 % kollidon® va64, 20 % hpmc and 34 % soluplus® along with different drugs (15 %) and several excipients like tween 80, gelucire 44/14, gelucire 48/16, triacetin, etc., to increase dissolution rate and printability of formulations. up to 80 % drug release in <1 h was obtained from several formulations for tablets with 50 % infill, while the dissolution from other formulations took longer. incorporation of additives it has been reported in the literature that the incorporation of certain fillers in fdm 3d-printed tablets could also increase the dissolution rate of drugs from tablets. okwuosa et al. [48] observed complete release of theophylline or dipyridamole in 30 min from tablets containing 50 % polyvinylpyrrolidone (pvp; mw 40,000 ) as the polymer, 12.5 % triethyl citrate (tec) as the plasticizer, 27.5 % talc as the filler, along with 10 % drug load. talc remained solid and fully crystalline, and the drugs were partially crystalline in the tablets. the exact mechanism of how the additives increased drug release rates has not been elucidated in the study. it has been postulated that the matrix structure of tablet influenced the drug release rate. sadia et al. [49] also observed similar results in the presence of tec and a different filler, tricalcium phosphate (tcp), where over 80 % release of four model drugs (5-acetyl salicylic acid, captopril, prednisolone, or theophylline) was obtained in <30 min from tablets containing eudragit® epo, tec, tcp, and drug at the ratio of 46.75:3.25:37.5:12.5. the amorphous tricalcium phosphate dispersed in the tablet as separate particles, and the drug was partially crystalline. the exact mechanism for the improvement of dissolution rates by these additives has also not been elucidated. in addition to increasing the drug dissolution rate, the presence of tcp in the formulation helped in the printing process by enhancing the solidification of different printing layers after extrusion. in another study of the effect of additives, fanous et al. [50] observed that the incorporation of the hydrophilic plasticizer xylitol (13.5 %] and the pore former maltodextrin (9.5 %) in a eudragit® epo-based formulation could enhance drug release from the fdm 3d-printed tablets. by using these additives along with the modification of infill density, they developed tablets with complete dissolution in 30 to 60 min. there are also reports on the increase in drug dissolution rates by the incorporation of polyethylene oxide as the additive [38,43,51,52]. for example, gultekin et al. [38] used mixtures of eudragit® epo and polyethylene oxide (peo) n80 at 70:30, 60:40 and 50:50 w/w ratios for a low-dose drug (0.25 to 1 mg) and observed complete drug release in a ph 1.2 medium (0.1m hcl) in 20 min. the authors reported that, when extruded at 130 °c, the filaments were flexible, smooth, and extrudable. it is, however, not clear from the paper whether the high ratios of polymer and peo used (70:30, 60:40 and 50:50 w/w) were miscible with each other or whether they were phase separated, which could be an issue on the physical stability of filaments and tablets. also, the dissolution testing was conducted at ph 1.2, where the polymer eudragit® epo is soluble; it is not known what the dissolution rates of formulations at higher intestinal ph conditions admet & dmpk 11(1) (2023) 33-55 development of rapidly dissolving fdm 3d-printed tablets doi: https://doi.org/10.5599/admet.1622 41 (ph 5.5 to 7.5) would be since the polymer is not soluble at such ph. unlike gultekin et al. [38], the experience of alhijjaj et al. [43] in using peo and related additives was quite mixed. they prepared fdm 3d-printed discs containing 10 % felodipine by using 50 % of either eudragit® epo or soluplus as the polymeric matrix and adding 15 % peo, 15 % tween 80 and 10 % polyethylene glycol (peg) 4000 to the formulations. the effects of peo and the other additives on drug release from eudragit® epo and soluplus®-based 3d-printed discs were different. while complete drug dissolution was observed from the eudragit® epo-based formulation at the gastric ph condition of 1.2 in <30 min, the drug release from the same formulation was much lower at the intestinal ph condition of 6.8 as only ~20 % drug dissolution was observed in 6h, thus showing the effect of ph on drug release. for the soluplus®-based discs, on the other hand, the drug release rates were extremely low at both ph 1.2 and 6.8 despite the presence of peo. moreover, peo and peg 400 were not miscible with polymers like eudragit® epo and soluplus® as differential scanning calorimetric (dsc) and powder x-ray diffraction (pxrd) analyses indicated the existence of separate phases in the tablets (discs). such results indicated that the incorporation of low-melting peo and other additives might not be the solution to the dissolution issue with fdm 3d-printed tablets. in a different study, kollamaram et al. [53] increased the dissolution rate and decreased the printing temperature of the fdm 3d-printed tablets by incorporating 20 % peg 1500 and 10 % mannitol in a formulation containing 3 % ramipril (drug) and 65 % kollidon® va64 or a mixture of kollidon® va64 and kollidon® 12pf. the authors reported almost complete drug release in <30 min when the filaments were extruded at 65-70 °c and the tablets were printed at 90 °c. however, since the glass transition temperature of both the polymers was 109 °c [54], which is much higher than the processing temperatures of filaments and tablets, it is not clear from the report whether the relatively low melting peg 1500 (m.p.: 43-49 °c) was miscible with kollidon® va64 and kollidon® 12pf at extrusion and printing temperatures used. it could be possible that peg and polymers were only physically mixed together in the filaments due to the melting of peg, and it could also be possible that the rapid drug release was observed because of the dispersion of the low concentration drug (3 %) in the higher amount of peg (20 %) and the polymers did not play any role in the dissolution process. attempts were also made to increase dissolution rates of drugs from the fdm 3d-printed tablets by the incorporation of such super-disintegrants as sodium starch glycolate, croscarmellose sodium, and crosslinked polyvinylpyrrolidone (kollidon® cl) with a mixture of pva and hpmc used as the polymeric matrix [55]. although the drug release rate in aqueous media could be increased from 58 % in 5 h to almost complete release in 2 h by this approach, it was not a sufficient increase in dissolution rate for the development of a rapidly dissolving tablet. a faster drug release could not be obtained possibly because when the disintegrants were extruded with the polymers, they lost their swelling properties and disintegration efficiency since the disintegrants were completely surrounded by the polymers preventing any penetration of water to the disintegrants and there was also no porosity within the filaments for swelling of the disintegrants. shi et al. [56] observed that the incorporation of release-modifying excipients like poly(vinyl alcohol) (pva), soluplus®, polyethylene glycol (peg) 6000, eudragit® rlpo/rspo, hpmc k4m/e10m/k100m, kollidon® va64, etc., may also not be the solution to slow drug release from fdm 3d-printed tablets. they used various release modifiers to modulate dissolution rates of ibuprofen from the fdm 3d-printed tablets containing ethyl cellulose as the polymeric matrix. it took over 24 h for the complete release of the drug because of the nature of the polymer and release modifiers used. the incorporation of additives is a promising and potentially viable approach for increasing the dissolution rates of drugs from the fdm 3d-printed tablets. indeed, krueger et al. [57] called the polymer-additive https://doi.org/10.5599/admet.1622 abu t.m. serajuddin admet & dmpk 11(1) (2023) 33-55 42 adducts hybrid materials since the additives can impart new properties to the polymers. resulting in better performance. however, further research is necessary for a better understanding of the interaction between polymers and different additives and its effects on the drug release rate. from the above brief review of the literature, it is apparent that certain additives can plasticize polymers to make them extrudable and printable, and some other additives, especially the solid ones like talc and tcp, may also improve the printability of filaments by making them harder. care must, however, be taken in selecting additives for such effects since they do not always exert desired effects. in one study, wei et al. [33] observed that when relatively low melting additives like peg 3350 or peg 8000 were mixed with polyvinyl alcohol for plasticization of the polymer during melt extrusion into filaments, the additives melted and oozed out the extruder before exerting any plasticization effects. there could also be phase separation of additives during the stability testing of filaments and tablets. for these reasons, it is essential that the additives, especially the low-melting ones, are miscible with the polymers. modification of tablet pattern and geometry as mentioned earlier, the fdm 3d-printed tablets are hard and non-disintegrating compacts that usually dissolve by erosion. therefore, one common strategy to increase the dissolution rate of drugs from the tablets is to increase the dissolving surface area of tablets by modifying tablet geometry. some of these approaches, as shown in figure 3, are described below: figure 3. different methods of increasing surface area by modifying tablet patterns and geometry: i, changing tablet infill; ii, geometric modification by connecting blocks of fill materials; iii, introduction of perforated channels; and iv, connection of parallel plates in a radiator-like design. changing tablet infill the most common approach to increase the dissolving surface area of the fdm 3dprinted tablet is to decrease the infill of tablets, where, as shown in fig. 3-i, the infill represents the volume of a tablet taken up by the fill material during the printing of the tablet. the infill of an fdm 3d-printed tablet may vary from less than 10 % to 100 %, indicating that a tablet with, for example, 30 % and 80 % infills will have 30 % and 80 % filled and the remaining 70 % and 20 % will, respectively, be empty spaces or voids within the tablets, while a tablet with 100 % infill will be totally filled and there would not be any empty space inside. thus, the same weight of tablet with a lower infill will have a higher dissolving surface and, therefore, a higher dissolution rate in the dissolution medium than a tablet with a higher infill [58]. although the lowering of infill density may increase dissolution rates of drugs from fdm 3d-printed tablets, it may not necessarily lead to a product that may be called rapid or immediate release. in a study of the effect of infill density on the admet & dmpk 11(1) (2023) 33-55 development of rapidly dissolving fdm 3d-printed tablets doi: https://doi.org/10.5599/admet.1622 43 dissolution of a weakly acidic compound, 2-(4-isobutylphenyl) propionic acid, from fdm 3d-printed tablets containing hydroxypropyl methyl acetate succinate (hpmcas-mg), thakkar et al. [58] observed that it took ~7h for 80 % drug release at ph 6.8 from tablets with 80 % infill. the drug release rate increased progressively when the fill densities were decreased to 60, 40 and 20 %. however, even at 20 % infill, the drug release rate was relatively slow (80 % drug release in ~2 h). considering that hpmcas-mg is soluble at ph 6.8, such release rates are not optimal for the development of rapidly dissolving drug products. in another study from the same laboratory, zhang et al. [59] obtained complete drug release from ibuprofen-hpmc-mg tablets in 3-4 h from fdm 3d-printed tablets with 25 to 75 % infill density. such slow-release rates were also observed by goyanes et al. [60] from fdm 3d-printed tablets of the organic dye fluorescein containing polyvinyl alcohol, a water-soluble polymer, as the matrix. they observed complete release of the dye in aqueous media from tablets with 10 % infill in 6 h, while only about 70 % release was observed at the same period from tablets with 50 and 90 % infills. geometric modification by connecting blocks of fill materials in addition to changing the infill density, several other geometrical modifications to increase drug release rates have also been reported in the literature. in one such geometric modification shown in fig. 3-ii, arafat et al. [61] used a blend of theophylline, hydroxypropyl cellulose and triacetin (50:45:5) for the development of fdm 3d-printed tablets, where they connected 9 blocks of fill materials with 3 bridges within a tablet by leaving gaps ranging from 0 to 1.2 mm between the blocks. the drug release rate from tablets depended on the size of the block, the number of bridges and the spacing between the blocks. it took almost 2 h for the complete dissolution of the drug when the gap in between blocks was 0.2 mm. on the other hand, when the gap was increased to 1 mm or higher, over 80 % theophylline dissolved in 30 min; the higher gap was necessary for the collapse of the bridge and the disintegration of tablet. introduction of perforated channels in another approach to increase the dissolution rate of tablets, sadia et al. [62] introduced perforated channels within the fdm 3d-printed tablets for exposure of the larger surface area to dissolution media (fig. 3-iii). the capsule-shaped tablets (caplets) contained channels of increasing width (0.2, 0.4, 0.6, 0.8 or 1.0 mm) and variable length that were aligned parallel or at a right angle to the long axis of tablet. over 80 % of drug release in 30 min at ph 1.2 was observed when the channel width was ≥ 0.6 mm. however, the formulation for the tablet was composed of a mixture of hydrochlorothiazide (hct), eudragit® e, tricalcium phosphate (tcp) and triethyl citrate (tec) at the ratio 12.5:46.75:25.37:3.25 w/w. as was discussed earlier, tcp plays a major role in increasing dissolution rates of drugs from 3d-printed tablets, and it is not clear what roles tcp played in increasing dissolution rates of hct from this formulation. further, eudragit® e is soluble under gastric ph conditions (ph<5) but insoluble at higher intestinal ph conditions. therefore, it is also essential to consider how much improvement in dissolution rate the perforated channels may provide at ph>5 from formulations containing eudragit® e. additionally, the dissolution rate improvements from tablets by low infill density and by perforated channels appear to be based on similar principles, and it is essential to determine whether one method would have an advantage over the other. connection of parallel plates in yet another geometric modification to increase dissolution rates of tablets, isreb et al. [63] connected parallel plates of infill in a radiator-like design with inter-plate spacings of 0.5, 1, 1.5, or 2.0 mm (fig. 3-iv). filaments used for this study contained 30:35:35 w/w mixtures of theophylline, polyethylene glycol (peg) 6000, and different grades of polyethylene oxide (peo) with molecular weights ranging from 100,000 to 900,000, where theophylline was not completely miscible with other components and remained mostly crystalline. a minimum inter-plate spacing of 1 mm was deemed necessary for increasing the drug release rate, where over 80 % of theophylline was released from fdm 3dhttps://doi.org/10.5599/admet.1622 abu t.m. serajuddin admet & dmpk 11(1) (2023) 33-55 44 printed tablets in <30 min at ph 1.2. the radiator-like design possibly helped in the swelling and disintegration of tablets, whereby crystalline theophylline was liberated in the dissolution medium. among other variations of tablet patterns and geometry, khaled et al. [64] introduced mesh structures within tablets to increase surface area and dissolution rates of 3d-printed paracetamol tablets that appear to be analogous to changing the infill density. other investigators changed tablet shapes and dimensions to increase the dissolution rates of tablets [65,66]. for example, goyanes et al. [65] printed tablets with five different geometries, namely, cube, pyramid, cylinder, sphere, and torus, and observed that the drug release rate from the tablets was dependent on the surface area to volume ratio since, as expected, the higher the surface to volume ratio was, so was the dissolution rate. raje et al. [67] introduced a new tablet geometry where they applied different numbers of outer shells to the fdm 3d-printed tablets to modulate drug release from tablets containing a water-soluble polymer, poly(2-ethyl-tetra-oxazoline) [petox], and two different drugs (dextromethorphan hydrobromide and hydrochlorothiazide). the drug release rate increased with the decrease in the number of outer shells. the release rates of dextromethorphan hydrobromide from one-shell and five-shell tablets were, respectively, >80 % and >60 % in 1 h, and because of the lower solubility of hydrochlorothiazide, the release rate was much lower as it took 2 h for 80 % drug release from a one-shell tablet. although the modification of tablet geometry could be an attractive approach to increase dissolution rates of fdm 3d-printed tablet, melocchi et al. [68] reported that computer-aided design (cad) of such tablets could be very challenging as they involve detailed and complex digital models, and the users may need extensive training and years of experience to master all the functionalities. this is possibly the reason why each geometric design of 3d-printed tablets reported in the literature, except for the infill density, is essentially one of a kind and there is not much follow-up of the technique by other investigators. hopefully, the technology will grow rapidly enough that the geometric modification will be easy to design and will become more user-friendly. in addition to the complexity of designing tablets, one should keep in mind that the geometric modification may improve the dissolution rate only to a certain extent, and it may not always lead to immediately dissolving or rapidly dissolving tablets, which also depends on types of polymers used and compositions of formulations. application of acid-base supersolubilization principle another promising approach to increase dissolution rates of fdm 3d-printed tablets that has been recently introduced by patel and serajuddin [20,21] is the application of acid-based supersolubilization or abs principle in the development of rapidly dissolving haloperidol-glutaric acid and haloperidol-malic acid tablets, where an otherwise brittle polymer, kollidon® va64, was made printable, the loading in the tablet was increased, melt extrusion and printing temperatures were lowered, and the drug release rate was increased. the abs principle, as described in figure 4, represents a novel strategy of greatly increasing the aqueous solubility of weakly basic drugs and thereby increasing drug dissolution rates by interaction with weak acids that would not normally form salts with the drugs [69-71]. this principle is based on classical phsolubility theory [72,73], where, as shown in fig. 4-i, the aqueous solubility of a basic drug increases according to line a→b when the ph of an aqueous medium is lowered by the addition of acid until phmax or the ph of maximum solubility (pont b), is reached. a salt is formed below phmax, where the drug solubility is dictated by the solubility of its salt (line b→c). however, as shown in fig. 4-ii, if the added acid is not strong enough to lower ph below phmax, no salt is formed, and the solubility of the basic drug keeps on increasing, leading to supersolubilization. singh et al. [70] observed that the aqueous solubility of the model drug, haloperidol, could be increased from the free base solubility of 1 µg/ml to >300 mg/ml when weak acids like malic, admet & dmpk 11(1) (2023) 33-55 development of rapidly dissolving fdm 3d-printed tablets doi: https://doi.org/10.5599/admet.1622 45 tartaric, and citric acids were used to decrease ph. when the highly concentrated solutions were dried, they formed amorphous solids. parikh et al. [74] demonstrated that such an acid-base interaction leading to the formation of amorphous drugs can also occur in a solid state when mixtures of basic drugs with weak acids are heated together or when the acid-base mixtures were melt extruded along with polymers. in contact with water, the amorphous drugs can exhibit supersolubilization and high dissolution rates. the application of such findings in greatly increasing dissolution rates of basic drugs from the fdm 3d-printed tablets will be illustrated later in case study 2. some recent findings in our laboratory demonstrated that the abs principle is also applicable to increasing dissolution rates of acidic drugs from fdm 3d-printed tablets by interaction with weak bases. figure 4. graphical illustration of acid-base supersolubilization (abs) principle for a basic drug, where graph i (top) represents a typical ph versus solubility profile resulting in salt formation at low ph, and graph 2 (bottom) shows a high increase in solubility (supersolubilization) with lowering of ph when no salt is formed by the addition of a weak acid. recent progress in development of fdm 3d-printed tablets – case studies as evident from the above review of literature, ensuring rapid drug release is a major challenge in the development of successful fdm 3d-printed tablets. most studies to increase drug release rates from such tablets reported in the literature have been preliminary and exploratory in nature. however, intense efforts towards the development of new strategies for the formulation of rapidly dissolving fdm 3d-printed tablets are continuing. two emerging strategies that may have major impacts in the development of rapidly dissolving fdm 3d-printed tablets in the future are described below by using case studies. https://doi.org/10.5599/admet.1622 abu t.m. serajuddin admet & dmpk 11(1) (2023) 33-55 46 case study 1: development of fdm 3d-printed tablets by systematic screening of polymers for rapid drug release as mentioned earlier, only a very limited number of pharmaceutically acceptable polymers are available for fdm 3d printing. most polymers used in pharmaceutical dosage forms may not be extrudable into filaments and even those that could be extruded may not be printed since the filaments produced may be either too brittle or too soft. even when a polymer is both extrudable and printable, it may have other disadvantages like lack of drug-polymer miscibility, slow drug release rate, and the need for high extrusion and printing temperatures where drug, polymer or both may degrade. therefore, it is essential that polymers and other additives are carefully selected for the successful development of fdm 3d-printed tablets. however, in most of the reported studies, the selection of polymers and other excipients appears to be either arbitrary or by trial and error. to address these issues, solanki et al. [19] developed a systematic approach for identifying polymers or polymeric blends for the development of rapidly dissolving fdm 3d-printed tablets. later, aho et al. [75] also developed a roadmap for the development of fdm 3d-printed tablets based on material properties and the processing behavior of drug-excipient combinations, but their focus was on the printing of tablets rather than drug release rate. the strategy developed by solanki et al. {19] for the identification of polymers for rapid drug release, as shown schematically in figure 5, involved the following steps: 1. hot melt extrusion of haloperidol with different pharmaceutical grade polymers to determine whether filaments would be formed and whether the drug is molecularly or amorphously dispersed in the polymers to form amorphous solid dispersion (asd), 2. evaluate drug release from crushed extrudates for a preliminary assessment of whether the polymer would provide immediate drug release and, if the drug dissolves rapidly from the extrudates, whether the filaments would be printable, 3. determine the miscibility of the drug with polymer or mixtures of polymers to identify the maximum drug load that may be used in the filaments identified in step 2 above, 4. print tablets by fdm using drug-loaded filaments with selected polymers or polymer blends using optimum drug loads determined by drug-polymer miscibility testing in step 3 above. 5. evaluate drug release from fdm 3d-printed tablets to determine whether they meet the target drug release. figure 5. systematic approach for the identification of polymers for the development of rapidly dissolving fdm 3d-printed tablets. admet & dmpk 11(1) (2023) 33-55 development of rapidly dissolving fdm 3d-printed tablets doi: https://doi.org/10.5599/admet.1622 47 solanki et al. [19] screened four commonly used polymers, namely, kollidon va64 (vinylpyrrolidone-vinyl acetate copolymer), kollicoat® ir (polyvinyl alcohol-polyvinyl glycol graft copolymer), affinisol hpmc hme 15cp (hydoxypropyl methylcellulose hme), and aquoat as-mg (hydroxypropyl methylcellulose acetate succinate) for their suitability in the development of rapidly dissolving fdm 3d-printed tablets by using a model basic drug, haloperidol (pka~8), with ph-dependent solubility (1-4 mg/ml at ph 1-3; 2-3 µg/ml at ph~7). in step 1 of the process, they observed that drug formulations with all four polymers were extrudable when they were extruded at a common temperature of 150 °c. in step 2, the dissolution testing of crushed filaments was conducted at ph 2 and 6.8, and the printability of whole (uncrushed) filaments was studied by printing actual tablets. it was observed that the crushed filaments containing kollidon va64 only provided the most rapid drug release rate, but the filaments were not printable. for this reason, 1:1-mixtures of kollidon® va64 with affinisol as well as hpmcas mg were also tested for dissolution and printability, and the results are presented in table 2. although slower than that with kollidon® va64 alone, the drug-loaded filaments containing the 1:1-mixtures of kollidon® va64 with affinisol was still quite rapid. filaments containing kollicoat® ir or 1:1-mixtures of kollidon® va64 with hpmcas mg exhibited delayed drug release rate. based on the results presented in table 2, filaments containing 1:1 kollidon va64 and affinisol were selected for further development. the miscibility of haloperidol with kollidon va64, affinisol and their mixtures was then studied in step 3 using the film casting technique previously developed by parikh et al. [32]. the miscibility of the drug was 20 % in kollidon va64, 10 % in affinisol and 15 % in the kollidon va64affinisol mixture. based on these results, the formulation composition selected for 3d printing was 10 % haloperidol, 45 % kollidon va64 and 45 % affinisol. table 1. the results of the screening of different polymers and polymer-polymer mixtures with 10 % w/w drug loads for drug release from crushed filaments at ph 2 and 6.8 and printability of whole filaments. polymer printability drug release from milled filaments 10 % haloperidol + 90 % kollidon va 64 breaks into the drive gear; not printable immediate 10 % haloperidol + 90 % kollicoat ir printable retarded 10 % haloperidol + 90 % affinisol 15 cp printable retarded 10 % haloperidol + 45 % hpmcas mg + 45 % kollidon va 64 printable retarded 10 % haloperidol + 45 % kollidon va 64 + 45 % affinisol 15 cp printable immediate the selected formulation was then printed into tablets at 210⁰c in step 4 of the process. two infill densities, 100 % and 60 %, were used to determine whether there would be any effect of infill density on drug release. as shown in figure 6, the dissolution testing of the fdm 3d-printed tablets in step 5 gave ~80 % of drug release from tablets with 60 % infill in 45 and 120 min at ph 1.2 and 6.8, respectively. from tablets with 100 % infill, over 80 % of the drug was released at ph 1.2 and 6.8 in 2 and 3 h, respectively. although the dissolution rate of haloperidol obtained by solanki et al. [19] did not meet the fda guideline of 85 % and higher drug release in 60 min for immediate drug release [76], the rate was still rapid enough for complete drug release in 2-3 h, indicating the potential for relatively rapid and complete drug release during the gastrointestinal residence time [77]. thus, such a systematic approach demonstrates how 3d-printed tablets with relatively rapid drug release may be identified. if the drug release rate of the fdm 3d-printed tablet thus obtained does not meet the target profile, modifications of the formulation may be made to https://doi.org/10.5599/admet.1622 abu t.m. serajuddin admet & dmpk 11(1) (2023) 33-55 48 further increase the drug release rate. later studies in our laboratory showed that the formulation was still extrudable into filaments if the amount of cellulose was reduced from 1:1 w/w kollidon® va64 to affinisol® 15cp ratio to 3:1 ratio of the same, and decreasing the amount of cellulose in the formulation resulted in a higher drug release rate from the tablets produced. figure 6. dissolution of haloperidol from fdm 3d-printed tablets containing 10 % drug load in 1:1 kollidon® va64-affinisol® 15 cp mixtures and having 100 % and 60 % infill densities at ph 2 and 6.8. figure 5 also identifies several issues that are common in the 3d printing process. the relatively high temperature (150 °c) used for melt extrusion of filaments and even higher temperature (210 °c) necessary for 3d printing of filaments into tablets could pose issues if the drug is thermolabile. there could also be issues with drug loading in tablets depending on the drug-polymer miscibility. although the miscibility may differ from drug to drug and polymer to polymer, it is essential that drug loading in tablets should be based on a thorough evaluation of the drug-polymer miscibility. thus, the systematic identification of such issues during the development program guides the formulators towards evaluating and possibly resolving them. case study 2: development of fdm 3d-printed tablets with rapid drug release by acid-base interactions this case study represents a recent advance in fdm 3d-printed technology that not only provides rapid drug release from the tablets, but, at the same time, also resolves two other major issues faced in the development of fdm 3d-printed tablets, namely, low drug-polymer miscibility and the need for high processing temperature. this case study is based on the acid-base supersolubilization (abs) principle described earlier. the application of the abs principle in increasing release rates of drugs from fdm 3d-printed tablets is shown schematically in figure 7 based on the work of patel and serajuddin [20], where the formulation contained haloperidol as the model basic drug, glutaric acid as the super-solubilizing agent, and kollidon® va64 as the polymer. in the presence of glutaric acid, the aqueous solubility of haloperidol increased to >300 mg/ml, which indicated that an interaction between haloperidol and glutaric acid occurred, leading to the supersolubilization of the drug. such an interaction between haloperidol and glutaric acid was also observed in the solid state, which reduced the melt extrusion temperature of the formulation from 150 °c to 115 °c since the haloperidol-glutaric mixture converted to a semisolid mass during melt extrusion that plasticized the polymer. while the haloperidol-kollidon® va64 mixture was not printable, the plasticizing effect of the acid-base mixture when glutaric acid was added to the formulation also made the filaments printable and lowered the printing temperature to as low as 120 °c as compared to ~200 °c required in the absence of glutaric acid. the drug was also highly miscible with the formulation to as high as 50 % w/w. admet & dmpk 11(1) (2023) 33-55 development of rapidly dissolving fdm 3d-printed tablets doi: https://doi.org/10.5599/admet.1622 49 figure 7. effects acid-base interaction between glutaric acid and haloperidol on melt extrusion and printing temperatures, printability, and drug-polymer miscibility of fdm 3d printed tablets. the drug release rates from tablets containing glutaric acid were rapid. as shown in fig. 8, almost all the drug dissolved at both ph 2 and 6.8 in 30 min, and because of the high dissolution rate of the drug from such a formulation, there was also no significant effect of the infill density. figure 8. dissolution of haloperidol from fdm 3d-printed tablets containing 15 % drug load from a formulation containing 1:2 molar ratio of drug to glutaric acid and having 100 % and 60 % infill densities at ph 2 and 6.8. kol, hal and ga represent kollidon® va64, haloperidol, and glutaric acid. similar results were also obtained for fdm 3d-printed tablets containing haloperidol-malic acid mixtures [21], where the kollidon® va64-haloperidol-malic acid ternary mixtures could be extruded into filaments at 120 °c and printed into tablets at 125 and 100 °c, respectively, at 15 % and 30 % drug loads. as high as 50 % w/w drug load could be achieved due to the high miscibility of the drug with the formulation; however, the filaments were too soft with 40 and 50 % w/w drug loads, and, therefore, formulations with only 15 and 30 % w/w drug loads were printed. over 80 % drug release could be achieved from 100% infill tablets in <30 min at ph 1.2 and 6.8. https://doi.org/10.5599/admet.1622 abu t.m. serajuddin admet & dmpk 11(1) (2023) 33-55 50 summary and conclusions although there has been a great interest and much research in recent years in the fdm 3d-printing technology for the development of tablets for personalized medicines, it faces many challenges, such as limited availability of suitable polymers, low drug-polymer miscibility, the need for high processing temperatures, and the slow drug release rate that impede the development of clinically efficacious and commercially viable fdm 3d printed tablets. these challenges are discussed in this article with a special focus on the drug release rate. the fdm 3d printing produces compact tablets that dissolve slowly by erosion, while rapid drug release is important for the clinical efficacy of most drug candidates. such strategies as using rapidly dissolving polymers or polymer-polymer mixtures, incorporation of additives, modification of tablet pattern and geometry, and supersolubilization by acid-base interaction to increase drug dissolution rates are reviewed and the pros and cons of these strategies are discussed. many polymers, such as pvps, kollidon® va64, etc., are highly water-soluble and could potentially increase dissolution rates of drugs from the fdm 3d-printed tablets, but their filaments are too brittle and cannot be printed. on the other hand, cellulosic filaments are printable, but the tablets produced are slowly dissolving. mixing cellulose with kollidon® va64 is a promising way of increasing dissolution rates of the formulations as the mixtures are printable and, at the same time, they can improve dissolution rates over that with celluloses alone. a polymer like eudragit® epo can also produce rapidly dissolving tablets; however, since it dissolves only at ph<5, the dissolution of tablets under intestinal ph conditions at ph>5 could be slow and incomplete. the printability of eudragit® epo could also be a concern since it has a relatively low glass transition temperature (~50 °c) and its filaments may require solid additives to make them harder for printing. the incorporation of additives can also increase the dissolution rates of drugs and improve the printability of tablets. however, the rationale for the use of different additives has not been fully elucidated in the literature, and the selection of appropriate additives will require careful consideration and further research. additionally, there could be phase separation and potential physical instability of filaments and tablets if the additives are not miscible with polymers. among different methods of modifying tablet patterns and geometry, changing infill density is the most used, and further research and technological progress are necessary for the implementation of other methods. there are, however, limitations to what extent the dissolution rate may be improved by this method. two emerging strategies for the development of rapidly dissolving fdm 3d-printed are elaborated by using case studies, where one provides a systematic approach for screening polymers for rapidly dissolving tablets, and the other involves the acid-base supersolubilization (abs) principle that not only increases solubility and dissolution rates of drugs but also increases drug-polymer miscibility and decreases processing temperatures for tablets. although very promising, most of the strategies applied to the development of fdm 3d-printed tablets are rather at their early stages of development, and further research is needed for the implementation of these strategies into the development of marketed drug products. many of the gaps in the current state of the fdm 3d-printing technology have been presented in this article, and a few strategies to fill those gaps have been highlighted. hopefully, they will provide guidance for the future development of the technology. conflict of interest: the author declares no conflict of interest. references [1] a. awad, a. goyanes, a. basit, a. zidan, c. xu, d. li, r. narayan, r.k. chen. a review of state-of-theart on enabling additive manufacturing processes for precision medicine. journal of manufacturing science and engineering 145 (2023) 010802. https://doi.org/10.1115/1.4056199 [2] k.v. wong, a. hernandez. a review of additive manufacturing. international scholarly research notices 2012 (2012) 208760. https://doi.org/10.5402/2012/208760 https://doi.org/10.1115/1.4056199 https://doi.org/10.5402/2012/208760 admet & dmpk 11(1) (2023) 33-55 development of rapidly dissolving fdm 3d-printed tablets doi: https://doi.org/10.5599/admet.1622 51 [3] spritam fda approval history. https://www.drugs.com/history/spritam.html (last accessed: november 12, 2022) [4] a. melocchi, m. uboldi, m. cerea, a. foppoli, a. maroni, s. moutaharrik, l. palugan, l. zema, a. gazzaniga. a graphical review on the escalation of fused deposition modeling (fdm) 3d printing in the pharmaceutical field. journal of pharmaceutical sciences 109(10) (2020) 2943-2957. https://doi.org/10.1016/j.xphs.2020.07.011 [5] j. wang, y. zhang, n.h. aghda, a.r. pillai, r. thakkar, a. nokhodchi, m. maniruzzaman. emerging 3d printing technologies for drug delivery devices: current status and future perspective. advanced drug delivery reviews 174 (2021) 294-316. https://doi.org/10.1016/j.addr.2021.04.019 [6] w. jamróz, j. szafraniec, m. kurek, r. jachowicz. 3d printing in pharmaceutical and medical applications–recent achievements and challenges. pharmaceutical research 35(9), (2018) 1-22. https://doi.org/10.1007/s11095-018-2454-x [7] c. parulski, o. jennotte, a. lechanteur, b. evrard. challenges of fused deposition modeling 3d printing in pharmaceutical applications: where are we now?. advanced drug delivery reviews. 175 (2021) 113810. https://doi.org/10.1016/j.addr.2021.05.020 [8] h. ragelle, s. rahimian, e.a. guzzi, p.d. westenskow, m.w. tibbitt, g. schwach, r. langer. additive manufacturing in drug delivery: innovative drug product design and opportunities for industrial application. advanced drug delivery reviews 178 (2021) 113990. https://doi.org/10.1016/j.addr.2021.113990 [9] r.d.p. reddy, v. sharma. additive manufacturing in drug delivery applications: a review. international journal of pharmaceutics 589 (2020) 119820. https://doi.org/10.1016/j.ijpharm.2020.119820 [10] i. seoane-viaño, s.j. trenfield, a.w. basit, a. goyanes. translating 3d printed pharmaceuticals: from hype to real-world clinical applications. advanced drug delivery reviews 174 (2021) 553-575. https://doi.org/10.1016/j.addr.2021.05.003 [11] s.j. trenfield, a. awad, a. goyanes, s. gaisford, a.w. basit. 3d printing pharmaceuticals: drug development to frontline care. trends in pharmacological sciences 39(5) (2018) 440-451. https://doi.org/10.1016/j.tips.2018.02.006 [12] l. zema, a. melocchi, a. maroni, a. gazzaniga. three-dimensional printing of medicinal products and the challenge of personalized therapy. journal of pharmaceutical sciences 106(7) (2017) 1697-1705. https://doi.org/10.1016/j.xphs.2017.03.021 [13] s. bandari, d. nyavanandi, n. dumpa, m.a. repka. coupling hot melt extrusion and fused deposition modeling: critical properties for successful performance. advanced drug delivery reviews 172 (2021) 52-63. https://doi.org/10.1016/j.addr.2021.02.006 [14] d.k. tan, m. maniruzzaman, a. nokhodchi. advanced pharmaceutical applications of hot-melt extrusion coupled with fused deposition modelling (fdm) 3d printing for personalised drug delivery. pharmaceutics 10(4) (2018) 203. https://doi.org/10.3390/pharmaceutics10040203 [15] s.a. khaled, j.c. burley, m.r. alexander, j. yang, c.j. roberts. 3d printing of five-in-one dose combination polypill with defined immediate and sustained release profiles. journal of controlled release 217 (2015) 308-314. https://doi.org/10.1016/j.jconrel.2015.09.028 [16] s.a. khaled, j.c. burley, m.r. alexander, j. yang, c.j. roberts. 3d printing of tablets containing multiple drugs with defined release profiles. international journal of pharmaceutics 494(2) (2015) 643-650. https://doi.org/10.1016/j.ijpharm.2015.07.067 [17] p.a. dickinson, f. kesisoglou, t. flanagan, m.n. martinez, h.b. mistry, j.r. crison, j.e. polli, m.t. cruañes, a.t. serajuddin, a. müllertz. optimizing clinical drug product performance: applying biopharmaceutics risk assessment roadmap (bioram) and the bioram scoring grid. journal of pharmaceutical sciences 105 (2016) 3243-3255. https://doi.org/10.1016/j.xphs.2016.07.024 [18] a. selen, p.a. dickinson, a. müllertz, j.r. crison, h.b. mistry, m.t. cruañes, m.n. martinez, h. lennernäs, t.l. wigal, d.c. swinney. the biopharmaceutics risk assessment roadmap for optimizing clinical drug product performance. journal of pharmaceutical sciences 103 (2014) 3377-3397. https://doi.org/10.1002/jps.24162 https://doi.org/10.5599/admet.1622 https://www.drugs.com/history/spritam.html https://doi.org/10.1016/j.xphs.2020.07.011 https://doi.org/10.1016/j.addr.2021.04.019 https://doi.org/10.1007/s11095-018-2454-x https://doi.org/10.1016/j.addr.2021.05.020 https://doi.org/10.1016/j.addr.2021.113990 https://doi.org/10.1016/j.addr.2021.113990 https://doi.org/10.1016/j.ijpharm.2020.119820 https://doi.org/10.1016/j.addr.2021.05.003 https://doi.org/10.1016/j.tips.2018.02.006 https://doi.org/10.1016/j.xphs.2017.03.021 https://doi.org/10.1016/j.addr.2021.02.006 https://doi.org/10.3390/pharmaceutics10040203 https://doi.org/10.1016/j.jconrel.2015.09.028 https://doi.org/10.1016/j.ijpharm.2015.07.067 https://doi.org/10.1016/j.xphs.2016.07.024 https://doi.org/10.1002/jps.24162 abu t.m. serajuddin admet & dmpk 11(1) (2023) 33-55 52 [19] n.g. solanki, m. tahsin, a.v. shah, a.t. serajuddin. formulation of 3d printed tablet for rapid drug release by fused deposition modeling: screening polymers for drug release, drug-polymer miscibility and printability. journal of pharmaceutical sciences 107 (2018) 390-401. https://doi.org/10.1016/j.xphs.2017.10.021 [20] n.g. patel, a.t. serajuddin. development of fdm 3d-printed tablets with rapid drug release, high drugpolymer miscibility and reduced printing temperature by applying the acid-base supersolubilization (abs) principle. international journal of pharmaceutics 600 (2021) 120524. https://doi.org/10.1016/j.ijpharm.2021.120524 [21] n.g. patel, a.t. serajuddin. improving drug release rate, drug-polymer miscibility, printability, and processability of fdm 3d-printed tablets by weak acid-base interaction. international journal of pharmaceutics (2023), in press [22] k. ilyés, n.k. kovács, a. balogh, e. borbás, b. farkas, t. casian, g. marosi, i. tomuță, z.k. nagy. the applicability of pharmaceutical polymeric blends for the fused deposition modelling (fdm) 3d technique: material considerations–printability–process modulation, with consecutive effects on in vitro release, stability and degradation. european journal of pharmaceutical sciences 129 (2019) 110123. https://doi.org/10.1016/j.ejps.2018.12.019 [23] r. govender, e.o. kissi, a. larsson, i. tho. polymers in pharmaceutical additive manufacturing: a balancing act between printability and product performance. advanced drug delivery reviews 177 (2021) 113923. https://doi.org/10.1016/j.addr.2021.113923 [24] j.m. nasereddin, n. wellner, m. alhijjaj, p. belton, s. qi. development of a simple mechanical screening method for predicting the feedability of a pharmaceutical fdm 3d printing filament. pharmaceutical research 35 (2018) 1-13. https://doi.org/10.1007/s11095-018-2432-3 [25] s. borandeh, b. van bochove, a. teotia, j. seppälä. polymeric drug delivery systems by additive manufacturing. advanced drug delivery reviews 173 (2021) 349-373. https://doi.org/10.1016/j.addr.2021.03.022 [26] s. cailleaux, n.m. sanchez-ballester, y.a. gueche, b. bataille, i. soulairol. fused deposition modeling (fdm), the new asset for the production of tailored medicines. journal of controlled release 330 (2021) 821-841. https://doi.org/10.1016/j.jconrel.2020.10.056 [27] s. henry, a. samaro, f.h. marchesini, b. shaqour, j. macedo, v. vanhoorne, c. vervaet. extrusion-based 3d printing of oral solid dosage forms: material requirements and equipment dependencies. international journal of pharmaceutics 598 (2021) 120361. https://doi.org/10.1016/j.ijpharm.2021.120361 [28] r. liu, water-insoluble drug formulation, 3rd edition ed., boca raton, 2018. https:/doi.org/10.1201/9781315120492. [29] g.g. pereira, s. figueiredo, a.i. fernandes, j.f. pinto. polymer selection for hot-melt extrusion coupled to fused deposition modelling in pharmaceutics. pharmaceutics 12 (2020) 795. https://doi.org/10.3390/pharmaceutics12090795 [30] g.m. khalid, n. billa. solid dispersion formulations by fdm 3d printing—a review. pharmaceutics 14 (2022) 690. https://doi.org/10.3390/pharmaceutics14040690 [31] l.a. pinho, a.l. lima, l.l. sa-barreto, t. gratieri, g.m. gelfuso, r.n. marreto, m. cunha-filho. preformulation studies to guide the production of medicines by fused deposition modeling 3d printing. aaps pharmscitech 22 (2021) 1-12. https://doi.org/10.1016/j.jpha.2021.09.010 [32] t. parikh, s.s. gupta, a.k. meena, i. vitez, n. mahajan, a.t. serajuddin. application of film-casting technique to investigate drug–polymer miscibility in solid dispersion and hot-melt extrudate. journal of pharmaceutical sciences 104 (2015) 2142-2152. https://doi.org/10.1002/jps.24446 [33] c. wei, n.g. solanki, j.m. vasoya, a.v. shah, a.t. serajuddin. development of 3d printed tablets by fused deposition modeling using polyvinyl alcohol as polymeric matrix for rapid drug release. journal of pharmaceutical sciences 109 (2020) 1558-1572. https://doi.org/10.1016/j.xphs.2020.01.015 [34] n. solanki, s.s. gupta, a.t. serajuddin. rheological analysis of itraconazole-polymer mixtures to determine optimal melt extrusion temperature for development of amorphous solid dispersion. https://doi.org/10.1016/j.xphs.2017.10.021 https://doi.org/10.1016/j.xphs.2017.10.021 https://doi.org/10.1016/j.ijpharm.2021.120524 https://doi.org/10.1016/j.ijpharm.2021.120524 https://doi.org/10.1016/j.ejps.2018.12.019 https://doi.org/10.1016/j.addr.2021.113923 https://doi.org/10.1007/s11095-018-2432-3 https://doi.org/10.1016/j.addr.2021.03.022 https://doi.org/10.1016/j.addr.2021.03.022 https://doi.org/10.1016/j.jconrel.2020.10.056 https://doi.org/10.1016/j.ijpharm.2021.120361 https://doi.org/10.1016/j.ijpharm.2021.120361 https://doi.org/10.1201/9781315120492 https://doi.org/10.1201/9781315120492 https://doi.org/10.3390/pharmaceutics12090795 https://doi.org/10.3390/pharmaceutics12090795 https://doi.org/10.3390/pharmaceutics14040690 https://doi.org/10.1016/j.jpha.2021.09.010 https://doi.org/10.1002/jps.24446 https://doi.org/10.1016/j.xphs.2020.01.015 admet & dmpk 11(1) (2023) 33-55 development of rapidly dissolving fdm 3d-printed tablets doi: https://doi.org/10.5599/admet.1622 53 european journal of pharmaceutical sciences 111 (2018) 482-491. https://doi.org/10.1016/j.ejps.2017.10.034 [35] s.s. gupta, t. parikh, a.k. meena, n. mahajan, i. vitez, a.t. serajuddin. effect of carbamazepine on viscoelastic properties and hot melt extrudability of soluplus®. international journal of pharmaceutics 478 (2015) 232-239. https://doi.org/10.1016/j.ijpharm.2014.11.025 [36] n.l. funk, s. fantaus, r.c.r. beck. immediate release 3d printed oral dosage forms: how different polymers have been explored to reach suitable drug release behaviour. international journal of pharmaceutics (2022) 122066. https://doi.org/10.1016/j.ijpharm.2022.122066 [37] s. thakral, n.k. thakral, d.k. majumdar. eudragit: a technology evaluation. expert opin drug deliv 10 (2013) 131-149. https://doi.org/10.1517/17425247.2013.736962. [38] h.e. gültekin, s. tort, f. acartürk. an effective technology for the development of immediate release solid dosage forms containing low-dose drug: fused deposition modeling 3d printing. pharmaceutical research 36 (2019) 1-13. https://doi.org/10.1007/s11095-019-2655-y [39] k. pietrzak, a. isreb, m.a. alhnan. a flexible-dose dispenser for immediate and extended release 3d printed tablets. european journal of pharmaceutics biopharmaceutics 96 (2015) 380-387. https://doi.org/10.1016/j.ejpb.2015.07.027 [40] y.m. than, v. titapiwatanakun. tailoring immediate release fdm 3d printed tablets using a quality by design (qbd) approach. international journal of pharmaceutics 599 (2021) 120402. https://doi.org/10.1016/j.ijpharm.2021.120402 [41] x. lin, l. su, n. li, y. hu, g. tang, l. liu, h. li, z. yang. understanding the mechanism of dissolution enhancement for poorly water-soluble drugs by solid dispersions containing eudragit® e po. journal of drug delivery science technology 48 (2018) 328-337. https://doi.org/10.1016/j.jddst.2018.10.008 [42] h. liu, p. wang, x. zhang, f. shen, c.g. gogos. effects of extrusion process parameters on the dissolution behavior of indomethacin in eudragit® e po solid dispersions. international journal of pharmaceutics 383 (2010) 161-169. https://doi.org/10.1016/j.ijpharm.2009.09.003 [43] m. alhijjaj, p. belton, s. qi. an investigation into the use of polymer blends to improve the printability of and regulate drug release from pharmaceutical solid dispersions prepared via fused deposition modeling (fdm) 3d printing. european journal of pharmaceutics biopharmaceutics 108 (2016) 111125. https://doi.org/10.1016/j.ejpb.2016.08.016 [44] x. wang, y. zhu, x. zhao, s. zhang, m. cao, x. wang, w. li. development and characterization of an amorphous curcumin-eudragit® e100 solid dispersions with improved solubility, stability, and pharmacokinetic properties. pharmaceutical development technology (2022) 1-16. https://doi.org/10.1080/10837450.2022.2141778 [45] n. washington, c. washington, c. wilson, physiological pharmaceutics: barriers to drug absorption, crc press, 2000, 0429204590. [46] y. yang, h. wang, x. xu, g. yang. strategies and mechanisms to improve the printability of pharmaceutical polymers eudragit® epo and soluplus®. international journal of pharmaceutics 599 (2021) 120410. https://doi.org/10.1016/j.ijpharm.2021.120410 [47] m. saydam, s. takka. improving the dissolution of a water-insoluble orphan drug through a fused deposition modelling 3-dimensional printing technology approach. european journal of pharmaceutical sciences 152 (2020) 105426. https://doi.org/10.1016/j.ejps.2020.105426 [48] t.c. okwuosa, d. stefaniak, b. arafat, a. isreb, k.-w. wan, m.a. alhnan. a lower temperature fdm 3d printing for the manufacture of patient-specific immediate release tablets. pharmaceutical research 33 (2016) 2704-2712. https://doi.org/10.1007/s11095-016-1995-0 [49] m. sadia, a. sośnicka, b. arafat, a. isreb, w. ahmed, a. kelarakis, m.a. alhnan. adaptation of pharmaceutical excipients to fdm 3d printing for the fabrication of patient-tailored immediate release tablets. international journal of pharmaceutics 513 (2016) 659-668. https://doi.org/10.1016/j.ijpharm.2016.09.050 [50] m. fanous, m. bitar, s. gold, a. sobczuk, s. hirsch, j. ogorka, g. imanidis. development of immediate release 3d-printed dosage forms for a poorly water-soluble drug by fused deposition modeling: study https://doi.org/10.5599/admet.1622 https://doi.org/10.1016/j.ejps.2017.10.034 https://doi.org/10.1016/j.ejps.2017.10.034 https://doi.org/10.1016/j.ijpharm.2014.11.025 https://doi.org/10.1016/j.ijpharm.2022.122066 https://doi.org/10.1517/17425247.2013.736962 https://doi.org/10.1007/s11095-019-2655-y https://doi.org/10.1016/j.ejpb.2015.07.027 https://doi.org/10.1016/j.ijpharm.2021.120402 https://doi.org/10.1016/j.jddst.2018.10.008 https://doi.org/10.1016/j.ijpharm.2009.09.003 https://doi.org/10.1016/j.ejpb.2016.08.016 https://doi.org/10.1080/10837450.2022.2141778 https://doi.org/10.1080/10837450.2022.2141778 https://doi.org/10.1016/j.ijpharm.2021.120410 https://doi.org/10.1016/j.ejps.2020.105426 https://doi.org/10.1007/s11095-016-1995-0 https://doi.org/10.1016/j.ijpharm.2016.09.050 https://doi.org/10.1016/j.ijpharm.2016.09.050 abu t.m. serajuddin admet & dmpk 11(1) (2023) 33-55 54 of morphology, solid state and dissolution. international journal of pharmaceutics 599 (2021) 120417. https://doi.org/10.1016/j.ijpharm.2021.120417 [51] h.e. gültekin, s. tort, f. acartürk. fabrication of three dimensional printed tablets in flexible doses: a comprehensive study from design to evaluation. journal of drug delivery science technology 74 (2022) 103538. https://doi.org/10.1016/j.jddst.2022.103538 [52] h. öblom, j. zhang, m. pimparade, i. speer, m. preis, m. repka, n. sandler. 3d-printed isoniazid tablets for the treatment and prevention of tuberculosis—personalized dosing and drug release. aaps pharmscitech 20 (2019) 1-13. https://doi.org/10.1208/s12249-018-1233-7 [53] g. kollamaram, d.m. croker, g.m. walker, a. goyanes, a.w. basit, s. gaisford. low temperature fused deposition modeling (fdm) 3d printing of thermolabile drugs. international journal of pharmaceutics 545 (2018) 144-152. https://doi.org/10.1016/j.ijpharm.2018.04.055 [54] n.g. patel, a.t. serajuddin. moisture sorption by polymeric excipients commonly used in amorphous solid dispersion and its effect on glass transition temperature: i. polyvinylpyrrolidone and related copolymers. international journal of pharmaceutics 616 (2022) 121532. https://doi.org/10.1016/j.ijpharm.2022.121532 [55] m. đuranović, m. madžarević, b. ivković, s. ibrić, s. cvijić. the evaluation of the effect of different superdisintegrants on the drug release from fdm 3d printed tablets through different applied strategies: in vitro-in silico assessment. international journal of pharmaceutics 610 (2021) 121194. https://doi.org/10.1016/j.ijpharm.2021.121194 [56] k. shi, j.p. salvage, m. maniruzzaman, a. nokhodchi. role of release modifiers to modulate drug release from fused deposition modelling (fdm) 3d printed tablets. international journal of pharmaceutics 597 (2021) 120315. https://doi.org/10.1016/j.ijpharm.2021.120315 [57] l. krueger, j.a. miles, a. popat. 3d printing hybrid materials using fused deposition modelling for solid oral dosage forms. journal of controlled release 351 (2022) 444-455. https://doi.org/10.1016/j.jconrel.2022.09.032 [58] r. thakkar, a.r. pillai, j. zhang, y. zhang, v. kulkarni, m. maniruzzaman. novel on-demand 3dimensional (3-d) printed tablets using fill density as an effective release-controlling tool. polymers 12 (2020) 1872. https://doi.org/10.3390/polym12091872 [59] j. zhang, r. thakkar, y. zhang, m. maniruzzaman. structure-function correlation and personalized 3d printed tablets using a quality by design (qbd) approach. international journal of pharmaceutics 590 (2020) 119945. https://doi.org/10.1016/j.ijpharm.2020.119945 [60] a. goyanes, a.b. buanz, a.w. basit, s. gaisford. fused-filament 3d printing (3dp) for fabrication of tablets. international journal of pharmaceutics 476 (2014) 88-92. https://doi.org/10.1016/j.ijpharm.2014.09.044 [61] b. arafat, m. wojsz, a. isreb, r.t. forbes, m. isreb, w. ahmed, t. arafat, m.a. alhnan. tablet fragmentation without a disintegrant: a novel design approach for accelerating disintegration and drug release from 3d printed cellulosic tablets. european journal of pharmaceutical sciences 118 (2018) 191-199. https://doi.org/10.1016/j.ejps.2018.03.019 [62] m. sadia, b. arafat, w. ahmed, r.t. forbes, m.a. alhnan. channelled tablets: an innovative approach to accelerating drug release from 3d printed tablets. journal of controlled release 269 (2018) 355-363. https://doi.org/10.1016/j.jconrel.2017.11.022 [63] a. isreb, k. baj, m. wojsz, m. isreb, m. peak, m.a. alhnan. 3d printed oral theophylline doses with innovative ‘radiator-like’design: impact of polyethylene oxide (peo) molecular weight. international journal of pharmaceutics 564 (2019) 98-105. https://doi.org/10.1016/j.ijpharm.2019.04.017 [64] s.a. khaled, m.r. alexander, d.j. irvine, r.d. wildman, m.j. wallace, s. sharpe, j. yoo, c.j. roberts. extrusion 3d printing of paracetamol tablets from a single formulation with tunable release profiles through control of tablet geometry. aaps pharmscitech 19 (2018) 3403-3413. https://doi.org/10.1208/s12249-018-1107-z https://doi.org/10.1016/j.ijpharm.2021.120417 https://doi.org/10.1016/j.jddst.2022.103538 https://doi.org/10.1208/s12249-018-1233-7 https://doi.org/10.1016/j.ijpharm.2018.04.055 https://doi.org/10.1016/j.ijpharm.2022.121532 https://doi.org/10.1016/j.ijpharm.2022.121532 https://doi.org/10.1016/j.ijpharm.2021.121194 https://doi.org/10.1016/j.ijpharm.2021.120315 https://doi.org/10.1016/j.jconrel.2022.09.032 https://doi.org/10.1016/j.jconrel.2022.09.032 https://doi.org/10.3390/polym12091872 https://doi.org/10.1016/j.ijpharm.2020.119945 https://doi.org/10.1016/j.ijpharm.2014.09.044 https://doi.org/10.1016/j.ijpharm.2014.09.044 https://doi.org/10.1016/j.ejps.2018.03.019 https://doi.org/10.1016/j.jconrel.2017.11.022 https://doi.org/10.1016/j.ijpharm.2019.04.017 https://doi.org/10.1208/s12249-018-1107-z https://doi.org/10.1208/s12249-018-1107-z admet & dmpk 11(1) (2023) 33-55 development of rapidly dissolving fdm 3d-printed tablets doi: https://doi.org/10.5599/admet.1622 55 [65] a. goyanes, p.r. martinez, a. buanz, a.w. basit, s. gaisford. effect of geometry on drug release from 3d printed tablets. international journal of pharmaceutics 494 (2015) 657-663. https://doi.org/10.1016/j.ijpharm.2015.04.069 [66] s. palekar, p.k. nukala, s.m. mishra, t. kipping, k. patel. application of 3d printing technology and quality by design approach for development of age-appropriate pediatric formulation of baclofen. international journal of pharmaceutics 556 (2019) 106-116. https://doi.org/10.1016/j.ijpharm.2018.11.062 [67] v. raje, s. palekar, s. banella, k. patel. tunable drug release from fused deposition modelling (fdm) 3d-printed tablets fabricated using a novel extrudable polymer. pharmaceutics 14 (2022) 2192. https://doi.org/10.3390/pharmaceutics14102192 [68] a. melocchi, f. briatico-vangosa, m. uboldi, f. parietti, m. turchi, d. von zeppelin, a. maroni, l. zema, a. gazzaniga, a. zidan. quality considerations on the pharmaceutical applications of fused deposition modeling 3d printing. international journal of pharmaceutics 592 (2021) 119901. https://doi.org/10.1016/j.ijpharm.2020.119901 [69] t. parikh, h.k. sandhu, t.t. talele, a. serajuddin. characterization of solid dispersion of itraconazole prepared by solubilization in concentrated aqueous solutions of weak organic acids and drying. pharmaceutical research 33 (2016) 1456-1471. https://doi.org/10.1007/s11095-016-1890-8 [70] a. shah, a. serajuddin, supersolubilization by using nonsalt-forming acid-base interaction, in amorphous solid dispersions, springer2014, p. 595-611. [71] s. singh, t. parikh, h.k. sandhu, n.h. shah, a.w. malick, d. singhal, a. serajuddin. supersolubilization and amorphization of a model basic drug, haloperidol, by interaction with weak acids. pharmaceutical research 30 (2013) 1561-1573. https://doi.org/10.1007/s11095-013-0994-7 [72] a.t. serajuddin. salt formation to improve drug solubility. advanced drug delivery reviews 59 (2007) 603-616. https://doi.org/10.1016/j.addr.2007.05.010 [73] a.t. serajuddin, c.i. jarowski. effect of diffusion layer ph and solubility on the dissolution rate of pharmaceutical bases and their hydrochloride salts i: phenazopyridine. journal of pharmaceutical sciences 74 (1985) 142-147. https://doi.org/10.1002/jps.2600740208 [74] t. parikh, a. serajuddin. development of fast-dissolving amorphous solid dispersion of itraconazole by melt extrusion of its mixture with weak organic carboxylic acid and polymer. pharmaceutical research 35 (2018) 1-10. https://doi.org/10.1007/s11095-018-2407-4 [75] j. aho, j.p. bøtker, n. genina, m. edinger, l. arnfast, j. rantanen. roadmap to 3d-printed oral pharmaceutical dosage forms: feedstock filament properties and characterization for fused deposition modeling. journal of pharmaceutical sciences 108 (2019) 26-35. https://doi.org/10.1016/j.xphs.2018.11.012 [76] guidance for industry dissolution testing of immediate release solid oral dosage forms, https://www.fda.gov/media/70936/download (november 12, 2022). [77] j.b. dressman, g.l. amidon, c. reppas, v.p. shah. dissolution testing as a prognostic tool for oral drug absorption: immediate release dosage forms. pharmaceutical research 15 (1998) 11-22. https://doi.org/10.1023/a:1011984216775 ©2023 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) https://doi.org/10.5599/admet.1622 https://doi.org/10.1016/j.ijpharm.2015.04.069 https://doi.org/10.1016/j.ijpharm.2015.04.069 https://doi.org/10.1016/j.ijpharm.2018.11.062 https://doi.org/10.1016/j.ijpharm.2018.11.062 https://doi.org/10.3390/pharmaceutics14102192 https://doi.org/10.1016/j.ijpharm.2020.119901 https://doi.org/10.1016/j.ijpharm.2020.119901 https://doi.org/10.1007/s11095-016-1890-8 https://doi.org/10.1007/s11095-013-0994-7 https://doi.org/10.1016/j.addr.2007.05.010 https://doi.org/10.1002/jps.2600740208 https://doi.org/10.1007/s11095-018-2407-4 https://doi.org/10.1016/j.xphs.2018.11.012 https://doi.org/10.1016/j.xphs.2018.11.012 https://www.fda.gov/media/70936/download http://creativecommons.org/licenses/by/3.0/ development of conducting paper-based electrochemical biosensor for procalcitonin detection doi: https://doi.org/10.5599/admet.1575 263 admet & dmpk 11(2) (2023) 263-275; doi: https://doi.org/10.5599/admet.1575 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper development of conducting paper-based electrochemical biosensor for procalcitonin detection yachana gupta and aditya sharma ghrera* applied science department, the northcap university, huda-sector 23a, gurugram, india *corresponding author: adityasghrera@gmail.com received: october 28, 2022; revised: january 29, 2022; published: january 30, 2023 abstract in the present research, an advanced cellulose fiber paper (cfp) based biosensor is developed. this sensor is modified with nanocomposites containing poly(3,4-ethylene dioxythiophene) polystyrene sulfonate (pedot:pss) as the main matrix and functionalized gold nanoparticles (pedot:pss-aunp@cfp) for the selective and sensitive detection of bacterial infection (bi)-specific biomarker procalcitonin (pct). scanning electronic microscopy, fourier transform infrared spectroscopy, and x-ray diffraction are used to characterize the pedot:pss-aunp nanocomposite. this biosensor exhibits a high sensitivity of 1.34 μa (pg ml-1)-1 in the linear detection ranges of 1-20×104 pg ml-1, and a 24-day life span for pct antigen detection. anti-pct antigenic protein is used for immobilization for pct quantification. the results of electrochemical response studies showed that this conductive paper bioelectrode had good reproducibility, stability, and sensitivity in physiological ranges (1-20×104 pg ml-1). further, the proposed bioelectrode is an alternative choice for point-of-care pct detection. ©2023 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords gold nanoparticles; pedot; pedot:pss; procalcitonin; monoclonal antibodies; chronoamperometry introduction paper is excellent as a basis for manufacturing analytical sensors due to its flexibility, disposability, copiousness, and minimal cost [1-3]. the cellulose fiber paper (cfp) is strong, has a very good bonding ability, and its application has specific demands. due to its good absorption capacity, particles can be retained on the surface for a longer time, and as a substrate, it exhibits strong antigen-antibody interaction [4,5]. this cfp substrate is used in the development of various sensors and other research applications at the nanoscale, such as membrane affinity, tissue engineering, and assay development, because it has a large surface area, one-dimensional configuration, and good coordination with surface molecules [6-8]. currently, research is focused on improving the conductivity of paper substrates compared to other conventional substrates such as ceramics and glass. conducting polymers such as polyaniline [9,10], polypyrrole [11], poly(phenylenevinylene) [12], polyacetylene [13], polythiophene [14], poly(para-phenylene) [15], poly(3,4-ethylene dioxythiophene) polystyrene sulfonate (pedot:pss) [16,17] and polyfuran [18], etc. are used for surface modification. among the several conducting polymers, pedot:pss is the most appropriate and multipurpose https://doi.org/10.5599/admet.1575 https://doi.org/10.5599/admet.1575 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:adityasghrera@gmail.com http://creativecommons.org/licenses/by/4.0/ y. gupta and a.s. ghrera admet & dmpk 11(2) (2023) 263-275 264 due to its advantages of being film-forming ability, good thermal stability, homogeneous set-up on paper, mechanical flexibility, and exhibiting remarkable electrochemical properties in the development of paperbased biosensors [19-22]. in addition, solvent treatment of pedot:pss can promote rapid electron transfer between the paper surface and the solution, thereby increasing conductivity. conductive polymers, when integrated with gold nanoparticles (aunp), silver, and other metals and their oxide nanoparticles, etc., can improve electrochemical, optical, and physical properties [23-25]. various studies have used nanoparticles with pedot:pss to improve surface properties and electrical conductivity [26-28]. several other electrochemical biosensor sensors have been established for the detection of bacterial infection (bi)-specific biomarker procalcitonin (pct) [29-32]. pct is a biomarker specific for bi detection and has gained much attention as a potential solution to the problems associated with determining appropriate antibiotic use [33]. it has a molecular weight of 13 kda and is composed of 116 amino acid proteins. thyroid c cells produce the pct, which is then converted to calcitonin before entering the bloodstream. in children, pct levels can rise under any circumstances. pct is rarely detected in detectable quantities in the blood of healthy individuals. according to numerous studies, serum pct levels rise in individuals with sepsis and reduce after a successful diagnosis of bi with an antibiotic. thus, the quantitative assessment of pct levels earlier and afterward bi diagnosis may contribute to the early detection of this malignant infection. a variety of conductive polymer-modified electrochemical sensors have been reported for pct detection [34-38], but these approaches suffer from limitations due to multifaceted pre-treatment steps and the use of exorbitant electrodes. as a result, a facile, profligate, flexible, and economical substrate is required for pct detection. in the present study, an electrochemical biosensor is developed for the detection of bi-specific biomarker pct using cfp as the electrode platform, and for the first time, a composite of aunp and conductive polymer (pedot:pss) is proposed for coating the cellulose fiber paper and for the signal amplification. the dip coating method is used to fabricate cellulose paper substrate with pedot:pss-aunp composite. in this method, a composite of pedot:pss and aunp was prepared to increase the electrical conductivity of the paper electrode. the developed cfp-based sensor is inexpensive and biodegradable compared to other expensive substrates such as glassy carbon electrodes, ito, fto, etc. on the other hand, paper-based methods are being developed because they have various advantages such as affordability, are disposable, and require a small sample volume. pedot:pss-aunp@cfp shows good antibody-antigen interaction with a low detection limit of 1 pg ml-1 and sensitivity of 1.34 μa (pg ml-1)-1 in linear ranges of 1 -20×104 pg ml-1, respectively. this bioelectrode also shows good reproducibility and improved stability. in this research, a cost-effective, userfriendly, portable, and disposable pedot:pss-aunp composite-modified cfp-based biosensor is evaluated as a new approach for pct detection. using the chronoamperometry method, efforts were made to achieve a satisfactory response of the developed biosensor. experimental required materials and apparatus pedot:pss was purchased from sigma aldrich, and ethylene glycol was purchased from cdh chemicals. bovine albumin serum (bsa), auric chloride, trisodium citrate, monobasic sodium phosphate, and disodium phosphate were purchased from cdh chemicals. tween-80 was purchased from thomas baker (chemicals). monoclonal antibody (mab) was purchased from www.mybiosource.com. cellulose fiber paper (21×260mm) was procured with advanced micro devices (mdi) membrane. for electrochemical characterization, a threeelectrode autolab pgstat204 potentiostat/galvanostat (eco chemie, netherlands) was employed. in 0.1 m kcl containing 10 mm [fe(cn)6]3-/4-, a cellulose fiber paper strip was applied as a working electrode, a platinum foil http://www.mybiosource.com/ admet & dmpk 11(2) (2023) 263-275 electrochemical biosensor for procalcitonin doi: https://doi.org/10.5599/admet.1575 265 as a counter electrode, and ag/agcl as a reference electrode. field emission scanning electron microscope (fesem) model a jeol jsm-7610f plus was used to investigate the nanocomposite morphology. to ensure that the pedot:pss-aunp composite is formed, fourier transform infrared spectroscopy (ftir) (model-bruka alpha) and x-ray diffraction (xrd) (model-rigaku smart lab studio) techniques were used. a low current source and a digital microvoltmeter (model: dmv 1001) were used to assess conductivity using four-probe techniques (model: lcs 102). synthesis of pedot-aunp nanocomposite a colloid solution of spherical-sized aunp was prepared by the earlier-mentioned citrate method [39]. before the synthesis of pedot:pss-aunp nanocomposite, an aqueous suspension of 2.4 wt% pedot:pss and 5 % of ethylene glycol were prepared and stirred for 30 minutes. after that, this suspension was mixed with a colloidal solution of aunp (ratio 1:1) and left for ultrasonication for 1 h. the prepared pedot:pss-aunp composite was kept at 4 °c for further use. fabrication of cfp electrode before fabricating the cfp electrodes, cfp was cut into a (1×1 cm2) dimension and washed three times in ethanol and deionized water. thereafter, tween 80 surfactant was used for stabilization and dried at 37 °c for 24 hours. these stabilized electrodes were dipped in a pedot:pss-aunp composite previously synthesized. the pedot:pss-aunp with cfp electrodes were kept for ultrasonication for 40 minutes. the electrode's color change from white to black indicates that pedot: pss-aunp composite has been deposited on the paper. after that, the electrodes were dried in a hot air oven at 40 °c for 30 minutes. the electrodes that resulted were stored in the freezer for future use. the pedot:pss-aunp@cfp electrode is now referred to as a cp (conducting paper) electrode. antibody immobilization to immobilize the antibody onto the cp electrode, a monoclonal antibody (anti-pct) which is specific to pct was used. 10 µl of 0.08 mg ml-1 concentration of anti-pct in pbs (7.4) was dropped at the cp electrode and dried for 30 minutes. after that, 10 µl of 0.25 % blocking agent bsa was dropped over the cp electrode to block the unspecific sites. thereafter it was kept in the undisturbed position for 30 minutes. now, a prepared bsa/anti-pct/cp electrode was used for the electrochemical response studies. for the electrochemical response studies, 1 cm of prepared bsa/anti-pct/cp was dipped into a redox probe solution. to ensure the anti-pct immobilization on to cp surface, electrochemical characterization techniques and spectroscopic techniques were involved. the fabricated pct/ bsa/anti-pct/cp schematic has been demonstrated in figure 1. results and discussion spectroscopic characterization pre-prepared pedot:pss-aunp nanocomposite solution was analyzed through ftir measuring instrument. figures 2a and 2b depict ftir and xrd spectra of pedot:pss-aunp nanocomposite films. ftir spectroscopy provides data at the microscopic level, i.e., the changes that take place in the pedot:pss backbone as a result of the existence of charged nanoparticles in their surroundings. pedot:pss-aunp has wide-ranging and strong absorption peaks in the ftir spectrum at 3276, 2952, and 2884 cm-1 [40], respectively, correlating to the asymmetric stretching of hydroxyl (–oh), carbonyl (–c = o), and epoxy (c-o) groups [41]. the c–c stretching from pedot's thiophene ring is indicated by the peaks at 1337 cm−1 and 1416 cm−1 [42]. the peaks at 1646 and 1458 cm−1 are associated with c-c stretching in pss aromatic rings, respectively. the c–s stretching of the thiophene ring in pedot coincides with the additional peaks at 882, https://doi.org/10.5599/admet.1575 y. gupta and a.s. ghrera admet & dmpk 11(2) (2023) 263-275 266 669, 569, and 516 cm−1 [42,43]. peak shifts in ftir bands are caused by nanoparticles as a result of signal transduction in both the pedot and pss chains. when charged nanoparticles appear, electrostatic attraction between nanoparticles and pss occurs, reducing the interaction between pedot and pss. as a result, the validation of pedot and pss chains differs, resulting in improved connectivity. the crystalline phase of the developed pedot:pss-aunp composite was investigated using the x-ray diffraction (xrd) technique, and the resulting xrd patterns are displayed in figure 2. pedot:pss-aunp composite exhibited three distinct peaks at 2 = 12.6°, 22.3°, and 31.6°. all three peaks corresponded to standard bragg reflections (020), (011), (021), and monoclinic cubic lattice. because it follows seven crystal system a ≠ b ≠ c, α = γ = β ≠90°. the considerable diffraction at the 22.3° peak implies that zero-valent gold's optimal growth configuration was set to (011) [44]. solids with a molecular scale and a 3d pattern of atoms or molecules with an average crystal size of 0.371 nm are referred to as molecular solids [43]. figure 1. schematic diagram of the fabrication of conducting paper-based pct biosensor. fesem measurements demonstrate the morphologies and structural features of the as-synthesized pedot:pss-aunp nanocomposite. the cfp substrate was properly cleaned and dried several times. after that, the pedot: pss-aunp solution was dropped at the cfp surface and dried. the coating of cfp with the pedot: pss-aunp solution was performed four times. after the coated cfp substrate had been completely dried, it was used for fesem characterization. the sem image of bare cellulose fiber paper substrate and prepared pedot:pss-aunp nanocomposite is shown in figures 2c and 2d, respectively. through the fesem, it was found that pedot:pss-aunp nanocomposite was successfully deposited onto the cfp substrate and can be used for further studies. the energy dispersive x-ray analysis (edax) spectra of aunps incorporated pedot:pss composite is shown in figure. 2e. the edax spectra and elemental mapping of the fesem images show that aunps are distributed uniformly in the pedot:pss composite film. the results of the edax analysis of pedot: pss-aunp are shown in table. 1. the edax spectrum of pelletized pedot:pss-aunp with a carbon and oxygen peak is shown in figure 2(e), and other peaks in the spectrum correspond to sodium, sulphur, iron, and gold. sulphur and iron were present as a result of the use of pss and sodium due to citrate as admet & dmpk 11(2) (2023) 263-275 electrochemical biosensor for procalcitonin doi: https://doi.org/10.5599/admet.1575 267 oxidizing agents. aunp corresponds to au. in table 1 the c:o mass ratios in pedot:pss-aunp films were estimated to be 42.6 and 41.3 %, respectively. the atom ratios of c: o. pedot:pss-aunp composite films were estimated to be 56.2 and 40.9 %, respectively. figure 2. (a) and (b) ftir and xrd spectra of pedot:pss-aunp, (c) fesem image of bare cfp substrate, (d) fesem image of cp, and (e) edax analytical performance of cp. table 1. results for edax analysis of cp electrode element line content, wt.% content, at.% c k 42.6 56.2 o k 41.3 40.9 na k 1.4 0.9 s k 1.2 0.6 fe k 1.3 0.4 au l 12.3 0.1 electrochemical characterization electrochemical impedance spectroscopy (eis), cyclic voltammetry (cv), and chronoamperometry were used to characterize the cp electrode. 10 mm [fe(cn)6]3-/4with 0.1 m kcl solution was used to monitor changes at each step. eis is an effective approach for determining the intercalation resistance at the conductive interface and investigating electrodes with surface functionalization [45-47]. figure 3a displays nyquist plots of several modified electrodes, generated by graphing the imaginary part (z′′) on the y-axis and the real component (z′) on the x-axis. usually, the charge transfer resistance (rct) at the electrode surface is determined by the nyquist plot of the diameter of the semicircle part, which indicates the electron transfer kinetics at the interface. due to the large impedances and complex features of the impedance spectra shown in figure 3a, rct could not be accurately determined but roughly estimated by the non-linear least square fitting of the experimental results assuming randles circuit [rs(rctcdl)] of the electrochemical cell.the alteration of the electrode surface was revealed by the decrease or rise in rct values. the rct value of cp was roughly estimated to be 4220 ω (curve i), and after immobilization of anti-pct/cp, to 4335 ω, which is higher than the cp electrode. when the blocking agent was used rct value of bsa/anti-pct/cp further increased and reached to 4400 ω [48-50]. this could be due to bsa's insulating properties, which prevented electron transport at the https://doi.org/10.5599/admet.1575 y. gupta and a.s. ghrera admet & dmpk 11(2) (2023) 263-275 268 electrode-electrolyte interface. all of these comparative fluctuations in the rct values of dissimilar electrodes support electrode modification. the electrochemical behavior of the designed bioelectrode (bsa/anti-pct/cp) with varying pct concentrations was also observed using the cyclic voltammetry (cv) technique in the potential range of -0.8 to 0.4 v and at a scan rate of 50 mv s-1 (figure 3b). cv is an efficient and important method for visual observation of surface morphology and analyzing the integrations of electrode materials [51]. as a result, cv was selected to examine the modification in electrode behavior caused by electrode modification. figure 3b demonstrates the cv of (i) cp, (ii) anti-pct/cp, and (iii) bsa/anti-pct/cp. all cvs exhibit resistive behavior without current peaks at 50 mvs-1, indicating no fast charge transfer occurs in the system. the highest currents were registered on bare cp material (curve i). curve (ii) shows a decrease in the current values, which may be attributed to the fact that antibodies have been successfully immobilized on cp. the antigenantibody complexes so formed have produced additional barriers and prevent electron exchange between the working electrode and the solution, thereby decreasing the peak current values [52]. the currents of the bsa/anti-pct/cp bioelectrode is even lower after modification with shielding bsa as the obstructing agent for the non-specific signaling pathways. as a result of its electrical insulation, bsa hinders the transfer of electrons between the electrode and medium, resulting in a reduction in peak current value [52]. all of these variabilities in peak current value assist and verify the fabrication of electrodes with biomolecules. the chronoamperometric approach was used to analyze the electrochemical behavior of (i) cp, (ii) antipct/cp, and (iii) bsa/anti-pct/cp bioelectrodes at a potential of 0.20 v. the chronoamperometric plot for different modifications on cfp electrodes is shown in figure 3c and it was found that for all electrodes after the initial transient period of about 10 s, the current assumed a steady value. for the particular instance of solution-processed cp, the current was 9.7 μa [(curve (i)]. due to the macromolecule size of anti-pct, there was a slight decrease in current for the anti-pct/cp electrode (8.7 μa; curve ii). furthermore, for bsa/antipct/cp, decreases in electrochemical current (5.7 μa; curve iii) were observed. this is due to bsa's insulating properties, which impeded the electron transference mediator's entry into the substrate. a b c figure 3. electrochemical (a) eis, (b) cv and, (c) chronoamperometry studies of unmodified and modified bioelectrodes electrochemical response studies chronoamperometry is also known as potentiostatic coulometry. a steady voltage is given to the working electrode throughout this procedure, and the current is measured over time. the analyte is reduced or oxidized at a distinct oxidation potential. chronoamperometry experiments were conducted to evaluate the admet & dmpk 11(2) (2023) 263-275 electrochemical biosensor for procalcitonin doi: https://doi.org/10.5599/admet.1575 269 physical transformation per mole of electron flow at the electrode-electrolyte interaction over a brief time span at a constant potential. the minimum detection potential 0.20 v was used to produce electrochemical response. the electrochemical response of the bsa/anti-pct/cp bioelectrode was measured in 20 ml of 10 mm [fe(cn)6]3-/4in 0.1 m kcl solution for various concentrations of pct (1-20 ×104 pg ml-1) over a ten-minute incubation time. the chronoamperometry was conducted at 0.20 v for 140 seconds to ensure the antibodyantigen interaction was completed. figure 4a demonstrates the difference in the response current as a function of pct concentration (1-20×104 pg ml-1). for this objective, the bsa/anti-pct/cp bioelectrode was estimated with 10 μl of pct solution and it was found that when pct concentration rises, the response current also increases [6,53]. the calibration plot (figure 4b) was constructed in the 1-20×104 pg ml-1 ranges, which was produced by displaying the pct concentration vs. response current at 140 s. from this plot, it was observed that the response current is dependent on concentration. equation (1) describes how the current varies with concentration. the lod of the developed sensor was found to be 103 fg ml-1, and the linear correlation resulted in a regression coefficient of 0.99. i = 1.34 cpct + 1.72 (1) a b figure 4. (a) chronoamperometry response study of pct/ bsa/anti-pct/cp bioelectrode in 10 mm [fe(cn)6]3-/4-redox probe with 0.1 m kcl solution; (b) calibration plot of the bioelectrode between the response current and pct concentrations. reproducibility and stability to evaluate the reproducibility of the bsa/anti-pct/cp bioelectrode, five distinct electrodes were fabricated under the same optimized condition, and the biosensing characteristics were assessed using chronoamperometry. figure 5a demonstrates that there was a minimal difference in current for the five distinct electrodes, with a relative standard deviation of 3.94 % (mean value 5.84), implying that the bsa/anti-pct/cp bioelectrode had good reproducibility. the assessment of the developed pct biosensor's long-term stability is essential to the biosensor's applicability for practical applications. the response current of the fabricated bsa/anti-pct/cp bioelectrode was measured every six days to study the stability of the biosensor. the fabricated bsa/antipct/cp bioelectrode electrode maintained approximately 89.5 % of the initial current response after 24 days, demonstrating that the bsa/anti-pct/cp bioelectrode has good stability (figure 5b). this improved stability is due to aunp/cfp's strong affinity for pedot:pss, which is presented as a result of repeated pre-treatment. in comparison to previous biosensors for pct detection, this related work demonstrated (table 2) pedot:pss:aunp composite fabricated cfp substrate exhibited promising detection range, sensitivity, and enhanced shelf life https://doi.org/10.5599/admet.1575 y. gupta and a.s. ghrera admet & dmpk 11(2) (2023) 263-275 270 a b figure 5. (a) reproducibility (b) stability of the bsa/anti-pct/cp in 0.1 m kcl containing 10 mm [fe(cn)6]3-/ 4 table 2. comparison of the electrochemical sensing performance of different nanocomposites modified electrodes towards pct detection. modified electrode substrate detection method concentration, pg ml-1 ref. linear range detection limit g-co@ ncnbs gce one-pot solvothermal strategy and self-catalyzed chemical vapor deposition 0.1  100,000 0.01 [54] rgo-au gce electrodeposition 1.00 to 20,000 0.43 [29] pani nrs/rgo-au gce electrochemiluminescence 0.10 to 50,000 54,000 [55] gc/mwcnts, cs, ga gce electrochemical immunoassay 10 to 350,000 0.5 [56] pedot:pss-aunp paper dip coating 1 to 200,000 1 this work electrical conductivity studies the four-probe method was used to evaluate the conductivity of cf paper by employing a digital microvoltmeter and a low-current source. each modification step is measured to explore the conductivity of cf paper electrodes. the conductivity value of aunp coated on cf paper was 3.45×10-5 s cm-1, while the conductivity value of pedot:pss-ethylene glycol suspension coated cf paper was 2.57×10-4 s cm-1. the zerooverlap semimetal structure of pedot:pss, which contains electrons and holes as charge carriers, accounts for its high electrical conductivity. furthermore, the pedot:pss-aunp composite was coated on paper to increase conductivity, which was found to be 1.09×10-3 s cm-1. pedot:pss-aunp can facilitate electron transfer between electroactive species and electrode substrates, indicating that the integrated platform can potentially further promote bioelectrode conductivity. conclusion the present article describes a facile strategy to detect bi-specific biomarker pct. the pedot:pss-aunp nanocomposite was used to fabricate a cfp-based electrochemical sensor. for the deposition of pedot:pss/aunp, cfp substrate is considered because of its good absorption capacity, particle retaining properties, cost-effectiveness, portable, disposable, and user friendly. modifications in the cfp surface were investigated using electrochemical characterization. deposition of pedot:pss-aunp composite on the cfp surface was studied using fesem and edax spectroscopy. through the xrd analysis, crystalline size was monitored and with ftir studies presence of functional groups was investigated. synergistic interaction of admet & dmpk 11(2) (2023) 263-275 electrochemical biosensor for procalcitonin doi: https://doi.org/10.5599/admet.1575 271 aunp with pedot:pss could lead to significantly improved sensitivity and lower electrode overpotentials. additionally, the pedot:pss/aunp modified cfp shows enhanced electrochemical properties rather than unmodified electrodes. the modified pct/bsa/anti-pct/cp bioelectrode demonstrated an excellent detection limit of detection (1 pg ml-1) with a linear detection range (1-20×104 pg ml-1). a-prepared bioelectrode exhibits excellent reproducibility and high stability for at least four weeks, with negligible change in electrochemical response. furthermore, a sensor of this type may also be realistically designed for real-time sample analysis with adequate sensitivity and efficiency. this implies that our well-known electrochemical biosensor would be ideal for bioanalytical research. conflict of interest: the authors declare no conflict of interest. acknowledgment: a.s. ghrera thanks financial support from science and engineering board (dst), india, under the young scientist project (yss/2015/001330). the authors wish to thank and acknowledge manipal university, jaipur, for the fesem, edax, xrd and ftir characterization and acknowledge delhi technological university for conductivity extents. references [1] a.t. singh, d. lantigua, a. meka, s. taing, m. pandher and g. camci-unal. paper-based sensors: emerging themes and applications. sensors (basel) 9 (2018) 2838. https://doi.org/10.3390/s18092838 [2] y. yang, e. noviana, m.p. nguyen, b.j. geiss, d.s. dandy and c.s. henry, paper-based microfluidic devices: emerging themes and applications. analytical chemistry 89 (2017) 71-91. https://doi.org/ 10.1021/acs.analchem.6b04581 [3] y. gupta, a.s. ghrera. disposable paper-based biosensing platform for procalcitonin detection. anaytical bioanalytical chemistry reearch 10 (2022) 15-23. https://doi.org/10.22036/abcr.2022.34 0432.1761 [4] y. peng, v.v. gelder, a. amaladoss, k.h. patel. covalent binding of antibodies to cellulose paper discs and their applications in naked-eye colorimetric immunoassays. journal of visual experiment, 54111 (2016). https://doi.org/10.3791/54111 [5] n. sahraei, m. mazloum-ardakani, and f. hoseynidokht. electrochemical paper-based biosensors for point-of-care diagnostics: detection methods and applications. journal of electrochemical science and engineering 12 (2022) 399-419. https://doi.org/10.5599/jese.1104 [6] d. chauhan, p.r. solanki. hydrophilic and insoluble electrospun cellulose acetate fiber-based biosensing platform for 25-hydroxy vitamin-d3 detection. acs appl. polym. mater. 1 (2019) 16131623. https://doi.org/10.1021/acsapm.9b00179 [7] g. wang, g. jiang, y. zhu, w. cheng, k. cao, j. zhou, h. lei, g. xu and d. zhao. developing cellulosic functional materials from multi-scale strategy and applications in flexible bioelectronic devices. carbohydrate polymer 283 (2022) 119160. https://doi.org/10.1016/j.carbpol.2022.119160 [8] s. ge, l. zhang, y. zhang, f. lan, m. yan, j. yu. nanomaterials-modified cellulose paper as a platform for biosensing applications. nanoscale 9 (2017) 4366-4382. https://doi.org/10.1039/c6nr08846e [9] n. ruecha, r. rangkupan, n. rodthongkum, o. chailapakul. novel paper-based cholesterol biosensor using graphene/polyvinylpyrrolidone/polyaniline nanocomposite. biosensor bioelectronics 52 (2014) 13-19. https://doi.org/10.1016/j.bios.2013.08.018 [10] n. ruecha, k. shin, o. chailapakul, n. rodthongkum. label-free paper-based electrochemical impedance immunosensor for human interferon gamma detection. sensor and. actuators b: chemical 279 (2019) 298-304. https://doi.org/10.1016/j.snb.2018.10.024 https://doi.org/10.5599/admet.1575 https://doi.org/10.3390/s18092838 https://doi.org/‌10.1021/acs.analchem.6b04581 https://doi.org/‌10.1021/acs.analchem.6b04581 https://doi.org/10.22036/abcr.2022.340432.1761 https://doi.org/10.22036/abcr.2022.340432.1761 https://doi.org/10.3791/54111 https://doi.org/10.5599/jese.1104 https://doi.org/10.1021/acsapm.9b00179 https://doi.org/10.1016/j.carbpol.2022.119160 https://doi.org/10.1039/c6nr08846e https://doi.org/10.1016/j.bios.2013.08.018 https://doi.org/10.1016/j.snb.2018.10.024 y. gupta and a.s. ghrera admet & dmpk 11(2) (2023) 263-275 272 [11] o. hamouma, n. kaur, d. oukil, a. mahajan, m.m. chehimi. paper strips coated with polypyrrolewrapped carbon nanotube composites for chemi-resistive gas sensing. synthetic metals 258 (2019) 116223. https://doi.org/10.1016/j.synthmet.2019.116223 [12] j. banerjee, k. dutta. a short overview on the synthesis, properties and major applications of poly(pphenylene vinylene. chemical papers 75 (2021) 5139-5151. https://doi.org/10.1007/s11696-020 01492-9 [13] y. hu, z. zhao, x. bai, x. yuan, x. zhang, t. masuda. organoborane-containing polyacetylene derivatives: synthesis, characterization, and fluoride-sensing properties. rsc advances 98 (2014) 55179-55186. https://doi.org/10.1039/c4ra11062e [14] r. ramkumar, m.v. sangaranarayanan. electrochemical sensing of anesthetics using polythiophene coated glassy carbon electrodes. chemistry select 4 (2019) 9776-9783. https://doi.org/10.1002/ slct.201901982 [15] z. hsine, s. blili, r. milka, h. dorizon, a.h. said, h. korri-youssoufi. sensor based on redox conjugated poly(para-phenylene) for the simultaneous detection of dopamine, ascorbic acid, and uric acid in human serum sample. analytical bioanalytical chemistry 412 (2020) 4433-4446. https://doi.org/ 10.1007/s00216-020-02686-6 [16] y. wang, m. mukaida, k. kirihara, l. lyu and q. wei. poly(3,4-ethylene dioxythiophene)/poly(styrene sulfonate) electrodes in electrochemical cells for harvesting waste heat. energytechnology 8 (2020) 1900998. https://doi.org/10.1002/ente.201900998. [17] y. tangal, d. coban, s. cogal. a wse2@ poly(3, 4-ethylenedioxythiophene) nanocomposite-based electrochemical sensor for simultaneous detection of dopamine and uric acid. journal of electrochemical science and engineering 12 (2022) 1251-1259. https://doi.org/10.5599/jese.1375 [18] l.g. lascane, e.f. oliveira, d.s. galvão, a. batagin-neto. polyfuran-based chemical sensors: identification of promising derivatives via dft calculations and fully atomistic reactive molecular dynamics. europian polymer journal 141 (2020) 110085. https://doi.org/10.1016/j.eurpolymj.2020.110085 [19] y.-k. yen, c.-h. chao and y.-s. yeh. a graphene-pedot:pss modified paper-based aptasensor for electrochemical impedance spectroscopy detection of tumor marker. sensors 20 (2020) 1372. https://doi.org/10.3390/s20051372 [20] s. kumar, m. umar, a. saifi, s. kumar, s. augustine, s. srivastava, b.d. malhotra. electrochemical paper based cancer biosensor using iron oxide nanoparticles decorated pedot:pss. analytica chimica acta 1056 (2019) 135-145. https://doi.org/10.1016/j.aca.2018.12.053 [21] s. kumar, s. kumar, c.m. pandey, b.d. malhotra. conducting paper based sensor for cancer biomarker detection. journal of physics: conference series 704 (2016) 012010. http://doi.org/10.1088/1742 6596/704/1/012010 [22] t. beduk, e. bihar, s.g. surya, a.n. castillo, s. inal, k.n. salama. a paper-based inkjet-printed pedot:pss/zno sol-gel hydrazine sensor. sensor and actuators b: chemical 306 (2020) 127539. https://doi.org/10.1016/j.snb.2019.127539 [23] i.s. kucherenko, o.o. soldatkin, d.y. kucherenko, o.v. soldatkina, s.v. dzyadevych. advances in nanomaterial application in enzyme-based electrochemical biosensors: a review. nanoscale advances 1 (2019) 4560-4577. https://doi.org/10.1039/c9na00491b. [24] c. garcia-cabezon, c. salvo-comino, c. garcia-hernandez, m.l. rodriguez-mendez, f. martin-pedrosa. nanocomposites of conductive polymers and nanoparticles deposited on porous material as a strategy to improve its corrosion resistance. surface coating technology 403 (2020) 126395. https://doi.org/ 10.1016/j.surfcoat.2020.126395 [25] y. gupta, kalpana, a.s. ghrera. electrochemical studies of lateral flow assay test results for procalcitonin detection. journal of electrochemical science and engineering 12 (2022) 265-274. https://doi.org/10.5599/jese.1127 [26] k. muro, m. watanabe, t. tamai, k. yazawa, k. matsukawa. pedot/pss nanoparticles: synthesis and properties. rsc advances 90 (2016) 87147-87152. https://doi.org/10.1039/c6ra16829a https://doi.org/10.1016/j.synthmet.2019.116223 https://doi.org/10.1007/s11696-020‌01492-9 https://doi.org/10.1007/s11696-020‌01492-9 https://doi.org/10.1039/c4ra11062e https://doi.org/10.1002/‌slct.201901982 https://doi.org/10.1002/‌slct.201901982 https://doi.org/10.1002/ente.201900998 https://doi.org/10.5599/jese.1375 https://doi.org/10.3390/s20051372 https://doi.org/10.1016/j.aca.2018.12.053 http://doi.org/10.1088/1742‌6596/704/1/012010 http://doi.org/10.1088/1742‌6596/704/1/012010 https://doi.org/10.1016/j.snb.2019.127539 https://doi.org/10.1039/c9na00491b https://doi.org/10.1016/j.surfcoat.2020.126395 https://doi.org/10.1016/j.surfcoat.2020.126395 https://doi.org/10.5599/jese.1127 https://doi.org/10.1039/c6ra16829a admet & dmpk 11(2) (2023) 263-275 electrochemical biosensor for procalcitonin doi: https://doi.org/10.5599/admet.1575 273 [27] a. vázquez-lópez, a. yaseen, d. maestre, j. ramírez-castellanos, e.s. marstein, s.z. karazhanov, a. cremades. synergetic improvement of stability and conductivity of hybrid composites formed by pedot:pss and sno nanoparticles. molecules 25 (2020) 695. https://doi.org/10.3390/molecules 25030695 [28] j. dong, d. gerlach, p. koutsogiannis, p. rudolf, g. portale. boosting the thermoelectric properties of pedot:pss via low-impact deposition of tin oxide nanoparticles. adv. electron. mat. 7 (2021) 2001284. https://doi.org/10.1002/aelm.202001284 [29] f. liu, g. xiang, r. yuan, x. chen, f. luo, d. jiang, s. huang, y. li, x. pu. procalcitonin sensitive detection based on graphene–gold nanocomposite film sensor platform and single-walled carbon nanohorns/hollow pt chains complex as signal tags. biosensors bioelectronics 142 (2019) 1115322. https://doi.org/10.1016/j.bios.2019.111532 [30] a.s. tanak, b. jagannath, y. tamrakar, s. muthukumar, s. prasad. non-faradaic electrochemical impedimetric profiling of procalcitonin and c-reactive protein as a dual marker biosensor for early sepsis detection. analytica chimica acta 3 (2019) 100029. https://doi.org/10.1016/j.acax.2019.100029 [31] á. molinero-fernández, m. moreno-guzmán, m.á. lópez, a. escarpa, magnetic bead-based electrochemical immunoassays on-drop and on-chip for procalcitonin determination: disposable tools for clinical sepsis diagnosis. biosensors 10 (2020) 66. https://doi.org/10.3390/bios10060066 [32] p. seshadri, k. manoli, n. schneiderhan-marra, u. anthes, p. wierzchowiec, k. bonrad, c. di franco, l. torsi. low-picomolar, label-free procalcitonin analytical detection with an electrolyte-gated organic field-effect transistor based electronic immunosensor. biosensensor bioelectronics 104 (2018) 113119. https://doi.org/10.1016/j.bios.2017.12.041 [33] y. gupta, a.s. ghrera. recent advances in gold nanoparticle-based lateral flow immunoassay for the detection of bacterial infection. archives of microbiology 203 (2021) 3767-3784. https://doi.org/ 10.1007/s00203-021-02357-9 [34] f. battaglia, v. baldoneschi, v. meucci, l. intorre, m. minunni and s. scarano. detection of canine and equine procalcitonin for sepsis diagnosis in veterinary clinic by the development of novel mip-based spr biosensors. talanta 230 (2021) 122347. https://doi.org/10.1016/j.talanta.2021.122347 [35] á. molinero-fernández, m. moreno-guzmán, l. arruza, m.á. lópez, a. escarpa. polymer-based micromotor fluorescence immunoassay for on-the-move sensitive procalcitonin determination in very low birth weight infants’ plasma. acs sensors 5 (2020) 1336-1344. https://doi.org/10.1021/ acssensors.9b02515 [36] e. murugan, a. dhamodharan. separate and simultaneous determination of vanillin, theophylline and caffeine using molybdenum disulfide embedded polyaniline/graphitic carbon nitrite nanocomposite modified glassy carbon electrode. diamond and related materials 120 (2021) 108684. https://doi.org/ 10.1016/j.diamond.2021.108684 [37] k. sadani, l. muthuraj, p. nag, m. fernandes, k. kondabagil, c. mukhopadhyay, s. mukherji. a point of use sensor assay for detecting purely viral versus viral-bacterial samples. sensor actuators b: chemical 322 (2020 )128562. https://doi.org/10.1016/j.snb.2020.128562 [38] h. ding, l. yang, h. jia, d. fan, y. zhang, x. sun, q. wei, h. ju. label-free electrochemical immunosensor with palladium nanoparticles functionalized mos2/nico heterostructures for sensitive procalcitonin detection. sensors actuators b: chemical 312 (2020) 1279809. https://doi.org/10.1016/j.snb.2020.12 7980 [39] gupta, y., k. ladi, and a. ghrera, electrochemical studies of lateral flow assay test results for procalcitonin detection: original scientific paper. journal of electrochemical science and engineering, 11(2021) https://doi.org/10.5599/jese.1127 [40] j.v. piovesan, e.r. santana, a. spinelli. reduced graphene oxide/gold nanoparticles nanocompositemodified glassy carbon electrode for determination of endocrine disruptor methylparaben. journal of electroanalytical chemistry 813 (2018) 163-170. https://doi.org/10.1016/j.jelechem.2018.02.025 [41] b. amanulla, s. palanisamy, s.-m. chen, t.-w. chiu, v. velusamy, j.m. hall, t.-w. chen, s.k. ramaraj. selective colorimetric detection of nitrite in water using chitosan stabilized gold nanoparticles https://doi.org/10.5599/admet.1575 https://doi.org/10.3390/‌molecu‌les‌25030695 https://doi.org/10.3390/‌molecu‌les‌25030695 https://doi.org/10.1002/aelm.202001284 https://doi.org/10.1016/j.bios.2019.111532 https://doi.org/10.3390/bios10060066 https://doi.org/10.1016/j.bios.2017.12.041 https://doi.org/‌10.1007/s00203-021-02357-9 https://doi.org/‌10.1007/s00203-021-02357-9 https://doi.org/10.1016/j.talanta.2021.122347 https://doi.org/10.1021/‌acssensors.9b02515 https://doi.org/10.1021/‌acssensors.9b02515 https://doi.org/10.1016/j.diamond.2021.108684 https://doi.org/10.1016/j.diamond.2021.108684 https://doi.org/10.1016/j.snb.2020.128562 https://doi.org/10.1016/‌j.snb.20‌20.12‌7980 https://doi.org/10.1016/‌j.snb.20‌20.12‌7980 https://doi.org/10.5599/jese.1127 https://doi.org/10.1016/j.jelechem.2018.02.025 y. gupta and a.s. ghrera admet & dmpk 11(2) (2023) 263-275 274 decorated reduced graphene oxide. scientific report 7 (2017) 14182. https://doi.org/10.1038/s41598-017-14584-6 [42] c. sriprachuabwong, c. karuwan, a. wisitsorrat, d. phokharatkul, t. lomas, p. sritongkham, a. tuantranont. inkjet-printed graphene-pedot:pss modified screen printed carbon electrode for biochemical sensing. journal of material chemistry 22 (2012) 5478-5485. https://doi.org/10.1039/ c2jm14005e [43] p. sakunpongpitiporn, k. phasuksom, n. paradee, a. sirivat. facile synthesis of highly conductive pedot:pss via surfactant templates. rsc advances 9 (2019) 6363-6378. https://doi.org/10.1039/ c8ra08801b [44] s. krishnamurthy, a. esterle, n.c. sharma, s.v. sahi. yucca-derived synthesis of gold nanomaterial and their catalytic potential. nanoscale research. letter 9 (2014) 627. https://doi.org/10.1186/1556-276x 9-627 [45] n. meddings, m. heinrich, f. overney, j.-s. lee, v. ruiz, e. napolitano, s. seitz, g. hinds, r. raccichini, m. gaberšček, j. park. application of electrochemical impedance spectroscopy to commercial li-ion cells: a review. journal of power sources 480 (2020) 228742. https://doi.org/10.1016/j.jpowsour.2020.228742 [46] t. pajkossy, r. jurczakowski. electrochemical impedance spectroscopy in interfacial studies. curr. op. electrochem. 1 (2017) 53-58. https://doi.org/10.1016/j.coelec.2017.01.006 [47] j. moškon, s. drvarič talian, r. dominko, m. gaberšček. advances in understanding li battery mechanisms using impedance spectroscopy. journal of electrochemical science and engineering, 10 (2020) 79-93 https://doi.org/10.5599/jese.734. [48] j. narang, n. malhotra, c. singhal, a. mathur, d. chakraborty, a. anil, a. ingle, c.s. pundir. point of care with micro fluidic paper based device integrated with nano zeolite-graphene oxide nanoflakes for electrochemical sensing of ketamine. biosensor bioelectronics 88 (2017) 249-257. https://doi.org/ 10.1016/j.bios.2016.08.043 [49] h. li, s. song, m. wen, t. bao, z. wu, h. xiong, x. zhang, w. wen, s. wang. a novel label-free electrochemical impedance aptasensor for highly sensitive detection of human interferon-gamma based on target-induced exonuclease inhibition. biosensor bioelectronics 142 (2019) 111532. https://doi.org/ 10.1016/j.bios.2019.111532 [50] s. kumar, a. sen, s. kumar, s. augustine, b.k. yadav, s. mishra, b.d. malhotra. polyaniline modified flexible conducting paper for cancer detection. applied physcis letters 108 (2016) 203702. https://doi.org/10.1063/1.4950961 [51] j.a. buledi, s. ameen, n.h. khand, a.r. solangi, i.h. taqvi, m.h. agheem, z. wajdan. cuo nanostructures based electrochemical sensor for simultaneous determination of hydroquinone and ascorbic acid. electroanalysis 32 (2020) 1600-1607. https://doi.org/10.1002/elan.202000083 [52] p. lv, l. min, r. yuan, y. chai, s. chen. a novel immunosensor for carcinoembryonic antigen based on poly(diallyldimethylammonium chloride) protected prussian blue nanoparticles and double-layer nanometer-sized gold particles. microchimica acta 171 (2010) 297-304. https://doi.org/10.1007/ s00604-010-0435-9 [53] s. nantaphol, o. chailapakul, w. siangproh. sensitive and selective electrochemical sensor using silver nanoparticles modified glassy carbon electrode for determination of cholesterol in bovine serum, sensors and actuators b chemical 207 (2015) 193-198. https://doi.org/10.1016/j.snb.2014.10.041 [54] x.y. ge, j.x. zhang, y.g. feng, a.j. wang, l.p. mei, j.j. feng. label-free electrochemical biosensor for determination of procalcitonin based on graphene-wrapped co nanoparticles encapsulated in carbon nanobrushes coupled with auptcu nanodendrites. microchimica acta 189 (2022) 110. https://doi.org/ 10.1007/s00604-022-05179-8 [55] l. yang, j. xue, y. jia, y. zhang, d. wu, h. ma, q. wei, h. ju. construction of well-ordered electrochemiluminescence sensing interface using peptide-based specific antibody immobilizer and n(aminobutyl)-n-(ethylisoluminol) functionalized ferritin as signal indicator for procalcitonin analysis. biosensors bioelectronics 142 (2019) 111562. https://doi.org/10.1016/j.bios.2019.111562 https://doi.org/10.1038/s41598-017-14584-6 https://doi.org/10.1038/s41598-017-14584-6 https://doi.org/10.1039/‌c2jm14005e https://doi.org/10.1039/‌c2jm14005e https://doi.org/10.1039/c8ra08801b https://doi.org/10.1039/c8ra08801b https://doi.org/10.1186/1556-276x‌9-627 https://doi.org/10.1186/1556-276x‌9-627 https://doi.org/10.1016/j.jpowsour.2020.228742 https://doi.org/10.1016/j.jpowsour.2020.228742 https://doi.org/10.1016/j.coelec.2017.01.006 https://doi.org/10.5599/jese.734 https://doi.org/10.1016/j.bios.2016.08.043 https://doi.org/10.1016/j.bios.2016.08.043 https://doi.org/10.1016/j.bios.2019.111532 https://doi.org/10.1016/j.bios.2019.111532 https://doi.org/10.1063/1.4950961 https://doi.org/10.1002/elan.202000083 https://doi.org/10.1007/‌s00604-010-0435-9 https://doi.org/10.1007/‌s00604-010-0435-9 https://doi.org/10.1016/j.snb.2014.10.041 https://doi.org/10.1007/s00604-022-05179-8 https://doi.org/10.1007/s00604-022-05179-8 https://doi.org/10.1016/j.bios.2019.111562 admet & dmpk 11(2) (2023) 263-275 electrochemical biosensor for procalcitonin doi: https://doi.org/10.5599/admet.1575 275 [56] y.s. fang, h.y. wang, l.s. wang, j.f. wang. electrochemical immunoassay for procalcitonin antigen detection based on signal amplification strategy of multiple nanocomposites. biosensors & bioelectronics 51 (2014) 310-316. https://doi.org/10.1016/j.bios.2013.07.035 ©2023 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) https://doi.org/10.5599/admet.1575 https://doi.org/10.1016/j.bios.2013.07.035 http://creativecommons.org/licenses/by/3.0/ manuscript doi: 10.5599/admet.3.3.208 274 admet & dmpk 3(3) (2015) 274-280; doi: 10.5599/admet.3.3.208 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper specific frequency bands of amplitude low-frequency fluctuations in memory-related cognitive impairment: predicting alzheimer’s disease yin tian 1 *, zechao ding 1 , kin yip tam 3 , zhongyan wang 1 , huiling zhang 1 , dechun zhao 1 , yi zhao 1 , wei xu 1 , and shuxing zheng 2 ; for the alzheimer’s disease neuroimaging initiative ζ 1 bio-information college, chongqing university of posts and telecommunications, chongqing 400065, china 2 colleg of computer science and technology, chongqing university of posts and telecommunications, chongqing 400065, china 3 faculty of health sciences, university of macau, macau, china ζ it is noted that the data used in this article were obtained from the alzheimer’s disease neuroimaging initiative (adni) database (adni.loni.ucla.edu). as such, the investigators within the adni contributed to the design and implementation of adni and/or provided data but did not participate in analysis or writing of this report. a complete listing of adni investigators can be found at: http://adni.loni.ucla.edu/wp-content/uploads/how to apply/adni acknowledgement list.pdf. *corresponding author: e-mail: tiany20032003@163.com; tel.: +86-23-62460536; fax: +86-23-62460536 received: august 09, 2015; revised: august 12, 2015; published: september 05, 2015 abstract resting-state functional magnetic resonance imaging was utilized to measure the amplitude low frequency fluctuations (alff) in human subjects with alzheimer’s disease (ad) and normal control (nc). two specific frequency bands (slow5: 0.01-0.027 hz and slow4: 0.027-0.073 hz) were analysed in the main cognitive control related four subregions of the right ventral lateral prefrontal cortex (vlpfc), i.e. ifj, posteriorvlpfc, mid-vlpfc, and anterior-vlpfc. differences in alff values between the ad and the nc group were found throughout the subregions of the right vlpfc. compared to normal control group, decreased alff values were observed in ad patients in the ifj (in two given frequency bands), and the mid-vlpfc (in slow5). in contrast, increased alff valued were observed in ad patients in the posteriorand anteriorvlpfc (in both slow5 and slow4), and also in the mid-vlpfc in slow4. moreover, significant alff differences between the ifj and three other subregions of the right vlpfc were found. furthermore, alff values in the right vlpfc showed significant correlations with the time course of disease. taken together, our findings suggest that ad patients have largely abnormalities in intrinsic neural oscillations which are in line with the ad pathophysiology, and further reveal that the abnormalities are dependent on specific frequency bands. thus, frequency-domain analyses of the alff may provide a useful tool to investigate the ad pathophysiology. keywords ad; alff; vlpfc; rs-fmri; top-down modulation introduction alzheimer’s disease (ad) is a neurodegenerative disorder leading to cognitive deficits. though the amyloid-β(aβ) peptides plague aggregation and tau hyperphosphorylation are thought to be two main http://www.pub.iapchem.org/ojs/index.php/admet/index admet & dmpk 3(3) (2015) 274-280 alffs of vlpfc predict ad doi: 10.5599/admet.3.3.208 275 causal roles in the pathogenesis of ad [1-2], the underlying neural mechanisms remain uncovered. previous neural imaging studies using positron emission tomography (pet) and single-photon emission (spe) have found that ad patients show abnormally low resting cerebral blood flow and a decreased cerebral metabolic rate for glucose in the widespread brain regions, including the temporal, parietal, and prefrontal cortex [3-6]. resting-state functional magnetic resonance imaging (rs-fmri) has also been considered for development as a biomarker and analytical tool for brain diseases evaluation such as ad and pd. the spontaneous low-frequency fluctuations of the blood oxygen level dependent (bold) in rsfmri were thought to reflect spontaneous neuronal activity [7-9]. the amplitude of low frequency fluctuations (alff) of bold signals, which were believed to be physiologically meaningful, was proposed to assess the amplitude of resting state spontaneous brain activity [10]. using resting-state spontaneous brain activity technology, researchers found that ad patients were abnormalities of lff in many brain regions including the hippocampus, the medial and lateral temporal lobes, the medial parietal lobe, the posterior cingulated cortex (pcc) and the medial and superior prefrontal cortex (mpfc) [11-15]. the function of the right ventral lateral prefrontal cortex (vlpfc) is crucial to the goal-directed cognition (i.e., top-down control), involving attention re/orienting, cognitive control, working memory, motor inhibition, and action updating [16-19]. more research evidence indicated that the right vlpfc could be divided into four subregions according to their cognitive function. however, few researches emphasize the changes of the vlpfc in ad patients though ad related disease could lead to learning and memory impairments, aprosexia, and executive dysfunction. in this study, we proposed that individual differences in ad might be captured based on spontaneous neuronal activity in the functional subregions of the vlpfc related to cognitive impairment identified by the alff of bold signals. gray matter (gm) related fluctuation amplitudes mainly occurred in two frequency bands, i.e. 0.01-0.027 hz (slow5) and 0.0270.073 hz (slow4). and these two given frequency bands contribute differently to the alff [12,20-21], suggesting that individual frequency bands could link to specific characteristics [22]. to validate the hypothesis underlying ad-related alff, the characteristic patterns of the changes of alff in ad patients will be compared with normal control subjects, especially slow5 and slow4. we aim to evaluate if the alff techniques could help us to reveal this intrinsic relationship, which may provide a foundation for further investigations of neural cognitive processes in ad. experimental methods subjects twenty two right-handed subjects with ad group (8 males and 3 females, mean±sd = 69.7±2.4 years) and nc group (11 males mean±sd=48.5±8.4 years) were selected in the current study. all subject data were obtained from the alzheimer’s disease neuroimaging initiative (adni, http:// www.loni.ucla.edu/adni) data acquisition and analysis three-dimensional (3d)-t1 was acquired using philips healthcare 3t scanner. high-resolution volumetric t1-weighted spoiled gradient-recalled (spgr) images were generated for each subject (tr=3000ms, te=30 ms, flip angle=80°, slice thickness=3.3 mm, fov=24×24 cm 2 , 64×64 matrix size, 6720 slices). data preprocessing were carried out using spm8 (http://www.fil.ion.ucl.ac.uk/spm), which included acquisition times correction, motion correction, image segmentation, and spatially smooth. then dparsf http://www.fil.ion.ucl.ac.uk/spm y. tian et al. admet & dmpk 3(3) (2015) 274-280 276 [23] was used to analyses the alff. briefly, we converted data from time-domain to frequency-domain using a fast fourier transform (fft). the square root of the power spectrum was computed and then averaged frequency ranged from 0.01 hz to 0.08 hz at each voxel. slow5 (0.01-0.027 hz) and slow4 (0.0270.073 hz) were selected to analyze. this averaged square root was taken as the alff. to reduce the global effects of variability across participants, the alff of each voxel was divided by the global mean alff value for each subject, resulting in a relative alff. figure 1. the t-score alff maps of ad and nc groups in rois. rois: regions of interesting; ad: alzheimer’s disease; nc: normal control; vlpfc: ventral lateral prefrontal cortex; a: posterior-vlpfc; b: middle-vlpfc; c: anterior-vlpfc; and d: ifj, inferior frontal junction. to explore the within-group alff patterns, one-sample t-test within the group were performed on the individual alff maps in a voxel-wise way for each group. two-sample t-test among groups, i.e. ad group vs. normal control (nc) group, was conducted. pearson’s correlation coefficient was implemented to analyze the correlations among the mean alff and the time course of disease with patients. the statistical threshold was set at p < 0.05. for regions of interest (rois), i.e. the ventrolateral prefrontal cortex (vlpfc) was divided four functional subregions. posterior region is the most caudal extent as shown in fig. 1-region a. middle region is rostral to the ascending ramus as shown in fig. 1-region b. anterior region is the horizontal ramus of the lateral sulcus as shown in fig. 1-region c. the most posterior and superior region is referred to the inferior frontal junction (ifj) as shown in fig. 1-region d. results and discussion figure 1 shows the alff maps of ad and nc groups in rois, i.e. the four functional subregions in the right ventral lateral prefrontal cortex. the detail information of regions-related mni coordinates is shown in table 1. the alff values of each subject including both ad (red circles) and nc (green circles) were shown in figure 2. a multi-anova statistical analysis within factors (regions: posterior-vlpfc vs. midvlpfc vs. anterior-vlpfc vs. ifj; frequency band: slow5 vs. slow4) and between groups (ad vs. nc) was admet & dmpk 3(3) (2015) 274-280 alffs of vlpfc predict ad doi: 10.5599/admet.3.3.208 277 conducted. it was found that only the main effect of regions was marginal significant (f=2.722, p=0.069, greeenhour-geisser correction). further t-test measurement showed that the alff in the ifj was significantly different in the other regions (posterior-vlpfc: p=0.025; mid-vlpfc: p=0.007; anterior-vlpfc: p=0.052, marginal). no any conditions achieve significant differences (all ps>0.05) between ad group and nc group. pearson correlation was adopted to measure the relation between time course of disease and alff for each of the rois within the right vlpfc (fig. 1). as shown in figure 3, a significant correlation between course of disease and the alff of the subregions in the right vlpfc were identified (negative correlation: regions including a, b, and d in slow5, and a in slow4; positive correlation: d region in slow4). apart from that, no other significant relationship was found. table 1. regions showing alff differences among the ad and the nc with mni coordinates. regions group mni coordinates brodmann area x y z frontal_lnf_oper_r (a) ad 57.7 16.1 12.2 44 frontal_lnf_tri_r (b) ad 56.41 27.27 8.891 45 frontal_lnf_orb_r (c) ad 52.54 37.64 -3.39 45 frontal_lnf_oper_r (d) ad 44.65 12.32 38.79 44 frontal_lnf_oper_r (a) nc 57.9 16.38 14.23 44 frontal_lnf_tri_r (b) nc 55.14 30.13 10.47 45 frontal_lnf_orb_r (c) nc 49.48 38.26 -9.554 44 frontal_lnf_oper_r (d) nc 46.95 20.18 31.16 44 ps. frontal_inf_oper_r: right inferior frontal operculum; frontal_inf_tri_r: right inferior frontal operculum; frontal_inf_orb_r: right inferior frontal operculum. a: posterior-vlpfc; b: mid-vlpfc; c: anterior-vlpfc; d: ifj. figure 2. the alff values of ad and nc groups in rois. rois: regions of interesting; ad: alzheimer’s disease; nc: normal control; vlpfc: ventral lateral prefrontal cortex; a: posterior-vlpfc; b: middle-vlpfc; c: anteriorvlpfc; and d: ifj, inferior frontal junction. in the present study the ad related changes in the intrinsic brain activity was correlated to the individual alff values of rs-fmri signals in the functional subregions of the right vlpfc (i.e., posteriory. tian et al. admet & dmpk 3(3) (2015) 274-280 278 vlpfc, mid-vlpfc, anterior-vlpfc, and ifj) and then the associated course of disease. we found that there were differences in alff values between the ad and the nc group throughout the subregions of the right vlpfc. decreased alff values were observed in ad patients in the ifj (in both slow5 and slow4), and the mid-vlpfc (in slow5). in contrast, increased alff valued were observed in ad patients in the posterior and anterior-vlpfc (in both slow5 and slow4), and also in the mid-vlpfc in slow4. moreover, significant alff differences between the ifj and three other subregions of the right vlpfc were found. further analysis revealed that the alff value for the right vlpfc can be applied to differentiate ad patients from normal controls with a high sensitivity. we also found that alff values in the right vlpfc showed significant correlations with the time course of disease. figure 3. relationships between the alff values and the course of disease (month) in rois. rois: regions of interesting; ad: alzheimer’s disease; nc: normal control; vlpfc: ventral lateral prefrontal cortex; a: posteriorvlpfc; b: middle-vlpfc; c: anterior-vlpfc; and d: ifj, inferior frontal junction. * :significant level. the broader evidence indicated that specific forms of cognitive control are linked with distinct functional subregions of the right vlpfc [16-17]. the ifj supports the detection of behavioral relevant stimuli. the top-down modulation mediated by the ifj plays a causal role between early attentional processes and subsequent memory performance [16-17, 24-25]. in our current study, decreased alff in several subregions in the right vlpfc including ifj (in both slow5 and slow4) was in accordance with previous studies that this region had structural and functional abnormalities. furthermore, we noticed that there were significant negative correlations between the alff of the ifj in slow5 and the course of disease, indicating that the decline alff for this region along with the course of disease were associated with memory-related cognitive impairment in these patients. in contrast, the significant positive correlation between the alff of the ifj in slow4 and the course of disease, suggesting the existence of compensation process for cognitive resources [12]. other subregions in right vlpfc are related to motor control, i.e. posterior-vlpfc is activated by the active plan updating and mid-vlpfc is responsible for decision uncertainty. anterior-vlpfc encoding higher-order goals usually bridge the longer time spans, and the midvlpfc (in slow5) [16]. in our study, the significant negative correlations between the alff of the posteriorvlpfc and the course of disease, suggest that the longer time after the onset of disease accompanied with the weaker motor control. admet & dmpk 3(3) (2015) 274-280 alffs of vlpfc predict ad doi: 10.5599/admet.3.3.208 279 conclusions in summary, we found that ad patients had abnormal alff in the subregions of right vlpfc, including ifj, posterior-vlpfc, mid-vlpfc, and anterior-vlpfc. these findings were similar with previous studies in ad. our findings indicated that the abnormal intrinsic neural oscillations in ad present different neural patterns in different frequency bands, predicting the memory related impairment with the course of disease. further investigations are important to reveal how they relate to the clinical and pathological correlations of alff changes in ad. acknowledgements: this study is supported by the national nature science foundation of china (#31100745), the cqcbe project kj110502 and the science and technology development fund, macao s.a.r (fdct) (project reference no.: 118/2013/a3). references [1] j. hardy, neuron 52 (2006) 3-13. [2] z. s. khachaturian, arch. neurol. 42 (1985) 1097-1105. [3] a.l. bokde, p. pietrini, v. ibáñez, m.l. furey, g.e. alexander, n.r. graff-radford, s.i. rapoport, m.b. schapiro, b. horwitz, arch. neurol. 58 (2001) 480–486. [4] m.j. de leon, a. convit, o.t. wolf, c.y. tarshish, s. desanti, h. rusinek, w. tsui, e. kandil, a.j. scherer, a. roche, a. imossi, e. thorn, m. bobinski, c. caraos, p. lesbre, d. schlyer, j. poirier, b. reisberg, j. fowler, proc. natl. acad. sci. usa 98 (2001) 10966–10971. [5] k. herholz, e. salmon, d. perani, j-c. baron, v. holthoff, l. frölich, p. schönknecht, k. ito, r. mielke, e. kalbe, g. zündorf, x. delbeuck, o. pelati, d. anchisi, f. fazio, n. kerrouche, b. desgranges, f. eustache, b. beuthien-baumann, c. menzel, j. schröder, et al., neuroimage 17 (2002) 302–316. [6] e. salmon, f. collette, c. degueldre, c. lemaire, g. franck, hum. brain mapp. 10 (2000) 39–48. [7] b. biswal, f.z. yetkin, v.m. haughton, j.s. hyde, magn. reson. med. 34 (1995) 537–541. [8] p. fransson, hum. brain mapp. 26 (2005) 15–29. [9] v. kiviniemi, j. jauhiainen, o. tervonen, e. pääkkö, j. oikarinen, v. vainionpää, h. rantala, b. biswal, magn. reson. med. 44 (2000) 373–378. [10] z. wang, c. yan, c. zhao, z. qi, w. zhou, j. lu, y. he. k li, hum. brain mapp. 32 (2011) 1720-1740. [11] f. bai, neurosci. lett. 438 (2008) 111–115. [12] y. he, l. wang, y. zang, l. tian, x. zhang, k. li, t. jiang, neuroimage 35 (2007) 488–500. [13] y. han, j. wang, z. zhao, b. min, j. lu, k. li, y. he, j. jia, neuroimage 55 (2011) 287-295. [14] k. wang, m. liang, l. wang, l. tian, x. zhang, k. li, t. jiang, hum. brain mapp. 28 (2007) 967–978. [15] h.y. zhang, s.-j. wang, j. xing, b. liu, z.-l. ma, m. yang, z.-j. zhang, g.-j. teng, behav. brain res. 197 (2009) 103–108. [16] b.j. levy, a.d. wagner, annals of the new york academy of sciences 1224 (2011) 40-62. [17] y. tian, s. liang, d. yao, neurosci. bull. 30 (2014) 141-152. [18] y.tian, d. yao, biol. psychol. 79 (2008) 171–178. [19] y. tian, s. liang,z. yuan, s. chen, p. xu, d. yao, neuroreport 25 (2014) 843-847. [20] x.n. zuo, a. di martino, c. kelly, z.e. shehzad, d.g. gee, d.f. klein, f.x. castellanos, b.b. biswal, m.p. milham, neuroimage 49 (2010) 1432–1445. [21] a.t. baria, m.n. baliki, t. parrish, a.v. apkarian, j. neurosci. 31 (2011) 7910–7919. [22] g .buzsáki, a. draguhn, science 304 (2004) 1926–1929. y. tian et al. admet & dmpk 3(3) (2015) 274-280 280 [23] c. yan, y. zang, front. syst. neurosci. 4 (2010) 13. [24] a. gazzaley, a.c. nobre, trends cogn. sci. 16 (2012) 129-135. [25] m. d’esposito, b. r. postle, annu. rev. psychol. 66 (2015) 115. ©2015 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ recent advances in nanoparticles as antibacterial agent doi: http://dx.doi.org/10.5599/admet.1172 115 admet & dmpk 10(2) (2022) 115-129; doi: https://doi.org/10.5599/admet.1172 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index review recent advances in nanoparticles as antibacterial agent murat ozdal*and sumeyra gurkok department of biology, science faculty, ataturk university, 25240 erzurum, turkey *corresponding author: e-mail: murat.ozdal@yahoo.com; tel.: +90-442-231-1648 received: november 14, 2021; revised: january 24, 2022; published: february 02, 2022 abstract recently, the rapid increase in antibiotic-resistant pathogens has caused serious health problems. researchers are searching for alternative antimicrobial substances to control or prevent infections caused by pathogens. different strategies are used to develop effective antibacterial agents, and in this respect, nanoparticles are undoubtedly promising materials. nanoparticles act by bypassing drug resistance mechanisms in bacteria and inhibiting biofilm formation or other important processes related to their virulence potential. nanoparticles can penetrate the cell wall and membrane of bacteria and act by disrupting important molecular mechanisms. in combination with appropriate antibiotics, nps may show synergy and help prevent the developing global bacterial resistance crisis. furthermore, due to characteristics such as enhanced biocompatibility and biodegradability, polymer-based nanoparticles enable the development of a wide range of medical products. antibacterial applications of nanoparticles range from antimicrobial synthetic textiles to biomedical and surgical devices when nanoparticles are embedded/loaded/coated into different materials. in this review, the antibacterial mechanisms of nanoparticles and their potential for use in the medical field are discussed. ©2021 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords nanoparticles; antimicrobials; biofilm; nanomedicine; antibiotic resistance introduction nanoparticles (nps) are particles in the 1-100 nm scale range. nps have certain physical and chemical properties that differ significantly from bulk materials. although nps can be synthesized by different physical and chemical methods, biological synthesis is an environmentally friendly green chemistry approach that is non-toxic, biocompatible and inexpensive [1,2]. antibiotics are very common and effective weapons in fighting infectious diseases. however, unconscious, excessive and inappropriate use of antibiotics has led to antimicrobial resistance, which is one of the serious threats affecting public health [3]. it is estimated that antimicrobial resistance may be responsible for approximately 10 million deaths per year by 2050 and will outnumber cancer deaths [4]. this situation has forced the search for new alternatives against bacterial infections. due to high cost and various difficulties related to exploration and design, an insufficient number of novel antibiotics have been discovered in the last quarter-century [5]. with the help of nanotechnological developments, nps that can be designed with desired properties have started to appear as promising tools. in particular, metal nps http://dx.doi.org/10.5599/admet.1172 https://doi.org/10.5599/admet.1172 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:murat.ozdal@yahoo.com http://creativecommons.org/licenses/by/4.0/ ozdal and gurkok admet & dmpk 10(2) (2022) 115-129 116 have attracted great interest as antimicrobial agents with their unique properties. various metal nps including silver (ag), copper (cu), selenium (se), nickel (ni), gold (au), zinc oxide (zno), titanium dioxide (tio2), and iron oxide (fe3o4), have been extensively studied for their antimicrobial effects [6]. nanotechnology provides a new possibility for biological applications by changing the physico-chemical properties of substances. in this regard, numerous nps have been discovered in recent years to be effective against various pathogens, including antibiotic-resistant microorganisms. the size, surface area, morphology, net charge and physicochemical properties of nps are among the important parameters that change their antimicrobial properties through multiple mechanisms [7]. as the size of the nps decreases, the surface-to-volume ratio increases significantly. large surface areas of nps provide better interaction with microorganisms and greatly influence their antimicrobial effects. among metal nps, positively charged ones have been reported to bind more tightly to negatively charged bacterial surfaces and show higher antimicrobial effects. the spherical nps have been shown to have a better antimicrobial effect as they allow more ions to be released due to their larger surface area [8]. slomberg et al. [9] argued that rod-like particles are more effective than spherical particles in delivering no and inducing greater antibacterial effects across the biofilm [9]. in another study, compared with spherical or rod-shaped agnps, truncated triangular-shaped agnps with the same surface areas have been reported to have higher antibacterial activity against escherichia coli [10]. furthermore, the antibacterial action of nps is affected by the type of capping agent used, as well as the ph and ionic strength of the medium. capping agents adsorbed on the surface of nps are extremely critical as stabilizers that prevent aggregation and overgrowth of nps. these agents affect the biological activities as well as certain structural and physico-chemical properties of nps. for example, agnps capped with solanum trilobatum extract, a stabilizing and reducing agent, have been reported to have enhanced antimicrobial properties on gram-negative and gram-positive bacterial species and fungal species [11]. bacteria use many different mechanisms to develop resistance to natural and synthetic antibiotics, rendering them ineffective and thus necessitating the development of new alternatives. however, due to the cost and complexity of developing a new antibiotic, very few new antibiotics have been commercialized in recent years. for this purpose, more attention should be paid to the development of new antimicrobial agents, which are difficult for bacteria to develop resistance against. the aim of this review is to explain the antimicrobial mechanisms of nps and the latest developments in antimicrobial applications. mechanisms of antibacterial properties of nps the mechanism of the antimicrobial properties of nps is still not fully elucidated. some of the prominent mechanisms that cause cell death are that the particles (i) bind to the bacterial cell, impair cell membrane permeability and respiration, (ii) cause toxicity by the release of free metal ions from the surface of nps, or (iii) cause oxidative stress by generating reactive oxygen species (ros). figure 1 summarizes the most proposed mechanisms. the mechanisms vary according to the np type. metal oxide nps including zno and tio2 are thought to exert bactericidal effects mostly by the generation of ros [12], whereas ag and au are thought to exert bactericidal effects predominantly by releasing metal ions [13]. the lethal impact of agnps has been linked to the direct interaction of nps with the bacterial cell wall, followed by penetration into the cytoplasm. nps accumulate on the bacterial cell wall and membrane, causing morphological changes such as shrinkage of the cytoplasm, detachment of membrane, formation of multiple electron-dense pits, and eventually membrane disruption [14]. np deposition in the cell wall of e. admet & dmpk 10(2) (2022) 115-129 antimicrobial nanoparticles against bacteria doi: http://dx.doi.org/10.5599/admet.1172 117 coli has been shown to create pits, which cause the release of lipopolysaccharide molecules and membrane proteins and thus the loss of outer membrane integrity, resulting in eventual cell death [14]. metals can be chemically oxidized in aqueous solutions to give metallic ions. due to their higher surfaceto-volume ratio, nps are able to release more ions under aerobic conditions than bulk materials. the released ions from the surface of nps play an important role in antibacterial activity. nps first anchor to the bacterial cell wall and, due to their nanoscale size, easily penetrate and pass through the cell wall and interact with the cell membrane. they cause structural changes in the cell membrane with the ions they release, thereby disrupting its integrity and increasing its permeability. as a result, they cause leakage of cell contents and eventually cell death. nps targeting the cell membrane also impair membrane potential and proton motive force, block oxidative phosphorylation, and decrease intracellular atp levels. such an antibacterial effect of spherical agnps against e. coli was revealed by proteomic studies, which showed that exposure to agnps has led to the build-up of envelope protein precursors implying disruption of proton motive force [15]. once the free metal ions reach the inside of the cell, they also interact with carbonyl, amino, phosphate and sulfhydryl (thiol) groups of the cellular biomolecules, such as dna, proteins, and lipids [16]. it was shown that free ag ions interact with bacterial nucleic acids preferentially via nucleosides rather than phosphate groups and degrade chromosomal dna or interrupt the dna replication [17]. metal ions, such as ag + , also interact with thiol groups of enzymes and proteins, altering their three-dimensional structure and blocking the active binding sites for their substrates [16]. interference with the proteins on the cell wall and membrane disrupts the bacterial cell wall, impairing the electron transport chain inhibiting the respiratory process, and growth of the cells [18]. they interact with cytoplasmic proteins required for atp production, inactivating them and impairing cellular functions. ions also prevent protein synthesis by denaturing ribosomal components and interfering with the binding of ribosome subunit to trna [19]. cui et al. [13] has shown by transcriptomic and proteomic approaches that aunps show antibacterial properties against e. coli by collapsing membrane potential and reducing atp levels via inhibiting atpase function and inhibiting the binding of ribosome subunit to trna. signal transduction in bacteria is another target of the antibacterial activity of nps. bacterial communities in a wide variety of environments are sensitive and respond to a variety of external stimuli in order to survive through their signal transduction mechanism. this mechanism, which is known to be necessary for the viability and growth of bacteria as well as the expression of virulence factors, is widely used as a target in antimicrobial drug design [20]. in this cell signaling system, a physical or chemical signal is transmitted through a cell as a cascade of molecular pathways, frequently protein phosphorylation mediated by protein kinases, which finally results in a cellular response. agnps 10–15 nm in size were suggested to interrupt the signal transduction by altering the phosphorylation pattern of tyrosine residues of the key proteins, which leads to apoptosis or growth inhibition in bacteria [21]. another pivotal mechanism of the antibacterial effect of the nps is the burst of ros in the bacterial cell. ros are generated during the regular oxygen metabolism and are crucial for different cellular signaling pathways. oxygen acts as the final acceptor of electrons transported by ets during oxidative phosphorylation and is reduced to the water molecule. some of these electrons are taken by molecular oxygen, resulting in the formation of o 2, that can then be transformed into h2o2 and •oh. however, when bacterial cells are exposed to nps, metal ions released from the surface of nps induce ros bursts by disrupting respiratory systems and can considerably increase intracellular ros production. released metal ions contribute to further increase in intracellular ros accumulation by causing disruption of membrane http://dx.doi.org/10.5599/admet.1172 ozdal and gurkok admet & dmpk 10(2) (2022) 115-129 118 integrity, inactivation of cellular enzymes, disruption of the electron transport system, and decreased membrane potentials. bacteria have natural antioxidant defense systems to deal with oxidative stress. they have natural antioxidants like carotenes and ascorbic acid, which prevent lipid peroxidation or other rosrelated stresses. in addition, they have enzymes such as catalase, peroxidase, and superoxide dismutase that convert toxic reactive oxygen forms into non-toxic or less toxic forms. however, upon exposure to nps, the accumulation of ros exceeds a certain level and bacteria cannot cope with detrimental changes in vital cellular structures such as cell wall, cell membrane, dna, and protein [7]. as a result, chemically highly reactive ros accumulation and ros-induced oxidative stress in the bacterial cell cause induced pore formation and lipid peroxidation in the cell membrane, damage to chromosomal dna and proteins, and ultimately cell death. tio2 and zno nps show their antibacterial activities through oxidative stress. they kill microorganisms with their potent oxidizing capacity through the formation of free radicals. they are able to induce oxidative stress and dna damage causing reduced viability of e. coli [12,22]. due to the differences in bacterial cell wall structures, the effects of nps on gram-positive and gramnegative bacterial species also change. in many studies, it has been reported that the antimicrobial effects of nps on gram-negative bacteria strains are stronger than gram-positive bacteria. gram-positive bacteria have a thick peptidoglycan layer consisting of many more layers, while gram-negative bacteria have a thin peptidoglycan layer and additional lipopolysaccharides (lps). manzoor et al. [23] found that the effect of nps was significantly more pronounced on gram-negative strains than on gram-positive organisms. there are also studies that suggest the opposite. premanathan et al. [24] investigated the antibacterial activity of zno-nps against gram-negative e. coli and pseudomonas aeruginosa and gram-positive staphylococcus aureus and reported that the antimicrobial effect was stronger in gram-positive bacterium. figure 1. antimicrobial mechanisms of nps synergistic effects of nps with antibiotics nps can be combined with antimicrobial agents to overcome antibiotic resistance and increase their effectiveness. in addition, they can reduce the dose and toxicity of antibiotics to be taken [5]. since nps act on bacteria through multiple targets and/or mechanisms, it is very difficult for the microorganism to admet & dmpk 10(2) (2022) 115-129 antimicrobial nanoparticles against bacteria doi: http://dx.doi.org/10.5599/admet.1172 119 acquire resistance. in other words, the probability of simultaneous mutations necessary for resistance formation is very low. moreover, this is even more unlikely when nps are combined with antimicrobials [25]. as a result, the use of nps with antibiotics is considered as a method that can be used to prevent bacterial resistance development [26]. antibiotics combined with nps are more effective against both gram-positive and gram-negative bacteria and even drug-resistant bacteria. aabed and mohammed [27] showed synergistic effects of agnps combined with bacitracin, ciprofloxacin, tetracycline, and cefixime against p. aeruginosa, e. coli, s. aureus, and candida albicans. abo-shama et al. [28] also demonstrated the synergistic effect of antibiotics (azithromycin, cefotaxime, cefuroxime, fosfomycin, and chloramphenicol) against e. coli was notably increased in the presence of agnps compared to the antibiotic used alone. in another study, a synergistic effect was observed when cuo nps combined with cephalexin against e. coli [29]. anti-quorum sensing and anti-biofilm activity of nps obstruction of quorum sensing (qs) is an influential alternative strategy to struggle with microbial infections. qs is a cell density-dependent regulatory process that regulates the expression of virulence factors in many pathogenic bacteria and relies on the release of extracellular small signal molecules known as autoinducers [30,31]. the qs system created by bacteria is known to be responsible for cell-to-cell communication and the formation of various virulence factors and biofilm formation [32]. several quorum quenching (qq) strategies have been suggested to disrupt communication between cells by degradation of a signaling molecule, inhibition of signaling molecule-receptor complex formation, inhibition of signaling molecule synthesis, inhibition of expression of qs regulated genes [33]. nps interfere with virulence factors (pigments, enzymes, exopolysaccharides, and toxins) in bacteria with their qq properties. garcía-lara et al. [34] showed the inhibitory effects of zno nps on qs by inhibiting the pyocyanin pigment, elastase and biofilm formation in p. aeruginosa. agnps have also been shown to inhibit the synthesis of p. aeruginosa virulence factors such as proteases, elastases, and pyocyanin [35,36]. in another study, qais et al. [31] reported that agnps successfully inhibited the qs regulated virulence agents of multiple bacterial pathogens, including p. aeruginosa, s. marcescens, and c. violaceum. it has also been shown that nps effectively prevent the formation of biofilms and destroy existing ones. biofilms, defined as well-organized multicellular microbial communities, are one of the survival strategies of microorganisms in nature. free-floating microorganisms anchor to a surface, begin to grow, and colonize the surface. they are organized in dense clusters in the extracellular polymeric substances (eps) they have formed [37]. this heterogeneous architecture is one of the main mechanisms that confer resistance to microorganisms against antimicrobial compounds. when bacteria form a biofilm, they become up to 1000 times more resistant to antibiotics than planktonic ones [38]. biofilm causes undesirable conditions such as treatment failure, the persistence of infection and the development of antibiotic resistance in the clinic [39]. at least 65 % of infectious diseases (such as otitis media, lung infections, mastitis, chronic wounds, urinary tract infections, rhinosinusitis, and gingivitis) are caused by biofilm [40]. among the antibiofilm strategies (enzymes, nps, phages, qs inhibitors, surfactants, phyto-compounds, and antimicrobial photodynamic therapy), nanostructured materials have gained importance in recent years [41]. ag + is known to have potent antibiofilm activity and is used to prevent biofilm development on medical surfaces, including heart valves, central venous catheters, orthopedic and dental implants [35,42]. however, it is easy to sequester them with chloride and phosphate or other cellular components. nps, on the other hand, are less exposed to sequestration due to their nano size and show a more permanent and http://dx.doi.org/10.5599/admet.1172 ozdal and gurkok admet & dmpk 10(2) (2022) 115-129 120 effective antibiofilm effect. the high surface area to volume ratio, inert structure, customizable physical features such as size and shape, biocompatibility, bacteriostatic or bactericidal capabilities at extremely low concentrations are all advantages of nps for antibiofilm applications [43]. another advantage is that nps are much smaller than 350 nm and can pass through biofilms. the interactions of nps with bacteria and biofilms depend on their surface charge. generally, positively charged nps better penetrate biofilms [44]. this can be explained by the higher interaction of positively charged nps with negatively charged biofilm structures (polysaccharide skeleton, proteins and dna) and bacterial cell wall. also, pointed and sharp triangular nps have more vertices and ends than spherical and rod-shaped nps, and therefore cause more damage to bacterial cells [45]. different metallic and non-metallic nps have been widely studied for their antibiofilm properties. the multifunctional properties of nps make it an attractive way to enhance the effect of antimicrobial agents to control infections. many types of nps such as nio [32], ag [46], au [47], se, te [48], si [49] and agcl-tio2 [50] block the mechanism of qs and thus hinder the formation of biofilm. zno and tio2 nps inhibit methicillin-resistant s. aureus biofilm formation [51]. agnps synthesized by solibacillus isronensis was shown to exert antibiofilm activity against e. coli and p. aeruginosa [52]. tio2 nanocomposites are frequently used for surface, dental and orthopedic implant coatings [53]. the combination of agnps and curcumin was also found to be effective in inhibiting biofilm formation and destroying established mature biofilm [54]. araujo et al. (2020) [55] studied the antimicrobial and antibiofilm effects of colloidal nanocarrier feo nps coated with chitosan containing chlorhexidine on c. glabrata and enterococcus faecalis associated with oral diseases. it was reported that aunps coated with the antimicrobial peptide indolicidin inhibited biofilm formation [56]. chitosan-coated feo nps were reported to inhibit biofilm formation in s. aureus [57]. when a cationic antimicrobial peptide polymyxin b was conjugated on the agnps surface, polymyxin b capped agnps showed a 3-fold higher biofilm reduction than pure agnps [58]. slomberg et al. [9] investigated the effectiveness of nitric oxide (no) releasing silica nps against gramnegative p. aeruginosa and gram-positive s. aureus biofilms, taking into account particle size and shape. they showed that particles with reduced size and increased aspect ratio released better no and were more effective against p. aeruginosa and s. aureus biofilms, and gram-negative strains were more sensitive to no. they also showed that rod-like particles were more effective than spherical particles in delivering no and inducing greater antibacterial action across the biofilm. biomedical applications of nps contributing to many areas of our lives, nanotechnology has also proven to have an important potential in the field of biomedicine, and as summarized in table 1, it has recently been used more widely. in order to prevent bacterial adhesion and biofilm formation, nano-engineering materials are developed by making surface modifications on implants and medical devices. medical implants and devices as next-generation nanomaterials are non-expensive and biocompatible. antibacterial film-based composite materials enable different uses, such as implant or catheter coatings and wound dressings. when designing composite films, natural polymers such as cellulose and collagen are more preferred due to their biocompatibility. on the other hand, synthetic polymers including polyethylene, polycaprolactone, and polyurethane are used in the production of antibacterial composite films due to their durability and ease of processing [59]. there are many studies on np-containing materials developed with antibacterial properties. the association of agnps with polyethylene typically reduces the wear of the polymer surface and increases the antibacterial properties of the polymer [60]. nanocomposites consisting of a mixture of agnps and admet & dmpk 10(2) (2022) 115-129 antimicrobial nanoparticles against bacteria doi: http://dx.doi.org/10.5599/admet.1172 121 polyethylene can be used as agents with antimicrobial and antibiofilm inhibitory potential in the food and health fields [61]. kim et al. [62] produced injectable agnp/methylcellulose nanocomposite hydrogel for topical antimicrobial applications that can be used on burn wounds. it has been reported that a variety of antibacterial nanocomposites such as β-chitin/zno nps [63], poly(vinyl alcohol)/zno [64], and collagendextran-zno-nps [65] can be used to heal infected wounds. kharaghani et al. [66] developed antibacterial contact lenses containing polyvinyl alcohol (pva) agnps and cunps. table 1. different nanomaterials used in biomedical fields with antibacterial properties biomaterials potential applications bacteria reference cotton/silk fabrics containing reduced graphene oxide (rgo) and ag/cu nps antimicrobial protective medical textiles p. aeruginosa e. coli s. aureus [67] polyvinyl alcohol containing ag/cu nps antibacterial contact lenses s. aureus p. aeruginosa [66] lysozyme-coated aunps in combination with the β-lactam diabetic wound healing s. aureus acinetobacter calcoaceticus p. aeruginosa e. coli klebsiella pneumoniae bacillus subtilis, b. cereus [68] keratin containing agnps skin wound healing and tissue recovery e. coli s. aureus [69] agnps-loaded bacterial cellulose hydrogels moist wound-healing hydrogels s. aureus p. aeruginosa [70] colloidal nps (zno, cuo, tio2 ag) antiseptic mouthwashes s. mutans streptococcus sangius [71] dextran/ceo2 nps against implant infections p. aeruginosa s. epidermidis [72] collagen conjugated with agnp repairing of infected bone vancomycin-resistant s. aureus [73] agnps with alginate-nano hydroxyapatite potential candidate for bone tissue repair and regeneration s. aureus [74] ag–zno@ carboxymethyl cellulose/k-carrageenan/graphene oxide/konjac glucomannan hydrogel nursing care for diabetic foot ulcers s. aureus e. coli [75] collagen chondroitin sulfate fibronectin agnps oral cavity lesions repair fusobacterium nucleatum porphyromonas gingivalis [76] pyrolytic carbon coated with agnp antibacterial artificial heart valve mrsa s. pyogenes e. coli k. pneumoniae p. aeruginosa, proteus vulgaris [77] face mask coated with colloidal agnps antimicrobial face masks s. aureus e. coli [78] nanoparticles as antimicrobial agent in np-drug conjugate system nps can be engineered and combined with other antimicrobial agents and they gain greater functionality in combating resistant microorganisms. due to the chemical properties of nanoparticles, they allow long-term binding to the target site of antibiotics and protection from enzymes. therefore, higher antibiotic requirements are avoided. it is important to create antibiotic nanoparticle conjugates to prevent multidrug-resistant pathogenic microbial infections. the preparation of conjugated nps is based on physical (hydrophobic, host-guest, and electrostatic http://dx.doi.org/10.5599/admet.1172 https://www.sciencedirect.com/topics/physics-and-astronomy/staphylococcus ozdal and gurkok admet & dmpk 10(2) (2022) 115-129 122 interactions) and chemical interactions (with amine, trans-cyclooctene, hydrazide, isothiocyanate, sulfhydryl, azide groups of drug) [79-81]. related possible pathways of some antibacterial nanoparticledrug conjugate formations are presented below.  the nanoparticle antibiotic combination provides great benefits in the solubility of poorly soluble drugs, drug half-life, systemic circulation, and drug release.  the negative charge of peptidoglycan layer, lipopolysaccharides, and teichoic acid promotes adhesion of nps and makes bacteria more sensitive to antimicrobial therapy.  hydrogenation of nps increases the stability of nps and impairs their function by binding to the negatively charged surface of the bacterial cell.  by binding to the proteins in the bacterial cell membrane, the nps increase the permeability and more antibiotics are passed into the bacterial cell. the active surface of nps causes membrane damage, disrupts protein−protein interactions and metabolic disorders in cells.  nps interact with sulfhydryl (-sh) groups in the cell wall to form r-s-s-r bonds and inhibit respiration resulting in cell death. when nps enter bacteria, they can affect cell membrane functions (permeability, respiration). since multiple simultaneous mutations are required in the same microorganism, antimicrobial resistance is unlikely to develop if antibiotics are combined with nps. there are still many unexplored conjugates. therefore, new antimicrobials are waiting to be discovered. pharmacology and toxicity of nanoparticles the pharmacokinetics of nps depends on various factors such as physicochemical properties (morphology, composition, size, charge, etc.), route of exposure (topical, intramuscular, intradermal, parenteral and subcutaneous), dosage, and animal species. absorption of aunp, agnp or tio2 nps is usually low by the oral, dermal or pulmonary route. inhalational absorption of aunps ranges from 0.06% to 5.5% and oral absorption is about 0.01–5% for aunps, 1–4.2% for agnps, and 0.01–0.05% for tio2 nps depending on the size [82-83]. excessive accumulation of nps in the target tissue is generally desired for therapeutic effects, and conversely, high levels of distribution or accumulation to non-target tissues can result in undesirable toxicity. further research is needed to understand better the pathway and dose differences in the pharmacokinetics of nps. contrary to the positive aspects of nanoparticles, they can cause adverse effects on living things. nanoparticles have found use in many industrial areas such as food, cosmetics, medicine, textiles, and automotive with their extraordinary physicochemical properties. in parallel with their increasing use in daily life, humans are constantly exposed to nanoparticles by inhalation, oral, dermal contact, and intravenous injection routes and nps can pose a potential threat in long-term exposures. the interaction of nps with body tissues and thus their toxic effect has not yet been fully elucidated. it is predicted that nps may have varying effects according to their type and physicochemical properties, exposure route, dose and time, and type of the cell line [84]. the toxicity of ag nps has been extensively studied in vitro and ag nps have been demonstrated to be more harmful to cell lines than other metal nps [85]. it has been observed that as the size of the np increase, the adverse health effects also increase [86]. nps injected intravenously can deposit in the colon, liver, spleen, and lymphatic system [87]. inhalation of nps might cause cytotoxicity in the lung [88]. once in the body, nps can easily enter bloodstream due to their small size and be transported to other parts of the body, such as the lungs, liver, kidneys, and reproductive organs. here, they can accumulate and admet & dmpk 10(2) (2022) 115-129 antimicrobial nanoparticles against bacteria doi: http://dx.doi.org/10.5599/admet.1172 123 interact with tissues leading cytotoxicity and dysfunction of the organs. moreover, due to their very small size, nps can even cross the blood-brain barrier and cause neurotoxicity [89]. another concern of using the antimicrobial properties of nps is the possibility that they can cause the death of beneficial human microflora. nps are known to be cytotoxic, carcinogenic, genotoxic, apoptosis inducer, and cell proliferation inhibitor [90]. negative effects caused by nps in living things usually occur by the destruction of cell membranes and organelles [91] or by binding to biomacromolecules and changing their structures and functions [92]. the toxicity of nps has been recognized by numerous in vivo studies. repeated 28‐day oral exposure of albino wistar rats to magnesium oxide (mgo) nps caused damage to dna, chromosomes, proteins and enzymes, redox balance and increased hepatic enzyme concentration in blood [93]. it has been reported in a study with rats that oxidative stress caused by cuo np interacts with cell components and induces hepatotoxicity and nephrotoxicity [94]. ag-nps and tio2-nps affect the central nervous system and cause neuroinflammation by inducing glial cell activation to release proinflammatory cytokines and produce ros and nitric oxide [89]. cytotoxic effects of agnps on osteoclasts and osteoblasts, the cardiovascular and respiratory systems, dna, and embryo development abnormalities, have been studied. nps can also cause hemolysis and disrupt the blood coagulating mechanisms [68]. since the side effects of nps are not fully understood, their use in clinical applications is limited and more studies are needed to benefit from nps more effectively. clarifying the toxicity of nps with detailed in vivo and clinical studies will pave the way for the routine use of nps in combating infections caused by multi-drug resistant bacteria. to reduce the harmful effects of nps on organisms, there are different approaches, such as the use of antioxidants. the beneficial effects of using antioxidant substances have been reported, especially in toxicities associated with np-induced oxidative stress formation [95]. in addition, proper adjustment of the threshold dose and exposure time that inhibit cell viability is vital for the safe delivery of nps. conclusions nanotechnology offers alternative sources to antibiotics. in order to struggle the infections by bacteria, nanoparticles exhibit multiple features such as inhibition of biofilms and/or increased intracellular accumulation of nps. the nanotechnological intervention provides new possibilities for the development of new therapeutic drug candidates to control the qs-regulated virulence profile, biofilm formation, and drug resistance profile. when nanoparticles are used together with appropriate antibiotics against pathogens, they reduce the amount of antibiotics to be applied, minimizing both the possibility of resistance development and toxicity. this synergistic effect of nanoparticles with antibiotics can be used against pathogenic bacteria in the near future. recent advances in nanotechnology have enabled the synthesis of new nanomaterials with multifunctional antimicrobial properties. when nanoparticles are embedded/loaded/coated into/on different materials, they can be used in a variety of applications, from antimicrobial synthetic textiles to biomedical and surgical equipment. biopolymer-based nanomaterials allow the development of many medicinal products due to their advantages, such as increased biocompatibility and biodegradability. in the light of increasing research in the field of nanomedicine, new uses of antibacterial nanoparticles will be revealed. conflict of interest: the authors declare no conflict of interest. http://dx.doi.org/10.5599/admet.1172 ozdal and gurkok admet & dmpk 10(2) (2022) 115-129 124 references [1] t. çakıcı, m. özdal, m. kundakcı, r. kayalı. znse and cuse np’s by microbial green synthesis method and comparison of i-v characteristics of au/znse/p-si/al and au/cuse/p-si/al structures. mater. sci. semicond. process. 103 (2019) 104610. https://doi.org/10.1016/j.mssp.2019.104610. [2] s. gurkok, m. ozdal. production of zinc selenide nanoparticles (znse-nps) via microbial green synthesis and their antibacterial activities, in world congress on applied nanotechnology, 2021, pp. 102–105. [3] m. makowski, í.c. silva, c.p. do amaral, s. gonçalves, n.c. santos. advances in lipid and metal nanoparticles for antimicrobial peptide delivery. pharmaceutics 11 (2019) 11. https://doi.org/10.3390/pharmaceutics11110588. [4] m. gautham, n. spicer, s. chatterjee, c. goodman. what are the challenges for antibiotic stewardship at the community level? an analysis of the drivers of antibiotic provision by informal healthcare providers in rural india. soc. sci. med. 275 (2021) 113813. https://doi.org/10.1016/j.socscimed.2021.113813. [5] m. hutchings, a. truman, b. wilkinson. antibiotics: past, present and future. curr. opin. microbiol.. 51 (2019) 72–80. https://doi.org/10.1016/j.mib.2019.10.008. [6] h.a. hemeg. nanomaterials for alternative antibacterial therapy. int. j. nanomedicine. 12 (2017) 8211–8225. https://doi.org/10.2147/ijn.s132163. [7] p.v. baptista, m.p. mccusker, a. carvalho, d.a. ferreira, m. mohan, m. martins, a.r. fernandes. nano-strategies to fight multidrug resistant bacteria-“a battle of the titans”. front. microbiol. 9 (2018) 1–26. https://doi.org/10.3389/fmicb.2018.01441. [8] r. shanmuganathan, d. mubarakali, d. prabakar, h. muthukumar, n. thajuddin, s.s. kumar, a. pugazhendhi. an enhancement of antimicrobial efficacy of biogenic and ceftriaxone-conjugated silver nanoparticles: green approach. environ. sci. pollut. res. 25 (2018) 10362–10370. https://doi.org/10.1007/s11356-017-9367-9. [9] d.l. slomberg, y. lu, a.d. broadnax, r.a. hunter, a.w. carpenter, m.h. schoenfisch. role of size and shape on biofilm eradication for nitric oxide-releasing silica nanoparticles. acs appl. mater. interfaces. 5 (2013) 9322–9329. https://doi.org/10.1021/am402618w. [10] s. pal, y.k. tak, j. m. song. does the antibacterial activity of silver nanoparticles depend on the shape of the nanoparticle? a study of the gram-negative bacterium escherichia coli. appl. environ. microbiol. 73 (2007) 1712–1720. https://doi.org/10.1128/aem.02218-06. [11] s. ramanathan, s. c. b. gopinath, p. anbu, t. lakshmipriya, f. h. kasim, c. g. lee. eco-friendly synthesis of solanum trilobatum extract-capped silver nanoparticles is compatible with good antimicrobial activities. j. mol. struct. 1160 (2018) 80–91. https://doi.org/10.1016/j.molstruc.2018.01.056. [12] y.h. leung, x. xu, a.p.y. ma, f. liu, a.m.c. ng, z. shen, l.a. gethings, m.y. guo, a.b. djurišić, p.k.h. lee, h.k. lee, w.k. chan, f.c.c. leung . toxicity of zno and tio2 to escherichia coli cells. sci. rep. 6 (2016) 1–13. https://doi.org/10.1038/srep35243. [13] y. cui, y. zhao, y. tian, w. zhang, x. lü, x. jiang. the molecular mechanism of action of bactericidal gold nanoparticles on escherichia coli. biomaterials 33 (2012) 2327–2333. https://doi.org/10.1016/j.biomaterials.2011.11.057. [14] i. sondi, b. salopek-sondi. silver nanoparticles as antimicrobial agent: a case study on e. coli as a model for gram-negative bacteria. j. colloid interface sci. 275 (2004) 177–182. https://doi.org/10.1016/j.jcis.2004.02.012. [15] c.-n. lok, c.-m. ho, r. chen, q.-y. he, w.-y. yu, h. sun, p.k.-h. tam, j.-f. chiu, c.-m. che. proteomic analysis of the mode of antibacterial action of silver nanoparticles. j. proteome res. 5 (2006) 916– 924. https://doi.org/10.1021/pr0504079. [16] s. tang, j. zheng. antibacterial activity of silver nanoparticles: structural effects. adv. healthc. mater. 7 (2018) 1–10. https://doi.org/10.1002/adhm.201701503. https://doi.org/10.1016/j.mssp.2019.104610 https://doi.org/10.3390/pharmaceutics11110588 https://doi.org/10.3390/pharmaceutics11110588 https://doi.org/10.1016/j.socscimed.2021.113813 https://doi.org/10.1016/j.socscimed.2021.113813 https://doi.org/10.1016/j.mib.2019.10.008 https://doi.org/10.2147/ijn.s132163 https://doi.org/10.3389/fmicb.2018.01441 https://doi.org/10.1007/s11356-017-9367-9 https://doi.org/10.1007/s11356-017-9367-9 https://doi.org/10.1021/am402618w https://doi.org/10.1128/aem.02218-06 https://doi.org/10.1016/j.molstruc.2018.01.056 https://doi.org/10.1016/j.molstruc.2018.01.056 https://doi.org/10.1038/srep35243 https://doi.org/10.1016/j.biomaterials.2011.11.057 https://doi.org/10.1016/j.biomaterials.2011.11.057 https://doi.org/10.1016/j.jcis.2004.02.012 https://doi.org/10.1021/pr0504079 https://doi.org/10.1002/adhm.201701503 admet & dmpk 10(2) (2022) 115-129 antimicrobial nanoparticles against bacteria doi: http://dx.doi.org/10.5599/admet.1172 125 [17] y. yakabe, t. sano, h. ushio, t. yasunaga. kinetic studies of the interaction between silver ion and deoxyribonucleic acid. chem. lett. 9 (1980) 373-376. [18] a. ahmad, y. wei, f. syed, k. tahir, a. u. rehman, a. khan, s. ullah, q. yuan. the effects of bacteriananoparticles interface on the antibacterial activity of green synthesized silver nanoparticles. microb. pathog. 102 (2017) 133–142. https://doi.org/10.1016/j.micpath.2016.11.030. [19] v. pareek, r. gupta, j. panwar. do physico-chemical properties of silver nanoparticles decide their interaction with biological media and bactericidal action? a review. mater. sci. eng. c. 90 (2018) 739– 749. https://doi.org/10.1016/j.msec.2018.04.093. [20] s.tiwari, s. b. jamal, s. s. hassan, p. carvalho, s. almeida, d. barh, p. ghosh, a. silva, t. l. p. castro, v. azevedo. two-component signal transduction systems of pathogenic bacteria as targets for antimicrobial therapy: an overview. front. microbiol. 8 (2017) 1–7. https://doi.org/10.3389/fmicb.2017.01878. [21] s. shrivastava, t. bera, a. roy, g. singh, p. ramachandrarao, d. dash. characterization of enhanced antibacterial effects of novel silver nanoparticles. nanotechnology. 18 (2007) 22. https://doi.org/10.1088/0957-4484/18/22/225103. [22] a. kumar, a.k. pandey, s.s. singh, r. shanker, a. dhawan. engineered zno and tio2 nanoparticles induce oxidative stress and dna damage leading to reduced viability of escherichia coli. free radic. biol. med. 51 (2011) 1872–1881. https://doi.org/10.1016/j.freeradbiomed.2011.08.025. [23] s. manzoor, d. j. bashir, k. imtiyaz, m.m.a. rizvi, i. ahamad, t. fatma, n.b. agarwal, i. arora, m. samim. biofabricated platinum nanoparticles: therapeutic evaluation as a potential nanodrug against breast cancer cells and drug-resistant bacteria. rsc adv. 11 (2021) 24900–24916. https://doi.org/10.1039/d1ra03133c. [24] m. premanathan, k. karthikeyan, k. jeyasubramanian, g. manivannan. selective toxicity of zno nanoparticles toward gram-positive bacteria and cancer cells by apoptosis through lipid peroxidation. nanomedicine nanotechnology biol. med. 7 (2011) 184–192. https://doi.org/10.1016/j.nano.2010.10.001. [25] n.y. lee, w. c. ko, p.r. hsueh. nanoparticles in the treatment of infections caused by multidrugresistant organisms. front. pharmacol. 10 (2019) 1–10. https://doi.org/10.3389/fphar.2019.01153. [26] y. zhao, x. jiang. multiple strategies to activate gold nanoparticles as antibiotics. nanoscale 5 (2013) 8340–8350. https://doi.org/10.1039/c3nr01990j. [27] k. aabed, a. e. mohammed. synergistic and antagonistic effects of biogenic silver nanoparticles in combination with antibiotics against some pathogenic microbes. front. bioeng. biotechnol. 9 (2021) 1–14. https://doi.org/10.3389/fbioe.2021.652362. [28] u.h. abo-shama, h. el-gendy, w.s. mousa, r.a. hamouda, w.e. yousuf, h.f. hetta, e.e. abdeen. synergistic and antagonistic effects of metal nanoparticles in combination with antibiotics against some reference strains of pathogenic microorganisms. infect. drug resist. 13 (2020) 351–362. https://doi.org/10.2147/idr.s234425. [29] y. zhang, l. wang, x. xu, f. li, q. wu. combined systems of different antibiotics with nano-cuo against escherichia coli and the mechanisms involved. nanomedicine 13 (2018) 339–351. https://doi.org/10.2217/nnm-2017-0290. [30] f.a. qais, m.s. khan, i. ahmad. nanoparticles as quorum sensing inhibitor: prospects and limitations. v. c. kalia (ed.), biotechnological applications of quorum sensing inhibitors, springer singapore, singapore (2018) 227-244. https://doi.org/10.1007/978-981-10-9026-4_11. [31] f.a. qais, a. shafiq, i. ahmad, f.m. husain, r.a. khan, i. hassan. green synthesis of silver nanoparticles using carum copticum: assessment of its quorum sensing and biofilm inhibitory potential against gram negative bacterial pathogens. microb. pathog. 144 (2020) 104172. https://doi.org/10.1016/j.micpath.2020.104172. [32] m. maruthupandy, g.n. rajivgandhi, f. quero, w.j. li. anti-quorum sensing and anti-biofilm activity of nickel oxide nanoparticles against pseudomonas aeruginosa. j. environ. chem. eng. 8 (2020) 104533. https://doi.org/10.1016/j.jece.2020.104533. http://dx.doi.org/10.5599/admet.1172 https://doi.org/10.1016/j.micpath.2016.11.030 https://doi.org/10.1016/j.msec.2018.04.093 https://doi.org/10.3389/fmicb.2017.01878 https://doi.org/10.3389/fmicb.2017.01878 https://doi.org/10.1088/0957-4484/18/22/225103 https://doi.org/10.1088/0957-4484/18/22/225103 https://doi.org/10.1016/j.freeradbiomed.2011.08.025 https://doi.org/10.1039/d1ra03133c https://doi.org/10.1039/d1ra03133c https://doi.org/10.1016/j.nano.2010.10.001 https://doi.org/10.1016/j.nano.2010.10.001 https://doi.org/10.3389/fphar.2019.01153 https://doi.org/10.1039/c3nr01990j https://doi.org/10.3389/fbioe.2021.652362 https://doi.org/10.2147/idr.s234425 https://doi.org/10.2217/nnm-2017-0290 https://doi.org/10.1007/978-981-10-9026-4_11 https://doi.org/10.1016/j.micpath.2020.104172 https://doi.org/10.1016/j.jece.2020.104533 ozdal and gurkok admet & dmpk 10(2) (2022) 115-129 126 [33] s. hayat, s. muzammil, b. aslam, m. h. siddique, m. saqalein, m. a. nisar. quorum quenching: role of nanoparticles as signal jammers in gram-negative bacteria. future microbiol. 14 (2019) 61–72. https://doi.org/10.2217/fmb-2018-0257. [34] b. garcia-lara, m.a. saucedo-mora, j.a. roldan-sanchez, perez-b. eretza, m. ramasamy, j. lee, r. coria-jimenez, m. tapia, v. varela-guerrero, r. garcia-contreras. inhibition of quorum-sensingdependent virulence factors and biofilm formation of clinical and environmental pseudomonas aeruginosa strains by zno nanoparticles. lett. appl. microbiol. 61 (2015) 299–305. https://doi.org/10.1111/lam.12456. [35] a.n. coman, a. mare, c. tanase, e. bud, a. rusu. silver-deposited nanoparticles on the titanium nanotubes surface as a promising antibacterial material into implants. metals 11 (2021) 1–16. https://doi.org/10.3390/met11010092. [36] s.g. ali, m.a. ansari, h.m. khan, m. jalal, a.a. mahdi, s.s. cameotra. crataeva nurvala nanoparticles inhibit virulence factors and biofilm formation in clinical isolates of pseudomonas aeruginosa. j. basic microbiol. 57 (2017) 193–203. https://doi.org/10.1002/jobm.201600175. [37] m.a. rather, k. gupta, m. mandal. microbial biofilm: formation, architecture, antibiotic resistance, and control strategies. brazilian j. microbiol. 52 (2021) 1701–1718. https://doi.org/10.1007/s42770021-00624-x. [38] m.e. olson, h. ceri, d.w. morck, a.g. buret, r.r. read. biofilm bacteria: formation and comparative susceptibility to antibiotics. can. j. vet. res. 66 (2002) 86–92. pmid: 11989739. [39] y. zhao, l. chen, y. wang, x. song, k. li, x. yan, l. yu, z. he. nanomaterial-based strategies in antimicrobial applications: progress and perspectives. nano res. 14 (2021) 4417–4441. https://doi.org/10.1007/s12274-021-3417-4. [40] l.k. vestby, t. grønseth, r. simm, l.l. nesse. bacterial biofilm and its role in the pathogenesis of disease. antibiotics 9 (2020) 59. https://doi.org/10.3390/antibiotics9020059. [41] s. qayyum, a. u. khan. nanoparticles: vs. biofilms: a battle against another paradigm of antibiotic resistance. medchem comm. 7 (2016) 1479–1498. https://doi.org/10.1039/c6md00124f. [42] k. wu, y. yang, y. zhang, j. deng, c. lin. antimicrobial activity and cytocompatibility of silver nanoparticles coated catheters via a biomimetic surface functionalization strategy. int. j. nanomedicine 10 (2015) 7241–7252. https://doi.org/10.2147/ijn.s92307. [43] p. singh, s. pandit, c. jers, a. s. joshi, j. garnæs, i. mijakovic. silver nanoparticles produced from cedecea sp. exhibit antibiofilm activity and remarkable stability. sci. rep. 11 (2021) 1–13. https://doi.org/10.1038/s41598-021-92006-4. [44] j. wu, f. li, x. hu, j. lu, x. sun, j. gao, d. ling. responsive assembly of silver nanoclusters with a biofilm locally amplified bactericidal effect to enhance treatments against multi-drug-resistant bacterial infections. acs cent. sci. 5 (2019) 1366–1376. https://doi.org/10.1021/acscentsci.9b00359. [45] j.m.v. makabenta, a. nabawy, c.h. li, s. schmidt-malan, r. patel, v.m. rotello. nanomaterial-based therapeutics for antibiotic-resistant bacterial infections. nat. rev. microbiol. 19 (2021) 23–36. https://doi.org/10.1038/s41579-020-0420-1. [46] s. shah, s. gaikwad, s. nagar, s. kulshrestha, v. vaidya, n. nawani, s.pawar. biofilm inhibition and anti-quorum sensing activity of phytosynthesized silver nanoparticles against the nosocomial pathogen pseudomonas aeruginosa. biofouling 35 (2019) 34–49. https://doi.org/10.1080/08927014.2018.1563686. [47] s. samanta, b.r. singh, a. adholeya. intracellular synthesis of gold nanoparticles using an ectomycorrhizal strain em-1083 of laccaria fraterna and its nanoanti-quorum sensing potential against pseudomonas aeruginosa. indian j. microbiol. 57 (2017) 448–460. https://doi.org/10.1007/s12088-017-0662-4. [48] b. gómez-gómez, l. arregui, s. serrano, a. santos, t. pérez-corona, y. madrid. selenium and tellurium-based nanoparticles as interfering factors in quorum sensing-regulated processes: violacein production and bacterial biofilm formation. metallomics 11 (2019) 1104–1114. https://doi.org/10.1039/c9mt00044e. https://doi.org/10.2217/fmb-2018-0257 https://doi.org/10.1111/lam.12456 https://doi.org/10.3390/met11010092 https://doi.org/10.1002/jobm.201600175 https://doi.org/10.1007/s42770-021-00624-x https://doi.org/10.1007/s42770-021-00624-x https://doi.org/10.1007/s12274-021-3417-4 https://doi.org/10.3390/antibiotics9020059 https://doi.org/10.1039/c6md00124f https://doi.org/10.2147/ijn.s92307 https://doi.org/10.1038/s41598-021-92006-4 https://doi.org/10.1021/acscentsci.9b00359 https://doi.org/10.1038/s41579-020-0420-1 https://doi.org/10.1080/08927014.2018.1563686 https://doi.org/10.1080/08927014.2018.1563686 https://doi.org/10.1007/s12088-017-0662-4 https://doi.org/10.1007/s12088-017-0662-4 https://doi.org/10.1039/c9mt00044e https://doi.org/10.1039/c9mt00044e admet & dmpk 10(2) (2022) 115-129 antimicrobial nanoparticles against bacteria doi: http://dx.doi.org/10.5599/admet.1172 127 [49] k.p. miller, l. wang, y.p. chen, p.j. pellechia, b.c. benicewicz, a.w. decho. engineering nanoparticles to silence bacterial communication. front. microbiol. 6 (2015) 1–7. https://doi.org/10.3389/fmicb.2015.00189. [50] k. naik, m. kowshik. anti-quorum sensing activity of agcl-tio2 nanoparticles with potential use as active food packaging material. j. appl. microbiol. 117 (2014) 972–983. https://doi.org/10.1111/jam.12589. [51] a. jesline, n.p. john, p. m. narayanan, c. vani, s. murugan. antimicrobial activity of zinc and titanium dioxide nanoparticles against biofilm-producing methicillin-resistant staphylococcus aureus. appl. nanosci. 5 (2015) 157–162. https://doi.org/10.1007/s13204-014-0301-x. [52] p. singh, s. pandit, v.r.s.s. mokkapati, j. garnæs, i. mijaković. a sustainable approach for the green synthesis of silver nanoparticles from solibacillus isronensis sp.and their application in biofilm inhibition. molecules 25 (2020) 2783. https://doi.org/10.3390/molecules25122783. [53] v. kumaravel, k.m. nair, s. mathew, j. bartlett, j. e. kennedy, h.g. manning, b.j. whelan, n.s. leyland, s.c. pillai. antimicrobial tio2 nanocomposite coatings for surfaces, dental and orthopaedic implants. chem. eng. j. 416 (2021) 129071. https://doi.org/10.1016/j.cej.2021.129071. [54] c.y. loo, r. rohanizadeh, p.m. young, d. traini, r. cavaliere, c. whitchurch, w.-h. lee. combination of silver nanoparticles and curcumin nanoparticles for enhanced anti-biofilm activities. j. agric. food chem. 64 (2016) 2513–2522. https://doi.org/10.1021/acs.jafc.5b04559. [55] h.c. araujo, a.c.g. da silva, l.i. paião, m.k.w. magario, s.c.t. frasnelli, s.h.p. oliveira, j.p. pessan, d.r. monteiro. antimicrobial, antibiofilm and cytotoxic effects of a colloidal nanocarrier composed by chitosan-coated iron oxide nanoparticles loaded with chlorhexidine. j. dent. 101 (2020) 103453. https://doi.org/10.1016/j.jdent.2020.103453. [56] e. de alteriis, a. falanga, s. galdiero, m. guida, v. maselli, e. galdiero. genotoxicity of gold nanoparticles functionalized with indolicidin towards saccharomyces cerevisiae. j. environ. sci. 66 (2018) 138–145. https://doi.org/10.1016/j.jes.2017.04.034. [57] s. f. shi, j. f. jia, x. k. guo, y. p. zhao, d. s. chen, y. y. guo, x. l. zhang. reduced staphylococcus aureus biofilm formation in the presence of chitosan-coated iron oxide nanoparticles. int. j. nanomedicine 11 (2016) 6499–6506. https://doi.org/10.2147/ijn.s41371. [58] p.r. lambadi, t.k. sharma, p. kumar, p. vasnani, s. m. thalluri, n. bisht, r. pathania, n. k. navani. facile biofunctionalization of silver nanoparticles for enhanced antibacterial properties, endotoxin removal, and biofilm control. int. j. nanomedicine 10 (2015) 2155–2171. https://doi.org/10.2147/ijn.s72923. [59] v.a. spirescu, c. chircov, a.m. grumezescu, b. ștefan vasile, e. andronescu. inorganic nanoparticles and composite films for antimicrobial therapies. int. j. mol. sci. 22 (2021) 4595. https://doi.org/10.3390/ijms22094595. [60] k. morley, p. webb, n. tokareva, a. krasnov, v. popov, j. zhang, c. roberts, s. howdle. synthesis and characterisation of advanced uhmwpe/silver nanocomposites for biomedical applications. eur. polym. j. 43 (2007) 307–314. https://doi.org/10.1016/j.eurpolymj.2006.10.011. [61] d. olmos, g. m. pontes-quero, a. corral, g. gonzález-gaitano, j. gonzález-benito. preparation and characterization of antimicrobial films based on ldpe/ag nanoparticles with potential uses in food and health industries. nanomaterials 8 (2018) 60. https://doi.org/10.3390/nano8020060. [62] m. h. kim, h. park, h. c. nam, s. r. park, j. y. jung, w. h. park. injectable methylcellulose hydrogel containing silver oxide nanoparticles for burn wound healing. carbohydr. polym. 181 (2018) 579–586. https://doi.org/10.1016/j.carbpol.2017.11.109. [63] p.t. sudheesh kumar, v.k. lakshmanan, m. raj, r. biswas, t. hiroshi, s.v. nair, r. jayakumar. evaluation of wound healing potential of β-chitin hydrogel/nano zinc oxide composite bandage. pharm. res. 30 (2013) 523–537. https://doi.org/10.1007/s11095-012-0898-y. [64] a. chaturvedi, a.k. bajpai, j. bajpai, s.k. singh. evaluation of poly (vinyl alcohol) based cryogel-zinc oxide nanocomposites for possible applications as wound dressing materials. mater. sci. eng. c 5 (2016) 408–418. https://doi.org/10.1016/j.msec.2016.04.054. http://dx.doi.org/10.5599/admet.1172 https://doi.org/10.3389/fmicb.2015.00189 https://doi.org/10.3389/fmicb.2015.00189 https://doi.org/10.1111/jam.12589 https://doi.org/10.1111/jam.12589 https://doi.org/10.1007/s13204-014-0301-x https://doi.org/10.3390/molecules25122783 https://doi.org/10.1016/j.cej.2021.129071 https://doi.org/10.1021/acs.jafc.5b04559 https://doi.org/10.1016/j.jdent.2020.103453 https://doi.org/10.1016/j.jes.2017.04.034 https://doi.org/10.2147/ijn.s41371 https://doi.org/10.2147/ijn.s72923 https://doi.org/10.2147/ijn.s72923 https://doi.org/10.3390/ijms22094595 https://doi.org/10.3390/ijms22094595 https://doi.org/10.1016/j.eurpolymj.2006.10.011 https://doi.org/10.3390/nano8020060 https://doi.org/10.1016/j.carbpol.2017.11.109 https://doi.org/10.1007/s11095-012-0898-y https://doi.org/10.1016/j.msec.2016.04.054 ozdal and gurkok admet & dmpk 10(2) (2022) 115-129 128 [65] g. păunica-panea, a. ficai, m. m. marin, s. marin, m. g. albu, v.d. constantin, c. dinu-pîrvu, z. vuluga, m.c. corobea, m.v. ghica. new collagen-dextran-zinc oxide composites for wound dressing. j. nanomater. 2016 (2016) 1-7. https://doi.org/10.1155/2016/5805034. [66] d. kharaghani, d. dutta, p. gitigard, y. tamada, a. katagiri, d.-n. phan, m. d. p. willcox, i. soo kim. development of antibacterial contact lenses containing metallic nanoparticles. polym. test. 79 (2019) 106034. https://doi.org/10.1016/j.polymertesting.2019.106034. [67] s. bhattacharjee, r. joshi, a.a. chughtai, c.r. macintyre. grapheneand nanoparticle-embedded antimicrobial and biocompatible cotton/silk fabrics for protective clothing. acs appl. bio mater. 4 (2021) 6175–6185. https://doi.org/10.1021/acsabm.1c00508. [68] s. kalita, r. kandimalla, a.c. bhowal, j. kotoky, s. kundu. functionalization of β-lactam antibiotic on lysozyme capped gold nanoclusters retrogress mrsa and its persisters following awakening. sci. rep. 8 (2018) 1–13. https://doi.org/10.1038/s41598-018-22736-5. [69] m. konop, j. czuwara, e. kłodzińska, a.k. laskowska, d. sulejczak, t. damps, u. zielenkiewicz, i. brzozowska, a. sureda, t. kowalkowski, r. a. schwartz, l. rudnicka. evaluation of keratin biomaterial containing silver nanoparticles as a potential wound dressing in full-thickness skin wound model in diabetic mice. j. tissue eng. regen. med. 14 (2020) 334–346. https://doi.org/10.1002/term.2998. [70] a. gupta, s.m. briffa, s. swingler, h. gibson, v. kannappan, g. adamus, m. kowalczuk, c. martin, i. radecka. synthesis of silver nanoparticles using curcumin-cyclodextrins loaded into bacterial cellulose-based hydrogels for wound dressing applications. biomacromolecules 21 (2020) 1802– 1811. https://doi.org/10.1021/acs.biomac.9b01724. [71] n. eslami, f. fasihi, a. jamalinasab, f. ahrari. biocompatibility of several colloidal solutions containing nanoparticles on human gingival fibroblasts. dent. res. j. 18 (2021) 8. https://doi.org/10.4103/17353327.310037. [72] e. alpaslan, b. m. geilich, h. yazici, t.j. webster. ph-controlled cerium oxide nanoparticle inhibition of both gram-positive and gram-negative bacteria growth. sci. rep. 7 (2017) 1–12. https://doi.org/10.1038/srep45859. [73] c. ying sun, y. jun che, s. jin lu. preparation and application of collagen scaffold-encapsulated silver nanoparticles and bone morphogenetic protein 2 for enhancing the repair of infected bone.biotechnol. lett. 37 (2015) 467–473. https://doi.org/10.1007/s10529-014-1698-8. [74] p.a. dalavi, a. prabhu, r.p. shastry, j. venkatesan. microspheres containing biosynthesized silver nanoparticles with alginate-nano hydroxyapatite for biomedical applications. j. biomater. sci. polym. ed. 31 (2020) 2025–2043. https://doi.org/10.1080/09205063.2020.1793464. [75] x.x. li, j.y. dong, y.h. li, j. zhong, h. yu, q.q. yu, m. lei. fabrication of ag–zno@ carboxymethyl cellulose/k-carrageenan/graphene oxide/konjac glucomannan hydrogel for effective wound dressing in nursing care for diabetic foot ulcers. appl. nanosci. 10 (2020) 729–738. https://doi.org/10.1007/s13204-019-01194-z. [76] o. craciunescu, a.m. seciu, o. zarnescu. in vitro and in vivo evaluation of a biomimetic scaffold embedding silver nanoparticles for improved treatment of oral lesions. mater. sci. eng. c 123 (2021) 112015. https://doi.org/10.1016/j.msec.2021.112015. [77] j.t.t. angelina, s. ganesan, t.m.r. panicker, r. narayani, m. paul korath, k. jagadeesan. pulsed laser deposition of silver nanoparticles on prosthetic heart valve material to prevent bacterial infection. mater. technol. 32 (2017) 148–155. https://doi.org/10.1080/10667857.2016.1160503. [78] c.b. hiragond, a.s. kshirsagar, v.v. dhapte, t. khanna, p. joshi, p.v. more. enhanced anti-microbial response of commercial face mask using colloidal silver nanoparticles. vacuum 156 (2018) 475–482. https://doi.org/10.1016/j.vacuum.2018.08.007. [79] v.a. spirescu, c. chircov, a.m. grumezescu, e. andronescu. polymeric nanoparticles for antimicrobial therapies: an up-to-date overview. polymers 13 (2021) 724. https://doi.org/10.3390/polym13050724. https://doi.org/10.1155/2016/5805034 https://doi.org/10.1016/j.polymertesting.2019.106034 https://doi.org/10.1021/acsabm.1c00508 https://doi.org/10.1038/s41598-018-22736-5 https://doi.org/10.1002/term.2998 https://doi.org/10.1021/acs.biomac.9b01724 https://doi.org/10.4103/1735-3327.310037 https://doi.org/10.4103/1735-3327.310037 https://doi.org/10.1038/srep45859 https://doi.org/10.1038/srep45859 https://doi.org/10.1007/s10529-014-1698-8 https://doi.org/10.1080/09205063.2020.1793464 https://doi.org/10.1007/s13204-019-01194-z https://doi.org/10.1007/s13204-019-01194-z https://doi.org/10.1016/j.msec.2021.112015 https://doi.org/10.1080/10667857.2016.1160503 https://doi.org/10.1016/j.vacuum.2018.08.007 https://doi.org/10.3390/polym13050724 https://doi.org/10.3390/polym13050724 admet & dmpk 10(2) (2022) 115-129 antimicrobial nanoparticles against bacteria doi: http://dx.doi.org/10.5599/admet.1172 129 [80] f. wahid, c. zhong, h.-s. wang, x.-h. hu, l.-q. chu. recent advances in antimicrobial hydrogels containing metal ions and metals/metal oxide nanoparticles. polymers 9 (2017) 636. https://doi.org/10.3390/polym9120636. [81] p. jelinkova, a. mazumdar, v.p. sur, s. kociova, k. dolezelikova, a.m.j. jimenez, z. koudelkova, p.k. mishra, k. smerkova, z. heger. nanoparticle-drug conjugates treating bacterial infections. j. control. release 307 (2019) 166–185. https://doi.org/10.1016/j.jconrel.2019.06.013. [82] n.-y. lee, w.-c. ko, and p.-r. hsueh. nanoparticles in the treatment of infections caused by multidrug-resistant organisms. front. pharmacol. 10 (2019) 1153. https://doi.org/10.3389/fphar.2019.01153. [83] z. lin, n.a. monteiro‐riviere, j.e. riviere. pharmacokinetics of metallic nanoparticles. wires nanomedicine and nanobiotechnology 7 (2015) 189–217. https://doi.org/10.1002/wnan.1304. [84] w.k. boyes, b.l.m. thornton, s.r. al-abed, c.p. andersen, d.c. bouchard, et al. a comprehensive framework for evaluating the environmental health and safety implications of engineered nanomaterials. crit. rev. toxicol. 47 (2017) 767-810. https://doi.org/10.1080/10408444.2017.1328400. [85] a. ivask, k. juganson, o. bondarenko, m. mortimer, v. aruoja, k. kasemets. mechanisms of toxic action of ag, zno and cuo nanoparticles to selected ecotoxicological test organisms and mammalian cells in vitro: a comparative review. nanotoxicology 5390 (2013) 1-15. https://doi.org/10.3109/17435390.2013.855831. [86] c.a. dos santos, m.m. seckler, a.p. ingle, i. gupta, s. galdiero, m. galdiero, a. gade, m. rai. silver nanoparticles: therapeutical uses, toxicity, and safety issues. j. pharm. sci. 103 (2014) 1931-1944. https://doi.org/10.1002/jps.24001. [87] w.i. hagens, a.g. oomen, w.h. de jong, f.r. cassee, a.j. sips. what do we (need to) know about the kinetic properties of nanoparticles in the body?. regul. toxicol. pharmacol. 49 (2007) 217-229. https://doi.org/10.1016/j.yrtph.2007.07.006. [88] s.e. leucuta. systemic and biophase bioavailability and pharmacokinetics of nanoparticulate drug delivery systems. curr. drug deliv. 10 (2013) 208-240. https://doi.org/10.2174/1567201811310020007. [89] t. wu, m. tang. the inflammatory response to silver and titanium dioxide nanoparticles in the central nervous system. nanomedicine 13 (2018) 233–249. https://doi.org/10.2217/nnm-2017-0270. [90] a. herman, a. p. herman. nanoparticles as antimicrobial agents: their toxicity and mechanisms of action. j. nanosci. nanotechnol. 14 (2014) 946–957. https://doi.org/10.1166/jnn.2014.9054. [91] y. wang, m. tang. dysfunction of various organelles provokes multiple cell death after quantum dot exposure. int. j. nanomed. 13 (2018) 2729-2742. https://doi.org/10.2147/ijn.s157135. [92] n. liu, m. tang. toxic effects and involved molecular pathways of nanoparticles on cells and subcellular organelles. j. appl. toxicol. 40 (2020) 16–36. https://doi.org/10.1002/jat.3817. [93] b. mangalampalli, n. dumala, r. perumalla venkata, p. grover. genotoxicity, biochemical, and biodistribution studies of magnesium oxide nano and microparticles in albino wistar rats after 28-day repeated oral exposure. environ. toxicol. 33 (2018) 396–410. https://doi.org/10.1002/tox.22526. [94] r. lei, c. wu, b. yang, h. ma, c. shi, q. wang, q. wang, y. yuan, m. liao. integrated metabolomic analysis of the nano-sized copper particle-induced hepatotoxicity and nephrotoxicity in rats: a rapid in vivo screening method for nanotoxicity. toxicol. appl. pharmacol. 232 (2008) 292-301. https://doi.org/10.1016/j.taap.2008.06.026. [95] v. mihailovic, j. s. katanic stankovic, d. selakovic, g. rosic. an overview of the beneficial role of antioxidants in the treatment of nanoparticle-induced toxicities. oxid. med. cell. longev. 2021 (2021) 1–21. https://doi.org/10.1155/2021/7244677. ©2022 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://dx.doi.org/10.5599/admet.1172 https://doi.org/10.3390/polym9120636 https://doi.org/10.3390/polym9120636 https://doi.org/10.1016/j.jconrel.2019.06.013 https://dx.doi.org/10.3389/fphar.2019.01153 https://dx.doi.org/10.3389/fphar.2019.01153 https://doi.org/10.1002/wnan.1304 https://doi.org/10.1080/10408444.2017.1328400 https://doi.org/10.1080/10408444.2017.1328400 https://doi.org/10.3109/17435390.2013.855831 https://doi.org/10.3109/17435390.2013.855831 https://doi.org/10.1002/jps.24001 https://doi.org/10.1016/j.yrtph.2007.07.006 https://doi.org/10.2174/1567201811310020007 https://doi.org/10.2174/1567201811310020007 https://doi.org/10.2217/nnm-2017-0270 https://doi.org/10.1166/jnn.2014.9054 https://doi.org/10.2147/ijn.s157135 https://doi.org/10.1002/jat.3817 https://doi.org/10.1002/tox.22526 https://doi.org/10.1016/j.taap.2008.06.026 https://doi.org/10.1155/2021/7244677 http://creativecommons.org/licenses/by/3.0/ manuscript doi: 10.5599/admet.3.3.199 235 admet & dmpk 3(3) (2015) 235-241; doi: 10.5599/admet.3.3.199 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index review transcranial direct current stimulation in patients with alzheimer’s disease: challenges and responses hong yuan* 1 , serik tabarak 1 , jing yu, xu lei faculty of psychology, southwest university; key laboratory of cognition and personality; key research institute of humanities & social science, chongqing, 400715, china. *corresponding author: e-mail: yuanyh@swu.edu.cn; tel.: +86-023-6836-7489; fax: +86-023-6825-3629 1 these authors contributed equally to this work. received: july 06, 2015; revised: august 18, 2015; published: september 05, 2015 abstract the use of transcranial direct current stimulation (tdcs) as a noninvasive therapeutic approach for alzheimer’s disease (ad) has gained increasing attention. research regarding the utility of tdcs in ad is inconsistent. in this study, we reviewed the importance of individual diversity among ad patients, starting with uninformative mean results. we also demonstrated variations among ad patients. highly educated patients seemed to benefit more from therapy; education also seems to modulate baseline measurements and results. individual cortical morphology also affects current distribution, which influences the effectiveness of stimulation. we suggest using structural mri to distinguish inter-individual variability; high-resolution modeling may also be used to predict current distributions and should be combined with cognitive training (ct) and tdcs. keywords individual diversity; cortical morphology; educational level; current distribution; cognitive training introduction alzheimer’s disease (ad) is a major form of age-related dementia and currently affects over 35 million people worldwide (http://www.alz.co.uk/research/world-report). it is associated with impaired episodic memory, as well as other cognitive abilities such as word retrieval, language comprehension, calculation, visuospatial orientation, learning capacity, abstract thinking and judgment [1]. moreover, it induces changes in personality and disturbs mood [2]. currently, there is no effective therapy for ad, and pharmacological treatments are associated with significant adverse effects [3]. the substantial costs and social burden of ad demand alternative or complementary and low-cost therapeutic treatments for this disease. in this context, non-pharmacological interventions have gained increasing attention in recent years. such interventions include non-invasive brain stimulation techniques such as transcranial direct current stimulation (tdcs) and transcranial magnetic stimulation (tms) [4], particularly in combination with cognitive training [5]. in tdcs, the current is passed through saline-soaked sponge electrodes on the scalp, penetrates the skull and influences neuronal function. cathodal tdcs (ctdcs) suppresses cortical activity, anodal tdcs (atdcs) mailto:yuanyh@swu.edu.cn http://www.alz.co.uk/research/world-report yuan et al. admet & dmpk 3(3) (2015) 235-241 236 has an activating effect, and sham current is used as a placebo [6]. as a type of non-invasive brain stimulation, tdcs helps to improve various cognitive functions, including motor activity [7], executive function and language ability [8]. compared with rtms, which is applied above the active threshold, tdcs modulates brain activity through low-intensity electrical currents (1 to 2 ma); tdcs also has fewer side effects and is less invasive, less expensive, easier to use, better controlled, and potentially portable [7]. moreover, with respect to randomized, double-blinded trials, it is difficult to distinguish real tdcs from sham stimulation because of its subthreshold modulation. the mechanism of tdcs in ad may be related to brain-derived neurotrophic factor (bdnf). bdnf is required for the survival and synaptic growth of cells compromised by ad [9]; decreased bdnf levels are related to oligomeric amyloid, the major toxic species in ad [10], and correlate with neurodegeneration [11-13]. decreased bdnf levels also cause defects in long-term potentiation and memory, which correlate with cognitive decline [14-16]. tdcs induces changes in the neuronal membrane resting threshold and contributes to plasticity changes [8]. these effects may occur via the bdnf gene [17]. the existing reviews of tdcs for ad have focused on study design, stimulus parameters and locations [3,8,18]. notably, before these reviews, all of the studies regarding tdcs in ad have demonstrated its effect. three additional studies regarding tdcs in ad were published after these reviews, but the results of these studies were inconsistent. herein, we discuss variations observed among affected individuals, which may determine the effectiveness of tdcs in the treatment of ad. these variations include differences in the local structure of both grey and white matter, educational level and cortical morphology. studying these factors may help us to better understand the scope of the application of tdcs and its limitations. variability of individual response mean results may be uninformative in behavioral experiments, particularly when the inclusion criteria for the subjects are based only on a cognitive neuropsychology test. researchers typically focus on changes in the mean response associated with an experimental manipulation. however, this averaging ignores large variations in individual responses [19]. according to our screening data (table 1) in which we used mean responses, the effects of tdcs in ad were mixed. three studies observed a primary effect when they compared anodal and either cathodal or sham groups. regarding these studies, one observed that atdcs significantly improved recognition memory. cathodal tdcs significantly decreased word recognition accuracy, whereas sham tdcs did not affect word accuracy. another study noted that atdcs significantly improved visual recognition memory compared with sham stimulation. moreover, atdcs also improved visual recognition memory by 8.99 % from baseline, whereas sham stimulation decreased visual recognition memory by 2.62 % [20-22]. the remaining three studies with parallel-sham controlled designs did not observe these effects. in one study, tdcs did not significantly ameliorate memory performance compared with sham tdcs [5]. the other study found no significant differences in apathy after 1 and 2 weeks compared with baseline [23]. the last study found that tdcs improved cognitive function via both anodal and cathodal interventions [24]. we believe that the differences in individual responses are an important source of variation. furthermore, these types of individual differences may be more obvious in older adults because individual differences in life experience and cognitive abilities increase with age [25]. these differences may be enhanced when parallel-sham controlled study designs are used [5,23,24], as presented in table 1. mri studies of the human brain have demonstrated inter-individual variability in a wide range of basic and higher cognitive functions, including perception, motor control, memory, aspects of consciousness and introspection. this variability may be predicted based on the local structure of grey and white matter, as admet & dmpk 3(3) (2015) 235-241 transcranial direct current stimulation in ad doi: 10.5599/admet.3.3.199 237 assessed by either voxel-based morphometry or diffusion tensor imaging [19]. because tdcs is a neuromodulatory tool that influences neuronal function by applying current to specific neural structures, it is important to adequately consider inter-individual differences in the aspects described below. table 1. list of patient backgrounds in studies of the use of tdcs as an intervention for ad. author & year study design subjects & sex (%, m) screening methods age diagnosis education & years results (cognitive task) ferrucci et al. 2008 crossoversham controlled 10, 30 mmse 22.7±1.8 75.2±7.3 nincds-adrda & dsm-iv [probable] 10.9±4.8 atdcs > basline ctdcs < basline sham (word recognition task) boggio et al. 2009 crossoversham controlled 10, 40 mmse 17±4.9 79.1±8.8 nincds-adrda [mild-moderate] 8.7±4.9 atdcs > sham (visual recognition task) boggio et al. 2012 crossoversham controlled 15, 53 mmse ①21.0±3.2 ②19.0±2.8 77.5±6.9 80.6±9.5 nincds-adrda & dsm-iv 13.3±4.8 15.7± 0.8 atdcs > basline sham < basline (picture recognition task) cotelli et al. 2014 parallelsham controlled 36, 19 mmse 20.1±2.4 76.6±4.6(ac) 74.7±6.1(pc) 78.2±5.2(am) nincds-adrda [mild-moderate] 5.5±2.4(ac) 8.9±5.1(pc) 6.3±2.6(am) ac > am pc > am (face naming task) suemoto et al. 2014 parallelsham controlled 40, 30 mmse 15.2±2.9 80.5±7.5 nincds-adrda; apathy scale>14 [moderate severe] 4.8±4.0 atdcs sham (apathy scores) khedr et al. 2014 parallelsham controlled 34, 55 mmse 18.1±3.3 69.7±4.8 nincds-adrda [mild-moderate] illiterate atdcs > sham ctdcs > sham (mmse scores) the following variables were extracted from all studies when pertinent and available: (a) study design, (b) demographic variability of aging, including the number of participants, mean age, gender, and educational level, and (c) screening baseline and neuropsychological diagnosis. nincds-adrda: national institute of neurology and communication disorders and stroke-the alzheimer's disease and related disorders association criteria; dsm-iv: diagnostic and statistical manual–iv; ①: patients from italy; ②: patients from brazil; mmse: mini-mental state examination; ac: atdcs plus ct; pc: sham tdcs plus ct; am: atdcs plus motor training. -: no effect. education education may serve as an important reference point for differences in individual life experience. older adults with higher education levels have benefited from tdcs stimulation during a working memory task, whereas individuals with lower education levels did not experience any benefits [26]. as presented in table 1, highly educated patients seem benefit more with tdcs [20-22]. one testable hypothesis for this difference is that highly educated patients employ a different strategy than less educated patients. this strategy may enable these highly educated patients to better recruit prefrontal cortex (pfc) structures during specifics tasks. support for this interpretation comes from a recent neuroimaging study that reported greater pfc activation in expert pilots but not novice pilots performing a track-following task [27]. this finding highlights the need for ct to develop different strategies to better recruit the pfc; we recommend combining tdcs with ct, which we will discuss in more detail below. effective stimulation and measurements are needed both at baseline and after treatment. the minijavascript:void(0); yuan et al. admet & dmpk 3(3) (2015) 235-241 238 mental status examination (mmse), which is used in the majority of studies, is a commonly administered measure of global cognitive functioning that is thought to be sensitive to dementia [5,20-24,28]. compared with the traditional cut-off score of 24 for highly educated individuals, a cut-off score of 27 or more generates better estimates among individuals of different ethnicities [29], as less educated and older individuals tend to receive lower scores [30,31], and among illiterate patients, as both atdcs and ctdcs facilitate improvements in mmse scores compared with sham tdcs [24]. therefore, educational level also modulates participant recruitment and outcome measures. to effectively compare results across studies, matching patients’ backgrounds, including their educations, is important. moreover, it may also be useful to conduct an iq test in addition to the mmse. neuropathological variables another critical issue is individual differences in cortical morphology, which influence current flow distribution [32-34]. in ad, cortical atrophy, synaptic damage and ventricular enlargement [35] lead to changes in the volume of cerebrospinal fluid (csf), which exhibits greater conductivity and therefore significantly alters current pathways [32]. the current parameters used in ad patients are based on models of healthy subjects and clinical trial outcomes, which ignore individual variations in cortical morphology [36]. moreover, an individual model is supposed to predict current distribution, so that it may effectively modulate the target area. the first modeling study used a cad-rendered current flow distribution in the cortex of both healthy and stroke subjects [33]. follow-up studies considered skull anisotropy, white matter, and gyri resolution and demonstrated the importance of anatomical influences on current flow [37,38]. nevertheless, these models are often not utilized, most likely because they are difficult to compute. among the studies of tdcs in ad patients that did not use the modeling approach, halko et al. observed significant correlations between region-specific fmri signals and the locations of stimulating electrodes, which were placed based on individualized mri head models [39]. for individual models to be routinely used in tdcs studies, they must be easy to derive, and researchers must understand the importance of accurate head models for the determination of current flow and distribution in each patient. this understanding will naturally lead to more studies that may be compared. to better promote individual models, more studies are needed in which predicted current flow distributions are compared with outcomes. this type of model also offers the possibility of reaching deeper structures such as the hippocampus. combining tdcs with cognitive training given the increasing use of ct in ad, combining tdcs with ct may be even more beneficial. the term "cognitive training" is based on the use of structured methods rather than unspecific and freely managed approaches. the choice of approach depends on the study’s purpose. ct may be divided into two basic categories as follows: compensatory and restorative [40,41]. compensatory strategies aim to teach new ways of performing cognitive tasks by working around cognitive deficits. restorative strategies attempt to improve functioning in specific domains, with the ultimate goal of returning function in these domains to premorbid levels. both strategies apply equally well to ad. to ameliorate memory impairment in ad and improve recall, the compensatory approach may involve taking relevant notes, and the restorative approach may involve repeated cuing and questioning [42]. in a study of mild to moderate ad patients, patients were randomly assigned to one of the following three study groups: anodal tdcs+individualized computerized (ic) memory training, sham tdcs+ic memory training, and anodal tdcs+motor training. tdcs was then administered during memory training of face– admet & dmpk 3(3) (2015) 235-241 transcranial direct current stimulation in ad doi: 10.5599/admet.3.3.199 239 name associations (a type of ct). there was general improvement in ic memory performance in both the anodal and sham groups after 2 weeks; this effect was still significant after 12 weeks in the sham group [5]. a case study compared tdcs+ct with sham+ct in individuals with mild ad for 2 months. the results revealed that the combination of tdcs+ct had limited effects but stabilized patients’ global cognitive function for approximately 3 months and therefore slowed their cognitive decline [28]. only two studies have combined tdcs with ct; one was a case study, a form of preliminary research. more studies are needed to determine the optimal combination of these two therapies, particularly with respect to the order and timing of the therapies. these two studies used different orders of stimulation; one used tdcs before ct [28], and the other used tdcs during ct [5]. older subjects seem to be more sensitive to time course differences compared with younger subjects. one of the studies demonstrated that stimulation before and during the task in young individual exerted equivalent effects, but improvement in picture naming was observed only in the elderly if tdcs was administered during the task [43]. anodal stimulation during the task may mediate activity-dependent calcium influx and therefore influence task-related plasticity, whereas anodal stimulation before the task may negate task-related plasticity [44]. nonetheless, these studies are a useful reference for future studies of different ct tasks. conclusion due to growing evidence that tdcs is a promising neurostimulation tool for rehabilitation, tdcs has been used in many studies. however, in studies of tdcs in ad, there has been a wide range of interindividual variability, including the education level of older adults; the mean results of behavioral evaluations may be insufficient to prove its effect. additionally, individual differences in brain anatomy are extensive; these differences are increased in ad due to the course of the disease and also result in different cognitive disorders in different patients. mri imaging of the brain is needed to understand individual differences before treatment to apply a precise dosage to the appropriate target region to optimize current flow. future research should analyze individual baselines and responses and use randomized block designs to explore its different levels. briefly, as an age-related disease among older adults, ad is suitable for the use of non-invasive tools with relatively small side effects such as tdcs, particularly in combination with ct. additionally, because it enables flexible dosing, is inexpensive and is well tolerated, we recommend that structural mri be combined with behavioral data to improve and optimize the effects of tdcs. acknowledgements authors are very much thankful to key research institute of humanities & social science (11skb21), southwest university, chongqing and national natural science foundation of china (nsfc 31300856) for providing financial assistance to carry out this research work. references [1] b.c. dickerson, h. eichenbaum, neuropsychopharmacology 35(1) (2010) 86-104. [2] j.l. cummings, h.v. vinters, g.m. cole, z.s. khachaturian, neurology 51(1 suppl 1) (1998) s2-17; discussion s65-67. [3] r. nardone, j. bergmann, m. christova, f. caleri, f. tezzon, g. ladurner, e. trinka, s. golaszewski, int. j. alzheimer's dis. 2012 (2012) 687909. [4] a. floel, l.g. cohen, neurobiol. dis. 37(2) (2010) 243-251. yuan et al. admet & dmpk 3(3) (2015) 235-241 240 [5] m. cotelli, r. manenti, m. brambilla, m. petesi, s. rosini, c. ferrari, o. zanetti, c. miniussi, front. aging neurosci. 6:38 (2014). [6] m.a. nitsche, w. paulus, j. physiol. 527 pt 3 (2000) 633-639. [7] m.a. nitsche, l.g. cohen, e.m. wassermann, a. priori, n. lang, a. antal, w. paulus, f. hummel, p.s. boggio, f. fregni, a. pascual-leone, brain stimul. 1(3) (2008) 206-223. [8] p.s. boggio, c.a. valasek, c. campanhã, a.c.a. giglio, n.i. baptista, o.m.lapenta, f. fregni, neuropsychol. rehabil. 21(5) (2011) 703-716. [9] j. yuan, b.a. yankner, nature 407(6805) (2000) 802-809. [10] m.p. lambert, a.k. barlow, b.a. chromy, c. edwards, r. freed, m. liosatos, t.e. morgan, i. rozovsky, b. trommer, k.l. viola, p. wals, c. zhang, c.e. finch, g.a. krafft, w.l. klein, proc. natl. acad. sci. usa 95(11) (1998) 6448-6453. [11] y. gong, l. chang, k.l. viola, p.n. lacor, m.p. lambert, c.e. finch, g.a. krafft, w.l. klein, proc. natl. acad. sci. usa 100(18) (2003) 10417-10422. [12] l.f. lue, y.-m. kuo, a.e. roher, l. brachova, y. shen, l. sue, t. beach, j.h. kurth, r.e. rydel, j. rogers, am. j. pathol. 155(3) (1999) 853-862. [13] c.a. mclean, r.a. cherny, f.w. fraser, s.j. fuller, m.j. smith, k. vbeyreuther, a.i. bush, c.l. masters, ann. neurol. 46(6) (1999) 860-866. [14] m. korte, p. carroll, e. wolf, g. brem, h. thoenen, t. bonhoeffer, proc. natl. acad. sci. usa 92(19) (1995) 8856-8860. [15] k. olofsdotter, o. lindvall, f. asztely, neuroscience 101(3) (2000) 531-539. [16] s.l. patterson, t. abel, t.a.s. deuel, k.c. martin, j.c. rose, e.r. kandel, neuron 16(6) (1996) 11371145. [17] a. antal, l. chaieb, v. moliadze, k. monte-silva, c. poreisz, n. thirugnanasambandam, m.a. nitsche, m. shoukier, h. ludwig, w. paulus, brain stimul. 3(4) (2010) 230-237. [18] c. freitas, h. mondragon-llorca, a. pascual-leone, exp. gerontol. 46(8) (2011) 611-627. [19] r. kanai, g. rees, nat. rev. neurosci. 12(4) (2011) 231-242. [20] p.s. boggio, r. ferrucci, f. mameli, d. martins, o. martins, m. vergari, l. tadini, e. scarpini, f. fregni, a. priori, brain stimul. 5(3) (2012) 223-230. [21] p.s. boggio, l.p. khoury, d.c.s. martins, o.e.m.s martins, e.c. de macedo, f. fregni, j. neurol. neurosurg. psychiatry 80(4) (2009) 444-447. [22] r. ferrucci, f. mameli, i. guidi, s. mrakić-sposta, m. vergari, s. marceglia, f. cogiamanian, s. barbieri, e. scarpini, a. priori, neurology 71(7) (2008) 493-498. [23] c.k. suemoto, d. apolinario, e.m. nakamura-palacios, l. lopes, r.e.p. leite, m.c. sales, r. nitrini, s.m. brucki, l.s. morillo, r.m. magaldi, f. fregni, brain stimul. 7(2) (2014) 308-313. [24] e.m. khedr, n.f.e. gamal, n.a. el-fetoh, h. khalifa, e.m. ahmed, a.m. ali, m. noaman, a.a. el-baki, a.a. karim, front. aging neurosci. 6:275 (2014). [25] a. gutchess, science 346(6209) (2014) 579-582. [26] m.e. berryhill, k.t. jones, neurosci. lett. 521(2) (2012) 148-151. [27] m. peres, p.f. van de moortele, c. pierard, s. lehericy, p. satabin, d. le bihan, c.y: guezennec, aviat. space environ. med. 71(12) (2000) 1218-1231. [28] b. penolazzi, s. bergamaschi, m. pastore, d. villani, g. sartori, s. mondini, neuropsychol. rehabil. (2014) 1-19. [29] c.c. spering, v. hobson, j.a. lucas, c.v. menon, j.r. hall, s.e. o’bryant, j. gerontol. a biol. sci. med. sci. 67(8) (2012) 890-896. [30] g.m. nys, m.j.e. van zandvoort, p.l.m. de kort, b.p.w. jansen, l.j. kappelle, e.h.f. de haan, arch. clin. neuropsychol. 20(5) (2005) 623-629. [31] t.n. tombaugh, n.j. mcintyre, j. am. geriatr. soc. 40(9) (1992) 922-935. [32] m. bikson, a. rahman, a. datta, f. fregni, l. merabet, neuromodulation 15(4) (2012) 306-315. admet & dmpk 3(3) (2015) 235-241 transcranial direct current stimulation in ad doi: 10.5599/admet.3.3.199 241 [33] t. wagner, f. fregni, s. fecteau, a. grodzinsky, m. zahn, a. pascual-leone, neuroimage 35(3) (2007) 1113-1124. [34] t. wagner, a. valero-cabre, a. pascual-leone, annu. rev. biomed. eng. 9 (2007) 527-565. [35] g.b. frisoni, n.c. fox, c.r. jack jr, p. scheltens, p.m. thompson, nat. rev. neurol. 6(2) (2010) 67-77. [36] s. mahdavi, f. yavari, s. gharibzadeh, f. towhidkhah, front. comput. neurosci. 8:72 (2014). [37] a. datta, v. bansal, j. diaz, j. patel, d. reato, m. bikson, brain stimul. 2(4) (2009) 201-207.e1. [38] h.s. suh, et al., conf. proc. ieee eng. med. biol. soc. (2009) 638-641. [39] m.a. halko, a. datta, e.b. plow, j. scaturro, m. bikson, l.b. merabet, neuroimage 57(3) (2011) 885891. [40] a.s. bellack, j.m. gold, r.w. buchanan, schizophr. bull. 25(2) (1999) 257-74. [41] w.d. spaulding, s.k. fleming, d. reed, m. sullivan, d. storzbach, m. lam, schizophr. bull. 25(2) (1999) 275-289. [42] d.i. sitzer, e.w. twamley, d.v. jeste, acta psychiatr. scand. 114(2) (2006) 75-90. [43] a. fertonani, m. brambilla, m. cotelli, c. miniussi, front. aging neurosci. 6:131 (2014). [44] u. ziemann, h.r. siebner, brain stimul. 1(1) (2008) 60-66. ©2015 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) recent advances in electrochemical sensors and biosensors for monitoring drugs and metabolites in pharmaceutical and biological samples doi: https://doi.org/10.5599/admet.1709 151 admet & dmpk 11(2) (2023) 151-173; doi: https://doi.org/10.5599/admet.1709 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index review recent advances in electrochemical sensors and biosensors for monitoring drugs and metabolites in pharmaceutical and biological samples charaf laghlimi*,1, abdelaziz moutcine2, abdelilah chtaini2, jalal isaad1, adil soufi1, younes ziat3, hassan amhamdi4 and hamza belkhanchi3 1erci2a, fsth, abdelmalek essaadi university, tetouan, morocco 2molecular electrochemistry and inorganic materials team, faculty of science and technology of beni mellal, sultan moulay slimane university, morocco 3laboratory of engineering and applied technologies, higher school of technology, sultan moulay slimane university, beni mellal, morocco 4applied chemistry team, fsth, abdelmalek essaadi university, tetouan, morocco *corresponding author: e-mail: charaf.cac.fbs@gmail.com;c.laghlimi@uae.ac.ma; tel.: +0706182632 received: february 15, 2023; revised: april 7, 2023; published: may 1, 2023 abstract various applications of electrochemical sensors and biosensors have been reported in many fields. these include pharmaceuticals, drug detection, cancer detection, and analysis of toxic elements in tap water. electrochemical sensors are characterised by their low cost, ease of manufacture, rapid analysis, small size and ability to detect multiple elements simultaneously. they also allow the reaction mechanisms of analytes, such as drugs, to be taken into account, giving a first indication of their fate in the body or their pharmaceutical preparation. several materials are used in the construction of sensors, such as graphene, fullerene, carbon nanotubes, carbon graphite, glassy carbon, carbon clay, graphene oxide, reduced graphene oxide, and metals. this review covers the most recent progress in electrochemical sensors used to analyze drugs and metabolites in pharmaceutical and biological samples. we have highlighted carbon paste electrodes (cpe), glassy carbon electrodes (gce), screen-printed carbon electrodes (spce) and reduced graphene oxide electrodes (rgoe). the sensitivity and analysis speed of electrochemical sensors can be improved by modifying them with conductive materials. different materials used for modification have been reported and demonstrated, such as molecularly imprinted polymers, multiwalled carbon nanotubes, fullerene (c60), iron(iii) nanoparticles (fe3o4np), and cuo micro-fragments (cuo mf). manufacturing strategies and the detection limit of each sensor have been reported. ©2023 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords pharmaceuticals and drugs; screen printed carbon electrodes; glassy carbon electrodes, carbon nanotubes introduction pharmacokinetics aims to determine a drug's pathway from administration to excretion [1]. for clinical diagnosis, there is a need for information on how medicines work in the human body and the dose that should be administered [2]. it is, therefore, necessary to quantify drugs in physiological fluids such as serum, urine, sweat and saliva, as well as in pharmaceutical tablets. several sensitive and accurate methods have been used https://doi.org/10.5599/admet.1709 https://doi.org/10.5599/admet.1709 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:charaf.cac.fbs@gmail.com mailto:c.laghlimi@uae.ac.ma http://creativecommons.org/licenses/by/4.0/ c. laghlimi et al. admet & dmpk 11(2) (2023) 151-173 152 to quantify and detect different types of drugs, such as spectrophotometry [3,4], fluorimetry [3], rp-hplc [4], hplc-diode array detector, tlc-densitometric methods [5] and capillary zone electrophoresis [6]. however, these techniques are very expensive and require the use of expensive solvents, specialised technicians, long sample preparation times and expensive instrumentation, all of which increase the cost and duration of the analysis [2]. for this reason, the pharmacokinetic community strives to provide alternative methods of drug surveillance that are easy, inexpensive, sensitive and precise [7]. in the last decades, electrochemical sensors have become an interesting alternative technique due to their high capacity to analyse drugs in different physiological fluids [8-10]. graphene has considerable sensitivity and selectivity and often has a large potential window to accommodate the redox process involved. it also has a significant electron transfer process due to its planar sites and the presence of σ and π bonds, making it an attractive material for sensor manufacture [11,12]. modification of the carbon paste electrode offers high sensitivity, low cost and rapid analysis without any pre-treatment [13-21]. the modification is an effective way to reduce the overpotential and increase the sensitivity of the sensors [22]. various electrodes based on carbon as a conductor have been used for the electroanalysis of drugs [23,25]. in this review, we report on a new advance in carbon graphite-based electrochemical sensors used to analyse different types of drugs in pharmaceutical tablets and a variety of physiological fluids such as serum, blood, sweat and urine. the different modifiers used to improve sensor performance and increase the electroactive surface area are studied. sensitivity, detection limit and method of preparation have been highlighted to provide comprehensive information on all these techniques and their use in drug analysis. electrodes based on modified carbon paste for detecting various drugs and metabolites carbon paste electrodes (cpe) have a very large electroactive surface area, which can be renewed for a variety of applications. in addition, it has low ohmic resistance and high stability, reproducibility and lifetime, allowing it to analyse various drugs, revealing the oxidation and reduction processes involved due to its wide potential range. the sensitivity and speed of analysis of cpe can be significantly improved by modifying them with conductive materials. different types of modifiers have been used. the modified cpe has been used to examine various drugs in various samples, including human urine, pharmaceutical formulations, plasma and serum. the carbon paste electrode modified with poly(ebt) [26] exhibits very high electrocatalytic and semiconducting for detecting methdilazine hydrochloride (mdh), an antihistamine drug, in dilosyn syrup and human urine using the swv method (figure 1). the mdh oxidation peak is observed at 0.675v with higher intensity compared to unmodified cpe. the surface area of poly-ebt/cpe is calculated to be 0.097 cm2, which is 2.30 times higher than that of cpe. the developed sensor shows good accuracy in both media (average recovery is 98.14 % in dilosyn syrup and 97.4 % in human urine). the developed sensor has a low lod of the order of 10-8 mol l-1 in the range of 0.1-50 μmol l-1 compared to other methods such as visible spectrophotometry (3.23 μmol l-1) [27], ultra-high performance reversed phase chromatography (0.254 μmol l-1 ) [28] and spectrophotometric method (1.62 μmol l-1 ) [29]. the researchers [30] reported that cpe/nanozeolite type x outperformed cpe/nanozeolite type a with a lod of 0.2 μmol l-1 for paracetamol and 8 mmol l-1 for epinine. a recent study described the analysis of the antifungal agent ketoconazole (ktc) in pharmaceutical and urine samples using ce-btc mof/il/cpe (cf. table 1) [31]. the modifiers are synthesised according to the method of liu [32]. the electrochemical behaviour is studied by chronoamperometry (ca), differential pulse voltammetry (dpv), cyclic voltammetry (cv) and linear sweep voltammetry (lsv). the lod of ce-btc mof/il/cpe is 0.04 μmol l-1 in the range of 0.1-110.0 μmol l-1 . the sensitivity is found to be 0.1342 μa μmol-1 l. admet & dmpk 11(2) (2023) 151-173 electrochemical sensors and biosensors for monitoring drugs and metabolites doi: https://doi.org/10.5599/admet.1709 153 figure 1. schematic of a) ebt/cpe preparation, b) polymerisation steps of modified cpe, c) proposed mechanism of eriochrome black t (ebt) polymerisation and d) electrochemical detection of mdh by ebt/cpe. carried out at the base of the reference [26] with an order license id of 1346497-1 j. zoubir et al. [33] used silver nanoparticles to modify carbon graphite by electrodeposition. the fabricated sensor was used to detect metronidazole in milk and tap water with a detection limit of 0.206 μmol l-1 in the range 1 -1000 μmol l-1 . o. vajdle et al. [34] used the drop coating method to modify cpe with gold nanoparticles. aunps/cpe were used to analyse four macrolide antibiotics by swv. aunps/cpes are used for azithromycin (azi) detection in hemomycin® by the swv method. the oxidation peak of the azi is identified at 0.77 v on the aunps/cpe, while the roxithromycin (rox) peak is found at 0.65 v by cpe in the runac®. both electrodes were found to be valid by comparison with hplc-dad measurements. on the other hand, nitric acid and sulphuric acid are used to increase -oh and -cooh groups on the surface of mwcnts, which increases the surface area of the material [35]. in addition, the numerous carboxyl groups inhibit π-π interactions, which leads to a decrease in the adsorbency of mwcnts [36]. ofloxacin (ofx) adsorbs on mwcnt by binding its fluorine group to the -oh group of the cnt [37-40]. flake graphite (fg) is used to increase the conductivity of mwcnts weakened by the adsorption of compounds owing to its high degree of crystallisation. for this reason, m. elfiky et al. [41] prepared an electrode by mixing graphite powder with fg and mwcnts. the sensor has a larger electroactive area of 2.08 cm2 compared to the unmodified cpes, with only 1.14 cm2. this finding is due to cracks on the irregular compact layers of the electrode. the electrode presented by the [10%fg/5%mw] cpe shows excellent performance in the analysis of ofloxacin in a commercial formulation (ofloxacin® tablets) and in human urine samples with a lod of 0.18 nmol l-1. ofx peak is observed at 0.85v by sw-adas the analysis of azi, an alkaline chemical, by hplc presents a difficulty due to its adsorption on si-oh chromatographic materials [42]. therefore, electrochemiluminescence detection is considered an alternative method for the detection of this antibiotic due to its sensitivity and high productivity [43,44]. electrochemiluminescence (ecl) is based on the redox process involved in the formation of excited states that are able to emit light. the concept of molecularly imprinted polymers (mip) is based on imprinting a molecular cavity on the surface of the polymer using a template molecule, which is then removed while the https://doi.org/10.5599/admet.1709 https://www.sciencedirect.com/topics/chemistry/azithromycin https://www.sciencedirect.com/topics/chemistry/roxithromycin c. laghlimi et al. admet & dmpk 11(2) (2023) 151-173 154 polymer is still imprinted. this mip can then be used to capture and detect a target molecule similar to the template molecule and specifically binds to a functional group in the cavity in the same way that an antibody binds to its antigen. there are many ways to make mips, such as soft lithography applied to giant molecules. l. hu et al. [45] successfully analysed azithromycin (azi) in urine and serum samples using mip/cp ecl sensor, with mip used as a recognisor. the oxidation peak of the amino groups of azi [31] is observed at 0.8 v with lod of 23 pmol l-1 in the range of 0.10 400 nmol l-1. the analytical results show the mip/cp ecl sensor and hplc agreement. clay has attractive characteristics, like high specific surface area and cation exchange capacity [46,47]. bakary tigana djonse justin et al. [48] recently fabricated a titanium dioxide modified carbon clay paste electrode (cpea/tio2/uv) that allows the analysis of ascorbic acid in pharmaceutical tablets using cv.the sensitivity of the electrode is enhanced by photoactivation of tio2 by light irradiation (with a 100 w lamp), which produces electron (e)/hole (h+) pairs upon absorption of appropriate light energy [49]. the detection limit is 0.732 μmol l-1in the range of 0.15-0.850 μmol l-1 . v. vinoth et al. [50] modified a graphite carbon electrode with cuo microflowers (mfs) by solubilising them in nafion to facilitate the modification. this electrode was named cuo mfs/nafion/gc. the cuo mfs/nafion/gc exhibited perfect electrocatalytic properties, selectivity in the presence of interfering molecules, stability and reproducibility. the amperometric analysis of glucose showed a sensitivity of 3.1 μa μmol-1 l cm-2 and lod of 6.48 μmol l-1 between 10 to 120 μmol l-1 at +500 mv. this detection limit is very good compared to other sensors used for the same purpose such as nio-swcnt (907 μmol l-1 ) [51]. table 1 shows the different types of modified cpes as well as the drugs detected and their detection medium, the limit of detection (lod), concentration range and other characteristics. table 1. evaluation of electrochemical properties of modified cpe electrode analyte and matrix method linear dynamic range lod year / ref poly-ebt/cpe mdh in dilosyn syrup and human urine swv 0.150 μm 0.0257 μm 2020 / [26] cpe/nanozeolite type x paracetamol and epinine lsv 0.5-70.0 μm para = 0.2 μm epi = 0.008 μm 2023 / [30] ce-btc mof/il/cpe ktc in pharmaceutical and urine samples chrono-amperometry; dpv; cv; lsv 0.1-110.0 μm 0.04 μm s = 0.1342 μa μm-1 2023 / [31] agnps@cpe mtz in milk and tap water 11000 μm 0.206 μm 2022 / [33] [10%fg/5%mw] cpe ofloxacin in pharmaceutical tablets human urine samples sw-adas 0.60 to 15.0 nm 0.18 nm 2019 / [41] mip/cp ecl sensor azithromycin urine and serum samples ecl 0.10 400 nm 0.023 nm 2018 / [45] cpea/tio2/uv aa in pharmaceutical tablets cv 150 850 nm 0.732 μm 2023 / [48] cuo mfs/nafion/gc glucose in real serum sample amperometric 10 120 μm 6.48 μm s=3.1 μa μm-1 cm-2 2018 / [50] dmip/cpe mtz in human serum, urine and drug dpv 0.4 200 μm 91 nm 2016 / [52] fe3o4/zif-67 /ilcpe smz in urine and water dpv 0.01 520.0 m 5.0 nm 2021 / [53] poly-ebt: poly-eriochrome black t; ce: cerium; btc: benzene tricarboxylic acid; mof : metal organic framework; il : ionic liquids; agnps: silver nanoparticles; fg: flake graphite; mw: multi-wall carbon nanotubes; mip: molecularly imprinted polymer; ecl: electrochemiluminescence; cuo mfs: cuo micro flowers; dmip: duplex molecularly imprinted polymer; zif-67: zeolitic imidazolate framework ; mdh: methdilazine hydrochloride; ktc: ketoconazole ; aa: ascorbic acid; mtz: metronidazole; smz: sulfamethoxazole; s: sensitivity admet & dmpk 11(2) (2023) 151-173 electrochemical sensors and biosensors for monitoring drugs and metabolites doi: https://doi.org/10.5599/admet.1709 155 the year of publication is indicated in each table in order to highlight the most recent publications and make them directly accessible to the reader. electrodes based on modified glassy carbon for detecting various drugs and metabolites glassy carbon is a type of carbon with an amorphous structure like glass and ceramics. glassy carbon is an interesting material for electrochemical sensors because of its chemical resistance, low density and low electrical resistance. glassy carbon electrodes (gce) have also played an important function in electrochemical sensors and biosensors for drug analysis. the electron transfer rate at the gce surface can be increased by various modifiers for the detection of medicines in diverse media. the gce/zno@ndcs/gox is an enzymatic biosensor used to analysis glucose in serum (cf. table 2) [54]. the biosensor had a reproducible sensitivity of 231.7 μa mol-1 l cm-2 with a lod of 6.3 μmol l-1 at 0.57 v. the sensor selectivity is demonstrated via the addition of some interfering potentials such as ascorbic acid, dopamine, fructose, uric acid and mannose (man) with glucose, resulting in only a small increase in current for each species. due to the easy transfer of electrons through this biosensor, the response time is very short (<3 s). the biosensor is well suited for glucose analysis with recovery between 99.73-100.14 % and retains about 95.3 % of the initial response during 50 days of storage in human blood. acyclovir (acv) (9-(2-hydroxyethoxymethyl)guanine) is a drug largely employed for the therapy of viral skin infections and neuritis [55-57]. overdoses lead to adverse effects in patients [58,59]. the femoo4 compound is able to catalytically fix nitrogen [60] due to its interesting electrocatalytic property [61]. it also has interesting redox kinetics, making it a desirable material for electrode formation. recently, y. wei et al. [62] prepared a sensor using graphene oxide composites loaded (ultrasonically) with ferrous molybdate (femoo4) for the analysis of acv in pharmaceuticals by lsv. the oxidation peak of acv is observed at 1.1 v. femoo4-go/gce and has two linearity intervals (0.1-10 and 10-100 μmol l-1 ) with a lod of 20 nmol l-1. the reaction of acv on the developed electrode is diffusion controlled with 2eand 2h+ transferred. the active surface area of femoo4-go/gce is 1.59 times larger than that of femoo4/gce, which clearly shows that the femoo4-go improves the performance of the sensor. y. huang et al. [63] fabricated a biosensor to study the human umami taste receptor (ht1r1) and the umami substances, such as monosodium glutamate (msg), using a multilayer material to modify the glassy carbon electrode (figure 2). during the fabrication of this electrode, a human umami taste receptor (ht1r1) was attached to the layers formed by the aunps [64]. horseradish peroxidase (hrp) is used for direct electron transfer to the multilayer material formed [65]. the researchers suggest that ht1r1 is likely a receptor used by the body to sense nitrogen, opening up a new way of studying nutrient and drug adsorption. in 2023, the same researchers [66] created another biosensor by attaching colon cancer and adjacent tissues to gce to visualise the kinetics of responding to c and n nutrient receptors such as glucose and sodium lactate. in order to do this, they mixed solutions of starch gum with an aldehyde base and sodium alginate, which were spread over two microporous polycarbonate membranes into which the colon tissues were placed to build a layered assembly aligned to the gce (figure 3). researchers found that the cells had different sensitivities to lactate, suggesting the possibility of using this nutrient to treat colon cancer. colon cancer tissue is insensitive to lactate, whereas adjacent tissue is sensitive. coco2o4 nanorods embedded in hexagonal boron nitride are used to modify the gce for ronidazole determination [67]. the electrode synthesis presents a nano-lod of 3 nmol l-1 between 0.01 and 1345 μmol l-1and a higher sensitivity of 5.845 µa μmol-1 l cm-2 using dpv. this sensitivity is explained by the enhanced adsorption and transport of mass ions via the formation of aggregates of hexagonal boron nitride (h-bn) and https://doi.org/10.5599/admet.1709 c. laghlimi et al. admet & dmpk 11(2) (2023) 151-173 156 spinel cobalt oxide nanorods (coco2o4 nrs). coco2o4 nrs are used for their low cost, good stability and other properties. the h-bn, formed from boron and nitrogen, has a structure similar to graphene [68], which makes it an electroactive material [69,70]. figure 2. schematic of the multilayer modification of the gce to study the electrochemical response of the human umami taste receptor ht1r1 towards umami substances such as sodium glutamate, disodium inosinate and disodium guanylate. carried out at the base of the reference [63] with an order license id of 1346500-1 figure 3. schematic of the multilayer modification of gce to study the detection kinetics of c and n nutrient receptors. based on the reference [66] methotrexate (mtx) is a drug that inhibits the growth of tumour cells. it is applied for the treatment of certain types of cancer, notably breast cancer [71] and pulmonary cancer [72]. go-nafion-gce sensor is constructed by dispersing graphite oxide in a solution containing nafion-ethanol and then using it to modify the gce [73]. modified electrode stabilisation is achieved by using a cv of 0.1 v/s and in the potential between 0.5 and 1.2 v. go-nafion-gce is able to detect methotrexate in mtx injection and urine with a lod admet & dmpk 11(2) (2023) 151-173 electrochemical sensors and biosensors for monitoring drugs and metabolites doi: https://doi.org/10.5599/admet.1709 157 of 9 nmol l-1 between 0.4 and 20 μmol l-1 , that is close to the lod of detection of the same drug by 3dpgcnt/gce [74]. the electrolyte solution used for the analysis of mxt is perchloric acid (0.03 mol l-1). a mixed adsorption-diffusion phenomenon controls the reaction process of mxt on the electrode surface. dopamine (da) and paracetamol (pa) detection in synthetic urine was performed using a gceergo/polycotapc electrode [75]. first, gce was modified by electrochemically reduced graphene oxide (ergo), which was simultaneously reduced and deposited from graphene oxide (go). finally, gceergo/polycotapc was formed by electropolymerisation of cobalt (ii) tetra-amino phthalocyanine. the developed electrode sensitivity was 1.32 µa mol-1 l cm-2 for pa and 8.39 µa mol-1 l cm-2 for da. lod is 0.10 µmol l-1 for pa and 0.095 µmol l-1 for da using dpv. the modification of gce by qds-p6lc-pedot:pss allowed the amoxicillin (amx) analysis in synthetic urine, whole milk and pharmaceuticals [76]. amx oxidation peak is 0.88v and the reaction process on qds-p6lcpedot:pss/gce is irreversible controlled by diffusion process with an equal number of proton and electron exchange. the researchers found that the alkaline medium facilitates the deprotonation of amx during the oxidation reaction. furthermore, the detection limit is found to be 50 nmol l-1 in the range 0.90-69 μmol l-1 . to form graphene (gr)-zno/gce, x. yue et al. [80] first polished the surface of gce with a suspension of al2o3 powder on a polishing cloth. then, 8 µl of gr-zno nanocomposite suspension is used to modify the glassy carbon electrode. gr-zno/gce was used for the simultaneous detection of sulfamethoxazole (smz) and trimethoprim (tmp) in urine and serum. smz and tmp oxidation peaks were successfully located at the ep (smz) = 0.85 v and ep (tmp) = 1.06 v by the dpv method, with a slight shift in peak potentials compared to those observed on the gce. the researchers explained this by the ability of the gr to enhance zno nanorods conductivity and the reciprocal ability of the zno nanorods to avoid aggregation of the gr by reducing the van der waals force. smz and tmp oxidation reactions are irreversible. they are controlled by adsorption. smz lod is 0.4 μmol l-1 between 1-220 and 0.3 μmol l-1of tmp in the range of 1-180 μmol l-1 . the same drugs have been detected by gce modified with go and ag nanoparticles (gc/rgo-agnp) [81] using dpv with lod (smz) = 0.6 μmol l-1 and lod (tmp) = 0.4 μmol l-1 between 1.0 and 10.0 mmol l-1. in addition, smz and tmp exhibit irreversible oxidation peaks with ∆etmp-smz= 1.14 0.92 = 0.22v. a further study has shown that the nickel ferrite/rgo (nife2o4/rgo) film is an excellent modifier of the gce. clenbuterol was analyzed in pig urine samples using the electrode produced [82]. the lod is 0.17 µmol l-1 between 0.99 and 18.03 µmol l-1 using dpv. due to the excellent selectivity of chitosan [83] and the high selectivity and ease of preparation of the mip method [84], y. wu et al. [85] fabricated a sensor called mip-mwcnts/gce for the detection of tryptophan (trp), a possible cause of schizophrenia [86], by depositing a printed chitosan film on the mwcnt premodified gce surface. this pre-modification with mwcnts aims to enhance the response of molecularly imprinted polymer electrodes [87]. mwcnts are a type of carbon nanotube with highly active sites due to the presence of more concentric tubes [88], which gives them high adhesive activity and good conductivity. y. wu et al. [85] demonstrated hydrogen bonding between chitosan and trp using fourier transform infrared (ftir) spectroscopy. the extraction reagent (ethanol) used during the extraction process causes the chitosan film to deflate and etch, forming a porous structure capable of efficiently binding trp molecules to these imprinted sites (figure 4). the oxidation process of trp involves the exchange of an identical number of e and h+. analysis of trp in human serum by mip-mwcnts/gce showed a very encouraging recovery between 96.5 and 102.5 %. in the presence of several interferents, the trp peak is very intense compared to the other substances. the mip-mwcnts/gce sensor has a low lod (1 nmol l-1) compared to the gce sensor modified with aunp and mwcnts [89]. https://doi.org/10.5599/admet.1709 c. laghlimi et al. admet & dmpk 11(2) (2023) 151-173 158 figure 4. schematic of mip-mwcnt/gce preparation and electrochemical detection of tryptophan. based on the reference [85] daunorubicin is an anticancer drug [90] that requires dose control [91] to avoid adverse effects, including cardiac arrest. h. karimi-maleh et al. [92] attempted to develop a glassy carbon electrode modified with a nanocomposite of pt/swcnts and ds-dna (biorecognition) to eventually form ds-dna/pt/swcnts/gce. dvp shows a positive displacement of the daunorubicin oxidation peak (from 847 to 882 mv), confirming the mutual intercalation reaction of ds-dna (guanine base) and daunorubicin. furthermore, the equilibrium constant of the association is 5.044×103 mol-1 l with a nanometre lod (1.0 nmol l-1) over a concentration between 4.0 nmol l-1 and 250.0 μmol l-1. using the fe3o4/mwcnt/gce electrode, t. bhengo et al. [93] obtained oxidation peaks of sulfamethoxazole and trimethoprim at 910 and 1120 mv, respectively, with detection limits of 11.0 nmol l-1 for smz and 21.0 nmol l-1 for tmp using the dpv method. similarly, rajasekhar chokkareddy et al. [94] deposited il-f-znonps@mwcntson the clean surface of gce by dispersing them in dimethylformamide (cf. table 2).to prepare a solution containing smz, the researchers crushed two sandoz co-trimoxazole (pharmaceuticals) tablets and dissolved part of the powder in methanol. after shaking, the solution was refined and diluted with pbs (ph 6.0). the resulting solution was then added to the pharmaceutical urine samples. the cyclic voltammetry oxidation peak of the smx was found to be at 620 mv. the reaction process on il-fznonps@mwcnts@gce is diffusion controlled and the detection limit is 0.1 ng/ml in the range of 0.1-10 ng ml-1. in addition, ag-mwcnt/mtoac/gce [95] was successfully used to detect smx in co-trimoxazole (mg per tablet), cotrim paediatric and human urine by dpv. the samples were diluted with phosphate solution to reduce the matrix effect and then introduced into the electrochemical cell. the irreversible oxidation of the smx is observed at +0.85 v and involves the exchange of an equal number of eand h+ and is controlled by adsorption. on the other hand, the simultaneous analysis of ascorbic acid (aa), dopamine and uric acid in human urine was investigated using a gce-based sensor modified with 2d titanium carbide nanoplatelets (mxene) [96]. all three analytes showed remarkably distinct oxidation peaks. aa was observed at 0.001 v, while dopamine and uric acid were at 210 mv and 330 mv, respectively. it should be noted that 2d titanium carbide (mxene) nanoplatelets are mentioned in the literature as having a high conductivity of 9880 s cm-1 [102], higher than admet & dmpk 11(2) (2023) 151-173 electrochemical sensors and biosensors for monitoring drugs and metabolites doi: https://doi.org/10.5599/admet.1709 159 that of graphene. in particular, mxene groups, such as fluor, oxygen and hydroxyl, increase its hydrophilicity and allow rapid access to the analyte. it is therefore used as an ideal material for sensors [103]. table 2 shows the different types of modified gces as well as the drugs detected and their detection medium, the limit of detection (lod), concentration range and other characteristics. table 2. evaluation of electrochemical properties of modified gce electrode analyte and matrix method linear dynamic range lod year / ref gce/zno@ndcs/gox glucose in human blood serum cv; dpv;cam 0.2 and 12 mm 6.3 μm s = 231.7 μa mm-1 cm-2 2018 / [54] femoo4-go/gce acyclovir in drug samples lsv 0.1-10 μm 10-100 μm 20 nm 2022 / [62] coco2o4 nrs/h-bn/gce ronidazole water samples dpv 0.01-1345 µm 3 nm s = 5.845 µa µm-1 cm-2 2023 / [67] go-nafion-gce mtx in methotrexa e injection an u ine samples cv 0.4 -20 µm 9 nm 2019 / [73] gce-ergo/polycotapc da and paracetamol in synthetic urine dpv 0.10 µm for pa 0.095 µm for da spa=1.32 µa µm-1 cm-2 sda= 8.39 µa µm-1 cm-2 2022 / [75] qds-p6lc-pedot:pss/gce ax in synthetic urine, whole milk and pharmaceuticals swv 0.90 to 69 µm 50 nm 2020 / [76] tio2/cmk3/aunps/nafion/gce ax in pharmaceutical product, mineral and environmental water cv ldr1= 0.5-2.5 μm ldr2 = 2.5-133 μm 0.3 µm s1 = 5071 μa mm−1 cm−2 s2 = 2971 μa mm−1 cm−2 2018 / [77] cb/dph/gce ax inbiological urine, lake and tap water swv 2.0-18.8μm 0.12 µm 2018 / [78] nim inbiological urine, lake and tap water swv 0.30-5.0 μm 0.016 µm aunps/pdnps/ergo/gce ax inhuman urine samples swv 30.0-350.0μm 9.0 µm s = 0.0376 μa μm-1 2017 / [79] lmf human urine samples swv 4-500 μm 81 nm s = 0.0759 μa μm-1 gr-zno/gce smz and tmp in urine and human serum dpv lrdsmz= 1-220 μm lrdtmp = 1-180 μm lodsmz = 0.4 μm lrdtmp = 0.3 μm 2020 / [80] gc/rgo-agnp smz and tmp in wastewaters samples dpv 1.0-10.0 mm lodsmz = 0.6 μm lrdtmp = 0.4 μm 2017 / [81] mip-mwcnts/gce trp in the human serum samples. cv, sdlsvs 2.0 nm-0.2 μm 0.2 -10 μm 10-100 μm 1.0 nm 2020 / [85] ds-dna/pt/swcnts/gce drn in daunorubicin injection dvp 4.0 nm to 250.0 μm 1.0 nm 2021 / [92] il-f-znonps@mwcnts@gce smz pharmaceutical urine samples cv 0.1-10 ng ml-1 0.1 ng ml-1 2022 / [94] https://doi.org/10.5599/admet.1709 c. laghlimi et al. admet & dmpk 11(2) (2023) 151-173 160 electrode analyte and matrix method linear dynamic range lod year / ref ag-mwcnt/mtoac/gce smz in pharmaceutical formulationsand human urine dpv 0.05-70 μm 0.01 μm 2018 / [95] ti-c-tx/gce aa in human urine samples cv, dpv 100-1000μm 4.6 μm 2021 / [96] da in human urine samples cv, dpv 0.5-50 μm 0.06 μm ua in human urine samples cv, dpv 0.5-4 μm 100-1500 μm 0.075 μm pvp-gr/gce aa in human urine samples lsv 4-1000 μm 0.80μm 2020 / [97] da in human urine samples lsv 0.02-100 μm 0.002 μm ua in human urine samples lsv 0.04-100 μm 0.02 μm rgo-zno/gce aain real plasma and urine samples dpv 50-2350μm 3.71μm 2016 / [98] ua in real plasma and urine samples dpv 1-70 μm 0.33 μm da in real plasma and urine samples dpv 3-330 μm 1.08 μm 3d-hpg/pth/gce mtz in pharmaceutical and real wate samples cv; dpv 0.05-70 μm 70−500 μm 1 nm 2018 / [99] polydopamine/mwcntsecoo h nanocomposites/gce mtz in pharmaceutical and biological samples dpv 5-5000 μm 0.25 μm 2018/ [100] ag-mwcnt/mtoac/gce smz in pharmaceutical formulations and human urine dpv 0.05-70 μm 0.01 μm 2018 / [101] ndcs: nitrogen-doped carbon sheets; gox: glucose oxidase; femoo4-go: ferrous molybdate-graphene oxide; coco2o4 nrs: spinel cobalt oxide nanorods; h-bn: hexagonal boron nitride; ergo: electrochemically reduced graphene oxide; polycotapc: polymer of cobalt (ii) tetra-amino phthalocyanine; qds: cadmium telluride quantum dots; p6lc: printex 6l carbon; pedot: poly(3,4-ethylenedioxythiophene; pss: polystyrene sulphonate film; cmk-3: mesoporous carbon cmk-3, aunps: gold nanoparticles; dph: dihexadecyl hydrogen phosphate; cb: carbon black; pdnps: palladium nanoparticles; gr: graphene; mip: molecularly imprinted polymer; ds-dna: guanine-based; il-f-znonps: ionic liquid functionalised zinc oxide nanoparticles; mtoac: methyltrioctyl ammonium chloride; ti-c-tx: titanium carbide (mxene) nanosheets; pvp: polyvinylpyrrolidone; rgo: reduced graphene; 3d: three-dimensional; hpg: graphene-like carbon architecture; pth: polythionine; mwcntsecooh: carboxylic mwcnt; rnz: ronidazole; mtx: methotrexate; da: dopamine; ax: amoxicillin, nim: nimesulide; lmf: lomefloxacin; smz: sulfamethoxazole; tmp: trimethoprim; trp: tryptophan; dnr: daunorubicin, aa: acide ascoorbic; da: dopamine; ua: uric acid; mtz: metronidazole; s: sensitivity; cam: chronoamperometry. electrodes based on modified graphene oxide for detecting various drugs and metabolites graphene oxide reduced (rgo) is obtained by thermal, electrochemical or chemical treatment of graphene oxide using substances such as nabh4 or aluminium powder. this reduction removes oxidised functional groups (which increases the sensitivity of the go and makes it insulating) and creates a defect structure (active sites) characteristic of rgo. these properties have enabled rgo to have high electrochemical activity compared to graphene and graphene oxide. f. zhou et al. [104] successfully fabricated a nafion/gox/au-zno/rgo/ito electrode. the researchers found that the rate and efficiency of electron transfer were augmented by the deposition of aunps and uv irradiation, allowing a significant increase in glucose sensitivity (cf. table 3). lod of this sensor is 0.2 μmol l-1 https://www.sciencedirect.com/topics/materials-science/titanium-carbide https://www.sciencedirect.com/topics/materials-science/nanosheet https://www.sciencedirect.com/topics/chemistry/chronoamperometry admet & dmpk 11(2) (2023) 151-173 electrochemical sensors and biosensors for monitoring drugs and metabolites doi: https://doi.org/10.5599/admet.1709 161 by amperometric method between 0 and 9.5 mmol l-1 with a higher michaelis-menten constant (15.54 mmol l-1) than that of non-uv irradiated nafion/gox/au-zno/rgo/ito. the sensors have exceptional detection accuracy for blood glucose measurement. regorafenib (reg) or regonix is an anticancer drug used in several types of cancer [105,106]. monitoring the concentration of reg in serum or blood seems to be of great interest as doses above certain limits can cause adverse effects on vital organs of the body [107]. reg is detected in real samples by pd-ru/rgo using the dpv technique [108] (cf. table 3). the acidic graphene neutrality is obtained from pomegranate peel extract (ppe). in fact, the redox current peak is linear with v (v = potential sweep rate), indicating that the reg reaction occurring at pd-ru/rgo is diffusion controlled, with a constant heterogeneous charge transfer rate equal to k0 = 2.29 s-1, calculated based on laviron's equation (cf. eq. (1), (2) and (3)) [109]. furthermore, the reg oxidation involves the exchange of 2eand 2h+. ( ) a 0 2.3 log 1 rt v ep e nf = + − (1) c 0 2.3 logrt v ep e nf = + (2) ( ) ( ) ( ) s 1 log log 1 1 log log 2.3 nf eprt k nfv rt       −   = − + − − −    (3) table 3 shows the different types of modified rgos as well as the drugs detected and their detection medium, limit of detection (lod), concentration range and other characteristics. table 3. evaluation of electrochemical properties of modified rgo electrode analyte and matrix method linear dynamic range lod other characteristics year/ref. nafion/gox/auzno/rgo/ito glucose in blood amperometric 0-9.5 mm 0.2 μm s = 10.93 μa mm-1 cm-2 2020 / [104] pd-ru/rgo regorafenib in human blood and plasma pharmaceutical formulation dpv 0.5-300 nm 1.6 nm ks = 2.29 s-1 2022 / [108] rgo/tmu-22 mof levodopa in human urine and tablet samples cv; swv 0.1-85 μm 0.1-85 µm 25 nm s = 0.58 µa µm−1 ks = 7.7 s−1 2020 / [110] aunps : gold nanoparticles ; gox: glucose oxidase; rgo: reduced graphene oxide;ito: indium tin oxide ; tmu-22 mof: metal-organic framework; pd-ru: palladium/ruthenium nanoparticles; s: sensitivity electrodes based on modified screen-printed carbon electrodes for detecting various drugs and metabolites the screen-printed electrode (spe) consists of a substrate (pvc, ceramic) onto which miniature electrodes can be printed using ink based on conductive materials. various substances are employed to increase the sensitivity of the screen-printed electrode, including fullerenes (c60) and iron(iii) nanoparticles (fe3o4np). the antibiotic drug furaltadone (flt) was detected by a screen-printed carbon electrode (spce) decorated with cu/ni/tio2/mwcnts nanocomposites [111]. flt reduction peak on cu/ni/tio2/mwcnts nanocomposites was located at -470 mv with a significant current density of -39.17 μa. the reduction of the r-no2 group to rnhoh was performed by electron transfer from homo to lumo in the fabricated electrode and the lod is equal to 0.0949 μmol l-1 with a high sensitivity of 1.9288 μa μmol-1 l cm-2 in water. we also find that erythronmycin (ery) is detected by sodium dodecyl sulfate-modified spces (spces/sds) by flow injection analysis with amperometric detection in pbs of ph 8.5 [112]. peak detection of ery is 0.7 v with an lod of 0.19 μmol l-1. a spe coated with c60, rgo and nafion (nf) is used to analysis antibiotic metronidazole (mtz) in real samples (cf. table 4) [113]. reduction of go with nabh4 results in the disappearance of oxygen groups. fullerenes are used because of their ability to operate at low potentials, which is advantageous for a sensor https://doi.org/10.5599/admet.1709 c. laghlimi et al. admet & dmpk 11(2) (2023) 151-173 162 to avoid interference effects [114,115]. to stabilise the response of the c60-rgo-nf/spe, the researchers applied two cyclic voltammetry cycles between 0 and -1500 mv at v = 20 mv s-1 in a 1 mol l-1 electrolyte solution of koh and then the electrode was placed in pbs (neutre) to apply v = 50 mv s-1 (scan rate) between 550 and -50 mv. the mtz reduction peak observed at -0.9 v, representing the exchange of 4eand 4h+. the intensity of the mtz reduction peak on c60-rgo-nf/spe is five times higher than that observed on spe. this sensitivity is explained by the porosity of c60, rgo and nf, which increases the surface area of c60-rgonf/spe. the lod by swv is 0.21 μmol l-1in the range 0.25-34.0 μmol l-1. gp-cac/pvc is a screen-printed electrode (spe) [116], estimated to cost 0.016 $ per unit compared to 7 $ for a commercial electrode, used to determine levodopa (l-dopa) in two commercial drugs, parkidopa® and ekson®. these drugs are prescribed to treat parkinson's disease by reducing the severity of symptoms. to maintain the conductive material of the electrode, researchers have used polymeric materials such as cellulose [117], polyvinyl [118], epoxy glue, nail polish [119] and cellulose acetate (cac) [120]. in this case, the conductive ink for the electrode is obtained by mixing cellulose acetate and graphite powder in attendance of acetone/cyclohexanone (apolar solvents). the adhesive stencil is used to draw the relief of three electrodes, which are then attached to the pvc. conductive ink is then applied to the stencil to print the electrodes on the pvc. the stencil is immediately removed to allow the ink to run off. the gp-cac/pvc has an electroactive surface area of 0.48 cm2 with a sensitivity of 0.101 μa μmol-1 l owing to the significant porosity of the cellulose acetate. the detection limit by swv is 0.06 μmol l-1 between 8.00 and 100 μmol l-1. the researchers found that at ph 5 (ph ≥ 5.0), l-dopa exhibits an irreversible reaction with an oxidation peak located at 0.25 v caused by exchanging the 2h+ and 2e-. furthermore, at ph 2 (ph ≥ 4.0), the behaviour of l-dopa changes towards a reversible system with ep(oxy) = 0.27v and ep(red) = -0.01v. on the other hand, spe modified with the cuo/co3o4 nanocomposite integrated mwcnts [121] has a very low lod of 0.223 pmol l-1 between 10-12 and 10-2 mol l-1 range for urea determination by electrochemical impedance spectroscopy (eis) (fig. 5). this performance is due to the combination of cuo and co3o4, which have higher electronic conductivity than co3o4 or cuo alone. figure 5. schematic of cuo/co3o4@mwcnts preparation and electrochemical detection of urea. based on the reference [121] admet & dmpk 11(2) (2023) 151-173 electrochemical sensors and biosensors for monitoring drugs and metabolites doi: https://doi.org/10.5599/admet.1709 163 table 4 shows the different types of modified spes as well as the drugs detected and their detection medium, limit of detection (lod), concentration range and other characteristics. table 4. evaluation of electrochemical properties of modified spe electrode analyte and matrix method linear dynamic range lod / other characteristics year / ref. cu/ni/tio2/mwcnts/spce flt in water cv; dpv 10-150 μm 0.0949 μm s = 1.9288 μa μm-1 cm-2 2022 / [111] spces/sds ery in water amperometric 1-15 mg l-1 0.19 μm 2019 / [112] c60-rgo-nf/spe mtz in serum and urine samples swv 0.25-34.0 μm 0.21 μm 2021 / [113] gp-cac/pvc levodopa in two commercial drugs swv 8-100 μm 0.06 μm 2021 / [116] cuo/co3o4@mwcnts/spe urea eis 10-12-10-2 m 0.223 pm 2023 / [121] nano-au/mwnts-zno/spe mtx in whole blood samples and pharmaceutical swv 0.02-1.00 μm 10 nm 2014 / [122] epi in whole blood samples and pharmaceutical swv 0.005-0.2 μm 2.5 nm mwcnt-aonp/spce aa in fresh oranges swv 160-640 nm 140 nm 2022 / [123] spce/cv-fe3o4np aa swv 10-100 μm 15.7μm 2021 / [124] sds: sodium dodecyl sulfate ; nf: nafion; c60: fullerene ; cac: cellulose acetate; gp: graphite powder; pvc: polyvinyl chloride; aonp: antimony oxide nanoparticle; cv: callistemon viminalis; fe3o4np: iron(iii) nanoparticle; flt: furaltadone; ery:erythromycin; mtz:metronidazole ; mtx : methotrexate; epi : epirubicin; aa: ascorbic acid. conclusions and perspectives this review discusses the different types of recently published electrochemical sensors for the detection of drugs and metabolites in pharmaceutical and biological samples. cpe has a very large electroactive surface area, which can be renewed for a variety of applications. in addition, it has low ohmic resistance and high stability, reproducibility and lifetime, giving it the ability to analyse various drugs, revealing the oxidation and reduction processes involved due to its wide potential range. the sensitivity and speed of analysis of cpe can be significantly improved by modifying them with conductive materials. various modifiers were used, including organic modifiers like poly-eriochrome black t (poly-ebt), benzene tricarboxylic acid (btc), metalorganic frameworks (mofs) or inorganic modifiers like silver nanoparticles (agnps), cuo microfragments (cuo mfs) and tio2. on the other hand, the improvement of cpe can be achieved by using a carbon nanomaterial modifier such as flake graphite (fg), multi-wall carbon nanotubes (mwcnts) and carbon quantum dots (cqds). the modified carbon past electrode has been used for the determination of various drugs such as methdilazine hydrochloride (mdh), ketoconazole (ktc), metronidazole (mtz), sulfamethoxazole (smz) in diverse samples such as human urine, pharmaceutical formulations, plasma and serum. modification with silver nanoparticles is also encouraged due to their biocompatibility, sensitivity, stability and ability to increase peak intensity due to their high conductivity. similarly, metal oxide nanomaterials such as cuo mfs have an essential contribution to the performance of cpe owing to their wide surface area. l. hu et al. fabricated a selective electrochemiluminescence (ecl) sensor by modifying an cpe with a mip. the sensor showed a very low lod of 0.023 nmol l-1 to analyse azithromycin (azi) in real samples.the very low lod of 0.18 nmol l-1 is also found by a cpe modified with lamellar graphite and mwcnt for the analysis of ofloxacin in pharmaceutical tablets and human urine samples. similarly, cpe modified with fe3o4/zif67/ilcpe nanocomposite was applied for the analysis of smz in urine and water with a lod of 5.0 nmol l-1. https://doi.org/10.5599/admet.1709 c. laghlimi et al. admet & dmpk 11(2) (2023) 151-173 164 the transfer rate of electrons at the surface of glassy carbon electrodes (gce) can be increased by various modifiers such as nitrogen-doped carbon sheets, glucose oxidase, cobalt (ii) polymer tetraamino phthalocyanine (polycotapc), methyltrioctyl ammonium chloride (mtoac) and guanine (ds-dna). in addition, inorganic modifiers such as iron molybdate (femoo4), spinel cobalt oxide nanorods (coco2o4 nrs), hexagonal boron nitride (h-bn), cobalt (ii) tetra-aminophthalocyanine polymer (polycotapc), cadmium telluride quantum dots (qds) and titanium carbide (mxene) nanosheets (ti-c-tx) have been used for the same purpose. similarly, a variety of carbon types have been reported in the literature to improve gce sensitivity, including go, electrochemically reduced graphene oxide (ergo), printex 6l carbon (p6lc), cmk-3 mesoporous carbon (cmk-3), carbon black (cb), graphene (gr), reduced graphene (rgo), graphene-like carbon architecture (hpg) and carboxylic mwcnt (mwcntsecooh). the modified gce was used to determine the drugs in the different matrices. the drugs analysed were: ronidazole, methotrexate, dopamine, amoxicillin, nimesulide, lomefloxacin, sulphamethoxazole, trimethoprim, tryptophan and daunorubicin. platinum nanoparticles have high electrocatalytic properties. similarly, dna nanostructures have high selectivity and affinity for the analyte, which can be used to modify the gce used for electrochemical drug analysis. the conductivity of the glassy carbon electrodes can be increased by using polymers such as polyaniline (pani). the combination of various modifiers with high conductivity electrochemical materials such as mips, mwcnts, swcnts, ds-dna and pt has further increased the sensitivity of the sensors. mip-mwcnts/gce and ds-dna/pt/swcnts/gce showed a very low lod of 1 nmol l-1 for drug determination. in addition, the modification with 3d-hpg/pth shows a similar nanometric detection limit. in 2023, a group of researchers successfully fabricated a glassy carbon electrode modified with coco2o4 nrs/h-bn used to analysis the mtr in pharmaceuticals, with a sensitivity of 5.845 µa µmol-1 l cm-2 and a lod of 3 nmol l-1. aunps have been employed as modifiers for rgo due to their considerable catalytic activity, biocompatibility and high electrochemical potential, which allows the redox process generated at the electrode surface to be detected. other materials are employed to improve the sensitivity of rgos, including glucose oxidase, indium tin oxide (ito), metal-organic framework (mof tmu-22) and palladium/ruthenium nanoparticles (pd-ru). the modification of rgo with pd-ru resulted in a very low lod of 1.6 nmol l-1 for the determination of regorafenib in various samples like human blood, plasma and pharmaceutical formulations. a variety of conductive compounds are available to increase the sensitivity of the screen-printed electrode, like sodium dodecyl sulphate (sds), nafion (nf), fullerene (c60), cellulose acetate, graphite powder, antimony oxide nanoparticles (aonp) and iron (iii) nanoparticles (fe3o4np). in addition, multi-wall carbon nanotubes (mwnts) have a high active surface area and conductivity. the specific structure of these nanomaterials gives them the potential to be an interesting candidate for drug detection electrochemical sensors. the combination of metal oxide and carbon nanotubes significantly increases the sensitivity of spe. recently, h. s. magar et al. fabricated an spe using a combination of cuo/co3o4 and mwcnts. the sensor was used for urea analysis with a very low lod of 0.223 pmol l-1. not all electrode types mentioned in this paper reported this value. in addition, a lod of 2.5 nmol l-1 was found for nano-au/mwnts-zno/spe used to analysis epi in whole blood and pharmaceutical samples. the use of nanomaterials and metal oxides seems promising to improve the sensitivity of the electrodes, but there are many problems, such as difficulties in fabrication, handling and characterisation, which encourage investment in finding simpler fabrication methods or extraction from plants and animal secretions. acknowledgements: we are grateful to mr. haroun laghlimi (my dear son), mrs. fatna nâna and mrs. ikram ait lebbadfor their encouragement and support. admet & dmpk 11(2) (2023) 151-173 electrochemical sensors and biosensors for monitoring drugs and metabolites doi: https://doi.org/10.5599/admet.1709 165 conflict of interest: the authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. list of abbreviations c60 buckminsterfullerenes ca chronoamperometry cpe carbon paste electrodes cpea carbon clay paste electrode cv cyclic voltammetry dpv differential pulse voltammetry ebt eriochrome black t ecl electrochemiluminescence eis electrochemical impedance spectroscopy ergo electrochemically reduced graphene oxide fg flake graphite ftir fourier transform infrared gce glassy carbon electrode go graphene oxide gp graphite powder gr graphene homo highest occupied molecular orbital hplc-dad high performance liquid chromatography-diode array detector lod limit of detection lsv linear scan voltammetry lumo lowest unoccupied molecular orbital mip molecularly imprinted polymer mwcnts multi-wall carbon nanotubes mxene m: transition metal, x: carbon and/or nitrogen, ene: similarity to graphene nf nafion pbs phosphate-buffered saline pvc polyvinyl chloride rgo reduced graphene oxide sdlsv second-order derivative linear sweep voltammetry spc screen-printed carbon spce screen-printed carbon electrode sw-adas square wave adsorptive anodic stripping voltammetry swcnt single-walled carbon nanotubes swv square wave voltammetry tlc-densitometric thin layer chromatography densitometric references [1] s.b. somvanshi, p.b. kharat, k.m. jadhav, n.d. thorat, h. townley, nanomaterials and pharmacokinetics, in nano-pharmacokinetics and theranostics, academic press, 2021, p. 1-14. https://doi.org/ 10.1016/b978-0-323-85050-6.00007-4 [2] a.a. khorshed, m. khairy, s.a. elsafty, c.e. banks. disposable screen-printed electrodes modified with uniform iron oxide nanocubes for the simple electrochemical determination of meclizine, an antihistamine drug. analytical methods 11 (2019) 282-287. https://doi.org/10.1039/c8ay02405g [3] f. ibrahim, a. el-yazbi, m. wagih, m. barary. sensitive inexpensive spectrophotometric and spectrofluorimetric analysis of ezogabine, levetiracetam and topiramate in tablet formulations using hantzsch condensation reaction. spectrochimica acta a 184 (2017) 47-60. https://doi.org/10.1016/j.saa.20 17.04.078 [4] n. tamilselvi, a. rajasekaran. stability-indicating rp-hplc method for the determination of ezogabine and identification of its degradation products. journal of pharmaceutical sciences and research 8 (2016) 19–23. [5] n.m. habib, m.m. abdelrahman, n.s. abdelwhab, n.w. ali. validated chromatographic methods for the analysis of two binary mixtures containing pyridoxine hydrochloride. journal of association of official agricultural chemists international 100 (2017) 414-421. https://doi.org/10.5740/jaoacint.16-0213 https://doi.org/10.5599/admet.1709 https://doi.org/10.1016/b978-0-323-85050-6.00007-4 https://doi.org/10.1016/b978-0-323-85050-6.00007-4 https://doi.org/10.1039/c8ay02405g https://doi.org/10.1016/j.saa.2017.04.078 https://doi.org/10.1016/j.saa.2017.04.078 https://doi.org/10.5740/jaoacint.16-0213 c. laghlimi et al. admet & dmpk 11(2) (2023) 151-173 166 [6] r. peraman, m. manikala, v.k. kondreddy, p.r. yiragamreddy. a stability-indicating rp-hplc method for the quantitative analysis of meclizine hydrochloride in tablet dosage form. journal of chromatographic science 53 (2015) 793-799. https://doi.org/10.1093/chromsci/bmu127 [7] m.a. mohamed, a.s. fayed, m.a. hegazy, n.n. salama, e.e. abbas. fully optimized new sensitive electrochemical sensing platform for the selective determination of antiepileptic drug ezogabine. microchemical journal 144 (2019) 130-138. https://doi.org/10.1016/j.microc.2018.08.062 [8] h. karimi-maleh, r. darabi, f. karimi, c. karaman, s.a. shahidi, n. zare, m. baghayeri, l. fu, s. rostamnia, j. rouhi, s. rajendran. state-of-art advances on removal, degradation and electrochemical monitoring of 4-aminophenol pollutants in real samples: a review. environmental research 222 (2023) 115338. https://doi.org/10.1016/j.envres.2023.115338 [9] a. hojjati-najafabadi, m. mansoorianfar, t. liang, k. shahin, h. karimi-maleh. a review on magnetic sensors for monitoring of hazardous pollutants in water resources. science of the total environment 824 (2022) 153844. https://doi.org/10.1016/j.scitotenv.2022.153844 [10] j.a. buledi, n. mahar, a. mallah, a.r. solangi, i.m. palabiyik, n. qambrani, f. karimi, y. vasseghian, h. karimi-maleh. electrochemical quantification of mancozeb through tungsten oxide/reduced graphene oxide nanocomposite: a potential method for environmental remediation. food and chemical toxicology 161 (2022) 112843. https://doi.org/10.1016/j.fct.2022.112843 [11] s. cheraghi, m.a. taher, h. karimi-maleh, f. karimi, m. shabani-nooshabadi, m. alizadeh, a. alothman, n. erk, p.k. yegya raman, c. karaman. novel enzymatic graphene oxide based biosensor for the detection of glutathione in biological body fluids. chemosphere 287 (2022) 132187. https://doi.org/10.1016/j.chemosphere.2021.132187 [12] j. mohanraj, d. durgalakshmi, r.a. rakkesh, s. balakumar, s. rajendran, h. karimi-maleh. facile synthesis of paper based graphene electrodes for point of care devices: a double stranded dna (dsdna) biosensor. journal of colloid and interface science 566 (2020) 463-472. https://doi.org/ 10.1016/j.jcis.2020.01.089 [13] a. moutcine, c. laghlimi, o. ifguis, m.a. smaini, s.e. el qouatli, m. hammi, a. chtaini. a novel carbon paste electrode modified by np-al2o3 for the electrochemical simultaneous detection of pb (ii) and hg (ii). diamond and related materials 104 (2020) 107747. https://doi.org/10.1016/j.diamond.20 20.107747 [14] a. moutcine, c. laghlimi, y. ziat, m.a. smaini, s.e. el qouatli, m. hammi, a. chtaini. preparation, characterization and simultaneous electrochemical detection toward cd (ii) and hg (ii) of a phosphate/zinc oxide modified carbon paste electrode. inorganic chemistry communications 116 (2020) 107911. https://doi.org/10.1016/j.inoche.2020.107911 [15] c. laghlimi, y. ziat, a. moutcine, m. hammi, z. zarhri, o. ifguis, a. chtaini. a new sensor based on graphite carbon paste modified by an organic molecule for efficient. chemical data collections 31 (2021) 100595. https://doi.org/10.1016/j.cdc.2020.100595 [16] c. laghlimi, y. ziat, a. moutcine, m. hammi, z. zarhri, r. maallah, o. ifguis, a. chtaini. analysis of pb (ii), cu (ii) and co (ii) in drinking water by a new carbon paste electrode modified with an organic molecule. chemical data collections 29 (2020) 100496. https://doi.org/10.1016/j.cdc.2020.100496 [17] m.a. smaini, i. smaini, m. ennachete, c. laghlimi, h. saâdane, a. moutcine, a. chtaini. electrochemical determination of adenosine by natural phosphate modified carbon paste electrode: analytical application in serum. sensing and bio-sensing research 23 (2019) 100272. https://doi.org/ 10.1016/j.sbsr.2019.100272 [18] o. ifguis, a. moutcine, c. laghlimi, y. ziat, r. bouhdadi, a. chtaini, a. moubarik, m. mbarki. biopolymer-modified carbon paste electrode for the electrochemical detection of pb (ii) in water. journal of analytical methods in chemistry 2022 (2022). https://doi.org/10.1155/2022/5348246 [19] r. maallah, a. moutcine, c. laghlimi, m.a. smaini, a. chtaini. electrochemical bio-sensor for degradation of phenol in the environment. sensing and bio-sensing research 24 (2019) 100279. https://doi.org/10.1016/j.sbsr.2019.100279 https://doi.org/10.1093/chromsci/bmu127 https://doi.org/10.1016/j.microc.2018.08.062 https://doi.org/10.1016/j.envres.2023.115338 https://doi.org/10.1016/j.scitotenv.2022.153844 https://doi.org/10.1016/j.fct.2022.112843 https://doi.org/10.1016/j.chemosphere.2021.132187 https://doi.org/10.1016/j.jcis.2020.01.089 https://doi.org/10.1016/j.jcis.2020.01.089 https://www.sciencedirect.com/journal/diamond-and-related-materials https://doi.org/10.1016/j.diamond.2020.107747 https://doi.org/10.1016/j.diamond.2020.107747 https://www.sciencedirect.com/journal/inorganic-chemistry-communications https://doi.org/10.1016/j.inoche.2020.107911 https://www.sciencedirect.com/journal/chemical-data-collections https://doi.org/10.1016/j.cdc.2020.100595 https://www.sciencedirect.com/journal/chemical-data-collections https://doi.org/10.1016/j.cdc.2020.100496 https://doi.org/10.1016/j.sbsr.2019.100272 https://doi.org/10.1016/j.sbsr.2019.100272 https://doi.org/10.1155/2022/5348246 https://doi.org/10.1016/j.sbsr.2019.100279 admet & dmpk 11(2) (2023) 151-173 electrochemical sensors and biosensors for monitoring drugs and metabolites doi: https://doi.org/10.5599/admet.1709 167 [20] a. moutcine, o. ifguis, m.a. samaini, m. ennachete, h. sâadane, c. laghlimi, a. chtaini. simultaneous electrochemical determination of heavy metals by an electrode modified cpe-np-al2o3. materials today: proceedings 53 (2022) 404-407. https://doi.org/10.1016/j.matpr.2022.01.388 [21] y. el hamdouni, s. el hajjaji, t. szabo, l. trif, i. felhősi, k. abbi, n. labjar, l. harmouche, a. shaban. biomass valorization of walnut shell into biochar as a resource for electrochemical simultaneous detection of heavy metal ions in water and soil samples: preparation, characterization, and applications. arabian journal of chemistry 15 (2022) 104252. https://doi.org/10.1016/j.arabjc.2022.104252 [22] m.a. mohamed, a.m. yehia, c.e. banks, n.k. allam. novel mwcnts/graphene oxide/pyrogallol composite with enhanced sensitivity for biosensing applications. biosensors and bioelectronics 89 (2017) 1034–1041. https://doi.org/10.1016/j.bios.2016.10.025 [23] s. lotfi, h. veisi. electrochemical determination of clonazepam drug based on glassy carbon electrode modified with fe3o4/r-sh/pd nanocomposite. materials science and engineering : c 103 (2019) 109754. https://doi.org/10.1016/j.msec.2019.109754 [24] m. fallah, m. rahimnejad, m. asghary, m. mashkour. an electrochemical sensor based on a carbon paste electrode for the determination of buserelin. analytical methods 12 (2020) 33-38. https://doi.org/10.1039/c9ay01760g [25] a. zarezadeh, h.r. rajabi, o. sheydaei, h. khajehsharifi. application of a nano-structured molecularly imprinted polymer as an efficient modifier for the design of captopril drug selective sensor: mechanism study and quantitative determination. materials science and engineering: c 94 (2019) 879-885. https://doi.org/10.1016/j.msec.2018.10.042 [26] n.p. shetti, d. ilager, s.j. malode, d. monga, s. basu, k.r. reddy. poly (eriochrome black t) modified electrode for electrosensing of methdilazine. materials science in semiconductor processing 120 (2020) 105261. https://doi.org/10.1016/j.mssp.2020.105261 [27] k. basavaiah, v.s. charan. the use of chloranilic acid for the spectrophotometric determination of three antihistamines. turkish journal of chemistry 26 (2002) 653-661. https://journals.tubit ak.gov.tr/chem/vol26/iss5/3/ [28] m.s. raghu, k.b. cijo, m. xavier, k.n. prashanth. development and validation of stability-indicating rp-uplc method for the determination of methdilazine in bulk drug and in pharmaceutical dosage form. international scholarly research network chromatography 2012 (2012) 1-8. https://doi.org/ 10.5402/2012/916932 [29] m.s. raghu, k. basavaiah. simple, sensitive and selective spectrophotometric methods for the determination of methdilazine in pharmaceuticals through charge transfer complex formation reaction. proceedings of the indian national science academy 79 (2013) 65-78. http://eprints.uni-mysore.ac.in/ 10806/ [30] j.c. kemmegne-mbouguen, f.p. tchoumi. synthesis of nanozeolites type a and x from quartz-rich cameroonian kaolin: application to the modification of carbon paste electrode for acetaminophen and epinine electrochemical sensing. journal of solid state electrochemistry 27 (2023) 1-15. https://doi.org/10.1007/s10008-022-05355-z [31] s. tajik, f. sharifi, b. aflatoonian, a. di bartolomeo. a new electrochemical sensor for the detection of ketoconazole using carbon paste electrode modified with sheaf-like ce-btc mof nanostructure and ionic liquid. nanomaterials 13 (2023) 523. https://doi.org/10.3390/nano13030523 [32] k. liu, h. you, g. jia, y. zheng, y. huang, y. song, m. yang, l. zhang, h. zhang. hierarchically nanostructured coordination polymer: facile and rapid fabrication and tunable morphologies. crystal growth & design 10 (2010) 790-797. https://doi.org/10.1021/cg901170j [33] j. zoubir, n. bougdour, w.e. hayaoui, c. radaa, a. idlahcen, a. assabbane, i. bakas. electrochemical detection of metronidazole using silver nanoparticle-modified carbon paste electrode. electrocatalysis 13 (2022) 386-401. https://doi.org/10.1007/s12678-022-00722-w [34] o. vajdle, s. šekuljica, v. guzsvány, l. nagy, z. kónya, m.a. ivić, d. mijin, s. petrović, j. anojčić. use of carbon paste electrode and modified by gold nanoparticles for selected macrolide antibiotics determination as standard and in pharmaceutical preparations. journal of electroanalytical chemistry 873 (2020) 114324. https://doi.org/10.1016/j.jelechem.2020.114324 https://doi.org/10.5599/admet.1709 https://doi.org/10.1016/j.matpr.2022.01.388 https://doi.org/10.1016/j.arabjc.2022.104252 https://doi.org/10.1016/j.bios.2016.10.025 https://doi.org/10.1016/j.msec.2019.109754 https://doi.org/10.1039/c9ay01760g https://doi.org/10.1016/j.msec.2018.10.042 https://doi.org/10.1016/j.mssp.2020.105261 https://journals.tubitak.gov.tr/chem/vol26/iss5/3/ https://journals.tubitak.gov.tr/chem/vol26/iss5/3/ https://doi.org/10.5402/2012/916932 https://doi.org/10.5402/2012/916932 http://eprints.uni-mysore.ac.in/10806/ http://eprints.uni-mysore.ac.in/10806/ https://www.researchgate.net/journal/journal-of-solid-state-electrochemistry-1433-0768?_sg=aoztind8gtkdcqg-_hswpoiqf4prf4s3wvjildyyzjwxne9onee-7c3sdwl5xupe9cbrj-fvrikt_hxbayer8mte37uybw.iqzmmr2f0dhkzfd2okwksbbisl6gkjbkxzlhxuj9olph107gugv6tjvtcwaihtq1l8-8d-ahcut70g1clrzhng https://doi.org/10.1007/s10008-022-05355-z https://doi.org/10.3390/nano13030523 https://pubs.acs.org/crystal https://pubs.acs.org/crystal https://doi.org/10.1021/cg901170j https://doi.org/10.1007/s12678-022-00722-w https://doi.org/10.1016/j.jelechem.2020.114324 c. laghlimi et al. admet & dmpk 11(2) (2023) 151-173 168 [35] s.a.c. carabineiro, t. thavorn-amornsri, m.f. pereira, j.l. figueiredo. adsorption of ciprofloxacin on surface-modified carbon materials. water research 45 (2011) 4583-4591. https://doi.org/ 10.1016/j.watres.2011.06.008 [36] g.d. sheng, d.d. shao, x.m. ren, x.q. wang, j.x. li, y.x. chen, x.k. wang. kinetics and thermosdynamics of adsorption of ionizable aromatic compounds from aqueous solutions by as-prepared and oxidized multiwalled carbon nanotubes. journal of hazardous materials 178 (2010) 505-516. https://doi.org/10.1016/j.jhazmat.2010.01.110 [37] h. peng, b. pan, m. wu, r. liu, d. zhang, d. wu, b. xing. adsorption of ofloxacin on carbon nanotubes: solubility, ph and cosolvent effects. journal of hazardous materials 211 (2012) 342-348. https://doi.org/10.1016/j.jhazmat.2011.12.063 [38] m. keiluweit, m. kleber. molecular-level interactions in soils and sediments: the role of aromatic πsystems. environmental science and technology 43 (2009) 3421-3429. https://doi.org/10.1021/es 8033044 [39] d. lin, b. xing. adsorption of phenolic compounds by carbon nanotubes: role of aromaticity and substitution of hydroxyl groups. environmental science and technology 42 (2008) 7254-7259. https://doi.org/10.1021/es801297u [40] l.s. lee, j.j. pignatello. evidence π–π for electron donor–acceptor interactions between π-donor aromatic compounds and π-acceptor sites in soil organic matter through ph effects on sorption. environmental science and technology 38 (2004) 4361-4368. https://doi.org/10.1021/es035379e [41] m. elfiky, n. salahuddin, a. hassanein, a. matsuda, t. hattori. detection of antibiotic ofloxacin drug in urine using electrochemical sensor based on synergistic effect of different morphological carbon materials. microchemical journal 146 (2019) 170-177. https://doi.org/10.1016/j.microc.2018.12.034 [42] y. chang, l.x. wang, y.p. li, c.q. hu. factors influencing the hplc determination for related substances of azithromycin. journal of chromatographic science 54 (2016) 187-194. https://doi.org/ 10.1093/chromsci/bmv127 [43] b. wu, y. guo, h. cao, y. zhang, l. yu, n. jia. a novel mesoporous molecularsieves-based electrochemilumenescence sensor for sensitive detection of azithromycin. sensors and actuators b: chemical 186 (2013) 219-225. https://doi.org/10.1016/j.snb.2013.05.080 [44] y.m. liu, y.m. shi, z.l. liu, w. tian. a sensitive method for simultaneous determination of four macrolides by ce with electrochemiluminescence detection and its applications in human urine and tablets. electrophoresis 31 (2010) 364-370. https://doi.org/10.1002/elps.200900302 [45] l. hu, t. zhou, j. feng, h. jin, y. tao, d. luo, s. mei, y.i. lee. a rapid and sensitive molecularly imprinted electrochemiluminescence sensor for azithromycin determination in biological samples. journal of electroanalytical chemistry 813 (2018) 1-8. https://doi.org/10.1016/j.jelechem.2018.02.010 [46] t.t. zhu, c.h. zhou, f.b. kabwe, q.q. wu, c.s. li, j.r. zhang. exfoliation of montmorillonite and related properties of clay/polymernanocomposites. applied clay science 169 ( 2019) 48-66. https://doi.org/10.1016/j.clay.2018.12.006 [47] c. tournassat, c.i. steefel, i.c. bourg, f. bergaya. surface properties of clay minerals. developments in clay science 6 (2015) 5-31. https://doi.org/10.1016/b978-0-08-100027-4.00001-2 [48] b.t.d. justin, n. blaise, h.g. valery. investigation of the photoactivation effect of tio2 onto carbonclay paste electrode by cyclic voltammetry analysis. heliyon 9 (2023) 13474. https://doi.org/ 10.1016/j.heliyon.2023.e13474 [49] m. pawar, s. topcu sendoğdular, p. gouma. a brief overview of tio2 photocatalyst for organic dye remediation: case study of reaction mechanisms involved in ce-tio2 photocatalysts system. journal of nanomaterials 2018 (2018) 13. https://doi.org/10.1155/2018/5953609 [50] v. vinoth, t.d. shergilin, a.m. asiri, j.j. wu, s. anandan. facile synthesis of copper oxide microflowers for nonenzymatic glucose sensor applications. materials science in semiconductor processing 82 (2018) 31-38. https://doi.org/10.1016/j.mssp.2018.03.032 https://www.researchgate.net/journal/water-research-0043-1354 https://doi.org/10.1016/j.watres.2011.06.008 https://doi.org/10.1016/j.watres.2011.06.008 https://www.researchgate.net/journal/journal-of-hazardous-materials-0304-3894 https://doi.org/10.1016/j.jhazmat.2010.01.110 https://pubmed.ncbi.nlm.nih.gov/?term=pan+b&cauthor_id=22264890 https://pubmed.ncbi.nlm.nih.gov/?term=wu+m&cauthor_id=22264890 https://pubmed.ncbi.nlm.nih.gov/?term=liu+r&cauthor_id=22264890 https://pubmed.ncbi.nlm.nih.gov/?term=zhang+d&cauthor_id=22264890 https://pubmed.ncbi.nlm.nih.gov/?term=wu+d&cauthor_id=22264890 https://pubmed.ncbi.nlm.nih.gov/?term=xing+b&cauthor_id=22264890 https://www.researchgate.net/journal/journal-of-hazardous-materials-0304-3894 https://doi.org/10.1016/j.jhazmat.2011.12.063 https://doi.org/10.1016/j.jhazmat.2011.12.063 https://www.researchgate.net/journal/environmental-science-and-technology-1520-5851 https://doi.org/10.1021/es8033044 https://doi.org/10.1021/es8033044 https://www.researchgate.net/journal/environmental-science-and-technology-1520-5851 https://doi.org/10.1021/es801297u https://www.researchgate.net/journal/environmental-science-and-technology-1520-5851 https://doi.org/10.1021/es035379e https://doi.org/10.1016/j.microc.2018.12.034 https://www.researchgate.net/journal/journal-of-chromatographic-science-1945-239x https://doi.org/10.1093/chromsci/bmv127 https://doi.org/10.1093/chromsci/bmv127 https://doi.org/10.1016/j.snb.2013.05.080 https://doi.org/10.1002/elps.200900302 https://doi.org/10.1016/j.jelechem.2018.02.010 https://doi.org/10.1016/j.clay.2018.12.006 https://doi.org/10.1016/b978-0-08-100027-4.00001-2 https://doi.org/10.1016/j.heliyon.2023.e13474 https://doi.org/10.1016/j.heliyon.2023.e13474 https://doi.org/10.1155/2018/5953609 https://doi.org/10.1016/j.mssp.2018.03.032 admet & dmpk 11(2) (2023) 151-173 electrochemical sensors and biosensors for monitoring drugs and metabolites doi: https://doi.org/10.5599/admet.1709 169 [51] n.q. dung, d.patil, h. jung, j. kim, d. kim. nio-decorated single-walled carbon nanotubes for highperformance nonenzymatic glucose sensing. sensors and actuators b: chemical 183 (2013) 381-387. https://doi.org/10.1016/j.snb.2013.04.018 [52] n. xiao, j. deng, j. cheng, s. ju, h. zhao, j. xie, d. qian, j. he. carbon paste electrode modified with duplex molecularly imprinted polymer hybrid film for metronidazole detection. biosensors and bioelectronics 81 (2016) 54-60. http://dx.doi.org/10.1016/j.bios.2016.02.041 [53] m. shahsavari, s. tajik, i. sheikhshoaie, h. beitollahi. fabrication of nanostructure electrochemical sensor based on the carbon baste electrode (cpe) modified with ionic liquid and fe3o4/zif-67 for electrocatalytic sulfamethoxazole detection. topics in catalysis 65 (2021) 1-10. https://doi.org/10.1007/s11 244-021-01471-8 [54] n. muthuchamy, r. atchudan, t.n.j.i. edison, s. perumal, y.r. lee. high-performance glucose biosensor based on green synthesized zinc oxide nanoparticle embedded nitrogen-doped carbon sheet. journal of electroanalytical chemistry 816 (2018) 195-204. https://doi.org/10.1016/j.jelechem.2018.03.059 [55] a. heidari, e. kalantar, m. dehghan, p. fallah. herpes simplex encephalitis: successful treatment with acyclovir. archives of advances in biosciences 10 (2019) 56-58. https://doi.org/10.22037/jps.v10i1.19886 [56] d.j. newman, g.m. cragg. natural products as sources of new drugs from 1981 to 2014. journal of natural products 79 (2016) 629-661. https://doi.org/10.1021/acs.jnatprod.5b01055 [57] x. hong, x. wang, z. wang. a rare case report of acyclovir-induced immune thrombocytopenia with tongue hematomas as the first sign, and a literature review. biomed central pharmacology and toxicology 18 (2017) 1-5. https://doi.org/10.1186/s40360-017-0120-2 [58] z. shen, q. yu, y. li, y. bao, h. lu. determination of acyclovir in renal microdialysis fluid and confirmation of renal function index. drug and chemical toxicology 43 (2020) 574-580. https://doi.org/10.1080/01480545.2018.1524474 [59] s.s. sen, a. si, u. chakraborty, a. chandra. stevens-johnson syndrome-toxic epidermal necrolysis: a fatal cutaneous adverse reaction to oral acyclovir. british medical journal case reports 13 (2020) 8. https://doi.org/10.1136/bcr-2020-238555 [60] k. chu, q. li, y. cheng, y. liu. efficient electrocatalytic nitrogen fixation on femoo4 nanorods. acs applied materials & interfaces 12 (2020) 11789-11796. https://doi.org/10.1021/acsami.0c00860 [61] l. tang, l. liu, f. yang, femoo4-graphene oxide photo-electro-catalyst for berberine removal and hydrogen evolution. international journal of hydrogen energy 44 (2019) 19755-19761. https://doi.org/10.1016/j.ijhydene.2019.03.078 [62] y. wei, l. yao, y. wu, x. liu, j. feng, j. ding, k. li, q. he. ultrasensitive electrochemical detection for nanomolarity acyclovir at ferrous molybdate nanorods and graphene oxide composited glassy carbon electrode. colloids and surfaces a: physicochemical and engineering aspects 641 (2022) 128601. https://doi.org/10.1016/j.colsurfa.2022.128601 [63] y. huang, d. lu, h. liu, s. liu, s. jiang, g.c. pang, y. liu. preliminary research on the receptor–ligand recognition mechanism of umami by an ht1r1 biosensor. food & function 10 (2019) 1280-1287. https://doi.org/10.1039/c8fo02522c [64] a.j. venkatakrishnan, x. deupi, g. lebon, c.g. tate, g.f. schertler, m.m. babu. molecular signatures of gproteincoupled receptors. nature 494 (2013) 185-194. https://doi.org/10.1038/nature11896 [65] l. wei, x. wang, d. lu, y. li, g. pang, j. xie. a novel staphylococcal enterotoxin q immunosensor prepared with self-assembly method based on horseradish peroxidase and double-layer gold nanoparticles. food analytical methods 10 (2016) 892-899. https://doi.org/10.1007/s12161-016-0632-1 [66] d. lu, d. liu, y. liu, x. wang, y. liu, s. yuan, r. ren, g. pang. comparative study on the sensing kinetics of carbon and nitrogen nutrients in cancer tissues and normal tissues based electrochemical biosensors. molecules 28 (2023) 1453. https://doi.org/10.3390/molecules28031453 [67] b. karupppaiah, a. jeyaraman, s.m. chen, y.c. huang. development of highly sensitive electrochemical sensor for antibiotic drug ronidazole based on spinel cobalt oxide nanorods embedded with hexagonal boron nitride. electrochimica acta 446 (2023) 142008. https://doi.org/10.1016/j.electacta.2023.14 2008 https://doi.org/10.5599/admet.1709 https://doi.org/10.1016/j.snb.2013.04.018 http://dx.doi.org/10.1016/j.bios.2016.02.041 https://doi.org/10.1007/s11244-021-01471-8 https://doi.org/10.1007/s11244-021-01471-8 https://doi.org/10.1016/j.jelechem.2018.03.059 https://doi.org/10.22037/jps.v10i1.19886 https://doi.org/10.1021/acs.jnatprod.5b01055 https://doi.org/10.1186/s40360-017-0120-2 https://doi.org/10.1080/01480545.2018.1524474 https://doi.org/10.1136/bcr-2020-238555 https://doi.org/10.1021/acsami.0c00860 https://doi.org/10.1016/j.ijhydene.2019.03.078 https://doi.org/10.1016/j.colsurfa.2022.128601 https://doi.org/10.1039/c8fo02522c https://doi.org/10.1038/nature11896 https://doi.org/10.1007/s12161-016-0632-1 https://doi.org/10.3390/molecules28031453 https://doi.org/10.1016/j.electacta.2023.142008 https://doi.org/10.1016/j.electacta.2023.142008 c. laghlimi et al. admet & dmpk 11(2) (2023) 151-173 170 [68] j. pu, k. zhang, z. wang, c. li, k. zhu, y. yao, g. hong. synthesis and modification of boron nitride nanomaterials for electrochemical energy storage: from theory to application. advanced functional materials 31 (2021) 1-32. https://doi.org/10.1002/adfm.202106315 [69] a. garcía-miranda ferrari, s.j. rowley-neale, c.e. banks. recent advances in 2d hexagonal boron nitride (2d-hbn) applied as the basis of electrochemical sensing platforms. analytical and bioanalytical chemistry 413 (2021) 663-672. https://doi.org/10.1007/s00216-020-03068-8 [70] m. rafiq, x. hu, z. ye, a. qayum, h. xia, l. hu, f. lu, p.k. chu. recent advances in structural engineering of 2d hexagonal boron nitride electrocatalysts. nano energy 91 (2022) 106661. https://doi.org/10.1016/j.nanoen.2021.106661 [71] h.k. ahn, b. han, s.j. lee, t. lim, j.m. sun, j.s. ahn, m.j. ahn, k. park. alk inhibitor crizotinib combined with intrathecal methotrexate treatment for non-small cell lung cancer with leptomeningeal carcinomatosis. lung cancer 76 (2012) 253-254. https://doi.org/10.1016/j.lungcan.2012.02.003 [72] e.e. cohen, l.f. licitra, b. burtness, j. fayette, t. gauler, p.m. clement, j.j. grau, j.m. del campo, a. mailliez, r.i. haddad, j.b. vermorken, m. tahara, j. guigay, l. geoffrois, m.c. merlano, n. dupuis, n. krämer, x.j. cong, n. gibson, f. solca, j.p. machiels. biomarkers predict enhanced clinical outcomes with afatinib versus methotrexate in patients with second-line recurrent and/or metastatic head and neck cancer. annals of oncology 28 (2017) 2526-2532. https://doi.org/10.1093/annonc/mdx344 [73] d. huang, h. wu, y. zhu, h. su, h. zhang, l. sheng, z. liu, h. xu, c. song. sensitive determination of anticancer drug methotrexate using graphite oxide-nafion modified glassy carbon electrode. international journal of electrochemical science 14 (2019) 3792-3804. https://doi.org/10.20964/2019.04.03 [74] e. asadian, s. shahrokhian, a. i. zad, f. ghorbani-bidkorbeh. glassy carbon electrode modified with 3d graphene–carbon nanotube network for sensitive electrochemical determination of methotrexate. sensors and actuators b 239 (2017) 617-627. https://doi.org/10.1016/j.snb.2016.08.064 [75] c. luhana, p. mashazi. simultaneous detection of dopamine and paracetamol on electroreduced graphene oxide–cobalt phthalocyanine polymer nanocomposite electrode. electrocatalysis 14 (2022) 1-12. https://doi.org/10.1007/s12678-022-00806-7 [76] a. wong, a.m. santos, f.h. cincotto, f.c. moraes, o. fatibello-filho, m.d. sotomayor. a new electrochemical platform based on low cost nanomaterials for sensitive detection of the amoxicillin antibiotic in different matrices. talanta 206 (2020) 120252. https://doi.org/10.1016/j.talanta.2019.120252 [77] a. pollap, p. knihnicki, p. kuśtrowski, j. kozak, m. gołda‐cępa, a. kotarba, j. kochana. sensitive voltammetric amoxicillin sensor based on tio2 sol modified by cmk‐3‐type mesoporous carbon and gold ganoparticles. electroanalysis 30 (2018) 2386-2396. https://doi.org/10.1002/elan.201800203 [78] p.b. deroco, r.c. rocha-filho, o. fatibello-filho. a new and simple method for the simultaneous determination of amoxicillin and nimesulide using carbon black within a dihexadecylphosphate film as electrochemical sensor. talanta 179 (2018) 115-123. https://doi.org/10.1016/j.talanta.20 17.10.048 [79] n. kumar, rosy, r.n. goyal. gold-palladium nanoparticles aided electrochemically reduced graphene oxide sensor for the simultaneous estimation of lomefloxacin and amoxicillin. sensors and actuators b: chemical 243 (2017) 658-668. https://doi.org/10.1016/j.snb.2016.12.025 [80] x. yue, z. li, s. zhao. a new electrochemical sensor for simultaneous detection of sulfamethoxazole and trimethoprim antibiotics based on graphene and zno nanorods modified glassy carbon electrode. microchemical journal 159 (2020) 105440. https://doi.org/10.1016/j.microc.2020.105440 [81] d.l.c. golinelli, s.a.s. machado, i. cesarino. synthesis of silver nanoparticle-graphene composites for electroanalysis applications using chemical and electrochemical methods. electroanalysis 29 (2017) 1014-1021. https://doi.org/10.1002/elan.201600669 [82] n.q. man, n.t.t. tu, n.t.v. hoan, h.x.a. vu, l.l. son, n.d.v. quyen, d.n. nhiem, n.h. phong, v.t. nguyen, t.n. tuyen, d.q. khieu. electrochemical determination of clenbuterol with nickelferrite/reduced-graphene-oxide-modified electrode. journal of nanoparticle research 25 (2023). https://doi.org/10.1007/s11051-023-05679-5 https://doi.org/10.1002/adfm.202106315 https://doi.org/10.1007/s00216-020-03068-8 https://doi.org/10.1016/j.nanoen.2021.106661 https://doi.org/10.1016/j.lungcan.2012.02.003 https://doi.org/10.1093/annonc/mdx344 https://doi.org/10.20964/2019.04.03 https://doi.org/10.1016/j.snb.2016.08.064 https://doi.org/10.1007/s12678-022-00806-7 https://doi.org/10.1016/j.talanta.2019.120252 https://doi.org/10.1002/elan.201800203 https://doi.org/10.1016/j.talanta.2017.10.048 https://doi.org/10.1016/j.talanta.2017.10.048 https://doi.org/10.1016/j.snb.2016.12.025 https://doi.org/10.1016/j.microc.2020.105440 https://doi.org/10.1002/elan.201600669 https://link.springer.com/article/10.1007/s11051-023-05679-5#auth-nguyen_thi_vuong-hoan https://link.springer.com/article/10.1007/s11051-023-05679-5#auth-ho_xuan_anh-vu https://link.springer.com/article/10.1007/s11051-023-05679-5#auth-le_lam-son https://link.springer.com/article/10.1007/s11051-023-05679-5#auth-nguyen_duc_vu-quyen https://link.springer.com/article/10.1007/s11051-023-05679-5#auth-dao_ngoc-nhiem https://link.springer.com/article/10.1007/s11051-023-05679-5#auth-nguyen_hai-phong https://link.springer.com/article/10.1007/s11051-023-05679-5#auth-vo_thang-nguyen https://link.springer.com/article/10.1007/s11051-023-05679-5#auth-vo_thang-nguyen https://link.springer.com/article/10.1007/s11051-023-05679-5#auth-tran_ngoc-tuyen https://link.springer.com/article/10.1007/s11051-023-05679-5#auth-dinh_quang-khieu https://doi.org/10.1007/s11051-023-05679-5 admet & dmpk 11(2) (2023) 151-173 electrochemical sensors and biosensors for monitoring drugs and metabolites doi: https://doi.org/10.5599/admet.1709 171 [83] s. wu, x. dai, t. cheng, s. li. highly sensitive and selective ion-imprinted polymers based on one-step electrodeposition of chitosan-graphene nanocomposites for the determination of cr(vi). carbohydrate polymers 195 (2018) 199-206. https://doi.org/10.1016/j.carbpol.2018.04.077 [84] l. chen, s. xu, j. li. recent advances inmolecular imprinting technology: current status, challenges and highlighted applications. chemical society reviews 40 (2011) 2922-2942. https://doi.org/ 10.1039/c0cs00084a [85] y. wu, p. deng, y. tian. rapid recognition and determination of tryptophan by carbon nanotubes and molecularly imprinted polymer modified glassy carbonelectrode. bioelectrochemistry 19 (2020) 107493. https://doi.org/10.1016/j.bioelechem.2019.107393 [86] q. he, j. liu, j. liang, x. liu, w. li, z. liu, z. ding, d. tuo. towards improvements for penetrating the blood–brain barrier—recent progress from a material and pharmaceuticalperspective. cells 7 (2018) 24. https://doi.org/10.3390/cells7040024 [87] d. duan, h. yang, y. ding, l. li, g. ma. a three-dimensional conductive molecularly imprinted electrochemical sensor based on mof derived porous carbon/carbon nanotubes composites and prussian blue nanocubesmediated amplification for chiral analysis of cysteine enantiomers. electrochimica acta 302 (2019) 137-144. https://doi.org/10.1016/j.electacta.2019.02.028 [88] h. wei, y. wang, j. guo, b. qiu, d. ding, s. wei, e.k. wujcik, z. guo. synthesis of multifunctional carbon nanostructures. handbook of carbon nano materials 7 (2015) 89. https://doi.org/ 10.1142/9789814678919_0003 [89] m. kooshki, h. abdollahi, s. bozorgzadeh, b. haghighi. second-order data obtained from differential pulse voltammetry: determination of tryptophan at a gold nanoparticles decorated multiwalled carbon nanotube modified glassy carbon electrode. electrochimica acta 56 (2011) 8618-8624. https://doi.org/10.1016/j.electacta.2011.07.049 [90] l. qiu, g. zhou, s. cao. targeted inhibition of ulk1 enhances daunorubicin sensitivity in acute myeloid leukemia. life sciences 243 (2020) 117234. https://doi.org/10.1016/j.lfs.2019.117234 [91] q. nong, c. zhang, q. liu, r. xie, m. dong. effect of daunorubicin on acute promyelocytic leukemia cells using nuclear magnetic resonance spectroscopy-based metabolomics. environmental toxicology and pharmacology 78 (2020) 103382. https://doi.org/10.1016/j.etap.2020.103382 [92] m. alizadeh, y. orooji, f. karimi, m. baghayeri, j. rouhi, s. tajik, h. beitollahi, s. agarwal, v.k. gupta, s. rajendran, s. rostamnia, l. fu, f.s. movahed, s. malekmohammadi. guanine-based dna biosensor amplified with pt/swcnts nanocomposite as analytical tool for nanomolar determination of daunorubicin as an anticancer drug: a docking/experimental investigation. industrial & engineering chemistry research 60 (2021) 816-823. https://doi.org/10.1021/acs.iecr.0c04698 [93] t. bhengo, m. moyo, m. shumba, o.j. okankwo. simultaneous oxidative determination of antibacterial drugs in aqueous solutions using an electrode modified with mwcnts decorated with fe3o4 nanoparticles. new journal of chemistry 42 (2018) 5014-5023. https://doi.org/10.1039/ c8nj00129d [94] r. chokkareddy, s. kanchi, g.g. redhi. a novel il-f-znonps@mwcnts nanocomposite fabricated glassy carbon electrode for the determination of sulfamethoxazole. journal of molecular liquids 359 (2022) 119232.https://doi.org/10.1016/j.molliq.2022.119232 [95] a. yari, a. shams. silver-filled mwcnt nanocomposite as a sensing element for voltammetric determination of sulfamethoxazole. analytica chimica acta 1039 (2018) 51-58. https://doi.org/ 10.1016/j.aca.2018.07.061 [96] n. murugan, r. jerome, m. preethika, a. sundaramurthy, a.k. sundramoorthy. 2d-titanium carbide (mxene) based selective electrochemical sensor for simultaneous detection of ascorbic acid, dopamine and uric acid. journal of materials science & technology 72 (2021) 122-131. https://doi.org/10.1016/j.jmst.2020.07.037 [97] y. wu, p. deng, y. tian, j. feng, j. xiao, j. li, j. liu, g. li, q. he. simultaneous and sensitive determination of ascorbic acid, dopamine and uric acid via an electrochemical sensor based on pvpgraphene composite. journal of nanobiotechnology 18 (2020) 112. https://doi.org/10.1186/s12951020-00672-9 https://doi.org/10.5599/admet.1709 https://doi.org/10.1016/j.carbpol.2018.04.077 https://doi.org/10.1039/c0cs00084a https://doi.org/10.1039/c0cs00084a https://doi.org/10.1016/j.bioelechem.2019.107393 https://doi.org/10.3390/cells7040024 https://doi.org/10.1016/j.electacta.2019.02.028 https://www.worldscientific.com/doi/abs/10.1142/9789814678919_0003 https://www.worldscientific.com/doi/abs/10.1142/9789814678919_0003 https://doi.org/10.1142/9789814678919_0003 https://doi.org/10.1142/9789814678919_0003 https://doi.org/10.1016/j.electacta.2011.07.049 https://doi.org/10.1016/j.lfs.2019.117234 https://doi.org/10.1016/j.etap.2020.103382 https://doi.org/10.1021/acs.iecr.0c04698 https://doi.org/10.1039/c8nj00129d https://doi.org/10.1039/c8nj00129d https://doi.org/10.1016/j.molliq.2022.119232 https://doi.org/10.1016/j.aca.2018.07.061 https://doi.org/10.1016/j.aca.2018.07.061 https://doi.org/10.1016/j.jmst.2020.07.037 https://doi.org/10.1186/s12951-020-00672-9 https://doi.org/10.1186/s12951-020-00672-9 c. laghlimi et al. admet & dmpk 11(2) (2023) 151-173 172 [98] x. zhang, y.c. zhang, l.x. ma. one-pot facile fabrication of graphene-zinc oxide composite and its enhanced sensitivity for simultaneous electrochemical detection of ascorbic acid, dopamine and uric acid. sensors and actuators b: chemical 227 (2016) 488-496. https://doi.org/10.1016/j.snb.2015.12.073 [99] m. yang, m. guo, y. feng, y. lei, y. cao, d. zhu, y. yu, l. ding. sensitive voltammetric detection of metronidazole based on three-dimensional graphene-like carbon architecture/polythionine modified glassy carbon electrode. journal of the electrochemical society 165 (2018) b530. https://doi.org/ 10.1149/2.1311811jes [100] s. tursynbolat, y. bakytkarim, j. huang, l. wang. ultrasensitive electrochemical determination of metronidazole based on polydopamine/carboxylic multiwalled carbon nanotubes nanocomposites modified gce. journal of pharmaceutical analysis 8 (2018) 124-130. https://doi.org/10.1016/ j.jpha.2017.11.001 [101] a. yari, a. shams. silver-filled mwcnt nanocomposite as a sensing element for voltammetric determination of sulfamethoxazole. analytica chimica acta 1039 (2018) 51-58. https://doi.org/ 10.1016/j.aca.2018.07.061 [102] c. zhang, y. ma, x. zhang, s. abdolhosseinzadeh, h. sheng, w. lan, a. pakdel, j. heier, f. nüesch. two‐dimensional transition metal carbides and nitrides (mxenes): synthesis, properties, and electrochemical energy storage applications. energy & environmental materials 3 (2020) 29-55. https://doi.org/10.1002/eem2.12058 [103] j. halim, s. kota, m.r. lukatskaya, m. naguib, m.q. zhao, e.j. moon, j. pitock, j. nanda, s.j. may, y. gogotsi, m.w. barsoum. synthesis and characterization of 2d molybdenum carbide (mxene). advanced functional materials 26 (2016) 3118-3127. https://doi.org/10.1002/adfm.201505328 [104] f. zhou, w. jing, s. liu, q. mao, y. xu, f. han, z. wei, z. jiang. electrodeposition of gold nanoparticles on zno nanorods for improved performance of enzymatic glucose sensors. materials science in semiconductor processing 105 (2020) 104708. https://doi.org/10.1016/j.mssp.2019.104708 [105] a. granito, a. forgione, s. marinelli, m. renzulli, l. lelasi, v. sansone, f. benevento, f. piscaglia, f. tovoli. regorafenib combined with other systemic therapies: exploring promising therapeutic combinations in hcc. journal of hepatocellular carcinoma 8 (2021) 477-492. https://doi.org/ 10.2147/jhc.s251729 [106] d. merzak, s.i. kaya, c. ahmet, k.b. nurgul, t. mohamed, a.o. sibel. detailed electrochemical behavior and thermodynamic parameters of anticancer drug regorafenib and its sensitive electroanalytical assay in biological and pharmaceutical samples. microchemical journal 170 (2021) 106717. https://doi.org/10.1016/j.microc.2021.106717 [107] h. kim, y. bang, m.a. lee, j.w. kim, h.k. jee, h.k. chon, b. kang, m.j. kang, i. kim, j. cheon, j.e. hwang, j.h. kang, s. byeon, j.y. hong, b.y. ryoo, b.y. lim, c. yoo. regorafenib in patients with advanced child-pugh b hepatocellular carcinoma: a multicentre retrospective study. liver international 40 (2020) 2544-2552. https://doi.org/10.1111/liv.14573 [108] b. sravani, s. kiranmai, g.r. reddy, j.p. park, y. v. reddy, g. madhavi. highly sensitive detection of anticancer drug based on bimetallic reduced graphene oxide nanocomposite. chemosphere 287 (2022) 132281. https://doi.org/10.1016/j.chemosphere.2021.132281 [109] e. laviron. general expression of the linear potential sweep voltammogram in the case of diffusionless electrochemical systems. journal of electroanalytical chemistry and interfacial electrochemistry 101 (1979) 19-28. https://doi.org/10.1016/s0022-0728(79)80075-3 [110] e. naghian, f. shahdostfard, e. sohouli, v. safarifard, m. najafi, m. rahiminasrabadi, a. sobhaninasab. electrochemical determination of levodopa on a reduced graphene oxide paste electrode modified with a metal-organic framework. microchemical journal 156 (2020) 104888. https://doi.org/10.1016/j.microc.2020.104888 [111] d. vasu, a. karthi keyan, s. sakthinathan, t.w. chiu. investigation of electrocatalytic and photocatalytic ability of cu/ni/tio2/mwcnts nanocomposites for detection and degradation of antibiotic drug furaltadone. scientific reports 12 (2022) 1-16. https://doi.org/10.1038/s41598-022-04890-z https://doi.org/10.1016/j.snb.2015.12.073 https://doi.org/10.1149/2.1311811jes https://doi.org/10.1149/2.1311811jes https://doi.org/10.1016/j.jpha.2017.11.001 https://doi.org/10.1016/j.jpha.2017.11.001 https://doi.org/10.1016/j.aca.2018.07.061 https://doi.org/10.1016/j.aca.2018.07.061 https://doi.org/10.1002/eem2.12058 https://onlinelibrary.wiley.com/action/dosearch?contribauthorraw=pitock%2c+jeremy https://onlinelibrary.wiley.com/action/dosearch?contribauthorraw=nanda%2c+jagjit https://onlinelibrary.wiley.com/action/dosearch?contribauthorraw=may%2c+steven+j https://onlinelibrary.wiley.com/action/dosearch?contribauthorraw=gogotsi%2c+yury https://onlinelibrary.wiley.com/action/dosearch?contribauthorraw=gogotsi%2c+yury https://doi.org/10.1002/adfm.201505328 https://doi.org/10.1016/j.mssp.2019.104708 https://doi.org/10.2147/jhc.s251729 https://doi.org/10.2147/jhc.s251729 https://doi.org/10.1016/j.microc.2021.106717 https://doi.org/10.1111/liv.14573 https://doi.org/10.1016/j.chemosphere.2021.132281 https://doi.org/10.1016/s0022-0728(79)80075-3 https://doi.org/10.1016/j.microc.2020.104888 https://www.researchgate.net/profile/te-wei-chiu?_sg%5b0%5d=zaznxxu7kp7icdumiaslerbdzbf_ob8wg3y_xn-_j724_tetcqj8zfnhin0pj2kdjj0lkei.jzmoguzpzsu_yzaaspsnb9yasujdj0kbsk7pnzunujvtloju7xg9rjxlor06w4_xi3huzjspvexyjre04iwlaa&_sg%5b1%5d=1za6ksix4cyitlwyxqlkxtsemkmll1bcmxzass3-fddcvml-le3jpid8g-xnex5vyz98vgg.ru5m5lhyymftjvlnnsmjhfxlkdpsrpox5vuslcgqykc67jkumdf9by_hebklh6r-1yayobmqjnbk34wrkcjndg https://doi.org/10.1038/s41598-022-04890-z admet & dmpk 11(2) (2023) 151-173 electrochemical sensors and biosensors for monitoring drugs and metabolites doi: https://doi.org/10.5599/admet.1709 173 [112] a. veseli, f. mullallari, f. balidemaj, l. berisha, ľ. švorc, t. arbneshi. electrochemical determination of erythromycin in drinking water resources by surface modified screen-printed carbon electrodes. microchemical journal 148 (2019) 412-418. https://doi.org/10.1016/j.microc.2019.04.086 [113] e.m. materón, a. wong, t.a. freitas, r.c. faria, o.n. oliveira. a sensitive electrochemical detection of metronidazole in synthetic serum and urine samples using low-cost screen-printed electrodes modified with reduced graphene oxide and c60. journal of pharmaceutical analysis 11 (2021) 646652. https://doi.org/10.1016/j.jpha.2021.03.004 [114] s. kurbanoglu, s.a. ozkan. electrochemical carbon based nanosensors : a promising tool in pharmaceutical and biomedical analysis. journal of pharmaceutical and biomedical analysis 147 (2018) 439-457. https://doi.org/10.1016/j.jpba.2017.06.062 [115] p. yáñez-sedeño,s. campuzano, j.m. pingarrón. fullerenes in electrochemical catalytic and affinity biosensing. c journal of carbon research 3 (2017) 21. https://doi.org/10.3390/c3030021 [116] j.h. carvalho, j.l. gogola, m.f. bergamini, l.h. marcolino-junior, b.c. janegitz. disposable and lowcost lab-made screen-printed electrodes for voltammetric determination of l-dopa. sensors and actuators reports 3 (2021) 100056. https://doi.org/10.1016/j.snr.2021.100056 [117] s.m.v. dos santos, p.r. de oliveira, m.c. de oliveira, m.f. bergamini, l.h. marcolino. eletrodos impressos construídos por serigrafia utilizando negro de fumo como material conductor. revista virtual de quimica 9 (2017) 626-640. https://doi.org/10.21577/1984-6835.20170037 [118] h. menon, r. aiswarya, k.p. surendran. screen printable mwcnt inks for printed electronics. royal society of chemistry advances 7 (2017) 44076-44081. https://doi.org/10.1039/c7ra06260e [119] a. gevaerd, e.y. watanabe, c. belli, l.h. marcolino-junior, m.f. bergamini. a complete lab-made point of care device for non-immunological electrochemical determination of cortisol levels in salivary samples. sensors and actuators b: chemical 332 (2021) 129532. https://doi.org/10.1016/ j.snb.2021.129532 [120] t. rungsawang, e. punrat, j. adkins, c. henry, o. chailapakul. development of electrochemical paperbased glucose sensor using cellulose-4-aminophenylboronic acid-modified screen-printed carbon electrode. electroanalysis 28 (2016) 462-468. https://doi.org/10.1002/elan.201500406 [121] h.s. magar, r.y.a. hassan, m.n. abbas. non-enzymatic disposable electrochemical sensors based on cuo/co3o4@mwcnts nanocomposite modified screen-printed electrode for the direct determination of urea. scientific reports 13 (2023) 2034. https://doi.org/10.1038/s41598-023-28930-4 [122] y. wang, j. xie, l. tao, h. tian, s. wang, h. ding. simultaneous electrochemical determination of epirubicin and methotrexate in human blood using a disposable electrode modified with nanoau/mwnts-zno composites. sensors and actuators b 204 (2014) 360-367. https://doi.org/ 10.1016/j.snb.2014.07.099 [123] p.c. motsaathebe, o.e. fayemi. electrochemical detection of ascorbic acid in oranges at mwcnt aonp nanocompositefabricated electrode. nanomaterials 12 (2022) 645. https://doi.org/10.3390/ nano12040645 [124] g. uwaya, o.e. fayemi. electrochemical detection of ascorbic acid in orange on iron(iii) oxide nanoparticles modified screen printed carbon electrode. journal of cluster science 33 (2022) 10351043. https://doi.org/10.1007/s10876-021-02030-7 ©2023 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) https://doi.org/10.5599/admet.1709 https://doi.org/10.1016/j.microc.2019.04.086 https://doi.org/10.1016/j.jpha.2021.03.004 https://doi.org/10.1016/j.jpba.2017.06.062 https://doi.org/10.3390/c3030021 https://doi.org/10.1016/j.snr.2021.100056 https://doi.org/10.21577/1984-6835.20170037 https://doi.org/10.1039/c7ra06260e https://doi.org/10.1016/j.snb.2021.129532 https://doi.org/10.1016/j.snb.2021.129532 https://doi.org/10.1002/elan.201500406 https://doi.org/10.1038/s41598-023-28930-4 https://doi.org/10.1016/j.snb.2014.07.099 https://doi.org/10.1016/j.snb.2014.07.099 https://doi.org/10.3390/nano12040645 https://doi.org/10.3390/nano12040645 https://doi.org/10.1007/s10876-021-02030-7 http://creativecommons.org/licenses/by/3.0/ manuscript doi: 10.5599/admet.3.4.249 352 admet & dmpk 3(4) (2015) 352-358; doi: 10.5599/admet.3.4.249 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper rapid identification of bioactive carbohydrazide reaction products by an lc-dad-spe-nmr approach iva habinovec, tomislav jednačak and predrag novak* university of zagreb, faculty of science, department of chemistry, horvatovac 102a, 10000 zagreb, croatia ihabinovec@chem.pmf.hr; tjednacak@chem.pmf.hr *corresponding author: e-mail: pnovak@chem.pmf.hr; tel.: +385-1-460-6184; fax: +385-1-460-6181 received: november 09, 2015; revised: december 24, 2015; published: december 30, 2015 abstract on-line coupling of high performance liquid chromatography with diode array detection to solid phase extraction combined with nuclear magnetic resonance (lc-dad-spe-nmr) was used to monitor carbohydrazide condensation reaction progress. first, a chromatographic method was developed and optimised and individual peak separation was readily achieved by using an isocratic acetonitrile-phosphate buffer mobile phase. subsequently, separated compounds were trapped on spe cartridges and dried with nitrogen gas. peak elution was then performed with deuterated acetonitrile and sent for nmr analysis. single and multiple trapping options were applied. oneand two-dimensional nmr spectra were recorded using a prodigy cryoprobe. it was demonstrated that lc-dad-spe-nmr setup was proved very useful for rapid and unambiguous identification of the reaction products and for determination of their structure. by using prodigy cryoprobe in nmr measurements we were able to detect and identify compounds present at microgram level thus proving a high sensitivity of this methodology for monitoring reactions of bioactive molecules and drugs. keywords carbohydrazide derivatives; reaction monitoring; lc-dad-spe-nmr; product identification; structure elucidation introduction aromatic imines derived from carbohydrazide are promising therapeutic agents with a broad spectrum of biological properties, such as antiviral, antitubercular, antifungal, cytotoxic and antimalarial activity [14]. furthermore, they were reported to be used as precursors in the synthesis of various nitrogen and sulphur containing heterocycles [5,6]. owing to their ability to act as chelate ligands and stabilise metal ions in different oxidation states, carbohydrazide schiff bases have also found applications as multitopic ligands for the targeted construction of bioactive coordination systems [7,8]. these compounds can easily be prepared by the condensation reaction between carbohydrazide and appropriate aldehyde or ketone. the condensation of carbohydrazide (i) with salicylaldehyde (ii) yielded a mixture of mono(iii) and bis(salicylidene)carbohydrazide (iv), as shown in figure 1 [9,10]. http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:ihabinovec@chem.pmf.hr mailto:tjednacak@chem.pmf.hr mailto:pnovak@chem.pmf.hr admet & dmpk 3(4) (2015) 352-358 carbohydrazide reaction monitoring by lc-dad-spe-nmr doi: 10.5599/admet.3.4.249 353 figure 1. the condensation reaction between carbohydrazide and salicylaldehyde. we have recently reported that in-line raman spectroscopy combined with multivariate data analysis could be applied for in-line monitoring of condensation reaction and for assessing the mixture content [11]. however, several drawbacks of this methodology, such as low sensitivity, complexity of the reaction mixture, and severe band overlapping, prevented precise identification of the compounds in the reaction mixture and determination of the reaction endpoints directly from the in-line spectra. in order to resolve these issues, we employed here an approach based on the hyphenated system comprising of high performance liquid chromatography with diode array detection (lc-dad), solid phase extraction (spe) and nuclear magnetic resonance spectroscopy equipped with a cryogenically cooled probe (cryo nmr). this lcdad-spe-nmr setup represents a state-of-the-art technology for rapid and accurate mixture analysis. the coupling of hplc with nmr spectroscopy has become one of the most efficient and powerful analytical technique to analyse complex mixtures. identification and structural characterization of compounds in a mixture without their physical separation has long been a dream of analytical chemists. lc-nmr hyphenation gives a wealth of structural information on small quantities of substances without the necessity of their isolation and purification [12-16]. a further coupling of lc-nmr to spe has considerably improved the sensitivity of this technique by post column peak trapping onto the spe cartridges enabling the separate compounds to be concentrated and eluted fully with deuterated solvents. lc-spe-nmr is now widely accepted tool for dereplication and discovery of natural products from complex mixtures of plants, metabolites, drug impurity profiling and environmental analysis [17-21]. on the other hand, it has been less frequently used to monitor chemical reaction products [22,23]. hence, in the present work we applied this hyphenated method to analyse carbohydrazide condensation reaction. oneand two-dimensional nmr experiments with a noesy-type double presaturation for suppressing acetonitirile and water resonances will be used for structure elucidation of mixture components. experimental chemicals salicylaldehyde (min. 98 %) and carbohydrazide (min. 98 %) were purchased from sigma-aldrich (mo, usa). piperidine (min. 99 %) was purchased from acros organics (nj, usa). methanol (min. 99.95 %) was obtained from carl roth gmbh (karlsruhe, germany). acetonitrile (hplc grade) was purchased from j. t. baker (pa, usa) and potassium dihydrogen phosphate (for chromatography) was obtained from kemika habinovec et al. admet & dmpk 3(4) (2015) 352-358 354 (zagreb, croatia). acetonitrile-d3 (d, 99.8 %) was purchased from cambridge isotope laboratories, inc. (ma, usa). synthesis carbohydrazide (0.60 g, 6.67 mmol) was gradually dissolved in 40 ml of methanol. after the addition of salicylaldehyde (0.71 ml, 6.67 mmol) and a few drops of piperidine, the reaction mixture was heated at 55 °c and stirred under reflux for 40 min [9]. during heating, a white crystalline product started to precipitate. the product was filtered off, washed with a small amount of cold methanol and dried in air. lc-spe analysis an agilent 1260 infinity hplc system equipped with quaternary pump, an autosampler, thermostatted column compartment and dad detector was used for chromatographic separation prior to post-column solid-phase extraction. separation was performed on an agilent poroshell 120 c18 column (250 mm x 4.6 mm) with 4 μm particle size in analytical mode. injection volume was 25 μl and the flow rate was 1.0 ml min –1 . elution was isocratic using a mobile phase consisting of 50 % acetonitrile and 50 % phosphate buffer (ph 7) at 298 k over 10 min. the automatic post-column spe was performed on a prospekt 2 lcspe-nmr interface (spark holland, emmen, the netherlands). both single and multi-trapping methods were used for the spe pre-concentration. compounds from reaction mixture were trapped on hysphere resin gp, c18 and c8 spe cartridges (10 x 2 mm). to increase the retention of the compounds on spe cartridges a post-column water addition was performed with a flow rate of 3.0 ml min –1 using the knauer k-120 hplc pump (knauer, berlin, germany). after the cartridges were dried with nitrogen gas for 59 min, compounds were eluted with 336 μl of deuterated acetonitrile into the 3 mm nmr tubes. nmr spectroscopy nmr spectra with noesy-type solvent suppression module were recorded on bruker avance iii hd 400 spectrometer operating at 400 mhz equipped with a broadband observed (bbo) 5 mm prodigy cryoprobe and z-gradient accessories. after lc separation and peak trapping on the spe unit, the separated components were dissolved in acetonitrile-d3 and measured in 3 mm tubes with match inserts for adapting the tubes to 5 mm bruker spinners. all experiments were carried out at 298 k using tms as an internal standard. proton spectra with spectral width of 6002 hz and a digital resolution of 0.36 hz per point were measured with 64 and 128 scans. in the gcosy experiment, 2048 points in the f2 dimension and 128 increments in the f1 dimension were used. for each increment, 16 scans and the spectral width of 6010 hz were applied. digital resolution was 5.87 and 93.78 hz per point in f2 and f1 dimensions, respectively. the ghsqc and ghmbc spectra were acquired with 64 and 128 scans, respectively. spectral width was 6009 hz in f2 and 20124 hz in f1 dimension for both experiments. 1k data points were applied in the time domains and for each data set 256 increments were collected. the resulting digital resolution was 5.86 hz per point in f2 dimension and 314.4 hz per point in f1 dimension. results and discussion the condensation reaction of carbohydrazide and salicylaldehyde was performed in methanol and the reaction products were subsequently analysed by lc-dad-spe-nmr method. prior to lc-dad-spe-nmr analysis, a chromatographic method was developed and optimised. chromatographic separation of the individual products and reactants was readily achieved by using an isocratic acetonitrile-phosphate buffer admet & dmpk 3(4) (2015) 352-358 carbohydrazide reaction monitoring by lc-dad-spe-nmr doi: 10.5599/admet.3.4.249 355 mobile phase. to identify and structurally characterise reaction components, we performed lc-dad-spenmr analysis. two different approaches were used. in the first one, we filtered the reaction mixture and dissolved the crude products iii and iv together with reactants in acetonitrile and injected 25 µl of the solution into the lc-spe system. the lc-dad chromatogram of this standard mixture is presented in figure 2a. in the second approach, we took aliquots directly from the reaction mixture (20 µl), added 480 μl of acetonitrile and then performed the analysis by injecting 25 µl of the solution. the chromatogram is shown in figure 2b. figure 2. lc-dad chromatogram of (a) the standard reaction mixture, (b) the reaction mixture aliquot collected 5 min after the reaction start. the peaks designated with asterisks are unknown compounds. aliquots were collected at the beginning of the reaction and after 3, 5, 10, 25 and 40 minutes. it is seen in the figure 2 that two products were formed, iii and iv with the retention times of 2.6 and 5.9 minutes, respectively. in chromatograms collected during the reaction course those peaks gradually increased while i and ii decreased and vanished in the interval between 5 and 10 minutes thus indicating the end of the habinovec et al. admet & dmpk 3(4) (2015) 352-358 356 reaction. this finding is somewhat different from that previously observed by in-line vibrational spectroscopy and statistical methods [11] where the reaction endpoint in methanol was estimated to be around 30 minutes. according to lc area peak integrals condensation reactions gave 48 % of iii and 52 % of iv which was similar to previous results. a close inspection of chromatogram depicted in figure 2b reveals two additional peaks with retention times around 4 minutes. unfortunately, those couldn’t be trapped on the cartridges used in this study and further method development is needed to resolve their identity and structure. to identify and structurally characterise the reaction products iii and iv, the contents of their lc peaks were sent to post column spe trapping for pre-concentration prior to nmr analysis in order to achieve a better signal-to-noise ratio. subsequently, the spe cartridges were dried with nitrogen gas before peak elution. deuterated acetonitrile was used to elute and transfer the peaks from the cartridges to 3 mm nmr tubes. to record one-dimensional proton and two-dimensional cosy spectra, a single trapping was sufficient for the compound iv (figures 3 and 4), while multiple trapping option was used to record heteronuclear two-dimensional spectra (hsqc and hmbc). figure 3. an expanded region of 1 h nmr spectrum of iv obtained in lc-spe-nmr mode for: (a) standard reaction mixture (initial mass 1.95 μg, 64 scans, 6 min) (b) reaction mixture aliquot collected after 5 min (64 scans, 6 min) and (c) (128 scans, 12 min). it is noteworthy to say that the initial amount of the compound iv to record the proton spectrum displayed in figure 3a was only 1.95 μg. the concentration of iv was even lower in the samples collected directly from the reaction mixture and satisfactory proton spectrum was obtained with 128 scans (figure 3c). furthermore, very good signal-to-noise ratio was observed also in the cosy spectrum of iv depicted in figure 4 with the initial amount of only 24 μg. this clearly demonstrates the advantage of using lc-spenmr setup with prodigy cryoprobe for detection and identification of low concentrated samples. the structure of the compounds was unambiguously confirmed by the analysis of the recorded one( 1 h) and two-dimensional nmr spectra (cosy, hsqc and hmbc). admet & dmpk 3(4) (2015) 352-358 carbohydrazide reaction monitoring by lc-dad-spe-nmr doi: 10.5599/admet.3.4.249 357 figure 4. cosy nmr spectrum obtained in lc-spe-nmr mode for iv isolated from the reaction mixture (16 scans, 1h and 16 min). conclusions condensation reaction of carbohydrazide with salicylaldehyde was successfully monitored by lc-dadspe-nmr with prodigy cryoprobe. application of this method made it possible to identify and structurally characterise reaction products in a fast and efficient manner. after the reaction mixture components were separated on the chromatographic column and trapped on spe cartridges, they were sent to nmr. one and two-dimensional nmr spectra were recorded and analysed. spectral analysis confirmed the structures of the reaction products, e.g. mono(iii) and bis(salicylidene)carbohydrazide (iv). references [1] s. barman, l. you, r. chen, v. codrea, g. kago, r. edupuganti, j. robertus, r.m. krug, e. v. anslyn, european journal of medicinal chemistry 71 (2014) 81-90. [2] v.n. telvekar, a. belubbi, v.k. bairwa, k. satardekar, bioorganic & medicinal chemistry letters 22 (2012) 2343-2346. [3] t. aboul-fadl, f.a.s. bin-jubair, o. aboul-wafa, european journal of medicinal chemistry 45 (2010) 4578-4586. [4] j. camacho, a. barazarte, n. gamboa, j. rodrigues, r. rojas, a. vaisberg, r. gilman, j. charris, bioorganic & medicinal chemistry 19 (2011) 2023-2029. [5] a.h. corvin, j.d. reinheimer, journal of american chemical society 73(3) (1951) 1184-1186. [6] b.f. abdel-wahab, a.-a. s. el-ahl, phosphorus, sulfur, and silicon, 185 (2010) 249-260. [7] m. sutradhar, t. r. barman, e. rentschler, inorganic chemistry communications 39 (2014) 140-143. [8] r. n. patel, inorganica chimica acta 363 (2010) 3838-3846. [9] p. novak, t. jednačak, j. parlov vuković, k. zangger, m. rubčić, n. galić, t. hrenar, croatica chemica acta 85 (4) (2012) 451-456. habinovec et al. admet & dmpk 3(4) (2015) 352-358 358 [10] m. rubčić, n. galić, i. halasz, t. jednačak, n. judaš, j. plavec, p. šket, p. novak, crystal growth & design 14 (2014) 2900-2912. [11] t. jednačak, p. novak, a. hodzic, o. scheibelhofer, j. g. khinast, j. plavec, p. šket, j. parlov vuković, acta chimica slovenica 61 (2014) 161-169. [12] s. görög, trends in analytical chemistry 69 (2015) 114-122. [13] p. novak, m. cindrić, p. tepeš, s. dragojević, m. ilijaš, k. mihaljević, journal of separation science 28 (2005) 1442-1447. [14] s. singh, t. handa, m. narayanam, a. sahu, m. junwal, r. p. shah, journal of pharmaceutical and biomedical anaysis 69 (2012) 148-173. [15] p. novak, p. tepeš, m. cindrić, m. ilijaš, s. dragojević, k. mihaljević, journal of chromatography a 1033 (2004), 299-303. [16] p. novak, p. tepeš, m. ilijaš, i. fistrić, i. bratoš, a. avdagić, v. gabelica marković, m. dumić, journal of pharmaceutical and biomedical anaysis 50 (2009) 68-72. [17] g. schlotterbeck, s.m. ceccarelli, bioanalysis 1 (2009) 549-559. [18] s. sturm, c. seger, journal of chromatography a 1259 (2012) 50-61. [19] j. w. jaroszewski, planta medica 71 (2005) 795-802. [20] s. schmidt, c. piechotta, m. godejohann, i. nihels, c. mügge, talanta, 82 (2010) 143-150. [21] s. christophoridou, p. dais, l. h. tseng, m. spraul, journal of agriultural and food chemistry 53 (2005) 4667-4679. [22] c. seger, m. godejohann, m. spraul, h. stuppner, f. hadacek, journal of chromatography a 1136 (2006) 82-88. [23] j. larsen, d. staerk, c. cornett, s. h. hansen, j. w. jaroszewski, journal of pharmaceutical and biomedical anaysis 49 (2009) 839–842. ©2015 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ mechanistically transparent models for predicting aqueous solu¬bility of rigid, slightly flexible, and very flexible drugs (mw<2000) accuracy near that of random forest regression doi: https://doi.org/10.5599/admet.1897 1 admet & dmpk 0(0) (2023) 000-000; doi: https://doi.org/10.5599/admet.1879 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index review mechanistically transparent models for predicting aqueous solubility of rigid, slightly flexible, and very flexible drugs (mw<2000) accuracy near that of random forest regression alex avdeef in-adme research, new york, ny 10128 usa *corresponding author: e-mail: alex@in-adme.com; tel.: +1-646-678-5713 received: may 8, 2023; revised: august 15, 2023; published: august xx, 2023 abstract yalkowsky’s general solubility equation (gse), with its three fixed constants, is popular and easy to apply, but is not very accurate for polar, zwitterionic, or flexible molecules. this review examines the findings of a series of studies, where we have sought to come up with a better prediction model, by comparing the performances of the gse to abraham’s solvation equation (absolv), and random forest regression (rfr) machine-learning (ml) method. large, well-curated aqueous intrinsic solubility databases are available. however, drugs may be sparsely distributed in chemical space, concentrated in clusters. even a large database might overlook some regions. test compounds from under-represented portions of space may be poorly predicted, as might be the case with the ‘loose’ set of 32 drugs in the second solubility challenge (2020). there appears to be still a need for better coverage of drug space. increasingly, current trends in predictions of solubility use calculated input descriptors, which may be an advantage for exploring properties of molecules yet to be synthesized. the risk may be that overall prediction approaches might be based on accumulated uncertainty. the increasing use of ml/ai methods can lead to accurate predictions, but such predictions may not readily suggest the strategies to pursue in selecting yet-to-be-synthesized compounds. based on our latest findings, we recommend predictions based on both ‘grouped’ absolv(grp) and ‘flexible acceptor’ gse(φ,b) models with the provided best-fit parameters, where φ is the kier molecular flexibility index and b is the abraham h-bond acceptor strength. for molecules with φ < 11, the prudent choice is to pick the consensus model, the average of absolv(grp) and gse(φ,b). for more flexible molecules, gse(φ,b) is recommended. ©2023 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords general solubility equation (gse); abraham solvation equation (absolv); flexible-acceptor gse(φ,b); consensus model; decision-tree exclusive or (xor) model; kier molecular flexibility index (φ); drug-like molecules; machine learning (ml); intrinsic solubility introduction the pursuit of accurate prediction of solubility of drugs from molecular structure is still evolving and continues to be challenging [1-7]. it had been proposed that shortfalls have been due to the lack of high-quality solubility data from the chemical space of drugs. but there has been some pushback to that view [6]. since 2011, we have been actively collating and harmonizing published values of ph-dependent aqueous solubility of https://doi.org/10.5599/admet.1897 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:alex@in-adme.com http://creativecommons.org/licenses/by/4.0/ a. avdeef admet & dmpk 00(0) (2023) 000-000 2 drug-like and drug-relevant molecules of importance in the discovery-to-development stages of pharmaceutical research, aggregated in the wiki-ps0tm database (in-adme research). a book tentatively entitled “predicting solubility of new drugs handbook of critically curated data for pharmaceutical research” is under review by a publisher [7]. it tabulates solubility data up to 2022. the collection comprises 3695 different substances, with 7619 entries of intrinsic solubility (uncharged form). considerable effort has been put into deciding that the data are of high quality, based on carefully selected published sources, guided in part by the best-practices ‘white papers’ recently published [8-12]. just about all entries in the database are referenced to primary sources. the reproducibility of statistical methods to predict solubility at best has hovered around the root mean square error (rmse) of 0.6 log unit but is typically rmse > 1 in many studies [1-5]. on the other hand, it has been firmly estimated that the average interlaboratory reproducibility can be as low as 0.18 log unit in carefully curated databases, which includes correcting reported solubility for ionization (i.e. deriving intrinsic solubility, s0) and by normalizing for temperature (by transforming measurements performed in the range 10-50 to 25 °c) [7,13-20]. consistent solubility unit conversions, methods of phase separation, and procedures for measuring ph also play critical underpinnings in data quality [8-12]. this review discusses a recent series of five interrelated publications [16-20], where the use of several computational methods to predict intrinsic solubility were explored: (a) yalkowsky general solubility equation (gse) [21-26], (b) abraham solvation equation (absolv) [27-29], (c) avdeef-kansy flexible-acceptor general solubility equation (a.k.a., gse(φ,b)) [18], (d) consensus of absolv and gse(φ,b) [19], and (e) breiman random forest regression (rfr) statistical machine learning (ml) method [30-33]. the above data-driven methods were trained with the wiki-ps0 database. the traditional gse is often considered pre-trained. it is popular for its simplicity and ease of use. new prediction methods are often benchmarked against the gse. this review concludes with the introduction of a new variant method induced from the above five studies, called the exclusive or (‘xor’) decision tree model, drawing on absolv and/or gse(φ,b) models, depending on the value of φ. the method may be useful when a large diverse database of intrinsic solubility values of drug-like or drug-relevant molecules is available. it mirrors the consensus model [19]. analytic continuity of methods to predict solubility of drug-like molecules the solubility (intrinsic, log molarity units) calculated by gse depends on the value of the octanol-water partition coefficient (measured log p or calculated clog p) and the measured (or calculated) melting point (mp / °c) of the molecule. no further training is required for this thermodynamically well-founded legacy equation. log s0gse = 0.5 log p -0.01 (mp-25) (1) the abraham and le [28] absolv equation to predict solubility takes the form: log s0absolv = a0 + a1a + a2b + a3sπ + a4e + a5v + a6ab (2) in the multiple linear regression (mlr) equation, the log s0 is the dependent variable (measured log intrinsic molar solubility) and the independent variables are the five solute solvation descriptors accounting for the energy of transfer of solute from solid to the solution phase: a is the sum of h-bond acidity (donor potential), b is the sum of h-bond basicity (acceptor potential), sπ is the dipolarity/polarizability (subscripted here, so as not to be confused with solubility), e is an excess molar refractivity in units of (cm3∙mol-1)/10, and v is the mcgowan characteristic molar volume in units of (cm3∙mol-1)/100. the a0 a6 constants in eq. (2) are determined by regression based on the training database of intrinsic solubility values. the five solvation descriptors may be calculated from 2d structure (provided in smiles notation or as coordinates in a ‘mol’ file) using the program absolv [29] (cf. www.acdlabs.com). http://www.acdlabs.com/ admet & dmpk 00(0) (2023) 000-000 mechanistically transparent models for predicting aqueous solubility doi: https://doi.org/10.5599/admet.1897 3 as detailed elsewhere [18-20], the flexible-acceptor model, gse(φ,b), was developed as a critical expansion of the legacy gse, following an exhaustive search of descriptors in a principal components analysis. for the original three constants (0.5, -1.0, -0.01) in eq. (1), the first two were fitted to an exponential function of the sum of two descriptors, φ (kier’s [34] molecular flexibility index) and b (abraham’s [27-29] h-bond acceptor potential). the third crystal-lattice contribution term was best characterized as a linear function of φ+b. log s0gse(φ,b) = c0 + c1· clog p + c2· (mp-25)/100 = {b0 + b1·exp(b2(φ+b))} + + {b3 + b4(1 exp(b5(φ+b))} clog p + {b6 + b7·(φ+b)}(mp-25)/100 (3) the molecular flexibility descriptor, φ, has been defined [34] in terms of structural attributes (chains, rings, branches, atom counts) as φ = 1κ· 2κ /nha, where 1κ and 2κ are the firstand second-order topological shape indices, and nha is the heavy atom count in the molecule (rdkit descriptors from http://www.rdkit.org). in the wiki-ps0 database, φ values ranged from 0.4 to 43. for molecules with molecular weight > 500 da, the 630 φ values range from 4.9 to 43, averaging 12. b (cf. eq. 2) for the large molecules ranges from 0.4 to 12.9, averaging 3.6. in the development of the sge(φ,b) model, several combinations of φ and single abraham descriptors were examined. the sum of φ and b improved performance over just φ in eq. (3) [18-20]. the eight b0-b7 constants in the gse(φ,b) model were determined in two steps. (i) the log s0 values in the training set were sorted on φ+b. the sorted data were then divided into 20 bins, each of near-equal values of φ+b. seven of the lowest-(φ+b) bins contained an average of nearly 1000 log s0 values per bin. the data in each bin were then analyzed to find the best-fit c0, c1 and c2 constants in the first line of eq. (3), using pls regression. (ii) the resultant 20 c0, c1, and c2 constants from the bins were then displayed in three plots against (φ,b) values, to suggest possible nonlinear fitting equations. from this, the analytical expressions for the c-coefficients were determined by standard nonlinear least-squares methods, as detailed elsewhere [18-20]. of the new ml statistical approaches, the rfr method is thought to be among the premier performers in prediction accuracy, although deep-learning neural network methods may be as good [3,4]. the rfr software is freely downloadable and is easy to use (cf. ’random forest’ library for the r statistical software [31-33]). the method works by constructing an ensemble of hundreds of decision trees based on a random selection of a portion of the training set of solubility measurements, using hundreds of randomly-grouped descriptors provided by the user. it is not possible to state a simple explicit equation, like eqs. (1)-(3), for the rfr method. in this study, all metrics expressed as the coefficient of determinations (r2) and root-mean square errors (rmse) are of the ‘validation type’ [35], unless otherwise indicated (see abbreviations and definitions). also, all logarithm functions are with reference to base 10. random forest regression trained with the wiki-ps0 database in the first of our recent solubility prediction studies [16], four test sets were examined: • test set 1 – yalkowsky-banerjee [22] set of 21 organic nonelectrolytes (6 solid and 3 liquid poorly-soluble pesticides, 11 older drugs, and a dye/laxative molecule), with the test set mean log s0 = -3.85 (log molarity). this set has been widely tested by other investigators, to assess the effectiveness of prediction models. • test set 2 – hopfinger et al. [2] set of 28 well-studied drugs (all ionizable), with mean log s0 = -4.03. this test was part of the first solubility challenge [1,2]. • test set 3 – llinas-avdeef [3] high-consensus ‘tight’ set of 100 drugs, with the mean log s0 = -4.03. this test was part of the second solubility challenge [3,4]. • test set 4 – llinas-avdeef [3] low-consensus ‘loose set’ set 32 drugs, with the mean log s0 = -5.24. this test was also used in the second solubility challenge [3,4]. it was found that all the predictions generated negative bias (-0.08 to -0.65), in many cases suggesting that solubility of low-soluble compounds was overestimated. rfr outperformed the gse and absolv methods on https://doi.org/10.5599/admet.1897 http://www.rdkit.org/ a. avdeef admet & dmpk 00(0) (2023) 000-000 4 all common metrics [35], averaged over the four test sets: r2avg = (0.69, 0.47, 0.42), rmseavg = (0.92, 1.17, 1.17), mppavg (measure of prediction performance [2]) = (51, 38 and 32 %), for rfr, absolv and gse, respectively. it is interesting that the rfr method predicted yalkowsky-banerjee set 1 only marginally well: r2 = 0.82, rmse = 0.83, mpp = 57 %, bias = -0.23. when 115 practically-insoluble agrochemicals were added to the wikips0 database (which had been devoid of such pesticides), the performance improved: r2 = 0.89, rmse = 0.63, mpp = 71 %, bias = +0.02. this confirmed the importance of matching the chemical space of test molecules with those in the training set. however, the ability to predict the solubility of drugs to the level of the quality of measured data remained out of reach (rmse 0.6 vs. 0.2). improvements in the methods (e.g. more effective descriptors) [6] and better coverage of the clustered [36] chemical space of drugs were called for. the principal component analysis (pca) based on the 30-most important descriptors identified in rfr yielded a scores plot of the first two principal components for the training set solubility, showing a very dense symmetrical distribution about the origin for molecules with molecular weights (mw<500 da). large molecules (mw>800 da) were sparsely (but near diagonally) represented in the bottom-right quadrant, giving the overall distribution a ‘comet-like’ head-tail appearance [16]. small molecules in the lipinski [36] rule of 5 (ro5) chemical space, populating the ‘head,’ thus appeared to have very different distributions than those beyond the ro5 (bro5) space in the ‘tail.’ small ro5 molecules can predict the solubility of large bro5 molecules, using the rfr method in our second study [17], the above ‘comet-like’ pca distribution enticed us to explore whether small molecules (ro5 space) in a training set could be used to predict the solubility of large molecules (bro5 space), although at first, we were not optimistic of a good outcome. the molecules with mw>800 da were selected as the test set, with the rest of the database molecules serving as the training set. • test set 5 – avdeef-kansy [17] test set of 31 ‘big’ drug-like/drug molecules with mw 802 1882 da (e.g. cyclosporine a, gramicidin a, leuprolide, nafarelin, oxytocin, vancomycin), with mean log s0 = -4.52. it was found the rfr method spectacularly distinguished itself from the simpler gse and absolv models: r2 = (0.37, -5.24, -3.82), rmse = (1.07, 3.36, 2.95), mpp = (42, 10 and 16 %), bias = (+0.30, +2.64, +0.16), for rfr, absolv and gse, respectively. in probing this further, the first two normally-fixed parameters in the gse(classic) model (+0.5 and -1.0) were ‘re-trained’, to yield log s0gsesmall = -0.28 -0.83 clogp -0.01·(mp 25). the modified versions of the yalkowsky’s equation used as the training set produced metrics comparable to those of the rfr method: r2 = 0.33, rmse = 1.10, mpp = 30 %, and bias = +0.04. a similar treatment for the large molecules yielded log s0gsebig = -1.77 -0.40 clogp -0.01·(mp 25). these modifications showed that c0 and c1 could be variable coefficients, rather than constants. it was this observation that pointed the way to the development of the flexible-acceptor model, gse(φ,b), which turned out to be quite an improvement over the classic equation, providing realistic coverage of the chemical space of drugs up to mw of about 2000 da. flexible-acceptor gse(φ,b) for predicting the solubility of large molecules when it became evident that the gse might be beneficially re-trained to predict a broader class of compounds, especially large drug-like molecules, a search was made for analytical functions which could represent the three traditional constants in terms of meaningful property descriptors, guided by the pca distribution mentioned above [16]. the novel flexible-acceptor model (cf. eq. (3)), gse(φ,b), was the outcome of the effort [18]. as the last section indicated, it had been a challenge to predict solubility of large molecules using simple but easily transparent and interpretable models such as gse and absolv (as indicated by r2 < 0 and rmse ≥ 3). rfr showed promise, but this ml method can be opaque, given that over 200 calculated descriptors are often used admet & dmpk 00(0) (2023) 000-000 mechanistically transparent models for predicting aqueous solubility doi: https://doi.org/10.5599/admet.1897 5 in the random mixing of portions of the learning set with randomly-selected subsets of descriptors. nevertheless, there are good reasons to attempt the prediction of the solubility of large molecules. many drugs (mostly derived from natural products) in immunosuppression, oncology, and for the treatment of infectious/viral diseases are large, lipophilic, and possess many h-bond acceptors. literature discussions highlight the promise of therapeutic opportunities for ‘beyond the rule of 5’ (bro5) molecules [37-44]. caron and colleagues [42-44] have paved the way for recognizing the importance of kier’s molecular flexibility index, φ, in predicting the physical properties of bro5 molecules. flexible molecules with the potential to form intramolecular h-bonds may possess enhanced solubility in polar media (i.e. water), by adopting hydrophilic ‘extended’ conformations, as well as enhanced permeability across apolar cell membranes, by adopting hydrophobic ‘folded’ conformations [37-44]. given that large molecules may pose pharmacokinetic (pk) risks due to low solubility, the need for caution is especially important. so, reliable, and actionable in silico models to predict solubility before such molecules are prioritized for synthesis could be a valuable contribution in pk risk assessment. • test set 6 – avdeef-kansy [18] – one additional molecule was added to set 5 to comprise test set 6 of 32 drugs with mw from 802 to 1882 da, mwavg = 1037 da. average log s0 = -4.61 (range -1.2 to -7.6); clogpavg = 3.3 (-3.6 to +17.9); bavg = 5.8 (1.9 11.6); φavg = 20 (11-41). using all the database molecules with mw < 800 da as the training set, the gse(φ,b) prediction of larger molecules yielded promising statistics: r2 = 0.40, rmse = 1.10, mpp = 41 %, bias = -0.08. by contrast, the traditional gse generated r2 < 0, rmse = 3.0, and mpp = 16 %. rfr, the ‘gold standard’ of accuracy in the minds of some computational chemists, generated r2 = 0.37, rmse = 1.07, mpp = 38 %, bias = +0.30. overall, rfr and gse(φ,b) performances were about the same. this encouraging result became the segway to our study of a more diverse class of test molecules: recently fda-approved drugs (2016-2022). gse(φ,b) works well for both big and small molecules, but the consensus model, based on the average of gse(φ,b) and absolv(grp) can be even slightly better than the rfr model the pharma r&d productivity trended downwards from a high point in 1996 to leveling off by 2010, judging by the count of new molecular entities (nmes) approved each year [45]. from 2011 to 2020, an upward recovery trend, albeit bumpy, can be discerned. more recently, a trend reversal may be taking place. of the drugs approved in 2020 and 2021, 72 % are considered ‘small molecule’ nmes. but even these are getting larger, less soluble, more lipophilic, and possessing more h-bond acceptors, when compared to older drugs. large molecules may be burdened with pk risks, as noted above. in our fourth [19] and fifth [20] studies, we directed our efforts to predict the solubility of newly-approved drugs, covering the period of 2016-2021. a few newly-approved drugs were added from 2022 [7]. it was of particular interest to see how the flexible-acceptor model, gse(φ,b), would perform, since many of the new drugs are large and draw on a diverse chemical space. in addition to predictions, the trends in physicochemical properties of these new drugs were quantitated [20] – property inflation was evident. to reduce method bias in absolv prediction, the training set data were divided into six groups. the new operational variant [20], called absolv(grp), is described below. ● test set 7 – the intrinsic solubility of 105 newly fda-approved drugs (2016 to 2022) were added to the wiki-ps0 database [7]. average log s0 = -4.64 (ranging -8.5 to +0.6); mwavg = 465 da (174 1215), clogpavg = 3.3 (-5.8 to +8.7); bavg = 2.2 (0.8 7.7); φavg = 6.7 (1.9 32.0). the gse(φ,b) was originally developed to predict the solubility of large flexible drug-like molecules. it was shown to predict the solubility of drugs beyond lipinski’s ‘rule of 5’ chemical space (bro5) to a precision matching that of the random forest regression (rfr) machine learning method [18]. surprisingly, the https://doi.org/10.5599/admet.1897 a. avdeef admet & dmpk 00(0) (2023) 000-000 6 gse(φ,b) appeared to work well also for ro5 drugs [19]. as before, to add context to the gse(φ,b) model, gse(classic), absolv(grp), and rfr models were also applied to predict log s0 of the newly-approve ‘small molecule’ nmes, for which useable reported solubility values could be accessed (the majority from fda new drug application published reports). the prediction models were retrained with an enlarged version of the wiki-ps0 database. gse(classic) was applied in its traditional form. rfr and gse(φ,b) outperformed the gse(classic) and absolv(grp) models in most of the metrics [35]: r2 = (0.64, 0.59, 0.47, 0.41), rmse = (1.09, 1.15, 1.33, 1.40), mpp = (37, 32, 29 and 31 %), and bias = (-0.11, -0.34, +0.01, -0.33) for rfr, gse(φ,b), absolv(grp), and gse(classic), respectively, for the 105 new drugs. the consensus model [19] (average of gse(φ,b) and absolv(grp)), performed just about as well as the rfr model, with the metrics: r2 = 0.63, rmse = 1.10, mpp = 34 %, and bias = -0.16. the near zero bias of the absolv(grp) model in the most recent studies was largely achieved by dividing the training set into six sub-classes. big molecules (mw > 800 da) and quaternary ammonium compounds were first removed from the training set and were each analyzed separately, with the a0-a7 coefficients in eq. (2) determined by pls regression. the remaining training set molecules were divided into four classes, based on their net charge at ph 7.4 : acids (-), bases (+), zwitterions (±), and neutrals (0). table 1 summarizes the results of the sub-class training. table 1. absolv(grp) coefficients determined by pls regression analysis of training seta log s0absolv = a0 + a1a + a2b + a3sπ + a4e + a5v + a6ab group a0 a1 a2 a3 a4 a5 a6 pearson r2 rmse n acids -0.30 0.26 1.07 0.04 -0.91 -1.80 0.43 0.65 1.17 1578 bases -0.39 -0.61 1.95 0.25 -1.67 -1.37 0.09 0.64 1.11 945 zwitterions 1.52 -1.44 0.88 -0.23 -1.02 -1.08 0.38 0.71 0.88 641 neutrals -0.45 -0.18 1.73 0.10 -1.48 -1.36 0.26 0.63 1.12 4246 big (mw > 800 da) -3.76 0.72 0.61 -0.03 -0.47 -0.38 -0.02 0.42 1.04 93 quaternaries 0.53 -1.23 0.74 -0.04 -0.62 -0.49 0.09 0.66 0.94 39 aa0-a6 coefficients in eq. (2). pls open-source package from https://cran.r-project.org/web/packages/pls. the gse(φ,b) model was further re-trained, as more data had been added to the wiki-ps0 database. for this step, the solubility data in the training set were sorted on φ+b and grouped into twenty bins of increasing values [19]. in each φ+b bin of about 700 solubility entries, the three gse coefficients, c0-c2 in eq. (3) were each determined by linear pls regression. the resultant three groups of c-coefficients showed recognizable forms as a function of φ+b. the trend in c0 was characteristic of a decreasing exponential function in φ+b, suggesting that the solubility of a liquid solute in model lipid (octanol) [16-18] decreases as φ+b increases. the trend in the c1 coefficients (lipophilicity factors) was that of an increasing exponential function, indicating a decreasing influence due to lipophilicity as φ+b increases. the slightly increasing c2 coefficients (crystal lattice effect) could be fit to an ascending linear form as a function of φ+b. apparently, crystal lattice contributions are not appreciably altered by molecular flexibility and h-bond acceptor character, and trend near the traditional value (-0.01) in eq. (1). evidently, solubility dependence on flexibility and h-bond acceptor strength are mediated by solution-phase interactions [46]. table 2 summarizes the most-recently trained gse(φ,b) parameters. table 2. flexible-acceptor model, gse(φ,b)a log s0gse(φ,b) = c0 + c1 clog p + c2 (mp-25)/100 coefficients pearson r2 rmse number of bins c0 = –4.456 + 6.049 e–0.0817(φ,b) 0.93 0.41 20 c1 = –1.326 + 1.058 (1 e–0.1226(φ,b) ) 0.89 0.09 20 c2 = –0.941 + 0.0389(φ,b) 0.55 0.27 20 ab0-b7 coefficients in eq. (3) determined for the wiki-ps0 database, using pls regression: open-source package from https://cran.r-project.org/web/packages/pls. https://cran.r-project.org/web/packages/pls https://cran.r-project.org/web/packages/pls admet & dmpk 00(0) (2023) 000-000 mechanistically transparent models for predicting aqueous solubility doi: https://doi.org/10.5599/admet.1897 7 the consensus prediction equation is simple in form and can be easily incorporated into spreadsheet calculations (using the parameters in tables 1 and 2), which is not the case for the rfr model. aside from opaqueness, there are other limitations to rfr. since the prediction is the average log s0 value of several training set molecules with descriptors most like those of the test compound, the rfr model, as it is currently implemented, cannot extrapolate beyond its training space. so, for molecules much less soluble than those in the training set, the prediction always overestimates the solubility. consequently, if a test molecule is also inadvertently included in the training set, rfr will very likely present the experimental value as the prediction. consensus vs. exclusive or (xor) models in prediction of the solubility of newly-approved drugs [19,20], the rfr and gse(φ,b) models outperformed the gse(classic) and absolv(grp) models in most of the metrics, as noted above. as a bonus, the consensus model based on the average of gse(φ,b) and absolv(grp), slightly outperformed the rfr method in one study [19]. in this last section, we put the rfr aside, after having used it as a valuable benchmark. instead, we focus on developing a decision tree to identify simple mechanistically transparent and easy-to-understand models based mainly on gse(φ,b) and absolv(grp) [5]. as we checked specific molecules over the entire database ad hoc (or in the cases of outliers [5], discussed below), we found that the consensus model was not always the best predictor. for example, we found that in cases of molecules with φ < 1.66 (182 very rigid molecules), the yalkowsky gse(classic) was the best performing model, but just slightly so (cf. table 3, and fig. 2). also, it is uncommon to find an example of a better simple predictor than the gse(φ,b) model for the space beyond φ > 10.83 (342 very flexible molecules) (cf. table 3, and fig. 2). the span between φ 1.66 and 10.83 (comprising 7144 entries– most of the database) revealed flipflop in performance between gse(φ,b) and absolv(grp). either the latter or the former was the best performer, i.e. an ‘exclusive or’ (xor) behavior. how could one justify choosing the individual models over the consensus model? insights to a possible answer to the above question may be revealed by further scrutinizing the performances of gse(classic), absolv(grp) and gse(φ,b) models in the entire wiki-ps0 database. to start, all the entries in the database were sorted on φ into 19 bins of near similar values of φ. in each bin for each of the models, the rmse value was calculated. table 3 shows the bin errors distributions. figure 1 illustrates the distribution of φavg counts in the bins. figure 1. distribution of the bin-average values of the molecular flexibility indices, φavg (table 3). https://doi.org/10.5599/admet.1897 a. avdeef admet & dmpk 00(0) (2023) 000-000 8 figure 2 compares the rmse values distribution of the three models: gse(classic) as red curves with triangle symbols, absolv(grp) as blue dotted curves with square symbols, and gse(φ,b) as green curves with circle symbols. a decision tree constructed based on the crossings in figure 2 is shown in figure 3. since the xor model in figure 3 is dependent on the types of molecules within a particular bin of near constant flexibility, and since different databases may have molecules with differing properties in each range of φ values, it may be prudent to select the consensus model as the ‘best’ prediction. in our case of model training, the cross-over points may be sufficiently grounded to advance the xor model as the ‘best’ for the middle ‘consensus domain’ (cf. zoom view in fig. 2). this is especially so, since this middle domain is comprised of many values of solubility, averaging over 1100 entries/bin for bins 2-7 (table 3). table 3. model selection as a function of kier flexibility, φ φavg in bin number of entries in bin rmse (gse) rmse absolv (grp) rmse gse(φ,b) ‘best’ model 1.0 182 1.42 1.57 1.50 gse(classic) 2.0 1882 1.03 0.987 0.990 absolv(grp) 3.0 1245 1.11 0.99 1.03 absolv(grp) 4.0 1928 1.06 1.00 0.95 gse(φ,b) 5.0 539 1.37 1.04 1.17 absolv(grp) 6.0 953 1.45 1.24 1.14 gse(φ,b) 7.0 122 1.49 1.40 1.19 gse(φ,b) 7.4 100 1.50 1.12 1.23 absolv(grp) 7.8 100 1.68 1.50 1.26 gse(φ,b) 8.3 100 2.10 1.37 1.68 absolv(grp) 9.0 100 2.09 1.60 1.53 gse(φ,b) 9.9 75 2.36 1.29 1.71 absolv(grp) 11.0 75 1.75 1.22 1.15 gse(φ,b) 12.7 75 2.12 1.88 1.64 14.6 50 1.04 1.22 0.93 15.7 50 1.12 1.28 0.91 18.1 50 1.82 0.93 0.85 23.6 25 2.55 2.30 1.01 31.2 17 5.42 1.69 1.53 figure 2. distribution of root-mean-square (rmse) values calculated for each of the bins in table 3, as a function of kier molecular flexibility index, φ. for φ < 1.66, the gse slightly outperforms the other two models (as demarked by the vertical dashed line). for φ > 10.83, the best performing simple model is gse(φ,b) – the flexible-acceptor model fit for flexible molecules. the zoom view on the left side of the drawing illustrates the gse(φ,b)-absolv(grp) flip-flop consensus domain region. admet & dmpk 00(0) (2023) 000-000 mechanistically transparent models for predicting aqueous solubility doi: https://doi.org/10.5599/admet.1897 9 figure 3. decision tree for the selection of the ‘best’ simple method for the prediction of intrinsic solubility. either the consensus model may be sought for 1.95 ≤ φ < 10.83 by taking the average of the two best values, or the xor model may be invokes, by picking best of absolv(grp) and gse(φ,b) models. testing the decision tree model: consensus vs. exclusive or (xor) variants the above decision tree considers the prediction potential from the view of the entire database. however, it is the performance of the predictions of test set molecules that really counts. so, we selected a slightly-increased set of 108 newly-approved fda drugs to examine this last point. • test set 8 – three additional log s0 values were added to test set 7 of newly fda-approved drugs as more approvals were announced in 2022. table 4 summarizes the prediction metrics of the three simple models. based on the properties of the newlyapproved drugs in the test set, the ranking of the various methods is suggested. for of r2 and rmse, the ranking of best performers follows: consensus > gse(φ,b) > xor decision tree > absolv(grp) > gse(classic). the model with the least bias is absolv(grp) and the model with the best mpp (more molecules within ±0.5 log of the correlation identity line) is the xor decision tree (cf. fig. 3). figure 4 illustrates the differences between the xor decision tree and consensus models for the new drug predictions. the general scatter in the consensus model is evidently less than that in the xor model. other minor differences are evident. figure 4. prediction of newly-approved drugs (2016-2022). (a) xor decision tree model predictions, compared to (b) consensus model. https://doi.org/10.5599/admet.1897 a. avdeef admet & dmpk 00(0) (2023) 000-000 10 the summary in table 4 may seem to be complicated, and different trends may be encountered for other test sets. in new test cases, the discussed simple models can be easily incorporated into an excel spreadsheet (using the refined parameters in tables 1 and 2) and compared for new cases. table 4. simple models for predicting solubility of newly-approved drugsa model r2 rmse bias mpp, % consensus 0.64 1.08 -0.15 32.4 gse(φ,b) flexible-acceptor 0.62 1.12 -0.32 33.3 xor decision tree 0.57 1.18 -0.18 35.2 absolv(grp) 0.51 1.26 0.02 32.4 gse(classic) 0.42 1.38 -0.36 31.5 atest set 8 with n = 108. the best metrics are highlighted in bold. another test example in model selection based on outliers oja et al. [5] responded to the second solubility challenge [3,4] with three data-driven mlr models for predicting intrinsic aqueous solubility, which were mechanistically transparent and easily understandable. they discussed the challenges posed by outlier molecules. table 5. simple-model prediction of solubility of outliers in the oja et al. [5] study a drug φ log s0(obs) [7] consensus [5] consensus [this work] xor model gse(φ,b) flexibleacceptor absolv (grp) gse (classic) obs-calc closest prediction folic acid 6.6 -5.91 ± 0.17 -3.88 -2.51 -4.07 -2.97 -2.05 -1.71 -1.84 cisapride 8.6 -6.78 ± 0.17 -4.21 -4.16 -2.97 -4.24 -4.07 -3.71 -2.54 amiodarone 9.2 -10.40 ± 0.50 -7.86 -7.21 -4.38 -6.48 -7.93 -7.75 -2.47 itraconazole 9.6 -8.71 ± 0.57 -7.27 -7.12 -7.93 -5.69 -8.54 -6.48 -0.17 rifabutin 13.1 -3.99 ± 0.43 -6.81 -5.05 -8.54 -5.21 -4.89 -5.63 0.90 cyclosporine a 31.5 -5.03 ± 0.16 -8.27 -4.49 -5.21 -4.38 -4.59 -4.03 -0.44 aclosest predictions are highlighted in bold table 5 shows a different pattern of performance ranking compared to that in table 4. absolv(grp) performed consistently in the first position for the four most flexible molecules. for the least flexible molecules, folic acid and cisapride, the xor and gse(φ,b) looked promising. our consensus model lagged the others in this outliers example. conclusion large, well-curated aqueous intrinsic solubility databases are available, with average interlaboratory reproducibility of less than 0.2 log unit. however, the distribution of drugs in the chemical space might not be uniform but may be sparsely populated in clusters [47]. even a massive database might miss some clusters. test compounds from such an under-represented portion of space may be poorly predicted, as the outliers in table 5 could suggest. also, the 32 ‘difficult-to-predict’ drugs in the second solubility challenge [3,4] may be good examples of underpopulated cluster space, for which better representation is needed. it should also be noted that the 32 molecules are also ‘difficult-to-measure’ drugs. the increasing use of ml/ai methods can lead to accurate predictions, as we have seen. however, these results may not readily suggest the steps to take to improve the properties of tested compounds. our five studies in solubility prediction have attempted to match the performance of the random forest regression method, using relatively simple, mechanistically transparent, and easily applied models [5]. increasingly, current trends in in-silico predictions of solubility use calculated input descriptors, which may be an advantage to explore properties of molecules yet to be synthesized. the risk may be that overall prediction approaches might be based on accumulated uncertainty, something that is often not emphasized [47]. admet & dmpk 00(0) (2023) 000-000 mechanistically transparent models for predicting aqueous solubility doi: https://doi.org/10.5599/admet.1897 11 the knowledge gained and predictive power applied to novel classes of test molecules can still be limited by the calculated descriptors. based on our latest findings, we recommend that both absolv(grp) and gse(φ,b) be calculated (e.g., by taking advantage of the refined parameters in tables 1 and 2). for molecules with φ < 11, the prudent choice is to pick the consensus model, the average of absolv(grp) and gse(φ,b). for more flexible molecules, gse(φ,b) is recommended. abbreviations and definitions s0 intrinsic aqueous solubility (i.e. solubility of the uncharged form of the compound, in molarity units) rmse correlation root-mean-square error: rmsebias = [ 1/(n-1) σi (yiobs bias yicalc)2 ]1/2, where yobs or ycalc is the measured or calculated value of log s0 , n = number of measurements of log s0. note: some statistical programs define rmse slightly differently: see discussion in ref. [35]. ‘validation’ type rmseval = [1/n σi (yiobs yicalc)2]1/2 r2 correlation coefficient of determination, r2bias = 1 σi(yiobs bias yicalc)2 / σi(yiobs <yobs>)2 , where y = log s0, and <yobs> is the mean value of measured log s0. note: some statistical programs define r2 slightly differently. ‘validation’ type [35] r2val = 1 σi(yiobs yicalc)2 / σi(yiobs <yobs>)2 bias intercept (a) in the correlation fit: yobs = a + bycalc, where the slope factor b is fixed at unity. sd standard deviation: sd = [1/n σi(yiobs <y>)2]1/2, where n = number of measurements, <y> = mean value of log s0. mpp measure of prediction performance [2]. it refers to the percent of ‘correct’ predictions, defined as the percentage of log residuals within 0.5 log unit of the identity line. mpp is represented as a pie chart in the correlation plots. abraham solvation descriptors a h-bond total acidity (donor potential) b h-bond total basicity (acceptor potential) sπ dipolarity/polarizability due to solute-solvent interactions between bond dipoles and induced dipoles e excess molar refraction (dm3 mol-1 / 10); which models dispersion force interaction arising from π and n-electrons of the solute v mcgowan molar volume (dm3 mol-1 / 100) a·b acid-base h-bonding product descriptor used in absolv solubility prediction availability of the wiki-ps0 database: a book tentatively entitled: “predicting solubility of new drugs handbook of critically curated data for pharmaceutical research” is under review by a publisher. it is anticipated to be available in april 2024. acknowledgment: collaborations and in-depth discussions of the prediction of drugs with dr. manfred kansy (freiburg, germany) are deeply appreciated. conflict of interest: none references [1] a. llinàs, r.c. glen, j.m. goodman. solubility challenge: can you predict solubilities of 32 molecules using a database of 100 reliable measurements? journal of chemical information and modeling 48 (2008) 1289-1303. https://doi.org/10.1021/ci800058v [2] a.j. hopfinger, e.x. esposito, a. llinàs, r.c. glen, j.m. goodman. findings of the challenge to predict aqueous solubility. journal of chemical information and modeling 49 (2009) 1-5. https://doi.org/10.1021/ci800436c [3] a. llinas, a. avdeef. solubility challenge revisited after ten years, with multi-lab shake-flask data, using tight (sd ∼ 0.17 log) and loose (sd ∼ 0.62 log) test sets. journal of chemical information and modeling 59 (2019) 3036-3040. https://doi.org/10.1021/acs.jcim.9b00345 https://doi.org/10.5599/admet.1897 https://doi.org/10.1021/ci800058v https://doi.org/10.1021/ci800436c https://doi.org/10.1021/ci800436c https://doi.org/10.1021/acs.jcim.9b00345 a. avdeef admet & dmpk 00(0) (2023) 000-000 12 [4] a. llinas, i. oprisiu, a. avdeef. findings of the second challenge to predict aqueous solubility. j. chem. inf. model. 60 (2020) 4791-4803. https://doi.org/10.1021/acs.jcim.0c00701 [5] m. oja, s. sild, g. piir, u. maran. intrinsic aqueous solubility: mechanistically transparent data-driven modeling of drug substances. pharmaceutics 14 (2022) 2248. https://doi.org/10.3390/pharmaceutics14102248 [6] d.s. palmer, j.b.o. mitchell. is experimental data quality the limiting factor in predicting the aqueous solubility of druglike molecules? molecular pharmaceutics 11 (2014) 2962-2972. https://doi.org/10.1021/mp500103r [7] a. avdeef. predicting solubility of new drugs handbook of critically curated data for pharmaceutical research. crc press, boca raton, fl, usa, 2024 isbn: 978-1032617671. https://www.barnesand noble.com/w/predicting-solubility-of-new-drugs-alex-avdeef/1143832638 [8] a. avdeef. suggested improvements for measurement of equilibrium solubility-ph of ionizable drugs. admet & dmpk 3 (2105) 84-109. https://doi.org/10.5599/admet.3.2.193 [9] a. avdeef, e. fuguet, a. llinàs, c. ràfols, e. bosch, g. völgyi, t. verbić, e. boldyreva, k. takács-novák. equilibrium solubility measurement of ionizable drugs – consensus recommendations for improving data quality. admet & dmpk 4 (2016) 117-178. https://doi.org/10.5599/admet.4.2.292 [10] a. veseli, s. žakelj, a. kristl. a review of methods for solubility determination in biopharmaceutical drug characterization. drug development and industrial pharmacy 45 (2019) 1717-1724. https://doi.org/10.1080/03639045.2019.1665062 [11] a. ono, n. matsumura, t. kimoto, y. akiyama, s. funaki, n. tamura, s. hayashi, y. kojima, m. fushimi, h. sudaki, r. aihara, y. haruna, m. jiko, m. iwasaki, t. fujita, k. sugano. harmonizing solubility measurement to lower inter-laboratory variance – progress of consortium of biopharmaceutical tools (cobito) in japan. admet & dmpk 7 (2019) 183-195. http://dx.doi.org/10.5599/admet.704 [12] m. vertzoni, j. alsenz, p. augustijns, a. bauer-brandl, c.a.s. bergström, j. brouwers, a. müllerz, g. perlovich, c. saal, k. sugano, c. reppas. ungap best practice for improving solubility data quality of orally administered drugs. european journal of pharmaceutical sciences 168 (2022) 106043. https://doi.org/10.1016/j.ejps.2021.106043 [13] n. sun, a. avdeef. biorelevant pka (37 oc) predicted from the 2d structure of the molecule and its pka at 25oc. journal of pharmaceutical and biomedical 56 (2011) 173-182. https://doi.org/10.1016/j.jpba.2011.05.007 [14] a. avdeef. solubility temperature dependence predicted from 2d structure. admet & dmpk 3 (2015) 298-344. https://doi.org/10.5599/admet.3.4.259 [15] a. avdeef. multi-lab intrinsic solubility measurement reproducibility in cheqsol and shake-flask methods. admet & dmpk 7 (2019) 210-219. http://dx.doi.org/10.5599/admet.698. [16] a. avdeef. prediction of aqueous intrinsic solubility of druglike molecules using random forest regression trained with wiki-ps0 database. admet & dmpk 8 (2020) 29-77. https://dx.doi.org/10.5599/admet.766 [17] a. avdeef, m. kansy. can small drugs predict the intrinsic aqueous solubility of ‘beyond rule of 5’ big drugs? admet & dmpk 8 (2020) 180–206. https://dx.doi.org/10.5599/admet.794 [18] a. avdeef, m. kansy. ‘flexible-acceptor’ general solubility equation for beyond rule of 5 drugs. mol. pharmaceutics 17 (2021) 3930-3940. https://doi.org/10.1021/acs.molpharmaceut.0c00689. [19] a. avdeef, m. kansy. predicting solubility of newly-approved drugs (2016-2020) with a simple absolv and gse(flexible-acceptor) consensus model outperforming random forest regression. journal of solution chemistry 51 (2022) 1020-1055. https://doi.org/10.1007/s10953-022-01141-7 [20] a. avdeef, m. kansy. trends in physchem properties of newly approved drugs over the last six years, predicting solubility of drugs approved in 2021. journal of solution chemistry 51 (2022) 1455-1481. https://doi.org/10.1007/s10953-022-01199-3. [21] s.h. yalkowsky, s.c. valvani. solubility and partitioning i: solubility of nonelectrolytes in water. journal of pharmaceutical sciences 69 (1980) 912-922. https://doi.org/10.1002/jps.2600690814. https://doi.org/10.1021/acs.jcim.0c00701 https://doi.org/10.3390/pharmaceutics14102248 https://doi.org/10.3390/pharmaceutics14102248 https://doi.org/10.1021/mp500103r https://doi.org/10.1021/mp500103r https://www.barnesandnoble.com/w/predicting-solubility-of-new-drugs-alex-avdeef/1143832638 https://www.barnesandnoble.com/w/predicting-solubility-of-new-drugs-alex-avdeef/1143832638 https://doi.org/10.5599/admet.3.2.193 https://doi.org/10.5599/admet.4.2.292 https://doi.org/10.1080/03639045.2019.1665062 http://dx.doi.org/10.5599/admet.704 https://doi.org/10.1016/j.ejps.2021.106043 https://doi.org/10.1016/j.jpba.2011.05.007 https://doi.org/10.1016/j.jpba.2011.05.007 https://doi.org/10.5599/admet.3.4.259 http://dx.doi.org/10.5599/admet.698 https://dx.doi.org/10.5599/admet.766 https://dx.doi.org/10.5599/admet.766 https://dx.doi.org/10.5599/admet.794 https://doi.org/10.1021/acs.molpharmaceut.0c00689 https://doi.org/10.1007/s10953-022-01141-7 https://doi.org/10.1007/s10953-022-01199-3 https://doi.org/10.1002/jps.2600690814 admet & dmpk 00(0) (2023) 000-000 mechanistically transparent models for predicting aqueous solubility doi: https://doi.org/10.5599/admet.1897 13 [22] s.h. yalkowsky, s. banerjee. aqueous solubility: methods of estimation for organic compounds. marcel dekker, inc.: new york. (1992) p. 142. https://api.semanticscholar.org/corpusid:92804093 [23] y. ran, n, jain, s.h. yalkowsky. prediction of aqueous solubility of organic compounds by the general solubility equation (gse). j. chem. inf. comput. sci. 41 (2001) 1208-1217. https://doi.org/10.1021/ci010287z. [24] n. jain, g. yang, s.g. machatha, s.h. yalkowsky. estimation of the aqueous solubility of weak electrolytes. international journal of pharmaceutics 319 (2006) 169-171. https://doi.org/10.1016/j.ijpharm.2006.04.022 [25] s.h. yalkowsky, y. he, p. jain. handbook of aqueous solubility data, second edition. crc press: boca raton, fl, (2010). https://doi.org/10.1201/ebk1439802458 [26] d. alantari, s. yalkowsky. comments on prediction of the aqueous solubility using the general solubility equation (gse) versus a genetic algorithm and a support vector machine model. pharmaceutical development and technology 23 (2018) 739-740. https://doi.org/10.1080/10837450.2017.1321663 [27] m.h. abraham. scales of hydrogen bonding their construction and application to physicochemical and biochemical processes. chemical society reviews 22 (1993) 73-83. https://doi.org/10.1039/cs9932200073 [28] m.h. abraham, j. le. the correlation and prediction of the solubility of compounds in water using an amended solvation energy relationship. journal of pharmaceutical sciences 88 (1999) 868-880. https://doi.org/10.1021/js9901007 [29] j.a. platts, d. butina, m.h. abraham, a. hersey. estimation of molecular linear free energy relation descriptors using a group contribution approach. journal of chemical information and computer sciences 39 (1999) 835-845. https://doi.org/10.1021/ci980339t [30] l. breiman. random forests. machine learning 45 (2001) 5-32. https://doi.org/10.1023/a:1010933404324 [31] d.s. palmer, n.m. o’boyle, r.c. glen, j.b.o. mitchell. random forest models to predict aqueous solubility. journal of chemical information and modeling 47 (2007) 150-158. https://doi.org/10.1021/ci060164k [32] w.p. walters. what are our models really telling us? a practical tutorial on avoiding common mistakes when building predictive models, in chemoinformatics for drug discovery. j. bajorath (ed.). john wiley & sons, hoboken, nj, 2014, pp. 1-31. https://doi.org/10.1002/9781118742785.ch1 [33] a. liaw. random forests what, why, and how. https://www.youtube.com/watch?v=xjnjlpw9w5a. (youtube lecture) https://nyhackr.blob.core.windows.net/presentations/random-forests-what-whyand-how_andy_liaw.pdf (accessed 23 nov 2022). [34] l.b. kier. an index of molecular flexibility from kappa shape attributes. quant. struct.-act. relat. 8 (1989) 221-224. https://doi.org/10.1002/qsar.19890080307 [35] a. avdeef. do you know your r2? admet & dmpk 9 (2021) 69-74. https://doi.org/10.5599/admet.888 [36] c.a. lipinski, f. lombardo, b.w. dominy, p.j. feeney. experimental and computational approaches to estimate solubility and permeability in drug discovery and development settings. advanced drug delivery reviews 23 (1997) 3-25. https://doi.org/10.1016/s0169-409x(00)00129-0. [37] b.c. doak, b. over, f. giordanetto, j. kihlberg. oral druggable space beyond the rule of 5: insights from drugs and clinical candidates. chemistry & biology 21 (2014) 1115-1142. https://doi.org/10.1016/j.chembiol.2014.08.013 [38] d.a. degoey, h.-j. chen, p.b. cox, m. d. wendt. beyond the rule of 5: lessons learned from abbvie’s drugs and compound collection. journal of medicinal chemistry 61 (2018) 2636-2651. https://doi.org/10.1021/acs.jmedchem.7b00717 [39] g. ermondi, m. vallaro, g. goetz, m. shalaeva, g. caron. experimental lipophilicity for beyond rule of 5 compounds. future drug. discov. 1 (2019) https://doi.org/10.4155/fdd-2019-0002 [40] g. caron, j. kihlberg, g. ermondi. intramolecular hydrogen bonding: an opportunity for improved design in medicinal chemistry. medicinal research reviews 39 (2019) 1707-1729. https://doi.org/10.1002/med.21562 https://doi.org/10.5599/admet.1897 https://api.semanticscholar.org/corpusid:92804093 https://doi.org/10.1016/j.ijpharm.2006.04.022 https://doi.org/10.1016/j.ijpharm.2006.04.022 https://doi.org/10.1201/ebk1439802458 https://doi.org/10.1080/10837450.2017.1321663 https://doi.org/10.1039/cs9932200073 https://doi.org/10.1039/cs9932200073 https://doi.org/10.1021/js9901007 https://doi.org/10.1021/ci980339t https://doi.org/10.1023/a:1010933404324 https://doi.org/10.1023/a:1010933404324 https://doi.org/10.1021/ci060164k https://doi.org/10.1021/ci060164k https://doi.org/10.1002/9781118742785.ch1 https://www.youtube.com/watch?v=xjnjlpw9w5a https://nyhackr.blob.core.windows.net/presentations/random-forests-what-why-and-how_andy_liaw.pdf https://nyhackr.blob.core.windows.net/presentations/random-forests-what-why-and-how_andy_liaw.pdf https://doi.org/10.1002/qsar.19890080307 https://doi.org/10.5599/admet.888 https://doi.org/10.1016/s0169-409x(00)00129-0 https://doi.org/10.1016/j.chembiol.2014.08.013 https://doi.org/10.1021/acs.jmedchem.7b00717 https://doi.org/10.4155/fdd-2019-0002 https://doi.org/10.1002/med.21562 a. avdeef admet & dmpk 00(0) (2023) 000-000 14 [41] g. ermondi, m. vallaro, g. goetz, m. shalaeva, g. caron. updating the portfolio of physicochemical descriptors related to permeability in the beyond the rule of 5 chemical space. european journal of pharmaceutical sciences 146 (2020) 105274. https://doi.org/10.1016/j.ejps.2020.105274 [42] g. caron, v. digiesi, s. solaro, g. ermondi. flexibility in early drug discovery: focus on the beyond-ruleof-5 chemical space. drug discovery today 25 (2020) 621-627. https://doi.org/10.1016/j.drudis.2020.01.012 [43] g. ermondi, v. poongavanam, m. vallaro, j. kihlberg, g. caron, g. solubility prediction in the bro5 chemical space: where are we right now? admet & dmpk 8 (2020) 207-214. https://doi.org/10.5599/admet.834 [44] d.g. jiménez, m.r. sebastiano, m. vallaro, v. mileo, d. pizzirani, e. moretti, g. ermondi, g. caron. designing soluble protacs: strategies and preliminary guidelines. journal of medicinal chemistry 65 (2022) 12639-12649. https://doi.org/10.1021/acs.jmedchem.2c00201 [45] a. mullard. 2021 fda drug approvals. the fda approved 50 novel drugs in 2021, including the first kras inhibitor for cancer and the first anti-amyloid antibody for alzheimer’s disease. nature reviews drug discovery 21 (2022) 83-88. https://doi.org/10.1038/d41573-022-00001-9. [46] l.d. hughes, d.s. palmer, f. nigsch, j.b.o. mitchell. why are some properties more difficult to predict than others? a study of qspr models of solubility, melting point, and log p. journal of chemical information and modeling 48 (2008) 220-232. https://doi.org/10.1021/ci700307p. [47] c.a. lipinski. drug-like properties and the causes of poor solubility and poor permeability. journal of pharmacological and toxicological methods 44 (2000) 235-249. https://doi.org/10.1016/s10568719(00)00107-6. ©2023 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) https://doi.org/10.1016/j.ejps.2020.105274 https://doi.org/10.1016/j.drudis.2020.01.012 https://doi.org/10.1016/j.drudis.2020.01.012 https://doi.org/10.5599/admet.834 https://doi.org/10.5599/admet.834 https://doi.org/10.1021/acs.jmedchem.2c00201 https://doi.org/10.1038/d41573-022-00001-9 https://doi.org/10.1021/ci700307p https://doi.org/10.1016/s1056-8719(00)00107-6 https://doi.org/10.1016/s1056-8719(00)00107-6 http://creativecommons.org/licenses/by/3.0/ manuscript doi: 10.5599/admet.3.1.159 15 admet & dmpk 3(1) (2015) 15-33; doi: 10.5599/admet.3.1.159 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper expression and transport of -synuclein at the bloodcerebrospinal fluid barrier and effects of manganese exposure christopher a. bates 1 , sherleen fu 1 , daniel ysselstein 2 , jean-christophe rochet 2 , and wei zheng 1 * 1 school of health sciences, purdue university, west lafayette, in 47907 2 department of molecular pharmacology and medicinal chemistry, purdue university, west lafayette, in 47907 *corresponding author: e-mail: wzheng@purdue.edu; tel.: (765) 496-6447; fax: (765) 496-1377 received: january 19, 2015; revised: march 16, 2015; published: march 31, 2015 abstract the choroid plexus maintains the homeostasis of critical molecules in the brain by regulating their transport between the blood and cerebrospinal fluid (csf). the current study was designed to investigate the potential role of the blood-csf barrier (bcsfb) in α-synuclein (a-syn) transport in the brain as affected by exposure to manganese (mn), the toxic metal implicated in parkinsonian disorders. immunohistochemistry was used to identify intracellular a-syn expression at the bcsfb. quantitative realtime pcr was used to quantify the change in a-syn mrna expression following mn treatments at the bcsfb in vitro. elisa was used to quantify a-syn levels following in vivo and in vitro treatments of mn, copper (cu), and/or external a-syn. thioflavin-t assay was used to investigate a-syn aggregation after incubating with mn and/or cu in vitro. a two-chamber transwell system was used to study a-syn transport by bcsfb monolayer.data revealed the expression of endogenous a-syn in rat choroid plexus tissue and immortalized choroidal epithelial z310 cells. the cultured primary choroidal epithelia from rats showed the ability to take up a-syn from extracellular medium and transport a-syn across the cellular monolayer from the donor to receiver chamber. exposure of cells with mn induced intracellular a -syn accumulation without causing any significant changes in a-syn mrna expression. a significant increase in a-syn aggregation in a cell-free system was observed with the presence of mn. moreover, mn exposure resulted in a significant uptake of a-syn by primary cells.these data indicate that the bcsfb expresses a-syn endogenously and is capable of transporting a-syn across the bcsfb monolayer; mn exposure apparently increases a-syn accumulation in the bcsfb by facilitating its uptake and intracellular aggregation. keywords a-synuclein transport; cerebrospinal fluid; blood-cerebrospinal fluid barrier; manganese exposure introduction parkinson’s disease (pd) is the second most common neurodegenerative disorder behind alzheimer’s disease and is characterized by dopaminergic neurodegeneration in the substantia nigra with the following hallmark symptoms: bradykinesia, rigidity, postural instability, and resting tremors [1]. dysfunction and aggregation of the protein, -synuclein (a-syn) has been associated with pd pathophysiology [2]. the protein a-syn is small (14.5 kda), natively unfolded, and expressed in a wide range of cell types. the exact http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:wzheng@purdue.edu bates et al. admet & dmpk 3(1) (2015) 15-33 16 function of wild-type a-syn is uncertain. some researchers have suggested potential functions of a-syn in synaptic plasticity [3], synaptic pool maintenance [4], regulation of dopamine synthesis [5], and vesicular dynamics, e.g. stabilization and exocytic fusion at the presynapse [2,6,7]. the aggregation of a-syn within neurons results in the efflux of aggregate species from the cell to the interstitial fluid (isf) [8,9], which merges with the cerebrospinal fluid (csf). the csf is mainly produced by the choroid plexus located in brain ventricles. the choroid plexus forms the blood-csf barrier (bcsfb), which separates the blood circulation from the csf. for its continuous cycle throughout brain structure, the csf functions as a “sink”, where waste products can be collected from the isf and removed from the brain via the bcsfb or subarachnoid drainage [10,11]. the a-syn molecules present in the csf are believed to originate primarily from neurons and neuroglia [12], and the a-syn concentration in the csf has been compared against a-syn in blood as potential biomarkers for pd progression [13-15]. literature evidence suggests that the transport of a-syn aggregates between neurons and astrocytes of the blood-brain barrier (bbb) is by endocytosis and exocytosis [8,9,16,17]. since the a-syn level in the blood is over 1000-fold higher than that in the csf in healthy humans [12,14,18], any possible transport of a-syn from the csf back to the blood must be energy dependent and against this large concentration gradient. these transport properties, however, have yet to be thoroughly investigated at the bcsfb or bbb. for detailed discussions, please see a recent review by bates and zheng [19]. manganese (mn) is an essential trace element utilized for a variety of functions in the brain including glutamate and gaba metabolism, astrocyte morphology and migration, and the neutralization of reactive oxidative species [20]. mn overexposure occurs primarily in the occupational setting for welders, smelters, and others with related professions [21,22], but it also occurs in premature newborns via medically administered paternal nutrition [20]. excess mn has been shown to accumulate primarily in the globus pallidus; more recent data have shown that mn also accumulates in the substantia nigra [23-25]. in humans and rats, excess mn has been associated with increased gaba levels at the basal ganglia, impaired iron and copper homeostasis, and mitochondrial dysfunction [26-30]. individuals suffering from mn intoxication display clinical symptoms including impaired balance and gait, kinetic tremors, dystonia, cognitive problems, memory loss, anxiety, and aggression [31,32]. previous findings in the literature showed that the choroid plexus accumulates mn to a great extent after mn exposure [22,33,34]. the bcsfb is known to favor mn influx to the brain with a slow rate of mn efflux from the csf [22,35,36]. this may contribute to the high mn storage and relatively long half-life of mn in the brain [37]. studies have also indicated that mn can interact with a-syn, promoting the aggregation of the protein into oligomers in vitro [38,39] and overexpression of a-syn in pc12 and sh-sy5y cells [40,41]. the question as to whether or not mn may interfere with a-syn production, aggregation and transport in the bcsfb remains unanswered. the purposes of the current study were to 1) determine the endogenous expression of a-syn in the bcsfb as affected by mn exposure, 2) investigate the uptake of a-syn by the bcsfb and the effect of mn exposure, 3) study the direct interaction between mn and a-syn in a cell-free system, and 4) explore the directional transport properties of wild-type a-syn monomer by the bcsfb and effects of mn exposure. the findings of these works establish a foundation for understanding the role of the bcsfb in regulating a-syn homeostasis in the brain, the effects of toxic mn exposure, and the potential implications of these relationships in parkinsonian disorders. admet & dmpk 3(1) (2015) 15-33 effects of manganese exposure on expression and transport of -synuclein doi: 10.5599/admet.3.1.159 17 experimental materials chemicals and reagents were acquired from the following sources: mn chloride tetrahydrate (mncl2•4h2o) from fisher scientific (pittsburgh, pa); sodium chloride (nacl) from avantor performance materials (center valley, pa); sodium phosphate (na2hpo4), potassium chloride (kcl), and potassium phosphate (kh2po4) from mallinckrodt chemicals (phillipsburg, nj); copper chloride (cucl2), paraformaldehyde (pfa), phenylmethylsulfonyl fluoride (pmsf), epidermal growth factor (egf), cis-4hydroxyl-l-proline (cis-hp), 0.25 % trypsin-edta from sigma (st. louis, mo); hank’s balanced salt solution (hbss), fluor alexa-555 conjugated goat anti-mouse secondary antibody, fluor alexa-488 conjugated goat anti-rabbit secondary antibody, trizol reagent, collagen i (rat tail), fetal bovine serum (fbs), penicillin, streptomycin, amphotericin, gentamycin, and human a-syn elisa kit from life technologies (carlsbad, ca); fetal bovine serum (fbs), and dulbecco’s modified eagle’s medium (dmem) from cellgro (manassas, va); pronase, protease, and protease inhibitor cocktail from calbiochem (san diego, ca); sodium dodecyl sulfate (sds), triton x-100, cdna synthesis kit, and itaq universal sybr green supermix from bio-rad (hercules, ca); collagen-coated transwell-col inserts from corning (cambridge, ma); rabbit polyclonal asyn primary antibody from cell signaling technologies (2642s) (danvers, ma). all reagents were of analytical grade, hplc grade, or the best pharmaceutical grade available. cell cultures and treatments the choroidal epithelial z310 cell line was derived from rodent choroid plexus by this laboratory; the culturing and maintenance procedures of this cell line have been previously described [42-44]. z310 cells were grown in dmem medium supplemented with 10 % fbs, 100 u/ml penicillin-streptomycin, 100 μg/ml gentamicin sulfate, and 10 ng/ml egf, in a humidified incubator with 95% air-5 % co2 at 37 °c. 0.25 % trypsin-edta was used to dissociate cells during maintenance. z310 cells were passaged twice a week. for culturing primary choroidal epithelial cells from rats, male sprague-dawley rats (3-week old, 35-49 g) were used. rats were purchased from harlan laboratories (indianapolis, in) and euthanized with co2. choroid plexuses from the lateral and third ventricles were dissected, minced down to cubes 1 mm in length, and digested in 1 ml hbss containing 0.2 % pronase at 37 °c for 5-10 min. digestion was stopped by adding 4 ml of hbss. cells were pelleted via centrifugation (800 x g for 5 min), washed with fresh hbss, and centrifuged again (800 x g for 5 min). cells were passed 13-14 times through a 20-gauge needle attached to a 10 ml syringe and counted, seeded on wells pre-coated with collagen (gibco/life technologies cat. no. a1048301), and incubated at 37 °c with 95 % air-5 % co2 for 2-3 days without disturbing. medium was replaced every other day with a fresh growth medium. on the 4 th day, cells were cultured with the growth medium supplemented with 25 μg/ml cis-hp to inhibit the growth of fibroblastic cells. this protocol has been used extensively in previous research performed by this lab [27,29,45-47]. autoclaved 25 mm mncl2 and 25 mm cucl2 stock solutions were prepared by dissolving mncl2•4h2o and cucl2 in sterile, double-deionized water, respectively. z310 cells were treated with 50 or 100 μm mncl2 or cucl2 for 24 h, or 100 μm mncl2 for 24 and 48 h for the immunohistochemical studies. z310 cells were treated with 25, 50, or 100 μm mncl2 for 24 h for the quantitative rt-pcr study. the dose regimen, used in the following studies, was chosen according to previous research performed in this lab or by other groups [27,29,42,48]. bates et al. admet & dmpk 3(1) (2015) 15-33 18 immunohistochemical (ihc) staining freshly harvested choroid plexus tissues were immediately fixed with 4 % paraformaldehyde (ph 7.4) for 10 min, permeabilized with 0.1 % triton x-100, and blocked in 1 % bovine serum albumin (bsa) at room temperature for 1 h. tissues were washed three times with pbs (10 min/wash) after the fixing and permeabilizing steps. tissues were incubated with polyclonal rabbit anti-a-syn antibody (1:100) in 1 % bsa at 4 °c for 48 h, followed by repeated washes with pbs (3 x 10 min/wash), and then incubated with alexa fluor -488 conjugated goat anti-rabbit secondary antibody (1:1000) in 1 % bsa at room temperature for 1 h. tissues were rinsed with pbs (3 x 10 min/wash) and then mounted to the objective slides using prolong gold anti-fade reagent (life technologies, carlsbad, ca) to prevent fluorescent bleaching for confocal microscopy examination. glass coverslips (0.17 mm in thickness) were sterilized with 75 % ethanol. z310 cells were seeded on glass coverslips at a density of 2 x 10 5 /well in a 6-well-plate. rat plexus primary cells were seeded on glass coverslips at a density of 5 x 10 4 /well in a 24-well plate and allowed to grow for 24 h to achieve 50 % confluence, and then treated with 100 μm mncl2 for 12, 24, or 48 h. at the end of treatment, cells were fixed in 4 % paraformaldehyde and permeabilized in 0.1 % triton x-100 at room temperature and washed with pbs (3 x 5 min/wash) after the fixing and permeating steps. cells were then blocked with 1% bsa for 1 h at room temperature followed by incubation with rabbit polyclonal a-syn antibody (1:1000) at 4 °c for 48 h. cells were washed with pbs (3 x 5 min/wash) before being incubated in alexa fluor-488 conjugated goat anti-rabbit secondary antibody (1:1000) for 24 h. glass coverslips were mounted to objective slides for confocal microscopy observation. slides were prepared using gold anti-fade to avoid fluorescent bleaching for confocal microscopic examination. all slides were dried in a hood, with no exposure to light, at room temperature overnight. the negative control was established by using only the secondary antibody to reflect non-specific staining on the background. confocal microscopy and quantification to acquire images, pre-dried slides containing experimental specimens were mounted on the stage of a nikon inverted confocal laser-scanning microscope and viewed through a 60x oil-immersion objective (plan apo, 60x/1.40 oil, ∞ 0.17, dic), with a 488-nm laser and a 562-nm laser source for excitation. lower laser intensity was used to avoid photo bleaching. each slide was examined under reduced transmitted-light illumination; the area containing undamaged epithelium with underlying vasculature was chosen for analyses. approximately 10-20 areas of epithelium were selected for image collection and analysis. image data reported were the results of a single experiment representative of three to four experiments. images were scanned with the nikon ez-c1 (version 3.90) confocal imaging program. signal intensities were quantified using the nis-elements br (version 3.10) program. each cell or group of cells was isolated as a region of interest (roi). cells that were partially imaged within the image frame were excluded to improve accuracy and reliability of quantification. mean intensity values were calculated as the sum of intensities within the corresponding roi divided by the area of the roi. a minimum of three images per treatment group was used for confocal quantification. a-synuclein mrna quantitation by quantitative real-time rt-pcr (qpcr) the transcription levels of mrna encoding rat a-syn (snca) were quantified using qpcr. z310 cells were seeded and cultured for 24 h prior to incubation with 25, 50, or 100 μm mncl2 for 24 h. total rna was isolated from the cells using trizol, following the manufacturer’s directions. rna samples were reverseadmet & dmpk 3(1) (2015) 15-33 effects of manganese exposure on expression and transport of -synuclein doi: 10.5599/admet.3.1.159 19 transcribed using the biorad iscript cdna synthesis kit (hercules, ca). the itaq universal sybr green supermix (hercules, ca) was used for qpcr analysis. snca cdna was amplified using the following nucleotide sequences: forward primer: 5’-aga tct gcc cag gtg ttc ttcc-3’ reverse primer: 5’-agg act ccg atc act gct gatg-3’ the primers were designed using the primer express 3.0 software. all primers were obtained from integrated dna technologies (coralville, ia). the amplification was performed in the cfx connecttm realtime pcr detection system (biorad, hercules, ca). thioflavin t assay recombinant human a-syn was produced using the methods by zhang et al. [49]. the protein was dialyzed in 1x pbs (137 mm nacl, 2.7 mm kcl, 10 mm na2hpo4, and 2 mm kh2po4, ph 7.4) overnight prior to use in order to remove excess salt. the recombinant human a-syn protein (70 μm) was incubated with 100 μm mncl2, cucl2, or a mixture of mncl2 and cucl2 (each at 100 μm) in a 200 μl solution of 20 mm 2-(n-morpholino)ethanosulfonate (mes) buffer containing 200 μm thioflavin t for 72 h at 37 °c under constant stirring with a teflon ball. the combination of mncl2 and cucl2 at 100 μm without a-syn are used as a negative control. fluorescence readings were measured with a tecan genios plate reader (san jose, ca) every 15 min (excitation: 440 nm; emission: 480 nm). tecan magellan software (san jose, ca) was used for recording the data. concentrations of metal ions and a-syn were chosen based on previous literature [38,39]. cellular uptake of a-syn and elisa quantification primary rat plexus cells were seeded and cultured as described above. once the cells reached 70-80 % confluence, they were exposed to 25, 50, or 100 μm mncl2 for 24 h. cells were then washed with pbs and incubated with 0.5 mg/ml recombinant human a-syn for 2 h in serum-free dmem medium. following treatments, cells were washed and collected. cell lysates were analyzed for a-syn content via elisa quantification. protein lysates were extracted from cell samples in a homogenization buffer containing 20 mm tris, ph 7.5, 5 mm egta, 1 % triton x-100, 0.1 % sds, and protease inhibitor cocktail (calbiochem, san diego, ca). samples were sonicated, centrifuged, and quantified for protein concentration using the bradford protein assay. medium and cell samples were diluted 25-fold in diluent buffer included in the human a-synuclein elisa kit (product khb0061, invitrogen, frederick, md) in accordance with the product’s protocol. samples were measured for fluorescence at 450 nm with an m2e plate reader (molecular devices, sunnyvale, ca) and the softmax pro analysis software (version 5.4, molecular devices, sunnyvale, ca) within 2 h of the addition of the stop solution. transport of a-syn across bcsfb monolayer in two-chamber transwell system please refer to fu et al. [29] for a more detailed description of this procedure. rat choroid plexus primary cells, obtained as previously described were seeded on collagen-coated laminin membranes in the inner chamber of a two-chamber transwell system. figure s1 illustrates the concept of this transport system. a 1.0 ml aliquot of growth dmem medium was added to the inner chamber. the inner chamber was then immersed in 1.0 ml of medium in the outer chamber. cells were cultured in the medium for 7-10 days to the confluence. wells were then divided into 2 main sets, i.e., one set for the influx studies and the bates et al. admet & dmpk 3(1) (2015) 15-33 20 other for the efflux studies. each set was further divided into 3 time points: 6, 12, and 24 h. each time point was comprised of a control (no treatment) and a mn-treated group. transepithelial electrical resistance (teer), an indicator of barrier tightness, was used to monitor the development and the tightness of the barrier monolayer on the laminin membrane. teer readings were taken every 2 days until the resistance reached 50-60 ω cm 2 , which confirms the presence of tight junctions throughout the barrier [27,29,46]. once the cell monolayer was confluent and completely developed, cells designated for mn treatment were treated with 100 μm mncl2 for 24 h. cells were incubated with 1.0 mg/ml a-syn dissolved in serum-free dmem medium. a-syn was added to the outer chamber for the “influx” study; for the “efflux” study, a-syn was added to the inner chamber only. a 20-μl aliquot of both chambers and cells was collected at each time point and immediately frozen at -80 °c until being thawed for elisa quantification. statistical analysis confocal quantification and the transwell chamber studies were analyzed using one-way or twoanova with tukey’s post-hoc test as necessary. qpcr data were analyzed with one-way anova and the dunnett’s post-hoc test. the a-syn uptake studies quantified with elisa were analyzed with one-way anova and the tukey’s post-hoc test. any p values equal to or less than 0.05 were considered statistically significant. results expression of a-syn in the choroidal epithelia and effects of mn exposure data from freshly isolated choroid plexus tissue (fig. 1a) and choroidal epithelial z310 cells (fig. 1b) showed that the choroid plexus cells expressed strong fluorescent signals representing a-syn; furthermore, the fluorescent signal was present primarily in the cytoplasm of these cells. this cytoplasmic expression of a-syn is consistent with the localization of a-syn in other cell types described previously in the literature [2,50]. our data clearly show that choroidal epithelial cells possess the ability to express endogenous a-syn. following incubation of z310 cells with 100 μm mncl2 for 48 h, some small, yet bright green dots appeared to form in the cytoplasm (fig. 2a). control z310 cells did not display these dots. however, quantitation of the intensity of a-syn signals in these cells did not reveal a significant difference between control and mn-treated groups at 24 h but did show a significant decrease at 48 h (fig. 2b). this decrease could be a result of the clearing of a-syn aggregates, which has been shown to occur in the presence of mn in neurons [50]. the decrease in a-syn signals, however, is contradictory to some literature reports, where a-syn overexpression following mn exposure was observed in other cell types [40, 41]. we further used qpcr to quantify a-syn mrna expression in z310 cells with various concentrations of mn. after a 24 h treatment, the qpcr analysis did not reveal any significant change in a-syn mrna expression (fig. 2c). these data were consistent with the confocal quantitation results, supporting the view that mn exposure did not induce a-syn overexpression at either the transcriptional or translational level. the changes in a-syn signals in z310 cells after mn exposure may reflect a direct interaction of mn with asyn. admet & dmpk 3(1) (2015) 15-33 effects of manganese exposure on expression and transport of -synuclein doi: 10.5599/admet.3.1.159 21 figure 1. expression of endogenous a-syn in choroid plexus. (a) presence of a-syn in rat choroid plexus tissue. choroid plexuses were dissected from rats and stained with primary anti-a-syn. a typical image from n = 4 is presented. (b) presence of a-syn in choroidal z310 cells. z310 cells were fixed and stained with primary anti-a-syn. a typical image from n = 4 is presented. in both tissues and cells, a-syn immunoreactivity is evident in the cytosol. scale bar: 20 μm. effects of cu treatment on endogenous expression of a-syn at the bcsfb recent findings in our lab have shown that mn accumulation in the choroid plexus can cause cu dyshomeostasis in the bcsfb [29]. cu itself has been shown to induce a-syn aggregation and dysfunction [38,39,51]. to compare with the mn effects outlined in the previous section, we sought to investigate the effect of cu on endogenous a-syn expression in the plexus cells. after z310 cells were incubated with 50 and 100 μm cucl2 for 24 h, bright dots were also observed in the cytoplasm of z310 cells, similar to those in the mn exposed group (fig. 3a). confocal quantitation of asyn signals revealed that cells exposed to 50 μm cucl2 displayed a significant increase in a-syn intensity (fig. 3b). a) b) bates et al. admet & dmpk 3(1) (2015) 15-33 22 figure 2. endogenous a-syn expression in z310 cells following mn exposure. (a) ihc staining of a-syn in z310 cells. z310 cells were treated with 100 μm mncl2 for 24 and 48 h. the white circle indicates that at 48 h after mn treatment, bright green puncta were formed in the cytoplasm. data represent a typical study of n = 4. scale bar: 20 μm. (b) confocal quantification of endogenous a-syn expression in z310 cells following mn exposure. data represent mean ±sd. n = 15-27. *: p<0.05 as compared to ctrl. ##: p<0.01 as compared to data at 24 h. (c) expression of endogenous a-syn mrna in z310 cells following mn exposure. z310 cells were treated with 25, 50, or 100 μm mncl2 for 24 h. the mrna levels were quantified by qpcr. data represent mean ± sd; n = 4. aggregation of a-syn following mn and cu treatment results so far appeared to suggest a direct, physicochemical interaction between a-syn, mn, and/or cu. thioflavin t (thiot) has been used extensively in previous studies to assess metal-induced aggregation of asyn [38,39]. in the current study, we incubated recombinant human a-syn with mn, cu, or a combination of mn and cu at a concentration of 100 μm [38,39] and used thiot to determine the metal effects on a-syn aggregation in vitro. results presented in fig. 4 showed a drastic decrease in the maximal (final) fluorescence of a-syn when incubated with mn, cu, or the combination of the two metals (data not shown). this quenching of fluorescence was an indication of a structural change of a-syn and its aggregated forms in the presence of these metals [38]. when the groups were normalized to their respective maximum fluorescence intensities, the results revealed that the emergence of a-syn fluorescence occurred much earlier, within 5 h when a-syn was incubated with mn or cu alone, compared to a-syn alone at about 10 h (fig. 4). unexpectedly, however, co-incubation of a-syn with both mn and cu did not show an additive effect, but rather it counteracted each metal’s individual effect by slowing the increase of a-syn fluorescence (fig. 4). this finding suggests that both metals likely bind to a-syn, inducing altered a) b) c) * ## admet & dmpk 3(1) (2015) 15-33 effects of manganese exposure on expression and transport of -synuclein doi: 10.5599/admet.3.1.159 23 conformations of a-syn aggregates. to substantiate this finding, experiments beyond thiot, e.g., nmr, should be conducted. figure 3. expression of a-syn in z310 cells following cu treatment. (a) ihc staining of a-syn in z310 cells. z310 cells were treated with 50 or 100 μm cucl2 for 24 h. the white circles indicate that after cu treatment, bright green puncta were formed in the cytoplasm. data represent a typical study of n = 4. (b) confocal quantification of endogenous a-syn expression in z310 cells following cu treatment. data represent mean ± sd. n = 12-20. **: p<0.01 as compared to control. green (a-syn) monochro./ dic merge control 50 μm cu 100 μm cu a) b) ## bates et al. admet & dmpk 3(1) (2015) 15-33 24 figure 4. thioflavin-t assay of a-syn aggregation in vitro. human recombinant a-syn was incubated in vitro with 100 μm mncl2, 100 μm cucl2, or a combination of 100 μm mn and 100 µm cu for 72 hours under constant stirring with teflon ball. data were normalized to the maximum fluorescent values achieved by each group (n = 3). uptake of a-syn by plexus cells and effect of mn exposure when primary cultures of choroid epithelial cells were exposed to 100 μm mncl2, without adding a-syn, the cells displayed a significant increase in a-syn signals after 48 h (fig. 5a). with added external a-syn, the signals became stronger and the presence of intracellular puncta containing a-syn became apparent (fig. 5a). confocal quantifications revealed an increased signal intensity at 48 h in the 50 μm group without added a-syn (p<0.05) and in both 50 and 100 μm groups with added a-syn (p<0.01) as compared to controls (fig. 5b). the a-syn signals in cells exposed to mn + a-syn were significantly greater than those in cells exposed to mn alone at the same mn concentration (supplemental fig. s2-s3). to directly quantify the effect of mn exposure on a-syn uptake, we used an elisa method to determine intracellular a-syn levels with or without mn exposure. primary choroidal epithelial cells treated with various concentrations of mn for 24 h followed by incubation with a-syn for an additional 2 h showed a gradual decrease in a-syn content with an increase in mn concentration (fig. 6). a significant decrease was observed at 100 μm mn (the highest mn concentration) compared to the control (p<0.05) (fig. 6). the observed decrease in a-syn uptake was a direct contrast to the confocal microscopy data. this discrepancy could be explained by the fact that the antibody used in the elisa method was only capable of detecting human a-syn monomers but not oligomers. nonetheless, a significant decrease of a-syn monomer detected by elisa, in fact, indirectly suggested an increase in aggregated a-syn within the cells. a-syn transport by bcsfb monolayer in a two-chamber transwell device to determine the directional transport of a-syn across the bcsfb and the effect of mn exposure, we adapted a two-chamber transwell transport model (fig. s1, efflux model is shown). one approach involved adding a-syn to the outer (donor) chamber and subsequently monitoring the appearance of a-syn in the inner (receiver) chamber to model the influx of a-syn from the blood to the csf. conversely, the a-syn was added to the inner chamber and the protein’s appearance in the outer chamber was monitored to model the efflux of a-syn from the csf to the blood. asyn+mn+cu asyn+mn asyn+cu asyn alone mn+cu admet & dmpk 3(1) (2015) 15-33 effects of manganese exposure on expression and transport of -synuclein doi: 10.5599/admet.3.1.159 25 figure 5. uptake of a-syn by primary choroidal epithelial cells in the presence of a-syn and after mn exposure. (a) ihc staining of a-syn in primary plexus cells. cells were treated with mn or saline (control) for 12, 24, or 48 h, followed by incubation with 0.5 μm recombinant human a-syn for 2 h. cells were stained with primary anti-a-syn. a typical study of n = 5 is presented. (b) confocal quantification of a-syn. “+mn, -a-syn”: cells treated with mn but without added a-syn; “+mn, +a-syn”: cells treated with mn followed by a-syn incubation. data represent mean ± sd. n = 10-24. *: p < 0.05, **: p < 0.01, ***: p < 0.001 as compared to controls at matching time point. results from the influx study showed that after a-syn was added to the outer chamber, there was a significant increase of a-syn content in the inner chamber between the 0-h and 6-h time points; the increase continued until 12-h in the inner chamber, followed by a gradual decline at 24-h (fig. 7a). accordingly, the a-syn concentration in the outer (donor) chamber decreased significantly between the 0-h and 6-h time points. interestingly, the a-syn concentration in the outer chamber began to increase to an apparent steady state between 12 and 24 h (fig. 7a). mn exposure had no significant effect on a-syn concentrations in either the inner or outer chambers at any given time point (fig. 7a). by quantifying a-syn concentrations in choroidal epithelial cells, we determined that there was a detectable amount of a-syn accumulated in the cells in the control group, and this accumulation appeared to continue to increase up to 24 h, although there was no statistical significance between 12 and 24 h (fig. 7b). mn treatment had no significant effect on intracellular a-syn concentrations of the choroidal epithelial monolayer (fig. 7b). a) b) * *** *** *** ** * bates et al. admet & dmpk 3(1) (2015) 15-33 26 figure 6. elisa quantification of a-syn uptake by primary choroidal epithelial cells following mn exposure. primary cells were pre-treated with 25, 50, or 100 μm mncl2 or with saline (‘ct’) for 24 h, followed by incubation with 0.5 mg/ml a-syn for 2 h. nct (negative control) was not treated with mncl2, saline, or a-syn. levels of a-syn were quantified by elisa. data represent mean ± sd, n = 5. *: p<0.05, **: p<0.01 as compared with controls. the efflux study showed an increase of a-syn in the outer (receiver) chamber from the 0-h time point to the 12-h time point, followed by a decrease at 24 h (fig. 8a). in the inner donor chamber, the a-syn levels were initially sharply decreased between the 0-h and 6-h time points, followed by a significant, albeit slight, increase in a-syn content between the 6-h and 12-h time points (fig. 8a). the a-syn concentration in the inner chamber appeared to reach a steady state between 12and 24-h time points (fig. 8a). similar to the influx study, mn exposure had no significant effect on the transport of a-syn by the bcsfb monolayer (fig. 8a). also similar to the influx study, the accumulation of a-syn in primary choroidal epithelial cells in controls was increased from 0-h to 24 h, although there was no significant change between 12 and 24 h (fig. 8b). mn exposure did not significantly alter the intracellular a-syn concentrations of the bcsfb monolayer (fig. 8b). the results from these influx/efflux studies suggested that a-syn molecules were likely transported in either direction across the bcsfb monolayer, and the bcsfb did not appear to have a preferred direction of a-syn transport. moreover, mn exposure had no effect on a-syn transport by the bcsfb or intracellular a-syn content. discussion the data presented in this report suggest that 1) a-syn is naturally expressed in choroid plexus tissues, particularly within epithelial cells; it is also present in the immortalized choroidal epithelial z310 cell line; 2) treatment of plexus cells with mn or cu induces a-syn accumulation within the bcb; 3) mn exposure does not significantly alter a-syn production at the transcriptional level; 4) in the test tube, incubation of a-syn with mn or cu causes the aggregation of a-syn; and 5) while mn exposure does not affect the transport of a-syn across the bcsfb monolayer, it increases a-syn uptake by a primary culture of choroidal epithelial cells. ** * admet & dmpk 3(1) (2015) 15-33 effects of manganese exposure on expression and transport of -synuclein doi: 10.5599/admet.3.1.159 27 figure 7. effect of mn exposure on a-syn influx transport (from the outer to inner chamber) by choroidal epithelial monolayer. primary choroidal cells were cultured in the inner chamber. an aliquot of a-syn was added to the outer chamber after the cells were treated with 100 μm mncl2 for 24 h. levels of a-syn in both chambers and in the cells were quantified by elisa. (a) time course. data represent mean ± sd, n = 6. means are compared between inner and outer chamber at each time point (6, 12, 24 h). ***: p < 0.001 as compared to 6-hour inner chamber/ct group. ###: p < 0.001 as compared to 6-hour outer chamber/ct group. (b) a-syn concentrations in cell monolayer following mn exposure. data represent mean ± s.d., n = 5-6. figure 8. effect of mn exposure on a-syn efflux transport (from the inner to outer chamber) by choroidal epithelial monolayer. primary choroidal cells were cultured in the inner chamber. an aliquot of a-syn was added to the inner chamber after the cells were treated with 50 μm mncl2 for 24 h. levels of a-syn in both chambers and in the cells were quantified by elisa. (a) time course. data represent mean ± sd. n = 6. means were compared between inner and outer chamber at each time point (6, 12, 24 h). ***: p < 0.001 as compared to 6-hour inner chamber/ct group ###: p < 0.001 as compared to 6-hour outer chamber/ct group. (b) a-syn concentrations in cell monolayer following mn exposure. data represent mean ± s.d., n = 5-6. endogenous expression of a-syn occurs across a variety of cell types, e.g. neurons, erythrocytes, etc. [52,53]. here we report the expression of endogenous a-syn in choroidal epithelial cells in brain ventricles. multiple functions have been proposed for a-syn in neurons [54]; but how endogenous a-syn may function in a cell type constituting a cellular barrier between blood and csf remains unknown. interestingly, exposure of these barrier cells to mn can lead to increased a-syn fluorescent signals within these cells. generally, intracellular protein levels increase as a result of (1) increased protein de-novo synthesis, (2) decreased protein degradation, and/or (3) an altered protein transport mechanism. our qpcr study clearly showed that the levels of mrnas encoding a-syn were not affected by mn exposure in the choroidal z310 cells. while this observation conflicts with previous reports, which show increased a-syn expression following mn exposure in neurons and pc12 cells [40,41], it is possible that different cell types have entirely different responses to mn treatment. in the current study, we did not study the a-syn degradation process. thus, it is unclear whether mn treatment inhibits or promotes the activity of the enzyme system a) *** ### ### *** *** ### ### bates et al. admet & dmpk 3(1) (2015) 15-33 28 that metabolizes a-syn molecules, which could alter intracellular a-syn levels. furthermore, the contrasting results between the decrease in a-syn signal following 48 h of mn exposure in z310 cells and the increase in a-syn signal in primary choroidal epithelia following mn exposure suggests that the ability to degrade asyn aggregates induced by mn exposure differs between the primary cells and the established cell line. additionally, our in vitro studies with primary choroidal cells showed a mn-induced increase of a-syn uptake in the presence of externally added a-syn. thus, it seems likely that mn exposure may act on some yet unknown mechanisms that transport a-syn into the cells. our observation of bright green dots in the plexus tissues appears to suggest an aggregation of a-syn within the plexus cells after mn exposure. this observation is in a good agreement with previous findings in the literature that a direct incubation of a-syn with mn and cu can result in a-syn aggregation [38,39,55,56]. the findings from the thiot assay provide further evidence to support the hypothesis that mn exposure induces a-syn aggregation in the bcsfb. remarkably, however, the aggregation rate of recombinant a-syn in the thiot assay did not increase, but rather decreased, upon concomitant incubation with both mn and cu, when the data were compared to that obtained for a-syn incubated with mn or cu, alone. cu has been shown to bind to multiple sites on the a-syn protein [51,55,57,58] and undergo redox reactions with a-syn directly, by swapping between cu(i) and cu(ii) oxidation states [58]. furthermore, cu has different preferred binding sites to a-syn, depending on its oxidation state [53,55]. these facts suggest that cu can induce a-syn dysfunction in a variety of ways. similar research on mn/a-syn interactions is practically non-existent. it is possible that mn may also have multiple binding sites in a-syn molecules, similar to cu and other metals [51]. in addition, mn might undergo direct redox reactions with a-syn, similar to cu. these potential interactions between metals themselves and among the metals and the protein may explain our in vitro a-syn aggregation data. clearly more research is necessary to depict the underlying mechanisms. the results from our ihc and elisa quantifications of a-syn in rat primary choroidal epithelia following mn exposure showed conflicting results. while the results from the ihc quantitation demonstrated a mninduced a-syn accumulation and significantly increased uptake of a-syn from the external environment (compared to controls), the results from the elisa quantification showed a decrease in a-syn concentrations following mn exposure. the elisa kit, however, is designed to detect a-syn monomer only. the confocal data from the uptake studies revealed a-syn accumulation in mn-treated cells, and the thiot assay clearly showed a-syn aggregation occurring in the presence of mn. therefore, we propose that the decrease in a-syn content after mn treatment determined by elisa may be the result of a loss of a-syn monomer caused by mn-induced aggregation of the protein. the two-chamber transwell model cultured with primary choroidal epithelial cells is a useful model to study the directional transport of substances across the bcsfb, because the monolayer of epithelial cells tends to form the polarized cellular architecture with microvilli protruding to the inner chamber culture medium [42]. the results from our transwell study showed that the bcsfb monolayer was capable of transporting a-syn in either direction without specific preference toward either chamber and that a-syn transport by the bcsfb was unaffected by mn exposure. it is possible that a severe leakage of the monolayer, either due to mn insults or a-syn toxicity, may render the barrier inefficient to passing molecules. however, in both influx and efflux studies, the levels of a-syn in the donor chambers were significantly lower than those in the receiver chambers at nearly all time points; these data appear to suggest an active transport, rather than a passive diffusion, of a-syn from the donor to receiver chamber. this hypothesis deserves further investigation such as by using atp production inhibitors. admet & dmpk 3(1) (2015) 15-33 effects of manganese exposure on expression and transport of -synuclein doi: 10.5599/admet.3.1.159 29 quantification of a-syn levels in chamber cells did not reveal any significant changes after mn exposure. this observation was in contrast to the results with the same primary plexus cells in the uptake study. in the uptake study, there was an increased accumulation of a-syn when the cells were exposed to mn. a plausible reason for this discrepancy is that the primary choroidal cells in the a-syn uptake experiment did not have a compartment on the basolateral side to expel the excess a-syn taken up by the cells. this would lead to excessive accumulation of a-syn in the cells. the primary cells cultured on the membrane of the transwell inserts, however, possess compartments on both sides of the bcsfb monolayer in the transwell system. therefore, the cells could effectively expel the excess a-syn in either direction to the extracellular culture medium. at present, the mechanism by which a-syn is expelled out of these bcsfb cells is unknown. thus, it has become absolutely necessary to investigate how a-syn is transported by the bcsfb and what are the clinical implications if these naturally present transport machineries are altered due to genetic defects, environmental insults, or both. in summary, our findings reveal that the bcsfb expresses a-syn endogenously and is capable of transporting a-syn between the blood and csf. exposure to mn increases the uptake of a-syn in choroid epithelial cells, but does not have any apparent effect on a-syn transport across the bcsfb. the clinical relevance of these observations to mn-induced parkinsonian disorder remains unclear. the current study represents the initial steps that have provided the foundation for further investigation of a-syn transport at the bcsfb in health and disease conditions. competing interests the authors declare that there are no financial or non-financial competing interests. authors’ contributions cb and wz designed experiments; cb, sf, and dy participated in the acquisition of the data. cb drafted and wrote the manuscript. wz and jr commented and revised the manuscript. all authors have read and approved the final version of the manuscript. acknowledgements: this research was supported in part by nih/national institute of environmental health sciences grants ro1-es008146-17. mr. christopher bates has been partially supported by nih/niehs promotion of diversity in health-related research program. references [1] f. kamel, science 341 (2013) 722-723. [2] p. auluck, g. caraveo, s. lindquist, ann. rev. cell dev. biol. 26 (2010) 211-233. [3] d.f. clayton, j.m. george, j. neurosci. res. 58 (1999) 120-129. [4] d. murphy, s.m. rueter, j.q. trojanowski, v.m.y. lee, j. neurosci . 20 (2000) 3214-3220. [5] r.g. perez, j.c. waymire, e. lin, j.j. liu, f. guo, m.j. zigmond, j. neurosci. 22 (2002) 3090-3099. [6] j. burré, m. sharma, t. tsetsenis, v. buchman, m. etherton,t.c. südhof, science 329 (2010) 16631667. [7] s.l. gordon, m.a. cousin, traffic 15 (2014) 245-254. [8] h. lee, s. patel,s. lee, j. neurosci. 2005. 25 (2005) 6016-6024. [9] h. lee, e. cho, k.w. lee, j. kim, s. cho, s. lee, exp. & mol. med. 45 (2013) 1-9. [10] w. zheng, m. aschner, j.f. ghersi-egea, toxicol. appl. pharmacol. 192 (2003) 1-11. [11] c.e. johansen, j.a. duncan, e.g. stopa, a. baird, pharma. res. 22 (2005) 1011-1037. bates et al. admet & dmpk 3(1) (2015) 15-33 30 [12] b. mollenhauer, e. trautmann, b. otte, j. ng, a. spreer, p. lange, f. sixel-döring, m. hakimi, j. vonsattel, r. nussbaum, c. trenkwalder, m.g. schlossmacher, j. neural. transm. 119 (2012) 739746. [13] z. hong, m. shi, k.a. chung, j.f. quinn, e.r. peskind, d. galasko, j. janković, c.p. zabetian, j.b. leverenz, g. baird, t.j. montine, a.m. hancock, h. hwang, c. pan, j. bradner, u.j. kang, p.h. jensen, j. zhang, brain 133 (2010) 713-726. [14] p.g. foulds, p. diggle, j.d. mitchell, a. parker, m. hasegawa, m. masuda-suzukake, d.m.a. mann, d. allsop. sci. rep. 3 (2013) 2540. [15] b. mollenhauer, e. trautmann, p. taylor, p. manniger, f. sixel-döring, j. ebentheuer, c. trenkwalder, m.g. schlossmacher, j. neurosci. lett. 532 (2013) 44-48. [16] h. lee, j. suk, e. bae, j. lee, s.r: paik, s. lee, int. j. biochem. & cell bio. 40 (2008) 1835-849. [17] p. desplats, h. lee, e. bae, c. patrick, e. rockenstein, l. crews, b. spencer, e. masliah, s. lee, pnas 106 (2009) 13010-13015. [18] m. wennstrom, y. surova, s. hall, c. nilsson, l. minthon, f. boström, o. hansson, h.m. nielsen, plos one 8 (2013) e53250. [19] c.a. bates, w. zheng, fluids barriers cns, 2014. (in press). [20] m.r. dewitt, p. chen, m. aschner, biochem. biophys. res. commun. 432 (2013) 1-4. [21] d.m. cowan, w. zheng, y. zou, x. shi, j. chen, f. rosenthal, q. fan, neurotoxicology 30 (2009) 1214-1222. [22] j. bornhorst, c.a. wehe, s. hüwel, u. karst, h.j. galla, t. schwerdtle, j. biol. chem. 287 (2012) 17140-17151. [23] d.p. perl, c.w. olanow, j. neuropathol. exp. neuro. 66 (2007) 675-682. [24] j.p. covy, b.i. giasson, neurotox 32 (2011) 622-629. [25] g. robison, t. zakharova, s. fu, w. jiang, r. fulper, r. barrea, m.a. marcus, w. zheng, y. pushkar, plos one 7 (2012) e48899 [26] t.r. guilarte, m.k. chen, neurotox 28 (2007) 1147-1152. [27] x. wang, g.j. li, w. zheng, exp. biol. med. 233 (2008) 1561-1571. [28] u. dydak, y. jiang, l. long, h. zhu, j. chen, w. li, r.a.e. edden, s. hu, x. fu, z. long, x. mo, d. meier, j. hareziak, m. aschner, j.b. murdoch, w. zheng, env. health perspect. 119 (2011) 219-224. [29] x. fu, y. zhang, w. jiang, a.d. monnot, c.a. bates,w. zheng, toxicol sci 139 (2014) 423-451. [30] z. long, y. jiang, x. li, w. fadel, j. xu, c. yeh, l. long, h. luo, j. harezlak, j.b. murdoch, w. zheng, u. dydak, neurotox 1669 (2014) 1-8. [31] j. crossgrove, w. zheng, nmr biomed 17 (2004) 544-553. [32] b.a. racette. neurotox (2013) [in press]. [33] a. takeda, t. akiyama, j. sawashita, s. okada, brain res. 640 (1994) 341-344. [34] w. zheng, y. jiang, y. zhang, w. jiang, x. wang, d.m. cowan, neurotox 30 (2009) 240-248. [35] r.a. yokel, j.s. crossgrove, b.l. bukaveckas, neurotox 24 (2003) 15-22. [36] c. schmitt, n. strazielle, p. richaud, a. boubon, j.f. ghersi-egea, j. neurochem. 117 (2011) 747-756. [37] a. takeda, j. sawashita, s. okada, brain res. 695 (1995) 53-58. [38] v.n. uversky, j. li, a.l. fink, febs letters 500 (2001) 105-108. [39] a. binolfi, r.m. rasia, c.w. bertoncini, m. ceolin, c. zweckstetter griesinger, t.m. jovin, c.o. fernândez, j. am. chem. soc. 128 (2006) 9893-9901. [40] t. cai, t. yao, g. zheng, y. chen, k. du, y. cao, x. shen, j. chen, w. luo, brain res. 1359 (2010) 201207. [41] y. li, l. sun, t. cai, y. zhang, s. lv, y. wang, l. ye, brain res. bull. 81 (2010) 428-433. [42] w. zheng, q. zhao, methods mol. biol. 188 (2002) 99-114. admet & dmpk 3(1) (2015) 15-33 effects of manganese exposure on expression and transport of -synuclein doi: 10.5599/admet.3.1.159 31 [43] g.j. li, q. zhao, w. zheng, toxicol. appl. pharmacol. 205 (2005) 188-200. [44] x. wang, g.j. li, w. zheng, brain res. 1097 (2006) 1-10. [45] w. zheng, q. zhao, j.h. graziano, in vitro cell dev. biol. 34 (1998) 40-45. [46] l.z. shi, w. zheng, brain res. 1057 (2005) 37-48. [47] a.d. monnot, m. behl , s. ho, w. zheng, toxicol. appl. pharmacol. 256 (2011) 249-257. [48] g. zheng, j. chen, w. zheng, toxicol. appl. pharamcol. 260 (2012) 285-293. [49] h. zhang, a. griggs, j.c. rochet, l.a. stanciu, biophys. j. 104 (2013) 2706-2713. [50] h. lee, f. khoshaghideh, s. patel, s. lee, j. neurosci. 24 (2004) 1888-1896. [51] y. lu, m. prudent, b. fauvet, h.a. lashuel, h.h. girault, acs chem. neurosci. 2 (2011) 667-675. [52] t. bartels, j.g. choi, d.j. selkoe, nature 477 (2011) 107-111. [53] b. fauvet, m.m. kamdem, m.b. fares, c. desobry, s. michael, m.t. ardah, e. tsika, p. coune, m. prudent, n. lion, d. eliezer, d.j. moore, b. schneider, p. aebischer, o.m. el-agnaf, e. masliah, h.a. lashuel, j. biol. chem. (2012) 1-35. [54] d. kuzdas-wood, n. stefanova, k.a. jellinger, k. seppi, m.g. schlossmacher, w. poewe, g.k. wenning, prog. neurobiol. 118 (2014) 19-35 [55] a. binolfi, a.a. valiente-gabioud, r. duran, m. zweckstetter, c. griesinger, c.o. fernandez, j. am. chem. soc. 133 (2011) 194-196. [56] t.r. guilarte, frontiers in aging neuroscience 5 (2013) 1-10. [57] a. ahmad, c.s. burns, a.l. fink, v.n. uversky, j.biomol. struc.dyn. 29 (2013) 825-842. [58] f. camponeschi, d. valensin, i. tessari, l. bubacco, s. dell’acqua, l. casella, e. monzani, e. gaggelli, g. valensin, inorg. chem. 52 (2013) 1358-1367. bates et al. admet & dmpk 3(1) (2015) 15-33 32 supplemental fugures figure s1. diagram of two-chamber transwell transport system. the primary choroidal epithelial cells are cultured on a porous member in the inner chamber; the inner chamber is inserted in the outer chamber. the monolayer epithelial cells have the tight junctions between each other, which forms a tight barrier between in the inner and outer chambers. an efflux experiment is illustrated by adding a-syn into the inner chamber and quantifying a-syn in the outer chamber. control 25 μm mn 50 μm mn 100 μm mn 1 2 h o u rs 2 4 h o u rs 4 8 h o u rs a) add a-syn into inner chamber determine a-syn in the outer chamber inner chamber outer chamber cells porous filter laminin admet & dmpk 3(1) (2015) 15-33 effects of manganese exposure on expression and transport of -synuclein doi: 10.5599/admet.3.1.159 33 figure s2. dose-time responses of mn effect on a-syn expression in primary choroidal epithelial cells in the absence or presence of a-syn. (a). ihc staining of a-syn in primary plexus cells without added a-syn. cells were treated with various concentrations of mn or saline (control) for 12, 24, or 48 h and stained for a-syn. (b). cells were treated with various concentrations of mn or saline for 12, 24, or 48 h, followed by incubation with 0.5 μm recombinant human asyn for 2 h. cells were stained with primary anti-a-syn. a typical study of n = 4 is presented. (c). confocal quantification of a-syn. for each time point, the data were compared between cells with or without added a-syn. data represent mean ± sd. n = 10-24. *: p < 0.01, **: p < 0.001 as compared to the data at the same time point without added a-syn. 1 2 h o u rs 2 4 h o u rs 4 8 h o u rs control + a-syn 25 μm mn + a-syn 50 μm mn + a-syn 100 μm mn + a-syn b) c) ** ** ** ** ** ** ** * manuscript doi: 10.5599/admet.3.3.195 242 admet & dmpk 3(3) (2015) 242-253; doi: 10.5599/admet.3.3.195 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index review animal models for alzheimer’s disease: a focused review of transgenic rodent models and behavioral assessment methods xiaoli wu, jing li, wei zhou, kin tam* faculty of health science, university of macau, macau, china *corresponding author: e-mail: kintam@umac.mo; tel.: +853-8822-4988; fax: +853-8822-2314 received: june 18, 2015; revised: august 20, 2015; published: september 05, 2015 abstract with the increased morbidity and unclear etiology of alzheimer’s disease (ad), there is an urgent nee d to put more effort to investigate the causes of the disease and develop novel drug to stop or reverse the disease progression. transgenic rodent models mimicking different types of ad-like pathologies are essential resources to discover potential drug targets and study the mechanisms of drug actions. the common symptom of ad is the cognitive deficits. the ultimate readout for any interventions should be evaluated by the test of learning and memory. although a multiply number of rodent models and behavioural assessment methods have been widely utilized in mechanism studies and screening of novel drug candidates, large variability still exists in the methodologies, especially in terms of how the rodent models are being utilized. to select suitable and valid models for supporting ad research, it is important to understand the characteristics and applicability of the rodent models and behavioural assessment methods. this review seeks to summarize and discuss the pathological feature of some transgenic rodent models that are commonly used in ad research (e.g. app, ps1/2 and tau gene mutations). moreover, the characteristics and applicability of some behavioural assessment methods (e.g. morris water maze and radial arm water maze) will be summarized. finally, we will discuss the applicability of these models and methods in ad research. keywords neurodegeneration; cognitive impairment; transgenic mice; behavioural alteration; morris water maze; radial arm water maze. 1. introduction alzheimer’s disease (ad) is a neurodegenerative dysfunction with the most common early symptom as memory loss. two histopathological hallmarks of ad are usually observed, which include: (1) extracellular amyloid plaques primary comprised 39 to 42 amino acids peptides or polymers that produced by proteolytic cleavage of amyloid precursor protein (app) with the beta and gamma secretases, and (2) intracellular neurofibrillary tangles (nfts) comprised of hyperphosphorylated tau protein aggregated in oligomeric structures. the presence and/or accumulation of these misfolded proteins are thought to be the reasons for neuronal apoptosis and synaptic deficits, which eventually lead to systemic cognitive impairments [1]. the incidence of ad exhibits a continuous growth in recent decades, partly due to the aging population http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:kintam@umac.mo admet & dmpk 3(3) (2015) 242-253 animal models for alzheimer’s disease doi: 10.5599/admet.3.3.195 243 as the incidence may increases as the percentage of people older than 65 increases [2]. it was predicted that there will be 13.2 million people suffer from this neurodegenerative dysfunction by 2050 [3]. accordingly, the worldwide societal cost of dementia especially ad was estimated to break through $422 billion in 2009 [4]. as incidence and cost are forecasted to rise at an alarming speed in the next decades, intensive and meticulous research on ad are urgently needed. despite decades of pathological studies on ad, the etiology and mechanism of ad are still unknown, and many basic questions also remain unanswered. continuing research into the underlying mechanism of ad as well as renewed effort in seeking for disease-modifying drugs are essential to address this problem. rodents have been extensively applied in ad research on account of the relatively high similarity in physical structure and cognitive system, as well as the availability and relatively low cost in comparison with primate system. drugs targeting aspects related to the pathophysiological mechanisms and diseasemodifying therapies are usually tested in rodent models before being advanced to clinical trials in humans. indeed, the most striking characteristic of ad is the cognitive impairment. as a consequence, the crucial aim for therapy is to prevent and/or ameliorate the cognitive dysfunction. in particular, the cognitive assays provide the advantage of targeting cognitive functions without the requirement of pathological hypothesis. so far, a multiply number of rodent models and behavioural assessment methods have been widely utilized in mechanistic studies and screening of novel drug candidates. however, large variability still exists in the methodologies and how the rodent models are utilized. hence, it is necessary to understand the characteristics and applicability of the rodent models and behavioural assessment methods for choosing suitable and valid models. this review summarized the pathological characteristics and the applicability of transgenic rodent models with app, presenilin 1/2 (ps1/2) and tau mutations mimicking the familial ad (fad) models. given the close relationship between rodent models and behavioural assessment methods, this review discussed the features, influential factors and applicability of two classical behavioural assessment methods including morris water maze (mwm) and radial arm water maze (rawm) which may be used to study the learning and spatial working memory as well as to assess damage on cerebral cortex and hippocampus. since hippocampus is closely related to cognitive function where ad is thought to initiate and develop, the hippocampus-dependent cognitive tests are ideally-suited for ad search to maximize the possibility of selecting a target or drug that is relevant to cognitive function in vivo. in brief, this review will summarize some transgenic rodent models and behavioural assessment methods that are commonly used in ad research. then we will discuss the applicability of these rodent models and behavioural assessment methods in ad research. 2. alzheimer’s disease pathology and the relevant transgenic mice the pathology of ad still remains unclear so far, although there are several hypotheses about ad’s pathology. among these hypotheses, aβ hypothesis and tau hypothesis appeared to be widely accepted [5]. amyloid plaques and nfts are the two striking hallmarks of ad. aβ and tau protein are the primary constituents of amyloid plaques and nfts, respectively. more importantly, mutations in app result in ad with 100 % penetration, and fad-related mutations of app bring about an enhancive formation or aggregation of aβ [6,7]. alternatively, the illumination of mutated tau protein in ftdp-17 definitively proved that the dysregulation of tau can bring about neurodegeneration and eventually leading to ad [8]. evidence of the abnormal aggregation of the aβ peptides and highly phosphorylated tau protein exemplified a crucial pathogenic characteristic of this disease [9,10]. all these findings support the aβ kin tam et al. admet & dmpk 3(3) (2015) 242-253 244 hypothesis and tau hypothesis. given the importance of having a validated animal model for drug discovery and mechanism research, a multiple number of mutations in the genes of app, ps1/2 and tau had been discovered in familial ad (fad) accordingly. although more than 95 % ad cases belong to late-onset ad (load), and fad only account for less than 5 % of the total ad cases. there exists very high phenotypic similarity between fad and load, suggesting that the knowledge about mechanism obtained from fad will also be directly relevant for load. therefore, deeper insight into the investigation about ad using transgenic rodent models is critically important. original transgenic mouse strain were developed to study ad carrying familial mutations of app [11,12], while subsequent models were relied on ps mutations [13], tau mutations [14], or a combination of 2 or 3 mutations [15]. 2.1. aβ hypothesis in ad’s process, amyloid plaques are the striking hallmark presenting in the autopsy of ad patient brain tissues. in many rodent models including transgenic and non-transgenic rodents, high levels of aβ peptides (> nanomolar) can dramatically weaken neuronal physiology and synaptic density, which were also distinctive feature of ad [12]. more importantly, excess generation of aβ peptides induced cognitive dysregulation impairing the animals’ performance in learning and memory tasks [16,17]. these findings supported the importance of aβ hypothesis in ad pathology. 2.1.1 transgenic rodent models with app or/and ps mutations serving as the fundamental principles for many genetically modified ad rodent models as well as responsible for the aβ hypothesis, a large amount of rodent models with app and ps1/2 gene mutations had been developed. all of these transgenic models were manifested to increase the ratio of aβ 42 to aβ 40 and enhance the aβ aggregation that resulted in higher levels of amyloid plaques [18]. so far, more than 30 app mutations and nearly 200 ps1/2 mutations had been authenticated and linked with ad [19]. as shown in table 1, mice from the strains with app gene mutations all display amyloid plaques at 2-9 months old, and the cognitive impairments are observed at 2-6 months old. unfortunately, nfts are not detected in these mice. it is worth noting that the strains with ps gene mutation lack of most of the ad-like phenotype except synaptic deficits. since app and ps1/2 mutations are important in the development of ad, combining app and ps mutations lead to an accelerated ad-like phenotype. prominent amyloid plaques depositions are observed as early as at 2-3 months along with a striking increased production of aβ 42 levels at 1-2 months old in these mice, which is developed before cognitive impairment appearing at 6 months old. 2.1.2 application in alzheimer’s disease these transgenic rodent models with app and ps mutations have been utilized in mechanism investigation and drug candidate discovery successfully. some therapeutic vaccine and antibodies had been developed using the pdapp mice model [29-31]. the tg2576 mice were used to investigate the role of reactive oxygen species on the cerebral amyloid angiopathy, which was an important cause of cognitive dysfunction in elderly patients with and without ad [32]. surprisingly, some anti-ulcer and anti-tumor drugs were found to be effective to ameliorate the amyloid-like pathology in app23 mice. since the safeties of these drugs have been confirmed in human, these drugs can be considered as candidates for the prevention of ad [33, 34]. in addition, some transgenic mice models were used to explore the novel pathway or therapeutic target for the treatment of ad pathogenesis. for example, an extracellular matrix admet & dmpk 3(3) (2015) 242-253 animal models for alzheimer’s disease doi: 10.5599/admet.3.3.195 245 protein called reelin, could delay delays amyloid plaque and rescue the recognition memory deficits in j20 mice [35]. so the reelin pathway deserved consideration as a novel therapeutic target for ad pathogenesis [35]. table 1. transgenic rodent models with app or/and ps mutations ` pdapp tg2576 app23 j20 ps1 psapp 5xfad mutation(s)/transgene app (indiana) app (swedish) app (swedish) app (swedish and indiana) ps1 (m146l) ps1 (m146l) and app (swedish) ps1 and app outcome enhanced cleavage by gamma-secretase; increased aβ 42:40 ratio; tau hyperphosphorylation enhanced cleavage by beta-secretase; expansive amyloid plaques deposition obvious amyloid plaques and cerebral amyloid angiopathy combination of effects on app processing/aβ enhanced aβ 42:40 ratio; dysfunction of intracellular calcium; without cognitive deficits accelerated phenotype and pathology; without formation of nfts accelerated amyloid plaques; increased production of aβ 42 levels; without formation of nfts amyloid plaques (age) yes (6–9) yes (9) yes (6) yes (6) none yes (3) yes (2) nfts (age) no no no no no no no neuronal loss (age) no no yes (14-18) no no no yes synaptic deficits (age) yes yes yes yes (2-4) yes yes yes memory deficits (age) yes (6) yes (4-6) yes (3) yes (2-4) na yes yes references [20] [21, 22] [23, 24] [24] [25] [26] [27, 28] 2.1.3 brief summary of transgenic mice with app or/and ps mutations transgenic rodent strains with app and /or ps mutations have obtained important information about aβ pathology in ad and strongly support the aβ hypothesis. however, these transgenic rodent strains do not display nfts or prominent neurological alterations that commonly observed in ad, making them incomplete ad models. both pathological changes and behavioral impairments are crucial judgment criteria of ad. because there exists a free clearance capacity for amyloid plaques in vivo, only a small amount of amyloid plaques deposition do not mimic the complete ad model. even so, these rodent models with app and/or ps mutations are considered as ideal models of aβ pathology and have their vital part in reflecting the effects of anti-aβ therapies on aβ dynamics. 2.2. tau hypothesis despite considerable attention had been paid to the aβ in ad research and drug discovery, more and more evidences demonstrated that the abnormal aggregation of tau protein also played a vital mediating role in the development of ad. this is referred to as the tau hypothesis [36]. recent study revealed that the reduced expression of tau lessened the neurotoxicity in aβ-treated cells and animal models [37, 38]. these results indicated that tau hypothesis was as important as aβ hypothesis in the rationalization of the neuropathology of ad. 2.2.1 transgenic rodent model with tau mutation tauopathy is a pathological change in ad with the pathological aggregation of tau protein named nfts in hippocampus of brain. the rate of neuronal loss was much higher than that of nfts, indicating that there might be a relevant mechanism between nfts formation and neuronal apoptosis [39]. the tau hypothesis was conclusively supported by the discovery of a tau gene mutation in chromosome 17 (ftdp-17) in frontotemporal dementia (ftd), elucidating that tau dysfunction or abnormality alone independently kin tam et al. admet & dmpk 3(3) (2015) 242-253 246 induced cognitive impairments and neurodegeneration [14]. as shown in table 2, mice from the strains with tau gene mutation all display nfts at 4.5-6.5 months old as well as neuronal loss and synaptic deficits. table 2. transgenic rodent model with tau mutation. characterization jnpl3 rtg4510 htau mutation(s)/transgene tau (p301l) tau (p301l inducible) human tau outcome nfts in several regions of the brain and spinal cord; without amyloid plaques and cognitive deficits nfts and neuronal apoptosis in the ca1 zone of the hippocampus; without amyloid plaques nfts and neuronal loss in the neocortex and hippocampus similar with human; without amyloid plaques amyloid plaques (age) none none none nfts (age) yes (4.5-6.5) yes (4.5-5.5) yes (6) neuronal loss (age) yes yes yes (15) synaptic deficits (age) yes yes yes memory deficits (age) na yes yes (12) references [40] [41] [42, 43] 2.2.2 application in alzheimer’s disease many transgenic rodent models with tau mutations have been developed and used in the pathogenesis research and drug discoveries, for example, p301l, rtg4510 and htau mice. the p301l mouse model was utilized to investigate the distribution of tau protein, finally to found that regulating the synaptosomal tau level might be a potential target for a therapeutic intervention directed at preventing neurodegeneration [44]. in addition, calpastatin was found to inhibit the activity of calpain to alleviate the taupathy in ad using p301l mice [45]. interestingly, it was reported that methylene blue could ameliorate tau-related neurodegeneration prevented behavioural deficits and reduce soluble tau levels in the brain rtg4510 mice [46]. however, it could not dissolve existing neurofibrillary tangles in rtg4510 mice [46]. furthermore, the rtg4510 mice were used to conduct the longitudinal evaluation of blood-brain barrier (bbb), finally to found that bbb was damage with progressive igg, t cell and red blood cell infiltration as the mice grow older [47]. 2.2.3 brief summary of transgenic mice with app and ps mutations the transgenic mice with tau mutation can display nfts which is one of the important hallmarks of ad. unfortunately, amyloid plaques pathology has not been detected, making them as limited models of ad. yet, these transgenic rodent models are ideally-suited for the investigation of tau dysregulation and neuronal apoptosis in ad. 2.3. transgenic rodent model with app, ps1/2 and tau mutations as shown in table 3, the multiple transgenic rodent models display accelerated ad-like phenotype with both amyloid plaques and nfts, which do not exist in other transgenic rodent models simultaneously. the first transgenic rodent strain with mutations both in app and tau is the tapp mouse, which is developed by the crossing with the tg2576 strain and the jnol3 strain [15]. these mice display amyloid plaques at 9 months old, which is similar to tg2576 mice in their distribution, development and severity. another transgenic rodent strain is the 3xtg-ad mouse, which is developed by simultaneously inserting swedish app mutation and the p301l tau mutation into ps1 mice [48]. these mice display relatively complete pathological changes and behavioural impairments including amyloid plaques, nfts, neuronal apoptosis, reduced synaptic density and cognitive deficit. in the first instance, soluble aβ is observed within neurons at 3–4 months old in the neocortex and at 6 months old in the hippocampus. amyloid plaques are developed at about 6 months old, and learning deficits occur prior to the formation of amyloid plaques at admet & dmpk 3(3) (2015) 242-253 animal models for alzheimer’s disease doi: 10.5599/admet.3.3.195 247 about 4.5 months old [49]. finally, nfts are developed in the hippocampus and amygdale regions at about 12 months old. table 3. transgenic rodent model with app, ps1/2 and tau mutations. 2.3.1 application in alzheimer’s disease these multiple transgenic mice models also have been utilized in the alzheimer’s disease research widely, since they can display both aβ plaques and nfts simultaneously. firstly, the happ mice, were used to evaluate the effect of o-linked n-acetyl glucosamine on ad [54]. the result showed that o-linked nacetyl glucosamine could improve the performance in the morris water maze, decrease amyloid plaques levels, which provided good support for o-linked n-acetyl glucosamine as a promising therapeutic target to alter disease progression in ad [54]. the 3xtg-ad mouse model is a more popular transgenic model in the research of ad. the old 3xtg-ad mice (21-24 months old) were used to evaluate the neuroprotective effect of quercetin (a flavonoid generally found in fruits and vegetables, such as onions and apples, and red wine) on ad [55]. the results showed that quercetin could decrease the extracellular β-amyloidosis, tauopathy, astrogliosis and microgliosis in the hippocampus and the amygdala, as well as improving the performances both in morris water maze and elevated plus maze [55]. in addition, this mouse model was used to investigate the effect of the endogenous tau on nfts and cognitive deficits by crossing 3xtg-ad with mtauko mice to obtain a novel transgenic strain, named 3xtg-ad/mtauko [11]. the data showed that the endogenous tau was contributed to the generation of nfts, but it could not affect the change of cognitive capacity [11]. moreover, some publications reported the age-dependent differences (reference and working memory deficits [56], amyloid plaques levels and nfts [57], synaptic dysfunction [58], mitochondrial dysfunction [59] and neurogenesis damage [60]) of this transgenic mouse model to illuminate the pathology development process. 3. behavioural assessment methods the hippocampus and the surrounding regions belong to limbic system, which is responsible for a variety of activities such as learning, memory and emotional behavior. in particular, hippocampus had been demonstrated to be critical for spatial memory and emotional behavior which were closely related to cognitive deficit in ad [15]. moreover, hippocampus was easy to damage at the earliest stages of ad. therefore, it was considered as a key region to understand the disease pathophysiology comprehensively [61]. indeed, most of the behavioral tasks are designed to assess the hippocampal-dependent memory. so far, many behavioral tasks have been developed to evaluate the behavioral alterations in ad. 3.1. morris water maze (mwm) the mwm [62] is a behavioural task that depends on hippocampus for assessing spatial learning and long-term memory (ltm) in rodents [63,64]. this task has been extensively utilized in behavioural characterization tapp 3xtg-ad mutation(s)/transgene app (swedish) and tau (p301l) ps1, app, and tau outcome accelerated phenotype and pathology; without cognitive deficits accelerated phenotype and pathology amyloid plaques (age) yes (9) yes (6) nfts (age) yes yes (12) neuronal loss (age) yes yes synaptic deficits (age) yes yes memory deficits (age) na yes (4.5) references [50, 51] [52, 53] kin tam et al. admet & dmpk 3(3) (2015) 242-253 248 assessment on account of its simple handling, high repeatability and reliability. mwm has been successfully utilized in testing the learning and memory capacities reflecting the therapeutic effect of novel drugs [64]. considering the widespread use of mwm in the neuroscience and psychology for studying spatial learning and memory, a mathematical model was established to understand the exact mnemonic and navigational demands of the task [65]. this model could identify different parameter values and suggest the activation of different neuronal pathways [65]. 3.1.1 mwm in the development of ad model a multiple of animal model have been developed to mimic the pathology of ad for the drug screening and mechanism research. considering that the importance of spatial learning and memory deficit in the criterion of ad, mwm has been applied in the judgement of many potential ad animal model. mwm was used to test the cognitive deficits of a novel animal model of ad, the hemizygous transgenic mcgill-r-thy1app (tg+/−) rat [66]. the data revealed that this transgenic rat exhibited spatial memory deficit in the mwm as early as 3 months old, which persisted at 6 and 12 months old when compared to wild-type rats [66]. in addition, mwm was used to test different ages of b6c3-tg (appswe/psen1de9) double-transgenic mice to provide the evidence as an ad model [67]. 3.1.2 mwm in novel drug candidate discoveries spatial memory deficit is a prominent feature of ad, and mwm is a standard task to test the spatial memory deficit [65]. so mwm has been applied in the novel drug candidate discoveries widely. for example, mwm was used to evaluate the effect of silibinin on ad, and the result showed that silibinin could improve the performance of app/ps1 transgenic mice after dosing with silibinin [68]. and the mechanism revealed that silibinin might act as a dual-target drug for the treatment of ad by inhibiting amyloid β peptide aggregation and the acetylcholinesterase activity [68]. for another example, lycopene was found to abrogate neuroinflammatory cascade in intra cerebroventricular injection of β-amyloid1–42induced learning and memory impairment mice using mwm [69]. similarly, a lipid amide named palmitoylethanolamide, was proved to decrease the escape latency in mwm task, which suggested to be a potential drug candidate not just to alleviate the symptoms but also to modify disease progression [70]. 3.1.3 mwm in the pathological mechanism research of ad although a lot of endeavours were made to the pathological research of ad, the mechanism is still unclear so far. mwm has been applied in the investigation of ad pathogenesis widely. it was reported that mwm was used to test the cognitive function in an estrogen receptor alpha (erα) knockout mice with aβ intracerebroventricular injection to understand the effect of estrogen in the development of ad [54]. furthermore, in the investigation of the relation between hippocampal long-term depression and spatial learning, mwm was utilized to test the spatial learning capacity [71]. 3.2. radial arm maze (ram) ram, is also a behavioural task depends on hippocampus for assessing spatial learning in rodents [72]. compared with mwm, ram is suitable for the assessment of stm concluding working memory and reference memory. the ram is a hippocampal-dependent task, where food-deprived mice must learn to locate maze arms baited with food rewards [73]. during this task, mice have to enter the arms to find food. then, each visit (entry of the full body, excluding the tail, into an arm) is scored as a correct visit or a wrong visit [73]. a correct visit was defined as an entry into a baited arm that had not been visited earlier in the same trial [73]. finally, the “wrong visits” were analysed as working memory errors and reference memory admet & dmpk 3(3) (2015) 242-253 animal models for alzheimer’s disease doi: 10.5599/admet.3.3.195 249 errors to determinate the basis of learning. a working memory error was determined as re-entering the baited arm that the rat already had visited. and a reference memory error was scored as entries into an arm that was not baited at first [74]. 3.2.1 ram in novel drug candidate discoveries similar to mwm, ram also has been applied in the novel drug candidate discoveries broadly. in the effect evaluation of a tumor necrosis factor-α inhibitor on ad, ram was used to test the learning and memory function of the triple transgenic mice after 10 weeks of treatment of this inhibitor [75]. for another example, curcumin was prepared into nanoparticles to alleviate the ad pathology of tg2576 mice, and ram was used to test cognitive performance of these mice after orally administered with curcumin nanoparticles for 3 months [76]. 3.2.2 ram in the pathological mechanism research of ad ram is also utilized in the pathological mechanism research of ad as wmw. for instance, ram was used to test the memory function of endothelial nitric oxide synthase deficient mice to find out the relation between endothelial nitric oxide and amyloid precursor protein [77]. and the results suggested that chronic loss of endothelial nitric oxide might be a crucial contributor to both aβ related pathology and cognitive deficit [77]. in addition, in the investigation of the effect of gabaa α5 positive allosteric modulators on ad, ram was used to evaluate the spatial learning and memory capacities change after dosing with these modulators [78]. furthermore, in the study of the relation between toll-like receptor 9 signalling and ad pathology, ram was used to test the spatial cognitive function in tg2576 ad model transgenic mice [79]. 4. future direction a multiple number of transgenic rodent models and behavioural assessment methods have been established to investigate the underlying mechanisms and screen novel drug candidates in ad. the evaluation criterion of therapeutic effects is based on the pathological and behavioural alterations. however, the high-profile failures of novel drugs in clinical trials strongly call for the need of choosing suitable and correct transgenic rodent models and behavioural assessment methods. generally speaking, a good model means that it can truly mimic the relevant aspects of ad including aetiology, symptomatology, treatment and physiological basis. unfortunately, none of the models satisfies all the aspects of the ad pathologies. the models discussed above represent some specific aspects of ad pathologies that may play a crucial role in the assessment of therapeutic effects of novel drugs. these models can also attribute to the research of underlying pathophysiological mechanisms of ad. indeed, it is difficult to conduct the evaluation process clinically on account of the morality, ethics and law. accordingly, rodent models may help to establish the relationship between drug candidates and human. moreover, behavioural assessment methods are also important in drug discoveries. behavioural assessment may serves as readout without the requirement of the disease pathogenesis that might be disproved by future studies. undoubtedly, behavioural assessment can offer useful information on the efficacy and validity of the drug candidates. however, it is insuperable to break out the species barrier between animal and human. nevertheless, rodent models are still essential for accessing ad-like pathology in vivo. in spite of many deficiencies in rodent model and behavioural assessment method, it is acceptable that rodent model is irreplaceable. ideally, at least three or more rodent models and three or more behavioural assessment methods are needed to evaluate novel drugs. for instance, if a compound is kin tam et al. admet & dmpk 3(3) (2015) 242-253 250 designed for inhibiting the aggregation of aβ, then the transgenic rodent models displaying increased aβ levels (e.g. app mice, ps1 mice and tg2576 mice, etc.) might be selected for studies. on the other hands, if a compound is designed for inhibiting the nfts, then the transgenic mice models displaying tau dysfunction (e.g. jnpl3 mice, tapp mice and 3xtg-ad mice, etc.) might be selected for studies. obviously some deficits still exist in the current transgenic rodent models. it is beneficial to apply a rodent model expressing relatively complete pathological changes and behavioural impairments (e.g. the 3xtg-ad model, see table 1) for ad study. acknowledgements: we thank the financial support from the science and technology development fund, macao s.a.r (fdct) (project reference no.: macau fdct project 118/2013/a3). references [1] s.t. ferreira, m.n. vieira, f.g. de felice, iubmb life 59 (2007) 332-345. [2] c.p. ferri, m. prince, c. brayne, h. brodaty, l. fratiglioni, m. ganguli, k. hall, k. hasegawa, h. hendrie, y. huang, the lancet 366 (2006) 2112-2117. [3] l.e. hebert, p.a. scherr, j.l. bienias, d.a. bennett, d.a. evans, archives of neurology 60 (2003) 11191122. [4] a. wimo, b. winblad, l. jönsson, alzheimer's & dementia 6 (2010) 98-103. [5] c.m. wischik, c.r. harrington, j. storey, biochemical pharmacology 88 (2014) 529-539. [6] m. citron, t. oltersdorf, c. haass, l. mcconlogue, a.y. hung, p. seubert, c. vigo-pelfrey, i. lieberburg, d.j. selkoe, nature 360 (1992) 672-674. [7] k.m. felsenstein, l.w. hunihan, s.b. roberts, nature genetics 6 (1994) 251-256. [8] m.g. spillantini, j.c. van swieten, m. goedert, neurogenetics 2 (2000) 193-205. [9] c. ballard, s. gauthier, a. corbett, c. brayne, d. aarsland, e. jones, the lancet 377 (2011) 1019-1031. [10] y. huang, l. mucke, cell 148 (2012) 1204-1222. [11] d. baglietto-vargas, m. kitazawa, e.j. le, t. estrada-hernandez, c.j. rodriguez-ortiz, r. medeiros, k.n. green, f.m. laferla, neurobiology of disease 62 (2014) 407-415. [12] h.a. born, j.-y. kim, r.r. savjani, p. das, y.a. dabaghian, q. guo, j.w. yoo, d.r. schuler, j.r. cirrito, h. zheng, the journal of neuroscience 34 (2014) 3826-3840. [13] k. duff, c. eckman, c. zehr, x. yu, c.-m. prada, j. perez-tur, m. hutton, l. buee, y. harigaya, d. yager, nature 383 (1996) 710-713. [14] j. lewis, e. mcgowan, j. rockwood, h. melrose, p. nacharaju, m. van slegtenhorst, k. gwinn-hardy, m.p. murphy, m. baker, x. yu, nature genetics 25 (2000) 402-405. [15] j. lewis, d.w. dickson, w.-l. lin, l. chisholm, a. corral, g. jones, s.-h. yen, n. sahara, l. skipper, d. yager, science 293 (2001) 1487-1491. [16] g.m. shankar, s. li, t.h. mehta, a. garcia-munoz, n.e. shepardson, i. smith, f.m. brett, m.a. farrell, m.j. rowan, c.a. lemere, nature medicine 14 (2008) 837-842. [17] d.m. watterson, v.l. grum-tokars, s.m. roy, j.p. schavocky, b.d. bradaric, a.d. bachstetter, b. xing, e. dimayuga, f. saeed, h. zhang, plos one 8 (2013) e66226. [18] o. murayama, t. tomita, n. nihonmatsu, m. murayama, x. sun, t. honda, t. iwatsubo, a. takashima, neuroscience letters 265 (1999) 61-63. [19] m. cruts, j. theuns, c. van broeckhoven, human mutation 33 (2012) 1340-1344. [20] d. games, d. adams, r. alessandrini, r. barbour, p. berthelette, c. blackwell, t. carr, j. clemens, t. donaldson, f. gillespie, et al., nature 373 (1995) 523-527. [21] k. hsiao, p. chapman, s. nilsen, c. eckman, y. harigaya, s. younkin, f. yang, g. cole, science (new york, n.y.) 274 (1996) 99-102. admet & dmpk 3(3) (2015) 242-253 animal models for alzheimer’s disease doi: 10.5599/admet.3.3.195 251 [22] j.a. saydoff, a. olariu, j. sheng, z. hu, q. li, r. garcia, j. pei, g.y. sun, r. von borstel, journal of alzheimer's disease : jad 36 (2013) 637-657. [23] c. sturchler-pierrat, d. abramowski, m. duke, k.h. wiederhold, c. mistl, s. rothacher, b. ledermann, k. burki, p. frey, p.a. paganetti, c. waridel, m.e. calhoun, m. jucker, a. probst, m. staufenbiel, b. sommer, proceedings of the national academy of sciences of the united states of america 94 (1997) 13287-13292. [24] m.e. calhoun, k.-h. wiederhold, d. abramowski, a.l. phinney, a. probst, c. sturchler-pierrat, m. staufenbiel, b. sommer, m. jucker, nature 395 (1998) 755-756. [25] k. duff, c. eckman, c. zehr, x. yu, c.-m. prada, j. perez-tur, m. hutton, l. buee, y. harigaya, d. yager, d. morgan, m.n. gordon, l. holcomb, l. refolo, b. zenk, j. hardy, s. younkin, nature 383 (1996) 710713. [26] l. holcomb, m.n. gordon, e. mcgowan, x. yu, s. benkovic, p. jantzen, k. wright, i. saad, r. mueller, d. morgan, s. sanders, c. zehr, k. o'campo, j. hardy, c.-m. prada, c. eckman, s. younkin, k. hsiao, k. duff, nature medicine 4 (1998) 97-100. [27] h. oakley, s.l. cole, s. logan, e. maus, p. shao, j. craft, a. guillozet-bongaarts, m. ohno, j. disterhoft, l. van eldik, r. berry, r. vassar, the journal of neuroscience: the official journal of the society for neuroscience 26 (2006) 10129-10140. [28] s.d. girard, m. jacquet, k. baranger, m. migliorati, g. escoffier, a. bernard, m. khrestchatisky, f. feron, s. rivera, f.s. roman, e. marchetti, hippocampus 24 (2014) 762-772. [29] f. bard, m. fox, s. friedrich, p. seubert, d. schenk, g.g. kinney, t. yednock, experimental neurology 238 (2012) 38-43. [30] z. luo, j. li, n.r. nabar, x. lin, g. bai, j. cai, s.f. zhou, c. cao, j. wang, journal of neuroimmune pharmacology : the official journal of the society on neuroimmune pharmacology 7 (2012) 640-655. [31] s. liu, d. shi, h.c. wang, y.z. yu, q. xu, z.w. sun, scientific reports 5 (2015) 7771. [32] b.h. han, m.l. zhou, a.w. johnson, i. singh, f. liao, a.k. vellimana, j.w. nelson, e. milner, j.r. cirrito, j. basak, m. yoo, h.h. dietrich, d.m. holtzman, g.j. zipfel, proceedings of the national academy of sciences of the united states of america 112 (2015) e881-890. [33] t. hoshino, k. suzuki, t. matsushima, n. yamakawa, t. suzuki, t. mizushima, plos one 8 (2013) e76306. [34] p. he, x. cheng, m. staufenbiel, r. li, y. shen, plos one 8 (2013) e55091. [35] l. pujadas, d. rossi, r. andres, c.m. teixeira, b. serra-vidal, a. parcerisas, r. maldonado, e. giralt, n. carulla, e. soriano, nature communications 5 (2014) 3443. [36] j. avila, j.j. lucas, m. perez, f. hernandez, physiological reviews 84 (2004) 361-384. [37] l.m. ittner, y.d. ke, f. delerue, m. bi, a. gladbach, j. van eersel, h. wölfing, b.c. chieng, m.j. christie, i.a. napier, cell 142 (2010) 387-397. [38] e.d. roberson, k. scearce-levie, j.j. palop, f. yan, i.h. cheng, t. wu, h. gerstein, g.q. yu, l. mucke, science 316 (2007) 750-754. [39] m.-c. chartier-harlin, f. crawford, h. houlden, a. warren, d. hughes, l. fidani, a. goate, m. rossor, p. roques, j. hardy, nature 353 (1991) 844-846. [40] j. lewis, e. mcgowan, j. rockwood, h. melrose, p. nacharaju, m. van slegtenhorst, k. gwinn-hardy, m.p. murphy, m. baker, x. yu, k. duff, j. hardy, a. corral, w.-l. lin, s.-h. yen, d.w. dickson, p. davies, m. hutton, nature genetics 25 (2000) 402-405. [41] k. santacruz, j. lewis, t. spires, j. paulson, l. kotilinek, m. ingelsson, a. guimaraes, m. deture, m. ramsden, e. mcgowan, c. forster, m. yue, j. orne, c. janus, a. mariash, m. kuskowski, b. hyman, m. hutton, k.h. ashe, science (new york, n.y.) 309 (2005) 476-481. [42] k. duff, h. knight, l.m. refolo, s. sanders, x. yu, m. picciano, b. malester, m. hutton, j. adamson, m. goedert, k. burki, p. davies, neurobiology of disease 7 (2000) 87-98. [43] c. andorfer, y. kress, m. espinoza, r. de silva, k.l. tucker, y.a. barde, k. duff, p. davies, journal of neurochemistry 86 (2003) 582-590. kin tam et al. admet & dmpk 3(3) (2015) 242-253 252 [44] n. sahara, m. murayama, m. higuchi, t. suhara, a. takashima, frontiers in neurology 5 (2014) 26. [45] m.v. rao, m.k. mcbrayer, j. campbell, a. kumar, a. hashim, h. sershen, p.h. stavrides, m. ohno, m. hutton, r.a. nixon, the journal of neuroscience : the official journal of the society for neuroscience 34 (2014) 9222-9234. [46] t.l. spires-jones, t. friedman, r. pitstick, m. polydoro, a. roe, g.a. carlson, b.t. hyman, neuroscience letters 562 (2014) 63-68. [47] l.j. blair, h.d. frauen, b. zhang, b.a. nordhues, s. bijan, y.c. lin, f. zamudio, l.d. hernandez, j.j. sabbagh, m.l. selenica, c.a. dickey, acta neuropathologica communications 3 (2015) 8. [48] s. oddo, a. caccamo, j.d. shepherd, m.p. murphy, t.e. golde, r. kayed, r. metherate, m.p. mattson, y. akbari, f.m. laferla, neuron 39 (2003) 409-421. [49] l.m. billings, k.n. green, j.l. mcgaugh, f.m. laferla, the journal of neuroscience 27 (2007) 751-761. [50] j. lewis, d.w. dickson, w.l. lin, l. chisholm, a. corral, g. jones, s.h. yen, n. sahara, l. skipper, d. yager, c. eckman, j. hardy, m. hutton, e. mcgowan, science (new york, n.y.) 293 (2001) 1487-1491. [51] e.m. ribe, m. perez, b. puig, i. gich, f. lim, m. cuadrado, t. sesma, s. catena, b. sanchez, m. nieto, p. gomez-ramos, m.a. moran, f. cabodevilla, l. samaranch, l. ortiz, a. perez, i. ferrer, j. avila, t. gomez-isla, neurobiology of disease 20 (2005) 814-822. [52] s. oddo, a. caccamo, j.d. shepherd, m.p. murphy, t.e. golde, r. kayed, r. metherate, m.p. mattson, y. akbari, f.m. laferla, neuron 39 (2003) 409-421. [53] l.m. billings, k.n. green, j.l. mcgaugh, f.m. laferla, the journal of neuroscience: the official journal of the society for neuroscience 27 (2007) 751-761. [54] c.j. hwang, h.m. yun, k.r. park, j.k. song, h.o. seo, b.k. hyun, d.y. choi, h.s. yoo, k.w. oh, d.y. hwang, s.b. han, j.t. hong, molecular neurobiology (2014). [55] a.m. sabogal-guaqueta, j.i. munoz-manco, j.r. ramirez-pineda, m. lamprea-rodriguez, e. osorio, g.p. cardona-gomez, neuropharmacology 93 (2015) 134-145. [56] l.m. stevens, r.e. brown, behavioural brain research 278 (2015) 496-505. [57] s. oddo, neurobiology of aging 24 (2003) 1063-1070. [58] s. oddo, a. caccamo, j.d. shepherd, m.p. murphy, t.e. golde, r. kayed, r. metherate, m.p. mattson, y. akbari, f.m. laferla, neuron 39 (2003) 409-421. [59] j. yao, r.w. irwin, l. zhao, j. nilsen, r.t. hamilton, r.d. brinton, proceedings of the national academy of sciences of the united states of america 106 (2009) 14670-14675. [60] j.j. rodriguez, v.c. jones, m. tabuchi, s.m. allan, e.m. knight, f.m. laferla, s. oddo, a. verkhratsky, plos one 3 (2008) e2935. [61] l. wang, j.p. miller, m.h. gado, d.w. mckeel, m. rothermich, m.i. miller, j.c. morris, j.g. csernansky, neuroimage 30 (2006) 52-60. [62] r.g. morris, p. garrud, j.n. rawlins, j. o'keefe, nature 297 (1982) 681-683. [63] c.v. vorhees, m.t. williams, nature protocols 1 (2006) 848-858. [64] j. mclaurin, m.e. kierstead, m.e. brown, c.a. hawkes, m.h. lambermon, a.l. phinney, a.a. darabie, j.e. cousins, j.e. french, m.f. lan, f. chen, s.s. wong, h.t. mount, p.e. fraser, d. westaway, p. st george-hyslop, nature medicine 12 (2006) 801-808. [65] d. fey, s. commins, e. bullinger, journal of computational neuroscience 30 (2011) 447-454. [66] p. galeano, p.v. martino adami, s. do carmo, e. blanco, c. rotondaro, f. capani, e.m. castano, a.c. cuello, l. morelli, frontiers in behavioral neuroscience 8 (2014) 321. [67] z. zhong, l. yang, x. wu, w. huang, j. yan, s. liu, x. sun, k. liu, h. lin, s. kuang, x. tang, journal of molecular neuroscience: mn 53 (2014) 370-376. [68] s. duan, x. guan, r. lin, x. liu, y. yan, r. lin, t. zhang, x. chen, j. huang, x. sun, q. li, s. fang, j. xu, z. yao, h. gu, neurobiology of aging 36 (2015) 1792-1807. [69] a.k. sachdeva, k. chopra, the journal of nutritional biochemistry 26 (2015) 736-744. admet & dmpk 3(3) (2015) 242-253 animal models for alzheimer’s disease doi: 10.5599/admet.3.3.195 253 [70] c. scuderi, c. stecca, m. valenza, p. ratano, m.r. bronzuoli, s. bartoli, l. steardo, e. pompili, l. fumagalli, p. campolongo, l. steardo, cell death & disease 5 (2014) e1419. [71] z. dong, y. bai, x. wu, h. li, b. gong, j.g. howland, y. huang, w. he, t. li, y.t. wang, neuropharmacology 64 (2013) 65-73. [72] d.m. diamond, c.r. park, k.l. heman, g.m. rose, hippocampus 9 (1999) 542-552. [73] c. corre, m. friedel, d.a. vousden, a. metcalf, s. spring, l.r. qiu, j.p. lerch, m.r. palmert, brain structure & function (2014). [74] f. khojasteh, a. nahavandi, s. mehrpouya, j.r. homberg, s. mirzamohammadi, s. raufi, m. soleimani, m. barati, brain research 1596 (2015) 58-68. [75] w.j. strittmatter, proceedings of the national academy of sciences of the united states of america 90 (1993) 1977-1981. [76] k.k. cheng, c.f. yeung, s.w. ho, s.f. chow, a.h. chow, l. baum, the aaps journal 15 (2013) 324-336. [77] s.a. austin, a.v. santhanam, d.j. hinton, d.s. choi, z.s. katusic, journal of neurochemistry 127 (2013) 691-700. [78] m.t. koh, s. rosenzweig-lipson, m. gallagher, neuropharmacology 64 (2013) 145-152. [79] h. scholtzova, r.j. kascsak, k.a. bates, a. boutajangout, d.j. kerr, h.c. meeker, p.d. mehta, d.s. spinner, t. wisniewski, the journal of neuroscience: the official journal of the society for neuroscience 29 (2009) 1846-1854. ©2015 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ anti-inflammatory activities of flavonoid derivates doi: https://doi.org/10.5599/adme t.1918 1 admet & dmpk 0(0) (2023) 000-000; doi: https://doi.org/10.5599/admet.1918 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper enhancement of apixaban's solubility and dissolution rate by inclusion complex (β-cyclodextrin and hydroxypropyl β-cyclodextrin) and computational calculation of their inclusion complexes zainab n. salman1, israa al-ani1,*, khaldun m. al azzam1,, bashar j. m. majeed1, hassan h. abdallah2 and el-sayed negim3,4 1pharmacological and diagnostic research center (pdrc), faculty of pharmacy, al -ahliyya amman university, amman 19328, jordan 2chemistry department, college of education, salahaddin university -erbil, iraq. 3school of materials science and green technologies, kazakh -british technical university, st. tole bi, 59, almaty 050000, kazakhstan. 4school of petroleum engineering, satbayev university, 22 satpayev street, almaty 050013, kazakhstan corresponding authors: e-mail:*ialani@ammanu.edu.jo; azzamkha@yahoo.com received: may 10, 2023; revised: july 31, 2023; published: august xx,2023 abstract background and purpose: apixaban (ap) is a factor x inhibitor, an orally active drug that inhibits blood coagulation for better prevention of venous thromboembolism. it has poor solubility, dissolution rate and low bioavailability. the aim of this study was to improve the aqueous solubility and dissolution rate of oral ap as a step to enhance its bioavailability by preparing it as an inclusion complex with beta and hydroxy propyl beta-cyclodextrin. experimental approach: a simple, rapid method of analysis of ap was developed using ultraviolet spectrophotometry (uv) and partially validated in terms of linearity, precision and accuracy, recovery, and robustness. ap was prepared as a complex with beta cyclodextrin (βcd) and hydroxy propyl beta cyclodextrin (hpβcd) in weight ratios 1:1, 1:2, and 1:3 by kneading, solvent evaporation and spray drying methods and characterized by fourier transfer infra-red (ftir), differential scanning calorimetry (dsc), and percent drug content in each of the prepared complex. using the computer simul ation, the interactions of ap with βcd and hpβcd were investigated. key results: the phase solubility study showed that the solubility of ap was greatly enhanced from 54×10 -3 mmol /l to 66 mmol/l using hpβcd with acceptable stability constant. computer docking supports the formation of a stable 1:1 complex between ap and cd’s. the dissolution test results showed that the complex gave a significantly higher percentage of drug release (95%) over one hour compared to the free ap (60%) (p<0.05). conclusion: aphpβcd complex in the ratio of 1:2 (w/w) can significantly improve the solubility and in vitro dissolution rate of ap. ©2023 by the authors. this article is an open -access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords apixaban; hpβcd, βcd, capsule, bioavailability, solubility introduction the problem of low solubility is a major challenge in the formulation of new drugs [1]. poor solubility might lead to slow and incomplete absorption and low bioavailability [2,3]. the enhancement of drug https://doi.org/10.5599/admet.1918 https://doi.org/10.5599/admet.1918 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:ialani@ammanu.edu.jo mailto:azzamkha@yahoo.com http://creativecommons.org/licenses/by/4.0/ z. n. sal man et al. admet & dmpk 00(0) (2023) 000-000 2 solubility, thereby enhancing its oral bioavailability, remains one of the most challenging steps of the drug development process, especially for oral administrated drug delivery systems. many techniques can be utilized for solubility enhancement, especially those of class ii and iv bcs (biopharmaceutical classification system) drugs, to improve their oral bioavailability. techniques include cosolvents, micronization, changing in the dielectric constant of the solvent, use of amorphous forms, chemical modification of the molecule, use of surfactants, solid dispersion techniques, eutectic mixtures, inclusion complex, manipulation of ph of solve nt, utilization of hydrates or solvates, utilization of soluble prodrugs, application of ultrasonic waves, functional polymer technology, and some others. novel drug delivery technologies used in recent years for solubility enhancement of insoluble drugs a re size reduction technologies, lipid-based delivery systems, micellar technologies, and porous microparticle technology. the selection of which method would be based on certain criteria such as properties of a drug, nature of excipients to be chosen, and nature of the desired dosage form as well as cost and industrial issues [4-6]. apixaban (ap) was developed by bristol-myers squibb using the trade name of “eliquis”, with high potency and selectivity, and is a very effective inhibitor for the blood coagulation factor xa [7]. it is an orally developed drug that inhibits blood coagulation for better prevention of venous thromboembolism (vte) after hip or knee replacement surgery as an alternative to previously used warfarin [8,9]. it is also indicated for the prevention and treatment of other thromboembolic diseases. ap is {1-(4-methoxyphenyl)-7-oxo-2)-4]-6-oxopiperidin-1-yl)phenyl]-5,4-dihydropyrazolo-[3,4-c]-pyridine3-carboxamide}, its chemical structure is shown in figure 1 [10]. ap has a molecular weight of 459.5 g/mol, and the chemical formula of c25h25n5o4. it is a white to pale yellow powder with a melting point equal to 326 °c. it has good solubility in dimethyl sulfoxide and dimethylformamide but low water solubility of (0.028 mg/ml at 24 °c). different references give different solubility values for ap [11]. figure 1. the chemical structure of ap. ap is absorbed along the gastrointestinal tract git with the distal part and descending colon contributes up to 55 % of total absorption. it also shows dissolution limited absorption, which also results in slow absorption, and so, cmax is achieved in 3 5 hours, while in rats, absorption was found to be faster and cmax was achieved in 1 2 hours [12]. evidence shows that ap is a p-gp substrate and its efflux is affected by p-gp inhibitors, like ketoconazole or cyclosporin on the caco-cell membrane [13]. ap mainly distributes e xtracellular fluids with a volume of distribution of 21 l and is 87 to 90 % bound to plasma protein, mainly albumin [14]. the half-life of ap is about 12 hours [15] and the elimination of ap involves metabolism, biliary excretion, and renal excretion of unchanged drug. the pathways of metabolizing ap include o-demethylation, hydroxylation, and sulfation of hydroxylated o-demethyl ap with a primary metabolism going on through cytochrome p450 (cyp) 3a4, with slight help from cyp2c9, cyp2c19, cyp2c8, and cyp1a2. a similar profile was noticed in rats in the preclinical study phase; the o-dimethyl ap was the major metabolite [16]. admet & dmpk 00(0) (2023) 000-000 enhancement of apixaban solubility by cyclodextrins doi: https://doi.org/10.5599/adme t.1918 3 cyclodextrins are cyclic oligomers capable of producing water-soluble inclusion complexes with small molecules and portions of large compounds. these biocompatible, cyclic oligosaccharides will not cause an immune reaction and have little toxicity in animals and humans. cyclodextrins have multiple purposes in the pharmaceutical industry, including increasing the bioavailability of drugs. cyclodextrin -containing polymers are studied and their use in drug delivery presented an interest in the use of cyclodextrin-containing polymers to provide a better ability in the delivery of nucleic acids [17]. the aim of this study is to prepare and evaluate βcd-ap and hpβcd-ap inclusion complexes to improve the solubility and dissolution rate of an oral dosage form of ap. experimental chemicals and instrumentation apixaban (purity 99.5 /sigma), βcd and hpβcd /sigma) a hitachi ratio beam u-1800 spectrometer, evisa, ependorff 5810 centrifuge with a maximum rpm of 11,000 (ny, usa), vortex mixer (labinco l46, netherland), eppendorf centrifuge 5810 r, balance mettler 96 (at300, sartorius, five decimals), ph meter (sartorius 7110), sonicator (elmasonic s100), water bath bm100 with digital thermostat, ir-prestige-21 ftir spectrometer (shimadzu, japan), buchi nano spray dryer b-90 hp were used throughout the study. development of analytical method based on uv absorbance ap is slightly soluble in water. absolute ethanol and methanol were tried in this study to prepare ap stock solution. the best solubility was obtained in methanol (1 mg/ml). good solubility was also obtained from ethanol (0.88 mg/ml) that covered the required and expected concentrations. due to methanol toxicity and methanol residue avoidance, ethanol was chosen as a solvent in addition to ethanol–water mixture for more dilute samples. the stock solution of ap in ethanol was prepared by dissolving 25 mg ap in 50 ml ethanol with gentle stirring and then filtration to get a clear solution. dilution to 50 and 100 µg/ml were prepared using ethanol and scanned from 200 400 nm in uv spectrophotometer (model u -2000, hitachi, tokyo, japan). quartz cuvette was used and ethanol as blank and control. βcd and hpβcd soluble in distilled water were also scanned between 200-400 nm to ensure the specificity of the method when used to measure ap in the prepared complexes. method validation different dilutions of ap were made from the stock solution in ethanol and used in different tests of validation according to the ich guideline of pharmaceutical analysis. linearity, precision and accuracy, recovery and robustness were performed to validate the developed method. calibration concentrations used in the linearity test were 15, 25, 30, 50, 60, 70, 80 and 90 µg/ml. although 5 µg/ml was also tested for linearity as a lower limit of quantification. absorbance was read trice and average ± sd was recorded at 278 nm. inter and intra-day precision of the study were examined and established by analysis of 6 samples (calibration concentrations) with three replicates. within-run accuracy and precision of the method were calculated by analyzing six samples with three replicates. relative standard deviation (rsd) was measured using the proportions of the standard deviation to the average. the comparison of practical amounts from the controls with actual values present in the samples provides the accuracy of this method. recovery is a ratio of the concentration of analyte available in or mixed with the analytical part of the test substance. this would be extracted and submitted for measurement (ich guidelines). the test powder and z. n. sal man et al. admet & dmpk 00(0) (2023) 000-000 4 capsule were used to test the recovery of the method. the formula contained ap beside other excipients prepared to compare the dissolution rate of the capsule containing the complex to the mixture of drug and excipients. one capsule contains 5 mg ap with the powder mix. a powder weight contain ing 1 mg ap in a powder blend was taken and dissolved in 5 ml ethanol, filtered through membrane filter 0.45 and 2 ml was diluted with ethanol to 10 ml to get theoretically 20 µg/ml solution. three solutions were prepared from 3 capsules and measured spectrophotometrically at 278 nm. the other two concentrations (c) were prepared, 50 and 100 µg/ml, following the same procedure. each time drug recovery was calculated by equation (1), in addition to rsd as a precision parameter: recovery = (cap measured / ctheoretical) 100 (1) robustness studies the effect of variation in conditions on the result. different conditions were chosen for measuring a sample of 30 µg/ml ap. conditions such as changing the wavelength ± 5 nm and warming of sample test to 30 °c and changing the instruments from other laboratories were used (medicinal chemistry lab in aau). other solvent was used (methanol). each sample was read in triplicate and accuracy and rsd were calculated. preparation of ap-βcd and ap-hpβcd inclusion complexes all the following methods were applied on βcd and hpβcd in the same way and 3 different ratios (1:1, 1:2, and 1:3 w/w). kneading method in the kneading method, hpβcd or βcd was weighed (25 mg) and put in a mortar, then mixed with enough amount of water-ethanol mixture to get a thick paste. to this paste, ap was gradually incorporated in a total of up to (25 mg). the kneading process was continued for 1 hour manually in one direction. later, enough solvent was mixed to have a consistent paste, and then it was dried at 40 °c for 48 hours in (memert oven). then, the dried mixture was crushed gently using the mortar and pestle. the scratched mixture was sieved using a mesh sieve (#65). then, the formed complex was stored in a closed container. the above procedure was repeated to make the weight ratios of 1:1, 1:2, 1:3 for both βcd and hpβcd. solvent evaporation method in this method, 25 mg ap was dissolved in 50 ml ethanol. the specified amount of βcd or hpβcd was dissolved in 25 ml of distilled water. the above-prepared solutions were mixed with the help of a magnetic stirrer, and then the mixture was stirred for 8 hours. later, most of the solvent was removed using a rotary evaporator (heating at 50 °c) at constant stirring. reduced pressure was applied to remove the residual solvent from the mixture. the final product was placed in an oven at 40 °c for 48 hours to eliminate the excess solvent. the mixture was scratched and sieved using a mesh sieve (#65) and then stored in a tightly closed container. the above procedure was repeated to make the ratios of 1:1, 1:2, 1:3 for both βcd and hpβcd. spray drying method buchi nano spray dryer b-90 hp was used to perform the spray drying method. ap was dissolved in ethanol (50 ml), then hpβcd or βcd was dissolved in distilled water (50 ml). the two mixtures were combined by sonication for 20 minutes to provide a clear final solution. later, the solution w as undergone for spray dry process, and the drying conditions were as follows: temperature (outlet); 90 °c, temperature (inlet); 168 °c, flow rate (solution) 1000 ml/h; flow rate (air) 400 ml/h. the final inclusion complex was crushed and sieved using sieve mesh size #65. after sieving, the inclusion complex was stored in a tight container. the above procedure was repeated to make the ratios (1:1, 1:2, 1:3) and for both βcd and hpβcd. admet & dmpk 00(0) (2023) 000-000 enhancement of apixaban solubility by cyclodextrins doi: https://doi.org/10.5599/adme t.1918 5 characterization of the prepared complex after the preparation of the eighteen samples of ap-cd and aphpβcd complexes, they were characterized by the following methods: fourier transform infrared spectroscopy (ftir) ftir was used as a characterization technique for the formation of complexes. ap, βcd and hpβcd, and the prepared complexes were scanned separately to detect the changes and identify peaks (only a representative chart is included). differential scanning calorimetry (dsc) dsc was performed on selected samples for further confirmation. dsc (model dt-60, shimadzu) was used to perform this test. ap, ap-βcd, and the ap-hpβcd complex were scanned from 0 – 360 oc at 10 oc/min. thermograms were recorded and examined. percent yield of the prepared complex the yield (%) of each prepared complex was calculated as follow s, equation (2): yield = (practical weight of the product/ theoretical weight (drug + cd)) 100 (2) percent drug content in the prepared complex this test will give the efficiency of the complexation process. it also reflects the efficiency of the method. for all types of complexes prepared, the percent ap in the complex was determined by dissolving an amount of complex equivalent to a specified amount of ap (25 mg) in 50 ml distilled water and shaking for 1 hour. several amounts were tried that gave a clear solution. then 50 ml ethanol was added and stirred for 2 hours to dissolve ap. suitable dilutions were made after filtration and then read s pectrophotometrically at 278 nm. phase solubility study a solubility study was conducted in distilled water with the higuchi and connors method [18]. different amounts of hpβcd (depending on the results of characterization) were dissolved in 20 ml distille d water and the excess amount of ap was added. samples were placed in a water bath at 25 °c for 24 hours with agitation at a rate of 50 rpm until equilibrium was achieved. samples of 3 ml were withdrawn and filtered through a 0.45 μm membrane filter. filtrate (100 μl) was diluted appropriately and assayed at 278 nm. the concentrations used are given in table 1. table 1. concentrations of ap (using excess amount) and hpβcd used in phase solubility study. sample code chpβcd / mmol l-1 sample 1 4.00 sample 2 10.0 sample 3 15.0 sample 4 20.0 the solubility of pure ap was also tested using distilled water and an excess of ap in the same conditions. a phase solubility graph was then constructed, and the apparent stability constants ( kst) were determined from the phase solubility analysis graph using equation (3): kst = slope/ (s0 (1 slope)) (3) where kst is the stability constant of the complex, s0 is the absolute solubility of ap in the absence of cds [18]. z. n. sal man et al. admet & dmpk 00(0) (2023) 000-000 6 computer simulation of inclusion complex the structure of the guest molecule, ap, was downloaded as a mol file from the chemspider database (www.chemspider.com) [19]. the structure was optimized using the am1 method using gaussian09 software [20]. on the other hand, the crystal structure of the host molecule βcd was downloaded from cambridge crystallographic data centre (ccdc) under the reference number. the structure was checked, and the crystalized molecules were removed. in addition, hydroxypropyl groups were added using gaussview [ 20] to produce the hydroxypropyl-βcd structure. the optimized structure of the guest molecule was then docked at the center of the host molecules, βcd and hpβcd. to prepare the host -guest complexes for docking calculations, kollman united atom (kua) charges were added in a grid box with 60x60x60 dimensions using autogrid as part of autodock 4 software (molinspiration database). all possible conformations were docked at the center of the host molecule using the lamarckian genetic algorithm with 250 runs for each guest molecule. in vitro dissolution of ap-hpβcds inclusion complexes versus free ap powder dissolution was performed to compare the dissolution of free ap with the complex prepared in a dissolution apparatus usp type ii (paddle type). the dissolution medium consists of 900 ml (ph 6.8 phosphate buffer) maintained at 37±1 °c, with paddle rotation at the speed of 50 rpm. three jars were used for free drug and 3 jars for the complex. five mg ap free was added in each jar specified for “free drug” and an amount of the dry complex powder equivalent to 5 mg ap was added to the jars specified to the “complex powder” and labeled. samples of 5 ml dissolution media were withdrawn at various time points and replaced with fresh buffer at time points: 5, 10, 20, 30, 45, and 60 minutes. samples were analyzed using the validated method and the percent cumulative ap release versus time point were calculated and plotted. preparation of ap capsules ap is found in the market as tablets contain 2.5 or 5 mg ap and (anhydrous lactose (bulking agent)), microcrystalline cellulose (bulking agent), croscarmellose sodium (disintegrant), sodium lauryl sulfate (surfactant), and magnesium stearate (lubricant) [21]. formulation of tablets and their parameters was left as future work. a simple capsule dosage form was prepared to compare a formulation and perform drug release on a dosage form. a simulated capsule containing 5 mg ap, 94 mg avicel ph102, and 1 mg sls in each capsule was prepared. no disintegrant or lubricant was needed. each test capsule contained an amount of complex equivalent to 5 mg ap, 1 mg sls and the weight was completed to 100 mg by avicel. a small batch of 20 capsules was prepared by mixing ingredients and filling in a lab-scale capsule filling machine. to ensure uniformity of weight, capsules were weighed individually and any capsule deviating by 5% and more from 100 mg was excluded. dissolution test of ap capsules the dissolution test was performed on the prepared ap capsule and the complex -containing capsules in the same conditions as the powder dissolution. each sample withdrawn was filtered through a nylon membrane filter (0.45 μm), diluted by methanol (blank) and analyzed using uv spectrophotometer at 278 nm. the percent cumulative of ap release of each experiment were calculated and plotted. to investigate the kinetic of drug release, several release models were applied. these models are shown in table 2. the aim is to get a straight line, and the highest correlation (r2) would be considered the closest model. all models were fitted using microsoft excel. http://www.chemspider.com/ admet & dmpk 00(0) (2023) 000-000 enhancement of apixaban solubility by cyclodextrins doi: https://doi.org/10.5599/adme t.1918 7 table 2. kinetic of drug release models tested. model name y-axis x-axis zero-order cumulative amount of drug release time first order log (cumulative amount of drug remained) time peppas model log (cumulative amount of drug release) log (time) higuchi model cumulative amount of drug release time1/2 hixon crowel model w01/3 – wt1/3* time *w0 is initial amount, wt is amount remaining statistical analysis statistical analysis was used in this study. for the method of analysis, means, standard deviations (sd), coefficient of variation (cv) and their percentages were all used. determination of co rrelation coefficient (r2) was also used. in terms of dissolution and phase solubility analysis study, a t-test was used to compare the target parameters of both drugs when given alone and in combination to test if any significant difference in ci 5% is present. the online “graphpad” or excel software program was used for this purpose. results and discussion method development and validation a sample of ap was scanned for the determination of the λmax. figure 2 shows the scanning result of a solution containing 50 µg/ml ap in ethanol. the wavelength of 278 nm gave the maximum absorbance. scanning of ethanol alone and different concentrations of βcd and hpβcd was also performed to ensure the specificity of the λmax chosen. these results proved that the solvent, βcd, and hpβcd have negligible absorbances at the selected λmax of ap, which is 278 nm (data are not shown). this λmax was selected to continue validation and other measurements in the work. wavelenght, nm figure 2. uv scanning of a solution containing 50 µg/ml ap in ethanol. the wavelength of 278 nm exhibited the highest absorption. method validation the method validation followed the ich guidelines in terms of linearity, precision, accuracy, re covery, and robustness. z. n. sal man et al. admet & dmpk 00(0) (2023) 000-000 8 linear regression was observed in the concentrations range of 15 90 μg/ml. the correlation coefficient for the linearity test was 0.998. figure 3 shows the linearity plot of ap. all tests used in the measurement of ap had the target of dilutions to produce concentrations within this range of 15 90 µg/ml to ensure linearity of readings versus concentrations. the back-calculation gave rsd 0.24-2.66 %. cap / g ml -1 figure 3. the linearity plot of ap shows the linear regression equation and correlation coefficient. the precision of an analytical procedure expresses the closeness of agreement (degree of scatter) between a series of measurements obtained from multiple sampling of the same homogeneous sample under the prescribed condition. precision is measured by the rsd of a series of readings which should not exceed 2.0 % according to the ich guidelines. while accuracy is expressed in “percent” measured concentration relative to the theoretical concentration. measurements were made over 2 days. table 3 shows measurements at day 1 (interday precision) and day 2 (intraday precision) of ap. the rsd for all readings ranged 0.88 to 1.2 %, which complies with the ich. the accuracy or recovery of ap from the samples ranged between 96.4 to 103.2 %, knowing that 90 to 110 % is accepted as per the ich guidelines. table 3. results of inter and intra-day precision and accuracy of ap. sample id ctheoretical / µg ml-1 absorbance (average of n=3) cmeasured / µg ml-1 accuracy, % rsd, % day 1 sample1 50 0.471 49.28 98.6 1.2 sample2 50 0.473 49.47 98.9 sample3 50 0.459 48.17 96.4 sample4 50 0.480 50.12 100.2 sample5 50 0.472 49.37 98.8 sample6 50 0.470 49.19 98.4 day 2 sample1 50 0.496 51.60 103.2 0.88 sample2 50 0.481 50.21 100.4 sample3 50 0.469 49.10 98.2 sample4 50 0.488 50.86 101.7 sample5 50 0.476 49.75 99.5 sample6 50 0.471 49.28 98.6 the powder mixture of the test capsule of ap was used to test the recovery of the method. the formula contained ap beside other excipients prepared to compare the dissolution rate of the capsule containing the complex to the mixture of drug and excipients. the recovery or accuracy of the previous test was calculated y = 0.001x – 0.0171 r2 = 0.9983 admet & dmpk 00(0) (2023) 000-000 enhancement of apixaban solubility by cyclodextrins doi: https://doi.org/10.5599/adme t.1918 9 on the pure ap dissolved in ethanol samples. while this test examines the ability of the method to quantify ap in the powder mixture of the formula. this necessitates high specificity to the compound and negligible effect of excipients. a good method would recover the amount of api in higher or lower concentration s using the same conditions with high accuracy. table 4 shows the results of the recovery test of ap, showing the accuracy of measurement and rsd as a measurement of precision. results gave percent recovery of ap from all formula s 97.3 98.2% with precision between 0.9 and 1, which fulfills the criteria of ich guidelines. table 4. results recovery test of ap from capsule dosage form showing accuracy and precision ap in test formulas. sample id ctheoretical / µg ml-1 measured absorbance cmeasured / µg ml-1 (from formula) accuracy, % average accuracy, % rsd, % (precision) sample1-1 40.0 0.359900 39.00 97.5 97.9 0.9 sample1-2 40.0 0.362060 39.20 98.0 sample1-3 40.0 0.362924 39.28 98.2 sample2-1 60.0 0.570500 58.50 97.5 98.2 1.0 sample2-2 60.0 0.574388 58.86 98.1 sample2-3 60.0 0.580220 59.40 99.0 sample3-1 90.0 0.900980 89.10 99.0 97.3 0.97 sample3-2 90.0 0.881540 87.30 97.0 sample3-3 90.0 0.879596 87.12 9.86 robustness studies the effect of variation in conditions on the result and tests how much the method is resistant to these changes. different conditions were chosen for measuring a sample of 50 µg/ml ap. the accuracy of measurement of ap ranged between 97.3 to 99.8 %, which indicated high resistance of the method to changes in the condition specified, as shown in table 5. the wavelength was changed from 278-275 nm one time to 285 another time to test the possibility of a result change. two instruments were used, and the temperature of the test solution was elevated to 30 °c as well as blank. these parameters showed a very slight difference in the accuracy of the measured concentration tested. in conclusion, the method successfully handled some changes in conditions within the ich guidelines' requirements. table 5. results of robustness of method ap (50 µg ml-1) test showing different conditions used, accuracy and precision. sample id absorban ce c/ µg ml-1 accuracy, % rsd, % theoretical measured (n=3) robustness + 5 nm (285 nm) 0.693 50 49.86 99.8 0.6 robustness – 5 nm (275 nm) 0.691 50 49.72 99.6 0.9 robustness-instrument 1 0.687 50 49.37 99.1 0.58 robustness-instrument 2 0.674 50 48.11 97.3 1.2 robustness 30oc 0.690 50 48.78 98.4 0.7 ap in methanol 50 49.1 98.2 0.9 characterization of the prepared complex the prepared complexes by the three mentioned methods were evaluated to confirm complex formation, percent yield and ap content in each powder complex prepared. based on the results, the method of preparation was evaluated, and the best complex was chosen for further work. ftir analysis was done on all the prepared complexes. below are a few examples of the prepared complexes. figure 4 shows the ftir spectrum of ap. figure 5 shows the ftir overlay spectra of ap (a), apβcd (b), and ap-hpβcd (c) results. z. n. sal man et al. admet & dmpk 00(0) (2023) 000-000 10 ap identifying peaks observed were: (1) 3325.0 cm -1 (n-h stretching of sulphonamido group), (2) 3300 to 3500 cm-1 (n-h stretching of amido group) (3) 2930 cm-1 (ch stretching of ch2) and (4) characteristic c=o of amide at 1628.9 cm-1. the prominent peaks for βcd alone are 3398 cm-1 (o-h), 2930 cm-1 (c-h stretching), 1628 to 1598 cm-1 (h-o-h bending), 1156 cm-1 (c-o stretching), 1028 cm-1 (c-o-c stretching). the ir-abundant peaks for hpβcd alone are 3395 cm-1 (o-h), 2930 cm-1 (c-h stretching), 1628 cm-1 (h-o-h bending), 1156 cm-1 (c-o stretching), 1032 cm-1 (c-o-c stretching). wavenumber, cm-1 figure 4. ftir spectrum of ap showing the major peaks (1) 3325.0 cm-1 (n-h stretching of sulphonamido gr.), (2) 3300-3500 cm-1 (n-h stretching of amido gr.) (3) 2930 cm-1 (ch stretching of ch2) and (4) characteristic c=o of amide at 1628.9 cm-1. wavenumber, cm-1 figure 5. overlay spectra of ap (a), ap-βcd (b), and ap-hpβcd (c). the most characteristic peaks in the two complexes ap-βcd (b), and ap-hpβcd (c) appeared with less intensity of the c=o amide group of ap, which might be attributed to the hidden nature of this group inside the cone of the cds. t ra n sm it ta n ce , % t ra n sm it ta n ce , % admet & dmpk 00(0) (2023) 000-000 enhancement of apixaban solubility by cyclodextrins doi: https://doi.org/10.5599/adme t.1918 11 yield and content of ap-βcds inclusion complexes ap-βcd and hpβcd complexes were prepared on a laboratory scale. however, percentage yield was important to determine how efficient the method and the skills used were. the percent yield of each prepared complex was calculated, as shown in table 6. all the methods used were under good control and gave a very high percent yield (89 – 99 %). spray drying gave the highest yield with minimum loss. measuring the content of ap included in each type of the complex formed generally gave a high loading efficiency. two variables can be explained: the weight ratio of ap to cd and the method of preparation. results in table 6 showed that the kneading method gave the highest loading efficiency of ap in the complex for both βcd and hpβcd. solvent evaporation and spray drying methods gave a bit less content drug inclusion. this difference is attributed to the method of preparation and how the molecule gets inside the cd cone. results also showed that increasing the weight ratio of ap from 1:1 to 1:2, both with βcd and hpβcd, respectively, increased the amount of drug loading, but a further increase in the cd ratio did not give a noticeable increase in the amount of drug loading. for that reason, ap-βcd 1:2 and ap-hpβcd 1:2 was chosen for further studies because there was no need to increase cd concentration in the formula if there was no prominent impact on the solubility of ap. table 6. yield and content of ap-βcd and ap-hpβcd complexes. method of preparation type of complex ap:cd weight ratio yield, % loading efficiency, % kneading βcd 1:1 91 90.6 kneading βcd 1:2 93 95.5 kneading βcd 1:3 94 95.6 kneading hpβcd 1:1 89 89.2 kneading hpβcd 1:2 92.5 97.2 kneading hpβcd 1:3 92 96.3 solvent evaporation βcd 1:1 96 88 solvent evaporation βcd 1:2 98 93 solvent evaporation βcd 1:3 98 92 solvent evaporation hpβcd 1:1 95 85 solvent evaporation hpβcd 1:2 96 93 solvent evaporation hpβcd 1:3 97 92 spray drying βcd 1:1 99 85 spray drying βcd 1:2 99 92 spray drying βcd 1:3 99 93 spray drying hpβcd 1:1 99 82 spray drying hpβcd 1:2 99 94 spray drying hpβcd 1:3 99 95 characterization of ap-βcd (1:2) and ap-hpβcd (1:2) complex by dsc further characterization of ap-hpβcd (1:2) complex was done by dsc for further confirmation. results are shown in figure 6 a (ap), -b (hpβcd), and -c (ap-hpβcd 1:2 complex), which was chose n for the formulation of the capsules. results show the melting point of ap of approximately 220 °c from dsc. figure 6 b also shows the dcs thermogram of hpβcd showing a broad peak in the range 280 – 340 °c. the complex thermogram shows that the sharp peak of ap is embedded in the hpβcd and is not clearly distinguished, indicating the formation of an inclusion complex between ap and hpβcd (c). phase solubility analysis of ap-βcd and ap-hpβcd inclusion complexes the phase solubility profile of ap-hpβcd is presented in figure 7. the solubility of pure ap was calculated approximately as 54.4x10-3 mmol (25 µg/ml) (28 µg/ml reported by [22]). using suitable dilutions, concentrations, mol/l of solubilized ap by different concentrations of hpβcd were calculated and given in table 7. z. n. sal man et al. admet & dmpk 00(0) (2023) 000-000 12 figure 6. dsc thermograms of ap (a), hpβcd (b) and ap -hpβcd complex ratio (ap:hpβcd) (1:2 w/w) (c). admet & dmpk 00(0) (2023) 000-000 enhancement of apixaban solubility by cyclodextrins doi: https://doi.org/10.5599/adme t.1918 13 table 7. results of phase solubility study. c / mmol l-1 hpβcd soluble ap soluble ap total s0 +scomplex soluble ap, by hpβcd 4.0 30.621 66.6764 66.622 10 .0 30.781 66.8434 66.789 15.0 30.871 67.1904 67.185 20.0 30.971 67.4084 67.403 chpcd / mm l-1 figure 7. phase solubility diagram of ap with hpβcd. from figure 7, the solubility of ap increased proportionally with an increase in the concentration of hpβcd (r2 = 0.97). however, the increase of molar concentration of hpβcd from 4 20 (5 folds) resulted in a very slight increase in the solubility of ap. 1:1 relation means that every molecule of the host compound incorporates one molecule of the guest (ap). the lowest concentration of hpβcd used (4 mm) resulted in 1224 fold increase in the solubility of ap in water (from 54.4 µm to 66.622 mm). further increase in hpβcd resulted in a slight increase in solubility. however, a straight line with a good correlation was obtained in figure 7, and the stability constant was calculated by equations (3) and (4): ( ) st 0 slope 1-slope k s = (3) ( ) st 0.0513 0.0544 1 0.0513 k − = 994.2 m-1 (4) the stability constant (kst) equals 994.2 m -1. it was reported that ks value between 50 and 5000 m −1 was considered the most suitable for the improvement of solubility and stability of poorly soluble drugs [2 3]. in conclusion, hpβcd increased the solubility of ap significantly with good stability of the complex in a ratio 1:1 host-guest molecules. computer calculations docking calculations offer essential information regarding the conformation of the guest -host complexes and produce a quantitative parameter of the binding affinity of different complexes. ap has shown the highest binding affinity in comple xing with hpβcd (-7.62 kcal/mol) followed by βcd (-6.81 kcal/mol), where it agrees with the experimental part attained. c s o lu b il e a p , m m l -1 , % z. n. sal man et al. admet & dmpk 00(0) (2023) 000-000 14 figures 8 and 9 give better insight into the binding interaction between the host-guest complexes. in both cases, the amide group is included inside the cone. the same result was obtained from the ftir analysis , where the amide peak was shortened, and its absorption was weakened. figure 8. interaction of ap with βcd as proposed by gaussian09 software. figure 9. interaction of ap with hpβcd as proposed by gaussian09 software. admet & dmpk 00(0) (2023) 000-000 enhancement of apixaban solubility by cyclodextrins doi: https://doi.org/10.5599/adme t.1918 15 dissolution studies the dissolution rate profiles of standard ap alone and ap -hpβcd (1:2) system in ph 6.8 at 37.0 ± 0.5 °c are displayed in figures 10 and 11. the release rate profiles were expressed as the amount (amount released×100 / total amount (5 mg)) of ap released from the complexes against time in minutes. figure 10. dissolution profile of ap free powder and ap -hpβcd (1:2) powder at 37 °c, 50 rpm, usp dissolution apparatus type ii. figure 11. dissolution profile of ap free ap-hpβcd (1:2) from capsule dosage form 37 °c, 50 rpm, usp dissolution apparatus type ii. figure 10 shows an improvement in the dissolution rate of ap released from ap -hpβcd (1:2) powder over ap pure powder. this indicates improvement of solubility of ap by complexation with hpβcd. the amount released in 10 minutes (t10) was taken as a comparison point and results showed a significant increase in ap dissolved from hpβcd complexes (52 ± 3.5) 5 compared to ap pure powder (30 ± 4.0) with (p<0.05). z. n. sal man et al. admet & dmpk 00(0) (2023) 000-000 16 ap release from capsule dosage form (free ap versus complex) is shown in figure 11. time started after the capsules were perforated in the media and the powder started to spread in the media in about 20 ±3 minutes. results showed the dissolution was significantly improved compared to the free -ap capsules. t10 from capsules containing the complex was 55±3.4 % compared to 32±1.5 %. these results are close to those of free powder. kinetics of drug release the in vitro drug release mechanism of ap from capsules in buffer ph 6.8 at 37.0±0.5 °c can be described by fitting the dissolution data in five different kinetic models such as zero -order, first-order, peppas, higuchi, and hixson–crowell (table 8). the linearity was calculated to determine the best-fit model and possible mechanism of drug release from the formula. drug release kinetics was studied for the capsules (reference and ap -hpβcd (1:2) containing capsules) using the models shown in table 8 using microsoft excel for the model fitting and calculations. the aim is to get a straight line that explains the linear relation between the specified variables. linearity is evaluated by correlation coefficient (r2) to choose the best-fit model (table 8). the equation that best suits the dissolution data where the values of r2 were >0.9. results of model fitting (table 8) showed that ap is released from the complex in a capsule dosage form following “first order” release kinetic with a total amount released in 1 hour of more than 90 %, which is a suitable behavior for immediate release (ir) dosage form indicating first-order release mechanism. these results also indicated that the complex significantly improved the drug release and dissolution behavior. table 8. results of model fitting for drug release from reference capsules and capsul es contain (ap-hpβcd). model zero-order first order peppas model higuchi model hixon crowel model ap release from the reference capsule r2 0.88 0.94 0.91 0.96 0.94 ap release from capsules contains (ap-hpβcd) r2 0.80 0.99 0.98 0.97 0.95 conclusions ap forms a complex with βcd and hpβcd with good loading efficiency using the kneading method, solvent evaporation method and spray drying method. the highest percentage was gained from the kneading method in a ratio of 1:2 ap-hpβcd. also, the phase solubility study showed that ap reacts with hpβcd in a ratio of 1:1 stoichiometry with a stability constant equal to 994 m-1.. computer simulation of the ap-cd interaction supports the formation of 1:1 complex. for characterization purposes, ftir and dsc showed a physical interaction between ap and hpβcd. additionally, drug dissolution was significantly higher from the powder of the complex and the formula of gelatin capsules compared to free ap and ap in capsules without the complex. list of abbreviations ap apixaban uv ultraviolet spectrophotometry βcd beta cyclodextrin and hpβcd hydroxy propyl beta cyclodextrin ftir fourier transfer infra-red dsc differential scanning calorimetry bcs biopharmaceutical classification system vte venous thromboembolism sd standard deviations cv coefficient of variation r2 correlation coefficient admet & dmpk 00(0) (2023) 000-000 enhancement of apixaban solubility by cyclodextrins doi: https://doi.org/10.5599/adme t.1918 17 ethics approval and consent to participate: no animal experiments have been conducted in this study. human and animal rights: no humans or animals were used in this study. availability of data and materials: the data used to support the findings of this study are available from the corresponding authors upon request. conflict of interest: the authors declare no conflict of interest, financial or otherwise. acknowledgements: the author gratefully thank to ysrafil, to conductiing this projects, apt. firzan nainu, m. biomed., sc., ph.d and talha bin emran, ph.d for supervised and grammar/structural editing of this manuscript. references [1] a. sharma, m. soni, s. kumar, g.d. gupta. solubility enhancement—eminent role in poorly soluble drugs. research journal of pharmacy and technology 2 (2009) 220-224. https://rjptonline.org/ abstractview.aspx?pid=2009-2-2-46 [2] j. hodgson. admet—turning chemicals into drugs. nature biotechnology 19 (2001) 722-726. https:// doi.org/10.1038/90761 [3] k.t. savjani, a.k. gajjar, j.k. savjani. drug solubility: importance and enhancement techniques. international scholary research notices 2012 (2012) 1–10. https://doi.org/10.5402/2012/195727 [4] c. lipiniski, c.j lipinski, c. lipinski, c.a. lipinski, c. lipinski, c.a. lipinski, lipinski, c.a. lipinski. poor aqueous solubility an industry wide problem in drug delivery. american pharmaceutical review 5 (2002) 82-85. https://www.scienceopen.com/document?vid=dbc999fb-0f66-463c-8166-d6576f8778d5 [5] z. guo, c. boyce, t. rhodes, l. liu, g.m. salituro, k.j. lee, a. bak, d.h. leung. a novel method for preparing stabilized amorphous solid dispersion drug formulations using acoustic fusion. international journal of pharmaceutics 592 (2021) 120026. https://doi.org/10.1016/j.ijpharm.20 20.120026 [6] v.r. vemula, v. lagishetty, s. lingala. solubility enhancement techniques . international journal of pharmaceutical sciences review and research 5 (2010) 41-51. https://globalresearchonline.net/jour nalcontents/volume5issue1/article -007.pdf [7] d.j. pinto, m.j. orwat, s. koch, k.a. rossi, r.s. alexander, a. smallwood, p.c. wong, a.r. rendina, j.m. luettgen, r.m. knabb, k. he, b. xin, r.r. wexler, p.y.s. lam. discovery of 1-(4-methoxyphenyl)-7-oxo-6-(4-(2-oxopiperidin-1-yl) phenyl)-4, 5, 6, 7-tetrahydro-1 h-pyrazolo [3, 4-c] pyridine3-carboxamide (apixaban, bms-562247), a highly potent, selective, efficacious, and orally bio available inhibitor of blood coagulation factor xa. journal of medicinal chemistry 50 (2007) 53395356. https://doi.org/10.1021/jm070245n [8] e.d. deeks. apixaban. drugs 72 (2012) 1271-1291. https://doi.org/10.2165/11209020-000000000-00000 [9] y. falck-ytter, c.w. francis, n.a. johanson, c. curley, o.e. dahl, s. schulman, t.l. ortel, s.g. pauker, c.w.c. jr. prevention of vte in orthopedic surgery patients: antithrombotic therapy and prevention of thrombosis: american college of ch est physicians evidence-based clinical practice guidelines. chest 141 (2012) e278s-e325s. https://doi.org/10.1378/chest.11-2404 [10] p.c. wong, d.j. pinto, d. zhang. d. preclinical discovery of apixaban, a direct and orally bioavailable factor xa inhibitor. journal of thrombosis and thrombolysis 31 (2011) 478-492. https://doi.org/ 10.1007/s11239-011-0551-3 [11] h. li, a. farajtabar, y. xie, z. li, h. zhao. apixaban (i) in several aqueous co-solvent mixtures: solubility, solvent effect and preferential solvation. the journal of chemical thermodynamics 150 (2020) 106200. https://doi.org/10.1016/j.jct.2020.106200 [12] w. byon, s. garonzik, r.a. boyd, c.e. frost. apixaban: a clinical pharmacokinetic and pharmacodynamic review. clinical pharmacokinetics 58 (2019) 1265-1279. https://doi.org/10.1007/ s40262-019-00775-z https://rjptonline.org/‌abstractview.aspx?pid=2009-2-2-46 https://rjptonline.org/‌abstractview.aspx?pid=2009-2-2-46 https://doi.org/10.1038/90761‏ https://doi.org/10.1038/90761‏ https://doi.org/10.5402/2012/195727 https://www.scienceopen.com/search#author/fcd542af-dc85-4d2c-9571-825818e2348e https://www.scienceopen.com/search#author/e81a7295-5246-4bad-9d68-5a116d434aa1 https://www.scienceopen.com/search#author/2241ee9e-ce4f-4541-a702-873e7b20da39 https://www.scienceopen.com/search#author/a5bf8c5b-225d-4670-be90-f6c632c3586d https://www.scienceopen.com/search#author/6347ed94-05e7-42f7-982d-035b0564bf25 https://www.scienceopen.com/search#author/ff880798-75d8-452a-bba3-5b0faede4a6f https://www.scienceopen.com/search#author/8ca745ce-ce3b-485b-b463-e9d54d03f29d https://www.scienceopen.com/document?vid=dbc999fb-0f66-463c-8166-d6576f8778d5 https://doi.org/10.1016/j.ijpharm.20‌20.120026 https://doi.org/10.1016/j.ijpharm.20‌20.120026 https://globalresearchonline.net/jour‌nalcontents/volume5issue1/article-007.pdf https://globalresearchonline.net/jour‌nalcontents/volume5issue1/article-007.pdf https://pubmed.ncbi.nlm.nih.gov/?term=wong+pc&cauthor_id=17914785 https://pubmed.ncbi.nlm.nih.gov/?term=rendina+ar&cauthor_id=17914785 https://pubmed.ncbi.nlm.nih.gov/?term=luettgen+jm&cauthor_id=17914785 https://pubmed.ncbi.nlm.nih.gov/?term=knabb+rm&cauthor_id=17914785 https://pubmed.ncbi.nlm.nih.gov/?term=he+k&cauthor_id=17914785 https://pubmed.ncbi.nlm.nih.gov/?term=xin+b&cauthor_id=17914785 https://pubmed.ncbi.nlm.nih.gov/?term=wexler+rr&cauthor_id=17914785 https://pubmed.ncbi.nlm.nih.gov/?term=lam+py&cauthor_id=17914785 https://doi.org/10.1021/jm070245n https://doi.org/10.2165/11209020-000000000-00000 https://pubmed.ncbi.nlm.nih.gov/?term=ortel+tl&cauthor_id=22315265 https://pubmed.ncbi.nlm.nih.gov/?term=pauker+sg&cauthor_id=22315265 https://pubmed.ncbi.nlm.nih.gov/?term=pauker+sg&cauthor_id=22315265 https://pubmed.ncbi.nlm.nih.gov/?term=colwell+cw+jr&cauthor_id=22315265 https://doi.org/‌10.1007/s11239-011-0551-3 https://doi.org/‌10.1007/s11239-011-0551-3 https://www.sciencedirect.com/science/article/abs/pii/s0021961420303694?via%3dihub#! https://www.sciencedirect.com/science/article/abs/pii/s0021961420303694?via%3dihub#! https://www.sciencedirect.com/science/article/abs/pii/s0021961420303694?via%3dihub#! https://www.sciencedirect.com/science/article/abs/pii/s0021961420303694?via%3dihub#! https://www.sciencedirect.com/science/article/abs/pii/s0021961420303694?via%3dihub#! https://www.sciencedirect.com/journal/the-journal-of-chemical-thermodynamics https://doi.org/10.1016/j.jct.2020.106200 https://doi.org/10.1007/‌s40262-019-00775-z https://doi.org/10.1007/‌s40262-019-00775-z z. n. sal man et al. admet & dmpk 00(0) (2023) 000-000 18 [13] d. zhang, k. he, j.j. herbst, j. kolb, w. shou, l. wang, p.v. balimane, y.h. han, j. gan, c.e. frost, w.g. humphreys. characterization of efflux transporters involved in distribution and disposition of apixaban. drug metabolism and disposition 41 (2013) 827-835. https://doi.org/10.1124/ dmd.112.050260 [14] a. jaber, i. al-ani, m. hailat, e. daoud, a. abu-rumman, z. zakaraya, b.j. majeed, o. al meanazel, w.a. dayyih. esomeprazole and apixaban pharmacokinetic interactions in healthy rats. heliyon 8 (2022) e11015. https://doi.org/10.1016/j.heliyon.2022.e11015 [15] c. ward, g. conner, g. donnan, a. gallus s. mcrae. practical management of patients on apixaban: a consensus guide. thrombosis journal 11 (2013) 1-8. https://doi.org/10.1186/1477-9560-11-27 [16] l. wang, n. raghavan, k. he, j.m. luettgen, w.g. humphreys, r.m. knabb, d.j. pinto, d. zhang. sulfation of o-demethyl apixaban: enzyme identification and species comparison. drug metabolism and disposition 37 (2009) 802-808. https://doi.org/10.1124/dmd.108.025593 [17] m.e. davis, m.e. brewster. cyclodextrin-based pharmaceutics: past, present and future. nature reviews drug discovery 3 (2004) 1023-1035. https://doi.org/10.1038/nrd1576 [18] m.m. doile, k.a. fortunato, i.c. schmücker, s.k. schucko, m.a.s. silva, p.o. rodrigues. physicochemical properties and dissolution studies of dexamethasone acetate -β-cyclodextrin inclusion complexes produced by different methods. aaps pharmscitech 9 (2008) 314-322. https://doi.org/10.1208/s12249-008-9042-z [19] chemspider database. http://www.chemspider.com/ (accessed date: 01/08/2023). [20] gaussian 09, revision d.01, m.j. frisch, g.w. trucks, h.b. schlegel, g.e. scuseria, m.a. robb, j.r. cheeseman, g. scalmani, v. barone, b. mennucci, g.a. petersson, h. nakatsuji, m. caricato, x. li, h.p. hratchian, a.f. izmaylov, j. bloino, g. zheng, j.l. sonnenberg, m. hada, m. ehara, k. toyota, r. fukuda, j. hasegawa, m. ishida, t. nakajima, y. honda, o. kitao, h. nakai, t. vreven, j.a. montgomery, j.e. jr., peralta, f. ogliaro, m. bearpark, j.j. heyd, e. brothers, k.n. kudin, v.n. staroverov, r. kobayashi, j. normand, k. raghavachari, a. rendell, j.c. burant, s.s. iyengar, j. tomasi, m. cossi, n. rega, j.m. millam, m. klene, j.e. knox, j.b. cross, v. bakken, c. adamo, j. jaramillo, r. gomperts, r.e. stratmann, o. yazyev, a.j. austin, r. cammi, c. pomelli, j.w. ochterski, r.l. martin, k. morokuma, v.g. zakrzewski, g.a. voth, p. salvador, j.j. dannenberg, s. dapprich, a.d. daniels, ö.farkas, j.b. foresman, j.v. ortiz, j. cioslowski, d.j. and fox. gaussian, inc., wallingford ct, 2009. https://www.scienceopen.com/document?vid=839f33cc-9114-4a55-8f1a-3f1520324ef5 [21] apixaban in: https://www.rxlist.com/eliquis-drug.htm (accessed date: 20/04/2023). [22] y. chen, l. li, j. yao, y.y. ma, j.m. chen, t.b. lu. improving the solubility and bioavailability of apixaban via apixaban–oxalic acid cocrystal. crystal growth and design 16 (2016) 2923-2930. https://doi.org/10.1021/acs.cgd.6b00266 [23] s. sip, a. gościniak, p. szulc, j. walkowiak, j. cielecka-piontek, j. assisted extraction with cyclodextrins as a way of improving the antidiabetic activity of actinidia leaves. pharmaceutics 14 (2022) 2473. https://doi.org/10.3390/pharmaceutics14112473 ©2023 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) https://doi.org/10.1124/‌dmd.112.050260 https://doi.org/10.1124/‌dmd.112.050260 https://doi.org/10.1124/dmd.112.050260 https://doi.org/10.1186/1477-9560-11-27 https://doi.org/10.1124/dmd.108.025593 https://doi.org/10.1038/nrd1576 https://doi.org/10.1208/s12249-008-9042-z http://www.chemspider.com/ https://www.scienceopen.com/document?vid=839f33cc-9114-4a55-8f1a-3f1520324ef5 https://www.rxlist.com/eliquis-drug.htm https://doi.org/10.1021/acs.cgd.6b00266 https://doi.org/10.3390/pharmaceutics14112473 http://creativecommons.org/licenses/by/3.0/ manuscript doi: 10.5599/admet.1.4.21 46 admet & dmpk 1(4) (2013) 46-47; doi: 10.5599/admet.1.4.21 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index editorial one year in circulation – an exciting journey for admet & dmpk kin tam editor: admet & dmpk e-mail: kin_tam@iapchem.org it is now almost a year since admet & dmpk launched. the current issue will be the final one for 2013. we are happy to report that it has been a very exciting year for the journal. in particular, we have published a range of articles, such as book reviews, original scientific papers, perspectives, opinions and mini reviews covering different areas of interest to the journal. this reflects the diversity of the type of articles that enables the exchange of knowledge and shared learning. we would like to take this opportunity to offer thanks for all the help and hard work from our editorial team for supporting the journal. meet the admet & dmpk editorial team mandić has written an excellent review on a book entitled “absorption and drug development. solubility, permeability and charge state. 2nd edition” by the leading expert in the area, alex avdeef [1]. he has given a good overview of the book, and remarked that “this book is essential literature on the desk of the pharmaceutical scientist”. next, gabelica marković has published a book review on “physico chemical methods in drug discovery and development”, edited by zoran mandić [2]. this book was indeed a collection from some selected high quality presentations summarizing the key methods and techniques in pharmaceutical research from the 2 nd world conference on physico chemical methods in drug discovery and development held in zadar in 2011. with regard to the original scientific article, takács-novák et al. have published their study of the acid-based properties of lisinopril, a tetraprotic ampholyte ace inhibitor [3]. nmr-ph titration was used to assign the constants to the functional groups and for the examination of site-specific, submolecular basicities of the molecule. the microconstants and macroconstants were reported and used to help the interpretation of the pharmacokinetic and pharmacodynamic properties of lisinopril. moreover, avdeef et al has described a versatile computer program, pdisol-x tm , to process salt solubility data [4]. this novel approach does not require any explicitly derived equations, and is able to deal with other equilibrium process, such as aggregate formation, as well as performs adjustment calculation in high ionic strength environment. we envisage that pdisol-x tm will be a powerful tool for our preformulation colleagues during salt selection. hong has written a perspective article to discuss the adme tests that should be conducted in preclinical studies, with a particular emphasis on the data needed to support ind filing [5]. in the article, the author has shared his views and opinion based on his many years of experience in drug discovery, cros and ind filing, which is a must-read for readers who want to learn http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:kin_tam@iapchem.org http://pub.iapchem.org/ojs/index.php/admet/about/editorialteam admet & dmpk x(y) (20xx) 000-000 running title doi: 10.5599/admet.0000.0000 47 more about pharmaceutical development. tam has published two mini review articles on the role of doxorubicin in hepatocellular carcinoma [6], and the estimation of first-in-man doses with particular emphasis on oncology drugs [7]. finally, wu et al. have published an opinion article concerning the regulatory control of medical device containing nanomaterial [8]. this is an emerging area where the product safety due to the presence of nanomaterial could be a concern, especially for manufacturing staff, customers and end users. the authors have pointed out that regulatory regulations are important for this kind of medical device. despite the promising development we have seen this year, many challenges remain. for instance, it is generally difficult for new journals to attract a good number of high quality paper submissions in the first few years. no doubt admet & dmpk is suffering from the same problem. hopefully this will become less of an issue next year, as some of our academic/industrial colleagues have already expressed interests and/or committed to submit their articles to our journal. it is believed that admet & dmpk has a unique position in the open literature [9]. to aid the rapid scientific discussion, the journal will seek to publish commentary/correspondence articles where research groups could engage in the debate on a particular topic of interest. moreover, we plan to publish a special issue highlighting the applications of electrochemical techniques in admet studies. it is our expectation that admet & dmpk will become a platform for all scientists working in admet and dmpk areas to publish their research findings in a timely manner. acknowledgements: we thank iapc publishing for supporting the running cost of the journal and hosting it on their website. references [1] z. mandić, amdet & dmpk 1 (2013) 3-5. [2] v.gabelica marković, amdet & dmpk 1 (2013) 17-18. [3] k. takács-novák, k. deák, s. béni, g. völgyi, amdet & dmpk 1 (2013) 6-16. [4] g. völgyi, a. marosi, k. takács-novák, a. avdeef, amdet & dmpk 1 (2013) 48-62. [5] h. wan, amdet & dmpk 1 (2013) 19-28. [6] k. tam, amdet & dmpk 1 (2013) 29-44. [7] k. tam, amdet & dmpk 1 (2013) 63-75. [8] e. wu, c. chan, a.t.w. li, amdet & dmpk 1 (2013) 76-81. [9] k. tam, amdet & dmpk 1 (2013) 1-2. ©2013 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ impact of gastrointestinal differences in veterinary species on the oral drug solubility, in vivo dissolution, and formulation of veterinary therapeutics doi: http://dx.doi.org/10.5599/admet.1140 1 admet & dmpk 10(1) (2022) 1-25; doi: https://doi.org/10.5599/admet.1140 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index perspective impact of gastrointestinal differences in veterinary species on the oral drug solubility, in vivo dissolution, and formulation of veterinary therapeutics # marilyn n. martinez1,*, mark g. papich2, raafat fahmy3 1office of new animal drug evaluation, center for veterinary medicine, us food and drug administration, rockville, maryland, marilyn.martinez@fda.hhs.gov 2college of veterinary medicine, north carolina state university, raleigh, north carolina, mark_papich@ncsu.edu 3office of new animal drug evaluation, center for veterinary medicine, us food and drug administration, rockville, maryland, raafat.fahmy@fda.hhs.gov *corresponding author received: october 15, 2021; revised: december 19, 2021; published: january 02, 2022 abstract many gaps exist in our understanding of species differences in gastrointestinal (gi) fluid composition and the associated impact of food intake and dietary composition on in vivo drug solubilization. this information gap can lead to uncertainties with regard to how best to formulate pharmaceuticals for veterinary use or the in vitro test conditions that will be most predictive of species-specific in vivo oral product performance. to address these challenges, this overview explores species-specific factors that can influence oral drug solubility and the formulation approaches that can be employed to overcome solubility-associated bioavailability difficulties. these discussions are framed around some of the basic principles associated with drug solubilization, reported species differences in gi fluid composition, types of oral dosage forms typically given for the various animal species, and the effect of prandial state in dogs and cats. this basic information is integrated into a question-and-answer section that addresses some of the formulation issues that can arise in the development of veterinary medicinals. ©2022 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords veterinary; gastrointestinal physiology; formulation considerations; solubility introduction drug thermodynamic solubility (cs) is one of the determinants of drug oral bioavailability. the us pharmacopoeia (usp) general chapter, gc<1236> [1], defines cs as “the maximum quantity of a substance that can be completely dissolved at a given temperature, pressure, and solvent ph”. it is only upon in vivo solubilization that a drug can penetrate biological membranes. once traversing the enterocyte membrane, it is either effluxed back into the intestine, metabolized within the enterocyte, or transported into the portal (or lymphatic) circulation [2,3]. ______ # the contents of this manuscript reflects the views of the author and should not be construed to represent fda’s views or policies http://dx.doi.org/10.5599/admet.1140 https://doi.org/10.5599/admet.1140 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:marilyn.martinez@fda.hhs.gov mailto:mark_papich@ncsu.edu mailto:raafat.fahmy@fda.hhs.gov http://creativecommons.org/licenses/by/4.0/ martinez, papich and fahmy admet & dmpk 10(1) (2022) 1-25 2 the conditions influencing in vivo drug solubility can differ markedly across an individual’s gastrointestinal (gi) tract, between individuals of a given species or between species. unfortunately, as compared to humans, there remain many gaps in our understanding both of species-specific differences in gi fluid composition and the impact of food intake and dietary composition on in vivo drug solubilization. furthermore, with the exception of cattle, there is little to no published information on gi fluid composition as a function of animal age, breed, or diet (see details below). these gaps can lead to uncertainties with regard to how best to formulate pharmaceuticals for veterinary use or the in vitro test conditions most predictive of in vivo product performance. this overview describes species-specific factors that can influence in vivo drug solubility and potential formulation approaches to overcome bioavailability challenges. discussions focus on the “major” target animal species within the us (dogs, cats, horses, pigs, swine, cattle and chickens and turkeys) www.farad.org/us-food-animals.html. general considerations the first step in understanding how species differences in gi fluid composition can influence in vivo product performance and drug solubility is to appreciate the difference between dissolution rate and substrate solubility (cs). in that regard, variables influencing dissolution rate can be described by the nernstbrunner equation:   s d d c ds c c t vh (1) where c is the solubilized concentration at time t, cs is the solubility of the substance in question, s is the surface area of the dissolving particle, d is the diffusion coefficient, h is the thickness of the diffusion layer, and v is the volume of the diffusing medium [4]. with regard to h and d, the species differences in fluid viscosity need to be considered. v can also vary as a function of species and prandial state. to appreciate how gi fluid composition may influence in vivo values for cs, it is necessary to recognize the factors that can influence the interaction between a drug, its salt form, drug solid-state characteristics, and the aqueous environment within which the drug molecule will dissolve [5]. in vivo factors to consider include:  fluid viscosity [6]. as discussed later, this can be a particularly important factor affecting oral drug bioavailability in dogs and cats when administered in the fed versus fasted state.  solubility-enhancing surfactants (including those incorporated into the formulation and natural surfactants [7-9]. in this regard, the differences in bile salt composition may influence the nature of drug solubilization across the various veterinary species (an important gap remaining in our understanding of physiological differences as a function of species, diet and breeds).  fluid volume versus dose. within veterinary medicine, formulations are typically approved for administration on the basis of an animal’s body weight (bw). however, bw and gastric fluid volume may not scale directly. this issue can be particularly problematic in dogs where although the body size can range from small miniature breeds to giant breeds (a range that can span 4 to 230 lbs), residual gastric fluid volume is typically limited to the small amounts of water consumed. estimates of gastric fluid volumes in food-producing animal species and horses as provided in cvm’s gfi #171 [10] are included in table 1. http://www.farad.org/us-food-animals.html admet & dmpk 10(1) (2022) 1-25 veterinary gastrointestinal differences influencing solubility doi: http://dx.doi.org/10.5599/admet.1140 3 table 1. estimates of gastric fluid volumes of food-producing species and of the horse species gastric fluid volume, l gastric residence time, hr temperature, °c cattle rumen: 47 8 36.7-39.3 swine 0.5 1 38.7-39.8 horse 1.5 0.25 37.2-38.2 chicken 0.01 (proventriculus + ventriculus) 2 40.6-43.0 turkey 0.04 (proventriculus + ventriculus) 2 40.6-41.5 oral dosage forms the development of species-specific formulations necessitates adapting dosage forms to species-specific constraints in drug delivery, whether attributable to palatability issues, husbandry practices or pet owner compliance. examples of veterinary dosage forms for the various species have been described in the merck veterinary manual [11] and are listed in drug formularies and handbooks [12]. dogs and cats: tablets, capsules, solutions, and suspensions. given the upsurge of interest in chewable dosage forms, this topic is further discussed in detail later in this review. horses: solutions, suspensions (typically administered by nasogastric tubes), pastes (applied to the tongue), syrups, and granules (as medicated feed). cattle: medicated feed, drinking water, and oral boluses. pigs: medicated feed and oral solutions. poultry: medicated feed and drinking water. unique dosage considerations in veterinary medicine administered dose in contrast to human medicine where the dose is titrated to effect, veterinary pharmaceuticals are typically administered on a mg/kg basis. an exception is the cat, where oral formulations are typically designed to be administered as either one tablet or capsule per cat or as liquid formulations that deliver a specific fluid volume (e.g., 1 ml) per cat. in both the mg/kg and unit/cat dosing paradigms, the importance of recognizing the relationship between gastric fluid volume and bw is underscored. in veterinary species, the fasted gastric fluid volume reflects residual liquid plus ad libitum water consumption. in contrast, the human gastric volume is assumed to reflect the consumption of 8 ounces (240 ml) of water. because the biopharmaceutics classification system (bcs) drug solubility assessments (typically expressed as a “dose number” (do)) are based on the highest approved human oral dose in 240 (or 250) ml of fluid, where do = (dose strength/240ml)/drug solubility, we cannot simply extrapolate human bcs drug solubility classifications to non-human species (e.g., papich and martinez, 2015) [13]. another challenge confounding veterinary therapeutics is that upon occasion, the same medication (same formulation) may be administered to cats and dogs. however, the substantial anatomic and physiologic differences between these two species can result in differences in drug solubility and the fraction of administered dose absorbed. examples of how this may influence drug oral drug solubility and absorption are discussed later in this review. canine-specific bcs challenges given the wide range of canine body weights (approximately 4 lbs for a chihuahua or pomeranian to 230 http://dx.doi.org/10.5599/admet.1140 martinez, papich and fahmy admet & dmpk 10(1) (2022) 1-25 4 lbs for a male english mastiff), the question is whether do needs to be determined on a breed-specific basis. the percent contribution of the gi tract to total bw can differ as a function of dog breed. for example, considering the miniature poodle versus the great dane, the gi tract comprises 3-4 % of the total bw of large breeds but comprises 6-7 % bw in smaller breeds [14]. moreover, breed differences in intestinal length may affect the relationship between the rate of in vivo dissolution and the concentration of solubilized drug driving absorption across the intestinal segments. we note that a significant positive correlation was observed between bw and fecal water content in 60-week-old dogs [15], raising the question of whether there may, in fact, be some difference in the fraction of drug dissolved within the colon of large versus small breed dogs (especially if rapid intestinal absorption is not available to maintain in vivo sink conditions). while puppies have been observed to exhibit a longer orocecal transit time (octt) than adults, this difference was only statistically significant in large canine breeds [16]. body size did not significantly influence the octt across breeds of adult dogs. in that study, octt was defined as the interval from ingestion of the meal to the time at which a marker compound was detectable in plasma. thus, it primarily reflected the duration of gastric emptying and small intestinal transit. in contrast, when measuring mean total gastrointestinal transit time (ttt, time from consumption to defecation), hernot et al. (2005, 2006) observed a positive correlation to canine bw [16,17]. this was attributable to size-associated increases in large intestinal transit time. assuming that most drug absorption occurs in the small intestine, it is likely that canine differences in bw, ttt, and fecal quality will not lead to drug solubility-associated differences in oral bioavailability (with the exception of situation involving colonic absorption as noted above). fed versus fasted state similar to humans, the canine and feline gi tracts can exist either in a fed or fasted state. while we know that canine intestinal transit times tend not to be influenced by the presence or absence of food [18,19], to date, there is an absence of published information on the postprandial gi fluid composition of dogs or cats. when evaluating human drug products, the us fda recommends that formulation effects be considered under both fed and fasted conditions because, in some cases, food can magnify formulation-associated differences in product bioavailability [20]. however, due to constraints encountered in veterinary medicine, in vivo bioequivalence studies are typically conducted in fasted dogs and cats (vich gl52) [21]. exceptions are those drug products specifically labeled for administration in the fed state. furthermore, certain oral dosage forms for dogs and cats are formulated as chewy treats or are administered in treat-like pouches, which itself is likely to induce a fed or semi-fed state upon ingestion. although formulation comparisons are typically generated in fasted animals, we know that a prandial state can alter feline and canine oral drug absorption. frequently, food will increase drug solubilization (reasons discussed below). however, as summarized by watson (1979, 1986) [22,23], food can negatively influence the oral bioavailability of penicillins, cephalosporins, and tetracyclines. hernot et al. (2015) showed that feeding significantly decreased the oral absorption of minocycline in dogs (figure 1) [24]. in cats, the absorption of chloramphenicol from the palmitate ester pro-drug was significantly impaired if the cats fasted, while that of chloramphenicol tablets was not [25]. as discussed below, these two molecules exhibit similar solubility characteristics. these observations point to a range of factors that my influence postprandial differences in product bioavailability. admet & dmpk 10(1) (2022) 1-25 veterinary gastrointestinal differences influencing solubility doi: http://dx.doi.org/10.5599/admet.1140 5 figure 1. minocycline oral absorption in dogs; comparison of fed vs. fasted. all dogs were administered the same dose in a crossover study (based upon data from hnot et al. 2015) [24]. examples of potential reasons for positive or negative food effects include [26,27]: positive: food-associated increase in gastric ph enhances drug solubility (e.g., weak acids); bile salts enhance drug solubility; food-induced increase in hepatic blood flow leads to saturated 1st pass metabolism (e.g., propranolol); food-induced enhancement of lymphatic uptake, thereby bypassing first-pass hepatic metabolism. negative: food-associated increase in ph decreases drug solubility (e.g., weak bases); prolonged gastric residence time and drug instability in gastric fluids; increased fluid viscosity, which limits drug diffusion from dissolving tablet (retarding drug dissolution process); food-drug interaction (e.g., calcium and tetracycline); higher drug affinity for the bile micelle than for the intestinal unstirred water layer (micellar entrapment). the extent of the negative dissolution will depend upon the magnitude of the counteracting force imposed by the release of bile salts. the fats or fiber in a meal can lead to drug entrapment for certain compounds. note that these effects are independent of ph-associated influences on in vivo drug solubilization. when considering the decrease in free water and increase in fluid viscosity associated with a meal, humans tend to exhibit a decrease in tablet disintegration due to the decreased free water content. the meal-induced higher fluid viscosity (80 to 800-fold greater than that in the fasted human stomach) and the resulting decreased diffusivity can negatively impact in vivo dissolution [27,28]. similarly, fluid viscosity can affect oral drug bioavailability in dogs. the nature of this viscosity-induced change varies as a function of the aqueous drug solubility and its site of absorption. reppas et al. (1998) [28] observed that for highly soluble compounds absorbed primarily in the upper part of the small intestine, an increase in luminal viscosity would lead to a reduction both in peak concentrations and extent of absorption (drugs administered with guar gum to fasted dogs). in contrast, highly soluble compounds but absorbed throughout the gi tract are likely to exhibit a reduction in cmax without a decrease in auc. however, the addition of guar gum had minimal impact on the auc and cmax of poorly soluble compounds. we can anticipate similar effects in cats. regarding the other veterinary species, although fed/fasted studies may be published in horses [29] and swine [30], species such as cattle, horses, poultry, and swine typically have levels of residual material within their stomachs throughout the day. in nature, horses rarely exist in a fasted state [31]. moreover, for foodproducing species, much of the oral medications are administered either in drinking water or in medicated http://dx.doi.org/10.5599/admet.1140 martinez, papich and fahmy admet & dmpk 10(1) (2022) 1-25 6 feed. thus, for these species, oral drug bioavailability is generally determined by conditions prevailing during the fed state (note that equine oral bioavailability comparisons are accepted under fasted conditions by several non-us countries as evidenced by published equine papers) [32]. palatability dogs and cats: bad tasting or high dose drugs are a challenge to formulate into tablets that are freely accepted by dogs and cats. in addition to initial taste, mouth-feel and after-taste must also be considered. this has resulted in research efforts to support taste-masking technologies. from an evolutionary perspective, it has been suggested that the canine ancestors may have relied not only upon animal prey but also upon plant materials when prey was scarce [33]. for this reason, dogs often consume foods containing either animal-derived or vegetable-derived flavors. while they prefer meat-based or complex flavor mixtures (including animal protein digests and hydrolysates, animal proteins, emulsified meats, amino acids, animal fats), they can also be attracted to sugars such as sucrose, glucose, fructose, and lactose (but not maltose) [34,35]. this behavioral observation is consistent with the neurophysiological response to these substances, the intensity of which is influenced by the presence of monovalent cations (e.g., na+), divalent cations (e.g., ca2+), and the amount of these ions relative that of the sugar [36]. in contrast, from an evolutionary perspective, cats remained dependent upon frequent meals of small prey. while both dogs and cats exhibit a carnivore pattern of taste preferences, cats further display a differential response pattern to certain animal acids (e.g., stimulated by l-lysine but inhibited by 2tryptophan) [33]. they prefer flavors such as fish, liver, meat, sour/acidic flavors (ph range of 4.5–5.5), brewer’s yeast, yeast extract, dairy (milk and cream), and amino acids [34]. they neither have an attraction nor aversion to sweet carbohydrates [37]. based upon studies of taste-induced electrophysiological nerve activity, as with dogs, this behavior is consistent with a lack of neuronal stimulation [37]. thombre, 2004 [34] noted that dosage form texture, shape, and size are considered to be more important to cats than dogs. this appears to be related to a cat’s tendency to nibble their food, while it is chewed and rapidly consumed by dogs. cats prefer the ‘‘cheerio’’ or ‘‘star’’ to a ‘‘fish’’ shape, a finding that influences the manufacture of cat food more so than it does medicinal products. unlike the flavorants used in human medicine, many of those used for dogs and cats are lipophilic and may impose manufacturing and stability problems. this is particularly problematic when large amounts of the flavorants are needed to cover a very bitter drug substance [38]. even when artificial flavors are used, many continue to pose lipophilicity concerns. in that regard, the greater the amount of lipophilic flavorants integrated into the tablet (particularly chewable formulations), the greater the risk of negatively impacting drug product dissolution (and stability). flavorants can either be natural or artificial. these terms are defined in 21 cfr 101, 22(a)(3). horses: the general preference tends toward sweet and salty flavors [39]. however, an in-depth analysis shows that taste preferences can be influenced by the animals' breed and sex. for example, pellets containing molasses were consumed more willingly by mares than stallions. feeds with the addition of apples or carrots were the favorite treats of all tested breeds [40]. although medications are frequently administered to horses as oral pastes or suspensions, tablet formulations have been approved. for example, in a study comparing firocoxib oral paste to an oral tablet, the peak exposure and time to peak exposure were slightly greater after the paste as compared to the tablets, although the extent of firocoxib oral bioavailability from these two formulations was comparable (nada 141458, https://www.equioxx.com/). https://www.equioxx.com/ admet & dmpk 10(1) (2022) 1-25 veterinary gastrointestinal differences influencing solubility doi: http://dx.doi.org/10.5599/admet.1140 7 cattle, pigs, and poultry: with regard to these other target animal species, the types of dosage forms used (see discussion above) limit the importance of palatability concerns to drug formulation. prescribing practices due to human food safety concerns, it is not legally permissible to administer drugs in an extralabel manner to food-producing animal species such as cattle, swine, and poultry. in contrast, it is legal (and common) to prescribe drugs in an extralabel manner to dogs and cats [41]. moreover, it is not unusual to find human generic drug formulations prescribed for administration to companion animals or for drugs approved for use in dogs to be prescribed for similar indications in cats. therefore, it is important to appreciate potential interspecies differences in oral bioavailability and in vivo product performance. differences in gastric acid secretion can be particularly important in the solubilization and subsequent absorption of weak bases [42]. since the dissolving free base can neutralize gastric acid, drug solubility may decrease during the gastric residence of a basic drug substance. this issue can be particularly problematic when extrapolating human drug solubility estimates because gastric acid secretion in dogs and cats is markedly lower than that of humans [43,44]. examples of recognized in vivo differences in oral drug bioavailability are provided below. examples of dog-cat differences in oral bioavailability there are many factors influencing oral drug absorption across veterinary species, including drug solubility, gi transit time, absorptive surface area, active transporters, and metabolizing enzymes (enterocyte and the liver). while it is not always possible to distinguish the magnitude of which solubility versus permeability and presystemic metabolism contribute to these dog/cat bioavailability differences, some published examples show that drug solubility may have had an important role. please note that the studies reported below have been conducted either using experimental formulations or drugs that may not have been fda approved for use in the investigated species. the drugs discussed below were used solely to explore pharmacokinetic (pk) differences between dogs and cats and are not intended to characterize or imply therapeutic uses.  chloramphenicol: chloramphenicol is a neutral molecule with an estimated water solubility of 2.5 mg/ml. similar solubility characteristics are associated with the palmitate ester. however, the palmitate oral suspension is associated with markedly lower oral bioavailability as compared to chloramphenicol tablets. this is particularly evident when administered to fasted cats [25]. in contrast, equivalent palmitate and crystalline chloramphenicol doses produce comparable blood levels in the fasted dog [45]. the author suggests that, at least in part, the poor bioavailability of chloramphenicol from chloramphenicol palmitate in fasted cats might be due to reduced secretion of digestive enzymes in the fasting state and consequently impaired hydrolysis of the ester. since the dog and the cat were both dosed with chloromycetin palmitate suspension by parke davis and company, the observed species differences could not be attributed to potential differences in chloramphenicol palmitate polymorphic form.  cannabidiol (cbd): there is tremendous interest in the use of cbd in pets. however, poorly soluble cbd has lower oral bioavailability in cats than dogs. when administered to dogs in the form of oral chews (soft chew treat made with a glycerol/starch/fiber base), cbd had markedly higher oral absorption than cbd-infused fish oil capsules administered to cats. both formulations were administered at a dose of 2 mg/kg. but the mean maximum concentration (cmax) was 301 and 43 ng/ml in dogs and cats, http://dx.doi.org/10.5599/admet.1140 martinez, papich and fahmy admet & dmpk 10(1) (2022) 1-25 8 respectively. the time to maximum plasma concentration (tmax) values were 1.4 + 0.2 hr (dog) and 2.0 + 0.6 hr (cats). species differences in terminal elimination half-life (t½) could not explain these outcomes as the t½ in dogs was 1 hr while that of cats was 1.5 hr. the corresponding area-under-thecurve (auc) was 1297 and 164 ng hr/ml for dogs and cats, respectively [46]. the extent to which formulation versus species impacted this tremendous difference in absorption is unclear. however, since cbd itself is a highly lipophilic substance, at least in part, one may expect that differences may have been associated with a higher affinity of the cdb for the fish oil in the feline capsule (thereby preventing its solubilization within the gi fluids), than for the fluids in the feline gi tract. other possibilities to consider include species differences in gi fluid composition, or, due to the use of an extruded formulation in dogs, could have induced a fed in the dog (thereby improving drug oral bioavailability).  itraconazole: this compound is notorious for its poor oral solubility. oral absorption of this azole antifungal is highly affected by its solubility in the gi tract. in dogs, the oral absorption of oral itraconazole capsules and the oral solution (sporanox, which is drug complexed in a hydroxylpropyl‐β‐ cyclodextrin vehicle) are comparable (figure 2). relative bioavailability of the oral capsules vs. solution in dogs was 85 % [47]. in cats, there was a much greater difference between the oral bioavailability of the two dosage forms (figure 2), likely reflecting species difference in gi ph, gi fluid volume (smaller in the cat) and potentially differences in endogenous surfactants (note that we currently lack information on bile salt composition in the cat). when the oral solution (sporanox) was administered to cats, the bioavailability was approximately 5x higher than the oral capsule [48]. the pka of itraconazole is 3.7 (a weak base).  fluoroquinolones: problems associated with the solubilization of ciprofloxacin in small fluid volumes is more pronounced in cats than in dogs. accordingly, ciprofloxacin oral bioavailability in cats (10 mg/kg of ciprofloxacin powder in hard gelatin capsules) was only 22 % (%cv=50), rendering it unsuitable for treating most feline bacterial infections [49]. in dogs, dosing the oral solution (approximately 7.5 mg/kg) resulted in an oral bioavailability of about 78 % (%cv 26.4 %) and about 61 % (%cv=57.7 %) for an immediate release tablet formulation (dosed at approximately 20 mg/kg) [50].  selamectin: in contrast to the other two examples, selamectin has markedly higher oral bioavailability in cats than in dogs [51]. the oral formulation contained 24 mg selamectin/ml in sesame seed oil. a single dose of 24 mg/kg was given by oral gavage to cats or by a stomach tube in dogs. as compared to an intravenous dose, oral bioavailability was 109 % in cats but only 62 % in dogs. nevertheless, tmax (7 hrs cat, 8 hrs dog) was comparable. again, whether permeability/presystemic metabolism versus in vivo solubility was a primary cause of these differences is unclear. selamectin is a p-gp substrate. while one may deduce the potential impact of species differences in intestinal p-gp, this appears unlikely since the tissue distribution of p-gp appears to be similar in dogs, humans, and cats [52]. summary of factors influencing drug solubility within the gi track the us major veterinary species dog a primary incentive for the amount of published research available on the canine gi tract relative to that available in other veterinary species has been the use of dogs as a preclinical species for the assessment of oral absorption, formulation strategies and the pk of medications being developed for human use [53]. however, because these preclinical studies are typically performed in fasted beagle dogs of approximately 10 kg bw, there is a lack of information on how the canine gi fluid environment may vary as a function of admet & dmpk 10(1) (2022) 1-25 veterinary gastrointestinal differences influencing solubility doi: http://dx.doi.org/10.5599/admet.1140 9 age, breed, diet, and prandial state. figure 2. oral absorption of itraconazole capsules vs. solution in cats (left) and dogs (right) from hasbach et al. 2017, and mawby, et al. 2018 [47,48]. under fasted conditions, the fluid volume of the beagle stomach tends to be low. using magnetic resonance imaging from 12 beagle dogs weighing 9–12 kg (6 male and 6 female), the average gastric fluid volume was found to be 24.0 ± 4.2 ml [54]. that estimate is within the range of volumes (6 ml to 35 ml) used to examine the applicability of the human bcs system application for dogs [13]. the oral absorption of fluoroquinolone antimicrobials illustrates the impact of human-canine solubility differences and, at least in part, its contribution to species differences in oral bioavailability.  ciprofloxacin tablets are formulated for people and not for dogs. although they can be administered extralabel manner, the oral absorption in dogs was shown to be lower and much more variable than that observed in people [55,56] (keep in mind that as discussed above, even lower oral bioavailability is expected to be associated with oral ciprofloxacin administration to cats). the authors suggest that one explanation for the low and variable oral absorption in dogs is their low gastric volume. the canine oral dose was administered with a 12 ml water flush. although ciprofloxacin is classified as a highly soluble drug (based on bcs criteria), this assumes a fluid volume of 250 ml. since its solubility is approximately 10 mg/ml, for even the largest sized human tablet (750 mg), the dose number (d0) for humans = 0.3. a d0 < 1.0 is considered highly soluble [57]. clearly, a different situation exists in dogs when they consume the tablet with a water volume of 12 ml (the volume of the oral flush). with that small fluid volume, the calculated d0 for ciprofloxacin in dogs was 2.08, which would classify ciprofloxacin as a poorly soluble drug in dogs. in contrast, when ciprofloxacin was administered as an oral solution in the same study, it was dissolved in a much larger volume of water (37 ml). the d0 calculated for this volume was 0.68, rendering it highly soluble. accordingly, a markedly higher fraction of administered dose was absorbed. the finding from that study is consistent with the work of martinez et al., 2017 [50], where it was observed that unless the ciprofloxacin dose had been fully solubilized by the time it reaches the upper portion of the small intestine, it bypasses the canine absorption window, leading to poor oral bioavailability.  a different situation exists for levofloxacin [58]. levofloxacin tablets formulated for humans and not for dogs were nearly 100 % absorbed in dogs and did not exhibit the variability seen in the previous studies using ciprofloxacin. the higher and more predictable oral absorption of levofloxacin can be explained by its higher water solubility. with a reported water solubility of approximately 200 mg/ml, levofloxacin succeeds in meeting the d0 criteria of highly soluble in dogs. http://dx.doi.org/10.5599/admet.1140 martinez, papich and fahmy admet & dmpk 10(1) (2022) 1-25 10 there is little published information available on the composition of fluids in the canine gi tract in the fed state. however, there is evidence that in contrast to humans, canine food consumption does not produce an initial increase in stomach ph. this human-canine difference may be attributable to the lower basal peak acid secretion known to occur in dogs [53]. the usp gc<1236> provides the components and composition of biorelevant media describing the gastric and intestinal fluids of the fasted dog. due to the variations in reported canine gastric ph values, the fasted gastric fluids are described for a ph range of 1.2 – 2.5 or 2.2 – 6.5. this is somewhat different from the ph range of 0.9 to 2.5 found using the bravor capsule in dogs [59,60]. this range of variable results (particularly if it reflects within and between-dog variability) can affect our predictions of canine drug solubility. the canine small intestines tends to exhibit a higher ph than that of humans. accordingly, weak acids are typically more soluble in the canine vs human fasted intestinal fluids. for neutral molecules and weak bases with a pka of less than 3, the difference is largely a function of the species-specific extent of bile micellization. above ph 4, the human-canine differences in the solubility of weak acids appear to be associated with the extent to which solubilization occurs when the drug is in both its ionized and its unionized forms [61]. a comparison of human and canine bile acid composition was reviewed by martinez et al., 2021 [7]. a fundamental difference between these two species is that while the ratio of human bile acid conjugation to glycine (primary) versus taurine (glycine/taurine) is 3, there is negligible conjugation to glycine in dogs. nearly all of the canine bile acid conjugation is to taurine. under fed conditions, intestinal bile salt and phospholipid content tend to be markedly higher in the dog (estimated in labradors as 18mm and 19.4 mm, respectively) as compared to humans (11.8 mm and 4.31 mm, respectively). one might anticipate that these differences in intestinal components could lead to a greater ability to solubilize certain lipophilic drugs in dogs as compared to humans [9]. furthermore, the intestinal buffering capacity (mm/∆ph) of the fasted dog is slightly higher than that of fasted humans (13.8 versus 12 for dog and human, respectively) and the osmolality (mosmol/kg) tends to be higher in humans as compared to dogs (181.6 versus 270 for dogs and humans, respectively) [61]. cats medication issues in cats have been reviewed by papich (2006) [62] and herve lefebvre cvm ph.d. decvpt and brice reynolds dvm) [63]. these reviews note that less is known about the gi fluid composition of cats as compared to dogs. however, recent evidence shows that in addition to differences in gi fluid composition and gastric volume, the gastric ph of cats differs from that of dogs. for example, tolbert et al. (2017) estimated the gastric ph of unanesthetized cats using bravo capsules that were retained in the stomach, permitting measurements for a duration of 12 hrs [64]. when averaged over the 12-hr test period, the gastric ph of healthy cats was approximately 1.6. although the cats were fed prior to capsule administration, there did not appear to be a change in gastric ph over the 12 hr period, even though cats fasted during this time. slightly different results were observed by telles et al. (2021) when using a bravo capsule to capture feline gi ph under fasted and fed conditions. in their study, the gastric and intestinal ph’s were higher under fasted versus fed conditions [65]. in addition, tremendous inter-animal variability was observed both for the gastric and intestinal phs (table 2). although not directly related to drug solubility, the following additional points need to be considered when formulating poorly soluble, slowly dissolving compounds for cats and may be of value to consider when developing in vitro dissolution procedures for feline solid oral dosage forms: admet & dmpk 10(1) (2022) 1-25 veterinary gastrointestinal differences influencing solubility doi: http://dx.doi.org/10.5599/admet.1140 11  the gastric emptying rate of 1.5 – 5 mm beads in the fed state is markedly longer than that in the fasted cat. the time to 90 % emptying (t90) was 0.74 hrs (1.5 mm bead) and 1.02 hrs (5 mm bead) in sedated fasted cats. in the fed state, the t90 of sedated cats was 6.65 hrs and 9.06 hrs for the 1.5 mm and 5.0 mm beads, respectively [66].  the short intestinal transit time of 1.5 – 5 mm beads is unaffected by the prandial state and may be slightly longer in the cat than the dog [63].  in cats, it is essential tablets are administered in a manner that prevents esophageal retention, which, in turn, can lead to esophageal stricture or esophagitis. this has been cited as an important problem for drugs formulated as acidic hydrochloride salts. the most often cited example is doxycycline hyclate (hydrochloride), which can cause severe esophageal lesions (figure 4) [67,68]. table 2. gi ph estimates in fed and fasted cats using a bravo capsule monitoring system [65] fasted fed median range median range esophageal ph 7 3.5-7.8 4.5 2.9-6.4 gastric ph 2.7 1.7-6.2 2 1.1-8.6 small intestine 8.2 7.4-8.7 8.3 7.9-8.6 large intestinal ph 8.5 7.0-8.9 7.8 6.3-8.7 when employing the bravo capsule monitoring system to estimate gi transit times (tt) in fasted and fed cats (crossover study), telles et al. (2021) [65] obtained the results given in table 3. table 3. gi tt estimates in fed and fasted cats using a bravo capsule monitoring system [65] fasted fed median range median range esophageal tt (min) 11 1-317 2 1-379 gastric tt (min) 94 1-4101 1068 484-5521 small intestinal tt (min) 1350 929-2961 1534 442-2538 total gi tt (min) 1733 1115-5741 2796 930-6590 note that while the direction of differences in fed versus fasted states was comparable to that described by chandler et al., 1997 [66], the observed transit times were somewhat different. this comparison is consistent with the influence of particle size and gi transit times.  the maximum fluid volume (l) in the small intestine and colon of the cat is markedly less than that of the dog (table 4) [69]. in addition, because of their inherent feeding habits (discussed above), cats have smaller stomachs and drink less water than dogs. table 4. small and large intestinal maximum fluid volumes of dogs and cats (bws not provided) [69]. species absolute, l relative% gi volume cat small intestine 0.11 14.6 colon 0.12 15.9 dog small intestine 1.62 23.3 cecum 0.09 1.3 colon 0.91 13.1 http://dx.doi.org/10.5599/admet.1140 martinez, papich and fahmy admet & dmpk 10(1) (2022) 1-25 12 cattle forage and feeds mix with the bovine saliva that contains sodium, potassium, phosphate, bicarbonate, and urea. the resulting bolus moves from the mouth to the reticulorumen, which can hold approximately 5 gallons of material in the mature cow (liquid plus food). the latter compartment acts in a manner similar to that of a fermentation vat, containing microbes, carbohydrates, and a variety of volatile fatty acids (refer to usp gc <1236>). the abomasum is the “true stomach” of a ruminant. it is the compartment that is most similar to a stomach in a nonruminant, with a ph of approximately 3.5 to 4.0 due to hydrochloric acid secretion [70]. when considering drug solubility in ruminants, conditions in the rumen are typically the predominating factor. for example, in a study conducted on sulfamethazine oral boluses, ruminal conditions, including the slow transit of material through the rumen, allowed for formulations exhibiting markedly different in vitro dissolution profiles to produce nearly identical oral bioavailability [71]. rumen ph is best measured at its lowest value (i.e., 2-4 hours after feeding a concentrate meal or 4-8 hours after offering a fresh total mixed ration). rumen ph can be determined using a portable ph meter. the normal ph of grass-fed ruminants is 6-7. a ph value of 5.5-6 is seen in cattle on high-grain diets or pasturefed cattle with early lactic acidosis [72]. as noted in usp gc <1236>, the normal ph of a healthy reticulo-rumen is in the range of 5.5–6.8. high grain diets typically result in a lower ruminal ph (~5.5), whereas high-forage diets result in a higher ruminal ph (~6.8). the usp gc <1236> also states that the ph of the abomasum (true stomach) is about 2–3 and that their intestinal ph is similar to that observed in monogastrics and humans. the ph at the pylorus is about 3.0 and increases to about 7.5 in the ileum ( usp gc <1236>). swine an extensive study on the postmortem gastric and intestinal fluid contents of fasted landrace pigs was published by henze et al. (2020) [73]. this information will soon be incorporated into usp gc <1236>. pigs were fasted for 24-hours prior to euthanasia. the authors compared the information obtained in their study to that published in dogs and humans (for references describing the gi fluid contents for humans, dogs, and minipigs to which the swine data were compared, please refer to the manuscript by henze et al, [73]). they observed the following:  ph gastric fluids: similar to humans, porcine fasted state ph varies from 1.7 to 3.4, with a mean value of 2.2 ± 0.7 (median: 1.9). these values are slightly higher than the fasted ph reported in yucatan minipigs (0.3–1.7). intestine fluids: values ranged between 6.3 and 7.9, with a mean ph of 7.0 ± 0.5 (median: 7.0). the observed inter-subject variability was low, indicating a consistent and well buffered intestinal ph.  buffer capacity gastric fluids: swine gastric buffer capacity is 6.1 ± 3.5 mmol * l-1 *△ph-1, which is lower than that reported for humans (14.3 ± 9.3 mmol* l-1 *△ph-1). intestinal fluids: swine small intestine buffer capacity is 19.4 ± 2.9 mmol* l-1 *△ph-1, which is 3.4 fold higher than humans and 6.9fold higher than dogs. admet & dmpk 10(1) (2022) 1-25 veterinary gastrointestinal differences influencing solubility doi: http://dx.doi.org/10.5599/admet.1140 13  osmolality gastric fluids: the osmolality in the swine stomach is 99.33 ± 53.08 mosm kg-1, which is lower than that of the human (mean of about 220 mosm kg-1) but higher than that of dogs (74.9 mosm kg-1) intestinal fluids: the osmolality in the swine small intestine is 387 ± 61 mosm kg-1samples of landrace pigs was 2-fold higher than that reported for humans (about 197 mosm kg-1) and that of dogs (69207 mosm kg-1).  bile: a major difference between the bile acids of humans versus pigs is that in humans, hyocoholic acid is only present in small amounts, whereas, in pigs, it is one of the key bile acids. in addition, while the most common bile acids in humans are cholic acid, chenodeoxycholic (cdc) and deoxycholic acid, in pigs, cdc and hyodeoxycholate acid are the major bile acids. 97.2 % of pig bile is conjugated with either glycine (69.3 %) or taurine (38.2 %). the total intestinal bile salt concentration ranged from 19.43 to 38.44 mm (median human estimate = 3.30 mm). henze et al. (2020) [73] provide highly detailed supplemental information on the specific amounts of each bile acid contained in gastric and intestinal fluids. gastric fluids: total bile acid content = 2.5 ± 1.7 mm intestinal fluids: total bile acid content = 28.3 ± 9.6 mm  phospholipids: the most prevalent phospholipid in the swine gi tract is phosphatidylcholine, which is hydrolyzed to lyso-phosphatidylcholine in the small intestinal lumen. the swine phospholipid concentration is somewhat lower than that of humans. gastric fluids: total phospholipids = 0.20 ± 0.18 mm intestinal fluids: total phospholipids = 0.37 ± 0.20 mm  cholesterol: the high cholesterol intestinal content was interpreted as a function of enterohepatic recirculation of bile juices that contain bile salts and cholesterol. it is similar to that reported in humans. gastric fluids: total phospholipids = 0.051 ± 0.060 mm intestinal fluids: total phospholipids = 1.442 ± 0.772 mm. regarding gi transit time, this was estimated in pigs in a semi-fasted state. the pigs were administered either labelled solution, tablets or pellets and the transit time was monitored via gamma scintigraphy. tablets and pellets were administered in a crossover study design. furthermore, the pigs were administered a drink of 200 ml strawberry milkshake containing a colloidal tc-99m colloidal solution. with this preparation, the time for 50 % gastric emptying of the liquid was approximately 1.4 hr, that of pellets was 2.2 hr, but the tablets movement from the stomach ranged between 5 and 6 hrs for 2 pigs and between 1.5 – 2 hr for another pig. the estimated total transit time for the various dosage forms was about 24-48 hours, except for one pig whose tablet total transit time was 72-96 hr [74]. horses horses are unique regarding oral absorption of medications. although this topic was reviewed several years ago, the author’s observations remain relevant today [75]. being herbivores, the equine gi tract is adjusted to a high carbohydrate and fiber diet. regarding the time for gastric emptying of tablets in fasted horses, a range of values have been reported. for example, the tmax of acetaminophen plasma concentrations (which is used as a marker to reflect gastric emptying across a http://dx.doi.org/10.5599/admet.1140 martinez, papich and fahmy admet & dmpk 10(1) (2022) 1-25 14 range of species, including humans) was estimated as 30 minutes in one study [76], but much betweenanimal variability was observed. in another study, corresponding fasted tmax values were estimated to range from 28 -87 minutes [77]. using nuclear scintigraphy, the time to 50 % versus 90 % gastric emptying was 30 and 138 minutes, respectively [78]. with respect to water, nearly complete movement from the stomach into the small intestine occurs within 10 minutes [79]. with regard to other sections of the gi tract, the small intestine comprises 30 % of the total digestive tract, but the passage of food is rapid at approximately 1 foot/minute going from duodenum to cecum. the hindgut is comprised of the cecum, large colon, small colon. of primary importance are the fermentation activities occurring in the hindgut, where volatile fatty acids (acetic, propionic, and butyric acids) are produced, and water is reabsorbed [80]. therefore, the “forgiving” solubilization conditions seen in cattle would not be applicable to the horse. the stomach of adult horses secretes approximately 1.5 liters of gastric juice per hour and has an acid output ranging from 4 to 60 mm hydrochloric acid per hour. the ph of gastric contents ranges from 1.5 to 7.0, depending on the region measured [81]. bermingham et al. (2020) discussed some of the issues associated with bioavailability/bioequivalence studies in horses. they note that horses are typically fed a high fiber diet (e.g., hay) which increases gastric emptying time, affects gastric ph and potentially adsorbs administered drugs [31]. specific values for gastric emptying time in fed horses depend upon what is being measured. based upon information posted online by an equine nutritionist, the horse’s stomach is mostly empty by about six hours after being fed, with nearly all the larger fibrous particles passing within 12 hours [82]. moreover, although it is possible to hold horses offfeed overnight to create a fasted state, intermittent feeding schedules that produce an empty stomach periodically may be one of the risk factors for equine gastric disease syndrome. nevertheless, while some in vivo bioequivalence studies employ fasted horses (typically fasted for 8 – 12 hours), such studies do not replicate product oral bioavailability under actual field conditions (see the previous discussion on the prandial state). care to avoid over-estimating oral drug bioavailability due to the tests being conducted in fasted horses rather than replicating prandial state under normal use conditions may be particularly important for antimicrobial agents where such over-estimations could potentially lead to ineffective treatments under field use conditions. an additional problem is a tendency for some drugs to adsorb onto ingesta (hay, feed), preventing its aqueous solubilization and delaying its oral absorption until the ingesta reach the cecum, where it can be released upon digestion. this phenomenon can lead to the “double peaks” sometimes observed in the pk profile of drugs administered to horses. baggot (1992) suggested that the double peak may reflect partial drug solubilization and absorption in the small intestine, while other portions require fermentative digestion prior to absorption in the colon or caecum [75]. water-soluble drugs have been examined for bioavailability in the equine [83]. the study focused on cephalexin, marbofloxacin, metronidazole, and fluconazole (providing a wide range of physicochemical properties). the investigators observed that drug solubility had little influence on the drug’s oral absorption in horses. rather, the authors noted that oral absorption was correlated more with the drugs’ lipophilicity (and ability to undergo transcellular absorption) than with its aqueous solubility. the exception was metronidazole which has a very low molecular weight and is likely absorbed by paracellular transport. thus, the magnitude to which drug solubilization influences the observed fraction absorbed remains unresolved. another difference unique to the horse is that of its bile flow. since horses lack a gall bladder, their bile flows continuously through the biliary tract into the duodenum. food undergoing digestion leaving the stomach will quickly rise to a ph of 7.0 or slightly above [84]. admet & dmpk 10(1) (2022) 1-25 veterinary gastrointestinal differences influencing solubility doi: http://dx.doi.org/10.5599/admet.1140 15 poultry although the majority of poultry research has been conducted in chickens, for the most part, the information generated for the chicken gi tract can be applied to turkeys. a description of the chicken stomach, which consists of two chambers (proventriculus and gizzard), has been summarized by zootecnica international [85]:  proventriculus: the site of acid secretion. its thick walls are lined with gastric glands that secrete hydrochloric acid, pepsin, and mucus. this is the primary site of drug solubilization.  gizzard: the site of the physical digestion phase where most gastric proteolysis occurs. the muscles of the gizzard are arranged to allow for both a rotary and crushing action during contractions, thereby grinding the ingested feed. the gizzard epithelium is coated with a koilin layer which protects it from acid, proteolytic enzymes, and physical damage. the flow of feed particles into the small intestines is regulated by the pyloric sphincter, which restricts exit based upon particle size. the retention time of feed in the gizzard is about 30-60 minutes. in a review by svihus (2011), he notes that the threshold size for being constrained from leaving the gizzard in chickens is between 0.5 and 1.5 mm [86]. he later notes that the gizzard content ph of broiler chickens varied between 1.9 and 4.5, with an average value of 3.5. however, due to the high calcium carbonate content in the diet, ph values for gizzard contents are commonly between 4 and 5 for layer hens, although a ph around 3.5 has also been reported for laying hens [87]. note that the transfer of digested materials between the proventriculus and the gizzard can occur up to 4 times per minute. clearly, the gi environment can vary as a function of poultry breed and diet. nevertheless, for estimation purposes, generalizations are often made based on broiler chickens. gauthier (2002) summarized the ph and transit time (min) of poultry gi segments [88]. table 5 represents the ph and mean transit time duration of all mash feed in different compartments of the broiler gut after 6-weeks of ad libitum feeding. additional information on the enzymes within each section is based upon a symposium presentation by rob porter, 2012 [89]. table 5. ph, transit time, and enzymes associated with the various portions of the broiler gi tract [88,89] gi segment transit time, min ph enzymes mouth 7-7.5 amylase crop 50 5.5 none mucous secretion proventriculous and gizzard 90 2.5-3.5 pepsin, lipase duodenum 5-8 5-6 amylase, trypsin, collagenase, bile, lipase jejunum 20-30 6.5-7 maltase, lactase, peptidases ileum 50-70 7-7.5 impact on formulation strategies: how species gi fluid differences may impact formulation and how formulation can influence in vivo product performance with these species-specific differences in mind, the following questions and answers provide perspectives on how formulations may need to be adjusted to optimize drug solubility (and therefore oral bioavailability) in veterinary medicine. it should be noted that unlike the literature available for formulation optimization in human medicine, there is negligible corresponding information published for veterinary species. therefore, this section reflects http://dx.doi.org/10.5599/admet.1140 martinez, papich and fahmy admet & dmpk 10(1) (2022) 1-25 16 the perspectives of the authors based upon general published information and formulation principles. where possible, we include citations to support our perspectives. how might poor drug solubility influence tablet formulation for dogs and cats? typically, formulation approaches used in human medicine [90] can be used to enhance the oral solubility of drugs administered to dogs and cats. however, the shorter gi tract and corresponding reduced time for drug absorption in dogs and cats introduce an additional challenge associated with formulation development. the other challenge is the smaller volume of fluid in the gi tract available for drug dissolution. this problem is even more pronounced in cats as compared to that of dogs. because solubilization must occur within a much shorter timeframe in dogs [7,91] and cats (see above), compounds exhibiting poor oral solubility may fail to attain the bioavailability needed to achieve therapeutic plasma drug concentrations. in these situations, it may be appropriate to administer the therapeutic moiety in a more readily soluble salt form [6] or as a prodrug. an example of this is the veterinary use of enrofloxacin versus ciprofloxacin [92]. ciprofloxacin is known to exhibit inconsistent and poor oral bioavailability in dogs [55,56] and cats [49], while orally administered enrofloxacin is almost completely bioavailable in both species (https://bayer.cvpservice.com/product/basic/view/1040011). what are special formulation and manufacturing considerations needed when forming chewy, treat-like formulations? although we do not have information available on tablet hardness preferences for dogs, we do have some insights on that subject for cats. when formulating feline chewable tablets, the generalized “ideal” hardness is in the range of 3–6 kp. when hardness exceeds 6 kp, tablet palatability tends to decrease. for example, while a tablet with a hardness of 6 kp may be associated with a 95 % free choice acceptance, this value is reduced to 50 % free choice when formulated with a 12 kp tablet hardness [93]. the texture is an issue for chewable medications and can make the tablet appealing or unappealing to dogs and cats. although chewable tablets can be made by direct compression, wet granulation (using either water or alcohol), dry granulation (slugging or roller compaction), extrusion, or a forming machine [93], frequently, the “edible soft chew” is made using an extrusion or forming process. both processes involve the mixing of the drug with tablet excipients (which include polymer, solvents such as polyethylene glycol (peg) 30, glycerin, vegetable oil), binder, filler, stabilizer and flavorant) prior to the introduction of this mixture into an extruder or forming machine. the critical differences between these two processes pertain to the temperature and pressure at which the chews are produced, both being much higher for the extruder than the forming machine. these can be important considerations for heat-labile drugs or when using heatsensitive solvents. however, both processes can be incorporated into the generation of tablets of varied sizes and shapes. with regard to shape preferences, refer to the earlier discussion on tablet palatability. how might flavorants influence in vivo drug solubility? flavorants are available as commodity items from houses specializing in flavors for cats and dogs. some potential flavorants are water-soluble (e.g., apple flavor or sugar or butter scotch flavor or lactose). however, others are poorly soluble, such as beef, liver, pork, chicken flavor, wheat germ, vegetable oil, or gelatin [94]. species specific taste preferences have been discussed above (see the section on palatability). the use of lipophilic flavorants can negatively impact the solubility of some compounds. in contrast, it https://bayer.cvpservice.com/product/basic/view/1040011 admet & dmpk 10(1) (2022) 1-25 veterinary gastrointestinal differences influencing solubility doi: http://dx.doi.org/10.5599/admet.1140 17 could also potentially enhance the solubility of certain lipophilic compounds by inducing a semi-fed state and a release of bile salts. in that case, the positive or negative effect on in vivo dissolution will depend upon fed/fasted factors, as previously discussed. how might flavorants be incorporated into the formulation? the manner in which the flavoring agents are integrated into the formulation depends upon the dosage form. for compressed tablets, if the flavorant comprises only a small percentage of the total tablet weight, no special manufacturing considerations are needed. however, the technique for taste masking may not simply be its addition as one of the tablet ingredients. for example:  if the active agent is extremely bitter, taste-masking may need to be accomplished by drug encapsulation. this may involve additional manufacturing steps to coat the tablet with a polymer, adding complexity to the drug formulation.  alternatively, the addition of more flavorant may be needed to encourage consumption of chewable products manufactured by such processes as extrusion or forming machines. if the formulation is hard chewable, it can be made by direct compression, wet granulation (using either water or alcohol), or dry granulation (slugging or roller compaction). in that case, the flavoring agents can be applied using techniques similar to that used for human tablets. are there excipients that may behave differently in dogs versus cats? of particular concern is the effect that interspecies differences in gi ph and volume can have on the choice of disintegrant. zhao and augsburger [95] studied the effect of the ph on the disintegrant and concluded that an acidic medium significantly reduces the liquid uptake rate and capacity of sodium starch glycolate (primojel) and croscarmellose sodium (ac-di-sol) but not of crospovidone nf (polyplasdone xl10). tablets containing croscarmellose sodium were less affected by the acidic medium than were those formulated with sodium starch glycolate. therefore, given the tendency towards a lower gastric ph in the cat than in the dog, disintegrants such as sodium starch glycolate and croscarmellose sodium may not function well when included in formulations intended for administration to cats. furthermore, since many formulations are used both in dogs and cats, it may be preferable to use a disintegrant such as crospovidone. it should also be noted that while excipients associated with enteric coating have been explored for use in dogs [96,97], the ph at which drug release occurs needs to be carefully evaluated to insure the ability to withstand fluctuations in canine gastric ph and the location of drug release does not compromise oral bioavailability (when considering the relatively rapid transit through the small intestine). clearly, more work is needed on this topic, with a focus on the relationship between formulation selection and the drug molecule. with regard to cats, although there have been discussions of nutraceuticals such as enteric-coated s-adenosylmethionine (same) in cats [98], there is a lack of pk-based evaluations of the impact of excipient selection on the in vivo performance of enteric-coated formulations. therefore, more data are needed to ascertain whether the optimal choice of enteric coating for a given drug will be the same for dogs and cats. how can crystalline form and particle size influence the drug solubility and in vivo performance of medicated premixes? given the effect of crystalline form on drug solubility [99], it is important to consider retention time in the stomach, rumen, or gizzard. while a food-induced delay in gastric emptying and release of bile salts may reduce the impact of this effect in ruminants and swine, it may be a concern in poultry, particularly if the particle size is about 1 mm or less (thereby enabling it to leave the gizzard undissolved [86]). should this occur, it could be associated with decreased therapeutic effects. an example of where this issue could be http://dx.doi.org/10.5599/admet.1140 martinez, papich and fahmy admet & dmpk 10(1) (2022) 1-25 18 problematic is coccidiosis in poultry where the locally acting drug needs to be fully solubilized upon entering the duodenum [100]. alternatively, given the prevailing conditions in the bovine rumen, in vivo drug solubilization may be more forgiving for formulations administered to cattle [101]. what are issues to be considered with regard to species differences in critical formulation variables with the possible exception of the dog, little work has been published on the influence of species-specific gi fluid environments on drug solubilization or on in vivo drug product performance. however, by combining what is known about the gi fluid conditions of each species, along with information obtained from studies supporting human drug development, we can anticipate some of the challenges that may be worthy of consideration. for horses, ruminants, and pigs, medications are often administered to fed animals or in medicated feed. therefore, drug adsorption onto the gi contacts may influence in vivo drug solubilization [102,103]. on the one hand, adsorption of drug molecules onto the surface of excipients or materials such as hay can reduce drug particle size and increase the surface area of the drug available to the dissolution medium. conversely, if the forces of attraction are high, desorption may be retarded, and absorption compromised. accordingly, the potential for drug adsorption may need to be considered when formulating dosage forms such as oral suspensions for horses or medicated articles for cattle and swine. for drugs administered in medicated feeds to poultry, drug particle size and crystalline form will be essential considerations in product manufacture. for dogs and cats, many of the critical formulation variables have been discussed above. additional points to consider are: ruminants: given the relatively high ph present in the rumen, there should be no problems in terms of the solubilization of acids. however, depending upon its pka, weak bases may exist in an unionized form. for the latter, the presence of endogenous surfactants may be anticipated to enhance drug solubilization. in addition, the long residence time of material within the rumen may help promote the solubilization of the drug. unless there is an excipient present that can affect drug permeability, intestinal transit time, or presystemic drug metabolism, once the dissolved medication moves into the small intestines, oral bioavailability will likely depend primarily on the drug’s pk properties. horses: in contrast to the other veterinary species, the horse has the capability to ferment food (with the corresponding presence of surfactants and volatile fatty acids) in the hindgut, which comprises more than 50 % of the gi tract capacity. therefore, along with its acidic ph, one may expect that additional solubilization can occur with this region [104]. the question is then the pka of the compound, the magnitude to which the drug has already been absorbed in the upper portion of the gi tract, and the extent to which drug absorption can occur within the lower gi tract of the horse. another key issue that has been explored in the horse is the efficiency of various types of formulations to prevent drug gastric exposure. for example, omeprazole can be degraded by gastric acids. in a study of 5 formulations, it was observed that while the enteric-coated formulations tended to have modestly higher cmax and auc values as compared to buffered suspensions when administered fasted horses, these differences were not statistically (or clinically) significant [105]. closing comments while species differences in gi fluid volume and composition can affect drug solubilization, this admet & dmpk 10(1) (2022) 1-25 veterinary gastrointestinal differences influencing solubility doi: http://dx.doi.org/10.5599/admet.1140 19 constitutes only one of the factors influencing species-specific drug absorption characteristics. other variables include gi transit time (dictating the time available for drug solubilization/product dissolution), and permeability. moreover, the absorption process itself can promote the presence of in vivo sink conditions by reducing the amount of dissolved drug within the gut, thereby promoting the dissolution of products containing drugs with borderline solubility. in addition, although not a topic for this review, enterocyte metabolizing enzymes, intestinal influx and efflux transporters, and the possible influence of microbial gut flora can impact the amount of free drug remaining within the gi tract (i.e., affecting the fraction absorbed). when considered in combination with species-appropriate dosage forms and palatability, these physiological characteristics influence the formulation and manufacturing considerations to be incorporated into the development of veterinary orally administered drug products. important information gaps continue to influence our ability to generate in vivo solubilization predictions. this includes: 1. little to no information on breed differences in physiological surfactants, buffers, ionic composition, ph, and fluid volume. studies to generate this information are greatly needed. 2. more information is needed on the influence of diet on gi fluid composition under fed conditions for nearly all veterinary species. this is particularly important for dogs and cats, where most oral bioavailability studies are conducted in the fasted state. 3. since the human-pet interaction with dogs, cats and horses can cover their entire life span, more information is needed on changes in companion animal gi fluid composition with age. 4. there is also the need for information on intestinal fluid composition as a function of the specific segment of the gi tract. while many of these issues will have less importance in species where oral drug products are administered in drinking water or in medicated feed, it will be an important consideration for dogs and cats. furthermore, it would be helpful to have more information generated on the gi fluid composition in food-producing species and potential variability as a function of breed and diet. such information could be of value in understanding how components of medicated feed or drug crystalline characteristics and particle size may influence in vivo drug absorption. within animal health, funding limitations challenge opportunities to explore these critical questions. clearly, each of these points can influence drug and species-specific formulation optimization or improve the prediction of oral bioavailability difficulties that can occur when dogs and cats are administered oral formulations in a manner other than that indicated on the fda-approved drug label. hopefully, there will be a greater emphasis on filling these information gaps in the future. conflict of interest: dr. mn martinez, r fahmy, and mg papich have no conflicts of interest to declare. references [1] usp general chapter <1236> published 12/2019. [2] d.r. brocks, n.m. davies. lymphatic drug absorption via the enterocytes: pharmacokinetic simulation, modeling, and considerations for optimal drug development. j. pharm. pharm. sci. 21 (2018) 254s270s. https://doi.org/10.18433/jpps30217. [3] s.a. peters, c.r. jones, a.l. ungell, o.j. hatley. predicting drug extraction in the human gut wall: assessing contributions from drug metabolizing enzymes and transporter proteins using preclinical models. clin. pharmacokinet. 55 (2016) 673-696. https://doi.org/10.1007/s40262-015-0351-6. http://dx.doi.org/10.5599/admet.1140 https://doi.org/10.18433/jpps30217 https://doi.org/10.1007/s40262-015-0351-6 martinez, papich and fahmy admet & dmpk 10(1) (2022) 1-25 20 [4] a. dokoumetzidis, p. macheras. a century of dissolution research: from noyes and whitney to the biopharmaceutics classification system. int. j. pharm. 321 (2006) 1-11. https://doi.org/10.1016/j.ijpharm.2006.07.011. [5] a. avdeef, e. fuguet, a. llinas, c. rafols, e. bosch, g. volgyi, t. verbic, e. boldyreva, k. takacs-novak. equilibrium solubility measurement of ionizable drugs – consensus recommendations for improving data quality. admet & dmpk 4 (2016 117-178. https://doi.org/10.5599/admet.4.2.292. [6] a.t. serajuddin. salt formation to improve drug solubility. adv. drug. deliv. rev. 59 (2007) 603-616. https://doi.org:10.1016/j.addr.2007.05.010. [7] m.n. martinez, j.p. mochel, s. neuhoff, d. pade. comparison of canine and human physiological factors: understanding interspecies differences that impact drug pharmacokinetics. aaps j. 23 (2021) 59. https://doi.org/10.1208/s12248-021-00590-0. [8] t.s. wiedmann, l. kamel. examination of the solubilization of drugs by bile salt micelles. j. pharm. sci. 91 (2002) 1743-1764. https://doi.org/10.1002/jps.10158. [9] n. pavlović, s. goločorbin-kon, m. ðanić, b. stanimirov, h. al-salami, k. stankov, m. mikov. bile acids and their derivatives as potential modifiers of drug release and pharmacokinetic profiles. front. pharmacol. 9 (2018) 1283. https://doi.org:10.3389/fphar.2018.01283. [10] cvm gfi #171 demonstrating bioequivalence for soluble powder oral dosage form products and type a medicated articles containing active pharmaceutical ingredients considered to be soluble in aqueous media, 2021. https://www.fda.gov/regulatory-information/search-fda-guidance-documents/cvm-gfi-171-demonstrating-bioequivalence-soluble-powder-oral-dosage-form-products-andtype-medicated (accessed 09/18/21). [11] p.t. reeves, c. roesch, m.n raghnaill (2015). routes of administration and dosage forms. in merck manual veterinary manual, https://www.merckvetmanual.com/pharmacology/pharmacologyintroduction/routes-of-administration-and-dosage-forms. merck & co., inc., kenilworth, nj, usa (accessed 09/18/21). [12] m.g. papich. papich handbook of veterinary drugs, 5th edition. elsevier, st. louis. 2020. 1056 pages. isbn 98780323709576. https://www.elsevier.com/books/papich-handbook-of-veterinary-drugs/papich/978-0-323-70957-6. [13] m.g. papich, m.n. martinez. applying biopharmaceutical classification system (bcs) criteria to predict oral absorption of drugs in dogs: challenges and pitfalls. aaps j. 17 (2015) 948-964. https://doi.org/10.1208/s12248-015-9743-7. [14] h. meyer, e. kienzle, j. zentek. body size and relative weights of gastrointestinal tract and liver in dogs. j. vet. nutr. 2 (1993) 31-35. https://pdfs.semanticscholar.org/b4fb/829d3ff7f4abe988592b17163fba42a13803.pdf [15] m.p. weber, l.j. martin, v.c. biourge, p.g. nguyen, h.j. dumon. influence of age and body size on orocecal transit time as assessed by use of the sulfasalazine method in healthy dogs. am. j. vet. res. 64 (2003) 1105-1109. https://doi.org/10.2460/ajvr.2003.64.1105. [16] d.c. hernot, v.c. biourge, l.j. martin, h.j. dumon, p.g. nguyen. relationship between total transit time and faecal quality in adult dogs differing in body size. j. anim. physiol. anim. nutr. (berl) 89 (2005) 189193. https://doi.org/10.1111/j.1439-0396.2005.00544.x. [17] d.c. hernot, h.j. dumon, v.c. biourge, l.j. martin, p.g. nguyen. evaluation of association between body size and large intestinal transit time in healthy dogs. am. j. vet. res. 67 (2006) 342-347. https://doi.org/10.2460/ajvr.67.2.342. [18] m. koziolek, m. grimm, t. bollmann, k.j. schäfer, s.m. blattner, r. lotz, g. boeck, w. weitschies. characterization of the gi transit conditions in beagle dogs with a telemetric motility capsule. eur. j. pharm. biopharm. 136 (2019) 221-230. https://doi.org:10.1016/j.ejpb.2019.01.026. [19] m. koziolek, f. schneider, m. grimm, c. modeβ, a. seekamp, t. roustom, w. siegmund, w. weitschies. intragastric ph and pressure profiles after intake of the high-caloric, high-fat meal as used for food effect studies. j. control. release 220 (2015) 71-78. https://doi.org/10.1016/j.jconrel.2015.10.022. https://doi.org/10.1016/j.ijpharm.2006.07.011 https://doi.org/10.1016/j.ijpharm.2006.07.011 https://doi.org/10.5599/admet.4.2.292 https://doi.org:10.1016/j.addr.2007.05.010 https://doi.org/10.1208/s12248-021-00590-0 https://doi.org/10.1002/jps.10158 https://doi.org:10.3389/fphar.2018.01283 https://www.fda.gov/regulatory-information/search-fda-guidance-documents/cvm-gfi-171-demonstrating-bioequivalence-soluble-powder-oral-dosage-form-products-and-type-medicated https://www.fda.gov/regulatory-information/search-fda-guidance-documents/cvm-gfi-171-demonstrating-bioequivalence-soluble-powder-oral-dosage-form-products-and-type-medicated https://www.fda.gov/regulatory-information/search-fda-guidance-documents/cvm-gfi-171-demonstrating-bioequivalence-soluble-powder-oral-dosage-form-products-and-type-medicated https://www.merckvetmanual.com/pharmacology/pharmacology-introduction/routes-of-administration-and-dosage-forms https://www.merckvetmanual.com/pharmacology/pharmacology-introduction/routes-of-administration-and-dosage-forms https://www.elsevier.com/books/papich-handbook-of-veterinary-drugs/papich/978-0-323-70957-6 https://www.elsevier.com/books/papich-handbook-of-veterinary-drugs/papich/978-0-323-70957-6 https://doi.org/10.1208/s12248-015-9743-7 https://pdfs.semanticscholar.org/b4fb/829d3ff7f4abe988592b17163fba42a13803.pdf https://pdfs.semanticscholar.org/b4fb/829d3ff7f4abe988592b17163fba42a13803.pdf https://doi.org/10.2460/ajvr.2003.64.1105 https://doi.org/10.1111/j.1439-0396.2005.00544.x https://doi.org/10.2460/ajvr.67.2.342 https://doi.org/10.1016/j.ejpb.2019.01.026 https://doi.org/10.1016/j.jconrel.2015.10.022 admet & dmpk 10(1) (2022) 1-25 veterinary gastrointestinal differences influencing solubility doi: http://dx.doi.org/10.5599/admet.1140 21 [20] guidance for industry. food-effect bioavailability and fed bioequivalence studies, cder, 2002. https://www.fda.gov/files/drugs/published/food-effect-bioavailability-and-fed-bioequivalencestudies.pdf (accessed 09/18/21). [21] guidance for industry, bioequivalence: blood level bioequivalence study, gl52, 2016. https://www.fda.gov/media/89840/download (accessed 09/18/21). [22] a.d.j. watson. some factors affecting bioavailability of antimicrobial drugs given by mouth in small animals. the veterinary annual, 19th issue, eds. g.s.c. grunsell, f.w.g. hill (eds). 1979, 217-222. [23] a.d. watson. influence of food on absorption of antimicrobial drugs. in: comparative veterinary pharmacology, toxicology and therapy, a. s. j. p. a. m. van miert, m. g. bogaert, m. debackere (eds). springer, nature, switzerland, 1986, pp. 93-104. isbn: 978-94-009-4153-3. https://link.springer.com/book/10.1007/978-94-009-4153-3. [24] m.l. hnot, l.k. cole, g. lorch, m.g. papich, p.j. rajala-schultz, j.b. daniels. evaluation of canine-specific minocycline and doxycycline susceptibility breakpoints for meticillin-resistant staphylococcus pseudintermedius isolates from dogs. vet. dermatol. 26 (2015) 334-338, e70-1. https://doi.org/10.1111/vde.12227. [25] a.d. watson. effect of ingesta on systemic availability of chloramphenicol from two oral preparations in cats. j. vet. pharmacol. ther. 2 (1979) 117-121. https://doi.org/10.1111/j.13652885.1979.tb00363.x. [26] c. pentafragka, m. symillides, m. mcallister, j. dressman, m. vertzoni, c. reppas. the impact of food intake on the luminal environment and performance of oral drug products with a view to in vitro and in silico simulations: a pearrl review. j. pharm. pharmacol. 71 (2019) 557-580. https://doi.org/10.1111/jphp.12999. [27] c. pentafragka, m. vertzoni, j. dressman, m. symillides, k. goumas, c. reppas. characteristics of contents in the upper gastrointestinal lumen after a standard high-calorie high-fat meal and implications for the in vitro drug product performance testing conditions. eur. j. pharm. sci. 155 (2020) 105535. https://doi.org10.1016/j.ejps.2020.105535. [28] c. reppas, g. eleftheriou, p. macheras, m. symillides, j.b. dressman. effect of elevated viscosity in the upper gastrointestinal tract on drug absorption in dogs. eur. j. pharm. sci. 6 (1998) 131-139. https:/doi.org/10.1016/s0928-0987(97)00077-8. [29] f.j. mendoza, j.m. serrano-rodriguez, a. perez-ecija. pharmacokinetics of meloxicam after oral administration of a granule formulation to healthy horses. j. vet. intern. med. 33 (2019) 961-967. https://doi.org/10.1111/jvim.15433. [30] p. nielsen, n. gyrd-hansen. bioavailability of oxytetracycline, tetracycline and chlortetracycline after oral administration to fed and fasted pigs. j. vet. pharmacol. ther. 19 (1996) 305-311. https://doi.org/10.1111/j.1365-2885.1996.tb00054.x. [31] e. bermingham, j.l. davis, t. whittem. study design synopsis: designing and performing pharmacokinetic studies for systemically administered drugs in horses. equine. vet. j. 52 (2020) 643650. https://doi.org/10.1111/evj.13312. [32] b.w. sykes, c. underwood, r. greer, c.m. mcgowan, p.c. mills. pharmacokinetics and bioequivalence testing of five commercial formulations of omeprazole in the horse. j. vet. pharmacol. ther. 39 (2016) 78-83. https://doi.org/10.1111/jvp.12240. [33] j.w. bradshaw. the evolutionary basis for the feeding behavior of domestic dogs (canis familiaris) and cats (felis catus). j. nutr. 136 (2006) 1927s-1931s. https://doi.org/10.1093/jn/136.7.1927s. [34] a.g. thombre. oral delivery of medications to companion animals: palatability considerations. adv. drug deliv. rev. 56 (2004) 1399-1413. https://doi.org/10.1016/j.addr.2004.02.012. [35] t.s. white, j.c. boudreau. taste preferences of the cat for neurophysiologically active compounds. physiol. psychol. 3 (1975) 405-410. https://doi.org/10.3758/bf03326850. [36] t. kumazawa, k. kurihara. large enhancement of canine taste responses to sugars by salts. j. gen. physiol. 95 (1990) 1007-1018. https://doi.org/10.1085/jgp.95.5.1007. http://dx.doi.org/10.5599/admet.1140 https://www.fda.gov/files/drugs/published/food-effect-bioavailability-and-fed-bioequivalence-studies.pdf https://www.fda.gov/files/drugs/published/food-effect-bioavailability-and-fed-bioequivalence-studies.pdf https://www.fda.gov/media/89840/download https://link.springer.com/book/10.1007/978-94-009-4153-3 https://doi.org/10.1111/vde.12227 https://doi.org/10.1111/vde.12227 https://doi.org/10.1111/j.1365-2885.1979.tb00363.x https://doi.org/10.1111/j.1365-2885.1979.tb00363.x https://doi.org/10.1111/jphp.12999 https://doi.org/10.1111/jphp.12999 https://doi.org10.1016/j.ejps.2020.105535 https://doi.org/10.1016/s0928-0987(97)00077-8 https://doi.org/10.1111/jvim.15433 https://doi.org/10.1111/j.1365-2885.1996.tb00054.x https://doi.org/10.1111/j.1365-2885.1996.tb00054.x https://doi.org/10.1111/evj.13312 https://doi.org/10.1111/jvp.12240 https://doi.org/10.1093/jn/136.7.1927s https://doi.org/10.1016/j.addr.2004.02.012 https://doi.org/10.3758/bf03326850 https://doi.org/10.1085/jgp.95.5.1007 martinez, papich and fahmy admet & dmpk 10(1) (2022) 1-25 22 [37] x. li, w. li, h. wang, d.l. bayley, j. cao, d.r. reed, a.a. bachmanov, l. huang, v. legrand-defretin, g.k. beauchamp, j.g. brand. cats lack a sweet taste receptor. j. nutr. 136 (2006) 1932s-1934s. https://doi.org/10.1093/jn/136.7.1932s. [38] r. fahmy, d. danielson, m.n. martinez. quality by design and the development of solid oral dosage forms. in: advances in delivery science and technology, ed: m.j. rathbone, a. mcdowell. 2013, 107129. springer, new york. http://doi.org/10.1007/978-1-4614-4439-8_1. [39] k. apfel, horse canada. december 25, 2019. accessed https://horse-canada.com/magazine/miscellaneous/10-amazing-facts-equine-sense-taste/. [40] i. janczarek, i. wilk, s. pietrzak, m. liss, s. tkaczyk. taste preferences of horses in relation to their breed and sex. j. equine. vet. sci. 64 (2018) 59-64. https://doi.org/10.1016/j.jevs.2018.02.010. [41] animal medicinal drug use clarification act of 1994 (amduca). https://www.fda.gov/animalveterinary/guidance-regulations/animal-medicinal-drug-use-clarification-act-1994-amduca (accessed 09/18/2021). [42] n. matsumura, a. ono, y. akiyama, t. fujita, k. sugano. bottom-up physiologically based oral absorption modeling of free weak base drugs. pharmaceutics 12 (2020) e844. https://doi.org/10.3390/pharmaceutics12090844. [43] s. emås, m.i. grossman. comparison of gastric secretion in conscious dogs and cats. gastroenterology 52 (1967) 29-34. https://www.gastrojournal.org/article/s0016-5085(67)80097-0/pdf. [44] j. tibbitts. issues related to the use of canines in toxicologic pathology--issues with pharmacokinetics and metabolism. toxicol. pathol. 31 (2003) 17-24. https://doi.org/10.1080/01926230390174896. [45] a.d. watson. plasma chloramphenicol levels in dogs after the administration of four different oral preparations of chloramphenicol. aust. vet. j. 49 (1973) 460-462. https://doi.org/10.1111/j.17510813.1973.tb09292.x. [46] k.a. deabold, w.s. schwark, l. wolf, j.j. wakshlag. single-dose pharmacokinetics and preliminary safety assessment with use of cbd-rich hemp nutraceutical in healthy dogs and cats. animals (basel) 9 (2019) e832. https://doi.org/10.3390/ani9100832. [47] a.e. hasbach, d.k. langlois, e.j. rosser, m.g. papich. pharmacokinetics and relative bioavailability of orally administered innovator-formulated itraconazole capsules and solution in healthy dogs. j. vet. intern. med. 31 (2017) 1163-1169. https://doi.org/10.1111/jvim.14779. [48] d.i. mawby, j.c. whittemore, l.e. fowler, m.g. papich. comparison of absorption characteristics of oral reference and compounded itraconazole formulations in healthy cats. j. am. vet. med. assoc. 252 (2018) 195-200. https://doi.org/10.2460/javma.252.2.195. [49] g.a. albarellos, v.e. kreil, m.f. landoni. pharmacokinetics of ciprofloxacin after single intravenous and repeat oral administration to cats. j. vet. pharmacol. ther. 27 (2004) 155-162. https://doi.org/10.1111/j.1365-2885.2004.00573.x. [50] m.n. martinez, b. mistry, v. lukacova, k.a. lentz, j.e. polli, s.w. hoag, t. dowling, r. kona, r.m. fahmy. exploring canine-human differences in product performance. part ii: use of modeling and simulation to explore the impact of formulation on ciprofloxacin in vivo absorption and dissolution in dogs. aaps j. 19 (2017) 712-726. https://doi.org/10.1208/s12248-017-0055-y. [51] s p. sarasola, a.d. jernigan, d.k. walker, j. castledine, d.g. smith, t.g. rowan. pharmacokinetics of selamectin following intravenous, oral and topical administration in cats and dogs. j. vet. pharmacol. ther. 25 (2002) 265-272. https://doi.org/10.1046/j.1365-2885.2002.00415.x. [52] s. van der heyden, k. chiers, r. ducatelle. tissue distribution of p-glycoprotein in cats. anat. histol. embryol. 38 (2009) 455-460. https://doi.org/10.1111/j.1439-0264.2009.00972.x. [53] j.b. dressman. comparison of canine and human gastrointestinal physiology. pharm. res. 3 (1986) 123-131. https://doi.org/10.1023/a:1016353705970. [54] c. wang, b. zhai, h. guo, p. wang, z. liu, j. gu, h. ho, p. langgut, k. li, c. wang, z. hao, in vivo measurement of gastric fluid volume in anesthetized dogs, j. drug deliv. sci. technol 55 (2020) p.101488. https://doi.org/10.1016/j.jddst.2019.101488. https://doi.org/10.1093/jn/136.7.1932s https://scholar.google.com/citations?user=lvts-2caaaaj&hl=en&oi=sra http://doi.org/10.1007/978-1-4614-4439-8_1 https://horse-canada.com/magazine/miscellaneous/10-amazing-facts-equine-sense-taste/ https://horse-canada.com/magazine/miscellaneous/10-amazing-facts-equine-sense-taste/ https://doi.org/10.1016/j.jevs.2018.02.010 https://www.fda.gov/animal-veterinary/guidance-regulations/animal-medicinal-drug-use-clarification-act-1994-amduca https://www.fda.gov/animal-veterinary/guidance-regulations/animal-medicinal-drug-use-clarification-act-1994-amduca https://doi.org/10.3390/pharmaceutics12090844 https://doi.org/10.3390/pharmaceutics12090844 https://www.gastrojournal.org/article/s0016-5085(67)80097-0/pdf https://www.gastrojournal.org/article/s0016-5085(67)80097-0/pdf https://doi.org/10.1080/01926230390174896 https://doi.org/10.1111/j.1751-0813.1973.tb09292.x https://doi.org/10.1111/j.1751-0813.1973.tb09292.x https://doi.org/10.3390/ani9100832 https://doi.org/10.1111/jvim.14779 https://doi.org/10.2460/javma.252.2.195 https://doi.org/10.1111/j.1365-2885.2004.00573.x https://doi.org/10.1111/j.1365-2885.2004.00573.x https://doi.org/10.1208/s12248-017-0055-y https://doi.org/10.1046/j.1365-2885.2002.00415.x https://doi.org/10.1111/j.1439-0264.2009.00972.x https://doi.org/10.1023/a:1016353705970 https://doi.org/10.1016/j.jddst.2019.101488 admet & dmpk 10(1) (2022) 1-25 veterinary gastrointestinal differences influencing solubility doi: http://dx.doi.org/10.5599/admet.1140 23 [55] m.g. papich. ciprofloxacin pharmacokinetics and oral absorption of generic ciprofloxacin tablets in dogs. am. j. vet. res. 73 (2012) 1085-1091. https://doi.org/10.2460/ajvr.73.7.1085. [56] m.g. papich. ciprofloxacin pharmacokinetics in clinical canine patients. j. vet. intern. med. 31 (2017) 1508-1513. https://doi.org/10.1111/jvim.14788. [57] l.x. yu, g.l. amidon, j.e. polli, h. zhao, m.u. mehta, d.p. conner, v.p. shah, l.j. lesko, m.l. chen, v.h. lee, a.s. hussain. biopharmaceutics classification system: the scientific basis for biowaiver extensions. pharm. res. 19 (2002) 921-925. https://doi.org/10.1023/a:1016473601633. [58] m. madsen, k. messenger, m.g. papich. pharmacokinetics of levofloxacin following oral administration of a generic levofloxacin tablet and intravenous administration to dogs. am. j. vet. res. 80 (2019) 957962. https://doi.org:10.2460/ajvr.80.10.957. [59] k. sagawa, f. li, r. liese, s.c. sutton. fed and fasted gastric ph and gastric residence time in conscious beagle dogs. j. pharm. sci. 98 (2009) 2494-2500. https://doi.org//10.1002/jps.21602. [60] m. akimoto, n. nagahata, a. furuya, k. fukushima, s. higuchi, t. suwa. gastric ph profiles of beagle dogs and their use as an alternative to human testing. eur. j. pharm. biopharm. 49 (2000) 99-102. https://doi.org/10.1016/s0939-6411(99)00070-3. [61] p.l. walsh, j. stellabott, r. nofsinger, w. xu, d. levorse, k. galipeau, f. kesisoglou. comparing dog and human intestinal fluids: implications on solubility and biopharmaceutical risk assessment. aaps pharmscitech 18 (2017) 1408-1416. https://doi.org/10.1208/s12249-016-0611-2. [62] papich mg. drug therapy in cats: precautions and guidelines. in: consultations in feline internal medicine, august j.r. (editor), elsevier saunders, st. louis, mo. chapter 31. page 279-290, 2006. isbn: 9780721604237 [63] h.p. lefebvre, b. reynold. a cat is not a dog: specific therapeutic considerations https://www.vin.com/apputil/content/defaultadv1.aspx?pid=11242&catid=31922&id=3860757. (accessed 09/18/2021). [64] m.k. tolbert, s. olin, s. maclane, e. gould, j.m. steiner, s. vaden, j. price. evaluation of gastric ph and serum gastrin concentrations in cats with chronic kidney disease. j. vet. intern. med. 31 (2017) 14141419. https://doi.org/10.1111/jvim.14807. [65] n.j. telles, b.t. simon, e.m. scallan, e.n. gould, m.g. papich, y. he, m.t. lee, j.a. lidbury, j.m. steiner, a. kathrani, k.m. tolbert. evaluation of gastrointestinal transit times and ph in healthy cats using a continuous ph monitoring system. journal of feline medicine and surgery december (2021). https://doi.org/10.1177/1098612x211062096. [66] m.l. chandler, g. guilford, c.r. lawoko. radiopaque markers to evaluate gastric emptying and small intestinal transit time in healthy cats. j. vet. intern. med. 11 (1997) 361-364. https://doi.org/10.1111/j.1939-1676.1997.tb00481.x. [67] b. carlborg, o. densert. esophageal lesions caused by orally administered drugs. an experimental study in the cat. eur. surg. res. 12 (1980) 270-282. https://doi.org/10.1159/000128132. [68] b. carlborg, o. densert, c. lindqvist. tetracycline induced esophageal ulcers. a clinical and experimental study. laryngoscope 93 (1983) 184-187. https://doi.org/10.1288/00005537-19830200000011. [69] t.t. kararli. comparison of the gastrointestinal anatomy, physiology, and biochemistry of humans and commonly used laboratory animals. biopharm. drug dispos. 16 (1995) 351-380. https://doi.org/10.1002/bdd.2510160502. [70] understanding the ruminant animal digestive system. http://extension.msstate.edu/publications/publications/understanding-the-ruminant-animal-digestive-system (accessed 09/18/2021). [71] m.n. martinez, j.c. kawalek, k.d. howard, j.l. ward, p. marroum, w. marnane, d. bensley, f.r. pelsor, s. hoag, a.s. tatavarti, l. xie, r. fahmy. comparison of bovine in vivo bioavailability of two sulfamethazine oral boluses exhibiting different in vitro dissolution profiles. j. vet. pharmacol. ther. 29 (2006) 459-467. https://doi.org/10.1111/j.1365-2885.2006.00781.x. http://dx.doi.org/10.5599/admet.1140 https://doi.org/10.2460/ajvr.73.7.1085 https://doi.org/10.1111/jvim.14788 https://doi.org/10.1023/a:1016473601633 https://doi.org/10.2460/ajvr.80.10.957 https://doi.org/10.1002/jps.21602 https://doi.org/10.1016/s0939-6411(99)00070-3 https://doi.org/10.1208/s12249-016-0611-2 https://www.vin.com/apputil/content/defaultadv1.aspx?pid=11242&catid=31922&id=3860757 https://www.vin.com/apputil/content/defaultadv1.aspx?pid=11242&catid=31922&id=3860757 https://doi.org/10.1111/jvim.14807 https://doi.org/10.1177/1098612x211062096 https://doi.org/10.1111/j.1939-1676.1997.tb00481.x https://doi.org/10.1111/j.1939-1676.1997.tb00481.x https://doi.org/10.1159/000128132 https://doi.org/10.1288/00005537-198302000-00011 https://doi.org/10.1288/00005537-198302000-00011 https://doi.org/10.1002/bdd.2510160502 http://extension.msstate.edu/publications/publications/understanding-the-ruminant-animal-digestive-system http://extension.msstate.edu/publications/publications/understanding-the-ruminant-animal-digestive-system https://doi.org/10.1111/j.1365-2885.2006.00781.x martinez, papich and fahmy admet & dmpk 10(1) (2022) 1-25 24 [72] w. grünberg, p.d. constable, chapter 6 function and dysfunction of the ruminant forestomach. in: food animal practice (fifth edition), d.e. anderson, d. m. rings (eds), w.b. saunders, missouri, 2009, pages 12-19, isbn 9781416035916, https://doi.org/10.1016/b978-141603591-6.10006-5 . [73] l.j. henze, n.j. koehl, r. jansen, r. holm, m. vertzoni, p.d. whitfield, b.t. griffin. development and evaluation of a biorelevant medium simulating porcine gastrointestinal fluids. eur. j. pharm. biopharm. 154 (2020) 116-126. https://doi.org:10.1016/j.ejpb.2020.06.009. [74] s.s. davis, l. illum, m. hinchcliffe. gastrointestinal transit of dosage forms in the pig. j. pharm. pharmacol. 53 (2001) 33-39. https://doi.org/10.1211/0022357011775163. [75] j.d. baggot. bioavailability and bioequivalence of veterinary drug dosage forms, with particular reference to horses: an overview. j. vet. pharmacol. ther. 15 (1992) 160-173. https://doi.org/10.1111/j.1365-2885.1992.tb01003.x. [76] t.j. doherty, f.m. andrews, m.k. provenza, d.l. frazier. acetaminophen as a marker of gastric emptying in ponies. equine. vet. j. 30 (1998) 349-351. https://doi.org/10.1111/j.20423306.1998.tb04109.x. [77] k.l. lohmann, a.j. roussel, n.d. cohen, d.m. boothe, p.c. rakestraw, m.a. walker. comparison of nuclear scintigraphy and acetaminophen absorption as a means of studying gastric emptying in horses. am. j. vet. res. 61 (2000) 310-3155. https://doi.org/10.2460/ajvr.2000.61.310. [78] a.m. stokes, n.l. lavie, m.l. keowen, l. gaschen, f.p. gaschen, d. barthel, f.m. andrews. evaluation of a wireless ambulatory capsule (smartpill®) to measure gastrointestinal tract ph, luminal pressure and temperature, and transit time in ponies. equine vet. j. 44 (2012) 482-486. https://doi.org/10.1111/j.2042-3306.2011.00533.x. [79] s.j. baker, e.l. gerring. gastric emptying of four liquid meals in pony foals. res. vet. sci. 56 (1994) 164169. https://doi.org/10.1016/0034-5288(94)90099-x. [80] digestive anatomy and physiology of the horse. https://www.extension.iastate.edu/equine/blog/drpeggy-m-auwerda/digestive-anatomy-and-physiology-horse. (accessed 09/18/2021). [81] f.m. andrews. equine gastric ulcer syndrome. american association of equine practitioners. https://aaep.org/horsehealth/equine-gastric-ulcersyndrome#:~:text=high%20acid%20in%20the%20stomach%20may%20predispose%20foals,from%20 1.5%20to%207.0%2c%20depending%20on%20region%20measured. (accessed 09/18/2021). [82] c. thune. ideal feed frequency for horses. https://thehorse.com/182673/ideal-feed-frequency-forhorses/. accessed 10/07/2021. [83] j.l. davis, d. little, a.t. blikslager, m.g. papich. mucosal permeability of water-soluble drugs in the equine jejunum: a preliminary investigation. j vet pharmacol ther 29 (2006) 379-385. https://doi.org:10.1111/j.1365-2885.2006.00757.x. [84] m. van den berg. anatomy of the horse’s digestive tract, part 2: the small intestine. (dated september 20, 2020). https://horsesandpeople.com.au/anatomy-of-the-digestive-tract-part-2-the-smallintestine/ (accessed 09/18/2021). [85] a. sacranie. the relationship between fee particle size, gizzard development and broiler performance. in: zootechnica internation. https://zootecnicainternational.com/poultry-facts/relationship-feedparticle-size-gizzard-development-broiler-performance/. accessed 09/18/21. [86] b. svihus. the gizzard: function, influence of diet structure and effects on nutrient availability. worlds poult. sci. j. 67 (2011) 207-224. https://doi.org/10.1017/s0043933911000249. [87] b. svihus. function of the digestive system. j. appl. poult. res. 23 (2014) 306–314. https://doi.org/10.3382/japr.2014-00937. [88] r. gauthier (2002). intestinal health, the key to productivity: the case of organic acids. presented at precongreso cientifico avicola iasa, puerto vallarta, jal. mexico, 30 april, 2002. https://www.researchgate.net/publication/313391701_intestinal_health_the_key_to_productivity_t he_case_of_organic_acids#fulltextfilecontent accessed 09/18/2021. https://doi.org/10.1016/b978-141603591-6.10006-5 https://doi.org:10.1016/j.ejpb.2020.06.009 https://doi.org/10.1211/0022357011775163 https://doi.org/10.1111/j.1365-2885.1992.tb01003.x https://doi.org/10.1111/j.1365-2885.1992.tb01003.x https://doi.org/10.1111/j.2042-3306.1998.tb04109.x https://doi.org/10.1111/j.2042-3306.1998.tb04109.x https://doi.org/10.2460/ajvr.2000.61.310 https://doi.org/10.1111/j.2042-3306.2011.00533.x https://doi.org/10.1111/j.2042-3306.2011.00533.x https://doi.org/10.1016/0034-5288(94)90099-x https://www.extension.iastate.edu/equine/blog/dr-peggy-m-auwerda/digestive-anatomy-and-physiology-horse https://www.extension.iastate.edu/equine/blog/dr-peggy-m-auwerda/digestive-anatomy-and-physiology-horse https://aaep.org/horsehealth/equine-gastric-ulcer-syndrome#:~:text=high%20acid%20in%20the%20stomach%20may%20predispose%20foals,from%201.5%20to%207.0%2c%20depending%20on%20region%20measured https://aaep.org/horsehealth/equine-gastric-ulcer-syndrome#:~:text=high%20acid%20in%20the%20stomach%20may%20predispose%20foals,from%201.5%20to%207.0%2c%20depending%20on%20region%20measured https://aaep.org/horsehealth/equine-gastric-ulcer-syndrome#:~:text=high%20acid%20in%20the%20stomach%20may%20predispose%20foals,from%201.5%20to%207.0%2c%20depending%20on%20region%20measured https://thehorse.com/182673/ideal-feed-frequency-for-horses/ https://thehorse.com/182673/ideal-feed-frequency-for-horses/ https://doi.org/10.1111/j.1365-2885.2006.00757.x https://horsesandpeople.com.au/anatomy-of-the-digestive-tract-part-2-the-small-intestine/ https://horsesandpeople.com.au/anatomy-of-the-digestive-tract-part-2-the-small-intestine/ https://zootecnicainternational.com/poultry-facts/relationship-feed-particle-size-gizzard-development-broiler-performance/ https://zootecnicainternational.com/poultry-facts/relationship-feed-particle-size-gizzard-development-broiler-performance/ https://doi.org/10.1017/s0043933911000249 https://doi.org/10.3382/japr.2014-00937 https://doi.org/10.3382/japr.2014-00937 https://www.researchgate.net/publication/313391701_intestinal_health_the_key_to_productivity_the_case_of_organic_acids#fulltextfilecontent https://www.researchgate.net/publication/313391701_intestinal_health_the_key_to_productivity_the_case_of_organic_acids#fulltextfilecontent admet & dmpk 10(1) (2022) 1-25 veterinary gastrointestinal differences influencing solubility doi: http://dx.doi.org/10.5599/admet.1140 25 [89] r. porter. avian digestive system (dated march 2012). https://www.pheasant.com/portals/0/avian%20digestive%20system%20-%20robert%20porter%20pdf.pdf?ver=2019-09-05-135223-670. (accessed 09/18/2021). [90] d.p. elder. effective formulation development strategies for poorly soluble active pharmaceutical ingredients (api. american pharmaceutical review october 2010. https://www.americanpharmaceuticalreview.com/featured-articles/114950-effective-formulation-developmentstrategies-for-poorly-soluble-active-pharmaceutical-ingredients-apis/ [91] e. eckernäs, c. tannergren. physiologically based biopharmaceutics modeling of regional and colonic absorption in dogs. mol. pharm. 18 (2021) 1699-1710. http://doi.org/10.1021/acs.molpharmaceut.0c01201. [92] c.c. cester, p.l. toutain. a comprehensive model for enrofloxacin to ciprofloxacin transformation and disposition in dog. j. pharm. sci. 86 (1997) 1148-1155. https://doi.org/10.1021/js9603461. [93] r.m. fahmy, m.n. martinez. principles of pharmaceutics and veterinary dosage forms. in: veterinary pharmacology and therapeutics, j.e. riviere and m.g. papich, eds., john wiley and sons, inc., hoboken, n.j., (2018) pp 87-110. [94] f. banova, flavoring options for veterinary compounding. https://www.pccarx.com/blog/flavoringoptions-for-veterinary-compounding accessed 12/03/2021. [95] n. zhao, l.l. augsburger. the influence of swelling capacity of superdisintegrants in different ph media on the dissolution of hydrochlorothiazide from directly compressed tablets. aaps pharmscitech. 6 (2005) e120-126. https://doi.org/10.1208/pt060119. [96] c. cui, j. sun, x. wang, z. yu, y. shi. factors contributing to drug release from enteric-coated omeprazole capsules: an in vitro and in vivo pharmacokinetic study and ivivc evaluation in beagle dogs. dose response 18 (2020). https://doi.org/10.1177/1559325820908980. [97] a. gaier, j. price, l. grubb, s. fitzgerald, m.k. tolbert mk. a prospective, randomized, masked, placebocontrolled crossover study for the effect of 10 mg omeprazole capsules on gastric ph in healthy dogs. j. vet. intern. med. 35 (2021) 887-891. https://doi.org/10.1111/jvim.16061. [98] s.a. center, j.f. randolph, k.l. warner, j. mccabe-mcclelland, p. foureman, w.e. hoffmann, h.n. erb. the effects of s-adenosylmethionine on clinical pathology and redox potential in the red blood cell, liver, and bile of clinically normal cats. j. vet. intern. med. 19 (2005) 303-314. https://doi.org/10.1892/0891-6640(2005)19[303:teosoc]2.0.co;2. [99] s.r. byrn, r.r.pfeiffer, j.g. stowell (1999). solid-state chemistry of drugs (2nd ed). ssci, inc., west lafayette, indiana. [100] s. nath. coccidiosis in poultry. coccidia of poultry (ndvsu.org). accessed 12/03/21. [101] m.n. martinez, m.d. apley. drug solubility classification in the bovine. j. vet. pharmacol. ther. 35 suppl 1 (2012) 93-97. https://doi.org/10.1111/j.1365-2885.2012.01369.x. [102] r.i. senderoff, m. mahjour, g.w. radebaugh. characterization of adsorption behaviour by solid dosage form excipients in formulation development. int. j. pharm. 83 (1992) 65–72. https://doi.org:10.1016/0378-5173(82)90009-6. [103] s.m. alsaidan, a.a. alsughayer, a.g. eshra. improved dissolution rate of indomethacin by adsorbents. drug dev. ind. pharm. 24 (1998) 389–394. https://doi.org/10.3109/03639049809085635. [104] iowa state university extension and outrach: digestive anatomy and physiology of the horse. https://www.extension.iastate.edu/equine/blog/dr-peggy-m-auwerda/digestive-anatomy-andphysiology-horse. accessed 12/06/2021. [105] b.w. sykes, c. underwood, r. greer, c.m. mcgowan, p.d. mills. pharmacokinetics and bioequivalence testing of five commercial formulations of omeprazole in the horse. j. vet. pharmacol. ther. 39 (2016) 78-83. https://doi.org/10.1111/jvp.12240. ©2022 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://dx.doi.org/10.5599/admet.1140 https://www.pheasant.com/portals/0/avian%20digestive%20system%20-%20robert%20porter%20pdf.pdf?ver=2019-09-05-135223-670 https://www.pheasant.com/portals/0/avian%20digestive%20system%20-%20robert%20porter%20pdf.pdf?ver=2019-09-05-135223-670 https://www.americanpharmaceuticalreview.com/featured-articles/114950-effective-formulation-development-strategies-for-poorly-soluble-active-pharmaceutical-ingredients-apis/ https://www.americanpharmaceuticalreview.com/featured-articles/114950-effective-formulation-development-strategies-for-poorly-soluble-active-pharmaceutical-ingredients-apis/ https://www.americanpharmaceuticalreview.com/featured-articles/114950-effective-formulation-development-strategies-for-poorly-soluble-active-pharmaceutical-ingredients-apis/ http://doi.org/10.1021/acs.molpharmaceut.0c01201 http://doi.org/10.1021/acs.molpharmaceut.0c01201 https://doi.org/10.1021/js9603461 https://www.pccarx.com/blog/flavoring-options-for-veterinary-compounding https://www.pccarx.com/blog/flavoring-options-for-veterinary-compounding https://doi.org/10.1208/pt060119 https://doi.org/10.1177/1559325820908980 https://doi.org/10.1111/jvim.16061 https://doi.org/10.1892/0891-6640(2005)19%5b303:teosoc%5d2.0.co;2 http://ndvsu.org/images/studymaterials/parasitology/coccidiosis-in-poultry-ppt.pdf https://doi.org/10.1111/j.1365-2885.2012.01369.x https://doi.org/10.1016/0378-5173(82)90009-6 https://doi.org/10.3109/03639049809085635 https://www.extension.iastate.edu/equine/blog/dr-peggy-m-auwerda/digestive-anatomy-and-physiology-horse https://www.extension.iastate.edu/equine/blog/dr-peggy-m-auwerda/digestive-anatomy-and-physiology-horse https://doi.org/10.1111/jvp.12240 http://creativecommons.org/licenses/by/3.0/ manuscript doi: 10.5599/admet.3.3.201 281 admet & dmpk 3(3) (2015) 281-286; doi: 10.5599/admet.3.3.201 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper a trend model for alzheimer’s mortality örjan hallberg* hallberg independent research, brattforsgatan 3, 123 50 farsta, sweden *corresponding author: e-mail: oerjan.hallberg@swipnet.se; tel.: +46-8-605-4998 received: july 07, 2015; revised: august 02, 2015; published: september 05, 2015 abstract in sweden, mortality rates from alzheimer’s disease have increased since early 90’s. in this study, we compared rates reported from 2006-2012 with projected trends determined previously and found a good fit. the objective of this study was to investigate if increased mortality can be modeled as a single exponential function of time lived in a new environment, where the risk of dying from alzheimer’s disease has been increased. the results demonstrated that the exponential model can be used to predict future mortalities for different scenarios, and that it can also project age-specific trends. we conclude that increasing mortality rates from alzheimer’s disease seem caused by an environmental change introduced since the 1990’s. since similar trend breaks also have been reported for different cancers, responsible authorities should seriously address this problem to pinpoint causative factors. keywords alzheimer; mortality; cell phones; model; exponential. introduction in 2009, this author published a model study on alzheimer mortality, where data up till 2006 were used to predict future mortality rates in sweden [1]. in that study, the annual use of cell phones was found to be a good base for the model. after parameters were varied to best fit reported age-standardized mortality rates, calculated age-specific rates also perfectly matched reported data after the year 1992. in this report we wanted to see if reported mortalities since 2006 (up till 2012) have been following the projected trends predicted in the earlier report. the model work from 2009 used the multiple “life matrix” model to cover all birth cohorts of interest to calculate age-specific mortalities. here, we investigated the usefulness of a much simpler exponential model, based on the time different age groups have been living older than a specific age, p, since a specific year yo when a causative environmental change started. we also wanted to project future mortality trends for different future exposure scenarios. experimental data on reported mortality due to alzheimer’s disease (ad) were retrieved from the national board of health and welfare in sweden. the “multiple life matrix” model used for the first report in 2009 [1] was later also useful in trend analysis of skin melanoma, lung cancer, and brain cancer. this time we used a less complicated exponential approach, which previously worked well in modeling melanoma incidence and mortality trends [2]. http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:oerjan.hallberg@swipnet.se örjan hallberg admet & dmpk 3(3) (2015) 281-286 282 the exponential model the model assumes that the mortality can be described by an exponential function of time lived in a new environment, according to equation (1) described below. the initial exponential increase is assumed to be similar for all age groups. thus, it was easy to create an exponential model that applies to all ages using the following key parameters:  a = age group, e.g. 70-74 years old  yo = the year in which a sudden environmental change occurred  y = calendar year  t = y-yo: the number of years since the environmental change took place; y > yo  p = age from when the environmental change has any effect, e.g. 45 years.  am = mortality in year yo and earlier for age group a  a = a constant  b = the slope used in the exponential function the mortality m is thus a function of t and a as given by equation (1), formatted as an excel ifexpression. 𝑀(𝑇, 𝐴) = if(𝑇 < 𝐴 − 𝑃; 𝐴𝑚 + 𝑎(exp(𝑏𝑇) − 1); 𝐴𝑚 + 𝑎(exp(𝑏(𝐴 − 𝑃)) − 1) (1) we used equation (1) to calculate the age-standardized rates (asrw) of ad mortality. in searching for best fit between calculated and reported rates, we varied the constant (a), the exponential slope factor (b), and the starting age (p) to determine the best fit. to test the validity of this approach, we then compared calculated age-specific incidences with reported data without any further parameter variations. to estimate the effect of completely stopping exposure from the radiation sources involved, we used the following calculation: if the exposure is stopped in year y1, then the mortality at time y=y1+t for age group a can be estimated according to equation (2): 𝑀(𝑇, 𝐴) = (𝐴 − 𝑡)𝑚 + 𝑎(exp(𝑏(𝑇 − 𝑡)) − 1) (2) where (a-t)m is the initial mortality before e.g. 1992 for age group a-t. results and discussion it turned out that reported age-specific mortalities from ad continued to follow the projected trends after 2006 quite well. figure 1 shows the original graph (figure 8 from ref [1]) as well as the reported mortality rates up until 2012. the exponential model regular registration of ad deaths became routine in sweden after 1987 (icd 9). based on the model analysis in 2009, [1], the actual mortality before 1987 was likely higher than reported at that time (figure 1). based on those results, we used reported rates for 1992 as am in equation (1) for each age group for the years before 1992. in 1987 cell phones operating mainly at 450 mhz had been used for a few years. the next generation working at 900 mhz became popular around 1992; while, the real boom occurred in late 1997 when the dual band phones, 900-1800 mhz were introduced. to design a simple exponential model based on the time lived as an alzheimer victim in a wireless world, we assigned 1992 as the starting year for population exposure to cell phone ghz radiation. equation (1), described in experimental section, was then used together with reported age-standardized mortality data from 1992 till 2012 to determine model parameters for the mortality trends. the reported and calculated mortality from ad if the whole cell phone system closed down in 2022 (figures 2 and 3) is an example of a scenario calculated according to equation (2). admet & dmpk 3(3) (2015) 281-286 alzheimer mortality model doi: 10.5599/admet.3.3.201 283 figure 1. reported and calculated age-specific mortality rates from ad in sweden. data up till 2006 are from ref [1], including the projections. reported rates up till 2012 have here been added to the graph using filled symbols. calculated data are marked c, reported up till 2006 by rep and reported after 2006 by rb. the age-standardized mortalities (world standard) are calculated based on equation (1) when all age groups are taken into account. thus, it was easy to compare reported and calculated age-specific rates, since they had already been used to calculate age-standardized mortalities. the age-specific rates reported, as well as those calculated using the same parameters used in calculating age standardized mortalities (figure 2, table 1), are shown in figure 3. figure 2. reported and calculated age-standardized mortality from ad. the black line indicates the calculated total mortality if all mobile phone systems were no longer used after 2022. örjan hallberg admet & dmpk 3(3) (2015) 281-286 284 table 1. optimum parameters for the exponential function used in the trend models shown in figures 2 and 3 constant a a*exp(bt) exponent b a*exp(bt) start age p (years) 25.17 0.1494 65.3 figure 3. reported (symbols) and calculated (lines) ad age-specific mortalities in sweden if all cell phone systems were closed down in year 2022. discussion the updated mortality data added to figure 1 show clearly that the original model projections [1] still fit well with reported rates. thus, we can state that the multi life matrix model seems to work well, and could be used to analyze data from other countries in a similar way. as depicted in figure 2, increasing ad mortality rates correspond with population exposure to cell phones and base station generated radiation from 1992 onwards. figure 2 also shows how reported mortalities increased fast from 1987 to around 1990, most likely as a result of new classification standards (icd 9) introduced in 1987. thus, mortalities reported in 1992 were used as base level for earlier years, resulting in the flat calculated level in figure 2 for years before 1992. the model also predicts that we will see a leveling-off of the age-standardized mortality quite soon since all age groups from 65.3 years just about now have been living in the new environment we got after 1992. this simple exponential model also projected age-specific rates that did fit well with reported data without any further parameter adjustments, as shown in figure 3. parameters giving best fit between calculated and reported age-standardized mortality are shown in table 1. equation (1) assumes that the mortality in a specific age group a will stabilize as soon as all people in that age group have been living their lives since age p in the new wireless world. reported data show that the age group of 72 years has a stable mortality from the 1990’s, while the age group 77 years flattens out a few years after year 2000. the age groups 82 and 85+ still have some years to wait for their stabilizations to occur. figures 1 and 3 clearly show that the more complex model from 2009 still works [1] and gives the most accurate projections of ageadmet & dmpk 3(3) (2015) 281-286 alzheimer mortality model doi: 10.5599/admet.3.3.201 285 specific mortalities. however, the much simpler exponential model seems as useful for age-standardized projections and to estimate the response to an improved environment. research on biological changes in the brain, which accelerate the mortality risk among ad patients, should be performed. possible connections between continuous base station radiation, cell phone use, and biological changes should also be investigated. in 2014, v. lauer published a theory and model about the interaction of t lymphocytes with electromagnetic fields [3]. his work further contributes to the understanding of biological effects from non-ionizing radiation. data on ad mortality from the usa and from the 21 different swedish counties in the time period 19992004 show that the mortality was higher in states or counties where the population density was lower [4,5]. data from 2010 in the usa had an even stronger linear association (p = 0.011) to population density than in 1999-2004 (p = 0.038), [6]. this suggests that the higher average output power from cell phone handsets in sparsely populated areas also might be an important factor to consider (figure 4). the fact, that mortality among dementia patients suddenly started to increase right after 1997, was noticed in a report from the swedish national health and welfare authority. no explanation was given except that doctors seemed to have become more inclined to classify deaths as due to dementia since 1987 [7]. figure 4. mean alzheimer mortality for men and women in sweden (se) and the usa vs. population density in counties (sweden) or states (usa) during 1997-2004. conclusions this short report shows that ad mortality can be modeled by different techniques with good fit to reported data. since both models used here clearly indicate that an environmental change from around 1992 has been an important factor, responsible authorities should seriously start addressing this problem. acknowledgements: this study was financed by internal funds within hallberg independent research. örjan hallberg admet & dmpk 3(3) (2015) 281-286 286 references [1] ö. hallberg, curr. alzheimer res. 6 (2009) 471-475. [2] ö. hallberg, int. j stats. med. res. 4 (2015) 65-71. [3] v. lauer, hyper articles en ligne, hal : hal-00975963, version 1; (2014), http://hal.archivesouvertes.fr/hal-00975963. [4] ö. hallberg, o. johansson, eur. biol. bioelectromag. 1 (2005) 225-246, http://sgll.nu/medpub/ref%2013%20ebb3b.pdf. [5] k. steenland, j. macneal, i. vega, a. levey, alzheimer dis. assoc. disord. 23 (2009) 165-170 http://www.ncbi.nlm.nih.gov/pubmed/?term=19484918. [6] b. tejada-vera, mortality from alzheimer’s disease in the united states. data for 2000 and 2010. nchs data brief, no 116. hyattsville, md: national center for health statistics, (2013), http://www.ncbi.nlm.nih.gov/pubmed/23742787. [7] official statistics of sweden, statistics – health and medical care, causes of death 2012. isbn 978-91-7555-090-9, article nr 2013-8-6. (figure 22), http://www.socialstyrelsen.se/lists/artikelkatalog/attachments/19175/2013-8-6.pdf. ©2015 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://hal.archives-ouvertes.fr/hal-00975963 http://hal.archives-ouvertes.fr/hal-00975963 http://sgll.nu/medpub/ref%2013%20ebb3b.pdf http://www.ncbi.nlm.nih.gov/pubmed/?term=19484918 http://www.ncbi.nlm.nih.gov/pubmed/23742787 http://www.socialstyrelsen.se/lists/artikelkatalog/attachments/19175/2013-8-6.pdf http://creativecommons.org/licenses/by/3.0/ doi: 10.5599/admet.4.2.251 91 admet & dmpk 4(2) (2016) 91-97; doi: 10.5599/admet.4.2.251 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper calendula officinalis l. (asteraceae) possess antioxidant properties on fe 2+ -initiated peroxidation of rat brain microsomes alejandro palacios, javier barberón, patricio leaden and pedro zeinsteger * cátedra bioquímica, facultad de ciencias veterinarias, universidad nacional de la plata, cc296 (1900) la plata, buenos aires, argentina *corresponding author: e-mail: pzeins@fcv.unlp.edu.ar; tel.: +54-221-592-4308 received: november 10, 2015; revised: april 15, 2016; published: june 29, 2016 abstract in this study the effects of calendula officinalis l. (asteraceae) extract (co) on the polyunsaturated fatty acid composition, chemiluminescence and unsaturation index of microsomes isolated from brain rat, are presented. after incubation of microsomes in an ascorbate (0.4 mm)-fe 2+ (2.15 µm) system (180 min at 37 °c) it was observed that the total cpm/mg protein originated from light emission:chemiluminescence was lower in brain microsomes obtained from co group compared to the control group (without extract supplementation). moreover, it was observed that the addition of the extract reduced chemiluminescence -measured as total cpmin a concentration dependent manner. the fatty acid composition of brain microsomes from control group was profoundly modified when subjected to non-enzymatic lipoperoxidation with a considerable decrease of arachidonic acid c20:46 and docosahexaenoic acid c22:63. as a consequence, the unsaturation index, a parameter based on the maximal rate of oxidation of specific fatty acids, was higher in the co group compared to controls. the simultaneous analysis of unsaturation index, chemiluminescence and fatty acid composition indicate that co may act as an antioxidant protecting rat brain microsomes from peroxidative damage. keywords marigold; lipoperoxidation; chemiluminescence introduction the polyunsaturated acids located in biological membranes are excellent targets for peroxidation with peroxides formation [1,2]. the consequence of peroxidation of unsaturated fatty acids membranes is severe, damage of membrane function, enzymatic inactivation, toxic effects on the cellular division, etc. [35]. such alterations have been implicated in causation of several diseases such as liver cirrhosis [6], inflammation and atherosclerosis [7]. duthie et al. have stated that microsomes may be considered as an interesting system for lipoperoxidation studies [8]. the microsome is a convenient experimental model for detailed studies of kinetic reaction and peroxidation mechanism, considering the injured microsome a reason of tissue alterations in many pathological processes [9]. the production of reactive oxygen species in the brain has been implicated as a common factor in the etiology of a number of neurodegenerative diseases [10]. the link between free radicals and disease processes has led to considerable research into http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:pzeins@fcv.unlp.edu.ar palacios et al. admet & dmpk 4(2) (2016) 91-97 92 nontoxic drugs that can scavenge the free radicals. several plant extracts and plant products have been shown to possess significant antioxidant potential [11,12]. calendula officinalis l. (asteraceae, “marigold”) extract (co) is rich in flavonoids, terpenoids and lutein and has both antioxidant and anti-inflammatory activities and has been linked to the reduction in the risk of chronic diseases such as macular degeneration, cancer, cardiovascular and neurodegenerative disorders [13]. present study was designed to determine whether microsomes from rat brain could be a target for non-enzymatic lipoperoxidation as well as to establish the level of protection of such membranes incubated with an extract of calendula officinalis. the degradative process was followed simultaneously by the determination of chemiluminescence and fatty acid composition of brain microsomes. the unsaturation index was used to evaluate the fatty acid alterations observed during the process [14]. experimental female wistar ah/hok rats were obtained from laboratory animal facility, facultad de ciencias veterinarias, universidad nacional de la plata. bsa (fraction v) was obtained from wako pure chemical industries ltd, japan. standards of fatty acids methyl esters were kindly supplied by nu check prep. inc, elysian, mn, usa. l(+) ascorbic acid and boron trifluoride-methanol complex were from merck laboratories. all other reagents and chemicals were of analytical grade from sigma-aldrich. co extract preparation fifteen grams of co dried flowers were milled until plant material passed through a 2 mm screen and put in a flask together with 50 ml methanol for extraction under mechanical agitation at 870 rpm and dim light during 12 h. after extraction co extract was submitted to vacuum filtration and then concentrated using a rotary evaporator (senco ltd, zhong shan nan yi rd., shanghai, china.) until total evaporation of the solvent. a residue of 3.15 g was obtained, which was re-suspended in 20 ml methanol. from the total methanolic extract 10 ml were used for the qualitative analysis of phytochemical constituents while the remaining volume was used to test the anti-oxidative properties. for the phytochemical analysis of the extract the original volume of 10 ml was fractioned in three aliquots using a 1-5 ml adjustable volumepipette: fraction a for testing the presence of flavonoids (zinc hydrochloride reduction test, shinoda´s test), tannins (ferric chloride test), lipids (iodine reaction) and hydrocarbons (phenol 5 % + concentrated h2so4); fraction b for the investigation of the presence of steroids (acetic anhydride + concentrated h2so4, liebermann-burchard reaction) and anthraquinones (sodium hydroxide test, bornträger´s test); and fraction c for the determination of alkaloids (potassium iodide-bismuth nitrated test, dragendorff´s reagent), cardenolides (dinitrobenzoic acid + sodium hydroxide, kedde´s reagent), steroids (liebermannburchard reaction) and leucoanthocyanins (concentrated hcl + amyl alcohol, rosenheim reaction) [15,16]. animals and membrane preparation female wistar ah/hok rats, 7 weeks old, weighing 120-137 g were used. all rats were fed commercial rat chow and water ad libitum. the rats were euthanized by cervical dislocation and the brain was rapidly removed, cut into small pieces and extensively washed with 0.15 m nacl. a 30 % (w/v) homogenate was prepared in a 0.25 m sucrose solution, 10 mm tris-hcl ph 7.4 using a potter-elvejhem homogenizer. the homogenate was spun at 10,000 x g for 10 min. the supernatant (3 ml) was applied to a sepharose 4b column (1.6 x 12 cm) equilibrated and eluted with 10 mm tris-hcl ph 7.4, 0.01 % nan3. the microsomal fraction appearing in the void volume (10-16 ml) was brought to 0.25 m sucrose by adding solid sucrose. all operations were performed at 4 °c and under dim light. the quality of microsomal preparation is similar in composition as regards concentrations and activities of certain microsomal enzymes to that obtained by admet & dmpk 4(2) (2016) 91-97 calendula officinalis antioxidant properties doi: 10.5599/admet.2.4.251 93 ultracentrifugation [17]. lipid peroxidation of rat brain microsomes chemiluminescence and lipid peroxidation were initiated by adding ascorbate to microsomes [18]. the microsomes (0.5 mg microsomes protein) with addition of co total extract (0.1, 0.2, 0.3 and 0.4 mg) were incubated at 37 °c with 0.01 m phosphate buffer ph 7.4, 0.4 mm ascorbate, final vol. 2 ml. phosphate buffer is contaminated with enough iron to provide the necessary ferrous or ferric iron (final concentration in the incubation mixture was 2.15 µm) for lipid peroxidation [19]. microsome preparations, which lacked ascorbate, were carried out simultaneously. membrane light emission was determined over a 180 min period, chemiluminescence was recorded as count per minute (cpm) every 10 min and the sum of the total chemiluminescence was used to calculate cpm/mg protein. chemiluminescence was measured as counts per min in liquid scintillation analyzer packard 1900 tr equipment with a program for chemiluminescence. fatty acid analysis microsomal lipids were extracted with chloroform/methanol (2:1 v/v) [20] from native or peroxidized membranes. fatty acids were transmethylated with f3b in methanol at 60 °c for 3 h. fatty acid methyl esters were analyzed with a gc-14a gas chromatograph (shimadzu, kyoto, japan) equipped with packed column (1.80 m x 4 mm i.d.) gp 10 % degs-ps on 80/100 supelcoport. nitrogen was used as the carrier gas. the injector and detector temperatures were maintained at 250 °c, at the column temperature was held at 200 °c. fatty acid methyl ester peaks were identified by comparison of the retention times with those of standards. protein determination proteins were determined by the method of lowry et al. [21] using bsa as standard. unsaturation index (ui) ui was calculated as the sum of the percentage of each fatty acid x the number of olefinic bonds, divided by the sum of the percentage of saturated fatty acids. statistical analysis results were expressed as mean ±s.d. of six independent determinations. data were statistically evaluated by one-way analysis of variance (anova) and tukey’s test. the statistical criterion for significance was selected at different p-values, which was indicated in each case. results and discussion phytochemical analysis of co extract table 1 shows the results of the phytochemical analysis of the co extract. qualitative chemical determinations were performed in order to assess the presence of those compounds that exert antioxidative properties, as mentioned in co extract preparation. light emission of rat brain microsomes during lipid peroxidation the incubation of rat brain microsomes in the presence of ascorbate-fe 2+ resulted in the peroxidation of membranes as evidenced by emission of light (chemiluminescence). after incubation of microsomes in an ascorbate-fe 2+ system at 37 °c for 180 minutes, the cpm originated from light emission was lower in the co group than in the control group in a concentration dependent manner (fig. 1) figure 2 shows the total light palacios et al. admet & dmpk 4(2) (2016) 91-97 94 emission obtained from co group and control group. the values were from 2,476.853 ± 312.987 cpm in the control group to 1,462.044 ± 142.044 cpm with the addition of 0.4 mg extract/mg protein. table 1. qualitative analysis of co extract. determination fraction a fraction b fraction c shinoda + ferric chloride + iodine + phenol 5% + conc. h2so4 + liebermann-burchard + + bornträger dragendorff kedde rosenheim + figure 1. lipid peroxidation of rat brain microsomes with different concentrations of co extract. figure 2. total chemiluminescence of rat brain microsomes induced by ascorbate fe 2+ system. fatty acid composition of brain microsomes and unsaturation index table 2 shows the fatty acid composition of brain microsomes from control group, native and peroxidized in ascorbate fe 2+ system. the lipid peroxidation of brain microsomes showed significant changes in the polyunsaturated fatty acid composition, significant decreases of arachidonic acid c20:46 and docosahexaenoic acid c22:63 were observed. there were marked differences when the unsaturation admet & dmpk 4(2) (2016) 91-97 calendula officinalis antioxidant properties doi: 10.5599/admet.2.4.251 95 index of peroxidized control group and native control group brain microsomes were compared. table 2. fatty acid composition (area %) of rat brain native and peroxidized microsomes without co extract. a a data are given as the mean ±sd of six experiments. statistically significant differences in fatty acids concentrations and ui of brain microsomes native and peroxidized are indicated by 1,4 p ˂ 0.01, 2 p ˂ 0.004, 3 p ˂ 0.03 the fatty acid composition from isolated brain microsomes after addition of co extract to the ascorbate fe 2+ system is presented in table 3. when peroxidized control group (table 2) is compared to co group after lipid peroxidation it is observed that arachidonic acid c20:46 and docosahexaenoic acid c22:63 increased in the co group (concentration dependent). polyunsaturated fatty acids ranged from 29.32 % in native microsomes, 8.3 % in peroxidized and 24.58 % in peroxidized + 0.4 mg extrac /mg protein. significant changes in the unsaturation index were observed when compared peroxidized control group against co group. the changes were less pronounced in membranes from co group. as a result, the unsaturation index of peroxidized membranes in the co group was significantly higher than in the control group. table 3. fatty acid composition of rat brain microsomes with different concentrations of co extract a fatty acid co extract concentration 0.1 mg 0.2 mg 0.3 mg 0.4 mg c16:0 25.285 ± 2.072 26.406 ± 2.325 28.333 ± 2.734 23.109 ± 3.109 c18:0 23.067 ± 1.439 24.579 ± 1.973 22.947 ± 1.995 20.926 ± 2.505 c18:1ω9 34.601 ± 2.231 36.869 ± 2.163 34.429 ± 3.029 31.389 ± 2.942 c18:2ω6 3.019 ± 0.178 2.514 ± 0.710 0.645 ± 0.225 4.343 ± 0.982 fatty acid brain microsomes native brain microsomes peroxidized c16:0 21.941 ± 1.018 28.006 ± 1.804 c18:0 16.496 ± 1.192 19.579 ± 1.032 c18:1ω9 30.991 ± 2.589 37.191 ± 2.879 c18:2ω6 0.717 ± 0.040 0.337 ± 0.053 c18:3ω3 2.404± 1.062 3.270 ± 1.527 c20:4ω6 10.370 ± 0.526 1 3.457 ± 0.225 1,b,c c22:6ω3 13.194 ± 0.487 2 0.173 ± 0.040 2, d, e saturated 38.346 ± 1.730 47.586 ± 2.094 monounsaturated 30.991 ± 2.589 37.171 ± 2.201 polyunsaturated 29.317 ± 1.991 4 8.297 ± 1.003 4 total unsat./sat. 60.309 ± 1.114 45.468 ± 2.205 saturated/unsaturate d 0.637 ± 0.039 1.115 ± 0.531 unsaturation index 173.445 ± 7.002 3 67.657 ± 2.391 3 palacios et al. admet & dmpk 4(2) (2016) 91-97 96 c18:3ω3 8.715 ± 1.322 6.447 ± 1.721 4.438 ± 1.045 7.450 ± 1.227 c20:4ω6 5.313 ± 0.621 3.184 ± 0.109 5.685 ±0.538 b 7.682 ± 0.301 c c22:6ω3 0.324 ± 0.057 1 0.435 ± 0.093 2 3.530 ± 0.611 d 5.102 ± 0.471 1,2, e saturated 48.352 ± 2.425 50.985 ± 3.094 51.280 ± 3.723 44.035 ± 3.429 monounsaturated 34.601 ± 1.327 36.689 ± 2.629 34.429 ± 2.408 31.389 ± 2.929 polyunsaturated 17.047 ± 1.005 12.145 ± 1.005 14.298 ± 1.300 24.577 ± 1.853 total unsat./sat. 51.640 ± 3.109 49.014 ± 2.116 48.727 ± 3.513 55.966 ± 4.003 saturated/unsaturated 0.936 ± 0.127 1.040 ± 0.057 1.052 ± 0.281 0.786 ± 0.217 unsaturation index 88.036±4.076 3 73.974 ±3.226 4 92.957 ± 4.007 123.765±4.216 3,4 a data are given as the mean ±sd of six experiments. statistically significant differences in fatty acids concentrations and ui of brain microsomes native and peroxidized are indicated by 1,2 p ˂ 0.008, 3,4 p ˂ 0.05. statistically significant differences between tables 1 and 2 are indicated by b, c p ˂ 0.05, d p ˂ 0.009, e p ˂ 0.006 conclusions rat brain microsomes incubated with co extract were protected against lipid peroxidation when compared to similar membranes from control group, as demonstrated by the results from chemiluminescence, polyunsaturated fatty acids composition and unsaturation index. as a result of the analysis of ui of peroxidized microsomes membranes, it was observed that co group was successful against lipid peroxidation compared to control group. these results are in concordance with previous reports of braga et al. [22]. in vitro lipid peroxidation studies are useful for the elucidation of possible mechanisms of peroxide formation in vivo [23], since the high concentration of polyunsaturated fatty acid membranes causes susceptibility to lipid peroxidative degradation [24]. in our assay it was found that arachidonic acid c20:46 and docosahexaenoic acid c22:63 from controls were more peroxidized than those from co group. this predisposition to peroxidation of arachidonic acid c20:46 and docosahexaenoic acid c22:63 was demonstrated in previous works [25]. although considerable research has already been performed to characterize the changes in structure, composition and physical properties of membranes subjected to oxidation [26,27], it is important to know how biological compounds with antioxidant properties contribute to the protection of specialized membranes against deleterious effects produced by reactive oxygen species and other free radicals. evidence leading to the recognition of the anticarcinogenic activity of calendula officinalis l. (asteraceae) has been reviewed [28]. hamzawy et al. [29] also reported the anticancer effects of c. officinalis extract due to chemoprotector properties as well as radical scavenging activity because of the presence of phenolic compounds. new data indicates that co has a potent antioxidant activity [29], being marigold extracts able to improve biochemical parameters and decrease the formation of inflammatory cytokines, thus preventing the oxidative stress. the phytochemical constituents present in our co extract were similar to those found in previous works, which also presented the chemical identity of such compounds [30,31]. in conclusion, our results are consistent with the hypothesis that calendula officinalis l. (asteraceae) extract may act as a physiological antioxidant in cell membranes. acknowledgements: to secretaría de ciencia y técnica universidad nacional de la plata, for the financial support. references [1] s. miyamoto, g.r. martinez, d. rettori, proc. natl. acad. sci. 103 (2006) 293-298. [2] s. miyamoto, g.e. ronsein, m. prado, iubmb life. 59 (2007) 322–331. [3] n.a. porter, j. org. chem. 78 (2013) 3511–3524. [4] l.k. lee, s. shahar, n. rajab, j. nutr. biochem. 24 (2013) 803-808. https://scholar.google.com.ar/citations?user=kgaaepaaaaaj&hl=es&oi=sra javascript:void(0); admet & dmpk 4(2) (2016) 91-97 calendula officinalis antioxidant properties doi: 10.5599/admet.2.4.251 97 [5] e. niki, free radic. biol. med. 47 (2009) 469-484. [6] s.h. ganji, b.s. kashyap, b.s. kamanna, metabolism 64 (2015) 982-990. [7] d.t. antoniak, m.j. duryee, t.r. mikuls, g.m. thiele, d.r. anderson, free radic. biol. med. 89 (2015) 409-418. [8] g. duthie, p. morrice, oxid. med. cell longev. 165127 (2012). doi: 10.1155/2012/165127. [9] s. asha, m. vidyavathi, appl. biochem. biotechnol. 160 (2010) 1699-1722. [10] t.h. sanderson, c.a. reynolds, r. kumar, mol. neurobiol. 47 (2013) 9-23. [11] m.m. cascaes, g.m. guilhon, e.h. andrade, m.g. zoghbi, l. santos, int. j. mol. sci. 16 (2015) 881-904. [12] y.j. zhang, r.y. gan, s. li, y. zhou, a.n. li, d.p. xu, h.b. li, molecules 20 (2015) 1138-1156. [13] e. guler, f.b. barlas, m. yavuz, colloids surf. b. biointerfaces.. 121 (2014) 299-306. [14] s. ricci, l. patoia, m. berrettini, stroke 18 (1987) 575-578. [15] j. bruneton, farmacognosia, acribia, zaragoza, spain, 2001, p. 1099. [16] j.b. harborne, phytochemcial methods. a guide to modern techniques and plant analysis, chapman and hall, london, england, 1998, p. 302. [17] l.a. witing, m.k. hornitt. j nutr. 82 (1964) 19-25. [18] j.r. wright, r.c. rumbaugh, h.d. colby, arch. biochem. biophys. 192 (1979) 344-351. [19] b. tadolini, g. hakim, free radic. res. 25 (1996) 221-227. [20] j. folch, m. lees, g.h. sloane stanley, j. biol. chem. 226 (1957) 497-509. [21] o.h. lowry, n.j. rosebrough, n.j. farr, j. biol. chem. 193 (1951) 265-275. [22] p.c. braga, m. dal sasso, m. culici, pharmacology. 83 (2009) 348-355. [23] c. mancuso, e. barone, p. guido, f. miceli, f. di domenico, m. perluigi, r. santangelo, p. preziosi, neurosci. lett. 518 (2012) 101-105. [24] m. shichiri, j. clin. biochem. nutr. 54 (2014) 151-160. [25] p.l. else, e. kraffe, biochim. biophys. acta. 1848 (2015) 417-421. [26] a. ayala, m.f. muñoz, s. argüelles, oxid. med. cell longev. 360438 (2014). doi: 10.1155/2014/360438. [27] t.i. omotayo, g.s. akinyemi, p.a. omololu, redox biol. 4 (2014) 234-241. [28] n. babaee, d. moslemi, m. khalilpour, daru j. pharm.sci. 21 (2013) 18-21. [29] m.a. hamzawy, e.s el-denshary, n.s. hassan, isrn nutr. 538427 (2013). doi: 10.5402/2013/538427. [30] d. arora, a. rani, a. sharma, pharmacogn rev. 7 (2013) 179-187. [31] b.p. muley, s.s khadabadi, n.b. banarase, trop. j. pharm. r. 8 (2009) 455-465. ©2016 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://www.sciencedirect.com/science/article/pii/s089158490900330x http://www.sciencedirect.com/science/journal/08915849/47/5 http://www.ncbi.nlm.nih.gov/pubmed/19582595 http://www.ncbi.nlm.nih.gov/pubmed/?term=folch%20j%5bauthor%5d&cauthor=true&cauthor_uid=13428781 http://www.ncbi.nlm.nih.gov/pubmed/?term=lees%20m%5bauthor%5d&cauthor=true&cauthor_uid=13428781 http://www.ncbi.nlm.nih.gov/pubmed/?term=sloane%20stanley%20gh%5bauthor%5d&cauthor=true&cauthor_uid=13428781 http://www.ncbi.nlm.nih.gov/pubmed/13428781 http://www.ncbi.nlm.nih.gov/pubmed/?term=lowry%20oh%5bauthor%5d&cauthor=true&cauthor_uid=14907713 http://www.ncbi.nlm.nih.gov/pubmed/?term=rosebrough%20nj%5bauthor%5d&cauthor=true&cauthor_uid=14907713 http://www.ncbi.nlm.nih.gov/pubmed/?term=dal%20sasso%20m%5bauthor%5d&cauthor=true&cauthor_uid=19420983 http://www.ncbi.nlm.nih.gov/pubmed/?term=culici%20m%5bauthor%5d&cauthor=true&cauthor_uid=19420983 http://www.ncbi.nlm.nih.gov/pubmed/19420983 http://www.hindawi.com/73827153/ http://www.hindawi.com/63287372/ http://www.hindawi.com/30128409/ http://creativecommons.org/licenses/by/3.0/ manuscript doi: 10.5599/admet.2.1.26 56 admet & dmpk 2(1) (2014) 56-62; doi: 10.5599/admet.2.1.26 open access: issn: 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper rapid screening of mupirocin skin permeation modification by micronized and nanonized alaptide radka opatrilova and josef jampilek* department of chemical drugs, faculty of pharmacy, university of veterinary and pharmaceutical sciences, palackeho 1/3, 61242 brno, czech republic *corresponding author: e-mail: josef.jampilek@gmail.com; tel.: +420-541-562-926 received: december 18,2013; revised: january 28, 2014; published: april 01, 2014 abstract the investigation deals with the affection of permeation of mupirocin from the bactroban® leciva ointment through full-thickness pig ear skin by alaptide that was applied in micronized or nanonized form as a potential excipient. alaptide, (s)-8-methyl-6,9-diazaspiro[4.5]decan-7,10-dione, was the original czech compound. it was discovered that micronized alaptide significantly enhanced the permeation of mupirocin within 1 h after administration (approx. 5-fold). on the other hand, nanonized alaptide almost completely inhibited permeation of mupirocin from bactroban® leciva through the skin. rapid primary screening showed that the two forms of alaptide differently influenced the depth and the rate of permeation/penetration of mupirocin into/through the skin, i.e., affect curative effect of mupirocin on/in skin immediately after application of drug formulation. keywords: alaptide; mupirocin; nanoparticles; bactroban® introduction development in the field of pharmaceutical dosage forms results in the discovery of additional highly sophisticated drug delivery systems that allow for the maintaining of constant levels of active substances in organisms. transdermal administration of drugs represents an alternative to conventional pharmaceutical dosage forms. however, transdermal drug delivery often faces the problem of insufficient or no penetration of active pharmaceutical substances through the skin, because skin itself is a physiological barrier that protects the body from toxins and microorganisms from the external environment. to solve this critical issue, various approaches to overcoming the skin barrier have been developed. these approaches can be classified as chemical (modification of drugs using transdermal chemical penetration enhancers) or physical (modification of drug particle size to nanosize) enhancement techniques [1−5 and refs. therein]. in general, transport mechanisms support nanomaterial penetration [6−9]; however, the promise of nanoparticle-mediated drug delivery into the epidermis and dermis without barrier modification has met with little success. the penetration of particulate materials into the skin is a very complex issue. therefore, the interaction of nanoparticles with the skin and especially skin models is an intriguing field of research. however, the data obtained do not show a clear image of the effect of nano-carriers. in particular, the http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:josef.jampilek@gmail.com admet & dmpk 2(1) (2014) 56-62 mupirocin permeation doi: 10.5599/admet.2.1.26 57 penetration of such particles is an open and controversially discussed issue. the literature reports different results primarily on pig or murine skin, showing either strong penetration (pig and mouse) or the opposite. looking only at the sizes of the particles, no conclusive picture can be obtained. the fact that nanoparticles did not penetrate/permeate or expressed decreased penetration/permeation was described and discussed in a number of recent papers [10−15]. our team focused on investigating a slightly different topic, namely the influence of a transdermal chemical penetration enhancer (cpe) in micronized and/or nanonized form on the modification of the permeation/penetration properties of a drug through the skin. in this investigation alaptide was chosen as a potential cpe and mupirocin was selected as a tested drug. alaptide (chemically, (s)-8-methyl-6,9diazaspiro[4.5]decan-7,10-dione, see fig. 1) is an original czech compound that was prepared by kasafirek et al. in the 1980s [16,17]. during biological assays, it was found that alaptide affected the behaviour and learning abilities of rodents, particularly rats and mice; as such, it can be classified as nootropic [18−20]. alaptide also expressed anti-proliferative activity [21] and immunomodulatory effect [22]. additionally, alaptide reduced the number and the extent of experimental gastric ulcers. the gastric juice quantity, acidity and enzymatic activity were not influenced [23,24]. alaptide was tested on the diploid cell line of a human embryonic lung lep-19, where it expressed a stimulating effect on the growth and breeding of cells without transformation changes in their morphology. in vitro tests also showed that alaptide increased cell proliferation; both the above mentioned effects can significantly contribute to the curative effect of alaptide. an influence of alaptide on epidermal regeneration was investigated in a number of tests. in vivo experiments were performed using domestic pigs to which alaptide was applied on experimental injury; faster skin regeneration was observed after alaptide application. similarly, alaptide accelerated curing of experimental skin injuries in rats [17,25]. as alaptide was designed as an analogue of melanocytestimulating hormone release-inhibiting factor (mif-1), it negatively affects the inhibition of the release of melanocyte-stimulating hormone and thus increases the concentration of melanocytes in the epidermis. melanocytes significantly influence the creation and function of keratinocytes by means of melanosomes [26−28]. metabolic studies in rats showed that alaptide is readily absorbed from the gastrointestinal tract and penetrates the blood-brain barrier [17,21]. it is excreted unchanged, mostly via urine (90%); a similar metabolic profile was also found in humans [29]. any toxicity (acute, subchronic and chronic) as well as genotoxic, teratogenic and embryotoxic effects of alaptide were not observed [17,21,30]. additionally, it was found that both alaptide enantiomers did not induce the biotransformation enzymes (1a1, 1a2, 1b1) of the cytochrome p450. these enzymes are critical in the bioactivation of procarcinogens and are upregulated, inter alia, by ultraviolet-b radiation in the skin [31,32]. based on the above mentioned facts and structural analogy with other cpes, it was suggested to evaluate alaptide as a potential cpe of many anti-inflammatory drugs, antimicrobial chemotherapeutics, sex hormones/genital system modulators, or drugs targeting the central/vegetative nervous system [4,32,33]. furthermore, the permeation of micronized (ala) and nanonized (nala) alaptide alone through the full-thickness pig ear skin from various semi-solid compositions was evaluated and different behaviour among both forms was confirmed. though permeation of nala through the skin was limited compared with ala [34], nala was also positively evaluated as a potential modifier of transdermal permeation [32,33]. thus, the aim of this primary rapid study was to evaluate the effect of alaptide in micronized and nanonized forms on drug permeation through the skin immediately after application of the drug formulation. opatrilova and jampilek admet & dmpk 2(1) (2014) 56-62 58 mupirocin (pseudomonic acid a, chemically 9-{[(2e)-4-[(2s,3r,4r,5s)-3,4-dihydroxy-5-{[(2s,3s)-3[(2s,3s)-3-hydroxybutan-2-yl]oxiran-2-yl]methyl}oxan-2-yl]-3-methylbut-2-enoyl]oxy}nonanoic acid, see fig. 1) is an antibiotic originally isolated from pseudomonas fluorescens and is structurally unrelated to any other antibiotics. it is used topically and primarily effective against gram-positive bacteria, such as staphylococcus aureus (including methicillin-resistant s. aureus) and beta-hemolytic streptococci, including s. pyogenes. it can be used for treatment of furuncles, impetigo, open wounds, etc. mupirocin is bacteriostatic at low concentrations and bactericidal at high concentrations. it has a unique mechanism of action, which is selective binding to bacterial isoleucyl-trna synthetase, which halts the incorporation of isoleucine into bacterial proteins. following intravenous or oral administration, mupirocin is rapidly metabolised. the principal metabolite is monic acid, which has no antibacterial activity; therefore, its indication is topical treatment of bacterial skin infections. mupirocin is mostly formulated as 2% macrogol ointment and was originally marketed by glaxosmithkline under trade name bactroban® [35−37]. in a rapid in vitro screening, the effect of both forms of alaptide on permeation of mupirocin from bactroban® leciva ointment through full-thickness pig ear skin using a franz cell was investigated. n h h3c o h n o alaptide o o cooh h3c oh ho o oh ch3 ch3 o mupirocin figure 1. structures of (s)-alaptide and mupirocin. experimental preparation of alaptide alaptide was synthesised according to the standard process [38]. all reagents were purchased from aldrich and merck. alaptide was micronized using a planetary ball mill pm 200, retsch (germany). the particle size of the used micronized alaptide was measured by a nikon optiphot 2 microscope with a vds ccd-1300f digital camera. the particle size distribution (x90) of microcrystalline alaptide was 40 μm. the suspension of alaptide (30 g), polyvinylpyrrolidone (30 g) and purified water (240 ml, diluted during milling by the addition of 150 ml) was initially mixed for 12 h at ambient temperature and then filtered through a mill sieve. the milling procedure was performed using a nanomill netzsch (germany) with glass beads (0.3 mm); the rotor speed was 986 rpm; the pump speed was 30 rpm; the temperature in the grinding chamber was within 17–20 °c. the rotor speed was increased to 1500 rpm after 6 h of milling. the total time of milling was 57.5 h. the content of alaptide in the suspension was 38.76 g/l (determined by rphplc, see below). the particle size of the prepared nanonized alaptide was measured by sympatec nanophox 0138 p (germany) and the particle size x90 was 770 nm. in vitro transdermal permeation experiments skin samples were obtained from a porcine ear. full thickness ear skin was cut in fragments and stored at −20°c until utilised. skin samples were slowly thawed (at 4 °c overnight and then at ambient temperature) prior to each experiment. the penetration enhancing effect of micronized alaptide (ala) and nanonized alaptide (nala) was evaluated in vitro, using a vertical franz diffusion cell (ses – analytical systems, germany) with a donor surface area of 0.6359 cm 2 and a receptor volume of 5.2 ml. the skin was mounted between the donor and receptor compartments of the franz diffusion cell with the epidermal side up. the receptor compartment was filled with phosphate buffered saline (ph 7.4) and maintained at admet & dmpk 2(1) (2014) 56-62 mupirocin permeation doi: 10.5599/admet.2.1.26 59 34±0.5 °c using a circulating water bath. the receptor compartment content was continuously stirred using a magnetic stirring bar. the skin was kept in contact with the receptor phase for 0.5 h prior to the experiment. bactroban® leciva (2% macrogol ointment produced by zentiva prague, czech republic) was purchased from a pharmacy. the ointment was mixed with ala or nala (1% related to drug); ala or nala dissolved in the ointment and the prepared monophasic compositions were applied to the skin surface as a donor sample and the donor compartment of the cell was covered by parafilm ® . bactroban® leciva was used as a control. samples (0.5 ml) of the receptor phase were withdrawn at four pre-determined time intervals (30, 60, 90 and 120 min) and the cell was refilled with an equivalent amount of fresh buffer solution. a minimum of five determinations was performed using skin fragments from a minimum of two animals for each compound and the data was expressed as means ± sd. the samples were immediately analysed by hplc. analysis of samples was performed using an agilent 1200 series hplc system equipped with a diode array detection (dad) system, a quaternary model pump and an automatic injector (agilent technologies, germany). data acquisition was performed using chemstation chromatography software. a zorbax eclipse xdb c18 5 μm, 4.6×150 mm (agilent, usa) chromatographic column was used. the total flow of the column was 0.5 ml/min; injection was 10 μl; column temperature was 30 °c; sample temperature was 10 °c. the detection wavelength of 210 nm was chosen and the time of analysis was eight minutes. a mixture of meoh (hplc grade, 75.0%) and h2o (hplc – mili-q grade, 25.0%) was used as a mobile phase. mupirocin standard was purchased from sigma-aldrich. the retention time (tr) of mupirocin was 3.4±0.05 min; the limit of detection (lod) was 9.5 ng/ml; the limit of quantification (loq) was 31.7 ng/ml. results and discussion mupirocin was chosen as a frequently applied antibiotic for topical treatment of bacterial skin infections that are often associated with various injuries of the skin, because alaptide itself showed significant curative effect on the skin [17,25]. it was supposed that alaptide could influence the time of mupirocin retaining on/in the skin and act synergistically with this antibiotic. the rapid in vitro skin permeation screening experiments were performed using static franz diffusion cells [39] in a time limit of two hours only. full-thickness pig ear skin was selected for in vitro evaluation of permeation. this tissue is a suitable in vitro model of human skin [40,41], because porcine skin has shown to be histologically and biochemically similar to human skin and has previously been used in numerous studies [42]. the permeation of mupirocin through full-thickness pig ear skin with and without 1% (w/w related to drug) ala or nala was tested using bactroban® leciva, i.e., macrogol ointment as a donor vehicle. the results are listed in table 1 and illustrated in figure 2. table 1. cumulative permeated amounts qt per unit area [μg/cm 2 ] of mupirocin (mup) from bactroban® leciva (macrogol ointment) with and without micronized (ala) or nanonized (nala) alaptide as a modifier of transdermal permeation achieved in in vitro transdermal permeation experiments using the franz diffusion cell. qt are expressed as mean ± sd (n = 5 experiments). time [h] cumulative permeated amounts qt per unit area[µg/cm 2 ] mup mup+ala mup+nala 0.5 301.1±5.8 730.4±6.9 105.8±4.7 1.0 403.7±6.0 1951.8±8.6 158.3±4.9 1.5 2861.7±15.0 2337.1±7.0 208.4±7.4 2.0 4062.7±14.2 3055.2±3.5 248.7±8.8 opatrilova and jampilek admet & dmpk 2(1) (2014) 56-62 60 figure 2. in vitro profile of cumulative permeated amounts qt per unit area [μg/cm 2 ] in time of mupirocin (mup) alone and after addition of 1% w/w of micronized (ala) or nanonized (nala) alaptide from bactroban® leciva through skin. qt values are expressed as mean ± sd (n = 5 experiments). as the current study represents a preliminary rapid in vitro screening, all the experiments were performed only in the time range from 0.5 to 2.0 h. the values obtained from the permeation experiments were expressed as the cumulative permeated amount of the drug (qt [μg]) per unit of skin surface area (0.6359 cm 2 , see table 1). due to the short time interval of experiments, steady-state permeation flux and lag time could not be calculated; however, this was not the aim of the experiments. the dependences of the cumulative permeated amount of the drug per unit of skin surface area in time are shown in figure 2. based on the results obtained, it can be stated that mupirocin without ala did not permeate within 1 h after application and the permeated amount of mupirocin started to increase rapidly after 1 h; in 1.5 h, the permeated concentration of mupirocin was slightly higher than with the addition of ala. on the other hand, in the system containing mupirocin with ala, the permeated amount of the drug increased approximately linearly with time (r = 0.988) during the whole experiment (0-2 h) and in 0.5 and 1h it reached approx. 2.5-fold or 5-fold higher values than in the system containing mupirocin without ala. thus, it can be concluded that the application of ala to mupirocin in bactroban® leciva caused a uniform linear increase within 1 h when the permeation through the skin with ala was faster than the permeation without ala. an absolutely different effect was observed with the system containing nala: the addition of 1% w/w of nala to bactroban® leciva led to a significant decrease of permeation (in fact, the blocking of permeation) of mupirocin from bactroban® leciva through the skin during the experiment (0-2h), indicating that both mupirocin and alaptide can act curatively only on the surface of the skin. although it was found that nanoparticles, especially solid lipid nanoparticles also enhanced skin penetration, primarily due to an increase in skin hydration caused by the occlusive film formed by these particles on the skin surface [43], and that nala promoted permeation of various anti-inflammatory drugs, analgesics and antimicrobial chemotherapeutics through the skin in different studies [32,33], in this case, nala blocked permeation of mupirocin from macrogol. the effect of the lipophilicity of the tested drug, as well as the effect of the applied medium/formulation used in this study on behaviour/properties of nala as a permeation modifier will be the aim of further investigation. 0 500 1000 1500 2000 2500 3000 3500 4000 4500 0.5 1 1.5 2 q t [ g /c m 2 ] time [h] mup mup+ala mup+nala admet & dmpk 2(1) (2014) 56-62 mupirocin permeation doi: 10.5599/admet.2.1.26 61 based on the findings above, it can be assumed that the evaluated pharmaceutical composition of mupirocin can be used with success in combination with alaptide. as mentioned above, alaptide administered topically/locally supports/induces much faster regeneration and healing of the skin; therefore, it can be supposed than alaptide can potentiate the curative effect of mupirocin. the selection of the micronized or nanonized form of alaptide can influence the depth and the rate of permeation to the skin, i.e., the curative effect. conclusions the ability of alaptide in micronized and nanonized forms to modify permeation of mupirocin into/through the porcine ear skin was examined using the franz cell in the rapid preliminary in vitro screening. micronized and nanonized alaptide applied in a concentration of 1% w/w related to the amount of mupirocin influenced the permeation/penetration of the drug from the bactroban® leciva macrogol ointment through the skin. according to the calculated cumulative permeated amount of the drug per unit of skin surface area, in time, mupirocin with the addition of micronized alaptide, permeated approx. 2.5fold or 5-fold more in 0.5 and 1.0 h than without alaptide, i.e., micronized alaptide can significantly accelerate the impending effect of mupirocin within 1 h, whereas nanonized alaptide inhibited the permeation of mupirocin from the macrogol ointment. this finding, as well as confirmation of the mechanism of action of alaptide, will be the aim of our next investigation. nevertheless, on the basis of the preliminary results of the rapid preliminary screening, it can be concluded that alaptide combined with mupirocin can influence the depth and the rate of permeation of mupirocin into/through the skin. moreover, alaptide can act as a skin protecting substance due to its skin curative effect. acknowledgements: this study was supported by the grant agency of the czech republic (czech science foundation); project number gacr p304/11/2246. references [1] h.y. thong, h. zhai, h.i. maibach, skin pharmacology and physiology 20 (2007) 272-282. [2] i.b. pathan, c.m setty, tropical journal of pharmaceutical research 8 (2009) 173-179. [3] r.k. subedi, s.y. oh, m.k. chun, h.k. choi, archives of pharmacal research 33 (2010) 339-351. [4] j. jampilek, k. brychtova, medicinal research reviews 32 (2012) 907–947. [5] j. jampilek, journal of bioequivalence & bioavailability 5 (2013) 233–235. [6] b. bhushan, handbook of nanotechnology, springer-verlag, berlin-heidelberg, 2004. [7] c. buzea, i.i. pacheco, k. robbie, biointerphases 2 (2007) mr17–mr71. [8] e. corredor, p.s. testillano, m.j. coronado, p. gonzalez-melendi, r.f. pacheco, c. marquina, m.r. ibarra, j.m. de la fuente, d. rubiales, a. perez-de-luque, m.c. risueno, biomed central plant biology 9 (2009) 11 pages. [9] a. verma, f. stellacci, small 6 (2010) 12–21. [10] m. schneider, f. stracke, s. hansen, u.f. schaefer, dermato-endocrinology 1 (2009) 197–206. [11] b. baroli, journal of pharmaceutical sciences 99 (2010) 21–50. [12] t.w. prow, j.e. grice, l.l. lin, r. faye, m. butler, w. becker, e.m. wurm, c. yoong, t.a. robertson, h.p. soyer, m.s. roberts, advanced drug delivery reviews 63 (2011) 470–491. [13] c.s.j. campbell, l.r. contreras-rojas, m.b. delgado-charro, r.h. guy, journal of controlled release 162 (2012) 201–207. [14] e. kimura, y. kawano, h. todo, y. ikarashi, k. sugibayashi, biological & pharmaceutical bulletin 35 (2012) 1476–1486. opatrilova and jampilek admet & dmpk 2(1) (2014) 56-62 62 [15] a.c. watkinson, a.l. bunge, j. hadgraft, m.e. lane, pharmaceutical research 30 (2013) 1943–1946. [16] e. kasafirek, j. vanzura, i. krejci, j. krepelka, a. dlabac, m. valchar, (united pharmaceutical works & research institute for pharmacy and biochemistry), belg. 897843 (1984) & cs 231227 (1986). [17] s. radl, e. kasafirek, i. krejci, drugs of the future 15 (1990) 445–447. [18] e. kasafirek, i. krejci, z. hlinak, m. valchar, k. dobrovsky, a. sturc a, (united pharmaceutical works & research institute for pharmacy and biochemistry), us 5318973 (1994). [19] j. nedvidkova, e. kasafirek, j. nedvidek, k. pacak, v. schreiber, endocrine research 20 (1994) 39-46. [20] z. hlinak, i. krejci, s. hynie, v. klenerova, neuroendocrinology letters 29 (2008) 917-923. [21] j. vanzura, k. kosar, e. kasafirek, toxicology letters 31 (1986) 189-193. [22] s. blazickova, j. rovensky, e. kasafirek, m. buc, folia biologica 42 (1996) 57-59. [23] l. korbova, j. cizkova, j. kohout, e. kasafirek, i. krejci, j. vanzura, journal of czech physicians 127 (1988) 1574-1577. [24] l. korbova, j. kohout, e. kasafirek, gastroenterology and hepatology 48 (1994) 170-182. [25] e. kasafirek, l. korbova, j. kohout, m. jiraskova, i. krejci, a. galatik, (united pharmaceutical works & research institute for pharmacy and biochemistry), cs 276270 (1992). [26] j.a. mcgrath, r.a. eady, f.m. pope, rook's textbook of dermatology, 7th ed. t. burns, s. breathnach, n. cox, c. griffiths, (eds)., blackwell publishing, oxford, 2004, pp. 3–7. [27] w. james, t. berger, d. elston. andrews' diseases of the skin: clinical dermatology, 10th ed., saunders-elsevier, philadelphia, 2006, pp. 5–6. [28] f.m. watt, bioessays 8 (1988) 163–167. [29] r. lapka, journal of pharmacy and pharmacology, 43 (1991) 874-876. [30] k. kosar, j. vanzura, pharmazie, 43 (1988) 715-716. [31] j. jampilek, r. opatrilova, a. rezacova, z oktabec, p pavek, v kral, j. dohnal, (faculty of pharmacy, university of veterinary and pharmaceutical sciences brno), cz pv 2011-232 (2011), http://spisy.upv.cz/applications/2011/ppvcz2011_0232a3.pdf. [32] j. jampilek, r. opatrilova, l. coufalova, a. cernikova, j. dohnal, (faculty of pharmacy, university of veterinary and pharmaceutical sciences brno), wo/2013/020527 a1 (2013), http://patentscope.wipo.int/search/en/wo2013020527. [33] j. jampilek, a. cernikova, r. opatrilova, j. dohnal, (faculty of pharmacy, university of veterinary and pharmaceutical sciences brno), cz pv 2013-1000 (2013). [34] r. opatrilova, a. cernikova, l. coufalova, j. dohnal, j. jampilek, the scientific world journal, 2013 (2013), article id 787283, 8 pages, http://www.hindawi.com/journals/tswj/2013/787283/. [35] a.t. fuller, g. mellows, m. woolford, g.t. banks, k.d. barrow, e.b. chain, nature 234 (1971) 416-417. [36] c.m. thomas, j. hothersall, c.l. willis, tj simpson, nature reviews microbiology 8, (2010) 281–289. [37] drug bank – mupirocin, http://www.drugbank.ca/drugs/db00410 (december 12, 2013). [38] e. kasafirek, m. rybak, i. krejci, a. sturs, e. krepela, a. sedo, life science 50 (1992) 187–193. [39] t.j. franz, journal of investigative dermatology 64 (1975) 190–195. [40] u. jacobi, m. kaiser, r. toll, s. mangelsdorf, h. audring, n. otberg, w. sterry, j. lademann, skin res technol 13 (2007) 19–24. [41] c. herkenne, a. naik, y.n. kalia, j. hadgraft, guy r.h., pharmaceutical research 23 (2006) 1850-1856. [42] w. meyer, k. schwarz, k.t. neurand, current problems in dermatology 7 (1978) 39–52. [43] h.a.e. benson, current drug delivery 2 (2005) 23–33. ©2014 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://spisy.upv.cz/applications/2011/ppvcz2011_0232a3.pdf http://patentscope.wipo.int/search/en/wo2013020527 http://www.hindawi.com/journals/tswj/2013/787283/ http://www.drugbank.ca/drugs/db00410 http://creativecommons.org/licenses/by/3.0/ manuscript doi: 10.5599/admet.3.2.183 122 admet & dmpk 3(2) (2015) 122-130; doi: 10.5599/admet.3.2.183 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index review perspectives on caenorhabditis elegans models of human parkinson’s disease martina rudgalvyte 1 and garry wong 1,2 * 1 department of neurobiology, a.i. virtanen institute for molecular sciences, pl 1627, university of eastern finland, kuopio 70211, finland 2 faculty of health sciences, university of macau, macau s.a.r., taipa 999078, china *corresponding author: e-mail: garrygwong@umac.mo; tel.: +853-8822-4979; fax: +853-8822-2314 received: may 01, 2015; revised: june 21, 2015; published: july 01, 2015 abstract caenorhabditis elegans is a 1 mm long nematode comprised of 959 cells in the adult hermaphrodite. through transgenic injection, neurotoxin treatment, or isolation of mutants, this roundworm has been used as an animal model for studies of human parkinson’s disease (pd). the ability to genetically manipulate this animal, its short reproductive cycle and transparent body type have allowed it to be treated pharmacologically and toxicologically and interrogated for features of pd including loss of dopaminergic neurons, aggregation of α-synuclein protein, basal slowing responses to food, and lifespan. this short review aims to capture some of the recent studies on caenorhabditis elegans pd models and highlight some aspects of absorption, distribution, metabolism, and excretion that make the worm a useful organism for studies in neurodegeneration. keywords keyword; dopamine, nematode, transgenic model, neurodegeneration, synuclein. introduction parkinson’s disease (pd) is a devastating neurological disorder for which there is currently no cure. symptoms of the disease include severe motor impairment such as tremor, slowness of movement, rigidity, and postural instability, and involuntary movement. the disease etiology is attributed to progressive loss of dopaminergic neurons from the basal ganglia. neuropathological hallmarks of pd consists of lewy bodies and lewy neurites which are proteinaceous inclusions containing α-synuclein. while there is a genetic basis for pd, the overwhelming number of cases arises sporadically. to date, at least 20 human genes (park1-20) have been identified that contribute to pd. several recent reviews have dealt in detail with the contributions of caenorhabditis elegans (c. elegans) models in understanding the development and progression, and role of genetic factors in pd [1-5]. those reviews cover in depth the genetic models available and their contributions to understanding the neuropathology of the disease. this short review aims to take a broad pharmacological perspective in order to demonstrate the utility of c. elegans pd models in studying not only basic mechanisms but also potential treatments. http://www.pub.iapchem.org/ojs/index.php/admet/index admet & dmpk 3(2) (2015) 122-130 caenorhabditis elegans models of parkinson’s disease doi: 10.5599/admet.3.2.183 123 c. elegans is a round worm nematode that naturally lives in soil. its adult length of approximately 1 mm contains exactly 959 cells in the hermaphrodite sex and 1031 in males. of these cells, 302 are neurons of which 8 are dopaminergic in the hermaphrodite. the male has 6 additional dopaminergic neurons in the tail. the basic body plan consists of an outer cuticle that acts as a flexible tube and an inner tube that comprises the intestine. c. elegans is transparent, so neurons can be visualized in living animals through differential contrast or fluorescence microscopy. the reproductive life cycle of 2.5 days and life span of 3 weeks also provides vast numbers of easily reproduced animals for experiments. finally, in the laboratory, c. elegans is cultured on the surface of agar plates with e. coli as a food source and can also be cultured in liquid making pharmacology and toxicology studies straightforward. the specific dopaminergic neuron toxins 6-hydroxy dopamine (6-ohda), and 1-methyl-4-phenyl-1,2,3,6tetrahydropyridine (mptp), the precursor of toxic 1-methyl-4-phenylpyridinium (mpp+) have been used widely in c. elegans [6,7]. the toxic effects of 6-ohda (10 mm) have been pharmacologically shown to be via uptake through the dopamine transporter since these effects can be blocked by inhibitors damphetamine (10 mm) or imipramine (1 mm) [6]. more recently, dopaminergic neuron degeneration has been shown with heavy metals such as methyl mercury [8], manganese [9,10], and aluminum [11]. more general toxins directed at mitochondrial electron transport and oxidative stress includes the pesticides paraquat and rotenone, and these have been used as well in c. elegans to generate neurodegeneration models [12]. transcriptomic studies have verified the role of oxidative stress and misfolded protein pathways in the toxicology of these agents [13]. in addition, the use of gfp protein expressed in dopaminergic neurons has greatly facilitated the scoring of neurodegeneration in these chemical models. this visualization in living animals allows for direct and speedy determination of the effectiveness of chemical treatments. of the 20 park genes, 11 have c. elegans orthologs [1-5]. the orthologs can be used to model human pd by transgenic overexpression if the mode of inheritance is dominant. alternatively, homozygous mutants or rnai knockdowns can be used if the mode is recessive. for park genes without a c.elegans ortholog, dominant wildtype and mutant versions of the gene can be over expressed ectopically. use of cell type specific promoters facilitates localization of transgenic proteins. recessive park genes without a c. elegans ortholog are not possible to study. one of the most useful transgenic models utilizes the overexpression of α-synuclein (park1), both wildtype and mutant forms in c. elegans dopaminergic neurons [14-15]. in several studies, dopaminergic neuron degeneration as well as alterations on lifespan and movement has been shown. another useful model overexpresses α-synuclein in body wall muscles [16-17]. while these are not neurons, the muscle cells are large and lend themselves for rapid and automated measurement of fluorescence desired in genetic and drug screens. another transgenic model is transgenic overexpression of human mutant lrrk2 (park8) using a pan neuronal promoter [12]. these animals showed increased loss of gfp fluorescence in dopaminergic neurons as well as dopamine levels. a specific mutant (g2019s) was also shown to be more vulnerable to rotenone induced neurodegeneration. finally, knockdown of lrk-1, the c. elegans ortholog of human lrrk2, showed reduced survival compared to wild type controls when treated with the mitochondria stressor rotenone (25 μm). in addition to lrk-1, some mutant c. elegans lines are available for study: pdr-1 (park2), pink-1 (park6), djr-1.1 (park7) (reviewed in [1]). rudgalvyte and wong admet & dmpk 3(2) (2015) 122-130 124 figure 1. microscopy images of c. elegans head region. a. transgenic animal expressing gfp in dopaminergic neurons and associated processes [dat-1p::gfp]. b. light micrograph of pharynx showing mouth, anterior bulb, isthmus, terminal bulb and connection to intestine. c. transgenic animal expressing gfp in excretory pore and excretory duct cells [hmp-1p::hmp-1::gfp + dlg-1p::dlg-1::dsred + prf4(rol-6(su1006))]. admet in c. elegans absorption c. elegans has principally 2 routes of absorption: orally via feeding and dermally through the cuticle. the worm takes food through a multi-part structure termed the pharynx which consists of: a buccal cavity, procorpus, metacorpus with an anterior bulb, istmus, terminal bulb, and valve to separate the pharynx from intestine [18]. food in the pharynx is crushed in the terminal bulb and passed to the intestine where nutrients are absorbed. the muscle cells of the pharynx contract ~200 times per minute allowing for a large amount of food relative to its body size to be passed through per unit time. within the intestine, proteases, peptidases, and lipases digest material passed from the pharynx [19]. the worm intestine is long, comprises about one-third of the worm mass and extends throughout the body terminating at the anal cavity. nutrients and substances are absorbed from the intestine to the pseudocoelomic space. similar to higher organisms, the worm intestine has microvilli structures which maximize surface area for absorption and extends into the intestinal lumen. in contrast to the intestine, which is organized to rapidly assimilate small molecules and peptides, the cuticle is a barrier designed to keep chemicals out. the c. elegans cuticle is an extracellular matrix that acts like skin and covers essentially the entire c. elegans body as an exoskeleton. the cuticle is composed of proteins typically found in skin such as cross linked collagens [20], but also has insoluble proteins termed cuticlins. while flexible to allow for worm movement, it also acts to protect the animal from both mechanical and chemical insults. the cuticle itself is covered by an epicuticle consisting of glycoproteins and lipids. molting during each of the 4 larvae stages of the c. elegans life cycle permits the cuticle to be exchanged during each stage. a less well known and minor route of absorption is via sensory cilia in the c. elegans head. these are specialized structures of neurons which send process from the cell body to interact with the environment. small molecules can pass from these ciliated endings toward neuronal cell bodies as evidenced by experiments using fluorescence dyes. distribution once nutrients are absorbed by the intestine they enter the pseudocoelom of the worm. this area, also known as the body cavity, is bounded by the basal lamina of hypodermal cells and the walls of the intestine and gonads [21]. epidermal cells of the intestine and gonads form belt junctions which prevent contents once in the space from passing back out between the pseudocoelom and these structures. the admet & dmpk 3(2) (2015) 122-130 caenorhabditis elegans models of parkinson’s disease doi: 10.5599/admet.3.2.183 125 pseudocoelom is under hydrostatic pressure. graphically, one can think of the pseudocoelom as the area between tubes in a body structure of a “tube within a tube”. there is no circulatory system in c. elegans and it is presumed that contents in the space are moved during physical motion of the worm. there is no barrier in the pseudocoelom to neurons located in the nerve ring located between the bulbs of the pharynx or to motor neurons located in the ventral nerve cord. therefore, once compounds pass into the pseudocoelomic space, exposure to neuronal cells can be immediate. c. elegans also has a family of multidrug resistant proteins to exclude unwanted chemicals from entering cells [22]. metabolism small molecule metabolism occurs through the c. elegans cytochrome p-450 system (cyp). c. elegans contains 77 known cyp genes [23]. they are expressed in different structures including cuticle, intestine, and also in neurons. cyp genes show similar characteristics to their orthologs in higher organism: they are inducible by xenobiotics, participate in signaling molecule biosynthesis, and for a vast majority encode a protein for which the endogenous or exogenous substrate is not known [23-24]. the best characterized cyps reside in the intestine where fat stores are located. the cyp35 family members are expressed in intestine and regulate fat storage and levels of endocannabinoids. cyp29a3 and cyp33e2 are involved in biosynthesis of fatty acid derived signaling molecules for hydroxylation and epoxylation of eicosapentaenoic acid [25]. in addition to cyps, c. elegans has 5 flavin monooxygenases (fmo) [26]. the role of these genes in c. elegans physiology are not known, but they appear to be highly inducible (>50fold) by exposure to methyl mercury [13]. excretion excretion in c. elegans occurs through 2 structures. the anus is a specialized structure controlled by a specialized muscle group containing an intestinal muscle, anal sphincter, and anal depressor via a neuromuscular gabaergic junction. when feeding, c. elegans undergoes a regular 45 sec defecation cycle. the defecation cycle consists of: an anterior body contraction, a posterior body contraction, opening of an anal sphincter, and expulsion of anal cavity contents [27]. the excretion route is supported by another structure that terminates as an excretory pore located underneath the isthmus in the pharynx. the excretory system is made up of 4 cells: 2 excretory gland cells, an excretory duct cell, and an excretory pore cell [28]. the excretory pore cell is very large and extends 2 parallel canals along the entire length of worm body. excretion materials are thus collected through the canals for distal sites. excretion of contents from the excretory pore or anus includes saline fluid for osmoregulation and secondary metabolism contents. prior to excretion, oxidized small molecules are conjugated to glutathione or glucuronic acid through a family of glutathione s transferases (gst) or glucuronosyltransferases (ugt), respectively. like their phase i metabolism members, many of these phase ii members are also highly inducible under oxidative stress conditions [13,29]. toxicology c. elegans is a robust organism from a biochemical perspective and can tolerate very high concentrations of toxic compounds. for example, the ld50 at 24h for ethanol is >50 mg/ml and atropine sulphate is >20 mg/ml [30]. animals also have heavy metaland multidrug resistant abc transporters. very good correlation (r=0.885) in acute toxicity of 21 different compounds has been shown between rat and c. elegans, and this correlation value was found to be even better than mouse versus rat [30]. the nematode is however very sensitive to physical environment. c. elegans larvae or adult worms live in a narrow temperature range near 20°c. the animals undergo heat shock at 30 °c and cannot be maintained at 4 °c, rudgalvyte and wong admet & dmpk 3(2) (2015) 122-130 126 however their embryos can be frozen in glycerol solution. animals also desiccate easily. under crowding, desiccation, or other extreme physical stress, c. elegans enters an alternative life cycle termed dauer. in this state, the animal produces a protective protein coat, moves slowly, and can extend its life span from 3 weeks to 4 months. several excellent reviews on the topic of dauer stage and life in nature are available [31,32]. recent insights on pd using c. elegans while there is currently no cure for pd, current human therapy aims at treating symptoms and providing supportive therapy. the first line medications used to improve the main symptoms of pd are levodopa, dopamine agonists, monoamine oxidase-b inhibitors, and catechol-o-methyltransferase inhibitors. recently, surgical procedures have proven to be enormously successful and include deep brain stimulation. as an experimental model system, novel compounds have been shown to attenuate many of neuropathological hallmarks of pd. thus the aims of recent pharmacological studies in c. elegans have been to prevent or attenuate dopamine neuron cell death. some studies have also measured effectiveness towards symptoms such as uncoordinated movement and changes in life-span. below, we describe these studies. pharmacological stuedies spermidine is a polyamine that is present endogenously and acts in cell proliferation and differentiation presumably via an autophagic pathway. previous studies have shown that it can extend life span and therefore it was tested as an anti-neurodegenerative compound. spermidine (5 mm) supplemented in food was shown to attenuate the loss of dopaminergic neurons caused by overexpression of α-synuclein [33]. the authors also showed concurrent increased production of autophagosomes in treated animals suggesting a mechanism for spermidine actions. valproic acid, an approved anti-epileptic drug that acts via gabaergic signaling and also has effects on voltage gated sodium channels and t-type calcium channels. it was shown to attenuate dopaminergic neurodegeneration caused by transgenic overexpression of α-synuclein [34]. valproic acid, dissolved in the agar plates used to culture animals, was effective at both 2 and 3 mm concentrations, but not 1 mm. using rnai, the authors could negate the effect of valproic acid with knockdown of the mek-2 gene suggesting a pharmacological action involving the erk-mapk signaling pathway. acetylcorynoline is a major alkaloid component of the traditional chinese medical herb corydalis bungeana. it has indications as an anti-inflammatory agent for upper respiratory tract infections, bronchitis, tonsillitis, and acute nephritis. acetylcorynoline dissolved in the agar (5mm) was able to decrease aggregation of α-synuclein expressed in muscle cells [35]. at the same concentration, it was also able to protect dopaminergic neurons from degeneration elicited by 6-ohda. furthermore, in 6-ohda treated animals, it restored the basal slowing response on food, a c. elegans behavior mediated through dopaminergic neurons. n-butylidenephthalide is an organic extraction product of angelica sinensis, and has been shown to have anti-inflammatory properties. with results similar to acetylcorynoline, it was also shown at 2 mm and 5 mm concentrations in agar to reduce α-synuclein aggregation in muscle cells, rescue loss of dopaminergic neurons and reverse loss of basal slowing response in 6-ohda treated animals [36]. synthetic medicinal chemistry approaches have also been used with c. elegans models providing an assay for novel compounds. 3-arycoumarin-tetracyclic tacrine hybrids have been developed as potential admet & dmpk 3(2) (2015) 122-130 caenorhabditis elegans models of parkinson’s disease doi: 10.5599/admet.3.2.183 127 anti-pd agents [37]. many members of this series of compounds are able to significantly decrease αsynuclein aggregation in muscle cells. molecular modeling of interaction between active compounds and αsynuclein suggests that more potent and selective derivatives might be possible as well as provide insight into the mechanism of anti-aggregation. xyloketal derivatives have also been developed as potential neuroprotective agents in pd [38]. in c. elegans animals with dopaminergic neuron degeneration produced by 1-methyl-4-phenylpyridinium (mpp+), the survival rate could be significantly increased with several derivatives. genetic studies genetic screens take advantage of the available genetic resources of the c. elegans community and the rapid reproductive cycle [39]. these screens aim to identify genes that when expressed enhance or suppress a pd phenotype in a c. elegans model. one rnai screen using α-synuclein-gfp fusion expressed in body wall muscle cells, identified 80 genes which influence α-synuclein inclusion formation in muscle cells during aging [17]. among the genes identified that suppress inclusion formation was sir2.1. not all 80 gene products could be druggable, but the rnai screening approach demonstrates perhaps the large number of possible proteins that affect a biochemical process such as protein aggregation. in contrast to rnai screens, forward genetic screens using random mutagenesis identify fewer genetic interactors, but assure that the gene product has been changed on the structural level. the mutagen ethyl methanesulfonate was used to randomly mutagenize the worm genome and identified tetraspanin (tsp17) as a protector of dopaminergic neuron degeneration elicited by 6-ohda [40]. tetraspanin was shown to inhibit the function of the dopamine transporter and this provided insight into dopamine neurotransmission. the nontargeted approach of screens highlights the insight and potential for new target discovery in pd. a more targeted approach has utilized existing knowledge that oxidative stress and its associated pathways play a key role in neurodegenerative processes. under oxidative conditions, the thioredoxin and glutaredoxin systems act to reduce disulfide bonds. these systems contain multiple proteins in both humans and c. elegans. mutants that lack thioredoxin reductase 1 (trxr-1) were shown to be more vulnerable to 6-ohda mediated dopaminergic cell death [41]. similarly, loss of glutaredoxin (grx-1) gene exacerbated dopaminergic neurodegeneration in a human lrrk2 mutant c. elegans model [42]. this effect was also seen in α-synuclein and tyrosine hydroxylase overexpressing animals. another perspective has been taken where oxidative stress via exposure to hydrogen peroxide (1mm) for 30 minutes has been shown to induce the transcription factor hlh-13, the c. elegans ortholog of p48 [43]. hlh-13 mutant animals were more vulnerable to dopaminergic cell loss 7 days after hydrogen peroxide treatment suggesting a protective role for this transcription factor. finally, rnai knockdown of c. elegans phosphotidylserine decarboxylase, an enzyme used in the synthesis of phosphatidylethnolamine, accelerates degeneration of α-synuclein over expressing dopaminergic neurons [44]. this effect can be rescued by ethanolamine which can be converted to phosphatidylethanolamine. these results highlight the importance of mitochondrial membrane integrity in pd but also may suggest ways to alleviate any deficiencies. rudgalvyte and wong admet & dmpk 3(2) (2015) 122-130 128 future perspectives it is now becoming clear that multiple model organism systems will be needed to tackle the challenging problems in understanding and ultimately treating pd. although at least 20 park genes have been identified [1], their expression or mutation in c. elegans has provided insight into disease etiology only with the assistance of and often in conjunction with other systems such as saccharomyces cerevisiae and drosophila melanogaster. several examples we highlight in this review follow this pattern. moreover, validation of c. elegans findings in human cell culture is proving to be more necessary, not only because pd is a human disease, but also because the genetic environment of a human cell more closely mimics that of the transgene. at the same time, while pd is an age-related neurodegenerative disease, it shares many of the same pathological features of alzheimer’s and huntington’s diseases. essentially, one principle feature of all these diseases is the presence of toxic protein aggregates. therefore, more attention has been paid to the ability of neuronal cells to maintain structural integrity of proteins in terms of folding, as well as aggregation and elimination. this field has been termed proteostasis and findings in this area will likely shed light across many age related disorders [45]. while genetic manipulation of c. elegans has made it a powerful model to study, chemical manipulation to mimic environmental risk factors in neurodegenerative disorders have now advanced to the forefront. in addition to heavy metals and pesticides reviewed earlier, other toxic compounds humans are likely to encounter during their lifetime have been administered to c. elegans with the aim to access any damage to neuronal systems. these include nanoparticles, plasticizers, and cyanobacteria toxins [46-48]. considering the ease at which toxicity studies can be conducted, and the high correlation between acute toxicity for compounds between rat and c. elegans, it is somewhat surprising that this list is not currently more extensive. finally, the wealth of genomic, transcriptomic, and proteomic data available in public databases will help to better design and validate studies using c. elegans in pd [49,50]. a paradigm is emerging in neurodegenerative disease research where human gwas and transcriptomic data informs on gene candidates that can be investigated in c. elegans. genetic studies in c. elegans is then validated or complemented by similar studies in other organisms. hypothesis generation from these studies to look retrospectively at human disease databases then completes the investigative loop. ultimately, for a simple model organism, c. elegans has in the past and will continue into the future to inform on a human neurodegenerative disease with currently no cure. acknowledgements the authors wish to thank members of the wong laboratories at the department of neurobiology, a.i. virtanen institute and faculty of health sciences at the university of macau for their support of this work and critical comments of this manuscript. the work was funded by the academy of finland (m.r.) and a start-up and multi-year research grant from the university of macau (g.w.) admet & dmpk 3(2) (2015) 122-130 caenorhabditis elegans models of parkinson’s disease doi: 10.5599/admet.3.2.183 129 references [1] p.m.chege and g. mccoll, front aging neurosci. 6 (2014) 89. [2] h. deng and l. yuan, ageing res rev. 15 (2014) 15:161-176. [3] j. vistbakka, n. vanduyn, g. wong and r. nass, cns neurol disord drug targets. 11 (2012) 957964. [4] a.j. harrington, s. hamamichi, g.a. caldwell and k.a. caldwell, dev dyn. 239 (2010) 1282-1295. [5] r. nass and s. przedborski (editors) parkinson's disease: molecular and therapeutic insights from model systems, academic press, waltham, massachusetts, u.s.a. (2008) pp. 1-686. [6] r. nass, d.h. hall, d.m. miller 3 rd and r.d. blakely, proc natl acad sci u s a. 99 (2002) 3264-3269. [7] e. braungart, m. gerlach, p. riederer, r. baumeister and m.c. hoener, neurodegener dis. 1 (2004) 175-183. [8] n. vanduyn, r. settivari, g. wong and r. nass, toxicol sci. 118 (2010) 613-624. [9] r. settivari, j. levora and r. nass, j biol chem. 284 (2009) 35758-35768. [10] a. benedetto, c. au, d.s. avila, d. milatović and m. aschner, plos genet. 6 (2010) e1001084. [11] n. vanduyn, r. settivari, j. levora, s. zhou, j. unrine and r. nass, j neurochem. 124 (2013) 147157. [12] s. saha, m.d. guillily, a. ferree, j. lanceta, d. chan, j. ghosh, c.h. hsu, l. segal, k. raghavan, k. matsumoto, n. hisamoto, t. kuwahara, t. iwatsubo, l. moore, l. goldstein, m. cookson and b. wolozin, j neurosci. 29 (2009) 9210-9218. [13] m. rudgalvyte, n. vanduyn, v. aarnio, l. heikkinen, j. peltonen, m. lakso, r. nass and g. wong,toxicol lett. 222 (2013) 189-196. [14] m. lakso m, s. vartiainen, a.m. moilanen, j. sirviö, j.h. thomas, r. nass, r.d. blakely and g. wong, j neurochem. 86 (2003) 165-172. [15] t. kuwahara, a. koyama, k. gengyo-ando, m. masuda, h. kowa, m. tsunoda, s. mitani and t. iwatsubo , j biol chem. 281 (2006) 334-340. [16] s. hamamichi, r.n. rivas, a.l. knight, s. cao, k.a. caldwell and g.a. caldwell, proc natl acad sci usa. 105 (2008) 728-733. [17] t.j. van ham, k.l. thijssen, r. breitling, r.m. hofstra, r.h. plasterk and e.a. nollen, plos genet. 4 (2008) e1000027. [18] s.e. mango, (2007) wormbook, ed. the c. elegans research community, wormbook, doi/10.1895/wormbook.1.129.1 [19] j.d. mcghee, (2007), wormbook, ed. the c. elegans research community, wormbook, doi/10.1895/wormbook.1.133.1 [20] a.p. page and i.l. johnstone, (2007), wormbook, ed. the c. elegans research community, wormbook, doi/10.1895/wormbook.1.138.1 [21] wormatlas, altun, z.f., herndon, l.a., crocker, c., lints, r. and hall, d.h. (eds) 2002-2015. http://www.wormatlas.org (date accessed april 28, 2015). [22] n. vanduyn and r. nass , j neurochem. 128 (2014) 962-974. [23] v. aarnio, m. lehtonen, m. storvik, j.c. callaway, m. lakso and g. wong, front pharmacol. 2 (2011) 12. [24] j. peltonen, v. aarnio, l. heikkinen, m. lakso and g. wong, j biochem mol toxicol. 27 (2013) 219228. [25] j. kulas, c. schmidt, m. rothe, w.h. schunck and r. menzel, arch biochem biophys. 472 (2008) 6575. [26] m.i. petalcorin, g.w. joshua, p.m. agapow and c.t. dolphin, gene 346 (2005) 83-96. [27] d.j. reiner and j.h. thomas, j neurosci. 15 (1995) 6094-6102. [28] f.k. nelson and d.l. riddle, j exp zool. 231 (1984) 45-56. http://www.wormatlas.org/ rudgalvyte and wong admet & dmpk 3(2) (2015) 122-130 130 [29] r. settivari, n. vanduyn, j. levora and r. nass, neurotoxicology 38 (2013) 51-60. [30] y. li, s. gao, h. jing, l. qi, j. ning, z. tan, k. yang, c. zhao, l. ma and g. li, toxicology res. 2 (2013) 403-412. [31] n. fielenbach and a. antebi , genes dev. 22 (2008) 2149-2165. [32] l. frézal and m.a. félix, elife 4 (2015) elife.05849. [33] s. büttner , f. broeskamp, c. sommer, m. markaki, l. habernig, a. alavian-ghavanini, d. carmonagutierrez, t. eisenberg, e. michael, g. kroemer, n. tavernarakis, s.j. sigrist and f. madeo, cell cycle 13 (2014) 3903-3908. [34] b.b. kautu, a. carrasquilla, m.l. hicks, k.a. caldwell and g.a. caldwell, neurosci lett. 541 (2013) 116-119. [35] r.h. fu, y.c. wang, c.s. chen, r.t. tsai, s.p. liu, w.l. chang, h.l. lin, c.h. lu, j.r. wei, z.w. wang, w.c. shyu and s.z. lin, neuropharmacology 82 (2014) 108-120. [36] r.h. fu, h.j. harn, s.p. liu, c.s. chen, w.l. chang , y.m. chen, j.e. huang, r.j. li, s.y. tsai, h.s. hung , w.c. shyu, s.z. lin and y.c. wang , plos one 9 (2014) e85305. [37] k.v. sashidhara, r.k. modukuri, p. jadiya, k.b. rao, t. sharma, r. haque, d.k. singh, d. banerjee, m.i. siddiqi and a. nazir, acs med chem lett. 5 (2014) 1099-1103. [38] s. li, c. shen, w. guo, x. zhang, s. liu, f. liang, z. xu, z. pei, h. song, l. qiu, y. lin and j. pang , mar drugs. 11 (2013) 5159-5189. [39] o. sin, h. michels and e.a. nollen, biochim biophys acta. 1842 (2014) 1951-1959. [40] n. masoudi, p. ibanez-cruceyra, s.l. offenburger, a. holmes and a. gartner, plos genet. 10 (2014) e1004767. [41] l. arodin, a. miranda-vizuete, p. swoboda and a.p. fernandes, free radic biol med. 73 (2014) 328336. [42] w.m. johnson, c. yao, , s.l. siedlak, w. wang, x. zhu, g.a. caldwell, a.l. wilson-delfosse, j.j. mieyal and s.g. chen, hum mol genet. 24 (2015) 1322-1335. [43] p. bou dib, b. gnägi, f. daly, v. sabado, d. tas, d.a. glauser, p. meister and e. nagoshi, plos genet. 10 (2014) e1004718. [44] s. wang , s. zhang, l.c. liou, q. ren, z. zhang, g.a. caldwell, k.a. caldwell and s.n. witt, proc natl acad sci u s a. 111 (2014) e3976-3985. [45] m. brehme, c. voisine, t. rolland, s. wachi, j.h. soper, y. zhu, k. orton, a. villella, d. garza, m. vidal, h. ge and r.i. morimoto, cell rep. 9 (2014) 1135-1150. [46] y. li, s. yu, q. wu, m. tang, y. pu and d. wang, j hazard mater. 220 (2012) 221-230. [47] j. ju, n. saul, c. kochan, a. putschew, y. pu, l. yin and c.e. steinberg, int j environ res public health 11 (2014) 4589-4606. [48] i.l. tseng, y.f. yang, c.w. yu, w.h. li and v.h. liao, plos one 8 (2013) e8265. [49] s. asikainen, m. rudgalvyte, l. heikkinen, k. louhiranta, m. lakso, g. wong and r. nass, j mol neurosci. 41 (2010) 210-218. [50] parkinson’s mutation database, http://www.molgen.vib-ua.be/pdmutdb/, (date accessed april 28, 2015). ©2015 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://www.molgen.vib-ua.be/pdmutdb/ http://creativecommons.org/licenses/by/3.0/ manuscript doi: 10.5599/admet.2.1.29 63 admet & dmpk 2(1) (2014) 63-70; doi: 10.5599/admet.2.1.29 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index preliminary work study of formulation variables influencing polymeric microparticles by experimental design rameshwar k. deshmukh and jitendra b. naik* department of pharmaceutical technology, university institute of chemical technology, north maharashtra university, jalgaon, 425 001, maharashtra, india *corresponding author: e-mail: jitunaik@gmail.com; tel.: +91257 2258441; fax: +91257 2258403 received: december 21,2013; revised: march 01, 2014; published: april 01, 2014 abstract the objective of this study was to prepare diclofenac sodium loaded microparticles using the single emulsion [oil-in-water (o/w)] solvent evaporation method. the 2 2 experimental design methodology was using the design-expert ® software used to evaluate the effect of two formulation variables on the properties of the microspheres in terms of particle size, morphology, encapsulation efficiency, and in vitro drug release. the graphical and mathematical analysis of the design showed that the independent variables had a significant effect on the encapsulation efficiency and drug release of microparticles. the low magnitudes of error and significant values of r 2 prove the high prognostic ability of the design. the microspheres showed high encapsulation efficiency with an increase in the amount of polymer and a decrease in the amount of pva in the formulation. the particles were found to be spherical with a smooth surface. prolonged drug release and enhancement of encapsulation efficiency of polymeric microparticles can be successfully obtained by applying experimental design techniques. keywords: microencapsulation; microparticles; statistical experimental design; solvent evaporation; single emulsion; drug delivery introduction biodegradable and biocompatible microparticles are popularly investigated drug delivery systems for therapeutic drugs [1-5]. they are capable of providing controlled release of the encapsulated drug over a long period of time [6,7]. encapsulation of the drug within biomaterial-based microparticles can increase the stability of the drug and facilitate controlled and sustained release of the compound [8]. recently, various different encapsulation techniques have been employed to produce fine particles of various active compounds. many microencapsulation processes are modifications of the three basic techniques: solvent extraction/evaporation, phase separation (coacervation), and spray-drying techniques [9]. solvent evaporation and organic phase separation techniques are widely used in the pharmaceutical industry for the preparation of microparticles [1,2,9-12]. these microparticles can reduce side effects, enhance the efficacy of therapy, and increase patient compliance. non-steroidal anti-inflammatory drugs (nsaids) are usually indicated for the treatment of musculoskeletal disorders where pain and inflammation are present. diclofenac sodium [sodium(o-((2,6http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:jitunaik@gmail.com deshmukh and naik admet & dmpk 2(1) (2014) 63-70 64 dichlorophenyl)-amino)-phenyl)-acetate] is a nsaid that is commonly used drug for the relief of pain and inflammation in arthritis, rheumatoid arthritis, osteoarthritis, and pain management in case of kidney stones. the dose is 100 to 150 mg daily with a short duration of action. therefore, repeated dosing of the drug is essential for effective pain management. a major disadvantage of nsaid therapy is the potential for upper gastrointestinal (gi) complications and renal complications [13-19]. due to the high usage of nsaids, a drug delivery formulation which will achieve the highest healing effect with minimal gi events is desirable. for that purpose, a sustained release formulation is essential. therefore, polymeric microparticles of diclofenac sodium might be beneficial to overcome the side effects and increase efficacy of therapy and patient compliance. diclofenac sodium and ethylcellulose were used as the model drug and polymer for the preparation of microparticles. ethyl cellulose, an ethyl ether of cellulose, is a long chain polymer of β-anhydroglucose units joined together by acetal linkages. ethylcellulose is a hydrophobic polymer widely used in the pharmaceutical field as a coating material for the preparation of microcapsules, microparticles, tablets, and for controlled release dosage forms due to its wide range of practically relevant properties, including low toxicity (fda approved for internal use), good film forming abilities, and relatively low cost. high viscosity grades are used in drug microencapsulation and for sustained release products [13, 20-23]. pharmaceutical scientists often using the single factor optimisation technique (one factor at a time) to optimise a formulation. multiple factor optimisation is more efficient and generates a maximum point (an optimised condition) compared with the single factor optimisation technique. in many process development and manufacturing applications of pharmaceutical formulations, the number of potential input variables (factors) is greater. these formulations often face the challenge of identifying the key input variables or process conditions that affect product quality. however, this challenge can be overcome by using statistical design of experiment tools to carry out the minimum number of experiments to estimate the influence of individual variables on response variables [24-34]. the aim of this study was to evaluate, by means of a statistical experimental design methodology, the influence of the formulation variables on the encapsulation efficiency of diclofenac sodium loaded microparticles in order to obtain a sustained release system to treat musculoskeletal disorders. experimental materials diclofenac sodium was provided by natco pharma limited, (hyderabad, india) as a gift sample. ethylcellulose was kindly donated by colorcon india pvt. ltd. (mumbai, india). polyvinyl alcohol (pva), (mw 140,000) was purchased from himedia laboratories pvt. ltd. (mumbai, india). dichloromethane was obtained from merck specialities pvt. ltd. (mumbai, india). methanol (hplc grade) was purchased from rfcl limited (new delhi, india). all other solvents and reagents were of analytical grade and used as provided. experimental design before application of the design, a number of preliminary trials were conducted to determine the conditions at which the process resulted in microparticles. the levels of factors were also determined by this procedure. the 2 2 (two factors and two levels) factorial design was employed for the optimisation of diclofenac sodium loaded microparticles using the design-expert® software (version-8.0.7.1, stat-ease inc., minneapolis, mn) which allows evaluation from four experiments. the purpose of the design was to admet & dmpk 2(1) (2014) 63-70 polymeric microparticles doi: 10.5599/admet.2.1.29 65 evaluate the effects of the formulation factors and identify the key one that influences the encapsulation efficiency (dependent variable). the amount of ethylcellulose (a, mg), and the amount of pva (b, % w/v) were selected as independent variables. the levels of screening variables evaluated in this study are listed in table 1 along with their low and high levels. statistical analysis was performed and the significance of the model was determined by the comparisons of statistical parameters [35-37]. table 1. screening variables and their levels (coded and actual) used for 2 2 experimental design independent variables level used, actual (coded) low, (-1) high, (+1) a = amount of ethylcellulose (mg) 250 750 b = amount of pva as a surfactant (% w/v) 0.05 0.2 preparation of encapsulated microparticles microparticles containing diclofenac sodium were prepared using the oil-in-water (o/w) single emulsion solvent evaporation method. briefly, 250 mg of diclofenac sodium was mixed with a mixture of dichloromethane (10 ml), and methanol (10 ml) containing a certain proportion of ethylcellulose polymer as per the factorial design. this organic phase was dispersed in 100 ml of distilled water containing the varying amount of pva as an emulsifier to produce an o/w single emulsion at a speed of 800 rpm using four bladed lab stirrers (remi electrotechnik limited, thane, india). the external aqueous phase was adjusted to ph 3.9 with acetic acid. stirring was continued for 3 hours until the evaporation of organic solvent leaving solid microparticles. the microparticles were collected by filtration, rinsed with n-hexane, air dried for 48 hours, and used for further studies. characterisation of the microparticles the amount of diclofenac sodium encapsulated into microparticles was determined by uv-vis spectrophotometer (hitachi u-2900, tokyo, japan). an accurately weighed 10 mg of microparticles were stirred with dichloromethane (2 ml) to dissolve the polymeric coat, and was then extracted in phosphate buffer solution (ph 6.8). stirring continued for 30 min at room temperature to facilitate the evaporation of organic solvent. the dispersion was filtered, and the residue was washed with phosphate buffer solution. the encapsulation efficiency was determined in the filtrate after appropriate dilution with a phosphate buffer solution at 276 nm using a uv-vis spectrophotometer. the encapsulation efficiency was expressed as the percentage of drug incorporated in the formulation relative to the total amount of drug used in the formulation. the encapsulation efficiency of diclofenac sodium was calculated using the following equation: [ ] [ ] of all the formulations, microparticles with a higher encapsulation efficiency were chosen for further characterisation study. the particle size and shape of the drug loaded microspheres was determined by capturing the series of digital images using motic b1 advanced series digital microscope (motic, china) fitted with imaging accessories equipped with computer-controlled image analysis software (motic images plus, 2.0 ml version). calibration of an object lens was done using a standard calibrated circle. the sample was put on the slide, and the diameter of the particles was determined. deshmukh and naik admet & dmpk 2(1) (2014) 63-70 66 drug release from microparticles was performed in vitro using phosphate buffer (ph 6.8) for 12 hours in a tablet dissolution tester (electrolab tdt-06 t, india) usp xxviii, type i (100 rpm, 37 ± 0.5°c). the dissolution medium of phosphate buffer (ph 6.8) was prepared according to the indian pharmacopoeia [38]. aliquots of the dissolution medium were withdrawn at predetermined intervals and replenished with fresh dissolution media to maintain the sink condition. the samples were filtered through a whatman filter paper no. 41. the diclofenac sodium content of each sample after suitable dilution was assayed by uv-vis spectroscopy at λmax of 276 nm using a 1 cm cell. the experiments were done in triplicate. the release data were evaluated kinetically to study the possible mechanism of drug release from the microparticles. results and discussion experimental design the statistical experimental design is a useful and efficient mathematical approach to evaluate the effect of a factor on a particular response generated by conducting a smaller number of experimental trials. each analysed variable was evaluated at two levels, high (+1) and low (-1). the evaluation consisted of analysing the response in all the conditions quoted in table 2. the polynomial model describing the correlation between the formulation process variables and encapsulation efficiency can be represented by the following equation: e.e. = 72.68 + 3.83a – 8.68b in the above equation a negative sign of variable b signifies the antagonistic effect, while a positive sign of variable a signifies a synergistic effect to the responses. it can be observed from the above equations that % e.e. was influenced by both factors. from the anova of design for encapsulation efficiency, p-value, f-value, mean square and r 2 of the model are given in table 3. table 2. the 2 2 experimental design matrix (in coded level) and experimental results run independent variables encapsulation efficiency a b 1 750 0.05 84.93 ± 0.06 2 750 0.2 68.08 ± 0.04 3 250 0.2 59.91 ± 0.12 4 250 0.05 77.78 ± 0.01 table 3. analysis of variance for encapsulation efficiency source sum of squares d.f. mean square f value p-value prob > f model 359.87 2 179.94 691.51 0.0269 (s) aethylcellulose 58.68 1 58.68 225.50 0.0423 (s) bpva 301.20 1 301.20 1157.52 0.0187 (s) residual 0.26 1 0.26 total (corrected) 360.13 3 correlation coefficient (r 2 ) 0.9993 adjusted r 2 0.9978 predicted r 2 0.9884 press 4.16 d.f. degree of freedom, s significant admet & dmpk 2(1) (2014) 63-70 polymeric microparticles doi: 10.5599/admet.2.1.29 67 standardised pareto charts, (fig. 1) representing the estimated effects of parameters on response, can allow us to check the statistical significance of experimental design. it consists of bars with a length proportional to the absolute value of the estimated effects divided by the standard error, which is the tvalue of the student’s t-test. it can be observed that ethylcellulose and pva have significant influence on the encapsulation efficiency. figure 1. pareto chart of standardised effects on the encapsulation efficiency characterisation of the microparticles the encapsulation efficiency of different experimental runs of microparticles is reported in table 2. two-dimensional (2d) contour plots and three-dimensional (3d) response plots resulting from the equations were constructed to visualise the effect of independent variables on the response using designexpert ® software [39]. it was observed that the encapsulation efficiency was higher with a high level of ethylcellulose and low level of pva. surface response plots and contour plots (fig. 2 a & b) represent the increase in % encapsulation efficiency with an increase in the amount of ethylcellulose and a decrease in the amount of pva. figure 2. response surface plot [a] and contour plot [b,] showing the effect of independent variables on response the shape, size, and surface morphology of the diclofenac sodium loaded microparticles were examined by motic microscope. in this study, the particle size of the drug loaded microspheres captured deshmukh and naik admet & dmpk 2(1) (2014) 63-70 68 using motic b1 advanced series digital microscope (fig. 3) was found to be in the range of 10–200 µm. the particles were found to be discrete with spherical geometry. figure 3. image of microparticles by motic microscope ethylcellulose is a water insoluble polymer able to control drug release from microparticles. fig. 4 shows the cumulative % drug release as a function of the dissolution time from the diclofenac sodium loaded microparticles. the release of diclofenac sodium was evaluated using phosphate buffer (ph 6.8) as the release media. prolonged drug release (92 %) (fig. 4) from the microparticles was observed over 12 hours. the dissolution data were fitted into the matrix model equation. the sample showed good linearity (r 2 : 0.9764) with a value of the slope (n) < 0.45. it indicates that fickian diffusion is the dominant mechanism of drug release from these microspheres. figure 4. cumulative % drug release from the microspheres conclusions this work shows that diclofenac sodium loaded ethylcellulose polymeric microparticles were successfully prepared. the effects of formulation variables had been screened out by employing 2 2 experimental design. the formulation variables have a significant effect on the encapsulation efficiency of microspheres. design and optimisation through statistical experimental designs work well in the admet & dmpk 2(1) (2014) 63-70 polymeric microparticles doi: 10.5599/admet.2.1.29 69 development of polymeric microparticles. microparticles prepared by such methods may represent a promising approach for the enhancement of encapsulation efficiency of diclofenac sodium. acknowledgements: authors are very much thankful to university grants commission, new delhi for providing financial assistance to carry out this research work. the authors gratefully acknowledge natco pharma limited, (hyderabad, india), evonik degussa india pvt. ltd. (mumbai, india) & colorcon india pvt. ltd. (mumbai, india) for providing the gift sample of diclofenac sodium & ethylcellulose respectively. references [1] a. lamprecht, h.r. torres, u. schafer, c.m. lehr, journal of controlled release 69 (2000) 445-454. [2] j. herrmann, t. bodmeier, international journal of pharmaceutics 126 (1995) 129-138. [3] h.s. yoo, colloids and surfaces b: biointerfaces 52 (2006) 47-51. [4] e. polard, p.s. corre, f. chevanne, r.l. verge, international journal of pharmaceutics 134 (1996) 3746. [5] j.b. naik, r.k. deshmukh, v.v. kamble, american journal of pharmtech research 3 (2013) 892-904. [6] k. avneshi, s.k. yadav, s.c. yadav, colloids and surfaces b: biointerfaces 75 (2010) 1-18. [7] s. raveendran, a.c. poulose, y. yoshida, t. maekawa, d.s. kumar, carbohydrate polymer 91 (2013) 22-32. [8] a. kochut, p. dersch, european journal of pharmaceutics and biopharmaceutics 84 (2013) 242-250. [9] s. freitas, h.p. merkle, b. gander, journal of controlled release 102 (2005) 313-332. [10] r. bodmeier, h. chen, p. tyle, p. jarosz, journal of controlled release 15 (1991) 65-77. [11] m. li, o. rouaud, d. poncelet, international journal of pharmaceutics 363 (2008) 26-39. [12] h. heiskanen, p. denifl, p. pitkanen, m. hurme. chemical engineering research and design 90 (2012) 1517-1526. [13] j.l. arias, m. lopez-viota, j. lopez-viota, a.v. delgado, international journal of pharmeutics 382 (2009) 270-276. [14] j.t. edelson, a.n. tosteson, p. sax, jama 264 (1990) 41-47. [15] c.h. barrier, b.i. hirschowitz, arthritis rheum 32 (1989) 926-932. [16] t. pincus, m. griffin, american journal of medicine 91 (1991) 209-212. [17] c. sostres, c.j. gargallo, m.t. arroyo, a. ganas, best practice in research and clinical gastrointerology 24 (2010) 121-132. [18] g. thiefin, l. beaugerie, joint bone spine 72 (2005) 286-294. [19] a. pilotto, d. sancarlo, f. addante, c. scarcelli, m. franceschi, surgical oncology 19 (2010) 167-172. [20] r.c. rowe, p.j. sheskey, m.e. quinn. handbook of pharmaceutical excipients. 6 ed. washington, dc, usa: pharmaceutical press and american pharmacists association; 2009. [21] h.l. lai, k. pitt, d.q.m. craig, international journal of pharmaceutics 386 (2010) 178-184. [22] m.a. alhnan, a.w. basit, european journal of pharmaceutical sciences 42 (2011) 30-36. [23] m. song, n. li, s. sun, l.r. tiedt, w. liebenberg, m.m. devilliers, ii farmaco 60 (2005) 261-267. [24] j. malakar, a.k. nayak, chemical engineering research and design 90 (2012) 1834-1846. [25] d. bas, i.h. boyaci, journal of food engineering 78 (2007) 836-845. [26] s. chopra, g.v. patil, s.k. motwani. european journal of pharmaceutics and biopharmaeutics 66 (2007) 73-82. [27] s. kehoe, n. kilcup, d. boyd, material letters 86 (2012) 13-17. [28] r.s. liu, y.j. tang, bioresource technology 101 (2010) 3139-3146. deshmukh and naik admet & dmpk 2(1) (2014) 63-70 70 [29] x. li, j. ouyang, y. xu, m. chen, x.y. song, q. yong. bioresource technology 100 (2009) 3139-3146. [30] n. kettaneh-wold, journal of pharmaceutical and biomedical analysis 9 (1991) 605-610. [31] r. ragonese, m. macka, j. hughes, p. petocz, journal of pharmaceutical and biomedical analysis 27 (2002) 995-1007. [32] s.a. martins, d.a. prazeres, l.p. fonseca, g.a. monteiro, analytical biochemistry 391 (2009) 17-23. [33] s. jose, j.f. fangueiro, j. smitha, t.a. cinu, a.j. chacko, k. premaletha, e.b. souto, colloids and surfaces b:biointerfaces 92 (2012) 175-179. [34] r.k. deshmukh, j.b. naik, journal of pharmaceutical innovation 8 (2013) 276-287. [35] l. al haushey, m.a. bolzinger, c. bordes, j.y. gauvrit, s. briancon, international journal of pharmaceutics 344 (2007) 16-25. [36] s. bolton, c. bon, pharmaceutical statitics practical and clinical application. 5 ed. new york, london: informa healthcare; 2009. [37] s.a. joshi, s.s. chavhan, k.k. sawant, european journal of pharmceutics and biopharmaceutics 76 (2010) 189-99. [38] indian pharmacopoeia commission ghaziabad, india. indian pharmacopoeia. 2010. pp. 562. [39] a.k. nayak, d. pal, j. pradhan, m.s. hasnain, international journal of biological macromolecules 54 (2013) 144-54. ©2014 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ manuscript doi: 10.5599/admet.3.4.209 359 admet & dmpk 3(4) (2015) 359-362; doi: 10.5599/admet.3.4.209 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index note application of the henderson-hasselbalch equation to solubility determination: nsc-639829 case study samuel yalkowsky*, raj patel and doaa al-antary department of pharmaceutical sciences, college of pharmacy, the university of arizona, tucson, arizona, usa *corresponding author: e-mail: yalkowsky@pharmacy.arizona.edu; tel.: +1-520-626-1289; fax: +1-520-626-2466 received: august 11, 2015; revised: october 04, 2015; published: december 30, 2015 __________________________________________________________________________ a number of publications which challenge the applicability of the henderson-hasselbalch equation to saturated solutions have appeared in the last few years (avdeef [1-3], butcher et al. [4], and volgyi et al. [5]). in the most recent of these, butcher et al. [4] suggested “the henderson-hasselbalch equation may not always be an accurate predictor of the ph dependence of solubility.” they claimed that the pka of 4.70 determined by jain et al. [6] for nsc-639829 is incorrect and that the value of 3.76, which they obtained by extrapolation of spectrophotometrically determined pka values in 22, 30, and 41 percent methanol-water solutions, is the correct value. we believe that 4.70 is the correct value and that there are several serious flaws in their analysis. these are described below. prediction schonherr et al. [7] calculated and measured the solubilities of 34 active pharmaceuticals. they found that all solubility-ph profiles showed the expected shape and appearance, which are in good agreement with the values calculated using the henderson-hasselbalch equation. it is clear from the observed data in figure 1 that a pka 4.70, (as reported by jain et al. [6]), for nsc639829 gives good estimates of the observed solubilities and that a pka of 3.76 (as proposed by butcher et al. [4]) will give solubility predictions that are nearly an order of magnitude less than the observed values. in order to justify their pka value butcher et al. have to postulate the existence of a mixed charge dimer (or higher order oligomers) using weighted, nonlinear regression which is then iterated. there is no way to calculate the solubility of nsc-639829 as a function of ph without using the value of 4.70. the postulation of a logk2 of 7.8, so that the logarithmic factor in equation a7 (from ref. [4]) is 0.94, only serves to bring the pka 3.76 up to 4.70, which is identical to value of jain et al. [6]. http://www.pub.iapchem.org/ojs/index.php/admet/index yalkowsky et al. admet & dmpk 3(4) (2015) 359-362 360 figure 1. solubility-ph profile for nsc-639829. (a) points and solid curve from jain et al. [6]. (b) the dashed curve from butcher et al. [4]. extrapolation the pka value proposed by butcher et al. [4] was obtained by extrapolation. while extrapolation is a valuable tool, it must be used with caution. the use of data from 22 41 % methanol to obtain the value in 0 % methanol represents a long extrapolation (roughly 100 % on the basis of methanol concentration or roughly 50 % on the basis of reciprocal dielectric constant). since self-association of nsc-639829 is minimal in methanol and dilute aqueous solution, the predicted pka or butcher et al. [4] is likely that of the monomeric species. pka of monomer and associated forms the pka of any compound is profoundly affected by its environment. the pka of a compound at infinite dilution, or even at a low concentration, is different from its value in an aggregate, whether the aggregate is a dimer, a micelle, or something in between. therefore, even if the pka of the monomer is 3.74, it is of no value for calculating the total drug concentration in a saturated solution. as stated by burns et al. “the pka … is an aggregation dependent value. therefore, a concentration-appropriate pka must be applied when interpreting the results of experiments ….” [8]. it is noteworthy that avdeef et al. [9] indicate that the observed pka of a dimer often differs from that of the monomer. as shown in figure 2, many physical properties change with increasing concentration, especially if there is self-association [10]. in fact, yalkowsky and zografi [11] showed that the pka of the carboxyl group of the decylcarnitine zwitterion (figure 3) is constant and independent of concentration below the critical micelle concentration (cmc), begins to increase at the cmc and asymptotes the micellar value at higher concentrations. the micellar pka is lower because the positive charge resulting from the partially neutralized acid of the micelle repels the hydronium ion. this phenomena has been reported for a number of other systems, including: alkylamine oxides by tokiwa and ohki [12], aliphatic bases by matulis and bloomfield [13], carboxylic acids by kanicky and shah [14], and resveratrol by lopez-nicolas and garciacarmona [15]. it is also responsible for the sharp increase in solubility of prostaglandin observed by roseman and yalkowsky [16]. admet & dmpk 3(4) (2015) 359-362 study of nsc-639829 solubility doi: 10.5599/admet.3.4.209 361 figure 2. physical property curves for sodium lauryl sulfate from preston [10]. figure 3. pka vs. concentration of decylcarnitine from yalkowsky and zografi [11]. conclusions the henderson-hasselbalch equation, as used by jain et al. [6] is an accurate predictor of the ph dependence of solubility, provided that the correct pka is used. the monomer and the aggregated forms of any compound can, and often do, have different values for many physical properties, including pka values. the pka of the monomer should be used for calculations involving dilute solutions, where association is minimal, and pka of the form that exists at saturation should be used for solubility studies, including the study of the solubility of nsc-639829. ch 3 o o n o o ch 3 ch 3 ch 3 + yalkowsky et al. admet & dmpk 3(4) (2015) 359-362 362 references [1] a. avdeef. admet & dmpk 2(1) (2014) 33-42. [2] a. avdeef. admet & dmpk 2(1) (2014) 43-55. [3] a. avdeef. admet & dmpk 3(2) (2015) 84-109. [4] g. butcher, j. comer, a. avdeef. admet & dmpk 3(2) (2015) 131-140. [5] g. völgyi, a.k.j. marosi, a. avdeef. admet & dmpk 1(4) (2013) 48-62. [6] n. jain, g. yang, s.e. tabibi, s.h. yalkowsky. int. j. pharm. 225 (2001) 41-47. [7] d. schönherr, u. wollatz, u. haznar-garbacz, u. hanke, k.j. box, r. taylor, r. ruiz, s. beato, d. becker, w. weitschies. eur. j. pharm. biopharm. 92 (2015) 155-170. [8] d.c. burns, d.a. ellis, h. li, c.j. mcmurdo, e. webster. environ. sci. technol. 42 (2008)9283-9288. [9] a. avdeef, d. voloboy, and a. foremn in comprehensive medicinal chemistry ii (eds. j.b. taylor and d.j. triggle) 5 (2007)399-423. [10] w.c. preston. j. phys. colloid. chem. 52 (1948) 84-97. [11] s.h.yalkowsky, g. zografi. j. pharm. sci. 59 (1970) 798-802. [12] f. tokiwa, k. ohki. j. phys. chem. 71 (1967) 1824-1829. [13] d. matulis, v.a. bloomfield. biophys. chem. 93 (2001) 37-51. [14] j.r. kanicky, d.o. shah. langmuir 19 (2003) 2034-2038. [15] j.m. lopez-nicolas, f.g. garcia-carmona. j. agr. food chem. 56 (2008) 7000-7005. [16] t.j. roseman, s.h. yalkowsky. j. pharm. sci. 62 (1973) 1680-1685. ©2015 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://www.sciencedirect.com/science/article/pii/s0939641115001101 http://creativecommons.org/licenses/by/3.0/ manuscript doi: 10.5599/admet.2.3.22 143 admet & dmpk 2(3) (2014) 143-156; doi: 10.5599/admet.2.3.22 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index review use of voltammetry for in vitro equilibrium and transport studies of ionisable drugs matěj velický 1, *, andrew n.j. rodgers 1 , robert a.w. dryfe 1 and kin tam 2 1 school of chemistry, university of manchester, oxford road, manchester m13 9pl, uk 2 faculty of health science, university of macau, macau, china *corresponding author: e-mail: matej.velicky@googlemail.com; tel.: +44-161-306-2770; fax: +44-161-275-4598 received: november 29, 2013; published: september 16, 2014 abstract in this review, we will briefly outline the voltammetric investigations of the transfer of ionisable drugs at the interface between two immiscible electrolyte solutions. the voltammetric techniques enable the determination of some key in vitro properties of ionisable drugs, including partition coefficient, diffusion coefficient and membrane permeability. some successful applications will be highlighted, together with the background methodologies. keywords cyclic voltammetry; partition coefficient; diffusion coefficient; permeability; ities. 1. introduction the transport properties of a pharmaceutical generally refer to its diffusion coefficient and membrane permeability, whereas the equilibrium state can be described with the use of the partition coefficient. these parameters underpin the physicochemical characteristics of a drug molecule, and provide means to assess the molecule’s transport through membranes and/or lipophilic media. for instance, the determination of the partition coefficient yields its lipophilicity, which represents the affinity of a compound for a lipidic environment. the diffusion coefficient, which is a size dependent parameter, determines how quickly the molecules diffuse within either aqueous or organic environments, while membrane permeability can be described by the rate of permeation of the molecule across an artificial membrane. these parameters are useful for estimation of drug disposition events such as passive absorption and modeling the drug distribution among different body compartments. therefore, the determination of these parameters is of considerable interest in drug discovery and development. over 60 % of marketed drugs are ionisable under physiological conditions [1]. the importance of ionisable forms, and their charge, on the overall transport properties of drugs have traditionally been underestimated mainly due to the lack of reliable methods for determination. the introduction of voltammetric techniques has offered new opportunities for evaluation of the transport properties and understanding of the transport processes of ionisable drugs. in this mini review, voltammetric investigations made on the transfer of ionisable drugs at the interface between two immiscible electrolyte solutions (ities) will be discussed. in particular, cyclic voltammetry has been used extensively to investigate http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:matej.velicky@googlemail.com velický et al. admet & dmpk 2(3) (2014) 143-156 144 the transfer characteristics of charged species. particular emphasis will be placed on the determination of partition coefficients, diffusion coefficients and membrane permeability. 2. determination of partition coefficients the admet (absorption, distribution, metabolism, excretion and toxicity) of a drug is generally governed by both its chemical structure and physiochemical properties, such as lipophilicity, solubility and ionisation. of these physiochemical properties, lipophilicity is widely regarded as the most important parameter to describe the affinity of a drug to a lipophilic phase and is used in both the design of drugs and assessment of their performance [2-4]. the study of drug partition across an interface between immiscible aqueous and organic phases is an established way to characterise the lipophilicity of drugs and the resulting partition coefficient (p) is a common measure of lipophilicity. a partition coefficient is simply the ratio of the activity of a solute in both phases in equilibrium, as expressed in eq. (1), in which and o are the activities of solute a in the aqueous and organic phases, respectively. o (1) due to the wide range of their values, partition coefficients are usually quoted on a logarithmic scale, as log p. shake-flask experiments and potentiometry are traditional methodologies used for determining partition coefficients in the standard water|n-octanol system [5-9]. however, the partition coefficients of ionised fractions of drugs were usually ignored. the increasing interest in partitioning of fully or partially ionised drugs prompted the search for new methodologies to determine partition coefficients [10-12]. one of the first applications of electrochemistry for the determination of partition coefficients was that of kontturi and murtomäki [10], in which cyclic voltammetry at the ities was used. studies at the ities are analogous to electrochemistry at the more common solid electrode|solution interface, with the solid electrode being replaced by a second liquid phase. the application of a potential across the ities, known as the galvani potential (δφ), produces current as a result of charge transfer across the ities. both liquid phases must be rendered conductive through the dissolution of an aqueous and an organic electrolyte, which remain in their respective phases, in order to apply a potential across the ities. conventionally, the galvani potential is defined as the potential difference between the aqueous and organic phase, with the sign of the galvani potential assigned to the aqueous phase, e.g. positive galvani potential corresponding to the aqueous phase being positively polarised with respect to the organic phase. a diagram of an ities setup is shown in fig. 1. as in conventional solid electrode|solution electrochemistry, the current can arise from electron transfer across ities and additionally it can also arise from the transfer of ions across ities [13-17]. this renders the ities ideally suited to the study of ion partitioning across two immiscible phases, which combined with traditional techniques allows for the contribution of ions and neutral compounds to be determined, as shown in the studies of bouchard et al. [18,19], reymond et al. [11], and nagatani et al. [20]. admet & dmpk 2(3) (2014) 143-156 voltammetry in drug transport studies doi: 10.5599/admet.2.3.22 145 figure 1. a schematic of the interface between two immiscible electrolyte solutions (ities), with an organic (green) and an aqueous (blue) phase, set-up for electrochemical measurement. each phase contains a counter and reference electrode (ce and re, respectively), which allow application of potential difference (voltage) between the two liquid phases. partition coefficients are simply a measure of solute’s affinity for either phase of a liquid|liquid system, as are gibbs energies of transfer across the ities, and therefore the two are directly related. this is demonstrated by eq.(2), where pi and pi 0 are the partition coefficient and standard partition coefficient, respectively, of ion i, δφ is the galvani potential, tr,i , o is the standard gibbs energy of transfer for ion i from the aqueous phase to the organic phase, zi is the charge of ion i, f is the faraday constant, r is the universal gas constant and t is the temperature [14,15]. i tr, i , o i i i (2) the gibbs energy of transfer for an ion is itself related to the standard transfer potential of the ion, which can be determined from the half-wave potential of the peak on a cyclic voltammogram (cv) corresponding to the transfer of the ion from one phase to the other. the standard partition coefficient (pi 0 ) is the partition coefficient of ion i when no potential is applied across the ities, while the partition coefficient (pi) is the partition coefficient of ion i at a given galvani potential, as defined by eq. (2). as a result of its poor polarisability, n-octanol is not a suitable organic solvent for ities studies. in the paper of kontturi and murtomäki [10], 1,2-dichloroethane (dce) was chosen as a replacement for noctanol, because of their similar electrical properties, including relative permittivity and dipole moment, and the fact that dce is sufficiently polarisable to allow the transfer of ions across the water|dce interface. the water|dce system has now become the standard for determining partition coefficients of drugs using the ities [20-23]. kontturi and murtomäki [10] determined partition coefficients at aqueous ph ≈ 2, at which the drugs under investigation existed in a protonated form. as mentioned previously, it is not possible to electrochemically determine the partition coefficients of neutral drugs using the ities. in order to circumvent this, kontturi and murtomäki calculated the tr , o for h + , converted it to the standard velický et al. admet & dmpk 2(3) (2014) 143-156 146 transfer potential and then subtracted its contribution from the overall half-wave transfer potential used for calculating the partition coefficient for each drug. therefore the calculated partition coefficients corresponded to the neutral forms of the drugs. n-octanol and dce are not identical solvents and partition coefficients determined using one solvent are not directly comparable with the other. this is problematic, as the water|n-octanol system has long been established as the standard system for the study of drug partitioning and large databases of partition coefficients have already been determined using this system. initially, kontturi and murtomäki plotted logpoct values from the literature against their own logpdce values and found two separate trends, both with good linearity, for proton donors and proton acceptors. bouchard et al. [21] adopted a similar approach, finding separate linear trends for h-bond-donor and non-h-bond-donor drugs, and also introduced the parameter δlogp n oct-dce, which is simply the difference between logpoct and logpdce for neutral drug molecules. the value of δlogp n oct-dce is positive when a drug is more attracted by n-octanol than dce and vice versa. it as found that δlogp n oct-dce was positive for h-bond-donor compounds and negative for nonh-bond-donor compounds, which is explained by both the greater solubility of water in n-octanol and the greater h-bonding capacity of n-octanol, relative to dce. thus δlogp n oct-dce can be used to characterise the h-bond-donor capacity of drugs, although care must be taken during analysis, as drugs which form intramolecular h-bonds acted as non-h-bond-donor drugs in this study. jing et al.[24] were able to sufficiently polarise a water|n-octanol nano-interface, supported at the tip of a nano-pipette, and determine partition coefficient for the anionic form of lauric acid. however, as the authors themselves acknowledge, the potential window obtained was only 400 mv, which is quite restrictive for investigation of a wide range of partition coefficients. gulaboski et al.[25] developed an interesting new methodology for studying the water|n-octanol interface electrochemically. a droplet of n-octanol containing an electrochemically active compound, usually decamethylferrocene (dmfc), is attached to a graphite electrode and immersed in an electrolytic aqueous solution, which contains an anion of interest. the dmfc is oxidised to dmfc + at the three-phase boundary between water|graphite|octanol and the anion the transfers across the liquid|liquid interface and into the n-octanol droplet in order to compensate the charge and maintain electroneutrality, as illustrated in fig. 2. this method requires that the gibbs energy of transfer of dmfc + from n-octanol to water is larger than that of the anion from water to n-octanol. once the standard redox potential of dmfc/dmfc + in n-octanol (which was calculated using extrapolated standard gibbs energy of transfer values of various halide anions from other n-aliphatic solvents) was known, the potential of the peak on the cv caused by the reaction was related to the standard transfer potential of the anion and thus the partition coefficient of the anion could be calculated. the methodology of gulaboski et al. was further tested by bouchard et al. [26], in a study which looked at 26 different anions. a good linearity was found between partition coefficients determined in their study and those previously determined using the water|dce system. when using the dmfc/dmfc + redox couple, the three-phase system is only suitable for determining the partition of anions across the water|n-octanol interface. quentel et al. [27] sought to address this by using lutetium bis(tetra-tert-butylphthalocyaninato) (lu[tbu4pc]2), which can be oxidised or reduced to the cation lu[tbu4pc]2 + or the anion lu[tbu4pc]2 , respectively, therefore allowing the study of both anion and cation transfer using the same redox species. both lu[tbu4pc]2 + and lu[tbu4pc]2 are strongly lipophilic and therefore suited for this method. the tr, i , -octanol was calculated for several alkali metal and organic cations and whilst these cations are not pharmaceutically relevant, the principle of using the three-phase admet & dmpk 2(3) (2014) 143-156 voltammetry in drug transport studies doi: 10.5599/admet.2.3.22 147 system for cation transfer was demonstrated well. the entire methodology is an elegant bridge between the water|n-octanol system and electrochemical methods for partition coefficient determination. figure 2. schematic showing the oxidation of decamethylferrocene (dmfc) in an n-octanol droplet supported on a graphite electrode immersed in an aqueous solution, followed by transfer of an anion x across the liquid|liquid interface in order to maintain electroneutrality. whilst the partitioning of neutral compounds is not dependent on either potential or ph, the partitioning of ions is dependent on both. with ph being an important condition in vivo, it is useful to have an understanding of how both potential and ph affect the partitioning of ions. reymond et al. [28] developed ionic partition diagrams for this very purpose. such diagrams are plots of galvani potential against ph, similar to the pourbaix diagrams used in corrosion science, and act as a visual representation of which ions predominate, and in which phase, at a given potential and ph. in addition, they give a description of the mechanisms via which the ions transfer between the two phases, be it simple iontransfer, or a proton-coupled transfer, as either potential or ph is varied [19, 29]. an ionic partition diagram for the zwitterionic drug cetirizine, from the study of bouchard et al. [19] is shown in fig. 3, with a more detailed diagram showing mechanisms of phase transfer over a smaller range of potentials and ph, shown in fig. 4. the theoretical lines in fig. 3 represent regions in which the concentrations of the species on either side of the line are equal. the horizontal lines are defined by the standard transfer potentials of the species xh3 2+ (w), xh2 + (w) and x (o), where x = cetirizine. the vertical lines are defined by the three pka values of cetirizine, 2.12, 2.90 and 7.98. the lines of finite gradient represent the transfer of species across the water|dce interface, which are coupled with a proton exchange, as can be seen from fig. 3. ionic-partition diagrams represent the culmination of electrochemical studies of ion-partitioning to this date and provide a valuable first port of call when wishing to understand the partitioning behaviour of an ionised drug under particular conditions of potential and ph. whilst there is still work to be done in bridging the water|n-octanol and water|dce systems, the contrasting information provided by these systems has actually proved useful in characterising the properties of drugs. in this respect, electrochemical techniques for studying ion partitioning can be viewed as complimentary to traditional techniques, rather than as a replacement for them, and combining the two has resulted in a large expansion in the information which can be gleaned from ion partition studies. velický et al. admet & dmpk 2(3) (2014) 143-156 148 figure 3. theoretical ionic partition diagram of cetirizine (x) based on analysis in [19]. figure 4. detailed ionic partition diagram of cetirizine (x), showing the mechanisms of phase transfer (based on analysis in [19]). 3. determination of diffusion coefficients in addition to the determination of partition coefficients, electrochemistry at the ities is ideally suited for measurement of the diffusion coefficients of ions. as charge transfer processes across the ities are diffusion controlled, the same equations that govern diffusion-controlled charge transfer at the solid electrode|solution interface also apply to the ities, namely the randles-ševčík equation (3) and the cottrell equation (4) [13-15]. admet & dmpk 2(3) (2014) 143-156 voltammetry in drug transport studies doi: 10.5599/admet.2.3.22 149 2/1 i 1/2 ii 2/1 i p 4463.0 cfadz rt fz i        (3) t dfacz i  iii  (4) where ip is the peak current observed from the voltammogram, zi is the charge of the ion i, f = 96 485 c mol 1 is the faraday constant, r = 8.31 j mol 1 k 1 is the universal gas constant, t is the temperature, a is the ities area, di is the diffusion coefficient of the ion i, ci is the concentration of the ion i, v is the scan rate, i is the time-dependent current and t is time. when determining the diffusion coefficient from the randlesševčík equation, cyclic voltammetry is used and the scan rate is varied. a plot of ip against v 1/2 then yields a linear plot, the slope of which can be used to calculate the diffusion coefficient of ion i. chronoamperometry is the technique employed when using the cottrell equation for diffusion coefficient determination. the galvani potential is stepped from a value at which the ion of interest does not transfer across the ities, to one at which it does. the second potential is then held for a predetermined amount of time and the diffusion controlled current decay is recorded. a plot of i against t 1/2 will then yield a linear plot, from which the diffusion coefficient of ion i can be calculated. although diffusion coefficients are important drug transport parameters, the number of studies determining their values using electrochemistry at the ities are negligible compared to partition coefficient studies. kontturi and murtomaki [10] did report diffusion coefficients in their original study, but stated that the determined values may not have been accurate due to a difference in drug concentration between the bulk and the liquid|liquid interface. since then, the diffusion coefficients of drugs have largely been ignored in work at the ities. no explanation was given as to why such a concentration difference may have existed, but it proved to be of critical importance in a later study by velický et al. [30]. in the ork of velický et al. [30], warfarin and propranolol were studied at ph 7.4, at which both drugs are almost entirely ionised (0.26% and 0.74% neutral form of warfarin and propranolol, respectively). using the ities setup described earlier, diffusion coefficients, which were 14and 46-fold lower than predicted by molar mass [31], were determined for warfarin and propranolol, respectively. it was observed that the magnitude of the current produced by ion-transfer, which is used to calculate the diffusion coefficient of the ion, gradually decreased over time after initial set-up of the experiment and without an applied potential. this was explained, and confirmed by shake-flask experiments, by a partitioning of the neutral fraction of the drug across the interface. the concentration of the neutral fraction is constantly replenished through the aqueous dissociation equilibrium (illustrated for warfarin), w (aq) + h + (aq) hw(aq) where w = warfarin. provided that the concentration of protons in the aqueous phase is sufficient, the concentration of the ionic form of the drug at the interface gradually depletes, resulting in lowered iontransfer current and thus a lower calculated diffusion coefficient than would be expected for the bulk concentration of drug. velický et al. admet & dmpk 2(3) (2014) 143-156 150 stirring of the aqueous phase was devised as a solution to this problem, therefore minimizing the effect of diffusive mass transport and maintaining a constant concentration of the ionic form of the drug at the interface. stirring was made possible by separation of the two liquid phases using a polyvinylidene fluoride (pvdf) membrane. a diagram of the experimental setup used in this study is shown in fig. 5. diffusion coefficients of warfarin and propranolol determined via this method were on the order of magnitude expected from their molar masses [31]. this setup offered a solution to a problem almost conspicuous in its absence from the literature, given the expected ease with which diffusion coefficients could be calculated in tandem with partition coefficients at the ities. figure 5. the schematic of the stirred ities setup used in [30], in which the two immiscible phases are separated by a rotating membrane. 1) pvdf membrane, 2) glass tube (aqueous phase), 3) glass cell (organic phase dce), 4) reference phase (aqueous), 5) counter electrodes (pt), 6) reference electrodes (ag/agcl), 7) rotation control. 4. voltammetry in membrane permeability studies traditional bio-relevant membrane mimics such as black lipid membranes (blm) and supported lipid membranes are invaluable substrates for studies of drug-membrane interaction and functions of biological membranes [32-34]. however, they are of a fragile nature, which inevitably limits their use. supported liquid membranes (slm) on the other hand, which typically contain a lipid solution of an organic solvent immobilised on a polymer membrane, are very stable and have intrinsic durability and are therefore suitable for use in the industrial setting and even large scale applications. slms were successfully applied in transfer studies of drug molecules, namely in prediction of drug absorption in humans using the artificial membrane permeability method parallel artificial membrane permeation assay (pampa) [35-37], cellmonolayer permeability assays such as human epithelial colorectal adenocarcinoma (caco-2)[38-40] or admet & dmpk 2(3) (2014) 143-156 voltammetry in drug transport studies doi: 10.5599/admet.2.3.22 151 madin—darby canine kidney (mdck) [41], drug lipophilicity determination [37,42,43] and extraction/removal of drugs [44,45]. polarisation of slms has attracted significant interest in the last few decades [46]. thompson et al. have reported one of the first electrochemical characterisations of drug transfer (amphotericin b and valinomycin) across various polymer filter membranes soaked in organic solvent and lipids and proposed the formation of micro-blms within the membrane [47]. the most common electrochemical methods employed are cyclic voltammetry and chronoamperometry, with applications ranging from drug partitioning studies and drug extraction to concentration of analytes in order to increase analytical sensitivity. fig. 6 shows a typical cyclic voltammogram of the transfer of ions obtained in a slm system. solvents traditionally used in the pharmaceutical industry, due to their good bio-mimetic properties, are not usually suitable for quantitative electrochemical experiments in slm, due to their low polarity (noctanol, as mentioned in section 2) or miscibility with water, problematic in thin (10-100 µm) membranes (dce). 1,2-dichlorobenzene (odcb), nitrobenzene (nb), nitrophenyl octyl ether (npoe) or 1,9-decadiene have proven to be good alternative solvents [48]. figure 6. cyclic voltammograms of transfer of tetraalkylammonium ions across npoe-based membrane at scan rates of 40 mv s 1 . the ions are present on both sides of the membrane hence the double-peak transfer is observed (adapted from [48]. electrochemistry in slm systems is usually performed using a four-electrode setup, similar to that for a two-phase ities system described in section 2 (fig.1). a pair of reference and counter electrodes placed in each aqueous phase, which are separated by the immiscible organic membrane, as shown in fig. 7. liquid|liquid extraction of drugs seems to be one of the most promising applications of slms in the pharmaceutical context. pedersen-bjergaard et al. have introduced a method of electro-kinetic migration of drugs across an npoe-based membrane. initially, a very large voltage of 300 v dc was applied and chemical tuning of the membrane composition used to extract both polar and non-polar drug molecules [49]. gradually, the group improved the method by decreasing the voltage to values as low as 1 v dc [50] and developing an on-chip device for online uv-vis and hplc detection of electrically extracted drugs [51]. while this method is suitable for quantitative extraction of molecules across the membrane and promising velický et al. admet & dmpk 2(3) (2014) 143-156 152 for quick analysis of complex samples, such as urine or blood [52-54], the relatively high applied voltage and absence of supporting electrolyte in the membrane present a problem for theoretical explanation of the ionised drug transfer. figure 7. schematic of a typical supported liquid membrane system. voltage (or potential difference) across the liquid membrane is applied via a pair of reference and counter electrodes (denoted r and c, respectively). the fundamental understanding of electrochemically induced transfer across slms stems from electrochemistry at ities, as discussed in previous sections, and the connection between the two inspired researchers to extend the know-how to systems with two liquid|liquid interfaces, where a small voltage (usually below 500 mv dc) is applied. samec et al. studied transfer of protonated local anesthetics across an npoe-based slm using voltammetry and correlated their pharmacological potency data [55,56] with the voltage required to transfer the drugs across the membrane, as shown in fig. 8 [57]. ulmeanu et al. used cyclic voltammetry and ph profiling to determine the partition coefficient of drugs using a commercial 96well micro-filter plate system, which by nature requires only a small amount of the organic solvent and studied drug [58]. other reports made significant contribution to the polarised membrane research, albeit not directly studying pharmaceutically relevant molecules [59-63]. significant efforts have also been spent on development of an electrochemical theory to describe the observed current-potential dependence, both for systems with and without membrane electrolytes [64-67]. furthermore, a number of research activities have focused on use of slms as ion-selective electrodes applied to drug molecules [68]. the permeability of lipophilic drugs can be limited by transport across the aqueous phase boundary layer. it has been shown that controlled hydrodynamics or stirring of the membrane improved the predictive power of in vitro permeability assays, thus mimicking the diffusion/convection conditions in a human intestine [69,70]. the traditional view of the widely accepted ph-partition hypothesis has also been challenged using a hydrodynamic approach to permeability [71]. it is interesting to note that several research groups independently pursued studies on the electrochemical response of slm systems under hydrodynamic control. manzanares et al. contributed an excellent research paper on determination of ion transfer kinetics in a rotated diffusion cell, i.e. with controlled rotation of the membrane phases, as well as analytical description of this system [72]. this ork as follo ed by murtomäki et al., who reported the effect of slm composition on transfer of tetraalkylammonium and tacrine cations in the same system [73]. most recently, permeation of an organic dye, crystal violet, was studied via uv-vis spectrophotometry and admet & dmpk 2(3) (2014) 143-156 voltammetry in drug transport studies doi: 10.5599/admet.2.3.22 153 electrochemistry, with the transfer mechanism via naked ions and ion pairs across the membrane elucidated [74]. figure 8. logarithm of the drug potency relative to procaine vs. the standard ion transfer potentials of the protonated anesthetics: (□) potency taken from ref. [54], (■) expressed as the inverse ratio of the inhibition constants ic50 [55]. there is yet to be a report to be published, in which all the above elements, i.e. membrane permeability, electrochemistry and hydrodynamic control, have been combined to study drug molecules and applied in a pharmaceutical setting. the strength of interdisciplinary collaboration has been proven in many research fields and so this opportunity, which could provide in-depth insight to drug-membrane interaction and significant improvement in prediction of drug in vivo behaviour, will hopefully not be overlooked. 5. conclusions in this mini review, we have shown the usefulness of ion-transfer voltammetry in studies of the distribution of ionisable drugs at the interface between two immiscible electrolyte solutions and related systems. successful examples, particularly the determination of partition coefficients, diffusion coefficients and membrane permeability have been highlighted. it has been noted that the aforementioned voltammetric techniques coupled with controlled hydrodynamics or stirring of the membrane improve the predictive power of in vitro permeability assays, thus mimicking the diffusion/convection conditions in a human intestine. this novel approach offers a new way to study the passive transport of pharmaceutical compounds across biological membranes, which may facilitate the development of screening methods in pharmaceutical research. acknowledgements: mv and anjr thank the epsrc for funding. velický et al. admet & dmpk 2(3) (2014) 143-156 154 references [1] j. comer and k. y. tam, in pharmacokinetic optimization in drug research: biological, physicochemical and computational strategies (b. testa, h. van de waterbeemd, g. folkers, and r. guy, eds.), verlag helvetica chimica cta, zürich, 2 1, p. 275. [2] c. a. lipinski, f. lombardo, b. w. dominy, and p. j. feeney, advanced drug delivery reviews 46 (2001) 3-26. [3] d. a. smith and h. van de waterbeemd, current opinion in chemical biology 3 (1999) 373-378. [4] b. testa, p. a. carrupt, p. gaillard, f. billois, and p. weber, pharmaceutical research 13 (1996) 335-343. [5] j. c. dearden and g. m. bresnen, quantitative structure-activity relationships 7 (1988) 133-144. [6] s. d. krämer, j. c. gautier, and p. saudemon, pharmaceutical research 15 (1998) 1310-1313. [7] a. leo, c. hansch, and d. elkins, chemical reviews 71 (1971) 525-616. [8] . mälkiä, l. murtomäki, . urtti, and k. kontturi, european journal of pharmaceutical sciences 23 (2004) 13-47. [9] k. takacs-novak and a. avdeef, journal of pharmaceutical and biomedical analysis 14 (1996) 14051413. [10] k. kontturi and l. murtomaki, journal of pharmaceutical sciences 81 (1992) 970-975. [11] f. reymond, p. a. carrupt, b. testa, and h. h. girault, chemistry a european journal 5 (1999) 39-47. [12] z. samec, j. langmaier, . trojánek, e. samcová, and j. málek, analytical sciences 14 (1998) 35-41. [13] j. koryta, electrochim acta 24 (1979) 293-300. [14] f. reymond, d. fermin, h. j. lee, and h. h. girault, electrochimica acta 45 (2000) 2647-2662. [15] z. samec, pure and applied chemistry 76 (2004) 2147-2180. [16] p. vanýsek, electrochimica acta 40 (1995) 2841-2847. [17] p. vanýsek and l. b. ramírez, journal of the chilean chemical society 53 (2008) 1455-1463. [18] g. bouchard, a. pagliara, p. a. carrupt, b. testa, v. gobry, and h. h. girault, pharmaceutical research 19 (2002) 1150-1159. [19] g. bouchard, a. pagliara, g. p. van balen, p. a. carrupt, b. testa, v. gobry, h. h. girault, g. caron, g. ermondi, and r. fruttero, helvetica chimica acta 84 (2001) 375-387. [20] h. nagatani, s. suzuki, d. j. fermín, h. h. girault, and k. nakatani, analytical and bioanalytical chemistry 386 (2006) 633-638. [21] g. bouchard, p. a. carrupt, b. testa, v. gobry, and h. h. girault, chemistry a european journal 8 (2002) 3478-3484. [22] s. fantini, j. clohessy, k. gorgy, f. fusalba, c. johans, k. kontturi, and v. j. cunnane, european journal of pharmaceutical sciences 18 (2003) 251-257. [23] g. bouchard, p. a. carrupt, b. testa, v. gobry, and h. h. girault, pharmaceutical research 18 (2001) 702-708. [24] p. jing, m. zhang, h. hu, x. xu, z. liang, b. li, l. shen, s. xie, c. m. pereira, and y. shao, angewandte chemie international edition 45 (2006) 6861-6864. [25] r. gulaboski, v. mirčeski, and f. scholz, electrochemistry communications 4 (2002) 277-283. [26] g. bouchard, a. galland, p. a. carrupt, r. gulaboski, v. mirceski, f. scholz, and h. h. girault, physical chemistry chemical physics 5 (2003) 3748-3751. [27] f. quentel, v. mirčeski, and m. l'her, journal of solid state electrochemistry 12 (2008) 31-39. [28] f. reymond, g. steyaert, p. a. carrupt, b. testa, and h. girault, journal of the american chemical society 118 (1996) 11951-11957. [29] f. reymond, v. chopineaux-courtois, g. steyaert, g. bouchard, p. a. carrupt, b. testa, and h. h. girault, journal of electroanalytical chemistry 462 (1999) 235-250. [30] m. velický, k. y. tam, and r. . w. dryfe, journal of electroanalytical chemistry 683 (2012) 94-102. [31] a. avdeef, expert opinion on drug metabolism & toxicology. 1 (2005) 325-342. admet & dmpk 2(3) (2014) 143-156 voltammetry in drug transport studies doi: 10.5599/admet.2.3.22 155 [32] . mälkiä, phd thesis, helsinki university of technology (2004) p.84. [33] r. p. richter, j. l. k. him, and a. brisson, materials today 6 (2003) 32-37. [34] g. van meer, d. r. voelker, and g. w. feigenson, nature reviews molecular cell biology 9 (2008) 112124. [35] m. bermejo, a. avdeef, a. ruiz, r. nalda, j. a. ruell, o. tsinman, i. gonzalez, c. fernandez, g. sanchez, t. m. garrigues, and v. merino, european journal of pharmaceutical sciences 21 (2004) 429-441. [36] k. sugano, h. hamada, m. machida, h. ushio, k. saitoh, and k. terada, international journal of pharmaceutics 228 (2001) 181-188. [37] f. wohnsland and b. faller, journal of medicinal chemistry 44 (2001) 923-930. [38] p. artursson and j. karlsson, biochemical and biophysical research communications 175 (1991) 880885. [39] a. avdeef and k. y. tam, journal of medicinal chemistry 53 (2010) 3566-3584. [40] v. fade, journal of pharmaceutical sciences 87 (1998) 1604-1607. [41] j. d. irvine, l. takahashi, k. lockhart, j. cheong, j. w. tolan, h. e. selick, and j. r. grove, journal of pharmaceutical sciences 88 (1999) 28-33. [42] b. faller, h. p. grimm, f. loeuillet-ritzler, s. arnold, and x. briand, journal of medicinal chemistry 48 (2005) 2571-2576. [43] m. h. abraham, h. s. chadha, r. a. e. leitao, r. c. mitchell, w. j. lambert, r. kaliszan, a. nasal, and p. haber, journal of chromatography a 766 (1997) 35-47. [44] z. lazarova, b. syska, and k. schügerl, journal of membrane science 202 (2002) 151-164. [45] y. yamini, c. t. reimann, . vatanara, and j. . jönsson, journal of chromatography a 1124 (2006) 5767. [46] k. kontturi, l. murtomäki, and j. . manzanares, ionic transport processes: in electrochemistry and membrane science, oxford, 2008. [47] m. thompson, r. b. lennox, and r. a. mcclelland, analytical chemistry 54 (1982) 76-81. [48] m. velický, k. y. tam, and r. . w. dryfe, analytical chemistry 86 (2014) 435-442. [49] a. gjelstad, k. e. rasmussen, and s. pedersen-bjergaard, journal of chromatography a 1124 (2006) 2934. [50] i. j. ø. kjelsen, . gjelstad, k. e. rasmussen, and s. pedersen-bjergaard, journal of chromatography a 1180 (2008) 1-9. [51] n. j. petersen, s. t. foss, h. jensen, s. h. hansen, c. skonberg, d. snakenborg, j. p. kutter, and s. pedersen-bjergaard, analytical chemistry 83 (2011) 44-51. [52] m. eskandari, y. yamini, l. fotouhi, and s. seidi, journal of pharmaceutical and biomedical analysis 54 (2011) 1173-1179. [53] s. nojavan and a. r. fakhari, journal of separation science 33 (2010) 3231-3238. [54] k. f. seip, j. stigsson, a. gjelstad, m. balchen, and s. pedersen-bjergaard, journal of separation science 34 (2011) 3410-3417. [55] l. m. hondeghem and r. d. miller, in basic and clinical pharmacology (b. g. katzung, ed.), prentice-hall int. inc., new york, 1992, p. 363. [56] e. t. mcneal, g. a. lewandowski, j. w. daly, and c. r. creveling, journal of medicinal chemistry 28 (1985) 381-388. [57] z. samec, . trojánek, j. langmaier, e. samcová, and j. málek, electroanalysis 12 (2000) 901-904. [58] s. m. ulmeanu, h. jensen, g. bouchard, p. a. carrupt, and h. h. girault, pharmaceutical research 20 (2003) 1317-1322. [59] m. h. barker, l. murtomäki, and k. kontturi, journal of electroanalytical chemistry 497 (2001) 61-68. [60] j. langmaier, k. stejskalová, and z. samec, journal of electroanalytical chemistry 496 (2001) 143-147. velický et al. admet & dmpk 2(3) (2014) 143-156 156 [61] s. m. ulmeanu, h. jensen, z. samec, g. bouchard, p. a. carrupt, and h. h. girault, journal of electroanalytical chemistry 530 (2002) 10-15. [62] j. langmaier and z. samec, electrochemistry communications 9 (2007) 2633-2638. [63] o. shirai, s. kihara, y. yoshida, and m. matsui, journal of electroanalytical chemistry 389 (1995) 61-70. [64] t. kakiuchi, electrochimica acta 44 (1998) 171-179. [65] . molina, j. . ortuño, c. serna, e. torralba, and j. gonzalez, electroanalysis 22 (2009) 1634-1642. [66] . molina, c. serna, j. gonzalez, j. . ortuño, and e. torralba, physical chemistry chemical physics 11 (2009) 1159-1166. [67] z. samec, . trojánek, j. langmaier, and e. samcová, journal of electroanalytical chemistry 481 (2000) 1-6. [68] j. guo and s. amemiya, analytical chemistry 78 (2006) 6893-6902. [69] a. avdeef, p. e. nielsen, and o. tsinman, european journal of pharmaceutical sciences 22 (2004) 365374. [70] m. velický, k. y. tam, and r. . w. dryfe, european journal of pharmaceutical sciences 44 (2011) 299309. [71] m. velický, d. f. bradley, k. y. tam, and r. a. w. dryfe, pharmaceutical research 27 (2010) 1644-1658. [72] j. a. manzanares, r. lahtinen, b. quinn, k. kontturi, and d. j. schiffrin, electrochimica acta 44 (1998) 59-71. [73] l. murtomäki, m. h. barker, j. . manzanares, and k. kontturi, journal of electroanalytical chemistry 560 (2003) 95-103. [74] m. velický, k. y. tam, and r. . w. dryfe, analytical chemistry 84 (2012) 2541-2547. ©2014 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ manuscript doi: 10.5599/admet.3.3.206 260 admet & dmpk 3(3) (2015) 260-273; doi: 10.5599/admet.3.3.206 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index review the impact of l2 learning on cognitive aging kaiwen cheng 1,2 ; yanhui deng 3 ; ming li 2 ; hong mei yan 1 * 1 key laboratory for neuroinformation of ministry of education, school of life science and technology, university of electronic science and technology of china, chengdu 610054, china 2 school of foreign languages, southwest jiaotong university, chengdu 611756, china 3 school of foreign languages, chengdu technological university, chengdu 611730, china * corresponding author: e-mail: hmyan@uestc.edu.cn; kevin.cheng78@163.com 1 these authors contributed equally to this work. received: august 07, 2015; revised: august 30, 2015; published: september 05, 2015 abstract it has become a multidisciplinary research area to overcome cognitive decline caused by aging. many factors can affect cognitive aging and the influence of second language learning (l2 learning) cannot be ignored. the recent decade has witnessed much pathological, behavior and neuroimaging research that l2 experience may help maintain the cognitive function in the elderly, resist cognitive decline, and even delay the onset of alzheimer's disease (ad). this work is to review available literature concerned and elucidate the neural mechanisms under which l2 learning (training) may modify or sculpt the brain from perspectives of cognitive reserve, plasticity and overlapping networks. future directions concerning length of learning, frequency of use, comparison with other cognitively stimulating activities are put forward so as to clarify the relationship between language experience and cognitive aging. keywords cognitive decline; l2 learning; cognitive reserve (cr); neuroplasticity; overlapping network introduction aging is increasingly serious with the rapid growth of the global population. according to a recent un report, people over 60 are the world’s fastest growing age group, accounting for 700 million so far, which will increase to 2 billion by 2050 [1]. cognitive aging has become an emergency as cases of neurodegenerative diseases, alzheimer's diseases (ad) in particular, increase considerably. currently, the prevalence of alzheimer's diseases is about 5% among those aged over 60, at the speed of which a conservative estimate is that over one hundred million old people will have been diagnosed with ad by 2050 [2]. caring for those who cannot provide for themselves has become a major burden of families and the society. therefore, in the next few decades, it is vital to promote successful cognitive aging for both individuals and public health institutions and overcoming cognitive decline due to aging is becoming a multidisciplinary research field and one of the biggest challenges we face. apart from age, many factors affect cognitive aging, including heredity, diseases, education, occupation, lifestyle, etc [3]. it is remarkable that language experience has been found to relate to cognitive advantage. over half a century, psycholinguists and neuroscientists consent at the benefits of l2 acquisition, rendering http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:hmyan@uestc.edu.cn mailto:kevin.cheng78@163.com admet & dmpk 3(3) (2015) 260-273 l2 learning and cognitive aging doi: 10.5599/admet.3.3.206 261 plausible explanations based on language processing and cognitive plasticity, which both enrich theories of bilingualism and shed a deeper insight into the relationship between language and cognition: language not only reflects the mind, but also shapes the brain to a certain degree [4]. most notable in recent years are several intriguing reports that bilingualism can significantly delay the onset of alzheimer's diseases. it provides new train of thought for both pathological research and clinical treatments of cognitive degeneration [5-10]. in view of these, our review is aimed to explore the impact of l2 learning on cognitive aging and possible mechanisms, and also present challenges as well as directions for future research. connotation and categories of cognitive aging under the framework of biology, cognitive aging reflects that the body’s protective function is no longer sufficient to offset a wide variety of pressure sources [11], the process involving damage of oxidative stress, loss of free radical detoxification, decline of mitochondrial function, accumulation of potentially harmful proteins, any one of which can lead to decreased neuronal membrane integrity, altered normal metabolic functions, even the death of neurons [12]. on the pattern of aging process, views of the past was that all the components of nervous system showed a similar degree of decline with the growth of age [13], while current researchers argue that some specific regions and networks are more vulnerable than others, such as gray matter in the prefrontal cortex (resulting in impaired executive function) or subcortical white matter (resulting in slow information processing speed), hippocampus (resulting in decline in memory) [14]. generally, cognitive aging is categorized into normal aging and pathological aging. normal aging, also known as “non-pathological aging” or “successful cognitive aging”, entails predictable changes to cognitive function along with the increase of age. that is, no symptoms of the diseases can be identified to have negative effects on the central nervous system (such as ad, cerebrovascular disease). otherwise, it is viewed as “pathological aging”. however, many people insist that aging process is a continuum with no clear boundary between the two categories. some even argue that dementia is still a part of normal aging due to the fact that quite a lot of old people aged over 85 would have clinical symptoms of dementia [15] and a large proportion of old people who are cognitively normal also manifest some degree of pathological phenomena (especially the amyloid protein or neurons winding) and cerebrovascular diseases. research has shown that mild cognitive impairment caused by the ad may start as early as ten years before clinical onset [16], which means that many senior people who seem normal have actually been accumulating pathological phenomena. so it is very difficult to decide whether cognitive decline is caused by diseases or the normal aging. however, it is well accepted that typical manifestations may be memory loss and decline of executive control, poor working memory and slow information processing. of course language and communication are also affected, including bad speech perception and production, reduced semantic comprehension in natural language environment, poor vocabulary memory, among others [17]. advances on the impact of l2 learning on cognitive aging early studies last century mainly concerned cognitive effects of bilingual experiences of children within so-called “critical period”, rather than young adults and the aged. not until the beginning of this century did scholars begin to pay close attention to the influence of bilingual experiences on cognitive aging. the main research wisdom is to determine the differences of cognitive function between the older bilinguals and monolinguals by comparing their performances on various behavior tasks (such as simon task, flanker task, stroop tasks, etc.). recently, with the aid of electrophysiological and neuroimaging technologies, such as event-related potentials(erp), functional magnetic resonance imaging (fmri), positron emission k. cheng et al. admet & dmpk 3(3) (2015) 260-273 262 tomography (pet), researchers can obtain more accurate data about brain function and structure in comparison to behavioral tasks, allowing deeper probe into the relationship between l2 learning and cognitive aging. below we will retrospect on current research progress along two channels: normal aging and pathological aging. bilingualism contributing to better cognitive ability in healthy elderly adults although a few studies concerning children and young adults have found weaknesses of speaking more than one language, for example, slower reaction time and higher error rate on the picture naming tasks, poor fluency on either language and so on, more research has shown that bilinguals do excel in learning strategies, problem solving, conflict resolution, attention regulation, executive control in comparison to their counterparts [18]. more importantly, positive effects are assumed to extend to the elderly, helping ward off age-related decline of the cognitive levels [19-22]. for example, by comparing the performances of the middle-aged and old bilinguals with those of their monolingual counterparts in simon task and the visual stroop task (executive control), bialystok et al. found bilingual advantage in both groups, especially in old age group [19,23]. second, through working memory span tasks (spatial wm) which are also known to be closely related to executive control, luo et al., found the bilinguals outperformed the monolinguals in spatial wm rather than in verbal wm, although there was no interaction effect of bilingual experience and aging. they explained that this might be because bilingual deficit in verbal process offsets the positive impact on working memory as a function of their superior executive control [24]. notably, kave et al. (2008), with a follow-up study of 814 healthy old man over a period of 13 years, found that the number of languages they spoke could predict the performance on various cognitive tests after controlling for some demographic variables (such as age, gender, place of birth, age and education level). they proposed for the first time that the more languages old people spoke on regular basis, the higher their cognitive levels would be [21]. it is known as the cumulative effect of languages, which is consistent with the results of luxembourg’s study that mastering one more language would increase the probability of protection against cognitive impairment by more than four times (of course, a threshold existed) and the earlier they became multilingual, the greater the protection would be [25]. it is a pity that this research didn’t involve monolinguals, which makes impossible the comparison between monlolinguals and bilinguals or multilinguals (people speaking more than two languages). recently, bak and his colleagues from university of edinburgh obtained encouraging results with lothian birth cohort (about 1100 irish born in 1936 in and near edinburgh) [26]. although this study was not meant to explore the influence of language, cognitive states were tracked toward the age of 70 for 853 of the participants, all of whom coincidentally spoke only english before the age of 11 and almost a third or 262 people, had learned to speak another one language ever since. this provides the absolute homogeneous sample for the longitudinal study on cognitive effects of foreign languages (the interference of the original intelligence ruled out). after conducting a series of cognitive tests for participants including intelligence test, and comparing the results with their own test scores at the age of 11, the researchers found that those who mastered two or more languages had significantly better cognitive ability, especially in terms of general intelligence and reading, irrespective of when to learn the second language. it may be inferred that the aging brain of adult learners will also benefit from speaking a second language even if their l2 cannot achieve the proficiency of their mother tongue, which turns out to be a great incentive for millions of adults around the world who are trying to learn a second language. admet & dmpk 3(3) (2015) 260-273 l2 learning and cognitive aging doi: 10.5599/admet.3.3.206 263 bilingualism protecting against the onset of ad at the end of last century, graves and his colleagues published a study regarding the influence of cultural factors on the pathological process of a group of older japanese americans [27]. after controlling for many variables, such as immigration history, years of formal education in japan, religion, diet, the proficiency of reading and writing in japanese and so on, they found that comprehensive level of japanese could serve as a most powerful indicator to predict the lower risk of cognitive decline among all participants]. this raised an interesting possibility that mastering one more language may affect cognitive decline [28]. the connection between language ability and aging echoed with the finding of another autopsy study by snowdon et al. that language ability in youth (english self-introduction at the age of 1927) was in significantly negative correlation with its ad pathology (age spots or nerve filament winding) in some cerebral areas (such as frontal lobe, temporal lobe and parietal lobe) among 74 elderly (ages 78-92), although their language experiences (bilingual or not) were mentioned in the study [28]. the breakthrough must be attributable to canadian scientists, bialystok, craik and freeman, who for the first time confirmed that l2 learning experience had significantly positive influence on pathological aging by analysing medical records of 184 old patients diagnosed with various types of dementia (70 % ad) in a toronto hospital. after interviewing patients and families, eliminating various factors which may affect the statistical results (age, gender, education, economy, immigration status, etc.), they were surprised to find that the average age of bilinguals diagnosed with ad were four years later than monolingual patients (71.4 years for monolingual while 75.5 years for bilinguals) [5]. once published in the well-known journal neuropsychologia, it immediately aroused great interest from scholars and language learners throughout the world. in 2010, with a larger sample of 211 patients (100 % ad), the same research team not only sustained their earlier conclusion but found those with bilingual or multilingual experience had onset time delayed by up to 5 years, after controlling for language level, occupation, gender, immigration history and other factors [6]. another group in canada, with an even larger cohort of 632 people in montreal (90% monolingual or multilingual immigrants), found that their results still mirrored those of bialystok’s team, especially in the immigrant population [7]. shortly afterwards, different research teams in different regions or countries replicated the results of bialystok et al. in 2011, gollan et al. studied 44 spain-english bilingual older people in disease research center of university of california and explored the relationship between proficiency of either of two languages and onset age of the dementia, finding that those with higher proficiency of second language are less likely to get symptoms of ad and other kinds of dementia (the higher proficiency, the lower the incidence) [8]. more recently, indian scientists continued to push forward the research theme in journal neurology by investigating language abilities of 648 indians diagnosed with different degrees of different types of dementia (ad and frontal temporal dementia, vascular dementia, etc.). they reported that those who could speak two languages developed dementia 4.5 years later than those who could speak one, irrespective of other potential confounding factors such as education, sex, occupation, and living environments [9]. in contrast to previous studies, all the participants in this study were natives of india, which ruled out of the confounding effect of immigration status completely. however, the subjects in most studies reviewed above are either patients diagnosed with dementia or healthy old people, with individuals in between unconsidered. as mentioned at the beginning, some scholars believe that there is a progressive transition stage between the healthy and demented elderly, known as amnestic mild cognitive impairment (amci) [30]. by examining the effect of bilingualism on the age of diagnosis in individuals with single-or multiple-domain amci through a battery of k. cheng et al. admet & dmpk 3(3) (2015) 260-273 264 neuropsychological tests and questionnaires, michigan scholars identified the early protective advantage of bilingual experience against amci, which was further narrowed down to single domain amci, one specific subtype most likely to progress to ad. hopefully, the study verified the inevitable link between l2 experience and the defense against ad in the whole temporal course and proved the continuity of the defense [4]. neuroimaging evidence the exciting behavioral and pathological results aforementioned were in line with a large body of neuroimaging studies. firstly, some researchers recently explored with fmri the impact of speaking a second language on brain functional connectivity in the elderly. relative to the monolingual counterparts, bilinguals performed better on perceptual switching tasks while with significantly less activation in three key regions of the brain responsible for executive control, left dlpfc (dorsolateral prefrontal cortex), left vlpfc (ventrolateral prefrontal cortex) and acc (anterior cingulate cortex) , with other conditions controlled for [31]. this suggests that the bilingual experience, especially constant daily switching between languages, may enhance the executive control system, resulting in less consumption of cognitive resources on the same cognitive tasks. secondly, schweizer et al. proved for the first time by ct (computed tomography) the differences of brain structure between bilingual and monolingual patients diagnosed with ad after matched on level of cognitive performance and years of education [32]. through linear analysis into ct images of 40 subjects, they found bilingual patients exhibited substantially greater amount of brain atrophy than monolinguals in traditionally ad related brain regions, specifically, medial temporal lobe. it is implied that bilinguals could tolerate more brain atrophy or brain structural changes without symptoms, or maintain comparable cognitive level with their counterparts but with less brain atrophy. therefore, bilingual experience could protect against neuropathology of ad, delaying the onset of the disease. the volume of gray matter (gmv) will decrease with normal aging [33], but the extent may be regulated by external experiences, such as speaking another language on daily basis. the left anterior temporal pole (latp) is considered as a conceptual hub and characteristic region of brain plasticity in bilingual speakers, because they require it to store and distinguish the lexical concepts of two different languages in both comprehension and production processes [28]. abutalebi et al. employed vbm (voxel-based morphometry) to compare the brain structure of 23 chinese/english bilingual elders (mandarin or cantonese) and monolingual italian elders [34]. after ruling out all kinds of confounding factors, they found a wider range of age-related decreases in gmv in monolingual brains and a significant increase in left temporal pole in bilingual brains. meanwhile, roi (region of interested) analysis showed that bilinguals’ performance in picture-naming tasks in second language was positively correlated with gmv in left temporal lobe. this indicates that the temporal pole is vulnerable during normal aging and speaking another language contributes a lot to holding up the development of age-related gmv decreases and maintaining the health of the elderly. additionally, the same team had found that the advantage of gmv in acc related to executive control in young adult bilinguals [35]. the decrease in white matter integrity is also known to accompany aging [36], and the extent may be also mitigated by bilingual experience. through dti (diffusion tensor imaging ), luk et al. found higher wm integrity in older bilinguals in the corpus callosum extending to the superior and inferior longitudinal fasciculi comparative to monolinguals controlling for some demographic and neuropsychological data. resting-state functional connectivity analysis showed stronger anterior to posterior functional connectivity admet & dmpk 3(3) (2015) 260-273 l2 learning and cognitive aging doi: 10.5599/admet.3.3.206 265 of bilingual frontal lobe regions adjacent to wm areas, which reflected stronger white matter (wm) connectivity between brain regions, facilitating information transfer and executive performance [37]. in a more recent study, both white matter (wm) integrity and gray matter volumes (gmv) were detected and compared between older adult lifelong bilinguals and monolinguals who were matched on a number of relevant cognitive test scores and controlled for some confounding variables such as education, socioeconomic status and intelligence. results showed that, when the two groups performed in a draw on a series of neuropsychological tasks, significantly lower wm integrity in above-mentioned corpus-callosumcentred tracts was observed in the bilingual group, such as the inferior longitudinal fasciculus and inferior fronto-occipital fasciculus bilaterally, while no significant difference in gmv [38]. this seems inconsistent with luk’s results. the authors interpreted that this inconsistency could be due to the fact that their bilingual subjects were more likely to be at a preclinical ad stage, who might have suffered moderate neuron degeneration but maintained a comparable normal cognitive function with monolinguals since the decrease of fa (fractional anisotropy) and the increase of dr (radial diffusivity) were well-recognized as the neuroimaging characteristics of amci [17,39]. they also suggested that bilinguals did not necessarily have increased wm, but might actually compensate for their wm degeneration through the use of their more efficient executive function networks [17]. mechanisms of l2 learning influencing cognitive aging the studies reviewed in former section provide plausible evidence that both epidemiological and neuroanatomical effects can be induced by long-term experience with learning a new language. most exciting are sufficient findings that speaking two or more languages can serve as a safeguard to maintain cognitive function in the elderly, delay cognitive decline and even protect against the onset of neurodegenerative diseases such as ad. what remains to be understood are the questions of why and how second language learning can affect cognitive aging. in this section we will explore the reasons and the neural mechanisms from the three prominent aspects which are cognitive reserve, neural plasticity and brain networks in order to provide a synthesis of available studies, and build up a basis for future research in this domain. cognitive reserve hypothesis in today’s world where the aging problem is increasingly serious, it is vital to study the issue of cognitive reserve [40]. the cognitive reserve hypothesis are proposed by stern in contrast to the twin concept of brain reserve [41], which is generally believed to be quantitatively measured by the number of synapses or brain size. for an individual it is assumed to have a critical threshold, beyond which either clinical or functional disorders will be manifested. but due to individual differences in brain sizes, the threshold varies a lot, such that the “passive” brain reserve cannot explain the common phenomena that many older individuals display distinct cognitive ability although with similar degree of brain atrophy or injury [41]. on the other hand, cr is dynamic and experience-dependant, which can better justify the discrepancy between the degree of neuropathology and clinical symptoms. it highlights the individuals’ ability that they employ existing brain reserves or alternative networks to complete the tasks flexibly and effectively when brain reserve is insufficient or optimal network is disrupted. it applies to both healthy old individuals and those with brain injury [42]. neural reserve and neural compensation are accepted as two main mechanisms by which cr functions. the former posits that some individuals with neuropathology maintain cognitive functioning by making more efficient use of the same networks engaged by healthy individuals [43]. in this case, individuals with greater neural reserve require less neural activation within the optimal k. cheng et al. admet & dmpk 3(3) (2015) 260-273 266 task-specific network in response to a given increase in task demand compared with those with less neural reserve, as a result of which those with greater neural reserve can withstand more neuropathology before the optimal task-specific network is disrupted [44]. on the other hand, neural compensation highlights individual differences in the capacity to recruit alternative networks or brain structures for the sake of maintaining cognitive functioning when the optimal task-specific networks of a given tasks have been disrupted [44-45]. that is, individuals with greater cr can produce new and compensatory networks to make up for the defects caused by neuropathology or brain damage. for example, a meta-analysis of more than 29,000 people showed that individuals with high cr were 46 % less likely to be diagnosed with dementia as opposed to those with lower cr [46]. much literature shows that quite a few indexes contribute to the construct of cr, such as childhood intelligence, education, occupation accomplishment, social economic status, leisure life style and personality [47,48]. however, it is well-acknowledged that these indicators are unable to stop the development of neuropathology or other brain injuries, but modify the clinical manifestations or pathological symptoms of diseases. and many of these indexes are practically inseparable, and overlap considerably to contribute to general construct of cr, which makes it an urgent future task to specify their respective contributions and causal mechanisms [28]. bilingualism and cognitive reserve in recent years, as one of the key indicators to measure cr, the leisure life style has become a hot research topic in cognitive aging [47]. the leisure activities are generally categorized into cognitive or intellectual training (such as reading, and playing cards), social networking (such as visiting friends) and physical activity (such as walking) [49,50]. with converging evidence that bilingual experience significantly delays onset of dementia, it is no wonder that learning and speaking a second language is featured as a cognitively stimulating activity that affect the cr, not to mention some even take it as an independent indicator of cr [28]. literature on bilingualism shows that bilinguals have to inhibit the unattended language while using the target one since both two languages are always active in most communication situations [51]. as a matter of fact, bilinguals are engaged in substantial exercise on brain subconsciously on daily basis over lifetime [52]. there arises the question how such exercise contributes to greater cr and further exerts positive impact on cognitive aging. much of the research addressing this question has focused on the assumption that countless inhibition exercise in bilinguals over a long period of time is bound to enhance the central executive control of the brain, which declines gradually with aging [53]. thus, speaking two languages from early years is supposed to forge a much stronger “executive control system” to protect against brain impairment or injury [54] since almost all the types of dementia even normal aging are associated with the impairment and dysfunction of central executive control system [55]. literature shows language switching network overlaps significantly with general executive function network (non-linguistic, such as perception switching network) in the lpfc (left prefrontal cortex) and acc (known as the hub of executive control) [56,57]. in view of this, it is justifiable that older bilinguals’ reduced activation in overlapping areas in perceptual taskswitching experiment by gold at al. are consistent with a negative interaction between language competence and cognitive control demands (better language, less demand on control) in another longitudinal study spanning a year with fifteen bilingual children [38,56]. the mechanism involved may be that l2 learning increases neural efficiency of executive control regions to benefit even nonlinguistic tasks. admet & dmpk 3(3) (2015) 260-273 l2 learning and cognitive aging doi: 10.5599/admet.3.3.206 267 on the other hand, schweizer and his colleagues’ research mentioned above provides direct evidence that neural compensation may help determine when symptoms of dementia may set on in old people [32]. more atrophy in medial temporal lobe (possibly more ad neuropathology) in bilingual patients indicates that a certain amount of brain regional atrophy or neural network failure does not mean the ever-lasting loss of functions or network collapse but can be compensated for through actively recruiting other remaining regions or networks in the whole brain. for example, luk et al. (2011) implied that the greater wm integrity around corpus callosum and the surrounding fiber bundles in bilinguals may compensate for the growing decrease of gmv due to aging or neurodegenerative diseases [37]. to sum up, lifelong bilingual experience may serve as a major deterrent to the onset of age-related cognitive decline, which provides the neural basis for the idea of “cognitive reserve”. l2 learning and neuroplasticity further support for the relationship between cognitive reserve and bilingualism comes from abundant literature that the enriched environment can influence the brain by modifying its physical structure and functional organization throughout the lifetime, the phenomenon known as neuroplasticity. it posits that structural changes may result from learning and experience, or responses to brain damage [58]. for instance, london taxi drivers show increased volume of the posterior hippocampus [59], a region which is vital for spatial memory. likewise, l2 learning can also bring about structural changes in brain in terms of increased gmv, increased cortical thickness (ct), or enhanced wm integrity. a number of recent studies have identified such changes induced by l2 learning in a series of brain regions in healthy adults, such as the prefrontal cortex (pfc), the temporo-parietal cortex, anterior cingulate cortex (acc), hippocampus, caudate nucleus as well as putamen [58]. one of the early prominent studies using vbm to examine gmv in language learners was conducted by mechelli and his colleagues, who found greater gmv in the left inferior parietal lobule (ipl) of bilinguals than that of monolinguals, and the effect was modulated by the age of l2 learning (the earlier they learnt, the more the gmv) and the proficiency (more proficient, more gm) [60]. subsequent studies replicated their results and confirmed that bilinguals had greater gm density than monolinguals in the temporoparietal cortex, such as inferior parietal lobule (ipl) and the posterior supramarginal gyrus (smg) [58, 61]. in addition, abutalebi’s team found young bilinguals had gm advantage over monolinguals in acc through flanker task [62], and in left putamen through a picture naming task[63]. interestingly, zou et al. found that advanced language learners (chinese sign language) had much greater gmv in the left caudate nucleus than monolinguals, which is so far the first attempt to illustrate the effect of different modal language learning [64]. in terms of wm integrity, luk et al. (2011) found higher white matter integrity in healthy bilingual older adults, primarily in the corpus callosum and increased anterior-posterior connectivity [37]. as discussed above, this result suggests bilingualism is associated with better maintenance of wm integrity in the course of cognitive aging [65]. however, cummine and boliek found greater integrity for adult monolinguals over bilinguals in the right inferior fronto-occipital fasciculus and the anterior thalamic radiation [66], which seems contradictory to luck’s findings (2011), but more consistent with those of mohades et al. (2012) in which no significant differences were observed between monolingual and bilingual children (aged 8 to 11) [67]. these three studies may reflect a developmental trajectory concerning the establishment of lifelong wm integrity, which may take as much as few decades of l2 experience to a stable degree [58]. most of the studies reviewed above are about bilingual speakers who have long-term (sometimes k. cheng et al. admet & dmpk 3(3) (2015) 260-273 268 lifelong) l2 experiences. confusion remains about whether short-term learning or intensive training lead to the same type of structural changes. several recent longitudinal studies have been intended to clear the confusion, and some exciting results are achieved. stein et al. (2012) found that college students who went to switzerland learning german over the course of 5 months increased gm density in the left ifg (inferior frontal gyrus) and the left anterior temporal lobe, two areas that are implicated to serve lexical access and semantic integration [58,68]. mårtensson et al. (2012) examined 14 young military students who went through intensive language training for 3 months in preparation for interpreters and found, as compared to controls matched for age and cognitive abilities, the future interpreters had increased ct in left pfc as well as increased right hippocampal volume [69]. meanwhile, schlegel et al. (2012) found greater wm density for 11 american college students who had had 9-month chinese intensive language course in opposition to 16 controls [70]. further, hosoda, et al. found, after a 16-week english vocabulary training in the laboratory, 24 college students showed both increased gm and wm density in the right ifg (inferior frontal gyrus), and an average increase of 6 % in the volume of right prefrontal lobe as compared to the controls [71]. in summary, as far as neural plasticity is concerned, the aforementioned studies with balanced or unbalanced young bilinguals (even beginners), help explore possible neural mechanism by which l2 may sculpt the brain. it could be concluded that a period of language learning or practice, whether short-term or long term, may lead to changes in the structure of the brain. accordingly, we attempt to believe that if such changes are maintained to a certain amount as brain reserve, individuals may be able to withstand reductions of the neurons, the deterioration of cell death or nerve fiber winding, which are accompanied by aging in the long run. overlapping between language and aging networks as discussed above with the connotation of cognitive aging, although the mechanisms of pathological aging and normal aging are not the same, they are both accompanied by decline in cognitive functions, ranging from the general (e.g., reduced working memory and executive control) to the language-specific (e.g., slower lexical retrieval and poor fluency). the cognitive decline maybe typically derive from brain atrophy, loss of neuronal synaptic connections (functional connectivity), and signs of neuropathology associated with dementia [71]. a few brain imaging studies have established that the brain network involved in cognitive aging is extensive, including prefrontal cortex, medial temporal lobe and subcortical structures such as the hippocampus, acc, etc. in normal aging without clinically significant neurodegenerative diseases, the shrinkage of gm rather than wm is found to be the principal cause of total brain volume reduction. between the ages of 30 to 90, the brain's gm will diminish by 15 % to 25 %, with the atrophy mostly occurring in frontal and temporal cortexes, such as the prefrontal cortex, in which gm reduction is said to be associated with the decline or damage of executive control in elders [72]. in the presence of cognitive impairment which is not sufficiently severe to meet criteria for a dementia diagnosis, hippocampal atrophy emerges as the most consistent symptom [73]. wolf et al. (2001) suggests that the turning point from normal cognitive aging to earlier alzheimer’s disease may be detected by hippocampal atrophy. meanwhile, wm density decreases in the pertinent brain regions responsible for information processing speed and executive function due to the effect of demyelinization caused by aging [17]. however, quite a few studies have found that language learning network overlaps extensively with the brain network engaged in cognitive aging [74]. the brain network involved in language learning in adulthood can be characterized into several sub-networks [75]. specifically, the left ifg (inferior frontal gyrus) and left mfg (middle frontal gyrus) in prefrontal cortex are key regions in the articulatory network. admet & dmpk 3(3) (2015) 260-273 l2 learning and cognitive aging doi: 10.5599/admet.3.3.206 269 additionally, medial temporal lobe is involved in acoustic–phonetic processes and the acquisition of meaning [76]. the hippocampus may be critically involved in the memory processes or vocabulary acquisition [77]. the learning of grammatical rules is linked to the frontal-striatal system, which connects frontal lobe regions with caudate nucleus and putamen [78]. l2 learning largely involves the same neural structures as the native language, except that more activity in left prefrontal areas is typically observed in those who acquire the second language later in life and whose proficiency has not reached native-like level [79]. as reviewed in last section, the anterior cingulated cortex (acc), the ipl (inferior parietal lobe), and subcortical regions including the basal ganglia, particularly the left caudate and the putamen are involved in bilingual control network [80,58]. the exact functional roles of these sub-networks are sometimes controversial, however, few would deny that the acquisition of a new language involves a large brain network with regions similar to those implicated in cognitive aging described above. therefore, we attempt to hypothesize that it is the overlapping of these two kinds of networks that make the l2 learners keep strengthening brain areas which are prone to cognitive decline. for example, wm density enhancement schlegel (2012) found in fiber bundles going through corpus callosum in young l2 learners [70], is amazingly consistent with luk’s suggestion (2011) that life-long language experience entails the maintenance of wm in similar regions of older healthy people [37]. conclusion taken together, given the possible mechanisms under which l2 learning may affect cognitive aging, we attempt to hypothesize that the ultimate result of second language learning may be that the integrity of the brain structures involved is maintained to attenuate atrophy or lesion, and more potential neural networks and sub-networks available could allow for compensation or substitute of age-related cognitive declines. however, there are still many researchers who have questioned this hypothesis. for example, in the two longitudinal studies with japanese americans who began to learned japanese as adults, crane et al. found there were no connection between the incidence of dementia later and japanese proficiency, either written or oral [81]. sanders et al. (2012) also failed to find speaking english may be a cognitive activity associated with the lower risk of developing dementia for non-native english speakers [82]. in a more recent study of a larger sample of hispanic immigrants in manhattan communities over a time-span of 23 years, still no protective effect of bilingualism against normal aging and dementia was identified in spite of the findings that bilingualism was associated with better initial performance on tests of memory, executive function, and task switching[83]. it is possible that the discrepancy in results between studies is due to some factors with regard to the methods to determine dementia, criteria to select participants, definition and measurements of bilingualism, the dissolution of language experience from education, among others. nonetheless, these contradictory results undoubtedly pose an unprecedented challenge to our hypothesis. urgency turns out to be clarification-making about circumstances under which second language learning can delay the cognitive aging and the roles some variables play in bilingualism as a moderator between neuropathology and clinical onset, such as migration status, language proficiency, frequency of use and so on [28]. more research in the future is needed to achieve a breakthrough in these respects. most of the studies discussed above concerns those early language learners who spoke two or more languages over the course of a lifetime, they have high proficiency of each language and the frequency of using it is also high, two variables known to play a crucial part in constituting bilingual advantage over monolinguals [8,60]. it is no doubt that the effect of language experience promoting healthy aging against cognitive decline become more obvious and stable in this population [5]. moreover, more exciting is much https://en.wikipedia.org/wiki/frontal_lobe k. cheng et al. admet & dmpk 3(3) (2015) 260-273 270 evidence that short-term foreign language learning or training in young adults can also lead to positive structural and functional changes [68-71]. it seems quite possible that, if these changes can be maintained to the elderly, they are bound to constitute cognitive reserve, suppressing the decrease of neurons or nerve fibers winding causing age-related cognitive decline. given that the aged brain retains somewhat plasticity [40], foreign language learning initiated in old age is also likely to have analogous effects, and be used as an effective therapy to alzheimer's disease [17]. however, up to present, few empirical studies have been conducted to give definite evidence to this end. therefore, more longitudinal studies are needed in future work to determine whether learning a foreign language in the elderly (aged 65 or above) can promote cognitive functions or even produce positive structural changes in their brain as in young adults. many studies have shown other cognitively stimulating activities, such as reading newspapers, bridge, chess, crossword puzzle, among others, can also improve the cognitive function and delay the symptoms of ad and mild cognitive impairment [84,85]. in order to seek optimal non-medicinal approaches to prevent or treat dementia, it is required that magnitude of cognitive improvement should be compared between varieties of cognitively stimulating activities. it remains to be verified whether l2 learning could serve as a major deterrent or preferable therapy to age-related cognitive decline, and whether stronger neural networks or connections are engaged in l2 learning than other forms of cognitive trainings that have been investigated. acknowledgements this review was supported by the national social science foundation of china (15byy068), humanities and social science project of education department in sichuan province, china (14sb0212, 15sa0199). references [1] department of economic and social affairs, world population ageing. united nations publications; new york: 2007. [2] b.l. plassman, k.m. langa, g.g. fisher, s.g. heeringa, d.r. weir, m.b. ofstedal, j.r. burke, m.d. hurd, g.g. potter, w.l. rodgers, d.c. steffens, r.j. willis, r.b. wallace, neuroepidemiology 29(1-2) (2007) 125-132. [3] l.l. drag, l.a. bieliauskas, journal of geriatric psychiatry and neurology 23(2) (2010) 75-93. [4] c. kaiwen, d. yanhui, y. dezhong, advances in psychological science 22(11) (2014) 1723-1732. [5] e. bialystok, f.i.m. craik, k.j. freedman, k.j. murphy, a.k. troyer, neuropsychologia 45(2) (2007) 459-464. [6] f.i.m. craik, e. bialystok, m. freedman, neurology 75(19) (2010) 1726-1729. [7] h. chertkow, v. whitehead, n. phillips, c. wolfson, j. atherton, h. bergman, alzheimer disease & associated disorders 24(2) (2010) 118-125. [8] t.h. gollan, d.p. salmon, r.i. montoya, d.r. galasko, neuropsychologia 49(14) (2011) 3826-3830. [9] s. alladi, t.h. bak, v. duggirala, b. surampudi, m. shailaja, a.k. shukla, j.r. chaudhuri, s. kaul, neurology 81(22) (2013) 1938-1944. [10] l. ossher, e. bialystok, f.i.m. craik, k.j. murphy, a.k. troyer, the journals of gerontology series b: psychological sciences and social sciences 68(1) (2013) 8-12. [11] l.j. whalley, i.j. deary, c.l. appleton, j.m. starr, ageing research reviews 3(4) (2004) 369-382. [12] m. tosato, v. zamboni, a. ferrini, m. cesari, clinical interventions in aging 2(3) (2007) 401-412. [13] u. lindenberger, p.b. baltes, psychology and aging 9(3) (1994) 339-355. admet & dmpk 3(3) (2015) 260-273 l2 learning and cognitive aging doi: 10.5599/admet.3.3.206 271 [14] k.r. daffner, journal of alzheimer's disease 19(4) (2010) 1101. [15] l.x. hy, d.m. keller, neurology 55(2) (2000) 198-204. [16] a. la rue, l.f. jarvik, the international journal of aging & human development 25(2) (1987) 79-89. [17] m. antoniou, g.m. gunasekera, p.c.m. wong, neuroscience & biobehavioral reviews 37(10) (2013) 2689-2698. [18] h. wenguang, c. baoguo, advances in psychological science 19(11) (2011) 1615-1624. [19] e. bialystok, f.i.m. craik, r. klein, m. viswanathan, psychology and aging 19(2) (2004) 290-303. [20] e. bialystok, f.i.m. craik, j. ryan, journal of experimental psychology: learning, memory and cognition 32(6) (2006) 1341-1354. [21] g. kavé, n. eyal, a. shorek, j. cohen-mansfield, psychology and aging 23(1) (2008) 70-78. [22] b.t. gold, c. kim, n.f. johnson, r.j. kryscio, c.d. smith, the journal of neuroscience 33(2) (2013) 387-396. [23] e. bialystok, f.i.m. craik, g. luk, journal of neurolinguistics 21(6) (2008) 522-538. [24] l. luo, f.i.m. craik, s. moreno, e. bialystok, psychology and aging 28(1) (2013) 28-34. [25] m. perquin, m. vaillant, a.-m. schuller, j. pastore, j.f. dartigues, m.l. lair, n. diederich, plos one 8(4) (2013) e62030. doi:10.1371/journal.pone.0062030. [26] t.h. bak, j.j. nissan, m.m. allerhand, i.j. deary, annals of neurology 75(6) (2014) 959-963. [27] a.b. graves, l. rajaram, j.d. bowen, w.c. mccormick, s.m. mccurry, e.b. larson, the journals of gerontology series b: psychological sciences and social sciences 54(3) (1999) s154-s161. [28] e. guzmán-vélez, d. tranel, neuropsychology 29(1) (2015) 139-150. [29] d.a. snowdon, s.j. kemper, j.a. mortimer, l.h. greiner, d.r. wekstein, w.r. markesbery, jama 275(7) (1996) 528-532. [30] r.c. petersen, the new england journal of medicine 364(23) (2011) 2227-2234. [31] b.t. gold, c. kim, n.f. johnson, r.j. kryscio, c.d. smith, the journal of neuroscience 33(2) (2013) 387-396. [32] t.a. schweizer, j. ware, c.e. fischer, f.i. craik, e. bialystok, cortex 48(8) (2012) 991-996. [33] n. raz, f. gunning-dixon, d. head, k.m. rodrigue, a. williamson, j.d. acker, neurobiology of aging 25(3) (2004) 377-396. [34] j. abutalebi, m. canini, p.a. della rosa, l.p. sheung, d.w. green, b.s. weekes, neurobiology of aging 35(9) (2014) 2126-2133. [35] j. abutalebi, p.a. della rosa, d.w. green, m. hernandez, p. scifo, r. keim, s.f. cappa, a. costa, cerebral cortex 2011: bhr287. [36] d.j. madden, i.j. bennett, a.w. song, neuropsychology review 19(4) (2009) 415-435. [37] g. luk, e. bialystok, f.i.m. craik, c.l. grady, the journal of neuroscience 31(46) (2011) 16808-16813. [38] b.t. gold, n.f. johnson, d.k. powell, neuropsychologia 51(13) (2013) 2841-2846. [39] b. bosch, e.m. arenaza-urquijo, l. rami, r. sala-llonch, c. junqué, c. solé-padullés, c. peña-gómez, n. bargalló, j.l. molinuevo, d. bartrés-faz, neurobiology of aging 33(1) (2012) 61-74. [40] e. bialystok, f.i.m. craik, g. luk, trends in cognitive sciences 16(4) (2012) 240-250. [41] r. katzman, neurology 43(1) (1993) 13-20. [42] n. scarmeas, y. stern, m.x. tang, r. mayeux, j.a. luchsinger, annals of neurology 59(6) (2006) 912921. [43] d. bartrés-faz, e.m. arenaza-urquijo, brain topography 24(3-4) (2011) 340-357. [44] j. steffener, a. reuben, b.c. rakitin, y. stern, brain imaging and behavior 5(3) (2011) 212-221. [45] y. stern, c. habeck, j. moeller, n. scarmeas, k.e. anderson, h.j. hilton, j. flynn, h. sackeim, r. van heertum, cerebral cortex 15(4) (2005) 394-402. [46] m.j. valenzuela, p. sachdev, psychological medicine 36(04) (2006) 441-454. k. cheng et al. admet & dmpk 3(3) (2015) 260-273 272 [47] b.r. reed, m. dowling, s. tomaszewski farias, j. sonnen, m. strauss, j.a. schneider, d.a. bennett, d. mungas, journal of the international neuropsychological society 17(04) (2011) 615-624. [48] c. sattler, p. toro, p. schönknecht, j. schröder, psychiatry research 196(1) (2012) 90-95. [49] l. fratiglioni, s. paillard-borg, b. winblad, the lancet neurology 3(6) (2004) 343-353. [50] n. scarmeas, y. stern, journal of clinical and experimental neuropsychology 25(5) (2003) 625-633. [51] y.j. wu, g. thierry, the journal of neuroscience 30(22) (2010) 7646-7651. [52] j. diamond, science(washington) 330(6002) (2010) 332-333. [53] a.f. kramer, s. hahn, n.j. cohen, m.t. banich, e. mcauley, c.r. harrison, j. chason, e. vakil, l. bardell, r.a. boileau, a. colcombe, nature 400(6743) (1999) 418-419. [54] e. bialystok, f.i.m. craik, current directions in psychological science 19(1) (2010) 19-23. [55] l.m. duke, a.w. kaszniak, neuropsychology review 10(2) (2000) 75-99. [56] j. abutalebi, d. green, journal of neurolinguistics 20(3) (2007) 242-275. [57] t. guo, h. liu, m. misra, j.f. kroll, neuroimage 56(4) (2011) 2300-2309. [58] p. li, j. legault, k.a. litcofsky, cortex 58 (2014) 301-324. [59] e.a. maguire, d.g. gadian, i.s. johnsrude, c.d. good, j. ashburner, r.s. frackowiak, c.d. frith, proceedings of the national academy of sciences 97(8) (2000) 4398-4403. [60] a. mechelli, j.t. crinion, u. noppeney, j. o'doherty, j. ashburner, r.s. frackowiak, c.j. price, nature 431(7010) (2004) 757-757. [61] p.a. della rosa, g. videsott, v.m. borsa, m. canini, b.s. weekes, r. franceschini, j. abutalebi, cortex 49(2) (2013) 605-608. [62] j. abutalebi, p.a. della rosa, d.w. green, m. hernandez, p. scifo, r. keim, s.f. cappa, a. costa, cerebral cortex 22(9) (2012) 2076-2086. [63] j. abutalebi, p.a. della rosa, g. ding, b. weekes, a. costa, d.w. green, cortex 49(3) (2013) 905-911. [64] l. zou, g. ding, j. abutalebi, h. shu, d. peng, cortex 48(9) (2012) 1197-1206. [65] m.h. davis, m.g. gaskell, philosophical transactions of the royal society b: biological sciences 364(1536) (2009) 3773-3800. [66] j. cummine, c.a. boliek, brain structure and function 218(2) (2013) 595-601. [67] s.g. mohades, e. struys, p. van schuerbeek, p. van de craen, r. luypaert, brain research 1435 (2012) 72-80. [68] m. stein, a. federspiel, t. koenig, m. wirth, w. strik, r. wiest, d. brandeis, t. dierks, cortex 48(4) (2012) 458-465. [69] j. mårtensson, j. eriksson, n.c. bodammer, m. lindgren, m. johansson, l. nyberg, m. lövdén, neuroimage 63(1) (2012) 240-244. [70] a.a. schlegel, j.j. rudelson, u.t. peter, journal of cognitive neuroscience 24(8) (2012) 1664-1670. [71] c. hosoda, k. tanaka, t. nariai, m. honda, t. hanakawa, the journal of neuroscience 33(34) (2013) 13663-13672. [72] f.m. gunning-dixon, n. raz, neuropsychologia 41(14) (2003) 1929-1941. [73] h. wolf, m. grunwald, f. kruggel, s.g. riedel-heller, s. angerhöfer, a. hojjatoleslami, a. hensel, t. arendt, h. gertz, neurobiology of aging 22(2) (2001) 177-186. [74] a. rodríguez-fornells, t. cunillera, a. mestres-missé, r. de diego-balaguer, philosophical transactions of the royal society of london b: biological sciences 364(1536) (2009) 3711-3735. [75] g. hickok, d. poeppel, nature reviews neuroscience 8(5) (2007) 393-402. [76] p.c.m. wong, m. ettlinger, j.p. sheppard, g.m. gunasekera, s. dhar, ear and hearing 31(4) (2010) 471-479. [77] m.h. davis, m.g. gaskell, philosophical transactions of the royal society b: biological sciences 364(1536) (2009) 3773-3800. admet & dmpk 3(3) (2015) 260-273 l2 learning and cognitive aging doi: 10.5599/admet.3.3.206 273 [78] k.l. sakai, science 310(5749) (2005) 815-819. [79] j. abutalebi, acta psychologica 128(3) (2008) 466-478. [80] y. li, j. yang, k.s. scherf, p. li, brain and language 127(3) (2013) 452-462. [81] p.k. crane, l.e. gibbons, k. arani, v. nguyen, k. rhoads, s.m. mccurry, l. launer, k. masaki, l. white, epidemiology (cambridge, mass.) 20(5) (2009) 766-774. [82] a.e. sanders, c.b. hall, m.j. katz, r.b. lipton, journal of alzheimer's disease 29(1) (2012) 99-108. [83] l.b. zahodne, p.w. schofield, m.t. farrell, y. stern, j.j. manly, neuropsychology 28(2) (2014) 238246. [84] j. verghese, r.b. lipton, m.j. katz, c.b. hall, c.a. derby, g. kuslansky, a.f. ambrose, m. sliwinski, h. buschke, new england journal of medicine 348(25) (2003) 2508-2516. [85] r.s. wilson, d.a. bennett, j.l. bienias, n.t. aggarwal, c.f. mendes de leon, m.c. morris, j.a. schneider, d.a. evans, neurology 59(12) (2002) 1910-1914. ©2015 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ doi: 10.5599/admet.2.4.210 84 admet & dmpk 4(2) (2016) 84-90; doi: 10.5599/admet.4.2.210 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index short communication metabolism of six cyp probe substrates in fetal hepatocytes abdul naveed shaik* 1,2 , sandeep kumar vishwakarma 1 and aleem ahmed khan 1 1 center for liver research and diagnostics, owaisi research center, deccan college of medical sciences, hyderabad, ts, india 2 department of pharmaceutical sciences, mcphs university, boston ma, usa *corresponding author. e-mail: naveedshaik@gmail.com; tel.: +1-407-303-7003; fax: +1-407-303-7030 received: august 12, 2016; revised: october 02, 2016; published: june 29, 2016 abstract cytochrome p-450 (cyp) are the most common drug metabolizing enzymes and are abundantly expressed in liver apart from kidney, lungs, intestine, brain etc. their expression levels change with physiological conditions and disease states. the expression of these cyps is less in human foetus and neonates compared to adults, which results in lower clearance of xenobiotics in infants and neonates compared to adults. hepatocytes are the cells which are largely used to study these cyps. we have isolated hepatocytes from aborted foetus to study the metabolism of six probe substrates: phenacetin, diclofenac, s-mephenytoin, dextromethorphan, nifedipine and testosterone. the results obtained show the expression of various cyps (cyp1a2, cyp2c19, cyp2c9, cyp2d6, and cyp3a4) in human foetus and their involvement in metabolism of cyp probe substrates. keywords fetal metabolism; cyp (cytochrome-p450); hepatocytes; metabolic stability; hplc; dextromethorphan; diclofenac; s-mephenytoin; nifedipine; testosterone. introduction the role of cytochromes p450 (cyp) mediated metabolism is well appreciated in the drug discovery setup, over the last few decades much emphasis has been laid on how different drugs are handled by the cyps. cyps are one of the major source of variability in pharmacokinetics (pk) and pharmacodynamics (pd) of the drugs [1]. the cyps are widely distributed and expressed in different organs including, liver, intestine, lungs, kidney, placenta, brain etc. the highest expression levels are observed in liver, and the major ones to be expressed in liver are cyps 3a4, 2c9, 2c8, 2e1, and 1a2 [2]. these five isoforms are responsible for the biotransformation of many of the drugs in market which makes up to 80 % of the total drugs in clinical use. expression of these cyp enzymes depends on varied factors including sex, age, disease states, etc [3]. the mechanism of these cyp isoforms and their contribution towards drug-drug interactions leading to either changes in label or complete withdrawal from the market has been well understood [4]. even though thorough investigations of these cyps including their polymorphism led to relatively safer prescription practices [5], involvement of fetal metabolism still remains highly disparaged [6-9]. role of cyp enzymes in clearance of drugs from foetus and neonates is explored since early 70’s, but the lack of microsomes and hepatocytes of fetal and neonatal origin has greatly limited our understanding of the drug behaviour in http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:naveedshaik@gmail.com admet & dmpk 4(2) (2016) 84-90 cyp substrates in fetal hepatocytes doi: 10.5599/admet.4.2.210 85 foetus and neonates. furthermore, lack of complete knowledge of adverse events related to drugs and scant data of pk parameters during pregnancy leads to either cautious prescription or complete avoidance of many drugs [10-12]. advances in drug metabolism, drug transport and drug-drug interactions in human foetus will greatly benefit the patients and clinicians alike [13]. hepatocytes are considered as one of the better in vitro tools to study metabolism [14, 15], therefore we explored the metabolic stability of six probe substrates: phenacetin (cyp1a2), diclofenac (cyp2c9), s-mephenytoin (cyp2c19), dextromethorphan (cyp2d6), nifedipine (cyp3a4) and testosterone (cyp3a4) in hepatocytes of fetal origin [16]. experimental chemicals and reagents diclofenac, 4-hydroxydiclofenac, dextromethorphan, dextrorphan, phenacetin, paracetamol, smephenytoin, 4-hydroxymephenytoin, nifedipine, oxidized nifedipine, testosterone, 6βhydroxytestosterone, celecoxib, dmso (dimethyl sulfoxide), dmem (dulbecco’s modified eagles medium), pbs (phosphate buffer saline) were obtained from sigma aldrich (st. louis, mo usa). hplc (high performance liquid chromatography) grade acetonitrile and methanol were obtained from e merck india limited (mumbai, india), and all the other chemicals were used of highest quality available. hepatocytes isolation, in vitro culture and characterization human fetal livers were obtained from spontaneous abortions from 10-14 weeks of gestation. all donors of the fetus were thoroughly screened serologically for syphilis, toxoplasmosis, rubella, hepatitis b and c, human immunodeficiency virus 1, cytomegalovirus, parvovirus, and herpes simplex types 1 and 2. liver cells were isolated using two-step collagenase perfusion method as described by habibullah et al [17]. briefly, resected liver from the aborted fetuses were collected under sterile conditions within 2h of the termination of pregnancy and then cut into thin slice (1-2 mm thickness). thereafter, each fetal liver was digested with 0.02 % collagenase type iv (sigmaaldrich, st. louis, usa) containing penicillin g (10 u/ml, streptomycin10 u/ml, amphoteracin b (0.025 μg/ml) at 37 °c for 30 min. single cell suspension of liver was collected by centrifugation at 250 g and then washed with sterile phosphate buffered saline (pbs ph 7.4). the viability of the cells was measured using trypan blue method. cells showing more than 80 % viability were used for the metabolism experiments. metabolism studies 1 ml of cell mixture containing 1 million cells per ml were transferred to a 12 well cell culture plate (costar) and 10 µl of test compound (in dmso water) was added to maintain a final volume of dmso below 0.2% v/v. the incubation plate was then transferred to an incubator set at 37 °c with a relative humidity of 95 % and co2 (carbon dioxide) at 5 %. the samples were incubated for 2 h and at the end of 2 h viability of the cells was checked using an aliquot of 20 µl mixture. the incubation was terminated using 4 ml of acetonitrile containing internal standard celecoxib (1 mg/ml), vortex mixed for 2 mins and centrifuged at 3000 rpm on a table top centrifuge (remi instruments, india). the supernatant was transferred to clean glass tube and the contents were dried under gentle stream of nitrogen using a multivap evaporator set at 40 °c (n-evap, orgiinomation, berlin, ma, usa). the residue was reconstituted in 200 µl of mobile phase (a:b, 1:1) and 100 µl of this solution was injected into the hplc. for 0 h samples, acetonitrile was added prior to addition of test compound to prevent the reaction followed by the above described sample preparation method. each of the test compounds was incubated in triplicates; the area of the drug was divided by the area of shaik et al. admet & dmpk 4(2) (2016) 84-90 86 the internal standard to get the area ratio. the area ratio obtained from 0 h was considered as 100 % and the area ratio of 2 h sample was calculated to get the metabolic stability of the test compound. each of the experiment was conducted at least three times on different days with new fetal hepatocytes. instrumentation and chromatographic conditions hplc analysis was performed using the previously published method [18]. in brief, hplc system consisted of a waters 2695 alliance separation module attached with a waters 2996 photodiode array (pda) detector, the instrument was setup to detect a range of 190-400 nm. a c18 inertsil ods 3v column (4.6 x 250 mm, 5 µm, gl sciences, inc., tokyo, japan) was used for the analysis. a ternary mobile phase gradient system was used consisting of a (0.01 m ammonium acetate, ph 5.0: acetonitrile; 90:10), b (0.01 m ammonium acetate buffer, ph 5.0: acetonitrile; 5:95) and c0.01 mm ammonium acetate buffer, ph 5.0:methanol; 5:95), the total run time was 40 minutes with gradient flow shown in table 1, analysis was performed by integrating the peak area of the individual peaks at uv (ultraviolet) maximum as described by rao et al [18] and shown in table 2. similarly, the peak area counts of internal standard was integrated. the area counts of test compound was divided by the area counts obtained from the internal standard (is) within the same analytical run to get the area ratios which were used to calculate the % depletion of the parent compound over a period of 2 h. data analysis the data was plotted using prism software (graphpad software inc., san diego, ca), all data are presented as the mean of each group ± s.e.m. (standard error of mean) as indicated in the figure legend. metabolic stability was calculated using the substrate depletion approach [19], using the equation below [20]. ratio of substrate in test sample % substrate remained in test sample = *100 ratio of substrate in control sample results and discussion isolation of hepatocytes from the fetal liver resulted in 80 % viable cell, isolation resulting in less than 80 % viability, or cells showing visible signs of deterioration were not considered for metabolism experiments. figure 1 shows microscopic image of the 14 week old healthy hepatocytes. six of the commonly used probe substrates, phenacetin at a concentration of 100 µm as a substrate of cyp1a2 [21], diclofenac at a concentration of 25 µm as a substrate of cyp2c9 [22], s-mephenytoin at a concentration of 5 µm as a substrate of cyp2c19 [23], dextromethorphan at a concentration of 50 µm as a substrate of cyp2d6 [24], nifedipine at a concentration of 5 µm as a substrate of cyp3a4 and testosterone at a concentration of 120 µm as a substrate of cyp3a4 [25], were used to evaluate the contribution towards fetal metabolism. these concentrations were chosen based on in-house km values generated using liver microsomes and hepatocytes and with reference to the concentrations suggested by the fda (food and drug administration) guidance for drug interaction studies 2006 [16, 26]. each of the marker reaction was monitored by comparing the uv peak of metabolite with the analytical standard of each metabolite. the peak area of all the test compound was obtained at lambda max (λ) of individual well resolved chromatographic peaks devoid of interference. table 2 shows uv max and retention time (rt) of all compounds. the following reactions were monitored, phenacetin-o-deethylation as marker reaction of cyp1a2, diclofenac-4-hydroxylation as marker reaction of cyp2c9, s-mephenytoin-4-hydroxylation as marker reaction of cyp2c19, dextromethorphan-o-demethylation as marker reaction of cyp2d6, nifedipine admet & dmpk 4(2) (2016) 84-90 cyp substrates in fetal hepatocytes doi: 10.5599/admet.4.2.210 87 oxidation as marker reaction of cyp3a4, and testosterone 6 β-hydroxylation as marker reaction of cyp3a4 to confirm the metabolism by the respective cyps. comparison of area ration between 0 h which was considered as 100 % against 2 h samples resulted in, 68.2 ± 8.13 %, 50.5 ± 4.29 %, 79.8 ± 6.11 %, 72.2 ± 6.12 %, 1.32 ± 0.21 %, and 4.26 ± 0.92 % respectively for phenacetin, diclofenac, s-mephenytoin, dextromethorphan, nifedipine and testosterone respectively. the results are shown as mean of three independent experiments with standard error of mean. depletion of all the substrates indicates maturation and involvement of cyp1a2, cyp2c9, cyp2c19, cyp2d6, and cyp3a4 in fetal liver. cyp3a4 showed maximum activity, resulting in more than 90 % depletion of both the cyp 3a4 substrates nifedipine as well as testosterone, followed by 50 % depletion of diclofenac, 32 % depletion of phenacetin, 28 % depletion of dextromethorphan and 20 % depletion of s-mephenytoin. these results clearly suggest the maturation and involvement of 5 different cyps in metabolism through human foetus. due to limitation in availability of fetal tissue other substrates cannot be evaluated. furthermore, studies to evaluate the kinetic parameters to generate km and vmax of each cyp using a probe substrate will be beneficial for the regulatory authorities and clinicians to optimize the dose and prevent drug-drug interactions in pregnant women and neonates. table 1. gradient time program for hplc with uv detection with ternary gradient mobile phase, method was reproduced from rao et. al. [18]. time, min flow, ml/min %a %b %c 0 1 100 0 0 2 1 90 0 10 9 1 50 20 30 12 1 50 25 25 25 1 30 50 20 28 1 10 85 5 35 1 10 85 5 36 1 100 0 0 40 1 100 0 0 a =90:10:: 10 µm ammonium acetate: acetonitrile, b =5:95:: 10 µm ammonium acetate: acetonitrile, and c =5:95:: 10 µm ammonium acetate: methanol table 2. retention times and uv max (λ max) of cyp probe substrates, their metabolites and internal standard. compound uv extracted, nm, retention time, min phenacetin 240 14.1 paracetamol 240 7.1 diclofenac 280 24.9 4-hydrxoydiclofenac 280 18.7 s-mephenytoin 250 15.9 4-hydroxymephenytoin 250 11.5 dextromethorphan 277 14.3 dextrorphan 277 10.3 nifedipine 240 22.9 hydroxynifedipine 240 21.2 testosterone 240 24.4 6β-hydroxytestosterone 240 15.7 celecoxib 250 30.1 shaik et al. admet & dmpk 4(2) (2016) 84-90 88 figure 1. a 10 x microscopic image of the well differentiated hepatocytes of human origin, obtained from 14 week old aborted foetus. figure 2. metabolism of six cyp probe substrates, the bars represent mean of percent remaining of the drug after 2 h obtained from three independent experiments, the bars represent standard error of mean. conclusions in conclusion, fetal hepatocytes were successfully isolated from different subjects with viability of greater than 80% using earlier reported method [17]. these cells were used to perform the metabolism studies of six probe cyp substrates, using parent depletion approach [19, 27]. the results obtained by the commonly used substrate depletion approach [19, 28, 29] prove differentiation, maturation and involvement of five cyps (cyp1a2, cyp2c9, cyp2c19, cyp2d6, and cyp3a4) in clearance of phenacetin, diclofenac, s-mepheytoin, dextromethorphan, nifedipine and testosterone from the fetal liver. furthermore, the percent depletion of testosterone and nifedipine is similar to adults (data not shown) % p a r e n t r e m a in in g d e x tr o m e to r p h a n d ic lo fe n a c m e p h e n y to in p h e n a c e ti n n if e d ip in e t e s to s te r o n e 0 2 0 4 0 6 0 8 0 1 0 0 admet & dmpk 4(2) (2016) 84-90 cyp substrates in fetal hepatocytes doi: 10.5599/admet.4.2.210 89 suggesting well differentiation of cyp3a4 as compared to other cyps which resulted in lesser depletion of other substrates. this study was limited by the availability of fetal hepatocytes to calculate the intrinsic clearance (clint) commonly used to represent in vitro metabolism [30, 31], and determine the mrna content of cyp450 for head on comparison with adult hepatocytes [32], which will be helpful in designing the doses of different drugs in pregnant women and neonates. note: abdul naveed shaik’s current affiliation: center for pharmacometric and systems pharmacology, university of florida, orlando fl references [1] u.m. zanger, m. schwab, pharmacology & therapeutics 138 (2013) 103-141. [2] j.h. ansede, d.r. thakker, journal of pharmaceutical sciences 93 (2004) 239-255. [3] j. george, c. liddle, m. murray, k. byth, g.c. farrell, biochemical pharmacology 49 (1995) 873-881. [4] a.s. kalgutkar, r.s. obach, t.s. maurer, current drug metabolism 8 (2007) 407-447. [5] s.c. preissner, m.f. hoffmann, r. preissner, m. dunkel, a. gewiess, s. preissner, plos one 8 (2013) e82562. [6] b. krauer, p. dayer, clinical pharmacokinetics 21 (1991) 70-80. [7] m.s. miller, m.r. juchau, f.p. guengerich, d.w. nebert, j.l. raucy, fundamental and applied toxicology : official journal of the society of toxicology 34 (1996) 165-175. [8] m. garland, k.m. abildskov, t.w. kiu, s.s. daniel, r.i. stark, drug metabolism and disposition: the biological fate of chemicals 33 (2005) 68-76. [9] m.r. juchau, nida research monograph 60 (1985) 17-24. [10] g.d. anderson, clinical pharmacokinetics 44 (2005) 989-1008. [11] h. malm, j. martikainen, t. klaukka, p.j. neuvonen, drug safety 27 (2004) 899-908. [12] r. sannerstedt, p. lundborg, b.r. danielsson, i. kihlstrom, g. alvan, b. prame, e. ridley, drug safety 14 (1996) 69-77. [13] g.g. briggs, r.k. freeman, s.j. yaffe, drugs in pregnancy and lactation: a reference guide to fetal and neonatal risk, lippincott williams & wilkins2012. [14] p. baranczewski, a. stanczak, k. sundberg, r. svensson, a. wallin, j. jansson, p. garberg, h. postlind, pharmacological reports : pr 58 (2006) 453-472. [15] r. gebhardt, j.g. hengstler, d. muller, r. glockner, p. buenning, b. laube, e. schmelzer, m. ullrich, d. utesch, n. hewitt, m. ringel, b.r. hilz, a. bader, a. langsch, t. koose, h.j. burger, j. maas, f. oesch, drug metabolism reviews 35 (2003) 145-213. [16] r. yuan, s. madani, x.x. wei, k. reynolds, s.m. huang, drug metabolism and disposition: the biological fate of chemicals 30 (2002) 1311-1319. [17] c.m. habibullah, i.h. syed, a. qamar, z. taher-uz, transplantation 58 (1994) 951-952. [18] m.n. rao, b. biju, a.k. ansar, s. mujeeb, m. ramesh, n.r. srinivas, xenobiotica; the fate of foreign compounds in biological systems 33 (2003) 1233-1245. [19] h.m. jones, j.b. houston, drug metabolism and disposition: the biological fate of chemicals 32 (2004) 973-982. [20] a.p. li, drug discovery today 6 (2001) 357-366. [21] k. hoshino, h. inouye, t. unokuchi, m. ito, n. tamaoki, k. tsuji, diabete & metabolisme 2 (1976) 157-158. [22] u. yasar, e. eliasson, c. forslund-bergengren, g. tybring, m. gadd, f. sjoqvist, m.l. dahl, european journal of clinical pharmacology 57 (2001) 729-735. shaik et al. admet & dmpk 4(2) (2016) 84-90 90 [23] s.m. de morais, g.r. wilkinson, j. blaisdell, k. nakamura, u.a. meyer, j.a. goldstein, the journal of biological chemistry 269 (1994) 15419-15422. [24] n.l. kerry, a.a. somogyi, f. bochner, g. mikus, british journal of clinical pharmacology 38 (1994) 243-248. [25] k.c. patki, l.l. von moltke, d.j. greenblatt, drug metabolism and disposition: the biological fate of chemicals 31 (2003) 938-944. [26] u. food and drug administration, guidance for industry drug interaction studies study design, data analysis, and implications for dosing and labeling http://www.fda.gov/ohrms/dockets/98fr/06d-0344-gdl0001.pdf, 2006. [27] r.s. obach, a.e. reed-hagen, drug metabolism and disposition: the biological fate of chemicals 30 (2002) 831-837. [28] h.s. brown, m. griffin, j.b. houston, drug metabolism and disposition: the biological fate of chemicals 35 (2007) 293-301. [29] m.a. mohutsky, j.y. chien, b.j. ring, s.a. wrighton, pharmaceutical research 23 (2006) 654-662. [30] r.s. obach, drug metabolism and disposition: the biological fate of chemicals 27 (1999) 1350-1359. [31] y. naritomi, s. terashita, a. kagayama, y. sugiyama, drug metabolism and disposition: the biological fate of chemicals 31 (2003) 580-588. [32] k.p. kanebratt, t.b. andersson, drug metabolism and disposition: the biological fate of chemicals 36 (2008) 1444-1452. ©2016 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://www.fda.gov/ohrms/dockets/98fr/06d-0344-gdl0001.pdf http://creativecommons.org/licenses/by/3.0/ doi: 10.5599/admet.4.3.305 261 admet & dmpk 4(3) (2016) 261-268; doi: 10.5599/admet.4.3.305 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index short communication evaluation of p-glycoprotein expression in pain relevant tissues: understanding translation of efflux from preclinical species to human renu singh dhanikula 1 *, cyrla hoffert 2 , rebecca grant 2 , denis projean 1 , rosemarie panetta 3 , dajan o’donnell 2 , marie roumi 1 1 drug metabolism and pharmacokinetics, 2 translational science, 3 bioscience, astrazeneca r&d montreal, montreal, québec, canada *corresponding author: e-mail: singrenu@yahoo.com; tel.: +1 (514) 651-3602; fax: +1-111-111-112 received: may 25, 2016; revised: july 18, 2016; published: september 30, 2016 abstract various efflux transporters, such as p-glycoprotein (p-gp) are now widely accepted to have profound influence on the disposition of substrates. nevertheless, there is paucity of information about their expression and functionality in the pain relevant tissues (such as brain, spinal cord and dorsal root ganglia (drg)) across various species. therefore, our attempts were directed at evaluating p-gp expression in these tissues to understand its effect on the central nervous system (cns) disposition. as a means of characterizing the normal tissue distribution of p-gp, immunohistochemistry was performed with two antibodies (c219 and h241) directed against different epitopes of mdr1 gene. notable expression of p-gp was detected in the drg of sprague dawley rat, beagle dog, cynomolgous monkey as well as human. the expression of p-gp was observed in the cns tissues with evident species differences, the expression of p-gp in human brain and spinal cord was lower than in rats and dogs but relatively comparable to that in monkeys. however, no species related differences were seen in the expression at the drg level. double labelling using an antibody against a marker of endothelial cells confirmed that p-gp was exclusively localized in capillary endothelial cells. this study highlights the cross species similarities and differences in the expression of p-gp and thus serves as a vital step in understanding the translation of exposure of p-gp substrates to human. keywords transporter, cns, brain, spinal cord, dorsal root ganglia, immunohistochemistry introduction developing cns therapeutics has been one of the most challenging areas of drug discovery and development in the pharmaceutical industry. this has been primarily due to the restricted delivery of the drug across the physiological barriers (e.g. the blood brain barrier etc.) and also due to the lack of understanding of the differences in expression of proteins affecting the transport of drugs [1]. transport of molecules across these barriers is a highly complex phenomenon driven by passive diffusion alone or in concert with efflux or influx transport [2]. abc efflux transporters such as p-glycoprotein (p-gp) are well known to have profound influence on brain concentration of their substrates [3]. studies on understanding the contribution of p-gp efflux on cns exposure have been largely conducted in rodents, studies in human http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:singrenu@yahoo.com renu singh et al. admet & dmpk 4(3) (2016) 261-268 262 and monkey have been sparse; as a consequence, there is limited knowledge about species differences in the efflux of drug molecules across the blood-brain barrier. syvänen et al., have reported that some p-gp substrates were effluxed out of the rodent brain but not by the primate brain [4]. the observed differences in brain exposure of these compounds in rodents vs primates could be attributed to existence of species difference in p-gp function. there have been some reports of species differences in the affinity of p-gp substrates amongst cell lines transfected with p-gp from different species [5-8]. these results indicate species differences in p-gp function despite the high degree of homology in the amino acid content across species. in vivo evidence for species differences in p-gp function has been demonstrated by cutler et al., who found that higher concentrations of the p-gp inhibitor gf-120918 were needed in guinea pigs than in rats to achieve the same increase in brain concentrations of an undisclosed p-gp substrate [9]. besides differences in substrate affinity, differences in p-gp expression levels could also result in differences in p-gp transport capacity. to our knowledge, there have been no reports on the differences in expression patterns of p-gp in brain tissues across various species. in addition, functionality and expression of p-gp at blood spinal cord barrier is controversial and influence of p-gp on spinal cord exposure has not been widely studied [10-13]. because the site of action for some pain indications are located in spinal cord and dorsal root ganglia (drg), the knowledge gained from the microenvironment of the spinal cord drg could be potentially harnessed to understand the role of p-gp in manipulating drug exposure to target proteins expressed in these regions and to aid in the translation of efficacy data of compounds from animals to human [14, 15]. thus, in this study our attempts were directed at evaluating the expression of p-gp in the brain, spinal cord and drg of various species. rat, dog and monkey were selected for evaluation because they are widely used preclinical species. expression of p-gp in the tissues of the above mentioned species was evaluated using indirect immunohistochemistry and the cellular localization was determined using immunofluorescence and confocal microscopy. experimental animal tissues tissues were collected from naïve male sprague dawley rats, male beagle dogs and male cynomolgus monkeys. tissues were collected from at least three different animals for all species evaluated. all procedures were carried out as per the institutional animal care and use committee as well as animal research: reporting of in vivo experiments (arrive) guidelines. briefly, prior to tissue excision dogs were anaesthetised with ketamine, xylamine and euthanyl. the tissues were excised and immediately frozen in liquid nitrogen. cynomolgus monkey (macaca fascicularis) weighed between 2-3 kg. one of the three monkeys was naïve, the other two monkeys were given a washout period of 4 weeks prior to tissue excision. for tissue excision, monkeys were anaesthesized by ketamine and sodium pentobarbital prior to being scarified. tissues were harvested and stored in liquid nitrogen and subsequently stored at -80 °c. transverse tissue sections were cut at 14 µm using a microm hm 500 om cryostat (microm, walldorf, germany) after embedding the tissues in tissue freezing medium (jung, nussloch, germany). human tissues human tissues were obtained from the university of minnesota as per the institutional ethical guidelines and the code of ethics of the world medical association (declaration of helsinki). tissues were collected from two male and three female individuals with age ranging between 18-56 years. none of the individual had reported neurological condition. tissue sectioning was performed as outlined in the previous section. admet & dmpk 4(3) (2016) 261-268 p-gp expression in pain relevant tissues doi: 10.5599/admet.4.3.305 263 immunohistochemistry expression of p-gp protein was mapped using indirect immunohistochemistry. all the experiments were conducted on histologically well preserved fresh frozen tissues. two distinct antibodies against p-gp were employed as primary antibodies, namely, c219 (mouse monoclonal antibody) and h241 (rabbit anti-human polyclonocal antibody) (calbiochem, san diego, ca, usa and santa cruz biotechnologies, dallas, texas, usa, respectively). frozen sections were thawed for 5 min followed by fixation with ice cold acetone for 5 min. tissue sections were washed 3 times in 0.1 m tbs (tris buffered saline) and then quenched with 0.5 % hydrogen peroxide. this was followed by washing steps with 0.1 m tbs, then sections were subsequently blocked with blocking buffer (5 % normal goat serum) for 1.5 h. the blocking medium was removed and sections were incubated overnight with appropriately diluted primary antibody. as a negative control, the primary antibodies were replaced by serum to determine the degree of artifactual labelling by the secondary antibodies (biotinalyted goat anti-mouse igg or biotinalyted goat anti-rabbit igg, vector laboratories, burlingame, ca, united states). additionally, the primary antibody was incubated with blocking peptide to delineate non-specific labelling by the primary antibody. sections were then rinsed three times with 0.1 % bsa (bovine serum albumin) in 0.1 m tbs, and then incubated for 30 min with avidin-biotinylated horseradish peroxidase complex (vectastain, vector laboratories, burlingame, ca, united states). diaminobenzidine (immpact dab, vector laboratories, burlingame, ca, united states) was used as a chromogen. after the treatment, sections were washed with distilled water, dehydrated with ethanol (50, 70, 95 and 100 %) and defatted in xylene and mounted with permount mounting medium. bright field images were captured using a leica dm4000b microscope (leica microsystems, wetzlar, germany). immunofluorescence and confocal microscopy cellular localization of p-gp was assessed by immunofluorescent co-localization with the c219 antibody and an antibody raised against endothelial cell marker, von-willebrand factor (vwf, rabbit polyclonocal antibody, abcam, cambridge, united kingdom). for double labelling, tissue sections were allowed to air dry for approximately 5 min. they were then fixed with ice cold acetone (4 °c) for 5 min. after being rinsed with 0.1 m tbs thrice, slides were incubated with 0.5 % hydrogen peroxide for 10 min at room temperature. slides were rinsed again and incubated with blocking buffer (5 % normal goat serum) for 1.5 h at room temperature. c219, p-gp primary antibody was added to each slide followed by overnight incubation at 4 °c. the next day, slides were washed slides with 0.1 % bsa in 0.1 m tbs and then incubated with secondary antibody (alexa 488 goat anti-mouse, vector laboratories, burlingame, ca, united states) for c219 for 2 h. subsequently, all sections were rinsed with 0.1 % bsa in 0.1 m tbs, blocked with blocking buffer (5 % normal donkey serum) for von willebrand factor (vwf) antibody for 1.5 h at room temperature and then incubated with vwf primary antibody for 4.5 h at 4 °c. slides were rinsed with 0.1 % bsa in 0.1 m tbs and subsequently incubated with secondary antibody for vwf (cy-3 conjugated donkey anti-rabbit, jackson laboratories, west grove, pa, united states) for 45 min. slides were coversliped using aqua polymount after being rinsed with 0.1% bsa in 0.1 m tbs. fluorescent signals from double labelled sections were detected and recorded using a confocal microscope (zeiss confocal laser scanning tcs spe microscope, carl zeiss microscopy gmbh, jena, germany). post-acquisition all fluorescent images were analysed by imagej analysis software for intensity and area of p-gp labelling/expression. renu singh et al. admet & dmpk 4(3) (2016) 261-268 264 results and discussion the published literature on the expression of p-gp in the cns largely focuses on its expression in the brain tissue; consequently, there is a paucity of information on its expression in other cns tissues. in particular, the reports on the expression of this trans-membrane efflux pump in the spinal cord and drg have been sparse. the lack of understanding of species differences in expression and functionality of transporters can serve as a bottleneck in the translation of drug disposition from preclinical species to human, particularly, for compounds with the site of action in cns [16]. therefore, in an attempt to examine species differences in p-gp expression in various pain relevant tissues, we have looked at the detailed micro-anatomic localization of p-gp in the normal brain, spinal cord and drg tissues of rat, dog, monkey and human by using immunohistochemistry, immunofluorescence and confocal microscopy. to the best of our knowledge, this is the first report in the literature wherein expression of p-gp has been shown in the drg and spinal cord across several species. nonetheless, expression of p-gp in various tissues has been examined previously by immunohistochemistry and several monoclonal antibodies have been described in the literature for immunohistochemical detection of its internal or external epitopes [17-19]. in this study, antibodies recognizing two distinct epitopes were used to minimize false-positive results due to binding of antibody to proteins with similar epitopes as well as minimize false-negative results from the failure of antibody binding due to inadequate tissue processing. we have selected c219 as one of the antibodies, because it is one of the most widely used antibody for p-gp immunodetection in both research and diagnosis [20, 21]. it recognizes amino acid sequence vqeald of the cytoplasmic region of the protein, which is conserved in all cytosolic domains of p-gp from rodents to human [20, 22, 23]. h-241, a polyclonal antibody that recognizes amino acids 1040-1280 at the p-gp c-terminal cytoplasmic domain of human, mouse and rat origin was another antibody selected for this study. the specificity of both antibodies has been shown by manufacturers using western blotting, immunoprecipitation and immunohistochemistry. we have additionally validated the specificity of these antibodies by testing in multidrug resistant cell lines (mdr1-mdck and dld-1). a brown immunoreactive labelling (figure 1) representing presence of p-gp protein was formed by reaction of diaminobenzidine chromogen with horseradish peroxidase complexed to c219 and h241 antibodies. the pattern of brown staining in the tissue sections was considered to be reflective of presence of p-gp protein in tissues. overall expression of p-gp was found to be similar with both antibodies in all species and tissues examined, suggesting absence of both false-positives and negatives (data not shown). in negative controls, wherein primary antibody was omitted or when the blocking peptide was used, p-gp labelling was not observed (data not shown). this finding together with similar labelling by c219 and h241 antibodies, indicate absence of non-p-gp specific imunnolabelling. for the simplicity of presentation we have reported results of only c219 antibody in the figures. the immunohistographs of the drgs and spinal cord sections showed prominent labelling of p-gp protein in the microvessels (figure 1). the level of expression of p-gp in drg appeared to be consistent across rat, dog, monkey and human as seen in figures 1a to 1d. immunoreactivity in the spinal cord was uniformly distributed in all the lamellas of the spinal cord with no specific expression pattern (figures 2a to 2d). spinal cord sections from thoracic, lumbar and sacral regions were examined and regional differences were not detected (data not shown). however, grey matter of the spinal cord had relatively higher p-gp expression compared to the white matter. we have also assessed p-gp expression in the brain of rat, dog, monkey and human. p-gp was expressed in the cortical tissues of all species (figures 3a to 3d) as has been reported previously [24, 25]. the localization of p-gp in the brain is controversial with some studies suggesting endothelial expression [24, 25], while others reporting expression in the astrocytes or glial cells [26-28]. nonetheless, even by using antibodies against admet & dmpk 4(3) (2016) 261-268 p-gp expression in pain relevant tissues doi: 10.5599/admet.4.3.305 265 distinct p-gp epitopes (c219 and h241), immunoreactivity was exclusively detected in capillary endothelial cells; neuronal expression of p-gp in the brain tissue was not observed in the species evaluated in the present study. distribution of p-gp in the microvessels was also confirmed by immunofluorescence technique (figure 3). the observation regarding endothelial localization of p-gp could be due probably due to the fact that the tissues used in the present study were from healthy animals and non-diseased human donors with no reported neurological conditions. figure 1. distribution of p-gp in sprague dawley rat (1a, 2a, 3a, 4a, 5a, 6a), beagle dog (1b, 2b, 3b), cynomolgus monkey (1c, 2c, 3c) and human (1d, 2d, 3d) dorsal root ganglia, spinal cord and brain cortex as determined by immunohistochemical staining. c219 was used a primary antibody and biotinylated goat antimouse igg was used as a secondary antibody. images 1a to 3d were acquired at 10x, while 4a to 6a were captured at 40x. data are representative of images captured for each species and tissue (n = 3). arrows indicate p-gp labelling. drg spinal cord brain r a t 4a 5a 6a 2c 2d 3c 3d 2a 2b d r g s p in a l c o rd b ra in rat dog monkey human 1c 1d 1b 1a 3a 3b renu singh et al. admet & dmpk 4(3) (2016) 261-268 266 immunofluorescence and confocal microscopy were done to understand the micro-anatomical localization of p-gp in the tissues, mainly localization in endothelial cells was evaluated. for this purpose, endothelial cells were labelled by vwf antibody (cy-3 fluorophore) and p-gp by c219 antibody (alexa 488 fluorophore) and images were acquired at wavelengths specific to both fluorophores. rich distribution of p-gp was observed in the microvessels of various cns tissues by immunofluorescence imaging. as seen in figures 2 and 3, the green fluorescent signal from the labelling of p-gp (c219) and red fluorescent signal from endothelial cells labelling (vwf) overlayed remarkably well, suggesting largely endothelial expression of p-gp. the analysis of the acquired images indicated approximately 95 % of all endothelial cells were labelled by p-gp antibodies in both brain and spinal cord. we did not observe any marked changes in the microvessel localization of p-gp with species. however, as evidenced by intense labelling in the immunohistographs, the highest level of p-gp expression in the microvessels was observed in the drg tissues of all species examined (figures 2c, 2f & 2i). figure 2. confocal microscope images of fresh frozen human tissues labelled with c219 and vwf antibodies for p-gp and endothelial cells, respectively. cy-3 goat anti-mouse was used a secondary antibody for p-gp and alexa-fluor 488 goat anti-rabbit for endothelial cells. the green panel (left) is indicative of p-gp expression in the tissues, while the red panel (middle) displays endothelial cell expression. the right panel is the overlay of both red and green signals demonstrating co-localization. images were acquired at 10 x. data are representative of images captured for each tissue (n = 3). brain (p-gp) brain (vwf) brain (overlay) spinal cord (p-gp) spinal cord (vwf) spinal cord (overlay) drg (p-gp) drg (vwf) drg (overlay) admet & dmpk 4(3) (2016) 261-268 p-gp expression in pain relevant tissues doi: 10.5599/admet.4.3.305 267 regional differences were observed within the brain cortex in all species, with the grey matter exhibiting higher protein levels relative to the white matter. these regional differences are considered to be primarily due the extent of vascularisation of these regions as can be seen from figures 2 and 3. the intensity and area of labelling in the cns tissues evidently decreased from rat to human. for instance, the expression of p-gp in the brain cortex was relatively higher in rat than in dog followed by monkey and human. it is important to note that p-gp expression in human and monkey was comparable. a similar trend was observed in the spinal cord expression of p-gp. interestingly, there were no species differences in the p-gp expression in the drg. figure 3. fluorescence images of fresh frozen human tissues labelled with c219 and vwf antibodies, acquired with a confocal microscope at 40x. the green signal is indicative of p-gp labelling in the tissues, while the red signal is for endothelial cell expression. the overlay of the two signals is shown in the right panel. conclusions in this study we have examined the expression and distribution of p-gp in the pain relevant tissues of species widely employed in preclinical research namely sprague dawley rat, beagle dog, cynomolgous monkey and have compared to that in human. we have also shown that p-gp is expressed in the drg and spinal cord, and that there are species specific and regional differences in p-gp expression in the brain cortex and spinal cord. the findings of this study serves as a step forward in understanding species differences in transporter expression and may help build our understanding of the translation of drug exposure and efficacy from preclinical species to human. acknowledgements: authors thank francoise mennicken for her advice and technical support. references [1] f. girardin, dialogues clin neurosci. 8 (2006) 311–321. [2] w. m. pardridge, neurorx. 2 (2005) 3–14. [3] p. l. golden, g. m. pollack, j. pharm. sci. 92 (2003) 1739–1753. [4] s. syvänen, o. lindhe, m. palner, b. r. kornum, o. rahman, b. långström, g. m. knudsen, m. hammarlund-udenaes, drug metab. dispos. 37 (2009) 635–643. [5] t. ohe, m. sato, s. tanaka, n. fujino, m. hata, y. shibata, a. kanatani, t. fukami, m. yamazaki, w. m. pardridge, p. l. golden, y. s. kang and u. bickel, j. neurochem. 68 (1997) 1278–1285. [6] m. katoh, n. suzuyama, t. takeuchi, s. yoshitomi, s. asahi, t. yokoi, j. pharm. sci. 95 (2006) 2673– 2683. [7] c. q. xia, g. xiao, n. liu, s. pimprale, l. fox, c. j. patten, c. l. crespi, g. miwa, l. s. gan, mol. pharm. 3 (2006) 78–86. 3a 3b 3c renu singh et al. admet & dmpk 4(3) (2016) 261-268 268 [8] s. baltes, a. m. gastens, m. fedrowitz, h. potschka, v. kaever, w. lo¨scher, neuropharmacology 52 (2007) 333–346. [9] l. cutler, c. howes, n. j. deeks, t. l. buck, p. jeffrey, j. pharm. sci. 95 (2006) 1944–1953. [10] n.j. abbott, a.a. patabendige, d.e. dolman, s.r. yusof, d.j. begley, neurobiol. dis 37 (2010) 13–25. [11] a.d. wong, m. ye, a.f. levy, j.d. rothstein, d.e. bergles, p.c. searson, front neuroeng. 6 (2013) 7. [12] t. ashraf, a. kao, r. bendayan, adv. pharmacol. 71 (2014) 45–111. [13] w. loscher, h. potschka, neurorx 2 (2005) 86–98. [14] e. c. delange, p. g. ravenstijn, d. groenendaal, t.j. van steeg, aaps j. 7 (2005) e532–e543. [15] m. danhof, e. c. de lange, o. e. della pasqua, b. a. ploeger, r. a. voskuyl, trends pharmacol. sci. 29 (2008) 186–191. [16] x. chu, k. bleasby, r. evers, expert opin drug metab toxicol. 9 (2013) 237-252. [17] p. m. chaudhary, e. b. mechetner, i. b. roninson, blood 80 (1992) 2735-2739. [18] a. h. schinkel, r. j. arceci, j. j. smit, e. wagenaar, f. baas, m. dollé, t. tsuruo, e.b. mechetner, i. b. roninson, p. borst, int j. cancer. 55 (1993) 478-484. [19] t. grogan, w. dalton, j. rybski, c. spier, p. meltzer, l. richter, m. gleason, j. pindur, a. cline, r. and scheper, lab invest. 63 (1990) 815-824. [20] j. m. van den elsen, d. a. kuntz, f. j. hoedemaeker, d. r. rose, proc. natl. acad. sci. usa 96 (1999) 13679–13684. [21] e. okochi, t. iwahashi, t. tsuruo, leukemia 11 (1997) 1119-1123. [22] e. georges, g. bradley, j. gariepy, v. ling, proc. natl. acad. sci. usa. 87 (1990) 152-156. [23] l. jetté, b. têtu and r. béliveau, biochim biophys acta. 1150 (1993) 147-154. [24] y. matsuoka, m. okazaki, y. kitamura, t. taniguchi, j. neurobiology 39 (1999) 383–392. [25] i. sugawara, h. hamada, t. tsuruo, s. mori, jpn. j. cancer res. 81 (1990) 727–730. [26] w.m. pardridge, p.l. golden, y.s. kang, u. bickel, j. neurochem. 68 (1997) 1278-1285. [27] x. decleves, a. regina, j. l. laplanche, f. roux, b. boval, j. m. launay, j. m. scherrmann, j. neurosci. res. 60 (2000) 594–601. [28] g. lee, l. schlichter, m. bendayan,r. bendayan, j. pharmacol. exp. ther. 299 (2001) 204–212. ©2016 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ doi: 10.5599/admet.3.3.184 190 admet & dmpk 3(3) (2015) 190-202; doi: 10.5599/admet.3.3.184 open access : issn : 1848-7718 review nose-to-brain delivery of insulin for alzheimer’s disease martina stützle 1,2,3 , johannes flamm 1,3 , stefan carle 1,3 and katharina schindowski 1,3 * 1 institute of applied biotechnlogy, dept. for molecular pharmacology, biberach university of applied science, karlstrasse 11, d-88400 biberach/riss, germany 2 faculty of medicine, university of ulm, albert-einstein-allee 11, d-89081 ulm, germany 3 faculty of natural science, university of ulm, albert-einstein-allee 11, d-89081 ulm, germany *corresponding author: e-mail: zimmermann@hochschule-bc.de, tel.: +49-7351-582 498 received: may 04, 2015; revised: august 04, 2015; published: september 05, 2015 abstract the transport of small molecules, peptides and proteins via the olfactory epithelium and along olfactory and trigeminal nerve pathways from the nasal cavity to the brain is very well known and clinically established for central nervous system (cns) active drugs like oxytocin, sumatriptan or insulin. insulin is a clinically well-established biopharmaceutical with a validated function in cognition. central supply with insulin via intranasal administration improves cognition in animal models and in human, making insulin a so-called cognitive enhancer. furthermore, dysregulation of insulin is implicated in the pathogenesis of alzheimer’s disease, which is associated with lower levels of insulin in the cerebrospinal fluid and is involved in amyloid-beta (a) regulation. clinical trials with intranasal insulin implicate positive effects on learning and memory, but a massive lack of pharmacokinetic and efficacy data hamper a pharmacokinetic – pharmcodynamic relation and a possible clinical development as cognition enhancer. a lack of such data also prevents resolving the mechanisms involved in directing insulin to the central or to the peripheral compartment. here we discuss the basic mechanism of nose-to-brain delivery, evidences for intranasal insulin as cognition enhancer, medical devices for intranasal delivery and safety aspects. keywords intranasal; insulin; blood-brain-barrier; medical device; efficacy; cognition; metabolism. cns delivery – an unmet medical need the world health organization (who) estimates that more than a billion people worldwide are suffering from diseases of the central nervous system (cns; [1,2]). alzheimer’s disease (ad) is the most common neurodegenerative dementia in the industrialized world, with prevalence rates well over 30 % in the over 80-years-old population [1,2]. ad causes enormous costs to the social healthcare systems, as well as personal tragedies for the patients, families and caregivers. like most neurodegenerative diseases, ad has a poor prognosis and only symptomatic therapy is currently available. efficient treatment strategies are still limited and an aging society in demographic change presents an enormous challenge to the health systems of industrialized nations. despite the extensive research and effort to uncover the mechanism of ad pathogenesis, more or less all drug candidates failed to demonstrate significant effects on cognition in clinical trials [3,4]. a highly critical point in that context is the low central availability of drugs. the passage of most cnsactive drugs and in particular of biopharmaceuticals is massively hampered by the blood–brain barrier mailto:zimmermann@hochschule-bc.de admet & dmpk 3(3) (2015) 190-202 intranasal insulin as hope for ad? doi: 10.5599/admet.3.3.184 191 (bbb). the bbb is located at the level of the cerebral microvasculature and is important for maintaining cns homeostasis. though, the bbb restricts the access of potentially neurotoxic substances into the brain, it impedes massively the delivery of therapeutic drugs to the cns. tight junctions between the endothelial cells in the cns block the natural transport and hence seal the blood compartment from the brain compartment [5]. a limited number of essential nutrients like glucose and amino acids, co-factors like iron, and peptide hormones like insulin are actively transported across the bbb [6-9]. molecules that do not possess a specific transport mechanism have only a chance to pass the bbb via passive diffusion. while there are at least some good examples of chemical modifications for small molecule drugs to enhance their central bioavailability, nearly all of the larger molecules such as peptides and proteins fail to cross the bbb [10]. currently, all biopharmaceuticals that are used or being evaluated in the clinics for cns disease act predominantly via peripheral mechanism: e.g. anti-amyloid- (a) capturing monoclonal/polyclonal antibodies in ad [3,4] or immune cell regulating biopharmaceuticals in multiple sclerosis [11]. however, candidate biopharmaceuticals reveal promising results in cellular models or in animal models when delivered intracerebroventricular. the current state-of-the-art to deliver drugs with a low central bioavailability is intrathecal, intracerebroventricular or intraparenchymal injections that deliver directly to the cerebrospinal fluid (csf) of the cns, some of them are given chronically via an implanted intrathecal micropump (e. g. synchromed®; [12,13]). although, such delivery systems are commonly used for baclofen for the treatment of spasticity or analgetics for the treatment of cancer pain [14], these routes of administration are invasive and provide a long list of adverse events and contraindications [15,16]. hence, a safe and efficient drug delivery platform technology for cns active molecules is needed [17]. nose to brain (n2b) intranasal delivery intranasal nose to brain (n2b) delivery to the upper third of the nasal cavity bypasses the blood-brain barrier to rapidly target therapeutics to the cns along the olfactory and trigeminal neural pathways (for excellent summary see [18]). the n2b route of administration provides a nonto minimal-invasive method of bypassing the bbb. anatomy and histology of the nasal cavity the nasal cavity is divided longitudinally by the nasal septum and extends from the nostrils to the nasopharynx (roughly 12–14 cm), but has an impressively large mucosa surface area (about 160 cm 2 ) [18,19]. the frontal and lateral views are shown in figure 1. three turbinates (also called conchae) are the cause for the large surface area and their biological function is to humidify, warm and filter the inspired air. nasal secretions and inhaled particles are transported to the nasopharynx via mucociliar clearance, where they are swallowed or expectorated. importantly, the nasal mucosa provides very important immune function since countless inhaled pathogens are filtered here and transported to the nasopharynxassociated lymphoid tissue [20]. therefore, intranasal vaccinations like flumist® replace more and more injection-associated vaccinations. the nasal mucosa consists of four different epithelia: respiratory, olfactory, squamous and transitional epithelium [21]. the squamous epithelium covers the nasal vestibule from the nostrils to the anterior part of the turbinates and harbours hairs and glands. the transitional epithelium is located at the transitions between the other three types of epithelium and appears not to have a relevant role in intranasal delivery. the nasal respiratory epithelium is a pseudostratified columnar secretory epithelium that is formed by ciliated cells, goblet cells, intermediate cells and basal cells. the tissue shares high similarity to the m. stützle, j. flamm, s. carle and k. schindowski admet & dmpk 3(3) (2015) 190-202 192 respiratory epithelium of the lung and covers up to 90 % of the nasal cavity in humans and roughly 50 % in rodents [18]. through close contact with inspired air, it warms and humidifies the air and air-borne particles and pathogens are removed. different serous glands produce the nasal mucus and nasal secretion that are propelled from the ciliated cells to the nasopharynx. the mucus is well characterized and creates a mild acidic and antibacterial milieu with a ph of 5.5 6.5 [19]. the nasal respiratory epithelium is innervated by the trigeminal nerve, the fifth of the cranial nerves. numerous intraepithelial trigeminal fibres are located near the basal region of the epithelium; some of them extending outward to nearly reach the epithelial surface stopping at the line of tight junctions about 1 µm from the surface [22]. the largely branched trigeminal nerve projects mainly but not exclusively to the brainstem and is highly implicated into n2b transport [22,23]. the respiratory epithelium has a large surface and is highly perfused; hence well suitable for the systemic absorption of drugs [24]. figure 1. the anatomy of the human nasal cavity. the olfactory region permits the transport of apis to the csf and the olfactory bulb. (a) frontal view with inferior, middle and superior turbinate. (b) lateral view with cribriform plate, olfactory bulb and csf in close vicinity to the olfactory region. the nasopharynx-associated lymphatic tissue (nalt) has immunological functions and is located in lower part of the nasal cavity close to the nasopharynx. the olfactory cleft at the roof of the nasal cavity up to the superior parts of the turbinates is covered with olfactory epithelium. in humans the olfactory region comprises up to 10 % of the surface area of the nasal epithelium while accounting for about 30 % of the surface area in rodents [18]. the olfactory epithelium is formed by columnar epithelial cells, olfactory neurons, supporting cells, basal cells and bowman’s glands [19]. olfactory neurons are the only neurons having their cell bodies located in a distal epithelium and their non-motile cilia processes extend into the mucus hence being in direct contact to the environment (figure 2). turbulences at the olfactory cleft increase the residence time of the inspired air to increase the interaction of the olfactory receptors with odorants. hence, small number of odorant substances can be detected [19]. the unmyelinated axons of olfactory neurons spread through the basal lamina and form the fila olfactoria nerve bundles enclosed by olfactory ensheathing cells and olfactory nerve fibroblasts [25]. the ensheathed nerve bundles travel through the cribriform plate of the ethmoid bone into the cns and terminate at the olfactory bulb that project directly to the piriform cortex, amygdala, entorhinal cortex and olfactory nuclei [26]. compared to the respiratory epithelium the olfactory region is less perfused, but still well vasculated. the serous mucus produced by the bowman's admet & dmpk 3(3) (2015) 190-202 intranasal insulin as hope for ad? doi: 10.5599/admet.3.3.184 193 glands is not well characterized in the literature. the lamina propria underneath the olfactory epithelium harbours blood and lymphatic vessels in addition to immune cells connected to the deep cervical lymph. figure 2. postulated n2b transport mechanism: dispersed drug particles (coloured in green) are inspired with the aid of a nasal pump spray or an aerosol generator. drug molecules depositing at the olfactory region diffuse through extracellular pathways (green arrows) to the csf (see left side) or being transported via intracellular pathways to the olfactory bulb (see right side) and from here to other parts of the cns. absorption and transport mechanism of the n2b route although, numerous studies have demonstrated central effects of n2b delivered drugs in rodents, most of these studies did not show pharmacokinetic (pk) data or evidence for brain uptake. hence, the pk and transport mechanism is still far from being elucidated. it appears that intranasal delivery is a combination of different pathways [18]. it was found that crossing of the epithelial barrier can include intracellular or extracellular pathways. intracellular pathways across the olfactory epithelium include endocytosis into olfactory neurons shown for several viruses and some proteins like peroxidase or albumin [27-32] or transcytosis across supporting cells to the lamina propria. in addition, the intracellular uptake by endocytosis and transcytosis across the respiratory epithelium into trigeminal nerve processes or basal lamina, respectively, has been observed [33,34]. paracellular diffusion through epithelial tight junctions to the underlying basal lamina is the dominant extracellular transport pathways across either the olfactory or respiratory epithelia (figure 2) [18]. the permeability of the nasal epithelium is comparable to the intestinal epithelium [35]. the use of absorption enhancers like the natural polymer chitosan or lipophilic m. stützle, j. flamm, s. carle and k. schindowski admet & dmpk 3(3) (2015) 190-202 194 additives can increase the paracellular passage [36]. after uptake to or through the epithelium, different transport pathways have been implicated. n2b delivery of radioactive labelled igf-1 and interferon- in rodents and monkeys seemed to occur along trigeminal and olfactory nerves and to reach first the brainstem and olfactory bulb before the proteins are distributed over the cns [37,38]. molecules being taken-up via intracellular pathways continue their passage via anterograde axonal transport. peroxidase is likely to be purely transported with axonal transport and kinetic studies match very well with mathematical predictions taking into account the transport rate [18]. drugs that have been taken-up by transcellular diffusion and convection can be adsorbed by the lymphatic or vascular system (and thereby having a low probability of entering the brain) or diffuse to perineural or perivascular spaces and thereby enter the cranial compartment. the perineural spaces of the olfactory and trigeminal nerves seem to allow transport to the csf of the subarachnoid space [39]. however, some studies suggest that n2b transported substances may be present in the brain without being detectable in the csf [37]. mathematical predictions, however, strongly suggest that convection/bulk flow along olfactory and trigeminal nerves is the most plausible mechanism [18]. is n2b insulin appropriate as therapy for neurodegenerative diseases? evidences from in vivo and clinical studies insulin has a molecular weight of 5.8 kda and is one of the oldest recombinant biopharmaceuticals. insulin has well-known peripheral metabolic effects and lowers the serum glucose concentration, but it also plays an important central role in cognition, learning and memory [40-43]. furthermore, dysregulation of insulin is involved in the pathogenesis of ad [44]. ad is associated with lower levels of insulin in the csf and is involved in amyloid-beta (a) regulation. in cellular and animal ad models, insulin reduces aoligomer formation and protects against a-toxicity [45]. in vitro studies demonstrated that insulin stimulates the formation of the insulin degrading enzyme (ide), which is capable of degrading a [46]. moreover, the activity of glycogen-synthase kinase-3-beta, a tau kinase associated with the formation of neurofibrillary tangles, has been reported to be down-regulated in response to insulin [47]. post mortem analyses of human ad brains have documented progressive disruption in central insulin regulation. radioactive-labelled insulin was distributed widely throughout the mouse brain 1 h following intranasal administration, with the highest levels detected in the trigeminal nerve and the olfactory bulb [43]. though insulin undergoes transcytosis at the bbb, intranasal administration reached significantly higher cns levels when comparing with subcutaneous administration at the same time. intranasal insulin also slowed development of cognitive decline in different disease models [43] and improved learning and memory in wild type mice [48]. very interestingly, incretins – a group of metabolic hormones – favour insulin release in the periphery and likewise they have a comparable central activity as insulin. exendin – an agonist at the glp-1 (glucagon-like peptide-1) receptor – improved learning [49] when delivered via n2b. higher brain levels of exendin were found after intranasal administration when compared to intravenous injection [50]. clinical trials with intranasal insulin have provided positive effects in small cohorts of ad patients and mild cognitive impairment in verbal, visospatial and episodic memory [40,45,51]. unfortunately, only one study in humans gives detailed pk data (plasma and csf) after n2b delivery of peptides including insulin [52]. interestingly, the plasma levels of insulin have not been altered by intranasal administration of 10 iu insulin while a clear peak in csf levels was detectable after 30 and 80 minutes. this study has been admet & dmpk 3(3) (2015) 190-202 intranasal insulin as hope for ad? doi: 10.5599/admet.3.3.184 195 conducted with a conventional nasal spray atomizer filled with insulin formulated for subcutaneous delivery. the same group observed that insulin delivered intranasally improved memory and mood [40,53]. in addition, n2b delivered insulin decreased food intake and decreased postprandial serum insulin [54,55]. in patients with amnesic mild cognitive impairment the treatment modulated plasma levels of a and improved memory, concluding intranasal delivery of insulin as a possible treatment for ad [56]. in ad patients a chronic treatment over 21 days with intranasal insulin (20 or 40 iu novolin r) improved memory, attention and functioning [51]. in this study, patients were placed in a supine position with the head tilted back and insulin or placebo saline was administered with a needle-less syringe into alternating nostrils with a total administration volume of 400 µl. a validation of this administration technique like the group of mori et al. did [57] was not presented. moreover, no determination or estimation of the volume that reached the olfactory cleft was published. in addition, the volume that was swallowed or aspirated during the procedure was not determined. hence, a calculation of the central bioavailability e.g. insulin levels in csf samples is due to the lack of data not possible. the serum levels of insulin and glucose at baseline and 45 minutes after administration are not altered, though, the insulin levels in all three arms were rather high for fasting non-diabetic subjects. however, a pk study focusing on the administration of intranasal insulin preparations for the treatment of diabetes mellitus showed that the serum insulin levels peaked rapidly after 15 minutes with 25 iu insulin and returned to the baseline after 45 minutes [58]. this data were meanwhile confirmed without adsorption enhancer and 160 iu intranasal insulin with diabetic and control subjects [59]. in addition, the authors specified in the procedure that the patients were fasting at administration, but did not specify this for blood collection [60]. hence, the published samples collected in the latter study are insufficient for pk-pd estimation. a four-month pilot study from the same group around craft et al. was designed with three arms delivering a daily dose of 20 iu or 40 iu intranasal insulin (novolin r) or placebo twice a day with the nasal drug delivery device vianase™ [61]. the 20 iu insulin group benefited from the treatment with improved delayed memory and both groups profited from improved daily function and cerebral metabolism determined by fdg-pet. no changes in insulin csf levels were observed. again this study had several limitations that were discussed by the authors. the csf and fdg-pet data were collected for only a subset of participants. insulin levels in csf were not collected directly after insulin administration, though it is known that they drop to baseline within 1 hour after administration [52]. compared to placebo, more patients suffered from nose bleeds in the verum group. a recently published study used the long-acting insulin (detemir, levemir®) delivered intranasally via the vianase™ device twice daily over 21 days in ad patients [62]. detemir binds to albumin resulting in a prolonged release and greater parenchymal penetration. however, peripheral administered detemir is not transported across the bbb to the brain [63]. here, the 40 iu group had the largest benefit for the memory composite compared with placebo. the effect was significantly modulated by apoe4 carriage a genetic risk factor for ad, but also metabolic diseases and baseline insulin resistance, both being associated with higher baseline insulin auc (area under the curve). apoe4 negative patients showed an increased insulin resistance after 21 days treatment with 40 iu insulin detemir. no effects were reported for daily or executive functioning. two recent proof-of-concept studies evaluated the acute effects on cerebral vasoreactivity and cognition of a single 40 iu dose of intranasal insulin (novolin®) via a vianase™ device compared with placebo in type 2 diabetes (t2dm; [64,65]). across all subjects, intranasal insulin administered improved visuospatial memory and increased resting-state functional connectivity in older adults with t2dm. m. stützle, j. flamm, s. carle and k. schindowski admet & dmpk 3(3) (2015) 190-202 196 furthermore, intranasal insulin administration was well tolerated. systemic glucose levels were not significantly altered, though a tendency towards decreased levels was evident in diabetic individuals. heni et al. reported improved peripheral insulin sensitivity via hypothalamus and parasympathetic outputs after 160 iu insulin dosed with a nasal pump spray without absorption enhancer [66]. the investigators used a hyperinsulinemic-euglycemic glucose clamp with an intravenous bolus injection of 6.25 mu/kg insulin 90 minutes prior to intranasal administration and a continuous intravenous infusion of 0.25 mu/kg/min over 210 minutes. even under this conditions plasma insulin peaked after 15 minutes compared to placebo. the glucose consumption needed to keep the serum glucose euglycaemic was significantly higher in the intranasal insulin group. the authors discuss the use of n2b insulin for central insulin resistance in obesity; ad is also associated with impaired central insulin resistance. one study with 20 iu could not resolve any significant effect either on cerebral glutamate concentration and on memory [67]. several other unpublished studies are summarized in table 1. according to the published studies, intranasal insulin administration generally neither causes nasal irritation nor destroys the olfactory function and projections [48,68]. though, no statistically significant effects on lipid metabolism were observed in the 3-week detemir-study, the supplementary data of the study implicates a dose-dependent tendency associating insulin treatment with higher total cholesterol, ldl and lower hdl levels. no critical discussion was found in any study, asking about the long-term mutagenic effect of the growth factor insulin that is highly associated with increased risk of cancer. table 1. a selection of unpublished completed clinical trials using intranasal insulin for either central or peripheral metabolic activity (source: www.clinicaltrials.gov). study/sponsor dose status safety and effectiveness study of intranasal insulin glulisine on cognitive and memory in mild-mod ad patients. (nct01436045) / healthpartners institute for education and research 20 iu completed (2013) memory and insulin in early alzheimer's disease (nct00581867) / university of kansas 40 iu completed (2013) safety study of intranasal insulin in type 1 diabetes and diabetic peripheral neuropathy (nct01469559) / university of calgary 20 iu completed (2012) a study to evaluate the effect of nasal insulin on postprandial glycemic control in type 2 diabetic patients (nct00624767) / nastech pharmaceutical company, inc. 30 iu completed (2008) medical devices and dosage forms for intranasal delivery one of the first studies by born et al. delivered insulin by using a conventional manual nasal pump spray atomizer [52] that generates droplets of roughly 50 to 100 µm dependent from many different factors like pressure and distance to the nozzle [69,70]. hence, the handling of device by different patients or caregivers may vary easily and result in deposition at different sites of the nasal mucosa. moreover, aerosols delivered by nasal pump sprays cleared rather quickly: about 50 % was cleared after 15 minutes, and after 6 hours, less than 5 % of activity was retained in the nose [69]. the above-mentioned clinical trials at washington university administered insulin via aerosols http://www.clinicaltrials.gov/ admet & dmpk 3(3) (2015) 190-202 intranasal insulin as hope for ad? doi: 10.5599/admet.3.3.184 197 generated with the electronic atomizer vianase™ (kurve technology inc.), a vortex-propelled nebulizer system. the vianase device generates nebulized particles between 9 and 11 μm in size to be inhaled with an occlusive nosepiece over 2 minutes [69,71]. according to the manufacturer, this electric device covers a higher content of the nasal mucosa with drug-containing droplets compared to nasal pump sprays without targeting the lungs [71,72]. a bidirectional breath-powered nasal delivery platform is optimist™ by optinose. the novel technology overcomes undesired pulmonary delivery by a breath-actuation mechanism. the mechanism is extraordinarily simple: the device has a nosepiece and a mouthpiece, which is inserted into one nostril and the patient blows into the mouthpiece. exhalation pressure releases drug as (solid or fluid) particles through the nosepiece into the nasal cavity. exhalation closes the soft palate and the air flow can exit through the other nostril as bidirectional flow reaching both parts of the nasal septum [73]. through the closed soft palate drug particles cannot enter the lungs [74]. optimist™ generates droplets of 43 μm diameter and targets larger initial and cumulative deposition in the upper posterior sector of the nasal cavities and significantly lower deposition in the anterior segment. the device is currently evaluated in a clinical study for the use with the peptide oxytocin. pressure operated devices aim to work as precision olfactory delivery (pod®, impel neuropharma) devices to deliver drugs to the upper nasal cavity for direct n2b transport. according to the manufacturer the pod device technology results in over 50 % deposition at the olfactory region and enhanced uptake into the cns in comparison to other nasal delivery devices, which are quickly cleared by respiratory epithelium and absorbed into systemic circulation [75]. in addition, standard medical nebulizers can be used for aerosol generation and delivery into the nasal cavity. however, most of them have been developed to target the lower airways, but newer models engineered for the delivery of the paranasal sinuses (e. g. pari sinus™) might be suitable for n2b delivery. the influence of liposomes, nanoparticle and gels as dosage forms for intranasal delivery is summarized in table 2. intranasal insulin for metabolic diseases n2b insulin delivery appears to be a clinically safe application method for the dissociation of central and peripheral insulin effects [76]. an intranasal insulin spray containing the adsorption enhancer cyclopentadecalactone (cpe-215) was developed as nasulin™ for the needle-free regimen of diabetes. the programme was stopped after disappointing results of a phase 2a proof-of-concept clinical trial [77]. the primary objective to demonstrate that subjects receiving nasulin™ would achieve a larger increase from baseline in the mean proportion of time spent in euglycemia than those receiving placebo failed statistical significance. like in the latter ad trials, no critical safety signals were detected with nasulin™. the most common adverse events were those attributable to administration site reactions associated with the nasal route of delivery, the majority of which were mild. the percentage of subjects reporting hypoglycemia was similar between both the nasulin™ and placebo groups. nevertheless, it seems that intranasal insulin is yet not dead. the fda gave a black box warning to the pulmonary administered insulin exubera® since clinical studies reported a four-fold increase in the incidence of lung cancer among patients with a history of smoking. in 2007 exubera® was discontinued and shifted back the interest to nasal insulin administration with very recent studies on beneficial metabolic effects of central insulin action [59,66,78]. several companies continue their nasal insulin development programmes for the regimen of diabetes, all of them m. stützle, j. flamm, s. carle and k. schindowski admet & dmpk 3(3) (2015) 190-202 198 using different absorption enhancers like dodecyl-β-d-maltoside (aegis therapeutics inc.; [79]), micro crystalline cellulose (snbl; [80]) and criticalsorb™ (critical pharmaceuticals; [81]). table 2. dosage forms used in intranasal delivery. liposomes study compound formulation particle size outcome source preclinical/ rodents galantamine hbr soy phosphatidylcholine, cholesterol, propylene glycol 112 ± 8 nm 2.8-fold increase of auc compared to free drug via n2b; 3.4-fold increase compared to oral delivery [82] preclinical/ rodents rivastigmine cholesterol, soy lecitihin 10 ± 2.8 µm 3-fold increase of auc compared to free drug via n2b; 5.5-fold increase compared to oral delivery [83] preclinical/ rodents ovalbumin as surrogate for protein therapeutics dioleoylphosphatidylcholine, cholesterol, stearylamine 299 ± 26.4 nm >4-fold higher aucbrain/aucblood for the liposomal formulation compared to the pbs preparation [84] nanoparticles study compound formulation particle size outcome source preclinical/ rodents insulin (for blood glucose management) glycopolymer poly(2-lactobionamidoethylmethacrylate-r-3acrylamidophenylboronic acid) p(lama-r-aapba) 2:1 289.4 ± 2.5 nm 80% decrease of peripheral blood glucose levels with nanoparticles, 4-fold increase compared to pbs formulation [85] gels study compound formulation outcome source preclinical (rodents) and clinical (humans) study insulin (for blood glucose management) carbopol 934p and hydroxypropyl methylcellulose prolonged effect on blood glucose levels in rodents; decrease of blood glucose levels by 20% with nasal gel, no control group for comparison. [86] preclinical/ rodents insulin (for blood glucose management) chitosan and polyvinyl alcohol (pva) prolonged 60% decrease of peripheral blood glucose levels with gel formulation [87] conclusive remarks about the lack of valuable pk-pd studies nasal insulin delivery using such above mentioned absorption enhancers shows rapid delivery of insulin to the serum [58,88]. in addition, comparative studies are needed to resolve the mechanisms needed to target insulin or other biopharmaceuticals either to the periphery or to the brain. for intranasal insulin that includes insulin and glucose levels in serum and csf to calculate pk-pd relations. the bioavailability and steady-state kinetics of intranasal insulin administered over 3 weeks to 4 months has hardly been evaluated. for a convincing safety profile this data should be provided. in addition, safety aspects and adverse events like hypoglycaemia or dyslipidaemia should be adequately discussed in the light of a chronic treatment. an excess of insulin that is distributed to the vasculature might mediate such adverse events. admet & dmpk 3(3) (2015) 190-202 intranasal insulin as hope for ad? doi: 10.5599/admet.3.3.184 199 though, any ad study that does not fail causes euphoria, the use and the transport mechanism of intranasal insulin with and without additives needs to be analysed very tightly and discussed more critically in relation to its safety profile. the number of ad cases increases with the demographic change and we need to act now. future ad patients are dependent on our analytical knowledge, pharmaceutical expertise and scientific creativity to develop candidates like n2b delivered insulin to a safe, suitable and validated drug therapy. a recent empirical analysis uncovered the massive lack of quantitative data in published n2b scientific publications: only 3 % of the studies determine and publish quantitative data like bioavailability or similar [89]. however, the development of a successful n2b technique requires quantitative pk-pd data to engineer a safe and efficient delivering strategy. acknowledgements: this analysis was supported by ulm and biberach joint graduate school in pharmaceutical biotechnology funded by the baden-württemberg state ministry of science, research and arts, from the graduate school “molecular medicine” of ulm university (to martina stützle) and from a fellowship from the stiftung der deutschen wirtschaft (to johannes flamm). special thanks to anne-sophie and jürgen zimmermann for support and help with the manuscript. references [1] world_health_organization_and_alzheimer’s_disease_international, dementia: a public health priority, 2012. [2] neurological disorders: public health challenges, http://www.who.int/mental_health/neurology/neurodiso/en/index.html. [3] j. moreth, c. mavoungou, k. schindowski, immunity & ageing 10:18 (2013). [4] c. schindowski, j. zimmermann, k. schindowski, degenerative neurological & neuromuscular disease 4 (2014) 121-130. [5] t.s. reese, m.j. karnovsky, the journal of cell biology 34 (1967) 207-217. [6] w.a. banks, bmc neurology 9 suppl 1 (2009) s3. [7] a.p. dick, s.i. harik, a. klip, d.m. walker, proceedings of the national academy of sciences of the united states of america 81 (1984) 7233-7237. [8] w.a. jefferies, m.r. brandon, s.v. hunt, a.f. williams, k.c. gatter, d.y. mason, nature 312 (1984) 162-163. [9] w.a. banks, regulatory peptides 149 (2008) 11-14. [10] d.j. begley, pharmacology & therapeutics 104 (2004) 29-45. [11] h. gensicke, d. leppert, o. yaldizli, r.l. lindberg, m. mehling, l. kappos, j. kuhle, cns drugs 26 (2012) 11-37. [12] r.m. levy, d. salzman, neurosurgical management of pain, (eds. r.b. north, r.m. levy) springer 1997, p. 302-324. [13] t.j. lamer, mayo clinic proceedings 69 (1994) 473-480. [14] r. gilmartin, d. bruce, b.b. storrs, r. abbott, l. krach, j. ward, k. bloom, w.h. brooks, d.l. johnson, j.r. madsen, j.f. mclaughlin, j. nadell, journal of child neurology 15 (2000) 71-77. [15] s. sgouros, c. charalambides, p. matsota, i. tsangaris, g. kostopanagiotou, pediatric neurosurgery 46 (2010) 62-65. [16] p. maino, e. koetsier, r.s. perez, regional anesthesia and pain medicine 39 (2014) 434-437. http://www.who.int/mental_health/neurology/neurodiso/en/index.html m. stützle, j. flamm, s. carle and k. schindowski admet & dmpk 3(3) (2015) 190-202 200 [17] e. neuwelt, n.j. abbott, l. abrey, w.a. banks, b. blakley, t. davis, b. engelhardt, p. grammas, m. nedergaard, j. nutt, w. pardridge, g.a. rosenberg, q. smith, l.r. drewes, the lancet neurology 7 (2008) 84-96. [18] j.j. lochhead, r.g. thorne, advanced drug delivery reviews 64 (2012) 614-628. [19] r.l. doty, handbook of olfaction and gustation, third edition. john wiley & sons inc., hoboken, new jersey, 2015. [20] h. kiyono, s. fukuyama, nature reviews immunology 4 (2004) 699-710. [21] j.r. harkema, s.a. carey, j.g. wagner, toxicologic pathology 34 (2006) 252-269. [22] t.e. finger, v.l. st jeor, j.c. kinnamon, w.l. silver, the journal of comparative neurology 294 (1990) 293-305. [23] f. bojsen-moller, the journal of comparative neurology 159 (1975) 245-256. [24] h.r. costantino, l. illum, g. brandt, p.h. johnson, s.c. quay, international journal of pharmaceutics 337 (2007) 1-24. [25] p. field, y. li, g. raisman, journal of neurocytology 32 (2003) 317-324. [26] s.t. carmichael, m.c. clugnet, j.l. price, the journal of comparative neurology 346 (1994) 403-434. [27] r.g. thorne, c.r. emory, t.a. ala, w.h. frey, brain research 692 (1995) 278-282. [28] r.l. doty, annals of neurology 63 (2008) 7-15. [29] k. kristensson, y. olsson, acta neuropathologica 19 (1971) 145-154. [30] r.d. broadwell, b.j. balin, the journal of comparative neurology 242 (1985) 632-650. [31] h. baker, r.f. spencer, experimental brain research 63 (1986) 461-473. [32] k. kristensson, nature reviews. neuroscience 12 (2011) 345-357. [33] f. anton, p. peppel, neuroscience 41 (1991) 617-628. [34] a.m. deatly, a.t. haase, p.h. fewster, e. lewis, m.j. ball, neuropathology and applied neurobiology 16 (1990) 213-223. [35] k. hosoya, h. kubo, h. natsume, k. sugibayashi, y. morimoto, s. yamashita, biopharmaceutics & drug disposition 14 (1993) 685-695. [36] s.s. davis, l. illum, clinical pharmacokinetics 42 (2003) 1107-1128. [37] r.g. thorne, g.j. pronk, v. padmanabhan, w.h. frey, neuroscience 127 (2004) 481-496. [38] r.g. thorne, l.r. hanson, t.m. ross, d. tung, w.h. frey, neuroscience 152 (2008) 785-797. [39] m. johnston, a. zakharov, c. papaiconomou, g. salmasi, d. armstrong, cerebrospinal fluid research 1 (2004) 2. [40] c. benedict, m. hallschmid, a. hatke, b. schultes, h.l. fehm, j. born, w. kern, psychoneuroendocrinology 29 (2004) 1326-1334. [41] e. shemesh, a. rudich, i. harman-boehm, t. cukierman-yaffe, the journal of clinical endocrinology and metabolism 97 (2012) 366-376. [42] l. plum, m. schubert, j.c. bruning, trends in endocrinology and metabolism: tem 16 (2005) 59-65. [43] g.j. francis, j.a. martinez, w.q. liu, k. xu, a. ayer, j. fine, u.i. tuor, g. glazner, l.r. hanson, w.h. frey, c. toth, brain: a journal of neurology 131 (2008) 3311-3334. [44] s. craft, g.s. watson, the lancet neurology 3 (2004) 169-178. [45] j. freiherr, m. hallschmid, w.h. frey, 2nd, y.f. brunner, c.d. chapman, c. holscher, s. craft, f.g. de felice, c. benedict, cns drugs 27 (2013) 505-514. [46] s. craft, archives of neurology 66 (2009) 300-305. [47] m. hong, v.m. lee, the journal of biological chemistry 272 (1997) 19547-19553. [48] d.r. marks, k. tucker, m.a. cavallin, t.g. mast, d.a. fadool, the journal of neuroscience: the official journal of the society for neuroscience 29 (2009) 6734-6751. admet & dmpk 3(3) (2015) 190-202 intranasal insulin as hope for ad? doi: 10.5599/admet.3.3.184 201 [49] m.j. during, l. cao, d.s. zuzga, j.s. francis, h.l. fitzsimons, x. jiao, r.j. bland, m. klugmann, w.a. banks, d.j. drucker, c.n. haile, nature medicine 9 (2003) 1173-1179. [50] w.a. banks, m.j. during, m.l. niehoff, the journal of pharmacology and experimental therapeutics 309 (2004) 469-475. [51] m.a. reger, g.s. watson, w.h. frey, 2nd, l.d. baker, b. cholerton, m.l. keeling, d.a. belongia, m.a. fishel, s.r. plymate, g.d. schellenberg, m.m. cherrier, s. craft, neurobiology of aging 27 (2006) 451-458. [52] j. born, t. lange, w. kern, g.p. mcgregor, u. bickel, h.l. fehm, nature neuroscience 5 (2002) 514516. [53] w. kern, j. born, h. schreiber, h.l. fehm, diabetes 48 (1999) 557-563. [54] c. benedict, s. brede, h.b. schioth, h. lehnert, b. schultes, j. born, m. hallschmid, diabetes 60 (2011) 114-118. [55] c. benedict, w. kern, b. schultes, j. born, m. hallschmid, the journal of clinical endocrinology and metabolism 93 (2008) 1339-1344. [56] m.a. reger, g.s. watson, p.s. green, c.w. wilkinson, l.d. baker, b. cholerton, m.a. fishel, s.r. plymate, j.c. breitner, w. degroodt, p. mehta, s. craft, neurology 70 (2008) 440-448. [57] e. mori, c. merkonidis, m. cuevas, v. gudziol, y. matsuwaki, t. hummel, european archives of otorhino-laryngology 272 (2015) doi:10.1007/s00405-015-3701-y. [58] a.c. leary, m. dowling, k. cussen, j. o'brien, r.m. stote, journal of diabetes science and technology 2 (2008) 1054-1060. [59] s. gancheva, c. koliaki, a. bierwagen, p. nowotny, m. heni, a. fritsche, h.-u. häring, j. szendroedi, m. roden, diabetes 64 (2015) 1966-1975. [60] m.s. dhamoon, j.m. noble, s. craft, neurology 72 (2009) 292-293; author reply 293-294. [61] s. craft, l.d. baker, t.j. montine, s. minoshima, g.s. watson, a. claxton, m. arbuckle, m. callaghan, e. tsai, s.r. plymate, p.s. green, j. leverenz, d. cross, b. gerton, archives of neurology 69 (2012) 29-38. [62] a. claxton, l.d. baker, a. hanson, e.h. trittschuh, b. cholerton, a. morgan, m. callaghan, m. arbuckle, c. behl, s. craft, journal of alzheimer's disease: jad 44 (2015) 897-906. [63] w.a. banks, j.e. morley, j.l. lynch, k.m. lynch, a.d. mooradian, peptides 31 (2010) 2284-2288. [64] v. novak, w. milberg, y. hao, m. munshi, p. novak, a. galica, b. manor, p. roberson, s. craft, a. abduljalil, diabetes care 37 (2014) 751-759. [65] h. zhang, y. hao, b. manor, p. novak, w. milberg, j. zhang, j. fang, v. novak, diabetes 64 (2015) 1025-1034. [66] m. heni, r. wagner, s. kullmann, r. veit, h. mat husin, k. linder, c. benkendorff, a. peter, n. stefan, h.-u. häring, h. preissl, a. fritsche, diabetes 63 (2014) 4083-4088. [67] mechanisms of insulin facilitation of memory, https://http://www.clinicaltrials.gov/ct2/show/results/nct01145482 (04/28/2015). [68] s. khafagy el, m. morishita, y. onuki, k. takayama, advanced drug delivery reviews 59 (2007) 15211546. [69] s. albu, drug design, development and therapy 6 (2012) 125. [70] k. inthavong, m.c. fung, w. yang, j. tu, journal of aerosol medicine and pulmonary drug delivery 28 (2015) 59-67. [71] p.g. djupesland, neurology 71 (2008) 864; author reply 864. [72] deposition comparison: a scintigraphic study: vianasetm with controlled particle dispersiontm (cpd), spray pump and nebulizer deposition comparison, http://www.kurvetech.com/pdf/depositioncomp_scintigraphicstudy.pdf. [73] p.g. djupesland, a. skretting, m. winderen, t. holand, the laryngoscope 116 (2006) 466-472. http://www.clinicaltrials.gov/ct2/show/results/nct01145482 http://www.kurvetech.com/pdf/depositioncomp_scintigraphicstudy.pdf m. stützle, j. flamm, s. carle and k. schindowski admet & dmpk 3(3) (2015) 190-202 202 [74] p.g. djupesland, a. skretting, m. winderen, t. holand, journal of aerosol medicine 17 (2004) 249259. [75] j.d. hoekman, r.j. ho, anesthesia and analgesia 113 (2011) 641. [76] a. kupila, j. sipila, p. keskinen, t. simell, m. knip, k. pulkki, o. simell, diabetes/metabolism research and reviews 19 (2003) 415-420. [77] cpex pharmaceuticals announces preliminary results from its phase 2a clinical trial of nasulin, http://www.bloomberg.com/apps/news?pid=newsarchive&sid=af8dtg4w0xt4 (03/31/2015). [78] s. dash, c. xiao, c. morgantini, k. koulajian, g.f. lewis, diabetes 64 (2015) 766-774. [79] e.t. maggio, us8772231 b2 (2006). [80] licensing agreement for nasal drug delivery technology platform – nasal dihydroergotamine for treatment of migraine headache, http://snbl.com/press-release/licensing-agreement-for-nasaldrug-delivery-technology-platform-nasal-dihydroergotamine-for-treatment-of-migraine-headache/ (04/16/2015). [81] criticalsorb is a proprietary absorption promoter that has the potential to provide a viable alternative to injection for a wide variety of biological and challenging small molecule drugs., http://www.criticalpharmaceuticals.com/technology/criticalsorb (04/16/2015). [82] w. li, y. zhou, n. zhao, b. hao, x. wang, p. kong, environmental toxicology and pharmacology 34 (2012) 272-279. [83] k. arumugam, g.s. subramanian, s.r. mallayasamy, r.k. averineni, m.s. reddy, n. udupa, acta pharmaceutica 58 (2008) 287-297. [84] m.m. migliore, t.k. vyas, r.b. campbell, m.m. amiji, b.l. waszczak, journal of pharmaceutical sciences 99 (2010) 1745-1761. [85] c. zheng, q. guo, z. wu, l. sun, z. zhang, c. li, x. zhang, european journal of pharmaceutical sciences 49 (2013) 474-482. [86] r. d’souza, s. mutalik, m. venkatesh, s. vidyasagar, n. udupa, aaps pharmscitech 6 (2005) e184e189. [87] a.k. agrawal, p.n. gupta, a. khanna, r.k. sharma, h.k. chandrawanshi, n. gupta, u.k. patil, s.k. yadav, die pharmazie 65 (2010) 188-193. [88] r. stote, m. miller, t. marbury, l. shi, p. strange, journal of diabetes science and technology 5 (2011) 113-119. [89] l. kozlovskaya, m. abou-kaoud, d. stepensky, journal of controlled release 189 (2014) 133-140. ©2015 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://www.bloomberg.com/apps/news?pid=newsarchive&sid=af8dtg4w0xt4 http://snbl.com/press-release/licensing-agreement-for-nasal-drug-delivery-technology-platform-nasal-dihydroergotamine-for-treatment-of-migraine-headache/ http://snbl.com/press-release/licensing-agreement-for-nasal-drug-delivery-technology-platform-nasal-dihydroergotamine-for-treatment-of-migraine-headache/ http://www.criticalpharmaceuticals.com/technology/criticalsorb http://creativecommons.org/licenses/by/3.0/ manuscript doi: 10.5599/admet.3.4.257 363 admet & dmpk 3(4) (2015) 363-365; doi: 10.5599/admet.3.4.257 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index opinion ethical considerations of preclinical testing allen goldenthal* faculty of health sciences, university of macau, taipa, macau *corresponding author: e-mail: aegoldenthal@umac.mo; tel.: +853-623-75280 received: november 17, 2015; revised: december 02, 2015; published: december 30, 2015 abstract the numbers of animal tests being conducted are on a sharp incline. much of this increase is directly due to our ability to generate transgenic models and knock-outs, thereby increasing the validity of the animal model but not necessarily correlating directly with any translational medical benefits to the human counterpart. in spite of our best efforts, there still exist species differences that prevent the application directly from animal to human, and in some examples having a completely different and adverse effect from that seen in the animal model. there are several ways in which we can improve the opportunity for a positive test outcome and at the same time reduce the animal usage which is associated with our current animal testing practices. the benefit of the 3r’s is that they encourage us not only to avoid wastage of life but that they require us to provide considerable foresight and extrapolated thought before directly engaging in the preclinical testing phase. keywords animal models; in vivo; regulatory; pharmaceuticals; ethics introduction it has been estimated that of every 5000 new drug development projects that make it through initial computer modelling and in-vitro testing, only 5 will ever successfully pass preclinical tests which are mandatory before clinical testing in humans. moreover, of those drug developments that successfully pass the preclinical testing phase, only 5 % ever make it to the market as a licensed treatment for humans. essentially, that suggests that there is less than a 0.1 % chance of a successful drug product being developed from inception to market. although these statistics justify why drug companies must charge accordingly to recover their cost outputs which includes all the costs of the majority of failed projects, it cannot ethically justify the number of animals that are sacrificed in order to achieve this high failure rate. there is this vast chasm between preclinical testing in animals and the eventual successful outcome of clinical trials in humans into which the majority of these drug trials disappear. in order to compensate or reduce this failure rate, additional safeguards must be placed on which drugs do make it to the animal testing phase to at least ensure that the percentage of drugs successfully passing preclinical trials is much higher than the current 0.1 %. we cannot do much at our current level of testing and knowledge to increase the translational level of success between animals and humans, since that is still not fully understood why a drug is successful in one species and not another, but we certainly can narrow the gap http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:aegoldenthal@umac.mo allen goldenthtal admet & dmpk 3(4) (2015) 363-365 364 between drugs that appear to have great potential in the computer modelling and in-vitro stage, only to fail when they are introduced into animals. major challenges the issue with safety and toxicity testing in a preclinical trial centres on dosage. when examining the weight ratio of a mouse at 25 g versus the average human at 60 kg, then the ratio of 1:2400 would suggest that we must administer to the mouse 1/2400 of the anticipated human dose on a dose/kg basis. this would require working with extremely small quantities which may in turn be so small that they are ineffective even on an animal of such small size. that certain drugs may have a threshold level before they have any efficacy is a commonly found factor, but in terms of trying to identify what that threshold might be, often numerous pilot studies are required which not only involve time and expense on behalf of the principal investigator but may be exorbitant in the number of animals sacrificed in order to find that level. even so, once established, will this threshold really correlate to the quantity that must be administered to humans? the dosages may be so high, that the equivalent dosage on a dose/kg basis may far exceed the safe levels for human administration. sensitivity to particular drugs and treatments will vary between species, simply because we are different as even the best humanized models can confirm. but the danger in relying on the dose established in the animal model as a guideline for even administering a much lower dose in humans is a minefield as was clearly evident in the limited clinical trial of tgn 1412, using a mousehuman hybrid antibody that resulted in a cytokine storm event [1]. what that particular case brings to mind is that the animal model itself was incapable of manifesting the adverse response evident in humans and therefore left a huge gap of possible or unknown consequences that only became clear when administered into human subjects. we can extrapolate this same concept into animal models of disease such as diabetes both type i and type ii, which although similar to their counterparts in humans, are not identical, and as we have come to appreciate, will not respond in the same manner in both species to the same drug. as can be heard in many institutions, “we have some excellent cures for diabetes in mice but unfortunately they have proven of little value in humans.” as some of these models are induced, rather than having an identical genetic predisposition or aetiology as in the human subject, even though we will classify them as type i or type ii diabetes they are in fact mimics of the disease but not the identical human disease. pre-test requirements and resolutions as such, it is important that prior to conducting a preclinical test for a drug, we must first demonstrate that the model we will use has been proven or shown to be a reflection of the human counterpart and therefore there is a higher value to any data generated from the study. this requires that we have both invitro and computational studies to verify the evidence and increase the probability of a successful outcome. without this initial testing, moving forward into animal testing should not be considered. secondly, we must have all the information at hand concerning the drug being tested, or a similar drug from the same family of compounds, especially where it concerns previous animal models tested. if similar or related class of drugs were reported in preclinical trials, having failed to move forward into human trials, or in fact failing in human trials, then it is ethically important to make the argument as to why the particular drug now being tested will have a greater chance to succeed than it did in the prior tests. if such an argument cannot be made and subsequently approved, then once again animal testing should not be considered. admet & dmpk 3(4) (2015) 363-365 ethical considerations of preclinical testing doi: 10.5599/admet.3.4.257 365 in this vein, a third issue is that much of preclinical research outcomes are never published and that distorts the evidence base on which to make decisions. when we say that researchers should be interpreting preclinical outcomes in light of how similar drugs have performed in the past, it is very difficult to do that if one simply cannot consult the record of prior drugs because the results were not published as a corporate or institutional decision. we would want to see measures going forward that encourage and incentivize publication of all preclinical results, even if those results were negative, as this is vitally important to reduce redundancy of experiments with no chance of success, a situation which is frequently happening within our current framework. fourthly, we should perform an up-down study using a minimum amount of animals in order to establish what we believe to be the noel (no observable effects level), loel (lowest observable effects level) and mtd (maximum tolerated dose). by pairing animals at a particular dose then observing for effects and based on those results determining if the next pair of animals receive a higher or lower administrated dosage, then we can establish more effectively the groupings to be used for any preclinical study. finally, we must give consideration to intercession time points when establishing the protocol or any preclinical studies. this practice will identify key critical decision points where the study should be terminated for reasons of either severe adverse effects or no indication of any desired effects. these require that a scale be predetermined of various levels of adverse effects, as well as one on determining data to quantify that there is absolutely no beneficial outcome before the preclinical study commences. conclusions it must be remembered that preclinical studies are all operated under the umbrella of good laboratory practices (glp) and therefore any principal investigator conducting such studies must be familiar with the glp and that their assistants are also equally aware of the regulatory requirements for conducting a preclinical study before they participate in the trial. straying from the glp not only places the study at risk but also any data generated from such a study since it may not accurately reflect the actual effects of the drug but may be unduly influenced and adversely affected by both poor husbandry and animal management. the obtained results from animal testing must undergo several checks for reliability before presuming there will be a correlating beneficial effect in humans. certain adverse effects of the therapy may go completely undetected because we do not have the ability to recognize subclinical or monitor low incidence effects in animals such as muscle aches, low grade fever, and even psychological manifestations that might be occurring within the animal even when being observed [2]. hence, it is imperative that we improve our skills in animal modelling, test design, and effect recognition, if we are to attain the benefits that animal testing can provide. references [1] h. attarwala. j. young pharm. 2(3) (2010) 332-336. [2] v. baumans. gene therapy 11 (2004) s64-s66. ©2015 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ manuscript doi: 10.5599/admet.1.4.19 76 admet & dmpk 1(4) (2013) 76-81; doi: 10.5599/admet.1.4.19 open access: issn: 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index opinion on the evaluation of the safety aspects of nanomaterials in medical devices – a regulatory perspective eva wu, christopher chan* and albert t.w. li office of medical device evaluation, center for measurement standards, industrial technology research institute, taiwan *corresponding author: e-mail: cchan@itri.org.tw; tel.: +886-03-5743836; fax: +886-03-5732299 received: november 20, 2013; november 30, 2013; published: december 16, 2013 abstract nanotechnology is widely used in many aspects of the design and manufacture of medical devices. to date, many of these new medical devices, referred to as nano medical devices, have been submitted to health authorities for premarket regulatory review. there are ongoing discussions between medical device manufacturers and regulatory authorities regarding the standards and methods required for the evaluation process. taking into consideration aspects including nano-toxicology and biocompatibility, the clinical effects of nanotechnology and risk management, there are issues yet to be resolved. in this article, we will discuss nano medical device safety from the regulatory control point of view. keywords nanotechnology; nanomaterial; regulatory; medical device; risk management; usability. 1. introduction there is no single official definition of nanotechnology in the application of medical use, but the us fda refers to nanotechnology as engineered materials or end products or the exhibition of properties or phenomena that are attributable to at least one of the dimensions in the range of 1 to 100 nanometers in general [1]. due to these dimensions, the physical and chemical, and thereby the biological properties of these materials may not necessarily be the same as the same material in the bulk configuration. the use of nanomaterial has been successfully explored in many areas such as chemical materials, electronics, it and the biomedical industry. in taiwan, nano-products are becoming more and more common in the domestic market, including nano-rags, nano-clothing, nano-silver deodorant liquids, ceramic nano-energy cups, nano-toilets, etc. the advertising of these products infers that they are better than their traditional counterparts with additional benefits from nanotechnology. it seems that all products designated as "nano" will offer some sort of magical and, more importantly, beneficial effects. consequently, the ministry of economic affairs of taiwan has commissioned the center for measurement standards of the industrial technology research institute (itri) to assess and verify the specifications of nano products. a “nanomark” will be granted to those nano products that meet these criteria. the purpose of this assessment is to ensure that products issued with the “nanomark” do in fact possess http://www.pub.iapchem.org/ojs/index.php/admet/index admet & dmpk 1(4) (2013) 76-81 nanomaterials safety doi: 10.5599/admet.1.4.19 77 certain pre-defined nanotechnology characteristics and scientific evidence to support claims regarding pollution control, antibacterial activity, deodorant properties, wear-resistance and other characteristics as described in the product specification, in order to protect the interests of consumers. with regard to nano medical devices (medical devices that involve the use of nanomaterials or nanotechnology in general), safety and efficacy are of paramount importance. it is necessary to point out that any medical device, in particular those that employ novel technology like nanotechnology in which the toxicity of nanomaterials is not yet fully understand or there is little history of a comprehensive evaluation of the technology, their performance and safety should be properly evaluated. it is for this reason that nano medical devices have raised the regulatory concerns of the health authorities regarding safety and efficacy. for example, nano-fiber socks that are advertised as possessing antibacterial, deodorant and other functions can attract the attention of customers and increase the selling price of the product. however, the use of such products for medical purposes has garnered the attention from the health authorities, resulting in additional safety assessments on the nanomaterial used before granting marketing authorization. in this context, it is crucial to ensure that the nanomaterial creates no new hazards to human health or the environment compared with the traditional product, and the product offers the features depicted in the specifications. for that, functions such as antibacterial and deodorant activity may not seem critical compared with a safety evaluation. the situation is further complicated by developments in nanotechnology, whereby nanotechnology applications in different disciplines are integrated into a single purpose, like combination products that combine any two or even three of nanotechnology utilizing medical devices, drugs and biologics, making product classification difficult. consequently, the product review can be problematic. a systematic and multi-disciplinary approach is therefore necessary, and particular attention should be paid during the regulatory review of these products. 2. medical applications of nanotechnology nanomedicine generally refers to the clinical application of nanotechnology to provide a novel therapy or diagnosis, including diagnostic equipment, drug delivery devices, gene therapy and tissue engineering. in 2003, miller recommends the definition of nanomedicine as "the monitoring, repair, construction and control of human biological systems at the molecular level, using engineered nanodevices and nanostructures" [2]. as nanotechnology manipulates nanostructures at the molecular level, it could potentially generate substantial changes in properties such as the melting point, electrical, magnetic, optical or surface properties and chemical activity and thereby have an impact on the biological properties within a limited space. these modifications could make products that are not just smaller but also lighter, more flexible, mechanically stronger, more sensitive, more selective or more efficient. the regulatory control of nano medical devices should consider the product’s safety, efficacy and quality. the safety of nano products is of particular importance, especially when in contact with the surface of the human body. additionally, it is unclear whether or how the absorption and distribution of nanomaterials in the human body would introduce new risks. with regard to product efficacy, the manufacturer should determine the material’s size, shape, structure and distribution to ensure that the nanomaterial reaches the nanoscale, resulting in the predetermined and desired new properties, features or enhancements. in terms of quality, manufacturers should pay attention to validation, verification, uses and manufacturing processes in relation to the nanomaterial. to sum up, the important issues related to the regulatory control of nano medical devices include the definition of the nano medical device, classification, premarket approval procedures, good manufacturing practices, labeling and postmarket authors admet & dmpk 1(4) (2013) 76-81 78 surveillance. in view of the complexity of the technology involved, it will be a challenge to establish a regulatory requirement and procedure to control nano medical devices, whether it is a nano medical device used on its own or in combination with other equipment or drugs. ultimately, the aim of regulatory authorities is to ensure that the product offers improved treatment to the benefit of patients without incurring unacceptable risk. there are already a variety of nano medical devices that have been allowed onto the market by the us fda [3]. generally, these products serve the same clinical purpose as the related traditional medical device. however, the incorporation of a nanomaterial improves its effectiveness, performance or functionality, such as in antimicrobial dressings, dental materials, orthopedic materials like bone cement, and in vitro diagnostic reagents [3]. it is the fda’s current thinking that regulations can only be based on the current best information, and the product has to be fundamentally safe and effective, independent of whether the product employs a nanomaterial [4-5]. the product should, therefore, comply with the basic, relevant premarket review procedure and registration regulations and, if necessary, comply with relevant nanotechnology evaluations like toxicology assessments. the assessment of the safety and effectiveness of nano medical device is very much an ongoing development process for health authorities as well as academia in many countries. in fact, many nano medical devices that are currently on the market were approved through the conventional medical device approval procedure without taken full consideration of the presence of the nanomaterial. having said that, some health authorities are becoming more aware that the traditional assessment approach may not be fully appropriate for nano medical devices, particularly in terms of assessing the risk associated with the nanomaterial. recent animal studies have also demonstrated that nanomaterials exhibit many unique toxicological properties. therefore, in contrast with traditional products, nano medical devices require more careful consideration in the safety assessment and the development of a standardized evaluation protocol. in taiwan, the ministry of health has expressed a high degree of concern regarding the regulatory control of nano medical devices, and is looking into the relevant regulatory procedures for control nano medical devices. future regulations for nano medical devices may include a pre-market review, quality management systems and post-market surveillance. of those topics, there are issues of particular interest, including risk management, safety and efficacy assessments, nano process validation, labeling and warning, adverse event reporting and recall, and monitoring. the risk assessment of the effects of nanomaterials on public health should be a long-term, evolving process. it is important to maintain communication with experts in the area to ensure the speedy exchange of information, and to plan and adjust the risk assessment/management procedure accordingly. at the moment, although the international community does not regard the development of new legislation for nano medical devices as being necessary, it is important to pay close attention to the development of any international standards in order to formulate or adjust the relevant review mechanisms. 3. risk assessment for nanomaterials for the regulatory control of nano medical devices, in addition to regulations and product pre-market review mechanisms, risk management forms the basis of nano medical device assessment and the establishment of nano medical device regulations. for many nano medical devices, there is a fine balance between risk and effectiveness, especially in the medical field where nanomaterials are emerging as a promising approach for diagnostic technology and new treatment modalities [6]. in contrast to bulk admet & dmpk 1(4) (2013) 76-81 nanomaterials safety doi: 10.5599/admet.1.4.19 79 materials, nanomaterials can exhibit very different physical, chemical and biological properties that differ from expectation, which could result in new risks that need proper evaluation and even control before marketing is permitted. there is currently no specific standard for risk management regarding nano medical devices; however, the international standard iso 14971 (medical devices – application of risk management to medical devices) could be used for this purpose. like traditional medical devices, the existing standard operating procedures can be used, with special attention paid to potential hazards related to the properties and phenomena that can result from the nanomaterial or nanostructure. these potential hazards should be first identified, estimated and evaluated according to the predefined risk acceptance criteria. if necessary, risk control measures should be implemented, thereby ultimately ensuring that the product’s benefits outweigh the risk [7]. the identification of nano-hazards requires in-depth and in some cases multidisciplinary knowledge of the product’s technology, application environment, regulatory requirements and even usability. nano medical devices may on first appearance look exactly the same as their traditional counterparts, but for this exact reason, their usability becomes very important and the use hazards should be properly evaluated if necessary, as usage error may result from failure to tell the nano device from the traditional product. the use of standards like iso/iec 62366 [8] and ansi/aami he75 [9] are just some of the best currently available tools to address the usability issue. the identification of hazards can be performed systematically according to the different characteristics of the nanomaterial. for a nanomaterial with a novel structure, the hazards are usually more difficult to predict (as compared with a micronized material such as metal oxide, titanium dioxide or zinc oxide). in cases where these nanomaterials are enclosed within a surface coating or matrix material, the biological reactivity and mobility will be contained. however, the durability of the enclosure would then need to be considered in order to avoid the unexpected leaching of these nanomaterials into the environment. toxicology is obviously a major risk of nanotechnology that needs to be addressed. nanotoxicology risks are still very much debatable issues, and most result from various pathways of adsorption into the body. therefore, a nanomaterial unexpectedly or uncontrollably entering the human body through the skin or by inhalation is a concern. research studies have shown that as a material’s particle size becomes smaller, the total surface area of the potentially inhaled or adsorbed particles becomes greater, leading to increased toxicity [10]. airborne nanomaterials are often the main culprit in terms of toxicity. in order to assess the effects of exposure, adsorption, transport and distribution within the human body, various factors need to be taken into consideration, including the particle concentration in air, particle size, shape, surface area and surface chemistry. 4. nano-toxicology some types of nanomaterials may be harmful to humans. they can be classified into the following two kinds of nanomaterial structure: 1. free form nanoparticles: this refers to the direct use of nanoparticles in the manufacturing process. this usually consists of a single element, compound or complex mixture type that can be released into water and air or accumulate in soil. 2. fixed form nanoparticles or nanocomposites: this refers to a nanomaterial that becomes part of a product, which may be nanocomposites or nano surface structures. authors admet & dmpk 1(4) (2013) 76-81 80 compared to fixed nanoparticles, it is imperative to explore the possibility that free nanoparticles could be harmful. since nanomaterials are quite different from the original material, the related toxicity is expected to be unlike that of the original material. generally, it is difficult to obtain a uniform particle size distribution (monodisperse) in a nanoparticle formulation. in the absence of a stabilizing agent, nanoparticle aggregation is likely to occur, which could deteriorate the properties of the product and introduce complications in particle size characterization. some nanomaterials, such as carbon nanotubes and metal oxides, tend to form clusters [10-11]. a larger particle size could potentially reduce membrane permeability and discourage dispersion throughout the human body, such as in the respiratory tract [1112]. there are four ways that nanoparticles can enter the human body: through the skin, by inhalation, oral ingestion or during a medical procedure (e.g. release from implants). exposure assessments are used to evaluate the risk of human in contact with the nanoparticle product. this should include the whole life cycle of the product starting from synthesis to disposal. although the main concerns are the hazard to the customer during product usage, assessments should also be done after disposal to evaluate nanoparticle release into the environment. furthermore, the manufacturing staff remains the most likely population to be exposed to nanoparticles, and should be properly evaluated during risk management. the nanoparticle exposure assessment could vary between different products, and should depend on the product design to assess any potential risks. the dimensions of the nanomaterial affect its toxicity. for nanoparticles, smaller particles are usually more active and potentially more toxic. moreover, one should take into account the shape factor. the human immune system is poorly defensive against nanoparticles, and does not recognize nanoparticles. for instance, the defense mechanism in the lung cannot deal with nanoparticles less than 70 nm in size [6,13]. nanoparticles less than 50 nm in size can easily enter and move into the nucleus of human cells, and 30 nm particles may damage the central nervous system [6,13]. it is evident that nanomaterial research and risk assessments are urgently needed to clarify the potential effects of these materials on humans, particularly in terms of cardiovascular, carcinogenic, reproductive, developmental, immunological and neurological aspects. although there are still many uncertainties and limitations in the current understanding of nanomaterials, it is essential to conduct risk assessments on nanomaterials and nano medical devices before the product is allowed to be marketed. 5. conclusions in view of rapid advancements in the field of nanotechnology, it is very challenging to come up with a set of nano medical device regulations that can cover all the relevant issues. for currently listed nano medical products, such as those containing nano silver antimicrobial dressings, the fda review process has been completed. this will serve as a classic example for the regulatory regulations on a nano medical device. however, it is questionable whether such an example can be directly extended to other kinds of nano medical devices because the risk and benefit factors are likely to be very different. for instance, nanoparticle drug delivery systems that work by targeting tumors and nano silver antimicrobial dressing products are based on completely different scientific principles. nevertheless, it is mandatory for the manufacture of nano medical devices to meet quality, testing and safety and other regulations in order to protect the manufacturing staff, customers and end users. equally important, device manufacturers are also required to submit product development, testing and registration applications with other related information to the relevant regulatory body for proper evaluation. because of the diversity of admet & dmpk 1(4) (2013) 76-81 nanomaterials safety doi: 10.5599/admet.1.4.19 81 nanomaterials that can be applied to medical devices, regulatory regulations are important, and frequent updates are needed to keep pace with rapidly emerging applications. acknowledgments the office of medical device evaluation (omde) would like to acknowledge the support of the food and drug administration of taiwan and the center for measurement standards of itri towards omde’s work in nano medical device risk assessment, and all our team members who made this work feasible. references [1] fda: considering whether an fda-regulated product involves the application of nanotechnology, http://www.fda.gov/regulatoryinformation/guidances/ucm257698.htm (accessed 12 nd nov 2013) [2] j miller, the columbia science and technology law review 4 (2003) article 5. http://www.stlr.org/cite.cgi?volume=4&article=5 (accessed 12 nd nov 2013) [3] fda: nanotechnology, http://www.fda.gov/scienceresearch/specialtopics/nanotechnology/default.htm (accessed 12 nd nov 2013) [4] nanotechnology. a report of the u.s. food and drug administration nanotechnology task force, july 25, 2007. [5] nanotehnology & fda-regulated products: the essential guide, the food and drug law institute. 2009. [6] j.s. tsuji, a.d. maynard, p.c. howard, j.t. james, c.w. lam, d.b. warheit, a.b. santamaria, toxicology sciences 89 (2006) 42-45. [7] iso 14971:2007 medical devices – application of risk management to medical devices (2007) [8] ansi/aami/iec 62366:2007 medical devices – application of usability engineering to medical devices (2007) [9] ansi/aami he 75:2009 human factors engineering – design of medical devices (2009) [10] a.d. maynard, e.d. kuempel, journal of nanoparticle research 7 (2005) 587-614. [11] a.d. maynard, p.a. baron, m. foley, a.a. shvedova, e.r. kisin, v. castranova, journal of toxicology and environmental health a 67 (2004) 87-107. [12] d.b. warheit, k.l. reed, t.r. webb, experimental lung research 29 (2003) 593-606. [13] etc group, no small matter ii: the case for a global moratorium size matters! (2003) ©2013 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://www.fda.gov/regulatoryinformation/guidances/ucm257698.htm http://www.stlr.org/cite.cgi?volume=4&article=5 http://www.fda.gov/scienceresearch/specialtopics/nanotechnology/default.htm http://creativecommons.org/licenses/by/3.0/ anomalous salting-out, self-association and pka effects in the practically-insoluble bromothymol blue doi: https://doi.org/10.5599/admet.1823 1 admet & dmpk 0(0) (2023) 000-000; doi: https://doi.org/10.5599/admet.1823 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper development of a highly sensitive voltammetric sensor for the detection of folic acid by using mos2 and ionic liquid-modified carbon paste electrode hadi soltani nejad1, fariba garkani nejad2 and hadi beitollahi2,* 1department of chemistry, graduate university of advanced technology, kerman, iran 2environment department, institute of science and high technology and environmental sciences, graduate university of advanced technology, kerman 7631818356, iran *corresponding author: e-mail: h.beitollahi@yahoo.com received: april 18, 2023; revised: june 21, 2023; published: july 22, 2023 abstract background and purpose: sensitive analytical determination of folic acid is important in clinical laboratories due to its versatile biological functions. experimental approach: a simple folic acid sensor was successfully fabricated based on two-dimensional transition metal dichalcogenide mos2 modified carbon ionic liquid paste electrode (mos2-cilpe). the electrochemical properties of the fabricated electrode were investigated by cyclic voltammetry (cv), differential pulse voltammetry (dpv), and chronoamperometry. key results: the fabricated sensor displayed excellent electroactivity towards folic acid using cv. under optimal conditions (0.1 m pbs (ph 7.0)), the dpv oxidation peak current was proportional to folic acid concentration in the range from 5.0 μm to 100.0 μm with an estimated limit of detection of 1.0 µm and limit of quantification of 5.0 µm. conclusion: the ability of the sensor for routine analyses was demonstrated by the detection of folic acid present in folic acid tablets and urine samples with appreciable recovery values. ©2023 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords vitamin b9; transition metal dichalcogenides; electrocatalysis introduction recently, significant progress has been made in the development of electrochemical sensors and their application in point-of-care diagnostics, environmental studies, food safety, drug screening, and security [1-3]. low cost, simplicity, high reproducibility, real-time measurements, rapid response, low detection limit and portable devices are some advantages that cause extensive interest in electrochemical methods. particularly, voltammetric techniques are extremely sensitive and selective for the detection of easily oxidizable analytes [47]. to achieve sensitivity and selectivity, the modification of the working electrode in voltammetry is a usual practice. these chemically modified electrodes gain considerable attention in electrochemical quantification studies due to the enhanced electron transfer rate as well as selectivity achieved due to modifications [8-11]. carbon paste electrode (cpe) has been widely used in the determination of drugs, vitamins and other species because of its specific properties like easy preparation and wider potential window. the modifier has an important effect on the performance of modified cpes for electrochemical measurement [12-15]. https://doi.org/10.5599/admet.1823 https://doi.org/10.5599/admet.1823 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:h.beitollahi@yahoo.com http://creativecommons.org/licenses/by/4.0/ hadi soltani et al. admet & dmpk 00(0) (2023) 000-000 2 recent activity has focused on the development of nanoscaled particles applied in analytical chemistry to obtain special physicochemical characteristics of electrodes [16-21]. for example, nanomaterials with a large surface area, good conductivity, and excellent biological compatibility can be used as signal amplification elements in electrochemical sensors. therefore, exploring new advanced nanomaterials is key to developing sensors with a high sensitivity and low detection limit [22-24]. 2d layered nanomaterials are an emerging but important class of materials. they refer to materials with one dimension restricted to a single-atom layer, including monolayer and few-layer nanomaterials [25-27]. in the 2d family, layered transition metal dichalcogenides (tmds), such as mos2, ws2, tis2, tas2, mose2, and wse2, are fundamentally and technologically intriguing [28]. layer-structured transition-metal dichalcogenides possess an unique layered structure of similar structure to graphite and large surface areas, as well as outstanding physical, chemical, optical, and electronic properties, which holds great potential for applications in catalysis, sensing, optics, and energy [29-31]. because of its ultra thin layer structure, specific electrochemical properties, band gap (1.9 ev), large active edges, and easy surface modification, mos2 becomes one of the fascinating candidates to construct electrochemical sensors with high performance. as one of the layer-structured transition-metal dichalcogenides, mos2 has an analogous structure to graphite, which is composed of three atom layers: a mo layer sandwiched between two s layers, and the triple layers are stacked and held together by weak van der waals interactions. there are recent reports on using ionic liquids to design high sensitive electrochemical sensors. ionic liquids possess high ionic conductivity, high chemical and thermal stabilities, and high viscosity, and they are promising candidate materials for the fabrication of electrochemical sensors [32-34]. folic acid (fa), (2s)-2-[(4-{[(2-amino-4-hydroxypteridin-6-yl) methyl]amino}phenyl)formamido]pentanedioic acid, also known as folate (the natural form in body), vitamin b9, vitamin bc (or folacin), pteroyl-l-glutamic acid, pteroyl-l-glutamate and pteroylmonoglutamic acid are essential for numerous bodily functions. since humans cannot synthesize folate, the consumption of natural sources such as some green-leafy vegetables or fortified food and tablets is necessary [35-38]. several chronic diseases, for example, gigantocytic anemia, leucopoenia, mentality devolution, psychosis, heart attack, and stroke, are related to the deficiency of fa. it has also been suggested that decreased folate concentration is associated with enhanced carcinogenesis as folic acid with vitamin b12 participates in the nucleotide synthesis, cell division and gene expression. besides, it is an essential nutrient for pregnant women to prevent neural tube defects in the fetus [39-42]. so, a sensitive determination of fa from a clinical viewpoint is very important. in this study, mos2 modified carbon ionic liquid paste electrode (mos2-cilpe) sensor was fabricated as a highly sensitive voltammetric sensor to determine the fa. the mos2-cilpe sensor showed an acceptable ability to determine the folic acid in folic acid tablets and urine samples. experimental chemicals and instrumentation all chemicals used were of analytical grade and were used as received without any further purification and were obtained from sigma-aldrich. orthophosphoric acid was utilized to prepare the phosphate buffer solutions (pbss), and sodium hydroxide was used to adjust the desired ph values (ph range between 2.0 and 9.0). all solutions were prepared with deionised water of millipore direct-q® 8 uv (ultra-violet) (millipore, germany). the ph was also measured and a buffer solution was prepared using a digital ph meter (metrohm, admet & dmpk 00(0) (2023) 000-000 detection of folic acid by using mos2 and modified carbon paste electrode doi: https://doi.org/10.5599/admet.1823 3 ph lab 713). voltammetric measurements were carried out using an autolab pgstat302n, potentiostat/galvanostat (made in netherlands). the system was run on a pc using general purpose electrochemical system (gpes) 4.9 software. a three-electrode system was used, including a platinum wire as the auxiliary electrode, an ag/agcl/kcl (saturated) as the reference electrode, and the mos2-cilpe as the working electrode. the synthesis and characterization of 2d mos2 nanosheets has been reported in our previous work [43]. preparation of mos2-cilpe mos2-cilpe was prepared by mixing 0.04 g of mos2 nanosheets with 0.96 g graphite powder and the appropriate amount of ionic liquid (1-butyl-3-methylimidazolium hexafluorophosphate) and paraffin oil (30/70 (w/w)) with a mortar and pestle. the paste was then packed into the end of a glass tube (3.4 mm inner diameter and 15 cm long). a copper wire inserted into the carbon paste provided the electrical contact. for comparison, carbon ionic liquids paste electrode (cilpe) in the absence of mos2 nanosheets, mos2-cpe consistent of mos2 nanosheets powder, graphite powder and paraffin oil, and bare cpe consisting of graphite powder and paraffin oil were also prepared in the same way. preparation of real samples five tablets of the fa purchased from a local pharmacy in kerman, iran (1 mg fa per tablet) were completely powdered in the mortar with a pestle. then, an accurately weighed amount of the homogenized fa powder was transferred into 100 ml 0.1 m pbs (ph 7.0). for better dissolution, the solutions inside the flasks were sonicated (20 min). after that, the resulting samples were filtered. finally, a specific volume of the prepared samples was transferred to volumetric flasks and diluted with 0.1 m pbs (ph 7.0). the diluted solutions were then put in the electrochemical cell for dpv analysis. the collected urine samples were stored in the refrigerator after collection. the urine sample was centrifuged for 5 minutes at 2000 rpm. then, the supernatant solution was filtered after phase separation and diluted with 0.1 m pbs (ph 7.0). the diluted solution was then put in the electrochemical cell for dpv analysis. the analytical experiments were performed using the standard addition method. results and discussion electrochemical behavior of fa on the mos2-cilpe mechanism of the fa oxidation on mos2-cilpe is suggested on the basis of the relationship between the oxidation potential and ph of supporting electrolyte. the effect of the electrolyte ph on the oxidation of 100.0 μm fa was investigated at mos2-cilpe using dpv measurements in the pbs in the ph range from 2.0 to 9.0. according to the results, the oxidation peak current of fa depends on the ph value. it increases with increasing ph until it reaches the maximum at ph 7.0, then decreases with higher ph values. the optimized ph corresponding to the higher peak current was 7.0 (fig. 1), indicating that protons are involved in the reaction of fa oxidation (2 electrons and 2 protons). the effect of mos2 nanosheets and ils in the modification process was investigated by recording cyclic voltammograms of 100.0 µm fa at the surface of cpe (curve a), mos2-cpe (curve b), ionic liquid modified carbon paste electrode (il-cpe) (curve c) and mos2-cilpe (curve d). the results are shown in figure 2. the oxidation current and potential for fa were detected at about 3.1 µa and 750 mv at the surface of cpe and 4.8 µa and 732 mv at the surface of mos2-cpe, respectively. on il-cpe, the oxidation peak was located at 725 mv with an oxidation peak height of 8.0 µa (curve c). it can be seen that the oxidation peak potential moved to the negative direction with a significant increase of the oxidation peak current attributed to the https://doi.org/10.5599/admet.1823 hadi soltani et al. admet & dmpk 00(0) (2023) 000-000 4 presence of ionic liquid as the modifier in the carbon paste electrode. the modification of cpe with mos2 nanosheets and ils improved the oxidation current of fa (11.7 µa) and decreased the oxidation potential of fa (700 mv) compared with the bare cpe. figure 1. plot of the oxidation peak current of 100.0 μm fa as a function of ph solution at mos2-cilpe in 0.1 m pbs at different ph value (2.0 9.0). figure 2. cyclic voltammetric response of 100.0 μm folic acid at (a) bare cpe, (b) mos2-cpe, (c) il-cpe and d) mos2-cilpe in 0.1 m pbs of ph 7.0 (scan rate = 50 mv s1) in the potential window of 350-820 mv. effect of scan rate the effect of the potential scan rates (10-100 mv s-1) on the electrochemical oxidation of fa was studied by cv. figure 3 shows the cv of 90.0 µm of fa in the 0.1 m pbs at the mos2-cilpe. these results show that the anodic current increased with an increasing scan rate. the oxidation current of fa increased linearly with the square root of the scan rate (figure 3, inset), demonstrating a diffusion-controlled electrochemical process. admet & dmpk 00(0) (2023) 000-000 detection of folic acid by using mos2 and modified carbon paste electrode doi: https://doi.org/10.5599/admet.1823 5 figure 3. cyclic voltammetric responses of 90.0 μm fa in 0.1 m pbs (ph 7.0) at scan rates of 10 to 100 mv s-1 at mos2-cilpe (a-h refers to 10, 20, 30, 40, 50, 60, 80, and 100 mv s-1) in the potential window of 280-800 mv. inset: plot of the square root of the scan rate vs. the oxidation peak current of fa. chronoamperometric analysis the chronoamperometric measurements of fa at the mos2-cilpe surface were done to estimate the apparent diffusion coefficient. figure 4 shows the current-time profiles obtained by setting the working electrode potential at 750 mv for different concentrations of fa. figure 4. the chronoamperograms obtained at mos2-cilpe in 0.1 m pbs at ph 7.0 for different concentrations of fa at step potential = 750 mv. noted that a–d related to 0.1, 0.5, 1.0, and 1.5 mm of fa. inset a: the i plot versus t-1/2 observed by chronoamperograms a to d. inset b: slope plot of the straight line vs. concentration of fa. https://doi.org/10.5599/admet.1823 hadi soltani et al. admet & dmpk 00(0) (2023) 000-000 6 at long enough experimental times (t = 0.3 to 3 s), where the electron transfer reaction rate of fa is more than its diffusion rate toward the working electrode surface, the current is diffusion controlled. figure 4, inset a, shows the experimental plots of i versus t-1/2 with the best fit for different concentrations of fa employed. the slopes of the resulting straight lines were then plotted versus the fa concentration (figure 4, inset b). based on the cottrell equation (the cottrell equation is i = nfac (d/πt)1/2, where d is the diffusion coefficient (cm2 s-1), c is the concentration in bulk solution (mm), a is the surface area of the electrode (cm2), f is faraday’s constant, t is the time (s), and n is the number of electrons transferred), the slope of this plot (figure 4 inset b) can be used to estimate of the diffusion coefficient of fa. from the slope of this plot, the value of d was found to be 5.7×10-6 cm2s-1 for fa. calibration plot and limit of detection since dpv has a much higher current sensitivity and better resolution than cv, dpv was used for the determination of fa. figure 5 shows the dpv curves of mos2-cilpe in the pbs with variable fa levels (step potential = 0.01 v and pulse amplitude =0 .025 v). it was found that the electrocatalytic peak currents of fa oxidation at the mos2-cilpe surface linearly depended on fa concentrations above the range of 5.0 to 100.0 µm. the limit of detection is estimated by using the following equation, lod = 3sb/m. in this equation, m is the slope of the calibration plot (0.0829 μa μm-1) and sb is the standard deviation of the blank response, obtained from 8 replicate measurements of the blank solution. the limit of detection was 1.0 μm and the limit of quantification (loq) was obtained 5.0 µm. a comparison of fa detection using various sensors is presented in table 1. figure 5. dpv response of fa at mos2-cilpe in the concentration range 5.0 to 100.0 μm in 0.1 m pbs of ph 7.0 (a-k refers to 5.0, 10.0, 20.0, 30.0, 40.0, 50.0, 60.0, 70.0, 80.0, 90.0, and 100.0 µm). inset: the peak current plot as a function of the fa concentration ranging from 5.0 to 100.0 μm. admet & dmpk 00(0) (2023) 000-000 detection of folic acid by using mos2 and modified carbon paste electrode doi: https://doi.org/10.5599/admet.1823 7 table 1. comparison of different dpv sensors for fa detection, . electrochemical sensor linear range, µm lod, µm ref. mos2-reduced graphene oxide hybrid/glassy carbon electrode 0.01-100 0.010 [44] ferrocene dicarboxylic acid/carbon nanotube paste 4-152 1.1 [45] methylene blue-reduced graphene oxide/glassy carbon electrode 4-167 0.500 [46] zro2 nanoparticles-carbon paste electrode 20–2500 9.86 [47] multi-walled carbon nanotubes-pt nanoparticles/glassy carbon electrode 0.2-100 0.050 [48] mos2-cilpe 5.0-100.0 1.0 this work interference studies to evaluate the selectivity of mos2-cilpe for fa, an investigation of the influence of potential interfering substances was performed under the optimized conditions. the dpv responses after adding interfering substances into 0.1 m pbs (ph 7.0) containing 50.0 µm fa were recorded. the tolerance limit was defined as the ratio of the concentration of the interfering species to the analyte, which led to a relative error of less than ±5.0 %. it was found that the 500-fold excess of glucose, glycine, methionine, histidine, alanine, glutamic acid, glycine, phenylalanine, 400-fold excess of ca2+, na+, mg2+, nh4+, cl-, so42-, 70-fold excess of urea, uric acid, and 10-fold of ascorbic acid did not remarkable interfere for fa determination. stability, reproducibility, and repeatability of the mos2-cilpe sensor for evaluation of the reproducibility of the prepared sensor, five mos2-cilpe were prepared independently and used in the determination of fa through dpv in pbs (0.1 m, ph 7.0). under the same experimental conditions, the calculated relative standard deviation (rsd) of peak currents was only about 4.1 %, indicating reliable reproducibility of the sensing platform. the storage stability of the mos2-cilpe was further examined by measurement of the fa oxidation peak current over the time interval of 12 days. no obvious decrease in the initial current value of fa was observed after 12 days, implying acceptable storage stability. the repeatability of the mos2-cilpe sensor was studied by 5 consecutive measurements of 50.0 μm fa with rsd of 3.5 %, indicating good repeatability of the sensor. analysis of real samples the real samples for the analysis were prepared and quantified by the dpv method. the developed sensor was applied to detect fa in folic acid tablets and urine samples. the results are summarized in table 2. each measurement was repeated five times. the fa acid content of each tablet was obtained at 1.003 mg. the recovery and relative standard deviation (rsd) values confirmed that the mos2-cilpe sensor has great potential for analytical application. table 2. determining fa in folic acid tablets and urine through mos2-cilpe. all the concentrations are in μm (n = 5). sample fa concentration, μm recovery, % spiked found folic acid tablet 0 3.5±0.01 1.0 4.4±0.015 97.8 2.0 5.6±0.02 101.8 3.0 6.7±0.012 103.1 4.0 7.4±0.01 98.7 urine 0 5.0 4.9±0.011 98.0 6.0 6.1±0.016 101.7 7.0 6.8±0.013 97.1 8.0 8.3±0.019 103.7 https://doi.org/10.5599/admet.1823 hadi soltani et al. admet & dmpk 00(0) (2023) 000-000 8 conclusion a sensitive and reliable electrochemical method based on mos2-cilpe was proposed for the determination of fa. due to the large surface area of mos2, high conductivity and catalytic activity of ionic liquid, the modified electrode exhibited good catalytic activity to fa with enhanced oxidation peak current and decreased oxidation overpotential. the voltammetric current response increased linearly with increasing fa concentration in the range of 5.0 to 100.0 μm and the detection limit of 1.0 μm was obtained. moreover, the mos2-cilpe sensor may provide a facile and effective analysis approach for the determination of fa in real samples. references [1] v. cerda, g. o. araijo rennan, s. l. ferreira, revising flow-through cells for amperometric and voltammetric detections using stationary mercury and bismuth screen printed electrodes. progress in chemical and biochemical research 5(4) (2022) 351-366. https://doi.org/10.22034/pcbr.2022.36 2520.1232. [2] s. tajik, m. a. taher, h. beitollahi, simultaneous determination of droxidopa and carbidopa using a carbon nanotubes paste electrode. sensors and actuators b 188 (2013) 923-930. https://doi.org/ 10.1016/j.snb.2013.07.085. [3] m. hassani, m. zeeb, a. monzavi, z. khodadadi, m. kalaee, response surface modeling and optimization of microbial fuel cells with surface-modified graphite anode electrode by zsm-5 nanocatalyst functionalized. chemical methodologies 6(3) (2022) 253-268. https://doi.org/10.22034/ chemm.2022.324312.1425. [4] j. mohanraj, d. durgalakshmi, r. a. rakkesh, s. balakumar, s. rajendran, h. karimi-maleh, facile synthesis of paper based graphene electrodes for point of care devices: a double stranded dna (dsdna) biosensor. journal of colloid and interface science 566 (2020) 463-472. https://doi.org/ 10.1016/j.jcis.2020.01.089. [5] h. roshanfekr, a simple specific dopamine aptasensor based on partially reduced graphene oxide– aunps composite. progress in chemical and biochemical research 6(1) (2023) 79-88. https://doi.org/ 10.22034/pcbr.2023.381280.1245. [6] s. mohammadi, h. beitollahi, a. mohadesi, electrochemical behaviour of a modified carbon nanotube paste electrode and its application for simultaneous determination of epinephrine, uric acid and folic acid. sensor letters 11(2) (2013) 388-394. https://doi.org/10.1166/sl.2013.2723. [7] h. pyman, design and fabrication of modified dna-gp nano-biocomposite electrode for industrial dye measurement and optical confirmation. progress in chemical and biochemical research 5(4) (2022) 391-405. doi: 10.22034/pcbr.2022.367576.1236. [8] s. cheraghi, m. a. taher, h. karimi-maleh, f. karimi, m. shabani-nooshabadi, m. alizadeh, a. alothman, n. erk, p. k. y. raman, c. karaman, novel enzymatic graphene oxide based biosensor for the detection of glutathione in biological body fluids. chemosphere 287 (2022) 132187. https://doi.org/ 10.1016/j.chemosphere.2021.132187 [9] m. vardini, n. abbasi, a. kaviani, m. ahmadi, e. karimi, graphite electrode potentiometric sensor modified by surface imprinted silica gel to measure valproic acid. chemical methodologies 6(5) (2020) 398-408. https://doi.org/10.22034/chemm.2022.328620.1437 [10] z. zhang, h. karimi-maleh, label-free electrochemical aptasensor based on gold nanoparticles/titanium carbide mxene for lead detection with its reduction peak as index signal. advanced composites and hybrid materials 6 (2023) 68. https://doi.org/10.1007/s42114-023-00652-1 [11] k. harismah, m. mirzaei, m. dai, z. roshandel, e. salarrezaei, in silico investigation of nanocarbon biosensors for diagnosis of covid-19 eurasian chemical communications 3(2) (2021) 95-102. http://dx.doi.org/10.22034/ecc.2021.267226.1120 https://doi.org/10.22034/pcbr.2022.362520.1232 https://doi.org/10.22034/pcbr.2022.362520.1232 https://doi.org/10.1016/j.snb.2013.07.085 https://doi.org/10.1016/j.snb.2013.07.085 https://doi.org/10.22034/chemm.2022.324312.1425 https://doi.org/10.22034/chemm.2022.324312.1425 https://doi.org/10.1016/j.jcis.2020.01.089 https://doi.org/10.1016/j.jcis.2020.01.089 https://doi.org/10.22034/pcbr.2023.381280.1245 https://doi.org/10.22034/pcbr.2023.381280.1245 https://doi.org/10.1166/sl.2013.2723 https://doi.org/10.1016/j.chemosphere.2021.132187 https://doi.org/10.1016/j.chemosphere.2021.132187 https://doi.org/10.22034/chemm.2022.328620.1437 http://dx.doi.org/10.22034/ecc.2021.267226.1120 admet & dmpk 00(0) (2023) 000-000 detection of folic acid by using mos2 and modified carbon paste electrode doi: https://doi.org/10.5599/admet.1823 9 [12] z. taleat, m. m. ardakani, h. naeimi, h. beitollahi, m. nejati, h. r. zare, electrochemical behavior of ascorbic acid at a 2,2'-[3,6-dioxa-1,8-octanediylbis (nitriloethylidyne)]-bis-hydroquinone carbon paste electrode. analytical sciences 24(8) (2008) 1039-1044. https://doi.org/10.2116/analsci.24.1039 [13] z. zhang, h. karimi-maleh, in situ synthesis of label-free electrochemical aptasensor-based sandwichlike aunps/ppy/ti3c2tx for ultrasensitive detection of lead ions as hazardous pollutants in environmental fluids. chemosphere 324 (2023) 138302. https://doi.org/10.1016/j.chemosphere.20 23.138302 [14] a. hosseini fakhrabad, r. sanavi khoshnood, m.r. abedi, m. ebrahimi, fabrication a composite carbon paste electrodes (cpes) modified with multi-wall carbon nano-tubes (mwcnts/n, n-bis (salicyliden)1,3-propandiamine) for determination of lanthanum (iii). eurasian chemical communications 3(9) (2021) 627-634. http://dx.doi.org/10.22034/ecc.2021.288271.1182 [15] s. saghiri, m. ebrahimi, m. bozorgmehr, electrochemical amplified sensor with mgo nanoparticle and ionic liquid: a powerful strategy for methyldopa analysis. chemical methodologies 5(3) (2021) 234239. https://doi.org/10.22034/chemm.2021.128530 [16] f. garkani nejad, h. beitollahi, r. alizadeh, sensitive determination of hydroxylamine on zno nanorods/graphene oxide nanosheets modified graphite screen printed electrode. analytical & bioanalytical electrochemistry 9(2) (2017) 134-144. [17] h. beitollahi, i. sheikhshoaie, novel nanostructure-based electrochemical sensor for simultaneous determination of dopamine and acetaminophen. materials science and engineering c 32(2) (2012) 375-380. https://doi.org/10.1016/j.msec.2011.11.009 [18] r. mohammadi. sonocatalytic degradation of methyl red by sonochemically synthesized tio2sio2/chitosan nanocomposite. journal of applied organometallic chemistry 2(4) (2022) 212-223. http://dx.doi.org/10.22034/jaoc.2022.154993 [19] c. karaman, o. karaman, p. l. show, y. orooji, h. karimi-maleh, utilization of a double-cross-linked amino-functionalized three-dimensional graphene networks as a monolithic adsorbent for methyl orange removal: equilibrium, kinetics, thermodynamics and artificial neural network modeling. environmental research 207 (2021) 112156. https://doi.org/10.1016/j.envres.2021.112156 [20] y. dessie, s. tadesse, a review on advancements of nanocomposites as efficient anode modifier catalyst for microbial fuel cell performance improvement. journal of chemical reviews 3(4) (2021) 320-344. http://dx.doi.org/10.22034/jcr.2021.314327.1128 [21] v. k. bale, h. r. katreddi, synthesis, characterization and catalytic activity of zinc oxide nanoparticles functionalized with metallo-thoiosemicarbazones. asian journal of nanosciences and materials 5(2) (2022) 159-173. doi: 10.26655/ajnanomat.2022.2.7 [22] h. karimi-maleh, r. darabi, f. karimi, c. karaman, s. a. shahidi, n. zare, m. baghayeri, l. fu, s. rostamnia, j. rouhi, state-of-art advances on removal, degradation and electrochemical monitoring of 4-aminophenol pollutants in real samples. environmental research 222 (2023) 115338. https://doi.org/10.1016/j.envres.2023.115338 [23] m. bijad, a. hojjati-najafabadi, h. asari-bami, s. habibzadeh, i. amini, f. fazeli, an overview of modified sensors with focus on electrochemical sensing of sulfite in food samples. eurasian chemical communications 3(2) (2021) 116-138. http://dx.doi.org/10.22034/ecc.2021.268819.1122 [24] m. mazloum-ardakani, h. beitollahi, z. taleat, h. naeimi, n. taghavinia, selective voltammetric determination of d-penicillamine in the presence of tryptophan at a modified carbon paste electrode incorporating tio2 nanoparticles and quinizarine. journal of electroanalytical chemistry 644(1) (2010) 1-6. https://doi.org/10.1016/j.jelechem.2010.02.034 [25] h. karimi-maleh, c. t. fakude, n. mabuba, g. m. peleyeju, o. a. arotiba, the determination of 2phenylphenol in the presence of 4-chlorophenol using nano-fe3o4/ionic liquid paste electrode as an electrochemical sensor. journal of colloid and interface science 554 (2019) 603-610. https://doi.org/ 10.1016/j.jcis.2019.07.047 [26] y. h. wang, k. j. huang, x. wu, recent advances in transition-metal dichalcogenides based electrochemical biosensors: a review. biosensors and bioelectronics 97 (2017) 305-316. https://doi.org/ 10.1016/j.bios.2017.06.011 https://doi.org/10.5599/admet.1823 https://doi.org/10.1016/j.chemosphere.2023.138302 https://doi.org/10.1016/j.chemosphere.2023.138302 http://dx.doi.org/10.22034/ecc.2021.288271.1182 https://doi.org/10.22034/chemm.2021.128530 http://dx.doi.org/10.22034/jaoc.2022.154993 https://doi.org/10.1016/j.envres.2021.112156 http://www.jchemrev.com/article_140627.html http://www.jchemrev.com/article_140627.html http://dx.doi.org/10.22034/jcr.2021.314327.1128 https://doi.org/10.1016/j.envres.2023.115338 http://dx.doi.org/10.22034/ecc.2021.268819.1122 https://doi.org/10.1016/j.jcis.2019.07.047 https://doi.org/10.1016/j.jcis.2019.07.047 hadi soltani et al. admet & dmpk 00(0) (2023) 000-000 10 [27] h. zhang, ultrathin two-dimensional nanomaterials. acs nano 9(10) (2015) 9451-9469. https://doi.org/10.1021/acsnano.5b05040 [28] h. karimi-maleh, y. liu, z. li, r. darabi, y. orooji, c. karaman, f. karimi, m. baghayeri, j. rouhi, l. fu, calf thymus ds-dna intercalation with pendimethalin herbicide at the surface of zif8/co/rgo/c3n4/ds-dna/spce; a bio-sensing approach for pendimethalin quantification confirmed by molecular docking study. chemosphere 332 (2023) 138815. https://doi.org/10.1016/j.chemosphe re.2023.138815 [29] c. tan, h. zhang, two-dimensional transition metal dichalcogenide nanosheet-based composites. chemical society reviews 44(9) (2015) 2713-2731. https://doi.org/10.1039/c4cs00182f [30] f. li, m. xue, two-dimensional transition metal dichalcogenides for electrocatalytic energy conversion applications, in two-dimensional materials-synthesis, characterization and potential applications, p. k. nayak ed., intech, 2016, 64-84. https://doi.org/10.5772/64760 [31] s. tajik, z. dourandish, f. g. nejad, h. beitollahi, p. m. jahani, a. di bartolomeo, transition metal dichalcogenides: synthesis and use in the development of electrochemical sensors and biosensors. biosensors and bioelectronics 216 (2022) 114674. https://doi.org/10.1016/j.bios.20 22.114674 [32] m. pumera, a. h. loo, layered transition-metal dichalcogenides (mos2 and ws2) for sensing and biosensing. trac trends in analytical chemistry 61 (2014) 49-53. https://doi.org/10.1016/j.trac.20 14.05.009 [33] j. ping, s. ru, k. fan, j. wu, y. ying, copper oxide nanoparticles and ionic liquid modified carbon electrode for the non-enzymatic electrochemical sensing of hydrogen peroxide. microchimica acta 171 (2010) 117-123. https://doi.org/10.1007/s00604-010-0420-3 [34] m. bijad, h. karimi-maleh, m. farsi, s. a. shahidi, an electrochemical-amplified-platform based on the nanostructure voltammetric sensor for the determination of carmoisine in the presence of tartrazine in dried fruit and soft drink samples. journal of food measurement and characterization 12 (2018) 634640. https://doi.org/10.1007/s11694-017-9676-1 [35] a. moghaddam, h. zamani, h. karimi-maleh, a new sensing strategy for determination of tamoxifen using fe3o4/graphene-ionic liquid nanocomposite amplified paste electrode. chemical methodologie 5(5) (2021) 373-380. https://doi.org/10.22034/chemm.2021.135727 [36] b. b. prasad, r. madhuri, m. p. tiwari, p. s. sharma, imprinted polymer–carbon consolidated composite fiber sensor for substrate-selective electrochemical sensing of folic acid. biosensors and bioelectronics 25(9) (2010) 2140-2148. https://doi.org/10.1016/j.bios.2010.02.016 [37] b. batra, v. narwal, v. kalra, m. sharma, j. s. rana, folic acid biosensors: a review. process biochemistry 92 (2020) 343-354. https://doi.org/10.1016/j.procbio.2020.01.025 [38] j. p. winiarski, r. rampanelli, j. c. bassani, d. z. mezalira, c. l. jost, multi-walled carbon nanotubes/nickel hydroxide composite applied as electrochemical sensor for folic acid (vitamin b9) in food samples. journal of food composition and analysis 92 (2020) 103511. https://doi.org/10.1016/ j.jfca.2020.103511 [39] h. beitollahi, j. b. raoof, r. hosseinzadeh, electroanalysis and simultaneous determination of 6thioguanine in the presence of uric acid and folic acid using a modified carbon nanotube paste electrode. analytical sciences 27(10) (2011) 991-997. https://doi.org/10.2116/analsci.27.991 [40] p. kalimuthu, s. a. john, selective electrochemical sensor for folic acid at physiological ph using ultrathin electropolymerized film of functionalized thiadiazole modified glassy carbon electrode. biosensors and bioelectronics 24(12) (2009) 3575-3580. https://doi.org/10.1016/j.bios.2009.05.017 [41] b. b. prasad, r. madhuri, m. p. tiwari, p. s. sharma, electrochemical sensor for folic acid based on a hyperbranched molecularly imprinted polymer-immobilized sol–gel-modified pencil graphite electrode. sensors and actuators b: chemical 146(1) (2010) 321-330. https://doi.org/10.1016/ j.snb.2010.02.025 https://doi.org/10.1016/j.chemosphere.2023.138815 https://doi.org/10.1016/j.chemosphere.2023.138815 https://doi.org/10.5772/64760 https://doi.org/10.1007/s11694-017-9676-1 https://doi.org/10.22034/chemm.2021.135727 https://doi.org/10.2116/analsci.27.991 admet & dmpk 00(0) (2023) 000-000 detection of folic acid by using mos2 and modified carbon paste electrode doi: https://doi.org/10.5599/admet.1823 11 [42] m. arvand, m. dehsaraei, a simple and efficient electrochemical sensor for folic acid determination in human blood plasma based on gold nanoparticles–modified carbon paste electrode. materials science and engineering c 33(6) (2013) 3474-3480. https://doi.org/10.1016/j.msec.2013.04.037 [43] s. tajik, z. dourandish, f. garkani-nejad, a. aghaei afshar, h. beitollahi, voltammetric determination of isoniazid in the presence of acetaminophen utilizing mos2-nanosheet-modified screen-printed electrode. micromachines 13(3) (2022) 369. https://doi.org/10.3390/mi13030369 [44] f. chekin, f. teodorescu, y. coffinier, g. h. pan, a. barras, r. boukherroub, s. szunerits, mos2/reduced graphene oxide as active hybrid material for the electrochemical detection of folic acid in human serum. biosensors and bioelectronics 85 (2016) 807-813. https://doi.org/10.1016/j.bios.2016.05.095 [45] a. a. ensafi, h. karimi-maleh, modified multiwall carbon nanotubes paste electrode as a sensor for simultaneous determination of 6-thioguanine and folic acid using ferrocenedicarboxylic acid as a mediator. journal of electroanalytical chemistry 640(1-2) (2010) 75-83. https://doi.org/10.1016/ j.jelechem.2010.01.010 [46] d. zhang, x. ouyang, w. ma, l. li, y. zhang, voltammetric determination of folic acid using adsorption of methylene blue onto electrodeposited of reduced graphene oxide film modified glassy carbon electrode. electroanalysis 28(2) (2016) 312-319. https://doi.org/10.1002/elan.201500348 [47] m. mazloum-ardakani, h. beitollahi, m. k. amini, f. mirkhalaf, m. abdollahi-alibeik, new strategy for simultaneous and selective voltammetric determination of norepinephrine, acetaminophen and folic acid using zro2 nanoparticles-modified carbon paste electrode. sensors and actuators b: chemical 151(1) (2010) 243-249. https://doi.org/10.1016/j.snb.2010.09.011 [48] z. kun, z. ling, h. yi, c. ying, t. dongmei, z. shuliang, z. yuyang, electrochemical behavior of folic acid in neutral solution on the modified glassy carbon electrode: platinum nanoparticles doped multiwalled carbon nanotubes with nafion as adhesive. journal of electroanalytical chemistry 677 (2012) 105-112. https://doi.org/10.1016/j.jelechem.2012.05.010 ©2023 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) https://doi.org/10.5599/admet.1823 https://doi.org/10.3390/mi13030369 https://doi.org/10.1016/j.bios.2016.05.095 https://doi.org/10.1016/j.jelechem.2010.01.010 https://doi.org/10.1016/j.jelechem.2010.01.010 https://doi.org/10.1002/elan.201500348 https://doi.org/10.1016/j.snb.2010.09.011 https://doi.org/10.1016/j.jelechem.2012.05.010 http://creativecommons.org/licenses/by/3.0/ synthetic routes to theranostic applications of carbon-based quantum dots doi: https://doi.org/10.5599/admet.1926 1 admet & dmpk 0(0) (2023) 000-000; doi: https://doi.org/10.5599/admet.1926 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index review liposomes: from august wassermann to vaccines against covid-19 ganna grygorieva1, daria pylypenko2* and yuriy krasnopolsky3 1si "institute of pharmacology and toxicology of nams of ukraine", 14, ezhene potier str., kyiv, 03680, ukraine 2state biotechnological university, 44, alchevskikh str., kharkiv, 61002, ukraine 3national technical university “kharkiv polytechnic institute”, 2, kyrpychova str., kharkiv, 61002, ukraine *corresponding author: e-mail: pdmforwork@gmail.com received: june 5, 2023; revised: june 27, 2023; published: june 30, 2023 abstract background and purpose: the development of vaccines against the sars-cov-2 virus has become a big challenge for many countries in 2020-2022. mrna vaccines were shown to be effective and safe and have been widely used worldwide in the fight against the covid-19 pandemic. the fundamental factor in creating mrna vaccines, which ensures effective delivery of mrna to the host cells, is the composition of lipid nanoparticles, namely the presence of ionized charged lipids, which ensures the binding of mrna molecules. however, the significant role of liposomes in the development of liposomal vaccines and identification of immunochemical reactions involving lipids should be assessed in the context of the development of the pioneering idea of august wassermann about the use of liposomal antigens in the diagnosis and immunoprophylaxis of serious human diseases. experimental approach: the review is devoted to the use of liposomal antigens as antigen-delivery systems for diagnosis and immunoprophylaxis. key results: studies of cardiolipin antigen in serodiagnosis of syphilis became the foundation of antibodies in diagnosing various infectious diseases and pathological conditions, such as tuberculosis, lupus erythematosus, covid19, borreliosis, etc. identification of antiphospholipid antibodies (mainly anticardiolipin) and today is the most important diagnostic tool for antiphospholipid syndrome. conclusion: the liposomal system first proposed in 1906 for the diagnosis of syphilis evolved more than a century later into mrna vaccines, which are used today in the fight against the covid-19 pandemic. ©2023 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords liposomal antigen delivery system; cardiolipin antigen; antilipid antibodies; wassermann reaction; serodiagnosis of syphilis; lipid nanoparticle; mrna vaccine introduction in modern concepts of drug delivery systems, a priority place is occupied by nanosized lipid structures, which for the last 40 years, have become the applied basis for liposomal drug and antigen delivery systems for pharmacotherapy, diagnostics, and immunoprophylaxis [1,2]. however, it should be recalled that the successful empirical use of lipid nanoparticles in laboratory diagnostics of infectious diseases dates back to the beginning of the 20th century. we are talking about the first serological test proposed in 1906 by a. https://doi.org/10.5599/admet.1926 https://doi.org/10.5599/admet.1926 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:pdmforwork@gmail.com http://creativecommons.org/licenses/by/4.0/ g. grygorieva et al. admet & dmpk 00(0) (2023) 000-000 2 wassermann, a. neisser and с. bruck for the diagnosis of treponema pallidum (syphilis) in humans using the complement fixation reaction with lipid antigens (ag), called “wassermann reaction” (figure 1) [3]. figure 1. principle of wassermann reaction: (a) positive result and (b) negative result. despite the development of methods for obtaining various lipid ags and modification of reactions involving these ags in subsequent years, only cardiolipin (diphosphatidylglycerol – dpg) was identified as an immunochemically active lipid ag for the serodiagnosis of syphilis fabrics due to its presence in both healthy and syphilitic tissues [4,5]. in addition to dpg itself, a composition of dpg and phosphatidylcholine (pc) was proposed as a lipid ag for syphilis serodiagnosis, the components of which were isolated from the heart muscle of cattle. note that in the second half of the 20th century, the unsatisfactory specificity of these ags in serodiagnosis initiated the study of a whole spectrum of cardiolipin ags (so-called vdrl-test), differing in qualitative and quantitative composition [6-8]. liposomal structure of the cardiolipin antigen in the serodiagnosis of syphilis the most important step in the study of cardiolipin ag was the confirmation of its liposomal structure in the aqueous phase, which is no longer in doubt today. a number of studies have shown that the high immunochemical activity of cardiolipin ag is a characteristic of the dpg liposomal composition. pc and cholesterol (chol) with a particle size of 1 to 10 µm (figure 2) [9-11]. using vdrl-test based on dpg, pc, and chol, the possibility of agglutination of antibodies with an emulsion of large bilayer liposomes containing not treponemal but lipid ag was shown [12]. when the emulsion is mixed with serum, the antibodies in the serum of a patient with syphilis react specifically and with high affinity with dpg, causing agglutination. a lipid composition including a negatively charged lipid (e.g. dpg, phosphatidylglycerol (pg), phosphatidylserine (ps) or phosphatidic acid (pa)), neutral lipid (e.g., pc) and chol was offered [13-15]. using 31p-nmr, it was established that cardiolipin ag is represented in the form of multilayer liposomes no less than 1 μm in size, consisting of a large number of bilayer phospholipid membranes [13-15]. the presence of phosphoglycerides (e.g., phosphatidylethanolamine, lysophosphatidylcholine, and products of lipid peroxidation) in the antigen leads to the appearance in liposomes of regions with differences in the bilayer admet & dmpk 00(0) (2023) 000-000 liposomes: from august wassermann to vaccines against covid-19 doi: https://doi.org/10.5599/admet.1926 3 structure. the violation of the integrity of the bilayer membrane results in a distortion of the biological activity of ag that decreases the specificity of serological reactions [16]. figure 2. (a) structure of cardiolipin, pc, and chol. (b) liposomal structure of the cardiolipin antigen based on cardiolipin, pc, and chol. dpg is the only component of ag capable of directly interacting with the active centers of antilipid antibodies [10,17]. in this case, pc and chol are considered structural components of ag, which allow the antigen-antibody complex to bind, complement or form a precipitate. perhaps, pc forms a phospholipid bilayer where the antigenic determinants of dpg are located, and chol provides rigidity to the lipid structure. the study of the ultrastructure of cardiopin ag showed that its sensitivity and specificity were significantly reduced in the absence of chol [18]. the study of the morphology of vdrl liposomes showed that most of chol is located outside the lamellar membranes of pc and dpg, and its role is in the mechanical surface dispergation of dpg and pc [19]. presumably, chol does not affect the structure of epitopes that react with antibodies in the serum of syphilis patients. antilipid antibodies in the infectious process lipid ags used to diagnose syphilis should functionally detect antibodies that specifically bind to antigenic determinants on the surface of liposomal nanoparticles. over the past 100 years, researchers in the field of immunology and lipidology have offered various explanations for the origin of antilipid antibodies in syphilis and other infectious diseases. the possibility of releasing lipid components from damaged treponema cells, https://doi.org/10.5599/admet.1926 g. grygorieva et al. admet & dmpk 00(0) (2023) 000-000 4 which naturally leads to an autoimmune reaction, was discussed. the predictable phenomenon of induction of the so-called "non-treponemal" lipid antibodies in the infected organism in response to the treponema pallidum lipid complex is also of interest. to confirm this hypothesis, the lipid composition of different strains of treponema pallidum was studied by reiter, kazans, nichols, and others [20-24]. in treponema pallidum, dpg, pc, and chol were identified as basic lipids, which content and ratio significantly depended on the nature of the strain, and the presence of varying amounts of sphingomyelin, phosphatidylethanolamine, phosphatidylinositol, and lysophosphatidylcholine was also determined. the average content of dpg in treponema pallidum is estimated at 13-18 %, except for the kazans strain, in which only trace amounts of dpg were determined at a relatively high content of pc (up to 40 %). perhaps, the presence of the lipid complex dpg, pc, and chol in the membrane of treponema pallidum causes the appearance of antilipid antibodies specific to dpg in patients with syphilis. when studying the localization of cardiolipin ag in treponema, it was found that lipids are present in the outer membrane of the causative agent of syphilis [25,26]. cardiolipin antigen in immunodiagnostics during immunization of animals with compositions based on dpg or pi, pc, and chol, antilipid antibodies in serum was detected using complement fixation reaction and microprecipitation reaction [27]. in this case, the immunogenicity of lipids is directly related to their liposomal organization. the possibility of obtaining antilipid antibodies against anionic phospholipids (dpg and pg) by long-term immunization of rabbits with cardiolipin ag and ag complexes with methylated bsa (bovine serum albumin) or cytochrome c has been shown [24,28]. antibodies interacting with ag in low titers are detected in the blood of animals that were immunized with individual lipids. the use of complexes of cardiolipin ag with proteins causes a significant increase in titers of antilipid antibodies, most significant for cytochrome c [29]. the composition of antilipid antibodies interacting with dpg is mainly represented by gand m-immunoglobulins, but iga is also detected in severe stages of the disease [30]. the high titer of antilipid antibodies in syphilitic infection is the result of the combined action of both treponema pallidum cardiolipin ag and cellular dpg from damaged host tissues. both factors can lead to the formation of anticardiolipin antibodies [20]. in recent years, this thesis has been confirmed not only for treponema pallidum, but also for other infections [31]. lipids play an irreplaceable structure-forming and functional role in various cellular processes, such as membrane fusion, division, endocytosis, transfer and functions of proteins, etc. at the same time, the influence of lipids on the development of the infectious process can be realized both as a result of interaction with the host membrane to facilitate the penetration of pathogens, and as a result of the support of certain functions of a pathogenic microorganism to increase its survival and proliferation. these features of the interaction with host lipids explain the appearance of antibodies against lipids, in particular, against phospholipids [32]. using antiphospholipid elisa, the immune response in the form of the production of antibodies against host phospholipids was studied in animals (mice) and humans [32]. several phospholipids (namely ps, pa, pc) are targets for antibodies that occur at the early stage of infection, and a humoral immune response occurs to the same phospholipids in patients with acute infection and infected mice. the use of the liposomal form of cardiolipin ag, which has already become quite routine in complement fixation, agglutination, or microprecipitation reactions, has recently been enriched by the creation of new methodological approaches to the detection of antilipid antibodies [33]. an innovative platform for the detection of anticardiolipin antibodies in the serum of patients with syphilis was proposed using a selfassembled (3-mercaptopropyl)-trimethoxysilane monolayer for immobilization of an antigenic suspension admet & dmpk 00(0) (2023) 000-000 liposomes: from august wassermann to vaccines against covid-19 doi: https://doi.org/10.5599/admet.1926 5 containing dpg, pc, and chol as a biorecognition element [34]. the biosensor platform with a detection limit of anticardiolipin antibodies at 1:1024 and high selectivity for nonspecific biomolecules was evaluated by cyclic voltammetry, spectroscopy, electrochemical impedance, and atomic force microscopy. most of the current research on the detection and quantification of antiphospholipid antibodies is focused on enzyme-linked immunosorbent assay (elisa) or radioimmunoassay (ria). the algorithm of elisa is based on the inhibition of the activity of certain enzymes (namely peroxidase and glucose oxidase) due to the interaction of the dpg-enzyme complex with antibodies. the use of cardiolipin ag in the detection of antilipid antibodies by elisa was proposed [35]. liposomal forms of antigens are used to diagnose infectious diseases and detect antibodies in various pathological conditions in which antilipid antibodies appear. studies of ag in serodiagnosis of syphilis actually became the foundation for using antilipid antibodies for the diagnosis of various infectious diseases and pathological conditions, such as tuberculosis [36,37], lupus erythematosus [38], covid-19 [39], borreliosis [32], etc. [41,42]. liposomal forms of antigens containing secretory proteins of mycobacteria are also proposed for the diagnosis of tuberculosis [43]. liposomal antigens are used to study a number of allergic diseases in their diagnosis and obtain antibodies in model experiments [44]. identification of antiphospholipid antibodies (mainly anticardiolipin) is the most important diagnostic tool for a serious disease called antiphospholipid syndrome [40,45]. in this case, the diagnosis is carried out using liposomal antigens, mainly with cardiolipin. antiphospholipid antibodies correlate with a predisposition to arterial or venous thrombosis, fetal loss, or thrombopenia. however, levels of antilipid antibodies, particularly anticardiolipin antibodies, do not correlate well with disease activity or specific features such as arthritis or kidney damage. nevertheless, the presence of antilipid antibodies serves as a marker of thromboembolic complications. for these studies, it is possible to use antigens in liposomal form [46]. thrombosis-associated antibodies detected in primary and secondary antiphospholipid syndrome also appear in lupus, myocardial infarction in patients with neurological symptoms, showing a clear correlation with the extent of the disease. it is important to note that antibodies in antiphospholipid syndrome have a different specificity compared to antibodies that accompany common infections, such as syphilis, malaria, parasitic diseases, and infectious mononucleosis [44]. liposome in covid-19 vaccine development in the second half of the 20th century the development of research in the field of biophysics and biochemistry of lipids not only introduced the concepts of "lipid nanoparticle" and "liposome" in relation to immunochemical reactions involving lipids [47,48] but also resulted in the development of a new direction in pharmacy and medicine associated with the design and use of a class of innovative liposomal drugs for diagnosis, pharmacotherapy and immunoprophylaxis [49-54]. in 2020-2021, the appearance of vaccines of various structures and compositions against the sars-cov-2 virus on the global pharmaceutical market became a milestone achievement in the field of liposomal technologies [55,56]. the high pace of vaccine development in the alarming conditions of the covid-19 pandemic was predetermined by the purposeful combination of the results of applied research on a wide class of lipids with fundamental information about the properties of messenger rna (mrna). mrna is synthesized during dna transcription, contains information about the primary structure of proteins, and is used during translation as a template for their subsequent synthesis [57]. mrna technology provides an innovative platform for in situ antigen expression with the advantage that mrna-based vaccines (unlike dna-based) do not integrate into chromosomes, avoiding the risks of oncogenesis and insertional mutagenesis [58-60]. https://doi.org/10.5599/admet.1926 g. grygorieva et al. admet & dmpk 00(0) (2023) 000-000 6 g. grigoriadis was the first to formulate significant advantages of the liposomal organization of mrnabased vaccines [61,62]. mrna encapsulated in the liposome is completely protected from nuclease attack in the bloodstream and penetrates the cell cytoplasm by endocytosis. at the same time, mrna in cationic liposomes escapes the lysosomotropic pathway and remains intact in the cytoplasm. in the cytoplasm, mrna is expressed as a spike protein, after which liposomes or lipid fragments exhibit their immunological adjuvant effect according to unclear mechanisms. these factors provide the possibility of quantitative functional encapsulation of mrna into liposomes and safe biodegradation of nanoparticles. encapsulation of mrna into lipid nanoparticles (lnps) in the development of vaccines ensures the capture of protected mrna by host cells and its delivery into the cytosol, where the mrna sequence is translated into s-protein in ribosomes [63]. after post-translational processing in host cells, the s protein as a membrane-bound antigen provides the target ag for b cells. intramuscular administration of mrna vaccine based on lnps leads to temporary punctate inflammation, which promotes the recruitment of neutrophils and antigen-presenting cells (apc) to the delivery site [64]. the mrna part of the genetic code of the virus leads to the synthesis of antigenic protein structures, in response to which specific antibodies appear in the human body [65]. after the administration of mrna during vaccination, viral spike proteins begin to be produced, and neutralizing antibodies against spike proteins and cellular immune responses can prevent infection. the mrna vaccine development platform based on lipid nanosystem was first used for covid-19 vaccine creation by pfizer-biontech and moderna [66]. unlike traditional vaccines, which can deliver an inactivated or attenuated version of the virus (such as a capsular protein), these vaccines deliver genetic information to the body, synthesizing a protein to generate an immune response. the mrna vaccine against covid-19 encodes the spike glycoprotein s spike protein of the sars-cov-2 virus, which is used by the virus to penetrate into human cells. in the composition of mrna vaccines against covid-19, the lipid component is represented by several lipids, among which so-called "ionized" lipid is functionally accentuated, and the rest stabilize the structure of lnps and provide the stability of the vaccine in the body (figure 3) [55]. figure 3. structure of the lnp in mrna vaccine [55] similarly to classical liposomes [2], the cationic charge of the main lipid contributes to obtaining the optimal structure of lnps. the pfizer-biontech covid-19 vaccine contains four lipid components [67,68]: 50 % ionized admet & dmpk 00(0) (2023) 000-000 liposomes: from august wassermann to vaccines against covid-19 doi: https://doi.org/10.5599/admet.1926 7 lipid ((4-hydroxybutyl)azandiyl)-bis(hexane-6,1-diyl)-bis(2-hexyl-decanoate)), 10 % dspc, 38.5 % chol and 1.5 % peg lipid ([(polyethylene glycol)-2000]-n,n-ditradecylacetamide). the lnp size ranges from 100 to 170 nm. the algorithm for polylipid composition design of mrna vaccine aims to form lnp and encapsulate the mrna into its structure due to the interaction of the anionic phosphate backbone of the mrna with the ionized lipid. in the design process, the mrna contacts four lipids in a water-organic environment under the control of ph level and hydrophobic-hydrophilic balance to ensure the binding of the mrna to the ionized lipid. bound mrna is identified in the inner cavity of multilayer lnps formed by complementary lipid components. in this case, dspc forms the main bilayer of the liposomal membrane, and the presence of peglipid and chol ensures optimal hydrophilicity, stability, and size of lnps with encapsulated mrna. a liposome-based vaccine candidate eg-covid differs from the composition of pfizer-biontech product, consists of dotap (1,2-dioleoyl-3-trimethylammonium-propane), dope (1,2-dioleoyl-sn-glycero-3-phosphoethanolamine) and chol, induces stable humoral and cellular immunity to the sars-cov-2 virus, and suppresses the sars-cov-2 viral infection in vero cells [69]. the optimization of the lnp lipid spectrum of the composition of the lyophilized form of eg-covid can increase compliance in the standardization of the product and the process of immunoprophylaxis. conclusions in conclusion, the significant role of liposomes in the development of liposomal vaccines against covid19 and the identification of immunochemical reactions involving lipids should be assessed in the context of the development of the pioneering idea of august wassermann about the diagnostic use of liposomal antigens in the creation of innovative liposomal preparations for the diagnosis and immunoprophylaxis of serious human diseases. thus, the liposomal system first proposed in 1906 for the diagnosis of syphilis (figure 2) evolved more than a century later into mrna vaccines (figure 3), which are used today in the fight against the covid-19 pandemic. conflict of interest: the author declares no conflict of interest. references [1] u. bulbake, s. doppalapudi, n. kommineni, w. khan. liposomal formulations in clinical use: an updated review. pharmaceutics 9 (2017) 12. https://doi.org/10.3390/pharmaceutics9020012 [2] v.i. shvets, yu.m. krasnopolsky, g.m. sorokoumova. liposomal forms of drugs: technological features of production and use in the clinic. remedium, moscow, 2016, p. 200. [3] a. wassermann, a. neisser, c. bruck. eine serodiagnostische reaktion bei syphilis. deutsche medicinische wochenschrift 32 (1906) 745-746. https://doi.org/10.1055/s-0028-1142018. [4] m.c. pangborn. isolation and purification of a serologically active phospholipid from beef heart. journal of biological chemistry 143 (1942) 247-256. https://doi.org/10.1016/s0021-9258(18)72683-5. [5] m.c. pangborn. the composition of cardiolipin. journal of biological chemistry 168 (1947) 351-361. https://doi.org/10.1016/s0021-9258(17)35123-2. [6] g.l. orlova, yu.m. krasnopolsky, i.i. golbets. method for obtaining highly purified cardiolipin. chemistry and technology of organic production 9 (1979) 86-91. [7] v.i. shvets, g.a. sennikov, i.i. holbets. production of purified lecithin. pharmaceutical journal 4 (1977) 79-81. [8] r.l. kahn, e.b. mcdermott. kahn reactions with cardiolipin antigen compared with kahn antigen, with a note on microflocculation procedure with cardiolipin antigen. university hospital bulletin 12 (1946) 81-84. [9] g.a. sennikov, l.s. reznikova, v.i. shvets, i.i. gol'bets, g.l. orlova. izuchenie optimal'nogo sostava kardiolipinovogo antigena dlia serodiagnostiki sifilisa study of the optimal composition of cardiolipin antigen for serodiagnosis of syphilis. vestnik dermatologii i venerologii 7 (1978) 48–52. https://doi.org/10.5599/admet.1926 https://doi.org/10.3390/pharmaceutics9020012 https://doi.org/10.1055/s-0028-1142018 https://doi.org/10.1016/s0021-9258(18)72683-5 https://doi.org/10.1016/s0021-9258(17)35123-2 g. grygorieva et al. admet & dmpk 00(0) (2023) 000-000 8 [10] g.a. sennikov, i.i. golbets, y.m. krasnopolskyi, g.l. orlova. chemical criteria in the assessment of the quality of cardiolipin antigen used in the serodiagnosis of syphilis. pharmaceutical journal 3 (1977) 5760. [11] v.t. kung, y.p. vollmer, f.j. martin. large liposome agglutination technique for the serological detection of syphilis. journal of immunological methods 90 (1986) 189-196. https://doi.org/10.1016/ 0022-1759(86)90075-x [12] v.t. rung, f.v. martin, y.p. volmer, (cooper lipotech inc.). large-liposome agglutination reagent and method. us 4605630 a (1980). [13] yu.m. krasnopol'skii, i.i. gol'bets, g.a. sennikov, v.i. shvets. immunology of lipids (survey). pharmaceutical chemistry journal 15 (1981) 455–465. [14] i.a. vasylenko, v.v. chupin, yu.m. krasnopolsky. relationship between the structure and properties of the cardiolipin antigen. chemical pharmaceutical journal 15 (1981) 14-18. [15] i.a. vasilenko, v.v. chupin, yu.m. krasnopol'skii, i.i. gol'bets, g.a. sennikov, v.i. shvets, r.p. evstigneeva. interrelationship of the structure and properties of the cardiolipin antigen. pharmaceutical chemistry journal 15 (1981) 71–74. https://doi.org/10.1007/bf01149144 [16] yu.m. krasnopolsky, i.i. golbets, g.l. orlova, influence of fatty acid composition and degree of lipid oxidation on the immunochemical activity of cardiolipin antigen. vestnik dermatologii i venerologii 8 (1986). 51-56. [17] i.i. golbets, yu.m. krasnopolsky, g.l. orlova. on the chemical composition of industrial lipid cardiolipin antigens. chemistry and technology of organic production 7 (1977) 24-26. [18] t. takashi, k. inoue, s. nojima. immune reactions of liposomes containing cardiolipin and their relation to membrane fluidity. journal of biochemistry 87 (1980) 679-685. https://doi.org/10.1093/oxford journals.jbchem.a132796 [19] o. orum, j.r. nielsen, a. birch-andersen. the effect of cholesterol on the morphology and reactivity of the mixture of lipids used in syphilis serology. apmis 98 (1990) 9-18. https://doi.org/10.1111/j.16990463.1990.tb00995.x [20] k. gao, x. shen, y. lin, x.z. zhu, l.r. lin, m.l. tong, y. xiao, h.l. zhang, x.m. liang, j.j. niu, l.l. liu, t.c. yang. origin of nontreponemal antibodies during treponema pallidum infection: evidence from a rabbit model. journal of infectious diseases 218 (2018) 835-843. https://doi.org/10.1093/infdis/jiy241 [21] d.r. schultz. antiphospholipid antibodies: basic immunology and assays. seminars in arthritis and rheumatism 26 (1997) 724–739. https://doi.org/10.1016/s0049-0172(97)80041-8 [22] r.c. johnson, b.p. livermore, h.m. jenkin, l. eggebraten. lipids of treponema pallidum kazan 5. infection and immunity 2 (1970) 606-609. https://doi.org/10.1128/iai.2.5.606-609.1970 [23] h.m. matthews, t.k. yang, h.m. jenkin. unique lipid composition of treponema pallidum (nichols virulent strain infection and immunity 24 (1979) 713-719. https://doi.org/10.1128/iai.24.3.713-719.1979 [24] v.i. shvets, yu.m. krasnopolsky. the main directions of lipid immunochemistry. ukrainian biochemical journal 56 (1984) 254-263. [25] q.l. li, q.y. xu, k. gao, h.l. zhang, l.l. liu, l.r. lin, j.j. niu, t.c. yang. membrane location of cardiolipin antigen in treponema pallidum: further study on the origin of nontreponemal antibodies. future microbiology 17 (2022) 873-886. https://doi.org/10.2217/fmb-2021-0245 [26] a. luchini, d. cavasso, a. radulescu, g. d'errico, l. paduano, g. vitiello. structural organization of cardiolipin-containing vesicles as models of the bacterial cytoplasmic membrane. langmuir 37 (2021) 8508-8516. https://doi.org/10.1021/acs.langmuir.1c00981 [27] m. guarnieri. reaction of cardiolipin and phosphatidylinositol antisera with phospholipid antigens. lipids 9 (1974) 692-695. https://doi.org/10.1007/bf02532177 [28] a.j. desiervo. anti-cardiolipin and anti-phosphatidylglycerol antibodies prepared against bacterial phospholipids. infection and immunity 9 (1974) 835-838. https://doi.org/10.1128/iai.9.5.835-838.1974 [29] yu.m. krasnopolsky, g.l. orlova, i.i. golbets. to the question of the immunogenicity of lipids. chemistry and technology of organic production 1x (1979) 80-85. https://doi.org/10.1007/bf01149144 https://doi.org/10.1093/oxfordjournals.jbchem.a132796 https://doi.org/10.1093/oxfordjournals.jbchem.a132796 https://doi.org/10.1016/s0049-0172(97)80041-8 https://doi.org/10.1128/iai.2.5.606-609.1970 https://doi.org/10.1128/iai.24.3.713-719.1979 https://doi.org/10.2217/fmb-2021-0245 https://doi.org/10.1021/acs.langmuir.1c00981 https://doi.org/10.1007/bf02532177 https://doi.org/10.1128/iai.9.5.835-838.1974 admet & dmpk 00(0) (2023) 000-000 liposomes: from august wassermann to vaccines against covid-19 doi: https://doi.org/10.5599/admet.1926 9 [30] v.i. shvets, g.a. sennikov, i.i. golbets, yu.m. krasnpolsky. immunochemical structural study of syphilitic reagins. voprosy meditsinskoi khimii 26 (1980) 55-58. [31] r. dadhich, s. kapoor. various facets of pathogenic lipids in infectious diseases: exploring virulent lipidhost interactome and their druggability. journal of membrane biology 253 (2020) 399-423. https://doi.org/10.1007/s00232-020-00135-0 [32] p.j. gwynne, l.h. clendenen, s.p. turk, a.r. marques, l.t. hu. antiphospholipid autoantibodies in lyme disease arise after scavenging of host phospholipids by borrelia burgdorferi. journal of clinical investigation 132 (2022) e152506. https://doi.org/10.1172/jci152506 [33] j. ngeh-ngwainbi, s.s. kuan, f. steinman, g.g. guilbault. the cardiolipin antigen: chemistry and composition. biotechnology and applied biochemistry 8 (1986) 553-563. [34] e.m.n. do egito, a.g. silva-júnior, r.p.s. lucena, m.d.l. oliveira, c.a.s. andrade. electrochemical platform for anti-cardiolipin antibody detection in human syphilitic serum. current research in biotechnology 104 (2022) 58-65. https://doi.org/10.1016/j.crbiot.2022.01.001 [35] t.m. dhason, e.l. jairaj, r. sankaralingam, b. mahendren, b. chilukuri, s. vengudusamy, m. seetharaman. role of anticardiolipin antibodies in bad obstetric history detected by elisa test in a tertiary care centre. journal of immunology and clinical microbiology 2 (2017) 43-47. https://doi.org/10.5455/jicm.24.25 9659 [36] yu.m. krasnopolsky, i.i. golbets, g.a. sennikov. antigenic activity of phospholipids of mycobacteria. problems of tuberculosis 11 (1985) 55-58. [37] c. ordoñez, h.p. savage, m. tarajia, r. rivera, c. weeks-galindo, d. sambrano, l. riley, p.l. fernandez, n. baumgarth, a. goodridge. both b-1a and b-1b cells exposed to mycobacterium tuberculosis lipids differentiate into igm antibody-secreting cells. immunology 154 (2018) 613–623. https://doi.org/10.1111/imm.12909 [38] a.s. sanoff, m.j. ostro, m.j. weiner, g. weissmann, j.r. seibold, (the lipodome company, inc.), liposome composition for lupus assay, us 4564599a, (1983). [39] a. hollerbach, n. müller-calleja, d. pedrosa, a. canisius, m.f. sprinzl, t. falter, h. rossmann, m. bodenstein, c. werner, i. sagoschen, t. münzel, o. schreiner, v. sivanathan, m. reuter, j. niermann, p.r. galle, l. teyton, w. ruf, k.j. lackner. pathogenic lipid-binding antiphospholipid antibodies are associated with severity of covid-19. journal of thrombosis and haemostasis 19 (2021) 2335-2347. https://doi.org/10.1111/jth.15455 [40] s. hörkkö, e. miller, d.w. branch, w. palinski, j.l. witztum. the epitopes for some antiphospholipid antibodies are adducts of oxidized phospholipid and beta2 glycoprotein 1 (and other proteins). proceedings of the national academy of sciences of the united states of america 94 (1997) 1035610361. https://doi.org/10.1073/pnas.94.19.10356 [41] c. landa-saldívar, a. reséndiz-mora, s. sánchez-barbosa, a. sotelo-rodríguez, g. barrera-aveleida, i. nevárez-lechuga, i. galarce-sosa, k. taniguchi-ponciano, o.d.r. cruz-guzmán, i. wong-baeza, a. escobar-gutiérrez, i. baeza, c. wong-baeza. liposomes bearing non-bilayer phospholipid arrangements induce specific igg anti-lipid antibodies by activating nk1.1+, cd4+ t cells in mice. membranes (basel) 12 (2022) 643. https://doi.org/10.3390/membranes12070643 [42] s.g. barreno-rocha, s. guzmán-silahua, s.d. rodríguez-dávila, g.e. gavilanez-chávez, e.g. cardonamuñoz, c. riebeling-navarro, b. rubio-jurado, a.h. nava-zavala. antiphospholipid antibodies and lipids in hematological malignancies. international journal of molecular sciences 23 (2022) 4151. https://doi.org/10.3390/ijms23084151 [43] d. tiwari, s. haque, r.p. tiwari, a. jawed, t. govender, h.g. kruger. fast and efficient detection of tuberculosis antigens using liposome encapsulated secretory proteins of mycobacterium tuberculosis. j microbiol immunol infect 50 (2017) 189-198. https://doi.org/10.1016/j.jmii.2015.05.014 [44] k.j. bednar, l. hardy, j. smeekens, d. raghuwanshi, s. duan, m.d. kulis, m.s. macauley. antigenic liposomes for generation of disease-specific antibodies. journal of visualized experiments 140 (2018) 58285. https://doi.org/10.3791/58285 https://doi.org/10.5599/admet.1926 https://doi.org/10.1007/s00232-020-00135-0 https://doi.org/10.1172/jci152506 https://doi.org/10.1016/j.crbiot.2022.01.001 https://doi.org/10.5455/jicm.24.259659 https://doi.org/10.5455/jicm.24.259659 https://doi.org/10.1111/imm.12909 https://doi.org/10.1111/jth.15455 https://doi.org/10.1073/pnas.94.19.10356 https://doi.org/10.3390/membranes12070643 https://doi.org/10.3390/ijms23084151 https://doi.org/10.1016/j.jmii.2015.05.014 https://doi.org/‌10.3791/‌58285 g. grygorieva et al. admet & dmpk 00(0) (2023) 000-000 10 [45] d. green. pathophysiology of antiphospholipid syndrome. thrombosis and haemostasis 122 (2022) 1085-1095. https://doi.org/10.1055/a-1701-2809 [46] a. radu, s.c. dudu, a. ciobanu, g. peltecu, g. iancu, r. botezatu, n. gica, a.m. panaitescu. pregnancy management in women with antiphospholidic syndrome. maedica (bucur) 14 (2019) 148-160. https://doi.org/10.26574/maedica.2019.14.2.148 [47] a.d. bangham, r.w. horne. negative staining of phospholipids and their structural modification by surface-active agents as observed in the electron microscope. journal of molecular biology 8 (1964) 660-668. https://doi.org/10.1016/s0022-2836(64)80115-7 [48] n. dusgunes, g. gregoriadis. introduction: the origins of liposomes: alec bangham at babraham. methods in enzymology 391 (2005). 1-3. https://doi.org/10.1016/s0076-6879(05)91029-x [49] d. guimarães, a. cavaco-paulo, e. nogueira. design of liposomes as drug delivery system for therapeutic applications. international journal of pharmaceutics 601 (2021) 120571. https://doi.org/ 10.1016/j.ijpharm.2021.120571 [50] t.m. allen, p.r. cullis. liposomal drug delivery systems: from concept to clinical applications. advanced drug delivery reviews 65 (2013) 36-48. https://doi.org/10.1016/j.addr.2012.09.037 [51] e. beltrán-gracia, a. lópez-camacho, i. higuera-ciapara, j.b velázquez-fernández, a.a. vallejocardona. nanomedicine review: clinical developments in liposomal applications. cancer nano 10 (2019) 11. https://doi.org/10.1186/s12645-019-0055-y [52] m. alavi, m. hamidi. passive and active targeting in cancer therapy by liposomes and lipid nanoparticles. drug metabolism and personalized therapy 34 (2019). https://doi.org/10.1515/dmpt-2018-0032 [53] m.a. mohamed, e. alaaeldin, a.k. hussein, h.a. sarhan. liposomes and pegylated liposomes as drug delivery system. journal of advanced biomedical and pharmaceutical sciences 3 (2020) 80-88. https://doi.org/10.21608/jabps.2020.22937.1068 [54] i.m. mikheytseva, g.s. grygorieva, n.v. pasyechnikova, s.g. kolomiichuk, t.i. siroshtanenko, n.f. konakhovych. ocular hypotensive efficacy of a new liposomal latanoprost formulation administered by different routes for experimental ocular hypertension. journal of ophthalmology 505 (2022) 37-41. [55] m.d. buschmann, m.j. carrasco, s. alishetty, m. paige, m.g. alameh, d. weissman. nanomaterial delivery systems for mrna vaccines. vaccines (basel) 9 (2021) 65. https://doi.org/10.3390/vaccines 9010065 [56] n. pardi, m.j. hogan, d. weissman. recent advances in mrna vaccine technology. current opinion in immunology 65 (2020) 14-20. https://doi.org/10.1016/j.coi.2020.01.008 [57] d.k. lvov, s.v. alkhovsky source of the covid-19 pandemic: ecology and genetics of coronaviruses (betacoronavirus: coronaviridae) sars-cov, sars-cov-2 (subgenus sarbecovirus), and mers-cov (subgenus merbecovirus). problems of virology 65 (2020) 62-70. https://doi.org/10.36233/0507-40882020-65-2-62-70 [58] l.a. brito, s. kommareddy, d. maione, y. uematsu, c. giovani, f. berlanda scorza, g.r. otten, d. yu, c.w. mandl, p.w. mason, p.r. dormitzer, j.b. ulmer, a.j. geall. self-amplifying mrna vaccines. advances in genetics 89 (2015) 179-233. https://doi.org/10.1016/bs.adgen.2014.10.005 [59] n. pardi, d. weissman. nucleoside modified mrna vaccines for infectious diseases. methods in molecular biology (clifton, n.j.) 1499 (2017) 109-121. https://doi.org/10.1007/978-1-4939-6481-9_6 [60] n. pardi, m.j. hogan, m.s. naradikian, k. parkhouse, d.w. cain, l. jones, m.a. moody, h.p. verkerke, a. myles, e. willis, c.c. labranche, d.c. montefiori, j.l. lobby, k.o. saunders, h.x. liao, b.t. korber, l.l. sutherland, r.m. scearce, p.t. hraber, i. tombácz, h. muramatsu, h. ni, d.a. balikov, c. li, b.l. mui, y.k. tam, f. krammer, k. karikó, p. polacino, l.c. eisenlohr, t.d. madden, m.j. hope, m.g. lewis, k.k. lee, s.l. hu, s.e. hensley, m.p. cancro, b.f. haynes, d. weissman. nucleoside-modified mrna vaccines induce potent t follicular helper and germinal center b cell responses. journal of experimental medicine 215 (2018) 1571-1588. https://doi.org/10.1084/jem.20171450 [61] d. davis, g. gregoriadis liposomes as immunological adjuvants in vaccines: studies with entrapped and surface-linked antigen. biochemical society transactions 6 (1986) 1036-1037. https://doi.org/10.1042/ bst0141036 https://doi.org/10.1055/a-1701-2809 https://doi.org/10.26574/maedica.2019.14.2.148 https://doi.org/10.1016/s0022-2836(64)80115-7 https://doi.org/10.1016/s0076-6879(05)91029-x https://doi.org/10.1016/j.addr.2012.09.037 https://doi.org/10.1186/s12645-019-0055-y https://doi.org/10.1515/dmpt-2018-0032 https://doi.org/10.21608/jabps.2020.22937.1068 https://doi.org/10.3390/‌vaccines‌9010065 https://doi.org/10.3390/‌vaccines‌9010065 https://doi.org/10.1016/j.coi.2020.01.008 https://doi.org/10.36233/0507-4088-2020-65-2-62-70 https://doi.org/10.36233/0507-4088-2020-65-2-62-70 https://doi.org/10.1016/bs.adgen.2014.10.005 https://doi.org/10.1007/978-1-4939-6481-9_6 https://doi.org/10.1084/jem.20171450 admet & dmpk 00(0) (2023) 000-000 liposomes: from august wassermann to vaccines against covid-19 doi: https://doi.org/10.5599/admet.1926 11 [62] gregoriadis g. liposomes and mrna: two technologies together create a covid-19 vaccine. medicine in drug discovery 12 (2021) 100104. https://doi.org/10.1016/j.medidd.2021.100104 [63] y. huang, c. yang, x.f. xu, w. xu, s.w. liu. structural and functional properties of sars-cov-2 spike protein: potential antivirus drug development for covid-19. acta pharmacologica sinica 41 (2020) 1141-1149. https://doi.org/10.1038/s41401-020-0485-4 [64] l. schoenmaker, d. witzigmann, j.a. kulkarni, r. verbeke, g. kersten, w. jiskoot, d.j.a. crommelin. mrna-lipid nanoparticle covid-19 vaccines: structure and stability. international journal of pharmaceutics 601 (2021) 120586. https://doi.org/10.1016/j.ijpharm.2021.120586 [65] x. hou, t. zaks, r. langer, y. dong. lipid nanoparticles for mrna delivery. nature reviews materials 6 (2021) 1078–1094. https://doi.org/10.1038/s41578-021-00358-0 [66] y. krasnopolsky, d. pylypenko. licensed liposomal vaccines and adjuvants in the antigen delivery system. biotechnologia (pozn) 103 (2022) 409-423. https://doi.org/10.5114/bta.2022.120709 [67] k.j. hassett, k.e. benenato, e. jacquinet, a. lee, a. woods, o. yuzhakov, s. himansu, j. deterling, b.m. geilich, t. ketova, c. mihai, a. lynn, i. mcfadyen, m.j. moore, j.j. senn, m.g. stanton, ö. almarsson, g. ciaramella, l.a. brito. optimization of lipid nanoparticles for intramuscular administration of mrna vaccines. molecular therapy. nucleic acids 15 (2019) 1-11. https://doi.org/10.1016/j.omtn.2019.01.013 [68] l.a. jackson, e.j. anderson, n.g. rouphael, p.c. roberts, m. makhene, et al. an mrna vaccine against sars-cov-2 preliminary report. new england journal of medicine 383 (2020) 1920-1931. https://doi.org/10.1056/nejmoa2022483 [69] h.c. hong, k.s. kim, s.a. park, m.j. chun, e.y. hong, s.w. chung, h.j. kim, b.g. shin, a. braka, j. thanappan, s. jang, s. wu, y.j. cho, s.h. kim. an mrna vaccine against sars-cov-2: lyophilized, liposome-based vaccine candidate eg-covid induces high levels of virus neutralizing antibodies. biorxiv preprint (2021) https://doi.org/10.1101/2021.03.22.436375. ©2023 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) https://doi.org/10.5599/admet.1926 https://doi.org/10.1016/j.medidd.2021.100104 https://doi.org/10.1038/s41401-020-0485-4 https://doi.org/10.1016/j.ijpharm.2021.120586 https://doi.org/10.1038/s41578-021-00358-0 https://doi.org/10.5114/bta.2022.120709 https://doi.org/10.1016/j.omtn.2019.01.013 https://doi.org/10.1056/nejmoa2022483 https://doi.org/10.1101/2021.03.22.436375 http://creativecommons.org/licenses/by/3.0/ manuscript doi: 10.5599/admet.4.1.258 35 admet & dmpk 4(1) (2016) 35-53; doi: 10.5599/admet.4.1.258 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper simulated rat intestinal fluid improves oral exposure prediction for poorly soluble compounds over a wide dose range jörg berghausen 1, *, frank h. seiler 2 , nathalie gobeau 1,3 and bernard faller 1 1 novartis institutes for biomedical research, novartis campus, ch-4002 basel 2 chemical and pharmaceutical profiling, novartis pharma ag, novartis campus, ch-4002 basel 3 current address: medicines for malaria venture, 20, route de pré-bois, 1215 geneva 15 *corresponding author: e-mail: joerg.berghausen@novartis.com; tel.: +41 61 69 62335 received: november 18, 2015; revised: march 12, 2016; published: march 31, 2016 abstract solubility can be the absorption limiting factor for drug candidates and is therefore a very important input parameter for oral exposure prediction of compounds with limited solubility. biorelevant media of the fasted and fed state have been published for humans, as well as for dogs in the fasted state. in a drug discovery environment, rodents are the most common animal model to assess the oral exposure of drug candidates. in this study a rat simulated intestinal fluid (rsif) is proposed as a more physiologically relevant media to describe drug solubility in rats. equilibrium solubility in this medium was tested as input parameter for physiologically-based pharmacokinetics (pbpk) simulations of oral pharmacokinetics in the rat. simulations were compared to those obtained using other solubility values as input parameters, like buffer at ph 6.8, human simulated intestinal fluid and a comprehensive dissolution assay based on rsif. our study on nine different compounds demonstrates that the incorporation of rsif equilibrium solubility values into pbpk models of oral drug exposure can significantly improve the reliability of simulations in rats for doses up to 300 mg/kg compared to other media. the comprehensive dissolution assay may help to improve further simulation outcome, but the greater experimental effort as compared to equilibrium solubility may limit its use in a drug discovery environment. overall, pbpk simulations based on solubility in the proposed rsif medium can improve prioritizing compounds in drug discovery as well as planning dose escalation studies, e.g. during toxicological investigations. keywords computational adme; preclinical pharmacokinetics; pbpk; modelling; solubility introduction the use of physiologically-based pharmacokinetic (pbpk) modeling in drug development has matured and aims at prediction of plasma-concentration time profiles based on both in silico and in vitro parameters. regardless of the advancement of the modeling software itself, the success of pbpk approaches to support decision making will strongly depend on the use of appropriate input parameters. this is particularly true in drug discovery, when the robustness of input parameters is naturally lower compared to data generated in later stage development. a recent overview on the use of pbpk models for oral dosage forms is given by kostewicz et al. [1]. many reports about its use and years of application at various stages of drug development have generated http://www.pub.iapchem.org/ojs/index.php/admet/index berghausen et al. admet & dmpk 4(1) (2016) 35-53 36 two main strategies for implementation of pbpk in early drug discovery. on the one hand simulation approaches can be compared to observed data as a strategy to build hypothesis regarding differences and sensitivity of input parameters. to prove or disprove such a hypothesis, results may be used to initiate further studies in order to identify the underlying reason for discrepancies between simulated and observed results. another strategy is to use pbpk approaches to perform pharmacokinetic prediction across species, contributing to increased confidence in the design of preclinical toxicology studies and ultimately reliable prediction of human pharmacokinetics. in order for pbpk approaches to have impact in a drug discovery setting we need to be confident of performance reliability. the quality of pbpk simulations during this early phase is often compromised by the lack of validated input data. there is greater confidence with human pbpk prediction based on the fact that at this stage of drug development, a broad set of robust input data is available. in addition, many aspects of human physiology have been studied in much greater detail compared to preclinical species. although rodents, especially rats, are still the most relevant screening model in drug discovery, many parameters relevant for drug adme properties are not known in detail or show a high variability. even within one defined rat strain, a broad range of values is typically reported in the literature [2,3]. it is clear that solubility can play a major role in the drug absorption process, particularly when dealing with poorly soluble compounds. as the majority of compounds in current research pipelines of pharmaceutical industry are classified into the bcs class ii category [4], i.e. drug absorption is limited by solubility, it is very important to assess the development risk of such drug candidates at a very early stage of the process when choices among different chemical structures are still possible. in order to provide pbpk simulation software with better solubility input parameters, the use of solubility values in human fassif (fasted-state simulated intestinal fluid) has been proposed before. many publications refer to human pk prediction [5], but fassif has been proposed as input parameters for rats, too [6]. as rats are lacking a gall bladder, their gi tract is exposed to a constant bile flow at concentrations higher than in any other typical preclinical species. reported bile salt concentrations in rats are in the range of 12-51 mm [7-9]. human bio-relevant fluids are undoubtedly relevant dissolution media in order to assess the performance of drug candidates in a clinical research and development environment. however, the decision to move a potential drug candidate forward is taken during the pre-clinical phase. up to the point of compound selection for clinical development, these compounds have only been tested in typical preclinical species, like rats, dogs, monkeys or minipigs. the role of solubility in the gastrointestinal tract of a particular species becomes evident, as soon as a compound of low solubility has shown reasonable pk parameters in rodents and is selected for testing of pk properties in a non-rodent species. potentially, solubility may be one of the main factors that limit human bioavailability prediction based on animal studies [10], in case solubility in human gi fluids is significantly lower than in gi fluids of preclinical species. solubilization of compounds in the gi fluids is mainly driven by bile salt and phospholipid contents. to what extent differences in fluid composition influence solubility across species has been demonstrated in a comparison of solubility of drug molecules in human and dog gi fluids [11]. for this reason, the impact of solubility on pbpk simulation results has been widely studied, especially the use of high-throughput vs. thermodynamic equilibrium solubility data. gao et al. [6] have described a way to generate solubility input parameters for pbpk simulations of rats by applying a full gi dissolution profile that takes into consideration the constant dilution by fluid secretion along the gi transit, as well as formulation options admet & dmpk 4(1) (2016) 35-53 simulated rat intestinal fluid doi: 10.5599/admet.4.1.258 37 that could lead to differences in supersaturation and re-precipitation behavior. however, their model is based on human fassif to mimic the dilution in the rat gi fluid. due to the sensitivity of solubility to bile salts, which is expected for many compounds in drug discovery, solubility tested in the proposed simulated rat fluid should improve the prediction power of pbpk simulations when rodents are the species of interest. in this publication, we describe a way to generate a simulated rat intestinal fluid (rsif) and study its impact on pbpk simulation for rats. the results are compared to simulations which use other solubility input values, like solubility in buffer ph 6.8, fassif and the dissolution curve generated by an in vitro gi dissolution assay [6]. in addition, bile salt dependent solubility estimated based on log p is included in this comparison [12]. materials and methods physico-chemical measurements passive permeability of the compounds was investigated using the pampa assay [13]. in the current work, the log pampa value has been used, which represents the highest value of effective permeability from three measured values at ph 4, ph 6.8 and ph 8. equilibrium solubility was determined using the shake-flask method using about 1.5 mg of compound. the mixture was shaken for 20 to 24 hours and the solid phase was separated by centrifugation for 15 min at 2000g. concentration of the compound in the supernatant was determined by hplc-uv (agilent 1200) based on an external calibration curve. the partition coefficient log p was determined for ionizable compounds by potentiometric titration (sirius t3, sirius instruments ltd.). non-ionizable compounds were analyzed by reverse-phase hplc. determination of the acid dissociation constant (pka) was done by potentiometric titration (sirius t3, sirius instruments ltd.). preparation of test media buffer solution from merck kgaa; ph 6.8; 0.1 m di-sodium hydrogen phosphate / potassium hydrogen phosphate was used. preparation of human fassif-v2 was done based on the suggestion by jantratid et al. [14]. to obtain one liter of clear fassif-v2, sodium taurocholate and sodium chloride were first dissolved in 400 ml of purified water, followed by the addition of 1 ml of 1 m hcl. after stirring for 30 min, lecithin was added and the mixture was sonicated until complete dissolution and stirred for another two hours. then, maleic acid and 500 ml of water were added. after overnight stirring, the ph was adjusted to 6.5 by addition of 1m naoh and the total volume was adjusted to 1 l. simulated rat intestinal fluid a rat simulating intestinal fluid (rsif) was developed as surrogate to mimic intestinal conditions in vivo with regard to bile salt and phospholipid concentrations, surface tension, ph, buffer capacity and osmolarity [15]. this was done to supplement previously published human [14] and canine [11] media. due to unavailability of sufficient ex vivo samples, published information was used to define the target properties of rsif. table 1 compares the main properties of rsif with human fassif-v2. berghausen et al. admet & dmpk 4(1) (2016) 35-53 38 table 1. main properties of rat simulated intestinal fluid (rsif) in comparison to human fassif-v2 rsif fassif-v2 ph 6.0 6.5 buffer capacity [mmol/l/ph] 15.2 10 surface tension [mn/m] 42 54.3 osmolality [mosmol/kg] 480 180 bile salt concentration [mm] 25 3 phospholipid concentration [mm] 5.16 0.2 due to the constant bile flow in rats due to a lack of a gall bladder and most typical housing conditions with constant access to food, a homostatic condition was assumed for rats. in consequence, fasted and fed states were not distinguished for the rat medium. rat intestinal ph was averaged to 6.0 based on values reported elsewhere [2,16,17]. an average was selected to reflect the main areas of absorption within the proximal intestine with reported ph values between 5.0 and 7.1. to allow higher throughput, use of multiple ph media was neglected. adjustment of surface tension had to be balanced with bile salt and phospholipid concentrations. addition of cholic acid and chenodeoxycholic acid sodium was required to achieve targeted surface tension values. it was found that combination of bile salts provided a synergistic effect and provided best reduction of surface tension. both substances are reported to be present in rat bile fluids and thus have physiological relevance [18]. surface tension values obtained for rsif are in line with recent data [19]. bile salt and phospholipid concentrations are slightly above values referenced and used for example in gastroplus models and other references [17]. the values for buffer capacity were estimated taking into account human fassif-v2 and the constant bile flow, as there was no reference available at the time of rsif development. recent data from merchant [19] indicate that this value appears to be too low. depending on new insights rsif may be further refined with regard to bile salt concentration, buffer capacity and osmolality. further refinement of media could also reflect more section specific ph and bile concentrations in analogy to human media published [14]. table 2. composition of rsif; 100 ml target volume conc. (mm) molecular weight (g/mol) weighted sample (mg) sodium taurocholate 5.00 537.7 269 nacl 18.70 58.0 108 lecithin 5.16 775.0 400 maleic acid 29.86 116.1 347 purified water q.s. naoh 1 n q.s. sodium oleate 0.26 304.4 8 sodium cholate hydrate 12.50 430.6 538 sodium chenodeoxycholic acid 7.50 414.6 311 glyceryl monooleate 1.67 356.5 60 admet & dmpk 4(1) (2016) 35-53 simulated rat intestinal fluid doi: 10.5599/admet.4.1.258 39 preparation of rsif in order to prepare 100 ml of rsif, a composition according to table 2 was used. sodium taurocholate, sodium cholate hydrate, sodium chenodeoxycholic acid and nacl were dissolved in 50 ml of purified water. after stirring for 30 min, lecithin, sodium oleate and glyceryl monooleate were added. subsequently, the mixture was sonicated for about 5-10 min and stirred overnight. a clear solution was obtained. maleic acid and 40 ml of purified water were added, resulting in a suspension of ph around 2.2. the ph was adjusted to 6.0 by 1 m naoh. the mixture was stirred until a clear solution was obtained. if needed, the ph was adjusted, again. finally, the volume was filled up to 100 ml by addition of purified water. the surface tension was determined in triplicates using a kibron delta-8 tensiometer (kibron inc., helsinki). an average value of 42.8 mn/m was obtained. pbpk simulations pbpk simulations were performed using the acat model of the gastroplus software (simulations plus inc., lancaster) version 8.6. the distribution and elimination were described by a compartmental pk model. the number of compartments and parameters of the model were obtained by fitting the in vivo i.v. time-concentration profiles in the pkplus module of gastroplus. the absorption was predicted by the acat model in gastroplus using the “rat physiological fasted” gi tract and based on the physico-chemical and in vitro adme properties of the compounds. the permeability of the compound was characterized by measurement in pampa. first, a calibration of the pampa assay to the gastroplus permeability peff input parameter was carried out. it was based on the curve of pampa versus human fraction absorbed for 83 marketed compounds. the curve of the gastroplus intrinsic permeability peff versus human fraction absorbed was obtained by varying the peff value, see figure 1. the correlation between pampa and peff was obtained by matching the two curves. to be consistent with other data of internal databases an animal weight of 250 g was used. this is slightly below the actual weight of animals in the current study. the difference in simulation outcome has been tested for selected cases (data not shown) and is rather small. no effect on the ranking of the simulation results has been observed. the animals had free access to food, thus the feeding state was uncontrolled, neither fasted nor fed. due to the lack of the gall bladder, the difference between fasted and fed state for rats is regarded as to be small. as tested (data not shown), the impact of physiology of the simulation model is marginal in comparison to the solubility impact. thus, the fasted rat physiology was used. the absorption scaling factor model, designed to take into account the regional changes in permeability, was the “opt logd model sa/v 6.1”. the scaling coefficient c4 in the colon was set to zero in order to avoid overestimation of colonic absorption as outlined in equation (1): c d asf c 4log 3 10  (1) berghausen et al. admet & dmpk 4(1) (2016) 35-53 40 figure 1: peff conversion based on log pampa measurements as a consequence, the original model led to high absorption for all lipophilic drugs regardless of their pampa (or caco-2) permeability, which did not match our observations. the property space of the training set used in the gastroplus software where poorly permeable compounds are typically hydrophilic molecules does not match the property space of our discovery molecules that are characterized by a higher average molecular weight and for which poor permeability is not exclusively driven by low lipophilicity. the liver first-pass extraction was estimated with the well-stirred model based on in vitro microsomal clearance without taking into account plasma protein binding or microsomal binding. this approach was found to be reasonable by parrot et al [12], germani et al [21] and jones et al [22]. it was compared with the estimation based on the in vivo clearance and provided better predictions (see figure 2). admet & dmpk 4(1) (2016) 35-53 simulated rat intestinal fluid doi: 10.5599/admet.4.1.258 41 a) liver first pass based on in vivo clearance b) liver first pass based on microsomal clearance figure 2: comparison of the performance of the model between the estimation of the liver first pass with the in vivo clearance (a) and microsomal clearance (b) for 61 proprietary compounds administered to rats. the octanol-water partition coefficient log p is based on measured values. for all suspensions, a particle radius of 7 m was assumed, which is a typical number based on our experience with compounds in the lead optimization phase. typically, the unknown solid state and the impact of the amorphous form are supposed to play a major role for particle dissolution, which may overrule the particle size effect. information about dose dependence of systemic clearance is typically not available in the lead optimization stage and the assumption made in this study is that clearance does not significantly change with the dose. pharmacokinetic studies pharmacokinetic studies have been performed in male sprague-dawley rats, originating from charles river wiga (germany). the experiment was performed according to the regulations effective in the canton basel-city, switzerland. six to four days before first drug administration, the rats (body weight approx. 270-330 g) were anesthetized and catheters were surgically implanted into the femoral artery (for blood collection) and femoral vein (for intravenous injection). the catheters were exteriorized at the neck where they were fixed via the tether and a flexible spring to a harvard swivel system, which allowed blood sampling and intravenous injections without disturbing the freely-moving animal. for analgesic treatment, animals received temgesic (10 µg kg -1 s.c.) before surgery and subsequently twice at appropriate times after surgery. animals were kept individually in macrolon cages, with free access to food and water throughout the experiment. dosing i.v. and p.o. was performed in two different animal groups. for per oral administration, the compounds were formulated as a suspension in methyl cellulose. for intravenous administration, the compounds were dissolved in an appropriate mixture of n-methyl pyrrolidone (nmp) and poly(ethylene glycol) 200 (peg). for sample analysis, approximately 50 µl of blood was taken. 30 µl of blood was mixed with 200 µl of acetonitrile and centrifuged at 4 °c. approx. 200 µl of supernatant was transferred into a microtiter plate and mixed with 200 µl 0.1 % formic acid. an aliquot of each sample was injected into the berghausen et al. admet & dmpk 4(1) (2016) 35-53 42 lc-ms/ms system for analysis. gi dissolution assay full gi dissolution was performed by a modified setup based on the proposal of gao et al. [6]. their gi dissolution model takes into consideration the transit times in the stomach and four additional segments in the intestine. gao et al. included the dilution that is supposed to happen during the transit through the gastrointestinal tract. in particular, dilution was based on basal acid output, the pancreatic juice secretion rate and the bile flow. gao et al. proposed a basal acid output of 0.8657 ml/4 h/100 g. in order to calculate the dilution in the small intestine, a pancreatic juice secretion rate of 0.033 ml/min/kg is suggested; bile flow is set to 70 ml/day/kg in their setup. the corresponding transit times are 0.25 h in the stomach and 0.2, 2.0, 4.5 and 8 h for duodenum, jejunum/ileum, cecum and colon, respectively. dilution in the stomach is supposed to be caused by basal acid output only. in the small intestine, the basal acid output adds to the secretion of pancreatic juice and bile. based on these assumptions, dilution factors of 1.5 for the stomach, 1.9 for the duodenum, 5.8 for jejunum/ileum and 2.9 for the cecum were proposed. no further dilution was considered in the colon, since most fluid had been absorbed. for improved differentiation of weakly basic compounds in our setup, the ph of the stomach compartment was lowered to ph 3. as dilution medium the rsif as proposed in table 2 was used instead of fassif-v2. results and discussion in vitro solubility studies in order to study the difference of equilibrium solubility in the proposed rat fluid (rsif) in comparison to buffer solubility and human fassif-v2, a set of proprietary compounds has been investigated by the shakeflask method. the compounds have been selected based on availability of in vivo data, and sufficient compound amount for running solubility and gi dissolution studies. also, only compounds exhibiting high in vitro permeability have been chosen, thus keeping the impact of permeability on drug absorption low. solubility results in buffer ph 6.8, human fassif-v2 and rsif media for the selected compounds are summarized in table 3. table 3. solubility (g/l) of selected compounds in buffer ph 6.8, fassif-v2 in comparison to rsif buffer ph 6.8 fassif-v2 rsif compound a 0.001 0.0041 0.105 compound b 0.004 0.0001 0.448 compound c 0.031 0.043 0.711 compound d 0.005 0.003 0.023 compound e <0.0005 0.0009 0.046 compound f 0.05 0.14 0.881 compound g 0.003 0.013 0.193 compound h 0.0002 0.0005 0.01 compound i 0.0025 0.002 0.015 mostly compounds exhibiting low solubility in both buffer ph 6.8 and fassif-v2 have been selected. solubility in rsif can differ vastly from the other two media investigated, thus leading to the assumption admet & dmpk 4(1) (2016) 35-53 simulated rat intestinal fluid doi: 10.5599/admet.4.1.258 43 that fassif-v2 solubility is not able to provide a realistic input parameter for pbpk simulations in the rat. compound f has been selected as an example of acceptable solubility in fassif-v2, but rsif still showing a 7-fold higher equilibrium value. this result will be used as a test case whether solubility in rsif would lead to over-prediction in pbpk simulations. bile salt dependent solubility there is little doubt that the use of thermodynamic solubility values in buffers in many cases will not be very useful to predict in vivo concentration-time profiles. with the exception of well soluble compounds with good wetting properties, buffer solubility values will most likely underestimate the in vivo solubilization of most molecules. since the introduction of the human fassif and fessif media, there is an option to obtain solubility values which are more appropriate as input parameters for pbpk simulations and fassif solubility maybe one of the physico-chemical parameters that is determined as one of the first after a compound is sent out for characterization. however, to properly assess the risk for clinical development it is necessary to understand the impact of solubility on in vivo absorption during lead optimization and thus in preclinical species. although the fassif and fessif media allow studying the potential solubility limitation for absorption in humans, they may not be appropriate input parameters in preclinical species. the bile salt concentration in gastrointestinal fluids in the fasted dog is known to be at the level of human fessif rather than fassif [11], and the situation in rats is substantially different. due to the absence of a gall bladder and thus a constant bile flow, factors limiting absorption in a rat may be rather different as compared to the situation in humans. as bile salt concentrations in rat gi fluids are reported up to a level of 51 mm [9] the rat is likely to be a better solubilizer for lipophilic compounds which might be better absorbed than in higher species. the gastroplus software includes an estimation of bile salt dependent solubility based on log p according to the equation of mithani et al. [20]: bs aq sc sr sc  (2) where scbs is defined as the solubilization capacity of bile salt (moles drug/mole bile salt) and scaq is defined as the solubilization capacity of water (moles drug/mole water). for their set of compounds, they found: sr plog 2.23 0.61log  (3) this approach was evaluated based on a dataset of proprietary compounds. for these compounds, buffer solubility and the solubility in buffer containing various amount of sodium taurocholate was determined, see table 4. to calculate scbs, the molar solubility at a defined bile salt concentration was divided by the actual bile salt concentration, i.e. 10, 50 or 100 mm in our data set. scaq, correspondingly, is obtained by dividing the molar solubility in buffer ph 6.8 by 55.5 mol/liter. calculation of the log p dependence of the solubility ratio (sr) as defined by mithani et al. led to a linear regression similarly to the equation as proposed by mithani, see figure 3. while this data principally berghausen et al. admet & dmpk 4(1) (2016) 35-53 44 supports the linear dependency of sr on log p, this equation is based on a fit through several orders of magnitude for both log p and log sr, which will compromise the ability to predict individual solubility values with the accuracy required for pbpk modeling. in order to estimate the predictive power of the mithani equation, the calculated values and the measured solubility at different bile salt concentrations are displayed in figure 3. for some of the compounds of the test set (see table 4) the variation in sr can be more than one order of magnitude within the different concentration of bile salts used to determine the solubility, i.e. 10, 50 or 100 mm. the maximum difference to the linear fit is found to be about 2 log units of sr. importantly, the differences to the linear fit do not appear to be markedly log p dependent. also, the increase in solubility with increasing bile salt concentration for many compounds is not a linear dependency, see table 4. table 4. solubility test set for taurocholate concentrations of 10, 50 and 100 mm. s indicates solubility in buffer at ph 6.8. calculated solubility ratio according to taurocholate concentrations (log sr10, log sr50, log sr100), based on the mithani equation (1). mw log p pka acid/base s / mm 10 mm 50 mm 100 mm log sr10 log sr50 log sr100 cpd 1 529.5 5.2 5.4b; 3.9b 0.0019 17.78 483.2 666.9 7.718 8.453 8.292 cpd 2 379.4 3.2 3.4b 0.022 0.04 0.26 0.34 4.004 4.118 3.933 cpd 3 560.5 4.1 8.2b; 3.7b 0.0054 0.42 3.95 5.49 5.639 5.913 5.755 cpd 4 558.2 4.6 9.1b; 4.1b 0.001 0.7 4.12 4.07 6.589 6.660 6.354 cpd 5 577.8 4.9 9.1b; 3.9b 0.3202 10.8 11.02 11.23 5.272 4.582 4.289 cpd 6 410.4 2.6 5.1b 0.0975 2.27 4.35 9.86 5.111 4.695 4.749 cpd 7 485.5 4.3 4.2b 0.001 <0.001 0.15 0.34 4.907 4.962 cpd 8 453.6 3.1 3.7b 0.0066 0.01 0.01 0.01 3.924 3.225 2.924 cpd 9 710.7 7.1 5.11a; 3.8b 0.0014 0.03 1.27 1.9 5.073 6.001 5.875 cpd 10 313.4 3.7 n.d. 0.16 0.14 1.6 3.686 4.045 cpd 11 361.3 3.9 11.7a 0.14 0.46 0.51 0.63 4.261 3.607 3.398 y = 0.672x + 2.1634 2 3 4 5 6 7 8 9 2 3 4 5 6 7 lo g s r logp logsr10 logsr50 logsr100 mithani prediction linear (logsr100) figure 3. log p dependence of solubility ratio at 10 mm (log sr10), 50 mm (log sr50) and 100 mm (logsr 100) taurocholate concentration. the linear fit is based on log sr100, the crosses represent the prediction according to equation (2). as a consequence, due to the significant deviations from the linear fit for many compounds, the estimation of bile salt solubility for individual compounds is regarded as not sufficient to support studying admet & dmpk 4(1) (2016) 35-53 simulated rat intestinal fluid doi: 10.5599/admet.4.1.258 45 the impact of solubility in pbpk simulations. the estimation of log p dependent solubilization ratio might be justified in cases where no bile salt dependent solubility can be determined, however, there remains a risk of being misled. this is especially true for low soluble compounds, as changes of solubility by only a factor 2 to 4 might strongly impact the absorption profile of a compound. thus, fassif solubility will not provide accurate enough information to further extrapolate the contribution of bile salt solubilization at concentrations that are assumed to mimic the in vivo situation in rats. pbpk simulations the importance of solubility input parameters for pbpk simulations of low soluble compounds is indisputable. at low dose, a significant portion of a compound administered might be dissolved in the stomach and subsequently be absorbed. however, at a higher dose, there is a larger portion remaining undissolved. thus, the solubilization power of the intestinal fluid is more relevant for the absorption of higher doses. as a consequence, gi dissolution profiles adapted to the relevant species should positively impact the prediction results, especially for cmax. to test this hypothesis, gastroplus simulations have been performed at different dose levels and simulation results are compared to in vivo results. compounds with no expected permeability limitation according to in vitro data (pampa) and low to moderate total blood clearance have been selected to investigate the impact of solubility parameters on pbpk simulations, see table 5. a systematic approach was chosen to compare the impact of solubility on the simulation results for the selected compounds. five different solubility input parameters have been chosen. first, the thermodynamic equilibrium solubility in buffer was used, without correction for bile salt solubilization. in a next step, the fassif solubilty value was used to calculate the solubilization ratio (sr). based on the data in table 4, the theoretical sr calculation as provided by the gastroplus software was tested in another approach. the rat sif fluid solubility was used without any correction for bile salt content, similar to the thermodynamic solubility value. according to the authors’ experience, correction of bile salt dependent solubility using the built-in in vitro sr calculation compromises the value of the solubility in the proposed rat sif medium. finally, if not limited by compound availability, the full dissolution curve was recorded to mimic the dissolution properties of the compound during its transit through the rat gastrointestinal tract, based on the modified procedure of gao et al.[6]. table 5. summary of input parameters for pbpk simulations. first pass extraction ratio (er%) based on rat liver microsomes. clearance as a result of fit to in vivo data by the pkplus module. the volume of distribution into the respective distribution model is given by vc, v2 and v3. in case only two numbers are given, they refer to vc and v2, respectively. first pass er% pka log pampa / cm/s log p clearance, l/h/kg vc, v2, v3 compound a 77 3.8 base -3.9 1.7 0.42 0.38, 0.15, 1.67 compound b 28 5.4 acid -4.0 3.7 1.10 3.28, 4.58 compound c 92 <3 base -3.6 3.6 3.48 2.77, 0.95, 1.74 compound d 25 4.7 base -3.6 3.5 0.98 4.32, 3.50, 17.62 compound e 56 2.7 base -3.5 4.2 0.42 0.88, 2.19 compound f 67 3.2 base -3.4 3.1 0.42 0.46, 0.47 compound g 55 3.4 base -4.6 3.9 3.36 4.50, 0.80 compound h 38 3.5 base -4.5 5.3 1.34 3.74, 7.19 compound i 37 1.7 base -4.2 2.8 5.19 2.57, 0.55 berghausen et al. admet & dmpk 4(1) (2016) 35-53 46 the pbpk simulation results based on the acat model have been split into lower and higher dose studies and are summarized in tables 6 to 9. simulation results are compared to in vivo data from compounds dosed orally as suspension. pbpk results lower range of p.o. doses the impact of various solubility input parameters on pbpk simulation results has been tested. as evaluation criteria for the success of pbpk simulations versus the in vivo pk results, cmax and auc were chosen, see table 6 and table 7. different solubility input parameters have been tested, as described above (see table 3). in addition, the full gi dissolution profile at the respective dose has been measured and the solubility values for each compartment have been entered according to the method section. table 6. pbpk results of cmax in g/ml for lower dose range based on various solubility input. cmax buffer, fassif, fassif with theoretical bile salt correction (cmax fassif bile) and rsif refer to equilibrium values, whereas gi disso comprises the input of the full gi dissolution profile. pred. vs. obs. denotes the ratio of the predicted values vs. the observed cmax. dose mg/kg cmax obs cmax buffer cmax fassif cmax fassif bile cmax rsif cmax gi disso compound a 10 5.041 0.035 0.273 0.089 2.185 pred. vs. obs. 0.0069 0.0542 0.0175 0.4333 compound b 3 0.138 0.012 0.012 0.040 0.238 pred. vs. obs. 0.0869 0.0869 0.2896 1.7246 compound c 3 0.134 0.028 0.043 0.046 0.045 pred. vs. obs. 0.2089 0.3208 0.3432 0.3358 compound d 3 0.213 0.053 0.057 0.124 0.167 0.204 pred. vs. obs 0.2488 0.2676 0.5822 0.7840 0.9577 compound e 10 2.211 0.069 0.071 0.609 1.502 1.517 pred. vs. obs 0.0312 0.0321 0.2754 0.6793 0.6861 compound f 10 3.753 2.743 2.743 2.742 2.743 2.308 pred. vs. obs 0.7309 0.7309 0.7306 0.7309 0.6150 compound g 3 0.020 0.004 0.030 0.022 0.074 0.073 pred. vs. obs 0.2000 1.500 1.100 3.700 3.650 compound h 3 0.082 0.0007 0.0007 0.031 0.037 0.071 pred. vs. obs 0.0085 0.0085 0.3780 0.4512 0.8659 compound i 3 0.056 0.011 0.008 0.020 0.062 pred. vs. obs 0.1964 0.1429 0.3571 1.107 results for the lower dose range (see table 6, table 7 and figure 4) support the assumption that the buffer solubility typically leads to an underprediction of both cmax and auc. the only exception is compound f, despite the fact that its solubility in rsif is much higher than in buffer and fassif-v2 media. apparently, its rather high permeability and sufficient solubility lead to nearly complete absorption, yet if the lowest measured solubility value is used. even a ten times higher solubility value, according to rsif in vitro results, does not change the predicted concentration-time profile for this compound for the 10 mg/kg dose. admet & dmpk 4(1) (2016) 35-53 simulated rat intestinal fluid doi: 10.5599/admet.4.1.258 47 table 7. pbpk results of auc (0-t) in g h/ml for lower dose range based on various solubility input. auc buffer, fassif, fassif with theoretical bile salt correction (auc fassif bile) and rsif refer to equilibrium values, whereas gi disso comprises the input of the full gi dissolution profile. pred. vs. obs. denotes the ratio of the predicted values vs. the observed auc. dose, mg/kg auc obs auc buffer auc fassif auc fassif bile auc rsif auc gi disso compound a 10 16.51 0.186 1.08 0.407 5.05 pred. vs. obs 0.0113 0.0654 0.0247 0.3059 compound b 3 1.928 0.146 0.146 0.414 1.867 pred. vs. obs 0.0757 0.0757 0.2147 0.9684 compound c 3 0.232 0.064 0.069 0.070 0.071 pred. vs. obs 0.2759 0.2974 0.3017 0.3060 compound d 3 2.631 0.570 0.607 1.154 1.570 1.802 pred. vs. obs 0.2166 0.2307 0.4386 0.5967 0.6849 compound e 10 26.4 0.758 0.795 4.803 9.351 9.335 pred. vs. obs 0.0287 0.0301 0.1819 0.3542 0.3536 compound f 10 13.4 7.812 7.812 7.823 7.812 7.734 pred. vs. obs 0.5830 0.5830 0.5838 0.5830 0.5772 compound g 3 0.131 0.020 0.116 0.090 0.285 0.260 pred. vs. obs 0.1527 0.8855 0.6870 2.176 1.985 compound h 3 1.06 0.007 0.007 0.228 0.312 0.546 pred. vs. obs 0.0066 0.0066 0.2151 0.2943 0.5151 compound i 3 0.571 0.038 0.033 0.062 0.19 pred. vs. obs 0.0665 0.0578 0.1086 0.3327 when the in vitro equilibrium solubility in fassif is used to correct the bile salt dependent solubility, an improved prediction is found in a number of cases. in order to allow for an estimate of bile salt dependent solubility in case fassif solubility data is not available, the use of the theoretical bile solubilization ratio may be an option. as shown in table 4 the results of this method might significantly deviate from the measured values. as the results in table 6 and table 7 suggest, for most of the investigated compounds the use of the theoretical bile salt ratio leads to an improved prediction, but underprediction for low solubility compounds can still be significant. when comparing the simulation results that were generated using the in vitro rat sif solubility, for nearly all compounds of the current test set an improved prediction result is achieved. this is reflected in a maximum deviation of the predicted value vs the observed value by a factor of at least 3. for some compounds even the range of 2-fold variation is reached. overprediction was observed only in one case (compound g), for which even fassif-v2 solubility resulted in a too high value of cmax. in case of compound d and h, the full gi dissolution further improves the simulation outcome. however, taking into account the time and effort to record the dissolution profile, it appears justified to use rsif equilibrium solubility instead. berghausen et al. admet & dmpk 4(1) (2016) 35-53 48 -2.5 -2 -1.5 -1 -0.5 0 0.5 1 lo g ( p re d ic te d v s . o b s e rv e d ) cmax buffer cmax fassif cmax fassif bile cmax rat fluid cmax gi disso cpd a cpd b cpd c cpd d cpd e cpd f cpd g cpd h cpd i -2.5 -2 -1.5 -1 -0.5 0 0.5 lo g ( p re d ic te d v s . o b s e rv e d ) auc buffer auc fassif auc fassif bile auc rat fluid auc gi disso cpd a cpd b cpd c cpd d cpd e cpd f cpd g cpd h cpd i figure 4. a) cmax and b) auc from simulations performed on studies comprising doses from a 3 to 10 mg/kg dose observations at higher p.o. doses compared to low p.o. doses, the solubility and dissolution properties of a compound are assumed to impact the overall pk profile in an even stronger way when higher doses area applied. pbpk results for doses from 30 to 300 mg/kg are summarized in table 8 and table 9. figure 5 displays the results as ratio of predicted vs. observed for both cmax and auc. based on rsif solubility, predicted cmax a) b) admet & dmpk 4(1) (2016) 35-53 simulated rat intestinal fluid doi: 10.5599/admet.4.1.258 49 of all compounds tested is found within a 3-fold range as compared to observed values, most of them even within 2-fold. the difference to all other equilibrium solubility values investigated is even more pronounced and full gi dissolution does not necessarily help improving the quality of pbpk prediction. the example of compound f reveals, again, that there is no influence of solubility for this compound, as already pointed out for the lower dose studies. table 8. pbpk results of cmax in g/ml for higher dose range based on various solubility input. cmax buffer, fassif, fassif with theoretical bile salt correction (cmax fassif bile) and rsif refer to equilibrium values, whereas gi disso comprises the input of the full gi dissolution profile. pred. vs. obs. denotes the ratio of the predicted values vs. the observed cmax. dose, mg/kg cmax obs cmax buffer cmax fassif cmax fassif bile cmax rat fluid cmax gi disso compound a 30 8.581 0.047 0.401 0.121 4.139 pred. vs. obs 0.0055 0.0467 0.0141 0.4823 compound a 100 18.3 0.042 0.468 0.127 4.02 pred. vs. obs 0.0023 0.0256 0.0069 0.2197 compound b 30 0.784 0.022 0.023 0.119 1.531 pred. vs. obs 0.0280 0.0293 0.1518 1.953 compound c 30 0.546 0.041 0.101 0.236 0.447 0.491 pred. vs. obs 0.0751 0.1850 0.4322 0.8187 0.8993 compound d 30 0.908 0.095 0.058 0.271 0.380 2.04 pred. vs. obs 0.1046 0.0639 0.2985 0.4185 2.2467 compound e 100 3.592 0.094 0.220 1.227 3.869 4.74 pred. vs. obs 0.0262 0.0612 0.3416 1.077 1.320 compound f 100 17.0 26.3 27.3 25.9 27.4 27.4 pred. vs. obs 1.547 1.610 1.524 1.612 1.612 compound g 30 0.813 0.005 0.055 0.038 0.274 0.263 pred. vs. obs 0.0062 0.0677 0.0467 0.3370 0.3235 prediction of auc in the dose range 30 to 300 mg/kg follows the same trend as for cmax. rsif solubility, besides the gi dissolution assay, provides best matches of pbpk results without a tendency to overpredict. for instance, for compound b at a dose of 3 mg/kg, without using the rsif solubility but the calculated solubilization ratio an under-prediction of auc by a factor of 5 was found. at the dose of 30 mg/kg, the under-prediction was nearly 10-fold, whereas the simulation based on rat sif fluid led to an excellent match for auc with a factor of 1.3-fold. there are still results where auc is underpredicted by a factor of more than five. one may hypothesize that for these cases, the impact of other parameters than solubility, like first pass extraction, might be more important in limiting exposure. our results obtained with nine compounds and across different dose ranges show that the rsif solubility can be regarded as a very useful biorelevant solubility parameter when performing pbpk simulations for rats. berghausen et al. admet & dmpk 4(1) (2016) 35-53 50 table 9. pbpk results of auc (0-t) in g h/ml for higher dose range based on various solubility input. auc buffer, fassif, fassif with theoretical bile salt correction (auc fassif bile) and rsif refer to equilibrium values, whereas gi disso comprises the input of the full gi dissolution profile. pred. vs. obs. denotes the ratio of the predicted values vs. the observed auc. dose, mg/kg auc obs auc buffer auc fassif auc fassif bile auc rat fluid auc gi disso compound a 30 41.3 0.514 2.057 0.861 15.2 pred. vs. obs 0.0124 0.0498 0.0209 0.3680 compound a 100 97.5 0.297 2.48 0.766 19.2 pred. vs. obs 0.0030 0.0254 0.0079 0.1969 compound b 30 11.4 0.274 0.274 1.274 14.75 pred. vs. obs 0.0240 0.0240 0.1118 1.2939 compound c 30 6.99 0.175 0.317 0.521 0.706 0.874 pred. vs. obs 0.0250 0.0454 0.0745 0.1010 0.1250 compound d 30 14.83 1.139 0.707 2.900 4.27 18.02 pred. vs. obs 0.0768 0.0477 0.1955 0.2879 1.215 compound e 100 56.9 1.14 4.80 15.1 35.8 44.0 pred. vs. obs 0.0200 0.0844 0.2654 0.6292 0.7733 compound f 100 268 78.0 78.1 78.2 78.1 78.1 pred. vs. obs 0.2910 0.2914 0.2918 0.2914 0.2914 compound g 30 6.25 0.033 0.279 0.202 1.249 1.225 pred. vs. obs 0.0052 0.0446 0.0323 0.1998 0.1960 admet & dmpk 4(1) (2016) 35-53 simulated rat intestinal fluid doi: 10.5599/admet.4.1.258 51 -2.5 -2 -1.5 -1 -0.5 0 0.5 lo g ( p re d ic te d v s . o b s e rv e d ) cmax buffer cmax fassif cmax fassif bile cmax rat fluid cmax gi disso cpd a cpd b cpd c cpd d cpd e cpd f cpd g -2.5 -2 -1.5 -1 -0.5 0 0.5 lo g ( p re d ic te d v s . o b s e rv e d ) auc buffer auc fassif auc fassif bile auc rat fluid auc gi disso cpd a cpd b cpd c cpd d cpd e cpd f cpd g figure 5. observed a) cmax and b) auc results from simulations performed on studies comprising doses from a 30 to 300 mg/kg dose conclusions biorelevant solubility parameters can be markedly different across species. for proper prediction of solubility mediated exposure limitation in human, one needs to understand the solubility limitation across in vivo species. a good exposure in rats will not necessarily translate into acceptable human drug absorption, if solubility is mainly driven by bile salt solubilization, or may lead to an undesired absorption profile, like a strong food effect. even good exposure in a non-rodent species, like dog, might not be indicative of the situation in humans, as the dog intestinal fluid still contains bile salts at a level that is close to the post-prandial situation in humans. a) b) berghausen et al. admet & dmpk 4(1) (2016) 35-53 52 a better understanding of the impact of solubility when drug absorption is studied in various species helps to lower the risk of failure in clinical studies. for this, a rat simulated intestinal fluid (rsif) is proposed to support understanding the role that solubility plays for absorption of a particular drug candidate. the proposed rat sif greatly improves the quality of pbpk predictions in rats. it explains why some compounds that are low soluble in aqueous media nevertheless show good p.o. exposure in rats and why this might not always translate to higher species. the results presented in this study allow for a better prediction of exposure in rats across the dose range of interest. in addition, a better understanding of solubility mediated exposure in rats will help to better anticipate the risk of solubility limited exposure in higher species and ultimately in humans. acknowledgements the authors gratefully thank daniel gosling and philippe monbaron for performing the solubility and dissolution experiments. further we greatly acknowledge the contribution of angus hibbins to the development of rsif. references [1] e.s. kostewicz, l. aarons, m. bergstrand, m.b. bolger, a. galetin, o. hatley, m. jamei, r. lloyd, x. pepin, a. rostami-hodjegan, e. sjögren, c. tannergren, d.b. turner, c. wagner, w. weitschies, j. dressman, european journal of pharmaceutical sciences 57 (2014) 300-321. [2] e.l. mcconnel, a.w basit, s. murdan, the journal of pharmacy and pharmacology 60(1) (2008) 6370. [3] r. holm, a. mullertz, h. mu, international journal of pharmaceutics 453(1) (2013) 44-55. [4] g. amidon, h. lennernäs, v. shah, j. crison, pharmaceutical research 12(3) (1995) 413-420. [5] h. jones, n. parrott, g. ohlenbusch, t. lavé, clinical pharmacokinetics 45(12) (2006) 1213-1226. [6] y. gao, r.a. carr, j.k. spence, w.w. wang, t.m. turner, j.m. lipari, j.m. miller, molecular pharmaceutics 7(5) (2010) 1516-1526. [7] m. hagio, m. matsumoto, m. fukushima, h. hara, s. ishizuka, journal of lipid research 50(1) (2009) 173-180. [8] j. m. dietschy, journal of lipid research 9(3) (1968) 297-309. [9] y. tanaka, t. hara, w. ryoichi, n. shunji, journal of pharmacy & pharmaceutical sciences 15(4) (2012), 510-518. [10] h. musther, a. olivares-morales, o.j. hatley, b. liu, a. rostami hodjegan, european journal of pharmaceutical sciences 57 (2014) 280-291. [11] m. arndt, h. chokshi, k. tang, n.j. parrott, c. reppas, j.b. dressman, european journal of pharmaceutics and biopharmaceutics 84(3) (2013) 633-641. [12] n. parrott, n. paquereau, p. coassolo, t. lave, journal of pharmaceutical sciences 94(10) (2005) 2327-2343. [13] f. wohnsland, b. faller, journal of medicinal chemistry 44(6) (2001) 923-930. [14] e. jantratid, n. janssen, c. reppas, j.b. dressman, pharmaceutical research 25(7) (2008) 1663-1676. [15] k. magwedere, s. mukaratirwa, international journal of applied research in veterinary medicine 6(3) (2008) 166-174. [16] j.m. desesso, a.l. williams, annual reports in medicinal chemistry 43 (2008) 353-371. [17] t.t. kararli, biopharmaceutics & drug disposition 16 (1995) 351-380. [18] d.m. heuman, z.r. vlahcevic, m.l. bailey, p.b. hylemon, hepatology 8(4) (1988) 892-897. admet & dmpk 4(1) (2016) 35-53 simulated rat intestinal fluid doi: 10.5599/admet.4.1.258 53 [19] h.a. merchant, s.c. rabbie, f.j.o. varum, f. afonso-pereira, a.w. basit, european journal of pharmaceutical sciences 62 (2014) 76-85. [20] s.d mithani, v. bakatselou, c.n. tenhoor, j. b. dressman, pharmaceutical research 13(1) (1996) 163167. [21] m. germani, p. crivori, m. rocchetti, p.s. burton, a.g. wilson, m.e. smith, i. pogessi, european journal of pharmaceutical sciences 31 (2007) 190-201. [22] h. m. jones, i. b. gardner, k. j. watson, the aaps journal 11(1) (2009) 155-166. ©2016 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ jese manuscript doi: 10.5599/admet.1.1.2 3 admet & dmpk 1(1) (2013) 3-5; doi: 10.5599/admet.1.1.2 open access : issn: 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index book review finding a needle in a haystack zoran mandić faculty of chemical engineering and technology, university of zagreb, marulićev trg 19, hr-10000 zagreb, croatia e-mail: zmandic@fkit.hr; tel.: +385-1-4597-164; fax: +385-1-3733-640 received: december 20, 2013; revised: january 10, 2013; published: january 14, 2013 absorption and drug development. solubility, permeability and charge state. 2nd edition. alex avdeef, john wiley & sons inc. 2012. in my early days in pharmaceutical research i was summoned into the boss's office where a meeting was taking place. i came right into a discussion about the determination of the partition coefficients of two substances of potential significance to the company. “this is a notorious time wasting”, i muttered to myself while trying to avoid being drawn into the discussion. what would have been more important was the synthesis of new chemical entities, which certainly had very good potential to become new blockbusters on the drugs market, especially those originating from my flask. little did i know at that time that the issue of accurate prediction of drug absorption in the early stages of drug discovery would become so important. it was only some time later, when a colleague threw a paper in front of me claiming excitedly that it was something we could and should do, that i gave the subject a second thought. it was one of a series of alex avdeef's papers about physico-chemical profiling of drug substances including the potentiometric determination of octanol–water partition coefficients. it took no time for the colleague of mine to assemble a potentiometric titration device from scraps and pieces of old, discarded equipment, and off we went... now, almost two decades later, while flicking through the pages of the second edition of alex's book absorption and drug development, i find myself astonished by how much the concept of drug absorption and its importance has progressed into a scientific discipline without which no contemporary pharmaceutical research could be considered. the author describes in a very simple way four key physicochemical processes, i.e. ionization, solubility, lipophilicity and permeability, which are of relevance to the absorption of most orally administered drugs, and explains various relationships among them. by the application of simple scientific concepts of physico-chemical profiling and continuously placing them in the drug development setting, readers can easily gain deep understanding of their significance. physicochemical parameters such as pka, log p or log s are no longer perceived by medicinal chemists or formulation and pre-formulation engineers as “good” or “bad” numbers, but provide valuable information revealing and predicting the whole pathway and the fate of the drug from its administration to entering the bloodstream. http://www.pub.iapchem.org/ojs/index.php/admet/index� mailto:zmandic@fkit.hr� zoran mandić admet & dmpk 1(1) (2013) 3-5 4 this book is based on the author’s many years of experience and on the experience of many other scientists who have tackled the subject of drug absorption and given it the significance it deserves. if the author’s goal was to “examine the components of multimechanistic processes related to charge state…. with the aim of advancing improved strategies for in vitro assays related to drug absorption”, then there is no shadow of doubt that he achieved it. however, the core strength of absorption and drug development is that it provides solid ground on which different worlds, embodied in the labs of medicinal chemists, pharmaceutical scientists, analytical chemists and even pharma managers, can understand each other and together rationalize the results obtained and their consequences. the book is conceptually organized into ten chapters. chapter 1 is an introductory chapter that places the topic appropriately in the pharmaceutical research context. chapter 2 gives a physico-chemical model of drug permeation based on the ph–partition hypothesis and fick’s laws of diffusion. this is followed by a description of biological environments relevant to drug absorption. molecular ionization and the acid-base properties of drugs are dealt with in chapter 3. this chapter is the most comprehensive coverage of the theory and practical aspects of pka determination by potentiometric techniques i have ever encountered in textbooks of physical chemistry. from the beginning of the chapter readers learn the importance of pka as a molecular property defining the charge state of drug in various parts of the body. the thermodynamics of ionization reactions are explained followed by a description of the practical aspects of pka measurements that should be taken into account while performing pka determinations with a glass electrode. data handling and data presentation are introduced with the emphasis on avoiding errors and false results. particular attention is paid to coping with challenges and obstacles that might be encountered in pka determination of low solubility substances of pharmaceutical interest. all topics are accompanied with many practical examples and tutorials. chapters 4 and 5 describe lipophilicity as a molecular property of considerable importance in drug absorption. octanol–water partitioning (chapter 4) and “biomimetic” liposome–water partitioning (chapter 5) and their determinations are presented. special emphasis is placed on ion-pair partitioning and its significance for drug absorption. an attempt is made to define new concepts and explain some of the anomalies in lipophilicity results. alex avdeef giving a lecture at the 1st world conference on physico-chemical methods in drug discovery and development, organized by iapc in rovinj, croatia, 2009. admet & dmpk 1(1) (2013) 3-5 book review doi: 10.5599/admet.1.1.2 5 an in-depth discussion of solubility and its measurements is presented in chapter 6. important concepts are illustrated with several case studies. ph-dependent solubility profiles are evaluated and highthroughput solubility measurements are tackled. chapters 7–9 deal with the concepts of permeability. the practical and theoretical background of prediction of human jejunal permeability and human intestinal absorption by in-vitro assays, most notably by pampa, is illustrated in a broad sense (chapter 7). cell-based assays, caco-2 and mdck are explored in chapter 8. the idea is to develop protocols and advanced data analysis for inter-laboratory cell-based transport studies. finally, fundamental concepts of permeability measurements using endothelial cell models of the blood–brain barrier are covered in chapter 9. throughout the book the rules and laws of physical chemistry are followed. they are explained in a simple way but with sufficient detail for readers to gain a complete insight into the topics. furthermore, this book gathers together a huge collection of relevant physico-chemical parameters determined by the author and his co-workers or found in other literature. more than 900 pka values, 350 octanol/water partition coefficients, 100 liposome/water partition coefficients, 60 solubility constants, 300 permeability coefficients, and almost 700 caco-2/mdck permeability coefficients as well as 200 permeability values related to the blood–brain barrier can be found in addition to many other parameters of pharmaceutical interest. this book is essential literature on the desk of the pharmaceutical scientist and is highly recommended to students and others interested in the topic of drug absorption. ©2013 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/� manuscript doi: 10.5599/admet.4.2.291 98 admet & dmpk 4(2) (2016) 98-113; doi: 10.5599/admet.4.2.291 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper defining desirable natural product derived anticancer drug space: optimization of molecular physicochemical properties and admet attributes deepika singh medicinal chemistry division, central institute of medicinal and aromatic plants, po cimap, lucknow 226015, india corresponding author: e-mail: deepika.sh25@yahoo.com. received: march 25, 2016; revised: may 14, 2016; published: june 29, 2016 abstract as part of our endeavor to enhance survival of natural product derived drug candidates and to guide the medicinal chemist to design higher probability space for success in the anti cancer drug development area, we embarked on a detailed study of the property space for a collection of natural product derived anti cancer molecules. we carried out a comprehensive analysis of properties for 24 natural products derived anti cancer drugs including clinical development candidates and a set of 27 natural products derived anti cancer lead compounds. in particular, we focused on understanding the interplay among eight physicochemical properties including like partition coefficient (log p), distribution coefficient at ph=7.4 (log d), topological polar surface area (tpsa), molecular weight (mw), aqueous solubility (log s), number of hydrogen bond acceptors (hba), number of hydrogen bond donors (hbd) and number of rotatable bonds (nrot) crucial for drug design and relationships between physicochemical properties, adme (absorption, distribution, metabolism, and elimination) attributes, and in silico toxicity profile for these two sets of compounds. this analysis provides guidance for the chemist to modify the existing natural product scaffold or designing of new anti cancer molecules in a property space with increased probability of success and may lead to the identification of druglike candidates with favorable safety profiles that can successfully test hypotheses in the clinic. keywords anticancer, admet, natural products, physicochemical property, pharmacokinetic-pharmacodynamic introduction cancer is one of the major disease causes of mortality worldwide and the numbers of cancer cases are increasing gradually [1]. cancer is a main public health burden in both developed and developing countries and affects the lives of millions of people. cancer is an abnormal growth of cells in the body, which underlies a collection of multiple genetic abnormalities through a multistep, mutagenic process. cancer cells usually invade and destroy normal cells in the body. factors responsible for cancers includes genetic predisposition, smoking, incorrect diet, infectious diseases and environmental factors. american cancer society has predicted ~27 million newly diagnosed individuals and ~17 million cancer related deaths globally by 2050 [2]. the key problem to cancer treatment is the recurrence of tumor and the side effects of chemotherapy drugs. hence, there is a potential demand to develop new and efficient anti-cancer drugs http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:deepika.sh25@yahoo.com admet & dmpk 4(2) (2016) 98-113 defining natural product derived anticancer drug space doi: 10.5599/admet.4.2.291 99 [3]. natural products have received increasing attention in the past 30 years for the discovery of novel cancer preventive and therapeutic agents [4]. natural products have been used for centuries for the treatment of several ailments. there are many basic ancient medicinal systems derived from dietary sources. nature has provided a plenty of natural products with potential anti-cancer activity in the last few decades. since 1940, approximately 175 small molecules have been approved as anti-cancer agents, of these, 48.6 % were a natural product or derivative [5]. currently, pharmaceutical industry faces large attrition rates of preclinical and clinical candidates due to toxicity or lag of optimal pharmacokinetics properties, resulting in high costs and increased timelines for the drug discovery process [6]. lead structures are compounds that typically exhibit suboptimal target binding affinity. pharmacological studies have shown that there is a difference exists between leads and drugs [7]. the present study is an approach to establish the difference between some selected potent anticancer natural compounds (leads) and fda approved natural product derived anti cancer drugs considering the distribution of physicochemical, adme (absorption, distribution, metabolism, and elimination) attributes and in silico toxicity endpoints. this data was examined with the goal of identifying trends and defining a set of property values that would best define the anticancer drug space associated with a higher probability of clinical success. several critical physicochemical properties of compounds like log p, log d, tpsa, mw, log s, hba, hbd and nrot proposed by various research groups should be considered for compounds with oral drug delivery as a concern [8]. the information obtained from this analysis could, in turn, be utilized to design anticancer drug molecules with optimum bioavailability and less or no toxicity based on the alignment of a set of key properties. the multi-parameter optimization (mpo) approach is very popular for providing guidance on how to design preferred molecules to reduce the attrition rate and increase the probability of prospectively designing molecules that survive preclinical safety studies and that possess optimal pharmacokinetic and pharmacodynamic properties to test hypotheses in the clinic [9]. tremendous progress has been made in recent years in terms of enabling the development of robust pharmacokineticpharmacodynamic (pk/pd) relationships for anticancer agents as well as in understanding how these relationships are influenced by molecular physicochemical properties. key physicochemical properties related to a drug like molecules have been described previously by various research groups [10]. the most important and well-known rule of five (ro5) was given by lipinski et al. in 1997 based on the database of clinical candidates that had reached phase ii trials or further [11]. ro5 provided the end points for four crucial physicochemical properties that described 90 % of orally active drugs: (a) molecular weight, mw < 500 da; (b) calculate of 1-octanol/water partition coefficient, clogp < 5; (c) number of hydrogenbond donors, (oh plus nh count) < 5; and (d) number of hydrogenbond acceptors, (o plus n atoms) < 10. these four physicochemical properties and their endpoints are associated with acceptable aqueous solubility and intestinal permeability, the important first step of oral bioavailability [11]. after the lipinski’s ro5, various other ways of predicting the drug-like space for rational design purposes have been introduced by other people. veber et al. 2002, showed that molecular weight cutoff at 500 da does not itself significantly separate compounds with poor oral bioavailability from those with acceptable values based on the oral bioavailability measurements in rats for glaxosmithkline database of almost 1100 drug candidates [12]. he suggested that compounds having two criteria of (1) number of rotatable bonds (nrot) ≤ 10, (2) tpsa ≤ 140 å 2 will have a high probability of good oral bioavailability. another effective range of physicochemical properties provided by ghose et al. 1999, based on comprehensive medicinal chemistry (cmc) database can be used in the design of drug-like deepika singh admet & dmpk 4(2) (2016) 98-113 100 combinatorial libraries [13]. to go beyond the properties associated with the ro5 and other drug-like filters, we became interested in developing a holistic understanding of physicochemical property space for anticancer molecules by carrying out a thorough analysis of properties for natural product derived anticancer drugs and a set of natural product lead anticancer molecules, as most of the anticancer drugs have been derived from the natural products [14]. herein, we present our efforts to develop a prospective mpo design tool for anti cancer molecules that does not focus on hard cutoffs or single end points but utilizes the eight essential physicochemical properties to prospectively align drug like attributes such as high permeability, low p-gp efflux liability, low metabolic clearance, and high safety into one molecule. in order to increase the flexibility in design and probability of identifying candidates with optimal pharmacokinetic and safety profile, we should not use hard cutoffs or focus on a single property, as it may restrict design space and may not align multiple attributes at once. experimental admet property calculations poor pharmacokinetic properties are one of the main reasons for terminating the development of drug candidates. computed physicochemical properties associated with compounds that have good oral bioavailability, less or no toxicity and optimum values of physicochemical properties are key parameters for the anti cancer drug discovery, and we need compounds with good pharmacokinetic properties [15, 16]. the drug set used in this study includes 24 natural product derived anti cancer drugs and structure of these drugs are mentioned in figure 1. to the best of our knowledge, all compounds in the drug set could be used as oral agents [17]. the lead candidates included in our analysis consisted of 27 natural products derived anti cancer compounds that collected from the literature belong to the several chemical classes as shown in figure 2 [18]. a complete list of the drugs and lead candidates used in the analysis appears in table 1. admet related physicochemical properties for 24 natural product derived anticancer drugs including clinical development candidates and 27 natural product lead anticancer compounds were predicted using osiris datawarrior version 4.2.2 software on a windows xp operating system [19]. datawarrior is able to calculate physicochemical properties, leador drug-likeness related parameters, ligand efficiencies, various atom and ring counts, molecular shape, flexibility and complexity as well as indications for potential toxicity. after calculating properties, these are automatically added as new columns to the data table. chemical structures for 24 natural product derived anticancer drugs including clinical development candidates and 27 natural product lead compounds were downloaded and saved individually in 3d sdf format from pubchem (www.pubchem.org). datawarrior is unable to optimize structures; therefore, geometry optimization of the molecules was performed in avogadro software prior to the prediction of physicochemical properties [20]. datawarrior software calculates the descriptors as inputs to independent mathematical models to estimate a range of admet values at relevant ph 7.4. physicochemical properties of interest included predicted lipophilicity (log p), predicted aqueous solubility (log s), topological polar surface area (tpsa), molecular weight (mw), hydrogen bond donor (hbd), hydrogen bond acceptor (hba), and number of rotatable bonds (nrot). specific adme properties of interest included predicted distribution coefficient at ph=7.4 (log d) (value predicted by acd/labs, admet & dmpk 4(2) (2016) 98-113 defining natural product derived anticancer drug space doi: 10.5599/admet.4.2.291 101 www.chemspider.com), predicted aqueous solubility (log s), quantitatively predicted apparent permeability (papp caco-2 cell), predicted effective permeability (peff), and predicted human intestinal absorption (hia). in order to evaluate the distribution of drugs and leads, we considered two important parameters including a fraction of unbound to plasma proteins (fu), and volume of distribution (vdss), a requirement of all clinical candidate through recently developed online admet calculation tool pkscm (http://bleoberis.bioc.cam.ac.uk/pkcsm/) [21]. figure 1. chemical structures of 24 natural products derived anti cancer drugs. to determine excretion routes, natural products anticancer drugs and leads we quantitatively predicted the total clearance and qualitatively predicted renal oct2 substrate. the safety profile of compounds is one of the most common factors in drug attrition (1). as part of our analysis of properties for natural products anticancer drugs and leads, we assess some of the major toxicity endpoints. we generated in silico data to assess potential for the following safety risks: drug-drug interactions (cyp inhibitions) including cyp3a4, cyp2c9, and cyp2d6, herg liability (inhibition of dofetilide binding), predicted ld50, predicted hepatotoxicity, predicted skin sensitization, cellular toxicity through pkscm tool and mutagenicity, tumorigenicity and irritant effects through datawarrior software. to access the likelihood of binding to transporter permeability-glycoprotein (p-gp), we used pgp_substrate model. we also calculated the three most crucial druglikeness filters including lipinski, ghose, and veber rules as well as the quantitative estimate of drug-likeness (qed) with the drug likeness tool (drulito) software (http://www.niper.gov.in/pi_dev_tools/drulitoweb/drulito_index.html). http://bleoberis.bioc.cam.ac.uk/pkcsm/) http://www.niper.gov.in/pi_dev_tools/drulitoweb/drulito_index.html deepika singh admet & dmpk 4(2) (2016) 98-113 102 figure 2. chemical structures of 27 natural products derived anti cancer lead molecules. results and discussion optimum physicochemical property space for anticancer molecules the 24 natural product derived anticancer drugs including clinical development candidates and 27 natural product lead anticancer compounds were evaluated against a set of eight calculated fundamental physicochemical properties that have gained wide acceptance as key parameters for drug design and development: (a) lipophilicity, calculated partition coefficient (log p); (b) distribution coefficient at ph=7.4 (log d); (c) molecular weight (mw); (d) topological polar surface area (tpsa); (e) number of hydrogen bond donors (hbd); (f) hydrogen bond acceptor (hba), (g) number of rotatable bonds (nrot) and predicted aqueous solubility (log s) [11, 22]. the calculated physicochemical properties value for the drugs and leads are mentioned in table 1. physicochemical property space as captured by these eight parameters was quite broad (figure 3). the mw values for the drugs varied from 246 to 853 with a median mw value of 514, while mw range for leads is much broad and varied from 114 to 975 with a median value of 336. the log p value of the drugs varied from 0.46 to 4.67 with a median log p value of 2.6. molecules in the lead set having the log p value from -2 to 16, which is quite broad with a median of 2.68. there was no significant difference in the median log p values between the two sets, although the drug set had a lower span of log p values. low hydrophilicity (e.g. high log p) values can cause poor absorption or permeation. our analysis suggests that for anticancer drugs, there may be a need to design compounds with further reduced log p or mw to better match the corresponding properties in the drug set. polarity, as described by polar surface area (tpsa), ranged from admet & dmpk 4(2) (2016) 98-113 defining natural product derived anticancer drug space doi: 10.5599/admet.4.2.291 103 about 29 å 2 to 224 å 2 with a median value of 118.25 å 2 for the drug set, while the polar surface area (tpsa) for lead molecules span from 0 å 2 to 266 å 2 . there was a significant difference in the tpsa values between lead candidates and drugs, almost ~60 % of the lead candidates having tpsa< 80 å 2 , oppositely 75 % drugs have tpsa ≥ 80 å 2 , which clearly suggests there is a huge need for the optimization of tpsa of lead candidates. the drugs and the lead candidates had a minimal number of hydrogen bond donors (hbd), with the median value of 2 for both sets. almost ~78 % of lead candidates and 83 % of drugs having the hbd value ≤ 3. lipinski’s ro5 identified hbd as a critical component of the drug property analysis and targets a hbd count (oh plus nh count) of < 5. based on the number of hbd associated with anticancer drugs and lead candidates, optimization of hbd to ≤ 3 may increase the probability of identifying better anticancer molecules. figure 3. physicochemical property distribution and statistics of drugs and lead candidates are shown for mw, log p, log s, hba, hbd, tpsa, nrot and log d. hydrogen bond acceptor is another valuable physicochemical property in ro5, drugs and the lead candidates having the median value of 10 and 4 respectively. only 11 out of 24 drugs in this study are following the hba < 10 rule of lipinski, on the other hand, most of the lead candidates ~89 % having the hba < 10, so anticancer lead candidates well following this ro5 compare to anticancer drugs. aqueous deepika singh admet & dmpk 4(2) (2016) 98-113 104 solubility is another very important parameter for the oral bioavailability. the recommended range for a molecule to be good oral bioavailable is (-6 ≤ log s ≤ 0.5), here all drugs are following this rule and almost ~95 % of lead candidates are also falling within the recommended range. the majority of this increased tpsa should originate from an increased number of hba, as hbd must be strictly controlled at ≤ 6 to avoid reducing oral bioavailability. log d provides a better measurement of lipophilicity for ionizable compounds; we know that hydrophilic molecules have higher solubility, but are less equipped to readily cross the cell membrane. hence, a compound is considered to be hydrophilic, if log d < 0, lipophilic if log d > 0, and molecule excessive lipophilic by molecules with log d ≥ 3.5. nearly ~85 % of the lead candidates having a log d value from 1 to 8 (1 ≤ log d ≤ 8), with a median of 3.15, which suggest maximum lead candidates are hydrophobic in nature. as expected for anticancer drugs, a similar but narrower range existed for log d, which varied from -0.86 to 4.88 with a median value of 2.57. in order to accurately access the differences between anticancer drugs and leads, four drug-like indices were utilized for comparison; lipinski’s ro5, ghose filter, vebers’s selective criteria for oral bioavailable drugs, and qed by drulito. the detailed results in percentage (%) of lead and drug molecules are following and violating the above mentioned most promising oral bioavailable rules are mentioned in the figure 4(a) and figure 4(b) respectively. figure 4. bar graph for percentage (%) of lead and drug molecules are (a) following, and (b) violating the lipinski’s ro5, ghose filter, vebers’s rule, qed, and all selected filters. the inspection of the bar graph shown in figure 4(a) reveals that leads are following most of the bioavailable rules greater than the drugs except the ghose filter. drugs and leads are showing almost equal percentage of molecules ~38 % following the all selected filters, which is obtained by considering all 4 bioavailable, filters (lipinski’s ro5, ghose filter, vebers’s rule, and qed) together. similarly, greater percentage of drugs are violating the bioavailable rules except the ghose filter. while the natural product derived anticancer drug space as defined by mw, log p, log s, hba, hbd, tpsa, nrot and log d, is broad. hence, our emphasis has been on defining physicochemical property rules for compounds to reduce attrition and increase the likelihood of candidates at various stages of anticancer drug development based on our analysis as well as earlier published various oral bioavailability rules by different research groups. our analysis shows the optimal property ranges (covering almost ~80 % or more of the anticancer drugs) used to select drug-like anticancer molecule for these properties are 200 < mw ≤ 800 da, 1< log p ≤ 5, -6 ≤ clog s ≤ -1, 5 ≤ hba ≤ 13, 1 ≤ hbd ≤ 5, 50 ≤ tpsa ≤ 180 å 2 , 0 ≤ nrot ≤ 10, log d=2.8, which may be very admet & dmpk 4(2) (2016) 98-113 defining natural product derived anticancer drug space doi: 10.5599/admet.4.2.291 105 table 1. important computed physicochemical properties for anticancer lead candidates and drugs. lead candidates mw log p log s hba hbd tpsa nrot log d 13-epi-sclareol 611 -1.41 -2.58 16 10 266 6 -2.42 6-gingerol 294 3.56 -3.25 4 2 67 10 3.64 ahwagandhanolide 975 5.36 -8.25 12 6 233 8 6.75 allicin 162 1.84 -1.22 1 0 62 5 2.01 anethol 148 2.68 -2.54 1 0 9 2 2.77 berberine 336 0.52 -4.67 5 0 41 2 3.96 beta carotene 537 13.87 -7.33 0 0 0 10 12.00 betulinic acid 457 6.37 -6.28 3 2 58 2 6.52 capsaicin 305 3.80 -3.32 4 2 59 9 3.90 catechins 290 1.51 -1.76 6 5 110 1 1.90 corchorusin-d 781 1.54 -5.21 13 8 208 6 1.31 curcumin 368 2.95 -3.62 6 2 93 8 3.55 diallyl sulfide 114 2.16 -2.01 0 0 25 4 2.03 diosgenin 415 4.88 -5.58 3 1 39 0 4.63 ellagic acid 302 1.28 -3.29 8 4 134 0 1.29 eugenol 164 2.27 -2.05 2 1 29 3 2.58 genistein 270 1.63 -2.73 5 3 87 1 1.12 indole-3-carbinol 147 1.10 -2.03 2 2 36 1 1.52 limonen 136 3.36 -2.54 0 0 0 1 3.50 drugs abiraterone 392 4.67 -4.75 3 0 39 3 4.72 amrubicin 483 0.84 -4.62 10 5 177 3 1.59 arglabin 246 1.48 -2.78 3 0 39 0 2.13 camptothecin 348 1.18 -2.74 6 1 80 1 2.02 carfilzomib 720 2.51 -4.57 12 4 158 20 4.43 combretastatin 334 2.17 -2.48 6 2 77 7 1.84 docetaxel 808 2.61 -5.81 15 5 224 13 2.59 ellipticine 246 3.90 -5.14 2 1 29 0 3.14 etoposide 589 0.67 -3.95 13 3 161 5 0.94 exemestane 296 3.61 -3.95 2 0 34 0 3.01 formestane 302 3.14 -4.04 3 1 54 0 2.55 homoharringtonine 546 3.65 -4.38 10 2 124 11 1.34 irinotecan 587 3.56 -4.50 10 1 113 5 3.25 paclitaxel 854 3.19 -6.29 15 4 221 14 3.06 podophyllotoxin 414 1.79 -3.84 8 1 93 4 1.95 rohitukine 305 1.62 -2.30 6 3 90 1 -0.86 roscovitine 354 2.58 -3.93 7 3 88 8 4.45 teniposide 657 1.35 -4.85 13 3 189 6 2.52 topotecan 421 0.46 -1.96 8 2 103 3 0.93 vinblastine 811 3.30 -5.08 13 3 154 10 4.88 vincristine 825 2.98 -5.53 14 3 171 10 3.97 vindesine 754 1.99 -4.62 12 5 165 7 2.38 vinflunine 817 3.95 -5.87 12 2 134 10 4.30 vinorelbine 779 3.59 -5.15 12 2 134 10 4.66 helpful in prospective design of anticancer molecules from natural products or identification of lead candidates from natural products that can successfully progress to the clinic and becomes better anticancer drug. for some of the physicochemical property, specifically mw, hba and nrot lead molecules deepika singh admet & dmpk 4(2) (2016) 98-113 106 are showing better optimal range compare to the drug candidates according to the ro5, which are quite significant for shaping the admet properties of potential anticancer drug candidates. profiling adme space of anticancer molecules potential therapeutic compounds are useless without having a good admet profile, and thus, it is essential to find the source of such diminished potency for developing a drug. significant advances in the development of ht in vitro adme assays have enabled computational scientist to make robust computational models to the earlier assessment of potential liabilities (low permeability, susceptibility to efflux transporters, etc.) associated with new potential lead compounds. in order to gain a better perspective on the adme properties of drugs and lead candidates, we evaluated the in silico profiling of these compounds to assess caco2 cell permeability, human intestinal absorption, and p-gp efflux liability [23]. we can classify the permeability of a molecule as low, or high based on the predictive model and its relative range of log papp (in 10 -6 cm/s) rates are, as follows: log papp > 0.9, considered to be high permeability, while log papp < 0.9, considered to be low permeability of the molecule. the caco2 cell permeability values for lead candidates and drugs are mentioned in table 2. in the caco2 cell permeability prediction, 70 % of the lead candidates show high log papp values; surprisingly the drugs had a lower percentage (30 %) with high log papp values. a similar discrepancy was observed when we assessed p-gp efflux liabilities for drugs and lead candidates. the p-gp efflux liability was assessed utilizing preadmet’s [https://preadmet.bmdrc.kr/] p-gp_substrate model. prediction of the likelihood of pgp efflux shows that all drugs and 60% of the lead candidates are considered to be p-gp efflux substrates; the predicted values from the pgp_substrate model for both drugs and lead candidates dataset are mentioned in table 2. an optimal clinical candidate could be achieved if it is possessed both high log papp and low p-gp efflux liability. the intestine is the primary site of absorption for the orally administered drugs; hence, we predicted the percentage (%) of human intestinal absorption of the drugs and lead candidates. systemic oral dosage requires compound properties that allow for dissolution and stability in the gastrointestinal (gi) tract, including the acidic environment of the stomach (ph 1–2 in fasted state, 3–7 in fed state) and the close to neutral environment (ph 4.4–6.6) of the small intestine [28. the % human intestinal absorption of the drugs and lead candidates was assessed by using pkcsm web server model [21]. all the drugs are showing good predicted human intestinal absorption > 60 %; while 93 % of the lead candidates predicted >70 % human intestinal absorption. the predicted % human intestinal absorption values for both drugs and lead candidates dataset are mentioned in table 2. many of the drugs in plasma will exist in equilibrium between an unbound state and a bound to serum proteins or whole blood proteins at various affinities. it is commonly accepted that only unbound drug may interact with anticipated molecular targets [24]; hence, the efficacy of a drug might affect by the degree to which it binds whole blood proteins. we have predicted the fraction unbound of both drugs and lead candidates through the predictive model of pkcsm, which was built using the measured free proportion of 552 compounds in human blood (fu). we also evaluated the steady-state volume of distribution (vdss) of drugs and lead candidates; another important parameter, which suggests the total dose of a drug would be required to be uniformly distributed to provide the similar concentration as in blood plasma. the values of predicted fraction unbound (fu) and vdss values for both drugs and lead candidates dataset are mentioned in table 2. evaluation of individual adme properties (papp, p-gp, fu) suggested that to increase the probability of success, the design should focus on optimizing all properties of a molecule. admet & dmpk 4(2) (2016) 98-113 defining natural product derived anticancer drug space doi: 10.5599/admet.4.2.291 107 table 2. computed adme properties for anticancer lead candidates and drugs. lead candidates caco2 permeability (log papp in 10 -6 cm/s) intestinal absorption (human) (% absorbed) vdss (human) (log l/kg) fraction unbound (human) p-gp substrate (yes/no) 13-epi-sclareol -0.791 28.495 -1.597 0.419 yes 6-gingerol 0.959 93.293 -0.061 0.248 yes ahwagandhanolide 0.197 82.511 -0.244 0.112 yes allicin 1.366 98.525 0.084 0.502 no anethol 1.391 98.814 0.066 0.322 no berberine 1.732 100 0.134 0.196 no beta carotene 1.397 91.234 1.319 0 no betulinic acid 1.294 91.58 0.418 0 yes capsaicin 1.429 92.542 0.084 0.172 yes catechins -0.38 71.562 -0.79 0.326 yes corchorusin-d -0.193 55.926 -0.401 0.318 yes curcumin 0.556 81.7 -0.677 0.103 yes diallyl sulfide 1.252 98.6 0.282 0.487 no diosgenin 1.245 96.426 0.931 0.09 yes ellagic acid -0.273 80.032 -1.214 0.27 yes eugenol 1.48 96.594 -0.071 0.375 no genistein 1.07 90.14 -0.836 0.192 yes indole-3-carbinol 1.31 94.062 -0.027 0.406 no limonen 1.248 98.048 0.503 0.422 no longimide 1.02 89.104 0.525 0 yes longitriol 1.277 90.662 0.857 0 yes lycopene 1.444 90.326 1.115 0 no methyl anolensate 1.241 100 0.04 0.14 yes resveratrol 1.294 89.885 -0.403 0.198 no s-allyl cyteine 0.375 88.593 0.026 0.711 no silymarin -0.363 74.315 -1.327 0.142 yes withaferin a 1.475 95.986 0.337 0.171 yes drugs camptothecin 1.118 99.412 -0.694 0.195 yes docetaxel -0.337 62.988 -1.068 0.138 yes etoposide -0.32 82.748 -1.185 0.194 yes irinotecan 0.986 93.659 -0.036 0.179 yes paclitaxel -0.239 71.535 -1.08 0.105 yes teniposide -0.255 90.274 -1.267 0.098 yes topotecan 0.668 83.641 -0.21 0.31 yes vinblastine 0.257 81.36 0.123 0.241 yes vincristine 0.186 78.558 -0.034 0.261 yes vinorelbine 1.413 90.783 0.282 0.193 yes abiraterone 1.20 98.16 0.67 0.05 yes amrubicin -0.26 66.22 -0.48 0.38 yes arglabin 1.59 100.00 0.49 0.41 yes carfilzomib 0.26 52.92 0.01 0.23 yes combretastatin 1.06 94.13 -0.71 0.21 yes ellipticine 1.35 94.47 -0.15 0.09 yes exemestane 1.52 100.00 0.69 0.18 yes formestane 1.39 96.29 0.50 0.21 yes homoharringtonine 0.57 79.16 0.11 0.35 yes podophyllotoxin 0.78 95.31 -0.82 0.13 yes rohitukine -0.22 73.76 0.18 0.49 yes roscovitine 1.01 90.21 1.97 0.59 yes vindesine 0.36 70.43 0.24 0.29 yes vinflunine 1.31 91.20 0.16 0.19 yes deepika singh admet & dmpk 4(2) (2016) 98-113 108 in order to understand the effect of physicochemical properties on adme attributes of the molecules, we analyzed adme attributes for both drugs and leads against all eight fundamental physicochemical properties. interestingly, tpsa, hbd, and hba showed good correlation with the caco2 cell permeability, with the correlation coefficient of 0.83, 0.8, and 0.7 respectively for the drug molecules. similarly, anticancer leads also showed slightly better correlation of physicochemical properties tpsa, hbd and hba with caco2 cell permeability, with the correlation coefficient of 0.83, 0.84, and 0.78, respectively. although, all three physicochemical properties (tpsa, hbd, hba) are inversely correlated with the caco2 cell permeability suggesting that lipophilicity is important for molecule to have good caco2 cell permeability and by optimizing the tpsa, hbd, and hba cell permeability of a molecule can be enhanced. human intestinal absorption also showed good correlation with the hbd, tpsa, and hba physicochemical properties for both anticancer drug and lead molecules. the correlation coefficient of % human intestinal absorption with hbd, tpsa, and hba was 0.88, 0.74, and 0.67 respectively for drugs and 0.91, 0.81, and 0.8 respectively for leads. this results clearly revealed the influence of physicochemical properties on adme attributes of the molecule for oral bioavailability and the key physicochemical properties especially hba, hbd, and tpsa need to be consider for further improvement in the adme profile of natural product derived anticancer leads. determining potential safety end points for anticancer molecules early prediction of the safety endpoints through in silico techniques screening have become regular practice for both designing new molecule and screening of the large chemical databases within pharmaceutical industries [25]. as part of our analysis of properties for anticancer drugs and lead candidates, we determine the potential toxicity end points for through pkcsm. most frequently measured end points to evaluate potential safety issues include inhibition of cytochrome p450 (cyps) monooxygenase enzymes to determine potential for drug-drug interactions [26], inhibition of herg potassium ion channel effects [27], lethal rat acute toxicity (ld50) and other crucial toxicity (ames toxicity, skin sensitization, and hepatotoxicity). all toxicity predictions for both drugs and lead candidates are presented in table 3. we qualitatively predicted the inhibition of cyp2d6 and cyp3a4 through pkcsm, which suggests the potential for drugs and leads candidates to mediate drug-drug interactions (ddi) through perturbation of clearance mechanisms for other drug substances. inhibition of the potassium herg channel might cause in prolongation of the qt interval of cardiac rhythm, which has resulted in the withdrawal of many clinical candidates from the market [28]. therefore, we have qualitatively predicted the potassium herg channel inhibition potential of drugs and lead candidates. the data obtained suggests that all of these drugs and lead candidates are non-inhibitor of the herg channel as mentioned in table 3. analysis of the inhibition data of cyp2d6 and cyp3a4 revealed that all the drugs are non-inhibitor of both cyp’s and all lead candidates are non-inhibitor of cyp2d6, while 89 % of lead candidates are non-inhibitor of cyp3a4. this data suggests that most of these drugs and lead candidates occupy desirable, low-risk space for ddi. drug metabolism and the drug excretion also have a significant role in the drug design process. issues related to metabolism have been commonly associated with the compounds failure in the clinical. understanding the metabolic pathways of drugs would be very helpful in predicting drug-drug interactions (ddi), toxicities, and pharmacokinetics [29]. many relationships between cyp family enzymes and in silico molecular properties have been available in the literature; the primary concern is inhibition of cyp3a4, which is correlated to increasing mw and log p [30]. this may lead to issues with clearance as well as drugdrug interactions. we have predicted the total clearance for drugs and lead candidates measured by the admet & dmpk 4(2) (2016) 98-113 defining natural product derived anticancer drug space doi: 10.5599/admet.4.2.291 109 proportionality constant and primarily occur as a combination of hepatic and renal clearance mentioned in table 3. hughes et al. have done the most considerable work with regard to the impact of molecular properties on in vivo toxicity, led to the “3/75 rule”, derived from an analysis of exploratory or dose-finding toxicology studies of 245 compounds at pfizer [31]. key finding emerged from this analysis was that compounds with a clog p < 3 and tpsa >75 å 2 were 2.5 times more likely to be non-toxic at the same total exposure. reversely, compounds those with high lipophilicity (clog p > 3) and low polar surface area (tpsa < 75 å 2 ) had an increased risk of widespread toxicities in short-term animal studies. one crucial elucidation of these results would be that promising lipophilic compounds with small polar functionality likely to have an increased chance of toxicity. a similar study by astrazeneca [32] on their compound failures showed a different profile, with the majority of failure happening with tpsa > 75 å 2 and clog p < 3. though, attrition in the high-log p–low-tpsa space can readily be rationalized via consideration of promiscuity and interactions across a range of systems. further eli lilly company study of > 400 (eli lilly) compounds supported the influence of compound lipophilicity on toxicology in rat toxicological studies [33]. in this analysis, there was a three-fold enrichment in toxic compounds when log p > 3, but tpsa had little or no influence. clearly, the benefits of establishing a link between important clinically relevant end points and simple descriptors such as log p and psa (which can be easily calculated before synthesis) are highly attractive. we also analyzed our natural product derived drugs and lead compounds predicted toxicity endpoints, to establish meaningful correlations between physicochemical properties and toxicity profile of compounds. predicted toxicities of drugs and leads have been categorized as “yes” or “no”. most of anticancer drugs in our dataset having the low lipophilicity (clog p < 3.5) are showing the hepatotoxicity (e.g. camptothecin, rohitukine, carfilzomib, docetaxel, etc.), out of which some drugs also having clog p < 3.5 and mw > 700 also showing hepatotoxicity (e.g. vinblastine, vincristine, vindesine, carfilzomib, and docetaxel, etc.). similarly, four drugs, showed the ames toxicity, also having the low lipophilicity (clog p < 2). this link between the low lipophilicity of compounds and toxicity is in line with the results of hughes et al. [31] and other research groups. furthermore, some drugs showing the toxicity but no specific correlations with physicochemical properties was found, possibly the toxicity was a consequence of the primary drug target mechanism or of a specific off-target pharmacology. examination of the relationship between physicochemical properties and other predicted toxicity end points, we found very good correlation for the drug molecules between the physicochemical properties and oral rat chronic toxicity (loael). the correlation coefficient of loael with mw, hba, hbd, tpsa, and nrot was 0.85, 0.84, 0.68, 0.84, and 0.77, respectively. all five physicochemical properties are positively correlated with the loael, suggesting the need for the optimization of these physicochemical parameters to avoid the loael toxicity. on the other hand, no correlation was observed between the loael and physicochemical properties for lead molecules, which is evident from the table 3, that drug molecules showed more toxicity endpoints compare to lead molecules. hence, establishing meaningful correlations between the physicochemical properties and toxicity of natural product derived oral anticancer drugs and leads might be useful for future anticancer drug discovery. admet & dmpk 4(2) (2016) 98-113 defining natural product derived anticancer drug space doi: 10.5599/admet.4.2.291 110 table 3. computed safety end points for anticancer lead candidates and drugs 1 cyp2d6 inhibitor cyp3a4 inhibitor total clearance renal oct2 substrate ames toxicity herg inhibitor oral rat acute toxicity (ld50) oral rat chronic toxicity (loael) hepatotoxicity skin sensitisation 13-epi-sclareol no no 0.183 no no no 1.526 2.231 no no 6-gingerol no no 1.369 no no no 1.861 2.381 no no ahwagandhanolide no yes -0.711 no no no 2.011 3.113 yes no allicin no no 0.721 no no no 2.423 1.317 no yes anethol no no 0.279 no no no 2.03 2.204 no yes berberine no no 1.324 no no no 2.587 1.998 no no beta carotene no no 1.024 no no no 1.766 0.666 no no betulinic acid no no 0.076 no no no 2.298 2.161 no no capsaicin no no 1.269 no no no 1.998 2.373 yes no catechins no no 0.215 no yes no 2.101 2.076 no no corchorusin-d no no -0.069 no no no 2.094 2.341 no no curcumin no yes 0.014 no no no 1.93 2.421 no no diallyl sulfide no no 0.562 no no no 2.274 1.689 no yes diosgenin no no 0.287 no no no 2.408 1.496 no no ellagic acid no no 0.539 no yes no 2.201 1.947 no no eugenol no no 0.28 no no no 1.994 2.304 no yes genistein no no 0.241 no yes no 2.309 2.075 no no indole-3-carbinol no no 0.555 no no no 2.301 1.836 no yes limonen no no 0.224 no no no 2.257 2.204 no yes longimide no no 0.036 no no no 2.347 1.445 yes no longitriol no no 0.239 no no no 2.218 2.035 no no lycopene no no 1.912 no no no 1.461 0.764 no no methyl anolensate no yes 0.265 no no no 2.594 1.088 no no resveratrol no no 0.147 no yes no 2.072 2.605 no no s-allyl cyteine no no 0.613 no no no 2.153 2.479 no no silymarin no no -0.092 no no no 2.184 2.593 no no withaferin a no no 0.37 no no no 2.294 1.983 no no deepika singh admet & dmpk 4(2) (2016) 98-113 doi: 10.5599/admet.4.2.291 111 table 3. conituned 2 drugs camptothecin no no 0.564 no yes no 2.465 1.845 yes no docetaxel no yes -0.172 no no no 1.617 2.808 yes no etoposide no no 0.041 no no no 2.071 2.331 no no irinotecan no yes 1.213 no no no 2.766 1.693 yes no paclitaxel no yes -0.121 no no no 1.708 3.081 no no teniposide no yes 0.409 no no no 2.162 2.689 no no topotecan no no 1.196 no yes no 2.58 1.867 yes no vinblastine no yes 0.618 no no no 2.251 2.588 yes no vincristine no yes 0.739 no no no 2.11 2.67 yes no vinorelbine no yes 0.622 no no no 2.343 2.5 yes no amrubicin no yes 0.4 no no no 2.531 1.721 no no arglabin no yes 1.059 no yes no 2.022 2.103 no no carfilzomib no no 0.839 no no no 1.986 1.482 no no combretastatin no yes 1.564 no no no 2.162 3.012 yes no ellipticine no yes 0.222 no no no 2.076 2.104 no no exemestane no yes 0.543 no no no 2.712 1.522 no no formestane no yes 0.832 no no no 2.062 1.822 no no homoharringtonine no yes 0.64 no no no 2.214 1.918 no no podophyllotoxin no yes 1.543 no no no 2.162 1.86 yes no rohitukine no yes 0.14 no no no 2.205 2.284 no no roscovitine no yes 0.495 no yes no 2.448 1.607 yes no vindesine no yes 1.174 no no no 2.761 1.683 yes no vinflunine no yes 0.467 no no no 2.414 2.729 yes no amrubicin no yes 0.365 no no no 2.311 2.267 no no admet & dmpk 4(2) (2016) 98-113 defining natural product derived anticancer drug space doi: 10.5599/admet.4.2.291 112 conclusions 3 improving the survival rate of clinical candidates and reducing the drug attrition is governed by multi-4 factors, and thus, a holistic strategy that addresses key attrition factors (safety, adme, and efficacy). 5 chemical space defined by physicochemical properties is vast, yet there are several design parameters that 6 medicinal chemists can follow when designing druglike compounds (e.g., lipinski’s rule of five) and 7 defining the parameters that increase the likelihood of identifying best in class molecules is of critical 8 importance. understanding the fundamental relationships between physicochemical properties and in vitro 9 and in vivo results is primary need to prospectively design compounds with an overall desired profile. as 10 part of our efforts to further build this understanding in the anticancer drug development space, we 11 undertook a thorough analysis of the physicochemical properties, adme attributes, and safety end points 12 for 24 natural product derived anticancer drugs and 27 natural product lead candidates. we examined a 13 comparison of eight fundamental physicochemical properties associated with these two sets of 14 compounds: log p, log d, mw, tpsa, hbd, hba, log s and nrot. the anticancer drug space defined by these 15 physicochemical properties is pretty broad, but our analysis identified the optimum ranges for each of 16 these properties. the optimal property ranges (covering almost ~80 % or more of the anticancer drugs) 17 were found to be 200 < mw ≤ 800 da, 1< log p ≤ 5, -6 ≤ clog s ≤ -1, 5 ≤ hba ≤ 13, 1 ≤ hbd ≤ 5, 50 ≤ tpsa ≤ 18 180 å 2 , 0 ≤ nrot ≤ 10, log d=2.8. analysis of in silico generated adme data reinforced that the majority of 19 anticancer drugs (70 %) are low permeable (caco2 of log papp (in 10 -6 cm/s) < 0.9), and also all drugs are 20 considered to be p-gp efflux substrates, and with low to moderate clearance rates. 21 on the other hand, our analysis showed that for anticancer drugs, there may be a need to optimize new 22 compounds with further reduced mw, hba, and nrot to better match the corresponding properties in the 23 marketed drug set. in addition, we have established meaningful correlations between the physicochemical 24 properties specially hba,hbd, and tpsa and adme attributes of the molecules that might be generally 25 applicable for the future anticancer drug development and optimization of the natural product derived 26 anticancer leads/clinical candidates. our study showed the meaningful correlations between 27 physicochemical properties and toxicity profile of compounds. log p and mw are most critical 28 physicochemical parameter and robust predictor of toxicity profile of anticancer leads/clinical candidates 29 we showed by our analysis that early prediction of physicochemical properties, adme attributes, and 30 safety attributes through in silico tools are all important parameters to enable better lead candidate 31 selections, saving considerable time and effort in the anticancer drug development. 32 33 acknowledgements: author is thankful to the director, central institute of medicinal and aromatic plants 34 (cimap-csir), lucknow. 35 36 references 37 [1] a. jemal, r. siegel, j. xu, e. ward. ca: a cancer journal for clinicians 60 (2010) 277-300. 38 [2] b.b. aggarwal, d. danda, s. gupta, p. gehlot. biochemical pharmacology 78 (2009) 1083–1094. 39 [3] s. coseri, mini-reviews in medicinal chemistry 9 (2009) 560-571. 40 [4] d.j. newman, journal of medicinal chemistry 51 (2008) 2589-2599. 41 [5] d.j. newman, g. m. cragg, journal of natural products 75 (2012) 311-335. 42 [6] m. s. kinch, a. haynesworth, s.l. kinch, d. hoyer, drug discovery today 19 (2014) 1033–1039. 43 deepika singh admet & dmpk 4(2) (2016) 98-113 doi: 10.5599/admet.4.2.291 113 [7] c. d. bradley, o. bjӧrn, g. fabrizio, k. jan, chemistry & biology 21 (2014) 1115 – 1142. 44 [8] n.a. meanwell, chemical research in toxicology 24 (2011) 1420–1456. 45 [9] p. barton, r.j. riley, drug discovery today 21 (2016) 72-81. 46 [10] m.p. gleeson, journal of medicinal chemistry 51 (2008) 817–834. 47 [11] c.a. lipinski, f. lombardo, b.w. dominy, p.j. feeney, advanced drug delivery reviews 23 (1997) 3–48 25. 49 [12] d.f. veber, s.r. johnson, h.y. cheng, b.r. smith, k.w. ward, k.d. kopple, journal of medicinal 50 chemistry 45 (2002) 2615-2623. 51 [13] a.k. ghose, v.n. viswanadhan, j.j. wendoloski journal of computational chemistry 1 (1999) 55-68. 52 [14] p. d. leeson, advanced drug delivery reviews 101 (2016) 22-33. 53 [15] e.c.a cornelis. hop attrition in the pharmaceutical industry: reasons, implications, and pathways 54 forward, ed. a. alex, c.j. harris, d.a. smith. (2016) john wiley & sons, inc. 55 [16] d.c. swinney, j. anthony, nature reviews drug discovery 10 (2011) 507–519. 56 [17] k.k. dholwani, a.k. saluja, a.r. gupta, d.r. shah, indian journal of pharmacology 40 (2008) 49–58. 57 [18] a. bhanot, r. sharma, m. n. noolvi, international journal of phytomedicine 3 (2011) 9-26. 58 [19] t. sander, j. freyss, m.v. korff, c. rufener, journal of chemical information and modeling, 55 (2015) 59 460–473. 60 [20] m.d. hanwell, d.e. curtis, d.c. lonie, t. vandermeersch, e. zurek, g.r. hutchison, journal of 61 cheminformatics 4 (2012) 1-17. 62 [21] e.v.p. douglas, t.l. blundell, d.b. ascher, journal of medicinal chemistry 58 (2015) 4066–4072. 63 [22] m. j. waring, bioorganic & medicinal chemistry letters, 19 (2009) 2844–2851 64 [23] d.a. smith, l. di, e.h. kerns, nature reviews drug discovery 9 (2010) 929-939. 65 [24] d.a. smith, h. van de waterbeemd, d.k. walker, (2001) pharmacokinetics and metabolism in drug 66 design. wiley–vch, weinheim, germany. 67 [25] k. a. houck, r.j. kavlock, toxicology and applied pharmacology 227 (2008) 163–178. 68 [26] v.p. miller, d.m. stresser, a. p. blanchard, s. turner, c. l. crespi, annals of the new york academy of 69 sciences 919 (2000) 26–32. 70 [27] m. deacon, d. singleton, n. szalkai, r. pasieczny, c. peacock, d. price, j. boyd, h. boyd, j.v. steidl-71 nichols, c. williams, journal of pharmacological and toxicological methods 55 (2007) 238–247. 72 [28] e.h. kerns, l. di, (2008). drug-like properties: concepts, structure design and methods: from adme 73 to toxicity optimization. (amsterdam, boston: academic press). 74 [29] c.m. hosey, l.z. benet, molecular pharmaceutics 12 (2015) 1456-1466. 75 [30] f. lovering, j. bikker, c. humblet, journal of medicinal chemistry 52 (2009) 6752–6756. 76 [31] j.d. hughes j. blagg, d.a. price, s. bailey, g.a. decrescenzo, r.v. devraj, e. ellsworth, y.m. fobian, 77 m.e. gibbs, r.w. gilles, n. greene, e. huang, t. krieger-burke, j. loesel, t. wager, l. whiteley, y. 78 zhang, bioorganic & medicinal chemistry letters 18 (2008) 4872–4875. 79 [32] d. muthas, s. boyera, c. hasselgren, medicinal chemical communications 4 (2013) 1058–1065. 80 [33] j.j. sutherland, j.w. raymond, j.l. stevens, t.k. baker, d.e. watson, journal of medicinal 81 chemistry 55 (2012) 6455–6466. 82 83 84 ©2016 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and 85 conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) 86 http://creativecommons.org/licenses/by/3.0/ doi: 10.5599/admet.4.3.314 232 admet & dmpk 4(3) (2016) 232-240; doi: 10.5599/admet.4.3.314 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper cross talk of c-jun n-terminal kinase and p38 map kinase modulate insulin and tnfα-stimulated vcam-1 expression in rat aorta endothelial cells gregory b. pott 1 , mark tsurudome 1 , jamie bui 1 , chelsea gardner 1 , and marc l. goalstone 1,2 * 1 eastern colorado health care system, denver colorado; 2 university of colorado, denver, colorado 80220 *corresponding author. e-mail: marc.goalstone@va.gov ; tel.: 303.399.8020 x 3610 received: june 06, 2016; revised: august 19, 2016; published: september 30, 2016 abstract vcam-1 cell surface expression was determined by flow cytometry in rat aorta endothelial cells stimulated with insulin and/or tumor necrosis factor-a in the absence or presence of short-hairpin rna inhibitors of c-jun nterminal kinase (jnk) and p38 map kinase. cells transfected with insulin alone exhibited moderate increases in cell surface vcam-1 expression, whereas cells stimulated with tnfα alone or in combination with insulin, exhibited a significant (p, 0.05) increase in vcam-1 surface expression. cells transfected with shjnk alone showed increased vcam-1 expression at the cell surface as compared to mock-transfected positive controls. in contrast, cells transfected with shp38 exhibited significant decreased insulin-stimulated vcam-1 expression at the cell surface, but only moderate decreased vcam-1 expression in cells stimulated by tnfα alone in combination with insulin. interestingly, in cells first tranfected with shjnk and then shp38, vcam-1 expression appeared to increase in an additive fashion. in contrast, cells transfected with shp38 first then shjnk, vcam -1 expression exhibited increased vcam-1 expression even in the presence of shp38. one can conclude that jnk is a potent negative regulator of insulinand tnfα-stimulated cell surface expression of vcam-1, whereas p38 is mild positive regulator of insulin-, but not tnfα-stimulated vcam-1 expression. jnk appears to be a more potent mediator of insulinand tnfα-stimulated cells surface vcam-1 expression than p38 map kinase. thus, it could be a therapeutic target for amelioration of inflammation-associated vcam-1 expression and its sequelae of events that lead to atherosclerosis. keywords jnk; p38, tnfα; vcam-1; hyperinsulinemia; atherosclerosis introduction the transduction of external signals to internal events in cells is mediated in part by a family of enzymes called protein kinases. kinases transfer the gamma phosphate of atp to their respective substrates, thereby activating these substrates and their downstream effectors. the super-family of mitogen-activated protein kinases (mapk) is one such group. interestingly, some kinases inhibit their downstream effectors by phosphorylation of their respective substrates. for example, the phosphorylation of glycogen synthase (gs) by glycogen synthase kinase deactivates gs [1]. the pathophysiology of many diseases involves abnormal kinase activity. thus, the characterization of kinase activity in cellular pathways and their pathophysiological sequelae have led to the better understanding of normal and pathological events, http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:marc.goalstone@va.gov admet & dmpk 4(3) (2016) 232-240 modulation of insulin and tnfα stimulated vcam-1 expression doi: 10.5599/admet.4.3.314 233 respectively. atherosclerosis is a major consequence of vascular dysfunction and leads to, but is not limited, to vascular smooth cell proliferation, lack of vascular compliance, endothelial cell proliferation, neointima formation and vascular inflammatory events. one particular characteristic of endothelial pathology is the increased expression of cellular adhesion molecules (cams) at the surface of endothelial cells, which are located at the margins of the intima in vessels [2-4]. the expression of the cams is mediated in part by the activity of intracellular endothelial cell kinases, which convey receptor activation at the cell surface to downstream effectors, which in turn upregulate the production and expression of cell surface molecules. recently, we reported that erk2 and akt are negative regulators of insulinand tnfα-stimulated vcam-1 [5] expression. it is not uncommon for kinases to "cross-talk" and modulate the activities of downstream effectors. we report here that c-jun n-terminal kinase (jnk) and p38 map kinase (p38) also modulate insulin and tnfα-stimulated vcam-1 expression, but in a different aspect. materials and methods materials all general lab reagents were purchased from sigma-aldrich (st. louis, mo). primary rabbit antibodies to jnk (9258s), p38 (9212s), and alpha-tubulin (2144s) were from cell signaling (boston, ma). the primary rabbit antibody to vcam-1 (nbp1-95622) was from novus biologicals (littleton, co) and goat anti-rabbitsecondary antibody irdye680rd (926¬68171) was from li-cor (lincoln, ne). dylight 488-conjugated antivcam-1 antibody was from thermo fisher scientific (pittsburgh, pa). four-well chamber slides were from thermo fisher and dapi mounting medium was from vector labs (burlingame, ca). rat aorta vascular endothelial cells (raec) (crl-1444) were from atcc (manassas, va) and culture medium was from life technologies (grand island, ny). jnk short-hairpin rna (shrna) (kr43333p) and p38 shrna (kr52703p) were obtained from sa biosciences/qiagen thermofisher (valencia, ca). transfection medium (sc108062) and transfection reagent (sc108061) were from santa cruz (dallas, tx). cell culturing raec were cultured in complete growth medium (cgm) [dmem with 4 mm l-glutamine, 4.5 g/lglucose and 1.5 g/l sodium bicarbonate) and supplemented with 10 % heat-inactivated fetal bovine serum (hi-fbs) (10438-026) (life technologies, grand island, ny) and 1 % antimycotic-antibiotic solution (15240062) (life technologies) and cultured at 37 °c, 5 % co2 atmosphere. western blot analysis sodium dodecyl sulfate polyacrylamide electrophoresis was performed on cleared lysates. western blot analysis was subsequently performed as previously described [6], with the following differences. after completion of protein transfer, membranes were washed in ultra-pure water for 5 min. membranes were then incubated in 3 % non-fat milk (milk) in tris-buffer saline (tbs) blocking solution for 1 h at room temperature and then incubated with a designated primary antibody solution (1:1000 in 3 % milk/tbs-t) overnight at 4 °c. membranes were washed 4 times with tbs plus tween (tbs-t) for 5 min at room temperature and then incubated with a goat anti-rabbit secondary antibody (1:5000 in 3 % milk/tbs-t) conjugated to fluorochrome ir680rd for 1 hour at room temperature. membranes were washed 4 times with tbs-t for 5 min each time at room temperature and then incubated with a rabbit anti-tubulin primary antibody solution (1:1000 in 3 % milk/tbs-t) for 3 h at room temperature. after washing the membranes 4 marc goalstone et al. admet & dmpk 4(3) (2016) 232-240 234 times with tbs-t, the membranes were again incubated with a goat anti-rabbit secondary antibody (1:5000 in 3 % milk/tbs-t) conjugated to fluorochrome ir680rd for 1 hour at room temperature. the membranes were washed 4 times with tbs-t and allowed to dry before performing densitometry. densitometry was performed using an odyssey licor system (lincoln, ne). alpha-tubulin protein was used to normalize vcam-1 signals. preparation of shrna stable cell lines raec were grown to 50-70 % confluence in cgm in 6-well culture plates. cells were transfected with shjnk or shp38 inhibitory plasmids alone or in combination as previously described [6]. cells were incubated in cgm containing 2 ug/ml of puromycin (sigma-aldrich) for 2-3 weeks for selection of puromycin resistant transformants. dual transfection of stable cell lines to examine the effect of simultaneous jnk and p38 knockdown on vcam-1 expression, the jnk shrna stable cell line (jnk kd) was transiently transfected with shp38 plasmid and the p38 shrna stable cell line (p38 kd) was transiently transfected with the shjnk plasmid. these two protocols were carried out in order to see if any differences occurred with respect to transfection sequence. stable cell lines were transiently transfected with vehicle or shrna as described above and incubated for 5 hr with the dna transfection mix. subsequently the transfection mix was aspirated and replaced with 2.0 ml cgm. stimulation of cells by insulin and/or tnfα occurred 48 hours after transient transfection was accomplished. stimulation of vcam-1 expression raec were cultured in cgm, whereas shrna stable cell lines (e.g., jnk kd and p38 kd) were cultured in cgm containing 2 ug/ml puromycin until assays were performed. after incubating the transfected cells for an additional 48 h, the cells were stimulated with or without insulin (10 nm) 1hour or tnfα (10 ng/ml) alone or in combination for six hours. thereafter we evaluated for vcam-1 expression by flow cytometry. flow cytometry mock transfected raec (control), stably transfected jnk knockdown (jnk kd) and p38 knockdown (p38 kd) cell lines were inoculated into 6-well tissue culture dishes, transiently transfected with vehicle, shp38 or shjnk plasmids, respectively, and allowed to settle for 48 hours. cells were then stimulated without or with insulin or tnfα alone, or combination as described above. the cells were washed twice with 2 ml of 1x pbs (gibco). the pbs was aspirated and 0.5 ml of cell dissociation solution non-enzymatic (sigmaaldrich) was added to each well. after incubating the cells at 37 °c and 5 % co2 for 30 min, 1 ml of 1 % bovine serum albumin (bsa, sigma-aldrich) in pbs was added to the cells and then were gently triturated into a single cell suspension. the cells were transferred to 5 ml falcon polystyrene round bottom tubes (thermo scientific) and centrifuged at 500 x g for 5 min. after aspirating the supernatants, the cells were resuspended in 3 ml 1 % bsa, pelleted at 500 x g by centrifugation, and the supernatants removed by aspiration. the cells were resuspended in 200 ul of 1 % bsa. two microliters of dylight 488-conjugated anti-vcam-1 antibody (life technologies, grand island new york) were added to each tube and the cells were resuspended by vortexing. the cells were incubated in the dark for 30 min at room temperature. the cells were centrifuged, washed twice with 3 ml 1% bsa and resuspended in 200 ul of 1 % paraformaldehyde (pfa, electron microscopy sciences, hatfield, pa). after incubating the cells for 5 min at room temperature, the cells were diluted with an additional 300 ul of pbs and analyzed using flow cytometry. the experiments were run on a bd lsrii (bd biosciences, san jose, ca). mfi and gating admet & dmpk 4(3) (2016) 232-240 modulation of insulin and tnfα stimulated vcam-1 expression doi: 10.5599/admet.4.3.314 235 percentages as part of data analysis was done using bd facsdiva v6 software. chamber slide cell preparation 2 x 10 5 of raec control, jnk kd or p38 kd stable cell lines were plated separately in 1 ml of cgm in each well of a 4-well chamber slide and allowed to grow for 24 h, 37 °c and 5 % co2. the medium was aspirated and 1.0 ml of fresh cgm was applied to the cells. jnk kd and p38 kd were mock transfected, or transiently transfected with shp38 or shjnk, respectively for 48 hours. cells were then treated with either vehicle, tnfα (10 ng/ml) or insulin (10 nm) alone, or tnfα (10 ng/ml) plus insulin (10 nm). the medium was aspirated and the cells were washed three times with pbs and then incubated in 400 µl of 4 % paraformaldehyde in pbs for 30 minutes. the medium was aspirated and washed three times with 1 ml of pbs. the final pbs wash was aspirated and 400 µl of a 1:1000 dylight anti-vcam-1 antibody solution in 1% bsa was added to each chamber and incubated for 30 minutes at room temperature. the cells were then washed three times with 500 µl of 1 % bsa. the chamber walls were removed and one drop of dapi mounting medium was added to each group of cells on the slide. cells were then sealed with a glass cover slip using clear nail polish. slides were kept in a dark refrigerator until microscopic visualization. confocal microscopy a single, non-confluent monolayer of cells was imaged with a leica tsc sp8x white light laser scanning confocal microscope (leica microsystems gmbh. ernst-leitz-straße 17-37 wetzlar, 35578 germany). all image acquisitions were carried out using the leica application suite x (version 1.1.0.12420, lasx af). excitation of the dapi channel was accomplished using a 405nm diode laser with an excitation intensity level of 2.67 %. emission signal was captured with standard pmt channel 1 and an emission gap of 430 – 480 nm. the leica supercontinuum white light excitation laser line (488nm) with 3 % intensity level was used to for alexa fluor 488. emission signals were captured with the leica hyd 2 detector (hybird 2 pmt) with an emission gap of 505 – 555 nm. data analysis data were analyzed by either unpaired student's t test (two groups) or anova with subsequent tukey posttest (several groups) as indicated. a "p" value of less than 0.05 was considered significant. results were expressed as the mean ± standard error of the mean (sem) of three or more independent experiments. results in our first set of experiments, we determined which shrna plasmid clone of each kinase would affect the greatest decrease in total cell content of that kinase (figure 1). clone #2 for jnk and clone #1 for p38 were the most effective in downregulating their respective kinases. using these two clones we then established stable raec cell lines that would constitutively express decreased levels of their respective kinases. raec cell lines stably transfected with shjnk and shp38 were denoted as jnk kd and p38 kd, respectively. in contrast, transiently transfected cells were named shjnk and shp38, correspondingly. additionally, we determined that not only single transfections effected down regulation of the respective proteins, but double transfections down regulated both proteins (figure 2). in previous studies, we noted that expression of total vcam-1 was not necessarily indicative of expression of vcam-1 at the cell surface [7,8]. additionally, since the thrust of these experiments was to define changes in amounts of vcam-1 protein at the cell surface (i.e., where adhesion to monocytes occurs), we performed only flow cytometry and confocal microscopy experiments and not western blot marc goalstone et al. admet & dmpk 4(3) (2016) 232-240 236 analysis. figure 1. transfection efficiency and knockdown by shjnk and shp38 clones in raec. cells were transiently transfected with individual shjnk or shp38 clones as described in methods. tubulin was used as loading controls. (a) western blot expression profiles of total jnk and p38 protein indicated changes in total jnk and p38 protein in the presence of designated shrna clones. total jnk and p38 kinase proteins were resolved by sds-page and determined by western blot analysis. (b) figure represents the expression of jnk and p38 in the presence of scrambled control (c) clones or designated numbered clones provided by the supplier. the numbers on the "x" axis represent the clone number. the expression of vcam-1 in mock transfected control (c) cells was set to 100%. expression of clones 1 4 are normalized to the expression of the scrambled shrna controls (c). figure 2. expression of jnk and p38 total protein in raec transfected with shjnk and shp38 plasmids. cells were transfected with clone-2 shjnk or clone-1 shp38 alone or both and cultured for 24 h. (a) total jnk and p38 proteins were resolved by sds-page and determined by western blot analysis. (b) graph represents the protein content of jnk (light gray bars) and p38 protein (solid black bars) as percent of controls (con) and results are expressed as the mean ± sem for four independent experiments. *, p < 0.05. we challenged cell groups with insulin (10 nm) or tnfα (10 mg/ml) alone or in combination and determined changes in surface vcam-1 expression via flow cytometry. interestingly, jnk kd cells transiently transfected with shp38 (figure 3a) exhibited a different profile from that seen with p38 kd transiently transfected with shjnk (figure 3b). in all instances, except in insulin only treated cells, vcam-1 significantly (p < 0.05) increased above positive controls. in comparison, vcam-1 was significantly (p < 0.05) decreased in the presence of insulin alone in p38 kd cells as compared to p38 kd positive controls. in contrast, vcam-1 surface expression neither increased nor decreased significantly in the presence of tnfα alone or in combination with insulin as compared to positive controls in p38kd cells. yet, in p38 kd + shjnk admet & dmpk 4(3) (2016) 232-240 modulation of insulin and tnfα stimulated vcam-1 expression doi: 10.5599/admet.4.3.314 237 cells, tnfα alone and in combination with insulin exhibited significant (p < 0.05) increases in vcam-1 expression as compared to positive controls. in order to corroborate the flow cytometry results, we performed confocal microscopy of control, jnk kd ± shp38 and p38 kd ± shjnk cells when stimulated with insulin or tnfα alone or in combination (figure 4). the confocal micrographs appeared to agree with our flow cytometry measurements with regard to expression of vcam-1 at the cell surface. figure 3. changes in expression of vcam-1 on the cell surface of raec were measured by flow cytometry. raec (controls), jnk kd and p38kd cells were transfected with either scramble control shrna, shjnk, shp38 or both, depending on the desired combinations. thereafter, the cells were stimulated with vehicle (pbs) or insulin (10 nm) for one hour or tnfα (10 ng/ml) for 6 hours or both. vcam-1 expression at the cell surface was analyzed by flow cytometry as described in methods. (a) cells were stably transfected with scrambled control shrna (open bars) or shjnk (jnk kd) (stippled bars) alone or in combination with transiently transfected shp38 (solid gray bars). (b) cells were stably transfected with scrambled control shrna (open bars) or shp38 (p38 kd) (light gray bars) alone or in combination with transiently transfected shjnk (solid black bars). graphs represent the percent of surface vcam-1 expression compared to unstimulated, mock-transfected cells and results are expressed as mean ± sem for five independent experiments.*, p < 0.05 vs. nonstimulated, mock-transfected controls.**, p < 0.05 vs. positive controls. $, p < 0.05 versus single-transfected cells. n = 5. figure 4. raec transfected with either shjnk (jnk kd) or shp38 (p38 kd) alone or in combinations exhibited differential surface expression of vcam-1 in cells stimulated with insulin and tnfα as determined by confocal microscopy. cells were plated and stimulated as described in methods. surface vcam-1 was detected by immunocytochemistry as described in methods and visualized by confocal microscopy. (con), controls, scrambled shrna. (jnk kd) cells stably transfected with shjnk. (p38 kd) cells stably transfected with shp38. (jnk kd + shp38) cells stably transfected with shjnk and transiently transfected with shp38. (p38kd + shjnk), cells stably transfected with shp38 and transiently transfected with shjnk. (no stim), no stimulation by insulin or tnfα. (ins only) cells stimulated with 10 nm insulin for 1 hour. (tnfα) cells stimulated with 10 ng/ml of tnfα for 6 hours. (ins + tnfα) cells stimulated with insulin for one hour and tnfα for six hours. marc goalstone et al. admet & dmpk 4(3) (2016) 232-240 238 discussion here we report that jnk is a negative mediator of insulin and tnfα-stimulated vcam-1 expression. in contrast, p38 kinase appears to be a positive regulator in the insulin signaling system, but plays a less significant role in the tnfα signaling pathway. additionally, jnk appears to be a more important mediator of vcam-1 expression than p38 kinase in the insulin and jnk signaling pathways. the presumed mechanisms for this phenomenon may lie in three regions of the cell: (1) the receptors for insulin and tnfα, (2) the respective proximal/downstream kinase mediators of these receptors and (3) common crosstalk of downstream mediators (i.e. the signal pathway kinases) of these upstream pathways. since there is no promiscuity between the receptors of insulin and tnfα the only additive effect that could exist would be via common downstream pathways and effectors. one of the most important mediators of the insulin receptor pathway is the adaptor protein src homology 2 (shc) protein [9] and the associated rodent associated sarcoma (ras) protein [10,11]. in turn these two proteins transduce external hormone information (insulin) to the internal signal pathway mediator jnk [12]. important mediators of the tnfα receptor are the tumor necrosis factor receptor type 1-associated death domain (tradd) protein, inhibitor of kb kinase (ikk) and jnk [13]. further downstream of these adaptor proteins are the proximal kinases p38 map kinase and apoptosis signalregulating kinase-1 (ask1). both jnk and p38 map kinase regulate their respective signal transducer and activator of transcription (stat) proteins [14]. there are two interesting results noted in this study: (1) in the jnk kd + shp38 or p38 kd + shjnk transfected cells, insulin plus tnfα elicited additive effects of vcam-1 expression as compared to vcam-1 expression in insulin or tnfα-stimulated control (mock-transfected) cells; and (2) there appears to be a lack of potentiation of insulin on tnfα-stimulated vcam-1 expression (figure 3). insulin alone in control stimulated cells does not significantly increase surface vcam-1 greater than that seen for non-stimulated controls (figure 3). in jnk kd cells insulin alone did not significantly (p = 0.08) increase vcam-1 expression greater than stimulated controls. in p38 kd cells, insulin significantly (p = 0.01) decreased vcam-1 expression as compared to insulin-stimulated controls, suggesting that p38 is a positive regulator of insulin-stimulated vcam-1 expression. however, as seen in jnk kd + shp38 cells and p38 kd + shjnk cells, insulin stimulated vcam-1 expression is significantly (p < 0.05) greater than that seen for stimulated controls. these data suggest that p38 regulation of insulinstimulated vcam-1expression is very different than that seen for tnfα-stimulated vcam-1 expression. tnfα-stimulated vcam-1 expression shows a different profile than insulin in the two cell lines. first, in control cells tnfα significantly stimulated an increase in vcam-1 compared to non-stimulated controls (figure 3). this suggests that tnfα is a positive regulator of vcam-1 expression. in jnk kd cells, tnfα stimulated and increase of vcam-1 significantly (p < 0.05) greater than that seen in stimulated control cells. interestingly this effect was not observed in p38 kd cells. however, in jnk kd + shp38 cells and p38 kd cells + shjnk, tnfα-stimulated vcam-1 expression was greater than that seen for jnk kd and p38 kd cells alone. this suggests again that jnk is a major regulator of tnfα-stimulated vcam-1 expression. another interesting phenomenon was seen in cells stimulated with insulin plus tnfα. neither in jnk kd cells nor in p38 kd cells did the combination of insulin and tnfα increase vcam-1 expression greater than tnfα alone. this phenomenon may have occurred because cells were not incubated in insulin for more than one hour. longer periods (e.g., 12 or 24 h) of insulin stimulation may show different vcam-1 admet & dmpk 4(3) (2016) 232-240 modulation of insulin and tnfα stimulated vcam-1 expression doi: 10.5599/admet.4.3.314 239 expression effects and graph profiles. thus, the detrimental aspects of insulin at short incubation times do not augment the effects of tnfα on vcam-1 expression. we have begun to address this issue and will report on these results in future communications. a third issue must be addressed. in our model (figure 5), p38 acts as a positive regulator of insulin towards the expression of vcam, but a negative regulator for tnfα. additionally, both p38 and jnk act negatively on the unknown signal pathway intermediate "a" in the insulin signaling model, but do not appear to have any influence on the intermediate "b". these possible scenarios would account for the differences seen in insulin-stimulated vcam-1 expression in jnk kd cells versus p38 kd cells. figure 5. model. surface vcam-1 expression in raec stimulated by insulin and tnfα is mediated in part by jnk and p38 kinase. a decrease in jnk alone has a reciprocal increase in vcam-1 expression. a decrease in p38 alone decreases vcam-1 expression in insulin-stimulated cells, but not tnfα-stimulated cells. a decrease in both jnk and p38 causes a significant increase in cell surface vcam-1 expression greater than that seen in jnk kd cells alone. finally, it is of interest to note that vcam-1 expression was significantly (p < 0.05) increased in nonstimulated jnk kd + shp38 cells and p38k kd + shjnk cells greater than that seen in non-stimulated controls (figure 3). this suggests that the lack of regulation from both jnk and p38 can, in and of itself, cause an increase in vcam-1 expression at the cell surface. this then would be detrimental to the vasculature. thus, the constitutive expression of jnk and p38 is important to the quiescent properties of vascular endothelial cells. diabetes is a risk factor for cardiovascular disease (cvd) [15,16] and within the spectrum of cvd are the effects of atherosclerosis. atherosclerosis is characterized by remodeling of the arterial vessel causing an increase in the endothelial layer, decrease in lumen diameter and decrease in blood flow; all of which are related to the inflammatory conditions that are regulated in part by hyperinsulinemia and tnfα [16,17]. here we report that insulin and tnfα stimulated vcam-1 expression of vcam-1 in raec is mediated in part by the intracellular kinases, jnk and p38. additionally, these two internal cellular pathway mediators appear to "cross-talk" with one another in regulating the expression of vcam-1 at the surface of aorta endothelial cells (figure 5). it is the increase in vcam-1 that accentuates the inflammatory process seen in arteries of patients with diabetes [18]. other cellular adhesion molecules, such as intercellular adhesion molecule (icam), may be involved with arterial inflammation [19-22]. future studies are targeted for icam experiments. thus, the activities of jnk and p38 are fodder for marc goalstone et al. admet & dmpk 4(3) (2016) 232-240 240 further research and may very well be therapeutic targets for the amelioration of inflammation in the arteries that is associated with atherosclerosis. acknowledgements we would also like to acknowledge the assistance of harsh pratap (flow cytometry technician for the mucosal and vaccine research colorado) and ron bouchard (microscopy core), both of the eastern colorado health care service (denver vamc). this work was supported by the research service of the department of veterans affairs (to m.l.g.), in which dr. goalstone is a recipient of a va merit award. references [1] m. maqbool, m. mobashir, n. hoda, eur. j. med. chem. 107 (2016) 63-81. [2] r. madonna, a. pandolfi, m. massaro, a. consoli, r. de caterina, diabetologia 47 (2004) 532-536. [3] m. okouchi, n. okayama, s. imai, h. omi, m. shimizu, t. fukutomi, m. itoh, diabetologia 45 (2002) 1449-1456. [4] p.a. watson, a. nesterova, c.f. burant, d.j. klemm, j.e. reusch, j. biol. chem. 276 (2001) 4614246150. [5] g.b. pott, m. tsurudome, n. bamfo, journal of inflammation 13 (2016) 6-14. [6] d.z. mackesy, m.l. goalstone, j. diabetes. 6 (2014) 595-602. [7] g.b. pott, m. tsurudome, n. bamfo, m.l. goalstone, j. inflamm. (london) 13 (2016) 6. [8] g.b. pott, m. tsurudome, j.d. bui, m.l. goalstone, advances in diabetes and endocinology 1 (2016) 7-13. [9] w.a. ricketts, d.w. rose, s. shoelson, j.m. olefsky, j. biol. chem. 271 (1996) 26165-26169. [10] m.l. goalstone, j.w. leitner, p. berhanu, p.m. sharma, j.m. olefsky, b. draznin, j. biol. chem. 276 (2001) 12805-12812. [11] m.l. goalstone, b. draznin, cell signal. 10 (1998) 297-301. [12] f. andreozzi, e. laratta, a. sciacqua, f. perticone, g. sesti, circ. res. 94 (2004) 1211-1218. [13] a.h. vy tran, s.h. hahm, s.h. han, j.h. chung, g.t. park, y.s. han, mutat. res. 777 (2015) 11-19. [14] t. matsuda, k. ferreri, i. todorov, y. kuroda, c.v. smith, f. kandeel, y. mullen, endocrinology 146 (2005) 175-185. [15] r.a. defronzo, e. ferrannini, diabetes care 14 (1991) 173-194. [16] k.j. reddy, m. singh, j.r. bangit, r.r. batsell, j. cardiovasc. med. (hagerstown) 11 (2010) 633-647. [17] d.z. mackesy, m.l. goalstone, j. inflamm. (lond.) 8 (2011) 34. [18] c. erbel, t.j. dengler, s. wangler, f. lasitschka, f. bea, n. wambsganss, m. hakimi, d. bockler, h.a. katus, c.a. gleissner, basic res. cardiol. 106 (2011) 125-134. [19] s. blankenberg, s. barbaux, l. tiret, atherosclerosis 170 (2003) 191-203. [20] y. huo, k. ley, acta. physio. scand. 173 (2001) 35-43. [21] i. kim, s.-o. moon, s.h. kim, h.j. kim, y.s. koh, g.y. koh, j. biol. chem. 276 (2001) 7614-7620. [22] t. watanabe, j. fan, int j cardiol. 66 suppl 1 (1998) s45-53; discussion s55. ©2016 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ manuscript doi: 10.5599/admet.2.1.31 43 admet & dmpk 2(1) (2014) 43-55; doi: 10.5599/admet.2.1.31 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper phosphate precipitates and water-soluble aggregates in re-analyzed solubility-ph data of twenty-five basic drugs alex avdeef in-adme research, 1732 first avenue, #102, new york, ny 10128, usa e-mail: alex@in-adme.com; tel.: +1 646 678 5713 received: february 15, 2014; revised: march 12, 2014; published: april 01, 2014 abstract the purpose of the study was to assess the stoichiometries of phosphate precipitates and determine the intrinsic solubilities, s0, of 25 basic drugs from their published solubility-ph profiles in the landmark study of bergström et al. (2004), where 0.15 m phosphate buffer media had been used. a secondary purpose of this study was to attempt to predict phosphate 1:1 and 2:1 solubility products, ksp, from knowledge of s0. the published data have been re-analyzed using a novel solubility-ph analysis computer program, pdisol-x tm . the program internally derives implicit solubility equations, given a set of proposed equilibria and constants (which are then iteratively refined by weighted nonlinear regression), and does not require explicit henderson-hasselbalch equations. the data were tested for the presence of phosphate precipitates of various stoichiometries, as well as the simultaneous presence of aggregated species, either cationic or neutral. the presence of particular species was suggested by the slope characteristics of the log s vs. ph curves. considerably different intrinsic solubility constants were found, compared to those originally reported, for several drugs (e.g., celiprolol, desipramine, haloperidol). the least soluble molecule, amiodarone, analyzed to have the extraordinarily low intrinsic solubility of 2 picograms/ml, a moderate salt solubility of 0.82 mg/ml at the gibbs-pka 5.4, corresponding to the species bh∙h2po4(s), and a substantial presence of the positively-charged pentameric aggregate, (bh)5. keywords sparingly-soluble drugs; ph-dependent solubility; salt solubility products; prediction of solubility products; solubility equations; aggregation; miniaturized shake-flask method. introduction when sparingly soluble basic drugs are added in substantial amount to phosphate-containing buffer solutions, there is likelihood that some drugs will precipitate as phosphate salts in neutral or acidic solutions. because the buffer can exist in three negatively charged overlapping forms in solution, different phosphate precipitates can form with positively charged ionizable drugs. the process is influenced by the solution ph in relation to the pka of the drug. if an appreciable amount of chloride is also present in solution, drug-chloride precipitates may form in lieu of, or along with phosphate precipitates. furthermore, in parallel to precipitation, formation of water-soluble oligomeric drug and micelle species http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:alex@in-adme.com avdeef admet & dmpk 2(1) (2014) 43-55 44 can substantially complicate the interpretation of the solubility measurements, especially when the drugs have more than one pka. although such systems had been studied on a specific case-by-case basis, general computational approaches suitable to tackle the above complexity had been scarce. perhaps as a result, the need for such analyses had not been widely appreciated. yet, quantitative analysis of such complex reactions clearly would have beneficial impact in early drug development in terms of clearer interpretations of mechanistic dissolution studies involving salts of sparingly soluble drugs in complex media with which the drugs can directly interact (e.g., such as simulated intestinal fluids) [1-5]. since salt solubility is a conditional constant, depending on the drug and the specific counterion concentrations, the determination of the thermodynamic solubility product allows for scalability of salt effects across a wide range of concentrations of substituents in formulation research. in addition, the formation of drug aggregates can have significant impact on measured in vitro permeability, since aqueous diffusivity of compounds depends on their size, e.g., as was suggested in the caco-2 study of retinoic acid in neutral solution [6]. there are numerous reports of aggregation forming systems [1,5-12]. for example, doxycycline [8], with intrinsic solubility s0 0.72 mg∙ml -1 , at 25 °c in 1 m nacl/hcl acidic solutions, forms dimers with k2 = 24 m -1 . such aggregates were found to have unusually high solubility, sensitively dependent on temperature. zhu and streng [9] found that the self-association of dolasetron to form cationic dimers and trimers was enthalpy driven (h-bond/aromatic ring rather than hydrophobic/electrostatic interactions), with the aggregation constants, k2-3, ranging from 4 to 50 m -1 at 25 °c. many nonsteroidal antiinflammatory drugs [10], such as indomethacin, diclofenac, ibuprofen, ketoprofen, naproxen, and sulindac, can self-associate by forming mixed-charge micelle or micelle-like structures. the reported cmc of diclofenac is about 25 mm and of ketoprofen is about 160 mm [10]. the re-analysis of the data of higuchi et al. [7] indicated that barbital and phenobarbital likely formed anionic dimers (reaching peak concentrations near ph 8), whereas oxytetracycline showed a pronounced tendency to form a cationic dimer (peaking near ph 2) [11]. saturated alkaline solutions of diprenorphine (ph > 9) appeared to suggest the formation of a mixedcharge anionic dimer [12]. in these and many other similar studies, it is evident that the solubility-ph profiles cannot be described accurately with the henderson-hasselbalch equation when aggregates form [1,5,11-13]. this manuscript is the third contribution in a series of studies applying a new salt solubility analysis, refinement, and simulation computer program, pdisol-x tm (in-adme research), to the study of nonhenderson-hasselbalch behaving solubility systems [11,12]. the main objective of this brief commentary is to assess the stoichiometries of phosphate precipitates and determine the intrinsic solubilities, s0, of 25 basic drugs from their published solubility-ph profiles in the landmark study of bergström et al. [13], where 0.15 m phosphate buffer media had been used. a parallel purpose of the study was to attempt to predict the 1:1 and 2:1 phosphate solubility products, ksp, from knowledge of s0, something that had not been done hitherto, as far as we know. experimental legacy data in the bergström et al. [13] study, the solubility experiments were performed at 23 °c, using a miniaturized shake flask method. each drug was added in excess to 0.05-0.2�ml of 0.15 m phosphate buffer. the ph of each drug suspension was then adjusted using solutions of h3po4 or koh to a selected ph (in some cases as low as 1.5 and as high as 12). the suspensions were then agitated at 300 rpm on a admet & dmpk 2(1) (2014) 43-55 phosphate precipitates & aggregates of basic drugs doi: 10.5599/admet.2.1.31 45 plate shaker for 24 h, the assigned equilibration time. after 24 h, the samples were centrifuged to separate the precipitate from the solution. the drug concentration in the supernatant was analyzed by hplc (uv/fluorescence detection). light scattering measurements did not find colloidal particulates or agglomerates in the supernatant. practically insoluble amiodarone, carvedilol and mifepristone could not be detected directly by hplc. subsequently, these drugs were studied in 6 – 24 % w/w methanol-water mixtures. the corresponding intrinsic solubilities, s0, were determined by extrapolation of the watercosolvent solubilities to zero cosolvent. in most cases, the value of s0 was assigned to the drug solubility at ph ≈ pka + 2, where the drug was expected to be predominantly uncharged. the assays appeared to be well designed. however, it would have been helpful presently to have access to the actual amounts of drug added per ml of buffer in the original study. in addition, it is plausible that some of the practically insoluble drugs had not reached true thermodynamic equilibrium in 24 h. this could be a topic of further study. refinement of intrinsic and salt solubility and aggregation constants the new data analysis method uses log s ph as measured input data (along with the standard deviations in log s) into the pdisol-x computer program, as described previously [11,12]. solubility values from literature sources are converted to molarity units for processing by the program. briefly, an algorithm was developed which considers the contributions of all species present in solution, including all components of buffers or mixtures thereof. the approach does not depend on any explicitly derived extensions of the henderson-hasselbalch equations [1]. the uniqueness of the computational algorithm is that it derives its own implicit equations internally, given any practical number of equilibria and estimated constants, which are subsequently refined by weighted nonlinear least-squares regression. drug-salt precipitates, -aggregates, -complexes, bile salt and -surfactant species can be modeled [1,5]. the presence of specific buffer-drug formed species can be tested. the computer program calculates the distribution of species corresponding to a sequence of additions of standardized strong-acid titrant hcl (or weak-acid titrants, e.g., h3po4) to simulate the speciation in the suspension down to ph ~ 0. subsequently, a sequence of perturbations with standardized naoh (or, e.g., koh) is simulated, and solubility calculated at each point (in ph steps of 0.005-0.2), up to ph ~ 13. the ionic strength is rigorously calculated at each step, and pka values (as well as solubility products, aggregation constants) are accordingly adjusted [11]. at the end of the speciation simulation, the calculated log s vs. ph curve is compared to actual measured log s vs. ph. a log s-weighted nonlinear least squares procedure refines the proposed equilibrium model, using analytical expressions for the differential equations. the process is repeated until the differences between calculated and measured log s values reach a minimum, as described elsewhere in detail [11]. the analysis of the bergström log s ph data tested for the presence of solid species b(s), bh∙cl(s), bhb∙cl(s), bh∙h2po4(s), (bh)2∙hpo4(s), bhb∙h2po4(s), (bhb)2∙hpo4(s), (bh)3∙h2po4∙hpo4(s), as well as water soluble cationic aggregates (bh)n and (bhb)n, n = 1-6, and the uncharged aggregate b2. up to three different solids could be tested in a single calculation. the presence of particular species was suggested by the slope characteristics of the log s ph curve [1,5,7-11]. avdeef admet & dmpk 2(1) (2014) 43-55 46 results and discussion based on the results of the analysis, the 25 drugs were divided into six groups, according to the type of deviations from the henderson-hasselbalch equation in the non-salt regions (ph > pka gibbs [14,15]) of the log s – ph curves, as discussed at length by avdeef [1].  case 0 (24 % of all drugs): follow henderson-hasselbalch equation, and apparent s0 = true s0.  case 1b (24 %): apparent pka < true pka and apparent s0 > true s0, which is consistent with the formation of water-soluble uncharged aggregates; degree of aggregation cannot be determined from ph-dependence.  case 2b (16 %): diagonal region of the log s – ph profile has non-unit slope, indicative of cationic aggregates, where the average degree of aggregation defined by the value of the slope; apparent s0 = true s0.  case 3b (16 %): diagonal region of the log s – ph profile (slope = -1) shifted to higher ph than predicted by henderson-hasselbalch equation (i.e., apparent pka > true pka), indicative of mixed-charge dimeric aggregates, bhb + , or higher order oligomers of the same stoichiometric ratio; apparent s0 = true s0.  two group of compounds showed combined case 1b+2b (8 %) and case 2b+3b (12 %). the results of the analysis of the bergström et al. data are summarized in table 1, with all constants presented in molarity or molarity related units. figure 1 compares intrinsic solubility, ps0 (-log10 s0), of the compounds determined in this study to that reported by bergström et al. [13]. amiodarone is not shown in figure 1, since the previously reported value was stated conditionally (table 1). the eight compounds in figure 1 represented as checkered circles are all examples of case 1b, and showed the largest variances from the reported intrinsic solubility values. on the other hand, the agreement between the filled circle compounds (corresponding to other cases) determined here and reported previously is very good: ps0 bergström = 0.02 + 1.007 ps0 pdisol-x , r 2 =0.99, s=0.17, f=1014, n=16. table 1. summary of the results of the re-analysis of the bergstrom et al. [13] data. admet & dmpk 2(1) (2014) 43-55 phosphate precipitates & aggregates of basic drugs doi: 10.5599/admet.2.1.31 47 figure 1. comparison of previously reported intrinsic solubility values, ps0 (molarity based units), to those determined here. aggregation reactions case 0 – no aggregation; henderson-hasselbalch equation accurate for ph > pka gibbs figure 2 shows the six compounds which show no appreciable deviation from the log s – ph curves calculated by the henderson-hasselbalch equation (non-salt region: ph > pka gibbs ). the compounds are arranged in the order of decreasing intrinsic solubility. carvedilol and chlorprothixene indicated the lowest intrinsic solubility, with values < 0.5 µg∙ml -1 . the most soluble molecule is lidocaine, with s0 = 2.5 mg∙ml -1 . the ph at the point of discontinuity between the (most-often flat) salt region and the diagonal slope = -1 curve is called pka gibbs [14,15]. for example, pka gibbs = 6.2, 9.2, 7.0, 2.9, 6.7, 6.0 for the drugs in figure 2, in the order of their appearance. with the exception of verapamil and carvedilol, simple 2:1 and/or 1:1 phosphate salts appear to form with this class of compounds. verapamil data in the salt region (ph < pka gibbs ) are consistent with the formation of a chloride salt with the mixed-charge dimer of verapamil, (bhb + )∙cl (s). the region below ph 7 displays a slope = -0.5, characteristic of such a stoichiometry. carvedilol salt region is best fit with a diphosphate stoichiometry: 3 bh + + h2po4 + hpo4 2  (bh)3h3(po4)2(s). verapamil and carvedilol (and mifepristone, below) illustrate why it is better to designate the ph of discontinuity as pka gibbs , rather than “phmax,” since the discontinuity can occur at a ph where the solubility is not at all at its maximum value [14,15]. avdeef admet & dmpk 2(1) (2014) 43-55 48 figure 2. case 0 compounds which show henderson-hasselbalch form log s – ph curves in the non-salt region. case 1b – neutral aggregates; henderson-hasselbalch equation not accurate figure 3 shows the six compounds with apparent pka < true pka, which is consistent with the formation of water-soluble uncharged aggregates. the henderson-hasselbalch equation does not accurately predict the log s – ph curves across the entire ph range. for case 1b drugs, the apparent s0 > true s0. the degree of aggregation cannot be determined from ph-dependence, since the aggregated compound is in the uncharged form. the compounds in figure 3 are arranged in the increasing order of aggregation strength. figure 3. case 1b compounds: log s – ph curve distortions consistent with the formation of neutral aggregates. admet & dmpk 2(1) (2014) 43-55 phosphate precipitates & aggregates of basic drugs doi: 10.5599/admet.2.1.31 49 the compounds in the figure 3 either (i) form stable water-soluble aggregates of the uncharged drug, or (ii) the compounds show a pattern of curve distortion, which is characteristic of a system that had not reached a true state of thermodynamic equilibrium at the 24 h time point. that is, neutral aggregates may be thermodynamically unstable and are very slowly coalescing into solid form. examples of the former (i) effect have been reported with molecules like piroxicam [1,5], and cases of the latter (ii) effect have been reported for terfenadine, which at 24 h showed more pronounced distortion that at 68 h [1,5] (it would be useful to repeat the measurements of the six drugs in figure 3 at much longer equilibration times than the 24 h actually used). it is evident that to analyze case 1b compounds, it is necessary to know the true pka accurately. it is with this class of compounds that there were the largest deviations in figure 1. apparently, the published intrinsic values [13] ascribed the reported intrinsic solubility to the apparent intrinsic values (the minimum red (solid) curve value rather than the minimum dashed curve value). the six examples in the above figure show flat curves in the salt region, corresponding to simple stoichiometry phosphate or chloride salts. case 2b – cationic aggregates; henderson-hasselbalch equation not accurate figure 4 shows four compounds which show non-unit slopes in a part of the diagonal region of the log s ph profile. this is consistent with the formation cationic aggregates. the average degree of aggregation is defined by the value of the slope. apparent s0 = true s0. figure 4. case 2b compounds which show log s – ph curve distortions consistent with the formation of cationic aggregates. bupivacaine analyzed as consisting of (bh + )2 cationic dimers for ph < 7, co-existing with the 1:1 phosphate precipitate. this raises the observed solubility in the salt region, due to competition between avdeef admet & dmpk 2(1) (2014) 43-55 50 the drug in the solid and aggregated states. consequently, greater amount of compound added to solution makes the compound more soluble in the salt region. the data shape of the other three molecules was most consistent with the presence of pentameric or hexameric cationic aggregates. it was possible to rationalize the amiodarone log s – ph profile as being distorted from the shape predicted by the henderson-hasselbalch equation due to the formation of the pentamer (bh + )5. if this were a valid interpretation of the data, then there are some very interesting consequences. in the analysis of the amiodarone data, it was assumed that 5 mg of drug were added to 1 ml of buffer. the ph < 5 region of solubility curve (fig. 4c) depends on both the presence of the cationic aggregate and the phosphate precipitate. since the concentration of the aggregate depends steeply on the amount of amiodarone added to the buffer, a series of simulations with decreasing amounts of amiodarone results in lowering the solubility in the salt region. in the extreme, maximum solubility can decrease from about 10 -2 (fig. 4c) to 10 -7 m (not shown). this suggests that a parenteral injectable solution can show solubility as high as 5 mg∙ml -1 in the ph 1-5 region, provided enough drug were present in the formulation. however, if a much smaller amount of drug were added, then the solubility could be limited by the apparent minimum pksp,1:1 (-log10 ksp,1:1), to about 14 ng∙ml -1 , with no influence of the aggregate, according to the pdisol-x simulation calculation. amiodarone is the drug of choice in the treatment of arrhythmia, but it has side effects that limit its long-term use. in the treatment of chronic arrhythmia, oral doses of amiodarone may take 30 days to develop full therapeutic effect because of the extreme hydrophobicity of the drug. once treatment is discontinued, it may take 1-2 months to wash out the drug completely. on the other hand, during recovery from heart surgery, the atrium can start to fibrillate dangerously. one remedy is to inject an amiodarone parenteral formulation [16] into a vessel carrying blood into the heart. within a few minutes, the drug can take effect. the analysis of the log s – ph profile of amiodarone can shed light on the above two seemingly incompatible circumstances. hydralazine was problematic to analyze. figure 4d shows a red (solid) curve based on the assumed presence of a hexameric (bh + )6 species, which is most consistent with the data that deviates most from the dashed line in the non-salt region. however, the consequence of such a species being present in parallel with the phosphate precipitate is that the pksp,1:1 would analyze as 5.12 ± 0.24 (the value in table 1). since the ps0 value is unaffected by the cationic aggregate, the relationship between the ksp and s0 can be dependent on the aggregation model. it will be suggested below that it may be more meaningful to assume that the data in the deviation region ph 6-7 is not at equilibrium. when the data are excluded from the analysis and no stable aggregate is assumed present, then pksp,1:1 analyzes as 2.74 ± 0.04. the latter value, which follows the relationship in figure 8a more closely, was used in the plot and regression analysis. case 3b – mixed-charge cationic dimers; henderson-hasselbalch equation not accurate figure 5 shows four examples of case 3b log s – ph profiles. these are characterized by the (slope = -1) diagonal region of the log s – ph profile shifted to higher ph (i.e., apparent pka > true pka) than predicted by henderson-hasselbalch equation, indicative of mixed-charge dimeric aggregates, bhb + . in such cases, the apparent s0 is the same as the true s0. the compounds are arranged in the order of increasing dimerization formation constant, with trimethoprim nearly showing henderson-hasselbalch characteristic of case 0. mifepristone appears to admet & dmpk 2(1) (2014) 43-55 phosphate precipitates & aggregates of basic drugs doi: 10.5599/admet.2.1.31 51 show a phosphate salt formation stoichiometry (bhb + )∙b∙(h2po4 )(s), characterized by an average slope -0.53 ± 0.06 in the ph 1.0 – 3.4 salt region, similar to the case of verapamil (fig. 2c). trimethoprim and promethazine.hcl show simple 1:1 and 2:1 phosphate salts. figure 5. case 3b compounds which show log s – ph curve distortions consistent with the formation of mixed-charge dimeric aggregates. mixtures of cases; henderson-hasselbalch equation not accurate figures 6 and 7 show examples of mixtures of case 1b+2b and 2b+3b, respectively. the case 1b distortion in thioridazine may be due to incomplete equilibration time. figure 6. two examples of compounds showing two overlapping distortion types: case 1b (neutral aggregates) and case 2b (cationic aggregates). avdeef admet & dmpk 2(1) (2014) 43-55 52 figure 7. three examples of compounds showing two overlapping distortion types: case 2b (cationic aggregates) and case 3b (mixed-charge dimers). relationships between intrinsic solubility and the two phosphate solubility products relationship between pksp,1:1 and ps0 just 18 of the 25 drugs considered were interpreted to include 1:1 bh∙h2po4 salt formations. the other seven drugs were either salts with complicated phosphate stoichiometry or chloride salts (for very soluble drugs introduced as hydrochlorides). the average of the ratio values pksp,1:1 / ps0 is 0.73 ± 0.34. figure 8a shows the empirical correlation between the pksp,1:1 constant and the intrinsic solubility ps0 for the 18 compounds. a trend is evident, in that the less soluble the uncharged form of the drug, the lower is the ksp value. although the variance is relatively large, the relationship can still be a valuable basis for prediction of the phosphate solubility product in the absence of measurement. figure 8. (a) correlations between the 1:1 salt solubility products and the intrinsic solubility constants for 18 of the studied compounds. (b) correlation between 1:1 and 2:1 drug-phosphate solubility products for 9 of the studied compounds. the bracket concentration products (molarity related units) in the labels have –log10 implied. relationship between pksp,1:1 and pksp,2:1 nine of the 25 drugs were interpreted to have both 1:1 and 2:1 phosphate salt precipitation. figure 8b shows an extraordinarily high correlation between the two salt solubility products. this is probably expected from equilibrium equation considerations, since the slope of 0.46 is so close to the one-to-two ratio of stoichiometries. the average ratio of the two pksp constants is 0.50 ± 0.04 in figure 8b. admet & dmpk 2(1) (2014) 43-55 phosphate precipitates & aggregates of basic drugs doi: 10.5599/admet.2.1.31 53 prediction equations for phosphate salts from the correlation relationships in figure 8, two useful empirical equations may be derived: pksp,1:1 = 0.525 + 0.554 ps0 (1) pksp,2:1 = 0.744 + 1.207 ps0 = -0.4002 +2.1796 pksp,1:1 (2) these equations were used to fill in the “missing” values in table 1, indicated in red italic numbers. chloride salts since some relatively soluble salts were introduced as drug hydrochlorides, the solubility data were tested for the possibility of chloride precipitates. the log s – ph plots for celiprolol, orphenadrine, terazosin and pramoxine appeared to be consistent with the formation of 1:1 chloride salts. the salt region in the solubility plot is expected to be a horizontal line, since the common ion effect is not expected when the ph is adjusted using h3po4. however, the above assignments are tentative without additional data collected at different added amounts of these drugs, especially with data taken to very low ph values. how constant are solubility products? the pdisol-x program calculates the ionic strength at each ph point in the processing of the log s – ph data, as described in detail elsewhere [11]. figure 9 shows the speciation analysis of disopyramide in the phosphate salt region, ph 2 – 9. figure 9b shows that the ionic strength varies from about 0.1 to 0.4 m as the ph changes from 2 to 9. the vertical dotted line near ph 6 in figure 9a indicates the transition ph between the 1:1 and the 2:1 phosphate salts. the ionic strength dependence of the two ksp constants differs, which affects the calculated concentrations. the result of this is that the solubility product above ph 6 is not exactly constant, due to the effect of ionic strength. the analysis program is particularly well suited to explore the consequences of the choices of various species tested in the solubility model. avdeef admet & dmpk 2(1) (2014) 43-55 54 figure 9. the effect of the ionic strength on the constancy of the solubility product. conclusions evidently, concentrated phosphate buffers (0.15 m) can dramatically influence the solubility profiles of sparingly soluble drugs. clearly, the bergström et al. [13] data for the 25 compounds studied could not be explained by the simple henderson-hasselbalch equation in all but six cases. apparently, the aggregated species proposed here were either not recognized originally, or were not fully explored. such anomalies may be common with sparingly soluble drugs, but are not always easy to recognize. pdisol-x could be a helpful new tool in early development to further aid in the analysis of dissolution mechanisms of sparingly soluble drugs, which depend on the stoichiometry and solubility of drug species, especially of salts. acknowledgements helpful discussions with christel bergström of uppsala university, and gergely völgyi and krisztina takács-novák of semmelweis university (budapest) are greatly appreciated. admet & dmpk 2(1) (2014) 43-55 phosphate precipitates & aggregates of basic drugs doi: 10.5599/admet.2.1.31 55 references [1] a. avdeef. absorption and drug development second edition, wiley-interscience, hoboken, nj, 2012, pp. 251-318. [2] a. glomme, j. märz, j.b. dressman. in: b. testa, s.d. krämer, h. wunderli-allenspach, g. folkers (eds.), pharmacokinetic profiling in drug research: biological, physicochemical, and computational strategies, wiley-vch, weinheim, 2006, pp. 259–280. [3] j.h. fagerberg, o. tsinman, k. tsinman, n. sun, a. avdeef, c.a.s. bergström. mol. pharmaceutics 7, (2010) 1419-1430. [4] a. avdeef, o. tsinman. pharm. res. 25 (2008) 2613-2627. [5] a. avdeef. adv. drug deliv. rev. 59 (2007) 568-590. [6] a. avdeef, k.y. tam. j. med. chem. 53 (2010) 3566-3584. [7] t. higuchi, m. gupta, l.w. busse. j. am. pharm. assoc. 42 (1953) 157-161. [8] j.b. bogardus, r.k. blackwood, jr. j. pharm. sci. 68 (1979) 188-194. [9] c. zhu, w.h. streng. int. j. pharm. 130 (1996) 159-168. [10] a. fini, g. fazio, g. feroci. int. j. pharm. 126 (1995) 95-102. [11] g. völgyi, a. marosi, k. takács-novák, a. avdeef. admet & dmpk 1(4) (2013) 48-62. [12] a. avdeef. admet & dmpk 2 (2014) 33-42 [13] bergström cas, luthman k, artursson p. eur. j. pharm. sci. 24 (2004) 387-398. [14] w.h. streng. int. j. pharm. 186 (1999) 137–140. [15] a. avdeef. pharm. pharmacol. commun. 4 (1998) 165-178. [16] amiodarone-containing parenteral solution patent: ep 1 267 865 b1 (15 feb 2006). ©2014 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ manuscript doi: 10.5599/admet.3.2.193 84 admet & dmpk 3(2) (2015) 84-109; doi: 10.5599/admet.3.2.193 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index review suggested improvements for measurement of equilibrium solubility-ph of ionizable drugs alex avdeef in-adme research, 1732 first avenue #102, new york, ny 10128 usa corresponding author: e-mail: alex@in-adme.com; tel.: +1-646-678-5713 received: june 14, 2015; revised: june 22, 2015; published: july 01, 2015 abstract the accurate prediction of solubility of drug-like molecules is difficult, and perhaps a satisfactory general model is not yet available. the most cited challenge to good prediction has been the lack of enough highquality and drug-relevant solubility data that adequately cover the chemical space of drugs. this review addresses data quality in solubility measurement. specifically, the “gold standard” shake-flask and related methods used to measure equilibrium solubility of ionizable drug-like compounds as a function of ph were reviewed. over 800 publications were examined. many factors affecting the quality of the measurement were recognized, and a number of suggestions are offered to improve the experimental methodology. some of the suggestions focus on improving methods for future measurements, and some refer to improvements in data mining, i.e., to ways of extracting more reliable information from existing data. by normalizing data for ph (i.e., deriving intrinsic solubility, s0) and for temperature (by transforming measurements performed in the range 20 – 40 °c to 25 °c), it is suggested that the 0.6-0.7 log unit currently expected interlaboratory reproducibility can be reduced to near 0.15. it is the aim of the review that the improvements in data quality would lead to better predictions of drug solubility using in silico methods. keywords shake-flask solubility; intrinsic solubility; henderson-hasselbalch equation; aggregates; oligomers; micelles. introduction since the mid-1990s there has been a heightened effort in drug discovery to predict drug-relevant aqueous solubility, described in at least a hundred publications (e.g., huuskonen et al. [1,2]; abraham and le [3]; jorgensen and duffy [4,5]; bergström et al. [6]; hou et al. [7]; delaney [8]; dearden [9]; balakin et al. [10]; taskinen and norinder [11]; jain and yalkowsky [12]; shayanfar and jouyban [13]; wang and hou [14]; elder and holm [15]; mcdonagh et al. [16]). the typical errors in drug solubility prediction are 0.7 – 1.0 log unit, and for low-soluble compounds, errors are considerably greater than a log unit (jorgensen and duffy [5]; palmer and mitchell [17]). according to faller and ertl [18], whose sentiment may be broadly shared, “no really satisfactory approach to (drug) solubility prediction is available yet,” in spite of the large number of prediction studies. two large aqueous solubility databases have been the sources for many of the in silico studies. first, the 1608-page handbook of aqueous solubility data, second edition (yalkowsky et al. [19]) contains over http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:alex@in-adme.com admet & dmpk 3(2) (2015) 84-109 equilibrium solubility measurements improvements doi: 10.5599/admet.3.2.193 85 18,000 solubility values covering 4661 molecules, collected from many chemical classes, including pharmaceuticals. second, the physprop database (howard and meylan [20,21]) from syracuse research corp. (http://www.srcinc.com/) had accumulated over 6000 measured solubility values, with substantial coverage of agrochemicals and potential environmental pollutants. in the two curated compilations, the ph of saturated solutions had not been reported for most of the ionizable compounds. in addition, smaller databases of crystalline drug-like molecules have been published (analytical profiles of drug substances [22], mcfarland et al. [23]; rytting et al. [24]; bergström et al. [25]; faller and ertl [18]; llinàs et al. [26]; hopfinger et al. [27]). these smaller databases consist largely of intrinsic solubility values, s0 – i.e., the solubility of the neutral species. the most cited challenge to good prediction has been the difficulty to access enough high-quality, drugrelevant, and sufficiently-diverse solubility data that adequately cover the chemical space of drugs and hopefully that of research compounds. in this context, the focus on drug molecules is important, since it is consistently shown that the best prediction models are devised from training sets most similar to the test sets (walters [28]). however, palmer and mitchell [17] posed a contrarian view that the challenge to good prediction might rest in the deficiency of current qspr methods. measuring high quality data is expensive and analytical-resource consuming. even with great costing, quality is not ensured when results are determined from poorly-designed assays. selecting training set molecules which are ionizable, but ignoring the effect of ph, can mitigate accurate prediction. drawing on data from a range of temperatures without adjustments also can be problematic. there are other factors that affect data quality. to address many of the above concerns, this review draws on our past experiences and also on the examination of over 800 publications to suggest ways (a) to improve the quality of future measurement of equilibrium solubility and (b) to normalize existing data for ph and temperature effects to extract intrinsic solubility values with improved accuracy. we focus on the shake-flask solubility measurement as a function of ph which is still the “gold standard” methodology in the minds of most experimentalists. also, two potentiometric methods are considered. the characterization of solid forms (crystalline, amorphous, nanoparticle, etc.) and their impact on the measured solubility are beyond the scope of this review, although some aspects (i.e., solvate, polymorph, racemate effects) are noted. in part, this review serves as background preparation for the “panel of experts” solubility session at the 4th world conference on physico chemical methods in drug discovery and development (pcmddd-4) in croatia, 21-24 september 2015 (http://www.iapchem.org/page.php?page_id=56). methods briefly stated, in the course of gathering published aqueous solubility training data, 803 publications have been examined and a largely fresh intrinsic solubility database containing 4557 entries for 2413 compounds has been assembled. in addition to this, 666 publications reporting pka values have been processed, to add to the in-adme research wiki-pka database, now consisting of 2651 qualified entries (www.in-adme.com/wiki_pka.php). it has been nearly a full-time project at in-adme research since 2011. the collection of training data is nearing a state suitable to support a new solubility prediction effort, which is planned to be the subject of another publication. we are energized to improve the accuracy of the prediction of intrinsic solubility from 2d structure, particularly of sparingly-soluble (or practically-insoluble) ionizable drug-relevant molecules. preliminary random forest regression modeling (walters [28]) has been tried using the 193 descriptors calculated by the open-source chemoinformatics and machine learning http://www.srcinc.com/ http://www.iapchem.org/page.php?page_id=56 avdeef admet & dmpk 3(2) (2015) 84-109 86 rdkit library of programs (landrum et al. [29]; http://rdkit.readthedocs.org/en/latest/), combined with the abraham five solvation descriptors (abraham et al. [30]), and lang and bradley [31] predicted melting points in the qsardb open repository of data and prediction tools (http://qsardb.org/repository/handle/10967/104). the most sensitive molecular descriptors will be further tested in partial least squares models, to better understand the impact of specific descriptors on the predicted solubility. in the database construction, intrinsic solubility (s0) values were mostly derived from (a) sw aqueous solubility determined in distilled water, often with ph not reported, (b) sph single-ph buffer values, and (c) multiple-ph buffer sph values (log s vs. ph). in cases (a) and (b), the henderson-hasselbalch equation (discussed below) was assumed to be valid. the pdisol-x program (in-adme research; www.in-adme.com/pdisol_x.html) was used to determine the s0 values from the reported sw and sph data, as described in völgyi et al. [32] and avdeef [33,34]. data the 4557 set of measured solubility values used to deduce the intrinsic solubility (s0) database were collected from four secondary sources (57 %) and the rest from primary sources (43 %):  physprop database (sep. 1999 version: 6356 measured water solubility, sw) [20,21]: 1327 shake-flask values were selected of molecules which were not appreciably ionized at ph~7. filters used to exclude compounds were: (a) melting point < 40 °c, (b) log sw < -8 or > 0, (c) surfactants or compounds with long aliphatic chains, (d) multi-ring aromatic hydrocarbons, (e) peroxides, (f) carboxylic acids, (g) salts/complexes with chloride, bromide, iodide, sulfate, phosphate, sodium, potassium, lithium, calcium, magnesium, zinc, copper, arsenic, mercury, lead, antimony, silver, silicon, tin, etc., (h) dyes or names containing color, and (i) herbicides, pesticides, insecticides, rodenticides, and acaricides (as indicated by “tags” at the royal society of chemistry chemspider website: http://www.chemspider.com/). of the selected compounds, the sw values of 1210 nonionizable/nonionized molecules were taken to be s0. the remaining 117 compounds were processed by pdisol-x to calculate s0 and phsat (ph of saturated aqueous solution) from the given sw and pka.  handbook of aqueous solubility data [19]: 1130 sw data were selected, with 776 values subjected to pdisol-x analysis to determine s0 values. a few of the same preceding filters were used in the selections.  analytical profiles of drug substances [22]: all 39 volumes of the series of monographs were searched for quantitative solubility data. monographs on 155 molecules were found to be useful. most of the reported solubility values of ionizable molecules were measured in pure water with unspecified saturation ph. for those ionizable molecules which were not salts, the intrinsic values were calculated by pdisol-x. unfortunately, the solubility accompanied scant experimental detail (e.g., temperature not always reported), and many entries were referenced as ‘personal communication.’ but there are several high-quality log s ph data sets in the monographs.  miscellaneous publications (shake-flask data, some from secondary sources): 852 s0 values (taken as reported).  cheqsol s0 data (potentiometric): 201 values for 151 molecules collected from several publications (stuart and box [35]; sköld et al. [36]; llinàs et al. [26,37]; box et al. [38]; hopfinger et al. [27]; narasimham et al. [39]; hsieh et al. [40]; comer et al. [41]; palmer and mitchell [17]; schönherr et al. [42]). http://rdkit.readthedocs.org/en/latest/ http://qsardb.org/repository/handle/10967/104 http://www.chemspider.com/ admet & dmpk 3(2) (2015) 84-109 equilibrium solubility measurements improvements doi: 10.5599/admet.3.2.193 87  dissolution template titration (dtt) s0 data (potentiometric): 75 published values were collected (avdeef [43,46]; avdeef et al. [44]; avdeef and berger [45]; faller and wohnsland [47]; bergström et al. [48]; fioritto et al. [49]; ottaviani et al. [50]).  rytting et al. [24] free-base and -acid (no salts used) shake-flask sw: solubility of 122 molecules were gathered, all measured in one laboratory, with s0 calculated by pdisol-x.  shake-flask measurements in two or more ph buffers (primary sources): 697 molecules with s-ph data were analyzed by pdisol-x to determine s0. figure 1a shows the frequency distribution of the multi-ph data. there were 101 studies (14 %) with 2 ph/assay and 164 studies (24 %) with 3-5 ph/assay. eight studies were reported, each with more than 40 ph points defining the log s-ph profile. the median in the set of 697 compounds was seven s-ph points per molecule. most of the published data were presented graphically as log s vs. ph plots. the process of digitizing data from plots introduced some small error. it was not possible to digitize the graphs entirely if the molecules were low soluble and the plots were in non-logarithmic s units vs. ph. a large fraction of the primary source data originated from five journals (in rank order): int. j. pharm., j. pharm. sci., pharm. res., j. chem. eng. data, and eur. j. pharm. sci. it is primarily from this group of 697 “gold standard” shake-flask multi-ph measurements that many of the suggestions below were formulated. the assembled 4557 intrinsic solubility set ranges in log s0 from -11.6 to +1.8 (log molarity). figure 1b shows a frequency distribution for the set. about 40 % of the compounds have log s0 between -6 and -3, the typical range of values for research compounds (faller and ertl [18]; walters [28]). the median log s0 of the distribution is -2.8. about 5 % of the molecules have log s0 < -6. the least-soluble molecules (log s0 < -8) in the set are amiodarone, cosalane, halofantrine, clofazimine, itraconazole, quinclorac, probucol, brodifacoum, epristeride, silafluofen, carbenoxolone, tamoxifen, fluotrimazole, moteretinide, esfenvalerate etofenprox, etretinate, and npc-1161c (which includes four agro-chemicals). the most soluble (log s0 > 0) substances are amino acids, simple carboxylic acids, and carbohydrates. the compounds in the 4557-set are solids at room temperature, with propofol and nitroglycerin at the border line with the melting points 19 and 14 °c, respectively. the shake-flask values were mainly clustered around room and physiological temperatures: 23 ± 3 °c (78 %) and 37 ± 5 °c (22 %). figure 1c shows the temperature frequency distribution. surprisingly, a large number of sources don’t state the temperature used in the assay. contextually, room temperature is a reasonable guess in most instances. other publications cite “room temperature.” where the precise temperature was not reported, it was taken to be 23 °c in the assembled database. figure 2 shows plots of six high-quality examples of multi-ph solubility data analyzed by pdisol-x. most of the compounds can be represented by simple henderson-hasselbalch equations (cf., below). the data from verapamil was best fit with a tricationic aggregate below ph 7. above ph 9 the free-base form of verapamil may participate in the formation of uncharged water-soluble aggregates or micelle-like structures, consistent with the discussion by surakitbanharn et al. [51]. avdeef admet & dmpk 3(2) (2015) 84-109 88 figure 1. (a) distribution of the number of multiple-ph shake-flask measurements per assay. (b) intrinsic solubility, log s0, distribution, as counts in 0.1 log unit intervals. (c) distribution of assay temperature in the 4557-set. on the average, log s0 increases by 0.134 log unit as temperature goes from 25 to 37 °c. admet & dmpk 3(2) (2015) 84-109 equilibrium solubility measurements improvements doi: 10.5599/admet.3.2.193 89 figure 2. examples of log s – ph profiles from well-designed assays for an acid (flufenamic acid), three bases (oxycodone, verapamil, and indinavir), and two ampholytes (dorzolamide and tyrosine). note that the tyrosine is based on measurements done in 1924. the solid (red) curves represent the best-fit pdisol-x model which rationalizes the measured solubility values (filled circle symbols). the dotted curves depict the ph interval where the compounds were fully-dissolved. the dashed curves were calculated from the hendersonhasselbalch equations. flufenamic acid, oxycodone, and dorzolamide were studied at concentrations high enough to effect salt precipitation. in all cases but verapamil, it was possible to refine the pka values. verapamil and flufenamic acid show some evidence of formation of aggregates. avdeef admet & dmpk 3(2) (2015) 84-109 90 results and discussion solubility units and conversion issues solubility measurements have been reported in a multiplicity of concentration units: mol/l (molarity, m), mm, μm, mol/kg (molality, m), mole fraction, mass fraction, weight/ml (e.g., mg/ml, μg/ml, ng/ml, pg/ml …), mg/100 ml, mg/dl, %w/v, g%ml, mg/ml%, mg%, “1 in 40 of water,” “soluble in 2 parts of water,” “3% soluble in water,” units of iu/ml, etc. mole fraction, mass fraction, and molality units are almost always used when solubility is determined over a wide range of temperatures, since the units do not depend on the density of the solutions. it is not clear why units such as mg% would be preferred. in certain publications such units are not explicitly defined. usually (but not always), mol fractions are precisely defined in the publication, so that conversion to molarity is straight forward if the solution density is known (which is slightly higher than that of pure water). the same cannot be said of the very frequently used weight/ml “practical” units. in the clearly presented papers, the equivalent molecular weight to use to convert the practical values to molarity is stated (e.g., “concentration is expressed as free base equivalent”) or is evident, but too often, the reader has to guess what the authors intended. for a given molecule, often the uncertain units became evident when compared to results from different publications. as different units are in common usage, consequently a useful sense of comparisons between different studies is a challenge. on a practical note, it is all too easy to make a mistake in converting the units to the preferred molarity scale of the database. it could be argued that solubility should be presented in log units (preferably based on molarity), since (a) direct values span over 12 orders of magnitude and cannot be accurately depicted in s-ph plots at the low end of the scale, and (b) since errors in log values do not depend on the magnitude of the log solubility (whereas they do when direct units are considered). for example, it should be possible to claim that shake-flask log s values have an average reproducibility of 0.1 log unit for a given molecule between values reported in different publications; it makes no obvious sense to present an average reproducibility value of s measurements that can span so many orders of magnitude. in the 4557-set database, values reported in molality units were noted, but not converted to those in molarity (by correcting for density), since the differences are small near the temperature range of interest (20-40 °c), and the since the actual solution density is seldom reported. impact of accuracy in the type of data used for prediction training sets most studies aimed to predict s0 values, whether explicitly stated or not. however, most studies used sw values in the training sets, which resulted in increased errors for low soluble ionizable molecules. the exception is with nonionizable molecules, where water and intrinsic solubility are the same: sw = s0. the majority of the aqueous solubility values, sw, were measured by adding excess solid (ideally, the free-acid or free-base, not the salt) to distilled water. frequently, ph of the saturated solution, phsat, was not reported (and probably not measured). when a moderately soluble weak acid or base is added to distilled water, the aqueous ph is altered by the ionizing molecule, in the direction where the molecule remains largely in the uncharged form: sw ≈ s0, provided the compound is added as a pure free acid or free base. for compounds added as salts, it is frequently not possible to deduce s0 and phsat from just sw and pka, since the total added amount of compound can affect the disposition of the saturated solution. for example, if not enough salt form of the admet & dmpk 3(2) (2015) 84-109 equilibrium solubility measurements improvements doi: 10.5599/admet.3.2.193 91 drug is added, the solid disproportionates to the free acid/base in the saturated solution, with phsat depending on the weight of salt added. if the ionizable compound is practically insoluble, then the value of sw can be quite different from s0, since not enough of the compound dissolves to alter the ph in the direction of maintaining a nearly uncharged molecule. in such cases, it is possible to calculate the phsat, as well as s0, provided the pka is known and that the henderson-hasselbalch equation accurately describes the solubility-ph relationship for a one-pka molecule (‘±’ in eq. 1: acid when ‘+’ and base when ‘-‘): log s = log s0 + log { 10 ±(ph–pka) + 1 } (1) more complex henderson-hasselbalch equations for ampholytes and multiprotic acids and bases have been tabulated elsewhere [52,57]. a feature of pdisol-x allows the calculation of phsat: when the ph is not reported with the sw value, a phsat of 7.0 is initially assumed, and the regular mass balance regression analysis is performed [32]. (fixed values of carbon dioxide may be included.) the ph titrant volume is calculated. if the volume is non-zero, the value of phsat is adjusted iteratively by the regression procedure until the calculated volume is driven to zero. at the same time, s0 is refined. variance increase due to pooling of training set solubility data determined at different temperatures it appears that many of the published prediction studies have drawn data from two large secondary sources, as noted above. these two compilations have data collected mostly in the range 20 – 40 °c, roughly in a bimodal distribution (cf., fig. 1c). some prediction papers state that only 25 °c data were used. many prediction papers are less clear, and it might be that some pooling from different temperatures takes place, which would contribute to increased variance in the experimental training set data. if the temperature dependence of solubility could be predicted, then the number of available training set values in the former case would increase, and in the latter case the variance in the measured data would decrease. that is, solubility values could be normalized to a single reference temperature, e.g., 25 °c. as far as we are aware, there are no publications where temperature dependence of solubility is predicted solely from 2d structures. we have collected a database of 626 values of enthalpies of solution, with 77 determined by calorimetric methods, and the rest by solubility methods (van’t hoff slopes from log s vs. 1/t plots). the calorimetric data are producing the most promising results, with r 2 > 0.5 using just the abraham solvation descriptors. this procedure will be described in a separate publication. as a preliminary observation, on the average, measured log s0 values increase by 0.13 log unit (cf., fig. 1c), as the temperature is raised from 25 to 37 °c. in cases where the training set data are pooled from multiple temperatures, the variance can be expected to increase by about 0.13 log unit. this can be avoided if the data mining process were to convert measurements to a single reference temperature. limitations of intrinsic solubility, s0, determined from a single measurement of sw when measured compounds contain protogenic impurities, the phsat may be affected, which could lead to a change in the measured sw. for this reason, it is highly recommended that phsat be measured and not just calculated. otherwise, the conversion of sw to s0 may be inaccurate. avdeef admet & dmpk 3(2) (2015) 84-109 92 abraham and le [3] discussed the relationship between the measured sw and the underlying s0 for ionizable molecules, and identified under which circumstances large errors could result for sw used (in place of the calculated s0) as the training set values. the authors derived useful plots of s0/sw vs. log sw for acids and bases over a range of pka values. for example, sw ≈ s0 for acids with pka 5 and sw > 0.001 m, or with pka 3 and sw > 0.1 m. with bases, sw ≈ s0 for pka 10 with sw > 0.01 m, or pka 8 with sw > 0.0001 m. there are further considerations. when the aqueous solubility of practically insoluble free bases (pka > 9) are measured in distilled water, the ph is only slightly affected by the extent to which the base dissolves. most notably, the observed ph often is regulated by a much stronger buffer present in water: ambient dissolved carbon dioxide. this is often overlooked. the ph of the drug-saturated solution can vary from 610, depending on how much co2 is dissolved in water and how insoluble the basic compound is. for example, rytting et al. [24] reported log sw = -4.87 (molarity) for terfenadine dissolved in water. if co2 content in solution were ignored, the calculated log s0 = -5.7 and phsat = 9.2. however, if [co2] = 20 μm (a common ambient level), the calculated log s0 = -8.3 and phsat = 6.6. the error in determining s0 of molecules like terfenadine in distilled water is expected to be enormous (as much as 3 log units), since it is very difficult to eliminate co2 entirely simply by bubbling an inert gas through solution. in fact, any protogenic drug impurity in solution under the circumstances would lead to large uncertainties in the intrinsic solubility of the drug substance of interest. the simple remedy might be to measure the ph of the saturated solution. however, measuring the ph accurately when the solution is essentially unbuffered is problematic, due to the effects of uncontrolled electrode junction potentials and other factors [52]. added buffer improves accuracy when measuring aqueous solubility (sph) at a single ph the above terfenadine water solubility example illustrates that high errors can result when the ph of the saturated solution is not known or whose measurement is problematic. the remedy is to buffer the solution (but not excessively) and actually measure the ph when the saturated solution reaches equilibrium. however, when only a single aqueous solubility at a known ph, sph, is measured, it is still necessary to assume that the henderson-hasselbalch equation accurately describes the log s ph profile, in order to calculate the s0, given an accurate independently-determined pka. measurement of aqueous solubility at several buffered ph, below and above the pka, for best accuracy by measuring solubility at several different values of ph (cf., fig. 2) below and above the pka of an ionizable molecule, it’s possible to overcome several of the above sources of error. the log s data as a function of ph can be analyzed, to determine the value of s0. the validity of the henderson-hasselbalch equation can be easily tested in such analysis, using mass-balance based solubility simulation software (e.g., pdisol-x). a universal buffer mixture (e.g., whose ph is nearly linearly controlled by added aliquots of standardized naoh solution) or several independent buffers may be used, but the ph still needs to be measured when the saturated solution reaches equilibrium. it can be misleading to assume that the buffer ph will remain unchanged in the course of the drug dissolution. nor is it a good idea to use excessively high concentrations of buffers, since drug-buffer complexes and precipitates may form, affecting the measured solubility (shoghi et al. [53]). in simply-behaving systems, even the pka can be determined from the log s – ph data (zimmermann [54]; cf., fig. 2), but this is not generally recommended [32-34]. it is far more reliable to use purposedesigned pka measurement techniques (e.g., potentiometric, spectrophotometric, or capillary admet & dmpk 3(2) (2015) 84-109 equilibrium solubility measurements improvements doi: 10.5599/admet.3.2.193 93 electrophoretic), under conditions where ionic strength is well controlled and complications due to sample complexation, self-aggregation, or precipitation are avoided. calibration of ph electrode glass ph electrodes are not all equal. there is no single standard method for calibrating electrodes. in solubility publications, the electrode calibration is virtually never described. the make of the electrode is hardly ever stated. in contrast, researchers who determine pka values using commercial instruments are well acquainted with the routinely-used four-parameter procedure (avdeef and bucher [55]). measurements of ph are sensitive to ionic strength effects, especially in poorly buffered solutions. in measurements of salt solubility, it is not uncommon to have the ionic strength reach 1-3 m. methods for electrode calibration to address these challenging conditions have been discussed in the literature (völgyi et al. [32]; wang et al. [56]). errors in the ph scale can easily exceed 0.1-0.2 units in the mid-scale region. for measurements at ph < 1 or ph > 12, the ph scale can be in error by as much as 0.5 log unit. the ph electrode practices routinely used in modern pka methodology would benefit solubility measurement and lead to improved quality of results. separating solid from saturated solution an advantage of the potentiometric methods is that they do not require phase separation. also, certain fiber optic probe methods can determine concentrations in turbid solutions (cf., appendix). however, most traditional methods require some sort of phase separation. of the 4557 entries in database, there were only 583 indications of the type of phase separation used. it’s surprising that not all primary sources identified the means of separation, although one might surmise that most of these employed filtration. none of the secondary sources lists such detail. of the indicated values, 44 % used some means of filtration, 14 % used centrifugation, and 9 % used sedimentation solely. the majority of those using filtration seemed to be aware of the problem of filter adsorption and discarded the initial filtrate solution or performed “double filtration” [52,57]. a few of those using filtration did not take heed of the chen et al. [58] recommendations against using nylon filters. a solid drug substance in equilibrium with its saturated solution is dynamically dissolving and precipitating at equal rates. since that equilibrium is maintained by the presence of the solid, the act of separation by filtration disrupts the process, to the extent that the solution concentrations may be altered by reactions of the substance with the container and filter surfaces. thus, avoiding filtration, if possible, can be recommended. however, it is such a popular procedure, that most non-potentiometric protocols use it. especially noteworthy, 13 % of separations were done by centrifugation, followed by filtration. this can be problematic with low soluble compounds, for the above reasons. if a saturated solution is devoid of excess solid (which centrifugation removes), then passing the weakly-poised (i.e., nearly subsaturated) solution through a filter can produce a significantly subsaturated solution. vessel surface adsorption may contribute to a further lowering of the amount of dissolved sample. in such a combined procedure, solubility may be significantly underestimated. the above practice is best avoided, especially with low soluble compounds. avdeef admet & dmpk 3(2) (2015) 84-109 94 figure 3. comparison of traditional shake-flask measurements to (a) the dissolution template titration (dtt) potentiometric method, (b) the cheqsol potentiometric method, and (c) the shake-flask method done in one laboratory using the same protocol. centrifugation is needed in cases of samples that form (a) opalescent saturated solutions (e.g., colloid systems), or (b) stable suspensions that do not sediment easily, or (c) small agglomerate particulates that can pass through commonly-used filtration membranes. it was shown by mcgovern et al. [59] that certain types of low soluble research compounds, dubbed “promiscuous inhibitors,” form agglomerates of the order of 0.1 μm in size. these can pass through most filters used for phase separation, leading to overestimates of the solubility. if such compounds are suspected, it may be a useful first to filter the suspension and then to ultracentrifuge it (in that order). equilibration time and identifying the final form of the solid in the saturated solution most of the equilibrium protocols reviewed here strive to determine the solubility of the most stable form of the solid, most likely crystalline (and sometimes hydrated). assay times are selected to be long enough to ensure that the measured solubility is no longer changing and that it has reached its lowest value. it is a common practice in pharmaceutical research to adopt 24 h for the equilibration time (“one-shoe-size-fits-all” approach). in most cases, 24 h is enough, but certain practically-insoluble compounds, which consequently have very slow dissolution rates may require much longer times [34]. loftsson et al. [60] allowed 3-7 days for equilibration to be reached for a variety of low soluble compounds. other researchers used 2-10 days equilibration times for practically-insoluble anticancer drugs (venkatesh et al. [61]; he et al. [62]). fini et al. [63] allowed diclofenac suspensions to incubate for 30 days. undeniably, the stability of the compound needs to be verified when such long times are used. all too often, at the end of equilibration, the actual form of the solid is not characterized in reported studies. often the monoor dihydrate is more stable (i.e., less soluble) than the anhydrous form of the solid. sometimes, multiple-hydrate forms of the solid precipitate in the crystalline form. but this is not always characterized in published solubility-ph studies. admet & dmpk 3(2) (2015) 84-109 equilibrium solubility measurements improvements doi: 10.5599/admet.3.2.193 95 solvates aside, different polymorphs of the same stoichiometry may form, depending on the specific assay protocol. pudipeddi and serajuddin [64] examined the differences between the solubility values of polymorphs of various substances and found that 0.3 log unit was a typical spread. a few substances showed differences as high as a log unit. dl-racemates of optically active molecules can have solubilities lower than those of pure enantiomers. for example, at 25 °c, dl-tartaric acid has log s0 = +0.18, but the l-form has +0.58, an increase of 0.40 log unit. fourteen amino acids were found with solubility values reported for both the dland lforms. on the average the log s0 of the dlenantiomer is 0.15 lower (more insoluble) than that of the lenantiomer. the least-soluble amino acids (tryptophan, tyrosine, and cystine) show the largest differences in the solubility between the dland the l-forms (i.e., dl-forms less soluble). very few chiral drugs have both dland l-/d form solubilities reported. as an example, dl-atropine has log s0 = -2.20 ± 0.03 (n=3); l-atropine has -1.84 ± 0.10 (n=2), an increase of 0.36 log unit. when solubility of chiral molecules is reported without mention of the enantiomeric state, racemic mixture is probably implied. as mentioned above, in many solubility studies, the solid added to solution is in the salt form (chloride, maleate, tartrate, sodium, etc.). the solid that is isolated at the end of equilibration may not be that of the original salt form, but rather of the free-acid/-base. this can lead to ambiguity in ascribing the reported solubility (anderson and flora [65]). consequently, crystal-form uncertainty translates to higher variances in prediction studies, where many literature values are pooled into the same training set. molecular descriptors based on 2d considerations may be difficult to assign to test compounds which may exhibit significant polymorphism/solvatomorphism. reproducibility of experimental solubility data interlaboratory reproducibility can be assessed for a few model compounds which are repeatedly used in solubility studies. for example, reported solubility of anthracene from 17 different laboratories indicate standard deviation, sd = 0.17 log unit (kishi and hashimoto [66]). additional examples for phenytoin, sd = 0.15 (n=16), flurbiprofen, sd = 0.19 (n=16), and diclofenac, sd = 0.20 (n=20) show similar reproducibility (this work). perhaps, these can be the expected interlaboratory reproducibility values for such relatively simple molecules. however, ionizable molecules with very low solubility might be expected to have poorer reproducibility. particularly poignant examples are indicated (this work) by indomethacin (which can be unstable at elevated ph), sd = 0.79 (n=12), and terfenadine (whose unbuffered sw is strongly affected by ambient co2), sd = 1.48 (n=12). the handbook of aqueous solubility data [19] is a massive source of values to estimate interlaboratory reproducibility, although a systematic analysis of the compiled data has not been published, as far as we are aware. for 411 compounds with reported replicate measurements, katritzky et al. [67] found average sd = 0.58. comparable values have been reported for research compounds. according to taskinen and norinder [11], an astrazeneca in-house database of solubility measurements of different batches of the same compound typically showed reproducibility of 0.49 log units. higher uncertainties had been suggested (jorgensen and duffy [5]; palmer and mitchell [17]). however, there are many compounds determined in different laboratories that indicate reproducibility much less than the above sd values. for example, barbital, hydrochlorothiazide, and lidocaine show sd = 0.08 log unit for n = 10-14 each; for hydrocortisone sd = 0.07 (n=11); for testosterone sd = 0.06 (n = 10), and for acetanilide, aminopyrine, threonine, alanine, 5-fluorouracil, budesonide, minoxidil, fluconazole, avdeef admet & dmpk 3(2) (2015) 84-109 96 corticosterone, phenobarbital, lidocaine, hydroflumethiazide, acetaminophen, serine, glycine, atenolol, sd = 0.01 – 0.05 (n= 4-11). in this review of shake-flask solubility measurement, it was decided to include data from two potentiometric methods (dtt and cheqsol – c.f., appendix). it was thus of interest to estimate how concordant the dtt and cheqsol data are with shake-flask (sf) results for molecules evaluated by the different approaches from different laboratories. figure 3a shows a correlation plot of log s0 (sf) vs. log s0 (dtt). although some replicate sf measurements possess large variances, the two sets of log s0 values correlated well: log s0 sf = -0.02 + 1.01 log s0 dtt , r 2 = 0.95, s = 0.36, f = 1093, n = 62. based on sf replicates, the average sd = 0.27, which is not very different for the correlation standard deviation, s = 0.36. the differences between the two log s0 sets were ≤ 0.25 for 66 % of the molecules. a weighted linear regression, using the individual inverse variances of the sf data as weights did not appreciable change the statistics of the sf dtt comparison. figure 3b shows a plot comparing sf to cheqsol measurements, indicating that the two sets appear comparable: log s0 sf = -0.13 + 1.00 log s0 cheqsol , r 2 = 0.90, s = 0.52, f = 1094, n = 125. there is a slight bias: s0 cheqsol /s0 sf = 1.35 (based on the intercept in the fit). the average sd = 0.25, based on sf replicates, which is notably less than the s = 0.52 from the correlation plot. the differences between the two sets of log s0 values were ≤ 0.25 for 61 % of the molecules. using sf-based variances in a weighted linear regression analysis produced: log s0 sf = -0.08 + 0.99 log s0 cheqsol , with goodness-of-fit = 2.4. it appears that the bias in the unweighted regression is influenced, in part, by the high variance of some of the sf values. to put the above two comparisons into perspective, we selected a one-source set of sf distilled-water sw measurements (with phsat and s0 calculated with pdisol-x), performed in the same laboratory using the same assay protocol for 122 free-acids/-bases (rytting et al. [24]). for 117 of the molecules, there were reported sf values from other laboratories. figure 3c shows the sf-to-sf correlation plot. as expected, the two sets of data compared reasonably well: log s0 sf = -0.02 + 1.01 log s0 rytting-sf , r 2 = 0.93, s = 0.39, f = 1637, n = 117. the average sd = 0.17, based on non-rytting sf replicates. (it is noteworthy that – with the exception of clofazimine – the rytting set of molecules were relatively simple.) the differences between log s0 values from the two sf sources were ≤ 0.25 log unit for 70 % of the molecules. from the correlation plots in figure 3, one may conclude that diverse sets of molecules show average reproducibility in the range 0.17 to 0.39 log unit by the “gold standard” shake-flask method. the highsolubility end of the range may be better determined, whereas the low-solubility end may have higher measurement errors. the expected reproducibility of the dtt method, s = 0.36, appears to match that of the sf method, while the cheqsol method shows a slightly higher value, s = 0.52 (similar to the katritzky et al. [67] estimate). in the entire 4557-set, there were 786 replicate log s0 values from different laboratories, where two or more values could be averaged. the mean sd of all the averaged values is 0.19 log unit. it is expected that the mean value of 0.19 could be further reduced, to near 0.15 log unit, if corrections were applied to normalize log s0 values to a common temperature (e.g., 25 °c). summary of the factors affecting reproducibility some of the factors affecting reproducibility of equilibrium solubility measurement discussed above may be summarized in the list:  incomplete dissolution over the equilibration time (e.g., latent supersatured conditions) admet & dmpk 3(2) (2015) 84-109 equilibrium solubility measurements improvements doi: 10.5599/admet.3.2.193 97  inappropriate phase separation (e.g., first centrifuging a saturated solution, then filtering the supernatant)  adsorption to the filter or assay vial surfaces  poor wettability  formations of drug aggregates/oligomers (dimers, trimers, …), micelles, and drug-buffer complexes  formation of ion-pair between ionic strength adjustor (e.g., nacl, kcl, etc.) and charged form of drug  polymorphs, hydrates, solvates, amorphous forms  stereoisomers (dl-, d-, l-)  cis-/transisomers  not using buffers with low-soluble ionizable drugs  not taking into account the effect of ambient co2 on the water solubility of low-soluble bases  using unnecessarily high buffer concentrations, possibly effecting drug-buffer complexation  not measuring the final ph of the equilibrated saturated solution of ionizable drugs  inadequate ph electrode calibration procedure at low and high ph and in salt solubility studies  effect of impurities, especially those which are ionizable  “promiscuous inhibitor” particles passing through filter  compound instability at the extremes of ph or over long saturation times  not sufficiently sensitive analytical methods used to determine very low drug concentration from the perspective of predicting solubility, the impact of many of the above factors can be minimized by employing good experimental and data analysis practices. however, some of the factors leading to variability in measured solubility, such as the formation of polymorphs, hydrates, solvates, amorphous solids, and the impact of stereoisomers will be harder to deal with. further in silico insights will be needed to address these challenges. recommended procedures for more accurate solubility-ph measurements solubility units, tabulation of results and the use of logarithmic plots following the format used in the handbook of aqueous solubility data (yalkowsky et al. [19]), it is recommended that solubility be tabulated both in molarity and in practical (mg/ml) units. standard deviations in the measured solubility (based on averaging three or more values) should be included in the table of values. additionally, a graphical display of log s vs. ph (but not s vs. ph) would be visually helpful. in the logarithmic forms, the plots can serve as templates [52,57], to recognize the presence of aggregates, incomplete equilibration, corrections for the presence of small quantities of dmso, etc. solubility methodology and the benefits of knowing the accurate pka the “gold standard” multiple-ph buffer shake-flask measurement is recommended for challenging ionizable molecules. other methods may also be satisfactory, provided that the henderson-hasselbalch relationship is independently validated. miniaturization can be recommended, as long as the protocols are rigorous and well validated. for ionizable molecules, the measurement of sw without stating the ph is not recommended. it is far better to measure sph values in well-qualified buffers (see below), at three or more ph values, bracketing the pka. avdeef admet & dmpk 3(2) (2015) 84-109 98 ionization constant the log s – ph data should be evaluated to estimate the value of the intrinsic solubility, s0. to do this, the independently-determined pka is needed. it is sometimes very inaccurate to use the solubility-ph data to determine the pka, because usually the required assumption is that the henderson-hasselbalch equation is valid, but it may not be so for some particular low soluble molecule, or when high concentrations are used to characterize salt solubility. it is recommended that methods specifically designed to determine the ionization constants of very poorly soluble molecules are used (e.g., state-of-the-art uv spectrophotometry, capillary electrophoresis, potentiometry). these methods are widely available and have been fine-tuned for the challenge. commercial pka instruments based on the above three technologies are generally well-supported by their manufacturers. temperature solubility is a function of temperature, so the assay temperature always needs to be reported. it is clear that “room temperature” can be different from laboratory to laboratory, and in some cases, seasonally variable. it is advisable to measure and record the actual temperature in the sample vessel during the equilibration period. or better yet, the measurement is performed in a thermostated vessel kept at 25/37 °c. ionic strength measured solubility can be affected by ionic strength (particularly when salt solubility is measured), so the ionic strength usually needs to be reported. not only do the ionic strength adjustor (nacl, kcl, etc.) and buffers contribute to the ionic strength, but so does the sample, especially when introduced in salt form. equilibration time and stirring protocol finely granulated crystals dissolve more quickly than large crystals, illustrating the surface area effect in the classical noyce-whitney equation (eq. 5 in the appendix). amorphous solids, which are generally more soluble than their crystalline counterparts, dissolve more quickly than crystals, illustrating the solubility effect – the rate of dissolution is proportional to the solubility (cf., noyce-whitney equation, eq. 5 in the appendix). other examples of this are: anhydrous crystals dissolve more quickly than hydrates; usually pure enantiomers dissolve more quickly than racemic mixtures. vigorous stirring during the equilibration period can hasten the rate of dissolution, allowing for equilibrium to be reached more quickly (cf., simple noyce-whitney equation). however, for particle less than 1-5 μm in diameter, stirring speed has little influence on the rate of dissolution. towards the end of the equilibration period, stopping (or slowing) the stirring and allowing the solid to sediment probably contributes to formation of better quality crystals. the 80 % stirring – 20 % sedimentation timing protocol describe by baka et al. [68] can be recommended. in regions of ph where the sample molecule is largely ionized, equilibration times as short as 1-6 hours may be adequate. but for practically insoluble molecules, in ph regions where the molecule is largely in its uncharged form, equilibration times of 24-72 h may be required to reach equilibrium. often, the conversion from less stable amorphous or anhydrous solids to more stable crystalline (often hydrate/solvate form) takes 12-24 h, and sometimes longer. in rigorous applications, the shake-flask method usually determines the equilibrium solubility of the most stable solid state form of the compound. equilibration times as long as several days have been used. admet & dmpk 3(2) (2015) 84-109 equilibrium solubility measurements improvements doi: 10.5599/admet.3.2.193 99 for example, terfenadine studies have used 3-4 d equilibration times. steroids have been allowed to incubate for 2 -12 d. probably 24 hours is a good average time, but it is a good idea to test longer times when measuring the equilibrium solubility of very poorly-soluble drugs: 48 h, or even 3 7 d. perhaps special cases like amiodarone (30 d for full therapeutic effect of oral dose) may warrant long equilibration times. ordinarily, such long times are not recommended. shortening the equilibration time the cheqsol method uses cycles of dissolution and re-precipitation by ph adjustment to shorten the time to reach equilibrium. presumably, the solid which re-precipitates in subsequent cycles is more active than the original crystalline material, perhaps being of small particle size (high surface area) and possibly amorphous. loftsson et al. [60] describes using temperature spiking cycles and seed crystal to hasten the equilibration period. the facilitated dissolution method of higuchi, described in the appendix can be used to speed up the rate of equilibration. the method is fundamentally rigorous and deserved much more attention than it has received. a comprehensive validation study would be welcomed. solution composition it is a good idea to keep the assay solutions simple and to define all components precisely! when studying salt solubility, it is particularly helpful (and perhaps necessary) to specify the actual weight of sample in each vial. it is unhelpful to see statements such as “excess solid was added.” it is not enough to state that “water was used” as the solvent. was there any added ionic strength adjustor (e.g., 0.15 m nacl, etc.)? was carbon dioxide purged out? low soluble bases such as terfenadine can indicate solubility over several magnitudes due to the effect of ambient co2 in unbuffered solutions. ph measurement using glass electrodes it is not a good practice to assume that the assay buffer ph is not altered by the addition of sample. it is highly recommended that the final ph of the saturated solution be carefully measured, using a properly standardized electrodes. particular attention should be given to the ph electrode calibration, especially when extreme ph is measured (ph < 1 or ph > 12) or when the ionic strength reaches high values, as in the case of salt solubility measurement. it is recommended that research-grade combination ph electrodes be used, with adequately described calibration procedure. ideally, the four-parameter electrode calibration is recommended [55]. those using the sirius titrators will recognize the procedure as “four-plus” method, and those using the pion titrators will recognize it as the “abc” method. simpler procedures may be satisfactory, as long as they are described precisely. the reported ph should be evidently either on the “operational” or the “concentration” scale; the “mixed” scale is not recommended [52]. a strong case can be made that general solubility equilibrium quotients are best formulated in the concentration scale (rather than activity), with ph electrodes standardized in a constant ionic medium, e.g., 0.15 m nacl [52]. this may be especially important when salt solubility measurements are performed. avdeef admet & dmpk 3(2) (2015) 84-109 100 how was ph adjusted (e.g., 0.1 m hcl, 1 m naoh)? statements like: “concentrated h3po4 was added drop-wise to adjust the ph” are not clear enough. it is far better to state the precise concentration of the acid used. record the precise volume of ph adjustor solution added. if one wishes to use ionizable titrants, such as phosphoric acid, acetic acid, maleic acid, tartaric acid, etc., it may be useful to add enough of the titrant so the starting ph is as low as desired in the assay. that way, the pre-acidified solution ph can be solely regulated by the amount of base titrant added. there is no real need to have a separate burette dedicated to the ionizable acid titrant. result computation will be greatly simplified. buffers sufficient buffering is needed so that reliable ph can be measured and that the titrant can be dispensed precisely enough to adjust the solution to the desired ph. ordinarily, relatively low buffer concentrations (e.g., 5-10 mm) can be recommended. the sample itself may be an adequate buffer. it is not beneficial to overload the assay protocol with added complications arising from possible buffer-drug interactions. the use of universal buffers (e.g., britton-robinson, sorensen, mcilvaine) can be very useful, but it may also be a source of unintended/unnecessary complications when it comes to processing log s – ph data, due to drug-buffer interactions, particularly when salt solubility is measured. some buffer components may cause difficulties in uv measurement. high concentration phosphate buffer is not recommended; however drug regulatory agencies may prescribe it. it is inadequate to state that “buffers were used”. precisely which buffer compounds were used? which salt forms of buffers were used? at what concentrations were they prepared? this information will be needed to calculate the total ionic strength. it may be a good idea to use zwitterionic buffers generally, or acid buffers with weak acid samples and basic buffers with base samples. phosphate buffers interact with weak base drugs. so, caution is needed when phosphate buffers are used with low-soluble basic drugs. purpose designed assays can better address the effect of phosphate anions on positively-charged sample molecules. separating solid from saturated solution filtration can be recommended, using hydrophilic filters, typically 0.2-0.45 μm pore sizes. chen et al. [58] found the hydrophilic pvdf (polyvinylidene fluoride) and pes (polyether sulfone) filters performing the best, and nylon the worst, in terms of excessive compound adsorption. it is useful to discard the first 1025 % filtered solution, to allow filters and surfaces to be saturated with adsorbed compound. the filtration step should not be rushed. if “promiscuous inhibitors” (mcgovern et al. [59]) are suspected, then the filtered solutions may be further subjected to ultracentrifugation. it is highly inadvisable to centrifuge first and then filter, for reasons discussed in the review. conclusions over 800 publications describing equilibrium solubility measurement of sparingly-soluble ionizable drug-like molecules by the shake-flask and related methods were examined. many factors affecting the quality of the measurement were identified, and a number of suggestions were offered to improve the methodology. some of the suggestions focused on improving methods for future measurements, and some referred to ways of extracting more reliable information from existing measurements. by normalizing data for ph (by using intrinsic solubility, s0, derived from water solubility, sw) and temperature effects (by admet & dmpk 3(2) (2015) 84-109 equilibrium solubility measurements improvements doi: 10.5599/admet.3.2.193 101 transforming measurements performed in the range 20 – 40 °c to the standard value of 25 °c), it can be demonstrated that the expected interlaboratory reproducibility of 0.7 log unit can be reduced to near 0.15. glossary dtt dissolution titration template potentiometric method used to determine intrinsic solubility, s0 sf shake-flask method, the “gold standard” solubility measurement method hh henderson-hasselbalch equation (e.g., eq. 1) kn aggregation constant, where n is the degree of aggregation pka negative logarithm of the ionization constant phsat the equilibrium ph of a saturated solution s solubility, ideally expressed in units of mol/l (m), μg/ml, or mg/ml s0 “intrinsic” solubility (i.e., the solubility of the uncharged form of the compound) sw “water” solubility, defined by dissolving enough pure free acid/base in distilled water (or water containing an inert salt as ionic strength adjustor) to form a saturated solution. the final ph of the suspension, phsat, and s0 can be calculated by the hh equation (when valid), provided the true pka is known. compound added as a salt form may disproportionate into free acid/base, depending on how much solid had been added. it is not generally possible to calculate the ph and s0 of such a drug salt suspension. sph “ph buffer” solubility (i.e., the total solubility of the compound at a well-defined phsat) acknowledgements helpful discussions with antonio llinàs (astrazeneca), krisztina takács-novák and gergely völgyi (semmelweis university, budapest), christel bergström (uppsala university), and tatjana verbić (university of belgrade) are much appreciated. we are greatly indebted to agustin g. asuero (univ. of seville), michael h. abraham (university of london), and william e. acree, jr. (univ. of n. texas) for pointing out many important log s-ph publications. uko maran (univ. of tartu), manfred kansy, holger fischer, and stefanie bendels (hoffman-la roche) have provided valuable insights and leads into the literature of chemoinformatics, for which we are grateful. references [1] j. huuskonen, m. salo, j. taskinen, j. chem. inf. sci. 38 (1998) 450-456. [2] j. huuskonen, j. rantanen, d. livingstone, eur. j. med. chem. 35 (2000) 1081-1088. [3] m.h. abraham, j. le, j. pharm. sci. 88 (1999) 868-880. [4] w.l. jorgensen, e.m. duffy, bioorg. med. chem. lett. 10 (2000) 1155-1158. [5] w.l. jorgensen, e.m. duffy, adv. drug deliv. rev. 54 (2002) 355–366. [6] c.a.s. bergström, u. norinder, k. luthman, p. artursson, pharm. res. 19 (2002) 182-188. [7] t.j. hou, k. xia, w. zhang, x.j. xu, j. chem. inf. comput. sci. 44 (2004) 266-275. [8] j.s. delaney, drug discov. today 10 (2005) 289-295. [9] j.c. dearden, expert opin. drug discov. 1 (2006) 31-52. [10] k.v. balakin, n.p. savchuk, i.v. tetko, curr. med. chem. 13 (2006) 223-241. [11] j. taskinen, u. norinder, in silico predictions of solubility. in: b. testa, h. van de waterbeemd (eds.). comprehensive medicinal chemistry ii, elsevier: oxford, uk, 2007, pp. 627-648. avdeef admet & dmpk 3(2) (2015) 84-109 102 [12] p. jain, s.h. yalkowsky, int. j. pharm. 385 (2010) 1–5. [13] a. shayanfar, a. jouyban, lat. am. j. pharm. 30 (2011) 1525-1530. [14] j. wang, t. hou, combi. chem. high throughput scr. 14 (2011) 1-11. [15] d. elder, r. holm, int. j. pharm. 453 (2013) 3-11. [16] j.l. mcdonagh, n. nath, l. de ferrari, t. van mourik, j.b. mitchell, j. chem. inf. model. 54 (2014) 844-856. [17] d.s. palmer, j.b.o. mitchell, mol. pharmaceutics 11 (2014) 2962-2972. [18] b. faller, p. ertl, adv. drug deliv. rev. 59 (2007) 533–545. [19] s.h. yalkowsky, y. he, p. jain, handbook of aqueous solubility data, second edition. crc press, boca raton, fl, 2010. [20] p. howard, w. meylan, physprop database. syracuse research corp., environmental science center, n. syracuse, ny, sept. 1999. [21] syracuse research corporation. the physical properties database (physprop) http://www.srcinc.com/what-we-do/environmental/scientific-databases.html (accessed 18 feb 2015). [22] analytical profiles of drug substances (analytical profiles of drug substances and excipients; profiles of drug substances, excipients and related methodology). k. florey (ed., vols. 1-20),h.g. brittain (ed., vols. 21-39). academic press, san diego, 1972-2014. [23] j.w. mcfarland, a. avdeef, c.m. berger, o.a. raevsky, j. chem. inf. comput. sci. 41 (2001) 13551359. [24] e. rytting, k.a. lentz, x.q. chen, f. qian, s. venkatesh, pharm res. 21 (2004) 237-244. [25] c.a.s. bergström, c.m. wassvik, u. norinder, k. luthman, p. artursson, j. chem. inf. comput. sci. 44 (2004) 1477-1488. [26] a. llinàs, r.c. glen, j.m. goodman, j. chem. inf. model. 48 (2008) 1289–1303. [27] a.j. hopfinger, e.x. esposito, a. llinàs, r.c. glen, j.m. goodman, j. chem. inf. model. 49 (2009) 1-5. [28] w.p. walters, what are our models really telling us? a practical tutorial on avoiding common mistakes when building predictive models. in: j. bajorath (ed.). chemoinformatics for drug discovery. john wiley & sons, hoboken, nj, 2014, pp. 1-31. [29] g. landrum, r. lewis, a. palmer, n. stiefl, a. vulpetti, j. cheminform. 3 (2011) 1-1. [30] m.h. abraham, a. ibrahim, a.m. zissimos, y.h. zhao, j. comer, d.p. reynolds, drug disc. today 7 (2002) 1056-1063. [31] a. sid lang, j-c bradley, ons melting point model 010. qdb archive, doi: 10.15152/qdb.104. qsardb content. property mpc. [32] g. völgyi, a. marosi, k. takács-novák, a. avdeef, admet & dmpk 1(4) (2013) 48-62. [33] a. avdeef, admet & dmpk 2(1) (2014) 33-42. [34] a. avdeef, admet & dmpk 2(1) (2014) 43-55. [35] m. stuart, k. box, anal. chem. 77 (2005) 983–990. [36] c. sköld, s. winiwarter, j. wernevik, f. bergström, l. engström, r. allen, l. box, j. comer, j. mole, a. hallberg, h. lennernäs, t. lundstedt, a-l. ungell, a. karlén, j. med. chem. 49 (2006) 6660-6671. [37] a. llinàs, j.c. burley, k.j. box, r.c. glen, j.m. goodman, j. med. chem. 50 (2007) 979-983. [38] k.j. box, j.e.a. comer, curr. drug metab. 9 (2008) 869-878. [39] l.y.s. narasimham, v.d. barhate, j. pharmacy res. 4(2) (2011) 532-536. [40] y.-l. hsieh, g.a. ilevbare, b. van eerdenbrugh, k.j. box, m.v. sanchez-felix, l.s. taylor, pharm. res. 29 (2012) 2738-2753. [41] j. comer, s. judge, d. matthews, l. towes, b. falcone, j. goodman, j. dearden, admet & dmpk 2(1) (2014) 18-32. [42] d. schönherr, u. wollatz, d. haznar-garbacz, u. hanke, k.j. box, r. taylor, r. ruiz, s. beato, d. http://www.srcinc.com/what-we-do/environmental/scientific-databases.html admet & dmpk 3(2) (2015) 84-109 equilibrium solubility measurements improvements doi: 10.5599/admet.3.2.193 103 becker, w. weitschies, eur. j. pharm. biopharm. 92 (2015) 155–170. [43] a. avdeef, pharm. pharmacol. commun. 4 (1998) 165-178. [44] a. avdeef, c.m. berger, c. brownell, pharm. res. 17 (2000) 85-89. [45] a. avdeef, c.m. berger, eur. j. pharm. sci. 14 (2001) 281-291. [46] a. avdeef, curr. topics med. chem. 1 (2001) 277-351. [47] b. faller, f. wohnsland, physicochemical parameters as tools in drug discovery and lead optimization. in: b. testa, h. van de waterbeemd, g. folkers, r. guy, r. (eds.). pharmacokinetic optimization in drug research. verlag helvetica chimica acta: zürich and wiley vch: weinheim, pp. 257-274 (2001). [48] c.a.s. bergström, m. strafford, l. lazarova, a. avdeef, k. luthman, p. artursson, j. med. chem. 46 (2003) 558-570. [49] a.f. fioritto, s.n. bhattachar, j.a. wesley, int. j. pharm. 330 (2007) 105-113. [50] g. ottaviani, d.j. gosling, c. patissier, s. rodde, l. zhou, b. faller, eur. j. pharm. sci. 41 (2010) 452457. [51] y. surakitbanharn, r. mccandless, j.p. krzyzaniak, r-m. dannenfelser, s.h. yalkowsky, j. pharm. sci. 84 (1995) 720-723. [52] a. avdeef, absorption and drug development second edition. wiley-interscience: new york. 2012. [53] e. shoghi, e. fuguet, e. bosch, c. rafols, eur. j. pharm. sci. 48 (2012) 290-300. [54] i. zimmermann, int. j. pharm. 31 (1986) 69-74. [55] a. avdeef, j.j. bucher, anal. chem. 50 (1978) 2137-2142. [56] z. wang, l.s. burrell, w.j. lambert, j. pharm. sci. 91 (2002) 1445-1455. [57] a. avdeef, adv. drug deliv. rev. 59 (2007) 568-590 [58] t-m. chen, h. shen, c. zhu, combi. chem. hts 5 (2002) 575-581. [59] s.l. mcgovern, e. caselli, n. grigorieff, b.k. shoichet, j. med. chem. 45 (2002) 1712-1722. [60] t. loftsson, d. hreinsdóttir, m. másson, int. j. pharm. 302 (2005) 18–28. [61] s. venkatesh, j. li, y. xu,r. vishnuvajjala, pharm. res. 13 (1996) 1453-1459. [62] y. he, s.e. tabibi, s.h. yalkowsky, j. pharm. sci. 95 (2006) 97-107. [63] a. fini, g. fazio, g. feroci, int. j. pharm. 126 (1995) 95-102. [64] m. pudipeddi, a.t.m. serajuddin, j. pharm. sci. 94 (2005) 929-939. [65] b.d. anderson, k.p. flora, preparation of water-soluble compounds through salt formation. in: c.g. wermuth, (ed.). the practice of medicinal chemistry. academic press, london, pp. 739-754 (1996). [66] h. kishi, y. hashimoto, chemosphere 18 (1989) 1749-1759. [67] a.r. katritzky, y. wang, s. sild, t. tamm, m. karelson, j. chem. inf. comput. sci. 38 (1998) 720-725. [68] e. baka, j.e.a. comer, k. takács-novák, j. pharm. biomed. anal. 46 (2008) 335-341. [69] a. avdeef, d. voloboy, a. foreman, dissolution and solubility. in: b. testa, h. van de waterbeemd (eds.). comprehensive medicinal chemistry ii, elsevier: oxford, uk, 2007, pp. 399-423. [70] a. avdeef, o. tsinman, pharm. res. 25 (2008) 2613-2627. [71] k. tsinman, a. avdeef, o. tsinman, d. voloboy, pharm. res. 26 (2009) 2093-2100. [72] a. avdeef, k. tsinman, o. tsinman, n. sun, d. voloboy, chem. biodiv. 11 (2009) 1796-1811. [73] j.h. fagerberg, o. tsinman, k. tsinman, n. sun, a. avdeef, c.a.s. bergström, mol. pharmaceutics 7 (2010) 1419-1430. [74] a. avdeef, s. bendels, o. tsinman, m. kansy, pharm. res. 24 (2007) 530-545. [75] j. alsenz, m. kansy, adv. drug deliv. rev. 59 (2007) 546–567. [76] j. alsenz, e. meister, e. haenel, j. pharm. sci. 96 (2007) 1748-1762. [77] c.a.s. bergström, k. luthman, p. artursson, eur. j. pharm. sci. 22 (2004) 387-398. avdeef admet & dmpk 3(2) (2015) 84-109 104 [78] a. glomme, j. märz, j.b. dressman, j. pharm. sci., 94 (2005) 1-16. [79] t. higuchi, f.m. shih, t. kimura, j.h. rytting, j. pharm. sci. 68 (1979) 1267-1272. [80] w.h. streng, s.k. his, p.e. helms, h.g.h. tan, j. pharm. sci. 73 (1984) 1679-1684. [81] f. wohnsland, b. faller, j. med. chem. 44 (2001) 923-930. [82] g. völgyi, e. baka, k.j. box, j.e.a. comer, k. takács-novák, anal.chim. acta 673 (2010) 40–46. admet & dmpk 3(2) (2015) 84-109 equilibrium solubility measurements improvements doi: 10.5599/admet.3.2.193 105 appendix brief summary of methods suitable for measuring equilibrium solubility the appendix briefly describes methods which can potentially be used to generate intrinsic solubility of sparingly-soluble drugs. kinetic methods are not suitable for such application, and are thus not considered. classical saturation shake-flask (sf) – still the ‘gold standard’ solubility measurement as a function of ph under equilibrium conditions usually requires long equilibration times (12 h 7 d). the shake-flask (sf) method is a relatively simple procedure: the drug is added as solid to a standard buffer solution (in a flask) until saturation occurs, indicated by excess undissolved drug. the thermostated saturated solution is shaken as the suspension equilibrates. after filtration or centrifugation, the concentration of the substance in the supernatant solution is then determined using hplc, usually with uv detection. for a solubility-ph profile, parallel measurements need to be performed in several different ph buffers. baka et al. [68] suggested concrete ways to improve the quality of the sf measurements. in high-quality sf measurements, (a) it is actually verified that the dissolution/precipitation of the solid has reached equilibrium, (b) the final (equilibrium) ph is measured, and (c) the equilibrated solid is isolated and identified (or characterized). the measured solubility is expected to be that of the most stable polymorph, which can often be a hydrate. micro-dissolution for polymorph/solvate solubility measurement (μdiss profiler) by usp standards, dissolution equipment typically uses 900-ml solution volumes, so it would be impractical to determine solubility of drugs, since an excessive amount of compound would be required to form a saturated solution. however, the miniaturized-volume dissolution apparatus, µdiss profiler (pion inc.), uses 1.0 ml working volumes and makes it practical to determine solubility (avdeef et al. [69]; avdeef and tsinman [70]; tsinman et al. [71]; avdeef et al. [72]; fagerberg et al. [73]). since the instrument uses in situ fiber optic uv (diode array) detection, the progression of dissolution may be followed in real time, making it possible to characterize the (transient) solubility of different polymorphs/solvates, over periods of days, if necessary. it has been demonstrated that 10-100 μg of powder can be characterized for both the transient polymorph solubility and equilibrium solubility of the most stable form of the solid. a hydroxypropyl-β-cyclodextrin phase-solubility study of itraconazole using the micro-dissolution apparatus estimated the active polymorph solubility of itraconazole to be < 9 ng/ml. equilibrium 96-well microtitre plate methods to increase throughput and decrease sample consumption, several pharmaceutical companies have transferred the larger-volume classical sf method to smaller-volume 96-well microtitre plate format linked to robotic liquid handling systems, using several distinct approaches (including those which avoid use of dmso). there are trade-offs in the high-throughput methods, and usually, the quality of the data are not expected to be as accurate as those obtained by the sf method. since the compound is usually introduced as aliquots of a 10 mm dmso stock solution, such methods have an upper solubility limit, typically less than 150 μm. also, the presence of the small amount of dmso in the final buffer solution increases the measured solubility values of the most insoluble compounds by up to 100-fold (e.g., glibenclamide, clotrimazole), compared to dmso-free conditions. this appears to be highly compound-dependent (chen et al. [58]). quality results require this dmso-effect to be factored out (e.g., pdisol-x method). avdeef admet & dmpk 3(2) (2015) 84-109 106 for acceptably reliable aqueous intrinsic solubility (s0) to be determined from single-ph determination (of compounds introduced from dmso stock solutions), it is necessary (a) to remove the dmso originating from the stock solution, (b) to know the pka value of the test compound, and (c) to assume that the henderson-hasselbalch (hh) equation is a valid description of the log s ph relationship. if the measurement is performed in the presence of ≤1 % v/v dmso, there is a way to determine s0, provided that the solubility measurement is done at several values of ph over a wide enough a ph range ([52]; pdisol-x method). if the pka is known, then it is not necessary to assume that the hh equation holds, as has been demonstrated by avdeef et al. [74]. lyophilization method (‘genevac’ 96-well dmso-removing device) some pharmaceutical companies have implemented methods where aliquots of 10 mm dmso stock solutions of test compounds are added to a microtitre plate. then, the dmso is removed by lyophilization (e.g., using genevac vacuum centrifugation apparatus), after which, a ph 6.5 or 7.4 buffer is added to the compound residues in the microtitre plate. the plate is sealed and allowed to incubate up to 24 h, usually at room temperature, during which time the plate may be shaken. for best results, the final (equilibrium) ph needs to be measured. with the initial dmso removed, measured solubility values can match those obtained by the traditional sf method. alsenz and kansy [75] described the lyophilization solubility assay (lysa) based on uv plate reader and hplc methods. equilibrium solubility assay (dmso-free ‘thesa’ method) alsenz and kansy [75] also described the thermodynamic solubility assay (thesa) microtitre plate method where < 1 mg of weighed solid sample is added a ph 6.5 buffer solution. the suspensions are stirred for 2 h and allowed to sit for another 20 h, before filtration. a provision for reading the final ph is included in the protocol. the advantage of the thesa method is that the compound is introduced in the original solid state. partially-automated solubility screening (dmso-free ‘pass’ method) alsenz et al. [76] described the microtitre plate partially-automated solubility screening (pass), where solid compounds (dmso-free) are suspended as slurries in volatile heptane, sonicated to increase fractionation and dispersion, and then dispensed quickly in small aliquots into microtitre plate wells. from the dispensed volume, the weight of drug is calculated. the advantage of the pass method is that the compound is introduced in the original solid state. the heptane is allowed to evaporate, before buffer is added. this procedure inherently can be used to assess relatively high solubility (several mg/ml) by a microtitre plate method. the comparison of pass solubility values to those from standard sf values was good. small-scale shake-flask method (ssf) bergström et al. [6] described the small-scale shake-flask (ssf) method, where crystalline compounds were added to 50-200 μl solutions whose ph was then adjusted to at least one ph below the pka for acids and one ph above the pka for bases, in an effort to measure s0 directly. after 24-72 h of stirring, the solutions were centrifuged to separate the phases. seventeen compounds were measured (0.7 ng/ml to 6 mg/ml). expanding the ssf approach, bergström et al. [77] further studied 25 bases over a wide ph range, using 0.15 m phosphate buffer adjusted with koh or h3po4. the suspensions were incubated for 24 h. phases were separated by centrifugation. an in-depth analysis of the 25 bases has been recently reported admet & dmpk 3(2) (2015) 84-109 equilibrium solubility measurements improvements doi: 10.5599/admet.3.2.193 107 [34], where it was tentatively suggested that 24 h may not be sufficient equilibration time for several of the sparingly-soluble bases studied. other concerns about ph control were raised by völgyi et al. [82]. miniaturized shake-flask (msf) glomme et al. [78] described the miniaturized shake-flask (msf) method. the sample is weighed (twice the in silico-estimated solubility) directly into the whatman uniprep filter chamber, which is then filled with 2 ml of buffer solution. the capped chambers are agitated for 24 h at 37 °c. afterwards, a chambermated plunger equipped with a filter is pushed over the sample solution to separate the solid. the final ph is read at the end of the equilibration period. dual-phase potentiometric methods the two related potentiometric methods described below are only suitable for ionizable compounds, whose pka can be determined in situ (dtt) or need to be determined in a separate method (cheqsol). dissolution template titration method (dtt) the dissolution template titration (dtt) semi-automated intrinsic solubility potentiometric method [43] takes the estimated pka and the octanol/water partition coefficient, log poct, as input parameters to predict s0. the dtt procedure then simulates the entire titration curve before the assay starts. this defines a “template” for data collection. titration is done in the direction of increasing dissolution, taking the saturated solutions past the point of complete dissolution. the noyes-whitney equation (eq. 5 below) is used to set the pace of data collection. the more insoluble the compound is (based on the template), the longer is the assay time (typically, 3 10 h). data are analyzed using bjerrum plots: ͞nh, the average number of bound protons, versus ph. since it is known how much strong acid, [hcl], and strong base, [koh], have been added to the solution at any point and since it is known how many dissociable protons, n, the sample adds to the solution, the total hydrogen ion concentration is defined. the difference between the latter and the free hydrogen (ph electrode reading) concentrations is equal to the concentration of the bound hydrogen, which, when divided by the sample concentration, c, results in ͞nh = { [hcl] [koh] + n c [h + ] + [oh ] } / c. the ph at half-integral value of ͞nh equals pka app , the apparent pka. the presence of precipitate is indicated by the pka app shifting to a higher value for acids and to a lower value for bases, compared to the true pka. the intrinsic solubility can be deduced by inspection of the curves and applying the relationship log s0 = log(c/2) -| pka app pka |. the graphically-estimated constant is subsequently refined by a massbalance weighted nonlinear regression procedure, which does not require that the henderson-hasselbalch relationship be valid (although, often, it is assumed to be so). the pka can be determined alongside intrinsic solubility. cheqsol potentiometric method stuart and box [35] embraced the ͞nh part of the dtt method and developed a valuable novel way to speed up the time to reach equilibration. focusing on the ph region where ͞nh is expected to be near halfintegral, the cheqsol instrument quickly dissolves the solid by raising the ph for acids (e.g., to 10-12) or lowering the ph for bases (e.g., to 1-2). afterwards, the original ph is re-established by adding acid/base titrant, whereupon the solid re-precipitates, possibly as an amorphous phase. the procedure of dissolution-precipitation is cycled several times, from which the ph corresponding to a known value of nh (near 0.5) is determined by interpolation. using the ddt nh relationships, the intrinsic s0 is calculated from avdeef admet & dmpk 3(2) (2015) 84-109 108 a single ph point, suggested by the value of the pka. since the method is not based on the general massbalance base regression analysis, cheqsol implicitly assumes that the henderson-hasselbalch equation is always valid, which in some cases could potentially be a source of systematic error. facilitated dissolution method (fdm) the facilitated dissolution method (fdm), devised by higuchi et al. [79], can be used to overcome extremely low equilibration rates during solubility measurement of practically-insoluble compounds. since it is rarely practiced, the method is described here in some detail (cf., [52]). in the fdm approach, a twofold excess of solid over that needed to saturate the solution is recommended. to overcome the expected slow dissolution, a small volume of an immiscible organic solvent (e.g., 2 % v/v hexadecane) is added to the aqueous solution of the sparingly-soluble compound. as long as the saturated system contains three distinct phases (solid, oil, and water), the solubility value is not altered by the presence of the water-immiscible oil. (this is sufficiently but not precisely true for ionizable compounds, as commented below.) to show this, consider a weak base example (e.g., terfenadine), for which the fdm equilibrium reactions are: b(s) ↔ b(org) sorg = [b(org)] (2) b↔ b(org) po/w = [b(org)] / [b] (3) where eq. 2 denotes the solubility of the compound in oil, sorg, and eq. 3 denotes the oil-water partition coefficient of the compound, po/w. by subtracting the partition reaction from the solubility-in-oil reaction, one gets the expected solubility-in-water equation, b(s) ↔ b sw = sorg / po/w (4) so, the presence of a small quantity of oil, into which the water sparingly-soluble compound can appreciably dissolve, does not affect the aqueous solubility value. according to the noyes-whitney dissolution rate (µg/cm 3 ∙s) equation under sink condition, d[b]/dt = (a / v) pabl sw (5) where a = powder surface area (cm 2 ), v = volume of aqueous solution (cm 3 ), pabl = permeability of the aqueous boundary layer adjacent to the surface of the solid particles (cm/s), and sw = aqueous solubility (µg/cm 3 ). the addition of the oil may also alleviate problems due to poor wettability of the original crystalline solid. let's consider the terfenadine fdm example explicitly. the intrinsic solubility of the weak base in water is 5.6x10 -8 m (log s0 = -7.25) at 25 °c (streng et al. [80]; data analyzed using pdisol-x). the measured pka is 9.91 at 25 °c, i = 0.15 m (origin-shifted yasuda-shedlovsky method [52]). consider 20 μg of terfenadine free base added to 1 ml of 10 mm taurine buffer in 0.15 m nacl at ph 9.0, containing 20 μl of hexadecane. let's assume 100 μm (diam.) spherical particles of solid in the suspension at the start. the hexadecanewater partition coefficient of terfenadine is estimated to be log phxd/w = 2.3 (using calibration standards from wohnsland and faller [81]). the dissolution simulation feature of pdisol-x was used to calculate the precise concentrations and quantities in each of the three phases. at ph 9.0, 0.302 μg of the drug is predicted to dissolve in the buffer in the absence of added hexadecane, with 19.70 μg remaining in the solid state. the calculated total surface area of the solid is initially 0.0086 cm 2 and decreases by 1.3 % at saturation. the (a/v) factor for eq. 5 is 0.0086 cm -1 . after the addition of 20 μl of hexadecane, the admet & dmpk 3(2) (2015) 84-109 equilibrium solubility measurements improvements doi: 10.5599/admet.3.2.193 109 calculated amount of the drug in water is still exactly 0.302 μg; the amount in the solid is now 19.59 μg, as 0.20 μg of the drug partitions into the 20 μl of hexadecane. the solubility in hexadecane is calculated to be 2.2x10 -7 m, compared to 6.4x10 -7 m in the buffer. the smaller volume of the oil phase increase the rate of dissolution into the oil phase: the (a/v) factor between the solid surface and the oil volume is 0.0086 cm 2 /0.02 cm 3 = 0.43 cm -1 . this suggests that the rate of dissolution into the oil phase will be about 50 times greater than the rate of dissolution into the water phase. what about the rate of transfer from the oil phase into water (to complete the transfer cycle in eq. 4)? this is expected to be high in a well stirred (e.g., 500 rpm) solution, dispersing the oil into tiny droplets, thus increasing oil-water interfacial surface area. hence, a small amount of adjunct oil substantially increases the overall dissolution rate associated with eq. 5, without affecting the equilibrium solubility of terfenadine. it is important to point out that the fdm approach does not work for ionizable compounds in poorlybuffered solutions, and generally cannot be applied in the dtt or the cheqsol method (since buffers are not generally used). using the pdisol-x program, simulations in buffer-free solutions show that the addition of 10 μl hexadecane can affect the water solubility of terfenadine, through a subtle interplay of the various ph-dependent equilibria. the effect is lessened if a very large excess of terfenadine were added. but this is not recommended, as discussed by higuchi et al. [79]. as far as we are aware, the fdm approach was last applied by venkatesh et al. [61] in the solubility determination of cosalane, a steroid derivative with aqueous intrinsic solubility of much less than 1 ng/ml. the pka values of cosalane have not been reported to date, so there may be some uncertainty in the intrinsic solubility of cosalane. since the fdm method is best applied to practically-insoluble molecules, probably lc/ms-ms resources would be required to measure the extremely low sample concentrations in saturated solutions. ©2015 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ manuscript doi: 10.5599/admet.0000.0000 47 admet & dmpk 5(1) (2017) 47-56; doi: 10.5599/admet.5.1.327 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper the inhibition of adenylate kinase by 2,4-thiazolidinedione evaluated by protein-ligand docking mihaela ileana ionescu iuliu hatieganu university of medicine and pharmacy, department of microbiology, 6 pasteur, cluj-napoca, romania corresponding author: e-mail: mionescu@umfcluj.ro received: july 14, 2016; revised: february 16, 2017; published: march 25, 2017 abstract due to its crucial role in nucleotide metabolism, adenylate kinase deserves a special attention in screening of potential inhibitors. herein, we report the assessment of the relative orientation of the ligand 2,4thiazolidinedione to adenylate kinase crystallized in closed conformation. protein-ligand docking was performed to estimate the binding energy and inhibition constant of 2,4-thiazolidinedione to the adenylate kinases’ active sites from different organisms. our results revealed the best orientation of 2,4thiazolidinedione is with gram-positive and acid fast bacteria adenylate kinase – ki = 0.76±0.1 mm and binding energy -4.26±0.08 kcal/mol. human adenylate kinases display unfavourable interactions, the binding affinity fluctuating among ki=0.84 mm and 8.8 mm (3.88±3.51); the energy binding -3.56±0.57. from the three human adenylate kinases analysed, only isoenzyme 2 shows a binding conformation similar to its counterpart from e. coli. adenylate kinase this small enzyme needed for survival of every organisms interacts differently with 2,4-thiazolidinedione, this selectivity being the most important evidence of the present study. keywords molecular docking; 2, 4-thiazolidinedione introduction bacterial resistance to antibiotics is an evolving problem of medical practice [1,2]. discovery of new antibacterial agents and/or reevaluating of different targets are impetuously needed [3], [4]. these kinds of studies are usually constructed by comparison of well characterized pathogens. escherichia coli (e. coli) is an opportunistic gram-negative pathogen, widely isolated from nosocomial infections [5]. streptococcus pneumoniae (s. pneumoniae) – a gram-positive cocci – is an extracellular organism responsible for pneumoniae, otitis media, meningitis and even sepsis in very young children and adults above 65 years [6,7]. tuberculosis treatment remains one of the most challenging problems for clinicians; multidrugresistant mycobacterium tuberculosis (m. tuberculosis) strains considerably diminish treatment options [8]. adenylate kinase (ak) – atp: amp phosphotransferase; ec 2.7.3.4 – belong to the nucleotide kinase super family which catalyzes conversion between adenylate nucleotide by the following reaction: mg 2+ atp + amp ↔ mg 2+ adp + adp. ak is widely studied due to its implication in nucleotide metabolism and to its ubiquities – it is described in archaea, prokaryotes and eukaryotes. some aks are very well characterized, http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:mionescu@umfcluj.ro mihaela i. ionescu admet & dmpk 5(1) (2017) 47-56 48 many of them have the crystal structures solved and biochemical parameters established. ak is a flexible protein, which can adopt different conformations during catalytic process [9,10]. ak is, due to its universality and its flexibility, an attractive target for screening of new inhibitors. thiazolidinediones (tzds) derivates – synthetic agonists for the peroxisome proliferator-activating receptorgamma receptor – were demonstrated as potential inhibitors of s. pneumoniae out-growth [11]. as we are interested in inhibition of ak by thiazolidine derivates [12], the present study is focused on 2,4thiazolidinedione interaction with the active site of bacterial and human aks. materials and methods data collection and design the sequences and annotations for the proteins examined were collected from ncbi protein database (http://www.ncbi.nlm.nih.gov/protein). aks x-ray 3d structures were downloaded from the protein data bank (http://www.rcsb.org/pdb). aks co-crystallized with the substrates or substrates analogs like the inhibitor p1,p5-di(adenosine-5')pentaphosphate (ap5a) (c20h29n10o22p5), were stored for further procedures. the structure of the ligand 2,4-thiazolidinedione (c3h3no2s) was retrieved from the open chemistry database – national center for biotechnology information pubchem compound database; cid=5437, https://pubchem.ncbi.nlm.nih.gov/compound/5437. this open database provided chemicals as sdf files. protein-ligand docking to analyze the binding of 2,4-thiazolidinedione, we performed protein-ligand docking by autodock4, a widely used non-commercial program, which provided useful instruments for accurate prediction of binding interaction [13–15]. the autodocktools (http://mgltools.scripps.edu/downloads) offer an excellent interface for non-bioinformaticians. interpreting the docking data was possible by employing of free 3d molecular visualization tools – biovia discovery studio visualizer 2016 (ds visualizer 2016); (http://accelrys.com/products/discovery-studio/visualization-download.php). aks included in the present study were presented in the table 1. firstly, from the 3d structure, substrate or substrate analogs (ap5a) and all solvent molecules were removed and a closed conformation was obtained, which allows an algorithm based on a rigid protein structure. in our case, this docking algorithm was sufficient to obtain reliable data. the lamarckian genetic algorithm (lga), with a maximum of 2,500,000 energy evaluations, was choose as searching parameter [16]. multiple sequence alignment multiple sequence alignment was performed using clustal omega program (http://www.ebi.ac.uk/tools/msa/clustalo), an open access tool provided by the european bioinformatics institute – embl-ebi [21,22]. statistical evaluations: data were expressed as mean ± sd. inclusion criteria: aks from any microorganism or from human origin with crystal structure determined by x-ray crystallography co-crystallized with substrate or substrate analogs, so a closed conformation is provided. exclusion criteria: aks with no crystal structure recorded in protein databases or 3d structure in open conformation. http://www.ncbi.nlm.nih.gov/protein http://www.rcsb.org/pdb http://pubchem.ncbi.nlm.nih.gov/search/#collection=compounds&query_type=mf&query=c3h3no2s&sort=mw&sort_dir=asc https://pubchem.ncbi.nlm.nih.gov/compound/5437 http://mgltools.scripps.edu/downloads http://accelrys.com/products/discovery-studio/visualization-download.php http://www.ebi.ac.uk/tools/msa/clustalo admet & dmpk 5(1) (2017) 47-56 adenylate kinase inhibition by 2,4-thiazolidinedione doi: 10.5599/admet.5.1.327 49 table 1 the main characteristics of the aks used in protein-ligand docking accession number source co-crystallized with substrate or substrate analogs no. of residues references bacterial source 3hpq e. coli ap5a 214 [17] 1ank e. coli amp, amppnp 214 [18] 4w5j s. pneumoniae ap5a, mg 2+ 217 [19] 1zip b. stearothermophilus ap5a, mn 2+ , and mg 2+ 217 [9] 2cdn m. tuberculosis two adp molecules and mg 2+ 201 [20] 2rgx aquifex aeolicus ap5a, zn 2+ 206 [10] homo sapiens 2c95 (chain a – ak1) bis(adenosine)-5'-tetraphosphate; malonate ion 196 unpublished 2c9y (chain a – ak2) bis(adenosine)-5'tetraphosphate; ethylene glycol 242 unpublished 2bwj (chain a – ak5) amp, cl , so4 2 199 unpublished results and discussion protein-ligand docking aks from different sources were subjected to docking with 2,4-thiazolidinedione. it is already established that ak is a natural flexible protein, binding and releasing of its substrates are followed by motions of the amp-binding domain and lid domain. as a result of sophisticated conformational transitions a so-called closed conformation results during catalytic process [23,24]. starting from these premises, rigid protein structures were considered for further docking algorithm. for each ak, ten conformations were obtained as a result of docking process. the conformation with the smallest binding energy (∆g) and the smallest inhibition constant (ki) was considered the best. as it is shown in the table 2, regarding aks from gram-positive and acid fast species, ∆g values are almost the same with a mean of -4.26±0.08 kcal/mol, and ki are clustered around the mean 0.76±0.1 mm. on the other hand, the human counterpart aks exhibit more diverse values. human ak isoenzyme 2 (2c9y) – mitochondrial origin – exhibits values similar to enzymes of bacterial sources. active sites residues which interact with the ligand 2,4-thiazolidinedione are shown in the table 3. the most striking observation is related to aks of human origin, as it is illustrated in the figures 1 and 2. the ak5 (2bwj) – expressed exclusively in the brain [25] – is clearly a poor target, none of the residues from the active site being involved in 2,4-thiazolidinedione binding. further, in the case of ak2 (2c9y), the ligand forms standard hydrogen bonds with two nonpolar residues from amp-binding region (gly100 and phe101) and with lys28 from atp-binding domain. besides, from the aks examined in the present study, ak2 and one ak from e. coli 3hpq are the only ones which bind 2,4-thiazolidinedione with p-loop domain residues (lys28 and lys13, respectively). notably, the same oxygen atom (o2) (scheme 1) of the ligand forms hydrogen bond with lysine but different atoms are involved – nz regarding 2c9y and ce for 3hpq (figures 3 and 4). on the contrary, apart from the residues involved in amp-binding, ak1 shows only a glycine from nmp-binding domain which binds the ligand. this might be a reasonable explanation of almost twofold value ki of ak1 (1.99 mm) compared with ak2 (0.84 mm). a unique detail is observed in the case of gram-positive and acid fast aks – a favorable interaction between pi-sulfur of the ligand and the nonpolar residue phe35 from nmp-binding domain (figures 5 and mihaela i. ionescu admet & dmpk 5(1) (2017) 47-56 50 6) might play an important role in stabilization of ak-2,4-thiazolidinedione complex [26]. multiple sequence alignment exhibits a conservative replacement of phe35 with leucine on other aks. moreover, in a. aeolicus’ ak, van der waals’ forces between leu35 and the only sulfur atom of the ligand are noticed. table 2. energy binding and inhibition constant (ki) of the best conformation of the complex ak2, 4-thiazolidinedione ak ∆g, kcal/mol ki, mm scheme 1. the numbering of 2,4-thiazolidinedione atoms gram-negative bacteria 3hpq -4.0 1.17 1ank -4.18 0.86 2rgx -4.22 0.81 mean ± sd -4.13±0.09 0.94±0.16 gram-positive and acid fast bacteria 1zip -4.36 0.64 4w5j -4.27 0.77 2cdn -4.16 0.89 mean ± sd -4.26±0.08 0.76±0.1 homo sapiens 2c95 -3.68 1.99 2c9y -4.19 0.84 2bwj -2.8 8.8 mean ± sd -3.56±0.57 3.88±3.51 total mean ± sd -3.98 ± 0.46 1.86 ± 2.48 table 3. residues of substrate binding domains, according to protein database (www.rcsb.org/pdb/home/home.do) details code color for interactions with 2,4-thiazolidinedione: in green are shown conventional hydrogen bonds; blue – van der waals’ forces, brown – carbon hydrogen bonds; orange – pi-sulfur, red – unfavorable acceptor-acceptor ak nmp-binding domain amp binding lid domain atp binding (p-loop) 3hpq ser30-val60 lys57-val59, gly85-(phe86, pro87)arg88, thr31, arg36, gln92, arg167 gly122-asp159 gly10-(lys13)-thr15, val132tyr133, arg119, arg123, lys200 1ank ser30-(leu58)val59 lys57-(leu58)-val59, gly85-(phe86)arg88, thr31, arg36, gln92, thr156, arg167 gly122-asp159 gly10-thr15, val132-tyr133, arg119, arg123, lys200 2rgx ser30-(thr31/thr31, leu35, leu58)-val59 glu57-val59, gly82, phe83, pro84, arg85, thr31/thr31, gln89, arg131 gly123-asp153 gly10-thr15, val133-tyr134, arg120, arg124, lys189 1zip ser30-(thr31, phe35, leu58)-val59 asp57-(leu58)-val59, gly85-arg88, ser30, arg36, gln92, gln159, arg171 gly126-asp163 gly10-thr15, thr136-tyr137, arg127, gln199 4w5j ser30-(phe35)val59 glu57-(leu58)-val59, gly85-(gly86, tyr87)-arg88, thr31, arg36, gln93, arg156, arg167 gly127-asp159 gly10-thr15, thr137-phe138, arg128, gln195 2cdn ser30-(phe35, leu58)val59 asp57-(leu58)-val59, gly85-(tyr86)arg88, thr31, arg36, gln92, arg129, arg140 gly126-asp132 gly10-thr15, arg127, gly166 2c95 ser38-(gly40)val67 gln65-val67, gly94-arg97, thr39, arg44, gln101, arg138, arg149 lys131-asp141 gly18-thr23, arg132, gly177 2c9y ala45-val74 lys72-val74, gly100, phe101, pro102, arg103, thr46, arg51, gln107, arg175, arg186 gly141-asp178 gly25-(lys28)-thr30, ser151tyr152, arg138, arg142, gln214 2bwj ser41-val70 asp68-val70, gly97-arg100, thr42, arg47, arg104, arg152 gln136-asp144 gly21-thr26, arg135, gly180 http://www.rcsb.org/pdb/home/home.do admet & dmpk 5(1) (2017) 47-56 adenylate kinase inhibition by 2,4-thiazolidinedione doi: 10.5599/admet.5.1.327 51 the interactions of 2,4-thiazolidinedione with aks of gram-negative bacteria a. aeolicus – a hyperthermophilic species – and two structures from e. coli are shown in the figures 3 and 4. in the latter cases, even though the aminoacid sequences and the structures are identical, different ligand orientations were observed. 2c95 2c9y 2bwj figure 1. human aks docked with 2,4-thiazolidinedione. 2c95-ak1; 2c9y-ak2; 2bwj-ak5 code color for interactions: in green are shown conventional hydrogen bonds; yellow – carbon hydrogen bonds; red – unfavorable bump 2c95 2c9y 2bwj figure 2. interaction 2,4-thiazolidinedione with of human aks 3hpq 1ank 2rgx figure 3. interaction 2,4-thiazolidinedione with of aks from gram-negative microorganisms mihaela i. ionescu admet & dmpk 5(1) (2017) 47-56 52 3hpq 1ank 2rgx figure 4. aks from gram-negative microorganisms docked with 2,4-thiazolidinedione. 3hpq-e. coli; 1anke. coli; 2rgxa. aeolicus. code color for interactions: in green are shown conventional hydrogen bonds; yellow – carbon hydrogen bonds; orange – sulfur-x; red – unfavorable acceptor-acceptor one explanation could reside on the fact that the two enzymes being co-crystallized with different substrate analogs (table 1). it is well known that ak is a flexible enzyme, the lid domain and nmp-binding domain being highly dynamic. therefore, the transition between open and closed states is characterized by important conformational fluctuations (dynamics). studies performed with ak mutated in the residues which do not alter its structure, demonstrated that the binding interactions of different ligands affect the dynamics of ak, without accompanying structural changes of the enzyme [23,27,28]. a. aeolicus’s ak (2rgx) shows particularly comparable docking with s. pneumoniae (4w5j) and m. tuberculosis (2cdn) counterparts. the exception is phe35 residue which is not present in gram-negative aks. on the contrary, 2rgk binds the sulfur from 2,4-thiazolidinedione structure to thr31 residue. moreover, glutamine, a polar residue strictly conserved in aks – the last residue of the consensus sequence motif [livmfywca] – [livmfyw] (2)–d–g–[fy]–p–r–x(3) –[nq] – promotes hydrogen bonds with o2 / o3 in a. aeolicus, e. coli (1ank), ak1, m. tuberculosis, and gram-positive aks. on these six aks, sulfur atom contributes to protein-ligand complex stabilization also. contrary to aks of gram-positive and acid fast bacteria where pi-sulfur interactions were observed, hydrogen bonds were noticed for the rest. regarding 1ank and ak1, the valine residue val59, respectively val67, strictly conserved in aks, participates in hydrogen bonding with sulfur atom. admet & dmpk 5(1) (2017) 47-56 adenylate kinase inhibition by 2,4-thiazolidinedione doi: 10.5599/admet.5.1.327 53 4w5j 1zip 2cdn figure 5. interaction 2,4-thiazolidinedione with of aks from gram-positive and acid fast microorganisms 4w5j 1zip 2cdn figure 6. aks from gram-positive and acid fast microorganisms docked with 2,4-thiazolidinedione. 4w5js. pneumoniae; 1zipb. stearothermophilus; 2cdnm. tuberculosis. code color for interactions: in green are shown conventional hydrogen bonds; yellow – carbon hydrogen bonds; orange – pi-sulfur in gram-negative microorganisms, the ligand 2,4-thiazolidinedione binds different residues. looking at e. coli’s aks, we noticed that gly85, another residue strictly conserved from the consensus sequence motif, promotes a hydrogen bonding with the sulfur in the instance of 3hpq, but failed to establish a valid interaction in 1ank. on the contrary, in human ak2, this residue (gly100) interacts with the o2 atom from the 2,4-thiazolidinedione structure. as it is demonstrated in the figures 5 and 6 and in multiple sequence alignment (figure 7), aks from gram-positive and acid fast microorganisms, bind almost in the same manner the residues involved in ampbinding (thr31, gln93, and phe35), a notable difference is for ak from b. stearothermophilus in this latter instance, the smallest binding affinity is noticed – ki = 0.64 mm. one can speculate other gram-positive pathogens aks’ have the same inhibition pattern. aks of human origin display unfavorable interactions, the binding affinity fluctuating among 0.84 mm mihaela i. ionescu admet & dmpk 5(1) (2017) 47-56 54 and 8.8 mm. mitochondrial isoenzyme 2 of ak behaves in the same manner as one e. coli’s ak 3hpq. as demonstrated in the figure 7, they bind the same residues and have comparable binding affinity, ki = 0.84 mm and 0.86 mm (table 2). on the contrary, even though ak2 has energy binding and ki comparable to gram-positive microorganisms, 2,4-thiazolidinedione binds amp-binding pocket and atp-binding domain simultaneously. considering these different binding patterns of human aks the selectivity of 2,4thiazolidinedione is an advantage in the context of antimicrobial treatment. multiple sequence alignment primary structure offers useful details; comparisons of even small identities could be exploited for finding interesting patterns. it would be of interest to speculate that enzymes with similar primary structure, behave the same with otherwise well characterized enzyme. sometimes, observing minute details on primary structure has a major impact in decipher intriguing mechanism of a particular protein. the ligand chooses for the present work, 2,4-thiazolidinedione, has proven to bind important residues involved in amp-binding. figure 7. multiple sequence alignment. in gray background is marked atp-binding domain – the walker a motif gxxxxgkgt/s; in blue background are shown leucine and valine residues, very well conserved in aks family; in green background is highlighted the consensus sequence [livmfywca] – [livmfyw] (2)–d–g–[fy]– p–r–x(3)–[nq], were aspartic acid and arginine, strictly conserved in aks family, are bolded; in red are colored the residues which interact with 2,4-thiazolidinedione the main limitation of the present work is the availability of aks’ crystal structures deposited in the public databases. even though some databases have some weaknesses – some of them being redundant – they remain a powerful tool in fundamental research. unfortunately, kinetic experiments were not admet & dmpk 5(1) (2017) 47-56 adenylate kinase inhibition by 2,4-thiazolidinedione doi: 10.5599/admet.5.1.327 55 performed, which was another limit of the present study. but the biochemical characterization of a particular enzyme, not always deciphers all capabilities of the protein of interest. to put these findings in perspective, two approaches could be taken in consideration. either the screening of library protein molecules [28] to find receptors to the highest affinity to 2,4-thiazolidinedione, or design novel tzds derivates with better binding affinity. inevitably, inhibitors used for bacterial infections could interact with human counterpart’s target. although these evidences are not strong enough to persuade us that 2,4-thiazolidinedione is selective inhibitors for bacterial aks, a novel perspective of ak potentiality is revealed. results obtained in this study lead us to the conclusion that refining the structure of 2,4-thazolidinedione could be a viable idea in developing potential inhibitors for vital enzymes like ak . conclusions this study demonstrated an effort to reveal interactions of ak of bacterial / human origin with 2,4thiazolidinedione. the most noteworthy evidences were achieved referring to gram-positive and acid fast bacteria’s aks; 2,4-thiazolidinedione interacts with the same residues belonging to nmp-binding region. further, the complex comprising aks of s. pneumoniae and m. tuberculosis exhibit the same ligand-protein pattern. human aks interactions with 2,4-thiazolidinedione are divergent; a single notable connection being observed in case of mitochondrial isoenzyme 2 and its counterpart from e. coli. ak, a small enzyme needed for survival, could be an attractive target for tzds derivates, at least in case of m. tuberculosis and gram-positive microorganisms. acknowledgements: the author would like to thank the referees for their courteously comments. references [1] v.d. rosenthal, am. j. infect. control 44(6) (2016) e81–e90. [2] j. perry, n. waglechner, g. wright, cold spring harb. perspect. med. 6(6) (2016) 1-9. [3] m. jamal, t. hussain, c. rajanna das, s. andleeb, jundishapur j. microbiol. 8(11) (2015) e25917. [4] g. dinos, c. athanassopoulos, d. missiri, p. giannopoulou, i. vlachogiannis, g. papadopoulos, d. papaioannou, d. kalpaxis, antibiotics 5(2) (2016) 1-21. [5] j. xia, j. gao, w. tang, biosci. trends 10(1) (2016) 14–21. [6] m.t. htar, d. christopoulou, h.-j. schmitt, bmc infect. dis. 15 (2015) 1-10. [7] c. trucchi, c. paganino, f. ansaldi, j. prev. med. hyg. 56(3) (2015) e144–e149. [8] t. maitre, a. aubry, v. jarlier, j. robert, n. veziris, and cnr-myrma, med. mal. infect. 47(1) (2016) 3-10. [9] m.b. berry, g.n.j. phillips, proteins struct. funct. genet, 32(3) (1998) 276–288. [10] k.a. henzler-wildman, v. thai, m. lei, m. ott, m. wolf-watz, t. fenn, e. pozharski, m.a. wilson, g.a. petsko, m. karplus, c.g. hübner, d. kern, nature 450(7171) (2007) 838–844. [11] m.e. stegenga, s. florquin, a.f. de vos, t. van der poll, crit. care med. 37(2) (2009) 614–618. [12] m.i. ionescu, a.z. costache, o.oniga, h.l. banciu, i.lupan, i. lupan, rom. rev. lab. med. 21(1) (2013) 93–99. [13] d.s. goodsell, a.j. olson, proteins struct. funct. genet. 8(3) (1990) 195–202. [14] g.m. morris, r. huey, w. lindstrom, m.f. sanner, r.k. belew, d.s. goodsell, a.j. olson, j. comput. chem. 30(16) (2009) 2785–2791. [15] s.m. rizvi, s. shakil, m. haneef, excli j. 23(12) (2013) 831–857. [16] g.m. morris, d.s. goodsell, r.s. halliday, r. huey, w.e. hart, r.k. belew, a.j. olson, j. comput. chem. 19(14) (2015) e1004586. mihaela i. ionescu admet & dmpk 5(1) (2017) 47-56 56 [17] t.p. schrank, d.w. bolen, v.j. hilser, proc. natl. acad. sci. u. s. a. 106(40) (2009) 16984–16989. [18] m.b. berry, b. meador, t. bilderback, p. liang, m. glaser, g.n. phillips, proteins 19(3) (1994) 183– 198. [19] t.t. thach, s. lee, acta crystallogr. sect. f struct. biol. commun. 70(11) (2014) 1468–1471. [20] m. bellinzoni, a. haouz, m. graña, h. munier-lehmann, w. shepard, p.m. alzari, protein sci. 15(6) (2006) 1489–1493. [21] w. li, a. cowley, m. uludag, t. gur, h. mcwilliam, s. squizzato, y.m. park, n. buso, r. lopez, nucleic acids res. 43 (2015) w580–584. [22] f. sievers, a. wilm, d. dineen, t.j. gibson, k. karplus, w. li, r. lopez, h. mcwilliam, m. remmert, j. söding, j.d. thompson, d.g. higgins, mol. syst. biol. 7, (2011) 539. [23] c.w. müller, g.e. schulz, j. mol. biol. 224(1) (1992) 159–177. [24] s. esteban-martín, r. b. fenwick, j. ådén, b. cossins, c. w. bertoncini, v. guallar, m. wolf-watz, x. salvatella, plos comput. biol. 10(7) (2014) e1003721. [25] a.s.l. ng, j. kramer, a. centurion, j. dalmau, e. huang, j.a. cotter, m.d. geschwind, j. neuroimmunol. 15(287) (2015) 31-35. [26] a.l. ringer, a. senenko, c.d. sherrill, protein sci. 16(10) (2007) 2216–2223. [27] v.v hemanth, g. rao, s. gosavi, plos comput biol. 10(11) (2014) e1003938. [28] y.g. chushak, r.r. chapleau, j.s. frey, c.a. mauzy, j.m. gearhart, toxicol. res. 4(2) (2015) 519–526. ©2017 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ synthetic routes to theranostic applications of carbon-based quantum dots doi: https://doi.org/10.5599/admet.1747 1 admet & dmpk 0(0) (2023) 000-000; doi: https://doi.org/10.5599/admet.1747 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index review synthetic routes to theranostic applications of carbon-based quantum dots pemula gowtham§, karthick harini§, anbazhagan thirumalai, pragya pallavi, koyeli girigoswami, agnishwar girigoswami* medical bionanotechnology, faculty of allied health sciences, chettinad hospital & research institute (chri), chettinad academy of research and education (care), kelambakkam, chennai, tn-603 103, india *corresponding author: e-mail: agnishwarg@gmail.com; tel. +91-9445 268 615 §equal contribution received: march 15, 2023; revised: april 30, 2023; published: may xx, 2023 abstract background and purpose: modern technologies are making advanced paths to address emerging issues. the development of carbon dots (cds) technology at a tiny level has been researched to have made impeccable strides in advancing the modern scientific field, especially in nanomedicine. experimental approach: researchers have gained much attention on cds of their unique properties in the synthesis, easy surface modifications, excellent optical properties, low toxicity, and water solubility. doping carbon dots with other elements makes them more convenient for their use in the medical sector. key results: the manuscript provides a detailed discussion of the two main methods, including the hydrothermal pathway. cds are synthesized bottom-up by building up molecules at the atomic scale and top-down by transforming large carbon particles into nanoscale dimensions. conclusion: the present article discussed the role, importance, and recent advancements in the synthesis of cds, by using various approaches giving importance to the hydrothermal process. recent investigations, their mechanism, and theranostic applications have also been reported. ©2023 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords bottom-up approach; carbon dots; hydrothermal synthesis; synthetic route; theranostic applications introduction carbon dots (cd) are specific quantum dots containing carbon nanocrystals with diameters smaller than 10 nm. due to their quantum confinement, these ultra-small carbon crystals show photoinduced electron transfer-driven tunable photoluminescence, chemiluminescence, and electrochemical luminescence properties [1]. compared to organic fluorophores, carbon dots exhibit high tunable flexibility, improved aqueous solubility, biocompatibility, bioavailability, chemical inertness, and considerably low photobleaching to apply in biomedical imaging and biosensing [2,3]. their high flexibility and biocompatible properties allow these nanosized carbon dots to be highly useful in bioimaging, drug delivery, and diagnosis [4,5]. owning to the electron transfer mechanism and good optical absorption, cds are regarded as a good photosensitizer due to their ability to split water molecules and reduce carbon dioxide and pollutants degradation upon https://doi.org/10.5599/admet.1747 https://doi.org/10.5599/admet.1747 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:agnishwarg@gmail.com http://creativecommons.org/licenses/by/4.0/ pemula gowtham et al. admet & dmpk 00(0) (2023) 000-000 2 irradiation with uv-vis-nir region. cds display photocatalytic properties in the presence of other photosensitizers and redox mediators like enzymes or nanocomposite that functions as electron shuttles in between cds and enzyme-bound substrates [6,7]. it was reported that cds, in combination with sio2, au, ag, and cu nanocomposites, are widely used in the oxidation and reduction of alkanes, cyclooctenes, nitrobenzenes, quinones, etc. in semi-biological photosynthesis, the enzyme fumarate reductase catalyzes photo-enzymatic degradation after combining with cds to convert fumarate to succinate [8]. the cd can be co-doped with other elements, such as n, s, cl, and br, and other metal ions, such as copper, titanium, zinc, etc., which improves photocatalytic activities [9,10]. these elemental amalgamations enhance the electron-exchange ability of carbon dots that take a major role in redox reactions at the surface. these induced redox properties are used in wastewater management and hydrogen generation. the graphite nanoparticles lacked molecular dynamics simulation studies. volker et al. designed fluorescent organic molecules whose surface is functionalized with surface-active groups to provide high solubility, including co-nh2, -coo, -oh, and -nh2 [11]. the density-functional theory and configuration-interaction calculations show the structure-activity relationships of qds [10, 12]. a study reported by song et al. [13] stated that both the ethylene diamine and citric acid precursors undergo a thermal reaction resulting in the formation of bright blue fluorophore products, allocated to understanding the mechanism of photoluminescence in the specific molecular species. different combinations of different materials with citric acid lead to the formation of different kinds of molecular fluorophore species [13]. the study by zhou et al. involves the preparation of the most refined form of high fluorescent graphite carbon nitride qds with urea and sodium citrate under high temperature in an autoclave, resulting in the formation of surface decorated with negatively charged oxygen-rich surface active groups [14]. synthesis mechanism of carbon dots there are two main approaches to synthesizing carbon dots: top-down and bottom-up (figure 1). laser ablation techniques and electrochemical methods are classified as more popular top-down approaches. in this method, the micron-sized carbon structures divide into (or) break down into nanosized crystals ranging less than 10 nm. the preparation of blue luminescent nanocrystals via the electrochemical method was reported by zhou et al. [15], for which multi-walled carbon nanotubes (mwcnts) were used as the precursor. the laser ablation technique is generally used to amalgamate nanoparticles in the liquid medium by modifying solid particles using laser pulses. this made it convenient and understandable for various laser parameters such as irradiation time, pulse duration, and wavelength in the solid and solvent features. reyes-contreras et al. [16] synthesized carbon nanoparticles with the size ranging from 4 to 20 nm using acetone as a liquid media by targeting graphite with a pulse of neodymium-doped yttrium aluminum garnet (nd-yag) laser pulse with an emission wavelength of 1064 nm. the excitation wavelength can be altered to covert this from blue to yellow. the nanosized crystal structures are synthesized via polymerization reaction or conversion of suitable atoms and molecules into large nanostructures using different synthesizing procedures such as microwave-mediated synthesis techniques, hydrothermal, pyrolysis synthesis, ultrasonication method, etc., classified under the bottom-up approach. zeng et al. [17] fabricated green-emitting carbon dots with a size of 2-6 nm utilizing citric acid and urea by microwave synthesis. cds are showing exuberant in fields of imaging and delivery, yet still, it is very important to synthesize the particle to attain a uniform shape and size. admet & dmpk 00(0) (2023) 000-000 theranostic applications of carbon-based quantum dots doi: https://doi.org/10.5599/admet.1747 3 figure 1. different approaches for carbon dot synthesis top-down synthetic approach electrochemical procedure the electrochemical method is the heterogeneous method of synthesizing nanoparticles with a highly efficient and controllable process to obtain the purified and controllable particle size and shape. the preparation of cds using this method utilizes the chemical cutting process of the carbon material. this technique mainly obtains a high quantum yield, higher purity, and cost-affordable cds. deng et al. [18] synthesized the red and blue cds by the electrochemical method utilizing ethanol, which is low molecular weight alcohol. oh plays a major role in obtaining the cds with high quantum yield in alkaline conditions naoh/etoh. the synthesized particle sizes were 2.1, 3.5, and 4.3 nm, and the particles were confirmed with a peak at 23° in xrd [18]. qingxiao et al. [19]. synthesized cds by an electrochemical method using ophenylenediamine (opod) as the carbon source during this process, it undergoes many changes to obtain the yellow color emission of cds at the platinum anode [19]. ray et al. [20] synthesized the cds using nitric acid oxidation of carbon soot, having a nanocrystal size ranging from 2-6 nm, by applying a temperature of 100 °c for 12 hours. qiao et al. [21] synthesized fluorescent carbon nanocrystals with a uniform size of 1.9±0.3 with blue color and 3.2±0.5 with yellow color using the graphite electrode and calomel electrode in nah2po4 as medium. liu et al. [22] synthesized the cds using the electrochemical method with an average size of 4.0±0.2 to help in the detection of the fe3+ ions in tap water. at room temperature, colorless cds convert to yellow upon oxidation. photoluminescence quenching of cds occurs due to the functional groups' presence on the cds, which helps the fe3+ to bind with the outer layer of carbon dots, leading to quenching of the compounds [22]. niu et al. [23] prepared nitrogen-doped cds using an electrochemical method applying pyrocatechol as the precursor for the detection of alkaline phosphatase and pyrophosphate anions. this method does not require harsh reaction conditions like higher temperatures or toxic chemicals; at the same time, good crystallinity of cds can be maintained. in this method, the yield of particles is very low due to intercalation, and the process is very slow due to the accumulation of ions. https://doi.org/10.5599/admet.1747 pemula gowtham et al. admet & dmpk 00(0) (2023) 000-000 4 laser ablation the laser ablation technique is used for the breakdown of bulk materials into simple nanoscale materials, which are present in the liquid or gaseous environment using the laser beam on the targeted material, which results in the formation of fine ablation of materials [24-26]. laser ablation is the technique that uses light sources to excite the atoms from the lower excitation region to the higher excitation region, and the atoms trigger the initial process for the formation of crystals. material ablation occurs when the material and the laser interact due to the coupling of electron and photon interactions with the free valency of electrons reaching the higher excited state. the size of the obtained materials can be tuned by controlling the parameters such as pulse repetition rate, pulse width, energy, and solubility factors for the formation of particles of different sizes [27]. carbon dots obtained by immersion of graphite in a liquid medium of ethylenediamine or poly(ethyleneimine) using the nd:yag laser obtained a particle size of 1 to 3nm, which is spherical in shape, was synthesized by kaczmarek et al. [28]. reyes et al. [29] synthesized the carbon dots targeting the carbon source in the acetone medium by inducing the nanosecond laser pulses of nd:yag laser, which was targeted to the specific part with the different wavelengths by using the infrared lens and uv-visible lens. nguyen et al. [30] synthesized blue and red cds using the ablation technique and an infrared laser with 800 nm in polyethylene glycol as the medium to obtain cds from graphite powder. the size of the cds depends on the laser dependence, and it emits high photoluminescence (pl) when treated with low laser influence upon them. the size of the carbon dots is inversely proportional to the irradiation time [30]. yu et al. [31] synthesized cds with toluene as a precursor in the presence of the nd:yag laser with the wavelength produced in different sizes by slightly different ranges in the pulse. formation from the higher was used for obtaining the nano-size particles from the large compounds by using a laser source that stimulates and triggers the photons to be used as energy for the amplification. the preparation of cds by this method has potential photoluminescence properties that are highly stable, tunable, and bright are the advantages. the disadvantages include complex procedures, low output, and the size distribution of the particles that are not uniform. the process needs additional passivation, oxidation, and irradiation to control the products. ultrasonic treatment the ultrasonic treatment mechanism is characterized by the frequency of ultrasound waves generated in the liquid medium, known as acoustic cavitation. when these waves pass in the liquid medium leads to generate microbubbles with high intensity to create the local hotspots by generating an inertial effect by continuous longitudinal waves and refractions. obtaining high pressure and temperature in the surroundings enhances the molecular breakdown of the compounds [32]. the ultrasonication method of synthesizing the cds was mainly dependent on the parameters like sonication time and the amplitude for obtaining the specific size. kumar et al. [33] synthesized yellow cds ranging between 2-9 nm using polyethylene glycol by one-step synthesis. the particle size varies as the sonication amplitude varies, and from which cds of different sizes can be obtained. as the sonication amplitude increases, emission intensity and curve increase [33]. dehvari et al. [34] synthesized nitrogen-doped carbon dots as a probe using a crab shell by ultrasonication irradiation. the one-step sonochemical approach produced a product with a particle size of 8 nm and 14.5 % of fluorescence quantum yield. the product was highly soluble with functional groups over the surface. due to the targeting moiety, it featured both therapeutic and diagnostic applications [34]. li et al. [35] used a one-step ultrasonic method for synthesizing blue cds ranging from the natural precursor. the size range was around 5 nm. the particle exhibited theranostic potential due to high loading capacity and enhanced photoluminescence properties [35]. wang et al. [36] synthesized upconversion cds using bisbo4 via an ultrasonication process to enhance the photocatalytic properties and chemical stability by redox properties, admet & dmpk 00(0) (2023) 000-000 theranostic applications of carbon-based quantum dots doi: https://doi.org/10.5599/admet.1747 5 which would, in turn, improve the degradation of organic pollutants. studies conducted on the top-down mediated synthesis of cds are compiled in table 1. ultrasonic treatment is a low-cost and green method to produce nanomaterials that have the disadvantages of getting ultrasonic reactors for the large-scale production of cds. table 1. summary of top-down approaches used for the preparation of cds for various biomedical applications carbon source/ precursor method size, nm emission / excitation, nm quantum yield, % application ref. histidine hydrochloride electrochemical 3-5 505/420 33.8 selective and sensitive for quantitation of cu2+ ions [37] acetonitrile/bmimpf6* electrochemical 3 422/365 13.3 imaging of cells and detection of ferric ion [38] sodium citrate and urea electrochemical 1-3.5 433/351 11.9 sensing of hg2+ [39] toluene laser ablation 4.2 475/330 18 monitoring fluorescence change to enable real-time analysis [31] graphite and acetone (liquid medium) laser ablation 4-20 520/440 biomedical applications, due to its tunable photoluminescence properties [16] graphite flakes laser ablation 3-13 430/365 12.2 [40] polyethylene glycol (peg) sonochemical process ultrasonic irradiation 2-7 460/370 16 promising in bioimaging and solar cell applications [33] glucose ultrasonic treatment <5 450/350 7 potential candidate in fluorescence markers, biosensors, drug delivery, and biomedical imaging. [35] ascorbic acid and ammonia ultrasonic treatment 10 514/385 to replace the current selfassembly device and its technology [41] gelatin ultrasonic treatment 3.8 450/365 33.8 cellular imaging and anticounterfeiting ink [42] graphite oxidative cleavage 1.4-4.2 500/350 photodetector device [43] *bmimpf6 1-butyl-3-methylimidazolium hexafluorophosphate bottom-up approach microwave-mediated synthesis microwaves range from 0.3-300 ghz, where microwave uses the heating theory for the organic reactions, which occurs in closed reactors using dielectric heating effects that accelerate the rate of chemical reactions [44]. the organic molecules mainly depend on the polarity to absorb the heat energy and to convert the molecules or the distribution of electric charge over the atoms. a dielectric material contains either permanent or induced dipoles when suspended between two electrodes acting as capacitors. the materials have either a positive or negative charge that is stored on the surface of the molecules, which undergo molecular rotation when irradiated with the microwave [45]. the microwave-mediated synthesis has advantages compared to conventional heating synthesis as it gives rise to more heating capacity when molecules undergo a chemical reaction. the conventional method needs a long period of time to make the reaction medium reach the required temperature, and at times it may also lead to the degradation of the reactants. in contrast, microwave-assisted synthesis offers a fast reaction time through bypasses the heating of microwaves on the reaction vessel. microwave-assisted synthesis of carbon dots has surface modification with different solubilized functional groups such as carboxyl, hydroxyl, or amines, making them highly https://doi.org/10.5599/admet.1747 pemula gowtham et al. admet & dmpk 00(0) (2023) 000-000 6 soluble. some of the heterocyclic aromatic amines are known to be carcinogens. lopez et al. [46] synthesized the cds using the lactose by microwave process, where hcl is used as precursors and given a microwave treatment for 15 min. the transmission electron microscope (tem) was used to determine the size of obtained cds and to study the details of functional groups on the surface. the cds were around 10 nm in size, while the surface shows -oh and c=o functional groups. these functional groups account for excitationdependent fluorescence with maximum emission at 450 nm and excitation at 350 nm with an intensity of 13 % [46]. zhao et al. [47] worked on the synthesis of cds using ethylenediamine as a nitrogen dopant via a microwave-mediated method by heating the solution in a microwave oven for around 10 mins. tem analysis of the cds revealed the size range, about 2-5 nm, and then the surface groups that showed the presence of o-h and n-h, c-h, c=o, and c-n functional groups, contributing to water dispersibility. the maximum emission and excitation spectra were 450 and 370 nm, respectively. as the ph values increase, the fluorescence intensity of cds decreases, and the presence of metal ions such as co2+ shows a brown-colored cd. at the same time, it remains the same for the other metals, which helps in the detection of water samples [47]. xu et al. [48] synthesized cds doped with sulfur and nitrogen using glycerol with one pot reaction mechanism, a high dielectric constant as precursors, and cystine as the source was heated in a microwave oven for 6 mins. a size range from 1.5-5.5 nm was obtained with the hydroxyl functional groups on the surface, leading to an aggregation-induced enhancement effect resulting in the real-time detection of hg2+ [48]. this method of processing cds involves no contact heating since the electromagnetic field interacts directly with molecules to create energy. in addition, cds are produced with a high yield due to lesser reaction time, environmental friendliness, uniformly distributed heating and energy saving modes. it is generally not possible to use microwave reactors for large-scale reactions, since they are confined to small volume. a low boiling point solvent is not suitable in this procedure due to the pressure and temperature limitations. thermal decomposition thermal decomposition is an endothermic reaction that generally uses heat as the reactant for the breakdown of chemical compounds into molecules, and this technique is used to synthesize cds through the bottom-up method. the major benefits of this method are less time taking process with inexpensive, efficient, and large-scale manufacturing [4]. ludmerczki et al. [49] fabricated luminescent cds with citric acid and its intermediate derivates, which helps produce highly fluorescent cds. here, citric acid monohydrate was heated to 180 °c, and the sample was collected from colorless to light yellow to orange-yellow. the surface modification was done using 3-(aminopropyl) triethoxysilane (aptes), which helps in ph shifting, usually from acidic to basic, and obtained a peak at 450 nm, and the size of the particle was around 5-15 nm [49]. wan et al. [50] utilized a thermal method to synthesize cds with 1-butyl 3-methyl bromide imidazolium and i-cysteine at 240 °c. with the afm study, the height of cds was found to be around 1.0-3.5 nm. this method was used to determine the cds from the small organic molecules [50]. the method is easy to perform, and the physicochemical properties, including optical properties, can easily be tuned. the formation of impurities during the decomposition process can affect cds' quality, which needs further purification. carbonization synthesis the carbonization process is adopted to synthesize large carbon content with solid residues formed from organic materials. this reaction usually takes place in an inert condition through pyrolysis. the carbonization process is the conversion by progressive heating of the organic macromolecular system to a macro-atomic network of carbon atoms. molecular precursors undergo carbonization, which is the best, cheapest, and onestep process for synthesizing cds [4]. shi et al. [51] synthesized n, s doped cds via carbonization technique. admet & dmpk 00(0) (2023) 000-000 theranostic applications of carbon-based quantum dots doi: https://doi.org/10.5599/admet.1747 7 this procedure involves the reaction of citric acid and l-cysteine to form tpca (5-oxo-3,5-dihydro-2hthiazolo[3,2-a]-pyridine-7-carboxylic acid) as a primary component. tpca is responsible for the production of cds, which emit strong fluorescence with a high quantum yield (figure 2). figure 2. schematic representation of the mechanism of tpca formation lin et al. [52] synthesized high fluorescence quantum yield cds from l-cysteine and citric acid with a size ranging from 2-4 nm, which helps increase fluorescence for cellular imaging (figure 3). figure 3. schematic representation of tpca formation using citric acid and l-cysteine as the precursor qin et al. [53] synthesized cds using uric acid by carbonization method. the obtained cds ranged from 1.3 to 7.9 nm with an excitation wavelength of 350 nm, followed by an emission wavelength of 402 nm with a quantum yield of 52.06 % [53]. using citric acid as a medium, wang et al. [54] synthesized the blue luminescent emitting carbon dots with a range of 4.8-9 nm by thermal reduction. this enhances the quantum yield compared to the normal non-reacted cds, which were analyzed by a thermogravimetric analyzer [54]. carbon dots are uniformly formed using a controlled carbonization reaction, and unlike other synthesis methods, they do not make many impurities. this method needs a longer reaction time, and cds yield is also very low. pyrolysis synthesis method the pyrolysis method is also a type of thermal decomposition where the chemical transformation of a sample occurs on heating to high temperatures in an inert environment condition. it is an irreversible process where the organic materials undergo decomposition under high temperatures and controlled pressure. during this process, the organic materials undergo both physical and chemical changes resulting in the formation of carbon residue. ma et al. [55] developed blue-emitting cds doped with nitrogen using citric acid and ethylenediamine at 170 °c through the pyrolysis method with an average diameter of 2-6 nm in the range with a quantum yield of 88 %. rimal et al. [56] fabricated cds via the pyrolysis technique of bottom-up synthesis using oleic acid as an organic substrate at 260 °c. the obtained cds were in the range of 5-10 nm https://doi.org/10.5599/admet.1747 pemula gowtham et al. admet & dmpk 00(0) (2023) 000-000 8 with a quantum yield of 50 %, which also combines with an acetate polymer matrix to form polymer composites that are stable at high temperatures [56]. using glycerol solvent as the precursor, a new type of cd was synthesized using the pyrolysis method, where it was heated up to 230 °c, ranging from 2-4 nm. the prepared cds were passivated by polyethylene glycol (peg), which increases biocompatibility. the thermogravimetric analyzer was used to identify the heat loss, and the intensity was recorded as 14.3 %, as studied by lai et al. [57]. tang et al. [58] synthesized blue-emitting cds using lactic acid ionic liquid via a onepot pyrolyzing technique. according to tem results, obtained cds ranged from 4.4 nm and the emission of 454 nm and excitation of 365 nm, which gives an intensity of 16 % [58]. studies that have been conducted to synthesize cds via bottom-up approaches are compiled in table 2. the pyrolysis method is cost-effective and does not need high-quality reagents and arrangments. the morphology and the size of cds can be controlled easily by varying the reaction conditions. scaling up the synthesis of cds and determining the growth temperature are the major issues of this method. table 2. summary of bottom-up approaches used for the preparation of cds for various biomedical applications carbon source (precursor) method size, nm emission/ excitation, nm quantum yield, % application ref. lactose, hcl microwave 10 450/350 chemical sensing of heterocyclic aromatic amines [46] glucose, glutamine and imidazolium microwave 7-15 efficient metal-free catalyst in the production of h2o2. [59] citric acid microwave 1.5-4.5 530/450 antimicrobial photodynamic therapeutic applications [60] citric acid and diethylenetriamine (deta) thermal decomposition 3-5.5 456/360 88.6 fluorescent chemosensor detection of chromium (vi) [61] dicyanamide (dcd) and citric acid thermal decomposition 8-16 528/410 73.2 detection of iron and fluorine ions [62] citric acid and aeapms thermal decomposition 1.4 450/360 47 for bioimaging and biosensing applications [63] l-dopa and n, n dimethylformamide carbonization 3.27 400/320 48 used as a probe for the live cell imaging [64] glucose carbonization 1-7 498/400 48 turn-on biosensor for ovalbumin and in vivo imaging applications. [65] sago waste pyrolysis 37 390/315 probe for heavy metal detection [66] n-hydroxysuccinimide pyrolysis 20-30 420/350 14-31 used in sensors and highperformance optoelectronic devices [67] forestry lignocellulosic waste pyrolysis 20-80 438-473/ 350 28 sensor and contrast agent [68] synthesis using hydrothermal method hydrothermal synthesis is referred to as synthesizing of nanoparticles using chemical reactions in an aqueous medium at high temperatures. generally, hydrothermal synthesis of chemical reactions occurs in a closed autoclave with high temperatures (100-1000 °c) and pressure (1-100 mpa). the hydrothermal synthesis interfaces with solids and solutions to form nanoparticles during phase transfer. the separation mechanism mainly depends on the reactants, chemical reactions, conditions, and the relation between the structures and properties of the compounds [69]. in hydrothermal synthesis, the reaction mechanism mainly depends on the reactant molecules that react with the solution giving rise to different products with different structures even if the same reactants were used [70,71]. this is also very useful in the chemical combination of compounds that cannot be done using solid-state synthesis. the particles synthesized using this method admet & dmpk 00(0) (2023) 000-000 theranostic applications of carbon-based quantum dots doi: https://doi.org/10.5599/admet.1747 9 have unique physical and chemical properties due to their aggregation, condensed state, valance state, and structure, making the compound grow in a single crystal with controlled size and structures (figure 4). depending on the temperature used in the hydrothermal synthesis, they are again classified into subcritical and supercritical synthesis. the subcritical synthesis reaction occurs between 100 and 300 °c, whereas the supercritical synthesis reaction goes up to 1000 °c [72,73]. different organic solvents have different soluble properties, which change due to high temperature and pressure. the hydrothermal system mainly accelerates the reaction rate and intense the hydrolyzation reaction, which leads to changes in the redox potential of the reactants. the medium used in hydrothermal synthesis plays a major role in the obtained products, as various organic solvents dissolve the reactants and form a solvent-reactant complex [74]. this affects the rate of the chemical reaction, and hence, selecting the organic solvent is critical since it plays an important role in the reaction process. the solvent can be selected based on molecular parameters such as melting point, boiling point, density, molecular volume, rate of evaporation, dielectric constant, molecular weight, and solvent polarity. the hydrothermal method has become the most important technique for the fabrication of nanomaterials, such as nanohybrids and nanocomposite materials. temperature, °c figure 4. schematic representation of the mechanism of hydrothermal mediated carbon dot synthesis the nanoparticles synthesized using this method possess chemical homogeneity and solubility with monodispersed size and shape. researchers developed many strategies for obtaining the desired nanostructures by controlling the hydrothermal conditions (table 3). many factors important in obtaining the desired structures are internal factors such as the nature of reactants, concentration, ph value of the reaction system, and pressure of the reaction system. other external factors, such as reactive time and temperature, also reflect in the structural aspect [74]. hydrothermal synthesis of cds is cost-efficient and easy to synthesize from different components such as amines, saccharides, organic acids, and their derivatives [4]. cds are modified with various chemical components to enhance their photoluminescent properties. photoluminescence mainly depends on two major principles: 1) the carbon core with modified functional groups and 2) the surface state, which is determined by the hybridization of the carbon backbone and connected chemical groups. in the hydrothermal method, various precursors have a dominant role in synthesizing cds to obtain different sizes and colors when doped with surface passivating agents. based on the surface passivated compounds, the absorption properties vary from one cd to another. cds are prepared with many precursors, including citric acid, urea, sodium citrate, ethylenediamine, and glucose. generally, the organic precursor will be taken in an autoclave and reacted with hydrothermal reactors at high temperatures [8]. the properties of cds depend on carbonization, an endothermic process that relays directly on the temperature. the temperature should not be too high or low, as the high temperature degenerates the surface structure due to over-oxidizing the molecules, whereas the low temperature leads to the formation of defective cds since the optical properties directly depend on the reaction time and temperature [5,75]. this part will discuss various studies involving a hydrothermal technique for synthesizing cd with a minimal size range and high quantum yield for various biomedical applications. https://doi.org/10.5599/admet.1747 pemula gowtham et al. admet & dmpk 00(0) (2023) 000-000 10 table 3. summary of the hydrothermal method used for the preparation of cds for various biomedical applications carbon source (precursor) size, nm emission / excitation, nm quantum yield, % application ref. citric acid, ethylenediamine, and methyl blue 1.86 440/350 68.0 fluorescent probes for sensing mercuric ions and chlorine monoxide ions in tap water samples [93] glucose, m-phenylenediamine 8 460/320 17.5 good candidates for fluorescent probes of fe3+, cro4-2 and as cell labeling reagent [94] thelephora ganbajun zang 1.6-7.5 444/356 9.44 environmental pollutant detection2,4dinitrophenol(2,4-dnp), and 4-nitrophenol (4-np) in water and soil samples [95] liquid products of biodegradation of coal (lpbc) 3.5 442/328 0.54 ph and temperature-based sensing applications. [96] cysteine, trisodium citrate dehydrate, and ethylenediamine 3.82 442/350 66 bioimaging of human ovarian cancer cells and fluorescent determination of baicalein [97] salvia hispanica l. (chia) 5.4 415/310 17.8 targeted treatment of kidney-related diseases [98] persimmon peels 2 350/256 [99] folic acid, glycerol 5.1 442/365 25.3 detection of cu2+ in water samples [100] citric acid and l-histidine 3.9 414/340 22 used to analyze chlorogenic acid in coffee and honeysuckle [101] gardenia fruit 2.1 450/360 10.7 detection of hg2+ and cysteine [102] sodium citrate, triethylenetetramine, rose bengal 525/440 chemosensors for hypochlorous acid detection [103] citric acid monohydrate, 2-aminopyridine 9 421/310 18 bioimaging of candida albicans [104] l-lysine, thiourea 6.86 365/300 53.19 detection of picric acid in water [105] p-phenylenediamine 3.2 580/520 a strong antibacterial agent inhibits the growth of s. aureus and e. coli [106] jackfruit peel and tamarind peel 5.3 430/350 13.04 multifunctional ability, including anticancer activity [107] diethylenetriamine and trans-aconitic acid 28 435/365 81 highly selective and sensitive towards fe3+ [108] highland barley 5.8 480/440 14.4 used for detection of hg2+ [109] biomass of dunaliella salina 4.7 415/340 8 used for intracellular detection of hg (ii) and cr (vi) [110] palm kernel shells, lphenylalanine 2 430/360 13.7 [111] l-arginine, phosphoric acid 2.4 444/340 18.38 efficient fluorescence sensor for detection of vitamin b12 [112] o-phenylenediamine and semicarbazide 4.71 625/540 23 optoelectronics and forensic science [113] fluorescent cds have recently gained tremendous attention, yet there is no proper evidence to support the origin of fluorescence in cds. citric acid is a more commonly used precursor in synthesizing carbon dots [76,77]. citric acid, when treated with ammonia, forms fluorophore citrazinic acid through a hydrothermal reaction process. the citrazinic acid derivatives play a major role in the photoluminescence property in cds. a study reported by schneider and team used citric acid and hexamethylenetetramine (hmta) for the preparation of cds via hydrothermal reaction, which released ammonia at high temperatures and formed citrazinic acid derivatives fluorophores [78]. but the obtained cd-hmta compounds exhibited low fluorescence, which was explained in terms of the decomposition rate and strength of nucleophile (nh3) (figure 5). both characteristics decide the fluorescence of a compound, and they are directly proportional. admet & dmpk 00(0) (2023) 000-000 theranostic applications of carbon-based quantum dots doi: https://doi.org/10.5599/admet.1747 11 figure 5. schematic representation of the reaction of citric acid with ammonia many reports drew a path suggesting that fluorescence cds peak can be varied by changing the excitation wavelength range and forming different emission paths, which was unclear. cds can be passivated with different functional groups, such as carbonyl and hydroxyl groups, which can be used as trapping agents, making cds emit various colors in different excitation wavelength ranges. temperature plays a major role in the surface groups' passivation and emitting fluorescence. in other cases, the emission also occurs through a radiative transition of sp2 carbons in the amorphous sp3 matrix. the cds with the rich source of amino-group are desirable for their controlled surface passivation and thus exhibit excitation-independent fluorescence due to the surface-coated functional groups [79]. zholobak and team fabricated cds with urea and citric acid as a precursor in a ratio of 1:3 at 133 °c for the formation of blue emission and green emission cds, where the blue emission cds were obtained from the citrazinic acid and the green emission was due to the formation of citrazinic amide [80]. urea acts as a carbon source and as a surface amino functional group on the surface of cds and can contribute to changes in the luminescence properties. kasprzyk et al. [81] obtained cds using citric acid and urea from the open and closed reaction systems. ammonia produced during this reaction plays a major role in the emission of cds (figure 6 ). figure 6. schematic representation of the reaction of citric acid and ammonia to form cds https://doi.org/10.5599/admet.1747 pemula gowtham et al. admet & dmpk 00(0) (2023) 000-000 12 the deeper and more detailed information about these cds' chemical structures and surface states has been studied using nmr spectroscopy and electrospray ionization mass spectrometry. this study confirms the correlation between the shifting of pyridone nitrogen and imide nitrogen with hydrogens in fluorophore structures. the interpretation of the chemical structure of the green fluorescence particle formed 4-hydroxy1h-pyrrolo[3,4-c]pyridine-1,3,6(2h,5h)-trione (hppt), which was first reported by this group [82]. hppt can be formed by using two main pathways (figure 7). one possible formation involves citrazinic amide or acid from citric acid and urea. further, the urea was broken-down to obtain isocyanic acid, which was then added to the pyridine ring at position 3. the hppt was then formed after closing the imide ring. the other route follows the creation of an intermolecular complex of amide between citrazinic and acid urea. after the formation of intermolecular amide, intramolecular condensation and cyclization occurs by excluding ammonia [82]. in the first mechanism, isocyanic acid is obtained from urea by the thermal-induced mechanism in the presence of water. in the second mechanism, an end product citrazinic acid, with a blue emission, is formed due to the reaction between citric acid and urea in an aqueous medium. it was then transformed into green-emitting hppt after the evaporation of water. in both mechanisms, citrazinic acid is a byproduct to form hppt, whereas it can also be obtained directly by heating urea and citrazinic acid [82]. figure 7. schematic representation of the mechanism of hppt formation song et al. [13] worked on the synthesis of fluorescent cds prepared with precursor such as citric acid and eda. the team utilized the hydrothermal method by heating the solution at a constant temperature of 140 °c for 10 h. the prepared cds ranging from 2-6 nm led to the formation of fluorescence molecules (1,2,3,5-tetrahydro-50oxo-imidazo[1,2-a] pyridine-7-carboxylic acid, ipca) via amidation reaction to form the branched polymers. further, it undergoes a carbonization reaction to obtain the amorphous matrix (figure 8) [13]. liu et al. [83] fabricated cds using metronidazole through a hydrothermal approach by heating them to 250 °c for 8 h. the prepared cds attained a size of 2.9 nm with a quantum yield of 28.1 %. the nitrogen-rich functional groups on the surface were responsible for antibacterial activity resulting from the inhibition of the growth of obligate anaerobes and favored variegated bioimaging [83]. admet & dmpk 00(0) (2023) 000-000 theranostic applications of carbon-based quantum dots doi: https://doi.org/10.5599/admet.1747 13 figure 8. schematic representation of the reaction of citric acid and eda to form ipca carbon dots were formed using citric acid and ethanolamine, which forms derivatives of citrazinic acid and exhibits high fluorescence due to the organic functional groups passivated on the surface of cds [84]. the photoluminescence behavior of the amorphous mixture containing sp2 and sp3 carbons supports the photogeneration of electron-hole pairs that induces the radiative recombination of localized trap carriers which locks the sp2 clusters bounded by sp3 defects and makes the mechanism remains stable in the presence of heteroatom groups. wang et al. [84] synthesized blue-emitting nitrogen-doped cds using ethanolamine ionic liquid as the precursor via the hydrothermal method by applying the temperature of 200-240 °c for 4 h. the dispersion was dialyzed using 1000 da for three days to obtain a particle size of 3.4 nm with rich hydroxyl and nitrogen-rich functional groups on the surface with a quantum yield of 24.7 %. the fluorescent probes were useful for detecting heavy metals in tap water [84]. li et al. [85] synthesized the cds using citric acid with polyethyleneimine via a one-pot hydrothermal reaction for the detection of the moraceae family compound called morin for double mode detection in human urine. the quantum yield of obtained cds was recorded as 48.3 %, emission range of 459 nm, and a size range from 3.0 to 6.0 nm when heated at 110 °c for 2 h [85]. lin et al. [86]synthesized cds doped with phosphorus and nitrogen using citric acid and o-phosphorylethanolamine by hydrothermal method for detecting cadmium ions in serum and urine samples based on the chelationenhanced fluorescence (chef) mechanism. the preparation involved heating up to 180 °c for 12 h, and it was further filtered using a membrane filter to obtain a particle size of about 1-2 nm [86]. a study by wang et al. [87] developed cds with hydrochloric acid as a catalyst using the combination of l-tryptophan and lphenylalanine at 200 °c for 2 h with a quantum yield of 21 %. the obtained particle measured about 4.8 nm, and the surface was passivated with hydrazine functional groups and emitted blue fluorescence at 435 nm [87]. liu et al. [88] synthesized blue-emitting nitrogen-doped cds using citric acid and ethylenediamine as precursors, which emit fluorescence at 465 nm. the mixture was heated to 150 °c for 2 h. the hydroxyl groups and carboxyl groups on the surface induced a negative charge. the quantum yield was recorded as 58.6 %, and the size of the particle was found to be 6 nm. the compound was synthesized as a chemosensor to detect metals in water samples [88]. another group studied the impact of temperature on the fluorescence emission range. the procedure utilized folic acid and sodium hydroxide. the mixture was then heated at a different temperature from 170 to 200 °c for 5 h. as expected, the fluorescence emission shifted from 447 to 462 due to temperature variation. the optimum temperature used was around 190 °c. the carbonization rate decreased due to a decrease in temperature. the particle size ranging from 3-6 nm was confirmed by using transmission electron microscopy [89]. wang et al. [90] synthesized cds hydrothermally, using catechol as the carbon source. in the presence of water, the compound was heated to 200 °c for 24 h and placed in a dialyzed membrane to obtain the nanoparticles in the 4.7 nm range. the surface of cds was passivated with the residual catechol, https://doi.org/10.5599/admet.1747 pemula gowtham et al. admet & dmpk 00(0) (2023) 000-000 14 carboxyl, and hydroxy functional groups, and the maximum emission of the particle was recorded at 345 nm with a quantum yield of 32 %. this work was done for the identification of hypochlorite in the water samples [90]. the carbon dots can also be synthesized via the green synthesis technique. atchudan et al. [91] worked with this idea and synthesized cds using banana peel with water as the solvent. the procedure involves an autoclave and heating up to 200ºc for 24h. the obtained particles emitted blue fluorescence at 365 nm and size was measured at about 5 nm with a quantum yield of 20 % [91]. the same group reported the green synthesis of cd using the juice of phyllanthus emblica (indian gooseberry). the nitrogen-doped blue emission cds were obtained when heated for 12 h at 200 °c. the emission maximum was seen at 411 nm with a particle size of 5 nm, and a quantum yield was calculated as 13.5 %. the obtained compounds were used as fluorescent ink in bioimaging and drug delivery applications [92]. applications of cds nanoparticles are gaining interest for their multimodal applicational approach in healthcare industries [114-116]. they are mostly used to formulate therapeutics and imaging agents to improve bioavailability and reduce toxicity [117-120]. carbon dots, with their high potential characteristic features such as biocompatibility, photostability, and easy surface modifications with high penetration into the cell surfaces, make them great alternatives to the current marketing drug carriers [5]. for the targeted imaging of cells and their structures, cds are prepared by surface modifications based on the specific targeted molecules to avoid immune responses to prevent them from detection as foreign bodies by the immune system (figure 9). these surface passivation of cds helps in vitro and in vivo applications. cancer cell imaging if left undiagnosed, cancer will lead to metastasis and death in a short period. as a consequence, it is very necessary to develop novel methods for identifying the early diagnosis of cancer. carbon dots were considered future candidates for imaging applications due to their unique solubility and low cytotoxicity properties. the surface-modified cds have the potential to pierce into various types of cancer cells by giving fluorescence imaging. several studies have reported the preparation of cd for diagnostic applications. zhou et al. [14] synthesized carbon nitride quantum dots using urea and citric acid, which were incubated with the hek-293t cells and monitored under the confocal microscope. red and green fluorescence was absorbed in the membrane and cytoplasmic areas and was excited at 543 and 488 nm [14]. zeng et al. [17] synthesized carboxylrich surface green-emitting cds, which showed high fluorescence in hepg2 tumor-bearing mice. this fluorescence also helped to deliver the drug and perform localized therapy [17]. zhong et al. [121] synthesized orange luminescent carbon dots using 1,2-benzenediamine and carbamaldehyde. the mixture was heated at 180 °c for 3 h, and the surface was passivated with the carboxyl groups, which react with the human lung cancer cells (a549). another study by lai et al. [57] used glycerol via a pyrolysis process to synthesize the cds surface modified using poly (ethylene glycol). this study used hela cancer cell lines incubated on a cover slip and observed under a confocal microscope at the wavelength of 405 nm. the cytoplasm exhibits blue fluorescence, which shows new paths for cellular imaging [57]. wang et al. [122] developed cds from beer for visualizing the mcf-7 breast cancer cell lines under a confocal microscope with excitation of 405 nm laser. there was no fluorescence in the control group, and a blue color appeared in the cell cytoplasm. the control group showed no fluorescence, and a blue color appeared in the cell cytoplasm. the nucleus was visualized in the range of 543 nm [122]. liu et al. [123] synthesized cds using folic acid as precursors to obtain a high quantum yield and target cancer cells. this helped in the receptor-mediated cellular uptake in hela cells as it binds to the surface and gives a high fluorescence intensity compared to the other cell lines. this study supports that cds can effectively target cancer cells by folate receptor-mediated processes [123]. admet & dmpk 00(0) (2023) 000-000 theranostic applications of carbon-based quantum dots doi: https://doi.org/10.5599/admet.1747 15 nucleus-targeted fluorescence imaging the nucleus is the major component of the cell as it carries hereditary information. visualization of a nucleus is more important to determine the structure and metabolic features of the cell. certain functional groups can be decorated on the surface to target and bind to specific molecules, tenable active targeting. the first study reported by datta et al. comprised synthesizing of cds with a positive charge and allowed them to combine with graphene to form a hybrid nanomaterial [124]. this helped in easy penetration into nih/3t3 cells through electrostatic interactions. in addition, ci et al. [125] synthesized nitrogen-doped carbon dots using ascorbic acid and poly(ethylenimine) via one-pot hydrothermal synthesis. the temperature here played a vital role in obtaining appropriate minimal particle size. the cds were co-incubated with hela cells in different concentrations and observed under the confocal microscope using cck8 assay. due to the presence of zwitter ions on the surface, the cds can easily penetrate the nucleus by leaving minor parts in the cytoplasm and other organelles. this shows the potential of the cds to be used as nucleus labeling [125]. another study reported by jung et al. [126]used beta-alanine as the zwitterionic agent on the surface of cds prepared with citric acid. the particles were able to target the nucleus, as the zwitterionic surface was more capable of being resistant to the non-specific adsorption of proteins and increased the cytoplasmic uptake and delivery to the nucleus region [126]. hill et al. [127] synthesized green emissive fluorescence cds using glucosamine and m-phenylenediamine via microwave synthesis procedure. 2,5-deoxyfructosazine was used as a passivating agent to target the nucleus of hela cells and human dermal fibroblasts. after the cells are internalized with cds, led-mediated selective killing takes place. the interaction with the dna causes an increase in intracellular temperature, leading to the breakdown of atp [127]. a green synthesis method reported that blue fluorescence emitting cds were prepared using citric acid and barley leaves. pk-15 and hela cell lines were used for cellular imaging. the results show that negatively charged cds are observed in the cytoplasm when the excitation is 405 nm, and the neutral cds are spread over the entire cell, including the nucleus. this study suggests that charge has a specific role in targeting organelles [128]. mitochondria, the powerhouse of the cell, help produce atp. the abnormalities in mitochondria may lead to many fatal diseases. cds have great potential to track cellular fate and mitochondrial disorders. a novel development of functionalized cds shows high permeability, solubility, and photostability. sun et al. [129] synthesized fluorescent carbon dots by microwave-assisted synthesizing with citric acid and n, n-dimethylaniline. the cds were labeled on hela cells and visualized under the confocal microscope with a laser wavelength of 488 nm. electrostatic interactions play a vital part in assembling the cationic charge of cds in the mitochondrial membrane, which has a high negative charge potential. when the mitochondria membrane potential (mmp) decreases, the cds migrate into the nucleus due to the interactions with nucleic acids [129]. figure 9. major biomedical applications of carbon dots cd as sensing materials carbon dots possess chemical stability, tunable surface functional properties, low toxicity, high solubility, and low-cost efficiency. all these properties make cds a great candidate in the biosensing field. when the https://doi.org/10.5599/admet.1747 pemula gowtham et al. admet & dmpk 00(0) (2023) 000-000 16 cds having excellent fluorescence properties are loaded with biosensor molecules, their tunable surface properties help to incorporate different functional groups, in turn, helps in the detection of specific biomolecules and shows excellent advantages in biomedical applications. zhang et al. [130] worked on the synthesis of cds using citric acid for the detection of fe3+. the nitrogen doped-cds show fluorescence quenching by non-radiative electron transfer, which can effectively detect fe3+. the detection limit was found to be 2.5 nm [130]. wang and team [131] synthesized cds hydrothermally with m-aminobenzoic acid for the identification of fe3+ in water samples. the surface functional groups directly correlate with fe3+ ions, and the detection limit was recorded as 0.05 µm [131]. mohammadi et al. [132] synthesized cds loaded with dna. citric acid and ethylenediamine were used as the precursors for the synthesis. the surface was passivated with dna probes for targeting the microrna155 in serum samples. the assay undergoes the fret mechanism, and the detection limit was 0.1 am. this study was conducted by using mcf-7 human breast cancer cell lines to measure microrna expression [132]. shan et al. [133] synthesized cds for the identification of hydrogen peroxide (h2o2). the surface of cds contains peroxidase-like activity for the identification of h2o2, and the same can be used for glucose detection at the detection limit of 10.0 mm [133]. drug/gene delivery carbon dots are used as potential drug carrier vehicles to the target site by making a slight modification over the surface of cds, and the pathways can be visualized due to their fluorescence nature. hence, these actively replace conventional fluorescent dyes. owning to this property, hailing et al. utilized the hydrothermal method for the synthesis of cds. the glycerol and pei-25k was used as precursor, while the surface was coated with polyethyleneimine to incorporate the anticancer drug over the outer layer of the cds through electrostatic interactions. hep3b and mhcc-97l cells were used to study the drug release, which confirms that the cell death rate increased in the cancer cells compared to the normal cells and showed strong inhibition against tumor growth [134]. chung et al. [135] have synthesized monodispersive cd-hydroxyapatite (hap) with sugarcane waste and sodium hydroxide via the hydrothermal procedure. the shape of the prepared cd was spherical. the obtained fluorescent cds were incorporated with the hap, which is used as a biomaterial to promote tissue regeneration. these cd-hap enhance the loading efficiency twice as compared to the normal carrier. the release profile was studied by the absorbance of the drug in phosphate buffer saline. the results show that the drug follows the diffusion mechanism, enhancing therapeutic efficacy [135]. sarkar et al. [136] synthesized fluorescence cds using aloe vera gel and chemically cross-linked them with polymeric hydrogels for controlled drug delivery of vancomycin, a tricyclic glycopeptide antibiotic to the gastrointestinal tract. the synthesized cds were further coated onto the surface of alginate film through the casting technique to improve thermal stability. the surface modifications on cds help to enhance more loading capacity of the drug onto the surface. the drug ß-vancomycin coated on cds was released through the ph-response in ph 1.5, similar to the stomach. the potential uptake of the drug was recorded as 96 %, whereas the release rate was recorded as 56 % in 120 h. this confirms that ß-cds have great potential as a drug delivery vehicle for oral administration [136]. tumor theranostics coupling therapeutic agents with ligands complementary to target biomarkers of the diseased site resulted in image-guided therapy to treat the condition better. adjuvant therapies like ptt, and pdt combined with an imaging system provide information about the size and site of the tumor but also engage in real-time monitoring admet & dmpk 00(0) (2023) 000-000 theranostic applications of carbon-based quantum dots doi: https://doi.org/10.5599/admet.1747 17 of the therapy (table 4). this type of monitoring of the therapeutic effect in the region will enable physicians to plan therapy better with increased therapeutic efficiency. decorating the surface of cds with appropriate functional groups will facilitate drug delivery at the targeted site. due to the significant fluorescence property of cds, they help to image the same. further, the cds can be modified to respond to external and internal stimuli, paving the way for more targeted delivery. this way, cds eliminate most of the adverse effects and are standalone nanoparticles. chen et al. [137] developed cds that could respond to dual stimuli and release the drug. cds were developed via the microwave-mediated method. the prepared particle showed redox and ph sensitivity. mesoporous silica nanoparticles/cd functionalized with multiple functional groups was developed as the carrier and loaded with dox as a sample drug. the comparison was made with the carrier with and without cargo. table 4. summary of carbon dots prepared for theranostic tumor applications no. carbon source (precursor) synthesis strategy doping cargo imaging modality therapy outcome ref. 1. edta hydrothermal technique nitrogen and nickel mri and photoacoustic and photothermal imaging ptt the as-prepared cds served as a probe for multimodal imaging and ptt agent for effective therapy against u14 tumor in a mouse model. the experiments showed that the particle was able to induce potent therapy against the tested cancer model. the particle possessed renal clearance, which could eliminate other long-term toxicity. this way, the particle assured biosafety inside a bodily system. [140] 2. p-phenylenediamine solvothermal method fluorescence and mr imaging chemodynamic therapy the cds was decorated with gadolinium and ferrous ions to obtain imaging capabilities. the particle was able to emit red fluorescence, which helped in fluorescence imaging. the ferrous ions helped to perform chemodynamic therapy by getting released at the tumor site. this was accomplished through a fenton reaction. the gadolinium ions were added to perform mr imaging. the study results revealed that the particle featured less toxicity and performed synergistic activity in the management of cancer. [141] 3. citric acid hydrothermal approach photosensitizer chlorine and copper ions fluorescence imaging pdt, ptt, and chemodynamic therapy the copper ions in the formulation helped to perform chemodynamic therapy, while the photosensitizer helped in pdt. the copper ions also achieved gsh depletion resulting in redox-responsive delivery. the in vivo and in vitro results indicated that the formulation performed flmediated tumor therapy. [142] 4. aminosalicylic acid hydrothermal method doxorubicin fluorescence imaging chemotherapy a yellow emissive cds was obtained hydrothermally. the cds was loaded with anticancer drug and coated with cancer cell membrane. this coating helped to escape immune cells such as macrophages and also to target the cells. the formulation exhibited redox sensitive drug release also. the results from the experiments showed an effective image-guided therapy against cancer. [143] https://doi.org/10.5599/admet.1747 pemula gowtham et al. admet & dmpk 00(0) (2023) 000-000 18 no. carbon source (precursor) synthesis strategy doping cargo imaging modality therapy outcome ref. 5. citric acid solvothermal method gadolinium as1411 aptamers fluorescence and mr imaging ptt the aptamer-conjugated cd nanostructure exhibited the emission of red fluorescence. the doping material made the nanostructure as an imaging probe suitable for mr and fluorescence imaging. the particle also performed ptt due to its photothermal conversion ability. [144] 6. cyanine dye solvothermal process fluorescence imaging ptt the in vivo and in vitro study results showed that the particle synthesized possessed high photostability. [145] 7. dhca and eda hydrothermal method gadolinium doxorubicin and ir825 mr imaging chemotherapy and ptt the as-synthesized cds emit blue fluorescence. the doping material made the formulation serve as a probe for mr imaging. the in vivo experiments conducted on breast cancer-induced mice model showed excellent ptt and chemotherapeutic effect with good tumor inhibition rate. [146] 8. o-phenylenediamine and lysine hydrothermal method n and s tp fluorescence imaging ptt and pdt the study compared the cds with different doping for the treatment and management of cancer. the results suggest doping of cds with n and s yields better therapeutic characteristics compared to the cds doped with n atoms alone. the efficiency of oxygen-producing ability and photothermal conversion of the cds co-doped with n and s atoms also was comparatively high, providing better theranostic efficiency. [147] 9. polydopamine and folic acid hydrothermal method n doping psmadirected fluorescence imaging ptt the as-synthesized particle showed good photothermal conversion ability and helped to perform photothermal therapy. the therapeutic approach was mediated via temperature-induced apoptosis. the in vitro study with various cell lines indicated that the n-doped cds functionalized with folic acid ligands induced high cytotoxicity to the prostate cancer cells. [148] 10. watermelon juice hydrothermal method optical imaging ptt upon irradiation with nir laser, the cds prepared exhibited excellent photothermal conversion efficiency and executed photothermal therapy against cancer in vivo. the superior optical properties of the obtained cds made synergistic optical imaging possible. [149] the cds produced in the work showed nir fluorescence that served as the probe for fluorescence imaging, where the surface functional groups provided targeted drug delivery leading to theranostic applications. the silica nanoparticles were used to provide biocompatibility [137]. luo et al. [138] designed a nanohybrid system with cds doped with iron through a solvothermal approach for tri-therapy with multimodal imaging. the particle was peglyated to improve their biocompatibility. the in vivo and in vitro study against the breast cancer model showed an increased tumor inhibition growth rate. the particle delivered the gene to the site through a photothermal mediated response, and the efficiency of this procedure conducted in vivo was around 6-fold higher [138]. fahmi et al. [139] synthesized cds through a pyrolysis technique coated with magnetite and admet & dmpk 00(0) (2023) 000-000 theranostic applications of carbon-based quantum dots doi: https://doi.org/10.5599/admet.1747 19 loaded with naproxen as a model drug. the drug release was found to be ph responsive. the cellular toxicity study done with breast cancer cell lines in vitro showed minimal toxic effects, and the delivery of the drug was targeted to cancerous cells alone. this way, the designed hybrid material showed effective theranostic applications [139]. conclusion and future perspective carbon dots (cds) have gained enormous attention in biology and the healthcare industry due to their good aqueous solubility, biocompatibility, inexpensive fabrication procedures, and excellent photoluminescence properties. due to optical properties and structure-specific properties, cds possess tunable surface chemistry. biocompatible cds were synthesized via various synthetic procedures, such as bottom-up and top-down procedures, and were discussed from small organic molecules and their applications in various fields. although various methods are proposed for synthesizing cds, we still lack an understanding of the properties and mechanism to obtain a well-defined structure and uniform size distribution, which plays a key role in deciding the toxicity and fluorescence properties that further affect biological applications. the hydrothermal synthesis approach of synthesizing carbon dots comes out with a well-defined size and structure with functional modification, which helps obtain a high quantum yield and desired photoluminescence properties. however, future research will still be required to gain a thorough understanding of the various fields where cds can have a significant impact and benefits. as the first step, the fabrication process of cds needs to be studied since the size of the particles accounts for their fluorescence performance. control over their size, morphology, and surface properties is crucial to get desired product. the product yield is another concern during the synthesis procedure. carbon dots prepared especially via the green-mediated strategies exhibit low product yield. regardless of the issues, cds still hold a promising position in healthcare, as a little more understanding could resolve most concerns. acknowledgment: authors acknowledge care for financial and infrastructural support. pp, kh & pg acknowledge care for fellowships too. authors contribution: pg & kh collected all required data and wrote the initial draft. pp, ig, at, and kg compiled it with additional data. the concept, overall representation, and final version were prepared by ag. conflict of interest: the author declares no conflict of interest. references [1] u. abd rani, l.y. ng, c.y. ng, e. mahmoudi. a review of carbon quantum dots and their applications in wastewater treatment. advances in colloid and interface science 278 (2020) 102124. https://doi.org/10.1016/j.cis.2020.102124. [2] k. brindhadevi, h.a. garalleh, a. alalawi, e. al-sarayreh, a. pugazhendhi. carbon nanomaterials: types, synthesis strategies and their application as drug delivery system for cancer therapy. biochemical engineering journal 192 (2023) 108828. https://doi.org/10.1016/j.bej.2023.108828. [3] f. barati, m. avatefi, n.b. moghadam, s. asghari, e. ekrami, m. mahmoudifard. a review of graphene quantum dots and their potential biomedical applications. journal of biomaterials applications 37 (2023) 1137-1158. https://doi.org/10.1177/08853282221125311. [4] a. sharma, j. das. small molecules derived carbon dots: synthesis and applications in sensing, catalysis, imaging, and biomedicine. journal of nanobiotechnology 17 (2019) 1-24. https://doi.org/10.1186/s12951-019-0525-8. [5] p. sharmiladevi, n. akhtar, v. haribabu, k. girigoswami, s. chattopadhyay, a. girigoswami. excitation wavelength independent carbon-decorated ferrite nanodots for multimodal diagnosis and stimuli responsive therapy. acs applied bio materials 2 (2019) 1634-1642. https://doi.org/10.1021/acsabm.9b00039. https://doi.org/10.5599/admet.1747 https://doi.org/10.1016/j.cis.2020.102124 https://doi.org/10.1016/j.bej.2023.108828 https://doi.org/10.1177/08853282221125311 https://doi.org/10.1186/s12951-019-0525-8 https://doi.org/https:/doi.org/10.1021/acsabm.9b00039 https://doi.org/https:/doi.org/10.1021/acsabm.9b00039 pemula gowtham et al. admet & dmpk 00(0) (2023) 000-000 20 [6] b. das, a. girigoswami, a. dutta, p. pal, j. dutta, p. dadhich, p.k. srivas, s. dhara. carbon nanodots doped super-paramagnetic iron oxide nanoparticles for multimodal bioimaging and osteochondral tissue regeneration via external magnetic actuation. acs biomaterials science & engineering 5 (2019) 3549-3560. https://doi.org/10.1021/acsbiomaterials.9b00571. [7] b. das, a. girigoswami, p. pal, s. dhara. manganese oxide-carbon quantum dots nano-composites for fluorescence/magnetic resonance (t1) dual mode bioimaging, long term cell tracking, and ros scavenging. materials science and engineering: c 102 (2019) 427-436. https://doi.org/10.1016/j.msec.2019.04.077. [8] s. cailotto, m. negrato, s. daniele, r. luque, m. selva, e. amadio, a. perosa. carbon dots as photocatalysts for organic synthesis: metal-free methylene–oxygen-bond photocleavage. green chemistry 22 (2020) 1145-1149. https://doi.org/10.1039/c9gc03811f. [9] p. gowtham, v. haribabu, a.d. prabhu, p. pallavi, k. girigoswami, a. girigoswami. impact of nanovectors in multimodal medical imaging. nanomedicine journal 9 (2022) 107-130. https://doi.org/10.22038/nmj.2022.63459.1664. [10] p. pallavi, p. sharmiladevi, v. haribabu, k. girigoswami, a. girigoswami. a nano approach to formulate photosensitizers for photodynamic therapy. current nanoscience 18 (2022) 675-689. https://doi.org/10.2174/1573413718666211222162041. [11] s.f. völker, s. uemura, m. limpinsel, m. mingebach, c. deibel, v. dyakonov, c. lambert. polymeric squaraine dyes as electron donors in bulk heterojunction solar cells. macromolecular chemistry and physics 211 (2010) 1098-1108. https://doi.org/10.1002/macp.200900670. [12] p. sharmiladevi, k. girigoswami, v. haribabu, a. girigoswami. nano-enabled theranostics for cancer. materials advances 2 (2021) 2876-2891. https://doi.org/10.1039/d1ma00069a. [13] y. song, s. zhu, s. zhang, y. fu, l. wang, x. zhao, b. yang. investigation from chemical structure to photoluminescent mechanism: a type of carbon dots from the pyrolysis of citric acid and an amine. journal of materials chemistry c 3 (2015) 5976-5984. https://doi.org/10.1039/c5tc00813a. [14] j. zhou, y. yang, c.-y. zhang. a low-temperature solid-phase method to synthesize highly fluorescent carbon nitride dots with tunable emission. chemical communications 49 (2013) 8605-8607. https://doi.org/10.1039/c3cc42266f. [15] j. zhou, c. booker, r. li, x. zhou, t.-k. sham, x. sun, z. ding. an electrochemical avenue to blue luminescent nanocrystals from multiwalled carbon nanotubes (mwcnts). journal of the american chemical society 129 (2007) 744-745. https://doi.org/10.1021/ja0669070. [16] d. reyes-contreras, m. camacho-lópez, m.a. camacho-lópez, s. camacho-lópez, r.i. rodríguezbeltrán, m. mayorga-rojas. influence of the per pulse laser fluence on the optical properties of carbon nanoparticles synthesized by laser ablation of solids in liquids. optics & laser technology 74 (2015) 48-52. https://doi.org/10.1016/j.optlastec.2015.05.010. [17] q. zeng, d. shao, x. he, z. ren, w. ji, c. shan, s. qu, j. li, l. chen, q. li. carbon dots as a trackable drug delivery carrier for localized cancer therapy in vivo. journal of materials chemistry b 4 (2016) 5119-5126. https://doi.org/10.1039/c6tb01259k. [18] j. deng, q. lu, n. mi, h. li, m. liu, m. xu, l. tan, q. xie, y. zhang, s. yao. electrochemical synthesis of carbon nanodots directly from alcohols. chemistry–a european journal 20 (2014) 4993-4999. https://doi.org/10.1002/chem.201304869. [19] q. an, q. lin, x. huang, r. zhou, x. guo, w. xu, s. wang, d. xu, h.-t. chang. electrochemical synthesis of carbon dots with a stokes shift of 309 nm for sensing of fe3+ and ascorbic acid. dyes and pigments 185 (2021) 108878. https://doi.org/10.1016/j.dyepig.2020.108878. [20] s. ray, a. saha, n.r. jana, r. sarkar. fluorescent carbon nanoparticles: synthesis, characterization, and bioimaging application. the journal of physical chemistry c 113 (2009) 18546-18551. https://doi.org/10.1021/jp905912n. [21] q.-l. zhao, z.-l. zhang, b.-h. huang, j. peng, m. zhang, d.-w. pang. facile preparation of low cytotoxicity fluorescent carbon nanocrystals by electrooxidation of graphite. chemical communications (2008) 5116-5118. https://doi.org/10.1039/b812420e. https://doi.org/10.1021/acsbiomaterials.9b00571 https://doi.org/10.1016/j.msec.2019.04.077 https://doi.org/10.1016/j.msec.2019.04.077 https://doi.org/10.1039/c9gc03811f https://doi.org/10.22038/nmj.2022.63459.1664 https://doi.org/10.2174/1573413718666211222162041 https://doi.org/10.1002/macp.200900670 https://doi.org/10.1039/d1ma00069a https://doi.org/10.1039/c5tc00813a https://doi.org/10.1039/c3cc42266f https://doi.org/10.1021/ja0669070 https://doi.org/10.1016/j.optlastec.2015.05.010 https://doi.org/https:/doi.org/10.1039/c6tb01259k https://doi.org/10.1002/chem.201304869 https://doi.org/10.1016/j.dyepig.2020.108878 https://doi.org/10.1021/jp905912n https://doi.org/10.1039/b812420e admet & dmpk 00(0) (2023) 000-000 theranostic applications of carbon-based quantum dots doi: https://doi.org/10.5599/admet.1747 21 [22] m. liu, y. xu, f. niu, j.j. gooding, j. liu. carbon quantum dots directly generated from electrochemical oxidation of graphite electrodes in alkaline alcohols and the applications for specific ferric ion detection and cell imaging. analyst 141 (2016) 2657-2664. https://doi.org/10.1039/c5an02231b. [23] f. niu, y.-l. ying, x. hua, y. niu, y. xu, y.-t. long. electrochemically generated green-fluorescent ndoped carbon quantum dots for facile monitoring alkaline phosphatase activity based on the fe3+mediating on-off-on-off fluorescence principle. carbon 127 (2018) 340-348. https://doi.org/10.1016/j.carbon.2017.10.097. [24] h. zeng, x.w. du, s.c. singh, s.a. kulinich, s. yang, j. he, w. cai. nanomaterials via laser ablation/irradiation in liquid: a review. advanced functional materials 22 (2012) 1333-1353. https://doi.org/10.1002/adfm.201102295. [25] v. gardette, v. motto-ros, c. alvarez-llamas, l. sancey, l. duponchel, b. busser. laser-induced breakdown spectroscopy imaging for material and biomedical applications: recent advances and future perspectives. analytical chemistry 95 (2023) 49-69. https://doi.org/10.1021/acs.analchem.2c04910. [26] r.m. altuwirqi. graphene nanostructures by pulsed laser ablation in liquids: a review. materials 15 (2022) 5925. https://doi.org/10.3390/ma15175925. [27] n. semaltianos. nanoparticles by laser ablation. critical reviews in solid state and materials sciences 35 (2010) 105-124. https://doi.org/10.1080/10408431003788233. [28] a. kaczmarek, j. hoffman, j. morgiel, t. mościcki, l. stobiński, z. szymański, a. małolepszy. luminescent carbon dots synthesized by the laser ablation of graphite in polyethylenimine and ethylenediamine. materials 14 (2021) 729. https://doi.org/10.3390/ma14040729. [29] d. reyes, m. camacho, m. camacho, m. mayorga, d. weathers, g. salamo, z. wang, a. neogi. laser ablated carbon nanodots for light emission. nanoscale research letters 11 (2016) 1-11. https://doi.org/10.1186/s11671-016-1638-8. [30] v. nguyen, l. yan, j. si, x. hou. femtosecond laser-induced size reduction of carbon nanodots in solution: effect of laser fluence, spot size, and irradiation time. journal of applied physics 117 (2015) 084304. https://doi.org/10.1063/1.4909506. [31] h. yu, x. li, x. zeng, y. lu. preparation of carbon dots by non-focusing pulsed laser irradiation in toluene. chemical communications 52 (2015) 819-822. https://doi.org/10.1039/c5cc08384b. [32] v.s. sutkar, a.v. mahulkar, a.b. pandit, efficiency assessment and mapping of cavitational activities in sonochemical reactors, in energy aspects of acoustic cavitation and sonochemistry, elsevier2022, p. 157-183. https://doi.org/10.1016/b978-0-323-91937-1.00015-3. [33] v.b. kumar, z.e. porat, a. gedanken. facile one-step sonochemical synthesis of ultrafine and stable fluorescent c-dots. ultrasonics sonochemistry 28 (2016) 367-375. https://doi.org/10.1016/j.ultsonch.2015.08.005. [34] k. dehvari, k.y. liu, p.-j. tseng, g. gedda, w.m. girma, j.-y. chang. sonochemical-assisted green synthesis of nitrogen-doped carbon dots from crab shell as targeted nanoprobes for cell imaging. journal of the taiwan institute of chemical engineers 95 (2019) 495-503. https://doi.org/10.1016/j.jtice.2018.08.037. [35] h. li, x. he, y. liu, h. huang, s. lian, s.-t. lee, z. kang. one-step ultrasonic synthesis of water-soluble carbon nanoparticles with excellent photoluminescent properties. carbon 49 (2011) 605-609. https://doi.org/10.1016/j.carbon.2010.10.004. [36] z. wang, q. cheng, x. wang, j. li, w. li, y. li, g. zhang. carbon dots modified bismuth antimonate for broad spectrum photocatalytic degradation of organic pollutants: boosted charge separation, dft calculations and mechanism unveiling. chemical engineering journal 418 (2021) 129460. https://doi.org/10.1016/j.cej.2021.129460. [37] y.-s. lin, y. lin, a.p. periasamy, j. cang, h.-t. chang. parameters affecting the synthesis of carbon dots for quantitation of copper ions. nanoscale advances 1 (2019) 2553-2561. https://doi.org/10.1039/c9na00137a. https://doi.org/10.5599/admet.1747 https://doi.org/10.1039/c5an02231b https://doi.org/10.1016/j.carbon.2017.10.097 https://doi.org/10.1016/j.carbon.2017.10.097 https://doi.org/10.1021/acs.analchem.2c04910 https://doi.org/10.1021/acs.analchem.2c04910 https://doi.org/10.3390/ma15175925 https://doi.org/10.1080/10408431003788233 https://doi.org/10.3390/ma14040729 https://doi.org/10.1186/s11671-016-1638-8 https://doi.org/10.1063/1.4909506 https://doi.org/10.1039/c5cc08384b https://doi.org/10.1016/b978-0-323-91937-1.00015-3 https://doi.org/10.1016/j.ultsonch.2015.08.005 https://doi.org/10.1016/j.ultsonch.2015.08.005 https://doi.org/10.1016/j.jtice.2018.08.037 https://doi.org/10.1016/j.jtice.2018.08.037 https://doi.org/10.1016/j.carbon.2010.10.004 https://doi.org/10.1016/j.cej.2021.129460 https://doi.org/10.1039/c9na00137a https://doi.org/10.1039/c9na00137a pemula gowtham et al. admet & dmpk 00(0) (2023) 000-000 22 [38] f. niu, y. xu, m. liu, j. sun, p. guo, j. liu. bottom-up electrochemical preparation of solid-state carbon nanodots directly from nitriles/ionic liquids using carbon-free electrodes and the applications in specific ferric ion detection and cell imaging. nanoscale 8 (2016) 5470-5477. https://doi.org/10.1039/c6nr00023a. [39] y. hou, q. lu, j. deng, h. li, y. zhang. one-pot electrochemical synthesis of functionalized fluorescent carbon dots and their selective sensing for mercury ion. analytica chimica acta 866 (2015) 69-74. https://doi.org/10.1016/j.aca.2015.01.039. [40] s. hu, j. liu, j. yang, y. wang, s. cao. laser synthesis and size tailor of carbon quantum dots. journal of nanoparticle research 13 (2011) 7247-7252. https://doi.org/10.1007/s11051-011-0638-y. [41] f. wang, s. wang, z. sun, h. zhu. study on ultrasonic single-step synthesis and optical properties of nitrogen-doped carbon fluorescent quantum dots. fullerenes, nanotubes and carbon nanostructures 23 (2015) 769-776. https://doi.org/10.1080/1536383x.2014.996287. [42] c. li, x. sun, y. li, h. liu, b. long, d. xie, j. chen, k. wang. rapid and green fabrication of carbon dots for cellular imaging and anti-counterfeiting applications. acs omega 6 (2021) 3232-3237. https://doi.org/10.1021/acsomega.0c05682. [43] f. sun, h. ghosh, z. tan, s. sivoththaman. top-down synthesis and enhancing device adaptability of graphene quantum dots. nanotechnology 34 (2023) 185601. https://doi.org/10.1088/13616528/acb7fb. [44] t.v. de medeiros, j. manioudakis, f. noun, j.-r. macairan, f. victoria, r. naccache. microwaveassisted synthesis of carbon dots and their applications. journal of materials chemistry c 7 (2019) 7175-7195. https://doi.org/10.1039/c9tc01640f. [45] s. faraji, f.n. ani. microwave-assisted synthesis of metal oxide/hydroxide composite electrodes for high power supercapacitors–a review. journal of power sources 263 (2014) 338-360. https://doi.org/10.1016/j.jpowsour.2014.03.144. [46] c. lópez, m. zougagh, m. algarra, e. rodríguez-castellón, b. campos, j.e. da silva, j. jiménez-jiménez, a. ríos. microwave-assisted synthesis of carbon dots and its potential as analysis of four heterocyclic aromatic amines. talanta 132 (2015) 845-850. https://doi.org/10.1016/j.talanta.2014.10.008. [47] c. zhao, x. li, c. cheng, y. yang. green and microwave-assisted synthesis of carbon dots and application for visual detection of cobalt (ii) ions and ph sensing. microchemical journal 147 (2019) 183-190. https://doi.org/10.1016/j.microc.2019.03.029. [48] y. xu, h. li, b. wang, h. liu, l. zhao, t. zhou, m. liu, n. huang, y. li, l. ding. microwave-assisted synthesis of carbon dots for" turn-on" fluorometric determination of hg (ii) via aggregation-induced emission. microchimica acta 185 (2018) 1-7. https://doi.org/10.1007/s00604-018-2781-y. [49] r. ludmerczki, s. mura, c.m. carbonaro, i.m. mandity, m. carraro, n. senes, s. garroni, g. granozzi, l. calvillo, s. marras. carbon dots from citric acid and its intermediates formed by thermal decomposition. chemistry–a european journal 25 (2019) 11963-11974. https://doi.org/10.1002/chem.201902497. [50] n. wang, s. koh, b.g. jeong, d. lee, w.d. kim, k. park, m.k. nam, k. lee, y. kim, b.-h. lee. highly luminescent silica-coated cds/cdse/cds nanoparticles with strong chemical robustness and excellent thermal stability. nanotechnology 28 (2017) 185603. https://doi.org/10.1088/13616528/aa6828. [51] l. shi, j.h. yang, h.b. zeng, y.m. chen, s.c. yang, c. wu, h. zeng, o. yoshihito, q. zhang. carbon dots with high fluorescence quantum yield: the fluorescence originates from organic fluorophores. nanoscale 8 (2016) 14374-14378. https://doi.org/10.1039/c6nr00451b. [52] h. lin, j. huang, l. ding. preparation of carbon dots with high-fluorescence quantum yield and their application in dopamine fluorescence probe and cellular imaging. journal of nanomaterials 2019 (2019). https://doi.org/10.1155/2019/5037243. [53] j. qin, l. zhang, r. yang. powder carbonization to synthesize novel carbon dots derived from uric acid for the detection of ag (i) and glutathione. spectrochimica acta part a: molecular and biomolecular spectroscopy 207 (2019) 54-60. https://doi.org/10.1016/j.saa.2018.08.066. https://doi.org/10.1039/c6nr00023a https://doi.org/10.1039/c6nr00023a https://doi.org/10.1016/j.aca.2015.01.039 https://doi.org/10.1007/s11051-011-0638-y https://doi.org/10.1080/1536383x.2014.996287 https://doi.org/10.1021/acsomega.0c05682 https://doi.org/10.1088/1361-6528/acb7fb https://doi.org/10.1088/1361-6528/acb7fb https://doi.org/10.1039/c9tc01640f https://doi.org/10.1016/j.jpowsour.2014.03.144 https://doi.org/10.1016/j.talanta.2014.10.008 https://doi.org/10.1016/j.microc.2019.03.029 https://doi.org/10.1007/s00604-018-2781-y https://doi.org/10.1002/chem.201902497 https://doi.org/10.1002/chem.201902497 https://doi.org/10.1088/1361-6528/aa6828 https://doi.org/10.1088/1361-6528/aa6828 https://doi.org/10.1039/c6nr00451b https://doi.org/10.1155/2019/5037243 https://doi.org/10.1016/j.saa.2018.08.066 admet & dmpk 00(0) (2023) 000-000 theranostic applications of carbon-based quantum dots doi: https://doi.org/10.5599/admet.1747 23 [54] j. wang, j. wei, s. su, j. qiu. novel fluorescence resonance energy transfer optical sensors for vitamin b 12 detection using thermally reduced carbon dots. new journal of chemistry 39 (2015) 501-507. https://doi.org/10.1039/c4nj00538d [55] c.a. ma, c. yin, y. fan, x. yang, x. zhou. highly efficient synthesis of n-doped carbon dots with excellent stability through pyrolysis method. journal of materials science 54 (2019) 9372-9384. https://doi.org/10.1007/s10853-019-03585-7. [56] v. rimal, s. shishodia, p. srivastava. novel synthesis of high-thermal stability carbon dots and nanocomposites from oleic acid as an organic substrate. applied nanoscience 10 (2020) 455-464. https://doi.org/10.1007/s13204-019-01178-z. [57] c.-w. lai, y.-h. hsiao, y.-k. peng, p.-t. chou. facile synthesis of highly emissive carbon dots from pyrolysis of glycerol; gram scale production of carbon dots/msio 2 for cell imaging and drug release. journal of materials chemistry 22 (2012) 14403-14409. https://doi.org/10.1039/c2jm32206d. [58] w. tang, b. wang, j. li, y. li, y. zhang, h. quan, z. huang. facile pyrolysis synthesis of ionic liquid capped carbon dots and subsequent application as the water-based lubricant additives. journal of materials science 54 (2019) 1171-1183. https://doi.org/10.1007/s10853-018-2877-0. [59] t.-n. pham-truong, t. petenzi, c. ranjan, h. randriamahazaka, j. ghilane. microwave assisted synthesis of carbon dots in ionic liquid as metal free catalyst for highly selective production of hydrogen peroxide. carbon 130 (2018) 544-552. https://doi.org/10.1016/j.carbon.2018.01.070. [60] m.p. romero, f. alves, m.d. stringasci, h.h. buzzá, h. ciol, n.m. inada, v.s. bagnato. one-pot microwave-assisted synthesis of carbon dots and in vivo and in vitro antimicrobial photodynamic applications. frontiers in microbiology 12 (2021) 1455. https://doi.org/10.3389/fmicb.2021.662149. [61] m. zheng, z. xie, d. qu, d. li, p. du, x. jing, z. sun. on–off–on fluorescent carbon dot nanosensor for recognition of chromium (vi) and ascorbic acid based on the inner filter effect. acs applied materials & interfaces 5 (2013) 13242-13247. https://doi.org/10.1021/am4042355. [62] j. hou, w. wang, t. zhou, b. wang, h. li, l. ding. synthesis and formation mechanistic investigation of nitrogen-doped carbon dots with high quantum yields and yellowish-green fluorescence. nanoscale 8 (2016) 11185-11193. https://doi.org/10.1039/c6nr02701f. [63] f. wang, z. xie, h. zhang, c.y. liu, y.g. zhang. highly luminescent organosilane‐functionalized carbon dots. advanced functional materials 21 (2011) 1027-1031. https://doi.org/10.1002/adfm.201002279. [64] y. xu, m. wu, y. liu, x.z. feng, x.b. yin, x.w. he, y.k. zhang. nitrogen‐doped carbon dots: a facile and general preparation method, photoluminescence investigation, and imaging applications. chemistry–a european journal 19 (2013) 2276-2283. https://doi.org/10.1002/chem.201203641. [65] x.-m. wei, y. xu, y.-h. li, x.-b. yin, x.-w. he. ultrafast synthesis of nitrogen-doped carbon dots via neutralization heat for bioimaging and sensing applications. rsc advances 4 (2014) 44504-44508. https://doi.org/10.1039/c4ra08523j. [66] x.w. tan, a.n.b. romainor, s.f. chin, s.m. ng. carbon dots production via pyrolysis of sago waste as potential probe for metal ions sensing. journal of analytical and applied pyrolysis 105 (2014) 157165. https://doi.org/10.1016/j.jaap.2013.11.001. [67] c.s. stan, c. albu, a. coroaba, m. popa, d. sutiman. one step synthesis of fluorescent carbon dots through pyrolysis of n-hydroxysuccinimide. journal of materials chemistry c 3 (2015) 789-795. https://doi.org/10.1039/c4tc02382j. [68] l. stan, i. volf, c.s. stan, c. albu, a. coroaba, l.e. ursu, m. popa. intense blue photo emissive carbon dots prepared through pyrolytic processing of ligno-cellulosic wastes. nanomaterials 13 (2023) 131. https://doi.org/10.3390/nano13010131. [69] x. wang, j. zhuang, q. peng, y. li. a general strategy for nanocrystal synthesis. nature 437 (2005) 121-124. https://doi.org/10.1038/nature03968. [70] g. ischia, m. cutillo, g. guella, n. bazzanella, m. cazzanelli, m. orlandi, a. miotello, l. fiori. hydrothermal carbonization of glucose: secondary char properties, reaction pathways, and kinetics. chemical engineering journal 449 (2022) 137827. https://doi.org/10.1016/j.cej.2022.137827. https://doi.org/10.5599/admet.1747 https://doi.org/10.1039/c4nj00538d https://doi.org/10.1007/s10853-019-03585-7 https://doi.org/10.1007/s13204-019-01178-z https://doi.org/10.1039/c2jm32206d https://doi.org/10.1007/s10853-018-2877-0 https://doi.org/10.1016/j.carbon.2018.01.070 https://doi.org/10.3389/fmicb.2021.662149 https://doi.org/10.1021/am4042355 https://doi.org/10.1039/c6nr02701f https://doi.org/10.1002/adfm.201002279 https://doi.org/10.1002/adfm.201002279 https://doi.org/10.1002/chem.201203641 https://doi.org/10.1039/c4ra08523j https://doi.org/10.1016/j.jaap.2013.11.001 https://doi.org/10.1039/c4tc02382j https://doi.org/10.3390/nano13010131 https://doi.org/10.1038/nature03968 https://doi.org/10.1016/j.cej.2022.137827 pemula gowtham et al. admet & dmpk 00(0) (2023) 000-000 24 [71] t. ghosh, r. sahoo, s.k. ghosh, p. banerji, n.c. das. simplistic hydrothermal synthesis approach for fabricating photoluminescent carbon dots and its potential application as an efficient sensor probe for toxic lead (ii) ion detection. frontiers of chemical science and engineering (2023) 1-12. https://doi.org/10.1007/s11705-022-2239-y. [72] h. hayashi, y. hakuta. hydrothermal synthesis of metal oxide nanoparticles in supercritical water. materials 3 (2010) 3794-3817. https://doi.org/10.3390/ma3073794. [73] t. adschiri, y. hakuta, k. sue, k. arai. hydrothermal synthesis of metal oxide nanoparticles at supercritical conditions. journal of nanoparticle research 3 (2001) 227-235. https://doi.org/10.1023/a:1017541705569. [74] s.-h. feng, g.-h. li, hydrothermal and solvothermal syntheses, in modern inorganic synthetic chemistry, elsevier2017, p. 73-104. https://doi.org/10.1016/b978-0-444-63591-4.00004-5. [75] v. haribabu, k. girigoswami, p. sharmiladevi, a. girigoswami. water–nanomaterial interaction to escalate twin-mode magnetic resonance imaging. acs biomaterials science & engineering 6 (2020) 4377-4389. https://doi.org/10.1021/acsbiomaterials.0c00409. [76] c. olla, a. cappai, s. porcu, l. stagi, m. fantauzzi, m.f. casula, f. mocci, r. corpino, d. chiriu, p.c. ricci. exploring the impact of nitrogen doping on the optical properties of carbon dots synthesized from citric acid. nanomaterials 13 (2023) 1344. https://doi.org/10.3390/nano13081344. [77] x. yao, y. wang, f. li, j.j. dalluge, g. orr, r. hernandez, q. cui, c.l. haynes. unconventional aliphatic fluorophores discovered as the luminescence origin in citric acid–urea carbon dots. nanoscale 14 (2022) 9516-9525. https://doi.org/10.1039/d2nr02361j. [78] j. schneider, c.j. reckmeier, y. xiong, m. von seckendorff, a.s. susha, p. kasák, a.l. rogach. molecular fluorescence in citric acid-based carbon dots. the journal of physical chemistry c 121 (2017) 2014-2022. https://doi.org/10.1021/acs.jpcc.6b12519. [79] y. wang, a. hu. carbon quantum dots: synthesis, properties and applications. journal of materials chemistry c 2 (2014) 6921-6939. https://doi.org/10.1039/c4tc00988f. [80] n.m. zholobak, a.l. popov, a.b. shcherbakov, n.r. popova, m.m. guzyk, v.p. antonovich, a.v. yegorova, y.v. scrypynets, i.i. leonenko, a.y. baranchikov. facile fabrication of luminescent organic dots by thermolysis of citric acid in urea melt, and their use for cell staining and polyelectrolyte microcapsule labelling. beilstein journal of nanotechnology 7 (2016) 1905-1917. https://doi.org/10.3762/bjnano.7.182. [81] w. kasprzyk, s. bednarz, p. żmudzki, m. galica, d. bogdał. novel efficient fluorophores synthesized from citric acid. rsc advances 5 (2015) 34795-34799. https://doi.org/10.1039/c5ra03226a. [82] w. kasprzyk, t. świergosz, s. bednarz, k. walas, n.v. bashmakova, d. bogdał. luminescence phenomena of carbon dots derived from citric acid and urea–a molecular insight. nanoscale 10 (2018) 13889-13894. https://doi.org/10.1039/c8nr03602k. [83] j. liu, s. lu, q. tang, k. zhang, w. yu, h. sun, b. yang. one-step hydrothermal synthesis of photoluminescent carbon nanodots with selective antibacterial activity against porphyromonas gingivalis. nanoscale 9 (2017) 7135-7142. https://doi.org/10.1039/c7nr02128c. [84] b. wang, h. tan, t. zhang, w. duan, y. zhu. hydrothermal synthesis of n-doped carbon dots from an ethanolamine–ionic liquid gel to construct label-free multifunctional fluorescent probes for hg 2+, cu 2+ and s 2 o 3 2−. analyst 144 (2019) 3013-3022. https://doi.org/10.1039/c9an00116f. [85] j.-y. li, y. liu, q.-w. shu, j.-m. liang, f. zhang, x.-p. chen, x.-y. deng, m.t. swihart, k.-j. tan. one-pot hydrothermal synthesis of carbon dots with efficient up-and down-converted photoluminescence for the sensitive detection of morin in a dual-readout assay. langmuir 33 (2017) 1043-1050. https://doi.org/10.1021/acs.langmuir.6b04225. [86] l. lin, y. wang, y. xiao, w. liu. hydrothermal synthesis of carbon dots codoped with nitrogen and phosphorus as a turn-on fluorescent probe for cadmium (ii). microchimica acta 186 (2019) 1-7. https://doi.org/10.1007/s00604-019-3264-5. https://doi.org/10.1007/s11705-022-2239-y https://doi.org/10.3390/ma3073794 https://doi.org/10.1023/a:1017541705569 https://doi.org/10.1023/a:1017541705569 https://doi.org/10.1016/b978-0-444-63591-4.00004-5 https://doi.org/10.1021/acsbiomaterials.0c00409 https://doi.org/10.3390/nano13081344 https://doi.org/10.1039/d2nr02361j https://doi.org/10.1021/acs.jpcc.6b12519 https://doi.org/10.1039/c4tc00988f https://doi.org/10.3762/bjnano.7.182 https://doi.org/10.3762/bjnano.7.182 https://doi.org/10.1039/c5ra03226a https://doi.org/10.1039/c8nr03602k https://doi.org/10.1039/c7nr02128c https://doi.org/10.1039/c9an00116f https://doi.org/10.1021/acs.langmuir.6b04225 https://doi.org/10.1007/s00604-019-3264-5 admet & dmpk 00(0) (2023) 000-000 theranostic applications of carbon-based quantum dots doi: https://doi.org/10.5599/admet.1747 25 [87] z. wang, b. fu, s. zou, b. duan, c. chang, b. yang, x. zhou, l. zhang. facile construction of carbon dots via acid catalytic hydrothermal method and their application for target imaging of cancer cells. nano research 9 (2016) 214-223. https://doi.org/10.1007/s12274-016-0992-2. [88] s. liu, j. cui, j. huang, b. tian, f. jia, z. wang. facile one-pot synthesis of highly fluorescent nitrogendoped carbon dots by mild hydrothermal method and their applications in detection of cr (vi) ions. spectrochimica acta part a: molecular and biomolecular spectroscopy 206 (2019) 65-71. https://doi.org/10.1016/j.saa.2018.07.082. [89] h. lin, l. ding, b. zhang, j. huang. detection of nitrite based on fluorescent carbon dots by the hydrothermal method with folic acid. royal society open science 5 (2018) 172149. https://doi.org/10.1098/rsos.172149. [90] s. wang, s.-h. wu, w.-l. fang, x.-f. guo, h. wang. synthesis of non-doped and non-modified carbon dots with high quantum yield and crystallinity by one-pot hydrothermal method using a single carbon source and used for clo− detection. dyes and pigments 164 (2019) 7-13. https://doi.org/10.1016/j.dyepig.2019.01.004. [91] r. atchudan, t.n.j.i. edison, m. shanmugam, s. perumal, t. somanathan, y.r. lee. sustainable synthesis of carbon quantum dots from banana peel waste using hydrothermal process for in vivo bioimaging. physica e: low-dimensional systems and nanostructures 126 (2021) 114417. https://doi.org/10.1016/j.physe.2020.114417. [92] r. atchudan, t.n. jebakumar immanuel edison, s. perumal, y.r. lee. indian gooseberry-derived tunable fluorescent carbon dots as a promise for in vitro/in vivo multicolor bioimaging and fluorescent ink. acs omega 3 (2018) 17590-17601. https://doi.org/10.1021/acsomega.8b02463. [93] h. lee, y.-c. su, h.-h. tang, y.-s. lee, j.-y. lee, c.-c. hu, t.-c. chiu. one-pot hydrothermal synthesis of carbon dots as fluorescent probes for the determination of mercuric and hypochlorite ions. nanomaterials 11 (2021) 1831. https://doi.org/10.3390/nano11071831. [94] t.-y. shen, p.-y. jia, d.-s. chen, l.-n. wang. hydrothermal synthesis of n-doped carbon quantum dots and their application in ion-detection and cell-imaging. spectrochimica acta part a: molecular and biomolecular spectroscopy 248 (2021) 119282. https://doi.org/10.1016/j.saa.2020.119282. [95] y. tu, s. wang, x. yuan, y. xiang, k. qin, y. wei, q. zhang, x. chen, x. ji. facile hydrothermal synthesis of nitrogen, phosphorus-doped fluorescent carbon dots for live/dead bacterial differentiation, cell imaging and two nitrophenols detection. dyes and pigments 184 (2021) 108761. https://doi.org/10.1016/j.dyepig.2020.108761. [96] j. zhang, s.s. zhao, z. yang, z. yang, s. yang, x. liu. hydrothermal synthesis of blue‐green emitting carbon dots based on the liquid products of biodegradation of coal. international journal of energy research 45 (2021) 9396-9407. https://doi.org/10.1002/er.6468. [97] j. luo, z. sun, w. zhou, f. mo, z.-c. wu, x. zhang. hydrothermal synthesis of bright blue-emitting carbon dots for bioimaging and fluorescent determination of baicalein. optical materials 113 (2021) 110796. https://doi.org/10.1016/j.optmat.2020.110796. [98] s. marouzi, m. darroudi, a. hekmat, k. sadri, r.k. oskuee. one-pot hydrothermal synthesis of carbon quantum dots from salvia hispanica l. seeds and investigation of their biodistribution, and cytotoxicity effects. journal of environmental chemical engineering 9 (2021) 105461. https://doi.org/10.1016/j.jece.2021.105461. [99] p. mohammad-jafarieh, a. akbarzadeh, r. salamat-ahangari, m. pourhassan-moghaddam, k. jamshidi-ghaleh. solvent effect on the absorption and emission spectra of carbon dots: evaluation of ground and excited state dipole moment. bmc chemistry 15 (2021) 1-10. https://doi.org/10.1186/s13065-021-00779-6. [100] a. meng, b. huangfu, l. sheng, x. hong, z. li. one-pot hydrothermal synthesis of boron and nitrogen co-doped carbon dots for copper ion assay and multicolor cell imaging using fluorescence quenchometric method. microchemical journal 174 (2022) 106981. https://doi.org/10.1016/j.microc.2021.106981. https://doi.org/10.5599/admet.1747 https://doi.org/10.1007/s12274-016-0992-2 https://doi.org/10.1016/j.saa.2018.07.082 https://doi.org/10.1098/rsos.172149 https://doi.org/10.1016/j.dyepig.2019.01.004 https://doi.org/10.1016/j.dyepig.2019.01.004 https://doi.org/10.1016/j.physe.2020.114417 https://doi.org/10.1021/acsomega.8b02463 https://doi.org/10.3390/nano11071831 https://doi.org/10.1016/j.saa.2020.119282 https://doi.org/10.1016/j.dyepig.2020.108761 https://doi.org/10.1002/er.6468 https://doi.org/10.1016/j.optmat.2020.110796 https://doi.org/10.1016/j.jece.2021.105461 https://doi.org/10.1186/s13065-021-00779-6 https://doi.org/10.1016/j.microc.2021.106981 https://doi.org/10.1016/j.microc.2021.106981 pemula gowtham et al. admet & dmpk 00(0) (2023) 000-000 26 [101] z. han, l. he, s. pan, h. liu, x. hu. hydrothermal synthesis of carbon dots and their application for detection of chlorogenic acid. luminescence 35 (2020) 989-997. https://doi.org/ https://doi.org/10.1002/bio.3803. [102] d. sun, t. liu, c. wang, l. yang, s. yang, k. zhuo. hydrothermal synthesis of fluorescent carbon dots from gardenia fruit for sensitive on-off-on detection of hg2+ and cysteine. spectrochimica acta part a: molecular and biomolecular spectroscopy 240 (2020) 118598. https://doi.org/10.1016/j.saa.2020.118598. [103] c. zhang, m. liu, t. li, s. liu, q. chen, j. zhang, k. zhang. one-pot hydrothermal synthesis of dualemission fluorescent carbon dots for hypochlorous acid detection. dyes and pigments 180 (2020) 108507. https://doi.org/10.1016/j.dyepig.2020.108507. [104] t.n.j.i. edison, r. atchudan, n. karthik, d. xiong, y.r. lee. facile hydrothermal synthesis of nitrogen rich blue fluorescent carbon dots for cell bio-imaging of candida albicans. process biochemistry 88 (2020) 113-119. https://doi.org/10.1016/j.procbio.2019.10.003. [105] z.m. khan, s. saifi, z. aslam, s.a. khan, m. zulfequar. a facile one step hydrothermal synthesis of carbon quantum dots for label-free fluorescence sensing approach to detect picric acid in aqueous solution. journal of photochemistry and photobiology a: chemistry 388 (2020) 112201. https://doi.org/10.1016/j.jphotochem.2019.112201. [106] z. ye, g. li, j. lei, m. liu, y. jin, b. li. one-step and one-precursor hydrothermal synthesis of carbon dots with superior antibacterial activity. acs applied bio materials 3 (2020) 7095-7102. https://doi.org/10.1021/acsabm.0c00923. [107] a. paul, m. kurian. facile synthesis of nitrogen doped carbon dots from waste biomass: potential optical and biomedical applications. cleaner engineering and technology 3 (2021) 100103. https://doi.org/10.1016/j.clet.2021.100103. [108] z. yi, x. li, h. zhang, x. ji, w. sun, y. yu, y. liu, j. huang, z. sarshar, m. sain. high quantum yield photoluminescent n-doped carbon dots for switch sensing and imaging. talanta 222 (2021) 121663. https://doi.org/10.1016/j.talanta.2020.121663. [109] y. xie, d. cheng, x. liu, a. han. green hydrothermal synthesis of n-doped carbon dots from biomass highland barley for the detection of hg2+. sensors 19 (2019) 3169. https://doi.org/10.3390/s19143169. [110] a.k. singh, v.k. singh, m. singh, p. singh, s.r. khadim, u. singh, b. koch, s. hasan, r. asthana. one pot hydrothermal synthesis of fluorescent np-carbon dots derived from dunaliella salina biomass and its application in on-off sensing of hg (ii), cr (vi) and live cell imaging. journal of photochemistry and photobiology a: chemistry 376 (2019) 63-72. https://doi.org/10.1016/j.jphotochem.2019.02.023. [111] y. newman monday, j. abdullah, n.a. yusof, s. abdul rashid, r.h. shueb. facile hydrothermal and solvothermal synthesis and characterization of nitrogen-doped carbon dots from palm kernel shell precursor. applied sciences 11 (2021) 1630. https://doi.org/10.3390/app11041630. [112] d. wang, x. mei, s. wang, j. li, c. dong. a one-pot synthesis of fluorescent n, p-codoped carbon dots for vitamin b 12 determination and bioimaging application. new journal of chemistry 45 (2021) 35083514. https://doi.org/10.1039/d0nj05597b. [113] p. varatharajan, i.s. banu, m.h. mamat, n. vasimalai. hydrothermal synthesis of orange fluorescent carbon dots and their application in fabrication of warm wleds and fluorescent ink. physica b: condensed matter 654 (2023) 414703. https://doi.org/10.1016/j.physb.2023.414703. [114] k. harini, k. girigoswami, d. ghosh, p. pallavi, p. gowtham, a. girigoswami. architectural fabrication of multifunctional janus nanostructures for biomedical applications. nanomedicine journal 9 (2022) 180-191. https://doi.org/10.22038/nmj.2022.65101.1682. [115] k. harini, k. girigoswami, a. girigoswami. nanopsychiatry: engineering of nanoassisted drug delivery systems to formulate antidepressants. international journal of nano dimension (2022). https://doi.org/10.22034/ijnd.2022.1955783.2133. https://doi.org/ https://doi.org/10.1002/bio.3803 https://doi.org/10.1016/j.saa.2020.118598 https://doi.org/10.1016/j.saa.2020.118598 https://doi.org/10.1016/j.dyepig.2020.108507 https://doi.org/10.1016/j.procbio.2019.10.003 https://doi.org/10.1016/j.jphotochem.2019.112201 https://doi.org/10.1021/acsabm.0c00923 https://doi.org/10.1016/j.clet.2021.100103 https://doi.org/10.1016/j.talanta.2020.121663 https://doi.org/10.3390/s19143169 https://doi.org/10.3390/s19143169 https://doi.org/https:/doi.org/10.1016/j.jphotochem.2019.02.023 https://doi.org/https:/doi.org/10.1016/j.jphotochem.2019.02.023 https://doi.org/10.3390/app11041630 https://doi.org/10.1039/d0nj05597b https://doi.org/10.1016/j.physb.2023.414703 https://doi.org/10.22038/nmj.2022.65101.1682 https://doi.org/10.22034/ijnd.2022.1955783.2133 admet & dmpk 00(0) (2023) 000-000 theranostic applications of carbon-based quantum dots doi: https://doi.org/10.5599/admet.1747 27 [116] k. harini, p. pallavi, p. gowtham, k. girigoswami, a. girigoswami. smart polymer-based reduction responsive therapeutic delivery to cancer cells. current pharmacology reports (2022) 1-7. https://doi.org/10.1007/s40495-022-00282-z. [117] p. sharmiladevi, m. breghatha, k. dhanavardhini, r. priya, k. girigoswami, a. girigoswami. efficient wormlike micelles for the controlled delivery of anticancer drugs. nanoscience & nanotechnologyasia 11 (2021) 350-356. https://doi.org/10.2174/2210681210999200728115601. [118] r. sakthi devi, a. girigoswami, m. siddharth, k. girigoswami. applications of gold and silver nanoparticles in theranostics. applied biochemistry and biotechnology (2022) 1-33. https://doi.org/10.1007/s12010-022-03963-z. [119] g. poornima, k. harini, p. pallavi, p. gowtham, k. girigoswami, a. girigoswami. rna–a choice of potential drug delivery system. international journal of polymeric materials and polymeric biomaterials (2022) 1-15. https://doi.org/10.1080/00914037.2022.2058946. [120] a. girigoswami, m. ramalakshmi, n. akhtar, s.k. metkar, k. girigoswami. zno nanoflower petals mediated amyloid degradation-an in vitro electrokinetic potential approach. materials science and engineering: c 101 (2019) 169-178. https://doi.org/10.1016/j.msec.2019.03.086. [121] y. zhong, j. li, y. jiao, g. zuo, x. pan, t. su, w. dong. one-step synthesis of orange luminescent carbon dots for ag+ sensing and cell imaging. journal of luminescence 190 (2017) 188-193. https://doi.org/10.1016/j.jlumin.2017.05.057. [122] z. wang, h. liao, h. wu, b. wang, h. zhao, m. tan. fluorescent carbon dots from beer for breast cancer cell imaging and drug delivery. analytical methods 7 (2015) 8911-8917. https://doi.org/10.1039/c5ay01978h. [123] h. liu, z. li, y. sun, x. geng, y. hu, h. meng, j. ge, l. qu. synthesis of luminescent carbon dots with ultrahigh quantum yield and inherent folate receptor-positive cancer cell targetability. scientific reports 8 (2018) 1-8. https://doi.org/10.1038/s41598-018-19373-3. [124] k. datta, o. kozak, v. ranc, m. havrdova, a. bourlinos, k. šafářová, k. hola, k. tomankova, g. zoppellaro, m. otyepka. quaternized carbon dot-modified graphene oxide for selective cell labelling– controlled nucleus and cytoplasm imaging. chemical communications 50 (2014) 10782-10785. https://doi.org/10.1039/c4cc02637c. [125] j. ci, y. tian, s. kuga, z. niu, m. wu, y. huang. one‐pot green synthesis of nitrogen‐doped carbon quantum dots for cell nucleus labeling and copper (ii) detection. chemistry–an asian journal 12 (2017) 2916-2921. https://doi.org/10.1002/asia.201700880. [126] y.k. jung, e. shin, b.-s. kim. cell nucleus-targeting zwitterionic carbon dots. scientific reports 5 (2015) 1-9. https://doi.org/10.1038/srep18807. [127] s.a. hill, s. sheikh, q. zhang, l.s. ballesteros, a. herman, s.a. davis, d.j. morgan, m. berry, d. benitoalifonso, m.c. galan. selective photothermal killing of cancer cells using led-activated nucleus targeting fluorescent carbon dots. nanoscale advances 1 (2019) 2840-2846. https://doi.org/10.1039/c9na00293f. [128] h. liu, y. bai, y. zhou, c. feng, l. liu, l. fang, j. liang, s. xiao. blue and cyan fluorescent carbon dots: one-pot synthesis, selective cell imaging and their antiviral activity. rsc advances 7 (2017) 2801628023. https://doi.org/10.1039/c7ra03167j [129] s. guo, y. sun, j. li, x. geng, r. yang, x. zhang, l. qu, z. li. fluorescent carbon dots shuttling between mitochondria and the nucleolus for in situ visualization of cell viability. acs applied bio materials 4 (2020) 928-934. https://doi.org/10.1021/acsabm.0c01408. [130] h. zhang, y. chen, m. liang, l. xu, s. qi, h. chen, x. chen. solid-phase synthesis of highly fluorescent nitrogen-doped carbon dots for sensitive and selective probing ferric ions in living cells. analytical chemistry 86 (2014) 9846-9852. https://doi.org/10.1021/ac502446m. [131] r. wang, x. wang, y. sun. one-step synthesis of self-doped carbon dots with highly photoluminescence as multifunctional biosensors for detection of iron ions and ph. sensors and actuators b: chemical 241 (2017) 73-79. https://doi.org/10.1016/j.snb.2016.10.043. https://doi.org/10.5599/admet.1747 https://doi.org/10.1007/s40495-022-00282-z https://doi.org/10.2174/2210681210999200728115601 https://doi.org/10.1007/s12010-022-03963-z https://doi.org/10.1080/00914037.2022.2058946 https://doi.org/10.1016/j.msec.2019.03.086 https://doi.org/10.1016/j.jlumin.2017.05.057 https://doi.org/10.1039/c5ay01978h https://doi.org/10.1039/c5ay01978h https://doi.org/10.1038/s41598-018-19373-3 https://doi.org/10.1039/c4cc02637c https://doi.org/10.1002/asia.201700880 https://doi.org/10.1038/srep18807 https://doi.org/10.1039/c9na00293f https://doi.org/10.1039/c9na00293f https://doi.org/10.1039/c7ra03167j https://doi.org/10.1021/acsabm.0c01408 https://doi.org/10.1021/ac502446m https://doi.org/10.1016/j.snb.2016.10.043 pemula gowtham et al. admet & dmpk 00(0) (2023) 000-000 28 [132] s. mohammadi, a. salimi. fluorometric determination of microrna-155 in cancer cells based on carbon dots and mno2 nanosheets as a donor-acceptor pair. microchimica acta 185 (2018) 1-10. https://doi.org/10.1007/s00604-018-2868-5. [133] x. shan, l. chai, j. ma, z. qian, j. chen, h. feng. b-doped carbon quantum dots as a sensitive fluorescence probe for hydrogen peroxide and glucose detection. analyst 139 (2014) 2322-2325. https://doi.org/10.1039/c3an02222f. [134] y. hailing, l. xiufang, w. lili, l. baoqiang, h. kaichen, h. yongquan, z. qianqian, m. chaoming, r. xiaoshuai, z. rui. doxorubicin-loaded fluorescent carbon dots with pei passivation as a drug delivery system for cancer therapy. nanoscale 12 (2020) 17222-17237. https://doi.org/10.1039/d0nr01236j. [135] h.k. chung, v. wongso, n.s. sambudi. biowaste-derived carbon dots/hydroxyapatite nanocomposite as drug delivery vehicle for acetaminophen. journal of sol-gel science and technology 93 (2020) 214223. https://doi.org/10.1007/s10971-019-05141-w. [136] n. sarkar, g. sahoo, r. das, g. prusty, s.k. swain. carbon quantum dot tailored calcium alginate hydrogel for ph responsive controlled delivery of vancomycin. european journal of pharmaceutical sciences 109 (2017) 359-371. https://doi.org/10.1016/j.ejps.2017.08.015. [137] z. chen, t. liao, l. wan, y. kuang, c. liu, j. duan, x. xu, z. xu, b. jiang, c. li. dual-stimuli responsive near-infrared emissive carbon dots/hollow mesoporous silica-based integrated theranostics platform for real-time visualized drug delivery. nano research 14 (2021) 4264-4273. https://doi.org/10.1007/s12274-021-3624-4. [138] t. luo, y. nie, j. lu, q. bi, z. cai, x. song, h. ai, r. jin. iron doped carbon dots based nanohybrids as a tetramodal imaging agent for gene delivery promotion and photothermal-chemodynamic cancer synergistic theranostics. materials & design 208 (2021) 109878. https://doi.org/10.1016/j.matdes.2021.109878. [139] m.z. fahmi, n. machmudah, p. indrawasih, a. wibrianto, m.a. ahmad, s.c. sakti, j.-y. chang. naproxen release from carbon dot coated magnetite nanohybrid as versatile theranostics for hela cancer cells. rsc advances 12 (2022) 32328-32337. https://doi.org/10.1039/d2ra05673a. [140] b. tian, s. liu, l. feng, s. liu, s. gai, y. dai, l. xie, b. liu, p. yang, y. zhao. renal‐clearable nickel‐doped carbon dots with boosted photothermal conversion efficiency for multimodal imaging‐guided cancer therapy in the second near‐infrared biowindow. advanced functional materials 31 (2021) 2100549. https://doi.org/10.1002/adfm.202100549. [141] j. li, z.-e. hu, y.-j. we, y.-h. liu, n. wang, x.-q. yu. multifunctional carbon quantum dots as a theranostic nanomedicine for fluorescence imaging-guided glutathione depletion to improve chemodynamic therapy. journal of colloid and interface science 606 (2022) 1219-1228. https://doi.org/10.1016/j.jcis.2021.08.114. [142] s. sun, q. chen, z. tang, c. liu, z. li, a. wu, h. lin. tumor microenvironment stimuli‐responsive fluorescence imaging and synergistic cancer therapy by carbon‐dot–cu2+ nanoassemblies. angewandte chemie 132 (2020) 21227-21234. https://doi.org/10.1002/ange.202007786. [143] y. guo, y. fan, g. li, z. wang, x. shi, m. shen. “cluster bomb” based on redox-responsive carbon dot nanoclusters coated with cell membranes for enhanced tumor theranostics. acs applied materials & interfaces 13 (2021) 55815-55826. https://doi.org/10.1021/acsami.1c15282. [144] m. jiao, y. wang, w. wang, x. zhou, j. xu, y. xing, l. chen, y. zhang, m. chen, k. xu. gadolinium doped red-emissive carbon dots as targeted theranostic agents for fluorescence and mr imaging guided cancer phototherapy. chemical engineering journal 440 (2022) 135965. https://doi.org/10.1016/j.cej.2022.135965. [145] m. zheng, y. li, s. liu, w. wang, z. xie, x. jing. one-pot to synthesize multifunctional carbon dots for near infrared fluorescence imaging and photothermal cancer therapy. acs applied materials & interfaces 8 (2016) 23533-23541. https://doi.org/10.1021/acsami.6b07453. [146] q. jiang, l. liu, q. li, y. cao, d. chen, q. du, x. yang, d. huang, r. pei, x. chen. nir-laser-triggered gadolinium-doped carbon dots for magnetic resonance imaging, drug delivery and combined photothermal chemotherapy for triple negative breast cancer. journal of nanobiotechnology 19 (2021) 1-15. https://doi.org/10.1186/s12951-021-00811-w. https://doi.org/10.1007/s00604-018-2868-5 https://doi.org/10.1039/c3an02222f https://doi.org/10.1039/d0nr01236j https://doi.org/10.1007/s10971-019-05141-w https://doi.org/10.1016/j.ejps.2017.08.015 https://doi.org/10.1007/s12274-021-3624-4 https://doi.org/10.1016/j.matdes.2021.109878 https://doi.org/10.1016/j.matdes.2021.109878 https://doi.org/10.1039/d2ra05673a https://doi.org/10.1002/adfm.202100549 https://doi.org/10.1016/j.jcis.2021.08.114 https://doi.org/10.1002/ange.202007786 https://doi.org/10.1021/acsami.1c15282 https://doi.org/10.1016/j.cej.2022.135965 https://doi.org/10.1021/acsami.6b07453 https://doi.org/10.1186/s12951-021-00811-w admet & dmpk 00(0) (2023) 000-000 theranostic applications of carbon-based quantum dots doi: https://doi.org/10.5599/admet.1747 29 [147] y. bai, j. zhao, s. wang, t. lin, f. ye, s. zhao. carbon dots with absorption red-shifting for two-photon fluorescence imaging of tumor tissue ph and synergistic phototherapy. acs applied materials & interfaces 13 (2021) 35365-35375. https://doi.org/10.1021/acsami.1c08076. [148] a.r. gul, t.n. le, m.w. kim, s.k. kailasa, k.t. oh, t.j. park. one-pot synthesis of carbon dots with intrinsic folic acid for synergistic imaging-guided photothermal therapy of prostate cancer cells. biomaterials science 7 (2019) 5187-5196. https://doi.org/10.1039/c9bm01228a. [149] y. li, g. bai, s. zeng, j. hao. theranostic carbon dots with innovative nir-ii emission for in vivo renalexcreted optical imaging and photothermal therapy. acs applied materials & interfaces 11 (2019) 4737-4744. https://doi.org/10.1021/acsami.8b14877. ©2023 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) https://doi.org/10.5599/admet.1747 https://doi.org/10.1021/acsami.1c08076 https://doi.org/10.1039/c9bm01228a https://doi.org/10.1021/acsami.8b14877 http://creativecommons.org/licenses/by/3.0/ doi: 10.5599/admet.2.1.37 1 admet & dmpk 2(1) (2014) 1-2; doi: 10.5599/admet.2.1.37 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index editorial special issue devoted to the 3rd world conference on physico chemical methods in drug discovery and development (pcmddd-3) kin tam 1 and zoran mandić 2 1 faculty of health science, university of macau, macau, china e-mail: kintam@umac.mo; tel.: +853-8397-8501; fax: +853-8397-8500 2 university of zagreb, faculty of chemical engineering and technology, hr-10000, zagreb, croatia e-mail: zmandic@fkit.hr; tel: +385 1 4597164 the present and following issues of admet and dmpk are dedicated to the 3rd world conference on physico-chemical methods in drug discovery and development (pcmddd-3) held in dubrovnik, croatia, 23-26 september 2013. pcmddd-3 is organized as a biannual event and is intended to provide a place and common ground for the scientific community whose work is closely related to the application of physical chemistry in adme and dmpk research to get together in an open and relaxing atmosphere to exchange ideas and discuss challenges. the topics of the last conference included: physico-chemical methods and instrumentation in the physico-chemical profiling of drug substances, adme and dmpk, determination and characterization of different solid forms, hydrates and polymorphs, separation and analytical techniques of importance in medicinal chemistry and pharmaceutical research, computational methods and modelling. more details of the pcmddd-3 can be found on the conference webpage: http://www.iapchem.org/page.php?page_id=34. almost 150 delegates actively participated at the pcmddd-3. most of them presented their work either orally or through poster communication resulting in a diverse but well-balanced four-day programme. these two issues review a small selection of about 15 papers representing a typical cross-section of the conference workings. the first part is composed of six papers. a review about multivariate analysis of hydrophobic parameters by stefan dove gives a historical overview and describes the importance of the multivariate approaches in the determination of hydrophobic descriptors. the author reviews his own work as well as the significant contributions of other authors in the field. three original scientific papers belong to the field of solubility of drug substances: the paper on the solubility of indomethacin was presented by the group led by john comer and alex avdeef contributed two very interesting papers about the anomalous solubility behaviour of acidic and basic drugs. the issue concludes with two papers dealing with drug formulations and their improvement. radka opatrilova and josef jampilek studied the effect of microand nano-ionized alaptide on the skin permeation of mupirocine. rameshwar deshmukh and jitendra naik presented their preliminary work on the preparation of diclofenac microparticles. http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:kintam@umac.mo mailto:zmandic@fkit.hr http://www.iapchem.org/page.php?page_id=34 admet & dmpk 2(1) (2014) 1-2 pcmddd-3 editorial doi: 10.5599/admet.2.1.37 2 we wish to thank all the authors of this and the following special issue for their high-quality papers. we would also like to thank the referees who critically evaluated the papers at short notice. finally, we hope readers will enjoy the articles and find them interesting, useful and beneficial for their work. doi: 10.5599/admet.5.1.346 39 admet & dmpk 5(1) (2017) 39-46; doi: 10.5599/admet.5.1.346 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper biorelevant dissolution of candesartan cilexetil lucie gruberová* and bohumil kratochvíl department of solid state chemistry, university of chemistry and technology prague, technická 5, 166 28 praha *corresponding author: e-mail: gruberol@vscht.cz; tel.: +420 220 443 692 received: september 26, 2016; revised: december 16, 2016; published: march 25, 2017 abstract the choice of an appropriate medium for dissolution tests is an essential step during a dosage form development. the adequate design of dissolution testing enables forecasting in vivo behavior of drug formulation. biorelevant media were developed for this purpose because dissolution media described in the international pharmacopoeia are not thoroughly suitable. therefore, we carried out solubility and dissolution tests in biorelevant media and we compared the results with data measured in compendial dissolution media. a shake-flask method and standard paddle apparatus were used. the concentration was measured by a uv-vis spectrophotometer. an oral solid dosage form with poorly soluble drug candesartan cilexetil was tested. significant differences in the solubility and dissolution profiles of candesartan cilexetil were observed. the study offers the overview of compendial and biorelevant media simulating fasted state that can be analyzed by a spectrophotometric technique. keywords dissolution test; compendial medium; biorelevant medium; fasted state; poorly soluble drug; candesartan cilexetil. introduction during a drug dosage form development it is essential to investigate factors which influence drug absorption, especially after oral administration. the prediction of limiting factors can be facilitated by in vitro tests [1]. in vitro dissolution tests should mimic a drug performance in a human proximal gastrointestinal tract (git). to establish reliable in vitro testing it is important that artificial environments simulate physiological conditions as closely as possible. the level of the simulation is dependent on many factors related to used equipment and medium. a physiological relevant dissolution medium is great contribution to in vitro dissolution tests [2]. dissolution media were initially intended mainly for quality control purposes and water was frequently used as the dissolution medium. however, later approach led to the development of new media which resemble gastrointestinal fluids. simulated gastric fluid (sgf) and simulated intestinal fluid (sif) were the first proposals [3]. both artificial fluids are described in the international pharmacopoeias. these media reflect roughly ph conditions and the enzyme activity in the stomach (sgf contains pepsin, ph 1.2) and small intestinal (sif contains pancreatin, ph 6.8). nevertheless, there are variants without enzymes – sgfsp (sine pepsin) and sifsp (sine pancreatin) [4]. sgf and sif do not contain any surfactants, although http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:gruberol@vscht.cz gruberová and kratochvil admet & dmpk 5(1) (2017) 39-46 40 surfactants have been commonly used to improve the solubility of low soluble drugs. therefore, dressman et al. [5] and galia et al. [6] proposed addition of synthetic surfactants – triton x®-100 (sgftriton) and sodium lauryl sulfate (sgfsls). these media induced better solubility of low soluble compounds but on the other hand they were not physiologically relevant [7]. although the composition of artificial fluids reflects specific parameters of git physiology, they are not adequate models of in vivo conditions [2]. thus, fasted state simulated gastric fluid (fassgf) and fasted state simulated intestinal fluid (fassif) have been regarded as biorelevant media. fassgf was proposed by vertzoni et al. in 2005 [8] and fassif by galia at al. in 1998 [4]. both biorelevant media contain natural surfactants: taurocholate sodium, which represents bile salts, and lecithin. the medium fassgf comprises the enzyme pepsin as the medium sgf but pepsin is contained in very low concentration in fassgf. unlike sif, fassif does not comprise the enzyme pancreatin. all reported media were developed for the simulation of fasting conditions in the upper git. media simulating fed state of stomach (fed state simulated gastric fluid, fessgf) and small intestine (fed state simulated intestine fluid, fessif) have been also developed. former researches about the simulation of postprandial gastric conditions tested homogenized breakfasts with milk and nutrition products [2]. the first proposal of fessif was published by galia at al. in 1998 [4]. updates media (snapshot media) were proposed by jantratid et al. in 2008 [9]. however, this study is not interested in fed state media since their composition hinders direct spectrophotometric analysis. the biorelevant dissolution properties of many drugs have been studied since the beginning of biorelevant media development. studies of solubility in biorelevant media were carried out to predict bioavailability particularly of ionisable molecules since the solubility of ionisable compounds depends on their pka and ph of the media [10]. the solubilizing effect of bile salts, lecithin and other food effects on drug solubilization were tested for instance by takács-novák et al. [10], frank et al. [11], dressman at al. [12], and kostewicz et al. [13]. figure 1. structure of candesartan cilexetil (cc) the precise imitation of the in vivo conditions is desirable particularly for poorly soluble drugs. as a case example of poorly soluble drug candesartan cilexetil (figure 1) was chosen which is practically insoluble in water (less than 0.05 µg/ml [14]). as prodrug, candesartan cilexetil (cc) is completely bioactivated by ester hydrolysis to candesartan during a gastrointestinal absorption. actually, candesartan is an angiotensin ii receptor antagonist and it is mainly indicated for the treatment of hypertension [15]. however, cc is classified as bcs class ii drug and its very low solubility across the physiological ph range brings about an incomplete absorption and it is a reason for low bioavailability (about of 15 % [14]). the compound of our interest is a weak acid (pka 6.0 [16]) with lower solubility in low phs due to deprotonation of a tetrazole group at phs higher than 6 [16]. the results from our study confirm an increase in cc solubility with admet & dmpk 5(1) (2017) 39-46 biorelevant dissolution of candesartan cilexetil doi: 10.5599/admet.5.1.346 41 increasing ph. moreover, the solubility of cc is enhanced using surfactant tween 20 and the surfactant is required to achieve sink conditions. the dissolution testing of cc for quality controls has been standardly carried out in 900 ml of 0.05 m phosphate buffer (ph 6.5) with 0.35 % or 0.7 0% (w/w) tween 20 (shortcut ppt20). the amount of tween 20 depends on dosing conditions [17]. a big challenge with cc is its low stability in solutions. the most important factors affecting the stability of cc in solution are ph and temperature. hoppe and sznitowska [18] found out that the degradation rate of cc was faster at elevated temperatures and low phs. according to their results, the half-life of degradation is only 35.91 hours in 0.1 m hcl (ph 1.2) and 150.7 hours in 0.05 m phosphate buffer (ph 6.5), both at 37 °c and without surfactants [18]. our aim was to study and compare the dissolution properties of cc in media simulating the fasted state of stomach and small intestine. this paper provides the solubility data and dissolution profiles of cc in various media (artificial media with or without surfactant and enzyme, biorelevant media) which will be helpful for prediction of the in vivo performance of cc after oral administration. experimental materials candesartan cilexetil and immediate release (ir) tablets carzap (product of zentiva k.s., czech republic) with 8 mg of cc were obtained as gift samples. all chemicals used in the study were of analytical grade. the composition of the prepared media is presented in table 1 and table 2. table 1. composition of gastric media. component fassgf sgf sgfsp sgfsls sgftriton sodium taurocholate (mm) 0.08 lecithin (mm) 0.02 sodium lauryl sulfate (%, w/v) 0.25 triton x-100 (%, w/v) 0.10 pepsin (mg/ml) 0.10 3.20 sodium chloride (mm) 34.20 34.20 34.20 34.20 34.20 ph (adjusted by hydrochloric acid) 1.6 1.2 1.2 1.2 1.2 table 2. composition of intestinal media. component fassif sifsp ppt20 sodium taurocholate (mm) 3.00 lecithin (mm) 0.75 tween 20 (%, w/w) 0.35 pancreatin (mg/ml) monobasic sodium phosphate (mm) 28.36 monobasic potassium phosphate (mm) 49.97 50.00 sodium hydroxide (mm) 8.70 45.00 sodium chloride (mm) 105.85 ph (adjusted by sodium hydroxide) 6.5 6.8 6.5 gruberová and kratochvil admet & dmpk 5(1) (2017) 39-46 42 methods solubility studies the solubility of cc was determined by the shake-flask method. cc was added in excess into specific solvent and shaken for 24 hours at 37 °c in a thermostated orbital platform shaker (heidoplh, unimax 1010 with incubator 1000) to obtain equilibrium solubility [18, 19]. due to the low stability of cc, a ‘shortened shake flask method’ was suggest, the shaken time of certain saturated solutions, namely with solvents sgf, fassgf and sgftriton, was only 2 hours or 12 hours. the concentration of cc was measured by a uv-vis spectrophotometer (schimadzu uv mini 1240) at 256 nm and before the analysis supernatant had been filtered through a 35 micron porous filter from uhmw polyethylene. all tests were carried out in triplicate. concentration of dissolved cc was determined from calibration curve. the stock solution of cc was prepared in the suitable quantity of methanol. the aliquot of this solution was diluted with the specific solvent to get the final concentration of standard solutions (10-30 µg/ml). the method obeyed lambert -beer law (r 2 >0.99). dissolution studies dissolution tests were performed in a paddle apparatus (usp ii). the stirring rate was 50 rpm. the dissolution medium was filtered through a 35 micron porous filter from uhmw polyethylene, pumped through a 10 mm flow cell by a peristaltic pump (ismatic, reglo digital ms-2/6) and analyzed by a uv-vis spectrophotometer (schimadzu uv mini 1240), wavelenght was 256 nm. all dissolution tests were conducted in triplicate. the tablets of cc were placed in 500 ml of dissolution medium (37±0.5 °c). the volume was set as the smallest amount which enables the performance of the dissolution testing in usp ii [2]. to simulate the fasted state, the volume of media in the range of 250-300 ml (stomach) or 300-500 ml (duodenum) is recommended [8], but the volume reduction was not feasible. despite the volume reduction, sink conditions were maintained in medium ppt20 regarding on our solubility data of cc. these data verify findings reported by hoppe and sznitowska [18]. results and discussion solubility study the equilibrium solubility data (table 3) confirm that cc is a poorly soluble and stable compound. the values of cc solubility in sgf after 24 hours of shaking were so small that it was not possible to record them by the uv-vis spectrophotometer. zero solubility in sgf and very low solubility in fassgf (0.72±0.19 µg/ml) were caused by the rapid cc degradation at the acid environment. due to the long measurement time, the large amount of cc degraded. because of the cc degradation, the shaking time was shortened to two hours for media sgf, fassgf and sgftriton. from the obtained data, we concluded that the shorter shaking time reduced the amount of emerging degradation products (cc saturation solubility was 2.76 µg/ml in sgf and 0.90 µg/ml in fassgf). but the equilibrium may not be always reached in such short time, as evidenced by almost ten times lower value of cc solubility in sgftriton (0.69 µg/ml after 2 hours compared to 6.00 µg/ml after 24 hours test). cc gained the greatest solubility in fassgf after 12-hour period of shaking (1.59 µg/ml). for these data, it can be concluded that 12 hours is an optimal period for the shake-flask method considering the solubility and degradation of cc in fassgf. very low amounts of cc were dissolved in sgfsp and sifsp, 1.35 and 0.65 µg/ml, respectively. in this case, the low solubility of cc was probably the main reason of such results. we observed that the powder of cc admet & dmpk 5(1) (2017) 39-46 biorelevant dissolution of candesartan cilexetil doi: 10.5599/admet.5.1.346 43 was almost dry after the tests because in the media without enzymes and surfactants the required wetting of the samples was not reached. satturwar et al. published the values of cc solubility in sgf (0.6 µg/ml) and sif (8.6 µg/ml) [20]. however, the preparations of these media and the conditions of solubility study were not described in the research paper. besides, we suppose that the used media did not contain enzymes, which means sgfsp and sifsp, due to other procedures mentioned in the paper. our solubility data of cc measured by a shake-flask method in sgfsp and sifsp do not correspond with data by satturwar et al. [20]. our cc solubility in sgfsp is twice as high and cc solubility in sifsp 13 times as low than satturwar’s data. table 3. saturation solubility of cc in used media. medium solubility of cc (µg/ml) shortened test standard test (24-hour) ppt20 93.00±0.10 fassgf 0.90±0.07* 1.59±0.02** 0.72±0.19 sgf 2.76±0.46* sgfsp 1.35±0.10 sgftriton 0.69±0.07* 6.00±0.25 sgfsls 147.63±0.35 sifsp 0.65±0.01 fassif 8.26±0.18 * 2-hour test **12-hour test results demonstrate that only in the media sgfsls (147.63 µg/ml) and ppt20 (93.00 µg/ml) sink conditions for 8 mg tablets of cc can be achieved in volume 500 ml. three liters of fassif and four liters of sgftriton were required to achieve minimum sink conditions (approximately 30 % of the saturation concentration of cc) in dissolution studies. however, so large volumes are not physiologically relevant. based on solubility studies, it is expected that in the biorelevant media for fasted state conditions are not sufficient to achieve complete dissolution of the minimum dose. dissolution study the dissolution properties of cc are influenced by ph due to the ionization effect of cc at phs above 6.0, and also by the presence of a surfactant. the similar amount of cc (12-13 %) was dissolved in sgf and fassgf media (figure 2) at ph 1.2 and 1.6, respectively. it was predictable from the results of saturation solubility. dissolution testing proved the same dissolution rate of cc. very small divergence between profiles of sgf and fassgf is caused by the close values of ph and the low concentration level of the natural surfactants in fassgf. despite the fast degradation of cc at acid ph, the decrease of the cc concentration was not observed during one hour. however, 5 % decrease of the cc concentration was observed throughout the dissolution tests in sgfsp (figure 3). less considerable, but still noticeable, 0.5% reduction occurred also in case of the dissolution cc in sifsp. these courses of the dissolution profiles in sgfsp and sifsp prove the extremely short-time stability of cc in aqueous environment without surfactants or enzymes. the reason why the medium sifsp instead of sif was used is that pancreatin coloured the medium and sif could not be analyzed by a spectrophotometric technique. the amounts of dissolved cc in 35 minutes in sgf (12 %) and sifsp (10 %) were close to each other until the degradation process in sifsp occured. similar dissolution profiles were achieved in spite of the low ph value in sgf (sgf-ph 1.2 and sifsp-ph 6.8, the highest ph of the used media) on the one hand and the low equilibrium solubility of cc in sifsp on the other hand. nevertheless, the ionization of cc molecules at phs above 6.0 probably occurred gruberová and kratochvil admet & dmpk 5(1) (2017) 39-46 44 during the dissolution test. the ionization resulted in solubility increase and subsequently the degradation of dissolved compound. figure 2. dissolution profiles of candesartan cilexetil in medium ppt20, fassgf, sgf, sgftriton, sgfsls and fassif the most promising medium for the dissolution testing of cc was considered fassif, the medium with ph 6.5 and the natural surfactants taurocholate sodium and lecithin. however, the cc solubility in this medium was not sufficient. for this reason, sink conditions were not obtained in the paddle apparatus and therefore, the dissolution rate and extent of the cc solubility (46 %) was reduced (figure 2). figure 3. dissolution profiles of candesartan cilexetil in medium sgf, sgfsp and sifsp the biorelevant dissolution profiles of cc are not accordant with the dissolution profiles of cc in the media with the artificial surfactants, especially with the surfactant tween 20 and sodium lauryl sulfate (figure 2). it was found for cc that solubilization plays much more determining role in solubility than ionization. the solubility and dissolution rate of cc in the media ppt20 (96 %) and sgfsls (72 %) are considerately higher. in the case of sgfsls it is despite the low ph value (ph 1.2). the increase of the solubility in the medium containing the surfactant triton x®-100 is not so pronounced (31 %). this indicates admet & dmpk 5(1) (2017) 39-46 biorelevant dissolution of candesartan cilexetil doi: 10.5599/admet.5.1.346 45 how the choice of a surfactant influences the resulting solubility of a drug. it is evident that the application of dissolution media with artificial surfactants would lead to false positive results because artificial surfactants induce greater solubilization effects than what would be physiologically relevant. conclusions the need for in vitro test which mimics in vivo conditions of the gastrointestinal tract led to the development of physiologically relevant dissolution media. the dissolution of the poorly soluble drug cc was studied in different type of dissolution media: the compendial medium (ppt20) commonly used for quality control, the artificial fluids with the synthetic surfactants (sgftriton, sgfsls) or without them (sgf, sgfsp, sifsp) and the fasted state simulating biorelevant media (fassgf, fassif). in this study, the dissolution media that are possible to analyse by a uv-vis spectrophotometer were examined. consequently, dissolution tests with medium sif and biorelevant media simulated fed state (fessgf and fessif) were not carried out. the amount of dissolved cc in the biorelevant media was significantly lower than the amount of cc dissolved during tests with the media containing the synthetic surfactants. whereas the biorelevant media contain natural surfactants as an alternative to the non-physiologically relevant surfactants, the dissolution profiles of cc corresponded with the in vivo behavior of cc closely in fassgf and fassif. the poor solubility and stability of cc in the biorelevant media resulted in its low bioavailability. the results can be utilized as an example that the biorelevant media are useful to forecast the oral absorption of a bcs class ii drug. list of abbreviations cc candesartan cilexetil fassgf fasted state simulated gastric fluid fassif fasted state simulated intestinal fluid fessgf fed state simulated gastric fluid fessif fed state simulated intestinal fluid git gastrointestinal tract ppt20 phosphate buffer with surfactant tween 20 sgf simulated gastric fluid sgfsp simulated gastric fluid sine pepsin sgfsls simulated gastric fluid without enzyme but with surfactant sodium lauryl sulfate sgftritin simulated gastric fluid without enzyme but with surfactant tritonx® 100 sif simulated intestinal fluid sifsp simulated intestinal fluid sine pancreatin acknowledgements: financial support from specific university research (msmt no 20-svv/2016) and teva czech industries s.r.o. references [1] j. b. dressman, c. reppas, european journal of pharmaceutical sciences 11 (2000) s73-s80. [2] s. klein, aaps journal 12 (2010) 397-406. [3] j. b. dressman, dissolution technologies 6 (2014) 6-10. gruberová and kratochvil admet & dmpk 5(1) (2017) 39-46 46 [4] e. galia, e. nicolaides, d. hörter, r. löbenberg, c. reppas, j. b. dressman, pharmaceutical research 15 (1998) 698-705. [5] j. b. dressman, g. l. amidon, c. reppas, v. p. shah, pharmaceutical research 15 (1998) 11-22. [6] e. galia, j. horton, j. b. dressman, pharmaceutical research 16 (1999) 1871-1875. [7] m. vertzoni, e. pastelli, d. psachoulias, l. kalantzi, c. reppas, pharmaceutical research 24 (2007) 909-917. [8] m. vertzoni, j. b. dressman, j. butler, j. hempenstall, c. reppas, european journal of pharmaceutics and biopharmaceutics 60 (2005) 413-417. [9] e. jantratid, n. janssen, c. reppas, j. b. dressman, pharmaceutical research 25 (2008) 1663-1675. [10] k. takács-novák, v. szöke, g. völgyi, p. horváth, journal of pharmaceutical and biomedical analysis 83 (2013) 279-285. [11] k. j. frank, u. westendt, k. m. rosenblatt, p. hölig, j. rosenberg, m. mägerlein, m. brandl, g. fricker, european journal of pharmaceutical sciences 47 (2012) 16-20. [12] j. b. dressman, m. vertzoni, l. goumas, c. reppas, advanced drug delivery reviews 59 (2007) 591602. [13] e. s. kostewicz, u. brauns, r. becker, j. b. dressman, pharmaceutical research 19 (2002) 345-349. [14] b. abrahamsson, j. ödman, http://patentscope.wipo.int/search/en/wo2008030161 (2008). [15] m. burnier, circulation 103 (2001) 904-912. [16] e. cagigal, l. gonzáles, r. m. alonso, r. m. jiménez, journal of pharmaceutical and biomedical analysis 26 (2001) 477-486. [17] u.s. food and drug administration, http://www.accessdata.fda.gov/scripts/cder/dissolution/dsp_getalldata.cfm (1. 9. 2014). [18] k. hoppe, m. sznitowska, aaps pharmscitech 15 (2014) 1116-1125. [19] e. baka, j. e. a. corner, k. takács-novák, journal of pharmaceutical and biomedical analysis 46 (2008) 355-341. [20] p. satturwar, m. n. eddone, f. ravenelle, j.-c. leroux, european journal of pharmaceutics and biopharmaceutics 65 (2007) 379-387. ©2017 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://patentscope.wipo.int/search/en/wo2008030161 http://www.accessdata.fda.gov/scripts/cder/dissolution/dsp_getalldata.cfm http://creativecommons.org/licenses/by/3.0/ manuscript doi: 10.5599/admet.4.4.352 302 admet & dmpk 4(4) (2016) 302-313; doi: 10.5599/admet.4.4.322 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper tyrosine and aurora kinase inhibitors diminish transport function of multidrug resistance-associated protein (mrp) 4 and breast cancer resistance protein (bcrp) rhiannon n. hardwick 1 , marina snellings 2 , brian c. ferslew 2 , yang lu 2 , kim l.r. brouwer 1,2* 1 curriculum in toxicology, unc school of medicine, north carolina, us 2 division of pharmacotherapy and experimental therapeutics, unc eshelman school of pharmacy, the university of north carolina at chapel hill, us *corresponding author: e-mail: kbrouwer@unc.edu ; tel.: +1-919-962-7030; fax: +1-919-962-0644 received: june 27, 2016; revised: december 07, 2016; published: december 26, 2016 abstract tyrosine and aurora kinases are important effectors in signal transduction pathways that are often involved in aberrant cancer cell growth. tyrosine (tki) and aurora (aki) kinase inhibitors are anti-cancer agents specifically designed to target such signaling pathways through tki/aki binding to the atp-binding pocket of kinases thereby leading to diminished kinase activity. some tkis have been identified as inhibitors of atp-binding cassette (abc) transporters such as p-glycoprotein and breast cancer resistance protein (bcrp), which are commonly upregulated in malignant cells. tki/akis have been investigated as abc transporter inhibitors in order to facilitate the accumulation of concomitantly administered chemo therapeutics within cancer cells. however, abc transporters are prominently expressed in the liver and other eliminating organs, and their inhibition has been linked to intracellular accumulation of drugs, altered disposition, and toxicity. the potential for tkis/akis to inhibit other important hepatic efflux transporters, particularly multidrug resistance-associated proteins (mrps), remains unknown. the aim of the current study was to compare the inhibitory potency of 20 selected tki/akis against mrp4 and bcrp through the use of inverted membrane vesicle assays. relative ic50 values were estimated by determining tki/aki inhibition of mrp4-mediated [ 3 h]-dehydroepiandrosterone sulfate uptake and bcrp-mediated [ 3 h]estrone sulfate uptake. to provide insight to the clinical relevance of tki/aki inhibition of abc efflux transporters, the ratio of the steady-state maximum total plasma concentration (css) to the ic50 for each compound was calculated with css/ic50 ratio >0.1 deemed potentially clinically relevant. such analysis identified several potentially clinically relevant inhibitors of mrp4: alisertib, danusertib, erlotinib, lapatinib, neratinib, nilotinib, pazopanib, sorafenib, and tozasertib. the potentially clinically relevant inhibition of bcrp was much more extensive and included alisertib, barasertib, danusertib, enzastaurin, erlotinib, gefitinib, imatinib, neratinib, nilotinib, pazopanib, selumetinib, sorafenib, sunitinib, tozasertib, and vandetanib. these findings indicate the significant potential for tki/akis to inhibit multiple abc efflux transporters. the resulting inhibition data could provide insight regarding the clinical interpretation of pharmacokinetic/pharmacodynamic outcomes when tki/akis are administered concomitantly with additional chemotherapeutic agents. keywords abc transporters; transport inhibition http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:kbrouwer@unc.edu admet & dmpk 4(4) (2016) 302-313 tkis and akis diminish transport of function of mrp4 and bcrp doi: 10.5599/admet.4.4.322 303 introduction small molecule kinase inhibitors have revolutionized the field of oncology by facilitating the targeting of specific kinases, which have been found to be over-expressed and/or mutated in specific tumor types. such targeting has enabled a more directed, individualistic treatment regimen for many patients [1-4]. these small molecule kinase inhibitors have been designed primarily to target specific receptor tyrosine kinases, namely the epidermal growth factor receptor (egfr and her2), vascular endothelial growth factor receptor (vegfr), platelet-derived growth factor receptor (pdgfr), and the bcr/abl (breakpoint cluster region/abelson) fusion protein; however, some off-target inhibition of intracellular kinases can occur. receptor tyrosine kinases are important mediators of intracellular signaling cascades [4-6]. specifically, upon binding of ligand to the receptor tyrosine kinase, autocatalytic phosphorylation events spark the propagation of one or more signaling cascades ultimately leading to transcription of genes. aberrant regulation of receptor tyrosine kinases can result in amplification of genes related to proliferation, apoptosis, and differentiation, and thus enable cancer cell growth and metastasis [3, 6]. aurora kinases are different from receptor tyrosine kinases in that they are primarily active during mitosis, but also have been found to be upregulated in many tumor types [7, 8]. as such, several small molecule inhibitors have been designed in an effort to mitigate the activity of aurora kinases during cancer cell growth. though kinase inhibitors are designed with a particular target in mind, cross-reactivity with off-target kinases can occur in nonmalignant cells, which may increase the potential for off-target, potentially toxic effects [4, 9]. tkis and akis exert their effects by acting as atp-mimics and binding to the atp-binding pocket of kinases to decrease the kinase catalytic activity [2]. in recent years, tkis have been identified as inhibitors of several atp-binding cassette (abc) efflux transporters that are commonly upregulated in cancer cells including p-glycoprotein (multidrug resistance protein 1), breast cancer resistance protein (bcrp), and multidrug resistance-associated protein (mrp) 1 [10-18]. abc efflux transporters consist of a large family of membrane-spanning proteins involved in the active extrusion of substrates such as endogenous molecules, drugs, and drug metabolites from the cell through hydrolysis of atp [19-21]. abc efflux transporters are expressed prominently in eliminating organs such as the liver and kidney, and thus can play a significant role in the overall disposition of a drug within the body [19]. upregulation of abc efflux transporters in cancer cells confers chemoresistance, and imparts a greater ability of the cell to limit exposure to, and diminish the effects of, chemotherapeutic agents [12, 22, 23]. the upregulation of abc transporters during chemoresistance has made them an attractive target in overcoming the clinical difficulty of maintaining intracellular exposure to chemotherapeutics. due to their ability to inhibit abc transporters commonly upregulated in cancer cells, kinase inhibitors recently have been investigated for their capability to reverse chemoresistance. in particular, bcrp can confer resistance to gefitinib, but in contrast also is inhibited by gefitinib, vandetanib, and neratinib [18]. similarly, it has been shown that sunitinib is an inhibitor of bcrp and can reverse chemoresistance to co-administered chemotherapeutics in bcrp over-expressing cells [16]. others have shown that dasatinib, nilotinib, and imatinib are competitive inhibitors of p-glycoprotein and bcrp, meaning that they are substrates at low concentrations, while they inhibit transport function at high concentrations [24]. although these data indicate the potential for kinase inhibitors to be used as tools to overcome chemoresistance through abc efflux transporter inhibition in vitro, several clinical trials employing kinase inhibitors for mitigation of chemoresistance have been unable to definitively demonstrate improvements in progression free survival and/or overall survival of patients [25-28]. studies seeking to utilize kinase inhibitors for mitigation of chemoresistance though inhibition of specific abc efflux transporters that are upregulated in cancer cells often fail to address a key r.n. hardwick et al. admet & dmpk 4(4) (2016) 302-313 304 complication: off-target inhibition of abc efflux transporters in eliminating organs such as liver, kidney, and intestine. several mrps including mrp2, mrp3, and mrp4 are differentially localized to the apical and basolateral membranes of hepatocytes, enterocytes, and proximal tubule cells, which enables them to direct the disposition of some chemotherapeutics [19, 20]. for example, mrp2 is localized to the apical membrane of hepatocytes while mrp3 is situated at the basolateral membrane. mrp2 and mrp3 share substrate specificity yet extrude drugs and drug metabolites into the bile and systemic circulation, respectively [19]. inhibition of mrp2 can shift the hepatobiliary disposition of a drug from bile, and subsequent elimination in feces, to the sinusoidal blood resulting in greater systemic exposure. similar to mrp3, mrp4 resides on the basolateral membrane of hepatocytes, while bcrp, like mrp2, is localized to the apical membrane with each differentially extruding drugs and metabolites into the systemic circulation and bile, respectively. mrp4 transports nucleosides, phase ii conjugation products, and conjugated bile acids [20, 29]. inhibition of mrp4-mediated transport has been associated with an increased risk of cholestatic potential, particularly among compounds that do not inhibit the bile salt export pump (bsep) [29]. similarities in the chemical properties of mrp4 inhibitors and bsep inhibitors have been identified [30]. collectively, data suggest that mrp4 inhibition may serve as a confounding factor in bsep-mediated drug-induced liver injury or, in some cases, lead to drug-induced liver injury in the absence of bsep inhibition. thus, inhibition of one or more abc efflux transporters can significantly affect the disposition of a co-administered drug, and in some cases, may impact the risk of hepatotoxicity. knowledge of the inhibitory potency of tkis and akis for multiple abc efflux transporters, not limited to those upregulated in cancer cells, is critically important to predict the pharmacokinetic and pharmacodynamic outcomes and potential for hepatotoxicity of tki/akis and co-administered chemotherapeutics. the purpose of the current study was to compare the inhibitory potency of 20 tkis and akis (see table 1 for selected compounds) against the abc efflux transporters mrp4 and bcrp using inverted membrane vesicle assays. additionally, half maximal inhibitory concentration (ic50) data were leveraged against reported steady state concentration (css) values to provide insight to the clinical relevance of tki/aki inhibition of mrp4and bcrp-mediated transport function. table 1. selected tyrosine (tki) and aurora (aki) kinase inhibitors. twenty tki/akis were selected for analysis of their potential to inhibit mrp4and bcrp-mediated transport. tki/akis are categorized by primary kinase target. egfr, epidermal growth factor receptor; vegfr, vascular endothelial growth factor receptor; pdgfr, platelet-derived growth factor receptor; bcr/abl, breakpoint cluster region/abelson fusion protein. primary target inhibitors egfr erlotinib, gefitinib, lapatinib, neratinib, vandetanib vegfr & pdgfr axitinib, pazopanib, sorafenib, sunitinib bcr/abl bosutinib, dasatinib, imatinib, nilotinib miscellaneous crizotinib (anaplastic lymphoma kinase), enzastaurin (protein kinase c isoforms), selumetinib (mitogen-activated protein kinase 1) aurora alisertib (aurora a), barasertib (aurora b), danusertib (aurora a/b/c), tozasertib (aurora a/b/c) experimental materials cell culture supplies were obtained from gibco (life technologies, grand island, ny), and hygromycin b from roche life sciences (indianapolis, in). microscint-20, [ 3 h]-dehydroepiandrosterone-sulfate (dheas; 70.5 ci/mmol) and [ 3 h]-estrone sulfate (e1s; 45 ci/mmol) were purchased from perkin elmer life and analytical sciences (waltham, ma). dmso (dimethylsulfoxide) was obtained from fisher scientific (fairlawn, nh). compounds were purchased from selleckchem (houston, tx) or cayman chemical (ann arbor, mi). admet & dmpk 4(4) (2016) 302-313 tkis and akis diminish transport of function of mrp4 and bcrp doi: 10.5599/admet.4.4.322 305 cell culture and membrane vesicle preparation hek293t cells stably overexpressing mrp4 (hek-mrp4), hek293t cells transfected with empty vector (hek-mock), mdckii cells stably overexpressing bcrp (mdckii-bcrp), and untransfected mdckii cells were cultured in dmem supplemented with 10 % fbs as previously described [29, 31]. mrp4 or bcrp overexpression was maintained by antibiotic selection with hygromycin b [29]. cells were grown to ~85-95 % confluence before membrane vesicle preparation. membrane vesicles were prepared as previously described [31-34]. vesicles were validated for maximal activity and used in transport studies within 1-2 weeks of preparation. western blot analysis of transporter expression transporter over-expression was verified in control and transporter over-expressing membrane vesicles via western blot analysis. lysates of membrane vesicles (15 μg/well) were prepared in tris-glycine sds sample buffer (life technologies) and separated by sds-polyacrylamide gel electrophoresis (sds-page) using 4-12 % tris-glycine gels (life technologies). proteins separated by sds-page were transferred to polyvinylidene difluoride (pvdf) membranes overnight. protein levels on membranes were determined by incubation with antibodies directed against mrp4 (m4i-10, generously provided by george l. scheffer, amsterdam, the netherlands) or bcrp (bxp-21, kamiya biomedical company). membrane vesicle transport assay tkiand aki-mediated inhibition of transport function was investigated in membrane vesicles so as to study the function of a single transporter in isolation without confounding variables such as cellular uptake or protein binding. to confirm transport based on previously developed assays, transport of [ 3 h]dheas (2 µm, 0.7 µci/ml) by mrp4 or [ 3 h]e1s (1 µm; 0.4 µci/ml) by bcrp was measured over time via incubation of control (hek-mock or mdckii) and transporter over-expressing (hek-mrp4 or mdckii-bcrp) membrane vesicles in tris-sucrose buffer (50 mm tris/hcl and 250 mm sucrose, ph 7.4) in the presence or absence of an atp-regenerating system (4 mm atp, 10 mm mgcl2, 10 mm creatine phosphate, and 0.1 mg/ml creatine kinase). transporter substrate concentrations were selected based on km values [29, 34]. amp was replaced with atp for incubations conducted in the absence of an atp-regenerating system. the difference between uptake with atp versus amp was used to determine the atp-dependent uptake of [ 3 h]-dheas and [ 3 h]-e1s in mrp4 and bcrp membrane vesicles, respectively. for transport inhibition assays, 20 different tkior aki-test compounds (alisertib, axitinib, barasertib, bosutinib, crizotinib, danusertib, dasatinib, enzastaurin, erlotinib, gefitinib, imatinib, lapatinib, neratinib, nilotinib, pazopanib, selumetinib, sorafenib, sunitinib, tozasertib, or vandetanib) solubilized in dmso (1 % dmso final reaction concentration) were added in separate experiments to the reaction mixture containing control or transporter over-expressing membrane vesicles in tris-sucrose buffer with the atp-regenerating system. test compounds were examined at a concentration range of 0.316-100 μm. for potent bcrp inhibitors, the concentration range was reduced to 0.00316 μm. for control samples, identical concentrations of the solvent (1 % dmso) were used. all uptake studies were performed at 37 °c. uptake was stopped by addition of ice cold tris-sucrose buffer followed by rapid vacuum filtration through glass fiber 96-well plates pre-soaked in tris-sucrose buffer (bcrp reactions) or tris-sucrose buffer with 3 mm glutathione and 10 mm dithiothreitol (mrp4). microscint-20 (75 μl) was added to each well and the plate was allowed to rotate on an orbital shaker for 15 minutes. radioactivity was quantified by liquid scintillation counting in a topcount nxt instrument (perkin elmer). incubations with test compounds were completed in at least three independent experiments with no replicates in individual experiments. positive control inhibitors (mk571 r.n. hardwick et al. admet & dmpk 4(4) (2016) 302-313 306 for mrp4; sulfasalazine and promethazine for bcrp) were included in each individual experiment for quality control. data analysis and statistics substrate uptake into vesicles was calculated as pmol/mg protein/min before conversion to percent of control inhibition. inhibition by tki/akis as a percentage of control was calculated as the difference between substrate uptake in control and transporter-overexpressing membrane vesicles compared to vehicle-treated membrane vesicles (mean ± standard deviation, sd). dose-response ic50 curves were fit via nonlinear regression analysis according to the four-parameter logistic model (4pl) with the bottom constraint set to >0 using graphpad prism 5. for analysis of potential clinical relevance of transporter inhibition, steady-state total plasma css values for each tki/aki were extracted from micromedex and manufacturer package inserts. results and discussion mrp4 over-expression was confirmed in hek-mrp4 membrane vesicles, and was significantly higher than hek-mock membrane vesicles (figure 1a). the timeand atp-dependent uptake of [ 3 h]-dheas in hek-mrp4 vesicles was linear up to 10 minutes (figure 1a), similar to previous studies [29]. [ 3 h]-dheas uptake in hek-mrp4 vesicles was at least 75-fold greater than in hek-mock vesicles. an incubation time of 3.5 minutes was chosen for experiments with hek-mrp4 vesicles to evaluate the inhibition potential of test compounds, similar to previous studies [29]. the pan-mrp inhibitor mk571 was assessed across a concentration range of 0.316-100 μm. mk571 produced a prototypical dose-response curve ranging from nearly full activity to complete inhibition (figure 1b) with an ic50 value of 2.04 μm (95 % confidence interval: 1.31-3.19 μm). morgan, et al. previously reported an ic50 value of 6 µm for mk571-mediated inhibition of 3 h-estradiol-17β-d-glucuronide uptake in mrp4 membrane vesicles derived from sf9 cells [35]. mk571 was included as a positive control with all subsequent test compound experiments for quality assurance of the mrp4 vesicle assay. figure 1. characterization of the mrp4 membrane vesicle assay. (a) mrp4 expression in hek-mock and hek-mrp4 cells. also shown is atp-dependent, time-dependent uptake of 2 μm [ 3 h]-dehydroepiandrosterone (dheas) in membrane vesicles. data are presented as mean ± sd, n=3 independent experiments. (b) doseresponse ic50 curve of the mrp4 positive control inhibitor mk571 (3.5 min incubation). data are presented as mean % of control uptake ± sd, n=7 independent experiments. the inhibition potential of 20 tki/akis was assessed in control and mrp4 over-expressing membrane vesicles. dose-response curves of tkis are shown in figure 2. of the 16 tkis examined, 6 exhibited ic50 values <10 μm (table 2), namely, the egfr inhibitors erlotinib, gefitinib, lapatinib and neratinib; the vegfr & pdgfr inhibitor sorafenib; and the bcr/abl inhibitor nilotinib. others have reported the inhibitory admet & dmpk 4(4) (2016) 302-313 tkis and akis diminish transport of function of mrp4 and bcrp doi: 10.5599/admet.4.4.322 307 potency of some tkis against mrp4-mediated transport, particularly dasatinib, gefitinib, imatinib, lapatinib, neratinib, pazopanib, and sunitinib [35]. we found a similar ic50 for gefinitib inhibition of mrp4 (4.48 μm) to that of morgan, et al. (4.6 μm), and a somewhat comparable ic50 for dasatanib (18.03 μm in the present study versus 27.3 μm). however, the ic50 values for the remaining tkis evaluated by morgan, et al. were notably different, which may reflect differences in the experimental conditions such as the specific transporter substrate and range of inhibitor concentrations selected for investigation (e.g. extrapolation of an ic50 value beyond the tested concentrations) [35, 36]. the majority of the potent mrp4 inhibitors fell within the egfr tki class with only one potent inhibitor each in the dual vegfr & pdgfr and the bcr/abl classes. four akis were tested for their potential to inhibit mrp4 (figure 3). two akis, danusertib and tozasertib, produced ic50 values <10 μm (table 2). in particular, tozasertib exhibited comparable inhibitory potency to mk571 (1.87 ± 0.81 μm versus 2.04 μm, respectively). in total, eight of the 20 tki/akis assessed were identified as potent inhibitors (ic50 ≤ 10 μm) of mrp4. figure 2. mrp4 inhibition by tyrosine kinase inhibitors. inhibition of mrp4-mediated [ 3 h]-dheas (2 μm; 3.5 min) uptake across an inhibitor concentration range of 0.316-100 μm by egfr, vegfr/pdgfr, bcr/abl inhibitors, and inhibitors with miscellaneous targets is shown. data are presented as mean % of control uptake ± sd, n=3-6 independent experiments. bcrp over-expression in mdckii-bcrp membrane vesicles was determined via western blot analysis and compared to control mdckii vesicles (figure 4a). bcrp over-expression was confirmed to be substantially greater in mdckii-bcrp vesicles. to establish the bcrp membrane vesicle transport assay, the timeand atp-dependent uptake of [ 3 h]-e1s was measured in mdckii and mdckii-bcrp vesicles (figure 4a). the uptake of [ 3 h]-e1s was linear up to 1.5 minutes and appeared to saturate by 3 minutes. an incubation time of 1.5 minutes was chosen for subsequent assessment of potential bcrp inhibitors. there was at least a 35fold increase in [ 3 h]-e1s uptake in mdckii-bcrp versus mdckii vesicles. several compounds ascribed as bcrp inhibitors were assessed for their potential as positive controls in the membrane vesicle inhibition assay. sulfasalazine and promethazine were discovered to have the most consistent inhibition profile across multiple independent experiments (n=9-12). the dose-response curves for sulfasalazine and r.n. hardwick et al. admet & dmpk 4(4) (2016) 302-313 308 promethazine are shown in figure 4b. promethazine exhibited an ic50 value of 9.83 μm (95 % confidence interval: 5.14-18.80 μm), while sulfasalazine proved to be a much more potent inhibitor of bcrp with an ic50 value of 0.79 μm (95 % confidence interval: 0.64-0.99 μm). others have reported an ic50 of 0.6 µm for sulfasalazine-mediated inhibition of 3 h-estrone-3-sulfate uptake in bcrp membrane vesicles derived from hek-bcrp cells [37]. figure 3. mrp4 inhibition by aurora kinase inhibitors. inhibition of mrp4-mediated [ 3 h]-dheas (2 μm; 3.5 min) uptake across an inhibitor concentration range of 0.316-100 μm is shown. data are presented as mean % of control uptake ± sd, n=3-6 independent experiments. figure 4. characterization of the bcrp membrane vesicle assay. (a) bcrp expression in mdckii and mdckiibcrp cells. also shown is atp-dependent, time-dependent uptake of 1 μm [ 3 h]-estrone sulfate (e1s) in membrane vesicles. data are presented as mean ± sd, n=3 independent experiments. (b) dose-response ic50 curves of the bcrp positive control inhibitors sulfasalazine and promethazine are shown (1.5 min incubation). data are presented as mean % of control uptake ± sd, n=9-12 independent experiments. of the 16 tkis assessed, the majority (the egfr inhibitors erlotinib, gefitinib, neratinib and vandetanib; the vegfr & pdgfr inhibitors axitinib, pazopanib, sorafenib and sunitinib; the bcr/abl inhibitors bosutinib, imatinib and nilotinib; and the protein kinase c isoform inhibitor enzastaurin) were found to be potent inhibitors of bcrp with ic50 values <10 μm (figure 5). notably, erlotinib, gefitinib, sorafenib, imatinib, nilotinib, and enzaustaurin produced ic50 values <1 μm. a previous report identified axitinib as admet & dmpk 4(4) (2016) 302-313 tkis and akis diminish transport of function of mrp4 and bcrp doi: 10.5599/admet.4.4.322 309 only a partial inhibitor of bcrp with an ic50 value of 4.4 μm; however the experiments were performed in mcdk-bcrp cell monolayers and may not be directly comparable to isolated membrane vesicle systems [14]. tki inhibition potency towards bcrp was not overrepresented in one inhibitor class versus another as was seen for egfr inhibitors with respect to mrp4. one tki, lapatinib, did not produce a dose-response curve and was deemed to not be an inhibitor of bcrp. dose-response curves for aki inhibition of bcrpmediated transport are shown in figure 6. in the aki class, alisertib, barasertib, and danusertib were identified as potent inhibitors of bcrp with ic50 values <10 μm (table 2). unlike the tkis, none of the akis produced ic50 values <1 μm and thus appeared to be slightly less potent inhibitors of bcrp. figure 5. bcrp inhibition by tyrosine kinase inhibitors. inhibition of bcrp-mediated [ 3 h]-e1s (1μm; 1.5 min) uptake across an inhibitor concentration range of 0.00316-100 μm (0.00316-1 μm or 0.0316-10 μm for potent inhibitors) by egfr, vegfr/pdgfr, bcr/abl inhibitors, and inhibitors with miscellaneous targets is shown. data are presented as mean % of control uptake ± sd, n=3-7 independent experiments. r.n. hardwick et al. admet & dmpk 4(4) (2016) 302-313 310 figure 6. bcrp inhibition by aurora kinase inhibitors. inhibition of bcrp-mediated [ 3 h]-e1s (1 μm; 1.5 min) uptake across an inhibitor concentration range of 0.316-100 μm is shown. data are presented as mean % of control uptake ± sd, n=3-4 independent experiments. to provide insight to the clinical relevance of tki/aki inhibition of mrp4and bcrp-mediated transport, steady-state total plasma css values were extracted from micromedex and manufacturer package inserts. the css values were expressed as a ratio to the ic50 for each transporter with a css/ic50 ratio >0.1 regarded as having potential for clinically relevant transporter inhibition [35, 38]. although it is well-recognized that drug concentrations within cells may become more concentrated than measured css in patient plasma, either due to transporter-mediated accumulation or organelle binding, the use of total plasma css values have been recommended and are used herein in the absence of unbound intracellular drug concentration data [35, 38, 39]. css/ic50 ratios are shown in table 2. nine of the 20 tki/akis tested exhibited css/ic50 ratios >0.1 for mrp4. in particular, alisertib, danusertib, erlotinib, lapatinib, neratinib, nilotinib, pazopanib, sorafenib, and tozasertib were identified as having potential for clinically relevant mrp4 inhibition. although the majority of potent mrp4 inhibitors (ic50 ≤ 10 μm) were found to be in the egfr inhibitor class, the tki/akis identified as having potential for clinical relevance were primarily stratified across both egfr and aurora kinase inhibitor classes. only 2 vegfr & pdgfr and 1 bcr/abl kinase inhibitors were found to be clinically relevant. similar to the report of morgan, et al., pazopanib was identified as having the potential for clinically relevant mrp4 inhibition [35]. additionally, though the studies of morgan, et al., also identified gefitinib as a potent inhibitor of mrp4, gefitinib-mediated inhibition of mrp4 transport activity was not deemed clinically relevant [35], which is similar to the current results. the majority of the tki/akis investigated were found to potentially be clinically relevant inhibitors of bcrp, with 15 of the 20 compounds having a css/ic50 ratio >0.1. specific compounds identified were alisertib, barasertib, danusertib, enzastaurin, erlotinib, gefitinib, imatinib, neratinib, nilotinb, pazopanib, selumetinib, sorafenib, sunitinib, tozasertib, and vandetanib. several compounds, notably alisertib, danusertib, erlotinib, neratinib, nilotinib, pazopanib, sorafenib, and tozasertib were found to be dual inhibitors of mrp4 and bcrp with potential clinical relevance for both. dual inhibition may further complicate drug-mediated transporter inhibition by minimizing compensatory mechanisms for export of compounds from within a cell. in summation, the analysis of css/ic50 ratios revealed a greater potential for clinically relevant bcrp inhibition admet & dmpk 4(4) (2016) 302-313 tkis and akis diminish transport of function of mrp4 and bcrp doi: 10.5599/admet.4.4.322 311 by tki/akis compared to mrp4, though eight tki/akis were found to be dual inhibitors of both transporters. table 2. summary of tki and aki inhibition of mrp4 and bcrp. data are presented as mean relative ic50 estimates (μm). total plasma cmax values (μm) were extracted from micromedex 2.0 (truven health analytics) and manufacturer package inserts. n/a = data not available. asterisk (*) indicates cmax/ic50 ratio >0.1 for an individual transporter. tki/aki steady-state total css (μm) mean relative mrp4 ic50 estimate (μm, ±sd) css/ic50 mrp4 mean relative bcrp ic50 estimate (μm, ±sd) css/ic50 bcrp alisertib 1.80 16.54 (±4.34) 0.109* 3.00 (±0.38) 0.600* axitinib 0.07 14.99 (±5.29) 0.005 1.56 (±1.07) 0.045 barasertib 2.30 86.30 (±81.93) 0.027 7.62 (±7.61) 0.302* bosutinib 0.38 16.75 (±14.88) 0.023 9.11 (±1.81) 0.042 crizotinib 1.09 n/a (outside concentration range) n/a 15.90 (±8.23) 0.069 danusertib 9.34 8.70 (±5.18) 1.074* 7.49 (±2.70) 1.248* dasatinib 0.13 18.03 (±4.42) 0.007 13.17 (±2.50) 0.010 enzastaurin 2.19 43.93 (±23.77) 0.050 0.97 (±0.34) 2.266* erlotinib 3.95 3.11 (±1.16) 1.270* 0.11 (±0.02) 36.127* gefitinib 0.43 4.48 (±1.64) 0.096 0.12 (±0.01) 3.728* imatinib 2.72 n/a (partial inhibitor) n/a 0.29 (±0.15) 9.292* lapatinib 4.18 8.33 (±5.68) 0.502* n/a n/a neratinib 2.95 6.65 (±0.66) 0.444* 1.77 (±1.44) 1.669* nilotinib 3.85 6.18 (±4.82) 0.623* 0.05 (±0.01) 81.828* pazopanib 122.58 22.92 (±13.82) 5.349* 9.43 (±2.85) 13.000* selumetinib 2.08 103.61 (±18.42) 0.020 15.04 (±6.21) 0.138* sorafenib 21.51 2.60 (±0.33) 8.261* 0.26 (±0.08) 82.078* sunitinib 0.23 91.53 (±40.46) 0.003 1.93 (±0.37) 0.119* tozasertib 2.67 1.87 (±0.81) 1.427* 13.74 (±3.98) 0.194* vandetanib 1.93 33.18 (±18.90) 0.058 8.70 (±3.20) 0.222* conclusions together, these data represent the first analysis of the inhibitory potential of multiple tki/akis against both mrp4 and bcrp, along with estimation of the clinical relevance of tki/aki-facilitated transport inhibition. the data herein may enable greater predictive power of the pharmacokinetic/pharmacodynamic and potential hepatotoxic effects of tki/aki use in co-treatment regimens. acknowledgements this work was supported by the national institute of general medical sciences of the national institutes of health (nih) under award number r01 gm041935. dr. hardwick was supported by the national institute of environmental health sciences under award number t32 es007126. dr. ferslew was supported by an amgen predoctoral fellowship in pharmacokinetics and drug disposition. mdckii and mdckii-bcrp cell lines were a generous gift from dr. markus grube, ernst moritz arndt university of greifswald, greifswald, germany. hek-mock and hek-mrp4 cell lines were created by dr. kathleen kӧck, unc-ch, using plasmids generously donated by dr. dietrich keppler, german cancer research center, heidelberg, germany. references [1] n.p. van erp, h. gelderblom, h.j. guchelaar, cancer treatment reviews 35 (2009) 692-706. [2] j. zhang, p.l. yang, n.s. gray, nature reviews. cancer 9 (2009) 28-39. [3] a. arora, e.m. scholar, the journal of pharmacology and experimental therapeutics 315 (2005) 971979. r.n. hardwick et al. admet & dmpk 4(4) (2016) 302-313 312 [4] s. niijima, a. shiraishi, y. okuno, journal of chemical information and modeling 52 (2012) 901-912. [5] j. andrae, r. gallini, c. betsholtz, genes & development 22 (2008) 1276-1312. [6] r.a. bradshaw, r.j. chalkley, j. biarc, a.l. burlingame, advances in biological regulation 53 (2013) 8796. [7] m. kollareddy, p. dzubak, d. zheleva, m. hajduch, biomedical papers of the medical faculty of the university palacky, olomouc, czechoslovakia 152 (2008) 27-33. [8] m.r. green, j.e. woolery, d. mahadevan, expert opinion on drug discovery 6 (2011) 291-307. [9] j.t. hartmann, m. haap, h.g. kopp, h.p. lipp, current drug metabolism 10 (2009) 470-481. [10] a. azzariti, l. porcelli, g.m. simone, a.e. quatrale, n.a. colabufo, f. berardi, r. perrone, m. zucchetti, m. d'incalci, j.m. xu, a. paradiso, cancer chemotherapy and pharmacology 65 (2010) 335-346. [11] a. brozik, c. hegedus, z. erdei, t. hegedus, c. ozvegy-laczka, g. szakacs, b. sarkadi, expert opinion on drug metabolism & toxicology 7 (2011) 623-642. [12] c. ozvegy-laczka, j. cserepes, n.b. elkind, b. sarkadi, drug resistance updates : reviews and commentaries in antimicrobial and anticancer chemotherapy 8 (2005) 15-26. [13] m. galetti, p.g. petronini, c. fumarola, d. cretella, s. la monica, m. bonelli, a. cavazzoni, f. saccani, c. caffarra, r. andreoli, a. mutti, m. tiseo, a. ardizzoni, r.r. alfieri, plos one 10 (2015) e0141795. [14] e.l. reyner, s. sevidal, m.a. west, a. clouser-roche, s. freiwald, k. fenner, m. ullah, c.a. lee, b.j. smith, drug metabolism and disposition: the biological fate of chemicals 41 (2013) 1575-1583. [15] x.k. wang, l.w. fu, current drug metabolism 11 (2010) 618-628. [16] s. shukla, r.w. robey, s.e. bates, s.v. ambudkar, drug metabolism and disposition: the biological fate of chemicals 37 (2009) 359-365. [17] y. wei, y. ma, q. zhao, z. ren, y. li, t. hou, h. peng, molecular cancer therapeutics 11 (2012) 16931702. [18] c. hegedus, k. truta-feles, g. antalffy, g. varady, k. nemet, c. ozvegy-laczka, g. keri, l. orfi, g. szakacs, j. settleman, a. varadi, b. sarkadi, biochemical pharmacology 84 (2012) 260-267. [19] n.d. pfeifer, r.n. hardwick, k.l. brouwer, annual review of pharmacology and toxicology 54 (2014) 509-535. [20] c.d. klaassen, l.m. aleksunes, pharmacological reviews 62 (2010) 1-96. [21] k. kock, k.l. brouwer, clinical pharmacology and therapeutics 92 (2012) 599-612. [22] f.j. sharom, pharmacogenomics 9 (2008) 105-127. [23] y. an, w.m. ongkeko, expert opinion on drug metabolism & toxicology 5 (2009) 1529-1542. [24] m. dohse, c. scharenberg, s. shukla, r.w. robey, t. volkmann, j.f. deeken, c. brendel, s.v. ambudkar, a. neubauer, s.e. bates, drug metabolism and disposition: the biological fate of chemicals 38 (2010) 1371-1380. [25] a. argiris, m. ghebremichael, j. gilbert, j.w. lee, k. sachidanandam, j.m. kolesar, b. burtness, a.a. forastiere, journal of clinical oncology : official journal of the american society of clinical oncology 31 (2013) 1405-1414. [26] l.g. paz-ares, b. biesma, d. heigener, j. von pawel, t. eisen, j. bennouna, l. zhang, m. liao, y. sun, s. gans, k. syrigos, e. le marie, m. gottfried, j. vansteenkiste, v. alberola, u.p. strauss, e. montegriffo, t.j. ong, a. santoro, n.r.e.u.s.i.s. group, journal of clinical oncology : official journal of the american society of clinical oncology 30 (2012) 3084-3092. [27] t.k. choueiri, r.w. ross, s. jacobus, u. vaishampayan, e.y. yu, d.i. quinn, n.m. hahn, t.e. hutson, g. sonpavde, s.c. morrissey, g.c. buckle, w.y. kim, d.p. petrylak, c.w. ryan, m.a. eisenberger, a. mortazavi, g.j. bubley, m.e. taplin, j.e. rosenberg, p.w. kantoff, journal of clinical oncology : official journal of the american society of clinical oncology 30 (2012) 507-512. [28] k.t. flaherty, s.j. lee, f. zhao, l.m. schuchter, l. flaherty, r. kefford, m.b. atkins, p. leming, j.m. kirkwood, journal of clinical oncology : official journal of the american society of clinical oncology 31 (2013) 373-379. admet & dmpk 4(4) (2016) 302-313 tkis and akis diminish transport of function of mrp4 and bcrp doi: 10.5599/admet.4.4.322 313 [29] k. kock, b.c. ferslew, i. netterberg, k. yang, t.j. urban, p.w. swaan, p.w. stewart, k.l. brouwer, drug metabolism and disposition: the biological fate of chemicals 42 (2014) 665-674. [30] m.a. welch, k. kock, t.j. urban, k.l. brouwer, p.w. swaan, drug metabolism and disposition: the biological fate of chemicals 43 (2015) 725-734. [31] m. grube, s. reuther, h. meyer zu schwabedissen, k. kock, k. draber, c.a. ritter, c. fusch, g. jedlitschky, h.k. kroemer, drug metabolism and disposition: the biological fate of chemicals 35 (2007) 30-35. [32] k. kock, a. koenen, b. giese, m. fraunholz, k. may, w. siegmund, e. hammer, u. volker, g. jedlitschky, h.k. kroemer, m. grube, the journal of biological chemistry 285 (2010) 11336-11347. [33] g. ghibellini, e.m. leslie, g.m. pollack, k.l. brouwer, pharmaceutical research 25 (2008) 1851-1860. [34] b.c. ferslew, k. kock, a.s. bridges, k.l. brouwer, drug metabolism and disposition: the biological fate of chemicals 42 (2014) 1567-1574. [35] r.e. morgan, c.j. van staden, y. chen, n. kalyanaraman, j. kalanzi, r.t. dunn, 2nd, c.a. afshari, h.k. hamadeh, toxicological sciences : an official journal of the society of toxicology (2013). [36] c.j. van staden, r.e. morgan, b. ramachandran, y. chen, p.h. lee, h.k. hamadeh, current protocols in toxicology / editorial board, mahin d. maines chapter 23 (2012) unit 23 25. [37] j.e. karlsson, c. heddle, a. rozkov, j. rotticci-mulder, o. tuvesson, c. hilgendorf, t.b. andersson, drug metabolism and disposition: the biological fate of chemicals 38 (2010) 705-714. [38] m.j. zamek-gliszczynski, c.a. lee, a. poirier, j. bentz, x. chu, h. ellens, t. ishikawa, m. jamei, j.c. kalvass, s. nagar, k.s. pang, k. korzekwa, p.w. swaan, m.e. taub, p. zhao, a. galetin, c. international transporter, clinical pharmacology and therapeutics 94 (2013) 64-79. [39] n.d. pfeifer, k.b. harris, g.z. yan, k.l. brouwer, drug metabolism and disposition: the biological fate of chemicals (2013). ©2016 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ manuscript doi: 10.5599/admet.3.4.261 296 admet & dmpk 3(4) (2015) 296-297; doi: 10.5599/admet.3.4.261 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index editorial pka values in solubility determination using the hendersonhasselbalch equation derek reynolds and kin tam admet & dmpk editorial board, email: admet_dmpk@iapchem.org. published: december 30, 2015 in this journal we publish only high quality peer reviewed scientific papers but sometimes not all scientists can agree with all aspects of a particular paper. we wish to encourage healthy scientific debate and when appropriate it is right to publish alternative interpretations of reported experiments. alex avdeef [1] and samuel yalkowsky [2] have independently studied the solubility and ionization properties of an investigational drug (nsc-639829). in the study conducted by avdeef and coworkers, the pka value derived from henderson-hasselbalch (hh) equation was 4.70, which was different from the value of 3.76 as determined experimentally by the yasuda-shedlovsky plot using the pka values obtained from co-solvent uv titrations. the authors then invoked an ionization-precipitation-aggregation model by assuming the formation of a positively charged dimer to generate the theoretical solubility-ph profile. in their implementation, the pka was fixed at 3.76, while the s0, and the aggregation equilibrium constant were treated as adjustable parameters. to this end, the authors obtained a reasonably good fit between the model and the experiment. the s0 value as determined using this approach appeared to be in line with the previously reported value [2]. in this issue, yalkowsky et al. [3] published a note to comment on the approach that avdeef et al. [1] developed in the study of nsc-639829. in particular, they have pointed out that the pka of 4.70 for nsc639829 was the correct value in the solubility determination study, and the analysis carried out by avdeef et al. [1] was problematical. yalkowsky et al. have highlighted that pka is a concentration dependent parameter, which was supported by several literature examples [3]. apparently, yalkowsky et al. have put forward some compelling arguments to justify their points of view in case of nsc-639829. generally speaking, pka is not an absolute molecular constant because it depends very much on the medium, conditions and ionic strength of determination. for low sample concentration and/or in the presence of cosolvent, the formation of aggregate in pka determination is less favorable. in case of nsc-639829, the pka determination was accomplished by long range extrapolation from co-solvent pka data (22-41 % methanolwater mixtures) [1]. again, this may introduce some uncertainty to the extrapolated pka value. it may be difficult to make direct comparison between the pka determined independently and the pka derived from the hh equation since the chemical/physical environments are unlikely to be the same in these two experiments. in our opinion, avdeef et al. [1] have developed very powerful software to analyze complex ionizationprecipitation-aggregation system, which we should applaud. on the other hand, yalkowsky et al. pointed http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:admet_dmpk@iapchem.org admet & dmpk 3(4) (2015) 296-297 editorial on henderson-hasselbalh equation doi: 10.5599/admet.3.4.261 297 out that effectively what the software does in the case of nsc-639829 was adding an arbitrary factor of 0.94 to bring the pka to 4.70. we see that these are just two different approaches to interpret the same dataset. perhaps the aggregate formation may well be true as proposed by avdeef et al. [1]. however, it is always difficult to prove the formation of the positively charged dimer in the equilibrium model solely based on mass balance calculation. to confirm the validity of the model, it is necessary to obtain independent structural or spectroscopic evidence. although this discussion centers around just one compound it is part of a wider debate regarding the applicability of the hh equation to the study of solubility where complications are likely to arise due to presence of aggregate species in solution. all the authors of these papers have made significant contributions to the field. we hope that many scientists with an interest in solubility and ionization will read both papers and learn from the differing views of these eminent researchers. references [1] g. butcher, j. comer, a. avdeef, admet & dmpk 3(2) (2015) 131-140. [2] n. jain, g. yang, s.e. tabibi, s.h. yalkowsky, int. j. pharm. 225 (2001) 41-47. [3] s. yalkowsky, r. patel, d. al-antary, admet & dmpk 3(4) (2015) 358-361. ©2015 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ 60 admet & dmpk 4(2) (2016) 60-61; doi: 10.5599/admet.4.2.320 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index editorial changing trends in use of hepatocytes and microsomes for evaluating metabolism studies in drug discovery abdul naveed shaik guest editor: admet & dmpk center for pharmacometrics and systems pharmacology, college of pharmacy, university of florida, orlando, fl email: naveedshaik@gmail.com; tel.: +1-217-413-7760. a superfamily of hemoprotein named cytochrome p450 (cyps) are the proteins which bring about catalysis of many endo biotics as well as xenobiotics. these are the membrane associated proteins located in the endoplasmic reticulum. as these proteins bring about oxidative metabolism of many xenobiotics, they have been since used to in drug discovery and development extensively to study the drug metabolism of new chemical entities (nces). in humans, these cyps are localized in different organs, such as liver, brain, intestine and kidney. in the liver, cyps are present in the endoplasmic reticulum of hepatocytes. ultracentrifugation of hepatocytes at a speed of 100,000 x g will generate microsomes. microsomes have been traditionally used in drug discovery to evaluate the metabolic properties of nces. they contain both phase i enzymes such as a cyps, flavin mono oxygenase (fmo) and phase ii enzymes such as uridine 5'-diphospho-glucuronosyltransferases (ugts), which can be regarded as a convenient metabolic tool in drug discovery. there are few limitations of microsomes. for instance, they have to be supplied with co-factors like nadph, or udpga for the metabolic reactions to initiate, and lack the cell membranes to mimic the physiological environment in hepatocytes. however, in the cell system the drugs never have easy access to the drug metabolizing enzymes. they have to pass through barriers which will have drug transporters facilitating the movement of these drugs across the membrane either concentrating the cell, granting access or effluxing the drug and denying the drugs access to the enzymes. due to the involvement of these drug transporters, in drug disposition and drug-drug interactions, there has been major interest in the past decade to study the role of these transporters in drug discovery and development. hepatocytes have a cellular organization containing, the necessary co-factors and membranes transporters, which represent a more complete system to study drug metabolisms. fresh cryopreserved hepatocytes have been made commercially available in recent years. there has been a change in trend to use hepatocytes in drug discovery and development. due to the high cost of hepatocytes and the complexity in conducting experiments with hepatocytes, microsomes still remain the tool of choice for early drug discovery studies. admet & dmpk devoted a special issue on use of microsomes and hepatocytes in drug discovery with the aim to analyze the trend in use of these tools in drug discovery and development. there have been few original articles as well as review articles submitted to this special issue, these articles covered varied areas including, fetal hepatocytes, pharmacogenetics profiling and metabolic activity in stem cell derived http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:naveedshaik@gmail.com 61 hepatocytes, microsomes in brain, admet properties of natural products and an opinion on use of hepatocytes and microsomes in various stages of drug discovery and development. clearly microsomes and hepatocytes are playing an important role in supporting the metabolic study in drug discovery. in the near future, we will call for more special issues related to drug metabolisms and drug transporters. ©2016 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ manuscript doi: 10.5599/admet.5.2.402 57 admet & dmpk 5(2) (2017) 57-58; doi: 10.5599/admet.5.2.402 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index editorial role of drug transporters in drug development: a qualitative and quantitative approach sumit basu, abdul naveed shaik * guest editors; admet & dmpk center for pharmacometrics and systems pharmacology, university of florida, orlando, florida. usa *corresponding author: e-mail: naveedshaik@gmail.com; tel.: +1-407-313-7009; fax: +1-407-313-7030 received: june 21, 2017; published: june 22, 2017 through last two decades, transporters have been recognized as one of the important elements which plays an important role to regulate the adme (absorption, distribution, metabolism and excretion) properties of drugs and xenobiotics. to evaluate the role of the transporters in drug development, an international transporter consortium (itc) discussed about the identification of the clinically relevant transporters, application of these transporters in different drug-drug interaction (ddi) studies and the different criteria to determine whether additional ddi studies are required. owing to the increase of transporter research studies in the last decade and the emphasis of the transporter mediated ddi by different regulatory agencies (fda, ema and pmda), the incorporation of transporter related information in the different drug “label” has increased from 24 % to 56 % within five years. in addition, the transporter related post marketing studies became quite important to address different questions regarding dose optimization and in the revision of drug product labelling. in general, the role of transporters varies depending on their function and localization within the gut lumen, liver, kidney and brain. depending on the directions of the transport, they can be classified as either efflux or influx/uptake. usually efflux transporters shuttle compounds out of the cell against the concentration gradient whereas the uptake transporters direct compounds into the cells along the concentration gradient. since many transporters have the ability to recognize more than one substrate, different types of ddi have been noticed due to the inhibition and induction of the function of the transporters. in addition, many transporters demonstrated genetic polymorphisms which may affect their function. traditionally, in vitro tools and preclinical animal experiments have often been used to assess the role of efflux transporters in drug development. carefully designed and properly planned studies can give a clear understanding of the intrinsic functions and attributes of the study. the accumulation of the current in vitro and preclinical animal knowledge base is a conscious effort to link the clinical endpoints with the early developmental work which ultimately will help to identify the effect of drug transporters in the clinical settings. however, owing to the broad substrate specificity as well as significant overlap of inhibitor specificity, one should always investigate the experimental design and the results to determine the combined effects of particular transporters. similarly, naturally occurring genetic variabilities arising from http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:naveedshaik@gmail.com s. basu and a.n. shaik admet & dmpk 5(2) (2017) 57-58 58 genetic polymorphisms can also affect both drug absorption and disposition which often results in inconclusive results. although the lack of ideal probes and inhibitors has confounded the transporter data interpretations for intestinal absorption, the role of efflux in other organs (blood-brain-barrier) is better understood. with the increasing interest in drug transporters from pharmaceutical companies and regulatory authorities alike and following the success of a previous special issue on drug metabolism [1] as well as a special issue on drug transporters and permeability (2015, volume 3, issue 1), admet & dmpk devoted a special issue on “role of drug transporters in drug development: a qualitative and quantitative approach” with the focus to explore the current trend in the uses of these transporters at various stages of drug discovery and development. we received a few original articles and two review articles, which covered various areas including, uric acid transporters and its inhibitors in gout, perillyl alcohol mediated abrogated membrane transport in candida albicans, use of different fluorescent organic cations for screening against oct2 transporters and expression of drug transporters in human skin and its uses in drug development. one particular submission of a great interest was “consensus rank orderings of molecular fingerprints illustrate the ‘most genuine’ similarities between marketed drugs and small endogenous human metabolites, but highlight exogenous natural products as the most important ‘natural’ drug transporter substrates” where the authors s. o’hagan and d.b. kell brought about the similarities between small endogenous metabolites and marketed drugs. these submissions clearly indicate the utilization of various drug transporters at different stages of drug discovery and development. as there has been significant interest in this special issue focussing on drug transporters we will follow with a second edition of the special issue to accommodate other submissions. references [1] a.n. shaik, changing trends in use of hepatocytes and microsomes for evaluating metabolism studies in drug discovery, admet and dmpk 4 (2016), 60-61. ©2017 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ manuscript doi: 10.5599/admet.1.4.24 48 admet & dmpk 1(4) (2013) 48-62; doi: 10.5599/admet.1.4.24 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper salt solubility products of diprenorphine hydrochloride, codeine and lidocaine hydrochlorides and phosphates – novel method of data analysis not dependent on explicit solubility equations gergely völgyi 1 , attila marosi 1 , krisztina takács-novák 1 , alex avdeef 2 * 1 semmelweis university, department of pharmaceutical chemistry h-1092 hőgyes endre street 9. budapest, hungary 2 in-adme research, 1732 first avenue, #102, new york, ny 10128, usa *corresponding author: alex avdeef; e-mail: alex@in-adme.com; tel.: +1 647 678 5713; received: december 6, 2013; revised: december 11, 2013; published: december 16, 2013 abstract a novel general approach was described to address many of the challenges of salt solubility determination of drug substances, with data processing and refinement of equilibrium constants encoded in the computer program pdisol x tm . the new approach was illustrated by the determinations of the solubility products of diprenorphine hydrochloride, codeine hydrochloride and phosphate, lidocaine hydrochloride and phosphate at 25 o c, using a recently-optimized saturation shake-flask protocol. the effects of different buffers (britton-robinson universal and sörensen phosphate) were compared. lidocaine precipitates were characterized by x-ray powder diffraction (xrpd) and polarization light microscopy. the ionic strength in the studied systems ranged from 0.25 to 4.3 m. codeine (and possibly diprenorphine) chloride were less soluble than the phosphates for ph > 2. the reverse trend was evident with lidocaine. diprenorphine saturated solutions showed departure from the predictions of the henderson-hasselbalch equation in alkaline (ph > 9) solutions, consistent with the formation of a mixed-charge anionic dimer. keywords sparingly-soluble drugs; ph-dependent solubility; salt solubility products, solubility equations; aggregation; shake-flask method. introduction salt selection in preformulation is an important step in the preparation of effective oral drug formulations [1-8]. salt solubility, being a conditional constant, takes on different values according to the concentrations and types of reactants used in a particular study. for this reason, laboratory-to-laboratory comparisons can be complicated, possibly leading to conflicting interpretations of in vitro dissolution studies in formulation development. on the other hand, salt solubility products are true equilibrium constants. but these are not often reported in published salt solubility studies. interpretation and scaling of salt solubility measurements can be very challenging for a number of reasons. http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:alex@in-adme.com admet & dmpk 1(4) (2013) 48-62 novel method of salt solubility analysis doi: 10.5599/admet.1.4.24 49  since drug salts are often much more soluble than the corresponding uncharged forms, salt solubility measurement is usually carried out in relatively concentrated solutions, with ionic strengths, i, often exceeding 1 m.  at these high levels, activity coefficients of ions are poorly controlled and cannot be accurately predicted by the traditional debye-hückel equation.  also, ph electrodes calibrated in buffers with i = 0.15 m (“physiological” level) may not be accurate at much higher ionic strengths, especially in the extreme regions of ph (< 1 or > 12) where electrode junction potentials may be very different from those characteristic of the physiological level.  the salt solubility is a conditional constant, which depends not only on the concentration of the drug but also on that of the counterion with which the charged drug precipitates. the counterion may originate from the buffer used or other unsuspecting solution additives.  under such high sample concentrations, many drugs, especially those likely to be surface-active, can form micelles, self-associated aggregates (dimers, trimers, or higher-order oligomers), or complexes with buffer species or other solution additives. all these effects can complicate the interpretation of the solubility data [9-18]. such complexity may not be evident unless a full solubility – ph profile is measured, over a ph range containing the uncharged and charged forms of the drug, preferably at more than one solid-sample excess. carefully planned experimental designs and complicated computations are often needed to correctly interpret the measured salt solubility. although explicit solubility equations (cf., appendix a) have been derived for many different cases of salt solubility and aggregation [19], it is hardly practical to derive such equations for the vast number of possible forms of salt and aggregation stoichiometry that can be encountered. for this reason, salt solubility analysis of data in the past had been done on a case by case basis, sometimes using incomplete explicit solubility equations. at times the impact of aggregation reactions had been recognized but not dealt with quantitatively, presumably because computational methods were not available at the time [9,10,16]. in this study we describe a novel general approach to address many of the challenges of salt solubility determination, with data processing and refinement of equilibrium constants encoded in the computer program pdisol-x tm (in-adme research). the new approach was illustrated by the determinations of the solubility products of three model drugs: diprenorphine hydrochloride, codeine hydrochloride and phosphate, lidocaine hydrochloride and phosphate. diprenorphine, whose solubility-ph profile had not been reported, is primarily used as an opioid antagonist to reverse the effects of etorphine and carfentanyl. codeine is a naturally-occurring analgesic opiate. lidocaine is a local anesthetic. the effects of different buffers (britton-robinson universal and sörensen phosphate) were compared. lidocaine precipitates were characterized by x-ray powder diffraction and polarization light microscopy. the ionic strength in the studied systems ranged from 0.25 to 4.3 m. experimental chemicals codeine hydrochloride, codeine phosphate, lidocaine free base and lidocaine hydrochloride were purchased from sigma and used without further purification. diprenorphine hydrochloride was an in-house synthesized compound and was of analytical grade. gergely völgyi et al. admet & dmpk 1(4) (2013) 48-62 50 potassium hydroxide and hydrochloric acid used for pka determination were from sigma. the koh titrant solution was standardized by titration against primary standards potassium hydrogen phthalate dissolved in 15 ml of 0.15 m potassium chloride. potassium hydrogen phthalate and potassium chloride were of analytical grade and were purchased from sigma. the hcl titrant solution was standardized by titrating a measured volume against the standardized koh. two buffer solutions (britton–robinson universal and sörensen phosphate) were used at ph 5 in the shake-flask experiments of codeine and lidocaine. britton-robinson buffer solution (mixture of acetic, phosphoric and boric acids, each at 0.04 m) was treated with 0.2 m naoh to give the required ph. the sörensen phosphate buffer solutions were prepared by mixing 0.067 m na2hpo4 and 0.067 m kh2po4 solutions to reach the required ph (4.8 – 8.5). phosphate-containing solutions between ph 1 and 12.5 were used for the equilibrium solubility measurements of diprenorphine, where 0.5 m koh, 0.5 m or 14.85 m h3po4 (standardized 85%) solutions were used to reach the desired ph values. distilled water of ph. eur. grade was used. all other reagents were of analytical grade. ph electrode standardization and compensation for large changes in ionic strength all of the equilibrium constants reported here are based on the concentration scale, i.e., the “constant ionic medium” thermodynamic standard state [19]. since the measured ph is based on the operational activity scale, these values need to be converted to the concentration scale, pch (=-log[h + ]). the procedure to calibrate and standardize the ph electrode is described in detail elsewhere [19]. briefly, the electrode is standardized by the blank titration method: to a 0.15 m kcl solution, enough standardized 0.5 m hcl is added to lower the ph to 1.8 (0.87 ± 0.03 ml when using a 20 ml solution); then the acidified solution is precisely titrated with standardized 0.5 m koh up to about ph 12.2 (which consumes about 2 ml 0.5 m koh). the blank titration ph data are fit to a four-parameter equation [20]: ph = α + ks pch + jh [h + ] + joh kw/[h + ] (1) where kw is the ionization constant of water. the jh term corrects ph readings for the nonlinear ph response due to liquid junction and asymmetry potentials in highly acidic solutions (ph < 2), while the joh term corrects for high-ph nonlinear effects [19]. typical values of the adjustable parameters at 25c and 0.15 m ionic strength are α = 0.09, ks = 1.002, jh = 0.5 and joh = -0.5. however, each electrode indicates its own characteristic set. since salt solubility measurements can be under conditions where ionic strength may reach values as high as 1 m, the experimentally-determined parameters in eq. (1), based on the “reference” ionic strength 0.15 m, are automatically compensated by the computational procedure in pdisol-x, for changes in ionic strength from the reference value 0.15 m to the actual values in a particular solubility assay, according to the empirically-determined relationships defined elsewhere [19]. pka determination the pka values of diprenorphine were determined by potentiometry using a glpka instrument (sirius, forest row, uk) equipped with a combination ag/agcl ph electrode. the titrations were carried out at constant ionic strength (i = 0.15 m kcl) and temperature (t = 25.0 ± 0.5 °c), and under nitrogen atmosphere. aqueous solutions of diprenorphine hydrochloride (10 ml, 0.7-0.8 mm) were pre-acidified to admet & dmpk 1(4) (2013) 48-62 novel method of salt solubility analysis doi: 10.5599/admet.1.4.24 51 ph 2 with 0.5 m hcl, and then titrated with 0.5 m koh to ph 12. three parallel measurements were carried out. the pka values were calculated by the refinementpro tm software (sirius, forest row, uk). solubility measurements by saturation shake-flask method in the shake-flask experiments, the ph of the solutions was measured by a radiometer phm 220 ph meter with combined ag/agcl glass electrode. the temperature of the samples was maintained at 25.0 ± 0.5 °c during the solubility measurements using a lauda thermostat. a heidolph mr 1000 magnetic stirrer was used to mix the two phases. the concentration in the supernatant was measured by spectroscopy using a jasco v-550 uv/vis spectrophotometer. to facilitate the measure of concentration by uv spectrophotometry, the specific absorbance (a1cm 1% , the absorbance of 1 g/100 ml solution over a 1 cm optical path length at a given wavelength) of each sample at the given ph values was determined separately at a selected wavelength using 12-18 points of a minimum of two parallel dilution series, from the linear regression equation (lambert-beer law). the equilibrium solubility of the samples at different ph values was determined by the new protocol of saturation shake-flask method [21]. the sample was added to the aqueous britton-robinson universal and sörensen phosphate buffers (codeine and lidocaine) or phosphate-containing solution (diprenorphine) until a heterogeneous system (solid sample and liquid) was obtained. the solution containing solid excess of the sample was stirred for a period of 6 hours (saturation time) at controlled temperature allowing it to achieve thermodynamic equilibrium. after a further 18 hours of sedimentation, the concentration of the saturated solution was measured by uv spectroscopy. three aliquots were taken out with a fine pipette (5500 μl) from the liquid and diluted with the solvent if necessary. at least three parallel concentration measurements were carried out and the result was calculated using 9-12 data points. the standard deviation varied from 1-13%. refinement of intrinsic and salt solubility and aggregation constants the new data analysis method uses logs ph as measured input data (along with the standard deviations in logs) into the computer program. the data can be uv-derived, potentiometric-derived, etc. an algorithm was developed which considers the contributions of all species present in solution, including universal buffer components (e.g., britton-robinson, sörensen, prideaux-ward, etc.). the approach does not depend on any explicitly derived extensions of the henderson-hasselbalch equations. the computational algorithm derives its own implicit equations internally, given any practical number of equilibria and estimated constants, which are subsequently refined by weighted nonlinear leastsquares regression. so, in principal, drug-salt precipitates, -aggregates, -complexes, bile salts, -surfactant can be accommodated [19]. specific buffer-drug formed species can be tested, as is often necessary with phosphate-, and sometimes, citrate-containing buffers. the program assumes an initial condition of a suspension of the solid drug in a solution often containing a background electrolyte, e.g., 0.15 m nacl, ideally with the suspension saturated over a wide range of ph. the computer program calculates the distribution of species corresponding to a sequence of additions of standardized strong-acid titrant hcl (or weak-acid titrants h3po4, h2so4, acetic acid, maleic acid, lactic acid) to simulate the suspension speciation down to ph ~ 0, the staging point for the next operation. a sequence of perturbations with standardized naoh (or koh) is simulated, and solubility calculated at each point (in ph steps of 0.005-0.2), until ph ~ 13 is reached. the ionic strength is rigorously calculated at each step (cf, appendix b), and pka values (as well as solubility products, aggregation and complexation constants) are accordingly adjusted (cf., appendix gergely völgyi et al. admet & dmpk 1(4) (2013) 48-62 52 b). nonlinear ph electrode standardization parameters (eq. 1) are included in the calculation [20], a feature that is especially important for measurement of accurate ph in the extreme ph (<1 or >12) regions. at the end of the speciation simulation, the calculated logs vs. ph curve is compared to actual measured logs vs. ph. a logs-weighted nonlinear least squares refinement commences to refine the proposed equilibrium model, using analytical expressions for the differential equations. the process is repeated until the differences between calculated and measured logs values reach a minimum. specifically, the logs ph data are refined to minimize the weighted residual function,    n i i calc i obs i w s ss r )(log )log(log 2 2  (2) where n is the measured number of solubility values used to test the model, and logsi calc is the calculated log solubility values. the estimated standard deviation in the observed logs, i, is estimated as 0.10 (log units), or is set equal to the values reported in the measurement. the overall quality of the refinement is assessed by the "goodness-of-fit" (gof), p w nn r gof   (3) where np is the number of refined parameters. if a proposed model fits the data well, and accurate weighting factors are used, then gof = 1 is the statistically expected value. results and discussion table 1 summarizes some of the physicochemical properties of the three model compounds, including the pka constants of diprenorphine determined in this work. table 2 lists the shake-flask solubility measurement details, including the actual weights of compounds added to form the saturated solutions. in critical salt solubility studies, it is often necessary to state the actual weight of compound added, and not just state that “excess solid was added.” this is because salt solubility constants are conditional, and in some cases require the actual weights to determine solubility products, especially when aggregates form in saturated solutions or when salt stoichiometries are complex. in table 2, the calculated volumes of 14.85 m h3po4 or 0.5 m koh titrants used to adjust the ph are normalized to 1 ml total solution volumes (actual volumes ranged from 0.5 to 15 ml). table 3 lists the equilibrium constants determined and critical concentrations (including ionic strength and buffer capacity) calculated in the various assays. the pka values of the model compounds are listed in table 1 at standard state conditions (iref = 0.15 m, 25 o c), but the primed constants in table 3 are those that were actually applied at the specific ionic strength (0.25 – 4.34 m), calculated as described in appendix b. those of diprenorphine changed slightly by 0.0 to +0.05 (pka1) and 0.0 to -0.04 (pka2); that of codeine +0.02 to +0.14; that of lidocaine +0.15 to +0.51. according to the simple debye-hückel theory, those of monoprotic bases are expected not to change, but eq. b5 predicts otherwise. also listed in table 3 are the actual pksp values used at the particular ionic strengths, compared to those at iref. admet & dmpk 1(4) (2013) 48-62 novel method of salt solubility analysis doi: 10.5599/admet.1.4.24 53 table 1. physicochemical properties, 25 °c. compound mw log p pka literature -log s0 = ps0 [m] diprenorphine 425.56 3.81 a 9.68 ± 0.01 d , 8.52 ± 0.01 d - codeine 299.36 1.19 b 8.24 b 1.52 e lidocaine 234.34 2.44 c 7.95 c 2.04 f , 1.88 g , 1.75 h , 1.71 i a calculated using adme boxes 4.9 (acd labs, toronto). b [22] c [23] d this work. e [24] f [16]. g [25] h [26] i [27] the solubility products are also affected significantly when i >> iref. for example, the phosphate pksp of codeine phosphate decreased from 1.00 {1.11} at iref to -0.07 {0.02} at i = 1.21 {1.07} m, where the braced value refers to britton-robinson universal and the unbraced value refers to sörensen buffer measurements (table 3). lidocaine showed similar large shifts due to the large i = 4.34 {3.81} m. the changes were much smaller for the codeine chloride pksp, since i = 0.33 {0.24} m. table 2. shake-flask solubility data, 25 °c a ph buffer v (14.85 m h3po4) /ml b v (0.5 m koh) /ml b s (mg ml -1 ) sample wt. (g ml -1 ) expected precipitate 0.83 sör 0.4998 0.0000 4.98 ± 0.04 0.0307 xh2 + h2po4 (s) 1.94 sör 0.0111 0.0000 54.1 ± 1.1 0.0601 xh2 + cl (s) 3.53 sör 0.0003 0.0000 45.2 ± 0.9 0.0634 xh2 + cl (s) 5.06 sör 0.0000 0.0082 45.1 ± 1.2 0.0637 xh2 + cl (s) + xh(s) 7.22 sör 0.0000 0.4969 0.249 ± 0.014 0.0050 xh(s) 8.34 sör 0.0000 0.5702 0.088 ± 0.021 0.0114 xh(s) 8.83 sör 0.0000 0.5759 0.032 ± 0.001 0.0007 xh(s) 10.38 sör 0.0000 0.6251 0.667 ± 0.008 0.0106 xh(s) 11.52 sör 0.0000 0.8274 4.79 ± 0.11 0.0051 xh(s) 5.00 br 0.0000 0.1376 52.0 ± 0.1 0.1050 bh + cl (s) 5.00 sör 0.0000 0.0044 53.0 ± 0.1 0.1535 bh + cl (s) 5.00 br 0.0000 0.1716 290 ± 38 0.4950 bh + h2po4 (s) 5.00 sör 0.0000 0.0390 308 ± 30 0.5305 bh + h2po4 (s) 5.00 br 0.2149 0.0000 310 ± 15 0.7700 bh + h2po4 (s) 5.00 sör 0.2940 0.0000 294 ± 10 1.0300 bh + h2po4 (s) 5.00 br 0.0000 0.0064 >1000 1.0300 bh + cl (s) 5.00 sör 0.0000 0.0042 >1000 1.1354 bh + cl (s) a sör = sörensen buffer (0.15 m phosphate, adjusted with naoh). br = britton-robinson universal buffer (acetic, phosphoric, and boric acids, each at 0.04 m, adjusted with naoh to ph 5). b calculated by pdisol-x. the pka values of the acetate, phosphate, and borate buffer components used were taken from wiki-pka website (<http://www.in-adme.com/wiki_pka.php/>). these were automatically adjusted for changes in the ionic strength by pdisol-x. the changes were most pronounced with phosphoric acid (0.17 – 0.62), and were minimal with acetic and boric acids (0.01-0.06). http://www.in-adme.com/ gergely völgyi et al. admet & dmpk 1(4) (2013) 48-62 54 diprenorphine hydrochloride figure 1a shows the logs vs. ph profile of diprenorphine hydrochloride (xh2 + cl ). the dashed line is calculated with the henderson-hasselbalch equation (eq. a5), using the two pka of diprenorphine. the solid curve was calculated based on the equilibrium model most consistent with the actual logs measurements at the various ph. for ph > pka gibbs (5.07; cf., appendix a), the precipitate is the uncharged ordinary ampholyte, showing the characteristic parabolic shape. the intrinsic solubility (molar units) was refined as ps0 = 4.60 ± 0.08 (s0 = 11.5 ± 0.4 µg ml -1 ). at ph below the gibbs pka, either the chloride or the phosphate salt precipitates (or possibly both). it is not easy to be certain which form precipitates unless several measurements are made at ph < 2, in order to exploit the common ion effect which would be evident for phosphate salt, since phosphoric acid was used to lower the ph. the solubility at ph 0.83 suggests that a phosphate precipitate may form at the very low ph. in contrast, the chloride model predicts that all solid dissolves at ph 0.83 (given the amount of sample added). the flat shape of the curve ph 2–5 is most consistent with chloride precipitate. a phosphate precipitate would be expected to show an upward curvature near the gibbs pka. the constant based on the assumed chloride precipitate is pksp = 2.06 ± 0.07 (sxh = 44.6 ± 1.5 mg ml -1 , in hydrochloride equivalents). at ph > 9, the logs – ph curve shows a shift to lower ph, compared that what would be predicted from the henderson-hasselbalch equation. the consistent interpretation of the shift is that a water-soluble mixed-charge anionic dimer forms, with the stoichiometry xhx [19]. similar species have been observed in numerous studies, and in some cases, lc/ms was able to corroborate the hypothesized aggregate formations [18]. the equilibrium constant for the reaction xh + x = xhx is log kxhx = 5.21 ± 0.18 m -1 . admet & dmpk 1(4) (2013) 48-62 novel method of salt solubility analysis doi: 10.5599/admet.1.4.24 55 figure 1. solubility-ph profiles of the three model drugs studied. (see text) gergely völgyi et al. admet & dmpk 1(4) (2013) 48-62 56 codeine hydrochloride and dihydrogenphosphate figure 1b shows the codeine hydrochloride (bh + cl ) and phosphate (bh + h2po4 ) logs – ph profiles. the curve was calculated assuming the intrinsic solubility, ps0 = 1.52 (s0 = 12 mg ml -1 , assuming phosphate salt molecular weight), reported by kuhne et al. [24]. although single ph measurements were made, the whole curve adds useful perspective to the expected solubility – ph relationship. the assignment of the types of salts formed is possible in the way the assays were designed. the codeine phosphate assay had no source of chloride, and thus revealed the phosphate pksp = 1.00 {1.11}. the codeine chloride assay indicated a significantly higher pksp = 1.59 {1.57}, suggesting that the salt precipitate was that of the chloride. furthermore, the solubility product for the phosphate salt in the case of the chloride was not exceeded (table 3), suggesting the absence of phosphate salt in the chloride assay. lidocaine free base and hydrochloride figure 1c shows the lidocaine phosphate (bh + h2po4 ) logs – ph profile. the chloride was very soluble with sbh>1 g ml -1 . the chloride pksp in table 3 refer to the minimum possible values: pksp bhcl = -0.99 {-0.85}, suggesting that the chloride salt is at least 50 times more soluble than the phosphate salt. the solid curve was calculated assuming the intrinsic solubility, ps0 = 2.04 (s0 = 2.5 mg ml -1 , assuming bh + cl molecular weight), reported by bergström et al. [16]. as in the case of the codeine assays, the assignment of the types of salts formed with lidocaine is possible in the way the assays were designed. the lidocaine free base assay had no source of chloride, and thus revealed the phosphate pksp = 0.85 {0.88}. the solubility product for the phosphate salt in the case of the chloride was not exceeded (table 3), suggesting the absence of phosphate salt in the chloride assay. the unfilled circle symbols in figure 1c are the solubility values reported by bergström et al. [16], using the miniaturized shake-flask method (25 o c, 0.15 m phosphate buffer, 24 h incubation). even though the pksp values of phosphate precipitate in the free base case were nearly identical for the two buffer systems used, the morphology of the crystals isolated was quite different, as shown in figure 2, with the sörensen buffer producing much larger and better-formed crystals, compared to the brittonrobinson buffer (approx. 800 vs. 60 µm, respectively). the x-ray powder diffractograms in figure 3 confirmed that the crystals formed in the free base assay were those of the same phosphate salt polymorph. admet & dmpk 1(4) (2013) 48-62 novel method of salt solubility analysis doi: 10.5599/admet.1.4.24 57 figure 2. microphotographs of lidocaine phosphate solids obtained from solubility experiments: (a) britton-robinson buffer, (b) sörensen buffer. 5 10 15 20 25 30 35 2theta (°) 0 500 1000 1500 2000 in te n s it y ( c p s ) figure 3. x-ray powder diffractograms of reference lidocaine base (blue), lidocaine phosphate from solubility measurement in britton-robinson buffer (brown) and in sörensen buffer (orange). gergely völgyi et al. admet & dmpk 1(4) (2013) 48-62 58 conclusion the design of salt solubility assays here and the data analysis capability of the new program, pdisol-x offered an opportunity to critically investigate issues related to the challenges of characterizing salt solubility products, such as the high ionic strengths, which can affect values of equilibrium constants and the calibration of ph electrodes. the “anomalies” in the shapes of logs – ph profiles that cannot accurately be predicted by the henderson-hasselbalch may be common with sparingly-soluble or practically-insoluble drugs, such as diprenorphine, but are not always easy to recognize unless accurately-determined pka values are available [28]. phosphate buffers can dramatically influence the solubility profiles of ionizable drugs, as shown here and elsewhere [16,29]. these and other similar complications may be common, but are not always easy to interpret quantitatively. in such instances, pdisol-x may be a helpful data analysis and simulation tool. it can further aid in the analysis of dissolution mechanisms which depend on the salt solubility of drugs. appendices a) explicit solubility ph equations the idealized relationship between solubility and ph can be readily derived for simple equilibrium models. the "model" refers to a set of equilibrium equations and the associated equilibrium constants. the following are examples of such derivations correspond to a diprotic ampholyte. precipitation of diprotic amphoteric drug, xh (henderson-hasselbalch equation) in the case of a diprotic amphoteric drug, a saturated solution can be defined by the equations and the corresponding constants xh2 +  h + + xh ka1 = [h + ][xh] / [xh2 + ] (a1) xh  h + + x – ka2 = [h + ][x – ] / [xh] (a2) hx(s)  xh s0 = [xh] (a3) solubility, s, at a particular ph is defined as the mass balance sum of the concentrations of all of the species dissolved in the aqueous phase: s = [x – ] + [xh] + [xh2 + ] (a4) where the square brackets denote molar concentration of species. the above equation can be transformed into an expression containing only constants and [h + ] (as the only variable), by substituting the ionization and solubility eqs. a1-a3 into eq. a4. logs = log(ka1/([xh][h + ]) [xh] [h + ] /ka1 + [xh] + [xh][h + ] / ka1 ) = log[xh] + log(ka2 / [h + ] + 1 + [h + ] / ka1 ) = logs0 + log(1 + 10 – pka2+ph +10 +pka1–ph ) (a5) eq. a5 is often called the henderson-hasselbalch (hh) equation for a diprotic ampholyte, and describes a u-shaped logs – ph curve. at the low-ph bend in the logs – ph curve, the ph equals the pka1; at the other bend, ph equals the pka2. salt precipitation of diprotic amphoteric drug, xh admet & dmpk 1(4) (2013) 48-62 novel method of salt solubility analysis doi: 10.5599/admet.1.4.24 59 eq. a5 considers only one precipitate, that of the uncharged species xh, defining the “intrinsic solubility” concentration, s0. if the solubility experiment were carried out as an acid-base (hcl/naoh) titration, over a wide range ph <<pka1 to ph >> pka2, with enough excess compound added that the solubility products of the salts xh2 + cl – (s) and na + x – (s) are exceeded, then the curve would have regions at low ph and high ph that would not behave as indicated by the hh equation. to describe these processes, two additional solubility equations need to be added to eqs. a1-a3: xh2 + cl – (s)  xh2 + + cl – ksp1 = [xh2 + ][cl – ] (a6) na + x – (s)  na + + x – ksp2 = [na + ][x – ] (a7) hydrochloride salt of the drug, xh2 + cl – (s) if enough compound is added to the suspension and the ph is gradually lowered, well below pka1, at a critical ph point, called pka1 gibbs [30,31], the hh equation is replaced by an approximately horizontal line. in the salt solubility region, ph < pka1 gibbs , eq. a4 still holds, but [xh2 + ] becomes constant, instead of [xh]. eq. a5 can be modified (using eqs. a1, a2 and a6) to reflect this. logs = log(ka2 ka1 [xh2 + ] / [h + ] 2 + ka1 [xh2 + ] / [h + ] + [xh2 + ] ) = log[xh2 + ] + log(1 + 10 –pka1–pka2+2ph +10 – pka1+ph ) = logksp1 – log[cl – ] + log(1 + 10 –pka1–pka2+2ph +10 –pka1+ph ) (a8) for ph << pka1, the right-most logarithmic term in eq. a8 becomes vanishingly small, and cationic-drug conditional solubility is described by s(+) = [xh2 + ] = ksp1 /[cl – ]. eq. a8 largely describes a horizontal line. however, since the hcl titrant both dilutes the solution and elevates the cl – , the horizontal line curves downward for ph < 2, noted as the chloride “common-ion” effect, according to eq. a8. sodium salt of the drug, na + x – (s) in the presence of a large excess of compound in alkaline ph >> pka2, at the second critical ph point, pka2 gibbs , the hh equation is replaced by another approximately horizontal line. in the salt solubility region, ph > pka2 gibbs , eq. a4 still holds, but [x – ] becomes constant, while [xh] and [xh2 + ] are variable. eq. a5 can be further modified (using eqs. a1, a2 and a7) as logs = log( [x – ] + [x – ][h + ] / ka2 + [x – ] [h + ] 2 / (ka1ka2) ) = log[x ] + log(1 + 10 + pka1+pka2–2ph +10 – pka2+ph ) = logksp2 – log[na + ] + log(1 + 10 + pka1+pka2–2ph +10 –pka2+ph ) (a9) for ph >> pka2, the right-most logarithmic term in eq. a9 becomes vanishingly small, and conditional anionic-drug solubility is described by s(–) = [x – ] = ksp2 / [na + ]. since the naoh titrant both dilutes the solution and raises the sodium concentration in solution, the horizontal line curves downward for ph > 12, as a result of the sodium “common-ion” effect, according to eq. a9. the complete solubility equation across the whole ph range, given enough compound excess, the complete solubility equation consists of three segments: (a) conditional hydrochloride salt region, ph < pka1 gibbs (eq. a8), (b) drug concentrationindependent “hh” region, pka2 gibbs > ph > pka1 gibbs (eq. a5), and (c) conditional sodium salt region, ph > pka2 gibbs (eq. (a9)). gergely völgyi et al. admet & dmpk 1(4) (2013) 48-62 60 gibbs pka values at the lower ph = pka1 gibbs , two solids co-precipitate: xh2 + cl – (s) and xh(s). at the higher ph = pka2 gibbs , two different solids co-precipitate: na + x – (s) and xh(s). more discussion of the topic may be found in the literature [19,30,31]. cationic, anionic and neutral self-aggregate formations in concentrated drug solutions or in solutions containing practically insoluble drugs, oligomeric species, xn n– , (xh2 + )n n+ , (xh)n, (xh.x – )n n, (xh.xh2 + )n n+ , …, can form. these can characteristically alter the shape of the solubility – ph profile. consequently, most of eqs. a1-a9 would need to be modified to accommodate aggregation equilibria. the resultant solubility equations often become very complex, and not all combinations of possible species have had the corresponding equations derived. examples of such derivations of aggregation-solubility equations may be found in the literature [19]. b) automatic ionic strength compensation generally, the ionic strength, i, changes in the course of an acid-base titration due to ionizations, additions of titrant, and dilution of concentrations. this change affects acid-base equilibrium constants. in solubility experiments designed to determine salt solubility products, ksp, the ionic strength can vary substantially during a titration, and sometimes reaches values as high as 1 m, or even higher. in contrast, ionization constants, pka, are determined at a nearly constant i = 0.15 m, under conditions where low sample concentrations (e.g., 10 -3 to 10 -6 m) are “swamped” by the added inert salt (e.g., 0.15 m nacl or kcl). the independently-determined pka constants are critical to the analysis of solubility data, and thus the above large differences in ionic strength need to be factored in, as described here. it is a reasonable practice to designate 0.15 m as the “reference” ionic strength, iref, (“physiological” level) to which all equilibrium constants are scaled in the solubility assay. in the older literature, the reference state was often chosen as zero, but in current pharmaceutical applications, 0.15 m is usually chosen, with no loss of thermodynamic rigor [19]. since the ionic strength at any given ph point in a titration is likely different from iref, all ionization constants need to be locally transformed (from reference iref to local i) for the calculation of local point concentrations. the procedure below describes such an adjustment of activity coefficients. consider a three-reactant system, based on reactants x, y and h (proton), whose charges are qx, qy, and +1. the concentration of the jth particular species, cj, is defined in terms of these reactants cj = [x qx ] exj [y qy ] eyj [h + ] ehj βj (b1) associated with the general equilibrium expression exj x + eyj y + ehj h  xexjyeyjhehj (b2) where βj is the cumulative formation constant [19], and exj, eyj, and ehj are the x, y, and h stoichiometric coefficients, respectively, of the j th species. at each ph point in a solubility assay, i is calculated precisely, according to the general formula [32]: i = [nacl] + [naoh] + [h + ] + ½{qx(1+qx) x + qy(1+qx) y + σ qj(1+qj) cj } + ½ {│qx + nx │ qx – nx } x + ½ {│qy + ny │ qy – ny )}y (b3) admet & dmpk 1(4) (2013) 48-62 novel method of salt solubility analysis doi: 10.5599/admet.1.4.24 61 the last two terms in braces in eq. b3 take into account any nx or ny number of counter ions introduced to the solution by drug substances in salt form (per xor y-compound, respectively). the reference set of β(iref) formation constants are locally transformed to the set β(i) according to the general expression:                                     )( )( log )( )( log )( )( log )( )( log)(log)(log refj j refh h hj refy y yj refx x xjrefjj if if if if e if if e if if eii  (b4) where β(i) refers to the ionic strength i, while β(iref) refers to iref. the ionic-strength-dependent activity coefficients of x, y, h, and j th species (cf., eq. b2) are denoted fx, fy, fh, and fj, respectively. a similar equation was introduced by avdeef [32], based on the davies-modified debye-hückel equation, which is reasonably useful up to i = 0.3 m. since much higher ionic strengths are reached in salt solubility experiments, eq. b4 in the current study is cast in an expanded activity coefficient equation based on the hydration theory proposed by stokes and robinson [32,33], further elaborated by bates et al. [34] and robinson and bates [35] to include single-ion activities, then slightly modified by bockris and reddy [36], and recently applied to solubility data of a drug-like molecule by wang et al. [7]:                    j jj j j iii chc cc ah ib ia qif oh oh oh )1( loglog å1 )(log 2 2 2 i 2 (b5) the first term on the right side of eq. b5 is the debye-hückel equation accounting for the ion-ion electrostatic interactions; the second term is related to the decrease in the activity of water due to the work done in immobilizing some of the bulk water to hydrate ions; the third term is related to free energy change of the ions, as their concentrations effectively increase when the volume of bulk water decreases upon hydration of the ions. the parameters at 25c (molar scale): dielectric constant of water, ε = 78.3, 76.8, 67.3 at i = 0, 0.15, and 1 m (nacl), respectively [37]; the debye-hückel slope, a = 1.825x10 6 (εt) -3/2 = 0.512, 0.528, 0.642 at i = 0, 0.15, and 1 m, resp.; b = 50.29 (εt) -1/2 = 0.329, 0.333, 0.355 at i = 0, 0.15, 1 m, resp.; t is the absolute temperature (k). the two adjustable parameters are åi, corresponding to the mean diameter of the i th hydrated ion [38], and hi, the hydration number of the i th ion [37]. the activity of water, ah2o = 1.000, 0.995, 0.967 at i = 0, 0.15, 1 m, respectively [36]. ch2o = 55.51 m (concentration of water in pure solvent form, i = 0). the summation symbols in eq. b5 are over all charged species (including the reactants) in the system. acknowledgements we thank sándor hosztafi (semmelweis university, dept. of pharmaceutical chemistry) for the synthesis of diprenoprhine hydrochloride sample. we also thank dénes janke for the xrpd measurements. references [1] s. miyazaki, h. inouie, t. nadai, t. arita, m. nakano, chem. pharm. bull. 27 (1979) 1441-1447. [2] s. miyazaki, m. oshiba, t. nadai, j. pharm. sci. 70 (1981) 594-596. [3] b.d. anderson, r.a. conradi, j. pharm. sci. 74 (1985) 815-820. [4] b.d. anderson, k.p. flora, preparation of water-soluble compounds through salt formation. in: c.g. wermuth (ed.). the practice of medicinal chemistry. academic press, london, 1996, pp. 739-754. [5] m.t. ledwidge, o.i. corrigan, int. j. pharm. 174 (1998) 187-200. gergely völgyi et al. admet & dmpk 1(4) (2013) 48-62 62 [6] a.t.m. serajuddin, m. pudipeddi, salt selection strategies, in: stahl ph, wermuth cg (eds.), handbook of pharmaceutical salts: properties, selection, and use, wiley-vch, weinheim, 2002, pp. 135-160. [7] z. wang, l.s. burrell, w.j. lambert, j. pharm. sci. 91 (2002) 1445-1455. [8] p.h. stahl, salt selection, in: r. hilfiker (ed.), polymorphism in pharmaceutical industry, wiley-vch, weinheim, 2006, pp. 309-322. [9] t. higuchi, m. gupta, l.w. busse, j. amer. pharm. assoc. (sci. ed.) 42 (1953) 157-161. [10] t.j. roseman, s.h. yalkowsky, j. pharm. sci. 62 (1973) 1680-1685. [11] d.j. attwood, j. gibson, j. pharm. pharmacol. 30 (1978) 176-180. [12] a. fini, g. fazio, g. feroci, int. j. pharm. 126 (1995) 95-102. [13] z. wang, k.r. morris, b. chu, j. pharm. sci. 84 (1995) 609-613. [14] y. surakitbanharn, r. mccandless, j.f. krzyzaniak, r.-m. dannenfelser, s.h. yalkowsky, j. pharm. sci. 84 (1995) 720-723. [15] w.h. streng, d.h.-s. yu, c. zhu, int. j. pharm. 135 (1996) 43-52. [16] c.a.s. bergström, k. luthman, p. artursson, eur. j. pharm. sci. 22 (2004) 387-398. [17] a. avdeef, adv. drug deliv. rev. 59 (2007) 568-590. [18] e. shoghi, e. fuguet, e. bosch, c. ràfols, eur. j. pharm. sci. 48 (2013) 291-300. [19] a. avdeef, absorption and drug development, second edition. wiley-interscience, hoboken, nj, 2012. [20] a. avdeef, j.j. bucher, anal. chem. 50 (1978) 2137-2142. [21] e. baka, j. comer, k. takács-novák, j. pharm. biomed. anal. 46 (2008) 335-341. [22] a. avdeef, d.a. barrett, p.n. shaw, r.d. knaggs, s.s. davis, j. med. chem. 39 (1996) 4377-4381. [23] a. avdeef, k.j. box, j.e.a. comer, c. hibbert, k.y. tam, pharm. res. 15 (1997) 208-214. [24] r. kuhne, r-u. ebert, f. kleint, g. schmidt, g. schuurmann, chemosphere 30 (1995) 2061-2077. [25] e. shoghi, e. fuguet, c. ràfols, e. bosch, chem. biodivers. 6 (2009) 1789-1795. [26] c.m. wassvik, a.g. holmén, r. draheim, p. artursson. logp-independent solubility of drugs molecular descriptors for solid-state limited solubility. in: c.m. wassvik. phd thesis, univ. uppsala, 2006. [27] s. pinsuwan, p.b. myrdal, y.-c. lee, s.h. yalkowsky, chemosphere 35 (1997) 2503-2513. [28] g. völgyi, e. baka, k. box, j. comer, k. takács-novák, anal. chim. acta 673 (2010) 40-46. [29] w.h. streng, s.k. hsi, p.e. helms, h.g.h. tan, j. pharm. sci. 73 (1984) 1679-1684. [30] a. avdeef, pharm. pharmacol. commun. 4 (1998) 165-178. [31] w.h. streng, int. j. pharm. 186 (1999) 137–140. [32] a. avdeef, j. pharm. sci. 82 (1993) 183-190. [33] r.h. stokes, r.a. robinson, j. amer. chem. soc. 70 (1948) 1870-1878. [34] r.g. bates, b.r. staples, r.a. robinson, anal. chem. 42 (1970) 867-871. [35] r.a. robinson, r.g. bates, marine chem. 6 (1978) 327-333. [36] j.o’m. bockris, a.k.n. reddy, modern electrochemistry, vol. 1. plenum publishing corp., new york, ny, 1973. [37] r.a. robinson, r.h. stokes, electrolytic solutions. 2 nd revised ed. dover publications, inc., mineola, ny, 2002. [38] j. kielland, j. amer. chem. soc. 59 (1937)1675-1678. ©2013 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ manuscript doi: 10.5599/admet.3.4.247 345 admet & dmpk 3(4) (2015) 345-351; doi: 10.5599/admet.3.4.247 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper permeation of indomethacin through skin using nanonized alaptide aneta cernikova, jiri dohnal and josef jampilek* department of chemical drugs, faculty of pharmacy, university of veterinary and pharmaceutical sciences brno, palackeho 1/3, 61242 brno, czech republic *corresponding author: e-mail: josef.jampilek@gmail.com; tel.: +420-541-562-926 received: october 28, 2015; revised: november 22, 2015; published: december 30, 2015 abstract alaptide, (s)-8-methyl-6,9-diazaspiro[4.5]decan-7,10-dione, is an original czech compound; in this paper it is used as an excipient. the investigation deals with the affection of the permeation of indomethacin through full-thickness pig ear skin using a franz diffusion cell from the donor vehicle of propylene glycol/water (1:1) using nanonized alaptide as a potential transdermal permeation enhancer. alaptide was applied in ratio 1:10 (w/w) related to the amount of indomethacin. nanonized alaptide showed an excellent rapid onset of enhancement effect, already at the 30 th minute after application, when the permeated amount of indomethacin was 5-fold more than in the formulation without alaptide. the enhancement ratio of nanonized alaptide was 5.6, which indicates that alaptide modifies skin structure, which results in significantly enhanced permeation at long-term application. keywords alaptide; indomethacin; nanoparticles; permeation; skin. introduction transdermal administration of drugs can be considered as an alternative to conventional pharmaceutical dosage forms. nevertheless, a frequent problem of transdermal drug delivery is insufficient or no penetration of active pharmaceutical substances through the skin, and thus various approaches to overcoming the skin barrier was to be developed. these approaches can be based on an optimization of a drug/vehicle system or modification of stratum corneum (which is the outermost layer of the skin responsible for barrier function and formed by corneocytes and an intercellular lipid matrix). modification (i.e. hydration, lipid fluidization and/or disruption) of stratum corneum is possible through: (i) application of transdermal chemical permeation enhancers (cpes); (ii) overall optimization of formulation using nonhydrophobic excipients; (iii) application of physical enhancement techniques (electrically assisted methods), such as iontophoresis, electroporation, acoustic methods, microneedles, etc.) [1−3 and refs. therein]. cpes can be considered as excipients specifically affecting intercellular space between corneocytes or modifying corneocytes by hydration or denaturation of keratin. the heterogeneity of molecular structures of cpes limits simple explanation of their action. several possible mechanisms of action were hypothesized, but exact mechanisms have not been elucidated. it is almost certain that cpes exhibit multiple effects: http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:josef.jampilek@gmail.com cernikova et al. admet & dmpk 3(4) (2015) 345-351 346 (i) interact with the intercellular lipid matrix; (ii) interact with proteins (influencing the conformation of keratin in corneocytes or proteins in desmosomes); (iii) promote partitioning (influencing the stratum corneum nature leads to raising the penetrant concentration gradient and thus increasing the flux, i.e. increasing the concentration of the drug in the skin) [1−3 and refs. therein]. alaptide, (s)-8-methyl-6,9-diazaspiro[4.5]decan-7,10-dione, (figure 1) is an original czech compound prepared by kasafirek et al. in the 80s of the 20 th century [4]. a class of similar compounds was designed as analogues of melanocyte-stimulating hormone release-inhibiting factor (mif-1) [5], i.e. they negatively affect the inhibition of the release of melanocyte-stimulating hormone, and thus increase the concentration of melanocytes in epidermis. in relation to the effect of alaptide as a potential cpe it is important to note that melanocytes significantly influence the creation and function of keratinocytes by means of melanosomes [6−8]. during biological assays it was found that alaptide showed significant curative effect in different therapeutic areas [5], for example, it reduced the number and the extent of gastric ulcers [9,10], increased cell proliferation and epidermal regeneration and significantly accelerated regeneration (curing) of skin injuries [5,11]. although bulk alaptide is absorbed from the gastrointestinal tract or permeates through the skin [5,12], nanonized alaptide (nala) permeates through the skin insignificantly [12]. alaptide is not metabolized and is excreted mostly via urine [13]. alaptide demonstrated very low acute toxicity; no subchronic and chronic toxicity, genotoxic, teratogenic and embryotoxic effects were observed [5,14,15]. alaptide enantiomers do not induce cytochrome p450 (1a1, 1a2, 1b1) [16]. indomethacin, [1-(4-chlorobenzoyl)-5-methoxy-2-methyl-1h-indol-3-yl]acetic acid, (figure 1) is a nonsteroidal anti-inflammatory drug (nsaid) with anti-inflammatory, analgesic and antipyretic activity. its pharmacological effect is mediated through inhibition of cyclooxygenase (cox), the enzyme responsible for catalysing the rate-limiting step in prostaglandin synthesis via the arachidonic acid pathway. indomethacin inhibits both isoforms cox-1 and cox-2, with greater selectivity for cox-1, which accounts for its increased adverse gastric effects relative to other nsaids. cox-1 is required for maintaining the protective gastric mucosal layer. the analgesic, antipyretic and anti-inflammatory effects of indomethacin occur as a result of decreased prostaglandin synthesis. its antipyretic effects may be due to the action on the hypothalamus, resulting in an increased peripheral blood flow, vasodilation and subsequent heat dissipation. indomethacin is commonly commercially available in oral, rectal and topical formulations [17]. this follow-up paper to the previous contributions [18−22] is aimed at the investigation of the effect of nanonized alaptide as a potential cpe on the permeation of indomethacin substance through the skin. figure 1. structures of (s)-alaptide and indomethacin. experimental preparation of nanonized alaptide alaptide was synthesized by the standard process [23]. all reagents were purchased from aldrich and merck. the suspension of alaptide (30 g), polyvinylpyrrolidone (30 g) and purified water (240 ml, during admet & dmpk 3(4) (2015) 345-351 skin permeation of indomethacin doi: 10.5599/admet.3.4.247 347 milling was diluted by addition of additional 150 ml) was initially mixed for 12 h at ambient temperature and then filtered through a mill sieve. the milling procedure was performed using a nanomill netzsch (germany) with glass beads (0.3 mm); the rotor speed was 986 rpm; the pump speed was 30 rpm; the temperature in the grinding chamber was within 17–20 °c. the rotor speed was increased to 1500 rpm after 6 h of milling. the total time of milling was 57.5 h. the content of alaptide in the suspension was 38.76 g/l (determined by rp-hplc [24]). the particle size x90 of the prepared nanonized alaptide measured by sympatec nanophox 0138 p (germany) was 770 nm. in vitro transdermal permeation experiments skin samples were obtained from porcine ear. full-thickness pig ear skin was cut in fragments and stored at −20°c until utilized. skin samples were slowly thawed (at 4 °c overnight and then at ambient temperature) before each experiment [25,26]. the penetration enhancing effect of nala was evaluated in vitro, using a vertical franz diffusion cell (ses – analytical systems, germany) with a donor surface area of 0.6359 cm 2 and a receptor volume of 5.2 ml. the skin was mounted between the donor and receptor compartments of the franz diffusion cell with the epidermal side up. the receptor compartment was filled with phosphate buffered saline (ph 7.4) and maintained at 34±0.5 °c [25,27], using a circulating water bath. the receptor compartment content was continuously stirred using a magnetic stirring bar. the skin was kept in contact with the receptor phase for 0.5 h prior to the experiment. indomethacin was purchased from sigma; all other reagents were purchased from merck. the donor samples were prepared by dissolving indomethacin (10 mg) and nala (1 mg of alaptide, i.e. 26 µl of suspension) in propylene glycol (0.5 ml). 0.5 ml of water was added to the mixture. this mixture was shaken vigorously and then sonicated for 10 minutes at 40 °c, then this stable system (dissolved indomethacin in enhancer emulsion) was applied to the skin surface and the donor compartment of the cell was covered by parafilm®. the control samples were prepared in the same manner without nala. samples (0.5 ml) of the receptor phase were withdrawn at pre-determined time intervals (30, 60, 90, 120, 240, 360, 480, 720 and 1440 min), and the cell was refilled with an equivalent amount of fresh buffer solution. a minimum of five determinations was performed using skin fragments from a minimum of 2 animals for each composition, and the data was expressed as means ± sd. the samples were immediately analysed by hplc. analysis of samples was performed using an agilent 1200 series hplc system, equipped with a diode array detection (dad) system, a quaternary model pump and an automatic injector (agilent technologies, germany). data acquisition was performed using chemstation chromatography software. a gemini c6-phenyl 110a 5 μm, 250×4.6 mm (phenomenex®, usa) chromatographic column was used. the total flow of the column was 1.0 ml/min; injection was 10 μl; column temperature was 40 °c; and sample temperature was 20 °c. the detection wavelength of 260 nm was chosen, the time of analysis was 5 min. a mixture of meoh (hplc grade, 49.0%) with formic acid p.a. (0.2 %), acetonitrile (hplc grade, 50.0 %) and h2o (hplc – mili-q grade, 1.0%) was used as a mobile phase. the retention time (tr) of indomethacin was 3.1±0.05 min; the limit of detection (lod) was 0.059 µg/ml; and the limit of quantification (loq) was 0.197 µg/ml. as a result of the sampling, the receptor compartment concentration of alaptide was corrected for sample removal and replenishment using equation: c'n = cn (vt/vt-vs) (c'n-1/cn-1); where c'n = corrected drug concentration in the n th sample, cn = measured drug concentration in the n th sample, c'n-1 = corrected drug concentration in the (n-1) th sample, cn-1 = measured drug concentration in the (n-1) th sample, vt = total volume of receptor solution, vs = volume of the sample, and c'1 = c1 [28]. the corrected data were expressed as the cumulative drug permeation (qt) per unit of skin surface area using equation: qt = c'n/a; cernikova et al. admet & dmpk 3(4) (2015) 345-351 348 where a = 0.6359 cm 2 in our experiment. from the slope of the linear portion of the curve [29] of the dependence the cumulative amount of drug per unit area (qt [μg]) on time (t [h]), steady-state permeation flux (j [μg/h/cm 2 ]) was determined. similarly, the lag time (tlag [h]) was determined by extrapolating the linear portion of cumulative amount of permeation per unit area (qt) versus time (t [h]) curve to the abscissa [30]. the permeability coefficient (kp [cm/h]) can be calculated according to kp = j/cd; where cd = drug concentration in the donor compartment. it is assumed that under sink conditions, drug concentration in the receptor compartment is negligible compared to that in the donor compartment [31,32]. the enhancement effect was expressed as an enhancement ratio (er) that was calculated by the formula er = jss-x/jss-k; where jss-x = steady-state permeation flux with cpe, jss-k = steady-state permeation flux without cpe [33]. all the calculated data are listed in tables 1 and 2 and illustrated in figure 2. results and discussion the high toxicity and the frequent occurrence of undesirable/side effects of indomethacin, on the one hand, and significant curative action, on the other hand, brought us to an idea to formulate this substance with alaptide for its skin permeation enhancement and thus prepare less irritable therapeutic system. in vitro skin permeation experiments were performed using static franz diffusion cells [27] within 24 hours. full-thickness pig ear skin was selected for in vitro evaluation of permeation. this tissue is a suitable in vitro model of human skin [34,35], because porcine skin has shown to be histologically and biochemically similar to human skin [36]. the permeation of indomethacin through the skin without and with nala was tested from the donor vehicle of propylene glycol/water (1:1). previous studies have indicated that propylene glycol by itself (or a propylene glycol/water system) does not interfere with membranes [37,38]. table 1. cumulative permeated amounts qt per unit area [μg/cm 2 ] of indomethacin (idm) from propylene glycol:water (1:1) without and with nanonized alaptide (nala) as potential cpe achieved in in vitro transdermal permeation experiments using franz diffusion cell. qt values are expressed as mean ± sd (n = 5 experiments). time [h] cumulative permeated amounts qt per unit area[µg/cm 2 ] idm idm+nala 0.5 0.9±0.2 4.6±0.6 1.0 1.7±0.3 6.2±0.5 1.5 2.8±1.0 7.5±0.3 2.0 5.1±2.0 9.3±0.8 3.0 7.6±2.2 13.1±2.5 4.0 10.5±2.5 17.0±2.9 6.0 15.1±3.3 32.8±3.4 8.0 18.9±4.6 65.2±3.6 12.0 32.0±6.2 128.9±6.3 24.0 83.2±6.5 384.1±5.2 the values obtained from the permeation experiments were expressed as the cumulative permeated amount of the drug (qt [μg]) per unit of skin surface area, see table 1. other permeation parameters of indomethacin without and with nala from propylene glycol:water (1:1), steady-state permeation fluxes admet & dmpk 3(4) (2015) 345-351 skin permeation of indomethacin doi: 10.5599/admet.3.4.247 349 (j [μg/h/cm 2 ]), lag times (tlag [h]), permeability coefficients (kp [cm/h]) and enhancement ratio (er) are mentioned in table 2. the dependences of the cumulative permeated amount of the drug per unit of skin surface area in time are illustrated in figure 2 that is divided into parts a and b for better lucidity. table 2. permeation parameters of indomethacin (idm) without and with nanonized alaptide (nala) from propylene glycol:water (1:1): steady-state permeation fluxes (j), corresponding lag times (tlag), permeability coefficients (kp) and enhancement ratio (er). all values are expressed as mean ± sd (n = 5 experiments). sample j [μg/cm 2 /h] tlag [h] kp×10 -3 [cm/h] er idm 2.9±0.4 1.0±0.6 0.3±0.05 – idm+nala 16.1±1.2 3.9±0.2 1.6±0.1 5.6±0.4 figure 2. in vitro profile of cumulative permeated amounts qt per unit area [μg/cm 2 ] in time of indomethacin (idm) alone and after addition of nanonized alaptide (nala) in ratio 10:1 (idm:nala) from propylene glycol/water (1:1) system through skin. qt values are expressed as mean ± sd (n = 5 experiments). the permeated amount of indomethacin with nala increased rapidly already at the 30 th minute, when it reached approx. 5-fold higher values than formulation without nala, see table 1. similarly sharp enhancement of transdermal permeation was found, e.g., for permeation of other nsaids or antipyretics, such as ibuprofen, nimesulide, acetylsalicylic acid or paracetamol [18,19,21]. it can be stated that indomethacin without nala permeated moderately in comparison with indomethacin with added nala; in the whole investigated time range in every time the corresponding qt values related to the system without and with nala were statistically different from each other, see figures 2a and 2b. the effectivity of alaptide as a potential cpe can also be confirmed by parameters mentioned in table 2, since significant increases of steady-state permeation flux and permeability coefficient (approx. about 5.3) were observed. enhancement ratio calculated from steady-state permeation fluxes is 5.6, which indicates that nanonized alaptide enhanced permeation of compounds through the skin. on the other hand, an absolutely different effect was observed for the combination of bactroban® leciva with nala when a significant decrease of permeation (in fact, blockade of permeation) of mupirocin from ointment through the skin occurred during the experiment, indicating that mupirocin can act curatively only on the surface of the skin [20]. based on the presented results, it can be assumed that the contribution of nala to the enhanced permeation of indomethacin through the skin is significant not only immediately after application but also for long-term application. the structure of alaptide can be classified as a hybrid between the derivatives of 0 5 10 15 20 25 30 35 40 0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 q t [ g /c m 2 ] time [h] id m id m+nala a 0 50 100 150 200 250 300 350 400 5.0 10.0 15.0 20.0 25.0 q t [ g /c m 2 ] time [h] id m id m+nala b cernikova et al. admet & dmpk 3(4) (2015) 345-351 350 urea and 2-pyrrolidone, therefore the supposed mechanism of enhancement action can be as follows. as an urea-like derivative it can demonstrate moisturizing effect on the stratum corneum [38−40], and, as a 2-pyrrolidone-like derivative it can exhibit interactions preferentially in the keratin region [39,41]. nevertheless, the exact mechanism of action of alaptide and effects of mutual interactions of indomethacin, alaptide and skin are under investigation, for example, using raman or infrared spectroscopy for evaluation of interactions of the formulation and alaptide with the skin structure and using thermal methods for understanding of interactions between indomethacin and alaptide. conclusions modification of the permeation of indomethacin through the full-thickness pig ear skin by nanonized alaptide was investigated by in vitro screening using the franz diffusion cell. nanonized alaptide applied in ratio 1:10 related to the amount of indomethacin significantly enhanced the permeation of the drug substance from the donor vehicle of propylene glycol/water (1:1) through the skin. the calculated parameters, such as steady-state permeation flux and permeability coefficient, confirmed the enhancement effect of nanonized alaptide. overall enhancement ratio er is 5.6. the permeated amount of indomethacin with nala increased rapidly already at the 30 th minute, when it reached more than 5-fold higher values than indomethacin without nala. thus, it can be summarized that nanonized alaptide can be successfully applied as a cpe, because it provided excellent rapid onset of enhancement effect, which is an important requirement for chemical permeation enhancers. it is evident that alaptide modifies skin structure, which results in significantly enhanced permeation at long-term application. acknowledgements: this study was supported by iga vfu brno 302/2015/faf and by the technology agency of the czech republic ta04010065. references [1] j. jampilek, k. brychtova. medicinal research reviews 32 (2012) 907–947. [2] j. jampilek. journal of bioequivalence & bioavailability 5 (2013) 233–235. [3] d. kaushik, p. batheja, b. kilfoyle, v. rai, b. michniak-kohn. expert opinion on drug delivery 5 (2008) 517–529. [4] e. kasafirek, j. vanzura, i. krejci, j. krepelka, a. dlabac, m. valchar. (united pharmaceutical works & research institute for pharmacy and biochemistry), belg. 897843 (1984) & cs 231227 (1986). [5] s. radl, e. kasafirek, i. krejci. drugs of the future 15 (1990) 445–447. [6] j.a. mcgrath, r.a. eady, f.m. pope. rook's textbook of dermatology, 7th ed. t. burns, s. breathnach, n. cox, c. griffiths, (eds)., blackwell publishing, oxford, 2004, pp. 3–7. [7] w. james, t. berger, d. elston. andrews' diseases of the skin: clinical dermatology, 10th ed., saunders-elsevier, philadelphia, 2006, pp. 5–6. [8] f.m. watt. bioessays 8 (1988) 163–167. [9] l. korbova, j. cizkova, j. kohout, e. kasafirek, i. krejci, j. vanzura. journal of czech physicians 127 (1988) 1574-1577. [10] l. korbova, j. kohout, e. kasafirek. gastroenterology and hepatology 48 (1994) 170-182. [11] e. kasafirek, l. korbova, j. kohout, m. jiraskova, i. krejci, a. galatik. (united pharmaceutical works & research institute for pharmacy and biochemistry), cs 276270 (1992). [12] r. opatrilova, a. cernikova, l. coufalova, j. dohnal, j. jampilek. the scientific world journal 2013 (2013), article id 787283 (8 pages), http://www.hindawi.com/journals/tswj/2013/787283/. [13] r. lapka. journal of pharmacy and pharmacology 43 (1991) 874-876. [14] j. vanzura, k. kosar, e. kasafirek. toxicology letters 31 (1986) 189-193. [15] k. kosar, j. vanzura. pharmazie 43 (1988) 715-716. http://www.hindawi.com/journals/tswj/2013/787283/ admet & dmpk 3(4) (2015) 345-351 skin permeation of indomethacin doi: 10.5599/admet.3.4.247 351 [16] j. jampilek, r. opatrilova, a. rezacova, z oktabec, j. dohnal. (faculty of pharmacy, university of veterinary and pharmaceutical sciences brno), wo/2014/019556 a1 (2014). [17] drug bank – indomethacin, http://www.drugbank.ca/drugs/db00328 (october 28, 2015). [18] j. jampilek, r. opatrilova, l. coufalova, a. cernikova, j. dohnal. (faculty of pharmacy, university of veterinary and pharmaceutical sciences brno), wo/2013/020527 a1 (2013). [19] j. jampilek, r. opatrilova, l. dvorakova, k. brychtova, j. dohnal. (faculty of pharmacy, university of veterinary and pharmaceutical sciences brno), cz 304915 b6 (2014). [20] r. opatrilova, j. jampilek. admet & dmpk 2 (2014), 56–62. [21] a. cernikova, r. opatrilova, j. jampilek. military medical science letters 83 (2014) 34–39, http://mmsl.cz/vicms/soubory/pdf/mmsl_2014_1_6__www.pdf. [22] a. cernikova, r. opatrilova, p. bobal, j. jampilek. admet & dmpk 2 (2014) 248-253. [23] e. kasafirek, m. rybak, i. krejci, a. sturs, e. krepela, a. sedo. life science 50 (1992) 187–193. [24] m. dousa, k. lemr. journal of separation science 34 (2011), 1402–1406. [25] oecd guidelines for the testing of chemicals, section 4, test no. 428: skin absorption: in vitro method, oecd publishing, paris, 2004, http://dx.doi.org/10.1787/9789264071087-en. [26] who. environmental health criteria (ehc 235) – dermal absorption. who press, geneva, switzerland, 2006, http://www.who.int/ipcs/features/2006/ehc235/en/. [27] t.j. franz. journal of investigative dermatology 64 (1975) 190–195. [28] h. wu, c. ramachandran, n.d. weiner, b.j. roessler. international journal of pharmaceutics 220 (2001) 63–75. [29] n. akhtar, m.u. rehman, h.m.s. khan, f. rasool, t. saeed, g. murtaza. tropical journal of pharmaceutical research 10 (2011) 281–288. [30] l. panigrahi, s. pattnaik, s.k. ghosal. aaps pharmscitech 6 (2005), article 25, e167–e173. [31] c.t. huang, m.j. tsai, y.h. lin, y.s. fu, y.b. huang, y.h. tsai, p.c. wu. international journal of nanomedicine 2013 (2013) 2295–2304. [32] y.s. rhee, j.g. choi, e.s. park, s.c. chi. international journal of pharmaceutics 228 (2001) 161–170. [33] s.a. ibrahim, s.k. li. journal of controlled release 136 (2009) 117–124. [34] u. jacobi, m. kaiser, r. toll, s. mangelsdorf, h. audring, n. otberg, w. sterry, j. lademann. skin research and technology 13 (2007) 19–24. [35] c. herkenne, a. naik, y.n. kalia, j. hadgraft, r.h. guy. pharmaceutical research 23 (2006) 1850-1856. [36] w. meyer, k. schwarz, k.t. neurand. current problems in dermatology 7 (1978) 39–52. [37] m.a. yamane, a.c. williams, b.w. barry. journal of pharmacy and pharmacology 47 (1995) 978–989. [38] a.c. williams, b.w. barry. international journal of pharmaceutics 56 (1989) 43–50. [39] a.c. williams, b.w. barry. advanced drug delivery reviews 56 (2004) 603–618. [40] h. trommer, r.h.h. neubert. skin pharmacology and physiology 19 (2006) 106–121. [41] b.w. barry. journal of control release 6 (1987) 85–97. ©2015 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://www.drugbank.ca/drugs/db00328 http://mmsl.cz/vicms/soubory/pdf/mmsl_2014_1_6__www.pdf http://dx.doi.org/10.1787/9789264071087-en http://www.who.int/ipcs/features/2006/ehc235/en/ http://creativecommons.org/licenses/by/3.0/ manuscript doi: 10.5599/admet.2.1.33 18 admet & dmpk 2(1) (2014) 18-32; doi: 10.5599/admet.2.1.33 open access: issn: 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper the intrinsic aqueous solubility of indomethacin john comer 1 *, sam judge 1 , darren matthews 1 , louise towes 1 , bruno falcone 2 , jonathan goodman 2 and john dearden 3 1 sirius analytical ltd., forest row, west sussex rh18 5dw, uk 2 unilever centre for molecular informatics, department of chemistry, university of cambridge, lensfield road, cambridge cb2 1ew, uk 3 school of pharmacy & biomolecular sciences, liverpool john moores university, byrom street, liverpool l3 3af, uk *corresponding author: e-mail: john.comer@sirius-analytical.com; tel.: +44 1342 820720 received: february 19, 2014; revised: april 01, 2014; published: april 01, 2014 abstract a value of 8.8 μg/ml was measured for the intrinsic solubility of indomethacin. evidence of a form with a solubility of about 77 μg/ml was also obtained. solubility measurements were conducted using the cheqsol and curve fitting methods using a maximum ph of 9. it is also demonstrated that a published intrinsic solubility of 410 μg/ml was in error due to decomposition of indomethacin at ph 12. the decomposition of indomethacin at ph 12 was investigated. decomposition products comprising pchlorobenzoic acid and 5-methoxy-2-methyl-3-indoleacetic acid were isolated and characterised. keywords: indomethacin, solubility, cheqsol, p-chlorobenzoic acid, decomposition introduction indomethacin is a widely-used non-steroidal anti-inflammatory drug (nsaid), despite its propensity to cause gastric irritation and ulceration. its structure is shown in figure 1. indomethacin can exist in several polymorphic solid forms and as an amorphous solid. yamamoto [1] reported in 1968 that he had isolated three polymorphs, and with slightly different melting points. borka [2] and lin [3] claimed to have found at least four polymorphic modifications. other authors recognise only the  and  polymorphs [4-6]. the polymorphism is believed to arise from different orientations between the aromatic indole and phenyl rings [7]. solvates are also known to exist [2,8]. figure 1. decomposition of indomethacin into 5-methoxy-2-methyl-3-indoleacetic acid (1) and p-chlorobenzoic acid at ph 12. n o cl o oh meo ph 12 h n o oh meo + o cl ho indomethacin 1 p-chlorobenzoic acid http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:john.comer@sirius-analytical.com admet & dmpk 2(1) (2014) 18-32 indomethacin solubility doi: 10.5599/admet.2.1.33 19 one difficulty associated with the measurement of the solubility of indomethacin is that the measured solubility can change with over time [2,10,11], suggesting that there is conversion from one form to another. for example, murdande et al. [10] found that amorphous indomethacin in aqueous solution changed almost completely to a mixture of the and  polymorphs over a 40-minute period, with solubility decreasing from about 26 μg/ml to about 9 μg/ml. numerous measurements of the aqueous solubility of indomethacin have been reported and results are presented in table 1. predicted indomethacin solubilities from a number of commercial and free-to-use software programs are also summarised in table 1. it can be seen from table 1 that the published solubilities for the  and  polymorphs are quite consistent. it is, however, difficult to say with assurance whether the so-called polymorph i is the or  polymorph, and polymorphs ii and iv are not obviously either or . table 1. reported measured values for the solubility of indomethacin, and a selection of values calculated by commercial software. rt = room temperature; n/a = not available. because amorphous solids do not have crystal lattices in which molecules can be held strongly, they are more soluble and are lower-melting than are their crystalline forms. murdande et al. [10] reported that the aqueous solubility ratio (amorphous/-polymorph) for indomethacin was 4.9. borka [2] found the amorphous form of indomethacin to have a melting point of 55-57 °c, in contrast to values ranging from 134 to 160 °c for various polymorphic forms. there is considerable variation among the unspecified form solubilities reported, which could indicate that different proportions of polymorphs and amorphous form have contributed to the reported solubilities. one reported solubility value, (410 μg/ml), is significantly higher than all the others and has been the source of some controversy [23]. this result was one of 132 measured by the sirius cheqsol method and cited in the cambridge solubility challenge. the cambridge solubility challenge invited readers to use 100 measured solubility values to train software for predicting solubility from the structure alone [24], and then form μg/ml temp. °c ref. form μg/ml temp. °c ref. form μg/ml ref. pol ymorph i 4.2 25 2 uns peci fi ed 9.5 rt 15 uns peci fi ed 17.3 39 pol ymorph ii 15.6 25 2 uns peci fi ed 0.94 25 16 "nati ve" 39.1 39 pol ymorph iv 20 25 2 uns peci fi ed 18.5 25 17 uns peci fi ed 0.94 40 α pol ymorph 8.7 35 4 uns peci fi ed 25.3 35 17 uns peci fi ed 4.3 41 γ pol ymorph 6.9 35 4 uns peci fi ed 3.9 25 18 uns peci fi ed 2.4 42 γ pol ymorph 5 25 9 uns peci fi ed 2.3 25 19 uns peci fi ed 3.1 43 amorphous 22.5 25 9 uns peci fi ed 40 25 20 uns peci fi ed 0.81 44 γ pol ymorph 5 25 10 uns peci fi ed 15 25 21 amorphous 24.5 25 10 uns peci fi ed 27.1 rt 22 α pol ymorph 9.4 35 12 uns peci fi ed 410 25 25 γ pol ymorph 6.9 35 12 uns peci fi ed 0.94 25 31 pol ymorph i 9.1 25 13 uns peci fi ed 1.16 25 35 pol ymorph ii 14.4 25 13 uns peci fi ed 3.09 25 36 α pol ymorph 4 n/a 14 uns peci fi ed 0.4 25 37 γ pol ymorph 6 n/a 14 uns peci fi ed <1 rt 38 amorphous 10 n/a 14 meas ured val ues cal cul ated val ues comer et al. admet & dmpk 1(2) (2014) 18-32 20 to use the trained software to predict the solubility of a further 32 molecules whose solubilities had been withheld from the publication. one of these 32 molecules was indomethacin. a total of 99 entries were received from software providers, and the results were reviewed in a subsequent publication [25]. although most of the 32 values were predicted well, software was not able to predict the solubility of indomethacin. in the current paper we used the cheqsol approach to reinvestigate the aqueous solubility of indomethacin and determined a value of 8.8 μg/ml. we also found evidence suggesting an amorphous form, as well as evidence to suggest that indomethacin might have decomposed during the measurement reported in [25]. we used nmr and hplc to characterise the decomposition, which occurs at high ph. the results reported in this paper together with the graphical representations of dissolution and precipitation events during the experiments enable us to gain improved understanding of the solubility of indomethacin. experimental indomethacin (i7378, ≥99 %) was purchased from sigma life science and used without further purification. the polymorphic form of this material was not specified. solubility and pka values were measured using the siriust3 automatic titration system (sirius analytical ltd., forest row, uk), using the software supplied with the instrument to calculate results from the experimental data. solubility experiments to prepare solid material for hplc and nmr studies were undertaken on the sirius glpka automatic titration system. solution and solid composition were investigated using nmr spectroscopy, lcms and hplc with uv detection. nmr spectral data (500 mhz 1 h nmr and 125 mhz 13 c) were recorded in a bruker avance 500 mhz broadband. 1 h and 13 c nmr chemical shifts are reported in parts per million (ppm) and referenced relative to their residual solvent peaks. assignments were determined by unambiguous chemical shift, analogue comparison, coupling patterns, or hmqc experiments. lcms data were collected with a waters zq mass spectrometer using electrospray ionisation, connected to a waters 2795 hplc and 996 dad. the column used was an agilent poroshell 120 c-18 4.6 x 30 mm, 2.7 µm particle size, and the solvent mixtures were a: 10 mm ammonium acetate with 0.1 % formic acid, and b: 95 % aqueous acetonitrile with 0.05 % formic acid. hplc/uv data were collected with a shimadzu spd m-20a pda connected to a shimadzu lc-10 hplc. the column used was a phenomenex kinetexc-18 4.6 x 50 mm, 2.6 µm particle size, and the solvent mixtures were a: 0.1 % formic acid, and b: acetonitrile. solubility measurements were carried out using the cheqsol and curve fitting methods. these methods and the associated bjerrum curves are described in detail elsewhere [26-28]. indomethacin samples of around 2 mg were weighed accurately into glass vials. measurements were performed in 1.5 ml of deionised water. the ph of the prepared sample was raised by adding standardised 0.5 m koh solution, under which conditions all the indomethacin dissolved in ionised form. the transmission of light through the sample was monitored at 500 nm by use of a spectroscopic dip probe connected to a diode array spectrophotometer. the spectroscopic data were used to detect the onset of precipitation as the ph was lowered by adding standardised 0.5 m hcl solution. in cheqsol assays, small aliquots of koh or hcl solution were added after precipitation to maintain the system close to equilibrium, and solubility results were calculated from rates of ph change vs. concentration. in curve fitting assays, only hcl was added, and solubility was calculated by fitting a precipitation bjerrum curve to the data. some curve fitting experiments commenced at low ph; these are discussed later. in order to calculate solubility results from the cheqsol and curve fitting data the pka of indomethacin is required, and a value of 4.13 ± 0.018, i = 0.058m, 25.2 °c was measured in triplicate by a uv-metric method admet & dmpk 2(1) (2014) 18-32 indomethacin solubility doi: 10.5599/admet.2.1.33 21 on the same instrument. some comment is required on this pka measurement. a ph-metric titration method is often chosen to measure the pka of compounds in which the ionisable group shows weak uv activity. however it proved to be surprisingly difficult to measure the pka reliably by the ph-metric method. because of its low aqueous solubility, indomethacin pka must be measured in cosolvent-water mixtures and an aqueous result obtained by extrapolation. this was first attempted in methanol-water solutions adjusted to an ionic strength of 0.15m with kcl. a result of 4.20 was obtained from the yasuda-shedlovsky extrapolation of four pska values measured at methanol percentages between 37.4 % and 52.7 %. however a further 14 methanol-water pska values were considered to be too unreliable to include in the extrapolation, and confidence in the methanol-water result was low. measurement was then attempted in solutions containing dioxane-water adjusted to an ionic strength of 0.15 m with kcl. a result of 4.22 was obtained from the yasuda-shedlovsky extrapolation of seven pska values measured at dioxane percentages between 37.4 % and 52.7 %. this value was used in the curve fitting solubility measurement made from low to high ph (figure 2). however the linear extrapolation of the same data yielded a result of 3.60 and the discrepancies between these two dioxane-water extrapolations led to some caution. it has been reported that indomethacin can self-aggregate in pure water and in the absence of ionic strength [29], and it seems likely that this tendency will be stronger at high ionic strength. this tendency for aggregation was not allowed for in the software used in the siriust3 to calculate pka values from titration data, and this could lead to errors in pka measurement. the behaviour of indomethacin during ph-metric pka measurement at 1 mm concentration when forming aggregates in the presence of 0.15 m kcl and varying percentages of solvent is likely to be hard to predict. the difficulty of accounting for aggregation might also explain why other reported measurements of indomethacin pka differ considerably (4.01 +/-0.09, i=0.05 m, 25 °c, ce procedure [30]), (4.17, i=0.5 m, 25 °c [31], (4.5, conditions not described [29]). the effects of aggregation should be less apparent in uv-metric pka measurement, where the sample concentration is around 30 μm. in parallel with these attempts to measure indomethacin pka by ph-metric titration it was observed during the solubility measurements that indomethacin did not fully dissolve at ph 9 in the presence of 0.15 m kcl, which would be the case if the potassium salt of indomethacin was poorly soluble. it is important in cheqsol experiments that the sample is fully dissolved in ionised form at the start of the experiment and that no precipitated salt is present [26]. it was found that indomethacin did dissolve when solutions were prepared in deionised water adjusted to ph 9, and the subsequent solubility measurements were therefore done in a low ionic strength background. in order that the pka value used in the solubility data sets were collected under similar conditions, it was measured in ph-adjusted deionised water (average ionic strength = 0.058m) by a uv-metric method. the carboxylic acid in indomethacin is not part of a chromophore but a small change in absorbance vs. ph was observed, and this was sufficient for reliable measurements to be made. it is understood that ph-metric experiments done at low ionic strength are susceptible to ph electrode calibration errors but these errors are most apparent at ph below 3 and above 11, and will have little effect on the solubility measurements reported here in which all relevant data were between ph 4 and 9. results and discussion evaluating the outlying published result the starting point for this research was to discover why a value of 410 μg/ml for solubility measured by cheqsol was so different from other reported values. comer et al. admet & dmpk 1(2) (2014) 18-32 22 the first step was to re-examine the data from which this result was derived. this experiment used a standard template for measurement of solubility of a monoprotic weak acid, which includes a dissolution phase wherein the sample solution is held at ph 12 for several minutes. the sample will be fully ionised at this ph and is expected to dissolve in aqueous solution. although an experiment starting at lower ph (e.g. between 7 and 10) would similarly ensure the compound is fully ionised, a value of 12 is normally chosen because water has a high buffer capacity at this ph and therefore resists the sample’s tendency to drag the ph down as it dissolves, helping to ensure complete dissolution of the sample. the existence of titration data between the ph of sample dissolution and the ph of precipitation generally improves the quality of cheqsol experiments because it aids the calculation of acidity error and concentration factor, and may in some circumstances provide experimental verification of the sample’s pka. a curve fitting experiment was then run at 25 °c in 0.15 m kcl solution on the siriust3, starting at low ph where the indomethacin was present as a suspension of crystalline solid as obtained from the manufacturer. the suspension was titrated slowly with koh solution and a precipitation bjerrum curve was calculated from the ph measured as the sample dissolved (figure 2). although the data below ph 3 deviates above a mean molecular charge of 0 and may indicate an electrode calibration error, this deviation occurs in a region where the sample is not undergoing ionization, and has no effect on the data above ph 3. it is, however significant that unionised indomethacin is poorly wettable and tends to float, which led to variable data quality. a result of 3.3 μg/ml was determined from the data points between the vertical red lines by the curve fitting procedure described below in the section “investigating the intrinsic solubility of indomethacin”; this value is close to the values listed in table 1. figure 2. curve fitting experiment to determine the solubility of indomethacin; 3.8 mg of indomethacin in 1.5 ml of 0.15 m kcl, titrated with 0.5 m koh at 25 °c. the solid blue triangles denote data points that are included in the curve fitting calculation. unfilled triangles are excluded. the comparison of solubilities measured in two experiments done in opposite ph directions provides a quick way to check for metastable behaviour of the solid state after precipitation. however a solubility of 410 μg/ml seemed too high to represent the solubility of a polymorphic form of indomethacin. subsequent analysis of the data from the “410 μg/ml” experiment indicated that the sample chased equilibrium during the assay, indicating that this result is not likely to correspond to an amorphous form. 1 3 5 7 9 11 ph (concentration scale) -1.0 -0.5 0.0 m e a n m o le c u la r c h a rg e admet & dmpk 2(1) (2014) 18-32 indomethacin solubility doi: 10.5599/admet.2.1.33 23 a careful literature search revealed that indomethacin can undergo base-catalysed hydrolysis [32] and it is shown here that it rapidly decomposes at ph 12 to p­chlorobenzoic acid and 5-methoxy-2-methyl-3indoleacetic acid (1). it is therefore likely that the result of 410 μg/ml is erroneous, and that the species that was observed to chase equilibrium during the cheqsol assay was the least soluble of these two decomposition products. this decomposition is illustrated in figure 1. the experimental work now proceeded in two stages: a study to investigate the decomposition of indomethacin and a series of new cheqsol experiments to re-investigate the intrinsic solubility of indomethacin. investigating the decomposition of indomethacin at high ph the “410 μg/ml” solubility experiment was repeated using the sirius glpka, which provides 15 ml of solution containing 20 mg of solid for examination. nmr shows the presence of peaks representing pchlorobenzoic acid and the substituted indole (1). separation of these two compounds was attempted using column chromatography but did not succeed due to the high polarity of the compounds. instead, methyl esters of the acids were formed and separation via column chromatography was successful. the ester of the indole product was subsequently hydrolysed back to the corresponding acid. these compounds were fully characterized, and it was found that indomethacin had decomposed into p-chlorobenzoic acid and the substituted indole (1) during the dissolution phase of the experiment (15 – 20 min). the pka values of p-chlorobenzoic acid (3.75) and the substituted indole 1 (4.42) were measured. for an experiment at 10 mm concentration, no precipitate was found for the indole (1), and its kinetic solubility was therefore expected to be higher than 10 mm. using the measured value for the pka of p-chlorobenzoic acid, the initial indomethacin decomposition cheqsol experiment was reanalysed and an accurate value for the solubility of p-chlorobenzoic acid was determined (302 μm; 42.4 μg/ml). this was possible because the indole (1) has a much higher solubility than the concentration present in the experiment. the nature of the precipitate was established as p-chlorobenzoic acid. details of the decomposition study and nmr results are given in the appendix. a re-analysis of the bjerrum titration curve of the “410 µg/ml” experiment leads to a conclusion that is consistent with the hypothesis that decomposition had occurred. figure 3 shows that twice as many protons were lost from the added acid as was expected. this arises from the fact that twice as much base has been added as would be necessary for the concentration of monoprotic indomethacin introduced in the experiment, thus suggesting that two acidic protons are present for each sample molecule. if the calculation is modified to account for the presence of two weak acids, namely the poorly soluble pchlorobenzoic acid that precipitated and the substituted indole (1) that remained in solution, then the calculated curve representing p-chlorobenzoic acid fits the experimental one (figure 4). comer et al. admet & dmpk 1(2) (2014) 18-32 24 figure 3. cheqsol solubility bjerrum curve for indomethacin starting from ph 12. the blue star indicates the starting point of the titration. the pink circle indicates the onset of precipitation. the mean molecular charge value of -2 is consistent with the presence of two titratable acids. figure 4. cheqsol solubility bjerrum curve for indomethacin starting from ph 12. the settings were modified into an assay for p-chlorobenzoic acid in the presence of one equivalent of the substituted indole 1. using these settings, a solubility result showing good agreement with the reported intrinsic solubility of p­chlorobenzoic acid was obtained. investigating the intrinsic solubility of indomethacin a new series of cheqsol and curve fitting experiments was conducted to investigate the intrinsic solubility of indomethacin. to check for stability, solutions of indomethacin were prepared at ph 7.4, 9, 12 and >12 and stored for 3 hours, and then compared by hplc/uv. sharp peaks after 4.2 minutes were observed for the solutions at ph 7.4 and 9. however the peaks occurred after 2.8 minutes for the solutions prepared at higher ph, suggesting that the composition of the solution was significantly different to the 1 3 5 7 9 11 ph (concentration scale) -2.0 -1.5 -1.0 -0.5 0.0 m e a n m o le c u la r c h a rg e 1 3 5 7 9 11 13 ph (concentration scale) -1.0 -0.5 0.0 m e a n m o le c u la r c h a rg e admet & dmpk 2(1) (2014) 18-32 indomethacin solubility doi: 10.5599/admet.2.1.33 25 composition of the solutions prepared at lower ph. with this evidence it was decided that decomposition could be avoided if experiments started at a ph of about 9. all experiments were conducted at 25 ± 0.5 °c. samples of indomethacin between 2±0.3 mg were weighed into glass vials. 50 μl of dmso was added manually to aid dissolution and the vials were then placed on the sirius t3 instrument, which added 1.5 ml of deionised water and then raised the ph to between 9.03 and 9.39 by adding about 10 μl of 0.5 m koh solution. in our experience the inclusion of 50 µl dmso in the sample solution (corresponding to a dmso concentration of about 2 % v/v) does not alter significantly the measured solubility values of most compounds. the vial plus contents was then sonicated for 5 minutes to ensure complete dissolution of the indomethacin in ionised form. the solution was then titrated with 0.5 m hcl until the onset of precipitation, which was detected by an in-situ uv probe. the concentration of neutral indomethacin in solution at the onset of precipitation is referred to as the kinetic solubility. after precipitation the experiments followed cheqsol or curve fitting protocols, which are described elsewhere [28]. results of these solubility measurements are listed in table 2. table 2. results of experiments to measure solubility of indomethacin. log s refers to the logarithm to base 10 of the solubility in units of molarity. the system chased equilibrium during the latter stages of all cheqsol experiments, suggesting that the precipitated material had crystallized. in one cheqsol experiment, a few crystals of the original solid were added after the system had begun chasing equilibrium but there was no obvious shift in solubility, suggesting that the system was measuring the same polymorphic form. in all of the cheqsol experiments the indomethacin precipitated initially in a form with mean kinetic solubility of 77 μg/ml that endured for between 5 and 15 minutes before converting to a form with mean intrinsic solubility of 8.8 μg/ml (n = 4). by analogy with published studies [33] it is likely that the initial precipitation is a liquid-liquid phase separation (llps) in which a disordered amorphous solid state is created that later crystallizes. this process is summarised in figures 5 and 6. although the value of 8.8 μg/ml was reproducibly measured in four experiments, it is not possible to claim that it represents the solubility of the least soluble polymorph. the effects of possible aggregation were not modelled in the software. there was 50 µl of dmso present in each experiment, which may affect the result and the crystal form. the curve fitting experiment starting at low ph with the original solid (figure 1) determined a solubility of 3.3 μg/ml, and other workers have reported lower values (table 1). it could be useful in the future to re-measure the solubility in experiments with longer duration (e.g. 24 hours) to check for further figure description ionic strength pka log s μg/ml log s μg/ml log s μg/ml 1 curve-fitting, ph2 up. original solid (3.8 mg) dissolved during experiment. 0.157 m 4.22 (n = 1) -5.0 3.3 5 cheqsol ph9 down. 2 mg + 50µl dmso 0.0065 m 4.13 average (n = 4) -3.6 76.9 -4.6 8.8 std. dev. 0.1 8.3 0.1 1.7 7 curve fitting, ph9 down. 2 mg + 50µl dmso 0.004 m 4.13 average (n = 3) -3.7 68.7 -3.7 79.8 std. dev. 0.0 5.4 0.0 5.8 8 curve fitting, ph9 down. 2 mg + 50µl dmso 0.004 m 4.13 (n = 1) -3.7 72.1 -4.4 13.2 intrinsickinetic amorphous comer et al. admet & dmpk 1(2) (2014) 18-32 26 conversion to a less soluble form; such changes have occasionally been observed in the sirius laboratory and are described in figure 6. figure 5. cheqsol solubility bjerrum curve for indomethacin starting from ph 9. the blue star denotes the start of the experiment. the red triangles denote the addition of hcl titrant. the pink circle indicates the onset of precipitation, and lies on the green line representing the solubility of the initial precipitated form. figure 6. data from figure 5 re-plotted to show the concentrations of the initial precipitated form (plateau on left hand side) and the crystalline form (points from 35 minutes onwards), to which a solid blue line representing the intrinsic solubility has been fitted. the changes in magnitude of the concentration changes associated with the lower plateau may indicate that crystals are consolidating by ostwald ripening. although not evident here, cheqsol experiments with other samples sometimes show concentrations dropping to a lower plateau after longer times, suggesting that a metastable crystalline form has converted to a more stable crystalline form. by contrast the precipitated sample persisted in the higher solubility form throughout three curve fitting experiments, as shown in figure 7. it may be useful to speculate why the sample remained amorphous in the curve fitting experiments but crystallized in the cheqsol. in curve fitting experiments ph is adjusted in one direction only and this often allows the sample to persist in the amorphous state. in admet & dmpk 2(1) (2014) 18-32 indomethacin solubility doi: 10.5599/admet.2.1.33 27 cheqsol experiments, successive aliquots of acid and base are added and it is believed this may stimulate the onset of crystallization after a short amorphous period. although the sample remained amorphous during three curve fitting experiments, it converted soon after precipitation in a fourth experiment to a less soluble form, as shown in figure 8. it is not understood why this conversion took place. figure 7. curve fitting experiment in which indomethacin persisted in a form that is probably amorphous. figure 8. curve fitting experiment in which indomethacin converted soon after precipitation a form that is probably amorphous to a form with lower solubility. it is important to point out that the sirius curve fitting protocol differs from the pion psol method [34]. in the psol method the solubility is calculated using an approach based on mass balance expressions constructed from the equilibrium equations and constants which iteratively derives the concentrations of all species present in solution and those which have precipitated. in the sirius curve fitting method, samples are dissolved in ionised form and the solutions are titrated with acid or base towards the ph where the samples are in neutral form. the solution is a user-supervised automated on-screen graphics exercise in which the user selects the data points to include, and a theoretical bjerrum curve representing the comer et al. admet & dmpk 1(2) (2014) 18-32 28 precipitation and calculated from the pka and proposed solubility result is manually fitted to the selected data points. data collection for curve fitting experiments is fast for compounds that precipitate in the amorphous (i.e. llps) form. this is because the so-called precipitation is actually a phase separation between an aqueous solution and a liquid or supercooled liquid phase. the ph quickly reaches a stable value after each addition of titrant, and the data generally fits the theoretical model well. curve fitting experiments are not suitable for compounds that quickly crystallise after precipitation because it may take many minutes for the ph to reach a stable value after each addition of titrant. these compounds are measured by the cheqsol method. indomethacin is an unusual compound because it tends to remain in amorphous form during curve fitting experiments yet quickly converts to a crystalline form during cheqsol experiments. conclusions indomethacin decomposes rapidly at ph 12. this invalidates measurements of its solubility that involved any exposure to high ph conditions, and illustrates the importance of selecting appropriate assay conditions when analysing acidor base-labile molecules using titration methods. any unexpectedly large mean molecular charge values should be investigated, as they may suggest the occurrence of decomposition. it is shown that in some cases cheqsol assays can be carried out successfully even for phunstable compounds if mild starting conditions are utilised. indomethacin is stable at ph 9. a value of 8.8 μg/ml for the intrinsic solubility of indomethacin was measured in experiments in which all data was collected at ph 9 or below; however, this result may not represent the least soluble form. these experiments also provided strong evidence for the existence of a form of indomethacin with a solubility of about 77 μg/ml, which persisted before crystallization for between 5 and 15 minutes. the authors would like to suggest the following topics for future research. any one of the following would be interesting: to create additional software for calculating solubility results from the ph-metric cheqsol data that includes equilibrium expressions to describe aggregation; to run the cheqsol experiments for longer times in case the form with solubility of 8.8 μg/ml converts to a less soluble form; to examine the precipitates with a polarising light microscope or other tools to provide evidence of their amorphous or crystalline form; to identify a target ph at which indomethacin precipitates as the ph is lowered and then to run controlled supersaturation experiments at higher ph to investigate the duration of supersaturation and the induction time when a form change occurred. who did what: sam judge and louise towes ran pka and solubility measurements using the siriust3. darren matthews ran hplc experiments to validate the sample integrity. bruno falcone and jonathan goodman characterised the decomposition of indomethacin and measured the pka and solubility of p-chlorobenzoic acid and the substituted indole (1). john dearden encouraged the other authors to write this paper and provided valuable literature searches and insights. john comer planned the solubility investigations, created the figures and wrote or edited the text. references [1] h. yamamoto. chem. pharm. bull. 16(1) (1968) 17-19. [2] l. borka. acta pharm. suecica. 11(3) (1974) 295-303. [3] s-y lin. j pharm sci. 81(6) (1992) 572-576. [4] n. kaneniwa, m. otsuka, t. hayashi. chem. pharm. bull. 33(8) (1985) 3447-3455. [5] v. andronis, g. zografi. j. non-cryst. solids. 271(3) (2000) 236-248. [6] j.m. aceves-hernandez, i. nicolás-vázquez, f.j. aceves, j. hinojosa-torres, m. paz, v.m. castaño. j. pharm. sci. 98(7) (2009) 2448-2463. admet & dmpk 2(1) (2014) 18-32 indomethacin solubility doi: 10.5599/admet.2.1.33 29 [7] c. aubrey-medendorp, m.j. swadley, t. li. pharm. res. 25(4) (2008) 953-959. [8] n. hamdi, y. feutelais, n. yagoubi, d. de girolamo, b. legendre. j. therm. anal. cal. 76 (2004) 9851001. [9] b.c. hancock, m. parks. pharm. res. 17(4) (2000) 397-403. [10] s.b. murdande, m.j. pikal, r.m. shanker, r.h. bogner. j. pharm. sci. 99(3) (2010) 1254-1264. [11] d. iohara, f. hirayama, t. ishiguro, h. arima, k. uekama. int. j. pharm. 354(1-2) (2008) 70-76. [12] m. otsuka, t. matsumoto, n. kaneniwa. chem. pharm. bull. 34(4) (1986) 1784-1793. [13] a. rezaei mokarram, a. kebriaee zadeh, m. keshavarz, a. ahmadi, b. mohtat. daru j. pharm sci. 18(3) (2010) 185-192. [14] t. ito, t. sugafuji, m. maruyama, y. ohwa, t. takahashi. j. supramol. chem. 1(4-6) (2001) 217-219. [15] o.a. deshpande, v.b. yadav. int. j. chem.tech. res. 1(4) (2009) 1312-1317. [16] s.h. yalkowsky, r.m. dannenfelser. aquasol database of aqueous solubility, version 5, college of pharmacy, university of arizona, tucson, az, usa, 1992. [17] f. martínez, m.a. peña, p. bustamante. fluid phase equilib. 308(1-2) (2011) 98-106. [18] a. nokhodchi. j. pharm. pharmaceu.t sci. 8(1) (2005) 18-25. [19] c.a.s. bergström, m. strafford, l. lazorova, a. avdeef, k. luthman, p. artursson. j. med. chem. 46(4) (2003) 558-570. [20] s. muchtar, m. abdulrazik, j. frucht-pery, s. benita. j. control. release 44(1) (1997) 55-64. [21] s. jambhekar, r. casella, t. maher. int. j. pharm. 270(1-2) (2004) 149-166. [22] r. bodmeier, h. chen. j. control. release. 12(3) (1990) 223-233. [23] m. hewitt, m.t. cronin, s.j. enoch, j.c. madden, d.w. roberts, j.c. dearden. j. chem. inf. model. 49(11) (2009) 2572-87. [24] a. llinas, r.c. glen, j.m. goodman. j. chem. inf. model. 48(7) (2008) 1289-303. [25] a.j. hopfinger, e.x. esposito, a. llinàs, r.c. glen, j.m. goodman. j. chem. inf. model. 49(1) (2009) 1-5. [26] m. stuart, k. box. anal. chem. 77(4) (2005) 983-990. [27] k.j. box, g. völgyi, e. baka, m. stuart, k. takács-novák, j.e.a. comer. j. pharm. sci. 95(6) (2006) 1298-1307. [28] k. box, j.e. comer, t. gravestock, m. stuart. chem. biodiversity 6(11) (2009) 1767-1788. [29] a. fini, g. fazio, g. feroci. int. j. pharm. 126 (1995) 95-102. [30] m. shalaeva, j. kenseth, f. lombardo, a. bastin, a. j. pharm. sci. 97(7) (2008) 2581-2606. [31] k.g. mooney, m.a. mintun, k.j. himmelstein, v.j. stella. j. pharm. sci. 70(1) (1981) 13-22. [32] b. tsvetkova, i. pencheva, a. zlatkov, p. peikov. int. j. pharm. sci. 4(supplement 3) (2012) 549. [33] s.a. raina, g.z. zhang, d.e. alonzo, j. wu,d. zhu,n.d. catron, y. gao, l.s. taylor. j. pharm. sci. 2014, doi: 10.1002/jps.23826 [34] a. avdeef. pharm. pharmacol. commun. 4 (1998) 165-178. [35] m.z. southard, d.w. green, v.j. stella, k.j. himmelstein. pharm. res. 9(1) (1992) 58-69. [36] k. okimoto, r.a. rajewski, k. uekama, j.a. jona, v.j. stella. pharm. res. 13(2) (1996) 256-64. [37] c.m. wassvik, a.g. holmen, c.a. bergstrom, i. zamora, p. artursson. eur. j. pharm. sci. 29(3-4) (2006) 294-305. [38] j.h. fagerberg, o. tsinman, n. sun, k. tsinman, a. avdeef, c.a. bergstrom. mol. pharm. 7(5) (2010) 1419-30. [39] simulations plus, inc. 42505 10th street west lancaster, ca 93534, usa. http://www.simulationsplus.com/ [40] stardrop: optibrium ltd. 7221 cambridge research park beach drive cambridge cb25 9tl, uk. http://www.optibrium.com/starvue/index.php http://www.simulations-plus.com/ http://www.simulations-plus.com/ http://www.optibrium.com/starvue/index.php comer et al. admet & dmpk 1(2) (2014) 18-32 30 [41] cslogws: chemsilico llc, 48 baldwin st. tewksbury, ma 01876, usa. http://www.chemsilico.com/ [42] alogps: i.v. tetko, j. gasteiger, r. todeschini, a. mauri, d. livingstone, p. ertl, v.a. palyulin, e.v. radchenko, n.s. zefirov, a.s. makarenko, v.y. tanchuk, v.v. prokopenko. j. comput. aid. mol. des 19 (2005) 453-463. [43] wskowwin: http://www.epa.gov/oppt/exposure/pubs/episuite.htm [44] waternt: http://www.epa.gov/oppt/exposure/pubs/episuite.htm http://www.chemsilico.com/ http://www.epa.gov/oppt/exposure/pubs/episuite.htm http://www.epa.gov/oppt/exposure/pubs/episuite.htm admet & dmpk 2(1) (2014) 18-32 indomethacin solubility doi: 10.5599/admet.2.1.33 31 appendix indomethacin decomposition experiment ionic strength adjusted water (10 ml, 0.15 m kcl) was added to indomethacin (130 mg, 0.36 mmol). the ph was brought to 12 by addition of koh solution (1.73 ml, 0.5 m) and the solution was stirred for 40 min under nitrogen. the mixture was titrated towards low ph until precipitation was detected. extraction of product of decomposition experiment the solution was brought to ph 1 by addition of aqueous hcl (3n). the aqueous layer was extracted with etoac (3 x 25 ml). the organic layers were combined, dried over na2so4 and the solvent was removed in vacuo to afford a white solid (100 mg). identification of products of decomposition experiment 1 h nmr (500 mhz, cdcl3, t = 298 k)  7.99 (2h, d, j = 8.5 hz), 7.43 (2h, d, j = 8.5 hz) corresponding to pchlorobenzoic acid. 1 h nmr (500 mhz, cdcl3, t = 298 k)  7.76 (1h, br), 7.14 (1h, d, j = 8.7 hz), 6.98 (1h, d, j = 2.2 hz), 6.78 (1h, dd, j = 8.7, 2.4 hz), 3.84 (3h, s), 3.68 (2h, s), 2.37 (3h, s) corresponding to the substituted indole (1). lcms electrospray ionisation: calc. for [m – h] − 218.08, found 218.4; calc. for p-chlorobenzoic acid c7h5 35 clo2 [m – h] − 154.99, found 155.2 (75%); calc. for p­chlorobenzoic acid c7h5 37 clo2 [m – h] − 156.99, found 157.2 (25%). esterification of decomposition products the mixture of decomposition products was dissolved in meoh (5 ml). hcl (1 m in meoh, 1 ml) was added and the mixture was heated under reflux for 3 h, stirred at room temperature overnight, and heated under reflux again for 3.5 h. the solvent was concentrated in vacuo. purification by flash column chromatography (sio2, 20:1 40-60 petroleum ether / etoac for fraction i, and 4:1 40-60 petroleum ether / etoac for fraction ii) afforded methyl p-chlorobenzoate (40 mg, fraction i), and (5-methoxy-2methyl-indol-3-yl) acetic acid methyl ester (2) (70 mg, fraction ii). 1 h nmr (500 mhz, cdcl3, t = 298 k)  7.92 (1h, s, nh), 7.07 (1h, d, j = 8.7 hz, h7), 7.01 (1h, d, j = 2.2 hz, h4), 6.78 (1h, dd, j = 8.7, 2.2 hz, h6), 3.87 (3h, s, ome), 3.68 (3h, s, coome), 3.68 (2h, s, ch2), 2.31 (3h, s, c2-me). 13 c nmr (125 mhz, cdcl3, t = 298 k)  172.8 (coo), 154.1 (c-5), 133.8 (c-2/3a/7a), 130.3 (c-2/3a/7a), 128.9 (c-2/3a/7a), 111.1 (c-7), 110.9 (c-6), 104.2 (c-3), 100.5 (c-4), 56.0 (ome), 52.0 (coome), 30.3 (ch2), 11.7 (me). hydrolysis of 2 a mixture of 2 (640 mg, 2.75 mmol) and lioh•h2o (1.15 g, 27.5 mmol) in 1:1 thf:water (10 ml) was stirred for 23 h. aqueous hcl (3 m, 5 ml) was added and the ph was brought to 4. the solution was saturated with nacl and extracted with etoac (3 x 25 ml). the organic fractions were combined, dried over na2so4 and concentrated in vacuo. the product was recrystallised twice from hot ethanol to afford (5methoxy-2-methyl-indol-3-yl) acetic acid (5-methoxy-2-methyl-3-indoleacetic acid, 1) (98 mg). 1 h nmr (500 mhz, cdcl3, t = 298 k)  7.72 (1h, br, nh), 7.14 (1h, d, j = 8.7 hz, h7), 6.96 (1h, d, j = 2.3 hz, h4), 6.78 (1h, dd, j = 8.7, 2.4 hz, h6), 3.84 (3h, s, ome), 3.65 (2h, s, ch2), 2.35 (3h, s, me). comer et al. admet & dmpk 1(2) (2014) 18-32 32 1 h nmr (500 mhz, cd3od, t = 298 k)  7.11 (1h, d, j = 8.7 hz, h7), 6.95 (1h, d, j = 2.4 hz, h4), 6.67 (1h, dd, j = 8.7, 2.4 hz, h6), 3.79 (3h, s, ome), 3.61 (2h, s, ch2), 2.34 (3h, s, me). 13 c nmr (125 mhz, cd3od, t = 298 k)  176.2 (coo), 155.0 (c5), 135.0 (c2/3a/7a), 132.1 (c2/3a/7a), 130.2 (c2/3a/7a), 111.9 (c7), 111.2 (c6), 104.8 (c3), 101.2 (c4), 56.3 (ome), 30.9 (ch2), 11.4 (me). ©2014 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ manuscript doi: 10.5599/admet.4.2.280 62 admet & dmpk 4(2) (2016) 62-83; doi: 10.5599/admet.4.2.280 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper pharmacogenetic profiling and metabolic activity of human embryonic stem cell derived hepatocytes: focus on cyp450mediated oxidation jenna l. voellinger and edward j. kelly department of pharmaceutics, university of washington, seattle, wa 98195. e-mail addresses: jenna l. voellinger, jennav2@uw.edu; edward j. kelly, edkelly@uw.edu. received: march 25, 2016; revised: may 14, 2016; published: june 29, 2016 abstract primary human hepatocytes are commonly used to evaluate liver drug metabolism and toxicity. pluripotent stem cell derived hepatocytes (scdhs) have the potential to overcome access and function-related limitations associated with primary hepatocytes. scdhs may also provide an improved system for evaluating genotype phenotype relationships, e.g. cytochrome p450 (cyp) gene polymorphisms and their impact on drug metabolism and toxicity. however, in order for scdhs to become routinely used in preclinical drug metabolism and toxicity screening, they must demonstrate reproducible activity of drug metabolism proteins, particularly the oxidative cyp enzymes. we characterized the differentiation status of scdhs, with emphasis on the expression and functional capacity of drug-metabolizing enzymes, and genetically profiled the commonly used wicell® hesc lines, focusing on cyp2d6 in our analysis. an assay of enzymatic function in scdhs using a cyp cockta il assay designed to measure 8 different isozymes revealed only minimal activity for cyp3a, below that observed in primary hepatocytes. with regard to cyp2d6 gene copy number variation, we found the h1 line has only one gene copy, which also harbors the cyp2d6*41 splicing defect, predictive of a cyp2d6 poor/intermediate metabolizer. we identified no cyp2d6 gene duplications, indicating no representative ultra-rapid metabolizer. the h7 and h14 lines are heterozygous for the non-functional cyp2d6*4 variant resulting in a predicted intermediate metabolizer phenotype. in addition, we compared the penetrance of the cyp2d6*41 splicing defect in scdhs and liver tissue via reverse-transcription pcr assay. we found incomplete penetrance of the cyp2d6*41 allele in liver tissue and variable penetrance in scdhs. based on gene expression profiling, scdhs most closely resemble fetal hepatocytes, especially with regards to afp, cyp3a7 and fmo1 expression. finally, these studies indicate a low degree of genetic diversity of pharmacogenetically-relevant genes in the wicell® hesc lines. keywords drug metabolism; adme; in vitro introduction preclinical screening and characterization of the metabolic pathways of a new drug is a necessary part of its development and can be used to aid in predicting drug bioavailability, drug efficacy, drug-drug interactions, and drug toxicity. the main family of enzymes responsible for oxidative drug metabolism is the cytochrome p450s (cyps), in particular members of just three subfamilies (cyp1, cyp2 and cyp3) are primarily responsible for the metabolism of most drugs [1]. the drug-metabolizing cyps are highly expressed in the liver, making it the primary organ to study for the metabolism, detoxification and http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:jennav2@uw.edu mailto:edkelly@uw.edu admet & dmpk 4(2) (2016) 62-83 pharmacogenetics and cyp activity doi: 10.5599/admet.4.2.280 63 elimination of drugs, and the focus of drug optimization and safety studies [2, 3]. pharmacogenetic variability, in drug-metabolizing enzymes such as the cyps, can influence an individual’s susceptibility to drug toxicity, contributing to observed adverse drug reactions (adrs) [4]. drug-drug interactions are a major concern of pharmaceutical companies, particularly with respect to adrs, a leading cause of preventable death in the united states [4]. primary human hepatocytes are the most physiologically relevant in vitro preclinical screening system used to evaluate drug metabolism in the liver [5]. this is due to their high-expression of phase i and ii drug-metabolizing enzymes, as well as phase iii transporters. however, primary human hepatocytes have certain features that limit their utility in preclinical drug development, including: high donor variability, uncertain availability, short-term survival in culture, and a rapid loss of hepatic functions in culture [5-7]. stem cell derived hepatocytes (scdhs) may provide an alternative or complementary in vitro tool to primary human hepatocytes for evaluating drug metabolism and liver toxicity characteristics. pluripotent stem cells have high self-renewal abilities, providing a potentially unlimited source of hepatocytes of defined genotype and phenotype [6, 8]. in order for scdhs to become routinely used in preclinical drug metabolism and toxicity screening, they must demonstrate reproducible activity of drug metabolism proteins, particularly the oxidative cyp enzymes. additionally, it would be ideal if scdhs produced metabolic activities comparable to that of primary human hepatocytes. since scdhs have defined genotypes, they may also provide an improved system for evaluating genotypephenotype relationships, e.g. cyp gene polymorphisms and their impact on drug metabolism and toxicity, and in particular drug-induced liver injury. overall, published studies demonstrate that scdhs generated from either human embryonic stem cells (hescs) or induced pluripotent stem cells (ipscs), do express cyps and other drug-metabolizing enzymes, however they seem to be at levels below those in primary human hepatocytes. in addition, the majority of studies have also reported scdhs as still expressing α-fetoprotein (afp), which is highly expressed in fetal livers and not typically expressed in healthy adult livers, indicating the cells may still be in an immature state. despite the advances made in generating scdhs, the biggest limitation in the field to date is the inability to reproducibly and consistently generate mature hepatocytes with high levels of functional drug metabolizing enzymes. in order for scdhs to be successfully used as an in vitro assay platform for metabolite profiling, this is something that must be overcome. this indicates there is a significant need for improvement on generating and characterizing scdhs, particularly with regards to their metabolic capacity as it compares to primary human hepatocytes. we sought to further investigate the state of scdhs as they compare to cryopreserved human hepatocytes, with emphasis on the expression and functional capacity of drug-metabolizing enzymes. this was done in part through gene expression profiling of pluripotency markers, hepatocyte markers, drug-metabolizing enzymes and transporters. to evaluate the functional capacity of cyp enzymes in scdhs, metabolite formation was assessed using a cyp cocktail approach and lc-ms/ms. additionally, to demonstrate the utility of scdhs in pharmacogenetic predictive screening, we sought to genetically profile the commonly used wicell® hesc lines h1, h7, h9, h13 and h14 using the affymetrix dmet™ plus chip array. this array covers approximately 2000 polymorphisms in absorption, distribution, metabolism and excretion (adme) relevant genes. we focused on cyp2d6 in our analysis in utilizing scdhs for genotype-phenotype predictions, given its highly polymorphic nature and role in metabolizing about 20 % of marketed drugs [9, 10]. in terms of testing genotype-phenotype relationships for cyp2d6 in scdhs, we assessed the penetrance of cyp2d6*41, a splicing defect which results in a cyp2d6 splice product lacking exon 6, through use of a reverse-transcription pcr assay. voellinger and kelly admet & dmpk 4(2) (2016) 62-83 64 experimental pluripotent stem cells the hesc lines h1 (wa01), h7 (wa07), h9 (wa09), h13 (wa13), and h14 (wa14) were obtained from wicell® research institute (madison, wi, http://www.wicell.org) and propagated at the university of washington (uw) institute for stem cell and regenerative medicine. all studies were conducted under approval of the university of washington embryonic stem cell research oversight committee. cryopreserved human hepatocytes and heparg cells cryopreserved human hepatocyte lots hu8110 and hu8114 were purchased from invitrogen (carlsbad, ca) and cultured according to manufacturer’s instructions. briefly, cells were thawed in chrm® media and plated in williams medium e supplemented with hepatocyte plating supplement pack from invitrogen on 24-well plates coated with rat collagen i. following a 6-hour incubation period to allow for cell attachment, the media was changed to williams medium e supplemented with hepatocyte maintenance supplement pack from invitrogen. cells were maintained in this hepatocyte maintenance media for 4 days with daily fluid changes. cryopreserved human hepatocyte lots hum4012 and hum4034 were purchased from triangle research labs (trl, research triangle park, nc) and cultured according to manufacturer’s instructions. briefly, cells were thawed in animal thawing and plating medium from trl and plated in human hepatocyte plating medium from trl supplemented with the plating supplement on 24-well plates coated with rat collagen i. following a 6-hour incubation period to allow for cell attachment, the media was changed to hepatocyte maintenance medium supplemented with the maintenance supplement from trl. cells were maintained in this hepatocyte maintenance media for 3 days with daily fluid changes. in addition, plated primary human hepatocytes were kindly gifted from dr. david eaton’s lab at the university of washington. undifferentiated heparg cells were obtained from biopredic international (saint-gregoire, france). undifferentiated heparg cells were cultured in williams medium e (invitrogen) supplemented with growth supplement (biopredic) for 2 weeks with fluid changes every 2-3 days. after 2 weeks in culture, heparg cells were differentiated using a 50:50 mix of media supplemented with growth supplement and williams medium e supplemented with differentiation supplement (biopredic). after 2 days in culture, cells were switched to 100 % media with differentiation supplement. cells were maintained in media with differentiation supplement for 2 weeks, with fluid changes every 2-3 days. hepatocyte differentiation hescs were cultured on mouse embryonic fibroblasts (mefs) before being transferred to matrigel™ (bd biosciences, san diego, ca)-coated plates using mtesr™1 media (stem cell technologies, vancouver, canada). hepatocyte differentiation was induced in three stages, as published by hay et al [11, 12]. briefly, differentiation was initiated at 60-70 % confluence by replacing the mtesr™1 media with priming medium: rpmi 1640 containing 1xb27 (invitrogen), 100 ng/ml activin a (peprotech, rocky hill, nj), and 50 ng/ml wnt3a (r&d systems, minneapolis, mn). after 72 hours, with daily media changes, the cells were switched to differentiation medium: knockout-dmem containing 20 % knockout serum replacement, 1 mm glutamine, 1 % nonessential amino acids, 0.1 mm β-mercaptoethanol (invitrogen), and 1 % dimethyl sulfoxide (dmso) (sigma, st louis, mo) for a further 5 days. lastly, the cells were switched to maturation media: hepatozyme containing 10 μm hydrocortisone 21-hemisuccinate, 2 mm glutamine, 10 ng/ml hepatocyte growth factor and 20 ng/ml oncostatin m (r&d systems). the media was changed every other day during maturation until the end of differentiation around day 17. to investigate the inductive capacity http://www.wicell.org/ admet & dmpk 4(2) (2016) 62-83 pharmacogenetics and cyp activity doi: 10.5599/admet.4.2.280 65 of cyp enzymes, cells were cultured in the presence of 10 μm rifampicin (sigma), 1 mm phenobarbital (sigma), or 0.1 % dmso as a vehicle control for 72 hours before the end of differentiation on day 17. additional h9 scdhs were a kind gift from dr. david hay at the university of edinburgh. scdhs used for the cyp cocktail assay were differentiated using a slightly different method. briefly, hescs were cultured on matrigel™ plates using mtesr™1 media as above. prior to the start of differentiation, cells were passaged with 0.05 % trypsin-edta (invitrogen) to single cells and 1 x 10 5 cells per well were plated on fresh matrigel™ coated 24-well plates in mtesr™1 media containing 10 μm y27632 (rock inhibitor) (selleck chemicals, houston, tx). cells were cultured in these conditions for 24 hours and then differentiation was initiated with priming medium as described above. fetal and adult liver tissue six human fetal liver tissue samples were obtained from the birth defects research laboratory at the university of washington, courtesy of dr. qingcheng mao. fetal liver tissues were obtained with use of non-identifiable biological specimens/data and so exempt from human subjects review. human liver tissue samples were obtained from the uw school of pharmacy human liver bank (seattle, wa) [13]. immunocytochemistry cells on day 17 of differentiation were fixed with ice-cold methanol for 10-15 minutes, washed with pbs, and permeabilized with 0.5 % triton x-100 for 2 minutes. following a pbs wash, cells were blocked with 2 % normal goat serum for 1 hour at room temperature. cells were then washed with pbs again and incubated overnight at 4 °c with the following primary antibodies: anti-afp at 1:600 (dako, carpinteria, ca); antialbumin at 1:1000 (a gift from dr. jean campbell, department of pathology, university of washington); anti-hnf4α at 1:100 (santa cruz biotechnology, dallas, tx); and anti-cyp3a at 1:100 (santa cruz biotechnology). after washing with pbs, cells were incubated with alexa fluor® 488 secondary antibodies at 1:1000 for 1 hour at room temperature (invitrogen). finally, cells were mounted with slowfade® gold antifade mount with dapi (invitrogen). dna isolation and genotyping genomic dna was isolated from the wicell® hesc lines using the qiagen (valencia, california) dneasy blood and tissue kit according to the manufacturer’s recommendations. dna was quantified on a biorad smartspec plus (hercules, ca) spectrophotometer, measuring absorbance and 260/280 nm ratio. genotyping of hesc lines h1, h7, h9, h13, and h14 was conducted using the affymetrix dmet™ plus array according to manufacturer’s recommendations (santa clara, ca, http://www.affymetrix.com). analysis was performed using affymetrix’s dmet™ console analysis software. quantitative real-time pcr (qrt-pcr) and cyp2d6 copy number assay rna was isolated using trizol reagent from invitrogen (carlsbad, ca) following manufacturer’s recommendations, and quantified on a biorad smartspec plus (hercules, ca) spectrophotometer, measuring absorbance and 260/280 nm ratio. complementary dna (cdna) was synthesized using 1 μg total rna and the taqman® reverse transcription reagents in 10 μl total volume (invitrogen). following synthesis, cdna was diluted to 5 ng/μl and 20 ng was used for qrt-pcr analysis. qrt-pcr for cyps 3a4, 3a7, 1a2, 1b1, 2a6, 2b6, 2c9, 2c19, 2c8, 2d6, 2e1, fmo1, fmo3, ugt1a1, ugt1a6, ugt2b4, ugt2b7, gstt1, gstm1, gstp1, oatp1b1, oatp1b3, mrp2, abcb1, abcg2, atp7b, nanog, sox2, oct4, alb, afp, hnf4α, car, pxr, ahr, and por was carried out using taqman® gene expression assays from applied biosystems (foster city, california) with gusb as the housekeeping gene. cyp2d6 copy number was voellinger and kelly admet & dmpk 4(2) (2016) 62-83 66 assessed using a taqman® copy number assay from applied biosystems (hs00010001_cn) with rnasep as a single copy gene calibrator, according to the manufacturer’s recommendations, using their sds 2.3 and copy caller v1.0 programs. one copy and more than two copy positive controls were identified in the uw school of pharmacy human liver bank (seattle, wa) through sequencing analysis as previously described. cyp2d6*41 reverse-transcription pcr assay total rna was isolated using trizol reagent (invitrogen), following manufacturer’s recommendations, and quantified on a biorad smartspec plus (hercules, ca) spectrophotometer, measuring absorbance and 260/280 nm ratio. cdna was synthesized using the taqman reverse transcription kit (invitrogen) from 1 μg of total rna. to detect the presence of exon 6 in cyp2d6, a 579-bp pcr fragment spanning a region from exon 5 to exon 8 was amplified with primers exon5/8for3 (5’-tccccgtcctcctgcata-3’) and exon5/8rev3 (5’-gtgttcggggtggaagcg-3’) (idt, coralville, iowa). if exon 6 is skipped due to the presence of cyp2d6*41, a smaller pcr fragment of 438-bp is expected. pcr was run under the following conditions, 95 °c for 5 minutes, 92 °c for 20 seconds, 59.6 °c for 30 seconds, 72 °c for 2 minutes, 72 °c for 10 minutes, 4 °c hold, repeating steps 2-4 for 40 cycles. pcr products were run on 1.5 % agarose gels and visualized by staining with 0.5 μg/ml ethidium bromide (biorad). human liver samples used for this study were obtained from the uw school of pharmacy human liver bank. genotyping for cyp2d6*41 was done using an existing taqman® assay (c_34816116) (applied biosystems). cyp cocktail assay and lc-ms/ms two cyp cocktail assays developed in dr. jashvant unadkat’s lab at the university of washington were used to assess functional cyp activity [14]. cocktail a assesses cyps 3a, 2c9, 1a, 2d6, 2a6, and cocktail b assesses cyps 2b6, 2c8, and 2e1. acetaminophen, amodiaquine dihydrochloride dihydrate, bupropion, chlorzoxazone, coumarin, phenacetin, and tolbutamide were purchased from sigma. dextromethorphan, d5-diazepam, dextrorphan, testosterone, n-desethylamodiaquine, hydroxy-bupropion, 4hydroxytolbutamide, 6-hydroxychlorzoxazone, and 6-β-hydroxytestosterone were purchased from cerilliant (round rock, tx). 7-hydroxycoumarin was a kind gift from dr. unadkat’s lab. for the analytical assays, a set of metabolite standards with nine calibration points and a separate set of three quality control (qc) samples were prepared. the calibrator concentration ranges were: acetaminophen, 0.031-8 μm; oh-bupropion, 0.002-0.5 μm; 4-oh-tolbutamide, 0.013-3.3 μm; 6-ohtestosterone, 0.31-80 μm; 6-oh-chlorzoxazone, 0.0625-16 μm; 7-oh-coumarin, 0.0042-1.08 μm; ndesethylamodiaquine, 0.0083-2.12 μm; dextrorphan, 0.0017-0.44 μm. the qc concentration ranges were: acetaminophen, 0.125-4 μm; oh-bupropion, 0.0078-0.25 μm; 4-oh-tolbutamide, 0.052-1.66 μm; 6-ohtestosterone, 1.25-40 μm; 6-oh-chlorzoxazone, 0.25-8 μm; 7-oh-coumarin, 0.017-0.54 μm; ndesethylamodiaquine, 0.033-1.06 μm; dextrorphan, 0.0068-0.22 μm. all calibrator and qc points were made in hepatozyme maturation media, an equal volume of calibrator or qc was then added to an equal volume of acetonitrile containing 0.06 ng/μl of internal standard d5-diazepam and centrifuged at 13,000 rpm for 10 minutes. the supernatant was removed and stored at 4 °c until analysis. cyp assay incubations were performed on day 17 scdhs and cryopreserved hepatocytes lots hum4012 and hum4034. incubations were performed as described by liu et al., with some modifications [14]. briefly, standard stock mixtures of cocktail a and b were prepared and diluted in hepatozyme maturation media for cell incubations to the following final concentrations: cocktail a, tolbutamide, 75 μm, testosterone, 100 μm, dextromethorphan, 5 μm, coumarin, 0.5 μm, phenacetin, 50 μm; cocktail b, bupropion, 100 μm, amodiaquine, 4 μm, chlorzoxazone, 25 μm. the final concentration of organic solvent in cell incubations was 1 %. cocktail incubations with scdhs were performed for 2 hours at 37 °c. at the end of the incubation admet & dmpk 4(2) (2016) 62-83 pharmacogenetics and cyp activity doi: 10.5599/admet.4.2.280 67 period, the media was collected and added to an equal volume of acetonitrile containing 0.06 ng/μl d5diazepam. cocktail incubations with cryopreserved hepatocytes were performed for 30 minutes or 1 hour at 37 °c, media was collected at the end of the incubation and added to an equal volume of acetonitrile containing d5-diazepam. all incubation samples were centrifuged at 13,000 rpm for 10 minutes. the supernatant was removed and stored at 4 °c until analysis. the cyp assay metabolites were analyzed using an agilent 6460 triple quadrupole mass spectrometer (agilent technologies, santa clara, ca) in the university of washington mass spectrometry center as described by liu et al., with minor modifications [14]. briefly, hplc separation was achieved using an agilent xdb c18 2.1 x 150 mm, 5 μm column. compounds were eluted at a flow rate of 0.25 ml/min, using the following gradient of 0.1 % formic acid in water (solvent a) and 0.1 % formic acid in acetonitrile (solvent b): for the first 1.5 minutes, 99 % solvent a, a linear decrease to 70 % solvent a until 4 minutes, maintained at 70 % solvent a until 6 minutes, a linear decrease to 25 % solvent a until 9 minutes and held at 25 % solvent a until 13 minutes to wash and cycled back to initial conditions. the column was allowed to reequilibrate for 4 minutes before the next sample was injected. results and discussion the university of wisconsin has made wicell® lines available to researchers for well over a decade [15]. this has made them some of the most widely used hesc lines in research, and in particular the most commonly used hesc lines in scdh research, specifically the h1 and h9 lines [11, 16-35]. to be useful in preclinical drug metabolism and toxicity screening, scdhs must demonstrate consistent and reproducible activity of drug metabolism proteins, in particular the oxidative cyp enzymes. in addition, it would be ideal if scdhs demonstrate metabolic activities comparable to human primary hepatocytes. we sought to further investigate the state of scdhs as they compare to cryopreserved human hepatocytes, with emphasis on the expression and functional capacity of drug-metabolizing enzymes. in addition, when evaluating cyp activity versus primary hepatocytes, it’s important to take pharmacogenetic variability into account. to demonstrate the utility of scdhs in pharmacogenetic predictive screening, we also sought to genetically profile the commonly used wicell® hesc lines, focusing on cyp2d6 in our analysis. hepatocyte differentiation and gene expression hepatocyte differentiation was performed using a published protocol [11, 12]. briefly, this protocol is broken into three stages to induce hepatocyte differentiation: first the priming stage primes the cells to the definitive endoderm lineage [11, 33, 36], secondly the differentiation stage induces hepatocyte differentiation to form hepatoblast cells, and finally the maturation stage allows for final differentiation and maturation of the resulting hepatocytes. this differentiation protocol yielded scdhs which morphologically resemble hepatocytes by displaying the hepatocyte cuboidal morphology and bi-nucleated cells (figure 1a and b). in addition, the scdhs were immunocytochemically positive for afp, albumin, hnf4α and cyp3a (figure 1c). positive immunocytochemical staining was uniform across the wells with the exception of colonies of cells that failed to differentiate. in order to evaluate the maturity level of scdhs and gain a better understanding of their potential utility in preclinical drug metabolism screens, gene expression profiling of select pluripotency markers, hepatocytes markers, transcription factors, drug-metabolizing enzymes and transporters was done. this expression profiling was performed for scdhs at each stage of differentiation, hescs, cryopreserved hepatocytes, heparg cells, human fetal liver tissue and human adult liver tissue. heparg cells are a cell line derived from a human adult hepatocarcinoma patient, which resemble the bi-potential hepatoblast cell voellinger and kelly admet & dmpk 4(2) (2016) 62-83 68 [37-39]. these cells can be cultured in their hepatoblast, or undifferentiated state, and then undergo a differentiation process to form both biliary epithelial cells and hepatocytes [37-39]. heparg cells express many of the drug-metabolizing enzymes and more closely resemble primary hepatocytes than the commonly used hepg2 cells [39]. to determine if scdhs retain stem cell characteristics, the pluripotency markers nanog, oct4 and sox2 were evaluated. all results are reported as fold change compared to differentiated heparg cells, with gusb as the housekeeping gene. from the gene expression profiling, nanog, oct4, and sox2 expression dropped as differentiation progressed from stem cell to scdhs (figure 2), suggesting the cells are undergoing a differentiation process and are not retaining their pluripotency. notably, expression of nanog and oct4 in scdhs was similar to expression in differentiated heparg cells, which was below the level of expression in fetal liver tissue (figure 2). sox2 expression in scdhs was more variable between the different cell lines, and ranged from 0.4to 15-fold compared to differentiated heparg cells, while sox2 expression in fetal liver tissue ranged from 0.4to 7-fold compared to differentiated heparg cells. expression of all three pluripotency markers was low in the adult human liver tissues evaluated. figure 1. scdh photomicrographs and immunocytochemistry. (a,b) photomicrographs depicting h1 and h9 scdhs, respectively, at each stage of differentiation; hesc, definitive endoderm (de), hepatoblast, and hepatocyte. (c) immunocytochemistry results for albumin, α-fetoprotein (afp), hepatocyte nuclear factor 4α (hnf4α), and cyp3a on day 17 scdhs. positive staining is indicated in green, dapi was used to stain nuclei, shown in blue. a similar expression pattern was observed for three hepatocyte markers (afp, alb and hnf4α), with expression increasing from stem cells through differentiation to scdhs (figure 2). for albumin, expression in scdhs was lower than the expression in differentiated heparg cells, fetal and adult liver tissues. for hnf4α, expression in scdhs was comparable to expression in fetal and adult liver tissues, though lower than differentiated heparg cells (figure 2). with regard to afp, scdhs expressed afp to a much greater extent than differentiated heparg cells, ranging from 2800to 13000-fold higher than heparg cells (figure 2). expression of afp in fetal liver tissue ranged from 3600to 30000-fold higher than heparg cells, indicating that scdhs expressed afp at levels comparable to fetal liver tissue. as expected, afp expression admet & dmpk 4(2) (2016) 62-83 pharmacogenetics and cyp activity doi: 10.5599/admet.4.2.280 69 figure 2. scdh gene exp. characterization. qrt-pcr results shown as fold change compared to differentiated heparg cells, which was set to 1. gus-b was used as a housekeeping gene. for all stem cell and scdh samples, error bars represent a pool of triplicate biological replicates and triplicate technical replicates. for all other samples, error bars represent triplicate technical replicates. heparg undiff represents undifferentiated heparg cells; heparg diff represents differentiated heparg cells; fl1, fl2, fl3, fl4, fl5, fl6 represent fetal liver tissue 1, 2, 3, 4, 5, 6, respectively; hl161 and hl167 represent human liver bank samples 161 and 167, respectively. voellinger and kelly admet & dmpk 4(2) (2016) 62-83 70 was minimally expressed in adult liver tissues. the persistent expression of afp in scdhs is in accordance with published reports indicating a high retention of afp expression in scdhs. overall, scdhs appear to express the hepatocyte markers afp, alb, and hnf4α at levels comparable to fetal liver tissue, indicating a close resemblance to a fetal state hepatocyte. transcription factors which regulate drug metabolizing enzymes and transporters were evaluated. the transcription factors constitutive androstane receptor (car), pregnane x receptor (pxr) and aryl hydrocarbon receptor (ahr) were selected because they regulate many of the hepatic phase i and ii drugmetabolizing enzymes and transporters [40, 41]. the expression of car, pxr and ahr in scdhs was minimal (supplemental figure 1). the expression of ahr in scdhs was comparable to that in fetal liver tissues, while the expression of car and pxr in scdhs was lower than that in fetal liver tissues. the expression of cyps and flavin containing monooxygenases (fmo) were evaluated. with regard to fmos, fmo1 is normally expressed to a greater extent in the fetal liver, and so can be used as another marker to evaluate how fetal-like scdhs are. indeed, from gene expression profiling fmo1 was highly expressed in fetal liver tissue, at levels 20to 50-fold higher than that in differentiated heparg cells, while fmo1 expression was undetected in adult liver tissues (figure 2). fmo1 expression in scdhs increased through differentiation, up to 7to 13fold higher than differentiated heparg cells (figure 2). while scdhs retain high expression of fmo1, it was lower than that in fetal liver tissue. the cyp enzymes cyp3a4 and cyp3a7 also exhibit differential expression between fetal and adult liver, with cyp3a7 being the predominant isoform in fetal liver and cyp3a4 being predominant in the adult liver. this pattern was also observed between our fetal and adult liver tissue samples. cyp3a4 was detected in only one scdh sample, with expression well below (0.0003-fold) that observed in heparg cells (figure 2). cyp3a7 was detected in two scdh samples, at 4.5and 11.8-fold higher than heparg cells. additionally, cyp3a7 expression in scdhs was below expression levels in fetal liver tissue (figure 2). from these data, scdhs retain expression of cyp3a7, though cyp3a7 expression was lower than fetal liver tissue, indicating an overall lack of cyp3a expression in scdhs. cyp 1a2, 2c19 and 2d6 were undetectable in scdhs (supplemental figure 1). cyp2a6 was expressed in scdhs, at levels slightly above adult liver tissue, though below those observed in fetal liver tissue (supplemental figure 1). cyp2b6 was undetectable in fetal liver tissue samples, but was expressed in scdhs, though at levels below adult liver tissue (supplemental figure 1). cyp 2c8, 2c9 and 2e1 were detectable in scdhs, but at levels much lower than expression in adult liver tissue, though comparable to expression in fetal liver tissue (supplemental figure 1). overall, there was minimal expression of the hepatic cyp enzymes in scdhs, with the exception of the fetal cyp3a7 enzyme. the low expression of cyp enzymes in scdhs would result in low functional activity and was unexpected given that hnf4α was expressed at a level consistent with fetal and adult liver tissue. hnf4α is known to have a role in supporting the constitutive expression of cyp enzymes in human hepatocytes [42]. given the relative expression of hnf4α, we would expect higher cyp expression in the scdhs. the expression of p450 oxidoreductase (por) was also evaluated. por is essential for normal function of the cyp enzymes it is needed for electron transfer from nadph to the cyp [43, 44]. por expression in scdhs was below that expressed in heparg differentiated cells, ranging from 0.07to 0.56-fold (figure 2). this expression was also well below that found in fetal and adult liver tissues (figure 2). interestingly, the undifferentiated hesc samples exhibited higher por expression than scdhs. a potential reason for the higher expression of por in stem cells could be its use in supporting other cyp enzymes. for example, cyp26 plays an important role in embryonic stem cells to promote cell differentiation through retinoid signaling [45, 46]. the low expression of por in scdhs may contribute to the observed low functional activity for the drugmetabolizing cyp enzymes. admet & dmpk 4(2) (2016) 62-83 pharmacogenetics and cyp activity doi: 10.5599/admet.4.2.280 71 to examine expression of phase ii drug-metabolizing enzymes, udp-glucuronosyltransferases (ugt) and glutathione-s-transferases (gst) were evaluated. expression of ugt2b7 in scdhs was comparable to fetal liver tissue (figure 2). expression of ugt1a1 in scdhs was comparable to expression in both fetal and adult liver tissue and ugt1a6 expression was comparable to that in adult liver tissue (supplemental figure 1). in contrast, expression of ugt2b4 in scdhs was below that of both fetal and adult liver tissue (supplemental figure 1). in general, for gstt1, gstm1 and gstp1 there was largely no difference in expression between stem cells and scdhs (supplemental figure 1). for the h1 and h9 cell lines evaluated in the gene expression profiling, no polymorphisms for the gsts were detected in the dmet genotyping analysis. to examine the expression of drug transporters in scdhs, p-gp, bcrp, mrp2, oatp1b1, oatp1b3, and atp7b mrna levels were measured. atp7b is a copper transporting atpase expressed in the liver, and is responsible for transporting copper from hepatocytes into the bile [47]. mutations in atp7b are associated with wilson’s disease, a disorder in copper metabolism with manifestations in the liver and the brain [47]. expression of atp7b was up-regulated in scdhs, compared to stem cells. the expression in scdhs was slightly higher than differentiated heparg cells, fetal and adult liver tissue (supplemental figure 1). p-gp was expressed in scdhs at levels comparable to that in fetal liver tissue and below heparg differentiated cells and adult liver tissue (figure 2). bcrp expression was higher in scdhs compared to heparg cells, fetal and adult liver tissue (figure 2). expression of oatp1b1 and oatp1b3 was minimal in scdhs, and well below that in fetal and adult liver tissue (supplemental figure 1). lastly, while mrp2 mrna was detected in scdhs, it was below the expression levels in heparg cells, fetal and adult liver tissue (supplemental figure 1). similar to the other results obtained for gene expression, some transporters were minimally expressed (oatp1b1, oatp1b3, mrp2), while some were similar to that in fetal liver tissue (p-gp and atp7b), with the exception of bcrp which was highly expressed in scdhs. figure 3. cyp3a induction. qrt-pcr results for scdhs incubated with the inducer rifampicin or phenobarbital, shown as fold change compared to scdhs incubated with dmso, which was set to 1. gus-b was used as a housekeeping gene. scdh rif represents scdhs incubated with rifampicin; scdh pheno represents scdhs incubated with phenobarbital. scdhs were incubated with the canonical inducers rifampicin and phenobarbital and effects on cyp3a4, cyp3a7, car and pxr expression were evaluated. results are reported as fold change compared to scdhs voellinger and kelly admet & dmpk 4(2) (2016) 62-83 72 incubated with a dmso vehicle control (figure 3). in this instance, basal expression of cyps 3a4 and 3a7 was detected in scdhs. cyp3a4 expression was induced by rifampicin and phenobarbital. phenobarbital induced cyp3a4 in scdhs to levels comparable with a fetal liver tissue sample. cyp3a7 and car expression was induced by phenobarbital. phenobarbital induced cyp3a7 in scdhs to levels comparable with an adult liver tissue sample. the gene expression results reported here indicate scdhs are capable of responding to the inducers rifampicin and phenobarbital. overall, gene expression profiling of scdhs indicates they are largely immature hepatocytes, especially with regard to the cyp enzymes. with regards to afp, cyp3a7 and fmo1 expression, scdhs appear to more closely resemble fetal hepatocytes. this suggests that even though scdhs are not currently in a fully differentiated, or mature, state they may be useful in studying drug-metabolism in neonates, or potentially as a test system to investigate the ontogeny of drug-metabolizing enzymes. however, further investigation of drug-metabolizing enzyme expression throughout the differentiation process and comparison to a more extensive sampling of fetal tissues at different ages than what was utilized in this study would be warranted. if scdhs are to be used in conjunction with, or as an alternative to, cryopreserved hepatocytes a differentiation protocol that generates more mature hepatocytes is needed. one potential method that may enhance scdh differentiation includes the use of small molecules in place of less chemically defined proteins. another method that may enhance scdh differentiation is the manipulation of micrornas (mirnas) during the cell culture process that contribute to hepatocyte development. mirnas regulate many genes, making them similar to transcription factors, and therefore potential regulators of cellular processes such as differentiation. for example, knockdown of mirnas that may contribute to maintenance of a pluripotent, or stem cell state, may allow further maturation of scdhs. alternatively, over-expression of mirnas known to normally be expressed in mature hepatocytes may also allow for further maturation of scdhs. lastly, there is the potential of 3d culture systems to enhance scdh differentiation and functional maturation. culturing the scdhs in a 3d, or microphysiological, culture system that more closely represents the true in vivo micro-environment of a hepatocyte is another potential method to enhance scdh differentiation. cyp cocktail assay to assess the metabolic function of the cyp enzymes in scdhs, a cyp cocktail assay was utilized and metabolite formation was measured using lc-ms/ms. for scdhs, all metabolites were below the lower limit of quantitation (lloq) with the exception of 6-oh-testosterone, indicating some measurable activity of cyp3a (table 1). 6-oh-testosterone was detectable for both h1 and h9 scdhs, with activity at 1.27 pmol/min/10 6 cells. in contrast, activity in primary hepatocytes was 92.9 pmol/min/10 6 cells. the lloq for 6-oh-testosterone was 0.312 μm. positive control primary and cryopreserved hepatocytes generated measurable metabolites for all compounds tested, indicating the incubation assay and lc-ms/ms method was working. based on the cyp cocktail assay, scdhs do not display detectable cyp activity, except for cyp3a, which was expected based on gene expression profiling. overall, these results demonstrate an absence of functional drug-metabolizing enzyme activity in scdhs, corroborating what has largely been reported for scdhs to date; that they are immature compared to cryopreserved hepatocytes, especially with regards to drug-metabolizing enzyme functions. genotyping we focused on cyp2d6 in our analysis in utilizing scdhs for genotype-phenotype predictions, given its highly polymorphic nature and role in metabolizing about 20 % of marketed drugs [9, 10]. cyp2d6 is the most polymorphic drug-metabolizing enzyme, with more than 63 functional gene alleles admet & dmpk 4(2) (2016) 62-83 pharmacogenetics and cyp activity doi: 10.5599/admet.4.2.280 73 (www.cypalleles.ki.se/cyp2d6.htm), including polymorphisms that result in gene deletion, defective splicing and gene duplication. having an in vitro cell system, such as genotyped scdhs, would allow for more thorough characterization of pharmacogenetic effects on drug metabolism, and potentially their impact on the pharmacokinetics of a new chemical entity early on in drug development. in particular, for compounds found to be highly metabolized by cyp2d6, where polymorphisms may lead to therapeutic failure or a drug toxicity risk, this may aid in making go/no-go decisions or help to inform possible inclusion of genotyping/phenotyping of patients in clinical studies. select genotyping results for cyp2d6 can be found in table 2. analysis of cyp2d6 copy number found h1 to have 1 copy, with all other lines having 2 copies of cyp2d6. no cyp2d6 duplications were found, indicating there is no representative of a cyp2d6 ultra-rapid metabolizer within these cell lines (table 2). table 1. cyp cocktail assay results. results are reported as mean (sd). scdh activity, n = 6; primary hepatocyte activity, n = 2. lloq = lower limit of quantitation lloq (μm) scdh activity (pmol/min/10 6 cells) primary hepatocyte activity (pmol/min/10 6 cells) cyp3a 0.312 1.27 (0.361) 92.9 (1.45) cyp1a 0.125 blq 14.9 (0.365) cyp2a6 0.0084 blq 0.0431 (0.0168) cyp2b6 0.0039 blq 1.62 (0.169) cyp2c8 0.017 blq 0.917 (0.066) cyp2c9 0.013 blq 2.85 (0.00544) cyp2d6 0.0017 blq 2.27 (0.122) cyp2e1 1 blq 12.1 (0.695) table 2. cyp2d6 genotyping results. a summary of the cyp2d6 copy number and genotyping results for wicell® hesc lines h1, h7, h9, h13, and h14. based on genotyping results a prediction for expected cyp2d6 phenotype of each cell line was made. pm = poor metabolizer, im = intermediate metabolizer, em = extensive metabolizer. common name probe set id h1 h7 h9 h13 h14 copy number 1 2 2 2 2 cyp2d6*2 2850c>t am_12261 t c/t c/t t/t c/t cyp2d6*3 2549dela am_12267 a a/a a/a a/a a/a cyp2d6*4 1846g>a am_12274 g g/a g/g g/g g/a cyp2d6*6 1707delt am_12276 t t/t t/t t/t t/t cyp2d6*9 2613_2615delaga am_12264 aga aga/aga aga/aga aga/aga aga/aga cyp2d6*10 100c>t am_12285 c c/t c/t c/c c/t cyp2d6*17 1023c>t am_12280 c c/c c/c c/c c/c cyp2d6*29 3183g>a am_12255 g g/g g/g g/g g/g cyp2d6*35 4180g>c am_12247 c c/c c/c c/c c/c cyp2d6*41 2988g>a am_12257 a g/g g/a g/g g/g predicted phenotype pm-im im im-em em im http://www.cypalleles.ki.se/cyp2d6.htm voellinger and kelly admet & dmpk 4(2) (2016) 62-83 74 lines h7 and h14 were both found to have one copy of the cyp2d6*4 null allele, with no other polymorphisms associated with aberrant function detected. with regard to the cyp2d6*41 allele, line h9 was found to be heterozygous for this allele, and the single gene copy of cyp2d6 in line h1 also contains the cyp2d6*41 allele. taking the copy number and genotyping data together, we have made predictions regarding cyp2d6 phenotype in these cell lines, as found in table 2. to summarize, we predict line h1 to be a cyp2d6 poor-intermediate metabolizer, lines h7 and h14 to be intermediate metabolizers, line h9 to be an intermediate-extensive metabolizer, depending on activity from the cyp2d6*41 allele, and line h13 to be an extensive metabolizer. select genotyping results for other cyp enzymes can be found in table 3. notably, all 5 hesc lines were found to be homozygous for cyp3a5*3, a null allele, meaning they are all cyp3a5 non-expressers. in addition, the h14 line is homozygous for cyp2c9*2 and h1 is heterozygous for cyp2c9*3; both of these alleles result in decreased cyp2c9 activity and have documented effects on in vitro and in vivo pharmacokinetics and clinical drug response, e.g. warfarin [48]. genotyping results for phase ii/iii enzymes can be found in table 4. lines h7, h13, and h14 are heterozygous for the ugt1a1*28 allele, which results in reduced activity of ugt1a1 [49]. clinically, reduced activity of the ugt1a1*28 allele is associated with development of toxicity following administration of the chemotherapeutic irinotecan (the pharmacogenomics knowledge base: https://www.pharmgkb.org) [50]. additionally, lines h1, h7, and h14 are heterozygous for the nat2*5 allele and h13 is homozygous for this allele, likely resulting in slow acetylator status for these cell lines. the affymetrix dmet™ array genotyping results in their entirety can be accessed from gene expression omnibus (geo): (http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc =gse53889; accession: gse53889). table 3. cyp genotyping results. a summary of cyp genotyping results for wicell® hesc lines h1, h7, h9, h13, and h14. common name probe set id h1 h7 h9 h13 h14 cyp1a2*1f -163c>a am_10785 a/a c/c a/c c/c a/c cyp2a6 deletion cn_cyp2a6 one or more one or more one or more one or more one or more cyp2b6*5 25505c>t am_11426 c/c c/t c/c c/t c/t cyp2b6*6 15631g>t am_11411 g/g no call no call g/g g/t cyp2c9*2 3608c>t am_10100 c/c c/c c/c c/c t/t cyp2c9*3 42614a>c am_10113 a/c a/a a/a a/a a/a cyp2c19*2 19154g>a am_10070 g/g g/g g/g g/g g/g cyp2c19*3 17948g>a am_10068 g/g g/g g/g g/g g/g cyp2c19*17 -806c>t am_10053 c/t c/c c/t c/c c/c cyp2e1*2 1132g>a am_10249 g/g g/g g/g g/g g/g cyp3a5*3 6986a>g am_14759 g/g g/g g/g g/g g/g cyp3a5*6 14690g>a am_14748 g/g g/g g/g g/g g/g cyp3a7*1c -232a>c am_14791 a/a a/a a/a a/a a/a cyp3a7*1c -284t>a am_14796 t/t t/t t/t t/t t/t cyp3a7*2 26041c>g am_14781 c/c c/c c/c c/c c/c https://www.pharmgkb.org/ http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc%20=gse53889 http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc%20=gse53889 admet & dmpk 4(2) (2016) 62-83 pharmacogenetics and cyp activity doi: 10.5599/admet.4.2.280 75 for use in preclinical drug metabolism screening assays, vendors commonly offer pooled donor lots of hepatocytes to mitigate the effects of interindividual variability. with this in mind, the h1 line has one copy of cyp2d6, making it a predicted poor metabolizer in human hepatocytes, meaning scdhs generated from hesc line h1 may not metabolize cyp2d6 probe drugs at a level comparable to primary hepatocytes. in this case it would be important to ensure that a lot of single donor human primary hepatocytes used for comparison with h1-derived scdhs have been characterized as a cyp2d6 poor metabolizer. regarding other cyp isozymes, a commonly used activity test in scdhs is an ethoxyresorufin-o-deethylase (erod) assay for cyp1a activity and inducibility [11, 19, 51, 52]. our genotyping results show the h1 line to be homozygous and h9 to be heterozygous for cyp1a2*1f, an allele which has been shown to influence the magnitude of inducibility of cyp1a2 [48, 53]. table 4. phase ii/iii genotyping results. a summary of genotyping results for phase ii and iii enzymes for wicell® hesc lines h1, h7, h9, h13, and h14. common name probe set id h1 h7 h9 h13 h14 aldh2 1510g>a am_10586 g/g g/g g/g g/g g/g gstm1 deletion cn_gstm 1 one or more 0 one or more 0 0 gstt1 deletion cn_gstt1 one or more one or more one or more one or more one or more nat1*14 560g>a am_14978 a/g g/g g/g g/g g/g nat2*5 341t>c am_15001 c/t c/t t/t c/c c/t sult1a1*2 638g>a am_11005 a/g a/g a/g g/g a/g tpmt*2 238g>c am_13986 g/g g/g g/g g/g g/g tpmt*3b 460g>a am_13980 g/g g/g g/g g/g g/g tpmt*3c 719a>g am_13973 a/a a/a a/a a/a a/a tpmt*4 626-1g>a am_13977 g/g g/g g/g g/g g/g ugt1a1*28 tata-box am_13024 (ta)5or6/ (ta)5or6 (ta)5or6/ (ta)7or8 (ta)5or6/ (ta)5or6 (ta)5or6/ (ta)7or8 (ta)5or6/ (ta)7or8 ugt2b7*2 802c>t am_13465 t/t t/t t/t c/t t/t abcb1 1236c>t am_14612 c/t c/t c/t c/t c/t abcb1 2677g>t>a am_14592 g/t g/t g/t g/t g/t abcb1 3435c>t am_14581 c/t c/t c/t c/t c/t abcg2 421c>a am_13688 c/c c/c c/c c/c c/c slco1a2 404a>t am_10533 a/a a/a a/a a/a a/a slco1a2 516a>c am_10531 a/a a/a a/a a/a a/a slco1b1*1b 388a>g am_10496 a/g a/a a/g a/a a/a slco1b1*4 463c>a am_10498 c/c c/c a/c c/c c/c slco1b3 334g>t am_10481 g/g g/g g/g g/g g/g slco1b3 699a>g am_10482 a/a a/a a/a a/a a/a voellinger and kelly admet & dmpk 4(2) (2016) 62-83 76 the genotyping data presented here is important to better understand the activity of metabolizing enzymes and transporters when evaluating metabolism of probe substrates in scdhs, particularly when being used as a marker to evaluate the efficiency of differentiation of scdhs. the affymetrix dmet™ array used for genotyping in this study provides a thorough evaluation of the major known polymorphisms in adme-related genes. however, there are examples when calls were not made (e.g. cyp2b6*6), such that if evaluating activity of these proteins, individual genotyping for these alleles may be necessary. this study only reports results for the five commonly used wicell® lines, however this approach to genotyping should be considered for all hesc and human ipsc lines used as a target cell source for models of drug metabolism and toxicity. in particular, if a bank of hesc or ipscs is to be established then the use of a genotyping array, such as the affymetrix dmet™ array, should be implemented with results readily available to researchers. as the cost of whole genome sequencing continues to decrease, it will become an attractive alternative to targeted arrays such as what we employed, with the caveat that mining these large data sets for specific adme-related polymorphisms is currently challenging. cyp2d6*41 reverse-transcription pcr assay aberrant splicing resulting from the cyp2d6*41 polymorphism was assessed ex vivo with human liver tissue and in vitro using scdhs. to assess differentiation in the scdhs used for cyp2d6*41 analysis, gene expression of afp, alb, hnf4α, cyp2d6, cyp3a4, nanog and oct4 was evaluated, results are reported as fold change compared to cryopreserved hepatocyte log hum4012 (figure 4a). while expression of hnf4α in scdhs is comparable to that seen in cryopreserved hepatocytes, expression levels of albumin and cyp3a4 were lower than that of cryopreserved hepatocytes. conversely, expression of afp in scdhs is higher than that expressed in cryopreserved hepatocytes, corroborating our previous gene expression results. expression of nanog and oct4 in scdhs was lower than expression in the undifferentiated hesc lines and comparable to cryopreserved hepatocytes. in the case of cyp2d6, expression levels were detectable in h9 scdhs, though below expression in cryopreserved hepatocytes, while h1 scdhs were below the limit of quantitation. this differential expression is in line with the genotyping results for these two cell lines. earlier identification and characterization of the 2988g>a snp, now referred to as cyp2d6*41 (human cyp allele nomenclature website: http://www.imm.ki.se/), revealed it to cause a splicing defect, resulting in exon 6 being skipped, and a predictor of cyp2d6 intermediate metabolizer phenotype [54-56]. toscano et al. identified the splice variant lacking exon 6 in liver samples genotyped for cyp2d6*41 (2988g>a) and found that 2988g>a causes favorable formation of a splice product that lacks exon 6, with up to 7-fold increased levels of the splice variant [55]. we wanted to further elucidate the penetrance of cyp2d6*41 aberrant splicing and, in addition, we wanted to demonstrate the potential utility of scdhs to evaluate genotype-phenotype relationships for cyp2d6. results from pcr amplification using primers exon5/8for3 and exon5/8rev3 are shown in figure 4b and c. the expected amplicons for samples homozygous for cyp2d6*41 are 438-bp, for heterozygous samples both 438-bp and 579-bp, and for non-carriers of cyp2d6*41 579-bp. as shown in figure 4c, certain samples (lanes 1 and 2) homozygous for *41 have a 579bp amplicon as well as the expected 438-bp alternatively spliced band, indicating the unexpected presence of exon 6, and demonstrating incomplete penetrance of this splicing defect. in addition, one sample (lane 7) heterozygous for *41 only displays the 579-bp amplicon, indicating presence of exon 6 in all mrna produced (or suppression of expression of the *41 allele). it should be noted that the aberrantly spliced band may be present, but simply was below the limit of detection in our rt-pcr assay. the h1 scdh sample, with one copy of cyp2d6 containing *41 displays the expected exon 6 spliced band at 438-bp, as seen in figure 4c, lane 13. no properly spliced band was detected, as expected. our scdh cyp2d6*41 http://www.imm.ki.se/ admet & dmpk 4(2) (2016) 62-83 pharmacogenetics and cyp activity doi: 10.5599/admet.4.2.280 77 heterozygous cell, line h9 (lane 14) also displays a band indicating aberrant exon 6 splicing, but only a weak band for the properly spliced transcript fragment. again, bands may be present, but simply were below the limit of detection in our rt-pcr assay; based on observations from the qrt-pcr that cyp2d6 expression in h1 scdhs is below the limit of quantitation and in h9 scdhs is 16 fold lower than that of cryopreserved hepatocytes based on ∆∆ct calculations. we are able to show similar results for cyp2d6*41 splicing in our scdhs as seen in the liver tissue samples. overall, our results corroborate the findings of toscano et al. and provide further evidence that cyp2d6*41 displays incomplete penetrance. figure 4. cyp2d6*41 reverse transcription pcr assay. (a) gene expression results shown as fold change compared to cryopreserved hepatocyte lot hum4012, which was set to 1. (b) schematic representation of the cyp2d6 gene. the polymorphism of interest is 2988g>a located between exons 6 and 7, and results in exon 6 skipping. to detect the presence of exon 6 in cyp2d6, a 579-bp pcr fragment spanning a region from exon 5 to exon 8 was amplified with primers exon5/8for3 (f primer) and exon5/8rev3 (r primer). if exon 6 is skipped due to the presence of cyp2d6*41, a smaller pcr fragment of 438-bp is expected. (c) agarose electrophoretic analysis of reverse transcription pcr products of the cyp2d6 region indicated in (b) from human liver tissue samples and day 17 scdhs generated from wicell® lines h1 and h9. lane m, 100bp marker; lane 1 *41/*41; lane 2 *41/*41; lane 3 *1a/*1a; lane 4 *1a/*1a; lane 5 *2a/*2a; lane 6 *1a/*1a; lane 7 *1a/*41; lane 8 *4a/*41; lane 9 *1a/*41; lane 10 *2a/*41; lane 11 *2a/*41; lane 12 *9a/*41; lane 13 h1 *41; lane 14 h9 *41 carrier. voellinger and kelly admet & dmpk 4(2) (2016) 62-83 78 conclusions pluripotent stem cells have a high self-renewal capability and can differentiate into any cell type in the body, making them an attractive tool for researchers in drug discovery and development. scdhs offer a promising new tool for use in hepatic drug metabolism and toxicity screens, adding to available human derived cell models. there has been extensive research on scdhs for this purpose, and while there is evidence scdhs express drug-metabolizing enzymes, they remain in a de-differentiated state. this indicates there is a significant need for improvement on generating and characterizing scdhs, particularly with regards to their metabolic capacity as it compares to primary hepatocytes. in this study, gene expression profiling and functional assays of cyp activity indicate scdhs are largely immature hepatocytes, lacking significant cyp expression and activity. based on gene expression profiling, scdhs most closely resemble fetal hepatocytes, especially with regards to afp, cyp3a7 and fmo1 expression. functional assay of enzyme activity in scdhs using a cyp cocktail revealed only minor cyp3a activity, below that observed in primary hepatocytes. these data demonstrates that more work needs to be done to promote the maturity of scdhs. enhanced expression and function of drug-metabolizing enzymes in scdhs would make them a more viable option to supplant primary hepatocytes in the pharmaceutical industry for drug metabolism/liver toxicity assays. in addition, scdhs may provide an improved in vitro system for evaluating genotype-phenotype relationships of the cyp enzymes and their impact on drug metabolism and toxicity. this is the first published report of genotyping in the wicell® hesc lines with regards to adme-relevant polymorphisms. in addition, we are able to demonstrate the cypp2d6*41 splicing defect displays incomplete penetrance, with similar results observed in scdhs and liver tissue. the pharmacogenetic data provided here is vital to understanding metabolism profiles when these hesc lines are used as a target tissue source for models of drug metabolism and toxicity. acknowledgements the authors would like to thank the university of washington (uw) institute for stem cell and regenerative medicine, including carol ware, christopher cavanaugh, jennifer hesson and savannah cook for their input on culturing human embryonic stem cells. we would like to thank david hay, sebastian greenhough and dagmara szkolnicka at the university of edinburgh for teaching us their hepatocyte differentiation protocol and providing some samples. we would like to thank sengkeo srinouanprachanh, jesse tsai, theo bammler, fred farin, and richard beyer for assistance in running the affymetrix dmet™ plus array. we would also like to thank james macdonald for assistance in submitting the array data to the gene expression omnibus. support was provided for this work in part from the drug metabolism transport and pharmacogenetics research fund in the school of pharmacy, university of washington (to ejk); es007033 (to ejk); institute of translational health sciences ignition award (to ejk); and t32gm007750 (to jlv). research reported in this publication was supported by the national center for advancing translational sciences of the national institutes of health under award number tl1tr000422 (to jlv). the content is solely the responsibility of the authors and does not necessarily represent the official views of the national institutes of health. references [1] f.p. guengerich, human cytochrome p450 enzymes, in: p.r.o.d. montellano (ed.) cytochrome p450: structure, mechanism, and biochemistry, kluwer academic/plenum publishers, new york, 2005, pp. 377-530. [2] h. zhang, c.d. davis, m.w. sinz, a.d. rodrigues, expert opin drug metab toxicol 3 (2007) 667-687. admet & dmpk 4(2) (2016) 62-83 pharmacogenetics and cyp activity doi: 10.5599/admet.4.2.280 79 [3] j.m. mckim, in vitro approaches for determining liver-specific toxicity of new drug candidates, in: a.g. wilson (ed.) new horizons in predictive toxicology: current status and application, the royal society of chemistry, cambridge, 2012, pp. 157-214. [4] i. johansson, m. ingelman-sundberg, toxicological sciences 120 (2010) 1-13. [5] j.c. davila, toxicological sciences 79 (2004) 214-223. [6] claire n. medine, s. greenhough, david c. hay, biochemical society transactions 38 (2010) 1033. [7] p. godoy, n.j. hewitt, u. albrecht, m.e. andersen, n. ansari, s. bhattacharya, j.g. bode, j. bolleyn, c. borner, j. böttger, a. braeuning, r.a. budinsky, b. burkhardt, n.r. cameron, g. camussi, c.-s. cho, y.-j. choi, j. craig rowlands, u. dahmen, g. damm, o. dirsch, m.t. donato, j. dong, s. dooley, d. drasdo, r. eakins, k.s. ferreira, v. fonsato, j. fraczek, r. gebhardt, a. gibson, m. glanemann, c.e.p. goldring, m.j. gómez-lechón, g.m.m. groothuis, l. gustavsson, c. guyot, d. hallifax, s. hammad, a. hayward, d. häussinger, c. hellerbrand, p. hewitt, s. hoehme, h.-g. holzhütter, j.b. houston, j. hrach, k. ito, h. jaeschke, v. keitel, j.m. kelm, b. kevin park, c. kordes, g.a. kullak-ublick, e.l. lecluyse, p. lu, j. luebke-wheeler, a. lutz, d.j. maltman, m. matz-soja, p. mcmullen, i. merfort, s. messner, c. meyer, j. mwinyi, d.j. naisbitt, a.k. nussler, p. olinga, f. pampaloni, j. pi, l. pluta, s.a. przyborski, a. ramachandran, v. rogiers, c. rowe, c. schelcher, k. schmich, m. schwarz, b. singh, e.h.k. stelzer, b. stieger, r. stöber, y. sugiyama, c. tetta, w.e. thasler, t. vanhaecke, m. vinken, t.s. weiss, a. widera, c.g. woods, j.j. xu, k.m. yarborough, j.g. hengstler, archives of toxicology 87 (2013) 1315-1530. [8] d. szkolnicka, w. zhou, b. lucendo-villarin, d.c. hay, annual review of pharmacology and toxicology 53 (2013) 147-159. [9] w.e. evans, science 286 (1999) 487-491. [10] l.c. wienkers, t.g. heath, nature reviews drug discovery 4 (2005) 825-833. [11] d.c. hay, j. fletcher, c. payne, j.d. terrace, r.c.j. gallagher, j. snoeys, j.r. black, d. wojtacha, k. samuel, z. hannoun, a. pryde, c. filippi, i.s. currie, s.j. forbes, j.a. ross, p.n. newsome, j.p. iredale, proceedings of the national academy of sciences 105 (2008) 12301-12306. [12] d. szkolnicka, s.l. farnworth, b. lucendo-villarin, d.c. hay, curr protoc stem cell biol 30 (2014) 1g 5 11g 5 12. [13] m.f. paine, m. khalighi, j.m. fisher, d.d. shen, k.l. kunze, c.l. marsh, j.d. perkins, k.e. thummel, the journal of pharmacology and experimental therapeutics 283 (1997) 1552-1562. [14] l. liu, g.m. mugundu, b.j. kirby, d. samineni, p.b. desai, j.d. unadkat, biopharmaceutics & drug disposition 33 (2012) 207-217. [15] g. vogel, science 287 (2000) 948-949. [16] n.l. magner, y. jung, j. wu, j.a. nolta, m.a. zern, p. zhou, stem cells 31 (2013) 2095-2103. [17] e. chiao, m. elazar, y. xing, a. xiong, m. kmet, m.t. millan, j.s. glenn, w.h. wong, j. baker, stem cells 26 (2008) 2032-2041. [18] j. cai, y. zhao, y. liu, f. ye, z. song, h. qin, s. meng, y. chen, r. zhou, x. song, y. guo, m. ding, h. deng, hepatology 45 (2007) 1229-1239. [19] h. basma, a. soto–gutiérrez, g.r. yannam, l. liu, r. ito, t. yamamoto, e. ellis, s.d. carson, s. sato, y. chen, d. muirhead, n. navarro–álvarez, r.j. wong, j. roy–chowdhury, j.l. platt, d.f. mercer, j.d. miller, s.c. strom, n. kobayashi, i.j. fox, gastroenterology 136 (2009) 990-999.e994. [20] s. agarwal, k.l. holton, r. lanza, stem cells 26 (2008) 1117-1127. [21] t.n. bukong, t. lo, g. szabo, a. dolganiuc, liver international 32 (2012) 732-741. [22] y. duan, x. ma, w. zou, c. wang, i.s. bahbahan, t.p. ahuja, v. tolstikov, m.a. zern, stem cells 28 (2010) 674-686. [23] m. ghaedi, y. duan, m.a. zern, a. revzin, journal of tissue engineering and regenerative medicine (2012) n/a-n/a. [24] d.c. hay, s. pernagallo, j.j. diaz-mochon, c.n. medine, s. greenhough, z. hannoun, j. schrader, j.r. black, j. fletcher, d. dalgetty, a.i. thompson, p.n. newsome, s.j. forbes, j.a. ross, m. bradley, j.p. iredale, stem cell research 6 (2011) 92-102. voellinger and kelly admet & dmpk 4(2) (2016) 62-83 80 [25] j. heo, e.-k. ahn, h.-g. jeong, y.-h. kim, s.-h. leem, s.-j. lee, e.-k. park, m. yang, biochemical and biophysical research communications 434 (2013) 235-240. [26] s. ogawa, j. surapisitchat, c. virtanen, m. ogawa, m. niapour, k.s. sugamori, s. wang, l. tamblyn, c. guillemette, e. hoffmann, b. zhao, s. strom, r.r. laposa, r.f. tyndale, d.m. grant, g. keller, development 140 (2013) 3285-3296. [27] t.s. ramasamy, j.s.l. yu, c. selden, h. hodgson, w. cui, tissue engineering part a 19 (2013) 360-367. [28] k. takayama, m. inamura, k. kawabata, k. katayama, m. higuchi, k. tashiro, a. nonaka, f. sakurai, t. hayakawa, m. kusuda furue, h. mizuguchi, molecular therapy 20 (2011) 127-137. [29] k. takayama, k. kawabata, y. nagamoto, k. kishimoto, k. tashiro, f. sakurai, m. tachibana, k. kanda, t. hayakawa, m.k. furue, h. mizuguchi, biomaterials 34 (2013) 1781-1789. [30] t. touboul, n.r.f. hannan, s. corbineau, a. martinez, c. martinet, s. branchereau, s. mainot, h. strickmarchand, r. pedersen, j. di santo, a. weber, l. vallier, hepatology 51 (2010) 1754-1765. [31] a. delaforest, m. nagaoka, k. si-tayeb, f.k. noto, g. konopka, m.a. battle, s.a. duncan, development 138 (2011) 4143-4153. [32] d.c. hay, d. zhao, a. ross, r. mandalam, j. lebkowski, w. cui, cloning and stem cells 9 (2007) 51-62. [33] d.c. hay, d. zhao, j. fletcher, z.a. hewitt, d. mclean, a. urruticoechea-uriguen, j.r. black, c. elcombe, j.a. ross, r. wolf, w. cui, stem cells 26 (2008) 894-902. [34] l. rambhatla, c.p. chiu, p. kundu, y. peng, m.k. carpenter, cell transplant 12 (2003) 1-11. [35] d. szkolnicka, s.l. farnworth, b. lucendo-villarin, c. storck, w. zhou, j.p. iredale, o. flint, d.c. hay, stem cells transl med 3 (2014) 141-148. [36] a.b. mclean, k.a. d'amour, k.l. jones, m. krishnamoorthy, m.j. kulik, d.m. reynolds, a.m. sheppard, h. liu, y. xu, e.e. baetge, s. dalton, stem cells 25 (2007) 29-38. [37] p. gripon, s. rumin, s. urban, j. le seyec, d. glaise, i. cannie, c. guyomard, j. lucas, c. trepo, c. guguen-guillouzo, proceedings of the national academy of sciences of the united states of america 99 (2002) 15655-15660. [38] c. aninat, a. piton, d. glaise, t. le charpentier, s. langouet, f. morel, c. guguen-guillouzo, a. guillouzo, drug metabolism and disposition: the biological fate of chemicals 34 (2006) 75-83. [39] s.n. hart, y. li, k. nakamoto, e.a. subileau, d. steen, x.b. zhong, drug metabolism and disposition: the biological fate of chemicals 38 (2010) 988-994. [40] h. li, h. wang, expert opin drug metab toxicol 6 (2010) 409-426. [41] a.h. tolson, h. wang, adv drug deliv rev 62 (2010) 1238-1249. [42] r. jover, r. bort, m.j. gomez-lechon, j.v. castell, hepatology 33 (2001) 668-675. [43] c.e. fluck, p.e. mullis, a.v. pandey, biochem biophys res commun 401 (2010) 149-153. [44] a.v. pandey, c.e. fluck, pharmacology & therapeutics 138 (2013) 229-254. [45] s. langton, l.j. gudas, dev biol 315 (2008) 331-354. [46] l.j. gudas, j.a. wagner, j cell physiol 226 (2011) 322-330. [47] s. lutsenko, r.g. efremov, r. tsivkovskii, j.m. walker, j bioenerg biomembr 34 (2002) 351-362. [48] m. ingelman-sundberg, s.c. sim, a. gomez, c. rodriguez-antona, pharmacology & therapeutics 116 (2007) 496-526. [49] h. smith, eaton, dl, bammler, tk, polymorphisms in xenobiotic conjugation, in: l. costa, eaton, dl (ed.) gene-environment interactions: the fundamentals of ecogenetics, wiley press, new york, 2006, pp. 127-158. [50] j.m. hoskins, r.m. goldberg, p. qu, j.g. ibrahim, h.l. mcleod, jnci journal of the national cancer institute 99 (2007) 1290-1295. [51] y. duan, a. catana, y. meng, n. yamamoto, s. he, s. gupta, s.s. gambhir, m.a. zern, stem cells 25 (2007) 3058-3068. [52] m. ek, t. söderdahl, b. küppers-munther, j. edsbagge, t.b. andersson, p. björquist, i. cotgreave, b. jernström, m. ingelman-sundberg, i. johansson, biochemical pharmacology 74 (2007) 496-503. admet & dmpk 4(2) (2016) 62-83 pharmacogenetics and cyp activity doi: 10.5599/admet.4.2.280 81 [53] r. ghotbi, m. christensen, h.-k. roh, m. ingelman-sundberg, e. aklillu, l. bertilsson, european journal of clinical pharmacology 63 (2007) 537-546. [54] s. raimundo, clinical pharmacology & therapeutics 76 (2004) 128-138. [55] c. toscano, k. klein, j. blievernicht, e. schaeffeler, t. saussele, s. raimundo, m. eichelbaum, m. schwab, u.m. zanger, pharmacogenet genomics 16 (2006) 755-766. [56] t. rau, s. diepenbruck, i. diepenbruck, t. eschenhagen, clin pharmacol ther 80 (2006) 555-558; author reply 558-560. ©2016 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ voellinger and kelly admet & dmpk 4(2) (2016) 62-83 82 supplemental material admet & dmpk 4(2) (2016) 62-83 pharmacogenetics and cyp activity doi: 10.5599/admet.4.2.280 83 supplemental figure 1. scdh gene expression characterization. additional qrt-pcr results shown as fold change compared to differentiated heparg cells, which was set to 1. gus-b was used as a housekeeping gene. for all stem cell and scdh samples, error bars represent a pool of triplicate biological replicates and triplicate technical replicates. for all other samples, error bars represent triplicate technical replicates. heparg undiff represents undifferentiated heparg cells; heparg diff represents differentiated heparg cells; fl1, fl2, fl3, fl4, fl5, fl6 represent fetal liver tissue 1, 2, 3, 4, 5, 6, respectively; hl161 and hl167 represent human liver bank samples 161 and 167, respectively. manuscript doi: 10.5599/admet.0000.0000 254 admet & dmpk 3(3) (2015) 254-259; doi: 10.5599/admet.3.3.207 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index review an overview on the role of telomere, telomerase in degenerative diseases and cancer churong li 1 , yin gang 1 * 1 department of radiotherapy, si chuan cancer hospital, chengdu, sichuan province, china *corresponding author: e-mail: cxqyguestc@163.com; tel.: +86 028-85420305; fax: +86 028-85420305 received: august 07, 2015; revised: august 20, 2015; published: september 05, 2015 abstract telomerase maintains the length of telomeric dna at the end of chromosomes; stabilizes the functions of chromosomes, protects the structure of chromosome dna and regulates normal cell growth. the length and stability of telomeres determine the cellular lifespan. moreover, they correlate with cellular aging and carcinogenesis. researches on the role of telomerase in degenerative diseases and cancers show that it regulates cellular senescence. although telomerase is not an oncogene, it does make tumor cells i mmortal by maintaining the stability of telomeric dna. telomerase-based therapies have become a hotspot in new anticancer and degenerative diseases treatments. this review aims to provide a comprehensive understanding of the role of telomerase and telomere in cellular senescence and cancer. keywords telomerase; telomere; degeneration diseases; cancer introduction telomeres are specialized structures located at the terminal ends of chromosomes and play a critical role in stabilization of chromosomes, maintaining the integrity of the genome [1,2], and regulation of normal cell growth. telomerase is a reverse transcriptase enzyme which synthesizes the g-rich tandem repeats that comprise telomeres using a template on the rna that is complementary to the telomeric repeat. by adding hexameric (ttaggg) repeats, the continued erosion of telomeres is compensated. over the past few decades, studies have shown that telomeres and telomerase are closely related to cellular malfunction, degenerative diseases and tumor progression. therefore, researches into the role of telomerase and telomere in degenerative diseases and cancer shed light on a better understanding of these diseases. telomerase-based therapies, including targeted agents will open up a whole new horizon for degenerative diseases and cancer treatments. telomeres, telomerase and their relationship telomere, from the greek word “telos,” (end), and “meros,” (part) was first named by american geneticist hermann j. muller, who observed through mutation experiments carried out during the 1930s [3]. telomeres which have specialized structures that repeat at the ends of chromosomes are thought to be important for the integrity and stability of chromosomes. this observation implied that cells contained an http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:cxqyguestc@163.com admet & dmpk 3(3) (2015) 254-259 telomere and telomerase in degenerative diseases doi: 10.5599/admet.3.3.207 255 enzyme that could catalyze the lengthening of chromosomal ends. in 1984, elizabeth blackburn discovered the enzyme telomerase in the ciliate tetrahymena, which has the ability to lengthen chromosomal ends by its reverse transcriptase activity [4]. telomeres dna sequences are highly conservative and genus-species specific. human telomeres contain an array of tandem dna repeats (5’-ttaggg-3’), which do not encode any gene product. the length of the telomere dna repeats varies from about 5～15kp. these special dna repeats combine their associated proteins to form the cap of chromosomes, which can prevent chromosome end-to-end fusion, degradation and rearrangement from happening, unless the free ends of the dna molecule are processed as double stranded breaks [5]. to avoid chromosome damage, the end of telomeres forms two internal loops (d-loop and t-loop) with the participation of the shelterin complex [6]. usually the shelterin complex is made up of six proteins which play different roles, including telomeric repeat-binding factors 1 and 2 (trf1 and trf2), trf1-interacting nuclear factor 2 (tin2), repressor activator protein 1 (rap1), tin2-interacting protein 1 (tint1) and protection of telomeres 1 protein (pot1). trf1 and trf2 bind to double stranded ttaggg repeats, and the pot1 binds to the 3’ g-rich overhang. the three remaining factors contained tin2, rap1 and tint1, which interact with the telomere-bound proteins to form a complex. this allows cells to distinguish telomeres from sites of dna damage [7-9]. telomeres maintain the integrity of chromosomes. as the cell divides, about 100-200 bp of telomeric dna sequence is lost. this phenomenon is so-called the “end replication problem”. when the dna repeats strand becomes too short to form loops, the chromosome loses its endpoint protection [9]. this is why the telomeric repeats shorten progressively with each cell division [10]. if the telomeres are shortened enough, telomerase is activated to compensate for the end-replication problem [11], so that the length of telomeres is maintained. human telomerase which is a ribonucleoprotein polymerase (rnp) composed of an essential rna and associated proteins has important function. when chromosome loses its endpoint protection, telomerase is activated to catalyze the addition of ttaggg nucleotide repeats to the 3’ terminus of the chromosome’s dna [12], so that it guarantees the length and function of telomeres. in generally, the telomerase holoenzyme consists of three components: telomerase reverse transcriptase (tert), the telomerase rna component (terc) and telomerase related proteins (tep). the tert and terc are most important. terc as a template can add a ttaggg nucleotide repeat to the 3’ end of the telomere’s leading strand; tert is taken to be the rate-limiting component of telomerase activity [13]. terc varies greatly among species, but tert is relatively conserved throughout evolution [14]. in humans, terc contains 451 nucleotides, which includes an 11-nucleotide template and several conserved regions [13,15]. it can provide the template for the synthesis of telomere dna repeats, and also plays essential roles in catalysis and accumulation of telomerase rnp. the tert gene which located on chromosome 5, consists of 16 exons and 15 introns spanning 35 kb, including 330 base pairs upstream of the translation start site, as well as 37 base pairs of exon 2 of the tert gene. the tert promoter which contains many sites for several transcription factors is gc-rich and lacks tata and caat. it is an important target for therapy of cancer and degenerative diseases. telomerase has other related proteins (tep), such as gar1, nhp2 and nop10. they are assembled in the telomerase complex and play distinctive roles in telomerase function [16]. the role of telomere and telomerase in degenerative diseases telomere length shortening has been evidenced as not only a marker of biological aging but also a mechanism with important functional consequences. it has been considered as a risk factor for multiple aging-associated diseases and conditions. the study of telomere and telomerase biology shows a higher prevalence of development in many neurodegenerative diseases, such as alzheimer’s disease (ad), c. li and y. gang admet & dmpk 3(3) (2015) 254-259 256 parkinson’s disease (pd), cvd, hypertension, atherosclerosis, heart failure, diabetes and so on. this review focuses on providing a general understanding of the role of telomere and telomerase in alzheimer’s disease (ad) and parkinson’s disease (pd). alzheimer’s disease and parkinson’s disease are the two main aging-associated progressive neurodegenerative disorders. studies show that they are related to telomere and telomerase. the relationship between telomere shortening and ad has been studied using in vitro and in vivo models in ad patients. oxidative stress is regarded as the most important influencing factor of telomere shortening. there are some studies connecting the telomere length to alzheimer’s disease. panossian l. a. et al. [17] first found that patients with ad had shorter telomere lengths in peripheral blood mononuclear cells than that of the age-matched control subjects. then, jenkins e. c et al. [18] found that older females with down syndrome and ad-type dementia had shorter t-cell telomeres than that of the age-matched counterpart. moreover, a case control study of 257 individuals revealed a shortened peripheral blood leukocyte (pbl) telomere length in ad patients compared with control subjects [19]. and among individuals with ad, short telomeres were associated with higher mortality rates. another study confirmed the association between short pbl telomere length and ad, and was the first to examine hippocampus telomere length in histopathologically validated ad patients [20]. surprisingly, telomere lengths in the hippocampus were found to be 49 % longer in ad patients than in control subjects (p<0.01). recently, j. nicholas lukens compared the telomere lengths in peripheral blood and cerebellum in alzheimer’s disease and found that reduced pbl telomere length in ad might not reflect reduced telomere length in bulk brain tissue, but may be a marker of changes in a subset of brain tissues or other tissues that affect the pathogenesis of ad [21]. telomerase activity in ad patients was significantly elevated compared with healthy controls and vascular dementia patients, possibly as a response to the telomere erosion accompanying the disease. in addition, the telomerase activity of lymphocytes was significantly correlated with the degree of dementia in ad patients, suggesting an accelerated telomere dysfunction in lymphocytes and impaired immune function of ad patients. the relationship between telomere and pd is not well clear. several studies have shown that telomere and telomerase are related to pd, but other studies indicate that there are no evidence for short telomere length in pd. wang et al. [22] proved that men with shorter telomeres had a lower risk for pd, which was in marked contrast with the observations in other aging-related diseases. in a study of 28 japanese male pd patients, leucocyte telomere length in pd patients was not significant different from those in controls. but mean telomere length in pd patients was found to be 5 kb shorter suggesting telomere shortening was accelerated in pd patients [23]. additionally, maede et al. [24] investigated the age-associated alterations of subtelomeric methylation in japanese pd patients. short telomeres with hypomethylated subtelomeres increased with aging in the healthy controls, but did not change in the pd patients. johanna eerola et al analyzed 131 pd patients and 115 controls. it was found that extremely short telomeres were not more frequent in pd patients than in controls, so they consider there was no evidence for shorter leukocyte telomere length in pd [25,26]. the role of telomere and telomerase in neurodegenerative diseases is not fully understood through above researches, despite the progress made during the last decade. further research is essential to clarify whether the ad and pd are triggered by telomere shortening. the role of telomere and telomerase in cancer telomere studies in human cancer have been extensively reported for diagnostics and prognostics. shortened telomere length demonstrates a significant association with poor cancer outcome [27]. admet & dmpk 3(3) (2015) 254-259 telomere and telomerase in degenerative diseases doi: 10.5599/admet.3.3.207 257 telomere maintenance is regarded as an important mechanism in evading senescence by cancer cells, and the role of dysfunctional telomere biology in cancer is acknowledged. many cancer cases were initiated by reactivating or up-regulating telomerase activity. in addition, in many types of cancer cell, there is a considerable shortening of telomeres despite telomerase expression [28-32]. based on these researches, several anti-cancer drugs and cancer vaccines that target the telomerase are used now in phase iii clinical trials. the relations between telomere, telomerase and cancer have been a hot research topic. cancers occur when a cell acquires multiple genetic mutations which cause the cell to escape from normal controls on replication and migration. once the cells and their offspring multiply uncontrollably, they can invade and damage nearby tissue. some may also break away and travel to parts of the body where they do not belong, resulting in new metastases at distant sites. the lack of telomerase in tumor cells may slow the growth with dividing cells, and lose their telomeres before they do much damage. on the other hand, if cancer cells were able to generate telomerase, they could retain telomeres and might growth indefinitely. researchers suspect that the loss of proliferative capacity observed in human cells that lacking telomerase may help us to avoid cancer. here, we direct our attention to digestive cancers, telomere and telomerase. telomerase activity is especially considered as the most important factor in carcinogenesis, progression and prognosis. fundamental researches showed that telomerase was activated in various cancer tissues and cell lines [33-35]. telomerase activity was shown to be positive in 85 %～100 % of cancer tissues and negative in almost all normal tissues, except for reproductive and haemopoietic cells. in digestive cancers, many studies regard telomerase activity and telomere length are associated with the occurrences of cancer, and as prognostic markers in malignancies. studies indicate telomerase activity is detected in 85 % of the hepatocellular carcinoma specimens [36], 85 %～88 % of gastric carcinomatous tissues [37], and 90 % of colorectal tumors [38]. in gastrointestinal cancer, human telomerase are expressed in various levels between all tumor specimens and normal mucosa. for instance, the expression of telomerase is 81% higher in the tumor than normal mucosa. meanwhile 35 % of normal mucosa showed telomerase activity and all of them contained intestinal metaplasia [39,40]. the study by engelhardt m indicated that increased telomerase activity occurs in colon cancer cells that have undergone extensive telomere shortening relative to surrounding normal tissues, in which telomerase-induced stabilization of telomeres may be critical for the continued proliferation of the malignant clone [38]. the link between telomerase activity and stage suggests that telomerase is up-regulated as a function of increased tumor cell invasion, tumor progression, and metastatic potential in colon cancer. on the research of prognosis, telomere length could be a potential unique biomarker predictive of clinical benefit (i.e. progression free survival) of mcrc patients treated with anti-egfr therapy [41]. the high level of telomerase activity may be an independent prognosis-predicting factor in the patients with colorectal cancer [42]. prospect to sum up, the related research shows that the occurrence and changes of degenerative diseases is closely related to telomere shortening. especially, the study for patients with alzheimer and parkinson shows that the dysfunction of telomerase activity and telomere shortening which may lead to the occurrence of these disease and relate to the severity of the disease. through the study of telomerase activity, it may be possible to find new ways of treatment for these two diseases. in addition, there are studies to explore the relationships and related mechanisms between tumor and telomere, telomerase. telomerase activation and telomere shortening are closely related to the occurrence and development of malignant tumor. currently, the study of telomerase targeted drugs for the treatment of cancer has c. li and y. gang admet & dmpk 3(3) (2015) 254-259 258 entered the phase iii clinical trials. hopefully findings from these studies will emerge as a new approach for cancer therapy. references [1] r.k. moyzis, j.m. buckingham, l.s. cram, m. dani, l.l. deaven, m.d. jones, j. meyne, r.l. ratliff, j.r. wu, proc. natl. acad. sci. usa 85 (1988) 6622-6626. [2] e.h. blackburn, nature 350 (1991) 569-573. [3] h.j. muller, the collecting net 13 (1938) 181-198. [4] c.w. greider, e.h. blackburn, cell 43(2pt1) (1985) 405-413. [5] m.g. ferreira, k.m. miller, j.p. cooper, mol. cell. 13 (2004) 7-18. [6] d. li, q. yuan, w. wang, j. int. med. res. 40 (2012) 1242-1250. [7] d. liu, m.s. o’connor, j. qin, z. songyang, j. biol. chem. 279 (2004) 51338-51342. [8] t. de lange, genes dev. 19 (2005) 2100-2110. [9] u. herbig, w.a. jobling, b.p. chen, d.j. chen, j.m. sedivy. mol. cell 14 (2004) 501-513. [10] j.d. watson, nat. new biol. 239 (1972) 197-201. [11] w.e. wright, m.a. piatyszek, w.e. rainey, w. byrd, j.w. shay, dev. genet. 18 (1996) 173-179. [12] c.w. greider, e.h. blackburn, cell 43 (1985) 405-413. [13] c. autexier, n.f. lue, annu. rev. biochem. 75 (2006) 493 517. [14] q. zhang, n.k. kim, j. feigon, proc. natl. acad. sci. usa 108 (2011) 20325-20332. [15] j. feng, w.d. funk, s.s. wang, s.l. weinrich, a.a. avilion, c.p. chiu, r.r. adams, e. chang, r.c. allsopp, j. yu, et al., science 269 (1995) 1236-1241. [16] k. collins, j.r. mitchell, oncogene 21 (2002) 564-579. [17] l.a. panossian, v.r. porter, h.f. valenzuela, x. zhu, e. reback, d. masterman, j.l. cummings, r.b. effros, neurobiol. aging. 24 (2003) 77-84. [18] e.c. jenkins, m.t. velinov, l. ye, h. gu, s. li, e.c. jenkins jr, s.s. brooks, d. pang, d.a. devenny, w.b. zigman, n. schupf, w.p. silverman, neurobiol. aging. 27 (2006) 941-945. [19] l.s. honig, n. schupf, j.h. lee, m.x. tang, r. mayeux, ann. neurol. 60 (2006) 181-187. [20] p. thomas, n.j. o'callaghan, m. fenech, mech. ageing dev. 129 (2008)183-190. [21] j. nicholas lukens, v. van deerlin, c.m. clark, s.x. xie, f.b. johnson, alzheimers dement. 5(6) (2009) 463-469. [22] h. wang, h. chen, x. gao, m. mcgrath, d. deer, i. de vivo, m.a. schwarzschild, a. ascherio, mov. disord. 23 (2008) 302-305. [23] j.z. guan, t. maeda, m. sugano, j. oyama, y. higuchi, t. suzuki, n. makino, j. gerontol. a biol. sci. med. sci. 63 (2008) 467-473. [24] t. maeda, j.z. guan, j. oyama, y. higuchi, n. makino, j. gerontol. a biol. sci. med. sci. 64 (2009) 949955. [25] g. hudson, d. faini, a. stutt, m. eccles, l. robinson, d.j. burn, p.f. chinnery, neurobiol aging. 32(11) (2011) 2107.e3-2107.5. [26] j. eerola, l. kananen, k. manninen, o. hellström, p.j. tienari, i. hovatta, j. gerontol. a biol. sci. med. sci. 65(11) (2010) 1181-1184. [27] c. zhang, x. chen, l. li, y. zhou, c. wang, s. hou, plos one (2015) e0133174. [28] e. hiyama, k. hiyama, t. yokoyama, y. matsuura, m.a. piatyszek, j.w. shay, nat. med. 1 (1995) 249255. [29] e. hiyama, t. yokoyama, n. tatsumoto, k. hiyama, y. imamura, y. murakami, t. kodama, m.a. piatyszek, j.w. shay, y. matsuura, cancer res. 55 (1995) 3258-3262. [30] e. hiyama, t. kodama, k. shinbara, t. iwao, m. itoh, k. hiyama, j.w. shay, y. matsuura, t. yokoyama, cancer res 57 (1997) 326-331. admet & dmpk 3(3) (2015) 254-259 telomere and telomerase in degenerative diseases doi: 10.5599/admet.3.3.207 259 [31] e. hiyama, k. hiyama, oncogene 21 (2002) 643-649. [32] j.w. shay, cancer j. sci. am. 4 (1998) 26-34. [33] n.w. kim, m.a. piatyszek, k.r. prowse, c.b. harley, m.d. west, p.l. ho, g.m. coviello, wright we, weinrich sl, shay jw. science, 266 (1994) 2011-2015. [34] c. chadeneau, k. hay, h.w. hirte, s. gallinger, s. bacchetti, cancer res. 55 (1995) 2533-2536. [35] h. tahara, t. nakanishi, m. kitamoto, r. nakashio, j.w. shay, e. tahara, g. kajiyama, t. ide, cancer res. 55 (1995) 2734-2736. [36] h. kojima, o. yokosuka, f. imazeki, h. saisho, m. omata, gastroenterology 112 (1997) 493-500. [37] w.e. wright, j.w. shay, m.a. piatyszek, nucleic acid res. 23 (1995) 3794-3795. [38] m. engelhardt, p. drullinsky, j. guillem, m.a. moore, clin. cancer res. 3(11) (1997) 1931-41. [39] a.a. avillion, m.a. piatyszek, j. gupta, j.w. shay, s. bacchetti, c.w. greider, cancer res. 56 (1996) 645-650. [40] h. kuniyasu, t. domen, t. hamamoto, h. yokozaki, w. yasui, h. tanara, jpn. j. cancer res. 88 (1997) 103-107. [41] t.a. augustine, m. baig, a. sood, t. budagov, g. atzmon, j.m. mariadason, s. aparo, r. maitra, s. goel, br. j. cancer 112(2) (2015) 313-318. [42] n. tatsumoto, e. hiyama, y. murakami, y. imamura, j.w. shay, y. matsuura, t. yokoyama, clin. cancer res. 6(7) (2000) 2696-701. ©2015 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license(http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ manuscript doi: 10.5599/admet.2.1.30 33 admet & dmpk 2(1) (2014) 33-42; doi: 10.5599/admet.2.1.30 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper anomalous solubility behavior of several acidic drugs alex avdeef in-adme research, 1732 first avenue, #102, new york, ny 10128, usa e-mail: alex@in-adme.com; tel.: +1 646 678 5713 received: january 27, 2014; revised: march 12, 2014; published: april 01, 2014 abstract the “anomalous solubility behavior at higher ph values” of several acidic drugs originally studied by higuchi et al. in 1953 [1], but hitherto not fully rationalized, has been re-analyzed using a novel solubilityph analysis computer program, pdisol-x tm . the program internally derives implicit solubility equations, given a set of proposed equilibria and constants (iteratively refined by weighted nonlinear regression), and does not require explicit henderson-hasselbalch equations. the re-analyzed original barbital, phenobarbital, oxytetracycline, and sulfathiazole solubility-ph data of higuchi et al. is consistent with the presence of dimers in saturated solutions. in the case of barbital, phenobarbital and sulfathiazole, anionic dimers, reaching peak concentrations near ph 8. however, oxytetracycline indicated a pronounced tendency to form a cationic dimer, peaking near ph 2. under the conditions of the original study, only barbital indicated a slight tendency to form a salt precipitate at ph > 6.8, with a highly unusual stoichiometry (consistent with a slope of 0.55 in the log s – ph plot): k + + a2h + 3ha  ka5h4(s). thus the “anomaly” in the higuchi data can be rationalized by invoking specific aggregated species. keywords sparingly-soluble drugs; ph-dependent solubility; salt solubility products; solubility equations; aggregation; shake-flask method. introduction aggregation and micelle formation of drugs in a saturated solution can complicate the interpretation of the solubility measurements, especially when the molecules are ionizable. there are many reports of such systems [1-6]. for example, doxycycline [3], with intrinsic solubility s0 = 0.72 mg/ml, at 25 o c in 1 m nacl/hcl solutions (ph 0 to 6), forms dimers with k2 = 24 m -1 . such aggregates were found to have unusually high and temperature-sensitive solubility. zhu and streng [4] found that the self-association of dolasetron to form cationic dimers and trimers was enthalpy driven (h-bond/aromatic ring rather than hydrophobic/electrostatic interactions), with the aggregation constants, k2-3, ranging from 4 to 50 m -1 at 25 o c. many nonsteroidal antiinflammatory drugs [5], such as indomethacin, diclofenac, ibuprofen, ketoprofen, naproxen, and sulindac, can self-associate by forming mixed-charge micelle or micelle-like structures. the reported cmc of diclofenac is about 25 mm and of ketoprofen is about 160 mm [5]. the solubility-ph profiles cannot be accurately described with the henderson-hasselbalch equation when aggregates form [2,6]. http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:alex@in-adme.com avdeef admet & dmpk 2(1) (2014) 33-42 34 the objective of this brief commentary is to address the “anomalous solubility behavior at higher ph values” of several ionizable drugs studied by higuchi et al. in 1953 [1], but hitherto not fully understood. the measured solubility-ph profiles of two barbiturates (phenobarbital, barbital) and two ampholytes (oxytetracycline and sulfathiazole) showed anomalous shapes, which could not be accurately described by the henderson-hasselbalch equations. analysis of the data using a new program, pdisol-x, suggested the presence of anionic (mixed-charge) dimers, aha , in the case of the barbiturates and sulfathiazole and very stable cationic (mixed-charge) dimers, xh2∙xh3 + , in the case of oxytetracycline. higuchi et al. noted that between ph 6.8 and 7.1, the slope in the log s vs. ph plot for barbital had a value of +0.55, but no concrete explanation was offered. they also noted the oxytetracycline between ph 2 and 5 had a slope of -0.91, which would not have been expected for a simple system. the values of the pka of the molecules were not reported in the study. had accurate pka values been discussed, then other anomalies in the solubility data would have been evident at that time. experimental legacy data in the higuchi et al. study [1], ph was adjusted by the addition of standardized hcl or koh to buffer-free suspensions containing 10-30 % excess solid over the saturation concentration. the ionic strength was well controlled with 0.15 m kcl, and the temperature was maintained at 25 °c (30 °c in one case). the equilibration times ranged from one to eight days, after which the final ph had been measured. refinement of intrinsic and salt solubility and aggregation constants detail of the mathematical approach in the pdisol-x computer program has been described by völgyi et al. [6]. briefly, the new data analysis method uses log s ph as measured input data from any analytical technique (along with standard deviations, sd(log s)) into pdisol-x. an algorithm was developed which considers the contribution of all species proposed to be present in solution, including universal buffer components (e.g., britton-robinson). the approach does not depend on any explicitly derived extensions of the henderson-hasselbalch equations. the computational algorithm derives its own implicit equations internally, given any practical number of equilibria and estimated constants, which are subsequently refined by weighted nonlinear least-squares regression [2,6]. therefore, in principal, drug-salt precipitates, -aggregates, -complexes, bile salts, -surfactant can be accommodated [2,6]. presence of specific bufferdrug species can be tested. the program assumes an initial condition of a suspension of the solid drug in a solution, ideally with the suspension saturated over a wide range of ph. the program calculates the distribution of species consequent to a sequence of additions of standardized strong-acid titrant hcl (or weak-acid titrants h3po4, h2so4, acetic acid, maleic acid, lactic acid) to simulate the suspension phspeciation down to ph ~ 0, the staging point for the subsequent operation. a sequence of perturbations with standardized koh is simulated, and solubility calculated at each point (in ph steps of 0.005-0.2), until ph ~ 13 is reached. the ionic strength is rigorously calculated at each step, and pka values (as well as solubility products, aggregation and complexation constants) are accordingly adjusted [6]. at the end of the ph-speciation simulation, the calculated log s vs. ph curve is compared to measured log s vs. ph. a log s-weighted nonlinear least squares refinement commences to refine the proposed equilibrium model, using analytical expressions for the differential equations. the process is repeated until the differences between calculated and measured log s values reach a minimum, as described elsewhere in detail [2,6]. admet & dmpk 2(1) (2014) 33-42 anomalous solubility of weak acids doi: 10.5599/admet.2.1.30 35 results and discussion the results of the analysis of the higuchi et al. data are summarized in table 1. all four drugs in the higuchi study indicated the presence of singly-charged dimers in saturated solution of the free acid. a salt precipitate also formed in the case of barbital. table 1. summary of the results of the re-analysis of the higuchi et al. [1] data. compound pka (25 o c, i =0.15 m) a -log s0 s0 (mg/ml) mixed-charge dimer log k2 (m -1 ) k2 (m -1 ) gof b barbital c 7.86 1.57 ± 0.01 5.00 ha+a  a2h 2.17 ± 0.03 148 ± 10 0.12 phenobarbital 7.41 2.55 ± 0.01 0.66 ha+a  a2h 2.43 ± 0.04 269 ± 22 0.22 oxytetracycline 8.82, 7.22, 3.23 3.20 ± 0.03 0.29 xh3 + +xh2  x2h5 + 3.42 ± 0.07 2634 ± 425 0.75 sulfathiazole 7.12, 2.30 2.83 ± 0.01 0.37 hx+x  x2h 1.50 ± 0.10 32 ± 7 0.12 a pka values taken from the "wiki-pka" database: <http://www.in-adme.com/wiki_pka.php/>. b gof = goodness-of-fit, based on assumed individual measurement errors of 0.1 log units. c 30 o c. the other three molecules were studied at 25 o c. barbital the solid curve in figure 1a is the log s vs. ph profile of barbital at 30 °c. the dashed curve in the figure was calculated using the henderson-hasselbalch equation, employing pka of 7.86. as can be seen, the latter equation does not predict the experimental points very well. figure 1b shows the simulated titration curve (volume of 1 m koh added per ml of suspension solution vs. ph) for barbital that best describes the higuchi data (discrete points). the amount of solid barbital added to 1 ml of water was taken as 250 mg in the pdisol-x data analysis. the flat region centered at ph 6.82, bounded by two discontinuities, is due to the co-precipitates of free acid, ha(s), and complex salt, ka5h4(s), i.e., [k + (ha)3 aha ](s). this fixed-ph region is defined by the so called pka gibbs [7,8]. the free acid is the only precipitate present for ph < pka gibbs . it takes about 0.3 ml koh titrant to convert the free acid precipitate entirely into the salt form, as shown in figure 1b. for ph > pka gibbs , ka5h4(s) is the sole precipitate, up to the peak ph 9.56, where the salt precipitate completely dissolves due to ionization leading to the formation of the anion, a (fig. 1d). note that the width of the flat region of the titration curve (fig. 1b), and hence the values of the pka gibbs and peak ph are conditional, since the extent of dimer formation depends on the amount of solid added to water. under the conditions of the study, only barbital appeared to show a tendency to form a salt precipitate, consistent with the very unusual stoichiometry (slope of +0.55 in the log s-ph plot): ka5h4(s)  k + + aha + 3ha, with ksp = 2.23 x10 -8 m 5 . the above salt stoichiometry best rationalizes the log s – ph curve of barbital for ph > pka gibbs . the solubility curve (solid line) for ph < pka gibbs (6.82) is horizontally displaced to lower ph by about one unit, without change in its shape, compared to that calculated using the henderson-hasselbalch equation (dashed line in fig. 1a). this is consistent with the formation of a mixed-charge dimeric species, aha (in the saturated solution), which may be visualized as a conjugate of a free acid, ha, and the ionize species, a (cf., appendix). the aha dominates in concentration over that of a up to ph 9 (fig. 1c). for ph > pka gibbs , the log s curve takes on a slope of about +0.55, as originally noted by higuchi et al. this can be explained if a new precipitate replaces ha(s). a reasonable choice might have been k + ∙aha (s), since aha is the dominant anion present in solution. however, the log s curve is not well fitted with such a model. further testing of the data could have assumed that the above salt precipitate also contained ha components. avdeef admet & dmpk 2(1) (2014) 33-42 36 after a trial-and-error process, the best fit was found where three molecules of the free acid, ha, and one molecule of aha combine with k + to form the putative solid with the unusual stoichiometry k + (ha)3aha (s). pdisol-x can easily consider new tentative species in the model development and quickly test the data against such species, without having to derive explicit new solubility equations. the method described in the appendix a in [6] certainly could have been applied here to derive the explicit equation corresponding to the formation of k + (ha)3 aha (s), but the resultant equation would have been quite complex and of limited general applicability. figure 1c shows the direct concentration of various barbital species as a function of ph, while figure 1d illustrates the log concentration vs ph relationships. in alkaline solution, the monomeric species, a , and the mixed-charge dimeric species, aha coexist, with the latter species reaching a maximum concentration at the peak ph 9.56 (fig. 1c) and declining in concentration with further increases in ph. under the conditions of the higuchi data, of the four compounds studied, only barbital indicated a the above case study of barbital is an example of a general methodology to rationalize complicated log s ph profiles. in a simple “henderson-hasselbalch” precipitation process, the log s curve does not depend on the quantity of solid added, as long as enough solid is added to maintain a saturated solution. however, if an aggregate, oligomer, or micelle-like drug forms, then the measured solubility can critically depend on the amount of solid added to solution. for example, suppose a micelle of the drug forms above the critical micelle concentration (cmc). a very poorly soluble drug can become better (self-) solubilized in high concentration (>cmc). for example, this could explain why a parenteral formulation of amiodarone (high concentration) can stop atrial fibrillation in a few minutes post injection, whereas an oral formulation may take 30 days to reach full therapeutic effectiveness [9]. being able to quantitate such concentration dependence (due to aggregate formation, etc.) could be a practically useful feature of pdisol-x in drug figure 1. barbital: (a) log s vs. ph; (b) titration curve; (c) concentration of species as a function of ph; (d) log concentration of species as a function of ph. admet & dmpk 2(1) (2014) 33-42 anomalous solubility of weak acids doi: 10.5599/admet.2.1.30 37 development, to help predict dosing effects of poorly soluble molecules which form aggregates in solution. phenobarbital phenobarbital, which is about eight times less soluble than barbital, shows a similar solubility-ph profile in figure 2a as that of barbital, except that no salt precipitate forms (the ratio of the octanol-water partition coefficients of the two molecules is also about eight). a dimeric species of the same stoichiometry forms as in the case of barbital, reaching maximum concentration at (conditional) ph of 7.82, as shown in figure 2b (the ph-speciation simulation assumed the 5 mg of solid had been added to 1 ml of water). the presence of the dimer is indicated by the shift of the log s – ph profile to the lower ph values above ph 5, compared to that predicted by the henderson-hasselbalch equation. oxytetracycline oxytetracycline is an ampholyte, whose ionizations can be described by three pka values (table 1). figure 3a shows the solubility-ph curve shifted to higher ph values below ph 5, compared to what is predicted by the henderson-hasselbalch equation. this is a pattern consistent with the presence of a cationic dimer formed from the neutral xh2 species and the cation xh3 + . many other shift patterns have been described in reference [2]. the cationic dimer maximized in concentration at (conditional) ph 1.63, as shown in figure 3b (the ph-speciation simulation assumed that 50 mg of solid had been added to 1 ml of water). figure 2. phenobarbital: (a) log s vs. ph; (b) concentration of species as a function of ph. avdeef admet & dmpk 2(1) (2014) 33-42 38 sulfathiazole the ampholyte sulfathiazole has the least tendency to form dimers, as suggested in figure 4. a small quantity of anionic dimer appears to be consistent with the high precision data, whose concentration reaches a maximum at ph 8.16. invoking the presence of a dimeric species under the circumstance would not be warranted ordinarily, were it not for the exceptionally high precision of the higuchi et al. data. conclusions in the pdisol-x analysis, the “anomaly” in the higuchi et al. 1953 data could be explained by presence of anionic dimers in three cases, and by the strong presence of a cationic dimer in the case of oxytetracycline, and also by the barbital salt precipitate that exhibits the slope of +0.55. apparently, these species were either not recognized originally, or were not fully characterized. such anomalies can be common with sparingly-soluble drugs, but are not always easy to recognize. pdisol-x could be a helpful new tool to further aid in the analysis of dissolution mechanisms of sparingly-soluble drugs, since dissolution is closely linked to solubility. figure 3. oxytetracycline: (a) log s vs. ph; (b) concentration of species as a function of ph. figure 4. sulfathiazole: (a) log s vs. ph; (b) concentration of species as a function of ph. admet & dmpk 2(1) (2014) 33-42 anomalous solubility of weak acids doi: 10.5599/admet.2.1.30 39 acknowledgements: helpful discussions with gergely völgyi and krisztina takács-novák of semmelweis university (budapest) are greatly appreciated. references [1] t. higuchi, m. gupta, l.w. busse. j. am. pharm. assoc. 42 (1953) 157-161. [2] a. avdeef. absorption and drug development second edition, wiley-interscience, hoboken, nj, 2012, pp. 251-318. [3] j.b. bogardus, r.k. blackwood, jr. j. pharm. sci. 68 (1979) 188-194. [4] c. zhu, w.h. streng. int. j. pharm. 130 (1996) 159-168. [5] a. fini, g. fazio, g. feroci. int. j. pharm. 126 (1995) 95-102. [6] g. völgyi, a. marosi, k. takács-novák, a. avdeef. admet & dmpk 1(4) (2013) 48-62. [7] w.h. streng. int. j. pharm. 186 (1999) 137–140. [8] a. avdeef. pharm. pharmacol. commun. 4 (1998) 165-178. [9] amiodarone-containing parenteral solution patent: ep 1 267 865 b1 (15 feb 2006). avdeef admet & dmpk 2(1) (2014) 33-42 40 appendix shape pattern interpretation in solubility profiles for sparingly-soluble ionizable weak acids an important part of the analysis of solubility profiles involves recognizing characteristic shapes and interpreting these in terms of an equilibrium model [2]. figure 5 shows the results of a series of simulations of a weak acid, mw 500 g/mol, having the same pka = 7 and ps0 = 3, but having different combinations of salt and aggregate formations. in the first frame, 50 mg of solid free acid were added to 1 ml of 0.15 m kcl solution. in the rest of the frames, 1500 mg were added. the calculated average ionic strength is indicated in red at the bottom right of each frame. the dashed curves in the figures are calculated from the henderson-hasselbalch equation using the pka and ps0 constants. the solid lines are the actual log s curves, where either the free acid or the salt is precipitated. the ph at which both the free acid and the salt coprecipitate is the pka gibbs (figs. 5b-d). in cases where ph is raised past the point where the entire solid dissolves, the curve is indicated by a dotted line (figs. 5a,e,f). figure 5a is the simplest case, with no salt precipitate and no aggregates. the henderson-hasselbalch equation describes the log s – ph profile as long as the solution remains saturated. figures 5b-d show the shape of the profiles as k + a (s) or k + aha (s) salts precipitate for ph ≥ pka gibbs . notice that in figure 5c, the maximum solubility does not occur at ph = pka gibbs . in older literature, it had been popular to refer to pka gibbs as “phmax”, but clearly, such nomenclature is misleading in some cases, such as when complex salts form, as in figure 5c [7,8]. in figures 5d-f cases, aggregates form, which elevates solubility in characteristic ways. in the case of neutral aggregates (fig. 5d), the region of the intrinsic solubility (ph < pka) is raised due to the enhanced solubility of the neutral species. the apparent pka is shifted to 8.4 from the true value of 7. in the cases of charged aggregates (fig. 5e,f), the solubility product for k + a (s) is no longer exceeded (for 1500 mg/ml), and the compound completely dissolves above about ph 8, due to the increased solubility in alkaline solution. figure 5e is an example of a mixed-charge dimer formation, aha , where the solubility figure 5. simulations of various cases of free acid log s – ph curves where salt precipitates or where aggregates form (see text). admet & dmpk 2(1) (2014) 33-42 anomalous solubility of weak acids doi: 10.5599/admet.2.1.30 41 curve in the slope = 1 region shifts to lower ph values compared to that predicted by the hendersonhasselbalch equation. in figure 5f, an anionic trimer, a3 3, is indicated by the slope = 3 in the diagonal region of the log s profile. example of a step by step model construction and refinement for ionizable sparingly-soluble weak acid the following is a brief description of a practical process for developing and refining an equilibrium model to describe the solubility behavior of sparingly soluble drugs. 1. before starting to develop an equilibrium model for an ionizable molecule, it is first necessary to know the accurate value of the pka. preferably, the pka is determined by an independent technique (e.g., by potentiometry, spectrophotometry, or capillary electrophoresis – but not by solubility). 2. plot the log s data as a function of ph. with an accurate pka value at hand, the log s profile is inspected for its shape characteristics, to see if the profile resembles any of the template frames in figure 5. 3. if the pka appears to be higher from the true value (as indicated by the ph in the middle of the bend in the log s curve), then either neutral aggregates form (fig. 5d) or not enough time had been allowed for equilibrium to be reached. the ps0 can be estimated (for acids) as ps0 app + (pka app – pka) [2]. select h + +a  ha, ha  ha(s), and 2ha  a2h2 as the three equilibria to consider in the model, and associate pka and ps0 with the first two equations. take a guess on the third constant, e.g., assume it to be 3 (log units). refinement will usually hone in on the appropriate value quickly enough if the equilibrium model is valid. 4. if the pka appears to be lower than the true value, but the slope in the diagonal portion of the log s profile is 1, then a mixed-charge dimer (or higher order oligomer) may be present (cf., fig 5e). the ps0 can be estimated as solubility in the acidic region of the curve (ph << pka). select h + +a  ha, ha  ha(s), and ha + a  aha as the three equilibria to consider in the model, and associate pka and ps0 with the first two equations. take a guess on the third constant, e.g., assume it to be 3 (log units). 5. if the pka appears to be slightly lower than the true value, but the slope in the diagonal portion of the log s profile is greater than 1, then an anionic oligomer may be present, with the degree of aggregation equal to the value of the slope (cf., fig 5f). the ps0 can be estimated as solubility in the acidic region of the curve (ph << pka). select h + +a  ha, ha  ha(s), and na  an n as the three equilibria to consider in the model, and associate pka and ps0 with the first two equations. take a guess on the third constant, e.g., assume it to be 3 (log units). 6. it is possible that more than one mechanism is associated with the solubility profile. therefore, a combination of the above steps may be needed. 7. if the curve levels off in the alkaline solution (ph >> pka), as in figure 5, then a salt may have precipitated in alkaline solution. it needs to be ascertained that the effect is not due to the complete dissolution of solid, as in figures 5a,e,f. enter the appropriate equilibrium expression along with a guess of the value of the constant. 8. having initial estimates of constants corresponding to proposed equilibrium reactions, the model is ready for weighted nonlinear regression, a procedure described elsewhere [2,6]. the iterative refinement will test the consistency between the data and the model. if the agreement is not good avdeef admet & dmpk 2(1) (2014) 33-42 42 (root-mean square error >> 0.1), then the model may be modified and the refinement process repeated, until no further improvements are evident. ©2014 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ manuscript doi: 10.5599/admet.5.2.388 126 admet & dmpk 5(2) (2017) 126-134; doi: 10.5599/admet.5.2.388 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper cellular energy status is indispensable for perillyl alcohol mediated abrogated membrane transport in candida albicans moiz a. ansari, zeeshan fatima* and saif hameed* amity institute of biotechnology, amity university haryana, gurgaon (manesar)-122413, india. *corresponding authors: dr. saif hameed and dr. zeeshan fatima, amity institute of biotechnology, amity university haryana, gurgaon (manesar)-122413, india. email: saifhameed@yahoo.co.in, drzeeshanfatima@gmail.com ; tel.: +91124-2337015, ext: 1116. received: april 21, 2017; revised: may 23, 2017; published: june 22, 2017 abstract the prevalence of fungal infections and their resistance patterns in fungal isolates fro m large number of patients with impaired immunity still remains poorly monitored. in spite of significant advances being made in the improvement of antifungal drugs, only a limited number of antifungal drugs are currently available. the present study aimed to gain further mechanistic insights into the previously described anticandidal activity of natural monoterpenoid, perillyl alcohol (pa). we found that cellular transport across cell membrane was abrogated in presence of pa. this was demonstrated by dose and time dependent enhanced cellular leakage accompanied by inhibited sodium and potassium cellular transport. in addition, we found disrupted ph homeostasis which was depicted by enhanced extracellular ph. we further observed that mitochondrial energy status is highly integrated with the antifungal activity of pa. this was evident from inhibited propidium iodide (pi) uptake in presence of sodium azide and di-nitro phenol (dnp) which showed no fluorescence when treated with pa. moreover, we observed that pa leads to disrupted mitochondrial membrane potential. additional cell death hallmarks in response to pa such as nuclear fragmentation was also observed with 4',6-diamidino-2-phenylindole (dapi) staining. taken together, pa is a novel candidate that deserves further attention to be exploited as effective antifungal agent of pharmacological interest. keywords perillyl alcohol; sodium transport; potassium transport; ph homeostasis; mitochondria; atp; dna damage introduction candida albicans being an opportunistic human fungal pathogen resides commensally within the host. however, under immune-compromised conditions such as organ transplantation, chemotherapy and diabetes, it turns up into a deadly pathogen [1]. according to the recent finding, around 1.5 million people died per annum due to invasive fungal infections [2] and among all fungal pathogens candida species leads the toll for the cause of the mortality related to the mycotic death throughout the world [3]. with the limited repertoire of current therapeutics which includes antifungal drugs from three major classes; azoles, polyenes and echinocandins, the utility of these drugs are restricted due to high cost and sometimes host toxicity. moreover, emergence of drug resistant strains due to the phenomenon of multidrug resistance http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:saifhameed@yahoo.co.in mailto:drzeeshanfatima@gmail.com admet & dmpk 5(2) (2017) 126-134 perillyl alcohol mediated abrogated membrane transport in c. albicans doi: 10.5599/admet.5.2.238 127 (mdr) has compounded the problem and became major obstacle against effective therapeutics [4]. therefore, the flux towards phytochemical research to circumvent the problem of mdr in need of new antifungals is in serious vigilance nowadays. we previously demonstrated that perillyl alcohol (pa) is a natural benzylic monoterpenoid (fig. 1), similar to its class terpenoid [5], it also showed immense antifungal potential against human pathogen c. albicans as well as non-albicans species of candida [6]. however, pa is immensely studied for its anticancerous activities and recently reported as suitable for brain tumor therapy [7-9]. moreover, pa is fda approved food additive that can be safely consumed without any toxic effect. the current study enhances the knowledge of mechanistic action of pa induced membrane disruption against c. albicans leading to the altered cationic concentration and demonstrates that cellular energy status is indispensable for functional anticandidal activity of pa. additionally, we explored that pa leads to disrupted mitochondrial membrane potential and nuclear fragmentation. oh ch2h3c figure 1. chemical structure of pa experimental materials yeast extract, di-nitro phenol (dnp) and peptone were purchased from himedia, india. d-glucose, dimethyl sulfoxide (dmso), rhodamine b, nan3 and kcl were purchased from thermofisher. fluconazole (flc) dapi and pa were purchased from sigma. cellular leakage analysis of loss of cellular material absorbing at 260 nm from candida cell was performed using the method described elsewhere with little modifications [10]. briefly, candida cells were grown in ypd at 30°c for 48 h and cell suspension (od600 as 1) was prepared in 10 ml pb. furthermore, this suspension was treated with different concentrations (1 ×, 2 ×, and 4 × mic) of pa and flc for 1 h or at fixed concentration (1× mic) for different time intervals (30, 60, 90 and 120 min). untreated sample was run as negative control. after treatment, samples were centrifuged at 10,000 rpm for 10 min at 4 ᵒc, and the absorbances of supernatants were read at 260 nm using uv–vis spectrophotometer (vsi-501, india). intracellular na + concentration intracellular sodium ion concentration was obtained as described elsewhere with slight modification [11]. briefly, cells were pre-grown at 30 ᵒc in ynb medium overnight. the cells were inoculated into 50 ml of ynb medium (with and without drug) to obtain od600 as 0.05 and grown at 30 ᵒc to obtain od600 between 0.17–0.3; and the intracellular sodium content was measured. to estimate the intracellular na + ansari, fatima and hameed admet & dmpk 5(2) (2017) 126-134 128 concentration, cell pellets, after harvesting and washing with water, were suspended in 18 ml of incubation buffer (20 mmol/l mes and 0.1 mmol/l mgcl2, ph 5.5 adjusted by solid ca(oh)2). cells were washed twice with washing buffer (20 mmol/l mgcl2) and lysed in 5 ml of extraction buffer (10 mmol/l mgcl2 and 0.2 mol/l hcl). flame photometer (systronics, india) was used to determine the sodium content in the extracts. k + leakage estimation estimation of extracellular k + leakage from the fungal cell was performed using the method described elsewhere with little modifications [11]. briefly, candida cells were grown in 100 ml ypd at 30 °c, 120 rpm for 24 h. cell pellets were harvested at 5000 rpm for 5 min and washed three times with sterile phosphate buffer (pb) and suspended in 10 ml of pb. 500 μl of the cell suspension was used to determine dry cell weight. 2 ml of cell suspension was mixed with 2 ml of pb containing desired concentration (1 ×, 2 × and 4 × mic) of test agents and incubated for 4 hr. after treatment samples were centrifuged at 10,000 rpm for 10 min, and supernatant was removed and stored for the determination of extracellular k + content released into the medium. the k + concentration was estimated by flame photometer (systronics, india) using potassium filter. pm-atpase mediated proton pumping the proton pumping activity of c. albicans was estimated by monitoring the glucose-induced acidification of the external medium due to ph changes as previously reported [12]. briefly, overnight cultures of c. albicans were grown in yepd broth for 18 h at 30 °c. the cells were collected by centrifugation at 3000 g for 5 min at 4 °c and washed with sterile distilled water and 50 mm kcl (ph 6.5). the washed cells were resuspended in 40 ml of 50 mm kcl (ph 6.5) and incubated at 4°c overnight to deplete their carbon reserves. the carbon-starved cells were harvested by centrifugation, and approximately 1.0 g wet weight of the pellet was resuspended in 40 ml of 50 mm kcl (ph 6.5). to a 40 ml aliquot of the cell suspension, pa at mic80 was added to obtain the required concentration and mixed well, and the volume was adjusted to 45 ml with 50 mm kcl. the cell suspension was incubated at room temperature with gentle stirring for 10 min, and then 5 ml of 20 % glucose (final concentration, 55 mm) was added and the ph of the external medium was monitored at regular intervals for 40 min at indicated time points. the experiment was performed in the presence of a comparable concentration of the solvent dmso (control) to measure the extent of acidification of the external medium in the absence of pa. propidium iodide uptake propidium iodide, is a dye impermeable to membrane and is widely used to differentiate cells that have damaged plasma membranes from healthy cells [13]. to evaluate the effect of pa on the fungal plasma membrane, c. albicans sc5314 cells (approximately 1×103 cfus/ml) were obtained from exponential phase and exposed to pa (225 µg/ml) for 3 h at 30 °c with gentle shaking. subsequently, cells were harvested, incubated with propidium iodide for 15 min, and observed by coslab fluorescence microscope. mitochondrial membrane potential (mtψm) rhodamine b dye (molecular probes) was used to detect the change of mitochondrial membrane potential [14]. c. albicans (sc5314) cells were treated with pa (225 µg/ml) overnight at 30 °c. the treated cells were washed in pbs and then suspended in 40 nm rhodamine b for 20 min at 37 °c. the cells were centrifuged at 500 g for 5 min and the pellet was re-suspended in 1 ml pbs. the cells were analyzed using coslab fluorescence microscope and fluorescent intensity was measured through image j software. all admet & dmpk 5(2) (2017) 126-134 perillyl alcohol mediated abrogated membrane transport in c. albicans doi: 10.5599/admet.5.2.238 129 observations were confirmed with 3 independent cell cultures. dapi staining cells were stained with dapi to detect nuclear fragmentation as discussed elsewhere [15]. the strains were grown in yepd overnight in presence of pa at 30 °c. fragmented nuclei were identified by staining with 0.1µg/ml dapi for 15 min. the stained cells were collected then washed twice with 1 % bsa in pbs for 5 min followed by a five minute rinse in 0.1 % bsa in pbs. cells were visualized with coslab fluorescence microscope. all observations were confirmed with 3 independent cell cultures. results and discussion membrane disruption by pa leads to cellular leakage in c. albicans in our previous study we have determined that anticandidal effect of pa is mediated through membrane perturbation [6]. in this study we have explored the severity of this damage in terms of transport across cell membrane. for this, cellular leakage across membrane was estimated by the leakage of intracellular components. a cellular component that absorbs light of 260 nm wavelength signifies for a particular class of out-flowing components, mainly nucleotides which include uracil compounds revealing strongest absorbance [10]. in the pa treated cells similar to flc, the absorbance of intracellular content was observed to be increased for pa in dose dependent manner (figure 2a). we further confirmed the cellular leakage by studying time kinetics at mic concentration of pa. we observed the elevation in absorbance from the cells treated with pa and flc with respect to time till 120 min (figure 2b). this proves that the pa treatment enhances the leakage of cellular contents through membrane disruption. figure 2. effect of pa and flc on 260nm absorbing material in c. albicans: (a) concentration dependent, (b) time dependent. pa affects na + transport and ph homeostasis in c. albicans enhanced cellular leakage in presence of pa prompted us to study the transport of ions across the membrane more closely. to verify that the leakage of cellular content due to pa treatment is also affecting ansari, fatima and hameed admet & dmpk 5(2) (2017) 126-134 130 cationic concentration, the intracellular na + concentration was evaluated. it was observed that the na + concentration was significantly decreased in the pa treated cells which suggests that pa induced membrane disruption possibly affects na + transport across the membrane (figure 3a). the two main transporters in yeast that are accountable for alkali-metal-cation permeability across the membrane are na + /h + antiporter and a na + -atpase. both transporters harmonize each other to support detoxification. the na + /h + antiporter is active during low extracellular ph however na + atpase is highly active at neutral and higher extracellular ph [16]. thus we checked that if the decrease in na + concentration is linked with some changes in extracellular ph. it was seen that the pa treatment leads to the significant increment in the extracellular ph in the candida cells (figure 3b). while convalescing themselves during starvation, candida cells pump intracellular protons outside the cell through plasma membrane atpase, utilizing available number of atp after the addition of glucose to non-growing nutrient deficient cells. thus, proton pumping action of the atpase could be the reason for acidification of the medium. this confirms that pa treated candida cells lost the ability to pump intracellular protons to the external medium may be due to inhibition in plasma membrane atpase thereby elevating the ph of the external medium and diminishing the intracellular na + concentration. 0 20 40 60 80 100 120 140 control nacl control pa nacl pa n a + in (n m /m g d ry m a ss ) * * 0 1 2 3 4 5 6 0 10 20 30 40 50 60 p h time (min) control pa a b * * * * * * figure 3. effect of pa on membrane permeability in c. albicans: (a) effect of pa on intracellular na + concentration (b) effect of pa on the glucose-dependent acidification of the medium by c. albicans cells. mean of ph ± sd of three independent sets of experiments are shown in absence (control) and presence of pa with respect to time (minutes) and *p value <0.05 is considered to be significant. pa affects k + transport in c. albicans as a known matter of fact, na + transport across cellular membrane is linked with k + transport. moreover, the concentration both of na + and k + plays a critical role to maintain the membrane potential in the cells [17]. in fact dysregulation of ion homeostasis forms the basis of cell death for many terpenoids. admet & dmpk 5(2) (2017) 126-134 perillyl alcohol mediated abrogated membrane transport in c. albicans doi: 10.5599/admet.5.2.238 131 thus we checked whether there are also some changes in k + transport in presence of pa. this was evident from the elevated level of intracellular and diminished levels of extracellular k + concentration respectively in pa treated cells (figure 4). this result suggests that the na + k + antiporters may be affected in presence of pa. from these observations, one may also speculate that enhanced membrane permeability due to pa could cause dissipation of ionic gradient across cell membrane leading to membrane depolarization, however, further work is needed to confirm. 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 control pa-mic pa 2xmic pa 4xmic k + c o n c (n m /m g d ry m a ss ) internal external figure 4. effect of pa on intracellular and extracellular k + concentration in c. albicans membrane perturbations due to pa is energy dependent sodium azide and dnp is known for atp reducing property by decreasing the amount of atp in candida cells targeting cytochrome oxidase of mitochondria and binds with the heme a3 iron and cub in the reduction site of oxygen [18,19]. f1 domain is also an important component of mitochondria for its atp generation ability. it is the assembly of alternatively arranged three α and three β subunits around the central stalk, which is made of γ, δ and ε subunits. azide binds to the f1 catalytic domain and inhibits atp hydrolysis [20]. to study whether the membrane disruptive action of pa is energy dependent, an assay was performed through pi uptake of sodium azide treated cells in the presence of pa (figure 5). we observed that in cells that were not treated by sodium azide or dnp, pi uptake was efficient and showed fluorescence when treated with pa. however in the sodium azide or dnp treated cells this effect of pa was reverted and no cell was stained with pi. these observations not only confirmed non-intact membrane confirming our previous hypothesis of membrane disruption [6] but also suggests that the intracellular localization and the membrane damaging effect of pa is energy dependent. pa alters mitochondrial membrane potential (mtψm) in c. albicans in our previous study we documented that pa inhibits the mitochondrial enzyme functioning [6]. moreover, mtψm is a sensitive indicator of the energy status of the mitochondria which is the result of an electrochemical gradient maintained through the electron transport chain and also plays key role in atp synthesis which is derived from mitochondrial oxidative phosphorylation [21]. hence, we study the effect of pa on mtψm by employing rhodamine b probe as described in material and methods. rhodamine b spreads throughout the biological membranes in response to the ψm. the change in mitochondrial potential (∆ψm) directly correlates with the increase or decrease in atp content of the cells, where higher the ∆ψm more will be the cellular atp level [14]. in the intact cells, the fluorescence of rhodamine b tends to decrease because of the formazan, quencher of rhodamine b which naturally accumulates in mitochondrial matrix, hence ∆ψm will be more. however in mitochondria disrupted cells the ∆ψm will be ansari, fatima and hameed admet & dmpk 5(2) (2017) 126-134 132 less due to no quenching in fluorescence depicted by red signals [14,22]. our result clearly demonstrates that pa treated candida cells showed more fluorescence (figure 6) which suggests that pa alters the mtψm in c. albicans. figure 5. effect of energy depletion on anticandidal activity of pa. fluorescent microscopy of pi in presence of 5mm nan3 and dnp. figure 6. effect of pa on mtψm: a) fluorescent microscopy of rhodamine b for analysis of mtψm in presence of pa b) fluorescent intensity of untreated (control) and pa treated cells and *p value <0.05 is considered to be significant. pa induces nuclear fragmentation in c. albicans mitochondria dysfunctioning is one of the major causes that lead to the dna damage [15]. moreover, the leakage of intracellular content primarily nucleotides from this study and hypersensitivity towards pa in presence of dna damaging agent from previous observation was enough evidence to study nuclear fragmentation [6]. hence we checked the nuclear fragmentation using dapi staining as described in material and methods. dapi is a fluorescent dye that binds to the at site between the minor groove of dna. the fluorescence can be evaluated with irregular margin and condensed chromatin [15]. we found that the cells treated with pa showed nuclear fragmentation which could be linked with dna damage (figure 7). since damaged dna cannot replicate owing to prevention of mutations, cell cycle is arrested [23]. this observation is also commensurate with our earlier findings which showed pa induced down admet & dmpk 5(2) (2017) 126-134 perillyl alcohol mediated abrogated membrane transport in c. albicans doi: 10.5599/admet.5.2.238 133 regulation of several dna repair genes and cell cycle arrest [6]. c o n tr o l p a bright fluorescence merged 10µm 10µm 10µm 5µm 5µm 5µm figure 7. effect of pa on nuclear fragmentation. fluorescent microscopy of dapi for detection of nuclear fragmentation in presence of pa. conclusion taken together, the antifungal potential of pa not only provides an opportunity to aid pa develop into pharmacologically acceptable and efficient antifungal but also helps in deciphering novel regulatory circuits governing drug resistance in c. albicans. this is evident by wide range of antifungal targets of pa in comparison to single mode of action of known antifungals such as azoles. further studies are warranted to competently employ phytotherapeutics such as pa in treating candida infections. acknowledgements s.h. thank for the financial assistance in the form of young scientist award (sr/ft/ls-12/2012) from science and engineering research board (serb), new delhi. we are grateful to joseph heitman for providing candida sc5314 reference strain as generous gift. we thank rajendra prasad, dean, faculty of science, engineering and technology for encouragement. references [1] s. singh, z. fatima, s. hameed, infez. med. 23(3) (2015) 211-23. [2] g.d. brown, d.w. denning, n.a. gow, s.m. levitz, m.g. netea, t.c. white, sci. transl. med. 4(165) (2012) 165rv113. [3] m.a. pfaller, d.j. diekema, crit. rev. microbiol. 36 (2010) 1-53. [4] j. tanwar, s. das, z. fatima, s. hameed, interdiscip. perspect. infect. dis. 2014 (2014) 541340. [5] g.b. zore, a.d. thakre, s. jadhav, s.m. karuppayil, phytomedicine 18(13) (2011) 1181-90. [6] m.a. ansari, z. fatima, s. hameed, plos one 11(9) (2016) e0162465. [7] t. c. chen, c.o. da fonseca, ah. schönthal, int. j. mol. sci. 17(9) (2016). [8] c.e. oturanel, i̇. kıran, ö. özşen, g.a. çiftçi, ö. atlı, anticancer agents med. chem. 17 (2017) doi: 10.2174/1871520617666170103093923. [9] a.c. gomes, a.l. mello, m.g. ribeiro, et al. arch. immunol. ther. exp. (warsz) (2017). [10] m.s. khan, i. ahmad, s.s. cameotra, amb express 3(1) (2013) 54. ansari, fatima and hameed admet & dmpk 5(2) (2017) 126-134 134 [11] h. elicharova, h. sychrova, med. mycol. 51(8) (2013) 785-94. [12] s. singh, z. fatima, s. hameed, arch. microbiol. 198(5) (2016) 459-72. [13] s.s. wong, r.y. kao, k.y. yuen, et al. plos one 9(1) (2014) e85836. [14] p. reungpatthanaphong, s. dechsupa, j. meesungnoen, c. loetchutinat, s. mankhetkorn, j. biochem. biophys. methods 57(1) (2003) 1-16. [15] b. hao, s. cheng, c.j. clancy, m.h. nguyen, antimicrob. agents chemother. 57(1) (2013) 326-32. [16] y. krauke, h. sychrova, curr. microbiol. 61 (2010) 335–339. [17] m.s. cyert, c.c. philpott, genetics 193(3) (2013) 677-713. [18] j. han, m.a. jyoti, h.y. song, w.s. jang, plos one 11(2) (2016) e0150196. [19] s. yoshikawa, k. shinzawa-itoh, r. nakashima, et al. science 280(5370) (1998) 1723-9. [20] m.w. bowler, m.g. montgomery, a.g. leslie, j.e. walker, proc. natl. acad. sci. usa 103(23) (2006) 8646-9. [21] s.w. perry, j.p. norman, j. barbieri, e.b. brown, h.a. gelbard, biotechniques 50(2) (2011) 98-115. [22] e. gottlieb, s.m. armour, m.h. harris, c.b. thompson, cell death differ. 10(6) (2003) 709-17. [23] h.l. lin, t.y. liu, c.w. wu, c.w. chi, cancer 92(3) (2001) 500-9. ©2017 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ doi: 10.5599/admet.3.3.196 203 admet & dmpk 3(3) (2015) 203-215; doi: 10.5599/admet.3.3.196 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index review the roles of genetics in aβ related alzheimer's diseases yin tian*, zhongyan wang, zechao ding, huiling zhang bio-information college, chongqing university of posts and telecommunications, chongqing 400065, china * corresponding author: e-mail: tiany20032003@163.com; tel.: +86-23-62460536; fax: +86-23-62460536 received: june 18, 2015; revised: august 09, 2015; published: september 05, 2014 abstract about 35 million people worldwide were suffered from alzheimer’s disease (ad) in 2014 and the number of patients was expected to increase by 4-fold in 2050. as a neurodegenerative disease impacting our society, the pathogenesis and prognosis of ad have not yet fully understood. senile plaque is generally regarded as one of the hallmarks of the disease, which may be due to the imbalance of aβ peptides in the brain. over the last decades, studies of early onset familial ad have led us to a deeper understanding of the genetics and molecular biology of ad. there is increasing evidence to suggest that the pathogenesis of ad is more likely to be caused by multiple genetic mutations. however, the precise genetic component leading to ad pathogenesis remains unclear. in this review, we will briefly introduce the classification of ad in the context of genetics. then we will discuss the gene mutations in chromosome 21, 19 and presenilin as well as their links to aβ peptides. imaging data will be discussed alongside to complement the associated structural and physiological changes in the brain. it is our hope that future research in genetics will continue to enhance our understanding of the pathogenesis of ad and the mechanisms leading to the formation of senile plaques. keywords neurobiological pathogenesis; senile plaque; aβ peptide; amyloid precursor protein; genetic mutation introduction alzheimer’s disease (ad) is a neurodegenerative disease in which the cognitive decline of the patient progresses over time. it is one of the most common diseases of senile dementia symptoms, including waning memory, and cognitive dysfunctions, such as time, place, character, emotion, as well as personality changes. severe ad patients could suffer from language obstacles leading to difficulties in daily communication. with the continuous improvement of living standard and life-span, the incidence rate of ad is rising. as of 2014, the number of ad patients has reached 35 million. in 2050, the number of ad patients will exceed 100 million. currently, the us has over 5 million ad patients, with an annual cost to the healthcare system over $ 200 billion. from now to 2050, the cost of treatment for ad will reach $ 20 trillion [1]. these represent a significant burden on the nation resource. clearly, a good understand on the pathogenesis and prognosis of ad is very important. classification of alzheimer’s disease in the context of genetics ad can be classified according to the age of the patient, namely age early and late ad. early onset ad http://www.pub.iapchem.org/ojs/index.php/admet/index y. tian et al. admet & dmpk 3(3) (2015) 203-215 204 (eoad) occurs in patients younger than 65 years old, while for the late onset ad (load) the patients are older than 65 years. one of the distinct pathological characteristics of ad is the presence of amyloid-β plaques in the brain [2]. regardless of eoad or load, the pathogenesis is more or less associated with genetics. gatz et al believe more than half (60 % 80 %) of the ad pathogenesis is influenced by genetics, which is referred to as the familial ad (fad) [3]. although the etiology and pathogenesis of ad is not completely understood, there are clear evidence suggest that gene mutations lead to extracellular deposition of amyloid-β (aβ) and subsequently the formation of senile plaque, which is one of one of the characteristic pathological changes in ad [4,5]. these ad disease genes are family associated and mostly autosomal dominant, with the fad gene located on chromosome 21, 19, 14 and 1. on the 21, 14, and 1 chromosome, the mutations are linked to eoad. on the 19 chromosome carrying apolipoprotein e (apoe), the mutations are linked to load [2]. most of the pathogenesis of ad is not solely determined by a single gene. instead, it is more likely to be caused by multiple genes and environmental factors, together with genetic mutation [6,7]. however, the precise genetic component leading to ad pathogenesis remains unclear. mechanisms of ad related to aβ based on the pathological mechanism of aβ in the formation of senile plaques, this review will explore the pathology evidence and functional evidence of ad in the context of genetics. literature evidence suggested that there exists dynamic equilibrium between brain extracellular aβ generation (production) and clearance (clearance). when the balance is upset, aβ accumulation (accumulation) and aggregation (aggregation) begin [8-10]. neurotoxic aβ aggregation leads to the formation of plaques of amyloid-β, resulting in ad patients with dementia and other cognitive disorders. in addition, we will discuss other genetic factors leading to ad pathogenesis and the prevention mechanisms (see fig. 1). figure 1. hypothesis on ad based on mutations associated with aβ peptides. the rationale behind this hypothesis is that genes and gene mutations affect the production, clearance, degradation and aggregation of aβ. these metabolic processes, if not in balance, may lead to cell death, the formation of senile plaques and increased ad risk. chromosome 21: formation of aβ autosomal dominant inheritance ad and/or family eoad are closely related to the amyloid precursor protein (app) on the chromosome 21. specifically, the hydrolysis of app by β-secretase and γ-secretase complex leads to the formation of hydrophobic amyloid-β [11]. at the same time, hydrolysis lead to the formation of water soluble app intracellular structure (aicd), which could result in apoptosis. in addition to admet & dmpk 3(3) (2015) 203-215 the roles of genetics in alzheimer’s disease doi: 10.5599/admet.3.3.196 205 the production of amyloid-β, app can be hydrolyzed by the α-secretase enzyme to form 3-kda product (p3), aicd and soluble app extracellular domain (sappα), which has a neuroprotective effect [12,13]. excessive sappα can also suppress the formation of toxic aβ [11]. as shown in fig. 2, sappβ, aβ and aicd are generated in the 671/672 locus of app via the β-secretase enzyme hydrolysis, while sappα, p3 and aicd are produced in the 687/688 locus of app by α-secretase. the formation of aβ40 (aβ42) is accomplished by hydrolysis of app with β-secretase through γ40 (γ42) in which γ40 and γ42 are kinds of γsecretase (see fig. 2). choy r et al reported that aβ40 is mainly produced in trans-golgi network (tng). after the detachment of tng from the nucleus, the app aggregation in endosomes reduced, so as the accumulation of aβ [13]. in app transgenic mice, it has been found that the loss of neuronal cysteine protease cathepsin b (catb) led to in increased amyloid plaque load and increased aβ1-42 to aβ1-40 ratios. in early app transgenic mice, the expression level of catb elevated. this leads to the denaturation of aβ, resulting in the inactivation of aβ oligomers and reduction of aβ fibers [14]. app gene mutations: which one affecting aβ? as shown in fig. 2, app is modified by mutation to form toxic aβ. accumulation of these toxic peptides increases the risk of getting ad. usually, the mutation occurred at the cleavage site β-secretase enzyme (aβ) near the swedish km670 / 671nl. with this mutation, the intracellular aβ formation is about 6 to 8fold more likely than wild type [15]. however, whole-genome sequencing revealed that a673t mutation in app was indeed due to the β site being replaced, which led to the decrease in extracellular aβ peptides. for the elderly people that do not suffer from ad, the knowledge on a673t mutation may offer predictive and/or preventive measures for cognitive decline [16]. figure 2. schematic to show the hydrolysis of app and the formation of aβ (app gene on chromosome 21 mutation, abnormal hydrolysis to produce toxic aβ). app is hydrolyzed into peptide fragments (with aβ marked pink), including soluble sappβ, aβ, aicd or sappα, p3 and aicd. the formation of aβ40 (aβ42) is accomplished by hydrolysis of app with β-secretase through γ40 (γ42). in case of homozygotes, a673v (valine-673) gene mutation increases the generation of aβ42 and aβ40 simultaneously, thereby increasing the accumulation of amyloid fibrils in vitro. however, the same mutation shows the opposite effects on aβ42 and aβ40 in the heterozygotes [17]. more sappα is generated via the hydrolysis of app at allele 690 and 692 by the substitution of alanine and phenylalanine, respectively. at the same time, the generation of p3 peptide is reduced, which facilitates the increase of y. tian et al. admet & dmpk 3(3) (2015) 203-215 206 aβ40 and aβ42 [18]. for arctic e693 mutation carriers, the levels of aβ42 and aβ40 in plasma are lower than that of wild type. yet, the rate of aβ generation in the mutants is faster, which may reflect another possible disease mechanism associated with ad [19]. at the γ cleavage sites, the a713t mutation can lead to cerebral amyloid angiopathy symptoms, which is also associated with ad [20]. in iranian fad, allele thr714ala may induce a remarkable increase in the aβ1-42 / aβ1-40 ratio (up to 11 times) [21]. recently, sieczkowski e found that i716f βapp is a mutation associated with glutamate-aβ, which increases the generation of senile plaques and nerve tangles [22]. the app 717 mutation increases the production of the soluble βapp, aβ42 and aβ38, which are associated with dementia [23-25]. the newly discovered k724 mutation in china (40-58 years old) has been shown to enhance the aβ1-42 / aβ1-40 ratio (over 2-fold). monitoring this mutation may provide preventive measures for early-onset fad in china [26]. another mutation k16n, which is situated at the α hydrolysis sites, may provide information on earlyonset ad risk. k16n affects the app alleles and the formation of longer aβ peptides (e.g. from aβ40 to aβ42). although the aβ peptide itself does not show cytotoxic effects, equimolecular mixture of both wild type aβ and mutant type aβ is far more toxic. in addition, k16n inactivates the aβ-degrading enzyme, neprilysin, affecting the elimination of aβ42 peptide [27]. the d7h mutation could increase the aβ1-42 production and the aβ1-42 / aβ1-40 ratio (approximately 2-fold), and the resulting aβ1-42 peptide becomes more toxic. in the presence of zn 2+ or cu 2+ , d7h-aβ aggregates to form oligomers. hence, d7h allele can be used to increase d7h-aβ the production and aggregation [28,29]. h6r mutations increase the toxicity of all subtypes (isoform) and their rate of fiber formation. due to the increase in aβ dimers and c-terminal monomers, aβ42 fiber forming time is shortened [30]. app mutations not only can direct the aβ generation (production), clearance (clearance), accumulation (accumulation), aggregation (aggregation) to influence the potential risk of ad, but also can increase the fat content of the body, indirectly enhancing the risk of ad [31]. app imaging modalities: changes of aβ in brain atrophy a wide range of imaging modalities, such as mri (magnetic resonance imaging), functional mri, 11cpib (carbon 11-labeled pittsburgh compound b) pet (positron emission tomography) imaging have been utilized in supporting ad research [32,33]. in particular, some of these techniques enable the in situ detection and monitoring of senile plagues in ad patients [34]. these may provide evidence on the research of app mutation and pathology evidence to enable the identification of the early signs of ad. mutation of app-e693δ was found in cases of early fad. the ad patients with the ad e693δ mutation exhibit a memory dysfunction and partial brain atrophy. however, senile plaques are rarely seen, which may be due to low level of accumulation of aβ in vivo [32]. according to 11c-labeled pib-pet and 18f-fdgpet imaging, a significant difference between the arctic app (apparc) and sweden (appswe) mutations on app has been observed [33]. apparc mutation carriers show low level of absorption of pib in cortex, while the rate of glucose metabolism in the brain and the aβ42 levels in cerebral spinal fluid (csf) are low. appswe mutation carriers absorbed more pib in cortex, with the cerebral glucose metabolic rate reduced, which was especially obvious in the striatum. the aβ42 level in csf was increased significantly [33]. in contrast with the cerebro-absorption of 11c-pib in sporadic ad patients, fad with app mutation carriers show large increases in absorption in striatum, posterior cingulate (caudate nucleus, putamen), while absorption in other cortical areas also show small increases [34]. in addition to the frontal lobe, the temporal lobe, occipital lobe and posterior cingulate cortex and other brain areas show increased absorption of 11c-pib [35,36]. according to fmri data, app allelic variations can lead to increased activity in brain middle temporal gyri (mtg) and left fusiform gyri regions [37]. in summary, app mutations may admet & dmpk 3(3) (2015) 203-215 the roles of genetics in alzheimer’s disease doi: 10.5599/admet.3.3.196 207 affect the accumulation and deposition of aβ and other cortical regions (see fig. 3). this may also change the rate of glucose metabolism in the cerebral cortex and other regions of the striatum, resulting in an increased risk of synaptic dysfunction in ad patients. further investigation will be required to understand the app mutations and the exact mechanisms in regulating aβ. chromosome 19: (apoe) the effects of apoe on the clearance of aβ: impacts of ε2, ε3, ε4 on aβ the app mutation mostly alters the generation of amyloid-β, which affects the ad pathogenesis. in marked contrast, the apolipoprotein e (apo-e) in chromosome 19 works by clearing the amyloid-β to reduce the risk of ad. if there is no mutation, apo-e plays the role of clearing the soluble amyloid-β. with the aid of apo-e, the activity of neprilysin is enhanced for the elimination of intracellular aβ. in extracellular environment, the elimination of toxic aβ is accomplished by apo-e related insulin degrading enzymes. thus, apo-e isoforms can enhance the degradation of aβ, while lipidated apo-e can also be useful to accelerate the aβ clearance [38]. unlike apo-e gene, apo-e has three alleles ε2, ε3, ε4, which can promote the formation of hydrophobic aβ and reduce the clearance of toxic aβ, which increases the potential risk of ad [39]. apoe may be combined with hydrophobic ß-ap, which induces the formation of ß-ap monofilament and precipitation. comparing with the allele ε3, ε4 forms a stronger bonding between the carrier protein apo-e4 and ß-ap. in cases of the apo-e ε4 homozygous individuals, the amount of ß-ap precipitation in brain is far more than that of the apo-e ε3 gene homozygous individuals. in the ß-ap fibrous structure formation, apo-e played an important role in the formation of senile plaques. for ad patients containing ε2, ε3, ε4 mixing genes in the brain, individuals with ε4 allele show higher incidence rate, while individuals with of ε2 allele show lower incidence rate [39-42]. for very elderly patients (90 years or more), apo-e ε2 associated with intact cognition (dementia associated risk reduced). however, the increase in apo-e ε2 associated with ad related neuropathology [40]. autopsy on ad patients revealed that most of the aβ peptide accumulation and precipitation genotypes are: apo-e ε3/ε4 (ε3/ε4 > ε3/ε3 > ε2/ε3). the apo-e ε4 homozygotes carriers are generally dementia [41]. however, the age of ad diagnosed is negatively correlated with the risk of incidence. apo-e ε4 gene is related to the lowest age of ad diagnosed, while apo-e ε2 gene is related to the highest age of ad diagnosed. for apo-e ε3 gene, the age of ad diagnosed and the risk of incidence are between that of ε4 and ε2 [42]. apo-e ε4 and other factors that affect ad it has been reported that the vast majority of ad patients (approximately 65% -75%) carry apo-e ε4 allele. the distribution of apo-e allele varies among different races and geographical locations. usually near the equator and in the arctic region, populations carry higher frequency of apo-e ε4 allele [43]. for the chinese people, the risk of ad incidence with apo-e ε4 allele was 3.93 times above average. in particular, ε4 homozygous genotype is a risk factor for ad; while apo-e ε3 carriers show the risk of ad incidence lower than average [44]. apo-e ε4 allele may also affect patient load by reducing (approximately halved) the insulin-degrading enzyme (ide) in the hippocampus. the main function of ide is the degradation and removal of amyloid-β in the brain. the involvement of apo-e ε4 and ide in amyloid-β metabolism increases the risk of ad [45]. y. tian et al. admet & dmpk 3(3) (2015) 203-215 208 before the onset ad (about 70 years), apo-e allele and brain-derived neurotrophic factor (bdnf) lead to the brain's cognitive decline. in 4.5-year time, comparing to aβ and ε4 noncarriers, aβ/ε4 carriers showed significant cognitive decline, while individuals carrying aβ without carrying ε4 show only vocabulary cognitive decline. in aβ carriers (4.5-year later), comparing with the ε4 noncarriers, and ε4 / bdnfmet carriers (brain derived neurotrophic factor met), the vocabulary and visual episodic memory decreased significantly faster.regardless of the status of bdnf and apo-e (4.5-year later), only little cognitive decline was observed in non-ε4 carriers [46]. in the mice blood-brain barrier (bbb) model, it has been shown that apo-e allele undermines the clearance of neurotoxic aβ peptides (destructive effects: apo-e4 is greater than that of the apo-e3 and apo-e2) [47]. in this context, aβ peptides bind to apo-e4, resulting in a rapid removal of free aβ1-40 / aβ1-42. the apo-e4 inside the bbb can form complex with aβ, which is significantly slower than that with lrp1 (ldl (low density lipoprotein) receptor-related protein 1). in contrast, the apo-e2 and apo-e3 on the bbb can form complex with aβ in vlvlr (vldl (very low density lipoprotein) receptor) and lrp1, where the clearance rate is faster than that of the aβ-apo-e4 complex [8]. the brain apo-e allele offers different models of clearance for aβ, which may help to understand the disease progression of ad in animal/human models. pet, mri, dti and fmri imaging for apo-e ε4 and their influences on brain networks as shown in fig. 3, using the florbetapir (18f-fdg) positron emission tomography (pet) imaging technique, it has been shown that the apoe alleles in chromosome 19 and butyrylcholinesterase (bche) can be regarded as aβ peptide deposition regulator [47]. in cognitively normal ad patients, the cortex glucose metabolism rate (cmrgi) is significantly reduced for apo-e ε4 carrier [ε4 +], especially in the front frontal lobe, parietal lobe, temporal lobe and posterior cingulate regions [48-52]. figure 3. schematic to shows the app, apoe, psen mutations causing changes in the brain function. app gene mutations resulted in the increase in the absorption of 11c-pib in striatum, posterior cingulate gyrus (pcg), temporal gyrus (tg), occipital gyrus (og), with the middle temporal gyri (mtg) and left fusiform gyri (ffg.l) activation increased. apo-e ε4 allele mutations significantly reduced the superior frontal gyrus (sfg), parietal gyrus (pg), tg, pcg cortex glucose metabolism rate (cmrgi) and altered functional connectivity in brain regions including: pg, left sfg (sfg.l) and anterior cingulate gyri (acg). psen1 gene mutation carriers showed significant increase in the absorption of pib in striatum, acg, pcg, thalamus and og cortex. the axonal density increased in left thalamus (tha.l), striatum, hippocampus (hip). network characteristics changed including the default network connection, reduction of the sfg connections and increase in the inferior frontal gyrus (ifg) connections. the task state and the activation strength of lsfg and og were reduced. admet & dmpk 3(3) (2015) 203-215 the roles of genetics in alzheimer’s disease doi: 10.5599/admet.3.3.196 209 apo-e ε4 dose is related with higher and earlier risk of dementia [50]. the hippocampal volume, the hippocampal glucose metabolism and posterior cingulate glucose metabolism are closely related to apo-e [ε4 +] gene. among these, the correlation of posterior cingulate glucose metabolism and apo-e [ε4 +] allele is significantly greater than the correlation of hippocampal volume, glucose metabolism and apo-e [ε4 +] allele [50]. mri imaging has been used to study the cortical thickness in vietnamese men. it has been shown that in the apo-e homozygote-carriers [ε3 +] and the brain regions of apo-e ε3/4 carriers, the frontal lobe, the left front and right rear volume central region are significantly thinner ; for apo-e ε2/3 carriers, the right parahippocampal cortex region is relatively thick, which may be due to natural aging [53]. diffusion tensor imaging (dti) of t1 structure imaging revealed that apo-e carrier [ε4 +] loss the interconnection faster than that due to aging, especially in the medial prefrontal orbital frontal cortex, precuneus and outer local parietal cortex region where the connectivity is reduced [54]. in a picture learning experiment using fmri monitoring, apo-e carriers [ε4 +] show enhanced bold signal during learning. this suggests that apo-e4 allele in older adults with ad exhibit greater genetic risk with cognitive-related episodic memory [55]. there are significant functional connectivity changes in parietal lobe, left frontal regions and anterior cingulate [56, 57]. for apo-e carriers [ε4 +], network connections were reduced in the left middle temporal gyrus, left parietal lobe, regional connectivity bilateral anterior temporal lobe, but were enhanced in bilateral insular cortex, cingulate, connectivity striatum and thalamus medial prefrontal cortex [56]. during memory activation task, apo-e carriers [ε4 +] show increased brain activation in the parietal lobe, left hippocampus and the prefrontal region. during the recall memory task, the signal density increases particularly in the hippocampus region with apo-e carrier [ε4 +] greater than [ε3 +]. the average number of whole-brain activation area also increases [57]. apo-e carrier [ε4 +] topology of the whole brain significantly impaired and the amount of parallel transmission of information reduced. moreover the connections between the rear default mode network (p-dmn) and other brain functions were decreased. ecn，p-dmn and sensory-motor system also showed abnormalities and global topology damaged [58]. thus, apo-e plays a role in the elimination of aβ. with apo-e genetic mutations, amyloid-β clearance is compromised, leading to aβ accumulation. regional, environmental, age and other factors could affect the apo-e genetic mutations. comparing with the apo-e carrier [ε3 +], apo-e carrier [ε4 +] is more likely to be an ad patient. presenilin (psen) mutations in psen, psen-1 and psen-2 presenilin (psen) is a transmembrane protein that functions as a part of the γ-secretase intramembrane protease complex. psen regulates protein transport and aberrant protein synthesis. clearly, protein dysfunctions can affect βapp, which could accelerate the development of ad [59,60]. in presenilin-1 (psen-1) gene and its homologous genes presenilin-2(psen-2), the most common mutations occur in chromosome 14 and chromosome 1, which could lead to eoad. however, the n141i mutation in the 342/343 position of psen-2 is linked to eoad and load [61]. in transgenic mice model, it has been shown that psen-2/n141i mutation can reduce the aβ40 level, but produce more neurotoxic aβ42 leading to the aβ1-42/aβ1-40 ratio increases, and undermining the aβ clearance [62]. the mutations in psen-2 mainly include: a85v, t122p, t122r, e126k, v148i, m174v, s175c, q228l, y. tian et al. admet & dmpk 3(3) (2015) 203-215 210 y231c, m239i, m239v, t430m and d439a. these mutations could lead to eoad [63-66]. among these, t122r mutation modifies nerve fibers and m239 increases the production aβ42 peptide, which are all dementia-related. overexpression of psen-2 may promote aβ42 fragments induced apoptosis [64,65]. for psen-1, mutation mainly changes the activity of γ-secretase thus enhancing the generation of aβ peptides. the m146l, l166p and n141i mutations in psen-1 could lead to a functional impairment of γ-secretase, and inhibition of endogenous generation of aβ1-40, leading to an increase of aβ1-42 and the aβ1-42/aβ1-40 ratio [67-69]. the psen-1/g384a mutation is due to the accumulation of intracellular aβ hydrolyzate (> aβ1-42), which slows down the rate of degradation of aβ in the cell membranes. after a long period of time, the aβ hydrolyzate eventually degrades. although this had no effect on the generation of aβ1-42, the aβ1-40 level was reduced [70]. in transgenic mice, heterozygous carriers with psen-1/r278i mutation express higher level of neurotoxic aβ1-43 peptide, thus inhibiting the expression level of normal aβ1-40 peptide. embrynoic lethality was observed in homozygous carriers, indicating that the psen-1/r278i mutation could lead to physiological defects [71]. psen-1/i213t and psen1/c410y mutations reduce the generation of aβ1-40 and aβ1-42 peptides [72,73], which is mainly due to the psen-1/i213t mutation destroys the app c-terminal fragment of the cleavage γsites. thus, the total reaction rate γ-secretase is reduced, resulting in an increase of aβ1-43, aβ1-45 exceeding the aβ1-46 production [72]. in psen-1 mutations, regardless of increase or decrease in the aβ1-40 and aβ1-42 levels, the aβ1-42/aβ1-40 ratio always increases, resulting in neurotoxicity and increased risk of eoad. 6.2 psen-1 mutations affect the degradation of aβ γ-secretase complex is one of key components for the hydrolysis of app into aβ peptides. in case of ad, psen-1 mutations significantly affect the activity of γ-secretase complex and alter aicd thereby increasing the aβ1-42/aβ1-40 ratio [74]. through the use of γ-secretase, psen-1 fad-linked mutations can degrade aβ peptides, with the degradation pathways as follows: aβ49 → aβ46 → aβ43 → aβ40 and aβ48 → aβ45 → aβ42 → aβ38 this increases the neurotoxicity by increasing the aβ1-42/aβ1-40 residue [75]. moreover, in vitro and cell internal psen / γ-secretase offer additional degradation pathway: aβ43 → aβ38.in contrast, the psen mutant acts on γ-secretase by inhibiting the degradation of longer aβ. psen / γ-secretase can minimize the functional loss of aβ42 and aβ43, which could in turn help to prevent ad [76]. of course, the lower the toxicity of aβ42 peptide and increase γ-secretase modulators (gsms) can be fed back to the tmd-1 (transmembrane domain) of psen-1 to change its structure. in tmd-1 abnormal secretions include aβ45 and longer aβ peptides. tmd-1 helical surface comprises the binding function of aβ45 and aβ48. this combination changes the encoding mutations of the gsms and tmd-1. at the same time, such a combination facilitates the formation of the hydrophobic aβ42 [77]. psen-1 imaging study and regional shrinking effects with the aid of cerebrospinal fluid, plasma markers, structural mri and fmri studies, it is possible to assess the risk factors of eoad in psen-1 autosomal mutations. in psen-1/e280a mutation carriers, it has been observed that the average age showing aβ toxicity was 28, while the average age showing cognitive admet & dmpk 3(3) (2015) 203-215 the roles of genetics in alzheimer’s disease doi: 10.5599/admet.3.3.196 211 impairment was 44 [78]. klunk we, et al. reported that in the psen-1/c410y fad-linked mutation carriers, striatal significant aβ deposition has been observed in striatum, which led to cortical lesion. however, the lesion area is significantly smaller than that of sporadic ad patients [79]. the use of 11c-pib pet measurements further confirmed that the psen-1 striatal lesions were indeed caused by mutations. compared with ad patients without psen-1, the striatum of the mutation carriers exhibited a considerably higher uptake of pib. beside, anterior cingulate, posterior cingulate, thalamus and occipital cortex cortical areas have also shown significant enhanced uptake of pib [80]. in presymptomatic fad, it has been shown that in psen-1 mutation carriers, the left thalamus and caudate nucleus shrank significantly. the density of axons, right hippocampus and right cingulate axon density were also reduced. in cases of fad, axon density in hippocampus and caudate nucleus increase. this may be due to the intensive axonal lesion area further extending to the entire brain networks [81]. for psen-1 mutation carriers, a decrease in the glucose metabolic rate has been observed in pcc, bilaterally inferior parietal lobule, superior temporal gyrus, hippocampus and left entorhinal cortex regions [82]. when psen-1 mutation carriers undergo a visual coding task, resting-state fmri shows the default network changes, which include a reduction in the number of connections in the frontal lobe, an increase in the number of connections at frontal rear increases. at the task state, the activation strength at the left prefrontal cortex and occipital are reduced, while posterior cingulate and precuneus activation are enhanced (see fig. 3) [83]. conclusion overall, the gene mutations in chromosome 21, 19, 14 and 1 in the context of ad are less than 30 %. more genetic risk factors leading to ad have yet to be discovered [5]. currently, chromosome 12, 10 and 9 are found to be genetically related to ad. chromosome 10 mutations increased intracellular aβ1-42 peptides and plasma aβ. chromosome 10 and 12 mutations also increased the risk of fad-linked load [84-85]. finally, research in ad pathology and function in the context of genetics will continue to enhance our understanding of the pathogenesis of ad and the formation of senile plaques. through the regulations of α-, βand γ-secretases, it may be possible to help restoring the homeostasis of aβ, which could prevent and/or reduce the incidence of ad. acknowledgements we thank the support from national nature science foundation of china (#31100745) and the cqcbe project kj110502. references [1] world alzheimer report 2014 dementia and risk reductio，http://www.alz.co.uk/research/world alzheimer report 2014.pdf (september 2014). [2] alzheimer's association, alzheimer's & dementia 10 (2014) 47-92. [3] m. gatz, c.a. reynolds, l. fratiglioni, b. johansson, j.a. mortimer, s. berg, a. fiske, n.l. pedersen, archives of general psychiatry 63 (2006) 168-174. [4] r.e. tanzi, cold spring harbor perspectives in medicine 2 (2012) 006-296. [5] a. goate, j. hardy, journal of neurochemistry 120 (2012) 3-8. [6] m. gatz, j.y. jang, i.k. karlsson, n.l. pedersen, advances in behavior genetics 1 (2014) 201-231. http://www.alz.co.uk/research/world%20alzheimer%20report%202014.pdf http://www.alz.co.uk/research/world%20alzheimer%20report%202014.pdf y. tian et al. admet & dmpk 3(3) (2015) 203-215 212 [7] h. amieva, h. mokri, m. le goff, c. meillon, h. jacqmin-gadda, a. foubert-samier, j.-m. orgogozo, y. stern, j.-f. dartigues, brain 137 (2014) 1167-1175. [8] r. deane, a. sagare, k. hamm, m. parisi, s. lane, m.b. finn, d.m. holtzman, b.v. zlokovic, the journal of clinical investigation 118 (2008) 4002-4013. [9] k.g. mawuenyega, w. sigurdson, v. ovod, l. munsell, t. kasten, j.c. morris, k.e. yarasheski, r.j. bateman, science 330 (2010) 1774-1774. [10] p.e. cramer, j.r. cirrito, d.w. wesson, c.y. daniel lee, j.c. karlo, a.e. zinn, b.t. casali, j.l. restivo, w.d. goebel, m.j. james, k.r. brunden, d.a. wilson, g. e. landreth, science 335 (2012) 1503-1506. [11] h. zhang, q. ma, y. zhang, h. xu, journal of neurochemistry 120 (2012) 9-21. [12] y. zhang, r. thompson, h. zhang,h. xu, 4:3 mol. brain (2011). [13] h. zheng, e.h. koo, mol. neurodegener. 6 (2011) 27. [14] r.w.y. choy, z. cheng, r. schekman, proceedings of the national academy of sciences 109 (2012) e2077-e2082. [15] g.m. cole, s.a. frautschy, neuron 51 (2006) 671-672. [16] m. citron, t. oltersdorf, c. haass, l. mcconlogue, a.y. hung, p. seubert, c. vigo-pelfrey, i. lieberburg, d.j. selkoe, nature 360 (1992) 672-674. [17] t. jonsson, j.k. atwal, s. steinberg, j. snaedal, p.v. jonsson, s. bjornsson, h. stefansson, p. sulem, d. gudbjartsson, j. maloney, k. hoyte, a. gustafson, y. liu, y. lu, t. bhangale, r.r. graham, j. huttenlocher, g. bjornsdottir, o.a. andreassen, e.g. jönsson, a. palotie, t.w. behrens, o.t. magnusson, a. kong, u. thorsteinsdottir, r.j. watts, k. stefansson, nature 488 (2012) 96-99. [18] g. di fede, m. catania, m. morbin, g. rossi, s. suardi, g. mazzoleni, m. merlin, a.r. giovagnoli, s. prioni, a. erbetta, c. falcone, m. gobbi, l. colombo, a. bastone, m. beeg, c. manzoni, b. francescucci, a. spagnoli, l. cantù, e. del favero, e. levy, m. salmona, f. tagliavini, science 323 (2009) 1473-1477. [19] c. haass, a.y. hung, d.j. selkoe, d.b. teplow, journal of biological chemistry 269 (1994) 1774117748. [20] c. nilsberth, a. westlind-danielsson, c.b. eckman, m.m. condron, k. axelman, c. forsell, c. stenh, j. luthman, d.b. teplow, s.g. younkin, j. näslund, l. lannfelt, nature neuroscience 4 (2001) 887-893. [21] g. rossi, g. giaccone, r. maletta, m. morbin, r. capobianco, m. mangieri, a.r. giovagnoli, a. bizzi, c. tomaino, m. perri, m. di natale, f. tagliavini, o. bugiani, a.c. bruni, neurology 63 (2004) 910912. [22] p. pasalar, h. najmabadi, a.r. noorian, b. moghimi, a. jannati, a. soltanzadeh, t. krefft, r. crook, j. hardy, neurology 58 (2002) 1574-1575. [23] e. sieczkowski, i. milenkovic, v. venkataramani, r. giera, t. ströbel, r. höftberger, p.p. liberski, e. auff, o. wirths, t.a. bayer, g.g. kovacs, journal of alzheimer's disease 44 (2015) 103-114. [24] c.r. muratore, h.c. rice, p. srikanth, d.g. callahan, t. shin, l.n.p. benjamin, d.m. walsh, d.j. selkoe, t.l. young-pearse, human molecular genetics 23 (2014) 3523-3536. [25] m.c. chartier-harlin, f. crawford, h. houlden, a. warren, d. hughes, l. fidani, a. goate, m. rossor, p. roques, j. hardy, m. mullan, nature 353 (1991) 844-846. [26] j.r. murrell, a.m. hake, k.a. quaid, m.r. farlow, b. ghetti, archives of neurology 57 (2000) 885-887. [27] x.l. peng, l. hou, s.h. xu, y. hua, s.-j. zhou, y. zhang, y.-p. zheng, y.-h. fu, q. xu, l.-s. zhang, j. wang, x.-t. guan, j.-s. he, neurobiology of aging 35 (2014) 2657. e1-2657.e6. [28] d. kaden, a. harmeier, c. weise, l.m. munter, v. althoff, b.r. rost, p.w. hildebrand, d. schmitz, m. schaefer, r. lurz, s. skodda, r. yamamoto, s. arlt, u. finckh, g. multhaup, embo molecular medicine 4 (2012) 647-659. [29] w.t. chen, c.j. hong, y.t. lin, w.-h. chang, h.-t. huang, j.-y. liao, y.-j. chang, y.-f. hsieh, c.-y. cheng, h.-c. liu, y.-r. chen, i. h. cheng, plos one 7 (2012) e35807. admet & dmpk 3(3) (2015) 203-215 the roles of genetics in alzheimer’s disease doi: 10.5599/admet.3.3.196 213 [30] p.m. truong, m.h. viet, p.h. nguyen, c.-k. hu, m.s. li, the journal of physical chemistry b 118 (2014) 8972-8981. [31] m.h. viet, p.h. nguyen, p. derreumaux, m.s. li, acs chemical neuroscience 5 (2014) 646-657. [32] k.l. puig, a.m. floden, r. adhikari, m.y. golovko, c.k. combs, plos one 7 (2012) e30378. [33] h. shimada, s. ataka, t. tomiyama, h. takechi, h. mori, t. miki, dementia and geriatric cognitive disorders 32 (2011) 45-54. [34] m. schöll, a. wall, s. thordardottir, d. ferreira, n. bogdanović, b. långström, o. almkvist, c. graff, a. nordberg, neurology 79 (2012) 229-236. [35] a.m. remes, l. laru, h. tuominen, s. aalto, n. kemppainen, h. mononen, k. någren, r. parkkola, j.o. rinne, archives of neurology 65 (2008) 540-544. [36] w.e klunk, h. engler, a. nordberg, y. wang, g. blomqvist, d.p. holt, m. bergström, i. savitcheva, g.f. huang, s. estrada, b. ausén, m.l. debnath, j. barletta, j.c. price, j. sandell, b.j. lopresti, a. wall, p. koivisto, g. antoni, c.a. mathis, b. långström, annals of neurology 55 (2004) 306-319. [37] n.m. kemppainen, s. aalto, i.a. wilson, k. någren, s. helin, a. brück, v. oikonen, m. kailajärvi, m. scheinin, m. viitanen, r. parkkola, j.o. rinne, neurology 67 (2006) 1575-1580. [38] m.n. braskie, l.d. medina, y. rodriguez-agudelo, d.h. geschwind, m.a. macias-islas, j.l. cummings, s.y. bookheimer, j.m. ringman, neurobiology of aging 33 (2012) 424.e11-424.e21. [39] q. jiang, c.y.d. lee, s. mandrekar, b. wilkinson, p. cramer, n. zelcer, k. mann, b. lamb, t.m. willson, j.l. collins, j.c. richardson, j.d. smith, t.a. comery, d. riddell, d.m. holtzman, p. tontonoz, g.e. landreth, neuron 58 (2008) 681-693. [40] c.c. liu, t. kanekiyo, h. xu, g. bu, nature reviews neurology 9 (2013) 106-118. [41] d.j. berlau, m.m. corrada, e. head, c.h. kawas, neurology 72 (2009) 829-834. [42] t. polvikoski, r. sulkava, m. haltia, k. kainulainen, a. vuorio, a. verkkoniemi, l. niinistö, p. halonen, k. kontula, new england journal of medicine 333 (1995) 1242-1248. [43] turner, jonathan d., konrad beyreuther, and franz theuring, springer science & business media 17 (1996). [44] s. kern, k. mehlig, j. kern, h. zetterberg, d. thelle, i. skoog, l. lissner, k. blennow, a. börjessonhanson, american journal of epidemiology 181 (2015) 214-217. [45] m. liu, c. bian, j. zhang, f.wen, scientific reports 4 (2014) 4383. [46] d.g. cook, j.b. leverenz, p.j. mcmillan, j.j. kulstad, s. ericksen, r.a. roth, g.d. schellenberg, l.-w. jin, k.s. kovacina, s. craft, the american journal of pathology 162 (2003) 313-319. [47] y.y. lim, v.l. villemagne, s.m. laws, r.h. pietrzak, p.j. snyder, d. ames, k.a. ellis, k. harrington, a. rembach, r.n. martins, c.c. rowe, c.l. masters, p. maruff, molecular psychiatry (2014). [48] v.k. ramanan, s.l. risacher, k. nho, s. kim, s. swaminathan, l. shen, t.m. foroud, h. hakonarson, m.j. huentelman, p.s. aisen, r.c. petersen, r.c. green, c.r. jack, r.a. koeppe, w.j. jagust, m.w. weiner, a.j. saykin, molecular psychiatry 19 (2014) 351-357. [49] e.m. reiman, r.j. caselli, l.s. yun, k. chen, d. bandy, s. minoshima, s.n. thibodeau, d. osborne, new england journal of medicine 334 (1996) 752-758. [50] e.m. reiman, k. chen, g,e, alexander, r.j. caselli, d. bandy, d. osborne, a.m. saunders, j. hardy, proceedings of the national academy of sciences of the usa 102 (2005) 8299-8302. [51] h.d. protas, k. chen, j.b.s. langbaum, a.s. fleisher, g.e. alexander, w. lee, d. bandy, m.j. de leon, l. mosconi, s. buckley, d. truran-sacrey, n. schuff, m.w. weiner, r.j. caselli, e.m. reiman, jama neurology 70 (2013) 320-325. [52] e.m. reiman, k. chen, x. liu, d. bandy, m. yu, w. lee, n. ayutyanont, j. keppler, s.a. reeder, j.b.s. langbaum, g.e. alexander, w.e. klunk, c.a. mathis, j.c. price, h.j. aizenstein, s.t. dekosky, r.j. caselli, proceedings of the national academy of sciences of the usa 106 (2009) 6820-6825. [53] e.m. reiman, k. chen, g.e. alexander, r.j. caselli, d. bandy, d. osborne, a.m. saunders, j. hardy, proceedings of the national academy of sciences of the usa 101 (2004) 284-289. y. tian et al. admet & dmpk 3(3) (2015) 203-215 214 [54] c. fennema-notestine, m.s. panizzon, w.r. thompson, c.-h. chen, l.t. eyler, b. fischl, c.e. franz, m.d. grant, a.j. jak, t.l. jernigan, m.j. lyons, m.c. neale, l.j. seidman, m.t. tsuang, h. xian, a.m. dale, w.s. kremen, journal of alzheimer's disease 26 (2011) 49-60. [55] j.a. brown, k.h. terashima, a.c. burggren, l.m. ercoli, k.j. miller, g.w. small, s.y. bookheimer, proceedings of the national academy of sciences of the usa 108 (2011) 20760-20765. [56] m.w. bondi, w.s. houston, l.t. eyler, g.g. brown, neurology 64 (2005) 501-508. [57] m.m. machulda, d.t. jones, p. vemuri, e. mcdade, r. avula, s. przybelski, b.f. boeve, d.s. knopman, r.c. petersen, c.r. jack, archives of neurology 68 (2011) 1131-1136. [58] s.y. bookheimer, m.h. strojwas, m.s. cohen, a.m. saunders, m.a. pericak-vance, j.c. mazziotta, g.w. small, new england journal of medicine 343 (2000) 450-456. [59] j. wang, x. wang, y. he, x. yu, h. wang, y. he, human brain mapping 36 (2015) 1828-1846. [60] g. struhl, i. greenwald, nature 398 (1999) 522-525. [61] y. ye, n. lukinova, m.e. fortini, nature 398 (1999) 525-529. [62] s. jayadev, j.b. leverenz, e. steinbart, j. stahl, w. klunk, c.-e. yu, t.d. bird, brain 133 (2010) 11431154. [63] s. kumar-singh, j. theuns, b. van broeck, d. pirici, k. vennekens, e. corsmit, m. cruts, b. dermaut, r. wang, c. van broeckhoven, human mutation 27 (2006) 686-695. [64] m. canevelli, p. piscopo, g. talarico, n. vanacore, a. blasimme, a. crestini, g. tosto, f. troili, g.l. lenzi, a. confaloni, g. bruno, neuroscience & biobehavioral reviews 42 (2014) 170-179. [65] g. binetti, s. signorini, r. squitti, a. alberici, l. benussi, e. cassetta, g.b. frisoni, l. barbiero, e. feudatari, f. nicosia, c. testa, o. zanetti, m. gennarelli, d. perani, d. anchisi, r. ghidoni, p.m. rossini, annals of neurology 54 (2003) 832-836. [66] g. marcon, g. giaccone, c. cupidi, m. balestrieri, c.a. beltrami, n. finato, p. bergonzi, s. sorbi, o. bugiani, f. tagliavini, journal of neuropathology & experimental neurology 63 (2004) 199-209. [67] u. müller, p. winter, c. bolender, d. nolte, journal of alzheimer's disease 42 (2014) 109-113. [68] y. qi, m. morishima-kawashima, t. sato, et al., biochemistry 42 (2003) 1042-1052. [69] p. koch, i.y. tamboli, j. mertens, p. wunderlich, j. ladewig, k. stüber, h. esselmann, j. wiltfang, o. brüstle, j. walter, the american journal of pathology 180 (2012) 2404-2416. [70] k. v. nguyen, biomolecular concepts 6 (2015) 11-32. [71] r. fluhrer, a. fukumori, l. martin, et al., journal of biological chemistry 283 (2008) 30121-30128. [72] t. saito, t. suemoto, n. brouwers, k. sleegers, s. funamoto, n. mihira, y. matsuba, k. yamada, p. nilsson, j. takano, m. nishimura, n. iwata, c. van broeckhoven, y. ihara, t.c. saido, nature neuroscience 14 (2011) 1023-1032. [73] m. shimojo, n. sahara, t. mizoroki, s. funamoto, m. morishima-kawashima,t. kudo, m. takeda,y. ihara, h. ichinose, a. takashima, journal of biological chemistry 283 (2008) 16488-16496. [74] d. xia, h. watanabe, b. wu, s.h. lee, y. li, e. tsvetkov, v.y. bolshakov, j. shen, r.j. kelleher, neuron 85 (2015) 967-981. [75] m. cacquevel, l. aeschbach, j. houacine, p.c. fraering, plos one 7 (2012) e35133. [76] m.a. fernandez, j.a. klutkowski, t. freret, m.s. wolfe, journal of biological chemistry 289 (2014) 31043-31052. [77] m. okochi, s. tagami, k. yanagida, m. takami, t.s. kodama, k. mori, t. nakayama, y. ihara, m. takeda, cell reports 3 (2013) 42-51. [78] y. ohki, n. shimada, a. tominaga, s. osawa, t. higo, s. yokoshima, t. fukuyama, t. tomita, t. iwatsubo, molecular neurodegeneration 9:7 (2014). [79] e.m. reiman, y.t. quiroz, a.s. fleisher, k. chen, c. velez-pardo, m. jimenez-del-rio, a.m. fagan, a.r. shah, s. alvarez, a. arbelaez, m. giraldo, n. acosta-baena, r.a. sperling, b. dickerson, c.e. stern, v. tirado, c. munoz, r.a. reiman, m.j. huentelman, g.e. alexander, j.b.s. langbaum, k.s. kosik, p.n. tariot, f. lopera, the lancet neurology 11 (2012) 1048. admet & dmpk 3(3) (2015) 203-215 the roles of genetics in alzheimer’s disease doi: 10.5599/admet.3.3.196 215 [80] w.e. klunk, j.c. price, c.a. mathis, n.d. tsopelas, b.j. lopresti, s.k. ziolko, w. bi, j.a. hoge, a.d. cohen, m.d. ikonomović, j.a. saxton, b.e. snitz, d.a. pollen, m. moonis, c.f. lippa, j.m. swearer, k.a. johnson, d.m. rentz, a.j. fischman, h.j. aizenstein, s.t. dekosky, the journal of neuroscience 27 (2007) 6174-6184. [81] j. koivunen, a. verkkoniemi, s. aalto, a. paetau, j.-p. ahonen, m. viitanen, k. någren, j. rokka, m. haaparanta, h. kalimo, j. o. rinne, brain 131 (2008) 1845-1853. [82] n.s. ryan, s. keihaninejad, t.j. shakespeare, m. lehmann, s.j. crutch, i.b. malone, j.s. thornton, l. mancini, h. hyare, t. yousry, g.r. ridgway, h. zhang, m. modat, d.c. alexander, m.n. rossor, s. ourselin, n.c. fox, brain 136 (2013) 1399-1414. [83] l. mosconi, s. sorbi, m.j. de leon, y. li, b. nacmias, p.s. myoung, w. tsui, a. ginestroni, v. bessi, m. fayyazz, p. caffarra, a. pupi, journal of nuclear medicine 47 (2006) 1778-1786. [84] r. sala-llonch, j. fortea, d. bartres-faz, b. bosch, a. llado, c. pena-gomez, a. antonell, f. castellanos-pinedo, n. bargallo, j.l. molinuevo, r. sanchez-valle, journal of alzheimer's disease 36 (2013) 165-175. [85] n. ertekin-taner, n. graff-radford, l.h. younkin, c. eckman, m. baker, j. adamson, j. ronald, j. blangero, m. hutton, s.g. younkin, science 290 (2000) 2303-2304. [86] w.k. scott, j.m. grubber, p.m. conneally, g.w. small, c.m. hulette, c.k. rosenberg, a.m. saunders, a.d. roses, j.l. haines, m.a. pericak-vance, the american journal of human genetics 66 (2000) 922-932. ©2015 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ manuscript doi: 10.5599/admet.4.2.317 114 admet & dmpk 4(2) (2016) 114-116; doi: 10.5599/admet.4.2.317 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index opinion is there a paradigm shift in use of microsomes and hepatocytes in drug discovery and development? sumit basu, abdul naveed shaik* center for pharmacometrics and systems pharmacology, department of pharmaceutics, university of florida. orlando fl. *corresponding author: e-mail: naveedshaik@gmail.com; tel.: +1-407-313-7009. received: june 13, 2016; revised: june 20, 2016; published: june 29, 2016 in last two decades, the field of drug metabolism and pharmacokinetics (dmpk) has been considerably improved due to various reasons, i.e., routine incorporation of dmpk into early drug discovery, better understanding of drug dispositional elements (drug metabolizing enzymes and uptake/efflux transporters) and implementation of optimized dmpk properties in preclinical drug discovery and development. in this regard, both in vitro and in vivo experimental tools play significant role to understand different critical absorption, distribution, metabolism and elimination (adme) parameters which accelerated tremendously the lead selection, lead optimization and the preclinical development of the drug candidate prior to the clinical trials. among these experimental tools, in vitro assays are quite advantageous in comparison to the in vivo models because of its convenient and economical way to address the specific questions in short time period. these in vitro experimental models are often employed to determine the nature of enzyme induction/inhibition (mainly cyps) in drug-drug interaction (ddi) studies, metabolite identification, reaction phenotyping and elucidation of the biotransformation/metabolic pathway of new chemical entities (nces) [1, 2]. although it has been shown that due to species related differences, preclinical species cannot always predict human cyp induction/inhibition; nonetheless a good in vitro-in vivo correlation has always been observed when human based in-vitro assays have been employed [3]. in case of metabolite identification and enzyme kinetic studies, the best way to investigate the metabolic profile of the drug in a certain tissue is to procure the tissue specimen and evaluate the activity. in case of liver, this particular method has been applied successfully and undoubtedly this is the method of choice to gain information on hepatic drug metabolism. however, in order to extrapolate the research findings in human, we need to consider additional factors such as ethics, sampling procedure and bioanalysis etc. in addition, although a normal liver is imperative for the research purpose, in reality most of the human liver material is pathological to some extent. depending on what kind of questions we want to address, a variety of different preparations can be used, such as a fully physiological preparation (e.g. a whole perfused liver) or a strictly biochemical preparation (e.g. cellular sub-fractions such as microsomes) or a compromise situation (e.g. liver slices, cubes and hepatocytes etc.). among all the established in vitro experimental tools, microsomes are the mostly widely used http://www.pub.iapchem.org/ojs/index.php/admet/index admet & dmpk 4(2) (2016) 114-116 use of microsomes and hepatocytes in drug discovery and development doi: 10.5599/admet.4.2.317 115 subcellular fractions for drug metabolism studies as it is inexpensive and easy to handle [4]. microsomal assays are the default assays for metabolism and ddi studies at the drug discovery stage [5]. in order to determine intrinsic clearance of nces, which is a useful parameter for screening process of stable compounds and for establishing an in vitro correlation between animals and humans, microsomal assays play significant role. in addition, this assay can also be utilized to assess extra-hepatic metabolism and determine the ddi potential of the drug candidate in preclinical phase [6]. apart from liver, microsomes can also be utilized to focus on the region-specific activity of the intestinal enzymes. in general, the experimental conditions utilizing microsomes can be easily controlled in comparison to in vivo and in situ models. moreover, these methods are theoretically applicable to all species and can be used to study phase i and phase ii drug metabolism to get the enzyme isoforms activity profile [7, 8]. there are few limitations of microsomes, for instance, they have to be supplied with co-factors like nadph, or udpga for the metabolic reactions to initiate, and lack the cell membranes to mimic the physiological environment in hepatocytes [7, 8]. in addition, it suffers from its non-physiological nature due to the absence of cellular metabolism, membrane transport and production of co-factor [3, 9]. furthermore, this tissue preparation cannot provide the information on the activities of the transporters along the intestine, which can be critical for the investigation of gut metabolism to rationalize the absorption and permeability of the compound. the preparations of intestinal microsomes are more prone to artefacts in comparison to the preparation of the liver microsomes. several attempts have been made to reach a mid-point between the physiological complexity and biochemical ease. for that effort, liver slices as well as isolated cells have been used. liver slices have the physiological cell to cell contact needed for the liver preparation but are somewhat unsatisfactory due to the inability to get access of the substrate and nutrient to the centre of the preparation. in comparison, isolated liver cells are regarded as the best alternate, as they can be stored in cultures for extended periods of time without losing their differentiated functions. due to the presence of various uptake or efflux transporters in hepatocyte cell membranes, this particular subcellular fraction can be used to determine the role of transporter on the disposition of nce in preclinical discovery phase. the primary culture of hepatocytes provides a physiological environment that closely mimics the in vivo conditions as it contains enzyme and co-factors at physiological concentrations. in drug discovery, hepatocytes are mainly used as sandwich cultures or suspensions which are primarily used for the study of metabolic stability or transporter-mediated uptake of drug candidates. like microsomes, hepatocytes can also be used in ddi predictions and in determining intrinsic clearance. however, the major application of cultured hepatocytes is in cyp induction studies as enzyme induction may involve gene transcription and translation [10]. despite hepatocytes playing a significant role in drug discovery, their application is to some extent limited because they are not readily available and difficult to reproduce the results due to high inter individual variability between human liver donors. of late, freshly isolated, pooled cryopreserved hepatocytes are being used to circumvent this problem which allows researchers to use the same lot of cells over a period of time. however, improving long term culture of hepatocyte has been an area of research focus in the last decade. in the cell system, drugs never have easy access to the drug metabolizing enzymes. they have to pass through barriers containing drug transporters which facilitate the movement of these drugs across the membrane either concentrating the cell via influx transporters, or dispersing the drug through efflux transporters denying access of enzymes to the drug [11]. due to the involvement of these drug transporters, in drug disposition and ddis, there has been major interest in the past decade to study the role of these transporters in drug discovery and development. hepatocytes have a cellular organization basu and shaik admet & dmpk 4(2) (2016) 114-116 116 containing the necessary co-factors and membranes transporters, which represent a more complete system to study drug metabolism. due to the high cost and the complexity in conducting experiments with hepatocytes, microsomes still remain the tool of choice for early drug discovery studies. due to the advancement of technology, commercial companies and research organizations can now obtain human livers, make and store deep frozen slices and isolated hepatocytes in research and drug testing. from the literature studies it is quite evident that instead of following a general method to procure and utilizing the tissue, a case by case approach has always been taken depending on what kind of questions we would like to answer, physiological or biochemical. for an example, in order to measure the change of phase i or ii drug metabolizing enzyme content, the usage of microsomes or other subcellular fraction is appropriate. on the other hand, to assess the impact of transporters on the disposition of nces in discovery phase as well as to investigate the presence or nature of hepatotoxicity, hepatocytes or whole liver preparation will be more useful. reverting back to the title of this opinion “is there a paradigm shift in use of microsomes and hepatocytes in drug discovery and development?”; based on the literature, even though utilization of hepatocytes in drug discovery and development has been increasing in the last decade, the choice of in vitro tool mainly depends on the type of study performed. references [1] a.s. kalgutkar, r.s. obach, t.s. maurer, current drug metabolism 8 (2007) 407-447. [2] a.n. shaik, t. bohnert, d.a. williams, l.l. gan, b.w. leduc, journal of pharmaceutical sciences 105 (2016) 1976-1986. [3] r.s. obach, drug metabolism and disposition: the biological fate of chemicals 27 (1999) 1350-1359. [4] d. zhang, g. luo, x. ding, c. lu, acta pharmaceutica sinica b 2 (2012) 549-561. [5] a.n. shaik, r. grater, m. lulla, d.a. williams, l.l. gan, t. bohnert, b.w. leduc, journal of chromatography b, analytical technologies in the biomedical and life sciences 1008 (2016) 164-173. [6] a.n. shaik, m.y. mahat, k. kandasamy, b.w. leduc, a.a. khan, azoles as novel inhibitors of monoamine oxidase (mao) a and b isozymes, drug metabolism reviews, taylor & francis ltd 4 park square, milton park, abingdon ox14 4rn, oxon, england, 2015, pp. 204-204. [7] b. wu, k. kulkarni, s. basu, s. zhang, m. hu, journal of pharmaceutical sciences 100 (2011) 36553681. [8] b. wu, s. basu, s. meng, x. wang, m. hu, current drug metabolism 12 (2011) 900-916. [9] d.m. cross, m.k. bayliss, drug metabolism reviews 32 (2000) 219-240. [10] g. luo, m. cunningham, s. kim, t. burn, j. lin, m. sinz, g. hamilton, c. rizzo, s. jolley, d. gilbert, a. downey, d. mudra, r. graham, k. carroll, j. xie, a. madan, a. parkinson, d. christ, b. selling, e. lecluyse, l.s. gan, drug metabolism and disposition: the biological fate of chemicals 30 (2002) 795804. [11] s. gao, s. basu, z. yang, a. deb, m. hu, current drug targets 13 (2012) 1885-1899. ©2016 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ manuscript doi: 10.5599/admet.4.1.276 1 admet & dmpk 4(1) (2016) 1-22; doi: 10.5599/admet.4.1.276 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index review an overall comparison of small molecules and large biologics in adme testing hong wan shanghai hengrui pharmaceutical co., ltd. dept. of dmpk/tox, shanghai, p.r. china e-mail: wanh@shhrp.com; tel.: +86(0)21-54759066-1313 received: march 12, 2016; revised: march 23; published: march 31, 2016 abstract biologics mainly monoclonal antibodies (mabs) and antibody-drug conjugates (adcs) as new therapeutics are becoming increasingly important biotherapeutics. this review is intended to provide an overall comparison between small molecules (sms) and biologics or large molecules (lms) concerning drug metabolism and pharmacokinetic (dmpk) or associated with absorption, distribution, metabolism and elimination (adme) testing from pharmaceutical industry drug discovery and development points of view, which will help design and conduct relevant adme testing for biologics such as mabs and adcs. recent advancements in the adme for testing biologics and related bioanalytical methods are discussed with an emphasis on adc drug development as an example to understand its complexity and challenges from extensive in vitro characterization to in vivo animal pk studies. general non-clinical safety evaluations of biologics in particular for adc drugs are outlined including drug-drug interaction (ddi) and metabolite/catabolite assessments. regulatory guidance on the adme testing and safety evaluations including immunogenicity as well as bioanalytical considerations are addressed for lms. in addition, the preclinical and human pk data of two marked adc drugs (adcetris, sgn-35 and kadcyla, t-dm1) as examples are briefly discussed with regard to pk considerations and pk/pd perspectives. keywords monoclonal antibody; antibody-drug conjugate; pharmacokinetics; bioanalysis; preclinical safety 1. introduction biologics or large molecules (lms) primarily monoclonal antibodies (mabs) and antibody-drug conjugate (adc) currently represent main stream therapeutics and continue to grow in number of new approvals and targets recently [1-4]. a majority of these biotherapeutics are mabs or mab-derived proteins, which are subject to transformation mechanisms such as deamidation, oxidation, and isomerization [1-4]. these processes usually result in relatively small structural changes in the parent drugs. such small structural changes may be difficult for a conventional immunoassay to differentiate, but they can still affect biological activity, pk and immunogenicity of a therapeutic protein [5]. whereas adcs are an emerging class of biotherapeutics that combines the target specificity of an antibody with the potent small-molecule drugs or cytotoxins, which can selectively deliver a potent cytotoxic drug to tumor cells via tumor-specific and/or over-expressed antigens with more favorable therapeutic window [1,6]. this new type of antibody-drug conjugate or antibody-linker-drug currently shows its great promising therapeutic options, which led to the recent fda approvals of adcetris (brentuximab vedotin, sgn-35) and kadcyla (ado-trastuzumab http://www.pub.iapchem.org/ojs/index.php/admet/index hong wan admet & dmpk 4(1) (2016) 1-22 2 emtansine, t-dm1) for the treatment of hodgkin’s lymphoma (hl) and anaplastic large-cell lymphomas (alcl), and her2-positive metastatic breast cancer, respectively, as well as a rich clinical pipeline of potential new cancer therapies [7]. throughout several decades of advancements and evolutions, admet profiling of small molecule (sm) has becomes a standardized paradigm in drug discovery and development in terms of in vitro screening, in vivo animal studies, lc-ms/ms based bioanalysis as well as regulatory considerations including ddi and drug metabolite safety testing etc. however, adme testing of lms lags behind that of sms due to the complex nature of the biological molecules and also lack of appropriate tools to study drug exposure, biotransformation and target engagement in the vascular and tissue spaces [2]. adme of lms or biologics are still based on the lessons learned from the sms and the tools that have applied to sm drugs [5]. in general, there is a similar high level pk/pd relation concept between the sms and lms, although they have different adme mechanisms and underlying admet determinants at different stages [5]. characterizing the absorption, distribution, metabolism, and excretion of these lm drugs (mabs and adcs) in preclinical animal models can better predict their efficacy and tolerability in clinic. accordingly, it is necessary to understand general characteristics and the difference between sms and lms (e.g., antibody and adc) in order to apply relevant approaches for adme testing and safety evaluation of lms as reviewed in a number of recent publications [2,3,5,8-11]. in a previous review, what adme tests should be conducted for small molecule drugs for preclinical studies was highlighted [12]. this review will provide an overall comparison between sm and lm properties with a particular focus on adc’s characteristics to gain a better understanding of in vitro and in vivo adme testing as well as toxicity evaluations. the in-depth information of adme will also be valuable for the designing of novel mab constructs and next generation of adcs with desirable pk profile and safety window. 2. general differences between sms and lms (mab and adc) in adme testing table 1 summarizes general differences between sms and lms with an overall comparison of mab and adc. as highlighted in table 1, due to the nature of various characteristics between sms and lms, the focus on adme studies of lms is thus different from sms in particular for drug metabolite safety testing and ddi evaluations. in these aspects, in vitro models and in vivo studies and related bioanalysis including transporter studies and safety evaluation and high-throughput screening approaches for sms have well been established across pharmaceutical industry [12-20]. typical in vitro adme tests for sms are metabolic stability by liver microsmoes or/and hepatocytes and passive permeability on cell-line models based on caco-2 or mdck assays, which are commonly utilized to predict in vivo clearance and absorption or bioavailability as well as potential drug-drug interaction (ddi) evaluations and metabolism pathway studies. in particular, a common consensus has been reached on drug metabolite testing across various species from in vitro models to in vivo studies following regulatory guidance, which clearly suggests critical criteria for decision-tree making in assessing key drug metabolite safety in human [21,22]. similarly, cytochrome p450 (cyp) enzymes based ddi as well as transporter-mediated ddi of sms are also defined and guided on the basis of many years industrious practice of drug discovery and development [23]. on the contrary, adme testing of biologics (mabs and adcs) can be rather diverse as highlighted in table 1 and figure 1. biodistribution of mabs and adcs are usually similar [24], but both have much lower vd than that of a typical sm drug as the structure of mabs or adcs is dominated by the antibody backbone with initial distribution limited to the vascular space or plasma, not organ tissues. while the drug metabolism of mabs is more complex due to receptor binding target-mediated drug disposition (tmdd), fcrn binding and lysosomal degradation, tissue protease, immune response antibody mediated metabolism etc in addition to admet & dmpk 4(1) (2016) 1-22 adme testing of large biologics doi: 10.5599/admet.4.1.276 3 other metabolic or metabolic-like biotransformation such modifications as deamidation, oxidation, isomerization, disulfide bond reduction or shuffling and proteolytic or glycolytic hydrolysis (ref. therein) [11]. the metabolism /catabolism of adcs can be more complicated than mabs due to a cytotoxic drug linked to the antibody via a linker. typically, the circulating unconjugated drug after the adc administration has metabolic properties of small-molecule compounds. hence, drug metabolite and ddi should be concerned for adc drugs, but maybe not for mabs wherein the ddi risk is presumably low or not as prominent as small molecules. once small-molecule drugs released from the adcs, they may be metabolized by cyp enzymes and thus subject to potential ddi from cyp enzyme inhibitors or inducers due to payload/small molecule component [9]. furthermore, one or more active small-molecule drugs may be released from an adc in vivo by additional catabolism mechanism. accordingly, both unconjugated smallmolecule drug and released catabolites small molecules may be considered for metabolite safety and ddi potential evaluations for adcs. this is considered to be a major differentiation of adcs from mabs regarding adme testing as the degraded small molecules of mabs are often amino acids, small peptides or small carbohydrates that are readily eliminated by renal excretion or return to the nutrient pool without biological effects or safety consideration [2]. however, it should be noted that examples of therapeutic protein (tp) and small-molecule drug (d) interactions in clinical studies were observed although the changes in exposures (auc, cmax) have not been as remarkable as with small-molecule drugs, and the types of study designs used to assess tp and d interactions were thereby outlined in cder’s special subject on therapeutic protein–drug interactions and implications for drug development [25]. on the other hand, for lms at early stage, more extensive in vitro characterizations are required as an example of screening cascade depicted in figure 1. these typical studies include antibody primary structure by chemical sequencing (edman degradation) as well as peptide mapping by means of esi-ms and maldims [26-28]; higher order structure by rp-hplc–esi–ms, ellman’s assay cd, ftir, hydrogen deuterium exchange (hdx)-ms and x-ray [28,29]; post-translational modifications (ptm) and charge variants by msion-exchange chromatography (iec), hydrophobic interaction chromatography (hic), capillary electrophoresis (ce) and isoelectric focusing (ief) [30]; glycan profile and variants and size heterogeneity by hplc-fluorescence, size-exclusion chromatography (sec), native gel, capillary electrophoresis, rp-hplc-ms and native intact ms [31], as well as solubility measurement by ultrafiltration or peg-induced precipitation methods [32,33]. moreover, the immunogenicity should be evaluated for both antibody and adcs [34,35], but unnecessary for sms. unlike the antibody, additional in vitro studies of adc include drug antibody ratio (dar) characterizations, conjugate site analysis, linker stability and toxin evaluations as highlighted in figure 1 may be required. furthermore, absorption needs to be understood as well if lms are targeting for an sc administration. despite distinct characteristics and differences of in vitro testing between antibody and adcs and sms as shown in table 1 and figure 1, similar principle of in vivo pk studies including mass balance and elimination as well as safety evaluations for sms are generally applicable to both antibody and adcs. however, different bioanalytical methods have to be applied for antibody and adcs, which will be discussed in more details in bioanalytical section. hong wan admet & dmpk 4(1) (2016) 1-22 4 table 1. characteristic comparison of small molecules (sms) and large molecules (lms)* *modified from references [8-10]. sms biologics or lms mabs adc mw ~ 200-500 ～150 kda (typically) ～150 kda (typically) test article chemical protein conjugate (protein-chemical) physicochemical properties mostly well-defined physicochemical properties complex physicochemical properties (e.g. tertiary structure, stability, ptm) antibody + toxin adme tools available/ extensive adme understanding understanding of adme still evolving understanding of adme still evolving dosing route oral often possible usually parenterally (iv, sc, and im), intravitreal injection usually parenterally (iv, sc, and im) dose interval daily (typically) intermittent dosing intermittent dosing half-life ( t1/2) short (typically several to 24 hrs) long (typically days or weeks) long but slightly shorter than antibody alone distribution (vd ) high vd, distribution to organs/tissues potential substrate of transporters lower vd , usually limited to plasma and/ or extracellular fluids lower vd , mainly target dependent metabolism pathway mainly by cyp enzymes and phase ii enzymes, metabolized to nonactive and active metabolites catabolism degraded to peptides or amino acids both catabolism and metabolism drug metabolite safety evaluation yes no toxin and released drug catabolites (may be concerned) excretion mainly biliary and renal excretion mostly recycled by body both clearance (cl) mostly linear pk; non-linearity mainly due to saturation of metabolic pathways slow clearance lower clearance, but slightly higher clearance than antibody potency and selectivity generally less selective high selectivity (affinity/ potency) high selectivity (affinity/ potency) pk analytes drug and metabolites antibody and ada adc or conjugated drug, total antibody, and unconjugated toxin/catabolites, ata pk bioanalysis lc-ms/ms methods mostly elisa (total antibody), recently with increased lcms/ms applications hybrid of elisa (total antibody, adc), high resolution lc-ms (dar) and highly selective lc-/ms/ms (un-conjugated toxin/catabolites) pd short acting long acting long acting pk/pd pk usually not driven by pd due to dominance of non-target mediated binding pk and pd mechanistically connected (tmdd) pk and pd mechanistically connected (tmdd) ddi many examples and pk and/or pd related (by cyp enzymes or transporters) sparse examples and mostly pd related sparse examples and mostly pd related toxin or released toxin/catabolites (should be concerned) herg yes no no immunogenicity no yes yes toxicity on-and off-target related toxicity typically exaggerated pharmacology typically antigen-independent can be more toxic than antibody due to toxin formulation complex and diverse simple formulation simple formulation api/production process synthesized ( uniform single entity) generics, bioequivalence culture-derived (generally nonuniform) no generics, biosimilar or comparability both + conjugation (mixtures of adcs with different dar no generics, biosimilar or comparability admet & dmpk 4(1) (2016) 1-22 adme testing of large biologics doi: 10.5599/admet.4.1.276 5 figure 1. typical adme/tox studies for sm and antibody and adc 2.1. adme testing of mabs fundamental adme characteristics and pk behaviors of biologics including therapeutic proteins and mabs have extensively been reviewed recently [2,11,25,36-39]. currently, there is lack of validated in vitro adme models for mabs. in addition to extensive characterization of antibody as depicted in figure 1, the major adme testing of mabs relies on in vivo studies such as pk and safety evaluations on relevant animal species. it should be noted that the selected species for in vivo pk and toxicity evaluations should preferably have cross-reactivity with human. another important aspect to consider is immunogenicity for adme testing of mabs as well as adcs, which is a key risk in biologic drug development, because therapeutic biotechnology-driven proteins must be as similar as possible in primary, secondary, and tertiary structures to its human homologues in order to avoid immunogenicity [2,40]. in general, the immunogenicity, specifically formation of anti-drug antibody (ada), is one of the major complicating issues in nonclinical and clinical programs for therapeutic proteins, which is a regulatory requirement as part of the safety assessment of biotherapeutic submission [34,38,41]. the ada production due to the continuous administration of mabs may affect the pk and/or the pd of mabs after they are administrated to patients [40]. when an immunogenic reaction mediated by adas takes place, it can increase blood clearance and decreases exposure (often relatively shorter half-life observed), thereby reducing its efficacy and likely in vitro potency/binding solubility metabolic stability permeability/transporter logd7.4 in vivo pk ppb cyp inhibition metabolite prof iling and id pk/pd toxicity and saf ety margin small molecule antibody adc excretion/mass balance preclinical candidate (pcc) in vivo pk & immunogenicity pk/pd pk/pd toxicity and saf ety margin toxicity (adc, toxin) dar distribution (in vitro) tissue distribution/mass balance metabolites/catabolites stability of linkers (plasma) aggregation & f ragmentation solubility primary structure & sequence oxidation and deamidation glycan prof ile &variants disulf ide bonds analysis peptide mapping analysis antibody formulation stability post-translational modif ications analysis free toxin, ddi charge variants aggregation & f ragmentation charge variants peptide mapping analysis conjugate site analysis in vivo pk & immunogenicity dar distribution (in vivo) plasma stability ddi evaluation hong wan admet & dmpk 4(1) (2016) 1-22 6 hampering its clinical application in some instances. although factors that affect ada are complex and the relationship of bioanalytical measures, for example, incidence and magnitude of ada responses to clinical efficacy and safety are uncertain, the ada incidence and magnitude should always be assessed relative to capacity of adas to neutralize the relevant biological activity of the therapeutic mab, in addition to impact on drug levels and clinical measures of efficacy and safety [42]. in principle, it would be acceptable for the biologic product to be less immunogenic than the reference product, provided that this did not modify the efficacy of the product or increase the incidence or severity of adverse reactions, although the magnitude of ada acceptance is not defined by regulators [42]. one should also bear in mind that the bioanalytical methods used to monitor immunogenicity are subject to various biases when interpreting ada data and assessing immunogenicity. in practice, a comparison of single dosing pk and multiple dosing pk with monitoring ada production over time is valuable for a better understanding of the pk/pd relation of mabs as well as the part of mab quality evaluation. to the best of our knowledge, once the ada occurs over the time, the plasma concentration at certain time point can drop dramatically. consequently, caution has to be taken to evaluate the pk parameters in particular calculate the half-life of mab with different compartmental models to be in line with its corresponding plasma concentrations for the correct interpretation of pd effect or its pk/pd relationship. additionally, mathematical modeling of ada response against mabs and its impact on mab pk/pd has been reviewed recently [40]. currently, ligand binding assay (lba) such as enzyme-linked immunesorbent assays (elisa) are generally regarded as the gold standard methods for bioanalysis of lms with sufficient sensitivity and specificity. of them, the sandwich and competition enzyme immunoassay technique are more widely utilized for detection and quantification of antibodies [43-45]. typically, various coat capture reagents, conjugate reagents (e.g. horseradish peroxidase, hrp), peroxidase substrates (e.g. tmb, stabilized hydrogen peroxide/tmb) and stop solution (e.g. sulfuric acid) are employed case by case. to date, the standard elisa platform has considerably been improved upon over the years, involving automation and improved detection methodologies, and new applications of this technique is further discussed in next bioanalysis section. 2.2. adme testing of adcs adcs are mabs with covalently bound cytotoxic drugs via a linker, which are designed to target tumor antigens selectively and offer the hope of cancer treatment to decrease the off-target toxicity, thereby, improving the therapeutic index of the cytotoxin [41,46]. adcs are usually comprised of 0 to 8 cytotoxic payloads with an average dar of 2 to 4 per mab, which are thus heterogeneous mixtures of conjugates. given that low payloads reduce potency and high payloads negatively impact pk, the dars can have a significant impact on adc efficacy [3]. structurally, the antibody component of the adc accounts for the majority of the therapeutic agent (approximately 98 % of total adc by molecular weight). biologically, the pk of adcs is strongly influenced by the underlying antibody backbone conferring properties such as target specific binding, neonatal fc receptor (fcrn)-dependent recycling and fc (fragment, crystallizable) effector functions [24]. compared with the unconjugated antibody, adcs can exhibit somewhat higher clearance due to introduction of an additional metabolic pathway (i.e. cleavage of the drug from the antibody by lysosomal peptide degradation). furthermore, linker stability, dar and site of drug conjugation can largely impact pk behaviors and distribution as well. conjugation with higher dar tends to have faster clearance than conjugation with lower dar as an example of cac10-vc mmae adcs with high dars exhibiting a faster total antibody clearance than lower dar adcs [47]. variable dars and attachment sites by a consequence of current random conjugation methods result in heterogeneous adcs with pk parameters that can vary substantially compared to the unconjugated antibody [24]. hamblett et al. compared adcs with different admet & dmpk 4(1) (2016) 1-22 adme testing of large biologics doi: 10.5599/admet.4.1.276 7 drug loading (mmae conjugated to the anti-cd30 mab) and concluded that the therapeutic index was increased by decreasing drug loading per antibody, demonstrating that drug loading is a key design parameter for antibody-drug conjugates [47]. it was shown that the terminal half-lives did not directly correlate with drug loading. when the drug loading increased, both clearance and volume of distribution increased, but the exposure (auc) of adc decreased, reflecting the significant attribute of dar to pk properties of adc. the systemic stability of the antibody−drug linker is also crucial for delivery of an intact adc to targetexpressing tumors. linkers stable in circulation but readily processed in the target cell are necessary for both safety and potency of the delivered conjugate [48]. in general, linker should be stable enough in systemic circulation to deliver drug into tumor cells, but labile to release this fully active payload drug inside the tumor cells. otherwise, the toxic drug could be released in the bloodstream, causing systemic toxicity. polson et al. demonstrated that non-cleavable linker (e.g. sgn-35) is more stable than cleavable linker (e.g. t-dm1) [49]. the non-cleavable linker may have slower deconjugation or lower clearance than cleavable linker and reduced systemic toxicity of adc in rats presumably due to the reduced release of free drug or other toxic metabolites into the circulation [49]. this is because the cleavable linkers may release drug by lysosomal proteases without the degradation of the mab component whereas non-cleavable linkers require catabolism of the mab backbones to release the drug [9 and ref. therein]. in addition, the selection of attachment site can affect stability of cleavable antibody-drug conjugates, as demonstrated in the design of systemically stable cleavable auristatin-based conjugates, offering the means to overcome inherent linker instability by optimal attachment site [48]. interestingly, junutula et al. [50] recently reported a site-specific conjugation method called thiomab technology enabling near-uniform, low-level conjugation of cytotoxins (dar=2) to antitumor antibodies, which displayed an increased tolerability or improved therapeutic index without compromising efficacy. this new type of thiomab-drug conjugates (tdcs) with nearly homogeneous composition of conjugates was better tolerated than conventional adc in monkeys, and also showed slower catabolism and deconjugation than adc in rats. the thiomab approach provides a general strategy for improved pk and superior therapeutic index of any adc generated by standard sufhydryl-directed maleimide chemistry [51]. likewise, dennler et al. more recently presented another site-specific modification of deglycosylated antibodies by microbial transglutaminase (mtgase) to form adcs with a defined homogeneous dar of 2 as well. this chemo-enzymatic approach was reported directly applicable to abroad variety of antibodies as it does not require prior genetic modifications of the antibody sequence [52]. clearly, the designing and forming a homogeneous adc can result in less complicated in vitro characterization and relatively easier in vivo pk bioanalysis as well as safety evaluation than a typical heterogeneous adc. apart from the dar characterization, some in vitro adme models and in vivo pk studies can be employed for adc evaluations as illustrated in figure 1. first adme testing of adcs is linker stability assessment both in in vitro or in vivo by lc-ms/ms method or radiolabeled adc, respectively, in order to explore the impact of linker on uptake and catabolism. secondly, free toxin was expected to undergo clearance mechanism studies for ddi potential including liver microsomal stability, metabolizing enzyme identification, cyp inhibition and induction, in vitro metabolite profiling studies [53]. then, general in vivo pk studies and non-clinical safety evaluations including immunogenicity with relevant animal species are necessary for pcc selection. finally, linker-containing drug catabolites should be identified and the plasma concentration of free drug and related catabolites should be determined by lc-ms/ms in both nonclinical and clinical studies for safety assessment of key metabolites or catabolites. hong wan admet & dmpk 4(1) (2016) 1-22 8 elisa was conventional method for pk/tk (toxicokinetics) study of antibody and conjugate. for instance, detection reagents for total and conjugated antibodies were goat antihuman igg antibody conjugated to horseradish peroxidase or goat antihuman fc conjugated to horseradish peroxidase [43]. coat capture reagents were used on a case by case basis, e.g. anti-mmae antibody (conjugated antibody), antibodies specific for drugs (total antibody). biotin-dig bridging elisa and surface plasma resonance (spr) techniques were developed for immunogenicity assessment to detect anti-therapeutic antibodies (atas) to the adc, including atas towards any of the adc molecular components such as the antibody, linker, drug or epitopes involving multiple adc components. finally, isotope-labelling method was used for the study of tissue distribution and mass balance [43,46,54]. recent advances in high-throughput formats and combined techniques for adc bioanalysis will be discussed in more details in the section 3. 2.3. safety evaluation of adcs with regard to preclinical safety assessment, toxicity of adcs is usually adc/drug-dependent and antigen-independent and it can be more toxic than mabs due to introduced cytotoxic drugs. as adcs consist of a monoclonal antibody, linker and cytotoxic components, the biological activity profiles of each should be considered when selecting the relevant and/or appropriate species for toxicity evaluations, typically in two relevant species (one rodent and one non-rodent) [55]. for a particular adc, the relevant toxicity species should be selected from its pharmacology and tissue-reactivity studies showing similar binding to human. primary considerations for the nonclinical safety assessment of adcs includes the evaluation of the entire adc or the various individual components (i.e., antibody, linker or the cytotoxin) to identify the onand off-target toxicities to enable first-in-human (fih) studies [41]. general toxicity studies are recommended and following specific toxicity studies may be considered as well for overall nonclinical safety evaluations [25,35,41,42,56,57]:  acute and chronic toxicity study on two relevant species (preferably rodent and non-rodent), or transgenic animals (i.e., animal modified to the human target) to attain maximum tolerated dose (mtd), non-observed adverse effect level (noael) or the highest non-severely toxic dose (hnstd).  tissue cross-reactivity (tcr) in tissue panels from humans and toxicity species for adc and unconjugated antibody  genotoxicity of cytotoxin and/or linker, and adc may be concerned  herg for unconjugated drug if concentration of the unconjugated cytotoxin are detected in the serum during toxicity and plasma stability testing  reproductive and developmental toxicity for embryofetal (efd) toxicity assessment of adc and unconjugated toxin  carcinogenicity  immunogenicity and anti-adc antibody  cyp inhibition and induction for ddi assessment  hemolysis, skin irritation and sensitization testing for biologics via injection administration. another important consideration would be the concentrations of unreacted linker that remain in the final drug product. if the cytotoxin is a novel chemical entity, then a more detailed assessment of metabolism and disposition (including pk/adme studies) may be necessary as the cytotoxic drug released from the adc may be associated with loss of efficacy or increased toxicity. however, if the concentrations of unconjugated cytotoxin in plasma are very low, for instance, the free dm1 after administrating the trastuzumab emtansine (t-dm1) was <10 ng/ml (<10 nm) at all doses [53], or maximal dm1 concentration did not exceed 25 ng/ml after repeated dosing of t-dm1 [58], the investigation of inhibition or induction of admet & dmpk 4(1) (2016) 1-22 adme testing of large biologics doi: 10.5599/admet.4.1.276 9 cyp isoenzymes to assess the potential ddi should not be necessary in this case [41]. in other words, the risk of ddi is presumably low due to very low concentrations of cytotoxic drug released from adcs. despite this hypothesis, cyp3a4 based inhibition studies for kadcyla and decetris have been conducted, respectively, as the sgn-35 is metabolized with a small fraction of mmae via oxidation by primarily cyp3a4 and cyp3a5, and kadcyla (t-dm1) mainly by cyp3a4 and a lesser extent by cyp3a5. nevertheless, an in vitro stability of adcs in plasma and in vivo monitoring toxin released from adc should be conducted to ensure the cytotoxic drug or potential catabolites under the minimal level as such without safety issues, in particular for major cyp metabolism involved adc drugs. on the other hand, a typical adc consists of a distribution of several entities containing different numbers of cytotoxic drugs on antibody, linked at different positions, thus, the inherent heterogeneity of adc complex with multiple components still remains a prominent challenge in understanding their properties in vivo during nonclinical development. for instance, conjugation through interchain disulfides can lead to these adcs with dar distribution ranging from 0 to 8, with each fraction potentially exhibiting a unique efficacy, pk property, and toxicity profile [43]. therefore, development of more sensitive and specific bioanalytical methods to differentiate a single adc and multiple adcs would be highly valuable to enable quantifying adc complex for a better understanding of attributes of each single adc component to pk and pd as well as safety profiles. although it would be ideal to monitor each individual adc species of a specific dar, this currently seems to be a technical challenge. therefore, the most common approach is to monitor all adc species, all antibody species or total antibody (tab), and the unconjugated small molecule. an alternative approach is to monitor conjugated small molecules as a surrogate for adc [9]. overall, a design goal of an adc is to maximize delivery of the cell-killing agent to the tumor tissue while minimizing delivery to normal tissues [54]. the choices of mabs, linker, toxin or cytotoxic drug are all important determinants of pk, efficacy and safety. as a result, design and conduct of relevant adme testing and fronting loading pk/pd studies as well as the key safety evaluations of lead molecules as early as possible will facilitate pcc’s selection for their successful development. 3. bioanalytical methods and challenges in the past several decades of development of sms, the bioanalytical strategies and approaches of sms have been well established from method development, validation as well as method transfer from instrument to instrument and lab to lab based on lc-ms/ms according to fda bioanaltical method guidance [59]. differently from sms, the most commonly used bioanalytical methods for lms are elisa based assays by either direct or bridging elisa measuring the concentration of the therapeutic over time in plasma or serum for in vivo studies case by case. in comprehensive reviews and white papers, several bioanalytical methods, considerations and strategies as well as challenges have extensively been discussed for lms [42,46,60-66]. this section will provide an overview on recent advancements in bioanalysis of mabs, adc, in particular with elisa and lc-ms/ms technologies and related new approaches dealing with matrix interference. 3.1 high-throughput bioanalysis from industry drug screening perspective, commercially automated meso scale discovery (msd) format and nonoscale immunoassay platform on gyrolab offer high-throughput capabilities of elisa bioanalysis, which enhances overall performance of traditional manual elisa assay with several advantages in terms of sample amount, broader dynamic linearity, throughput, higher sensitivity and reproducibility [67,68]. moreover, both the antibody and ada of an antibody or adc and ata from pk plasma samples can be hong wan admet & dmpk 4(1) (2016) 1-22 10 quantified simultaneously. in particular, the nanoscale volume of microfludic platform on gyros has made it durable to gain the pk data of antibody from one single mouse study by whole blood serial sampling with significantly reducing study cost and animal usage, offering comparable pk data as conventional composite sampling [69]. furthermore, this microfludic approach has been applied to high-throughput quantification of host cell protein impurities for bioprocess development [70]. it demonstrated an improved throughput (5-10 times faster), broader dynamic range (100-times) and decreased sample consumption, hands on time and duration for assay development compared with tecan plate-based elisa [70]. with those automated instruments, elisa based bioanalytical methods become easier transferring from preclinical to clinical between labs and cros, which can considerably speed up large molecule drug discovery process and clinical development. furthermore, traditional elisa approach was transferred to an automated microfluidics immunoassay platform based on nanoscale streptavidin bead columns enable highthroughput bioanalysis of a human mab in preclinical pk samples with enhanced bioanalysis capacity [71]. additionally, a high-throughput capillary electrophoresis based microfluidic device (labchip gxii) was applied to obtain pharmacokinetics (pk) of a fluorescently labeled human mab directly in single dose rat pk studies [72]. 3.2 lc-ms/ms and bla bioanalysis of mabs as mentioned above, lba-based elisa is the commonly used gold standard method for bioanalysis of mabs. recent advances in instrumentation technology have significantly increased the sensitivity and versatility of lc-ms/ms, making it an alternative tool for large molecule drug development. with appropriate sample preparation (based on the mw and concentration of the target analyte of interest), immmunocapture lc-ms/ms methods can achieve levels of high sensitivity down to 0.01 to 0.1 nm closer to elisa assays with a large dynamic range and orthogonal specificity that is generally unaffected by crossreaction issues [61]. in recent comparative studies, peng et al. developed and validated a simple, sensitive, specific and precise lc–ms/ms assay for quantitation of infliximab pk in human serum by using isotopelabeled signature peptide as the internal standard (is), which achieved an excellent correlation with elisa detection [73]. park et al. applied lc-ms/ms quadruple time-of-flight mass spectrometric method (qtof) for the determination of trastuzumab in rat plasma with good sensitivity (llod 0.5 µg/ml) and results in consistence with elisa assay (r 2 =0.9104) [74]. zhang et al. reported a generic automated lc-ms/ms method for the pharmacokinetic study of a mab in cynomolgus monkey with comparable data by the immunoassay data as well [75]. li et al. developed a general lc-ms/ms method approach employing an uniformly heavy-isotope labeled common whole mab is and a common immunocapture for sample processing for qualification of four igg(2) and four igg(1) mabs. this general lc-ms/ms method approach overcomes the limitations of current methods to reduce time and resources required for preclinical studies [76]. the key benefits of lc-ms/ms over elisa based method are its high degree of specificity and its ability to simultaneously resolve, detect, and quantify multiple peptides in biological samples or their extracts, as well as even able to identify key metabolites of biologics, in-depth structural characterization and functional insights of therapeutic mabs, e.g., differentiating biosimilar to originator mabs as the case of trastuzumab and cetuximab, which will be valuable for biobetters and next generation antibodies design and optimization [77]. highly selective lc-ms/ms made it possible for simultaneous quantification of several co-administered human antibodies (mabs), mab-a and mab-b of igg4 subclass in cynomolgus monkey serum with lloq around 5-25 µg/ml [78], which cannot be obtained by elisa assay. as demonstrated in recent examples, xu et al. successfully applied a multiplexed hybrid lc-ms/ms pharmacokinetic assay to measure two co-administrated mabs in a clinical study without requiring admet & dmpk 4(1) (2016) 1-22 adme testing of large biologics doi: 10.5599/admet.4.1.276 11 stringent affinity capture reagents [79]. lebert and coworkers also demonstrated absolute and multiplexed lc-ms/ms method combining a strategy so-called psaq tm (protein standard absolute quantification or stable isotopically-labeled full-length mabs) as the iss for quantification of three mabs variants with very similar sequences, e.g., igg1, igg2 and igg4 isotypes of a lead mab in rat serum, which can be employed for both preclinical and clinical studies [80]. moreover, a versatile immunoaffinity lc-ms/ms method was developed to quantify total receptor activator of nuclear factor-kappab ligand (rankl) in the presence of denosumab, a humanized monoclonal antibody (mab) specific to rankl with an lloq of rankl down to 3.13 ng/ml for mouse plasma pk, which was not able to determine the total rankl because the interference of denosumab decreased the recovery of rankl with commercial elisa kit [81]. additionally, 2d-lc (lc×lc)–ms/ms methodology with improved ms detection limit of approximately two orders of magnitude over direct lc–ms/ms was recommended [82], as well as other lc-ms formats and approaches such as capillary lc-ms gaining ms sensitivity were reviewed [66], offering new opportunities in the analysis of monoclonal antibodies. on the other hand, mass spectrometry methods (maldi-tof, q-tof, ltq-orbitrap) have been reported for metabolite identification or profiling (metid) (e.g. deamidation of asparaginyl and glutaminyl residues), preclinical and clinical pk/tk studies with no cross-reaction [61]. direct analysis of therapeutic monoclonal antibodies, without size reduction through enzyme digestion, is more challenging since these molecules may have a molecular weight up to 150 kda, forming a large distribution of charge states during electrospray ionization (esi), thus resulting in a complex mass spectrum with decreased sensitivity. to overcome these difficulties, the strategy consists of decreasing their size by proteolytic digestion in order to form 1000 to 2000 da peptides with few charge states (between 2 and 4) which can be readily quantified in mrm mode on esi-mass spectrometers. in another recent protocol, detailed mass analysis of structural heterogeneity in monoclonal antibodies at the intact protein level under pseudo-native conditions using native mass spectrometry was reported [83]. this method can be used for different applications such as the analysis of mixtures of mabs, drug-antibody conjugates and the analysis of mab ptms, including the qualitative and quantitative analysis of mab glycosylation, and it offers several advantages in terms of speed, sensitivity and specificity and high-throughput analysis. overall, there is a need for general and reliable lc-ms assays capable of supporting the bioanalysis of a variety of human monoclonal antibodybased therapeutics such as reported universal peptide approach to the bioanalysis of human monoclonal antibody protein drug candidates in animal pk/tk studies [84]. 3.3 lc-ms/ms and bla bioanalysis of adcs due to the fact that the adcs are complex mixtures incorporating large and small-molecule characteristics, this poses more unique bioanalytical challenges for adc than mab to measure adcs and their catabolites in plasma and serum [46]. these challenges include adc bioanalysis such as quantifying adc and dar in serum/plasma for pk studies and strategies for assessing immunogenicity. since adc species with different dar distribution may display different potencies, the measured plasma concentrations may not accurately reflect the associated pd effect. as a general rule, both total antibody assay and conjugate antibody assay based on elisa are needed to measure plasma concentrations over the time. in addition, adc complexity may increase in in vivo due to biotransformation by catabolism or metabolism, leading to additional changes in dar or dynamically changing mixtures. therefore, a standard calibration curve consisting of the reference standard may not be appropriate for quantification of analytes in vivo. biotransformation may result in adc analytes in vivo that differ from those in the reference standard in vitro. a notable example was characterization of adc (trastuzumab emtansine) drug distribution in a cynomolgus monkey pk study by her2 extracellular domain affinity capture lc–ms, which hong wan admet & dmpk 4(1) (2016) 1-22 12 showed the adr distribution shifts to lower values over time, e.g., dar=2.76 at day 1, 1.86 at day 10, 0.68 at day 28, respectively [46]. apparently, the dar composition of adc in vivo can vary due to drug deconjugation or metabolic clearance. therefore, it is important to assess the reference standard calibration curve appropriately for all pk time points to gain accurate dar and pk data for a better understanding of pd effect. in this case, as suggested by the adc working group of aaps and genentech scientists, a comparative analysis of assay performance for samples prepared with unconjugated antibody and samples prepared with adc preparations with varying dar values may be conducted [46,62]. both types of qc samples should ideally produce back-calculated concentrations within the expected acceptable range of the assay (e.g., ±20 %). in addition, recoveries for qc samples generated using unconjugated antibody and adc preparations with varying dar values may be compared [62]. likewise, dere et al. demonstrated a comparison of serum t-dm1 concentrations obtained by using the total-trastuzumab and conjugated-trastuzumab elisa assays using the t-dm1 as the reference standard, which showed an excellent linear regression (r 2 =0.995 and a slope =1.04) [60]. often diverse bioanalytical methods, a combination of lba, high resolution lc-ms and high sensitive lcms/ms method, are applied to measure adc related analytes such as drugantibody conjugate, unconjugated antibody and free toxin as well as metabolite/catabolites. selecting the appropriate lcms/ms method of a large molecule largely depends on its molecular features and the required assay sensitivity or expected concentration range in the sample matrix to be analyzed. the affinity capture lc-ms and affinity capture hydrophobic interaction chromatography (hic) methods developed for adc characterization provide powerful tools for understanding the fate of adcs in vivo [85]. 3.4 matrix interference another bioanalytical challenge is matrix interference with reagents in biological samples in lba when evaluating pk or ada, e.g., circulating drug may interfere with the detection of ada and drug target, or ada may interfere with quantitation of drug levels in pk/tk bioanalysis. in addition to the most commonly used sample dilution method, the emergent technologies action program committee (etapc) working group has recently discussed several emerging technologies such as the singulex® erenna® platform, quanterix’s proprietary simoa™ technology (single molecule array), anp technologies’ npx4000 nanoparticles, sqi diagnostic’s ig plex™ and genalyte’s maverick™ technology, the maverick detection system, to deal with matrix effect on possible false positive or negative ada detection [86]. moreover, zoghbi et al. have developed a breakthrough novel method using precipitation and acid dissociation (panda) to overcome drug interference in the ada assay, showing significant improvement over the current approaches in detecting of ada for two igg1 and igg4 drugs as the examples [87]. it was thus claimed that the principle of this novel assay could be used not only for ada assays but also pk and biomarker (drug target) analysis in the presence of interference factors. 4. regulatory considerations regulatory guidance on metabolites and ddi as well as bioanalytical considerations for sms is well defined [21-23,39,59,88]. in the recent fda guidance [89], bioanalytical consideration for biopharmaceutics has also been included for industry bioanalytical method validation. general guidelines on preclinical safety evaluations including immunotoxicity of biotechnology-derived pharmaceuticals are available [34,55,90]. currently, it seems no requirement or guidance for metabolic investigation of mabs from regulatory agencies probably because it is generally accepted that mabs are unlikely to undergo significant biotransformation. the products of lysosomal degradation of mabs are assumed to be small peptides, amino acids and small carbohydrates that are readily eliminated by renal excretion or return to the nutrient admet & dmpk 4(1) (2016) 1-22 adme testing of large biologics doi: 10.5599/admet.4.1.276 13 pool without biological effects. however, this assumption may not be true for mab derivatives such as fusion proteins and antibody–drug conjugates (adcs) [2]. indeed, therapeutic protein-drug interactions and implications for drug development have been observed in clinic [25,91]. as a consequence, ddi studies on biologics have been outlined in the draft guidance [23], which becomes a major concern for safety evaluations of biologic for investigational new drug (ind) filing. unlike the mabs, an additional consideration for adcs is the adme and safety evaluations of novel cytotoxic compounds since the cytotoxic drug or its derivatives can be regarded as the api rather than intermediate during adc synthesis from regulatory compliance and cmc perspectives. for instance, the unreacted cytotoxic drug or linker might remain in the final adc drug product. if the concentrations of these impurities are minimal (e.g., <0.2 %), then independent testing of this component may not be necessary [92,93], and vice versa. according to recent suggestions by cder and aaps experts [94], free drug related impurities in clinical lot should be qualified relative to data from toxicology studies. comparable drug/antibody ratios should be maintained between the toxicology lot and the clinical lot. characterization of the impurity profile of drug/linker intermediates including structure determination of individual impurities (even likely intermediates) at levels >0.1 % is recommended prior to pivotal clinical trials. again, given the examples of two approved adc drugs, e.g. adcetris and kadcyla, adme testing and non-clinical safety assessments of the adc as well as the cytotoxic drug according to the cmc criteria are essential for ind filing. regardless of the regulatory pathway, characterization, comparability, release and stability assays need to be appropriate for the molecule to be analyzed. before more detailed guidelines on lm bioanalysis are available [89], an alternative lc-ms/ms technique is anticipated to serve as a complementary technique to quantify the antibody or adc in plasma for more applications of pk determinations as well as the quality control of lms. given the fact of increasingly growing interest in biotherapeutics, development and validation of cutting-edge bioanalytical methods are highly desirable for novel mabs and next generation of adcs. strong regulatory guidance and standard industry best practices for admet testing of lms and related bioanalytical considerations are expected to assess and manage the potential risk of biotherapeutics. 5. analysis and interpretation of preclinical animal pk of adcs and human pk prediction one ultimate purpose of admet testing is to understand the metabolism pathway of target molecule in various in vitro and in vivo species enable to better predict human pk. in general, the human clearance of sms is more predictable combining in vitro and in vivo animal data by means of well-stirred model and commonly used allometric scaling etc [95-99]. the examples applying allometric scaling have been reviewed for predictions of clearance and volume distribution of therapeutic proteins recently [37]. however, it was argued recently that the utility of allometric scaling and body surface area (bsa) to translate dosage from animal models to human clinical trials are inappropriate for human pk prediction due to interspecies difference in drug metabolism clearance and absorption [100]. for mabs exhibiting a linear pk without the tmdd, allometric scaling approach applying different scaling exponents is applicable for clearance and volume distribution predictions [101]. alternatively, pbpk modeling is considered as more accurate human pk predictions for large molecules with the linear pk [102], but it requires a number of input parameters to enable accurate human pk predictions, which limits its applicability in drug development. a recent survey shows that minimal pbpk (mpbpk) model offers a more mechanistic approach using the major structural features of full pbpk models for mabs in specifically analyzing mab pk than found in compartment models and provides an intermediary method if a full pbpk model is not available [103]. comparative evaluations of prediction approaches for projecting human pk of mabs in early hong wan admet & dmpk 4(1) (2016) 1-22 14 drug development prior to the first-in-human (fih) have been reviewed and decision tree in support of mab human pk projection has been proposed recently [104]. for lms with a nonlinear pk involved in the tmdd process, the predictability of plasma and tissue pk can be much more challenging particularly for adcs. despite this complexity, a mechanistic framework based on tmdd model was proposed to describe the pk of adcs and simulate the pk of t-dm1 as the example [105]. chen et al. [106] have utilized the pbpk modeling as a tool to predict mmae-based ddi potential drug interactions of adcs in good agreement with the observed data from clinical ddi data. in our preliminary analysis of two approved adcs, it seems no linear regression between log(cl) and log(bw) cross species when performing an allometric scaling for either sgn-35 or t-dm1 based on data in table 2 (further analysis is under way). this may highlights an example of an unpredictable human clearance (cl) from preclinical animal pk data by allometric scaling approach, or at least for these two marked adc drugs examined herein, due to in different clearance mechanisms of adcs and sms. furthermore, in view of preclinical animal pk data and human pk behaviors of adcetris (sgn-35) and kadcyla (t-d1) at two representative doses as summarized in table 2, it appears that the half-lives of both adcs in human are relatively shorter than in rodent species such as mouse and rats, but closer to monkey. also, it should be pointed out that some of reported t1/2 data are not consistent with the calculated t1/2 values resulting from a relationship of vd and cl, i.e., (t1/2=0.693×vd/cl), reflecting the different ways in calculating and reporting t1/2. bear in mind that the application of different compartmental models can result in various half-lives, e.g., distribution half-life, elimination or terminal half-life, which should be clarified in pk calculations. thus, it is important to ensure the reported t1/2 data in line with the decline of plasma concentration in order to correctly interpret dosing regimen as well as pk/pd relation as addressed above. moreover, as shown in table 2, an increased dose generally results in an increased t1/2 as a result of a decreased cl and slightly increased vd in both preclinical species and human. this may be explained by complex clearance mechanisms of biologics by their specific interactions with cellular receptors or described as “tmdd” [11], in addition to possible cyp metabolism resulting from cytotoxin. more specifically, once bound, drugs mainly undergo endocytosis and subsequent lysosomal degradation, dependent on the ratio of receptor/drug concentration, i.e., as the drug level increases or the number of receptors diminishes, the proportion of drug internalized and metabolized decreases, contributing to a lower clearance and consequently longer half-lives. furthermore, both auc and cmax’s increment of two adcs is greater than dose proportional when dose increases in both preclinical species and human. however, no obvious accumulations were observed in tk analysis, which is different from small molecule drugs wherein drug accumulation could often occurs when the cyp metabolism enzymes could be saturated at higher doses by non-linear exposure especially for low clearance compounds. on the other hand, from pk/pd perspective, it is generally accepted that small drug molecules are rapidly equilibrated at steady state with the same free concentration (unbound) between plasma and tissues, which means that pk/pd can be assessed based on free plasma concentrations without measuring tissue levels. in contrast, therapeutic biologics generally with limited and variable distribution in tissues, and serum levels are not necessarily predictive for tissue levels. added to that, biologics are administered mainly parenterally, so, adme testing generally focus on biodistribution or the relationship between tissue concentration and efficacy. as mentioned above, adcs may have a dose proportional pk of in rats probably due to no specific binding/no cross-reactivity, but a non-linear pk in cynomolgus monkeys due to crossreactivity/specific binding. this has raised a challenge in predicting human pk as well as human dose based on traditional approaches applied for small molecules owing to limited number of species and crossreactivity. in other words, how will preclinical data be translated to clinic in term of human pk and dose admet & dmpk 4(1) (2016) 1-22 adme testing of large biologics doi: 10.5599/admet.4.1.276 15 prediction? in order to better understand the pk/pd relationship and toxicity profile of adcs, total mab, adc or released payload might need to be measured in both plasma and tumor. in addition, conducting pd evaluations requires a clear understanding of the interaction of mab/adc with the target at early stage as possible, which may help pk/pd for human efficacy and dose prediction. pk/pd modeling and simulation (m&s) may provide an excellent tool to overcome these challenges, as it can simultaneously integrate the pk/pd of adcs and their components in a quantitative manner [111,112]. additionally, the computational pk/pd models can also serve as a cornerstone for the model-based drug development and preclinical-toclinical translation of adcs. for instance, shah et al. has developed a mechanistic model able to predict clinical response by integrating all preclinical biomeasures and pk/pd data of brentuximab-vedotin [113]. khot et al. highlighted the applications of m&s to gain distinct insights into adc development for pk/pd and toxicodynamic data by integrating a diverse array of in vitro, preclinical and clinical data generated at different stages of adc development [114]. table 2. preclinical and human pk data of adc drugs (adcetris and kadcyla)* drug mouse rat monkey human adcetris (brentuximab vedotin, sgn-35 ) dose (mg kg -1 ) 10 — 0.5 5.0 0.3 1.0 1.2 2.7 cmax (µg ml -1 ) — — 13 171 6.9 29.2 18.9 45.7 auc(µg ml -1 day -1 ) 2313 — 20 253 11 55.3 46.1 125.8 t1/2 (days) 16.9 — 14.6 8.5 1.8 2.7 3.8 6.0 t1/2 (calculated) 16.9 — 5.1 4.7 1.7 2.5 3.8 6.0 cl (ml day -1 kg -1 ) 4.4 — 25 20 27.7 18.5 26.0 21.5 vss (ml kg -1 ) 107 — 183 135 68 67.4 142 186 kadcyla (adotrastuzumab emtansine, tdm1) dose (mg kg -1 ) 0.3 15 0.3 20 0.3 3.0 0.3 4.8 cmax (µg ml -1 ) — — — — 7.4 74.2 9.63 130 auc(µg ml -1 day -1 ) — — — — 7.5 180 14.5 673 t1/2 (days) 4.2 13.1 4.9 5.4 0.9 2.9 1.3 4.1 t1/2 (calculated) 2.2 2.3 5.0 4.7 0.8 2.5 1.2 4.0 cl (ml day -1 kg -1 ) 13.0 19.2 10.1 22.1 40.4 16.5 21.1 7.1 vss (ml kg -1 ) 40.5 62.9 72.9 149 44.2 60.7 35.7 41.2 *adcetris/sgn-35 mouse pk data from ref. [47], rat and monkey pk data from ref. [107] and human pk data from ref. [108]. kadcyla/tdm1 mouse and rat pk data from ref. [109], monkey pk data from ref. [110] and human pk data from ref. [58]. t1/2 (calculated) data obtained by the relationship between vd and cl data (t1/2=0.693vd/cl). despite considerable differences between preclinical pk data and unpredictable human pk as well as complex pk/pd implication, saber and leighton recently reported an fda analysis of adcs on fih dose selection [115]. based on an fda oncology analysis of inds for adcs using preclinical data (plasma stability, toxicities in animals, and toxicology study designs), it was concluded that adcs can share the same small molecule drug concerning fih dose selection. more specifically, selecting a fih dose such as 1/6th the hong wan admet & dmpk 4(1) (2016) 1-22 16 hnstd in cynomolgus monkeys or 1/10th the std10 in rodents scaled according to bsa generally resulted in the acceptable balance of safety and efficient dose-escalation in fih trial. this certainly provides an encouraging guideline to support the safety evaluation of adcs and fih dose selection for clinical development by similar approach as small molecule drugs in this aspect. 6. concluding remarks biologic large molecules hold unique characteristics from small chemical entities. overall, the principles of adme testing for sms can be applied to testing biologics such as mabs and adcs. however, there are remarkable differences between sms and lms in adme assays associated with bioanalytical strategies. in comparison with adme testing of sms, there are limited in vitro models for adme testing of lms. also, fewer species with cross-reactivity can be utilized for in vivo pk studies as well as safety evaluations for lms. for initial adc evaluation, in vitro plasma stability assay combined with monitoring the free toxin released from adc in in vivo plasma can be utilized as a surrogate before conducting complicated and labintensive bioanalytical method development for pk quantification. from adme perspective, the design of novel linkers and the application of new conjugation technologies for uniform dar can not only further improve adc stability and consequent pk profiles as well as tolerability in systemic circulation of adcs, but also make bioanalysis of adc relatively easier. general adme/pk testing and nonclinical toxicity studies should be conducted for both adc and cytotoxin as appropriate, as exemplified in two approved adc drugs (adcetris and kadcyla). owing to fact that many elements such as antibody, free toxin, linker stability as well as product attributes can be adc toxicity determinants, adme testing and safety evaluations of adc are much more complex with significant development challenges, requiring unique and customized nonclinical adme approaches that differs from classical adme studies for sms. it appears that safety evaluation of key drug metabolite for adc drugs may be considered as a minor issue as long as the released cytotoxic concentration is far below its mtd concentration. however, the ddi studies on mabs or adcs (both adc and cytotoxic drug) are strongly recommended by fda guidance because of increased use of biotherapeutics and clear evidence of proteindrug interactions observed in clinic. relevant tools need to be developed for early ddi assessment of mabs or adcs with small molecules and other biologics for predicting clinical risk. additionally, it is anticipated that new methods are likely to be developed for a much better understanding of the immunotoxic potential from preclinical observations to clinic relevance. pk of lms is often pd-dependent and it might have a dose proportional pk owing to no specific binding/no cross-reactivity or otherwise non-linear pk due to specific binding/cross-reactivity. currently, the prediction of human pk and translation of preclinical pk/pd models to clinic are challenging especially for adcs due to limited species with cross-reactivity as well as complex tmdd and cyp enzymes based metabolism pathways. in addition, it requires simultaneous quantitative understanding about the pk/pd properties of three different molecular species, i.e., the monoclonal antibody, the drug, and the conjugate in both serum and tissues. further understanding pk prediction models of lms and refining pk/pd based on m&s are needed for human pk and efficacy predictions and dose optimization. currently, lba-based elisa is primarily preferred method for bioanalysis of mabs on available highthroughput formats. whilst lc-ms/ms can be utilized as an alternative technology of unique advantages for simulations quantification of co-administrated mabs or in the case of low recovery observed due to matrix interference with elisa method. although the routine usage of lc-ms is mainly hampered by the relatively time-consuming method development due to complex sample preparations, the availability of isotopically labeled proteins as iss, immunocapture and enzyme digestion of lms as well as limited sensitivity as admet & dmpk 4(1) (2016) 1-22 adme testing of large biologics doi: 10.5599/admet.4.1.276 17 compared to a typical bla, the highly selective lc-ms has been emerging as a viable alternative for qualitative and quantitative applications of lms, which can achieve comparable detection sensitivity with elisa for pk quantification of mabs as demonstrated in a number of recent applications. several multiple approaches have proposed to mitigate interference issues in lba, including the use of lc-ms for bioanalysis of mabs. however, it is desirable to deploy only one assay, rather than multiple assays, especially in later stages of clinical development. due to the heterogeneous nature of adcs and potential biotransformation, the hybrid technologies of best lba and lc-ms/ms are imperative for molecular characterization, pk/tk bioanalysis including total antibody, total adc, antibody-conjugated payload and unconjugated payload and immunogenicity evaluation as well as biotransformation and safety assessments of adcs. acknowledgements author would like to thank his colleagues xiaoyan zhu, zhendong xue, jun feng, kaijie he and yuchang mao for helping with some references, especially dr. weikang tao for his valuable comments and critical review of this manuscript, and dr. kin tam for final proof-reading of this manuscript. declaration of interest this article solely reflects author’s personal opinions. the author declares no competing conflicts of interest and receives no financial support from any affiliations in preparing this manuscript. references [1] a. beck, j.m. reichert. mabs 6 (2014) 15-17. [2] j.w. lee. bioanalysis 5 (2013) 2003-2014. [3] h.l. perez, p.m. cardarelli, s. deshpande, s. gangwar, g.m. schroeder, g.d. vite, r.m. borzilleri. drug discov. today 19 (2014) 869-881. [4] j.m. reichert. mabs 7 (2015) 1-8. [5] t. prueksaritanont, c. tang. aaps j 14 (2012) 410-419. [6] k. thudium, s. bilić, d. leipold, w. mallet, s. kaur, b. meibohm, h. erickson, j. tibbitts, h. zhao, m. gupta. mabs 5 (2013) 5-12. [7] a. mullard. nat. rev. drug discov. 12 (2013) 329-332. [8] j.a. cavagnaro. nat. rev. drug discov. 1 (2002) 469-475. [9] t.h. han, b. zhao. drug metab. dispos. 42 (2014) 1914-1920. [10] s. shi. curr. drug metab. 15 (2014) 271-290. [11] e. ezan, f. becher, f. fenaille. expert opin. drug metab. toxicol. 10 (2014) 1079-1091. [12] h. wan. admet & dmpk 1 (2013) 19-28. [13] t. tuntland, b. ethell, t. kosaka, f. blasco, r.x. zang, m. jain, t. gould, k. hoffmaster. front. pharmacol. 5 (2014) 174. [14] b. feng, m.v. varma, c. costales, h. zhang, l. tremaine. expert opin. drug discov. 9 (2014) 873-890. [15] v. ahuja, s. sharma. j. appl. toxicol. 34 (2014) 576-594. [16] m. meng, l. wang, t. voelker, s. reuschel, k.c. van horne, p. bennett. bioanalysis 5 (2013) 91-115. [17] l. li, d. tian, f. chen, j. yang, k. yu, y. sun. curr. drug metab. 13 (2012) 1206-1212. [18] h. wan, a.g. holmen. comb. chem. high throughput screen. 12 (2009) 315-329. [19] h. wan, m.j. rehngren. j. chromatogr. a 1102 (2006) 125-134. [20] h. wan, m. rehngren, f. giordanetto, f. bergström, a. tunek. j. med. chem. 50 (2007) 4606-4615. hong wan admet & dmpk 4(1) (2016) 1-22 18 [21] fda guidance for industry: safety testing of drug metabolites. (2008) available from: http:// www.fda.gov/downloads/drugs/guidancecomplianceregulatoryinformation/guidances/ucm079266.p df (accessed on march 23, 2016). [22] ich guidance on nonclinical safety studies for the conduct of human clinical trials and marketing authorization for pharmaceuticals. (2010) available from: http://www.fda.gov/downloads/drugs/guidancecomplianceregulatoryinformation/guidances/ucm07 3246.pdf (accessed on march 23, 2016). [23] fda guidance for industry: drug interaction studies – study design, data analysis, implications for dosing, and labeling recommendation. (2012) available from: http://www.fda.gov/downloads/drugs/guidancecomplianceregulatoryinformation/guidances/ucm29 2362.pdf (accessed on march 23, 2016). [24] k. lin, j. tibbitts. pharm. res. 29 (2012) 2354-2366. [25] s.m. huang, h. zhao, j.i. lee, k. reynolds, l. zhang, r. temple, l.j. lesko. clin. pharmacol. ther. 87 (2010) 497-503. [26] i.a. kaltashov, c.e. bobst, r.r. abzalimov, g. wang, b. baykal, s. wang. biotechnol. adv. 30 (2012) 210-222. [27] d. nebija, h. kopelent-frank, e. urban, c.r. noe, b. lachmann. j. pharm. biomed. anal. 56 (2011) 684-691. [28] j. visser, i. feuerstein, t. stangler, t. schmiederer, c. fritsch, m. schiestl. biodrugs 27 (2013) 495507. [29] m.m. sun, k.s. beam, c.g. cerveny, k.j. hamblett, r.s. blackmore, m.y. torgov, f.g.m. handley, n.c. ihle, p.d. senter, s.c. alley. bioconjug. chem. 16 (2005) 1282-1290. [30] r.j. harris, b. kabakoff, f.d. macchi, f.j. shen, m. kwong, j.d. andya, s.j. shire, n. bjork, k. totpal, a.b. chen. j. chromatogr. b biomed. sci. appl. 752 (2001) 233-245. [31] j. vlasak, r. ionescu. curr. pharm. biotechnol. 9 (2008) 468-481. [32] s.r. trevino, j.m. scholtz, c.n. pace. j. pharm. sci. 97 (2008) 4155-4166. [33] t.j. gibson, k. mccarty, i.j. mcfadyen, e. cash, p. dalmonte, k.d. hinds, a.a. dinerman, j.c. alvarez, d.b. volkin. j. pharm. sci. 100 (2011) 1009-1021. [34] fda. guidance for industry: assay development for immunogenicity testing of therapeutic proteins. (2009) available from: http://www.fda.gov/downloads/drugs/guidancecomplianceregulatoryinformation/guidances/ucm19 2750.pdf. (accessed on march 23, 2016). [35] fda. guidance for industry: immunotoxicology evaluation of investigation new drug. (2002) available from: http://www.fda.gov/downloads/drugs/guidancecomplianceregulatoryinformation/guidances/ucm07 9239.pdf. (accessed on march 23, 2016). [36] e.d. lobo, r.j. hansen, j.p. balthasar. j. pharm. sci. 93 (2004) 2645-2668. [37] j.h. lin. curr. drug metab. 10 (2009) 661-691. [38] y. vugmeyster, x. xu, f.p. theil, l.a. khawli, m.w. leach. world j. biol. chem. 3 (2012) 73-92. [39] r. deng, f. jin, s. prabhu, s. iyer. expert opin. drug metab. toxicol. 8 (2012) 141-160. [40] j.d. gomez-mantilla, i.f. troconiz, z. parra-guillen, m.j. garrido. j. pharmacokinet. pharmacodyn. 41 (2014) 523-536. [41] s.a. roberts, p.a. andrews, d. blanset d, k.m. flagella, b. gorovits, c.m. lynch, p.l. martin, k. kramer-stickland, s. thibault, g. warner. regul. toxicol. pharmacol. 67 (2013) 382-391. [42] p. chamberlain. bioanalysis 5 (2013) 561-574. [43] c.a. boswell, e.e. mundo, c. zhang, d. bumbaca, n.r. valle, k.r. kozak, a. fourie, j. chuh, n. koppada, o. saad, h. gill, b.q. shen, b. rubinfeld, j. tibbitts, s. kaur, f.p. theil, p.j. fielder, l.a. khawli, k. lin. bioconjug. chem. 22 (2011) 1994-2004. admet & dmpk 4(1) (2016) 1-22 adme testing of large biologics doi: 10.5599/admet.4.1.276 19 [44] e. chertova, c. bergamaschi, o. chertov, r. sowder, j. bear, j.d. roser, r.k. beach, j.d. lifson, b.k. felber, g.n. pavlakis. j. biol. chem. 288 (2013) 18093-18103. [45] j. gaudreault, d. fei, j. rusit, p. suboc, v. shiu. invest. ophthalmol. vis. sci. 46 (2005) 726-733. [46] s. kaur, k. xu, o.m. saad, r.c. dere, m. carrasco-triguero. bioanalysis 5 (2013) 201-226. [47] k.j. hamblett, p.d. senter, d.f. chace, m.m.c. sun, j. lenox, c.g. cerveny, k.m. kissler, s.x. bernhardt, a.k. kopcha, r.f. zabinski, d.l. meyer, j.a. francisco. clin. cancer res. 10 (2004) 70637070. [48] m. dorywalska, p. strop, j.a. melton-witt, a. hasa-moreno, s.e. farias, m.g. casas, k. delaria, v. lui, k. poulsen, c. loo, s. krimm, g. bolton, l. moine, r. dushin, t.t. tran, s.h. liu, m. rickert, d. foletti, d.l. shelton, j. pons, a. rajpal. bioconjug. chem. 26 (2015) 650-659. [49] a.g. polson, j. calemine-fenaux, p. chan, w. chang, e. christensen, s. clark, f.j. de sauvage, d. eaton, k. elkins, j.m. elliott, g. frantz, r.n. fuji, a. gray, k. harden, g.s. ingle, n.m. kljavin, h. koeppen, c. nelson, s. prabhu, h. raab, s. ross, j.p. stephan, s.j. scales, s.d. spencer, r. vandlen, b. wranik, s.f. yu, b. zheng, a. ebens. cancer res. 69 (2009) 2358-2364. [50] j.r. junutula, h. raab, s. clark, s. bhakta, d.d. leipold, s. weir, y. chen, m. simpson, s.p. tsai, m.s. dennis, y. lu, y.g. meng, c. ng, j. yang, c.c. lee, e. duenas, j. gorrell, v. katta, a. kim, k. mcdorman, k. flagella, r. venook, s. ross, s.d. spencer, w.l. wong, h.b. lowman, r. vandlen, m.x. sliwkowski, r.h. scheller, p. polakis, w. mallet. nat. biotechnol. 26 (2008) 925-932. [51] n.k. damle. nat. biotechnol. 26 (2008) 884-885. [52] p. dennler, a. chiotellis, e. fischer, d. brégeon, c. belmant, l. gauthier, f. lhospice, f. romagne, r. schibli. bioconjug. chem. 25 (2014) 569-578. [53] p.f. peddi, s.a. hurvitz. future oncol. 9 (2013) 319-326. [54] h.k. erickson, j.m. lambert. aaps j. 14 (2012) 799-805. [55] ich. guidance for industry: s6(r1): preclinical safety evaluation of biotechnology-derived pharmaceuticals. (2012) available from: http://www.fda.gov/downloads/drugs/guidancecomplianceregulatoryinformation/guidances/ucm 194490.pdf. (accessed on march 23, 2016). [56] m.b. hock, k.e. thudium, m. carrasco-triguero, n.f. schwabe. aaps j. 17 (2015) 35-43. [57] fda. guidance for industry: s9 nonclinical evaluation for anticancer pharmaceuticals. (2010) available from: http://www.fda.gov/downloads/drugs/guidancecomplianceregulatoryinformation/guidances/ucm08 5389.pdf. (accessed on march 23, 2016). [58] s. girish, m. gupta, b. wang, d. lu, i.e. krop, c.l. vogel, h.a. burris iii, p.m. lorusso, j.h. yi, o. saad, b. tong, y.w. chu, s. holden, a. joshi. cancer chemother. pharmacol. 69 (2012) 1229-1240. [59] fda. guidance for industry: bioanalytical method validation. (2001) available from: http://www.fda.gov/downloads/drugs/guidancecomplianceregulatoryinformation/guidances/ucm0701 07.pdf. (accessed on march 23, 2016). [60] r. dere, j.h. yi, c. lei, o.m. saad, c. huang, y. li, j. baudys, s. kaur. bioanalysis 5 (2013) 1025-1040. [61] e. ezan, m. dubois, f. becher. analyst 134 (2009) 825-834. [62] b. gorovits, s.c. alley, s. bilić, b. booth, s. kaur, p. oldfield, s. purushothama, c. rao, s. shord, p. siguenza. bioanalysis 5 (2013) 997-1006. [63] y.m. wang, v. jawa, m. ma. bioanalysis 6 (2014) 79-87. [64] l. stevenson, l. amaravadi, h. myler, l. salazar-fontana, b. gorovits, s. kirshner, l. xue, f. garofolo, s.c. alley, t. thway, a. joyce, s. bansal, c. beaver, a. bergeron, x.y. cai, l. cojocaru, b. desilva, i.dumont, e. fluhler, s. fraser, d. gouty, s. gupta, s. haidar, r. hayes, b. ingelse, a. ishii-watabe, s. kaur, l. king, o. laterza, s. leung, a. lévesque, m. ma, c. petit-frere, r. pillutla, m. rose, g. schultz, j. smeraglia, s. swanson, a. torri, f. vazvaei, j. wakelin-smith, a. wilson, e. woolf, t.y. yang. bioanalysis 6 (2014) 3355-3368. hong wan admet & dmpk 4(1) (2016) 1-22 20 [65] l. stevenson, f. garofolo, b. desilva, i. dumont, s. martinez, m. rocci, l. amaravadi, m. brudnykloeppel, a. musuku, b. booth, c. dicaire, l. wright, l. mayrand-provencher, m. losauro, d. gouty, m. arnold, s. bansal, s. dudal, d. dufield, j. duggan, c. evans, e. fluhler, s. fraser, b. gorovits, s. haidar, r. hayes, s. ho, r. houghton, r. islam, r. jenkins, n. katori, s. kaur, m. kelley, m. knutsson, j. lee, h. liu, s. lowes, m. ma, a. mikulskis, h. myler, b. nicholson, t. olah, e. ormsby, s. patel, v. pucci, c. ray, g. schultz, j. shih, r. shoup, c. simon, a. song, j.t. neto, v. theobald, t. thway, j. wakelin-smith, j. wang, l. wang, j. welink, e. whale, e. woolf, r. xu. bioanalysis 5 (2013) 29032918. [66] d. l. chappell, m.e. lassman, t. mcavoy, m. lin, d.s. spellman, o.f. laterza. bioanalysis 6 (2014) 1843-1857. [67] y.q. xiao. moj immunol. 1 (2014) 4. [68] quantitation of antibody-drug conjugates, cyros application note, d0020478. [69] a.p. joyce, m. wang, r. lawrence-henderson r, c. filliettaz, s.s. leung, x. xu, d.m. o’hara. pharm. res. 31 (2014) 1823-1833. [70] j. h. heo, x. mou, f. wang, j.m. troisi, c.w. sandifer, s. kirby, d. driscoll, s. mercorelli, d.j. pollard. pharm. bioprocess. 2 (2014) 129-139. [71] c.m. hall, j.t. pearson, v. patel, l.c. wienkers, r.j. greene. j. immunol. methods 393 (2013) 70-73. [72] r. piparia, d. ouellette, w.b. stine, c. grinnell, e. tarcsa, c. radziejewski, i. correia. mabs 4 (2012) 521-531. [73] x. peng, b. liu, y. li, h. wang, x. chen, h. guo, q. guo, j. xu, h. wang, d. zhang, j. dai, s. hou, y. guo. chromatographia 78 (2015) 521-531. [74] m.h. park, m.w. lee, y.g. shin. biomed. chromatogr. 2015. [75] q. zhang, d.s. spellman, y. song, b. choi, n.g. hatcher, d. tomazela, m. beaumont, m. tabrizifard, d. prabhavalkar, w. seghezzi, j. harrelson, k.p. bateman. anal. chem. 86 (2014) 8776-8784. [76] h. li, r. ortiz, l. tran, m. hall, c. spahr, k. walker, j. laudemann, s. miller, h. salimi-moosavi, j.w. lee. anal. chem. 84 (2012) 1267-1273. [77] a. beck, s. sanglier-cianferani, a. van dorsselaer. anal. chem. 84 (2012) 4637-4646. [78] h. jiang, j. zeng, c. titsch, k. voronin, b. akinsanya, l. luo, h. shen, d.d. desai, a. allentoff, a.f. aubry, b.s. desilva, m.e. arnold. anal. chem. 85 (2013) 9859-9867. [79] k. xu, l. liu, m. maia, j. li, j. lowe, a. song, s. kaur. bioanalysis 6 (2014) 1781-1794. [80] d. lebert, g. picard, c, beau-larvor, l. troncy, c. lacheny, b. maynadier, w. low, n. mouz, v. brun, c. klinguer-hamour, m. jaquinod, a. beck. bioanalysis 7 (2015) 1237-1251. [81] i. onami, m. ayabe, n. murao, m. ishigai. j. chromatogr. a 1334 (2014) 64-71. [82] k. sandra, p. sandra. bioanalysis 7 (2015) 2843-2847. [83] s. rosati, y. yang, a. barendregt, a.j.r. heck. nat. protoc. 9 (2014) 967-976. [84] m.t. furlong, s. zhao, w. mylott, r. jenkins, m. gao, v. hegde, j. tamura, a. tymiak, m. jemal. bioanalysis 5 (2013) 1363-1376. [85] a. liu, a. kozhich, d. passmore, h. gu, r. wong, f. zambito, v.s. rangan, h. myler, a.f. aubry, m.e. arnold, j. wang. j. chromatogr. b analyt. technol. biomed. life sci. 1002 (2015) 54-62. [86] r.m. crisino, l. luo, b. geist, j. zoghbi, f. spriggs. bioanalysis 6 (2014) 1033-1036. [87] j. zoghbi, y. xu, r. grabert, v. theobald, s. richards. j. immunol. methods 426 (2015) 62-69. [88] fda. guidance for industry: nonclinical studies for the safety evaluation of pharmaceutical excipients. (2005) available from: http://www.fda.gov/downloads/drugs/guidancecomplianceregulatoryinformation/guidances/ucm07 9250.pdf. (accessed on march 23, 2016). [89] fda. guidance for industry: bioanalytical method validation, biopharmaceutics. revision 1. (2013) available from: http://www.fda.gov/downloads/drugs/guidancecomplianceregulatoryinformation/guidances/ucm 368107.pdf. (accessed on march 23, 2016). admet & dmpk 4(1) (2016) 1-22 adme testing of large biologics doi: 10.5599/admet.4.1.276 21 [90] k.l. hastings, int. immunopharmacol. 2 (2002) 1613-1618. [91] k. seitz, h. zhou. j. clin. pharmacol. 47 (2007) 1104-1118. [92] ich. guidance for industry: q3b(r2) impurities in new drug products.2006. available from: http://www.fda.gov/downloads/drugs/guidancecomplianceregulatoryinformation/guidances/ucm07 3389.pdf. (accessed on march 23, 2016). [93] ich. guidance for industry: q3a impurities in new drug substances.2008. available from: http://www.fda.gov/downloads/drugs/guidancecomplianceregulatoryinformation/guidances/ucm07 3385.pdf. (accessed on march 23, 2016). [94] regulatory considerations for antibody-drug conjugates , ops, cder, aaps (october 18, 2012). available from: http://www.fda.gov/downloads/aboutfda/centersoffices/officeofmedicalproductsandtobacco/cder/ ucm341177.pdf. (accessed on march 23, 2016). [95] i. mahmood. clin. pharmacokinet. 51 (2012) 527-542. [96] h. wan, p. bold, l.o. larsson, j. ulander, s. peters, b. lofberg, a.l. ungell, m. nagard, a. llinas. curr. drug metab. 11 (2010) 583-594. [97] v.k. sinha, s.s. de buck, l.a. fenu, j.w. smit, m. nijsen, r.a.h.j. gilissen, a. van peer, k. lavrijsen, c.e. mackie. clin. pharmacokinet. 47 (2008) 35-45. [98] q. huang, j.e. riviere. expert opin. drug metab. toxicol. 10 (2014) 1241-1253. [99] l. liu, k.s. pang. eur. j. pharm. sci. 29 (2006) 215-230. [100] o.l. blanchard, j.m. smoliga. faseb j. 29 (2015) 1629-1634. [101] l. diao, b. meibohm. expert opin. drug metab. toxicol. 11 (2015) 1115-1125. [102] h.m. jones, y. chen, c. gibson, t. heimbach, n. parrott, s.a. peters, j. snoeys, v.v. upreti, m. zheng s.d. hall. clin. pharmacol. ther. 97 (2015) 247-262. [103] y. cao, w.j. jusko. j. pharmacokinet. pharmacodyn. 41 (2014) 571-580. [104] j. wang, s. iyer, p.j. fielder, j.d. davis, r. deng. biopharm. drug dispos. 2015. [105] l. gibiansky, e. gibiansky. j. pharmacokinet. pharmacodyn. 41 (2014) 35-47. [106] y. chen, d. samineni, s. mukadam, h. wong, b.q. shen, d. lu, s. girish, c. hop, j.y. jin, c. li. clin. pharmacokinet. 54 (2015) 81-93. [107] adcetris, fda cder (center for drug evaluation and research), pharmacology review, application number 125399orig1s000, 2011. available from: http://www.accessdata.fda.gov/drugsatfda_docs/label/2011/125388s000,125399s000lbl.pdf. [108] adcetris, fda cder (center for drug evaluation and research), clinical pharmacology and biopharmaceutics review, application number 125399orig1s000, 2011. available from:http://www.accessdata.fda.gov/drugsatfda_docs/label/2011/125388s000,125399s000lbl.pdf. (accessed on march 23, 2016). [109] kacyla ema-assessment report on kadcyla-ema749228-2013. available from: http://www.ema.europa.eu/docs/en_gb/document_library/epar__public_assessment_report/human/002389/wc500158595.pdf. (accessed on march 23, 2016). [110] kadcyla, fda cder (center for drug evaluation and research), pharmacology review, application number 125427orig1s000, 2013. available from: http://www.accessdata.fda.gov/drugsatfda_docs/label/2013/125427lbl.pdf. (accessed on march 23, 2016). [111] a.p. singh, y.g. shin, d.k. shah. pharm. res. 32 (2015) 3508-3525 . [112] p. sapra, a. betts, j. boni. expert rev. clin. pharmacol. 6 (2013) 541-555. [113] d.k. shah, n. haddish-berhane, a. betts. j. pharmacokinet. pharmacodyn. 39 (2012) 643-659. [114] a. khot, s. sharma, d.k. shah. bioanalysis 7 (2015) 1633-1648. [115] h. saber, j.k. leighton. regul. toxicol. pharmacol. 71 (2015) 444-452. hong wan admet & dmpk 4(1) (2016) 1-22 22 abbreviations: aaps american association of pharmaceutical scientists ada anti-drug antibody adc antibody-drug conjugates adme absorption, distribution, metabolism and elimination api active pharmaceutical ingredient ata anti-therapeutic antibodies auc area under the curve bsa body surface area cder center for drug evaluation and research cyp cytochrome p450 dar drug-antibody ratio, payload ddi drug-drug interaction etapc the emergent technologies action program committee elisa enzyme-linked immunesorbent assay fih first-in-human herg human ether-a-go-go related gene hic hydrophobic interaction chromatography hnstd highest non-severely toxic dose lba ligand binding assay igg immunoglobulin g im intramuscular ind investigational new drug is internal standard iv intravenous lc-ms liquid chromatography-mass spectrometry lc-ms/ms liquid chromatography with tandem mass spectrometry lms large molecules mab monoclonal antibody mmae monomethyl auristatin e m&s modeling and simulation mtd maximum tolerated dose mw molecular weight pbpk physiologically-based pharmacokinetics pcc preclinical candidate pd pharmacodynamics pk pharmacokinetics pk/pd pharmacokinetics-pharmacodynamics pk/tk pharmacokinetics-toxicokinetics ptm post-translational modifications sc subcutaneous sms small molecules tdc thiomab-drug conjugates t-dm1 trastuzumab-emtansine tmb 3, 3', 5, 5' tetramethylbenzidine tmdd target-mediated drug disposition vd volume of distribution ©2016 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ manuscript doi: 10.5599/admet.4.1.275 23 admet & dmpk 4(1) (2016) 23-34; doi: 10.5599/admet.4.1.275 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper transport of six tyrosine kinase inhibitors: active or passive? richard j. honeywell, sarina hitzerd, ietje kathmann, godefridus j. peters* department of medical oncology, vu university medical center, amsterdam, the netherlands *corresponding author: e-mail: gj.peters@vumc.nl; tel.: +31 (20) 444 26 33 received: february 22, 2016; published: march 31, 2016 abstract transport of erlotinib, gefitinib, sorafenib, sunitinib, dasatinib and crizotinib can be active or passive, which was studied by measuring uptake at low (4 °c; passive) and normal temperature (37 °c; active and passive) and by the use of specific organic cation transporter (oct) inhibitors. intracellular accumulation was determined using caco-2 as monolayers, while for gut permeation we used differentiated caco-2 as model for intestinal epithelium in the transwell system. sorafenib and crizotinib uptake are likely to be dependent on passive transport. gefitinib, dasatinib and sunitinib uptake seem to be active. erlotinib’s transport also seems to be active. this study suggests that hocts might be involved in the apical to basolateral transport of gefitinib and crizotinib. overall it can be concluded that the accumulation and transport of these six tkis are very different, despite the fact that they are all tyrosine kinase inhibitors. keywords tyrosine kinase; drug transport; erlorinib; gefitinib; sorafenib; sunitinib; dasatinib; crizotinib introduction protein kinases are key regulator enzymes of cell function via the transfer of a phosphate group from atp to a protein in a cell [1]. one of the subgroups is the tyrosine kinases (tk) family. these kinases attach phosphate groups to the amino acid tyrosine of a protein, controlling a wide range of properties in proteins such as, but not limited to, differentiation, apoptosis, enzyme activity, subcellular localization and interaction between molecules [2]. subsequently, tks function in many signaling transduction cascades, controlling fundamental cellular processes including cell cycle, migration, survival, proliferation and metabolism [3]. receptor tks consist of an extracellular ligand binding domain which is connected to the cytoplasmic domain by a single trans-membrane helix. the binding of a ligand to the receptor induces dimerization resulting in internal activation of the rtk [4,5]. mutations in the atp-binding site can cause some tks to become constitutively active which results in a non-stop functional state that may contribute to initiation or progression of cancer [6]. small molecule tyrosine kinase inhibitors (tkis) were developed to inhibit tk function. in this study, six different tkis were investigated. erlotinib and gefitinib are reversible inhibitors of the epidermal growth factor receptor tyrosine kinase (egfr) [7] and are approved for treatment of non-small-cell lung cancer (nsclc) [8-10]. sorafenib is a multiple kinase inhibitor which inhibits multiple intracellular (craf, (mutant) braf) and receptor cell surface kinases (kit, flt-3, ret, vegfr-1/2/3, and pdgfr-β) [11]. sorafenib is approved for liver (hepatocellular carcinoma (hcc)) and kidney (renal cell carcinoma (rcc)) cancer [12,13]. sunitinib is also a multi-targeted receptor tki which is http://www.pub.iapchem.org/ojs/index.php/admet/index honeywell et al. admet & dmpk 4(1) (2016) 23-34 24 known to inhibit vegfr, pdgfr, kit, ret, csf-1 and flt3 [14]. sunitinib is fda approved for multiple cancer types, namely gastrointestinal stromal tumor (gist), kidney cancer (rcc) and pancreatic neuroendocrine tumors (pnet) [15-17]. dasatinib is a bcr-abl inhibitor, but can also inhibit pdgfr, c-kit, src and the ephrin a receptor [18]. dasatinib was fda approved for the treatment of chronic myeloid leukemia (cml) [19]. crizotinib is a met/alk multi-targeted (alk, ros1, cmet) receptor tki approved for nsclc [20,21]. generally, tkis are poorly soluble in aqueous solutions, leading to poor bioavailability [22,23]. this may also affect uptake and efflux of these compounds. efflux is mediated by atp-dependent efflux pumps of the atp-binding cassette (abc) family, such as bcrp (abcg2), p-gp (abcb1) and multidrug resistance protein 1 (mrp1) [24-26]. influx pumps include the organic cation transporters (octs). since several tkis have a cationic charge and high lipophilicity, they are substrates for octs [22,27,28]. tkis are oral drugs that have to pass the intestine barrier to enter the systemic circulation. therefore, transport proteins on the intestinal epithelia are very important in the drug bioavailability and resistance mechanisms [22,29]. the human colon carcinoma cell line caco-2 mimics the epithelial monolayer, which covers the inner intestinal wall and has been used as a model system for drug investigation [30] (figure 1). the hoct1 (basolateral) and hoct3 (apical) are known to be functional present at the intestinal level (caco2 cells) [31,32]. we recently validated the caco-2 permeable cell culture system for permeation of several tkis [33] and applied it to determine the role of various drug transporters in permeability and absorption of erlotinib, gefitinib, sorafenib, sunitinib, dasatinib and crizotinib, by using several specific inhibitors of these pumps. materials and methods materials erlotinib, gefitinib, sorafenib, sunitinib, dasatinib, crizotinib were from lc laboratories (woburn, usa). amantadine, β-estradiol, cimetidine and verapamil were purchased from sigma (st. louis, mo, usa). desipramine was from icn biochemicals (aurora, oh, usa). all above drugs were ultimately diluted in dmso. dulbecco’s modified eagle medium (dmem), trypsin-edta, penicillin/streptomycin (p/s; 10000 u/ml) and 1 m hepes buffer (in 0.85 % nacl) were purchased from lonza benelux bv (breda, the netherlands). phosphate buffered saline (pbs) was purchased from b. braun medical bv (oss, the netherlands). fetal bovine serum (fbs) was purchased from paa laboratories gmbh (pasching, austria). bovine serum albumin fraktion v (bsa) was purchased from roche diagnostics gmbh (mannheim, germany). analytical grade solvents like acetonitrile, methanol and isopropanol were purchased from biosolve bv (valkenswaard, the netherlands). bio-rad protein assay was purchased from bio-rad laboratories gmbh (münchen, germany). milliq water was supplied via a milliq water purification system (millipore, the netherlands). the biocoat® hts caco-2 assay system was purchased from becton dickinson bv (breda, the netherlands). breathe-easy microplate sealing film was purchased from diversified biotech bv (ulvenhout, the netherlands). the trans epithelial electrical resistance (teer) meter (millicell® – ers) was provided by millipore (amsterdam, the netherlands). lc-ms/ms analyses were performed using a dionex ultimate 3000 system coupled with an api 3000 mass spectrometer (applied biosciences sciex). for this system the following software was used: analyst version 1.5.2 from applied biosciences in combination with dionex, chromeleon lc modules version 6.8, controlled by dionex mass link (dms) version 2.10. admet & dmpk 4(1) (2016) 23-34 transport mechanism of tk inhibitors doi: 10.5599/admet.4.1.275 25 cell culture the caco-2 cell line, originating from a colorectal adenocarcinoma, was cultured in dmem supplemented with 10 % fbs and 20 mm hepes at 37 °c, 5 % co2 and 100 % humidity [31,33]. under normal culture conditions caco-2 cells behave like a colon cancer cell in which no polarization at apical or basolateral side is found. using a special coating in a transwell, the cell line differentiates and when confluent, forms a cellular layer with specific transporter expression at each side (figure 1). characterization of passive and active transport active and passive transport were characterized by culturing the cells at a low temperature (4 °c; passive transport) or a physiological temperature (37 °c; active and passive transport). caco-2 cells were seeded in 2 ml dmem culture medium with 1 % penicillin/streptomycin (p/s) in 6-well plates at a density of 0.2 x 10 5 cells/cm 2 . cells were allowed to attach for 36 hours at 37 °c, 5 % co2 and 100 % humidity. subsequently, cells were incubated at 4 °c or at 37 °c for two hours prior to the experiment. attached cells were than exposed to 10 µm erlotinib, gefitinib, sorafenib, sunitinib, crizotinib and 1 µm dasatinib for 2, 4 and 6 hours at both 4 °c and 37 °c. after drug exposure medium was aspirated and cells were washed three times with 5 ml ice cold pbs. trypsin/edta (200 µl) was added to each well and incubated for about 5-10 minutes at 37 °c, 5 % co2 and 100 % humidity. when cells were detached, the cells were re-suspended with 1 ml ice cold pbs and transferred into eppendorf tubes. next, cell suspension was spun down for 5 minutes at 425 g at 4 °c. at last the supernatant was aspirated, the pellet was snap frozen in liquid nitrogen and stored at -80 °c, until analysis by the liquid chromatography coupled to mass spectrometry (lc-ms-ms) [34]. the protein content was determined using the bio-rad protein assay. figure 1. caco-2 cell culture system with an apical (intestine) and basolateral (blood) side[35]. effect of specific oct inhibitors transporter inhibition studies were performed to determine whether the human organic cation transport (hoct) family is involved in the transport and accumulation of erlotinib, gefitinib, sorafenib, sunitinib, dasatinib or crizotinib. to achieve this, caco-2 cells were cultured as described above at 37 °c. after 36 hours, to allow proper attachment, cells were exposed to either 100 µm amantadine (hoct1&2), 100 µm cimetidine (hoct1&2&3) [37], 100 µm desipramine (hoct1&2&3), 10 µm β-estradiol (hoct1&3) or 10 µm verapamil (hoct1&p-gp) [28,35,36] for 15 min. then attached cells were exposed to 10 µm erlotinib, gefitinib, sorafenib, sunitinib, crizotinib or 1 µm dasatinib for 2 hours at 37 °c. after drug exposure cells were harvested and processed as described above. honeywell et al. admet & dmpk 4(1) (2016) 23-34 26 differentiated caco-2 cells as a model for gut epithelial transport in transwell transwell studies were performed to investigate the transport of tkis from apical (a) to basolateral (b) side through a confluent caco-2 monolayer. the optimization and validation of this model was recently described [33]. shortly, the becton dickinson bioscience biocoat ® hts caco-2 assay system was used [37,38]. caco-2 cells were plated at a density of 200,000 cells per insert (6.6 x 10 5 cells/cm 2 ) in 500 l basal seeding medium supplemented with 0.1 % lyophilized powder mito + serum extender and 1 % p/s (bsm supplemented). the transwell inserts were placed into a multiwell tm feeder tray, covered with breatheasier cell culture foil, and incubated for 22-24 h at 37 °c, 5 % co2 and 100 % humidity; the medium was aspirated from both apical and basolateral side and replaced with entero-stim tm supplemented with 0.1 % lyophilized powder mito + serum extender and 1 % p/s (entero-stim tm supplemented) for 44-48 hours enabling cells to differentiate. for the permeation the multiwell tm feeder tray was replaced for a falcon ® 24-well plate at 37 °c, 5 % co2 and 100 % humidity. the medium at the apical site was replaced with 300 l hank’s buffered salt solution (hbss), 100 m cimetidine, desipramine or 10 m verapamil (apical side) and 1 ml hbss (basolateral side). trans epithelial electral resistance (teer) was measured using a millicell-ersvoltmeter to verify the quality of the caco-2 cell monolayers. monolayers with teer values determined below 165 ω cm 2 were discarded [35]. next, the hbss and hoct inhibitor solution were replaced for 300 l of 20 m erlotinib, gefitinib, sorafenib, sunitinib, dasatinib or crizotinib alone or in combination with 100 m cimetidine, desipramine or 10 m verapamil (apical side) and 1 ml hbss + 5 % bsa (basolateral side). immediately after the addition of the solutions samples of 20 l were taken from the apical side (volume was not replaced). the plate was incubated under gently shaking (100 rpm) at 37 °c, 5 % co2 and 100 % humidity. samples of 50 l were taken after 15, 30, 60, 90, 120 and 180 min from the basolateral side (volume was replaced by hbss + 5% bsa). finally, samples of 20 l were taken from the apical side at 180 minutes and the cells were harvested from the filter inserts, by washing twice with pbs (room temperature) on both apical (400 l) and basolateral side (1 ml). pbs was replaced by 50 l of trypsin/edta to allow cells to detach; then cells were re-suspended with 150 l pbs, transferred into eppendorf tubes and spun down for 5 minutes at 10621 g at 4 °c. the supernatant was aspirated and the pellet was snap frozen in liquid nitrogen and stored at -80 °c until analysis by lc-ms/ms. the protein content was determined using the bio-rad protein assay. data analysis by lc-ms/ms all samples of the accumulation and transwell studies were measured using liquid chromatography coupled to mass spectrometry to determine the amount of drug (tki) present [34]. sample pellets were resuspended in 200 l of ice cold water and homogenized by aspiration via a pipette tip. cell pellets of the caco-2 monolayer were also transferred into qiagen shredder tubes (containing qia shredder mini spin columns) and spun down at 15800 g for 2 min to prepare a homogeneous sample. subsequently, 20 l of each sample was transferred to 96-well plates and 80 l of ice cold acetonitrile added. each plate was covered with a plastic foil, sonicated for 30 seconds and centrifuged at 1500 g at 4 °c for 10 min. next, 50 l of each sample was transferred to clean 96-well plates and covered with an nunc well plate seal. lastly, 1 l of each sample was injected into the lc-ms/ms. compounds were separated on a phenomenex prodigy ods3, 100x2 mm, 3 m column using a mobile phase containing 0.578 g ammonium acetate, 375 ml milliq, 125 ml methanol, 0.1 % isopropanol and 1000 ml acetonitrile with a ph of 7.8 at 500 l/min. final results of the lc-ms/ms (in ng/ml tki) were expressed in pmol (transwell) or combined with the results of the protein content determination (mg/ml protein), obtaining data expressed in ng drug per mg protein. admet & dmpk 4(1) (2016) 23-34 transport mechanism of tk inhibitors doi: 10.5599/admet.4.1.275 27 calculations apparent permeability coefficients were estimated by the following equation: app q v p t c a0    where papp is the apparent permeability coefficient (cm/sec), q/t is the rate of accumulative mass transport in the receiver chamber over time (m/sec), v is the volume in the receiver chamber (ml), c0 is the concentration of tki in the donor chamber (m) and a is the diffusion area (cm 2 ). results temperature dependent accumulation in caco-2 cells caco-2 cells were exposed to the six different tkis under different temperature conditions to determine whether the main transport was driven by passive diffusion (4 °c) or by active carrier transport (37 °c), (figure 2). the 2-hr time point is considered to be relevant for initial uptake; at later time-points other processes (efflux, metabolism, accumulation in cellular organelles) seem more important. after 2-hr erlotinib, gefitinib and dasatinib showed a much higher accumulation at 37 °c than at 4 °c, (figures 2a,b and 2f, respectively); for sunitinib, crizotinib and sorafenib this difference is much less and not significantly different (figures 2c, 2d and 2e respectively). at later time-points the pattern changes, erlotinib accumulation at 4 °c continued to increase, but at 37 °c the accumulation decreased (figure 2a), indicating an active efflux process. sunitinib accumulation at 37 °c continued to increase, possibly due to a lysosomal accumulation (figure 2c) [22,39]. for crizotinib, the accumulation continued to increase under both conditions (figure 2e), but for sorafenib no change was observed (figure 2d). transporter dependent hoct dependent accumulation in caco-2 cells caco-2 cells were exposed to several inhibitors of the hoct family in combination with the six tkis to determine whether hocts are involved in the active uptake of the six tkis (figure 3). the hoct family can be inhibited by amantadine (hoct1&2), cimetidine (hoct1&2&3) [40], desipramine (hoct1&2&3), βestradiol (hoct1&3) and verapamil (hoct1&p-gp) [27,35,36]. all five hoct inhibitors decreased gefitinib accumulation, demonstrating that hoct is involved in the uptake of gefitinib. desipramine, cimetidine and β-estradiol also decreased gefitinib accumulation, indicating that hoct3 is involved in the uptake of gefitinib. erlotinib shows a completely different effect on the incubation with hoct inhibitors. verapamil, desipramine and cimetidine induced a high increase of erlotinib accumulation, in contrast to gefitinib. amantadine and β-estradiol only induced a slight decrease of erlotinib accumulation. therefore hoct is not involved in erlotinib uptake. the effect of hoct inhibitors on sunitinib and crizotinib was the same, except for verapamil. verapamil caused a slight increase in sunitinib accumulation, but a slight decrease in crizotinib accumulation. desipramine decreased sunitinib and crizotinib accumulation, which indicates that hoct1, 2 and 3 are involved in sunitinib and crizotinib uptake. dasatinib accumulation was inhibited equally by all of the hoct inhibitors, except for verapamil. besides hoct1, verapamil also inhibits p-gp, an efflux transporter, resulting in a 3-fold increase of dasatinib accumulation. these results indicate that hocts are involved in dasatinib influx and p-gp is involved in dasatinib efflux. the influence of the hoct inhibitors on sorafenib accumulation is completely different, compared to dasatinib accumulation. amantadine and β-estradiol did not show any effect and verapamil only induced a slight increase, whereas desipramine and cimetidine decreased sorafenib accumulation. honeywell et al. admet & dmpk 4(1) (2016) 23-34 28 2 4 6 0 20 40 60 80 4°c 37°c time (hours) a c c u m u la ti o n ( n g / m g p ro te in ) 2 4 6 0 500 1000 1500 2000 4°c 37°c time (hours) 2 4 6 0 5000 10000 15000 4°c 37°c time (hours) a c c u m u la ti o n ( n g / m g p ro te in ) 2 4 6 0 2000 4000 6000 4°c 37°c time (hours) 2 4 6 0 2000 4000 6000 8000 10000 4°c 37°c time (hours) a c c u m u la ti o n ( n g / m g p ro te in ) 2 4 6 0 50 100 150 200 4°c 37°c time (hours) figure 2. influence of temperature (4 °c and 37 °c) on the cellular accumulation of 10 µm erlotinib (a); gefitinib (b); sunitinib (c); sorafenib (d); crizotinib (e) and 1 µm dasatinib (f) in caco-2 cells. all values are reported as the mean ± sem of n=2 and expressed as the ng accumulated per µg of sample protein. effect of hoct inhibition basolateral transport within an intestinal model to further investigate the influence of hoct on the uptake of the six tkis, accumulation and transport studies were performed with the becton dickinson bioscience biocoat® hts caco-2 assay system. verapamil, desipramine and cimetidine demonstrated the most pronounced effect in the accumulation study (figure 3), these hoct inhibiters were used in the transwell system. none of the three hoct inhibitors had any effect on the apical to basolateral transport of erlotinib (figure 4a), which is in line with the minor effect on the accumulation in caco-2 cells (figure 2a). however, cellular erlotinib accumulation was decreased at the end of the incubation (figure 5). in contrast to erlotinib, gefitinib permeation was affected by oct inhibitors. desipramine increased apical to basolateral transport of gefitinib (figure 4b), yet decreased cellular gefitinib accumulation (figure 5). therefore, there may be increased mass transport from apical to basolateral side. cimetidine did not affect permeation and cellular accumulation of gefitinib, while verapamil increases cellular accumulation and permeation. a f d b c e admet & dmpk 4(1) (2016) 23-34 transport mechanism of tk inhibitors doi: 10.5599/admet.4.1.275 29 figure 3. relative accumulation after the exposure of 10 μm erlotinib, gefitinib, sorafenib, sunitinib, crizotinib and 1 μm dasatinib with and without hoct inhibitors in caco-2 cells (100 μm amantadine, cimetidine or desipramine; 10 μm -estradiol and verapamil. values are presented as percent accumulation observed relative to control (without hoct inhibitor) of one experiment. cellular accumulation was normalized to cell protein amount. 15 30 60 90 120 180 0 500 1000 1500 2000 control verapamil desipramine cimetidine a c c u m a lt iv e m a s s t ra n s p o rt ( p m o l) 15 30 60 90 120 180 0 100 200 300 400 500 control verapamil desipramine cimetidine 15 30 60 90 120 180 0 100 200 300 400 500 control verapamil desipramine cimetidine time (minutes) a c c u m a lt iv e m a s s t ra n s p o rt ( p m o l) 15 30 60 90 120 180 0 100 200 300 400 500 control verapamil desipramine cimetidine time (minutes) 15 30 60 90 120 1800 100 200 300 400 500 control verapamil desipramine cimetidine time (minutes) 15 30 60 90 120 180 0 100 200 300 control verapamil desipramine cimetidine figure 4. accumulative mass transport from the apical to basolateral side after the exposure of 10 μm of erlotinib (a), gefitinib (b), sunitinib (f), sorafenib (d) and crizotinib (c). dasatinib (e) was used a concentration of 1 µm due to its higher activity in translational cell culture. all culturing was performed with and without the addition of either 10 μm verapamil, 100 μm desipramine or 100 µm cimetidine. values presented are from one representative experiment. cellular accumulation was normalized to cell protein amount. the hoct inhibitors desipramine, verapermil and cimetidine had little effect on the permeation of sunitinib, sorafenib and dasatinib (figures 4f, 4d and 4e, respectively). however, both crizotinib (figure 4c) and sunitinib demonstrated a small monolayer accumulation with verapamil and cimetidine whereas desipramine had little or no effect, sorafenib demonstrate little to no change in accumulation erlotinib gefitinib sorafenib sunitinib dasatinib crizotinib 0 100 200 300 400 verapamil desipramine cimetidine ß-estradiol amantadine r e la ti v e a c c u m u la ti o n ( % o f c o n tr o l) a f d b e c honeywell et al. admet & dmpk 4(1) (2016) 23-34 30 characteristics with all three inhibitors. however, crizotinib’s permeation was increased by desipramine and cimetidine unlike sunitinib. desipramine decreased cellular dasatinib accumulation in the transwell system while verapamil increased it, cimetidine had no effect. erlotinib gefitinib sorafenib sunitinib dasatinib crizotinib 0 50 100 150 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000 11000 12000 13000 control verapamil desipramine cimetidine m o n o la y e r a c c u m u la ti o n (n g d ru g / m g p ro te in ) figure 5. following the 3 hour exposure of the monolayer to erlotinib, gefitinib, sunitinib and sorafenib, crizotinib and dasatinib the cells were harvested and analysed for tki accumulation. exposure was determined both with and without the addition of 10 μm verapamil, 100 μm desipramine or 100 µm cimetidine. values presented are from one representative experiment. cellular accumulation was normalized to cell protein amount. discussion this study demonstrated that tkis can be transported, either passively or by an active process; for the latter both influx and efflux transporters are involved. caco-2 cells were used by two different approaches, as a colon cancer cell line to determine the passive and active accumulation of tki in the cells and as a model for intestinal epithelium using a transwell system as described earlier [33]. the temperature study demonstrated that uptake of crizotinib, sorafenib and sunitinib seems to be a passive process, but that of sunitinib, crizotinib and sorafenb seems to be passive. galetti et al. [41] also studied temperature dependency of gefitnib at 4 °c and 37 °c, but over a limited time period of sixty minutes with multiple points, showing a clear difference. next to gefitinib, dasatinib uptake seemed to be predominantly an active process. however, giannoudis et al. [42] postulated that the uptake is likely passive diffusion because of the lipophilic properties of dasatinib. on the other hand, they state that the role of other uptake transporters could not be excluded [43]. in our hands there is a clear difference between 4 °c and 37 °c, indicating an active transport. hu et al. [44] already reported that both sunitinib and sorafenib do not appear to rely on active transport to enter the cell. these drugs are also not a high affinity substrate for the abcfamily transporters, which is in agreement with the data reported by this study. however, passive transport can be masked by the lysosomal uptake leading to more accumulation in the cell at 37 °c as seen with sunitinib [22,39]. admet & dmpk 4(1) (2016) 23-34 transport mechanism of tk inhibitors doi: 10.5599/admet.4.1.275 31 cationic charge and high lipophilicity are requirements of oct dependent transport of drugs [28]. thomas et al. [27] already indicated the involvement of the human hoct1 in the active uptake of imatinib. however, since several of the six tkis (gefitinib, sunitinib and dasatinib) are positively charged [22] and all of the six tkis have high lipophilic properties, we hypothesized that some of the tkis are (partly) dependent on hoct uptake. the results indicate that the relative accumulation decreased for all tkis after the addition of one or more of the hoct inhibitors except for erlotinib. both erlotinib and dasatinib showed high relative accumulation after incubation with verapamil. this is probably due to the inhibition of p-gp, also induced by verapamil. de vries et al. [45] also suggested the involvement of p-gp in the efflux of erlotinib. giannoudis et al. [42] and gromicho et al. [43] suggested that p-gp is also involved in dasatinib dependent efflux. the effect of hoct inhibitors on sunitinib and crizotinib was similar, except for verapamil. both desipramine and cimetidine inhibit all three hocts. for most of the tkis, desipramine induced the highest decrease of relative accumulation. the different effects of cimetidine and desipramine, despite their same inhibitor function of hoct, might be due to concentration effects. some hoct inhibitors can be less effective at certain concentrations. only cimetidine and desipramine decreased the relative accumulation of sorafenib. it is possible that all three hoct should be inhibited to see an effect in relative accumulation of sorafenib, although hu et al. [44] reported that sorafenib does not appear to rely on active transport to enter the cell. nevertheless, none of the used hoct inhibitors are specific. therefore it cannot be concluded that the tkis are exclusively transported by hoct. for a better definition of the role of hoct transporter activity in the transport of tkis it is recommended to use more specific hoct inhibitors such as sirnas. in the caco-2 gut epithelial model system permeation is a sequential process, consisting of passive or active membrane transport at the apical site, possible trapping in the cell, and active or passive efflux at the basolateral site. moreover transport is not limited to the apical to basolateral transport determined in this series of experiments but basolateral to apical movement from the cell could play a crucial role in the absorption of these compounds. cimetidine, desipramine and verapamil were investigated in a model gut epithelial system to elucidate some of the relevance of the hoct role. due to the various locations of the different hocts in the intestinal model compared to the tumour model, different effects were observed. the hoct inhibitors did not have any effect on apical to basolateral transport of erlotinib. nevertheless an effect is seen in the accumulation of erlotinib in the caco-2 intestinal monolayer. therefore, verapamil, desipramine and cimetidine might act on different mechanisms within the cell. elmeliegy et al. [46] stated that erlotinib is a substrate for hoct2. in addition, minematsu et al. [28] showed that erlotinib had inhibitory effects on hoct1. in contrast to erlotinib, the hoct inhibitor desipramine affected apical to basolateral transport of gefitinib. since the inhibition of hocts by desipramine induced a higher apical to basolateral mass transport, it is suggested that the inhibition of hoct1 (on basolateral side) had a greater effect than the inhibition of hoct3 (on apical side), which means that there is less transport back into the cell from basolateral to apical side. the lower accumulation of gefitinib observed in the caco-2 intestinal monolayer after the incubation with desipramine supports this suggestion. the transport of sunitinib and crizotinib from apical to basolateral side differed. the hoct inhibitors showed only marginal effect on sunitinib transport from apical to basolateral side. in contrast, hocts seem involved in the apical to basolateral transport of crizotinib. desipramine increased mass transport of crizotinib, just like gefitinib, from apical to basolateral site. in addition, cimetidine had a similar effect as desipramine on crizotinib mass transport. the higher sunitinib and crizotinib accumulation in the caco-2 intestinal monolayer after the incubation of verapamil and cimetidine is probably masked by lysosomal accumulation [22,39]. just like sunitinib, sorafenib does not seem to be dependent on hoct transport from apical to basolateral side. the honeywell et al. admet & dmpk 4(1) (2016) 23-34 32 hoct inhibitors also showed only marginal effect on the accumulation of sorafenib in the caco-2 intestinal monolayer. just like sunitinib and sorafenib, dasatinib does not seem to be dependent on hoct transport from apical to basolateral side. dasatinib efflux is likely to be p-gp dependent and therefore an increase in dasatinib accumulation in the caco-2 intestinal monolayer is seen after the incubation with verapamil. desipramine decreased dasatinib accumulation in the caco-2 intestinal monolayer, so hocts might be involved in basolateral to apical transport of dasatinib. nevertheless, giannoudis et al. [42] suggested that the uptake of dasatinib is not likely to be dependent on hoct1 (on basolateral side). also gromicho et al. [43] suggested that effective uptake of dasatinib is not likely to occur with the help of hoct1. forchap et al. [47] concluded that dasatinib is not primarily transported by hoct1. it can be concluded that erlotinib is actively transported out of the cell, gefitinib and dasatinib uptake is also active while sorafenib uptake is passive. this study suggests that hocts might be involved in the uptake of gefitinib, sunitinib, crizotinib, sorafenib and dasatinib in the tumor model. in addition, hocts might be involved in the apical to basolateral transport of gefitinib and crizotinib in the intestinal model but to differing degree. overall, it can be concluded that the six tkis are very different in accumulation and transport. references [1] f. broekman, e. giovannetti, g.j. peters, world j. clin. oncol. 2(2) (2011) 80–93. [2] j. schlessinger, cell 103(2) (2000) 211–225. [3] s.r. hubbard, w.t. miller, curr. opin. cell biol. 19(2) (2007) 117–123. [4] a. gschwind, o.m. fischer, a. ullrich, nat. rev. cancer 4(5) (2004) 361–370. [5] m.a. lemmon, j. schlessinger, cell 141(7) (2010) 1117–1134. [6] p.b. allen, l.m. wiedemann, j. biol. chem. 271(32) (1996) 19585–19591. [7] d.r. duckett, m.d. cameron, expert opin. drug metab. toxicol. 6(10) (2010) 1175-1193 . [8] m.h. cohen, j.r. johnson, y.-f. chen, r. sridhara, r. pazdur, oncologist 10(7) (2005) 461–466. [9] m.h. cohen, oncologist 8(4) (2003) 303–306. [10] w.s. siegel-lakhai, j.h. beijnen, j.h.m. schellens, oncologist 10(8) (2005) 579–589. [11] e. lierman, i. lahortiga, h. van miegroet, n. mentens, p. marynen, haematologica 92(1) (2007) 27– 34. [12] r.c. kane, a.t. farrell, r. madabushi, b. booth, s. chattopadhyay, r. sridhara, r. justice, r. pazdur, oncologist 14(1) (2009) 95–100. [13] r.c. kane, a.t. farrell, h. saber, s. tang, g. williams, j.m. jee, c. liang, b. booth, n. chidambaram, d. morse, r. sridhara, p. garvey, r. justice, r. pazdur, clin. cancer res. 12(24) (2006) 7271–7278. [14] r. kumar, m.-c. crouthamel, d.h. rominger, r.r. gontarek, p.j. tummino, r.a. levin, a.g. king, br. j. cancer 101(10) (2009) 1717–1723. [15] e.p. rock, v. goodman, j.x. jiang, k. mahjoob, s.l. verbois, d. morse, r. dagher, r. justice, r. pazdur, oncologist 12(1) (2007) 107–113. [16] a. strimpakos, m.w. saif, k.n. syrigos, cancer metastasis rev. 27(3) (2008) 495–522. [17] m.w. saif, jop. j. pancreas 12(4) (2011) 316–321. [18] s.j. baker, e.p. reddy, mt. sinai j. med. 77(6) (2010) 573–586. [19] m. brave, v. goodman, e. kaminskas, a. farrell, w. timmer, s. pope, r. harapanhalli, h. saber, d. morse, j. bullock, a. men, c. noory, r. ramchandani, l. kenna, b. booth, j. gobburu, x. jiang, r. sridhara, r. justice, r. pazdur, clin. cancer res. 14(2) (2008) 352–359. [20] s.h.i. ou, drug des. devel. ther. 5 (2011) 471–485. admet & dmpk 4(1) (2016) 23-34 transport mechanism of tk inhibitors doi: 10.5599/admet.4.1.275 33 [21] i.b. muller, a.j. de langen, r.j. honeywell, e. giovannetti, g.j. peters, expert rev. anticancer ther. 16(2) (2016) 147–157. [22] da silva, c. g.; honeywell, r. j.; dekker, h.; peters, g. j. physicochemical properties of novel protein kinase inhibitors in relation to their substrate specificity for drug transporters. expert opin drug metab toxicol. 2015, 11 (1744-7607 (electronic)), 703–717. [23] p. di gion, f. kanefendt, a. lindauer, m. scheffler, o. doroshyenko, u, fuhr, j. wolf, u. jaehde, clin. pharmacokinet. 50(9) (2011) 551–603. [24] c. lemos, g. jansen, g.j. peters, br.j cancer 98(5) (2008) 857–862. [25] c. lemos, i. kathmann, e. giovannetti, c. calhau, g. jansen, g.j. peters, br.j cancer 100(7) (2009) 1120–1127. [26] s. shukla, z.-s. chen, s.v. ambudkar, drug resist. updat. 15(1-2) (2012) 70–80. [27] j. thomas, l. wang, r.e. clark, blood 104(12) (2004) 3739–3745. [28] t. minematsu, k.m. giacomini, mol. cancer ther. 10(3) (2011) 531–539. [29] k.m. giacomini, s.m. huang, d.j. tweedie, l.z. benet, k.l.r. brouwer, x. chu, a. dahlin, r. evers, v. fischer, k.m. hillgren, k.a. hoffmaster, t. ishikawa, d. keppler, r.b. kim, c.a. lee, m. niemi, j.w. polli, y. sugiyama, p.w. swaan, j.a. ware, s.h. wright, s.w. yee, m.j. zamek-gliszczynski, l. zhang, nat. rev. drug discov. 9(3) (2010) 215–236. [30] p. artursson, k. palm, k. luthman, adv. drug deliv. rev. 46(1-3) (2001) 27–43. [31] c. lemos, a. faria, m. meireles, f. martel, r. monteiro, c. calhau, eur. j. pharmacol. 682(1-3) (2012) 37–42. [32] m. roth, a. obaidat, b. hagenbuch, br. j. pharmacol. 165(5) (2012) 1260–1287. [33] r. honeywell, c. fatmawati, m. buddha, s. hitzerd, i. kathman, g. peters, admet and dmpk 3(1) (2015) 51–67. [34] r. honeywell, k. yarzadah, e. giovannetti, n. losekoot, e.f. smit, m. walraven, j.s. w.lind, c. tibaldi, h.m. verheul, g.j. peters, j. chromatogr. b anal. technol. biomed. life sci. 878(15-16) (2010) 1059–1068. [35] c. pauli-magnus, o. von richter, o. burk, a. ziegler, t. mettang, m. eichelbaum, m.f. fromm, j. pharmacol. exp. ther. 293(2) (2000) 376–382. [36] h. koepsell, b.m. schmitt, v. gorboulev, rev. physiol. biochem. pharmacol. 150 (2003) 36–90. [37] s. chong, s.a. dando, r.a. morrison, pharm. res. 14(12) (1997) 1835–1837. [38] s. yamashita, k. konishi, y. yamazaki, y. taki, t. sakane, h. sezaki, y. furuyama, j. pharm. sci. 91(3) (2002) 669–679. [39] k.j. gotink, h.j. broxterman, m. labots, r.r. de haas, h. dekker, r.j. honeywell, m.a. rudek, l.v. beerepoot, r.j. musters, g. jansen, a.w. griffioen, y.g. assaraf, r. pili, g.j. peters, h.m.w. verheul, clin.cancer res. 17(23) (2011) 7337–7346. [40] w.-k. lee, m. reichold, b. edemir, g. ciarimboli, r. warth, h. koepsell, f. thévenod, am. j. physiol. renal physiol. 296(6) (2009) f1504–f1513. [41] m. galetti, r.r. alfieri, a. cavazzoni, s. la monica, m. bonelli, c. fumarola, p. mozzoni, g. de palma, r. andreoli, a. mutti, m. mor, m. tiseo, a. ardizzoni, p.g. petronini, biochem. pharmacol. 80(2) (2010) 179–187. [42] a. giannoudis, a. davies, c.m. lucas, r.j. harris, m. pirmohamed, r.e. clark, blood 112(8) (2008) 3348–3354. [43] m. gromicho, j. dinis, m. magalhães, a.r. fernandes, p. tavares, a. laires, j. rueff, a.s. rodrigues, leuk. lymphoma 52(10) (2011) 1980–1990. [44] s. hu, z. chen, r. franke, s. orwick, m. zhao, m. rudek, a. sparreboom, s.d. baker, clin. cancer res. 15(19) (2009) 6062–6069. [45] n. de vries, t. buckle, j. zhao, j.h. beijnen, j.h.m. schellens, o. van tellingen, invest. new drugs 30(2) (2012) 443–449. honeywell et al. admet & dmpk 4(1) (2016) 23-34 34 [46] m.a. elmeliegy, a.m. carcaboso, m. tagen, f. bai, c.f. stewart, clin. cancer res. 17(1) (2011) 89– 99. [47] s.l. forchap, m. pirmohamed, r.e. clark, leukemia 26(3) (2012) 490–498. ©2016 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ manuscript doi: 10.5599/admet.5.2.390 75 admet & dmpk 5(2) (2017) 75-84; doi: 10.5599/admet.5.2.390 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index review expression of drug transporters in the human skin: comparison in different species and models and its implication for drug development hanan osman-ponchet * , alexandre gaborit, jean-michel linget and claire e. wilson nestlé skin health, les templiers , 2400 route des colles, bp 87, f-06902 sophia-antipolis *corresponding author: e-mail: hanan.osman-ponchet@galderma.com; tel.: +33-492-386-786; fax: +33-493-957-071 received: may 02, 2017; revised: june 15, 2017; published: june 22, 2017 abstract it is clear that many drug transporters (both abcs and slcs) are present in the human skin. different in vitro skin models can be used to investigate the role of drug transporters in the skin despite quantitative differences in expression profile across species. p-gp was shown to have an important influence on transdermal drug absorption in the skin and to function in “absorptive” transport, carrying substrate drugs from the skin surface to the dermis. this observation might be used to modulate drug distribution inside the skin. if drugs can be retained in the epidermis compartment by inhibition of the transporters, such property of the drug would be beneficial for treatment of dermatological diseases. therefore, it might be feasible to control transdermal delivery of drugs to specific locations in the skin, by modulating the function of the transporters in the skin. we are at the dawn of an exciting period where drug transporters might be novel targets for improvement of drug delivery to the skin and for pharmacological intervention. keywords solute-carrier transporter; atp-binding cassette transporter; skin absorption introduction the skin acts as an effective barrier against intrusion from the outside. the barrier may come in two ways: a physical one mainly through the stratum corneum, and a biochemical one as it has long been recognized that drug metabolizing enzymes are present in the skin: both phase i and phase ii enzymes are expressed and skin metabolism has been the subject of numerous reviews [1-3] mentioning specifically the exhaustive one from oesch et al., [4]. however, the skin is far from being a tight barrier for small lipophilic molecules as it is well known that there is a myriad of applications using the skin as the port of entry into the systemic circulation in the form of a patch, lotion, cream, ointment… [5]. the permeation through the skin of a toxic chemical may lead to serious adverse effects as the most unfortunate example of hexachlorophene poisoning in infants due to contaminated talc powder has shown [6]. so now during drug development of a new dermatological product it is mandatory to assess the uptake in the systemic circulation [7]. normal keratinocytes express drug metabolizing enzymes [8, 9] but the expression of these enzymes is http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:hanan.osman-ponchet@galderma.com h. osman-ponchet et al. admet & dmpk 5(2) (2017) 75-84 76 much higher in liver and intestine than in skin [10-13]. for most drugs the limiting factor for topical bioavailability is the ability to permeate through the skin and not the biochemical degradation by metabolism [14, 15]. however, it should be noted that for a compound like testosterone skin metabolism reduces sharply the topical bioavailability [2, 16]. transporters are by definition vectors, as they may have an intake or efflux function. but that notion of influx or efflux is a relative one as it depends on the cell or organ we are referring to (the efflux from one organ is the intake of another one). xenobiotic transporters are not specific, they may have several (or many) substrates and as many inhibitors (albeit with a different affinity). drug transporters in the skin belonging to both the atp-binding cassette (abc) [17-23] and solute carrier (slc) families [17, 23-25] have been identified in man and other animal species but our knowledge of these transporters is just at the beginning and is not as advanced as for drug metabolizing enzymes. we would like to review in the present paper our knowledge on drug transporters in the skin not by giving an exhaustive list of everything that has been done but on focusing on what is known and not known on expression and function of these transporters in the perspective of drug development. we will discuss the potential implication of the presence of these transporters on drug development both on the aspect of topical absorption and of being the target of pharmacological intervention. skin models different skin models can be used to characterize drug transporters in the skin, among them, skin organ-culture, 3d skin microtissue and 3d reconstructed epidermis models. skin organ-culture model skin organ-culture model was previously described [20, 24]. briefly, biopsies of fresh human skin samples excised during cosmetic surgery are collected and cultivated in chemically defined skin culture medium. 3d human skin micro-tissue model 3d human skin micro-tissue is a spherical in vitro skin model consisting of spherical skin equivalent (up to 500 μm diameter) containing fibroblasts (dermal core-like) surrounded by keratinocytes (epidermis-like) (figure 1a). this in vitro model provided in a 96-well format actually available as beta-testing for screening in dermatology is provided by insphero (switzerland). 3d reconstructed human epidermis model reconstructed human epidermis (rhe) consists of an air-lifted, living, multi-layered tissue construct produced in polycarbonate inserts in serum-free and chemically defined medium, featuring normal ultrastructure and functionality similar to human tissue in vivo (figure 1b). the rhe model episkin (france) is generated from primary normal human keratinocytes. this model most closely mimics normal skin, allowing the topical application of a great variety of products, solid, liquid, gel, cream… however, this skin model contains only one cell type, lacks skin appendages, and therefore is insufficient to adequately mimic the complexity of human skin. admet & dmpk 5(2) (2017) 75-84 expression of drug transporters in the human skin doi: 10.5599/admet.5.2.390 77 figure 1. 3d in vitro skin models. (a) 3d human skin microtissue. (b) 3d reconstructed human epidermis. gene expression of transporters in different species expression in human skin several studies have reported the expression of transporters in human skin [17-25]. their objectives and methods were not identical, but their results which cannot be exactly compared, show clearly the presence of abc and slc transporters in human skin. the semi-quantitative expression of human abc transporters in skin has been reported [21], and in this study most of the abca, abcb, abcc, abcd, abce, abcf, and abcg family members were found to be slightly (for abcg) to moderately-highly expressed for the other family members. it should be noted that in this study a significant inter-individual variability in the expression was observed as no expression was detected in some human samples: abca5 and abcf1 being the exception. in another study [20] mrna expression of abcb1, abcc1, abcc2 and abcg2 was measured in human skin organ-culture using rt-pcr. it was found that expression of abcc1 (mrp1) was very high compared to those of abcb1, abcc2, and abcg2 [20]. comparatively to the liver or kidney the expression of abc transporters in human skin is usually lower but we found that for abcc1 the reverse is true as there is more than 30 fold higher expression in human skin compared to liver [20]. many slc and slco genes were found to be expressed in human skin [23-25]. one more recent study assessed the expression of slc transporters in the human skin, liver and small intestine. while most of these transporters were moderately to highly expressed in the liver, their expression was limited in the skin and small intestine with, however, a rather high expression of slco2b1, slco4a1, slco16a1 and slc16a4 in the skin [25]. in another study, of 11 transporters assessed slc47a1 (mate1) was found with the highest expression in the skin [24]. comparatively with the liver it was shown that quantitatively the expression in the skin is at least 10 times lower than in the liver [24]. the study of transporters in the human skin is still in its infancy and in the 2 recent studies the expression was assessed by mrna levels and not by the protein level nor by transporter activity. the issue of interindividual variability observed in the expression is still an open one as the number of skin samples studied so far is rather small. the normal variability has to be known before assessing for example the observation that expression of abcc1 is 26 fold lower in the skin from skin cancer patients [21]. expression in 3d human skin models the main abc and slc transporters expressed in full thickness human skin in organ-culture are also expressed in both 3d human skin microtissue (figure 2a) and 3d reconstructed human epidermis (figure 2b). indeed, abcc1 (mrp1) and slc47a1 (mate1) are clearly expressed in those 2 models, indicating that both 3d in vitro skin models are suitable to evaluate drug transporters in the skin by studying their role, function, and regulation in the skin [26]. a b h. osman-ponchet et al. admet & dmpk 5(2) (2017) 75-84 78 figure 2. constitutive expression of abc and slc transporters in 3d in vitro human skin models. (a) 3d human skin microtissue. (b) 3d reconstructed human epidermis. quantitative real-time reverse transcription-pcr was used to investigate the constitutive expression of abc and slc transporters in 3d human skin models. gene expression was normalized to 3d human skin microtissue and reconstructed human epidermis gapdh to ensure equality of loading. all analyses were performed in triplicate. expression in skin from animal species abcc1 is the main abc transporter expressed in human skin, in göttingen mini pig skin, and in sprague dawley rat skin (figure 3a). slc47a1 is the main slc transporter expressed in human skin and in rat skin (figure 3b). moreover, the expression of slc47a1 is higher in human skin compared to rat skin. of the four slc transporters tested, none was expressed in mini pig skin, however, it should be noted that at present time, the gene sequence of slc47a1 is not yet known in mini pig. overall, these results show that expression profile of abc and slc transporters is qualitatively similar in skin of human and animal species tested. a b admet & dmpk 5(2) (2017) 75-84 expression of drug transporters in the human skin doi: 10.5599/admet.5.2.390 79 figure 3. constitutive expression of abc and slc transporters in skin of different animal species. (a) expression of abc transporters in rat, mini pig and human skin. (b) expression of slc transporters in rat and human skin. quantitative real-time reverse transcription-pcr was used to investigate the constitutive expression of abc and slc transporters in animal skin. gene expression was normalized to gapdh to ensure equality of loading. all analyses were performed in triplicate. regulation of transporter expression in the skin the effect of rifampicin on the expression of abcc1 and slc47a1, the most expressed transporters in the skin, was studied in human skin organ-culture after a 72 hour treatment period [20, 24]. rifampicin induced a significant decrease of expression of abcc1 [20] and slc47a1 [24]. expression of slc47a2 was also decreased but not to a significant level due to the variability of expression. interestingly uv light exposure decreased the expression of both slc47a1 and slc47a2, and again there is a large interindividual variability and the decrease is not significant for slc47a2 [24]. it is difficult at this stage to draw firm conclusions as the variability in the expression blunts the results, but there is nevertheless a trend towards regulation of a few transporters in the skin. this aspect should receive more attention in the future. localization and function of transporters in the skin the skin is not a homogeneous organ, so a differential expression of transporters in different skin compartments would not be unexpected. expression of abcb1 in human skin was found to be mainly restricted to dermal components [22], sweat ducts, vessels, nerve sheaths and muscle [27], confirming an earlier observation in which p-gp in neonatal murine skin were present on the surface of basal a b h. osman-ponchet et al. admet & dmpk 5(2) (2017) 75-84 80 keratinocytes and outer root sheath cells of hair follicles [28]. as a monoclonal antibody for human mrp1 is available we studied the localization of this transporter by immunohistochemistry, and it was found that mrp1 is mainly localized at the interface of epidermis-dermis and in the dermis in the sweat gland [20] and hair follicle (figure 4). the concentration of expression of mrp1 in the hair follicle was confirmed in a more recent study [19]. on the other hand, bcrp was immunohistochemically identified in human epidermis and dermal endothelial cells [21] and in the hair follicle [22]. there are less data about the localization of slc transporters in the skin; in one study slc21a9/oatp-b was detected in the epidermis of skin specimens by immunofluorescence and immunohistochemistry [9]. figure 4. localization of mrp1 in the hair follicle by immunohistochemistry. potential function of drug transporters in the skin mdr1/p-gp is thought to be an efflux transporter so it was a kind of a surprise when this transporter was found to be involved in the absorptive drug transport of rhodamine 123 and itraconazole in the murine skin [18]. the penetration was lower in knockout mice (mdr1a/1b-/-) than in wild-type mice, and propranolol, a p-gp inhibitor, decreased the concentration of itraconazole in the skin of wild-type mice but not in gene knockout mice [18]. this absorptive function was confirmed with flurbiprofen and indomethacin [17]. recently p-gp in skin was found to contribute to the transdermal absorption of topical corticoids and ingenol-3-angelate [29, 30]. on the other hand, bcrp was shown to contribute to the transdermal transport of rhodamine 123, a substrate of p-gp/mdr1 [22]. there was still the possibility that abc transporters (other than mdr1) expressed in the skin might play a role to prevent xenobiotics from entering the systemic circulation, however, [ 14 c]-grepafloxacin was shown to be transported in mouse from the skin back to the blood [31]. in another study [20], skin absorption of rhodamine 123 and vinblastine were significantly reduced in ex vivo human skin by verapamil, an inhibitor of mrp1 [20]. moreover, mrp1 was also found to actively reduce the accumulation of substrates within the hair follicle [19]. mdr1/p-gp, mrp1 and bcrp facilitate the flux of substrates towards the inner layers of the skin and not the reverse. so the simple concept of the skin being a barrier preventing the organ from the environment is clearly not entirely valid. the exact and final function of these transporters is still a matter of conjecture. these early studies all point out to a higher concentration of these transporters around the sweat gland and the hair follicle. the expression density of these transporters varied with the transporter being rather uniformly expressed for abcb1 to specifically expressed in the sweat gland and the dermal papilla for abcg2 [19]. even though the function of this differential expression has yet to be unravelled, it is admet & dmpk 5(2) (2017) 75-84 expression of drug transporters in the human skin doi: 10.5599/admet.5.2.390 81 conceivable that mrp1 provides protection against both xenobiotic and oxidative insult in the hair follicle thereby preventing melanocyte depletion and hair greying [19], and that bcrp may protect the hair follicle against cytotoxic insults from chemotherapeutics, thereby preventing or reducing the associated alopecia [19]. there are fewer studies about slc transporters in the skin, but the uptake of large organic cations such as drugs in keratinocytes was shown to be an active process mediated by members of the slc transporter family, which can be modulated by specific competitors in vivo [9]. overall, the real function of transporters in the skin is still largely unknown. implication for drug development absorption through the skin is thought to be largely dependent upon the physicochemical properties of the drug and to follow the concentration gradient across the skin according the fick’s first law. this is a rough approximation as the pharmacokinetics of drug permeation is actually far more complex and should be modelized by a multicompartmental system [32]. the presence of transporters would either facilitate or hamper the topical absorption depending upon the direction of the flux. furthermore, potential drug-drug interactions would modulate the permeation process and could either induce or inhibit the flux through a skin layer creating a local zone of abnormal concentrations. most drugs are likely not to be affected, but in a few instances we may get a clinically significant effect. there is now evidence that substrates of p-gp have a facilitated absorption through the skin. we do not know if these theoretical considerations actually play a clinical role in the skin permeation of drugs, but in theory we could modulate the flux to either improve or decrease the systemic exposure and to increase or decrease the skin concentrations. in theory, at least, inhibitors of p-gp would reduce the systemic exposure of corticoids [29] while inducers might improve the delivery of chemotherapeutic agents into the skin. we are at the early stage of transporter research in the skin, and we have to obtain more data about the quantitation of functional transporters and not just looking at the mrna expression. the interindividual variability in the expression and activity of transporters is another point to clarify. do we have genetic or disease-induced effects in the expression? so the usual drug-drug interactions regulatory guidances for other tissues (liver, intestine, kidney …) [34, 35] may not apply until more data are obtained. several studies have pointed out the unusually high expression of transporters around the hair follicle and sweat gland [19, 20, 22]. the function of these transporters is not yet clear, but more research is actively needed in order to clarify the potential role in the excretory function or the protection of these organelles [19], as this may have a great potential for intervention. although mate transporters belong to slc family, they are an efflux transporter and play a role in renal and biliary excretion of endogenous and exogenous compounds including xenobiotics and drugs [36, 37]. cis-inhibition studies have suggested that mate transporters may be involved in the transport of endogenous steroid hormones like testosterone [38, 39]. skin is known as steroidogenic organ and skin cells produce steroid hormones like testosterone [40]. we suppose that mate transporters are responsible for the secretion of steroid hormones through the plasma membrane in the skin. moreover, mate transporters are down-regulated by rifampicin which suggests a role in drug-drug interactions. of the hypothesis that mini pig studies may better reflect human drug-induced toxicities than other species, it is mandatory to know why none of slc transporters is expressed in the skin of göttingen mini pig. h. osman-ponchet et al. admet & dmpk 5(2) (2017) 75-84 82 one of the limitations of most of these studies is that the expression of transporters has been compared by quantifying the expression of the transporter mrnas. this method has its shortcomings, as for transporters, the mrna expression does not correlate well with protein expression of transporters and protein activity [41, 42]. quantification of protein concentration of transporter in the skin, using surrogate signature peptides and liquid chromatography-mass spectrometry, overcomes these limitations [43-45]. conclusions it is clear that many drug transporters (both abcs and slcs) are present in human skin and different in vitro skin models can be used to investigate the role of drug transporters in the skin. despite a significant difference in expression profile across species, we are at the dawn of an exciting period where the functional aspects need to be investigated. drug transporters were shown to have an important influence on transdermal drug absorption in the skin and to function in “absorptive” transport, carrying drugs from the skin surface to the dermis. this observation might be used to modulate drug distribution inside the skin. if drugs can be retained in the epidermis compartment by inhibition of the transporters, such property of the drug would be beneficial for treatment of dermatological diseases. therefore, it might be feasible to control transdermal delivery of drugs to specific locations in the skin, by modulating the function of the transporters in the skin. thus, drug transporters might be a novel target for improvement of drug delivery to the skin. references [1] c.k. svensson, drug metabolism and disposition, 37 (2009) 247-253. [2] a. mavon, c. jacques-jamin, cosmetics & toiletries, 128 (2013) 98-103. [3] n. manevski, p swart, k.k. balavenkatratam, b. bertschi, g. camenish, o. kretz, h. schiller, m. walles, b. ling, r. wettstein, d.j. schaefer, p. itin, j. ashton-chess, f. pognan, a. wolf, k. litherland, drug metabolism and disposition, 43 (2015) 126-139. [4] f. oesch, e. fabian, k. guth, r. landsiedel, archives of toxicology, 88 (2014) 2135-2190. [5] https://en.wikipedia.org/wiki/topical_medication, (april 2017). [6] g. martin-bouyer, m. toga, r. lebreton, p.d. stolley, j. lockhart, the lancet, 319 (1982) 91-95. [7] non clinical safety evaluation of reformulated drug products and products for administration by an alternate route, fda guidance for industry and review staff , oct 2015. [8] j.m. baron, d. holler, r. schiffer, s. frankenberg, m. neis, h.f. merck, f.k. jugert, journal of investigative dermatology, 116 (2001) 541-548. [9] r. schiffer, m. neis, d. höller, f. rodriguez, a. geier, c. gartung, f. lammert, a. dreuw, g. zwadloklarwasser, h. merck, f. jugert, j.m. baron, journal of investigative dermatology, 120 (2003) 285291. [10] h. schaeffer, t. redelmeier, skin barrier: principles of percutaneous absorption, s. karger, basel, ch, 1996. [11] h. swanson, chemico-biological interactions, 149 (2004) 69-79. [12] a. williams, transdermal and topical drug delivery, pharmaceutical press, london, 2003. [13] m. pham, j. magdalou, m. totis, s. fournel-gigleux, g. siest, b. hammock, biochemical pharmacology, 38 (1989) 2187-2194. [14] k. madison, journal of investigative dermatology, 121 (2003) 231-241. [15] m. prausnitz, s. mitragotri, r. langer, nature reviews drug discovery, 3 (2004) 115-124. [16] c. jacques, e. perdu, e.l. jamin, j.p. cravedi, a. mavon, h. duplan, d. zalko, skin pharmacol physiol, 27 (2014) 188-200. admet & dmpk 5(2) (2017) 75-84 expression of drug transporters in the human skin doi: 10.5599/admet.5.2.390 83 [17] q. li, h. tsuji, y kayo, y. sai, y. kubo, a. tsuji, journal of controlled release, 110 (2006) 542-546. [18] k. ito, h.t. nguyen, y. kato, t. wakayama, y. kubo, s. iseki, a. tsuji, journal of controlled release, 131 (2008) 198-204. [19] i.s. haslam, c. el-chami, a. shahmalak, c.a. o’neill, r. paus, british journal of dermatology, 172 (2015) 1562-1572. [20] h. osman-ponchet, a. boulai, m. kouidhi, k. sevin, m. alriquet, a. gaborit, b. bertino, p. comby, b. ruty, drug metabolism and drug interactions, 29 (2014) 91-100. [21] s. takenaka, t itoh, r. fujiwara, pharmacology research & perspectives, 1 (2013) 1-8. [22] n. hashimoto, n. nakamichi, s. uwafuji, k. yoshida, t. sugiura, a. tsuji, y. kato, journal of controlled release, 165 (2013) 54-61. [23] k. bleasby, j.c. castle, c.j. roberts, c. cheng, w.j. bailey, j.f. sina, a.v. kulkarni, m.j. hafey, r.evers, j.m. johnson, r.g. ulrich, j.g. slatter, xenobiotica, 36 (2006) 963-988. [24] m. alriquet, k. sevrin, a. gaborit, p. comby, b. ruty, h. osman-ponchet, admet & dmpk, 3 (2015) 34-44. [25] r. fujiwara, s. takenada, m. hashimoto, t. narawa, t, itoh, scientific reports, 4 (2014) 1-8. [26] h. osman-ponchet, 7th euro-global summit on toxicology and applied pharmacology, october 2426, 2016 rome, italy. [27] c. skazik, j. wenzel, y. marquardt, a. kim, h.f. merk, d.r. bickers; j. m. baron, experimental dermatology, 20 (2011) 450-452. [28] m.a. sleeman, j.d. watson, j.g. murison, journal of investigative dermatology, 115 (2000) 19-23. [29] n. hashimoto, n. nakamichi, e. yamazaki, m. oikawa, y. mauo, a.h. schinkel, y. kato, international journal of pharmaceutics, 521 (2017) 365-373. [30] l. li, s. shukla, a. lee, s.h. garfield, d. j. maloney, s.v. ambudkar, s.h. yuspa, cancer research, 70 (2010) 4509-4519. [31] q. li, y. kato, y. sai , t. imai, a. tsuji, pharmaceutical research, 22 (2005) 842-846. [32] y. dancik, o.g. jepps, m.s. roberts, in dermal absorption and toxicity chapter 11, university of queensland, 2013, pp 179-207. [33] f.g.m. russel, enzyme and transporter-based drug-drug interactions. progress and future challenges. pang, k.s.rodrigues, a.d. peter, r.m. (eds). 2010, hardcover. [34] guideline on the investigation of drug interactions ema, cpmp/ewp/560/95 rev.1 june 2012. [35] guidance for industry. drug interaction studies. studies design, data analysis, implications for dosing and labeling recommendations fda, cder, feb 2012. [36] m.h. brown, i.t. paulsen, r.a. skurray, molecular microbiology, 31 (1999) 394-395. [37] h. omote, m. hiasa, t. matsumoto, m. otsuka, y. moriyama, trends in pharmacological sciences, 27 (2006) 587-593. [38] m. otsuka, t. matsumoto, r. morimoto, s. arioka, h. omote, y. moriyama, proceedings of the national academy of sciences usa, 102 (2005), 17923-17928. [39] m. hias, t. matsumoto, t. komatsu, y. moriyama, american journal of physiology cell physiology, 291 (2006) 678-686. [40] a. slominski, j. wortsman, endocrine reviews, 21 (2000) 457-487. [41] s. ohtsuki, o. schaefer, h. kawakami, t. inoue, s. liehner, a. saito, n. ishiguro, w. kishimoto, e. ludwig-schwellinger, t. ebner, t. terasaki, drug metab dispos, 40 (2012) 83-92. [42] b. prasad, r. evers, a. gupta, c.e. hop, l. salphati, s. shukla, s.v. ambudkar, j.d. unadkat, drug metab dispos, 42 (2014) 78-88. [43] l. wang, b. prasad, l. salphati, x. chu, a. gupta, c.e. hop, r. evers, j.d. unadkat, drug metab dispos, 43 (2015) 367-374. h. osman-ponchet et al. admet & dmpk 5(2) (2017) 75-84 84 [44] b. prasad, k. johnson, s. billington, c. lee, g.w. chung, c. d. brown, e.j. kelly, j. himmelfarb, j.d. unadkat, drug metab dispos, 44 (2016) 1920-1924. [45] h.j. burt, a.e. riedmaier, m.d. harwood, h.k.crewe, k.l. gill, s. neuhoff, drug metab dispos, 44 (2016) 1550-1561. ©2017 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ manuscript doi: 10.5599/admet.4.3.348 182 admet & dmpk 4(3) (2016) 182-185; doi: 10.5599/admet.4.3.348 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index editorial adme/dmpk in the development and use of tyrosine and serine-threonine kinase inhibitors godefridus j. peters * , kin tam + editors: admet & dmpk * department of medical oncology, vu university medical center, amsterdam, the netherlands. gj.peters@vumc.nl. + faculty of health sciences, university of macau, taipa, macau. kin_tam@umac.mo. since the introduction of the tyrosine kinase inhibitor (tki) imatinib, for treatment of chronic myeloid leukemia (cml) about 30 new related compounds have been registered by the us fda, the emea and other registration authorities for various types of cancers, both solid tumors and hematological malignancies. some drugs have also been registered for other diseases, such as inflammation. these compounds do not only target tyrosine kinases, but also serine-threonine kinases [1, 2]. in a recent review we showed that these compounds show a large variation in their physico-chemical properties [3], which affect their adme properties. in general these small compounds (mostly < 500 kd) are poorly soluble, which may lead to a solubility limited absorption, while uptake through the gut is dependent on protein binding [4]. this leads to a large variability in pharmacokinetics (pk) [5]. it appeared that all these inhibitors obey all common parameters that dictate the variability of pk (figure 1). therefore for several of these drugs, it would be better to give them with food; despite this experience pharmaceutical companies often advise to give the drugs without food, leading to an even worse bioavailability [6]. figure 1. sources of variability in pharmacokinetics (pk). cam, complementary and alternative medicine. http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:gj.peters@vumc.nl mailto:kin_tam@umac.mo admet & dmpk 4(3) (2016) 182-185 tki special issue editorial doi: 10.5599/admet.4.3.348 183 protein kinases play important roles in signal transductions cascades that govern many cellular events (fig. 2), including cellular stability, selective growth disadvantage or apoptotic cell death [1]. these compounds are often referred to as targeted therapy, since the drugs are designed to inhibit a specific pathway in the cancer cell, usually a pathway which is essential for the cancer cell to survive. this property is often referred to as oncogene addiction. examples are the above-mentioned imatinib, which is targeted against the bcr-abl kinase, which only occurs in cml. imatinib is also targeted against c-kit, which is specifically expressed in gastrointestinal stromal tumors (gist). another early example includes the epidermal growth factor receptor (egfr) which is increased in adenocarcinoma, a subtype of non-small lung cancer (nsclc). adenocarcinomas with an activating mutation in egfr are sensitive to erlotinib and gefitinib [7]. another example is the alk-eml translocation in adenocarcinoma [8], against which crizotinib is targeted. to develop selective tkis for a particular tumor a thorough understanding of cellular signaling cascades is of utmost importance, not only in the tumor but also in its neighboring stroma, endothelial cells and immune cells such as various macrophage populations. also, communication between these cells needs to be elucidated. clearly the research and development along this direction will continue in the coming decades. figure 2. simplified scheme of the action of a receptor tyrosine kinase. the receptor is stimulated by a ligand, often leading to a dimerization, such as with egfr. the intracellular tyrosine site is phosphorylated with atp as the phosphate donor, leading a cascade of signals due to the transfer of a signal to one or more kinases. finally the last effector (usually a transcription factor) is activated and passes the nuclear membrane, resulting in a transcriptional modification. a tyrosine kinase inhibitor binds to the atp binding site resulting in inhibition of this pathway. in the early discovery / development phases of these compounds, the adme and dmpk aspects were sometimes neglected, since the focus was often on potency and selectivity against a particular kinase target of interest. for the first tki to be registered, imatinib, it is now clear that its efficacy is dependent on its pharmacokinetics, a certain trough level (about 1000 ng/ml, about 2 μm) is associated with a better response, although elevated trough levels (>1500 ng/ml) are associated with toxicity [9, 10]. also for sorafenib, a multikinase inhibitor of the vascular endothelial growth factor receptor (vegfr) and the wildtype ras-raf pathway, it was shown that the efficacy was related to its pharmacokinetics. moreover, some of its metabolites are substrates for several influx and efflux pumps [10, 11]. these properties lead to suboptimal efficacy and unexpected toxicity. for some drugs this resulted in development failures in the g. peters and k. tam admet & dmpk 4(3) (2016) 182-185 184 late phases. a recent example is the discontinuation of rociletinib [7], an inhibitor of the t790m mutation of efgr. this t790m mutation commonly develops after treatment with erlotinib or gefitinib. rociletinib was granted breakthrough therapy status from the fda for accelerated development, but in the phase iii trial severe grade 3 toxicity (hyperglycemia) was observed in 22 % of the patients due to inhibition of the insulin-like growth factor receptor (igfr). thirty-five percent of the patients had to take a glucose lowering drug (usually metformin). furthermore some responders could not be verified at an independent evaluation, with a drop in response rate from 59 to 32 %. this compound would have benefited from a more extensive evaluation. another specific inhibitor of the t790m mutation of egfr, osimertinib, did not suffer from this problem, and was recently registered by the fda after a similar breakthrough therapy status and an accelerated development program. in this special issue of admet and dmpk we intended to illustrate a number of issues important in the development of various protein kinase inhibitors, with a special focus on the admet and dmpk aspects of these compounds, and how to properly incorporate the basic admet and dmpk information to optimize the design of clinical trials. the papers will be divided between the september and december issues. thompson et al [12] give an overview on the mechanism of action of currently used egfr and alk directed compounds, describing the issues mentioned above. shah [13] describes a less common side-effect of these compounds, cardiovascular toxicity. this type of toxicity is also quite common for several of the chemotherapeutic agents and oncologists usually know how to deal with these patients. according to shah the hallmark of cardiovascular function, qt aberrations, are often overemphasized. this toxicity was also observed for rociletinib, but not a reason for its withdrawal. pott et al [14] describe a cross-talk between two signaling pathways. in the forthcoming issue interaction between the cytotoxic drug gemcitabine and the c-met inhibitor crizotinib will be described, as well as the development of a new formulation for a novel polo-like inhibitor, which had to take into account all above-described problems, such as poor solubility, low permeability and the crystalline nature. next another paper will focus on an alternate target for rock inhibitors, ocular disease. earlier the anti-angiogenic antibody bevacizumab was already indicated for ocular diseases as well. other aspects which will receive attention are the role of various efflux pumps for various tyrosine kinases and how to use docking models to develop kinase inhibitors. with this selection of papers we want to emphasize that with proper attention to the role of admet the use of the protein kinase inhibitors can be improved and extended to other applications, while for some drugs development should have been discontinued earlier. it can also be concluded that dose adaptation is sometimes necessary due to poor admet properties. references [1] f. broekman, e. giovannetti, g.j. peters. world j. clin. oncol. 2, (2011) 80-93. [2] p. wu, t.e. nielsen, m.h. clausen. trends in pharmacological sciences 36 (2015) 422-439. [3] c.g. da silva, r.j. honeywell, h. dekker, g.j. peters. expert opinion drug metab. toxicol. 11 (2015) 703-717. [4] r.j. honeywell, c. fatmawati, m. buddha, s. hitzerd, i. kathman, g.j. peters. admet & dmpk 3 (2015) 51-67. [5] a. sparreboom, j. verweij. clin pharmacol. ther 85 (2009) 113-117 [6] m.j. ratain. j clin oncol. 29 (2011) 3958-9. [7] n. van der steen, c. caparello, c. rolfo, p. pauwels, g.j. peters, e. giovannetti. oncotarget & therapy (2016) in press. [8] i.b. muller, a.j. de langen, r.j. honeywell, e. giovannetti, g.j. peters. exp review anticancer therapy 16 (2016) 147-157. admet & dmpk 4(3) (2016) 182-185 tki special issue editorial doi: 10.5599/admet.4.3.348 185 [9] c. delbaldo, e. chatelut, m. ré, a. deroussent, s. séronie-vivien, a. jambu, p. berthaud, a. le cesne, j.y. blay, g. vassal. clin cancer res. 12(20 pt 1) (2006) 6073-8. [10] s.d. baker, s. hu. clin pharmacol ther 85 (2009) 208-211. [11] e.i. zimmerman, s. hu, j.l. roberts, a.a. gibson, s.j. orwick, l. li, a. sparreboom, s.d. baker. clin cancer res. 19 (2013) 1458-66. [12] j.r. thompson, s.p. menon, g.k. dy. admet & dmpk 4 (2016) 186-211. [13] r. shah. admet & dmpk 4 (2016) 212-231. [14] g.b. pott, m. tsurudome, j. bui, c. gardner, m.l. goalstone. admet & dmpk 4 (2016) 232-240. ©2016 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ doi: 10.5599/admet.4.2.325 179 admet & dmpk 4(2) (2016) 179-181; doi: 10.5599/admet.4.2.325 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index education explore postgraduate biomedical engineering course integration between medical signal processing and drug development: example for drug development in brain disease yin tian*, li yang, zhongyan wang, huiling zhang, wei xu, shuxing zheng, haying zhang, dechun zhao bio-information college, chongqing university of posts and telecommunications, chongqing 400065, china * corresponding author: e-mail: tiany20032003@163.com; tel.: +86-23-62460536; fax: +86-23-62460536 received: june 28, 2016; published: july 03, 2016 abstract medical signal processing is a compulsory course in our university’s undergraduate biomedical engineering programme. recently, application of medical signal processing in supporting new drug development has emerged as a promising strategy in neurosciences. here, we discuss the curriculum reformation in biomedical signal processing course in the context of drug development and application in central nervous system, with a particular emphasis in knowledge integration. keywords biomedical signal processing; drug development; course integration; curriculum reformation 1. introduction modern medicine development technology is a multi-disciplinary topic and has received considerable attention in recent years. in order to develop a solid foundation, wide knowledge, high cultural quality, and strong adapt ability of multi-disciplinary talent, we are carrying education reform on personnel training mode and optimized programme curriculum in the major of biomedical engineering. the new curriculum aims to enhance strong links between theories and practical applications with wide range of content covering multi-disciplinary topic areas. without a carefully planned curriculum, it is difficult to cover this course with the allocated contact hours. therefore, this article will explore the theoretical and experimental aspects of combining biomedical signal processing and central nervous system (cns) drug development [1]. 2. curriculum integration the drug development technology using medical signal processing is an international frontier in the field of modern medical technology. we propose to focus on the application of medical signal processing on pharmacokinetics and pharmacodynamics of the drug development in teaching, highlighting combination with the implementation principles, the technical methods and applications. pharmacokinetic (pk) is http://www.pub.iapchem.org/ojs/index.php/admet/index yin tian et al. admet & dmpk 4(2) (2016) 179-181 180 mainly a quantitative study on drug processing in organisms (including absorption, distribution, metabolism and excretion) and described the dynamic regularity of drug in the body by the mathematical principle and methods. pharmacodynamics (pd) examines the effect, mechanism and quantity discipline of the drug on the body, namely under the action of drugs, the body organs physiological function and metabolic activity of cells. both pk and pd make up the foundation of modern pharmacology study. combination of these disciplines could indicate the time process of pharmacological effects caused by a given certain dosage. study of pk/pd relationship not only helps to guide the clinical medication correctly, but also explore mechanisms of drug action, new drug evaluation and the development of new preparations [2]. recently, a variety of integrated application of brain imaging technology has become a development trend of modern drug research and development, especially in the cranial nerve disease (such as alzheimer's and parkinson's disease) and mental illness (such as schizophrenia/depression) in drug development for biomarkers. electroencephalogram (eeg) and magnetoencephalogram (meg) can provide high temporal resolution of cognitive activity information. functional magnetic resonance imaging (fmri) and positron emission layer tracing technique (pet) can provide high spatial resolution information. mapping brain structure by registration and fusion of these two kinds of information (eeg/meg and fmri /pet) has become an important technical route. this course may be considered to offer at the postgraduate level. the students have already learned in their undergraduate courses in physiological basis, signal and system, biomedical signal processing, pet imaging technology, as well as biomedical electronics and detection technology. the study of the theory of these courses has laid a solid foundation. the setting of the proposed course contents are as follows: (1) brain disease of physiology and disease, (2) drug development basis on pk and pd, (3) eeg signal acquisition and analysis and (4) combination of eeg and cns drug development. 3. combining with teaching and experiment from the curriculum setting, this course focuses on the biomedical signal processing as a means of technical analysis. during the evaluation of drug performance in the course of central nervous system (cns) drug development, eeg signals and metabolic signals elicited by brain activity are recorded and analysed in the time domain and frequency domain using the time-frequency analysis method associated with cognitive activities to brain connectomes. these involve strong theoretical knowledge and practical techniques in biomedical signal processing and medical image processing. to improve students' initiative to acquire knowledge and problem solving ability, it is important to development a holistic teaching method. to this end, we combine both the teaching in theory and experiments as well as substantiate further with the training on preliminary scientific research ability and practical applications of the medical instrument. 4. existing problems and solving suggestion the difficulties in the integration of cns drug development and biomedical signal processing are: wide course content span involving multidisciplinary sciences (e.g., signal and system, signal processing and physiology, etc.), too close to the current most advanced science and technology and insufficient contact hours. many students might have already forgotten the basic knowledge they have learned in prerequisite courses taken in the previous semesters. it may be difficult for them to grasp the new contents in the first instance. therefore, the teacher should offer revision lessons at the beginning of the course, so that the students can easily get up to speed in their learning. interactive teaching is highly recommended. teacher can then adjust the speed of the lessons to ensure effectiveness of the teaching with sufficient explanation admet & dmpk 4(2) (2016) 179-181 medical signal processing and drug development doi: 10.5599/admet.4.2.325 181 in the course materials in order to effectively help students to digest and consolidate what they learned. at the same time, through the hands on experiment data processing, the student’s ability to analyse actual data could be improved, and conducive to the follow-up student’s self-experimental design. in addition to explain the new knowledge points in the teaching, the teacher are encouraged to be proactive to help students to solve practical pk/pd combined with signal processing analysis. the use of existing programming software (matlab), open software emse, eeglab analysis software to carry out the signal processing analysis is highly recommended and should be accomplished in an interactive manner. this helps not only reducing the analysis time, but also minimizing the needs to perform tedious numerical operations. at the same time, the students could save time in the interpretation of the analysis results. 5. conclusion multidisciplinary curriculum integration is the current focus of the teaching reform in colleges and universities. using a postgraduate biomedical engineering course as an example, we have explored the educational reform by the integration of both biomedical signal processing and pd/pd on cns drug development. acknowledgements we thank the support from cqupt’s educational projects (#ykc2015007 and #xjg1519). references [1] y. tian, z.d. ding, z.y. wang, h.l. zhang, 4 th world conference on physico-chemical methods in drug discovery and development (september, 2015). [2] l.h. reddy, d. bazile, advanced drug delivery reviews 71 (2013) 34-57. ©2016 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ manuscript doi: 10.5599/admet.1.2.3 6 admet & dmpk 1(2) (2013) 6-16; doi: 10.5599/admet.1.2.3 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper physico-chemical profiling of the ace-inhibitor lisinopril: acidbase properties krisztina takács-novák*, katalin deák, szabolcs béni, gergely völgyi semmelweis university, department of pharmaceutical chemistry h-1092 hőgyes endre street 9. budapest, hungary *corresponding author: krisztina takács-novák; e-mail: novak.krisztina@pharma.semmelweis-univ.hu; tel.: +36 1 215 5241; fax: +36 1 217 0891 received: january 7 th , 2013; revised: february 1 st , 2013; published: february 15 th , 2013 abstract the acid-base chemistry of a tetraprotic, ampholyte ace inhibitor, lisinopril, was studied by different methods. potentiometry in aqueous medium and a co-solvent technique in methanol-water mixtures as well as 1 h nmr-ph titration were applied for the highly precise measurement of protonation macroconstants. the log k values of lisinopril (at 25.0°c and 0.15 m ionic strength) were found: log k1 = 10.75 ± 0.01, log k2 = 7.13 ± 0.01, log k3 = 3.13 ± 0.01, log k4 = 1.63 ± 0.01, calculated as an average of the best two values obtained by independent methods. nmr -ph titration was used to assign the constants to the functional groups and for the examination of site-specific, submolecular basicities of the molecule (determination of protonation microconstants). in the first two well-separated protonation steps, the macroand microconstants were identical and assigned to the primary amino group (log k1 = log k a ) and to the secondary amine basicity (log k2 = log k b a), respectively. the two carboxylates exhibited overlapping protonation characterised first by microconstants (log d abk = 2.15 log c ab k = 3.10), revealing that the carboxylate on the proline ring has nine times greater intrinsic basicity than the carboxylate on the side chain. the distribution of protonation species (lis 2; hlis ; h2lis; h3lis + ; h4lis 2+ ) and microspecies (abc; abd) as a function of ph was calculated and used to interpret the pharmacokinetic and pharmacodynamic properties of lisinopril. keywords proton speciation; logk; potentiometry; nmr-ph titration introduction angiotensin converting enzyme (ace) inhibitors serve as fundamental medicines in the treatment of hypertension, one of the most prevalent chronic diseases nowadays. out of the numerous currently available ace-inhibitors, lisinopril – (2s)-1-[(2s)-6-amino-2[[(1s)-1-carboxy-3-phenylpropyl]amino]hexanoyl]-pirrole-2-carboxylic acid (fig. 1) – belongs to the proline-containing structures. http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:novak.krisztina@pharma.semmelweis-univ.hu admet & dmpk 1(2) (2013) 6-16 lisinopril speciation doi: 10.5599/admet.1.2.3 7 figure 1. chemical structure of the ace inhibitor lisinopril due to its safe and effective properties, lisinopril is widely used in the therapy of essential hypertension, symptomatic and asymptomatic left ventricular systolic dysfunction, post-myocardial infarction, renal failure and diabetic nephropathy [1]. lisinopril is administered orally usually in a daily dose of 2.5-40 mg. its oral bioavailability is between 25-30 % [2]. absorption from the gastrointestinal tract does not depend on nutrition. six hours are needed to reach the maximum plasma concentration. it hardly binds to plasma proteins. lisinopril is not metabolised, and the absorbed drug is primarily excreted invariably in the urine [3]. a vast literature describes the pharmacology, mechanism of action, pharmacokinetics and analytics of lisinopril [3-7]. however, surprisingly little information is available on its physico-chemical properties. lisinopril is a polyfunctional, ampholyte molecule, containing two basic and two acidic moieties. various data have been published on its acid-base chemistry. the dissociation constants (pka values) were reported by bennion et al. [8] (1.68, 3.29, 7.01, 11.12) and ip et al. [9] (2.5, 4.0, 6.7, 10.1), without an indication of the experimental error and with no assignment of basicities to moieties. gonzalez et al. [10] assigned the measured pka values (1.4 ± 0.1, 3.00 ± 0.01, 7.10 ± 0.01, 10.78 ± 0.01) to the functional groups, but only on the basis of chemical evidence. these three datasets of pka values are significantly different, particularly the lowest and highest constants, which remain uncertain. the difference between them exceeds one order of magnitude. the acid/base character determines the charge of a molecule in solution at a particular ph (characterised by the dissociation/protonation constant, pka/log k). further on in this paper, we use the log k term and consider the ionisation process as an association with a proton in all acid-base equilibria. this information is important in the estimation of adme (absorption, distribution, metabolism, excretion) parameters and the interpretation of pharmacokinetic (pk) properties. log k values can be used to better understand the binding mechanism in therapeutic events and also for optimisation of chemical reactions and analytical methods. several analytical methods have been applied for the determination of lisinopril in biological samples and pharmaceutical preparations such as alkalimetry [11], spectrophotometry [12], high performance liquid chromatography (hplc) [13], high performance thin-layer chromatography (hptlc) [14], and capillary electrophoresis (ce) [15]. obviously, upon the application of these methods, exact knowledge of acid-base chemistry is essential (for pharmacopoeial methods) or at least favourable (for chromatographic techniques and ce). the acidity/basicity of monovalent compounds can be quantified in terms of macroscopic log k parameters (macroconstants). for multiprotic compounds, macroconstants characterise the basicity of the molecule as a whole. they refer to the stoichiometric composition of the species, but they fail to provide information on specific proton-binding sites. site-specific, submolecular basicities can be obtained when krisztina takács novák et al. admet & dmpk 1(2) (2013) 6-16 8 microconstants are determined. microconstants measure the proton affinity of individual groups, while the protonation states of all other sites are definite in the molecule [16]. the macroscopic and microscopic basicities provide authentic information on propensities of intermolecular interactions both in pharmacokinetic (pk) and pharmacodynamic (pd) stages [17]. as a part of our studies on the physico-chemical properties of ace inhibitors, we investigated the acidbase chemistry of lisinopril. we characterised both the overall and the group-specific basicities of lisinopril. validated potentiometric methods and 1 h nmr spectroscopy resulted in more precise macroconstants than in previous investigations. here, we first studied the site-specific (microscopic) protonation equilibria of lisinopril considering the overlapping protonation of carboxylates. the results were used to interpret the pk/pd properties of the compound. experimental materials and instrumentation lisinopril dihydrate was generously supplied by gedeon richter ltd. (budapest, hungary) and used without further purification. distilled water was of pharmacopoeial grade [11] and all other reagents of analytical grade were purchased from commercial suppliers. potentiometric titrations were carried out on a glpka automated pka and log p analyser (sirius analytical instr. ltd. forest row, uk). nmr measurements were carried out on a 600 mhz varian inova spectrometer (palo alto, ca), equipped with a broadband inverse detection pulse field gradient probehead. potentiometric logk determination electrode calibration the four-parameter procedure was used for electrode standardisation in both aqueous and semiaqueous media [18]. hcl solutions of known concentration, containing 0–45.68 wt% methanol, were titrated with standardised koh at 25.0 ± 0.1 o c, at 0.15 m ionic strength using kcl, under an n2 atmosphere, in the ph interval of 1.8-12.2. the operational ph reading was related to pch values by the standard multiparametric equation: ph = a + spch + jh[h + ] + johkw/[h + ] (1) where a corresponds to the negative logarithm of the activity coefficient of [h + ] at working temperature and ionic strength and s is the ratio between the electrode slope and the nernst slope. the jh and joh terms correct the electrode junction effects at low and high ph, respectively. the parameters were determined by a weighted non-linear least squares procedure (refinement pro tm 2.2 software sirius analytical instr. ltd. forest row, uk). titration in aqueous medium six millilitres of 1.97–2.10 mm aqueous solutions of the sample were preacidified to ph 1.8 with 0.5 m hcl and then titrated alkalimetrically to ph 12. the titrations were carried out at 25.0 ± 0.1 c, at 0.15 m ionic strength using kcl under an n2 atmosphere. three separate titrations were performed. the initial admet & dmpk 1(2) (2013) 6-16 lisinopril speciation doi: 10.5599/admet.1.2.3 9 estimates of log k values were obtained from bjerrum difference plots ( n vs. ph, when n is the average number of bound protons) and were refined by a weighted non-linear least squares procedure using refinement pro tm 2.2 software (sirius analytical instr. ltd. forest row, uk). titration in solvent mixtures three semi-aqueous solutions of lisinopril containing 14.21, 28.54 and 43.82 wt% methanol were titrated under the same experimental conditions as in aqueous medium. the apparent protonation constants (logs k) were calculated from the difference (bjerrum) plot in a similar manner as the aqueous log k values. the yasuda-shedlovsky procedure was applied to estimate the aqueous log k values. the yasuda-shedlovsky extrapolation method is based on the linear relation between logs k and the dielectric constant (ε) of a solvent mixture: logs k + log [h2o] = a / ε + b (2) where a is the slope, b is the intercept (fitting constant) of the linear equation, ε is the dielectric constant of the methanol-water mixture and log [h2o] is the molar water concentration of the given solvent mixture. this method is a widely used procedure in co-solvent techniques [19,20]. 1 h nmr titration with in situ ph monitoring a single nmr sample solution of 0.6 ml was prepared containing 8 mm lisinopril, 140 mm nacl (to ensure a constant ionic strength of 0.15 m) in a 9/1 h2o/d2o solvent mixture. this sample also contained 2 mm of the following ph indicator molecules: dichloroacetic acid, chloroacetic acid, acetic acid, imidazole, tris(hydroxymethyl)-aminomethane (tris) and trimethylamine hydrochloride (tma) in order to determine the actual ph of the sample in each titration step, according to the electrodeless single tube nmr titration method [21]. as an internal chemical shift reference, 0.5 mm of sodium 3-(trimethylsilyl)-1propanesulfonate (dss) was applied. the 1 h nmr spectra were recorded at 25.0  0.1 c. the water signal was suppressed either by the double pulse field gradient spin echo (dpfgse) or the selective presaturation (presat) sequence. the in situ ph value was deduced from the chemical shift of the appropriate indicator by the following equation: obs indind hind obs ind ind loglogph δδ δδ k    (3) where log kind is the protonation constant of the indicator and ind, hind are the limiting chemical shifts of its non-protonated and protonated forms (determined in separate experiments) [22,23]. the nmr-ph datasets were fitted with the opium computer program [24] to determine protonation macroconstants and chemical shift values specific to each macrospecies (dhilis). results and discussion among the ace inhibitor drugs, lisinopril has the most interesting and complex acid-base chemistry. there are four proton-binding sites in the molecule: two carboxylates as well as a primary and a secondary krisztina takács novák et al. admet & dmpk 1(2) (2013) 6-16 10 amine group. the protonation processes of the two carboxylate groups are expected to be highly overlapping. for the investigation of the ionisation/protonation properties of the molecule, potentiometry in aqueous medium and in methanol-water mixtures as well as nmr-ph titration were applied. protonation macroconstants the stepwise protonation equilibria (fig. 2) of the whole molecule are characterised in terms of protonation macroconstants, log k values. figure 2. stepwise protonation equilibria of lisinopril generally, potentiometry in aqueous medium is the method of choice for logk determination for molecules with solubility greater than 0.8 mm in the entire ph interval of interest. the solubility of lisinopril (0.22 m, [9]) allows the determination of macroconstants by the standard method. the four obtained log k values along with the standard deviations calculated from three parallel titrations (3 x 15 points) are listed in table 1. since the log k4 value falls below to the lower applicability limit (log k < 2) of ph-metric titration, we also measured the protonation macroconstants using the co-solvent method. the effect of methanol on the protonation constant is known to depend on the charge of the basic site. for acids, the apparent logs k value increases with the increasing weight percent of methanol, while a decrease is usually observed for bases [25]. using this co-solvent method, we obtained a more reliable value for logk4 since the logs k4 values in methanol-water mixtures shifted up to the measurable ph range. at the same time, this cosolvent method allows for assigning logk values to the acidic and basic moieties of the molecule. according to the slopes of the regression lines on the yasuda-shedlovsky plot (fig. 3), log k1 and log k2 characterise amine functions, while log k3 and log k4 the carboxylate groups of the molecule, respectively. for the exact proton speciation of lisinopril, 1 h nmr-ph titrations were carried out using the most similar possible experimental conditions as in potentiometry. fig. 4 shows the aliphatic part of one spectrum from the titration series. admet & dmpk 1(2) (2013) 6-16 lisinopril speciation doi: 10.5599/admet.1.2.3 11 figure 3. yasuda-shedlovsky plot where open diamonds denote logs k1 (slope: -138.4, intercept: 14.271), open squares denote logs k2 (slope: -44.6, intercept: 9.532), filled diamonds denote logs k3 (slope: 82.6, intercept: 3.891) and filled squares denote logs k4 (slope: 29.9, intercept: 2.976). figure 4. representative 1 h nmr spectrum of lisinopril at ph 5.07 in 9/1 h2o/d2o. the assignment is based on conventional 2d nmr techniques. at certain ph values, the assignment of overlapping multiplets to methylene and methine protons was verified by tocsy and hsqc experiments. in fig. 5, the chemical shifts of the methine protons h(2), h(2’) and h(2”) along with the methylenes h(6’), h(3”) and h(4”) are depicted as functions of ph. these six datasets were fitted simultaneously to the tetraprotic macroscopic model function described elsewhere [26]. the obtained macroconstants are shown in table 1. krisztina takács novák et al. admet & dmpk 1(2) (2013) 6-16 12 table 1. protonation macroconstants measured by different methods. method protonation macroconstants log k1 ± sd log k2 ± sd log k3 ± sd log k4 ± sd potentiometry in aqueous medium 10.74 ± 0.01 7.14 ± 0.01 3.10 ± 0.01 1.54 ± 0.05 potentiometry in solvent mixtures 10.75 ± 0.01 7.22 ± 0.03 3.21 ± 0.02 1.62 ± 0.01 nmr-ph titration 10.53 ± 0.03 7.12 ± 0.01 3.15 ± 0.01 1.63 ± 0.01 average of the best two values* 10.75 ± 0.01 7.13 ± 0.01 3.13 ± 0.01 1.63 ± 0.01 * see details in text figure 5. nmr-ph titration curves with computer fitted solid line the macroconstants (table 1) obtained by different methods show generally good agreement. the average deviation between the log k values was 0.09 units. the ca. 0.2 unit discrepancy in log k1 between data acquired by the nmr and potentiometric methods may arise from the greater ambiguity of in situ ph values. specifically, trimethylamine has a log kind = 9.90 [23] and thus can monitor ph values higher than admet & dmpk 1(2) (2013) 6-16 lisinopril speciation doi: 10.5599/admet.1.2.3 13 10.9 with a lower precision (see the error estimations in [21] for details). in contrast, the log k4 = 1.63 value by nmr can be considered as more reliable than data obtained by potentiometry in aqueous medium, because below ph 2, the acidity error of the electrode is considerable, while here the nmr-ph titration as described above is an electrodeless technique. this constant could be determined with excellent precision from the extrapolation of semi-aqueous logs k values due to the reasons mentioned above. however, we think it reasonable to make an average of the best two experimentally measured logk values based on the following principle. we omit the log k value with the highest uncertainty (the highest sd value). the calculated, most reliable protonation macroconstants are indicated in the last row of table 1. comparing the obtained logk values with those reported in [9], the difference was greater than 0.4 log units in all log k values. both the log k1 determined by benion et al. [8] and the log k4 reported by gonzalez et al. [10] differ remarkably from our data. this is not surprising because those constants were deduced from conventional ph-potentiometry in aqueous medium, while the methods used in our study (potentiometry in methanol-water mixtures, 1 h nmr-titration) give more reliable protonation constants at ph extremes. protonation microconstants nevertheless, macroscopic constants could not directly be assigned to functional groups, since the difference in their logarithms did not exceed 3. the microscopic protonation scheme in fig. 6 should be considered for lisinopril instead, where the superscript on microconstant k indicates the group protonating in the equilibrium in question, whereas the subscript (if any) refers to already protonated group(s). figure 6. site-specific (microscopic) protonation scheme of lisinopril in the first two well-separated protonation steps, the macroand microconstants are identical. since the first protonation is reported most sensitively by h(6’), this protonation step can be assigned to the primary amino group (log k1 = log k a ). similarly, log k2 (= log k b a) accounts for the secondary amine basicity. in the case of lisinopril, the carboxylates denoted by c and d exhibit overlapping protonation, so selective monitoring of at least one of them is a prerequisite for microconstant determination. according to the normalised nmr titration curves, h(3”) and h(4”) can be assumed as selective sensors of d carboxylate protonation. thus, the experimental chemical shift profile of h(3”) was fitted to the following model equation: krisztina takács novák et al. admet & dmpk 1(2) (2013) 6-16 14 2 2 obs ][][1 ][][ )(       hkkhk hkkhk δδδδ 433 43 d ab lishlishlish 2 2 42 (4) the microscopic evaluation resulted in log d abk = 2.15 ± 0.05 (n= 2) which, in turn could be converted to the remaining three microconstants, using the following hessian constraints: 3 c ab d ab kkk  (5) 43 d abc c ab c abd d ab kkkkkk  (6) we obtained log c abdk = 2.63, log c ab k = 3.10 and log d abck = 1.68 (with estimated uncertainties of 0.05). the main pathway of protonation includes the abc microspecies, while its minor protonation isomer (abd) has nine times lower abundance at all ph values. in other words, the carboxylate on the proline ring exhibits nine times greater intrinsic basicity than the carboxylate on the side chain with an adjacent, electron-attracting protonated secondary amine (-nh2 + -) group. the mutual basicity-modifying effect of these moieties is quantified by the interactivity parameter, defined as the difference in the corresponding microconstants: pecd = 2.15 – 1.68 = 3.10 – 2.63 = 0.47. this pe value suggests an interaction exceeding the random, coulombic value of 0.2-0.3 [27]; presumably, these sites communicate through space with a simultaneous change in their geometric positions. a conformational change at ph < 4.5 is also substantiated by the changing multiplet patterns in the nmr spectra, but a detailed analysis of vicinal 3 jhh values holding conformational information is beyond the scope of this article. distribution of protonation species using log k and log k values, the percentage of various protonated species can be calculated for any arbitrary ph value. the distribution ratios of each species as a function of ph are shown in fig. 7. figure 7. distribution curves of the macrospecies (hilis) and microspecies (abc and abd) of lisinopril admet & dmpk 1(2) (2013) 6-16 lisinopril speciation doi: 10.5599/admet.1.2.3 15 table 2 summarises the relative concentration of lisinopril protonation species at the most relevant ph values in the body. in the stomach, the monocation (h3lis + ) and the dication (h4lis 2+ ) forms are present at about equal levels. in the gastrointestinal tract, the double protonated h2lis form predominates, while at the ph of plasma, the monoanion (hlis ) is the dominant species. table 2. the percentage concentration of the species in the stomach (ph 1.5), gastrointestinal tract (jejunum ph 6.5) and plasma (ph 7.4). form of the species percentage concentration stomach ph 1.5 gi tract (jejunum) ph 6.5 plasma ph 7.4 lis 2 0.00 0.00 0.03 hlis 0.00 18.98 65.04 h2lis 0.99 80.98 34.93 h3lis + 42.15 0.03 0.00 h4lis 2+ 56.86 0.00 0.00 conclusions in this study, the acid-base chemistry of tetraprotic lisinopril was characterised by protonation macro and microconstants. the highly precise log k values obtained by two independent analytical methods provided a better interpretation of the pk/pd properties. as fig. 7 indicates, under the ph conditions of the gi tract – the most likely site of oral absorption – lisinopril is predominantly present in the h2lis form. although the net charge of this species is zero, the molecule exists in solution as a double zwitterion: the two amine functions are protonated while the two carboxylates are not. this structure represents a highly polar dipole that is unfavourable for passive transport through lipoid membranes, which explains the low bioavailability of the molecule. recently, the intestinal dipeptide transporter system (dts) has been reported to be involved in the active transport mechanism of oral absorption of ace inhibitors [28]. at the ph of the plasma, the monoanion (hlis ) is the dominant form (more than 60%), which favours receptor binding. as is known, therapeutically useful ace inhibitors exhibit three-point binding to the zn ion-containing carbopeptidase ace enzyme [2]. the proline c2-carboxylate of lisinopril binds to the positively charged arg 146 , while the other carboxylate in the side chain participates in complex formation with the zn ion. the third binding site is represented by ser-oh where the c1’=o group forms an h-bond. the species distribution revealed in this study explains the good receptor binding of lisinopril. references [1] j. wong, r.a. patel, p.r. kowey, progress in cardiovascular diseases 47 (2004) 116-130. [2] m. harrold, in: d.a. williams, t.l. lemke (eds), foye’s principal of medicinal chemistry, 5th edn. lippincott williams & wilkins, philadelphia, 2002, p. 533. [3] b. beermann, american journal of medicine 85 (3b) (1988) 25-30. [4] d.j. eveson, t.g. robinson, j.f. potter, american journal of hypertension 20 (2007) 270-277. [5] n.k. hollenberg, d.a. anzalone, b. falkner, n.d.l. fisher, p.n. hopkins, w. hsueh, h. hutchinson, r.m. krauss, d.a. price, p. raskin, g.m. reaven, american journal of hypertension 14 (2001) 218223. krisztina takács novák et al. admet & dmpk 1(2) (2013) 6-16 16 [6] c.m. white, pharmacotherapy 18 (1998) 588-599. [7] k.l. goa, j.a. balfour, g. zuanetti, drugs 52 (1996) 564-588. [8] c. bennion, r.c. brown, a.r. cook, c.n. manners, d.w. payling, d.h. robinson, journal of medicinal chemistry 34 (1991) 439-447. [9] d.p. ip, j.d. demarco, m.a. brooks, analytical profiles of drug substances 21 (1992) 233-276. [10] e.b. gonzalez, e. farkas, a.a. soudi, t. tan, a.i. yanovsky, k.b. nolan, journal of chemical society dalton transactions (1997) 2377-2379. [11] european pharmacopoeia 7.0, strasbourg. [12] n. rahman, n. anwar, m. kashif, il farmaco 60 (2005) 605-611. [13] c.a. beasley, j. shaw, z. zhao, r.a. reed, journal of pharmaceutical and biomedical analysis 37 (2005) 559-567. [14] a. el-gindy, a. ashour, l. abdel-fattah, m.m. shabana, journal of pharmaceutical and biomedical analysis 25 (2001) 923-931. [15] s. hillaert, w. van den bossche, journal of pharmaceutical and biomedical analysis 25 (2001) 775783. [16] b. noszál, in: k. burger (ed.), biocoordination chemistry, coordination equilibria in biologically active systems. ellis horwood, chichester, 1990, p. 18. [17] k. takács-novák, b. noszál, m. tőkés-kövesdi, gy. szász, journal of pharmacy and pharmacology 47 (1995) 431-435. [18] a. avdeef, j.j. bucher, analytical chemistry 50 (1978) 2137-2142. [19] k. takács-novák, k.j. box, a. avdeef, international journal of pharmaceutics 151 (1997) 235-248. [20] k. deák, k. takács-novák, k. tihanyi, b. noszál, medicinal chemistry 2 (2006) 385-389. [21] z. szakács, g. hägele, r. tyka, analytica chimica acta 522 (2004) 247-258. [22] z. szakács, sz. béni, z. varga, l. örfi, gy. kéri, b. noszál, journal of medicinal chemistry 48 (2005) 249-255. [23] m. boros, j. kökösi, j. vámos, b. noszál, journal of pharmaceutical and biomedical analysis 43 (2007) 1306-1314. [24] m. kyvala, i. lukes, (1995) opium computer program, http://www.natur.cuni.cz/~kyvala/opium.html [25] a. avdeef, j.e.a. comer, s.j. thomson, analytical chemistry 65 (1993) 42-49. [26] z. szakács, m. kraszni, b. noszál, analytical and bioanalytical chemistry 378 (2004) 1428-1448. [27] b. noszál, z. szakács, journal of physical chemistry b 107 (2003) 5074-5080. [28] v.a. moore, w.j. irwin, p. timmins, p.a. lambert, s. chong, s.a. dando, r.a. morrison, international journal of pharmaceutics 210 (2000) 29-44. ©2013 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ adme-tox profiling of some low molecular weight water soluble chitosan derivatives doi: 10.5599/admet.5.3.423 192 admet & dmpk 5(3) (2017) 192-200; doi: http://dx.doi.org/10.5599/admet.5.3.423 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper adme-tox profiling of some low molecular weight water soluble chitosan derivatives adriana isvoran* 1 , alecu aurel ciorsac 2 and vasile ostafe 1 1 department of biology-chemistry and advanced environmental research laboratories, west university of timișoara, timișoara, romania 2 department of physical education and sport, university politehnica timișoara, timișoara, romania; *corresponding author: e-mail: adriana.isvoran@e-uvt.ro; tel.: +40-256-592-634; fax: +40-256-592-620 received: august 23, 2017; revised: september 26, 2017; published: september 29, 2017 abstract within this study we use a few computational tools for predicting absorption, distribution, metabolism, excretion and toxicity (adme-tox), pharmacokinetics profiles, toxic/adverse effects, carcinogenicity, cardiotoxicity and endocrine disruption of some of low molecular weight water soluble derivatives of chitosan that are used in wound healing. investigated compounds do not possess drug-like properties, their pharmacokinetics profiles reveal poor gastrointestinal absorption and low skin penetration. chitosan derivatives cannot pass the blood-brain barrier and they are not able to inhibit the enzymes of the cytochrome p450 that are involved in the metabolism of xenobiotics. they do not reflect carcinogenicity and cardiotoxicity and reveal only a low probability to be endocrine disruptors. the main side effects in humans of the investigated compounds are: weight loss, acidosis, gastrointestinal toxicity, respiratory failure. this information is especially important for professional exposure and accidental contamination with these compounds. keywords pharmacokinetics; toxic/adverse effects; carcinogenicity; cardiotoxicity; endocrine disruption. introduction chitosan is a cationic polysaccharide obtained from chitin and composed by glucosamine and n-acetyl glucosamine units, via β-(1,4) linkages. it differs from chitin by having free amino groups. chitosan is considered to have favourable biological properties revealing nontoxicity, mucoadhesivity, biocompatibility and biodegradability, but it is only soluble in acidic environments [1-3]. water soluble derivatives of chitosan (wsdc) are considered to have the similar favourable properties with chitosan, but they can be dissolved in aqueous media. wsdc are interesting biomaterials with numerous applications in bioengineering and biopharmaceutical fields [4], their preparation being one of the most commonly used chemical modifications of chitosan for biomedical purposes. as pharmaceutical applications, literature data reveal that chitosan and wsdc are used as carriers for radioactive elements, carriers for drug delivery and release systems and they have antimicrobial properties [2]. some wsdc are used to accelerate wound healing [5]. despite literature data revealing their use in biomedical applications, chitosan and its water soluble derivatives are not approved by the us food and drug administration (fda) for drug delivery, they http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:name@address.domain admet & dmpk 5(3) (2017) 192-200 admet of chitosan derivatives doi: 10.5599/admet.5.3.423 193 are only approved to be used for wound healing [6] and as dietary additives in japan, italy and finland [7]. there are limited data in specific literature concerning the effects of chitosan and chitosan derivatives on humans, most in vivo data being obtained from studies performed on animals, usually rodents [3]. the aim of this study is to predict the adme-tox profiles, pharmacokinetics and toxic/adverse effects of some low molecular weight wsdc that can be used for wound healing and to compare them with those of deacetylated chitosan and partially deacetylated chitosan. methodology in vitro studies revealed that chitosan and some of its water soluble derivatives have antifungal activity and may be used for wound healing purposes [5, 8-11]. according to these findings, we considered in this study 14 low molecular weight wsdc known to be useful for wound healing. in addition, for comparison purposes, we consider deacetylated chitosan and partially deacetylated chitosan, all investigated compounds being presented in table 1. their simplified molecular-input line-entry system (smiles) structures were obtained using acd/chemsketch [12] and are used for further predictions. when the structural data files (sdf) of compounds are needed for predictions, the online translation of smiles to sdf files is used [13]. table 1. molecules of wscd considered in this study compound common name smiles deacetylated chitosan (2 units) o[c@h]2c(o[c@@h]1oc(co)[c@@h](o)[c@h](o)c1n)[c@h](co)oc( o)c2n partially deacetylated chitosan (2 units) cc(=o)nc2c(o)[c@h](o[c@@h]1oc(co)[c@@h](o)[c@h](o)c1n)c(c o)o[c@h]2o n-carboxymethyl chitosan o[c@@h]1oc(co)[c@@h](o)[c@h](o)c1ncc(=o)o o-carboxymethyl chitosan oc1[c@h](o)c(cocc(=o)o)o[c@@h](o)c1n n,n-carboxymethyl chitosan o[c@@h]1oc(co)[c@@h](o)[c@h](o)c1n(cc(=o)o)cc(=o)o n,o-carboxymethyl chitosan o[c@@h]1oc(cocc(=o)o)[c@@h](o)c(o)c1ncc(=o)o n,n,o-carboxymethyl chitosan o[c@@h]1oc(cocc(=o)o)[c@@h](o)[c@h](o)c1n(cc(=o)o)cc(=o)o n,n,n -trimethyl-chitosan c[n+](c)(c)c2[c@@h](o)[c@h](o)c(co)o[c@h]2o[c@@h]1c(co)o[c @@h](o)c(n)c1o chitosan-lactate o[c@@h]1oc(co)[c@@h](o)[c@h](o)c1nc(=o)c(c)o hydroxyethyl chitosan cc(=o)nc2c(o)[c@h](o[c@@h]1oc(co)[c@@h](o)[c@h](o)c1ncco) c(cocco)o[c@h]2o hydroxypropyl chitosan cc(=o)nc2c(o)[c@h](o[c@@h]1oc(co)[c@@h](o)[c@h](o)c1ncc(c )o)c(cocc(c)o)o[c@h]2o dihydroxypropyl chitosan cc(o)cn(cc(c)o)c2[c@@h](o)[c@h](o)c(co)o[c@h]2o[c@@h]1c(c o)o[c@@h](o)c(nc(c)=o)c1o n-(2-hydroxy)propyl-3trimethylammonium chitosan c[n+](c)(c)cc(o)cnc2[c@@h](o)[c@h](o)c(co)o[c@h]2o[c@@h]1c( co)o[c@@h](o)c(nc(c)=o)c1o o-(2-hydroxy)propyl-3trimethylammonium chitosan c[n+](c)(c)cc(o)cocc2o[c@@h](o)c(nc(c)=o)c(o)[c@@h]2o[c@@h ]1oc(co)[c@@h](o)[c@h](o)c1n n-(4-carboxybutyroyl) chitosan cc(=o)nc2c(o)[c@h](o[c@@h]1oc(co)[c@@h](o)[c@h](o)c1nc(=o )ccc)c(co)o[c@h]2o n-trimethyl chitosan cc(=o)nc2c(o)[c@h](o[c@@h]1oc(co)[c@@h](o)[c@h](o)c1[n+](c )(c)c)c(co)o[c@h]2o for the compounds considered in this study we use the following computational tools: (i) swissadme [14] and fafdrugs [15, 16] to assess their pharmacokinetics and to predict their adme-tox profiles, (ii) pass online [17] for envisaging their toxic/side effects, (iii) carcinopred-el [18] to predict carcinogenicity, pred-herg4.0 [19] to assess cardiotoxicity and endocrine disruptome [20] to evaluate their endocrine disruption potential. these tools have been selected among the numerous computational utilities destined isvoran, ciorsac, ostafe admet & dmpk 5(3) (2017) 192-200 194 to predict the adme-tox profile and pharmacokinetics because of their continuous update, accessibility by free-login on the websites and of their robustness, accompanied by easy inputs and interpretations. swissadme is a free available on-line computational tool allowing online computation of the physicochemical properties of one or multiple molecules and prediction of pharmacokinetics, drug-likeness and medicinal chemistry friendliness [14]. it computes for every molecule: the chemical structure, physicochemical properties (such as molecular weight, topological polar surface area considering phosphorous and sulphur as polar atoms, molar refractivity, etc), lipophilicity, water solubility, bioavailability radar, pharmacokinetics, the skin permeability coefficient, drug-likeness and medicinal chemistry. lipophilicity is expressed by the partition coefficient between n-octanol and water (log po/w) which is computed under swissadme tool by implementing five different methods based on distinct approaches and displaying both the values obtained using these methods and a consensus log po/w as the arithmetic mean of the predicted values. concerning the skin permeability coefficient, it is expressed in log units and a more negative log kp value corresponds to a less skin permeant molecule. as an example, the skin permeability coefficient of diclofenac (that is known to have a good skin permeability) is predicted as log kp = -4.96 cm/s and those of ouabain (that is not able to penetrate the skin) is predicted as log kp = -10.94 cm/s [21]. when displaying the bioavailability radar, six physicochemical properties are taken into account: size, polarity, lipophilicity, solubility, flexibility and saturation. concerning the pharmacokinetics, drug-likeness and medicinal chemistry predictions, the swissadme classifiers reveal a similar accuracy of prediction as the other tools, demonstrating strong correlation between predicted and experimental values for molecules belonging to the training sets [14]. fadrugs is another free available online computational toll allowing analysing or filtering of molecules with eight predefined physicochemical filters based on several simple adme-tox rules [15, 16]. this tool allows the following procedures: (i) a data curation stage by desalting molecules and removing large molecules and compounds containing some types of inorganic atoms or duplicates, (ii) the computation of several physicochemical properties of molecules and their filtering considering several adme-tox rules, (iii) the detection of potential toxic groups, aggregators and pains [15, 16]. fafdrugs tool has been validated by using a list of 40 documented toxicophores and a list of 778 annotated oral drugs used in 19 therapeutic areas. an alert was detected in 36 among the 40 toxicophores with 100 % efficacy with regard to the nature of the alerts and among the 778 drugs, 81 % satisfied the rules [15, 16]. prediction of activity spectra of substances (pass) is a computational tool allowing prediction of biological activity and/or toxic and side effects of a chemical compound starting from its structural formula, the mean accuracy of prediction being about 90 % [17]. pass algorithm estimates the probability that the investigated compound belongs to a particular class of active compounds (pa) or inactive compounds (pi). the value of pa reveals the similarity of the molecule under investigation with the structures of those molecules within the training set which are the most typical in a subset of “actives”. correspondingly, pi reflects the similarity of the investigated compound with the molecules within the training set that belongs to the subset of “inactive”. the two probabilities values (pa and pi) vary from 0 to 1 and are independent. only those activities with pa > pi are considered as promising for a particular compound and a good accuracy of prediction is obtained when pa > 0.7 [17]. carcinopred-el is a free prediction online server based on twelve different molecular fingerprints and ensemble machine learning methods allowing identification of the structural features related to carcinogenic effects of organic compounds [18]. it uses 1003 compounds with known carcinogenicity against rats (494 carcinogens and 509 non-carcinogens) that are found in the carcinogenic potency admet & dmpk 5(3) (2017) 192-200 admet of chitosan derivatives doi: 10.5599/admet.5.3.423 195 database [22] as the training dataset to develop models for predicting the carcinogenicity and an external validation dataset of 40 compounds (23 carcinogens and 17 non-carcinogens) from the isscan database [23] that do not duplicate the training dataset and the accuracy of prediction is about 70 %. pred-herg is a free accessible web tool that builds predictive qsar (quantitative structure activity relationship) models of herg k+ channel blockage. the predictions are built using the largest free available dataset of structurally diverse compound with 16932 associated bioactivity records for the herg liability that are contained in the chembl database [24] with an accuracy up to 89 % [19]. endocrine disruptome is a free web service able to assess the endocrine disruption potential of a chemical compound by using the molecular docking method based on autodock vina algorithm [25]. investigated compound is docked to 18 crystallographic structures available in the protein data bank [26] for 14 distinct human nuclear receptors: androgen receptor; estrogen receptors α and β; glucocorticoid receptor; liver x receptors α and β; mineralocorticoid receptor; peroxisome proliferator activated receptors α, β/δ, and γ; progesterone receptor; retinoid x receptor α; thyroid receptors α and β [20]. the service was validated on small drug-like molecules with molecular weight under 600. results furnished by endocrine disruptome are dived in 4 classes based on three thresholds of sensitivity (se) that are set per structure and are used to obtain the docking scores: one class with high probability of binding (se < 0.25), two classes with medium probability of binding (0.25 < se < 0.50 and 0.50 < se < 0.75) and one class with low probability of binding. (se > 0.75) [20]. results and discussion adme-tox profiles generated using the swissadme tool and the drug-like and respiratory filters of implemented under fafdrugs tool reflect that compounds considered in this study do not possess druglike characteristics, some of their physicochemical properties (partition coefficient – log p, number of hydrogen bonds donors hbd, total number of hydrogen bonds donors and acceptors – hbd_hba, topological polar surface area tpsa) being outside the permitted limits for drugs (figures 1 and 2), and such as violating many rules (table 2). figures 1a and 1b illustrate the use of fafdrugs tool to obtain the positioning of the physicochemical properties of n-carboxymethyl chitosan (cmc) (a) and n,n,n-trimethyl chitosan (tmc) (b) respectively by compassion to the drug-like properties. figures 2a and 2b show the positioning of the physicochemical properties of cmc (a) and tmc (b) respectively when the respiratory filter is applied under fafdrugs tool. figure 1. radar pictures of the positioning of the physicochemical properties of investigated compounds (dark blue line) by comparison to the properties corresponding to drug-like molecules (light blue region) (a) ncarboxymethyl chitosan, (b) n,n,n-trimethyl-chitosan. (a) (b) isvoran, ciorsac, ostafe admet & dmpk 5(3) (2017) 192-200 196 figure 2. radar pictures of the positioning of the physicochemical properties of investigated compounds (dark blue line) by comparison to the properties corresponding to molecules that can be used as respiratory drugs (light blue region) (a) n-carboxymethyl chitosan, (b) n,n,n-trimethyl-chitosan. table 2. druglikeness of the wscd considered in this study compound number of violations lilly medchem rules bioavailability score lipinski ghose (bayer) veber egan muege deacetylated chitosan (2 units) 2 1 1 1 4 pass 0.17 partially deacetylated chitosan (2 units) 2 3 1 1 4 pass 0.17 n-carboxymethyl chitosan 1 1 0 1 2 pass 0.55 o-carboxymethyl chitosan 0 1 1 1 1 pass 0.55 n,n-carboxymethyl chitosan 1 1 1 1 3 pass 0.11 n,o-carboxymethyl chitosan 1 1 1 1 3 pass 0.11 n,n,o-carboxymethyl chitosan 2 1 1 1 4 do not pass 0.11 n,n,n-trimethyl-chitosan 1 1 1 1 3 pass 0.55 chitosan-lactate 1 1 0 1 1 pass 0.55 hydroxyethyl chitosan 2 1 2 1 4 pass 0.17 hydroxypropyl chitosan 2 3 2 1 4 pass 0.17 dihydroxypropyl chitosan 2 3 2 1 4 pass 0.17 n-(2-hydroxy)propyl-3trimethylammonium chitosan 2 3 2 1 4 do not pass 0.17 o-(2-hydroxy)propyl-3trimethylammonium chitosan 2 3 2 1 4 do not pass 0.17 n-(4-carboxybutyroyl) chitosan 2 1 1 1 4 pass 0.17 n-trimethyl chitosan 2 1 1 1 3 do not pass 0.17 with the exception of n-carboxymethyl chitosan, o-carboxymethyl chitosan and n,n,n-trimethylchitosan, chitosan and the other wsdc reveal a low oral bioavailability. from the point of view of safety, fafdrugs tool exposes that all these compounds reflect no toxicity, but some of them are able to induce phospholipidosis. it is not an unexpected result, as it is generally recognized that phospholipidosis is a response to cells exposed to cationic amphiphilic substances [27]. pharmacokinetics of investigated compounds has been obtained using swissadme tool and is revealed in table 3. the results presented in table 3 indicate that all investigated compounds present a low gastrointestinal absorption, low skin permeation and they are not able to inhibit any of the cytochromes p450 (cyp) involved in metabolism of xenobiotics. these compounds are predicted to be substrates for (a) (b) admet & dmpk 5(3) (2017) 192-200 admet of chitosan derivatives doi: 10.5599/admet.5.3.423 197 p-gp, another pharmacokinetics-relevant protein, emphasizing their active efflux through biological membranes. table 3. pharmacokinetics of water soluble chitosan derivatives: gigastrointestinal absorption, bbb – blood brain barrier penetration, p-gp – substrate of the p-gp protein, cypcytochrome p450, log kp –skin permeation coefficient compound gi bbb p-gp cyp1a2 cyp2d6 cyp2c9 cyp2c19 cyp3a 4 log kp (cm/s) deacetylated chitosan low no yes no no no no no -12.10 partially deacetylated chitosan low no yes no no no no no -12.18 o-carboxymethyl chitosan low no yes no no no no no -11.69 n,n-carboxymethyl chitosan low no yes no no no no no -12.08 n,o-carboxymethyl chitosan low no yes no no no no no -11.96 n,n,o-carboxymethyl chitosan low no yes no no no no no -12.27 n-carboxymethyl chitosan low no yes no no no no no -11.77 n,n,n-trimethylchitosan low no yes no no no no no -11.66 chitosan-lactate low no yes no no no no no -12.27 hydroxyethyl chitosan low no yes no no no no no -12.94 hydroxypropyl chitosan low no yes no no no no no -12.50 dihydroxypropyl chitosan low no yes no no no no no -12.55 n-(2-hydroxy)propyl-3trimethylammonium chitosan low no yes no no no no no -12.92 o-(2-hydroxy)propyl-3trimethylammonium chitosan low no yes no no no no no -12.92 n-(4-carboxybutyroyl) chitosan low no yes no no no no no -11.94 n-trimethyl chitosan low no yes no no no no no -11.74 results concerning side/toxic effects of compounds considered in this study were obtained using pass online and are presented in table 4. when considering the predictions made by pass online tool, we only took into account those results with the probability of a compound to be active higher than 0.9. these results illustrate that wsdc reflect fewer and less important side/toxic effects than deacetylated and partially deacetylated chitosan. carcinopred-el tool reflects that all considered compounds belong to the non-carcinogen class of substances. pred-herg results also indicate that the investigated compounds are non-blockers for the herg k + channels and do not reflect cardiotoxicity. endocrine disruptome tool emphasizes that wscd considered in this study have a low inhibitory potential of endocrine system-specific proteins. isvoran, ciorsac, ostafe admet & dmpk 5(3) (2017) 192-200 198 table 4. toxic effects of compounds considered in this study compound toxic effects deacetylated chitosan 2 units weight loss, acidosis, nausea, gastrointestinal toxicity, hematotoxicy, respiratory failure, reproductive dysfunction, neurotoxicity partially deacetylated chitosan 2 units weight loss, acidosis, nausea, gastrointestinal toxicity, nephrotoxicity, hematotoxicity, thrombocytopenia, pain, respiratory failure, cyanosis, necrosis, reproductive dysfunction, neurotoxicity o-carboxymethyl chitosan (1 unit) pa<0.9 n,n-carboxymethyl chitosan (1 unit) pa<0.9 n,o-carboxymethyl chitosan (1 unit) pa<0.9 n,n,o-carboxymethyl chitosan pa<0.9 n-carboxymethyl chitosan weight loss, acidosis, respiratory failure, inflammation n,n,n-trimethyl-chitosan weight loss, acidosis, respiratory failure, gastrointestinal toxicity chitosan-lactate acidosis, weight loss, nausea, hematotoxicity, gastrointestinal toxicity, hyperglycemic, respiratory failure, cyanoses, leukopenia hydroxyethyl chitosan hematotoxicity, nephrotoxicity, respiratory failure, acidosis, thrombocytopenia, allergic dermatitis, gastrointestinal toxicity, nausea, asthma, necrosis, cyanosis, leukopenia, neurotoxicity hydroxypropyl chitosan hematotoxicity, acidosis, nausea, gastrointestinal toxicity, dermatitis, nephrotoxicity dihydroxypropyl chitosan acidosis, nausea, gastrointestinal toxicity, respiratory failure n-(2-hydroxy)propyl-3trimethylammonium chitosan n/a o-(2-hydroxy)propyl-3trimethylammonium chitosan n/a n-(4-carboxybutyroyl) chitosan weight loss, acidosis, hematotoxicity, nausea, gastrointestinal toxicity, neurotoxicity, respiratory failure n-trimethyl chitosan n/a conclusions as there are limited experimental in vivo studies concerning the pharmacological and toxicological effects of chitosan and its wsd on humans, computational predictions of their pharmacokinetics and adverse effects may provide useful information. chitosan and its derivatives considered in this study reflect low skin permeability and low inhibitory potential for the proteins specific to the endocrine system. they are not able to penetrate the blood brain barrier and to affect the central nervous system and they are not predicted to have an inhibitory potential for the enzymes of the cytochrome p450 involved in drugs and other xenobiotics metabolism. they are predicted to be substrates for the p-gp protein and it illustrate that their systemic exposure could be reduced. by comparison to chitosan and deacetylated chitosan, some wsdc show good oral bioavailability. furthermore, the investigated compounds do not reveal carcinogenic potential and cardiac toxicity, but they are predicted to induce phospholipidosis and to have some adverse effects: weight loss, acidosis, gastrointestinal toxicity, respiratory failure. some of the predicted pharmacological and toxicological profiles of the considered wsdc are in good agreement with a few experimental studies revealing that wsdc have no or low cytotoxicity [28, 29], no genotoxicity [28], have fatand cholesterol-binding capacities producing weight loss [30]. their side effects increase with the molecular weight and degree of deacetylation [31]. this agreement strengthens the outcomes of this study and illustrates the utility of the methods. admet & dmpk 5(3) (2017) 192-200 admet of chitosan derivatives doi: 10.5599/admet.5.3.423 199 to the best of our knowledge, this is a first study dealing with the adme-tox properties of chitosan and some of its water soluble derivatives that are used for wound healing. chitosan and its water-soluble derivatives are approved by the us food and drug administration only to be used for wound healing. as they reflect low skin penetration, people being treated with these compounds reveal only a low degree of exposure. the obtained information is important especially for people manufacturing and handling these compounds being occupationally exposed to these chemicals. acknowledgements: this work is supported by grant pniii-p3-284, chitowound biotechnological tools implementation for new wound healing applications of byproducts from the crustacean seafood processing industry. references [1] r. riva, h. ragelle, a. des rieux, n. duhem, c. jérôme, v. préat, advanced polymers science 244 (2011) 19–44; https://link.springer.com/chapter/10.1007%2f12_2011_137. [2] r.c. cheung, t.b. ng, j.h. wong, w.y. chan, marine drugs 13 (2015) 5156-5186; http://www.mdpi.com/1660-3397/13/8/5156. [3] m.d. gades, j.s. stern, obesity research 11 (2003) 683-688; http://onlinelibrary.wiley.com/doi/10.1038/oby.2003.97/full. [4] q.-x. wu, d.-q. lin, s.-j.yao, marine drugs 12 (2014) 6236–6253; http://www.mdpi.com/1660-3397/12/12/6236. [5] v. patrulea, v. ostafe, g. borchard, o. jordan, european journal of pharmaceutics and biopharmaceutics 97 (2015) 417–426; http://linkinghub.elsevier.com/retrieve/pii/s0144861716300091. [6] i.wedmore, j.g. mcmanus, a.e. pusateri, j.b. holcomb, journal of trauma 60 (2006) 655–658; https://www.ncbi.nlm.nih.gov/pubmed/16531872. [7] l. illum, pharmaceutical research 15 (1998) 1326–1331; https://www.ncbi.nlm.nih.gov/pubmed/9755881. [8] f. seyfarth, s. schliemann, p. elsner, u.c. hipler, international journal of pharmaceutics 353 (2008) 139–148; http://europepmc.org/abstract/med/18164151. [9] k. azuma, r. izumi, t. osaki, s. ifuku, m. morimoto, h. saimoto, s.o minami, y. okamoto, journal of functional biomaterials 6 (2015) 104-142; https://www.ncbi.nlm.nih.gov/pmc/articles/pmc4384104/. [10] m. sobol, a. bartkowiak, b. de haan, p. de vos, journal of biomedical materials research part a 101a (2013) 1907–1914; http://onlinelibrary.wiley.com/doi/10.1002/jbm.a.34500/abstract. [11] j. macossay, lsu historical dissertations and theses (1995) 6119; http://digitalcommons.lsu.edu/gradschool_disstheses/6119. [12] acd chemsketch 12.00 and chemicalize from chemaxon; https://chemicalize.com; accessed 01.08.2017. [13] online smiles translator and structure file generator; https://cactus.nci.nih.gov/translate/; accessed 06.08.2017 and 24.09.2017. [14] a. daina, o. michielin, v. zoete, scientific reports 7 (2017) 42717; https://www.nature.com/articles/srep42717?wt.feed_name=subjects_physical-sciences. https://link.springer.com/chapter/10.1007%2f12_2011_137 http://www.mdpi.com/1660-3397/13/8/5156 http://onlinelibrary.wiley.com/doi/10.1038/oby.2003.97/full http://www.mdpi.com/1660-3397/12/12/6236 http://linkinghub.elsevier.com/retrieve/pii/s0144861716300091 https://www.ncbi.nlm.nih.gov/pubmed/16531872 https://www.ncbi.nlm.nih.gov/pubmed/9755881 http://europepmc.org/abstract/med/18164151 https://www.ncbi.nlm.nih.gov/pmc/articles/pmc4384104/ http://onlinelibrary.wiley.com/doi/10.1002/jbm.a.34500/abstract http://digitalcommons.lsu.edu/gradschool_disstheses/6119 https://chemicalize.com/ https://cactus.nci.nih.gov/translate/ https://www.nature.com/articles/srep42717?wt.feed_name=subjects_physical-sciences isvoran, ciorsac, ostafe admet & dmpk 5(3) (2017) 192-200 200 [15] d. lagorce, o. sperandio, j.b. baell, m.a. miteva, b.o. villoutreix, nucleic acids research 43 (2015) w200-207; https://www.ncbi.nlm.nih.gov/pmc/articles/pmc4489254/. [16] d. lagorce, l. bouslama, j. becot, m.a. miteva, b.o. villoutreix, bioinformatics (2017) accepted manuscript; https://doi.org/10.1093/bioinformatics/btx491. [17] a.v. stepanchikova, a.a. lagunin, d.a. filimonov, v.v. poroikov, current medicinal chemistry 10 (2003) 225-233; https://pdfs.semanticscholar.org/33af/3c27932acc01595715a322d2795ac45ae2cd.pdf. [18] l. zhang, h. ai, w. chen, z.yin, h. hu, j. zhu, j. zhao, q. zhao, h. liu, scientific reports 7 (2017) 2118; https://www.nature.com/articles/s41598-017-02365-0. [19] r.c. braga; v.m. alves, m.f.b. silva; e. muratov, d. fourches, l.m. liao, a. tropsha, c.h. andrade, molecular informatics 34 (2015) 698-701; http://onlinelibrary.wiley.com/doi/10.1002/minf.201500040/suppinfo. [20] k. olšek, j. mavri, m sollner dolenc, s. gobec, s. turk, journal of chemical information and modeling 54 (2014) 1254-1267; http://pubs.acs.org/doi/abs/10.1021/ci400649p. [21] r.o. potts, r.h. guy, pharmaceutical research 5 (1992) 663-669. [22] l. s. gold, t. h. slone, n. b. manley, g. b. garfinkel, e. s. hudes, l. rohrbach, b. n. ames, environmental health perspectives 96 (1991) 11-15; http://pubmedcentralcanada.ca/pmcc/articles/pmc1568255/pdf/envhper00416-0018.pdf. [23] r. benigni, c. bossa, a. m. richard, c. yang, annali dell'istituto superiore di sanita 44 (2008) 48–56; http://www.iss.it/publ/anna/2008/1/44148.pdf. [24] a.p. bento, a. gaulton, a. hersey, l.j. bellis, j. chambers, m. davies, f.a. krüger, y. light, l. mak, s. mcglinchey, m. nowotka, g. papadatos, r. santos, j.p. overington, nucleic acids research 42 (2014) 1083-1090; https://academic.oup.com/nar/article-lookup/doi/10.1093/nar/gkt1031. [25] o. trott, a. j. olson, journal of computational chemistry 31 (2010) 455-461; https://www.ncbi.nlm.nih.gov/pmc/articles/pmc3041641/. [26] h.m. berman, j. westbrook, z. feng, g. gilliland, t.n. bhat, h. weissig, i.n. shindyalov, p.e. bourne, nucleic acids research 28 (2000) 235-242; https://www.ncbi.nlm.nih.gov/pmc/articles/pmc102472/. [27] j. a. shayman, a. abe, biochimica and biophysica acta 1831 (2013) 602–611; https://www.ncbi.nlm.nih.gov/pmc/articles/pmc3528828/. [28] e.m. costa, s. silva, m.r. costa, m. pereira, d.a. campos, j. odila, a.r. madureira, a. cardellecobas, f.k. tavaria, , a.s. rodrigues, carbohydrate polymers 111 (2014) 385–392; http://www.sciencedirect.com/science/article/pii/s0144861714004032?via%3dihub. [29] c.y. chen, y.c. chung, journal of applied oral science 20 (2012) 620–627; https://www.ncbi.nlm.nih.gov/pmc/articles/pmc3881855/. [30] q. jin, h. yu, x. wang, k. li, p. li, peerj 5 (2017) e3279; https://dx.doi.org/10.7717/peerj.3279. [31] t. kean, m. thanou, advanced drug delivery reviews, 62 (2010) 3–11; http://www.sciencedirect.com/science/article/pii/s0169409x0900283x?via%3dihub. ©2017 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) https://www.ncbi.nlm.nih.gov/pmc/articles/pmc4489254/ https://doi.org/10.1093/bioinformatics/btx491 https://pdfs.semanticscholar.org/33af/3c27932acc01595715a322d2795ac45ae2cd.pdf https://www.nature.com/articles/s41598-017-02365-0 http://onlinelibrary.wiley.com/doi/10.1002/minf.201500040/suppinfo http://pubs.acs.org/doi/abs/10.1021/ci400649p http://pubmedcentralcanada.ca/pmcc/articles/pmc1568255/pdf/envhper00416-0018.pdf http://www.iss.it/publ/anna/2008/1/44148.pdf https://academic.oup.com/nar/article-lookup/doi/10.1093/nar/gkt1031 https://www.ncbi.nlm.nih.gov/pmc/articles/pmc3041641/ https://www.ncbi.nlm.nih.gov/pmc/articles/pmc102472/ http://www.sciencedirect.com/science/article/pii/s0144861714004032?via%3dihub https://www.ncbi.nlm.nih.gov/pmc/articles/pmc3881855/ https://dx.doi.org/10.7717/peerj.3279 http://www.sciencedirect.com/science/article/pii/s0169409x0900283x?via%3dihub http://creativecommons.org/licenses/by/3.0/ anti-inflammatory activities of flavonoid derivates doi: https://doi.org/10.5599/admet.1918 1 admet & dmpk 0(0) (2023) 000-000; doi: https://doi.org/10.5599/admet.1918 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index review anti-inflammatory activities of flavonoid derivates ysrafil ysrafil1,*, zulfiayu sapiun2, nangsih sulastri slamet2, fihrina mohamad2, hartati hartati2, sukmawati a damiti3, francisca diana alexandra1, sudarman rahman4, sri masyeni5,6, harapan harapan7,8,9, sukamto s. mamada10, talha bin emran11,12 and firzan nainu10 1department of pharmacotherapy, faculty of medicine, universitas palangka raya, palangka raya 73111, indonesia 2department of pharmacy, politeknik kesehatan kementerian kesehatan gorontalo, gorontalo 96135, indonesia 3department of midwivery, politeknik kesehatan kementerian kesehatan palangka raya 73111, palangka raya, indonesia 4faculty of mathematics and natural sciences, universitas palangka raya, palangka raya 73111, indonesia 5department of internal medicine, faculty of medicine and health sciences, universitas warmadewa, denpasar, bali 80235, indonesia 6department of internal medicine, sanjiwani hospital, denpasar, bali 80235, indonesia 7medical research unit, school of medicine, universitas syiah kuala, banda aceh 23111, indonesia 8tropical disease centre, school of medicine, universitas syiah kuala, banda aceh 23111, indonesia 9department of microbiology, school of medicine, universitas syiah kuala, banda aceh 23111, indonesia 10department of pharmacy, faculty of pharmacy, hasanuddin university, makassar 90245, indonesia 11department of pathology and laboratory medicine, warren alpert medical school & legorreta cancer center, brown university, providence, ri 02912, usa 12department of pharmacy, faculty of allied health sciences, daffodil international university, dhaka 1207, bangladesh *corresponding author: e-mail: ysrafil0155@gmail.com received: may 16, 2023; revised: june 29, 2023; published: july 26,2023 abstract background and purpose: flavonoids are a group of phytochemicals found abundantly in various plants. scientific evidence has revealed that flavonoids display potential biological activities, including their ability to alleviate inflammation. this activity is closely related to their action in blocking the inflammatory cascade and inhibiting the production of pro-inflammatory factors. however, as flavonoids typically have poor bioavailability and pharmacokinetic profile, it is quite challenging to establish these compounds as a drug. nevertheless, progressive advancements in drug delivery systems, particularly in nanotechnology, have shown promising approaches to overcome such challenges. review approach: this narrative review provides an overview of scientific knowledge about the mechanism of action of flavonoids in the mitigation of inflammatory reaction prior to delivering a comprehensive discussion about the opportunity of the nanotechnology-based delivery system in the preparation of the flavonoid-based drug. key results: various studies conducted in silico, in vitro, in vivo, and clinical trials have deciphered that the anti-inflammatory activities of flavonoids are closely linked to their ability to modulate various biochemical mediators, enzymes, and signalling pathways involved in the inflammatory processes. this compound could be encapsulated in nanotechnology platforms to increase the solubility, bioavailability, and pharmacological activity of flavonoids as well as reduce the toxic effects of these compounds. conclusion: in summary, we conclude that flavonoids and their derivates have given promising results in their development as new anti-inflammatory drug candidates, especially if they formulate in nanoparticles. ©2023 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords inflammation treatment; bioactive compounds; plants-base drug; nanotechnology https://doi.org/10.5599/admet.1918 https://doi.org/10.5599/admet.1918 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:ysrafil0155@gmail.com http://creativecommons.org/licenses/by/4.0/ y. ysrafil et al. admet & dmpk 00(0) (2023) 000-000 2 introduction inflammation is a complex and important protective response of the body to a stimulus in the form of stimulation by microorganisms, physical injuries, chemicals, allergic reactions or the presence of endogenous signals due to cell damage. in a physiological state, inflammation aims to eradicate inflammatory agents as well as repair the injured tissue and wounds [1-3]. in the aspect of pathomechanism, inflammation involves the body's immune system response, both natural and adaptive, although in most cases, the response is mainly initiated by the natural immune system. inflammation is initially stimulated by the introduction of highly conserved pathogenor damagedassociated molecular patterns (pamps and damps) by pattern recognition receptors (prrs) of the immune cells [4-6]. this will trigger the subsequent activation of the inflammatory cells such as neutrophils, eosinophils, mononuclear phagocytes, and macrophages, leading to the excessive production of nitric oxide (no), prostaglandin e2 (pge2), c-reactive protein (crp), chemokines and proinflammatory cytokines such as interleukins (il)-1β, il-6, and tumour necrosis factor (tnf)-α. the production of proinflammatory cytokines generally occurs due to the induction of nuclear factorkappa b (nf-κb), a transcription factor that plays a key role in the regulation of inflammation. in addition to the release of several molecules above, the inflammatory process is also accompanied by the release of a number of bioactive lipid mediators such as thromboxane, prostaglandins and prostacyclins on the initiative of the phospholipid conversion cascade by phospholipase and cyclooxygenase enzymes. the release of cytokines, chemokines, and mediators and these inflammatory molecules then triggers the appearance of inflammatory signs such as calor, dolor, rubor, tumour and functio laesa, and if it continues, it can cause a worse impact [3,7]. however, the protective function of inflammation can shift to the detrimental effects caused by the uncontrolled and excessive inflammatory response. in this condition, inflammation may disturb the homeostasis of the body's physiological processes and eventually develop into an inflammatory disorder [1-3]. several studies have demonstrated that inflammation is associated with developing and worsening various non-communicable disease-associated inflammation disorders, such as autoimmune diseases, neurodegenerative diseases, cancer, diabetes mellitus, and cardiovascular diseases [1,8]. these findings support the development of anti-inflammatory drugs to prevent the progression of the diseases. advances in the last few decades have revealed many facts about the benefits of various plant metabolites against inflammation. curcumin, resveratrol, and capsaicin have been previously reported to have the ability to decrease the production of the proinflammatory cytokines il-1β, il-6, and tnf-α, inhibit cyclooxygenase2 (cox-2) and the inflammatory pathways (e.g., nf-κb, mitogen-activated protein kinase (mapk), and activator protein 1 (ap-1) [9,10]. of several plant metabolites with potential anti-inflammatory action, flavonoids have attracted much interest. several studies have reported the activity of flavonoids in interfering with inflammatory signalling pathways and key enzymes, leading to the inhibition of the release of pro-inflammatory proteins and mediators [8,11]. however, the current problem is that these compounds have poor bioavailability and pharmacokinetic profiles. advances in nanotechnology have brought promising approaches in improving the bioavailability and increasing the effectiveness of these compounds as anti-inflammatory drug candidates [12,13]. in our review, we summarize the latest scientific findings regarding the benefits of flavonoids, pharmacokinetics, and their development based on nanomedicine, as this technology has become a solution in the delivery and improvement of drug effectiveness, including drugs used to impede inflammatory conditions. admet & dmpk 00(0) (2023) 000-000 anti-inflammatory activities of flavonoid derivates doi: https://doi.org/10.5599/admet.1918 3 mechanisms of inflammation as a fundamental biological process, inflammation is the protective response to tackle invading pathogens, antigens or damaged cells or tissues. it provides a defence mechanism to eliminate harmful substances and perform tissue repair [14,15]. at the molecular and cellular level, inflammation is characterized by five major signs: redness, heat, pain, swelling and functional loss. these changes can be triggered by two main causes, including infectious and non-infectious-related causes. the presence of these causes in the body will be recognized by receptors in the body that are subsequently followed by the activation of the inflammatory cascade, the release of the markers, and the recruitment of the inflammatory cells [16]. table 1. various causes that can trigger the activation of the inflammatory cascade. infectious causes non-infectious causes bacterial flagellin, meso-diaminopimelic acid, lipopeptides, muramyl dipeptide, lipopolysaccharides, peptidoglycans, microbial dna, toxoid. physical ionizing radiation, burnt, physical injury, the presence of foreign bodies, trauma, frostbite, blunt object. fungi zymosans chemical toxins, fatty acids, acidic environment, alcohol, and chemical irritants. viruses single or double-stranded rna biological damaged or death cells allergen pollen, chemical compounds psychological excitement the recognition of pamps and damps by prrs tissue injury, cell-infected pathogens, microbial invasion, or death cells associated with necroptosis or pyroptosis will release pamps or damps. pamps are unique structures in microbes or pathogens (e.g., lps, β-glucan, flagellin, spike, dna, rna) that can be recognized by immune cells and trigger their activation [17,18]. meanwhile, damps are endogenous molecules that are hidden from recognition by the immune system and will only be released when cell damage occurs as a signal to activate innate immune cells several damps released by damaged cells include high mobility group box 1 (hmgb1), histones, atp, heat shock proteins, fibrinogen, versican, uric acid, and mitochondrial components such as n-formylated peptides, cphosphate-g (cpg), dna repeats, and mitochondrial dna (mtdna) [19]. both pamps and damps will be recognized by prrs presented on innate immune cells such as neutrophils, monocytes, dendritic cells, macrophages, and other inflammatory cells [18,20]. neutrophils are the first leukocytes migrating to the injured or infected tissue to kill pathogens through the phagocytic mechanism and granule secretion. this process involves the release of reactive oxygen species (ros), hydrolytic enzymes, antimicrobial peptides, chemokines/cytokines, lipid mediators, as well as neutrophil extracellular traps (nets) to initiate inflammation and monocyte activation. upon this process, monocytes and macrophages will also migrate to the site of inflammation within a few hours to further promote inflammation as well as be involved in tissue repair [17,20]. innate immune cells have various prrs that recognize damps and pamps, trigger phagocytic processes, and mediate inflammation. there are 4 classes of prrs, including toll-like receptors (tlrs), nod-like receptors (nlrs), c-type lectin receptors (clrs), and rig-i-like receptors (rlrs) [21,22]. tlrs are a highly conserved subclass of prrs that are localized in transmembrane (tlr1, 2, 4, 5, 6, and 10) and in intracellular (tlr3, 7, 8, and 9). they can recognize various damps and pamps to trigger further activation of the myd88 signalling pathway or trif-dependent signalling pathway to induce the production of proinflammatory y. ysrafil et al. admet & dmpk 00(0) (2023) 000-000 4 cytokines and interferon-gamma. nlrs and rlrs have sensors that are located in the cytoplasm. however, they play different functions. rlrs are generally more specific for recognizing genome viruses (ssrna and dsrna) and trigger the production of ifns and inflammatory cytokines. meanwhile, nlrs tend to recognize pamps associated with bacteria (e.g., peptidoglycans, gamma-d-glutamyl-meso-diaminopimelic acid, and muramyl dipeptide) and some damps (e.g., uric acid, histones, biglycan, hyaluronan). recognition by these nlrs triggers the activation of the nf-kb signalling pathway [19,23]. furthermore, clrs are unique prrs that are located on the transmembrane and are characterized by the presence of a carbohydrate-binding domain in their receptors. they can recognize several pathogenic molecules (e.g., carbohydrates presented on bacteria, fungi, and viruses) as well as damaged cells (e.g. sap130, f-actin). this recognition will trigger the activation of the nf-kb signalling pathway either through tlr modulation or directly through spleen tyrosine kinase (syk) and raf1 pathways. the interaction between clrs and syk also triggers the activation of map kinase [19,23]. lipid-based signaling: arachidonic acid-derived eicosanoids recognition of damps and pamps by tlrs leads to the rapid phosphorylation of mapk. the presence of mapk will lead to the activation of cytosolic phospholipase a2 (cpla), an enzyme that can cleave membrane phospholipids to produce arachidonic acid (aa) as the primary eicosanoid precursor [24,25]. cytoplasmic aa can further be metabolized by cox, lipoxygenases (lox), or cytochrome enzymes (cyp) to produce oxygenated derivatives of eicosanoids including prostaglandins (pgs), thromboxanes (txs), leukotrienes (lts), lipoxins (lxs), hydroxyeicosatetraenoic acid (hetes), epoxyeicosatrienoic acids (eets), and eicosatetraenoic acids (etes). in addition to using arachidonic acid, eicosanoids can also be synthesized from other polyunsaturated fatty acids (pufas) such as di-homo-γ-linolenic acid (dgla), α-linolenic acid (ala), docosahexaenoic acid (dha), or eicosapentaenoic acid (epa). these eicosanoids can increase nerve pain signals and vasodilation in tissues that cause pain and swelling in the injured or invaded area [25-29]. cox is a highly conserved enzyme that plays an important role in the inflammatory process, making it a popular target for therapy like non-steroidal inflammatory drugs (nsaids). they exist in two isoforms, namely cox-1 and 2, and are responsible for the conversion of aa to various prostaglandins and thromboxanes [25,27]. furthermore, lox (such as 5-lox, 8-lox, 12-lox, and 15-lox) is an aa-metabolizing enzyme that works by inserting molecular oxygen on its substrate to produce hydroperoxy eicosatetraenoic acids (hpete). hpete is an intermediate product that can be further converted to various eicosanoids, including hetes, lxs, and hepoxilins. meanwhile, cyp is associated with the metabolism of aa into hetes and eet on the regulation of inflammation in certain organs such as the heart, blood vessels, and several other organs. however, despite this, drug development through this route is in lack of interest [29]. it is noteworthy that in addition to these afore mentioned enzymes, drug development is also carried out by targeting another enzyme called cytosolic phospholipase a2. release of proinflammatory chemokines and cytokines recognition and binding of pathogen and damaged cell molecules by prrs induce innate immune cells to release small signalling molecules, namely chemokines and cytokines, as a cascade of protective responses against invaded pathogens and repair of tissue injury. cytokines are small proteins secreted by immune cells to facilitate interaction and communication between cells. these proteins exist in several terms based on their origin, including lymphokine (secreted by lymphocytes) and monokine (secreted by monocytes), interleukins (secreted by certain lymphocytes but acting on other lymphocytes) as well as their activity, e.g. chemokines (cytokines with chemotactic activities) [30,31]. cytokines play an important role in inflammation both as inflammatory stimulants (proinflammatory cytokines) such as il-1β, il-6, and tnf-α and those inhibitions of inflammatory (anti-inflammatory cytokines), e.g., il-10, il-1ra, il-4, il-11, and il-13. admet & dmpk 00(0) (2023) 000-000 anti-inflammatory activities of flavonoid derivates doi: https://doi.org/10.5599/admet.1918 5 proinflammatory cytokines are highly secreted by activated and polarized macrophages for the upregulation of inflammatory reactions and mediate remote adaptive immunity during tissue or cell injury, invasion of the pathogen, and inflammation. some evidence reported that il-1β can also be secreted by monocytes, fibroblasts, and endothelial cells during the stimulation of prrs. expressed il-1β will be further activated by il1-converting enzyme (ice) or caspase-1 and followed by forming a complex with their receptor. this situation causes signal transduction leading to the activation of mapk and nf-κβ to produce proinflammatory cytokines. furthermore, similar to il-1β, tnf-α is also secreted by other cells, including monocytes, t cells, mast cells, natural killer (nk) cells, keratinocytes, fibroblasts, and neurons. their proinflammatory effect predominantly appears after their binding to tumour necrosis factor receptor 1 (tnfr1) and further facilitates activation of the c-jun n-terminal kinase (jnk), nf-κb, and mapk signalling pathways. meanwhile, il-6, which is also expressed by neutrophils, fibroblasts, both t and b cells, endothelial cells, keratinocytes, hepatocytes, and bone marrow cells, first binds to il-6r to mediate inflammation through activation of signal transduction via the gp130 proteins and lead activation of the jak/stat signalling pathway [32]. the other highlighted cytokine in inflammation is pro-inflammatory chemokine. chemokines, also known as chemotactic cytokines, are proteins of the cytokine family that play an important role in the chemoattraction and migration of leukocytes to various sites of tissue injury. inflammatory chemokines will be secreted when there is an inflammatory stimulus (both pathogens or damaged cells) to mediate further the innate and adaptive immune response [30,33]. chemokines induce the expression of integrins such as 2-integrin into leukocytes to facilitate diapedesis of these cells to the site of injury. cytokine signalling can be transduced by binding to the seventransmembrane g protein-coupled receptor (gpcr) found throughout the body [32,33]. involvement of mapk and nf-κβ signalling pathways in response to the introduction of prrs to pamps or damps activating the innate immune cells like macrophages, a series of intracellular signals are also activated to produce inflammatory mediators such as il-1β and tnfα. they will trigger a signalling cascade involving the adapter protein myd88 to lead to activation of the host mapk pathway, which is further followed by activation of transcription factor nf-κb. this, in turn, induces the production of proinflammatory cytokines [34-36]. under normal conditions, the presence of iκb proteins in the cytoplasm could inhibit nf-κb. the presence of a stimulus in prrs triggers signal transduction leading to the formation and activation of iκb kinase (ikk) that is composed of ikkα and ikkβ, as well as ikkγ as the regulatory subunit. iκb will be further phosphorylated to lead activation of nf-κβ. taken together, both signalling pathways play an important role in the inflammatory processes because they are the suppliers of proinflammatory cytokines and thus serve as targets for drug action [15,21,36]. flavonoid and its derivates flavonoids are a class of compounds with a low molecular weight based on the 2-phenyl-chromone nucleus (figure 1). figure 1. the basic structure of flavonoids. y. ysrafil et al. admet & dmpk 00(0) (2023) 000-000 6 the basic structure of flavonoids consists of aglycones, but their presence in nature is usually bound to glycosides and methylated derivatives. they are produced via the shikimic acid pathway using acetic acid/phenylalanine derivatives as the precursor. flavonoids are traditionally classed by oxidation degree, ring c annularity, and ring b connectivity [1,37]. they comprise 6 subclasses, as presented in table 2 [38,39]. among higher plant families and genera, the flavones and flavonols are the most commonly highly conserved encountered and marked by the presence of the 2-phenylchromen-4-one (2-phenyl-1-benzopyran-4-one) structures in their backbone [40]. table 2. subclasses of flavonoids subclass structure type ref. flavones apigenin luteolin acacetin diosmetin chrysoeriol [41,46-50] flavonols quercetin kaempferol galangin myricetin fisetin [51-56] flavanols catechin epicatechin gallocatechin [57] flavanones naringenin naringin hesperidin hesperetin eriodictyol silybin [43,58] isoflavones genistein daidzein glycitein formononetin biochanin a [59] anthocyanins cyanidin malvidin peonidin delphinidin pelargonidin petunidin [60] flavones are distinct from other flavonoids because their skeleton contains a double bond between c2 and c3. also, there is no substitution at their c3 position, while flavonols possess hydroxyl substitution at that position. further, flavones are oxidized at the c4 position [41,42]. flavanones and flavanols have saturated bonds between c2 and c3 and frequently coexist in plants with flavones and flavonols [37]. flavanones have a flavan nucleus formed by two aromatic rings (a and b) linked by a dihydropyrone ring (c). the presence of a saturated c2–c3 bond, a chiral carbon atom at the c2 position, and no substitution at the c3 position of the c ring distinguish flavanones from the other two classes of flavonoids, i.e., flavones and flavanols [43]. flavanols (flavan-3-ols) are heterocyclic compounds that contain a saturated heterocyclic ring, a single bond between c2 and c3, and a hydroxyl group at the c3 position. unlike other flavonoids, flavanols are only found in food as aglycones (the glycosylated state is excluded). additionally, they are found in monomeric form as catechins and epicatechins, as well as in polymeric form as tannins [44]. isoflavones are admet & dmpk 00(0) (2023) 000-000 anti-inflammatory activities of flavonoid derivates doi: https://doi.org/10.5599/admet.1918 7 secondary metabolites formed by rhizobial bacteria and leguminous plant defence responses [45]. isoflavones, like daidzein, include 3-phenylchromones. chalcones are ring c-opening isomers of dihydroflavones that are responsible for plant colour. anthocyanidins are a category of significant chromene pigments that exist as ions. other flavonoids without the c6—c3—c6 structure include biflavones, furan chromones, and xanthones. glycosides are the most common extant flavonoid form. the structure of aglycones dictates preferred glycosylation sites [37]. pharmacokinetics and bioavailability bioavailability to predict and explain the effect of flavonoids at their site of action, knowledge of the pharmacokinetics, form, and concentration of these compounds in plasma must be known. animal and human studies have found that these compounds have poor oral bioavailability, which might be attributed to the loss of the compounds during the absorption and metabolism phases. in addition, these compounds also tend to have poor water solubility, low permeability, as well as poor stability profile [13,61]. therefore, various studies have been carried out to engineer their structure to increase their water solubility and bioavailability and thereby increase their activity as a drug candidate as have been demonstrated in the treatment of several diseases [13,46,62]. absorption absorption of flavonoids in the intestine can occur through several mechanisms, namely active transport, passive diffusion, or both. some flavonoids, such as quercetin, are absorbed completely in the small intestine by involving sodium-dependent glucose transporters 1 (sglt1). these transporters are located in the apical membrane of intestinal epithelial cells that utilize na+/k+-atpase pumps to transport this flavonoid. in addition, the absorption of this compound is also carried out by glucose transporter 2 (glut2) located on the basolateral membrane of the intestine. however, the exact mechanism used by glut2 in transporting flavonoids is still unclear [63-65]. further, quercetin is hydrolysed by the intracellular enzymes called lactasephloridzin hydrolase (lph) and cytosolic β-glucosidase (cbg) in the aglycone form so that it is easily absorbed through the small intestine. in addition to quercetin, these enzymes can work on other flavonoid compounds. however, the effectiveness of these enzymes depends on the glycosides present in the flavonoid structure [41,64,66,67]. it should be noted that each compound grouped as flavonoids have different maximum absorption times. a study that was conducted in rats revealed that the absorption of some flavonoid derivatives such as apigenin, luteolin, and glucosides generally occurred rapidly with a maximum absorption time (tmax) of about 1 hour for maximum plasma concentration (cmax) of 1 to 100 μmol/l and it will change depends on co-consumed food [41,68]. distribution in circulation, flavonoids can bind to several blood proteins [69]. gecibesler & aydin [70] demonstrated that flavonoids could bind strongly to human serum albumin (hsa). hsa is an essential protein in blood plasma found abundantly in the body [71]. this soluble protein can bind and transport various metabolites and organic compounds such as flavonoids, unesterified fatty acids, hormones, and metal ions to various tissues throughout the body [72,73]. research by boer et al. [74] revealed that flavonoid compounds, in particular quercetin and its derivatives, are widely distributed in rat tissues with the highest levels in the lungs, testes, and kidneys, while the brain, white fat, and spleen showed the lowest levels. this group also reported that this compound was deposited in several other organs in the rat, including the thymus, heart, liver, brown fat, bone, and muscle. meanwhile, y. ysrafil et al. admet & dmpk 00(0) (2023) 000-000 8 in pigs, the deposition of quercetin was more abundant in the liver and kidney tissue after a quercetin diet of 500 mg/kg with a lesser extent in the brain, heart, and spleen . a similar finding was reported by beiger et al. [75] demonstrating that the distribution of quercetin and some of its derivatives was found to be highest deposited in the liver, kidney, jejunum, lung, and muscle (longissimus dorsi) after a single dietary quercetin dose (25 mg/kg) in pigs. furthermore, quercetin deposition was more distributed when administered at a dose of 50 mg/kg per day for four weeks, including colon (higher), kidney, jejunum, liver, lymph node, mesentery, lung, white adipose tissue, muscle (diaphragm), muscle (longissimus dorsi), and brain (as the smaller deposition). metabolism after ingestion, approximately 10 % of flavonoid glycosides are absorbed in the upper gastrointestinal tract, while the remaining 90 % pass through the small intestine to reach the large intestine as unmetabolized and unabsorbed flavonoids. the unabsorbed flavonoids undergo a further enzymatic transformation in the small intestine, such as oxidation, reduction, and decarburization, as the preparatory steps before entering the large intestine. once in the large intestine, they are hydrolysed and cleaved to remove glycosides and produce flavonoid aglycones, which are further metabolized to ring fission products. this process is carried out by the involvement of the lactase-phlorizin hydrolase (lph) and colonic enzymes produced by the intestinal microbiota [41,76-78]. flavonoids then extensively undergo two metabolic phases, namely phases 1 and 2, which occur in enterocytes and hepatocytes [79,80]. in the first phase, which takes place in hepatocytes, flavonoids will be oxidized by the cytochrome p450 (cyp450) enzyme to produce a minor contribution to flavonoid clearance. different isoforms of the cyp450 enzyme play the metabolic process of several flavonoid-derived compounds in this phase. isoform cyp2c9 is the most efficient enzyme for the metabolism of galangin followed by cyp1a3 and cyp1a1. meanwhile, most of the oxidation of galangin is determined by the action of cyp1a2 and followed by cyp2c9 and cyp1a1 [81]. another investigation reported that apigenin was metabolized by cyp1a1, cyp2b, and cyp2e to form three monohydroxylated derivatives [82]. in contrast to the first phase, phase ii metabolism occurs in enterohepatic and enteric and is predominant in the direction of flavonoid disposition. in this stage, flavonoids undergo two main processes, namely glucuronidation and sulfation. while the former is mediated by udp glucuronosyltransferases (e.g., ugt1a1, ugt1a8, and ugt1a9), the latter is catalysed by sulfotransferases (e.g., sult1a1 and sult1e1). methylation can also occur in this phase. the process is mediated by a group of enzymes known as methyltransferases mostly found in many tissues, including the liver and intestines. the most common methylation reaction associated with flavonoid metabolism is o-methylation which occurs in the liver and is catalysed by catecholo-methyltransferase (comt) [41,64,69,79,83,84]. the products of these metabolic processes will then be circulated and deposited in various body components to carry out their actions, including their antiinflammatory activities, as elaborated in the section “flavonoid derivates as anti-inflammatory agents and their mechanisms of action”. excretion flavonoids are mainly excreted through the urine. a study reported that about 60-80 % of the metabolites of anthocyanins are excreted in the urine [85,86]. the excretion pathway of flavonoids depends on the conjugate form and the site of elimination. glucuronide formed in the intestine tends to enter the systemic circulation directly and is excreted in the urine, while the products formed in the liver are mainly eliminated admet & dmpk 00(0) (2023) 000-000 anti-inflammatory activities of flavonoid derivates doi: https://doi.org/10.5599/admet.1918 9 through the bile. this indicates that glucuronidation in enterocytes favors the loss of flavonoids from the enterohepatic circulation (ehc) into the systemic circulation. ehc-undergoing flavonoid compounds may also be further sulphated in the intestine or hepatocytes, leading to urinary excretion (without bypassing biliary excretion). another study in isolated human faecal colonic bacteria found the presence of flavonoids such as quercetin, indicating the importance of the faecal route in facilitating the excretion of these compounds [76,86,87]. to sum up, the biological activities of flavonoids are diminished by the fact that these compounds have low aqueous solubility and poor bioavailability, as depicted in their pharmacokinetic profile described above. this leads to the development of various strategies to increase their solubility and oral bioavailability, including the development of nanoparticle-based delivery systems, which will be discussed in the section “application of nanomedicine as delivery system of flavonoid derivates”. figure 2. pharmacokinetics of flavonoid. flavonoid derivates as anti-inflammatory agents and their mechanisms of action various studies of in silico, in vitro, in vivo, and clinical trials have demonstrated that flavonoids have potential activities in mitigating inflammation [11,88]. they can modulate various key points in regulating inflammation in the early and late stages. as displayed in table 1, these compounds consist of several subclasses, and they can exert their anti-inflammatory properties via various mechanisms, including inhibition of the various inflammatory signaling pathways such as nf-κb [88], mapk [89], blockage of proinflammatory enzymes (cox-1, cox-2, and 5-lox) [11], inhibition of proinflammatory cytokines release (tnfα, il-1β, and il-6), and suppression of other inflammatory proteins [62]. the summary of the putative anti-inflammatory mechanisms of flavonoids is presented in table 2 and figure 3. y. ysrafil et al. admet & dmpk 00(0) (2023) 000-000 10 table 3. anti-inflammatory activity of various flavonoids. class of flavonoid sources flavonoid compound study target molecules references flavanol black and green tea, red wine, red grapes, bananas, apples, blueberries, peaches, and pears catechin, epigallocatechin-3galate, epicatechin gallate. in vitro and in vivo nf-κb, il-6, inos, cox-2, il-1β, mapk pathway, nox, ros, ikk, il8, jnk1/2, p38, jak/stat1 pathway, nrf2, and pi3k/akt [8,89-93] flavanone paulownia tomentosa fruits tomentodiplacone o in silico, in vitro and in vivo cox-1, cox-2, and 5-lox [11] citrus fruits hesperetin in vitro, ex vivo and vivo tlr4, nf-κb, tnf-α, no, pge2, inos, i-κb, jnk1/2, p38, mapk and cox-2 [62,94] grapefruits, sour orange, tart cherries, tomatoes, greek oregano. naringenin in vitro and in vivo tnf-α, il-1β, il-6, nf-κb, nrf2, trpv1, trpm3, trpm8, caspase-3, bad, bax, tlr4, inos, cox2, nox2, mapk, tgf-β1, hsp70, mmp-3, il-33, and ho-1 [95-97] flavanone citrus fruits naringin in vitro, and in vivo tnf-α, il-1β, il-6, il-8, nf-κb, smad-7, tgf-β, inos, vegf, pol-γ, caspase-3, smad-3, and bax [98-102] citrus fruits hesperidin in vitro and in vivo ros, cd14, nfκβ, e-selectin, tnf‑α, il‑1β, inos, il‑6, mcp‑1, icam-1, cox-2, ifn-γ, il-2, il-4, il10, p65, foxo1, foxo3 and nrf2 signaling pathway, mmp-3, mmp9, il-10, timp-1, and sox9 [103-110] citrus fruits eriodictyol in vitro and in vivo nf-κb, p38, mapk, erk1/2, jnk, tnf-α, il-6, nrf2 pathway, mip-2, inos, il-8, foxo1, pi3k/akt pathway, cox-2, pge2, iκbα, pp65, erk1/2, jnk, no and il-1β [111-114] milk thistle silybin in vitro and in vivo erk, mek, mcp-1, il-5, il-10, ifnγ, il-17a, gm-csf, il-1β, tgf-β, il6, ik-bα and raf [115,116] flavonol berries, grapefruit, onion, olive oil, and red wine fisetin, in vitro, and in vivo myd88 and nf-κb signaling pathways, no, pge2, il-6, tnf-α, inos, cox-2a, il-1β, ser9, βcatenin, il-2, il-4, ifn-γ, il-18 and il-5 [117,118] kaempferol in silico, in vitro, in vivo, and clinical trial myd88 and nf-κb signaling pathways, il-6, il-1β, il-18, il-8, tnf-α, akt, src, syk, irak1, irak4, nrf2, tlr4, inos, il-10, il-12, p70, lox, ros, icam-1, e-selectin, crp, cox1 and cox2 [51, 119-124] myricetin in vitro, and in vivo no, inos, pge2, cox-2, tnf-α, il6, nf-κb, iκbα, nrf2, stat1, i fn-β, il-12, il-1β, jak/stat1, nox2/p47 phox, rankl/rank, mapk, malat1, akt, ikk, akt, and mtor. [55, 125-127] quercetin in vitro, in vivo and clinical trial tnf-α, il-6, il-8, il-1β, il-18, il-12, nf-κb, jnk1/2, c-jun, erk1/2, mcp-1, no, inos and cox-2, ampk, sirt1, cd80, cd86, dabs, src, pi3k, akt, mhc-ii. [128-133] flavone magnolia officinalis, fruit peels, red peppers, and tomato skins apigenin in vitro, and in vivo cox-2, nf-κb, tlr4, tgf-β, tnf-α, il-1β, il-6, il-2, il-8, no, inos, ap1 (c-jun, c-fos, and junb), erk, mcp-1, iκbα, icam-1, ros, akt, mtor, jnk, and p38-mapk [134-139] chrysin in vitro, and in vivo il-1β, il-6, nlrp3, tgf-β, tnf-α, nf-кb [140-142] luteolin in silico in vitro, and in vivo nf-κb, mapk, ap1, socs3, stat3, il-1β, il-6, il-2, il-8, il-12, il-10, il-17, tnf-α, ifn-β, ccl2, cxcl2, cxcl8, and cxcl9, nlrp3, tgf-β, nf-кb, inos, no, cox− 2, pge2, gm-csf, mcp− 1. mmp− 2 and mmp− 9. [46] admet & dmpk 00(0) (2023) 000-000 anti-inflammatory activities of flavonoid derivates doi: https://doi.org/10.5599/admet.1918 11 class of flavonoid sources flavonoid compound study target molecules references isoflavone soya bean and peanuts daidzin in vitro and in vivo tnf-α, il6, il-1β, il-8, no, inos, cxcl2, pparα/γ, p38, iκb-α, stat1, ccl2, and jnk [143-146] genistein in vitro and in vivo il-6, il-1β, il-8, il-12, il-20, il-23, ifn-γ, tnf-α, vegfa, ccl2, tnf-α, nf-κb, pgs, inos, ros, ampk, mapk, jnk, tlr4, iκbα, ikk, cox2, mcp-1, kc, icam-1 and vcam-1 [147,148] anthocyanins cherry, elsberry, blueberry, hibiscus plants, and strawberry cyanidin, delphinidin, malvidin in vitro and in vivo nf-κb, mapk, mek1/2-erk1/2, jnk, il-1β, il-6, tnf-α, cox-2, pge2, inos, no, mcp-1, icam-1, vcam-1, and cinc-1. [149-154] flavanonols milk thistle silibinin in vitro, in vivo, and clinical trial nf-κb, mapk, tnf-α, il-6, il-8, il1β, il-10, mmp-9, cox-2, pge2, pgf2α, th17, sirt1, cyclin d1, bcl2, inos, no, vegf, mmp-1, mmp3, mmp-13, pi3k/akt [88,155-159] figure 3. mechanism of action of flavonoids as an anti-inflammatory agent. flavonoid inhibit eicosanoids metabolism and function prostaglandins and leukotrienes are potent lipid mediators derived from phospholipase-released arachidonic acid that play a critical role in inflammation [160]. they are produced from locally damaged cell membranes with the help of several enzymes such as phospholipase a2 (pla2), cyclooxygenase, and lipoxygenases as previously explained (in section lipid-based signaling: arachidonic acid-derived eicosanoids). several approved drugs such as corticosteroid-including drugs (such as methylpredisolone etc.) and nonsteroidal anti-inflammatory drugs (nsaids; e.g., mefenamic acid, indomethacin, ibuprofen) have targeted these enzymes for the treatment of several inflammation-related diseases such as pain, injury, asthma, allergies, and inflammation. elevated cox-2 levels have indeed shown adverse effects on inflamematory conditions and even on some other inflammation-related diseases such as tumor development, so their suppression could significantly suppress worsening prognosis [161]. y. ysrafil et al. admet & dmpk 00(0) (2023) 000-000 12 different studies demonstrated that flavonoid-derived compounds can inhibit enzymes, including phospholipase a2, cyclooxygenase, and lipoxygenases, improving inflammatory disorders [52,160,162,163]. several pieces of evidence have proposed their effect to block the pathway and mediate improvement in an intestinal inflammatory response. a study by lópez-posadas [162] reported that flavonoids could inhibit the formation of prostaglandins from the aa pathway by suppressing epithelial cox-2 expression during inflammation in intestinal epithelial cells. this finding is supported by another study reported by serra and colleagues that advocated c3g to provide strong inhibitory activity against cox-2 and lead downregulation of several inflammatory mediators in the colonic carcinoma cell [164]. apart from intestinal inflammation, several flavonoid derivative compounds have also been reported to be effective in inhibiting eicosanoid production in other inflammatory cases. a study by husain et al. suggested that epigallocatechin-3-gallate (egcg) can downregulate cox-2 in human prostate carcinoma lncap and pc-3 cells in a dose-dependent fashion [165]. similar findings demonstrated the efficacy of egcg could downmodulation of cox-2 were further lead reduction of pge2 levels [166,167]. in another finding, giving egcg results in the inhibition of nf-κb that in order leads to the inhibition of cox-2 promoter activity [168]. several other flavonoid derivative compounds, such as luteolin, kaempferol, hesperetin, and naringin, have been shown to have inhibitory effects on this cox-2 enzyme [161,169]. topical pre-treatment of kaempferol could attenuate uvb-induced cox-2 expression in mouse skin. this effect is hypothesized to arise through the modulation of src kinase activity in the mouse skin [161]. another promotion study demonstrated by he and colleagues [170], isolated and elucidated flavonoid compounds from hosta plantagine and found twelve 12 different ones that showed significant inhibitory effects to both enzyme isoforms from 53.00 to 80.55 % (in cox-1) and 52.19 to 66.29 % (to cox-2) in concentration 50 μm. the increase in eicosanoids is also mediated by the cox-1 isoform, especially in some cases of neuroinflammation. elevated cox-1 levels in microglia may explain pge2 levels in the cerebrospinal fluid of creutzfeldtjacob disease (cjd) patients associated with shorter survival [171]. regarding that, it has been reported that quercetin and tomentodiplacone-o could strongly inhibit cox-1 [11]. similar to this finding, kaempferol has been advocated to have an inhibitory effect on cox-1 in vitro. in addition to the cyclooxygenase pathway, several researchers have reported that this compound can inhibit the expression of lipooxygenase (lox) enzymes which play an important role in the metabolism of arachidonic acid into leukotrienes, which is involved in various inflammatory-related diseases including asthmatics, rheumatoid arthritis, psoriasis, inflammatory bowel disease etc. [51,172]. administration of several other flavonoid derivative compounds, such as quercetin and luteolin has also been shown to have a strong inhibitory effect on these enzymes to lead to improvement in symptoms of inflammation-related leukotrienes expression [169]. the effectiveness of flavonoids in inhibiting the formation of eicosanoids in the aa-related pathway has also been proposed through the pla2 enzyme inhibition mechanism. this enzyme plays an important role in the process of converting membrane phospholipids into product intermediates, namely arachidonic acid, before being further converted into lipid mediators. many recent studies have reported the activity of various flavonoid derivative compounds having good activity in inhibiting this enzyme and treating inflammation [173,174]. a study conducted by enechi and reported colleagues [175] that flavonoid-rich extract from peltophorum pterocarpum stem-bark provide a strong inhibitory effect in pla2 and generate downmodulate of the inflamematory symptoms of paw oedema induced by egg albumin, in comparison with prednisolone as well approve anti-inflammatory drugs that act in pla2. in accordance, several other studies have also demonstrated compounds such as quercetin, kaempferol, and galangin showing promising inhibitory activity against pla2 [173,174,176]. admet & dmpk 00(0) (2023) 000-000 anti-inflammatory activities of flavonoid derivates doi: https://doi.org/10.5599/admet.1918 13 flavonoid inhibit mapk and nf-κβ signalling pathways various evidence has suggested the crucial role of nuclear factor kappa beta (nf-κβ) and mitogen-activated protein kinases (mapks) as two main signaling pathways that are interrelated in underlying inflamemation [177]. nf-κβ accumulation has been implicated with increased transcription and expression of various inflammatory mediators. meanwhile, mapks, consisting of p38 kinase, erk, and c-jun nh2-terminal kinase (jnk), have been advocated to regulate the transcription of cox-2, inos, and inflammatory cytokines in inflammation. therefore, increasing evidence has studied and targeted these signaling pathways in the development of anti-inflammatory drugs [177-179]. a study conducted by hour and colleagues [177] reported that dihydromyricetin, a flavonoid isolated from ampelopsis grossedentata has been proposed to have good and promising inflamematory potential by inhibiting the phosphorylation of nf-κb and iκbα as well as p38 and jnk. this suppression simultaneously leads to a downmodulation effect on various proinflammatory cytokines such as tnf-α, il-1β, and interleukin-6, as well as increasing levels of il-10 as an anti-inflammatory cytokine. furthermore, inhibition of these two signaling pathways also reduces the levels of inos, tnf-α, and cox-2. in another finding, an in vivo study conducted by ichikawa and colleagues [180] had proposed epigallocatechin gallate (egcg) to have a significant effect of reducing il-6 expression, ikk activation, iκbα degradation, and activation of nf-κb which in turn inhibited inflammatory events that mediated by this pathway. the inhibition also reduces myocardial damage after ischemia and reperfusion in rats [181]. furthermore, these compounds show promising effects in inhibiting iκbα degradation and the action of nf-κb in binding to dna that leads to suppressing il-12p40 and inos expression in lps-activated macrophages [180]. in addition to egcg, quercetin has been reported to have good anti-inflammatory activity through the same mechanism as egcg, inhibiting iκbα phosphorylation and downmodulating the nf-κb pathway [182]. growing evidence has been demonstrated that compounds that can inhibit jnk, p38, and ikkβ generally exhibit anti-inflammatory effects and suggest a relationship between mapk and nfκb in inflamematory events. flavonoid derivatives have been reported to affect the mapk signalling pathway at various stages. a study has proven that egcg is able to influence this pathway through the inhibition of erk1/2, p-jnk, and pp38 in human ra synovial fibroblasts (rasfs) [183]. another study reported that kaempferol, luteolin, chrysin, and apigenin exert antiinflammatory activity by inhibiting mapk signalling pathways such as jnk, p38, and erk [181]. application of nanomedicine as delivery system of flavonoid derivates herbs have been widely used to promote human health since ancient times. advances in phytochemistry and pharmacology have made it possible to determine the composition and bioactivity of many medicinal products. it has been proved that the effectiveness of many herbal medicines depends on the delivery of the active compounds. however, their success in clinical trials has been less impressive, partly due to the compound's poor bioavailability [184,185]. furthermore, the efficacy of many drugs obtained from natural sources, including flavonoids in their micro/macro formulations, has also been shown to have poor bioavailability and pharmacokinetics following their oral administration, leading to their less effectivity [184,186]. researchers have developed other drug delivery methods while avoiding potential causes of decreased bioavailability, such as firstpass metabolism and drug malabsorption in the digestive tract [187,188]. recently, the use of nanotechnology has shown tremendous success in the field of drug delivery, including natural compound-based drugs for the treatment of inflammation. this technique formulates the natural compounds in the form of nanosized particles (1 to 100 nm) [189,190]. it has been suggested that the encapsulation of flavonoids into nanocarriers is a beneficial technique to protect drug molecules from y. ysrafil et al. admet & dmpk 00(0) (2023) 000-000 14 changes in their size, shape, and surface characteristics that might be caused by environmental factors. it has been proved that this method can increase the bioavailability of flavonoids and improve the targeted and controlled release profile of the natural products, which in turn increases the drug's effectiveness [185]. furthermore, nano-delivery systems could protect the therapeutic agents from being enzymatically metabolized, resulting in their increased stability and circulation time [191-195]. the smaller size of the nanoparticles has also been proven to increase the penetration of the compounds compared to conventional topical formulations [196]. recently, the most common types of nanoparticles used for drug delivery are polymer nanoparticles, including solid lipid nanoparticles, liposomes, micelles, crystalline nanoparticles, and dendrimers [185,197,198]. solid lipid nanoparticles solid lipid nanoparticles (slns) are nanometer colloidal carriers composed of the solid particle lipid core in which active substances are trapped and stabilized by surfactants [199]. their matrix is typically composed of solid lipids such as glycerides, sterols, fatty acids, and waxes. in particular, sln has recently been proposed for oral and dermal administration of phenolic compounds to protect them from chemical degradation [200]. tan et al. [201] demonstrated that the formulation of total flavonoid-based solid lipid nanoparticles (tf-sln) could produce spherical particles in shape with a uniform size distribution of 104.83 nm and a zeta potential of 28.7 mv. furthermore, this formulation was proved to have a better myocardial protection effect than flavonoids alone. another study showed that the formulation of quercetin-loaded solid lipid nanoparticles (qslns) exhibited an increase in their bioavailability by 3.5-fold compared to quercetin alone in rats [12]. polymeric nanoparticles natural or synthetic biodegradable polymer nanoparticles have become prominent in the field of nanomedicine for targeted drug delivery to improve biocompatibility, bioavailability, safety, increased permeability, preferable retention time, and less toxicity [186,202]. drugs that are conjugated to macromolecules such as synthetic polymers, natural polysaccharides, or proteins through covalent bonds provide a direct and efficient approach to modifying pharmacokinetic performance, promoting the stability of drugs, and preventing drug leakage or explosive release during transport [203]. chitosan is the most frequently used polymer for the formulation of nano-based drugs for their delivery to their target of action. wang et.al., [204] reported that the administration of the inhaled baicalein encapsulated with chitosan-nanoparticle could mediate delivery of this drug and lead to the reduction of il5 level, enhancement of il-12 and controlled inflammation-related asthma in mice. in addition, polymer nanoparticles are also effective for mediating drug delivery given topically. a study by nan et.al. [205] revealed that the formulation of a topical preparation of quercetin-loaded chitosan nanoparticles led to the increase of percutaneous absorption and retention of quercetin in the skin and improve its effects against ultraviolet b radiation. nanoliposomes nanoliposomes are phospholipid vesicles containing one or more bilayers surrounding an aqueous core. nanoliposome-based drugs are effective for the development of transdermal drugs, topical drugs, drugs used to target hair follicles and other uses. nanoliposomes usually consist of phospholipids with cholesterol often added to improve the stability of the liposome bilayer by filling the gaps caused by incomplete packaging. some of the literature report that the use of nanoliposomes in natural product-based drug, especially flavonoids, is effective in delivering drugs well to the targeted sites leading to the generation of desired effects. a study by wang et al. [206] explained that the formulation of nanoliposome of quercetin could admet & dmpk 00(0) (2023) 000-000 anti-inflammatory activities of flavonoid derivates doi: https://doi.org/10.5599/admet.1918 15 induce type iii-programmed cell death in c6 glioma cells. furthermore, another study by jin et al. [207] exhibited that treatment of apigenin-loaded d-alpha-tocopheryl polyethylene glycol 1000 succinate (tpgs) liposomes could deliver apigenin well and exert inhibitory effects on tumor growth in a549 cell-bearing mice. nanoemulsions nanoemulsions are formed by combining two mechanically shear immiscible liquid phases with a surfactant having a droplet size of less than 100 nm. the formulation of natural product-based drugs with this method is able to overcome the weakness of flavonoid compounds that have problems with their bioavailability by improving their pharmacokinetic profile. a recent study by fuior et al. [208] demonstrated that the formulation of flavonoid-loaded lipid nanoemulsions (flns) generated good characteristics (size of 200 nm, negative zeta potential, a high encapsulation efficiency >80 %, good in vitro stability, and steady release). further, this formulation facilitated the action of the flavonoid in reducing endothelial inflammation. this finding is supported by a study conducted by zain et al. [209] investigating the relationship between the delivery of nanoemulsion flavonoid-enriched oil palm and its wound-healing activity. self-nano-emulsifying drug delivery system (sneeds) snedds is a nanoemulsion pre-concentrate in which the drug is encapsulated in an oil phase in the presence of a surfactant/cosurfactant, capable of forming very small nanoscale droplets/nanoemulsion when gently mixed with an aqueous medium. self-emulsifying drug delivery systems are isotropic mixtures of oils, surfactants, solvents, and cosolvents/surfactants which are developed to improve the physicochemical as well as pharmacological properties of highly lipophilic drugs, including flavonoids [210-212]. it has been demonstrated that the loading of quercetin into snedds could effectively enhance the hepatoprotective activity of quercetin in mice suffering from hepatotoxicity [213]. another study performed in wistar rats confirmed that the formulation of apigenin into snedds enhanced the oral bioavailability of the compound [214]. concluding remarks flavonoids are secondary metabolites found abundantly in various plants. they exist in several subclasses, including flavanols, flavanones, flavonols, flavones, isoflavones, anthocyanins, and flavanonols. it has been demonstrated that flavonoids possess potential pharmacological activities, including their potency in mitigating inflammation. a number of studies conducted in silico, in vitro, in vivo, and clinical trials have deciphered that the anti-inflammatory activities of flavonoids are closely linked to their ability to modulate various biochemical mediators, enzymes, and signalling pathways involved in the inflammatory processes. however, these anti-inflammatory activities have been facing challenges as flavonoids have poor solubility and bioavailability. to tackle these challenges, the application of nanotechnology in developing better pharmaceutical dosage forms has brought promising results. various nanotechnology platforms, including solid nanoparticles, polymeric nanoparticles, nanoliposomes, nanoemulsion, self-nano emulsifying drug delivery systems, have been reported to have better action in delivering flavonoids to their targeted sites. these platforms can potentially increase the solubility, bioavailability, and pharmacological activity of flavonoids and reduce the toxic effects of these compounds. conflict of interest: there is no conflict of interest associated of this article as well as no significant support that influenced the content. acknowledgements: the author gratefully thank to ysrafil, to conductiing this projects, apt. firzan nainu, m. biomed., sc., ph.d and talha bin emran, ph.d for supervised and grammar/structural editing of this manuscript. y. ysrafil et al. admet & dmpk 00(0) (2023) 000-000 16 references [1] r. ginwala, r. bhavsar, d.i. chigbu, p. jain, z.k. khan. potential role of flavonoids in treating chronic inflammatory diseases with a special focus on the anti-inflammatory activity of apigenin. antioxidants 8(2) (2019). https://doi.org/10.3390/antiox8020035. [2] s.j. maleki, j.f. crespo, b. cabanillas. anti-inflammatory effects of flavonoids. food chemistry 299 (2019) 125124. https://doi.org/10.1016/j.foodchem.2019.125124. [3] t.y. nguyen, d.c. to, m.h. tran, j.s. lee, j.h. lee, j.a. kim, m.h. woo, b.s. min. anti-inflammatory flavonoids isolated from passiflora foetida. natural product communications 10(6) (2015) 929-931. [4] h. herwald, a. egesten. on pamps and damps. journal of innate immunity 8(5) (2016) 427-428. https://doi.org/10.1159/000448437. [5] k.c. ma, e.j. schenck, m.a. pabon, a.m.k. choi. the role of danger signals in the pathogenesis and perpetuation of critical illness. american journal of respiratory and critical care medicine 197(3) (2018) 300-309. https://doi.org/10.1164/rccm.201612-2460pp. [6] m.g. netea, f. balkwill, m. chonchol, f. cominelli, m.y. donath, e.j. giamarellos-bourboulis, d. golenbock, m.s. gresnigt, m.t. heneka, h.m. hoffman, r. hotchkiss, l.a.b. joosten, d.l. kastner, m. korte, e. latz, p. libby, t. mandrup-poulsen, a. mantovani, k.h.g. mills, k.l. nowak, l.a. o'neill, p. pickkers, t. van der poll, p.m. ridker, j. schalkwijk, d.a. schwartz, b. siegmund, c.j. steer, h. tilg, j.w.m. van der meer, f.l. van de veerdonk, c.a. dinarello. a guiding map for inflammation. nature immunology 18(8) (2017) 826-831. https://doi.org/10.1038/ni.3790. [7] t. schmid, b. brüne. prostanoids and resolution of inflammation beyond the lipid-mediator class switch. frontiers in immunology 12 (2021) 714042. https://doi.org/10.3389/fimmu.2021.714042. [8] k.w. choy, d. murugan, x.f. leong, r. abas, a. alias, m.r. mustafa. flavonoids as natural antiinflammatory agents targeting nuclear factor-kappa b (nfκb) signaling in cardiovascular diseases: a mini review. frontiers in pharmacology 10 (2019) 1295. https://doi.org/10.3389/ fphar.2019.01295. [9] y. peng, m. ao, b. dong, y. jiang, l. yu, z. chen, c. hu, r. xu. anti-inflammatory effects of curcumin in the inflammatory diseases: status, limitations and countermeasures. drug design, development and therapy 15 (2021) 4503-4525. https://doi.org/10.2147/dddt.s327378. [10] r. fürst, i. zündorf. plant-derived anti-inflammatory compounds: hopes and disappointments regarding the translation of preclinical knowledge into clinical progress. mediators of inflammation 2014 (2014) 146832. https://doi.org/10.1155/2014/146832. [11] z. hanáková, j. hošek, z. kutil, v. temml, p. landa, t. vaněk, d. schuster, s. dall'acqua, j. cvačka, o. polanský, k. šmejkal. anti-inflammatory activity of natural geranylated flavonoids: cyclooxygenase and lipoxygenase inhibitory properties and proteomic analysis. journal of natural products 80(4) (2017) 999-1006. https://doi.org/10.1021/acs.jnatprod.6b01011. [12] n. ahmad, v.t. banala, p. kushwaha, a. karvande, s. sharma, a.k. tripathi, a. verma, r. trivedi, p.r. mishra. quercetin-loaded solid lipid nanoparticles improve osteoprotective activity in an ovariectomized rat model: a preventive strategy for post-menopausal osteoporosis. rsc advances 6(100) (2016) 97613-97628. https://doi.org/10.1039/c6ra17141a. [13] g.e. batiha, a.m. beshbishy, m. ikram, z.s. mulla, m.e.a. el-hack, a.e. taha, a.m. algammal, y.h.a. elewa. the pharmacological activity, biochemical properties, and pharmacokinetics of the major natural polyphenolic flavonoid: quercetin. foods 9(3) (2020). https://doi.org/10.3390/foods9030374. [14] s. tasneem, b. liu, b. li, m.i. choudhary, w. wang. molecular pharmacology of inflammation: medicinal plants as anti-inflammatory agents. pharmacological research 139 (2019) 126-140. https://doi.org/10.1016/j.phrs.2018.11.001. [15] h.m.a. ullah, t.h. kwon, s. park, s.d. kim, m.h. rhee. isoleucilactucin ameliorates coal fly ashinduced inflammation through the nf-κb and mapk pathways in mh-s cells. international journal of molecular sciences 22(17) (2021). https://doi.org/10.3390/ijms22179506. https://doi.org/10.3390/antiox8020035 https://doi.org/10.1016/j.foodchem.2019.125124 https://doi.org/10.1159/000448437 https://doi.org/10.1164/rccm.201612-2460pp https://doi.org/10.1038/ni.3790 https://doi.org/10.3389/fimmu.2021.714042 https://doi.org/10.2147/dddt.s327378 https://doi.org/10.1155/2014/146832 https://doi.org/10.1021/acs.jnatprod.6b01011 https://doi.org/10.1039/c6ra17141a https://doi.org/10.3390/foods9030374 https://doi.org/10.1016/j.phrs.2018.11.001 https://doi.org/10.3390/ijms22179506 admet & dmpk 00(0) (2023) 000-000 anti-inflammatory activities of flavonoid derivates doi: https://doi.org/10.5599/admet.1918 17 [16] a.l. souto, j.f. tavares, m.s. da silva, f. diniz mde, p.f. de athayde-filho, j.m. barbosa filho. antiinflammatory activity of alkaloids: an update from 2000 to 2010. molecules 16(10) (2011) 8515-8534. https://doi.org/10.3390/molecules16108515. [17] t. ito. pamps and damps as triggers for dic. the journal of clinical investigation 2(1) (2014) 67. https://doi.org/10.1186/s40560-014-0065-0. [18] s. watanabe, m. alexander, a.v. misharin, g.r.s. budinger. the role of macrophages in the resolution of inflammation. journal of intensive care 129(7) (2019) 2619-2628. https://doi.org/10.1172/ jci124615. [19] j.s. roh, d.h. sohn. damage-associated molecular patterns in inflammatory diseases. immune network 18(4) (2018) e27. https://doi.org/10.4110/in.2018.18.e27. [20] y. su, j. gao, p. kaur, z. wang. neutrophils and macrophages as targets for development of nanotherapeutics in inflammatory diseases. pharmaceutics 12(12) (2020). https://doi.org/10.3390/ pharmaceutics12121222. [21] l. chen, h. deng, h. cui, j. fang, z. zuo, j. deng, y. li, x. wang, l. zhao. inflammatory responses and inflammation-associated diseases in organs. oncotarget 9(6) (2018) 7204-7218. https://doi.org/ 10.18632/oncotarget.23208. [22] t. kawasaki, t. kawai. toll-like receptor signaling pathways. frontiers in immunology 5 (2014) 461. https://doi.org/10.3389/fimmu.2014.00461. [23] o. takeuchi, s. akira. pattern recognition receptors and inflammation. cell 140(6) (2010) 805-820. https://doi.org/10.1016/j.cell.2010.01.022. [24] m. murakami, t. hirano. the molecular mechanisms of chronic inflammation development. frontiers in immunology 3 (2012) 323. https://doi.org/10.3389/fimmu.2012.00323. [25] a.e.f. sheppe, m.j. edelmann. roles of eicosanoids in regulating inflammation and neutrophil migration as an innate host response to bacterial infections. infection and immunity 89(8) (2021) e0009521. https://doi.org/10.1128/iai.00095-21. [26] p.c. calder. omega-3 polyunsaturated fatty acids and inflammatory processes: nutrition or pharmacology? british journal of clinical pharmacology 75(3) (2013) 645-662. https://doi.org/10.11 11/j.1365-2125.2012.04374.x. [27] u.n. das. essential fatty acids and their metabolites in the pathobiology of inflammation and its resolution. biomolecules 11(12) (2021). https://doi.org/10.3390/biom11121873. [28] y. jang, m. kim, s.w. hwang. molecular mechanisms underlying the actions of arachidonic acidderived prostaglandins on peripheral nociception. journal of neuroinflammation 17(1) (2020) 30. https://doi.org/10.1186/s12974-020-1703-1. [29] b. wang, l. wu, j. chen, l. dong, c. chen, z. wen, j. hu, i. fleming, d.w. wang. metabolism pathways of arachidonic acids: mechanisms and potential therapeutic targets. signal transduction and targeted therapy 6(1) (2021) 94. https://doi.org/10.1038/s41392-020-00443-w. [30] g. ramesh, a.g. maclean, m.t. philipp. cytokines and chemokines at the crossroads of neuroinflammation, neurodegeneration, and neuropathic pain. mediators of inflammation 2013 (2013) 480739. https://doi.org/10.1155/2013/480739. [31] j.m. zhang, j. an. cytokines, inflammation, and pain. international anesthesiology clinics 45(2) (2007) 27-37. https://doi.org/10.1097/aia.0b013e318034194e. [32] m.d. turner, b. nedjai, t. hurst, d.j. pennington. cytokines and chemokines: at the crossroads of cell signalling and inflammatory disease. biochimica et biophysica acta 1843(11) (2014) 2563-2582. https://doi.org/10.1016/j.bbamcr.2014.05.014. [33] i. shachar, n. karin. the dual roles of inflammatory cytokines and chemokines in the regulation of autoimmune diseases and their clinical implications. journal of leukocyte biology 93(1) (2013) 51-61. https://doi.org/10.1189/jlb.0612293. [34] b. kaminska. mapk signalling pathways as molecular targets for anti-inflammatory therapy--from molecular mechanisms to therapeutic benefits. biochimica et biophysica acta 1754(1-2) (2005) 253262. https://doi.org/10.1016/j.bbapap.2005.08.017. https://doi.org/10.3390/molecules16108515 https://doi.org/10.1186/s40560-014-0065-0 https://doi.org/10.4110/in.2018.18.e27 https://doi.org/10.3389/fimmu.2014.00461 https://doi.org/10.1016/j.cell.2010.01.022 https://doi.org/10.3389/fimmu.2012.00323 https://doi.org/10.1128/iai.00095-21 https://doi.org/10.3390/biom11121873 https://doi.org/10.1186/s12974-020-1703-1 https://doi.org/10.1038/s41392-020-00443-w https://doi.org/10.1155/2013/480739 https://doi.org/10.1097/aia.0b013e318034194e https://doi.org/10.1016/j.bbamcr.2014.05.014 https://doi.org/10.1189/jlb.0612293 https://doi.org/10.1016/j.bbapap.2005.08.017 y. ysrafil et al. admet & dmpk 00(0) (2023) 000-000 18 [35] y.l. yang, m. liu, x. cheng, w.h. li, s.s. zhang, y.h. wang, g.h. du. myricitrin blocks activation of nfκb and mapk signaling pathways to protect nigrostriatum neuron in lps-stimulated mice. journal of neuroimmunology 337 (2019) 577049. https://doi.org/10.1016/j.jneuroim.2019.577049. [36] h. yu, l. lin, z. zhang, h. zhang, h. hu. targeting nf-κb pathway for the therapy of diseases: mechanism and clinical study. signal transduction and targeted therapy 5(1) (2020) 209. https://doi.org/10.1038/s41392-020-00312-6. [37] t.y. wang, q. li, k.s. bi. bioactive flavonoids in medicinal plants: structure, activity and biological fate. asian journal of pharmaceutical sciences 13(1) (2018) 12-23. https://doi.org/10.1016/ j.ajps.2017.08.004. [38] c. di lorenzo, f. colombo, s. biella, c. stockley, p. restani. polyphenols and human health: the role of bioavailability. nutrients 13(1) (2021). https://doi.org/10.3390/nu13010273. [39] s. kumar, a.k. pandey. chemistry and biological activities of flavonoids: an overview. scientific world journal 2013 (2013) 162750. https://doi.org/10.1155/2013/162750. [40] f. dajas, a.c. andrés, a. florencia, e. carolina, r.m. felicia. neuroprotective actions of flavones and flavonols: mechanisms and relationship to flavonoid structural features. central nervous system agents in medicinal chemistry 13(1) (2013) 30-35. https://doi.org/10.2174/1871524911313010005. [41] g.l. hostetler, r.a. ralston, s.j. schwartz. flavones: food sources, bioavailability, metabolism, and bioactivity. advances in nutrition 8(3) (2017) 423-435. https://doi.org/10.3945/an.116.012948. [42] n. zhang, f. bi, f. xu, h. yong, y. bao, c. jin, j. liu. structure and functional properties of active packaging films prepared by incorporating different flavonols into chitosan based matrix. international journal of biological macromolecules 165(pt a) (2020) 625-634. https://doi.org/ 10.1016/j.ijbiomac.2020.09.209. [43] d. barreca, g. gattuso, e. bellocco, a. calderaro, d. trombetta, a. smeriglio, g. laganà, m. daglia, s. meneghini, s.m. nabavi. flavanones: citrus phytochemical with health-promoting properties. biofactors 43(4) (2017) 495-506. https://doi.org/10.1002/biof.1363. [44] r.k. singla, a.k. dubey, a. garg, r.k. sharma, m. fiorino, s.m. ameen, m.a. haddad, m. al-hiary. natural polyphenols: chemical classification, definition of classes, subcategories, and structures. journal of aoac international 102(5) (2019) 1397-1400. https://doi.org/10.5740/jaoacint.19-0133. [45] k.p. ko. isoflavones: chemistry, analysis, functions and effects on health and cancer. asian pacific journal of cancer prevention 15(17) (2014) 7001-7010. https://doi.org/10.7314/apjcp.2014.15.17.70 01. [46] n. aziz, m.y. kim, j.y. cho. anti-inflammatory effects of luteolin: a review of in vitro, in vivo, and in silico studies. journal of ethnopharmacology (2018) 342-358. https://doi.org/10.1016/j.jep.20 18.05.019. [47] d. barreca, g. mandalari, a. calderaro, a. smeriglio, d. trombetta, m.r. felice, g. gattuso. citrus flavones: an update on sources, biological functions, and health promoting properties. plants 9(3) (2020). https://doi.org/10.3390/plants9030288. [48] s.f. nabavi, n. braidy, o. gortzi, e. sobarzo-sanchez, m. daglia, k. skalicka-woźniak, s.m. nabavi. luteolin as an anti-inflammatory and neuroprotective agent: a brief review. brain research bulletin 119(pt a) (2015) 1-11. https://doi.org/10.1016/j.brainresbull.2015.09.002. [49] b. salehi, a. venditti, m. sharifi-rad, d. kręgiel, j. sharifi-rad, a. durazzo, m. lucarini, a. santini, e.b. souto, e. novellino, h. antolak, e. azzini, w.n. setzer, n. martins. the therapeutic potential of apigenin. international journal of molecular medicine 20(6) (2019). https://doi.org/10.3390/ ijms20061305. [50] s. singh, p. gupta, a. meena, s. luqman. acacetin, a flavone with diverse therapeutic potential in cancer, inflammation, infections and other metabolic disorders. food and chemical toxicology 145 (2020) 111708. https://doi.org/10.1016/j.fct.2020.111708. [51] w. alam, h. khan, m.a. shah, o. cauli, l. saso. kaempferol as a dietary anti-inflammatory agent: current therapeutic standing. molecules 25(18) (2020). https://doi.org/10.3390/molecules 25184073. https://doi.org/10.1016/j.jneuroim.2019.577049 https://doi.org/10.1038/s41392-020-00312-6 https://doi.org/10.1016/j.ajps.2017.08.004 https://doi.org/10.1016/j.ajps.2017.08.004 https://doi.org/10.3390/nu13010273 https://doi.org/10.1155/2013/162750 https://doi.org/10.2174/1871524911313010005 https://doi.org/10.3945/an.116.012948 https://doi.org/10.1002/biof.1363 https://doi.org/10.5740/jaoacint.19-0133 https://doi.org/10.3390/plants9030288 https://doi.org/10.1016/j.brainresbull.2015.09.002 https://doi.org/10.3390/ijms20061305 https://doi.org/10.3390/ijms20061305 https://doi.org/10.1016/j.fct.2020.111708 admet & dmpk 00(0) (2023) 000-000 anti-inflammatory activities of flavonoid derivates doi: https://doi.org/10.5599/admet.1918 19 [52] k.p. devi, d.s. malar, s.f. nabavi, a. sureda, j. xiao, s.m. nabavi, m. daglia. kaempferol and inflammation: from chemistry to medicine. pharmacological research 99 (2015) 1-10. https://doi.org/10.1016/j.phrs.2015.05.002. [53] h.n. lee, s.a. shin, g.s. choo, h.j. kim, y.s. park, b.s. kim, s.k. kim, s.d. cho, j.s. nam, c.s. choi, j.h. che, b.k. park, j.y. jung. anti‑inflammatory effect of quercetin and galangin in lps‑stimulated raw264.7 macrophages and dncb‑induced atopic dermatitis animal models. international journal of molecular medicine 41(2) (2018) 888-898. https://doi.org/10.3892/ijmm.2017.3296. [54] r.t. magar, j.k. sohng. a review on structure, modifications and structure-activity relation of quercetin and its derivatives. journal of microbiology and biotechnology 30(1) (2020) 11-20. https://doi.org/10.4014/jmb.1907.07003. [55] x. song, l. tan, m. wang, c. ren, c. guo, b. yang, y. ren, z. cao, y. li, j. pei. myricetin: a review of the most recent research. biomedicine & pharmacotherapy 134 (2021) 111017. https://doi.org/10.1016/j.biopha.2020.111017. [56] z. song, m.k. shanmugam, h. yu, g. sethi. butein and its role in chronic diseases. advances in experimental medicine and biology 928(2016) 419-433. https://doi.org/10.1007/978-3-319-413341_17. [57] t. ohishi, s. goto, p. monira, m. isemura, y. nakamura. anti-inflammatory action of green tea. antiinflammatory & anti-allergy agents in medicinal chemistry 15(2) (2016) 74-90. https://doi.org/10.2174/1871523015666160915154443. [58] r. wen, h. lv, y. jiang, p. tu. anti-inflammatory flavanones and flavanols from the roots of pongamia pinnata. planta medica 84(16) (2018) 1174-1182. https://doi.org/10.1055/a-0626-7356. [59] l. křížová, k. dadáková, j. kašparovská, t. kašparovský. isoflavones. molecules 24(6) (2019). https://doi.org/10.3390/molecules24061076. [60] y.m. lee, y. yoon, h. yoon, h.m. park, s. song, k.j. yeum. dietary anthocyanins against obesity and inflammation. nutrients 9(10) (2017). https://doi.org/10.3390/nu9101089. [61] s.h. thilakarathna, h.p. rupasinghe. flavonoid bioavailability and attempts for bioavailability enhancement. nutrients 5(9) (2013) 3367-3387. https://doi.org/10.3390/nu5093367. [62] h.l. yang, s.c. chen, k.j. senthil kumar, k.n. yu, p.d. lee chao, s.y. tsai, y.c. hou, y.c. hseu. antioxidant and anti-inflammatory potential of hesperetin metabolites obtained from hesperetinadministered rat serum: an ex vivo approach. journal of agricultural and food chemistry 60(1) (2012) 522-532. https://doi.org/10.1021/jf2040675. [63] y. fang, w. cao, m. xia, s. pan, x. xu. study of structure and permeability relationship of flavonoids in caco-2 cells. nutrients 9(12) (2017). https://doi.org/10.3390/nu9121301. [64] s. li, j. liu, z. li, l. wang, w. gao, z. zhang, c. guo. sodium-dependent glucose transporter 1 and glucose transporter 2 mediate intestinal transport of quercetrin in caco-2 cells. food & nutrition research 64 (2020). https://doi.org/10.29219/fnr.v64.3745. [65] p.v. röder, k.e. geillinger, t.s. zietek, b. thorens, h. koepsell, h. daniel. the role of sglt1 and glut2 in intestinal glucose transport and sensing. plos one 9(2) (2014) e89977. https://doi.org/10.1371/ journal.pone.0089977. [66] p.c.h. hollman. absorption, bioavailability, and metabolism of flavonoids. pharmaceutical biology 42(sup1) (2009) 74-83. https://doi.org/10.3109/13880200490893492. [67] y. zhang, g. zeng, h. pan, c. li, y. hu, k. chu, w. han, z. chen, r. tang, w. yin, x. chen, y. hu, x. liu, c. jiang, j. li, m. yang, y. song, x. wang, q. gao, f. zhu. safety, tolerability, and immunogenicity of an inactivated sars-cov-2 vaccine in healthy adults aged 18-59 years: a randomised, double-blind, placebo-controlled, phase 1/2 clinical trial. the lancet infectious diseases 21(2) (2021) 181-192. https://doi.org/10.1016/s1473-3099(20)30843-4. [68] j. cao, y. zhang, w. chen, x. zhao. the relationship between fasting plasma concentrations of selected flavonoids and their ordinary dietary intake. the british journal of nutrition 103(2) (2010) 249-255. https://doi.org/10.1017/s000711450999170x. https://doi.org/10.1016/j.phrs.2015.05.002 https://doi.org/10.3892/ijmm.2017.3296 https://doi.org/10.4014/jmb.1907.07003 https://doi.org/10.1016/j.biopha.2020.111017 https://doi.org/10.1007/978-3-319-41334-1_17 https://doi.org/10.1007/978-3-319-41334-1_17 https://doi.org/10.2174/1871523015666160915154443 https://doi.org/10.1055/a-0626-7356 https://doi.org/10.3390/molecules24061076 https://doi.org/10.3390/nu9101089 https://doi.org/10.3390/nu5093367 https://doi.org/10.1021/jf2040675 https://doi.org/10.3390/nu9121301 https://doi.org/10.29219/fnr.v64.3745 https://doi.org/10.3109/13880200490893492 https://doi.org/10.1016/s1473-3099(20)30843-4 https://doi.org/10.1017/s000711450999170x y. ysrafil et al. admet & dmpk 00(0) (2023) 000-000 20 [69] g.b. gonzales, g. smagghe, c. grootaert, m. zotti, k. raes, j. van camp. flavonoid interactions during digestion, absorption, distribution and metabolism: a sequential structure-activity/property relationship-based approach in the study of bioavailability and bioactivity. drug metabolism reviews 47(2) (2015) 175-190. https://doi.org/10.3109/03602532.2014.1003649. [70] i.h. gecibesler, m. aydin. plasma protein binding of herbal-flavonoids to human serum albumin and their anti-proliferative activities. biochimica et biophysica acta general subjects 92(1) (2020) e20190819. https://doi.org/10.1590/0001-3765202020190819. [71] d. barreca, g. laganà, g. toscano, p. calandra, m.a. kiselev, d. lombardo, e. bellocco. the interaction and binding of flavonoids to human serum albumin modify its conformation, stability and resistance against aggregation and oxidative injuries. biochim biophys acta gen subj 1861(1 pt b) (2017) 35313539. https://doi.org/10.1016/j.bbagen.2016.03.014. [72] c.z. lin, m. hu, a.z. wu, c.c. zhu. investigation on the differences of four flavonoids with similar structure binding to human serum albumin. journal of pharmaceutical analysis 4(6) (2014) 392-398. https://doi.org/10.1016/j.jpha.2014.04.001. [73] b. wang, q. qin, m. chang, s. li, x. shi, g. xu. molecular interaction study of flavonoids with human serum albumin using native mass spectrometry and molecular modeling. analytical and bioanalytical chemistry 410(3) (2018) 827-837. https://doi.org/10.1007/s00216-017-0564-7. [74] v.c. de boer, a.a. dihal, h. van der woude, i.c. arts, s. wolffram, g.m. alink, i.m. rietjens, j. keijer, p.c. hollman. tissue distribution of quercetin in rats and pigs. the journal of nutrition 135(7) (2005) 1718-1725. https://doi.org/10.1093/jn/135.7.1718. [75] j. bieger, r. cermak, r. blank, v.c. de boer, p.c. hollman, j. kamphues, s. wolffram. tissue distribution of quercetin in pigs after long-term dietary supplementation. the journal of nutrition 138(8) (2008) 1417-1420. https://doi.org/10.1093/jn/138.8.1417. [76] r.k. al-ishaq, a. liskova, p. kubatka, d. büsselberg. enzymatic metabolism of flavonoids by gut microbiota and its impact on gastrointestinal cancer. cancers 13(16) (2021). https://doi.org/ 10.3390/cancers13163934. [77] k. murota, y. nakamura, m. uehara. flavonoid metabolism: the interaction of metabolites and gut microbiota. bioscience, biotechnology, and biochemistry 82(4) (2018) 600-610. https://doi.org/ 10.1080/09168451.2018.1444467. [78] r. pei, x. liu, b. bolling. flavonoids and gut health. current opinion in biotechnology 61 (2020) 153159. https://doi.org/10.1016/j.copbio.2019.12.018. [79] z. chen, m. chen, h. pan, s. sun, l. li, s. zeng, h. jiang. role of catechol-o-methyltransferase in the disposition of luteolin in rats. drug metabolism and disposition: the biological fate of chemicals 39(4) (2011) 667-674. https://doi.org/10.1124/dmd.110.037333. [80] z. chen, s. zheng, l. li, h. jiang. metabolism of flavonoids in human: a comprehensive review. . current drug metabolism 15(1) (2014) 48-61. https://doi.org/10.2174/138920021501140218125020. [81] a. gradolatto, m.c. canivenc-lavier, j.p. basly, m.h. siess, c. teyssier. metabolism of apigenin by rat liver phase i and phase ii enzymes and by isolated perfused rat liver. drug metabolism and disposition: the biological fate of chemicals 32(1) (2004) 58-65. https://doi.org/10.1124/ dmd.32.1.58. [82] w. jiang, m. hu. mutual interactions between flavonoids and enzymatic and transporter elements responsible for flavonoid disposition via phase ii metabolic pathways. rsc advances 2(21) (2012) 7948-7963. https://doi.org/10.1039/c2ra01369j. [83] t. akazawa, y. uchida, e. miyauchi, m. tachikawa, s. ohtsuki, t. terasaki. high expression of ugt1a1/1a6 in monkey small intestine: comparison of protein expression levels of cytochromes p450, udp-glucuronosyltransferases, and transporters in small intestine of cynomolgus monkey and human. molecular pharmaceutics 15(1) (2018) 127-140. https://doi.org/10.1021/acs.molphar maceut.7b00772. https://doi.org/10.3109/03602532.2014.1003649 https://doi.org/10.1590/0001-3765202020190819 https://doi.org/10.1016/j.bbagen.2016.03.014 https://doi.org/10.1016/j.jpha.2014.04.001 https://doi.org/10.1007/s00216-017-0564-7 https://doi.org/10.1093/jn/135.7.1718 https://doi.org/10.1093/jn/138.8.1417 https://doi.org/10.3390/cancers13163934 https://doi.org/10.3390/cancers13163934 https://doi.org/10.1016/j.copbio.2019.12.018 https://doi.org/10.1124/dmd.110.037333 https://doi.org/10.2174/138920021501140218125020 https://doi.org/10.1124/dmd.32.1.58 https://doi.org/10.1124/dmd.32.1.58 https://doi.org/10.1039/c2ra01369j admet & dmpk 00(0) (2023) 000-000 anti-inflammatory activities of flavonoid derivates doi: https://doi.org/10.5599/admet.1918 21 [84] x. lv, j. hou, y.l. xia, j. ning, g.y. he, p. wang, g.b. ge, z.l. xiu, l. yang. glucuronidation of bavachinin by human tissues and expressed ugt enzymes: identification of ugt1a1 and ugt1a8 as the major contributing enzymes. drug metabolism and pharmacokinetics 30(5) (2015) 358-365. https://doi.org/10.1016/j.dmpk.2015.07.001. [85] l. bredsdorff, t. obel, c. dethlefsen, a. tjønneland, e.b. schmidt, s.e. rasmussen, k. overvad. urinary flavonoid excretion and risk of acute coronary syndrome in a nested case-control study. the american journal of clinical nutrition 98(1) (2013) 209-216. https://doi.org/10.3945/ajcn.112.046169. [86] c.d. kay. aspects of anthocyanin absorption, metabolism and pharmacokinetics in humans. nutrition research reviews 19(1) (2006) 137-146. https://doi.org/10.1079/nrr2005116. [87] j.p. spencer, m.m. abd-el-mohsen, c. rice-evans. cellular uptake and metabolism of flavonoids and their metabolites: implications for their bioactivity. archives of biochemistry and biophysics 423(1) (2004) 148-161. https://doi.org/10.1016/j.abb.2003.11.010. [88] v.s. giorgi, m.t. peracoli, j.c. peracoli, s.s. witkin, c.f. bannwart-castro. silibinin modulates the nfκb pathway and pro-inflammatory cytokine production by mononuclear cells from preeclamptic women. journal of reproductive immunology 95(1-2) (2012) 67-72. https://doi.org/10.1016/j.jri.20 12.06.004. [89] s.s. syed hussein, m.n. kamarudin, h.a. kadir. (+)-catechin attenuates nf-κb activation through regulation of akt, mapk, and ampk signaling pathways in lps-induced bv-2 microglial cells. the american journal of chinese medicine 43(5) (2015) 927-952. https://doi.org/10.1142/s0192415 x15500548. [90] f.y. fan, l.x. sang, m. jiang. catechins and their therapeutic benefits to inflammatory bowel disease. molecules 22(3) (2017). https://doi.org/10.3390/molecules22030484. [91] a.b. lagha, d. grenier. tea polyphenols inhibit the activation of nf-κb and the secretion of cytokines and matrix metalloproteinases by macrophages stimulated with fusobacterium nucleatum. scientific reports 6 (2016) 34520. https://doi.org/10.1038/srep34520. [92] t. nakanishi, k. mukai, h. yumoto, k. hirao, y. hosokawa, t. matsuo. anti-inflammatory effect of catechin on cultured human dental pulp cells affected by bacteria-derived factors. european journal of oral sciences 118(2) (2010) 145-150. https://doi.org/10.1111/j.1600-0722.2010.00714.x. [93] e. nakano, d. kamei, r. murase, i. taki, k. karasawa, k. fukuhara, s. iwai. anti-inflammatory effects of new catechin derivatives in a hapten-induced mouse contact dermatitis model. european journal of pharmacology 845 (2019) 40-47. https://doi.org/10.1016/j.ejphar.2018.12.036. [94] a. khan, m. ikram, j.r. hahm, m.o. kim. antioxidant and anti-inflammatory effects of citrus flavonoid hesperetin: special focus on neurological disorders. antioxidants 9(7) (2020). https://doi.org/10.3390/antiox9070609. [95] m.f. manchope, r. casagrande, w.a. verri, jr. naringenin: an analgesic and anti-inflammatory citrus flavanone. oncotarget 8(3) (2017) 3766-3767. https://doi.org/10.18632/oncotarget.14084. [96] b. salehi, p.v.t. fokou, m. sharifi-rad, p. zucca, r. pezzani, n. martins, j. sharifi-rad. the therapeutic potential of naringenin: a review of clinical trials. pharmaceuticals 12(1) (2019). https://doi.org/ 10.3390/ph12010011. [97] s.j. tsai, c.s. huang, m.c. mong, w.y. kam, h.y. huang, m.c. yin. anti-inflammatory and antifibrotic effects of naringenin in diabetic mice. journal of agricultural and food chemistry 60(1) (2012) 514521. https://doi.org/10.1021/jf203259h. [98] m.i. amaro, j. rocha, h. vila-real, m. eduardo-figueira, h. mota-filipe, b. sepodes, m.h. ribeiro. anti-inflammatory activity of naringin and the biosynthesised naringenin by naringinase immobilized in microstructured materials in a model of dss-induced colitis in mice. food research international 42(8) (2009) 1010-1017. https://doi.org/10.1016/j.foodres.2009.04.016. [99] r.a. hassan, w.g. hozayen, h.t. abo sree, h.m. al-muzafar, k.a. amin, o.m. ahmed. naringin and hesperidin counteract diclofenac-induced hepatotoxicity in male wistar rats via their antioxidant, anti-inflammatory, and antiapoptotic activities. oxidative medicine and cellular longevity 2021 (2021) 9990091. https://doi.org/10.1155/2021/9990091. https://doi.org/10.1016/j.dmpk.2015.07.001 https://doi.org/10.3945/ajcn.112.046169 https://doi.org/10.1079/nrr2005116 https://doi.org/10.1016/j.abb.2003.11.010 https://doi.org/10.1016/j.jri.2012.06.004 https://doi.org/10.1016/j.jri.2012.06.004 https://doi.org/10.3390/molecules22030484 https://doi.org/10.1038/srep34520 https://doi.org/10.1111/j.1600-0722.2010.00714.x https://doi.org/10.1016/j.ejphar.2018.12.036 https://doi.org/10.3390/antiox9070609 https://doi.org/10.18632/oncotarget.14084 https://doi.org/10.1021/jf203259h https://doi.org/10.1016/j.foodres.2009.04.016 https://doi.org/10.1155/2021/9990091 y. ysrafil et al. admet & dmpk 00(0) (2023) 000-000 22 [100] a.d. kandhare, j. alam, m.v. patil, a. sinha, s.l. bodhankar. wound healing potential of naringin ointment formulation via regulating the expression of inflammatory, apoptotic and growth mediators in experimental rats. pharmaceutical biology 54(3) (2016) 419-432. https://doi.org/10.3109/1388 0209.2015.1038755. [101] a.d. kandhare, p. ghosh, s.l. bodhankar. naringin, a flavanone glycoside, promotes angiogenesis and inhibits endothelial apoptosis through modulation of inflammatory and growth factor expression in diabetic foot ulcer in rats. chemico-biological interactions 219 (2014) 101-112. https://doi.org/10.1016/j.cbi.2014.05.012. [102] s. mohanty, a.k. sahoo, v.b. konkimalla, a. pal, s.c. si. correction to naringin in combination with isothiocyanates as liposomal formulations potentiates the anti-inflammatory activity in different acute and chronic animal models of rheumatoid arthritis. acs omega 6(4) (2021) 3434. https://doi.org/10.1021/acsomega.0c06294. [103] a.s. abuelsaad, g. allam, a.a. al-solumani. hesperidin inhibits inflammatory response induced by aeromonas hydrophila infection and alters cd4+/cd8+ t cell ratio. mediators of inflammation 2014 (2014) 393217. https://doi.org/10.1155/2014/393217. [104] s.a. adefegha, d.b. rosa leal, a.a. olabiyi, g. oboh, l.g. castilhos. hesperidin attenuates inflammation and oxidative damage in pleural exudates and liver of rat model of pleurisy. redox report 22(6) (2017) 563-571. https://doi.org/10.1080/13510002.2017.1344013. [105] c. meng, z. guo, d. li, h. li, j. he, d. wen, b. luo. preventive effect of hesperidin modulates inflammatory responses and antioxidant status following acute myocardial infarction through the expression of ppar‑γ and bcl‑2 in model mice. molecular medicine reports 17(1) (2018) 1261-1268. https://doi.org/10.3892/mmr.2017.7981. [106] h. parhiz, a. roohbakhsh, f. soltani, r. rezaee, m. iranshahi. antioxidant and anti-inflammatory properties of the citrus flavonoids hesperidin and hesperetin: an updated review of their molecular mechanisms and experimental models. phytotherapy research 29(3) (2015) 323-331. https://doi.org/10.1002/ptr.5256. [107] k. tamilselvam, j. nataraj, u. janakiraman, t. manivasagam, m.m. essa. antioxidant and antiinflammatory potential of hesperidin against 1-methyl-4-phenyl-1, 2, 3, 6-tetrahydropyridineinduced experimental parkinson's disease in mice. international journal of nutrition pharmacology neurological diseases 3(3) (2013) 294. https://doi.org/10.4103/2231-0738.114875. [108] s. tejada, s. pinya, m. martorell, x. capó, j.a. tur, a. pons, a. sureda. potential anti-inflammatory effects of hesperidin from the genus citrus. international journal of nutrition pharmacology neurological diseases 25(37) (2018) 4929-4945. https://doi.org/10.2174/0929867324666170718 104412. [109] y.-f. tsai, y.-r. chen, j.-p. chen, y. tang, k.-c. yang. effect of hesperidin on anti-inflammation and cellular antioxidant capacity in hydrogen peroxide-stimulated human articular chondrocytes. process biochemistry 85 (2019) 175-184. https://doi.org/10.1016/j.procbio.2019.07.014. [110] s. xiao, w. liu, j. bi, s. liu, h. zhao, n. gong, d. xing, h. gao, m. gong. anti-inflammatory effect of hesperidin enhances chondrogenesis of human mesenchymal stem cells for cartilage tissue repair. journal of inflammation 15 (2018) 14. https://doi.org/10.1186/s12950-018-0190-y. [111] z. deng, s. hassan, m. rafiq, h. li, y. he, y. cai, x. kang, z. liu, t. yan. pharmacological activity of eriodictyol: the major natural polyphenolic flavanone. evidence-based complementary and alternative medicine 2020 (2020) 6681352. https://doi.org/10.1155/2020/6681352. [112] j.k. lee. anti-inflammatory effects of eriodictyol in lipopolysaccharide-stimulated raw 264.7 murine macrophages. archives of pharmacal research 34(4) (2011) 671-679. https://doi.org/10.1007/s12 272-011-0418-3. [113] y. wang, y. chen, y. chen, b. zhou, x. shan, g. yang. eriodictyol inhibits il-1β-induced inflammatory response in human osteoarthritis chondrocytes. biomedicine & pharmacotherapy 107 (2018) 11281134. https://doi.org/10.1016/j.biopha.2018.08.103. https://doi.org/10.3109/13880209.2015.1038755 https://doi.org/10.3109/13880209.2015.1038755 https://doi.org/10.1016/j.cbi.2014.05.012 https://doi.org/10.1021/acsomega.0c06294 https://doi.org/10.1155/2014/393217 https://doi.org/10.1080/13510002.2017.1344013 https://doi.org/10.3892/mmr.2017.7981 https://doi.org/10.4103/2231-0738.114875 https://doi.org/10.1016/j.procbio.2019.07.014 https://doi.org/10.1186/s12950-018-0190-y https://doi.org/10.1155/2020/6681352 https://doi.org/10.1007/s12272-011-0418-3 https://doi.org/10.1007/s12272-011-0418-3 https://doi.org/10.1016/j.biopha.2018.08.103 admet & dmpk 00(0) (2023) 000-000 anti-inflammatory activities of flavonoid derivates doi: https://doi.org/10.5599/admet.1918 23 [114] g.f. zhu, h.j. guo, y. huang, c.t. wu, x.f. zhang. eriodictyol, a plant flavonoid, attenuates lpsinduced acute lung injury through its antioxidative and anti-inflammatory activity. experimental and therapeutic medicine 10(6) (2015) 2259-2266. https://doi.org/10.3892/etm.2015.2827. [115] m. trappoliere, a. caligiuri, m. schmid, c. bertolani, p. failli, f. vizzutti, e. novo, c. di manzano, f. marra, c. loguercio, m. pinzani. silybin, a component of sylimarin, exerts anti-inflammatory and antifibrogenic effects on human hepatic stellate cells. journal of hepatology 50(6) (2009) 1102-1111. https://doi.org/10.1016/j.jhep.2009.02.023. [116] h.l. yang, x.w. shi. silybin alleviates experimental autoimmune encephalomyelitis by suppressing dendritic cell activation and th17 cell differentiation. frontiers in neurology 12 (2021) 659678. https://doi.org/10.3389/fneur.2021.659678. [117] w. huang, m.l. li, m.y. xia, j.y. shao. fisetin-treatment alleviates airway inflammation through inhbition of myd88/nf-κb signaling pathway. international journal of molecular medicine 42(1) (2018) 208-218. https://doi.org/10.3892/ijmm.2018.3582. [118] i.m.n. molagoda, j. jayasingha, y.h. choi, r. jayasooriya, c.h. kang, g.y. kim. fisetin inhibits lipopolysaccharide-induced inflammatory response by activating β-catenin, leading to a decrease in endotoxic shock. scientific reports 11(1) (2021) 8377. https://doi.org/10.1038/s41598-021-87257-0. [119] s. hosseinpour-niazi, p. mirmiran, a. fallah-ghohroudi, f. azizi. non-soya legume-based therapeutic lifestyle change diet reduces inflammatory status in diabetic patients: a randomised cross-over clinical trial. the british journal of nutrition 114(2) (2015) 213-219. https://doi.org/10.1017/ s0007114515001725. [120] o. kadioglu, j. nass, m.e. saeed, b. schuler, t. efferth. kaempferol is an anti-inflammatory compound with activity towards nf-κb pathway proteins. anticancer research 35(5) (2015) 2645-2650. [121] m.r. khazdair, a. anaeigoudari, g.a. agbor. anti-viral and anti-inflammatory effects of kaempferol and quercetin and covid-2019: a scoping review. asian pacific journal of tropical biomedicine 11(8) (2021) 327-334. https://doi.org/10.4103/2221-1691.319567. [122] s.h. kim, j.g. park, j. lee, w.s. yang, g.w. park, h.g. kim, y.s. yi, k.s. baek, n.y. sung, m.j. hossen, m.n. lee, j.h. kim, j.y. cho. the dietary flavonoid kaempferol mediates anti-inflammatory responses via the src, syk, irak1, and irak4 molecular targets. mediators of inflammation 2015((2015) 904142. https://doi.org/10.1155/2015/904142. [123] c.l. saw, y. guo, a.y. yang, x. paredes-gonzalez, c. ramirez, d. pung, a.n. kong. the berry constituents quercetin, kaempferol, and pterostilbene synergistically attenuate reactive oxygen species: involvement of the nrf2-are signaling pathway. food and chemical toxicology 72((2014) 303-311. https://doi.org/10.1016/j.fct.2014.07.038. [124] x.l. tang, j.x. liu, w. dong, p. li, l. li, j.c. hou, y.q. zheng, c.r. lin, j.g. ren. protective effect of kaempferol on lps plus atp-induced inflammatory response in cardiac fibroblasts. inflammation 38(1) (2015) 94-101. https://doi.org/10.1007/s10753-014-0011-2. [125] b.o. cho, h.h. yin, s.h. park, e.b. byun, h.y. ha, s.i. jang. anti-inflammatory activity of myricetin from diospyros lotus through suppression of nf-κb and stat1 activation and nrf2-mediated ho-1 induction in lipopolysaccharide-stimulated raw264.7 macrophages. bioscience, biotechnology, and biochemistry 80(8) (2016) 1520-1530. https://doi.org/10.1080/09168451.2016.1171697. [126] h. lee da, c.s. lee. flavonoid myricetin inhibits tnf-α-stimulated production of inflammatory mediators by suppressing the akt, mtor and nf-κb pathways in human keratinocytes. european journal of pharmacology 784 (2016) 164-172. https://doi.org/10.1016/j.ejphar.2016.05.025. [127] b. wang, d. hao, z. zhang, w. gao, h. pan, y. xiao, b. he, l. kong. inhibition effects of a natural inhibitor on rankl downstream cellular signalling cascades cross-talking. journal of cellular and molecular medicine 22(9) (2018) 4236-4242. https://doi.org/10.1111/jcmm.13703. [128] j.i. dower, j.m. geleijnse, l. gijsbers, c. schalkwijk, d. kromhout, p.c. hollman. supplementation of the pure flavonoids epicatechin and quercetin affects some biomarkers of endothelial dysfunction and inflammation in (pre)hypertensive adults: a randomized double-blind, placebo-controlled, crossover trial. the journal of nutrition 145(7) (2015) 1459-1463. https://doi.org/10.3945/ jn.115.211888. https://doi.org/10.3892/etm.2015.2827 https://doi.org/10.1016/j.jhep.2009.02.023 https://doi.org/10.3389/fneur.2021.659678 https://doi.org/10.3892/ijmm.2018.3582 https://doi.org/10.1038/s41598-021-87257-0 https://doi.org/10.4103/2221-1691.319567 https://doi.org/10.1155/2015/904142 https://doi.org/10.1016/j.fct.2014.07.038 https://doi.org/10.1007/s10753-014-0011-2 https://doi.org/10.1080/09168451.2016.1171697 https://doi.org/10.1016/j.ejphar.2016.05.025 https://doi.org/10.1111/jcmm.13703 https://doi.org/10.3945/jn.115.211888 https://doi.org/10.3945/jn.115.211888 y. ysrafil et al. admet & dmpk 00(0) (2023) 000-000 24 [129] l.a. forney, n.r. lenard, l.k. stewart, t.m. henagan. dietary quercetin attenuates adipose tissue expansion and inflammation and alters adipocyte morphology in a tissue-specific manner. international journal of molecular sciences 19(3) (2018). https://doi.org/10.3390/ijms19030895. [130] f. javadi, a. ahmadzadeh, s. eghtesadi, n. aryaeian, m. zabihiyeganeh, a. rahimi foroushani, s. jazayeri. the effect of quercetin on inflammatory factors and clinical symptoms in women with rheumatoid arthritis: a double-blind, randomized controlled trial. journal of the american college of nutrition 36(1) (2017) 9-15. https://doi.org/10.1080/07315724.2016.1140093. [131] w. lin, w. wang, d. wang, w. ling. quercetin protects against atherosclerosis by inhibiting dendritic cell activation. molecular nutrition & food research 61(9) (2017). https://doi.org/10.1002/mn fr.201700031. [132] a. overman, c.c. chuang, m. mcintosh. quercetin attenuates inflammation in human macrophages and adipocytes exposed to macrophage-conditioned media. international journal of obesity 35(9) (2011) 1165-1172. https://doi.org/10.1038/ijo.2010.272. [133] s. sato, y. mukai. modulation of chronic inflammation by quercetin: the beneficial effects on obesity. journal of inflammation research 13 (2020) 421-431. https://doi.org/10.2147/jir.s228361. [134] a. kim, c.s. lee. apigenin reduces the toll-like receptor-4-dependent activation of nf-κb by suppressing the akt, mtor, jnk, and p38-mapk. naunyn-schmiedeberg's archives of pharmacology 391(3) (2018) 271-283. https://doi.org/10.1007/s00210-017-1454-4. [135] r.h. patil, r.l. babu, m. naveen kumar, k.m. kiran kumar, s.m. hegde, r. nagesh, g.t. ramesh, s.c. sharma. anti-inflammatory effect of apigenin on lps-induced pro-inflammatory mediators and ap-1 factors in human lung epithelial cells. inflammation 39(1) (2016) 138-147. https://doi.org/ 10.1007/s10753-015-0232-z. [136] r. soares, i. azevedo. apigenin: is it a proor anti-inflammatory agent? the american journal of pathology 168(5) (2006) 1762; author reply 1762-1763. https://doi.org/10.2353/ajpath.20 06.060087. [137] j. wang, y.t. liu, l. xiao, l. zhu, q. wang, t. yan. anti-inflammatory effects of apigenin in lipopolysaccharide-induced inflammatory in acute lung injury by suppressing cox-2 and nf-kb pathway. inflammation 37(6) (2014) 2085-2090. https://doi.org/10.1007/s10753-014-9942-x. [138] y.c. wang, k.m. huang. in vitro anti-inflammatory effect of apigenin in the helicobacter pyloriinfected gastric adenocarcinoma cells. food and chemical toxicology 53 (2013) 376-383. https://doi.org/10.1016/j.fct.2012.12.018. [139] q. zhou, h. xu, w. yu, e. li, m. wang. anti-inflammatory effect of an apigenin-maillard reaction product in macrophages and macrophage-endothelial cocultures. oxidative medicine and cellular longevity 2019 (2019) 9026456. https://doi.org/10.1155/2019/9026456. [140] a. ahad, a.a. ganai, m. mujeeb, w.a. siddiqui. chrysin, an anti-inflammatory molecule, abrogates renal dysfunction in type 2 diabetic rats. toxicology and applied pharmacology 279(1) (2014) 1-7. https://doi.org/10.1016/j.taap.2014.05.007. [141] y.h. chang, y.f. chiang, h.y. chen, y.j. huang, k.l. wang, y.h. hong, m. ali, t.m. shieh, s.m. hsia. anti-inflammatory and anti-hyperuricemic effects of chrysin on a high fructose corn syrup-induced hyperuricemia rat model via the amelioration of urate transporters and inhibition of nlrp3 inflammasome signaling pathway. antioxidants 10(4) (2021). https://doi.org/10.3390/anti ox10040564. [142] j.j. ramírez-espinosa, j. saldaña-ríos, s. garcía-jiménez, r. villalobos-molina, g. ávila-villarreal, a.n. rodríguez-ocampo, g. bernal-fernández, s. estrada-soto. chrysin induces antidiabetic, antidyslipidemic and anti-inflammatory effects in athymic nude diabetic mice. molecules 23(1) (2017). https://doi.org/10.3390/molecules23010067. [143] e.y. choi, j.y. jin, j.y. lee, j.i. choi, i.s. choi, s.j. kim. anti-inflammatory effects and the underlying mechanisms of action of daidzein in murine macrophages stimulated with prevotella intermedia lipopolysaccharide. journal of periodontal research 47(2) (2012) 204-211. https://doi.org/10.11 11/j.1600-0765.2011.01422.x. https://doi.org/10.3390/ijms19030895 https://doi.org/10.1080/07315724.2016.1140093 https://doi.org/10.1038/ijo.2010.272 https://doi.org/10.2147/jir.s228361 https://doi.org/10.1007/s00210-017-1454-4 https://doi.org/10.2353/ajpath.2006.060087 https://doi.org/10.2353/ajpath.2006.060087 https://doi.org/10.1007/s10753-014-9942-x https://doi.org/10.1016/j.fct.2012.12.018 https://doi.org/10.1155/2019/9026456 https://doi.org/10.1016/j.taap.2014.05.007 https://doi.org/10.3390/molecules23010067 https://doi.org/10.1111/j.1600-0765.2011.01422.x https://doi.org/10.1111/j.1600-0765.2011.01422.x admet & dmpk 00(0) (2023) 000-000 anti-inflammatory activities of flavonoid derivates doi: https://doi.org/10.5599/admet.1918 25 [144] m.h. liu, y.s. lin, s.y. sheu, j.s. sun. anti-inflammatory effects of daidzein on primary astroglial cell culture. nutritional neuroscience 12(3) (2009) 123-134. https://doi.org/10.1179/147683009 x423274. [145] y. peng, y. shi, h. zhang, y. mine, r. tsao. anti-inflammatory and anti-oxidative activities of daidzein and its sulfonic acid ester derivatives. journal of functional foods 35 (2017) 635-640. https://doi.org/10.1016/j.jff.2017.06.027. [146] y. sakamoto, j. kanatsu, m. toh, a. naka, k. kondo, k. iida. the dietary isoflavone daidzein reduces expression of pro-inflammatory genes through pparα/γ and jnk pathways in adipocyte and macrophage co-cultures. plos one 11(2) (2016) e0149676. https://doi.org/10.1371/journal.po ne.0149676. [147] n. buathong, s. poonyachoti, c. deachapunya. anti-inflammatory effect of genistein in human endometrial cell line treatment with endotoxin lipopolysaccharide. journal of the medical association of thailand 99 suppl 8 (2016) s134-s141. [148] y.x. goh, j. jalil, k.w. lam, k. husain, c.m. premakumar. genistein: a review on its anti-inflammatory properties. frontiers in pharmacology 13 (2022) 820969. https://doi.org/10.3389/fphar.2022.82 0969. [149] h. jung, h.-k. kwak, k.t. hwang. antioxidant and antiinflammatory activities of cyanidin-3-glucoside and cyanidin-3-rutinoside in hydrogen peroxide and lipopolysaccharide-treated raw264. 7 cells. food science and biotechnology 23(6) (2014) 2053-2062. https://doi.org/10.1096/fasebj.28.1_sup plement.830.23. [150] w.y. huang, y.m. liu, j. wang, x.n. wang, c.y. li. anti-inflammatory effect of the blueberry anthocyanins malvidin-3-glucoside and malvidin-3-galactoside in endothelial cells. molecules 19(8) (2014) 12827-12841. https://doi.org/10.3390/molecules190812827. [151] t. sogo, n. terahara, a. hisanaga, t. kumamoto, t. yamashiro, s. wu, k. sakao, d.x. hou. antiinflammatory activity and molecular mechanism of delphinidin 3-sambubioside, a hibiscus anthocyanin. biofactors 41(1) (2015) 58-65. https://doi.org/10.1002/biof.1201. [152] s.w. min, s.n. ryu, d.h. kim. anti-inflammatory effects of black rice, cyanidin-3-o-beta-d-glycoside, and its metabolites, cyanidin and protocatechuic acid. international immunopharmacology 10(8) (2010) 959-966. https://doi.org/10.1016/j.intimp.2010.05.009. [153] d.q. masheta, s.k. al-azzawi, iop conference series: materials science and engineering, 2018, p. 012061 1757-899x. https://doi.org/10.1088/1757-899x/454/1/012061. [154] a. bastin, a. sadeghi, m.h. nematollahi, m. abolhassani, a. mohammadi, h. akbari. the effects of malvidin on oxidative stress parameters and inflammatory cytokines in lps-induced human thp-1 cells. journal of cellular physiology 236(4) (2021) 2790-2799. https://doi.org/10.1002/jcp.30049. [155] w.w. tong, c. zhang, t. hong, d.h. liu, c. wang, j. li, x.k. he, w.d. xu. silibinin alleviates inflammation and induces apoptosis in human rheumatoid arthritis fibroblast-like synoviocytes and has a therapeutic effect on arthritis in rats. scientific reports 8(1) (2018) 3241. https://doi.org/ 10.1038/s41598-018-21674-6. [156] r. lim, c.j. morwood, g. barker, m. lappas. effect of silibinin in reducing inflammatory pathways in in vitro and in vivo models of infection-induced preterm birth. plos one 9(3) (2014) e92505. https://doi.org/10.1371/journal.pone.0092505. [157] j. chen, d.l. li, l.n. xie, y.r. ma, p.p. wu, c. li, w.f. liu, k. zhang, r.p. zhou, x.t. xu, x. zheng, x. liu. synergistic anti-inflammatory effects of silibinin and thymol combination on lps-induced raw264.7 cells by inhibition of nf-κb and mapk activation. phytomedicine 78 (2020) 153309. https://doi.org/10.1016/j.phymed.2020.153309. [158] k. raina, c. agarwal, r. agarwal. effect of silibinin in human colorectal cancer cells: targeting the activation of nf-κb signaling. molecular carcinogenesis 52(3) (2013) 195-206. https://doi.org/ 10.1002/mc.21843. https://doi.org/10.1016/j.jff.2017.06.027 https://doi.org/10.3390/molecules190812827 https://doi.org/10.1002/biof.1201 https://doi.org/10.1016/j.intimp.2010.05.009 https://doi.org/10.1088/1757-899x/454/1/012061 https://doi.org/10.1002/jcp.30049 https://doi.org/10.1038/s41598-018-21674-6 https://doi.org/10.1038/s41598-018-21674-6 https://doi.org/10.1371/journal.pone.0092505 https://doi.org/10.1016/j.phymed.2020.153309 y. ysrafil et al. admet & dmpk 00(0) (2023) 000-000 26 [159] w. zheng, z. feng, y. lou, c. chen, c. zhang, z. tao, h. li, l. cheng, x. ying. silibinin protects against osteoarthritis through inhibiting the inflammatory response and cartilage matrix degradation in vitro and in vivo. oncotarget 8(59) (2017) 99649. https://doi.org/10.18632/oncotarget.20587. [160] c.d. funk. prostaglandins and leukotrienes: advances in eicosanoid biology. science 294(5548) (2001) 1871-1875. https://doi.org/10.1126/science.294.5548.1871. [161] k.m. lee, k.w. lee, s.k. jung, e.j. lee, y.s. heo, a.m. bode, r.a. lubet, h.j. lee, z. dong. kaempferol inhibits uvb-induced cox-2 expression by suppressing src kinase activity. biochemical pharmacology 80(12) (2010) 2042-2049. https://doi.org/10.1016/j.bcp.2010.06.042. [162] r. lópez-posadas, i. ballester, c. mascaraque, m.d. suárez, a. zarzuelo, o. martínez-augustin, f. sánchez de medina. flavonoids exert distinct modulatory actions on cyclooxygenase 2 and nfkappab in an intestinal epithelial cell line (iec18). british journal of pharmacology 160(7) (2010) 17141726. https://doi.org/10.1111/j.1476-5381.2010.00827.x. [163] t. vezza, a. rodríguez-nogales, f. algieri, m.p. utrilla, m.e. rodriguez-cabezas, j. galvez. flavonoids in inflammatory bowel disease: a review. nutrients 8(4) (2016) 211. https://doi.org/10.3390/ nu8040211. [164] d. serra, j. paixão, c. nunes, t.c. dinis, l.m. almeida. cyanidin-3-glucoside suppresses cytokineinduced inflammatory response in human intestinal cells: comparison with 5-aminosalicylic acid. plos one 8(9) (2013) e73001. https://doi.org/10.1371/journal.pone.0073001. [165] t. hussain, s. gupta, v.m. adhami, h. mukhtar. green tea constituent epigallocatechin-3-gallate selectively inhibits cox-2 without affecting cox-1 expression in human prostate carcinoma cells. international journal of cancer 113(4) (2005) 660-669. https://doi.org/10.1002/ijc.20629. [166] j. hong, t.j. smith, c.t. ho, d.a. august, c.s. yang. effects of purified green and black tea polyphenols on cyclooxygenaseand lipoxygenase-dependent metabolism of arachidonic acid in human colon mucosa and colon tumor tissues. biochemical pharmacology 62(9) (2001) 1175-1183. https://doi.org/10.1016/s0006-2952(01)00767-5. [167] j.k. kundu, h.k. na, k.s. chun, y.k. kim, s.j. lee, s.s. lee, o.s. lee, y.c. sim, y.j. surh. inhibition of phorbol ester-induced cox-2 expression by epigallocatechin gallate in mouse skin and cultured human mammary epithelial cells. the journal of nutrition 133(11 suppl 1) (2003) 3805s-3810s. https://doi.org/10.1093/jn/133.11.3805s. [168] g. peng, d.a. dixon, s.j. muga, t.j. smith, m.j. wargovich. green tea polyphenol (-)-epigallocatechin3-gallate inhibits cyclooxygenase-2 expression in colon carcinogenesis. molecular carcinogenesis 45(5) (2006) 309-319. https://doi.org/10.1002/mc.20166. [169] j.h. lee, g.h. kim. evaluation of antioxidant and inhibitory activities for different subclasses flavonoids on enzymes for rheumatoid arthritis. journal of food science 75(7) (2010) h212-217. https://doi.org/10.1111/j.1750-3841.2010.01755.x. [170] j.w. he, l. yang, z.q. mu, y.y. zhu, g.y. zhong, z.y. liu, q.g. zhou, f. cheng. anti-inflammatory and antioxidant activities of flavonoids from the flowers of hosta plantaginea. rsc advances 8(32) (2018) 18175-18179. https://doi.org/10.1039/c8ra00443a. [171] s.h. choi, s. aid, f. bosetti. the distinct roles of cyclooxygenase-1 and -2 in neuroinflammation: implications for translational research. trends in pharmacological sciences 30(4) (2009) 174-181. https://doi.org/10.1016/j.tips.2009.01.002. [172] s.d. brain, t.j. williams. leukotrienes and inflammation. pharmacology & therapeutics 46(1) (1990) 57-66. https://doi.org/10.1016/0163-7258(90)90035-z. [173] j. lättig, m. böhl, p. fischer, s. tischer, c. tietböhl, m. menschikowski, h.o. gutzeit, p. metz, m.t. pisabarro. mechanism of inhibition of human secretory phospholipase a2 by flavonoids: rationale for lead design. journal of computer-aided molecular design 21(8) (2007) 473-483. https://doi.org/10.1007/s10822-007-9129-8. [174] v. thibane, a. ndhlala, j. finnie, j.j.s.a.j.o.b. van staden. modulation of the enzyme activity of secretory phospholipase a2, lipoxygenase and cyclooxygenase involved in inflammation and disease https://doi.org/10.18632/oncotarget.20587 https://doi.org/10.1126/science.294.5548.1871 https://doi.org/10.1016/j.bcp.2010.06.042 https://doi.org/10.1111/j.1476-5381.2010.00827.x https://doi.org/10.1371/journal.pone.0073001 https://doi.org/10.1002/ijc.20629 https://doi.org/10.1016/s0006-2952(01)00767-5 https://doi.org/10.1093/jn/133.11.3805s https://doi.org/10.1002/mc.20166 https://doi.org/10.1111/j.1750-3841.2010.01755.x https://doi.org/10.1039/c8ra00443a https://doi.org/10.1016/j.tips.2009.01.002 https://doi.org/10.1016/0163-7258(90)90035-z https://doi.org/10.1007/s10822-007-9129-8 admet & dmpk 00(0) (2023) 000-000 anti-inflammatory activities of flavonoid derivates doi: https://doi.org/10.5599/admet.1918 27 by extracts from some medicinal plants used for skincare and beauty. south african journal of botany 120 (2019) 198-203. https://doi.org/10.1016/j.sajb.2018.06.001. [175] o.c. enechi, e.s. okeke, o.e. awoh, c.o. okoye, c.k. odo. inhibition of phospholipase a2, platelet aggregation and egg albumin induced rat paw oedema as anti-inflammatory effect of peltophorun pterocarpus stem-bark. clinical phytoscience 7(1) (2021) 1-8. https://doi.org/10.1186/s40816-02100310-3. [176] c.a. cotrim, s.c. de oliveira, e.b. diz filho, f.v. fonseca, l. baldissera, jr., e. antunes, r.m. ximenes, h.s. monteiro, m.m. rabello, m.z. hernandes, d. de oliveira toyama, m.h. toyama. quercetin as an inhibitor of snake venom secretory phospholipase a2. chemico-biological interactions 189(1-2) (2011) 9-16. https://doi.org/10.1016/j.cbi.2010.10.016. [177] x.l. hou, q. tong, w.q. wang, c.y. shi, w. xiong, j. chen, x. liu, j.g. fang. suppression of inflammatory responses by dihydromyricetin, a flavonoid from ampelopsis grossedentata, via inhibiting the activation of nf-κb and mapk signaling pathways. journal of natural products 78(7) (2015) 1689-1696. https://doi.org/10.1021/acs.jnatprod.5b00275. [178] g.l. johnson, r. lapadat. mitogen-activated protein kinase pathways mediated by erk, jnk, and p38 protein kinases. science 298(5600) (2002) 1911-1912. https://doi.org/10.1126/science.1072682. [179] a. xagorari, c. roussos, a. papapetropoulos. inhibition of lps-stimulated pathways in macrophages by the flavonoid luteolin. br j pharmacol 136(7) (2002) 1058-1064. https://doi.org/10.1038/ sj.bjp.0704803. [180] d. ichikawa, a. matsui, m. imai, y. sonoda, t. kasahara. effect of various catechins on the il-12p40 production by murine peritoneal macrophages and a macrophage cell line, j774.1. biological & pharmaceutical bulletin 27(9) (2004) 1353-1358. https://doi.org/10.1248/bpb.27.1353. [181] t. behl, t. rana, g.h. alotaibi, m. shamsuzzaman, m. naqvi, a. sehgal, s. singh, n. sharma, y. almoshari, a.a.h. abdellatif, m.s. iqbal, s. bhatia, a. al-harrasi, s. bungau. polyphenols inhibiting mapk signalling pathway mediated oxidative stress and inflammation in depression. biomedicine & pharmacotherapy 146 (2022) 112545. https://doi.org/10.1016/j.biopha.2021.112545. [182] m. comalada, d. camuesco, s. sierra, i. ballester, j. xaus, j. gálvez, a. zarzuelo. in vivo quercitrin antiinflammatory effect involves release of quercetin, which inhibits inflammation through downregulation of the nf-kappab pathway. european journal of immunology 35(2) (2005) 584-592. https://doi.org/10.1002/eji.200425778. [183] t. behl, t. upadhyay, s. singh, s. chigurupati, a.m. alsubayiel, v. mani, c. vargas-de-la-cruz, d. uivarosan, c. bustea, c. sava, m. stoicescu, a.f. radu, s.g. bungau. polyphenols targeting mapk mediated oxidative stress and inflammation in rheumatoid arthritis. molecules 26(21) (2021). https://doi.org/10.3390/molecules26216570. [184] b.v. bonifácio, p.b. silva, m.a. ramos, k.m. negri, t.m. bauab, m. chorilli. nanotechnology-based drug delivery systems and herbal medicines: a review. int j nanomedicine 9((2014) 1-15. https://doi.org/10.2147/ijn.s52634. [185] r. watkins, l. wu, c. zhang, r.m. davis, b. xu. natural product-based nanomedicine: recent advances and issues. international journal of nanomedicine 10 (2015) 6055-6074. https://doi.org/10.2147/ ijn.s92162. [186] a. george, p.a. shah, p.s. shrivastav. natural biodegradable polymers based nano-formulations for drug delivery: a review. international journal of pharmaceutics 561 (2019) 244-264. https://doi.org/10.1016/j.ijpharm.2019.03.011. [187] s. cherukuri, u.r. batchu, k. mandava, v. cherukuri, k.r. ganapuram. formulation and evaluation of transdermal drug delivery of topiramate. international journal of pharmaceutical investigation 7(1) (2017) 10-17. https://doi.org/10.4103/jphi.jphi_35_16. [188] d.s. shaker, r.a.h. ishak, a. ghoneim, m.a. elhuoni. nanoemulsion: a review on mechanisms for the transdermal delivery of hydrophobic and hydrophilic drugs. scientia pharmaceutica 87(3) (2019) 17. https://doi.org/10.3390/scipharm87030017. https://doi.org/10.1016/j.sajb.2018.06.001 https://doi.org/10.1186/s40816-021-00310-3 https://doi.org/10.1186/s40816-021-00310-3 https://doi.org/10.1016/j.cbi.2010.10.016 https://doi.org/10.1021/acs.jnatprod.5b00275 https://doi.org/10.1126/science.1072682 https://doi.org/10.1038/sj.bjp.0704803 https://doi.org/10.1038/sj.bjp.0704803 https://doi.org/10.1248/bpb.27.1353 https://doi.org/10.1016/j.biopha.2021.112545 https://doi.org/10.1002/eji.200425778 https://doi.org/10.3390/molecules26216570 https://doi.org/10.2147/ijn.s52634 https://doi.org/10.1016/j.ijpharm.2019.03.011 https://doi.org/10.4103/jphi.jphi_35_16 https://doi.org/10.3390/scipharm87030017 y. ysrafil et al. admet & dmpk 00(0) (2023) 000-000 28 [189] m. auffan, j. rose, j.y. bottero, g.v. lowry, j.p. jolivet, m.r. wiesner. towards a definition of inorganic nanoparticles from an environmental, health and safety perspective. nature nanotechnology 4(10) (2009) 634-641. https://doi.org/10.1038/nnano.2009.242. [190] y. ysrafil, i. astuti. chitosan nanoparticle-mediated effect of antimirna-324-5p on decreasing the ovarian cancer cell proliferation by regulation of gli1 expression. bioimpacts 12(3) (2022) 195-202. https://doi.org/10.34172/bi.2021.22119. [191] r. jeetah, a. bhaw-luximon, d. jhurry. nanopharmaceutics: phytochemical-based controlled or sustained drug-delivery systems for cancer treatment. journal of biomedical nanotechnology 10(9) (2014) 1810-1840. https://doi.org/10.1166/jbn.2014.1884. [192] a. mukerjee, j.k. vishwanatha. formulation, characterization and evaluation of curcumin-loaded plga nanospheres for cancer therapy. anticancer research 29(10) (2009) 3867-3875. [193] r.s. mulik, j. mönkkönen, r.o. juvonen, k.r. mahadik, a.r. paradkar. apoptosis-induced anticancer effect of transferrin-conjugated solid lipid nanoparticles of curcumin. cancer nanotechnology 3(1-6) (2012) 65-81. https://doi.org/10.1007/s12645-012-0031-2. [194] j.s. nam, a.r. sharma, l.t. nguyen, c. chakraborty, g. sharma, s.s. lee. application of bioactive quercetin in oncotherapy: from nutrition to nanomedicine. molecules 21(1) (2016) e108. https://doi.org/10.3390/molecules21010108. [195] a.r. sharma, s.k. kundu, j.s. nam, g. sharma, c.g. priya doss, s.s. lee, c. chakraborty. next generation delivery system for proteins and genes of therapeutic purpose: why and how? biomed research international 2014 (2014) 327950. https://doi.org/10.1155/2014/327950. [196] y.c. cheng, t.s. li, h.l. su, p.c. lee, h.d. wang. transdermal delivery systems of natural products applied to skin therapy and care. molecules 25(21) (2020). https://doi.org/10.3390/molecu les25215051. [197] l.d. rizkita, y. ysrafil, r. martien, i. astuti. chitosan nanoparticles mediated delivery of mir-106b-5b to breast cancer cell lines mcf-7 and t47d. international journal of applied pharmaceutics 13(1) (2021) 129-134. https://doi.org/10.22159/ijap.2021v13i1.39749. [198] r.b. suardi, y. ysrafil, s.l. sesotyosari, r. martien, t. wardana, i. astuti, s.m. haryana. the effects of combination of mimic mir-155-5p and antagonist mir-324-5p encapsulated chitosan in ovarian cancer skov3. asian pacific journal of cancer prevention 21(9) (2020) 2603-2608. https://doi.org/ 10.31557/apjcp.2020.21.9.2603. [199] j. pardeike, a. hommoss, r.h. müller. lipid nanoparticles (sln, nlc) in cosmetic and pharmaceutical dermal products. international journal of pharmaceutics 366(1-2) (2009) 170-184. https://doi.org/ 10.1016/j.ijpharm.2008.10.003. [200] s.s. hallan, m. sguizzato, m. drechsler, p. mariani, l. montesi, r. cortesi, s. björklund, t. ruzgas, e. esposito. the potential of caffeic acid lipid nanoparticulate systems for skin application: in vitro assays to assess delivery and antioxidant effect. nanomaterials 11(1) (2021). https://doi.org/ 10.3390/nano11010171. [201] m.e. tan, c.h. he, w. jiang, c. zeng, n. yu, w. huang, z.g. gao, j.g. xing. development of solid lipid nanoparticles containing total flavonoid extract from dracocephalum moldavica l. and their therapeutic effect against myocardial ischemia-reperfusion injury in rats. international journal of nanomedicine 12 (2017) 3253-3265. https://doi.org/10.2147/ijn.s131893. [202] y. ysrafil, i. astuti, s.l. anwar, r. martien, f.a.n. sumadi, t. wardhana, s.m. haryana. microrna-1555p diminishes in vitro ovarian cancer cell viability by targeting hif1α expression. advanced pharmaceutical bulletin 10(4) (2020) 630-637. https://doi.org/10.34172/apb.2020.076. [203] q. wang, x. qin, j. fang, x. sun. nanomedicines for the treatment of rheumatoid arthritis: state of art and potential therapeutic strategies. acta pharmaceutica sinica b 11(5) (2021) 1158-1174. https://doi.org/10.1016/j.apsb.2021.03.013. [204] d. wang, e. mehrabi nasab, s.s. athari. study effect of baicalein encapsulated/loaded chitosannanoparticle on allergic asthma pathology in mouse model. saudi journal of biological sciences 28(8) (2021) 4311-4317. https://doi.org/10.1016/j.sjbs.2021.04.009. https://doi.org/10.1038/nnano.2009.242 https://doi.org/10.34172/bi.2021.22119 https://doi.org/10.1166/jbn.2014.1884 https://doi.org/10.1007/s12645-012-0031-2 https://doi.org/10.3390/molecules21010108 https://doi.org/10.1155/2014/327950 https://doi.org/10.3390/molecules25215051 https://doi.org/10.3390/molecules25215051 https://doi.org/10.22159/ijap.2021v13i1.39749 https://doi.org/10.3390/nano11010171 https://doi.org/10.3390/nano11010171 https://doi.org/10.2147/ijn.s131893 https://doi.org/10.34172/apb.2020.076 https://doi.org/10.1016/j.apsb.2021.03.013 https://doi.org/10.1016/j.sjbs.2021.04.009 admet & dmpk 00(0) (2023) 000-000 anti-inflammatory activities of flavonoid derivates doi: https://doi.org/10.5599/admet.1918 29 [205] w. nan, l. ding, h. chen, f.u. khan, l. yu, x. sui, x. shi. topical use of quercetin-loaded chitosan nanoparticles against ultraviolet b radiation. frontiers in pharmacology 9 (2018) 826. https://doi.org/10.3389/fphar.2018.00826. [206] g. wang, j.j. wang, g.y. yang, s.m. du, n. zeng, d.s. li, r.m. li, j.y. chen, j.b. feng, s.h. yuan, f. ye. effects of quercetin nanoliposomes on c6 glioma cells through induction of type iii programmed cell death. international journal of nanomedicine 7 (2012) 271-280. https://doi.org/10.2147/ijn.s26935. [207] x. jin, q. yang, y. zhang. synergistic apoptotic effects of apigenin tpgs liposomes and tyroservatide: implications for effective treatment of lung cancer. international journal of nanomedicine 12 (2017) 5109-5118. https://doi.org/10.2147/ijn.s140096. [208] e.v. fuior, m. deleanu, c.a. constantinescu, d. rebleanu, g. voicu, m. simionescu, m. calin. functional role of vcam-1 targeted flavonoid-loaded lipid nanoemulsions in reducing endothelium inflammation. pharmaceutics 11(8) (2019). https://doi.org/10.3390/pharmaceutics 11080391. [209] m.s.c. zain, s.l. edirisinghe, c.-h. kim, m. de zoysa, k. shaari. nanoemulsion of flavonoid-enriched oil palm (elaeis guineensis jacq.) leaf extract enhances wound healing in zebrafish. phytomedicine plus 1(4) (2021) 100124. [210] s. nagaraja, g.m. basavarajappa, m. attimarad, s. pund. topical nanoemulgel for the treatment of skin cancer: proof-of-technology. pharmaceutics 13(6) (2021). https://doi.org/10.3390/pharma ceutics13060902. [211] t. ponto, g. latter, g. luna, v.r. leite-silva, a. wright, h.a.e. benson. novel self-nano-emulsifying drug delivery systems containing astaxanthin for topical skin delivery. pharmaceutics 13(5) (2021). https://doi.org/10.3390/pharmaceutics13050649. [212] z. sapiun, a.k. imran, s.t.r. dewi, d.f.m. pade, w. ibrahim, r. tungadi, w.s. abdulkadir, y. banne, y. rifai, s. sartini. formulation and characterization of self nano-emulsifying drug delivery system (snedds) fraction of n-hexane: ethyl acetate from sesewanua leaf (clerodendrum fragrans wild.). international journal of applied pharmaceutics 15(2) (2023) 72-77. https://doi.org/10.22159/ ijap.2023v15i2.46365. [213] o.a. ahmed, s.m. badr-eldin, m.k. tawfik, t.a. ahmed, k.m. el-say, j.m. badr. design and optimization of self-nanoemulsifying delivery system to enhance quercetin hepatoprotective activity in paracetamol-induced hepatotoxicity. journal of pharmaceutical sciences 103(2) (2014) 602-612. https://doi.org/10.1002/jps.23834. [214] m. kazi, a. alhajri, s.m. alshehri, e.m. elzayat, o.t. al meanazel, f. shakeel, o. noman, m.a. altamimi, f.k. alanazi. enhancing oral bioavailability of apigenin using a bioactive self-nanoemulsifying drug delivery system (bio-snedds): in vitro, in vivo and stability evaluations. pharmaceutics 12(8) (2020). https://doi.org/10.3390/pharmaceutics12080749. ©2023 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) https://doi.org/10.3389/fphar.2018.00826 https://doi.org/10.2147/ijn.s26935 https://doi.org/10.2147/ijn.s140096 https://doi.org/10.3390/pharmaceutics13060902 https://doi.org/10.3390/pharmaceutics13060902 https://doi.org/10.3390/pharmaceutics13050649 https://doi.org/10.3390/pharmaceutics12080749 http://creativecommons.org/licenses/by/3.0/ infections associated with sars-cov-2 exploited via nanoformulated photodynamic therapy doi: https://doi.org/10.5599/admet.1883 1 admet & dmpk 0(0) (2023) 000-000; doi: https://doi.org/10.5599/admet.1883 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index review infections associated with sars-cov-2 exploited via nanoformulated photodynamic therapy pragya pallavi1,†, karthick harini1,†, noureddine elboughdiri2,3, pemula gowtham1, koyeli girigoswami1, agnishwar girigoswami1,* 1medical bionanotechnology, faculty of allied health sciences, chettinad hospital and research institute (chri), chettinad academy of research and education (care), kelambakkam, chennai, tn-603103, india 2chemical engineering department, college of engineering, university of ha'il, p.o. box 2440, ha'il 81441, saudi arabia 3chemical engineering process department, national school of engineers gabes, university of gabes, gabes 6029, tunisia †equal contribution *corresponding author: e-mail: agnishwarg@gmail.com; tel. +91-9445 268 615 received: may 08, 2023; revised: june 30, 2023; published: june xx,2023 abstract background and purpose: the pandemic of covid-19 has highlighted the need for managing infectious diseases, which spreads by airborne transmission leading to serious health, social, and economic issues. sars-cov-2 is an enveloped virus with a 60–140 nm diameter and particle-like features, which majorly accounts for this disease. expanding diagnostic capabilities, developing safe vaccinations with long-lasting immunity, and formulating effective medications are the strategies to be investigated. experimental approach: for the literature search, electronic databases such as scopus, google scholar, medline, embase, pubmed, and web of science were used as the source. search terms like 'nano-mediated pdt,' 'pdt for sars-cov-2', and 'nanotechnology in treatment for sars-cov-2' were used. out of 275 initially selected articles, 198 were chosen after the abstract screening. during the full-text screening, 80 papers were excluded, and 18 were eliminated during data extraction. preference was given to articles published from 2018 onwards, but a few older references were cited for their valuable information. key results: synthetic nanoparticles (nps) have a close structural resemblance to viruses and interact greatly with their proteins due to their similarities in the configurations. nps had previously been reported to be effective against a variety of viruses. in this way, with nanoparticles, photodynamic therapy (pdt) can be a viable alternative to antibiotics in fighting against microbial infections. the protocol of pdt includes the activation of photosensitizers using specific light to destroy microorganisms in the presence of oxygen, treating several respiratory diseases. conclusion: the use of pdt in treating covid-19 requires intensive investigations, which has been reviewed in this manuscript, including a computational approach to formulating effective photosensitizers. ©2023 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords pdt; covid-19; photosensitizers; vaccines; drug delivery introduction https://doi.org/10.5599/admet.1883 https://doi.org/10.5599/admet.1883 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:agnishwarg@gmail.com http://creativecommons.org/licenses/by/4.0/ p. pallavi et al. admet & dmpk 00(0) (2023) 000-000 2 sars is an airborne viral infection that spreads through mild droplets of saliva, much like the influenza virus and the common cold [1,2]. the outbreak of covid-19 (coronavirus disease-2019) has transformed into a global pandemic [3]. the name of the virus causing covid-19 is given by the international committee on taxonomy of viruses (ictv) [4,5]. this became the first novel virus to emerge in the twenty-first century as both a severe and easily transmissible infection. the sars-cov-2 (severe acute respiratory syndrome coronavirus 2) virus is a single-stranded rna virus that contains 29,903 nucleotides with a positive sense [6,7]. the transmission of these viruses was within animals until the end of 2019. as the seafood market in wuhan (china), known for selling live animals like frogs, bats, birds, and snakes, has sold a virally infected organism, the infection of covid-19 was transmitted to humans [8]. several terrible disasters occurred throughout this phase. china's national health commission has provided information about the pandemic, which has been linked to viral pneumonia. eventually, a new coronavirus (sars-cov2) was discovered following on study based on affected individuals and genetic sequence research. furthermore, with the discovery of genetic similarities between a new coronavirus and sars-like bat viruses, researchers hypothesized that bats might be the sole reservoirs [9-11]. in order to design preventive tactics to fight an infection, the origin and transmission channels must be recognized. the majorities of infected people who are identified with covid-19 begin the viral process three days before the beginning of the symptoms and can last up to a week after the onset of symptoms. incubation time varies from 2-14 days in individuals who become symptomatic [12]. the overall symptoms, means of communication, diagnosis, and therapy, including prevention, are represented in figure 1. figure 1. overview of the infection and theranostic approach against covid-19. fever or chills, cough, sore throat, nausea, rhinorrhea, diarrhea, weariness, exhaustion, headache, dyspnea, myalgia, dysgeusia, or anosmia are the prevalent symptoms, especially if the patients have been exposed recently [13,14]. in the majority of patients, the sickness is self-limiting, but in some cases, patients require hospitalization due to hypoxia, which is more common among those with concurrent conditions. acute respiratory distress syndrome affects around 5 % of the population, and the infected symptomatic patient requires oxygen supplement intubation and other invasive treatment. currently, there are many studies aiming at the development of vaccines and antiviral medicines for covid-19. there are some antiviral drugs being prescribed for treating the covid-19 infection, including remdesivir, steroids, tocilizumab, favipiravir, ivermectin, lopinavir admet & dmpk 00(0) (2023) 000-000 infections associated with sars-cov-2 exploited via photodynamic therapy doi: https://doi.org/10.5599/admet.1883 3 /ritonavir [15-17]. studies revealed that in comparison to other antiviral drugs, remdesivir showed a better therapeutic effect. a summary of the current treatment options and clinical features of covid-19 is provided here, along with the role and progress of nanotechnology-based photodynamic therapy in treating infections. methods a comprehensive search initially yielded a total number of 275 research articles related to the topic, while some specific keywords like 'nano-mediated pdt', 'pdt for sars-cov-2', and 'nanotechnology in treatment for sars-cov-2' were placed in the online database search. popular electronic databases such as scopus, google scholar, medline, embase, pubmed, and web of science were used to collect the articles. instead of merely using pdt as a keyword, pdt was sometimes referred to as 'photodynamic therapy'. after conducting an abstract screening, 198 articles were selected for further consideration. during the full-text screening, 80 articles were excludeed leaving a subset of relevant articles. additionally, 18 more papers were deemed unsuitable for inclusion in the manuscript during the data extraction process. while prioritizing articles published from 2018 onwards, a few older references were included to furnish the basic concepts utilizing their valuable and unchanging particulars. vaccines vaccine development is a difficult, time-consuming, and expensive procedure. trained professionals are required with multiple procedures, pauses for inspections, and data analysis to develop licensed work. the development of an ideal vaccine needs technical flexibility and adaptability with large-scale manufacturing, high purity, easy transportation, and storage. the foremost considerations while designing a vaccine are the stability, route of administration, and adverse effects. the present efforts are to create a vaccination against covid-19 as quickly as possible. the most challenging aspect of vaccine designing is identifying the causative agent, wherein the incidence of covid-19 was not found quickly after the outbreak began [18]. based on the knowledge gathered from past epidemics of corona-mediated viral infections, specifically from sars-cov and mers-cov, many international funding organizations for vaccines and immunization are sponsoring to support and continue the creative attempts to design a vaccine. the protein contents of sars-cov-2, as well as their mode of action, had already been recognized. sars-cov-2 is a positive sensed and single-stranded rna virus [19]. proteins e, s, m, and n are some structural and functional protein that surrounds the sars-cov-2 virus [20,21]. the current approaches for designing vaccines constitute viral vector-based vaccines, virus-like protein particles (vlp), viral protein s subunit, and novel non-viral nucleic acid (rna and dna) based vaccines [22,23]. due to the previous research in 2003 for developing a vaccine against sars-cov by targeting the s subunit, vaccination against sars-cov-2 went straight into clinical researchhuman trial with the previous results [24,25]. sars-coronavirus possess unique rna proofreading capabilities by 3'5' exonuclease activities of nsp12 and nsp14 [26]. this decides the two major characteristics of the virus: (1) high replication efficiency and (2) minimal new mutation rate and estimated to be around 2x105. due to the unusual faults during viral genome replication, it may sometimes result in the formation of variations, which greatly impacts the virus infectivity, causes antigenic drift, and alters the host's response towards the viral antigen. this antigen drift could hinder the construction of a vaccine that is active for a longer duration. single amino acid mutation in the s protein, like d614g, has emerged as the widespread genotype in the present pandemic. according to preliminary studies, the d614g variation does not affect the effectiveness of vaccines targeting the s protein, as this variation has no impact on the antigenic characteristics of the s protein [27,28]. dna-based vaccine dna-based covid-19 vaccine is a defining attempt for nucleic acid technology, mostly disregarded throughout the pandemic. a novel covid-19 vaccine entices the immune system against the sars-cov-2 https://doi.org/10.5599/admet.1883 p. pallavi et al. admet & dmpk 00(0) (2023) 000-000 4 virus by using circular strands of dna [29]. researchers have hailed the clearance of dna vaccines for humans. dna vaccines have a number of attractive properties that include easy production with excellent quality, cost-effectiveness, high safety concerns, and stability [30]. dna vaccines can be delivered via intramuscular or intradermal inoculation, as well as electroporation. the immunogenicity and tolerance rate of dna vaccines were found to be huge; thus, they are being developed to combat various infections. researchers developed a dna vaccine in order to decrease the level of viral rna, which was tested against a sars-cov-2 immunized monkey model and shown to exhibit an effective immune response [31,32]. similarly, recombinant adenovirus expressing whole spike protein vaccine was developed and tested. many vaccinated candidates displayed t-cell responses and dose-dependent antibody activity. the adverse effects were moderate, and no severe complications were reported [33,34]. mrna-based vaccine the mrna vaccine largely varies from conventional vaccines, where it stimulates the host immune system by using any one of two major components: an inactivated organism or an antigen, which is the protein of that organism [35]. an rna polymerase was used to transcribe an orf-containing mrna from a dna template. the advantages of mrna-based vaccinations include simplicity, safety, and scalability of laboratory manufacturing, depending on the translational machinery of the host, and loss of integration in the genome [36]. however, there is a significant barrier to successful delivery, including the stability of mrna during storage, the stability inside the host, an unpredictable immune response, and storage instability until frozen. vaccination with mrna is considered a relatively productive and time-saving option against covid-19. as a result, rna has emerged as the key component in the development of vaccines against covid-19. spike protein is the prime target for mrna vaccines, as the structural aspect is similar to sars-cov and sars-cov-2 [37,38]. neutralizing activity of antibodies against the n-terminal and receptor binding domain (rbd) of s protein was studied in covid-19 patients and proved that these are the components that hinder their use as the target for vaccine development [39,40]. the major problem is the entry route of sars-cov-2 into the cell with a mechanism that is not dependent on ace2. rna vaccine has some more challenges like the stability of mrna, absorption efficiency by the host cells, and release rate into the cell cytoplasm to target the protein [41]. nanoparticle used for vaccine rna and dna vaccines can generate immunity against a specific disease while reducing the risk of infection. the delivery of these immunogens to the site of action, together with other immunogens, as part of a vaccination regimen is a major difficulty. the limitations can be resolved by employing an efficient delivery method to deliver the vaccine to the target location and the adjuvant while preserving it against deterioration in a hostile environment. the delivery mechanism should have an immunogenic effect that lasts without causing any negative effects. nanoparticles based delivery systems are popular nowadays for theranostic applications [42-45]. nano-delivery methods might meet the criteria and allow for the long-term sustained release of vaccine molecules without being damaged by proteases [46,47]. the use of nanocarriers for vaccine components improves cellular absorption, resulting in enhanced innate, humoral, cellular, and mucosal immune responses [48,49]. nanoparticle vaccines are powerful, safe, and simple to make. in comparison to vaccinations that retain parts of the virus, attenuated viral vaccines are more effective, but they take a longer time to reach the site, have special storage conditions (subzero temperatures), and might carry the risk of adverse effects. vaccines containing nucleic acids (rna and dna) are easy to make, but they are expensive and may require many doses. there have been reports of nanoparticle vaccinations for covid-19 eliciting an effective immunological response in mice after a single dosage. the mrna is encapsulated in lipid-based nanoparticles or other substances and can be injected into the host via im (intramuscular) route to increase its stability and prevent it admet & dmpk 00(0) (2023) 000-000 infections associated with sars-cov-2 exploited via photodynamic therapy doi: https://doi.org/10.5599/admet.1883 5 from rapid degradation by host ribonucleases [50]. the mrna-1273 vaccine is the first mrna vaccine with lipid microparticle capsule-based modified viral rna against covid-19 [51]. diagnosis the diagnosis is made depending on a range of factors, including epidemiology, clinical findings, in vitro assays, and the nucleic acid amplification test. the most accurate assay for detecting sars-cov-2 is presently being reliably detected by real-time reverse transcriptase-pcr (rrt-pcr), which is reinforced by other auxiliary assays like serology and radiography [52]. commercial kits have been brought into the field by validating numerous molecular and immunological aspects by the fda (food and drug administration) and icmr (indian council of medical research) [53,54]. however, given the drawbacks in terms of sensitivity and specificity, as well as the gaps in monitoring the virus spread, there is an urgent need to develop a novel diagnostic technique that is both rapid and accurate with high safety concerns to deploy them on a wide scale in order to halt the global outbreak. nanoparticle-based diagnosis because of its simplicity, high sensitivity, and high specificity based on exponential growth in rna generated throughout the operation, rt-pcr was used in most viral rna detection procedures [55]. although rt-pcr techniques are commonly accepted as the gold standard for coronavirus detection, they do have certain drawbacks, such as limited extraction efficiency, time-consuming procedure, and false-positive result due to contamination. to improve the method of detection of virus efficacy, ultra-small nanoparticles have been used not only in rt-pcr but also in other viral detection techniques such as an enzyme-linked immunosorbent assay (elisa) and rt-lamp (reverse transcription loop-mediated isothermal amplification) [56,57]. metal nanoparticles, carbon nanotubes, silica nanoparticles, quantum dots (qds), and polymeric nanoparticles have been examined in the context of viral detection. colorimetric, fluorescent, electrochemical, and optical imaging techniques are used as most diagnostic methods. metal nanoparticles and qds with distinct optical characteristics offer increased sensitivity for optical biosensing, and magnetic properties are used in the extraction process of the virus [58-61]. nanoparticle-based viral detection method has been identified as a suitable target for sars-cov-2. nanoparticles will play a major role in enhancing not only coronavirus detection efficiency but also biological pathogen diagnoses, with improvement in study and development (figure 2) [62,63]. figure 2. nanoparticles used for prevention to treatment against sars-cov-2 https://doi.org/10.5599/admet.1883 p. pallavi et al. admet & dmpk 00(0) (2023) 000-000 6 treatment researchers are already into the development of a variety of pharmaceuticals to combat covid-19. drugs being investigated for the treatment against covid-19 follow either one of the two mechanisms: (1) to control the symptoms caused or (2) to inactivate by targeting the replication cycle of the virus [64,65]. since no proper drug is developed, many antimicrobial drugs are being provided to the patients as a first-line treatment is tabulated below (table 1). table 1. summary of antiviral drugs being used for covid-19 to suppress aggressiveness. no name of the drugs class of the drug route of administration mechanism of action adverse effect 1. remdesivir antiviral drug intravenous (iv) injection possess inhibitory activity by inhibiting viral rna-dependent polymerase. chest tightness, back pain, dark-colored urine, light-colored stools, flushing, headache, nausea, and vomiting. 2. dexamethasone corticosteroid oral / intravenous / intramuscular (im) / intraarticular inhibit a proinflammatory gene encoding cell adhesion molecules (cam). glaucoma, cataracts, fluid retention, hypertension, osteoporosis, mood swings, confusion, and irritation. 3. tocilizumab antiviral drug iv infusion binds to both sil-6r and mil-6r to inhibit the action. stuffy nose, sore throat, headache, hypertension, and injection site reactions. the rare, more serious side effects include gastrointestinal perforations. 4. favipiravir antiviral drug oral inhibit replication of influenza a and b, ebola, and other pathogenic viral infections, but the drug has shown promise in the treatment of avian influenza. diarrhea, hyperuricemia, reduced neutrophil count, abdominal pain, nausea, and vomiting. 5. ivermectin antiparasitic drug subcutaneous / oral it targets the impα component and blocks the nuclear transport of viral proteins. headache, dizziness, muscle pain, nausea, swollen lymph nodes, eye swelling, weakness, vision changes, and itching. 6. lopinavir /ritonavir antiretroviral oral inhibit the action of 3clpro therapy to disrupt viral replication. muscle weakness, diarrhea, heart burning, difficulty in sleep, muscle pain, nausea loss of appetite .7. baricitinib antirheumatic drug oral inhibit jak1/jak2 to modulate downstream inflammatory responses. anemia, liver disease, nausea, loss of appetite, abdominal pain, yellowing eyes/skin, dark urine, fever, fainting. 8. chloroquine and hydroxychloroquine antimalarial drug subcutaneous / intramuscular inhibit endosomal acidification to block glycosylation and break down the production of viral proteins. dizziness, loss of appetite, diarrhea, blurred vision, sensitivity to light, ringing in ears, muscle weakness, and mental change. admet & dmpk 00(0) (2023) 000-000 infections associated with sars-cov-2 exploited via photodynamic therapy doi: https://doi.org/10.5599/admet.1883 7 no name of the drugs class of the drug route of administration mechanism of action adverse effect 9. nafamostat and comostat antiviral drug ora l/ iv the antagonist of serine protease tmprss2 and inhibit sars-cov 2 rash, pruritus, nausea, abdominal discomfort, and liver enzyme level elevation 10. fluvoxamine anti-inflammatory, antiviral, and antidepressant oral blocks reuptake of serotonin enhances its action difficult breathing, skin rash, blisters, hives, fever, joint pain, swelling of the throat 11. umifenovir antiviral oral inhibits membrane fusion of influenza virus and prevents contact between the virus and host tissues nausea, vomiting, diarrhea 12. famotidine histamine (h2) receptor antagonist oral activate the h2 receptor and catalyze proton pump action to secrete acid muscle cramps, abdominal discomfort, fatigue, anorexia, rash, dry mouth photodynamic therapy (pdt) the use of photosensitizer (ps), light, and molecular oxygen (o2) has enhanced through time, and the term pdt has developed [66,67]. pdt is a non-invasive emerging therapeutic method that works by activating the photosensitizer (ps) with a certain wavelength of light. in the process of illumination, the ps transfers its energy to the molecular oxygen when it is stimulated, which in turn generates cytotoxic molecules, namely reactive oxygen species (ros), which have the ability to disrupt the cell wall of the microorganisms in a variety of diseases and infections. the singlet oxygen produced can efficiently oxidize the main cellular macromolecules that destroy the microorganisms [67]. this therapy has been implemented to treat cancer, bacterial infection, fungal and viral diseases, as well as photodynamic diagnostics in the area of dentistry. pdt does not pose harmful risks to the biological system. it is clear that compared with traditional treatment modalities, pdt has its advantages due to its limited invasiveness and negligible cumulative toxicity. therefore, pdt aims to boost the quality of life of patients. jablonski's schematic diagram displays the intricacies of photodynamic activity, defined as an array of photophysical and photochemical reactions (figure 3). pdt needs the synchronous combination of the ps, proper wavelength of visible light, and molecular oxygen. the photosensitizing molecule possesses two electrons (opposite spins) in its ground state, and the symbol s0 is denoted as total spin, which is zero. the reaction includes absorption of light by ps and triggers an array of photochemical reactions leading to ros production. the ros, also termed singlet oxygen (1o2) in the lowest energy state, can cause serious oxidative injury to pathogenic microorganisms like viruses, bacteria, fungi, and parasites. during the pdt process, it produces other 1o2, such as superoxide ions (o2-•), hydrogen peroxide (h2o2), and hydroxyl radicals (oh•) [68,69]. according to the mechanism of pdt, the ps excites from the low energy state to a higher energy state with a limited life cycle ranging from nanoseconds to less after the absorption of light. after excitation, any one of the two mechanisms can occur: (1) fall back to grounds state and emits fluorescence or (2) reach triplet state by undergoing the crossing in intersystem. the triplet state lifetime provides adequate time for the excited ps to interact with the oxygen molecule or other tissue substrates [70]. during this mechanism, the proton gets transferred due to this direct interaction between triplet ps and substrate, which leads to the formation of radical anion or cation, which in turn tends to react with molecular oxygen and produces oxygenated compounds like superoxide anion radical, hydrogen peroxides, and hydroxide radicals. this method is termed a type 1 reaction of pdt. in type ii reaction, the direct energy transfer of the excited ps to molecular oxygen, in turn, forms singlet oxygen (1o2). both the type i and type ii reactions involve the https://doi.org/10.5599/admet.1883 p. pallavi et al. admet & dmpk 00(0) (2023) 000-000 8 formation of some final products, which decides the therapeutic index of the whole therapy. in a pdt reaction, both the types i and ii reactions can result simultaneously [71,72]. in accordance with the category of ps, the amount of substrate in tissues, and the molecular oxygen concentration, the occurrence rate of type i and type ii reaction differs. most of the studies suggest that 1o2 plays a significant role in pdt. figure 3. schematic representation of the mechanism of pdt against viral infections pdt for the treatment of viral infections infections are one of the major public health issues being studied by diverse research teams. the number of novel antimicrobials and their target structures has steadily reduced, so there is a need for alternative infection treatment [73,74]. photodynamic therapy is considered an effective and alternative treatment option that has been shown to have great potential against a variety of disease states caused by an infectious microorganism and may be a viable option for treatment against sars-cov-2 infections, which can be studied further to improve the treatment success rate. the outbreak of covid-19 has expanded enormously due to the insufficiency of immunizations also therapeutic interventions for the prevention and management of viral infections. because of the severity and urgency of covid-19, a new technique needs to be developed for the prevention and treatment of the virus [75,76]. antiviral photodynamic therapy (apdt) is likely to be useful as a potential therapy for coronavirus inhibition and reduction. the effectiveness of apdt in inactivating mammalian viruses has been established in several research outcomes (table 2). table 2. recent findings on the effectiveness of pdt against viral infections no ps wavelength, nm target finding remarks 1. 5aminolevulinic acid (5-ala) 635 juvenile laryngeal papilloma after pdt treatment, laryngeal papilloma was almost cured, and no recurrence was observed during 6-24 months of continuous physical assessment. pdt treatment with the combination of co2 laser therapy can be helpful for the common benign tumor, which can be related to human papillomavirus (hpv) infections. 2. curcumin 450 pharyngo tonsillitis ps was able to get delivered into the pharynx and tonsils. the microbial reduction in 1 log10 of the colonyforming unit was observed after therapy, and also recolonization was not observed after 24 h. the use of photodynamic therapy to treat pharyngo tonsillitis causes only minimal side effects. admet & dmpk 00(0) (2023) 000-000 infections associated with sars-cov-2 exploited via photodynamic therapy doi: https://doi.org/10.5599/admet.1883 9 no ps wavelength, nm target finding remarks 3. methylene blue (mb) 660 herpes simplex virus (hsv-1) after pdt, the first session showed betterment in antiinflammatory activity and itching and increased the healing process of wounds. treating herpes simplex infections in the nasal wing with pdt and photobiomodulation was beneficial and delivering drug (ps) via nebulization can overcome drug release problems, minimizing side effects caused due to iv administration. 4. methylene blue (mb) and radahlorin 662 vero e6 cells and sarscov 2 in the presence of mb and radahlorin, the ic50 of 102 tcid50 of sars cov-2 was found as 0.22 µg/ml and 0.33 µg/ml. inhibitory concentration-50% for radahlorin and mb was observed at 0.6 µg/ml and 2 µg/ml. in vitro, when pdt is coupled with radahlorin and mb, high antiviral efficacy against sars-cov-2 was observed. 5. phenothiazines, methylene blue (mb), porphyrins, protoporhphyrineix (pp-ix) covid19 apdt treatment using cost-effective ps could assist in mitigating covid-19 by treating infected individuals by developing functional photoactive textiles, auto-disinfecting surfaces, and disinfecting water and air. photosensitizers mb and ppix are new technical methods to treat covidaffected patients in a less costly and safer way by disinfecting infected water and water surface which cause the covid disease. 6. methylene blue 660 lip lesions 0.1 % methylene blue (mb) photosensitizer was applied all over lip lesions and treated with apdt and photomodulation showed complete healing in 4 days. covid-19-related labial lesions can be efficiently treated by combining phototherapy modalities. 7. curcumin polylactic-coglycolide acid sars-cov-2, vero cells apdt had anti-covid action in vitro without causing cytotoxicity in vero cells cultured with sars-cov-2 treated plasma in vitro, covid-19 action was observed in treated plasma containing sarscov2 without apoptosis in vero cell or any negative effect on plasma quality in photodynamic therapy. 8. 5aminolevulinic acid 635 hsil (high-grade intraepithelial lesion)/cin2 (cervical squamous intraepithelial neoplasia 2) with high-risk human papillomavirus (hhpv) 5-ala pdt possessed higher safety and efficacy profile for the hsil/cin2 patients needing reproductive treatment. it can be a promising alternative surgical technique. ala-pdt is effective for treating cervical hsil and high-risk hpv infections, and the effect may be maintained for a long time. https://doi.org/10.5599/admet.1883 p. pallavi et al. admet & dmpk 00(0) (2023) 000-000 10 no ps wavelength, nm target finding remarks 9. porphyrin sars-cov-2 coating of siloxane copolymer onto the textiles produces excitation of porphyrin that shows 1o2 oxidative crosslinking of the polysiloxane with amine links to create hydrophobic fabric. the development of a treated mask can be a barrier and provide better inactivation of viral structures for the transmission of sars-cov-2 and also have broad-antiviral activity application against human-enveloped viruses such as influenza a virus and other coronaviruses. 10. methylene blue influenza virus, sars-cov2, and mers (middle east respiratory syndrome) as studies suggest, mb-mediated apdt can minimize covid-19 impact in the absence of therapy for covid-19. methylene blue as ps for apdt can take control of coronavirus and helps to eradicate it. more research is required for the use of antiviral pdt as a treatment against covid-19. examples such as adenovirus, hepatitis viruses a, b, and c, hiv (human immunodeficiency virus), parvovirus b19, herpes virus, etc., causing infections have been studied to get treated with pdt, and certain studies have found that the viruses with the outer envelope are more susceptible to apdt than non-enveloped viruses. apdt is being utilized for inactivating multiple viruses in a variety of biological fluids, including blood. treating superficial viral lesions with apdt is also very effective [77]. the ros will be produced when a photoinactivation method is performed on a variety of significant biomolecule targets, such as dna, lipids, and proteins. ros produces oxidative stress and causes irreparable damage to the cellular structure of the virus. this oxidative stress induces apoptosis and necrosis to kill the virus without causing damage to the surrounding healthy tissues. both intracellular as well as extracellular ros can be released by apdt. ps is considered as one of the essential components of pdt. colors of porphycenes and phenothiazines, such as azure, methylene blue (mb), toluidine blue o (tbo), chlorine, porphyrin, and phthalocyanine derivatives are suitable as ps for the apdt [78]. covid-19 is now treated with chloroquine (c18h26cln3), and it has a structural similarity with mb (c16h18cln3s). according to the preliminary data from a recent trial, mb might be an effective ps for the therapy for flu-like conditions like covid-19. studies proved that mb efficiently inhibits the spike protein and ace2 (angiotensin-converting enzyme 2) receptor of sars-cov-2 to inactivate the viral replication [79]. nanoformulated pdt management of sars-cov-2 infections is still in progress, and some vaccines are in use in an emergency. nanotechnology has been employed to improve the efficacy of those vaccines to target covid-19 (table 3). researchers currently focus on nanomaterials-based antiviral photodynamic therapy treatment. that has been proven as an innovative strategy to inhibit bacterial and viral infections like papilloma and herpes. nanomaterials can improve solubility, increase blood circulation time, limit enzymatic degradation, lesser unwanted side effects, and improve medication bioavailability [76]. in antimicrobial pdt, employing nanoparticles can improve the solubility, photophysics, and photochemistry of photosensitizers with targeting features. admet & dmpk 00(0) (2023) 000-000 infections associated with sars-cov-2 exploited via photodynamic therapy doi: https://doi.org/10.5599/admet.1883 11 table 3. nanoenabled immunization against covid-19 no. vaccine vaccine type administrat ion regimen target virus target disease nanoformulation mode of action company 1. bnt162 bnt162a1 uridine mrna (urna) vaccine prime/ boost (p/b) regimen sarscov-2 covid-19 lipid nanoparticles (lnp) na biontech and pfizer bnt162b1 nucleoside -modified rna vaccine p/b regimen targets the spike protein of membranebound sars-cov-2 and encodes for its prefusion state. bnt162b2 or comirnaty nucleoside -modified rna vaccine p/b regimen targets the spike protein of sars-cov-2 and encodes their receptor-binding domain, thereby eliciting human antibody and th1 t cell responses bnt162b3 nucleoside -modified rna vaccine p/b regimen na bnt162c2 selfamplifying mrna (sarna) vaccine p/b regimen or single dose (sd) regimen na 2. lunar-cov19 self-transcribing and replicating mrna vaccine p/b regimen or sd regimen sarscov-2 covid-19 lnp encodes the sarscov-2 spike protein and an alphavirusbased replicon. arcturus therapeutics 3. spikevax or mrna1273 or elasomeran nucleoside -modified rna vaccine two dose (td) regimen sarscov-2 covid-19 lnp targets the spike glycoprotein of sarscov-2 and encodes for its prefusion state. moderna in this process, ps can be encapsulated in different nanostructures or nanoparticles like micelles, reverse micelles, liposomes, metal and metal-oxide nanoparticles, polymer nanoparticles, carbon nanotubes, and ceramic-based nanoparticles. in this regard, the dendrimers are also promising to conjugate or load the ps molecules for better apdt [80]. the branch-like structure and lipophilicity of dendrimers enhance the solubility of ps and, at the same time, help in the improvement of cellular uptake. the other advantages include the higher production of ros in nanoformulations. ps molecules are well separated by the nanostructures like liposomes, dendrimers, etc., to stop self-quenching [81]. recently up-conversion nanoparticles-based pdt has gained popularity for its absorption band in the nir zone. after absorbing nir radiation, the up-conversion nanoparticles (ucnp) emit visible light, which is absorbed by the ps molecules to react with molecular oxygen for the generation of ros [82,83]. lim et al. showed that the ucnp-mediated ros generation could inactivate the dengue virus serotype 2 (denv2) and adenovirus type 5 (ad5v) [84]. this strategy can be well applicable to the management of sars-cov-2 infections. the surface modification strategy of nanoformulated ps with targeting moieties as specific ligands or antibodies can help to increase the specificity of apdt against covid-19 protecting healthy cells [85]. carbon nanotubes have been proposed for use as novel photosensitizers in photodynamic therapy. carbon nanotubes that have been functionalized, conjugated, or encapsulated with other photosensitizers are powerful for pdt against infectious diseases. banerjee et al. investigated ppix conjugated to multi-walled and https://doi.org/10.5599/admet.1883 p. pallavi et al. admet & dmpk 00(0) (2023) 000-000 12 single-walled carbon tubes, and this conjugation was found to inhibit the influenza a virus [86]. for recently discovered covid-19, surface-modified nanoparticles could target the receptor present on the lung cell with the particular antibodies causing the infected cells to be destroyed by apdt. recent research suggests that apdt may be effective in the treatment of covid-19 with lesser side effects and pharmacological interactions. computational approach the computational approach of the combination of chemistry and biology has evolved as one of the essential parts of drug discovery. new lead molecules can be significantly developed by computational approach at minimal cost within a limited period of time. the efficacy of the method of testing and analyzing the developed products can be effectively improved [87]. certain tools like molecular docking, virtual screening, protein modeling, admet (adsorption, distribution, metabolism, excretion, and toxicity), molecular dynamics, and qsar (quantitative structure-activity relationship) are capable of providing valid predictions and has raised as key part of the computational approach. the photochemical reactivity of ps needs to be well understood for an effective yield and for which molecular orbital calculations are used as they can predict the mechanism of action and anticipate the results. wang et al. recently designed a ps for pdt as cancer treatment using computer-aided drug designing [88]. the team used two drugs where the similarity in molecular properties and the intermolecular interactions between the two drugs were identified with a dynamic simulation model and molecular docking, respectively. the efficiency of the particle for cancer therapy was greatly improved due to the screening test done via a computational approach [88]. fedorov et al. utilized coarse graining molecular dynamic method to understand the molecular interaction between the developed ps and the surface structure of the virus, and with the same, the major and minor binding sites were also studied [89]. similarly, sharshov et al. studied the chemistry behind the interaction between ps and virus using dynamic simulation methods [90]. table 4 summarizes the other computational approaches to formulate the ps. table 4. the summary of the computer-aided designing and analysis of ps. no ps designed method used purpose ref. 1. palladium-platinum complex gaussian 16 program two-photon pdt cancer therapy [91] 2. imidazolium-based porous organic polymer photosensitizer non-adiabatic molecular dynamics simulation anticancer pdt [92] 3. benzothiazole-based organic ps gaussian 16 program novel design of organic ps that is operative in the nir region [93] 4. halogen-free boron dipyrromethene photosensitizer density functional theory (dft) with the dependence on time (tddft) bioimaging and anticancer pdt [94] 5. propolis-benzofuran a photosensitizer swissadme tool, osiris and pro-tox ii pdt for monkeypox virus [95] 6. cy7photolabile protecting groups gaussview 6.0 facile synthesis for biomedical applications [96] 7. meso-tetraphenylporphyrin tddft anticancer pdt [97] 8. dicyanomethylene-4h-chromene based ps mesra software with gaussian 16 program image-guided anticancer pdt [98] 9. fucoxanthin/graphene complex dft to study the photosensitization features for various applications [99] 10. porphyrin-based ps with two ethanethioate dft and tddft calculation of the excited state properties of the novel design [100] admet & dmpk 00(0) (2023) 000-000 infections associated with sars-cov-2 exploited via photodynamic therapy doi: https://doi.org/10.5599/admet.1883 13 conclusion and future perspective covid-19 is rapidly spreading around the world, resulting in an increase in affected cases and deaths. we are confronted with many unknowns, and we must continue to monitor and research. until now, no appropriate therapy against sars-cov-2 for a complete cure is available. to resolve this issue, researchers are focusing on developing viable vaccinations and treatments. nanotechnology is now emerging as a potential field for diagnosing and treating several disorders. due to the small size and exponential surface chemistry, the nanoparticle can be an effective tool in manufacturing vaccines and developing treatment strategies against covid-19 infection. pdt is a non-invasive therapeutic strategy based on the activation of a ps with the help of light in a certain wavelength. in pdt, the nanoparticle can act as a drug carrier or a drug by itself. in order to overcome the drawback of ps, nanoparticles can be used as a carrier of ps. studies revealed nanoparticle-based antiviral medicines and nano-based-pdt would sufficiently combat the covid-19 pandemic. authors contribution: pp & kh collected all required data and wrote the initial draft. nae, pg, and kg compiled it with additional data. the concept, overall representation, and final version were prepared by ag. conflict of interest: the authors declare that there is not any conflict of interest. acknowledgements: authors acknowledge care for financial and infrastructural support. pp, kr & pg acknowledge care for fellowships too. references [1] a. valsamatzi-panagiotou, r. penchovsky. environmental factors influencing the transmission of the coronavirus 2019: a review. environmental chemistry letters 20 (2022) 1603-1610. https://doi.org/ 10.1007/s10311-022-01418-9. [2] a. girigoswami, k. girigoswami. versatile applications of nanosponges in biomedical field: a glimpse on sars-cov-2 management. bionanoscience 12 (2022) 1018-1031. https://doi.org/10.1007/s126 68-022-01000-1. [3] g.s. el-sayyad, d. elfadil, m.s. gaballah, d.m. el-sherif, m. abouzid, h.g. nada, m.s. khalil, m.a. ghorab. implication of nanotechnology to reduce the environmental risks of waste associated with the covid-19 pandemic. rsc advances 13 (2023) 12438-12454. https://doi.org/10.1039/ d3ra01052j. [4] m. rademaker, c. baker, p. foley, j. sullivan, c. wang. advice regarding covid‐19 and use of immunomodulators, in patients with severe dermatological diseases. the australasian jjournal of dermatology (2020). https://doi.org/10.1111/ajd.13295. [5] s. aftab, m.z. iqbal, s. hussain, h.h. hegazy. recent advances in nanomaterials‐based fets for sars‐ cov‐2 (covid‐19 virus) diagnosis. advanced functional materials (2023) 2301007. https://doi.org/10.1002/adfm.202301007. [6] n. agharezaei, f. forouzesh. sars-cov-2: history, genetics, and treatment. journal of arak university of medical sciences 23 (2020) 666-685. https://doi.org/10.32598/jams.23.cov.5712.2. [7] a. tharayil, r. rajakumari, a. kumar, m.d. choudhary, p. palit, s. thomas. new insights into application of nanoparticles in the diagnosis and screening of novel coronavirus (sars-cov-2). emergent materials 4 (2021) 101-117. https://doi.org/10.1007/s42247-021-00182-w. [8] p. parthasarathy, s. vivekanandan. an extensive study on the covid-19 pandemic, an emerging global crisis: risks, transmission, impacts and mitigation. journal of infection and public health 14 (2021) 249-259. https://doi.org/10.1016/j.jiph.2020.12.020. [9] y. fan, k. zhao, z.-l. shi, p. zhou. bat coronaviruses in china. viruses 11 (2019) 210. https://doi.org/ 10.3390/v11030210. https://doi.org/10.5599/admet.1883 https://doi.org/‌10.1007/s10311-022-01418-9 https://doi.org/‌10.1007/s10311-022-01418-9 https://doi.org/10.1007/s126‌68-022-01000-1 https://doi.org/10.1007/s126‌68-022-01000-1 https://doi.org/10.1039/‌d3ra01052j https://doi.org/10.1039/‌d3ra01052j https://doi.org/10.1111/ajd.13295 https://doi.org/10.1002/adfm.202301007 https://doi.org/10.32598/jams.23.cov.5712.2 https://doi.org/10.1007/s42247-021-00182-w https://doi.org/10.1016/j.jiph.2020.12.020 https://doi.org/‌10.3390/v11030210 https://doi.org/‌10.3390/v11030210 p. pallavi et al. admet & dmpk 00(0) (2023) 000-000 14 [10] ss bandyopadhyay, a.k. halder, s. saha, p. chatterjee, m. nasipuri, s. basu. assessment of go-based protein interaction affinities in the large-scale human–coronavirus family interactome. vaccines 11 (2023) 549. https://doi.org/10.3390/vaccines11030549. [11] a.s. ray, k. bhattacharya. an overview on the zoonotic aspects of covid-19. proceedings of the national academy of sciences, india section b: biological sciences (2023) 1-5. https://doi.org/ 10.1007/s40011-023-01445-8. [12] b. lou, t.-d. li, s.-f. zheng, y.-y. su, z.-y. li, w. liu, f. yu, s.-x. ge, q.-d. zou, q. yuan. serology characteristics of sars-cov-2 infection after exposure and post-symptom onset. european respiratory journal 56 (2020). https://doi.org/10.1183/13993003.00763-2020. [13] g.m. vahey, k.e. marshall, e. mcdonald, s.w. martin, j.e. tate, c.m. midgley, m.e. killerby, b. kawasaki, r.k. herlihy, n.b. alden. symptom profiles and progression in hospitalized and nonhospitalized patients with coronavirus disease, colorado, usa, 2020. emerging infectious diseases 27 (2021) 385. https://doi.org/10.3201/eid2702.203729. [14] e.a. meyerowitz, a. richterman, i.i. bogoch, n. low, m. cevik. towards an accurate and systematic characterisation of persistently asymptomatic infection with sars-cov-2. the lancet infectious diseases 21 (2021) e163-e169. https://doi.org/10.1016/s1473-3099(20)30837-9. [15] j.a. siordia, m. bernaba, k. yoshino, a. ulhaque, s. kumar, m. bernaba, e. bergin. systematic and statistical review of coronavirus disease 19 treatment trials. sn comprehensive clinical medicine 2 (2020) 1120-1131. https://doi.org/10.1007/s42399-020-00399-6. [16] m. hossen, m.a. barek, n. jahan, m. safiqul islam. a review on current repurposing drugs for the treatment of covid-19: reality and challenges. sn comprehensive clinical medicine 2 (2020) 17771789. https://doi.org/10.1007/s42399-020-00485-9. [17] p. pagliano, g. scarpati, c. sellitto, v. conti, a.m. spera, t. ascione, o. piazza, a. filippelli. experimental pharmacotherapy for covid-19: the latest advances. journal of experimental pharmacology 13 (2021) 1. https://doi.org/10.2147/jep.s255209. [18] g.j. soufi, a. hekmatnia, m. nasrollahzadeh, n. shafiei, m. sajjadi, p. iravani, s. fallah, s. iravani, r.s. varma. sars-cov-2 (covid-19): new discoveries and current challenges. applied sciences 10 (2020) 3641. https://doi.org/10.3390/app10103641. [19] r. ranjbar, h. mahmoodzadeh hosseini, f. safarpoor dehkordi. a review on biochemical and immunological biomarkers used for laboratory diagnosis of sars-cov-2 (covid-19). the open microbiology journal 14 (2020). https://doi.org/10.2174/1874434602014010290. [20] y. huang, c. yang, x.-f. xu, w. xu, s.-w. liu. structural and functional properties of sars-cov-2 spike protein: potential antivirus drug development for covid-19. acta pharmacologica sinica 41 (2020) 1141-1149. https://doi.org/10.1038/s41401-020-0485-4. [21] c.r. carlson, j.b. asfaha, c.m. ghent, c.j. howard, n. hartooni, m. safari, a.d. frankel, d.o. morgan. phosphoregulation of phase separation by the sars-cov-2 n protein suggests a biophysical basis for its dual functions. molecular cell 80 (2020) 1092-1103. e1094. https://doi.org/10.1016/j.mol cel.2020.11.025. [22] j. w schott, m. galla, t. godinho, c. baum, a. schambach. viral and non-viral approaches for transient delivery of mrna and proteins. current gene therapy 11 (2011) 382-398. https://doi.org/10.2174/ 156652311797415872. [23] t. velikova, t. georgiev. sars-cov-2 vaccines and autoimmune diseases amidst the covid-19 crisis. rheumatology international 41 (2021) 509-518. https://doi.org/10.1007/s00296-021-04792-9. [24] y.-f. tu, c.-s. chien, a.a. yarmishyn, y.-y. lin, y.-h. luo, y.-t. lin, w.-y. lai, d.-m. yang, s.-j. chou, y.p. yang. a review of sars-cov-2 and the ongoing clinical trials. international journal of molecular sciences 21 (2020) 2657. https://doi.org/10.3390/ijms21072657. [25] f. oroojalian, a. haghbin, b. baradaran, n. hemmat, m.-a. shahbazi, h.b. baghi, a. mokhtarzadeh, m.r. hamblin. novel insights into the treatment of sars-cov-2 infection: an overview of current clinical trials. international journal of biological macromolecules 165 (2020) 18-43. https://doi.org/10.1016/j.ijbiomac.2020.09.204. https://doi.org/10.3390/vaccines11030549 https://doi.org/‌10.1007/s40011-023-01445-8 https://doi.org/‌10.1007/s40011-023-01445-8 https://doi.org/10.1183/13993003.00763-2020 https://doi.org/10.3201/eid2702.203729 https://doi.org/10.1016/s1473-3099(20)30837-9 https://doi.org/10.1007/s42399-020-00399-6 https://doi.org/10.1007/s42399-020-00485-9 https://doi.org/10.2147/jep.s255209 https://doi.org/10.3390/app10103641 https://doi.org/10.2174/1874434602014010290 https://doi.org/10.1038/s41401-020-0485-4 https://doi.org/10.1016/j.mol‌cel.2020.11.025 https://doi.org/10.1016/j.mol‌cel.2020.11.025 https://doi.org/10.2174/‌156652311797415872 https://doi.org/10.2174/‌156652311797415872 https://doi.org/10.1007/s00296-021-04792-9 https://doi.org/10.3390/ijms21072657 https://doi.org/10.1016/j.ijbiomac.2020.09.204 admet & dmpk 00(0) (2023) 000-000 infections associated with sars-cov-2 exploited via photodynamic therapy doi: https://doi.org/10.5599/admet.1883 15 [26] l. yan, y. yang, m. li, y. zhang, l. zheng, j. ge, y.c. huang, z. liu, t. wang, s. gao. coupling of n7methyltransferase and 3′-5′ exoribonuclease with sars-cov-2 polymerase reveals mechanisms for capping and proofreading. cell 184 (2021) 3474-3485. e3411. https://doi.org/10.1016/j.cell.20 21.05.033. [27] g. freer, m. lai, p. quaranta, p.g. spezia, m. pistello. evolution of viruses and the emergence of sarscov-2 variants. new microbiologica 44 (2021) 191-204. https://doi.org/europepmc.org/article/med/ 34942015. [28] s. qiu, y. hu. are covid-19 susceptibility genes related to lung cancer? journal of infection 83 (2021) 607-635. https://doi.org/10.1016/j.jinf.2021.08.032. [29] c. karthika, r. swathy krishna, m. rahman, r. akter, d. kaushik. covid-19, the firestone in 21st century: a review on coronavirus disease and its clinical perspectives. environmental science and pollution research 28 (2021) 64951-64966. https://doi.org/10.1007/s11356-021-16654-9. [30] m. verdecia, j.f. kokai-kun, m. kibbey, s. acharya, j. venema, f. atouf. covid-19 vaccine platforms: delivering on a promise? human vaccines & immunotherapeutics 17 (2021) 2873-2893. https://doi.org/10.1080/21645515.2021.1911204. [31] y. li, y. bi, h. xiao, y. yao, x. liu, z. hu, j. duan, y. yang, z. li, y. li. a novel dna and protein combination covid-19 vaccine formulation provides full protection against sars-cov-2 in rhesus macaques. emerging microbes & infections 10 (2021) 342-355. https://doi.org/10.1080/222217 51.2021.1887767. [32] y.b. seo, y.s. suh, j.i. ryu, h. jang, h. oh, b.-s. koo, s.-h. seo, j.j. hong, m. song, s.-j. kim. soluble spike dna vaccine provides long-term protective immunity against sars-cov-2 in mice and nonhuman primates. vaccines 9 (2021) 307. https://doi.org/10.3390/vaccines9040307. [33] b. doroftei, a. ciobica, o.-d. ilie, r. maftei, c. ilea. mini-review discussing the reliability and efficiency of covid-19 vaccines. diagnostics 11 (2021) 579. https://doi.org/10.3390/diagnostics11040579. [34] p. intapiboon, p. seepathomnarong, j. ongarj, s. surasombatpattana, s. uppanisakorn, s. mahasirimongkol, w. sawaengdee, s. phumiamorn, s. sapsutthipas, p. sangsupawanich. immunogenicity and safety of an intradermal bnt162b2 mrna vaccine booster after two doses of inactivated sars-cov-2 vaccine in healthy population. vaccines 9 (2021) 1375. https://doi.org/10.3390/ vaccines9121375. [35] w. ho, m. gao, f. li, z. li, x.q. zhang, x. xu. next‐generation vaccines: nanoparticle‐mediated dna and mrna delivery. advanced healthcare materials 10 (2021) 2001812. https://doi.org/ 10.1002/adhm.202001812. [36] s.s. rosa, d.m. prazeres, a.m. azevedo, m.p. marques. mrna vaccines manufacturing: challenges and bottlenecks. vaccine 39 (2021) 2190-2200. https://doi.org/10.1016/j.vaccine.2021.03.038. [37] r. yadav, j.k. chaudhary, n. jain, p.k. chaudhary, s. khanra, p. dhamija, a. sharma, a. kumar, s. handu. role of structural and non-structural proteins and therapeutic targets of sars-cov-2 for covid-19. cells 10 (2021) 821. https://doi.org/10.3390/cells10040821. [38] y. valdes-balbin, d. santana-mederos, f. paquet, s. fernandez, y. climent, f. chiodo, l. rodríguez, b. sanchez ramirez, k. leon, t. hernandez. molecular aspects concerning the use of the sars-cov-2 receptor binding domain as a target for preventive vaccines. acs central science 7 (2021) 757-767. https://doi.org/10.1021/acscentsci.1c00216. [39] t. li, w. xue, q. zheng, s. song, c. yang, h. xiong, s. zhang, m. hong, y. zhang, h. yu. crossneutralizing antibodies bind a sars-cov-2 cryptic site and resist circulating variants. nature communications 12 (2021) 1-12. https://doi.org/10.1038/s41467-021-25997-3. [40] n. wang, y. sun, r. feng, y. wang, y. guo, l. zhang, y.-q. deng, l. wang, z. cui, l. cao. structurebased development of human antibody cocktails against sars-cov-2. cell research 31 (2021) 101103. https://doi.org/10.1038/s41422-020-00446-w. [41] z. wu, t. li. nanoparticle-mediated cytoplasmic delivery of messenger rna vaccines: challenges and future perspectives. pharmaceutical research 38 (2021) 473-478. https://doi.org/10.1007/s11095021-03015-x. https://doi.org/10.5599/admet.1883 https://doi.org/10.1016/j.cell.20‌21.05.033 https://doi.org/10.1016/j.cell.20‌21.05.033 https://doi.org/europepmc.org/article/med/‌34942015 https://doi.org/europepmc.org/article/med/‌34942015 https://doi.org/10.1016/j.jinf.2021.08.032 https://doi.org/10.1007/s11356-021-16654-9 https://doi.org/10.1080/21645515.2021.1911204 https://doi.org/10.1080/222217‌51.2021.1887767 https://doi.org/10.1080/222217‌51.2021.1887767 https://doi.org/10.3390/vaccines9040307 https://doi.org/10.3390/diagnostics11040579 https://doi.org/10.3390/‌vaccines9121375 https://doi.org/10.3390/‌vaccines9121375 https://doi.org/‌10.1002/adhm.202001812 https://doi.org/‌10.1002/adhm.202001812 https://doi.org/10.1016/j.vaccine.2021.03.038 https://doi.org/10.3390/cells10040821 https://doi.org/10.1021/acscentsci.1c00216 https://doi.org/10.1038/s41467-021-25997-3 https://doi.org/10.1038/s41422-020-00446-w https://doi.org/10.1007/s11095-021-03015-x https://doi.org/10.1007/s11095-021-03015-x p. pallavi et al. admet & dmpk 00(0) (2023) 000-000 16 [42] a. girigoswami, w. yassine, p. sharmiladevi, v. haribabu, k. girigoswami. camouflaged nanosilver with excitation wavelength dependent high quantum yield for targeted theranostic. scientific reports 8 (2018) 1-7. https://doi.org/10.1038/s41598-018-34843-4. [43] v. haribabu, p. sharmiladevi, n. akhtar, as farook, k. girigoswami, a. girigoswami. label free ultrasmall fluoromagnetic ferrite-clusters for targeted cancer imaging and drug delivery. current drug delivery 16 (2019) 233-241. https://doi.org/10.2174/1567201816666181119112410. [44] p. sharmiladevi, m. breghatha, k. dhanavardhini, r. priya, k. girigoswami, a. girigoswami. efficient wormlike micelles for the controlled delivery of anticancer drugs. nanoscience & nanotechnologyasia 11 (2021) 350-356. https://doi.org/10.2174/2210681210999200728115601. [45] p. sharmiladevi, k. girigoswami, v. haribabu, a. girigoswami. nano-enabled theranostics for cancer. materials advances 2 (2021) 2876-2891. https://doi.org/10.1039/d1ma00069a. [46] s.w. vedakumari, r. senthil, s. sekar, c.s. babu, t.p. sastry. enhancing anti-cancer activity of erlotinib by antibody conjugated nanofibrin-in vitro studies on lung adenocarcinoma cell lines. materials chemistry and physics 224 (2019) 328-333. https://doi.org/10.1016/j.matchemphys.2018.11.061. [47] lm lazer, b. sadhasivam, k. palaniyandi, t. muthuswamy, i. ramachandran, a. balakrishnan, s. pathak, s. narayan, s. ramalingam. chitosan-based nano-formulation enhances the anticancer efficacy of hesperetin. international journal of biological macromolecules 107 (2018) 1988-1998. https://doi.org/10.1016/j.ijbiomac.2017.10.064. [48] f. hou, z. teng, j. ru, h. liu, j. li, y. zhang, s. sun, h. guo. flower-like mesoporous silica nanoparticles as an antigen delivery platform to promote systemic immune response. nanomedicine: nanotechnology, biology and medicine 42 (2022) 102541. https://doi.org/10.1016/j.nano.20 22.102541. [49] c.t. perciani, l.y. liu, l. wood, s.a. macparland. enhancing immunity with nanomedicine: employing nanoparticles to harness the immune system. acs nano 15 (2020) 7-20. https://doi.org/10.1021/acsnano.0c08913. [50] l. schoenmaker, d. witzigmann, j.a. kulkarni, r. verbeke, g. kersten, w. jiskoot, d.j. crommelin. mrna-lipid nanoparticle covid-19 vaccines: structure and stability. international journal of pharmaceutics 601 (2021) 120586. https://doi.org/10.1016/j.ijpharm.2021.120586. [51] j. van praet, m. reynders, d. de bacquer, l. viaene, m.k. schoutteten, r. caluwé, p. doubel, l. heylen, a.v. de bel, b. van vlem. predictors and dynamics of the humoral and cellular immune response to sars-cov-2 mrna vaccines in hemodialysis patients: a multicenter observational study. journal of the american society of nephrology 32 (2021) 3208-3220. https://doi.org/10.1681/ asn.2021070908. [52] e. hamidi-asl, l. heidari, j.b. raoof, t.p. richard, s. farhad, m. ghani. a review on the recent achievements on coronaviruses recognition using electrochemical detection methods. microchemical journal (2022) 107322. https://doi.org/10.1016/j.microc.2022.107322. [53] c.o. priyanka, i. singh. diagnosis of sars-cov-2: a review on the current scenario and future outlook. acta virologica 64 (2020) 396-408. https://doi.org/10.4149/av_2020_402. [54] s. yadav, m.a. sadique, p. ranjan, n. kumar, a. singhal, a.k. srivastava, r. khan. sers based lateral flow immunoassay for point-of-care detection of sars-cov-2 in clinical samples. acs applied bio materials 4 (2021) 2974-2995. https://doi.org/10.1021/acsabm.1c00102. [55] d. thompson, y. lei. mini review: recent progress in rt-lamp enabled covid-19 detection. sensors and actuators reports 2 (2020) 100017. https://doi.org/10.1016/j.snr.2020.100017. [56] j. lukose, s. chidangil, s.d. george. optical technologies for the detection of viruses like covid-19: progress and prospects. biosensors and bioelectronics 178 (2021) 113004. https://doi.org/10.1016/j.bios.2021.113004. [57] y. duan, w. wu, q. zhao, s. liu, h. liu, m. huang, t. wang, m. liang, z. wang. enzyme-antibodymodified gold nanoparticle probes for the ultrasensitive detection of nucleocapsid protein in sftsv. international journal of environmental research and public health 17 (2020) 4427. https://doi.org/10.3390/ijerph17124427. https://doi.org/10.1038/s41598-018-34843-4 https://doi.org/10.2174/1567201816666181119112410 https://doi.org/10.2174/2210681210999200728115601 https://doi.org/10.1039/d1ma00069a https://doi.org/10.1016/j.matchemphys.2018.11.061 https://doi.org/10.1016/j.ijbiomac.2017.10.064 https://doi.org/10.1016/j.nano.20‌22.102541 https://doi.org/10.1016/j.nano.20‌22.102541 https://doi.org/10.1021/acsnano.0c08913 https://doi.org/10.1016/j.ijpharm.2021.120586 https://doi.org/10.1681/‌asn.2021070908 https://doi.org/10.1681/‌asn.2021070908 https://doi.org/10.1016/j.microc.2022.107322 https://doi.org/10.4149/av_2020_402 https://doi.org/10.1021/acsabm.1c00102 https://doi.org/10.1016/j.snr.2020.100017 https://doi.org/10.1016/j.bios.2021.113004 https://doi.org/10.3390/ijerph17124427 admet & dmpk 00(0) (2023) 000-000 infections associated with sars-cov-2 exploited via photodynamic therapy doi: https://doi.org/10.5599/admet.1883 17 [58] h.k. choi, m.-j. lee, s.n. lee, t.-h. kim, b.-k. oh. noble metal nanomaterial-based biosensors for electrochemical and optical detection of viruses causing respiratory illnesses. frontiers in chemistry 9 (2021). https://doi.org/10.3389/fchem.2021.672739. [59] s. ghosh, k. girigoswami, a. girigoswami. membrane-encapsulated camouflaged nanomedicines in drug delivery. nanomedicine: nanotechnology, biology, and medicine 14 (2019) 2067-2082. https://doi.org/10.2217/nnm-2019-0155. [60] v. haribabu, k. girigoswami, a. girigoswami. magneto-silver core–shell nanohybrids for theragnosis. nano-structures & nano-objects 25 (2021) 100636. https://doi.org/10.1016/j.nanoso.2020.100636. [61] g. agraharam, a. girigoswami, k. girigoswami. nanoencapsulated myricetin to improve antioxidant activity and bioavailability: a study on zebrafish embryos. chemistry 4 (2021) 1-17. https://doi.org/10.3390/chemistry4010001. [62] r. medhi, p. srinoi, n. ngo, h.-v. tran, t.r. lee. nanoparticle-based strategies to combat covid-19. acs applied nano materials 3 (2020) 8557-8580. https://doi.org/10.1021/acsanm.0c01978. [63] c. durmus, s.b. hanoglu, d. harmanci, h. moulahoum, k. tok, f. ghorbanizamani, s. sanli, f. zihnioglu, s. evran, c. cicek. indiscriminate sars-cov-2 multivariant detection using magnetic nanoparticle-based electrochemical immunosensing. talanta (2022) 123356. https://doi.org/ 10.1016/j.talanta.2022.123356. [64] t. dube, a. ghosh, j. mishra, u.b. kompella, j.j. panda. repurposed drugs, molecular vaccines, immune‐modulators, and nanotherapeutics to treat and prevent covid‐19 associated with sars‐ cov‐2, a deadly nanovector. advanced therapeutics 4 (2021) 2000172. https://doi.org/10.1002/ adtp.202000172. [65] s. mouffak, q. shubbar, e. saleh, r. el-awady. recent advances in management of covid-19: a review. biomedicine and pharmacotherapy 143 (2021) 112107. https://doi.org/10.1016/j.bio pha.2021.112107. [66] k. girigoswami, p. pallavi, a. girigoswami, targeting cancer stem cells by nanoenabled drug delivery, in cancer stem cells: new horizons in cancer therapies, springer2020, p. 313-337. https://doi.org/10.1007/978-981-15-5120-8_17. [67] m. vimaladevi, k.c. divya, a. girigoswami. liposomal nanoformulations of rhodamine for targeted photodynamic inactivation of multidrug resistant gram negative bacteria in sewage treatment plant. journal of photochemistry and photobiology b: biology 162 (2016) 146-152. https://doi.org/ 10.1016/j.jphotobiol.2016.06.034. [68] l.v. chekulayeva, i.n. shevchuk, v.a. chekulayev, k. ilmarinen. hydrogen peroxide, superoxide, and hydroxyl radicals are involved in the phototoxic action of hematoporphyrin derivative against tumor cells. journal of environmental pathology, toxicology and oncology 25 (2006). https://doi.org/10.1615/jenvironpatholtoxicoloncol.v25.i1-2.40. [69] m. price, j.j. reiners, a.m. santiago, d. kessel. monitoring singlet oxygen and hydroxyl radical formation with fluorescent probes during photodynamic therapy. photochemistry and photobiology 85 (2009) 1177-1181. https://doi.org/10.1111/j.1751-1097.2009.00555.x. [70] e.s. shibu, m. hamada, n. murase, v. biju. nanomaterials formulations for photothermal and photodynamic therapy of cancer. journal of photochemistry and photobiology c: photochemistry reviews 15 (2013) 53-72. https://doi.org/10.1016/j.jphotochemrev.2012.09.004. [71] a.p. castano, t.n. demidova, m.r. hamblin. mechanisms in photodynamic therapy: part three— photosensitizer pharmacokinetics, biodistribution, tumor localization and modes of tumor destruction. photodiagnosis and photodynamic therapy 2 (2005) 91-106. https://doi.org/10.1016/ s1572-1000(05)00060-8. [72] j.m. dąbrowski, l.g. arnaut. photodynamic therapy (pdt) of cancer: from local to systemic treatment. photochemical & photobiological sciences 14 (2015) 1765-1780. https://doi.org/ 10.1039/c5pp00132c. [73] p. hillemanns, m.h. einstein, o.e. iversen. topical hexaminolevulinate photodynamic therapy for the treatment of persistent human papilloma virus infections and cervical intraepithelial neoplasia. https://doi.org/10.5599/admet.1883 https://doi.org/10.3389/fchem.2021.672739 https://doi.org/10.2217/nnm-2019-0155 https://doi.org/10.1016/j.nanoso.2020.100636 https://doi.org/10.3390/chemistry4010001 https://doi.org/10.1021/acsanm.0c01978 https://doi.org/‌10.1016/‌j.talanta.2022.123356 https://doi.org/‌10.1016/‌j.talanta.2022.123356 https://doi.org/10.1002/‌adtp.202000172 https://doi.org/10.1002/‌adtp.202000172 https://doi.org/10.1016/j.bio‌pha.2021.112107 https://doi.org/10.1016/j.bio‌pha.2021.112107 https://doi.org/10.1007/978-981-15-5120-8_17 https://doi.org/‌10.1016/j.jphotobiol.2016.06.034 https://doi.org/‌10.1016/j.jphotobiol.2016.06.034 https://doi.org/10.1615/jenvironpatholtoxicoloncol.v25.i1-2.40 https://doi.org/10.1111/j.1751-1097.2009.00555.x https://doi.org/10.1016/j.jphotochemrev.2012.09.004 https://doi.org/10.1016/‌s1572-1000(05)00060-8 https://doi.org/10.1016/‌s1572-1000(05)00060-8 https://doi.org/‌10.1039/c5pp00132c https://doi.org/‌10.1039/c5pp00132c p. pallavi et al. admet & dmpk 00(0) (2023) 000-000 18 expert opinion on investigational drugs 24 (2015) 273-281. https://doi.org/10.1517/13543784.20 15.990150. [74] ld. dias, v.s. bagnato. an update on clinical photodynamic therapy for fighting respiratory tract infections: a promising tool against covid-19 and its co-infections. laser physics letters 17 (2020) 083001. https://doi.org/10.1088/1612-202x/ab95a9. [75] v.a. svyatchenko, s.d. nikonov, a.p. mayorov, m.l. gelfond, v.b. loktev. antiviral photodynamic therapy: inactivation and inhibition of sars-cov-2 in vitro using methylene blue and radachlorin. photodiagnosis and photodynamic therapy 33 (2021) 102112. https://doi.org/10.1016/j.pdpdt.20 20.102112. [76] k. khorsandi, s. fekrazad, f. vahdatinia, a. farmany, r. fekrazad. nano antiviral photodynamic therapy: a probable biophysicochemical management modality in sars-cov-2. expert opinion on drug delivery 18 (2021) 265-272. https://doi.org/10.1080/17425247.2021.1829591. [77] c.p. sabino, a.r. ball, m.s. baptista, t. dai, m.r. hamblin, m.s. ribeiro, a.l. santos, f.p. sellera, g.p. tegos, m. wainwright. light-based technologies for management of covid-19 pandemic crisis. journal of photochemistry and photobiology b: biology 212 (2020) 111999. https://doi.org/ 10.1016/j.jphotobiol.2020.111999. [78] h. mahmoudi, a. bahador, m. pourhajibagher, m.y. alikhani. antimicrobial photodynamic therapy: an effective alternative approach to control bacterial infections. journal of lasers in medical sciences 9 (2018) 154. https://doi.org/10.15171/jlms.2018.29. [79] m. gheblawi, k. wang, a. viveiros, q. nguyen, j.-c. zhong, a.j. turner, m.k. raizada, m.b. grant, g.y. oudit. angiotensin-converting enzyme 2: sars-cov-2 receptor and regulator of the renin-angiotensin system: celebrating the 20th anniversary of the discovery of ace2. circulation research 126 (2020) 1456-1474. https://doi.org/10.1161/circresaha.120.317015. [80] m. klausen, m. ucuncu, m. bradley. design of photosensitizing agents for targeted antimicrobial photodynamic therapy. molecules 25 (2020) 5239. https://doi.org/10.3390/molecules25225239. [81] s. sansaloni-pastor, j. bouilloux, n. lange. the dark side: photosensitizer prodrugs. pharmaceuticals 12 (2019) 148. https://doi.org/10.3390/ph12040148. [82] p.c.a. swamy, g. sivaraman, r.n. priyanka, s.o. raja, k. ponnuvel, j. shanmugpriya, a. gulyani. near infrared (nir) absorbing dyes as promising photosensitizer for photo dynamic therapy. coordination chemistry reviews 411 (2020) 213233. https://doi.org/10.1016/j.ccr.2020.213233. [83] a. mavridi-printezi, m. guernelli, a. menichetti, m. montalti. bio-applications of multifunctional melanin nanoparticles: from nanomedicine to nanocosmetics. nanomaterials 10 (2020) 2276. https://doi.org/10.3390/nano10112276. [84] m.e. lim, y.-l. lee, y. zhang, j.j.h. chu. photodynamic inactivation of viruses using upconversion nanoparticles. biomaterials 33 (2012) 1912-1920. https://doi.org/10.1016/j.biomaterials.20 11.11.033. [85] r. youf, m. müller, a. balasini, f. thétiot, m. müller, a. hascoët, u. jonas, h. schönherr, g. lemercier, t. montier. antimicrobial photodynamic therapy: latest developments with a focus on combinatory strategies. pharmaceutics 13 (2021) 1995. https://doi.org/10.3390/pharmaceutics13121995. [86] i. banerjee, m.p. douaisi, d. mondal, r.s. kane. light-activated nanotube–porphyrin conjugates as effective antiviral agents. nanotechnology 23 (2012) 105101. https://doi.org/10.1088/09574484/23/10/105101. [87] i.j.d.s. nascimento, t.m. de aquino, e.f. da silva-júnior. the new era of drug discovery: the power of computer-aided drug design (cadd). letters in drug design & discovery 19 (2022) 951-955. https://doi.org/10.2174/1570180819666220405225817. [88] q. wang, m. sun, c. li, d. li, z. yang, q. jiang, z. he, h. ding, j. sun. a computer-aided chemphotodynamic drugs self-delivery system for synergistically enhanced cancer therapy. asian journal of pharmaceutical sciences 16 (2021) 203-212. https://doi.org/10.1016/j.ajps.2020.04.002. https://doi.org/10.1517/13543784.20‌15.990150 https://doi.org/10.1517/13543784.20‌15.990150 https://doi.org/10.1088/1612-202x/ab95a9 https://doi.org/10.1016/j.pdpdt.20‌20.102112 https://doi.org/10.1016/j.pdpdt.20‌20.102112 https://doi.org/10.1080/17425247.2021.1829591 https://doi.org/‌10.1016/j.jphotobiol.2020.111999 https://doi.org/‌10.1016/j.jphotobiol.2020.111999 https://doi.org/10.15171/jlms.2018.29 https://doi.org/10.1161/circresaha.120.317015 https://doi.org/10.3390/molecules25225239 https://doi.org/10.3390/ph12040148 https://doi.org/10.1016/j.ccr.2020.213233 https://doi.org/10.3390/nano10112276 https://doi.org/10.1016/j.bioma‌terials.20‌11.11.033 https://doi.org/10.1016/j.bioma‌terials.20‌11.11.033 https://doi.org/10.3390/pharmaceutics13121995 https://doi.org/10.1088/0957-4484/23/10/105101 https://doi.org/10.1088/0957-4484/23/10/105101 https://doi.org/10.2174/1570180819666220405225817 https://doi.org/10.1016/j.ajps.2020.04.002 admet & dmpk 00(0) (2023) 000-000 infections associated with sars-cov-2 exploited via photodynamic therapy doi: https://doi.org/10.5599/admet.1883 19 [89] v. fedorov, e. kholina, s. khruschev, i. kovalenko, a. rubin, m. strakhovskaya. electrostatic map of the sars-cov-2 virion specifies binding sites of the antiviral cationic photosensitizer. international journal of molecular sciences 23 (2022) 7304. https://doi.org/10.3390/ijms23137304. [90] k. sharshov, m. solomatina, o. kurskaya, i. kovalenko, e. kholina, v. fedorov, g. meerovich, a. rubin, m. strakhovskaya. the photosensitizer octakis (cholinyl) zinc phthalocyanine with ability to bind to a model spike protein leads to a loss of sars-cov-2 infectivity in vitro when exposed to far-red led. viruses 13 (2021) 643. https://doi.org/10.3390/v13040643. [91] x. wei, w.-b. cui, f.-y. sun, h. li, j.-f. guo, x.-l. hao, a.-m. ren. photophysical properties of pt (ii) and pd (ii) complexes for two-photon photodynamic therapy: a computational investigation. dyes and pigments 215 (2023) 111283. https://doi.org/10.1016/j.dyepig.2023.111283. [92] j. shin, d.w. kang, j.h. lim, j.m. an, y. kim, j.h. kim, m.s. ji, s. park, d. kim, j.y. lee. wavelength engineerable porous organic polymer photosensitizers with protonation triggered ros generation. nature communications 14 (2023) 1498. https://doi.org/10.1038/s41467-023-37156-x. [93] bk. kundu, g. han, y. sun. derivatized benzothiazoles as two-photon-absorbing organic photosensitizers active under near infrared light irradiation. journal of the american chemical society 145 (2023) 3535-3542. https://doi.org/10.1021/jacs.2c12244. [94] r. prieto-montero, a.d. andres, a. prieto-castañeda, a. tabero, a. longarte, a.r. agarrabeitia, a. villanueva, m.j. ortiz, r. montero, d. casanova. halogen-free photosensitizers based on mesoenamine-bodipys for bioimaging and photodynamic therapy. journal of materials chemistry b 11 (2023) 169-179. https://doi.org/10.1039/d2tb01515c. [95] m. pourhajibagher, a. bahador. virtual screening and computational simulation analysis of antimicrobial photodynamic therapy using propolis-benzofuran a to control of monkeypox. photodiagnosis and photodynamic therapy 41 (2023) 103208. https://doi.org/10.1016/ j.pdpdt.2022.103208. [96] g. alachouzos, a.m. schulte, a. mondal, w. szymanski, b.l. feringa. computational design, synthesis, and photochemistry of cy7‐ppg, an efficient nir‐activated photolabile protecting group for therapeutic applications. angewandte chemie 134 (2022) e202201308. https://doi.org/10.1002/ ange.202201308. [97] ns. kuzmina, v.f. otvagin, a.a. maleev, m.a. urazaeva, a.v. nyuchev, s.k. ignatov, a.e. gavryushin, a.y. fedorov. development of novel porphyrin/combretastatin a-4 conjugates for bimodal chemo and photodynamic therapy: synthesis, photophysical and tddft computational studies. journal of photochemistry and photobiology a: chemistry 433 (2022) 114138. https://doi.org/10.1016/j.jphoto chem.2022.114138. [98] e. kilic, z. elmazoglu, t. almammadov, d. kepil, t. etienne, a. marion, g. gunbas, s. kolemen. activitybased photosensitizers with optimized triplet state characteristics toward cancer cell selective and image guided photodynamic therapy. acs applied bio materials 5 (2022) 2754-2767. https://doi.org/10.1021/acsabm.2c00202. [99] a. saedi, a. mashinchian moradi, s. kimiagar, h.a. panahi. photosensitization of fucoxanthingraphene complexes: a computational approach. main group chemistry 21 (2022) 1065-1075. https://doi.org/10.3233/mgc-210188. [100] x. wang, h. lv, y. sun, g. zu, x. zhang, y. song, f. zhao, j. wang. new porphyrin photosensitizers— synthesis, singlet oxygen yield, photophysical properties and application in pdt. spectrochimica acta part a: molecular and biomolecular spectroscopy 279 (2022) 121447. https://doi.org/10.1016/ j.saa.2022.121447. ©2023 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) https://doi.org/10.5599/admet.1883 https://doi.org/10.3390/ijms23137304 https://doi.org/10.3390/v13040643 https://doi.org/10.1016/j.dyepig.2023.111283 https://doi.org/10.1038/s41467-023-37156-x https://doi.org/10.1021/jacs.2c12244 https://doi.org/10.1039/d2tb01515c https://doi.org/10.1016/j.pdpdt.2022.103208 https://doi.org/10.1016/j.pdpdt.2022.103208 https://doi.org/10.1002/‌ange.202201308 https://doi.org/10.1002/‌ange.202201308 https://doi.org/10.1016/‌j.jphoto‌chem.2022.114138 https://doi.org/10.1016/‌j.jphoto‌chem.2022.114138 https://doi.org/10.1021/acsabm.2c00202 https://doi.org/10.3233/mgc-210188 https://doi.org/10.1016/‌j.saa.2022.121447 https://doi.org/10.1016/‌j.saa.2022.121447 http://creativecommons.org/licenses/by/3.0/ doi: 10.5599/admet.4.3.328 269 admet & dmpk 4(3) (2016) 269-279; doi: 10.5599/admet.4.3.328 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper evaluation of pharmacokinetic and pharmacodynamic interaction between repaglinide and atazanavir in healthy, diabetic and hepatic impaired rats: possible inhibition of cyp3a, oatp, and p-glycoprotein transporters thirumaleswara goud* 1 , srinivas maddi 2 , nayakanti devanna 3 , thatipamula rajendra prasad 4 1 creative educational society’s college of pharmacy, chinnatekur, kurnool, andhra pradesh, india 2 gvk bio sciences, hyderabad, telangana, india 3 department of chemistry, jntu anantapur, anantapuramu, andhra pradesh, india 4 jawaharlal nehru technological university, anantapur, anantapuramu, andhra pradesh, india. *corresponding author. e-mail: thirumalram@gmail.com; tel.: +919705058169 received: july 14, 2016; revised: september 11, 2016; published: september 30, 2016 abstract the metabolic syndrome in hiv infected patients is particularly associated with the use protease inhibitors. atazanavir is an inhibitor of the cytochrome p 450 (cyp) system, in particular cyp3a4 and cyp2c9 which can affect the metabolism of several drugs. to treat metabolic syndrome in hiv patients repaglinide is used and it is a short acting insulin secretagogues undergoing metabolism with cyp 3a4 and cyp 2c8 enzyme system. the purpose of this study was to assess the possible pharmacokinetic and pharmacodynamic drug interaction of repaglinide and atazanavir in healthy, diabetic and impaired hepatic function rats. human oral therapeutic doses of atazanavir and repaglinide were extrapolated to rats based on the body surface area. the pharmacokinetic parameters and blood glucose concentrations of repaglinide were determined after oral administration of repaglinide alone (0.5 mg/kg) and in the presence of atazanavir (36 mg/kg) in normal, diabetic and hepatic impaired rats. the pharmacokinetics (pk) and blood glucose concentrations of repaglinide were significantly altered in the presence of atazanavir. the peak plasma concentration (c max), area under the plasma concentration time profile (auc) and elimination half-life of repaglinide were significantly (p<0.0001) increased. the repaglinide clearance (cl) was significantly (p<0.0001) decreased in the presence of atazanavir treatment. in the presence of atazanavir, repaglinide hypoglycaemic activity was increased significantly (p<0.0001) when compared with the repaglinide control group. the present study demonstrated the significant difference in the pk/pd changes due to the enhanced bioavailability and decreased total body clearance of repaglinide may be due to the inhibition of the cyp p450 metabolic system, oatp and p-gp transporters by atazanavir. keywords aids; drug drug interaction; hypoglycaemic activity; liver dysfunction; peak plasma concentration cmax; area under the curve auc; clearance; elimination half life. introduction the prevalence and incidence of metabolic syndrome, particularly diabetes mellitus, was increasing in hiv patients. treatment of hiv-infected patients with hiv-1 protease inhibitors (pis) as part of highly active http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:thirumalram@gmail.com thirumaleswara goud et al. admet & dmpk 4(3) (2016) 269-279 270 antiretroviral therapy (haart) has contributed considerable reductions in hiv viral load and increased in cd4 lymphocyte numbers, thereby slowing disease progression and improving patient survival. however, despite this clinical success, it is recognized that pi-based therapy is correlated with a number of significant metabolic complications, including lipodystrophy, hyperlipidaemia, and insulin resistance. around 80 % of patients who receive pis develop insulin resistance leading to diabetes mellitus [1]. drugs are primarily eliminated from the body by metabolism and excretion. in these two biologic processes, the liver plays a vital role in the metabolism of parent drugs and formation of metabolites prior to their excretion by the kidneys. the overall capacity of the liver to carry out its metabolic role is primarily dependent on three factors: activity of the metabolizing enzymes within the smooth endoplasmic reticulum and the cytosol of the hepatocytes; the degree of protein binding in the blood, which affects the amount of unbound drugs available for uptake into the hepatocytes; and liver blood flow, which delivers drugs to the hepatocytes via the portal vein for orally administered drugs and via the systemic circulation for all administered drugs. in hepatic impairment, patient-specific factors that affect enzymatic activity, protein binding, or liver blood flow would potentially result in significant alterations in drug disposition and therapeutic response. an understanding of the pharmacokinetic basis of hepatic drug elimination is helpful to conceptualize and quantify altered drug disposition in patients with liver dysfunction [2]. in hepatic dysfunction, the disruption of the liver vascular architecture may lead to increased blood flow resistance, which limits blood flow through the liver and causes portal vein pressure to rise. as a consequence, the formation of portocaval shunts may occur that allow the drug to bypass the first pass in the liver, which increase the systemic bioavailability of drugs. chronic hepatic disease causes damage to hepatocytes; this in turn may cause a decreased intrinsic clearance of the drug metabolizing liver enzymes. different cyp may be affected differently by the hepatocytes damage. in addition, the damage may also be different in the different regions of the liver. cholestasis will impair the elimination of drugs that are excreted in the bile. liver dysfunction causes a decrease of albumin in serum, which implies variation of the binding of drugs to the circulating proteins and can potentially affect the distribution volume of certain drugs [3]. pharmacotherapy in patients with impaired hepatic function is an important and difficult aspect of medical practice. the elimination of many drugs takes place by inactive, active, or toxic metabolites may be significantly altered by hepatic dysfunction and dosage must be evaluated carefully. the presence of liver dysfunction that leads to increased plasma drug or metabolite concentrations may increase the prospect of toxicity which results in more complex drug-drug interaction. hence, the preferred approach to prescribing medications for hepatic impaired patients is to select drugs that have predictable pharmacokinetic properties, are minimally affected by hepatic impairment and have a low potential for drug interactions with other drugs [4]. safe pharmacological treatment of these complications requires an understanding of the drug-drug interactions between antiretroviral drugs and the drugs used in the treatment of diabetes [5]. repaglinide is short acting anti diabetic drug used to normalize postprandial glucose concentrations in patients with type ii diabetes mellitus [6]. repaglinide is a substrate of the influx transporter oatp1b1 and metabolized mainly by cyp3a4 and cyp2c8. pharmacokinetic drug interactions were observed in repaglinide with drugs which are inhibitors of cyp3a4, cyp2c8 and oatp1b1. recently, a number of ddis caused by the inhibition and induction of drug transporters have also been reported. among these transporters, oatp1b1 is responsible for the influx of many therapeutic drugs in the liver; there have been a large number of reports on ddi caused by inhibition and induction of this transporter [7-8]. protease admet & dmpk 4(3) (2016) 269-279 pk and pd interaction between repaglinide and atazanavir doi: 10.5599/admet.4.3.328 271 inhibitors are the potent inhibitors of cyp 3a4 and oatp transporters. several ddis were noticed when protease inhibitors (pis) co-administered with drugs metabolized by cyp 3a4. repaglinide also has an affinity for p-gp and it can contribute significantly to potential drug-drug interactions with other p-gp substrates or inhibitors, hence the co-administration of repaglinide with the known p-gp inhibitor, cyclosporine a resulted in a significant increase in the plasma concentrations of repaglinide in human [9]. the protease inhibitors (pis) are also potent mechanism based inhibitors, of which atazanavir is most potent substrate and potent inhibitor of the cytochrome p450 (cyp) system, in particular cyp3a4 and cyp2c9 and affect the metabolism of several drugs [10]. atazanavir is an azapeptide hiv-i protease inhibitor approved by fda for the combination of hiv-i infection. atazanavir is metabolized by cyp3a4, as a substrate; however atazanavir interacts with drug transporter proteins. the study by mothanje barabara lucia et al indicated that atazanavir is a possible inhibitor for p-glycoprotein and competitively inhibits their efflux activity in a dose dependent manner [11]. in the present study, we hypothesize that the repaglinide elimination would be effected by oatp and pgp transporters, which makes available for the metabolism by cyp3a4. atazanavir can inhibit cyp 3a4, cyp2c9 enzymes and it is oatp, p-gp inhibitor. hence, the inhibition of transporter mediated hepatic uptake and metabolism may change the pharmacokinetics of repaglinide. however, there seem to be no published studies reported so far, so in the current study we have evaluated the influence of atazanavir on the pharmacokinetics and pharmacodyamics of repaglinide in normal, diabetic and hepatic impaired rats. experimental drugs and chemicals repaglinide a gift sample obtained from dr. reddy’s lab (hyderabad, india) and atazanavir from aurobindo labs (hyderabad, india). alloxan monohydrate was purchased from sigma aldrich, bangalore (bangalore, india). glucose kits purchased from agappe diagnostics, (mumbai, india). acetonitrile hplc grade and formic acid were obtained from merck chemicals (mumbai, india). hplc system: agilent technologies (california, usa) and ms/ms api-3200 mass spectrometer sciex technologies (foster city, ca, usa), hplc column: hypersil gold c18, 5µ column 4.6 * 100 mm internal diameter thermo scientific, (north america). vortex mixer and centrifuge were obtained from remi laboratory instruments (mumbai, india). hematocrit heparinized rat bleeding capillaries was purchased from top-tech lab equipments pvt ltd (hyderabad). pharmacokinetics and pharmacodynamic interaction study in normal rats male albino wistar rats weighing 200-250 gm were obtained from the mahaveera enterprises (hyderabad, india). they were housed under standard conditions and were maintained under a 12 hour light/dark cycle in the laboratory animal recourses facilities, creative educational society’s college of pharmacy, kurnool, india. rats were fasted overnight before dosing and until 4 hours after dosing. water was allowed ad libitum during the fasting period. the experiments were performed after prior approval of the study protocol by the institutional animal ethics committee. the study was conducted in accordance with the guidelines provided by the committee for the purpose of control and supervision of experiments and animals (cpcsea). rats were divided into 5 groups (n=6, each): group-1: repaglinide oral control group administered 0.5 mg/kg [12], group-2: atazanavir oral control group administered 36 mg/kg [13], thirumaleswara goud et al. admet & dmpk 4(3) (2016) 269-279 272 group-3: atazanavir (36 mg/kg) administered, followed by repaglinide (0.5 mg/kg) to normal rats, group-4: atazanavir (36 mg/kg) administered, followed by repaglinide (0.5 mg/kg) to diabetic induced rats, group-5: atazanavir (36 mg/kg) administered, followed by repaglinide (0.5 mg/kg) to a hepatic impaired rats. required amount of drugs (repaglinide and atazanavir) were weighed and placed in two separate motor’s triturated with pestle, to this required amount of tween 80 was added as a wetting agent and triturated well until the whole compound was wet, then 0.5 % of methyl cellulose was added in gravimetric dilution method and triturated uniformly. this suspension was administered to the respective group of animals through oral gavage [14]. blood samples collected retro-orbital puncture at 0, 0.25, 0.5, 1, 2, 4, 6, 8, 12 and 24 h [15]. plasma was separated by centrifugation using remi research centrifuge at 4000 rpm for 10-20 min. blood glucose levels were determined using glucose oxidase peroxidase (god-pod) method by measuring optical density spectrophotometrically at 510 nm [16] and the remaining samples were stored in vials at -80 °c until lc/msms analysis. pharmacokinetics and pharmacodynamic interaction study in diabetic rats diabetes was induced in rats by the administration of alloxan monohydrate in ice cold normal saline, in two doses, that is 100mg and 50 mg/kg body wt intraperitoneally for two consecutive days [17]. after 72 h, the samples were collected from rats by retro-orbital puncture of all surviving animals and the plasma was analyzed for glucose levels. rats with blood glucose levels of 200 mg/dl and above were considered as diabetic and selected for the study. similarly a separate set of one group (n=6) was used for the pharmacodynamic interaction study with pretreated atazanavir (36 mg/kg) followed by 0.5 mg/kg repaglinide in diabetic rats. pharmacokinetics and pharmacodynamic interaction study in hepatic impaired rats hepatic impairment was induced in rats by administration of carbon tetrachloride in olive oil (2ml/kg) intraperitoneally for one day [18]. after 24 h, blood samples were collected from rats by retro-orbital puncture of all surviving animals and the serum was analysed for total bilirubin (>2mg/dl), alanine transaminase (>150 mg/dl), aspartate transaminase (>200mg/dl) and albumin (<3 mg/dl) were selected for the study [19]. similarly, a separate set of one group (n=6) was used for the pharmacodynamic interaction study with pretreated atazanavir (36 mg/kg) followed by 0.5 mg/kg repaglinide in hepatic impaired rats. liquid chromatography-mass spectrometry analysis in vivo samples were prepared by protein precipitation with the addition of 200 µl acetonitrile containing an internal standard (rosuvastatin) to 50 µl sample volume. samples were vortexed and centrifuged at 4000rpm for 10 min prior to evaporation of the supernatant. atazanavir and repaglinide concentrations from plasma samples were measured using validated liquid chromatography/mass spectrometry (lc-ms/ms) method. positive-ion multiple reaction monitoring was used for the tandem mass spectrometric detection of repaglinide, atazanavir and rosuvastatin. the selected [m+h] + precursor ions were m/z 453.2 for repaglinide, m/z 705.5 for atazanavir, and m/z 482.3 for rosuvastatin, and the product ions monitored were at m/z 230.3, 168.2 and 258.15 for repaglinide, atazanavir and rosuvastatin (internal standard), respectively. the high-performance liquid chromatography column was a hypersil gold c18 (4.6x100mm,5µ) maintained at 40 °c with a flow rate of 1.3 ml/min. mobile phase consisted of 0.1 % formic acid in milliq water (a) and acetonitrile (b). a gradient elution was used. the overall run time was 6.0 min. an aliquot of 5 μl was injected at a mobile phase flow rate of 1300 μl/min. the liquid admet & dmpk 4(3) (2016) 269-279 pk and pd interaction between repaglinide and atazanavir doi: 10.5599/admet.4.3.328 273 chromatography was interfaced to api 3200 (absciex, ca) ion-trap mass spectrometer operated in the positive ion electrospray and full tandem mass spectrometry mode. before analyzing plasma, atazanavir and repaglinide standard solutions prepared in acetonitrile were run on lc to optimize the analysis and standardize calibration. standard concentrations were chosen on the basis of levels found in plasma samples. the peak area ratio of analyte (repaglinide, atazanavir) to internal standard was plotted against analyte concentrations, and standard curves were fitted by weighted (1/x 2 ) least-squares linear regression in the concentration of 9-20000 ng/ml for repaglinide and 19.5–20000 ng/ml for atazanavir. a correlation of 0.994 was desirable for all the calibration curves. the limit of quantitation for the purposes of this assay was 9 ng/ml and 19.5 ng/ml for repaglinide and atazanavir, respectively [20-21]. pharmacokinetic and statistical analysis the pharmacokinetic parameters were obtained by fitting the plasma concentration-time data to noncompartmental model by using (phoenix -v6.3.0.395; pharsight, mountain view, ca). the maximum plasma concentration (cmax) and the time to reach the same (tmax), the half-life of plasma drug elimination (t1/2) was the ratio of 0.693 to the slope obtained by log-linear regression of the terminal phase of the drug plasma profile and the area under the concentration-time curve to terminal time (auc0-t),area under the concentration-time curve to infinite time (auc0-∞) was calculated by the linear/log trapezoidal rule and the total plasma clearance (cl) were calculated by phoenix software. data were expressed as a mean±standard deviation (sd). the significance was determined by applying one way anova analysis. the results were considered to be statistically significant when p<0.0001. glucose reduction calculations percentage reduction in bgl = {(ibgl – fbgl) / ibgl} x 100 where bgl = blood glucose level; ibgl = initial blood glucose level; fbgl = final blood glucose level [22]. results lc-ms/ms analytical method validation the analysis of plasma samples was completed after thorough validation of lc-ms/ms method. no interfering peaks were observed in blank plasma chromatograms at atazanavir and repaglinide retention time indicating the selectivity of the present method. the calibration curves and linearity of repaglinide (19.5-20000 ng/ml) and atazanavir (19-20000 ng/ml) samples was determined using weighted (1/x 2 ) linear regression analysis. the correlation coefficients were 0.994 for calibration curves. the retention times for atazanavir, repaglinide and internal standard were 3.0, 3.5 and 3.3 min, respectively. quality control (qc) samples were prepared at low, medium, and high concentrations to assess accuracy, precision and recovery. intraday precision from a measurement of a minimum of six replicates was <7.1 % in rat plasma. the inter-day precision was determined from analysis of standard samples on three consecutive days and the relative standard deviation was found to be <6.9 % and <9 % for atazanavir and repaglinide, respectively. the overall accuracy of qc samples was in the range of 90-110 %. percentage recovery was calculated as a ratio of the peak areas of drug in the matrix and the corresponding peak area of drug in acetonitrile. in plasma samples, greater than 85 % drug recovery was accomplished for both atazanavir and repaglinide. the lower limit of quantification (loq) was determined based on intra-run accuracy of loq replicates. the tested loq values were 9 ng/ml and 19.5 ng/ml for repaglinide and atazanavir, respectively, in rat plasma. thirumaleswara goud et al. admet & dmpk 4(3) (2016) 269-279 274 effect of atazanavir on the pharmacokinetics and pharmacodynamics of repaglinide in normal rat mean plasma concentration of repaglinide in the absence and presence of atazanavir in normal rats is shown in figure 1. mean pharmacokinetic parameters of repaglinide alone and in the presence of atazanavir in normal rats were shown in table 1 (column i and ii). the pharmacokinetic parameters of repaglinide like auc, cmax, and clearance were altered significantly with single dose treatment of atazanavir in normal rats when compared to repaglinide treated group. compared to the repaglinide control group (repaglinide alone) atazanavir increased 2-fold peak plasma concentration (cmax), 11-fold area under the plasma concentration time curve (auc0-t and auc0-∞) of repaglinide and atazanavir relative decrease in the total body clearance by 6 folds. repaglinide produced a hypoglycemic effect which was indicated by percentage reduction levels in normal rats. the hypoglycemic effect of repaglinide enhanced when given in combination with atazanavir. it was indicated by a significant increase in percentage glucose reduction in comparison to repaglinide alone treated group in normal rats were shown in table 2 and figure 4. figure 1. plasma concentration-time curves of repaglinide following its oral administration (0.5 mg/kg) in control and atazanavir (36 mg/kg) pretreated normal rats. data are expressed as mean ±sd in six rats. table 1. plasma concentration time profiles of repaglinide following its oral administration at 0.5 mg/kg in control and atazanavir (36 mg/kg) pretreated normal, diabetic and hepatic impaired rats. data are expressed as mean ±s.d. in (n=6) rats. pharmacokinetic parameters repaglinide alone (control) repaglinide in the presence of atazanavir in normal rats repaglinide in the presence of atazanavir in diabetic rats repaglinide in the presence of atazanavir in hepatic impaired rats cmax (ng/ml) 536 ± 30 1089 ± 123*** 2753 ± 7 *** 3410 ± 64 *** tmax (h) 0.25 ± 0.00 4.0 ± 0.0 ** 4.0 ± 0.0 ** 0.5 ± 0.0 auc0-t (ng/ml.h) 678 ± 51 8031 ± 274*** 12602 ± 227*** 18301 ± 151*** auctotal(ng/ml.h) 726 ± 55 8044 ± 265*** 12722 ± 209 *** 19101 ± 184*** oral t1/2 (h) 3.0 ± 0.1 2.6 ± 0.1 3.5± 0.1 6.2 ± 0.3 ** oral cl (ml/min/kg) 11 ± 1 1.03 ± 0.03*** 0.44 ± 0.01*** 0.65 ± 0.01 *** *** significant at p<0.0001 compared to repaglinide control (alone) group. **significant at p<0.0001 compared to repaglinide control (alone) group admet & dmpk 4(3) (2016) 269-279 pk and pd interaction between repaglinide and atazanavir doi: 10.5599/admet.4.3.328 275 effect of atazanavir on the pharmacokinetics and pharmacodynamics of repaglinide in diabetic rats mean plasma concentration of repaglinide in the absence and presence of atazanavir in diabetic rats is shown in figures 2 & 3. mean pharmacokinetic parameters of repaglinide and atazanavir in diabetic rats are shown in table 1 (column i and iii). the pharmacokinetic parameters of repaglinide like auc, cmax, and clearance were altered significantly with single dose treatment of atazanavir in diabetic rats when compared to repaglinide treated group. compared to the repaglinide control group (repaglinide alone) atazanavir increased 5-fold peak plasma concentration (cmax), 18-fold area under the plasma concentration time curve (auc0-∞) of repaglinide and atazanavir relatively decreases in the total body clearance by 25fold. repaglinide produced a hypoglycemic effect indicated by percentage reduction levels in diabetic rats. the hypoglycemic effect of repaglinide exaggerated when given in combination with atazanavir, it was indicated by a significant increase in percentage glucose reduction in comparison to repaglinide alone treated group in normal rats were shown in table 2 and figure 4. table 2. mean percentage blood glucose reduction of repaglinide following its oral administration at 0.5mg/kg in control and atazanavir (36 mg/kg) pretreated normal, diabetic and hepatic impaired rats. data are expressed as mean ±s.d. in (n=6) rats. repaglinide alone (control) repaglinide in the presence of atazanavir in normal rats repaglinide in the presence of atazanavir in diabetic rats repaglinide in the presence of atazanavir in hepatic impaired rats time, h mean±sd mean±sd mean±sd mean±sd 0.25 37.07±3.80 17.70±5.91 15.76±4.77*** 43.95±2.32*** 0.50 25.12±5.84 37.21±22.51* 33.27±24.79*** 64.62±3.31*** 1 19.76±5.99 43.12±15.56** 28.25±2.38*** 57.15±2.04*** 2 7.03±5.02 44.50±6.84** 53.88±2.93*** 46.54±1.72*** 4 9.63±3.65 59.22±7.72** 65.89±1.27*** 38.45±5.03*** 6 12.57±5.11 47.72±5.50** 31.13±3.25*** 23.54±4.18*** 8 3.47±3.67 38.38±3.03** 16.86±2.29*** 14.47±15.70*** 12 2.28±6.66 27.75±6.10** 10.69±3.81*** 9.14±10.03*** 24 0.34±5.34 12.24±3.84* 4.82±3.24*** 3.71±11.59*** *** significant at p<0.0001 compared to repaglinide control (alone) group. ** significant at p<0.005 compared to repaglinide (alone) control group. effect of atazanavir on the pharmacokinetics and pharmacodynamics of repaglinide in hepatic rats mean plasma concentration of repaglinide in the absence and presence of atazanavir in hepatic impaired rats has shown in figures 2 & 3. mean pharmacokinetic parameters of repaglinide in the presence of atazanavir in hepatic impaired rats were shown in table 1 (column i and iv). the pharmacokinetic parameters of repaglinide like auc, cmax, t1/2, clearance were altered significantly by a single treatment of atazanavir in normal rats when compared to repaglinide treated group. in a single dose study with atazanavir, increases in pharmacokinetic parameters of repaglinide are as follows: auc (26-fold), cmax (6 thirumaleswara goud et al. admet & dmpk 4(3) (2016) 269-279 276 fold) with a relative decrease in total body clearance of 16-fold. the hypoglycemic effect of repaglinide enhanced when given in combination with atazanavir, it was indicated by a significant increase in percentage glucose reduction in comparison to repaglinide alone treated group in normal rats (see table 2 and figure 4). figure 2. plasma concentration time curves of repaglinide following its oral administration at 0.5 mg/kg in control and atazanavir (36 mg/kg) pretreated a) diabetic and b) hepatic impaired rats. data are expressed as mean ±s.d. in (n=6) rats. figure 3. plasma concentration time curves of repaglinide following its oral administration at 0.5 mg/kg in control and atazanavir (36 mg/kg) pretreated normal, diabetic and hepatic impaired rats. data are expressed as mean ±s.d. in (n=6) rats. discussion the most challenging aspect in treating hiv patients are to give the safe combination of drugs during the therapy because the hiv therapy regimen contains one nucleoside reverse transcriptase inhibitor, one non-nucleoside reverse transcriptase inhibitor and two protease inhibitors to reduce the viral load in the patients. the hiv patients will also have several other co-morbid conditions like kaposi’s sarcoma, lymphoma, and tuberculosis. apart from this the highly active antiretroviral is encountered with the admet & dmpk 4(3) (2016) 269-279 pk and pd interaction between repaglinide and atazanavir doi: 10.5599/admet.4.3.328 277 increase in prevalence of diabetes mellitus includes insulin resistance and glucose tolerance. the development of these metabolic syndromes like glucose and lipid disturbances presents a pharmacological challenge because of the possible pharmacokinetic drug-drug interactions associated with oral hypoglycemics and antiretroviral drugs. there are limited published data on drug interactions between antidiabetic medications and antiretroviral agents. in this investigation we studied the influence of atazanavir on the pharmacokinetics and pharmacodynamics of repaglinide at therapeutic doses in healthy, diabetic and hepatic impaired rats. the healthy rat model served to quickly identify the interaction and the diabetic and hepatic impaired model to validate the same response in the disease condition. repaglinide is extensively metabolized by the hepatic cytochrome p450 enzyme system, particularly with cyp3a4; cyp2c8 with less than 2 % of an oral dose excreted unchanged in humans. among this cyp3a4 has been identified as an important enzyme in the in-vitro metabolism of repaglinide [23]. the repaglinide pharmacokinetics was further complicated because of active hepatic uptake of repaglinide by oatp uptake transporter and it is a substrate for the oatp transporter [24]. repaglinide also act as a substrate for p-glycoprotein which can significantly contribute to potential drug-drug interactions with other p-gp substrate or inhibitors [25]. figure 4. mean percentage blood glucose reduction curves of repaglinide following its oral administration at 0.5 mg/kg in control and atazanavir (36 mg/kg) pretreated normal, diabetic and hepatic impaired rats. data are expressed as mean ±s.d. in (n=6) rats. hiv protease inhibitors have identified as a substrates, inhibitors or inducers of cyp3a4, oatp and p-gp. atazanavir is an extensively metabolized mainly by cytochrome p450 enzyme system which have capable of altering both p-gp and cyp 3a4 activity in-vitro [26]. atazanavir inhibits p-gp mediated transport and acts as a potent mechanism based inhibitor of cyp3a4 and it also inhibit the oatp transport particularly oatp1b1, oatp1b3 and oatp2b1 mediated transport in liver [27]. the pharmacokinetics (cmax, auc0-t, auct-∞, t1/2, clearance) and pharmacodynamics (percent blood glucose reduction) of repaglinide were significantly altered in the atazanavir treated normal, diabetic and hepatic impaired rats. the significant improvement of cmax and auc of repaglinide in healthy, diabetic and hepatic impaired rats could be due to cyp enzyme and p-gp mediated transport inhibition during the first pass metabolism. compared to the control group the atazanavir significantly decreased the clearance of thirumaleswara goud et al. admet & dmpk 4(3) (2016) 269-279 278 repaglinide in healthy, diabetic and hepatic impaired rats, this could be due to the inhibition of cyp enzyme and oatp transporter in the liver. the magnitude change in the exposure of repaglinide is more in hepatic impaired rats could be due to the synergistic effect of oatp inhibition by atazanavir and liver dysfunction the clearance of repaglinide was significantly decreased in normal and diabetic rats, which was due to the inhibition of cyp enzyme and oatp, p-gp transport by atazanavir. however, in hepatic impaired rats the magnitude change in the exposure of repaglinide was more, which could be due to the synergistic effect of oatp inhibition by atazanavir and liver dysfunction. in hepatic impaired rats t1/2 of repaglinide was increased significantly when compared to the repaglinide alone control group due to decrease in clearance of repaglinide. the blood glucose levels were decreased significantly when repaglinide is given in combination with atazanavir in normal, diabetic and hepatic impaired groups. the increased bioavailability (auc and cmax) of repaglinide when administered with atazanavir increases the plasma levels of repaglinide and decreases elimination which results in enhanced mean blood glucose percentage which may further precipitate hypoglycemic action. our findings were consistent with the report by shitara, et al reported gemfibrozil increased the auc of repaglinide approximately 8-fold and prolonged its t1/2 from 1.3 to 3.7 in healthy subjects [28]. gemfibrozil is a strong inhibitor of cyp2c8 in vivo mainly due to its 1-o-ß-glucuronide metabolite. gemfibrozil and its glucuronide inhibit the oatp1b1 in vitro, suggesting that the involvement of drug-drug interaction between gemfibrozil and repaglinide [28]. kajossari et al reported that the co-administration of repaglinide along with the known p-glycoprotein inhibitor cyclosporine a raised the plasma concentrations of repaglinide significantly, probably by inhibiting its cyp3a4-catalyzed biotransformation and oatp1b1mediated hepatic uptake. cyclosporine may enhance the blood glucose-lowering effect of repaglinide and increase the risk of hypoglycemia in humans [29]. conclusions in this study, atazanavir enhanced the bioavailability of repaglinide after oral administration, this enhanced bioavailability of repaglinide might be mainly due to the possible inhibition of cyp enzyme mediated metabolism and p-glycoprotein effect in the intestine and in the liver the reduced the total body clearance of repaglinide is due to cyp enzyme and oatp mediated transport inhibition. the current study has raised awareness of potential drug-drug interaction between the repaglinide and atazanavir in normal, diabetic and hepatic impaired groups. the clinical significance of this study should be further evaluated in clinical studies. references [1] j.c. koster, m.s. remedi, h. qiu, c.g. nichols, p.w. hruz, diabetes 52 (2003) 1696-1700. [2] j.t. dipiro, r.l. talbert, g.c. yee, g.r. matzke, pharmacotherapy: a pathophysiologic approach, seventh edition, mcgraw-hill education, llc. page no: 693-704. [3] m. gonzalez, l. goracci, g. cruciani, i. poggesi. expert opinion drug metabolism toxicology 10 (2014) :1397-1408. [4] g.m. ferron, r.a. preston, r.j. noveck, p. pockros, p. mayer, j. getsy, m. turner, m. abell, j. paul, clinical therapeutics 23 (2001) 1180-1192. admet & dmpk 4(3) (2016) 269-279 pk and pd interaction between repaglinide and atazanavir doi: 10.5599/admet.4.3.328 279 [5] s.k. mastan, g. chaitanya, k. r. reddy, k.e. kumar. international journal of pharmtech research 1 (2009) 613-622. [6] c.r. culy, b. jarvis. drugs 61 (2001) 1625–60. [7] a. tornio, m. niemi, p.j. neuvonen, j.t. backman. trends in pharmaceutical sciences 33 (2012) 312321. [8] t.b. bidstrup, i. bjørnsdottir, u.g. sidelmann, m. søndergård, t. kristian, t. hansen. british journal of clinical pharmacology 56 (2003) 305–314. [9] j.s. choi, i. choi, d.h. choi. pharmacological reports 65 (2013) 1422-1430. [10] e.s. perloff, s.x. duan, p.r. skolnik, d.j. greenblatt, l.l. von moltke. american society of pharmacology and experimental therapeutics 3 (2005) 764-770. [11] m.b. lucia, c. golotta, s. rutella, e. rastrelli, a. savarino, r. cauda. journal of acquired immune deficiency syndrome 39 (2005) 635-637. [12] d.-h. choi, j.-s. choi, korean journal of clinical pharmacology 21 (2011) 215-223. [13] s.k. mastan, k.e. kumar. international journal of diabetes mellitus 2 (2010) 56-60. [14] s.n. bhattachar, d.m. bender, s.a. sweetana, j.a. wesley, discovering and developing molecules with optimal drug-like properties. a.c. templeton, editor. discovery formulations: approaches and practices in early preclinical development, discovering and developing molecules with optimal drug-like properties. springer new york: aaps press, 2015 page no: 49-94. [15] v. riley, proceedings of the society for experimental biology and medicine 104 (1960)751–754. [16] p. trinder. journal of clinical pathology 22 (1969) 158–161. [17] r.e. heikkila, european journal of pharmacology 44 (1977) 191-193. [18] s. robin, k. sunil, a.c. rana, s. nidhi, international research journal of pharmacy 3 (2012) 86-95. [19] c. johnson-delaney, zoological education network (1996) 220-250. [20] l. xiao-rong, d. xiao-jian, z. yi-fan, d. jue-fang, c. xiao-yan, z. dafang, acta pharmaceutica sinica 48 (2013) 547 −553. [21] g. singh, r.s. pai, s. mustafa. journal of chromatographic separation techniques 5 (2015) 1-5. [22] p. neerati, j. gade. european journal of pharmaceutical sciences 42 (2011) 285–289. [23] p.n. van heiningen, v. hatrop, k. kramer nielsen, k.t. hansen, j.j. van lier, n.c. de merbel, b. oosterhuis, j.h.g. jonkman. european journal of clinical pharmacology 55 (1999) 521-525. [24] c. sall, j. brian houston, a. galetin. the american society for pharmacology and experimental therapeutics 40 (2012) 1279-1289. [25] c.k. lee, j.s. choi, j.s. bang. korean journal of physiology and pharmacology 17 (2013) 245-251. [26] e.s. perloff, s.x. duan, p.r. skolnik, d.j. greenblatt, l.l. von moltke. the american society for pharmacology and experimental therapeutics 33 (2005) 764-770. [27] l. griffin, p. annaert, k.l.r. brower. journal of pharmaceutical sciences 100 (2011) 3636-3654. [28] y. shitara, m. hirano, h. sato, y. sugiyama, journal of pharmacology and experimental therapeutics 311 (2004) 228–236. [29] l.i. kajosaari, m. niemi, m. neuvonen, j. laitila, p.j. neuvonen, j.t. backman. clinical pharmacology and therapeutics 78 (2005) 388-399. ©2016 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://www.ncbi.nlm.nih.gov/pubmed/12817528 http://link.springer.com/search?facet-creator=%22allen+c+templeton%22 http://creativecommons.org/licenses/by/3.0/ manuscript doi: 10.5599/admet.4.4.331 280 admet & dmpk 4(4) (2016) 280-301; doi: 10.5599/admet.4.4.331 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index review rock inhibitors in ocular disease éva halász, ellen townes-anderson* rutgers new jersey medical school, department of pharmacology, physiology and neuroscience, 185. s. orange ave, newark, new jersey, 07103 *ellen townes-anderson: e-mail: andersel@njms.rutgers.edu ; tel.: +1-973-972-7392; fax: +1-973-972-5059 received: july 29, 2016; revised: november 30, 2016; published: december 26, 2016 abstract rho kinases (rocks) have a crucial role in actin-cytoskeletal reorganization and thus are involved in broad aspects of cell motility, from smooth muscle contraction to neurite outgrowth. the first marketed rock inhibitor, called fasudil, has been used safely for treatment of cerebral vasospasm since 1995 in japan. during the succeeding decades rock inhibitors have been applied in many pathological conditions from central nervous system disorders to cardiovascular disease as potential therapeutic agents or experimental tools to help understand the underlying (patho)mechanisms. in 2014, a fasudil derivate named ripasudil was accepted for clinical use in glaucoma and ocular hypertension. since rock kinases are widely expressed in ocular tissues, they have been implicated in the pathology of many ocular conditions such as corneal dysfunction, glaucoma, cataract, diabetic retinopathy, age-related macular degeneration, and retinal detachment. this paper aims to provide an overview of the most recent status/application of rock inhibitors in the field of eye disease. keywords rho kinase; corneal endothelial wound healing; glaucoma; diabetic retinopathy; retinal detachment introduction in the past few years, rock inhibitors have undergone considerable structural, physico-chemical and pharmacokinetic development. a summary from last year listed more than 170 different compounds used in various research areas from cancer to respiratory disease [1], and since then the number of substances has only increased. in the case of ocular disease, ripasudil was approved in 2014 for the treatment of glaucoma and ocular hypertension (oht) [2]. there are other promising candidates for ocular disease in phase ii or phase iii clinical trials; however, a number of drugs have also already failed [1,3] (see table 1). most rock inhibitors target the atp-binding site of the kinase in its active conformation and thereby reversibly compete with atp [4]. the classic rock inhibitors are isoquinoline-derived fasudil or pyridinebased y27632 or y39983. fasudil’s specificity for rock was reported to be less than y27632, however both compounds also inhibit other kinases including protein kinase c, protein kinase a, and myosin light chain kinase (mlck) [5]. because of these broader interactions, and especially in the case of in vivo studies using higher concentrations, it is not always possible to determine whether the results obtained are due to the on-target or the off-target effects. http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:andersel@njms.rutgers.edu admet & dmpk 4(4) (2016) 280-301 rock inhibitors in ocular disease doi: 10.5599/admet.4.4.331 281 table 1. potential rock inhibitors for human ocular disease (n/a: not available). compound synonyms structure company ic50 values and [ref.] therapeutic application clinical trials fasudil* ha-1077 asahi kasei rock i: 330 nm [5] rock ii: 330 nm [1] cerebral vasospasm approved (china, japan) raynaud syndrome; scleroderma phase iii (nct00498615) atherosclerosis hypercholesterolemia phase ii (nct00120718) amyotrophic lateral sclerosis phase ii (nct01935518) diabetic macular edema phase iii (nct01823081) ripasudil* k-115 kowa company rock i: 51 nm [6] rock ii: 19 nm [6] glaucoma approved (japan) diabetic retinopathy phase ii (japan, cti:142456) chronic obstructive pulmonary disease (nct02219360) y27632* university of kyoto rock i: 140 nm [5] rock ii: 54 nm [8] corneal endothelial disorders phase i/ii (japan) y39983* rki-983, snj-1656 senju/novartis rock i: n/a rock ii: 3.6 nm [1,8] glaucoma phase ii (discontinued) ar13324 rhopressa aerie pharmaceuticals rock i-ii: n/a glaucoma phase iii pg324 roclatan (ar13324+ latanoprost) n/a aerie pharmaceuticals rock i-ii: n/a glaucoma phase iii ar12286* aerie pharmaceuticals rock i-ii:n/a glaucoma phase ii (discontinued) pg286 (ar12286+ travoprost) n/a aerie pharmaceuticals rock i-ii: n/a glaucoma (discontinued) ar13154* n/a aerie pharmaceuticals rock i-ii:n/a choroidea neovascularisation animal model ama0076 structure not disclosed amakem rock i: 3.7 nm [8] rock ii: 2.3 nm [8] glaucoma phase ii ama0428* structure not disclosed amakem rock i-ii:n/a amd animal model ats907 altheos rock i: 36 nm [1] rock ii: 37 nm [1] glaucoma phase ii (discontinued) ins117548 n/a inspire /merck rock i: 5 nm [1] rock ii: 14 nm [1] glaucoma phase i (discontinued) de104 n/a santen rock i-ii:n/a glaucoma phase ii (discontinued) kd025* kadmon rock i: >10 µm [12] rock ii: 59 nm [12] psoriasis phase ii (nct02106195) idiopathic pulmonary fibrosis phase ii (nct02688647) graft versus host disease (gvhd) phase ii (nct02841995) *these drugs are still being used for in vitro or animal research éva halász, ellen townes-anderson admet & dmpk 4(4) (2016) 280-301 282 small modifications in structure lead to remarkable changes. for example, the fasudil derivative ripasudil at 50 % of its effective concentration against rock i in cell free assays (ic50(ki)=0.051 µm) [6] has at least a 2-6 times higher binding affinity than that of other rock inhibitors such as y27632 (ic 50(ki) =0.14 µm) or fasudil (ic 50(ki) =0.33 µm) [5] at the same effective concentration. ama0076 and y39983 are both structurally related to y27632 [7]. in in vitro assays, ama0076 (ic50=2.3+/-0.9 nm) and y39983 (ic50=4.3+/-2.1 nm) showed similar on-target potency for rock ii, but they were at least ten times more potent than y27632 (ic50=54+/-23 nm). in normotensive rabbits ama0076 also was slightly more effective in reducing the fluid-based pressure of the eyeball, a measure used to diagnose glaucoma and called the intraocular pressure (iop), than y39983 [8]. stability in solution can also be disparate. both fasudil and y27632 were effective at reducing iop in animal models [6], however the stability of the solutions was poor [9,10]. the y27632-derivative y39983 was stabile in solution and shown to have 30 times more rock inhibition and to be 10 fold more effective at lowering iop than y27632 in animal models [9]. nonetheless, clinical trials with y39983 were discontinued after phase ii, probably because of treatment-related side effects [11]. fasudil-derived ripasudil proved to be safe and stabile for topical instillation and, as noted, has been approved for treatment of glaucoma [2]. one important direction for future rock inhibition development is the development of rock i and rock ii specific drugs. kd025 is an orally available, potent and highly selective small molecule rock ii inhibitor (ic 50=~60 nm/l) [12], developed by kadmon company and currently in phase ii clinical trials for chronic pulmonary fibrotic and systemic autoimmune disease (clinicaltrials.gov: nct02106195; nct02688647). bior tri-functional small molecule inhibitors are also being developed. these drugs usually are comprised of a rock inhibitor and one or more drugs with additional therapeutically relevant targets. for example, rhopressa, a bi-functional drug, is made up of a rock inhibitor and a norepinephrine transporter inhibitor [13], and roclatan, a tri-functional drug, is made up of a fixed combination of rhopressa with latanoprost [14]. both of them were developed by aerie pharmaceuticals and are in phase iii clinical trials [15]. a drug approach with less side effects has been developed by the amakem company for localized applications of rock inhibitors. in this approach, the drug is active where it is applied but once the drug enters the systemic circulation, the compounds are metabolized into non-toxic inactive metabolites through the activity of an esterase. these drugs are known as soft drugs. adding (and modifying) ester moieties to an otherwise classical rock inhibitor results in variability in drug stability, with plasma half-life times from less than 5 minutes to more than 120 minutes [7]. ama0076 lowered intraocular pressure in new zealand white rabbits with minimal hyperemia, and was more efficient than y39983 at the same concentration [8]. this soft drug was in phase ii clinical trials in 2014, however results from the trials have not yet been disclosed. another soft rock inhibitor, ama0428, has been shown to not only reduce neoangiogenesis but also to block inflammation and fibrosis in an animal model of neovascular age-related macular degeneration (amd) [16]. rock and the rho signaling pathway rocks (also called coiled-coil containing protein kinases) are small 160 kda serine/threonine kinases. they were first described as rhoa-gtp interacting proteins [17]. two isoforms of rock have been identified: rock i and rock ii. in humans the two isoforms share more than 60 % identity in their amino admet & dmpk 4(4) (2016) 280-301 rock inhibitors in ocular disease doi: 10.5599/admet.4.4.331 283 acid sequence and around 90 % homology of their kinase domain [18,19]. structurally rocks are composed of the following major parts: a kinase domain at the n terminal, which phosphorylates the protein targets, a rho-binding domain (rbd), where the upstream activator binds, and the c-terminal, which contains a pleckstrin homology (ph) and cysteine rich domain (crd). rho a, rho b and rho c are isoforms, small gtpases of the rho family, and are usually collectively known as rho [17]. rocks can be activated by the gtp-bound form of rho. binding of rho-gtp opens the loop formation of the enzyme, and the activated form then phosphorylates downstream targets [17] (figure 1). among other substrates, rock directly phosphorylates myosin light chain (mlc) and the myosin binding subunit of myosin light chain phosphatase, thereby enhancing actin-myosin-mediated contractility and promoting formation of stress fibers and focal adhesions [20-22]. lim kinases are also downstream targets of rock, and these kinases have a pivotal role in cofilin-mediated actin filament disassembly. cofilin binds to actin and enhances actin depolymerization, whereas activated lim kinases phosphorylate the cofilin and inhibit its filament binding activity [23]. rocks also phosphorylate other actin binding proteins, for instance adducin, which interacts with spectrin-actinnetworks and erm (ezrin-radixin-moesin) proteins which are cross linkers between actin filaments and membrane proteins [17,24]. taken together, rho pathway activation leads to a concerted series of events resulting in increased actin-myosin contractility and cytoskeletal change. figure 1 . regulation of rho kinase (rock) and downstream effectors. (rbd: rho-binding domain, ph: pleckstrin homology domain, crb: cysteine-rich region domain, gef: guaninenucleotide exchange factor, gap: gtpase activating proteins, gdp: guanosine diphosphate, gtp: guanosine triphosphate, mlcp: myosin light chain phosphatase, mlc: myosin light chain, limk: lim kinases, marck: myristoylated alanine-rich c kinase substrate, nf-l: neurofilament l protein, crmp2: collapsin response mediator protein-2, nhe-1: sodium/hydrogen exchanger-1, erm: ezrin-radixin-moesin) genetic deletion of rock i and ii in mice helped explore specific functions of rock i and rock ii (during development). homozygous rock i -/-mice were born in the expected mendelian ratio, however they had a developmental defect in eyelid and ventral body closure, resulting in different degrees of eye opening at birth (eob) and omphalocele [25]. in contrast, for rock ii -/mice, approximately 90 % of the embryos died in utero. the newborns that did survive were runts, albeit subsequently fertile without particular anatomical abnormality [26]. double rock i +/and rock ii +/heterozygous mice were also born with open eyelids and incomplete ventral body closure [27]. histological examination revealed impaired formation of actomyosin cables, demonstrating that both kinases promote actin bundling in vivo. however, the results suggest that the function of two enzymes are not 100% identical; in these systems, the isoforms éva halász, ellen townes-anderson admet & dmpk 4(4) (2016) 280-301 284 do not compensate for the loss of each other. both rock isomers are present in ocular tissues [28] and aberrant regulation of rock levels plays a role in the pathogenesis of glaucoma, diabetic retinopathy, and amd, to mention the most frequently investigated areas. corneal wound healing corneal endothelial cells are the innermost layer of the cornea and have a crucial role in maintaining cornea transparency. in late adulthood, the cell density of the healthy cornea endothelium is approximately 2500 cells/mm 2 , however if for some reason density falls under the critical level of 500 cells/mm 2 compensatory changes begin and the cornea exhibits edema and haziness, notably decreasing visual acuity. typical reasons for endothelial cell loss can be trauma caused by cataract surgery, pseudophakic bullous keratopathy and fuchs endothelial dystrophy. since human endothelial cells have poor in vivo proliferative potency, these conditions lead to compensatory enlargement of the remaining endothelial cells and corneal endothelial dysfunction. long term, chronic inflammation and cytokine imbalance can also cause corneal neovascularization, which affects different layers of the cornea. at present definitive clinical therapy for effective treatment of corneal endothelial dysfunction has not been developed. in advanced phases, corneal transplantation can be performed, however several problems can arise associated with the operation including primary graft failure, graft rejection or inflammation. however, in the past few years rock inhibitors have been investigated to elucidate their role and potential applicability as treatment of corneal endothelial disorders. intriguing recent findings suggest that rock inhibitors have a beneficial effect in corneal wound healing. in 2009, okumara et al. [29] showed that different concentrations of y27632 (1, 10, 33, 100 µm) enhanced the survival of monkey corneal endothelial cells, however, only the 10 µm concentration resulted in a significant increase in cell survival. these in vitro studies indicated that the optimal concentration of y27632 also promoted cell adhesion and increased proliferative capacity [29]. based on these findings, in vivo studies using topically administered y27632 were performed. in a primate endothelial injury model, artificially injured eyes were treated with 10 mm y27632 using eye drops, six times a day. after two days the difference between the treated and the non-treated eye in the size of the wound was remarkable. furthermore, after one week non-contact specular microscopy showed normal cell morphology and cell density (of 3000 cell/mm 2 ) in the treated group, in contrast with the control group, where the examined area exhibited enlarged corneal endothelial cells (cec) and a cell density of only 1500 cells/mm 2 [30]. descemet’s membrane is the basement membrane of the corneal endothelium. it can detach spontaneously, or be removed during cataract surgery, or other injuries. the importance of an intact descemet’s membrane in corneal endothelial recovery has been investigated in in vitro studies [31,32]. in a rabbit organ culture model (treated with 10 µm y27632 for 48 hours) there was no difference between the presence and absence of descemet’s membrane, and the rock inhibitor increased cec proliferation to approximately the same level in both samples [31]. in contrast, in human cec culture (treated with 10 µm y27632 for 14 days), results suggest that the presence of an intact descemet’s membrane does affect corneal endothelial wound healing [32]. soh et al. [32] observed that the maximum endothelial recovery in culture without descemet’s membrane was around 20 %, whereas with intact descemet’s membrane recovery was approximately 80 % due to increased endothelial migration. these differing results may be due to species differences or differences in the experimental models. for instance, in the case of the rabbit organ culture the epithelium was also removed, thus enhancing the access of the drug to the corneal endothelium. another approach to the treatment of corneal endothelial disorders is injection of cultured allogeneic admet & dmpk 4(4) (2016) 280-301 rock inhibitors in ocular disease doi: 10.5599/admet.4.4.331 285 corneal endothelial cells (cecs) into the anterior chamber. intracameral injection of cecs followed by facedown positioning for 3 hours was performed with or without supplementation of a rock inhibitor (y27632) in rabbit and monkey corneas with the endothelium removed but an intact descemet’s membrane. in the rabbit, intracameral injection of cecs with 100 µm y27632 (for 2 weeks) helped in the recovery of corneal transparency and decreased edema. the central corneal thickness in the y27632 supplemented group was <800 µm, whereas without drug it was >1200 µm during the 2-week period. in the monkey model 3 months after allogeneic transplantation of cecs, the cell density in the drug supported group was 3 times higher compared to the “pure” cec-injected group. in both animal models, reconstructed endothelium with y27632 treatment exhibited monolayer hexagonal cell shapes with normal expression of functionallyrelated endothelial markers. in contrast, in the absence of rock inhibitor, cells showed a stratified fibroblastic phenotype, and reduced levels of endothelial functional markers [33]. in a feline model of endothelial dysfunction, bostan et al. examined the contribution of injected cecs to the new endothelium and whether repair with rock inhibition worked through increased adherence of the injected cells, stimulation of cec migration to the injured area, or both. surprisingly, the best endothelial recovery occurred in a scenario where y27632 was injected without cecs and with limited endothelial scraping allowing the host to provide sufficient peripheral endothelial cells to migrate towards and cover the wound [34]. y27632 has not yet been approved for human application. a phase 1 clinical trial was performed in japan involving 10 healthy volunteers. this study confirmed that y27632, applied as 10 mm eye drops 6 times daily for 7 days, did not cause any systemic or local side effects [30]. eight patients who suffered from corneal deficiency were subsequently enrolled to test the feasibility of the drug in pathological conditions. after 6 months, y27632-treated participants with central corneal edema showed slight reductions in the central corneal thickness; however, for those patients with diffuse edema there was no difference between the preand post-treatment thickness. among the 8 patients, there were a few cases where the best corrected visual acuity also improved, but for most, vision remained the same [30]. a case report of lateonset fuchs corneal dystrophy of a 52-year-old patient who received 10 mm y27632 topically six times daily for one week showed recovered corneal clarity and improved visual acuity after two weeks. these positive changes were maintained up to the final examination, two years later in this case [35]. last, but not least, in 2015, okumara et al. [31] reported on 3 patients who had severe corneal endothelial damage due to cataract surgery and received y27632 eye drops. two of them received 1 mm y27632 eye drops administered 6 times daily in the first 4 months, and 4 times daily during the following 2 months; the third patient used the drug four times daily only for 3 months. the first two patients had spontaneous detachment of descemet’s membrane during surgery, which could have further impaired the healing of the wound. however, in contrast with the in vitro findings [32], in these cases the corneal endothelium was reconstructed on bare corneal stroma and the corneal clarity recovered. when the first patient was referred to the cornea clinic she was only able to count fingers with her right eye. by three months, her vision was 20/20. the third patient had an intact descemet membrane but the corneal endothelial cell density was critically low (508 cells/mm 2 ). contrary to expectations, by the third month, the cornea became clear and the visual acuity increased from the initial 20/63 to 20/25 [31]. although these human studies have a number of limitations, for example patients without treatment were not followed and the results are very preliminary, taken together they suggest that rock inhibitors may be effective in reducing corneal edema, and/or improve wound healing. another rock inhibitor, y39983, has been applied to in vitro and in vivo corneal endothelial wound healing models and seems to be more effective for corneal endothelial cell proliferation than y27632. éva halász, ellen townes-anderson admet & dmpk 4(4) (2016) 280-301 286 according to results in human endothelial cell culture, y39983 increased the percentage of proliferating cells in a dose dependent manner; application of 3 µm y39983 achieved a significant increase compared to the control group. in addition, in monkey corneal endothelial cell culture administration of only 0.3 or 3 µm y39983 resulted in the same level of cec proliferation that was achieved with 10 µm y27632. topical administration of 0.095 mm y39983 six times daily for two days in a rabbit model significantly decreased the size of the corneal endothelial wound and increased corneal endothelial cell proliferation compared to the non-treated control group [36]. given the recently reported promising outcomes with rock inhibitors in glaucoma (see the next section), okumara et al. [37] examined the applicability of ripasudil eye drops as a potential therapeutic agent in corneal endothelial injuries. in human cec cultures all concentrations of ripasudil (0.3, 1, 3, 10, 30 µm) enhanced the proliferative potential of the corneal endothelium at least the same or more effectively than 10 µm y27632 or 10 µm fasudil. in an in vivo model of corneal wounding, rabbits received 0.4% ripasudil eye drops three times daily for two weeks; five of six ripasudil-treated corneas became transparent again and the central corneal thickness decreased to a normal value, whereas the control corneas did not change. moreover, the corneal endothelial cell-density and the expression of pump function-related markers were higher in the ripasudil-treated animals compared to the vehicle-treated group. according to their findings, proliferating cells were not observed after the injured area was fully covered by cecs, which indicates that rock inhibitors do not induce cell proliferation after the endothelium is successfully reconstructed [37]. in light of the fact that the cornea is normally avascular, it can be a useful organ in the evaluation of angiogenesis. in an in vivo mouse model, it has been reported that fasudil can inhibit vegf-induced corneal neovascularization without adverse effects such as corneal edema or inflammation. however, in this experimental model the concentration of the effective dose in the cornea could not be measured precisely, because they administered the drug in the form of hydron pellets [38]. in live rabbits, alkali-burn-induced corneal fibrosis and neovascularization were also significantly reduced by fasudil (50 µl, 3 nm) applied topically twice daily for three days (corneas were harvested after two weeks) compared to the basic salinetreated control group. the histological examination confirmed the initial biomicroscopical findings: the fasudil-treated corneas exhibited significantly less staining for fibrotic markers (smooth muscle actin, fibronectin, f-actin) than the control group [39]. zeng et al. [40] also demonstrated that rock inhibitors (fasudil) can promote the healing of corneal epithelial defects and inhibit the formation of neovascularization. in their alkali burn mouse cornea model they used different concentrations of fasudil (30, 100, 300, 1000 µm eye drops, applied 4 times daily, for 14 days). although only the 100 µm dose inhibited the corneal neovascularization significantly at all time points examined compared to the control group, corneal re-epithelization was significantly accelerated by 100, 300, and 1000 µm of fasudil. according to current understanding, inflammation and reactive oxygen species (ros) contribute, at least in part, to neo-angiogenesis. in an in vivo corneal neovascularization murine model, 100 µm topical fasudil appeared to serve as a potential antioxidant because there was decreased ros activity, as well as reduced expression of angiogenesis related genes (vegf, tnf-alfa, mmp-8, mmp-9) and increased levels of protective heme oxygenase 1 (ho-1). in addition, the number of inflammatory cells in the cornea was reduced [40]. thus, rock inhibition has been shown to have effects on corneal endothelial migration and proliferation and on corneal neovascularization. admet & dmpk 4(4) (2016) 280-301 rock inhibitors in ocular disease doi: 10.5599/admet.4.4.331 287 glaucoma glaucoma is a progressive optic neuropathy characterized by ganglion cell death which leads to concomitant thinning of the retinal nerve fiber layer and visual field loss. in spite of the fact that glaucoma is the leading cause of irreversible blindness worldwide, the etiology of the disease is still not fully understood. present clinical therapeutic strategies focus on lowering intraocular pressure (iop) via pharmacologic agents or surgical procedures. rock inhibitors may have potential in glaucoma therapy in at least three ways: lowering the intraocular pressure, reducing the scarring after filtration surgery, and neuroprotection (neuroprotection is discussed in the retina section). proposed iop lowering mechanism: the cause of elevated iop levels originates from the imbalance between production of aqueous humor and drainage of aqueous humor. the aqueous humor is traditionally considered to drain out of the eye through either the conventional pathway via the trabecular meshwork (tm) or the unconventional pathway via the uveoscleral route. normally, the tm is a spongy connective tissue containing collagen and elastin fibers surrounded by trabecular cells, which display smooth-muscle like properties such as actin-myosin contractility. contraction of tm cells or ciliary muscle results in reduced aqueous humor outflow and increases the iop. activated rock takes part in regulation of ciliary muscle and trabecular meshwork cell contraction through phosphorylation of the myosin binding subunit of myosin light chain phosphatase [28] and direct phosphorylation of myosin light chain [41]. as previously described, lim kinases are downstream targets of rock, and their activation leads to stabilization of actin polymerization by phosphorylated cofilin. limk inhibition (either directly or through a rock inhibitor) is considered to induce trabecular meshwork relaxation via depolymerization of actin filaments [42]. aqueous humor in patients with primary open angle glaucoma contains elevated levels of endothelin 1 (et-1) [43] and tissue growth factor-β (tgf-β) [44] which are upstream activators of the rho-rock pathway [3,45]. rock inhibitors have been shown to act directly on primate tm and schlemm’s canal (sc), lowering the iop by increasing the outflow of the aqueous humor through the conventional pathway. in normotensive monkeys, ar13324 increased the outflow facility by 53 % and reduced the iop by more than 20 % compared to the contralateral vehicle-treated eye [46]. primate trabecular meshwork cells treated with different rock inhibitors (y27632, fasudil, ripasudil) exhibited rounding and a decreased number of actin bundles. these changes returned two hours after drug removal. all three of these rock inhibitors enhanced schlemm’s canal endothelial cell permeability probably via tight junction disruption [6]. in human tm and sc cell culture, y27632 caused reversible changes in cell shape, decreased the presence of actin stress fibers and focal adhesions, and significantly increased the schlemm’s canal monolayer permeability (figure 2). the iop lowering effect of different rock inhibitors has been examined in many other in vivo and in vitro animal models. in rabbits, a single instillation of 0.4 % ripasudil doubled the conventional outflow facility, although there was no effect on the uveoscleral route [47]. enucleated porcine eyes perfused with y27632 (10, 50, 100 µm) showed in a significant increase in aqueous humor outflow facility, and structurally the entire tm appeared to be distended and exhibited widening of extracellular spaces [41]. using optical coherence tomography in the living mouse, the real-time drug effects on tm and schlemm’s canal were observed. li at al. [48] reported that topically applied rock inhibitor (ar13324) expanded the tm and increased the cross-sectional area of the schlemm’s canal and thereby lowered the iop through increased perfusion through the conventional outflow tissues. éva halász, ellen townes-anderson admet & dmpk 4(4) (2016) 280-301 288 figure 2. proposed mechanism of rock inhibition in lowering intraocular pressure. (mlcp: myosin light chain phosphatase, mlc: myosin light chain, mlck: myosin light chain kinase, limk: lim kinases) filtration surgery: accumulating data suggests that rock inhibitors enhance the success rate of glaucoma filtration surgery through inhibition of tgf-β-induced scarring. a frequent reason for failed filtration surgery is excessive postoperative scarring, which closes off the bypass for aqueous humor. as previously mentioned, tgf-β is elevated in glaucoma. the cytokine has a crucial role in fibroblastmyofibroblast differentiation [49], which leads to fibrosis after surgery. on the basis of in vitro studies with human fibroblasts from the tenon capsule of the eye, y27632 has been proposed to interfere with postoperative bleb formation through inhibition of cell contraction and fibroblast activation (transdifferentiation to myofibroblasts) [50]. recent in vitro results indicate that ripasudil also significantly diminishes the tgf-β2-induced human conjunctival fibroblast activation, extracellular matrix expression, and collagen gel contraction [51]. honjo et al. [50] performed a sclerostomy in rabbits and demonstrated that 10 mm y27632 applied topically for seven days improved the surgical outcome by inhibiting fibrosis and decreasing collagen deposition in the surgical area compared to vehicle-treated eyes. recent status of rock inhibitors in clinical application given the fact that ripasudil is marketed in japan as a possible second-line treatment for glaucoma and ocular hypertension [2], the recent clinical trials for rock-inhibitor based anti-glaucoma drugs may be promising. in completed clinical trials, twice daily application of 0.4 % ripasudil proved to be safe and effective from the viewpoint of lowering the iop [52–54]. furthermore, an 8 week study (with four hundred and thirteen patients) pointed out that ripasudil has additional pressure-lowering effects when combined with betablockers or prostaglandin analogues [55]. although ripasudil with beta-blockers and prostaglandin analogues has an additive effect, pilocarpine (another anti-glaucoma eyedrop) seems to diminish the ripasudil iop lowering effect in mice eyes in the case of concomitant administration of ripasudil and pilocarpine [56]. a common adverse effect of rock inhibitor treatment is mild or moderate hyperemia, which resolves spontaneously in hours. hyperemia is attributed to the vasodilatatory effect of rock inhibitors on conjunctival blood vessels. in addition to the relatively harmless hyperemia, a one-year study [57] revealed that after 8 weeks of ripasudil instillation there is an occurrence of allergic blepharitis (20 %) and conjunctivitis (17 %); some patients left the study due to these adverse events. moreover, lens opacity was observed in mammals treated with high concentrations of ripasudil in preclinical studies. cataract admet & dmpk 4(4) (2016) 280-301 rock inhibitors in ocular disease doi: 10.5599/admet.4.4.331 289 progression in humans was not noted as drug-related; clinicians diagnosed lens opacity change as a natural progression of age-related senile cataract [57]. the same one-year study did not report any adverse corneal effects. nakagawa et al. [58] however, observed reversible pseudo-guttae-like corneal endothelial changes during one week twice daily topical instillation of 0.4 % ripasudil in healthy humans. one and a half hours after ripasudil administration, non-contact specular microscopy showed drug-related morphological changes in corneal endothelium, which disappeared within 6 hours. these findings were in line with an animal model [59]: rabbits exhibited indistinct corneal endothelial cell borders one hour after instillation of a rock inhibitor, similar to human subjects. these abnormalities resolved to normal within hours. using scanning and transmission electron microscopy, protrusions along the intercellular cell borders were observed, probably caused by the decreased actin-myosin contractility of corneal endothelial cells from rock inhibition. physicians should be aware that ripasudil can cause transient guttae-like formation in the cornea, in order to avoid misdiagnosing patients as having chronic corneal endothelial disease such as fuchs-endothelial dystrophy. taken together, it appears that further studies are needed to explore the possible long-term side effects of ripasudil treatment. among the ongoing clinical trials with other rock inhibitors the most promising is ar13324 (also known as netarsudil or rhopressa), developed by aerie pharmaceuticals [13]. ar13324 is a so called dual-acting drug; the rock inhibitor part diminishes the iop by increasing the outflow through the trabecular meshwork and the norepinephrine transporter (net) inhibitor decreases the production of the aqueous humor. from animal models, the iop-lowering effect of once daily 0.04 % ar13324 was certified as effective [60,46]. however, during the first human study, a 0.02 % concentration of ar13324 seemed to be the most potent and produced the best dose-response curve from the viewpoint of iop reduction [61]. eighteen healthy volunteers were involved in a phase i clinical study to evaluate the ocular and systemic safety of the drug. a once daily 0.02 % ar13324 application in the morning for eight days was without any systemic side effect. moreover, the pressure-lowering effect was maintained for more than 24 hours after the end of the treatment, which is unique. surprisingly the drug was notably less effective on day one compared to day eight [62]. two hundred and thirteen patients with primary open-angle glaucoma and ocular hypertension completed a 28-day trial with ar13324. once daily 0.02 % ar13324 was a little less effective than latanoprost (0.005 %) in pressure-lowering efficacy [63]. ar13324 (0.02 %) once and twice daily was also compared to timolol instilled twice daily in 3 month-long trials (called rocket i and rocket ii). rocket ii achieved the primary endpoint and was non-inferior to timolol application. two additional trials are in process: rocket iii is a one-year safety study in canada and rocket iv is a three month non-inferiority study comparing once daily ar13324 to twice-daily timolol with a six month follow-up safety evaluation [15]. aerie phamaceuticals has also developed a triple-action anti-glaucoma eye drop named pg324 (synonym roclatan), which is a fixed-dose combination of ar13324 (0.02%) with latanoprost (0.005 %). two hundred and ninety-two patients completed a 28-day trial, where 0.01 % and 0.02 % roclatan proved to be a clinically and statistically more efficacious and safe hypotensive agent relative to its single components [14]. there are several phase iii clinical trials with roclatan expected to begin or already enrolling patients. mercury i is a one-year safety trial and mercury ii is a three months efficacy study, both of them comparing roclatan to ar13324 and latanoprost in the united states. mercury iii will evaluate six months safety and efficacy of the new drug relative to another clinically used fixed-dose combination hypotensive eye-drop in europe [15]. ama0076, designed by amakem, is called a soft rock inhibitor (see definition in introduction), because its effects are localized thereby resulting in fewer side effects. the compound contains a carboxylic ester éva halász, ellen townes-anderson admet & dmpk 4(4) (2016) 280-301 290 group and exhibits high rock efficacy (rock ii ic50=2.5 nm) [7]. topically instilled drug, once it gets through the cornea into the aqueous humor, remains stable. however, drug on the surface of the eye (in conjunctiva and other tissues) undergoes rapid metabolic inactivation by esterase. a notable advantage of this rock inhibitor is that the degree of hyperemia is significantly less than any of the other candidates, thus improving the tolerability profile. in in vitro human trabecular meshwork cultures, ama0076 exposure temporarily altered cell shape through significant reduction in actin stress fibers and focal adhesions. in new zealand white rabbits, 0.1, 0.3 and 0.5 % eye-drop concentrations all proved to be more efficient in lowering intraocular pressure than the same concentrations of y39983 [8]. human trials were completed recently (clinicaltrials.gov identifier: nct02003547, nct01693315, nct02136940). in the first human trials, ama0076 proved to be safe and efficient in iop reduction with no significant conjunctival hyperemia associated with the treatment; results from the last, four week-long phase iia clinical study have not yet been published. despite the early favorable outcomes of animal or human studies, development of many rock inhibitors for clinical use in glaucoma have failed or been suspended. for instance, ar12286 (developed by aerie pharmaceuticals) completed a phase iia human study, and produced significant dose-dependent iop reduction [64], however no further human trials have been initiated. y39983 (also called rki-983, snj-1656) proved to be safe and an efficient hypotensive agent in rabbits [8], in monkeys [9] and in humans [65]. however, a seven day-long phase ii clinical trial with 66 patients revealed unexpected side effects related to the drug instillation such as punctate keratitis, headache, and hepatic dysfunction [11]. ins117548 (developed by inspire pharmaceuticals) was abandoned after completion of a phase i clinical trial (clinicaltrials.gov identifier: nct00767793), probably because the twice daily administration was reported to cause ocular burning and stinging in a dose-dependent manner [66]. a phase iia clinical trial was completed with ats907 (clinicaltrials.gov identifier: nct01668524), and the next trial begun (clinicaltrials.gov identifier: nct01520116). however, the company closed in february 2013, and no results from this trial have been disclosed. nonetheless, the number of rock inhibitors being tested for patient use provides hope that a new therapeutic drug will be achieved soon. retina diabetic retinopathy in the past few years, rhoa and rock has been implicated in the pathogenesis of diabetic retinopathy. the basic problem is the long-term glucose homeostasis imbalance, which leads to an accumulation of advanced glycation end products in the vessels. this accumulation leads to structural changes including vascular rigidity and hyperpermeability. the damaged vessels are not able to provide sufficient blood supply, potentially causing hypoxia. under these circumstances new vessel growth is promoted, however the structure of the new vessels is different than the normal, physiological vasculature. in vitro studies on rhesus macaque retinal endothelial cells showed that high glucose levels induced increased rhoa activity and endothelial permeability, through altered expression of tight junction proteins; both fasudil and y27632 reversed the high glucose-related changes [67]. higher rock expression accompanied by increased stiffness occurred in diabetic mouse retinal capillaries and also in high glucosetreated retinal endothelial cell cultures [68]. the stiffer high glucose-treated retinal endothelial cells also exhibited lower expression and activity of the mechanosensitive ion channel trpv4 (transient receptor potential vanilloid). rock inhibition resulted in significant recovery of trpv4 expression and activity, as admet & dmpk 4(4) (2016) 280-301 rock inhibitors in ocular disease doi: 10.5599/admet.4.4.331 291 well as enhanced endothelial no production and decreased monocyte adhesion to the endothelium [68]. in diabetes, endothelial nitric oxide synthase activity is reduced. fasudil can significantly reverse this change through increased enos phosphorylation and thereby contributes to vasodilatation and vasculoprotection [69,70]. changes in leukocyte adhesion to endothelium are a feature of the early stages of the disease. bhaskaran et al. [71] suggest that high glucose-induced endothelial monocyte adhesion contributes, at least in part, to increased stiffness of retinal endothelial cells and subendothelial matrix. human retinal endothelial cells co-cultured with activated monocytes for 24 and 72 hours significantly increased rock expression and activity along with impaired trpv4 activity and expression [71]. in vitro and intravitreal administration of fasudil in a diabetic animal model mitigated the leukocyte-induced microvascular damage through decreasing both icam-1 (intercellular adhesion molecule-1) expression on the endothelium and cd 18a/ 11b presentation on the neutrophil granulocytes [69,70]. thus rock inhibition reverses many of the changes seen in the retinal vessels in diabetes. monickaraj et al. [72] reported that patients suffering from manifest diabetic macular edema, a consequence of diabetic retinopathy, had elevated cathepsin d levels in their serum. the protease caused mechanical disruption of the endothelial barrier through increased rhoa-rock-dependent contractility in human retinal endothelial cells. application of a rock inhibitor (y27632) diminished the effect of monocyte-derived cathepsin d on endothelial cell permeability. rothschild et al. [73] hypothesized that rho-pathway activation is involved in the pathogenesis of diabetic macular edema through breakdown of the retinal pigment epithelial (rpe) cell barrier. in their diabetic rat model, rho pathway activation caused cytoskeletal remodeling (apical constriction and stress fiber formation) in rpe cells, leading to intercellular junction opening. intravitreally applied fasudil reversed the morphologic abnormalities and improved the barrier integrity. in later phases, neovascularization and proliferative vitreoretinopathy (pvr) are hallmarks of the progression of the disease. cicatricial contraction of the fibrous membrane can lead to tractional retinal detachment with severe vision loss. patients suffering from proliferative retinal disease have a significantly higher vitreal concentration of tgf-β2, than patients with non-proliferative diseases [74,75]. activated tgfβ2 has been implicated in contraction of the fibrous membrane through mlc phosphorylation mediated by rock. in an experimental pvr rabbit model, intravitreal injection of fasudil (30 µm) effectively suppressed this destructive progression, even after the proliferative membranes formed or connected to the retina [76]. furthermore, in another pvr rabbit model, multiple intravitreal injections of y27632 (targeted intraocular concentration was 50 µm) decreased the development of tractional retinal detachment (figure 3) [77]. because vascular endothelial growth factor (vegf) has been shown to play a pivotal role in macular edema and retinal neovascularization, anti-vegf therapy has been widely used for the treatment of agerelated macular degeneration and diabetic retinopathy in clinical studies. however, in many cases, satisfying visual or anatomical improvements are not achieved. moreover, the treatment has many possible adverse local and systemic effects, such as iop elevation, inflammation, hypertension, myocardial infarcts or cerebrovascular accidents. in 2008 was the first report, from in vivo mouse cornea, that fasudil had inhibitory potential in ocular angiogenesis [38]. additional evidence from human foreskin [78] and retinal microvascular endothelial cell cultures [10] has demonstrated that rhoa or rho kinases are involved in vegf-induced cytoskeletal changes and cell migration. increased rhoa activity was accompanied by elevated phosphorylation and activity of the vegf receptor-2, which might be the connection between rhoa pathways and vegf effects [78]. in 2014, in an experimental diabetes rabbit model, the effect of a éva halász, ellen townes-anderson admet & dmpk 4(4) (2016) 280-301 292 single intravitreal injection of fasudil or bevacizumab on vegf levels was evaluated. according to celik et al. [79], fasudil reduced the vegf levels, however not as efficiently as bevacizumab. there are pilot clinical studies which have evaluated the feasibility of fasudil combined with bevacizumab as an intravitreal injection in diabetic macular edema. during one short-term study, 15 patients suffering from diabetic macular edema received an intravitreal injection of 0.025 mg fasudil combined with 1.25 mg bevacizumab; the targeted intraocular concentration of the drug was 10 µm/l. one month later visual acuity and central macular thickness were measured. the best corrected visual acuity improved on average by 40 %, and the central macular thickness decreased by approximately 20 %, without any adverse complications [80] . these results confirmed findings shown in a previous smaller trial [81]. moreover, 0.4 % and 0.8 % ripasudil eye drops applied topically 3 times daily for 5 days decreased not just the neovascular area, but the avascular area also, in a hypoxia-induced diabetic retinopathy mice model. in contrast, 0.4 % fasudil eye drops did not show any effect. additional experiments which tested normal vascularization found that rock inhibitors did not affect physiological angiogenesis, although they did modify pathological vascular growth in a retinopathy model [10]. figure 3. proposed rho-rock pathway in the pathogenesis of diabetic retinopathy age-related macular degeneration (amd) age-related macular degeneration (amd) is a degenerative disorder and one of the most frequent causes of irreversible visual loss in industrialized countries. the etiology of the disease has been extensively investigated over the last decades, however we are still far from a complete understanding. the conventional classification distinguishes two different subtypes: dry amd with atrophic lesions and an exudative, wet form with neovascularization. the crucial difference between the two subtypes is the development of choroidal neovascularization (cnv), the formation of new vessels characterized by inflammation and fibrosis, in wet amd. zandi et al. [82] suggest that rocks may contribute to the amd pathomechanism through macrophage differentiation in the aging eye. their theory is supported by the following results. intravitreally injected undifferentiated m0 macrophages did not affect the area of neovascularization (cnv) in mice eyes, however m1 macrophages decreased the area of the cnv whereas m2 macrophages increased the size of the lesion. in these animals repeated intravitreal injections of fasudil (rock i and rock ii inhibitor) or kd025 (specific rock ii inhibitor) decreased the area of the cnv, however only the selective rock ii inhibitor increased markers of m1-type macrophages. taken together the authors assume that with aging the fundus cytokine environment changes stimulated by increased rock ii signaling, which may shift the admet & dmpk 4(4) (2016) 280-301 rock inhibitors in ocular disease doi: 10.5599/admet.4.4.331 293 macrophage transformation towards the m2 subtype, also called proangiogenic madams (macular degeneration associated macrophages). selective rock ii inhibitors may help to restore the macrophage balance to that found in healthy, young eyes and reduce cnv. on the other hand, in monkeys, intravitreal injection of fasudil (30 µm/l 3 times per week) still significantly decreased the thickness of and vascular leakage from the cnv membrane [82]. a novel and potent rock inhibitor ama0428, from amakem therapeutics, has been shown to not only reduce the angiogenesis, but also inhibit inflammation and fibrosis in a mouse neovascular amd model. in various human cell cultures (human umbilical vein endothelial cells=huvecs, human brain microvascular endothelial cells=hbmecs, human brain vascular pericytes =hbvps) administration of ama0428 inhibited the vegf-stimulated migration of endothelial cells (huvecs, hbmecs) and stimulated pericyte recruitment. these in vitro data support the concept that rock inhibition contributes to the maturation of blood vessels, and might be able to reduce new blood vessel formation. in a mouse laser-induced cnv model, a single intravitreal injection of ama0428 (100, 1000 ng) administered either at time 0 or day 3 significantly reduced the inflammation examined on day 5, while the vegf-2 receptor inhibitor (dc101) had no effect. fourteen days after cnv induction, repeated intravitreal injections of ama0428 (100 or 1000ng on days 0, 4, 10) had a remarkable inhibitory effect on angiogenesis and vessel leakage similar to anti-vegf (dc101) treatment. moreover, one month after the laser injury, multiple intravitreal administrations of this rock inhibitor (100 or 1000 ng, on days 0, 4, 10, 20) significantly reduced fibrosis in the mouse cnv model whereas the vegf-receptor 2 inhibitor (dc101) had no effect [16]. ar13154 is a selective multi-kinase inhibitor, which inhibits rock, janus kinase (jak) and plateletderived growth factor-ß (pdgf-ß). this compound significantly decreased neovascularization in laserinduced rat and hypoxia-induced mouse models. in rats, cnv was induced on day 0, and then ar13154 or vehicle was injected intravitreally on days 1, 4, 10; aflibercept (800 ug/ml) was used as a positive control and injected into the vitreous cavity only on the first day. ar13154 reduced the mean lesion size by 35 %, whereas the aflibercept administration achieved a 23 % reduction. in a mouse hypoxia-induced optic retinopathy model, topical administration of 0.06 % ar13154 decreased the area of neovascularization by 37 %, a level similar to the intraperitoneally applied aflibercept. the combination of the two drugs resulted in a 57 % reduction of lesion [83]. additional combinations of drugs, including selective rock ii inhibitors, may be of interest in the future. optic nerve blood flow, neuroprotection, retinal detachment besides the previously described mechanisms, clinical application of rock inhibitors has other potential benefits. for instance, rock inhibitors cause vasodilatation through smooth muscle relaxation and regulation of endothelial nitric oxide synthase (enos). impaired blood flow around the optic disc and in the posterior segment has been reported in several conditions such as glaucoma, non-arteritic anterior ischemic optic neuropathy (naion) and diabetic retinopathy. the optic disc blood flow is mainly supplied by the posterior ciliary arteries. topical and intravenously applied fasudil ameliorated the enos inhibitorinduced optic nerve blood flow impairment in rabbits, while no significant effect on normal blood flow was detected [84]. also in rabbits, rock inhibitors (y27632, y39983) showed a vasodilatatory effect on ciliary artery segments, even when the no mediated route of vasodilatation was blocked [85]. naion affects the anterior part of the optic nerve and is related to an altered blood supply of this area. intravitreal fasudil (0.025 mg/0.05 ml, calculated to be equivalent to 10 µm/l) was used in a pilot study with 13 recent-onset naion patients and seemed to facilitate a cure from the disease. more than 70 % of the patients showed significant recovery of visual acuity within one month after the injection, which is an outstanding result éva halász, ellen townes-anderson admet & dmpk 4(4) (2016) 280-301 294 compared to previous reports of other treatment options for naion. the small sample size and the lack of a control group is a limitation of this study, nonetheless, the outcomes are promising and should encourage further investigation [86]. in the posterior segment, intravitreal ripasudil (calculated intravitreal concentrations were 1 and 100 µm) in cats increased retinal blood flow and blood velocity without showing effect on vessel diameter [87]. neuroprotection: yang et al. [88] demonstrated in rats that two weeks after optic nerve injury rhoa, rock i and rock ii expression were increased. these increases were diminished by y39983 (applied intravitreally 20 µm on day 0 and 7). but in addition, the rock inhibitor enhanced the number of surviving retinal ganglion cells, thereby preventing axonal degeneration 500 µm from the crush site. in 2007, sagawa et al.[89] compared two rock inhibitors in feline retinal cultures. whereas 10 µm y39983 facilitated the extension of neurites and glial processes, y27632 had a similar effect only when the dose was increased to 100 µm. based on their in vitro results they continued the experiment in an in vivo cat optic nerve crush model in which 10 µm y39983 was injected intravitreally. not only the immediate injection, but also the repeated intravitreal administration on day seven significantly increased (at least doubled) the number of regenerated axons which were able to grow through the crush site. on the other hand, at week six, the number of axon fibers at the crush site was almost the same as that observed at week two. thus repeated injections of y39983 seemed to facilitate axon regeneration in the first two weeks after the nerve trauma; perhaps additional administration of the rock inhibitor would have been beneficial. however, reduction in activated rhoa levels [89] may reduce the protection to neurons of longer treatments. ichikawa et al. [90] conducted similar experiments comparing the effect of different doses of y27632 and fasudil in in vitro and in vivo cat optic nerve injury models. interestingly, in culture conditions both 10 µm and 30 µm fasudil or y27632 enhanced neurite extension, however 100 µm fasudil application resulted in neurite outgrowth inhibition, while 100 µm y27632 exhibited the highest number of processes. in in vivo experiments, fasudil treated eyes did not show newly regenerated axon fibers in the crush site, whereas 10 and 100 µm y27632 significantly facilitated neurite outgrowth beyond the crush site. in rabbits intravenous fasudil ameliorated optic nerve injury related degenerative molecular and morphological changes compared to the dexamethasone and saline treated group. in this experiment, nogoa (known as a nerve growth inhibitor) and caspase 3 (known to take part in apoptosis) levels were measured; expression of both factors was decreased by fasudil along with rhoa and rock ii levels. moreover, at every time point, (on day 3, 7, 14, 21) but especially after one week, the optic nerve trauma related changes, such as retinal ganglion cell nuclei degeneration and cell loss and thinning of the retina, were less severe in the fasudiltreated group [91]. in salamander retinal cultures 100 µm y27632 had different effects on photoreceptors depending on cell type [92]. for rods, it inhibited axonal retraction and slightly enhanced the rod process outgrowth; but for cones, it significantly increased the number of new process and the development of presynaptic varicosities. temporary or permanent ischemia or trauma to the optic nerve leads to oxidative stress with elevated levels of reactive oxygen species contributing to neurodegeneration. yamamoto et al. [93] described, in a mouse optic nerve crush model using orally administered (1 mg/kg/day) fasudil or ripasudil, the amelioration of retinal ganglion cell death. the specificity of ripasudil for rock was found to be 2-18 times higher than fasudil. on the basis of their results, they proposed a pathway leading to retinal ganglion cell death prevention which started with suppression of nox1 (nadph oxidase) expression by rock inhibition and resulted in decreased synthesis of reactive oxygen species, and thereby, neuroprotection. retinal detachment: in the above examples the purpose of using rock inhibitors was to increase the admet & dmpk 4(4) (2016) 280-301 rock inhibitors in ocular disease doi: 10.5599/admet.4.4.331 295 plasticity and growth of axons after the injury. for retinal detachment the goal is to prevent structural remodeling and plasticity in order to decrease disruption of synapses [94]. photoreceptor degeneration was shown to be one of the early changes after detachment in the feline retina [95]. however, all photoreceptors did not respond in the same way. while rods rapidly started to retract their synaptic terminals towards their cell bodies thereby breaking their synaptic connection to horizontal and bipolar cells, cones exhibited degenerative signs later. cones did not show retraction [95], but instead changed the shape of their terminals [94]. similar changes were described in human retina [96]. disruption of the first synapse in the visual pathway between the photoreceptors and the bipolar cells may be one reason for the imperfect visual recovery seen after successful reattachment surgery. rhoa was shown to be present in the inner and the outer retina of salamanders [97], and pigs [94]. from both in vitro and in vivo experiments, the rhoa-rock pathway was shown to take part in structural remodeling of photoreceptors, including rod axon retraction and cone neuritic outgrowth. in salamander retinal cultures, rods retract their axons by 24 hours. however, rods maintained their morphology and showed no axonal retraction with 100 µm y27632 pretreatment [97] and in similar conditions with even lower concentrations (10, 30, 100 µm) in a dose-dependent manner [98]. an in vitro retinal detachment model in porcine eyes demonstrated that rhoa activity increased immediately after detachment [94,99]; recent experiments, therefore, have explored short time periods after the detachment injury. porcine retinal explants treated with 1, 10 and 100 µm of y27632 significantly reduced the number of the rod synaptic terminals which retracted into the outer nuclear layer, moreover six hour-delayed administrations of 100 µm y27632 also reduced synaptic breakage [94]. in a live pig model of retinal detachment, y27632 or fasudil was injected subretinally. one and 10 mm of y27632 and 10 mm of fasudil achieved significant prevention of rod axonal remodeling in the detached area of the treated eye compared to the detached area of the fellow non-treated eye. in this model y27632 also seemed to be more effective than fasudil: 10 mm y27632 had significant effect on prevention of synaptic changes in both detached and non-detached retinal areas, compared to the control areas from the drug-free eye, whereas fasudil prevented synaptic retraction only in the detached area of retina [99], [unpublished data]. in vitro findings also suggest that other kinases in the rhoa/rock pathway, specifically lim kinase (limk), play a role in structural remodeling of photoreceptors and could be future therapeutic targets after retinal detachment [98]. éva halász, ellen townes-anderson admet & dmpk 4(4) (2016) 280-301 296 conclusions this review has collected the publications from the last few years related to the application of rock inhibitors in various ocular diseases. their potential beneficial effects are presented in table 2. table 2. beneficial effects of rock inhibitors in various ocular disorders. cornea ↑ corneal endothelial cell survival, migration, adhesion, proliferation ↑ endothelial markers related to pump function ↓ size of corneal endothelial wound ↓ corneal edema (central corneal thickness) ↓ cornea neovascularization ↑ healing of corneal epithelial defect glaucoma ↓ iop ↑outflow facility ↑cross sectional area and permeability of sclemm’s canal ↓ fibrosis and postoperative scarring at surgical area ↑ neuroprotection of retina retina diabetic retinopathy ↓ leukocyte adhesion, endothelial permeability ↑ eno level ↓ neovascularization ↓ proliferative membrane contraction amd ↓ angiogenesis, fibrosis and inflammation optic nerve injury ↑ increasing optic nerve blood flow ↓ degenerative and morphological changes ↑ axon regeneration retinal detachment ↑∕↓ extension of neurite and glial processes ∕ neurite outgrowth ↑ photoreceptor process outgrowth ↓ rod axon retraction for corneal endothelial wound healing and glaucoma further/continued investigation is necessary to address pharmacodynamics, safety and tolerability. but already accumulating evidence from in vitro and in vivo studies shows that fasudil, ripasudil and other rock inhibitors may be promising in prevention or treatment of these and other retinal disorders (i.e., diabetic retinopathy, amd) or injuries (e.g., retinal detachment), in part because the rho/rock pathway is deeply involved in their pathogenesis. moreover, the benefit of rock inhibition may be increased by the fact that the inhibitors often target pathological change specifically and have no effect on normal processes. for instance, rock inhibition has no effects on control cornea [37], normal blood flow [84] and normal angiogenesis [10]. constitutive rock activity may be low in many ocular tissues. however, there are important unanswered questions. first of all, we need to determine the optimal therapeutic intraocular concentration of rock inhibitor. then we must consider the best and most feasible method of drug administration. because of the short biological half-life [2,7,100] of rock inhibitors, repeated applications of intravitreal injections or eye drops will be necessary. intravitreal implantation of a slowly releasing drug delivery system may offer an admet & dmpk 4(4) (2016) 280-301 rock inhibitors in ocular disease doi: 10.5599/admet.4.4.331 297 alternative solution. some inhibitors may prove more useful than others because of their tissue interactions. for instance, topically instilled 0.4 % and 0.8 % ripasudil were shown to reach the posterior segment in sufficient concentration to affect neovascularization, whereas 0.4 % fasudil did not. preparation of 0.8 % fasudil eye drops was not feasible because of poor solubility at this concentration [10]. further chemical modification of drugs such as liposomal encapsulation [101,102], or other types of conjugations [103], may be able to enhance in vivo efficacy in the future. furthermore, the timing of the treatment may be important. since anti-vegf therapy is used in advanced phases of neovascular disease, rock inhibitors may be useful to prevent structural changes in earlier stages and may affect a broader spectrum of viable targets. in most disease models we also still need to learn about the time course of rhoa and/or rock activity. in the case of our retinal detachment model we are aiming to answer similar questions. although we know rhoa activation starts quickly we do not know, for in vivo injury, when it peaks and how long activity is above normal. for instance, how many injections of a rock inhibitor will be necessary to not only prevent photoreceptor synaptic remodeling but allow restoration of normal vision after detachment and what is the optimal time-window and therapeutic dose for treatment. we also are investigating whether more than one component of the rhoa/rock pathway can be targeted simultaneously and with what effect. early studies indicate that a combination of rock and limk inhibition may be a more effective treatment than either alone in our retinal detachment model [98,104]. however, in spite of remaining questions, improvements in treatment of ocular disease with rock inhibition seem imminent. acknowledgements: the authors gratefully acknowledge support from nih grant ey021542. references [1] y. feng, p. v. lograsso, o. defert, r. li, j. med. chem. 59 (2016) 2269–2300. [2] k. p. garnock-jones, drugs 74 (2014) 2211–2215. [3] s. k. wang, r. t. chang, clin. ophthalmol. 8 (2014) 883–890. [4] t. ishizaki, m. uehata, i. tamechika, j. keel, k. nonomura, m. maekawa, et al., mol. pharmacol. 57 (2000) 976–983. [5] m. uehata, t. ishizaki, h. satoh, t. ono, t. kawahara, t. morishita, et al., nature 389 (1997) 990–4. [6] y. kaneko, m. ohta, t. inoue, k. mizuno, t. isobe, s. tanabe, et al., sci. rep. 6 (2016) 19640. [7] s. boland, o. defert, j. alen, a. bourin, k. castermans, n. kindt, et al., bioorg. med. chem. lett. 23 (2013) 6442–6446. [8] s. van de velde, t. van bergen, d. sijnave, k. hollanders, k. castermans, o. defert, et al., invest. ophthalmol. vis. sci. 55 (2014) 1006–1016. [9] h. tokushige, m. inatani, s. nemoto, h. sakaki, invest. ophthalmol. vis. sci. 48 (2007) 3216–3222. [10] m. yamaguchi, s. nakao, r. arita, y. kaizu, m. arima, y. zhou, et al., invest. ophthalmol. vis. sci. 57 (2016) 2264–2276. [11] t. inoue, h. tanihara, h. tokushige, m. araie, acta ophthalmol. 93 (2015) e393–e395. [12] j. h. lee, y. zheng, d. von bornstadt, y. wei, a. balcioglu, a. daneshmand, et al., ann. clin. transl. neurol. 1 (2014) 2–14. [13] j. m. sturdivant, s. m. royalty, c. w. lin, l. a. moore, j. d. yingling, c. l. laethem, et al., bioorganic med. chem. lett. 26 (2016) 2475–2480. [14] r. a. lewis, b. levy, n. ramirez, c. c kopczynski, d. w. usner, g. d. novack, br. j. ophthalmol. 100 (2016) 339–344. [15] j. bacharach, glaucoma today (2016) 43–45. éva halász, ellen townes-anderson admet & dmpk 4(4) (2016) 280-301 298 [16] k. hollanders, t. van bergen, n. kindt, k. castermans, d. leysen, e. vandewalle, et al., invest. ophthalmol. vis. sci. 56 (2015) 1335–1348. [17] k. riento, a. j. ridley, nat. rev. mol .cell. biol. 4 (2003) 446–456. [18] t. leung, x. q. chen, e. manser, l. lim, mol. cell. biol. 16 (1996) 5313–5327. [19] m. f. olson, curr. opin. cell biol. 20 (2008) 242–248. [20] a. p. somlyo, a. v somlyo, j. physiol. 522 pt 2 (2000) 177–85. [21] m. amano, m. ito, y. fukata, k. chihara, t. nakano, y. matsuura, et al., j. biol. chem. 271 (1996) 20246–20249. [22] k. kimura, m. ito, m. amano, k. chihara, y. fukata, m. nakafuku, et al., science 273 (1996) 245–248. [23] m. maekawa, science (80-. ). 285 (1999) 895–898. [24] m. amano, y. fukata, k. kaibuchi, exp. cell res. 261 (2000) 44–51. [25] y. shimizu, d. thumkeo, j. keel, t. ishizaki, h. oshima, m. oshima, et al., j. cell biol. 168 (2005) 941– 953. [26] d. thumkeo, j. keel, t. ishizaki, m. hirose, k. nonomura, h. oshima, et al., mol. cell. biol. 23 (2003) 5043–55. [27] d. thumkeo, y. shimizu, s. sakamoto, s. yamada, s. narumiya, genes to cells 10 (2005) 825–834. [28] c. fukiage, k. mizutani, y. kawamoto, m. azuma, t. r. shearer, biochem. biophys. res. commun. 288 (2001) 296–300. [29] n. okumura, m. ueno, n. koizumi, y. sakamoto, k. hirata, j. hamuro, et al., invest. ophthalmol. vis. sci. 50 (2009) 3680–3687. [30] n. okumura, n. koizumi, e. p. kay, m. ueno, y. sakamoto, s. nakamura, et al., invest. ophthalmol. vis. sci. 54 (2013) 2493–2502. [31] n. okumura, r. inoue, y. okazaki, s. nakano, h. nakagawa, s. kinoshita, et al., invest. ophthalmol. vis. sci. 56 (2015) 6067–6074. [32] y. q. soh, g. peh, b. l. george, x. y. seah, n. k. primalani, k. adnan, et al., invest. opthalmology vis. sci. 57 (2016) 338-348. [33] n. okumura, n. koizumi, m. ueno, y. sakamoto, h. takahashi, h. tsuchiya, et al., am. j. pathol. 181 (2012) 268–277. [34] c. bostan, m. thériault, k. j. forget, c. doyon, j. d. cameron, s. proulx, et al., invest. ophthalmol. vis. sci. 57 (2016) 1620–34. [35] n. koizumi, n. okumura, m. ueno, h. nakagawa, j. hamuro, s. kinoshita, cornea 32 (2013) 1167– 1170. [36] n. okumura, s. nakano, e. p. kay, r. numata, a. ota, y. sowa, et al., invest. ophthalmol. vis. sci. 55 (2014) 318–329. [37] n. okumura, y. okazaki, r. inoue, k. kakutani, s. nakano, s. kinoshita, et al., invest. ophthalmol. vis. sci. 57 (2016) 1284–1292. [38] y. hata, m. miura, s. nakao, s. kawahara, t. kita, t. ishibashi, jpn. j. ophthalmol. 52 (2008) 16–23. [39] m. k. fink, r. r. mohan, s. gupta, s. ebers, e. crider, m. possin, et al., rock inhibitor ha1077: potently inhibits corneal fibrosis and neovascularization. in annu. meet. assoc. res. vis. ophtalmol., seattle, united states of america, 2016, abstract no.3481. [40] p. zeng, r. pi, p. li, r. chen, l. lin, h. he, et al., mol. vis. 21 (2015) 688–698. [41] p. vasantha rao, p. f. deng, j. kumar, d. l. epstein, invest. ophthalmol. vis. sci. 42 (2001) 1029– 1037. [42] b. a. harrison, z. y. almstead, h. burgoon, m. gardyan, n. c. goodwin, j. healy, et al., acs med. chem. lett. 6 (2015) 84–88. [43] w. noske, graefe’s arch. clin. exp. ophthalmol. 235 (1997) 551–552. [44] r. c. tripathi, j. li, w. f. chan, b. j. tripathi, exp. eye res. 59 (1994) 723–727. admet & dmpk 4(4) (2016) 280-301 rock inhibitors in ocular disease doi: 10.5599/admet.4.4.331 299 [45] t. inoue, h. tanihara, prog. retin. eye res. 37 (2013) 1–12. [46] r. wang, j. e. williamson, c. kopczynski, j. b. serle, j. glaucoma 24 (2015) 51–54. [47] t. isobe, k. mizuno, y. kaneko, m. ohta, t. koide, s. tanabe, curr. eye res. 39 (2014) 813–22. [48] g. li, d. mukherjee, i. navarro, n. e. ashpole, j. m. sherwood, j. chang, et al., eur. j. pharmacol. (2016) 1–12. [49] f. klingberg, b. hinz, e. s. white, j. pathol. 229 (2013) 298–309. [50] m. honjo, h. tanihara, t. kameda, t. kawaji, n. yoshimura, m. araie, invest. ophthalmol. vis. sci. 48 (2007) 5549–5557. [51] a. futakuchi, t. inoue, t. fujimoto, m. inoue-mochita, m. kawai, h. tanihara, exp. eye res. 149 (2016) 107–115. [52] h. tanihara, t. inoue, t. yamamoto, y. kuwayama, h. abe, m. araie, jama ophthalmol. 131 (2013) 1288–1295. [53] h. tanihara, t. inoue, t. yamamoto, y. kuwayama, h. abe, m. araie, am. j. ophthalmol. 156 (2013) 731–736.e2. [54] h. tanihara, t. inoue, t. yamamoto, y. kuwayama, h. abe, h. suganami, et al., acta ophthalmol. 93 (2015) e254–e260. [55] h. tanihara, t. inoue, t. yamamoto, y. kuwayama, h. abe, h. suganami, et al., jama ophthalmol. 133 (2015) 755–761. [56] r. yamagishi, m. honjo, y. aoyama, t. uchida, m. aihara, alterered interactions of concomitant administrations of rock inhibitor and parasympathetic agonist in mouse iop and outflow dynamics, in annu. meet. assoc. res. vis. ophthalmol., seattle, united states of america, 2016, abstract no.3011-a0360. [57] h. tanihara, t. inoue, t. yamamoto, y. kuwayama, h. abe, a. fukushima, et al., acta ophthalmol. 94 (2016) e26–e34. [58] h. nakagawa, n. koizumi, n. okumura, h. suganami, s. kinoshita, plos one 10 (2015) 1–15. [59] n. okumura, y. okazaki, r. inoue, s. nakano, n. j. fullwood, s. kinoshita, et al., invest. opthalmology vis. sci. 56 (2015) 7560. [60] j. w. kiel, c. c. kopczynski, j. ocul. pharmacol. ther. 31 (2015) 146–151. [61] m. weiss, b. levy, c. kopczynski, t. van haarlem, g. novack, a.-c. study group, invest. ophthalmol. vis. sci. 54 (2013) 754. [62] b. levy, n. ramirez, g. d. novack, c. kopczynski, am. j. ophthalmol. 159 (2015) 980–985. [63] j. bacharach, h. b. dubiner, b. levy, c. c. kopczynski, g. d. novack, ophthalmology 122 (2015) 302– 307. [64] r. d. williams, g. d. novack, t. van haarlem, c. kopczynski, am. j. ophthalmol. 152 (2011) 834– 841.e1. [65] h. tanihara, m. inatani, m. honjo, h. tokushige, j. azuma, m. araie, arch. ophthalmol. 126 (2008) 309–315. [66] j. chen, s. a. runyan, m. r. robinson, clin. ophthalmol. 5 (2011) 667–677. [67] q.-y. lu, w. chen, l. lu, z. zheng, x. xu, int. j. clin. exp. pathol. 7 (2014) 7268–7277. [68] x. yang, a. bhaskaran, h. scott, s. ardekani, j. xu, u. mohideen, et al., rho/rock-mediated retinal endothelial stiffening impairs trpv4 signaling and promotes diabetic retinal inflammation, in annu. meet. assoc. res. vis. ophthalmol., seattle, united states of america, 2016, abstract no.3220-c0020. [69] r. arita, y. hata, s. nakao, t. kita, m. miura, s. kawahara, et al., diabetes 58 (2009) 215–226. [70] r. arita, s. nakao, t. kita, s. kawahara, r. asato, s. yoshida, et al., invest. ophthalmol. vis. sci. 54 (2013) 2373–2383. [71] a. bhaskaran, x. yang, h. scott, k. ghosh, activated monocytes contribute to the mechanical control of diabetic retinal endothelial activation, in annu. meet. assoc. res. vis. ophtalmol., seattle, united states of america, 2016, abstract no.3219-c0019. éva halász, ellen townes-anderson admet & dmpk 4(4) (2016) 280-301 300 [72] f. monickaraj, p. g. mcguire, c. franco nitta, k. ghosh, a. das, faseb j. 30 (2016) 1670–1682. [73] p.-r. rothschild, s. salah, e. gelize, c. bergin, m. berdugo, c. klein, et al., diabetes-induced rock activation contributes to outer blood retinal barrier breakdown:implications for diabetic macular edema, in annu. meet. assoc. res. vis. ophthalmol., seattle, united states of america, 2016, abstract no. 6144. [74] t. kita, y. hata, k. kano, m. miura, s. nakao, y. noda, et al., diabetes 56 (2007) 231–238. [75] t. b. connor, a. b. roberts, m. b. sporn, d. danielpour, l. l. dart, r. g. michels, et al., j. clin. invest. 83 (1989) 1661–6. [76] t. kita, y. hata, r. arita, s. kawahara, m. miura, s. nakao, et al., proc. natl. acad. sci. u. s. a. 105 (2008) 17504–9. [77] y. zheng, h. bando, y. ikuno, y. oshima, m. sawa, m. ohji, et al., invest. ophthalmol. vis. sci. 45 (2004) 668–674. [78] g.p. van nieuw amerongen, p. koolwijk, a. versteilen, v. w. m. van hinsbergh, arterioscler. thromb. vasc. biol. 23 (2003) 211–217. [79] f. celik, f. ulas, z. g. ozunal, t. firat, s. celebi, u. dogan, int. j. ophthalmol. 7 (2014) 57–61. [80] r. nourinia, h. ahmadieh, m. h. shahheidari, s. zandi, s. nakao, a. hafezi-moghadam, j. ophthalmic vis. res. 8 (2013) 337–340. [81] h. ahmadieh, jama ophthalmol. 131 (2013) 923–924. [82] s. zandi, s. nakao, k. h. chun, p. fiorina, d. sun, r. arita, et al., cell rep. 10 (2015) 1173–1186. [83] c.-w. lin, j. m. sturdivant, m. a. delong, c. kopczynski, effectiveness of ar-13154 monotherapy and combination therapy in animal models of wet age-related macular degeneration and proliferative diabetic retinopathy, in annu. meet. assoc. res. vis. ophthalmol., seattle, united states of america, 2016, abstract no.287-b0320. [84] t. sugiyama, m. shibata, s. kajiura, t. okuno, m. tonari, h. oku, et al., invest. ophthalmol. vis. sci. 52 (2011) 64–69. [85] h. watabe, s. abe, t. yoshitomi, jpn. j. ophthalmol. 55 (2011) 411–417. [86] n. sanjari, m. pakravan, r. nourinia, h. esfandiari, a. hafezi-moghadam, s. zandi, et al., j. clin. pharmacol. (2015) 1–5. [87] s. nakabayashi, m. kawai, t. yoshioka, y. s. song, t. tani, a. yoshida, et al., exp. eye res. 139 (2015) 132–135. [88] z. yang, j. wang, x. liu, y. cheng, l. deng, y. zhong, oncol. rep. 29 (2013) 1140–1146. [89] h. sagawa, h. terasaki, m. nakamura, m. ichikawa, t. yata, y. tokita, et al., exp. neurol. 205 (2007) 230–240. [90] m. ichikawa, j. yoshida, k. saito, h. sagawa, y. tokita, m. watanabe, brain res. 1201 (2008) 23–33. [91] j. yu, l. lin, x. luan, x. jing, 8 (2015) 14717–14724. [92] a. m. fontainhas, e. townes-anderson, invest. ophthalmol. vis. sci. 49 (2008) 4177–4187. [93] k. yamamoto, k. maruyama, n. himori, k. omodaka, y. yokoyama, y. shiga, et al., invest. ophthalmol. vis. sci. 55 (2014) 7126–7136. [94] a. m. fontainhas, e. townes-anderson, invest. ophthalmol. vis. sci. 52 (2011) 579–587. [95] p. a. erickson, s. k. fisher, d. h. anderson, w. h. stern, g. a. borgula, invest. ophthalmol. vis. sci. 24 (1983) 927–942. [96] s. k. fisher, g. p. lewis, vision res. 43 (2003) 887–897. [97] a. m. fontainhas, e. townes-anderson, invest. ophthalmol. vis. sci. 49 (2008) 4177–4187. [98] w. wang, e. townes-anderson, invest. ophthalmol. vis. sci. 56 (2015) 7847–7858. [99] j. wang, m. zarbin, i. sugino, i. whitehead, e. townes-anderson, invest. ophthalmol. vis. sci. 57 (2016) 3892-3906. [100] m. li, y. huang, a. a. k. ma, e. lin, m. i. diamond, neurobiol. dis. 36 (2009) 413–420. admet & dmpk 4(4) (2016) 280-301 rock inhibitors in ocular disease doi: 10.5599/admet.4.4.331 301 [101] m. li, d. yasumura, a. a. k. ma, m. t. matthes, h. yang, g. nielson, et al., plos one 8 (2013) 1–8. [102] v. gupta, n. gupta, i. h. shaik, r. mehvar, i. f. mcmurtry, m. oka, et al., j. control. release 167 (2013) 189–199. [103] k. nahar, s. absar, n. gupta, v. r. kotamraju, i. f. mcmurtry, m. oka, et al., mol. pharm. 11 (2014) 4374–4384. [104] w. wang, e. townes-anderson, neural regen. res. 11 (2016) 1029-1032. ©2016 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ doi: 10.5599/admet.2.1.35 3 admet & dmpk 2(1) (2014) 3-17; doi: 10.5599/admet.2.1.35 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index review multivariate analysis of hydrophobic descriptors stefan dove* institute of pharmacy , university of regensburg, d-93040 regensburg, germany *corresponding author: e-mail: stefan.dove@chemie.uni-regensburg.de; tel.:+49-941-9434673;fax:+49-941-9434820 received: march 03, 2014; revised: march 06, 2014; published: april 01, 2014 abstract multivariate approaches like principal component analysis (pca) are powerful tools to investigate hydrophobic descriptors and to discriminate between intrinsic hydrophobicity and polar contributions as hydrogen bonds and other electronic effects. pca of log p values measured for 37 solutes in eight solventwater systems and of hydrophobic octanol-water substituent constants  for 25 metaand parasubstituents from seven phenyl series were performed (re-analysis of previous work). in both cases, the descriptors are reproduced within experimental errors by two principal components, an intrinsic hydrophobic component and a second component accounting for differences between the systems due to electronic interactions. underlying effects were identified by multiple linear regression analysis. log p values depend on the water solubility of the solvents and hydrogen bonding capabilities of both the solute and the solvents. results indicate different impacts of hydrogen bonds in nonpolar and polar solventwater systems on log p and their dependence on isotropic and hydrated surface areas. in case of the values, the second component (loadings and scores) correlates with electronic substituent constants. more detailed analysis of the data as -values of disubstituted benzenes xphy has led to extended symmetric bilinear hammett-type models relating interaction increments to cross products x y, y x and x y which are mainly due to mutual effects on hydrogen-bonds with octanol. keywords partition coefficients; hydrophobic substituent constant; principal component analysis. introduction per iupac definition [1], hydrophobicity "is the association of nonpolar groups or molecules in an aqueous environment which arises from the tendency of water to exclude nonpolar molecules". in the strict sense, this operational definition specifies "hydrophobic effects" and "hydrophobic bonding" as intramolecular or intermolecular interactions in an aqueous phase due to attracting forces (van der waals forces based on orientation, induction and dispersion) and structural reorganization of water adjacent to nonpolar groups. ordered water molecules in hydration shells with enthalpically stronger h-bonds are transferred into bulk water with a lower degree of order, leading to an increase of both enthalpy and entropy (for review, see [2]). with a predominating entropy term (which is not necessarily the case if restriction of flexible solutes is taken into acccount) hydrophobic interactions are stronger than the attracting forces themselves. in this context, hydrophobic effects are closely related to the interacting nonpolar surface and volume fractions of groups or molecules. http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:stefan.dove@chemie.uni-regensburg.de dove multivariate analysis of hydrophobicity doi: 10.5599/admet.2.1.35 4 hydrophobicity is operationally discriminated from lipophilicity which "represents the affinity of a molecule or a moiety for a lipophilic environment", commonly "measured by its distribution behavior in a biphasic system" (iupac [1]). with this definition, experimental aspects come into play: hydrophobic or lipophilic effects are quantified by measurement of partition coefficients p in solvent-water systems or retention indices in rp-hplc or tlc (for reviews of methods, see [3] and articles therein). these quantities do not simply reflect "hydrophobicity" as defined above, but depend on the interactions of the complete solute molecules with both phases and on the phase transfer. i.e., not only the nonpolar surface fraction of the solute plays a role, but also polar effects like dipole-dipole interactions as well as formation and breaking of hydrogen bonds (see [4] and references therein). different conformations and intramolecular interactions in both phases may also be of impact. hydrophobic (or lipophilic) descriptors are therefore rather complex. nevertheless, in a first approximation they may be regarded as consisting of a nonpolar and a polar component, making factorization an attractive tool to better understand their nature. the introduction of hydrophobic effects and parameters into systematic quantitative structure-activity relationships (qsar) analysis was an essential part of the pioneering work of corwin hansch, albert leo and their coworkers. on one hand, they collected or measured a huge number of partition coefficients in different solvent-water systems [5, 6], documented as linear free-energy related quantities log p, derived the hydrophobic substituent constant  [7] in analogy to hammett's electronic  parameter, and developed a constructional fragment method of calculating partition coefficients [8]. on the other hand, the exploration of numerous qsar on different levels of integration led to a substantial advancement of the theoretical background, namely of the role of hydrophobic effects in admet (especially in transport, distribution, membrane passage) and in protein-ligand interactions ("hydrophobic bonding"). beginning with the work of collander [9], the correlation of log p values from different solvent-water systems and the decomposition of log p into more fundamental molecular descriptors like solubility, surface fractions, polarizability and hydrogen-bond strengths have contributed to quantitative structureproperty analysis (qspr) of hydrophobic effects (for review, see [4, 10]). however, already in 1964 octanolwater partition coefficients were implemented as standard in qsar analysis because of the similarity of noctanol and lipophilic biophases [7]. qspr with focus on octanol-water log p were integral parts of the foundation of log p calculation software as rekker's fragment additivity method [11, 12], clogp (for review, see [10]), acdlabs [13], several atomistic approaches and the recent surface-integral model using local properties from semiempirical mo-calculations and their integrals over the molecular surface [14]. based on large log p databases – in 1995, hansch, leo and hoekman documented ca. 17,000 values [6] – these methods have become more and more predictive, but also rather intransparent for an ordinary user with respect to the underlying qspr, i.e., to the "factors", the specific interand intramolecular interactions and forces which affect the log p under consideration. thus, in addition to available papers and manuals, detailed qspr of hydrophobic descriptors may be helpful to interprete and validate calculated quantities. at this point, multivariate analysis may come into play. the linear decomposition of correlated hydrophobic descriptors into uncorrelated "inner variables" (factors or principal components, pcs) by factor or principal component analysis (pca) yields the underlying, "inner" data structure (dimensionality, common and specific components). correlation of the pcs with physicochemical parameters identifies basic effects accounting for the multivariate qspr. with this feedback, hydrophobic descriptors may be modeled as linear functions of nonpolar and polar components by comparative multiple regression analysis. in the following sections, previous multivariate approaches jointly investigating log p values from diverse solventwater systems and substituent constants  derived from different aromatic scaffolds will be reviewed, accompanied by recalculations based on more recent data if available. dove multivariate analysis of hydrophobicity doi: 10.5599/admet.2.1.35 5 experimental the present study is based on principal component analysis and multiple linear regression analysis of hydrophobic descriptors. in brief, pca factorizes correlated data from m variables (systems) and n objects (compounds) into uncorrelated pcs according to the model: etay ijkj p 1k ikij    where aik are system-specific pc loadings and tkj compound-specific pc scores. the number of significant pcs p yields the dimensionality of the data, i.e., their recombination by the model except for an error eij. scaling of the variables (original data, normalized with zero means, or standardized with zero means and standard deviations of one) determines whether the cross product, the covariance or the correlation matrix is diagonalized. the pcs 1 to p are calculated via successive extraction of the maximal (residual) "correlation", i.e., via the eigenvalues and eigenvectors of the matrix under consideration. present recalculations were performed with in-house programs fapca and regre. multivariate analysis of partition coefficients log p from different solvent-water systems the correlation of log p-values from different solvent-water systems was first reported by collander [9]. the collander-equation log psolvent 1 = a0 + a1 log psolvent 2 (1) is restricted to homologous series or purely nonpolar solutes and models the different contribution of nonpolar solute-solvent interactions in the two solvents. the more hydrophobic solvent 1, the higher the slope a1. the intercepts a0 are positively correlated with the water solubility of solvent 1, i.e., hydrophilic solutes (log p < 0) result in higher log p if solvent 1 is more polar than solvent 2. pcas of log p-values from such restricted series extract just one significant pc. the loadings increase with the hydrophobicity of the solvent, and the scores are strongly correlated with the nonpolar surface of the solutes. with variable polar solute moieties, the situation becomes more complex. then the phase transfer comes along with different contributions of electrostatic interactions and, in particular, of broken solutewater and newly formed solute-solvent hydrogen bonds. pcas of such extended series commonly lead to two significant pcs, a "hydrophobic" and a "polar" one. all analyses known from the literature comply with this rule, namely pcas of log p-values  from 18 solutes in six solvent-water systems (n-octanol, diethylether, chloroform, benzene, toluene, cyclohexane) [15];  from 28 solutes [16], 50 solutes [17], and 69 solutes [18], respectively, in six solvent-water systems (n-octanol, diethylether, chloroform, carbon tetrachloride, benzene, n-hexane) whereas dove et al. [15] applied the standard pca method, diagonalization of the correlation matrix, the results of the group of bill dunn [16, 17, 18] were based on diagonalizing the matrix of cross products implied in the simca software pocket. in the following section, a newly calculated pca of log p-values from 37 solutes will be presented to exemplify common principles and results. the series is a subset of the solutes analyzed by koehler et al. [18], the number of solvents was extended to eight (by toluene and cyclohexane). data, either taken from ref. [18] or from the tables of hansch and leo [5, 6], are presented in table 1. dove multivariate analysis of hydrophobicity doi: 10.5599/admet.2.1.35 6 table 1. pca of log p-values from different solvent-water systems: data and pc scores. nr solute octanol diethyl ether chloro form benzene toluene ccl4 n-hexane c-hexane t1 t2 1 methanol -0.77 -1.15 -1.26 -1.89 -2.15 -2.10 -2.80 -2.80 -2.65 -0.77 2 ethanol -0.31 -0.57 -0.85 -1.62 -1.70 -1.40 -2.10 -2.10 -2.07 -0.66 3 n-propanol 0.25 -0.02 -0.40 -0.70 -0.82 -0.82 -1.52 -1.52 -1.38 -0.46 4 n-butanol 0.88 0.89 0.45 -0.12 -0.30 -0.40 -0.70 -0.70 -0.63 -0.09 5 n-pentanol 1.56 1.20 1.05 0.62 0.51 0.40 -0.40 -0.26 0.02 0.05 6 n-hexanol 2.03 1.80 1.69 1.30 1.29 0.99 0.46 0.70 0.78 0.05 7 n-heptanol 2.41 2.40 2.41 1.91 1.84 1.67 1.01 1.49 1.45 0.07 8 propionic acid 0.33 0.27 -0.96 -1.35 -1.47 -1.60 -2.14 -2.54 -1.93 0.50 9 acetone -0.24 -0.21 0.24 -0.05 -0.31 -0.30 -0.91 -0.96 -0.97 -1.47 10 trimethylamine 0.27 -0.26 0.54 -0.29 -0.36 -0.09 -0.48 -0.44 -0.75 -1.47 11 butylamine 0.74 0.11 0.99 0.14 0.30 -0.04 -0.62 -0.29 -0.42 -1.04 12 diethylamine 0.57 -0.07 0.81 -0.05 -0.09 0.03 -0.48 -0.34 -0.54 -1.26 13 pyridine 0.65 0.08 1.43 0.41 0.29 0.23 -0.21 -0.31 -0.23 -1.37 14 aniline 0.90 0.85 1.42 1.00 0.78 0.60 -0.30 0.02 0.12 -0.85 15 phenol 1.46 1.64 0.37 0.36 0.15 -0.36 -0.70 -0.81 -0.36 0.81 16 2-cl-phenol 2.15 2.05 1.36 1.46 1.37 1.19 0.85 0.86 0.92 0.19 17 3-cl-phenol 2.50 2.10 1.02 1.12 1.05 0.49 -0.07 0.08 0.49 1.10 18 4-cl-phenol 2.39 2.22 1.01 1.13 1.08 0.48 -0.11 0.08 0.49 1.11 19 2-me-phenol 1.95 1.70 1.23 1.14 1.14 0.67 -0.05 0.15 0.46 0.35 20 2,4-dime-phenol 2.30 2.40 1.50 1.34 1.26 0.78 0.34 0.34 0.76 0.83 21 2,5-dime-phenol 2.33 2.40 1.59 1.52 1.43 1.00 0.38 0.56 0.90 0.71 22 3,5-dime-phenol 2.35 2.43 1.60 1.33 1.29 0.82 0.32 0.27 0.79 0.87 23 2-naphthol 2.70 1.77 1.74 1.74 1.68 0.99 0.30 0.32 0.92 0.57 24 2-oh-benzoic acid 2.26 2.37 0.58 0.45 0.31 0.00 -0.57 -0.50 0.01 1.61 25 4-oh-benzoic acid 1.58 1.42 -0.50 -1.07 -1.17 -1.38 -1.82 -1.77 -1.31 1.72 26 2-oh-anisole 1.32 1.44 1.70 1.32 1.26 0.98 0.36 0.48 0.60 -0.60 27 2-omebenzoic acid 1.59 0.78 2.53 2.68 2.59 2.70 1.65 2.15 1.71 -2.18 28 2-no2-phenol 1.79 2.18 2.35 2.32 2.28 1.91 1.40 1.45 1.53 -0.65 29 3-no2-phenol 2.00 2.18 0.60 0.48 0.34 -0.64 -1.40 -1.22 -0.34 1.80 30 4-no2-phenol 1.91 2.01 0.20 0.17 -0.06 -0.92 -2.00 -1.70 -0.72 2.04 31 2-no2-aniline 1.85 1.95 2.13 1.78 1.64 1.08 0.25 0.36 0.89 0.02 32 3-no2-aniline 1.37 1.71 1.61 1.31 1.19 0.45 -0.62 -0.42 0.28 0.18 33 4-no2-aniline 1.39 1.48 1.23 0.93 0.78 -0.14 -1.14 -1.00 -0.14 0.53 34 vanillin 1.21 0.96 1.42 0.82 0.62 0.20 -0.72 -0.70 -0.09 -0.21 35 o-vanillin 1.37 1.35 2.30 1.87 1.73 1.40 0.53 0.65 0.94 -0.96 36 i-vanillin 0.97 0.82 1.18 0.74 0.46 0.04 -0.85 -0.82 -0.26 -0.32 37 p-toluidine 1.39 1.35 1.99 1.43 1.35 1.11 0.44 0.65 0.73 -0.76 dove multivariate analysis of hydrophobicity doi: 10.5599/admet.2.1.35 7 table 2 shows pca results with respect to the solvents as well as the means and variances of the solvent columns which had to be additionally considered since merely the correlation structure of the data was analyzed by our pca. table 2. pca results for the solvents and water solubility parameters used for correlations mean variance pc loadings extracted variance water solubility a1 a2 log  h 1 n-octanol 1.39 0.75 0.775 0.609 97.1 % 3.01 5.8 2 diethyl ether 1.24 0.95 0.724 0.675 98.0 % 1.28 2.2 3 chloroform 1.04 0.91 0.951 -0.194 94.3 % 0.75 2.8 4 benzene 0.69 1.14 0.983 -0.035 96.8 % 0.65 1.0 5 toluene 0.58 1.19 0.989 -0.039 98.0 % 0.56 1.0 6 carbon tetrachloride 0.27 1.03 0.964 -0.253 99.3 % -0.07 0.3 7 n-hexane -0.39 1.02 0.946 -0.243 95.3 % -1.87 0.0 8 cyclohexane -0.29 1.19 0.953 -0.240 96.7 % -0.90 0.1 total 83.8 % 13.1 % 96.9 % the order of the means and variances reflects the general rules derived above from the collander equation. the means are highly correlated with the water solubility of the solvents expressed in terms of log  [19]: mean = 0.43 (± 0.16) log  + 0.39 (± 0.23) r 2 = 0.88, s = 0.23,  = 0.001 (2) the variances are inversely related to the hydrogen bonding component of the solvent solubility in water, h [20] (data from ref. [21]): variance = -0.065 (± 0.042) h + 1.13 (± 0.10) r 2 = 0.70, s = 0.08,  = 0.009 (3) thus, log p scales from water soluble solvents capabable of forming hydrogen bonds show higher means and lower variances compared to log p scales from nonpolar solvents. two pcs account for 96.9 % of the data variance (first pc, 83.8 %, second pc, 13.1 %). thus, pca of log pvalues again results in a two-component model as in the case of the previous analyses [15, 16, 17, 18]. the loadings aik (table 2) represent correlation coefficients between log p from solvent i and scores tk (table 1). obviously all nonpolar solvents, in particular benzene and toluene, are sufficiently described by the first pc which, however, extracts only ca. 55 % of the variance of the hydrogen bonding solvents n-octanol and diethylether. therefore, the first pc represents "pure" hydrophobic effects due to the transfer of solutes from water into inert solvents. for polar solvents, a second pc accounting for hydrogen bonds and electrostatic interactions between solutes and solvents is necessary. this relationship may be modeled by correlation of a2 with the water solubility of the solvents: a2 = 0.21 (± 0.17) log  + 0.05 (± 0.24) r 2 = 0.61, s = 0.24,  = 0.021 (4) large positive loadings of highly water-soluble solvents are in contrast to negative loadings of carbon tetrachloride, n-hexane and cyclohexane. inspection of the scores tk and their correlation with suitable solute descriptors will enable more detailed insights into the qspr. figure 1 presents a plot of t1 vs. t2 accounting for "purely hydrophobic" effects and polar corrections, respectively, as described above. a homologous series as aliphatic alcohols is characterized by a flat line nearly parallel with the abszissa, i.e., by variation of mainly the hydrophobic component. amines under consideration are clustered, their contribution to pc2 is significantly negative. dove multivariate analysis of hydrophobicity doi: 10.5599/admet.2.1.35 8 aniline (14) resides close to the other amines, whereas phenol (15) has a positive t2 value, indicating strong interactions with hydrogen-bond acceptor solvents like n-octanol and diethylether. figure 1.plot of the scores from pca. remarkable solutes are highlighted by colored symbols and/or numbers (cp. table 1): red – alcohols, green – amines, brown – acids, redorange – nitrophenols, orange – 2-oh-anisole, yellow – chlorphenols, purple – nitroanilines. positional effects in disubstituted benzenes are evident. on one hand, m-nitroanilines and mnitrophenols are slightly more hydrophobic than their p-substituted isomers (compare 29, 30 and 32, 33, respectively). on the other hand, o-substituted benzenes are located down to the right with respect to their mand p-isomers as obvious from figure 1 for hydroxybenzoic acids (24 vs. 25), chlorphenols (16 vs. 17 and 18), nitrophenols (28 vs. 29 and 30) as well as nitroanilines (31 vs. 32 and 33). the rightward shift is due to intramolecular hydrogen bonds and/or proximity effects, both increasing the nonpolar surface and reducing hydrogen bonds as well as electrostatic interactions with water. however, these effects are much more significant in nonpolar solvents. diethylether and n-octanol are strong hydrogen bond acceptors (n-octanol additionally a weak donor), preventing the formation of internal hydrogen bonds in solutes similarly as water (for review, see [22]). the downward shift of o-substituted isomers is a consequence of this phenomenon which is most pronounced in the case of o-methoxybenzoic acid (27). also o-hydroxyanisole (26) shows a small negative "ortho-factor" in polar solvents rather due to twist than to a hydrogen bond effect [23]. suitable descriptors for the identification of scores from pca were provided by dunn et al. [16, 17, 18] who defined and calculated the isotropic surface area, isa [24], of solutes as the surface of the molecule accessible to nonspecific interactions with the solvent. the surface area of the solutes involved in specific hydrogen bonds with water, hsa, was excluded from the isa. for calculation of isa and hsa, dunn et al. constructed hydrated solutes, "supermolecules", from empirical hydration rules based on crystallographic data, quantum-chemical approaches, solution modeling and experimental data from solute-gas phase equilibria (see [25, 26] and references therein). dove multivariate analysis of hydrophobicity doi: 10.5599/admet.2.1.35 9 the scores of pc1, t1, are highly correlated with the isotropic surface area (data from [18]): t1 = 2.15 (± 0.25) isa 4.96 (± 0.58) r 2 = 0.90, s = 0.32,  < 0.001 (5) thus, the first pc accounts for the well known dependency of log p values on the nonpolar surface fraction of solutes based on entropic effects of water exclusion and nonspecific solute-solvent interactions. in the pca approach of koehler et al. [18], the matrix of cross products was diagonalized, leading to orthogonal but correlated scores which also depend on the means and variances of log p in six analyzed water-solvent systems. the correlation of t1 from ref. [18] with isa was weaker than that shown in eq. 5 (r 2 = 0.81). the scores of pc2, t2, are significantly related to the hydrated surface area (r 2 = 0.46) [24], but even if four outliers (compounds 11, 12, 27, 36) are excluded from the analysis, the correlation remains rather weak: t2 = 2.67 (± 0.86) hsa 1.56 (± 0.60) r 2 = 0.56, s = 0.63,  < 0.001 (6) i.e., the hydrated surface area plays a role in increasing log p in polar solvents (positive pc2 loadings) and decreasing log p in nonpolar solvents (negative pc2 loadings), but hsa is not sufficient to quantitatively describe this effect. compared to eq. 6, koehler et al. [18] obtained a better correlation of the scores from the second pc with the hydrated fraction of the solvent accessible surface area, f(hsa) (69 solutes, r 2 = 0.74) [24]. in our pca approach, hsa is superior to f(hsa). however, the scores t2 from ref. [18] account only for nonpolar solvent-water systems as evident from correlation coefficients between log p and t2 (in analogy to loadings a2 from our pca): octanol, 0.23, diethylether, 0.12, chloroform, 0.84, carbon tetrachloride, 0.88, benzene, 0.78, hexane, 0.89. thus, the correlation of t2 with f(hsa) in the paper of koehler et al. [18] mainly reflects a negative impact of the hydrated surface area fraction on the hydrophobicity of solutes in nonpolar solvents. the discriminative effect of f(hsa) on log p-values from different solvent-water systems may be explored in more detail by regression analysis of log p as function of isa and f(hsa) (see table 3). table 3. regression equations of log p-values (table 1) as function of surface area descriptors [18]. regression coefficients s r 2 r 2 [18] b isa f(hsa) a intercept 1 n-octanol 2.16 (± 0.61) 6.20 (± 3.00) -4.96 (± 1.96) 0.54 0.61 0.57 2 diethyl ether 2.32 (± 0.74) 6.69 (± 3.61) -5.59 (± 2.35) 0.65 0.56 0.48 3 chloroform 1.94 (± 0.24) -0.89 (± 1.17) -3.26 (± 0.76) 0.21 0.95 0.88 c 4 benzene 2.31 (± 0.37) 0.23 (± 1.82) -4.68 (± 1.19) 0.33 0.91 0.86 5 toluene 2.35 (± 0.38) 0.18 (± 1.86) -4.88 (± 1.21) 0.34 0.91 6 carbon tetrachloride 1.80 (± 0.37) -2.18 (± 1.81) -3.40 (± 1.18) 0.33 0.90 0.88 7 n-hexane 1.56 (± 0.46) -3.09 (± 2.25) -3.32 (± 1.47) 0.40 0.84 0.80 8 cyclohexane 1.81 (± 0.49) -2.64 (± 2.40) -3.89 (± 1.56) 0.43 0.84 a regression coefficients in italics are not significant ( > 0.05). b corresponding correlations of the whole dataset in ref. [18] (n = 69). c regression coefficient of -1.59 for f(hsa) significant ( = 0.01). all equations for nonpolar solvents provide a sufficient decomposition of log p into surface area terms. in contrast, equations for n-octanol and diethylether explain only ca. 60 % of the data variance. whereas regression coefficients of isa and intercepts do not significantly differ, effects of f(hsa) are distinctive with respect to the solvent class: polar solvents are characterized by a positive, nonpolar solvents by a negative or no impact of the hydrated surface area fraction of solutes on log p. solutes with a high hydration potential are poorly transferred just into carbon tetrachloride, n-hexane and cyclohexane. koehler et al. dove multivariate analysis of hydrophobicity doi: 10.5599/admet.2.1.35 10 [18] correlated log p-values calculated from loadings and scores with isa and f(hsa). because of the nonsignificance of f(hsa) in equations for n-octanol and diethylether, they sugggested that the solutes partition into the solvent as the hydrated "supermolecule" due to the water solubility of these solvents which compete with water for the hydrogen bonding sites of the solutes, leading to displacement of water from the "supermolecule". however, f(hsa) is significant if measured n-octanol and diethylether log pvalues are correlated (table 3), albeit both equations do not sufficiently model hydrophobicity. therefore, additional descriptors must be taken into account for polar solvents. the free energy of solute partition from water into n-octanol and diethylether depends on the difference of hydrogen bond interactions in both phases. suitable descriptors considering these effects on log p have been derived by taft, kamlet and abraham et al. [27, 28, 29, 30, 31]. based on the solvatochromic model [32, 33, 34], log p-values are factorized into four parameters: the molecular volume v for nonpolar interactions, the solute's dipolarity/polarizability * for orientation and induction forces, as well as  and  for the hydrogen bond donor acidity and acceptor basicity, respectively. for example, solvatochromic analysis of octanol-water partition coefficients for 103 solutes resulted in the following equation [35]: log p = 5.15 (± 0.16) v/100 1.29 (± 0.16) * 3.60 (± 0.18)  + 0.45 (± 0.12) r 2 = 0.98 s = 0.16 (7) from the series of koehler et al. [18], solvatochromic descriptors were available for 45 compounds [27, 30]. also in case of this subset, the correlations of n-octanol and diethylether log p-values with isa and hsa (r 2 : 0.77 and 0.63, respectively) or with isa and f(hsa) (r 2 : 0.74 and 0.60, respectively) are not sufficient. combining these parameters with the solvatochromic descriptors leads to the following equations for noctanol: log p = 1.88 (± 0.21) isa + 1.59 (± 0.34) hsa 2.59 (± 0.59)  2.99 (± 0.70) r 2 = 0.92 s = 0.25 (8) log p = 2.42 (± 0.31) isa + 5.82 (± 1.40) f(hsa) 2.53 (± 0.64)  4.32 (± 1.06) r 2 = 0.91 s = 0.27 (9) and for diethylether: log p = 1.80 (± 0.18) isa + 1.19 (± 0.30) hsa 4.19 (± 0.52)  1.81 (± 0.62) r 2 = 0.95 s = 0.22 (10) log p = 2.24 (± 0.25) isa + 4.55 (± 1.12) f(hsa) 4.13 (± 0.51)  3.08 (± 0.85) r 2 = 0.95 s = 0.22 (11) the volume, dipolarity/polarizability and hydrogen bond acidity descriptors v, * and , respectively, are not significant. however, both hydrated surface area descriptors show a multiple collinearity with  and : hsa = 1.05 (± 0.20)  + 1.00 (± 0.39)  0.22 (± 0.24) r 2 = 0.74 s = 0.14 (12) f(hsa) = 0.37 (± 0.08)  + 0.38 (± 0.16)  0.08 (± 0.10) r 2 = 0.66 s = 0.06 (13) taken together, these equations represent the different impacts of hydrogen bonds in nonpolar and polar solvent-water systems. the hydrated surface area reflects hydrogen bond donor acidity and acceptor basicity in equal parts and is a suitable descriptor of the detrimental effect of solute-water hydrogen bonds on log p-values in nonpolar solvents as chloroform, n-hexane and cyclohexane. in contrast, n-octanol and diethylether are strong hydrogen bond acceptors themselves. donor solutes are favored, i.e., log p increases if a large hydrated surface area is mainly due to a high -term. the net effect of  on log p is negative (compare eqs. 8-11 with 12-13). accordingly, hydrogen bond acceptor solutes are less hydrophobic in these solvents, in particular in diethylether since n-octanol is also a hydrogen bond donor, but weak compared to water. in conclusion, the decomposition of log p-values must always consider differences of solute-water and solute-solvent interactions. dove multivariate analysis of hydrophobicity doi: 10.5599/admet.2.1.35 11 multivariate analysis of hydrophobic substituent constants  from disubstituted benzenes the hydrophobic substituent constant  was first introduced by hansch et al. [36] as difference of the octanol-water log p of substituted and unsubstituted phenoxyacetic acid. following this concept, numerous -values were derived for ortho-, metaand para-substituents x in various aromatic systems phy as benzenes, nitrobenzenes, anilines, phenols, phenylacetic acids and phenoxyacetic acids [7]. whereas metapara positional effects on x and x-differences of inert substituents were only marginal, x-values of hydrogen-bonding, electron-attracting or -releasing substituents depend significantly on the nature of the "functional group" y. thus, log p values of disubstituted benzenes xphy are not simply the sum of log p (benzene), x and y from the benzene system, but include "interaction increments" due to electronic effects. in a first approximation, their nature was identified by correlations of the type [7]:  = x (phy) – x (phh) = k x (14) i.e., the difference of the x-values from a series phy and benzene depends on the electronic properties of x, described by hammett's  constant, and on the specific impact of y on x, reflected by k. to further investigate these effects, principal component analyses of xy-values from different meta and para-substituted aromatic series phy [5, 7, 37] were performed by franke et al. [38, 39]. both approaches with separate [38] and simultaneous [39] consideration, respectively, of 27 metaand parasubstituents (phy: benzenes, nitrobenzenes, anilines, phenols, benzoic acids, phenylacetic acids, phenoxyacetic acids, piperidinoacetanilides) resulted in two significant pcs. the first pc accounted for the "average" hydrophobicity of the substituents, and the second pc was due to electronic interactions between x and y and correlated with x. however, these pcas suffered from too many unknown xy values (36 of 216, 17%) which had to be estimated by regression analysis in order to obtain a full data matrix. consideration of more recent experimental log p-values [6] and withdrawal of -values of piperidinoacetanilides and the ch2oh group enables a substantial reduction of calculated data (10 of 175, 6%). with this update and some substitutions by more reliable values [6, 40], the simultaneous pca of metaand paradisubstituted benzenes was recalculated. the data matrix is shown in table 4. table 5 presents pca results for the systems phy. two pcs account for 98.7 % of the data variance (first pc, 92.6 %, second pc, 6.1 %). the loadings ayk as correlation coefficients between xy from series y and scores tk (table 4) show that -values from benzenes, phenoxyacetic, phenylacetic and benzoic acids are sufficiently reproduced by the first pc, indicating only weak effects of x on y and vice versa in these systems. thus, pc 1 represents hydrophobicity of substituents x largely unaffected by interaction with y. the hydrogen bonding acceptor system nitrobenzene and the donor-acceptor systems aniline and phenol bear considerable, opposed loadings in the second pc which is significantly correlated with p-values of the "functional groups" y: a2 = 0.49 (± 0.26) yp + 0.02 (± 0.12) r 2 = 0.83, s = 0.11,  = 0.005 (15) the correlation with ym is only weak (r 2 = 0.63). these results must be interpreted in context with the scores t1 and t2 (table 4). figure 2 presents a plot of t1 vs. t2 accounting for "unaffected" hydrophobicity of substituents x and x-y interactions, respectively. there is obviously no positional effect, corresponding m and p-substituents overlap apart from small differences in the case of oh, ome, come, cn and br. thus, the joint analysis of metaand para-disubstituted benzenes is justified. the arrangement of the substituents along the abscissa (pc 1) corresponds to the common hydrophobicity scale (polar, hydrogenbonding substituents < h < me, f < cl < br < i). dove multivariate analysis of hydrophobicity doi: 10.5599/admet.2.1.35 12 table 4. pca of xy-values a from different aromatic series phy: data, pc scores and x. nr substituent benzene phoac acid phac acid benzoic acid aniline nitro benzene phenol t1 t2 x 1 h 0.00 0.00 0.00 0.00 0.00 0.00 0.00 -0.51 0.38 0.00 2 m-f 0.14 0.22 0.19 0.28 0.40 0.05 0.47 -0.08 -0.33 0.34 3 p-f 0.14 0.15 0.14 0.19 0.25 -0.05 0.31 -0.23 -0.25 0.06 4 m-cl 0.71 0.76 0.68 0.83 0.98 0.61 1.04 0.89 0.05 0.37 5 p-cl 0.71 0.70 0.70 0.78 0.93 0.54 0.93 0.81 0.06 0.23 6 m-br 0.86 0.94 0.91 0.99 1.20 0.79 1.17 1.20 0.18 0.39 7 p-br 0.86 1.02 0.90 0.98 1.36 0.70 1.13 1.22 -0.16 0.23 8 m-j 1.12 1.15 1.22 1.28 1.46 1.09 1.47 1.68 0.44 0.35 9 p-j 1.12 1.26 1.23 1.14 1.44 1.02 1.45 1.65 0.25 0.18 10 m-me 0.56 0.51 0.54 0.52 0.50 0.57 0.56 0.43 0.86 -0.07 11 p-me 0.56 0.52 0.45 0.42 0.49 0.52 0.48 0.35 0.81 -0.17 12 m-cf3 0.88 1.07 1.16 1.07 1.39 0.77 1.49 1.43 -0.39 0.43 13 p-cf3 0.88 1.13 1.04 1.23 1.49 0.70 1.36 1.43 -0.48 0.54 14 m-ome -0.02 0.12 0.04 0.14 0.03 0.31 0.12 -0.32 0.93 0.12 15 p-ome -0.02 -0.04 0.01 0.08 0.05 0.18 -0.12 -0.47 0.90 -0.27 16 m-oh -0.67 -0.49 -0.52 -0.38 -0.73 0.15 -0.66 -1.31 1.87 0.12 17 p-oh -0.67 -0.61 -0.66 -0.30 -0.86 0.11 -0.87 -1.44 2.19 -0.37 18 m-no2 -0.28 0.11 -0.01 -0.05 0.47 -0.36 0.54 -0.43 -1.66 0.71 19 p-no2 -0.28 0.24 -0.04 0.02 0.49 -0.39 0.50 -0.40 -1.72 0.78 20 m-cooh -0.28 -0.15 -0.27 -0.19 -0.18 -0.02 0.04 -0.77 0.31 0.37 21 p-cooh -0.28 -0.22 -0.49 -0.05 -0.22 0.03 0.12 -0.78 0.42 0.45 22 m-cn -0.57 -0.30 -0.28 -0.37 0.17 -0.68 0.22 -1.00 -1.88 0.56 23 p-cn -0.57 -0.32 -0.35 -0.31 -0.06 -0.66 0.14 -1.06 -1.50 0.66 24 m-come -0.55 -0.28 -0.83 -0.31 -0.04 -0.43 -0.07 -1.15 -0.78 0.38 25 p-come -0.55 -0.37 -0.73 -0.26 -0.08 -0.36 -0.11 -1.13 -0.50 0.50 log p (phy) 2.13 1.26 1.41 1.87 0.90 1.85 1.46 a xy values in italics were not available and therefore calculated as means of at least three regression equations of the type xy1 = f (xy2, x) table 5. pca results for the aromatic systems phy and descriptors used for correlations y pc loadings extracted variance yh ym yp a1 a2 1 benzene h 0.986 0.103 98.3 % 0.00 0.00 0.00 2 phenoxyacetic acid och2cooh 0.995 -0.055 99.3 % -0.87 0.30 -0.33 3 phenylacetic acid ch2cooh 0.985 0.012 97.1 % -0.72 0.15 -0.07 4 benzoic acid cooh 0.992 0.066 98.8 % -0.28 0.37 0.45 5 aniline nh2 0.957 -0.276 99.2 % -1.23 -0.16 -0.66 6 nitrobenzene no2 0.869 0.489 99.3 % -0.28 0.71 0.78 7 phenol oh 0.947 -0.301 98.8 % -0.67 0.12 -0.37 total 92.6 % 6.1 % 98.7 % dove multivariate analysis of hydrophobicity doi: 10.5599/admet.2.1.35 13 figure 1.plot of the scores from pca. substituent types are highlighted by colored symbols: red – electron-attracting, redorange – electron-releasing, green – halogens, blue – h, me. the spread of pc 2 is due to electronic properties. electron-releasing substituents (oh, ome) are characterized by positive, electron-attracting groups (come, cn, no2, cf3) by negative scores t2. these relationships may be quantified by the following correlations with "unaffected" -values from benzenes (xh) and xp: t1 = 1.63 (± 0.12) xh 0.25 (± 0.07) r 2 = 0.97, s = 0.17,  < 0.001 (16) t2 = 2.55 (± 0.56) xp + 0.58 (± 0.23) r 2 = 0.80, s = 0.46,  < 0.001 (17) eq. 17 is better than the correlation with position-dependent xmand xp-values (r 2 = 0.68). taken together, eqs. 15 and 17 reflect an electronic x-y interaction increment described by the product xp yp. the hydrophobicity of a substituent x increases if its electronic effects on the phenyl nucleus are counterbalanced by y (electron-attracting x combined with electron-releasing y and vice versa). these findings obtained from the multivariate pca approach may be explored in more detail by individual consideration of the series. instead of correlating x with x as in eq. 14 [7], the xy-values were directly related to xh and x by multiple regression analysis (see table 6). calculated xy-values were omitted. since correlations with xp and position-dependent xmand xp-values led to approximately equivalent equations, the latter, "correct" descriptors were used. all regression equations except that for the nitrobenzene series result in an intercept of approximately zero and explain more than 95 % of the data variance. as expected from the pca (eqs. 15, 17), the regression coefficients for x depend on electronic properties of y. however, the coefficients for xh and in particular their deviation from unity seem to follow the same trend. correlation of the regression coefficients with y results in: c (xh) = -0.36 (± 0.11) ym + 1.01 (± 0.04) r 2 = 0.94, s = 0.03,  < 0.001 (18) c (x) = -0.83 (± 0.53) yp + 0.28 (± 0.24) r 2 = 0.76, s = 0.23,  = 0.01 (19) dove multivariate analysis of hydrophobicity doi: 10.5599/admet.2.1.35 14 table 6. regression equations of xy-values as function ofxh and x (table 4). regression coefficients r 2 s n b xh x intercept a 1 benzene 1.00 0.00 0.00 1.00 0.00 25 2 phenoxyacetic acid 0.92 (± 0.06) 0.34 (± 0.12) 0.05 (± 0.04) 0.98 0.07 23 3 phenylacetic acid 0.95 (± 0.06) 0.27 (± 0.12) 0.03 (± 0.05) 0.98 0.07 21 4 benzoic acid 0.91 (± 0.07) 0.18 (± 0.14) 0.12 (± 0.06) 0.97 0.10 25 5 aniline 1.04 (± 0.12) 0.84 (± 0.23) 0.07 (± 0.09) 0.95 0.15 22 6 nitrobenzene 0.72 (± 0.14) -0.44 (± 0.28) 0.22 (± 0.12) 0.86 0.18 24 7 phenol 0.99 (± 0.06) 0.91 (± 0.13) -0.07 (± 0.06) 0.98 0.09 25 a intercepts in italics are not significant ( > 0.05) b nr. of xy-values used for analysis. in both cases, the correlation with mand p-values, respectively, is significantly better than with hammett-constants of the other position (r 2 = 0.71, 0.53). eqs. 18 and 19 indicate that xy-values include two x-y interaction increments depending on the products xhymand xp yp. however, the drawback of this model is that it has been derived from separate analyses of the seven systems. a common model for all systems must be based on equivalent, "symmetric" consideration of substituents x and y in metaand para-disubstituted benzenes. fujita [40] published such a model relying on bidirectional hammett-type relationships:  = xy – xh = y x + x y (20) in this equation, y is the difference of the susceptibility contants y(octanol) – y(water) of hydrogen bonding association between the respective solvent and the fixed substituent y to the effect of variable substituents x, and x is the equivalent difference for the impact of y on x. thus, the transmission of electronic effects of substituents from x to y is assumed to be independent of transmission from y to x. to make the hammett-type relationships from eqs. 18, 19 and table 6 bidirectional implies the introduction of an additional x-y interaction increment yhx. the following common models for metaand para-disubstituted benzenes, respectively, were derived from the data in table 4 (calculated data omitted, known xy values of x = ch2oh included): xym = 0.98 (± 0.08) xh 0.25 (± 0.20) xhym 0.34 (± 0.22) yhxm 0.56 (± 0.20) xp yp + 0.13 (± 0.06) r 2 = 0.95 s = 0.14  < 0.001 n = 72 (21) xyp = 0.93 (± 0.06) xh 0.24 (± 0.12) xhyp 0.42 (± 0.16) yhxp 0.41 (± 0.21) xp yp + 0.11 (± 0.04) r 2 = 0.95 s = 0.15  < 0.001 n = 69 (22) both equations are equivalent and without collinearities of descriptors. the regression coefficients of xh are close to unity, and the intercepts may be neglected. small differences between the terms xhy and yhx are possibly due to the imbalance of the numbers of x and y substituents. for meta-disubstituted benzenes, the correlation with xp yp is better than with xm ym (r 2 = 0.93). eqs. 21 and 22 resemble the hammett formalism for bidirectional inductive interaction of any two fragments [41]: gxy = 00 + 01 x + 10 y + 11 x y (23) i.e., the contribution of the interaction to a property as, e.g., solvation energy is a function of the electronic effects of the fragments and  constants depending on the skeleton between them. dove multivariate analysis of hydrophobicity doi: 10.5599/admet.2.1.35 15 comparing these models with such bidirectional approaches [40, 41] indicates that the "susceptibility constants" x and y depend on xhand yh, respectively. polar substituents (negative -values)increase xy in combination with a second, electron-attracting substituent. by this, polarity is reduced and interactions with the strong hydrogen bond acceptor octanol become slightly more favorable. in case of hydrophobic substituents (positive -values), the interaction of a second, electron-releasing group with octanol is favored by the same reason. the xp yp cross term may be due to electron-releasing hydrogen bond donor substituents as oh and nh2. combined with electron-attracting groups, their hydrogen bonds with octanol are facilitated. this effect is even underestimated by the cross term since the greatest differences between measured and calculated xy-values of +0.3 to +0.4 occur just in case of phenols and anilines with strongly electron-withdrawing substituents as no2, cf3 and cn. combination of two hydrogen bond acceptors with positive xp-values reduces xy most. in contrast to previous suggestions [38], electronic effects on the phenyl nucleus play a minor role since otherwise meta-disubstituted benzenes should be correlated with a xm ym cross term. taken together, the hammett-type relationships represented by eqs. 21 and 22 indirectly account for mutual interactions of substituents favorable or detrimental for hydrogen bonds with octanol. models of this type may be used for the calculation of log p values of metaand para-disubstituted benzenes: log p (xphy) = log p (benzene) + yh + xy  (24) where xy is the difference between log p (xphy) and log p (phy) and comprises all x-y interactions. thus: log p (xphy) = 2.13 + xh + yh cx xhym cy yhxm cxy xp yp (25) where cx (ca. 0.25), cy (ca. 0.4) and cxy (ca. 0.5) quantify the influence of the three hammett-type increments. this is an extension of the method applied in clogp [10] where electronic interactions in metaand para-disubstituted benzenes are considered by factors fxy = y x (here, x is no hammett constant, but derived from log p values). for ortho-disubstituted benzenes, additional factors come into play representing an ortho-effect (fortho), intramolecular hydrogen bonding (fhb) and alkyl-aryl interaction (fa), so that in this case fxy = y x + fortho + fhb + fa (26) conclusions multivariate, simultaneous analysis of hydrophobic descriptors by pca may provide valuable information about the data structure (dimensionality of two, two common components). correlated parameters from different solvent-water systems and phenyl series have been transformed into uncorrelated "inner" variables discriminating between the systems and leading to suggestions about underlying interactions. via identification of such interactions per multiple linear regression analysis, the different impact of hydrogen bonds in nonpolar and polar solvent-water systems on log p values and their dependence on isotropic and hydrated surface areas has become obvious. the analysis of -values of meta and para-disubstituted benzenes has led to extended symmetric bilinear hammett-type models relating interaction increments to three cross products xy, yx and x y. the resulting models from both approaches provide detailed insight into the nature of hydrophobic descriptors and fall into line with numerous other theoretical investigations on the background of hydrophobicity and lipophilicity. dove multivariate analysis of hydrophobicity doi: 10.5599/admet.2.1.35 16 references [1] glossary of terms used in computational drug design (iupac recommendations 1997), http://goldbook.iupac.org/src_pac1997691137.html (dec 13, 2013) [2] n. muller, trends in biochemical sciences 17 (1992) 459-463. [3] v. pliška, b. testa, h. van de waterbeemd, lipophilicity in drug action and toxicology, vch verlagsgesellschaft mbh, weinheim, 1996. [4] m.s. tute, in: lipophilicity in drug action and toxicology, v. pliška, b. testa, h.v.d. waterbeemd, (eds.), vch verlagsgesellschaft mbh, weinheim, 1996, pp. 7-26. [5] c. hansch, a. leo, substituent constants for correlation analysis in chemistry and biology, wiley & sons, new york, 1979. [6] c. hansch, a. leo, d. hoekman, exploring qsar: hydrophobic, electronic, and steric constants, american chemical society, washington, dc, 1995. [7] t. fujita, j. iwasa, c. hansch, journal of the american chemical society 86 (1964) 5175-5180. [8] a. leo, p.y.c. jow, c. silipo, c. hansch, journal of medicinal chemistry 18 (1975) 865-868. [9] r. collander, acta chemica scandinavica 3 (1949) 717-747. [10] a.j. leo, chemical reviews 93 (1993) 1281-1306. [11] g.g. nys, r.f. rekker, chimie therapeutique 8 (1973) 521-535. [12] g.g. nys, r.f. rekker, european journal of medicinal chemistry 9 (1974) 361-375. [13] a.a. petrauskas, e.a. kolovanov, perspectives in drug discovery and design 19 (2000) 99-116. [14] c. kramer, b. beck, t. clark, journal of chemical information and modeling 50 (2010) 429-436. [15] s. dove, r. franke, r. kühne, in: chemical structure biological actvity relationships: quantitative approaches, j. knoll, f. darvas, (eds.), akademiai kiado, budapest 1980, pp. 247 255. [16] w.j. dunn, s. wold, acta chemica scandinavica series b-organic chemistry and biochemistry 32 (1978) 536-542. [17] w.j. dunn, m.g. koehler, s. grigoras, journal of medicinal chemistry 30 (1987) 1121-1126. [18] m.g. koehler, s. grigoras, w.j. dunn, quantitative structure-activity relationships 7 (1988) 150159. [19] j. hine, p.k. mookerjee, journal of organic chemistry 40 (1975) 292-298. [20] c.m. hansen, k. skaarup, journal of paint technology 39 (1967) 511-514. [21] a.f.m. barton, chemical reviews 75 (1975) 731-753. [22] b. testa, p.-a. carrupt, p. gaillard, r.-s. tsai, in: lipophilicity in drug action and toxicology, v. pliška, b. testa, h.v.d. waterbeemd, (eds.), vch verlagsgesellschaft mbh, weinheim, 1996, pp. 4971. [23] a. leo, in: lipophilicity in drug action and toxicology, v. pliška, b. testa, h.v.d. waterbeemd, (eds.), vch verlagsgesellschaft mbh, weinheim, 1996, pp. 157-172. [24] to get appropriate scales, isa and hsa values from table 1 of the paper of koehler et al. were divided by 100; f(hsa) was calculated as hsa/(isa+hsa). [25] s. grigoras, a new semiempirical method for the calculation of the partition coefficient, phd thesis, university of illinois at chicago, chicago, 1985. [26] m.g. koehler, a new direction in partition coefficient calculations: a molecular hydration approach, phd thesis, university of illinois at chicago, chicago, 1987. [27] m.h. abraham, chemical society reviews 22 (1993) 73-83. [28] m.j. kamlet, j.l.m. abboud, m.h. abraham, r.w. taft, journal of organic chemistry 48 (1983) 2877-2887. [29] m.j. kamlet, m.h. abraham, r.m. doherty, r.w. taft, journal of the american chemical society 106 (1984) 464-466. [30] m.j. kamlet, r.m. doherty, m.h. abraham, y. marcus, r.w. taft, journal of physical chemistry 92 (1988) 5244-5255. [31] m.j. kamlet, r.w. taft, journal of the american chemical society 98 (1976) 377-383. [32] n. el tayar, b. testa, p.a. carrupt, journal of physical chemistry 96 (1992) 1455-1459. [33] r.w. taft, m.h. abraham, r.m. doherty, m.j. kamlet, nature 313 (1985) 384-386. [34] b. testa, p. seiler, arzneimittel-forschung/drug research 31-2 (1981) 1053-1058. [35] d.e. leahy, journal of pharmaceutical sciences 75 (1986) 629-636. dove multivariate analysis of hydrophobicity doi: 10.5599/admet.2.1.35 17 [36] c. hansch, p.p. maloney, t. fujita, nature 194 (1962) 178-180. [37] r. hagemann, a. meyer, a. barth, r. franke, pharmazie 32 (1977) 526-529. [38] r. franke, s. dove, r. kühne, european journal of medicinal chemistry 14 (1979) 363-372. [39] r. franke, a. gruska, in: chemometric methods in molecular design, h.v.d. waterbeemd, (ed.), vch verlagsgesellschaft mbh, weinheim, 1995, pp. 113-163. [40] t. fujita, in: progress in physical organic chemistry, r.w. taft, (ed.), john wiley & sons, new york, 1983, pp. 75-113. [41] v.a. palm, quantitative theory of organic reactions, khimiya, leningrad, 1977. ©2014 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ manuscript doi: 10.5599/admet.2.2.34 80 admet & dmpk 2(2) (2014) 80-97; doi: 10.5599/admet.2.2.34 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper a physiologically-oriented mathematical model for the description of in vivo drug release and absorption renzo del cont 1 , michela abrami 1 , dritan hasa 2 , beatrice perissutti 2 , dario voinovich 2 , anna barba 3 , gaetano lamberti 4 , gabriele grassi 5 , italo colombo 1 , davide manca 6 , mario grassi 1 * 1 dept. of engineering and architecture, trieste university, via alfonso valerio, 6/a, trieste, i-34127 italy 2 dept. of chemical and pharmaceutical sciences, trieste university, piazzale europa 1, trieste, i-34127, italy 3 dept. of pharmacy, salerno university, via giovanni paolo ii, 132, fisciano (sa), i 84084 italy 4 dept. of industrial engineering, salerno university, via giovanni paolo ii, 132, fisciano (sa), i 84084 italy 5 dept. of life sciences, cattinara university hospital, trieste university, strada di fiume 447, trieste, i-34149 italy 5 dept. of life sciences, cattinara university hospital, trieste university, strada di fiume 447, trieste, i-34149 italy 6 dept. of chemistry, materials and chemical engineering "giulio natta" politecnico di milano, p.zza leonardo da vinci, 32 – milano, i-20133 italy. *corresponding author: e-mail: mariog@dicamp.univ.trieste.it; tel.: +39-040-558-3435; fax: +39-040-569823 received: march 02, 2014; revised: july 16, 2014; published: july 18, 2014 abstract this paper focuses on a physiologically-oriented mathematical model aimed at studying the in vivo drug release, absorption, distribution, metabolism and elimination (adme). to this purpose, the model accounts for drug delivery from an ensemble of non-eroding poly-disperse polymeric particles and the subsequent adme processes. the model outcomes are studied with reference to three widely use d drugs: theophylline, temazepam and nimesulide. one of the most important results of this study is the quantitative evaluation of the interplay between the release kinetics and the subsequent adme processes. indeed, it is usually assumed that in vivo drug release coincides with in vitro so that the effect exerted by the adme processes is neglected. in addition, the proposed model may be an important tool for the design of delivery systems since, through proper changes, it could also account for different oral delivery systems. keywords mathematical modelling, in vitro release, in vivo release, pharmacokinetics. introduction quite recently, an initiative of the pharmaceutical research and manufacturers of america (phrma) tried to evaluate the reliability of predictive models in terms of of drug efficacy, safety and properties estimation [1-3]. for this purpose, 108 clinical compounds (22 % acids, 46 % bases, 18 % zwitterions and 14 % neutrals), supplied by 12 phrma member companies, have been considered. the overall goal of this study was to assess the predictability of human pharmacokinetics (pk) from preclinical data and to provide comparisons of available prediction methods from literature. in particular, physiologically based pk (pbpk) models, despite their limited application in drug discovery and development, due to their mathematical complexity and requirement of many experimental data, have been considered. indeed, these models have http://www.pub.iapchem.org/ojs/index.php/admet/index admet & dmpk 2(2) (2014) 80-97 in vivo drug release and absorption doi: 10.5599/admet.2.2.34 81 become more attractive for the prediction of human pk profiles thanks to the prediction of hepatic metabolic clearance and tissue distribution from in vitro and in silico data as well as the advent of commercially available and user-friendly software [4]. pbpk models have showed, in 69% of the cases, a medium to high degree of accuracy in simulating the time course of the drug plasma concentration in human following intravenous administration. on the contrary, the accuracy decreases to 23% in case of oral administration [5]. accordingly, the authors of this study concluded that i) pbpk models can be a useful basis for a more rational selection of first-in-human phase dose ranges and that ii) there is a need for better predictions of human pharmacokinetics following oral administration. indeed, predicting the time course of drug concentration in plasma/blood following oral administration presents the not negligible difficulty of simulating both the delivery and the absorption stages which is obviously not needed with intravenous administration. when efforts have been focused on taking into account one of these two stages, they were directed to the absorption [6], neglecting the delivery stage. as a matter of fact, both the phrma study and common commercial software consider the release stage either by assuming a fixed release kinetics or by providing the model with an in vitro release profile assumed to be representative of in vivo release [5]. in this light and considering that oral administration is by far the most common route of drug administration [5], we believe it is worth working on the improvement of pbpk models devoted to simulate pk following oral administration. in addition, we are convinced that the new challenge in the drug delivery field, for mathematical modelling, is to combine of mechanistic theories able to realistically describe the simultaneous processes of drug release and the subsequent adme processes within the human body [7,8]. indeed, this step is a fundamental prerequisite to get the final objective of building up a mathematical model able to predict pharmacodynamics, i.e. the clinical effect of a drug released from a specific delivery system [9]. nowadays, many powerful commercial software exist for the simulation of pk phenomena. among them, simcyp, gastroplus and pk-sim can be cited. these commercial pbpk models combine a variety of functions for defining release from oral formulations with sophisticated models for the description of adme processes. in particular, simcyp is able to account for population variability of parameters and their covariation where known. our mathematical model, on the contrary, tries to establish a deep connection between the adme phenomena, described by a recently developed simple pbpk model [10], and the drug delivery step, described by a model [11,12] that proved to be reliable. in particular, the considered delivery system consists of an ensemble of non-eroding poly-dispersed polymeric particles hosting the drug inside polymeric meshes. particulate systems, indeed, present remarkable advantages over the single unit devices including easier dispersion within the stomach that can result in an appreciable reduction of local drug concentration, and reduction of the insurgence of gastric irritation [13]. moreover, particulate systems are very versatile, one more reason for their wide use. mathematical model pbpk model the mathematical representation of the human body is essentially based on the pbpk model presented by di muria and co-workers [10]. the gastrointestinal (gi) tract is described by a continuous hollow cylinder subdivided into three zones: the gastric lumen (gl) the small intestine lumen (sil) and the large intestine lumen (lil) (see figure 1). the different physiological and morphological conditions (for example ph and internal radius) in the gi tract, account for different permeability (pgl, psil and plil, respectively), drug solubility (csgl, cssil and cslil, respectively) and internal radius ( i glr , i silr and i lilr , respectively) for each zone. the gi tract is crossed by the release environment, an aqueous volume (vr), containing the delivery system del cont et al. admet & dmpk 2(2) (2014) 80-97 82 (an ensemble of drug loaded, non-eroding, poly-dispersed polymeric particles), that, due to peristaltic movements, flows downwards from the gl to the sil and eventually to the lil. for modelling purposes, it is assumed that the release environment shape is that of a hollow cylinder whose external radius coincides with i glr , i silr or i lilr , while the internal radius (r i ) can span from zero to i glr , i silr or i lilr . figure 1. schematic representation of the physiologically-oriented mathematical model. the release environment (volume vr), due to peristalsis, flows from the gastric environment (gl) to the small (sil) and eventually the large (lil) intestine. once it leaves the delivery system (a non-eroding poly-disperse ensemble of polymeric particles), the drug reaches the release environment fluids and, then, it can permeate, by passive absorption, the gl, sil and lil mucosa to reach the gastrointestinal circulatory system (gics). here it reaches the liver through the portal vein (pv). once in the liver, it can be eliminated by metabolic or biliary clearance (clh) or it may reach the plasma through the hepatic vein (hv). once in the plasma, it may be exchanged with the less perfused tissues, it may return to the liver through the hepatic artery (ha) or it can go to the cigs following the portal vein (pv). adapted from [10]. when r i is equal to zero, the release environment is simply a cylinder of radius i glr , i silr or i lilr depending on the gi position reached by the release environment. when, on the contrary, r i is > 0, the release environment assumes the shape of a hollow cylinder. in both cases, the length of the release environment is defined by (see figure 1):     2 2 i i j r jπl v r r           j = gl, sil, lil (1) admet & dmpk 2(2) (2014) 80-97 in vivo drug release and absorption doi: 10.5599/admet.2.2.34 83 it is, thus, evident that the choice of r i serves to define the (time/position dependent) contact area between the release environment and the gi tract mucosa. consequently, r i can be considered a model fitting parameter. lacking further physiological information, in this work, the most obvious r i = 0 condition was considered. due to the well-known differences in the morphology/physiology of each zone, it is supposed that the release environment speed (vre) can vary along the gi tract and this is indirectly considered for by setting three different residence (transit) times (tgl, tsil and tlil, respectively), one for each zone. release environment velocity (vre) and position in the gi tract (pos) are connected by the following relation: pos = ∫0 t vre*dt, where t is time. thus, assuming a constant vre, the transit time (tt) of the whole gi tract is given by tt = pos/vre where, now, pos indicates the gi tract length. alternatively, always assuming a constant vre, after time t, the position gained by the release environment is given by: pos = vre*t. once released, the drug can undergo elimination in the release environment (elimination constant re elk , dimensionally time -1 ) and it can cross, due to passive diffusion, the gi mucosa according to the local permeability that depends on the vr position in the gi tract. then, the drug reaches the gastrointestinal circulatory system (gics), i.e. the ensemble of blood vessels around the gi tract. the compartment named gics (gastro intestinal circulatory system) is new to pharmacokinetic modeling, firstly introduced by [10]. basically, it corresponds to the ensemble of the mesenteric artery, the portal vein and the ensemble of microcirculatory gastrointestinal vessels. this section of the whole circulatory system behaves differently from the remaining. in the case of parenteral administration, the mesenteric artery transports the drug towards the gastrointestinal tract. in the case of enteral administration, on the contrary, the portal vein transports the drug from the gastrointestinal tract towards the liver (where the drug experience the socalled first-pass effect) to finally get the remaining part of the circulatory system (here called “plasma”) and the other tissues. the introduction of the gics has allowed to effectively transform the very detailed pharmacokinetic model proposed by [14], (consisting in 21 compartments and 38 ordinary differential equations with roughly one hundred parameters), into a simple and effective model (7 compartments and 7 ordinary differential equations, with about twenty parameters). this effort saw several researchers spending lot of work, with limited and ineffective results (see, for example (see, for example, [15]). the gics is assumed to be a well-stirred environment characterized by its own constant volume (vgics) and time dependent drug concentration (cgics). the portal vein ensures the convective drug transport (blood volumetric flow qpv) from the gics to the liver (volume vl). in this well-stirred environment, drug concentration (cl) varies due to hepatic clearance (clh; dimensionally a volumetric flow) and convective transport by the hepatic vein and the hepatic artery. the hepatic vein connects the liver to the highly perfused tissues and organs (where drug concentration is assumed equal to plasma’s) by a volumetric blood flow qhv, while the hepatic artery brings back the drug from the plasma environment to the liver (volumetric blood flow qha). once in the plasma, the drug can undergo elimination ( * elk is the elimination constant, dimensionally a volumetric flow), convective transport back to the gics according to the portal vein (again, blood volumetric flow qpv) and exchange with the poorly perfused tissues and organs, indicated as “tissues” in figure 1. drug exchange between plasma and tissues is ruled by the forward mass transfer coefficient * ptk (from the plasma to the tissues; dimensionally a volumetric flow) and the reverse term * tpk (from tissues to plasma; dimensionally a volumetric flow). although * elk , * tpk and * ptk constants are not commonly used in the pk field, their use is del cont et al. admet & dmpk 2(2) (2014) 80-97 84 mandatory in this case. indeed, in order ensure that all the kinetic equations ruling drug concentration in the different environments (gics, liver, plasma and tissues) represent true mass balances referring to the drug, * elk , * tpk and * ptk must be volumetric flows [12]. however, as this is not a common choice and it can generate confusion in the reader, model equations have been re-arranged so that more common drug elimination (kel, time -1 ) and plasma – tissues exchange constants (kpt, time -1 ; ktp, time -1 ) appear in place of * elk , * tpk and * ptk . the reminiscence of * elk , * tpk and * ptk in eqs. (3) –(5) is represented by parameter f (see eq. (8)) that expresses the ratio between the plasma (vp) and the tissues (vt) volumes. accordingly, the pbpk model discussed and reported in figure 1, is mathematically represented by the following system of ordinary differential equations where t is the independent variable, time:     j jgics pv r gics p gics gics gics d d a pc q c c c c t v v     gics (2) l pv ha hv lh gics p l l l l d d c q q q c c c c t v v v           liver (3)  p hv pv hal p el pt p tp t p p d d c q q q c c k k c k c t v v             plasma (4) t pt p tp t d d c k fc k fc t   tissues (5) reel r el r lh l el p p d d m v k c c c k v c t    (6) where: j j j2πra l j = gl, sil, lil j , ,gl sil silp p p p (7) * el el pk k v * pt pt pk k v * tp tp pk k v p tf v v (8) eq. (2) allows the computing of the drug concentration in the gics, eq. (3) evaluates the liver drug concentration, eq. (4) determines the evaluation of the plasma drug concentration, eq. (5) calculates the tissues drug concentration and, finally, eq. (6) allows calculation of the amount of drug eliminated in the release environment, in the liver, and in the plasma. inspection of eq. (2) reveals that the diffusive drug flux from the release environment to the gics requires knowledge of the drug concentration in the release environment (cr) that is evaluated in the following section. delivery model briefly, our model assumes that the drug release kinetics are essentially determined by (i) polymeric particles size distribution in the dry state (ii) particle swelling upon contact with the release environment fluid (solvent) (particle erosion is considered negligible, this being true for highly crosslinked polymeric networks) (iii) drug dissolution and diffusion inside particles and (iv) possible drug re-crystallisation upon contact with the release environment fluid. the first point is achieved by assuming that dry particle size distribution can be conveniently represented by a weibull distribution function [16]: admet & dmpk 2(2) (2014) 80-97 in vivo drug release and absorption doi: 10.5599/admet.2.2.34 85 p min 2 0 1 e r r v v               (9) where rp and rmin are, respectively, the generic particle radius and the minimum particle radius,  and  are two parameters regulating the weibull size distribution, while v0 and v arethe total volume occupied by the ensemble of polymeric particles and the volume occupied by particles having a radius lower than or equal to rp, respectively. in order to account for particles polydispersity, the weibull size distribution (eq. (9)) is subdivided into nc classes (from i = 1 to nc) on the basis of the particles radius. accordingly, particles of class i th share the same radius rpi. for each class, the swelling process can be evaluated by the following mass balance referred to the swelling agent (solvent): pi 2 i i2 i i 1c r j t r r         (10) rifii jjj  , pi fi 0 i c j d r     , pi ri ri r i τ c j j d r t        (11) where jfi and jri indicate the fickian and nonfickian component of the solvent flux, respectively. jri represents a delayed flux induced by the polymer/solvent viscoelastic character. d0 and dr are, the fickian and the non-fickian solvent diffusion coefficient, respectively, while  is the polymer-solvent relaxation time:  pi peq r eq 0 g c c d d e d       pipeq eq ccf e   (12) where cpeq is the solvent concentration in the completely swollen polymeric network (thermodynamic equilibrium), eq and deq are, respectively, the relaxation time and the solvent diffusion coefficient in equilibrium conditions while g and f are two model parameters evaluated in [11] by data fitting. the existence of two contributions in the solvent flux is motivated by the fact that the “polymer + solvent” system is viscoelastic, i.e. polymeric chains undergo rearrangements that develop with time (relaxation) upon solvent uptake. in order to account for this phenomenon, camera-roda and sarti successfully proposed the approach expressed by eq. (11) [17]. eq. (10) can be solved assuming that, initially, particles are solvent-free, that solvent flux, for symmetry reasons, is zero in the particle centre and that solvent concentration at the particle surface (cpint) is defined by:       pint pint peq pint int d d c t c c c t t    (13) the swelling part of this model is completed by assuming that the solvent-polymer mixing is ideal, i.e., upon swelling, particle volume can be simply computed anytime by summing polymer and solvent volume. inspection of eqs. (10) – (13) clearly reveals that solvent diffusion is assumed independent of the presence of drug, this being rigorously true only for low drug concentrations in the particles. drug transport in the polymeric network (point iii) is modelled according to the following equations: am nc mc 2 di di dii i i2 i i i 1 c c cc c d r t t t tr r r                           (14) del cont et al. admet & dmpk 2(2) (2014) 80-97 86 i i pi1 i ds s , 1 ρ cd e d              s = 1 g/cm 3 (15)  am amam s i didi am di 0 0 0 k c c if cc t if c           (16)   ram am mc mcs s0 s s k t c c c e c     (17)  nc ncnc s i didi nc di 0 0 0 k c c if cc t if c           (18)  mc mcmc s i didi mc di 0 0 0 k c c if cc t if c           (19)  mcc r rb r sc d d m v k c c t   (20) eq. (14) represents the drug mass balance relative to the “i th ” class of particles, i.e. the class of particles characterised by radius rpi. this equation implies the existence of dissolved and undissolved drug inside the swelling polymeric network. while undissolved drug cannot move, dissolved drug mobility is controlled by its diffusion coefficient d that depends on the local solvent concentration cpi according to the peppas – reinhart theory [18] (see eq. (15)). in eq. (15), dds represents the drug diffusion coefficient in the pure solvent while i is the local polymer volume fraction and s indicates solvent density. while ci is the position-dependent concentration of the dissolved drug, amdic , nc dic and mc dic are, respectively, the positiondependent concentrations of the undissolved drug in the amorphous, nano-crystalline and microcrystalline state. indeed, in polymeric matrices, depending on the drug loading technique adopted, the drug can be found in amorphous, nanocrystalline and microcrystalline state [16]. as nano-crystals and amorphous drug are characterised by a higher solubility with respect to that of microcrystals [19], their presence can have a significant effect upon drug bioavailability. whatever the state is, the drug dissolution rate (eqs. (16), (18), (19)) is assumed to be proportional (k being the proportionality constant) to the difference between the local drug concentration ci and drug solubility ( am sc , nc sc , mc sc for amorphous, nanocrystal and microcrystal drug, respectively). in principle, the model could also account for the ph dependence of amsc , nc sc , mc sc . this, of course, would require the definition of a proper mathematical function linking solubility and ph. as the amorphous drug is often not stable and, upon dissolution, generally transforms into the more thermodynamically stable microcrystal condition (re-crystallisation; point iv)), its solubility reduces with time according to nogami theory [20] (see eq. (17)) where kr is the recrystallisation constant and ams0c is the maximum amorphous drug solubility. obviously, drug recrystallization can also take place in the release environment and eq. (20) serves to account for this phenomenon. in particular, mc and krb are, respectively, the amount of re-crystallized drug up to time t and the re-crystallization constant in the release environment. the solution of eq. (14) is performed by assuming that, initially, no dissolved drug exists inside the particles (only solid or amorphous drug can be found in the polymeric network) and that the solid and/or amorphous drug concentration are uniform inside the particles (the model could easily account for more complex drug distributions). finally, for admet & dmpk 2(2) (2014) 80-97 in vivo drug release and absorption doi: 10.5599/admet.2.2.34 87 symmetry reasons, drug flux in the particle center is set to zero and the usual drug partitioning condition at the particle/solvent interface is considered (kp is the drug partition coefficient). in order to deal with the balance between problem unknowns (cgics, cl, cp, ct, cpi, ci, cr) and equations, it is necessary to add a further equation for the definition of cr. in particular, the overall drug mass balance made on drug is considered:         pi c am nc mc 2 0 r r pi i i di i di i di i i i gics gics l l p p t t c el i 1 0 4π d rn m v c n c r c r c r c r r r v c v c v c v c m m                 (21) where m0 is the administered drug dose, nc is the number of classes into which the continuous particle size distribution has been subdivided into (eq. (9)) and npi is the number of particles belonging to the “i th ” class (npi can be simply evaluated on the basis of eq. (9)). eq. (21) simply states that, at anytime, the summed drug amount in the release environment (first addendum in eq. (21) right hand side), in the particles (sum of integrals in eq. (21) right hand side), in the gics (third addendum in eq. (21) right hand side), in the liver (fourth addendum in eq. (21) right hand side), in plasma (fifth addendum in eq. (21) right hand side), inside tissues (sixth addendum in eq. (21) right hand side) plus the drug amount recrystallized (mc) and eliminated (mel) must be equal to the administered dose m0. interestingly, the use of eq. (21) is allowed only if the other model kinetic equations (2-5) are expression of drug mass balance in the different body compartments considered (gics, liver, plasma and tissues). due to the simultaneous presence of ordinary and partial differential equations, the solution of the model calls for a numerical algorithm. in particular, the ordinary differential equations have been converted into linear algebraic equations according to the implicit euler method [21] while the implicit control volume [22] (10 particles classes, 15 radial subdivisions) has been considered for the partial differential equations conversion into linear and non linear algebraic equations. in order to ensure numerical solution accuracy, the integration time step was set to 1 s. the relaxation method [21] has been used for the iterative solution of our algebraic equations system. the solution of each iteration was carried out according to the lu decomposition approach [21]. a home-made fortran code was built up for the numerical solution. results and discussion one of the critical factors connected to the use of pbpk models is the availability of reliable values for the model parameters. of course, this is not a trivial problem [1-5] because different values for the same parameter can be found in the literature due, in some cases to considerable inter-subjects variability [23 – 29]. accordingly, in order to simplify comparison of the three different drugs considered in this paper (theophylline, temazepam and nimesulide), we decided to rely exclusively on the information reported in [10] as source of pk parameters values. the first drug considered, theophylline, also known as 1,3-dimethylxanthine, is a methylxanthine drug that finds a wide clinical use in the treatment of respiratory diseases such as asthma. its chemical/physical and pk characteristics are reported in table 1. figure 2a shows the results of model simulations when assuming an orally administered dose of 100 mg dispersed in a polymeric carrier with a volume of 1 cm 3 . in particular, the continuous thick solid line refers to theophylline concentration in the plasma (cp) in the case of only one spherical matrix of 6200 m radius. the initial part of the curve, characterised by zero drug concentration, is related to drug residence in the stomach where permeability has been set equal to zero (pgl = 0). then, cp slowly increases to a maximum followed by a subsequent drop off till the complete drug disappearance from the plasma. del cont et al. admet & dmpk 2(2) (2014) 80-97 88 table 1. model parameters referring to microcrystalline theophylline (release environment vr = 250 cm 3 ; polymeric carrier density 1.2 g/cm 3 ). pk parameters [10] vp(cm 3 ) vl(cm 3 ) vgics(cm 3 ) vt(cm 3 ) 11000 1500 9.96 24000 qpv(cm 3 /s) qha(cm 3 /s) qhv(cm 3 /s) clh(cm 3 /s) 16.25 5.41 21.6 16.46 re elk (1/s) kel(1/s) ktp(1/s) ktp(1/s) 0.0 0.0 10 -6 10 -6 tgl(s) rgl(cm) iglr (cm) pgl(cm/s) 2880 7.5 0.0 0.0 tsil(s) rsil(cm) isilr (cm) psil(cm/s) 11520.0 2.5 0.0 7.0*10 -4 til(s) rlil(cm) ililr (cm) plil(cm/s) 115200.0 2.5 0.0 3.5*10 -4 delivery parameters [11] ceq(g/cm 3 ) eq(s) deq(cm 2 /s) d0(cm 2 /s) 0.31 0.1 10 -7 10 -8 g(cm 3 /g) f(cm 3 /g) dds(cm 2 /s) kp(-) 5.0 12.0 7.2*10 -6 1.0 mc sc (g/cm 3 ) k(1/s) kr(1/s) krb(1/s) 12495 1.0 0.0 0.0 the cp increase/decrease trend corresponds to the drug absorption in the small and large intestine and is the result of both the simultaneous action of drug release from the matrix and the following adme phenomena. the main reason for the low cp values is the very slow drug delivery kinetics (low ratio between the release surface and the matrix volume, typical of big spheres) from the matrix as depicted in figure 2b (see the continuous thick solid line; cr indicates the time evolution of drug concentration in release environment). by keeping constant the delivery system volume (1 cm 3 ), its splitting into all equal smaller particles leads to a much faster cp increase. in particular, the thin solid line and the dashed line of figure 2a represent, respectively, the cp trend referred to particles characterised by a radius equal to 620 m (about 10 3 particles) and 62 m (about 10 6 particles). it can be seen that, in both cases, the drugconcentration rise is fast and cp reaches about 5 g/cm 3 after 3.5 h in the first case (620 m) and 6.5 g/cm 3 after 2.5 h in the second case (62 m). further reductions of particles size do not reflect in an appreciable variation of the cp trend. indeed, as shown in figure 2b, when particles radius is around 62 m, the drug, totally, and very rapidly, dissolves in the release environment (stomach fluids) to get the maximum value compatible with the administered dose, i.e. 400 g/cm 3 . in this case, no control on drug release is exerted and the delivery system simply ensures a prompt and total release in the stomach. in other words, no differences occur in the cp trend induced by this formulation or by a drug solution of the same dose. it is worth mentioning that the cp trend of the 62 m particles can be approximately obtained by adopting an ensemble of poly-dispersed particles described by eq.(9) considering the following parameters:  = 1000 m,  = 2, rmin = 0 m and rmax = 1200 m (see the grey line in figure 2a). the difference of using a poly-dispersed distribution relies on a smaller drug concentration (cr, see grey line in figure 2b) in the stomach fluids, which is advantageous in the case of stomach wall aggressive drugs. admet & dmpk 2(2) (2014) 80-97 in vivo drug release and absorption doi: 10.5599/admet.2.2.34 89 finally, it is worthwhile underlying that, based on the pk parameters set in table 1, the time evolution of drug concentration in the cigs (cgics) and in the liver (cl) is substantially equal to that of cp. this depends on the small cigs (vgics = 10 cm 3 ) and liver (vl = 1500 cm 3 ) volume in comparison to the plasma (vp = 11000 cm 3 ) and tissues (vt = 24000 cm 3 ) volumes and the relatively high volumetric flow connecting the various compartments (qhv, qpv and qha; see figure 1 and table 1). figure 2. (a) theophylline plasma concentration cp referring to different spherical delivery systems characterized by the same volume (1 cm 3 ) and dose (100 mg). the number close to the curves indicates the diameter of the sphere(s) considered while the gray line refers to a poly-disperse ensemble of spherical particles whose diameter spans from 0 to 1240 m. (b) theophylline concentration in the release environment (cr) referring to the four delivery systems shown in (a). in the 6200 m, 620 m and 62 m cases, the numerical solution of the model implied the subdivision of the generic particle in 500 control volumes. in the 0 to 1240 m case, the distribution was subdivided in 10 classes and 15 control volumes were considered for the generic particle of each class. on the contrary, drug concentration (ct) in the scarcely perfused tissues and organs (tissues in figure 1) monotonically increases reaching, for all formulations, values that are about two orders of magnitude smaller than cp. this is essentially due to the high value of tissues volumes (vt = 24000 cm 3 ). interestingly, for smaller particles, cp falls inside the theophylline therapeutic window that corresponds to 5 – 15 g/cm 3 . the second drug considered, temazepam (tem), a benzodiazepine derivative characterised by sedative and hypnotic actions, was chosen for the possibility of loading it into a polymeric carrier (such as crosslinked polyvinyl-pyrrolidone particles) in a completely amorphous form [11]. this, of course, can be highly advantageous as amorphous tem is characterised by a much higher solubility (≈ 8649 g/cm 3 ) with respect to the original macro-crystal solubility (≈ 164 g/cm 3 ). accordingly, we would like to theoretically evaluate the pk advantage of using amorphous tem in place of its native macro-crystal condition. in absence of precise values of tem pk parameters, we adopted those of a similar drug [10] while all other model parameters were set according to our previous study [11] (see table 2). figure 3a makes clear that amorphous tem (thick line) shows, in pk terms, a better performance with respect to crystalline tem (thin line). indeed, not only it yields to a higher maximum cp (0.16 g/cm 3 amorphous tem, 0.14 g/cm 3 , crystalline tem) but, also, maximum cp occurs after about 2 h instead of about 3.5 h (crystalline tem). this simulation was performed assuming that amorphous tem re-crystallisation did not occur inside the stomach although we have evidence of the contrary in water [11]. actually, as re-crystallisation kinetics can strongly depend also on the physical/chemical properties of the dissolving fluids (stomach fluids differ a lot from simple water), lacking precise information about tem re-crystallisation in the stomach fluids, induced us to assume that re-crystallisation did not occur. in so doing, we estimated the maximum possible effect of amorphous tem over the native crystalline tem. figure 3b shows that, in the amorphous case, drug del cont et al. admet & dmpk 2(2) (2014) 80-97 90 concentration (cr) rapidly reaches the maximum admissible concentration (400 g/cm 3 ) compatible with the administered dose (100 mg). this, obviously, takes place in the stomach (t < 0.8 h) while, once in the sil and, eventually, in the lil, cr decreases up to disappearing. in the crystalline case, on the contrary, cr increase is smaller and the maximum value is well below that reached in the amorphous case. table 2. model parameters referring to temazepam (release environment vr = 250 cm 3 , polymeric carrier density 1.2 g/cm 3 ) pk parameters [10] vp(cm 3 ) vl(cm 3 ) vgics(cm 3 ) vt(cm 3 ) 229000 1500 9.96 152000 qpv(cm 3 /s) qha(cm 3 /s) qhv(cm 3 /s) clh(cm 3 /s) 16.25 5.41 21.6 16.46 re elk (1/s) kel(1/s) ktp(1/s) ktp(1/s) 0.0 0.0 6.4*10 -5 10 -6 tgl(s) rgl(cm) iglr (cm) pgl(cm/s) 2880 7.5 0.0 0.0 tsil(s) rsil(cm) isilr (cm) psil(cm/s) 11520.0 2.5 0.0 6.0*10 -4 til(s) rlil(cm) ililr (cm) plil(cm/s) 115200.0 2.5 0.0 3.0*10 -4 delivery parameters [11] ceq(g/cm 3 ) eq(s) deq(cm 2 /s) d0(cm 2 /s) 0.31 0.35 10 -7 10 -10 g(cm 3 /g) f(cm 3 /g) dds(cm 2 /s) kp(-) 52.0 12.0 7.2*10 -6 1.0 mc sc (g/cm 3 ) k(1/s) kr(1/s) krb(1/s) 164 0.1 0.007 0.0 nc sc (g/cm 3 ) amsc (g/cm 3 ) 224 8649 at the same time, cr reduction in the sil and in the lil is slower thus ensuring higher cp from 4 hours onwards. in the case of a system characterised by 50 % amorphous tem and 50 % crystalline tem, both the cp and the cr trends are, more or less, in between those competing to the completely amorphous and the totally crystalline tem as depicted by the grey lines in figure 3a and 3b. finally, it is interesting to notice that cr fastest variations occur when the delivery system leaves the stomach (0.8 h) and the lil (3.2 h) (figure 3b) and this is determined by the gi tract permeability variation. the last drug considered is nimesulide, a typical non-steroidal anti-inflammatory drug that has been largely employed in clinical practice [30]. like tem, it is scarcely soluble in aqueous media but its solubility is ph dependent as it is a weak acid (pka = 6.46 [31]). thus, water solubility at 37 °c is around 10 g/cm 3 for ph < pka, while solubility increases to about 100 g/cm 3 with higher ph [32]. obviously, nimesulide solubility can be improved by using it in its amorphous/nano-crystalline state [19]. in addition, the firstpass liver effect for nimesulide is negligible [30]. admet & dmpk 2(2) (2014) 80-97 in vivo drug release and absorption doi: 10.5599/admet.2.2.34 91 figure 3. (a) temazepam plasma concentration cp referring to amorphous (thick black line) and crystalline (thin black line) drug. the gray line refers to a system containing 50% amorphous and 50% crystalline temazepam. delivery systems volume is equal to 1 cm 3 and dose is equal to 100 mg. (b) temazepam concentration in the release environment (cr) referring to the three delivery systems shown in (a). the numerical solution of the model implied the subdivision of the particles size distribution in 10 classes and 15 control volumes were considered for the generic particle of each class. while some pk parameters [10,16,30] and all the delivery parameters [11,19] can be deduced from literature, we could not find reliable estimation of its solubility in the gastric environment and its permeability through the gl, sil and lil mucosa. thus, our model was fitted (four fitting parameters) to experimental data referring to the oral administration (three humans) of a commercial prompt action formulation containing nimesulide (100 mg) [33]. fitting procedure was performed assuming that nimesulide is embodied in macro-crystal form inside a polymeric carrier showing the physical-chemical characteristics reported in table 3 (they are similar to those competing to polyvinyl-pyrrolidone particles considered in the tem case). in addition, i glr was set to 7.3 cm so that the stomach absorption surface equals the one normally found in humans (around 1 m 2 [16]). figure 4a reports the very good agreement between model best fitting (solid thin line) and experimental data (open circles) being fitting parameters mc sc = 39 g/cm 3 , pgl = 6*10 -4 cm/s, psil = 8.3*10 -5 cm/s and plil = 7*10 -5 cm/s. these permeability values indicate that nimesulide is rapidly absorbed by the gastric mucosa while its absorption in the sil and lil is slower due to the reduction of permeability (one order of magnitude). these results seem reasonable. indeed, the acid environment of the stomach is particularly favorable for the absorption of nimesulide that, at low ph (< 3; nimesulide is a weak acid with pka = 6.46) is completely un-dissociated (un-ionized) and this condition favors its passive absorption by the lipid mucosal membrane. on the contrary, for higher ph, such as that found in the small intestine, nimesulide is partially or totally dissociated (ionized) and this condition hinders nimesulide crossing of the lipophilic mucosal membrane. accordingly, chemical and physical properties should be responsible for the nimesulide different permeability through the gi tract rather than other reasons such as the effect of intestinal wall transporters. these chemico-phyiscal considerations have been supported by studies about the regional absorption of nimesulide [34, 35], led on healthy volunteers (gamma scintigraphy). these studies proved that stomach and proximal small bowel account for 40 % of nimesulide absorption. del cont et al. admet & dmpk 2(2) (2014) 80-97 92 table 3. model parameters referring to nimesulide (release environment vr = 250 cm 3 , polymeric carrier density 1.2 g/cm 3 ) pk parameters [10, 16 14, 30 28] vp(cm 3 ) vl(cm 3 ) vgics(cm 3 ) vt(cm 3 ) 5000 1500 9.96 13000 qpv(cm 3 /s) qha(cm 3 /s) qhv(cm 3 /s) clh(cm 3 /s) 16.25 5.41 21.6 0.49 re elk (1/s) kel(1/s) ktp(1/s) ktp(1/s) 0.0 0.0 10 -6 10 -6 tgl(s) rgl(cm) iglr (cm) pgl(cm/s) 2880 7.5 7.3 6.0*10 -4 tsil(s) rsil(cm) isilr (cm) psil(cm/s) 11520.0 2.5 0.0 8.3*10 -5 til(s) rlil(cm) ililr (cm) plil(cm/s) 115200.0 2.5 0.0 7.0*10 -5 delivery parameters [11, 19 17] ceq(g/cm 3 ) eq(s) deq(cm 2 /s) d0(cm 2 /s) 0.31 0.35 10 -7 10 -8 g(cm 3 /g) f(cm 3 /g) dds(cm 2 /s) kp(-) 5.0 12.0 10 -6 1.0 am sc (g/cm 3 ) k(1/s) kr(1/s) krb(1/s) 4108 2.0 0.06 0.0 mc sc (g/cm 3 ) ph > 6.4 mcsc (g/cm 3 ) ph < 6.4 100 39 as in the tem case, nimesulide absorption can be improved by considering its amorphous form. figure 4a shows the predicted plasma concentration cp (thick line) in the case of amorphous nimesulide assuming, as done in the tem case, that no re-crystallisation takes place (krb = 0). thus, this prediction should represent the maximum expected effect of amorphous nimesulide. it can be seen that the maximum concentration is considerably increased and, up to about 5 hours, drug concentration (thick line) is higher than that of the macro-crystal case (thin line). this behavior in the plasma is explained by figure 4b where drug concentration in the release environment (cr) is reported for the amorphous drug (thick line) and the macro-crystalline drug (thin line). basically, in the amorphous case, nimesulide is rapidly released in the first 6 minutes and its concentration becomes about one order of magnitude higher than that corresponding to the macro-crystal case. this means that its onset of action is considerably increased, a desirable feature for a rapid pain relief. also in this case, the time evolutions of drug concentration in the cigs (cgics) and in the liver (cl) are substantially equal to the cp one and the reasons are the same as the theophylline case’s. again, drug concentration (ct) in the scarcely perfused tissues and organs monotonically increases reaching, for all formulations, values that are about two order of magnitude smaller than cp. the abrupt increase of cr occurring after 1 h, in both the crystalline and amorphous case, is simply due to the reduced value of drug permeability in the lil. indeed, permeability reduction implies the decrease of the drug flux admet & dmpk 2(2) (2014) 80-97 in vivo drug release and absorption doi: 10.5599/admet.2.2.34 93 leaving the release environment and this, in turn, determines a drug concentration increase in the release environment. 0 2 4 6 8 10 12 0 4 8 12 16 20 24 c p ( g /c m 3 ) t(h) amorphous crystalline a figure 4. (a) nimesulide plasma concentration cp referring to amorphous (thick line) and crystalline (thin line) drug. open circles indicate experimental data (average of three subjects; vertical bars indicate standard error) on which the model was fit to (thin line). delivery systems volume is equal to 1 cm 3 and dose is equal to 100 mg. (b) nimesulide concentration in the release environment (cr) referring to the two delivery systems shown in (a). the numerical solution of the model implied the subdivision of the particles distribution in 10 classes and 15 control volumes were considered for the generic particle of each class. conclusions although the problem of a reliable prediction of drug concentration in plasma following oral administration remains a challenging task, a considerable development in this direction is represented by mathematical models able to simultaneously account for the in vivo drug release and the subsequent absorption, distribution, metabolism and elimination processes (adme). in this frame, we built up a reasonable mathematical model resulting from merging a physiologically oriented pk model and a delivery model. wherever the final aim is to get a reliable in vivo simulation, at present, we set up a theoretical tool that enables to rationally compare either different formulations of the same drug or the same formulation for different drugs. indeed, our model allows the evaluation of the effect of different doses, different mean particles size and particle sizes distribution, and the drug solid states, i.e. amorphous, nano-crystalline and macro-crystalline drug. according to us, our model demonstrates that in vivo release kinetics can be different from the in vitro one due to the effect of living tissues. in other words, while the mutual effect between drug release and adme processes is well known, this paper represents one of the first attempts of clearly evaluating and quantifying it. in conclusion, our model represents a rational tool for the designing of modern delivery systems and it should constitute an important step on the way to reliable simulation of plasma drug concentration following oral administration. acknowledgements: financial support from the italian ministry of education fund “prin 2010-2011 (20109plmh2)” is acknowledged. del cont et al. admet & dmpk 2(2) (2014) 80-97 94 references [1] p. poulin, h.m. jones, r.d. jones, j.w.t. yates, c.r. gibson, j.y. chien, b.j. ring, k.k. adkison, h. he, r. vuppugalla, p. marathe, v. fischer, s. dutta, v.k. sinha, t. björnsson, t. lavé, m.s. ku, journal of pharmaceutical sciences 100 (2011) 4051 – 4073. [2] r.d. jones, h.m. jones, m. rowland, c.r. gibson, j.w.t. yates, j.y. chien, b.j. ring, k.k. adkison, m.s. ku, h. he, r. vuppugalla, p. marathe, v. fischer, s. dutta, v.k. sinha, t. björnsson, t. lavé, p. poulin, journal of pharmaceutical sciences 100 (2011) 4074 – 4089. [3] b.j. ring, j.y. chien, k.k. adkison, h.m. jones, m. rowland, r.d. jones, j.w.t. yates, m.s. ku, c.r. gibson, h. he, r. vuppugalla, p. marathe, v. fischer, s. dutta, v.k. sinha, t. björnsson, t. lavé, p. poulin, journal of pharmaceutical sciences 100 (2011) 4090 – 4110. [4] r. vuppugalla, p. marathe, h. he, r.d. jones, j.w.t. yates, h.m. jones, c.r. gibson, j.y. chien, b.j. ring, k.k. adkison, m.s. ku, v. fischer, s. dutta, v.k. sinha, t. björnsson, t. lavé, p. poulin, journal of pharmaceutical sciences 100 (2011) 4111 – 4126. [5] p. poulin, h.m. jones, r.d. jones, j.w.t. yates, c.r. gibson, m. rowland, j.y. chien, b.j. ring, k.k. adkison, m.s. ku, h. he, r. vuppugalla, p. marathe, v. fischer, s. dutta, v.k. sinha, t. björnsson, t. lavé, j.w. t. yates, journal of pharmaceutical sciences 100 (2011) 4127 – 4157. [6] b. agoram, w.s. woltosz, m.b. bolger, advanced drug delivery reviews 50 suppl 1 (2001) s41 – 67. [7] j. siepmann, f. siepmann, f. international journal of pharmaceutics 314 (2008) 101–119. [8] m. grassi, g. lamberti, s. cascone, g. grassi. international journal of pharmaceutics 418 (2011) 130–141. [9] d. evans, 2010. scientia pharmaceutica 78 (2010) 588 – 588. [10] m. di muria, g. lamberti, g. titomanlio, iec research 49 (2010) 2969 – 2978. [11] m. grassi, i. colombo, r. lapasin, journal of controlled release 68 (2000) 97–113. [12] g. grassi, d. hasa, d. voinovich, b. perissutti, b. dapas, r. farra, e. franceschinis, m. grassi, molecular pharmaceutics 7, (2010) 1488 – 1497. [13] c. tapia, g. buckton, j.m. newton, international journal of pharmaceutics 92 (1993) 211–218. [14] r. jain, l. gerlowski, j. weissbrod, j. wang, r. pierson, annals of biomedical engineering 9 (1981) 347-361. [15] i. a. nestorov, l. j. aarons, p. a. arundel, m rowland, journal of pharmacokinetics and biopharmaceutics 26 (1998) 21-46. [16] m. grassi, g. grassi, r. lapasin, i. colombo, understanding drug release and absorption mechanisms: a physical and mathematical approach, crc press, boca raton, usa, 2006. [17] g. camera-roda, g.c. sarti, aiche journal 36 (1990) 851 – 860. [18] n. a. peppas, c. t reinhart, journal of membrane science 15 (1983) 275 – 287. [19] n. coceani, l. magarotto, d. ceschia, i. colombo m. grassi, chemical engineering science 71 (2012) 345 – 355. [20] h. nogami, t. nagai, t. youtsuyanagi, chemical and pharmaceutical bulletin 17 (1969) 499 – 509. [21] s.c. chapra, r. p. canale, numerical methods for engineers,mcgraw-hill, boston, usa, 1998 third edition. [22] s. v. patankar, numerical heat transfer and fluid flow, hemisphere publishing, new york, usa, 1986. [23] f. podczeck, international journal of pharmaceutics 395 (2010) 1 – 1. [24] j. m. newton, international journal of pharmaceutics 395, (2010) 2 – 8. [25] k. h. yuen, international journal of pharmaceutics 395 (2010) 9 – 16 [26] c. g. wilson, international journal of pharmaceutics 395 (2010) 17 – 25. [27] f. j. o. varum, h. a. merchant, a. w. basit, international journal of pharmaceutics 395 (2010) 26 – 36. admet & dmpk 2(2) (2014) 80-97 in vivo drug release and absorption doi: 10.5599/admet.2.2.34 95 [28] a. bowles, j. keane, t. ernest, d. clapham, c. tuleua, international journal of pharmaceutics 395 (2010) 37 – 43. [29] j. f. pinto, international journal of pharmaceutics 395 (2010) 44 – 52 [30] a. bernareggi, clinical pharmacokinetics 35 (1998) 247 – 274. [31] p. r. b. fallavena, e. e. s. schapoval, international journal of pharmaceutics 158 (1997) 109 – 112. [32] m. grassi, n. coceani, l. magarotto, international journal of pharmaceutics 239 (2002) 157 – 169. [33] r. del cont, graduating thesis, trieste university, dept. of engineering and architecture (2012). [34] a. bernareggi, k. d. rainsford, in: k. d. rainsford k.d., ed: nimesulide-actions and use. 2005, birkhäuser basel. [35] ppl 322 1999 (pharmaceutical profiles-ppl 322 1999, intelsite). list of symbols adme adsorption, distribution, metabolism, elimination aj contact area between the release environment and the j = gl, sil and lil gi tracts am dic concentration of the undissolved amorphous drug nc dic concentration of the undissolved nano-crystalline drug mc dic concentration of the undissolved micro-crystalline drug cgics drug concentration in the gastro intestinal circulatory system ci drug concentration in the i th polymeric particle cl drug concentration in the liver clh hepatic clearance cp drug concentration in the plasma cpi solvent concentration in the polymeric particle of class i th cpint solvent concentration at the polymeric particle/solvent interface cpeq equilibrium solvent concentration in the polymeric particles cr drug concentration in the release environment csgl drug solubility in the gastric lumen cssil drug solubility in the small intestine lumen cslil drug solubility in the large intestine lumen am sc time dependent amorphous drug solubility in the solvent am s0c maximum value of the amorphous drug solubility in the solvent nc sc nano-crystalline drug solubility in the solvent mc sc macro-crystalline drug solubility in the solvent ct drug concentration in tissues d drug diffusion coefficient in the swelling polymeric network d0 fickian solvent diffusion coefficient dds drug diffusion coefficient in the pure solvent deq equilibrium diffusion coefficient of solvent dr non-fickian solvent diffusion coefficient f ratio between plasma and tissues volumes del cont et al. admet & dmpk 2(2) (2014) 80-97 96 f parameter appearing in eq.(12) g parameter appearing in eq.(12) ha epatic artery hv epatic vein gi gastro-intestinal gics gastro intestinal circulatory system gl gastric lumen k drug dissolution constant * elk drug elimination constant in the plasma (volumetric flow) kel drug elimination constant in the plasma (time -1 ) * ptk mass transfer coefficient between plasma and tissues (volumetric flow) kpt mass transfer coefficient between plasma and tissues (time -1 ) * tpk mass transfer coefficient between tissues and plasma (volumetric flow) ktp mass transfer coefficient between tissues and plasma (time -1 ) re elk drug elimination constant in the gi tract kp partition coefficient (particle/solvent) kr drug re-crystallisation constant in the polymeric network krb drug re-crystallisation constant in the release environment ji solvent flux (i th polymeric particle class) jfi fickian component of ji jri non-fickian component of ji lj release environment length in j = gl, sil, lil lil large intestine lumen mc amount of the drug re-crystallised in the release environment mel amount of drug eliminated in the gi tract, in the plasma and in the liver m0 drug dose nc number of classes characterised by polymeric particles having the same radius npi number of particles belonging to i th class pbpk physiologically based pharmacokinetic pk pharmacokinetic pgl drug permeability in the gastric lumen plil drug permeability in the large intestine lumen pos release environment position in the gi tract psil drug permeability in the small intestine lumen pv portal vein qhv blood volumetric flow of the hepatic vein qha blood volumetric flow of the hepatic artery qpv blood volumetric flow of the portal vein r i j internal radius of the release environment in the j = gl, sil and lil ri radial coordinate (i th particle class) i glr internal radius of the gastric lumen admet & dmpk 2(2) (2014) 80-97 in vivo drug release and absorption doi: 10.5599/admet.2.2.34 97 i lilr internal radius of the large intestine lumen i silr internal radius of the small intestine lumen rp polymeric particle radius rpi radius of the i th polymeric particle class rmax maximum value of rp rmin minimum value of rp t time tem temazepam tgl gastric lumen transit time tlil large intestine lumen transit time tsil small intestine lumen transit time tt gi tract transit time sil small intestine lumen v volume of the polymeric particles with radius lower than or equal to rp v0 volume of all the polymeric particles vgics volume of the gastro intestinal circulatory system vl liver volume vp plasma volume vre release environment speed vr release environment volume vt tissues volume greek letters s solvent density  weibull model parameter (eq.(9))  weibull model parameter (eq.(9))  relaxation time of the polymer-solvent system (eq.(11)) eq relaxation time of the polymer-solvent system (eq.(12)) ©2014 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ manuscript doi: 10.5599/admet.2.3.47 179 admet & dmpk 2(3) (2014) 179-184; doi: 10.5599/admet.2.3.47 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper electrochemical synthesis, structure elucidation and antibacterial evaluation of 9a-aza-9a-chloro-9ahomoerythromycin a zoran mandić a, *, gorjana lazarevski b , zlatko weitner c , predrag novak d , nataša maršić e , ana budimir f a university of zagreb, faculty of chemical engineering and technology, marulićev trg 19, hr-10000, zagreb, croatia b ilica 134a, hr-10000, zagreb, croatia c hrvatskih iseljenika 1, hr-10000, zagreb, croatia d university of zagreb, faculty of science, department of chemistry, horvatovac 102a, hr-10000, zagreb, croatia e hraste and partners attorneys at law, ribnjak 40, hr-10000, zagreb, croatia f university of zagreb, school of medicine, salata 3, 10 000 zagreb *corresponding author: e-mail: zmandic@fkit.hr; tel.: +385-1-4597-164; fax: +385-1-3733-570 received: july 20, 2014; revised: august 26, 2014; published: september 16, 2014 abstract electrochemical synthesis, structure elucidation and antibacterial evaluation of 9a-aza-9a-chloro-9ahomoerythromycin a were carried out. it was found that the anodic oxidation of 9a -aza-9a-homoerythromycin a via electrogenerated reactive chlorine species leads to the chlorination of lactam nitrogen in high yield provided the ph of the reaction mixture is maintained above 3. noesy spectra reveal the existence of the mixture of two conformational families in the solution, the "folding-out" conformer being slightly more abundant comparing to 9a-aza-9ahomoerythromycin a. the chlorine substitution of lactam hydrogen resulted in improved antimicrobial potency against sreptococcus pyogenes pscb0542, moraxella catarrhalis atcc 25238, haemophilus influenzae atcc 49247 and enteroccocus faecalis atcc 29212. keywords azithromycin; electrochemical chlorination; n-chloro lactam; antbacterial activity introduction 9a-aza-9a-homoerythromycin a, 1, belongs to a class of 15-membered ring macrolides which are important intermediates for the preparation of novel antibacterial compounds with improved antibiotic properties. it consists of 15-membered aglycone ring with cladinose and desosamine sugars attached to it at positions 3 and 5, respectively. the preparation of 1 is described in the literature [1]. as a part of the electrochemical derivatisation of macrolide antibiotics [2-6], we carried out the oxidation of 1 with electrogenerated reactive chlorine species in order to introduce chlorine atom into the lactam nitrogen. n-chlorolactam derivative should exhibit conformational changes in comparison with the starting compound what might in a greater or lesser extent affect biological activity. in addition, nhttp://www.pub.iapchem.org/ojs/index.php/admet/index mailto:zmandic@fkit.hr mandić et al. admet & dmpk 2(3) (2014) 179-184 180 chlorolactam derivative is a source of amydil radical [7] making it an important precursor for novel compounds with biological activity. although n-chloroamides and n-chlorolactams can be prepared by the number of chemical methods, the literature on the electrochemical halogenation of amides or lactams is sparse. krishnamoorthy et al. [8] conducted a thorough evaluation of optimal conditions for the electrochemical bromination and chlorination of succinimide. they found that the reaction is highly sensitive on experimental parameters. lyalin and petrosyan [9] succeded in preparative electrosynthesis of sodium salts of n-chloroamides of arylsulfonic acids. in this paper we report on the electrochemical halogenation of 1, thorough structural characterization and biological evaluation of the resulting product. experimental electrochemical synthesis of 2 electrochemical oxidation was carried out under constant current condition in the conventional twocompartment electrochemical cell. compound 1 (20-100 mg) was dissolved in 20 ml of 0.1 mol dm -3 licl solution in methanol and solution was electrolyzed with the constant current of 2.5 ma cm -2 . working electrode (anode) was platinum sheet and as cathode served graphite rod. the electrolysis was carried out at room temperature and the anolyte was stirred with a magnetic stirrer. the progress of the electrolysis was monitored by thin layer chromatography. after the completion of the electrolysis, the anolyte is separated from the reaction cell, evaporated to a dry residue and subjected to chromatography on a slilica gel column (ch2cl2-ch3oh-nh4oh = 90 : 9 : 1.5). the title compound is obtained in 60% yield with the following spectral data: ms (es+): m/z 797.4 (calc. 797.43); max(kbr)/cm -1 = 3444, 2973, 2937, 1731, 1650, 1463, 1381, 1169, 1055, 1011, 736; 1 h nmr (cdcl3):  = 2.67 (1h, h-2), 3.54 (1h, h-3), 1.92 (1h, h-4), 3.72 (1h, h-5), 3.33 (1h, t, h-8), 4.81 (1h, m, h-10), 3.58 (1h, d, h-11), 1.38 (3h, d, h-10-ch3), 4.85 (1h h-13); 13 c nmr (cdcl3):  = 44.1 (c-2), 71.8 (c-3), 41.7 (c-4), 78.8 (c-5), 30.8 (c-8), 179.1 (c-9), 52.0 (c-10), 11.2 (c-10-ch3),71.8 (c11). nmr all 1d and 2d ( 1 h, apt, gcosy, ghsqc and ghmbc) nmr spectra were recorded at ambient temperature on the avance drx500 spectrometer working at 500.13 mhz for 1 h and equipped with a 5 mm diameter inverse detection probe with z-gradient. the sample concentration in cdcl3 was 20 mg ml -1 with tms as the internal standard. 1d 1 h and apt nmr spectra were obtained with 4000 hz and 32000 hz spectral window, respectively, using 64k data points. digital resolution was 0.12 hz and 0.95 hz per point, respectively. 2d gcosy spectra were acquired with a sweep width of 6666 hz in both dimensions into 2k data points with 512 increments. spectra were zero-filled in the f1 dimension to 1k and processed using an unshifted sine bell window function. digital resolution was 3.26 hz per point and 16.58 hz per point in f2 and f1 dimensions, respectively. the inverse 1 h13 c correlation experiment ghmqc was recorded at 125.77 mhz using data matrices of 2k x 256 with 4 scans and processed with a shifted sine bell window function and linear prediction. hmbc spectra were recorded using a transfer delay for the evolution of long range c-h couplings of 60 ms with 256 increments into a matrix of 4kx2k data points, with a sweep width of 7000hz in f2 dimension and admet & dmpk 2(3) (2014) 179-184 n-chlorination of macrolide derivative doi: 10.5599/admet.2.3.47 181 31500 hz in f1 dimension. digital resolution was 1.7 hz per point and 30.70 hz per point in f2 and f1, respectively. antibiotic susceptibility test antibiotic susceptibility data given in table 1 were obtained by microdilution test in mueller-hinton media as described by nccls [10] except that test substances and standards were dissolved in dmf (merck). also for streptococcus medium, blood was substituted with 5 % horse serum. results and discussion formation of 2 by indirect electrochemical oxidation oxidation of chloride ions in methanolic solutions gives rise to the formation of reactive chlorine species as a powerful chlorinating agent. in the first step of the reaction dihalogen is formed which rapidly disproportionates into methyl hypochlorite, ch3ocl, and hydrochloric acid (equations 1 and 2). the formation of n-cl bond is an electrophilic substitution reaction and is believed to proceed via direct transfer of cl from methyl hypochlorite to lactam nitrogen (equation 3) as in the case of chlorination of α-amino acids by naocl in aqueous media [11]. 2nacl 2e → 2na + + cl2 (1) cl2 + ch3oh  ch3ocl + h + + cl (2) the crucial feature of the electrochemical chlorination of 1 is that the formation of 2 is sensitive to ph of the solution and it was necessary to maintain the ph during electrolysis above 3. the anolyte acidifies during the electrolysis which is due to in-situ formation of hydrochloric acid and, in addition, to the less than 100% current efficiency of dihalogen formation. namely, a small fraction of current is spent on oxygen evolution resulting in the liberation of protons. when the ph drops below 3, the equilibrium of reaction 2 is displaced to the left and the reaction stops. therefore, in order to carry out the chlorination reaction in high yield, the ph of the reaction mixture was continuously adjusted by adding a few drops of the naoh solution. o o o n o oh h oh o o oh oh n o o o o o o n o oh cl oh o o oh oh o n o o 1 2 -2e licl / ch3oh1 1 9 9 (3) mandić et al. admet & dmpk 2(3) (2014) 179-184 182 structural elucidation the chemical structure of the compound 2 was determined by means of ir and nmr spectroscopies and ms spectrometry. ir spectrum has shown no band characteristic for n-h  stretching vibration at 1530 cm -1 (amide ii) observed for the starting compound 1. a broad band in the region 800-600 cm -1 assigned to n-h wagging disappeared as well. these findings indicate that there is no hydrogen atom attached to the lactam nitrogen and that the reaction took place at 9a position. in the ms spectrum a precursor ion peak was observed at m/z 797.4 which is exactly by 34 da higher than the mass of the starting compound. isotopic pattern of precursor ion and some fragment ions revealed that chlorine atom is present in the molecule. the carbon and proton chemical shifts of 2 were assigned by the combined use of the standard oneand two-dimensional nmr experiments. a disappearance of the lactam proton at 6.24 ppm as observed for 1 provides further evidence for chlorine substitution at nitrogen. accordingly, changes in proton and carbon chemical shifts of the neighbouring atoms took place (table 1). a characteristic down-field effect of atoms h8, h10, h11 and h10me was observed due to the presence of the chlorine atom. changes in neighbouring carbon atoms chemical shifts (both upand down-field shifts) are the consequence of the chlorine substitution. in the apt spectrum the carbon multiplicity pattern is the same as that observed for the parent compound which excludes the substitution at the carbon atom. conformational studies on macrolides [12-14] have shown the existence of two major conformational families: folded-out and folded-in, referring to the outward and inward folding of the macrocycle ring fragment c3-c5. those studies demonstrated that vicinal coupling constants 3 jh2h3 and noe proton-proton contacts h4-h11 and h3-h11, respectively, can serve as good indicators of the ring folding. in that respect, lower values of 3 jh2h3 (2-3 hz) and h3-h11 noe cross peaks indicate the folded-in conformations whereas much higher coupling constant values (up to 10 hz) and h4-h11 noe contacts are characteristic for the folded-out ones. the ratio of the two conformational families was found to be dependent on the solvent and temperature. recent results have shown that the ribosome-bound conformations of some 6-o-methyl homoerythromycins were found to be very similar to the free ones with some small differences observed [15]. x-ray structural analysis showed that 1 adopted a folded-in conformation in the solid state as was also the case in solution [15]. there was no indication of intramolecular hydrogen bonding involving the lactam nitrogen atom. according to the 3 jh2h3 value measured for the solution of 2 (4.5 hz) the abundance of the folded-out conformers has slightly increased in comparison with 1 (5.5 hz) as a consequence of a chlorine substitution. both h3-h11 and h4-h11 noe cross-peaks were observed in the noesy spectrum of 2 which pointed towards a mixture of the two conformational families. table 1. chemical shift change from the starting compound, 1, to product 2 atom chemical shift change, ppm h8 2.23  3.33 h10 4.16  4.81 h11 3.22  3.58 h10me 1.16  1.38 c8 35.4  30.8 c9 177.1  179.1 c10 45.0  52.0 c10me 13.5  11.2 admet & dmpk 2(3) (2014) 179-184 n-chlorination of macrolide derivative doi: 10.5599/admet.2.3.47 183 antibiotic susceptibility evaluation the comparative mics of 1 and 2 against different gram-positive and gram-negative pathogens are presented in table 2. it appears that 2 possesses slightly improved potency against some respiratory pathogens. 2 was eight times more potent than 1 against s. pyogenes pscb0542, and twice as active against m. catarrhalis atcc 25238, h. influenzae atcc 49247 and e. faecalis atcc 29212. however, against erythromycin-susceptible streptococcus pneumoniae compound 2 slightly lost its potency comparing to 1. against the rest of tested strains (staphylococcus aureus pscb0329, s. aureus pscb0538, s. aureus pscb0330, s. aureus pscb0331, s. pneumoniae pscb0328, s. pneumoniae pscb0326, s. pyogenes pscb0543, s. pyogenes pscb0544, s. pyogenes pscb0545, e. coli atcc 25922) 2 had equal potency as 1. according to the nccls breakpoints for azithromycin in s. pyogenes, the change in mic from 1 to 0.125 is significant, and therefore the new compound, unlike 1, could be adequate for treatment of infections caused by these strains. in h. influenzae, breakpoint according to nccls is 4 μg/ml, and for clarithromycine is 8 μg/ml, and mic 16 is interpreted as intermediary resistant. resistance to macrolides in gram-positive organisms whether inducible or constitutive, is due to methylation of an adenine residue in rrna in 50s ribosomal subunit. 23s rna is binding site for macrolides. the equal potency of 1 and 2 against macrolideresistant gram-positive strains (strains with mls and m phenotype), whether it was constitutive or inducible was expected, since the slight conformational difference between 1 and 2, probably had no effect on the mechanism of action. table 2. mic values (µg/ml) for 1 and 2 strains phenotype 1 2 s. aureus pscb0329 erys 4 4 s. aureus pscb0538 imls >64 >64 s. aureus pscb0330 cmls >64 >64 s. aureus pscb0331 m >64 >64 s. pneumoniae pscb0541 erys <=0.125 0.25 s. pneumoniae pscb0328 cmls >64 >64 s. pneumoniae pscb0326 m 64 64 s. pyogenes pscb0542 erys 1 0.125 s. pyogenes pscb0543 imls 64 64 s. pyogenes pscb0544 cmls >64 >64 s. pyogenes pscb0545 m 64 64 m. catarrhalis atcc 25238 0,25 <=0.125 h. influenzae atcc 49247 16 4 e. faecalis atcc 29212 16 4 e. coli atcc 25922 >64 >64 conclusions electrochemical indirect oxidation of 9a-aza-9a-homoerythromycin a leads to the chlorination of lactam nitrogen, a novel compound in the class of macrolide antibiotic. the conformation of the chlorinated product in solution indicates more abundant presence of the “folded-out” conformer comparing to starting compound. in addition, it possesses slightly improved potency against some respiratory pathogens. however, the compounds described in this paper are not considered to be suitable candidates for drug development or to become lead compounds. mandić et al. admet & dmpk 2(3) (2014) 179-184 184 references [1] g. lazarevski, g. kobrehel, ž. kelnerić, us patent, 6,110,965, (2000) [2] z. mandić and n. lopotar, electrochem. comm., 7 (2005) 45-48 [3] d. iveković, n. lopotar, k. brajša, z. mandić, eur. j. pharm. sci., 18 (2003) 323-328 [4] z. mandić, z. weitner, m. ilijaš, j. pharm. biomed. anal., 33 (2003) 647-654 [5] z. mandić, a. naranđa, p. novak, k. brajša, m. đerek, d. iveković, j. antibiot., 55 (2002) 807-813 [6] z. mandić, a. naranđa, n. lopotar, lj. duić, d. iveković, m. tkalčec, j. antibiot., 52 (1999) 1143-1145 [7] b. daoust and j. lessard, tetrahedron, 55 (1999) 3495-3514 [8] s. krisnamoorthy, r. kanakam srinivasan and m. noel, bull. electrochem., 16 (2000) 544-550 [9] b.v. lyalin and v.a. petrosyan, russ. j. electrochem. 31 (1995) 1146-1150 [10] perfomance standards for antimicrobial susceptibility testing; nineteenth informational supplement. m100-s19. vol 29. no 3. clinical and laboratotory standards institute, 2009. [11] x.l. armesto, m. canle l., j.a. santaballa, "α-amino acids chlorination in aqueous media", tetrahedron, 49 (1993) 275-284. [12] j. r. everett, j. w. tyler, j. chem. soc., perkin trans. 2, (1987) 1659-1667 [13] p. novak, z. banić tomišić, p. tepeš, g. lazarevski, j. plavec, g. turkalj, org. biomol. chem. 3 (2005) 39–47. [14] p. novak, i. tatić, p. tepeš, s. koštrun, j. barber, j. phys. chem. a 110 (2006) 572-579. [15] n. košutić-hulita, d. matak-vinković, m. vinković, p. novak, g. kobrehel, g. lazarevski, croat. chem. acta, 74 (2001) 327-341. ©2014 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ manuscript doi: 10.5599/admet.2.2.40 71 admet & dmpk 2(2) (2014) 71-79; doi: 10.5599/admet.2.2.40 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index short communication increasing the biorelevance of simulated intestinal fluids for better predictions of drug equilibrium solubility in the fasted upper small intestine nikolaos koumandrakis, maria vertzoni, christos reppas* department of pharmaceutical technology,faculty of pharmacy, national and kapodistrian university of athens, panepistimiopolis, 15771 zografou, greece *corresponding author: e-mail: reppas@pharm.uoa.gr; tel.: +30 210 727 4678; fax: +30 210 727 4027 received: may 25, 2014; published: july 18, 2014 abstract to date the importance of luminal species other than bile salts and phosphatidylcholine on drug equilibrium solubility in the fasted upper small intestine has been evaluated to a very limited extent. in this communication the importance of fatty acids, cholesterol, and proteins on solubility of four model lipophilic compounds was evaluated by including these components into previously proposed simulated intestinal fluids. data were compared with ex vivo solubility data in aspirates reflecting the mean and the median luminal composition in the upper small intestine. it is concluded that estimation of solubility in aspirates reflecting the median luminal composition is better estimated when the presence of cholesterol and fatty acids is also simulated. in contrast, estimation of solubility in aspirates reflecting the mean luminal composition requires consideration of additional factors (e.g. buffer species identity, non-micellar colloidal structures, and lyso-phosphatidylcholine content). keywords: human intestinal fluids; biorelevant media; danazol; indomethacin; cinnarizine; ketoconazole introduction drug equilibrium solubility in the contents of upper intestinal lumen is an important factor for the oral absorption process as it determines the maximum dose which can be absorbed and it affects luminal drug dissolution rates. in case the active pharmaceutical ingredient (api) is lipophilic, luminal colloidal species (e.g. mixed bile salt micelles) greatly enhance luminal solubility and relevant information is crucial for predictions of drug plasma levels after oral administration of conventional dosage forms, e.g. when physiologically based pharmacokinetic modelling approaches are applied [1]. luminal equilibrium solubility can be estimated by data in simulated intestinal fluids [e.g. 2,3]. composition of relevant fluids is a balance between the actual luminal environment, the importance of specific luminal components, and, sometimes, the costs associated with their reproducible preparation methods and storage conditions. to date, the importance of certain luminal components which may have an impact on luminal solubility of lipophilic apis in the fasted state, such as fatty acids, cholesterol and http://www.pub.iapchem.org/ojs/index.php/admet/index koumandrakis et.al. admet & dmpk 2(2) (2014) 71-79 72 proteins, has not been considered. such components have been recently shown to have a role in luminal precipitation of lipophilic weak bases [4]. the objective of this communication was to evaluate the usefulness of various fasted state simulating intestinal fluids in predicting solubility in the upper small intestine of four model lipophilic compounds (table 1) with emphasis to components which are present in the upper small intestine but their importance has not been evaluated to date. simulated intestinal fluids were prepared by using crude bile salt material which had been previously analyzed in our laboratory for its exact bile salt content. data in simulated intestinal fluids were compared with data in human aspirates. since it is not clear whether the average luminal composition should be considered on the basis of mean or of median luminal composition [5], in the present study two pooled aspirates samples that reflected the mean and the median luminal composition were used. table 1. physicochemical characteristics of the model compounds used in this study cinnarizine [6-8] ketoconazole [9,10] indomethacin [9] danazol [11] molecular weight 368.5 531.4 357.8 337.5 clogp 5.19 3.65 3.66 4.5 pka 7.78 2.94 and 6.51 (both alkaline) 4.5 aqueous solubility (µg/ml) 0.29 0.0866 0.937 1.1 experimental cinnarizine was from janssen cilag aebe (lot # 1001105726, athens, greece), ketoconazole was from recordati espana s.l. (lot # 03000051, beniel, spain), indomethacin was from f.i.s. fabrica italiana sintetici s.p.a. (lot # 0907105123, vicenza, italy) and danazol was from sanofi-aventis (batch # f510357, fawdon, france). egg lecithin (lipoid e pc, > 98 % phosphatidylcholine) was from lipoid gmbh (ludwigshafen, germany). ox bile salt extract was from fluka chemie gmbh, germany (product no 86340, lot # 386645/1 12601). according to our hplc-cad analysis [12] the total bile salts content of the crude material used in the presenet study is 71.5 % w/w, and the following bile acids are present: taurocholic 22.2 % w/w, glycocholic 24.7 % w/w, taurochenodeoxycholic 1.7 % w/w, ursodeoxycholic 7.1 % w/w, glycochenodeoxycholic 1.3 % w/w, cholic 7.5 % w/w, glycodeoxycholic 7.1 % w/w. [3]. albumin from bovine serum, biochemika fraction v was from fluka (cat # 05488, lot # s31293-466). methanol and acetonitrile of hplc grade were from e. merck (darmstadt, germany). water purified with labconco water pro ps system (kansas city, missouri, usa) was used in all procedures. all other chemicals were of analytical grade. individual human aspirates were collected from the ligament of treitz after receiving approval by the ethical committee of the red cross hospital of athens (a.p. 497 / 06.12.2011). samples were selected and pooled so that the pooled sample had composition similar to the mean composition of intraluminal contents [5], hifmean. in addition, samples were selected and pooled so that the pooled sample had composition similar to the median composition of intraluminal contents [5], hifmedian (table 2). due to multiple freeze – thawing processes, ph of both hifmean and hifmedian used in the solubility measurements was increased from 6.5 (aspiration time) to 7.5. fassif [13] and fassif-v2 [2] reflect the mean environment in the fasted upper small intestine. recently, fassif-v2plus was proposed for predicting the precipitation of weak bases in the upper small admet & dmpk 2(2) (2014) 71-79 increasing the biorelevance of simulated intestinal fluids doi: 10.5599/admet.2.2.40 73 intestine [4]. in the present investigation, all these media were prepared by using crude bile material (ox bile salt extract) and used for measuring the solubility of the four model apis (table 1). the importance of phosphatidylocholine concentration (0.75, 1, 1.5 mm) and of bovine serum albumin concentration (bsa 0.5-4 mg/ml) on the predictability of luminal solubility of ketoconazole, indomethacin and danazol were also evaluated. compositions of simulated intestinal fluids used in the presented study is shown in table 2. finally, solubilities of ketoconazole, indomethacin and danazol were also measured in buffer solutions of phosphates and maleates having ph values of 6.5. solubilities were measured in triplicate using the shake-flask method. for solubility measurements in human intestinal fluids, 1 ml of hιf and pure drug powder in excess (1 mg for cinnarizine, 15 mg for ketoconazole, 5 mg for indomethacin and 1 mg for danazol) were transferred into 2-ml polypropylene vials. vials were then sealed with polypropylene caps, and put in an oscillating water bath (37 °c). equilibration times were experimentally determined or had already been determined previously using similar media volumes (1-5 ml) and the same in vitro setup [3,10,14,15]. equilibration times were 2 h (cinnarizine), 8 h (ketoconazole), 2 h (indomethacin) and 4 h (danazol). after reaching equilibrium, and since hιfs could not be filtered through 0.45 μm filters, samples were centrifuged for 10 min at 11500 g at 37 °c. after measuring the ph, the supernatant was appropriately diluted with acetonitrile, centrifuged again (10 min, 11500 g, 10 °c) and assayed for drug concentration using validated hplc-uv methods. for solubility measurements in simulated intestinal fluids, 5 ml of the medium and pure drug powder in excess (5 mg for cinnarizine, 175 mg for ketoconazole, 25 mg for indomethacin and 5 mg for danazol) were transferred into 10-ml polypropylene vials. vials were then sealed with high density polypropylene caps and put in an oscillating water bath (37 °c) for periods identical to those used in the experiments in human intestinal fluids (please see above). after reaching equilibrium, samples were filtered through 0.45 μm regenerated cellulose filters (adsorption to filters was measured to be negligible), the filtrates were diluted appropriately diluted with acetonitrile and assayed for their drug content using validated hplc-uv methods. apis were evaluated for their stability in solubility media prior to solubility measurements. ketoconazole, indomethacin and danazol were found to be stable in all media tested in this study. for cinnarizine, a portion of the pooled sample reflecting hifmean was transferred in a test tube, cinnarizine in excess was added, the tube was sealed and it was incubated at 37 °c in a shaking water bath. four hours after the initiation of shaking the contents of the tube were centrifuged (11500 g, 37 o c, 10 min). ph of the supernatant was measured to be 7.8. a sample from the supernatant was analyzed immediately for its cinnarizine content (content at time t=0) and the remaining was transferred in another tube, sealed and incubated for 20 hours at 37 °c in the same shaking water bath. samples were obtained 0.5, 1, 1.5, 2, 3.2 and 20 hours after the initiation of the incubation of the filtrate. all model compounds were assayed with hplc-uv methods (table 3). for each drug/medium combination, quantification was based on linear standard curves constructed using the same medium and identical sample handling procedure. koumandrakis et.al. admet & dmpk 2(2) (2014) 71-79 74 t a b le 2 . m e a n a n d m e d ia n c o m p o si ti o n o f fl u id s co ll e ct e d f ro m t h e f a st e d u p p e r sm a ll i n te st in e o f h e a lt h y a d u lt s b a se d o n p re vi o u s d a ta [ 1 1 ], ( h if m e a n a n d h if m e d ia n , re sp e ci ti ve ly ), c o m p o si ti o n o f th e p o o le d s a m p le s u se d i n t h e p re se n t st u d y to r e se m b le t h e c o m p o si ti o n o f h if m e a n a n d h if m e d ia n , a n d c o m p o si ti o n o f si m u la te d in te st in a l fl u id s u se d i n t h e p re se n t st u d y f a s s if cr v 2 p lu sm e d ia n 1 .8 * 0 .0 5 * * 0 .2 0 .1 6 .5 m a le a te s 6 8 .6 2 * c ru d e o x b il e w a s u se d * * b o v in e a lb u m in w a s u se d a t co n ce n tr a ti o n s ra n g e 1 -4 m g /m l f a s s if cr v 2 p lu s 3 * 0 .2 * * 0 .5 0 .2 6 .5 m a le a te s 6 8 .6 2 f a s s if cr v 2 3 * 0 .2 * * 6 .5 m a le a te s 6 8 .6 2 f a s s if cr 3 * 0 .7 5 * * 6 .5 p h o sp h a te s 1 0 5 .8 5 p o o le d s a m p le re fl e ct in g h if m e d ia n a n d u se d i n t h e p re se n t st u d y 1 .6 1 0 .0 9 0 .9 9 0 .3 2 0 .0 9 7 .6 b ic a rb o n a te s n o t m e a su re d h if m e d ia n n = 3 4 5 [5 ] 1 .9 0 .0 5 2 1 .1 6 0 .1 8 0 .0 8 4 6 .2 b ic a rb o n a te s n o t m e a su re d p o o le d s a m p le re fl e ct in g h if m e a n a n d u se d i n t h e p re se n t st u d y 3 .1 9 0 .4 2 .1 5 0 .6 2 0 .2 7 7 .6 b ic a rb o n a te s n o t m e a su re d h if m e a n n = 3 4 5 [ 5 ] 3 .1 3 0 .2 4 1 .3 1 0 .5 1 0 .2 3 6 .2 b ic a rb o n a te s n o t m e a su re d t o ta l b il e s a lt s (m m ) p h o sp h a ti d y lch o li n e (m m ) t o ta l p ro te in c o n te n t (m g /m l) f re e f a tt y a ci d s (m m ) c h o le st e ro l (m m ) p h p ri m a ry b u ff e r s p e ci e s s o d iu m c h lo ri d e (m m ) admet & dmpk 2(2) (2014) 71-79 increasing the biorelevance of simulated intestinal fluids doi: 10.5599/admet.2.2.40 75 table 3. the chromatographic conditions used for the analysis of drugs in the present study chromatographic conditions cinnarizine ketoconazole indomethacin danazol column fortis c18 (150×4.6mm, 5μm) bds c18 (250×4.6mm, 5μm) bds c18 (250×4.6mm, 5μm) bds c18 (250×4.6mm, 5μm) mobile phase methanol-watertriethylamine-acetic acid 75:25:0.4:0.6 v/v/v/v methanol-waterdiethylamine 74:26:0.1 acetonitrile-wateracetic acid 70:30:05 v/v/v acetonitrile-water 55:45 v/v flow rate (ml/min) 1 1 1 1.3 detection wavelength (nm) 254 240 266 286 injection volume (μl) 20 20 20 20 results and discussion cinnarizine was unstable in the pooled hif sample reflecting hifmean. figure 1 shows the cinnarizine degradation profile in the pooled hif sample at 37 °c together with the characteristics of the best fitted line drawn by using the data up to 3.20h. estimated degradation half-life is 1.44h. interestingly, solubility of cinnarizine has been previously reported to be 32.011.8 μg/ml and 47.317.4 μg/ml in hifs and fassif, respectively [16]. after reaching equilibrium in plain phosphate and plain maleate buffers filtrates were incubated at 37 °c for 48h at which time their cinnarizine content was measured to be zero. therefore, stability of cinnarizine in simulated intestinal media is also an issue. during the equilibration period, concentration peaked at 4h after initiation of oscillation, and this was considered as “apparent cinnarizine solubility”. apparent solubility in fassifcr-v2plus was significantly higher than in fassifcr-v2 (figure 2), i.e. inclusion of sodium oleate and cholesterol in increased the solubility of cinnarizine. this is in line with the high lipophilicity of the compound (table 1) and previously collected data on the effects of lipids on the solubilization capacity of mixed bile salt micelles for highly lipophilic compounds [17]. figure 1. cinnarizine concentration in the pooled hif sample reflecting hifmean as a function of time after initiation of incubation (37 o c). y = 18.834e -0.481x r 2 = 0.996 0 5 10 15 20 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 time (h) c o n c e n tr a ti o n (μ g /m l) koumandrakis et.al. admet & dmpk 2(2) (2014) 71-79 76 0 10 20 30 40 50 60 1 2 3 s o lu b il it y ( μ g /m l) fassifcr-v2plusfassifcr-v2fassifcr 0 10 20 30 40 50 60 1 2 3 s o lu b il it y ( μ g /m l) fassifcr-v2plusfassifcr-v2fassifcr figure 2. mean+sd (n=3) apparent cinnarizine solubility in simulated intestinal fluids (37 o c). please see in table 2 and in the text for compositions of simulated intestinal fluids. fassifcr-v2 was less useful than fassifcr in the prediction of ketoconazole, indomethacin, and danazol solubility in the pooled sample reflecting hifmean (figure 3). unlike with cinnarizine, inclusion of biorelevant concentrations of cholesterol and sodium oleate in fassifcr-v2 (i.e. fassifcr-v2plus) did not have significant impact on solubility of ketoconazole and danazol. also, the impact of bsa on ketoconazole, indomethacin, and danazol solubility in fassifcr and fassifcr-v2 was minimal (data not shown). this is in line with previous data indicating that luminal proteins may not be important to luminal drug solubility or their global simulation should be done cautiously [3]. in contrast, by increasing phosphatidylocholine concentration in fassif-v2 to 1mm, ketoconazole and danazol solubility in the pooled sample reflecting hifmean were successfully estimated (fassifcr-v21,pc, figure 3). fassifcr-v2plusmedian led to successful estimation of luminal solubility of ketoconazole and danazol in the pooled sample reflecting hifmedian (figure 4). for indomethacin, solubility was underestimated in all cases, due to the higher than average ph of the pooled samples reflecting hifmean and hifmedian; ph has much greater impact on indomethacin solubility than the rest of apis tested in the present study, due to its specific ionization properties (table 1, figures 3 and 4). admet & dmpk 2(2) (2014) 71-79 increasing the biorelevance of simulated intestinal fluids doi: 10.5599/admet.2.2.40 77 0 10 20 30 40 s o lu b il it y ( μ g /m l) hif mean fassif c -v2plusfassif cr -v2fassif cr phosphates maleates a fassif cr -v2 1,pc 0 10 20 30 40 s o lu b il it y ( μ g /m l) hif mean fassif c -v2plusfassif cr -v2fassif cr phosphates maleates a fassif cr -v2 1,pc 0 300 600 900 1200 1500 s o lu b il it y ( μ g /m l) hif mean fassif c -v2plusfassif cr -v2fassif cr phosphates maleates b fassif cr -v2 1,pc 0 300 600 900 1200 1500 s o lu b il it y ( μ g /m l) hif mean fassif c -v2plusfassif cr -v2fassif cr phosphates maleates b fassif cr -v2 1,pc 0 2 4 6 8 10 s o lu b il it y ( μ g /m l) hif mean fassif c -v2plusfassif cr -v2fassif cr phosphates maleates c fassif cr -v2 1,pc 0 2 4 6 8 10 s o lu b il it y ( μ g /m l) hif mean fassif c -v2plusfassif cr -v2fassif cr phosphates maleates c fassif cr -v2 1,pc figure 3. mean+sd (n=3) values for solubility of ketoconazole (a), indomethacin (b), and danazol (c) in the pooled hif sample reflecting hifmean, and in various simulated intestinal fluids (37 o c). please see table 2 and text for composition of simulated intestinal fluids. koumandrakis et.al. admet & dmpk 2(2) (2014) 71-79 78 0 10 20 30 40 s o lu b il it y ( μ g /m l) hifmedian fassifcr-v2plusmedian a 0 10 20 30 40 s o lu b il it y ( μ g /m l) hifmedian fassifcr-v2plusmedian a 0 300 600 900 1200 1500 s o lu b il it y ( μ g /m l) hifmedian fassifcr-v2plusmedian b 0 300 600 900 1200 1500 s o lu b il it y ( μ g /m l) hifmedian fassifcr-v2plusmedian 0 300 600 900 1200 1500 s o lu b il it y ( μ g /m l) hifmedian fassifcr-v2plusmedian b 0 10 20 30 40 s o lu b il it y ( μ g /m l) hifmedian fassifcr-v2plusmedian c 0 10 20 30 40 s o lu b il it y ( μ g /m l) hifmedian fassifcr-v2plusmedian c figure 4. mean+sd (n=3) values for solubility of ketoconazole (a), indomethacin (b) and danazol (c) in the pooled hif sample reflecting hifmedian and in fassifcr-v2plusmedian. concluding remarks estimation of solubility in median environment of upper small intestine in the fasted state seems adequate by data collected in fassif-v2plusmedian. in contrast, if solubility in the mean environment of upper small intestine is to be estimated, consideration of additional luminal characteristics is required, e.g. buffering species identity [18], non-micellar colloidal lipid structures [19], and lyso-phosphatidylcholine content [20]. perhaps a practical methodology to balance the lack of simulation of certain luminal components in fassif-v2 might be the use of fassif-v2 with high phosphatidylcholine concentration, e.g. 1mm. admet & dmpk 2(2) (2014) 71-79 increasing the biorelevance of simulated intestinal fluids doi: 10.5599/admet.2.2.40 79 acknowledgement part of this work was presented as a poster at the 9 th world meeting on pharmaceutics, biopharmaceutics and pharmaceutical technology, 31 march to 3 april 2014, lisbon, portugal. references [1] e.s. kostewicz, l. aarons, m. bergstrand, m.b. bolger, a. galetin, o. hatley, m. jamei, r. lloyd, x. pepin, a. rostami-hodjegan, e. sjögren, c. tannergren, d.b. turner, c. wagner, w. weitschies, j. dressman. eur. j. pharm. sci. 57 (2014) 300-321. [2] e. jantratid, n. janssen, c. reppas, j.b. dressman. pharm. res. 25 (2008) 1663-76. [3] m. vertzoni, a. diakidou, m. chatzilias, e. söderlind, b. abrahamsson, j.b. dressman, c. reppas. pharm. res. 27 (2010) 2187-2196. [4] d. psachoulias, m. vertzoni, j. butler, d. busby, m. symillides, j. dressman, c. reppas. pharm. res. 29 (2012) 3486-3498. [5] d. psachoulias, m. vertzoni, k. goumas, v. kalioras, s. beato, j. butler, c. reppas. pharm. res. 28 (2011) 3145-3158. [6] http://www.vcclab.org/lab/alogps/ (accessed 21 may 2014) [7] s. clarysse, j. brouwers, j. tack, p. annaert, p. augustijns eur. j. pharm. sci. 43 (2011) 260-269. [8] c.h. gu, d. rao, r.b. gandhi, j. hilden, k. raghavan. j. pharm. sci. 94 (2005) 199–208. [9] http://redpoll.pharmacy.ualberta.ca/drugbank/index.html (accessed 21 may 2014) [10] m. vertzoni, e. pastelli, d. psachoulias, l. kalantzi, c. reppas. pharm. res. 24 (2007) 909-917. [11] e.m. persson, a.s. gustafsson, a.s. carlsson, r.g. nilsson, l. knutson, p. forsell, g. hanisch, h. lennernäs, b. abrahamsson. pharm. res. 22 (2005):2141-2151. [12] m.vertzoni, h. archontaki, c. reppas. j. lipid res. 49 (2008) 2690-2695. [13] e. galia, e. nicolaides, d. hoerter, r. loebenberg, c. reppas, j.b. dressman pharm. res. 15 (1998) 698-705. [14] m. vertzoni, c. markopoulos, m. symillides, c. goumas, g. imanidis, c. reppas. mol. pharm. 9 (2012) 1189-1198. [15] l. kalantzi, e. persson, b. polentarutti, b. abrahamsson, k. goumas, j.b. dressman, c. reppas. pharm. res. 23 (2006) 1373-1381. [16] p. augustijns, b. wuyts, b. hens, p. annaert, j. butler, j. brouwers. eur j pharm sci. 57 (2014) 322332. [17] g.a. kossena, w.n. charman, b.j. boyd, c.j. porter. j. pharm. sci. 94 (2005) 481-492. [18] g. garbacz, b. kołodziej, m. koziolek, w. weitschies, s. klein. eur. j. pharm. sci. 51 (2014) 224-231. [19] a. müllertz, d. fatouros, j. smith, m. vertzoni, c. reppas. mol. pharm. 9 (2012) 237-247. [20] a. fuchs, b. kloefer, m. leigh, j. dressman. aaps annual meeting, november 10-14, 2013, san antonio, texas, usa (poster). ©2014 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://www.vcclab.org/lab/alogps/ http://redpoll.pharmacy.ualberta.ca/drugbank/index.html http://creativecommons.org/licenses/by/3.0/ manuscript doi: 10.5599/admet.4.1.277 54 admet & dmpk 4(1) (2016) 54-59; doi: 10.5599/admet.4.1.277 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper reversed phase parallel artificial membrane permeation assay for log p measurement zihao song, katsuhide terada, kiyohiko sugano* department of pharmaceutics, faculty of pharmaceutical sciences, toho university, 2-2-1, miyama, funabashi, chiba, 274-8510, japan *corresponding author: e-mail: kiyohiko.sugano@phar.toho-u.ac.jp; tel.: +81 47 472 1494 received: march 17, 2016; published: march 31, 2016 abstract a reversed phase parallel artificial membrane permeation assay (rp-pampa) was newly invented for log p measurement. an oil/water/oil sandwich was constructed using a conventional pampa instrument. 1 % agarose was used to improve the physical stability of the water phase. a linear correlation between log p and the apparent permeability was observed in the -0.24 < log p < 2.85 region (r 2 = 0.98). rp-pampa was also applied to pka measurement. keywords octanol, partition, parallel artificial membrane permeation assay, pka, drug introduction high throughput physicochemical profiling of a drug is still challenging in early drug discovery. various methods have been proposed for octanol – water partition coefficient (log p), solubility, and pka measurements [1]. the parallel artificial membrane permeation assay (pampa) has been widely used in drug discovery as pampa is compatible with high throughput screening (hts) [2,3]. in normal phase (np-) pampa methods, a lipid phase is immobilized on a filter (usually a 96 well filter plate) and the permeability of a drug across the lipid membrane is measured. previously, faller et al. applied np-pampa to log p measurement [4]. the apparent permeability (papp) across the octanol impregnated filter membrane was found to correlate with logp. however, the papp – log p curve showed a bell-shaped relationship with a plateau around log p = 1. therefore, it was impossible to estimate the log p values around log p = 1. a bellshaped relationship between papp and lipophilicity of drugs is usually observed in np-pampa [5]. the purpose of the present study was to overcome the drawback of np-pampa for log p measurement. a reversed phase pampa (rp-pampa) method for log p measurement was newly invented. in rp-pampa, an oil/water/oil sandwich was constructed using a conventional pampa instrument (figure 1). in addition, rp-pampa was applied for pka measurement, especially for low solubility compounds. http://www.pub.iapchem.org/ojs/index.php/admet/index admet & dmpk 4(1) (2016) 54-59 reversed phase pampa doi: 10.5599/admet.4.1.277 55 figure 1. schematic configuration of rp-pampa materials and methods materials octanol, pentoxifylline, 1-naphthol, dipyridamole, and acid blue 9 were purchased from tokyo chemical industry (tokyo, japan). agar powder, agarose s, agarose h, prednisone, sulfamethoxazole, carbamazepine, caffeine, chlormphenicol, ethanol, trisodium citrate, sodium dihydrogenphosphate, disodium hydrogenphosphate, sodium hydroxide solution, propranolol hydrochloride, warfarin sodium, piroxicam, and ketoprofen were purchased from wako pure chemicals (tokyo, japan). phenacetin was purchased from yamamoto corporation (osaka, japan). the other reagents were of analytical grade. papp measurement the schematic configuration of rp-pampa is shown in figure 1. in rp-pampa, the water phase (water membrane) was immobilized on the hydrophilic filter with the aid of agarose. agarose s was dissolved in hot water or a buffer at 1.0 % and then poured into a hydrophilic filter (multi screen-hv, pore size 0.48 μm, low protein binding, millipore). a model drug was dissolved in octanol at 10 mm and added to the donor plate (downside, 300 μl). the filter plate was then put on the donor plate. the filter plate was filled with 200 μl of octanol. after 16 hour incubation at room temperature (25 ± 1 °c), both octanol phases were diluted tenfold by ethanol and the drug concentrations in the donor and accepter sides were measured by uv spectroscopy. for pka measurement sodium phosphate and sodium citric acid buffers (100 mm of anion species) were used to construct the water membrane. papp was calculated as previously reported [6]: c c p a v v t a equilibrium app d a ln(1 / ) (1/ 1/ )       (1)  c c v c v v vequilibrium d d a a d a/( )     (2) where ca and cd are the drug concentrations in the donor and acceptor phases at time t, respectively. vd and va are the volumes of the donor and acceptor phases, respectively. a is the membrane surface area (0.28 cm 2 ). results construction of rp-pampa we first investigated the stability of the water phase (water membrane) constructed on the hydrophilic z. song et al. admet & dmpk 4(1) (2016) 54-59 56 filter with the aid of agarose. 1.0 % concentration was selected to enable pipetting of the hot sol phase while maintaining the physical strength of the agarose gel. it was found that at least 30 μl of 1 % agarose was required to provide sufficient physical strength for papp measurement. the ager powder was found to form a less stable water phase compared to agarose s and h. in a preliminary study, it was found that more than 10 hours were required to achieve a steady – state flux across the water phase (data not shown). therefore, the agarose s 30 μl water membrane and 16 hour incubation time were employed in the following studies. log papp – log p relationship the log papp and log p data are summarized in table 1 [1,7,8]. figure 2 shows the correlation between log papp and log p. a linear correlation was observed in the -0.24 < log p < 2.85 region (r 2 = 0.98). the slope of the log-log plot was -0.48. table 1. papp and log p drug log p (literature) a log papp (cm s -1 , mean ± s.d., n = 6) phenacetin 1.58 -6.38 ± 0.04 caffeine 0.10 -5.53 ± 0.02 carbamazepine 2.1 -6.52 ± 0.04 prednisone 1.56 -6.25 ± 0.01 chloramphenicol 1.14 -6.13 ± 0.01 sulfamethoxazole 0.70 -5.96 ± 0.02 1-naphthol 2.85 -6.95 ± 0.03 pentoxifylline 0.38 -5.78 ± 0.02 a refs. [1,7,8] figure 2. log p – log papp relationship. ph papp profile the ph papp profiles were shown in figure 3 and table 2. the pka values were obtained as the intersection of the slope and the horizontal lines. estimated and literature pka values are shown in table 3 [1,9]. y = -0.4835x 5.5584 r² = 0.9842 -7.0 -6.8 -6.6 -6.4 -6.2 -6.0 -5.8 -5.6 -5.4 -5.2 -5.0 0 1 2 3 lo g p a p p ( cm /s e c) logp admet & dmpk 4(1) (2016) 54-59 reversed phase pampa doi: 10.5599/admet.4.1.277 57 figure 3. ph – papp relationship. mean ± s.d. n = 3. table 2. papp of dissociable drugs at each ph a compound ph papp (10 -6 cm sec -1 ) compound ph papp (10 -6 cm sec -1 ) ketoprofen 3.1 0.17 ± 0.03 dipyridamole 3.9 0.31 ± 0.02 (log p = 3.2) 3.5 0.17 ± 0.07 (log p = 3.9) 4.5 0.30 ± 0.01 4.1 0.15 ± 0.05 5.1 0.20 ± 0.02 4.5 0.18 ± 0.04 5.5 0.11 ± 0.02 5.0 0.28 ± 0.09 5.9 0.07 ± 0.02 5.5 0.44 ± 0.10 6.5 0.05 ± 0.01 6.0 1.33 ± 0.18 7.0 0.05 ± 0.03 6.5 2.00 ± 0.23 7.5 0.04 ± 0.00 piroxicam 3.0 0.81 ± 0.07 propranolol 8.0 1.05 ± 0.06 (log p = 2.0) 3.5 0.81 ± 0.19 (log p = 2.9) 8.5 0.72 ± 0.08 4.0 0.77 ± 0.07 9.0 0.63 ± 0.10 4.5 0.68 ± 0.07 9.5 0.42 ± 0.04 5.0 0.81 ± 0.15 10.0 0.40 ± 0.12 5.5 0.98 ± 0.13 10.5 0.39 ± 0.04 6.0 1.62 ± 0.08 11.0 0.38 ± 0.01 6.5 2.37 ± 0.24 11.5 0.38 ± 0.03 warfarin 3.0 0.16 ± 0.01 (log p = 3.1) 3.5 0.16 ± 0.01 4.0 0.16 ± 0.02 4.5 0.23 ± 0.02 5.0 0.22 ± 0.04 5.5 0.23 ± 0.02 6.0 0.34 ± 0.04 6.5 0.80 ± 0.05 a mean ± s.d. n = 3. measured at 25 °c. the buffer concentration was 100 mm. z. song et al. admet & dmpk 4(1) (2016) 54-59 58 table 3. pka values drug this study a literature b ketoprofen 4.9 4.0 piroxicam 4.8 4.7 warfarin 5.7 5.0 dipyridamole 5.9 6.1 propranolol 9.4 9.5 a measured at 25 °c. the buffer concentration was 100 mm. b refs. [1], [9]. discussion in this study, rp-pampa for log p measurement was investigated for the first time. 1.0 % agarose was used to improve the physical stability of the water membrane. the mesh size of agarose is significantly larger than the size of drug molecules so that it does not affect the diffusion coefficient of drugs [10]. by using rp-pampa, log p in the -0.24 < log p < 2.85 range can be accurately measured. the measurable range can be expanded by using a more sensitive quantitation method such as lc-ms. the slope of the papp – log p relationship was 0.45, which is significantly smaller than 1. if papp follows the solubility – partition theory for membrane permeation, i.e., papp = pd/h where d is the diffusion coefficient and h is the thickness of the membrane, the slope of the log-log plot should be unity [11]. the reason for this deviation is not clear. previously, kwon et al. reported a poly(dimethylsiloxane)(pdms) permeation assay, which might be regarded as a kind of reversed phase membrane permeation assay [12]. however, the configuration of the pdms permeation assay was largely different from the one used in the present study that is usually referred as pampa. in the pdms permeation assay, a side-by-side single diffusion chamber was employed. a pdms membrane was put between two chambers filled with aqueous bulk fluids. in addition, pdms disks were added to both the donor and acceptor sides as dosing and sampling (extracting) phases, respectively. the aqueous phases were stirred by magnetic stirrers. in the pdms permeation assay, a good correlation was observed between log papp and log p in the range of log p > 3 even though pdms was used instead of octanol as the oil phase. for low solubility drugs, it has been difficult to measure pka by using conventional methods such as ph titration. the ph – solubility profile can be used to estimate pka for low solubility drugs [13,14]. however, this method may not be accurate due to aggregate formation, low detection limit, etc. in rp-pampa, a drug is solubilized in the organic solvent phase. therefore, it would become possible to measure pka for low solubility drugs by using rp-pampa. as the ph of the water membrane was changed in rp-pampa, the ph – papp relationship should become a mirror image of that for np-pampa. however, the ph – papp relationship deviated from the henderson – hasselbalch equation. therefore, the pka values of the model drugs were estimated as the intersection of the slope and the horizontal lines. the pka values of acidic drugs were underestimated by the rp-pampa method. the reason for this deviation is not clear. one possible reason may be that the incubation time of 16 hours might not be sufficient to achieve a steady state at ph > pka for acids. the pka of diclofenac has been reported to be ca. 4.0 in most cases in the literature. however, pka of 5.7 was obtained from the ph-solubility profile [15]. the papp value at a ph where a drug molecule is undissociable (intrinsic water permeability, pw,int) also correlated with log p. however, pw,int deviated from the log papp – log p line for the undissociable drugs about 0.3 log unit. the difference of the water membrane (pure water vs. a buffer) could be a reason for the discrepancy. the present rp-pampa method admet & dmpk 4(1) (2016) 54-59 reversed phase pampa doi: 10.5599/admet.4.1.277 59 needs to be improved for pka measurement in the future. in conclusion, in the present study, rp-pampa for log p measurement was constructed for the first time. 1.0 % agarose can be used to stabilize the water membrane. rp-pampa was applied to log p and pka measurements. as pampa is compatible with the current hts instrument, rp-pampa will be a useful tool in drug discovery. references [1] a. avdeef. absorption and drug development. hoboken: wiley-interscience, nj; 2003. [2] m. kansy, f. senner, k. gubernator. j. med. chem. 41 (1998) 1007-1010. [3] a. avdeef, s. bendels, l. di, b. faller, m. kansy, k. sugano, y. yamauchi. j. pharm. sci. 96 (2007) 2893-2909. [4] b. faller, h.p. grimm, f. loeuillet-ritzler, s. arnold, x. briand. j. med. chem. 48 (2005) 2571-2576. [5] m. kansy, h. fischer, k. kratzat, f. senner, b. wagner, i. parrilla. high-throughput artifical membrane permeability studies in early lead discovery and development. in: testa b, van de waterbeemd h, folkers g, guy r, editors. pharmacokinetic optimization in drug research. zürich: wiley-vch; 2001. p. 447-464. [6] k. sugano, h. hamada, m. machida, h. ushio. j. biomol. screen. 6 (2001) 189-196. [7] k.j. box, j.e. comer. curr. drug metab. 9 (2008) 869-878. [8] drugbank. http://www.drugbank.ca/. [9] k. sugano. biopharmaceutics modeling and simulations: theory, practice, methods, and applications. new jersey: john wiley & sons, inc.; 2012. [10] n. fatin-rouge, k. starchev, j. buffle. biophys. j. 86 (2004) 2710-2719. [11] t. xiang, y. xu, b. d. anderson. j. membr. biol. 165 (1998) 77-90. [12] j.-h. kwon, b.i. escher. environ. sci. technol. 42 (2008) 1787-1793. [13] h. krebs, j. speakman. j. chem. soc. (1945) 593-595. [14] i. zimmermann. int. j. pharm. 13 (1982) 57-65. [15] a. pobudkowska, u. domańska. chem. ind. chem. eng. q. 20 (2014) 115-126. ©2016 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://www.drugbank.ca/ http://creativecommons.org/licenses/by/3.0/ manuscript doi: 10.5599/admet.2.2.44 98 admet & dmpk 2(2) (2014) 98-106; doi: 10.5599/admet.2.2.44 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper high throughput determination log po/w/pka/log do/w of drugs by combination of uhplc and ce methods joan marc cabot, xavier subirats, elisabet fuguet and martí rosés* departament de química analítica – institut de biomedicina, universitat de barcelona, martí i franquès 1-11, 08028 barcelona *corresponding author: e-mail: marti.roses@ub.edu; tel.: +34-934039275; fax: +34-934021233 received: july 03, 2014; revised: july 11, 2014; published: july 18, 2014 abstract in 1997 valkó et al. developed a generic fast gradient hplc method, based on the calculation of the chromatographic hydrophobicity index (chi) from the gradient retention times, in order to measure lipophilicity. we have employed the correlations between chi and log po/w and adapted the rapid gradient hplc method to uhplc obtaining excellent resolution and repeatability in a short analysis time (< 4min). log po/w values can be easily obtained from these chi measurements but, unfortunately, these correlations are only valid for non-ionized compounds. consequently, in order to determine the effective log po/w value at a particular ph, a fast high-throughput method for pka determination was required. the is-ce method, based on the use of internal standards (is) and capillary electrophoresis (ce), is a fast and attractive alternative to other methods for pka determination, since it offers multiple advantages compared to them: low amounts of test compounds and reagents are needed, high purity is not required, specific interacti ons between test compounds and buffers are corrected, etc. in addition, it allows the determination of a pk a value in less than 5 minutes. both chi and is-ce have been combined in order to describe a high throughput alternative in the determination of the lipophilicity profiles of bioactive compounds. keywords: lipophilicity, drug discovery, acidity constants, chromatographic hydrophobicity index, chi, internal standard introduction the drug discovery and development process requires the high-throughput determination of physicochemical parameters of drug candidates. parameters measuring lipophilicity, acidity, solubility or protein interaction must be determined in a fast way for a high number of compounds of potential interest to select the most promising compounds for further development. one of the most important physical properties affecting the biological activity of substances is lipophilicity, traditionally expressed as the logarithm of the octanol-water partition coefficient (log po/w). in the common case of drug candidates with acid-base properties, the effective partition rate at a particular ph is noted as log do/w. although being these the most widely used liphophilicity indexes, experimental reference procedures are usually time consuming and require a high purity and a relatively high amount of sample. in order to decrease the analysis time and overcome these limitations, instrumental analytical http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:marti.roses@ub.edu admet & dmpk 2(2) (2014) 98-106 high throughput determination of log do/w doi: 10.5599/admet.2.2.44 99 methods have been developed to determine lipophilicity, and the values therefrom obtained in a particular lipophilicity scale have been correlated to log po/w or log do/w scale. about twenty year ago klára valkó and péter slégel [1] proposed a new hydrophobicity index measured by a reversed-phase hplc method, named ϕ0, defined as the concentration of organic modifier in the mobile phase (methanol or acetonitrile) which is required to obtain a retention time (tr) double than the dead time (t0). in these conditions, the logarithm of the retention factor equals to zero (log k’=log ((tr-t0)/t0)=0) and, therefore, the compound is equally distributed between the mobile and the stationary phases. the higher the value, the more hydrophobic is the compound. this index is characteristic of a compound and independent of the column dimensions, the mobile phase composition and flow-rate. it only depends on the stationary phase, the particular organic modifier employed, the temperature and, for acidic or basic compounds, the ph. however, ϕ0 values were experimentally determined in isocratic mode using several mobile phases containing low fractions of methanol or acetonitrile, which involved long run times for hydrophobic substances. a few years later, valkó and co-workers [2] proposed a new highthroughput chromatographic hydrophobicity index (chi), based on ϕ0 but derived from retention times observed in a fast gradient reversed-phase hplc method. compounds of previously determined chi values can be used as calibration set, and the retention times of new compounds can be correlated to their corresponding chi values. in that work the c18 column employed was an ods2-ik5 inertsil with the dimensions of 150 × 4.6 mm, and the gradient program lasted for about 15 min. in a paper published in 2001 in collaboration with michael h. abraham [3] a much shorter column was used, a 50 x 4.6 mm phenomenex luna c-18(2), and therefore gradient times were reduced to 5 minutes. in addition, correlations between chi and log po/w were improved by means of hydrogen-bond acidity descriptors, since it was found that the major difference between these scales is their sensitivity towards the hydrogen bond acidity of the compounds. however, in the case of compounds with acid/base properties, poor correlations between chi and log do/w have been obtained so far. in 2002 stephen f. donovan and mark c. pescatore [4] used a 20 x 4.2 mm octadecyl–poly(vinyl alcohol cartridge and a fast gradient methanolic elution in order to estimate log po/w from chromatographic data. due to the ph stability of this polymeric column, the ph of the mobile phase could be adjusted to the desired value to ensure the ionizable analyte to be in its neutral form. nowadays, uhplc technology presents also an excellent stability of c18 columns over a wide range ph, besides higher resolution and sensitivity, and shorter equilibration times within consecutive runs in gradient elution. since mostly of the potential drug candidates are acids or bases, log do/w is a fundamental parameter to be determined in the drug discovery process in order to estimate the pharmacokinetics of a compound of interest. especially relevant is the log do/w at ph 7.4, the physiological value, which indicates the lipophilicity of a drug in the blood plasma. as long as this effective partition rate at any particular ph can be estimated from the measured log po/w for the neutral substance and the pka of the compound, we propose a highthroughput methodology based on the combination of uhplc for chi measurements and ce for pka determination in order to define the log do/w profile of drug candidates. in a series of previous works [5-7], we have proposed a high-throughput method to determine acidity constants of weak acids and bases by capillary zone electrophoresis. this method is based on the use of an internal standard (is) of similar nature and acidity constant as the test compound, which is injected into the capillary just after the sample compound and analyzed in the same electrophoretic run. this method requires only two electrophoretic runs: a first one at a ph where both the test compound and the is are totally ionized, and a second one at another ph at which both are partially ionized. thus, from the mobilities of the compounds at these two ph values and the pka value of the is, the pka of the analyte can cabot et al. admet & dmpk 2(2) (2014) 98-106 100 be easily calculated [5,8]. besides its rapidness, the main advantages of this method are the compensation of systematic errors due to simultaneous analysis of analyte and is, and the unnecessity of an accurate ph measurement. experimental apparatus for uhplc measurements, a shimadzu (kyoto, japan) nexera uhplc system was used. the system was equipped with two lc-30ad high-pressure pumps, a dgu-20a5 online degasser, a cto-10asvp oven thermostatized at 25 o c, a sil-30ac autosampler, a spd-m20a diode array detector and a cbm-20alite controller. retention data were obtained from a waters (milford, ma, usa) acquity beh c18 column, 50 mm × 2.1 mm, which allowed the study of basic drugs because of their extended working ph range (up to 12). injected samples were prepared in dmso at a concentration of 0.5 mg/ml and thus, due to the melting point of dmso (19 o c), the autosampler temperature was set at 20 o c. aqueous buffers were, in all cases, 50 mm ammonium acetate at the desired ph value. acidic buffers were prepared from glacial acetic acid and the ph was adjusted with small volumes of concentrated ammonia (25%), and basic buffers were prepared inversely. medium acidic and basic buffers were obtained by solving the salt and adjusting the ph with concentrated ammonia or glacial acetic acid. capillary electrophoresis experiments were performed using a p/ace mdq beckman instrument (palo alto, ca, usa), equipped with a diode array detector. capillary was made of fused silica 50 µm i.d., 375 µm o.d., 35.2 cm length (25 cm to the detector) and purchased from composite metal services ltd (shipley, west yorkshire, uk). the temperature of the capillary was 25.0±0.1 o c and the applied voltage during separation was 20 kv. is and analyte samples were sequentially injected at 0.5 psi for 3 s and analyzed in the same run. additional 0.5 psi of hydrodynamic flow during separation was applied. the capillary was initially conditioned with 1 m naoh (2.0 min), water (0.5 min), and running buffer (2.0 min). between replicates with the same buffer the capillary was not rinsed. 20 psi of pressure was applied during rinsing processes. running buffers were prepared at different ph values and 50 mm ionic strength as described elsewhere [8]. injected samples were prepared in methanol/water mixtures at a concentration of 0.1 mg/ml, containing dmso as neutral marker. appropriate is were selected from the set proposed in previous works [8]. all samples and buffers were filtered through a nylon mesh 0.45 µm porous size (whatman, maidstone, uk) before use. ph measurements were performed with a combined crison (alella, spain) 5014 electrode in a crison glp22 ph meter. the electrode system was standardized with ordinary aqueous buffers of ph 4.01, 7.00 and 9.21. chemicals acetonitrile hplc gradient grade was purchased from vwr (west chester, pa, usa). water was purified by the milli-q® plus system from millipore (billerica, ma, usa) with a resistivity of 18.2 mω cm. the chemicals used for buffer preparation were anhydrous sodium acetate (>99.6%) and glacial acetic acid from j.t. baker (deventer, the netherlands), 2-(cyclohexylamino)ethanesulfonic acid (ches, >99%), and 3-(cyclohexylamino)-1-propanesulfonic acid (caps, >98%) from sigma (st. louis, mo, usa), 2,2-bis(hydroxymethyl)-2,2’,2’’-nitrilotriethanol (bistris, >99.9%) from fluka (buchs, switzerland), sodium hydroxide, hydrochloric acid and ammonia from merck (darmstadt, germany). dimethyl sulfoxide (dmso, >99.9%) was purchased from j.t. baker. the injected compounds were purchased from merck, sigma– aldrich and fluka, or obtained from almirall (barcelona, spain), all of them in high purity grade. admet & dmpk 2(2) (2014) 98-106 high throughput determination of log do/w doi: 10.5599/admet.2.2.44 101 results and discussion measurement of chimecn values by uhplc the hplc fast gradient method developed by valkó and co-workers [3] was transferred to our uhplc system taking into consideration the column and particle sizes, the flow rate, the injection volume, and the dwell time of the uhplc instrument. the final uhplc conditions, together with the original hplc ones, are shown in table 1. dmso was used as sample solvent because its ability to solve hydrophobic compounds sparingly soluble in mecn/water mixtures. in the present work mecn was used as organic modifier as long as this solvent allows better correlations between log po/w and chi than methanol. several substances covering a wide range of chimecn and log po/w values were tested as calibration standards. finally, the representative compounds shown in table 2 were selected. they are mainly unionizable substances that can be used at any ph value (acetanilide and phenones), but 4-hydroxybenzyl alcohol is a phenol that should be used in its neutral form below ph 8. when using aqueous buffers above this ph value, caffeine was used instead of 4-hydroxybenzyl alcohol. a representative chromatogram of the calibration set is shown in figure 1. after plotting the chimecn values against the retention times in the fast gradient elution, a quadratic calibration curve is obtained with a typical coefficient of determination (r 2 ) higher than 0.998 and a root mean square error (rmse) lower than 1.5. table 1. comparison between hplc and uhplc conditions to measure chi values hplc [3] uhplc (present work) column c18, 50 x 4.6 mm, 5 μm beh c18, 50 x 2.1 mm, 1.7 μm flow rate 2.0 ml/min 0.5 ml/min temperature 25 o c 25 o c fast gradient program % mecn % mecn 0.0 min -> 0% 0.0 min -> 0% 0.5 min -> 0% 0.4 min -> 0% 3.0 min -> 100% 2.5 min -> 100% 3.5 min -> 100% 2.9 min -> 100% 3.7 min -> 0% 3.1 min -> 0% 4.5 min -> 0% 3.8 min -> 0% injection volume and solvent 3 μl mecn/aqueous buffer 0.2 μl dmso similarly to log do/w scale, chi values of acidic or basic compounds depend on the ionization degree. the lower the ionization, the higher the chi value [9,10]. therefore, acidic substances present the highest chi when aqueous buffers of low ph are used, and consequently basic compounds exhibit the reversed trend. on the other hand, non-ionizable analytes show nearly the same chi values independently of the ph. in the present work all studied analytes were injected at three different ph values (3.0, 7.4 and 11.0), and a representative collection supporting the above statements is presented in table 3. table 2. compounds used as calibration standards. compound chimecn [11] log po/w [12] pka [12] 4-hydroxybenzyl alcohol a 18.91 0.25 9.83 caffeine b 24.45 -0.07 0.60 acetanilide 42.38 1.16 acetophenone 63.26 1.58 propiophenone 76.93 2.19 butyrophenone 87.14 2.66 valerophenone 96.53 3.28 hexanophenone 105.13 3.79 heptanophenone 112.79 4.32 a) used below ph 8. b) used above ph 8. cabot et al. admet & dmpk 2(2) (2014) 98-106 102 figure 1. chromatogram of the calibration set after subtraction of the dmso blank. uhplc conditions according to table 1, 50 mm ammonium acetate ph 7.4 as aqueous buffer. table 3. chimecn values obtained at ph 3.0, 7.4 and 11.0. compounds chimecn analyte behavior ph 3.0 ph 7.4 ph 11.0 paracetamol 21.2 20.7 -137.2 acidic theophylline 21.4 20.2 -55.9 acidic atenolol 19.1 18.6 32.1 basic colchicine 43.9 43.4 43.8 salicylic acid 46.0 21.1 -12.5 acidic benzoic acid 48.0 -64.4 -143.2 acidic hydrocortisone 50.2 50.0 50.3 metoprolol 38.1 35.9 61.5 basic procaine 24.9 30.1 62.9 basic propranolol 54.2 51.0 80.9 basic warfarin 82.9 41.8 25.9 acidic ketoconazole 70.8 83.4 83.9 basic lidocaine 35.4 77.9 86.4 basic haloperidol 70.1 66.4 88.9 basic thymol 90.0 89.9 89.5 acidic indomethacin 90.7 54.5 37.7 acidic napththalene 97.6 97.6 97.5 phenothiazine 98.4 98.4 98.2 reserpine 82.7 95.6 98.7 basic anthracene 111.5 111.7 112.3 determination of log po/w/pka/log do/w as mentioned in the introduction, valkó and co-workers proposed an equation to correlate chimecn and log po/w indexes, which involved the abraham’s solute hydrogen-bond acidity descriptor (a, also noted in the original paper as σα2 h ) [3]:  o/w mecn log 0.054 chi 1.319 a 1.877 86, 0.970, 0.29, 655        p n r s f (1) admet & dmpk 2(2) (2014) 98-106 high throughput determination of log do/w doi: 10.5599/admet.2.2.44 103 table 4. validation set used for the determination of log po/w and log do/w. behavior ref. compound chimecn a log po/w (chi) pka (is-ce) log d7.4 (chi+is-ce) log d7.4 (reference) non-ionizable 1 acetanilide 40.70 0.41 0.86 0.86 1.19 a 2 acetophenone 62.50 0.00 1.50 1.50 1.58 a 3 anthracene 112.30 0.00 4.19 4.19 4.49 a 4 butyrophenone 87.90 0.00 2.87 2.87 2.65 a 5 caffeine 25.90 0.00 -0.48 -0.48 -0.10 a 6 colchicine 43.90 0.26 0.83 0.83 1.09 a 7 heptanophenone 112.10 0.00 4.18 4.18 4.46 a 8 hexanophenone 105.20 0.00 3.80 3.80 3.69 a 9 hydrocortisone 50.30 0.73 1.80 1.80 1.58 a 10 naphthalene 97.50 0.00 3.39 3.39 3.19 a 11 phenothiazine 98.40 0.13 3.61 3.61 4.11 a 12 propiophenone 77.20 0.00 2.29 2.29 2.24 a 13 valerophenone 97.00 0.00 3.36 3.36 3.40 a basic 14 papaverine 66.93 0.00 1.74 6.41 1.69 15 reserpine 98.70 0.31 3.86 6.64 3.79 3.89 a 16 2,4-lutidine 49.35 0.00 0.79 6.81 0.69 17 ketoconazole 83.90 0.00 2.65 6.81 2.55 3.42 a 18 trazodone 76.34 0.00 2.25 6.84 2.14 2.54 b 19 pilocarpine 20.86 0.00 -0.75 7.08 -0.92 20 2,4,6-trimethylpyridine 57.17 0.00 1.21 7.51 0.85 21 lidocaine 86.40 0.26 3.13 7.89 2.52 1.61 a 22 clonidine 51.30 0.42 1.44 8.10 0.67 0.62 b 23 bupivacaine 101.75 0.26 3.96 8.19 3.11 24 quinine 66.42 0.23 2.01 8.45 0.93 25 haloperidol 88.90 0.31 3.34 8.56 2.15 2.77 a 26 procaine 62.90 0.23 1.82 9.08 0.14 0.24 a 27 propranolol 80.90 0.29 2.88 9.48 0.83 1.23 a 28 metoprolol 61.50 0.29 1.83 9.53 -0.26 -0.25 a 29 atenolol 32.10 0.78 0.89 9.67 -1.33 -1.81 a acidic 30 paracetamol 21.20 0.91 0.47 9.58 0.46 0.35 a 31 2-chlorophenol 63.64 0.33 2.00 8.50 1.96 32 theophylline 21.40 0.35 -0.27 8.39 -0.31 -0.07 a 33 methylparaben 52.33 0.66 1.82 8.35 1.78 34 3,5-dinitrophenol 85.97 0.83 3.86 8.18 3.80 35 4-hydroxybenzaldehyde 38.36 0.66 1.07 7.61 0.86 36 phenobarbital 51.84 0.52 1.61 7.53 1.37 1.51 b 37 vanillin 41.37 0.44 0.94 7.36 0.61 38 4-nitrophenol 55.31 0.67 1.99 7.09 1.51 39 warfarin 82.90 0.31 3.01 5.17 0.83 0.89 a 40 benzoic acid 48.00 0.57 1.47 4.22 -1.40 -1.27 a 41 indomethacin 90.70 0.57 3.77 4.16 0.88 a) shake-flask method, ref. [13]. b) experimental value, ref. [14]. with the aim of validating the uhplc method proposed, chimecn were measured for several nonionizable, acidic and basic compounds in the log po/w range between -0.07 and 4.45. all analytes were injected using aqueous mobile phases of ph 3.0, 7.4, and 11.0, and the highest value was considered for log po/w determination. these 41 compounds are listed in table 4, together with the chimecn values of neutral species determined in the present work, the a solute descriptors obtained from the acd/percepta platform [15], and the log po/w determined from chimecn and a according to eq. (1). there is a good correlation between literature log po/w values (mlog p values compiled in bio-loom database [12]) and the ones measured in the present work (slope: 0.99(±0.06), intercept: -0.15(±0.15), r: 0.940, rmse: 0.45). basic cabot et al. admet & dmpk 2(2) (2014) 98-106 104 solutes correlate slightly worse than non-ionizable and acidic compounds. nevertheless, from the 16 studied bases only 6 presented deviations from literature values higher than 0.7 log po/w units. in four of those cases the determined log po/w values were lower than the bibliographic ones (compounds 14, 16, 17, and 19 in table 4), and in two cases (solutes 21 and 29) the reverse trend was shown. as previously commented, poor correlations are obtained between log do/w and chimecn values for ionized compounds. however, the effective distribution rate of a drug at a particular ph value can be estimated from its acidity constant and the partition coefficients of the neutral and fully ionized species [16]. the current state of the art does not allow the direct measurement of the partition coefficient of ions but it can be assumed that a typical value is 3.15 units lower than that of the neutral species, according to the work published by donovan and pescatore [4]. thus, log do/w values of monoprotic compounds can be calculated from the following expressions at any desired ph value:    a aph-p ph-po/w o/w o/wacidic compound: log =log +( -3.15)·10 log 1+10k kd p p (2)    a ap -ph p -pho/w o/w o/wbasic compound: log =log + ( -3.15)·10 log 1+10k kd p p (3) consequently, the measurement of aqueous acidity constants was required in order to estimate log do/w values and the is-ce fast method was presented as a very appropriate option. this method allows an accurate determination of acidity constants, which normally present an excellent match with reference values found in literature. in the present work, the correlation between the measured pka values for some of the studied acidic and basic compounds (table 4) and the ones found in literature was remarkable (slope: 1.01(±0.01), intercept: -0.07(±0.09), r: 0.998, rmse: 0.11). table 4 shows the log do/w values at ph 7.4 resulting from log po/w determinations and pka measurements (eqs. (2-3)) and those obtained from reference methods [13,14]. as shown in figure 2, there is a good agreement between both sets of values, with the only exception of lidocaine, which presented yet one of the worst matches in log po/w determination. figure 2. correlation between reference log do/w values at ph 7.4 (table 4) and those obtained from log po/w and pka measurements (eqs. (2-3)). linear regression (continuous lines) and the statistics are also shown, together with the lines corresponding to ±2·rmse (dashed lines). admet & dmpk 2(2) (2014) 98-106 high throughput determination of log do/w doi: 10.5599/admet.2.2.44 105 figure 3 shows the log do/w profile obtained in the present work for vanillin and clonidine, which present acidic and basic nature, respectively. for the sake of comparison, the curve calculated by acd/labs software is also presented, together with their corresponding calculated and measured pka values. in the case of vanillin, the acd/labs calculated partition ratio (1.32) is moderately higher than the measured one (0.94), whereas both pka values, estimated and experimental, match very well (7.30 and 7.36, respectively). thus, the acd/labs estimated curve shifts up from the experimental one. for clonidine, the acd/labs calculated log po/w (2.05) is higher than the chimecn measured one (1.44) by 0.6 units, whereas the pka (7.90) is lower than the is-ce determined one (8.10). therefore, the acd/labs estimated curve moves up and left from the measured one, being the values determined in the present work more accurate according to literature (8.05 and 1.43, for pka and log po/w, respectively [12]). figure 3. log do/w profile for vanillin and clonidine according to eqs. (2-3) (continuous line) and calculated profile using acd/labs software (dotted lines), together with their corresponding measured and calculated pk a values. conclusions in the present work the combination of two high-throughput methods have been proposed in order to estimate the log do/w of an ionizable compound at any desired ph. on the one hand, a fast gradient reversed-phase methodology was successfully transferred into uhplc in order to determine log po/w from chromatographic retention times, based on the formerly developed chi lipophilicity index. after calibration of the system, a single fast-gradient run performed within 4 minutes is enough to determine the chimecn value of a compound, and then the value can be transformed into log po/w lipophilicity scale by means of the established equation. in the case of analytes with acid/base properties, runs performed at 3 different ph values (3.0, 7.4 and 11) are recommended, and the highest chimecn is selected. on the other hand, the is-ce allows the determination of a pka value from only two runs, lasting less than 5 minutes (about 2 min each). finally, from both the log po/w of the neutral species (uhplc-chi procedure) and the pka value (is-ce method), the effective distribution rate log do/w can be estimated at any particular ph value. cabot et al. admet & dmpk 2(2) (2014) 98-106 106 acknowledgements: financial support from the ministerio de ciencia e innovación of the spanish government and the fondo europeo de desarrollo regional (feder) of the european union (project ctq2010-19217/bqu) is acknowledged. jmc also thanks the secretaria d'universitats i recerca of the departament d'economia i coneixement of the catalan government for a supporting scholarship (2013fi_b1 00209). references [1] k. valkó, p. slégel, j. chromatogr. a 631 (1993) 49-61. [2] k. valkó, c. bevan, d. reynolds, anal. chem. 69 (1997) 2022-2029. [3] k. valkó, c.m. du, c. bevan, d.p. reynolds, m.h. abraham, curr. med. chem. 8 (2001) 1137-1146. [4] s.f. donovan, m.c. pescatore, j. chromatogr. a 952 (2002) 47-61. [5] e. fuguet, c. ràfols, e. bosch, m. rosés, j. chromatogr. a 1216 (2009) 3646-3651. [6] j.m. cabot, e. fuguet, c. ràfols, m. rosés, j. chromatogr. a 1217 (2010) 8340-8345. [7] e. fuguet, c. ràfols, m. rosés, j. chromatogr. a 1218 (2011) 3928-3934. [8] j.m. cabot, e. fuguet, c. ràfols, m. rosés, j. chromatogr. a 1279 (2013) 108-116. [9] e. fuguet, c. ràfols, e. bosch, m. rosés, j. chromatogr. a 1173 (2007) 110-119. [10] e. fuguet, c. ràfols, e. bosch, m. rosés, j. chromatogr. a 1216 (2009) 7798-7805. [11] k. valkó, c.m. du, c.d. bevan, d.p. reynolds, m.h. abraham, j. pharm. sci. 89 (2000) 1085-1096. [12] bio-loom, biobyte corp. (claremont, ca, usa), version 1.5. [13] a. andrés, m. rosés, c. ràfols, e. bosch, s. espinosa, v. segarra, j.m. huerta, manuscript in preparation. [14] a. avdeef, absorption and drug development: solutility, permeability, and charge state, wiley, hoboken, nj, usa, 2003. [15] acd/labs, advanced chemistry development, inc. (toronto, canada), build 2203 (29-jan-2013). [16] c. horváth, w. melander, i. molnár, anal. chem. 49 (1977) 142-154. ©2014 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ spectroscopic examination and release of microencapsulated oregano essential oil doi: 10.5599/admet.5.4.426 224 admet & dmpk 5(4) (2017) 224-233; doi: http://dx.doi.org/10.5599/admet.5.4.426 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper spectroscopic examination and release of microencapsulated oregano essential oil ioannis partheniadis 1 , panagiota karakasidou 1 , souzan vergkizi 2 , ioannis nikolakakis 1 * 1 department of pharmaceutical technology, school of pharmacy, faculty of health sciences, aristotle university of thessaloniki, thessaloniki, 54124, greece 2 department of microbiology, school of medicine, faculty of health sciences, aristotle university of thessaloniki, thessaloniki, 54124, greece *corresponding author: e-mail: yannikos@pharm.auth.gr; tel.: +302310997635; fax: +302310997652 received: august 29, 2017; revised: september 28, 2017; published: december 24, 2017 abstract oregano essential oil (eo) of greek origin with high carvacrol content (86.84 %) was encapsulated by spray drying using arabic gum, modified starch and maltodextrin (75:12.5:12.5) as wall materials. the spray-dried product (eosd) consisted of roundish particles with narrow size distribution. ft-ir and raman spectroscopy identified the eo in eosd, with raman spectra showing more distinct peaks and a small shift of the peak at 1260 cm -1 (assigned to the stretching vibration of the bond of c-o of the phenol), implying only minor chemical interaction with the wall materials. release of the eo from eosd was described by the hixson-crowell equation (r 2 =0.986) with apparent diffusion coefficient 8.3x10 -10 m 2 /s. these findings indicate that microencapsulation by spray drying did not affect the quality of the oregano eo and provided relatively fast and complete release. keywords oregano essential oil; microencapsulation; spray drying; ft-ir; raman spectroscopy; in vitro release introduction essential oils are complex mixtures of natural, aromatic and volatile compounds synthesized by aromatic plants. selected oils have been shown to act on microbial cell surface causing disruption of the cell wall and the cytoplasmic membrane leading to lysis and leakage of intracellular compounds [1]. because of the increasing problem of bacterial resistance to several antibiotics and their accepted safety profile, essential oils may be interesting candidates against microbial infections [2]. oregano essential oil (eo) has been receiving attention because of its antimicrobial activity against both gram negative and gram positive bacteria and to its colicidal and colistatic properties due to the presence of carvacrol and thymol in its composition [3-5]. since oregano eo contains volatile and easily oxidized active ingredients protection from environmental factors is important for retaining its activity in the marketed product. this problem may be tackled by spray drying, which is an industrially established continuous process that in combination with emulsification offers a method for eo microencapsulation [69]. for this purpose, an eo in water emulsion is initially prepared and the eo droplets are stabilized by a http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:yannikos@pharm.auth.gr admet & dmpk 5(4) (2017) 224-233 the release of microencapsulated oregano essential oil doi: 10.5599/admet.5.4.426 225 mixture of gums and carbohydrates [10-11]. the emulsion is then spray-dried under appropriate conditions of temperature and feed rate, and a powder product consisting of single or agglomerated particles is obtained where the eo is encapsulated and thus protected from humidity and light that normally lead to oxidation of alcohols, reduced antimicrobial activity and even inactivation [6, 12-14]. moreover, microencapsulation is expected to increase the solubility of the eo by facilitating wetting of the wall materials and re-emulsification of sprayed dried product (eosd), a process that possibly occurs when in contact with gastric fluids [15]. the aim of this work was to prepare a spray-dried encapsulated product of oregano eo of greek origin that had high carvacrol content, using a composition of arabic gum/starch/maltodextrin [7] as wall materials. in particular, the purpose was to identify the presence of eo in the product and possible interactions of the eo with the wall materials using fourier transform infrared (ftir) and raman spectroscopy and to elucidate the mechanism of in vitro release of the eo in aqueous medium by dialysis dissolution method. the essential oil of oregano was used instead of a methanol extract since it has been reported that the former has stronger and broader spectrum of antimicrobial activity [16]. arabic gum was used as the major of the three encapsulating materials because it forms a low viscosity colloidal dispersion suitable for spray drying, stabilizes eo emulsions in water across a wide ph range as encountered in the gastrointestinal tract, has good retention of volatile compounds and is compatible with other encapsulants [12, 17]. modified starch was used since besides protection from oxidation and loss of volatiles, it also has emulsification ability. maltodextrin with high dextrose equivalent (de 20-23) was added to protect from oxidation and also to facilitate spray drying and improve product wettability [7, 12, 18]. experimental materials oregano essential oil (eo), obtained from oreganum vulgare (heracleoticum) was a gift from ecopharm hellas, kilkis, greece batch 0614 and had in its composition: 86.84 % carvacrol, 2.82 % thymol, 0.96 % γterpinene, 3.46 % p-cymene, 1.82 % β-caryophyllene and 4.11 % other terpenes and phenols (manufacturer’s data). the encapsulating materials: arabic gum (ag, spraygum ab) was purchased from nexira, france. maltodextrin (md, glucidex 21) and modified starch (ms, clearam ch20 20, food grade acetylate di-starch adipate (e1422), waxy maize basis) were from roquette italia, both gifts from interallis chemicals, sindos, greece. preparation of encapsulated oregano eo a 30 % w/w concentration of dispersed phase of sprayed emulsion used for encapsulation was selected after preliminary trials since its viscosity allowed feeding and passing through the nozzle of the spray drier with minimal losses to the walls of the drying chamber at the applied conditions. the emulsion consisted of 3 % w/w oregano eo as the internal dispersed phase, 27 % w/w encapsulating materials (ag 75 %, ms 12.5 %, md 12.5 %) and 70 % deionised water as the external phase [7]. for the preparation of the emulsion, ag and md were hydrated overnight in deionised water at 4-5 °c and then heated to 82 °c. ms was added to this and the mixture homogenized at 15000 rpm for 20 min using an ultra turrax (ika, germany). the mixture was cooled to about 5 °c and oregano eo was added, followed by further homogenization for 5 min. the prepared emulsion was spray-dried using a mini spray dryer (b-191, büchi, switzerland) operated under the following conditions: feed rate 5 ml min -1 , inlet air temperature 180 °c, outlet temperature 117 °c, aspiration rate 100 % and airflow 600 ml/min. the nominal content of oregano eo in the final sprayi. nikolakakis et al. admet & dmpk 5(4) (2017) 224-233 226 dried product was 10 %. three replicate batches were prepared and used for the determination of physicochemical properties and drug release. a spray-dried product without eo was also prepared for comparison purposes. particle size and moisture content the spray-dried product consisted of reasonably well formed particles with a narrow normal size distribution and particle diameters in the range d10=3.9 μm to d90=16.9 with mean value 8.1 μm as previously reported [19]. the moisture content was measured by heating at 105 °c to constant weight and was found to be 3.4±0.2 expressed as % of weight loss on a dry basis (table 1). table 1. technological properties of spray-dried oregano eo product (mean ± standard error, n=3) property value refractive index of unprocessed eo 1.51387±5.2x10 -5 refractive indexof eo after spray drying 1.51566±8.5x10 -5 particle size (μm) d10 4.7 d50 11.0 d90 26.0 moisture content (%) 3.4 ± 0.2 refractive index the refractive index of the unprocessed eo and after its extraction from eosd was measured using a refractometer (bellingham and stanley, kent, england), illuminated with a sodium d1 (yellow) lamp at 589.6 nm and 20 °c. the units read on the scale of the instrument were converted to refractive index from the manufacturer’s conversion table. ft-ir and raman spectroscopy ft-ir spectra were obtained using a shimadzu ft-ir-prestige-21 spectrometer (shimadzu corporation) attached to a horizontal golden gate mkii single-reflection atr system (specac, kent, uk) equipped with a diamond/znse crystal (45° angle to infrared beam, 1.66 μm at 1000 cm -1 depth of penetration, 2.4 refractive index and 525 cm -1 long wavelength cut-off). a few drops of oregano eo or a small amount of the eosd powder product were placed on the diamond disk and 64 scans were collected over the range of 4000–400 cm -1 at resolution 4 cm -1 using appropriate software (shimadzu irsolution 1.3). for the spraydried powder product a sapphire anvil was used to restrain the powder in the path of the beam. raman spectra of samples placed in standard vials were recorded using a bench top raman spectrophotomer (agility, dual band 785/1064 nm model, bayspec, ca, usa) and supporting software (agile 20/20). the laser excitation line was 1064 nm selected due to strong fluorescence of the sample at lower wavelengths, the resolution 12 cm −1 , the exposure time 2 s, the power of incident laser beam 350 mw and the recorded spectra were the average of 100 runs. in vitro release of oregano eo a dialysis method was used to test the release of encapsulated eo from eosd. 100-mg sample corresponding to 10 mg nominal eo but in fact to about 7-8 mg due to losses during spray drying [7], was admet & dmpk 5(4) (2017) 224-233 the release of microencapsulated oregano essential oil doi: 10.5599/admet.5.4.426 227 wavelength / nm placed in a dialysis cellulose tubing (molecular weight cut-off 12500, sigma-aldrich), closed at the two ends and immersed in 200 ml phosphate buffer (pbs, ph 6.8, 37 °c), under magnetic stirring. since the cut-off mw was greater than that previously used for similar systems [20, 21], no interference of the membrane pores was expected in the measured values of eo. for the construction of the standard reference curve, the absorption peaks of the oregano eo were first determined from spectra of oregano eo obtained using a uv-vis spectrophotometer (pharma spec uv-1700 shimandzu, japan). a major peak appeared at wavelength 273 nm (figure 1) and was used for the determination of released eo. in fact, this was very close to the peak obtained from the pure carvacrol (sigma-aldrich) (figure 1) indicating its predominance in the composition of eo (86.84 % according to the manufacturer’s data). the reference curve of eo in pbs was then constructed using standard solutions of unprocessed oregano eo prepared by diluting stock eo solutions in methanol with phosphate buffer (pbs) ph 6.8, to give final concentrations in the range 0.01125 – 0.1125 mg/ml. the standard reference curve was c = (abs 0.0192) / 11.525), where c is in mg ml -1 and was used for the determination of eo in the aliquots taken at timely intervals from the dissolution medium. figure 1. uv-vis spectra of unprocessed oregano eo (black line) and carvacrol (red line) results the values of the refractive indices of the unprocessed and extracted eo are shown on table 1. it can be seen that they were only slightly different indicating that the oregano eo essentially retained its composition after spray drying. particle size, shape and moisture content particle size distribution data of the spray-dried product (eosd) are presented in table 1 as diameters d10, d50 and d90 corresponding to 10 %, 50 % and 90 % of the distribution. it can be seen that the particles formed a narrow and normal size distribution of low span [(d90-d10)/d50]=2.0, with mean diameter 11.0 μm. the narrow particle size distribution of eosd as well as the 3.4 % moisture content (table 1) is within the expected ranges for spray-dried products prepared using equipment of similar a b so rb a n ce i. nikolakakis et al. admet & dmpk 5(4) (2017) 224-233 228 capacity and operating conditions [11]. ft-ir and raman spectroscopy ft-ir spectra of carvacrol, oregano eo and of the spray-dried product without eo or with eo (eosd) are presented in figure 2. carvacrol and oregano eo spectra looked the same even in the fingerprint region 1600-600 cm -1 , confirming that the consistency did not change after spray drying, and the carvacrol content remained very high. several of the characteristic eo peaks present in the carvacrol and the oregano eo spectra (figure 2a-b) also appeared in the eosd spectrum (figure 2c) at the same wavelengths: 2960 cm -1 , 1420 cm -1 and at 1249 cm -1 (indicated by vertical dotted lines), although at much lower intensities because of the relatively low eo content (10 %) in the product. these peaks, however, were absent from the spectrum of the spray-dried product without eo (figure 2d), confirming that these were specific to oregano eo. similarly, from the raman spectra shown in figure 3 it can be seen that the peaks were almost identical for carvacrol and oregano eo reflecting the high carvacrol content in the eo. several characteristic peaks present in the carvacrol and the oregano eo spectra (figure 3a-b) also appeared distinctly in eosd (figure 3c) at the same wavelengths (marked by vertical dotted lines): 1621 cm -1 , 760 cm -1 and 570 cm -1 [22]. an additional peak at 1260 cm -1 in the carvacrol and eo spectra, appears to be shifted to the left at 1285 cm -1 (higher wavelength) in the spectrum of the eosd, implying possible interaction of the eo with the wall materials. the above peaks were absent from the spectrum of the spray-dried product with wall materials alone (without eo). in vitro release of oregano eo the dialysis membrane method was used for testing the release of encapsulated eo from eosd in order to avoid passage of colloidal particles of the high molecular weight (mw) polysaccharides into the dissolution medium and subsequent interference with uv absorption of eo. the release vs time profile of eo from the eo encapsulated spray-dried product (eosd) is shown in figure 4. as it can be seen, plateau is reached after about 60 min corresponding to release of about 7-8 mg, as expected due to loss during spray-drying [7]. the data were analyzed using different kinetic models. eo release should involve diffusion from the interior of the particles to the surface and additionally, relaxation and swelling of the hydrophilic wall materials due to hydration. therefore, the simple power equation of peppas and ritger [23] for swelling-controlled release systems was initially tried but poor fitting of the data was found implying that a more complex release mechanism operates which, besides diffusion and swelling also involves progressive dissolution of the wall material. for this reason the cube-root law [24-25], equation (1) was applied: 1/3 1/3 mo mt kt  (1) mt is the weight of unreleased eo at time t and mo the initial weight of eo in the sample. this is based on the assumptions that a) the dissolving particles are spherical and of narrow size distribution b) sink conditions prevail and c) there is no influence of stagnant layers forming in the neighborhood of the particles. these assumptions were obeyed in this study as demonstrated by the low content of eo in the dissolution fluid (about 7-8 mg in 200 ml) compared to the solubility of carvacrol (1.250 mg ml -1 ) and, by providing sufficient agitation. admet & dmpk 5(4) (2017) 224-233 the release of microencapsulated oregano essential oil doi: 10.5599/admet.5.4.426 229 figure 2. ft-ir spectra of (a) carvacrol, (b) oregano eo, (c) encapsulated eo spray-dried product and (d) spraydried product without eo. vertical dotted lines show presence of eo peaks in the eosd product. figure 3. raman spectra of (a) carvacrol, (b) oregano eo, (c) encapsulated eo spray-dried product and (d) spraydried product without eo. vertical dotted lines show presence of oregano eo peaks in the eosd product. the results were plotted according to equation (1) in figure 5 where it can be seen that the points fall on a straight (r 2 =0.986) with slope line k=0.0213 (mg 1/3 / min) representing the rate constant of the hixsoncrowell equation known as ‘cube-root law’ [24], expressed by equation: 1/3 (4 / 3 ) ( / )     app k n d cs h (2) ρ=0.95 g/cm 3 is the density of oregano eo, cs its solubility in water (taken as that of carvacrol 1.250 (d) (a) (c) (b) 2960 1420 1249 wavenumber, cm -1 t ra n sm it ta n ce raman shift, cm -1 r a m a n i n te n si ty i. nikolakakis et al. admet & dmpk 5(4) (2017) 224-233 230 mg/ml), h the thickness of diffusion layer (taken as the mean particle radius 5.5 μm, table 1, assuming uniform eo distribution in the particles) and n=1.387x10 4 the number of eo releasing particles in the 100 mg sample, calculated from the mean particle diameter and its density (1.34 g/cm 3 , equal to the weighted average of the components). dapp is an ‘apparent’ diffusion coefficient representing the composite release mechanism involving diffusion, swelling and particle dissolution. by substituting the values of the above parameters into equation (2), dapp=8.3x10 -10 m 2 /s. although the estimation of diffusion may be affected due to the passage of eo through the pores of the dialysis membrane, this effect is not expected to be significant due to the low mw of the ingredients of eo (i.e. carvacrol 155.22, p-cymene 139.21, γ-terpinene 136.23 and thymol 150.22). besides, previous estimations of the diffusion coefficients [20, 21] were also obtained using the dialysis method and therefore comparisons of the present with the previous data can be made. figure 4. release-time profile of oregano eo from the spray–dried powder product (error bars show standard deviation, n=3) figure 5. release data of oregano eo from the spray-dried powder product plotted according to the hixsoncrowell ‘cube-root law’ equation. discussion antimicrobial resistance is an emerging health problem which is enlarged by the shortage of new antibiotic agents. at the same time, interest in antimicrobials obtained from natural sources has increased admet & dmpk 5(4) (2017) 224-233 the release of microencapsulated oregano essential oil doi: 10.5599/admet.5.4.426 231 due to their accepted safe status [1-2, 3-4]. essential oils derived from aromatic plants and, oregano eo in particular fall into a group of antimicrobials that has attracted major attention as shown by the numerous reports in the literature [26-27]. however, since pure oregano eo is highly irritating to taste and cannot be easily consumed as such, its microencapsulation into a spray-dried product offers a convenient way for oral administration besides providing protection from environmental factors as well. the peaks that are common in the ft-ir spectra of carvacrol, oregano eo and eosd (figure 2) are ascribed at: 2960 cm -1 to antisymmetrical -ch3 stretching vibration, 1420 cm -1 to antisymmetric -ch3 bending and 1249 cm -1 to c-o-c stretching. since these peaks are absent from the spectrum of the spraydried product without eo they could be potentially used (especially the more distinct at 1420 cm -1 and 1249 cm -1 ), for identification purposes. raman spectroscopy is complementary to ft-ir and was used to add more information about the state of the eo in the product. the peak at 760 cm -1 in the eosd corresponds exactly to the position in the carvacrol spectrum, with no shifting towards lower wavenumbers that might be caused by the presence of thymol confirming carvacrol as the main constituent and that substances present at low percentages do not influence the eo spectrum [25]. the distinct peaks that are common in the raman spectra of carvacrol, unprocessed oregano eo and eo in the eosd (figure 3) are ascribed at: 1621 cm -1 to conjugated c=c stretching vibration, 760 cm -1 to breathing vibration mode and 570 cm -1 to aromatic ring vibration. since they are absent from the spectrum of the spray-dried product without eo, they can be used for identification of eo [22]. the shifting of the peak at 1260 cm -1 , which is assigned to the stretching vibration of the c-o bond of the phenol, to the left at 1285 cm -1 implies interaction of the eo and the wall material, probably between the aromatic –oh of carvacrol and the hydroxyl groups of maltodextrin, resulting in lower vibration frequency and shifting of the peak to higher wavelength. from the above discussion it appears that raman spectroscopy provides more information on the state of eo in the product by revealing more distinct peaks of eo in the product and possible interactions, thus offering a better alternative for identification and the state of eo in the eosd than ftir. contrary to previous works [20, 21] the results in the present study could not be described by fick’s diffusion models due to progressive dissolution of the wall material resulting in change of available particle surface. for this reason the release model cube-root law of hixson-crowell [24] was applied which is derived on the basis of surface change, expressed in equation (1) as mass change with time. the good fitting of the data to this model (r 2 =0.986) indicates distribution of eo in the wet jelly matrix of spray-dried particles during release, so that mass change at different times corresponds to release of proportional amounts of eo. the rather high value of the apparent diffusion coefficient dapp = 8.3x10 -10 m 2 /s in equation (2) compared to the values of about 10 -13 to 10 -16 in previous works [20-21], is attributed to the different wall materials used and different types of diffusion mechanisms. also, in their studies [20], interactions between eo and the wall material were reported whereas only minor interactions were seen in the present work. conclusions microencapsulated oregano eo by spray drying using arabic gum, starch and maltodextrin as wall materials has been previously reported by da costa et. al [7] to offer good retention and encapsulation efficiency. it was therefore considered worthwhile investigating the state of microencapsulated eo and its release profile from the spray-dried product (eosd). ft-ir and raman spectroscopy clearly identified the presence of eo in the eosd, with a small shift of the peak at 1260 cm -1 in the raman spectrum to 1285 cm -1 , i. nikolakakis et al. admet & dmpk 5(4) (2017) 224-233 232 implying only minor chemical interaction with the wall materials. release of the eo from eosd was completed within 1 – 1.5 h and was well described by the hixson-crowell equation (r 2 =0.986) with apparent diffusion coefficient 8.3x10 -10 m 2 /sec. the above findings encourage formulation of oregano eo into spray-dried product for oral administration for therapeutic purposes. references [1] s. burt, essential oils: their antibacterial properties and potential applications in foods-a review, international journal of food microbiology 94 (2004) 223–253. [2] j. c. lopez-romero, h. gonzález-ríos, a. borges, m. simões, antibacterial effects and mode of action of selected essential oils components against escherichia coli and staphylococcus aureus, evidence-based complementary and alternative medicine 2015 (2015) 1-9. [3] r. j. lambert, p. n. skandamis, p. j. coote, g. j. nychas, a study of the minimum inhibitory concentration and mode of action of oregano essential oil, thymol and carvacrol, journal of applied microbiology 91 (2001) 453-462. [4] a. béjaoui, h. chaabane, m. a. jemli, a. boulila, m. boussaid, essential oil composition and antibacterial activity of origanum vulgare subsp. glandulosum desf. at different phenological stages, journal of medicinal food 16 (2013) 1115-1120. [5] s. a. burt, r. d. reinders, antibacterial activity of selected essential oils against escherichia coli o157:h7, letters in applied microbiology 36 (2003) 162-167. [6] g. a. reineccius, the spray drying of food flavors, drying technology 22 (2004) 1289–1324. [7] j. m. g. da costa, s.v. borges, a. a. hijo, e. k. silva, g. r. marques, m. â. cirillo, v. m. de azevedo, matrix structure selection in the microparticles of essential oil oregano produced by spray dryer, journal of microencapsulation 30 (2013) 717–727. [8] a. m. bakry, a. shabbar, a. barkat, h. majeed, m. y. abouelwafa, a. mousa, l. liang, microencapsulation of oils: a comprehensive review of benefits, techniques, and applications, comprehensive reviews in food science and food safety 15 (2016) 143-182. [9] j. baranauskaite, l. ivanauskas, r. masteikova, d. kopustinskiene, a. baranauskas, j. bernatoniene, formulation and characterization of turkish oregano microcapsules prepared by spray-drying, pharmaceutical development and technology 22 (2017) 792-803. [10] m. i. ré. microencapsulation by spray drying, drying technology 16 (1998) 11195-1236. [11] d. a. botrel, v. s. borges, r. v. de barros fernandes, a. d. viana, j. m. g. da costa, g. r. marques, evaluation of spray drying conditions on properties of microencapsulated oregano essential oil, international journal of food science & technology 47 (2012) 2289–2296. [12] s. m. jafari, e. assadpoor, y. he, b. bhandari, encapsulation efficiency of food flavours and oils during spray drying, drying technology 26 (2008) 816–835. [13] l. p. fernandes, i. c. turatti, n. p. lopes, j. c. ferreira, r. c. candido, w. p. oliveira, volatile retention and antifungal properties of spray-dried microparticles of lippia sidoides essential oil, drying technology 26 (2008) 1534–1542. [14] r. v. de barros fernandes, s. v. borges, d. a. botrel, e. k. silva , j. m. g. da costa, f. queiroz, microencapsulation of rosemary essential oil: characterization of particles, drying technology 31 (2013) 1245–1254. [15] r. deshmukh, p. wagh, j. naik, solvent evaporation and spray drying technique for microand nanospheres/particles preparation, drying technology 34 (2016) 1758–1772. [16] f. şahin, m. güllüce, d. daferera, α. sökmen, m. sökmen, m. polissiou, g. agar, h. özer, biological activities of the essential oils and methanol extract of origanum vulgare ssp. vulgare in the eastern anatolia region of turkey, food control 15 (2004) 549–557. [17] a. imeson, food stabilisers, thickeners and gelling agents, wiley-blackwell publishing ltd, new jersey, usa, 2009, p. 368. admet & dmpk 5(4) (2017) 224-233 the release of microencapsulated oregano essential oil doi: 10.5599/admet.5.4.426 233 [18] b. f. mcnamee, l. e. white, e. d. o'riordan, m. o'sullivan, effect of partial replacement of gum arabic with carbohydrates on its microencapsulation properties, journal of agricultural and food chemistry 7 (2001) 3385-3388. [19] s. vergkizi-nikolakaki, p. karakasidou, i. nikolakakis. evaluation of antimicrobial activity of oregano oil by agar disk diffusion method. 10th world meeting on pharmaceutics, biopharmaceutics and pharmaceutical technology glasgow, scotland 4-7 april 2016. [20] s. beirão da costa, c. duarte, a.i. bourbon, a. c. pinheiro, m. i. n. januário, a. a. vicente, m. l. beirão da costa, i. delgadillo, inulin potential for encapsulation and controlled delivery of oregano essential oil, food hydrocolloids 33 (2013) 199-206. [21] s. beirão da costa, c. duarte, a. i. bourbon, a. c. pinheiro, a. t. serra, m. m. martins, m. i. n. januário, a. a. vicente, i. delgadillo, c. duarte, m. l. beirão da costa, effect of the matrix system in the delivery and in vitro bioactivity of microencapsulated oregano essential oil, journal of food engineering 110 (2012) 190–199. [22] n. g. siatis, a. c. kimbaris, c. s. pappas, p. a. tarantilis, d. j. daferera, m. g. polissiou, rapid method for simultaneous quantitative determination of four major essential oil components from oregano (oreganum sp.) and thyme (thymus sp.) using ft-raman spectroscopy, journal of agricultural and food chemistry 53 (2005) 202-206. [23] a. n. peppas, j. j. sahlin, a simple equation for the description of solute release. iii. coupling of diffusion and relaxation, international journal of pharmaceutics 57 (1989) 169-172. [24] a. w. hixson, j. h. crowell, dependence of reaction velocity upon surface and agitation i theoretical consideration, industrial & engineering chemistry 23 (1931) 923–931. [25] d. j. daferera, p. tarantilis, m.g. polissiou. characterization of essential oils from lamiaceae species by fourier transform raman spectroscopy, journal of agricultural and food chemistry 50 (2002) 5503-5507. [26] m. m. cowan, plant products as antimicrobial agents, clinical microbiology reviews 12 (1999) 564582. [27] j. s. franklyne, a. mukherjee, n. chandrasekaran, essential oil microand nanoemulsions: promising roles in antimicrobial therapy targeting human pathogens, letters in applied microbiology 63 (2016) 322-334.. ©2017 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ feasibility of the estimation of octanol-water distribution coefficients of acidic drugs by microemulsion electrokinetic chromatography doi: 10.5599/admet.6.1.510 55 admet & dmpk 6(1) (2018) 55-60; doi: http://dx.doi.org/10.5599/admet.6.1.510 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index short communication feasibility of the estimation of octanol-water distribution coefficients of acidic drugs by microemulsion electrokinetic chromatography alejandro fernández-pumarega 1 , susana amézqueta 1, *, elisabet fuguet 1,2 , martí rosés 1 1 departament de química analítica and institute of biomedicine of the university of barcelona (ibub), universitat de barcelona, martí i franquès 1-11, 08028, barcelona, spain 2 serra húnter programme, generalitat de catalunya *corresponding author: e-mail: samezqueta@ub.edu; tel.:+34934021277 received: february 19, 2018; revised: march 18, 2018; published: march 25, 2018 abstract previous studies have shown that a microemulsion electrokinetic chromatography (meekc) system can estimate the logarithm of the octanol-water partition coefficient (log po/w) of neutral solutes. in the present work, the applicability of the method to partially and fully ionized acids has been evaluated. naproxen, a monoprotic acid, has been used as test solute. the retention factor (k) of this compound has been measured in meekc at several values of ph and the retention factor-ph profile has been established. as log po/w correlates with log kmeekc for neutral compounds, this correlation has been used to estimate the logarithm of the octanol-water partition coefficient of the neutral (log po/w(ha)), and the fully ionized (log p o/w(a-)) forms of naproxen. then, the logarithm of the octanol-water distribution coefficient (log do/w) of the partially ionized form of the acid has been estimated. the comparison of the estimated values with the ones obtained experimentally using the classical procedures, such as the shake-flask method, shows differences under 0.4 log do/w units either if the acid is partially ionized or in its neutral form in the most part of the ph range. however, the method overestimates the log do/w of the highly (>99.5 %) or fully ionized form of naproxen. keywords capillary electrophoresis; meekc; log do/w; lipophilicity introduction lipophilicity, which is defined as the ability of a compound to be dissolved in lipids or non-polar solvents, is a key factor in drug discovery. this property, among others, is related to the capacity of a substance to pass through different biological membranes, which are formed mainly by a lipid bilayer. therefore, the viability of a compound as a drug candidate is clearly linked to its lipophilicity. in order to estimate the lipophilicity of a substance, different parameters can be evaluated. among these, the most widely used is the octanol-water partition coefficient (po/w). however, the direct evaluation of this parameter (following the shake-flask method) is time-consuming and it is not fully automated. because of these drawbacks, the development of faster high-throughput methods is of a great interest. http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:samezqueta@ub.edu susana amézqueta et al. admet & dmpk 6(1) (2018) 55-60 56 ishihama and coworkers [1] introduced a method able to evaluate the lipophilicity of a substance through a system based on microemulsion electrokinetic chromatography (meekc). the meekc system used is composed of two main phases, an aqueous phase (buffer) and a pseudostationary phase (the microemulsion). further studies [2,3] followed the method and correlated satisfactorily the logarithm of the octanolwater partition coefficient (log po/w) to the logarithm of the retention factor in meekc (log kmeekc). abraham et al. [2] correlated log po/w to log kmeekc measured in a system with a microemulsion composed of 1.44 % (w/w) sds, 6.49 % (w/w) 1-butanol, and 0.82 % (w/w) heptane, obtaining the following equation: log po/w= 1.542 + 1.276·log kmeekc , r 2 =0.99; sd=0.096; n=53 . (1) subirats et al. [3] performed the same procedure using a microemulsion consisting of 1.3 % (w/v) sds, 8.15 % (v/v) 1-butanol, 1.15 % (v/v) heptane, and 5 % (v/v) acetonitrile (at ph 7.4, in a 10 mm phosphate buffer). acetonitrile was added to overcome the co-elution of the microemulsion marker with the most lipophilic compounds. the resulting equation of the correlation and the statistics are: log po/w= 1.48 (±0.05) + 1.48 (±0.05)·log kmeekc , r 2 =0.96; n=32 . (2) taking into account equations 1 and 2 it can be concluded that this method is able to predict the octanol-water partition coefficient of new compounds in their neutral form through electrokinetic measurements. considering that most part of pharmaceutical drugs are acids or bases, a fast determination of this parameter for ionic compounds is of great interest. therefore, the aim of this work is to test the applicability of meekc measurements to estimate the logarithm of the octanol-water distribution coefficient (log do/w) of acidic solutes at several degrees of ionization. naproxen has been employed as test solute since it is an acid that presents a pka close to 4 (within the electrophoretic working ph range), its lipophilicity-ph profile is well-defined in the literature [4,5], and it contains chromophore groups (therefore detectable by uv-vis). experimental a capillary electrophoresis system (ce 7100) equipped with a diode array from agilent technologies (santa clara, ca, usa) was employed to do the electrophoretic measurements. the fused-silica capillary used was from polymicro technologies (lisle, il, usa) with an effective length of 25 or 30 cm (depending on the ph value) and an internal diameter of 50 μm. the microemulsion used as the pseudostationary phase was prepared in a phosphate or acetate buffer at 0.05 m ionic strength. first, the surfactant (sds) was dissolved in the buffer at 1.30 % (w/v), then the co-surfactant (1-butanol) was added at 8.15 % (v/v), followed by the addition of the oil (heptane) at 1.15 % (v/v). the solution was stirred and sonicated until the turbidity was no longer observed [3]. the analyses were performed using two different solutions (in the microemulsion and in a buffer without the pseudostationary phase) at 25 °c, applying a voltage ranging from 8 to 15 kv, and an internal pressure ranging from 0 to 50 mbar (depending on the buffer and the ph of each experiment). the solutes were injected with a pressure of 50 mbar during 5 s. dodecanophenone was employed as the microemulsion marker and dmso as the electroosmotic flow marker. the solute to be injected was disolved in microemulsion/methanol mixture at 9:1 ratio, in meekc measurements, and in water/methanol mixture at 9:1 ratio, in plain buffer measurements, with a final concentration of approximately 200 mg/l. admet & dmpk 6(1) (2018) 55-60 estimation of octanol-water log do/w of acidic drugs by meekc doi: 10.5599/admet.6.1.510 57 theory the migration behavior of an acidic compound when it is partially or totally ionized is a contribution of two main factors [6]: the hydrophobic interaction of the compound with the pseudostationary phase (the microemulsion), and the electrophoretic mobility of the anionic acid by its electrostatic attraction to the anode (with positive charge). retention factor of ionized compounds can be calculated through the next equation: 𝑘 = 𝜇−𝜇0 𝜇me−𝜇 , (3) where, k is the retention factor of the solute, µ the overall observed mobility of the compound, µme the mobility of the microemulsion phase (obtained by the mobility of the microemulsion marker), and µ0 the mobility of the compound in the aqueous buffer without the microemulsion, i.e., in capillary zone electrophoresis (cze). in parallel studies of the research group, it has been observed that the viscosity of the buffer can affect the electrophoretic mobility of the compound. so, in order to determine the retention factor of ionizable acids, as solutions with different viscosities have been used, a viscosity correction factor has to be introduced [7]. results and discussion in order to estimate the log do/w values of naproxen at several degrees of ionization through electrophoretic measurements a calibration curve that correlates log po/w vs. log kmeekc has been performed. 20 neutral compounds with known log po/w values [8,9] have been used to establish the curve. the substances that have been chosen present log po/w values uniformly distributed between approximately 0 and 4 units. the equation of the calibration curve obtained and their statistics are as follows: log po/w = 1.60 (±0.11)·log kmeekc + 1.51 (±0.08) , r 2 =0.92; sd=0.33; n=20 , (4) as expected, equation 4 shows a good correlation between log po/w and log kmeekc for neutral compounds, and it is very similar to equations 1 and 2. as the retention factor of acidic compounds is the weighted average of the retention factor of the neutral (k(ha)) and the fully ionized (k(a-)) forms of the compound: 𝑘 = (1 − α) · 𝑘(𝐻𝐴) + α · 𝑘(𝐴−) , (5) and the ionization degree (α) can be calculated using the following formula: α = 10ph−p𝐾a 1+10ph−p𝐾a , (6) an equation that relates retention factor to ph can be easily obtained [6]: 𝑘 = 𝑘(𝐻𝐴)+𝑘(𝐴−)·10 ph−p𝐾a 1+10ph−p𝐾a . (7) to study the applicability of the method to ionizable acidic compounds, the retention factor of naproxen has been measured in the meekc system at 6 different ph values, ranging from 2 to 7 (covering all of the degrees of ionization of the acid). then equation 7 has been fitted to the measured retention factors. figure 1 shows the experimental values (x) and the retention-ph profile obtained by fitting equation 7 to the experimental data (solid line). susana amézqueta et al. admet & dmpk 6(1) (2018) 55-60 58 figure 1. retention factor-ph profile of naproxen. the parameters obtained from the retention profile in figure 1, with their respective statistics are: k(ha) = 8.24 (±0.13); k(a-) = 0.40 (±0.11); pka = 4.38 (±0.04); (r 2 = 0.999; f = 1223; sd = 0.16). using equation 4, and k(ha) and k(a-) obtained from the fit of equation 7 (figure 1), the logarithm of the octanol-water partition coefficient of the neutral and fully ionized forms of the acid can be calculated: log po/w(ha)=2.97 and log po/w(a-)=0.87, respectively. similarly to the retention factor, the octanol-water distribution coefficient (do/w) is also related to the ionization degree of the compound through the following equation: 𝐷o/w = (1 − α) · 𝑃o/w(ha) + α · 𝑃o/w(a−) . (8) so, using equation 8 and considering the pka value of the compound (pka=4.38, obtained from figure 1 data) it is possible to estimate the log do/w of naproxen at any degree of ionization. table 1 presents the experimental log do/w values reported in the literature determined by the shake-flask [4] or the ph-metric methods [5], and the log do/w estimated values which have been calculated using equation 8. table 1. differences between literature and estimated log do/w values of naproxen at several ph values. ph log do/w (literature) α (1-α)·do/w(ha) α·do/w(a-) log do/w (estimated) log dlit.-log dest. 1.00 a 3.35 0.000 942.36 0.00 2.97 0.38 2.00 a 3.34 0.004 938.84 0.03 2.97 0.37 3.35 a 3.05 0.085 862.28 0.63 2.94 0.11 4.50 a 2.68 0.569 406.67 4.22 2.61 0.07 5.50 a 2.07 0.929 66.47 6.90 1.87 0.20 5.5 b 2.1 0.929 66.47 6.90 1.9 0.2 6.5 b 1.1 0.992 7.10 7.37 1.2 -0.1 6.70 a 0.79 0.995 4.49 7.39 1.07 -0.28 7.40 a 0.33 0.999 0.90 7.42 0.92 -0.59 7.4 b 0.3 0.999 0.90 7.42 0.9 -0.6 8.20 a 0.27 1.000 0.14 7.43 0.88 -0.61 9.20 a 0.25 1.000 0.01 7.43 0.87 -0.62 a) data obtained from ref. [4]. b) data obtained from ref. [5]. 0.0 2.0 4.0 6.0 8.0 10.0 1.0 3.0 5.0 7.0 9.0 k ph admet & dmpk 6(1) (2018) 55-60 estimation of octanol-water log do/w of acidic drugs by meekc doi: 10.5599/admet.6.1.510 59 the differences between estimated (obtained by meekc measurements) and literature log do/w data are in most cases under 0.4. higher differences have been observed only at those values of ph where the acid is highly or fully ionized (>99.5 %). this higher error could be explained due to the possible formation of different ion-pairs, depending on the experimental conditions, or due to the low retention of fully ionized acid species in the meekc system (meaning higher experimental errors as the magnitude of k is smaller than that of the neutral species). moreover, these differences could also be attributed to a surface between the aqueous and the lipid phase bigger in the microemulsion than in the octanol-water system. in this way, the retention of the compounds in meekc will be larger even if they are charged as they will be attached to the hydrophobic phase even if they do not partition into it. further research is needed to confirm these differences, clarify them, and propose additional correction for the ionized forms of the acidic drugs. the direct applicability of the method to ionizable basic compounds is doubtful because of the possible ion-pair formation between the protonated base and the anionic surfactant. work is in progress to clarify this point. conclusions accurate predictions of log do/w have been obtained for naproxen at different degrees of ionization using the meekc system presented in this work. at those ph values where the compound is in its neutral form or only partially ionized, the error obtained is in the order of the error obtained for the estimation of the log po/w of neutral compounds and the one of experimental determinations. only at very high ionization degrees (>99.5 %) an overestimation of the log do/w is observed. this work shows that the method based in meekc is an alternative to the classical shake-flask procedure to determine log do/w values for ionized species of this acid. a deeper study is necessary to check if this behavior can be extrapolated to other compounds. acknowledgements: financial support from the ministerio de economía y competitividad from the spanish government (ctq2014-56253-p) and the catalan government (2017 sgr 1074) is acknowledged. afp wishes to thank the university of barcelona for his phd fellowship. references [1] y. ishihama, y. oda, k. uchikawa, n. asakawa. evaluation of solute hydrophobicity by microemulsion electrokinetic chromatography. analytical chemistry 67 (1995) 1588-1595. [2] m.h. abraham, c. treiner, m. roses, c. rafols, y. ishihama. linear free energy relationship analysis of microemulsion electrokinetic chromatographic determination of lipophilicity. journal of chromatography a 752 (1996) 243-249. [3] x. subirats, h.p. yuan, v. chaves, n. marzal, m. rosés. microemulsion electrokinetic chromatography as a suitable tool for lipophilicity determination of acidic, neutral, and basic compounds. electrophoresis 37 (2016) 2010-2016. [4] f. barbato, g. caliendo, m.i. larotonda, c. silipo, g. toraldo, a. vittoria. distribution coefficients by curve fitting application to ionogenic nonsteroidal antiinflamatory drugs. quantitative structureactivity relationships 5 (1986) 88-95. [5] s. winiwarter, n.m. bonham, f. ax, a. hallberg, h. lennernäs, a. karlén. correlation of human jejunal permeability (in vivo) of drugs with experimentally and theoretically derived parameters. a multivariate data analysis approach. journal of medicinal chemistry 41 (1998) 4939-4949. [6] m.g. khaledi, s.c. smith, j.k. strasters. micellar electrokinetic capillary chromatography of acidic solutes migration behavior and optimization strategies. analytical chemistry 63 (1991) 1820-1830. [7] a. fernández-pumarega, s. amézqueta, e. fuguet, m. rosés, manuscript in preparation (2018). susana amézqueta et al. admet & dmpk 6(1) (2018) 55-60 60 [8] a. andrés, m. rosés, c. ràfols, e. bosch, s. espinosa, v. segarra, j.m. huerta. setup and validation of shake-flask procedures for the determination of partition coefficients (logd) from low drug amounts. european journal of pharmaceutical sciences 76 (2015) 181-191. [9] a. avdeef. absorption and drug development: solubility, permeability, and charge state, second edition, john wiley & sons, inc., hoboken, nj, usa, 2012, p. 201. ©2018 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license(http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ special issue devoted to the 6th iapc meeting: joint events comprising 6th world conference on physico-chemical methods in drug discovery and development and 3rd world conference on admet and dmpk doi: 10.5599/admet.6.1.522 1 admet & dmpk 6(1) (2018) 1-3; doi: http://dx.doi.org/10.5599/admet.6.1.522 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index editorial special issue devoted to the 6 th iapc meeting: joint events comprising 6 th world conference on physico-chemical methods in drug discovery and development and 3 rd world conference on admet and dmpk kin tam 1 and zoran mandić 2 1 faculty of health sciences, university of macau, macau, china e-mail: kintam@umac.mo; tel.: +853-8822-4988; fax: +853-8822-2314 2 university of zagreb, faculty of chemical engineering and technology, hr-10000, zagreb, croatia e-mail: zmandic@fkit.hr; tel: +385 1 4597164 the present issue of admet and dmpk is dedicated to the 6 th iapc meeting which was organized as a joint event consisting of 6 th world conference on physico-chemical methods in drug discovery and development (pcmddd-6) and 3 rd world conference on admet and dmpk (admet-3). the meeting took place in hotel westin, zagreb, croatia, september 04-07, 2017. iapc meetings are organized as annual events in alternating european and east asia locations. the topics covered the broad range of methods used in successful drug candidate identification and development. determination of adme/tox properties through the in vitro and in vivo assays was discussed. the particular attention was paid to the evaluation and improvement of critical drug parameters which determine the fate of the drug, from its administration over its remedial action to its excretion. roughly, six sessions were organized: drug development, solid state, pharmaceutical cocrystals, physchem and permeability, admet and dmpk and drug discovery. special evening session “pharmaceutical cocrystals – physicochemical properties and formulations” was organized and moderated by alex avdeef. the session highlighted issues related to the measurement of physicochemical properties, particularly dissolution and solubility as a function of ph, in support of formulation development of oral drug products with improved bioavailability. almost 150 delegates actively participated at the iapc-6 meeting. most of them presented their work either orally or through poster communication resulting in a diverse but well-balanced four-day programme. a small selection of papers giving a typical cross section of the conference workings was published in the present issue as well as in two previous issues (vol. 5, nos. 3 & 4) and will also be published in the next issue (vol. 6, no. 2). mario jug et al [1] contributed with the review about in vitro dissolution methods for the mucosal delivery systems, while adriana isvoran et al. [2] used a few computational tools for predicting absorption, distribution, metabolism, excretion and toxicity, pharmacokinetics profiles, toxic/adverse effects, carcinogenicity, cardiotoxicity and endocrine disruption of some of low molecular weight water soluble derivatives of chitosan that are used in wound healing. ioannis nikolakakis et al [3] investigated the release of microencapsulated oregano essential oil by spectroscopy. solubility-ph profiles of sibutramine and was http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:kintam@umac.mo mailto:zmandic@fkit.hr admet & dmpk 6(1) (2018) 1-3 iapc-6 meeting editorial doi: 10.5599/admet.6.1.522 2 published as an extended abstract [4]. the present issue consists of a review paper about the role of pharmacology in anticancer drug development two original scientific articles and three short communications. peters et al. [5] evaluated in front of the european organization for research and treatment of cancer (eortc) the process of selection of european potential anti-cancer compounds to be tested in national cancer institute on the existing cell lines representative of nine major tumour types. two very interesting original articles deal with the pharma industry view on in silico adme in drug design and the use of the low-field nmr for the characterization of gels and biological issues. in particular, winiwarter et al [6] concentrated on the pharmacokinetic properties of the compounds and suggested using in silico adme models to predict what potency would be required for a specific compound to enable coverage during the whole dosing interval. two short communications come from the clara rafols and her group, one deals with the biomolecular interactions using molecular fluorescence and another compares the dissolution rates of ciprofloxacin and its cocrystal with resorcinol. one short communication deals with the potential utilization of microemulsion electrokinetic chromatography for the determination of octanol/water distribution coefficients. from the four outstanding papers which are going to be published in the next issue of the journal two were presented in the special cocrystal session at the conference. one is the review paper “pharmaceutical cocrystals – physicochemical properties and formulations” in which all speakers from the special session contributed. another cocrystal paper comes from abu serajuddin group. we wish to thank all the authors of this and the following special issue for their high-quality papers. we would also like to thank the referees who critically evaluated the papers at short notice. finally, we hope readers will enjoy the articles and find them interesting, useful and beneficial for their work. figure 1. the opening ceremony of iapc-6 meeting, zagreb, croatia. references [1] m. jug, a. hafner, j. lovrić, m. lusina kregar, i. pepić, ž. vanić, b. cetina-čižmek, j. filipović-grčić. in vitro dissolution/release methods for mucosal delivery systems. admet and dmpk 5 (2017) 173-182. admet & dmpk 6(1) (2018) 1-3 iapc-6 meeting editorial doi: 10.5599/admet.6.1.522 3 [2] a. isvoran, a.a. ciorsac, v. ostafe. adme-tox profiling of some low molecular weight water soluble chitosan derivatives. admet and dmpk 5 (2017) 192-200. [3] i. partheniadis, p. karakasidou, s. vergkizi, i. nikolakakis. spectroscopic examination and release of microencapsulated oregano essential oil. admet and dmpk 5 (2017) 224-233. [4] d. lucero-borja, ò. castilla, r. barbas, x. subirats, c. ràfols. solubility-ph profiles of a free base and its salt: sibutramine as a case study. admet and dmpk 5 (2017) 253-256. [5] g.j. peters, a.s. govaerts, h.r. hendriks, for the eortc-pharmacology and molecular mechanism group. the role of pharmacology in anticancer drug development. admet and dmpk 6 (2018) 4-14. [6] s. winiwarter, e. ahlberg, e. watson, i. oprisiu, m. mogemark, t. noeske, n. greene. in silico adme in drug design – enhancing the impact. admet and dmpk 6 (2018) 15-33. manuscript doi: 10.5599/admet.2.2.45 107 admet & dmpk 2(2) (2014) 107-114; doi: 10.5599/admet.2.2.45 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper evaluation of log po/w values of drugs from some molecular structure calculation software juan m. pallicer 1 , martí rosés 1 , clara ràfols 1 , elisabeth bosch 1* , rosalia pascual 2 , adriana port 2 1 departament de química analítica and institut de biomedicina (ibub), universitat de barcelona, martí i franquès 111, 08028 barcelona. 2 esteve, baldiri i reixac, 4-8, 08028 barcelona. *corresponding author: elisabeth bosch, e-mail: e.bosch@ub.edu; tel.: +34934021284; fax: +34934021233 received: july 04, 2014; revised: july 11, 2014; published: july 18, 2014 abstract predictive software packages to estimate the lipophilicity of molecules have become key tools in the new drug design. six different well-known computational programs including the classical biobyte-clogp and the galas algorithm offered by acdlabs were evaluated through a set of 103 drugs with different structures and functionalities. to evaluate the predictions accuracy, reliable experimental log po/w values for the whole testing set were carefully selected. the best estimations are performed by galas/logp based on the fragmental method, corrected according to the similarity with compounds included in the software training set. keywords: lipophilicity of drugs, lipophilicity prediction software introduction lipophilicity, expressed by the logarithm of octanol-water partition coefficient log p, or distribution log d if ionised molecular species are present, constitutes a physicochemical property of paramount importance in medicinal chemistry and in overall drug discovery. optimal lipophilicity in compounds should be targeted in the early phases of drug research since it contributes on individual admet (absorption, distribution, metabolism, elimination and toxicology) [1,2], including blood-brain barrier penetration and clearance [3]. predictive software to estimate lipophilicity has also evolved as a master piece in the structure design for new chemical entities [4]. different methods for log p prediction have been developed and they can be divided into two main categories, substructure-based and property-based methods. substructure-based methods work decomposing the 2d structure of the compound into fragments (fragmental approaches) [5] or into single atoms (atom-based approaches) [6, 7]. the resulting log p value is obtained as a summation of terms, being the difference among the different fragmental methods, the set of fragments to identify in the target molecule, the contribution constant of each of those identified fragments, and the different correction factors to apply depending on the fragment’s environment. in the http://www.pub.iapchem.org/ojs/index.php/admet/index pallicer et al. admet & dmpk 2(2) (2014) 107-114 108 case of the atom-based methods there are commonly no correction factors in the summation and log p values are obtained by adding the contribution of the different atom types present in the compound, where different methods have different atom type definition and contribution values. in all cases the different fragment references, atom type definitions and values are the result of different training sets and analysis techniques used for the development of each method. property-based approaches [8] estimate log p values using calculated descriptors that account for the entire molecule. many of those techniques need the 3d structure of the compounds to calculate the required descriptors, and variability of results due to conformational uncertainty has to be taken into account. some methods imply as well quantum mechanical semi-empirical calculations and others molecular dynamics, which makes them resource demanding and slow. to overcome those difficulties some linear models based on 2d-descriptors have been developed as well, however substructure-based methods continue being the most widespread in the pharmaceutical world thanks to their speed and easy implementation in medchem desktop tools and molecular modelling programs. in this work, a set of 103 drugs which has been selected for representing a broad collection of pharmaceutical compounds with different structures and functionalities, and a range of log po/w values from 0 to 7, has been used in order to evaluate the accuracy of the log po/w values obtained through both classical and well accepted programs such as the biobyte-clogp and more recently developed ones as the galas algorithm by acd. calculation methods a list of 103 pharmaceutical compounds with a wide variety of functionalities was selected to predict log po/w values through six different substructure-based methods: alogp, which is an atom-based method published by ghose and crippen [9] and implemented in pipelinepilot (accelrys) [10], where carbon, hydrogen, oxygen and nitrogen are classified into 120 atom types; clogp, a fragmental method with correction factors that take structural and interaction factors into account, developed at the pomona college medchem project [11], licensed from biobyte corp and incorporated in chemfinder; chemprop/logp developed by cambridge soft and implemented in chemfinder [12], it uses 3 fragmentation methods that can handle molecules containing different atoms; classic acd/logp [13], which is based on contributions of separate atoms, structural fragments and intramolecular interactions that have been derived form acd/labs internal database that comprises information from reference books, articles and public sources; galas/logp [14], recently developed by acd where its name stands for global, adjusted locally according to similarity, the global method is as well fragmental and the baseline result is adjusted depending on the performance obtained for similar compounds from the training set; and consensus/logp [14] which is a consensus result on classic and galas algorithms provided as well by acd, where dynamic adaptive coefficients are assigned to each model according to the corresponding indications of prediction quality. results and discussion the log po/w values obtained for the 103 compounds are gathered in table 1. the reference log po/w are experimental values previously determined [15], complemented by the values recommended in bioloom online database (bioloom) and the “gold standard” list selected by avdeef [16]. in order to discover the admet & dmpk 2(2) (2014) 107-114 evaluation of log p values doi: 10.5599/admet.2.2.45 109 accuracy in used softwares, deviation obtained through experimental reference and predicted values are listed. colored values represent outliers where deviations are higher of 0.6 units. this limit value was selected because of the experimental variability in log po/w measurements that led the aoac to admit differences of 0.3 units between log po/w values measured from replicates using the shaking flask reference method as well as the variety of values shown for most compounds in the leo et al [17] and bioloom databases [18]. table 1. predicted and experimental (reference) log p values. predicted log po/w name log po/w ref classic (acd) deviation galas (acd) deviation consensus (acd) deviation chemprop (camdridge soft) deviation clogp (biobyte) deviation alogp (accelrys) deviation vildagliptin -0.16 [15] -0.14 -0.02 1.39 -1.55 0.79 -0.95 -0.32 0.16 0.69 -0.85 0.18 -0.34 theophylline -0.02 [18] -0.17 0.15 0.25 -0.27 0.13 -0.15 -1.03 1.01 -0.57 0.55 -0.31 0.29 pilocarpine 0 [18] -0.10 0.10 0.62 -0.62 0.39 -0.39 0.54 -0.54 -0.20 0.20 0.97 -0.97 atenolol 0.06 [15] 0.10 -0.04 0.29 -0.23 0.24 -0.18 0.50 -0.44 -0.11 0.17 0.67 -0.61 amiloride 0.1 [18] 2.89 -2.79 -1.02 1.12 1.03 -0.93 -0.71 0.81 -2.22 2.32 -0.75 0.85 levetiracetam 0.14 [18] -0.67 0.81 -0.76 0.90 -0.74 0.88 -0.75 0.89 -0.34 0.48 -0.30 0.44 pyricarbate 0.24 [18] -0.25 0.49 0.51 -0.27 0.34 -0.10 0.67 -0.43 -0.42 0.66 0.65 -0.41 sotalol 0.24 [18] 0.32 -0.08 0.28 -0.04 0.29 -0.05 -0.06 0.30 0.23 0.01 0.49 -0.25 ranitidine 0.26 [15] 1.23 -0.97 0.22 0.04 0.37 -0.11 0.00 0.26 0.67 -0.41 1.04 -0.78 4-aminophenazone 0.33 [18] 0.76 -0.43 1.18 -0.85 1.10 -0.77 0.36 -0.03 0.57 -0.24 0.99 -0.66 pyridoxine 0.33 [18] -1.10 1.43 -0.39 0.72 -0.62 0.95 -0.49 0.82 -0.35 0.68 -0.49 0.82 paracetamol 0.51[18] 0.34 0.17 0.43 0.08 0.40 0.11 0.89 -0.38 0.49 0.02 0.71 -0.20 barbital 0.65 [18] 0.64 0.01 0.79 -0.14 0.76 -0.11 0.75 -0.10 0.66 -0.01 0.74 -0.09 nadolol 0.71 [18] 1.29 -0.58 1.22 -0.51 1.24 -0.53 1.00 -0.29 0.38 0.33 1.15 -0.44 sulfamethoxazole 0.89 [18] 0.89 0.00 0.58 0.31 0.65 0.24 0.86 0.03 0.56 0.33 0.71 0.18 ephedrine 0.93 [18] 1.05 -0.12 1.33 -0.40 1.26 -0.33 1.38 -0.45 0.89 0.04 1.23 -0.30 procainamide 1.09 [18] 1.10 -0.01 1.30 -0.21 1.25 -0.16 1.05 0.04 1.42 -0.33 1.13 -0.04 codeine 1.14 [18] 1.20 -0.06 1.22 -0.08 1.21 -0.07 1.45 -0.31 0.98 0.16 1.64 -0.50 triamterene 1.22 [18] 1.34 -0.12 1.16 0.06 1.20 0.02 2.11 -0.89 1.61 -0.39 1.02 0.20 cortisone 1.47 [18] 1.44 0.03 1.46 0.01 1.46 0.01 0.82 0.65 1.30 0.17 1.24 0.23 phenobarbital 1.47 [18] 1.67 -0.20 1.30 0.17 1.41 0.06 1.52 -0.05 1.37 0.11 1.32 0.15 morin 1.56 [18] 1.61 -0.05 1.78 -0.22 1.74 -0.18 0.35 1.21 1.13 0.43 1.63 -0.07 clonidine 1.57 [16] 1.41 0.16 2.05 -0.48 1.80 -0.23 2.78 -1.21 1.73 -0.16 2.35 -0.78 hydrocortisone 1.61 [18] 1.43 0.18 1.72 -0.11 1.66 -0.05 0.50 1.11 1.70 -0.09 1.28 0.33 sulfadimethoxine 1.63 [18] 1.48 0.15 1.45 0.18 1.46 0.17 1.39 0.24 1.98 -0.35 1.36 0.27 ketorolac 1.68 [18] 2.08 -0.40 2.71 -1.03 2.58 -0.90 1.64 0.04 1.62 0.06 2.83 -1.15 sulfaquinoxaline 1.68 [18] 1.30 0.38 1.64 0.04 1.56 0.12 1.52 0.16 1.69 -0.01 1.57 0.11 bromazepam 1.69 [18] 1.65 0.04 2.18 -0.49 2.09 -0.40 2.10 -0.41 1.70 -0.01 2.33 -0.64 acebutolol 1.71 [18] 1.95 -0.24 1.58 0.13 1.68 0.03 0.94 0.77 1.71 0.00 1.62 0.10 benzthiazide 1.73 [18] 2.63 -0.90 1.75 -0.02 2.04 -0.31 2.30 -0.57 2.11 -0.38 1.97 -0.24 omeprazole 1.8 [16] 2.17 -0.37 2.07 -0.27 2.09 -0.29 2.17 -0.37 2.02 -0.22 2.90 -1.10 atropine 1.83 [18] 1.53 0.30 1.95 -0.12 1.86 -0.03 1.53 0.30 1.30 0.53 1.72 0.11 metoprolol 1.88 [18] 1.79 0.09 1.88 0.00 1.85 0.03 1.72 0.16 1.49 0.39 1.76 0.12 corticosterone 1.94 [18] 1.76 0.18 2.20 -0.26 2.10 -0.16 1.26 0.68 2.32 -0.38 2.02 -0.08 bendroflumethiazide 1.95 [18] 2.07 -0.12 1.32 0.63 1.49 0.46 2.22 -0.27 1.73 0.22 1.86 0.09 mepivacaine 1.95 [18] 2.04 -0.09 1.86 0.09 1.91 0.04 2.62 -0.67 2.10 -0.15 2.98 -1.03 milnacipran 2.03 [18] 1.23 0.80 1.94 0.09 1.66 0.37 1.63 0.40 1.91 0.12 1.29 0.74 flunitrazepam 2.06 [18] 1.44 0.62 2.07 -0.01 2.03 0.03 1.42 0.64 1.78 0.28 2.61 -0.55 oxprenolol 2.1 [18] 2.29 -0.19 2.29 -0.19 2.29 -0.19 2.22 -0.12 2.09 0.01 2.23 -0.13 pentobarbital 2.1 [18] 2.05 0.05 2.01 0.09 2.02 0.08 1.91 0.19 2.11 -0.01 1.91 0.19 clobazam 2.12 [18] 1.69 0.43 2.44 -0.32 2.29 -0.17 2.57 -0.45 2.44 -0.32 2.74 -0.62 carbamazepine 2.19 [18] 2.67 -0.48 2.17 0.02 2.28 -0.09 2.93 -0.74 2.38 -0.19 2.68 -0.49 spironolactone 2.26 [18] 3.12 -0.86 2.68 -0.42 2.78 -0.52 2.90 -0.64 2.65 -0.39 3.73 -1.47 clonazepam 2.41 [18] 2.34 0.07 2.57 -0.16 2.53 -0.12 2.18 0.23 2.38 0.03 2.86 -0.45 chlordiazepoxide 2.44 [18] 2.16 0.28 2.39 0.05 2.36 0.08 1.72 0.72 3.79 -1.35 2.30 0.14 naringenin 2.52 [18] 3.19 -0.67 2.49 0.03 2.62 -0.10 1.63 0.89 2.44 0.08 2.37 0.15 rosuvastatin 2.52 [15] 0.42 2.10 1.90 0.62 1.36 1.16 0.59 1.93 1.90 0.62 2.43 0.09 clofibric acid 2.57 [18] 2.72 -0.15 2.74 -0.17 2.74 -0.17 2.62 -0.05 2.82 -0.25 2.76 -0.19 furosemide 2.6 [18] 3.10 -0.50 2.21 0.39 2.35 0.25 0.74 1.86 1.90 0.70 1.40 1.20 hesperetin 2.6 [18] 2.90 -0.30 2.57 0.03 2.66 -0.06 1.50 1.10 2.29 0.31 2.36 0.24 cinchonidine 2.62 [18] 3.35 -0.73 2.79 -0.17 2.92 -0.30 2.60 0.02 2.49 0.13 2.75 -0.13 quinine 2.64 [18] 3.46 -0.82 2.86 -0.22 3.02 -0.38 2.60 0.04 2.99 -0.35 2.99 -0.35 lidocaine 2.65 [18] 2.36 0.29 2.32 0.33 2.33 0.32 2.41 0.24 1.95 0.70 2.63 0.02 tramadol 2.7 [15] 2.51 0.19 2.55 0.15 2.54 0.16 2.53 0.17 3.10 -0.40 2.70 0.00 lormetazepam 2.72 [16] 2.31 0.41 2.61 0.11 2.57 0.15 3.88 -1.16 2.40 0.32 3.71 -0.99 xipamide 2.8 [18] 2.50 0.30 2.99 -0.19 2.88 -0.08 2.73 0.07 1.93 0.87 2.71 0.09 trazodone 2.82 [18] 1.66 1.16 3.21 -0.39 2.77 0.05 2.94 -0.12 3.85 -1.03 2.42 0.40 diltiazem 2.84 [15] 3.63 -0.79 3.38 -0.54 3.43 -0.59 2.64 0.20 3.65 -0.81 3.09 -0.25 brompheniramine 2.88 [18] 3.57 -0.69 3.23 -0.35 3.33 -0.45 3.89 -1.01 3.30 -0.42 3.78 -0.90 alprenolol 2.89 [18] 2.88 0.01 2.74 0.15 2.78 0.11 2.70 0.19 2.65 0.24 2.64 0.25 propranolol 2.98 [18] 3.10 -0.12 3.32 -0.34 3.26 -0.28 2.65 0.33 2.75 0.23 2.54 0.44 diazepam 2.99 [18] 2.91 0.08 2.92 0.07 2.92 0.07 2.98 0.01 2.96 0.03 3.17 -0.18 pallicer et al. admet & dmpk 2(2) (2014) 107-114 110 apigenin 3.02 [18] 2.10 0.92 2.59 0.43 2.45 0.57 1.90 1.12 2.91 0.11 2.41 0.61 venlafaxine 3.05 [15] 2.91 0.14 3.33 -0.28 3.15 -0.10 2.81 0.24 3.27 -0.22 3.02 0.03 rosiglitazone 3.1 [15] 2.56 0.54 2.87 0.23 2.75 0.35 3.21 -0.11 3.02 0.08 3.27 -0.17 pyrilamine 3.12 [18] 2.75 0.37 3.14 -0.02 3.05 0.07 3.30 -0.18 3.23 -0.11 3.09 0.03 quetiapine 3.13 [15] 1.57 1.56 2.95 0.18 2.37 0.76 3.91 -0.78 2.99 0.14 2.66 0.47 ketoprofen 3.14 [18] 2.81 0.33 3.27 -0.13 3.12 0.02 3.31 -0.17 2.76 0.38 3.36 -0.22 chlorphenamine 3.17 [18] 3.39 -0.22 2.95 0.22 3.09 0.08 3.62 -0.45 3.15 0.02 3.70 -0.53 fenbufen 3.2 [18] 3.13 0.07 3.22 -0.02 3.21 -0.01 2.34 0.86 3.14 0.06 2.93 0.27 naproxen 3.24 [15] 3.00 0.24 2.98 0.26 2.98 0.26 2.86 0.38 2.82 0.42 2.85 0.39 warfarin 3.25 [18] 3.42 -0.17 2.99 0.26 3.11 0.15 2.97 0.28 2.90 0.35 2.83 0.42 diphenhydramine 3.27 [18] 3.66 -0.39 3.72 -0.45 3.71 -0.44 3.25 0.02 3.45 -0.18 3.38 -0.11 bupivacaine 3.41 [18] 3.64 -0.23 2.87 0.54 3.08 0.33 3.86 -0.45 3.69 -0.28 4.31 -0.90 clotiazepam 3.49 [18] 3.05 0.44 3.01 0.48 3.02 0.47 3.91 -0.42 3.03 0.46 3.77 -0.28 chrysin 3.52 [18] 2.88 0.64 3.24 0.28 3.11 0.41 2.29 1.23 3.56 -0.04 2.65 0.87 valsartan 3.59 [18] 4.75 -1.16 3.52 0.07 3.87 -0.28 0.00 3.59 3.63 -0.04 4.54 -0.95 haloperidol 3.82 [18] 3.01 0.81 3.61 0.21 3.48 0.34 3.49 0.33 3.85 -0.03 3.89 -0.07 flurbiprofen 3.84 [15] 4.12 -0.28 3.71 0.13 3.82 0.02 3.18 0.66 3.75 0.09 3.68 0.16 progesterone 3.87 [18] 4.04 -0.17 3.58 0.29 3.72 0.15 3.78 0.09 3.78 0.10 3.86 0.01 celecoxib 3.91 [18] 4.21 -0.30 3.02 0.89 3.24 0.67 4.34 -0.43 4.37 -0.46 4.43 -0.52 flurazepam 3.94 [18] 3.99 -0.05 3.47 0.47 3.55 0.39 3.81 0.13 4.22 -0.28 4.21 -0.27 glimepiride 3.97 [15] 2.94 1.03 3.44 0.53 3.17 0.80 2.59 1.38 3.96 0.01 3.78 0.19 ibuprofen 3.97 [18] 3.72 0.25 3.21 0.76 3.37 0.60 3.14 0.83 3.68 0.29 3.61 0.36 atorvastatin 4.08 [15] 4.13 -0.05 4.43 -0.35 4.36 -0.28 5.05 -0.97 4.46 -0.38 5.56 -1.48 indometacin 4.1 [15] 3.10 1.00 4.48 -0.38 4.02 0.08 3.69 0.41 4.18 -0.08 4.24 -0.14 nortriptyline 4.36 [18] 5.65 -1.29 4.41 -0.05 4.76 -0.40 4.25 0.11 4.32 0.04 4.24 0.12 loratadine 4.4 [18] 5.94 -1.54 5.15 -0.75 5.32 -0.92 4.13 0.27 5.05 -0.65 5.00 -0.60 fluoxetine 4.42 [15] 4.09 0.33 4.35 0.07 4.27 0.15 4.27 0.15 4.57 -0.15 4.03 0.39 diclofenac 4.5 [18] 4.06 0.44 4.61 -0.11 4.48 0.02 4.12 0.38 4.73 -0.23 4.37 0.13 clopidogrel 4.52 [15] 4.23 0.29 4.20 0.32 4.21 0.31 3.74 0.78 4.21 0.31 3.74 0.79 duloxetine 4.54 [15] 3.73 0.81 4.17 0.37 4.03 0.51 4.33 0.21 4.26 0.28 3.85 0.69 penbutolol 4.62 [16] 4.17 0.45 3.84 0.78 3.97 0.65 3.44 1.18 3.64 0.98 3.57 1.05 maprotiline 4.67 [18] 4.51 0.16 5.23 -0.56 4.95 -0.28 4.59 0.08 4.52 0.15 4.12 0.55 imipramine 4.8 [18] 4.80 0.00 4.65 0.15 4.69 0.11 4.32 0.48 5.04 -0.24 4.39 0.41 amitriptyline 4.92 [18] 4.92 0.00 4.67 0.25 4.72 0.20 4.42 0.50 4.85 0.07 4.77 0.15 cyproheptadine 4.92 [18] 6.41 -1.49 4.65 0.27 5.05 -0.13 3.97 0.95 5.30 -0.38 4.36 0.56 sertraline 5.17 [15] 4.81 0.36 5.40 -0.23 5.18 -0.01 5.03 0.14 5.35 -0.18 5.00 0.17 flufenamic acid 5.19 [15] 5.62 -0.43 5.26 -0.07 5.34 -0.15 3.98 1.21 5.53 -0.34 3.95 1.24 chlorpromazine 5.27 [15] 5.20 0.07 5.42 -0.15 5.36 -0.09 4.24 1.03 5.30 -0.03 4.74 0.53 miconazole 5.34 [18] 5.93 -0.59 6.28 -0.94 6.20 -0.86 5.09 0.25 5.81 -0.47 5.65 -0.31 rimonabant 5.57 [18] 6.01 -0.44 5.27 0.30 5.45 0.12 6.28 -0.71 6.47 -0.90 6.61 -1.04 clofazimine 6.3 [18] 7.26 -0.96 7.49 -1.19 7.44 -1.14 5.39 0.91 7.70 -1.40 7.14 -0.84 mean = -0.01 -0.02 -0.01 0.22 0.00 -0.07 sd = 0.65 0.43 0.41 0.71 0.47 0.55 according to the differences of table 1, clogp, classic acd/logp, consensus/logp, and galas/logp are the most accurate methods with mean deviations very close to 0. the precision of clogp, consensus/logp, and galas/log p is also very good with a standard deviation between 0.4 and 0.5 logp unities. the standard deviation of classic/logp is somewhat larger (0.65). alogp is less accurate and precise (mean of -0.07 and standard deviation of 0.55) and chemprop has the poorest accuracy and precision (mean of deviations of 0.22 and standard deviation of 0.71). figure 1 shows the correlation between the results of the calculation by the 6 different methods and reference log po/w values and table 2 the regression coefficients (r 2 ) and standard deviations obtained for those correlations. all methods give good linear correlations with r 2 values between 0.79 and 0.93.the results presented in table 2 are consistent with those of table 1. consensus/logp and galas/logp give the best results: they give exact results because the slopes and intercepts are not different from 1 and 0, respectively, at 95 % confidence level, but also the most precise (with the lowest overall standard deviations, 0.43 and 0.41). alogp, classic acd/logp and chemprop/logp have also slopes and intercepts not significantly different from 1 and 0, giving thus exact results but less precise (sd of 0.55, 0.66 and 0.70, respectively). the number of outliers for these last methods is higher too (table 3). clogp gives quite precise results (sd = 0.46) but it shows a slope slightly higher than one and an intercept lower than zero. thus, clogp may predict slightly negative deviations for hydrophilic compounds but positive deviations for hydrophobic ones. in fact, all methods have a lower prediction power for extreme ranges of log po/w values, admet & dmpk 2(2) (2014) 107-114 evaluation of log p values doi: 10.5599/admet.2.2.45 111 being the range of compounds with log p higher than 5 where accuracy tends to be poorer. this can be seen in table 3 which reports the number of outliers in the respective log p ranges. figure 1. comparative linear relationship between log p reference and the different predictive log p softwares (n =103). table 2. regression coefficients and standard deviations for the different predictive log p softwares. classic galas consensus chemprop clogp alogp r 2 0.84 0.92 0.93 0.79 0.93 0.87 sd 0.66 0.43 0.41 0.70 0.46 0.55 y=a+bx a (sd) 0.00 (±0.13) 0.09 (±0.09) 0.09 (±0.08) 0.02 (±0.14) -0.23 (±0.09) 0.23 (±0.11) b (sd) 1.00 (±0.04) 0.97 (±0.03) 0.97 (±0.03) 0.91 (±0.05) 1.09 (±0.03) 0.94 (±0.04) table 3. number of outliers with deviations higher than 0.6 units for the different predictive log p softwares and the different ranges of log po/w reference values. n * classic galas consensus chemprop clogp alogp log pref>5 6 1 2 2 4 2 3 5<log pref>1 81 22 7 8 30 9 19 log pref<1 16 4 6 5 4 4 6 * n is the total number of compounds in the log p range although galas/logp and consensus/logp values seem to be the most accurate methods for this set of compounds it has to be taken into account that the result of the fragmental method in those programs is just a baseline value, which is corrected adding a similarity weighted average from the differences between the baseline predictions and the experimental data for the most similar compounds in the training set to the submitted compound. this strategy must achieve to best fit the training set data, but the quality of the results for other compounds will depend a lot on the similarity of those to the training set. this may be pallicer et al. admet & dmpk 2(2) (2014) 107-114 112 quite high in our case, being all the 103 compounds known pharmaceutical compounds with public experimental values. in fact we have observed that at least for 87 compounds out of the 103, the same compound could be found in the training set, being their initial prediction therefore corrected to the experimental value with the highest weight. this handicap is overcome in clogp where results obey exclusively to the general formula, fragments, correction factors and coefficients developed for the method. in fact this method shows almost the same predictability than galas/logp, the least number of outliers or unacceptable predictions (>0.6) among all methods, and it has better results than classic acd/logp and alogp for this set of compounds. finally, chemprop shows the poorest performance according to accuracy, precision, and number of outliers. in order to observe the performance of galas/logp and consensus/logp for compounds that are not part of the training set, we have evaluated the prediction for the remaining 16 compounds, being the results shown in figure 2. we show the comparison as well for the remaining methods although in those cases we do not know whether those 16 compounds are part of the training set or not. we can see in table 4 that accuracy of predictions for galas/logp has diminished while clogp, classic acd/logp and alogp predictions retain their predictability levels. figure 2. comparative linear relationship between log p reference and the different predictive log p softwares (n =16: compounds not in acd training set). table 4. regression coefficients and standard deviations for the different predictive logp softwares (n =16: compounds not in acd training set). classic galas consensus chemprop clogp alogp r 2 0.78 0.83 0.84 0.58 0.93 0.80 sd 0.82 0.48 0.54 1.25 0.42 0.69 y=a+bx a (sd) -0.37 (±0.52) 0.96 (±0.31) 0.48 (±0.34) -0.20 (±0.79) 0.20 (±0.27) 0.51 (±0.44) b (sd) 1.02 (±0.15) 0.72 (±0.09) 0.82 (±0.10) 0.98 (±0.22) 0.98 (±0.07) 0.92 (±0.12) in any case to better estimate the performance of the methods, it would be desirable to compare calculated versus experimental log po/w value for a larger set of highly diverse compounds that had low structural similarity to compounds in the training set. it is worth to mention that to leverage that, admet & dmpk 2(2) (2014) 107-114 evaluation of log p values doi: 10.5599/admet.2.2.45 113 galas/logp offers the possibility to enlarge the training set with new compounds and their determined experimental data, which might be of use when predicting lipophilicity for analogs of on-going chemistry within a discovery program. at that point we would like to stress that all the predictive software are initiated from experimental values obtained from the literature and not in all cases literature values of different sources coincide. to study those differences in some depth we used the mentioned 87 compounds that are at the same time part of acd’s training set and compared our reference experimental log po/w values as summarize in table 1 with the acd reference experimental log po/w values. we have centered our comparison on those acd values as the experimental log po/w values used to train the acd models can be directly obtained within the software, however other training sets used to derive clogp or alogp are not so easily available. we observe that the values do not show a perfect correlation (figure 3). a slope of 0.92 and a y-intercept of 0.2 are obtained. although the variability in the experimental determination is a fact fully known, it is important to have that correlation value between experimental measures in mind when evaluating prediction methods, as in light of this, we cannot expect to obtain a correlation between experimental and predicted values higher than the 0.95. y = 0.9169x + 0.2185 r 2 = 0.9452 1 0 1 2 3 4 5 6 7 8 1 0 1 2 3 4 5 6 7 8 acd logpo/w ref u se d l o g p o / w r e f figure 3. linear relationship between log po/w used in acd database and the log p o/w reference used in this study (n=87). conclusions all tested substructure-based methods have an acceptable accuracy and precision for estimating the lipophilicity of pharmaceutical compounds, where predictions tend to loose accuracy in the extreme ranges. in the case of our dataset consensus/logp and galas/logp show the best results, however being the globally predicted values corrected depending on the performance of the method for the most similar compounds, and being our 103 compounds known pharmaceutical compounds, an analysis for a set of highly diverse compounds that had low structural similarity to compounds in the training set would be needed to verify its precision in evaluating log p values of a general dataset of compounds. among the other methods that do not apply any structural analog approach clogp is performing best whereas chemprop/logp is showing the lowest predictability. pallicer et al. admet & dmpk 2(2) (2014) 107-114 114 acknowledgements: financial support from the ministerio de ciencia e innovación of the spanish government and the fondo europeo de desarrollo regional (feder) of the european union (project ctq2010-19217/bqu) is acknowledged. references [1] y. henchoz, b. bard, d. guillarme, p.a. carrupt, j.l. veuthey, s. martel, analytical and bioanalytical chemistry 394 (2009) 707-729 [2] m.j. waring, expert opinion on drug discovery 5 (2010) 235-248 [3] e.h. kerns, l. di, drug discovery today: technologies 1 (2004) 343-348 [4] r. mannhold, g.i. poda, c. ostermann, i.v. tetko, journal of pharmaceutical sciences 98 (2009) 861-893 [5] r.f. rekker, r. mannhold, calculation of drug lipophilicity. weinheim: vch, 1992 [6] broto p, moreau g, vandycke c. european journal of medicinal chemistry 19 (1984) 71–78. [7] r. wang, y. gao, l.h. lai, perspectives in drug discovery and design 19 (2000) 47–66. [8] r. mannhold, g.i. poda, c. ostermann, i.v. tetko, journal of pharmaceutical sciences, 98 (2009) 861–893. [9] a.k. ghose, g.m. crippen, journal of computational chemistry, 7 (1986) 565–577 [10] accelrys software inc., discovery studio modeling environment, release 3.5, san diego: accelrys software inc., 2012 [11] a. leo, chemical review, 93 (1993) 1281-1306 [12] cambridgesoft, chemfinder ultra version 10.0.3. [13] a.a. petrauskas, e.a. kolovanov, perspectives in drug discovery and design 19 (2000) 99116. [14] acd percepta, acd/labs 2014 release. [15] j.m. pallicer, c. calvet, a. port, m. rosés, c. ràfols, e. bosch, journal of chromatography a 1240 (2012) 113-122 [16] a. avdeef, partitioning into octanol. in: absorption and drug development. john wiley & sons inc., hoboken, nj, usa, 2003, p. 42-66 [17] c. hansch, a. leo, exploring qsar, hydrophobic, electronic and steric constants. american chemical society, washington, dc. usa, 1995 [18] bioloom online database: http://www.davidhoekman.com/bioloom/bin/details ©2014 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://rd.springer.com/search?facet-author=%22yveline+henchoz%22 http://rd.springer.com/search?facet-author=%22bruno+bard%22 http://rd.springer.com/search?facet-author=%22davy+guillarme%22 http://rd.springer.com/search?facet-author=%22pierre-alain+carrupt%22 http://rd.springer.com/search?facet-author=%22jean-luc+veuthey%22 http://rd.springer.com/search?facet-author=%22sophie+martel%22 http://www.davidhoekman.com/bioloom/bin/details http://creativecommons.org/licenses/by/3.0/ manuscript doi: 10.5599/admet.3.3.189 155 admet & dmpk 3(3) (2015) 155-181; doi: 10.5599/admet.3.3.189 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index review nanotechnology platforms in parkinson’s disease rishi rajat adhikary 1 , puja sandbhor 1 and rinti banerjee * 1 research scholar, department of biosciences and bioengineering, indian institute of technology bombay, powai, mumbai, india400076. both authors contributed equally to the writing of this review paper. *department of biosciences and bioengineering, indian institute of technology bombay, powai, mumbai, india 400076. *corresponding author: e-mail: rinti@iitb.ac.in; tel.: +912225767868; fax: +912225723480 received: june 04, 2015; revised: august 21, 2015; published: september 05, 2015 abstract parkinson’s disease (pd) remains a serious concern due to its effects on the quality of life of patients and its socioeconomic burden to society. present day management of pd has limitations in both diagnosis and treatment. nanotechnology may provide smart solutions to this problem. the present review highlights the recent advancements in the development of nanotechnology platforms for pd. the review focuses on the use of such platforms in diagnostics, treatments, deep brain stimulation, neurosurgery and other modalities of management and the role of nanotechnology in each of these fields. the review also sheds light on the translation of technologies from labs to clinics and the essential advantages as well as concerns that accompany the translation. keywords smart materials; nanoparticles; sensors; nanotechnology; drug delivery; blood brain barrier. ____________________________________________________________________________________________ introduction parkinson’s disease (pd), in the present day, ranks second after alzheimer’s disease among the most common neurodegenerative diseases and has characteristic motor and non-motor symptoms [1]. globally, the prevalence of pd has been found to be close to 0.3 % of the total population and the crude incidence rate is 4.5–19 per 1,00,000 population per year [2,3]. the world health organization computed the disability-adjusted life years (dalys) or the number of years lost of healthy life, due to pd to be 1,617,000 globally in 2005 and projected it to increase to 2,015,000 years by 2030 [3]. it would mean that more and more people would be leading lives with disabilities due to parkinson’s disease. besides, there is a huge cost that is incurred globally for the management of pd. in developed countries like the united states, the national economic burden of pd was more than usd 14.4 billion in 2010 (approximately usd 22,800 per patient) [4]. in developing countries like india, patients with pd were found to spend nearly 16 % to 41.7 % of the average indian gross national income as direct costs in the treatment of pd [5]. thus, pd is a universal disorder and in addition to the clinical features, the socioeconomic impact of pd is a cause of concern. the management of pd has several limitations at present. the clinical diagnosis of pd originates to the http://www.pub.iapchem.org/ojs/index.php/admet/index r. banerjee et al. admet & dmpk 3(3) (2015) 155-181 156 acclaimed neurologists like parkinson and charcot who observed that the presence of two or three motor features including bradykinesia, rigidity and tremor was characteristic of parkinson’s disease [6]. two centuries later, the issue with the clinical diagnosis of parkinson’s disease today is that it fails to differentiate it from the many other forms of parkinsonism [7]. also, it requires expert clinicians for the evaluation of these subjective features of the disease. thus, there is a need for better biomarkers that can provide for an objective assessment of the presence or absence of pd. beyond the diagnosis, standard treatment modalities comprise of therapy using dopamine agonists and dopamine precursors to improve clinical symptoms associated with disease progression, patient’s care and quality of life. however, short half-lives, treatment intolerance together with side effects like levodopa (l-3,4-dihydroxyphenylalanine or l-dopa) induced dyskinesias limit the long term treatment efficacy. further, therapy is limited by the presence of the “blood-brain-barrier”. thus, in the sphere of treatment for pd, there is a need to develop efficient therapies with minimal side-effects and increased bioavailability in the central nervous system. nanotechnology, an emerging tool, has the potential to improve these circumstances by introducing novel carrier-based platforms that will target selective release of drug payload with on-demand and controlled release kinetics and increased reach via modulating or by-passing the blood-brain-barrier. the present review describes the use of nanotechnology-based platforms for diagnosis, drug delivery and other therapeutic approaches. epidemiology of parkinson’s disease prospective and well-designed population based studies reported that prevalence and incidence rates of pd were age-specific i.e. onset of disease is rare below 50 years of age whereas sharp rise (~1-1.5 %) is seen over 60 year of age resulting in prevalence less than 0.04 % below the age 40 years, which rises to 19 % above the age of 80 years [2,8]. the prevalence of pd varies by gender (men are more prone to pd than female, possibly due to the neuroprotective effect of estrogens), race/ ethnicity (incidence rates were found to be highest among hispanics>non-hispanic>whites>asian>blacks) [9]. differences in environmental factors like occupational exposures (pesticides, heavy metals) and cigarette smoking; dietary factors (anti-oxidants like vitamin e, fatty acids mainly unsaturated fatty acids, dietary iron etc.), genetic factors like α-synuclein (park1), parkin (park2), genes involved in dopamine metabolism, homocysteine metabolism and polymorphism in mitochondrial dna etc. are responsible for pd progression with variations in its prevalence and incidence [9]. etiology and pathophysiological hallmarks of parkinson’s disease for designing of successful diagnostic and therapeutic strategies against pd it is important to understand the etiology and pathophysiology behind development and progression of pd. pd is a chronic, progressive neurodegenerative disease characterized by impairment in motor and non-motor activities with symptoms like akinesia, tremor, muscle rigidity, postural instability, freezing of gait and depression, delirium, dementia, sleep disorders, autonomic symptoms (bradycardia, sialorrhea, sexual dysfunctions and dry eyes syndromes etc.) and other symptoms like olfactory disturbances, fatigue, weight changes etc. the classical pathophysiological hallmark of pd is the presence of lewy bodies in the dopaminergic neurons that are left behind after degeneration of these neurons in substantia nigra pars compacta. this results in dopamine depletion in the striatum and basal ganglia nuclei due to innervations of dopamine neurons in two segments of globus pallidus i.e. external and internal parts of basal ganglia nuclei and also auxiliary to subthalamic nucleus and thalamus [10]. the etiological factors include toxins mainly mptp (1methyl-4-phenyl-1,2,3,6-tetrahydropyridine) which produced mpp+ (1-methyl-4-phenylpyridinium), a admet & dmpk 3(3) (2015) 155-181 nanotechnology platforms in parkinson’s disease doi: 10.5599/admet.3.3.189 157 neuro-toxic precursor by monoamine oxidase-b (mao-b) that further interferes in mitochondrial metabolism leading to free radical generation and subsequently causes dopaminergic neuron degeneration in the substantia nigra [11]. various other toxins include taclo (formed from interaction between sedative agents like chloral or organic solvents like trichloroethylene with endogenous tryptamine) were also reported to be toxic to dopaminergic neurons [12]. other pathological hallmarks includes degeneration of non-dopaminergic neurons leading to depletion of nor-adrenaline, serotonin (depletion or hyperactivity of 5-hydroxy tryptamine in the brain giving rise to symptoms such as depression and hallucinations), glutamic acid, gamma-amino butyric acid (gaba), neuropeptides like endorphins (enkephalins and dynorphins) and acetylcholine etc. [13-16]. the pathophysiology of pd remains ambiguous, and is under continuous active investigations till date. olanow and tatton (1999) have reported that potential genes like glutathione peroxidase, superoxide dismutase-1, 2 and dopamine receptors (d2-4) etc. contribute to 5-10 % familial pd with autosomal dominant or recessive inheritance [17]. the pathogenesis of pd includes oxidative stress in substantia nigra through production of reactive oxygen species (ros) and hydrogen peroxide by extensive oxidative metabolism of dopamine. collectively, the increased turnover of dopamine, decreased glutathione levels and presence of reactive iron promotes generation of free radicals which may be responsible of disease development and progression [17]. previous studies reported oxidative damage to the dna {elevated levels of 8-hydroxyguanine (8-ohg)}, in substantia nigra of pd patients due to free species like peroxynitrite, hydroxyl radicals etc. [18]. in 1990, schapira and co-workers reported the role of increased mitochondrial dysfunction in pd patients and association with defects in mitochondrial complex-i of the respiratory chain leading to nigrostriatal neuronal degeneration through depletion of atp synthesis [19]. several studies demonstrated that different genes like α-synuclein, parkin and uch-l1 (gene encoding ubiquitin carboxyl-terminal hydrolase l1) act concomitant with an increased mitochondrial dysfunction and oxidative damage in pd patients [20]. song et al. (2004) reported a co-relation between α-synuclein and mitochondrial function, since α-synuclein transgenic mice were susceptible to neurotoxic effects induced by mptp whereas, resistance was observed in α-synuclein knockout mice [20]. sheng et al. reported the up-regulation or mutation in α-synuclein leads to formation of abnormal and misfolded proteins through series of events like oligomerization, fibrilization and aggregation within the nerve cells (formation of lewy bodies) that responsible for mitochondrial dysfunction, impaired neuronal homeostasis and lead to subsequent neurodegeneration [21]. several studies reported the role of genetic factors in pathogenesis of pd such as extensive generation of lewy bodies one of the pathophysiological hallmarks associated with mutated α-synuclein causing early and late onset of pd with rapid disease progression whereas, mutated lrrk2 showed late onset of pd (at 80 years of age) with lewy bodies formation, however dementia was not observed and also mutated atp13a2 cause an atypical pd with dementia [22-24]. another study demonstrated that activated macrophages and microglia responsible for secretion of different neurotrophic factors like brain derived neurotrophic factor (bdnf), and glial derived neurotrophic factor (gdnf) cause dopaminergic sprouting in nigrostriatal injury in pd patients [25]. paul and colleagues reported the significant role of ubiquitin–proteasome system (ups) and autophagy (macroautophagy) pathways in lysosomal and proteasomal dysfunction, another pathophysiological hallmark of pd [26]. further studies reported the damage to ups and lysosomal pathways (increased protein accumulation due to impaired unwanted protein clearance mechanism) through mutated genes like parkin, lrrk2 causing neuronal toxic insult in pd [27]. r. banerjee et al. admet & dmpk 3(3) (2015) 155-181 158 impaired iron metabolism concomitant with different genetic (α-synuclein aggregation) and environmental (tobacco, cigarette smoking) factors was observed in pathogenesis of sporadic or familial pd [28]. mitochondrial byproducts (peroxide and superoxide) can reactive with iron by generating most damaging hydroxyl radicals causing dopaminergic cell death in substantia nigra. these findings further supported by chronic mptp treated mouse studies correlating iron levels with selective dopamine degeneration through lipofuscin aggregation and up-regulated α-synuclein [29]. pla2g6 and atp13a2 are two important genetic factors associated with increased neurodegeneration in brain. iron accumulating in substantia nigra of pd patients have been reported by several studies however their precise mechanism of action is not yet clear [30,31]. despite the significant progress in molecular and cellular studies, the detailed pathophysiology of pd is not yet clear. present modalities of management and their limitations current techniques for diagnosis of pdclinical diagnosis and imaging accurate diagnosis plays an important role in the precise management of pd. currently, diagnosis of pd is done on the basis of different clinical symptoms of parkinsonism like bradykinesia, tremor, postural imbalance, dementia, sleep disorders, sialorrhea, dry eyes syndromes and olfactory disturbances, fatigue, weight changes [32]. presence of lewy bodies in substantia nigra, hypothalamus, basal nuclei, and cranial motor nerve and its nuclei etc. is an important biomarker for pd diagnosis. different genetic and imaging testing were carried out for diagnosis of pd. identification of genetic mutations revealed that α-synuclein, nurr1, lrrk2 and uch-l1 are responsible for autosomal-dominant inheritance whereas, dj1, parkin and pink1 are responsible for autosomal-recessive inheritance [32]. the cranial ct scan, mri and pet (fluorine18 labeled dopa) imaging, spect (single photon-emission computed tomography) and transcranial ultrasound imaging etc. are valuable techniques for differentiation and diagnosis of pd from other neurodegenerative disorders. for instance, spect tracer used for presynaptic dopamine transporter imaging like fp-cit (n-ω-fluoropropyl-2β-carbomethoxy-3β-(4-iodophenyl) nortropane binds to the dopamine transporters (dat) protein. thus, integrity of nigrostriatal dopaminergic transporter can be used as a diagnostic marker for detection of early pd progression [33]. however, accessibility and cost of these techniques restrict their usefulness and application in diagnosis of pd. pharmacotherapy and management of parkinson’s disease currently different therapies are available for treatment of motor and non-motor complications associated with pd. pharmacotherapy of pd includes treatment with drugs like l-dopa, dopamine receptor agonist, catechol-o-methyltransferase (comt) inhibitors, antagonists of mao-b and glutamate and amantadine etc. l-dopa is most widely used therapy for treatment of pd because it easily crosses bbb and is converted into dopamine in presence of an enzyme dopa-decarboxylase. however, long term treatment with l-dopa results in various side effects like dyskinesia (involuntary and non-rhythmic choreic or choreodystonic movements), alterations in on-off periods giving rise to motor fluctuations, nausea and vomiting [34]. therefore, in clinical practice l-dopa is co-administered with carbidopa (dopa-decarboxylase inhibitor) in which carbidopa itself does not cross bbb but reduces conversion of l-dopa to dopamine in peripheral nervous system and thereby increase its delivery to the brain [35]. comt inhibitors (entacapone) are ineffective when administered alone, therefore they are administered as adjuvants with levodopa and carbidopa in pd patients to avoid end-off-dose alterations [36]. early pd symptoms are treated with mao-b inhibitors like selegiline or rasagiline. recently, schapira et al. (2013) reported the dual action of safinamide (selective and reversible mao-b inhibitors that decrease glutamate release via admet & dmpk 3(3) (2015) 155-181 nanotechnology platforms in parkinson’s disease doi: 10.5599/admet.3.3.189 159 blocking voltage depended sodium channels) by preventing dyskinesia and motor fluctuations in early pd patients [37]. further randomized and double blind phase ii/iii studies demonstrated the anti-parkinson efficacy of safinamide therapy in pd patients [38]. dopamine (post-synaptic d2 receptors) agonists such as non-ergot derivatives (pramipexole, ropinirole and rotigotine) are also used in early pd patients (above age of 60 years). however, treatment with dopamine agonists is not free from side effects like confusion, hallucinations and psychosis etc. recent studies described the role of adenosine a2a receptor antagonists in treatment of pd through increased activation of dopamine (d2) receptors via inhibiting adenosine receptors in striatum and basal ganglia nuclei. for instance, adenosine a2a receptor antagonist (istradefylline) an adjuvant with levodopa has significant potential to prevent levodopa induced off-time fluctuation and dyskinesia [39]. glutamate, an excitatory neurotransmitter in brain, has a significant role in pathogenesis of pd; hence glutamate receptors (subtypes of glutamate receptors like n-methyl-d-aspartic acid (nmda), α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (ampa)) antagonists can be coadministered with levodopa to alleviate dyskinesia and other motor symptoms. recent clinical trials with afq056 (mglur5 i.e. metabotropic glutamic receptor antagonist) demonstrated an anti-dyskinetic therapy with improvement in motor symptoms. however, glutamate receptor antagonists are associated with various side effects like fatigue, and gastrointestinal symptoms [40]. different antioxidants (coenzyme q10 and n-acetyl-cysteine) that scavenge reactive oxygen species may have potential in preventing oxidative stress and dopaminergic neuronal degeneration in the nigrostriatal pathway in pd [41]. neuroprotective therapies include use of various neurotropic factors like gdnf may induce dopamine differentiation/regeneration, and improved motor functions [42]. however, unsatisfactory pre-clinical outcomes were observed since neurotropic factors have limited bbb permeability when administered systemically. their direct administration through intraputaminal implantation into basal ganglia has been reported in a randomized clinical trial by lang and colleagues [43]. however, patient-friendly and non-invasive approaches need to be developed for effective and safe neurotropic therapy in pd in the future [43]. neuronal signaling in the central nervous system occurs through a combination of different electrical and chemical signals which maintain the homeostasis by precise regulation of ionic composition across the system. however, this neuronal regulation is defined by presence of barrier layers across the system. there are three different types of barriers at the interface of blood and the central nervous system i.e. bloodbrain-barrier (bbb), blood-cerebrospinal fluid-barrier (bcsfb) and the arachnoids barrier. the bbb is widely studied as it is the major limiting factor for exchange between blood and brain. bbb with average surface area of about 200cm 2 g-1 is formed by the brain microvascular capillary endothelial cells (bmvec) comprising tight junctions (tjs) formed by proteins like occludins, claudins, zonula occludens etc., adherent junctions (aj) proteins like catenins, cadherins; and junctional adhesion molecules [44], astrocytes (bbb endothelium enclosing glial cells), basement membrane and pericytes (undifferentiated connective tissue cells that promote proliferation, migration and differentiation of endothelial cells). together these structural components play a dynamic role in maintaining brain homeostasis, barrier function and integrity of the bbb [45]. tight junctions maintain high transepithelial electrical resistance (1500-2000ω cm 2 over other peripheral microvessels) which shields the brain from different neurotoxins. the bmvec lacks fenestrations and is reported to have a higher mitochondrial load as compared to other endothelial cells consequently playing an important role in the active transport of nutrients to the brain [46]. different efflux transporters (p-glycoprotein, multidrug resistance associated proteins), enzymes (γ-glutamyl transpeptidase, alkaline phosphatase) and receptors (low density lipoproteins receptor, epidermal growth factor receptor) are expressed on either luminal or abluminal membrane surfaces of bbb to prevent r. banerjee et al. admet & dmpk 3(3) (2015) 155-181 160 transport of substances across the barrier [47,48]. the high electrical resistance of tight junctions restricts the entry of macromolecules whereas ions and other polar compounds can only access the bbb via paracellular diffusion. there are various specific and selective pathways for transport of different nutrient molecules across bbb (like solute carriers, transporters for glucose, amino acids etc. however, presence of efflux transporters like p-gp and tight junctions within the bbb are the additional challenges for effective drug delivery to the cns [49,50]. standard pharmacotherapy for pd has a poor therapeutic efficacy, with only a symptomatic relief i.e. they do not modify or identify the progress or root cause of the disease and show treatment failure often associated with various side effects. the limitations of current therapeutic modalities, as outlined in fig. 1 need to be addressed for the appropriate treatment of pd. figure 1. limitations in the treatment of parkinson’s disease neurostimulation by surgical approaches different surgical approaches have been under clinical practices for the management of pd. deep brain stimulation is one of the surgical approaches that involve implantation of ct/mri guided electrodes in the subthalamic nucleus or in the globus pallidus. the impulse generator creates electrical stimulation affecting the firing pattern of neurons in the implanted area along with the release of neurotransmitters glutamate and adenosine and improves the motor symptoms in pd with neurogenesis in addition to causing an increased blood circulation in the stimulated regions [51]. several studies have reported that deep brain stimulation as an adjuvant with pharmacotherapy showed significant improvements in patient daily activities and quality of life by reducing motor disability, levodopa induced motor fluctuations [52]. however, it is often associated with many adverse effects like intracranial hemorrhage, cognitive impairment, anxiety, infection and post-operative seizures. [53,54]. thalamotomy, is an invasive approach performed using a probe in region contralateral to the tremor and lesion in severe pd cases. recently transcranial mri guided focus ultrasound technology was use to perform thalamotomy which significantly reduced tremor, postural disability and improved the quality of life over a period of one year posttreatment. however, side effects include cognitive impairment, paresthesia of tongue and figure, headache and nausea [55]. gamma knife (gk) is another non-invasive surgical procedure where the emitted gamma admet & dmpk 3(3) (2015) 155-181 nanotechnology platforms in parkinson’s disease doi: 10.5599/admet.3.3.189 161 radiations intersect within the target brain region within 15-30min under anesthesia and led to a significant improvement in tremor and daily activities after several months of treatment. however, gk procedure suffers from lack of target region accuracy hence demands guided assist for more precise neurostimulation against pd [56]. nanotechnology as a viable alternative nanotechnology and nanoscience are the interdisciplinary areas involving the development of nanosized materials. the foundations of these areas and the concept of miniaturization involved were introduced by richard feynman half a century ago in his renowned talk titled: “there's plenty of room at the bottom.” [58]. principally, nanoscience is directed toward manipulation of materials at atomic or subatomic levels whose properties differ significantly from those of bulk matter whereas nanotechnology aims to exploit such manipulated materials for designing, characterizing, and production of improved structure, devices and systems with controlled size and shape (1-100 nm) for various applications [57]. nanotechnology has found numerous applications in biomedical sciences referred to as nanomedicine. for the purposes of nanomedicine, the definition of a nanomaterial is not restricted to those below 100 nm in size but extend to various submicron sized materials. the key aspect is to exploit the size dependant change in properties of submicron sized materials that can be exploited for altered cellular responses. nanostructures have been applied for diagnosis of diseases, imaging, as devices, for tissue regeneration as well as for drug delivery and therapeutics. nanotechnology which involves the manipulation of various systems at the nanoscale can have potential in the management of parkinson’s disease. a few advantages of nanotechnology platforms in pd are enlisted below: 1. nanotechnology can help in the development of sensors that are able to sense various biomarkers at low concentrations in the presence of other analytes and help in the development of affordable diagnostic devices. 2. nanotechnology-based drug delivery platforms play an important role to achieve better therapeutic efficacy and enhance bioavailability across bbb for the treatment of cns disorders. the different platforms like active targeting using receptor mediated endoor transcytosis, stimuli responsive nano-carriers and macrophage mediated passive targeting etc. are few of the approaches that can improve the treatment of pd. 3. nanotechnology can also improve neural prosthetic devices for deep brain stimulation with better contact with the brain. 4. nanotechnology can also help improve precision in neurosurgery through nano-level precision of laser axotomies. 5. lastly, nanotechnology may be employed to develop platforms for effective stem cell therapy through various tissue engineering approaches. the role of nanotechnology in the development of smart platforms for the diagnosis and treatment in pd are discussed in the subsequent parts of the review. nanotechnology-based platforms for diagnosis a brief outline of the technologies developed for the diagnosis of pd is shown in figure 2. nanotechnology is thus being infused into the in vitro diagnostics, novel biomarker identification and bioimaging modalities for the diagnosis of pd. r. banerjee et al. admet & dmpk 3(3) (2015) 155-181 162 figure 2. use of nanotechnology in diagnosis of parkinson’s disease parkinson’s disease faces the yet unmet need for development of well-established and validated biomarkers [58]. hence, most of the present day end-points in the assessment of patient response in the clinics as well as in large clinical trials depend on subjective measures of patient response and scores obtained using scales like the unified parkinson's disease rating scale [58]. as an improvement of the classical clinical features, these scales that included features like response to levodopa treatment along with motor evaluation were developed to improve accuracy of a clinical diagnosis [7]. however these scales still needed an expert clinician. in pd specifically, the clinical characters can be substantially masked due to compensatory mechanisms for many years and it is during this long prodromal period that there is further progression of parkinson’s disease. [7]. hence, a framework needs to be developed for incorporating clinical knowledge, pathology, genetics and molecular mechanisms along with objective tests for defining the criteria for the diagnosis of parkinson’s disease [7]. this has led to effort like the parkinson’s progression markers initiative (ppmi) funded by the michael j. fox foundation as an effort to meet the unmet need of finding suitable biomarkers. various platforms to detect and quantify these biomarkers, pd-associated pathologies, imaging modalities and clinical abnormalities are being considered in the diagnosis of parkinson’s disease. a variety of nanoparticles manufactured from zinc oxide, carbon nanotubes, fullerenes etc. are described here which have the potential to be used in the diagnosis of pd. detection of biomarkers the ppmi study is performing clinical tests on blood, urine and csf samples; performing neuroimaging to identify better markers of the earlier and accurate diagnosis of pd and its progression [59]. the quantification of dopamine has been extensively used and has traditionally been estimated using spectrophotometry and chromatography. nanotechnology platforms for detection of dopamine: there are a number of nanotechnology-based approaches that simultaneously detect dopamine with other biomarkers. tashkhourian et al. developed a modified carbon-paste electrode for simultaneous quantification of dopamine and ascorbic acid [60]. the electrodes were modified to incorporate silver nanoparticles and carbon nanotubes, which led to more admet & dmpk 3(3) (2015) 155-181 nanotechnology platforms in parkinson’s disease doi: 10.5599/admet.3.3.189 163 efficient electrochemical catalysis with high selectivity and sensitivity and a detection limit in micromolars for both the compounds [60]. in another study by yue and colleagues, zinc oxide (zno) nanowire arrays (each nanowire being sub-100 nm in diameter and a couple of microns in height) were prepared upon a three dimensional foam made of graphene and the assembly could selectively detect dopamine along with urate and ascorbate by a method called differential pulse voltammetry. the addition of nanowires lead to high electrical conductivity and enhancement of sensitivity of electrochemical biosensors leading to reduction in the detection limit to 1 nanomolar for dopamine and uric acid in the serum of pd patients [61]. the field testing of this array lead to the interesting finding of a reduced ua level in 7 pd patients than in healthy individuals which can make uric acid a potential biomarker in the diagnosis of pd [61]. in yet another modification, kurzatkowska et al. developed ion-channel mimetic self-assembled monolayers of macrocyclic polyamines deposited onto gold electrodes for electrochemical determination of dopamine [62]. the corrole molecules covalently attached on the surface of gold electrodes was an example of biomimicry of ligand-gated ion channels making the electrode surface semi-permeable to a redox marker. upon addition of dopamine, a corrole-dopamine complex is formed on the monolayer leading to the recording of a positive charge on the electrode [62]. this allowed the detection limits to go down to the picomolar range [62]. thus, the detection of dopamine can be made more sensitive and the detection limits can be reduced by orders of magnitude with the incorporation of nanotechnology-based smart platforms. mitochondrial dysfunction: as another example, pd has been associated with mitochondrial dysfunction and the inhibition of complex i comprising of nadh:ubiquinone oxidoreductase has been implicated in the demise of dopaminergic neurons [19]. ma et al. developed ubiquinone-quantum dot (cdse/zns) bioconjugates in which complex i and nadh could modulate the emission from the nanocomplexes by simulating the electron-transfer system part of the mitochondrial respiratory chain [63]. the authors showed that the bioconjugates could be used to monitor changes in fluorescence to trace complex i levels in human neuroblastoma sh-sy5y cells [63]. though in the proof of concept stage, this technique has the potential to serve as an in vitro biosensor for the early detection of pd and in monitoring disease progression. α-synuclein detection: as described earlier, α-synuclein is a very important neuronal protein associated with pd and present day strategies for its quantitative detection include sophisticated, time-consuming instruments including nmr spectroscopy, fluorescence measurements, western blotting and size-exclusion chromatography [64]. an et al. developed highly ordered microfabricated arrays using gold-doped tio2 nanotubes for photoelectrochemical detection of α-synuclein [64]. the arrays were effective platforms for the immobilization of primary antibodies while retaining their stability and α-synuclein binding [64]. then, the attachment of secondary antibody and gold nanoparticle-conjugated glucose oxidase, allowed excellent sensitivity by signal amplification. glucose oxidase catalyzed the conversion of glucose into gluconic acid and hydrogen peroxide. upon irradiating the other side of the titanium foil, the holes that were formed within the valence band of the nanotubes could be scavenged by the peroxide leading to a photocurrent proportional to concentrations of α-synuclein with a detection concentration in the range of pg/ml [64]. other novel concepts: there are a few other novel diagnostic concepts that are emerging. the use of breath tests to diagnose neurodegenerative disorders is one such advancement. a number of volatile organic compounds like styrene, butylated hydroxytoluene and hexadecane have been found in the breath of pd patients [65]. as an example of the use of nanotechnology in the sensing of such novel compounds, r. banerjee et al. admet & dmpk 3(3) (2015) 155-181 164 organically functionalized nanoparticles like carbon nanotubes and gold nanoparticles were used by tisch and co-workers [66]. the sensors were effective in distinguishing the breath prints of pd from healthy states with an accuracy of 78 % [66]. similarly, the technology could identify alzheimer’s disease and differentiate it from pd. improved imaging technologies imaging in pd is dependent on a number of sophisticated techniques like cranial ct, mri, pet and spect which are mostly available at advanced healthcare centres and are expensive. a number of experimental approaches are being developed and have potential for the imaging and earlier diagnosis of pd. a few of these have been elaborated below. iron metabolism: homeostasis of various metal ions which are found to be altered characterize neurodegeneration in pd. ortega and colleagues used this for chemical nano-imaging in vitro using synchrotron x-ray fluorescence nanoprobe (88 nm beam) which allowed detection of 10 −18 g of iron within cellular structures around 100 nm diameter [67]. results suggested that pd showed elevation of iron in the substantia nigra pars compacta, with loss of tyrosine hydroxylase activity that may lead to reduced iron-dopamine binding making the dopaminergic neurons more prone to iron toxicity [67]. thus, nanoimaging forms an important part in the detection of newer pathways for neurodegeneration in pd. theranostics: gdnf can be delivered as a neuroprotective agent using microbubbles which are ultrasound responsive [68]. the use of such particles in simultaneous imaging and therapy can give rise to a theranostic approach in the management of parkinson’s disease. similar therapies have been developed for the treatment of cancers [69] and is very promising. a similar approach is the development of magnetic-field responsive nanoparticles that are capable of crossing anatomical barriers like the bbb and respond to an external magnetic field by releasing the encapsulated therapeutic payloads [70]. the field of theranostics is promising and yet to be well explored in the treatment of parkinson’s disease. detection of cell loss: in 1999, damier et al. described a protocol based on immunostaining a protein existing in afferent striatonigral fibers called calbindin d28k [71]. using this technique in post-mortem specimens, researchers could delineate 60% of all neurons containing dopamine present in the substantia nigra pars compacta inside the zone that has high calbindin (nigral matrix), and the remaining 40% was found to exist, packed together as the zones poor in calbindin (nigrosomes) in the form of invaginated pockets within the nigral matrix [71]. the group identified a mean reduction of more than half of the total dopamine-containing neurons in pd as compared to controls [71]. besides, the substantia nigra pars compacta showed that the degree of loss of dopamine-containing neurons correlated with the duration of disease [71]. thus, the pattern of cell loss in pd was found using this experimental tool. further advancements in the imaging of cells using quantum dots and gold nanoparticles can advance this imaging modality to track pd progression. monitoring clinical abnormalities one of the interesting applications in the diagnosis of pd is the detection of clinical abnormalities. arrays of stretchable strain sensors based on silicon nanomembranes that conform well to the skin have been proposed as wearable devices for detection of clinical abnormalities [72]. nanofabrication allows the manufacture of such sensors coupled with added features like presence of ultrathin serpentine interconnects to improve the precision of diagnosis [72]. exposure to tension and compression on a human wrist are used for detection of frequency of tremors and can be of use in the continuous monitoring of pd. admet & dmpk 3(3) (2015) 155-181 nanotechnology platforms in parkinson’s disease doi: 10.5599/admet.3.3.189 165 other interesting clinical abnormalities including change in handwriting have also been explored for the diagnosis of pd [73]. nanotechnology platforms for drug delivery active targeting by trojan horse approach nanotechnology based active targeting plays a crucial role in directing the neurotherapeutics at the target site through specific pathways. nanoparticles are surface functionalized/modified with specific ligands like peptides (insulin, insulin like growth factor), antibodies or antigens for receptors/transporters that are expressed on bbb therefore, facilitating enhanced bioavailability through receptor mediated endocytosis into the brain. the peptidomimetic antibodies decorated on nanoparticles can be used as molecular trojan horses to ferry large molecules like drugs and genes across the bbb [74,75]. previous studies showed expression of different receptors mainly low-density lipoprotein (ldl) and low density lipoprotein related protein-1 and 2 (lrp-1,2) transferrin, p-gp efflux transporter on bbb [76]. the receptor mediated endocytosis involves ligand-receptor interaction resulting in endocytosis at the luminal side of bbb followed by translocation through the endothelial cytoplasm to the abluminal side and finally exocytosis allowing release of payload at the target site [77]. the payloads may be in the form of a drug or a gene. gene therapy is becoming a popular approach to treat pd with some drawbacks in currently available therapies like poor permeability of the gene vector across the bbb after intravenous administration which limits the therapeutic efficacy. local intracerebral administration of the gene vector at the target site is often limited by lack of patient compliance due to high invasiveness and repetitive injection schedule. schlachetzki and colleagues discussed trans-vascular delivery, a novel approach for gene therapy to the brain using targeted nano-carriers [74]. this approach can be of two typesthe more invasive one involves bbb disruption, using the intracarotid infusion of vasoactive compounds or hyperosmolar solutions and the less invasive one involves the targeting of certain endogenous transport systems within the bbb. as an example of the latter, zhang et al. delivered tyrosine hydroxylase gene to the brain using a pegylated immunoliposome system with a targeting moiety in the form of a murine monoclonal antibody (mab) to the rat transferrin receptor [78]. receptors like transferrin play an important role in transporting iron to the brain through receptor mediated endocytosis. the technology functions as a trojan horse to ferry the gene across the bbb and demonstrated reversible normalization of tyrosine hydroxylase (th) activity in the striatum along with favorable behavioral effects in experimental models of parkinsonism in rats [78]. carroll and colleagues (2010) have reported the transferrin receptor antibody (ox26) conjugated plga nanoparticles loaded with antioxidant has significant potential against oxidative stress induced neurodegenerative diseases [79]. the in vitro study showed higher cellular uptake and decreased cell viability of rg2 rat glioma cells with targeted nanoparticles as compared to nontargeted nanoparticles through receptor mediated endocytosis. these findings suggest the potential of transferrin receptor in the treatment of other neurodegenerative diseases mediating the transport of drug across bbb [79]. lactoferrin is a cationic iron binding glycoprotein from one of the transferrin receptor family. previous studies on lactoferrin receptors reported that their expression on bbb facilitates the transport of lactoferrin to the brain [80]. ji and colleagues in 2006 have demonstrated the greater brain uptake of lactotransferrin than that of transferrin and ox26 (an anti-transferrin receptor antibody) through endocytosis mediated by receptors in an in vitro bbb study. [81].using this concept, hu et al. in 2009 demonstrated the improved brain targeting potential of lactoferrin conjugated plga nanoparticles as r. banerjee et al. admet & dmpk 3(3) (2015) 155-181 166 compared to unmodified nanoparticles [81]. here lactoferrin acts as the ferrying moiety across the bbb to allow the trojan horselike delivery of payloads [81]. this finding showed higher brain uptake (3-fold) of lactoferrin modified nanoparticles in mice over non-modified nanoparticles [81]. interestingly, in 2011, hu et al.; further demonstrated the successful brain delivery of urocortin (corticotrophin releasing hormone related peptide) loaded lactoferrin conjugated polyethylene glycol-polylactide-polyglycolide (peg-plga) nanoparticles [82]. in vivo pharmacokinetic study of lactoferrin conjugated nanoparticles showed 2.49times increase in auc than that of non-conjugated nanoparticles in 24h. the in vitro cellular study on immortalized mouse brain endothelial cell line (b.end3) revealed that the uptake of conjugated nanoparticles occurred through clathrin mediated endocytosis with minimal adverse effect on cell viability [82]. further the conjugated nanoparticles showed significant reduction of striatum lesions in 6-ohda (6hydroxy dopamine) induced rat pd model. together these findings revealed that lactoferrin is a promising targeting ligand to enhance targeting and accumulation of active to brain and lactoferrin decorated nanoparticles can help to deliver drugs more effectively in pd. huang et al; (2010) demonstrated long term, non-invasive targeted gene therapy in rotenone induced chronic pd model using lactoferrin functionalized nanoparticles encapsulated with human glial cell line derived neurotrophic factor gene (hgdnf). the study revealed that intravenously administered lactoferrin decorated nanoparticles showed successful gene delivery to the brain through higher and long lasting gdnf expression in brain in experimental animals than that of single injection [83]. the lrp-1 and 2 receptors interact with various other molecules like apolipoprotein-e (apoe) in addition to lactoferrin and melanotransferrin which have been widely studied for targeting against pd. wanger et al. in 2012 have reported the potential of low density lipoprotein receptor related protein (lrp-2) for specific and effective apolipoprotein mediated nanoparticulate transport and uptake by brain capillary endothelial cells [84]. a brief representation of the fate of actively targeted nanoparticles is provided in figure 3. figure 3. use of targeted nanoparticles for active targeting in drug delivery for parkinson’s disease admet & dmpk 3(3) (2015) 155-181 nanotechnology platforms in parkinson’s disease doi: 10.5599/admet.3.3.189 167 routes of administration to enhance drug delivery across bbb the presence of the bbb, limits the permeability of therapeutic agents to reach target sites in the cns, resulting in insufficient bioavailability as well as poor treatment outcomes against neurodegenerative disorders. several attempts have been made for effective and safe delivery of active into the cns. local drug delivery strategies include intrathecal, intraparenchymal and intracerebral injections directly at the target site for transporting the therapeutic active to the cns. direct drug delivery systems have potential to circumvent the bbb. pillay et al.(2009) studied the binary cross-linked alginate scaffold embedded polymeric (cellulose acetate phthalate) nanoparticles loaded with dopamine which after intracranial implantation showed enhanced and sustained dopamine delivery within csf in sprague–dawley rat model with reduced peripheral side effects as compared to orally administered l-dopa [85]. worly et al. (2008) developed acrylated 3′-sialyllactose (bioactive epitope recognized by various growth factor receptors like egfr) embedded within poly[n-(2-hydroxypropyl methacrylamide)] (phpma) hydrogels for intracranial implantation [86]. on intracranial implantation in 6-ohda pd model, the hydrogel served as a three dimensional substrate promoting neuronal tissue remodeling and angiogenesis with high affinity for receptors [86]. these finding demonstrated that polymeric nanoparticulate hydrogels can act as effective carriers for local delivery of drugs in pd. nevertheless, these techniques are limited due to their invasive nature [77,87]. recently, intranasal routes through olfactory or trigeminal pathways have been explored for transporting actives to the brain. intranasal administration has the potential to circumvent the bbb via intracellular or extracellular pathways. the intracellular pathways include endocytosis (adsorptive or receptor mediated) into olfactory sensory neurons followed by intraneuronal transport (transcytosis) into trigeminal nerves whereas, extracellular pathways include paracellular diffusion to the lamina propria through perineural or perivascular channels [88,89]. intranasal administrations offer benefits like rapid onset of action and reduced systemic toxicities for delivering the therapeutics to the brain. however, mucocilliary clearance limits the residence time when drug solutions are delivered intranasally. [90]. recently, wen et al. (2011) demonstrated that intranasal administration showed improved therapeutic efficacy of urocortin (corticotrophin releasing factor related peptide) when encapsulated in odorranalectin (bioadhesive from lectin family) functionalized peg-plga nanoparticles in hemiparkinsonian rats [91]. when administered intranasally, the odorranalectin functionalized nanoparticles showed higher brain uptake and enhanced neuroprotection in experimental animals through direct pathway. similarly, migliore et al. (2014) reported the neurotropic and neuroprotective effect of gdnf following intranasal administration of gdnf loaded cationic liposomes (made up of dioleoylphosphatidylcholine (dopc), cholesterol, and stearylamine) [92]. intranasal gdnf protected the dopamine neurons within substantia nigra from acute 6-ohdp induced neuronal toxic insult [92]. sharma et al. (2013) reported intranasal delivery of pluronic f-127 based thermoreversible gel containing levodopa encapsulated chitosan nanoparticles [93]. chitosan being a muchoadhesive polymer plays an important role in improving drug absorption through nasal mucosa and can lead to opening of tight junctions across bbb. the in vivo findings revealed that intranasal administration showed maximum recovery of levodopa within the brain with chitosan nanoparticles [93]. md and colleagues (2013) have studies the potential of intranasal route to enhance the brain delivery of bromocriptine (brc) loaded chitosan nanoparticles in haloperidol induced pd mice [94]. when administered intranasally the brc loaded nanoparticles showed pronounced reversion of catalepsy and akinesia induced by haloperidol. the biodistribution study revealed higher brain/blood ratio of brc loaded nanoparticles following intranasal administration than that of intravenous route at 0.5 hours, indicating direct nose to brain transport of brc circumventing the bbb. the gamma scintigraphy r. banerjee et al. admet & dmpk 3(3) (2015) 155-181 168 study revealed higher brain uptake and enhanced antioxidant potential of brc following intranasal administration [94]. a flowchart of the various routes and approaches to cross the bbb has been outlined in figure 4. figure 4. novel approaches for drug delivery across the bbb (the approaches are mentioned within boxes. those in red are invasive approaches and those in black are relatively less invasive approaches for drug delivery.) passive targeting (macrophage loaded) nanotechnology platforms possess the ability to accumulate in a specific organ without active targeting moieties on its surface due to certain inherent characteristics. for example, in cancers, small size of nanoparticles allow them to preferentially accumulate at tumor sites by the enhanced permeation and retention effect (epr effect) which leverages the leakiness of blood vessels in a tumor and absence of lymphatics to deliver the nanoparticles [95]. in parkinson’s disease, the bbb is an important barrier that needs to be crossed for the delivery of therapeutics. this can be achieved passively through macrophages. in pd there is a chemokine-gradient that gets induced through neuroinflammatory responses which allow concentration of macrophages in the brain as outlined in figure 5. hence, it is possible to use macrophages as drug carriers for the delivery of various therapeutic payloads [96]. batrakova et al. delivered catalase using bone-marrow-derived macrophage (bmm) systems to pd-affected brain regions in an animal model [97]. the self-assembled catalase/polymer complexes referred to as nanozymes that were 60-100 nm in size were taken up by the bmms and showed sustained release for about 24 hours as compared to immediate degradation of the free enzyme [97]. this had implications in decomposition of microglial hydrogen peroxide that could reduce oxidative stress in pd. it was found that the nanozymes loaded into bmms increased deposition of labeled enzyme in tissues in vivo in mptp-treated mice as compared to free enzymes. the development of such nanoparticles loaded into macrophages can help in increased delivery admet & dmpk 3(3) (2015) 155-181 nanotechnology platforms in parkinson’s disease doi: 10.5599/admet.3.3.189 169 of therapeutics to the brain reducing the need and the costs associated with active targeting of nanoparticles. figure 5. use of nanotechnology for passive targeting through macrophages for drug delivery in parkinson’s disease stimulus responsive nanoparticles nanoscale biomedical imaging for image guided therapy and diagnosis against neurodegenerative diseases is recently emerging. one of the widely studied multipurpose systems are magnetic nanoparticles (mnps) with mri contrast for applications like cellular and molecular imaging of various metastases, hyperthermia, drug delivery and tissue repair. in case of neurodegenerative diseases the presence of structural barrier likes bbb; restrict mnps access to the cns. therefore smart engineering of mnps with desired size and appropriate physicochemical properties is important to circumvent the bbb to gain access to the brain [70, 98]. recently qiao et al. (2012) developed target specific theranostics to facilitate the access of nanoparticles across bbb through endocytosis mediated by lactoferrin receptor. pegylated magnetic nanoparticles (fe3o4) surface modified with lactoferrin serve as a brain mri guided delivery probe[99]. the in vitro porcine bbb model showed that amphiphilic peg coating and lactoferrin conjugation facilitate the transport of mnps across bbb [99]. the vascular imaging after intravenous injection of targeted mnps confirm lactoferrin receptor mediated transcytosis with interaction between targeted nanoparticles and lactoferrin receptors expressed on microvascular endothelial cells [99]. recently hwang et al. (2009) reported ultrasound responsive stable perfluorocarbon nanobubbles (pns) loaded with dopamine receptor agonist (apomorphine) for targeting the brain. the studies showed the improved stability and higher plasma residence time of apomorphine loaded pns in vivo attributable to rigid cholesterol and phospholipid membrane. ultrasound trigger altered the release profiles on demand. erythrocyte hemolysis study demonstrated the safety profile (<10% hemolysis) of pns [100]. altogether these studies revealed the potential of stimulus sensitive nano-carriers for improving stability, administration frequency and on-demand release kinetics of therapeutics against pd. a brief representation of the use of nanobubbles as stimulus-responsive nanoparticles is provided in figure 6. r. banerjee et al. admet & dmpk 3(3) (2015) 155-181 170 figure 6. use of stimulus responsive nanoparticles for drug delivery in parkinson ’s disease with ultrasound (stimulus) responsive nanobubbles (nanoparticles) as an example neuroprotection neuroprotection is the ability for a treatment to prevent neuronal cell death by inhibition of the cascade resulting in cell dysfunction and eventually cell death. the relevance to pd lays in the fact that the pathogenesis of pd involves oxidative stress in substantia nigra through production of reactive oxygen species (ros) and hydrogen peroxide by extensive oxidative metabolism of dopamine. neurotrophic factors (like glial cell derived neurotrophic factor (gdnf)), fullerenes and nanoceria, show neuroprotection in pd [101,102]. the response from direct infusion of gdnf to the putamen had initial success but later could not demonstrate any significant benefits in controlled clinical trials due to development of antibodies and cerebellar degeneration [103]. this called for better modes of controlled delivery of gdnf. wang et al. described the use of microbubbles loaded with gdnf which released gdnf on low-frequency ultrasound stimulation in the treatment of pd-induced in rats [68]. gdnf levels augmented in the striatum following low frequency ultrasound stimulation after the infusion of microbubbles and resulted in favorable biochemical and behavioral changes suggesting neuroprotective effects [68]. thus, ultrasound responsive nanoparticles can be of use in the theranostics of pdtherapy by drug delivery and imaging due to the contrast provided by the microbubbles. in another approach, a number of non-viral gene delivery vectors are being developed that have reduced toxicities and have comparable efficiencies in transferring gdnf genes into the target cells. as has been mentioned earlier, lactoferrin-conjugated nanoparticles have been used successfully for this purpose in pd rat models [104]. similarly, angiopep with a sequence of tffyggsrgkrnnfkteeyc as the targeting ligand to mediate bbb transport and cellular internalization was conjugated to poly-l-lysine-based dendrimer nanoparticles (dendrigraft poly-l-lysine) via hydrophilic polyethylene glycol (peg) [105]. these nanoparticles constructs allowed effective gene expression both in vitro and in rat models of pd [105]. in addition to the delivery of neurotrophic factors, the use of fullerenes and nanoceria is of great interest. pretreatment with polyhydroxylated fullerene derivative c(60)(oh)(24) at ≥20 µm concentrations were used by cai et al. for the purposes of neuroprotection [106]. in an in vitro cellular model of parkinson's disease, significant protective effects as reflected by cell viability, mitochondrial function; levels of reactive oxygen species decreased and dna and proteins damage were reduced with polyhydroxylated fullerenes [106]. also, fullerenes in the form of carboxyfullerenes (c3) showed neuroprotection in macaque fascicularis monkeys [107]. the technology leveraging the antioxidant properties of fullerenes in the treatment of parkinson's disease is being commercialized and the leading companies involved are the admet & dmpk 3(3) (2015) 155-181 nanotechnology platforms in parkinson’s disease doi: 10.5599/admet.3.3.189 171 bronx project, inc. (tbp), tego biosciences corporation (tego) and luna innovations incorporated (luna) [108]. in addition to this, cerium oxide (ceo2) is being used for its capability of neutralizing and reducing harmful reactive oxygen species in a redox reaction [102]. due to the oxygen vacancies on the ceo2 surface and presence of both trivalent and tetravalent states of cerium ions, during the redox reaction the ions can flip-flop between the two oxidation states [102]. leveraging this scavenging role of cerium oxide, pinna et al. showed a protective role of nanoceria both in vitro on pc12 cells survival and dopamine metabolism in manganese-induced parkinson’s disease [102]. other approaches in management of parkinson’s disease based on nanotechnology neural prosthetic electrodes for deep brain stimulation bilateral deep brain stimulation (dbs) of the subthalamic nucleus (stn) through electrical stimulation using implanted neural prosthetic electrodes is a proven and fda-approved treatment modality for advanced parkinson's disease [109]. deep brain stimulation has been found to be better than medical therapy in producing improvements in motor symptoms [109,110]. however, the precision of dbs is an important issue that needs to be tweaked to address the ethical implications of altering brain circuits with this technique [111]. for this purpose, neural stimulation has employed a wide range of metals and metal alloys as materials for fabrication of electrodes in neural stimulation. these include noble metals (platinum), titanium oxide, activated iridium, tantalum oxide, and silicon [112]. however, these materials include limitations in the form of low charge injection limit allowing only large electrodes with low current density to be used [113], faradaic reactions (consisting of electron transfer between electrode and electrolyte resulting in either reduction or oxidation of a particular chemical species in a given electrolyte) leading to diffusion of toxic products or the corrosion of electrodes [114] as well as issues with biocompatibility with the electrode materials [114]. carbon nanotubesone of the most versatile nanotechnology options explored for neural prosthetic electrodes are carbon nanotubes (cnts) which have the unique combination of being strong (young’s modulus more than 1 tpa) as well as flexible. both of these are extremely important for fabrication of penetrating electrodes in neural prostheses [114]. besides, to serve as an electrode for neural stimulation it should be able to deal with the critical trade-off between high charge injection limit and absence of faradaic reactions. wang et al. used multiwalled carbon nanotube (cnt) pillars to fabricate microelectrodes [112] and successfully created a neural interface prototype device using cnt microarrays containing cnt microelectrodes. the neurons could grow and differentiate on the interface as well as could be repeatedly excited. the important parameters that were identified included the high charge injection limit offered by these functionalized hydrophilic cnt microelectrodes without faradic reactions [114]. this ensured the efficient and safe electrical stimulation of neurons and had several uses in diagnosis and therapy of neurological diseases like pd [114]. conducting polymersanother group of innovations in this field is the development of coatings on neural prosthetic implants using conductive polymers containing alternating double and single bonds in their structure [115]. polypyrrole (ppy), poly(ethylene dioxythiophene) (pedot), polyaniline and polyterthiophene are common examples of such polymers [115]. george et al. found polypyrrole implants used in microfabricated neural prosthetics allowed the growth of neural connections on all surfaces [116]. however, these implants showed tissue reactions in the brain in the form of gliosis [116] which was one of the major causes of failure of these electrode arrays within a few weeks after implantation [117]. to tackle this problem, wadhwa et al. proposed ppy based coatings on the electrodes r. banerjee et al. admet & dmpk 3(3) (2015) 155-181 172 with electrically controlled local delivery of dexamethasone (dex), a drug which is known to reduce tissue reactions and gliosis [117]. micromachining and one step electropolymerization was used for the deposition of a thin film (several nm) of ppy and the drug on the gold coated electrode surface for the fabrication of this smart electrode capable of responsiveness to electrical stimulus [117]. the added advantage of use of polypyrrole in such smart platforms is the fact that redox reactions cause the charging and discharging of these polymers and can lead to the movement of hydrated ions, drugs and growth factors in and out of the bulk of the polymer [117]. the drug loaded into the electrode was released in the buffer solution in the presence of an electrical trigger that was applied through cyclic voltammetry [117]. conductive polymer hydrogelsthe increased biocompatibility of these drug-eluting coatings on smart microelectrodes led to research towards increasing the drug loading and stimulus responsiveness of the electrodes. this lead to conducting polymers being incorporated inside hydrogel-support matrices, a class of materials called conductive polymer hydrogels (cph) [118]. such a matrix can allow more drug incorporation along with faster release [117]. these properties can be further enhanced through the increase in surface area by creating nano-fibrous morphology employing nano-templating techniques [117]. such devices can serve as mechanical buffers in the interface between the hard electrodes and the soft brain tissue improving contact with viable neurons while increasing biocompatibility and limiting protein adsorption to the neural prosthetic electrodes [117,118]. according to a study by nguyen et al., the incorporation of a mechanically-compliant implant in the brain parenchyma can reduce the inflammatory response in comparison to stiffer systems [119]. thus, cph coated electrodes are biocompatible by themselves and can be additionally used to deliver drugs and growth factors locally. . kim et al. used ppy grown electrochemically on the scaffolds made of hydrogels followed by deposition on the surface of neural prosthetic devices [120]. the coating was found to reduce impedance and improve electrical properties of the electrodes [120]. however, hydrogels have the inherent property of a high swelling ratio and its swelling in situ can lead to increase in the distance from the brain surface and the microelectrode reducing the efficiency of deep brain stimulation. to deal with this problem, bio-inspired hybrid microelectrodes were fabricated by de faveri et al. in which they used highly biocompatible fibrin coatings [121]. the fibrin coatings reduced host reaction within 7 days of implantation and allowed growth of both neurons and astrocytes [121]. but most importantly, the fibrin coating displayed a smaller swelling than other types of hydrogels and allowed the control of coating thickness [121]. thus, the problem of hydrogel swelling could be tackled with suitable bio-inspired coatings. thus, nanotechnology is being increasingly employed for the development of better neural prosthetic electrodes for deep brain stimulation. further advances in the fabrication and stimulus responsiveness of these technologies can help incorporate dbs as a safer treatment modality in the management of parkinson’s disease. nanotechnology in neurosurgery after the introduction of levodopa half a century ago, neurosurgery was abandoned for the management of parkinson’s disease. however, the past decade saw the resurgence of the surgical treatment of pd using pallidotomy and thalamotomy [103]. this was due to the inherent drawbacks of drug therapy in which the prolonged administration of drugs lead to drug-induced motor symptoms [103]. with the advent of the age of nanotechnology, there have been great strides in the precision in neurosurgery. laser axotomyone of the most intriguing techniques in the field of nanotechnology is the use of admet & dmpk 3(3) (2015) 155-181 nanotechnology platforms in parkinson’s disease doi: 10.5599/admet.3.3.189 173 femtosecond laser axotomy which is being developed as a neurosurgical tool with 100% efficiency, precision in the range of nanometers and high speed. the most widely cited example of this precise technique is by yanik et al. (2011) where femtosecond laser axotomy was used to cut single axons inside the roundworm c. elegans by using focused nano-scale near-infrared laser pulse energies (10–40 nanojoules) of 200-femtoseconds [122]. this resulted in the vaporization of femtolitres of axon volumes and subsequent cutting of the axon processes [122]. besides, the technique was robust allowing the use of confocal fluorescence imaging and laser scanning brightfield for real-time imaging from the first incision to the end [122].the surgical technique has improved understanding of a number of processes in nerve cell injury [123] and has also paved way for an all new concept of nanomanipulation. vibrating micropipetteanother widely popular technology is the use of a rapidly vibrating (100 hz) micropipette (tip diameter in nanometers) which when dragged across the dendrite of a selected neuron can help in dissection of apical dendrites [124]. this provided a simple system for cutting dendrites from single neurons without damaging the cell [124]. nanotweezersthe word nanotweezers is being widely used with different meanings but with the same purpose of precise manipulation in the nanometer range. the nanotweezers may be of nanoscale electromechanical systems based on carbon nanotubes fabricated on pulled glass micropipettes as described by kim et al. [120]. voltages that are applied to these electrodes could open or close the free ends of the nanotubes allowing grabbing and manipulation of submicron clusters and nanowires [120]. the mechanical capabilities of such nanotweezers can be put to use for grabbing and manipulating neurons as well. a rather higher level of precision in manipulation is by the use photoswitchable crosslinkers which in response to light can manipulate the normal physiological signal or induce structural changes in peptides and ion channels allowing molecular level manipulation of cells [125]. the use of light is being increasingly used for the study of disease circuitry in parkinson’s disease especially through optogenetics which is a novel technique of gene delivery through encoding of proteins capable of instilling light sensitivity to neurons [126]. these technologies once translated can improve the precision of neurosurgery to molecular ranges. nanotechnology and the use of smart nanoparticles can greatly improve the precision of such surgical devices from cutting neuronal processes to modifying ion channels using light. as these devices are gradually becoming popular in the lab, their translation to bedside is possible in the near future. the essential precision that these nanotechnological tools add, can lead to a safer neurosurgical option for patients with pd as with many other neurological diseases. tissue engineering approaches finding the ideal platform the dopaminergic neurons in the substantia nigra responsible for cortical and thalamic regulation are lost in pd leading to motor dysfunction (tremor, rigidity, and bradykinesia) and non-motor symptoms (depression, anxiety) [103]. the transplantation of fetal tissue in the form of human fetal ventral mesencephalon was one of the most promising treatments in the 1980s and stimulated a number of clinical trials with varied protocols and endpoints [127,128]. there were reports of clinical improvement and two decades later, after sufficient evidence of safety and efficacy of these treatments, the national institutes of health, usa started two clinical trials in the early 2000s [127]. the trials found the therapy to be of little benefit to patients and caused some of the patients to become worse with symptoms like graftinduced dyskinesias (gid) [127,129]. however, a decade later, the debate is still on regarding the efficiency of stem-cell grafting [128,129]. besides the concerns in the clinical-trial design and ethical issues, the r. banerjee et al. admet & dmpk 3(3) (2015) 155-181 174 technology needed to be safe [127,128]. these concerns are more pronounced with embryonic stem cells (escs) because of their ability to differentiate into a number of cell types [127]. a potential solution is the use of patient derived cells and then induced pluripotent stem cells (ipscs) [127]. however, the concern still lies in the fact that the patient’s own cells may be susceptible to the pd pathology [127]. thus, it is the trade-off between optimal ability to proliferate while dodging the pathogenesis of pd that forms the basis of development of a safe treatment for pd. direct infusion of growth factors like glial cell–derived neurotrophic factor (gdnf) to the putamen in patients with pd has been found to be favorable and are being seen as prospective agents for neuroprotection [103]. however, these approaches are still to be verified as clinical trials reflect conflicting evidence [103]. tissue engineering tries to address these issues using stem cells and growth factors in suitable scaffolds for the controlled regeneration of various tissues. it is being increasingly used for neural stem cell engineering for parkinson’s disease. nanoparticles for delivery of versatile payloads: one of the approaches in tissue engineering for pd is the use of small molecules as triggers to allow the differentiation of stem cells into the neurons. santos et al. used retinoic acid-loaded polymeric nanoparticles to induce such differentiation in subventricular zone neural stem cells [130]. the nanoparticles leveraged the electrostatic interaction of polyethylenimine (polycation) with retinoic acid and dextran sulfate (polyanion) [130]. the intracellular delivery of retinoic acid using nanoparticles increased the differentiation of stem cells into neurons as was confirmed using gene expression profiles. such technologies can be of great use if nanoparticles containing retinoic acid are manufactured for targeting stem cell sequences to control in time and space the differentiation of stem cells [130]. scaffold design: in addition to using small molecules for the differentiation into a particular lineage, there can be changes made in the substrate too allowing nanoknitting of neurons on scaffolds [131]. ni et al. developed a self-assembling peptide nanofiber scaffold (sapns) 6–10 nm in diameter to enhance differentiation of escs as well as ipscs into dopaminergic neurons in a 3-dimensional culture [132]. this was confirmed using gene expression, maturation, immunocytochemistry and dopamine release all of which proved genetically and functionally differentiated dopaminergic neurons in the scaffold [132]. hence, these materials have the potential to be used for neuroprostheses. in addition to this, materials used in neural prosthetic electrodes have also been used in tissue engineering. conductive polymer hydrogels (cph) have the advantage of allowing cell immobilization while increasing the surface area and electrical properties necessary for nerve growth and regeneration [117,118,120]. thus, nanotechnology-based platforms are increasingly being used to tackle the challenges of stem-cell delivery. the present day platforms based on nanotechnology would allow a more precise and accurate delivery of stem cells for the management of parkinson’s disease. a brief summary of the use of nanotechnology in these therapies is represented in figure 7. admet & dmpk 3(3) (2015) 155-181 nanotechnology platforms in parkinson’s disease doi: 10.5599/admet.3.3.189 175 figure 7. nanotechnology-based platforms for the management of parkinson’s disease translation, future prospects and concerns the increased awareness of pd and the accelerated research towards its management has lead to the infusion of a number of technologies into the scenario. nanotechnology is a novel and versatile addition to the fight against pd. there have been efforts in accelerating research in the identification of novel biomarkers and strategies for the cure for pd. this has led to a lot of research reach from labs to the bedside. nanotechnology has the ability to develop smarter sensors for the diagnosis of pd using these novel biomarkers and imaging modalities. it is also being used for the targeted treatment of pd, in neuroprotection and in stem cell therapies. multinational technology companies are also entering the arena of nanotechnology to treat parkinson’s disease. google inc. has recently filed a patent for a wearable device in the form of a watch or bracelet that can direct functionalized nanoparticles (like magnetic nanoparticles through the magnetic field generated by the device) into the subsurface vasculature present close to the device [133]. the patent claims that there is a potential for the use of such devices in the treatment of pd by targeting certain proteins [133]. besides, the clinical translation of research has lead to the development of a number of commercial entities. the bronx project, inc. (tbp), tego biosciences corporation (tego) and luna innovations incorporated (luna) [108] are involved in the development of fullerenes for neuroprotection in pd. parkinson’s pen (by manus neurodynamica ltd.) is another patented technology that uses the changes in handwriting in combination with a sensor device on the pen to analyze the force exerted by each of the fingers, and the changes in handwriting [73]. such pressure sensors and detectors can be of use to detect pd. parkinson’s pen itself in its initial testing was able to distinguish 10 people with pd from ageand sexmatched controls with 90% sensitivity and 80 % specificity [134]. a clinical trial with a larger sample size is underway to confirm the sensitivity and specificity of the promising device [134]. this technology will have applications in diagnostics, pre-symptomatic screening, disease monitoring, treatment response and rehabilitation. the company is currently conducting clinical trials to this end. thus, there is a lot of translation that is taking place from diagnosis to rehabilitation of pd patients. r. banerjee et al. admet & dmpk 3(3) (2015) 155-181 176 another area that is developing a lot in pd is the use of stem cells. nanotechnology affords a targeted and controlled delivery and differentiation of stem cells for the therapy of pd. however, one of the biggest concerns in the present day is the development of treatments of pd using unproven cell-based treatments which pose serious threats to research translation [128]. these therapies pose increased risk of tumor formation [126]. such therapies continue to be performed without intensive testing in vitro and in vivo and pose a serious risk to jeopardize the spark of hope that is gradually infused in the sphere due to the advent of tissue engineering and nanotechnology [128]. in addition to this, there are ethical concerns raised regarding neurology and nanotechnology and their potential impact for which there is a need for a multidisciplinary bioethical enquiry [135]. also, there are conflicting reports regarding toxicity of certain nanoparticles which cross the blood brain barrier [136-138]. this has called for further evaluation of nanoparticles for surface properties, size and constituents for finetuning the therapeutic effects while avoiding toxic effects [136]. for drug delivery formulations, it is imperative to increase the therapeutic window which is the range of dosage that can effectively treat disease while staying in the safety range and without inducing adverse effects [136]. comprehensive testing for efficacy and toxicity must be performed before the introduction of nanotechnology-based systems into the market [136]. according to the royal society/royal academy of engineering report (2004), carbon nanotubes should be comprehensively tested as “new substances” for their potential toxicities under uk and european health, safety and environment legislation before they are introduced in the market [57]. however, each nanoparticle is unique and it is necessary to understand the specific issues associated with each type of nanoparticle for evaluation of toxicity issues. use of green chemistry, degradable materials, and regulatory approvals for nanotechnology-based products are the positive signals towards an optimistic future with nanotechnology, the effects of which can be of potential benefit to patients with pd. conclusions parkinson’s disease has a huge impact on the lives of the patients with tremendous socioeconomic effects. nanotechnology can serve as a viable alternative to the present day diagnosis and treatment of pd by providing novel, affordable and accessible devices. nanotechnological platforms are being increasingly developed with applications in the management of pd. the optimism in the translation of these technologies from labs to bedside can be seen with increasing interest by grant agencies, companies and physicians towards their development. the tremendous potential to tweak nanoparticles in the nanoscale can prove to be of great help in the fight against parkinson’s disease. references [1] d. longo, a. fauci, d. kasper, s. hauser, j. jameson, j. loscalzo, harrison's principles of internal medicine, mcgraw-hill education, 2011. [2] t. pringsheim, n. jette, a. frolkis, t.d.l. steeves, mov. disord. 29 (2014) 1583-1590. [3] world health organization, neurological disorders : public health challenges, world health organization, geneva, switzerland, 2006. [4] s.l. kowal, t.m. dall, r. chakrabarti, m.v. storm, a. jain, mov. disord. 28 (2013) 311-318. [5] m. ragothaman, s.t. govindappa, r. rattihalli, d.k. subbakrishna, u.b. muthane, mov disord 21 (2006) 1755-1758. [6] c.g. goetz, mov. disord. 1 (1986) 27-32. admet & dmpk 3(3) (2015) 155-181 nanotechnology platforms in parkinson’s disease doi: 10.5599/admet.3.3.189 177 [7] d. berg, a.e. lang, r.b. postuma, w. maetzler, g. deuschl, t. gasser, a. siderowf, a.h. schapira, w. oertel, j.a. obeso, c.w. olanow, w. poewe, m. stern, lancet neurol. 12 (2013) 514-524. [8] l.m. de lau, m.m. breteler, lancet neurol. 5 (2006) 525-535. [9] s.k. van den eeden, c.m. tanner, a.l. bernstein, r.d. fross, a. leimpeter, d.a. bloch, l.m. nelson, am. j. epidemiol. 157 (2003) 1015-1022. [10] a. benazzouz, o. mamad, p. abedi, r. bouali-benazzouz, j. chetrit, front. aging neurosci. 6 (2014) 87. [11] a.m. bonnet, j.l. houeto, biomed. pharmacother. 53 (1999) 117-121. [12] m. storch, a. storch, m.a. collins, neurotoxic factors in parkinson’s disease, related disorders, springer us, 2001. [13] d.c. german, k.f. manaye, c.l. white, 3rd, d.j. woodward, d.d. mcintire, w.k. smith, r.n. kalaria, d.m. mann, ann. neurol. 32 (1992) 667-676. [14] h.m. van praag, s. de haan, psychiatry res. 1 (1979) 219-224. [15] j.t. greenamyre, c.f. o'brien, arch. neurol. 48 (1991) 977-981. [16] s. haber, s.j. watson, neuroscience 14 (1985) 1011-1024. [17] c.w. olanow, w.g. tatton, annu. rev. neurosci. 22 (1999) 123-144. [18] z.i. alam, a. jenner, s.e. daniel, a.j. lees, n. cairns, c.d. marsden, p. jenner, b. halliwell, j. neurochem. 69 (1997) 1196-1203. [19] a.h.v. schapira, j.m. cooper, d. dexter, j.b. clark, p. jenner, c.d. marsden, j. neurochem. 54 (1990) 823-827. [20] d.j. moore, a.b. west, v.l. dawson, t.m. dawson, annu. rev. neurosci. 28 (2005) 57-87. [21] z.h. sheng, q. cai, nat. rev. neurosci. 13 (2012) 77-93. [22] a. thaler, e. ash, z. gan-or, a. orr-urtreger, n. giladi, j. neural. transm. 116 (2009) 1473-1482. [23] a.b. singleton, m.j. farrer, v. bonifati, mov. disord. 28 (2013) 14-23. [24] l.j. ozelius, g. senthil, r. saunders-pullman, e. ohmann, a. deligtisch, m. tagliati, a.l. hunt, c. klein, b. henick, s.m. hailpern, r.b. lipton, j. soto-valencia, n. risch, s.b. bressman, n. engl. j. med. 354 (2006) 424-425. [25] h. asada, n. ip, l. pan, n. razack, m. parfitt, r.j. plunkett, j. neurosci. res. 40 (1995) 22-30. [26] p.m. antony, n.j. diederich, r. kruger, r. balling, febs j. 280 (2013) 5981-5993. [27] y. chu, h. dodiya, p. aebischer, c.w. olanow, j.h. kordower, neurobiol. dis. 35 (2009) 385-398. [28] s.l. rhodes, b. ritz, neurobiol. dis. 32 (2008) 183-195. [29] z. lv, h. jiang, h. xu, n. song, j. xie, j. neural. transm. 118 (2011) 361-369. [30] s.a. schneider, c. paisan-ruiz, n.p. quinn, a.j. lees, h. houlden, j. hardy, k.p. bhatia, mov. disord. 25 (2010) 979-984. [31] a. mcneill, curr. drug targets 13 (2012) 1204-1206. [32] e. tolosa, g. wenning, w. poewe, lancet neurol. 5 (2006) 75-86. [33] j. booij, j.d. speelman, m.w. horstink, e.c. wolters, eur. j. nucl. med. 28 (2001) 266-272. [34] b.s. connolly, a.e. lang, jama 311 (2014) 1670-1683. [35] t. nagatsua, m. sawadab, parkinsonism relat. disord. 15 suppl 1 (2009) s3-8. [36] h.m. ruottinen, u.k. rinne, j. neurol. 245 (1998) p25-34. [37] a.h. schapira, f. stocchi, r. borgohain, m. onofrj, m. bhatt, p. lorenzana, v. lucini, r. giuliani, r. anand, eur. j. neurol. 20 (2013) 271-280. [38] f. stocchi, r. borgohain, m. onofrj, a.h. schapira, m. bhatt, v. lucini, r. giuliani, r. anand, mov. disord. 27 (2012) 106-112. [39] w. bara-jimenez, a. sherzai, t. dimitrova, a. favit, f. bibbiani, m. gillespie, m. morris, m. mouradian, t. chase, neurology 61 (2003) 293-296. r. banerjee et al. admet & dmpk 3(3) (2015) 155-181 178 [40] f. stocchi, o. rascol, a. destee, n. hattori, r.a. hauser, a.e. lang, w. poewe, m. stacy, e. tolosa, h. gao, j. nagel, m. merschhemke, a. graf, c. kenney, c. trenkwalder, mov. disord. 28 (2013) 18381846. [41] p. jenner, ann. neurol. 53 suppl 3 (2003) s26-36; discussion s36-38. [42] l.f. lin, d.h. doherty, j.d. lile, s. bektesh, f. collins, science 260 (1993) 1130-1132. [43] a.e. lang, j.w. langston, a.j. stoessl, m. brodsky, d.j. brooks, v. dhawan, w.j. elias, a.m. lozano, e. moro, j.g. nutt, m. stacy, d. turner, g.f. wooten, lancet neurol. 5 (2006) 200-202. [44] n.d. doolittle, l.e. abrey, w.a. bleyer, s. brem, t.p. davis, p. dore-duffy, l.r. drewes, w.a. hall, j.m. hoffman, a. korfel, r. martuza, l.l. muldoon, d. peereboom, d.r. peterson, s.d. rabkin, q. smith, g.h. stevens, e.a. neuwelt, clin. cancer res. 11 (2005) 421-428. [45] y. persidsky, s.h. ramirez, j. haorah, g.d. kanmogne, j. neuroimmune pharmacol. 1 (2006) 223236. [46] w.h. oldendorf, m.e. cornford, w.j. brown, ann. neurol. 1 (1977) 409-417. [47] w. loscher, h. potschka, prog. neurobiol. 76 (2005) 22-76. [48] w.m. pardridge, mol. biotechnol. 30 (2005) 57-70. [49] n.j. abbott, a.a. patabendige, d.e. dolman, s.r. yusof, d.j. begley, neurobiol. dis. 37 (2010) 13-25. [50] r. kortekaas, k.l. leenders, j.c. van oostrom, w. vaalburg, j. bart, a.t. willemsen, n.h. hendrikse, ann. neurol. 57 (2005) 176-179. [51] f.i. tarazi, z.t. sahli, m. wolny, s.a. mousa, pharmacol. ther. 144 (2014) 123-133. [52] w. schuepbach, j. rau, k. knudsen, j. volkmann, p. krack, l. timmermann, t. hälbig, h. hesekamp, s. navarro, n. meier, n. eng. j. med. 368 (2013) 610-622. [53] e. coley, r. farhadi, s. lewis, i.r. whittle, br. j. neurosurg. 23 (2009) 179-183. [54] k.a. sillay, p.s. larson, p.a. starr, neurosurgery 62 (2008) 360-6; discussion 366-367. [55] e. martin, d. jeanmonod, a. morel, e. zadicario, b. werner, ann. neurol. 66 (2009) 858-861. [56] c. ohye, y. higuchi, t. shibazaki, t. hashimoto, t. koyama, t. hirai, s. matsuda, t. serizawa, t. hori, m. hayashi, t. ochiai, h. samura, k. yamashiro, neurosurgery 70 (2012) 526-535; discussion 535536. [57] royal society, royal academy of engineering, nanoscience and nanotechnologies: opportunities and uncertainties, royal society, 2004. [58] w.g. meissner, m. frasier, t. gasser, c.g. goetz, a. lozano, p. piccini, j.a. obeso, o. rascol, a. schapira, v. voon, d.m. weiner, f. tison, e. bezard, nat. rev. drug discov. 10 (2011) 377-393. [59] s. sharma, c.s. moon, a. khogali, a. haidous, a. chabenne, c. ojo, m. jelebinkov, y. kurdi, m. ebadi, neurochem. inter. 63 (2013) 201-229. [60] j. tashkhourian, m.r.h. nezhad, j. khodavesi, s. javadi, j. electroanal. chem. 633 (2009) 85-91. [61] h.y. yue, s. huang, j. chang, c. heo, f. yao, s. adhikari, f. gunes, l.c. liu, t.h. lee, e.s. oh, b. li, j.j. zhang, t.q. huy, n.v. luan, y.h. lee, acs nano 8 (2014) 1639-1646. [62] k. kurzatkowska, e. dolusic, w. dehaen, k. sieroń-stołtny, a. sieroń, h. radecka, anal. chem. 81 (2009) 7397-7405. [63] w. ma, l.-x. qin, f.-t. liu, z. gu, j. wang, z.g. pan, t.d. james, y.-t. long, sci. rep. 3 (2013) 1537. [64] y. an, l. tang, x. jiang, h. chen, m. yang, l. jin, s. zhang, c. wang, w. zhang, chem. eur. j. 16 (2010) 14439-14446. [65] y.y. broza, h. haick, nanomedicine (lond) 8 (2013) 785-806. [66] u. tisch, i. schlesinger, r. ionescu, m. nassar, n. axelrod, d. robertman, y. tessler, f. azar, a. marmur, j. aharon-peretz, h. haick, nanomedicine (lond) 8 (2013) 43-56. [67] r. ortega, p. cloetens, g. devès, a. carmona, s. bohic, plos one 2 (2007) e925. [68] x. wang, g. cui, x. yang, z. zhang, h. shi, j. zu, f. hua, x. shen, brain res. bull. 103 (2014) 60-65. admet & dmpk 3(3) (2015) 155-181 nanotechnology platforms in parkinson’s disease doi: 10.5599/admet.3.3.189 179 [69] b. geers, i. lentacker, n.n. sanders, j. demeester, s. meairs, s.c. de smedt, j.control. release 152 (2011) 249-256. [70] o. veiseh, j.w. gunn, m. zhang, adv. drug deliv. rev. 62 (2010) 284-304. [71] p. damier, e.c. hirsch, y. agid, a.m. graybiel, brain, 122(pt 8) (1999) 1437-1448. [72] d. son, j. lee, s. qiao, r. ghaffari, j. kim, j.e. lee, c. song, s.j. kim, d.j. lee, s.w. jun, s. yang, m. park, j. shin, k. do, m. lee, k. kang, c.s. hwang, n. lu, t. hyeon, d.h. kim, nat. nanotechnol. 9 (2014) 397-404. [73] r.c. zietsma, (manus neurodynamica ltd), wo 2011141734 a1 (2011). [74] f. schlachetzki, y. zhang, r.j. boado, w.m. pardridge, neurology 62 (2004) 1275-1281. [75] w.m. pardridge, curr. opin. pharmacol. 6 (2006) 494-500. [76] l. costantino, d. boraschi, drug discovery today 17 (2012) 367-378. [77] w.m. pardridge, j. cereb. blood flow met. 17 (1997) 713-731. [78] y. zhang, f. calon, c. zhu, r.j. boado, w.m. pardridge, hum. gene ther. 14 (2003) 1-12. [79] r.t. carroll, d. bhatia, w. geldenhuys, r. bhatia, n. miladore, a. bishayee, v. sutariya, j. drug target 18 (2010) 665-674. [80] y.a. suzuki, v. lopez, b. lonnerdal, cell mol. life sci. 62 (2005) 2560-2575. [81] b. ji, j. maeda, m. higuchi, k. inoue, h. akita, h. harashima, t. suhara, life. sci. 78 (2006) 851-855. [82] k. hu, y. shi, w. jiang, j. han, s. huang, x. jiang, int. j. pharm. 415 (2011) 273-283. [83] r. huang, w. ke, y. liu, d. wu, l. feng, c. jiang, y. pei, j. neurol. sci. 290 (2010) 123-130. [84] s. wagner, a. zensi, s.l. wien, s.e. tschickardt, w. maier, t. vogel, f. worek, c.u. pietrzik, j. kreuter, h. von briesen, plos one 7 (2012) e32568. [85] s. pillay, v. pillay, y.e. choonara, d. naidoo, r.a. khan, l.c. du toit, v.m. ndesendo, g. modi, m.p. danckwerts, s.e. iyuke, int. j. pharm. 382 (2009) 277-290. [86] s. woerly, s. fort, i. pignot-paintrand, c. cottet, c. carcenac, m. savasta, biomacromolecules 9 (2008) 2329-2337. [87] j.g. nutt, k.j. burchiel, c.l. comella, j. jankovic, a.e. lang, e.r. laws, jr., a.m. lozano, r.d. penn, r.k. simpson, jr., m. stacy, g.f. wooten, neurology 60 (2003) 69-73. [88] j.j. lochhead, r.g. thorne, adv. drug deliv. rev. 64 (2012) 614-628. [89] h.l. wong, x.y. wu, r. bendayan, adv. drug deliv. rev. 64 (2012) 686-700. [90] t.k. vyas, a. shahiwala, s. marathe, a. misra, curr. drug deliv. 2 (2005) 165-175. [91] z. wen, z. yan, k. hu, z. pang, x. cheng, l. guo, q. zhang, x. jiang, l. fang, r. lai, j. control. release 151 (2011) 131-138. [92] m.m. migliore, r. ortiz, s. dye, r.b. campbell, m.m. amiji, b.l. waszczak, neuroscience 274 (2014) 11-23. [93] s. sharma, s. lohan, r.s. murthy, drug dev. ind. pharm. 40 (2014) 869-878. [94] s. md, r.a. khan, g. mustafa, k. chuttani, s. baboota, j.k. sahni, j. ali, eur. j. pharm. sci. 48 (2013) 393-405. [95] r. banerjee, nanomedicine 6 (2011) 1657-1660. [96] i. kadiu, j.g. glanzer, j. kipnis, h.e. gendelman, m.p. thomas, neurotox. res. 8 (2005) 25-50. [97] e.v. batrakova, s. li, a.d. reynolds, r.l. mosley, t.k. bronich, a.v. kabanov, h.e. gendelman, bioconjug. chem. 18 (2007) 1498-1506. [98] o.c. farokhzad, r. langer, adv. drug deliv. rev. 58 (2006) 1456-1459. [99] r. qiao, q. jia, s. huwel, r. xia, t. liu, f. gao, h.j. galla, m. gao, acs nano 6 (2012) 3304-3310. [100] t.l. hwang, y.k. lin, c.h. chi, t.h. huang, j.y. fang, j. pharm. sci. 98 (2009) 3735-3747. [101] m. srikanth, j.a. kessler, nat. rev. neur. 8 (2012) 307-318. r. banerjee et al. admet & dmpk 3(3) (2015) 155-181 180 [102] a. pinna, l. malfatti, g. galleri, r. manetti, s. cossu, g. rocchitta, r. migheli, p.a. serra, p. innocenzi, rsc advances 5 (2015) 20432-20439. [103] c. wiener, a. fauci, e. braunwald, d. kasper, s. hauser, d. longo, j. jameson, j. loscalzo, harrison's principles of internal medicine, self-assessment and board review, mcgraw-hill education, 2008. [104] r. huang, w. ke, y. liu, d. wu, l. feng, c. jiang, y. pei, j. neur. sci. 290 (2010) 123-130. [105] r. huang, h. ma, y. guo, s. liu, y. kuang, k. shao, j. li, y. liu, l. han, s. huang, s. an, l. ye, j. lou, c. jiang, pharm. res. 30 (2013) 2549-2559. [106] x. cai, h. jia, z. liu, b. hou, c. luo, z. feng, w. li, j. liu, j. neurosci. res. 86 (2008) 3622-3634. [107] l.l. dugan, l. tian, k.l. quick, j.i. hardt, m. karimi, c. brown, s. loftin, h. flores, s.m. moerlein, j. polich, s.d. tabbal, j.w. mink, j.s. perlmutter, ann. neurol. 76 (2014) 393-402. [108] z. zhou, pharmaceutics 5 (2013) 525-541. [109] g. deuschl, c. schade-brittinger, p. krack, j. volkmann, h. schäfer, k. bötzel, c. daniels, a. deutschländer, u. dillmann, w. eisner, d. gruber, w. hamel, j. herzog, r. hilker, s. klebe, m. kloß, j. koy, m. krause, a. kupsch, d. lorenz, s. lorenzl, h.m. mehdorn, j.r. moringlane, w. oertel, m.o. pinsker, h. reichmann, a. reuß, g.h. schneider, a. schnitzler, u. steude, v. sturm, l. timmermann, v. tronnier, t. trottenberg, l. wojtecki, e. wolf, w. poewe, j. voges, n. eng. j. med. 355 (2006) 896-908. [110] f.m. weaver, k. follett, m. stern, k. hur, c. harris, w.j. marks jr, j. rothlind, o. sagher, d. reda, c.s. moy, r. pahwa, k. burchiel, p. hogarth, e.c. lai, j.e. duda, k. holloway, a. samii, s. horn, j. bronstein, g. stoner, j. heemskerk, g.d. huang, jama j. 301 (2009) 63-73. [111] m. christen, m. bittlinger, h. walter, p. brugger, s. müller, ajob neuroscience 3 (2012) 37-43. [112] k. wang, h.a. fishman, h. dai, j.s. harris, nano lett. 6 (2006) 2043-2048. [113] d.b. mccreery, w.f. agnew, t.g. yuen, l.a. bullara, ann. biomed. eng. 16 (1988) 463-481. [114] d.r. merrill, m. bikson, j.g.r. jefferys, j. neurosci. met. 141 (2005) 171-198. [115] u.a. aregueta-robles, a.j. woolley, l.a. poole-warren, n.h. lovell, r.a. green, front. neuroeng. 7 (2014) 15. [116] p.m. george, a.w. lyckman, d.a. lavan, a. hegde, y. leung, r. avasare, c. testa, p.m. alexander, r. langer, m. sur, biomaterials 26 (2005) 3511-3519. [117] r. wadhwa, c.f. lagenaur, x.t. cui, j. control. release 110 (2006) 531-541. [118] y. zhao, b. liu, l. pan, g. yu, energ. environ. sci. 6 (2013) 2856-2870. [119] j.k. nguyen, d.j. park, j.l. skousen, a.e. hess-dunning, d.j. tyler, s.j. rowan, c. weder, j.r. capadona, j. neural. eng. 11 (2014) 056014. [120] d.h. kim, m. abidian, d.c. martin, j. biomed. mater. res. a 71 (2004) 577-585. [121] s. de faveri, e. maggiolini, e. miele, f. de angelis, f. cesca, f. benfenati, l. fadiga, front. neuroeng. 7 (2014) 7. [122] m.f. yanik, h. cinar, h.n. cinar, a.d. chisholm, y. jin, a. ben-yakar, nature 432 (2004) 822-822. [123] t. wu, s. mohanty, v. gomez-godinez, l.z. shi, l.-h. liaw, j. miotke, r.l. meyer, m.w. berns, j. roy. soc. interface 9 (2012) 535-547. [124] e.d. kirson, y. yaari, j. neurosci. met 98 (2000) 119-122. [125] p. gorostiza, e. isacoff, mol biosys. 3 (2007) 686-704. [126] v. gradinaru, m. mogri, k.r. thompson, j.m. henderson, k. deisseroth, science 324 (2009) 354359. [127] p.c. buttery, r.a. barker, j. comp. neur. 522 (2014) 2802-2816. [128] r.a. barker, j. barrett, s.l. mason, a. björklund, lancet neur. 12 (2013) 84-91. [129] c.w. olanow, c.g. goetz, j.h. kordower, a.j. stoessl, v. sossi, m.f. brin, k.m. shannon, g.m. nauert, d.p. perl, j. godbold, t.b. freeman, ann. neurol. 54 (2003) 403-414. admet & dmpk 3(3) (2015) 155-181 nanotechnology platforms in parkinson’s disease doi: 10.5599/admet.3.3.189 181 [130] t. santos, r. ferreira, j. maia, f. agasse, s. xapelli, l. cortes, j. bragança, j.o. malva, l. ferreira, l. bernardino, acs nano 6 (2012) 10463-10474. [131] r.g. ellis-behnke, y.-x. liang, s.-w. you, d.k.c. tay, s. zhang, k.-f. so, g.e. schneider, proceed. nat. acad. sci. usa 103 (2006) 5054-5059. [132] n. ni, y. hu, h. ren, c. luo, p. li, j.-b. wan, h. su, plos one 8 (2013) e84504. [133] a.j. conrad, (google inc.), us20150065821a1 (2015). [134] r.w. walker, r. zietsma, w.k. gray, expert rev. med. devices 11 (2014) 243-245. [135] s. alpert, neuroethics 1 (2008) 55-68. [136] w.h. de jong, p.j. borm, int. j. nanomedicine 3 (2008) 133-149. [137] j. kreuter, p. ramge, v. petrov, s. hamm, s.e. gelperina, b. engelhardt, r. alyautdin, h. von briesen, d.j. begley, pharm. res. 20 (2003) 409-416. [138] j.c. olivier, l. fenart, r. chauvet, c. pariat, r. cecchelli, w. couet, pharm. res. 16 (1999) 18361842. ©2015 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ lipophilicity determination of acidic compounds: meekc as a reliable high-throughput methodology doi: 10.5599/admet.491 1 admet & dmpk 6(2) (2018) 153-161; doi: http://dx.doi.org/10.5599/admet.491 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper lipophilicity determination of acidic compounds: meekc as a reliable high-throughput methodology xavier subirats, lídia redón and martí rosés* institute of biomedicine (ibub) and department of chemical engineering and analytical chemistry, universitat de barcelona, martí i franquès 1-11, 08028 barcelona, spain *corresponding author: e-mail: marti.roses@ub.edu; tel.: +34-934-039-275; fax: +34-934-021-233 received: december 15, 2017; revised: march 05, 2018; available online: april 06, 2018 abstract in the present study a pressure-assisted meekc method with reversed-polarity using a conventional ce instrument with uv detection and uncoated fused silica capillaries is validated as a high-throughput methodology for the lipophilicity determination of the neutral species of acidic compounds (pka > 3.5). after the calibration of the system with four standard compounds of known log po/w, mass distribution ratios (log kmeekc) of new molecules can be directly converted into log po/w values by means of a simple linear equation (log po/w=a·log kmeekc+b). the method was internally and externally validated for a log po/w range between -1.54 and 4.75, with higher accuracies than conventional liquid chromatographic methods. keywords capillary electrophoresis; hydrophobicity; log po/w; microemulsion electrokinetic chromatography introduction meekc as high-throughput surrogate model for the determination of lipophilicity according to iupac [1], lipophilicity represents the affinity of a molecule or a moiety for a lipophilic environment that is commonly measured by its distribution behavior in a biphasic system. since lipophilicity plays a fundamental role in the processes of absorption, distribution, metabolism, excretion, and toxicity (admet) of chemical compounds in biological systems, it is a relevant physicochemical property to be determined in the drug discovery and design process [2]. different lipophilicity indexes can be obtained depending on the particular biphasic system used, but the most widely used is the n-octanol/water partition coefficient (log po/w, also indicated as log ko/w). moreover, since lipophilicity is a critical parameter for chemical safety assessment, according to the reach regulation ((ec) no 1907/2006) log po/w must be reported for any organic compound produced in quantities of one tonne or more per year. thus, two test procedures are described in the test methods regulation ((ec) no 440/2008): a direct measurement via shake-flask methods [3] and a correlation approach by means of hplc [4]. however, other experimental methods can be used provided that they show an acceptable level of quality assurance [5]. this is clearly the case of dual-phase potentiometric titration procedures [6,7], commonly used in pharmaceutical research for ionizable drugs with pka values in a measurable ph range (2-12), which provide reliable and accurate log po/w values [8]. http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:marti.roses@ub.edu marti rosés et al. admet & dmpk 6(2) (2018) 153-161 154 several methods based on chromatographic retention have been proposed in order to measure lipophilicity, mainly using reversed-phase columns and acetonitrile/aqueous buffer mobile phases [9]. these approaches are much more time-efficient compared to shake-flask and potentiometric methods, with the additional benefit of a separation technique that does not require high purity samples. however, the accuracy of the estimated log po/w values depends to a great extent on the similarity between the calibration standards and the sample compounds. for instance, octanol and c18 phases exhibit a different interaction with hydrogen-bond donor solutes; in the chromatographic system solutes with h-bond acidity are discouraged from partitioning out of the hydro-organic mixture and into the poorer h-bond acceptor c18 phase, whereas octanol (either containing water or other h-bond acceptor moieties) can form hydrogen bonds. therefore, the introduction of molecular descriptors, either experimental or calculated, allows better correlations between chromatographic retention and log po/w values [9–12], but is limited to unionized molecules. however, a lower accuracy from chromatographic methods (±0.5 log units [4]) can be expected in relation to the shake-flask method (±0.3 [3]). in previous papers [13,14] an approach based on microemulsion electrokinetic chromatography (meekc) was successfully demonstrated as an indirect method for log po/w determination for compounds of pharmaceutical interest, using uv detection or hyphenated to an ms with an atmospheric pressure photoionizaion (appi) source [15]. meekc is a chromatographic technique (microemulsion (me) droplets acting as pseudo-stationary phase), and requires significantly shorter analysis times than reference shakeflask and potentiometric methods, avoiding in addition the requirement of high-purity samples. although it can only be applied to neutral species, the broad ph range of application (from ph 2 to 12) allows, in most cases, the finding of experimental conditions which ensure the analyte is in its neutral form. this is clearly an advantage over conventional chromatographic methods, since column stability might be a critical factor at such extreme ph values. in addition, and this is maybe the most significant advantage over traditional chromatographic methods, meekc measurements can be accurately correlated with log po/w without the need of molecular descriptors since me are better surrogates of n-octanol/water systems than c18 stationary phases. additionally, running costs of capillary electrophoresis techniques are lower because of the lower price of fused silica capillaries in relation to chromatographic columns and the lower solvent consumption. last but not least, meekc is a very robust technique for log po/w determination, since variations in the me composition (ph, buffer nature, surfactant type and concentration, and even the addition of organic modifiers) [14] were shown to produce insignificant changes in its predictive capacity. in 2000 poole and co-workers [16] published a study about lipophilicity determination by meekc that included acidic compounds using a running buffer of ph 3, and employing sulfonated silica capillaries in order to provide an adequate electroosmotic flow at such an acidic ph. in a previous study involving buffers over a wide range of ph values [14], we proposed a seminal pressure-assisted method with reversed polarity allowing measurements to be made at ph 2 and using conventional uncoated fused silica capillaries, which are much less expensive than the sulfonic acid coated capillaries. in this paper, a systematic study of structurally diverse compounds is presented, proposing a high-throughput validated method for the lipophilicity determination of acids with pka values as low as 3.5. the partition of analytes between the aqueous bulk solvent and the pseudo-stationary oil phase in the me is measured by the logarithm of the mass distribution ratio (log kmeekc) [17]:  r me r eof meekc eof log log 1 t t t t k t    (1) admet & dmpk 6(2) (2018) 153-161 lipophylicity determination of acidic compounds by meekc doi: 10.5599/admet.491 155 where tr and teof are the migration times of the analyte and the electroosmotic flow (eof) marker (e.g. dmso), respectively, and tme is the migration time of a very lipophilic compound used as me marker (e.g. dodecanophenone). thus, log kmeekc values obtained for unionized species can be used to estimate their lipophilicity according to the following linear equation: o/w meekclog ·logp a k b  (2) where a and b are the slope and intercept, respectively, of the linear regression calculated by the least square method. internal and external validation of the method the ability of the model to reproduce the data included in the set of compounds (the goodness-of-fit) is measured by the determination coefficient of the model (r 2 ), which in our case estimates the proportion of the variation in the predicted log po/w that can be explained according to the model. the robustness of the model can be assessed by cross-validation, in which a number of compounds are iteratively excluded from the set of substances used for model development and the results then allow the prediction to be made for the left-out compounds. thus, the cross-validated correlation coefficient (q 2 ), which is a measure of the goodness of the internal power to predict, can be calculated by the formula:     2 i i/i 2 1 2 i 1 ˆ 1 n i n i y y q y y         (3) where iy is the observed response (literature log po/w) for the i th object, /ˆi iy is the response of the i th object estimated by the model obtained without using the i th object (log po/w from log kmeekc measurements), and y is the mean value of observed responses for the n elements of the complete data set. the size of the group of excluded chemicals at every step is normally in the range of only one (leave-one-out, loo) to 50 % of the whole set of compounds (leave-many-out, lmo). high mean q 2 values in loo (q 2 loo) and lmo (q 2 lmo) validations (> 0.7) are necessary but are not sufficient conditions for a model to be robust and internally predictive [18]. the external validation is performed by splitting the whole data set into two different sets, a training set used for method development and a test set for the assessment of the predictive capacity. predicted values are correlated with the experimental ones, and the linear regression is expected to be as close as possible to that of unity slope and null intercept. this closeness can be calculated as (r 2 -r0 2 )/r 2 [19], where r0 2 is the determination coefficient of the regression line forced to pass through the origin, or by means of the concordance correlation coefficient (ccc) [20], which can be calculated by the formula:          i i 1 2 2 2 i i 1 1 2 1 n i n n i i x x y y ccc x x y y n x y                (4) where x and y correspond to the abscissa and ordinate values of the correlation plot, x and y are mean values and n is the number of compounds in the test set. the main advantage of ccc is the independence of the closeness value in relation to the disposition of the axes. thus, a model can be assessed as externally predictive if r 2 and r0 2 are close to 1, slopes of linear regressions are in the range between 0.85 and 1.15, and (r 2 -r0 2 )/r 2 < 0.1 [19], or alternatively if ccc ≥ 0.85 [20]. marti rosés et al. admet & dmpk 6(2) (2018) 153-161 156 experimental instrumentation a g1600 capillary electrophoresis (agilent, waldbronn, germany) with uv detection and polyimide coated capillaries of 50 μm id, 375 μm od and 57.0±0.1/48.5±0.1 cm of total/effective lengths (polymicro technologies, phoenix, usa) were used. the cassette temperature was set to 25 °c (forced air) and samples were injected hydrodynamically by application of a pressure of 50 mbar for 10 s. separations were carried by out applying a voltage of -24 kv (inlet, cathode; outlet, anode) and an external pressure of 50 mbar (on the inlet vial). current intensities were typically in the range between 30 and 40 µa. capillary preconditioning was performed by me for 2 min and postconditioning by 1 m sodium hydroxide and water for 2 min each. ph was measured with a crison glp 22 ph meter (barcelona, spain) using a 5014 combination glass electrode and a reference electrode with a 3.0 mol l −1 kcl solution in water salt bridge. mes were sonicated in a j.p. selecta (barcelona, spain) ultrasonic bath at a power of 360 w. reagents water was deionized to a resistivity of 18.2 mω cm by a milli-q® plus system (millipore, billerica, ma, usa), and buffers were prepared for analysis from phosphoric acid (merck, 85 %), and sodium hydroxide (sigma-aldrich, pellets, >98%). me consisted of sds (≥99.0 %), 1-butanol (hplc grade), and heptane (hplc grade), all from sigma-aldrich. injected compounds were purchased (table 1) from j.t. baker, carlo erba, fluka, merck, sigma-aldrich, and schuhardt; all of high purity grade. the neutral and me markers were dmso (merck, for analysis) and dodecanophenone (sigma-aldrich, 98%), respectively. microemulsion and sample preparation 20 mm aqueous buffers were prepared from phosphoric acid by adjusting the ph to 2.0 by the addition of small volumes of a 3 m sodium hydroxide solution prepared shortly before use. 1.3% w/v of sds was dissolved in the aqueous buffer at room temperature and stirred by magnetic stirrer until a transparent colorless solution was obtained, and the ph then adjusted if necessary. afterwards, 1-butanol (sigmaaldrich, ≥ 99.4%) was added up to 8.15% (v/v), followed by heptane (sigma-aldrich, ≥ 99%) up to 1.15% (v/v). both organic solvents were slowly added with a burette. at this point, the solution became white and turbid. magnetic stirring was maintained for 5 minutes and then the me was sonicated until it became clear again. finally, the solution was left to stand at room temperature for at least 1 hour. immediately before use the me was filtered using a 0.45 μm nylon syringe filter (simplepure, membrane-solutions, usa). sample solutions were prepared by dissolving the me marker (dodecanophenone, 0.5 mg/ml) directly into the me by sonication, followed by the addition of the eof marker (dmso, 0.1% in volume) and the analytes (0.5 mg/ml from a stock solution of 10 mg/ml in methanol). results and discussion method development: pressure-assisted meekc with reversed polarity using uncoated fused silica capillaries firstly, it was necessary to establish the experimental conditions (ph, sds content, capillary length, applied voltage and pressure) to allow the determination of a wide range of lipophilicity values in relatively short analysis times. thus, mixtures of substances with known log po/w values were injected in order to test the system behavior. these compounds were n,n-dimethylacetamide, 1-phenylthiourea, acetophenone, butyrophenone, propylbenzene, and pentachloronitrobenzene, and their corresponding measured log po/w admet & dmpk 6(2) (2018) 153-161 lipophylicity determination of acidic compounds by meekc doi: 10.5599/admet.491 157 values are -0.77, 0.73, 1.58, 2.66, 3.72, and 5.10, respectively [21]. these compounds were easily detected by uv and they behave as neutral species nearly over all of the ph range. only 1-phenylthiourea could behave as an ampholyte, but basic and acidic groups were expected to be too extreme to cause a significant ionization at ph 2.0 (their calculated pka values are 0.7 and 13.1, respectively [22]). under the conditions of assay in the present work, pentachloronitrobenze was too lipophilic to be sufficiently resolved from the me marker (dodecanophenone). it must be pointed out that at very acidic conditions the direction of the eof is found to be reversed in relation to neutral and alkaline running buffers. therefore, in order to overcome this eof issue the instrument polarity was reversed, with the anode in the destination vial (outlet) and the cathode in source vial (inlet). thus, the negatively charged me droplets (because of the sds) were the first to reach the detector window, while the eof marker migrated last. the application of an external pressure, which pushes the me filling the capillary towards the detector, is fundamental to achieving a good separation of the standard mixture in relatively short analysis times. table 1. experimental log po/w from the literature [21] and the calculated pka values (galas [22]) of the acidic compounds included in the study, together with the estimated molar percentage of the neutral species at ph 2.0. compound log po/w pka %neutral compound log po/w pka %neutral piracetam -1.54 13.6 100% 4-nitrobenzoic acid 1.89 3.3 95% barbituric acid -1.41 4.0 99% 4-nitrophenol 1.91 7.2 100% 2,4-dithiouracil -0.86 6.7 100% furosemide 2.03 3.5 97% hydroxypropyltheophylline -0.77 13.6 100% pentobarbital 2.10 7.9 100% thymine -0.62 9.7 100% 2-methyl-4,6-dinitrophenol 2.13 4.2 99% thiouracil -0.28 7.5 100% cinnamic acid 2.13 4.4 100% chlorothiazide -0.24 6.7 100% xipamide 2.19 4.8 100% hydrochlorothiazide -0.07 8.9 100% 2,4-dimethylphenol 2.30 10.0 100% theophylline -0.02 8.7 100% methyl salicylate 2.34 9.8 100% pyrogallol *0.29 9.0 100% 3-methylbenzoic acid 2.37 4.1 99% mandelic acid 0.62 3.4 96% 2-methylbenzoic acid 2.40 3.7 98% barbital 0.65 7.9 100% estriol 2.54 10.0 100% gallic acid 0.70 4.3 100% clofibric acid 2.57 4.0 99% resorcinol 0.80 9.6 100% warfarin 2.70 4.9 100% catechol 0.88 10.0 100% 2,4-dimethylbenzoic acid *2.82 3.7 98% acetylsalicylic acid 1.19 3.5 97% 3-bromobenzoic acid 2.87 3.6 98% vanillin 1.21 7.3 100% 1-naphthoic acid 3.10 3.5 97% 2,6-dinitrophenol 1.37 3.5 97% ketoprofen 3.12 4.1 99% phenol 1.47 10.0 100% 3,5-dichlorobenzoic acid 3.29 3.3 98% phenobarbital 1.47 7.5 100% naproxen 3.34 7.2 100% carbromal 1.54 10.8 100% ibuprofen 3.50 3.5 100% morin 1.54 7.4 100% 2,4,5-trichlorophenol 3.72 7.9 100% 4-hydroxybenzoic acid 1.58 4.6 100% 4-tert-butylbenzoic acid 3.85 4.2 99% 2,4-dinitrophenol 1.67 4.2 99% estradiol 4.01 4.4 100% 3-nitrobenzoic acid 1.83 3.4 96% diclofenac 4.50 4.8 100% benzoic acid 1.87 4.1 99% *data from ref. [23] selected set of substances a set of 51 acidic substances, structurally different, was selected for the validation of the proposed method (table 1). they covered a wide region of chemical space, with lipophilicity values covering 5 log po/w units (between -1.5 and 4.5) and calculated pka values in the range between 3.3 and 13.6. stronger acids had to be excluded, since the me employed had a ph value of 2.0 and this approach required the solutes to be in their neutral form (a molar fraction of 5 % of ionized species was set as threshold). marti rosés et al. admet & dmpk 6(2) (2018) 153-161 158 internal validation as shown in figure 1, there was an excellent correlation between experimental log po/w values found in the literature and log kmeekc measurements performed in the present work. the model successfully explained 92.75 % of the log po/w variation in the selected set of compounds, and the summarized overall error in the prediction of the model (0.38) is only slightly higher than the admitted differences from log po/w replicates by the shake-flask methods [3,24]. only three compounds presented higher errors than two times the standard error of the fitting; these were 2,4-dithiouracil, hydroxypropyltheophylline, and morin. in all cases, predicted log po/w values were higher (1.2, 0.9, and 2.5, respectively) than those found in the literature. figure 1. correlation between log po/w and log kmeekc values for the acidic compound studied (table 1). statistical parameters for the fitting (correlation coefficient, root mean square error, and the fisher statistic) are also shown. legend: (■) pka ≤ 4.0, (●) 4.0 < pka ≤ 6.0, (▲) 6.0 < pka ≤ 8.0, and (▼) pka > 8.0. since the validated model presents only one independent variable (log kmeekc), as expected the crossvalidated correlation coefficients q 2 loo (0.9211) was only slightly lower than the fitting parameter r 2 (0.9275), being all of them significantly much higher than the 0.7 threshold. the mean q 2 lmo, which was calculated from 2000 iterations excluding randomly 50 % of the chemicals of the set, presented a value of 0.915 with a standard deviation of 0.023, being thus very close to q 2 loo and with a low dispersion. therefore, internal validation demonstrated that the model was stable and internally predictive, and thus ready for the external validation step. external validation the application of this method for log po/w determination requires first the calibration of the response of the electrophoretic system according to equation (2). thus, it would be convenient to define a small set of standards in order to allow system calibration using a single injection of a mixture of these compounds. with the aim of finding a suitable set, the three compounds identified as possible outliers (2,4-dithiouracil, hydroxypropyltheophylline, and morin) were left out and a new linear regression log po/w vs. log kmeekc was calculated. four compounds covering a good range of lipophilicities and showing very little residuals were then selected as candidates for the calibration curve and, consequently, as the training set for the external validation. the chosen analytes and the calibration plots obtained with the equation are shown in figure 2. after developing the model with the training set, the 44 remaining compounds (excluding the three admet & dmpk 6(2) (2018) 153-161 lipophylicity determination of acidic compounds by meekc doi: 10.5599/admet.491 159 outliers) were used as a test set. in a strict sense this was not a proper external validation, since the chemicals belonging to the test set had been included in the previous step leading to the selection of the four training set compounds, thus not being completely new molecules. however, the aim of this validation was to prove the predictive capacity of the standard compounds selected for the calibration of the lipophilicity response of the meekc system. as shown in figure 3, there was an excellent correspondence between predicted and observed log po/w values, the slopes of the normal regressions and that forced to the origin being not different from 1. moreover, (r 2 -r0 2 )/r 2 presented a value of 0.002, suggesting that the origin ordinate was not significantly different from 0. in addition, the value for ccc was 0.974, pointing out the very good accuracy of the model in terms of precision (scattering of observation in relation to the fitting line) and trueness (closeness of the regression to the full correspondence represented by a line of slope 1 and intercept 0). figure 2. electropherogram of the mixture of compounds selected as training set for the external validation and the calibration plot used as model development (dashed line). figure 3. correlation between observed and predicted log po/w from external validation. dashed line represents the regression forcing the null origin. marti rosés et al. admet & dmpk 6(2) (2018) 153-161 160 conclusions in contrast to hplc methods, meekc measurements can be accurately correlated with log po/w without the additional need of molecular descriptors. thus, a high-throughput pressure-assisted meekc methodology with reversed polarity for log po/w determination of acidic compounds (pka > 3.5) has been proposed and validated (internally and externally), using a conventional ce instrument with uv detection and uncoated fused silica capillaries. 3-methylbenzoic acid, phenobarbital, barbital, and thiouracil are proposed as calibration standards, allowing the measurement of log po/w values in the range between -1.54 and 4.75 with a prediction accuracy of ±0.4. acknowledgements: this work is supported by the ministry of economy and competitiveness of spain (project ctq2014-56253-p). prof. paola gramatica is acknowledged for the free license of qsarins software [25]. references [1] c.g. wermuth, c.r. ganellin, p. lindberg, l. a. mitscher, glossary of terms used in computational drug design, pure appl. chem. 70 (1998) 1129–1143. doi:10.1351/pac197951081725. [2] j.a. arnott, s.l. planey, the influence of lipophilicity in drug discovery and design, expert opin. drug discov. 7 (2012) 863–875. doi:10.1517/17460441.2012.714363. [3] oecd, test no. 107: partition coefficient (n-octanol/water): shake flask method, oecd publishing, 1995. [4] oecd, test no. 117: partition coefficient (n-octanol/water), hplc method, oecd publishing, 2004. [5] echa, guidance on information requirements and chemical safety assessment. chapter r.7a: endpoint specific guidance: 2016. doi:10.2823/2611. [6] a. avdeef, ph-metric log p. part i. difference plots for determining ion-pair octanol-water partition coefficients of multiprotic substances, quant. struct. relationships. 11 (1992) 510–517. doi:10.1002/qsar.2660110408. [7] a. avdeef, ph-metric log p. ii: refinement of partition coefficients and ionization constants of multiprotic substances, j. pharm. sci. 82 (1993) 183–190. doi:10.1002/jps.2600820214. [8] k. takács-novák, a. avdeef, interlaboratory study of log p determination by shake-flask and potentiometric methods, j. pharm. biomed. anal. 14 (1996) 1405–1413. doi:10.1016/07317085(96)01773-6. [9] k. valkó, physicochemical and biomimetic properties in drug discovery, wiley, hoboken, nj, usa, 2014. [10] j.m. pallicer, s. pous-torres, j. sales, m. rosés, c. ràfols, e. bosch, determination of the hydrophobicity of organic compounds measured as log po/w through a new chromatographic method, j. chromatogr. a. 1217 (2010) 3026–3037. doi:10.1016/j.chroma.2010.02.051. [11] j.m. pallicer, r. pascual, a. port, m. rosés, c. ràfols, e. bosch, the contribution of the hydrogen bond acidity on the lipophilicity of drugs estimated from chromatographic measurements, eur. j. pharm. sci. 48 (2013) 484–493. doi:10.1016/j.ejps.2012.12.008. [12] x. subirats, m. rosés, e. bosch, high-throughput log po/w determination from uhplc measurements: revisiting the chromatographic hydrophobicity index, j. pharm. biomed. anal. 127 (2015) 26–31. doi:10.1016/j.jpba.2015.12.015. [13] y. ishihama, y. oda, k. uchikawa, n. asakawa, evaluation of solute hydrophobicity by microemulsion electrokinetic chromatography, anal. chem. 67 (1995) 1588–1595. doi:10.1002/elps.200410141. admet & dmpk 6(2) (2018) 153-161 lipophylicity determination of acidic compounds by meekc doi: 10.5599/admet.491 161 [14] x. subirats, h.-p. yuan, v. chaves, n. marzal, m. rosés, microemulsion electrokinetic chromatography as a suitable tool for lipophilicity determination of acidic, neutral, and basic compounds, electrophoresis. 37 (2016) 2010–2016. doi:10.1002/elps.201600080. [15] y. henchoz, s. romand, j. schappler, s. rudaz, j.l. veuthey, p.a. carrupt, high-throughput log p determination by meekc coupled with uv and ms detections, electrophoresis. 31 (2010) 952–964. doi:10.1002/elps.200900540. [16] s.k. poole, d. durham, c. kibbey, rapid method for estimating the octanol-water partition coefficient (log pow) by microemulsion electrokinetic chromatography, j. chromatogr. b biomed. sci. appl. 745 (2000) 117–126. doi:10.1016/s0378-4347(00)00072-4. [17] m.-l. riekkola, j.å. jönsson, terminology for analytical capillary electromigration techniques, int. union pure appl. chem. 76 (2001) 1–8. doi:10.1351/pac200476020443. [18] p. gramatica, principles of qsar models validation: internal and external, qsar comb. sci. 26 (2007) 694–701. doi:10.1002/qsar.200610151. [19] a. golbraikh, a. tropsha, beware of q 2 !, j. mol. graph. model. 20 (2002) 269–276. doi:10.1016/s1093-3263(01)00123-1. [20] n. chirico, p. gramatica, real external predictivity of qsar models: how to evaluate it? comparison of different validation criteria and proposal of using the concordance correlation coefficient, j. chem. inf. model. 51 (2011) 2320–2335. doi:10.1021/ci200211n. [21] bio-loom. biobyte corp. claremont, ca, usa. www.biobyte.com. [22] acd/labs. advanced chemistry development, inc. toronto, on, canada. www.acdlabs.com. [23] chemspider. royal society of chemistry, uk. www.chemspider.com. [24] epa, product properties test guidelines. oppts 830.7550. partition coefficient (n-octanol/water), shake flask method, u.s. environmental protection agency, prevention, pesticides and toxic substances, united states, 1996. [25] p. gramatica, n. chirico, e. papa, s. cassani, s. kovarich, qsarins: a new software for the development, analysis, and validation of qsar mlr models, j. comput. chem. 34 (2013) 2121– 2132. doi:10.1002/jcc.23361. ©2018 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ manuscript doi: 10.5599/admet.5.1.387 59 admet & dmpk 5(2) (2017) 59-74; doi: 10.5599/admet.5.2.387 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index review uric acid transporter inhibitors for gout philip k. tan and jeffrey n. miner* pharmadiscovery consulting llc, san diego, ca 92122 *corresponding author: e-mail: pharmadiscoveryconsulting@gmail.com; received: april 20, 2017; revised: may 21, 2017; published: june 22, 2017 abstract gout is a common inflammatory arthritis that is caused by chronically-elevated serum uric acid (sua) levels (hyperuricemia). in humans, sua levels are predominantly controlled by a variety of transporters that mediate the elimination of uric acid through the kidneys and intestines, a process that is altered in most gout patients. in this review, we highlight our current understanding of uric acid handling in healthy individuals and gout patients, therapies for gout that target uric acid transporters, and the mechanism of other therapies that alter sua levels through interactions with uric acid transporters. keywords gout; hyperuricemia; uricase; urat1; coevolution; urate transporters for urate homeostasis overview gout and hyperuricemia gout is a debilitating inflammatory arthritis of increasing prevalence that is caused by chronically elevated levels of serum uric acid (sua), or hyperuricemia, defined as concentrations exceeding 6.8 mg/dl (408 μm). in healthy nongouty individuals, serum urate levels are normally in the range of 3.5-7 mg/dl (210 – 420 μm) [1]. prolonged hyperuricemia can initiate the precipitation of the uric acid in joints and other tissues, and these deposits can trigger an acute and painful immune response known as a gout flare. hyperuricemia is also strongly and independently associated with a number of other important disorders including hypertension, cardiovascular disease and metabolic syndrome [2]. a number of therapies for gout that lower sua levels target transporters for uric acid, some which are also important drug transporters. uric acid homeostasis in humans in health and disease uric acid is produced mainly in the liver and gastrointestinal tract by the degradation of endogenous and dietary purines. sources of endogenous purines include nucleotides such as atp, and dna and rna that are recovered from dying cells. these purines are degraded into uric acid by xanthine oxidase in the liver. dietary purines are absorbed in the gut by the cnt2 transporter and are quantitatively metabolized to uric acid by endogenous xanthine oxidase within intestinal enterocytes (see figure 1b). in non-primate mammals, uric acid is converted to more highly soluble allantoin by urate oxidase (uricase). in primates including humans, however, uric acid is the end product of purine catabolism due to the progressive http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:pharmadiscoveryconsulting@gmail.com tan and miner admet & dmpk 5(2) (2017) 59-74 60 inactivation of uricase in primate evolution [3] hence primates have sua levels that are 3-10 times higher than other mammals. the xanthine oxidase inhibitor (xoi) class of urate-lowering therapies for gout (allopurinol and febuxostat) block the enzyme xanthine oxidase that produces uric acid in the last step of purine catabolism in primates [4]. uric acid production is balanced by elimination, which in humans occurs mainly through the kidneys (70 % of the total eliminated uric acid) but also through the intestines (30 % of the total eliminated uric acid). with a pka of 5.35 [5], uric acid is the negatively charged organic anion urate at physiological ph and so requires transporters for passage through cell membranes. genetic studies show that a large number of urate transporters (glut9, abcg2, npt1/3/4, urat1, oat4, mrp4) are important for urate homeostasis in humans [6-8] which are expressed in the organs that eliminate uric acid. in the kidney nephrons, uric acid handling occurs in a complex process involving complete and free filtration into the glomerulus, followed by near complete 90-95 % reabsorption in the proximal tubules back to the bloodstream (figure 1a). uric acid that is not reabsorbed in the proximal tubule (5-10 % of the glomerular-filtered uric acid) passes through the nephron and is excreted into the urine, and there is little evidence that uric acid can pass through the walls of the more distal segments of the nephron. the transporters involved in the proximal tubule handling of uric acid are shown in figure 1a and are described in detail below. a) b) figure 1. organization of urate transporters in polarized epithelial cells of the kidney (a) and intestine (b). transporters involved in reabsorption/absorption of urate are shown in green, while those involved in secretion are shown in red. transporters in ovals are genetically-associated with sua levels, while transporters in boxes have no genetic association with sua levels, but transport urate in vitro. specific urate transporters in the intestine, particularly on the basolateral membrane, are unknown. intestinal cnt2 is involved in import of dietary purines from the intestinal lumen, which are efficiently metabolized to urate by the enterocytes. admet & dmpk 5(2) (2017) 59-74 uric acid transporter inhibitors for gout doi: 10.5599/admet.5.1.387 61 in the intestine, uric acid permeability from mucosal to serosal (absorption) and serosal to mucosal (secretion) has been documented, but the identities of the responsible transporters in humans are, for the most part, unknown (figure 1b). the important drug transporter breast cancer resistance protein (bcrp, encoded by the abcg2 gene) is highly expressed on the apical surface of intestinal epithelial cells and is also an important urate transporter involved in intestinal urate secretion [9-12]. dietary purines are imported from the lumen of intestinal enterocytes by the nucleoside transporter cnt2/slc28a2, which are then converted to uric acid by intestinal xanthine oxidase. inhibitors of cnt2 have been shown to reduce sua levels in non-human primates and are therefore possible candidates for new urate-lowering therapies [13]. the chronically elevated concentrations of sua in gout patients are caused by increased uric acid production, decreased uric acid elimination, or a combination of the two. in most gout patients, high sua levels are due to decreased uric acid elimination in the kidneys, recently confirmed with clinical trial results [14]. we also discovered that gout patients have an altered renal urate reabsorption profile in response to xoi urate lowering therapies, relative to healthy individuals. these differences in the renal urate reabsorption in gout patients are likely due to differences in the activities of renal urate transporters. it is possible that transporters involved in reabsorption (colored green in figure 1a) are more active, leading to enhanced reabsorption of urate, or that transporters involved in urate secretion (colored red in figure 1a) are less active, or some combination of the above two scenarios [14]. it is also now known that many drugs and metabolic anomalies increase or decrease sua levels by altering the activities of urate transporters, and this is explained in the following section. uric acid transporters and drug interactions urat1 the renal uric acid transporter type 1 (urat1/slc22a12) is expressed on the apical membranes of renal proximal tubule epithelial cells (rptecs) and is required for the first step in reabsorption of uric acid from the lumen of the proximal tubule to the bloodstream (figure 1a). individuals carrying inactivating mutations of urat1 have high levels of urinary uric acid excretion and low levels of sua (renal hypouricemia) due to reduced uric acid reabsorption from the proximal tubules, showing that urat1 is essential for efficient renal reabsorption of uric acid [15]. the uricosuric class of urate-lowering therapies (lesinurad, probenecid, and benzbromarone) for the treatment of gout lower sua concentrations by inhibiting urat1. other drugs indicated primarily for other therapies (losartan, hypertension; fenofibrate, hyperlipidemia) are considered secondary uricosurics because they also inhibit urat1 [16, 17]. urat1 is a member of the organic anion transporter (oat) subtype of solute carrier (slc) type 22a family (slc22a) of transporters, which typically function as organic anion exchangers. specifically, urat1 imports urate in exchange for export of small anionic counterions such as lactate (figure 2). the activity of urat1 depends on these counterions, and certain pathophysiological conditions and drugs alter sua levels by altering the activity of urat1. increased concentrations of counterions can stimulate the counterion transport of urat1, leading to increased uric acid transport of urat1 in a process called trans-stimulation. alcohol poisoning (lactic acidosis) and diabetic ketoacidosis elevate sua levels due to elevated levels of the urat1 counterions lactate and ketoacids (acetoacetate and β-hydroxybutyrate), respectively [18, 19]. furthermore, salicylate (at low doses), niacin or nicotinate (for treatment of hyperlipidemia), and pyrazinoate (the primary metabolite of the tuberculosis therapy pyrazinamide) function as pharmacological urat1 counterions and also elevate sua levels through trans-stimulation of urat1. high doses of tan and miner admet & dmpk 5(2) (2017) 59-74 62 salicylate, on the other hand, inhibit the urate transport activity of urat1 (cis-inhibition) to lower sua levels. therefore, the paradoxical effect of salicylate dosage on sua levels is explained through different dose-dependent interactions with urat1. figure 2. effects of xenobiotics and endobiotics on urate transport in the kidney proximal tubule. for each transporter, endogenous substrates are in normal case, while drugs are in bold (nicotinate is an endogenous substrate for urat1 but is also used as a therapy for hyperlipidemia). the proposed direction of urate transport is shown for each transporter. drugs that stimulate transport of urate are not italicized, while drugs that inhibit urate transport are italicized. see text for more details. α-kg, alpha-ketoglutarate. urat1 is classified as a tertiary active transporter, as its activity is dependent on counterion and sodium gradients generated by secondary active transporters (na+/monocarboxylate cotransporters smct1/2) and the primary active transporter (na+/k+ atpase), respectively (figure 3). urat1 and smct1/2 also interact with pdzk1, a protein involved in the localization of apical membrane proteins in kidney and intestinal epithelial cells. these transporters, including many other urate transporters, interact with pdzk1 through specific c-terminal binding sequences (table 1). pdzk1 has a strong genetic association with sua levels, suggesting that pdzk1 functions as a scaffolding protein to assemble the various transporters into a “uric acid transportasome ” for efficient control of urate transport and homeostasis (figure 3) [20]. the urate transport activity of urat1 depends on a system of other transporters (yellow in figure 3). the primary active transporter na + /k + atpase provides an inward sodium ion gradient (higher concentration of extracellular sodium ions). the secondary active transporters, the sodium monocarboxylate cotransporters smct1/slc5a8 and smct2/slc5a12, use the sodium ion gradient for the inward cotransport of lactate and other monocarboxylates. finally, urat1 is the tertiary active transporter in this system, requiring the counterion lactate/monocarboxylate gradient generated by smct1/2 for the inward transport of urate. furthermore, we suggest that the sodium/proton exchanger nhe3/slc9a3, which imports sodium ions and exports protons, may also be important for urat1 activity. this system of transporters that is indirectly involved in import of urate from the lumen of the proximal tubule are admet & dmpk 5(2) (2017) 59-74 uric acid transporter inhibitors for gout doi: 10.5599/admet.5.1.387 63 sensitive to certain drugs that elevate sua levels by increasing the urate transport activity of urat1. for example, diuretics reduce the urinary excretion of urate and elevate sua levels, possibly through enhanced sodium reabsorption in the proximal tubule. this effect would elevate the intracellular levels of lactate within the renal proximal epithelial cells (through slc5a8/12) and also lower the extracellular ph (through nhe3), enhancing the reabsorption of urate by urat1. figure 3. the scaffolding protein nherf3/pdzk1 clusters apical membrane urate transporters into a “urate transportasome” for more efficient urate handling in the kidney. transporters indicated in yellow do not transport urate, but are important for the transport of urate by urat1 (see text for further details). for simplicity, certain transporters are grouped together (e.g. npt1 and npt4). glut9 the slc family 2a, facilitated glucose transporter member 9 (glut9/slc2a9), like urat1, is also essential for the efficient reabsorption of urate from the renal proximal tubule (figure 1a). individuals carrying inactivating mutations in glut9 have renal hypouricemia and very high levels of urinary uric acid excretion. in some individuals with inactivating glut9 mutations, the fractional excretion of uric acid exceeds 100 % (meaning that the amount of renally excreted uric acid exceeds the amount of glomerularfiltered uric acid), providing evidence that uric acid is actively secreted in the kidneys (red arrow in figure 1). the long isoform of glut9 (glut9a or -l) is localized to the basolateral membrane of rptecs, and transports uric acid along its concentration gradient from inside the cells to the blood, completing the process of uric acid reabsorption that is initiated by urat1 (figure 1a). unlike urat1, there are no medications that target glut9. similar to other members of the slc2 family, glut9 has been shown to transport glucose and other sugars, although this has recently been questioned [21]. however, sugars may influence the transport of urate by glut9 [22], suggesting that dietary sugars can influence sua levels through glut9. a short isoform of glut9 (glut9b or –s) differs at the amino terminus from the longer isoform due to alternative splicing. an antibody raised against the unique 21 amino terminal amino acids of glut9b detected this variant on membranes of epithelial cells in the collecting duct [23], suggesting the possibility that this distal segment of the nephron handles urate. interestingly, this antibody did not detect glut9b on the proximal tubule. tan and miner admet & dmpk 5(2) (2017) 59-74 64 oat4 oat4/slc22a11 is expressed on the apical membrane of rptecs (figure 1a), has a 51 % amino acid identity to urat1, and has a strong genetic association with sua levels. however, oat4 is a weak transporter of uric acid in vitro, and its role in vivo in uric acid transport and homeostasis is unclear. inactivating mutations of oat4 have been reported [24] but the sua levels in the individuals carrying these mutations is unknown. similar to urat1, oat4 is an organic anion exchanger, and diuretics such as hydrochlorothiazide were shown to enhance urate import by oat4 (trans-stimulation), explaining a possible oat4-dependent mechanism for diuretic-induced hyperuricemia [25]. however, to our knowledge these studies have not yet been replicated, and mixed findings have been reported on the genetic association of oat4 with diuretic-associated gout [26, 27]. oat4 is not expressed in mice, and no specific oat4 inhibitors are known (lesinurad and probenecid inhibit oat4, but also other transporters), hampering efforts to study oat4 using genetic or pharmacological methods. oat10 oat10/slc22a13 was initially characterized as a urate transporter that is expressed on apical membranes of proximal tubule cells and trans-stimulated by cyclosporine, explaining a mechanism for cyclosporine-induced hyperuricemia [28]. this localization is supported by a subsequent report using different antibodies [29], but is inconsistent with a third study which shows localization on basolateral membranes of the collecting duct [30]. oat10, however, has a consensus c-terminal pdz recognition motif (table 1) for localization to apical membranes of proximal tubule cells. no genetic association of oat10 with sua levels has been found, and the role of oat10 in urate transport and homeostasis needs to be further characterized. table 1. carboxyl (c-) terminal sequences of transporters known or presumed to interact with pdzk1. location transporter c-terminal sequence reference kidney urat1/slc22a12 stqf anzai 2004 [31] oat4/slc22a11 stsl miyazaki 2005 [32] oat10/slc22a13 styf bahn 2008 [28] a smct1/slc5a8 gtrl kanai 2008 [33] smct2/slc5a12 tthf kanai 2008 [33] npt1/slc17a1 htrl gisler 2003 [34] npt4/slc17a3 ltrl jutabha 2010 [35] a mrp2/abcc2 stkf kocher 1999 [36] mrp4/abcc4 etal park 2014 [37] nhe3/slc9a3 sthm gisler 2003 [34] intestine npt5/slc17a4 fthl togawa 2012 [38] b bcrp/abcg2* kkys shimizu 2011 [39] *bcrp may interact with pdzk1 in a non-classical way via an internal sequence. a,b apical membrane localization in rptec a or enterocytes b . presumed to interact with pdzk1 based on c-terminal sequence. admet & dmpk 5(2) (2017) 59-74 uric acid transporter inhibitors for gout doi: 10.5599/admet.5.1.387 65 oat1 and oat3 oat1/slc22a6 and oat3/slc22a8 are important transporters involved in the renal secretion of many drugs. they are localized to the basolateral membranes of rptecs and transport urate in vitro, suggesting that they are involved in urate secretion in vivo [40]. however, oat1 and oat3 show no genetic association with sua levels, and there is no solid evidence supporting a role in uric acid homeostasis in humans. oat1 and oat3 are in the same transporter family and have a 46 % and 42 % amino acid identity with urat1 [41]. the urat1 inhibitor probenecid also inhibits oat1 and oat3, leading to oat1/3dependent drug-drug interactions that limits its use. the urat1 inhibitor lesinurad has detectable interactions with oat1 and oat3 in vitro, but in contrast to probenecid has no clinically relevant oat1/3dependent drug-drug interactions [42, 43]. because of the relatedness to urat1, all urat1 inhibitors under evaluation for treatment of hyperuricemia associated with gout should be evaluated against oat1 and oat3. npt1, npt4 and npt5 npt1/slc17a1, npt4/slc17a3, and npt5/slc17a4 (formerly known as npt-like) occur in a gene cluster with a strong genetic association with sua levels [7]. npt1 and npt4 localize to the apical membrane of rptecs and are hypothesized to export urate into the tubule lumen as part of the urate secretory pathway. support for this functional role comes from the identification and analysis of missense mutations involved in hyperuricemia and/or gout that exhibit altered urate transport activities [35, 44-47]. moreover, it has been suggested that hyperuricemia induced by diuretics may be due to inhibition of npt4 [45]. meanwhile, npt5 is a urate exporter that expressed on the apical surface of intestinal epithelial cells, suggesting a role in intestinal urate secretion [38]. mrp4 multi-resistance protein 4 (mrp4/abcc4) transports urate in vitro, localizes to the apical membranes of rptec, and was recently shown to be genetically associated with gout and renal excretion of uric acid [8]. consistent with its proposed role as an efflux transporter involved in the renal secretion of urate, a nonsynonymous variant p1036l of mrp4 had reduced uric acid transport activity [6]. diuretics have been proposed to elevate sua levels through inhibition of mrp4 [48]. abcg2 atp-binding cassette sub-family g member 2 (abcg2, also known as breast cancer-resistance protein or bcrp, encoded by the abcg2 gene) is an atp-dependent transporter that is expressed on the apical membranes of rptec and intestinal epithelia, involved in the export of urate (secretion), and with a strong genetic association with sua levels and gout. humans carrying a common genetic variant, q141k, which exhibits reduced expression and activity, have elevated sua levels and elevated renal excretion of uric acid, supporting a role for abcg2 in the intestinal secretion of uric acid [49]. abcg2 is strongly expressed in the intestine and is weakly expressed in the kidneys, suggesting that it primarily functions to transport urate into the intestinal lumen. the strong genetic association with sua and gout suggests that drugs that activate abcg2 could serve as gout therapies. evolution of gout in humans uric acid is a small, seemingly innocuous waste compound found mainly in the excreta from many different species. biochemically, this simple compound is merely one step in the removal pathway for tan and miner admet & dmpk 5(2) (2017) 59-74 66 nitrogenous wastes in reptiles and birds and for purines in humans. the degradation of purines from nucleic acid results in the production of uric acid, which in most mammals is instantly converted to allantoin by the enzyme uricase. however, about 43 million years ago, for unknown but likely interesting reasons, selective pressure mounted for higher blood levels of uric acid. primates carrying mutations in the uric acid handling pathways that resulted in higher uric acid levels experienced a selective advantage. this pressure continued for 20 million years during simian evolution steadily decreasing uricase activity and increasing renal uric acid transporter affinity for uric acid, resulting in humans with no uricase enzyme activity, a high affinity renal uric acid reabsorbing transporter and the highest uric acid levels of any animal in the class mammalia [50]. the hypotheses explaining the rise in sua are many [3], and may involve enhanced energy storage in resource-constrained environments due to biochemical effects of uric acid on metabolism. these same selective pressures no longer exist today and humans are left with a uric acid load that in a time of resource excess has created an epidemic of gout and potentially other comorbidities associated with uric acid [51]. despite the strength of the association between uric acid and cardiovascular disease, hypertension and chronic kidney disease, these studies do not provide causation, thus it is unclear if uric acid levels are the culprit in driving these illnesses. more recent studies utilizing mendelian randomization and interventional urate lowering drugs have suggested that some aspects of the comorbidities are not caused by uric acid [52, 53] and others indicate that uric acid is causing these comorbidities [54, 55]. the data for these findings is actively being developed thus the overall picture will certainly change in the coming years. this is important because if uric acid is participating in these comorbidities, then urate lowering therapy could well be an important additional treatment for vascular disease, which could work in addition to current therapies. in humans, uric acid is produced by the degradation of endogenous purines and the purines in food. there are three primary sources of uric acid in the body: 1) metabolism, whereby free circulating purines are degraded in the liver and intestines by xanthine oxidase into uric acid, 2) ongoing cellular degradation (releasing dna and rna), which are degraded into uric acid, and 3) diet has a smaller impact through the ingestion of purines and nucleic acids and subsequent metabolism of these compounds in the intestinal mucosa by resident xanthine oxidase. the effects of dietary purines on serum uric acid levels is relatively mild with even strict diets affecting uric acid levels marginally [56]. purine increases from resulting from fructose, beer (from yeast) and other purine rich foods is of primary concern. fructose ingestion depletes atp from the action of fructokinase. the depleted atp (amp) is converted to uric acid through the purine catabolic pathway. once produced, urate cannot be degraded by the human body so all urate must be eliminated either by the kidney or the intestine. two-thirds of urate elimination is through the kidneys, the remainder through the intestines. the balance of production and elimination determines the concentration of uric acid in the serum as described above. genetics of uric acid genome-wide association studies (gwas) have identified 28 separate loci that control hyperuricemia and gout. many of these code for transporters directly involved in the regulation of uric acid levels in humans. these include transporters found in the kidney, liver and intestine with functions spanning both uric acid secretion and resorption [57]. gwas studies also show influence of urate production (glucokinase regulator or gckr, inhibin-activin network, insulin) – links to glucose homeostasis, diabetes, feeding status, energy expenditure. admet & dmpk 5(2) (2017) 59-74 uric acid transporter inhibitors for gout doi: 10.5599/admet.5.1.387 67 interestingly, xanthine oxidase, and other enzymes in the uric acid biosynthetic pathway are not found in gwas studies suggesting that despite a clear role, apparently genetically available variations in the activity of these enzymes does not lead to alterations in uric acid levels. the work in this area is proceeding rapidly with ongoing resequencing projects and expression analysis in clinical samples. renal handling of uric acid in early studies on renal apical bbmvs (brush-border membrane vesicles) and basolateral vesicles, transport activities were characterized [58] and eventually led to the cloning of uric acid transporters from renal tissue and the discovery of a multiprotein transportasome functioning to regulate uric acid levels at the apical membrane [59, 60]. in a normal human kidney, clearance of urate is a function of glomerular filtration rate (gfr) and fractional excretion of uric acid. urate is freely filtered at the glomerulus and is subsequently almost entirely reabsorbed within the proximal tubule. secretion of urate also occurs, either concomitantly with reabsorption or subsequent to reabsorption. the rate of urate excretion as a percentage of the rate of filtration of urate is termed the fractional excretion of uric acid (feua). low feua correlates with increased sua and the characterization of the efficiency of urate secretion is typically done by measuring feua. gout patients often exhibit significantly decreased feua. in addition, decreased gfr also increases sua, so patients with low feua, or impaired renal function are in danger of hyperuricemia and gout renal uric acid reabsorption most uric acid that is filtered by the glomerulus is reabsorbed in the proximal tubule. the main transporter on the apical side of the tubule is urat1, however oat4 and oat10 are also possible contributors to uric acid resorption across the apical membrane. it is likely that the sole transporter on the basolateral side is glut9. this is because the phenotype of glut9 mutations in humans results in an feua at or above 100 %, suggesting complete loss of resorption [61]. in contrast, the effect of urat1 mutations on feua is less severe >50 % [62]. other transporters may be present on the apical side but it is believed that only glut9 functions to export urate to the interstitium (blood side) across the basolateral membrane. the sensitivity of the kidney to alterations in transporter function is also detected at the level of gene expression of the transporter urat1. in women, urat1 expression is likely down regulated by estrogens resulting in fewer transporters at the apical surface and gfr is increased by estrogen through renal vasodilation. in mice, urat1 and glut9 expression is decreased in the presence of estradiol consistent with premenopausal women having significantly lower uric acid levels that post-menopausal women [63, 64]. further confirmation of this effect is obtained through testing of uric acid levels in patients undergoing gender reassignment. addition of estrogen lowers uric acid levels and addition of androgen raises them in these patients [65]. reabsorption transporters cyclical uric acid transport at the proximal tubule membrane the result of import and export transporters coexpressed and functioning on both the apical and tan and miner admet & dmpk 5(2) (2017) 59-74 68 basolateral sides of the tubular membrane means that uric acid likely cycles through the kidney multiple times during the day and perhaps within the same tubule. the renal system filters approximately 180 liters of water each day, similarly for uric acid each molecule of uric acid is likely subjected to multiple rounds of filtration, resorption and /or secretion. the relative affinity and capacity curves for these transporters govern the percentage of uric acid that is resorbed, secreted and the percentage that finally ends up excreted in urine [41]. a similar process is likely also occurring in the intestine. urat1 inhibitors used for the treatment of gout (figure 4) benzbromarone, (3,5-dibromo-4-hydroxyphenyl)-(2-ethyl-1-benzofuran-3-yl)methanone benzbromarone was originally derived from benziodarone, an iodinated compound synthesized initially for coronary dilation. benziodarone was withdrawn from the market in europe due to instances of jaundice in patients treated with this compound. prior to the withdrawal, it was noted that benziodarone was an effective uric acid lowering agent. the brominated analog of benziodarone was synthesized as a follow up for uric acid lowering in patients with gout (benzbromarone) [66]. this compound was withdrawn from the market later due to idiosyncratic liver toxicity and is only used in europe on a named patient basis. benzbromarone is capable of blocking urat1 potently and selectively, it is relatively weak on both oat4 and on glut9. the pharmacokinetics of benzbromarone provide for a cmax of approximately 5 μm [67]. this is well within the potency on urat1 at 100 nm, however, the high μm potency on other transporters (glut9 and oat4) in vitro, coupled with the relatively poor renal excretion profile suggests that benzbromarone will primarily inhibit urat1 to produce its uricosuric effect. n s oh o n n nn s oh o br o br br oh o o o n h cl cf 3 o o n s oho oo verinurad (rdea3170) lesinurad benzbromarone arhalofenate probenecid figure 4. structure of urat1 inhibitors for the treatment of gout. verinurad and arhalofenate are currently under evaluation in clinical trials. probenecid, 4-[(dipropylamino) sulfony1] benzoic acid probenecid is a compound originally developed by bayer to reduce the renal excretion of penicillin given the difficulty in obtaining sufficient quantities of antibiotic to cover the overwhelming need at that time. probenecid is an inhibitor of oat1 and oat3, transporters responsible for renal excretion of many drugs not just antibiotics, thus use of probenecid is limited due to the significant drug-drug interactions. admet & dmpk 5(2) (2017) 59-74 uric acid transporter inhibitors for gout doi: 10.5599/admet.5.1.387 69 after the use of probenecid became widespread in combination with antibiotics, it was discovered that it also reduced serum uric acid levels by inhibition of renal reabsorption, thus making it useful for the treatment of gout. the compound is still used for this purpose today, though the drug – drug interactions, multiple daily dosing and generally weak efficacy limit its use overall [68]. lesinurad (rdea594), 2-((5-bromo-4-(4-cyclopropylnaphthalen-1-yl)-4h-1,2,4-triazol-3-yl)thio)acetic acid lesinurad was discovered as an active metabolite from an unrelated clinical program. patients given the parent compound experienced a profound uric acid lowering effect. this effect was due to drug-mediated alterations in fractional excretion of uric acid [43]. this result suggested renal transporter inhibition as the mechanism. one metabolite from the parent was preferentially excreted to urine (rdea594). purification and testing for urat1 inhibition provided evidence that rdea594 (lesinurad) was responsible and likely mediated its effects through inhibition of urat1. this was later confirmed with human pk / pd studies in phase 1 [43]. lesinurad is a molecule capable of significant uric acid lowering in the presence of an xo inhibitor via the inhibition of urat1. this inhibition increases the feua in patients taking this medicine which targets the defect found in most gout patients. combination therapy showed a beneficial profile compared to monotherapy allopurinol, however significant numbers of patients experienced creatinine elevations while on lesinurad thus careful renal monitoring by prescribing physicians is required. verinurad (rdea3170), 2-((3-(4-cyanonaphthalen-1-yl)pyridin-4-yl)thio)-2-methylpropanoic acid verinurad is a selective uric acid reabsorption inhibitor with nm potency against urat1 [69]. it is in clinical development for the treatment of gout and hyperuricemia in combination with a xanthine oxidase inhibitor. in constrast to lesinurad, in phase 2 studies in combination with a xanthine oxidase inhibitor, verinurad treatment has not resulted in serum creatinine elevations, likely due to its pharmacodynamic/pharmacokinetic profile and the action of the xanthine oxidase inhibitor reducing the urinary uric acid concentration and thereby reducing the incidence of uric acid crystallization and scr elevations. the combination of verinurad and a xanthine oxidase inhibitor provided greater reductions in sua than by a xanthine oxidase inhibitor alone even after doubling the dose of the xoi [70]. arhalofenate, 2-acetamidoethyl (2r)-2-(4-chlorophenyl)-2-[3-(trifluoromethyl)phenoxy]acetate this urat1 inhibitor is currently in clinical trials for gout [71]. arhalofenate when dosed at 800 mg per day reduced serum uric acid similarly to 300 mg of allopurinol. interestingly, this treatment also decreased gout flares significantly compared to allopurinol at a dosage of 300 mg. arhalofenate was well tolerated in early trials. in addition to its uricosuric properties, it also appears to be capable of reducing inflammation. this effect could lead to reductions in flares during treatment though the data to date is inconclusive. combination studies are underway and development is continuing in the us. binding of known inhibitors the interactions between uricosuric agents and urat1 has been studied using modeling and chimeric mapping studies [41]. this work has demonstrated that the urat1 inhibitors (lesinurad, probenecid, benzbromarone and rdea3170 (verinurad) all bind to the same site in the transporter [41]. there are interesting differences in the specific amino acids required for binding for each compound but this is not surprising given the structural differences between them. it’s likely that there is a specific characteristic of this location within the transporter that makes it susceptible to inhibition by a wide variety of compounds. tan and miner admet & dmpk 5(2) (2017) 59-74 70 one of these characteristics may be that uric acid itself appears to interact with this pocket. modeling of the relevant amino acids onto and oat1 scaffold suggests that these amino acids are located within the interior channel of the transporter, ideally positioned to bind either substrate or inhibitor [41]. conclusions human uric acid level in serum is regulated by a complex array of multi-protein complexes found in several different organs. evolutionary pressure to increase uric acid levels likely resulted in the need to regulate these levels. the kidney plays the key role by sensing current uric acid levels from the blood in the filtrate and adjusting reabsorption accordingly. the intestine is also likely capable of sensing and responding to serum uric acid levels, however, it is less well characterized. the risks of elevated uric acid are well characterized and a number of therapies have been developed to reduce these levels in patients with gout. urat1 inhibitors represent one of the two main approaches for reduction of serum uric acid and work by increasing the efficiency of uric acid excretion. the other relies on the blockade production of uric acid via inhibition of xanthine oxidase. combination therapy with both an excretion enhancer and a production inhibitor is a powerful approach to efficient urate lowering therapy. references [1] d. khanna, j.d. fitzgerald, p.p. khanna, s. bae, m.k. singh, t. neogi, et al. arthritis care & research 64 (2012) 1431-1446. [2] s.g. mallat, s. al kattar, b.y. tanios, a. jurjus, current hypertension reports 18 (2016) 74. [3] j.t. kratzer, m.a. lanaspa, m.n. murphy, c. cicerchi, c.l. graves, p.a. tipton, e.a. et al. proceedings of the national academy of sciences of the united states of america 111 (2014) 3763-3768. [4] m.b. bridgeman, b. chavez, expert opinion on pharmacotherapy 16 (2015) 395-398. [5] k. sakhaee, journal of nephrology 27 (2014) 241-245. [6] m. kolz, t. johnson, s. sanna, a. teumer, v. vitart, m. perola, et al. plos genetics 5 (2009) e1000504. [7] a. kottgen, e. albrecht, a. teumer, v. vitart, j. krumsiek, c. hundertmark, et al. nature genetics 45 (2013) 145-154. [8] c. tanner, j. boocock, e.a. stahl, a. dobbyn, a.k. mandal, m. cadzow, a.j. phipps-green, r.k. topless, j. harre hindmarsh, l.k. stamp, n. dalbeth, h.k. choi, d.b. mount, t.r. merriman, arthritis & rheumatology (2017). [9] x. xu, c. li, p. zhou, t. jiang, pharmaceutical biology 54 (2016) 3151-3155. [10] t. takada, k. ichida, h. matsuo, a. nakayama, k. murakami, y. yamanashi, h. kasuga, n. shinomiya, h. suzuki, nucleosides, nucleotides & nucleic acids 33 (2014) 275-281. [11] a. nakayama, h. matsuo, t. takada, k. ichida, t. nakamura, y. ikebuchi, k. ito, t. hosoya, y. kanai, h. suzuki, n. shinomiya, nucleosides, nucleotides & nucleic acids 30 (2011) 1091-1097. [12] a. hosomi, t. nakanishi, t. fujita, i. tamai, plos one 7 (2012) e30456. [13] m. hiratochi, k. tatani, k. shimizu, y. kuramochi, n. kikuchi, n. kamada, f. itoh, m. isaji, european journal of pharmacology 690 (2012) 183-191. [14] s. liu, f. perez-ruiz, j.n. miner, joint, bone, spine : revue du rhumatisme 84 (2017) 183-188. [15] a. enomoto, h. kimura, a. chairoungdua, y. shigeta, p. jutabha, s.h. cha, et al. nature 417 (2002) 447-452. [16] p.h. gois, e.r. souza, the cochrane database of systematic reviews (2013) cd008652. [17] a.s. kydd, r. seth, r. buchbinder, l. falzon, c.j. edwards, d.m. van der heijde, c. bombardier, the journal of rheumatology. supplement 92 (2014) 33-41. admet & dmpk 5(2) (2017) 59-74 uric acid transporter inhibitors for gout doi: 10.5599/admet.5.1.387 71 [18] c.s. lieber, d.p. jones, m.s. losowsky, c.s. davidson, the journal of clinical investigation 41 (1962) 1863-1870. [19] j. padova, g. bendersky, the new england journal of medicine 267 (1962) 530-534. [20] r.l. george, r.t. keenan, current rheumatology reports 15 (2013) 309. [21] k. ebert, m. ludwig, k.e. geillinger, g.c. schoberth, j. essenwanger, j. stolz, h. daniel, h. witt, the journal of membrane biology 250 (2017) 171-182. [22] k. witkowska, k.m. smith, s.y. yao, a.m. ng, d. o'neill, e. karpinski, j.d. young, c.i. cheeseman, american journal of physiology. renal physiology 303 (2012) f527-539. [23] t. kimura, m. takahashi, k. yan, h. sakurai, plos one 9 (2014) e84996. [24] j.e. shima, t. komori, t.r. taylor, d. stryke, m. kawamoto, s.j. johns, e.j. carlson, t.e. ferrin, k.m. giacomini, american journal of physiology. renal physiology 299 (2010) f767-775. [25] y. hagos, d. stein, b. ugele, g. burckhardt, a. bahn, journal of the american society of nephrology: jasn 18 (2007) 430-439. [26] m.a. mcadams-demarco, j.w. maynard, a.n. baer, l.w. kao, a. kottgen, j. coresh, annals of the rheumatic diseases 72 (2013) 701-706. [27] y. bao, g. curhan, t. merriman, r. plenge, p. kraft, h.k. choi, annals of the rheumatic diseases 74 (2015) 1394-1398. [28] a. bahn, y. hagos, s. reuter, d. balen, h. brzica, w. krick, b.c. burckhardt, i. sabolic, g. burckhardt, the journal of biological chemistry 283 (2008) 16332-16341. [29] l. fagerberg, b.m. hallstrom, p. oksvold, c. kampf, d. djureinovic, j. odeberg, et al. molecular & cellular proteomics : mcp 13 (2014) 397-406. [30] c. schulz, c. fork, t. bauer, s. golz, a. geerts, e. schomig, d. grundemann, the biochemical journal 457 (2014) 243-251. [31] n. anzai, h. miyazaki, h. endou, seikagaku. the journal of japanese biochemical society 76 (2004) 101-110. [32] h. miyazaki, n. anzai, s. ekaratanawong, t. sakata, h.j. shin, p. jutabha, t. hirata, x. he, h. nonoguchi, k. tomita, y. kanai, h. endou, journal of the american society of nephrology : jasn 16 (2005) 3498-3506. [33] y. kanai, nihon rinsho. japanese journal of clinical medicine 66 (2008) 659-666. [34] s.m. gisler, s. pribanic, d. bacic, p. forrer, a. gantenbein, l.a. sabourin, a. tsuji, z.s. zhao, e. manser, j. biber, h. murer, kidney international 64 (2003) 1733-1745. [35] p. jutabha, n. anzai, k. hayashi, m. domae, k. uchida, h. endou, h. sakurai, journal of pharmacological sciences 116 (2011) 392-396. [36] o. kocher, n. comella, a. gilchrist, r. pal, k. tognazzi, l.f. brown, j.h. knoll, laboratory investigation; a journal of technical methods and pathology 79 (1999) 1161-1170. [37] j. park, j.o. kwak, b. riederer, u. seidler, s.p. cole, h.j. lee, m.g. lee, journal of the american society of nephrology : jasn 25 (2014) 726-736. [38] n. togawa, t. miyaji, s. izawa, h. omote, y. moriyama, american journal of physiology. cell physiology 302 (2012) c1652-1660. [39] t. shimizu, t. sugiura, t. wakayama, a. kijima, n. nakamichi, s. iseki, d.l. silver, y. kato, drug metabolism and disposition: the biological fate of chemicals 39 (2011) 2148-2154. [40] s.k. nigam, k.t. bush, g. martovetsky, s.y. ahn, h.c. liu, e. richard, v. bhatnagar, w. wu, physiological reviews 95 (2015) 83-123. [41] p.k. tan, t.m. ostertag, j.n. miner, scientific reports 6 (2016) 34995. [42] j. miner, p.k. tan, d. hyndman, s. liu, c. iverson, p. nanavati, d.t. hagerty, k. manhard, z. shen, j.l. girardet, l.t. yeh, r. terkeltaub, b. quart, arthritis research & therapy 18 (2016) 214. [43] z. shen, c. rowlings, b. kerr, v. hingorani, k. manhard, b. quart, l.t. yeh, c. storgard, drug design, development and therapy 9 (2015) 3423-3434. tan and miner admet & dmpk 5(2) (2017) 59-74 72 [44] p. jutabha, n. anzai, t. kimura, a. taniguchi, w. urano, h. yamanaka, h. endou, h. sakurai, journal of pharmacological sciences 115 (2011) 249-253. [45] p. jutabha, n. anzai, m.f. wempe, s. wakui, h. endou, h. sakurai, nucleosides, nucleotides & nucleic acids 30 (2011) 1302-1311. [46] t. chiba, h. matsuo, y. kawamura, s. nagamori, t. nishiyama, l. wei, et al. arthritis & rheumatology 67 (2015) 281-287. [47] m. sakiyama, h. matsuo, s. nagamori, w. ling, y. kawamura, a. nakayama, t. higashino, t. chiba, k. ichida, y. kanai, n. shinomiya, nucleosides, nucleotides & nucleic acids 35 (2016) 536-542. [48] a.a. el-sheikh, j.j. van den heuvel, j.b. koenderink, f.g. russel, british journal of pharmacology 155 (2008) 1066-1075. [49] n. dalbeth, m.e. house, g.d. gamble, b. pool, a. horne, l. purvis, a. stewart, m. merriman, m. cadzow, a. phipps-green, t.r. merriman, arthritis research & therapy 16 (2014) r34. [50] p.k. tan, j.e. farrar, e.a. gaucher, j.n. miner, molecular biology and evolution 33 (2016) 21932200. [51] r.j. johnson, y.y. sautin, w.j. oliver, c. roncal, w. mu, l. gabriela sanchez-lozada, b. rodrigueziturbe, t. nakagawa, s.a. benner, j comp physiol b 179 (2009) 67-76. [52] a.j. ahola, n. sandholm, c. forsblom, v. harjutsalo, e. dahlstrom, p.h. groop, g. finndiane study, kidney international (2017). [53] i. sluijs, m.v. holmes, y.t. van der schouw, j.w. beulens, f.w. asselbergs, j.m. huerta, et al. diabetes 64 (2015) 3028-3036. [54] d. yan, j. wang, f. jiang, r. zhang, t. wang, s. wang, d. peng, z. he, h. chen, y. bao, c. hu, w. jia, international journal of cardiology 214 (2016) 194-199. [55] j. white, r. sofat, g. hemani, t. shah, j. engmann, c. dale, et al. the lancet. diabetes & endocrinology 4 (2016) 327-336. [56] t. yamamoto, h. yokoyama, y. moriwaki, s. takahashi, m. suda, t. hada, k. higashino, european journal of clinical nutrition 44 (1990) 659-664. [57] t.r. merriman, arthritis research & therapy 17 (2015) 98. [58] t. dan, h. koga, european journal of pharmacology 187 (1990) 303-312. [59] e. leal-pinto, w. tao, j. rappaport, m. richardson, b.a. knorr, r.g. abramson, the journal of biological chemistry 272 (1997) 617-625. [60] n. anzai, y. kanai, h. endou, current opinion in rheumatology 19 (2007) 151-157. [61] d. dinour, n.k. gray, s. campbell, x. shu, l. sawyer, w. richardson, et al. journal of the american society of nephrology : jasn 21 (2010) 64-72. [62] a. ochi, t. takei, a. ichikawa, c. kojima, t. moriyama, m. itabashi, t. mochizuki, a. taniguchi, k. nitta, clinical and experimental nephrology 16 (2012) 316-319. [63] y. takiue, m. hosoyamada, m. kimura, h. saito, nucleosides, nucleotides & nucleic acids 30 (2011) 113-119. [64] f.m. anton, j. garcia puig, t. ramos, p. gonzalez, j. ordas, metabolism: clinical and experimental 35 (1986) 343-348. [65] e. lioudaki, e.s. ganotakis, d.p. mikhailidis, d.r. nair, current pharmaceutical design 16 (2010) 3815-3822. [66] c.g. wermuth, the practice of medicinal chemistry, 2nd ed., academic, amsterdam ; london, 2003. [67] i. walter-sack, j.x. de vries, a. von bubnoff, v. pfleilschifter, r. raedsch, european journal of medical research 1 (1995) 16-20. [68] n. robbins, s.e. koch, m. tranter, j. rubinstein, cardiovascular toxicology 12 (2012) 1-9. [69] p.k. tan, s. liu, e. gunic, j.n. miner, scientific reports 7 (2017) 665. admet & dmpk 5(2) (2017) 59-74 uric acid transporter inhibitors for gout doi: 10.5599/admet.5.1.387 73 [70] c. diaz-torne, n. perez-herrero, f. perez-ruiz, current opinion in rheumatology 27 (2015) 164-169. [71] j. poiley, a.s. steinberg, y.j. choi, c.s. davis, r.l. martin, c.a. mcwherter, p.f. boudes, i. arhalofenate flare study, arthritis & rheumatology 68 (2016) 2027-2034. ©2017 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ tan and miner admet & dmpk 5(2) (2017) 59-74 74 supplementary figure figure sp1. photomicrograph under polarized light of negatively birefringent monosodium uric acid crystals, the underlying cause of gouty inflammation manuscript doi: 10.5599/admet.5.2.389 135 admet & dmpk 5(2) (2017) 135-145; doi: 10.5599/admet.5.2.389 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper fluorescent organic cations for human oct2 transporters screening: uptake in cho cells stably expressing hoct2 malachy c. ugwu 1, 2 , ryan pelis 3 , charles o. esimone 2 & remigius u. agu 1* 1 biopharmaceutics and drug delivery lab college of pharmacy 5968 college street, po box 15000, halifax, ns b3h 4r2, canada 2 department of pharmaceutical microbiology & biotechnology faculty of pharmaceutical sciences nnamdi azikiwe university, near nafdac zonal lab, agulu, anambra state, nigeria. 3 department of pharmacology faculty of medicine dalhousie university p.o. box 15000
 halifax, ns 
 b3h 4r2 *corresponding author: e-mail: remigius.agu@dal.cab; tel.: +1-902-494-2092; fax: +1-902-494-1396 received: april 28, 2017; revised: june 05, 2017; published: june 22, 2017 abstract the aim of this study was to assess the suitability of amiloride, rhodamine 6g and rhodamine 123 as non radioactive substrates for characterizing hoct2 using cho cells. the uptake characteristics of these compounds were compared in wild-type (wt) and human organic cation transporter 2 (hoct2)-stably transfected chinese hamster ovary (cho) cells. all the compounds were accumulated by the cho-hoct2 cells. intracellular uptake of the compounds was higher in cho cells stably-expressing hoct2 compared to the wt. the uptake was concentration–dependent and saturable (except for rhodamine 123). the affinities of the compounds for the hoct2 (in descending order) were: amiloride (km = 72.63 12.02 µm) > rhodamine 6 g (km = 82.47  29.15 µm). uptake of amiloride in transfected cells was ph -dependent and significantly inhibited by hoct2 inhibitors (quinine, verapamil and quinidine). based on our kinetic data and other considerations, we recommend the use of amiloride for characterizing hoct2 transporters. keywords transporters; oct2; cho cells; amiloride; rhodamine; fluorescence introduction organic cations are positively charged amines or organic molecules with transient or permanent net positive charge(s) at physiological ph [1]. a wide variety of clinically used drugs (e.g. cimetidine, procainamide, quinidine, salbutamol) and endogenous bioactive amines (e.g. dopamine, histamine, nmethylnicotinamide, choline) are organic cations [1-2]. these charged molecules are transported across cell epithelia by carrier-mediated transport proteins, mainly organic cation transporters (octs) and proton/cation antiporters (mates) [3-6]. these transporters play pivotal roles in the absorption, distribution and disposition of cationic drugs in various tissues. the octs are members of the solute carrier family, slc22a [3-4, 7]. the slc22 family is sub-grouped according to substrates and based on the driving force for the cation transport mechanisms [1, 3-4]. one of such subgroups, which translocate organic http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:remigius.agu@dal.cab remigius agu et al. admet & dmpk 5(2) (2017) 135-145 136 cations and weak bases in an electrogenic-, sodium chlorideand proton gradient-independent manner, is the slc22a1-3 (oct1-3). these facilitative diffusion systems transport organic cations in both directions across the plasma membrane. other slc22 transporter subgroups are the sodiumand ph-dependent cation/carnitine transporters, slc22a4 (octn1) and slc22a5 (octn2) [8-10]. functionally, the octs and octns play a crucial role in the regulation of cellular functions. their involvement in cationic solute absorption, transport and excretion in the intestinal, placental, nasal and renal epithelia is well documented [10-11]. among the organic cation transporters, the oct2 is the transporter with the potential to significantly alter drug disposition. based on this fact, the us food and drugs administration agency (fda) and european medicines agency (ema) require that new drug candidates be screened for oct2 activity during early stages of drug development if active renal excretion is envisaged [12-13]. the transporter is mainly expressed in the basolateral membrane of renal proximal tubules and is also expressed in small intestine, lung, placenta, thymus, brain and the inner ear [2, 14-15]. although the role of the transporter in other tissues than the kidney is still under investigation, the oct2 is critical for the detoxification and elimination of drugs and endogenous compounds from the systemic circulation and thus a significant determinant of drug response and sensitivity [2, 4]. the most commonly used method for membrane transporter characterization is radioactivity. however, this approach is neither convenient nor cheap due isotope handling, disposal measures and regulatory restrictions. furthermore, there is potential health risk associated with the use of radiolabelled substrates if handled incorrectly. currently, there is a growing interest in the development of low-risk screening technologies for drug transporters characterization [16-17]. unlike radioactivity, fluorescent assays significantly improve laboratory safety with the opportunity for real-time monitoring using fluorimeters. currently, 4-di-1-asp is the main fluorescent probe for screening organic cation transporters. it has a permanent positive charge, high fluorescent intensity and stable over a range of ph and temperature. however, the compound is transported by all oct isoforms (oct1-2, octn1 and octn2). it is not among the compounds recommended by the international transporter consortium for investigating oct2 [18]. it is therefore relevant to explore the use of other fluorescent organic cation probes that may specifically interact with oct2. the objective of this study was to investigate the potential use of amiloride, rhodamine 123 and rhodamine 6g as specific fluorescent substrates for hoct2 using cho cells stably expressing the transporter. to the best of our knowledge, there is no published study on potential interaction between hoct2 stably expressed in cho cells and the fluorescent substrates described in this study. experimental chemicals recovery™ cell culture freezing medium, kaighns f12, 0.25% trypsin-edta, phenol red, zeocin™ selection reagent, hygromycin b, dmem-f12, salmon sperm dna, fetal bovine serum and phosphatebuffered-saline (pbs) were supplied by life technologies (burlington, on canada). triton x-100(r), verapamil, quinine, quinidine, corticosterone, bovine serum albumin (bsa), isopropanol, hanks’ balanced salt (hbss), and penicillin/streptomycin were purchased from sigma (st. louis, mo, usa). rhodamine 123, rhodamine 6g and amiloride were procured from santa cruz biotech inc. (dallas, texas, usa). bicinchoninic acid (bca) protein assay kit was from millipore (billerica, ma, usa) for protein assay according to the method described by smith et al [19]. admet & dmpk 5(2) (2017) 135-145 fluorescent organic cations for human oct2 transporters screening doi: 10.5599/admet.5.2.389 137 chinese hamster ovary (cho) cell culture and stable expression of hoct2 cho cell was chosen for stable expression of hoct2 as it is one of the models / cultured cell lines recommended by the international transporter consortium for studying drug transporter interactions so as to assess uptake and or efflux by single recombinant transporters [18]; and there is no published study on potential interaction between hoct2 stably expressed in cho cells and the fluorescent substrates .the cho cells used were from atcc (manassas, va, usa). for stable expression of hoct2, cells were electroporated with 10 µg of hoct2-pcdna3.1 plasmid dna and 10 µg of salmon sperm dna in a cuvette (4-mm gap) using a btx ecm 630 electroporator (harvard apparatus, st. laurent, quebec, canada) with settings of 1050 microfarad, 260 v, and at no resistance. forty -eight hours after the electroporation, positively transfected cells were selected with 200 µg/ml hygromycin b. the cho wild type were not transfected with hoct2 as the cells will be used as the control. the cho cells were grown in ham's f-12 (kaighn's modification) containing 10 % fbs, 1 % penicillinstreptomycin, 200 µg/ ml hygromicin b (transfected) and 100 µg /ml zeocin (wild-type), respectively. both cell types were maintained in a humidified atmosphere (5 % co2/95 % o2). the cells were passaged every 2-3 days. cells within passage 4–35 were used for the studies. uptake studies confluent cho-hoct2 and wild –type cells cultured in 24, 48or 96 well plates were used for uptake experiments. the uptake buffer was prepared with hbss supplemented with 25 mm glucose and 20 mm hepes (ph 7.4). the cells in each well were pre-incubated in substrate-free uptake buffer (200 µl) at 37 °c for 15 min. time-dependent studies were conducted by incubating cell monolayers with 10 µm cationic compounds for up to 60 min at 37 c. uptake studies were initiated by replacing the pre-incubation substrate-free buffer with buffer solutions containing 200 µl of rhodamine 123, rhodamine 6g, and amiloride, respectively. based on preliminary studies, the experiments were terminated at 1, 2, 5, 15, 30 and 60 min by immediately removing the compounds and washing the cells three times with ice-cold hbss. subsequently, the cells were dissolved in 200 μl of 1% triton x-100r containing 0.1 m naoh on a shaker for 1–2 hours at 37 °c. furthermore, uptake of the fluorescent substrates was investigated at different concentrations (1-3000µm) and ph values (5.0-8.5). in studies involving inhibition, the cells were preincubated at 37 °c with or without inhibitors for 15 min prior to the uptake studies. each set of experiments was performed with cells of the same age and passage. sample analyses rhodamine-123, rhodamine-6g, and amiloride in samples were quantified with modulus single tube multimode florescence reader (fluorimeter model 9200, turner bio systems, sunnyvale, ca, usa) using the green, blue and uv filter configurations for rhodamines, and amiloride, respectively. protein content of cell lysates was measured using bovine serum albumin as standards according to bca protein assay protocol so as to normalize the cell population/density in each plate [19]. data analyses. the saturable uptake of the fluorescent substrates was analyzed by assuming michaelis-menten type carrier-mediated transport. the maximum transport rate (vmax), the michaelis constant (km), and ic50 for the inhibition studies were determined by non-linear regression using prism(r) 5.0 (graphpad, san diego, ca, usa). passive diffusion was accounted for by subtracting the rate constant for the compounds passive diffusion in cho-wild cells from cho-hoct2-expressing cells. unless stated otherwise, all experiments remigius agu et al. admet & dmpk 5(2) (2017) 135-145 138 were performed in triplicates and data presented as mean ± sd. where appropriate, statistical significance of the results was determined with anova using instat® 3.0 (graphpad, san diego, ca, usa). p < 0.05 was considered significant. the specific uptake of the dyes was calculated by subtracting the unspecific uptake in wild cho cells from the total uptake in cho-hoct2 cells. results and discussion considering the fluorescence characteristics of rhodamine-123 and rhodamine-6g (high quantum yield, long emission and excitation wavelengths) their extensive current biomedical applications, and the fact that these compounds are organic cations [17], we selected them for this study. furthermore, rhodamine 123 was used based on the fact that it is an organic cation transporter used for characterizing oct2 [20] amiloride was chosen because it is the only oct2 substrate with desirable fluorescent characteristics listed by the international transporters consortium for oct2 transporter characterization [18]. the hoct2 is a very important oct isoform that exhibits clinically relevant genetic polymorphisms, participates in clinical drug–drug interactions and alters the disposition and excretion of some drugs [18]. figure 1. time-dependent uptake of the fluorescent compounds into cho-hoct2. cell layers were incubated with (10 m) of each compound as indicated. subsequently, the cells were washed with ice-cold uptake buffer followed by extraction with triton x-100/0.1 m naoh. data represent mean ± sd; n = 3. figure 1 shows the time–dependent uptake of the substrates by cho cells stably expressing hoct2. uptake of the three compounds was linear up to approximately 7 min; after which a gradual plateau was observed. based on these observations, an incubation period of 5 min was chosen for subsequent experiments. the effect of extracellular ph on the uptake of amiloride, rhodamine 6g and rhodamine 123 by cho-hoct2 and cho-wt cells is summarized in figure 2. the three compounds accumulated in the cells in a ph-dependent manner. highest uptake was observed at alkaline ph range of 8.0 8.5 for cho cells stably-expressing hoct2. conversely, decreased uptake was seen in acidic ph range. for cho-wt cells, no clear pattern of ph-dependency was deciphered. the observed ph effect suggests that ph-sensitive organic cation transporters may have contributed to the observed solute transport as no clear ph-dependent admet & dmpk 5(2) (2017) 135-145 fluorescent organic cations for human oct2 transporters screening doi: 10.5599/admet.5.2.389 139 pattern was seen in cho-wt cells. the ph effect for amiloride was comparable and in agreement with published data [9,17,21]. these studies involving the uptake of 4-di-1-asp by octs in calu-3, a549, 16hbe14oand caco-2 cells demonstrated higher substrate accumulation under alkaline extracellular ph range. figure 2. effect of extracellular ph on uptake in cho cells stably expressing hoct2 and cho wild type. uptake of the compounds (10m) was evaluated at 37 °c for 7 min using cho-hoct2, and cho-wt cells. subsequently, the cells were washed with ice-cold uptake buffer followed by extraction with triton x-100/0.1 m naoh. data represent mean ± sd; n = 3. if the uptake of the fluorescent cations was mediated by hoct2, concentration-dependent and saturable uptake would be expected. figure 3 shows the effect of concentration on amiloride, rhodamine 123, and rhodamine 6g uptake by cho-hoct2 and cho-wt cells. a clear concentration–dependent effect was observed for cho-hoct2 cells. at various concentrations, amiloride, rhodamine 6g and rhodamine 123 were accumulated to a greater amount by cho-hoct2 cells compared to cho-wt cells. uptake of the compounds in cho-hoct2 cells (except rhodamine 123) was saturable. the calculated kinetic parameters are: km = 72.6312.02 µm, vmax = 0.0270.001 µm/mg protein/min (amiloride) and km = 82.4729.15 µm, vmax= 0.0190.002 µm/mg protein/min (rhodamine 6g), respectively. we were unable to determine the kinetic parameters for rhodamine 123 due to excessive non-specific binding to the 96-well plates at higher concentrations. these results indicate that amiloride and rhodamine 6g appear to be better substrates for hoct2 compared to rhodamine 123. the rank order of affinity of the dyes based on michaelis-menten modelling is: amiloride (km = 72.63  12.02 µm) > rhodamine 6g (km = 82.47  29.15 µm) > rhodamine 123 (km = nd). as km is a measure of transporter-substrate affinity [22-23], higher km implies poorer binding/affinity. biermann et al., [2] used km to compare the affinities of different fluorescent compounds to hoct2 in transfected hek-293. according to the authors, amiloride and 4-di-1-asp had significantly higher uptake rates in hoct2 cells compared to wild type cells. the km obtained in this study was remigius agu et al. admet & dmpk 5(2) (2017) 135-145 140 comparable to the value (95 µm) obtained by biermann et al using hek-293 stably expressing hoct2. amiloride is not only taken up by cells, but can be transported. this behaviour may be significantly affected by the degree of ionization of the compound. the guanidine moiety of the compound is protonated at physiologic ph, and it is this positively charged species that makes it an organic cation [24]. as a weak base, the degree of ionization plays a critical role in its binding to organic cation transporters [25]. this implies that permeation at ph much lower than the pka of the compound may be by passive diffusion. for the cationic rhodamines investigated, there was no clear-cut ph effect within the range investigated. this was not surprising because we previously showed that fluorescence intensity of rhodamine-6g is characteristically acid-dependent due to molecular changes and transitions [17]. under our experimental conditions, it was not possible to differentiate the effect of this phenomenon on uptake from the impact of ph changes on uptake. figure 3. concentration-dependent intracellular uptake of the fluorescent compounds in cho-hoct2 and cho-wt. uptake was evaluated at 37 °c for 7 min. subsequently, the cells were washed with ice-cold uptake buffer followed by extraction with triton x-100/0.1 m naoh. data represent mean ± sd; n = 3. if the uptake of the cationic fluorochromes were mediated by hoct2, one would expect hoct2 inhibitors to differentially reduce the uptake of these compounds in transfected cells (figures 4-7). we used six inhibitors (figure 4) at concentration of 1mm for initial screening after which the three most potent (quinine, verapamil and quinidine hcl) were selected for a concentration dependent inhibition studies (figure 5). these inhibitors have been reported by other published literature as potent and oct specific inhibitors [9, 10, 17]. figure 4 shows the effect of hoct2 and octn1-2 inhibitors and substrates on amiloride, rhodamine 6g and rhodamine 123 uptake in cho-hoct2 cells. at 1mm, quinidine, quinine and verapamil consistently inhibited amiloride, and rhodamine 6g uptake. expectedly, octn inhibitors (nmn, choline and guanidine) had no significant effect on the uptake of these compounds (p0.5), an indication that their uptake was likely due to hoct2 transporters. to confirm the affinity data obtained with michaelis-menten kinetics, concentration-dependent inhibition of amiloride, rhodamine 6g, and rhodamine 123 using hoct2 inhibitors (quinine, quinidine, and verapamil) were investigated (figures 5-7). figure 5 shows the effect of increasing concentrations of quinidine on the fluorescent compounds uptake. admet & dmpk 5(2) (2017) 135-145 fluorescent organic cations for human oct2 transporters screening doi: 10.5599/admet.5.2.389 141 the ic50s for the inhibition were 891 µm (amiloride), 315 µm (rhodamine 123) and 1137 µm (rhodamine 6g). similarly, figure 6 summarizes concentration-dependent inhibition of the cationic dyes by quinine. the ic50s for the inhibition were 309.6, 279, and 206 µm for amiloride, rhodamine 123 and rhodamine 6g, respectively. data for inhibition studies with verapamil is shown in figure 7. verapamil concentrationdependently inhibited amiloride (ic50 = 486 µm), rhodamine 123 (ic50 = 302 µm) and rhodamine 6g (ic50 = 583 µm). inhibition of the cationic compounds by quinine, quinidine and verapamil, suggests that these compounds may share the same substrate recognition sites. previous reports from our laboratory and others have shown an interaction between quinine, verapamil, quinidine and organic cation transporters [2, 9, 10, 17]. quinine blocked et+ uptake by cells overexpressing hoct1 or hoct2 with high affinity [26]. similarly, quinine and verapamil inhibited hoct1 and hoct2-mediated uptake of rhodamine 123 with higher inhibitory effects observed for hoct2 [20]. the apparent discrepancy in the ic50 values could be attributed to the fact that amiloride, and rhodamines bind to different, but partially overlapping sites in the binding pockets of the hoct2. the hoct2 was previously reported to be polyspecific, as such can accept compounds of different sizes and molecular shapes [2, 4, 27]. thus it exhibits large variations in affinity and turnover for different compounds. the observed ic50 values for quinine and verapamil varied with the reported ic50 of 23 µm of quinine [28] and 85 µm of verapamil [29] in hek 293 cells expressing oct 2 using n-methylpyridinium as substrate. similarly the observed ic50 of quinidine is different from 13.3 µm in a cho cell expressing oct2 using n-methylpyridinium as substrate [30]. the presence of various substrate and inhibitor binding sites and the complex interactions between different sites explains why largely different ic50 values are obtained for individual transporters when different substrates are used for transport measurements [2,4]. volk et al. [31] showed that the substrate binding site of roct2 was complex, and it exposes different interactive domains for different substrates and inhibitors. a similar analogy could be true for hoct2 because human oct2 show about 90 % amino acid identities to the rat oct2 [32]. several of these amino acids in the transporter transmembrane helix contain highand lowaffinity substrate and/or inhibitor binding sites that are involved in substrate and/or inhibitor binding of the oct transporters [4,22]. in line with the conformational analogy, thévenod and co-workers maintained that octs acquire minimum of three conformational states: one state with an outward-open binding cleft containing an innermost binding pocket where transported cations bind and additional more peripherally localized binding sites for organic cations which may overlap with domains in the innermost binding pocket [31]. they reported that structurally different transported cations induce substrates due to the interaction of different substrates with partially different amino acids [31]. in addition to overlapping sites, they also reported that the inhibitor potencies of oct ligands are highly dependent on the substrate employed for uptake studies. the uptake of the fluorescent compounds (amiloride, rhodamine 6g and rhodamine 123) was oct2mediated. this was evident in the concentrationdependent uptake of the compounds, as well as their uptake inhibition by classical organic cationic inhibitors. remigius agu et al. admet & dmpk 5(2) (2017) 135-145 142 figure 4. inhibition of the fluorescent compounds uptake in cho cells stably expressing hoct2 by oct/octn inhibitors. cells were pre-incubated with inhibitors at 1 mm for 25 min followed by an additional 7 min incubation with the respective fluorescent compounds. subsequently, the cells were washed with ice-cold uptake buffer followed by extraction with triton x-100/0.1 m naoh. data represent mean ± sd; n = 3. figure 5. inhibition of the fluorescent compounds uptake in cho cells stably expressing hoct2 by quinidine. cells were pre-incubated with inhibitor (quinidine) for 25 min followed by additional 7 min incubation with the respective dyes. subsequently, the cells were washed with ice-cold uptake buffer followed by extraction with triton x-100/0.1 m naoh. data represent mean ± sd; n = 3. admet & dmpk 5(2) (2017) 135-145 fluorescent organic cations for human oct2 transporters screening doi: 10.5599/admet.5.2.389 143 figure 6. inhibition of the fluorescent compounds uptake in cho cells stably expressing hoct2 by quinine. cells were pre-incubated with inhibitor (quinine) for 25 min followed by an additional 7 min incubation with the respective dyes. subsequently, the cells were washed with ice-cold uptake buffer followed by extraction with triton x-100/0.1 m naoh. data represent mean ± sd; n = 3. figure 7. inhibition of the fluorescent compounds uptake in cho cells stably expressing hoct2 by verapamil. cells were pre-incubated with inhibitor (verapamil) for 25 min followed by additional 7 min incubation with the respective dyes. subsequently, the cells were washed with ice-cold uptake buffer followed by extraction with triton x-100/0.1 m naoh. data represent mean ± sd; n = 3. remigius agu et al. admet & dmpk 5(2) (2017) 135-145 144 conclusions in this study we investigated the potential use of amiloride, rhodamine 6g and rhodamine 123 as nonradioactive probes for characterizing human organic cation transporter isoform, hoct2 using cho cells stably-expressing the transporter. our data suggest that the investigated fluorescent organic cations were efficiently taken up by the cho cells. however, based on their fluorescence, affinities (km/ic50s), aqueous solubility, ease of use at high concentrations and recognition as hoct2 substrates by fda and european union and the possibility to be used for in vivo human studies, we recommend the use of amiloride as an acceptable non-radioactive substrate for screening hoct2 expression in cells using fluorimeters. although amiloride is a substrate for organic cation transporters, it also inhibits sodium reabsorption in the kidney via the blockade of na + /h + antiporter located on the apical side of the proximal tubules. from drug interaction perspective, care should be taken in geriatrics or patients with hyperkalaemia or those receiving other potassium-sparing agents when using this drug. furthermore, considering the mechanism of action of amiloride and the role and wide expression of oct2 in kidney, the use of this drug along with clinically used potent oct2 inhibitors (e.g. metformin), should be used with caution to avoid clinically significant interactions – in this case hypoglycaemic (reduced metformin clearance) or elevated amiloride concentration and duration (due to metformin inhibition), which may be an advantage considering amiloride’s short duration of action. it is also important to note that oct2 transporters are mainly expressed basolaterally in the kidney tissues, which makes individuals with polymorphisms of this transporter prone to clinically significant drug toxicity due to reduced renal clearance. overall, observed clinical manifestation of genetic polymorphisms involving amiloride may depend on whether the genes involved affects na+/h+ antiporter located apically or oct2 located basolaterally. acknowledgements: this project was funded with a grant from dalhousie university pharmacy endowment fund. references [1] g. ciarimboli. xenobiotica, 38 (2008) 936–971. [2] j. biermann, d. lang, v.gorboulev, h. koepsell, a. sindic, r. schroter, a. zvirbliene, h.pavenstadt, e.schlatter, g.ciarimboli. am. j. physiol .cell physiol . 290 (2006) 1521–1531. [3] u. anand, a. parikh, m.c. ugwu, r. u. agu. future med. chem. 6(12) (2014) 1381–1397. [4] a.t. nies, h. koepsell, k.damme, m. schwab. handbook exp. pharmacol. 201 (2011) 106-134. [5] w. m.suhre, s.ekins, c. chang, p. w. swaan, s. h.wright. mol. pharmacol. 67 (2005) 1067–1077. [6] a.yonezawa, k. inui. biochem. pharmacol. 81 (2011) 563–568. [7] m.k. kim, c. shim. arch. pharm. res. 29(7) (2006) 605-616. [8] g. ciarimboli, k. struwe, p.arndt, v. gorboulev, h. koepsell, e. schlatter, j.r. hirsch. j. cell. physio. 201 (2004) 420–428. [9] j. j.salomon, s.endter, g.tachon, f.falson, s. t. buckley, c. ehrhardt. eur. j. pharm. biopharm. 81 (2012) 351–359. [10] d. shao, e. massoud, u. anad, a.parikh, e. cowley, d. clark r.u. agu. ther. drug del. 4 (2013) 1-13. [11] c. bosquillon. j. pharma. scs. 99 (2010) 2240–2255. [12] ema guideline on the investigation of drug interactions, cpmp/ewp/560/95/rev. 1 corr. 2. www.ema.europa.eu/docs/en_gb/document...guideline/2010/.../wc500090112.pdf (accessed on april 27, 2017). [13] fda guidance for industry drug interaction studies study design, data analysis, implications for dosing, and labelling recommendations. https://www.fda.gov/downloads/drugs/guidances/ucm292362.pdf (accessed on april 27, 2017). https://www.fda.gov/downloads/drugs/guidances/ucm292362.pdf admet & dmpk 5(2) (2017) 135-145 fluorescent organic cations for human oct2 transporters screening doi: 10.5599/admet.5.2.389 145 [14] m.r koehler, b. wissinger, v. gorboulev, h. koepsell and m. schmidt. cytogenet. cell genet. 79 (1997) 198–200. [15] h. motohashi, y. sakurai, h. saito, s. masuda, y. urakami, m. goto, a., fukatsu, o. ogawa, k. inui. j am. soc. nephrol. 13 (2002) 866–874. [16] e. rytting, j.bryan, m. southard, k.l. audus. biochem. pharmacol. 73 (2007) 891– 900. [17] m.c. ugwu, a. oli, c.o. esimone, r.u. agu. j. pharmacol. toxicol. methods 82 (2016) 9–19. [18] k.m. giacomini, s.m. huang, d.j. tweedie, l.z. benet, k.l. brouwer, x. chu, a. dahlin, r. evers, v. fischer, k.m. hillgren, k.a. hoffmaster, t. ishikawa, d. keppler, r.b. kim, c.a. lee, m. niemi, j.w. polli, y. sugiyama, p.w. swaan, j.a. ware, s.h. wright, s.w. yee, m.j. zamek-gliszczynski, l. zhang. nat. rev. drug discov. 9 (2010) 215-236. [19] p. k. smith, r. i. krohn, g. t. hermanson, a. k. mallia, f. h. gartner, m. d. provenzano, e. k. fujimoto, n. m. goeke, b. j olson, d. c. klenk. anal. biochem. 150 (1985) 76-85. [20] e. jouan, m. levee, c. denizot, g. d. violante, o. fardel. fundam. clin. pharmacol. 28 (2014) 65–77. [21] c. macdonald, d. shao, a. oli, r. u. agu. j. drug targeting 21 (2013) 97–106. [22] c.volk, v.gorboulev, a. kotzsch, t.d.muller, h. koepsell. mol. pharmacol. 76 (2009)275–289. [23] t.yasujima, k. ohta, k.inoue, m. ishimaru, h. yuasa. drug metabol. disp. 38 (2010) 715-721. [24] t.r. kleyman, e.j. cragoe. j. membr biol. 105 (1988) 1-21. [25] w.m. barendt, s.h. wright. j. biol. chem. 277 (2002) 22491-22496. [26] w.lee, m. reichold, b.edemir, g.ciarimboli, r.warth, h. koepsell, f.the´venod. am j. renal physiol. 296 (2009) 1504–1513 [27] h. koepsell, k. lips, c. volk. pharm. res. 24 (2007)1227–1251. [28] j. müller, k.s. lips, l. metzner, r.h. neubert, h. koepsell, m. brandsch. biochem. pharmacol. 70 (2005) 1851-1860. [29] o. zolk, t.f. solbach j. könig m.f. fromm. naunyn schmiedebergs arch. pharmacol. 379 (2009) 337348. [30] x. ming, w. ju, h.wu , r.r. tidwell, j.e hall, d.r thakker. drug metab. dispos. 37 (2009) 424-430. [31] c.volk, v. gorboulev, t. budiman, g. nagel, h. koepsell. mol . pharmacol. 64 (2003) 1037–1047. [32] m. hayer-zillgen, m. bruess, h.b., nisch. br.. j. pharmacol. 136(2002) 829 -836. ©2017 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ doi: 10.5599/admet.4.3.337 186 admet & dmpk 4(3) (2016) 186-211; doi: 10.5599/admet.4.3.337 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index review tyrosine kinase inhibitors for egfrand alk-mutated non-small cell lung cancer jonathan r.thompson 1 , smitha p. menon 2 and grace k. dy 3 * 1 mazie froedtert willms &sue froedtert cancer fellow at froedtert hospital, division of hematology and oncology, medical college of wisconsin, milwaukee, wi 2 division of hematology and oncology, medical college of wisconsin, milwaukee, wi 3 department of medicine, roswell park cancer institute, buffalo, ny *corresponding author. e-mail: grace.dy@roswellpark.org. received: august 15, 2016; revised: september 11, 2016; published: september 30, 2016 abstract discovery of the epidermal growth receptor (egfr) activating mutations and anaplastic lymphoma kinase (alk) rearrangements has expanded the therapeutic landscape in non-small cell lung cancer (nsclc). survival outcomes for patients with these mutations have improved dramatically with egfr and alk tyrosine kinase inhibitors (tkis). multiple generations of egfr and alk tkis have been rapidly developed, and patients and clinicians now have several options for firstand second-line treatments. while these small molecule tkis have some similarities in therapeutic and pharmacologic profiles, the differences can be clinically substantial, allowing tailored treatment for each unique patient. this review details the clinical efficacy, pharmacology, safety profiles, cns penetration, and mechanisms of resistance of the four egfr tkis and three alk tkis that are currently approved by the united states food and drug administration (us fda). keywords epidermal growth factor receptor; anaplastic lymphoma kinase; targeted therapy introduction lung cancer is the leading cause of cancer mortality worldwide, leading to 1.6 million deaths annually [1]. non-small cell lung cancer (nsclc) is a heterogeneous disease that accounts for 85 % of all lung cancer diagnoses [2]. most patients are diagnosed at an advanced stage, where 5-year survival is less than 5 % [3]. within the past decade, numerous somatic molecular mutations have been discovered that drive oncogenesis in nsclc. protein kinase inhibitors that target these mutations improve response rates and survival compared to standard systemic chemotherapy. the most common clinically relevant “driver” mutations encountered are the epidermal growth factor receptor (egfr) activating mutations, which are present in 30-40 % of asian nsclc patients and 10 % of caucasian patients with nsclc [4,5]. rearrangements in the anaplastic lymphoma kinase (alk) gene lead to fusion proteins, which drive cellular proliferation. the alk gene rearrangement occurs in 3 to 7 % of all nsclc patients [6]. egfr activating mutations and alk gene rearrangements create constitutively active protein kinases for which there are multiple highly active tyrosine kinase inhibitors (tkis) available for treatment. these tkis have dramatically impacted clinical outcomes for patients with nsclc who previously carried poor http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:grace.dy@roswellpark.org admet & dmpk 4(3) (2016) 186-211 tki for egfrand alk-mutated cell cancer doi: 10.5599/admet.4.3.337 187 prognoses. the standard overall survival for all patients with advanced nsclc ranges from 4 to 12 months [7-10], but median overall survival in patients with egfror alk-mutated nsclc now approaches 3 to 4 years or more [11-15]. currently, there are four egfr tkis and three alk tkis that are widely approved for clinical use. the first-generation egfr tkis, gefitinib and erlotinib, reversibly inhibit the intracellular catalytic domain of the egfr tyrosine kinase. afatinib is a second-generation egfr tki that irreversibly inhibits the egfr tyrosine kinase domain. osimertinib is a third-generation egfr tki specifically designed to inhibit t790m-mutated egfr as well as other egfr activating mutations. crizotinib was the first tki to show activity in alk-positive nsclc, however, it is more potent at inhibiting met than alk. ceritinib and alectinib are tkis that are more potent and specific for the alk tyrosine kinase. in this article, the relevant egfr and alk mutations and the tkis which inhibit these mutated proteins will be reviewed. the clinical efficacy, safety, and pharmacology of these tkis will be detailed in order to promote understanding of this promising and rapidly evolving field. egfr tyrosine kinase inhibitors egfr protein and activating mutations egfr is a tyrosine kinase receptor of the erbb receptor family. it is composed of an extracellular domain (n-terminus) that binds ligands including egf and tgf-α, a transmembrane domain, and an intracellular domain (c-terminus) that is responsible for intracellular signaling through tyrosine kinase activity [16,17]. upon binding to its ligand on the extracellular domain, the receptor dimerizes, leading to autophosphorylation of the intracellular tyrosine-kinase domain. autophosphorylation of the intracellular domain is followed by binding of adaptor proteins and signaling through down-stream pathways. these pathways include the ras/raf/mek/erk, phosphoinositide 3-kinase (pi3k)/protein kinase b (akt), and janus kinase (jak)/signal transducer and activator of transcription (stat) pathways, which contribute to cell migration, survival, and proliferation [18,19]. activating mutations in the egfr gene were originally reported in 2004, and were associated with female sex, asian descent, absence of tobacco exposure and adenocarcinoma histology [4,20,21]. the egfr gene exists on chromosome 7, and activating mutations are predominantly located on exons 18 to 22. at diagnosis, 90 % of patients with an activating egfr mutation will have either an in-frame deletion in exon 19 or a l858r missense mutation on exon 21. these mutations all cause prolonged, ligand-independent signaling of the egfr [22]. due to the dependence on the constitutively activated egfr, these malignant clones are particularly susceptible to egfr tkis. the receptor targets, clinical indications, pharmacokinetics, pharmacodynamics and cns penetration of the fda approved egfr tkis are summarized in table 1. erlotinib clinical efficacy prior to the widespread knowledge of egfr activating mutations as predictive biomarker of treatment sensitivity to egfr tkis, erlotinib gained fda approval in november 2004 for treatment of advanced nsclc, regardless of tumor mutation status, after failure of firstor second-line chemotherapy. in a phase iii, placebo-controlled randomized study, erlotinib provided advantages in progression-free survival (pfs) (2.2 months vs 1.8 months, hazard ratio, 0.61;p<0.001) and overall survival (os) (6.7 months vs 4.7 months, hazard ratio, 0.71; p<0.001) compared to placebo [25]. erlotinib was next fda-approved for maintenance treatment of nsclc in april 2010 for patients whose disease had not progressed after four cycles of platinum-based chemotherapy based on the saturn trial. the randomized, double-blind, grace dy et al. admet & dmpk 4(3) (2016) 186-211 188 placebo-controlled study of 889 patients with stage iiib/iv nsclc who received erlotinib or placebo after receiving first-line platinum-based chemotherapy demonstrated improvements in pfs (12.3 weeks vs 11.1 weeks, hr 0.71; p<0.001) as well as os (12 months vs 11.1 months, hr 0.81; p=0.0088) for erlotinib maintenance. patients with activating mutations of egfr experienced more dramatic pfs and os benefits [26]. table 1. egfr tki characteristics egfr tki name erlotinib gefitinib afatinib osimertinib primary targets wild type egfr deletion 19 l858r wild type egfr deletion 19 l858r wild type egfr deletion 19 l858r t790m her2/3/4 deletion 19 l858r t790m type of inhibition reversible reversible irreversible irreversible clinical approval first-line treatment of advanced egfrmutated nsclc first-line treatment of advanced egfrmutated nsclc first-line treatment of advanced egfrmutated nsclc second-line treatment of advanced egfrmutated nsclc egfr mutation ic50 (in vitro)[23] 23 nm (del 19) 39 nm (l858r) 30 (del 19) 100 nm (l858r) 0.2 nm (del19) 0.2 nm (l858r) 9 nm (del 19) 12 nm (l858r) t790m ic50 (in vitro) [22,24] 1600 nm (del 19) >10000 nm (l858r) >5000 nm 141 nm (del 19) 196 nm (l858r) 3 nm (del 19) 13 nm (l858r) drug interactions proton pump inhibitors h2 antagonists cyp3a4 inducers/inhibitors proton pump inhibitors h2 antagonists cyp3a4 & cyp2d6 inducers/inhibitors p-gp inducers/inhibitors cyp3a4 inducers/inhibitors p-gp inducers/inhibitors food interactions food increases bioavailability. take on empty stomach. none high-fat meals decrease bioavailability. take on empty stomach. none metabolism liver cyp3a4 liver cyp3a4 cyp2d6 non-enzyme catalyzed michael adduct formation to electron-rich small molecules and proteins liver cyp3a4 method of elimination 83 % feces (1 % unchanged drug) 8 % urine 86% feces 4% urine 85% feces (89% unchanged) 4% urine 68 % feces (2 % unchanged) 14 % urine cns penetration ~5 % cns response achieved in patients refractory to gefitinib. ~1 % csf concentration of 1nm in single case report 0.2-1 % in case reports. higher clinical response witnessed than first generation tkis. abbreviations. ic50: concentration at which drug inhibits 50 % of receptor’s activity; nm: nanomolar; cyp: cytochrome p450; p-gp: p-glycoprotein as knowledge of the egfr activating mutations developed, development of companion diagnostic assays for a predictive biomarker of treatment sensitivity utilized in registrational studies facilitated fdaapproval of this agent for first-line treatment of patients with metastatic nsclc whose tumors harbor egfr exon 19 deletions or the l858r mutation in may 2013. eurtac, the open-label, randomized phase 3 trial demonstrated improvement in pfs (9.7 months vs 5.2 months, hr 0.37; p<0.001) for erlotinib compared admet & dmpk 4(3) (2016) 186-211 tki for egfrand alk-mutated cell cancer doi: 10.5599/admet.4.3.337 189 with standard first-line chemotherapy. a benefit in os was not demonstrated, but most patients in the chemotherapy group crossed over to receive erlotinib upon disease progression [27]. multiple subsequent phase iii trials have shown that erlotinib improves pfs compared to conventional platinum-based chemotherapy in first-line treatment of nsclc [28,29]. clinical safety skin rash is the most common toxicity encountered with erlotinib use. all-grade rash occurs in 61 to 80 % and grade 3 or 4 rash occurs in 2 to 13 % of patients. the classic erlotinib (or any egfr-inhibitor) rash is an acneiform rash that occurs on the scalp, cheeks, nose, perioral area, and upper trunk. the rash typically manifests within the first 2 weeks of treatment but can occur as late as 2 months after starting therapy [30]. less commonly, bullous eruptions can occur. toxic epidermal necrolysis (ten) and stevensjohnson syndrome(sjs) have developed in patients taking erlotinib. diarrhea is another common adverse event caused by erlotinib. any grade diarrhea occurs in 21 to 57 % and grade 3 or 4 diarrhea occurs in 1-6 %. rarely (<5 %), patients can develop cerebrovascular events, gastrointestinal perforation, fatal pneumonitis, liver/renal failure, or corneal perforation. patients on warfarin should have inr monitored closely as fatal hemorrhage has occurred in patients on both erlotinib and warfarin. dose reduction is necessary in 19 to 21 % of patients (7-12 % rash, 5-7 % diarrhea), and treatment interruption occurs in 27 % (14 % rash, 6 % diarrhea). treatment discontinuation is rarely necessary, occurring in only 3 to 6 % of patients [11,25,27,28]. pharmacology erlotinib hydrochloride, 6, 7-bis(2-methoxyethoxy)-n-(3-ethynyl-phenyl) quinazolin-4-amine [cp-358774 , osi-t774, tarceva, (genentech)] reversibly inhibits the egfr tyrosine kinase at the atp binding domain with high selectivity and potency. erlotinib inhibits purified egfr kinase with an ic50 of 2 nanomolar (nm) and inhibits egfr autophosphorylation with an ic50 of 20 nm in intact tumor cells. erlotinib binds more strongly to egfr harboring l858r or del 19 than it does to wild-type egfr [31], but it binds to several offtarget kinases, including cyclin g-associated kinase (gak), ste20-like kinase (slk) and serine/threonine kinase 10 (stk10). the inhibition of stk10 enhances lymphocyte migration and cytokine release, which contributes to the dermatologic toxicity of the medication [32]. erlotinib carries a recommended oral starting dose of 150 mg daily. the drug reaches peak plasma levels approximately 4 hours after administration, and the volume of distribution is 232 l [33]. bioavailability of erlotinib is 60 % but increases to nearly 100 % if administered with food [34]. the effect of food on erlotinib absorption appears to have inter-and intra-individual variability, and erlotinib is to be taken at least 1 hour prior to or at least 2 hours after eating [35]. erlotinib has multiple drug interactions, and van leeuwen et al. have previously reviewed common tki drug-drug interactions in detail [170]. gastric ph is a significant determinant in erlotinib absorption. as a weak base, the drug can exist in either ionized or non-ionized forms. at an elevated ph, the drug shifts to its less-soluble, non-ionized form, and drug absorption decreases. therefore, proton pump inhibitors and other drugs that increase gastric ph lead to significantly decreased serum levels of erlotinib and can decrease erlotinib auc by 46 % and maximum erlotinib serum concentration by 61 %. interestingly, this can be overcome by administering erlotinib with acidic beverages like cola [36]. upon absorption, erlotinib is heavily protein-bound (93-95 %) to albumin and α-1 acid glycoprotein, and its median half-life is 36 hours [37]. there is no significant association with drug clearance and gender, weight, or age [33]. erlotinib is primarily metabolized to its active metabolite, desmethyl erlotinib, in the liver through cyp3a4 and to a lesser degree by cyp1a2 and extrahepatic cyp1a1. therefore, induction and grace dy et al. admet & dmpk 4(3) (2016) 186-211 190 inhibition of cyp3a4 activity through medications like rifampin and ketoconazole, respectively, impacts serum erlotinib concentrations. tobacco is a well-known cyp inducer and increases erlotinib clearance by 24 % [37]. serum erlotinib levels have been successfully monitored in patients who are on concomitant medications that interact with cyp3a4, and levels correlate with disease response [38]. caution is recommended when erlotinib is used in patients with hepatic failure; however, pk and safety profiles are similar between patients with liver impairment and normal liver function [39]. erlotinib is a substrate for pglycoprotein (p-gp/abcb1) and breast cancer resistance protein (brcp/abcg2). p-gp and brcp are efflux transporters that negatively regulate intestinal absorption of erlotinib [40]. in patients with genotypes that produce low p-gp or brcp expression, higher trough concentrations of erlotinib and higher rates of clinical toxicity exist [41,42]. after metabolism to desmethyl erlotinib, oxidation and glucuronidation occur [43]. erlotinib and its metabolites are eliminated predominantly in the feces (83 %) and less than 2 % of the recovered drug is unchanged, indicating extensive metabolism prior to excretion [44]. given the limited role of renal metabolism and excretion of erlotinib, it can be safely administered in patients who have chronic renal failure and who are undergoing hemodialysis [45]. cns penetration central nervous system (cns) metastases occur commonly in nsclc. erlotinib crosses the blood-brainbarrier (bbb) to a low degree. the efflux transporter proteins p-gp and brcp are expressed on the bbb and decrease cns accumulation of the erlotinib. the brcp 421a polymorphism leads to decreased efflux of erlotinib and higher cns penetration [42]. in a small study of nsclc patients with brain metastases who received erlotinib, the mean csf concentration of erlotinib was 54 ng/ml. the mean csf penetration rate of erlotinib was 5.1 %. even at these concentrations, however, disease response in the cns was witnessed [46]. in a case series of 23 korean patients with untreated brain metastases who were treated with erlotinib 150 mg daily or gefitinib 250 mg daily, intracranial disease response was achieved in 69.6 % and intracranial disease control was witnessed in 82.6 % of patients [47]. high dose erlotinib given at 600 mg every 4 days or 300 mg every other day has demonstrated efficacy in controlling cns metastases with tolerable side effects in 2 case reports [46,48,49]. pulsatile dosing of erlotinib has also been investigated. in a retrospective analysis of 9 patients with egfr-mutated nsclc and brain metastases, a median erlotinib dose of 1500 mg weekly was utilized. partial cns response was witnessed in 67 % of patients with a median time to cns progression of 2.7 months and median os of 12 months [50]. mechanisms of resistance the most common mechanism of resistance to erlotinib is a secondary mutation in the egfr gene at exon 20 that leads to substitution of methionine for threonine at position 790 (t790m) causing a mutation in the egfr kinase domain. t790m occurs as the method of erlotinib resistance in approximately 50 % of patients upon disease progression [51,52]. erlotinib forms an important hydrogen bond with the threonine residue at position 790, which is positioned in the hydrophobic atp-binding pocket of the catalytic region [53]. erlotinib-resistance with t790m-mutated egfr is driven by increased binding affinity for atp at the atp binding pocket. the l858r mutation reduces egfr affinity for atp, thereby increasing susceptibility to erlotinib inhibition, but the t790m mutation restores egfr atp affinity to near wild-type levels (km[atp]: l858r 148 µm, t790m 5.9 µm, wt egfr 5.2 µm), and decreases the competitive advantage of the atp competitive inhibitor erlotinib. irreversible inhibitors of the egfr atp binding domain, such as afatinib, can overcome resistance to t790m in vitro because they are not in competitive equilibrium with atp [51]. met amplification (see gefitinib-resistance below), small cell transformation and her2 amplification have all admet & dmpk 4(3) (2016) 186-211 tki for egfrand alk-mutated cell cancer doi: 10.5599/admet.4.3.337 191 been less commonly observed as causes for erlotinib resistance [54]. a case report of a patient with egfrmutated nsclc on erlotinib, who developed met amplification as a resistance mechanism, showed tumor response with the addition of crizotinib [55]. gefitinib clinical efficacy gefitinib was the first egfr tki to reach the market, and originally received accelerated approval by the fda for marketing in may 2003 for patients with nsclc who were refractory to platinumand docetaxelbased chemotherapy [56]. the approval was based on a randomized, double-blind, phase ii trial of 216 patients comparing gefitinib 250 mg/day with gefitinib 500 mg/day. the overall response was 12 % in the 250 mg dose and 9 % in the 500 mg dose with a 1 year os of 25 % [57]. however, a survival benefit in this patient population was unable to be demonstrated and astrazeneca suspended promotion of gefitinib. access to the drug was limited to patients who were previously responding to gefitinib. in september 2011, per fda request, astrazeneca voluntarily withdrew gefitinib from the us market due to failure to demonstrate a survival benefit for the drug. while gefitinib did not appear to have meaningful activity in an unselected nsclc patient population, multiple subsequent large phase iii studies performed in asian populations consistently demonstrated clinical benefit for first-line gefitinib for patients with exon 19 and l858r mutations. in these studies, gefitinib improved pfs by 3 to 5 months compared with conventional platinum-based chemotherapy [5861]. given this data, gefitinib attained widespread approval outside of the united states. in july 2015, the fda once again approved gefitinib under orphan product designation for first-line treatment of patients with advanced nsclc with the exon 19 or l858r egfr-activating mutations. the approval was based on a phase iv study of gefitinib for first-line treatment of caucasian patients with exon 19 or l858r egfr mutations. in the 106 patients in this study, the overall response rate was 69.8 % with a median pfs of 9.7 months and median os of 19.2 months [5]. clinical safety gefitinib causes less dose-dependent dermatologic toxicity than erlotinib. this is likely due to the comparatively lower standard therapeutic dose of gefitinib commonly administered to patients compared to other egfr tkis. the approved clinical dose of gefitinib (250 mg daily) is only 33 % of the maximum tolerated dose, whereas erlotinib is clinically dosed at its maximum tolerated dose of 150 mg daily. there does not appear to be a difference in pfs or os for lower dose gefitinib (250 mg daily) vs higher dose gefitinib (500 mg daily) [62].for gefitinib, all-grade rash ranges from 44 to 77 % but grade 3 to 5 rash is very rare (<5 %) with gefitinib. skin rash on gefitinib manifests similarly compared to erlotinib, and ten and sjs have occurred in patients on gefitinib as well. like erlotinib, diarrhea with gefitinib is relatively frequent but not typically severe (all-grade: 32-47 %, grade 3-5: 1-4 %). grade 3 or 4 transaminitis occurs in up to 11 % of patients. rare fatal hepatotoxicity and interstitial pneumonitis have occurred with gefitinib [59,60,63,64]. pharmacology gefitinib, 4-quinazolinamine,n-(3-chloro-4-flurophenyl)-7-methoxy-6-[3-(4-morpholinyl)propoxy] [zd1839, iressa (astrazenaca)] reversibly inhibits egfr by competing with atp at the tyrosine kinase domain. the recommended starting dose is 250 mg orally daily. gefitinib has an ic50 of 0.03-0.1µm for egf, but demonstrates similar ic50 values for other growth factors, including pdgf and igf [56]. maximum plasma grace dy et al. admet & dmpk 4(3) (2016) 186-211 192 levels are achieved between 3 and 7 hours after administering gefitinib with oral bioavailability of approximately 60 % [65,66]. unlike erlotinib, absorption of gefitinib is not impacted by food [66]. however, the drug has significantly decreased solubility and absorption at higher gastric ph, therefore, use of proton pump inhibitors and h2-receptor antagonists should be avoided when possible. upon absorption, gefitinib is 90 % protein bound to alpha1-acid glycoprotein and albumin and the volume of distribution is about 1700 l [67]. gefitinib is metabolized predominantly in the liver by cyp3a4 and cyp2d6 and to a lesser degree by cyp3a5. therefore, gefitinib interacts with similar cyp3a4 inducers and inhibitors as erlotinib as well as cyp2d6 inhibitors/inducers [68]. although there is no specific dose adjustment recommended for patients with hepatic impairment, caution is recommended, as serum gefitinib levels increase in this setting. the primary metabolite of gefitinib, o-desmethyl gefitinib, is inactive [69]. only 4 % of the drug is renally excreted, and no dose-adjustments are recommended for patients with renal failure. cns penetration gefitinib does not penetrate the bbb well. in a study of 8 patients with egfr activating mutations with cns metastases, the csf gefitinib concentration was 3.7+/1.9 ng/ml and the csf penetration of the drug was 1.13+/-0.36 %. despite only 1 % of the drug penetrating into the csf, a partial response in cns metastases was demonstrated in one patient. erlotinib does have superior cns penetration to gefitinib, and erlotinib has successfully induced cns response to previously gefitinib-refractory cns metastases [70]. high-dose gefitinib has been attempted in order to improve control of cns disease. in a case report, gefitinib was increased to 1000 mg daily. this resulted in higher csf concentration of the drug as well as radiographic and symptomatic response in the patient. however, transaminitis and somnolence developed due to the high dose of the drug. of note, at autopsy, the patient’s extracranial tumors all harbored t790m mutation, but the brain metastases did not carry this mutation [71]. mechanism of resistance like erlotinib, the main mechanism egfr of resistance to gefitinib is through the secondary t790m mutation at the intracellular catalytic domain [52,72]. in a small series of 18 patients, 4 (22 %) developed gefitinib resistance through met amplification. this leads to erbb3 (her3)-dependent pi3k/akt activation, which bypasses the need for signaling through the inhibited egfr. the t790m mutation was present in about half of the samples in which met amplification occurred [54,73]. afatinib clinical efficacy afatinib was the first irreversible egfr tki available on the market. the drug was first fda approved on july 2013 for first-line treatment of patients with metastatic nsclc whose tumors harbor exon 19 or l858r activating egfr mutations. the approval was based on results from the lux-lung 3 trial, a randomized phase iii study comparing afatinib with cisplatin plus pemetrexed in patients with metastatic egfr-mutated nsclc. a total of 345 patients were randomized in this trial. overall response rate was significantly higher in patients receiving afatinib (56 % vs 23 %; p=.001). median pfs was 13.6 months with afatinib and 6.9 months with chemotherapy (hr 0.47; 95 % ci, 0.34 to 0.65; p=.001). both asian and non-asian patients were enrolled in the study, but the study was underpowered to detect a pfs difference in non-asian patients [74]. similarly, the lux-lung 6 trial demonstrated a significant pfs benefit for first-line afatinib versus cisplatin and gemcitabine in asian patients [75]. a combined analysis of lux-lung 3 and lux-lung 6 demonstrated an overall survival benefit for patients with exon 19 deletion but not with the l858r point mutation [76]. in the randomized phase iib lux-lung 7 trial, which compared first-line afatinib vs gefitinib admet & dmpk 4(3) (2016) 186-211 tki for egfrand alk-mutated cell cancer doi: 10.5599/admet.4.3.337 193 in egfr-mutated nsclc, afatinib produced superior pfs at 12, 18 and 24 months [77]. afatinib also recently gained fda approval in april 2016 for second-line treatment of advanced squamous cell nsclc. compared with erlotinib, afatinib prolonged both pfs and os in patients with advanced squamous cell lung cancer [78]. patients who develop acquired resistance to erlotinib or gefitinib have little response (8 %) to single-agent afatinib [79]. however, dual inhibition of egfr with afatinib and cetuximab induced responses in patients who previously progressed on erlotinib or gefitinib. responses of 32 % and 25 % were witnessed in t790m-positive and t790m-negative tumors, respectively [80]. clinical safety diarrhea occurs commonly with afatinib. all-grade diarrhea has been reported as high as 95 % with grade 3 to 5 diarrhea of 14 % in clinical trials. rash/acne (all-grade: 89 %, grade 3-5: 16 %) and stomatitis (all-grade: 51-72 %, grade 3-5: 8 %) are frequently encountered [74,75]. rare adverse effects include decrease in left ventricular ejection fraction, interstitial lung disease and fatal hepatotoxicity. pharmacology afatinib, (e)-n-[4-(3-chloro-4-fluoroanilino)-7-[(3s)-oxolan-3-yl]oxyquinazolin-6-yl]-4-dimethylami-no)but-2-enamide [bibw 2992, gilotrif, (boehringer ingelheim)], irreversibly inhibits erbb1 (egfr), erbb2 (her2), and erbb4 (her4). as opposed to first-generation egfr tkis, afatinib forms an irreversible covalent bond at the atp-binding site of egfr. the acrylamide group permits afatinib to form a covalent bond at the tyrosine kinase site and is important in its activity against t790m, her2, and erbb4 [81]. afatinib demonstrates a similar ic50 for wild-type egfr and l858r-mutated egfr, but it is 100-times more potent against t790m-mutated egfr. however, clinically meaningful inhibition of t790m with higher dose afatinib is unable to be achieved without substantial toxicity [80]. afatinib also inhibits her3 activity. her3 activates the pi3k/akt survival pathway in nsclc. this receptor lacks intrinsic kinase activity and its activity depends upon transphosphorylation by heterodimerization with egfr, her2 and met. outside of the erbb family of receptors, afatinib does not significantly inhibit other tyrosine or serine/threonine kinases [82]. the recommended starting dose of afatinib is 40 mg administered orally daily, and peak plasma concentrations are reached 2 to 5 hours later. absorption is decreased with high-fat meals and the drug should not be taken with food. solubility of afatinib is not affected by physiologic gastric ph levels; therefore, antacids do not affect absorption. upon absorption, afatinib is heavily protein bound (~95 %), bioavailability is 92 %, and half-life is 37 hours [83]. unlike the reversible egfr tkis, afatinib is not significantly metabolized through the cyp pathway, and therefore, cyp inhibitors and inducers have little impact on plasma levels of afatinib. metabolism of afatinib is primarily mediated by non-enzyme catalyzed michael adduct formation to nucleophilic, electron-rich small molecules and proteins. however, 89 % of afatinib is excreted as the unchanged parent compound (85 % feces, 4 % urine) [84]. afatinib serves as a substrate and inhibits both p-gp and brcp. inhibitors of p-gp, such as ritonavir can increase concentrations of afatinib, while inducers of p-gp reduce afatinib levels. the clinical consequences of concurrent use of afatinib with p-gp inducers/inhibitors are questionable, and these drugs can be administered 6-12 hours apart to avoid significant interactions [85]. in patients with renal impairment with egfr of 15-29 ml/minute/1.73 m 2 , a dose reduction to 30 mg daily is recommended. canadian labeling recommends against use of afatinib if crcl is less than 30 ml/min. the drug has not been well-studied in patients undergoing hemodialysis. caution is advised in patients with hepatic impairment, but no specific dose adjustments are recommended. canadian labeling advises against grace dy et al. admet & dmpk 4(3) (2016) 186-211 194 afatinib use in patients with child-pugh class c cirrhosis. cns penetration the lux-lung trials allowed enrollment of patients with stable brain metastases. the lux-lung 3 trial enrolled 35 patients with brain metastases who were treated with either first-line afatinib or cisplatin and pemetrexed. the median pfs was 11.1 months in the afatinib group and 5.4 months in the chemotherapy group (hr, 0.52; p=0.13). in an analysis 541 patients treated on the afatinib compassionate use program, 100 patients who had brain or leptomeningeal metastases were identified (74 % had documented egfr activating mutation). thirty-five percent of evaluable patients developed a cns response with median duration response of 120 days. one patient who developed a response on 40 mg of oral afatinib, achieved a csf afatinib concentration of about 1 nm [86]. mechanisms of resistance a study evaluating tumor samples from 42 patients who experienced disease progression on afatinib shed light on acquired resistance mechanisms to afatinib. t790m egfr mutation was detected in 47.6 % of specimens. there was no association with previous first-generation egfr tki use and detection of t790m after resistance to afatinib. there were no mutations detected in pik3ca, her2, braf, kras, nras, mek1, akt2, jak2 or lbk1 [87]. pre-clinical models have suggested that met amplification, upregulation of the stat3 pathway, and fgfr1-induced survival signals contribute to afatinib resistance [88-90]. osimertinib clinical efficacy osimertinib was the first “third-generation” egfr tki to make it to market. the third generation tkis are designed to irreversibly target mutant egfr (exon 19 deletion, l858r and t790m), while sparing wildtype egfr. in mouse models of egfr-mutated nsclc, osimertinib had similar antitumor activity as afatinib against l858r but more activity against t790m nsclc [91]. the phase i component of the aura trial had dose escalation and expansion cohorts, which enrolled 253 patients with locally advanced or metastatic egfr-mutated nsclc who had disease progression on an egfr tki. a total of 127 (50.2 %) patients were confirmed to have the t790m mutation. in all patients, the response rate to osimertinib was 51 %. in patients harboring a t790m mutation, the response rate to osimertinib was 61 %, but in patients without detectable t790m, the response rate was only 21 %. the median pfs in t790m-positive patients was 9.6 months and 2.8 months in t790m-negative patients [12]. preliminary data from the 201 patients enrolled in the phase ii extension of aura demonstrated an orr of 58 %. the phase ii aura2 trial also has preliminary data demonstrating a 64 % orr with osimertinib. both phase ii trials have not yet reached sufficient maturity to determine a median pfs [92]. in november 2015, osimertinib was granted accelerated fda approval for treatment of egfr-mutated nsclc that has progressed on a previous egfr tki. osimertinib is currently in phase iii development for adjuvant treatment as well as firstand secondline treatment of advanced egfr-mutated nsclc. clinical safety overall, osimertinib is well-tolerated. all-grade diarrhea occurs in 42 % of patients, with only 1 % experiencing grade 3 or 4 diarrhea. similarly 41 % of patients will experience a rash but only 0.5 % will develop grade 3 or 4 rash. dose reductions are necessary in only 4 % of patient on osimertinib. prolongation of the qtc interval was the most common reason for dose reduction, and occurred in 2.2 % of patients. as with other egfr tkis, fatal pneumonitis has occurred in patients with osimertinib but is rare admet & dmpk 4(3) (2016) 186-211 tki for egfrand alk-mutated cell cancer doi: 10.5599/admet.4.3.337 195 (1 %). fatal stroke and pneumonias have also occurred in patients on the drug. cardiomyopathy, with a decline in left ventricular function by >10 % was also rarely encountered. discontinuation of osimertinib, due mostly to pneumonitis and stroke, was necessary in 5.6 % of patients enrolled in the key clinical trials to date [12,92,93]. pharmacology osimertinib, n-[2-[2-(dimethylamino)ethyl-methylamino]-4-methoxy-5-[[4-(1-methylindol-3-yl)-pyrimidin-2-yl]amino]phenyl]prop-2-enamide [azd-9291, tagrisso, (astrazeneca)], irreversibly inhibits mutated egfr by covalently bonding with the cysteine-797 residue in the atp binding site. the drug’s affinity for t790m may be related to hydrophobic interactions between the methionine moiety on t790m-mutated egfr and the pyrimidine component of osimertinib. it has nearly 200-fold greater affinity for egfr with l858r/t790m mutation than wild-type egfr in vitro [91]. osimertinib also inhibits her2, her3, her4, ack1 and blk in vitro [94]. the manufacturer-recommended oral dose of osimertinib is 80 mg daily. antacids do not impact drug absorption. intake of high fat foods prior to osimertinib administration mildly increases osimertinib plasma concentration by about 14 % [95]. after administration, maximum plasma concentration is reached in approximately 6 hours, the volume of distribution is 986 l, and the half-life is 48 hours [12,94]. metabolism primarily occurs via oxidation in the liver through cyp3a4. the two active metabolites of osimertinib are az7550 and az5104, the latter of which causes more wild-type egfr inhibition than the parent drug [91]. there is no difference in absorption or metabolism of osimertinib based on age, ethnicity, or tobacco use. osimertinib is a substrate for p-gp and brcp, and it inhibits brcp but not p-gp. osimertinib is eliminated in feces (68 %) and in the urine (14 %) with only 2 % of the drug being unchanged at elimination. moderate renal impairment (creatinine clearance >30 ml/min) and mild hepatic impairment do not appear to impact drug levels or toxicity. osimertinib use in severe renal impairment (crcl<30 ml/min) or moderate to severe hepatic impairment has not been studied [94]. cns penetration in a mouse model, osimertinib’s distribution to the brain was 10-times higher than gefitinib and response in brain metastases were witnessed with osimertinib. in aura and aura2, the csf concentrations of osimertinib in two patients were 0.2 and 1 % of the predicted plasma concentrations [94]. case reports of two patients t790m-mutated egfr and untreated brain metastases showed that osimertinib 80 mg daily induced sustained partial responses to both cns and systemic disease [96]. the phase i bloom study evaluated osimertinib 160 mg daily for patients with egfr-mutated nsclc with leptomeningeal disease. all patients had previously received an egfr tki. preliminary data from this trial showed that in 12 patients who had 12-week imaging assessment, high-dose osimertinib produced radiologic improvement in cns disease in 7 patients (58.3 %), stable disease in 2 patients (16.7 %) and 3 patients were not evaluable. overall, the higher dose of osimertinib was well-tolerated, but 5 of 20 patients required dose interruptions and only 2 required dose reductions to 80 mg daily. grade 3 neutropenia was observed in one patient on the higher dose of osimertinib, but this resolved after 3 days of holding the drug and did not recur after reducing the dose to 80 mg [97]. resistance mechanisms currently, the most commonly reported acquired resistance mutation to osimertinib is the missense mutation of cysteine for serine at position 797 (c797s). this substitution prevents osimertinib from forming a covalent bond at the kinase-binding site [98]. the c797s mutation accounts for up to 40 % of acquired resistance to osimertinib in the limited data on this topic to date [99]. clinically, loss of the t790m grace dy et al. admet & dmpk 4(3) (2016) 186-211 196 mutation has been observed after treatment with osimertinib [99]. in pc9 and nci-h1975 cell lines, acquired nras and kras mutations were noted after exposure to osimertinib. a number of these resistant cell lines were sensitive to combination of osimertinib and the mek inhibitor selumetinib [100]. investigational agents (table 2) another third generation egfr tki, olmutinib (bi 1482694/hm61713), gained fda breakthrough therapy designation in december 2015. like osimertinib, olmutinib targets mutant egfr and t790m while sparing wild-type egfr. in the phase i/ii hm-emsi-101 trial, patients with egfr-mutated nsclc who progressed on a previous egfr tki received olmutinib. as of june 2016, 76 patients had received the drug at the recommended phase ii dose of 800 mg daily. the drug produced an orr of 62 % with a dcr of 91 % by independent assessments. few grade 3 or 4 toxicities were observed, including a 5 % risk of grade 3 or 4 skin rash [101]. the phase ii eluxa 1 (hm-emsi-202) trial is open for patient recruitment and is investigating the efficacy of olmutinib in patients who have developed t790m after exposure to first-line egfr tkis [102]. multiple monotherapy trials of olmutinib are being planned, including phase iii studies comparing olmutinib with platinum-based chemotherapy in a second-line setting and comparing olmutinib with afatinib in the first-line treatment of egfr-mutated nsclc. investigators are also planning on trials evaluating olmutinib in combination with various agents, including pembrolizumab, bevacizumab, nintedanib, afatinib, and the igf ligand-neutralizing antibody bi836845. table 2. egfr tkis in development drug name targets clinical activity select ongoing/future trials olmutinib (bi 482694/hm61713) [102] mutant egfr t790m phase i/ii (n=76) orr 62%, dcr 91% ongoing phase ii eluxa 1 (nct02485652) planned: olmutinib vs platinum regimen in 2 nd line; olmutinib vs afatinib in 1 st line; combinations with: pembrolizumab, bevacizumab, nintedanib, afatinib, bi836845 asp8273 [103] mutant egfr t790m phase i orr 50% (n=36); in t790m orr 80% (n=15) ongoing phase ii (nct02192697); ongoing 1 st line study (solar) (nct02588261) asp8273 vs erlotinib or gefitinib egf816 [104] mutant egfr t790m phase i/ii (n=22) orr 55%, dcr 86% ongoing phase i/ii (nct02108964) azd3759 [106] mutant egfr high cns penetration phase i (n=2) 1 intracranial pr and 1 intracranial sd ongoing phase i nct02228369) epitinib (hpml-813) [105] mutant egfr high cns penetration phase i (n=12) 71% tki naïve pts achieved pr in cns; 100% pts w/previous tki exposure achieved sd in cns ongoing phase i (nct02590952) tarloxotinib bromide (th-4000) [107] wild-type egfr mutant egfr (nont790m) her2 n/a ongoing phase ii (nct02454842) admet & dmpk 4(3) (2016) 186-211 tki for egfrand alk-mutated cell cancer doi: 10.5599/admet.4.3.337 197 other third-generation egfr tkis in development, including asp8273 and egf816, have selective activity against mutant egfr (l858r, del exon 19) as well as t790m without affecting wild-type egfr. these compounds have produced orrs of 35.5 to 44 %, dcrs of 65 to 91 % and median pfs of 6.7 months to 9.2 months, respectively [103,104]. epitinib (hmpl-813) is an egfr tki that was specifically designed for high cns penetration. in a phase i study (nct02590952), the drug was well-tolerated, with grade 3 elevation in liver function tests in 2 to 5 % of patients. in the dose expansion arm of the study, 12 patients were evaluable as of october 2015. of these, 5 patients who were tki naïve achieved a pr both systemically and in the cns. in 5 patients who had previous tki exposure, all had stable disease in the brain [105]. azd3759 is another egfr tki designed to have high bbb penetration. in a phase i trial (nct02228369) of patients with egfr-mutated nsclc and brain metastases, of the 2 evaluable patients with brain metastases reported to date, 1 patient achieved an intracranial pr and the other had stable cns disease. the trough csf concentration of azd3759 was between 6 and 7.7 nm, which is close to the mutant egfr ic50 of azd3759 [106]. tarloxotinib bromide (th-4000) is a prodrug that is converted to an irreversible tki of wild-type egfr, mutant egfr and her2 under hypoxic conditions. as the tumor environment is typically hypoxic, it is hypothesized that greater intra-tumoral tki exposure and a greater therapeutic index will be achieved with this drug, and egfr tki resistance has been overcome with th-4000 in animal models. it is currently being investigated in a phase ii study (nct02454842) for patients with egfr-mutated nsclc who have nont790m mediated resistance to first-line egfr tki [107]. alk fusion protein in 2007, another key molecular driver of nsclc was identified. an inversion within chromosome 2p causes the fusion of the echinoderm microtubule-associated protein-like 4 (eml4) gene and the anaplastic lymphoma kinase (alk) gene [6,108]. at least 8 different eml4-alk fusion variants have been identified. all variants harbor a breakpoint at intron 19 of alk and variable intron breakpoints in eml4 [109,110]. the eml4-alk oncogene generates the eml4-alk fusion protein and all variants of this protein contain an eml4 transmembrane domain and an intracellular alk tyrosine kinase domain. the eml4 domain promotes self-association of the alk tyrosine kinase domain driving autophosphorylation and downstream signaling, promoting oncogenesis through pi3k/akt and jak/stat pathways [6,111]. the eml4-alk fusion mutation occurs in approximately 3 to 7 % of nsclc (typically adenocarcinoma) with a predilection for younger patients with low tobacco exposure [112]. currently, there are three alk tkis, which have widespread clinical approval for the treatment of alk-positive nsclc (table 3). alk tkis crizotinib clinical efficacy crizotinib first gained accelerated fda approval in august 2011 for all patients with advanced alkpositive nsclc based on two small phase i/ii clinical trials, which demonstrated response rates of 50 to 60 % and pfs of over 9 months [113,114]. regular approval for the same indication was granted by the fda in november 2013 due to evidence presented in the phase iii, open label profile 1007 trial that compared crizotinib to docetaxel or pemetrexed for use in advanced alk-positive nsclc. in 347 previously treated alk-positive patients, crizotinib produced superior response rates compared with chemotherapy (orr 65 grace dy et al. admet & dmpk 4(3) (2016) 186-211 198 vs 20 %, respectively). the median pfs with crizotinib was 7.7 months vs 3.0 months with chemotherapy (hr for progression or death with crizotinib, 0.49; 95 % ci, 0.37-0.64; p<0.001) [115]. in the first-line treatment setting, profile 1014 compared crizotinib to pemetrexed plus platinum demonstrating superior orr (74 % vs 45 %; p<0.0001) and pfs (10.9 vs 7 months; hr 0.45;95 % ci, 0.35-0.60) with crizotinib [116]. table 3. alk tki characteristics alk tki name crizotinib ceritinib alectinib primary targets alk c-met ros1 alk (including crizotinib resistance mutations) igf-1r insr ros1 alk (including crizotinib resistance mutations) ret type of inhibition atp-competitive atp-competitive atp-competitive clinical approval first-line treatment of advanced alkpositive nsclc second-line treatment of advanced alkpositive nsclc second-line treatment of advanced alkpositive nsclc alk ic50 (cell free assay)[23] 3.6 nm 0.15 nm 1.9 nm drug interactions cyp3a4 inducers/inhibitors p-gp inducers/inhibitors cyp3a4 inducers/inhibitors p-gp inducers/inhibitors cyp3a4 minor substrate, no significant interactions with inducers/inhibitors food interactions take with or without food. high fat meal increases bioavailability. take on empty stomach. high fat meal increases bioavailability. take with food. metabolism liver cyp3a4/5 liver cyp3a4 liver cyp3a4 method of elimination 63% feces (53% unchanged) 22% urine (2% unchanged drug) 92% feces (68% unchanged) 1% urine 98% feces (84% unchanged) <1% urine cns penetration 0.06-0.26% cns penetration intracranial disease control 80%. csf concentrations unpublished to date. 86% cns penetration in 4 patient study clinical safety crizotinib produces unique visual disturbances in up to 70 % of patients, though these are rarely severe. the visual disturbances are pronounced with changes in ambient lighting and can present as floaters, trails or flashes of light in the peripheral vision. vision changes tend to be mild, brief, and self-limited and no associated ophthalmologic changes were identified in patients on the profile 1005 study. dose adjustment is rarely needed for vision changes [113,116,117]. diarrhea, peripheral edema, and vomiting occur in approximately 50 to 60 % of patients, but reach grade 3 or 4 toxicity in less than 5 % of patients. grade 3 or 4 elevations of aminotransferase levels occur in up to 15 % of patients receiving crizotinib and can be managed with dose interruptions or reductions. treatment discontinuation is rarely necessary for liver function abnormalities [113,116]. symptomatic hypogonadism has been linked to crizotinib use in male patients. in one study, 84 % of male patients were found to have low free testosterone levels while taking crizotinib with concomitant decline in fsh and lh, suggesting that crizotinib may induce secondary hypogonadism [118]. therefore, monitoring of free testosterone may be necessary in males taking admet & dmpk 4(3) (2016) 186-211 tki for egfrand alk-mutated cell cancer doi: 10.5599/admet.4.3.337 199 crizotinib, particularly in patients with symptoms suggestive of androgen deficiency. rare fatal interstitial lung disease/pneumonitis has been reported [119]. bradycardia and q-t prolongation can occur in about 5 % of patients taking crizotinib [120]. pharmacology crizotinib, 3-[(1r)-1-(2,6-dichloro-3-fluorophenyl)ethoxy]-5-(1-piperidin-4-ylopyrazol-4-yl)pyridine-2amine [pf-2341066, xalkori, (pfizer)] competes with atp at the catalytic domain of the alk tyrosine kinase and mesenchymal-epithelial transition factor (c-met) kinase [111]. identified as a potent inhibitor of alk and c-met in biochemical enzymatic screens, crizotinib inhibits tyrosine phosphorylation of npm-alk (a fusion protein present in some cases of anaplastic large cell lymphoma) with mean ic50 values between 25 to 50 nm and c-met with mean ic50 values of 5 to 20 nm [121-123]. the eml4-alk fusion protein is inhibited less potently by crizotinib, with ic50 values between 250 to 340 nm [124,125]. the recommended dose of crizotinib is 250 mg administered twice daily. food does not have a significant impact on absorption. bioavailability of crizotinib is 43 % and peak plasma concentration is reached at a median of 4 hours. the drug is heavily protein bound (91 %) with a volume of distribution of 1772l and a terminal halflife of 42 hours [126]. crizotinib undergoes extensive hepatic metabolism through the cyp3a4/5 pathway, and cyp3a4 inducers and inhibitors interact with crizotinib [126,127]. steady-state is achieved within 15 days, and crizotinib exhibits a non-linear pk, meaning the rate of clearance decreases with multiple dosing. the geometric mean values for cl/f declined from 100 l/hr after a single dose of crizotinib to 64.5 l/hr and 60.1 l/hr after 15 and 28 days of dosing, respectively. this phenomenon may be due to autoinhibition of cyp3a4. age, gender, race and body weight do not appear to influence the pk of crizotinib [128]. 63 % of the drug is eliminated via the feces (53 % unchanged) and 22 % in the urine (2.3 % unchanged) [126]. given the extensive hepatic metabolism of crizotinib, caution is advised in patients with liver dysfunction, though no specific dose adjustments are recommended. likewise, crizotinib has not been well-studied in patients with renal impairment, and close monitoring is advised in this setting. cns penetration overall, crizotinib has poor cns penetration. in a 29-year-old patient with metastatic alk-positive nsclc, new brain and leptomeningeal metastases developed 7.5 months after starting crizotinib. his plasma crizotinib concentration was 237 ng/ml, but his csf concentration was only 0.616 ng/ml, with a csf-to-plasma crizotinib ratio of 0.0026 [125]. despite poor cns penetration, crizotinib has been reported to induce responses in previously untreated cns disease. in a 58-year-old patient with an asymptomatic solitary brain metastasis and leptomeningeal carcinomatosis, crizotinib was initiated, and a complete cns response was achieved despite a crizotinib csf-to-serum ratio of 0.0006 [129]. a retrospective analysis of alk-positive patients with asymptomatic cns metastases who received crizotinib on the profile 1005 and 1007 trials showed that 275 (31 %) of the 888 had asymptomatic brain metastases upon enrollment on these trials. in patients with untreated brain metastases, crizotinib produced intracranial disease control in 56 % at 12 weeks. the median intracranial time to progression was 7 months in this population. in patients who had received brain radiotherapy prior to enrollment in these trials, cns disease control was attained in 62 % with median time to progression of 13.2 months. however, due to poor cns penetration of crizotinib, among patients without brain metastases at enrollment, 20 % developed cns disease while taking crizotinib [130]. second-generation alk inhibitors have better cns penetration and activity. resistance mechanisms alk-positive tumors can develop alk dependent and alk independent mechanisms of crizotinib grace dy et al. admet & dmpk 4(3) (2016) 186-211 200 resistance. crizotinib resistance predominantly occurs through acquired mutations in the alk kinase domain at the atp-binding pocket, which prevents binding of the drug. these mutations include the l1196m point mutation, which is homologous to the tki resistance mutations t790m in egfr-mutated nsclc and t315i in chronic myelogenous leukemia [131,132]. other acquired alk kinase mutations that confer resistance to crizotinib include c1156y, l1152r, i1171t, s1206y, f1174l and g1269a [131,133-136]. the g1202r crizotinib-resistance mutation also confers resistance to second generation alk tkis [137]. increase in alk copy number has also been demonstrated in patients who develop crizotinib resistance [132]. in humans, alk independent mechanisms of resistance include kras mutations and development of egfr mutation without evidence of persistent alk gene rearrangement [132]. in a mouse model, crizotinib resistance was induced through activation of the egfr signaling pathway, which could be overcome by combination therapy with crizotinib and the heat shock protein 90 (hsp90) inhibitor ganetespib. hsp90 is a chaperone protein that regulates the folding and stability of multiple “client proteins,” including egfr, eml4-alk, her2, and braf and can therefore target multiple pathways of drug resistance [138]. insulinlike growth factor 1 receptor activation has been identified as a possible resistance mechanism to crizotinib, and serves as a target of the second-generation alk inhibitor ceritinib (detailed below) [139]. ceritinib clinical efficacy ceritinib was the first second-generation alk tki approved for clinical use in the us. preclinical data demonstrated that certinib overcomes crizotinib-resistance mutations, particularly l1196m, g1269a, s1206y and i1171t but not g1202r and f1174c [136]. in april 2014, the fda granted accelerated approval to ceritinib for the treatment of patients with metastatic alk-positive nsclc who were intolerant to or experienced disease progression on crizotinib. this approval was based on the single-arm, multicenter phase i ascend-1 trial that enrolled a total of 130 patients with advanced alk-rearranged nsclc, including patients who had experienced disease progression during crizotinib treatment. the orr for patients who received at least 400 mg of ceritinib daily was 58 %, while the response rate for patients who had previously received crizotinib was 56 %. median pfs was 7.0 months (95 % ci, 5.6 to 9.5) [14]. in this trial, tumor response was noted regardless of the underlying mechanism of acquired crizotinib resistance, and patients with alk gene amplification and alk tyrosine kinase domain mutations developed responses to ceritinib. given the excellent responses witnessed in crizotinib-treated patients, the main mechanism of crizotinib resistance appears to be alk dependent, and may be overcome with a more potent alk inhibitor, such as ceritinib [14]. updated data from this trial, including 246 patients, showed an orr of 58.5 % in all patients, with a median duration of response of 9.7 months, time to first response of 6.1 weeks and median pfs of 8.2 months [140]. ascend-2 was a single-arm, multicenter phase ii trial of 140 patients that evaluated efficacy and safety of ceritinib in alk-positive nsclc who progressed after chemotherapy and crizotinib. the orr was 38.6 % with median pfs of 5.7 months. the trial also allowed patients with untreated brain metastases to participate (71.4 % had baseline brain metastases, 28 % had untreated brain metastases). median pfs was 5.4 months in patients with brain metastases versus 11.3 months in patients without brain metastases [141]. ascend-3 examined ceritinib use for metastatic, alkpositive nsclc patients who were alk tki-naïve, including those with brain metastases. at presentation, 124 patients had enrolled in the trial. in this study population, the orr was 63.7 % with disease control rate of 89.5 %, and the median pfs was 11.1 months [142]. admet & dmpk 4(3) (2016) 186-211 tki for egfrand alk-mutated cell cancer doi: 10.5599/admet.4.3.337 201 clinical safety like many other tkis, ceritinib induces nausea, diarrhea, and vomiting. while 60 to 80 % of patients experience these symptoms on ceritinib, only 5 to 7 % of patients experience grade 3 or 4 nausea, diarrhea or vomiting. approximately 58 % of patients taking ceritinib at the recommended starting dose of 750 mg daily required at least one dose reduction [143]. ceritinib was discontinued entirely in only 7 to 8 % of patients on clinical trial due to adverse events [14,140-142]. severe interstitial lung disease, which may be related to ceritinib has also been reported, but the incidence is less than 5 %. grade 3 or 4 transaminitis occurs in approximately 21 % of patients. like crizotinib, low risks of bradycardia and q-t prolongation have been observed. ceritinib also affects the insulin-like growth factor 1 receptor (igf-1r) and insulin receptor (insr), and grade 3 or 4 hyperglycemia was witnessed in up to 13 % of patients. underlying diabetes may exacerbate this effect. elevated lipase and fatal pancreatitis have been reported in clinical trials [14,140]. pharmacology ceritinib, 5-chloro-2-n-(5-methyl-4-piperidin-4-yl-2-propan-2-yloxyphenyl)-4-n-(2-propan-2-ylsulfo-nylphenyl)-pyrimidine-2,4-diamine [ldk378, zykadia, (novartis)] is a second-generation alk inhibitor that competitively inhibits the alk tyrosine kinase domain through phosphorylation. it also inhibits igf-1r, insr and the ros1 receptor. unlike crizotinib, it does not inhibit c-met [136,143]. ceritinib more potently inhibits alk than crizotinib, as ceritinib produces ic50 values of 0.15 nm in enzymatic assays (crizotinib ic50 3.6 nm) and ic50 of 25nm in cell lines [136,144]. the recommended starting dose is 750 mg administered orally once daily. its absorption is increased when administered with a meal, with peak plasma activity increasing by 41 to 43 % compared to a fasting state. the drug should, therefore, be taken on an empty stomach. oral bioavailability is unknown, but peak plasma levels are achieved at 4 to 6 hours after dosing. ceritinib is 97 % protein-bound upon absorption, has a volume of distribution of 4230l, and reaches steady state at 15 days. it is primarily metabolized in the liver by cyp3a4 and serves as a substrate for pglycoprotein. the drug may inhibit cyp3a and cyp2c9 at the recommended dosing. ceritinib has a terminal half-life of 41 hours, and like crizotinib, has a nonlinear pk over time, with a cl/f of 88.5 l/hr after one dose and a cl/f of 33.2l/hr at steady state. following administration, 92.3 % of the drug is eliminated in the feces (68 % unchanged) [143,144]. no dose adjustments for hepatic or renal failure have been recommended. cns penetration the ascend trials allowed patients with untreated, asymptomatic brain metastases to enroll. in ascend-2, intracranial disease control was attained in 80 % of patients with brain lesions. in 5 of the 6 patients with brain lesions that had not previously been treated, 2 complete responses and 3 partial responses were achieved with ceritinib [141]. likewise, in ascend-3, the intracranial disease control rate was 80 %, and 2 partial responses were achieved with ceritinib. all responses in the brain matched or exceeded the responses achieved systemically [142]. to date, csf measurements of ceritinib after oral administration are unavailable. resistance mechanisms in 10 patients who developed disease progression while taking ceritinib (median duration of treatment 7.5 months), tumor analysis revealed several novel alk resistance mutations in the tyrosine kinase domain. a total of 11 samples were obtained in the 10 patients, and in 5 of 11 samples, secondary mutations in the alk tyrosine kinase domain were identified. of these mutations, 3 were g1202r, 1 was f1174c, and 1 was grace dy et al. admet & dmpk 4(3) (2016) 186-211 202 f1174v. two of the patients had s1206y and g1269a mutations prior to ceritinib, which were no longer present on post-ceritinib tumor samples. alk amplification was not witnessed in this population [145]. alectinib clinical efficacy alectinib is another potent, second-generation oral alk tki that, like ceritinib, inhibits wild-type alk as well as the crizotinib resistance mutations l1196m, l1152r, g1296a and c1156y [146,147]. in a phase i/ii study performed on a japanese population of alk tki naïve patients, alectinib at a dose of 300 mg twice daily produced an orr of 93.5 % [148]. another phase i/ii study examining alectinib 600 mg twice daily for alk-positive patients who had previously progressed on or were intolerant to crizotinib showed orr of 55 % [149]. based on these data, in december 2015, the fda granted accelerated approval to alectinib for treatment of alk-positive nsclc patients who experience disease progression on or intolerance to crizotinib. subsequently, a phase ii trial of 87 patients evaluating alectinib for alk-positive nsclc patients who had progressed on either chemotherapy or crizotinib showed an orr of 48 [150]. recently, results of the j-alex study, a randomized open-label phase iii trial were presented at the american society of clinical oncology annual meeting. this trial randomized 207 japanese patients with advanced alk-positive nsclc who were alk tki naïve to alectinib 300 mg twice daily or crizotinib 250 mg twice daily. the primary endpoint of the study was pfs, and alectinib produced a statistically superior pfs with a pfs hr of alectinib compared to crizotinib of 0.34 (99.6826 % ci: 0.17-0.70, stratified log-rank p<0.0001). the median pfs of alectinib was not reached (95 % ci: 20.3-not estimated) and the median pfs of crizotinib was 10.2 months (95 % ci: 8.2-12) [151]. based on these results, enthusiasm exists for moving alectinib to front-line treatment for alk-positive nsclc, and a global, randomized phase iii trial comparing alectinib to crizotinib is underway (nct02075840). clinical safety alectinib carries a favorable side effect profile with significantly less gi toxicity than crizotinib. fatigue is the commonest adverse event encountered and occurs in approximately 30 % of patients (all grade 1-2). grade 1-2 myalgias, peripheral edema, increased creatine kinase, nausea, elevated aminotransferases, constipation and photosensitivity occur in 13-36 % of patients. grade 3 or 4 events are rare and occur at an incidence of less than 5 % with the exception of aminotransferase abnormalities occurring in 5 to 6 % of cases. fatal hemorrhage was witnessed in one patient who was on anticoagulants. in the phase ii trial, dose interruption was needed for 36 % of patients and dose reduction occurred in 16 % of patients. the medication was discontinued in 2 to 9 % of patients due to adverse events [148-150]. pharmacology alectinib, 9-ethyl-6,6-dimethyl-8-(4-morpholin-4-ylpiperidin-1-yl)-11-oxo-5h-benzo[b]carbazole-3carbonitrile [ch5424802, alecensa, (roche)] competitively inhibits the tyrosine kinase domain of the alk fusion protein. it is a more potent alk inhibitor than crizotinib with an in vitro kinase ic50 of 1.9 nm and ic50 of 3.0 nm in cell lines. it demonstrates high target selectivity with weak or no inhibition of over 20 other kinases, including ros1 and met [152]. however, recently, it was discovered that alectinib potently inhibits ret kinase activity with an ic50 of 4.8 nm, and is active against oncogenic ret-rearrangements observed in nsclc [153]. the recommended dose is 600 mg twice daily with food. it is metabolized to its major active metabolite, m4, in the liver by cyp3a4 [154]. the maximum plasma concentration at steady-state is 665ng/ml (44 %) and for m4 is 246 ng/ml (45 %), and the drug reaches steady-state after 7 days. under fed conditions, the bioavailability of alectinib is 37 %, and a high-calorie, high-fat meal increased exposure admet & dmpk 4(3) (2016) 186-211 tki for egfrand alk-mutated cell cancer doi: 10.5599/admet.4.3.337 203 of the drug by 3-fold [154]. the volume of distribution for alectinib is 4,016 l and 10,093l for the m4 metabolite. alectinib also demonstrates nonlinear pk with apparent clearance of 81.9l/hr for alectinib and 217 l/hr for m4. after administration, 84 % of the drug is eliminated in the feces as unchanged alectinib and 6 % as m4 [148,149,154]. despite its metabolism by cyp3a, co-administration with strong cyp3a inhibitors and inducers have not been found to significantly alter alectinib levels. alectinib does inhibit p-gp and brcp. no dose adjustments have been recommended for hepatic or renal impairment [154]. cns penetration alectinib has higher cns responses than crizotinib. in one of the phase i/ii studies, 21 patients were enrolled with cns metastases. of these patients, 12 (57 %) had progressive cns disease at enrollment. in the 21 patients with cns disease, alectinib produced a cns orr of 52 % with six (29 %) complete responses and five (24 %) partial responses. overall cns disease control rate was 90 %. in the four patients who had not received prior brain radiotherapy, a complete response was achieved in two patients and one had a partial response. in this study, at the 600 mg twice daily dose of alectinib, the csf trough concentration of alectinib was 2.69 nm/l, which is nearly the same as the unbound systemic trough of alectinib of 3.12 nm/l, and exceeds the ic50 for alk inhibition for alectinib in cell-free assays (1.9 nm/l) [149]. in the japanese phase i/ii alectinib study, 15 patients had known brain metastases, 12 of whom had previous brain radiation. no cns disease progression was witnessed on alectinib. of the 3 patients who did not receive radiation to the brain, 2 continued on study for more than 300 days without cns disease progression [148]. in the phase ii study, 16 patients had measurable cns disease at baseline. overall, 100 % of patients achieved disease control in the cns and 75 % of patients achieved intracranial disease response. four (25 %) of these patients achieved complete cns response and eight (50 %) developed partial response in the cns. the median duration of cns response was 11.1 months [150]. therefore, durable cns control and even complete cns responses are attainable with alectinib. resistance mechanisms (table 4) in a cell line study, an acquired v1180l mutation was inducible with alectinib exposure. this mutation exists in the alk tyrosine kinase domain at the back of the atp binding pocket, and sterically interferes with the ability of both alectinib and crizotinib to bind effectively to the kinase and mediates a high-level resistance. an i1171t mutation was isolated in a tumor specimen in a patient who had relapsed on alectinib. this mutation is thought to disrupt the hydrogen bond between alectinib and e1167. compared with v1180l, the i1171t mutation confers an intermediate level of alectinib resistance [155]. however, ceritinib inhibits i1171t [136]. g1202r is an acquired mutation that drives crizotinib resistance but has also been demonstrated to confer high-level in vitro and in vivo resistance to alectinib and ceritinib [137]. a case of met amplification has been reported in an alectinib-resistant tumor. it is unclear if met was the driver of resistance, however, the patient did have a 5 month duration of response to the c-met inhibitor crizotinib [156]. given the variety of secondary alk mutations encountered, and the varying susceptibilities to second generation and investigational alk tkis, it is important to consider serial tumor molecular profiling at each step of disease progression. grace dy et al. admet & dmpk 4(3) (2016) 186-211 204 table 4. selected alk resistance mutations [157-162] alk mutation treatment for mutation l1196m alectinib, ceritinib, brigatinib, lorlatinib, jh-viii-157-02* g1269a alectinib, ceritinib, brigatinib, lorlatinib, jh-viii-157-02* 1151tins alectinib, lorlatinib i1171t/n/s ceritinib, brigatinib s1206y alectinib, ceritinib, brigatinib, lorlatinib, jh-viii-157-02* s1206f brigatinib c1156y alectinib, ceritinib, lorlatinib, jh-viii-157-02* l1152r alectinib, lorlatinib l1198f crizotinib f1174c brigatinib f1174l/v alectinib, brigatinib, lorlatinib, jh-viii-157-02* r1275q alectinib, brigatinib g1202r brigatinib, lorlatinib, jh-viii-157-02* *structural analogue of alectinib, not currently in clinical testing investigational agents (table 5) brigatinib (ap26113) is a second-generation alk tki that also inhibits egfr as well as various alk mutations resistant to crizotinib, alectinib and ceritinib, including the g1202r mutation. preclinically, it has demonstrated inhibition of egfr t790m resistance mutation as well as potent activity against ros1 with ic50 of 1.9 nm [161]. early clinical data showed disappointing activity for brigatinib in egfr-mutated patients, and it has been predominantly developed as an alk inhibitor [163]. a phase i trial investigating brigatinib for advanced alk-positive patients who had disease progression after crizotinib or chemotherapy showed an orr of 74 % and median pfs of 13.4 months [163]. the phase ii study (nct02094573) is ongoing, but preliminary results were recently presented. as of december 2015, 222 patients were randomized to either 90 mg/day or 180 mg/day of brigatinib for the treatment of crizotinib-resistant alkpositive nsclc. the orr was 46 % and the median pfs was 8.8 months for those who received 90 mg/day of brigatinib versus an orr of 54 % and median pfs of 11.1 months for those who received 180 mg/day. in the higher dose group, grade 3 or higher toxicity for increased cpk (8 %) and pneumonitis (3 %) was noteworthy. many of the pulmonary events occurred within 7 days of treatment initiation. a phase iii study is planned comparing brigatinib versus crizotinib for tki-naïve, advanced alk-positive nsclc [162]. lorlatinib (pf-06463922) is a third-generation alk/ros1 tki with activity against most alk crizotinibresistance mutations that was designed to have high cns penetration. in an ongoing phase i/ii study (nct01970865), 54 patients who had advanced alk or ros1 positive nsclc received lorlatinib, producing an orr of 50 % with an intracranial orr of 44 % (22 % cr). response was witnessed in patients with the g1202r mutation [164]. grade 3 or 4 hypercholestrolemia occurred in 9 % of patients. the phase ii component of the trial is currently enrolling patients. the l1198f mutation confers resistance to lorlatinib. interestingly, this mutation resensitizes the alk kinase domain to crizotinib, and a clinical response with crizotinib has been witnessed in a patient with a l1198f lorlatinib-resistance mutation [160]. x-396 is another tki that potently inhibits alk, including the crizotinib-resistance mutations, l1196m and c1156y. in a multicenter phase i/ii trial, among 6 alk-positive nsclc patients who received at least 200 mg of x-396, pr was achieved in 83 % with the remainder achieving stable disease. enrollment for this admet & dmpk 4(3) (2016) 186-211 tki for egfrand alk-mutated cell cancer doi: 10.5599/admet.4.3.337 205 trial is ongoing (nct01625234) [165]. asp3026 is another alk inhibitor that achieved an orr of 44 % in 15 alk-positive patients who progressed on crizotinib in a phase i trial (nct01401504) [166]. table 5. alk tkis in clinical development drug name targets clinical activity select ongoing/future trials brigatinib (asp26113) [162,163] alk egfr alk l1196m egfr t790m phase i: orr 74%, median pfs 13.4 mo; phase ii: orr 46-54% median pfs 8.8-11.1 mo ongoing phase ii (nct02094573); planned phase iii trial brigatinib vs crizotinib in tki naïve patients lorlatinib (pf-06463922) [164] alk ros1 crizotinib-resistance mutations phase i/ii: orr 50% systemic intracranial orr 44% ongoing phase ii (nct01970865) x-396 [165] alk crizotinib-resistance mutations phase i/ii (n=6) orr 83% ongoing phase ii (nct01625234) asp3026 [166] alk crizotinib-resistance mutations phase i (n=15) orr 44% no currently active trials entrectinib (rxdx-101, nms-e628) [169] alk ros1 trka, trkb, trkc phase i 1 of 1 pt with alk+ nsclc developed pr phase i/ii startk-1 (nct02097810) tsr-011 [168] alk trka, trkb, trkc phase i/ii 120 mg: orr 3 of 3 pt <120 mg: orr 5 of 9 pt ongoing phase i/ii (nct02048488) entrectinib (rxdx-101, nms-e628) and tsr-011 are inhibitors of alk as well as the tropomysin receptor kinases (trk) trka, trkb and trkc. recently, gene fusions between the kinase domain of the ntrk1 gene and other genes, including mprip and cd74 have been discovered in up to 3 % of lung cancer patients, and appear to be independent of other driver mutations. these fusions create constitutively active trk proteins that drive tumorigenesis and serve as targets for trk inhibitors [167]. a phase i/ii trial (nct02048488) is ongoing for tsr-011, but preliminary results in alk-positive patients showed 3 of 3 evaluable patients developed a response at the 120mg or above daily dose and 5 of 9 patients achieved stable disease at lower doses [168]. multiple early phase trials for entrectinib are also in progress. preliminary data from the phase i alka-372-001 trial demonstrated a partial response in an alk-positive nsclc patient [169]. another phase i/ii trial examining the safety and efficacy of entrectinib for advanced malignancies positive for ntrk1, ntrk2, ntrk3, ros1 or alk is underway (startk-1, nct02097810). immunotherapy and the future of egfr and alk tkis the therapeutic landscape for nsclc has recently expanded with the introduction of immunotherapeutic agents, including the programmed cell death protein (pd-1) inhibitors nivolumab and pembrolizumab. these drugs are now commonly used for second-line treatment of advanced nsclc, as they both prolong overall survival compared to standard second-line chemotherapy [171-173]. nivolumab also appears to have activity in first-line treatment of advanced nsclc [174,175]. clinical trials assessing neoadjuvant and adjuvant uses of immune checkpoint inhibitors are currently in progress (nct02595944, nct02504372, nct02818920, nct02486718, nct02273375, nct02572843). the future of nsclc will be closely tied to immunotherapy, and the roles of egfr and alk tkis could be called into question. however, currently, the efficacy of immunotherapy for the treatment of egfrand alk-mutated nsclc is unclear. grace dy et al. admet & dmpk 4(3) (2016) 186-211 206 subgroup analyses from check mate 057 and keynote-010 suggest patients with egfr mutations may not benefit as much from nivolumab and pembrolizumab as wild-type egfr patients. the data should be interpreted with caution as the subgroups were small and the confidence intervals on the hazard ratios were wide [171,173]. the mutational burden and neoantigen expression of tumors appear to be correlated to response to immune checkpoint blockade, which can be driven by tobacco exposure [176,177]. egfr and alk-mutated nsclcs are associated with low tobacco exposure, and these tumors can have lower mutational loads, which may partially explain lower response to immune checkpoint blockade. however, multiple studies have demonstrated that both egfr and alk activation can increase expression of pd-l1 [178-181]. combination therapy with immune checkpoint blockade and egfr tki is under investigation. in the multi-arm phase ib tatton trial, 34 patients received a combination of the anti-pd-l1 monoclonal antibody durvalumab and osimertinib. at the time of the study presentation at the 2016 european lung cancer conference (elcc), 31 patients were evaluable for response. the combination had promising activity, with partial response witnessed in 64.5 % of patients, however, grade 3/4 interstitial lung disease was reported in 14.7 % of patients [182]. much is to be learned about immunotherapy and alkand egfrmutated lung cancer, therefore, tkis will continue to represent the core treatment of egfrand alkmutated nsclc for the foreseeable future. conclusion discovery of the egfr and alk activating mutations and their respective tkis have substantially improved outcomes for a select group of nsclc patients. despite these advances, challenges remain for these patients. immunotherapy, which has recently advanced the treatment of nsclc, does not appear to be as active in alk and egfr mutated patients. acquired mutations universally lead to drug resistance to alk and egfr tkis, resulting in disease progression and death. with improvement in tumor genome sequencing technology, novel resistance mutations will continue to be identified. inevitably, new tkis will be developed that will target the most commonly encountered resistance mutations as well as disease relapse in cns, and indeed, many egfr and alk tkis are currently in development addressing these twin issues. however, innovative treatment approaches are necessary to continue to improve survival. combination therapies with tkis and novel agents, including immunotherapy, may be needed to overcome the multiple molecular pathways that ultimately drive carcinogenesis. references [1] s. mcguire, world cancer report 2014. geneva, switzerland: world health organization, international agency for research on cancer, who press, 2015. adv. nutr. 7 (2016) 418-419. [2] r.l. siegel, k.d. miller, a. jemal, ca cancer. j. clin. 66 (2016) 7-30. [3] k. chansky, j.p. sculier, j.j. crowley, et al. j. thorac. oncol. 4 (2009) 792-801. [4] j.g. paez, p.a. janne, j.c. lee, et al. science 304 (2004) 1497-1500. [5] j.y. douillard, g. ostoros, m. cobo, et al. br. j. cancer. 110 (2014) 55-62. [6] m. soda, y.l. choi, m. enomoto, et al. nature 448 (2007) 561-566. [7] k. cetin, d.s. ettinger, y.j. hei, c.d. o'malley, clin. epidemiol. 3 (2011) 139-148. [8] g.v. scagliotti, p. parikh, j. von pawel, et al. j. clin. oncol. 26 (2008) 3543-3551. [9] j.d. patel, m.a. socinski, e.b. garon, et al. j. clin. oncol. 31 (2013) 4349-4357. [10] n. thatcher, f.r. hirsch, a.v. luft, et al. lancet oncol. 16 (2015) 763-774. [11] y.l. wu, c. zhou, c.k. liam, et al. ann. oncol. 26 (2015) 1883-1889. [12] p.a. janne, j.c. yang, d.w. kim, et al. n. engl. j. med. 372 (2015) 1689-1699. admet & dmpk 4(3) (2016) 186-211 tki for egfrand alk-mutated cell cancer doi: 10.5599/admet.4.3.337 207 [13] l.v. sequist, j.c. soria, j.w. goldman, et al. n. engl. j. med. 2015;372:1700-1709. [14] a.t. shaw, d.w. kim, r. mehra, et al. n. engl. j. med. 370 (2014) 1189-1197. [15] j.f. gainor, d.s. tan, t. de pas, et al. clin. cancer res. 21 (2015) 2745-2752. [16] s. cohen, dev. biol. 12 (1965) 394-407. [17] j. mendelsohn, j. baselga, oncogene 19 (2000) 6550-6565. [18] j. downward, p. parker, m.d. waterfield, nature 311 (1984) 483-485. [19] m. scaltriti, j. baselga, clin. cancer res. 12 (2006) 5268-5272. [20] t.j. lynch, d.w. bell, r. sordella, et al. n. engl. j. med. 350 (2004) 2129-2139. [21] w. pao, v. miller, m. zakowski, et al. proc. natl. acad. sci. u.s.a. 101 (2004) 13306-13311. [22] s.v. sharma, d.w. bell, j. settleman, d.a. haber, nat. rev. cancer. 7 (2007) 169-181. [23] t. hirano, h. yasuda, t. tani, et al. oncotarget. 6 (2015) 38789-38803. [24] d.b. costa, s.t. schumer, d.g. tenen, s. kobayashi, j. clin. oncol. 26 (2008) 1182-4; author reply 1184-6. [25] f.a. shepherd, j. rodrigues pereira, t. ciuleanu, et al. n. engl. j. med. 353 (2005) 123-132. [26] f. cappuzzo, t. ciuleanu, l. stelmakh, et al. lancet oncol. 11 (2010) 521-529. [27] r. rosell, e. carcereny, r. gervais, et al. lancet oncol. 13 (2012) 239-246. [28] c. zhou, y.l. wu, g. chen, et al. lancet oncol. 12 (2011) 735-742. [29] y.l. wu, c. zhou, c.k. liam, et al. ann. oncol. 26 (2015) 1883-1889. [30] n.m. de noronha e menezes, r. lima, a. moreira, et al. eur. j. dermatol. 19 (2009) 248-251. [31] k.d. carey, a.j. garton, m.s. romero, et al. cancer res. 66 (2006) 8163-8171. [32] n. yamamoto, m. honma, h.suzuki, mol. pharmacol. 80 (2011) 466-475. [33] j.r. johnson, m. cohen, r. sridhara, et al. clin. cancer res. 11 (2005) 6414-6421. [34] m. ranson, h. shaw, j. wolf, et al. cancer chemother. pharmacol. 66 (2010) 53-58. [35] j. ling, s. fettner, b.l. lum, m. riek, a. rakhit, anticancer drugs. 19 (2008) 209-216. [36] r.w. van leeuwen, r. peric, k.g. hussaarts, et al. j. clin. oncol. 34 (2016) 1309-1314. [37] j.f. lu, s.m. eppler, j. wolf, et al. clin. pharmacol. ther. 80 (2006) 136-145. [38] o. mir, b. blanchet, f. goldwasser, n. engl. j. med. 365 (2011) 379-380. [39] c.l. o'bryant, p. haluska, l. rosen, et al. cancer chemother. pharmacol. 69 (2012) 605-612. [40] s. marchetti, n.a. de vries, t. buckle, et al. mol. cancer. ther. 7 (2008) 2280-2287. [41] a. hamada, j. sasaki, s. saeki, et al. pharmacogenomics 13 (2012) 615-624. [42] m. fukudo, y. ikemi, y. togashi, et al. clin. pharmacokinet. 52 (2013) 593-609. [43] j. li, m. zhao, p. he, m. hidalgo, clin. cancer res. 13 (2007) 3731-3737. [44] j. ling, k.a. johnson, z. miao, et al. drug metab. dispos. 34 (2006) 420-426. [45] y. togashi, k. masago, m. fukudo, et al. j. thorac. oncol. 5 (2010) 601-605. [46] y. togashi, k. masago, m. fukudo, et al. j. thorac. oncol. 5 (2010) 950-955. [47] j.e. kim, d.h. lee, y. choi, et al. lung cancer 65 (2009) 351-354. [48] n. dhruva, m.a. socinski, j. clin. oncol. 27 (2009) e31-2. [49] a. hata, r. kaji, s. fujita, n. katakami, j. thorac. oncol. 6 (2011) 653-654. [50] c. grommes, g.r. oxnard, m.g. kris, et al. neuro oncol. 13 (2011) 1364-1369. [51] c.h. yun, k.e. mengwasser, a.v. toms, et al. proc. natl. acad. sci. u.s.a. 105 (2008) 2070-2075. [52] w. pao, v.a. miller, k.a. politi, et al. plos med. (2005) 2:e73. [53] j. stamos, m.x .sliwkowski, c. eigenbrot, j. biol. chem. 277 (2002) 46265-46272. [54] j. bean, c. brennan, j.y. shih, et al. proc. natl. acad. sci. u.s.a. 104 (2007) 20932-20937. [55] j.f. gainor, m.j. niederst, j.k. lennerz, et al. j. thorac. oncol. 11 (2016) e83-5. grace dy et al. admet & dmpk 4(3) (2016) 186-211 208 [56] m.h. cohen, g.a. williams, r. sridhara, et al. clin. cancer res. 10 (2004) 1212-1218. [57] m.g. kris, r.b. natale, r.s. herbst, et al. jama 290 (2003) 2149-2158. [58] t.s. mok, y.l. wu, s. thongprasert, et al. n. engl. j. med. 361 (2009) 947-957. [59] t. mitsudomi, s. morita, y. yatabe, et al. lancet oncol. 11 (2010) 121-128. [60] m. maemondo, a. inoue, k. kobayashi, et al. n. engl. j. med. 362 (2010) 2380-2388. [61] j.y. han, k. park, s.w. kim, et al. j. clin. oncol. 30 (2012) 1122-1128. [62] c. xue, s. hong, n. li, et al. sci. rep. 5 (2015) 10648. [63] m. fukuoka, s. yano, g. giaccone, et al. j. clin. oncol. 21 (2003) 2237-2246. [64] j.y. douillard, g. ostoros, m. cobo, et al. br. j. cancer. 110 (2014) 55-62. [65] h. swaisland, a. laight, l. stafford, et al. clin. pharmacokinet. 40 (2001) 297-306. [66] h.c. swaisland, r.p. smith, a. laight, et al. clin. pharmacokinet. 44 (2005) 1165-1177. [67] j. li, j. brahmer, w. messersmith, m. hidalgo, s.d. baker, invest. new drugs. 24 (2006) 291-297. [68] d. mckillop, a.d. mccormick, a. millar, g.s. miles, p.j. phillips, m.hutchison, xenobiotica 35 (2005) 39-50. [69] d. mckillop, s.p. guy, m.p. spence, et al. xenobiotica 36 (2006) 29-39. [70] y. togashi, k. masago, s. masuda, et al. cancer chemother. pharmacol. 70 (2012) 399-405. [71] d.m. jackman, a.j. holmes, n. lindeman, et al. j. clin. oncol. 24 (2006) 4517-4520. [72] s. kobayashi, t.j. boggon, t. dayaram, et al. n. engl. j. med. 352 (2005) 786-792. [73] j.a. engelman, k. zejnullahu, t. mitsudomi, et al. science 316 (2007) 1039-1043. [74] l.v. sequist, j.c. yang, n. yamamoto, et al. j. clin. oncol. 31 (2013) 3327-3334. [75] y.l. wu, c. zhou, c.p. hu, et al. lancet oncol. 15 (2014) 213-222. [76] j.c. yang, y.l. wu, m. schuler, et al. lancet oncol. 16 (2015) 141-151. [77] k. park, e.h. tan, k. o'byrne, et al. lancet oncol. 17 (2016) 577-589. [78] j.c. soria, e. felip, m. cobo, et al. lancet oncol. 16 (2015) 897-907. [79] n. katakami, s. atagi, k. goto, et al. j. clin. oncol. 31 (2013) 3335-3341. [80] y.y. janjigian, e.f. smit, h.j. groen, et al. cancer. discov. 4 (2014) 1036-1045. [81] f. solca, g. dahl, a. zoephel, et al. j. pharmacol. exp. ther. 343 (2012) 342-350. [82] d. li, l. ambrogio, t. shimamura, et al. oncogene 27 (2008) 4702-4711. [83] s. wind, m. schmid, j. erhardt, r.g. goeldner, p. stopfer, clin. pharmacokinet. 52 (2013) 1101-1109. [84] p. stopfer, k. marzin, h. narjes, et al. cancer chemother. pharmacol. 69 (2012) 1051-1061. [85] s. wind, t. giessmann, a. jungnik, et al. clin. drug investig. 34 (2014) 173-182. [86] p. hoffknecht, a. tufman, t. wehler, et al. j. thorac. oncol. 10 (2015) 156-163. [87] s.g. wu, y.n. liu, m.f. tsai, et al. oncotarget. 7 (2016) 12404-12413. [88] s.m. kim, o.j. kwon, y.k. hong, et al. mol. cancer. ther. 11 (2012) 2254-2264. [89] k. azuma, a. kawahara, k. sonoda, et al. oncotarget. 5 (2014) 5908-5919. [90] s. huang, s. benavente, e.a. armstrong, c. li, d.l. wheeler, p.m. harari, cancer res. 71 (2011) 70717079. [91] d.a. cross, s.e. ashton, s. ghiorghiu, et al. cancer discov. 4 (2014) 1046-1061. [92] r. lawrence, et al. j. thorac. oncol. 10 (2015) s66-s890. [93] j. yang, s.s. ramalingam, p.a. janne, m. cantarini, t. mitsudomi, j. thorac. oncol. 11 (2016) s152-3. [94] s.l.greig, drugs 76 (2016) 263-273. [95] d. planchard, k.h. brown, d.w. kim, et al. cancer chemother pharmacol. 77 (2016) 767-776. [96] b. ricciuti, r. chiari, p. chiarini, et al. clin. drug investig.36 (2016) 683-686. admet & dmpk 4(3) (2016) 186-211 tki for egfrand alk-mutated cell cancer doi: 10.5599/admet.4.3.337 209 [97] j.c. yang, d.w. kim, s.w. kim, et al. osimertinib activity in patients (pts) with leptomeningeal (lm) disease from non-small cell lung cancer (nsclc): updated results from bloom, a phase i study. american society of clinical oncology annual meeting 2016. [98] h.a. yu, s.k. tian, a.e. drilon, et al. jama oncol. 1 (2015) 982-984. [99] k.s. thress, c.p. paweletz, e. felip, et al. nat. med. 21 (2015) 560-562. [100] c.a. eberlein, d. stetson, a.a. markovets, et al. cancer res. 75 (2015) 2489-2500. [101] k. park, j.s. lee, j.y. han, et al. j. thorac. oncol. 2016;11:s113-0864(16)30243-x. epub 2016 apr 15. [102] k. park, j.y. han, d.w. kim, et al. j. thorac. oncol. 2016;11:s139-0864(16)30299-4. epub 2016 apr 15. [103] h.a. yu, a. spira, l. horn, et al. antitumor activity of asp8273 300 mg in subjects with egfr mutation-positive non-small cell lung cancer: interim results from an ongoing phase 1 study. american society of clinical oncology annual meeting 2016. [104] d.s. tan, j.c. yang, n. leighl, et al. updated results of a phase 1 study of egf816, a third-generation, mutant-selective egfr tyrosine kinase inhibitor (tki), in advanced non-small cell lung cancer (nsclc) harboring t790m. american society of clinical oncology annual meeting 2016. [105] q. zhou, b. gan, l. yuan, et al. the safety profile of a selective egfr tki epitinib (hmpl-813) in patients with advanced solid tumors and preliminary clinical efficacy in egfrm+ nsclc patients with brain metastasis. american society of clinical oncology annual meeting 2016. [106] d.w. kim, j.c. yang, k. chen, et al. azd3759, an egfr inhibitor with blood brain barrier (bbb) penetration for the treatment of non-small cell lung cancer (nsclc) with brain metastasis (bm): preclinical evidence and clinical cases. american society of clinical oncology annual meeting 2015. [107] s. liu, c. aggarwal, c. brzezniak, et al. phase 2 study of tarloxotinib bromide (trlx) in patients (pts) with egfr-mutant, t790m-negative nsclc progressing on an egfr tki. american society of clinical oncology annual meeting 2016. [108] k. rikova, a. guo, q. zeng, et al. cell 131 (2007) 1190-1203. [109] j.m. heuckmann, h. balke-want, f. malchers, et al. clin. cancer res. 18 (2012) 4682-4690. [110] r. bayliss, j. choi, d.a. fennell, a.m. fry, richards mw. cell mol. life sci. 73 (2016) 1209-1224. [111] a.t. shaw, b. solomon, clin. cancer res. 17 (2011) 2081-2086. [112] e.l. kwak, y.j. bang, d.r. camidge, et al. n. engl. j. med. 363 (2010) 1693-1703. [113] d.r. camidge, y.j. bang, e.l. kwak, et al. lancet oncol. 13 (2012) 1011-1019. [114] l. crino, d.w. kim, g.j. riely. initial phase ii results with crizotinib in advanced alk-positive nonsmall cell lung cancer (nsclc): profile 1005. american society of clinical oncology annual meeting 2011. [115] a.t. shaw, d.w. kim, k. nakagawa, et al. n. engl. j. med. 368 (2013) 2385-2394. [116] b.j. solomon, t. mok, d.w. kim, et al. n. engl. j. med. 371 (2014) 2167-2177. [117] r. salgia, b.j. solomon, a.t. shaw. visual effects in anaplastic lymphoma kinase (alk)-positive advanced non-small cell lung cancer (nsclc) patients treated with crizotinib. american society of clinical oncology annual meeting 2012. [118] a.j. weickhardt, r.c. doebele, w.t. purcell, et al. cancer 119 (2013) 2383-2390. [119] a. tamiya, i. okamoto, m. miyazaki, s. shimizu, m. kitaichi, k. nakagawa, j. clin. oncol. 31 (2013) e15-7. [120] s.h. ou, w.p. tong, m. azada, c. siwak-tapp, j. dy, j.a. stiber, cancer 119 (2013) 1969-1975. [121] h.y. zou, q. li, j.h. lee, et al. cancer res. 67 (2007) 4408-4417. [122] j.g. christensen, h.y. zou, m.e. arango, et al. mol. cancer ther. 6 (2007) 3314-3322. [123] j.j. cui, m. tran-dube, h. shen, et al. j. med. chem. 54 (2011) 6342-6363. [124] u. mcdermott, a.j. iafrate, n.s. gray, et al. cancer res. 68 (2008) 3389-3395. [125] d.b. costa, s. kobayashi, s.s. pandya, et al. j. clin. oncol. 29 (2011) e443-5. grace dy et al. admet & dmpk 4(3) (2016) 186-211 210 [126] xalkori (crizotinib) [package insert]. new york, ny: pfizer inc.; 2012. [127] m.p. curran, drugs 72 (2012) 99-107. [128] c. li, c. alvey, a. bello, k.d. wilner, w. tan. pharmacokinetics (pk) of crizotinib (pf-02341066) in patients with advanced non-small cell lung cancer (nsclc) and other solid tumors. american society of clinical oncology annual meeting 2011. [129] g. metro, g. lunardi, p. floridi, et al. j. thorac. oncol. 10 (2015) e26-7. [130] d.b. costa, a.t. shaw, s.h. ou, et al. j. clin. oncol. 33 (2015) 1881-1888. [131] y.l. choi, m. soda, y. yamashita, et al. n. engl. j. med. 363 (2010) 1734-1739. [132] r.c. doebele, a.b. pilling, d.l. aisner, et al. clin. cancer res. 18 (2012) 1472-1482. [133] t. sasaki, j. koivunen, a. ogino, et al. cancer res. 71 (2011) 6051-6060. [134] s. kim, t.m. kim, d.w. kim, et al. j. thorac. oncol. 8 (2013) 415-422. [135] d. huang, d.w. kim, a. kotsakis, et al. genomics 102 (2013) 157-162. [136] l. friboulet, n. li, r. katayama, et al. cancer. discov. 4 (2014) 662-673. [137] s.h. ignatius ou, m. azada, d.j. hsiang, et al. j. thorac. oncol. 9 (2014) 549-553. [138] j. sang, j. acquaviva, j.c. friedland, et al. cancer. discov. 3 (2013) 430-443. [139] c.m. lovly, n.t. mcdonald, h. chen, et al. nat. med. 20 (2014) 1027-1034. [140] e. felip, s. orlov, k. park, et al. efficacy and safety of ceritinib in patients (pts) with advanced anaplastic lymphoma kinase (alk)-rearranged non-small cell lung cancer (nsclc): an update of ascend-1. 2014. [141] t.s. mok, d.r. spigel, a.t. shaw. ascend-2: a single-arm, open-label, multicenter phase ii study of ceritinib in adult patients (pts) with alk-rearranged (alk+) non-small cell lung cancer (nsclc) previously treated with chemotherapy and crizotinib (crz). american society of clinical oncology annual meeting 2015. [142] e. felip, s. orlov, k. park, et al. ascend-3: a single-arm, open-label, multicenter phase ii study of ceritinib in alki-naïve adult patients (pts) with alk-rearranged (alk+) non-small cell lung cancer (nsclc). american society of clinical oncology annual meeting 2015. [143] zykadia (ceritinib) [package insert]. east hanover, nj: novartis pharmaceuticals; 2014. [144] t.h. marsilje, w. pei, b. chen, et al. j. med. chem. 56 (2013) 5675-5690. [145] j.f. gainor, l. friboulet, r. katayama, et al. evolution of resistance in alk-positive patients treated with alk tyrosine kinase inhibitors (tkis). american society of clinical oncology annual meeting 2014. [146] t. kodama, t. tsukaguchi, m. yoshida, o. kondoh, h. sakamoto, cancer lett. 351 (2014) 215-221. [147] h. sakamoto, t. tsukaguchi, s. hiroshima, et al. cancer cell. 19 (2011) 679-690. [148] t. seto, k. kiura, m. nishio, et al. lancet oncol. 14 (2013) 590-598. [149] s.m. gadgeel, l. gandhi, g.j. riely, et al. lancet oncol. 15 (2014) 1119-1128. [150] a.t. shaw, l. gandhi, s. gadgeel, et al. lancet oncol. 17 (2016) 234-242. [151] h. nokihara, t. hida, m. kondo, et al. alectinib (alc) versus crizotinib (crz) in alk-inhibitor naive alk-positive non-small cell lung cancer (alk+ nsclc): primary results from the j-alex study. american society of clinical oncology annual meeting 2016. [152] k. kinoshita, k. asoh, n. furuichi, et al. bioorg. med. chem. 20 (2012) 1271-1280. [153] t. kodama, t. tsukaguchi, y. satoh, et al. mol. cancer. ther. 13 (2014) 2910-2918. [154] alecensa (alectinib) [package insert]. basel, switzerland:roche; 2015. [155] r. katayama, l. friboulet, s. koike, et al. clin. cancer res. 20 (2014) 5686-5696. [156] t. gouji, s. takashi, t. mitsuhiro, i. yukito, j.t horac. oncol. 9 (2014) e27-8. [157] h.y. zou, l. friboulet, d.p. kodack, et al. cancer. cell. 28 (2015) 70-81. [158] d. fontana, m. ceccon, c. gambacorti-passerini, l. mologni, cancer. med. 4 (2015) 953-965. admet & dmpk 4(3) (2016) 186-211 tki for egfrand alk-mutated cell cancer doi: 10.5599/admet.4.3.337 211 [159] j.m. hatcher, m. bahcall, h.g. choi, et al. j. med. chem. 58 (2015) 9296-9308. [160] a.t. shaw, l. friboulet, i. leshchiner, et al. n. engl. j. med. 374 (2016) 54-61. [161] w.s. huang, s. liu, d. zou, et al. j. med. chem. 59 (2016) 4948-4964. [162] d.w. kim, m. tiseo, m.j. ahn, et al. brigatinib (brg) in patients (pts) with crizotinib (crz)-refractory alk+ non-small cell lung cancer (nsclc): first report of efficacy and safety from a pivotal randomized phase (ph) 2 trial (alta). american society of clinical oncology annual meeting 2016. [163] d.r. camidge, l.a. bazhenova, r. salgia, et al. updated results of a first-in-human dose-finding study of the alk/egfr inhibitor ap26113 in patients with advanced malignancies. american society of clinical oncology annual meeting 2013. [164] b.j. solomon, t.m. bauer, e. felip, et al. safety and efficacy of lorlatinib (pf-06463922) from the dose-escalation component of a study in patients with advanced alk+ or ros1+ non-small cell lung cancer (nsclc). american society of clinical oncology annual meeting 2016. [165] l. horn, j. infante, g.r. blumenschein, et al. a phase i trial of x-396, a novel alk inhibitor, in patients with advanced solid tumors. american society of clinical oncology annual meeting 2014. [166] m. maitland, s.h. ou, a. tolcher, et al. safety, activity, and pharmacokinetics of an oral anaplastic lymphoma kinase (alk) inhibitor, asp3026, observed in a “fast follower” phase 1 trial design. american society of clinical oncology annual meeting 2014. [167] a. vaishnavi, m. capelletti, a.t. le, et al. nat. med. 19 (2013) 1469-1472. [168] h. arkenau, j.c. sachdev, m.m. mita, et al. phase (ph) 1/2a study of tsr-011, a potent inhibitor of alk and trk, in advanced solid tumors including crizotinib-resistant alk positive non-small cell lung cancer. american society of clinical oncology annual meeting 2015. [169] f. de braud, m. niger, s. damian, et al. alka-372-001: first-in-human, phase i study of entrectinib – an oral pan-trk, ros1, and alk inhibitor – in patients with advanced solid tumors with relevant molecular alterations. american society of clinical oncology annual meeting 2016. [170] r. van leeuwen, t. van gelder, r. mathijssen, et al. lancet oncol. 15 (2014) 315-26. [171] j. brahmer, k. reckamp, p. baas, et al. n. engl. j. med. 373 (2015) 123-35. [172] h. borghaei, l. paz-ares, l. horn, et al. n. engl. j. med. 373 (2015) 1627-39. [173] r. herbst, p. baas, d. kim, et al. lancet 387 (2016) 1540-50. [174] s. gettinger, n. rizvi, l. chow, et al. j. clin. oncol. 34 (2016) 2980-7. [175] n. rizvi, m. hellmann, j. brahmer, et al. j. clin. oncol. 34 (2016) 2969-79. [176] a. snyder, v. marakov, t. merghoub, et al. n. engl. j. med. 371 (2014) 2189-99. [177] n. rizvi, m. hellmann, a. snyder, et al. science 348 (2015) 124-8. [178] e. akbay, s. koyama, j. carretero, et al. cancer discov. 3 (2013) 1355-63. [179] n. chen, w. fang, j. zhan, et al. j. thorac. oncol. 10 (2015) 910-23. [180] k. ota, k. azuma, a. kawahara, et al. clin. cancer res. 21 (2015) 4014-21. [181] s. hong, n. chen, w. fang, et al. oncoimmunology 5 (2015) e1094598. [182] m. ahn, j. yang, h. yu, et al. osimertinib combined with durvalumab in egfr-mutant non-small cell lung cancer: results from the tatton phase ib trial. european lung cancer conference 2016. ©2016 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ causal connectivity abnormalities of regional homogeneity in children with attention deficit hyperactivity disorder: a rest-state fmri study doi: 10.5599/admet.5.4.485 242 admet & dmpk 5(4) (2017) 242-252; doi: http://dx.doi.org/10.5599/admet.5.4.485 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper causal connectivity abnormalities of regional homogeneity in children with attention deficit hyperactivity disorder: a reststate fmri study dechun zhao + , shuxing zheng + , li yang, yin tian* college of computer science and technology, bio-information college, chongqing university of posts and telecommunications, chongqing 400065, china *corresponding author: e-mail: tiany20032003@163.com; tel.: +86-23-62460536; fax: +86-23-62460536 + the authors have equal contribution to this work received: december 03, 2017; revised: december 08, 2017; published: december 24, 2017 abstract the present study aimed to investigate individual differences of causal connectivity between brain regions in attention deficit hyperactivity disorder (adhd) which was a psychiatric disorder. resting-state functional magnetic resonance imaging (r-fmri) data of typically-developing controls (tdc) children group and combined adhd (adhd-c) children group were distinguished by the support vector machine (svm) with linear kernel function, based on regional homogeneity (reho), amplitude of low frequency fluctuation (alff) and fractional alff (falff). the highest classification accuracy yielded by reho was 90.91 %. furthermore, the granger causality analysis (gca) method based on the classified weight map of regions of interesting (rois) showed that five causal flows existed significant difference between tdc and adhd-c. that is, the averaged gca values of three causal connections (i.e. left vlpfc  left cc1, right pocg  left cc1, and right pocg  right cc2) for adhd-c were separately stronger than those for tdc. and the other two connections (i.e. right fef  right sog and right cc1  right sog) were weaker for adhd-c than those for tdc. in addition, only two causality flows (i.e. left vlpfc  left cc1 and right pocg  right cc2) presented that their gca values were positively correlation with adhd index scores, respectively. our findings revealed that adhd children represented widespread abnormalities in the causality connectivity, especially involved in the attention and memory related regions. and further provided evidence that the potential neural causality flows could play a key role in characterizing individual’s adhd. keywords adhd; reho; gca; svm; fmri introduction attention deficit hyperactivity disorder (adhd) was a mental disorder with age-inappropriate symptoms, which was mainly characterized by poor sustained attention, impulsiveness (adhd-i), and hyperactivity (adhd-h). many functional neuroimaging studies had been carried out to discover the pathologies underlying the disorder. convergent evidences have demonstrated that adhd induced abnormalities both in brain structure (i.e. reduced volume and cortical thickness in frontal, anterior cingulate cortex [1], amygdala [2], hippocampus [3], cerebellum [4, 5] and occipital [6, 7]) and brain function (i.e. functional connectivity alterations among frontal-occipital, temporo-occipital [8, 9], and http://www.pub.iapchem.org/ojs/index.php/admet/index admet & dmpk 5(4) (2017) 242-252 alffs of vlpfc predict ad doi: 10.5599/admet.5.4.485 243 frontal-striatal circuits networks [10, 11]). recently, resting-state functional magnetic resonance imaging (r-fmri) has attracted increasing attention for mapping large-scale neural network function and dysfunction. seed-based methods were used to detect network alterations of adhd patients, especially the abnormalities in default mode network and visual network [12]. although, these studies provided valuable insights into the pathological mechanism of adhd, some significant limitations were clearly found. firstly, the obtained information was limited to the selected regions of interest (rois) through seed-based methods and made it difficult to examine functional connectivity patterns on a whole-brain scale [13]. secondly, traditional group-level statistical methods were not able to provide evidence for evaluating the discriminative power of the identified connections at the individual level [14]. pattern analysis based on the group level statistical methods can complement seed-based and univariate statistical analyses, which helped researchers discover diagnostic markers of disease and provide additional help for clinical diagnosis. the causal connection such as granger causality analysis (gca) reflected the directional flow of information between brain regions [15]. the causal relationship and adjustment patterns between different activation regions could be observed using gca, which provided new insights for adhd. taken together, we studied the causal connectivity of brain regions in combined adhd (adhd-c) children group using multivariable variable gca methods. firstly, we used machine learning methods to distinguish adhd-c children and typically-developing controls (tdc) children for the purpose of obtaining a more convincing difference in brain regions. secondly, we calculated the classified weight map and the rois (seeds) were selected according to the peak points of the weight map. then on the basis of gca, a network analysis was conducted. finally, we studied the correlation between adhd index scores and gca values. the results of this study can provide strong support for the cognitive process of adhd-c. materials and methods subjects forty right-handed subjects with tdc (mean age =9.61; 29 males, 11 females; mean index =45.11) and twenty-six right-handed subjects with adhd-c (mean age = 9.8, 15 males, 11 females; mean adhd index = 73.04) were investigated in the current study. children with comorbidities were not included in the study. all data were obtained from the adhd dataset of new york university (http://fcon_1000.projects.nitrc.org/indi/adhd200/). data acquisition and analysis functional images were acquired on a siemens trio 3-tesla scanner. functional images were obtained axially using echo planer imaging (epi) sequence (tr= 2000ms, te= 15ms, 33 slices, slice thickness= 4.0mm, fov= 24 ×24 cm 2 , flip angle= 90°, resolution= 256×256). the acquire time of functional imaging was 360 seconds. to facilitate the localization of functional images, high-resolution t1-weighted was spoiled gradient-recalled whole-brain volume (tr= 2530ms, te= 3.25ms, slices= 256, thickness/gap =1.33.0/0mm, fov=256×256mm 2 , resolution=256×256, flip angle=7°). data pre-processing was carried out using data processing assistant for resting-state fmri (dparsf) [16], which was based on statistical parametric mapping (spm8) (http://www.fil.ion.ucl.ac.uk/spm) and resting-state fmri data analysis toolkit [17]. pre-processing was performed as follows: 1) removal of the first ten volumes and slice timing correction; 2) the functional images were spatial normalized to montreal http://fcon_1000.projects.nitrc.org/indi/adhd200/ dechun zhao et al. admet & dmpk 5(4) (2017) 242-252 244 neurological institute (mni) space applying the unified segmentation parameters; 3) the linear trend, head motion parameter (measured by friston-24 model), white matter (wm), and cerebrospinal fluids (csf) signals were further regressed out as nuisance covariates; 4) after removing of linear trend, band-pass filtering (0.01-0.08 hz) was performed on the time course of each voxel for further regional homogeneity (reho) analysis. the reho was calculated according to the described procedure in zang et al. [18]. moreover, the spatial smoothing (fwhm = 6 mm) was conducted after reho calculation as described by previous studies. in addition, the calculation process of alff and falff could be explained by the work of zuo [19]. feature extraction, classification and gca brain network analysis for each participant, the corresponding whole brain parameters have been calculated as above. we used the rest toolbox (version 1.8. http://www.restfmri.net/forum/rest v1.8) to compare the difference between different brain images (i.e. reho, alff and falff) of adhd-c group and tdc group. two-sample t-tests (two-tailed, p<0.05) and multiple comparisons with gaussian random field (grf) theory (voxellevel p = 0.01, cluster-level p =0 .05, two-tailed) were applied for statistical testing. based on the results of correction, the voxel values in differential brain regions were used as feature for classification. the support vector machine (svm) classifier with linear kernel function was used to evaluate discrimination of adhd-c from tdc. the accuracy, sensitivity (se) and specificity (sp) were estimated by leave-one-out crossvalidation (loocv) to assess the performance of the classifier. in order to further study the causal relationship between differential brain regions of adhd-c , rois were constructed based on peak points of the weight map [20] and the information flow between brain regions of two groups was analysed by gca [21]. finally, we investigated the correlation between significant gca values and adhd scale scores in two groups. the whole processing step was shown in figure 1. figure 1. the main steps of data analysis including basic data preprocessing, parameter feature calculation (reho, alff, falff), feature selection, classification, rois selection, network construction, gca network analysis and adhd scale analysis. results by two-sample t-tests and multiple comparison correction, the differential brain regions were used as the features for classification. the classification results were shown in table 1 which indicated that the reho yielded a higher classification rate than other features (accuracy: 90.91 %, sensitivity: 92.5 %, specificity: 88.46 %). receiver operating characteristic curve (roc) was a comprehensive index reflecting admet & dmpk 5(4) (2017) 242-252 alffs of vlpfc predict ad doi: 10.5599/admet.5.4.485 245 the continuous variables of sensitivity and specificity. hence, the roc curve of reho was presented in figure 2, and it revealed the relationship between sensitivity and specificity by the method of composition. in this study, auc (area under the curve) was equalled to 0.94. table 1. classification results using different imaging feature svm_method metrics accuracy (%) sensitivity (%) specificity (%) linear kernel reho 90.91 92.5 88.46 alff 66.67 72.5 57.69 falff 74 87.5 53 figure 2: the roc curve based on the reho value. differentiated regions the significant activated regions of reho revealed by two-sample t-tests and multiple comparison correction were showed in figure 3. compared with the tdc group, the activation of left middle occipital gyrus (mog), right superior occipital gyrus (sog), bilateral frontal eye field (fef), bilateral crus i of cerebellar (cc1) and right crus ll of cerebellar (cc2) was greater in adhd. however, the decreased activation of the left ventrolateral prefrontal cortex (vlpfc), right middle frontal gyrus (mfg), right postcentral gyrus (pocg) and right temporo-parietal junction (tpj) was found in adhd. figure 3. regions with significant activation of reho map. (a) vlpfc, ventrolateral prefrontal cortex; (b) mog, middle occipital gyrus; (c) mfg, middle frontal gyrus; (d) pocg, postcentral gyrus; (e)tpj, temporo-parietal junction; (f) sog, superior occipital gyrus; (g) fef, frontal eye field; (h) left cc1, crus i of cerebellar; (i) cc2, crus ll of cerebellar; (j) right cc1, crus i of cerebellar. warm and cold colors indicated the increased and decreased activation of regions related to adhd-c respectively. here, thresholds with voxel-wise were set at p < 0.01, and thresholds with cluster-wise of gaussian random field (grf) were set at p<0.05. and cluster size >20 voxels. l: left, r: right. dechun zhao et al. admet & dmpk 5(4) (2017) 242-252 246 classified weight map based on roi the classified weight map based on roi was shown in table 2. the regions with significant contribution to classify the tdc group from adhd-c group were right pocg, right tpj, right mfg and left vlpfc. while the regions contributed to distinguish adhd-c group from tdc group were bilateral fef, bilateral cc1, right cc2, right sog and left mog. the peak coordinates of these regions were shown in table 2. overall, the differences of brain regions between two groups were mainly concentrated on the frontal lobe, occipital lobe, and cerebellum, which was consistent with the previous study [22]. table 2. the weight map of roi regions label hemisphere weight (%) voxel mni_coordinate tdc>adhd pocg r 11.76 49 (54,-12,33) tpj r 8.17 42 (54,-54,30) mfg r 7.18 62 (54,21,27) vlpfc l 6.37 37 (-54,36,6) adhd>tdc fef r 16.71 25 ( 9,18,48) cc2 r 15.61 40 (30,-78,-42) cc1 l 11.10 32 (-27,-72,-30) cc1 r 10.67 32 (9,-78,-24) mog l 7.58 29 (-39,-84,18 sog r 4.85 45 (27,-87,15) granger causality analysis based on roi in order to further study the relationships between these differential brain regions, the rois were selected according to the peak points of weight map. each roi was centered at the peak point of weight map with a radius of 8 mm. accordingly, 10 rois were taken from the weight map and the directed adjacency matrix was calculated with gca method (model order: p = 1). two-sample t-tests was carried out for the adjacency matrix of two groups (p<0.05, fdr corrected). as a result, there were five significant causal flows (left vlpfc  left cc1, right pocg  left cc1, right pocg  right cc2, right fef  the right sog, right cc1  right sog) for adhd-c and tdc groups. the average values of five significant causal flows were reported in figure 4(a) and figure 4 (b) revealed the difference via the averaged gca values of five causal flows in adhd-c minus those in tdc. the mean gca values of three causal flows (i.e. left vlpfc  left cc1, right pocg  left cc1, and right pocg  right cc2) were positive and greater in adhd group than in tdc group, while the other causal connectivity (i.e. right fef  right sog and right cc1  right sog) was negative. the positive causal effect meant an excitatory effect, whereas the negative causal effect represented an inhibitory effect [23]. notably, these results revealed that the symptoms of adhd may be caused by the dysfunction from frontal cortex to the cerebellum. admet & dmpk 5(4) (2017) 242-252 alffs of vlpfc predict ad doi: 10.5599/admet.5.4.485 247 figure 4: the significant causal flows discovered by gca in adhd-c and tdc groups. (a) the mean gca values of five significant causal flows in two groups. (b) the causal flow differences obtained by the mean gca values of adhd-c minus that of tdc. the red edge represented positive difference of adhd-c compared to tdc group. the blue edge represented the negative difference compared to tdc. correlation between the mean gca value and adhd index scores furthermore, the correlation was analysed between gca value and the adhd index scores. adhd index scores were positively correlated with gca values of left vlpfc  left cc1 and right pocg  right cc2 only in adhd group (see figure 5). figure 5: granger causal influence between regions as functions of adhd index scores in two groups. (a) left vlpfc  the right cc1; (b) right pocg  the right cc2. here, the significant correlations were observed only in adhd-c group. discussion in this paper, we tried to implement classification for both tdc and adhd-c children without any other complication illness. we found that the highest classification rate (90.91 %) was captured by reho values. the sensitivity and specificity of the model were evaluated as well. the gca analysis demonstrated that the information flow of adhd group from left vlpfc to left cc1, right pocg to left cc1 and right pocg to right cc2 was enhanced compared to tdc group. however, the information flow from right fef to the right pocg and right cc1 to the right pocg was reduced in adhd-c group. these regions were mainly concentrated on the frontal, parietal, and cerebellum, which were included in attention network. this may suggest that the etiology of adhd may be the dysfunction from frontal cortex to the cerebellum. it was noted that the attention network model was composed of three sub-networks: alertness, orientation and execution control. these sub-networks were both independent and organic in anatomy dechun zhao et al. admet & dmpk 5(4) (2017) 242-252 248 and function and this view has been supported by empirical researches in different fields [24]. alertness was defined as achieving and maintaining an alert state and orientation was required when stimuli occurred outside the current focus of attention. the executive control was defined as resolving conflicts between responses [25]. as an important part of attention network, alertness was considered as the basis for maintaining cognitive function. the role of alertness was to maintain preparation or wake-up state to accept information transmission [26]. some studies found that the alert system was associated with the thalamus, prefrontal lobe, and parietal lobe [27, 28]. attention orientation was responsible for selecting information from a large number of sensory inputs, and the main function was to shift attention to stimulus that need to be chosen or concerned. attention orientation could be divided into external orienting and internal orienting: external orienting was usually accompanied by eye movements which directed attention towards target stimulation, while internal orientation involved more paid attention of individuals to internal signals usually without obvious eye movements [25]. attention orienting was thought to be achieved by a network composed of the right temporo-parietal cortex and the right inferior frontal gyrus, which was modulated by the cholinergic system. the executive control was mainly responsible for monitoring and resolving conflicts between stimuli or reactions [26], which was related to the activities of the anterior cingulate gyrus, the lateral prefrontal cortex, and the right frontal gyrus. in our study, the abnormalities of left vlpfc, right pocg, right tpj, left mog, right sog, bilateral fef and cerebellum were found in adhd-c group. these regions were mainly included in attention network. it was often assumed that neural activity in subcortical regions was modulated in a “top-down” fashion to facilitate processing of stimuli at attended locations. in particular, the parietal and frontal regions of the right hemisphere were regarded as the source of top-down bias [29, 30]. our results showed that compared to the tdc group, the left vlpfc, right mfg, right pocg and right tpj were supressed in adhd-c, while the activations of left mog, right sog and cerebellar region were increased. the strength of the causal connection from left vlpfc to left cc1, right pocg to left cc1 and right pocg to right cc2 was enhanced in adhd-c group. the cerebellum was emerging as a key structure in several neuropsychiatric disorders, particularly in adhd [31, 32]. previous studies have shown that cerebellum mainly involved in motion control and advanced cognitive processes such as learning, attention shifting, visual spatial processing, working memory and emotion management [33-36]. the increased amount of information flowing into the cerebellum may explain the enhanced activation of our study. the vlpfc and pocg of the alert network were suppressed, which may suggest that adhd-c patients could not remain a preparation and waken state to accept information transmission and the attention was poorly sustained. taken together, the decreased activation of vlpfc and pocg may reflect that adhd-c patients suffer from the defects in maintaining a preparation and waken state. in this way, the cerebellum was more active for the higher demands of receiving and storing information. this idea was also evidenced by the increased granger influences from left vlpfc to left cc1, right pocg to left cc1 and right pocg to right cc2. the neurons of fef were associated with the inhibition of saccade production. the activation of fef decreased rapidly when the stop-signal was given in a stop-signal paradigm, while it increased for fixation [37]. in addition, the fef has been implicated in responding preferentially to attended stimuli and exerted the top-down shifts of attention to bias the processing of task-related information as evidenced by a number of studies [38]. the occipital cortex (right superior occipital gyrus) regulated by top-down control was mainly recruited during diverse tasks (i.e. auditory, visual and tactile modalities) and played an important role in visual-spatial attention, working memory. compared to the tdc group, the decrease in the strength of the causal connection between fef and sog suggested that fef may reduce control over occipital cortex in adhd-c. this reduced top-down control may suggest that adhd-c had a deficit in biasing admet & dmpk 5(4) (2017) 242-252 alffs of vlpfc predict ad doi: 10.5599/admet.5.4.485 249 the processing of behaviorally relevant objects and filtering the irrelevant information [39]. this interpretation was evidenced by neurophysiological studies. for example, attention deficit hyperactivity disorder has suffered from behavioral deficits such as a poor sustained attention, impulsiveness and hyperactivity, which explaining the reduced top-down control from symptoms [40]. attention reorienting was reflected by the activation of fronto-striatal-insular network and reduced functional connectivity of the adhd children in the fronto-striato-parieto-cerebellar network was reported by rubia [41, 42]. furthermore, the decreased fronto-parietal connectivity was found when adhd conducted the working memory and interference suppression task [43, 44]. the inputs from the cerebellum were thought to reflect the frequency and timing of events. furthermore, the cerebellum was involved in bottom-up neural systems and send bottom-up signal to posterior cortical systems for enabling the objects perceived selectively [45]. fmri studies have indicated that the abnormality of cerebellum was related to attentional impairments such as autism, attention deficit hyperactivity disorder [46, 47]. the decreased granger causality of between cerebellum and occipital cortex may suggest that adhd-c had deficient in sensory tracking related to attention [48]. this consistent with the idea that adhd had difficulty in representing transiently relevant information for attention and memory. the decreased functional connectivity was found in cerebellum and superior parietal lobule compared to the controls when adhd patients conducted a working memory task [49]. the decreased granger influence from fef to sog and from cerebellum to sog may be responsible for the dysfunction of attention orienting in adhd. in summary, we speculated that the core symptoms of adhd might derive from abnormal functions of alert network and orientation network from the gca analysis. conclusion we concluded that adhd-c children had a frontal lobe occipital cerebellar abnormality derived from the r-fmri study of reho. the classification analyses were done for distinguishing the adhd-c patients from the tdc group. the highest classification rate (90.91%) was captured by the reho with the 92.5% sensitivity and 88.46% specificity. compared to the tdc group, the attention networks were significant suppressed, whereas the cerebellar region activated which was revealed by two-sample t-tests. on the basis of gca analysis, the strength of the causal connection from left vlpfc to left cc1, right pocg to left cc1 and right pocg to right cc2 was enhanced, while the connection from right fef to the right sog and right cc1 to right sog were weakened. these results suggested that adhd have abnormal functions in alert network and orientation network for maintaining preparation state and attention. however, the cerebellum activated more for compensation mechanism. in conclusion, the above findings supported the hypothesis that adhd patients were disordered in the loop of frontal lobe to the cerebellum. acknowledgements: this research is supported by the chongqing research program of basic science and frontier technology (no. cstc2017jcyjbx0007; no. cstc2015jcyja10024) and the national natural science foundation of china (#61671097) references [1] j.c. bledsoe, m. semrudclikeman, s.r. pliszka. anterior cingulate cortex and symptom severity in attention-deficit/hyperactivity disorder. journal of abnormal psychology 122 (2013) 558-565. [2] t. frodl, j. stauber, n. schaaff, n. koutsouleris, j. scheuerecker, m. ewers, m. omerovic, m. opgenrhein, h. hampel, m. reiser. amygdala reduction in patients with adhd compared with major depression and healthy volunteers. acta psychiatrica scandinavica 121 (2010) 111-118. dechun zhao et al. admet & dmpk 5(4) (2017) 242-252 250 [3] f. amico, j. stauber, n. koutsouleris, t. frodl. anterior cingulate cortex gray matter abnormalities in adults with attention deficit hyperactivity disorder: a voxel-based morphometry study. psychiatry research neuroimaging 191 (2011) 31-35. [4] j.c. bledsoe, m. semrud-clikeman, s.r. pliszka. neuroanatomical and neuropsychological correlates of the cerebellum in children with attention-deficit/hyperactivity disorder--combined type. journal of the american academy of child & adolescent psychiatry 50 (2011) 593-601. [5] l.g. montes, j. ricardo-garcell, d.l.t. lb, h.p. alcántara, r.b. garcía, d.a. acosta, a.f. bouzas. cerebellar gray matter density in females with adhd combined type: a cross-sectional voxel-based morphometry study. journal of attention disorders 15 (2011) 368. [6] j. ahrendts, n. rüsch, m. wilke, a. philipsen, s.b. eickhoff, v. glauche, e. perlov, d. ebert, j. hennig, l.t.v. elst. visual cortex abnormalities in adults with adhd: a structural mri study. world journal of biological psychiatry the official journal of the world federation of societies of biological psychiatry 12 (2011) 260-270. [7] n. makris, l.j. seidman, a. brown, e.m. valera, j.r. kaiser, c.r. petty, l. liang, m. aleardi, d. boriel, c.s. henderson. further understanding of the comorbidity between attention-deficit/hyperactivity disorder and bipolar disorder in adults: an mri study of cortical thickness. psychiatry research 202 (2012) 1-11. [8] l. cocchi, i.e. bramati, a. zalesky, e. furukawa, l.f. fontenelle, j. moll, g. tripp, p. mattos. altered functional brain connectivity in a non-clinical sample of young adults with attentiondeficit/hyperactivity disorder. journal of neuroscience the official journal of the society for neuroscience 32 (2012) 17753-17761. [9] j.s. jensen, c. bradshaw. management of mycoplasma genitalium infections can we hit a moving target? bmc infect dis 15 (2015) 343. [10] f.x. castellanos, e.j.s. sonugabarke, m.p. milham, tannock, rosemary. characterizing cognition in adhd: beyond executive dysfunction. trends in cognitive sciences 10 (2006) 117-123. [11] f.x. castellanos, e. proal. large-scale brain systems in adhd: beyond the prefrontal-striatal model. trends in cognitive sciences 16 (2012) 17-26. [12] e.g. willcutt. the prevalence of dsm-iv attention-deficit/hyperactivity disorder: a meta-analytic review. neurotherapeutics 9 (2012) 490-499. [13] a. elton, s. alcauter, w. gao. network connectivity abnormality profile supports a categoricaldimensional hybrid model of adhd. human brain mapping 35 (2014) 4531-4543. [14] j. choi, b. jeong, s.w. lee, h.j. go. aberrant development of functional connectivity among resting state-related functional networks in medication-naïve adhd children. plos one 8 (2013) e83516. [15] y. tian, l. yang, s. chen, d. guo, z. ding, k.y. tam, d. yao. causal interactions in resting-state networks predict perceived loneliness. plos one 12 (2017) e0177443. [16] y. chao-gan, z. yu-feng. dparsf: a matlab toolbox for “pipeline” data analysis of resting-state fmri. frontiers in systems neuroscience 4 (2010). [17] x.-w. song, z.-y. dong, x.-y. long, s.-f. li, x.-n. zuo, c.-z. zhu, y. he, c.-g. yan, y.-f. zang. rest: a toolkit for resting-state functional magnetic resonance imaging data processing. plos one 6 (2011) e25031. [18] y. zang, t. jiang, y. lu, y. he, l. tian. regional homogeneity approach to fmri data analysis. neuroimage 22 (2004) 394-400. [19] x.n. zuo, m.a. di, c. kelly, z.e. shehzad, d.g. gee, d.f. klein, f.x. castellanos, b.b. biswal, m.p. milham. the oscillating brain: complex and reliable. neuroimage 49 (2010) 1432. [20] j. schrouff, j. monteiro, m. joao rosa, l. portugal, c. phillips, j. mouraomiranda. can we interpret linear kernel machine learning models using anatomically labelled regions? organization for human brain mapping, 2014. [21] c.w.j. granger. investigating causal relations by econometric models and cross-spectral methods. econometrica 37 (1969) 424-438. admet & dmpk 5(4) (2017) 242-252 alffs of vlpfc predict ad doi: 10.5599/admet.5.4.485 251 [22] q. cao, y. zang, l. sun, m. sui, x. long, q. zou, y. wang. abnormal neural activity in children with attention deficit hyperactivity disorder: a resting-state functional magnetic resonance imaging study. neuroreport 17 (2006) 1033. [23] j.p. hamilton, g. chen, m.e. thomason, m.e. schwartz, i.h. gotlib. investigating neural primacy in major depressive disorder: multivariate granger causality analysis of resting-state fmri time-series data. molecular psychiatry 16 (2011) 763-772. [24] a. raz, j. buhle. typologies of attentional networks. nature reiview neuroscience 7 (2006) 367. [25] m.i. posner, j. fan. attention as an organ system. neurobiology of perception & communication: from synapse to society de lange conference iv, 2007. [26] p. bartolomeo, the attention systems of the human brain, springer london2014. [27] j. fan, b.d. mccandliss, j. fossella, j.i. flombaum, m.i. posner. the activation of attentional networks. neuroimage 26 (2005) 471-479. [28] a. galvao-carmona, j.j. gonzález-rosa, a.r. hidalgo-muñoz, d. páramo, m.l. benítez, g. izquierdo, m. vázquez-marrufo. disentangling the attention network test: behavioral, event related potentials, and neural source analyses. frontiers in human neuroscience 8 (2014) 1261-1273. [29] g.r. fink, p.w. halligan, j.c. marshall, c.d. frith, r.s.j. frackowiak, r.j. dolan. where in the brain does visual attention select the forest and the trees? nature 382 (1996) 626-628. [30] s. kastner, m.a. pinsk, p.d. weerd, r. desimone, ungerleider, l. g. increased activity in human visual cortex during directed attention in the absence of visual stimulation. neuron 22 (1999) 751. [31] s. durston, b.j. van, z.p. de. differentiating frontostriatal and fronto-cerebellar circuits in attentiondeficit/hyperactivity disorder. biological psychiatry 69 (2011) 1178. [32] c.j. o'halloran, g.j. kinsella, e. storey. the cerebellum and neuropsychological functioning: a critical review. journal of clinical & experimental neuropsychology 34 (2012) 35-56. [33] m. steinlin. cerebellar disorders in childhood: cognitive problems. cerebellum 7 (2008) 607. [34] h. golla, p. thier, t. haarmeier. disturbed overt but normal covert shifts of attention in adult cerebellar patients. brain a journal of neurology 128 (2005) 1525. [35] r.b. ivry, r.m. spencer, h.n. zelaznik, j. diedrichsen. the cerebellum and event timing. annals of the new york academy of sciences 978 (2002) 302–317. [36] c.j. stoodley, e.m. valera, j.d. schmahmann. functional topography of the cerebellum for motor and cognitive tasks: an fmri study. neuroimage 59 (2012) 1560-1570. [37] d.d. burman, c.j. bruce. suppression of task-related saccades by electrical stimulation in the primate's frontal eye field. journal of neurophysiology 77 (1997) 2252. [38] t.j. buschman, e.k. miller. serial, covert, shifts of attention during visual search are reflected by the frontal eye fields and correlated with population oscillations. neuron 63 (2009) 386. [39] m. corbetta, g. patel, g.l. shulman. the reorienting system of the human brain: from environment to theory of mind. neuron 58 (2008) 306. [40] r.a. barkley. behavioral inhibition, sustained attention, and executive functions: constructing a unifying theory of adhd. psychological bulletin 121 (1997) 65-94. [41] k. konrad, s. neufang, c. hanisch, g.r. fink, b. herpertz-dahlmann. dysfunctional attentional networks in children with attention deficit/hyperactivity disorder: evidence from an event-related functional magnetic resonance imaging study. biological psychiatry 59 (2006) 643. [42] k. rubia, a. cubillo, a.b. smith, j. woolley, i. heyman, m.j. brammer. disorder‐specific dysfunction in right inferior prefrontal cortex during two inhibition tasks in boys with attention‐deficit hyperactivity disorder compared to boys with obsessive&ndash;compulsive disorder. human brain mapping 31 (2010) 287-299. [43] f. edin, t. klingberg, t. stödberg, j. tegnér. fronto-parietal connection asymmetry regulates working memory distractibility. journal of integrative neuroscience 6 (2007) 567-596. dechun zhao et al. admet & dmpk 5(4) (2017) 242-252 252 [44] s. neufang, g.d. fink, b, k. willmes, k. konrad. developmental changes in neural activation and psychophysiological interaction patterns of brain regions associated with interference control and time perception. neuroimage 43 (2008) 399-409. [45] b.j. casey, s. durston. from behavior to cognition to the brain and back: what have we learned from functional imaging studies of attention deficit hyperactivity disorder? american journal of psychiatry 163 (2006) 957. [46] g. allen, e. courchesne. differential effects of developmental cerebellar abnormality on cognitive and motor functions in the cerebellum: an fmri study of autism. american journal of psychiatry 160 (2003) 262-273. [47] f.x. castellanos, j.n. giedd, w.l. marsh, s.d. hamburger, a.c. vaituzis, d.p. dickstein, s.e. sarfatti, y.c. vauss, j.w. snell, n. lange. quantitative brain magnetic resonance imaging in attention-deficit hyperactivity disorder. archives of general psychiatry 53 (1996) 607. [48] m.g. paulin. the role of the cerebellum in motor control and perception. brain behavior & evolution 41 (1993) 39-50. [49] r.c. wolf, m.m. plichta, f. sambataro, a.j. fallgatter, c. jacob, k.p. lesch, m.j. herrmann, b.j. connemann, g. grön, n. vasic. regional brain activation changes and abnormal functional connectivity of the ventrolateral prefrontal cortex during working memory processing in adults with attention-deficit/hyperactivity disorder. human brain mapping 30 (2009) 2252. ©2017 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ manuscript doi: 10.5599/admet.5.1.373 14 admet & dmpk 5(1) (2017) 14-38; doi: 10.5599/admet.5.1.373 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper in vitro membrane binding and protein binding (iam mb/pb technology) to estimate in vivo distribution: applications in early drug discovery klara valko* 1 , simon teague 2 , and charles pidgeon 3 1 bio-mimetic chromatography, unit 5b business & technology centre, stevenage, sg1 2dx hertfordshire united kingdom 2 glaxosmithkline, stevenage, united kingdom 3 independent researcher, affiliate of regis technologies inc, morton grove, il 60052 usa *corresponding author: e-mail: klara_valko@bio-mimetic-chromatography.com; tel.: +44-7521989558; received: february 07, 2017; revised: march 14, 2017; published: march 25, 2017 abstract the drug discovery process can be accelerated by chromatographic profiling of the analogs to model in vivo distribution and the major non-specific binding. a balanced potency and chromatographically determined membrane and protein binding (iam mb/pb) data enable selecting drug discovery compounds for further analysis that have the highest probability to show the desired in vivo distribution behavior for efficacy and reduced chance for toxicity. although the basic principles of the technology have already a ppeared in numerous publications, the lack of standardized procedures limited its widespread applications especially in academia and small drug discovery biotech companies. in this paper, the standardized procedures are described that has been trademarked as regis iam mb/pb technology®. comparison between the drug efficiency index (dei=pic50-logvdu+2) and generally used ligand lipophilicity efficiency (lle) has been made, demonstrating the advantage of measured iam and hsa binding over calculated log p. the power of the proposed chromatographic technology is demonstrated using the data of marketed drugs. keywords drug efficiency; biomimetic properties by hplc; unbound volume of distribution; in vivo distribution introduction medicinal chemists face multi-factorial challenge problems when designing drug molecules that can reduce the impact or cure a pathological condition; drug discovery scientists seek the smallest possible dose with minimal side effects [1-4]. screening cascades often generate vast quantities of primarily in-vitro data, and in some cases, in vivo data on compounds that form a decision on which compounds are selected for further progression. this paper documents an alternative screening triage which can help to avoid the problem of analysing too much discovery data when choosing lead compounds to pursue. chemical structure design of discovery molecules usually focuses on elucidating the structure – activity relationships after identifying the target enzyme/receptor followed by the development of a potency screening method. the active molecules are then tested for a variety of assays for affinity and http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:name@address.domain admet & dmpk 5(1) (2017) 14-38 in vitro membrane and protein binding doi: 10.5599/admet.5.1.373 15 developability, such as enzyme assays, selectivity assays, cellular assays, caco2 or mdck cell permeability, solubility [5], lipophilicity, microsomal stability, hep g cell hepatotoxicity, cellular toxicity, [6] etc. all future reference to receptors will have dual meaning referring to either receptor or enzyme targets, for simplicity. project teams usually design screening cascades to guide the filtering process of compounds after each measurement, to ensure those compounds that do not meet pre-defined criteria in a particular assay are eliminated from further screening. this sequential screening process frequently excludes compounds with good in vivo drug properties; marketed drugs have appropriate in vivo properties for the disease they are being developed to treat, along with sufficient potency to engage the target and produce the desired effect. the discovery challenge is to recognise that discovery compounds may exhibit high potency, but they may lack acceptable in vivo distribution enabling the compound to achieve a therapeutic free drug concentration at the target receptor in the target tissue(s) in vivo. in vivo properties, particularly distribution, are as important as high potency or receptor occupancy. regardless of this fact, none of the properties which are typically measured during compound screening is on their own sufficient to predict the compound’s fate in vivo or in human clinical trials. this must ultimately be measured. several other properties should also be measured, which include target binding, selectivity, absorption, distribution, metabolism and elimination (adme), pharmacokinetics and pharmacodynamics (pk/pd) profile, and of course the safety profile. in order to find good absorption and bioavailability, chemists have to design molecules which will have good solubility and permeability, and in general lipophilicity. designing a drug candidate which has good clinical adme properties, is often much different than designing discovery compounds with simply high receptor binding or potency. various ligand efficiency parameters have been introduced and frequently used by medicinal chemists that relate the measured potency to some calculated properties of the compounds such as, size and lipophilicity as it has been recently reviewed [7-8]. this manuscript describes how chromatographic analysis can efficiently predict or help select which compounds have the greater probability of becoming lead molecules by balancing the receptor binding or potency with optimum compound distribution in vivo. obviously, achieving good oral absorption alone is not sufficient when choosing a drug candidate. overcoming the body’s natural defense mechanisms is also necessary to achieve relevant oral bioavailability as nature has evolved animal and human protection against potentially harmful xenobiotics. some of these defense mechanisms we need to consider when developing nce’s are: (i) gut metabolism, (ii) efflux processes and (iii) first pass metabolism. thus, ensuring that discovery molecules are not subject to active efflux and quick elimination or metabolism is current mainstream thinking in most drug discovery processes. the latest studies show that the most significant problem with the late-stage attrition of discovery molecules is related to safety and efficacy. many safety-related issues have been directly attributed to the lipophilicity of compounds [9-12] especially when it is measured using immobilized artificial membranes (iams). high iam binding is indicative of the compound’s binding to multiple targets or receptors, this is often referred to as compound promiscuity. this may cause unwanted pharmacology or toxicity. it has previously been shown that compounds that require lower clinical dose (ranging between 1-100 mg) and therefore lower efficacious plasma concentrations are less likely to cause toxicity [9,13]. the root cause of these problems is partially dominated by the compounds binding or partitioning into tissue which is predominantly comprised of membrane phospholipids; when this happens, the drug is sequestered into non-specific phospholipid binding sites and is then less available as a free drug near the valko, teague and pidgeon admet & dmpk 5(1) (2017) 14-38 16 target of interest. plasma proteins can also reduce the amount of free drug available to interact with enzymes/receptors (and when this happens an increased dose is essential). thus, the non-specific binding/partitioning of compounds to phospholipids and proteins reduce the available free concentration at the site of action. thus, increased doses are needed for activity, but, this increases the possibility of undesired pharmacology or toxicity. this paper focuses on the role of distribution in the absorption, distribution, metabolism and elimination (adme) process, without neglecting the absorption and clearance that regulate the magnitude and the frequency of dosing. we propose to design and triage nces based on their in vitro distribution behavior that determines the amount/concentration of compounds in various tissue compartments, and the free unbound compartments that are responsible for safety issues and efficacy, respectively. this can be achieved by using some of the major protein components which make up tissues in the body as chromatographic stationary phases, this is called biomimetic chromatography and has been recently reviewed [14]. we also propose that the unbound volume of distribution of compounds is an important parameter to consider, as this in vivo parameter describes the proportion of the dose relative to the free plasma concentration of compounds in steady state [15]. based on the free drug hypothesis, [16] the free plasma concentration is the same as the free tissue concentration, when no permeability barrier or active transport distorts the thermodynamic equilibrium, between plasma and tissues which contain the therapeutic targets. though active transporters may have a significant effect on the free concentration of a drug on the two sides of the membrane [17], the concentration difference can be explained by strong binding of the drug to intracellular components. the free concentration difference is usually not more than an order of magnitude on the two sides of the membrane due to active transporters that require constant energy supply to keep the concentration difference. distorting the thermodynamic equilibrium by two to three orders of magnitude would require constant energy supply (atp), and the active transporters may become easily saturated. the unbound volume of distribution can be considered as a proportional parameter to the drug partition coefficient between the free and bound compartments. in this respect, it is very similar to the recently introduced drug efficiency parameter that relates the free bio-phase concentration to the dose. the drug efficiency concept that was introduced by braggio [18] highlights the importance of the free concentration of the drug in the biophase near the target relative to the dose to achieve efficient pk/pd of a drug molecule. the drug efficiency index, dei highlights the importance of balancing the potency and drug efficiency of the molecules and is proposed as an alternative parameter to simply focusing on potency, as other ligand efficiency parameters tend to be biased towards [19]. the dei concept also helps designing drug molecules with a low therapeutic dose that in turn reduce general toxicity. table 1 shows the abbreviations and their meaning of the properties that are investigated in this paper, while table 2 shows the equations how they are calculated and how they are related to each other. the drug distribution properties are most often characterised by lipophilicity. lipophilicity has been recognised for a long time as the principal parameter that influences solubility [20,21], permeability [22], tissue binding, protein binding [23,24], toxicity [10], promiscuity [1], clearance [25] etc. several ligand efficiency parameters contain the lipophilicity and propose to consider the potency relative to the lipophilicity of the compounds, such as ligand lipophilicity efficiency, lle [7,26]. recently, the chromatographic lipophilicity has been combined with the number of aromatic rings in the molecules to derive the property forecast index (pfi) [27]. it has been suggested that having a pfi value less than 6 for admet & dmpk 5(1) (2017) 14-38 in vitro membrane and protein binding doi: 10.5599/admet.5.1.373 17 drug discovery compounds increases the probability of improving drug attrition rates. table 1. variables, abbreviations used throughout the text iam immobilized artificial membrane (hplc stationary phase_ hsa human serum albumin agp alpha-1-acid-glycoprotein ppb total plasma protein binding mb membrane binding (measured by iam) mb/pb membrane binding and protein binding vdss steady state volume of distribution (dose/plasma concentration) vdu steady state unbound volume of distribution (dose/free plasma concentration= vd/fu) fu unbound fraction in plasma de drug efficiency (100* free biophase conc/dose) demax drug efficiency measured by mb/pb by hplc (100*free plasma conc/dose) dei drug efficiency index, potency plus log demax (pic50 +log de) iam mb membrane binding index, previously known as chi(iam) clogp calculated logarithm of octanol/water partition coefficient log d/p logarithm of distribution/partition coefficient table 2. equations used to predict distribution properties log k(iam); eq. (1) = 0.29*e (0.026mb+0.42) +0.7 log k(hsa); eq. (2) = e log k(hsa) log k(hsa) = log(%hsabound/(101%hsa bound)) estimated log vdss; eq. (3) = 0.44*log k(iam) -0.22*log k(hsa) – 0.62 estimated log vdu; eq. (4) = 0.23*log k(hsa) +0.43*log k(iam) -0.72 demax; eq. (7) = 100/vdu lle ligand lipophilicity efficiency (pic50 -clogp) log k (ppb) (8) 0.87 ∗ log 𝑘(𝐻𝑆𝐴) + 0.17 ∗ log 𝑘(𝐴𝐺𝑃) + 0.06 ∗ 𝑐𝑀𝑅 − 0.27 lipophilicity has been historically characterised by the octanol/water partition coefficient, log p, for the neutral form of the molecules and by the distribution coefficient, log d, measured at different phs, mainly at ph 7.4. log p is widely used since the work of hansch [28] who suggested log p to describe the so-called “random” walk of the drug molecule that reduces its free concentration near the receptor. lipophilic compounds energetically do not favour residing in the aqueous environment and have a preference to “stick” to any hydrophobic surface in the body. the octanol/water system is able to partially mimic the polar groups with hydrogen bond donor and acceptor properties that may be present in the lipophilic environment in the body and therefore serve as a good model to describe biological distribution. however, the octanol/water lipophilicity and any other lipophilicity measure that does not include cellular membranes and proteins, fails to explain the biological distribution of molecules when they have charge or special shape. an example is the comparison of nifedipine and amlodipine (see figure 1). nifedipine is more lipophilic at physiological ph but still, the volume of distribution of amlodipine is 30 times higher, indicating 30 times higher partitioning to tissues then nifedipine. the clearance of nifedipine is 9.8 ml/min/kg [29], and clearance of amlodipine is in a similar range, 7 ml/min/kg [30]. the major difference between the two valko, teague and pidgeon admet & dmpk 5(1) (2017) 14-38 18 molecules is a basic amine group on amlodipine. the basic group results in amlodipine having a stronger binding to phospholipids than albumin, and this results in a large difference in their volumes of distribution. figure 1. physico-chemical and biomimetic properties of nifedipine and amlodipine besides measuring the octanol/water partition coefficients we propose a method of measuring the protein and phospholipid binding of the compounds using reverse phase high-performance liquid chromatography in conjunction with the membrane binding/protein binding technology (mb/pb technology®) registered to regis technologies inc. using this approach, we can measure compounds interaction with phosphatidylcholine membrane (iam), human serum albumin (hsa) and alpha-1-acid glycoprotein (agp). these are commercially available hplc stationary phases that contain or mimic these constituents of the body. these include (i) iams, (ii) chiral hsa and (iii) chrial agp. the iam concept was introduced by pidgeon et al. [31-33] and it was the first attempt to emulate the biological membrane on a solid surface. hplc columns with iam stationary phase where the phosphatidylcholine moiety is chemically bonded to a solid surface mimic the density of phosphatidylcholine in the biological membrane bi-layer. other approaches do exist for measuring membrane binding such as using liposome partition measurements or micellar electrophoresis [35-36]. however, they have their limitations and are also very time-consuming. biomimetic stationary phases have been validated by comparing the retention times obtained on the commercially available chiralpak-hsa [37-39] and chiralpak-agp [40-41] stationary phases result in binding values that are proportional to the albumin and agp binding of compounds obtained by equilibrium dialysis. when using biomimetic stationary phases the retention time of the compounds is directly proportional to the dynamic equilibrium constant between the mobile phase (buffer at physiological ph) and the actual body component (membrane and proteins) in the stationary phase. the process is very similar to the biological distribution processes that are also dynamic (never in real equilibrium unless at steady state) and occurs on the surfaces of the biomimetic stationary phases. the measured membrane binding parameters reflect the threedimensional nature of the molecule’s interaction within an in vivo system (the drug, membrane and protein admet & dmpk 5(1) (2017) 14-38 in vitro membrane and protein binding doi: 10.5599/admet.5.1.373 19 stationary phase) that has been demonstrated in the literature [14]. the chromatographic technique to measure these biomimetic properties has several other technical advantages. it can be easily automated; there is no need for concentration determination and measurements. the important thing is, however, to standardise the retention times and binding data by measuring a calibration set of compounds with each compound set to be able to get normalised retention times for each of the stationary phases being used to generate binding data (iam, hsa and agp). a detailed description of the standardised methodology for each of the iam mb/pb stationary phases has recently been published [14]. there are several publications establishing the usefulness of the iam mb/pb technology to model in vivo drug distribution, such as the volume of distribution model [42], unbound volume of distribution model [43-44] and how they can be used to screen compounds in early drug discovery [45]. it was found that strong iam mb (chi iam > 50) can be related to phospholipidosis [46-47], as well as high volume of distribution and tissue partitioning [43]. recently, it was found that promiscuity (i.e., a discovery compound binding to multiple receptors) showed good correlation to the iam mb of drug discovery compounds. the iam mb data showed a good correlation with the intracellular concentration of compounds, indicating that membrane binding is important to get compounds into the cell [48]. estimating the potential clinical dose as early as possible is an important aspect of the drug discovery process to help select the optimal compound profile and highest probability of successful progression to the clinic [49-50]. valko et al. [51] highlighted that the maximum achievable drug efficiency (demax) that can be obtained assuming (i) 100 % bioavailability, (ii) no permeability barrier and (iii) no active transport could be calculated using in vitro biomimetic measurements, essentially the regis iam mb/pb technology® that we are proposing in this paper as a useful tool in early drug discovery [51]. the drug efficiency concept as a design parameter has the advantage that it can be estimated from in vitro measurements, and can be monitored during the drug development process to see how the early estimation performs when in vivo measurements become available pre-clinically, or from human clinical trials. in this paper, the intention is to show the models and the applicability of the regis iam mb/pb technology® in early drug discovery. it is also important to have the ability to estimate these in vitro properties from the chemical structure thus helping the design stage of the molecules too. it has been shown that the retention of compounds obtained on the proposed biomimetic stationary phases can be estimated in silico [52-55], however it is worth noting that the in silico models use 2d molecular descriptors that are not sufficient to estimate the 3d contribution of the molecules binding to proteins, therefore they can be used only as a rough estimations. phospholipids constitute approximately 40 % of the human body, therefore, the phospholipid binding of compounds are a very important parameter that should not be ignored in the drug design process. it was found that mb iam values above 50 indicate promiscuous binding [57], and higher phospholipidotic potential [56-59] for discovery compound. it was also found that mb iam showed a good correlation with the total cellular concentration of discovery compounds [60]. in this paper, we present the models for human clinical unbound volume of distribution and drug efficiency using measured biomimetic properties, and we describe a protocol how to use these data in early lead optimisation. experimental the values for dose, potency (pic50), the volume of distribution, clearance, the half-life of the known drugs were obtained from various databases (www.drugbank.ca/drugs and www.drugs.com) and listed in table a1. the measured iam mb/pb data are published in the book by valko [45] and listed in table a2. table a2 contains the estimated volume of distribution (iam mb/pb log vdss) and the estimated unbound http://www.drugbank.ca/drugs http://www.drugs.com/ valko, teague and pidgeon admet & dmpk 5(1) (2017) 14-38 20 volume of distribution (iam mb/pb log vdu) using the published models [42-43], and the regis iam mb/pb technology®. the same technology has been applied to estimate the maximum drug efficiency (hplc demax) value, that represents the maximum drug efficiency that can be achieved when the bioavailability is 100 %, and there is no permeability barrier or active transport to disturb the steady state equilibrium. this can be expressed as the reciprocal value of the steady-state unbound volume of distribution [51]. the iam mb data were obtained using iam pc.dd2 hplc column (regis technologies, inc., il, usa) with the dimensions of 100 x 3 mm. mobile phase a was 50 mm ammonium acetate adjusted to ph 7.4, while mobile phase b was 100 % acetonitrile. the flow rate was 1.5 ml/ min, the linear gradient was 0 to 80 % acetonitrile in 0 to 5 min, 5 to 5.25 min 80 % acetonitrile then back to 0 % acetonitrile by 5.5 min. the total run time is 7 min. the calibration set of compounds are listed in table 3 and the calibration test solution is available from bio-mimetic chromatography, uk (www.bio-mimetic-chromatography.com). a typical chromatogram is shown in figure 2. it is essential to obtain a straight line between the retention times and the iam mb index of the calibration set of compounds. moreover, it is essential to run a system suitability test of compounds (2 neutral, 2 basic and 2 acidic compounds) listed in table 4, and check that the measured iam mb data are the same +/5 units listed in the table. this ensures that the iam phospholipid immobilised phase is maintained, preserving the natural phospholipid density on the column. figure 2. a typical chromatogram obtained on iam pc.dd2 100 x 3 mm hplc column. flow rate: 1.5 ml/min, mobile phase a; 50 mm ammonium acetate adjusted to ph 7.4, b: acetonitrile. run time 7 min; gradient: 0 to 5 min 0 to 80 % b, 5 to 5.25 min 80 % b, 5.25 to 5.5 min 0 % b it was found that the log k(iam) shows a non-linear relationship with log p octanol/water values [42]. to convert the log k(iam) scale to the log p scale we need the transformation as shown by equation (1). the converted log k(iam) data is equivalent to the partition coefficient of the compound between the ph 7.4 buffer and the phospholipid phase. log k(iam)= 0.29* e (0.046*mb iam+0.42) + 0.70 (1) this data are then be used in the models for the volume of distribution, unbound volume of distribution and the in vitro maximum drug efficiency (hplc demax). http://www.bio-mimetic-chromatography.com/ admet & dmpk 5(1) (2017) 14-38 in vitro membrane and protein binding doi: 10.5599/admet.5.1.373 21 table 3. the calibration set of compounds and their standard iam mb score on iam.pc.dd2 column. the iam mb data approximate to the acetonitrile concentration in the mobile phase when the compound elutes using the gradient conditions. the log k(iam) values obtained as 0.046iam mb+0.42 where the constants have been derived by plotting isocratic log k(iam) data extrapolated to 0 % acetonitrile mobile phase using isocratic measurements [52]. compound typical tr (min) iam mb log k(iam) octanophenone 3.18 49.4 2.69 heptanophenone 3.07 45.7 2.52 hexanophenone 2.94 41.8 2.34 valerophenone 2.79 37.3 2.14 butyrophenone 2.58 32 1.89 propiophenone 2.35 25.9 1.61 acetophenone 2.04 17.2 1.21 acetanilide 1.85 11.5 0.95 paracetamol 1.62 2.9 0.55 table 4. system suitability test compounds and their expected iam mb data. the measured values should be within 5 iam mb value agreements with the values in the table. further drug molecules data can be found in reference [45]. compound acid/base iam mb carbamazepine neutral 26.5 colchicine neutral 18.0 warfarin acidic 16.0 indomethacin acidic 24.5 nicardipine basic 45.1 propranolol basic 46.8 the pb data are obtained on chiralpack-hsa and chiralpack agp columns with the dimensions of 50 x 3 mm. the flow rate was 1.5 ml/min and 2-propanol was used as mobile phase b, while mobile phase a is 50 mm ammonium acetate adjusted ph to 7.4. the calibration set of compounds and their log k(hsa) and log k(agp) data are listed in table 5a and 5b. these data are derived from literature %hsa and %agp binding data obtained by other methodology (equilibrium dialysis and ultrafiltration). these data are used to calibrate the gradient retention times on the protein column so that the binding data obtained from the retention times are comparable with binding data obtained by other methodologies. the typical retention times were obtained using a 6 min gradient run: 0 to 3 min 0 to 35 % 2-propanol, 3 to 4 min 35 % 2propanol, 4 to 4.2 min back to 0 % 2-propanol. typical calibration plots and chromatograms have been published previously [14]. it was found that the obtained log k(hsa) showed a non-linear relationship with the log p values of acetophenone homologues [42]. to convert the log k(hsa) data to the log p scale we need to transform them using equation 2. the so obtained log k(hsa) data are used in the models. log k(hsa) = e log k(hsa) (2) the published in vivo models used the above-described mb/pb data using the following equations below: mb/pb log vdss= 0.44*log k(iam)-0.22*log k(hsa) 0.66 (3) n=179 r 2 =0.76 s=0.33 valko, teague and pidgeon admet & dmpk 5(1) (2017) 14-38 22 where mb/pb log vdss is the estimated in vivo steady state volume of distribution based on 179 known drugs human clinical data [42]. the model for the unbound volume of distribution has also been published previously using the available data for 70 marketed drugs [43] and shown by equation 4. mb/pb log vdu = 0.23 log k(hsa) + 0.43 log k(iam) 0.72 (4) n=70 r 2 = 0.84 s=0.32 the hplc demax data were calculated using the reciprocal value of the mb/pb log vdu data and validated against human clinical demax data [51]. table 5. the literature % binding data and the log k(hsa) (a) and log k(agp) (b) data for the calibration set of compounds. the logarithmic retention times measured under a given condition should result in a straight line when plotted against the log k data. the regression coefficient should be above 0.96 and the warfarin enantiomers should be baseline separated in order to pass the suitability of the column for protein binding determination. (a) bio-mimetic chromatography calibration set of compounds typical tr (min) log tr % hsa from literature plasma protein binding log k(hsa) (as log(%hsa/(101%hsa)) warfarin 3.267 0.51 97.90 1.50 paracetamol 0.285 -0.55 14.00 -0.79 nizatidine 0.293 -0.53 20.40 -0.60 trimethoprim 0.512 -0.29 37.60 -0.23 propranolol 0.895 -0.05 66.60 0.29 carbamazepine 1.216 0.08 75.00 0.46 nicardipine 2.374 0.38 95.00 1.20 indomethacin 4.117 0.61 99.50 1.82 diclofenac 3.879 0.59 99.80 1.92 (b) bio-mimetic chromatography calibration set of compounds typical tr min log tr %agp (obtained by ultrafiltration [45]) log k(agp) (expressed from %agp as log (%agp/(101-%agp)) warfarin 3.72 0.57 83.2 0.67 acetaminophen 1.16 0.06 3.2 -1.49 nizatidine 2.19 0.34 37.1 -0.24 trimethoprim 2.50 0.40 46.2 -0.07 propranolol 3.86 0.59 86 0.76 carbamazepine 3.21 0.51 73.2 0.42 nicardipine 4.08 0.61 92 1.01 indomethacin 2.98 0.47 52.9 0.04 diclofenac 3.09 0.49 69.3 0.34 for the calculation of the stepwise regression equations, the academic version of jmp (sas institute) software has been used. the calculation of the physicochemical parameters the chemaxon and acd software were used. for creating the plots the sentira version 1.0.0.6 (optibrium 2014) software has been used. admet & dmpk 5(1) (2017) 14-38 in vitro membrane and protein binding doi: 10.5599/admet.5.1.373 23 results and discussion the importance of the unbound volume of distribution and the drug efficiency in the drug discovery process has been discussed in several recent papers focusing on the clearance and plasma protein binding [15-16]. there is a general consensus that the free concentration of the compound at the site of action drives clinical efficacy together with the high affinity for the target receptor. however, a significant debate remains regarding what determines the free concentration; the dose, dosing frequency, intrinsic clearance, plasma protein binding or volume of distribution, all are important factors that should be considered. we can depict a simplified model as it was described by stepensky [15,61] and consider that the dosing amount and frequency required to balance the elimination rate of a compound determines the amount of drug in the body. while the volume of distribution describes how the total amount of drug distributes between the tissue and plasma compartment, unbound volume of distribution describes the drug distribution between the free and bound compartments regardless of the location where the drug binds (tissues or plasma proteins, see figure 3). while investigating known drugs it was clear that the volume of distribution showed no correlation alone to clearance or half-life. however, it did show a good correlation with the product term (or the sum of their logarithmic values clearance and half-life) as is shown in figure 4a, b and c. the data supports the statement that the distribution properties of compounds are an independent parameter from the clearance. figure 3. the schematic absorption, distribution and elimination process the tissue/plasma partition coefficient of a compound can be defined as the tissue concentration divided by the plasma concentration of the compounds. the concentration can be expressed as the amount of drug in the tissue and plasma volume respectively, so ktissue/plasma can be described by equation 5: 𝐾𝑡𝑖𝑠𝑠𝑢𝑒 𝑝𝑙𝑎𝑠𝑚𝑎⁄ = (𝐷𝑜𝑠𝑒−𝐴𝑝) 𝑉𝑡⁄ 𝐴𝑝 𝑉𝑝⁄ (5) where the amount of drug in tissues equals the dose minus the amount in plasma (ap), vt is the tissue volume, vp is the plasma volume. as the steady-state volume of distribution (vdss) equals the dose over the valko, teague and pidgeon admet & dmpk 5(1) (2017) 14-38 24 plasma concentration, introducing that into equation 5 we can express the tissue plasma partition as a proportionality term to the steady state volume of distribution shown in equation 6. 𝐾𝑡𝑖𝑠𝑠𝑢𝑒 𝑝𝑙𝑎𝑠𝑚𝑎⁄ = 𝑉𝑑𝑠𝑠−𝑉𝑝 𝑉𝑡 (6) in turn, the volume of distribution of marketed drugs could be modeled by the difference in the membrane and the protein binding of compounds as described by equation 3 and shown in figure 5 for the investigated compounds. there are only 40 compounds that have been included in the training set, the majority of the compounds were not included in the original model. that explains that the statistics are slightly worse (r 2 = 0.76; root mean square error = 0.33 in the original model, while here the r 2 = 0.57 and root mean square error = 0.40). figure 4. the plot of log vdss vs log in vivo cl (a), log vdss vs log t½ (b) and log vdss vs the sum of log cl and log t½. the unbound volume of distribution is in principle the reciprocal value of the maximum drug efficiency (demax) defined by braggio et al. [18], that is defined by the proportion of the free bio-phase concentration and the dose. if the free drug hypothesis is true, which means the free plasma concentration if similar to (a) (b) (c) admet & dmpk 5(1) (2017) 14-38 in vitro membrane and protein binding doi: 10.5599/admet.5.1.373 25 the free tissue concentration of the compounds when no permeability barrier or active transport distorts the equilibrium, then the model for the unbound volume of distribution can be used to estimate the maximum drug efficiency (demax) of the compounds. based on the definition of the vdu and demax, their relationship can be described by equation 7. 𝐷𝐸𝑚𝑎𝑥 = 100 𝑉𝑑𝑢 (7) figure 6 shows the unbound volume of distribution and the estimated demax of the marketed drugs listed in table a1. in this case, only 10 compounds overlapped with the original training set. most of the compounds can be considered as a test set. the root-mean-square error increased from 0.32 to 0.45 while the r 2 dropped from 0.84 to 0.66. the statistics for the training set is r 2 = 0.76 and the standard error of the estimate is s = 0.33. the estimates can be considered as very predictive, as we have just used two major binding components in the body to describe the total non-specific binding of the compounds to the phospholipids and albumin type of proteins. figure 5. the observed in vivo log vdss vs the estimated vdss by the iam mb/pb technology® using equation 3. figure 6. the plot of the human unbound volume of distribution and the estimated demax using the iam mb/pb technology and equation 4. valko, teague and pidgeon admet & dmpk 5(1) (2017) 14-38 26 it is interesting to observe that the estimated demax of the marketed drugs is typically 1 % or just above. this also means that the unbound volume of distribution of marketed drugs is typically less than 100 l/kg. thus, when we plot the potency (pic50) values in the function of the drug efficiency index (dei), which is the sum of the pic50 and the log de, the marketed drugs are around the line of unity (see figure 7). it has been observed [51] that discovery compounds usually have very narrow potency range (pic50 between 7 to 9), but a much broader range of drug efficiency index. based on the retrospective analysis it turned out that compounds that are on the right side of the line of unity survived the strict candidate selection procedure and showed a good pk/pd profile [51]. figure 7. the plot of potency in the function of dei. the line of unity represents when the drug efficiency de=1 %. the logarithm of 1 is zero, thus, dei = pic50. the yellow squares show project compounds [51], having the same good potency but a range of dei. the candidate molecules (red stars) are on the right-hand side of the line where most of the marketed drugs are. it is interesting to note that we could not find any marketed drugs with drug efficiencies higher than 1015 %. this suggests that there should be an optimum proportion of the dose and free bio-phase concentration that is less than 10-15 % drug efficiency. it is very likely, that when the high proportion of the administered amount of compound is free then the elimination rate/clearance will be high and the various total tissue concentrations may be low. therefore, we can optimise drug discovery compounds and triage them to have reasonably good potency if the drug efficiency is approximately 1 to 5 %. the regis iam mb/pb technology® that use biomimetic hplc stationary phases to measure the dynamic equilibrium constants of the compounds with the major binding components of the body enables the estimation of volume distribution and unbound volume of distribution without using in vivo experimentation for the fraction of the cost and time. it is also interesting to note that the demax values representing the sum of the albumin and phospholipid binding showed very good inverse correlation with clogp (much better than with clogd) [51]. this is because the log d drops whatever charge is present on the molecule at ph 7.4, however the sum of albumin and phospholipid binding change very little with the presence of charge. it is because positively charged compounds bind more strongly to iam (phospholipids) while negatively charged compounds have admet & dmpk 5(1) (2017) 14-38 in vitro membrane and protein binding doi: 10.5599/admet.5.1.373 27 a stronger affinity to albumin type of proteins. that is why log p is a better model to use than log d [51]. therefore, there is a connection between dei and lle. demax determined by the iam mb/pb technology is a better representation of the nonspecific binding of the compounds than log p as it is based on real albumin and iam affinity of the compounds. in our earlier paper, we have shown that iam mb/pb deimax showed a better correlation with in vivo log deimax than lle as it shown in figure 8a and 8b. figure 8. comparison of the relationship between in vivo log deimax vs lle (a) and iam mb/pb log deimax (b), respectively based on the analysis in reference [51]. the better fit is not surprising, as log p was intended to model the non-specific binding in the body caused by lipophilicity of the compounds. the iam mb/pb technology provides the direct measurements of the major types of body components. though agp was not used in the models described above, it is mainly due to statistical reasons. the agp binding and iam mb binding showed a significant correlation, so it was not possible to use both as independent variables in the regression equations. agp binds also the positively charged compounds, however, there is a different steric hindrance for molecules that have bulky substitutions near the positive charge [62]. although agp can be found in human plasma at low concentrations (approximately 1.5 %) it may be significant in clinical investigations as its concentration depends on the disease state and can increase significantly [63]. therefore compounds’ strong binding to agp should be avoided as it may cause variable free concentrations of the drug in patients with various diseases. the model to estimate total plasma protein binding (log k(ppb)) includes the agp binding as is shown by equation 8 [64]. log 𝑘 (𝑃𝑃𝐵) = 0.87 ∗ log 𝑘(𝐻𝑆𝐴) + 0.17 ∗ log 𝑘(𝐴𝐺𝑃) + 0.06 ∗ 𝑐𝑀𝑅 − 0.27 (8) n = 55 r 2 = 0.85 s=0.36 cmr is the calculated molar refraction of the compounds related to the size of the molecules and accounts for the non-specific binding of the compounds to immunoglobulins in the plasma. agp also represents important cellular components such as glycoprotein in general. there has been observed overlapping binding between agp and p-glycoprotein [65]. also, the mucus in human airways is composed predominantly of glycoproteins. several tissue binding models include the agp binding, such as valko, teague and pidgeon admet & dmpk 5(1) (2017) 14-38 28 mucus binding showed good correlation to agp binding (r 2 = 0.85) [66] and lung tissue binding showed good correlations with the weighted sum of hsa and agp bindings [67]. conclusions established screening procedures in pharmaceutical drug discovery companies cause significant delays in getting a compound to market. one reason is that the initial discovery phase typically prioritises optimising discovery molecules to bind with high affinity to the target receptor. drug analogues typically possess high receptor occupancy, which is anticipated to have better efficacy and selectivity, i.e., reduced off-target pharmacology. the insight lacking during this process is that all the potent analogues may not have the adme properties to elicit sufficient free concentration of the drug near the receptor. we have presented an hplc based technology by which the drug discovery process can be accelerated by chromatographic profiling of the analogues to model the in vivo drug disposition. the proposed methodology is able to identify compounds that have a greater probability of having acceptable in vivo properties. the methodology has been compared to ligand lipophilicity efficiency (lle) parameter that is used early in the lead optimization process. there is an inverse correlation between the maximum drug efficiency and clogp, thus lle incorporates similar efficiency metrics as dei. while clogp can be easily calculated and the drug efficiency requires simple hplc based measurements, the advantage is that it mimics the biological non-specific binding more accurately than calculated octanol/water lipophilicity. the measured iam mb/pb incorporates the effect of charge and shape of the molecules on the binding to real body components which are not reflected in their distribution in octanol and water. the technology is suitable for estimating the tissue binding, plasma protein binding, volume of distribution, unbound volume of distribution, drug efficiency, promiscuity, phospholipidotic potential, etc. therefore, this technology is recommended for use in early drug discovery programmes to aid the compound triage process so that only compounds with a higher probability of having good affinity along with good adme properties are selected for progression to expensive animal studies. the technology is commercially available; the models have been published in a book titled “physicochemical and biomimetic properties in drug discovery: chromatographic techniques for lead optimisation” [45] and are encouraged to be routinely used in the early drug discovery process. the conclusion is that in vivo data should not be generated until the compounds have been analysed by chromatographic techniques, and only compounds that are predicted to have optimal in vivo properties should be progressed for further analysis. the proposed alternate drug discovery process should include (i) potency evaluation or receptor binding and (ii) predictive in vivo properties of each analogues. only after both of these steps are completed should compounds be chosen for further development by extensive testing both in vitro and in vivo. using this approach will eliminate numerous candidates from the pool of active analogues and has the potential to help save considerable amounts of time and money during the process of getting nce’s to market. even if only 5-10 compounds are immediately eliminated as drug candidates, this potentially saves millions of dollars in the cost to get a drug to market. references [1] e. perola. j. med. chem. 53(7) (2010) 2986–2997. [2] a. cheng, d.j. diller, s.l. dixon, w.j. egan, g. lauri, k.m. merz. j. comput. chem. 23(1) (2002) 172– 183. [3] h. van de waterbeemd, d.a. smith, k. beaumont, d.k. walker. j. med. chem. 44(9) (2001) 1313– 1333. [4] i. collins, p. workman. nat. chem. biol. 2(12) (2006) 689–700. admet & dmpk 5(1) (2017) 14-38 in vitro membrane and protein binding doi: 10.5599/admet.5.1.373 29 [5] a. avdeef. admet dmpk 3(2) (2015) 84–109. [6] j. hughes, s. rees, s. kalindjian, k. philpott. br. j. pharmacol. 162(6) (2011) 1239–1249. [7] c.w. murray, d.a. erlanson, a.l. hopkins, g.m. keserü, p.d. leeson, d.c. rees, c.h. reynolds , n.j. richmond. acs med. chem. lett. 5(6) (2014) 616-618. [8] a.l. hopkins, g.m. keserü, p.d. leeson, d.c. rees, c.h. reynolds. nature rev. drug discov. 13(2) (2014) 105-121. [9] j.d. hughes, j. blagg, d.a. price, s. bailey, g.a. decrescenzo, r.v. devraj, et al. bioorg. med. chem. lett. 18(17) (2008) 4872–4875. [10] a.a. nava-ocampo, a.m. bello-ramírez. clin. exp. pharmacol. physiol. 31 (2004) 116–118. [11] n.a. meanwell. chem. res. toxicol. 24(9) (2011) 1420–1456. [12] t.j. ritchie, s.j.f. macdonald. drug discov. today 14(21-22) (2009) 1011–1020. [13] d. smith, e.f. schmid. curr. opin. drug discov. devel. 9(1) (2006) 38–46. [14] k.l. valkó. j. pharm. biomed. anal. 130 (2016) 35–54. [15] d. stepensky. eur. j. pharm. sci. 42(1-2) (2011) 91–98. [16] d.a. smith, l. di, e.h. kerns. nat. rev. drug discov. 9(12) (2010) 929–939. [17] p.d. dobson, d.b. kell. nat. rev discov. 7(3) (2008) 205–20. [18] s. braggio, d. montanari, t. rossi, e. ratti. expert opin. drug discov. 5 (2010) 609–618. [19] d. montanari, e. chiarparin, m.p. gleeson, s. braggio, r. longhi, k. valko k, et al. expert opin. drug discov. 6(9) (2011) 913–920. [20] k.a. youdim, a. avdeef, n.j. abbott. drug discov. today 8(21) (2003) 997–1003. [21] e.m. del amo, l. ghemtio, h. xhaard, m. yliperttula, a. urtti, h. kidron. plos one 8(10) (2013) 1-12. [22] p. thansandote, r.m. harris, h.l. dexter, g.l. simpson, s. pal, r.j. upton, et al. bioorg. med. chem. 23(2) (2015) 322–327. [23] p.d. leeson, b. springthorpe. nat. rev. drug discov. 6(11) (2007) 881–890. [24] b. testa, p. crivori, m. reist, p. carrupt. perspect. drug discovery des. 19 (2000) 179–211. [25] l. huang, l. berry, s. ganga, b. janosky, a. chen, j. roberts, et al. drug metab. dispos. 38(2) (2010) 223–331. [26] g.m. keserü, g.m. makara. nat. rev. drug discov. 8(3) (2009) 203–212. [27] r.j. young, d.v.s. green, c.n. luscombe, a.p. hill. drug discov. today 16(17-18) (2011) 822–830. [28] c. hansch, t. fujita. p -σ-π analysis. j. am. chem. soc. 86(8) (1964) 1616–1626. [29] m.d. donovan. j. womens health 14(1) (2005) 30–37. [30] j. faulkner, d. mcgibney, l. chasseaud, j. perry, i. taylor. br. j. clin. pharmacol. 22(1) (1986) 21–25. [31] c. pidgeon, s. ong, h. choi, h. liu. anal. chem. 66(17) (1994) 2701–2709. [32] c.y. yang, s.j. cai, h. liu, c. pidgeon. adv. drug deliv. rev. 23(96) (1996) 229–256. [33] c. pidgeon. ep0408585a1 immobilized artificial membranes, int pub patent no wo 89/08/30, filing: 21 february 1989. [34] s.k. wiedmer, m.l. riekkola, m.s. jussila. trends anal. chem. 23(8) (2004) 562–582. [35] y. kumada, m. nogami, n. minami, m. maehara, s. katoh. j. chromatogr. a 1080(1) (2005) 22–28. [36] x.y. liu, q. yang, n. kamo, j. miyake. j. chromatogr. a 913 (2001) 123–131. [37] k. valko, s. nunhuck, c. bevan, m.h. abraham, d.p. reynolds. j. pharm. sci. 92(11) (2003) 2236–2248. [38] j. reilly, d. etheridge, b. everatt, z. jiang, c. aldcroft, p. wright, et al. j. liq. chromatogr. relat. technol. 34(4) (2011) 317–327. [39] f. beaudry, m. coutu, n.k. brown. biomed. chromatogr. 13(6) (1999) 401–406. [40] m. chrysanthakopoulos, t. vallianatou, c. giaginis, a. tsantili-kakoulidou. eur. j. pharm. sci. 60 (2014) 24–31. [41] s. taheri, l.p. cogswell, a. gent, g.r. strichartz. j. pharmacol. exp. ther. 304(1) (2003) 71–80. valko, teague and pidgeon admet & dmpk 5(1) (2017) 14-38 30 [42] f. hollosy, k. valko, a. hersey, s. nunhuck, g. keri, c. bevan, et al. j. med. chem. 49(24) (2006) 6958– 6971. [43] k.l. valkó, s.b. nunhuck, a.p. hill. j. pharm. sci. 100(3) (2011) 849–862. [44] x. zhao, w. chen, z. zhou, q. wang, z. liu, r. moaddel, et al. j. chromatogr. a 1407 (2015) 176–183. [45] k. valko. physicochemical and biomimetic properties in drug discovery: chromatographic techniques for lead optimization. wiley hoboken nj; 2014. 226 p. [46] k.m. wasan, d.r. brocks, s.d. lee, k. sachs-barrable, s.j. thornton. nat. rev. drug discov. 7(1) (2008) 84–99. [47] a. casartelli, m. bonato, p. cristofori, f. crivellente, g. dal negro, i. masotto, et al. cell biol. toxicol. 19(3) (2003) 161–176. [48] l.j. gordon, m. allen, p. artursson, m.m. hann, b.j. leavens, a. mateus, et al. j. biomol. screen. 21(2) (2016) 156–164. [49] k.m. page. mol. pharm. 13(2) (2016) 609-620. [50] k. tam. admet dmpk 1(4) (2013) 63–75. [51] k. valko, e. chiarparin, s. nunhuck, d. montanari. j. pharm. sci. 101(11) (2012) 4155–4169. [52] k. valko, c.m. du, c.d. bevan, d.p. reynolds, m.h. abraham, j. pharm. sci. 89(8) (2000) 1085-1096. [53] g. russo, l. grumetto, f. barbato, g. vistoli, a. pedretti. eur. j. pharm. sci. 99 (2017) 173–184. [54] g. colmenarejo. med. res. rev. 23(3) (2003) 275–301. [55] g. lambrinidis, t. vallianatou, a. tsantili-kakoulidou. adv. drug deliv. rev. 86 (2015) 27–45. [56] p.n. mortenson, c.w. murray. j. comput. aided mol. des. 25(7) (2011) 663–667. [57] u.m. hanumegowda, g. wenke, a. regueiro-ren, r. yordanova, j.p. corradi, s.p. adams. chem. res. toxicol. 23(4) (2010) 749–755. [58] d. price, j. blagg, l. jones, n. greene, t. wager. expert opin. drug metab. toxicol. 5(8) (2009) 921– 931. [59] a. casartelli, a. lanzoni, r. comelli, f. crivellente, r. defazio, r. dorigatti, et al. toxicol. pathol. 39(2) (2011) 361–371. [60] l.j. gordon, m. allen, p. artursson, m.m. hann, b.j. leavens, a. mateus, et al. j. biomol. screen. 21(2) 2016 156-164. [61] d. stepensky. expert opin. drug metab. toxicol. 7(10) (2011) 1233–1243. [62] r. kaliszan, a. nasal, m. turowski. biomed. chromatogr. 9(5) (1995) 211–215. [63] t. fournier, n. medjoubi, d. porquet. biochim. biophys. acta 1482 (2000) 157–171. [64] k. valko. physicochemical and biomimetic properties in drug discovery: chromatographic techniques for lead optimization. wiley hoboken nj, (2014), chapter 11, p. 294-310. [65] f. zsila. curr. drug metab. 8 (2007) 563–593. [66] k. valko. physicochemical and biomimetic properties in drug discovery: chromatographic techniques for lead optimization. wiley hoboken nj, (2014) chapter 11, p. 295. [67] k. valko. physicochemical and biomimetic properties in drug discovery: chromatographic techniques for lead optimization. wiley hoboken nj, (2014), chapter 11, p. 296. admet & dmpk 5(1) (2017) 14-38 in vitro membrane and protein binding doi: 10.5599/admet.5.1.373 31 appendix table a1. the in vivo human data for the investigated marketed drugs. training set marked from the vd model [42]. drug t e st s e t / tr a in in g s e t a ct u a l d a il y d o se pic 50 d o si n g /d a y in v iv o h u m a n v d ss ( l/ k g ) in v iv o l o g v d ss in v iv o v d u (l /k g ) in v iv o l o g v d u in v iv o c l m l/ m in /k g in v iv o l o g c l t1/2 (h) log t1/2 (h) abacavir test 600 7.2 2 0.84 -0.08 1.68 0.23 13.00 1.11 1.00 0.00 acebutolol hydrochloride test 400 6.8 2 1.7 0.23 2.30 0.36 10.00 1.00 3.50 0.54 acetylsalicylic acid train. 1200 5.5 3 0.22 -0.66 0.32 -0.49 12.00 1.08 acrivastine test 32 8.5 4 1.50 0.18 albuterol sulfate test 8 7.4 4 1.9 0.28 2.07 0.31 7.80 0.89 2.40 0.38 allopurinol test 300 5.3 1 0.58 -0.24 0.60 -0.22 11.00 1.04 0.80 -0.10 alosetron hydrochloride train. 1 7.3 2 1.1 0.04 6.11 0.79 8.70 0.94 1.60 0.20 amiloride hydrochloride test 5 5.5 1 5 0.70 7.50 0.88 aminoglutethimide test 500 4.9 2 12.50 1.10 amitriptyline hydrochloride test 50 6.2 1 8.7 0.94 124.29 2.09 6.10 0.79 17.00 1.23 amoxapine train. 300 6.5 3 aripiprazole test 10 9.7 1 4.9 0.69 490.00 2.69 0.83 -0.08 75.00 1.88 atomoxetine hydrochloride1 test 40 8.7 1 0.85 -0.07 42.50 1.63 9.30 0.97 5.20 0.72 bendroflumethiazide test 3 3.4 1 8.50 0.93 bicalutamide test 50 6.1 1 144.00 2.16 bupropion hydrochloride test 300 6.3 1 24.00 1.38 cabergoline test 0 10 0 45.71 1.66 65.00 1.81 caffeine test 150 4.7 3 0.63 -0.20 0.98 -0.01 1.40 0.15 4.90 0.69 candesartan cilexetil test 8 10.5 1 0.13 -0.89 13.00 1.11 0.37 -0.43 9.00 0.95 carbamazepine train. 800 4.6 1 17.00 1.23 celecoxib test 400 8.5 2 6.13 0.79 204.33 2.31 6.59 0.82 11.00 1.04 cilostazol test 200 6.7 2 12.00 1.08 citalopram hydrobromide test 20 8.3 1 12 1.08 60.00 1.78 4.30 0.63 33.00 1.52 clomipramine hydrochloride test 25 7.9 1 13 1.11 433.33 2.64 8.20 0.91 26.00 1.41 clonidine hydrochloride train. 0 8.5 1 3.3 0.52 5.89 0.77 4.00 0.60 7.60 0.88 dapsone test 100 4.8 1 0.83 -0.08 3.32 0.52 0.48 -0.32 22.00 1.34 desloratadine test 5 9.4 1 50.00 1.70 valko, teague and pidgeon admet & dmpk 5(1) (2017) 14-38 32 diazoxide train. 150 4.7 3 0.21 -0.68 3.50 0.54 0.06 -1.22 48.00 1.68 didanosine test 250 8 2 0.77 -0.11 0.81 -0.09 11.00 1.04 1.40 0.15 diflunisal train. 1000 4.2 2 0.1 -1.00 0.10 -1.00 10.00 1.00 diphenhydramine hydrochloride train. 75 8.1 3 6.5 0.81 34.21 1.53 9.80 0.99 9.30 0.97 dipyridamole train. 200 6.7 2 1.75 0.24 175.00 2.24 3.00 0.48 0.70 -0.15 domperidone test 30 9.8 3 3.4 0.53 42.50 1.63 9.50 0.98 7.50 0.88 donepezil hydrochloride test 5 8.2 1 12 1.08 300.00 2.48 2.17 0.34 70.00 1.85 efavirenz test 600 8.9 1 47.00 1.67 etodolac test 600 6.2 2 0.39 -0.41 39.00 1.59 0.82 -0.09 7.30 0.86 felbamate train. 1200 3.4 3 0.76 -0.12 1.09 0.04 0.50 -0.30 22.00 1.34 felodipine train. 5 9.8 1 4.4 0.64 11.00 1.04 10.00 1.00 fexofenadine hydrochloride test 60 7.3 2 14.40 1.16 finasteride train. 5 7.3 1 0.89 -0.05 5.56 0.75 4.70 0.67 3.00 0.48 fluoxetine hydrochloride train. 60 8.6 1 32.5 1.51 541.67 2.73 48.00 1.68 flurbiprofen train. 200 6.2 4 0.12 -0.92 5.00 0.70 flutamide test 750 5.9 3 6.00 0.78 fluvoxamine maleate test 50 8.3 1 25 1.40 125.00 2.10 15.60 1.19 furosemide train. 80 5 2 0.12 -0.92 12.00 1.08 1.60 0.20 2.50 0.40 glibenclamide test 3 5.3 1 0.13 -0.89 13.00 1.11 1.30 0.11 1.60 0.20 glimepiride test 1 5.4 1 0.19 -0.72 19.00 1.28 0.50 -0.30 10.00 1.00 glipizide train. 3 5.5 1 0.16 -0.80 8.00 0.90 0.56 -0.25 3.30 0.52 granisetron hydrochloride test 2 9.9 2 3.7 0.57 10.57 1.02 9.10 0.96 5.20 0.72 guanabenz acetate test 8 8.2 2 6.00 0.78 haloperidol test 2 10 2 17 1.23 212.50 2.33 7.80 0.89 35.00 1.54 hydrochlorothia zide test 50 4.7 2 8.00 0.90 imipramine hydrochloride train. 75 6.6 1 12 1.08 150.00 2.18 13.00 1.11 16.00 1.20 indapamide test 1 4.2 1 14.00 1.15 indomethacin train. 75 6.5 3 0.1 -1.00 10.00 1.00 1.30 0.11 1.40 0.15 irbesartan test 150 9.3 1 0.94 -0.03 9.40 0.97 2.30 0.36 14.00 1.15 isradipine test 5 6.6 2 1.5 0.18 37.50 1.57 26.00 1.41 3.30 0.52 ketoconazole train. 200 4.7 1 2.00 0.30 ketoprofen train. 225 7.6 3 0.13 -0.89 1.63 0.21 1.60 0.20 2.10 0.32 lamotrigine test 25 4 1 1.1 0.04 2.44 0.39 0.58 -0.24 25.00 1.40 lansoprazole test 15 7.1 1 0.28 -0.55 14.00 1.15 4.40 0.64 1.00 0.00 leflunomide test 100 4.9 1 0.13 -0.89 13.00 1.11 336.00 2.53 admet & dmpk 5(1) (2017) 14-38 in vitro membrane and protein binding doi: 10.5599/admet.5.1.373 33 letrozole test 3 7.9 1 1.9 0.28 4.63 0.67 0.57 -0.24 45.00 1.65 lorazepam train. 3 8.9 3 1.3 0.11 14.44 1.16 1.00 0.00 17.00 1.23 lovastatin test 10 9.5 1 0.87 -0.06 21.75 1.34 7.20 0.86 1.40 0.15 loxapine succinate test 40 8.3 2 4.00 0.60 maprotiline hydrochloride train. 75 4.4 3 45 1.65 409.09 2.61 14.00 1.15 51.00 1.71 mebendazole train. 200 6.7 2 1.2 0.08 13.33 1.12 15.00 1.18 1.10 0.04 mercaptopurine test 3 4.8 1 1 0.00 1.18 0.07 15.00 1.18 1.00 0.00 methylprednisolone train. 4 6.3 1 1.2 0.08 5.22 0.72 6.10 0.79 2.30 0.36 metoclopramide hydrochloride test 20 8.4 2 3.2 0.51 5.33 0.73 5.70 0.76 7.20 0.86 metolazone test 3 5.7 1 14.00 1.15 mianserin train. 60 8.1 2 13.50 1.13 mibefradil dihydrochloride test 50 6.2 1 3.1 0.49 310.00 2.49 4.00 0.60 13.00 1.11 mifepristone test 600 8 1 18.00 1.26 mycophenolic acid test 1440 8 2 0.77 -0.11 38.50 1.59 2.00 0.30 12.00 1.08 nadolol train. 40 8.9 1 1.9 0.28 13.57 1.13 2.90 0.46 9.20 0.96 naproxen train. 1000 5.2 2 15.00 1.18 neostigmine bromide train. 15 7.4 1 0.74 -0.13 9.20 0.96 1.30 0.11 nevirapine test 200 8 1 1.3 0.11 4.06 0.61 0.30 -0.52 53.00 1.72 nicardipine hydrochloride train. 60 5.6 2 1 0.00 100.00 2.00 11.00 1.04 4.10 0.61 nimesulide test 400 4.4 2 2.80 0.45 nimodipine test 120 9.6 4 1.1 0.04 55.00 1.74 15.00 1.18 1.30 0.11 nisoldipine test 20 9.8 1 5.5 0.74 15.00 1.18 11.00 1.04 nitrendipine test 20 7.6 2 6.1 0.79 305.0 2.48 25.00 1.40 8.20 0.91 nortriptyline hydrochloride train. 75 9 3 22 1.34 183.33 2.26 10.00 1.00 30.00 1.48 olanzapine test 10 7.5 1 14.3 1.16 204.29 2.31 7.14 0.85 37.50 1.57 pentoxifylline train. 800 3.7 2 1.8 0.26 6.00 0.78 39.00 1.59 1.20 0.08 pergolide mesylate test 0 8.9 1 27.00 1.43 phenytoin train. 90 4.2 1 22.00 1.34 pimozide test 8 9.6 2 29.30 1.47 pioglitazone hydrochloride test 15 4.5 1 0.63 -0.20 63.00 1.80 5.00 0.70 piroxicam train. 20 6.7 1 0.14 -0.85 2.33 0.37 58.00 1.76 pravastatin sodium test 40 8.2 1 0.46 -0.34 0.92 -0.04 14.00 1.15 0.80 -0.10 prazosin hydrochloride train. 2 8.9 3 0.73 -0.14 12.17 1.09 4.70 0.67 2.00 0.30 primaquine phosphate test 15 5.4 1 4 0.60 5.80 0.76 7.10 0.85 valko, teague and pidgeon admet & dmpk 5(1) (2017) 14-38 34 probenecid train. 1000 3.8 2 0.13 -0.89 1.00 0.00 0.25 -0.60 5.90 0.77 prochlorperazin e maleate test 15 8.5 3 22 1.34 16.00 1.20 9.00 0.95 procyclidine hydrochloride test 8 8.6 3 0.74 -0.13 0.86 -0.07 12.00 1.08 promethazine hydrochloride test 50 8.2 2 14 1.15 155.56 2.19 14.00 1.15 14.00 1.15 propafenone hydrochloride test 450 5.7 3 2.2 0.34 55.00 1.74 16.00 1.20 2.10 0.32 propranolol hcl train. 160 9.4 2 3.1 0.49 23.85 1.38 12.00 1.08 3.40 0.53 protriptyline hydrochloride test 15 5.6 3 22 1.34 quinine sulfate train. 1800 6.3 6 1.8 0.26 6.00 0.78 1.90 0.28 11.00 1.04 riluzole test 100 4.4 2 12.00 1.08 roxithromycin test 450 6.8 3 0.00 12.00 1.08 saquinavir test 2000 8.3 2 3.6 0.56 120.00 2.08 13.00 1.11 13.00 1.11 simvastatin test 40 8.4 1 3.00 0.48 sulfinpyrazone train. 100 3.3 1 0.12 -0.92 6.00 0.78 0.34 -0.47 6.20 0.79 sumatriptan test 75 8.1 3 1.7 0.23 2.05 0.31 19.00 1.28 1.70 0.23 tamsulosin hydrochloride test 2 11 2 0.21 -0.68 21.00 1.32 0.62 -0.21 6.80 0.83 trazodone hydrochloride train. 150 6.3 1 0.52 -0.28 1.40 0.15 7.30 0.86 venlafaxine hydrochloride test 75 7.7 2 4.4 0.64 6.03 0.78 14.00 1.15 5.00 0.70 verapamil hydrochloride test 120 6.9 3 3.7 0.57 41.11 1.61 18.00 1.26 2.80 0.45 table a2. measured biomimetic properties and estimated volume of distribution, unbound volume of distribution based on the published models. %hsa binding is obtained from the log tr vs log k(hsa) calibration plot using %hsa= 101*10 logk(hsa) /(1+10 logk(hsa) ), log k(hsa)= exp(logk(hsa); log k(iam) obtained from measured iam mb as described by equation (1); iam mb/pb demax is calculated from vdu as 100/vdu where log vdu is obtained as a sum of log k(iam) and log k(hsa) according to equation (4); iam mb/pb log vdss is obtained from the difference between log k(iam) and log k(hsa) as described by equation 3. the test/training marks are based whether it was used in the model building for vdu [43]. drug test or training set of compound %hsa binding log k(has) converted to log p scale log k (iam) converted to log p scale iam mb/pb demax (%) iam mb/pb log vdu iam mb/pb logvdss acid/base character abacavir test 31.9 0.71 1.77 60.9 0.21 -0.04 weak base acebutolol hydrochloride test 32.7 0.73 1.59 72.1 0.13 -0.12 basic acetylsalicylic acid test 66.4 1.33 acidic acrivastine test 82.5 1.91 2.68 13.2 0.87 0.10 weak base admet & dmpk 5(1) (2017) 14-38 in vitro membrane and protein binding doi: 10.5599/admet.5.1.373 35 albuterol sulfate test 22 0.57 1.55 81.6 0.08 -0.10 basic allopurinol 19.2 0.53 2 54.9 0.26 0.10 acidic alosetron hydrochloride test 75.5 1.60 2.68 15.8 0.80 0.17 weak base amiloride hydrochloride training 38.6 0.81 2.56 27.3 0.57 0.29 acidic aminoglutethimide training 25.9 0.63 1.53 81 0.08 -0.13 weak base amitriptyline hydrochloride test 90.1 2.50 6.3 0.308 2.56 1.56 basic amoxapine test 88.4 2.33 6.64 0.24 2.67 1.75 basic aripiprazole test 97.8 4.42 6.08 0.129 2.91 1.04 basic atomoxetine hydrochloride1 test 87.5 2.25 5.28 0.922 2.07 1.17 basic bendroflumethiazide test 64.4 1.28 2.95 12.7 0.84 0.36 weak acid bicalutamide test 96.8 3.91 3.21 2.66 1.56 -0.11 neutral bupropion hydrochloride test 73.9 1.55 3.21 9.7 1.02 0.41 basic cabergoline test 83.6 1.98 6.41 0.366 2.49 1.73 basic caffeine test 26.8 0.64 1.22 108 -0.05 -0.26 neutral candesartan cilexetil test 97.9 4.48 3.8 1.07 1.94 0.03 acidic carbamazepine test 79.9 1.78 2.34 19.7 0.70 -0.02 neutral celecoxib test 97.1 4.04 3.9 1.24 1.89 0.17 neutral cilostazol test 89.9 2.48 3.02 6.97 1.15 0.12 citalopram hydrobromide test 72.4 1.50 4.85 2.12 1.71 1.14 basic clomipramine hydrochloride test 94.4 3.18 7.6 0.06 3.28 1.99 basic clonidine hydrochloride test 37.7 0.80 1.96 48.9 0.31 0.03 basic dapsone test 80.6 1.82 2.07 25 0.59 -0.15 neutral desloratadine test 88.5 2.34 6.13 0.39 2.45 1.52 basic diazoxide training 77.6 1.68 2.05 27.3 0.55 -0.13 acidic didanosine test 27.5 0.65 weak acid diflunisal test 98.8 5.22 2.49 2.57 1.55 -0.71 acidic diphenhydramine hydrochloride test 68.7 1.39 2.98 13.3 0.88 0.35 basic dipyridamole test 88.2 2.31 4.03 2.93 1.54 0.60 weak base domperidone test 92.2 2.77 3.96 2.43 1.62 0.47 basic donepezil hydrochloride test 86.1 2.14 3.68 4.45 1.36 0.49 basic efavirenz test 97.3 4.14 4.34 0.777 2.10 0.34 etodolac test 95.6 3.48 2.51 6.52 1.16 -0.32 acidic felbamate test 68.7 1.39 1.76 42.3 0.36 -0.19 neutral felodipine test 95.9 3.58 4.45 0.965 2.02 0.51 neutral valko, teague and pidgeon admet & dmpk 5(1) (2017) 14-38 36 fexofenadine hydrochloride test 74.1 1.55 2.64 16.6 0.77 0.16 basic finasteride test 88.4 2.33 3.34 5.55 1.25 0.30 fluoxetine hydrochloride test 91.2 2.63 6.19 0.305 2.55 1.48 basic flurbiprofen test 98.6 5.02 1.99 4.77 1.29 -0.89 acidic flutamide test 94.3 3.15 3.48 3.11 1.50 0.18 neutral fluvoxamine maleate test 72.3 1.49 5.15 1.59 1.84 1.28 basic furosemide test 89.7 2.46 2.14 12.4 0.77 -0.26 acidic glibenclamide test 98 4.55 2.7 2.99 1.49 -0.47 acidic glimepiride test 98 4.55 2.53 3.61 1.41 -0.55 acidic glipizide test 95.8 3.54 1.83 12 0.88 -0.63 acidic granisetron hydrochloride test 69 1.40 4.34 3.57 1.47 0.94 basic guanabenz acetate test 88.1 2.30 4.85 1.36 1.90 0.97 basic haloperidol test 90.4 2.54 4.6 1.46 1.84 0.81 basic hydrochlorothiazide test 45.4 0.92 1.62 63.3 0.19 -0.15 imipramine hydrochloride test 86.3 2.16 4.23 2.63 1.59 0.73 basic indapamide test 75.1 1.59 2.6 17.1 0.76 0.13 weak acid indometacin test 98.6 5.02 2.34 3.33 1.44 -0.74 acidic irbesartan test 96.1 3.64 2.17 8.18 1.05 -0.51 acidic isradipine test 94.4 3.18 3.48 3.05 1.51 0.17 neutral ketoconazole training 94.6 3.22 3.59 2.69 1.56 0.21 weak base ketoprofen test 98.4 4.85 2.18 4.3 1.33 -0.77 acidic lamotrigine test 59.5 1.17 1.99 38.5 0.40 -0.04 weak base lansoprazole test 90.5 2.55 2.66 9.45 1.01 -0.05 leflunomide test 92.6 2.84 3.37 4.07 1.38 0.20 neutral letrozole test 59.3 1.17 2.42 25.7 0.59 0.15 neutral lorazepam test 91.1 2.62 3.16 5.67 1.24 0.15 neutral lovastatin test 95.5 3.45 4.09 1.47 1.83 0.38 neutral loxapine succinate test 92.7 2.85 5.28 0.658 2.21 1.04 basic maprotiline hydrochloride training 86 2.13 6.7 0.25 2.65 1.82 basic mebendazole test 92.9 2.89 3.07 5.25 1.26 0.06 weak base mercaptopurine test 41.2 0.85 methylprednisolone test 71.9 1.48 2.6 18.2 0.74 0.16 neutral metoclopramide hydrochloride test 58.9 1.16 3.54 8.87 1.07 0.64 basic metolazone test 83.8 1.99 2.49 15.2 0.81 0.00 neutral mianserin test 91.9 2.73 5.47 0.595 2.26 1.15 basic mibefradil dihydrochloride test 93.6 3.01 6.3 0.231 2.68 1.45 basic mifepristone test 95.5 3.45 4.2 1.33 1.88 0.43 admet & dmpk 5(1) (2017) 14-38 in vitro membrane and protein binding doi: 10.5599/admet.5.1.373 37 mycophenolic acid test 95.5 3.45 1.83 12.7 0.86 -0.61 acidic nadolol test 39.7 0.83 1.76 58.5 0.23 -0.07 naproxen test 99.9 7.09 1.96 1.55 1.75 -1.36 acidic neostigmine bromide test 87.6 2.26 1.68 28.1 0.52 -0.42 basic nevirapine test 54 1.06 1.89 45 0.34 -0.06 neutral nicardipine hydrochloride test 95.8 3.54 4.23 1.21 1.91 0.42 weak base nimesulide test 98 4.55 2.32 4.4 1.32 -0.64 acidic nimodipine test 92 2.74 3.09 5.64 1.24 0.10 neutral nisoldipine test 91 2.61 3.16 5.67 1.24 0.16 neutral nitrendipine test 93.9 3.07 3.59 2.88 1.53 0.24 neutral nortriptyline hydrochloride training 86.2 2.15 5.66 0.673 2.21 1.36 basic olanzapine test 86.2 2.15 5.06 1.2 1.95 1.09 basic pentoxifylline test 25.9 0.63 1.47 85.5 0.06 -0.15 neutral pergolide mesylate test 85.2 2.08 5.71 0.672 2.21 1.40 basic phenytoin test 83.4 1.97 2.51 15.1 0.81 0.01 weak acid pimozide test 98.6 5.02 5.61 0.142 2.85 0.70 basic pioglitazone hydrochloride test 97.9 4.48 2.76 2.97 1.50 -0.43 piroxicam test 97.3 4.14 1.9 8.14 1.05 -0.73 acidic pravastatin sodium test 40.3 0.84 1.83 54 0.26 -0.04 acidic prazosin hydrochloride test 85.2 2.08 2.47 14.6 0.82 -0.03 weak base primaquine phosphate test 79.2 1.75 4.6 2.31 1.66 0.98 basic probenecid training 95.4 3.43 1.81 13.2 0.85 -0.62 acidic prochlorperazine maleate test 96.9 3.95 7.67 0.037 3.49 1.85 basic procyclidine hydrochloride test 86.1 2.14 4.49 1.72 1.70 0.84 neutral promethazine hydrochloride test 92.5 2.82 7.2 0.107 3.02 1.89 basic propafenone hydrochloride training 88 2.29 4.38 2.1 1.69 0.76 basic propranolol hcl test 72.7 1.51 4.45 3.04 1.54 0.97 basic protriptyline hydrochloride test 83.8 1.99 5.42 0.929 2.07 1.29 basic quinine sulfate test 78.1 1.70 5.19 1.35 1.90 1.25 basic riluzole test 94.2 3.13 3.29 3.71 1.41 0.10 weak base roxithromycin test 46.6 0.93 5.37 1.88 1.80 1.50 saquinavir test 95.2 3.37 3.77 2.07 1.68 0.26 weak base simvastatin test 96.6 3.82 4.49 0.823 2.09 0.47 sulfinpyrazone training 97.2 4.09 2.11 6.83 1.13 -0.63 acidic sumatriptan test 28 0.66 2.68 26.5 0.58 0.37 basic valko, teague and pidgeon admet & dmpk 5(1) (2017) 14-38 38 tamsulosin hydrochloride test 68.9 1.39 2.84 14.8 0.82 0.28 basic trazodone hydrochloride test 92 2.74 3.04 5.81 1.22 0.07 weak base venlafaxine hydrochloride test 35 0.76 3.19 15.5 0.83 0.58 basic verapamil hydrochloride test 88.1 2.30 3.59 3.93 1.35 0.41 basic zafirlukast test 99.1 5.57 3.67 0.789 2.14 -0.27 acidic zalcitabine test 23.6 0.60 zwitterionic zidovudine test 11.9 0.42 1.17 127 -0.12 -0.24 weak acid zileuton training 89.7 2.46 2.53 10.46 0.93 -0.09 neutral ziprasidone hydrochloride training 97.2 4.09 4.64 0.596 2.22 0.48 weak base zolmitriptan training 61.4 1.21 2.78 16.3 0.75 0.30 basic ©2017 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ manuscript doi: 10.5599/admet.4.4.341 314 admet & dmpk 4(4) (2016) 314-326; doi: 10.5599/admet.4.4.341 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper preclinical formulation for the pharmacokinetics and efficacy of gbo-006, a selective polo like kinase 2 (plk2) inhibitor for the treatment of triple negative breast cancer srinivas maddi 1 *, ravi akkireddy 1 , srinivas lenkalapelly 1 , pratima srivastava 1 , joshodeep boruwa 1 , chandra deb 1 , arnab roy chowdhury 1 , duraiswamy a. jeyaraj 1 , ramana reddy 2 , premkumar reddy 3 , manoj maniar 4 , sachin bansal 5 and jang b. gupta 1 1 gvk biosciences pvt ltd., nacharam, hyderabad 500076, india 2 department of structural and chemical biology, mount sinai school of medicine, new york, ny 10029 3 department of oncological sciences, mount sinai school of medicine, new york, ny 10029 4 onconova therapeutics, newtown, pa 18940 5 5asb life sciences pvt ltd., nacharam, hyderabad 500076, india *corresponding author: email: srinivasarao.maddi@gvkbio.com; tel: 919701166299. received: september 23, 2016; revised: december 13, 2016; published: december 26, 2016 abstract gbo-006 was shown to be a highly specific and selective plk2 inhibitor that promoted mitotic arrest in various cancer cell lines, subsequently resulting in their apoptotic death. intraperitoneal alternate day dosing of gbo-006 using 100 % dmso as formulation showed significant tumor regression in xenograft models, demonstrating proof of concept of plk2 inhibition in vivo. these studies necessitated the development of a suitable and gras (generally considered as safe) preformulation for ph armacokinetic and efficacy studies. gbo-006 possesses challenging physicochemical and biopharmaceutical properties like poor solubility in aqueous media, low permeability and a crystalline nature. different methods like cosolvency, complexation and micellar solubilization were employed to improve the solubility of gbo-006. a strategy of co-solvency is used to solubilize the gbo-006 up to 10 mg/ml. a formulation with 20 % dmso, 40 % peg 400, 30 % of 100 mm citrate buffer (ph 3.0) and 10 % solutol displayed clear solution without any visual precipitation of the drug even after 2 weeks of storage. gbo-006 showed moderate clearance in rat and high systemic clearance in mouse and dog. it showed poor oral bioavailability across all species. intraperitoneal dosing of gbo-006 demonstrated the linear exposure. gbo-006 showed significant inhibition of tumor progression. keywords gbo-006; plk2 inhibitor; pharmacokinetics; efficacy; triple negative breast cancer introduction the up regulation of polo like kinase plk1 plays a key role in mitosis but little is known about the oncogenic significance of plk2 [1-2]. plk2 kinase activity is essential for centriolar duplication and is also believed to play a regulatory role in the survival pathway [3-5]. protein tyrosine hyper-phosphorylation and the up regulation of serine/threonine kinases associated with cell cycle progression have been linked to the http://www.pub.iapchem.org/ojs/index.php/admet/index admet & dmpk 4(4) (2016) 314-326 pharmacokinetics and efficacy of plk2 inhibitor doi: 10.5599/admet.4.4.341 315 etiology of cancer. in our attempt to develop atp-mimetic compounds that are cytotoxic against a panel of cancer cell lines, we identified several sulfonyl pyridopyrimidines that exhibit cyto-toxicity at nanomolar concentrations [6-8]. the most potent of these compounds, gbo-006 (2-(1h-indol-5-ylamino)-6-(2,4difluorophenylsulfonyl)-8-methylpyrido[2,3-d]pyrimidin-7(8h)-one) was found to be a specific plk2 inhibitor when profiled against a panel of 288 wild-type, 55 mutant and 12 special kinases. in vitro testing revealed that gbo-006 was found to be plk2 specific and selective inhibitor with nanomolar potency and no inhibitory activity against plk1, plk3 and plk4. the cytotoxic effect of the drug is mediated by apoptosis as evidenced by the induction of caspase 3/7 activity and by the cleavage of parp in a dose dependent manner [9]. it affects cell cycle progression by blocking tumor cells in the g2/m phase. intraperitoneal alternate day dosing of gbo-006 using 100 % dmso as formulation lead to significant tumor regression in xenograft models, demonstrating proof of concept of plk2 inhibition in vivo. however, gbo-006 showed poor solubility in aqueous media (<0.1 μg/ml), had a low permeability and a crystalline nature, making it very hard to have a feasible formulation with moderate concentration, prompting studies to develop a preclinical formulation. in this paper, we have presented the use of co-solvent based formulation to increase the solubility of the compound. the developed formulation was subsequently used to perform pharmacokinetic studies (mice, rats and dogs) and efficacy studies, with gbo-006 as a single agent in scid beige mouse xenograft models. materials and methods chemicals and reagents gbo-006 was synthesized with the purity of 99.5 % in-house (gvk biosciences pvt. ltd, hyderabad, india). all co-solvents, surfactants, solutol, glycerin, complexing agents, lipids and oils were purchased from sigma-aldrich (st. louis, usa). liquid chromatography grade acetonitrile, methanol and ethanol were purchased from e. merck limited (mumbai, india). multi-screen solvinert inert filter plates (0.45 μm, low binding hydrophilic ptfe) were purchased from millipore. all aqueous solutions were prepared in mili-q water (millipore, ma). all other chemicals were of analytical grade. matrigel was procured from sigmaaldrich (dlw356235). solubility studies solubilization approach using co-solvent(s) and surfactant(s) were tried to optimize the best formulation for pharmacokinetic and efficacy studies. to estimate solubility qualitatively, known amount of gbo-006 was taken in clear glass vial. to this, known volume of test media was added and observed for clarity. gbo006 was dissolved in nearly 30-80 % of mixtures of co-solvents with the aid of heating and bath sonication as and when required. aqueous medium was added only if solution was obtained in co-solvents. aqueous medium consisted of solution of anti-nucleating polymer and/or surfactant in distilled water or buffer. if required, sample was intermittently bath sonicated and heated at 60-80 °c. test media was added either till a solution was obtained or till theoretical concentration was achieved (data not shown). the final sample was observed both visually and also by using polarized light microscopy for better clarity (in terms of number of visible particles). pharmacokinetics pharmacokinetics in rats male sprague dawley (sd) rats weighing 240-260 g were obtained from the taconic laboratories (vivo biotech, hyderabad, india). they were housed under standard conditions and were maintained under a 12 h light/dark cycle in the laboratory animal resources facilities, gvk biosciences pvt. ltd, hyderabad, india srinivas maddi et al. admet & dmpk 4(4) (2016) 314-326 316 (aaalac accredited). they were allowed access to commercial rodent chow and water ad libitum. rats were fasted overnight before dosing and then until 3 h after dosing. water was allowed ad libitum during the fasting period. rats were cannulated at the jugular vein for serial bleeding. three male rats per group were dosed via oral (po, 5 mg/kg) and intravenous (iv, 1 mg/kg) routes for the determination of pharmacokinetics of gbo-006. after dosing, 0.3 ml specimens of blood were collected by serial sampling from the jugular cannula at 0.25, 0.5, 1, 2, 4, 8 and 24 h after po, and 0.08, 0.25, 0.5, 1, 2, 4, 8, 24 h after iv route, respectively. after collection, blood was processed to plasma by centrifugation at 10000 rpm for 5 min, and samples were analyzed by lc-ms/ms. pharmacokinetics in mice swiss albino mice were obtained from taconic laboratories (vivo biotech, hyderabad, india). they were housed under standard conditions and were maintained under a 12 h light/dark cycle in the laboratory animal resources facilities, gvk biosciences pvt. ltd, hyderabad, india (aaalac accredited). they were allowed access to commercial rodent chow and water ad libitum. mice were fasted overnight before dosing and then until 3 h after dosing. water was allowed ad libitum during the fasting period. the bodyweights of individual mice, ranging from 25 to 30 g, were determined on the morning of the study. nine mice for oral route and twelve mice for i.v. route were dosed. after dosing, 0.1 ml specimens of blood were collected by stagger sampling with two bleeds per mouse from the retro orbital plexus at 0.5, 1, 2, 4, 8, 24 h after po route and 0.08, 0.25, 0.5, 1, 2, 4, 8, 24 h after iv route respectively. after collection, blood was processed to plasma by centrifugation at 10000 rpm for 5 min, and samples were analyzed by lc-ms/ms. dose dependent intraperitoneal pharmacokinetics in mice gbo-006 was administered in mice via ip route at 10, 30, 75 and 100 mg/kg dose to understand the dose dependent exposure. after dosing, 0.3 ml specimens of blood were collected by stagger sampling with two bleeds per mouse from the retro orbital plexus at 0.5, 1, 2, 4, 8, 24 h. blood was processed for plasma collection by the process mentioned in the above pharmacokinetics section. pharmacokinetics in dogs dog studies were performed at palamur biosciences pvt. ltd (hyderabad, india). six male beagle dogs were housed under standard conditions and were maintained under a 12 h light/dark cycle in the laboratory animal resources facilities. they were allowed access to water ad libitum. dogs were fasted overnight before po dosing and then until 4 h after dosing. water was allowed during the fasting period. the body weight of the individual dogs ranged from 9.6 to 14.3 kg. three male dogs per group were dosed per oral (po), intravenous (iv) at 5 and 1 mg/kg for the determination of pharmacokinetics of gbo-006. after dosing, 0.3 ml specimens of blood were collected by serial sampling from the saphenous vein at 0.167, 0.5,1, 2, 4, 8, 12, 24 h after po route and 0.033, 0.167, 0.5,1, 2, 4, 8, 12, 24 h after iv route, respectively. after collection, blood was processed to plasma by centrifugation at 10000 rpm for 5 min, and samples were analyzed by lc-ms/ms. liquid chromatography mass spectrometry analysis in vivo samples were prepared by protein precipitation with the addition of 200 µl acetonitrile containing internal standard (telmisartan; 200ng/ml) to 25 µl sample volume. samples were vortexed and centrifuged at 4000 rpm for 10 min. then 100 µl of supernatant was diluted with 100 µl of methanol: water mixture vortexed and injected for analysis. standard and quality control samples were prepared by adding 1 µl of aqueous known standard / quality control samples to 24 µl of blank plasma and processed as mentioned above. the selected [m+h] + and product ions monitored were m/z 468.2, 290.2 for gbo-006 admet & dmpk 4(4) (2016) 314-326 pharmacokinetics and efficacy of plk2 inhibitor doi: 10.5599/admet.4.4.341 317 (figure 1) and m/z 515.3, 276.1 for telmisartan, respectively. the high-performance liquid chromatography column was a c18-xbridge column (4.6x50 mm, 3.5 µm) maintained at 400 c with a flow rate of 1.0 ml/min with the split ratio of 60:40. the mobile phase consisted of 0.1 % formic acid (a) and methanol (b). the initial mobile phase composition was 80% a/20% b and continued until 1.0 min, then it was changed to 5 % a/95% b over 2.1 minutes. after 2.1 min, the mobile phase is slowly shifted to 80% a/20% b by 2.2 min following which it was held at same for 3.2 minutes. the high performance liquid chromatography (hplc) was interfaced to api 4000 (ab sciex, ca) mass spectrometer operated in the positive ion electrospray with multiple reaction monitoring (mrm) mode. the peak area ratio of analyte to internal standard was plotted against analyte concentrations, and the standard curve and was fitted by weighted (1/x 2 ) least-squares linear regression in the concentration ranged from 1 to 1000 ng/ml. a correlation of 0.995 was desirable for the calibration curve. the limit of quantitation for the purposes of this assay was 1 ng/ml. figure 1. lc-ms spectra representing the product ion of gbo-006 pharmacokinetic analysis the pharmacokinetic parameters were obtained by fitting the plasma concentration-time data to a noncompartmental model by using phoenix winnonlin-v6.3.0.395; (pharsight, mountain view, ca). the maximum plasma concentration (cmax) and the time to reach the same (tmax) were determined by visually inspecting the profiles of plasma drug levels versus time. the half-life of plasma drug elimination (t1/2) was the ratio of 0.693 to the slope obtained by log-linear regression of the terminal phase of the drug plasma profile. the area under the concentration-time curve to terminal time auc0-t), (area under the concentration-time curve to infinite time (auc0-∞) and the mean residence time (mrt) were also evaluated. auct was calculated by the linear/log trapezoidal rule. other pharmacokinetic variables, including the total plasma clearance (cl), steady-state volume of distribution (vss), and oral bioavailability (%f) were also calculated. xenograft study all animal studies were performed according to the guidelines and approval of the appropriate animal care committees. female scid beige mice 4–6 weeks of age were purchased from taconic laboratories (vivo biotech, hyderabad, india) and housed in a pathogen-free environment under controlled conditions +ms2 (468.30): 40.000 volts from sample 2 (tunesampleid) of gbo-006-q3.wiff (turbo spray), centroided max. 6.9e5 cps. 60 80 100 120 140 160 180 200 220 240 260 280 300 320 340 360 380 400 420 440 460 480 500 m/z, da 0.0 5.0e4 1.0e5 1.5e5 2.0e5 2.5e5 3.0e5 3.5e5 4.0e5 4.5e5 5.0e5 5.5e5 6.0e5 6.5e5 6.9e5 i n t e n s i t y , c p s 468.0 261.9 290.2 221.3 468.8 403.1 263.3 278.0228.8 srinivas maddi et al. admet & dmpk 4(4) (2016) 314-326 318 of light and humidity and allowed free access to food and water. tumors were developed from mda-mb231 cells inoculated subcutaneously (sc) into each mouse as previously described [10]. mda-mb-231 cells were grown in rpmi 1640 w/ 10 % fbs, 2 % p/s, 1.25 % l-glut, and 1 % sodium pyruvate to 80 % confluency. cells were trypsinized with 0.25 % trypsin for 3-5 minutes, washed and suspended at 100 million cells per ml. animals received 5 million cells in matrigel (1:1 ratio of media and matrigel) and injected subcutaneously at the flank. tumors were measured three times a week with vernier calipers and tumor volume measured using the formula length x width 2/2. treatment was initiated when the tumors reached an average volume of 100 mm 3 . one group was dosed with the formulation {(20 % v/v dmso + 40 % w/v peg 400 + 10 % w/v solutol + 30 % v/v citrate buffer (100mm)} with the dose volume of 10 ml/kg as a vehicle control, whereas the two other groups were dosed via intraperitoneal route with 30 and 75/mg/kg gbo-006. all dosing were done once daily (qd) for 24 days. animal body weights were taken every alternate day, and they were monitored for general health status and any possible signs of treatment related toxicity. results development of formulation of gbo-006 various percentage compositions of water soluble organic cosolvents, surfactants were evaluated for an optimized gbo-006 formulation (table 1). gbo-006 concentrations in these formulations were kept fixed at 10 mg/ml. among these formulations (#a-k), gbo-006 was initially dissolved and subsequently exhibited various degrees of precipitation after 24 h storage at room temperature. however, formulation #l with 20 % dmso, 40 % peg 400, 30 % of 100mm citrate buffer (ph 3.0) and 10 % solutol displayed clear solutions without any visual precipitation of the drug even after 2 weeks of storage. therefore, this formulation was selected for in vivo pharmacokinetic and efficacy studies. table 1. formulation trails for gbo-006 at 10 mg/ml by solvation strategy using co-solvents and surfactants formula vehicle composition appearance at crt a 20% v/v nmp + 20% v/v ethanol + 20% v/v peg 400 + 10% w/v tpgs + 1% w/v poloxamer 188 + purified water q.s. precipitation after 8h b 20% v/v nmp + 20% w/v glycerin+ 20% v/v peg 400 +10% w/v solutol hs 15 + 10% w/v poloxamer 188 +purified water q.s. precipitation after 48h c 20% v/v nmp + 20% v/v ethanol+ 20% v/v peg 400 +10% w/v solutol hs 15 + 10% w/v tpgs + purified water q.s. precipitation after 48h d 20% v/v nmp + 20% w/v glycerin+ 20% v/v ethanol +10% w/v tpgs + 10% w/v poloxamer 188 + purified water q.s. precipitation after 12h f 20% v/v nmp + 10% v/v dmso + 30% v/v peg 400 + 20% w/v cremophor el + + purified water q.s precipitation after 24h g 20% v/v nmp + 10% v/v dmso + 30% v/v peg 400 + 10% w/v solutol hs 15 + + purified water q.s precipitation after 48h h 20% v/v nmp + 10% v/v dmso + 30% v/v peg 400 + 20% w/v tpgs + purified water q.s precipitation after 48h i 20% v/v nmp + 10% v/v dmso + 30% v/v peg 400 + 20% w/v poloxamer 188 + purified water q.s precipitation after 48h j 20% v/v nmp + 20% v/v dmso + 30% v/v peg 400 + 20% v/v labrasol + purified water q.s precipitation after 72 h k 20% v/v dmso + 30% w/v peg 400 + 10% w/v solutol + 30 % v/v citrate buffer (100mm) precipitation after 1 week l 20% v/v dmso + 40% w/v peg 400 + 10% w/v solutol + 30 % v/v citrate buffer (100mm) no visual precipitation up to 2 weeks we observed that the solubility of gbo-006 in dmso, a pharmaceutically acceptable strong solubilizing agent, was quite high. hence, dmso was chosen as the solubilizing agent in our formulation. the peg 400 (at 40 % v/v) was found to be a suitable additive in the formulation, as we observed a clear solution of high miscibility and compatibility with water. however, at a lower concentration of 10-30 % v/v, peg 400 was found to be insufficient to prevent precipitation of free gbo-006, observed after 24 h. therefore, peg 400 was fixed at a concentration of 40 % v/v. admet & dmpk 4(4) (2016) 314-326 pharmacokinetics and efficacy of plk2 inhibitor doi: 10.5599/admet.4.4.341 319 among the different percentage compositions of the nonionic surfactants considered, including tween 80, cremophor, labrasol, transcutol and solutol, the latter surfactant was picked as an ingredient in the optimal formulation at a concentration of 10 % w/v. this was purely based on the solubility, clarity and miscibility of solutol at this concentration with the rest of the additives in the formulation. it was also found that a percentage concentration of 5 % w/v is insufficient to prevent the precipitation of the free gbo-006. hence, we decided to optimize the formulation with at a percentage concentration of 10 % w/v. lc-ms/ms analytical method no interfering peaks were observed in blank plasma chromatograms at gbo-006 retention time (figure 2). the retention times for gbo-006 and internal standard were 1.58 and 1.80 min, respectively. the calibration curves and linearity was determined using weighted (1/x 2 ) linear regression analysis in the following ranges of 1 to 1000 ng/ml in rat, mouse and dog plasma. the correlation coefficients were > 0.995 for calibration curves. the batch was accepted once the low, medium and high quality control (qc) samples and standard samples met the acceptance criteria. pharmacokinetics in rats a plasma concentration and time profile of gbo-006 obtained is shown in figure 3 and corresponding pharmacokinetic parameters are shown in table 2. after intravenous bolus injection, the plasma concentrations declined by 99 % over the first 4h. gbo-006 showed moderate clearance (equivalent to 55 % of rat hepatic blood flow) with half-life of 0.6 h and extra vascular distribution. it showed poor oral bioavailability. pharmacokinetics in mice following iv and po administration in mouse, gbo-006 showed high clearance (77 ml/min/kg, equivalent to 85 % of mouse hepatic blood flow), short terminal half-life (0.5 h) and poor oral bioavailability (table 3). similar to rat, after intravenous bolus injection the plasma concentrations declined by 99 % over the first 4 h (figure 4). dose dependent pharmacokinetic parameters following ip administration at three different dose levels are listed in table 4. as the dose increased from 1 to 3, 7.5 and 10 fold, the area under the curve of plasma concentration versus time profile increased 1 to, 5, 8 and 10 fold. as demonstrated in figure 5, gbo-006 plasma concentrations escalates as the dose increased, thus displaying linear pharmacokinetics. srinivas maddi et al. admet & dmpk 4(4) (2016) 314-326 320 figure 2. representative mrm chromatograms of gbo-006 obtained from blank plasma sample (a) and calibration standard plasma sample of 1 ng/ml, lloq (b). figure 3. mean plasma concentration vs. time curve of gbo-006 following its intravenous and oral administration in male sprague dawley rat. each point (n=3) represents the mean ± sd. sample name: "bl" sample id: "" file: "001.wiff" peak name: "gbo-0061" mass(es): "468.200/290.200 da" comment: "" annotation: "" sample index: 3 sample type: unknown concentration: n/a calculated conc: 0.00 ng/ml acq. date: 12/2/2016 acq. time: 1:14:50 pm modified: no 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6 2.8 3.0 time, min 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 in te ns ity , c ps 1.57 2.54 2.68 2.792.31 1.68 1.991.90 2.95 2.43 2.21 2.10 1.48 0.44 1.04 0.50 0.71 sample name: "bl" sample id: "" file: "001.wiff" peak name: "telmisartan(is)" mass(es): "515.400/276.100 da" comment: "" annotation: "" sample index: 3 sample type: unknown concentration: 1.00 ng/ml calculated conc: n/a acq. date: 12/2/2016 acq. time: 1:14:50 pm modified: no proc. algorithm: analyst classic bunching factor: 1 noise threshold: 20.00 cps area threshold: 100.00 cps ,num. smooths: 4 sep. width: 0.20 sep. height: 1.00 exp. peak ratio: 5.00 exp. adj. ratio: 4.00 exp. val. ratio: 3.00 rt window: 30.0 sec expected rt: 1.78 min use relative rt: no int. type: base to base retention time: 1.82 min area: 413742 counts height: 1.32e+005 cps start time: 1.73 min end time: 2.01 min 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6 2.8 3.0 time, min 0.0 5000.0 1.0e4 1.5e4 2.0e4 2.5e4 3.0e4 3.5e4 4.0e4 4.5e4 5.0e4 5.5e4 6.0e4 6.5e4 7.0e4 7.5e4 8.0e4 8.5e4 9.0e4 9.5e4 1.0e5 1.1e5 1.1e5 1.2e5 1.2e5 1.3e5 1.3e5 in te ns ity , c ps 1.82 sample name: "gbo-006-std1" sample id: "" file: "001.wiff" peak name: "gbo-0061" mass(es): "468.200/290.200 da" comment: "" annotation: "" sample index: 5 sample type: unknown concentration: n/a calculated conc: 0.00 ng/ml acq. date: 12/2/2016 acq. time: 1:23:22 pm modified: no proc. algorithm: analyst classic bunching factor: 1 noise threshold: 20.00 cps area threshold: 100.00 cps ,num. smooths: 10 sep. width: 0.20 sep. height: 1.00 exp. peak ratio: 5.00 exp. adj. ratio: 4.00 exp. val. ratio: 3.00 rt window: 30.0 sec expected rt: 1.57 min use relative rt: no int. type: base to base retention time: 1.58 min area: 2771 counts height: 7.91e+002 cps start time: 1.51 min end time: 1.66 min 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6 2.8 3.0 time, min 0 20 40 60 80 100 120 140 160 180 200 220 240 260 280 300 320 340 360 380 400 420 440 460 480 500 520 540 560 580 600 620 640 660 680 700 720 740 760 780 800 in te n si ty , c p s 1.58 sample name: "gbo-006-std1" sample id: "" file: "001.wiff" peak name: "telmisartan(is)" mass(es): "515.400/276.100 da" comment: "" annotation: "" sample index: 5 sample type: unknown concentration: 1.00 ng/ml calculated conc: n/a acq. date: 12/2/2016 acq. time: 1:23:22 pm modified: no proc. algorithm: analyst classic bunching factor: 1 noise threshold: 20.00 cps area threshold: 100.00 cps ,num. smooths: 4 sep. width: 0.20 sep. height: 1.00 exp. peak ratio: 5.00 exp. adj. ratio: 4.00 exp. val. ratio: 3.00 rt window: 30.0 sec expected rt: 1.78 min use relative rt: no int. type: base to base retention time: 1.80 min area: 639887 counts height: 2.05e+005 cps start time: 1.69 min end time: 2.07 min 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6 2.8 3.0 time, min 0.0 5000.0 1.0e4 1.5e4 2.0e4 2.5e4 3.0e4 3.5e4 4.0e4 4.5e4 5.0e4 5.5e4 6.0e4 6.5e4 7.0e4 7.5e4 8.0e4 8.5e4 9.0e4 9.5e4 1.0e5 1.1e5 1.1e5 1.2e5 1.2e5 1.3e5 1.3e5 1.4e5 1.4e5 1.5e5 1.5e5 1.6e5 1.6e5 1.7e5 1.7e5 1.8e5 1.8e5 1.9e5 1.9e5 2.0e5 2.0e5 in te n si ty , c p s 1.80 1 10 100 1000 10000 0 2 4 6 8 10 12 p la sm a c o n ce n tr a ti o n ( n g /m l) time (h) iv po (a) (b) admet & dmpk 4(4) (2016) 314-326 pharmacokinetics and efficacy of plk2 inhibitor doi: 10.5599/admet.4.4.341 321 table 2. pharmacokinetic parameters of gbo-006 following its intravenous and oral route administration in male sprague dawley rat. data are expressed as mean ± sd in three rats. iv po dose (mg/kg) 1 5 cmax (ng/ml) 22 tmax(h) 1 auc0-24h (ng h /ml) 535 87 dnauc 535 17 cl (ml/min/kg) 30 thalf (h) 0.6 vdss (l/kg) 1.5 mrt last (h) 0.5 bioavailability (%f) 3 figure 4. mean plasma concentration vs. time following intravenous and oral administration of gbo-006 in male swiss albino mice. table 3. pharmacokinetic parameters of gbo-006 following its intravenous and oral route of administration in mouse. iv po dose (mg/kg) 2 5 cl (ml/min/kg) 77 thalf (h) 0.5 vdss (l/kg) 1.7 mrt last (h) 0.4 cmax, po (ng/ml) 64 tmax, po (h) 0.25 auc0-24h (ng h /ml), iv/po 431 25 bioavailability (%f) 2 1 10 100 1.000 10.000 0 2 4 6 8 10 12 p la sm a c o cn e tr a ti o n (n g /m l) time (h) iv po srinivas maddi et al. admet & dmpk 4(4) (2016) 314-326 322 table 4. pharmacokinetic parameters of gbo-006 following intraperitoneal administration in swiss albino mouse at 10, 30, 75, 100 mg/kg dose (mg/kg) 10 30 75 100 formulation strength (mg/ml) 1 3 7.5 10 cmax (ng/ml) 952 1413 1354 1652 tmax (h) 0.25 4 0.25 8 auc0-24h, (ng/ml h) 2451 13530 20177 25592 auc0-∞(ng/ml h) 2531 13807 27067 33744 dose normalized auc0-∞ (ng /ml h) 253 460 360 337 figure 5. mean plasma concentration vs. time following intraperitoneal administration of gbo-006 at different doses in swiss albino mice. pharmacokinetics in dogs a plasma concentration and time profile of gbo-006 obtained in dog is shown in figure 6 and corresponding pharmacokinetic parameters are shown in table 5. similar to results in mice, gbo-006 showed high clearance (47 ml/min/kg, more than dog hepatic blood flow) and short plasma elimination half-life (0.7 h) in dogs indicating that the compound is rapidly eliminated from the body. the volume of distribution of 2.4 l/kg provides significant distribution into extra vascular space. poor bioavailability (6 %) after po dosing showed that the compound did not show any absorption through gastrointestinal tract. 10 100 1000 10000 0 4 8 12 16 20 24 p la sm a c o n ce n tr a ti o n (n g /m l) time (h) 100 mg/kg 10 mg/kg 30 mg/kg 75mg/kg admet & dmpk 4(4) (2016) 314-326 pharmacokinetics and efficacy of plk2 inhibitor doi: 10.5599/admet.4.4.341 323 figure 6. mean plasma concentration vs. time following intravenous and oral administration of gbo-006 in male beagle dog. each point (n=3) represents the mean±sd. table 5. pharmacokinetic parameters of gbo-006 following its intravenous and oral route administration in beagle dog. data are expressed as mean ± sd in three dogs. iv po dose (mg/kg) 1 5 cl (ml/min/kg) 47 thalf (h) 0.7 vdss (l/kg) 2.4 mrt last (h) 0.8 cmax, po (ng/ml) 70 tmax, po (h) 0.8 auc0-24h (ng h /ml), 440 160 bioavailability (%f) 6 xenograft model gbo-006 reduced tumor volume by 52 and 58 % at doses of 30 and 75 mg/kg respectively (figure 7). the tumor growth inhibition at 30 mg/kg and 75 mg/kg dose was almost similar. the lack of dose dependent efficacy was not known at this movement. no statistically significant weight loss was found upon completion of the experiment in the treated groups compared to the vehicle group. there were no abnormalities found in liver, heart, lungs and kidneys suggesting that the compound was well tolerated up to 75 mg/kg. 1 10 100 1000 0 4 8 12 p ls a m a c o n ce n tr a ti o n ( n g /m l) time (h) iv po srinivas maddi et al. admet & dmpk 4(4) (2016) 314-326 324 figure 7. scid mice bearing established tumors (100 mm 3 ) from mda-mb-231 cells were treated with gbo006 at 30 mg/kg (triangle), 75 mg/kg (square), or vehicle (circle) for approximately 24 days. serial tumor volumes (mm 3 ) were measured at the indicated days. each data point represents the mean tumor volume of 8 tumors. discussion gbo-006 is a small plk2 selective kinase inhibitor which showed cytotoxicity against the entire panel of 16 tumor cell lines, with little or no cytotoxicity towards normal cells. kinase inhibition assays against a panel of 285 kinases revealed that this compound had a remarkable specificity towards plk2/snk, a kinase involved in centrosome duplication and mitotic progression [11-13]. most importantly, plk2 has recently been implicated as one of the kinases that links cellular metabolism to cell cycle. mitochondrial dysfunction with increased dependence on glycolysis is frequently observed in cancer cells (known as the warburg effect) and identification of pathways that promote cell survival under conditions of mitochondrial dysfunction has therapeutic implications [14]. intraperitoneal alternate day dosing of gbo-006 using 100 % dmso as formulation caused significant tumor regression in xenograft models, demonstrating proof of concept of plk2 inhibition in vivo. these studies necessitated the need to develop a formulation for pharmacokinetic and efficacy studies. the present communication describes the development of preclinical formulation to characterize the pharmacokinetic properties and efficacy of gbo-006 in xenograft models to determine the degree of inhibition of plk2 that is required for inhibition of tumor growth and assess how concentrations are related to antitumor activity of the compound and to evaluate the validity of using pbmn as surrogates. gbo-006 possesses challenging physicochemical and biopharmaceutical properties like poor solubility in aqueous media, low permeability and a crystalline nature. the poor solubility of gbo-006 in non-aqueous media including lipids & oils, limited any further usage of lipid-based technologies. gbo-006 was found to melt with degradation at very high temperatures (~346 °c), which limited the usage of any amorphous based strategies. in order to understand the solubility, initial screening was conducted in different concentrations and combinations of cosolvents and surfactants. among the various combinations and various solvent compositions tried in our laboratory, a cosolvent system was chosen consisting of dmso, peg 400, solutol and 100 mm citrate buffer at 20:40:10:30 (v/v/v/v) ratios. this system showed 10 mg/ml drug solubility without drug precipitation up to storage of 24 h. thus, this formulation was selected for in 0 50 100 150 200 250 300 350 400 450 0 2 4 6 8 10 12 14 16 18 20 22 24 26 a v e ra g e t u m o r v o lu m e ( m m 3 ) days vehicle 30 mg/kg 75 mg/kg admet & dmpk 4(4) (2016) 314-326 pharmacokinetics and efficacy of plk2 inhibitor doi: 10.5599/admet.4.4.341 325 vivo pharmacokinetic and efficacy evaluations. dmso and peg 400 are the major pharmaceutical cosolvents with strong solubilizing properties and widely used in oral and parenteral formulations. dmso is a water miscible polar aprotic solvent with high interfacial activity as a drug solubilizer and a penetration enhancer [15]. it also belongs to the class of dipolar aprotic solvents which includes dimethylformamide and dimethylacetamide. the cosolvent approach is based on the fact that low aqueous solubility of drugs is due to great difference in polarity of the two components. thus when a cosolvent having a polarity value less than that of water is added, the polarity difference between the drug and the water-co solvent system reduces thereby increase the solubility of drug [16]. gbo-006 pharmacokinetic studies provided an overview of the absorption and disposition behavior in animal models. gbo-006 showed very poor oral bioavailability (2-6 %) across mice, rats and dogs (figures 3, 4 and 6). potential causes for the low oral bioavailability could be due to metabolism in the gut wall, and first-pass metabolism in the liver [17-19]. following intravenous dosing to mice, rats and dogs, gbo-006 is rapidly eliminated from plasma within the first several hours with half-life estimates ranging from 0.5 to 0.7 h over this period. overall, gbo-006 pharmacokinetics was comparable across the three species tested. clearance, the critical connection between the administered dose and drug exposure (auc), was very high (which is more than hepatic blood volume) and consistent among mouse and dog. single day dose escalation study by intraperitoneal route demonstrated the dose proportional increase in auc and cmax, thus displaying linear pharmacokinetics. since linear exposure was observed in mice, 30 and 75 mg/kg were selected for xenograft study, where 30 mg/kg dose showed greater than 50 % tumor regression. further studies are required to compare the efficacy dosage in mice to adverse dosage effects in rat, for non-clinical risk assessment. current efforts are focusing on formulation studies that would enable us to carry out glp tox at higher doses of gbo-006. acknowledgements we gratefully acknowledge the thoughtful suggestions and constructive criticism from e. p. reddy was supported by ca158209. conflict of interest the authors declare that they have no conflict of interest. references [1] n. takai, r. hamanaka, j. yoshimatsu, i. miyakawa, oncogene 24 (2005) 287–291. [2] p. chopra, g. sethi, s.g. dastidar, a. ray, expert opin. investig. drugs 19 (2010) 27-43. [3] e.m. matthew, l.s. hart, a. astrinidis, a. navaraj, n.g. dolloff, d.t. dicker, e.p. henske, w.s. eldeiry, cell cycle 8 (2009) 4168–4175. [4] t.f. burns, p. fei, k.a. scata, d.t. dicker, w.s. el-deiry, mol. cell biol. 23 (2003) 5556–5571. [5] s. warnke, s. kemmler, r.s. hames, h.l. tsai, u. hoffmann-rohrer, a.m. fry, curr. biol. 14 (2004) 1200–1207. [6] t. lu, c.a. laughton, s. wang, t.d. bradshaw, mol. pharmacol. 87 (2015) 18-30. [7] j.p. zhang, j. huang, c. liu, x.f. lu, b.x. wu, l. zhao, n. lu, q.l. guo, z.y. li, c. jiang, chinese chemical letters 25 (2014) 1025–1028. [8] s. cockerill, c. stubberfield, j. stables, m. carter, s. guntrip, k. smith, s. mckeown, r. shaw, p. topley, l. thomsen, k. affleck, a. jowett, d. hayes, m. willson, p. woollard, d. spalding, bioorganic & medicinal chemistry letters 11 (2001) 1401–1405. srinivas maddi et al. admet & dmpk 4(4) (2016) 314-326 326 [9] m.v. reddy, b. akula, s. jatiani, r. vasquez-del carpio, v.k. billa, m.r. mallireddigari, s.c. cosenza, d.r. venkata subbaiah, e.v. bharathi, v.r. pallela, p. ramkumar, r. jain, a.k. aggarwal, e.p. reddy, bioorg. med. chem. 24 (2016) 521-544. [10] m.k. you, m.s. kim, k.s. jeong, e. kim, y.j. kim, h.a. kim, nutrition research and practice 10 (2016) 139-147. [11] v. archambault, d.m. glover, nat. rev. mol. cell biol. 10 (2009) 265-275. [12] k. strebhardt, nat. rev. drug discov. 9 (2010) 643-6600. [13] o. cizmecioglu, a. krause, r. bahtz, l. ehret, n. malek, i. hoffmann. j. cell sci. 125 (2012) 981-992. [14] t. matsumoto, p.y. wang, w. ma, h.j. sung, s. matoba, p.m. hwang, proc. natl. acad. sci. u s a. 106 (2009) 14542-14546. [15] p.k. gupta, the science and practice of pharmacy (21st edition), remington, philadelphia, usa, 2005. [16] m. sun, l. si, x. zhai, z. fan, y. ma, et al., drug dev. ind. pharm. 37 (2011) 986-994. [17] b.s. shin, s.d.yoo, t.h. kim, drug metab. dispos. 42 (2014) 974–982. [18] j.m. smoliga, o. blanchard, molecules 19 (2014) 17154-17172. [19] d.i. draganov, d.a. markham, d. beyer, et.al., toxicology 333 (2015) 168–178. abbreviations definitions dmso dimethyl sulfoxide dma n, n-dimethyl acetamide lloq lower limit of quantitation nmp n-methyl pyrrolidine pbmn peripheral blood mononuclear cells peg polyethylene glycol tpgs d alpha tocopheryl polyethylene glycol 1000 succinate ©2016 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ association of vancomycin with lipid vesicles doi: 10.5599/admet.5.3.400 183 admet & dmpk 5(3) (2017) 183-191; doi: http://dx.doi.org/10.5599/admet.5.3.400 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper association of vancomycin with lipid vesicles xiaohui hu and kin yip tam * drug development core, faculty of health sciences, university of macau, macau s.a.r., china *corresponding author: e-mail: kintam@umac.mo; tel.: +853-88224988; fax: +835-88222314 received: june 15, 2017; revised: july 22, 2017; published: september 29, 2017 abstract antibiotics play a pivotal role in modern medicine for the treatment of bacterial infection in patients. membrane defines the boundary between single cell and its environment and is a main target for antibacterial agents. to better understand the mechanism of antibiotics action on microbes, we utilized liposome as membrane mimic model to study antibiotics interaction with bacterial membrane by variety of biophysical methods. isothermal calorimetry and fluorescence photometry experiments were performed to examine interaction between antibiotics and liposome. we found that vancomycin, one of the most important antibiotics for the treatment of serious infections by gram-positive bacteria, binds to the liposome. the association between the drug and the liposome does not involve the tail part of the lipids. moreover, the binding affinity increases along with the increment of liposome size. of three major lipid components, phosphatidylglycerol is the preferential target for vancomycin binding. we also showed that vancomycin associates with vesicle derived from staphylococcus aureus membrane in a similar manner as the binding to liposome. our data suggested that vancomycin associates with bacterial membrane through direct interaction with lipid head groups with the extent of the association depending very much on specific type of lipids and curvature of local membrane structure. keywords liposome; membrane; binding; phospholipids; antibiotics introduction pathogenic bacteria pose a major threaten to public health, especially in under-developed region of the world. large scale production and use of antibiotic penicillin during the world war ii is a hallmark of modern antimicrobial therapy. since then, a plethora of antimicrobial agents have been developed and commercialized, playing significant roles in improvement of human health, highlighted by gradual increasing of average life span worldwide. prevailing of antibiotic treatment improves human health on one hand, but also triggers resistance of microbes to the drugs [1,2]. depending on their structures and target sites within microorganism, different antibacterial drugs have distinct modes of action [3]. of them, related to cell membrane are: inhibition or block of cell wall synthesis, inhibition of membrane protein or lipid synthesis and interference of cell membrane function. some antibacterial peptides [4,5], peptidomimics [6] and small proteins [7] could interact with or insert into cell membrane, where they interfere membrane formation or alter local membrane structure, thus reducing bacteria growth or reproduction. bacterial membrane is surrounded by a layer of peptidoglycan cell wall. this unique feature had rendered great interest in searching for glycopeptide and http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:kintam@umac.mo hu and tam admet & dmpk 5(3) (2017) 183-191 184 lipoglycopeptide [8], which mimics bacterial cell wall synthesis substrates or binds to synthesis intermediates, thus inhibits cell wall synthesis, while having little effect on host cells. glycopeptide antibiotic vancomycin has been used to treat a number of infections caused by broad spectrum of gram-positive bacteria [9,10]. it inhibits cell wall formation by binding to d-ala-d-ala terminus of peptidoglycan precursors and preventing them from cross-linking [11,12]. to cope with rising of antibiotic resistance including vancomycin, development of novel anti-bacterial agents with multiple action modes had been a promising strategy [13–15]. herein, we studied the binding property of vancomycin to artificial and bacterial membranes using different biophysics methods with the aim to obtain a better understanding of the binding process for expanding its antibiotic repository. experimental materials 1-palmitoyl-2-oleoyl-sn-glycero-3-phospho-(1'-rac-glycerol) (sodium salt) (popg), 1-hexadecanoyl-2-(9zoctadecenoyl)-sn-glycero-3-phosphoethanolamine (pope), 1-oleoyl-2-[12-[(7-nitro-2-1,3-benzoxadiazol-4yl)amino]dodecanoyl]-sn-glycero-3-phosphocholine (nbd pc) and cardiolipin (cl) were purchased from avanti polar lipids. vancomycin and nafcillin were obtained from sigma. lysozyme and rnasea were purchased from merck. staphylococcus aureus was from atcc (strain #43300) and cultured in tryptic soy broth medium (becton, dickinson and company). other chemicals were obtained from sigma. liposome preparation lipids of known mass in chloroform were well-mixed, then dried at room temperature in a glass tube by continuous purging of high pure nitrogen gas. being thoroughly dried in a speed vacuum, the lipids were redissolved in phosphate buffer saline (pbs) buffer (10 mm na2hpo4, 1.8 mm kh2po4, 137 mm nacl, 2.7 mm kcl, ph 7.4) pre-warmed over 45 °c. after repeated pipetting, the hydrated lipids were applied to a lipex extruder (northern lipids inc.) equipped with polycarbonate membrane for more than 10 times. scattering profile of the liposome was then measured on a bi-200sm goniometer (brookhaven instruments) while being excited by a laser passing 637 nm filter with 100 micron pinhole from 90 degree angle at 25 °c. data was collected and processed by the dynamic light scatter software coming with the equipment. isolation of membrane from staphylococcus aureus isolation of membrane was carried out as described in the literature [16]. briefly, bacteria were cultured till late exponential phase before harvest. cells was then re-suspended in 50 mm potassium phosphate buffer, ph 8.0 and digested by lysozyme, deoxyribonuclease-1 and ribonuclease for 30 minutes in the presence of 10 mm mgso4 at 37 °c. the reaction was stopped abruptly by 15 mm edta for 1 minute. additional mgso4 was then added to the system to bring its final concentration back to 10 mm. digested cells was centrifuged at room temperature, 25000 g for 45 minutes to pellet the membrane, which was then re-suspended in pbs and applied through 18-g needle for more than 20 times. resultant membrane vesicle was quantified by measuring protein concentration (pierce™ bca protein assay, thermo scientific), assuming equivalent amount of protein and lipid. size distribution of the vesicles was assessed on a bi200sm goniometer as describe in previous section. isothermal calorimetry measurement in a typical isothermal calorimetry (itc) experiment carried out on a microcal itc200 calorimeter (malvern instruments), 50 mm antibiotics were titrated into liposome with 5 mm lipid (total or individual) in pbs buffer at 25 °c. after each injection, system was allowed to regain thermal balance between sample admet & dmpk 5(3) (2017) 183-191 association of vancomycin with lipid vesicles doi: 10.5599/admet.5.3.400 185 and reference chambers for 180 seconds with continuous stir of 750 rpm. acquired data were processed by itc-costumed origin7 software using one-site model to avoid complication. reference subtractions were performed against the data obtained by titrating antibiotic into pbs buffer. fluorescence spectroscopy spectra of the fluorescent lipid incorporated liposome were collected on an ls55 fluorescence spectrometer (perkinelmer) at room temperature. fluorescent spectra of freshly prepares 2-4 mm liposome containing 0.5 % ndb pc in pbs buffer was recorded in a cuvette. the sample was excited by xenon lamp centred at 485 nm with 2.5 nm slit width. a fraction of antibiotics with high concentration in pbs buffer was then introduced to minimize volume change. after mixed well, the spectrum of the sample was measured with the same setting. results vancomycin binds to liposome with a millimolar affinity. we first utilized liposome as membrane mimic to explore if vancomycin associates with the cell membrane at all. we performed isothermal calorimetry experiment by titrating antibiotics into liposome extruded through polycarbonate membrane (0.1 µm diameter). we found that vancomycin, but not βlactam antibiotic nafcillin, introduced significant amount of heat change in liposome prepared from a mixture of pope, popg and cardiolipin (cl) (75:20:5) (figure 1), a model mimic for bacterial membrane [17]. fitting of the data into one-site model gave rise to millimolar level affinity (dissociate constant, kd), suggesting that vancomycin weakly, but definitely interact with liposome. the observation is in line with previous studies, where weak binding of vancomycin to model membrane were assessed by itc [18] and surface plasmon resonance [19]. figure 1. titration of vancomycin into liposome. raw data and fitting curve of itc experiment (left); comparative fitting plot of naficillin and vancomycin titration with liposome (right). hu and tam admet & dmpk 5(3) (2017) 183-191 186 vancomycin binds to vesicle derived from staphylococcus aureus membrane. now that we showed vancomycin association with membrane mimic liposome, it is tempting to ask whether the antibiotic can also bind to vesicle derived from biological sample. we thus isolated membrane from gram-positive bacterium staphylococcus aureus by means of enzymatic digestion, mechanical disruption and syringe needle homogenization [16]. resultant vesicle was then undergone calorimetric titration with vancomycin as the titrant. according to the fitting result of itc data, vancomycin also bound to the bacterial derived vesicle (figure 2), but with much higher affinity (sub-millimolar) than that of liposome (figure 2). although we cannot completely rule out the possibility of residual peptidoglycan remaining on the membrane, lysozyme digestion step of the cells should remove most, if not all of them. apparently, vancomycin binds to the bacterial membrane derived vesicle with higher affinity. figure 2. titration of vancomycin into vesicle derived from staphylococcus aureus membrane. raw data and fitting curve of itc experiment (left); comparative fitting plot of vancomycin titration with liposome and vesicle from membrane (top right); size distribution of liposome and membrane vesicle determined by dynamic light scattering (bottom right). vancomycin strengthens the association with liposome as its size increases. in light of differential binding between liposome and membrane-derived vesicle to vancomycin and the fact that the later measured much larger in diameter than the former (figure 2), we speculated that the size of liposome/vesicle might play a role in the antibiotic association with membrane mimic. to test the hypothesis, we performed serial calorimetric titration of vancomycin into liposome prepared with polycarbonate membranes of various sizes (0.1, 0.4 and 1 µm). we found that the binding affinity of vancomycin increased along with liposome size increment. (figure 3), which indicated that cell surface landscape or/and membrane curvature could be an important factor in vancomycin association. admet & dmpk 5(3) (2017) 183-191 association of vancomycin with lipid vesicles doi: 10.5599/admet.5.3.400 187 figure 3. size-dependent binding of liposome with vancomycin. data and fitting curve from titrations of liposome were plotted in same figure for comparison (left); ka value derived from fitting were plotted against the sizes of the liposome (right). vancomycin binds preferentially to popg over cardiolipin and pope. because the lipids used in the study are similar structurally in fatty acid moiety but vary in head groups, we reasoned that their difference in head group could contribute differently to the association with vancomycin. to explore binding preference of vancomycin among lipids, we performed itc measurement of the antibiotic with liposomes prepared from individual lipid. based on the ka value derived from data fitting, binding affinity of vancomycin to popg is significantly higher than that to pope and cl (figure 4), suggesting that popg is the major contributor to the association of vancomycin with liposome. the observation could be explained by careful examination of the structure of vancomycin [20,21] and the lipids (figure 4). positive charged amine in the head group of pope does not appear to favour the interaction with uncharged vancomycin. moreover, embedded hydroxyl group on hydrocarbon chain within two phosphoryl group is unlikely to facilitate binding in the case of cardiolipin. on the other hand, the two hydroxyls on the head group of the popg could potentially interact with vancomycin via hydrogen bond formation, which promotes the glycopeptide association with lipid membrane [22]. vancomycin interaction with liposome does not involve acyl terminus of the lipids. we have shown that vancomycin associates with liposome through its interaction with the head group of the lipids. but it remains unclear if the antibiotic could introduce any change inside the lipid bilayer. fluorescence spectroscopy was performed to gain more insight into the binding of vancomycin to liposome. we incorporated trace amount (0.5 %) of tail fluorescence-labeled lipid (ndb-pc) into liposome. addition of stoichiometric amount of vancomycin did not induce significant change in the spectrum of the liposome (figure 5), suggesting that binding of vancomycin to liposome might not involve the acyl terminus of the lipid. in other words, it is unlikely for the antibiotic to insert deep into the lipids bilayer. otherwise, alteration in fluorescence intensity or/and peak position shift would be obvious because of fluorophore disturbing by the insertion, as seen in a control experiment, where pore formation by depsipeptide antibiotics, valinomycin [23,24] reduced liposome fluorescence intensity substantially. this is consistent with the fact that vancomycin and other glycopeptides generally exhibit low membrane permeability [3,22]. hu and tam admet & dmpk 5(3) (2017) 183-191 188 figure 4. vancomycin binds preferentially to popg over pope and cardiolipin. itc raw data and fitting of vancomycin titration with popg (top left), pope (top right) and cardiolipin (bottom left); fitting of titrations were plotted for comparison (middle left); structure of three lipids were shown (bottom right). discussion we have shown that glycopeptide antibiotics vancomycin associate with biomimetic membrane, a fact have been largely ignored in the field because of its weak interaction [18,19]. furthermore, we showed for the first time that it binds to membrane vesicle derived from gram-positive bacteria staphylococcus aureus admet & dmpk 5(3) (2017) 183-191 association of vancomycin with lipid vesicles doi: 10.5599/admet.5.3.400 189 in a similar fashion. these observations could potentially open a new avenue for the prototype antibiotics to tackle infectious microbes by blocking cell wall synthesis and interfering cell membrane function simultaneously. multiple mechanism of action is thought to be an effective strategy to combat with prevalent resistance nowadays rising among pathological microorganism [14,15,25]. figure 5. fluorescence spectra of liposome with fluorescent label. left panel, spectra of 3.5 mm liposome containing 0.5 % ndb pc in the absence (green line) and presence (red line) of 5 mm vancomycin; right panel, spectra of 1.5 mm liposome containing 0.5 % ndb pc in the absence (green line) and presence (red line) of 2 mm valinomycin. we also found that vancomycin appears to bind to the membrane from outside and the extent of association depends much on local phospholipid composition and the curvature of the membrane. popg appears to be the favourite target for vancomycin binding. moreover larger size (flattened surface) of the liposome enhances the binding. these findings could provide some design ideas for further modification of the antibiotics to optimize its binding to bacterial membrane, which may improve antibacterial potency. to this end, a further investigation of vancomycin structure and bacterial membrane composition is required. in addition, self [20,21] and ligand-induced dimerization [26,27] of vancomycin might also play an important role as a larger hydrophobic interface once shielded by dimeric interaction has to be covered by membrane association. conclusions we have studied the binding of vancomycin to biomimetic membranes based on three phospholipid components and bacterial membranes by using a variety of biophysical methods, including isothermal calorimetry and fluorescence photometry. it has been shown that the ways of association of vancomycin to biomimetic membranes is very similar to that of vesicle derived from staphylococcus aureus membrane. our results clearly indicated that vancomycin associates to the cell membrane from outside, and does not permeate through the cell membrane, which is consistent with the mechanisms of the drug action. by varying the size and composition of the biomimetic membrane vesicles, it has been demonstrated that the extent of association depends on specific type of lipids and curvature of local membrane structure. these observations could potentially open a new avenue to revive the prototype antibiotic to fight infectious microbe through the design of specific tight binders to the outer surface of bacterial membranes to impede cell wall synthesis. hu and tam admet & dmpk 5(3) (2017) 183-191 190 acknowledgements: we thank the financial support from the university of macau (grant no.: srg201300055-fhs). we thank prof jun zheng (university of macau) for providing a bacteria strain of staphylococcus aureus. references [1] l.l. silver. challenges of antibacterial discovery. clin. microbiol. rev., 24 (2011) 71–109. [2] m.a. fischbach, c.t. walsh. antibiotics for emerging pathogens. science, 325 (2009) 1089–93. [3] m.a. kohanski, d.j. dwyer, j.j. collins. how antibiotics kill bacteria: from targets to networks. nat. rev. microbiol., 8 (2010) 423–35. [4] m. zasloff. antimicrobial peptides of multicellular organisms. nature, 415 (2002) 389–395. [5] k.a. brogden. antimicrobial peptides: pore formers or metabolic inhibitors in bacteria?. nat. rev. microbiol., 3 (2005) 238–250. [6] t. velkov, p.e. thompson, r.l. nation, j. li. structure-activity relationships of polymyxin antibiotics. j. med. chem., 53 (2010) 1898–1916. [7] s. mukherjee, h. zheng, m.g. derebe, k.m. callenberg, c.l. partch, d. rollins, d.c. propheter, j. rizo, m. grabe, q.-x. jiang, l. v hooper. antibacterial membrane attack by a pore-forming intestinal c-type lectin. nature, 505 (2014) 103–7. [8] d. kahne, c. leimkuhler, w. lu, c. walsh. glycopeptide and lipoglycopeptide antibiotics. chem. rev., 105 (2005) 425–448. [9] d.l. stevens. the role of vancomycin in the treatment paradigm. clin. infect. dis., 42 suppl 1 (2006) s51–7. [10] e. breukink, b. de kruijff. lipid ii as a target for antibiotics. nat. rev. drug discov., 5 (2006) 321– 332. [11] u.s. eggert, n. ruiz, b. v falcone, a.a. branstrom, r.c. goldman, t.j. silhavy, d. kahne. genetic basis for activity differences between vancomycin and glycolipid derivatives of vancomycin. science, 294 (2001) 361–364. [12] g.d. wright, c.t. walsh. d-alanyl-d-alanine ligases and the molecular mechanism of vancomycin resistance. acc. chem. res., 25 (1992) 468–473. [13] l. otvos, c. snyder, b. condie, p. bulet, j.d. wade. chimeric antimicrobial peptides exhibit multiple modes of action. int. j. pept. res. ther., 11 (2005) 29–42. [14] v. yarlagadda, s. samaddar, k. paramanandham, b.r. shome, j. haldar. membrane disruption and enhanced inhibition of cell-wall biosynthesis: a synergistic approach to tackle vancomycinresistant bacteria. angew. chem. int. ed. engl., 54 (2015) 13644–13649. [15] f. guilhelmelli, n. vilela, p. albuquerque, l. da s. derengowski, i. silva-pereira, c.m. kyaw. antibiotic development challenges: the various mechanisms of action of antimicrobial peptides and of bacterial resistance. front. microbiol., 4 (2013) 353. [16] r.k. poole. the isolation of membranes from bacteria. methods mol. biol., 19 (1993) 109–122. [17] j.k.l. and e.j.p. mary t. le. biomimetic model membrane systems serve as increasingly valuable in vitro tools. in: a. george (ed.), adv. biomimetics, intech, 2011. [18] s. al-kaddah, k. reder-christ, g. klocek, i. wiedemann, m. brunschweiger, g. bendas. analysis of membrane interactions of antibiotic peptides using itc and biosensor measurements. biophys. chem., 152 (2010) 145–152. [19] h. kinouchi, h. arimoto, k. nishiguchi, m. oka, h. maki, h. kitagawa, h. kamimori. binding properties of antimicrobial agents to lipid membranes using surface plasmon resonance. biol. pharm. bull., 37 (2014) 1383–1389. [20] z. jia, m.l. o’mara, j. zuegg, m.a. cooper, a.e. mark. vancomycin: ligand recognition, dimerization and super-complex formation. febs j., 280 (2013) 1294–1307. admet & dmpk 5(3) (2017) 183-191 association of vancomycin with lipid vesicles doi: 10.5599/admet.5.3.400 191 [21] p.j. loll, a.e. bevivino, b.d. korty, p.h. axelsen. simultaneous recognition of a carboxylatecontaining ligand and an intramolecular surrogate ligand in the crystal structure of an asymmetric vancomycin dimer. j. am. chem. soc., 119 (1997) 1516–1522. [22] o. domenech, g. francius, p.m. tulkens, f. van bambeke, y. dufrêne, m.-p. mingeot-leclercq. interactions of oritavancin, a new lipoglycopeptide derived from vancomycin, with phospholipid bilayers: effect on membrane permeability and nanoscale lipid membrane organization. biochim. biophys. acta biomembr., 1788 (2009) 1832–1840. [23] t.e. andreoli. the effect of valinomycin on the ionic permeability of thin lipid membranes. j. gen. physiol., 50 (1967) 2527–2545. [24] p. bhattacharyya, w. epstein, s. silver. valinomycin-induced uptake of potassium in membrane vesicles from escherichia coli. proc. natl. acad. sci., 68 (1971) 1488–1492. [25] s. gardete, a. tomasz. mechanisms of vancomycin resistance in staphylococcus aureus. j. clin. invest., 124 (2014) 2836–2840. [26] d.mcphail, a. cooper, s. j. hammond, j. p. mackay, m. s. westwell, d. a. beauregard, m. f. cristofaro. thermodynamics and kinetics of dissociation of ligand-induced dimers of vancomycin antibiotics. j. chem. soc. faraday trans. 93 (1997) 2283–2289. [27] m. rekharsky, d. hesek, m. lee, s.o. meroueh, y. inoue, s. mobashery. thermodynamics of interactions of vancomycin and synthetic surrogates of bacterial cell wall. j. am. chem. soc. 128 (2006) 7736–7737. ©2017 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ use of low field nmr for the characterization of gels and biological tissues doi: 10.5599/admet.6.1.430 34 admet & dmpk 6(1) (2018) 34-46; doi: http://dx.doi.org/10.5599/admet.6.1.430 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper use of low field nmr for the characterization of gels and biological tissues michela abrami 1 , gianluca chiarappa 1 , rossella farra 1 , gabriele grassi 2 , paolo marizza 3 , mario grassi 1* 1 department of engineering and architecture, trieste university, via valerio 6, trieste, i-34127, italy 2 department of life sciences, trieste university, strada di fiume 447, trieste, i-34149, italy 3 department of microand nanotechnology, technical university of denmark (dtu), ørsteds plads bygning 345ø, kongens lyngby 2800, denmark *corresponding author: e-mail: mario.grassi@dia.units.it; tel.: +39-040-558-3435; fax: +39-040-569823 received: august 30, 2017; revised: march 16, 2018; published: march 25, 2018 abstract the focus of this paper is on the theoretical interpretation of low field nuclear magnetic resonance (lfnmr) data regarding hydrogels architecture and on the most interesting applications of lf-nmr presented by this research group at the 6 th iapc symposium held in zagreb (hr) on september 2017. particular attention is devoted to the determination of the mesh size distribution of gels polymeric network and the determination of the pore size distribution of microporous systems such as scaffolds, bones, and porous gels. in addition, we report on a very recent application of lf-nmr for monitoring lung functioning in patients suffering from chronic pulmonary diseases like cystic fibrosis. the main findings of this work consist in providing a very simple and accurate approximation of a general theory devoted to evaluating the relation existing among four fundamental polymeric network parameters, i.e. the polymer volume fraction inside the hydrogel, mesh size, hydraulic radius, and the radius of the cylinder ideally embedding each polymeric network chain. furthermore, we demonstrated the potentiality of lf-nmr in the characterization of different polymeric systems among which the sputum of patients suffering from chronic pulmonary diseases appears the most innovative application for its simplicity, rapidity, effectiveness, and potential impact in the everyday clinic. keywords low field nmr; mesh size distribution; pores size distribution; biomedical applications introduction despite the widespread use of high-field nuclear magnetic resonance (7 – 37 t) for the study of chemical structures, low-field nuclear magnetic resonance (lf-nmr; 0.37 – 2.4 t) is quite uncommon. nevertheless, the fundamental works of brownstein and tarr [1,2], mitra et al. [3], torrey et al. [4], and chui et al. [5] clearly demonstrated the potentiality of lf-nmr for the study of different materials. accordingly, lf-nmr finds interesting applications in many fields like food chemistry. for instance, hills et al. [6] explored water distribution in skimmed milk and apple, aroulmoji et al. [7] investigated the hydration properties of aqueous sugar solutions, and de’nobili et al. [8] delved pectin/alginates films for the release of ascorbic acid in agar hydrogels mimicking food materials. moreover, lf-nmr proved to be very useful in http://www.pub.iapchem.org/ojs/index.php/admet/index admet & dmpk 6(1) (2018) 34-46 lf-nmr for the characterization of biological tissues doi: 10.5599/admet.6.1.430 35 determining the size distribution of oil droplets in oil-water emulsions [9,10], an important aspect for different kinds of edible creams like mayonnaise. at the same time, lf-nmr resulted to be an extremely important characterization tool in the petroleum industry as it may be applied not only to predicting the viscosity of crude oils and crude oil emulsions [11] but also to profitably evaluating the pores size distribution of the rocks containing petroleum [12-14]. the capability of estimating the pore size distribution rendered lf-nmr a valuable approach also in the biomedical field. indeed, many authors employed lf-nmr to characterize bones [15-18], a very particular porous material, while grassi et al. [19] and fiorentino et al. [20] focussed on the characterization of polymeric scaffolds for tissue engineering. finally, abrami et al. [21] dealt with a very new application of lf-nmr concerning the characterization of expectorate of patients affected by chronic pulmonary diseases like cystic fibrosis (cf). undoubtedly, however, one of the most common lf-nmr applications regards the characterization of hydrogels whose three-dimensional architecture is able to affect water hydrogens magnetic relaxation. for example, ghi et al. [22] studied the architecture of phema hydrogels, barbucci et al. [23] considered hyaluronan-based hydrogels, hietalahti et al. [24] investigated the crosslinking in unsaturated polyester solutions, brand et al. [25] focussed on the gel point determination of gelatin, dobies et al. [26] explored the gelation process of aqueous low methoxyl pectin solutions, while abrami et al. [27] delved the properties of an interpenetrated network made of alginate and pluronic f127. despite the great variety of lf-nmr applications, it is important to remember that, whatever the system considered is, the guiding principle allowing lf-nmr characterization relies on the effect of solid surfaces (polymeric chains, bones, rocks, and others) on the relaxation process of water hydrogens subjected to a sudden variation of an external magnetic field. the higher the ratio between system solid surface and system volume is, the faster is the hydrogens relaxation process. based on this physical evidence, it is possible to obtain interesting information concerning the three-dimensional architecture of gels network (mesh size distribution) and the pores size distribution of porous materials. the focus of this paper is on 1) some considerations on the theoretical interpretation of lf-nmr data regarding hydrogels architecture (see theoretical background section) and 2) the most interesting lf-nmr applications presented by this research group at the 6 th iapc symposium held in zagreb (hr) on september 2017 (see results and discussion section). in particular, applications concern the characterization of polymeric gels, scaffolds, and biological tissues such as bones and sputum of cf patients. theoretical background the principle of lf-nmr relies on the effect of a static magnetic field b0 on the dipole moments of some atoms like hydrogens (protons). indeed, hydrogens dipole moments do not align themselves with the magnetic field direction, but their axes start a precession movement around b0 direction. this precession is characterized by the resonance frequency 0 and the angle () formed between dipole moments and b0 direction. as all dipole moments react in the same manner to the presence of b0, they generate the global magnetization vector m, whose components are mz, conventionally aligned with b0, mx, and my lying in the xy plane perpendicular to b0. since dipole moments are randomly oriented on x and y axes, they cancel themselves and the overall mxy = 0 (no phase coherence). in addition, as some dipole moments are able to start a precession movement around the direction opposite that of b0 (but they are the minority for energetic reasons), mz module depends on the difference between the parallel and antiparallel oriented dipole moments. the application of a radiofrequency field b1, rotating at the resonance frequency and normal to b0, allows the rotation of m around b1 of the angle  = *b1*tp, where  is the gyromagnetic ratio (whose value depends on the particular atom) and tp is the pulsation time of b1. when  = 90°, mxy = m. grassi et al. admet & dmpk 6(1) (2018) 34-46 36 m0 and mz = 0. thus, upon b1 removal just after tp, all dipole moments will return to their original orientation (m parallel to b0) by rotating around b0 with a decreasing angle up to the attainment of . this re-orientation implies mxy reduction and mz increase up to mxy = 0 and mz = m. the parameter used to evaluate the speed of mxy disappearance is a constant called average relaxation time t2m (spin-spin relaxation time). at the same time, the increase of the mz component is characterized by another average relaxation time (t1m) called spin-lattice relaxation time. the ratio mxy/m is called fid (free induction decay). fid (= i(t)) temporal evolution may be described according to the following sum of exponential functions [5]:   2 1 exp() m i i i i t a t t   m m 2m i 2i i i=1 i=1 =t at a  ( 1 𝑇2 ) 𝑚 = ∑ 𝐴𝑖 𝑇2𝑖 𝑛 𝑖=1 ∑ 𝐴𝑖 𝑛 𝑖=1 ⁄ (1) where t is time, ai are pre-exponential factors (dimensionless), proportional to the number of protons relaxing with the relaxation times t2i and t2m, is the average relaxation time of protons. cpmg sequence unfortunately, due to the unavoidable presence of magnetic field inhomogeneities, after the application of the b1 pulse, magnetization m undergoes the fanning-out phenomenon consisting in the clockwise and anticlockwise rotation of the faster and the slower m components. indeed, m is owed to dipole moments rotating with different frequencies due to the above-mentioned field inhomogeneities. ultimately, fanningout makes virtually impossible fid determination. in order to properly overcome the fanning-out effect, the first b1 pulse (provoking a /2 rotation of m) is followed, after a time , by a second b1 pulse causing a further m rotation of . by doing so, the faster and the slower m components invert their rotation direction thus meeting after a time of 2 from the first b1 pulse, when it is possible to measure fid. the sequence 90°--180°-(echo) is called hahn spin-echo sequence. in order to follow the entire relaxation process, it is necessary to apply n times the b1 pulse (each one separated by a time interval ) thus provoking a  rotation of m: 90°[--180°-(echo)]n. this is the so-called cpmg (carr–purcell–meiboom–gill) pulse sequence [28]. surface effect on t2m t2m depends on many factors such as temperature, magnetic field intensity, and the interaction among hydrogens and other possible substances present in the environment [29]. in particular, hydrogens interactions with solid surfaces (for instance dispersed/solubilized polymeric chains and system boundaries) are one of the most important causes of t2m variation. indeed, the relaxation time of water protons near the solid surface (bound water protons) is lower (fast relaxation) than that (slow relaxation) of bulk water protons (free water protons) that are unaffected by solid surface [5]. accordingly, the average relaxation time (t2m) of protons will depend on the ratio between the solid surface area (s), proportional to the number of protons close to s, and system volume (v), proportional to the total number of protons belonging to the system, as demonstrated by brownstein and tarr [1] in the case of solid porous systems. in particular, the relation between s, v, and the average relaxation time may be represented by [29]: ( 1 𝑇2 ) 𝑚 = 1 𝑇2h2o + 𝑆 𝑉 𝜆 𝑇2s , (2) where t2h2o is the bulk protons relaxation time (i.e. the free water protons relaxation time t2h2o ≈ 3700 ms at 37 °c, 20 mhz [30]),  is the thickness of the water layer close to the solid surface (bound water), while t2s is the relaxation time of bound water protons. equation (2) clarifies that, when the ratio s/v rises, admet & dmpk 6(1) (2018) 34-46 lf-nmr for the characterization of biological tissues doi: 10.5599/admet.6.1.430 37 (1/t2)m increases and t2m decreases. indeed, relying on the equation (1), it follows that (1/t2)m ≠ 1/t2m, but it is easy to mathematically verify that, when (1/t2)m increases, t2m decreases. as the evaluation of t2s is very complex, by starting from equation (2), chui and co-workers [5] derived the following relation: ( 1 𝑇2 ) 𝑚 = 1 𝑇2h2o + 2 ℳ 𝑅ℎ , h =2 v r s , (3) where rh is the system hydraulic radius, while m is a physical parameter (relaxivity, length/time) considering the surface effect on proton relaxations. indeed, m represents the ratio between the thickness and the relaxation time of the bound water layer adhering to the solid surface (thus m = /t2s). m depends not only on the solid surface chemistry but also on temperature, hydrogenated fluid type, and magnetic field strength [29]. typically, at 20 mhz, in the temperature range 25-37 °c, polymeric materials are characterized by m values ranging between 10 -7 – 10 -5 m/s [31]. in the case of nano-structured systems like gels, chui [5] expressed the hydraulic radius as a function of three very important network parameters, i.e. the average mesh size a, the polymer volume fraction (), and the radius (rf) of the cylinder ideally containing the single polymeric chain of the gel network: rh = f(a,rf). accordingly, equation (3) becomes: ( 1 𝑇2 ) 𝑚 = 1 𝑇2h2o + 2 ℳ 𝑓(𝜉𝑎,ϕ,𝑅𝑓) . (4) while chui’s expression [5] for f(a,rf) relies on the assumption that the polymeric network is nothing but a very long fiber, scherer [32], by providing a more sophisticated and physically sound schematization of the polymeric network (cubic, octahedral or tetrahedral spatial arrangement), obtained a different expression for f(a,rf). nevertheless, also in the simplest case of a cubic arrangement, f(a,rf) is, mathematically speaking, not particularly user-friendly. as a result, scherer [32] proposed the following empirical approximation holding whatever the network arrangement considered is (cubic, octahedral or tetrahedral): 𝑅ℎ = 𝑅𝑓 1−0.58φ 𝜑 . (5) this expression ensures a very good approximation of the original scherer model (relative error < 10 % whatever the network spatial arrangement is) for  ≤ 0.56, a very high value that is never encountered in real gels. equation (5) represents an optimal starting point to deduce a useful approximate relation among a, , and rf. indeed, the original scherer equation [32] yields the following cubic equation in a: 𝜉𝑎 3 ( 2𝐶0 𝑅ℎ𝑅𝑓 ) − 2𝐶1𝜉𝑎 (1 + 𝑅𝑓 𝑅ℎ ) + 𝐶2𝑅𝑓 (2 𝑅𝑓 𝑅ℎ + 3) = 0 , (6) where c0, c1, and c2 are constants depending on the specific spatial arrangement of the network (see table 1). as, in the light of equation (5), we may assume that rf/rh ≈ 0 (at least for  ≤ 0.1), equation (6) may be re-written as: 𝜉𝑎 3 ( 2𝐶0 𝑅ℎ𝑅𝑓 ) − 2𝐶1𝜉𝑎 + 3𝐶2𝑅𝑓 = 0 , (7) or: 𝑅ℎ (𝜉𝑎 3 ( 2𝐶0 𝑅ℎ 2𝑅𝑓 ) − 2𝐶1𝜉𝑎 𝑅ℎ + 3𝐶2 𝑅𝑓 𝑅ℎ ) = 0 . (8) m. grassi et al. admet & dmpk 6(1) (2018) 34-46 38 table 1. constants of the scherer’s model [32] network structure c0 c1 c2 cube 1 3 8√2 octahedron √2 3 12 𝑎𝑟𝑐𝑠𝑖𝑛 ( √6 3 ) 16√2 tetrahedron √2 12 6 𝑎𝑟𝑐𝑠𝑖𝑛 ( √3 3 ) 8 by assuming, again, that rf/rh ≈ 0, equation (8) becomes: 𝑅ℎ (𝜉𝑎 3 ( 2𝐶0 𝑅ℎ 2𝑅𝑓 ) − 2𝐶1𝜉𝑎 𝑅ℎ ) = 𝑅ℎ 𝜉𝑎 (𝜉𝑎 2 ( 2𝐶0 𝑅ℎ 2𝑅𝑓 ) − 2𝐶1 𝑅ℎ ) = 0 . (9) it is easy to observe that, apart from the trivial solution (a = 0), the only physically sound solution of equation (9) is: 𝜉𝑎 = √ 𝐶1 𝐶0 𝑅ℎ 𝑅𝑓 . (10) by substituting rh for its approximation (eq.(5)), the final approximate expression for a is obtained: 𝜉𝑎 = 𝑅𝑓 √ 𝐶1 𝐶0 1−0.58φ 𝜑 . (11) on the basis of equation (11) and equation (5), we are also able to derive the direct relation between rh and a: 𝑅ℎ = 𝜉𝑎 √ 𝐶0 𝐶1 1−0.58φ 𝜑 . (12) thus, equation (4) may be satisfactorily approximated by: ( 1 𝑇2 ) 𝑚 = 1 𝑇2h2o + 2 ℳ 𝑅𝑓 1−0.58φ 𝜑 or ( 1 𝑇2 ) 𝑚 = 1 𝑇2h2o + 2 ℳ 𝜉𝑎√ 𝐶0 𝐶1 1−0.58φ 𝜑 . (13) in the case of a cubic arrangement of the network (similar considerations also apply in the case of octahedron and tetrahedron), figure 1 shows the very good agreement between equation (11) (solid line) and scherer’s original expression (equation (6), open circles). it is possible to observe (dotted line; right vertical axis) that whatever is, the relative error is always < 5 %. figure 2 demonstrates that a good agreement between scherer’s model (equation (6), open circles) and our approximation (equation (12)) occurs (relative % error < 10 %) up to  < 0.61. in the case of porous systems (such as zeolites, porous rocks, and microporous gels), the evaluation of the s/v ratio is considerably simpler if pores are assumed spherical. indeed, with this hypothesis, s/v = a/3, so that equation (3) becomes: ( 1 𝑇2 ) 𝑚 = 1 𝑇2h2o + 6 ℳ 𝜉𝑎 . (14) since t2h2o is known and (1/t2)m may be experimentally determined by fitting equation (1) to experimental relaxation data, equation (14) may be employed to evaluate m once a has been evaluated by an independent approach. interestingly, a may be determined resorting to the pgse sequence (pulsed gradient spin echo) [33] which is the cpmg sequence where two gradient fields are applied after the two pulses b1(/2) and b1(). the presence of the two gradients of intensity g allows determining the water self-diffusion coefficient dw dependence on the diffusion time [34]: admet & dmpk 6(1) (2018) 34-46 lf-nmr for the characterization of biological tissues doi: 10.5599/admet.6.1.430 39 figure 1. comparison between scherer’s model (equation (6), open circles) and our approximation (equation (11), solid line, cubic arrangement). the dotted line indicates the relative % error (right vertical axis),  is the polymer volume fraction characterizing the network, rf is the radius of the cylinder able to contain each polymeric chain in its linear configuration, while a is the average mesh size. figure 2. comparison between scherer’s model (equation (6), open circles) and our approximation (equation (12), solid line, cubic arrangement). the dotted line indicates the relative % error (right vertical axis),  is the polymer volume fraction characterizing the network, rh is the hydraulic radius, and a is the average mesh size.  2 2 2 0 ln / 3)w i d g i            , (15) where  is the hydrogen gyromagnetic ratio,  is the gradient length,  is the time elapsed from the beginning of the first gradient and the one of the second gradient, while td = -/3 is the diffusion time -20 -18 -16 -14 -12 -10 -8 -6 -4 -2 0 1 10 100 0 0.2 0.4 0.6 0.8 1 r e la ti v e e rr o r %  a /r f  scherer model eq.(11) relative error % -30 -25 -20 -15 -10 -5 0 5 10 0.1 1 10 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 r e la ti v e e rr o r % r h / a  scherer model eq.(12) relative error % m. grassi et al. admet & dmpk 6(1) (2018) 34-46 40 (time allowed for hydrogens diffusion). for small td values, dw is connected to a according to the following equation [35]: 𝐷𝑤(𝑡𝑑) 𝐷𝑤𝑓 ≈ 1 − 4 9√𝜋 √𝐷𝑤𝑓𝑡𝑑 6 𝜉𝑎 , (16) where dwf is the free water self-diffusion coefficient. fitting equation (16) to experimental dw vs td data allows the determination of a. mesh/pore size distribution obviously, a real polymeric network/porous system is composed of meshes/pores of different sizes that may be grouped into different main classes, each one characterized by its proper dimension i. interestingly, different i correspond to different s/v values, i.e. different relaxation times t2i (that may be determined by fitting equation (1) to experimental relaxation data). by assuming that m is independent of i [5], for each mesh/pore class, an equation similar to equations (13)/(14) may be written: 1 𝑇2𝑖 = 1 𝑇2h2o + 2 ℳ 𝜉𝑖√ 𝐶0 𝐶1 1−0.58φ 𝜑 mesh, 1 𝑇2𝑖 = 1 𝑇2h2o + 6 ℳ 𝜉𝑖 pore. (17) as i is the only unknown, equation (17) allows determining i for each mesh/pore class, thus providing a deep characterization of system topology. indeed, equation (1) and equation (17) allow the determination of the discrete mesh/pores size distribution (ai, i). finally, it is important to remind that the theoretical procedure presented above for the determination of mesh/pores size distribution strictly holds only in fast diffusion (or exchange) conditions occurring when water molecules mobility (represented by their self-diffusion coefficient, dw) is considerably higher than the relaxation rate (rr = 1/t2s) of their protons, i.e. when: 𝑎2 𝑅𝑅 𝐷𝑤 = 𝑎2 𝑇2𝑠𝐷𝑤 = 𝑎ℳ 𝐷𝑤 ≪ 1 , (18) where a is the thickness of the bound water layer. for fibrous systems like gels, it is easy to verify that fast diffusion conditions are usually accomplished. indeed, chui [5] provided the following evaluation for a: 𝑎 ≈ 𝑅𝑓 √𝜑 . (19) by remembering that, typically, rf ≤ 1 nm and  ≥ 0.01, a results to be lower than 10 nm (10 -8 m). in addition, as, in the temperature range 20-40 °c, dwf~ 2.5*10 -9 m 2 /s [36] and, in the same temperature range, t2s (20 mhz) ≥ 0.01 ms (10 -5 s), it is possible to conclude that m is approximately 10 -6 m/s and am/dw is always lower than 4*10 -3 , i.e. << 1. unfortunately, for porous systems such as some hydrogels and scaffolds, where pores size may range from 1 up to 1000 m, fast diffusion conditions may not be always met. this implies that the process of protons relaxation of each class of pores may no longer be described by only one relaxation time, but more (theoretically infinite) relaxation times are required. thus, equation (1) should be replaced by [1]: 𝐼(𝑡) = ∑ 𝐴𝑖 ∑ 𝐼𝑖𝑗 𝑒𝑥𝑝(−𝑡 𝑇2𝑖𝑗⁄ ) ∞ 𝑗=1 𝑚 𝑖=1 , ( 1 𝑇2 ) 𝑚 = ∑ 𝐴𝑖 ∑ 𝐼𝑖𝑗 𝑇2𝑖𝑗⁄ ∞ 𝑗=1 𝑚 𝑖=1 , (20) where: 𝑇2𝑖𝑗 = (𝜉𝑖 2⁄ ) 2 𝐷𝑤𝑋𝑖𝑗 and 𝐼𝑖𝑗 = 12(𝑠𝑖𝑛(𝑋𝑖𝑗)−𝑋𝑖𝑗𝑐𝑜𝑠(𝑋𝑖𝑗)) 2 𝑋𝑖𝑗 3 (2𝑠𝑋𝑖𝑗−𝑠𝑖𝑛(2𝑋𝑖𝑗)) (21) admet & dmpk 6(1) (2018) 34-46 lf-nmr for the characterization of biological tissues doi: 10.5599/admet.6.1.430 41 being xij the positive roots of the following equation: 1 − 𝑋𝑖𝑗 𝑐𝑜𝑡(𝑋𝑖𝑗 ) = ℳ𝜉𝑖 2𝐷𝑤 . (22) consequently, equations (14) and (17) become, respectively: ( 1 𝑇2 ) 𝑚 = ∑ 𝐴𝑖 ∑ 𝐼𝑖𝑗 𝑇2𝑖𝑗 =∞𝑗=1 𝑚 𝑖=1 1 𝑇2h2o + 6 ℳ 𝜉𝑎 , (23) ( 1 𝑇2𝑖 ) 𝑚 = ∑ 𝐼𝑖𝑗 𝑇2𝑖𝑗 =∞𝑗=1 1 𝑇2h2o + 6 ℳ 𝜉𝑖 . (24) as t2ij are not independent of each other (on the basis of equation (21), it follows t2ij = t2i1(xi1/xij) 2 ), equation (20) fitting parameters are t2i1, ai, and mrdi (=mi/2dw) for a total of 3m parameters. once these parameters are known by fitting equation (20) to experimental data, equation (23) may be employed to evaluate m. in turn, equation (24) may evaluate i, so that the discrete pores size distribution (ai, i) may be obtained. results and discussion halib et al. [31] applied the lf-nmr characterization to a porous hydrogel composed of acrylic acid and bacterial cellulose extracted from nata de coco. the gel was realized by dispersing cellulose powder in water (1% w/v) and, then, by adding acrylic acid in a ratio 1:4 (w/w) compared to cellulose. the mixture was poured into a plastic container (size 12 x 12 x 1 cm) and, then, irradiated in the air with an electron beam of 35 kgy intensity (5 kgy per pass, eps 3000, japan). the freshly prepared gel was dried and then left to swell at 37 °c in distilled water for 48 h to reach the equilibrium swelling. lf-nmr characterization (bruker minispec mq20 (0.47 t, 20 mhz)) was performed by measuring the water protons transverse relaxation time inside the hydrogel (t2m) and the dependence of the water self-diffusion coefficient inside the hydrogel (dw) on the diffusion time (td). in the case of relaxation experiments, cpmg sequence {90°[-180°-(echo)]n-tr} with an 8.36-s wide 90° pulse,  = 250 s, tr (sequences repetition rate) equal to 5 s and 36 repetitions was adopted. dw determination was achieved by the execution of pgse sequence (performed in triplicate at 37 °c) where the two gradients (duration  = 500 s) occurred, respectively, 100 s after the application of the first and the second pulse. figure 3 displays equations (16) and (20) best fit to experimental diffusion (dw) and relaxation data (fid = i(t)). while equation (16) best fit provides a = 14.6 m, equation (20) best fit allows concluding that four classes (a1 = 38 %, a2 = 47 %, a3 = 9 %, a4 = 6 %) are required to describe fid decay. in addition, it is noticeable that, in the second summation of equation (20), for each class, the first relaxation time (t2i1) always played a predominant role (ii1 ~ 1). finally, the use of equations (23) and (24) allowed concluding that m = 1.1*10 -5 m/s and that pores dimension may be subdivided into the following four classes: 1 = 91 m, a1 = 38 %; 2 = 33 m, a2 = 47 %; 3 = 10 m, a3 = 9 %; 4 = 1.3 m, a4 = 6 %. interestingly, these findings are close to those obtained by image analysis performed on environmental sem pictures. indeed, this analysis revealed that pores characterized by a smaller diameter than 30 m represent 4 % of the whole pores volume, those with a diameter comprised between 30 and 80 m are 29 %, while the remaining 67 % are pores characterized by a diameter ranging between 80 and 120 m. furthermore, as sem pictures did not reveal the presence of pores close to 1 m, we assumed that the fourth class identified by lf-nmr analysis simply represents the water fraction outside pores swelling the hydrogel continuous structure (meshes). by relying on this hypothesis, it could be possible to conclude that the majority of hydrogel volume (81 %) was constituted by micrometric pores filled by water, while only 19 % of m. grassi et al. admet & dmpk 6(1) (2018) 34-46 42 the whole hydrogel volume was represented by polymeric meshes swollen by water. figure 3. porous hydrogel composed of acrylic acid and bacterial cellulose (37 °c). a) left vertical axis (fid) and bottom horizontal axis (time t): equation (20) (solid black line) best fit to relaxation data (open circles). b) right vertical axis (water self-diffusion coefficient dw) and top horizontal axis (square root of the diffusion time td): equation (16) (solid gray line) best fit to experimental data (gray circles). dashed arrows help to catch the correct vertical and horizontal axis. adapted from [37]. fiorentino et al. [20] focussed the attention on poly(l-lactic acid) (plla) scaffolds for tissue engineering applications prepared through a thermally induced phase separation (tips) protocol. lf-nmr characterization, performed by means of the cpmg and pgse (the two gradients of duration  = 1000 s were applied, respectively, 1000 s after the first and second pulses) sequences previously illustrated, revealed that m = 2.4*10 -5 m/s, a = 92 m, and that 47 % of pores showed a diameter  = 367 m, 44 %  = 101 m, and 9 %  = 17 m. this evaluation is based, respectively, on equations (16) and (20) best fit to experimental diffusion (dw) and relaxation data (fid = i(t)) displayed in figure 4. interestingly, lf-nmr evaluation of pores size distribution was not so far from that descending from sem picture evidencing pores ranging from 400 m to 100 m. a further qualitative support for the lfnmr findings was provided by ct (x-ray microcomputed tomography characterization) that revealed the presence of pores ranging from 400 m to 100 m also in the scaffold bulk. the comparison between figure 3 and figure 4 demonstrates that larger pores are associated with a longer relaxation process (~ 2000 ms in figure 3 and ~ 4500 ms in figure 4) and a higher water self-diffusion coefficient. moreover, in the case of larger pores (figure 4), it was observed that approximately three relaxation times (t2i1, t2i2, t2i3) are required in the second summation of equation (20). fiorentino [37], in her ph.d. thesis, applied lf-nmr characterization (cpmg and pgse sequences set up equal to that used for figure 4 data) to study the time evolution of the topology of alginate/hydroxyapatite scaffolds for bone regeneration. in particular, she compared the time evolution (21 days) of the pores size distribution of empty scaffolds and of cells seeded scaffold (osteosarcoma mg-63 human cells) filled with dulbecco’s modified eagle’s medium (dmem). interestingly, she noticed that the fate of smaller pores (diameter < 110 m) was the same in empty and seeded scaffolds. this sounds reasonable as mg-63 cells 1e-09 1.5e-09 2e-09 2.5e-09 3e-09 0 0.05 0.1 0.15 0 10 20 30 40 50 60 70 80 90 0 500 1000 1500 2000 2500 d w (m 2 /s ) td 0.5(s0.5) fi d = i (t ) t(ms) admet & dmpk 6(1) (2018) 34-46 lf-nmr for the characterization of biological tissues doi: 10.5599/admet.6.1.430 43 are unable to grow in smaller pores than 100 m. on the contrary, she could verify that the diameter of larger pores (200 – 600 m) decreased more rapidly in seeded scaffold with respect to empty scaffolds and this seemed correct due to cells growth and/or to the production of extracellular matrix. interestingly, this characterization proved to be safe for what concerns cell survival. figure 4. plla scaffold (37 °c). a) left vertical axis (fid) and bottom horizontal axis (time t): equation (20) (solid black line) best fit to relaxation data (open circles). b) right vertical axis (water self-diffusion coefficient dw) and top horizontal axis (square root of the diffusion time td): equation (16) (solid gray line) best fit to experimental data (gray circles). dashed arrows help to catch the correct vertical and horizontal axis. adapted from [20]. abrami [38], in her ph.d. thesis, explored the possibility of finding structural differences among samples of the femoral head withdrawn from osteoarthritis and osteoporosis patients (cpmg and pgse sequences set up was equal to that used for figure 3 data). interestingly, she found that not only the average diameter of osteoarthritic samples is smaller than that of osteoporotic samples but also pores size distributions were different. osteoporotic samples were characterized by a pores size distribution (ai, i) shifted towards larger pores with respect to the osteoarthritic samples one. in addition, she found that the porosity of osteoporotic samples was smaller than that of osteoarthritic samples. abrami et al. [27] employed lf-nmr to gain an insight into the nanostructure of a highly biocompatible hydrogel constituted by crosslinked alginate and pluronic f127 (pf127), a synthetic poly(oxyethyleneoxypropyleneoxyethylene) tri-block copolymer able to generate micelles that organize themselves to yield a soft gel when added to water. by means of relaxation and diffusion tests, she discovered that pf127 pluronic micelles, organizing themselves in cubic domains, induce the distortion of the alginate network so that larger meshes, filled with ordered micelles domains, coexist with smaller meshes which may host only single pf127 micelles. this result was compatible with the evidence provided by tem pictures. abrami et al. [21] applied lf-nmr to establish a relation between the average relaxation time t2m of the sputum of patients suffering from chronic obstructive pulmonary diseases (copd, like cystic fibrosis) in relation to the patient's clinical conditions. this study was conducted by adopting the cpmg sequence set up described for the figure 3 data. 0 0.05 0.1 0.15 0.2 0.25 1e-09 1.5e-09 2e-09 2.5e-09 3e-09 0 10 20 30 40 50 60 70 80 0 1000 2000 3000 4000 5000 td 0.5 (s0.5) d w (m 2 /s ) fi d = i (t ) t(ms) m. grassi et al. admet & dmpk 6(1) (2018) 34-46 44 figure 5. sputum mean relaxation time (t2m) referring to healthy subjects and three patients suffering, respectively, from asthma, copd (lung bacterial infection), and cystic fibrosis (37 °c). vertical bars indicate data standard error. figure 5 clearly demonstrates that, while healthy sputum is characterized by the highest t2m, t2m decreases with increasing disease severity (asthma, copd, cf). this is essentially due to the increasing presence of substances such as dna, polymers (usually bacterial alginates), mucin, and albumin, which are typical of the considered pathologies. indeed, all these substances act as a solid substrate that, as discussed in the “surface effect on t2m” section, induce the reduction of the water hydrogens relaxation time in a concentration-dependent fashion. it is also worth mentioning that the differences among healthy volunteers and patients are unconfined to the decreasing t2m values. indeed, tab. 2 clarifies that, while only one relaxation time is sufficient to describe water hydrogens relaxation in healthy volunteers, the required number of relaxation times increases when considering asthma, copd, and cf patients. this evidence is connected to the increasing heterogeneity of the solid component spatial organization inside the liquid sputum part. thus, we may conclude that not only t2m value matters for the evaluation of clinical conditions but also an equally important role is played by the relaxation spectrum characteristics (ai, t2i). table 2. components of the mean sputum relaxation time t2m. healthy asthma copd cf t2m(ms) 3223 ± 200 2225 ± 5 908 ± 25 268 ± 2 a1%(-) 100 84 ± 5 32 ± 2 17 ± 1 t21(ms) 3213 ± 200 2405 ± 144 1633 ± 98 658 ± 39 a2%(-) 16 ± 1 32 ± 2 57 ± 3 t22(ms) 1257 ± 75 794 ± 48 230 ± 14 a3%(-) 36 ± 2 26 ± 2 t23(ms) 370 ± 22 103 ± 6 a further support for the relation between lf-nmr characterization and patients clinical conditions is provided by figure 6 showing a linear correlation between t2m and fev1 (forced expiratory volume in the first second), a typical test used to evaluate patients lung clinical conditions [21]. indeed, the correlation coefficient r = 0.836 (tr(16, 0.95)< 6.10) and the correlation ratio  = 0.901 (t(16, 0.95)< 8.34) are both statistically higher than zero and statistically equal each other (ftest: f(7, 9, 0.95) > 0.77). 0 500 1000 1500 2000 2500 3000 3500 4000 healthy asthma copd cf t 2 m (m s) admet & dmpk 6(1) (2018) 34-46 lf-nmr for the characterization of biological tissues doi: 10.5599/admet.6.1.430 45 figure 6. relation between fev1 (forced expiratory volume in the first second) and sputum mean relaxation time (t2m) referring to cf patients. black circles represent experimental data, while the solid line is the linear interpolant. gray large dot indicates healthy subjects. conclusions we believe that we have demonstrated the potentiality of lf-nmr for the characterization of different systems such as polymeric gels, scaffolds, and biological tissues (bones and sputum of cf patients, for instance). clearly, lf-nmr potentiality may be significantly improved by using this technique in conjunction with others such as rheology, sem, and tem. acknowledgments: this work has been supported by the “fondazione beneficentia stiftung” vaduz, by fondo di ateneo fra 2016 – trieste university, biofluo project, por fesr 2014-2020, fvg, italy. references [1] k.r. brownstein, c.e. tarr, physical review a 19 (1979) 2446–2453. [2] k.r. brownstein, c.e. tarr, journal of magnetic resonance 26 (1977) 17-24 [3] p. mitra, p. n. sen, l.m. schwartz, physical review b 47 (1993) 8565–8574 [4] h. c. torrey, j. corringa, d.o. seevers, j. uebersfeld, phys. rev. lett. 3 (1959) 418-419. [5] m.m. chui, r. j. phillips, m.j., mccarthy, journal of colloid and interface science 174 (1995) 336–344. [6] b.p. hills, s.f. takacs, p.s. belton, food chemistry, 37 (1990) 95-111. [7] v. aroulmoji, m. mathlouthi, l. feruglio, e. murano, m. grassi, food chemistry 132 (2012) 1644– 1650. [8] m.d. de’nobili, a.m. rojas, m. abrami, r. lapasin, m. grassi, journal of food engineering 165 (2015) 82–92. [9] g.j.w. goudappel, j.p.m. van duynhoven, m.m.w. mooren, journal of colloid and interface science 239 (2001) 535-542. [10] j.p.m. van duynhoven, g.j.w. goudappel, g. van dalen, p.c. van bruggen, j.c.g. blonk, a.p.a.m. eijkelenboom, magnetic resonance in chemistry 40 (2002) 551-559. [11] j. bryan, a. kantzas, c. bellehumeur, sca 39 (2002) 1-12. [12] g. h. sorland, k. djurhuus, h. c. wideroe, j.r. lien, a. skauge, diffusion fundamentals 5 (2007) 4.14.15. 0 20 40 60 80 100 120 140 0 1000 2000 3000 4000 fe v 1 (% ) t2m (ms) m. grassi et al. admet & dmpk 6(1) (2018) 34-46 46 [13] t. stait-gardner, s.a. willis, n.n. yadav, g. zheng, w.s. price, diffusion fundamentals 11 (2009) 15.115.22. [14] m.d. hurlimann, k.g. helmer, l.l. latour, c.h. sotak, journal of magnetic resonance series a 111 (1994) 169-178. [15] p. fantazzini, r.j.s. brown, g.c. borgia, magnetic resonance imaging 21 (2003) 227-234. [16] x. wang, q. ni, journal of orthopaedic research 21 (2003) 312-319. [17] c.s. rajapakse, m.b. zadeh, c. li, w. sun, a.c. wright, f.w. wehrli, radiology 276 (2015) 526-534. [18] q. ni, j. d. king, x. wang, measurement science and technology 15 (2004) 58-66. [19] m. grassi, s. fiorentino, r. farra, b. dapas, g. grassi in polysaccharide hydrogels: characterization and biomedical applications. pan stanford, singapore, chapter 4, 2015, pp. 139-166. [20] s.m. fiorentino, f. carfì pavia, v. la carrubba, v. brucato, m. abrami, r. farra, g. turco, g. grassi, m. grassi, international journal of polymeric materials and polymeric biomaterials 66 (2017) 469-477. [21] m. abrami, f. ascenzioni, e.g. di domenico, m. maschio, a. ventura, m. confalonieri, s. di gioia, m. conese, b. dapas, g. grassi, m. grassi, magnetic resonance in medicine 79 (2018) 2323–2331. [22] p.y. ghi, d.j.t. hill, a.k. whitaker, biomacromolecules 3 (2002) 991-997. [23] r. barbucci, g. leone, a. chiumiento, m. e. di cocco, g. d’orazio, r. ginaferri, m. delfini, carbohydrate research 341 (2006) 1848-1858. [24] k. hietalahti, m. skrifvars, a. root, f. sundhoolmp, journal of applied polymer science 68 (1999) 671680. [25] t. brand, s. richter, s. berger, journal of physical chemistry b 110 (2006) 15853-15857. [26] m. dobies, s. kusmia, s. jurga, acta physica polinica a 108 (2005) 33-36. [27] m. abrami, i. dagostino, g. milcovich, s. fiorentino, r. farra, f. asaro, r. lapasin, g. grassi, m. grassi, soft matter 10 (2014) 729-737. [28] s. meiboom, d. gill, the review of scientific instruments 29 (1958) 688-691. [29] d.p. gallegos, k. munn, d.m. smith, d.l. stermer, journal of colloid and interface science 119 (1987) 127-140. [30] t. coviello, p. matricardi, f. alhaique, r. farra, g. tesei, s. fiorentino, f. asaro, g. milcovich, m. grassi, express polymer letters 7 (2013) 733-746. [31] n. halib, m. amin, i. ahmad, m. abrami, s. fiorentino, r. farra, g. grassi g, f. musiani, r. lapasin, m. grassi. european journal of pharmaceutical science 63 (2014) 326–333. [32] g.w. scherer, journal of sol-gel science and technology 1 (1994) 285-291 [33] e.o. stejskal, j.e. tanner, journal of chemistry physics 42 (1965) 288-292. [34] k.r. brownstein, c.e. tarr, journal of magnetic resonance 26 (1977) 17-24. [35] l.l. latour, p.p. mitra, r.l. kleinberg, c.h. sotak, journal of magnetic resonance a 101 (1993) 342346. [36] m. holz, s.r. heil, a. sacco, physical chemistry chemical physics 2 (2000) 4740–4742. [37] s.m. fiorentino, ph.d. thesis, xxvii cycle, university of trieste, dept of engineering and architecture, 2015. [38] m. abrami, m. ph.d. thesis, xxix cycle, university of trieste, dept of life sciences, 2017. ©2018 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ dissolution rates of ciprofloxacin and its cocrystal with resorcinol doi: 10.5599/admet.6.1.497 61 admet & dmpk 6(1) (2018) 61-70; doi: http://dx.doi.org/10.5599/admet.6.1.497 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index short communication dissolution rates of ciprofloxacin and its cocrystal with resorcinol clara ràfols 1, *, hanan fael 1 , elisabet fuguet 1 , breeze outhwaite 2 , samuel lee 2 , rebeca ruiz 2 1 departament de química analítica and institut de biomedicina (ibub), universitat de barcelona; martí i franquès 111, 08028 barcelona, spain 2 sirius analytical, forest row, east sussex, uk * corresponding author: e-mail: crafols@ub.edu; tel.: +34 934021284; fax: +34 934021233 received: january 12, 2018; revised: march 16, 2018; published: march 25, 2018 abstract the synthesis of cocrystals is presented as an alternative to improve the properties of active pharmaceutical ingredients, especially those related to solubility and dissolution rate. in this work the dissolution rate of ciprofloxacin, a zwitterionic fluoroquinolone antibiotic, has been compared to its cocrystal with resorcinol. to this end, dissolution rate has been determined at several biorelevant ph values, and also in two simulated gastrointestinal fluids (fessif and fassif). results show that both, ciprofloxacin and the cocrystal, dissolve more slowly as ph increases (from 2.0 to 7.4), as ionization degree of ciprofloxacin decreases. in addition, dissolution is not enhanced by the components of the gastrointestinal fluids. keywords dissolution rate; cocrystal; ciprofloxacin; solid state; fassif; fessif; simulated gastrointestinal fluid introduction the dissolution rate plays an important role in drugs with limited solubility. in fact, for drugs with very poor aqueous solubility, the rate at which they dissolve is often the slowest step, and it exerts a ratelimiting effect on drug bioavailability. currently about 60 % of the drugs coming from synthesis are poorly soluble and hence have a low oral bioavailability [1]. diverse strategies have been developed to improve solubility, dissolution rate and subsequently the bioavailability of these drugs. these include formation of salts, inclusion complexes with cyclodextrines, amorphization, formulation of solid dispersions with hydrophilic polymers, changing the crystal form (polymorphs, hydrates, etc.) [2-5]. a promising strategy is the crystallisation of the drug with a suitable coformer to obtain a cocrystal, which may have different physical properties compared to the drug, such as melting point, moisture sorption, compressibility, solubility, and dissolution rate, without altering the pharmacological effect of the drug candidate [4-7]. a cocrystal can be defined as a stoichiometric multi-component system connected by non-covalent interactions where all the components present are solid under ambient conditions [5]. a pharmaceutical cocrystal is composed of an active pharmaceutical ingredient (api) and a suitable and pharmaceutically http://www.pub.iapchem.org/ojs/index.php/admet/index c. ràfols et al. admet & dmpk 6(1) (2018) 61-70 62 accepted molecule called a coformer. cocrystals are constructed from intermolecular interactions such as van der waals contact forces, π-π interactions, and especially hydrogen bonding. there are several methods to obtain cocrystals in the literature, such as slow solvent evaporation, slurry crystallization, solid state grinding, and melting [8-13]. the api selected in this study is ciprofloxacin (cip), a poorly soluble fluoroquinolone antibiotic active against both gram-positive and gram-negative bacteria, which is mainly used in the treatment of lower respiratory and urinary tract infections. the aim of this study is to compare the dissolution rate of cip in different media and ph values to that of the prepared cocrystal with resorcinol. in addition, possible changes in the solid state of the drug and its cocrystal during the dissolution process are also investigated. materials and methods instruments dissolution rate assays were carried out using a glpka titrator from sirius analytical instruments ltd (forest row, uk), equipped with a sirius d-pas spectrometer, a bifurcated fibre-optic dip probe from hellma analytics (müllheim, germany) with path length of 1 cm, and a two channels solvent degasser from smi-labhut ltd. (churcham, uk). the apparatus was controlled from a computer running the refinementpro2 software. acidity constant determinations were done with the same instrument as described elsewhere [14]. the powder x-ray diffraction (pxrd) characterization was performed using a panalytical x’pert pro mpd / powder diffractometer of 240 mm radius equipped with a pixcel detector from panalytical b.v. (almelo, the netherlands). the apparatus was set in a configuration of convergent beam with a focalizing mirror and a transmission geometry, with flat samples sandwiched between low absorbing films. the detector active length was 3.347º. work power was 45 kv – 40 ma with a defined beam height of 0.4 mm. five repeated scans were done from 2 to 60 2º with a step size of 0.026 2º and a measuring time of 40 seconds per step. calorimetric analysis of the samples was performed on differential scanning calorimetry (dsc) (mettler toledo dsc822e, switzerland). samples equivalent to 2-5 mg of the drug or cocrystal were loaded into aluminum crucibles. the thermal behavior of each sample was investigated in a temperature range of 30300 °c with 10 °c/min heating rate under a continuous flow of dry nitrogen at 50 ml/min. reagents ciprofloxacin (>98 %), resorcinol (>99 %), toluene (99.9 %), and potassium chloride (>99 %) were from sigma-aldrich (st. louis, mo, usa). sodium acetate anhydrous (>99 %), potassium dihydrogen phosphate (>99.5 %), and 0.5 m naoh standard solution (titrisol®) were for merck (darmstad, germany). fassif (fasted state simulated intestinal fluid) and fessif (fed state simulated intestinal fluid) powders were from biorelevant.com (london, uk). the simulated gastrointestinal solutions of fassif and fessif were prepared as specified by the manufacturer. water was purified by a milli-q plus system from millipore (bedford, ma, usa) with resistivity of 18.2mω cm. methods ciprofloxacin cocrystal was prepared using slurry crystallization method. 100 mg of ciprofloxacin was admet & dmpk 6(1) (2018) 61-70 dissolution rates of ciprofloxacin and its cocrystal with resorcinol doi: 10.5599/admet.6.1.497 63 mixed with 42 mg of resorcinol in 2 ml of toluene. the suspension was kept under stirring overnight and solid was collected next day by filtration under vacuum for 30 min. then, it was characterized by dsc and powder x-ray diffraction (pxrd) (figure 1). it can be observed that the signals corresponding to the cocrystal are different than the ones of the pure components. the formed cocrystal had 1:1 stoichiometry. figure 1: (a) dsc curve for resorcinol, the cocrystal, and cip. (b) pxrd diffractogram for resorcinol, the cocrystal, and cip. for dissolution rate measurements the “gi dissolution method” has been followed [15]. briefly, disks of 3 mm diameter containing 8-10 mg of drug or cocrystal (8 mg of cocrystal containing around 5 mg of cip, according to a 1:1 stoichiometry) were prepared by applying a constant pressure of 0.1 ton. a sirius glpka system was used to perform the measurements. temperature was 25±1 °c and ionic strength was kept constant at 0.15 m. study was performed at different ph sectors (2.0, 4.0, 5.5, and 7.4) which are typically encountered in the gi tract. to perform experiments 1.5 ml of a 0.125 m acetate and 0.125 m phosphate buffer solution at ph = 1.6 was introduced into the sample vial. then, the instrument added 13.5 ml of 0.15 m aqueous kcl solution, which raised initial ph around 2. dissolution started and the medium was stirred at a constant rate throughout the experiment. after 30 minutes, 0.5 m koh solution was automatically dispensed to adjust to the next ph, and so on. in all experiments uv-vis spectra were recorded at fixed intervals, normally 30 s. dissolution rate was determined individually at each ph sector, and then experiments were repeated in a full sequence of phs (from 2.0 to 7.4) staying 30 minutes at each ph. additionally, dissolution rates were also determined in the acetate/phosphate buffer at ph 5.0 and c. ràfols et al. admet & dmpk 6(1) (2018) 61-70 64 6.5, and in 15 ml of two simulated gastrointestinal fluids (fessif, ph 5.0 and fassif, ph 6.5). the concentration of sample in solution at each time point is determined from the spectroscopic data in the 320-410 nm wavelength range, using previously determined molar extinction coefficients. in this range solutions of fassif, fessif and resorcinol present a minimum absorbance so that they don’t interfere in the quantification of ciprofloxacin. data corresponding to saturated signals were excluded for the calculation. equation 1 was used to calculate the dissolution rate of the compounds,   ( )0dx = 1-e k t t st       (1) where [x]t is the weight (in grams) of the compound in solution at time t (min), s is the extrapolated solubility (g) of the drug, kd is the rate constant for dissolution (min -1 ), and t0 is a term allowing for a temporal offset. the dissolution rate (g min -1 ) is given by the product kd·s. after the dissolution rate experiments, the solid state of the remaining disks was analysed using pxrd. results and discussion dissolution behaviour of ionisable compounds is strongly dependent on ph value of the surrounding medium. as drug moves throughout the gastrointestinal tract (git) it is exposed to different ph environments that affect its degree of ionization, solubility, and thus dissolution or precipitation, according to the drug ionization constants. cip is a zwitterionic compound with pka1 = 6.20±0.04 and pka2 = 8.56±0.06 (values provided at 25 °c and 0.15 m ionic strength) [16]. however, cip exists in four different microspecies (x , xh ± , xh 0 , and xh2 + ) in solution. völgyi et al. [17] studied the acid base equilibria of cip and determined its protonation macroconstants and microconstants. figure 2 shows the species distribution curve of this compound according to the values provided in the study. the zwitterionic species is always predominant relative to the non-charged species independently of the ph of the solution. cip is positively charged at low ph values, and its ionization degree decreases as ph increases. the maximum percentage of zwitterionic form and neutral forms exist at ph 7.3, where cip has a log s0 around -3.7 (-3.62 [18], -3.72 [19], and -3.76 measured in this work by the shake-flask method according to [20]). at higher ph values, ionization degree rises again due to the deprotonation of the basic group. on the contrary, resorcinol is a highly soluble very weak diacid which remains non‐ionized (neutral) in the whole ph range between 2 and 7.4. figure 3 shows the aqueous dissolution profile at separate ph sectors for ciprofloxacin and its cocrystal with resorcinol (cip-r). the dissolution rate reached a maximum value at ph 2, where both cip and cip-r totally dissolve. as ph of the medium increases, the percentage of cip dissolved and its dissolution rate start to decrease, as stated in table 1. admet & dmpk 6(1) (2018) 61-70 dissolution rates of ciprofloxacin and its cocrystal with resorcinol doi: 10.5599/admet.6.1.497 65 figure 2: species distribution diagram for cip, according to the acidity macroconstant and microconstantvalues provided in [17]. table 1: percentage of cip dissolved, and dissolution rate of cip and cip-r in the dissolution experiments at different ph sectors. ph % cip dissolved dissolution rate (mg/min) ciprofloxacin cocrystal ciprofloxacin cocrystal 2.01 ± 0.04 97.5 ± 0.1 100.0 ± 0.3 1.2 ± 0.2 0.56 ± 0.06 4.01 ± 0.05 41 ± 3 98.0 ± 0.5 0.13 ± 0.01 0.23 ± 0.9 5.41 ± 0.03 7.1 ± 0.7 11 ± 1 0.017 ± 0.006 0.021 ± 0.001 7.41 ± 0.08 3.7 ± 0.2 6 ± 1 0.0033 ± 0.0004 0.003 ± 0.002 this is the expected behavior according to the ionization state of cip. the percentage of dissolved cip decreases dramatically when changing ph from 2 to 4, with only 41% dissolved after 30 minutes (table 1), and drops to around 6 and 4% at ph 5.5 and 7.4, respectively. instead, the cocrystal dissolves totally after 25 minutes at ph 4, but its dissolution also decreases significantly at ph 5.5 and 7.4 (11 and 6% of cip dissolved) respectively. dissolution rate is directly related to the slope of the dissolution curve. comparison of dissolution rates between cip and its cip-r evidences that at very low ph values (ph 2) the dissolution rate of cip is more than twice the one of the cocrystal, probably due to the effect of resorcinol (neutral at this ph), which slows dissolution of cip-r down. however, as ph increases the dissolution rate of cip and cip-r tend to be the same, reaching a minimum value of 0.003 mg min -1 at ph 7.4. on the contrary, the full ph experiment (figure 4) does not reflect the actual dissolution rate of the compound at each ph, as both cip and cip-r dissolve almost totally at the first sector (ph 2). then solutions stay supersaturated at the following ph sectors (4.0, 5.5, and 7.4). in case of cip there is evidence of precipitation at ph 7.4, since the signal falls to zero. although this kind of experiment does not allow knowing the dissolution behavior of the compounds at a given ph because each sector is influenced by the previous one, it can simulate better the processes that compounds experience when moving along the gastrointestinal tract, where different ph environments are encountered. c. ràfols et al. admet & dmpk 6(1) (2018) 61-70 66 figure 3: dissolution profile at separate ph sectors for cip (●) and its cip-r (○): (a) ph = 2; (b) ph = 4; (c) ph = 5.5; (d) ph = 7.4. x-ray analysis of the surface of the solid remaining in the disks indicates that approximately 50% of the solid remaining in cip tablets is transformed to cip·3h2o, whatever the working ph. meanwhile, the surface of the cocrystal tablets contain about 90% of the original cocrystal, and only 10% of cip·3h2o. these transformations in the tablet surface may modify, to some extent, the dissolution of the compounds, especially for cip where the percentage of transformation to the trihydrated form is higher. in general, the appearance of hydrated forms slows the dissolution process down because of the lower ion-dipole interaction energy liberated on its dissolution [21]. only at ph 4.0 a slightly decrease in the dissolution rate of cip compared to cip-r was observed. nevertheless, there are not enough evidences to know the reasons behind cip-r higher dissolution rate. on one hand the cocrystal itself could dissolve faster than the pure api, and on the other hand there could be a decrease in the dissolution of the api caused by the transformation into cip·3h2o on the surface of the tablet. results of dissolution rate determinations in simulated gastrointestinal media are shown in table 2 and figure 5. experiments have been done not only in the two gastrointestinal fluids, but also in acetate/phosphate buffer at the same ph. in this way it is possible to evaluate the real effect of the gastric media components (mainly bile salts and lecithin) to dissolution behavior of the compounds, independently of ph. table 2 and figure 5a show a different behavior of cip and cip-r at ph 5. the cocrystal dissolves faster than the pure api. however the differences cannot be attributed to fessif components since in both cases dissolution rates are practically the same in fessif and in the plain buffer. again, the reason can be due to a faster dissolution of the cocrystal, or due to the transformation of the api into the trihydrated form, as stated by pxrd. in fact, x-ray analysis of the remaining tablets indicated that the surface of the cocrystals were hardly modified (90% cocrystal and 10% cip·3h2o) in these two media, whereas the tables of cip alone were strongly modified (10% cip and 90% cip·3h2o). as expected, admet & dmpk 6(1) (2018) 61-70 dissolution rates of ciprofloxacin and its cocrystal with resorcinol doi: 10.5599/admet.6.1.497 67 dissolution rate values of cip and cip-r at ph 5 are in between the ones determined at ph 4.0 and 5.5 (table 1). figure 4: full ph dissolution experiment for cip (●) and cip-r (○). results in fassif and the corresponding plain buffer (ph 6.4) (figure 3b) show that either dissolution rate or percentage of cip dissolved for cip and cip-r are equivalent. again, these results are located in between the results at ph 5.5 and 7.4 in table 1 for both compounds. thus, apparently there is no effect of fassif or fessif components on the dissolution rate; only the ionization degree seems to play an important role in their dissolution. the zwitterionic nature of cip may be one of the reasons that gastrointestinal simulated fluids have no solubilizing effect on cip, since other studies [22-24] point out that the micelles formed in these fluids have negatively charged surface, what means attraction for cations, repulsion for anions, and may be no effect on zwitterions. table 2: percentage of cip dissolved, and dissolution rate of cip and cip-r; in fessif, fassif and in standard buffer at the same ph (5.0 and 6.5 respectively). ph % cip dissolved dissolution rate (mg/min) ciprofloxacin cocrystal ciprofloxacin cocrystal fessif (ph = 4.94 ± 0.05) 23 ± 4 94 ± 2 0.07 ± 0.01 0.25 ± 0.04 buffer ph = 4.94 ± 0.05 26 ± 1 105 ± 4 0.067 ± 0.001 0.29 ± 0.1 fassif (ph = 6.37 ± 0.05) 1.7 ± 0.1 2.5 ± 0.5 0.0038± 0.0004 0.005 ± 0.002 buffer ph = 6.43 ± 0.04 2.2 ± 0.2 2.10 ± 0.07 0.005 ± 0.001 0.003 ± 0.001 conclusions dissolution rate of ciprofloxacin and its cocrystal with resorcinol have been determined at several ph values of pharmaceutical interest in an acetate-phosphate buffer. results show that dissolution rate of ciprofloxacin decreases according to ionization state of ciprofloxacin, i.e., it is maximum at ph 2 and has the minimum value at ph 7.4, where ciprofloxacin is mainly in its zwitterionic form. the same behavior is c. ràfols et al. admet & dmpk 6(1) (2018) 61-70 68 observed with the cocrystal. however, ciprofloxacin dissolves faster than the cocrystal at ph 2, but the cocrystal dissolves faster than ciprofloxacin at intermediate ph values (4.0, 5.0 and 5.5). the reason can be attributed either to a better dissolution of the cocrystal itself, or a decrease in the dissolution of ciprofloxacin due to the formation of the more insoluble trihydrated form in the surface of the tablets, as stated by pxrd. gastric fluid components such as lecithin and bile salts have not shown any improvement in the dissolution rate of the studied compounds. figure 5: dissolution profile in simulated gastrointestinal fluids and buffer at the same ph. (a) cip in fessif (●), cip in buffer at ph 5 (▲), cip-r in fessif (○), cip-r in buffer at ph 5 (∆); (b) cip in fassif (●), cip in buffer at ph 6.5 (▲), cip-r in fassif (○), cip-r in buffer at ph 6.5 (∆). acknowledgements the authors from the university of barcelona are grateful for the financial support of the spanish government (project ctq2014-56253-p). references [1] e. merisko, g.g. liversidge, nanocrystals: resolving pharmaceutical formulation issues associated with poorly water-soluble compounds. in: particles; j.j. marty (ed.). marcel dekker, orlando, fl, usa, 2002. [2] s. sareen, g. mathew, l. joseph. improvement in solubility of poor water soluble drugs by solid dispersion. international journal of pharmaceutical investigation 2 (2012) 12–17. admet & dmpk 6(1) (2018) 61-70 dissolution rates of ciprofloxacin and its cocrystal with resorcinol doi: 10.5599/admet.6.1.497 69 [3] s.v. kurkov, t. loftsson. cyclodextrins. international journal of pharmacy 453 (2013) 167–180. [4] n.j. babu, a. nanjia. solubility advantage of amorphous drugs and pharmaceutical co-crystals. crystal growth and design 11 (2011) 2662–2679. [5] n. schultheiss, a. newman. pharmaceutical cocrystals and their physicochemical properties. crystal growth and design 9 (2009) 2950–2967. [6] n. qiao, m. li, w. schlindwein, n. malek, a. davies, g. trapitt. pharmaceutical co-crystals: an overview. international journal of pharmacy 419 (2011) 1–11. [7] r.d.b. walsh, m.w. bradner, s.g. fleischman, l.a. morales, b. moulton, n. rodriguez-hornedo, m.j. zaworotko. crystal engineering of the composition of pharmaceutical phases. chemical communications (2003) 186-187. [8] k. fucke, s. a. myz, t.p. shakhtshneider, e.v. boldyreva, u.j. griesser. how good are the crystallisation methods for co-crystals? a comparative study of piroxicam. new journal of chemistry 36 (2012) 1969-1977. [9] d.r. weyna, t. shattock, p. vishweshwar, m.j. zaworotko. synthesis and structural characterization of cocrystals and pharmaceutical cocrystals: mechanochemistry vs slow evaporation from solution. crystal growth and design 9 (2009) 1106–1123. [10] n. takata, k. shiraki, r. takano, y. hayashi, y. terada. co-crystal screening stanolone and mestanolone using slurry crystallization. crystal growth and design 8 (2008) 3032-3037. [11] t. friščić, w. jones. recent advances in understanding the mechanism of cocrystal formation via grinding. crystal growth and design 9 (2009) 1621-1637. [12] a. v. trask, w. jones. crystal engineering of organic co-crystals by the solid-state grinding approach. topics in current chemistry 254 (2005) 41-70. [13] n. blagden, d.j. berry, a. parkin, h. javed, a. ibrahim, p.t. gavan, l.l. de matos, c.c. seaton. current directions in co-crystal growth. new journal of chemistry 32 (2008) 1659-1672. [14] e. shoghi, e. fuguet, e. bosch, c. ràfols. solubility-ph profiles of some acidic, basic and amphoteric drugs. european journal of pharmaceutical sciences 48 (2013) 291-300. [15] t. gravestock, k. box, j. comer, e. frake, s. judge, r. ruiz. the “gi dissolution” method: a low volume, in vitro apparatus for assessing the dissolution/precipitation behaviour of an active pharmaceutical ingredient under biorelevant conditions. analytical methods 3 (2011) 560-567. [16] c. ràfols, x. subirats, j. rubio, m. rosés, e. bosch. lipophilicity of amphoteric and zwitterionic compounds: a comparative study of determination methods. talanta 162 (2017) 293-299. [17] g. völgyi, g. vizserálek, k. tackáks-novák, a. avdeef, k.y. tam. predicting the exposure and antibacterial activity of fluoroquinolones based on physicochemical properties. european journal of pharmceutical sciencies 47 (2012) 21-17. [18] d.l. ross, c.m. riley. aqueous solubilities of some variously substituted quinolone antibacterials. international journal of pharmaceutics 63 (1990) 237–250. [19] a. avdeef. physocochemical profiling (solubility, permeability and charge state). current topics in medicinal chemistry 1 (2001) 277-351. [20] a. avdeef, e. fuguet, a. llinàs, c. ràfols, e. bosch, g. völgyi, t. verbić, e. boldyreva, k. takácsnovác. equilibrium solubility measurement of ionizable drugs – consensus recommendations for improving data quality. admet & dmpk 4 (2016) 117-178. [21] d. giron, d.j.w. grant. evaluation of solid-state properties of salts (chapter 2) in: handbook of pharmaceutical salts. properties, selection, and use. p. h. stahl, c. g. wermuth (eds.). wiley-vch, weinheim, germany, 2002, pp. 19-40. [22] k. tackáks-novák, v. szöke, g. völgyi, p. horváth, r. ambrus, p. szabó-révész. biorelevant solubility of poorly soluble drugs: rivaroxaban, furosemide, papaverine and niflumic acid. journal of pharmceutical and biomedical analysis 83 (2013) 279-285. c. ràfols et al. admet & dmpk 6(1) (2018) 61-70 70 [23] j. fagerberg, o. tsinman, n. sun, k. tsinman, a. avdeef, c. bergström. dissolution rate and apparent solubility of poorly soluble drugs in biorelevant dissolution media. molecular pharmacy 7 (2010) 1419-1430. [24] g. ottaviani, d. gosling, c. patissier, s. rodde, l. zhou, b. faller. what is modulating solubility in simulated intestinal fluids? european journal of pharmceutical sciencies 41 (2010) 452-457. ©2018 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ utilizing in vitro transporter data in ivive-pbpk: an overview doi: 10.5599/admet.5.4.441 201 admet & dmpk 5(4) (2017) 201-211; doi: http://dx.doi.org/10.5599/admet.5.4.441 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index review utilizing in vitro transporter data in ivive-pbpk: an overview pankajini mallick college of pharmacy, department of pharmacological and pharmaceutical sciences, university of houston, houston, tx 77030, usa *corresponding author: e-mail: pankajini.mallick@gmail.com; tel.: 832-842-8343 received: september 18, 2017; revised: september 30, 2017; published: december 24, 2017 abstract in vitro-in vivo extrapolation (ivive) integrated in physiologically-based pharmacokinetic (pbpk) models have been increasingly used during drug discovery and development processes to predict human pharmacokinetic (pk) parameters. drug transporters can influence drug pharmacokinetics and are key aspects contributing to the development of a successful drug. this review provides a snapshot of challenges or shortcomings of in vitro and in vivo techniques for understanding the contribution of drug transporters to a drug’s pharmacokinetics. the paper also describes the potential of ivive-pbpk models as prospective approaches to predict the role of drug transporters in drug discovery and development. keywords pgp; bile acid transporter; drug transporters; drug metabolizing enzymes; drug induced liver injury introduction the role of drug transporters in drug efficacy and safety has been of much interest; drug transports play key roles in drug absorption, distribution, and excretion [1-3]. in drug discovery and development, drug transporters expressed in the intestine, liver, kidneys, and brain have received considerable attention [4-7]. many of the drug transporters that have been cloned and characterized belong to two major superfamilies, atp-binding cassette (abc) and solute carrier (slc). major efflux transporters of the abc family include mdr1 (p-glycoprotein, abcb1), bcrp (abcg2), and mrp2 (abcc2), and are localized to barrier tissues in the body, such as the intestine, liver, kidneys, blood-brain barrier, and placenta [8-11]. absorption of a drug is a key component to achieve good bioavailability, and for oral drugs, the majority of drug absorption occurs in the small intestine where the presence of villi and microvilli greatly increases the surface area for optimal absorption. thus, drug-transporter interaction in the intestine often contributes to poor absorption, resulting in low bioavailability. given that oral delivery is the preferred route of administration for drugs, intestinal transporters are potential targets to achieve optimal clinical oral plasma exposure for hydrophilic and polar drugs [12]. although the significance of drug transporters in absorption, distribution, metabolism, and elimination (adme) of drugs is well recognized, a knowledge gap exists in terms of change in drug transporter functionality (their amount and activity) in response to disease, and how pathophysiological pathways change transporter expressions [13]. because of the central role of the liver in drug metabolism, the role of transporters in the liver in drug interactions and drug responses, as well as drug-induced liver injury, has been in focus of late. http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:pankajini.mallick@gmail.com pankajini mallick admet & dmpk 5(4) (2017) 201-211 202 despite a steady growth in the field of transporter biology, which has bolstered transporter-mediated therapeutics, the physiological functions of many transporters are still largely unknown. these growing trends bring challenges for drug development, as well as for regulatory agencies, with regard to identification of membrane transporters that influence the disposition and safety of drugs, and also optimization of membrane permeability to improve oral absorption of new chemical entities. thus, utilization of in vitro transporter assays has become a critical tool for assessing a drug’s potential in vivo absorption properties and drug-drug interactions [14]. use of in vitro–in vivo extrapolation (ivive) integrated in physiological-based pharmacokinetic (pbpk) models, together known as ivive-pbpk models, has gained recent momentum in the pharmaceutical industry. since variation in protein expression of drug metabolizing enzymes and transporters is known to be a major complicating factor in ivive, the accurate quantification of these proteins in complex biological systems is essential for improving ivive-based pharmacokinetic predictions [15]. this review focuses on the use and shortcomings of qualitative and quantitative methods used to create drug transporter-based ivive-pbpk models in the liver, intestine, and brain. pitfalls of in vitro and in vivo systems used in drug transporter studies in vitro cell-based systems are used routinely to elucidate passive and active transporter processes acting to influence drug permeability. different in vitro systems include use of membrane vesicles or transfected cell lines for drug transporter studies; however, there are several pitfalls in estimating transporter kinetics using these conventional approaches [16]. considering that one drug can be transported by more than one uptake or efflux transporter, a limitation of these systems is the presence of a single transporter of interest. further, given that there exists an interplay between drug metabolizing enzymes and transporters, cell lines lacking metabolism machinery pose another limitation to their use. in vitro experiments can lead to biased estimation of apparent km, the biochemical constant in experiments that assumes similar km for both apical and basal transport directions [17-20]. another critical factor whether intracellular concentrations that drive efflux transport are accounted for [21]. drug concentrations within the donor and acceptor compartments are dynamic; thus, the assumption that unbound drug concentrations on either side of a membrane are in thermodynamic equilibrium [22, 23] is not applicable for drugs that are poorly permeable, actively transported, or extensively metabolized. a host of in vitro assays have been developed to mimic the complexity of biological systems; however, many questions remain regarding the translation of in vitro information to in vivo. therefore, in vivo systems provide the ultimate determination of the effect of drug transporters on disposition of drugs. transgenic and mutant animal models have provided an important tool for assessing drug efflux transporter activity. to this end, the efflux transporters p-glycoprotein (p-gp) and multidrug resistanceassociated protein (mrp1) have been the most frequently investigated in transgenic animals [24-26]. like in vitro systems, in vivo systems also have limitations. mice generally have two genes to encode transporter proteins, and silencing one of these genes may not sufficiently suppress the expression of the protein [25]. moreover, efflux transporters may work through a compensatory mechanism, where absence of one gene might be compensated for by another transporter gene [27]. furthermore, due to the ubiquitous expression of drug efflux transporters throughout the body, the removal of a gene or genes affects tissues other than the organ of interest. this can lead to undesirable outcomes such as altered pharmacokinetics, lack of viability, or systemic toxicity. to get around these limitations, specific transporters can be modulated with various agents [28]. the advantage of using drug transporter modulators is that these studies are generally performed with wild-type animals, and thus, the effect of a compensatory mechanism admet & dmpk 5(4) (2017) 201-211 utilizing in vitro transporter data in ivive-pbpk doi: 10.5599/admet.5.4.441 203 is not so important. however, due to overlap in substrate specificities, and lack of specificity, drug efflux transporter modulation is not as effective as gene knockout models [28]. further, if drug transporter modulators are dosed orally, it is often difficult to get an effective dose to the targeted organ; this is likely be due to first pass metabolism. even with intraperitoneal dosing, the limited effective dose of the modulator could be because of physiochemical properties of the modulator, such as protein binding and/or unfavorable tissue distribution [29]. importantly, the effects of modulators on drug transporters are transient and concentration dependent; therefore, proper dosing and study time points becomes critical variables in these studies. there exists potential species-specific differences in drug transporter activities and contributions, as exhibited in mdr1 functional studies, in which some compounds showed affinity for mdr1a/b activity in mouse mdr1 cells but not in human cells [30]. currently available in vitro and in vivo models have a limited capability to quantitatively predict the impact of transporters on intracellular drug concentrations. as pbpk modeling approaches consider physicochemical properties of drugs, in vitro cell-based passive permeability, and cellular uptake/efflux, as well as in vivo animal and human pharmacokinetic data to estimate unbound intracellular drug concentrations in different tissues, this approach provides an effective platform over currently available in vitro and in vivo models. however, these models involve a number of assumptions. use of in vitro models, such as three-dimensional cultured-hepatocytes that retain the activity of both apical and basolateral hepatic drug transporters, drug metabolizing enzymes, or in vivo multiple gene/tissue-selective knockouts, serves as an improved medium to bolster refinement of pbpk models for better prediction in humans. ivive-pbpk pbpk models offer a quantitative mechanistic platform for predicting drug pharmacokinetics. when modeled for pharmacodynamics, these models can be used to predict clinical efficacy and safety with varying drug doses. the ivive approach is well established for many drug metabolizing enzymes [31, 32]; however, the integration of drug transporters into pbpk models is lacking, with few positive examples reported [33-36]. ivive-pbpk models link in vitro systems to the in vivo system through algorithms and scaling factors [37, 38]. therefore, for an effective ivive-pbpk model to predict the behavior of a drug in vivo, accurate and relevant in vitro data in conjunction with the model system parameters of passive permeability and transporter-mediated flux are necessary. static models used alongside pbpk models in drug development are simple and require limited input data; however, due to simplification, these often result in conservative estimates, a lack of ability to capture parallel clearance and elimination pathways, are not suitable for non-linear kinetics, and lack physiology inputs. although pbpk models require extensive in vitro and in vivo data, they can capture transporter-mediated disposition, describe physiology, and can be applied for special populations such as pediatric patients, in which data is generally very limited. pbpk for the prediction of transporter-mediated drug induced liver toxicity membrane transporters expressed on hepatocytes and enterocytes play a critical role in homeostasis of endobiotics, e.g., bile acids. homeostasis of bile acid is maintained through enterohepatic circulation, which allows movement of bile acid molecules from the liver to the small intestine and back to the liver. bile acid transporter, also known as apical sodium-dependent bile acid transporter, is considered a major determinant of bile acid homeostasis in the body, and is an essential regulator of lipid and cholesterol homeostasis [39-41]. impaired bile acid export may lead to increased concentrations of liver bile acids, causing hepatocellular apoptosis and/or necrosis. thus, drugs with the potential to inhibit these pankajini mallick admet & dmpk 5(4) (2017) 201-211 204 transporters can disturb the disposition of both co-administered drugs and bile acid, contributing to the development of cholestatic drug-induced liver injury (dili) [42, 43]. in vitro studies with isolated membrane vesicles have demonstrated an association between dili and the inhibitory effects of drugs on bile acid efflux transporters [43-45]. in vitro findings may not translate directly to in vivo hepatotoxicity risk for various reasons, including complexity of bile acid homeostasis, feedback regulation of bile acid synthesis and transport, and dynamic drug/metabolite concentrations in the system [46, 47]. in vivo studies report that preclinical animals are less sensitive to dili caused by bile acid transporter inhibition, as compared to humans [48, 49]. therefore, mechanistic models that integrate physiological information and in vitro experimental data to evaluate dili mechanisms may be useful to prospectively predict hepatotoxic potential of new drug candidates [50]. the development of a mechanistic model to predict hepatic bile acid transport mediated dili of troglitazone (tgz) [51] demonstrated the significant contribution of pbpk modeling to the prediction of hepatotoxicity due to the inhibition of drug transporters, and the potential of a transporter-mediated pbpk model to aid drug development. the mechanistic pbpk model used to study the involvement of bile acid homeostasis in tgz hepatoxicity incorporated drug/metabolite disposition, bile acid physiology/pathophysiology, and hepatocyte life cycle. large population variability in bile acid exposure makes prediction of bile acid-mediated hepatotoxicity challenging; hence, the model included population-based analysis to overcome population variability and low incidence of hepatotoxicity. in a rat pbpk model, tgz was not hepatotoxic; however, in a stimulated human pbpk model, tgz resulted in a delayed increase in serum alanine aminotransferase (alt) (marker of liver injury). thus, the variability added to the parameters describing drug disposition, body weight, and sensitivity of atp synthesis to hepatic bile acid accumulation, allowed prediction of the incidence of tgz hepatotoxicity and the species specific bile acid-mediated dili mechanism. intestinal transporters linked to ivive-pbpk model for oral absorption the bioavailability of an orally administered drug is primarily dependent on the transporters expressed on the apical/basolateral side of enterocytes that influence both the fraction of drug absorbed [52, 53] and the fraction escaping gut metabolism. it is difficult to reproduce the complexity of the human gut within a laboratory setting; therefore, efforts are being made to incorporate drug absorption into pbpk models by taking into account several physiological elements of the gastro-intestinal tract, e.g., ph, active transporters, and metabolizing enzymes [54]. to mechanistically develop a pbpk model, the intestine essentially accounts for the region-specific anatomical and physiological differences. advanced models such as the advanced dissolution, absorption, and metabolism (adam) model and the advanced compartmental absorption and transit model (acat) are based on adaptations of the original compartmental absorption and transit (cat) model which accounts for small intestinal transit time (si), permeability, and radii. most importantly, these advanced models incorporate gastrointestinal transporter–metabolism interplay [55, 56]. a study by bruyere and co-workers has highlighted the importance of obtaining regional intestinal transporter expression data for incorporation in pbpk models [57]. this study estimated the scaling factor based on protein estimation of both drug metabolizing enzymes and transporters across the small intestine; these were used for predicting intestinal clearance and for the development of an ivive-pbpk model. the results demonstrated that accounting for the mdr1 distributions enhanced pbpk predictions of bioavailability for the development of the compound under study. likewise, a study by darwich and coworkers [56], which incorporated regional mdr1 and cyp3a expression into an adam model, demonstrated two points: non-uniform regional distribution of mdr1 in the small intestine and colon, and disparity between mrna and protein expression. this highlights that protein and mrna expression together with drug parameter data leads to successful capture of drug disposition. the mechanistic pbpk admet & dmpk 5(4) (2017) 201-211 utilizing in vitro transporter data in ivive-pbpk doi: 10.5599/admet.5.4.441 205 absorption model of oxybutynin (oxy) was based on a cat model that predicts the fraction absorbed based on permeability data in order to predict oxy’s oral bioavailability, for two different formulations: immediate-release (ir) and modified-release (mr) [58]. as the primary objective is to predict the fraction absorbed, this multi-compartmental absorption model was a simplified version of a cat model used to divide the small intestine into three compartments instead of seven, including the duodenum, jejunum, and ileum. the model allowed the drug amount to be modelled, either in solid or dissolved state, for each segment. assumptions made included: well mixed gi compartments, no drug degradation in the gi tract, only the dissolved drug to be absorbed, and absorption to be a non-saturable process. transfer of mass between adjacent segments was considered to follow first-order kinetics, controlled by a rate constant of gastric emptying. the dissolution rate of oxy was modeled based on spherical particles dissolving over time [55]. further, oxy’s regional intestinal metabolism was modeled through an enterocyte compartment, with the assumptions that the metabolism in the liver and intestine was equal, and there was no binding to enterocytes. the model showed that the fraction absorbed from the formulation was reduced compared to ir, despite higher intestinal availability of the formulation. considering the assumption that the fraction remaining in the liver and intestine is the same, the model prediction supported the hypothesis that higher bioavailability with the oxy formulation is due to increased intestinal availability. ivive-pbpk model for the central nervous system given the high failure rate of drugs developed for central nervous system (cns) indications [59], early and accurate prediction of drug penetration across the blood–brain barrier (bbb) is vital during drug development. further, the prediction of brain penetration for non-cns drugs is important during drug development, in order to avoid unwanted neurotoxicity. human pharmacokinetic data from the brain is highly restricted; thus, pbpk modeling is of significant value as it can be used to predict the target site concentrations in inter-species and inter-disease situations [60-62]. however, until recently, there has been relatively few published ivive-pbpk models to predict bbb permeability, mostly due to limitations in the current in vitro systems used to represent the complex and transporter-rich structure of the in vivo bbb, and due to difficulty in obtaining the appropriate in vitro–in vivo scaling factors [63]. a generic multicompartmental cns distribution model structure proposed by yamamoto et al [64], based on the compounds acetaminophen, atenolol, methotrexate, morphine, paliperidone, phenytoin, quinidine, remoxipride, and risperidone, successfully described the pharmacokinetics in plasma and different cns compartments [64]. prediction of this multi-compartmental brain pk model matched well with the observed concentration-time profiles, adequately describing the data for all compounds studied. the model structure comprised physiologically relevant brain compartments including brain extracellular fraction (brainecf), brain intracellular fluid compartment (brainicf), compartment of cerebrospinal fluid (csf) in lateral ventricle (csflv), compartment of csf in third and fourth ventricle (csftfv), compartment of csf in cisterna magna (csfcm), and csf subarachnoid space (csfsas). thus, it could successfully describe the pk in plasma and different cns compartments using microdialysis data. recently, there has been an increasing focus on translational modelling approaches compared to the traditional practice of fitting model parameters to preclinical in vivo data [65, 66]. ivive-pbpk models reported by fenneteau et al [67] and ball et al [68] demonstrate promising examples. this review has addressed imperative differences among various pbpk models in use, including details on brain compartmentalization and parameterization. the general approach of using in vivo data to obtain drug specific parameters and physiological parameter estimates by fitting to in vivo data has poor predictivity. similar to empirical pbpk models, the ivive-pbpk model uses a bottom-up approach through ivive, with the use of appropriate physiological scaling factors. fenneteau and colleagues scaled in vitro passive and efflux permeabilities of domperidone to in vivo pankajini mallick admet & dmpk 5(4) (2017) 201-211 206 intrinsic permeabilities, and then using literature-based physiological values for in vivo membrane surface area, they further scaled the permeabilities to whole organ permeabilities [67]. the in vitro intrinsic p-gp efflux permeability determined in caco-2 cells was corrected using the relative fraction of mdr1a/1b messenger rna expression measured in the brain compared to that in the intestine. two methods of ivive used for pbpk modelling of morphine and oxycodone describe different strategies for ivive of drug transporters at the bbb, which can be used in different phases of drug development [69]. the morphine ivive-pbpk model was based on data from in vitro caco-2 permeability and in vivo total concentrations in brain homogenate. on the other hand, the oxycodone ivive-pbpk model used data from complicated experiments generally performed in later stage of drug development, which included in vivo unbound extracelluar fraction (ecf) concentrations obtained from microdialysis and in vitro uptake kinetics in rat brain microvessel endothelial cells. relative activity factor (raf) was used to account for the difference between in vitro and in vivo data. since it is very difficult to obtain human brain homogenate or primary brain endothelial cells from isolated microvessels, determination of drug-specific parameters from in vitro systems of human origin is not always feasible. brain microdialysis and cerebrospinal fluid (csf) sampling is another approach to determine drug pharmacokinetics, but due to the invasive nature of these techniques, they are only carried out on diseased patients [70], which introduces significant challenges for the evaluation of model predictions with the human cns pbpk model. linking pbpk/pd models to biomarkers or observed in vivo pharmacodynamics are alternative approaches to evaluate human cns pbpk predictions. however, the success of these approaches depends on selection of appropriate biomarkers and the establishment of a reliable pk/pd relationship. thus, there is need for more ivive-pbpk models to fully explore and evaluate their applicability for the prediction of drug pharmacokinetics in the brain. ivive-pbpk model for transporter mediated drug-drug interaction traditional pharmacokinetic modeling lacks good predictability for complex drug-drug interactions (ddis) because in most cases, only in vitro and limited clinical data are available in early drug development. as ivive-pbpk enables scaling of early in vitro and animal data for human predictions, it can aid ddi risk assessment. the effect of transporter-enzyme interplay on drug bioavailability and hepatic disposition is well recognized [71, 72].there are examples of the involvement of transporter-enzyme interplay in ddis. lau et al reported that the inhibitory effect of rifampin on atorvastatin kinetics involves transporter-enzyme interplay [73]. rifampin is a potent inhibitor of organic anion-transporting polypeptide (oatp) transporters and atorvastatin is primarily metabolized by cyp3a in the liver, via its lactone group. in the study by lau et al [73], rifampin demonstrated inhibition of atorvastatin uptake into the liver and a change in the ratio of atorvastatin lactone to atorvastatin acid, thereby changing the amount of substrate available to the enzymes in the liver. grillo et al [74] demonstrated the use of ivive-pbpk modeling to predict potential clinically relevant ddi between rivaroxaban and a combination of p-gp and a moderate cyp3a4 inhibitor in renal impaired patients. thus, bottom-up pbpk models are a powerful tool for the prediction of transporter-mediated drug interactions, and accurate quantitation of protein concentrations in in vitro and in vivo systems will further enhance their accuracy. given that in vitro systems do not completely reproduce the transport rates and affinities like in vivo, the middle out parameter optimization approach is a necessary step in model verification. conclusions ivive requires information from numerous processes and organs. each has its advantages and limitations in translating data to the clinic. ivive-pbpk modeling is a powerful approach when full kinetic admet & dmpk 5(4) (2017) 201-211 utilizing in vitro transporter data in ivive-pbpk doi: 10.5599/admet.5.4.441 207 characterization of transporters is implemented. thus, understanding the influence of drug transporters on drug pharmacokinetics can be obtained through a combination of approaches. the future of drug transporter-based ivive-pbpk models lies in their potential to contribute to pharmacogenetic studies. drug transporters and metabolizing enzymes differ significantly based on various factors, including gender, age, race/ethnicity, genetics (polymorphisms), disease state, diet, and lifestyle. the impact of polymorphisms in drug metabolizing enzymes on pharmacokinetic variation is largely studied and known; however, knowledge regarding the impact of polymorphisms in drug transporters is limited. studies undertaken to date report significant changes in intracellular drug levels due to polymorphisms in drug transporters, which suggests potential alterations in the systemic pharmacokinetics of drugs. most drug transporters (e.g., p-gp and mrp-1) follow a developmental ontogeny and like most drug metabolizing enzymes, the activities of membrane transporters are immature in neonates. thus, future research is needed to drive system-dependent scaling factors for predicting transporter contributions of new molecular entities, in different disease states and among special populations, using ivive-pbpk modeling approaches. conflict of interest: none references [1] r.p. oude elferink, d.k. meijer, f. kuipers, p.l. jansen, a.k. groen, g.m. groothuis. hepatobiliary secretion of organic compounds; molecular mechanisms of membrane transport. biochim biophys acta 1241 (1995) 215-268. [2] l. zhang, c.m. brett, k.m. giacomini. role of organic cation transporters in drug absorption and elimination. annu rev pharmacol toxicol 38 (1998) 431-460. [3] h. kusuhara, y. sugiyama. role of transporters in the tissue-selective distribution and elimination of drugs: transporters in the liver, small intestine, brain and kidney. j control release 78 (2002) 43-54. [4] m. muller, p.l. jansen. molecular aspects of hepatobiliary transport. am j physiol 272 (1997) g1285-1303. [5] h. koepsell. organic cation transporters in intestine, kidney, liver, and brain. annu rev physiol 60 (1998) 243-266. [6] d.k. meijer, g.j. hooiveld, a.h. schinkel, j.e. van montfoort, m. haas, d. de zeeuw, f. moolenaar, j.w. smit, p.j. meier. transport mechanisms for cationic drugs and proteins in kidney, liver and intestine: implication for drug interactions and cell-specific drug delivery. nephrol dial transplant 14 suppl 4 (1999) 1-3. [7] h. suzuki, y. sugiyama. transporters for bile acids and organic anions. pharm biotechnol 12 (1999) 387-439. [8] d. lautier, y. canitrot, r.g. deeley, s.p. cole. multidrug resistance mediated by the multidrug resistance protein (mrp) gene. biochem pharmacol 52 (1996) 967-977. [9] g.j. hooiveld, j.e. van montfoort, d.k. meijer, m. muller. function and regulation of atp-binding cassette transport proteins involved in hepatobiliary transport. eur j pharm sci 12 (2001) 525-543. [10] g. lee, s. dallas, m. hong, r. bendayan. drug transporters in the central nervous system: brain barriers and brain parenchyma considerations. pharmacol rev 53 (2001) 569-596. [11] n. mizuno, t. niwa, y. yotsumoto, y. sugiyama. impact of drug transporter studies on drug discovery and development. pharmacol rev 55 (2003) 425-461. [12] m.v. varma, c.m. ambler, m. ullah, c.j. rotter, h. sun, j. litchfield, k.s. fenner, a.f. el-kattan. targeting intestinal transporters for optimizing oral drug absorption. curr drug metab 11 (2010) 730-742. [13] j.d. schuetz, p.w. swaan, d.j. tweedie. the role of transporters in toxicity and disease. drug metab dispos 42 (2014) 541-545. pankajini mallick admet & dmpk 5(4) (2017) 201-211 208 [14] t.d. bjornsson, j.t. callaghan, h.j. einolf, v. fischer, l. gan, s. grimm, j. kao, s.p. king, g. miwa, l. ni, g. kumar, j. mcleod, r.s. obach, s. roberts, a. roe, a. shah, f. snikeris, j.t. sullivan, d. tweedie, j.m. vega, j. walsh, s.a. wrighton, r. pharmaceutical, g. manufacturers of america drug metabolism/clinical pharmacology technical working, f.d.a.c.f.d. evaluation, research. the conduct of in vitro and in vivo drug-drug interaction studies: a pharmaceutical research and manufacturers of america (phrma) perspective. drug metab dispos 31 (2003) 815-832. [15] x. qiu, h. zhang, y. lai. quantitative targeted proteomics for membrane transporter proteins: method and application. aaps j 16 (2014) 714-726. [16] l.g. zhang d., ding x., lu c. preclinical experimental models of drug metabolism and disposition in drug discovery and development. acta pharmaceutica sinica b, 2012, pp. 549–561. [17] k. horie, f. tang, r.t. borchardt. isolation and characterization of caco-2 subclones expressing high levels of multidrug resistance protein efflux transporter. pharm res 20 (2003) 161-168. [18] m.d. troutman, d.r. thakker. novel experimental parameters to quantify the modulation of absorptive and secretory transport of compounds by p-glycoprotein in cell culture models of intestinal epithelium. pharm res 20 (2003) 1210-1224. [19] a. balakrishnan, n. hussainzada, p. gonzalez, m. bermejo, p.w. swaan, j.e. polli. bias in estimation of transporter kinetic parameters from overexpression systems: interplay of transporter expression level and substrate affinity. j pharmacol exp ther 320 (2007) 133-144. [20] y. shirasaka, t. sakane, s. yamashita. effect of p-glycoprotein expression levels on the concentration-dependent permeability of drugs to the cell membrane. j pharm sci 97 (2008) 553565. [21] k.r. korzekwa, s. nagar, j. tucker, e.a. weiskircher, s. bhoopathy, i.j. hidalgo. models to predict unbound intracellular drug concentrations in the presence of transporters. drug metab dispos 40 (2012) 865-876. [22] d.a. smith, l. di, e.h. kerns. the effect of plasma protein binding on in vivo efficacy: misconceptions in drug discovery. nat rev drug discov 9 (2010) 929-939. [23] t. yoshikado, k. toshimoto, t. nakada, k. ikejiri, h. kusuhara, k. maeda, y. sugiyama. comparison of methods for estimating unbound intracellular-to-medium concentration ratios in rat and human hepatocytes using statins. drug metab dispos 45 (2017) 779-789. [24] m.k. degorter, r.b. kim. use of transgenic and knockout mouse models to assess solute carrier transporter function. clin pharmacol ther 89 (2011) 612-616. [25] a.h. schinkel, j.j. smit, o. van tellingen, j.h. beijnen, e. wagenaar, l. van deemter, c.a. mol, m.a. van der valk, e.c. robanus-maandag, h.p. te riele, et al. disruption of the mouse mdr1a pglycoprotein gene leads to a deficiency in the blood-brain barrier and to increased sensitivity to drugs. cell 77 (1994) 491-502. [26] c.b. washington, h.r. wiltshire, m. man, t. moy, s.r. harris, e. worth, p. weigl, z. liang, d. hall, l. marriott, t.f. blaschke. the disposition of saquinavir in normal and p-glycoprotein deficient mice, rats, and in cultured cells. drug metab dispos 28 (2000) 1058-1062. [27] a.h. schinkel, u. mayer, e. wagenaar, c.a. mol, l. van deemter, j.j. smit, m.a. van der valk, a.c. voordouw, h. spits, o. van tellingen, j.m. zijlmans, w.e. fibbe, p. borst. normal viability and altered pharmacokinetics in mice lacking mdr1-type (drug-transporting) p-glycoproteins. proc natl acad sci u s a 94 (1997) 4028-4033. [28] j.w. polli, j.l. jarrett, s.d. studenberg, j.e. humphreys, s.w. dennis, k.r. brouwer, j.l. woolley. role of p-glycoprotein on the cns disposition of amprenavir (141w94), an hiv protease inhibitor. pharm res 16 (1999) 1206-1212. [29] y. zhang, c. bachmeier, d.w. miller. in vitro and in vivo models for assessing drug efflux transporter activity. adv drug deliv rev 55 (2003) 31-51. [30] m. yamazaki, w.e. neway, t. ohe, i. chen, j.f. rowe, j.h. hochman, m. chiba, j.h. lin. in vitro substrate identification studies for p-glycoprotein-mediated transport: species difference and predictability of in vivo results. j pharmacol exp ther 296 (2001) 723-735. admet & dmpk 5(4) (2017) 201-211 utilizing in vitro transporter data in ivive-pbpk doi: 10.5599/admet.5.4.441 209 [31] j.b. houston, a. galetin. methods for predicting in vivo pharmacokinetics using data from in vitro assays. curr drug metab 9 (2008) 940-951. [32] j.o. miners, p.i. mackenzie, k.m. knights. the prediction of drug-glucuronidation parameters in humans: udp-glucuronosyltransferase enzyme-selective substrate and inhibitor probes for reaction phenotyping and in vitro-in vivo extrapolation of drug clearance and drug-drug interaction potential. drug metab rev 42 (2010) 196-208. [33] h.m. jones, h.a. barton, y. lai, y.a. bi, e. kimoto, s. kempshall, s.c. tate, a. el-kattan, j.b. houston, a. galetin, k.s. fenner. mechanistic pharmacokinetic modeling for the prediction of transportermediated disposition in humans from sandwich culture human hepatocyte data. drug metab dispos 40 (2012) 1007-1017. [34] m.j. zamek-gliszczynski, c.a. lee, a. poirier, j. bentz, x. chu, h. ellens, t. ishikawa, m. jamei, j.c. kalvass, s. nagar, k.s. pang, k. korzekwa, p.w. swaan, m.e. taub, p. zhao, a. galetin, c. international transporter. itc recommendations for transporter kinetic parameter estimation and translational modeling of transport-mediated pk and ddis in humans. clin pharmacol ther 94 (2013) 64-79. [35] s. bosgra, e. van de steeg, m.l. vlaming, k.c. verhoeckx, m.t. huisman, m. verwei, h.m. wortelboer. predicting carrier-mediated hepatic disposition of rosuvastatin in man by scaling from individual transfected cell-lines in vitro using absolute transporter protein quantification and pbpk modeling. eur j pharm sci 65 (2014) 156-166. [36] s. neuhoff, k.r. yeo, z. barter, m. jamei, d.b. turner, a. rostami-hodjegan. application of permeability-limited physiologically-based pharmacokinetic models: part ii prediction of pglycoprotein mediated drug-drug interactions with digoxin. j pharm sci 102 (2013) 3161-3173. [37] a. rostami-hodjegan, g.t. tucker. simulation and prediction of in vivo drug metabolism in human populations from in vitro data. nat rev drug discov 6 (2007) 140-148. [38] s. neuhoff, k.r. yeo, z. barter, m. jamei, d.b. turner, a. rostami-hodjegan. application of permeability-limited physiologically-based pharmacokinetic models: part i-digoxin pharmacokinetics incorporating p-glycoprotein-mediated efflux. j pharm sci 102 (2013) 3145-3160. [39] j.m. dietschy. mechanisms for the intestinal absorption of bile acids. j lipid res 9 (1968) 297-309. [40] s.n. marcus, c.d. schteingart, m.l. marquez, a.f. hofmann, y. xia, j.h. steinbach, h.t. ton-nu, j. lillienau, m.a. angellotti, a. schmassmann. active absorption of conjugated bile acids in vivo. kinetic parameters and molecular specificity of the ileal transport system in the rat. gastroenterology 100 (1991) 212-221. [41] p.a. dawson, t. lan, a. rao. bile acid transporters. j lipid res 50 (2009) 2340-2357. [42] r.e. morgan, m. trauner, c.j. van staden, p.h. lee, b. ramachandran, m. eschenberg, c.a. afshari, c.w. qualls, jr., r. lightfoot-dunn, h.k. hamadeh. interference with bile salt export pump function is a susceptibility factor for human liver injury in drug development. toxicol sci 118 (2010) 485-500. [43] s. dawson, s. stahl, n. paul, j. barber, j.g. kenna. in vitro inhibition of the bile salt export pump correlates with risk of cholestatic drug-induced liver injury in humans. drug metab dispos 40 (2012) 130-138. [44] k. kock, b.c. ferslew, i. netterberg, k. yang, t.j. urban, p.w. swaan, p.w. stewart, k.l. brouwer. risk factors for development of cholestatic drug-induced liver injury: inhibition of hepatic basolateral bile acid transporters multidrug resistance-associated proteins 3 and 4. drug metab dispos 42 (2014) 665-674. [45] j.m. pedersen, p. matsson, c.a. bergstrom, j. hoogstraate, a. noren, e.l. lecluyse, p. artursson. early identification of clinically relevant drug interactions with the human bile salt export pump (bsep/abcb11). toxicol sci 136 (2013) 328-343. [46] k. yang, k. kock, a. sedykh, a. tropsha, k.l. brouwer. an updated review on drug-induced cholestasis: mechanisms and investigation of physicochemical properties and pharmacokinetic parameters. j pharm sci 102 (2013) 3037-3057. pankajini mallick admet & dmpk 5(4) (2017) 201-211 210 [47] a.d. rodrigues, y. lai, m.e. cvijic, l.l. elkin, t. zvyaga, m.g. soars. drug-induced perturbations of the bile acid pool, cholestasis, and hepatotoxicity: mechanistic considerations beyond the direct inhibition of the bile salt export pump. drug metab dispos 42 (2014) 566-574. [48] p.b. watkins. idiosyncratic liver injury: challenges and approaches. toxicol pathol 33 (2005) 1-5. [49] e.m. leslie, p.b. watkins, r.b. kim, k.l. brouwer. differential inhibition of rat and human na+dependent taurocholate cotransporting polypeptide (ntcp/slc10a1)by bosentan: a mechanism for species differences in hepatotoxicity. j pharmacol exp ther 321 (2007) 1170-1178. [50] j.l. woodhead, k. yang, s.q. siler, p.b. watkins, k.l. brouwer, h.a. barton, b.a. howell. exploring bsep inhibition-mediated toxicity with a mechanistic model of drug-induced liver injury. front pharmacol 5 (2014) 240. [51] k. yang, j.l. woodhead, p.b. watkins, b.a. howell, k.l. brouwer. systems pharmacology modeling predicts delayed presentation and species differences in bile acid-mediated troglitazone hepatotoxicity. clin pharmacol ther 96 (2014) 589-598. [52] x. cao, l.x. yu, c. barbaciru, c.p. landowski, h.c. shin, s. gibbs, h.a. miller, g.l. amidon, d. sun. permeability dominates in vivo intestinal absorption of p-gp substrate with high solubility and high permeability. mol pharm 2 (2005) 329-340. [53] c.l. cummins, l. salphati, m.j. reid, l.z. benet. in vivo modulation of intestinal cyp3a metabolism by p-glycoprotein: studies using the rat single-pass intestinal perfusion model. j pharmacol exp ther 305 (2003) 306-314. [54] k.s. pang. modeling of intestinal drug absorption: roles of transporters and metabolic enzymes (for the gillette review series). drug metab dispos 31 (2003) 1507-1519. [55] m. jamei, d. turner, j. yang, s. neuhoff, s. polak, a. rostami-hodjegan, g. tucker. population-based mechanistic prediction of oral drug absorption. aaps j 11 (2009) 225-237. [56] a.s. darwich, s. neuhoff, m. jamei, a. rostami-hodjegan. interplay of metabolism and transport in determining oral drug absorption and gut wall metabolism: a simulation assessment using the "advanced dissolution, absorption, metabolism (adam)" model. curr drug metab 11 (2010) 716729. [57] a. bruyere, x. decleves, f. bouzom, k. ball, c. marques, x. treton, m. pocard, p. valleur, y. bouhnik, y. panis, j.m. scherrmann, s. mouly. effect of variations in the amounts of p-glycoprotein (abcb1), bcrp (abcg2) and cyp3a4 along the human small intestine on pbpk models for predicting intestinal first pass. mol pharm 7 (2010) 1596-1607. [58] a. olivares-morales, a. ghosh, l. aarons, a. rostami-hodjegan. development of a novel simplified pbpk absorption model to explain the higher relative bioavailability of the oros(r) formulation of oxybutynin. aaps j 18 (2016) 1532-1549. [59] m.s. alavijeh, m. chishty, m.z. qaiser, a.m. palmer. drug metabolism and pharmacokinetics, the blood-brain barrier, and central nervous system drug discovery. neurorx 2 (2005) 554-571. [60] m. feher, e. sourial, j.m. schmidt. a simple model for the prediction of blood-brain partitioning. int j pharm 201 (2000) 239-247. [61] s. vilar, m. chakrabarti, s. costanzi. prediction of passive blood-brain partitioning: straightforward and effective classification models based on in silico derived physicochemical descriptors. j mol graph model 28 (2010) 899-903. [62] j.a. platts, m.h. abraham, y.h. zhao, a. hersey, l. ijaz, d. butina. correlation and prediction of a large blood-brain distribution data set--an lfer study. eur j med chem 36 (2001) 719-730. [63] n.j. abbott, a.a. patabendige, d.e. dolman, s.r. yusof, d.j. begley. structure and function of the blood-brain barrier. neurobiol dis 37 (2010) 13-25. [64] y. yamamoto, p.a. valitalo, d.j. van den berg, r. hartman, w. van den brink, y.c. wong, d.r. huntjens, j.h. proost, a. vermeulen, w. krauwinkel, s. bakshi, v. aranzana-climent, s. marchand, c. dahyot-fizelier, w. couet, m. danhof, j.g. van hasselt, e.c. de lange. a generic multicompartmental cns distribution model structure for 9 drugs allows prediction of human brain target site concentrations. pharm res 34 (2017) 333-351. admet & dmpk 5(4) (2017) 201-211 utilizing in vitro transporter data in ivive-pbpk doi: 10.5599/admet.5.4.441 211 [65] a.m. palmer, m.s. alavijeh. translational cns medicines research. drug discov today 17 (2012) 1068-1078. [66] a.k. deo, f.p. theil, j.m. nicolas. confounding parameters in preclinical assessment of blood-brain barrier permeation: an overview with emphasis on species differences and effect of disease states. mol pharm 10 (2013) 1581-1595. [67] f. fenneteau, p. poulin, f. nekka. physiologically based predictions of the impact of inhibition of intestinal and hepatic metabolism on human pharmacokinetics of cyp3a substrates. j pharm sci 99 (2010) 486-514. [68] k. ball, f. bouzom, j.m. scherrmann, b. walther, x. decleves. physiologically based pharmacokinetic modelling of drug penetration across the blood-brain barrier--towards a mechanistic ivive-based approach. aaps j 15 (2013) 913-932. [69] k. ball, f. bouzom, j.m. scherrmann, b. walther, x. decleves. development of a physiologically based pharmacokinetic model for the rat central nervous system and determination of an in vitro-in vivo scaling methodology for the blood-brain barrier permeability of two transporter substrates, morphine and oxycodone. j pharm sci 101 (2012) 4277-4292. [70] n. notkina, c. dahyot-fizelier, a.k. gupta. in vivo microdialysis in pharmacological studies of antibacterial agents in the brain. br j anaesth 109 (2012) 155-160. [71] w.c.-y. benet lz, hebert mf, wacher vj. intestinal drug metabolism and antitransport processes: a potential paradigm shift in oral drug delivery. journal of controlled release 2 (1996) 139–143. [72] v.j. wacher, l. salphati, l.z. benet. active secretion and enterocytic drug metabolism barriers to drug absorption. adv drug deliv rev 46 (2001) 89-102. [73] y.y. lau, y. huang, l. frassetto, l.z. benet. effect of oatp1b transporter inhibition on the pharmacokinetics of atorvastatin in healthy volunteers. clin pharmacol ther 81 (2007) 194-204. [74] j.a. grillo, p. zhao, j. bullock, b.p. booth, m. lu, k. robie-suh, e.g. berglund, k.s. pang, a. rahman, l. zhang, l.j. lesko, s.m. huang. utility of a physiologically-based pharmacokinetic (pbpk) modeling approach to quantitatively predict a complex drug-drug-disease interaction scenario for rivaroxaban during the drug review process: implications for clinical practice. biopharm drug dispos 33 (2012) 99-110. ©2017 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ subcellular localization of several structurally different tyrosine kinase inhibitors doi: 10.5599/admet.514 258 admet & dmpk 6(3) (2018) 258-266; doi: http://dx.doi.org/10.5599/admet.514 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific article subcellular localization of several structurally different tyrosine kinase inhibitors richard j. honeywell*, sarina m. hitzerd, g.a.m. kathmann, godefridus j. peters dept of medical oncology, vu university medical center, po box 7057, 1007 mb amsterdam, the netherlands *corresponding author: e-mail: r.honeywell@vumc.nl. received: march 06, 2018; revised: may 21, 2018; available online: july 16, 2018 abstract protein tyrosine kinases form an important target for a new class of anticancer drugs, the tyrosine kinase inhibitors (tkis). recently we demonstrated that sunitinib, an inhibitor of the membrane-associated vascular endothelial growth factor receptor (vegfr), is trapped in lysosomes which isolates the drug from its intended target. therefore we investigated whether this also holds for other tkis, targeted against different protein kinases. for this purpose we used the proteoextractr kit, which enables a subcellular extraction separating cellular proteins into four distinct fractions covering the cytosol, membranes and membrane organelles (including lysosomes), nuclear proteins and the cytoskeleton. since tkis are 98-100 % protein bound we used this property to study their subcellular distribution and used caco-2 cells as a model. as expected after 2 hours exposure sunitinib was trapped in cytosol (58 %) and organelles (42 % including lysosomes). crizotinib, an inhibitor of alk-eml4, showed a similar distribution. however, erlotinib, an inhibitor of the epidermal growth factor receptor (egfr) showed a very low cellular accumulation and was limited to the organelle fraction. in contrast, the other egfr inhibitor, gefitinib was predominantly located in the cytosolic (39 %) and membrane fraction (44 %). sorafenib, another vegfr inhibitor was predominantly located in the organelle fraction (85 %) and cytosol (15 %) after 2 hours, while after 24 hours distribution decreased (9.9 fold) with a slight shift. dasatinib, an inhibitor of bcr-abl was located only in the cytosol (100 %). in general localization after 24 hours was comparable, albeit several small changes were seen. in conclusion protein fractionation with the proteoextractr subcellular proteome extraction kit demonstrated large differences in tki levels in various cellular organelles, with a pattern in agreement with lysosomal accumulation of sunitinib. keywords tyrosine kinase inhibitors; lysosomal accumulation; subcellular distribution; sunitinib; erlotinib; dasatinib introduction protein tyrosine kinases (ptks) regulate important signaling processes specific for the cancer cell [1]. because of this important role a new class of drugs, the tyrosine kinase inhibitors (tkis) have been developed [2,3]. the targets for these drugs are receptors located either on or in the membrane or intracellularly. this holds both for tumor cells and tumor-associated endothelial cells [4]. a decreased uptake or increased efflux may limit the cellular drug accumulation, and the efficacy of these drugs [5]. moreover, inhibition may be reversible or irreversible; especially novel third or fourth generation tkis display irreversible enzyme inhibition [6]. recently we demonstrated that sunitinib, an inhibitor of the http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:r.honeywell@vumc.nl admet & dmpk 6(3) (2018) 258-266 subcellular localization of several structurally different tkis doi: 10.5599/admet.514 259 vascular endothelial growth factor receptor (vegfr) accumulates in lysosomes [7], while preliminary experiments demonstrated a similar distribution for crizotinib, an alk-eml4 and cmet inhibitor [5]. in contrast, erlotinib an inhibitor of the epidermal growth factor receptor (egfr) showed a very low cellular uptake limited to the membrane [8]. therefore, we wondered whether drug accumulation or drug distribution would be related to their efficacy. since tkis are 98-100 % plasma protein bound [9,10] we also wondered whether this is reflected cellularly and whether this would affect the distribution and accumulation of a tki. for this purpose, we used methodology, which is commonly applied in the field of proteomics, in which the isolation and analysis of proteins based on their cellular location is a rapidly expanding field. however, the different proteins that make up a cellular environment differ greatly and are difficult to isolate due to size, hydrophobicity and basicity. using the differential solubility of different proteins in the various subcellular compartments has led to the separation of four distinct fractions, the cytosolic fraction, the membrane and organelle fraction, cellular nucleus, and the cytoskeleton, since we used this approach to successfully characterize the cellular distribution of a lipophilic prodrug of cytarabine [11] we applied the same methodology to determine the cellular distribution of tkis. since tkis are almost completely bound to proteins, a cellular fractionation of cells incubated with drug for a suitable time period should also provide fractionation of the accumulated tki. simple extraction of each fraction should then provide data on the distribution of the tki within the cell. using a previously developed analytical liquid chromatography – mass spectrometry method [12] the exact proportion of tki can be determined in each fraction with a high degree of specificity. using this technique, we determined the subcellular distribution of six tkis (sunitinib, crizotinib, erlotinib, gefitinib, sorafenib and dasatinib) within the colon cancer cell line, caco-2. these six inhibitors were chosen based on their fda status and the extent to which they are used in the clinic. each of the selected molecules is the first line treatment in their specific area and regularly used for clinical patients. materials and methods cell culture the caco-2 cell line, originating from a colorectal adenocarcinoma, was cultured at 37 °c, 5 % co2 and 100 % humidity, in dulbecco’s modified eagle medium (dmem) supplemented with 10 % fetal bovine serum (fbs) and 20 mm hepes [13]. for the fractionation caco-2 cells were seeded in 6 ml dmem culture medium with 1 % penicillin/streptomycin in 25 cm 2 flasks. cells were allowed to attach and grow for 72 hours (aiming at an ultimate cell number of 3-5x10 6 cells). attached cells were than exposed to either 10 µm erlotinib, gefitinib, sorafenib, sunitinib, crizotinib or 1 µm dasatinib for 2 and 24 hours. next, medium was aspirated and cells were washed three times with 10 ml ice cold pbs. to each flask 500 µl of trypsin/edta was added and cells were incubated for about 5-10 minutes at 37 °c, 5 % co2 and 100 % humidity. when cells were detached, the cells were resuspended with 1.5 ml ice cold phosphate buffered saline and transferred into eppendorf tubes. after centrifugation, the supernatant was discarded and the pellets resuspended in 200 µl of water for fractionation. cellular fractionation the proteoextract®, subcellular proteome extraction kit from calbiochem (cat. no. 539791), was used to fractionate the cell pellets. the differential solubility of proteins in the various subcellular fractions was used to isolate these fractions by the addition of four sequential extraction buffers. each buffer specifically targets a subcellular compartment while maintaining structural integrity of the cell [13]. when correctly applied these fractions are distinct and can separate the different subcellular compartments with minimal richard honeywell et al. admet & dmpk 6(3) (2018) 258-266 260 to no mixing. cell suspensions were centrifuged at 100-300 g for 10 minutes at 4 °c, then the supernatant was removed. next, the cell pellets were washed with 2 ml ice cold washing buffer, gently resuspended, incubated for 5 minutes at 4 °c and centrifuged at 100-300 g at 4 °c for 10 minutes. the supernatant was discarded and the washing step was repeated once more. the resulting cell pellets were used to separate 4 distinct cellular compartments. the initial stage of the extraction releases the cytosolic proteins from the cell, by addition of 1 ml of ice cold extraction buffer i with 5 µl protease inhibitor cocktail (pic) to each cell pellet, which were incubated for 10 minutes at 4 °c and centrifuged at 500-1000 g at 4 °c for 10 minutes. the supernatant was then transferred quantitatively into a clean labelled tube without disturbing the pellet (fraction 1, representing the cytosolic proteins) and stored at -80 °c for future analysis. subsequently, 1 ml of ice cold extraction buffer ii with 5 µl pic was added to each pellet to solubilize the membrane and organelles fraction, incubated for 30 minutes at 4 °c and then centrifuged at 5000-6000 g at 4 °c for 10 minutes. the supernatant was transferred quantitatively into a clean tube without disturbing the pellet (fraction 2 representing the membrane and organelle proteins) and stored at -80°c for future analysis. next, 500 µl of ice cold extraction buffer iii with 5 µl pic and 1.5 µl benzonase nuclease was added to each pellet, incubated for 10 minutes at 4 °c and centrifuged at 6800 g at 4 °c for 10 minutes. the supernatant was transferred into a clean tube without disturbing the pellet, containing nuclear proteins (fraction 3) and stored at -80 °c for future analysis. the last stage was to add 500 µl of extraction buffer iv with 5 µl pic to each cell pellet, obtaining the components of the cytoskeleton (fraction 4) and stored at -80 °c for future analysis. data analysis by lc-ms/ms all samples of the subcellular fractionation were measured using liquid chromatography coupled to mass spectrometry to determine the amount of drug (tki) present using a previously developed method for biological materials. briefly, 20 µl of each supernatant fraction was transferred quantitatively into a round bottomed 96-well plate and 80 µl of ice cold acetonitrile added. each plate was covered with a plastic foil and allowed to stand for 10 minutes. subsequently the 96-well plate was centrifuged at 1500 g at 4 °c for 10 minutes. next, 50 µl of each sample was transferred to a clean conical 96-well plate and covered with a nunc well plate seal. lastly, 1 µl of each sample was injected into the lc-ms/ms under the following conditions. a mobile phase containing 0.578 g ammonium acetate, 375 ml milliq water, 125 ml methanol, 0.1 % isopropanol and 1000 ml acetonitrile with a ph of 7.8 was used at 350 µl/min. before use, the mobile phase was filtered using a nylon filter paper of 0.2 µm and subsequently degassed for 10 minutes. chromatography was performed on a phenomenex prodigy ods 3 column (100x2 mm, 3 µm). data analysis concentration of compounds determined by lc-ms/ms in each of the fractions was normalized to total volume of the collected fraction and quantified as accumulated drug within each fraction (fmol). protein content of each fraction was determined and data normalised to fmol/mg protein. the percentage of drug observed in each fraction was then determined in relation to the total drug accumulation (the sum total of all fractions per compound). fold changes between the 2 hour and 24 hour accumulation time points were determined as the ratio of total drug accumulated at the 2 hour mark to the total accumulated at the 24 hour time mark. admet & dmpk 6(3) (2018) 258-266 subcellular localization of several structurally different tkis doi: 10.5599/admet.514 261 results using the proteome extraction kit we managed to separate the proteins bound into representative fractions of the cellular compartments, even when accumulated at very low levels. sunitinib has previously been shown to accumulate within the lysosomal compartment of the cellular makeup, where it can reach very high concentrations. hence, sunitinib was used as a control for the occurrence of lysosomal accumulation. as expected after 2 hours sunitinib was predominantly located in the cytosol (57.5 %), membranes and membrane organelles (42.1 %) (e.g. lysosomes). after 24 hours no distribution change was observed, although the absolute amount of sunitinib decreased 3.9 fold (tables 1 and 2; figure 1). table 1. subcellular distribution of 6 tyrosine kinase inhibitors time (hours) sunitinib crizotinib erlotinib gefitinib sorafenib dasatinib cytosolic fraction (1) 2 57.5 58.3 10.1 38.8 15.3 100 24 56.6 41.7 0.0 56.3 0.0 55.6 membrane fraction (2) 2 42.1 41.7 7.1 44.4 84.7 0.0 24 40.2 56.4 0.0 30.0 60.4 27.9 organelles fraction (3) 2 0.4 0.0 3.6 16.8 0.0 0.0 24 2.7 2.0 0.0 11.0 13.9 0.0 cytoskeletal fraction (4) 2 0.0 0.0 79.2 0.0 0.0 0.0 24 0.5 0.0 100.0 2.8 25.7 16.5 values are percentages (%) of total cellular accumulation after 2 and 24 hr of analyses performed in duplicate table 2. time dependence of total cellular accumulation of 6 tkis drug total cellular accumulation (fmol/mg protein) 2 hours 24 hours sunitinib 25692.8 4218.4 crizotinib 5581.4 19867.6 erlotinib 314.9 88.4 gefitinib 1572.8 462.9 sorafenib 2056.2 267.1 dasatinib 244.4 97.1 values are means of analyses performed in duplicate. sem was below 15 % crizotinib showed a similar distribution pattern compared to sunitinib. after 2 hours it was highly distributed into the cytosol (58.3 %), membranes and membrane organelles (41.7 %). although the distribution remained consistent after 24 hours a 4.5-fold increase in accumulation was observed. one tki, which showed a totally different distribution pattern, was erlotinib (table 1). a distribution of erlotinib throughout the whole cell was observed after 2 hours exposure, although most of the accumulated erlotinib was located in cytoskeleton components. after 24 hours total accumulation decreased 20-fold and was located solely in the cytoskeleton. gefitinib, which also inhibits egfr showed a completely different pattern both after 2 and 24 hours exposure (table 1). after 2 hours gefitinib was located in the cytosol (38.8 %), membranes and membrane richard honeywell et al. admet & dmpk 6(3) (2018) 258-266 262 organelles (44.4 %) and nuclear proteins (16.8 %), but not in the cytoskeleton as erlotinib, which hardly changed after 24 hours. 5 7 . 5 0 % c y t o s o l e 4 2 . 1 0 % c e l l u l a r m e m b r a n e 2 h r s u n i t i n i b e x p o s u r e 0 . 4 0 % m i t o c h o n d r i a a n d n u c l e u s 2 4 s u n i t i n i b h o u r e x p o s u r e 4 0 . 2 0 % c e l l u l a r m e m b r a n e 2 . 7 0 % m i t o c h o n d r i a a n d n u c l e u s 5 6 . 6 0 % c y t o s o l e 1 0 0 . 0 0 % c y t o s o l e 2 h r d a s a t i n i b e x p o s u r e 2 4 h o u r d a s a t i n i b e x p o s u r e 2 7 . 9 0 % c e l l u l a r m e m b r a n e 1 6 . 5 0 % c y t o s k e l a t a l 5 5 . 6 0 % c y t o s o l e figure 1. pie chart of the cellular distribution of sunitinib and dasatinib after 2 and 24 hours exposure to 10 μm of either sunitinib or dasatinib respectively both sorafenib and dasatinib do not correspond with the other tkis and each other’s distribution pattern (table 1; figure 1). after two hours exposure, sorafenib is solely distributed in the membranes, membrane organelles (84.7 %) and cytosol (15.3 %). after 24 hours accumulation decreased 9.9-fold (table 2) and sorafenib distribution shifted to membranes and membrane organelles (60.4%), cytoskeleton (25.7 %) and nuclear proteins (13.9 %). accumulated dasatinib was located solely in the cytosol after two hours exposure (figure 1; table 1). admet & dmpk 6(3) (2018) 258-266 subcellular localization of several structurally different tkis doi: 10.5599/admet.514 263 similar to sorafenib, the distribution of dasatinib changed completely after 24 hours. some of the accumulated dasatinib shifted from the cytosol to membranes and membrane organelles (27.9 %) and cytoskeleton (16.5 %). there was only a small change in total accumulated drug. interestingly the total accumulation of the drugs was quite different both in the amount and the timecourse (table 2). the absolute amount of sunitinib accumulated in the cells decreased by 3.9-fold from 2 to 24 hours which is in agreement with the decline of sunitinib accumulation in the temperature dependent accumulation study after 24 hours [14]. in contrast to sunitinib, crizotinib demonstrated a 4.5-fold increase after 24 hours. accumulation of erlotinib was very low and even decreased after 24 hours to 5.3 % of that after 2 hours. also, total sorafenib accumulation decreased considerably from 2 to 24 hours. in contrast to the five other tkis, the total absolute amount of dasatinib accumulation in the cell hardly changed after 24 hours compared to 2 hours. discussion the data show a large difference in both the cellular uptake and distribution. structurally the tkis investigated differ significantly (figure 2) with a smile correlations of less than 0.4 with a median of 0.206538 (range 0.154453 – 0.389155) as calculated by chemmine, table 3 [15]. no relationship could be determined between the cellular localization of the molecules to their structural makeup. the nature of the accumulation of these compounds is highly dependent on the degree to which they are lysosomally accumulated, our data demonstrates that sunitinib and crizotinib are in agreement with the physical properties of lysosomal accumulation. for sunitinib, the lysosomal accumulation has also been demonstrated by using its fluorescent properties which co-localized with the lysotracker, a marker for lysosomes [7]. moreover, pre-incubation with bafilomycin-a prevented trapping of sunitinib in the lysosomes [7,16]. sunitinib achieves very high cellular concentrations since it has been shown that 90% of accumulated drug is held within lysosomes within the cell structure. the cell survives this accumulation since the bulk of the drug is isolated away from its targets. the relationship between physiochemical properties and the observed absorption characteristics is currently under investigation and out of the scope of this paper. table 3. comparison of molecular structures for crizotinib, dasatinib, erlotinib, gefitinib, sorafenib and sunitinib using chemmine correlation of smiles. crizotinib dasatinib erlotinib gefitinib sorafenib sunitinib crizotinib 0.231458 0.196302 0.285714 0.156522 0.203147 dasatinib 0.231458 0.208279 0.296345 0.154453 0.275956 erlotinib 0.196302 0.208279 0.389155 0.17601 0.206538 gefitinib 0.285714 0.296345 0.389155 0.18059 0.262319 sorafenib 0.156522 0.154453 0.17601 0.18059 0.191546 sunitinib 0.203147 0.275956 0.206538 0.262319 0.191546 richard honeywell et al. admet & dmpk 6(3) (2018) 258-266 264 cl hn os n n h nn n n ho dasatinib exact mass: 487.16 molecular weight: 488.01 protein binding: 96% gibbs energy: 836.07 [kj/mol] log p: 5.06 n nh2 o cl cl fn n hn log p: 3.39 crizotinib exact mass: 449.12 molecular weight: 450.34 protein binding: <90% hn n n o o o o erlotinib exact mass: 393.17 molecular weight: 393.44 protein binding: ~95% gibbs energy: 473.87 [kj/mol] log p: 3.3 f cl hn n n o o n o gefitinib exact mass: 446.15 molecular weight: 446.9 protein binding: ~90% gibbs energy: 313.35 [kj/mol] log p: 3.87 o nh n h f o nh n sunitinib exact mass: 398.21 molecular weight: 398.47 hn o n o n h o n h cl f f f sorafenib exact mass: 464.09 molecular weight: 464.82 log p: 2.03 gibbs energy: -225.59 [kj/mol] log p: 3.76 figure 2. the molecular structures and know physical chemical properties for the six tyrosine kinase inhibitors under investigation also, for crizotinib we demonstrated that bafilomycin prevented uptake into lysosomes (unpublished data). trapping of these drugs is possibly mediated by a selective uptake of the drugs via multidrug resistance proteins, subsequently the drugs are protonated and cannot subsequently be effluxed anymore [17,18]. for erlotinib it is likely that it is transported into the membrane, binds to its target, but is effluxed before it can enter the cell [19]. in contrast, the other egfr targeted drug (gefitinib) is taken up by an active transporter [13,20,21] and subsequently trapped in the cell. sorafenib is not only actively being transported into the cell but is possibly also trapped in the cell, either as the parent drug or as a metabolite, since it is substrate for various phase i and phase ii enzymes [22,23]. it seems to be effluxed efficiently from the cell. lastly, dasatinib is taken up by the cell, but is not further distributed and is not effluxed efficiently either. however, its cytosolic localization is in line with its high efficacy since its target bcr-abl (only in chronic myeloid leukemia) is located in the cytosol [1]. admet & dmpk 6(3) (2018) 258-266 subcellular localization of several structurally different tkis doi: 10.5599/admet.514 265 conclusion protein fractions contained considerable tki levels that changed markedly over time. this could be demonstrated by using the proteoextractr subcellular proteome extraction kit. in addition, it can be concluded that the distribution of the six tkis differs a lot and also changed over time. this may be related to the different physio-chemical properties of the various tkis, favouring either a lysosomal accumulation or a rapid efflux from the cell [10]. nevertheless, sunitinib and crizotinib distribution show comparable patterns, probably due to the lysosomal accumulation. references [1] f. broekman, e. giovannetti, g.j. peters. tyrosine kinase inhibitors: multi-targeted or singletargeted? world j. clin. oncol. 2 (2011) 80–93. [2] r. chebib, l. verlingue, n. cozic, m. faron, p. burtin, v. boige, a. hollebecque, d. malka. angiogenesis inhibition in the second-line treatment of metastatic colorectal cancer: a systematic review and pooled analysis. seminars in oncology 44 (2017) 114–128. [3] p. wu, t.e. nielsen, m.h. clausen. fda-approved small-molecule kinase inhibitors. trends in pharmacological sciences 36 (2015) 422–439. [4] m.p. pinto, g. i. owen, i. retamal, m. garrido. angiogenesis inhibitors in early development for gastric cancer. expert opinion on investigational drugs 26 (2017) 1007–1017. [5] c.g. da silva, r. j. honeywell, h. dekker, g. j. peters. physicochemical properties of novel protein kinase inhibitors in relation to their substrate specificity for drug transporters. expert opin. drug metab. toxicol. 11 (2015) 703-717. [6] n. van der steen, c. caparello, c. rolfo, p. pauwels, g.j. peters, e. giovannetti. new developments in the management of non-small-cell lung cancer, focus on rociletinib: what went wrong? oncotargets and therapy 9 (2016) 6065–6074. [7] k.j. gotink, h.j. broxterman, m. labots, r.r. de haas, h. dekker, r.j. honeywell, m.a. rudek, l.v. beerepoot, r.j. musetrs, g. jansen, a.w. griffioen,y.g. assaraf, r. pili, g.j. peters, h.m.w. verheul. lysosomal sequestration of sunitinib: a novel mechanism of drug resistance. clin. cancer res. 17 (2011) 7337-7346. [8] e. giovannetti, m. labots, h. dekker, e. galvani, j.s.w. lind, r. sciarillo, r. honeywell, e.f. smit, h.m.w. verheul, g.j. peters. molecular mechanisms and modulation of key pathways underlying the synergistic interaction of sorafenib with erlotinib in non-small-cell-lung cancer (nsclc) cells | benthamscience. curr. pharm. des. (2013) 927–939. [9] n.p. van erp, h. gelderblom, h.-j. guchelaar. clinical pharmacokinetics of tyrosine kinase inhibitors. cancer treat rev 35 (2009) 692–706. [10] k. valko. bio-mimetic chromatography to predict drug distribution in vivo. european pharmaceutical review (2010). [11] a.d. adema, k. smid, n. losekoot, r. honeywell, h.m.w. verheul, f. mihren, m.l. sandvold, g.j. peters. metabolism and accumulation of the lipophilic deoxynucleoside analogs elacytarabine and cp-4126. invest. new drugs 30 (2012) 1908–1916. [12] r.j. honeywell, d. sarkisjan, i. kathmann, m.h. kristensen, g.j. peters. sensitive liquid chromatography mass spectrometry (lc-ms) assay reveals novel insights on dna methylation and incorporation of gemcitabine, its metabolite difluorodeoxyuridine, deoxyuridine, and rx-3117 into dna. nucleosides, nucleotides and nucleic acids 35 (2016) 652-662. [13] r. honeywell, c. fatmawati, m. buddha, s. hitzerd, i. kathman, g.j. peters. adaptation of a human gut epithelial model in relation to the assessment of clinical pharmacokinetic parameters for selected tyrosine kinase inhibitors. admet dmpk 3 (2015) 51–67. [14] t. rabilloud, c. adessi, a. giraudel, j. lunardi. improvement of the solubilization of proteins in twodimensional electrophoresis with immobilized ph gradients. electrophoresis 18 (1997) 307–316. richard honeywell et al. admet & dmpk 6(3) (2018) 258-266 266 [15] t.w.h. backman, y. cao, t. girke. chemmine tools: an online service for analyzing and clustering small molecules. nucleic acids res. 39 (2011) w486–w491. [16] k. azijli, k.j. gotink, h.m. verheul. the potential role of lysosomal sequestration in sunitinib resistance of renal cell cancer. j. kidney cancer vhl 2 (2016) 195-203. [17] b. zhitomirsky, y.g. assaraf. lysosomes as mediators of drug resistance in cancer. drug resist. updat. 24 (2016) 23–33. [18] b. zhitomirsky, y.g. assaraf. lysosomal sequestration of hydrophobic weak base chemotherapeutics triggers lysosomal biogenesis and lysosome-dependent cancer multidrug resistance. oncotarget 6 (2015) 1143–1156. [19] r.j. honeywell, s. hitzerd, i. kathmann, g. j. peters. transport of six tyrosine kinase inhibitors: active or passive? admet dmpk 4 (2016) 23-34. [20] c. neul, e. schaeffeler, a. sparreboom, s. laufer, m. schwab, a.t. nies. impact of membrane drug transporters on resistance to small-molecule tyrosine kinase inhibitors. trends in pharmacological sciences 37 (2016) 904–932. [21] m. galetti, p. petronini, c. fumarola, d. cretella, s. la monica, m. bonelli, a. cavazzoni, f. saccani, c. caffarra, r. andreoli, a. mutti, m. tiseo, a. ardizzoni, r.r. alfieri. effect of abcg2/bcrp expression on efflux and uptake of gefitinib in nsclc cell lines. plos one 10 (2015) e0141795. [22] e.i. zimmerman, s. hu, j.l. roberts, a.a. gibbson, s.j. orwick, l. li, a. sparreboom, s.d. baker. contribution of oatp1b1 and oatp1b3 to the disposition of sorafenib and sorafenib-glucuronide. clin. cancer res. 19 (2013) 1458–66. [23] s.d. baker, s. hu. pharmacokinetic considerations for new targeted therapies. clin. pharmacol. ther. 85 (2009) 208–11. ©2018 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ manuscript doi: 10.5599/admet.3.3.194 216 admet & dmpk 3(3) (2015) 216-234; doi: 10.5599/admet.3.3.194 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index review chemical and physical approaches for the treatment of alzheimer's disease jing li, wei zhou, xiaoli wu and kin tam* faculty of health sciences, university of macau, macau, china *corresponding author: e-mail: kintam@umac.mo; tel.: +853-8822-4988; fax: +853-8822-2314 received: june 18, 2015; revised: august 20, 2015; published: september 05, 2015 abstract alzheimer’s disease (ad) is a destructive neurodegenerative disorder, which threatens elderly people in their mental health irreversibly, causing cognitive deficit such as learning difficulties and memory loss. neurobiological pathogenesis leading to the neuronal dysfunction in ad are mainly rationalized by two hypotheses: 1) extracellular aggregation of the β-amyloid (aβ) peptide, which impacts on the normal neuronal pathways through excessive oxidative stress and damages neurons; and 2) intracellular hyperphosphorylation of tau protein, which forms the paired helical filaments, then gravely impacts axonal transport. in this review, we firstly introduce the neurobiological concept of ad, and its pathogenesis. then we review the chemical approaches pointedly directing at aβ peptide and tau protein, and the clinical reality of these pharmacological strategies. thirdly, we discuss a physical approach, deep brain stimulation (dbs), which is receiving considerable attention in recent years due to its success in treating parkinson disease clinically. dbs delivers current pulses, which are generated by a implanted pacemaker, through electrodes into dysfunctional brain structures to influence neural activities with the possibility to improve the cognitive function. we will summarize the clinical applications of dbs to restore cognitive impairments due to ad, and animal studies related with dbs. finally, we will discuss the current study obstacles and future research development of dbs in preclinical and clinical investigations of ad. keywords alzheimer’s disease; neurobiological pathogenesis; β-amyloid peptide; tau protein; pharmacological strategies; deep brain stimulation 1. introduction a growing number of people suffer from ad due to the aggravation of social aging and the dilation of average life expectancy, which brings about heavy burden to the families, the society and the whole world [1]. therefore numerous researchers and clinicians have spent considerable effort to find a cure. the initial histopathologic hallmarks of ad are the presence of intracellular neurofibrillary tangles (nfts) and extracellular depositional β-amyloid (aβ) plaques [2]. other histopathologic distinctions differing from healthy people are loss of hippocampal neurons and degeneration of synapses [3]. ad is difficult to diagnosis for its inconspicuous initial symptoms. the disruption of episodic memory, emerging firstly in ad patients [4,5], as well as other subsequent cognitive impairments, especially in attention and execution, semantic memory, language expression, and orientation [6,7] may not draw http://www.pub.iapchem.org/ojs/index.php/admet/index admet & dmpk 3(3) (2015) 216-234 chemical and physical approaches for the treatment of ad doi: 10.5599/admet.3.3.194 217 attention of patients and their family members. the clinical diagnosis of ad would not be definitive until the cognitive impairments become serious enough to affect normal social function. pathogenesis of ad is considered to be the result of the accumulation of the aβ, which is triggered by the overproduction of aβ and the lack of clearance mechanisms. aβ self-aggregates into oligomers, and then forms neuritic plaques in brain [8]. the potent synaptotoxin of insoluble aβ plaques is endangering proteasome function, mitochondrial activity, neuronal function and so forth. on the other hand, hyperphosphorylated tau destabilizes microtubules, resulting in the damage of axonal transport and then leading to neuronal dysfunction [9]. moreover, hyperphosphorylated tau can self-accumulate into nfts which are also harmful to neurons. furthermore, the cleanup process of hyperphosphorylated tau is suppressed by the actions of aβ [10]. therefore, aβ cascade and hyperphosphorylated tau protein represent two essential therapeutic targets for researchers and clinicians. the effective pharmacological treatment and prevention of ad is exactly the disease-modifying therapies which researchers and clinicians are now trying to develop. as aβ cascade and hyperphosphorylated tau protein are generally recognized as vital therapeutic targets, the processes of developing disease-modifying therapies are concentrated on them [11,12]. here we address the 4 main strategies focusing on aβ cascade from different directions, and 5 diverse means aiming at hyperphosphorylated tau in details. it is important to note that we have already faced many failures in the course of developing diseasemodifying therapies, while it is nice to see some successes in the development of symptomatic treatment therapies [13]. available drugs nowadays for treating alzheimer’s disease are directing on cholinesterase and n-methyl-d-aspartate (amda) receptor, which are donepezil, galantamine, rivastigmine and huperzine a, as well as memantine [1,12,13]. in addition to the traditional ad-oriented chemical methods, one physical method, which is called deep brain stimulation (dbs), appears to be a promising approach to modulate the symptoms of some neurodegenerative diseases [14]. existing for years, dbs is a fairly minimally invasive neurological procedure which involves the implantation of a brain pacemaker sending mild electrical impulses to target encephalic regions, with the aim to treat certain affective and movement disorders [15]. the origins of dbs are connected to the beneficial effects of electrical stimulation in certain deep brain areas, and performed during the stereotactic functional neurosurgery to identify the right position of the electrodes for treating movement disorders and tremor in parkinson disease (pd) [16]. as the significantly advantageous outcomes and favourable actions result from treatment of pd which is conducted by dbs, it could be beneficial to apply dbs in the context of ad, especially to mitigate the symptoms resulting from cognitive impairment. the objective of this review seeks to unfold a retrospection of the chemical anti-ad researches and clinical reality of chemical treatment, as well as the physical anti-ad strategy in both preclinical and clinical studies. 2. chemical methods to treat ad 2.1 pharmacological strategies of disease-modifying therapeutic methods 2.1.1 aβ-targeting strategies aβ is produced by sequential proteolysis, where the overexpressed amyloid precursor protein (app) is successively cleaved by β-secretase in extracellular region, and γ-secretase in transmembrane domain [17]. jing li et al. admet & dmpk 3(3) (2015) 216-234 218 alternatively, aβ would not be generated if app is cleaved by the α-secretase, and another soluble neurotrophic, sappα is the hydrolysis product in this case [17]. details are shown in figure 1. however, due to the high expression level of β-secretase, amyloidogenic pathway is more likely to occur than nonamyloidogenic pathway in neurons [18]. aβ isoforms are proteins containing 36-43 amino acid residues, where aβ40 and aβ42 are the two major isoforms. aβ42 is predominant in the brain parenchyma of ad patients, which may result from fibrillogenic feature. if aβ cascade is chosen as therapeutic target, four main options seem to be feasible to tackle the problem: (1) suppression of aβ production, (2) stimulation of aβ clearance, (3) prevention of aβ aggregation into amyloid plaques, and (4) raising brain resistance to aβ [1, 13]. figure 1. aβ cascade and tau process: (a) amyloidogenic pathway of app proteolysis, (b) non-amyloidogenic pathway of app processing, (c) interaction between aβ cascade and tau process, (d) tau process. 2.1.1.1 three strategies of decreasing aβ production the extracellular accumulation of senile plaques composes of the aβ peptide, which represents one of the two defining lesions in ad brain. as a vital step of ad pathogenesis, aβ production is a potentially crucial target for disease-modifying therapeutic methods which attempt to restrain the disease at early stages [12,19]. 2.1.1.1.1 β-secretase inhibition β-secretase is an obvious therapeutic target because it initiates the amyloidogenic processing. however, the development of β-secretase inhibitors does not appear to be straight forward since the physiological roles and functions of this enzyme are not fully understood [20]. besides, it could be challenging to develop a blood brain barrier permeable drug which can bind effectively to the large catalytic pocket of the enzyme. the oral drugs for type 2 diabetes, rosiglitazone and pioglitazone show the ability to inhibit β-secretase by activating the peroxisome-proliferator activated receptor-γ (pparγ), which can suppress the expression of app and β-secretase [21]. rosiglitazone has been studied in a large phase 3 trial with ad patients, but the admet & dmpk 3(3) (2015) 216-234 chemical and physical approaches for the treatment of ad doi: 10.5599/admet.3.3.194 219 result was negative [22]. and it is also warned by fda that rosiglitazone may be associated with cardiac risks. pioglitazone exhibited its therapeutic effect of cognitive and functional improvements, and the effect of disease stabilization in diabetic patients with ad in a japanese open-label study [23]. and one phase 2 study showed no unanticipated or serious adverse events attributable to pioglitazone over a long-term exposure in nondiabetic patients with ad, though no significant treatment effect was observed [24]. an ongoing phase 3 trial (5,800 cognitively normal participants in the united states, australia, europe, and russia) is underway to evaluate pioglitazone's ability to delay this diagnosis in people deemed high-risk with ad [25]. another β-secretase inhibitor, cts-21166 has been evaluated in a phase 1 study. preliminary clinical results suggested that this inhibitor was able to reduce the aβ levels in human plasma [26]. other βsecretase inhibitors such as gsk188909 and pms777 have been reported to reduce the level of aβ42 in vitro [27, 28], but no clinical trial is conducted now. with more promising β-secretase inhibitors come to the horizon, it is expected that some of them might be developed into candidate drugs showing positive indications in future clinical evaluations for the modulation of cognitive decline in ad patients [29]. 2.1.1.1.2 γ-secretase modulation/inhibition similar to β-secretase, γ-secretase also has many other substrates besides app, including the notch receptor and other neuronal substrates. notch, involving in cell differentiation, is necessary for growth and development. the side effects of γ-secretase inhibition which is associated with notch (gastrointestinal and haematological toxicity, skin cancer, and changes in hair colour) have been hindering the development of γsecretase inhibitors [30]. several γ-secretase inhibitors linked with ad, like semagacestat (ly-450139), pf-3084014, avagacestat (bms-708163), begacestat (gsi-953), and nic5-15 have reached clinical trial. semagacestat is the first γsecretase inhibitor to have been taken into phase 3 clinical trials. however, the trials were suspended because of no efficacy on cognition or function and increased risk of infections and skin cancer [31]. a phase 1 clinical trial with healthy subjects showed that pf-3084014 with high selectivity for app had good activity exhibiting a dose-dependent decrease in plasma aβ concentrations, though its effects on cerebrospinal fluid (csf) were low [32]. avagacestat, begacestat and nic5-15 are notch-sparing γ-secretase inhibitors, which can be considered to have the effect of decreasing csf aβ concentration without arousing notch-related toxicity. however, most patients ceased the phase 2 trials with avagacestat because of gastrointestinal and dermatological side effects such as diarrhea, nausea, vomiting, rash, and itching skin, and even nonmelanoma skin cancers [33]. begacestat has been tested in a phase 1 trial with the consequence of reducing aβ concentrations in the plasma (with delayed rebound) [34], without significant reduction of csf aβ40. nic5-15 is a small and single monosaccharide exited in soy and other plants and fruits. it has been tested in ad patients in a finished phase 2 trial to assess its safety and efficacy, and the preliminary results have shown that this compound is safe and well tolerated [35].the other phase 2 trial was initiated in 2012, and no study results have been reported [36]. with selective inhibition of app proteolysis and no notch-based adverse effects, γ-secretase modulators have been given much attention, which can shift the γ-secretase cleavage point to obtain nontoxic, shorter aβ isoforms. several non-steroidal anti-inflammatory drugs (nsaids) were reported as aβ42-lowering γsecretase modulators, including sulindacsulfide, indomethacin, and ibuprofen. they can bind to app to jing li et al. admet & dmpk 3(3) (2015) 216-234 220 reduce the generation of aβ40 and aβ42, and enhance the production of aβ38 [37]. as a promising candidate, tarenflurbil, the r enantiomer of fluriprofen, failed in phase 3 trial for ad patients. disappointingly, ad patients treated with this compound showed progressive deterioration in cognitive functions and daily activities. this negative result might be ascribed to the low γ-secretase modulator potency and poor cns penetration [38]. and another compound called chf-5074 showed antiinflammatory benefit and improvement of executive function in phase 2 trial, with adverse events of mild diarrhoea, dizziness and backache [39]. not unexpectedly, another phase 2 trial with chf-5074 was withdrawn from clinicaltrials.gov prior to enrolment [40]. 2.1.1.1.3 α-secretase activation since α-secretase competes for cleaving app with β-secretase, enhancement of α-secretase activity can upregulate non-amyloidogenic pathway, which may decrease aβ generation and have therapeutic potential. a phase 2a trial with etazolate (eht-0202) has elaborated that this drug is safe for ad patients and well tolerated, which offer a good support for further studies of this compound [41]. bryostatin-1 is another compound which has entered phase 2 trials to evaluate its safety, tolerability and potential effectiveness, for it showed encouraging capacities of inhibiting aβ40 and aβ42, and improving behavioural assessment in ad mouse model [42]. however, the process of this compound in one phase 2 trial is uncovered [43], and one was terminated [44], and another is not open yet for participant recruitment [45]. it remains to be seen if this compound shows positive indications. another compound, exebryl-1, also showed positive results with substantial reduction of aβ formation and aggregation in ad transgenic mice, it has been regarded as another promising candidate to cure ad [46]. accordingly, clinical evaluation is expected to take place in china and us. 2.1.1.2 promotion of aβ clearance 2.1.1.2.1 immunotherapy active (vaccination) and passive (monoclonal antibodies) immunizations have been focused on to remove soluble and aggregated aβ for their promising data from in vitro researches and animal studies. one of the premier active immunization trials was conducted using human aβ42 (an-1792) and a thelper adjuvant (qs-21) in combination. however, the phase 2 study was terminated because of severe side effect, meningoencephalitis, and low probability of predetermined antibody response [47]. the development of new vaccines without the amino acid parts that stimulating t cells has been considered to avoid neuroinflammation and toxicity. as cad-106 has shown a good safety profile and acceptable antibody response in a phase 1 trial [48], it is now being tested in phase 2 trials. acc-001 and v950 are conjugate of multiple short aβ fragments, which can cause an immune response based on the bcell epitope that is included in the n-terminal aβ fragment [13]. v950 was tested in a phase 1 trial as an aluminium-containing adjuvant [49], which was completed in 2012. acc-001 was being tested in multiple phase 2 trials, with or without qs-21. however, acc-001 has been discontinued from development since may 9, 2013 [50]. a phase 1/2 trial of aci-24 began in the year of 2009, to explore the safety, efficacy, and immunogenicity of this liposome vaccine, which showed positive consequence in ad transgenic mice [51]. another compound ub311 was undergone a phase 1 study in taiwan to address the safety, tolerability and immunogenicity issues with favourable results obtained, including positive antibody responses and elevated neuropsychological outcomes without any serious side effects [52]. the other phase 1 trial held by the same group to monitor long-term efficacy and immunogenicity of the same vaccine was completed in the year of 2011 but no result has reported [53]. a phase 2 trial of ub311 was reported being initiated by admet & dmpk 3(3) (2015) 216-234 chemical and physical approaches for the treatment of ad doi: 10.5599/admet.3.3.194 221 united biochemical, inc. recently [54]. some severe side effects induced by vaccination and passive immunotherapy have been considered an alternative immunotherapeutic strategy, which are based on monoclonal antibodies or polyclonal immunoglobulins to promote the clearance of aβ. however, revascularization would be a potential issue for long term treatment of passive immunization, for these therapies could lead to cerebral microhemorrhages related to vascular amyloid deposits [55]. nevertheless, some monoclonal antibodies have already been tested in ad patients, like solanezumab (ly-2062430), bapineuzumab (aab-001), and gantenerumab (rg1450), ponezumab (pf-04360365), gsk-933776, mabt-5102a and immunoglobulins (ivig) [13]. serious adverse effect, vasogenicedema, occurred in the phase 2 trial of bapineuzumab, with no significant improvements of cognition and daily activities [56]. non-significant efficacy and cognitive enhancement were found in apoe ε4 non-carriers, so two phase 3 trials of this humanized monoclonal antibody were initiated using apoe ε4 carriers and non-carriers respectively in mild-to-moderate ad patients. unfortunately, the consequences were disillusionary and bapineuzumab did not improve clinical outcomes in ad patients [57]. the results of phase 1 and 2 trials showed that solanezumab might have pharmacological function in ad [58, 59], and two phase 3 trials were held to further explore the efficacy of this humanized monoclonal antibody. however, the disappointing results showed that it failed to provide improvement of cognition and functional ability [60]. as a fully human anti-aβ antibody, gantenerumab could reduce small plaques by recruiting microglia, and prevents new plaque formation [61]. a phase 3 trial of gantenerumab in 1,000 mild ad patients was initiated in the year of 2014 [62]. ponezumab is a humanized igg2δa monoclonal antibody which can bind to aβ40. phase 1 trial manifested its good tolerability in patients with mild-to-moderate ad [63], but two phase 2 trials showed no effect on the primary endpoints of change in brain or cerebrospinal fluid (csf) aβ burden [64]. the safety and tolerability, as well as pharmacokinetics and pharmacodynamics of gsk-933776 were investigated by two phase 1 trials with positive results [65, 66]. further clinical study is expected to take place in due course. natural anti-amyloid antibodies have been identified in human intravenous ivig, which were isolated from pooled plasma of healthy blood donors. a phase 1 trial has been carried out with eight patients treated with ivig, and been completed with seven patients. the exciting results showed that ivig prevented the decline of cognitive function in all seven patients and even improved cognition in six [67]. a phase 3 clinical study with more than 360 ad patients was started in 2009. the results from this clinical trial may provide conclusive evidence as to whether ivig can be used as a treatment option for ad [67]. 2.1.1.2.2 aβ-degrading enzymes recent studies have showed that the decreased clearance of aβ peptide played a more important role than its increased production in sporadic ad [68, 69], and the way to upgrade or activate aβ-degrading enzymes would be another potential approach to conquer ad. several proteinases, such as, neprilysin (nep) [70, 71], angiotensin-converting enzyme (ace) [1], plasmin [1], insulin degrading enzyme (ide) [1], and endothelin converting enzyme (ece) 1 and 2 [1], were shown to be able to degrade aβ peptides in vitro or in animal models. 2.1.1.3 suppression of aβ aggregation jing li et al. admet & dmpk 3(3) (2015) 216-234 222 the aβ-induced neurotoxicity has been shown on a number of occasions. potentially it may be possible to elicit a therapeutic effect by inhibiting the aβ peptide aggregation. as the only aggregation inhibitor entering phase 3 trial, tramiprosate (3aps) showed negative clinical efficacy, which resulted in the discontinuation of clinical trial of this drug [72]. though showing positive results in phase 2 trials, the development of clioquinol (pbt1) was suspended by its manufacturer [73]. another compound with the same production line, pbt2, failed to achieve its targets in phase 2 trials, and failed to show any benefits [74, 75]. primary clinical efficacy outcomes of a phase 2 trial of another compound, elnd005 (scylloinositol), were not significant [76]. 2.1.1.4 raising brain resistance to aβ for increasing brain resistance to aβ, one target is the group iv phospholipase a2 (giva-pla2), which is involved in neurovirulence of aβ. studies showed genetic reduction or ablation of giva-pla2 inhibits aβinduced cognitive deficit in animal models [77]. however specific drugs to inhibit giva-pla2 remain to be confirmed. the other target is tau protein level since it may involve the mediation of aβ-dependent neuronal dysfunction. it has been showed that genetic reduction or ablation of tau suppresses aβ-induced cognitive deficit, but does not affect the plaque burden or aβ oligomer levels in animal models [78]. 2.1.2 tau-targeting strategies tau protein is a cytoplasmatic microtubule-associated protein that is more frequently found in neurons than non-neuronal cells [79]. the main function of tau is stabilizing axonal microtubules through interacting with tubulin during its polymerization [80]. abnormally hyperphosphorylated tau develops the paired helical filaments (phf), which has severe toxicity impairing axonal transport gravely. tau hypothesis is the main competitor of the aβ hypothesis in ad pathologies [81], but it is also interacts with aβ cascade. the strategies concentrated on tau protein mainly based on its phosphorylation, aggregation and misfolding. 2.1.2.1 prevention of tau phosphorylation the phosphorylation of tau protein, which mainly mediated by microtubule-associated regulatory kinase (mark), cyclin-dependent kinase-5 (cdk5), and glycogen synthase kinase 3β (gsk-3β), increases dramatically in ad patients [82, 83], showing that the inhibitors of tau kinases could have the ability of antiad. valproate (vpa) showed its inhibitory effects on tau hyperphosphorylation in both ad transgenic mice and sh-sy5y cell model, which might be realized through both regulating gsk3β and cdk5 signalling pathways [84]. as the first drug aiming at tau hyperphosphorylation has entered phase 3 trial, valproate showed disappointingly negative results for no improvements on cognitive and functional performance of ad patients [85, 86]. the most studied compound having capacity of suppressing gsk3β is lithium [87], but it did not show its ability of improving working memory in aged 3xtg-ad mice [88], so the most efficacious treatment might be the combination of lithium and other anti-aβ interventions. unfortunately, the clinical consequence of a phase 2 trial showed us the negative side of this drug, which manifested no effect on gsk-3 activity and no improvement in global cognitive performance [89]. besides inhibiting the action of mapk, protein phosphatase pp-2a could also increase tau dephosphorylation. the inhibitors of cdk5 seem to impair the development of pathology in tau transgenic mice [90]. the m1 muscarinic agonist af267b (ngx267) can inhibit gsk-3β [1]. also, minocycline could inhibit tau aggregation, which associated with the suppression of aβ-induced neuronal death and cognitive impairment in animal models [91]. 2.1.2.2 prevention of tau aggregation though it is an ordinarily soluble protein, tau forms insoluble, filamentous accumulation in the course of nfts formation. inhibitors of tau aggregation independent of phosphorylation have been found and tested admet & dmpk 3(3) (2015) 216-234 chemical and physical approaches for the treatment of ad doi: 10.5599/admet.3.3.194 223 in cell cultures [92], including n-phenylamines, phenothiazines, thiazolyl-hydrazides, anthraquinones, thiacarbocyanine dyes, polyphenols, aminothienopyridazines, rhodanines, quinoxalines and so on [93]. 2.1.2.3 prevention of tau misfolding besides tau aggregation, the misfolding of hyperphosphorylated tau is conducive to the pathology of ad. it has been suggested that the misfolding of tau and the development of nfts might be prevented by molecular chaperones [94]. moreover, heat shock proteins could activate chaperones to contribute to prevent misfolding of tau [95]. 2.1.2.4 anti-phospho tau antibodies it would be possible to prevent the aggregation of hyperphosphorylated tau in order to limit its neurotoxic effect. sigurdsson and his colleagues suggested that anti-phospho tau antibodies could inhibit brain aggregated tau and suppress progression of tangle-related behavioural phenotype in mouse models [96]. 2.1.2.5 compensation for tau function since hyperphosphorylated tau lose the physiological cellular function of binding and stabilizing microtubules, possessing the function of stabilizing microtubules, bms-241027 may compensate for the loss of normal tau, and then suppress neurodegeneration and succeeding cognitive decline [97]. bms241027 showed its functions of reversing behavioral and cognitive deficiencies and restraining neuron loss in mouse models [98, 99]. a phase 1 trial to evaluate the tolerability and pharmacology of bms-241027 was completed in 2013, but no consequences have been demonstrated [100]. 2.2 symptomatic therapies: approved drugs for treatment in alzheimer’s disease until now, though there are no available drugs which can cure ad, some drugs have the ability to improve symptoms, or slow down their progression in some patients. current drugs for treating ad have been categorized into two types: one is inhibitors of the acetylcholinesterase (acheis), and the other one is n-methyl-d-aspartate (nmda) receptor antagonists, which target glutamatergic and cholinergic neurotransmission to improve the symptoms [13], though their neuroprotective activity nowadays is still controversial. as acheis can increase the levels of acetylcholine, a key neurotransmitter involved in memory, they are now utilized as long-term symptomatic treatments for ad patients [101]. ad damages or destroys neurons which produce and use acetylcholine (ach), then reducing the amount available to carry messages. an achei slows the breakdown of ach by blocking the activity of ache. by maintaining ach levels, the drug may help compensate for the loss of functioning neurons. it has also been shown that acheis intervene with app cleavage and decrease aβ-induced toxicity through some mechanisms, including interference of the aβ production, alteration of the aβ levels, and formation of the soluble form of app [101]. acheis can also modulate the expression of ache isoforms; increase the expression of nicotinic receptors, via which nicotine can show its protective effect by suppress aβ toxicity and aβ [102]. although modest, acheis still exhibit a significant effect on the cognitive functions of ad patients, and also show a positive effect on behavior and mood [103]. in general, acheis are well tolerated, but the possibility of adverse effects still exists. different acheis may have different safety profiles [104]. it is believed that the benefits of the usage of acheis outweigh their risks and costs. acheis now are considered as primary therapy for the patients with mild to moderate ad [12]. four different acheis drugs have been approved by u.s. fda to treat ad symptoms, which are jing li et al. admet & dmpk 3(3) (2015) 216-234 224 donepezil, rivastigmine, galantamine and tacrine. donepezil is approved to treat all stages of ad [101,105,106]; rivastigmine and galantamine are approved to treat mild to moderate stages [101]. although tacrine is the first achei which was approved by the fda in 1993, some side effects including hepatotoxic effects [107] lead to its rare usage. the fifth alzheimer's drug, memantine, is an nmda receptor antagonist, which has the capability of suppressing aβ toxicity and microglia-associated inflammation; preventing hyperphosphorylation of tau; increasing the release of neurotrophic factors from astroglia [108-110]. memantine works via decreasing glutamate excitotoxicity. glutamate itself is a vital neurotransmitter which involved in learning and memory in the brain [111]. nmda receptors permit calcium to enter the cell when combining with glutamate, which is imperative for cell signaling, and then for learning and memory [112]. however, excess glutamate can be released from impaired neurons in ad, resulting in chronic overexposure to calcium, which can in turn accelerate neuron impairment [113]. memantine can partially block the nmda receptors, and then prevent this destructive chain of events [108, 109]. 3. physical methods to treat ad 3.1 clinical studies of deep brain stimulation although dbs has been approved to treat some progressive neurodegenerative illness, such as parkinson disease (pd), the quantity of researches concentrating on the potential therapeutic effect of clinical dbs on the dementia has been still quite limited till now. in brief, the dbs technique involves stereotactically guided implantation of one or more electrodes into targeted brain region, and the electric pulses which are generated by a pacemaker implanted in infraclavicular area will be sent through wires and then electrode into brain to influence neural activities [114,115] (see figure 2). figure 2.diagrammatic sketch of deep brain stimulation. dbs is a fairly minimally invasive neurological procedure which involves the implantation of a brain pacemaker sending mild electrical impulses to targeted encephalic regions to treat certain affective and movement disorders the first clinical dbs treatment that approved by the food and drug administration in 1997 was for admet & dmpk 3(3) (2015) 216-234 chemical and physical approaches for the treatment of ad doi: 10.5599/admet.3.3.194 225 essential tremor, later for parkinson’s disease in 2002, then for dystonia in 2003, and obsessive-compulsive disorder (ocd) in 2009 [116]. dbs can also be used in research studies of major depression [117,118] and other affective disorders [119], as well as in treatment of chronic pain [120,121], though none of these newer indications have been approved by fda till now, which may probably due to high complication rates and side effects. so far, only two feasibility studies focusing on fornix dbs [122, 123] with published consequences and one case study [124] and one phase 1 trial [125] of nucleus basalis of meynert (nbm) dbs have been undertaken to evaluate if this method could improve the cognitive functions in pd or ad patients. 3.1.1 fornix stimulation fornix, a c-shaped bundle of fibers (axons) in brain, is a key element of the memory circuitry. it carries signals from the hippocampus to the anterior thalamic nuclei [126, 127]. the main characters of ad are the damages of neurons and neural circuits which relate to cognitive functions. fornix dbs could have the potential to modulate the neurophysiological activities, which may offer good clinical outcomes. in one study, this treatment was applied to a patient with morbid obesity. it has been shown that hypothalamic/fornix dbs modulated limbic activity of that patient and improved his memory [128]. then a phase 1 trial with 6 mild ad patients was investigated by laxton and his colleagues [122]. electrodes were located 2 mm anterior and parallel to the vertical portion of the fornix bilaterally. the neural activities in entorhinal cortex and hippocampus, which was part of the memory circuit, were enhanced by fornix dbs. the damaged glucose utilization in the temporal lobe and parietal lobe was sustained and reversed strikingly after the clinical dbs. moreover, the improvements of cognitive functions and/or the reduction in the rate of cognitive decline in some patients were clearly demonstrated. minor side effects in some patients were observed, encompassing minor heat production, sweating, increases in blood pressure and heart rate. a feasibility study with nine patients was conducted by fontaine and his colleagues to explore the effect and safety of dbs in ad patients who got mild cognitive deficit [123]. however, only one patient agreed to accept the surgery, and accomplished the study. the patient received stimulation bilaterally for 12 months, and the electrode target was the same as laxton’s study. the memory scores were stabilized comparing to control conditions, and a subjective view of memory improvement was delivered by the only patient herself, without significant 225ehavioural improvement or mood enhancement, however. furthermore, the glucose metabolism in medial temporal lobes slightly increased. unfortunately these results are not enough to demonstrate the feasibility of the technique due to the great limitation of the extremely small numbers of the patients. another combination of phase 1 and 2 trial of fornix dbs is now ongoing with about 50 mild probable ad patients to evaluate the acute and long-term safety of this system [129]. 3.1.2 nucleus basalis of meynert (nbm) stimulation it has been shown that nbm lesions can result in neurodegenerative changes in hippocampal mossy fibers and dentate gyrus [130,131], as well as the nadph-diaphorase enhancement in the dorsal ca1-ca3 of the hippocampus [132]. characterized by the intrinsic organization, anatomy, and cholinergic innervation, nbm is thought to be an auspicious target structure for clinical dbs in dementia. there is plentiful experimental evidence showing that the nbm projects cholinergic innervation to the whole cortex [133-136], complemented by gabaergic projections and glutaminergic projections from basal forebrain [124,137]. jing li et al. admet & dmpk 3(3) (2015) 216-234 226 in the case study of freund and his colleagues, nbm dbs was applied to a patient who got severe parkinson-dementia syndrome (pdd) to investigate the potential of this new dbs strategy in modifying the cognitive functions [124]. the only patient who received this surgery is a 71-year-old man with pdd. the target areas were bilateral nbm and subthalamic nucleus (stn). the stn dbs led to the only improvement of motor symptoms, while the nbm dbs resulted in significant improvement cognitive functions, including the attention, alertness, concentration, spontaneity, and drive. therefore the patient’s executive and social functions had been markedly improved. the enhancements in 226ehavioural and cognitive performance might be connected with the effects of irritating residual cholinergic projections and neuronal bodies in nbm. a clinical study with six mild-to-moderate ad patients was conducted by kuhn and his colleagues through applying bilateral low-frequency nbm dbs [125]. the whole trial encompassed a 4-week randomized sham-controlled stimulation period and a subsequent 11-month continued open stimulation phase. the consequences were assessed by cognitive subscale of the ad assessment scale and other psychological tests. electroencephalogram (eeg) and positron emission tomography (pet) were also applied to access the changes before and psychological test after the surgery. nearly no adverse effect showed the expected inner restlessness that one patient declared. although the patients’ average quality of life didn’t change, it could also be considered as a safe procedure obviously without significant stimulation-induced adverse effects, since we have to take all limitations of a pilot study into consideration. 3.2 animal researches related with deep brain stimulation there are several relevant animal studies concentrating on dbs-induced changes on behavior, memory, cognition, and/or neurogenesis. 10 brain regions have been explored in different animal models to investigate their functions in memory enhancement, which are fornix, nbm, hippocampus enthorinal cortex (ec), midline thalamic nuclei (mtn), anterior thalamic nucleus (an), perforant path of hippocampus, dorsal striatum (ds), anterior caudate nucleus, central thalamus (ct), and lateral hypothalamus (lh). 3.2.1 fornix and nucleus basalis of meynert (nbm) fornix dbs could antagonize the memory impairing effects resulting from scopolamine injection in rats [138]. the important favorable role of nbm dbs on memory acquisition was investigated in rats [139], in which investigators considered dbs could increase ach release from nbm neurons that in return helpfully influenced neural plasticity mechanisms. one study suggested that pertaining nbm electrical stimulation facilitates the memory acquisition [140], which supported the fact that nbm could play a part in the early stages of memory formation. another study also manifested the positive consequences of nbm dbs which were assistance of maintain neuronal plasticity and protective effect on the cortex [141]. 3.2.2 hippocampus enthorinal cortex (ec) application of acute high-frequency dbs in hippocampus ec of mice could lead to a transiently promoted proliferation in the dentate gyrus (dg) [142], and these dbs-induced neurons merged with hippocampal circuits once they were mature. 3.2.3 midline thalamic nuclei (mtn) mtn dbs was found to increase short-term memory in the ca1 of hippocampus in transgenic mice [143] with the results of enhancing object recognition memory during behavioral studies and promoting synaptic plasticity in ca1 of hippocampal slices. 3.2.4 anterior nucleus (an) of the thalamus admet & dmpk 3(3) (2015) 216-234 chemical and physical approaches for the treatment of ad doi: 10.5599/admet.3.3.194 227 the role of an dbs in memory might depend on current intensity [144]. high current treatment improved hippocampal activity, while relatively high current disrupted memory acquisition. interestingly, the results of the other study of the same research group showed that the positive role of an dbs on memory and neurogenesis seemed to be dependent on long-term plastic changes, leading to memory enhancement when the training was conducted at longer intervals [144]. another study also demonstrated that an dbs in rats had the ability to improve the hippocampal neurogenesis and reverse experimentally inhibited neurogenesis [145]. 3.2.5 perforant path of hippocampus the perforant path is the connection which originates from the hippocampus ec, and projects into all areas of the hippocampus. it triggers θ-phase resetting and the release of ach, the necessary condition of long-term potentiation (ltp) [14]. one study investigated whether perforant path dbs intervened with the circuits related to encoding and retrieving the information of the received stimuli in the working memory task [146]. the result could be explained through this way: θ-phase resetting might synchronize stimuli with the neural circuits in hippocampus to ensure the favorable conditions to facilitate the encoding of the received information into memory. 3.2.6 other brain regions there are other four brain regions related to dementia where dbs could play a role, which are dorsal stritum, anterior caudate nucleus, central thalamus and lateral hypothalamus. dorsal stritum has been shown to relate to procedural memory [147]. anterior caudate nucleus is in charge of implementing selective adjustments to the ‘associative weights’ between sensory cues and motor responses during the learning period [148]. dbs in central thalamus could lead to a complicated situation in arousal, behavior, and cognition [149-151]. with regard to the lateral hypothalamus, the post-training icss (the intracranial self-stimulation) could enhance the hippocampus-dependent memories via improving molecular and cellular mechanisms associated with synaptic plasticity, neurogenesis, and neuroprotection [152-155]. 4. conclusions researches on ad therapy have had some success in the field of symptomatic treatments, but there have also been many failures in the development of disease-modifying drugs. in spite of the frustrating research progress, more and more researchers and clinicians are dedicating all their energy to fighting with this complicated disease, and still more and more clinical and experimental studies are now underway. it would be beneficial to explore different approaches to narrow the gap between successful preclinical investigations and unsuccessful clinical studies via the rationalization of the disease progression mechanisms. chemical approaches we have discussed here are divided into two categories, one is disease-modifying therapeutic approaches, and the other is symptomatic treatments, in which the former is focused on the aβ cascade and hyperphosphorylated tau protein, and the latter is concentrated on cholinesterase and amda receptor. many clinical studies of disease-modifying drugs failed, and still many are now being kept ongoing, but the symptomatic therapies have received some successes. it is difficult for us to estimate the when the successful treatment of ad will be available, but it is easy for us to tell that the consequences from clinical researches are not in accordance with optimistic preclinical studies. the pessimistic outcomes of most clinical studies may be partly due to diverse errors, which can be categorized into the choice of patients, the choice of drugs, the outcome measurements, trial protocols, and optimization of resources [13]. clearly, the problem is largely due to our limited knowledge to the pathogenesis of ad. as ad is a jing li et al. admet & dmpk 3(3) (2015) 216-234 228 complex multifactorial disorder [156], the details of its pathogenesis are not yet fully understood to support the development of viable drug targets and therapeutic agents. how to solve this problem? on one hand, we need to do further research to explore the real pathogenesis of ad, not only limited to aβ cascade and hyperphosphorylated tau; on the other hand, the strategies of one drug directing at one disease hypothesis to develop ad therapies need to be revised, for a single pathogenic pathway for ad is not possible to be identified [13]. physical approach using dbs seems to be a feasible and safe treatment for patients with dementia, in consideration of the general low complication rate in about 100,000 pd patients treated with dbs all over the world [157,158]. currently, the fornix is the most frequently chosen targeting area for treating dementia, including ad and pdd, according to the extremely small number of cases. the potential of the nbm for dbs treatment in dementia is auspicious, which may be due to its characteristics in the anatomy, connectivity of the cholinergic nucleus, and intrinsic organization. other brain structures we have already mentioned are still lack of human data. the situation in human brain is much more complicated and intricate. the recruitment of patients in clinical evaluation in dbs study on ad was not easy. clearly there is an important role in the preclinical animal research of dbs for ad, which is extremely limited at present, though. with this data, it may be possible to move toward the rationalization of the mechanisms leading to the disease progression. it must be emphasized that continuing preclinical animal experiments and clinical studies are utmost important to demonstrate the promising efficacy of dbs in future treatment of ad. apparently, dbs is being used in the general population now. yet, the strategies to find satisfactory pharmacotherapeutics breakthrough are quite limited. therefore, dbs may be a good option for some ad patients if it is proved to be a valid treatment. as more clinical data emerge, it will be possible to properly evaluate the applications of dbs in the treatment of ad and other dementia. acknowledgement we thank the financial support from the science and technology development fund, macao s.a.r (fdct) (project reference no.: 118/2013/a3). references [1] y. hong-qi, s. zhi-kun, c. sheng-di, translational neurodegeneration 1 (2012) 21. [2] j. hardy, neuron 52 (2006) 3-13. [3] z. s. khachaturian, archives of neurology 42 (1985) 1097-1105. [4] k. a. welsh, n. butters, j. p. hughes, r. c. mohs, a. heyman, archives of neurology 49 (1992) 448452. [5] s. artero, m. c. tierney, j. touchon, k. ritchie, acta psychiatrica scandinavica 107 (2003) 390-393. [6] r. j. perry, j. r. hodges, brain 122 (1999) 383-404. [7] m. a. l. ralph, k. patterson, n. graham, k. dawson, j. r. hodges, brain 126 (2003) 2350-2362. [8] d. j. selkoe, journal of alzheimer’s disesae 3 (2001) 75-80. [9] p. h. reddy, brain research 1415 (2011) 136-148. [10] b. p. tseng, k. n. green, j. l. chan, m. blurton-jones, f. m. laferla, neurobiology of aging 29 (2008) 1607-1618. [11] t. e. golde, journal of neurochemistry 99 (2006) 689-707. [12] d. galimberti, e. scarpini, therapeutic advances in neurological disorders 4 (2011) 203-216. [13] f. mangialasche, a. solomon, b. winblad, p. mecocci, m. kivipelto, lancet neurology 9 (2010) 702716. [14] k. hardenacke et al., frontiers in psychiatry 4 (2013) 159. admet & dmpk 3(3) (2015) 216-234 chemical and physical approaches for the treatment of ad doi: 10.5599/admet.3.3.194 229 [15] m. l. kringelbach, n. jenkinson, s. l. owen, t. z. aziz, nature reviews neuroscience 8 (2007) 623-635. [16] j. m. schwalb, c. hamani, neurotherapeutics 5 (2008) 3-13. [17] w. sue, t. griffin, american journal of clinical nutrition 83 (2006) 470s-474s. [18] r. vassar, journal of molecular neuroscience 23 (2004) 105-114. [19] t. e. golde, brain pathology 15 (2005) 84-87. [20] c. venugopal, c. m. demos, k. s. rao, m. a. pappolla, k. sambamurti, cns & neurological disorders drug targets 7 (2008) 278-294. [21] g. landreth, q. g. jiang, s. mandrekar, m. heneka, neurotherapeutics 5 (2008) 481-489. [22] m. gold, c. alderton, m. zvartau-hind, s. egginton, a.m. saunders, m. irizarry, s. craft, g. landreth, u. linnamägi, s. sawchak, dementia and geriatric cognitive disorders 30 (2010) 131-146. [23] t. sato, h. hanyu, k. hirao, h. kanetaka, h. sakurai, t. iwamoto, neurobiology of aging 32 (2011) 1626-1633. [24] d. s. geldmacher, t. fritsch, m. j. mcclendon, g. landreth, archives of neurology 68 (2011) 45-50. [25] biomarker qualification for risk of mild cognitive impairment (mci) due to alzheimer's disease (ad) and safety and efficacy evaluation of pioglitazone in delaying its onset, https://www.clinicaltrials.gov/ct2/show/nct01931566?term=pioglitazone%2c+alzheimer&rank=1 (october 2014). [26] j. hey, g. koelsch, g. bilcer, a. jacobs, m. tolar, j. tang, international conference on alzheimer's disease (icad), chicago, il, usa (2008) 26-31. [27] i. hussain, j. hawkins, d. harrison, c. hille, g. wayne, l. cutler, t. buck, d. walter, e. demont, c. howes, a. naylor, p. jeffrey, m.i. gonzalez, c. dingwall, a. michel, s. redshaw, j.b. davis, journal of neurochemistry 100 (2007) 802-809. [28] j. li, j. hu, b. shao, w. zhou, y. cui, c. dong, j.m. ezoulin, x. zhu, w. ding, f. heymans, h. chen, cellular and molecular neurobiology 29 (2009) 589-595. [29] a. k. ghosh, m. brindisi, j. tang, journal of neurochemistry 120 suppl 1 (2012) 71-83. [30] g.t. wong, d. manfra, f.m. poulet, q. zhang, h. josien, t. bara, l. engstrom, m. pinzon-ortiz, j.s. fine, h.j.j. lee, l. zhang, g.a. higgins, e.m. parker, journal of biological chemistry 279 (2004) 1287612882. [31] r.s. doody, r. raman, m. farlow, t. iwatsubo, b. vellas, s. joffe, k. kieburtz, f. he, x. sun, r.g. thomas, p.s. aisen; alzheimer's disease cooperative study steering committee, e. siemers, g. sethuraman, r. mohs; semagacestat study group, new england journal of medicine 369 (2013) 341350. [32] r. n. soares h, sikpi m, et al., alzheimer’s & dementia 5 (2009) 252. [33] v. ćorić, c.h. van dyck, s. salloway, n. andreasen, m. brody, r.w. richter, h. soininen, s. thein, t. shiovitz, g. pilcher, s. colby, l. rollin, r. dockens, c. pachai, e. portelius, u. andreasson, k. blennow, h. soares, c. albright, h.h. feldman, r.m. berman, archive of neurology 69 (2012) 1430-1440. [34] r.l. martone, h., zhou, k. atchison, t. comery, j.z. xu, x. huang, x. gong, m. jin, a. kreft, b. harrison, s.c. mayer, s. aschmies, c. gonzales, m.m. zaleska, d.r. riddell, e. wagner, p. lu, s.c. sun, j. sonnenberg-reines, a. oganesian, k. adkins, m.w. leach, d.w. clarke, d. huryn, m. abou-gharbia, r. magolda, j. bard, g. frick, s. raje, s.b. forlow, c. balliet, m.e. burczynski, p.h. reinhart, h.i. wan, m.n. pangalos, j.s. jacobsen, journal of pharmacology and experimental therapeutics 331 (2009) 598-608. [35] m. g. grossman h, luo x, leroith d, sano m, pasinetti g, alzheimer’s & dementia 5 (2009) 259. [36] a single site, randomized, double-blind, placebo controlled trial of nic5-15 in subjects with alzheimer's disease, https://clinicaltrials.gov/ct2/show/record/nct01928420?term=nic5-15&rank=1 (august 2013). [37] t. tomita, expert review of neurotherapeutics 9 (2009) 661-679. [38] r.c. green, l.s. schneider, d.a. amato, a.p. beelen, g. wilcock, e.a. swabb, k.h. zavitz, journal of the american medical association 302 (2009) 2557-2564. https://www.clinicaltrials.gov/ct2/show/nct01931566?term=pioglitazone%2c+alzheimer&rank=1 https://clinicaltrials.gov/ct2/show/record/nct01928420?term=nic5-15&rank=1 jing li et al. admet & dmpk 3(3) (2015) 216-234 230 [39] j. ross, s. sharma, j. winston, m. nunez, g. bottini, m. franceschi, e. scarpini, e. frigerio, f. fiorentini, m. fernandez, s. sivilia, l. giardino, l. calza, d. norris, h. cicirello, d. casula, b.p. imbimbo, current alzheimer research 10 (2013) 742-753. [40] evaluation of effects of multiple dose regimens of chf 5074 on potential biomarkers of neurodegeneration in subjects with mild cognitive impairment (ct05), https://clinicaltrials.gov/ct2/show/nct01723670?term=chf-5074&rank=7 (february 2015). [41] b. vellas, o. sol, p.j. snyder, p.j. ousset, r. haddad, m. maurin, j.c. lemarié, l. désiré, m.p. pando, current alzheimer research 8 (2011) 203-212. [42] r. etcheberrigaray, m. tan, i. dewachter, c. kuipéri, i. van der auwera, s. wera, l. qiao, b. bank, t.j. nelson, a.p. kozikowski, f. van leuven, d.l. alkon, proceedings of the national academy of sciences of the united states of america 101 (2004) 11141-11146. [43] safety, efficacy, pharmacokinetics, and pharmacodynamics study of bryostatin 1 in patients with alzheimer's disease, https://clinicaltrials.gov/ct2/show/nct00606164?term=bryostatin1+alzheimer&rank=1 (january 2008). [44] study to evaluate the preliminary safety, efficacy, pk and pd of bryostatin 1 in patients with alzheimer's disease, https://clinicaltrials.gov/ct2/show/nct02221947?term=bryostatin1+alzheimer&rank=2 (april 2015). [45] a study assessing bryostatin in the treatment of moderately severe to severe alzheimer's disease, https://clinicaltrials.gov/ct2/show/nct02431468?term=bryostatin-1+alzheimer&rank=3 (april 2015). [46] a.d. snow, j. cummings, t. lake, q. hu, l. esposito, j. cam, m. hudson, e. smith, s. runnels, alzheimer’s & dementia 5 (2009) 418. [47] s. gilman, m. koller, r.s. black, l. jenkins, s.g. griffith, n.c. fox, l. eisner, l. kirby, m.b. rovira, f. forette, j.m. orgogozo, an1792(qs-21)-201 study team, neurology 64 (2005) 1553-1562. [48] b. winblad, n. andreasen, l. minthon, a. floesser, g. imbert, t. dumortier, r.p. maguire, k. blennow, j. lundmark, m. staufenbiel, j.m. orgogozo, a. graf, lancet neurology 11 (2012) 597-604 [49] d. l. brody, d. m. holtzman, annual review in neuroscience 31 (2008) 175-193. [50] pfizer pipeline, http://www.pfizer.com/sites/default/files/product-pipeline/pipeline_080913_0.pdf (august 9, 2013). [51] a. muhs, d.t. hickman, m. pihlgren, n. chuard, v. giriens, c. meerschman, i. van der auwera, f. van leuven, m. sugawara, m.c. weingertner, b. bechinger, r. greferath, n. kolonko, l. nagel-steger, d. riesner, r.o. brady, a. pfeifer, c. nicolau, proceedings of the national academy of sciences of the united states of america 104 (2007) 9810-9815. [52] c. y. w. s. lynn*, c. l. finstad, s.-n. hsu, y. kao, c.-f. wu, p.-n. wang, m.-j. chiu, 28th international conference of alzheimer's disease international (2013) 40. [53] study to evaluate safety, tolerability and immunogenicity of vaccine (ub 311) in subjects with alzheimer's disease, https://clinicaltrials.gov/ct2/show/nct00965588?term=ub311&rank=1 (august 2011). [54] b. winblad, a. graf, m. e. riviere, n. andreasen, j. m. ryan, alzheimer's research & therapy 6 (2014) 7. [55] d. m. wilcock, p. t. jantzen, q. li, d. morgan, m. n. gordon, neuroscience 144 (2007) 950-960. [56] s. salloway, r. sperling, s. gilman, n.c. fox, k. blennow, m. raskind, m. sabbagh, l.s. honig, r. doody, c.h. van dyck, r. mulnard, j. barakos, k.m. gregg, e. liu, i. lieberburg, d. schenk, r. black, m. grundman; bapineuzumab 201 clinical trial investigators, neurology 73 (2009) 2061-2070. [57] s. salloway, r. sperling, n.c. fox, k. blennow, w. klunk, m. raskind, m. sabbagh, l.s. honig, a.p. porsteinsson, s. ferris, m. reichert, n. ketter, b. nejadnik, v. guenzler, m. miloslavsky, d. wang, y. lu, j. lull, i.c. tudor, e. liu, m. grundman, e. yuen, r. black, h.r. brashear; bapineuzumab 301 and 302 clinical trial investigators, new england journal of medicine 370 (2014) 322-333. [58] e.r. siemers, s. friedrich, r.a. dean, c.r. gonzales, m.r. farlow, s.m. paul, r.b. demattos, clinical neuropharmacology 33 (2010) 67-73. https://clinicaltrials.gov/ct2/show/nct01723670?term=chf-5074&rank=7 https://clinicaltrials.gov/ct2/show/nct00606164?term=bryostatin-1+alzheimer&rank=1 https://clinicaltrials.gov/ct2/show/nct00606164?term=bryostatin-1+alzheimer&rank=1 https://clinicaltrials.gov/ct2/show/nct02221947?term=bryostatin-1+alzheimer&rank=2 https://clinicaltrials.gov/ct2/show/nct02221947?term=bryostatin-1+alzheimer&rank=2 https://clinicaltrials.gov/ct2/show/nct02431468?term=bryostatin-1+alzheimer&rank=3 http://www.pfizer.com/sites/default/files/product-pipeline/pipeline_080913_0.pdf https://clinicaltrials.gov/ct2/show/nct00965588?term=ub311&rank=1 admet & dmpk 3(3) (2015) 216-234 chemical and physical approaches for the treatment of ad doi: 10.5599/admet.3.3.194 231 [59] m. farlow, s.e. arnold, c.h. van dyck, p.s. aisen, b.j. snider, a.p. porsteinsson, s. friedrich, r.a. dean, c. gonzales, g. sethuraman, r.b. demattos, r. mohs, s.m. paul, e.r. siemers, alzheimer’s & dementia 8 (2012) 261-271. [60] r.s. doody, r.g. thomas, m. farlow, t. iwatsubo, b. vellas, s. joffe, k. kieburtz, r. raman, x. sun, p.s. aisen, e. siemers, h. liu-seifert, r. mohs; alzheimer's disease cooperative study steering committee; solanezumab study group., new england journal of medicine 370 (2014) 311-321. [61] b. bohrmann, k. baumann, j. benz, f. gerber, w. huber, f. knoflach, j. messer, k. oroszlan, r. rauchenberger, w.f. richter, c. rothe, m. urban, m. bardroff, m. winter, c. nordstedt, h. loetscher, journal of alzheimer’s disease 28 (2012) 49-69. [62] a study of gantenerumab in patients with mild alzheimer disease, https://clinicaltrials.gov/ct2/show/study/nct02051608?term=gantenerumab+phase+3&rank=1, (june 2015). [63] j.w. landen, q. zhao, s. cohen, m. borrie, m. woodward, c.b. billing jr, k. bales, c. alvey, f. mccush, j. yang, j.w. kupiec, m.m. bednar, clinical neuropharmacology 36 (2013) 14-23. [64] ponezumab, http://www.alzforum.org/therapeutics/ponezumab (2015). [65] t. leyhe, n. andreasen, m. simeoni, a. reich, c.a. von arnim, x. tong, a. yeo, s. khan, a. loercher, m. chalker, c. hottenstein, h. zetterberg, j. hilpert, p. mistry, alzheimer’s research and therapy 6 (2014) 19. [66] n. andreasen, m. simeoni, h. ostlund, p.i. lisjo, t. fladby, a.e. loercher, g.j. byrne, f. murray, p.t. scott-stevens, a. wallin, y.y. zhang, l.h. bronge, h. zetterberg, a.k. nordberg, a.j. yeo, s.a. khan, j. hilpert, p.c. mistry, plos one 10 (2015) e0098153. [67] r. dodel, f. neff, c. noelker, r. pul, y. du, m. bacher, w. oertel, drugs 70 (2010) 513-528. [68] k.g. mawuenyega, w. sigurdson, v. ovod, l. munsell, t. kasten, j.c. morris, k.e. yarasheski, r.j. bateman, science 330 (2010) 1774. [69] s. d. mulder, r. veerhuis, m. a. blankenstein, h. m. nielsen, experimental neurology 233 (2012) 373379. [70] c. bauer, r. pardossi-piquard, j. dunys, m. roy, f. checler, journal of alzheimer’s disease 27 (2011) 511-520. [71] n.n. nalivaeva, n.d. belyaev, d.i. lewis, a.r. pickles, n.z. makova, d.i. bagrova, n.m. dubrovskaya, s.a. plesneva, i.a. zhuravin, a.j. turner, journal of molecular neuroscience 46 (2012) 569-577. [72] h. amijee, d. i. c. scopes, journal of alzheimers disease 17 (2009) 33-47. [73] p.a. adlard, r.a. cherny, d.i. finkelstein, e. gautier, e. robb, m. cortes, i. volitakis, x. liu, j.p. smith, k. perez, k. laughton, q.x. li, s.a. charman, j.a. nicolazzo, s. wilkins, k. deleva, t. lynch, g. kok, c.w. ritchie, r.e. tanzi, r. cappai, c.l. masters, k.j. barnham, a.i. bush, neuron 59 (2008) 43-55. [74] there is no evidence that mpacs (pbt1 or pbt2) are of benefit in alzheimer's dementia, http://www.cochrane.org/cd005380/dementia_there-is-no-evidence-that-mpacs-pbt1-or-pbt2-areof-benefit-in-alzheimers-dementia (february 21, 2014). [75] pbt2 takes a dive in phase 2 alzheimer's trial, http://www.alzforum.org/news/research-news/pbt2takes-dive-phase-2-alzheimers-trial (apr 01, 2014). [76] s. salloway, r. sperling, r. keren, a.p. porsteinsson, c.h. van dyck, p.n. tariot, s. gilman, d. arnold, s. abushakra, c. hernandez, g. crans, e. liang, g. quinn, m. bairu, a. pastrak, j.m. cedarbaum; elnd005-ad201 investigators, neurology 77 (2011) 1253-1262. [77] r. o. sanchez-mejia, j.w. newman, s. toh, g.q. yu, y. zhou, b. halabisky, m. cissé, k. scearce-levie, i.h. cheng, l. gan, j.j. palop, j.v. bonventre, l. mucke, nature neuroscience 11 (2008) 1311-1318. [78] e.d. roberson, k. scearce-levie, j.j. palop, f. yan, i.h. cheng, t. wu, h. gerstein, g.q. yu, l. mucke, science 316 (2007) 750-754. [79] t. a. schoenfeld, r. a. obar, international review of cytology 151 (1994) 67-137. [80] d. couchie, c. mavilia, i.s. georgieff, r.k. liem, m.l. shelanski, j. nunez, proceedings of the national academy of sciences of the united states of america 89 (1992) 4378-4381. https://clinicaltrials.gov/ct2/show/study/nct02051608?term=gantenerumab+phase+3&rank=1 http://www.alzforum.org/therapeutics/ponezumab http://www.cochrane.org/cd005380/dementia_there-is-no-evidence-that-mpacs-pbt1-or-pbt2-are-of-benefit-in-alzheimers-dementia http://www.cochrane.org/cd005380/dementia_there-is-no-evidence-that-mpacs-pbt1-or-pbt2-are-of-benefit-in-alzheimers-dementia http://www.alzforum.org/news/research-news/pbt2-takes-dive-phase-2-alzheimers-trial http://www.alzforum.org/news/research-news/pbt2-takes-dive-phase-2-alzheimers-trial jing li et al. admet & dmpk 3(3) (2015) 216-234 232 [81] a. mudher, s. lovestone, trends in neuroscience 25 (2002) 22-26. [82] g. v. johnson, w. h. stoothoff, journal of cell science 117 (2004) 5721-5729. [83] r. medeiros, d. baglietto-vargas, f. m. laferla, cns neuroscience and therapeutics 17 (2011) 514524. [84] j.p. hu, j.w. xie, c.y. wang, t. wang, x. wang, s.l. wang, w.p. teng, z.y. wang, brain research bulletin 85 (2011) 194-200. [85] p. n. tariot, p. s. aisen, journal of clinical psychiatry 70 (2009) 919-921. [86] a. p. tariot, p. aisen, j. cummings, l. jakimovich, l. schneider, r. thomas, l. becerra, r. loy, alzheimer’s & dementia 5 (2009) 84–85. [87] a. martinez, d. i. perez, journal of alzheimers disease 15 (2008) 181-191. [88] a. caccamo, s. oddo, l. x. tran, f. m. laferla, american journal of pathology 170 (2007) 1669-1675. [89] h. hampel, m. ewers, k. bürger, p. annas, a. mörtberg, a. bogstedt, l. frölich, j. schröder, p. schönknecht, m.w. riepe, i. kraft, t. gasser, t. leyhe, h.j. möller, h. basun, journal of clinical psychiatry 70 (2009) 922-931. [90] w. noble, v. olm, k. takata, e. casey, o. mary, j. meyerson, k. gaynor, j. lafrancois, l. wang, t. kondo, p. davies, m. burns, a. veer, r. nixon, d. dickson, y. matsuoka, m. ahlijanian, l.f. lau, k. duff, neuron 38 (2003) 555-565. [91] y. choi, h.s. kim, k.y. shin, e.m. kim, m. kim, h.s. kim, c.h. park, y.h. jeong, j. yoo, j.p. lee, k.a. chang, s. kim, y.h. suh, neuropsychopharmacology 32 (2007) 2393-2404. [92] m. pickhardt, z. gazova, m. von bergen, i. khlistunova, y. wang, a. hascher, e.m. mandelkow, j. biernat, e. mandelkow, journal of biological chemistry 280 (2005) 3628-363. [93] c. ballatore, k.r. brunden, j.q. trojanowski, v.m. lee, a.b. smith 3rd, d.m. huryn, current topics in medicinal chemistry 11 (2011) 317-330. [94] f. dou, w.j. netzer, k. tanemura, f. li, f.u. hartl, a. takashima, g.k. gouras, p. greengard, h. xu, proceedings of the national academy of sciences of the united states of america 100 (2003) 721-726. [95] c. a. dickey, j. dunmore, b. lu, j.w. wang, w.c. lee, a. kamal, f. burrows, c. eckman, m. hutton, l. petrucelli, faseb journal 20 (2006) 753-755. [96] e. m. sigurdsson, journal of alzheimer’s disease 15 (2008) 157-168. [97] k. r. brunden, b. zhang, j. carroll, y. yao, j.s. potuzak, a.m. hogan, m. iba, m.j. james, s.x. xie, c. ballatore, a.b. smith 3rd, v.m. lee, j.q. trojanowski, journal of neuroscience 30 (2010) 13861-13866. [98] b. zhang, j. carroll, j.q. trojanowski, y. yao, m. iba, j.s. potuzak, a.m. hogan, s.x. xie, c. ballatore, a.b. smith 3rd, v.m. lee, k.r. brunden, journal of neuroscience 32 (2012) 3601-3611. [99] d. e. hurtado, l. molina-porcel, m. iba, a.k. aboagye, s.m. paul, j.q. trojanowski, v. m.y. lee, american journal of pathology 177 (2010) 1977-1988. [100] study to evaluate the safety, tolerability and the effect of bms-241027 on cerebrospinal fluid biomarkers in subjects with mild alzheimer's disease, https://clinicaltrials.gov/ct2/show/nct01492374?term=bms-241027&rank=1 (october 2013). [101] a. nordberg, alzheimer disease and associated disorders 20 (2006) s12-s18. [102] l. r. fodero, s.s. mok, d. losic, l.l. martin, m.i. aguilar, c.j. barrow, b.g. livett, d.h. small, journal of neurochemistry 88 (2004) 1186-1193. [103] j. birks, cochrane database of systematic reviews (2006) cd005593. [104] m. mehta, a. adem, m. sabbagh, international journal of alzheimer's disease 2012 (2012) 728983. [105] b. winblad, l. kilander, s. eriksson, l. minthon, s. båtsman, a.l. wetterholm, c. jansson-blixt, a. haglund; severe alzheimer's disease study group, lancet 367 (2006) 1057-1065. [106] b. winblad, american journal of alzheimer's disease & other dementias 24 (2009) 185-192. [107] p. b. watkins, h. j. zimmerman, m. j. knapp, s. i. gracon, k. w. lewis, journal of the american medical association 271 (1994) 992-998. [108] g. l. wenk, c. g. parsons, w. danysz, behavioural pharmacology 17 (2006) 411-424. https://clinicaltrials.gov/ct2/show/nct01492374?term=bms-241027&rank=1 admet & dmpk 3(3) (2015) 216-234 chemical and physical approaches for the treatment of ad doi: 10.5599/admet.3.3.194 233 [109] j. j. pei, m. sjogren, b. winblad, current opinion in psychiatry 21 (2008) 555-561. [110] h.m. wu, n.s. tzeng, l. qian, s.j. wei, x. hu, s.h. chen, s.m. rawls, p. flood, j.s. hong, r.b. lu, neuropsychopharmacology 34 (2009) 2344-2357. [111] g. riedel, b. platt, j. micheau, behavioural brain research 140 (2003) 1-47. [112] a. sanchez-perez, m. llansola, o. cauli, v. felipo, cerebellum 4 (2005) 162-170. [113] u. wojda, e. salinska, j. kuznicki, iubmb life 60 (2008) 575-590. [114] e. bell, g. mathieu, e. racine, surgical neurology 72 (2009) 577-586; discussion 586. [115] p. e. holtzheimer, h. s. mayberg, annual review of neuroscience 34 (2011) 289-307. [116] s. miocinovic, s. somayajula, s. chitnis, j. l. vitek, journal of the american medical association neurology 70 (2013) 163-171. [117] r.j. anderson, m.a. frye, o.a. abulseoud, k.h. lee, j.a. mcgillivray, m berk, s.j. tye, neuroscience and biobehavioral reviews 36 (2012) 1920-1933. [118] t.e. schlaepfer, m.x. cohen, c. frick, m. kosel, d. brodesser, n. axmacher, a.y. joe, m. kreft, d. lenartz, v. sturm, neuropsychopharmacology 33 (2008) 368-377. [119] b. velasques, c. diniz, s. teixeira, c. cartier, c. peressutti, f. silva, m. de carvalho, a. novaes, j. bittencourt, a.e. nardi, e. cheniaux, l. basile, m. cagy, r. piedade, p. ribeiro, cns & neurological disorders-drug targets 13 (2014) 961-971. [120] r. f. young, t. brechner, cancer 57 (1986) 1266-1272. [121] b. jenkins, s. j. tepper, headache 51 (2011) 1408-1418. [122] a.w. laxton, d.f. tang-wai, m.p. mcandrews, d. zumsteg, r. wennberg, r. keren, j. wherrett, g. naglie, c. hamani, g.s. smith, a.m. lozano, annals of neurology 68 (2010) 521-534. [123] d. fontaine , a. deudon, j.j. lemaire, m. razzouk, p. viau, j. darcourt, p. robert, journal of alzheimer’s disease 34 (2013) 315-323. [124] h.j. freund, j. kuhn, d. lenartz, j.k. mai, t. schnell, j. klosterkoetter, v. sturm, archives of neurology 66 (2009) 781-785. [125] j. kuhn, k. hardenacke, d. lenartz, t. gruendler, m. ullsperger, c. bartsch, j.k. mai, k. zilles, a. bauer, a. matusch, r.j. schulz, m. noreik, c.p. bührle, d. maintz, c. woopen, p. häussermann, m. hellmich, j. klosterkötter, j. wiltfang, m. maarouf, h.j. freund, v. sturm, molecular psychiatry 20 (2015) 353360. [126] j. p. aggleton, r. desimone, m. mishkin, journal of comparative neurology 243 (1986) 409-421. [127] n. neave, s. lloyd, a. sahgal, j. p. aggleton, behavioural brain research 65 (1994) 89-101. [128] c. hamani, m.p. mcandrews, m. cohn, m. oh, d. zumsteg, c.m. shapiro, r.a. wennberg, a.m. lozano, annals of neurology 63 (2008) 119-123. [129] advance dbs-f in patients with mild probable alzheimer's disease, https://clinicaltrials.gov/ct2/show/nct01608061?term=fornix+dbs&rank=4 (march 2015). [130] f. amenta, e. bronzetti, m.g. caporali, e. ciriaco, g.p. germana, t. niglio, a. ricci, a.s. de carolis, archives of gerontology and geriatrics 12 (1991) 49-58. [131] i. panocka, s. sagratella, a. scotti de carolis, y. c. zeng, f. amenta, neuroscience letter 190 (1995) 207-211. [132] m. sabbatini et al., mechanisms of ageing and development 107 (1999) 147-157. [133] v. parikh, m. sarter, molecular and biophysical mechanisms of arousal, alertness, and attention 1129 (2008) 225-235. [134] b. j. everitt, t. w. robbins, annual review of psychology 48 (1997) 649-684. [135] m. sarter, j. p. bruno, brain research reviews 23 (1997) 28-46. [136] l. zaborszky, k. pang, j. somogyi, z. nadasdy, i. kallo, advancing from the ventral striatum to the extended amygdala 877 (1999) 339-367. https://clinicaltrials.gov/ct2/show/nct01608061?term=fornix+dbs&rank=4 jing li et al. admet & dmpk 3(3) (2015) 216-234 234 [137] a. montero-pastor, a. vale-martinez, g. g. guillazo-blanch, m. marti-nicolovius, behavioural brain research 154 (2004) 41-54. [138] s. hescham, l.w. lim, a. jahanshahi, h.w. steinbusch, j. prickaerts, a blokland, y. temel, brain stimulation 6 (2013) 72-77. [139] a. montero-pastor, a. vale-martinez, g. guillazo-blanch, m. marti-nicolovius, behavioural brain research 154 (2004) 41-54. [140] n. boix-trelis, a. vale-martinez, g. guillazo-blanch, d. costa-miserachs, m. marti-nicolovius, learn memory 13 (2006) 783-793. [141] h. hotta, f. kagitani, m. kondo, s. uchida, neuroscience research 63 (2009) 122-128. [142] s.s.d. stone, c.m. teixeira, l.m. devito, k. zaslavsky, s.a. josselyn, a.m. lozano, p.w. frankland, journal of neuroscience 31 (2011) 13469-13484. [143] i. arrieta-cruz, c. pavlides, g. m. pasinetti, translational neuroscience 1 (2010) 188-194. [144] c. hamani, s. s. stone, a. garten, a. m. lozano, g. winocur, experimental neurology 232 (2011) 100104. [145] h. toda, c. hamani, a. p. fawcett, w. d. hutchison, a. m. lozano, journal of neurosurgery 108 (2008) 132-138. [146] h. mccartney, a. d. johnson, z. m. weil, b. givens, hippocampus 14 (2004) 684-687. [147] a. schumacher, a. p. de vasconcelos, l. lecourtier, a. moser, j. c. cassel, behavioural brain research 222 (2011) 368-374. [148] z. m. williams, e. n. eskandar, nature neuroscience 9 (2006) 562-568. [149] a.c. smith, s.a. shah, a.e. hudson, k.p. purpura, j.d. victor, e.n. brown, n.d. schiff, journal of neuroscience methods 183 (2009) 267-276. [150] s. a. shah, j. l. baker, j. w. ryou, k. p. purpura, n. d. schiff, conference proceeding ieee engineering in medicine and biology society. 2009 (2009) 3314-3317. [151] r. g. mair, j. r. hembrook, journal of neuroscience 28 (2008) 14293-14300. [152] c. soriano-mas, d. redolar-ripoll, l. aldavert-vera, i. morgado-bernal, p. segura-torres, behavioural brain research 160 (2005) 141-147. [153] c. soriano-mas, d. redolar-ripoll, g. guillazo-blanch, i. morgado-bernal, p. segura-torres, brain research bulletin 74 (2007) 51-57. [154] g. huguet, l. aldavert-vera, e. kádár, s. peña de ortiz, i. morgado-bernal, p. segura-torres, neuroscience 162 (2009) 359-374. [155] e. kadar, l. aldavert-vera, g. huguet, d. costa-miserachs, i. morgado-bernal, p. segura-torres, genes, brain and behavior 10 (2011) 69-77. [156] h. w. querfurth, f. m. laferla, new england journal of medicine 362 (2010) 329-344. [157] g. deuschl, c. schade-brittinger, p. krack, j. volkmann, h. schäfer, k. bötzel, c. daniels, a. deutschländer, u. dillmann, w. eisner, d. gruber, w. hamel, j. herzog, r. hilker, s. klebe, m. kloss, j. koy, m. krause, a. kupsch, d. lorenz, s. lorenzl, h.m. mehdorn, j.r. moringlane, w. oertel, m.o. pinsker, h. reichmann, a. reuss, g.h. schneider, a. schnitzler, u. steude, v. sturm, l. timmermann, v. tronnier, t. trottenberg, l. wojtecki, e. wolf, w. poewe, j. voges; german parkinson study group, neurostimulation section, new england journal of medicine 355 (2006) 896-908. [158] j. voges, y. waerzeggers, m. maarouf, r. lehrke, a. koulousakis, d. lenartz, v. sturm, journal of neurology, neurosurgery & psychiatry 77 (2006) 868-872. ©2015 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license(http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ solubility-ph profiles of a free base and its salt: sibutramine as a case study ‡ this extended abstract was presented at iapc-6 meeting, zagreb, croatia, september 4-6, 2017 253 admet & dmpk 5(4) (2017) 253-256; doi: http://dx.doi.org/10.5599/admet.5.4.452 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index extended abstract solubility-ph profiles of a free base and its salt: sibutramine as a case study ‡ diego lucero-borja 1 , òscar castilla 1 , rafael barbas 2 , xavier subirats 1 , clara ràfols 1, * 1 institute of biomedicine (ibub) and department of chemical engineering and analytical chemistry, universitat de barcelona martí i franquès, 1-11, 08028 barcelona, spain. 2 polymorphism and calorimetry unit, scientific and technological centers, universitat de barcelona – travessera de les corts, 131-159, 08028 barcelona, spain. *corresponding author: e-mail: crafols@ub.edu; tel.: +34-934-034-874; fax: +34-934-021-233 received: september 29, 2017; revised: november 26, 2017; published: december 24, 2017 abstract in the present study the solubility-ph profiles of sibutramine free base and its hydrochloride salt were determined in the ph range between 2.0 and 9.5 by means of the recommended shake -flask method, and the solids collected were dried and studied by x-ray diffraction in order to elucidate their free base or salt structure. above phmax (or gibbs pka) the solid collected was always identified as free base, whatever the sibutramine species (free base or hydrochloride salt) initially solved. however, in the ph range below ph max different solids were isolated depending on the buffers employed. keywords buffer; free base; ph profile; potentiometry; salt; shake-flask, solubility. introduction in drug formulation it is well known that salts from basic compounds are generally more stable and soluble than their corresponding free bases and, therefore, the salt form is preferred for use as active ingredient in a pharmaceutical form. thus, the study of the solubility of a compound as a function of ph, the solubility-ph profile, provides very useful information about the physicochemical properties of a compound and its rate of success as a drug candidate. in the case of weakly basic active pharmaceutical ingredients, at higher ph values (ph >> pka) the solubility is constant and independent of ph, and the solid phase in equilibrium with the saturated solution is the free base. in contrast, at acidic ph values (ph << pka) the solid in equilibrium with the cationic base is a salt, whose composition depends on the anionic components of buffering species. the particular salt formed is relevant because of its influence on the maximum achievable concentration of the saturated solution. between this acidic and basic ph regions there is a particular ph value, known as phmax or pka gibbs, for which the solution is saturated with both, the free and the cationic species, and the solid in equilibrium contains also free base and salt [1]. the crystalline racemic compound sibutramine hydrochloride monohydrate was used as the active pharmaceutical ingredient of medicines for the management of obesity from 1999 to 2010. its withdrawal http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:crafols@ub.edu clara ràfols et al. admet & dmpk 5(4) (2017) 253-256 254 from european and us markets was due to cardiovascular side effects. however, the physicochemical properties of sibutramine (molecular weight, acidity, solubility, lipophilicity…) make it a very interesting model for the solubility study of basic compounds. experimental the solubility-ph profiles of sibutramine (free base) and its hydrochloride salt have been determined at 25 o c in the ph range between 2.0 and 9.5 by means of the shake-flask method using the recommended protocol [2]. briefly, a mass spectrometry-friendly minimalist universal buffer (ms-mub) consisting of 25 mm acetic acid, 25 mm ethylenediamine, and 75 mm trifluoroacetic acid (tfa) was used, incubation was carried out for 48 h (24 h stirring and 24 h sedimentation), ph was measured after 4 h of stirring and readjusted if necessary with sodium hydroxide or hydrochloric acid, and finally the ph was measured again before the phase separation by centrifugation. sibutramine concentration was determined by hplc with uv detection and the solid was vacuum filtered and dried for 24 h in order to elucidate their free base or salt structure by powder x-ray diffraction (pxrd). with the aim of evaluating the effect of the ms-mub components on the precipitated solid, simple solutions prepared from acetic acid, trifluoroacetic acid, and hydrochloric acid were used. finally, the henderson-hasselbalch (hh) profile was constructed from the spectrophotometrically determined pka [3] and the intrinsic solubility measured by the cheqsol method [4]. results and discussion ph change during stirring step in the shake-flak procedure a universal buffer was selected in the present work in order to keep nearly constant the ionic strength of the buffering solution across the studied ph range, but intentionally avoiding phosphate ions because they are well known to interact with positively charged bases. besides these features, it is always interesting to count on a ms-friendly buffer, since liquid chromatography coupled to ms is a powerful technique for the measurement of low concentrations of poorly soluble substances. therefore, ms-mub was used, providing in the whole ph working range an ionic strength of 96 mm but being thought to be minimally intrusive in relation to anionic buffering species. the ph measurements after 4 hours of stirring revealed significant variations (above 1 ph unit) in relation to the initial ph of the buffer before the addition of the solid, which depends on the nature of starting solid. in fact, there is a remarkable increase in the ph value of acidic solutions (ph < 4) when the free base is used as starting species, and the reverse trend is observed for the hydrochloride salt at basic ph values. in the rest of cases, free base/basic ph and salt/acidic ph, ph variations are not significant during the stirring step. effect of buffer components on solubility-ph profile figure 1 shows the solubility-ph profile of sibutramine according to the buffering species used in the shake-flask procedure, and figure 2 relates the solubility with the solid species initially selected for the experiments (free base or hydrochloride salt) and the solid collected and dried after centrifugation. there is a good agreement between the shake-flask results and the hh profile in the ph range above phmax, independently of the sibutramine form initially weighted and the buffering species employed, and in this ph region the solid collected was always identified by pxrd as the free base. in contrast, a salt was admet & dmpk 5(4) (2017) 253-256 solubility-ph profiles of sibutramine doi: 10.5599/admet.5.4.452 255 figure 1. solubility-ph profile of sibutramine according to the buffering species used in the shake-flask procedure. legend: () ms-mub, () hydrochloric acid (unbuffered), (▲) tfa (unbuffered). figure 2. solubility-ph profile of sibutramine according to the initial weighted solid form and the collected solid after shake-flask procedure identified by pxrd. legend: circles and squares refer to initial species, free base or sibutramine·hcl, respectively; solid, empty, and crossed symbols indicate collected solids as free base, hcl salt, and tfa salt, respectively. expected in the ph range below phmax, but in fact two different solids were isolated depending on the buffers employed. when ms-mub or its single component tfa were used the dried solid was identified as sibutramine·tfa salt, whereas when only hydrochloric acid was employed the corresponding solid was the hydrochloride salt. interestingly, the solubility of the compound in equilibrium with the solid tfa is slightly lower than that of hcl salt. the tfa salt was formed independently of the starting sibutramine species, free base or hydrochloride salt. clara ràfols et al. admet & dmpk 5(4) (2017) 253-256 256 conclusions it is very convenient to periodically measure and readjust (if necessary) the ph of the saturated solution of drug, especially during the first hours of the stirring step. the sign and extent of the variation depends on the initial form of the compound (free base or salt) and the particular working ph. single anionic components of universal buffers, for instance tfa in ms-mub, may interact with positively charged basic drugs leading to unexpected salts in equilibrium with the saturated solution, affecting thus the solubility of the compound. acknowledgements: this work was supported by the ministry of economy and competitiveness of spain (project ctq2014-56253-p). references [1] p.h. stahl, c.g. wermuth, eds., handbook of pharmaceutical salts. properties, selection, and use, verlag helvetica chimica acta, zürich, switzerland, 2002. [2] a. avdeef, e. fuguet, a. llinàs, c. ràfols, e. bosch, g. völgyi, t. verbić, e. boldyreva, k. takácsnovák, equilibrium solubility measurement of ionizable drugs – consensus recommendations for improving data quality, admet dmpk 4 (2016) 117–178. doi:10.5599/admet.4.2.292. [3] k.y. tam, k. takács-novák, multi-wavelength spectrophotometric determination of acid dissociation constants: a validation study, anal. chim. acta. 434 (2001) 157–167. doi:10.1016/s0003-2670(01)00810-8. [4] m. stuart, k. box, chasing equilibrium: measuring the intrinsic solubility of weak acids and bases, anal. chem. 77 (2005) 983–990. doi:10.1021/ac048767n. ©2017 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ doi: 10.5599/admet.4.3.295 212 admet & dmpk 4(3) (2016) 212-231; doi: 10.5599/admet.4.3.295 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index review cardiovascular safety of protein kinase inhibitors: putting their “qt-phobia” in perspective rashmi r shah pharmaceutical consultant, 8 birchdale, gerrards cross, buckinghamshire, uk e-mail: clinical.safety@hotmail.co.uk, tel . + (44) 1753 886348 received: may 17, 2016; revised: july 12, 2016; published: september 30, 2016 abstract many potentially valuable drugs, including protein kinase inhibitors (pki), risk being dropped from further development, without exploration of their clinical benefits, if early studies show these drugs to inhibit herg channel and therefore, to have a potential for prolonging ventricular repolarisation (qt interval). this qtphobia results from a perceived possibility of the clinical risks of qt-related ventricular proarrhythmia, further aggravated by uncertainties surrounding the regulatory evaluation of the risk and either approvability or restrictive labelling of the drug concerned. in reality, qt interval prolongation per se is only an imperfect surrogate of the proarrhythmia risk which is much smaller than perceived and compared to their other cardiovascular and non-cardiovascular risks. pki-induced clinical hepatotoxicity, also evaluated on the basis of surrogate markers (serum transaminases and bilirubin) is another risk that far exceeds any risk arising from pkiinduced qt interval prolongation. this review of the currently approved 28 pkis places the qt-phobia surrounding the development of pkis in its perspective by juxta-positioning their potential to induce ventricular dysfunction, arterial thrombotic events and hepatotoxicity. available evidence suggests that herg channel may prove to be a valuable therapeutic target in oncology. therefore, the development, approval and labelling of such vital oncology drugs requires careful assessment of their benefits and their risk/benefit generally, without being overtly consumed by their potential qt-liability, in terms of their more direct consequences on clinically relevant endpoints of morbidity, mortality and quality of life. keywords arterial thrombotic events; cardiotoxicity; hepatotoxicity; herg channel; left ventricular dysfunction; proarrhythmias, qt interval; transaminases; torsade de pointes; protein kinase inhibitors introduction treatment of various cancers has been significantly revolutionized by the development of small molecule protein kinase inhibitors (pkis), a vast majority of them in current clinical use being tyrosine kinase inhibitors. the perceived benefits of these highly targeted agents have led regulatory authorities to approve a majority of them on an expedited or priority basis, often on limited preliminary data indicative of their safety, efficacy and a favourable risk/benefit ratio [1-3]. often, such approvals are associated with requirements for appropriate post-approval studies to substantiate these early expectations. as of 31 may 2016, 27 such agents have been approved by the us food and drug administration (fda) for use in oncology indications, beginning with the first approval of imatinib in may 2001. two additional pkis have also been approved by the fda but for non-oncology indications – tofacitinib and nintedanib for http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:clinical.safety@hotmail.co.uk admet & dmpk 4(3) (2016) 212-231 pkis and qt-phobia doi: 10.5599/admet.4.3.295 213 the treatments of rheumatoid arthritis and of idiopathic pulmonary fibrosis, respectively. in the european union (eu), however, nintedanib is also approved for non-small cell lung cancer and at the time of writing, one of the 29 agents (alectinib) approved by the fda was still under review, whereas tofacitinib had been rejected for approval, by the european medicines agency (ema). in addition to the development of new pkis, the agents already approved for one cancer are also being tested for their potential therapeutic benefits in other cancers and the range of oncology indications for many of these previously approved agents is gradually expanding [4-7]. for example, imatinib, initially approved for use in chronic myeloid leukaemia, now enjoys no less than seven distinct indications. the approval dates and the broad cancer-specific indications of the agents approved by the fda and/or the ema are shown in table 1. unfortunately, pkis are also associated with a number of adverse side effects, many potentially fatal, on cardiovascular system (e.g. qt interval prolongation of the surface electrocardiogram (ecg), hypertension, left ventricular (lv) dysfunction, arterial thrombotic events (ates), venous thrombo-embolic events, bleeding and symptomatic bradycardia) and non-cardiovascular systems (such as hepatic, pulmonary, dermatological thyroid, ocular and renal) [8]. many of these adverse effects are on-target effects which correlate with efficacy and therefore, difficult to avoid without compromising therapeutic benefit by simple dose reductions [8-17]. it is beyond the scope of this review to discuss the full safety profile of pkis and the interested reader is referred to other reviews [2, 3, 18-39]. despite this constellation of side effects associated with pkis, the one that has engendered a phobia, often leading to early termination from further development, is the potential of a pki to prolong the qtc interval. in rare cases, prolonged qt interval may predispose the patient to potentially fatal ventricular tachyarrhythmias. the fda guidance on pre-marketing risk evaluation and assessment recommends the sponsor to address, as a part of the new drug application (nda) for all new small molecule drugs, their potential for six serious adverse effects, one of them being drug-related qtc prolongation [40]. the others are drug-related liver toxicity, nephrotoxicity, myelotoxicity, interactions with other drugs and polymorphic drug metabolism. this review aims to put in perspective the “qt-phobia” that surrounds the development of pkis by highlighting just two of their other more serious and relatively more frequent cardiovascular effects, namely lv dysfunction and ates. it will focus on 28 oncology-related pkis including nintedanib which enjoys an oncological indication in the eu. for the purpose of this review, these 28 pkis approved during the last 15-years since 2001 are divided into two groups with reference to the dates of their first approval by the us fda. the dates selected are the fda approval date because typically, the fda is also the authority that has led the way in first approvals of each agent [1]. one group includes 16 (57 %) agents approved over 11.5 years from may 2001 to september 2012 (henceforth referred to as period 1) and the other includes 12 (43 %) agents approved more recently during the last 3.7 years from october 2012 to may 2016 (period 2). the pace of development of these agents is self-evident; 1.39 agents per year in period 1 and 3.24 agents per year in period 2. the labels of approved pkis continue to evolve as more data become available and therefore, the information reviewed herein is sourced from a variety of documents referred to in earlier reviews [1-3], the most current updated labels and the author’s overall assessment of the data available. rashmi shah admet & dmpk 4(3) (2016) 212-231 214 table 1. approval dates and current indications of protein kinase inhibitors (pkis) as of 31 may 2016 pki fda approval date eu approval date broad-term indication(s) (for use in selected patients with) afatinib 12 july 2013 25 sept. 2013 non-small cell lung cancer alectinib 11 dec. 2015 under review non-small cell lung cancer axitinib 27 january 2012 13 sept. 2012 renal cell carcinoma bosutinib 4 sept. 2012 27 march 2013 cml cabozantinib 29 nov. 2012 21 march 2014 thyroid cancer, renal cell carcinoma ceritinib 29 april 2014 6 may 2015 non-small cell lung cancer cobimetinib 10 nov. 2015 20 nov. 2015 melanoma crizotinib 26 august 2011 23 oct. 2012 non-small cell lung cancer dabrafenib 29 may 2013 26 august 2013 melanoma dasatinib 28 june 2006 20 nov. 2006 cml, all erlotinib 18 nov. 2004 19 sept. 2005 non-small cell lung cancer, pancreatic cancer gefitinib 5 may 2003 20 june 2009 non-small cell lung cancer ibrutinib 13 nov. 2013 21 oct. 2014 mantle cell lymphoma, cll, waldenström’s macroglobulinaemia imatinib 10 may 2001 11 nov. 2001 cml, all, gist, mastocytosis, hypereosinophilic syndrome, myelodysplastic/myeloproliferative disease, dermatofibrosarcoma protuberans lapatinib 13 march 2007 10 june 2008 breast cancer lenvatinib 13 february 2015 28 may 2015 thyroid cancer nilotinib 29 october 2007 19 nov. 2007 cml nintedanib 17 october 2014 (not applicable) 15 jan. 2015 21 nov. 2014 idiopathic pulmonary fibrosis non-small cell lung cancer osimertinib 13 nov. 2015 2 feb. 2016 non-small cell lung cancer pazopanib 19 october 2009 14 june 2010 renal cell carcinoma, soft tissue sarcoma ponatinib 14 dec. 2012 1 july 2013 cml, all regorafeniib 27 sept. 2012 26 august 2013 colorectal cancer, gist ruxolitinib 16 nov. 2011 23 august 2012 myelofibrosis, polycythaemia vera rubra sorafenib 20 dec. 2005 19 july 2006 hepatocellular carcinoma, renal cell carcinoma, thyroid cancer sunitinib 26 january 2006 19 july 2006 renal cell carcinoma, gist, pancreatic neuroendocrine tumour trametinib 29 may 2013 30 june 2014 melanoma vandetanib 6 april 2011 17 feb. 2012 thyroid cancer vemurafenib 17 august 2011 17 feb. 2012 melanoma tofacitinib 6 november 2012 refused on 26 april 2013 rheumatoid arthritis all = acute lymphoblastic leukaemia; cml = chronic myeloid leukaemia; gist = gastrointestinal stromal tumours; cll = chronic lymphocytic leukaemia admet & dmpk 4(3) (2016) 212-231 pkis and qt-phobia doi: 10.5599/admet.4.3.295 215 background to qt-phobia the qt interval measured on the ecg varies with the heart rate and therefore requires correction to compute a heart rate-corrected qt interval (qtc interval) to determine the effect of an intervention. the correction formula most widely used for regulatory submissions and safety assessment is the fridericia, study population-specific or individual subject-specific correction formula to compute qtcf, qtcp or qtci interval, respectively. qtc interval prolongation per se is not harmful. however, at the level of an individual patient, prolonged qtc interval, when excessive (typically significantly greater than 500 ms) and/or in presence of certain risk factors, induces a potentially fatal form of ventricular tachyarrhythmia, known as torsade de pointes (tdp) [41,42]. although typically of short duration, asymptomatic and self-terminating, tdp is infrequently sustained and symptomatic in a small proportion of cases and in 20 % or so of these patients, it degenerates into ventricular fibrillation, leading to death [43,44]. historically, this proarrhythmia (or the potential thereof) has been responsible for about a third of the drug withdrawals from the market [45, 46]. the approval of many others has been either delayed or associated with prescribing restrictions in terms of narrow indication(s), contraindications and/or warnings and monitoring precautions [47]. arising from this historical experience with a large number of drugs across a wide range of pharmacological and therapeutic classes including oncology, the potential of new drugs to prolong the qtc interval and induce tdp is one principal safety issue that has continued to worry the regulatory authorities over the last two decades, although qtc interval prolongation per se does not correlate well with torsadogenic potential. this imperfect correlation between qt prolongation and risk of tdp is explained by the fact that whereas qt prolongation results from blockade of human ether-a-go-go (herg) channel and prolongation of action potential duration (apd), tdp is a triggered activity arising from early afterdepolarizations (eads) and the risk of eads followed by tdp is modulated by other ancillary properties of the drug, such as blockade of calcium or late sodium currents or autonomic receptors [48-51]. ranolazine (an antianginal drug) and sertindole (a neuroleptic agent) best illustrate the role of late sodium current and α-blocking activities, respectively, in modulating the risk of tdp following qt interval prolongation [52-56]. in essence, therefore, tdp and fatality following drug-induced qtc prolongation are events that are much rarer than would be anticipated. for example, laksman et al. [57] have reported that, among 172 inhospital patients with qtc interval > 550 ms, in-hospital mortality was 29 %, with only 4 % of patients experiencing arrhythmic deaths, all of which were attributed to secondary causes. in a vast majority of cases, drug-induced prolongation of the qtc interval results from delayed ventricular repolarization due to inhibition of the (major) outward repolarizing potassium current mediated by herg subunit of the rapid component of the delayed rectifier channel (ikr) [58]. non-antiarrhythmic drugs can be characterised for this off-target undesirable pharmacological property in nonclinical in vitro and in vivo studies as well as clinical in vivo studies. not surprisingly, therefore, regulatory authorities have reacted to this potential by promulgating guidelines for pre-approval characterisation of all new drugs with systemic bioavailability for their “qt-liability” [47, 59-62]. this is the only adverse drug effect that has called for specifically targeted and internationally harmonised detailed guidelines [60, 61]. clinically, this qt-liability is determined in a formal and dedicated study, referred to as the thorough qt (tqt) study, that includes two dose levels of the investigational drug, a placebo and an active control. in order to avoid any false negatives and as a matter of precaution, the threshold of regulatory concern for a maximum mean increase in qtc interval from baseline in the tqt study has been (arbitrarily) set low at around 5 ms as evidenced by an upper bound of the 95 % confidence interval (ci) around the mean effect on qtc of 10 ms rashmi shah admet & dmpk 4(3) (2016) 212-231 216 [61]. however, qtc prolongation is an imperfect surrogate of torsadogenesis and therefore, it is acknowledged by all concerned (including the regulatory authorities) that a drug endowed with qt-liability does not necessarily mean that the drug is proarrhythmic; a breach of the above regulatory threshold simply calls for more diligent and intensive ecg monitoring in subsequent clinical trials. nevertheless, the sponsors have frequently reacted by terminating further development of new agents, often on the basis of an early in vitro herg study or first-in-human study, because of the fears and regulatory uncertainties surrounding the possible finding of a modest but often clinically irrelevant breach of the regulatory threshold by the drug concerned in what is a cost-ineffective and resource-intensive tqt study. qtc-liability of protein kinase inhibitors overall, as shown in table 2, 11 (39 %) of the 28 pkis have the potential to prolong the qtc interval to a variable extent. these 11 agents are considered to have the qt-liability on the basis that their qtc effect had breached the above very conservative threshold of regulatory concern; as a result, the labels of two (nilotinib and vandetanib) carry a black box warning. interestingly, qtc-liability is reported for eight (50 %) of the 16 agents approved in period 1 in contrast to only 3 (25 %) of the 12 approved in period 2, possibly suggesting early termination of pkis from development during period 2 due to this pharmacological property. however, experience with qtc-prolonging drugs that are known to be torsadogenic and nontorsadogenic has shown that drugs associated with maximum mean placebo-corrected increases of 6-10 ms are unlikely to be torsadogenic whereas drugs associated with increases of 16-20 ms are probably torsadogenic [63]. lin and kung have also summarized data which suggest that a mean qtc increase of 19.3 ms associated with strong torsadogens is significantly greater than the 8.0 ms for borderline torsadogens [64]. by comparison with known non-pki torsadogens, the qtc effect of majority of the pkis that prolong the qtc interval (often determined at supratherapeutic doses) appears to be relatively mild/modest with a maximum mean effect of around 10 ms (95 % upper bound around their mean effects being around 15 ms). six pkis (osimertinib, ceritinib, sunitinib, lapatinib, nilotinib and vandetanib) appear to have somewhat larger effects with a maximum mean effect ranging from 13.5 to 34.7 ms (95 % upper bound around the mean effect being 17.6, 22.2, 22.4, 23.4, 25.8 and 36.4 ms respectively) [2,3]. however, since some of these effects are not placebo-corrected, the true extent of the drug’s qtc-prolonging effect remains uncertain. therefore, it is interesting to note that the number of patients receiving any of these pkis who developed a qtc interval prolongation > 500 ms is typically well below 0.5 % except lapatinib (6 %) and vandetanib (4.3 %) [2]. in pre-approval clinical trials, only pazopanib and vandetanib were each associated with two cases of tdp, although the causal association with pazopanib in one case appeared uncertain since the patient was also receiving amiodarone [2]. in a dedicated qt study as required by the regulatory guidance, the maximum mean increase in baseline-adjusted, time-matched qtcf interval in pazopanibtreated patients versus placebo was only 4.4 ms (95 % upper bound 11.2 ms) [65]. mixed-effects modelling did not indicate a significant concentration-dependent effect of pazopanib or its metabolites on qtcf interval. admet & dmpk 4(3) (2016) 212-231 pkis and qt-phobia doi: 10.5599/admet.4.3.295 217 table 2. trend analysis of selected toxic effects of protein kinase inhibitors (pkis) with reference to approval period protein kinase inhibitor period 1 (may 2001 – september 2012) number of pkis approved = 16 period 2 (october 2012 – may 2016) number of pkis approved = 12 qt effect hepatotoxicity lv dysfunction arterial effects qt effect hepatotoxicity lv dysfunction arterial effects afatinib **   alectinib **  axitinib *     bosutinib *   cabozantinib**   ceritinib **   cobimetinib**   crizotinib *     dabrafenib **   dasatinib *   erlotinib *   gefitinib *   ibrutinib ** imatinib *     lapatinib *      lenvatinib **    nilotinib *        nintedanib **   osimertinib **   pazopanib *      ponatinib **      regorafeniib *    ruxolitinib * sorafenib *     sunitinib *        trametinib **  vandetanib *     vemurafenib *   pkis with the effect 50 % 75 % 63 % 63 % 25 % 58 % 58 % 33 % * pki approved in period 1; ** pki approved in period 2;  with black box warning;  post-approval addition or revision the rarity of qt-related proarrhythmia risk of pkis is further corroborated by post-marketing data. during the post-marketing use of 16 pkis approved in period 1 and for which some data are available, there were a total of 463 reports of qt interval prolongation but only 20 reports of tdp; only vandetanib was associated with significant risk of tdp [3]. spontaneous reporting data must of course be interpreted with great caution but although crizotinib, dasatinib, vandetanib and nilotinib were associated with 14, 35, 57 and 246 reports of qtc interval prolongation during their post-marketing use, the corresponding number of reports of tdp were only 0, 1, 2 and 3, respectively, thus further emphasising that not all events rashmi shah admet & dmpk 4(3) (2016) 212-231 218 of qt interval prolongations degenerate into tdp and the rarity of this potentially fatal proarrhythmia. although there were a total of 220 reports of sudden death, there was similar discordance between the qt-liability of a drug and the number of reports of sudden death associated with it. kloth et al. [66] have also reported a post-marketing observational study of 363 patients who were eligible for the analyses of qtc interval before and during treatment with erlotinib, gefitinib, imatinib, lapatinib, pazopanib, sorafenib, sunitinib or vemurafenib. the median on-treatment time before the ecg was performed was 43 days. mean (range) qtc intervals were 401 (388–415) ms at baseline and 415 (397–431) ms following therapy. a total of 33 patients (9.1 %) were characterised by an increased grade of common terminology criteria for adverse events (ctcae). only two individuals passed from grade 1 to grade 2 or 3 and nine patients (2.5 %) had a decrease in their ctcae grade for qtc interval. however, 321 (88.4 %) patients did not have a change in their ctcae grade after the start of pki treatment. only five patients (1.4 %) developed qtc  500 ms after starting the therapy, with all of them experiencing an increase of 100 ms from baseline. despite this marked increase from baseline, none was reported to have a proarrhythmia. one potential explanation for the above discordance, also observed all too often with drugs in other pharmacological or therapeutic classes, is that the pkis concerned may have a risk-mitigating inhibitory effect on other ion currents involved in apd, such as the calcium and/or late sodium currents as explained earlier [48-51]. whether or not pkis generally inhibit these ion channels is at present unclear since they are not routinely studied for these effects. protein kinases activate downstream signalling via phosphoinositide 3-kinase (pi3k). the complexity of the interactions is demonstrated by one study reporting that chronic but not acute exposure to nilotinib prolonged the apd in canine cardiomyocytes, an effect which was reversed by intracellular dialysis with phosphatidylinositol 3,4,5-trisphosphate (which is a downstream effector of pi3k) [67, 68]. these investigators also observed that apd prolongation by nilotinib was mediated by decreases in both the rapid (ikr) and slow (iks) components of the delayed rectifier potassium current, as well as by an increase in late sodium current (an effect which in theory ought to aggravate the proarrhythmia risk [69]); however, nilotinib also reduced the calcium current. one obvious and hitherto unquantified by-product of the “qt-phobia” is that many new candidate compounds are abandoned very early in their development without fully exploring their potential for significant therapeutic benefit. the unfortunate reality, however, is that the “qt-liability” of a drug is only an imperfect predictor of the proarrhythmia risk, and just one component of its much wider cardiovascular and overall safety. of equal concern with regard to determining the duration of qtc interval in routine clinical practice is the fact that the qtc interval value depends on the correction formula applied to the measured qt interval. the most common formula applied in routine clinical practice is the bazett’s formula (that computes qtcb interval) which has been shown to be highly imprecise when there are significant changes in heart rate [70-73]. although regulatory documents and drug labels provide guidance on dose adjustment in patients who develop, or are at risk of developing qtc prolongation, they provide no guidance on the formula to be used. similarly, the us national cancer institute, which has recently modified the criteria for qtc interval in the latest version of ctcae (v4.03) [74], provides no guidance on the choice of a formula for computing qtc interval duration. it is therefore timely to contrast and put in perspective the foregoing rarity of adverse clinical outcomes from pki-related effects on qtc interval with their effects on two specific cardiovascular effects, lv dysfunction and ates, which are of much greater concern and have a significant direct impact on morbidity, mortality and risk/benefit. admet & dmpk 4(3) (2016) 212-231 pkis and qt-phobia doi: 10.5599/admet.4.3.295 219 left ventricular dysfunction due to protein kinase inhibitors left ventricular (lv) dysfunction is now a well-recognised toxicity of a number of pkis. it can range from asymptomatic ecg changes (in qrs and t-waves and st-segment) through decrease in lv function (detectable noninvasively only by echocardiography or radionuclide techniques) to clinically manifest severe congestive heart failure (chf) and generalised fluid retention [75-77]. although not precisely known or understood, the potential mechanism underpinning this effect has been reviewed previously by others [2, 78-81]. post-approval, the us labels of a number of pkis have been revised well after their initial approval to include new warnings and precautions regarding the risks of lv dysfunction and/or fluid retention. these are axitinib, dabrafenib, imatinib, nilotinib and pazopanib, although there are some question marks concerning the potential of imatinib to induce cardiac dysfunction. during its post-marketing use as of 25 february 2015, about 14,700 patients worldwide had received crizotinib since its approval [82]. forty cases of cardiac failure had been reported in the post-marketing setting. a review by european medicines regulators of data from clinical trials and reports from clinical practice concluded that this side effect is common, occurring in 1-10 % of the patients taking crizotinib [82]. in most cases, cardiac failure occurred within one month of starting the treatment and affected patients with or without pre-existing cardiac disorders. the reports included cases where the evidence of symptoms of cardiac failure resolved after stopping crizotinib and recurred when it was reintroduced. some cases had a fatal outcome. it is noteworthy that in contrast to the revised european union prescribing information, the most current fda label of crizotinib (dated 29 april 2016) warns about symptomatic bradycardia but does not make any reference to cardiac failure [83]. currently, 17 (60 %) of the 28 pkis, approved as of 31 may 2016, are known to induce lv dysfunction, often loosely termed as “cardiomyopathy” or “cardiotoxicity”, and cardiac failure. it is evident that 10 (63 %) were among the 16 approved in period 1 and 7 (58 %) among the 12 approved in period 2. this trend is in contrast to the substantial reduction in the number of qt-prolonging drugs approved during period 2 compared to period 1 (from 50 % to 25 % for qtc-prolonging drugs versus from 63 % to 58 % for drugs affecting cardiac dysfunction). not only that but also the total number of pkis with an adverse effect on cardiac function (n=17; 60 %) is much higher than those with a qt-liability (n=11; 39 %). the complexity surrounding the estimation of incidence and interval to onset has been reviewed before since these depend on the dose of the drug and the criteria used to define lv dysfunction as well as previous chemotherapy. in broad terms, asymptomatic decrease in lv ejection fraction is by far the most frequent. however, overt chf is not uncommon with an incidence as high as 5-8 % with sunitinib and pazopanib. a number of recent meta-analyses of clinical trials have quantified the risk of pki-induced symptomatic cardiac failure [21,22,84]. these events are typically categorised as alland high-grade (grade 3 and higher) as defined by ctcae. grade 3 chf events require intervention, and grade 4 chf events usually include life-threatening dysfunction. these meta-analyses reveal that overall, the incidences of allgrade and high-grade chf are about 2.8 % and 1.1 %, respectively, with a relative risk (rr) of 2.5 and 1.5, respectively. the risk appears independent of tumour type or the pki used. the outcome has been fatal in a number of cases whereas others require anti-failure therapy and/or withholding the culprit pki. in routine clinical oncology, the scale of the problem is likely much higher as patients with cardiac disease who are the most at risk of cardiotoxicity have traditionally been excluded from pre-approval clinical trials. as stated above, four (23 %) of the 17 pkis with a potential for inducing cardiac failure revealed this risk rashmi shah admet & dmpk 4(3) (2016) 212-231 220 during their post-marketing use. currently, there are no regulatory guidelines for pre-approval characterisation of a drug for its effect on lv function. however, as yang and papoian from the fda point out, these reports of cardiac toxicity following the clinical use of pkis are unexpected and not well predicted by nonclinical studies [85] and clinical findings have exposed gaps in current nonclinical drug testing for predicting the development of cardiac toxicities in humans. they have suggested the use of cultured cardiomyocytes and application of isolated perfused heart methodology to chronic or sub-chronic rodent studies or including echocardiography in chronic large animal toxicity studies as potentially valuable tools to study the likelihood of pki-mediated cardiotoxicity. available data suggest that nonclinical studies, when adequately designed to address a specific issue, are capable of detecting these effects as was seen in studies with some pkis approved more recently. for example, in vivo cardiac safety pharmacology studies with osimertinib did suggest equivocal findings of decreased contractility dogs and guinea pigs whereas increased findings of cardiomyopathy compared to findings in control animals were reported in the 4-week rat toxicology study with cobimetinib though no other cardiac effects were reported in animal studies [86, 87]. with regard to afatinib, a decrease in left ventricular function was noted at a dose of 30 mg/kg in a single continuous intravenous administration study in domestic pigs [88]. use of echocardiography unequivocally identified sunitinib-induced cardiac lv dysfunction but failed to detect an effect of bosutinib [89, 90]. although electrocardiography is widely used to study qtc effects, echocardiography has not been used as frequently to study lv function in nonclinical studies. there clearly lies a significant challenge in pre-approval characterisation of a drug for its effect on lv function. recent studies have shown that a multi-parameter approach examining both cardiac cell health and function in human induced pluripotent stem cell-derived cardiomyocytes provides a comprehensive and robust assessment that can aid in determining potential cardiotoxic liability [91]. arterial thrombotic events due to protein kinase inhibitors among the most serious adverse effects of the pkis are their prothrombotic effects as illustrated by ponatinib [reviewed in 3]. following the post-marketing experience with ponatinib (see below), ates have now emerged as a major safety concern with pkis and the previously approved labels of a few have required revisions to reflect this risk. ten (63 %) of the 16 pkis approved during period 1 have been reported to be associated with arterial thrombosis (axitinib, dasatinib, erlotinib, imatinib, nilotinib, pazopanib, regorafenib, sorafenib, sunitinib and vandetanib). among the 12 newer pkis approved in period 2, four (33 %) have been reported to be associated with ates (cabozantinib, lenvatinib, nintedanib, and ponatinib, with the last one carrying a black box warning) [3]. the rates of ates appear to be comparable for pkis from both periods and the ate events typically include cerebral infarction, cerebral ischaemia, cerebrovascular accidents (cva), myocardial infarction and myocardial ischaemia. in pre-approval clinical trials, the incidences of ates were 5 % with lenvatinib, 2.5 % for nintedanib and 2 % for cabozantinib, the corresponding placebo rates being 2 %, 0.8 % and 0 % respectively. thus, the mean drug-emergent effect was of the order of 2.2%. myocardial infarction was the most common adverse reaction, occurring in 1.5 % of nintedanib-treated patients compared with 0.4 % of placebo-treated patients. the incidence of grade 3 or higher events was 3 % in lenvatinib-treated patients compared with 1 % in the placebo group. one meta-analysis involving 10,255 patients receiving sunitinib and sorafenib revealed an incidence of ates to be 1.4 % (95 % ci: 1.2–1.6) [23]. another meta-analysis involving a total of 9,711 patients from 19 trials concluded that the overall incidence of ates was 1.5 % (95 % ci: 1.0–2.3) following the use of vegfr admet & dmpk 4(3) (2016) 212-231 pkis and qt-phobia doi: 10.5599/admet.4.3.295 221 pkis [24]. the most common ates were myocardial ischaemia/infarction (67.4 %), central nervous system ischaemia (7.9 %) and cva (6.7 %). the odds ratio (or) was significantly increased when compared with controls (or 2.26, 95 % ci: 1.38–3.68; p = 0.001) and this did not vary significantly with tumour types (p = 0.70), vegfr pkis (p = 0.32), treatment regimens (p = 0.76), phase of trials (p = 0.37) and sample size (p = 0.89). thus, the overall incidence of ates with vegfr inhibitors is about 1.50 % with an rr in the region of 2.6. available limited evidence suggests that egfr inhibitors may not be associated with this risk. in a post-marketing observational study by srikanthan et al. [92], three agents were studied: erlotinib, sorafenib and sunitinib. patients treated with these agents were compared to 128,415 age and gender matched individuals without cancer who served as controls. of the 1,642 pki-treated patients followed up, 1.1 % developed a myocardial ischaemic event requiring hospitalization, 0.7 % developed a cva requiring hospitalization and 1,184 (72.1 %) died. 61 % of the myocardial ischaemic event events and 73 % of the cva events were associated with erlotinib, and these proportions closely mirrored the relative frequency of drug use in the population. cardiovascular events predominantly occurred late in follow-up. when patients with and without baseline ischaemic heart disease were compared, 3.3 % versus 0.5 %, respectively, developed a myocardial ischaemic event and 1.2 % versus 0.5 %, respectively, developed cva. however, the mortality rates were no different (72.5 % vs. 72.0 %, respectively). these investigators summarised that individuals treated with pkis have a significantly higher hazard of death relative to the general population but cause-specific hazards of ischaemic heart disease and cerebrovascular accidents are not increased. jang et al. [93] have recently reported cardiovascular events in 171 (26 %) of the 670 patients aged > 65 years treated with sunitinib or sorafenib. the incidence rates for chf or cardiomyopathy, acute myocardial infarction and stroke were 0.87, 0.14, and 0.14 per 1000 person-days, respectively. the use of either agent was associated with an increased risk of cardiovascular events (hazard ratio (hr) 1.38; 95 % ci: 1.02-1.87) and stroke (hr 2.84; 95 % ci: 1.52-5.31) in comparison with 788 patients diagnosed with advanced renal cell carcinoma who did not receive either agent. case of ponatinib when first approved in december 2012, ponatinib was already strongly associated with cardiovascular, cerebrovascular, and peripheral vascular thrombosis, including fatal myocardial infarctions and strokes [94]. overall, fifty-one (11 %) of the 449 patients experienced an arterial thrombotic event of any grade. myocardial infarction or worsening coronary artery disease was the most common event which occurred in 21 patients (5 %) of ponatinib-treated patients. the range of events included congestive heart failure (concurrent or subsequent to the myocardial ischaemia), cerebrovascular events, hemorrhagic conversion of the initial ischemic event, stenosis of large arterial vessels of the brain and peripheral arterial events with digital or distal extremity necrosis. serious arterial thrombosis occurred in 34 (8 %) of the 449 of ponatinib-treated patients, as a consequence of which 21 patients required various revascularization procedures. thirty of these 34 patients had one or more of the well recognised cardiovascular risk factors. patients with cardiovascular risk factors appeared to be at increased risk for arterial thrombosis following treatment with ponatinib. as a condition of approval, the fda had required the sponsor to characterize the safety of ponatinib and submit long-term safety data over a follow-up period of at least 12 months from all ongoing patients in a specific randomized controlled trial (referred to as ap24534-12-301) that adequately isolated the effect of the drug [95]. however, just over 10 months later as of 31 october 2013, approximately 24 % of patients in one phase ii clinical trial (median treatment duration 1.3 years) and approximately 48 % of patients in a phase i clinical trial (median treatment duration 2.7 years) had experienced serious adverse vascular events [96]. a rashmi shah admet & dmpk 4(3) (2016) 212-231 222 number of these patients required urgent surgical procedures to restore blood flow. in some patients, fatal and serious adverse events occurred as early as 2 weeks after starting ponatinib therapy. since a safe dose or duration of exposure could not be identified, the sponsor agreed to the request from the fda to suspend marketing and sales of ponatinib [96]. following a thorough assessment of all available data, the fda on 20 december 2013 required several new safety measures to be implemented, including restricted indication and additional warnings and precautions, before resumption of marketing to appropriate patients [97]. the fda also required a risk evaluation and mitigation strategy (rems) and the sponsor to conduct post-marketing investigations to further characterise the dose and the safety of the drug. the most current label for ponatinib, dated 2 june 2016 [98], includes a more detailed account of vascular occlusion with age-related increase in incidence (18 %, 33 % and 56 % for those aged  49, 50-74 and  75 years, respectively and prior history of risk factors and 12 %, 18 % and 46 % respectively for those without prior risk factors. the overall incidence is computed to be 24 % for vascular occlusion and 20 % for arterial thrombosis and occlusion. it was also found to cause renal artery stenosis, associated with worsening, labile or treatment-resistant hypertension in some patients. however, the dose that was approved initially has remained unchanged over time at 45mg once daily [94, 98]. leaving aside the rather atypical example of ponatinib, it is still sufficiently evident that pki-induced ates, rather than their qt-liability, have a far greater impact on quality of life, morbidity and mortality of patients receiving pki therapy. protein kinase inhibitors and fatalities due to non-qt related cardiovascular adverse events from the foregoing, it is not surprising that non-qt related cardiovascular effects feature heavily in various studies of fatal adverse events (fae) associated with pkis, especially those agents that target angiogenesis (vegfr inhibitors). vegfr inhibitors three large and independent meta-analyses [99-101], summarised in table 3, involving 12,870 patients receiving vegfr inhibiting pki and 11,114 control patients have quantified the incidence of fae to be in the range of 1.50-2.26 % with a rr in the range of 1.64-2.23. the pkis studied were sunitinib, sorafenib, pazopanib, vandetanib, axitinib, cabozantinib, lapatinib and regorafenib. adverse effects such as cardiac failure, myocardial infarction, stroke, thrombo-embolism, haemorrhage and cardiopulmonary insufficiency feature prominently in the list of these faes. other often reported events included hepatic failure, respiratory events, intestinal perforation and renal failure as well as sudden death. yang et al. [102] have recently reported that patients treated with sorafenib had a significantly greater risk of mortality than those in placebo/control groups, with an rr of 1.75. among different vegfr-pkis, sorafenib and sunitinib were associated with a significant risk of death when compared with control arms, respectively. combination of vegfr-pkis with other antineoplastic agents, but not vegfr-pki monotherapy, significantly increased the risk of treatment-related deaths. subgroup analyses have revealed some interesting results. whereas schutz et al. [100] reported no difference in the rate of faes found between different vegfr pkis or tumour types, hong et al. [101] reported a significantly increased risk of death in patients with non-small cell lung cancer and colorectal cancer. this indication-related increased risk of death following treatment with sorafenib has also been reported by yang et al. [102] who found an overall incidence of sorafenib-associated mortality to be 3.3 % admet & dmpk 4(3) (2016) 212-231 pkis and qt-phobia doi: 10.5599/admet.4.3.295 223 but patients with renal cell carcinoma and thyroid cancer had treatment-related mortality ≥5 %. recently, jean et al. [103] have also reported a distinct increase in the rate of occurrence of adverse effects of sorafenib when used in differentiated thyroid cancer compared with renal and hepatocellular cancer. while many theoretical explanations have been advanced to explain this indication-related difference in toxicity profile, the exact mechanism for this remains unclear. table 3. summary of three meta-analyses of fatal adverse events (faes) associated with protein kinase inhibitors (pkis) study pki group n = control group n = pkis risk (95% ci) incidence (%) cv events ref. sivendran et al. 2,762 2,401 sunitinib sorafenib pazopanib vandetanib rr = 1.64 (1.16 – 2.32) p = 0.01 2.26 vs 1.26 cardiac failure pulmonary embolism [99] schutz et al. 2,461 2,218 sunitinib sorafenib pazopanib rr = 2.23 (1.12 – 4.44) p = 0.023 1.50 haemorrhage myocardial infarction cardiac failure stroke pulmonary embolism sudden death [100] hong et al. 7,644 6,495 axitinib cabozantinib lapatinib pazopanib regorafenib sunitinib sorafenib vandetanib or = 1.85 (1.33 – 2.58) p = 0.01 1.90 cardiopulmonary insufficiency thrombo-embolism haemorrhage sudden death [101] egfr inhibitors in contrast to vegfr-pkis, qi et al. [104] reported a meta-analysis of 7,508 patients treated with two widely used egfr-pkis (erlotinib and gefitinib) and compared them with 6,317 control patients to determine the incidence and risk of faes. the overall incidence of faes was 1.9 % (95 % ci: 1.2 2.9), and the rr was 0.99 (95 % ci: 0.70 1.41; p = 0.97). no increase in faes was detected in any pre-specified subgroup. this analysis suggests that the use of egfr-pkis does not increase the risk of faes in patients with advanced solid tumours, and egfr-pkis are safe and tolerable by cancer patients, especially for those previously treated patients. in the context of this finding, it is worth noting that typical egfr-pkis are not known to induce cardiovascular adverse events that are so typical of vegfr-pkis, thus emphasising the impact of these events on the risk/benefit of pkis. pki in combination with cytotoxic agents in a more comprehensive meta-analysis of 43 trials involving 16,011 patients (8,460 on pkis and 7,551 controls) that balanced the risks versus the benefits of pkis active at vegfr and egfr, funakoshi et al. [105] evaluated the safety and efficacy of combining cytotoxic chemotherapy with pkis, which were divided into two subgroups: vegfr pki (axitinib, cabozantinib, pazopanib, regorafenib, sorafenib, sunitinib and vandetanib), and egfr-family pki (erlotinib, gefitinib and lapatinib). they found that compared with chemotherapy alone, the addition of a pki was associated with a significant improvement in progression free survival (hazard ratio 0.82; 9 5% ci: 0.76–0.89), but not overall survival (hazard ratio 0.99; 95 % ci: 0.95–1.03). however, the addition of a pki significantly increased the risk of faes (rr = 1.63, 95 % ci: 1.32– 2.01), treatment discontinuation (rr = 1.80, 95 % ci: 1.58–2.06), and any severe adverse event (rr = 1.25, 95 % ci: 1.16–1.36). surprisingly, the rr associated with addition of a pki was 1.49 (95 % ci: 1.16-1.90) with rashmi shah admet & dmpk 4(3) (2016) 212-231 224 vegfr inhibitors compared to 2.04 (95 % ci: 1.38-3.01) with egfr inhibitors. these findings serve to caution the physicians to weigh the risk of toxicity versus the modest benefit in terms of progression free survival associated with chemotherapy plus a pki in patients with solid cancers. safety of protein kinase inhibitors in routine clinical practice in routine clinical practice, however, the risk/benefit is likely to be inferior to that determined from highly controlled clinical trials. whereas patients in pre-approval clinical trials are carefully selected, treatment of wider and less selected patient population in routine oncologic practice may increase the likelihood of toxicity and lower the probability of benefit [106]. the problem may be further compounded by off-label use of oncology drugs. one study from switzerland reported a total of 985 consecutive patients receiving 1,737 anticancer drug treatments and found that overall, 32.4 % of all patients received at least one off-label drug, corresponding to 27.2 % of all anticancer drugs administered [107]. following a detailed analysis of the adverse drug reactions of targeted anticancer agents from their reporting in pivotal randomized clinical trials and subsequently updated drug labels, seruga et al. [108] concluded that many rare but serious and potentially fatal adverse drug reactions associated with these agents are not reported in clinical trials. one study on cancer drugs in japan reported that of the 111 fatal adverse drug reactions detected in the eight post-marketing surveillances, only 28 (25.0 %) and 22 (19.6 %) were described on the initial global and the initial japanese drug label, respectively, and 58 (52.3 %) fatal adverse drug reactions were first described in the all-case post-marketing surveillance reports [109]. herg blockade in oncology – a friend or a foe? most of the 14 drugs withdrawn from the market due to their qt-related proarrhythmic proclivity were either old drugs with more effective newer alternatives or were indicated for relatively benign indications. as a result, their risk/benefit was considered unfavourable and it was prudent to have withdrawn them form the market. however, for the pkis, the risk/benefit components are different. not only are they indicated for life-threatening conditions with potentially no alternatives but also, their post-marketing performance suggests that the risk of qt-related proarrhythmia may have been over-estimated. critically, as explained earlier, the risk assessment is based on a parameter (qt prolongation) that is known to be a poor predictor of the risk. if pkis can be approved and continue to be marketed and used clinically despite clinically the most relevant risks of lv dysfunction and arterial thrombotic events, there seems no reason why they should be abandoned early during the course of their development because of the phobia about qt-related over-estimated risks of proarrhythmias. since both ich s7b and ich e14 are principally “herg-centric” and “qt-centric”, respectively, a new paradigm, referred to as the comprehensive in vitro proarrhythmia assay (cipa), is now gathering momentum among all the stake holders, including the regulatory authorities [110]. this paradigm, aiming to characterise the risk of proarrhythmias as opposed to qtc prolongation, recognises the critical role of multiple ion channels blockade and among other recommendations, calls for all new drugs to be studied for their effects on multiple ion channels and incorporation of these effects in an in silico computer modelling of human ventricular electrophysiology to predict a drug’s proarrhythmic potential. of greater concern is the real possibility that discarding a pki because it inhibits herg channel may prove to be counter-productive since herg is expressed in a variety of malignancies [111-113]. early indications are that herg channel blockers attenuate the progression of both hematologic malignancies and solid tumours [114-119]. not surprisingly, herg channel has been suggested as a potential target for admet & dmpk 4(3) (2016) 212-231 pkis and qt-phobia doi: 10.5599/admet.4.3.295 225 anticancer drugs [112, 120]. among the interesting such drugs is astemizole, withdrawn from the market in 1999 for its torsadogenic potential [46], which is not only a herg blocker but also an antihistamine [121] and histamine favours proliferation of normal and cancerous cells. if herg channel were to prove to be a valuable therapeutic target in oncology, new paradigms will have to evolve regarding the management of pki-induced qt interval prolongation. discussion and conclusions in contrast to their qt-liability, the hepatotoxic potential of pkis is far greater. nineteen (68 %) of the 28 pkis are labelled as potentially hepatotoxic with five them carrying a black box warning (table 2). twelve (75 %) of the 16 pkis approved in period 1 and 7 (58 %) of the 12 approved in period 2 are deemed to be hepatotoxic. four of the 12 from period 1 and one of the 7 from period 2 carry black box warning (lapatinib, pazopanib, regorafenib and sunitinib, and ponatinib, respectively). although the potential for clinically relevant hepatic injury is typically evaluated on the basis of the magnitude of increases in serum aspartate and alanine transaminases (ast and alt), alkaline phosphatase (alp) and bilirubin, the risk assessment and labelling are based on the number of cases that meet “hy’s rule”. according to this rule, a significant risk of severe hepatotoxicity is associated with medications which cause hepatic injury (elevation in alt) together with reduction in hepatic function (the synthesis and transportation of bilirubin) in absence of any evidence of biliary obstruction (e.g. elevation of alp) or of other causes that can reasonably explain these elevations in alt and bilirubin. hy’s rule has been validated and confirmed in two large studies of drug-induced liver injury in which approximately 10 percent of subjects with hyperbilirubinaemia or jaundice died or needed liver transplant [122, 123]. finding one hy’s rule case in the clinical trial database is worrisome; finding two is considered highly predictive of the drug having the potential to cause severe drug-induced liver injury when given to a larger population [124]. in pre-approval clinical trials of relevant pkis, there were cases that met hy’s rule or fatal cases of hepatic failure. attribution of the risk of clinically relevant hepatotoxicity has been justified by numerous post-marketing reports of pki-induced hepatotoxicity that required revisions to the labelling of several pkis. interestingly, however, although bosutinib is labelled as potentially hepatotoxic, a post-approval observational study in which 248 patients received bosutinib for a median duration of 27.5 months revealed no cases in the bosutinib arm that led to hospitalization, were associated with permanent hepatic injury or liver-related deaths, or met hy’s rule criteria [125]. thus, although the risks of proarrhythmia and hepatotoxicity are both determined on the basis of surrogate markers, the assessment of hepatotoxicity is based on a combination of clinically relevant and validated endpoints. there are also significant challenges to evaluating qtand transaminase-based clinical risk. whereas qt prolongation does not correlate well with the risk of proarrhythmias for the reasons already explained earlier, diagnosing drug-induced liver injury can be difficult in presence of metastasis. development of newer pkis should proceed on the basis of assessment of clinically meaningful risks and therapeutic benefits in terms of morbidity, mortality and quality of life if potentially valuable agents are not to be discarded early in their development programme without exploring their clinically meaningful benefits. safe and effective use of these valuable drugs requires close collaboration between the oncologists and their colleagues in other specialties such as the collaboration which has been forged by oncologists and cardiologists [3]. funding no sources of funding were used to assist in the preparation of this review. rashmi shah admet & dmpk 4(3) (2016) 212-231 226 conflict of interest the author has no conflicts of interest that are directly relevant to the content of this review. until retirement in december 2004, he was a senior clinical assessor at the medicines and healthcare products regulatory agency (mhra), london, uk, and the ich e14 topic leader, representing the eu. he now provides expert consultancy services relating to cardiovascular safety of medicines to a number of pharmaceutical companies. references [1] r. r. shah, s. a. roberts, d. r. shah. br. j. clin. pharmacol. 76(3) (2013) 396-411. [2] r. r. shah, j. morganroth, d. r. shah. drug saf. 36(5) (2013) 295-316. [3] r. r. shah, j. morganroth. drug saf. 38(8) (2015) 693-710. [4] j. zhang, l. zhang, y. wang, g. zhao. cancer chemother. pharmacol. 77(5) (2016) 905-926. [5] m. rask-andersen, j. zhang, d. fabbro, h. b. schiöth. trends pharmacol. sci. 35(11) (2014) 604-620. [6] p. wu, t. e. nielsen, m. h. clausen. drug discov. today 21(1) (2016) 5-10. [7] k. l. miller, m. lanthier. nat. rev. drug discov. 14(2) (2015) 83. [8] d. r. shah, r. r. shah, j. morganroth. drug saf. 36(6) (2013) 413-426. [9] r. dienstmann, i. braña, j. rodon, j. tabernero. oncologist 16(12) (2011) 1729-1740. [10] f. di fiore, o. rigal, c. ménager, p. michel, c. pfister. br. j. cancer 105(12) (2011) 1811-1813. [11] h. y. small, a. c. montezano, f. j. rios, c. savoia, r. m. touyz. can. j. cardiol. 30(5) (2014) 534-543. [12] s. liu, r. kurzrock. semin. oncol. 42(6) (2015) 863-875. [13] m. schmidinger, u. m. vogl, m. bojic, w. lamm, h. heinzl, a. haitel, m. clodi, g. kramer, c. c. zielinski. cancer 117(3) (2011) 534-544. [14] a. nearchou, a. valachis, p. lind, o. akre, p. sandström. clin. genitourin. cancer 13(4) (2015) 280286. [15] e. b. bailey, s. k. tantravahi, a. poole, a. m. agarwal, a. m. straubhar, j. a. batten, s. b. patel, c. e. wells, d. d. stenehjem, n. agarwal. clin. genitourin. cancer 13(3) (2015) e131-137. [16] g. lombardi, f. zustovich, p. farina, p. fiduccia, a. della puppa, v. polo, r. bertorelle, m. p. gardiman, a. banzato, p. ciccarino, l. denaro, v. zagonel. anticancer drugs 24(1) (2013) 90-97. [17] a. poprach, t. pavlik, b. melichar, i. puzanov, l. dusek, z. bortlicek, r. vyzula, j. abrahamova, t. buchler: czech renal cancer cooperative group. ann. oncol. 23(12) (2012) 3137-3143. [18] r. b. cohen, s. oudard. invest. new drugs 30(5) (2012) 2066-2079. [19] m. schmidinger. ejc suppl. 11(2) (2013) 172-191. [20] w. x. qi, a. n. he, z. shen, y. yao. br. j. clin. pharmacol. 76(3) (2013) 348-357. [21] p. ghatalia, c. j. morgan, y. je, p. l. nguyen, q. d. trinh, t. k. choueiri, g. sonpavde. crit. rev. oncol. hematol. 94(2) (2015) 228-237. [22] w. x. qi, z. shen, l. n. tang, y. yao. br. j. clin. pharmacol. 78(4) (2014) 748-762. [23] t. k. choueiri, f. a. schutz, y. je, j. e. rosenberg, j. bellmunt. j. clin. oncol. 28(13) (2010) 2280-2285. [24] w. x. qi, z. shen, l. n. tang, y. yao. crit. rev. oncol. hematol. 92(2) (2014) 71-82. [25] w. x. qi, d. l. min, z. shen, y. j. sun, f. lin, l. n. tang, a. n. he, y. yao. int. j. cancer 132(12) (2013) 2967-2974. [26] g. sonpavde, y. je, f. schutz, m. d. galsky, r. paluri, j. e. rosenberg, j. bellmunt, t. k. choueiri. crit. rev. oncol. hematol. 87(1) (2013) 80-89. [27] w. x. qi, l. n. tang, y. j. sun, a. n. he, f. lin, z. shen, y. yao. ann. oncol. 24(12) (2013) 2943-2952. [28] g. sonpavde, j. bellmunt, f. schutz, t. k. choueiri. curr. oncol. rep. 14(4) (2012) 295-306. [29] r. r. shah, j. morganroth, d. r. shah. drug saf. 36(7) (2013) 491-503. admet & dmpk 4(3) (2016) 212-231 pkis and qt-phobia doi: 10.5599/admet.4.3.295 227 [30] r. iacovelli, a. palazzo, g. procopio, m. santoni, p. trenta, a. de benedetto, s. mezi, e. cortesi. br. j. clin. pharmacol. 77(6) (2014) 929-938. [31] p. ghatalia, y je, n. e. mouallem, p. l. nguyen, q. d. trinh, g. sonpavde, t. k. choueiri. crit. rev. oncol. hematol. 93(3) (2015) 257-276. [32] w. x. qi, y. j. sun, z. shen, y. yao. j. chemother. 27(1) (2015) 40-51. [33] o. abdel-rahman, h. elhalawani. future oncol. 11(7) (2015) 1109-1122. [34] p. r. massey, j. s. okman, j. wilkerson, e. w. cowen. support care cancer 23(6) (2015) 1827-1835. [35] o. abdel-rahman, m. fouad. expert rev. anticancer ther. 14(9) (2014) 1063-1073. [36] z-f zhang, t. wang, l-h liu, h-q guo. plos. one 9(3): (2014) e90135. [37] s. r. hayman, n. leung, j. p. grande, v. d. garovic. curr. oncol. rep. 14(4) (2012) 285-294. [38] a. yuan, s. l. kurtz, c. m. barysauskas, a. p. pilotte, a. j. wagner, n. s. treister. oral oncol. 51(11) (2015) 1026-1033. [39] p. ghatalia, c. j. morgan, t. k. choueiri, p. rocha, g. naik, g. sonpavde. crit. rev. oncol. hematol. 94(1) (2015) 136-145. [40] food and drug administration. guidance for industry: premarketing risk assessment (march 2005). food and drug administration, rockville, md 20857, usa. available at: http://www.fda.gov/downloads/drugs/guidancecomplianceregulatoryinformation/guidances/ucm 072002.pdf , [accessed on 1 march 2016]. [41] g. michels, m. kochanek, r. pfister. med. klin. intensivmed. notfmed. 2015 sep 4. [epub ahead of print]. [42] w. haverkamp, g. breithardt, a. j. camm, m. j. janse, m. r. rosen, c. antzelevitch, d. escande, m. franz, m. malik, a. moss, r. shah. cardiovasc. res. 47(2) (2000) 219-233. [43] p. salle, j. l. rey, p. bernasconi, j. c. quiret, m. lombaert. ann. cardiol. angeiol. (paris). 34(6) (1985) 381-388. french. [44] d. k. wysowski, a. corken, h. gallo-torres, l. talarico, e. m. rodriguez. am. j. gastroenterol. 96(6) (2001) 1698-1703. [45] r. r. shah. pharmacogenomics 7(6) (2006) 889-908. [46] n. stockbridge, j. morganroth, r. r. shah, c. garnett. drug saf. 36(3) (2013) 167-182. [47] r. r. shah. drug saf. 30(12) (2007) 1093-1110. [48] l. johannesen, j. vicente, j. w. mason, c. erato, c. sanabria, k. waite-labott, m. hong, j. lin, p. guo, a. mutlib, j. wang, w. j. crumb, k. blinova, d. chan, j. stohlman, j. florian, m. ugander, n. stockbridge, d. g. strauss. clin. pharmacol. ther. 99(2) (2016) 214-223. [49] m. c. makielski. trends cardiovasc. med. 26(2) (2016) 115-122. [50] z. j. huang, d. z. dai, n. li, t. na, m. ji, y. dai. clin. exp. pharmacol. physiol. 34(4) (2007) 310-317. [51] l. johannesen, j. vicente, j. w. mason, c. sanabria, k. waite-labott, m. hong, p. guo, j. lin, j. s. sørensen, l. galeotti, j. florian, m. ugander, n. stockbridge, d. g. strauss. clin. pharmacol. ther. 96(5) (2014) 549-558. [52] g. antoons, a. oros, j. d. beekman, m. a. engelen, m. j. houtman, l. belardinelli, m. stengl, m. a. vos ma. j. am. coll. cardiol. 55(8) (2010) 801-809. [53] s. jia, j. lian, d. guo, x. xue, c. patel, l. yang, z. yuan, a. ma, g. x. yan. br. j. pharmacol. 164(2) (2011) 308-316. [54] d. rampe, m. k. murawsky, j. grau, e. w. lewis. j. pharmacol. exp. ther. 286(2) (1998) 788-793. [55] l. eckardt, g. breithardt, w. haverkamp. j. pharmacol. exp. ther. 300(1) (2002) 64-71. [56] k. titier, p. o. girodet, h. verdoux, m. molimard, b. bégaud, w. haverkamp, m. lader, n. moore. drug saf. 28(1) (2005) 35-51. [57] z. laksman, b. momciu, y. w. seong, p. burrows, s. conacher, j. manlucu, p. leong-sit, l. j. gula, a. c. skanes, r. yee, g. j. klein, a. d. krahn. am. j. cardiol. 115(7) (2015) 907-911. http://www.fda.gov/downloads/drugs/guidancecomplianceregulatoryinformation/guidances/ucm072002.pdf http://www.fda.gov/downloads/drugs/guidancecomplianceregulatoryinformation/guidances/ucm072002.pdf rashmi shah admet & dmpk 4(3) (2016) 212-231 228 [58] d. m. roden. j. physiol. 594(9) (2016) 2459-2468. [59] committee for proprietary medicinal products (cpmp). points to consider: the assessment of the potential for qt interval prolongation by non-cardiovascular medicinal products (cpmp/986/96). european medicines agency, london, 17 december 1997. available at: http://www.fda.gov/ohrms/dockets/ac/03/briefing/pubs/cpmp.pdf, [accessed on 5 april 2012] [60] international conference on harmonisation (ich). ich note for guidance on: the nonclinical evaluation of the potential for delayed ventricular repolarization (qt interval prolongation) by human pharmaceuticals (ich s7b). international conference on harmonisation, geneva, may 2005. available at: http://www.ich.org/fileadmin/public_web_site/ich_products/guidelines/safety/s7b/step4/s7b_gu ideline.pdf, [accessed on 5 march 2016] [61] international conference on harmonisation (ich). ich note for guidance on: the clinical evaluation of qt/qtc interval prolongation and proarrhythmic potential for non-antiarrhythmic drugs (ich e14). international conference on harmonisation, geneva, may 2005. available at: http://www.ich.org/fileadmin/public_web_site/ich_products/guidelines/efficacy/e14/e14_guideli ne.pdf , [accessed on 5 march 2016] [62] international council for harmonisation (ich). ich e14 implementation working group. ich e14 guideline: the clinical evaluation of qt/qtc interval prolongation and proarrhythmic potential for non-antiarrhythmic drugs. questions & answers (r3). international council for harmonisation, geneva, december 2015. available at: http://www.ich.org/fileadmin/public_web_site/ich_products/guidelines/efficacy/e14/e14_q_as_r 3__step4.pdf, [accessed on 5 march 2016] [63] r. r. shah. br. j. clin. pharmacol. 54(2) (2002) 188-202. [64] y. l. lin, m. f. kung. pharmacoepidemiol. drug saf. 18(3) (2009) 235-239. [65] e. i. heath, j. infante, l. d. lewis, t. luu, j. stephenson, a. r. tan, s. kasubhai, p. lorusso, b. ma, a. b. suttle, j. f. kleha, h. a. ball, m. m. dar. cancer chemother. pharmacol. 71(3) (2013) 565-573. [66] k. s. kloth, a. pagani, m. c. verboom, a. malovini, c. napolitano, w. h. kruit, s. sleijfer, n. steeghs, a. zambelli, r. h. mathijssen. br. j. cancer 112(6) (2015) 1011-1016. [67] z. lu, c. y. wu, y. p. jiang, l. m. ballou, c. clausen, i. s. cohen, r. z. lin. sci. transl. med. 4(131) (2012) 131ra50. [68] l. m. ballou, r. z. lin, i. s. cohen. circ. res. 116(1) (2015) 127-137. [69] j. s. lowe, d. m. stroud, t. yang, l. hall, t. c. atack, d. m. roden. cardiovasc. res. 95(3) (2012) 300307. [70] d. h. staniforth. br. j. clin. pharmacol. 16(6) (1983) 615-621. [71] d. h. staniforth. br. j. clin. pharmacol. 19(6) (1985) 862. [72] m. malik. pacing clin. electrophysiol. 25(2) (2002) 209-216. [73] m. desai, l. li, z. desta, m. malik, d. flockhart. br. j. clin. pharmacol. 55(6) (2003) 511-517. [74] national cancer institute. common terminology criteria for adverse events (ctcae): version 4.0 published: may 28, 2009 (v4.03: june 14, 2010). available at: http://evs.nci.nih.gov/ftp1/ctcae/ctcae_4.03_2010-06-14_quickreference_5x7.pdf, [accessed on 2 march 2016] [75] m. schmidinger, c. c. zielinski, u. m. vogl, a. bojic, m. bojic, c. schukro, m. ruhsam, m. hejna, h. schmidinger. j. clin. oncol. 26(32) (2008) 5204-5212. [76] p. s. hall, l. c. harshman, s. srinivas, r. m. witteles. jacc heart fail. 1(1) (2013) 72-78. [77] s. moustafa, t. h. ho, p. shah, k. murphy, b. k. nelluri, h. lee, s. wilansky, f. mookadam. j. clin. ultrasound 44(4) (2016) 221-230. [78] c. cheng, t. force. prog. cardiovasc. dis. 53(2) (2010) 114-120. [79] t. force, k. l. kolaja. nat. rev. drug discov. 10(2) (2011) 111-126. http://www.fda.gov/ohrms/dockets/ac/03/briefing/pubs/cpmp.pdf http://www.ich.org/fileadmin/public_web_site/ich_products/guidelines/safety/s7b/step4/s7b_guideline.pdf http://www.ich.org/fileadmin/public_web_site/ich_products/guidelines/safety/s7b/step4/s7b_guideline.pdf http://www.ich.org/fileadmin/public_web_site/ich_products/guidelines/efficacy/e14/e14_guideline.pdf http://www.ich.org/fileadmin/public_web_site/ich_products/guidelines/efficacy/e14/e14_guideline.pdf http://www.ich.org/fileadmin/public_web_site/ich_products/guidelines/efficacy/e14/e14_q_as_r3__step4.pdf http://www.ich.org/fileadmin/public_web_site/ich_products/guidelines/efficacy/e14/e14_q_as_r3__step4.pdf http://evs.nci.nih.gov/ftp1/ctcae/ctcae_4.03_2010-06-14_quickreference_5x7.pdf admet & dmpk 4(3) (2016) 212-231 pkis and qt-phobia doi: 10.5599/admet.4.3.295 229 [80] m. svoboda, a. poprach, s. dobes, i. kiss, r. vyzula. cardiovasc. toxicol. 12(3) (2012) 191-207. [81] f. jacob, a. y. yonis, f. cuello, p. luther, t. schulze, a. eder, t. streichert, i. mannhardt, m. n. hirt, s. schaaf, j. stenzig, t. force, t. eschenhagen, a. hansen. plos. one 11(2) (2016) e0145937. [82] pfizer limited. dear healthcare professional letter (13 october 2015). “inclusion of a new warning regarding cardiac failure”. available at: https://assets.digital.cabinetoffice.gov.uk/media/56435e1340f0b674d3000024/xalnp36__uk_xalkori_dhcp_letter_cardiac_failure__2_.pdf, [accessed on 22 march 2016] [83] food and drug administration. crizotinib (xalkori  ) label (29 april 2016). food and drug administration, silver spring, md 20993, usa. available at: http://www.accessdata.fda.gov/drugsatfda_docs/label/2016/202570s017lbl.pdf, [accessed on 22 march 2016]. [84] o. abdel-rahman, m. fouad. crit. rev. oncol. hematol. 92(3) (2014) 194–207. [85] b. yang, t. papoian. j. appl. toxicol. 32(12) (2012) 945-951. [86] food and drug administration. pharmacology review of osimertinib (tagrisso  ). food and drug administration, silver spring, md 20993, usa. available at: http://www.accessdata.fda.gov/drugsatfda_docs/nda/2015/208065orig1s000pharmr.pdf, [accessed on 22 march 2016]. [87] food and drug administration. pharmacology review of cobimetinib (cotellic  ). food and drug administration, silver spring, md 20993, usa. available at: http://www.accessdata.fda.gov/drugsatfda_docs/nda/2015/206192orig1s000pharmr.pdf, [accessed on 22 march 2016]. [88] food and drug administration. pharmacology review of afatinib (gilotrif  ). food and drug administration, silver spring, md 20993, usa. available at: http://www.accessdata.fda.gov/drugsatfda_docs/nda/2013/201292orig1s000pharmr.pdf, [accessed on 22 march 2016]. [89] food and drug administration. pharmacology review of sunitinib (sutent  ). food and drug administration, silver spring, md 20993, usa. available at: http://www.accessdata.fda.gov/drugsatfda_docs/nda/2006/021938_s000_sutent_pharmr.pdf, [accessed on 22 march 2016]. [90] food and drug administration. pharmacology review of bosutinib (bosulif  ). food and drug administration, silver spring, md 20993, usa. available at: http://www.accessdata.fda.gov/drugsatfda_docs/nda/2012/203341orig1s000pharmr.pdf, [accessed on 22 march 2016]. [91] k. r. doherty, d. r. talbert, p. b. trusk, d. m. moran, s. a. shell, s. bacus. toxicol. appl. pharmacol. 285(1) (2015) 51-60. [92] a. srikanthan, j. l. ethier, a. ocana, b. seruga, m.k. krzyzanowska, e. amir. plos. one. 10(3) (2015) e0122735. [93] s. jang, c. zheng, h. t. tsai, a. z. fu, a. barac, m. b. atkins, a. n. freedman, l. minasian, a. l. potosky. cancer 122(1) (2016) 124-130. [94] food and drug administration. ponatinib (iclusig  ) label (14 december 2012). food and drug administration, silver spring, md 20993, usa. available at: http://www.accessdata.fda.gov/drugsatfda_docs/label/2012/203469lbl.pdf, [accessed on 22 march 2016]. [95] food and drug administration. ponatinib (iclusig  ) approval letter (14 december 2012). food and drug administration, silver spring, md 20993, usa. available at: http://www.accessdata.fda.gov/drugsatfda_docs/appletter/2012/203469orig1s000ltr.pdf, [accessed on 22 march 2016]. [96] food and drug administration. fda drug safety communication: fda asks manufacturer of the leukemia drug iclusig (ponatinib) to suspend marketing and sales. food and drug administration, https://assets.digital.cabinet-office.gov.uk/media/56435e1340f0b674d3000024/xalnp36_-_uk_xalkori_dhcp_letter_cardiac_failure__2_.pdf https://assets.digital.cabinet-office.gov.uk/media/56435e1340f0b674d3000024/xalnp36_-_uk_xalkori_dhcp_letter_cardiac_failure__2_.pdf https://assets.digital.cabinet-office.gov.uk/media/56435e1340f0b674d3000024/xalnp36_-_uk_xalkori_dhcp_letter_cardiac_failure__2_.pdf http://www.accessdata.fda.gov/drugsatfda_docs/label/2016/202570s017lbl.pdf http://www.accessdata.fda.gov/drugsatfda_docs/nda/2015/208065orig1s000pharmr.pdf http://www.accessdata.fda.gov/drugsatfda_docs/nda/2015/206192orig1s000pharmr.pdf http://www.accessdata.fda.gov/drugsatfda_docs/nda/2013/201292orig1s000pharmr.pdf http://www.accessdata.fda.gov/drugsatfda_docs/nda/2006/021938_s000_sutent_pharmr.pdf http://www.accessdata.fda.gov/drugsatfda_docs/nda/2012/203341orig1s000pharmr.pdf http://www.accessdata.fda.gov/drugsatfda_docs/label/2012/203469lbl.pdf http://www.accessdata.fda.gov/drugsatfda_docs/appletter/2012/203469orig1s000ltr.pdf rashmi shah admet & dmpk 4(3) (2016) 212-231 230 silver spring, md 20993, usa. available at: http://www.fda.gov/drugs/drugsafety/ucm373040.htm, [accessed on 22 march 2016]. [97] food and drug administration. fda drug safety communication: fda requires multiple new safety measures for leukemia drug iclusig; company expected to resume marketing. food and drug administration, silver spring, md 20993, usa. available at: http://www.fda.gov/drugs/drugsafety/ucm379554.htm, [accessed on 22 march 2016]. [98] food and drug administration. ponatinib (iclusig  ) label (2 june 2016). food and drug administration, silver spring, md 20993, usa. available at: http://www.accessdata.fda.gov/drugsatfda_docs/label/2016/203469s021lbl.pdf, [accessed on 22 march 2016]. [99] s. sivendran, z. liu, l. j. portas jr, m. yu, n. hahn, g. sonpavde, w. k. oh, m. d. galsky. cancer treat. rev. 38(7) (2012) 919-925. [100] f. a. schutz, y. je, c. j. richards, t. k. choueiri. j. clin. oncol. 30(8) (2012) 871-877. [101] s. hong, w. fang, w. liang, y. yan, t. zhou, t. qin, x. wu, y. ma, y. zhao, y. yang, z. hu, c. xue, x. hou, y. chen, y. huang, h. zhao, l. zhang. onco. targets ther. 7 (2014) 1851-1867. [102] x. yang, x. pan, x. cheng, y. cheng, y. kuang. int. j. clin. pharm. 37(6) (2015) 1047-1056. [103] g. w. jean, r. m. mani, a. jaffry, s. a. khan. jama oncol. 2(4) (2016) 529-534. [104] w. x. qi, l. n. tang, a. n. he, y. yao, z. shen. respir. med. 107(8) (2013) 1280-1283. [105] t. funakoshi, a. latif, m. d. galsky. cancer treat. rev. 40(5) (2014) 636-647. [106] s. niraula, b. seruga, a. ocana, t. shao, r. goldstein, i. f. tannock, e. amir. j. clin. oncol. 30(24) (2012) 3012-3019. [107] m. joerger, c. schaer-thuer, d. koeberle, k. matter-walstra, j. gibbons-marsico, s. diem, b. thuerlimann, t. cerny. eur. j. clin. pharmacol. 70(6) (2014) 719-725. [108] b. seruga, l. sterling, l. wang, i. f. tannock. j. clin. oncol. 29(2) (2011) 174-185. [109] j. mori, t. tanimoto, y. miura, m. kami. jpn. j. clin. oncol. 45(6) (2015) 588-594. [110] g. gintant, p.t. sager, n. stockbridge. nat. rev. drug discov. 15 (2016) 457-471. [111] x. d. shao, k. c. wu, x. z. guo, m. j. xie, j. zhang, d. m. fan. cancer biol. ther. 7(1) (2008) 45-50. [112] j. i. vandenberg, m. d. perry, m. j. perrin, s. a. mann, y. ke, a. p. hill. physiol. rev. 92(3) (2012) 13931478. [113] j. j. babcock, m. li. acta pharmacol. sin. 34(3) (2013) 329-335. [114] g. a. smith, h. w. tsui, e. w. newell, x. jiang, x. p. zhu, f. w. tsui, l. c. schlichter. j. biol. chem. 277(21) (2002) 18528-18534. [115] a. arcangeli, o. crociani, e. lastraioli, a. masi, s. pillozzi, a. becchetti. curr. med. chem. 16(1) (2009) 66-93. [116] g. glassmeier, k. hempel, i. wulfsen, c. k. bauer, u. schumacher, j. r. schwarz. pflugers. arch. 463(2) (2012) 365-376. [117] a. arcangeli, a. becchetti. drug resist. updat. 2015 jul-aug;21-22:11-9. [118] e. lastraioli, t. lottini, l. bencini, m. bernini, a. arcangeli. biomed. res. int. 2015 (2015) 896432. [119] l. leanza, a. managò, m. zoratti, e. gulbins, i. szabo. biochim biophys acta. 1863 (2016) 1385-1397. [120] v. r. rao, m. perez-neut, s. kaja, s. gentile. cancers 7(2) (2015) 849-875. [121] j. garcia-quiroz, j. camacho. anti-cancer agents med. chem. 11(3) (2011) 307–314. [122] r. j. andrade, m. i. lucena, m. c. fernández, g. pelaez, k. pachkoria, e. garcía-ruiz, b. garcía-muñoz, r. gonzález-grande, a. pizarro, j. a. durán, m. jiménez, l. rodrigo, m. romero-gomez, j. m. navarro, r. planas, j. costa, a. borras, a. soler, j. salmerón, r. martin-vivaldi: spanish group for the study of drug-induced liver disease. gastroenterology 129(2) (2005) 512-521. [123] e. björnsson, r. olsson. hepatology 42(2) (2005) 481-489. http://www.fda.gov/drugs/drugsafety/ucm373040.htm http://www.fda.gov/drugs/drugsafety/ucm379554.htm http://www.accessdata.fda.gov/drugsatfda_docs/label/2016/203469s021lbl.pdf admet & dmpk 4(3) (2016) 212-231 pkis and qt-phobia doi: 10.5599/admet.4.3.295 231 [124] food and drug administration. guidance for industry drug-induced liver injury: premarketing clinical evaluation (july 2009). food and drug administration, silver spring, md 20993-0002, usa. available at: http://www.fda.gov/downloads/drugs/guidancecomplianceregulatoryinformation/guidances/ucm 174090.pdf, [accessed on 1 march 2016]. [125] t. h. brümmendorf, j. e. cortes, c. a. de souza, f. guilhot, l. duvillié, d. pavlov, k. gogat, a. m. countouriotis, c. gambacorti-passerini. br. j. haematol. 168(1) (2015) 69-81. ©2016 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://www.fda.gov/downloads/drugs/guidancecomplianceregulatoryinformation/guidances/ucm174090.pdf http://www.fda.gov/downloads/drugs/guidancecomplianceregulatoryinformation/guidances/ucm174090.pdf http://creativecommons.org/licenses/by/3.0/ manuscript doi: 10.5599/admet.4.4.335 327 admet & dmpk 4(4) (2016) 327-334; doi: 10.5599/admet.4.4.335 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper tissue distribution of crizotinib and gemcitabine combination in a patient-derived orthotopic mouse model of pancreatic cancer richard j honeywell 1 , amir avan 2 , elisa giovannetti 1,3 and godefridus j peters 1,* 1 department of medical oncology, vu university medical center, po box 7057, 1007 mb amsterdam, the netherlands 2 molecular medicine group, department of modern sciences and technologies, faculty of medicine, mashhad university of medical sciences, mashhad, iran 3 cancer pharmacology lab, airc start-up unit, dipint, pisa, italy *corresponding author: e-mail: gj.peters@vumc.nl. received: august 10, 2016; revised: december 09, 2016; published: december 26, 2016 abstract pharmacokinetics focuses on the question whether a drug actually reaches its target in therapeutic concentrations or accumulates elsewhere, potentially causing toxicological or unpredictable side effects. we determined tissue distribution of gemcitabine, an antimetabolite, and crizotinib, a tyrosine-kinase inhibitor targeted against the anaplastic lymphoma kinase (alk) and mesenchymal-epithelial transition factor (c-met) receptors, in a validated orthotopic mouse model for pancreatic cancer. mice with pancreatic cancer were treated with either oral crizotinib at 25 mg/kg, gemcitabine at 100 mg/kg or with their combination. two hours after the last gemcitabine dose mice were sacrificed and all available blood/organs/tissues were sampled. tissue was subsequently analyzed for drug concentrations using a validated liquid chromatography-mass spectrometry (lc-ms/ms) technique. in whole blood gemcitabine was about 1.0 µm and crizotinib 2.4 µm in the single treatment, whereas in the combination crizotinib increased the levels of gemcitabine. crizotinib was found in all major tissues, being highest in the intestine. comparison of crizotinib alone to the gemcitabine-crizotinib combination showed that crizotinib tissue concentrations were 3-6 fold lower in liver, lung, kidney and spleen, 30-fold lower in the skin, heart and pancreas and 200-fold lower in the brain. tissue gemcitabine was highest in spleen and skin, being about 510 fold higher than in the other tissues, including brain, which still had a relatively high accumulation. in conclusion, both gemcitabine and crizotinib accumulate at clinically active but variable levels in tissues, possibly relating to the effects exerted by these drugs. keywords crizotinib; gemcitabine; tissue distribution introduction pharmacokinetics usually focuses on the behavior of a single drug in plasma in order to obtain information on its uptake, distribution and elimination [1]. in cancer patients this information can give insight in the mechanism of action, the relationship between dose, systemic concentration and biological effect (beneficial or otherwise) [2]. however, limited or insufficient attention is usually given to the question whether the drug actually reaches its target (the tumor) in therapeutic concentrations or accumulates in normal tissues potentially leading to toxicity. usually drug interactions of combinations receive limited interest. earlier we investigated the interaction between gemcitabine and crizotinib http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:gj.peters@vumc.nl g.j. peters et al. admet & dmpk 4(4) (2016) 327-334 328 (figure 1) and demonstrated a synergistic effect between the drugs in in vitro and in vivo models of pancreatic ductal adenocarcinoma (pdac) [3]. in the present study we aimed to analyze their interaction in normal tissues, in order to determine whether this information can give an explanation for toxic side effects seen with these drugs in patients. crizotinib is registered for the treatment of non-small cell lung cancer that expresses the abnormal eml4-alk fusion gene, and has been evaluated for use with tumors expressing high c-met [4-6]. in turn, gemcitabine is standard first line treatment in a wide range of carcinomas including non-small cell lung cancer, pancreatic cancer, bladder cancer and breast cancer [7]. in this paper we describe the tissue distribution of both crizotinib and gemcitabine, given alone or together in a patient-derived orthotopic mouse pdac model. figure 1. structural formulae of crizotinib (a) and gemcitabine (b). model and methods mouse model orthotopic primary-pdac mouse models were developed as reported by avan et al. [3] by implantation of primary tumor cells into 6-8 week old female athymic nude mice (harlan, host, the netherlands). tumor growth was monitored via bioluminescence, as described previously [3]. the mice were treated with oral crizotinib at 25 mg/kg, or gemcitabine at 100 mg/kg intraperitoneally 4 times at 3-day intervals. the third group received a combination of both drugs. blood was taken via the tail vein at 1 and 2 h after drug administration. thereafter mice were sacrificed and all available organs/tissue was sampled. sampled tissue was immediately frozen in liquid nitrogen and stored at -80 °c. drug analysis and equipment tissue and whole blood concentrations of gemcitabine and crizotinib were determined by adaptation of a validated liquid chromatography-mass spectrometry (lc-ms/ms) methodology previously reported [8,9] using a sciex api3000 mass spectrometer coupled with a dionex ultimate 3000 capillary lc system. precision, accuracy, robustness and stability were established for crizotinib and gemcitabine prior to tissue analysis in whole blood, plasma, serum and cell culture medium [8]. parameters were determined with software “analyst” version 6.2 (sciex bv, breda) and dmslink ver 2.10 (thermo scientific dionex, breda). whole blood, plasma and serum were extracted by adding 20 µl whole blood aliquots to 80 µl acetonitrile and sonicating for 15 seconds. after centrifugation at 2500 g for 10 minutes / 4 °c, the supernatant was taken and analyzed by lc-ms/ms. tissue sample preparation was performed by taking triplicate portions of each tissue (5-30 mg), weighed into 1.5 ml cryogenic vials which were subsequently desiccated using the freeze drying processes for 48 h. samples were then homogenized for 10 min with acetonitrile: water (5:1), sonicated and centrifuged. the subsequent supernatant was then analyzed by lcms/ms. (a) (b) admet & dmpk 4(4) (2016) 327-334 tissue distribution of crizotinib and gemcitabine in mouse model doi: 10.5599/admet.4.4.335 329 crizotinib optimization crizotinib analysis was optimized for mass spectroscopic detection as described earlier for tyrosine kinase inhibitors. briefly, 1 µg/ml solution was infused at 3 µl per min into the turbo ionisation source, the parent ion for crizotinib (mw-449) was established as 450, 451 and 452 amu with characteristic pattern of a chlorine containing molecule (figure 2a). the major fragment of the 450 amu response was 259.9 (figure 2a). mass spectrometric parameters for the transition quadrupole pairing (q1/q3) of 450.0/259.1 are: declustering potential (dp), 61 v; focusing potential, 270 v; entrance potential, 10 v; collision energy, 31 v; cell exit potential, 18 v. crizotinib 0 1 2 3 0 500 1000 1500 time (min) c o u n ts p e r s e c o n d gemcitabine 0 1 2 3 0 50 100 150 200 time (min) deoxycytidine 0 1 2 3 0 500 1000 1500 time (min) crizotinib extracted from whole blood 0 2000 4000 6000 0 50000 100000 150000 200000 250000 mouse whole blood human whole blood concentration (ng/ml) i n t e n s i t y ( c p s ) + q1 spectrum of crizotinib in mobile phase 420 440 460 480 0.0 5000000.0 1.010 07 1.510 07 2.010 07 2.510 07 450.0 451.0 452.0 n nh 2 o cl cl f n n hn c p s product ion spectrum crizotinib 250 300 350 400 4500 500000 1000000 1500000 2000000 259.9 450.0 369.9 c p s a b c figure 2. ionisation spectrum and chromatography of crizotinib, gemcitabine and of the internal control deoxycytidine. (a) ionisation spectrum of the parent compound (crizotinib) and the product ion spectrum of the [m+h] + ion for crizotinib in chromatographic mobile phase, consisting of acetonitrile, methanol, isopropyl alcohol and aqueous ammonium acetate at ph 7.80. (b) chromatography of crizotinib, gemcitabine and deoxycytidine. (c) linearity of crizotinib analysis in mouse and human whole blood. extraction and chromatography crizotinib chromatography was performed using the optimized chromatographic system for the multiple tyrosine kinase inhibitors (figure 2b) [8]. under these set conditions crizotinib exhibited less sensitivity g.j. peters et al. admet & dmpk 4(4) (2016) 327-334 330 when compared to erlotinib, sunitinib and gefitinib. however, the crizotinib assay was linear up to 30 µg/ml in both plasma and serum (r 2 = 0.986, n= 10), and reproducible over a 3 day period. crizotinib was more soluble in aqueous solutions than other reported tkis. for whole blood the extraction was best performed in reverse, where the 20 µl sample of whole blood was pipetted directly into round bottomed 96 well already containing 80 µl of acetonitrile. this ensured each sample was homogeneously mixed prior to centrifugation and transfer of the supernatant for lc-ms/ms analysis. linearity linear regression for whole blood preparations was similar to plasma / serum / cell culture medium in the 1 – 5000 ng/ml and was acceptable (r 2 >0.98). the standard error was within ± 15 % for each calibration point used. the limit of quantitation for crizotinib was 50 ng/ml, due to lack of stock mouse whole blood preparations of human whole blood were compared to mouse whole blood standard curves. no difference was observed in the linear regression parameters determined (figure 2c), hence human whole blood was used as an alternative stock matrix for the samples to be analyzed. to limit the standard requirements per analytical run, standard curves for crizotinib and gemcitabine, and its metabolite difluorodeoxyuridine (dfdu) were prepared in either whole blood or cell culture medium. separate extractions for gemcitabine/dfdu and crizotinib were just not possible because the sample volume available for the mouse blood samples ranged from 5-20 µl per time point. therefore, extraction using acetonitrile precipitation was used for gemcitabine and dfdu. gemcitabine/ dfdu sensitivity was lower (especially for dfdu in tissues) compared to the previously published procedure but was linear in response across multiple days. the limit of quantification for gemcitabine was 10 ng/ml. linear regression for gemcitabine was not affected by the presence of crizotinib and chromatographically all compounds were separated (figure 2c). tissue analysis preparation of a standard curve within a tissue matrix was not always feasible due to the lack of sufficient tissue from untreated mice. therefore cell culture medium was compared with blank mouse tissue (heart), which was available in sufficient quantity for the validation procedure. both were prepared as per protocol. separate standard curves were prepared by spiking with crizotinib and gemcitabine dilutions in acetonitrile:water (5:1). no difference was observed. therefore, for all tissue analysis cell culture medium was used for the standard matrix. recovery of crizotinib, gemcitabine and dfdu was verified using the extracted tissue of the combination treated animal, by testing whether subsequent extraction would yield any detectable drug. therefore, tissues were extracted with acetonitrile:water (5:1) for a second and third time. after analysis no crizotinib, gemcitabine or dfdu could be observed within the second and third extracts, indicating that close to 100 % had been recovered on the first extraction (data not shown). results whole blood analysis limited sampling was performed at 1 and 2 hour, since the peak concentration of gemcitabine was expected around these time intervals [10,11]. whole blood analysis of the treated animals demonstrated measurable concentrations for gemcitabine (2 h: 1.0 ± 0.3 µm), dfdu (2 h: 5.6 ± 1.4 µm) and crizotinib (2 h: 2.4 ± 0.8 µm) (figure 3). admet & dmpk 4(4) (2016) 327-334 tissue distribution of crizotinib and gemcitabine in mouse model doi: 10.5599/admet.4.4.335 331 tissue analysis gemcitabine was found in all tissues that could be analyzed, when given alone or in combination. the highest levels were found in skin and spleen. the concentration in the brain was relatively high compared to the other tissue types and even comparable to liver (figure 4), showing a good brain penetration of gemcitabine. however, the primary metabolite of gemcitabine, dfdu, was not detectable for this group of samples. this was most likely due to the lack of sensitivity of uridine based compounds to mass spectrometric detection under the chromatographic conditions used, since we did not use the optimal lcms conditions as used earlier for tissues [14]. in the combination the highest levels of gemcitabine were found in the same tissues as at single drug treatment, with increased drug levels in blood, but lower levels in lung and brain (figure 4). d ru g c o n c e n tr a ti o n ( µ m ) crizotinib gemcitabine dfdu 0 5 10 gemcitabine + crizotinib gemicitabine alone crizotinib alone nd nd nd d ru g c o n c e n tr a ti o n ( µ m ) crizotinib gemcitabine dfdu 0 5 10 gemcitabine + crizotinib gemicitabine alone crizotinib alone nd nd nd 1 hr 2 hr figure 3. concentrations of crizotinib, gemcitabine and dfdu in whole mouse blood. blood was taken at 1 and 2 h after drug administration. values are means ± sd of 3 animals. p m o l/ m g t is s u e g e m c it a b in e p an cr ea s l iv er l u n g b ra in in te st in e k id n ey s ki n h ea rt s p le en 0 5 10 20 30 40 50 p m o l/ m g t is s u e g e m c it a b in e p an cr ea s l iv er l u n g b ra in s ki n 0 10 20 30 40 a b figure 4. tissue analysis (at 2 hour after drug administration) of gemcitabine at single drug administration (a) and in combination with crizotinib (b). values are means ± sd of 3 animals. at single treatment crizotinib accumulated at high levels in lung, liver and pancreatic tissues, and at very low levels in the brain (100-fold lower) (figure 5a). in the combined treatment crizotinib (pmol/mg tissue) was also found in all major tissues at varying concentrations; intestine demonstrated the highest levels g.j. peters et al. admet & dmpk 4(4) (2016) 327-334 332 (1084) followed by liver (215), lung (172) and kidney (167). in addition, crizotinib was found in the spleen (133), skin (28), heart (34), pancreas (48) and brain (4). crizotinib demonstrated an increased accumulation compared to the single treatment in 4 of 5 tissues tested (figure 5b and table 1). p m o l/ m g t is s u e c ri z o ti n ib p an cr ea s l iv er l u n g b ra in s ki n 0 5 10 15 20 50 100 150 200 250 p m o l/ m g t is s u e c ri z o ti n ib p an cr ea s l iv er l u n g b ra in in te st in e k id n ey s ki n h ea rt s p le en 0 2 4 6 500 1000 100 a b figure 5. tissue analysis (at 2 hour after drug administration) of crizotinib at single drug administration (a) and in combination with gemcitabine (b). values are means ± sd of 3 animals. table 1. effect of combination treatment on drug levels in normal tissues, the pancreatic tumor and blood. blood pancreas liver lung brain skin tumor crizotinib 1.4 1.4 2.0 0.8 4.5 1.6 1.4 gemcitabine 2.0 1.3 0.4 0.2 0.2 1.2 1.3 values are ratios of the combination versus single drug. tumor data are from [7] discussion tissue distribution of crizotinib alone and of the gemcitabine and crizotinib combination has not been reported previously. only limited data are available for gemcitabine distribution in the range of tissues we have tested. early data on gemcitabine tissue accumulation were obtained using radioactively labeled drug [12,13]; however, these data does not differentiate between gemcitabine and dfdu, which could overestimate the tissues exposure to gemcitabine. as expected, gemcitabine is accumulated into all tested tissues, in agreement with previous results in humans and animals [14-17], as also shown with 19f-nmr15. the use of the sensitive lc-ms/ms technique enabled acquisition of reproducible blood pharmacokinetics for all three compounds (crizotinib, gemcitabine and dfdu) simultaneously without interference [8,9], but tissue analysis of dfdu was not possible since we had to adapt the protocol to have one extraction procedure for all drugs. using the optimal lc-ms/ms procedure for dfdu we could demonstrate accumulation of dfdu in human tissue [14]. whole blood analysis proved consistent with expected pharmacokinetics of gemcitabine 1 and 2 hours after dosing [10,11]. gemcitabine levels demonstrated a rapid decrease in concentration over the 2 hour time period corresponding to an increase in observable dfdu in agreement with an earlier pharmacokinetic study in mice [11]. for crizotinib, literature reported admet & dmpk 4(4) (2016) 327-334 tissue distribution of crizotinib and gemcitabine in mouse model doi: 10.5599/admet.4.4.335 333 data are limited, but since the methodology was adequate for blood and tissue pharmacokinetics of gemcitabine as well as blood crizotinib, it can be assumed that the crizotinib results also reflect the true situation. crizotinib demonstrated an elevated concentration after 1 hour which was maintained after 2 hours indicating a steady state condition had probably been reached. crizotinib accumulated into all tested tissues at levels consistent with therapeutic concentrations, except for brain. crizotinib accumulation in the brain tissue has not been previously reported, but low levels were expected since it is a good substrate for efflux pumps [18]. hence, crizotinib does not show efficacy in patients with either brain tumors of liver metastases [19]. the combination of crizotinib with gemcitabine appears to enhance crizotinib exposure in the tissues of the pancreas, liver, brain and skin. the accumulation of crizotinib in the orthotopic tumor was about 37.4±30.3 pmol/mg tissue, while that of gemcitabine was about 5.84±2.41 pmol/mg tissue, demonstrating excellent tumor uptake. c-met overexpression is observed in tumors from many different tissues such as brain, kidney, ovarian, breast and gastrointestinal cancers [4,20]. the use of crizotinib might be considered as an additional resource in these areas, especially in combination with gemcitabine. crizotinib also appears to enhance the exposure of pancreas and skin towards gemcitabine, possibly by inhibiting the gemcitabine degradation enzyme cytidine deaminase [3]. the high accumulation of crizotinib in tissues might not reflect the actually accumulation in cells, particularly in the cytoplasm. earlier we demonstrated in cancer cell lines that crizotinib might concentrate in lysosomes [18], similar to sunitinib [21]. hence, in tissue with a high content of lysosomes the accumulation might be a lysosomal accumulation. it is not clear what these data mean in view of toxicology, however, in general crizotinib is tolerated well, although the drug may induce gastrointestinal toxicity, edema and visual impairment [5,22,23]. the toxicity of gemcitabine in patients is predominantly hematological [24] which is in line with the high concentration observed in the spleen. these results clearly show that distribution of crizotinib and gemcitabine occurs over a broad range of tissue types achieving therapeutic concentrations for both drugs. the accumulation in specific tissues (spleen, intestine) seems to be in line with the observed toxicity in patients. moreover, in certain tissue types crizotinib appears to enhance the exposure to gemcitabine suggesting the need for further development of this combination, especially with regards to pdac. acknowledgements this research was supported by a grant of the cancer center amsterdam to elisa giovannetti and g.j. peters. references [1] n.p. van erp, h. gelderblom, h.j. guchelaar, cancer treat. rev. 35 (2009) 692-706. [2] c. carmi, m. mor, p.g. petronini, r.r. alfieri, biochem. pharmacol. 84 (2012) 1388-1399. [3] a. avan, v. caretti, n. funel, et al., cancer res. 73 (2013) 6745-6756. [4] r.e. shackelford, m. vora, k. mayhall, j. cotelingam, genes cancer 5 (2014) 1-14. [5] s.h. ou, drug des. devel. ther. 5 (2011) 471-485. [6] a.t. shaw, u. yasothan, p. kirkpatrick, nat. rev. drug. discov. 10 (2011) 897-898. [7] m. el-naggar, e. giovannetti, g.j. peters, current pharm. design 18 (2012) 2811-2829. [8] r. honeywell, k. yarzadah, e. giovannetti, et al., j. chromatogr. b. analyt. technol. biomed. life sci. 878 (2010) 1059-1068. g.j. peters et al. admet & dmpk 4(4) (2016) 327-334 334 [9] r.j. honeywell, e. giovannetti, g.j. peters, nucleosides nucleotides nucleic acids 30 (2011) 12031213. [10] j.r. kroep, g.j. peters, r.a. nagourney, in: cancer drug discovery and development: deoxynucleoside analogs in cancer therapy (ed. g.j. peters), humana press inc., (2015) 253-288. [11] g. veerman, v.w. ruiz van haperen, j.b. vermorken, et al., cancer chemother. pharmacol. 38 (1996) 335-42. [12] y. esumi, k. mitsugi, h. seki, et al., xenobiotica 24 (1994) 957-964. [13] y. esumi, k. mitsugi, a. takao, et al. xenobiotica 24 (1994) 805-817. [14] j. sigmond, r.j. honeywell, t.j. postma, et al. ann. oncol. 20 (2009) 182-7. [15] p.e. kristjansen, b. quistorff, m. spang-thomsen, h.h. hansen, ann. oncol. 4 (1993) 157-60. [16] j. turk, k. bemis, w. colbert, et al., arzneimittelforschung 44 (1994) 1089-92. [17] l.a. shipley, t.j. brown, j.d. cornpropst, m. hamilton, w.d. daniels, h.w. culp, drug metab. dispos. 20 (1992) 849-55. [18] c.g. da silva, r.j. honeywell, h. dekker, g.j. peters., expert opinion drug metab. toxicol. 11 (2015) 703-717. [19] i.b. muller, a.j. de langen, r.j. honeywell, e. giovannetti, g.j. peters, exp. review anticancer therapy 16 (2016) 147-157. [20] a. avan, m. maftouh, n. funel, et al., curr. med. chem. 21 (2014) 975-89. [21] k.g. gotink, h.j. broxterman, m. labots, r.r. de haas, h. dekker, r.j. honeywell, m.a. rudek, l.v. beerepoot, r.j. musters, g. jansen, a.w. griffioen, y.g. assaraf, r. pili, g.j. peters, h.m.w. verheul, clin. cancer res. 17 (2011) 7337-7346. [22] l. mologni, expert opin. investig. drugs 21 (2012) 985-94. [23] y.p. mosse, m.s. lim, s.d. voss, et al., lancet oncol. 14 (2013) 472-80. [24] j.r. kroep, g.j. peters, c.j. van moorsel, et al., ann. oncol. 10 (1999)1503-10. ©2016 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ the role and impact of high throughput biomimetic measurements in drug discovery doi: 10.5599/admet.530 74 admet & dmpk 6(2) (2018) 74-84; doi: http://dx.doi.org/10.5599/admet.530 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index review the role and impact of high throughput biomimetic measurements in drug discovery shenaz bunally* and robert j. young* drug design & selection, glaxosmithkline medicines research centre, stevenage, sg1 2ny, uk. *corresponding authors e-mail: shenaz.b.bunally@gsk.com, rob.j.young@gsk.com; tel.: +44 1738 551372 received: march 29, 2018; revised: may 16, 2018; available online: may 25, 2018 abstract during the early phase of drug discovery, it is becoming increasingly important to acquire the full physicochemical profile of molecules. for this purpose, there is a strong interest in developing efficien t and cost-effective platforms for fast and reliable measurements of physicochemical properties. we have developed an automated physchem platform which ensures that consistent, comprehensive, and high quality physicochemical property measurements and derived property information for 100's of compounds per week are available alongside potency data at the right time to guide compound progression decisions. we discuss the routine assessments of biomimetic properties using high throughput automated high performance liquid chromatography (hplc) platforms, with details of the methods and hardware employed, also with illustrations of the quality and impact of the data generated. keywords lipophilicity; physicochemical properties; biomimetic chromatography, drug efficiency; candidate quality introduction the use of biomimetic/physicochemical measurements, such as lipophilicity and protein/artificial membrane binding, to help rationalise the behaviour of experimental molecules in biological environments is an important facet of modern drug discovery [1,2]. such measurements can be used not only as surrogates to model and predict behaviour, but also to estimate the quality of a given molecule and thus its chances of progression [2,3]; indeed, high clinical attrition rates have been attributed to sub-optimal physicochemical properties [4,5]. high quality, high throughput methods for biomimetic measurements are key elements of these approaches. the partitioning and distribution of drug molecules between bio-phases are fundamental to drug action [6], modelling and understanding these processes provides insight into absorption, distribution, metabolism and excretion, adme [6], the key elements of pharmacokinetics, the science of what the body does to a drug. the partition and distribution coefficients of drug molecules between 1-octanol and aqueous buffers (ow) are well-established standards, and these biomimetic estimates of lipophilicity/hydrophobicity demonstrably influence adme profiles and other outcomes. the negative impact of excessive lipophilicity on the chances of progression of experimental molecules has come under particular scrutiny over the past decade and changed practices in drug discovery are evident by recent improvements in the physicochemical quality of molecules [7]. lipophilicity measurements (such as log10 [ow-partition], log p or the distribution at a given ph, log dph) are demonstrably unreliable for poorly http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:shenaz.b.bunally@gsk.com mailto:rob.j.young@gsk.com admet & dmpk 6(2) (2018) 74-84 high throughput biomimetic measurements in drug discovery doi: 10.5599/admet.530 75 soluble compounds [8]. however, fast gradient reversed phase high pressure liquid chromatographic (hplc) methodologies using c-18 columns provide an effective and reliable replacement [9], irrespective of solubility [10]. analyses of data generated in this manner gave an improved resolution of adme outcomes; together with enhanced log p/log d predictions that have enabled the construction of better structurebased in silico predictive models. other non-silica based polymeric stationary phases are being investigated to provide insight into non-polar environments and the potential for intramolecular hydrogen bonding [11]. the grafting of stationary phases other than c-18, onto hplc columns, enables the high throughput assessment of additional pertinent adme interactions [12], in particular the plasma proteins human serum albumin (hsa) and alpha-1-acid-glycoprotein (agp), plus phosphatidylcholine, which acts as a surrogate immobilised artificial membrane (iam). the availability of these various biomimetic columns within fully automated hplc platforms enables the high throughput and cost-effective gathering of sets of pertinent and reliable data, which can provide valuable insight on the likely behaviours in biological systems. progressing compounds with good physicochemical properties are fundamental to pharmaceutical companies’ aspirations for the objective assessment of the qualities of lead and candidate molecules [13]. an automated platform ensures that consistent, comprehensive, and high-quality physicochemical property measurements and derived property information are available at the right time to guide compound progression decisions. the role and impact of biomimetic methods high throughput workflow for hplc based assays at gsk the physicochemical/biomimetic assays are bundled to provide kinetic solubility [8], lipophilicity and biomimetic binding data on the majority of project compounds during lead discovery and optimisation. the process for preparing sample plates which are “ready to run” on the hplc systems is shown in figure 1. at gsk all experimental compounds are routinely dissolved and stored at 10 mm in dimethyl sulphoxide (dmso) solution, for ease of automated handling. 50 l samples of this stock are dispensed in 96 well master plates for the kinetic solubility assay, merged to give fully populated plates. 5l of the solutions in the master plates are dispensed into daughter plates; standards and blanks are added to these plates and all the samples are diluted using the appropriate solvents such as dmso and iso-propanol/water (50/50 v/v) to produce daughter plates for the lipophilicity and biomimetic binding assays respectively. each daughter plate has a unique barcode used to generate the plate map. to ensure production of data of high quality and integrity, system and assay suitability checks are embedded in the sample process and data analysis. calibration data is monitored before and after running the samples. a set of “check” standards with known lipophilicity and biomimetic binding data are run during each sequence of samples and the data checked to ensure they meet the defined criteria. an in-house developed application (figure 2) retrieves the relevant compound information from the barcodes and creates a “worklist” which is uploaded into the hplc systems and used for running the samples. the application extracts the pertinent data from the raw data and places it into an excel spreadsheet where further data analysis is conducted. the automatic extraction of data is based on a set of user-defined rules which interrogate the chromatograms and flags anomalous data, such as multiple peaks. this allows more robust and efficient data processing and analysis. the generation of such volume of data has enabled the building of high quality in-house predictive models (discussed later in this publication) which are used for quality control. comparison of measured data with predicted data using these models is routinely performed to highlight any anomalous data. s. bunnally and r.j. young admet & dmpk 6(2) (2018) 74-84 76 figure 1. the gsk high throughput physchem sample preparation workflow in addition to the kinetic solubility data, this process enables the generation of high-quality lipophilicity and biomimetic binding data for 100’s of compounds to be generated weekly. high-performance liquid chromatographic (hplc) based assays horváth et al was amongst the first who used hplc data for hydrophobicity measurements of amino acids [14]. it has been well described that chromatographic retention is related to the compound’s dynamic distribution between the stationary and mobile phases and this is governed by a compound’s hydrophobicity [15-17]. hence hplc offers an excellent automated platform to determine distribution coefficients of biologically active compounds between aqueous mobile phases and various non-polar and biomimetic stationary phases through measurements of retention times. chromatographic hydrophobicity index from fast gradient c-18 hplc: setting new standards in lipophilicity/hydrophobicity determinations the lipophilicity values of virtually all new compounds at gsk is measured by reversed phase hplc using a c-18 column (50 x 2 mm 3 µm gemini nx c18, phenomenex, uk), at each of ph 2, 7.4 and 10.5, using buffered fast gradient acetonitrile-water mobile phases. the retention-time derived chromatographic hydrophobicity index (chi) values are derived directly from the gradient retention times by using a calibration line obtained for standard compounds [9]. translation of chi values into chrom log d values at the given ph is achieved using empirically-derived equation 1 [10]. there is a deliberate offset on the scale to differentiate the data from the traditional octanol-water measurements, but there is a high correlation between the two (for soluble compounds). figure 3 indicates how the charge profile of each compound can be estimated based on the changes in logd across the 3 ph values. for neutral molecules, the 3 are the same (i.e. the partition coefficient, chrom log p); additionally, the highest distribution constant value (for non-zwitterionic compounds) is usually a reliable estimate of the chrom log p of the molecule. chrom log d = (0.0857)*chi 2.00 . (1) admet & dmpk 6(2) (2018) 74-84 high throughput biomimetic measurements in drug discovery doi: 10.5599/admet.530 77 figure 2. use of the physchem application for data analysis charge characteristic chrom log d at ph: neutral weak acid strong acid weak base strong base 2 = x = x >x <x = x 7.4 x x x x x 10.5 = x <x = x = x >x figure 3. patterns of changes in log dph, at high or low ph, compared to the measured value, x, at ph 7.4, used to classify the strength of charged motifs in compounds. the influence and impact of these chromatographic measurements [1,10] reflect other observations on the crucial role of modulating lipophilicity in drug discovery [2,18], both through their impact in building rational understandings of both admet outcomes and chances of successful compound progression. increasingly, appreciation of the impact of maximising lipophilic ligand efficiency is driving drug discovery thinking; this embodies the “minimum lipophilicity principle” proposed by hansch [19], who proposed that “compounds should be made as hydrophilic as possible without loss of efficacy” by subtracting lipophilicity from potency (usually expressed as ligand lipophilicity efficiency, lle = pic50 – log p) [20]. furthermore, the principle of concurrently minimising lipophilicity and aromaticity [21] is represented by the property forecast index (pfi), the summation of aromatic ring count [22,23] and a lipophilicity measure. the measured pfi (chrom log d7.4 + #ar) is an integral part of gsk candidate quality aspirations [13], based on analyses of marketed drugs and internal attrition; ideally an oral candidate should have pfi <6, fasted-state simulated intestinal fluid (fassif) solubility > 100 g/ml and a predicted dose of < 100 mg. s. bunnally and r.j. young admet & dmpk 6(2) (2018) 74-84 78 protein binding assay chemically bonded hsa hplc stationary phase with column dimension of 50 x 3 mm (chiral technologies, france) are used for measuring compounds’ binding to plasma proteins by applying linear gradient elution up to 30 % iso-propanol with 50 mm ammonium acetate buffer, ph 7.4 [24]. the gradient retention times are standardised using a calibration set of mixtures. the %hsa bound gives a reliable indication of the free fraction of the compound in plasma when compared to more complex pharmacokinetic methods. the %hsa is converted to the affinity constant, log khsa, using equation 2: log khsa = log [hsa% / (101-hsa%)] (2) figure 4. a) levels of hsa binding by hplc measurement, log khsa plotted versus chrom log d7.4 with charges highlighted (see figure 3) and b) box-whisker plot of log khsa vs binned measured pfi for gsk compounds, wherein khsa = [%hsa binding/(101-%hsa binding)] using the %binding values derived from chromatographic measurements. analysis of data derived from the hsa measurements show a clear increase in hsa binding with increasing lipophilicity (figure 4a); when separated by charge, the increased propensity for binding by acidic a) b) admet & dmpk 6(2) (2018) 74-84 high throughput biomimetic measurements in drug discovery doi: 10.5599/admet.530 79 compounds, over and above their lipophilicity, is evident. the impact of aromaticity on hsa binding is also clear, given higher binding as pfi increases (figure 4b). phospholipid binding assay the binding of compounds to the immobilised artificial membrane (iam) [25] is measured using commercially available immobilised phosphatidylcholine (pc dd2 100 x 4.6mm 10 µm, regis analytical, west lafayette, usa) hplc columns [26]. gradient retention times obtained by applying an acetonitrile gradient up to 85 % are converted to chromatographic hydrophobicity indices (chiiam) using a calibration set of compounds. the chiiam values are converted to the logarithmic retention factors using the following formula: log kiam = 0.046*chiiam + 0.42. chiiam binding gives an indication of the compound’s likely binding to tissues and further insights are emerging [27], notably semi-quantitative indicators of the risk of phospholipidosis [28], a cytotoxicity outcome characterised by the breakdown of phospholipids [29], due to cationic amphiphilic drugs (cads). the gsk model is based on the equation cad likeness = chiiam + delta chi, whereby delta chi = (chiph10.5 –chiph7.4) as measured in the c-18 assays at the given ph values. unsurprisingly, given the hydrophobic chains of the phosphatidylcholine, iam binding is driven by lipophilicity, but, in contrast to acid-binding hsa, the net negative charge of the phosphates leads to enhanced binding of basic molecules (figure 5). figure 5. levels of iam binding by hplc measurement (expressed as log kiam = 0.046*chiiam + 0.42) versus chrom log d7.4 with bases highlighted drug efficiency, hplc demax drug efficiency (de, equation 3) is a concept based on measured pharmacokinetic parameters, designed to guide lead optimisation and developability assessment, which reflects the free plasma concentration at the site of action (expressed as the fraction of the dose) [30]. measured de data correlate with hplc demax values (equation 4), generated biomimetically [31], using a combination of hsa and iam columns figure 6. the empirically-derived model generated from the data [32] is based on the notion that the unbound concentration is influenced by both plasma protein binding (hsa data) and the volume of distribution, for which the hplc iam data provides an excellent surrogate for the contributory tissue binding [33]. increasingly, these measurements are having an impact in decision making, through estimation of clinical dose [34], and are being generated prospectively with in silico models available at gsk. the influence of s. bunnally and r.j. young admet & dmpk 6(2) (2018) 74-84 80 hplc demax data on the selection of compounds for progression is illustrated in figure 7 for a gsk programme, where most compounds had similar potencies (pic50 7 to 8) but a range of hplc demax values. the candidates selected from this set have hplc demax > 1 % and are in the same space as the profiled drugs. these hplc measurements, pertinent to demax, are increasingly being gathered in programmes at gsk and are starting to influence thinking, design and decision making. biophaseconc (µg/ml) drug eff %= x100 dose (µg/g) (3) log hplc demax = 2 – (0.23 * log khsa + 0.43 * log kiam -0.72) (4) drug efficiency max (demax) is the maximum in vivo drug efficiency that could theoretically be achieved assuming 100 % oral absorption, no clearance, free permeability, and no active transport. high potency plus high drug efficiency = lower dose lower dose leads to reduced off-target risks. this contributes to reduced attrition. figure 6. the plot of log (in vivo de) vs log (hplc demax) values for the training set of known drugs an additional parameter, the drug efficiency index (dei) [29], can be generated by the summation of pxc50 and log10 (hplc demax); dei gives an estimate of the likely effective activity at the site of action, i.e. potency corrected for the free concentration. impact of measurements to enable, validate and improve predictive methods data collected by the various biomimetic measurements has enabled the building of high quality inhouse predictive models of each descriptor (e.g. figure 8 for chrom log d7.4). good practice exploits iterative prediction/measurement cycles to build confidence in each series under optimisation; this also enables refinement of models on an ad hoc local basis in the rare cases that the global model does not perform well for a given structural series. the next level is to use these predictions as part of multivariate and other predictive models of various dmpk parameters (including drug efficiency), which are demonstrably improved by the enhanced predictions of physicochemical descriptors. together, the output admet & dmpk 6(2) (2018) 74-84 high throughput biomimetic measurements in drug discovery doi: 10.5599/admet.530 81 of these initiatives is enabling an aspiration to predict by first intent in the physicochemical design process with demonstrable impacts. figure 7. plot of potency vs log (hplc demax) values for programme compounds (green) overlaid with the training set of known drugs blue, with the candidates are chosen for progression in red. figure 8. trellised plot of calculated vs measured chrom log d7.4 for compounds in 6 distinct chemical series with lines of best fit and unity; the r 2 values illustrate the quality of the predictions s. bunnally and r.j. young admet & dmpk 6(2) (2018) 74-84 82 conclusions the extensive use of high throughput biomimetic measurements impacts on the drug discovery process in many ways. chromatographic lipophilicity measurements are at the core of medicinal chemistry programmes and can be used to predict outcomes, design better compounds and as quality indicators. the complementary measurements from other stationary phases such as hsa and iam are now routinely used; increasing awareness with demonstrations of their utility and predictive impact. this should give an enhanced influence to programme progression in the future. acknowledgements: we gratefully acknowledge the contributions of many colleagues in the gsk phys chem group, present and, largely past, notably klara valko and alan hill, plus many colleagues in computational and medicinal chemistry for model development and access to data and examples. references [1] r.j. young, physical properties in drug design, in: n.a. meanwell (ed.) tactics in contemporary drug design, springer berlin heidelberg, berlin, heidelberg, 2014, pp. 1-68. [2] n.a. meanwell. improving drug candidates by design: a focus on physicochemical properties as a means of improving compound disposition and safety. chemical research in toxicology 24 (2011) 1420-1456. [3] m.p. gleeson, p.d. leeson, h.van der waterbeemd, chapter 1. physicochemical properties and compound quality, the handbook of medicinal chemistry, the royal society of chemistry2015, pp. 1-31. [4] p.d. leeson, b. springthorpe. the influence of drug-like concepts on decision-making in medicinal chemistry. nat rev drug discov 6 (2007) 881-890. [5] p.d. leeson, r.j. young. molecular property design: does everyone get it? acs medicinal chemistry letters 6 (2015) 722-725. [6] m.p. gleeson, a. hersey, d. montanari, j. overington. probing the links between in vitro potency, admet and physicochemical parameters. nat rev drug discov 10 (2011) 197-208. [7] r.j. young, p.d. leeson. mapping the efficiency and physicochemical trajectories of successful optimizations. journal of medicinal chemistry in press (2018). [8] a.p. hill, r.j. young. getting physical in drug discovery: a contemporary perspective on solubility and hydrophobicity. drug discov today 15 (2010) 648-655. [9] k. valko, c. bevan, d. reynolds. chromatographic hydrophobicity index by fast-gradient rp-hplc: a high-throughput alternative to log p/log d. analytical chemistry 69 (1997) 2022-2029. [10] r.j. young, d.v. green, c.n. luscombe, a.p. hill. getting physical in drug discovery ii: the impact of chromatographic hydrophobicity measurements and aromaticity. drug discov today 16 (2011) 822830. [11] g. caron, m. vallaro, g. ermondi, g.h. goetz, y.a. abramov, l. philippe, m. shalaeva. a fast chromatographic method for estimating lipophilicity and ionization in nonpolar membrane-like environment. molecular pharmaceutics 13 (2016) 1100-1110. [12] k.l. valkó. lipophilicity and biomimetic properties measured by hplc to support drug discovery. journal of pharmaceutical and biomedical analysis 130 (2016) 35-54. [13] m.k. bayliss, j. butler, p.l. feldman, d.v. green, p.d. leeson, m.r. palovich, a.j. taylor. quality guidelines for oral drug candidates: dose, solubility and lipophilicity. drug discov today 21 (2016) 1719-1727. [14] c. horváth, w. melander, i. molnár. solvophobic interactions in liquid chromatography with nonpolar stationary phases. journal of chromatography a 125 (1976) 129-156. admet & dmpk 6(2) (2018) 74-84 high throughput biomimetic measurements in drug discovery doi: 10.5599/admet.530 83 [15] f. lombardo, m.y. shalaeva, k.a. tupper, f. gao. elogdoct:  a tool for lipophilicity determination in drug discovery. 2. basic and neutral compounds. journal of medicinal chemistry 44 (2001) 24902497. [16] a. vailaya, c. horváth. solvophobic theory and normalized free energies of nonpolar substances in reversed phase chromatography. the journal of physical chemistry b 101 (1997) 5875-5888. [17] k. valkó, l.r. snyder, j.l. glajch. retention in reversed-phase liquid chromatography as a function of mobile-phase composition. journal of chromatography a 656 (1993) 501-520. [18] m.j. waring. lipophilicity in drug discovery. expert opin drug discov 5 (2010) 235-248. [19] c. hansch, j.p. bjorkroth, a. leo. hydrophobicity and central nervous system agents: on the principle of minimal hydrophobicity in drug design. journal of pharmaceutical sciences 76 (1987) 663-687. [20] a.l. hopkins, g.m. keseru, p.d. leeson, d.c. rees, c.h. reynolds. the role of ligand efficiency metrics in drug discovery. nat rev drug discov 13 (2014) 105-121. [21] t.j. ritchie, s.j. macdonald. physicochemical descriptors of aromatic character and their use in drug discovery. journal of medicinal chemistry 57 (2014) 7206-7215. [22] t.j. ritchie, s.j. macdonald. the impact of aromatic ring count on compound developability--are too many aromatic rings a liability in drug design? drug discov today 14 (2009) 1011-1020. [23] t.j. ritchie, s.j. macdonald, r.j. young, s.d. pickett. the impact of aromatic ring count on compound developability: further insights by examining carboand hetero-aromatic and -aliphatic ring types. drug discov today 16 (2011) 164-171. [24] k. valko, s. nunhuck, c. bevan, m.h. abraham, d.p. reynolds. fast gradient hplc method to determine compounds binding to human serum albumin. relationships with octanol/water and immobilized artificial membrane lipophilicity. journal of pharmaceutical sciences 92 (2003) 22362248. [25] c. pidgeon, s. ong, h. liu, x. qiu, m. pidgeon, a.h. dantzig, j. munroe, w.j. hornback, j.s. kasher. iam chromatography: an in vitro screen for predicting drug membrane permeability. journal of medicinal chemistry 38 (1995) 590-594. [26] k. valko, c.m. du, c.d. bevan, d.p. reynolds, m.h. abraham. rapid-gradient hplc method for measuring drug interactions with immobilized artificial membrane: comparison with other lipophilicity measures. journal of pharmaceutical sciences 89 (2000) 1085-1096. [27] f. tsopelas, t. vallianatou, a. tsantili-kakoulidou. advances in immobilized artificial membrane (iam) chromatography for novel drug discovery. expert opinion on drug discovery 11 (2016) 473488. [28] j.-m. alakoskela, p. vitovič, p.k.j. kinnunen. screening for the drug–phospholipid interaction: correlation to phospholipidosis. chemmedchem 4 (2009) 1224-1251. [29] n. anderson, j. borlak. drug-induced phospholipidosis. febs letters 580 (2006) 5533-5540. [30] s. braggio, d. montanari, t. rossi, e. ratti. drug efficiency: a new concept to guide lead optimization programs towards the selection of better clinical candidates. expert opinion on drug discovery 5 (2010) 609-618. [31] k.l. valkó, s.b. nunhuck, a.p. hill. estimating unbound volume of distribution and tissue binding by in vitro hplc-based human serum albumin and immobilised artificial membrane-binding measurements. journal of pharmaceutical sciences 100 (2011) 849-862. [32] k. valko, e. chiarparin, s. nunhuck, d. montanari. in vitro measurement of drug efficiency index to aid early lead optimization. journal of pharmaceutical sciences 101 (2012) 4155-4169. [33] k.l. valko, s.p. teague, c. pidgeon. in vitro membrane binding and protein binding (iam mb/pb technology) to estimate in vivo distribution: applications in early drug discovery. admet and dmpk 5 (2017) 14-38. s. bunnally and r.j. young admet & dmpk 6(2) (2018) 74-84 84 [34] s. teague, k. valko. how to identify and eliminate compounds with a risk of high clinical dose during the early phase of lead optimisation in drug discovery. european journal of pharmaceutical sciences 110 (2017) 37-50. ©2018 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ doi: 10.5599/admet.4.3.286 241 admet & dmpk 4(3) (2016) 241-260; doi: 10.5599/admet.4.3.286 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper dose optimization of ceftriaxone-vancomycin combination using fractional inhibitory concentration kinetics in resistant bacteria vishnu d sharma*, aman singla, manu chaudhary, mukesh kumar, anuj bhatnagar, shailesh kumar, manish taneja* venus medicine research centre, hill top estate, bhatoli kalan, h.p., india *corresponding authors. e-mail: vishnudutt070788@gmail.com & manishtaneja24@gmail.com; tel.: +91-9871681044 received: may 02, 2016; revised: july 12, 2016; published: september 30, 2016 abstract the notorious staphylococcus aureus resistant strains with ever changing resistance patterns have limited treatment options and have led to substantial number of deaths. almost dried antibiotic pipeline has led us to look into combinations of already approved antibiotics for tackling rising incidence of antibacterial resistance. recommended use of vancomycin and ceftriaxone together for treating severe infections involving resistant s. aureus is limited by dose adjustments and different dose frequencies. we have developed a pharmacodynamically synergistic fixed dose combination (fdc) of ceftriaxone and vancomycin (2:1), for eliminating individual component dose adjustments and frequencies. for identification of optimum exposureresponse of fdc, one compartment in vitro system was used for dose escalation, fractionation and doseresponse studies. the in-silico pharmacokinetic/pharmacodynamic (pk/pd) modeling, simulations and validations were done. the results suggested % t>miccomb (percentage of time fractional inhibitory concentrations of the drugs combined remained above the miccomb [minimum inhibitory concentration for fdc]) followed by auccomb/miccomb (ratio of area under fractional inhibitory curves to miccomb) can predict the exposure (dose of fdc)-response (reduction in bacterial load) relationships effectively (r2 >0.9). total exposure of 6 g in two divided doses (3 g each) was identified to be optimum. monte carlo simulations were performed to evaluate the effect of increasing doses against different mics. clinical breakpoint of the fdc was identified to be 4 µg/ml, which was 2 fold higher than that of vancomycin suggesting better antibacterial coverage. keywords pk/pd modelling; fixed dose combination; monte carlo simulations; antibacterial resistance introduction the bacteria methicillin-resistant staphylococcus aureus (mrsa) is one of most common pathogen responsible for nosocomial infections. it is associated with an increased risk of mortality and morbidity, requiring prolonged duration of treatment and increased cost [1]. it is reported that out of 10,800 deaths per year by infections in united states, 5,500 deaths are linked to mrsa. the prevalence of mrsa is extrapolated to 1 billion people by year 2022 [1]. vancomycin is the empirical choice for treating these infections. however, extensive vancomycin use has led to emergence of mrsa with reduced vancomycin susceptibility. in 2006, the minimum inhibitory http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:vishnudutt070788@gmail.com mailto:manishtaneja24@gmail.com vishnu sharma et al. admet & dmpk 4(3) (2016) 241-260 242 concentration (mic) breakpoints for s. aureus were lowered (from ≤4 µg/ml to ≤2 µg/ml for “susceptible,” from 8–16 µg/ml to 4–8 µg/ml for “intermediate,” and from ≥32 µg/ml to ≥ 16 µg/ml for “resistant”) and vancomycin was less likely to respond when mics were ≥4µg/ml [2]. the clinical relevance of change in breakpoint was further highlighted by the observation wherein subjects with high grade infections required high vancomycin mics leading to therapy failures or increase in mortality. mechanistically, vancomycin binds to the terminal d-alanyl-d-alanine moieties of the n-acetyl muramic acid/n-acetyl glucosaminepeptides, subsequently prevents the action of penicillin binding proteins and stops peptidoglycan growth (leading to bacterial killing). however, in case of resistant bacteria (s. aureus), the last d-alanyl moiety is replaced by d-lactate, which makes bacteria less permeable to vancomycin (vancomycin entrapment and increase membrane thickness). this “permeability-mediated” resistance manifested clinically by treatment failure or delayed resolution of s. aureus infections [3,4]. addition of an antibiotic that can resolve the permeability issue is a rationale choice. it was demonstrated by our group previously [5] that ceftriaxone plus vancomycin formulation (2:1 w/w ratio), exploits the virgin territory of the vancomycin resistant bacteria i.e. penicillin binding proteins. ceftriaxone bind irreversibly with these proteins and thus inhibit the transpeptidation step of peptidoglycan synthesis to stop bacterial cell wall growth. ceftriaxone-induced fragility in the peptidoglycan layer increases the penetrability of the vancomycin and restores (to some extent) its affinity towards the exposed d-alanine moieties. the 'double insult' caused by the antibiotic fdc makes bacterial cell susceptible to death. after pharmacodynamic selection of drugs combination to combat mrsa, the second step is to optimize dose regimens using their pharmacokinetic/pharmacodynamic (pk/pd) relationships [6,7]. the dosing regimens of antibiotics are based on mic estimates (a surrogate pd marker for antibacterial response characterization) and the quantification of exposure (changes in concentration of individual components of dose combination) response (the reduction of bacterial count) relationship [6,7]. the exposure-response relationships help in evaluating the pk/pd indices [8,9] and further identifying the efficacy drivers for understanding the potential activity of antibacterial agents. the three pk/pd indices that are routinely used for explaining the therapeutic efficacy in terms of achieving mic and mechanism based action of antibiotics include: 1) % t>mic time (% time for which the levels of antibiotic in serum/plasma exceed the mic); 2) area under the concentration-time curve (auc)/mic ratio; 3) peak plasma concentration (cmax)/mic ratio [10-12]. maximum literature for these kinds of pk/pd studies is available for single antibacterial agents [13,14]; the effect of combination therapy on the pk and pd, and subsequently on pk/pd indices is very limited [15]. the active components of the fixed dose combination (fdc) i.e. vancomycin (auc/mic ratio is the preferred index) and ceftriaxone (%t>mic is the preferred index) have similar pk profiles but their optimized pk/pd indices correlative of therapeutic efficacy are different. the parameters which are usually used to show antibiotic interactions for combination therapy are fic (fractional inhibitory concentration) indices. through this approach, the synergism or antagonism expressed as the maximum effect of the antibiotic combination can be calculated (fic index identified for this combination was <0.5 suggesting synergistic pharmacodynamic action [5], and estimations of pk/pd index on the basis of fics are considered as composite predictors. the purpose of the present study is to use in-silico pk/pd modeling and simulations strategies to identify the fic based pk/pd indices that can predict the therapeutic efficacy of combination therapy and further help in optimization of dosage regimen (figure 1) [16]. after identification of best pk/pd index driving the therapeutic outcome, monte carlo simulations were performed to understand the clinical admet & dmpk 4(3) (2016) 241-260 dose optimization of fdc against s. aureus resistant bacteria doi: 10.5599/admet.4.3.286 243 benefit of the fdc and 1000 subjects’ simulated data was generated using population pk parameters of individual components of the fdc [17]. the pta (probability of target attainment) of pk/pd index in the simulated data was calculated and the clinical breakpoint of the fdc was identified. the breakpoint analysis will not only help us in optimizing the use of antibiotic combination, but also in identifying the differences in resistance pattern for ceftriaxone/vancomycin (2/1) in comparison to vancomycin alone. figure 1. schematics of the approach used in the study. in vitro studies were performed to study exposureresponse relationships and pk/pd driver identification. using the information of in vitro studies, three approaches i.e. in vitro studies, pk/pd modeling and monte carlo simulations were employed for dose recommendations. materials and methods antibiotics, bacteria, and media the fdc of ceftriaxone/vancomycin (2/1) vials (vancoplus) were obtained from venus remedies ltd, h.p., india. genotypically characterized mrsa clinical isolates with reduced susceptibility to vancomycin were taken from clinical isolate bank of venus medicine research centre, baddi. the media mueller-hinton broth (becton dickinson, sparks, md) was used to perform all in vitro studies involving mic determination, bacterial kinetics, dose ranging, fractionation, and response studies. quantification of bacteria and antibiotic the bacterial density (cfu [colony forming units]) and mic determinations were conducted according to clinical laboratory standard institute guidelines [18]. the fic index was determined using checkerboard method [5]. the media samples were analyzed for ceftriaxone and vancomycin concentrations using high performance liquid chromatography, as mentioned in our previous article [19]. one-compartment chemostat infection model single-compartment chemostat infection system as described previously [14,20,21] was used for the in vitro modeling of the fdc. in summary, the chemostat system assembly consisted of a 500 ml glass central reservoir chamber with five ports for the addition and removal of media via silicone tubes connected to peristaltic pumps, injection of drug (antibiotic combination) solution, and removal of medium samples. prior to each experiment, mrsa colonies were grown overnight to obtain a starting inoculum of 108 cfu/ml in 500 ml central reservoir flask containing media. an aqueous solution of fdc was prepared. the vishnu sharma et al. admet & dmpk 4(3) (2016) 241-260 244 “in” and “out” flow rates from the central reservoir were adjusted to obtain half-life of ~5-6 h [22-24]. samples from the central reservoir were collected at different time points from 0 to 24 h post-fdc addition. the samples were analysed for individual component concentrations and cfu determination for bacterial densities. the bacteria growth control experiments were also performed using the same experimental set-up without adding any drug. net pd effects of all doses of fdc were then reported after accounting the bacterial dynamics of the growth control experiment. pre and post fdc exposure, mics were determined for evaluating any changes in mics due to fdc exposure. also the pk parameters calculated in in vitro chemostat infection model for individual drugs were in concordance with in vivo literature values (table 1) supporting the validity of in vitro systems. dose escalation, fractionation and response curve studies using in vitro chemostat model the drug concentrations were varied from 0.5-200 folds mics to determine the most efficacious exposure based on their pk/pd parameters. the best exposure from dose escalation study was then fractionated twice-a-day and thrice-a-day to determine a dosage regimen of antibiotic which can enhance bacterial killing at lower concentration. the best dosage regimen was then used to construct dose response curve. the dosage regimen at 75, 50, 38, and 25 % of original concentration were utilized and the corresponding change in bacterial density in 24 h was measured. the dose response curve was plotted between the dose percentage and log reduction in bacterial density and ec50 was calculated. pharmacokinetic analysis one compartment model was used to fit the time concentration values and pk parameters of ceftriaxone and vancomycin were calculated. the parameters included the area under the plasma concentration-time curve (auc), maximal concentration reached (the peak; cmax), elimination rate constant (ke), half-life (t1/2), volume of distribution (vd), and clearance (cl). these parameters were calculated for all dose escalations and dose fractionation studies. the parameters obtained from concentration profile data were compared with reported population pk parameters in the literature (table 1). pharmacokinetic/pharmacodynamic analysis the fic curves for ceftriaxone and vancomycin combinations were generated. briefly, the concentrations of each antibiotic at every time-point were divided by their respective mic contributions towards miccomb to obtain ficceftriaxone and ficvancomycin values. these values were added (ficcomb) for each time points and plotted against time. the resultant fic-time profile was fitted to one compartment model to obtain pk/pd parameters i.e. auccomb, cmax-comb, %t>miccomb and %t>ficcomb [9,16]. in addition, the pk/pd indices i.e. auc/mic, cmax/mic and %t>mic were also calculated for both ceftriaxone and vancomycin using their respective pk parameters. admet & dmpk 4(3) (2016) 241-260 dose optimization of fdc against s. aureus resistant bacteria doi: 10.5599/admet.4.3.286 245 figure 2. the inter-relationship between the time dependent changes in drug concentrations (pk-intravenous (iv) infusion) and the bacterial growth inhibition/bacterial killing (pd) after antibacterial treatment. the parameters are c, the concentration of combination or individual component of fdc in central drug compartment; ke, drug elimination rate constant; kgrowth and kdeath, rate constants for multiplication and degradation of bacteria, respectively and b is bacterial load at a given time point. semi-mechanistic pharmacokinetic/pharmacodynamic modeling and validation bacterial modeling (control experiments): the bacterial kinetics determination involves modeling of a single bacterial compartment (b) with first-order rate constants for bacterial multiplication (kgrowth) and bacterial death (kdeath) [25]. the equation 1 explains the observed exponential growth of bacteria until it reaches a stationary bacterial level without the addition of antibiotics (control experiments) and where b and bmax are initial bacterial density and maximum bacterial density [26]. (1) pharmacokinetic/pharmacodynamic modeling: the bacterial load was evaluated at 8, 9.5, 12 and 24 h post antibiotic exposures. the 8, 12 and 24 time-points were selected to evaluate bacterial load in case of thrice-a-day, twice-a-day and once-a-day dose regimens respectively; whereas, 9.5 h time-point was selected for internal evaluation as the second inoculum was added to in vitro system at 9.5 h to mimic harsher in vivo conditions. the relationship between the effect and the corresponding pk/pd indices was evaluated according to a sigmoidal emax type function as described in equation 2 [9,6]. the emax model describes the concentration–effect relationship in terms of a baseline effect or e0, hence taking into account the baseline condition which can significantly affect the maximum activity of antibiotic i.e. pdmax, a descriptor of direct response of antibiotic. additionally, the function asymptotes to an upper limit of inhibition by an antibiotic on bacterial load. (2) where, e is the pd endpoint i.e. bacterial density calculated as change in log10 cfu/ml after 8, 9.5, 12 or 24 h of treatment, e0 is the baseline effect i.e. pd endpoint without antibiotic treatment; x is pk/pd index; vishnu sharma et al. admet & dmpk 4(3) (2016) 241-260 246 pdmax is a maximum effect; ex50 is magnitude of x that is needed to achieve 50 % of the pdmax; hill is the sigmoidicity factor, reflecting the steepness of the relationship. all data from the different dosing regimens was fitted to above-mentioned model (equation 2). curve fitting was performed in graphpad prism (version 4.01, graphpad software, san diego, ca) using the nonlinear regression analysis. the coefficients of determination (r 2 ), sigmoidicity factor, and the visual inspection of observed versus predicted values graphs were used to select the best pk/pd index and the best predicted endpoint of antibacterial effect. anti-bacterial-pk/pd modeling: basic assumption for the dependence of antibacterial effect is generally based on non-linear relationship with concentration data of antibiotics. higher the sigmoidicity factor, lesser is the predictability of pd effect with respect to pk/pd index. the concentration-effect relationships were incorporated in the bacterial model (equation 1) to predict the bacterial count from pk/pd model [26-29] (equation 3). in the constructed model, the antibacterial effect of fdc of ceftriaxone/vancomycin is hypothesized as a combination of bacterial growth inhibition (kgrowth) and bacterial killing enhancement (kdeath) (figure 2). (3) model evaluation all data from the different drug exposure and dosage regimen was fitted to above-mentioned model (equation 3) for the time period of 8 and 24 h respectively using scientist (micromath, version 3.0, saint louis, missouri, usa). the 95 % confidence intervals were drawn for all the predicted values and compared with the observed values (obtained from in vitro system). additionally, the coefficients of determination (r2) between predicted and observed data values were determined and evaluated for the validation of the antibacterial-pk/pd model. monte carlo simulations monte carlo simulations for 1000 adult subjects were performed to determine how likely the fdc dose of 0.75, 1.5, 3, 6, and 9 gm would achieve auccomb > 400 and 60 %t>miccomb at different values of mic i.e. 1, 4, 8, 16, 3, 64 µg/ml. the population pk parameter utilized in simulations was clearance of 0.88 +/ 0.431 l/h [30] and 4.8 +/-1.9 l/h [31] for ceftriaxone and vancomycin, respectively. the patient weight is an important covariate on drug’s clearance, which was assumed to be 70 kg in the analysis. for auccomb, each dose was divided by the clearance values of simulated subjects at all mic values. the pta for auccomb was defined as percentage of the simulated subjects in which auccomb was more than desired values (i.e. 400 in present studies). for %t>miccomb, the concentration-profiles were generated using cl and vd of each simulated subjects and %t>miccomb was calculated for all simulated subjects for different doses at all mic values. the pta was then defined as percentage of simulated subjects showing %t>miccomb of more than 60. data analysis descriptive statistics were used for reporting all pk variables and summary tables were prepared using mean, standard deviation (sd), median, and range (whichever appropriate). log transformed data was used wherever applicable. the statistical analysis was done using graphpad prism (version 4.01, graphpad software, san diego, ca). non-linear regression analysis and logistic regression was done for bacterial admet & dmpk 4(3) (2016) 241-260 dose optimization of fdc against s. aureus resistant bacteria doi: 10.5599/admet.4.3.286 247 kinetics using scientist (micromath, version 3.0, saint louis, missouri, usa). monte carlo simulations were done using macros in ms excel 2010. results and discussion the major problem with antibacterial resistance is limited treatment options. with limited antibiotic research, fdc of already approved antibiotic may still provide certain answers to tackle the resistance [32]. however, the major concern with antibiotic fdcs utilization is their dose and frequency of dosing. focusing on dose optimization, it mainly depends upon the pharmacokinetic, and pharmacodynamic properties of a drug. the complexity increases with addition of another active drug in the formulation, similar to the one presented in the study. the current formulation is a fdc of ceftriaxone and vancomycin in 2:1 (w/w), results in reduction of combined mic from 16 to 4 µg/ml against a mrsa bacterial strain used for this study. the fic index was <0.5 again confirms the synergism offered by fdc [5]. the present study was focused on dose optimization of the fdc using in vitro studies, pk/pd modeling and monte carlo simulations. exposure-response relationship assessment using in vitro studies the in vitro models allow good prediction of in vivo bacterial growth in vivo, helps in comparing different dosing regimens of drugs, and thus play an important role in dose optimization [21]. the selection of a specific in vitro model is determined by the objectives of a pk/pd study. the usual method for testing in vitro antibiotic combinations involves the exposure of a given bacterial inoculum to a static concentration of the anitbiotics. however, bacteria at the site of an infection are exposed in vivo to varying concentrations of antibiotics. therefore, in the present study, we have employed dynamic dilution in vitro system i.e. one-compartment chemostat infection model to mimic the in vivo conditions with respect to the changing drug and medium much more closely. chemostat infection model provides optimal bacterial growth conditions by replenishing nutrients via fresh media, and removing drug/wastes/old media, hence mimicking closely the biological dynamics of bacteria during human infection. the in vitro system simulates 1st order drug elimination leading to exponential decrease in drug concentration. as both vancomycin and ceftriaxone concentration-time plots follow one compartment pk model, the model was apt to fit the first order kinetics of both drugs. the in vitro model for dose optimization requires pk and pd inputs to evaluate the pk/pd of a given exposure of drug against an infection. for pk input, the pk parameters reported in literature for ceftriaxone and vancomycin were employed (table 1). focusing on pd target, mrsa isolates were selected as a prototype of mrsa infections. microbiological studies were then performed for the selection of the target strain against which exposure-response relationship can be evaluated. it was observed that mrsa strain of mic 4 µg/ml exhibits dose-dependent behaviour against the fdc, which underlines its suitability in exposure-response studies (results not shown). the pd effect was defined as logarithmic reduction in mrsa bacterial densities over 24 h time period. the mrsa infection was treated with seven different exposures (0.5-200 folds of mic 4 µg/ml) of the fdc for 24 h to find pk/pd driver that can best describe the pd effect. it was noted that bacterial killing increase with increase in drug exposure till the saturation is achieved (figure 3 and table s1). after saturation (>50 fold mic), there was no substantial increase in antibacterial effect. thus 50 folds of mic (133.5 µg/ml of ceftriaxone and 66.75 µg/ml of vancomycin) was the lowest exposure for which maximum pd effect was observed; and thus the exposure (50x mic) was selected for further dose optimization studies. vishnu sharma et al. admet & dmpk 4(3) (2016) 241-260 248 figure 3. total exposure in terms of fold in mics was plotted against log fold change in bacterial density (cfu). downward arrow represents the 50 fold mic where the saturated pharmacodynamic response was observed. the same drug exposure was thus used in dose fractionation studies for dose optimization using in vitro studies. table 1. pharmacokinetic parameters estimated in chemostat infection model and its comparison to reported values in literature parameters as reported in literatures estimated from in vitro system ceftriaxone vancomycin ceftriaxone vancomycin clearance (l/h) 0.88 ± 0.431 [30] 4.8 ± 1.9 [31] 1.53 ± 0.11 3.59 ± 0.11 volume of distribution (l) 10.1 ± 1.0 [24] 25.8 [33] 10.79 ± 1.14 22.3 ± 0.67 half life (h) 5.9 ± 0.7 [24] 4-6 [34] 4.83 ± 0.71 4.3 ± 0.33 selection of pharmacokinetic/pharmacodynamic driver that can predict fdc’s antibacterial efficacy the routine pk/pd indices i.e. auc/mic, cmax/mic and %t>mic were calculated for ceftriaxone and vancomycin individually, for all exposures. to understand of pk/pd driver responsible for therapeutic potential of the fdc, the fic-time profiles were generated from concentration-time profile of both drugs for all doses as described above (see pk/pd analyses in materials and method section). the pk/pd indices i.e. auccomb, cmax-comb, %t>miccomb and %t>fic were calculated from the fic-time curves. all pk/pd indices and their respective logarithmic transformed values (excluding %t>miccomb and %t>fic) were plotted against the logarithmic reduction in bacterial densities of mrsa in 8, 9.5, 12 and 24 h. the time periods of 8, 12 and 24 h were chosen considering thrice-a-day, twice-a-day and once-a-day regimen of the fdc. the time point of 9.5 h was included because second inoculum was added at this time point which might affect the net reduction in bacterial density. sigmoidal emax type function (see materials and methods, equation 2) was used for fitting the exposureresponse dat; and coefficient of correlation (r 2 ), and sigmoidal factor (γ) were computed. sigmoidal factor is the steepness of the sigmoidal curve and can suggest the predictability of the sigmodial model. for instance, when the steepness is high, the exposure-response relationship becomes steep enough for the all-or-none effect, which is not desired in identification of pk/pd drivers which is intended to correlate/simulate the dose-dependent sigmoidal drug effect. it must be noted that the intial admet & dmpk 4(3) (2016) 241-260 dose optimization of fdc against s. aureus resistant bacteria doi: 10.5599/admet.4.3.286 249 concentration or cmax is predecided when the drug was injected in in vitro system, as shown by higher coefficients of correlation of log cmax vs log δcfus (figure 4b). thus, cmax curves were not further used to identify the pk/pd driver of the current fdc. ceftriaxone individually (r 2 of 0.82 for %t>mic) showed the time-dependent antibacterial effect of the drug, which also confirms with the literature [35] (figure s1a). since, auc is directly dependent upon the drug concentration and time, auc/mic (r 2 =0.95) also showed good correlation with change in bacterial density (figure s1a). for vancomycin, lower r 2 value of 0.38 for %t>mic vs log δcfus and higher r 2 value of 0.97 for log auc/mic vs log δcfus suggest exposuredependent antibacterial effect and this is in agreement with the literature (figure s1b) [33, 36-37]. figure 4. the pk/pd indices i.e. auccomb, cmax-comb, %t>miccomb and %t>fic were calculated for fdc using ficcomb curves. the log transformations were done for auc/miccomb and cmax-comb/mic. the log transformed pk/pd indices and %t>mic were plotted against log fold change in bacterial density. the data was fitted with a sigmoidal emax type function. model fitted values of fdc’s pk/pd index auccomb (a), log cmax-comb (b), %t>miccomb (c), and %t>fic (d), and its relationship with log fold change in bacterial density were shown. since the pk/pd driver of ceftriaxone and vancomycin were different, fics were used in order to combine the contribution of both drugs as per their concentration and mic contribution in the fdc. the objective was to find the pk/pd index that can predict the antibacterial effect of the ceftriaxone/vancomycin (2/1) as a single unit. from fic-t profiles, auccomb, cmax-comb , %t>miccomb, and %t>fic were calculated and plotted against bacterial killing (figure 4). as shown in figure 4c, the r 2 is highest for %t>miccomb i.e. 0.98 with high predictability (γ = 2.262). the second best r 2 of 0.97 was obtained from the plot of log auccomb vs log δcfu (γ =13.51; good predictability) (figure 4a). for %t>fic, r 2 was 0.97 but γ was very high at 153.8 (figure 4d). high value of sigmoidicity factor suggest all-or-none function of the system and thus decrease the predictability of model. therefore, %t>miccomb was considered as a best pk/pd driver followed by auccomb, to predict the pd effect of ceftriaxone/vancomycin (2/1) in the study. vishnu sharma et al. admet & dmpk 4(3) (2016) 241-260 250 probability target attainment levels after selection of pk/pd driver (i.e. %t>miccomb and auccomb), the next objective was to identify a desired value of %t>miccomb and auccomb that one should target to assure good therapeutic effect of the fdc. thus, plot of %t>mic vs. δ log10cfu was evaluated to estimate desired %t>mic value (figure 4c). it was clearly observed that pd effect of the fdc become saturated after 60 %t>mic, prompting us to select 60 as a target value of pk/pd i.e. %t>miccomb for dose optimization. in other words, one can assure good therapeutic effect of the fdc, if the drug concentration in body remains above mic for more than 60% of time of drug exposure. using the same rationale and figure 4a, the target value of 400 (~antilog of 2.6) was used for the second best pk/pd driver i.e. auccomb. dose optimization and recommendation the pk/pd drivers of fdc and its desired values were further utilized for dose optimization and recommendation using three approaches i.e. in vitro approach, pk/pd modeling and monte carlo simulations. these approaches use same piece of information and provide inferential data for dose optimization which vary from each other in terms of specificity of information. the first approach (in vitro system) mainly tells about the antibacterial potential of the fdc under a specific mic. the pk/pd model is generally employed for dose individualization; whereas, monte carlo simulations broadly recommends dose regimen based on mic of the target infection. all three approaches were utilized to understand the holistic view and are explained below in detail. dosage optimization using in vitro system: the pd effect being driven by the time-dependence of fdc, encouraged us to fractionate the dose, with an objective to increase the duration of exposure (above mic) and decrease the overall concentration (to reduce side effects). for dose fractionation studies, dose exposure of 50x mic i.e. 133.5/66.75 µg/ml (ceftriaxone/vancomycin) was selected. considering the antibiotic half-life and mic of drug against mrsa strains, the dose was fractionated to mimic a dosage regimen of once-a-day, twice-a-day, and thrice-a-day (figure 5). similar to dose escalation study, the in vitro system was used for these dosage regimens and results are shown in figure 5 and table s1. the antibacterial effect of the fdc was increased from once-a day dosing to twice-a-day dosage regimen in 24 hr time period (4.058 fold to 5.550 fold reduction in bacterial density). further fractionation of dose in thrice-a-day dosing decreased the drug concentration below mic for longer time period resulting in reduced ‘effective’ exposure time i.e. %t>mic and translated into sub-optimal 2.32 fold reduction in bacterial density. the selected dosage regimen roughly corresponds to 3 g fdc of ceftriaxone (2 g) and vancomycin (1 g), after correcting for protein binding of 80 and 40 %; and volume of distribution of 10.1 l and 25.8 l of both components i.e. ceftriaxone and vancomycin respectively was observed (table 1). the mic values remained same during 24 h drug-bacterial incubation period for all drug exposures (data not shown), thus ruling out the possibility of antimicrobial resistance development in mrsa isolates during in vitro studies. admet & dmpk 4(3) (2016) 241-260 dose optimization of fdc against s. aureus resistant bacteria doi: 10.5599/admet.4.3.286 251 figure 5. selected 50 fold mic concentrations of the fixed dose combination of ceftriaxone/sulbactam (2/1) were fractioned into once-a-day (od), twice-a-day (bd) and thrice-a-day (tid) dosage regimens, and were plotted against log fold change in bacterial density (cfu). dosage optimization using pk/pd modeling: pk/pd model development: the pk/pd model was developed using equation 3 of sub-section section 'antibacterial-pk/pd modeling' under 'materials and method' section. the net bacterial growth (knet) was determined from growth kinetics of bacteria without adding drug. for the determination of emax, ec50 and γ, dose response curve was constructed and analyzed using in vitro dilution system. depending upon the therapeutic action needed, the dose can be adjusted for enhanced pd effects using dose-response relationship (these relationships are reflections of the pharmacological actions of the drug). the mrsa infections were exposed to 25, 38, 50, 75 and 100 % of dosage regimen in in vitro chemostat system and the corresponding logarithmic change in bacterial densities of 0.08, 4, 4.6, 4.48, and 5.55 were observed respectively. the dose response curve was plotted (figure s2) and the data was fitted with sigmoidal function and ec50 of 37.46 % was calculated (defined as a percentage of dose required to attain the 50 percent of pd effect). this ec50 value was used along with other parameters i.e. emax, γ to create the semi-mechanistic antibacterial bacterial-pk/pd model. the model might help in determining the antibacterial response at a given concentration (and in clinical settings at given dose) and could be extrapolated to estimate dose requirements for an identified patient’s bacterial load in clinical settings. pk/pd model validation: the final pk/pd equation was validated to check its suitability in dose recommendation using scientist. the experimental data for dose escalation studies were fitted using equation-3. the goodness of fit and 95% confidence interval of simulated data were evaluated for first 8 h (before the addition of second inoculum). simulations after 8 h of drug exposure were not shown in case of dose-escalation study in order to focus mainly on initial exposure of drug, where most of the bacterial killing occurred. the semi-mechanistic-pk/pd model predictions were improved with an increase in dose exposure. for instance, r 2 values for the fitted values of dose exposures of 20x, 50x, 100x and 200x mic were 0.633, 0.856, 0.964 and 0.965 respectively (figure 6a, 6b, 6c and 6d) suggesting better predictability of model at comparitively higher exposures. additionally, most of the experimental data points were under 95 % confidence interval of simulated data, which further supports the validity of pk/pd model. vishnu sharma et al. admet & dmpk 4(3) (2016) 241-260 252 figure 6. the experimental data was fitted with the antibacterial-pk/pd model using non-linear least-square regression program scientist. (i) validation based on dose escalation studies data (a, b, c, d):the data of dose escalation from 20 folds to 200 folds of mic was validated with the model; (ii) validation based on dose fractionation studies data: the data of dose fractionation i.e. once-a-day (e), twice-a-day(f) and thrice-a-day (g) was validated with the model. the model (equation 3) was further applied to experimental data of dose fractionation. since, dose frequency is a main factor in dose fractionation data, model predictions were extended to 24 hours. the observed values of experiments and predicted values of model were plotted for all three dosage regimens i.e. once-a-day, twice-a-day and thrice-a-day (figure 6e, 6f and 6g). the addition of second inocula at 9.5 h had raised the cfu counts, as shown by corresponding downward peak in all three cases. also as mentioned before, model was weak for lower concentrations of drug, as observed in 20-24 h of once-a-day dosing (figure 6e) and 16-24 h of thrice-a-day dosing (figure 6g) of ceftriaxone/vancomycin (2/1). therefore, weak correlation was observed in all three cases as compared to dose escalation studies. the correlation coefficents increases from 0.6682 of once-a-day dosing to 0.6927 of twice-a-dosing of the fdc of antibiotics. further increase in dose fractionation reduces drastically the correlation coefficient to 0.3643, which suggests the model weakness in predicton of lower drug concentrations. however, the model was appropiate for the twice-a-day dose frequency, which was also the best dosage regimen for ceftriaxone/vancomycin (2/1) fdc combination. the same conclusion was derived from the 95% confidence interval of simulated data. most of the data of twice-a-day dosing were in the upper and lower limits of confidence interval, followed by once-a-day dosing (figure 6). again in case of thrice-a-day dose frequency, predicted values were poorly correlated with the experimental values (figure 6g). dose optimization using monte carlo simulations: the pk-pd studies were done using the fixed values of primary pk parameters and mic; which did not account the pk variability of a clinical population. in order to account for the randomness/variability in human pks and mics, monte carlo simulations of 1000 subjects were done to determine how likely the given exposure of fdc would achieve the target %t>miccomb or auccomb (reported as pta). the pk parameters utilized were clearance of 0.88 +/0.431 l/h and 4.8 +/-1.9 l/h and elimination rate constant of 0.113 and 0.11 h -1 for ceftriaxone and vancomycin respectively, on the basis of literature [30,31]. monte carlo simulations were done for both pk/pd indices i.e. auccomb and %t>miccomb (figure 7a and 7b). the selection of the pta target set-points (as described previously) were done based on the auccomb and %t>miccomb values which correlates best with maximal bacterial killing effect (4 fold log reduction in bacterial density). in the present study, the pta for auccomb admet & dmpk 4(3) (2016) 241-260 dose optimization of fdc against s. aureus resistant bacteria doi: 10.5599/admet.4.3.286 253 was defined as percentage of the simulated subjects in which auccomb was more than 400 [2]. for %t>miccomb, the pta was defined as percentage of simulated subjects showing %t>miccomb of more than 60. focusing on pta of auccomb ≥400, ceftriaxone/vancomycin (2/1) had achieved almost 98 % pta for all doses against the strains for which it exhibits mic of 1 µg/ml (figure 7a). also for the mic of 4 µg/ml, the lowest dose that can achieve the pta of more than 90 % was 1.5 g twice dosing. these strains were categorized as susceptible to the fdc. further increase in mic i.e. 8 µg/ml lowered the pta (31% at 1.5g twice-a-day) and could be improved with higher dose of the fdc (≈90% pta at 3.0g twice-a-day). however, strains corresponding to mic of 16, 32 and 64 µg/ml required large amount of fdc dose (>9g) to attain moderate pta (figure 7a). these strains were resistant to the therapeutic dose of fdc. figure 7. clinical breakpoint estimation (a, b): monte carlo simulations for 1000 subjects and pta of both pk/pd indices i.e. auccomb (a) and %t>miccomb (b) was evaluated in the simulated subjects. the pta for auccomb was defined as percentage of the simulated patients in which auccomb was more than 400. similarly pta for %%t>miccomb was 60. it was determined that how likely the fdc dose of 0.75, 1.5, 3, 6, and 9 gm would achieve auccomb/miccomb > 400 and 60%t>miccomb at different values of mic i.e. 1, 4, 8, 16, 3, 64 µg/ml. the same results were mirrored with more clarity in %t>miccomb based monte carlo simulations (figure 7b). it must be mentioned that the pk/pd index i.e. %t>miccomb was best pk/pd driver for fdc. as shown in figure 7b, the combination was very effective with pta of ≈99 % at all dose strengths, for the strains corresponding to mic of 1 and 4 µg/ml. for mic of 8 and 16 µg/ml, higher dose of fdc was required to revive its efficacy against the bacterial infection. interestingly, the pd effect against the strains of mic of 8 µg/ml, was tilted more toward susceptible range of mic (1 and 4 µg/ml), whereas strains corresponding to mic of 16 µg/ml was closer to resistant ones. following the same trend observed in case of auccomb (figure 7a), the fdc was ineffective at all drug exposures for the mic of 32 and 64 µg/ml. for these resistant strains, standard dose of 1.5 g and 3.0 g twice-a-dose were sub-optimal and higher doses were needed to achieve the antibacterial effect. thus, the strains corresponding to the mic of 1-4 µg/ml, 8-16 µg/ml and 32-64 µg/ml were susceptible, intermediate and resistant respectively to the fdc of ceftriaxone/vancomycin (2/1). this categorization is also defined as clinical breakpoints, which separates resistant bacterial strains from the sensitive ones against a drug in terms of its mic values. the clinical breakpoints obtained from monte carlo simulation, were compared with mic breakpoints obtained from in-house study on 500 clinical isolates of mrsa (mic and ast breakpoint estimation; table-s2). the results were in agreement with each other, suggesting promising outcomes in future clinical studies of the fdc ceftriaxone/vancomycin (2/1). these clinical breakpoints give a significant advantage over infections wherein vancomycin alone is vishnu sharma et al. admet & dmpk 4(3) (2016) 241-260 254 showing therapeutic limitations. for example the susceptibility of mrsa for fdc (vancomycin + ceftriaxone) is 2 fold higher than vancomycin (mic ≤ 4 µg/ml for fdc and ≤2 µg/ml for vancomycin) alone. in addition to mrsa susceptibility, the fdc provides broader antibacterial coverage (active against both gram positive and negative bacteria). the better clinical reach will help in tackling vancomycin mediated resistance more effectively. the cost-effectiveness study of the fdc with respect to other commercially available antibiotics is currently under investigation, as previously done by sharma et al. on similar fdcs [38]. our study mainly focused on in vitro pk/pd relationships of fdc antibiotics, however generalizing the inferences of this study to all fdc antibiotics will not be prudent. also, the in vitro-in vivo translation, and finally its applicability in clinical settings still needs to be evaluated. we have focused on only one bacterial isolate, and we understand that when more clinical isolates with different resistance patterns will be studied, the exposure targets may get modified. however, the results of this study not only give the starting point to use the fdc clinically but will also refine the clinical study and design for further clinical research work. conclusions to summarize, in vitro system was used to estimate pk parameters and pd effects for different exposure of the fdc of ceftriaxone/vancomycin (2/1). the best antibacterial effect was obtained from the exposure of 50x miccomb which when fractioned to twice-daily dosing showed maximum reduction in bacterial densities. the semi-mechanistic pk/pd model incorporating bacterial kinetics and drug’s pk/pd relationship showed good predictability (very week predictability at lower exposures) at therapeutic exposures. taking into the consideration the randomness and pk variability of subjects, monte carlo simulations revealed that clinical breakpoints of fdc for susceptible, intermediate and resistant strains were ≤ 4, 8-16, and ≥ 32 µg/ml, respectively. these breakpoints were 2 fold higher than vancomycin alone treatments. the same breakpoint values in in vitro (mic and ast breakpoint estimation) and simulated clinical population suggests that fdc combination exhibits same pd effect in vitro that it should show in clinical population. thus, the study corroborates the validity of antibacterial effect of fdc ceftriaxone/vancomycin (2/1 w/w/) in simulated population with unique advantage of giving both components together in similar dosing frequencies. acknowledgements: this research was funded by venus remedies ltd. special thanks are extended to analytical and microbiological department of venus remedies for analyzing the samples. references [1] e. klein, d.l. smith, r. laxminarayan, emerging infectious disease 13(12) (2007) 1840-1846. [2] f.c. tenover, r.c. moellering, clinical infectious diseases 44(9) (2007) 1208-1215. [3] b.p. howden, p.d. johnson, p.b. ward, t.p. stinear, j.k. davies, antimicrobial agents and chemotherapy 50(9) (2006) 3039-3047. [4] p.g. charles, p.b. ward, p.d. johnson, b.p. howden, m.l. grayson, clinical infectious diseases 38(3) (2004) 448-451. [5] e.m. o'shaughnessy, j. meletiadis, t. stergiopoulou, j.p. demchok, t.j. walsh, journal of antimicrobial chemotherapy 58(6) (2006) 1168-1176. [6] e.i. nielsen, l.e. friberg, pharmacological reviews 65(3) (2013) 1053-1090. [7] e.i. nielsen, a. viberg, e. lowdin, o. cars, m.o. karlsson, m. sandstrom, antimicrobial agents and chemotherapy 51(1) (2007)128-136. admet & dmpk 4(3) (2016) 241-260 dose optimization of fdc against s. aureus resistant bacteria doi: 10.5599/admet.4.3.286 255 [8] t. katsube, y. yano, y. yamano, t. munekage, n. kuroda, m. takano, journal of pharmaceutical sciences 97(9) (2008) 4108-4117. [9] e.i. nielsen, o. cars, l.e. friberg, antimicrobial agents and chemotherapy 55(10) (2011) 4619-4630. [10] b.d. vanscoy, r.e. mendes, m. castanheira, j. mccauley, s.m. bhavnani, r.n. jones, l.v. friedrich, j.n. steenbergen, p.g. ambrose, antimicrobial agents and chemotherapy 58(10) (2014) 6024-6031. [11] v.d. sharma, a. singla, m. chaudhary, m. taneja, international journal of pharmaceutical sciences and research 7(4) (2016) 1503-1512. [12] s.a. van wart, d.r. andes, p.g. ambrose, s.m. bhavnani, diagnostic microbiology and infectious disease 63(4) (2009) 409-414. [13] a. ogden, m. kuhn, m. dority, s. buist, s. mehrens, t. zhu, d. xiao, j.r. miller, d. hanna, antimicrobial agents and chemotherapy 56(1) (2012) 124-129. [14] k.d. leuthner, c. vidaillac, c.m. cheung, m.j. rybak, antimicrobial agents and chemotherapy 54(9) (2010) 3799-3803. [15] v.d. sharma, a. singla, m. chaudhary, m. taneja, international journal of pharmacy and pharmaceutical sciences 8(7) (2016) 346-353. [16] j.g. den hollander, j.w. mouton, h.a. verbrugh, antimicrobial agents and chemotherapy 42(4) (1998) 744-748. [17] v.d. sharma, a. singla, m. chaudhary, m. taneja, aaps pharmscitech 17(5) (2016) 1192-1203. [18] clsi. performance standards for antimicrobial disk susceptibility tests. clsi m02-a12 (2015) [19] a.s. tariq, j. kumar, d. reddy, m. chaudhary, r.k. singh, science asia (36) (2010) 297-304. [20] b.j. mcgrath, k.c. lamp, m.j. rybak, antimicrobial agents and chemotherapy 37 (9) (1993) 19311937. [21] j. gloede, c. scheerans, h. derendorf, c. kloft, journal of antimicrobial chemotherapy 65(2) (2010) 186-201. [22] a.j. larsson, k.j. walker, j.k. raddatz, j.c. rotschafer, journal of antimicrobial chemotherapy 38(4) (1996) 589-597. [23] j.h. martin, r. norris, m. barras, j. roberts, r. morris, m. doogue, g.r. jones, clinical biochemist reviews 31(1) (2010) 21-24 [24] i.h. patel, s. chen, m. parsonnet, m.r. hackman, m.a. brooks, j. konikoff, s.a. kaplan, antimicrobial agents and chemotherapy 20(5) (1981) 634-641. [25] w.j. jusko, journal of pharmaceutical sciences 60(6) (1971) 892-895. [26] j.w. mouton, a.a. vinks, n.c. punt, antimicrobial agents and chemotherapy 41(4) (1997) 733-738. [27] y. harigaya, j.b. bulitta, a. forrest, g. sakoulas, a.j. lesse, j.m. mylotte, b.t. tsuji, antimicrobial agents and chemotherapy 53(9) (2009) 3894-3901. [28] a. nolting, t. dalla costa, k.h. rand, h. derendorf, pharmaceutical research 13(1) (1996) 91-96. [29] j.b. bulitta, n.s. ly, j.c. yang, a. forrest, w.j. jusko, b.t. tsuji, antimicrobial agents and chemotherapy 53(1) (2009) 46-56. [30] d. garot, r. respaud, p. lanotte, n. simon, e. mercier, s. ehrmann, d. perrotin, p.f. dequin, c. le guellec, british journal of clinical pharmacology 72(5) (2011) 758-767. [31] j.y. chung, s.j. jin, j.h. yoon, y.g. song, journal of korean medical sciences 28(1) (2013) 48-54. [32] v.h. tam, m. nikolaou, plos computational biology 7(1) (2011) e1001043. [33] m.j. rybak, clinical infectious diseases 42 (2006) suppl 1:s35-39. [34] r.c. baselt, disposition of toxic drugs and chemicals in man. 9 edn. (2011) biomedical publications, foster city, ca [35] j.d. turnidge, clinical infectious diseases 27(1) (1998) 10-22 [36] s.j. vandecasteele, a.s. de vriese, e. tacconelli, journal of antimicrobial chemotherapy 68(4) (2013) 743-748. vishnu sharma et al. admet & dmpk 4(3) (2016) 241-260 256 [37] g.r. matzke, g.g. zhanel, d.r. guay, clinical pharmacokinetics 11(4) (1986) 257-282. [38] v.d. sharma, m. chaudhary, m. taneja, european journal of pharmaceutical and medical research 3(6) (2016) 454-462. ©2016 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ admet & dmpk 4(3) (2016) 241-260 dose optimization of fdc against s. aureus resistant bacteria doi: 10.5599/admet.4.3.286 257 supplemental materials supplement table 1: bacterial density reduction at various antibiotic exposures concentration used (ceftriaxone/vancomycin µg/ml) folds of mic log δcfu 1.335/0.667 0.5 -0.234 13.35/6.675 5 0.716 26.7/13.35 10 0.977 53.4/26.7 20 4.000 133.5/66.75 50 4.058 267.0/133.5 100 3.609 534.0/267.0 200 4.719 66.75/33.375 bd (total exposure: 133.5/66.75) 50 5.55 44.5/22.25 tid (total exposure: 133.5/66.75) 50 2.32 the concentrations of individual component of fdc and bacterial cfu-time profiles were obtained from in vitro chemostat model. bacterial killing increase with increase in drug exposure till the saturation is achieved i.e. from 0.5 -50 folds of mic and after saturation, further increase in drug concentration from 50 to 200 folds of mic, there was no substantial increase in anti-bacterial effect of antibiotic combination. further, selected total exposure (133.5/66.75 µg/ml) of 50 fold mic was fractionated into twice daily and thrice daily regimens. the twice daily regimen showed enhanced bacterial killing. cfu colony forming units; fdc fixed dose combination; mic minimum inhibitory concentration supplement table 2: clinical, mic and ast breakpoints of 150 bacterial isolates of mrsa breakpoints susceptible intermediate resistant clinical (µg/ml) ≤4 8-16 ≥32 mic (µg/ml) ≤4 8-16 ≥32 ast (mm) <14 14-17 >17 the in vitro antibacterial activity of fdc was evaluated against mrsa. the clinical breakpoint was estimated from monte carlo simulations. the mic breakpoints were determined based on target attainment of pk/pd index %t>mic (≈50% t>mic). the ast breakpoints were determined using least square regression analysis of zone diameter values with the corresponding mic values of the fdc combination against 150 bacterial isolates of mrsa. ast antimicrobial susceptibility test; cfu colony forming units; fdc fixed dose combination; mic minimum inhibitory concentration; mrsa methicillin resistant staphylococcus aureus; pk/pd pharmacokinetic/pharmacodynamic; % t>mic % time for which the levels of antibiotic in serum/plasma exceed the mic. vishnu sharma et al. admet & dmpk 4(3) (2016) 241-260 258 supplement figure 1: pharmacokinetic/pharmacodynamic indices of individual components of fdc and their relationship with log fold change in bacterial density the pk/pd indices i.e. auc/mic, cmax/mic and %t>mic were calculated for both ceftriaxone and vancomycin using their respective pk parameters. the log transformations were done for auc/mic and cmax/mic. the log transformed pk/pd indices and %t>mic were plotted against log fold change in bacterial density. the data was fitted with a sigmoidal emax type function . the selection of pk/pd driver was based on coefficient of correlation (r 2 ), and sigmoidal factor (γ: hill factor) values. a. model fitted values of ceftriaxone pk/pd indices and its relationship with log fold change in bacterial density; b. model fitted values of vancomycin pk/pd indices and its relationship with log fold change in bacterial density e pd endpoint i.e. bacterial density calculated as change in log10 cfu/ml; e0 the baseline effect i.e. pd endpoint without antibiotic treatment; pdmax, maximum effect; x pk/pd index; ex50 magnitude of x that is needed to achieve 50% of the pdmax; hill sigmoidicity factor; fdc fixed dose combination; pd pharmacodynamic; pk pharmacokinetic; mic minimum inhibitory concentration, % t>mic % time for which the levels of antibiotic in serum/plasma exceed the mic; auc/mic area under the concentrationtime curve/mic ratio; cmax/mic peak plasma concentration/mic ratio. admet & dmpk 4(3) (2016) 241-260 dose optimization of fdc against s. aureus resistant bacteria doi: 10.5599/admet.4.3.286 259 supplement figure 2: dose response curve of optimized twice daily regimen of fdc the best exposure from dose escalation study fractionated twice daily dosage regimen of fdc enhanced bacterial killing. the twice daily regimen at 75, 50, 38, and 25 % of original concentration and the change in bacterial density in 24 h was plotted to prepare the drc. the drc was plotted between the dose percentage and log reduction in bacterial density and sigmoidal function was used for curve fitting. the model showed good correlation (r 2 =0.9633) and ec50 of 37.46 % was estimated. ec50 percentage of dose required to attain the 50% of pd effect; fdc fixed dose combination; drc dose response curve; pd pharmacodynamics. supplement figure 3: breakpoint analysis of fixed dose combination of ceftriaxone and vancomycin against 150 mrsa strains the in vitro susceptibility tests categorize the bacterial isolates into three categories: susceptible (s), intermediate (i), and resistant (r). the mic breakpoints was determined on the basis of attainment of pk/pd index % t>mic ≈50%t>mic. the plot of mic values of various strains against the zone diameter (ast) for mrsa and segregated based on the mic breakpoints of 4 µg/ml (s) and 32 µg/ml (r). vishnu sharma et al. admet & dmpk 4(3) (2016) 241-260 260 ast antimicrobial susceptibility; mic minimum inhibitory concentration, % t>mic % time for which the levels of antibiotic in serum/plasma exceed the mic ©2016 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ in vitro dissolution/release methods for mucosal delivery systems doi: 10.5599/admet.5.3.425 173 admet & dmpk 5(3) (2017) 173-182; doi: http://dx.doi.org/10.5599/admet.5.3.425 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index review in vitro dissolution/release methods for mucosal delivery systems mario jug* 1 , anita hafner 1 , jasmina lovrić 1 , maja lusina kregar 2 , ivan pepić 1 , željka vanić 1 , biserka cetina-čižmek 2 , jelena filipović-grčić 1 1 university of zagreb, faculty of pharmacy and biochemistry, department of pharmaceutical technology, a. kovačića 1, 10 000 zagreb, croatia 2 pliva croatia ltd, teva group member, prilaz baruna filipovića 25, 10000 zagreb, croatia *corresponding author: e-mail: mjug@pharma.hr, tel.: +385-1-6394-764; fax: +385-1-4612-691 received: august 29, 2017; revised: september 19, 2017; published: september 29, 2017 abstract in vitro dissolution/release tests are an indispensable tool in the drug product development, its quality control and the regulatory approval process. mucosal drug delivery systems are designed to provide both local and systemic drug action following ocular, nasal, oromucosal, vaginal or rectal administration. they exhibit significant differences in formulation design, physicochemical characteristics and drug release properties. therefore it is not possible to devise a single method which would be suitable for release testing of such versatile and complex dosage forms. different apparatuses and techniques for in vitro release testing for mucosal delivery systems considering the specific conditions at the administration site are described. in general, compendial apparatuses and methods should be used as a first approach in method development when applicable. however, to assure adequate simulation of conditions in vivo, novel biorelevant in vitro dissolution/release methods should be developed. equipment set up, the selection of dissolution media and volume, membrane type, agitation speed, temperature, and assay analysis technique need to be carefully defined based on mucosal drug delivery system c haracteristics. all those parameters depend on the delivery system and physiological conditions at the site of application and may vary in a wide range, which will be discussed in details. keywords in vitro dissolution/release test; nasal drug delivery; ocular drug delivery; oromucosal drug delivery; rectal drug delivery; vaginal drug delivery introduction the drug dissolution/release test is a key test of dosage form performance both during formulation development as well as for quality control purposes. for majority of mucosal drug delivery systems (i.e. formulations aimed to provide local and/or systemic drug action via nasal, ocular, oral, rectal and vaginal mucosa) the methods and apparatuses for dissolution/release testing are not standardized and described in pharmacopoeial monographs, indicating the need for further development and systematisation of the currently used methods [1]. http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:mjug@pharma.hr mario jug et al. admet & dmpk 5(3) (2017) 173-182 174 it should be noted that the terms drug dissolution and drug release are often not appropriately distinguished in the literature and even in regulatory documents. the drug dissolution process refers to all formulations in which the drug is initially present in the solid state and encompasses 5 major mass transport steps: (i) the wetting of the particle surface with water; (ii) breakdown of solid state bonds in drug particle; (iii) solvation of the individualized species such as ions or molecules; (iv) their diffusion through the liquid unstirred boundary layer and (v) convection within the well-stirred bulk fluid. if drug dissolution process is the rate-limiting step in the overall release process, than the terms drug dissolution and drug release can be considered as synonyms. in all other cases, drug release is the more appropriate term. in fact, the drug release is a more complex phenomenon, where drug dissolution is just one of its steps. upon the contact with the aqueous dissolution medium, water penetrates into the matrix of the delivery system and dissolves the drug. the dissolved drug species subsequently diffuse out of the matrix of the delivery system due to concentration gradient. additionally, the matrix of the delivery system might also undergo several changes such as swelling and consequent dissolution in the aqueous medium, all contributing to the overall drug release process [2,3]. considering the complexity of mucosal delivery systems, often containing nanoparticulate carriers included in soluble or insoluble matrices, in vitro drug release seems to be the appropriate term when describing the liberation process of the drug from such formulations. in methods listed in usp and ph. eur. monographs, usp dissolution database and fda dissolution database (table 1), the apparatuses traditionally used for oral formulations are also often applied for mucosal delivery systems. however, many noncompendial methods have been developed with intention to take into account the specific physiochemical properties of the formulation and the physiological environment at which the drug should be released. the goal is to develop a test able to provide predictive estimation of in vivo drug product performance, with preferably level a of in vitro-in vivo correlation (ivivc) [4]. development of noncompendial methodologies requires standardization of test parameters and procedures in order to ensure reproducible and reliable results. besides the equipment set up, selection of the dissolution media and volume, membrane type, agitation speed, temperature, and assay analysis technique need to be carefully defined based on mucosal drug delivery system characteristics and physiological conditions at the site of application [5]. this article is aimed to provide useful guidelines and recommendations in this regard. the dissolution medium selection as shown in table 1, apparatuses traditionally used for oral formulations are often applied for mucosal delivery systems. generally, the relatively high volume of the dissolution medium used (500-1000 ml) and the hydrodynamics provided by such apparatuses is not in line with the in vivo conditions at the mucosal administration sites. to mimic the physiological conditions at mucosal administration sites, a low volume of aqueous dissolution medium in the physiological ph range should be used. some of the commonly used dissolution media for mucosal delivery systems are listed in table 2. on the other hand, sink conditions should be maintained. in order to achieve sink conditions for a water-insoluble active compound it may be necessary to add surfactants (such as tween 80 or sodium lauryl sulphate), complexing agents (such as cyclodextrins) or organic solvents (ethanol, methanol etc.) into the dissolution medium [14]. if organic solvents need to be included in dissolution medium, they have to be compatible with the formulation, and their concentration needs to be adjusted not to disrupt the formulation. in addition, in case of membrane diffusion methods, the selected solvent should also be compatible with the membrane used [15]. the most commonly applied temperature for oromucosal, rectal and vaginal formulations testing is 37 °c, while in admet & dmpk 5(3) (2017) 173-182 in vitro release testing of mucosal drug delivery systems doi: 10.5599/admet.5.3.425 175 case of ocular and nasal formulations, temperature may wary between 32-34 and 33-37 °c, respectively [3]. table 1. apparatuses used/recommended for in vitro release testing of mucosal formulations according to monographs and methods published by usp, ph. eur. and fda. reprinted from ref. [3] with permission of elsevier. formulation type apparatus reference semisolid dosage forms (creams, gels, ointments, lotions) vertical diffusion cell [6] immersion cell [6] flow through cell with adapter for semisolid dosage forms [6] paddle apparatus [7] oral, buccal and sublingual films paddle over disk apparatus [7–10] basket apparatus (small-volume configuration) [7,8] sublingual and buccal tablets paddle apparatus (standard or small-volume configuration) [7,8,11] basket apparatus (standard or small-volume configuration) [7,8] lozenges basket apparatus [7,8] paddle apparatus [7,8,11] reciprocating cylinder [7,8] medicated chewing gums dissolution apparatus for chewing gums [7,8,12] suppositories (hydrophilic) paddle apparatus [7,8,11] basket apparatus (standard or palmieri type basket) [7,8] flow-through cell [8] suppositories (lipophilic) dual chamber flow-through cell [8,13] vaginal tablets and vaginal inserts paddle apparatus [7,8] basket apparatus [7,8,11] vaginal rings incubator shaker [7] mucosal suspensions paddle apparatus (standard or small-volume configuration) [7,8] mucosal emulsions paddle apparatus [8] vertical diffusion cell [8] ocular systems reciprocating shaker [11] periodontal systems tube rotator [11] table 2. some frequently used dissolution media for in vitro drug release testing of mucosal drug delivery systems. reprinted from ref. [3] with permission of elsevier. medium composition ph reference simulated tear fluid 2.18 g nahco3, 6.78 g nacl, 0.063 g cacl2, 1.38 g kcl and distilled water up to 1000 ml 7.4 [16] simulated nasal fluid 8.77 g nacl, 2.98g kcl, 0.59 g cacl2 and distilled water up to 1000 ml 5.5 [17] simulated nasal electrolyte solution 7.45 g nacl, 1.29 g kcl, 0.32 cacl2x2h2o and distilled water up to 1000 ml 5.5-6.5 [18] artificial nasal electrolyte 0.8 g nacl, 3.0 g kcl, 0.45 g cacl2 and distilled water up to 1000 ml 6.8 [19] nasal fluid simulant 7.13 g nacl, 2.98 g kcl, 2.14 g nah2po4xh20, 1.21 g na2hpo4x7h2o and distilled water up to 1000 ml 7.4 [20] simulated saliva 2.38 g na2hpo4, 0.19 g kh2po4, 8 g nacl and distilled water up to 1000 ml 6.8 [21] modified simulated saliva 1.63 g kh2po4, 2.32 g nacl, 0.22 g cacl2 and distilled water up to 1000 ml 6.2 [22] vaginal fluid simulant 3.51 g nacl, 1.4 g koh, 0.222 g ca(oh)2, 0.018 g bovine serum albumin, 2 g lactic acid, 1 g acetic acid, 0.16 g glycerol, 0.4 g urea, 5 g glucose and distilled water up to 1000 ml 4.2 [23] mario jug et al. admet & dmpk 5(3) (2017) 173-182 176 selection of the apparatus type and setup basket apparatus basket apparatus (apparatus 1; figure 1) has been applied for the in vitro release testing of different mucosal delivery systems, such as nasal microparticles [24] and inserts [25] using relatively large volume (200-400 ml) of the dissolution medium and moderate stirring rate (50 rpm). similar apparatus, operating with 250 ml of the simulated saliva and 50 or 100 rpm stirring rate were used to investigate the release of tizanidine hydrochloride [26] and buspirone hydrochloride [27] from freeze-dried buccal sponges/wafers. those drugs are classified as class i drugs (high solubility and permeability) according to the biopharmaceutical classification systems (bcs) and in both cases, level a of ivivc was obtained by such instrumental setup [26,27]. basket apparatus has been also applied for the in vitro drug release testing of both solid and in situ gelling rectal suppositories. in latter case, the formulation was first thermostated at 37 °c in suppository moulds to achieve solidification of the formulation prior its placing into basket [28]. another approach, where a liquid suppository formulation was directly introduced into the basket covered with gauze, was also described [29]. compendial monographs, usp dissolution database and fda dissolution database also recommend the use of basket apparatus, both in standard and the small-volume configuration, for the in vitro testing of mucosal delivery systems (table 1). . figure 1. schematic representation of the basket apparatus (apparatus 1). reprinted from ref. [3] with permission of elsevier. paddle apparatus and modifications conventional paddle apparatus (apparatus 2; figure 2a) with 900 ml of the dissolution media stirred at 50 rpm has been occasionally used for the in vitro release testing from buccal tablets [30], sublingual formulations [31], vaginal tablets [32] and self-microemulsifying rectal suppositories [33]. according to the monographs and methods published by usp, ph. eur. and fda (table 1), the paddle apparatus is used for a variety of mucosal drug delivery systems. however, the general intention of the researchers in the field is to reduce the volume of the dissolution medium applied [34,35]. in cases where the volume is significantly reduced to only 10-25 ml, the agitation is performed by the use of shaking incubators. using such apparatus, level a ivivc with correlation coefficient of 0.909 was obtained for carbamazepine release from buccal multi-composite constructs [34]. the same level of ivivc was obtained in the case of vaginal nuvaring®, using similar approach [36]. while analysing the in vitro release of buccal and sublingual films by paddle apparatus, a formulation is commonly attached to a glass slide with cyanoacrylate glue [37] or two sided adhesive tape [38] and then immersed into dissolution vessel or glued onto the inner side of the dissolution vessel [37]. considering the porosity and the thickness of the formulation, the permeation of admet & dmpk 5(3) (2017) 173-182 in vitro release testing of mucosal drug delivery systems doi: 10.5599/admet.5.3.425 177 the glue into the formulation could be presumed with consequent modification of the drug release process. in this regard, the use of metal clamps, supports with wire mesh or compendial apparatus 5 (paddle over disc, figure 2b) is preferred. furthermore, the position of the formulation (bottom or the side of the dissolution vessel) influences greatly on the observed in vitro drug release kinetic, due to differences in hydrodynamics inside the dissolution vessel [37]. (a) (b) (c) figure 2. schematic representation of the paddle apparatus (a), paddle over disc apparatus (b) and immersion cell in combination with the small-volume paddle apparatus (c). reprinted from ref. [3] with permission of elsevier. membrane diffusion methods membrane diffusion (i.e. dialysis) method is considered as one of the most convenient techniques for determining the drug release profiles of nano-sized drug delivery systems like nanoparticles, liposomes, nanosuspensions and emulsions [3,39]. the most commonly used setups are dialysis bag, flow-through apparatus (with a membrane-containing adapter), franz-diffusion cell and immersion cell. the advantages of dialysis over other methods are in the ease of sampling and replacement of receptor media due to the physical separation of the formulation (by semipermeable dialysis membrane) from the receptor media. however, it is important to emphasize that non-critical interpretation of the obtained results can lead to flawed conclusions about the kinetics of drug release from the colloidal carriers [40,41] and other formulation types [42–46]. the presence of the drug in the receptor medium is the result of drug release from the carrier to the continuous phase in the donor compartment and diffusion of the drug through the membrane from the donor to the receptor compartment. each of these two processes can limit the overall drug release rate. if diffusion through the dialysis membrane is the limiting factor, then the results obtained do not say much about the kinetics of the drug release from the carrier. therefore, to account for the resistance of the dialysis membrane to drug diffusion and its influence on the overall drug release rate, an in vitro release study has to be performed with the drug solution too. furthermore, the possibility of reversible binding of the released drug onto a carrier in the donor compartment must be considered, which may reduce the thermodynamic activity of the drug in the donor compartment and thus the diffusion rate. this may lead to the wrong conclusions about the extended release of the drug. finally, it is very important to consider that as drug diffuses through the diffusion membrane, the ions diffuse from one compartment to another as well. this is of the most importance for ion activated in situ gelling systems for which the gelation as well as gel strength is related to the concentration of crosslinking ion such as ca 2+ [3]. dialysis bag with molecular weight cut-off in range from 8 to 14 kda is the most frequently used noncompendial setup among membrane diffusion methods. the volume of receptor media is usually between 20 and 100 ml and stirring conditions are ranging from 50 to 150 rpm. using dialysis method, a good ivivc mario jug et al. admet & dmpk 5(3) (2017) 173-182 178 was found for gatifloxacin release from ocular inserts [47]. level a ivivc was also achieved for paracetamol release from suppository formulation using similar apparatus setup [48]. dialysis bag is also widely used to study the in vitro release properties of semisolid and nanoparticulate formulations aimed for ocular, nasal and vaginal drug delivery [3]. in some cases, the dialysis bag is introduced to the basket of apparatus 1 or attached onto the paddle of the apparatus 2. another setup includes the use of a glass cylinder containing the formulation, sealed at the bottom with the semipermeable membrane and vertically immersed into the dissolution medium. such apparatus is used for in vitro release testing of rectal suppositories [49]. the compendial flow through apparatus (apparatus 4; figure 3) can be used in combination with a membrane containing adapter. this setup is widely employed for in vitro drug release testing from ocular inserts [50] and contact lenses [51]. sometimes is also used in case of semisolid dosage forms [6], microparticles for nasal application [52] and thermosensitive gels for buccal application [53]. this apparatus allows the drug release testing under physiological flow rate of the dissolution medium and offers ability to maintain the sink conditions while operating in an open system setup. besides compendial, some self-made flow through apparatuses were also used [54]. however, the flow through method shows some disadvantages, including instrument cost and set-up, filter clogging, drug adsorption to the apparatus, problems to maintain a constant flow rate and consequent high variability in the results [39]. figure 3. schematic representation of the flow through cell. reprinted from ref. [3] with permission of elsevier. franz-diffusion cell (i.e. vertical diffusion cell, figure 4) was originally developed for in vitro drug release testing from dermal creams, ointments and gels, but nowadays is widely used to test in vitro drug release from various mucosal delivery systems. the formulation loaded into upper donor compartment is separated from the receptor medium present in the lower compartment by a semipermeable membrane. franz-diffusion cell was used to assess the drug release properties from numerous formulations aimed for nasal application, such as thermoresponsive soluble gels [55], ion activated in situ gels [56], microparticles [57], nanostructured lipid carriers [58], gel containing microspheres [59], lipidic emulsomes [60], in situ gelling microemulsions [61] and solid lipid nanoparticles [62]. this method is advantageous over other compendial and non-compendial membrane diffusion methods particularly in case of dry powders as it allows them to hydrate slowly, and gel eventually, in humid environment conditions designed to be similar to those encountered in the nasal cavity [63]. it has been also used in case of nano sized and gelling ophthalmic formulations [64,65] as well as for films [66], wafers [67] and tablets [68] aimed for oromucosal administration. also, in majority of the in vitro drug release studies from nanosystem-in-hydrogel type vaginal formulations franz diffusion-cell method was applied [69]. the restricted volume of acceptor compartment is a drawback of the franz diffusion apparatus, which might impair significantly the observed drug release profile, especially in the case of poorly soluble drugs. admet & dmpk 5(3) (2017) 173-182 in vitro release testing of mucosal drug delivery systems doi: 10.5599/admet.5.3.425 179 figure 4. schematic representation of the franz-diffusion cell. reprinted from ref. [3] with permission of elsevier. the immersion cell (figure 2c), another compendial membrane diffusion method mostly used for semisolid preparations, was successfully applied to study the drug release kinetic from microparticles aimed both for nasal and dermal application [70]. recently, some membrane-less diffusion methods have been reported to study the drug release from ocular inserts [64] and nasal in situ forming gels [20], providing the direct contact of the formulation with the dissolution medium. agarose method is another membrane-less diffusion model which allows to monitor the release of both the released drug and liposomally associated drug in environment simulating the conditions presented at the vaginal site [71]. sample and separate method sample and separate method is simple and provides a direct approach to determine the drug release. dosage form is introduced into the release medium maintained at a constant temperature and the drug release is assessed by sampling of the release medium at defined time, separated by filtration [72] or centrifugation [73] and quantified by a suitable analytical technique [3]. the volume of the receptor medium should be adjusted to maintain sink conditions. this approach is applicable if the drug release lasts much longer (in hours) than the particle separation (in minutes) process and is often used to test in vitro drug release from nasal microand nano-sized delivery systems [72,74]. as the particle size decreases, the difficulties of separation increase. complete separation of nanoparticles requires ultracentrifugation or ultrafiltration. the application of such a strong force for particle separation may impair their integrity and thus affect the drug release profile. therefore, the conditions of separation procedure should be clearly stated in the published results. in (ultra)filtration the adsorption of the released drug to the filter should be considered. also, the particles may clog the filter pores resulting in slow filtration and limited volume of filtered sample. another disadvantage of this method is the loss of drug-loaded particles due to sampling, thereby resulting in an incomplete drug release profile. another important obstacle is aggregation of particles which may decrease the release rate. loss in volume because of filtration during sampling and buffer replacement is a concern when the amount of release media is small [75]. conclusions mucosal drug delivery systems differ significantly in formulation design and their physicochemical and release characteristics. therefore versatile in vitro release methods are currently used for their characterisation, considering the specific conditions at the administration site. compendial methods are used as a first approach in method development whenever applicable. further progress in this field is focused towards the development of novel biorelevant methods, which would be able to predict more mario jug et al. admet & dmpk 5(3) (2017) 173-182 180 closely the in vivo performance of the formulation. acknowledgements: this work was supported by a project “modelling of the pharmaceutical spray drying process of the emulsions in laboratory and pilot scale” in collaboration with the industrial partner pliva croatia ltd. references [1] c.k. brown, h.d. friedel, a.r. barker, l.f. buhse, s. keitel, t.l. cecil, j. kraemer, j.m. morris, c. reppas, m.p. stickelmeyer, c. yomota, v.p. shah, aaps pharmscitech 12 (2011) 782–794. [2] j. siepmann, f. siepmann, international journal of pharmaceutics 453 (2013) 12–24. [3] m. jug, a. hafner, j. lovrić, m. lusina kregar, i. pepić, ž. vanić, b. cetina-čižmek, j. filipović-grčić, journal of pharmaceutical and biomedical analysis (2017) doi: http://dx.doi.org/10.1016/j.jpba.2017.06.072 . [4] j. shen, d.j. burgess, journal of controlled release 219 (2015) 644–651. [5] m. siewert, j. dressman, c.k. brown, v.p. shah, j.-m. aiache, n. aoyagi, d. bashaw, c. brown, w. brown, d. burgess, j. crison, p. deluca, r. djerki, j. dressman, t. foster, k. gjellan, v. gray, a. hussain, t. ingallinera, j. klancke, j. kraemer, h. kristensen, k. kumi, c. leuner, j. limberg, p. loos, l. margulis, p. marroum, h. moeller, b. mueller, m. mueller-zsigmondy, n. okafo, l. ouderkirk, s. parsi, s. qureshi, j. robinson, v. shah, m. siewert, r. uppoor, r. williams, aaps pharmscitech 4 (2003) 43–52. [6] usp <1724> in:, united states pharmacopeia usp 40/nf 35, the united states pharmacopeial convention, rockville, usa, 2017, pp. 2055–2067. [7] fda dissolution methods database, https://www.accessdata.fda.gov/scripts/cder/dissolution/index.cfm. (accessed 17 march 2017). [8] usp <1004> in:, united states pharmacopeia usp 40/nf 35, the united states pharmacopeial convention, rockville, usa, 2017, pp. 829–832. [9] usp <724> in:, united states pharmacopeia usp 40/nf 35, the united states pharmacopeial convention, rockville, usa, 2017, pp. 600–607. [10] ph. eur. 2.9.4. in:, european pharmacopeia 9.0, european directorate for the quality of medicines and healthcare, strasbourg, france, 2017, pp. 309–311. [11] usp dissolution methods database, http://www.usp.org/usp-nf/overview/compendial-tools/uspdissolution-methods-database (accessed 8 march 2017) [12] ph. eur. 2.9.25. in:, eur. pharmacop. 9.0, european directorate for the quality of medicines and healthcare, strasbourg, france, 2017, pp. 340–344. [13] ph. eur. 2.9.42. in:, eur. pharmacop. 9.0, european directorate for the quality of medicines and healthcare, strasbourg, france, 2017, p. 377. [14] a. olejnik, j. goscianska, i. nowak, journal of pharmaceutical sciences 101 (2012) 4032–4045. [15] g.l. flynn, v.p. shah, s.n. tenjarla, m. corbo, d. demagistris, t.g. feldman, t.j. franz, d.r. miran, d.m. pearce, j.a. sequeira, j. swarbrick, j.c.t. wang, a. yacobi, j.l. zatz, pharmaceutical research 16 (1999) 1325–1330. [16] p. zhang, x. liu, w. hu, y. bai, l. zhang, carbohydrate polymers 149 (2016) 224–230. [17] s. maiti, a. chakravorty, m. chowdhury, international journal of biological macromolecules 68 (2014) 241–246. [18] c. karavasili, n. bouropoulos, l. sygellou, e.p. amanatiadou, i.s. vizirianakis, d.g. fatouros, material science and engineering: c 59 (2016) 1053–1062. [19] s. lungare, j. bowen, r. badhan, journal of pharmaceutical sciences 105 (2016) 1209–1220. [20] c. li, c. li, z. liu, q. li, x. yan, y. liu, w. lu, international journal of pharmaceutics 474 (2014) 123– 133. http://dx.doi.org/10.1016/j.jpba.2017.06.072 https://www.accessdata.fda.gov/scripts/cder/dissolution/index.cfm http://www.usp.org/usp-nf/overview/compendial-tools/usp-dissolution-methods-database http://www.usp.org/usp-nf/overview/compendial-tools/usp-dissolution-methods-database admet & dmpk 5(3) (2017) 173-182 in vitro release testing of mucosal drug delivery systems doi: 10.5599/admet.5.3.425 181 [21] a. miro, i. d’angelo, a. nappi, p. la manna, m. biondi, l. mayol, p. musto, r. russo, m.i. la rotonda, f. ungaro, f. quaglia, international journal of pharmaceutics 452 (2013) 283–291. [22] a.b. nair, r. kumria, s. harsha, m. attimarad, b.e. al-dhubiab, i.a. alhaider, journal of controlled release 166 (2013) 10–21. [23] d.h. owen, d.f. katz, contraception 59 (1999) 91–95. [24] e. gavini, g. rassu, l. ferraro, a. generosi, j. v. rau, a. brunetti, p. giunchedi, a. dalpiaz, journal of pharmaceutical sciences 100 (2011) 1488–1502. [25] a.g. sonje, h.s. mahajan, material science and engineering: c 64 (2016) 329–335. [26] g.m. el-mahrouk, o.n. el-gazayerly, a.a. aboelwafa, m.s. taha, international journal of pharmaceutics 467 (2014) 100–112. [27] m.a.a. kassem, a.n. elmeshad, a.r. fares, international journal of pharmaceutics 463 (2014) 68– 80. [28] e. pásztor, á. makó, g. csóka, z. fenyvesi, r. benko, m. prosszer, s. marton, i. antal, i. klebovich, drug development and industrial pharmacy 37 (2011) 1–7. [29] f. din, o. mustapha, d.w. kim, r. rashid, j.h. park, j.y. choi, s.k. ku, c.s. yong, j.o. kim, h.g. choi, european journal of pharmaceutics and biopharmaceutics 94 (2015) 64–72. [30] e. meng-lund, j. jacobsen, a. müllertz, e.b. jørgensen, r. holm, international journal of pharmaceutics 515 (2016) 125–131. [31] s.a. fouad, r.n. shamma, e.b. basalious, m.a. el-nabarawi, s.a. tayel, international journal of pharmaceutics 505 (2016) 212–222. [32] l. wang, x. tang, international journal of pharmaceutics 350 (2008) 181–187. [33] d. gugulothu, s. pathak, s. suryavanshi, s. sharma, v. patravale, aaps pharmscitech 11 (2010) 1179–1184. [34] o.a. adeleke, y.e. choonara, l.c. du toit, v. pillay, journal of pharmaceutical sciences 103 (2014) 1157–1169. [35] m. jug, i. kosalec, f. maestrelli, p. mura, carbohydrate polymers 90 (2012) 1794–1803. [36] p.j. marroum, in: pharmaceutical product development: in vitro-in vivo correlation, d. murthy chilukuri, g. sunkara, d. young (eds.), informa healthcare usa, inc, new york, london, 2007, pp. 177–195. [37] l. perioli, c. pagano, colloids surfaces b biointerfaces 102 (2013) 915–922. [38] j. xu, s. strandman, j.x.x. zhu, j. barralet, m. cerruti, biomaterials 37 (2015) 395–404. [39] d’souza suzan, advances in pharmaceutics 2014 (2014) 1–12. [40] s. modi, b.d. anderson, molecular pharmaceutics 10 (2013) 3076–3089. [41] y. zambito, e. pedreschi, g. di colo, international journal of pharmaceutics 434 (2012) 28–34. [42] h. onishi, k. yumoto, o. sakata, international journal of pharmaceutics 468 (2014) 207–213. [43] r. kumria, a.b. nair, b.e. al-dhubiab, drug development and industrial pharmacy 40 (2014) 625– 631. [44] y. yuan, y. cui, l. zhang, h.p. zhu, y.s. guo, b. zhong, x. hu, l. zhang, x.h. wang, l. chen, international journal of pharmaceutics 430 (2012) 114–119. [45] r. kamel, m. basha, s.h. abd el-alim, journal of liposome research 23 (2013) 28–36. [46] a.e.s.f. abou el ela, a. a. allam, e.h. ibrahim, drug delivery 7544 (2014) 1–10. [47] a.p. gorle, s.g. gattani, chemical and pharmaceutical bulletin 57 (2009) 914–919. [48] d. chicco, i. grabnar, a. škerjanec, d. vojnović, v. maurich, n. realdon, e. ragazzi, a. belič, r. karba, a. mrhar, international journal of pharmaceutics 189 (1999) 147–160. [49] a.s. zidan, s.e. emam, t.m. shehata, f.s. ghazy, aaps pharmscitech 16 (2015) 645–655. [50] y. sultana, m. aqil, a. ali, acta pharmaceutica 55 (2005) 305–314. [51] a. mahomed, j.s. wolffsohn, b.j. tighe, contact lens anterior eye 39 (2016) 97–105. mario jug et al. admet & dmpk 5(3) (2017) 173-182 182 [52] a. dalpiaz, m. fogagnolo, l. ferraro, a. capuzzo, b. pavan, g. rassu, a. salis, p. giunchedi, e. gavini, antiviral research 123 (2015) 146–157. [53] n. zeng, g. dumortier, m. maury, n. mignet, v. boudy, international journal of pharmaceutics 467 (2014) 70–83. [54] p. mura, n. mennini, i. kosalec, s. furlanetto, s. orlandini, m. jug, carbohydrate polymers 121 (2015) 231–240. [55] p. pandey, p.j. cabot, b. wallwork, b.j. panizza, h.s. parekh, european journal of pharmaceutical sciences 96 (2017) 499–507. [56] s.r. salunke, s.b. patil, international journal of biological macromolecules 87 (2016) 41–47. [57] a.d. kulkarni, d.b. bari, s.j. surana, c. v. pardeshi, journal of drug delivery science and technology 31 (2016) 108–117. [58] t.b. devkar, a.r. tekade, k.r. khandelwal, colloids and surfaces b: biointerfaces 122 (2014) 143– 150. [59] s. jose, c.r. ansa, t.a. cinu, a.j. chacko, n.a. aleykutty, s. v. ferreira, e.b. souto, international journal of pharmaceutics 441 (2013) 516–526. [60] g.m. el-zaafarany, m.e. soliman, s. mansour, g.a.s. awad, international journal of pharmaceutics 503 (2016) 127–140. [61] r. pathak, r. prasad dash, m. misra, m. nivsarkar, acta pharmaceutica sinica b 4 (2014) 151–160. [62] b. shah, d. khunt, h. bhatt, m. misra, h. padh, european journal of pharmaceutical sciences 78 (2015) 54–66. [63] a. martinac, j. filipović-grčić, b. perissutti, d. voinovich, ž. pavelić, journal of microencapsulation 22 (2005) 549–561. [64] i.d. rupenthal, c.r. green, r.g. alany, international journal of pharmaceutics 411 (2011) 69–77. [65] a. mahor, s.k. prajapati, a. verma, r. gupta, a.k. iyer, p. kesharwani, journal of colloids and interface sciences 483 (2016) 132–138. [66] r. kapil, s. dhawan, s. beg, b. singh, drug development and industrial pharmacy 39 (2012) 466– 480. [67] j.s. boateng, i. ayensu, drug development and industrial pharmacy 40 (2014) 611–618. [68] m.a. el-nabarawi, a.a. ali, h.m. aboud, a.h. hassan, a.h. godah, drug design, develepment and therapy 10 (2016) 4031–4045. [69] ž. vanić, n. škalko-basnet, european journal of pharmaceutical sciences 50 (2013) 29–41. [70] m. lusina kregar, m. dürrigl, a. rožman, ž. jelčić, b. cetina-čižmek, j. filipović-grčić, international journal of pharmaceutics 485 (2015) 202–214. [71] ž. pavelić, n. škalko-basnet, r. schubert, international journal of pharmaceutics 219 (2001) 139– 149. [72] b. deutel, f. laffleur, t. palmberger, a. saxer, m. thaler, a. bernkop-schnürch, european journal of pharmaceutical sciences 81 (2016) 157–161. [73] m. abdel mouez, n.m. zaki, s. mansour, a.s. geneidi, european journal of pharmaceutical sciences 51 (2014) 59–66. [74] g. shahnaz, a. vetter, j. barthelmes, d. rahmat, f. laffleur, j. iqbal, g. perera, w. schlocker, s. dünnhaput, p. augustijns, a. bernkop-schnürch, international journal of pharmaceutics 428 (2012) 164–170. [75] s.s. d’souza, p.p. deluca, pharmaceutical research 23 (2006) 460–474. ©2017 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ erratum to the paper: in vitro membrane binding and protein binding (iam mb/pb technology) to estimate in vivo distribution: applications in early drug discovery doi: 10.5599/admet.568 267 admet & dmpk 6(3) (2018) 267-268; doi: http://dx.doi.org/10.5599/admet.568 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index erratum erratum to the paper: in vitro membrane binding and protein binding (iam mb/pb technology) to estimate in vivo distribution: applications in early drug discovery klara valko* 1 , simon teague 2 , charles pudgeon 3 1 bio-mimetic chromatography ltd, unit 5b business & technology centre, stevenage, sg1 2dx hertfordshire united kingdom 2 glaxosmithkline, stevenage, united kingdom 3 independent researcher, affiliate of regis technologies inc, morton grove, il 60052 usa *corresponding author: e-mail: klara_valko@bio-mimetic-chromatofraphy.com; tel.: +44-7521-989558 received: july 11, 2018; revised: july 12, 2018; available online: july 15, 2018 the authors of the original paper published with the same title [1] apologize for the error in the plot on figure 5. the correct plot, statistics and explanation in figure 5 on page 25 using the published data in the appendix table a1 and a2 by equation 3 are shown below. “in turn, the volume of distribution of marketed drugs could be modeled by the difference in the membrane and the protein binding of compounds as described by equation 3 and shown in figure 5 for the investigated compounds. there are only 40 compounds that have been included in the training set, the majority of the compounds were not included in the original model. that explains that the statistics are slightly worse (r 2 = 0.76; root mean square error = 0.33 in the original model, while here the r 2 = 0.65 and root mean square error = 0.40). “ it is considered as an excellent prediction of the in vivo behaviour of the compounds using only two measured biomimetic properties, namely the iam membrane binding and the hsa protein binding (see the original paper for details). http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:klara_valko@bio-mimetic-chromatofraphy.com klara valko et al. admet & dmpk 6(3) (2018) 267-268 268 figure 5. the observed in vivo log vdss vs the estimated vdss by the iam mb/pb technology® using equation 3. references [1] k. valko, s. teague, c. pudgeon. in vitro membrane binding and protein binding (iam mb/pb technology) to estimate in vivo distribution: applications in early drug discovery. admet & dmpk 5(1) (2017) 14-38. https://dx.doi.org/10.5599/admet.5.1.373 ©2018 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) https://dx.doi.org/10.5599/admet.5.1.373 http://creativecommons.org/licenses/by/3.0/ in silico adme in drug design – enhancing the impact doi: 10.5599/admet.6.1.470 15 admet & dmpk 6(1) (2018) 15-33; doi: http://dx.doi.org/10.5599/admet.6.1.470 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper in silico adme in drug design – enhancing the impact susanne winiwarter 1, *, ernst ahlberg 1 , edmund watson 1 , ioana oprisiu 1 , mickael mogemark 1 , tobias noeske 1 , nigel greene 2 1 drug safety and metabolism, imed biotech unit, astrazeneca, gothenburg, sweden 2 drug safety and metabolism, imed biotech unit, astrazeneca, boston, us *corresponding author: e-mail: susanne.winiwarter@astrazeneca.net received: november 08, 2017; revised: december 12, 2017; published: march 25, 2018 abstract each year the pharmaceutical industry makes thousands of compounds, many of which do not meet the desired efficacy or pharmacokinetic properties, describing the absorption, distribution, metabolism and excretion (adme) behavior. parameters such as lipophilicity, solubility and metabolic stability can be measured in high throughput in vitro assays. however, a compound needs to be synthesized in order to be tested. in silico models for these endpoints exist, although with varying quality. such models can be used before synthesis and, together with a potency estimation, influence the decision to make a compound. in practice, it appears that often only one or two predicted properties are considered prior to synthesis, usually including a prediction of lipophilicity. while it is important to use all information when deciding which compound to make, it is somewhat challenging to combine multiple predictions unambiguously. this work investigates the possibility of combining in silico adme predictions to define the minimum required potency for a specified human dose with sufficient confidence. using a set of drug discovery compounds, in silico predictions were utilized to compare the relative ranking based on minimum potency calculation with the outcomes from the selection of lead compounds. the approach was also tested on a set of marketed drugs and the influence of the input parameters investigated. keywords drug discovery; pharmacokinetics; in silico predictions; multi-parameter ranking introduction drug design is a multi-parameter optimization process, with pharmacokinetic properties describing absorption, distribution, metabolism and excretion (adme) of a compound being important to consider early on. these parameters describe the pharmacokinetics of a compound and are key to determining the dose required for efficacy. parameters such as lipophilicity, solubility, and metabolic stability can be measured in a high throughput manner in vitro and are, thus, often used as early adme screens. during lead identification and lead optimization phases a substantial number of compounds is synthesized and characterized in such screening assays for selection towards additional, more costly in vitro and in vivo experiments to be able to finally select a single candidate drug. however, a compound needs to be physically available to be subjected to such assays. thus, drug industry generates thousands of compounds per month out of which many do not show desirable adme properties. http://www.pub.iapchem.org/ojs/index.php/admet/index susanne winiwarter et al. admet & dmpk 6(1) (2018) 15-33 16 in silico models for such adme endpoints, on the other hand, have been available for a long time, though with varying quality and usability [1-6]. ideally, such models would be used before synthesis and, together with any potency estimation for the specific case, influence the decision to make a compound. consequent usage of such in silico predictions has the potential to considerably reduce the number of synthesized compounds with inadequate adme properties. in practice, it seems that most often only a few predicted properties are really considered before synthesis, usually including lipophilicity as important parameter. while it is understood that all available knowledge should be used to select which compound to make, it is not easy to define how the outcome of various predictions can be combined unambiguously. lately, the use of multi-parameter optimization and scoring tools has been proposed and shown to be of value [7, 8]. however, the definition of the scoring functions can be difficult and may be a bit arbitrary. as physiologically meaningful scoring function the predicted dose to man (d2m) has been suggested to utilize as early as possible. several reports validated early d2m predictions both from preclinical in vivo data [9] as well as from in vitro data combined with in silico predictions [10]. however, using in silico predictions only was not found to be reliable enough, since both potency and pharmacokinetic properties would need to be accurately predicted. here we concentrate on the pharmacokinetic properties only and suggest to use in silico adme models to predict what potency would be required for a specific compound to enable coverage during the whole dosing interval. by reverting the d2m equation and setting the target dose to, for example, 100 mg once daily, the achievable plasma concentration can be calculated from in silico predicted parameters and the minimum required potency (“threshold pic50”) defined as ranking score for virtual compounds. we show how such a ranking score could be used in the discovery setting on the example of a set of 27 compounds leading towards a candidate structure [11]. we also test the concept on a set of known drugs with information about dosing and human pharmacokinetics [9] and compare the results from purely in silico predictions to those from in vitro data and from human in vivo data. we also investigate how the in silico derived parameters influence the outcome of the equations, highlighting concepts to help drug discovery to improve pharmacokinetic properties during design. experimental inverse dose-to-man (d2m) prediction equations used are standard pharmacokinetic equations [9, 12, 13] based on a one compartment model considering immediate absorption to estimate the average (cave, eq. 1), minimum (cmin, eq. 2) and maximum (cmax, eq. 3) concentration of a drug at steady state. (1) (2) (3) where f is bioavailability (eq. 4), dpo is the daily oral dose, cl is the plasma clearance estimated using the well-stirred model (eq. 5),  is the dosing interval, typically set to 24 h, v is the volume of distribution (predicted) and ke is the elimination constant (eq. 6). 𝐹 = 𝐹abs ∙ 𝐹g ∙ 𝐹h , (4) 𝐶ave = 𝐹∙𝐷po∙ 𝜏 24⁄ 𝐶𝑙∙𝜏 𝐶min = 𝐹∙𝐷po∙ 𝜏 24⁄ 𝑉∙(𝑒𝑘𝑒∙𝜏−1) 𝐶max = 𝐹∙𝐷po∙ 𝜏 24⁄ 𝑉∙(1−𝑒−𝑘𝑒∙𝜏) admet & dmpk 6(1) (2018) 15-33 in silico adme in drug design doi: 10.5599/admet.6.1.470 17 where fabs (fraction absorbed) is estimated from permeability (eq. 7), fg (fraction escaping gut metabolism) is set to 1 and fh (fraction escaping hepatic metabolism) is estimated from cl (eq. 8). (5) with clb being blood clearance, qh the liver blood flow (20 ml/min/kg), clintin vivo the in vivo intrinsic clearance estimated from in vitro clint (eq. 9), and fub the fraction unbound in blood. cl = clb ∙𝑅b , (5a) where rb is the blood plasma ratio and set to 1 for neutrals and bases, whereas set to 0.55 for acids and zwitterions. 𝑓ub = 𝑓u/𝑅b , (5b) where fu is fraction unbound in plasma (predicted). 𝑘e = cl/𝑉 (6) fabs = 1, if papp > 1 x 10 -6 cm/s, otherwise fabs = papp , (7) where papp is apparent permeability measured in caco2 cells, utilising the intrinsic caco2 permeability assay [14] (predicted). 𝐹ℎ = 1−𝐶𝑙𝑏/𝑄ℎ (8) 𝐶𝐿𝑖𝑛𝑡𝑖𝑛 𝑣𝑖𝑣𝑜 = 𝐶𝐹 ∙𝐶𝐿𝑖𝑛𝑡𝑠𝑐, (9) where cf is an empirical correction factor adopted for the simplified regression offset approach (cf=3, see below), and clintsc is the scaled intrinsic clearance from the human hepatocyte incubation (eq. 9a). 𝐶𝐿𝑖𝑛𝑡𝑠𝑐 = 𝐶𝐿𝑖𝑛𝑡𝐻𝐻 ∙ 𝑆𝐹𝐻𝐻/𝑓𝑢𝑖𝑛𝑐, (9a) where clinthh is the intrinsic clearance in human hepatocytes (predicted), sfhh are the scaling factors (120 million cells per g liver x 1680 g liver weight / 70 kg body weight = 2.88) and fuinc is the fraction unbound in the hepatocyte incubation (predicted). the minimum potency value, e.g., “threshold pic50” based on cmin, is then calculated as the negative logarithm of the predicted concentration considering the specific fold coverage, typically set to 1 or 3 (eq. 10). threshold pic50 = −log(𝐶min fold coverage⁄ ) (10) simplified regression offset approach – correction factor (cf) the regression offset approach [15] was shown to improve in vitro to in vivo scaling of clearance, especially when experiments where compared from different laboratories (sites) and using different batches of hepatocytes or microsomes. extensive in-house data analysis (data not shown) indicated that overall the regression slope can be set to 1 and the offset to about 0.5, giving an equivalent level of correct predictions in three species, human, rat and dog. thus, in vivo clint can be estimated by multiplying the scaled clint with a correction factor of 3. here we adopt this procedure to estimate in vivo clint directly from in silico predictions. in silico models five in silico models were used as input for the above calculations. human volume of distribution at 𝐶𝑙𝑏 = 𝑄ℎ∙𝐶𝐿𝑖𝑛𝑡𝑖𝑛 𝑣𝑖𝑣𝑜 ∙𝑓𝑢𝑏 𝑄ℎ+𝐶𝐿𝑖𝑛𝑡𝑖𝑛 𝑣𝑖𝑣𝑜 ∙𝑓𝑢𝑏 susanne winiwarter et al. admet & dmpk 6(1) (2018) 15-33 18 steady state, intrinsic caco2 permeability, human plasma protein binding, human hepatocyte intrinsic clearance, and fraction unbound in the hepatocyte incubation. all but the volume of distribution model used astrazeneca in-house experimental data. all models are available within astrazeneca. 1. human volume of distribution (v) human volume of distribution data for about 700 compounds were collated from literature [16-18] and randomly split into a training and test set (n = 544 and 144, respectively). both acids, bases, neutral and zwitterionic compounds are included in both datasets. it is a random forest model [19] using physicochemical descriptors [20] including acd log p and log d [21] and clog p [22]. the experimental data is modelled in its logarithmic form. the prediction explains ~70 % of the variance of the test set, and has an error in prediction of 0.4 log units. 2. intrinsic caco2 permeability (papp) the model is built on in-house data of intrinsic caco2 permeability, apparent permeability measured in caco2 cells in presence of a defined transport inhibitor cocktail as described earlier [14], and updated about every sixth month. the present training set consists of >4,000 data points. about 100 simple physicochemical descriptors [20] including acd log p and log d [21] and clog p [22] are calculated from the compound structures. the data is modelled as log papp using random forest regression [19] as implemented in scikit-learn [23]. the latest temporal test set comprises about 300 compounds and shows an r 2 of 0.4 and a root mean squared error of predictions (rmsep) of 0.6 (log scale), with about 60 % of the compounds being within 3-fold of the experimental value. the model can be used to distinguish between high and low permeability compounds with classification accuracy above 0.8. 3. human plasma protein binding as fraction unbound (fu) the model is built on in-house data of human plasma protein binding generated using equilibrium dialysis in high throughput assays as described earlier [24-26]. it is updated monthly. the present training set consists of almost 90,000 compounds. the data is modelled as log k (= log[(fraction unbound)/(fraction bound)]). the modelling procedure utilizes support vector machines [27] with a linear kernel, signature descriptors [28] and the conformal prediction framework [29] as implemented in cpsign from genettasoft [30]. the latest temporal test set comprising 750 compounds shows an r 2 of 0.7 and an rmsep of 0.4, with about 80 % of the compounds being within 3-fold of the experimental value of fu. 4. human hepatocyte intrinsic clearance (clinthh) the model is built on in-house data of human hepatocyte intrinsic clearance generated in high throughput assays using incubations of either cryopreserved or fresh human hepatocytes at 37 °c for up to 120 min as described earlier [26, 31, 32]. it is updated about every sixth month. the present training set contains more than 11,000 compounds and the data is modelled as log clint. the modelling procedure utilizes support vector machines [27] with a linear kernel, signature descriptors [28] and the conformal prediction framework [29] as implemented in cpsign from genettasoft [30]. the latest temporal test set comprising almost 200 compounds shows an r 2 of 0.2 and an rmsep of 0.4, with ~75 % of the compounds being within 3-fold of the experimental clint value. 5. fraction unbound in the hepatocyte incubation (fuinc) the model is built on in-house binding data measured in cryopreserved rat hepatocyte incubations as described earlier [33]. the model is updated approximately every sixth month and the present data set contains about 1,700 compounds. the data is modelled as log k (= log[(fraction unbound)/(fraction admet & dmpk 6(1) (2018) 15-33 in silico adme in drug design doi: 10.5599/admet.6.1.470 19 bound)]). the modelling procedure utilizes support vector machines [27] with a radial basis function kernel, signature descriptors [28] and the conformal prediction framework [29] as implemented in cpsign from genettasoft [30]. the latest temporal test set comprising about 200 compounds shows an r 2 of 0.5 and an rmsep of 0.5, with ~75 % of the compounds being within 3-fold of the experimental fuinc value. global sensitivity analysis global sensitivity analysis was performed by a quasi-monte carlo method using the fourier amplitude sensitivity test (fast) and sobol’ sensitivity, which are implemented in the global sensitivity analysis toolbox (gsat) [34] in matlab [35] using 20,000 sample points. local sensitivity analysis local sensitivity analysis was performed by taking jacobian matrices of the model with respect to the model parameters and other parameter points [36]. these matrices contain actual parameter values which can be examined, or plotted graphically against other parameters, to show which have the greatest effect at each point. 𝑆(𝑞,𝑝) = ( 𝜕𝑥(𝑞1,𝑝) 𝜕𝑝1 ⋯ 𝜕𝑥(𝑞,𝑝) 𝜕𝑝𝑛 ⋮ ⋱ ⋮ 𝜕𝑥(𝑞𝑚,𝑝) 𝜕𝑝1 ⋯ 𝜕𝑥(𝑞𝑚,𝑝) 𝜕𝑝𝑛 ) , (11) where 𝑞 = (𝑞1,…,𝑞𝑚) is a vector of parameter points to be examined, 𝑝 = (𝑝1,…,𝑝𝑛) is a vector of other parameters which are being examined, and x is a scalar output from the model. this was performed in matlab [35] using automatic differentiation of the parameters via the myad toolkit [37]. data sets two data sets were used for the present analysis. data set 1 was a set of 27 drug discovery compounds exemplifying important compounds leading towards candidate selection [11]. data set 2 was a set of 21 marketed drugs, for which potency data, dosing information and human pharmacokinetic data had been collated and used to show feasibility of early d2m predictions [9]. susanne winiwarter et al. admet & dmpk 6(1) (2018) 15-33 20 results and discussion calculation of required minimum potency (threshold pic50) for a set of drug discovery compounds as proof of concept a set of drug discovery molecules leading towards a candidate structure [11] was evaluated. intrinsic clearance in human hepatocytes, fraction unbound in the incubation, fraction unbound in plasma, volume of distribution and intrinsic caco2 permeability were predicted for all compounds using the present in-house models (see table 1) and minimum and maximum plasma concentrations for a once daily oral dose of 100 mg calculated using equations 2 and 3. the project aimed at minimum plasma concentrations with 3-fold coverage over potency measured as pic50 in a whole blood assay. thus, threshold pic50 values were calculated from cmin,total/3 (see table 2). blood plasma ratios for these mainly neutral compounds were considered to be 1. protein binding was considered in the scaling approach, but threshold pic50s were based on resulting total plasma concentration to match the potency measurement. the data indicated that some of the compounds had very short half-life and thus a once daily dose resulted in a high threshold pic50 to cover potency over 24 hours and a high cmax/cmin ratio. using a 10 times higher dose would reduce the threshold pic50 by one unit, but show the same cmax/cmin ratio, since the applied model assumes linear kinetics. considering only compounds with a threshold pic50 below 9 and a cmax/cmin ratio below 100 about 12 compounds remained. this set of compounds included compound 15b, the compound finally selected, and three of the remaining four compounds of higher interest for which rat in vivo results were reported. compound 22, which according to the present analysis was least favourable, was actually found in the rat study to have short half-life and, thus, not suitable to progress. in summary, it seems that for this data set the threshold pic50 together with the predicted cmax/cmin ratio could be used for ranking the compounds and correctly identify those with inferior pharmacokinetics. the results also highlight the danger in deprioritizing compounds just by applying a cut-off for, for example, metabolic stability. here, three compounds had (predicted) clint values higher than 20 µl/min/10 6 cells, compounds 12, 14a and 19. for 12, a reference compound, threshold pic50 was one of the lowest in the whole set based on a predicted blood clearance below 5 ml/min/kg, whereas 14a showed an intermediate threshold pic50 (8.2) and blood clearance just above 5 ml/min/kg. only for compound 19 the threshold pic50 was estimated to be on the high end within the present set, even though the blood clearance was again predicted as just above 5 ml/min/kg. 22, the compound with the highest threshold pic50 has almost double the blood clearance despite a lower hepatocyte clint (~ 8 µl/min/10 6 cells). rank order differences between in vitro (or in silico) clint and predicted in vivo clearance can be explained by the different binding properties: high binding in the incubation, i.e., low fuinc values, will potentiate the metabolic instability measured in the incubation, whereas high binding in plasma, i.e., low fu, can to some extent mitigate low metabolic stability seen in vitro. furthermore, minimum concentrations mainly depend on a compound’s half-life, determined by clearance and volume of distribution. thus, volume of distribution is another important factor that again may change the rank order. admet & dmpk 6(1) (2018) 15-33 in silico adme in drug design doi: 10.5599/admet.6.1.470 21 table 1. predicted properties for compound set 1 [11] using astrazeneca in-house in silico adme models cmpd id a clinthh b (l/min/10 6 cells) fuinc c %free fu d %free papp e (10 -6 cm/sec) v f (l/kg) mw g (g/mol) ion class 12 20.2 37.7 1.1 20 2.0 447.6 neutral 13a 15.0 38.1 2.5 5 2.1 433.6 neutral 13b 6.5 25.4 3.4 9 2.0 433.6 neutral 14a 27.3 80.2 2.4 26 1.7 447.5 neutral 14b 7.5 72.4 15.3 5 2.0 446.5 base 14c 9.5 73.1 10.5 36 1.3 431.5 neutral 14d 10.3 83.5 9.4 10 0.9 444.5 neutral 14e 6.8 74.9 4.3 27 1.5 447.5 neutral 15a 10.3 70.6 7.5 2 1.2 433.5 neutral 15b 3.7 56.9 10.2 7 1.4 433.5 neutral 15c 6.7 55.7 5.3 12 1.5 432.5 neutral 15d 3.6 47.4 10.7 7 1.4 433.5 neutral 15e 6.3 55.9 4.2 11 1.4 447.5 neutral 15f 6.1 47.6 5.0 7 1.4 431.5 neutral 19 32.3 70.9 1.8 10 0.8 416.4 neutral 20 8.1 76.1 4.3 12 1.0 458.5 neutral 21 12.9 76.0 2.5 13 1.2 415.5 neutral 22 7.9 39.2 10.3 6 0.9 386.4 neutral 29 4.7 49.8 8.0 9 1.5 447.5 neutral 30 5.8 47.5 6.7 9 1.6 447.5 neutral 31 4.1 47.0 9.6 10 1.5 447.5 neutral 32 5.4 52.2 11.5 9 1.5 445.5 neutral 37 2.6 40.6 8.8 8 1.3 469.5 neutral 41 3.7 56.9 10.2 7 1.4 433.5 neutral 42 4.6 46.6 7.7 8 1.7 461.6 neutral 47 2.3 67.1 24.2 7 1.3 433.5 neutral 49 4.0 49.0 16.9 8 1.4 421.5 neutral a cmpd id: compound identification as originally defined [11]; b clinthh: intrinsic clearance in human hepatocytes; c fuinc: fraction unbound in rat hepatocyte incubation; d fu: fraction unbound in human blood plasma; e papp: apparent permeability in caco2 cells with transport inhibition; f v: human volume of distribution at steady state; g mw: molecular weight susanne winiwarter et al. admet & dmpk 6(1) (2018) 15-33 22 table 2. predicted pk properties for compound set 1 [11] based on in silico adme model predictions, plasma concentrations calculated for a once daily dose of 100 mg (1.4 mg/kg), sorted by threshold pic50 cmpd id a cl blood b (ml/min/kg) cave c (m) cmin d (m) cmax e (m) threshold pic50 from cmin f cmax/cmin g 14e 2.88 0.66 0.12904 1.90 7.4 14.7 12 4.17 0.42 0.06645 1.32 7.7 19.9 15e 3.41 0.54 0.06363 1.90 7.7 29.8 21 3.13 0.64 0.06105 2.44 7.7 39.9 15c 4.33 0.42 0.02902 1.73 8.0 59.5 20 3.30 0.55 0.02517 2.56 8.1 101.6 42 4.92 0.33 0.02259 1.38 8.1 61.0 13b 5.45 0.31 0.02254 1.25 8.1 55.4 37 3.96 0.43 0.02136 1.96 8.1 91.5 15b 4.47 0.40 0.01900 1.84 8.2 96.8 41 4.47 0.40 0.01900 1.84 8.2 96.8 15f 4.37 0.41 0.01864 1.93 8.2 103.4 14a 5.22 0.31 0.01757 1.39 8.2 79.2 13a 6.03 0.26 0.01675 1.13 8.3 67.7 29 4.88 0.34 0.01431 1.64 8.3 114.7 30 5.27 0.31 0.01154 1.53 8.4 132.3 31 5.36 0.30 0.00998 1.54 8.5 154.0 15d 5.26 0.32 0.00709 1.78 8.6 250.9 47 5.34 0.31 0.00599 1.80 8.7 300.1 14b 8.10 0.16 0.00259 0.97 9.1 374.3 32 6.83 0.21 0.00171 1.45 9.2 846.0 15a 6.41 0.24 0.00074 1.91 9.6 2573.8 49 7.51 0.20 0.00060 1.53 9.7 2537.8 14c 7.44 0.19 0.00046 1.58 9.8 3398.1 19 5.18 0.34 0.00045 3.01 9.8 6751.1 14d 6.65 0.22 0.00004 2.46 10.9 58276.0 22 9.46 0.14 0.00000 2.20 12.8 4832468.9 a cmpd id: compound identification as defined in ref [11]; b cl blood: predicted human blood clearance; c cave: predicted total average plasma concentration at steady state; d cmin: predicted total minimum plasma concentration at steady state; e cmax: predicted total maximum plasma concentration at steady state; f threshold pic50 from cmin: minimum required pic50 estimated from cmin/3 for a 3fold coverage over 24 hours; g cmax/cmin: predicted ratio between maximum and minimum concentration at steady state admet & dmpk 6(1) (2018) 15-33 in silico adme in drug design doi: 10.5599/admet.6.1.470 23 calculation of threshold potency for a set of known drugs as second test, we investigated the approach for a set of marketed drugs where information on dosing and human pharmacokinetic data had been collated by mcginnity et al. [9]. this dataset was previously used to evaluate the validity of early dose to man predictions [9, 10]. table 3 shows in silico adme parameters for the 21 drugs predicted with the present models at astrazeneca. table 3. predicted properties for 21 known drugs [9] using astrazeneca in-house in silico adme models drug name clinthh a (l/min/10 6 cells) fuinc b %free fu c %free papp d (10 -6 cm/sec) v e (l/kg) mw f (g/mol) ion class acebutolol 7.7 79.0 80.8 2.5 1.5 336.4 base atenolol 1.6 89.7 92.4 0.3 1.5 266.3 base betaxolol 10.0 82.1 51.0 17.6 3.8 307.4 base bisoprolol 9.5 82.1 63.8 12.6 2.3 325.4 base carvedilol 25.0 19.5 0.6 4.8 1.9 406.5 base cimetidine 5.2 77.7 81.2 4.8 1.0 252.3 base cyclosporin 738.4 97.4 0.5 1.4 2.0 1203 neutral desloratidine 9.2 11.9 7.8 19.4 7.8 310.8 base diazepam 6.5 14.6 3.7 55.3 1.1 284.7 neutral diclofenac 55.2 52.8 1.8 33.8 0.3 296.2 acid famotidine 2.5 77.6 78.1 0.7 0.9 337.5 base indomethacin 7.5 39.7 0.6 74.1 0.2 357.8 acid irbesartan 15.1 17.7 0.8 15.6 0.9 428.5 acid metoprolol 4.9 87.3 73.1 13.4 2.4 267.4 base nadolol 4.4 79.8 76.9 0.4 2.1 309.4 base nifedipine 14.3 30.7 3.5 45.0 1.2 346.3 neutral pirenzepine 8.5 86.8 50.7 0.5 1.7 351.4 base prazosin 6.5 58.8 9.9 10.1 0.8 383.4 neutral propranolol 9.5 59.5 18.8 31.0 3.0 259.3 base ranitidine 5.6 69.4 80.2 2.2 1.2 314.4 base ritonavir 75.3 19.1 1.1 4.0 0.9 721 neutral a clinthh: intrinsic clearance in human hepatocytes; b fuinc: fraction unbound in rat hepatocyte incubation; c fu: fraction unbound in human blood plasma; d papp: apparent permeability in caco2 cells with transport inhibition; e v: human volume of distribution at steady state; f mw: molecular weight. to take into account that these drugs are not necessarily given once daily only, we adopted the dosing regimen suggestion by mcginnity et al. [9] when calculating the minimum required potency, threshold px (see table 4). the resulting threshold px values are below 10 for all but four of the compounds, thus indicating that the approach is able to estimate pharmacokinetic behaviour to some extent. additionally, the cmax/cmin ratio was estimated as below 100 for most of the compounds, even though it showed values above 1000 for three of them (ritonavir, bisoprolol and diclofenac). the availability of both in vitro [10] and in vivo [9] data for the compounds prompted us to investigate, to what extent the usage of experimental data would change the results (see table 5 and figure 1). susanne winiwarter et al. admet & dmpk 6(1) (2018) 15-33 24 table 4. dosing interval, predicted plasma concentrations and threshold px based on a daily dose of 100 mg (1.4 mg/kg), given as 1-6 doses per day, for 21 known drugs [9] drug name dosing interval a (h) cmin,total b (m) cmin,free c (m) cmax,total d (m) threshold px from cmin,free e cmax/cmin f acebutolol 8 0.00165 0.00134 0.21153 8.9 127.9 atenolol 12 0.00742 0.00686 0.35626 8.2 48.0 betaxolol 12 0.01102 0.00563 0.17910 8.2 16.3 bisoprolol 24 0.00004 0.00002 0.44702 10.6 11795.4 carvedilol 8 0.19649 0.00114 0.65630 8.9 3.3 cimetidine 4 0.01081 0.00878 0.28422 8.1 26.3 cyclosporin 12 0.00155 0.00001 0.12384 11.2 79.9 desloratidine 24 0.01218 0.00095 0.17440 9.0 14.3 diazepam 12 0.00572 0.00021 1.35487 9.7 236.9 diclofenac 4 0.00002 0.00000 1.26227 12.4 55501.6 famotidine 6 0.00631 0.00493 0.42939 8.3 68.1 indomethacin 6 0.16284 0.00095 5.90360 9.0 36.3 irbesartan 8 0.01792 0.00015 0.82347 9.8 46.0 metoprolol 8 0.02235 0.01635 0.29316 7.8 13.1 nadolol 8 0.00547 0.00421 0.10158 8.4 18.6 nifedipine 6 0.04706 0.00164 0.54780 8.8 11.6 pirenzepine 12 0.00055 0.00028 0.18872 9.6 345.6 prazosin 6 0.04517 0.00448 0.84533 8.3 18.7 propranolol 8 0.05111 0.00961 0.32040 8.0 6.3 ranitidine 6 0.00271 0.00218 0.26131 8.7 96.3 ritonavir 8 0.00021 0.00000 0.26073 11.6 1239.4 a dosing interval as suggested by mcginnity et al. [9], based on their predicted t1/2: t1/2>8h  24h; t1/2 4-8 h  12h; t1/2 2-4h  8h; t1/2 1-2 h  6h; t1/2 < 1h  4h; b cmin,total: predicted minimum total concentration at steady state; c cmin,free: predicted minimum concentration at steady state corrected for plasma protein binding; d cmax,total: predicted maximum total concentration at steady state; e threshold px from cmin,free: required minimum px estimated from free cmin for coverage over dosing interval; f cmax/cmin: predicted ratio between maximum and minimum total concentration at steady state. admet & dmpk 6(1) (2018) 15-33 in silico adme in drug design doi: 10.5599/admet.6.1.470 25 table 5. threshold px based on a daily dose of 100 mg (1.4 mg/kg), given as 1-6 doses per day, for 21 known drugs[9], derived from in silico data (as above), in vitro data [9, 10, 14], and actual human in vivo data [9] drug name threshold px from in silico a threshold px from in vitro b threshold px from in vivo c acebutolol 8.9 6.9 d 7.8 atenolol 8.2 7.8 e,f 6.5 betaxolol 8.2 7.6 e 6.4 bisoprolol 10.6 7.6 e 6.6 carvedilol 8.9 9.7 d 9.3 cimetidine 8.1 6.8 d,f 7.1 cyclosporin 11.2 nv g 8.9 desloratidine 9.0 7.5 d,h nv g diazepam 9.7 6.9 d 6.8 diclofenac 12.4 8.9 i,j 16.0 famotidine 8.3 7.8 d,f 7.4 indomethacin 9.0 9.3 e,f 9.5 irbesartan 9.8 8.8 e 8.8 metoprolol 7.8 6.6 e 7.9 nadolol 8.4 7.4 d,f 6.7 nifedipine 8.8 9.0 d 8.8 pirenzepine 9.6 8.0 e 7.2 prazosin 8.3 7.4 d 7.8 propranolol 8.0 7.7 e 8.4 ranitidine 8.7 7.3 e 7.7 ritonavir 11.6 8.6 k 8.6 a threshold px calculated from in silico values as in table 4; b threshold px calculated from in vitro values using the same procedure as in table 4 (clinthh and fu values from ref [10], papp values from [9] and [14] as specified, and in-house values for fuinc; c threshold px calculated from in vivo values [9] using the same procedure as in table 4; d papp value from ref [9]; e papp value from ref [14]; f fuinc value from in-house in silico model; g no value (not enough experimental data available for calculation); h in-house value for clinthh; i fu value from ref [17]; j papp from in-house in silico model; k in-house value for papp. most threshold px values can be found within the range of 7-9, with a couple of values, especially in vivo derived, below 7 and others, most often in silico derived, above 9. only one of the compounds for which human in vivo pk data was available showed an in vivo derived threshold px values above 10, whereas all in vitro derived threshold px values were actually below 10. the one compound with a higher in vivo data derived value, diclofenac, was earlier recognized as not being properly described by the one compartment pk model employed here [9]. for about a third of the compounds all three values are very close within about one log unit, e.g., carvedilol or nitrendipine. other compounds have a somewhat higher spread, most often the in vitro value closer to the in vivo derived value, e.g., acebutolol or betaxolol. for a few compounds the in silico derived value is clearly differing from the other two, e.g., bisoprolol, diazepam and ritonavir. to better understand why we find the in silico predictions differ, we investigated how well the in silico and in vitro data can predict human in vivo properties (see figures 2-4). note that cyclosporine and desloratidine were excluded from this analysis, since there was not enough in vitro data for the former and no in vivo data for the latter compound. susanne winiwarter et al. admet & dmpk 6(1) (2018) 15-33 26 figure 1. comparison of threshold px values calculated from purely in silico (open circles), in vitro (black circles) or in vivo (black triangles) data as described in text. figure 2 shows that in silico data tends to overpredict human clearance. the two compounds on the left, diazepam and ritonavir, are obvious extremes, and their higher estimate of the threshold potency value are most likely related to the clearance misprediction. note that clearance predictions from in vitro are clearly closer to the line of unity. however, we find two compounds, diclofenac and metoprolol, to be under predicted by more than 3-fold using the present scaling approach. this can also be seen in a more optimistic threshold px value when compared to in vivo (see figure 1). overall, about two thirds of the compounds have in silico predicted clearance values within 3-fold from the experimental in vivo values, and all but two compounds are within 3-fold utilizing in vitro values. figure 2. human in vivo plasma clearance plotted against predicted clearance based purely on in silico property calculations (open circles) and on in vitro data for human hepatocyte clint, human protein binding and hepatocyte incubation binding (black circles). full line indicates unity, dashed lines 3-fold limits. admet & dmpk 6(1) (2018) 15-33 in silico adme in drug design doi: 10.5599/admet.6.1.470 27 figure 3. human in vivo plasma half-life plotted against predicted half-life based upon in silico volume of distribution and clearance calculated from in silico property calculations (open circles) or clearance calculated from in vitro data (black circles). full line indicates unity, dashed lines 3-fold limits. also the underprediction of half-life from in silico data is mostly due to the clearance prediction, while half-life prediction from in vitro, using the same in silico derived value for distribution volume, is clearly enhanced. human fraction absorbed seems to be the most difficult parameter to predict correctly (see figure 4). while both in silico and in vitro predictions identify the two compounds with fabs below 0.4, compounds with fabs between 0.4 and 0.6 were only recognized in two or three cases from in silico and in vitro data, respectively. however, this failure is not necessarily a concern in the present case, since the relationship between fabs and plasma concentrations is assumed to be linear and a 2-fold change in fraction absorbed will lead to a 2-fold change in plasma concentration and thereby only to a 0.3 log units change for the threshold px. actually, page [10] suggested, that estimation of fabs as either 1 (acids) or 0.5 (all other ion classes) is sufficiently accurate for early d2m predictions. figure 4. human fraction absorbed plotted against predicted fraction absorbed from in silico caco permeability prediction (black circles) or in vitro caco permeability data (grey circles). full line indicates unity, dashed lines 3-fold limits. susanne winiwarter et al. admet & dmpk 6(1) (2018) 15-33 28 influence of basic parameters – intrinsic clearance, fraction unbound in the hepatocyte incubation, plasma protein binding and volume of distribution using pharmacokinetic equations assuming a one compartment model both dose and fraction absorbed are linearly related to the derived plasma concentrations and it is easily understood how either of these parameters will influence the threshold pic50 calculation. the remaining four input parameters, clinthh, fuinc, fu, and v, on the other hand are more intricately interlinked, especially since the first three also are included in the scaling approach converting in vitro (in silico) clint to in vivo clearance using the well-stirred liver model (equations 5 and 9). in order to better understand how these parameters influence the outcome, threshold pic50 values for hypothetical compounds with hepatocyte clint values varying from 1 to 200 l/min/10 6 cells, values for fraction unbound in the incubation varying from 0.01 to 0.7, values for protein binding (fu) varying from 0.001 to 0.3, and volume of distribution varying from 0.2 to 3 l/kg were calculated (see figure 5). as expected, lower intrinsic clearance as measured in hepatocytes leads to a lower threshold pic50 estimate. however, the sensitivity of this correlation varies both with the fuinc/fub ratio and the volume of distribution: lower volume of distribution leads to higher sensitivity towards intrinsic clearance. additionally, higher fuinc/fub ratios lead to lower sensitivity of the intrinsic clearance as well as of volume of distribution (see figure 5, panel in upper right corner). lower v leads to higher threshold pic50 values and higher fuinc/fub ratios essentially lead to lower threshold pic50s, at least as long as potency in blood assays, i.e., defined from total concentrations, is considered. the threshold pic50 shifts up by three log units in the uppermost panel row (fu = 0.001) when free plasma concentrations are taken into account, whereas the threshold pic50 in the lower panel row (fu = 0.3) shifts only by about 0.5 log units. sensitivity analysis sensitivity analysis is a useful tool for determining which parameters are important. global sensitivity analysis is used to explore the entire parameter space, considering physiologically relevant values to determine the relative importance of each of them. here it was shown that volume of distribution is the most influential parameter by far (see figure 6). local sensitivity analysis, on the other hand, can define how sensitive the calculation is to a parameter, when another parameter is scanned over its defined range. here, we checked the influence of the remaining parameters for different values of volume of distribution (see figure 7). clinthh and fu have overlapping influence, considering a 10 % change, whereas fuinc is exacty opposite. since in vivo clint is multiplied by fu, when calculating clb using the well stirred model (eq. 5) and is calculated from clinthh divided by fuinc (eqs 9 and 9a) this is not really surprising. note that the influence of the parameters is highest at low volume of distribution values, decreases quickly and does not change further as soon as v reaches a value of 2 l/kg. considering typical volume ranges for acids, bases and neutrals with median values of 0.2 l/kg, 2.9 l/kg and 1.3 l/kg, respectively [10, 17], as indicated in figure 7, it is clear that especially for acids a small change in either hepatocyte clint or binding properties can make a big difference for the prediction of threshold pic50 determined from cmin. admet & dmpk 6(1) (2018) 15-33 in silico adme in drug design doi: 10.5599/admet.6.1.470 29 figure 5. threshold pic50 estimated from minimum plasma concentration for hypothetical compounds as calculated from human hepatocyte clint values (1-200 µl/min/10 6 cells, x-axis), fraction unbound in the incubation (fuinc = 0.01 – 0.7; panel columns), fraction unbound in blood (fub = 0.001 – 0.3; panel rows), and volume of distribution (v = 3 l/kg, black, v = 2 l/kg, dark grey, v = 1 l/kg, medium grey, v = 0.4 l/kg, light grey, v = 0.2 l/kg, very light grey). values in the corner of each panel indicate the fuinc/fub ratio. lower left panels left empty since compounds are usually less bound in the hepatocyte incubation than in plasma. figure 6. global sensitivity towards the four parameters to determine threshold pic50. susanne winiwarter et al. admet & dmpk 6(1) (2018) 15-33 30 figure 7. sensitivity of remaining three parameters for different v values (clinthh and fu full line, fuinc dashed line; clint and fu have exactly same influence and are on top of each other, fuinc has opposite effect), areas of typical distribution volume values for acids, bases and neutrals indicated. influence of in silico model quality the quality of the in silico models needs to be considered for the present analysis. here we use astrazeneca’s in-house models for the five parameters, two of which have rather extensive training sets with more than 10,000 data points each, gathered over a long period of time (>10 years), whereas two, fuinc and papp, have intermediate training set size, and the fifth, volume of distribution, is based on literature data and uses only about 500 compounds. for the four models employing in-house data, about 70-75 % of the compounds within a temporal test set, i.e., compounds that have not been available when the model was built, were found to be within 3-fold of the experimental value. these results are rather encouraging when considering that the experimental variability in general is assumed to be about 2-fold [38]. the models are being regularly updated, as it was shown earlier that continuous updates of models enables new chemistry to be well represented and, thus, likely better predicted by the models [39]. it should be emphasized that the predictivity of the models needs to be investigated for the drug discovery project or chemical series in question when this approach is to be applied. note that, as was shown in figure 2, clearance prediction are reasonably accurate from in silico for most compounds in data set 2 and can be directly verified as soon as in vitro data is available for a compound. the remaining parameter, human distribution volume, is less easily available. the qsar model used here is based on human literature data and cannot be updated as straightforwardly as models using data from in-house screens. also, external data is not necessarily relevant to newest in-house chemistry, and model outcome can usually only be verified at later stages since in vivo data is required. nevertheless, the admet & dmpk 6(1) (2018) 15-33 in silico adme in drug design doi: 10.5599/admet.6.1.470 31 predictivity was considered good, with an error of prediction to be 0.4 log units (2.5-fold), and scrutinizing literature or databases for new data [40-42] in fairly regular intervals should ensure continued high quality of the model. conclusions the present study suggests the usage of in silico adme predictions to estimate which plasma concentration a new chemical structure may achieve when given orally using a defined dose and dosing regimen. from the plasma concentration a minimum required potency can be deduced, here referred to as threshold potency (or threshold pic50). while it was postulated, that early dose predictions from in silico was not yet possible [10], the idea here is to utilise predictions only to consider a compound’s summarized pharmacokinetic properties for ranking compounds within a series early on in the design process. combining the in silico predictions to a physiologically meaningful score, reduces the risk of deselecting a compound just because one of the parameters is outside an arbitrarily chosen limit. using the approach for a set of drug discovery compounds, it was shown that the threshold pic50 was able to define which compounds had a higher or lower chance of success, with the finally selected compound in the former set. it was assumed that the approach is most useful in lead optimisation stage, when the requirements for potency coverage are known. additionally, it was shown that the threshold potency values calculated for a set of marketed drugs were in a reasonable range for most of the compounds, when appropriate dosing regimens were considered for each. these results were also compared with the outcome, when experimental in vitro or human in vivo data were considered as input for the same procedure. in silico outcome was in many cases similar to both in vitro and in vivo, and when there were bigger differences usually the in vitro was closer to the in vivo outcome. thus, in silico prediction can be easily verified by in vitro experiments as soon as a compound is made. the examples shown here based threshold potency values on oral dosing and minimum concentration, i.e., required coverage over the whole dosing period. using such a dosing regimen, and based on a one compartment pharmacokinetic model considering immediate absorption, volume of distribution becomes the most important factor. special care needs to be taken for low volume compounds such as acids. obviously, the equations can be adapted to other coverage periods or dosing regimen, as required by the specific drug discovery project. references [1] s. ekins, j. rose. in silico adme/tox: the state of the art. journal of molecular graphics and modelling 20 (2002) 305-309. [2] h. van de waterbeemd, e. gifford. admet in silico modelling: towards prediction paradise? nature reviews drug discovery 2 (2003) 192-204. [3] m.p. gleeson, a. hersey, s. hannongbua. in-silico adme models: a general assessment of their utility in drug discovery applications. curr top med chem 11 (2011) 358-381. [4] l. tao, p. zhang, c. qin, s.y. chen, c. zhang, z. chen, f. zhu, s.y. yang, y.q. wei, y.z. chen. recent progresses in the exploration of machine learning methods as in-silico adme prediction tools. advanced drug delivery reviews 86 (2015) 83-100. [5] f. lombardo, e. gifford, m.y. shalaeva. in silico adme prediction: data, models, facts and myths. mini reviews in medicinal chemistry 3 (2003) 861-875. [6] f. lombardo, p.v. desai, r. arimoto, k.e. desino, h. fischer, c.e. keefer, c. petersson, s. winiwarter, f. broccatelli. in silico absorption, distribution, metabolism, excretion, and pharmacokinetics susanne winiwarter et al. admet & dmpk 6(1) (2018) 15-33 32 (adme-pk): utility and best practices. an industry perspective from the international consortium for innovation through quality in pharmaceutical development. journal of medicinal chemistry 60 (2017) 9097-9113. [7] m.d. segall, a.p. beresford, j.m.r. gola, d. hawksley, m.h. tarbit. focus on success: using a probabilistic approach to achieve an optimal balance of compound properties in drug discovery. expert opinion on drug metabolism & toxicology 2 (2006) 325-337. [8] m.d. segall. multi-parameter optimization: identifying high quality compounds with a balance of properties. current pharmaceutical design 18 (2012) 1292-1310. [9] d.f. mcginnity, j. collington, r.p. austin, r.j. riley. evaluation of human pharmacokinetics, therapeutic dose and exposure predictions using marketed oral drugs. current drug metabolism 8 (2007) 463-479. [10] k.m. page. validation of early human dose prediction: a key metric for compound progression in drug discovery. molecular pharmaceutics 13 (2016) 609-620. [11] m. lemurell, j. ulander, s. winiwarter, a. dahlen, o. davidsson, h. emtenas, j. broddefalk, m. swanson, d. hovdal, a.t. plowright, a. pettersen, m. ryden-landergren, j. barlind, a. llinas, m. herslof, t. drmota, k. sigfridsson, s. moses, c. whatling. discovery of azd6642, an inhibitor of 5lipoxygenase activating protein (flap) for the treatment of inflammatory diseases. journal of medicinal chemistry 58 (2015) 897-911. [12] j.b. houston. utility of in vitro drug metabolism data in predicting in vivo metabolic clearance. biochemical pharmacology 47 (1994) 1469-1479. [13] l.z. benet, p. zia-amirhosseini. basic principles of pharmacokinetics. toxicologic pathology 23 (1995) 115-123. [14] l. fredlund, s. winiwarter, c. hilgendorf. in vitro intrinsic permeability: a transporter-independent measure of caco-2 cell permeability in drug design and development. molecular pharmaceutics 14 (2017) 1601-1609. [15] a.-k. sohlenius-sternbeck, c. jones, d. ferguson, b.j. middleton, d. projean, e. floby, j. bylund, l. afzelius. practical use of the regression offset approach for the prediction of in vivo intrinsic clearance from hepatocytes. xenobiotica 42 (2012) 841-853. [16] m.p. gleeson, a. hersey, s. hannongbua. in-silico adme models: a general assessment of their utility in drug discovery applications. current topics in medicinal chemistry 11 (2011) 358-381. [17] r.s. obach, f. lombardo, n.j. waters. trend analysis of a database of intravenous pharmacokinetic parameters in humans for 670 drug compounds. drug metabolism and disposition 36 (2008) 13851405. [18] l.s. goodman, a.g. gilman, the pharmacological basis of therapeutics, 9th ed., mcgraw-hill publishers: new york, 1996. [19] v. svetnik, a. liaw, c. tong, j.c. culberson, r.p. sheridan, b.p. feuston. random forest: a classification and regression tool for compound classification and qsar modeling. journal of chemical information and computer sciences 43 (2003) 1947-1958. [20] p. bruneau. search for predictive generic model of aqueous solubility using bayesian neural nets. journal of chemical information and computer sciences 41 (2001) 1605-1616. [21] acd/labs software, 2015, advanced chemistry development, inc., toronto, ontario, canada. [22] clogp, 4.3, pomona college and biobyte, inc., claremont, ca, us. [23] f. pedregosa, g. varoquaux, a. gramfort, v. michel, b. thirion, o. grisel, m. blondel, p. prettenhofer, r. weiss, v. dubourg, j. vanderplas, a. passos, d. cournapeau, m. brucher, m. perrot, é. duchesnay. scikit-learn: machine learning in python. journal of machine learning research 12 (2011) 2825-2830. [24] h. wan, f. bergström. high throughput screening of drug protein binding in drug discovery. journal of liquid chromatography & related techniques 30 (2007) 681-700. admet & dmpk 6(1) (2018) 15-33 in silico adme in drug design doi: 10.5599/admet.6.1.470 33 [25] n.j. waters, r. jones, g. williams, b. sohal. validation of a rapid equilibrium dialysis approach for the measurement of plasma protein binding. journal of pharmaceutical sciences 97 (2008) 4586. [26] m.c. wenlock, l.a. carlsson. how experimental errors influence drug metabolism and pharmacokinetic qsar/qspr models. journal of chemical information and modeling 55 (2015) 125134. [27] n. cristianini, j. shawe-taylor, an introduction to support vector machines: and other kernel-based learning methods, university press, cambridge, 2000. [28] j.l. faulon, d.p. visco, jr., r.s. pophale. the signature molecular descriptor. 1. using extended valence sequences in qsar and qspr studies. journal of chemical information and computer sciences 43 (2003) 707-720. [29] v. vovk, a. gammerman, g. shafer, algorithmic learning in a random world, springer, new york, 2005. [30] genettasoft, http://www.genettasoft.com/ (2017-10-25). [31] a.-k. sohlenius-sternbeck, l. afzelius, p. prusis, j. neelissen, j. hogstraate, j. johansson, e. floby, a. bengtsson, o. gissberg, j. sternbeck, c. petersson. evaluation of the human prediction of clearance from hepatocytes and microsome intrinsic clearance for 52 drug compounds. xenobiotica 40 (2010) 637-649. [32] d.g. temesi, s. martin, r. smith, c. jones, b. middleton. high-throupghput metabolic stability studies in drug discovery by orthogonal acceleration time-of-flight (oatof) with analogue-to-digital signal capture (adc). rapid communications in mass spectometry 24 (2010) 1730-1736. [33] r.p. austin, p. barton, s. mohmed, r.j. riley. the binding of drugs to hepatocytes and its relationship to physicochemical properties. drug metabolism and disposition 33 (2005) 419-425. [34] f. cannavó. sensitivity analysis for volcanic source modeling quality assessment and model selection. computers & geosciences 44 (2012) 52-59. [35] matlab. the mathworks inc., natick, massachusetts, us, 2014. [36] m. fink, j.j. batzel, h. tran. a respiratory system model: parameter estimation and sensitivity analysis. cardiovascular engineering 8 (2008) 120-134. [37] http://uk.mathworks.com/matlabcentral/fileexchange/15235-automatic-differentiation-for-matlab (2017-10-11). [38] s. winiwarter, b. middleton, b. jones, p. courtney, b. lindmark, k.m. page, a. clark, c. landqvist. time dependent analysis of assay comparability: a novel approach to understand intraand intersite variability over time. journal of computer-aided molecular design 29 (2015) 795-807. [39] a.m. davis, d.j. wood. quantitative structure-activity relationship models that stand the test of time. molecular pharmaceutics 10 (2013) 1183-1190. [40] f. lombardo, y. jing. in silico prediction of volume of distribution in humans. extensive data set and the exploration of linear and nonlinear methods coupled with molecular interaction fields descriptors. journal of chemical information and modeling 56 (2016) 2042-2052. [41] m. olah, r. rad, l. ostopovici, a. bora, n. hadaruga, d. hadaruga, r. moldovan, a. fulias, m. mractc, t.i. oprea, wombat and wombat-pk: bioactivity databases for lead and drug discovery, chemical biology, wiley-vch verlag gmbh2008, pp. 760-786. [42] pharmapendium pk fact sheet. elsevier, https://www.elsevier.com/__data/assets/pdf_file/0009/80694/r_dsolutions_pharmapendium_fcat-sheet_-pharmacokinetics_digital.pdf (2017-10-24). ©2018 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://www.genettasoft.com/ http://uk.mathworks.com/matlabcentral/fileexchange/15235-automatic-differentiation-for-matlab https://www.elsevier.com/__data/assets/pdf_file/0009/80694/r_d-solutions_pharmapendium_fcat-sheet_-pharmacokinetics_digital.pdf https://www.elsevier.com/__data/assets/pdf_file/0009/80694/r_d-solutions_pharmapendium_fcat-sheet_-pharmacokinetics_digital.pdf http://creativecommons.org/licenses/by/3.0/ manuscript doi: 10.5599/admet.5.1.352 1 admet & dmpk 5(1) (2017) 1-8; doi: 10.5599/admet.5.1.352 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index mini review tyrosine kinase inhibitors: new molecules in non-small cell lung cancer (egfr and alk) fernando franco, míriam méndez*, lourdes gutierrez, mariano provencio puerta de hierro universitary hospital, madrid, spain *corresponding author: e-mail: mirimega@hotmail.com. received: october 25, 2016; revised: february 22, 2017; published: march 25, 2017 abstract lung cancer is a tumor pathology that includes a group of neoplasms with different biological characteristics, which determine the course of the disease and the potential treatment response depending on the subtype diagnosed. non-small lung carcinoma corresponds to 85 % of all lung cancers. adenocarcinoma is the most common histologic subtype of this group. in the united states approximately 130,000 new cases a year are diagnosed and in the world around a million. adenocarcinoma is the most likely subtype, with possible molecular alterations (egfr mutations and translocations in alk), which are a therapeutic target. knowledge of the molecular profile of the tumor and its biology has led us to identify therapeutic targets as protein tyrosine kinase, whose inhibition have demonstrated impact in terms of survival and quality of life for patients with these molecular alterations. however, the complex biology of non-small cell lung carcinoma gives the malignant cell the ability to generate mechanisms of resistance to these drugs and has forced us to extend the studies to control the disease by blocking these mechanisms. keywords lung cancer; egfr inhibitors, alk inhibitors introduction non-small lung carcinoma corresponds to 85 % of lung cancers. adenocarcinoma is the most common histologic subtype. in the united states approximately 130,000 new cases are reported per year [1]. patients with advanced non-small cell lung cancer (nsclc) usually receive systemic therapy. after histopathological diagnosis, an extensive molecular study is required to characterize the tumor. regardless of the clinical characteristics of the patient, at least the mutational status of the epidermal growth factor receptor (egfr) and alk should be analyzed [2-4]. activating mutations in egfr tyrosine kinase are observed in approximately 15 percent of nsclc adenocarcinoma in the united states and occur more frequently in nonsmokers [5]. in asian populations, the incidence of these egfr mutations is substantially higher, up to 62 percent [6]. in unselected nsclc populations, the alk rearrangement is a relatively rare event. the overall incidence of alk gene rearrangements has been about 4 percent. the presence of alk gene rearrangements in nsclc tumors tends to occur independently of egfr mutations. similar frequencies of alk gene rearrangements have been reported in asian and western populations. http://www.pub.iapchem.org/ojs/index.php/admet/index m. mendez et al. admet & dmpk 5(1) (2017) 1-8 2 egfr inhibitors first and second generation of egfr inhibitors patients with molecular alterations constitute a subpopulation with clinical characteristics, prognosis and response to different treatment (chemotherapy vs tyrosine kinase inhibitors) [5,7-9]. first-generation egfr tkis (including gefitinib and erlotinib) selectively target the receptor, through a competitive, reversible binding at the tyrosine kinase domain. however, despite good response there is evidence that after a variable period of time (5-13 months), the patients develop acquired resistance to egfr-tki therapy that negatively impacts their survival. second generation tkis (afatinib, dacomitinib) differ from first generation egfr-tkis because they form an irreversible link to the egfr kinase domain and inhibit receptors as her2 and her4 and are active agents particularly in those with exon 19 deletion. the increasing knowledge in tumor biology and the mechanisms, by which resistance develops, have helped to develop new molecules that have already demonstrated a positive impact on response rates and survival [9-11]. third generation of egfr inhibitors the t790m egfr mutation is present in over 60 % of cases of resistance to tkis first and second generation. it is the resistance mechanism of egfr most commonly known. this acquired resistance is mediated by a threonine-to-methionine mutation in egfr at position 790 (egfr t790m), activation of a bypass-signalling pathway such as mesenchymal-epithelial transition (met) pathway, human epidermal growth factor 2 (her2), or mitogen-activated protein kinase 1 (mapk1) amplification, and histologic transformation to small-cell lung cancer or an epithelial-to-mesenchymal transition [12-15]. following the acquisition of the mutation, the median survival is less than two years, so there is a need to develop new drugs that will control tumor growth in patients positive egfr t790m. osimertinib (azd9291) azd9291 approved by the fda in november 2015, is a new molecule tki third generation, which binds irreversibly to the residue kinase egfr at the site of atp binding by covalent bond formation, in patients who have acquired t790m mutation. after oral administration, it is metabolized azd9291 that has two pharmacologically active forms az5104 and az7550. az7550 showed a comparable potency and selectivity profile similar to osimertinib, whereas az5104 in addition exhibited greater potency against exon 19 deletion and t790m mutants and wild-type egfr. the first phase i dose escalation study was the aura, made with egfr mutated patients in progression with gefitinib or erlotinib treatment [11, 16-18]. the doses used were between 20 and 240 mg per day. in the expansion phase, the analysis of the t790m mutation patients demonstrated an objective response rate of 67 % (95 % ci 52-70 %) compared to 21 % in patients without the mutation. the rate of disease control in the overall analysis of all aura studies was 91 %. the median progression-free survival was 9.6 months versus 3 months in patients without the mutation. the profile of side events is favourable, since they are few and mostly grades 1 and 2. the most frequent were rash, diarrhea and nausea. since no significant difference in response to higher doses was observed, 80 mg/day is recommended until disease progression or unacceptable toxicity [16-18]. if it is necessary to reduce the dose, this should be done at 40 mg per day, since tablets are 40 and 80 mg. its elimination is mainly hepatic and has not been studied in patients with renal insufficiency [16]. current studies of liquid biopsy are presented as an option for detection of mutations in peripheral blood. this probably serves not only for diagnosis but also for monitoring; however, more studies are needed in this line. admet & dmpk 5(1) (2017) 1-8 tkis in non-small cell lung cancer doi: 10.5599/admet.5.1.352 3 rociletinib (co-16086) rociletinib, is a new oral selective covalent molecule egfr inhibitor, with demonstrated activity against the t790m mutation, deletionn 19 and l858r [19-20]. initial studies were performed in patients with acquired resistance to gefitinib and erlotinib. the subsequent inclusion of patients with mutation t790m has shown the particular activity of the drug in this subgroup of patients. partial results of studies of a phase ii study presented at asco 2015 reflect an objective response rate in this subgroup of 59 % (95 % ci 45-73) and rates of disease control 85 %. however, new combined analyses of the tiger-x and tiger-2 studies have shown that the actual response rate is only around 30 %. among the most common adverse events are diarrhea and skin toxicity, most of them in grades 1 and 2 [21]. cases of hyperglycemia and qt prolongation (> 10 %) were frequently presented, in the latter case without reaching documented arrhythmias. dose reductions occurred in 51 % of patients, most commonly due to hyperglycemia (22 %) and qtc prolongation (11 %). fifty-seven percent of patients had dose interruptions, which were most often due to hyperglycemia (22 %), qtc prolongation (10 %), and nausea (10 %). the average survival in patients was 8 months with the t790m mutation in patients where the 500-650 mg dose was administered. although initially seeming a promising drug, the adverse event profile and response rates have prompted the fda to give a negative opinion for accelerated approval of the drug [21-22]. hm61713 (bi 1482694) hm61713 is an irreversible kinase inhibitor that binds to a cysteine residue near the egfr kinase domain. preliminary studies in cell lines, has shown activity against t790m mutations and l858r. studies are currently underway with phase i / ii with a dose of 800 mg / day in patients with t790m mutation. initial results show an overall response rate of 62 % in this subgroup of patients. disease control rate was 91 % by independent assessment. the most common treatment-related adverse events included diarrhoea (55 %), nausea (37 %), rash (38 %) and pruritus (36 %), the majority were mild-to-moderate [23]. it seems a promising agent; however further studies to confirm therapeutic activity are needed. eluxa 1 is an ongoing pivotal phase ii clinical trial, designed to further investigate the efficacy and safety of bi 1482694 in patients with nsclc whose tumours acquired t790m-mediated resistance after first-line egfr tkis. asp8273 asp8273 is another small molecule, irreversible egfr tk inhibitor targeting egfr mutations including t790m. in mouse xenograft models, asp8273 achieved complete remission of tumors after 14 days of treatment. phase i / ii studies have established 400 mg / day as the optimal dose. the most common adverse events were gastrointestinal toxicity and thrombocytopenia. dose-limiting toxicities were reported at levels higher doses (400-600 mg). maximum tolerated dose was established as 300 mg. the objective response rate in patients with t790m was 25 % [24-25]. pf-06747775 and egf816 they are other new small molecules that inhibit t790m; however, they are in early stages of study and results are expected to verify their antitumor activity in the clinic [26-30]. similarly, there are clinical phase i-ii trials to evaluate the effectiveness of combinations of third generation tkis with various molecules, including met and mek inhibitors, as well as immunotherapy, among others. m. mendez et al. admet & dmpk 5(1) (2017) 1-8 4 alk inhibitors anaplastic lymphoma kinase (alk) is a member of the insulin receptor tyrosine kinase family (rtk). the echinoderm microtubule-associated protein-like 4-anaplastic lymphoma kinase (eml4-alk) fusion oncogene arises from an inversion on the short arm of chromosome 2 that joins eml4 to alk. the resulting chimeric protein, eml4-alk, contains an n-terminus derived from eml4 and a c-terminus containing the entire intracellular tyrosine kinase domain of alk. alk rearrangements occur in 3–7 % of patients with nsclc and are more common among patients with a non / light smoking history, adenocarcinoma histology, a younger age, female gender and in tumours wild type for egfr and kras [31]. anaplastic lymphoma kinase (alk) gene rearrangements may be detected using fluorescence in situ hybridization (fish) or immunohistochemistry (ihc). the gold standard assay for diagnosing is fish. to be declared alk positive, at least 15 % of 50 evaluated tumor cells must harbour split signals. ihc may be a more practical approach to detecting alk-positive nsclc. moreover, the absence of alk expression in normal lung tissue makes ihc less cumbersome to interpret than fish. in 2015, an ihc test using the d5f3 alk antibody (ventana) was approved after demonstration of concordance with the break apart fish test, with 94 % overall agreement [32]. first generation of alk inhibitors crizotinib (pf-02341066, xalkori) crizotinib is the first-generation alk inhibitor approved for alk-positive nsclc (4). it also has activity against c-met and ros1 tyrosine kinases. it was approved in first line and in pretreated alk positive nsclc patients [33-34]. response rates to crizotinib are approximately 60 %, which means there are a 40 % of primary resistances. one reason is because some alk translocations may not generate functional rearrangements in all patients [35]. there are secondary causes of resistance thought alk kinase mutations and it may occur in 30 % of the cases: l1196m, f1174l, and g1202r or c1156y mutation [36] activation of other oncogenes resulting in bypass signaling cascades such as egfr, ckit and kras or copy number gains of the alk fusion gene [37] and pharmacokinetic failings of the drug itself due to inadequate systemic levels or poor cns penetration and about 50 % of patients develop cns metastases during treatment with crizotinib [38-39] and in many cases, no definitive resistance mutation can be identified [40]. second generation of alk inhibitors these drugs were developed to overcome crizotinib-resistant mutations and to improve activity in cns disease (table 1). ceritinib ceritinib is 20-fold more potent than crizotinib in terms of alk selectivity [41]. it strongly inhibits resistance mutations, and especially l1196m, g1269a, i1171t and s1206y mutations. however, ceritinib was not effective against g1202r and f1174c crizotinib-resistant mutations. the updated ascend-1 report on the evaluation of the efficacy of ceritinib at a daily dose of 750 mg was recently presented. a total of 246 patients with alk+ nsclc were enrolled, including 163 pretreated with an alk inhibitor (crizotinib or alectinib) and 83 alk inhibitor naïve patients. the overall rate responses (orrs) were 56.4 %, 72.3 % and 61.8 % and the median of progression free survival (pfs) was 18.4, 9.0 and 6.9 months in the group pretreated with an alk inhibitor, alk inhibitor-naive patients and the overall population, respectively. 124 patients had measurable brain metastases at baseline: 28 patients pretreated with an alk inhibitor and 8 admet & dmpk 5(1) (2017) 1-8 tkis in non-small cell lung cancer doi: 10.5599/admet.5.1.352 5 patients in the alk inhibitor naive group. the intracranial orrs were 36 % (95 % ci 19–56 %) in patients pretreated with an alk inhibitor and 63 % (95 % ci 25–92 %) in the alk inhibitor-naive group [33,34]. the ascend-3 trial enrolled 124 patients, 40.3 % with brain metastases, 46 % of whom had no prior brt (brain radiotherapy). the orr was 63.7 % (95 % ci54.6–72.2 %). the intracranial orr was 20 % (95 %ci 2.5– 55.6 %) [35]. two randomized phase iii trials are currently open to evaluate ceritinib versus first-line chemotherapy (ascend-4) and second-line chemotherapy (ascend-5). alectinib alectinib is a highly selective alk inhibitor, in preclinical studies it showed activity against crizotinibresistant alk mutations (l1196m, c1156y and f1174l) and ret but not against insr, igf-r1 and ros1. japan was the first country to approve alectinib in patients with advanced alk-rearranged nsclc, based on the results of the phase i/ii af-001jp trial targeting patients with alk+ nsclc who had not previously been treated with crizotinib or other alk inhibitors [42,43]. in the phase ii trial, 46 patients were treated and 43 of them (93.5 %) achieved an orr (95 % ci 82.1–98.6). the 2-year pfs rate was 76 % (95 % ci 60–86 %) and the 2-year os was 79 % (95 % ci 63–89 %). a phase i/ii trial with 47 alk+ nsclc patients who progressed on or were intolerant to crizotinib. the orr was 55 % and the subset analysis of 21 patients with cns metastases at baseline showed a disease control rate approximating 90 %. the recommended dose for phase ii studies was 600 mg bid [44-46]. brigatinib brigatinib is a novel potent oral drug that has demonstrated ability to overcome crizotinib resistance mutations, including g1202r and activity against ros1 and inhibits mutant egfr, including t790m [47-49]. brigatinib showed antitumor activity in patients treated with crizotinib with an objective response rate of 71 % and in crizotinib-naive patients of 100 %. the median pfs was 13.4 months in the subgroup with prior crizotinib therapy [50]. a pivotal, randomized, phase 2 trial of brigatinib in patients with crizotinib-resistant alk+ nsclc has completed accrual. entrectinib entrectinib is a novel, orally available, selective tyrosine kinase inhibitor of the trk family of tyrosine kinases, ros1 and alk. rearrangements in ntrk1 were recently described in approximately 3 % of nsclc that were negative for other oncogenic driver mutations [51]. entrectinib has demonstrated in vitro and in vivo antitumour activity against various trk, ros1 or alk-driven xenograft models of different human cancers (npm-alk-driven lymphoma and eml4-alk-driven nsclc) entrectinib has adequate oral bioavailability and can efficiently cross the blood–brain barrier in mice with intracranially injected ncih2228 eml4-alk cells [52]. entrectinib 400 mg/m2 is the dose recommended for the trials phase ii [53]. third generation of alk inhibitors lorlatinib (pf-06463922) lorlatinib has been identified as an orally novel reversible available atp-competitive selective targeting of a specific residue leucine at position 1198 (l1198) detected in only approximately 25 % of kinases, brain penetrant inhibitor of alk (including all known mutations) and the c-ros oncogene 1 (ros1). in vitro and preclinical studies it is ∼10-fold more potent against wildtype eml4-alk and ∼40-fold more potent against eml4-alk l1196m compared with crizotinib [54]. importantly, lorlatinib has demonstrated activity in patients whose tumors harbor the highly resistant mutation alk g1202r. this mutation confers resistance m. mendez et al. admet & dmpk 5(1) (2017) 1-8 6 to other next-generation alk inhibitors. other drugs for crizotinib resistant tumor asp3026 this is an alk inhibitor active against crizotinib-resistant alk-positive advanced tumors [55]. x-376 and x-396 (xcovery) these are more potent inhibitors of alk but less potent inhibitors of met compared to crizotinib. moreover, x-396 could potently inhibit alk with two point mutations (l1196m and c1156y) associated with acquired resistance to crizotinib and there is preclinical data about cns penetration [56]. cep-28122 cep-29122 is a potent and selective alk inhibitor which showed selectivity in preclinical studies against alk compared to various types of tyrosine kinases, including insr, igf-r1 and c-met. cep37440 is an alk inhibitor and focal adhesion kinase (fak) [57-58]. hsp90 inhibitors another possibility to overcome resistance to alk inhibitors includes the combination with hsp90 inhibitors such as retaspimycin hydrochloride (ipi-504) and ganetespib (sta-9090) that have demonstrated clinical activity in alk+ nsclc patients [60-62]. several clinical trials are evaluating hsp90 inhibitors in combination with alk inhibitors. crizotinib is combined with ganetespib in crizotinib-naive patients and in crizotinib-resistant disease. in other studies crizotinib is combined with onalespib (at13387), while ceritinib is combined with auy922. table 1. main molecular characteristics of second and third-generation alk inhibitors. drug targets other than alk l1196m resistance mutation c1156y resistance mutation g1202r resistance mutation ceritinib igf-r1, insr, ros1 sens. res. res. alectinib ltk, gak sens. sens. res. brigatinib ros1, egfr sens. sens. sens. entrectinib trka, trkb, trkc, ros1 sens. sens. nd lorlatinib ros1 sens. nd sens. tsr-011 trka, trkb, trkc sens. nd nd asp3026 ros1, ack sens. nd nd x-396 met sens. sens. nd conclusions current knowledge of the biology of nsclc, allows the development of new molecules that are able to control the disease and have an impact on disease control, survival and quality of life. the new thirdgeneration egfr tkis demonstrate efficacy in patients with egfr resistance mutations patients while the new alk inhibitors show activity in patients with alk rearrangements and mutations. these benefits alone have already allowed the registration of some of them. results from other studies are pending, as well as those related to combinations with other drugs. admet & dmpk 5(1) (2017) 1-8 tkis in non-small cell lung cancer doi: 10.5599/admet.5.1.352 7 references [1] r. siegel, j. ma, z. zou, a. jemal. ca cancer j. clin. 64(1) (2014) 9-29. [2] n. lindeman, p. cagle, m.b. beasley et al. arch. pathol. lab. med. 137(6) (2013) 828–860. [3] t.j. lynch, d.w. bell, r. sordella, et al. n. engl. j. med. 350(21) (2004) 2129-2139. [4] e.l. kwak, y.j. bang, d.r. camidge, et al. n. engl. j. med. 363(18) (2010) 1693-1703. [5] r. rosell, e. carcereny, r. gervais, et al. lancet oncol. 13(3) (2012) 239–246. [6] f.-y. niu, y.-l.wu. exp. hematol. oncol. 3(1) (2014) 2. [7] n.b. leighl, n. rekhtman, w.a. biermann, et al. j. clin. oncol. 32(32) (2014) 3673. [8] a. chi, s. remick, w. tse. biomark. res. 1(1) (2013) 2. [9] c.k. lee, y.-l. wu, p.n. ding, et al. j. clin. oncol. 33(17) (2015) 1958–1965. [10] a.o. walter, r.t. sjin, h.j. haringsma, et al. cancer discov. 3(12) (2013) 1404–1415. [11] d.a. cross, s.e. ashton, s. ghiorghiu, et al. cancer discov. 4(9) (2014) 1046–1061. [12] s. kobayashi, t.j. boggon, t. dayaram, et al. n. engl. j. med. 352 (2005) 786–792. [13] j.a. engelman, k. zejnullahu, t. mitsudomi, et al. science 316 (2007) 1039–1043. [14] k. takezawa, v. pirazzoli, m.e. arcila, et al. cancer discov. 2 (2012) 922–933. [15] t.m. kim, a. song, d.-w. kim, et al. j. thorac. oncol. 10 (2015) 1736–1744. [16] p.a. jänne, j.c. yang, d.w. kim, et al. n. engl. j. med. 372(18) (2015) 1689. [17] s.s. ramalingam, j. yang, c. lee, et al. j. thoracic. oncol. 10(9_suppl 2) (2015) s319. [18] c.a. eberlein, d. stetson, a.a. markovets, et al. cancer res. 75(12) (2015) 2489–500. [19] a.o. walter, r.t. sjin, h.j. haringsma, et al. cancer discov. 3 (2013) 1404-1415. [20] l.v. sequist, j.c. soria, j.w. goldman, et al. n. engl. j. med. 372(18) (2015) 1700–1709. [21] j.c. soria, d.r. camidge. n. engl. j. med. 374 (2016) 2296-2297 [22] fda briefing document oncologic drugs advisory committee meeting. nda 208542 rociletinib. april 12, 2016 [23] k.-o. lee, m.y. cha, m. kim, et al. cancer res. 74(19) (2014) lb-100. [24] h. sakagami, s. konagai, h. yamamoto, et al. cancer res. 74(19 supplement) (2014) 1728. [25] y. jia, j. juarez, m. manuia, g. lelais, et al. cancer res. 74(19 supplement) (2014) 1734. [26] s. kasibhatla, j. li, c. tompkins, et al. cancer res. 74(19 supplement) (2014) 1733. [27] y. goto, h. nokihara, h. murakami, et al. asco meet abstr. 33(15_suppl) (2015) 8014. [28] g. lelais, r. epple, p.-y. michellys, et al. cancer res. 75(15 supplement) (2015) 2585. [29] study for patients with nsclc egfr mutations (del 19 or l858r +/t790m). clinicaltrials.gov identifier: nct02349633. [30] a. ullrich, j. schlessinger. cell 61(2) (1990) 203–212. [31] d. camidge, s. kono, a. flacco, et al. clin. cancer res. 16(22) (2010) 5581–5590. [32] j.j. cui, m. tran-dubé, h. shen, et al. j. med. chem. 54(18) (2011) 6342–6363. [33] a.t. shaw, s.h. ou, y.j. bang, et al. n. engl. j. med. 371(21) (2014) 1963–1971. [34] b.j. solomon, t. mok, d.w. kim, et al. n. engl. j. med. 371(23) (2014) 2167–2177. [35] e. felip, s. orlov, k. park, et al. j. clin. oncol. 33 (2015) 8060. [36] j.m. heuckmann, h. balke-want, f. malchers, et al. clin. cancer res. 18(17) (2012) 4682–4690. [37] y.l. choi, m. soda, y. yamashita, et al. n. engl. j. med. 363(18) (2010) 1734–1739. [38] r. katayama, a.t. shaw, t.m. khan et al. sci. transl. med. 4(120) (2012) 120ra17. [39] j. gainor, s.h. ou, j. logan, l.f. borges, a.t. sahw. j. thorac. oncol. 8(12) (2013) 1570–1573. [40] d. costa, a.t. shaw, s.h. ou, et al. j. clin. oncol. 33(17) (2015) 1881–88. [41] r. doebele, a. pilling, d. aisner, et al. clin. cancer res. 18(5) (2012) 1472–1482. m. mendez et al. admet & dmpk 5(1) (2017) 1-8 8 [42] l. friboulet, n. li, r. katayama, et al. cancer discov. 4(6) (2014) 662–673. [43] k. kinoshita, k. asoh, n. furuchi, et al bioorg. med. chem. 20(3) (2012) 1271–1280. [44] t. seto, k. kiura, m. nishio, et al. lancet oncol. 14(7) (2013) 590-598. [45] t. tamura, t. seto, k. nakagawa, et al. radiat. oncol. 90 (2014) s6. [46] s. gadgeel, l. gandhi, g. riely, et al. lancet oncol. 15 (2014) 1119–1128. [47] j. wu, j. savooji, d. liu. j. hematol. oncol. 9 (2016) 9. [48] r. squillace, r. anjum, d. miller, et al. cancer res. 73(8 suppl) (2013) abstract #5655. [49] v. rivera, f. wang, r. anjum, et al. cancer res. 72(8 suppl) (2012) abstract #1794. [50] d. camidge, l. bazhenova, r. salgia, et al. j. clin. oncol. 33 (2015) 8062. [51] a. vaishnavi, m. capelletti, a.t. le, et al. nat. med. 19(11) (2013) 1469–1472. [52] e. ardini, m. menichincheri, c. de ponti, et al. mol. cancer ther. 8(12 suppl) (2009) a243. [53] m. patel, t. bauer, s. liu, et al. j. clin. oncol. 33 (2015) 2596. [54] t. johnson, p. richardson, s. bailey, et al. j. med. chem. 57(11) (2014) 4720–4744. [55] t. li, p. lorusso, m. maitland, et al. j. hematol. oncol. 9 (2016) 23. [56] c. lovly, j. heuckmann, e. de stanchina, et al. cancer res. 71(14) (2011) 4920–4931. [57] m. cheng, m. quail, d. girgrich, et al. mol. cancer ther. 11 (2012) 670–679. [58] i. sullivan, d. planchard. ther. adv. med. oncol. 8(1) (2016) 32–47. [59] l. sequist, s. gettinger, n. sanzer, et al. j. clin. oncol. 28(33) (2010) 4953–4960. [60] e. normant, g. paez, k. west, et al. oncogene 30(22) (2011) 2581–2586. [61] m. socinski, j. goldman, i. el-harary, et al. clin. cancer res. 19(11) (2013) 3068–3077. ©2017 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) . http://creativecommons.org/licenses/by/3.0/ doi: 10.5599/admet.2.2.36 115 admet & dmpk 2(2) (2014) 115-125; doi: 10.5599/admet.2.2.36 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper diagnosing solubility limitations – the example of hydrate formation jörg berghausen*, bernard faller novartis institutes for biomedical research, novartis campus, ch-4002 basel *corresponding author: e-mail: joerg.berghausen@novartis.com; tel.: +41-61-69-62335 received: march 18, 2014; revised: june 05, 2014; published: july 18, 2014 abstract solubility is regarded as one of the key challenges in many drug discovery projects. thus, it is essential to support lead finding and optimization efforts with appropriate solubility data. in silico solubility prediction remains challenging and therefore a screening assay is used as a first filter, followed by selected follow-up assays to reveal what causes the low solubility of a specific compound or chemotype. results from diagnosing the underlying reason for solubility limitation are discussed. as lipophilicity and crystal lattice forces are regarded as main contributors to limiting solubility, changes in solid state are important to be recognized. solubility limitation by various factors will be presented and the impact of the solid -state is exemplified by compounds that are able to form hydrates. keywords: polar surface area; lipophilicity; crystal lattice energy introduction the importance of solubility for a molecule to become a successful drug has been debated for decades, as compounds with insufficient solubility bear a higher risk of attrition and are linked to higher cost in drug development. solubility issues have been associated with several factors, like increased use of highthroughput screening procedures and thus the tendency to higher lipophilicity and molecular weight [1]. furthermore, there is a shift of therapeutic targets from g-protein coupled receptors and enzymes to more challenging ones, such as kinases, ion channels, nuclear receptors and protein-protein interactions. for these targets, often a higher lipophilicity [2] or strong intermolecular interactions, such as intermolecular hydrogen bonds, is required. both lipophilicity and intermolecular bonds tend to negatively affect compound solubility, and the aim of this work is to present a way to distinguish them using an example in which the solid state contribution is impacted by hydrate formation. in fact, solubility still remains one of the key challenges in many drug discovery projects. thus, it is essential to support the lead finding and optimization efforts with appropriate solubility data. despite advancement in in silico solubility prediction, it remains a challenging field, mainly due to the low success of prediction of solid state contribution to the thermodynamic solubility of a crystalline drug compound. as a consequence, a screening assay is used as a first filter in a modern drug discovery environment, followed by selected follow-up assays to reveal what causes the low solubility of a specific compound or chemotype. solubility high-throughput screening assays typically allow the determination of solubility with a reasonable experimental error and within short turn-around time. their main drawback is that the solidhttp://www.pub.iapchem.org/ojs/index.php/admet/index mailto:joerg.berghausen@novartis.com berghausen and faller admet & dmpk 2(2) (2014) 115-125 116 state of the investigated compounds is either not characterized at all or at best by methods that allow a qualitative assessment only, like e.g. birefringence or polarized microscopy [3]. such high-throughput solubility data allows tracking the development of a chemical series, e .g. by creation of qsar models [4]. nevertheless, to overcome solubility issues beyond lipophilicity, guidance for the chemistry teams on what is the root cause of solubility limitation is inevitable. for that, secondary assays need to be applied that allow conclusions to be drawn on the interplay of the main solubility limiting factors. for most of the compounds in the chemical space of low molecular weight drug candidates these factors comprise lipophilicity, crystal lattice forces and ionization. in addition, other factors like common ion effect and compound aggregation in solution may further limit the solubility of a molecule. as a drawback, these investigations require the availability of about 20 mg of crystalline material and are carried out at medium throughput. consequently, this kind of solubility diagnosis can be performed on selected molecules only. nevertheless, when performed early enough during lead optimization, solubility issues of a chemical series can be revealed and the knowledge acquired can be applied to the optimization strategy. acceleration of solubility diagnosis could rather profit from advancements in in silico prediction of properties that are currently not accessible at the desired level of quality. these include the melting point and enthalpy of fusion as a surrogate for crystal lattice energy, as well as the lipophilicity. although prediction of lipophilicity is broadly used with sufficient accuracy in general chemical space, the applicability of the selected prediction tool for a specific chemical series needs to be confirmed by high quality measurements. prediction of melting point and lattice energy is currently not available at a level beyond classification (high/intermediate/low) [5]. additionally, it has been attempted to express the contribution of crystal lattice by chemical descriptors [6]; likewise improvements to the ‘general solubility equation’ (gse) by including the psa have been reported [7]. in the present work, the procedure to reveal the root cause of solubility limitation is exemplified by various compounds, including some that are capable of forming hydrates. typically, the impact of the solid state on the performance of a chemical entity as a drug is investigated in the early development phase in the context of clinical formulation development, rather than during lead optimization. however, solid state limitations pose a risk that needs to be identified before compounds are selected for clinical development. as a consequence for oral administration, thermodynamic driving forces might limit the absorption of a compound in the human gastro-intestinal tract, despite any formulation efforts. experimental solubility assays high-throughput equilibrium solubility sample preparation for high-throughput equilibrium solubility determination is performed in microtiter plate format on an automated platform that is based on a starplus module (hamilton). in addition, the platform comprises a centrifuge (rotanta, hettich), two plateshakers (timix 5, edmund bühler), a heat sealer (alps-3000, thermo scientific) as well as an evaporator (combidancer, hettich). plates are handled by a robotic arm (tx60 l, stäubli) that allows the transfer between the starplus module and the various instruments. sample analysis is performed by mass spectrometry using a rapid-fire-q-tof combination (agilent). results are stored in a proprietary laboratory information system and data evaluation is performed by excel (microsoft). admet & dmpk 2(2) (2014) 115-125 diagnosing solubility limitations doi: 10.5599/admet.2.2.36 117 secondary assays equilibrium solubility is determined using the shake-flask method. about 1.5 mg of compound is put into a vial and 0.75 ml of buffer is added. the mixture is shaken for 20 to 24 hours and the solid phase is separated by centrifugation. concentration of the compound in the supernatant is determined by hplc-uv (agilent 1200) based on a calibration curve. the partition coefficient between aqueous phase and 1-octanol (log d) is determined by various methods. ionizable compounds are analyzed by potentiometric titration (sirius t3, sirius instruments ltd.). non-ionizable compounds are either analyzed by a shake flask method or by reverse-hplc (elogp) [8]. determination of the acid dissociation constant (pka) is also done by potentiometric titration (sirius t3, sirius instruments ltd.). 0.7-1.2 mg of compound (final concentration 1 mm) is dissolved in meoh/h2o (0.15 m kcl). titration is performed in 60/50/40 or 40/30/20 % w/v meoh depending on the solubility. the pka is extrapolated to 0 % water with the yasuda-shedlovsky extrapolation [9]. intrinsic solubility (i.e. solubility of the neutral species) is determined by a potentiometric titration approach [10]. 0.7-1.2 mg of compound (final concentration 1 mm) is dissolved in aqueous buffer. the intrinsic solubility is determined by mathematical analysis of the titration curve based on measured pka values. ionization species distribution is calculated by the henderson–hasselbalch equation. polar surface area (psa) is calculated by the topological psa method according to ertl et al. [11]. melting point and enthalpy of fusion are determined by differential scanning calorimetry (dsc) with a dsc q2000 (ta instruments). a standard heating rate of 10 k/min was used. solubility diagnosis the solubility diagnosis tool is an advancement of the analysis published by faller and ertl [12]. the relationship between log p and log 1/s0 revealed that only a minority of compounds show higher solubility than determined by their lipophilicity. many potential drug candidates do not fall on the unity line between log p and log 1/s0, but show rather lower solubility than predicted by log p. therefore, the authors [12] introduced the sl parameter (eq. 1) as a descriptor of additional interactions that limit solubility. (1) in order to quantify the contribution of the various factors that contribute to solubility limitation, a table has been created that summarizes the solubility in the relevant buffer medium, ionization effects, the solute-solvent interactions as well as crystal lattice energy contributions (see table 1). table 1. parameters generated for solubility diagnosis a b c d e f g h i log 1/s6.8 (measured) log 1/s6.8 (extrapolated) log 1/s0 ip6.8 elogp sl tm hf,m cf the solubility diagnosis mainly helps to compare different factors that contribute to solubility limitation. column a and b in table 1 compare the measured solubility at ph 6.8 to the extrapolated solubility value, based on intrinsic solubility (column c) and the compound’s ionization state. column d contains the ionization parameter, which is a logarithmic parameter derived from the ionized fraction. the latter, in turn, is calculated taking into account the compound formula (column i). the sl parameter in column f, berghausen and faller admet & dmpk 2(2) (2014) 115-125 118 according to eq. (1), describes the solubility lowering effect according to other factors than log p. one main factor that limits solubility beyond lipophilicity is the contribution of crystal lattice energy, especially for compounds exhibiting rather high melting temperature. table 2. description of parameters used for solubility diagnosis parameter description remark s0 intrinsic solubility s6.8 (measured) solubility at ph 6.8 measured in buffer medium s6.8 (extrapolated) solubility at ph6.8 derived from experimental pka and logs0 using the henderson-hasselbalch equation fi fraction ionized ip6.8 ionization parameter at ph 6.8 ip = -log (1-fi) clogp calculated logp elogp experimental logp sl difference between solubility and lipophilicity sl = log 1/s0 – log p tm melting point hf,m heat of fusion at the melting temperature cf compound formula ah: acid, b: base, 0: neutral, xh: ampholyte, zh: zwitterion in order to obtain guidance in analyzing the solubility limiting factors of a given compound at a defined ph, table 3 summarizes some guidelines that help interpreting the results. table 3. guidelines to analyze the solubility diagnosis dataset column guidance a and b if log 1/s6.8 (measured) >> log 1/s6.8 (extrapolated) then solubility is limited by solubility product and/or aggregation c, e and f if sl > 2 then solubility is limited by other factors than log p b and c if log 1/s6.8 (measured) or log 1/s6.8 (extrapolated) are close to log 1/s0 then solubility is limited by ionization f and g if tm>200 or sl >2 then solubility is limited by high crystal lattice energy f and g tm>200 but sl <2 indicates possible supersaturation in the solubility assay i if compound formula = xh then solubility is usually lowest at physiological ph i if compound formula = ah then solubility is low in the stomach admet & dmpk 2(2) (2014) 115-125 diagnosing solubility limitations doi: 10.5599/admet.2.2.36 119 results and discussion what factors do influence compound solubility? the solubility of a compound is mainly influenced by its solute-solvent interaction, its solid-state interaction [13] as well as its ionization state. from these three main factors, the solute-solvent interactions are typically described by the partition coefficient log p [14]. the log p value is an intrinsic property of a molecule and thus depends on its chemical constitution. in contrast, solid state properties are accessible to changes of the solid form. various solid forms comprise the thermodynamically most stable crystalline form as well as further polymorphic forms that may exist, including the amorphous state as well as solvates and hydrates. it should therefore be obvious that the nature of the solid form has to be well characterized in order to understand the solid state contribution to solubility limitation. the ionization state of a compound is defined by its chemical structure and consequently by its pka values. generally, the ionization state of a molecule depends on the pka value of the surrounding solvent, too. in an aqueous environment, its ionization state is a function of the ph value. thus, for a drug-like molecule, a bio-relevant ph value needs to be defined that is relevant for its solubility at the site of absorption. here, a ph value of 6.8 is chosen even though a different value might be more appropriate based on the ph gradient in the gastro-intestinal tract or when a non-oral route of administration is considered. solid-state and lipophilicity – which one is easier to handle? which of the hurdles contributing to solubility limitations can be overcome more easily? in the case of solute-solvent interactions, the only way to impact the log p is during lead optimization. certainly, the room for maneuver is often limited for the medicinal chemist due to the need to balance molecular properties, especially not to lose potency on the relevant target. in other words, medicinal chemistry will not be able to avoid a certain level of lipophilicity. as other factors will add up which further decrease solubility of a molecule [12], it appears a reasonable strategy to avoid further contributions that negatively affect compound solubility, including solid state contributions. it is worth mentioning that solid-state properties are determined by molecular constitution, as well. thus, many lead structures favor strong molecular interactions and increasing their solubility by lead optimization will remain a challenging task [15][16]. despite the fact that the solid state can be impacted by formulation efforts, it has been shown that compounds that show mainly solvation limited solubility appear to have a higher chance to successfully reach the market [17]. organic hydrates in solubility diagnosis in the lead optimization phase, solid state properties are typically not evaluated in detail, if at all. bearing in mind the awareness of solid state contributions to solubility limitation, an attempt was made to understand the impact of such solid form changes on the solubility diagnosis results. as an example, hydrate formation was selected. as the performance of drug compounds is investigated in aqueous environment where the formation of hydrate forms can negatively impact its solubility, the stability of the solid form under investigation may implicate the conclusions drawn on solubility limitation. for this, selected compounds from in-house lead optimization projects as well as marketed drugs were examined, which all of them are known to form hydrates. for proprietary compounds, hydrate formation has been confirmed by in-depth solid state analysis. investigation results on nitrofurantoin and carbamazepine are publicly available. the following table (table 4) summarizes the solubility diagnosis for the selected compounds. in addition, the psa values are included to facilitate discussion of figure 2. berghausen and faller admet & dmpk 2(2) (2014) 115-125 120 table 4. examples of hydrates in solubility diagnosis a b c d e f g h i psa log1/s6.8( measured) log1/s6.8( extrapolated) log1/s0 ip6.8 elogp sl tm hf,m cf cpd1 4.4 4.4 4.9 0.48 3.8 1.1 97.1 12.6 b 113 cpd2 4.9 3.9 3.9 0.0 5.0 -1.1 243.7 22.1 zh2 57 cpd3 4.0 3.5 5 1.5 4.1 0.9 177.2 29.7 ah 134 cpd4 6 6.1 6.1 0.0 3.1 3.0 133.6 8.0 b 108 nitrofurantoin 3.1 3.1 3.3 0.2 -0.1 3.4 269.3 14.2 ah 120 carbamazepine 3.1 3.1 3.1 0.0 1.7 1.4 190.1 25.5 0 46 compound cpd1. the compound is a base with an experimentally determined log p value of 3.8. melting point determination by dsc revealed a rather low melting temperature of 97.1 °c. measured and extrapolated solubility are in agreement, the fraction ionized for this compound is 0.67. solubility diagnosis for this compound reveals that solubility is mainly limited by lipophilicity, as ls is far below 2 and melting temperature below 200 °c. the impact of the hydrate to solubility limitation is therefore not obvious. compound cpd2. this compound is practically not ionized at ph 6.8, its extrapolated solubility is ten times higher than the value determined by the shake flask method. it’s a zwitterion with a log p of 5.0 and the sl value reveals additional solubility limitation by roughly 1 logarithmic unit. however, the melting point above 200°c can be regarded as strong indicator for solubility limitation by crystal lattice forces. compound cpd3. in this example, an acidic compound was investigated. again, low sl and low melting temperature indicate no significant solubility limitation by solid-state contributions. compound cpd4. in this case, the compound is not ionized, but it is rather insoluble. a quite high sl of 3.0 indicates a thousand times lower solubility than what could be expected based on its lipophilicity (log p 3.1). again, the melting temperature of 133.6°c does not indicate a strong crystal lattice contribution. nitrofurantoin. solubility of this compound appears to be controlled mainly by other factors than lipophilicity. the compound is characterized by a log p of -0.1 and a quite high melting temperature around 269°c. carbamazepine. carbamazepine is not ionized at ph 6.8, the compound is not very lipophilic, as its log p of 1.7 suggests. neither the sl parameter of 1.0 nor the melting point at 190 °c necessarily indicate a significant solid-state contribution to solubility limitation. overall, within the solubility diagnosis parameter set, there is no direct investigation of the nature of the solid state. if a compound has been received in the laboratory as hydrated or generally solvated form, one admet & dmpk 2(2) (2014) 115-125 diagnosing solubility limitations doi: 10.5599/admet.2.2.36 121 may realize from dsc measurement that solvent evaporation has taken place or other solid form changes have most likely occurred. more difficult to judge, however, is a potential solid form change during equilibration in aqueous environment. if compounds arrive for solubility diagnosis in anhydrated form, characterization of solid form after equilibration would be necessary to identify changes in solid state. given the amount of material and the technology needed, such investigations are far beyond the scope of lead optimization activities. additionally, salts from the buffer solutions complicate solid form investigations after equilibration experiments. therefore, it would be desirable to classify compounds for their possibility of hydrate formation by other means that require less experimental effort, in order to cope with the time and material constrains during compound optimization phase. ideally, structure based chemical descriptors would allow at least to point out to a certain possibility of hydrate formation. according to infantes et al. [18], the probability of hydrate formation is not correlated to the ratio of hbond donors to acceptors, but to the sum thereof. as the sum of h-bond donors and acceptors is highly correlated to the psa, the frequency of hydrate formation is found to increase with psa, too. in order to review the contributions that h-bond donors and acceptors provide to classify compounds during the lead optimization process without detailed analysis of solid state properties, the currently available in-house dataset of solubility diagnosis was analyzed. it comprises about 200 proprietary compounds, for which the number of known hydrates has been counted. indeed, analysis revealed that the frequency of hydrate formation increases with increasing sum of h-bond donors and acceptors and thus also with increasing psa (see figure 1). the results are in accordance with the results from infantes et al. [18], and based on our analysis, an increasing number of hydrated compounds is found when the sum of h-bond donors and h-bond acceptors is greater than 9. according to our dataset, nearly 20% of the compounds with a sum of 10 are found to form hydrates. although in this dataset the number of compounds with a sum of more than 11 was rather low, the frequency distribution as percentage within the respective binning showed a steady increase with increasing sum of donors and acceptors. however, the total number of confirmed hydrates at this stage of drug development is limited, thus the absolute frequency of hydrate formation is likely to be underestimated. this may explain why the overall percentage of hydrates in our proprietary dataset is lower as compared to results in reference [18]. what can we learn from hydrates about solid state contribution to solubility limitation? since the idea of solubility diagnosis is to identify the main factors that contribute to solubility limitation, the in-house dataset was analyzed in regard of the role of psa for compounds that are either identified as solubility limited by lipophilicity or by crystal lattice forces. psa is preferred over the sum of hbond donors and acceptors since it can be easily calculated as a descriptor for qsar modeling. because an increase of the sum of h-bond donors and acceptors, and thus psa, has been found to lead to a higher frequency of hydrate formation, an additional effect of solid-state contribution with increasing polar interactions could also be suspected. thus, the role of polar interactions on solubility limiting contributions within our proprietary dataset should be elucidated. for this, compounds that were classified by solubility diagnosis as being mainly limited by either lipophilicity or crystal lattice forces have been analyzed with respect to their psa values. berghausen and faller admet & dmpk 2(2) (2014) 115-125 122 (b) (a) figure 1. (a) frequency distribution of hydrates vs the sum of h-bond donor (hbd) s and h-bond acceptors (hba) in the proprietary solubility diagnosis dataset; total numbers. (b) percentage of confirmed hydrates within the respective sum of h-bond donors and acceptors. table 5 summarizes the average and median values of compounds from these classes. the analysis of both average and median psa values strongly suggests that compounds with a significant crystal lattice contribution to solubility limitation are characterized by a higher psa value. as illustrated by the histogram in figure 2 (a), the majority of compounds that reveal a significant crystal lattice contribution to solubility limitation are characterized by a psa of 75 å 2 or more. in case of compounds with lipophilicity being the main limiting factor, this distribution is shifted towards lower psa values, see figure 2 (b). here, the majority of compounds are characterized by a psa of less than 75 å 2 . notably, only few compounds in the current dataset reveal a psa of greater than 100 å 2 in cases where lipophilicity dominates solubility limitation. table 5. average and median values of psa for compounds classified by type of solubility limitation crystal lattice limited lipophilicity limited average psa / å 2 82.3 71.4 median psa / å 2 83.1 69.8 admet & dmpk 2(2) (2014) 115-125 diagnosing solubility limitations doi: 10.5599/admet.2.2.36 123 (a) (b) figure 2. (a) frequency distribution of psa for compounds being mainly solubility limited by crystal lattice fores (b) frequency distribution of psa for compounds with solubility mainly limited by lipophilicity. compounds that have undergone solubility diagnosis so far represent a molecular weight range of approximately 255 to 600 g/mol and calculated log p values in the range of approximately 0.4 to 7. although increasing molecular weight leads to increasing psa, see figure 3 (a), most of the compounds cluster between a psa of 60 å 2 and 120 å 2 and comprise the full range of molecular weight represented in the dataset. similarly, compounds with higher psa by trend are characterized by a lower calculated log p, see figure 3 (b). however, there is no strong relationship found within the dataset that would allow classifying the type of solubility limitation based on molecular weight only. however, based on above mentioned findings, it becomes evident that the ‘general solubility equation’ (gse) can be successfully refined by including the psa [7]. (a) (b) figure 3. (a) molecular weight as a function of psa for the solubility diagnosis dataset. (b) calculated logp as a function of psa for the solubility diagnosis dataset conclusions solubility diagnosis has proven to be a very useful tool for understanding the solubility limiting factors of drug-like molecules. by comparing contributions from lipophilicity, crystal lattice energy and ionization, the dominant factors that limit solubility are identified. however, the solid state of a compound is typically not berghausen and faller admet & dmpk 2(2) (2014) 115-125 124 well understood during lead optimization efforts. thus, changes of solid state can significantly impact the conclusions drawn on its contribution to solubility limitation. in this work, solubility limitations of hydrated forms of drug-like compounds were investigated. it is concluded that the contribution of hydrates to solubility limitation might be underestimated and that hydrates and solvates are special cases that are not necessarily sufficiently characterized during solubility diagnosis. this is due to the fact, that the melting temperature is used as a surrogate for crystal lattice energy. since hydrates may either decompose during dsc measurements or may transform into anhydrates during heating, melting temperature no longer allows conclusions on solid state contributions. the tendency of a compound to exist in hydrated, or more generally solvated, forms might help to understand the combination of a strong solubility limitation beyond lipophilicity, i.e. a high sl value, and a rather low melting temperature. therefore, the polar surface area (psa) was introduced as additional parameter to deal with the possibility of hydrate formation. beyond that, analysis of the dataset comprising about 200 proprietary compounds that were run through the solubility diagnosis revealed that the polar surface area is not only a valuable descriptor to estimate the possibility of hydrate formation, but can also provide guidance to distinguish the type of solubility limitation. a cutoff value for psa around 100å 2 may be used to classify compounds with respect to their type of solubility limitation. data analysis has shown that for compounds with a psa above this value, cases of lipophilicity dominated solubility limitation are rather rare. instead, solubility limitation by crystal lattice forces is dominant for such compounds, at least in the chemical space represented by the investigated dataset. acknowledgements: data generation and solid state characterization was greatly supported by stephane rodde and daniel gosling, and the authors would like to acknowledge their contribution to this publication. references [1] c.a. lipinski, journal of pharmacological and toxicological methods 44 (2000) 235-249 [2] m.j. waring, expert opinion on drug discovery 5 (2010) 235-248 [3] c. saal, a.c. petereit, european journal of pharmaceutical sciences 47 (2012) 589-595 [4] j.g. cumming, a.m. davis, s. muresan, m. haeberlein, h. chen, nature reviews drug discovery 12 (2013) 948-962 [5] c.a.s. bergström, c.m. wassvik, u. norinder, k. luthman, p. artursson, journal of chemical information and computer sciences 44 (2004), 1477-1488 [6] c.m. wassvik, a.g. holmén, r. draheim, p. artursson, c.a.s. bergström, journal of medicinal chemistry 51 (2008) 3035-3039 [7] j. ali, p.camilleri, m.b. brown, a.j. hutt, s.b. kirton, journal of chemical information and modeling 52(2) (2012) 420-428 [8] f. lombardo, m. shalaeva, k. tupper, f. gao, m. abraham, journal of medicinal chemistry 43 (2000), 2922-2928 [9] a. avdeef, k.j. box, j.e.a. comer, m. gilges, m. hadley, c. hibbert, w. patterson and k.y. tam, journal of pharmaceutical and biomedical analysis, 20 (1999), 631-641 [10] a. avdeef, v.m. berger, c. brownell, pharmaceutical research 7 (2000) 85-89 [11] p. ertl, b. rohde, p. selzer, journal of medicinal chemistry 43 (2000) 3714-3717 [12] b. faller, p. ertl, advanced drug delivery reviews 59 (2007) 533-545 admet & dmpk 2(2) (2014) 115-125 diagnosing solubility limitations doi: 10.5599/admet.2.2.36 125 [13] c.m. wassvik, a.g. holmén, c.a.s. bergström, i. zamora, p. artursson, , european journal of pharmaceutical sciences 29 (2006) 294-305 [14] t. suzuki, journal of computer-aided molecular design 5 (1991) 149-166 [15] f. lovering, j. bikker, c. humblet, journal of medicinal chemistry 52 (2009) 6752-6756 [16] m. ishikawa, y. hashimoto, journal of medicinal chemistry 54 (2011) 1539-1554 [17] c.a.s. bergström, c.m. wassvik, k. johansson, i. hubatsch, journal of medicinal chemistry 50 (2007) 5858-5862 [18] l. infantes, l. fabian, w.d.s. motherwell, crystengcomm 9 (2007) 65-71 ©2014 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ the tales of two organic cation transporters, oct-1 and oct-2, in caenorhabditis elegans doi: 10.5599/admet.5.3.394 146 admet & dmpk 5(3) (2017) 146-158; doi: http://dx.doi.org/10.5599/admet.5.3.394 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index review the tales of two organic cation transporters, oct-1 and oct-2, in caenorhabditis elegans dindial ramotar maisonneuve-rosemont hospital research center, and the université de montréal, faculty of medicine, department of medicine, montréal, quebec, canada. *corresponding author: e-mail: dindial.ramotar@umontreal.ca; tel.: (514) 252-3400 ext. 4684; fax: (514) 252-3430 received: may 25, 2017; revised: july 13, 2017; published: september 29, 2017 abstract solute carrier transporters, previously thought to perform roles in the transport of ions and various nutrients are now assigned broader functions. these transporters have recently been shown to permit entry of therapeutic drugs into cells. there is growing interest to understand the broad spectrum of drugs and chemical compounds that are recognized by these transporters such that specific ligands can be used as therapeutics to target definite physiological pathways. to facilitate this investigation, simpler and cost effective model systems are needed, one of such is the live whole model animal caenorhabditis elegans (c. elegans) that offers a multitude of advantages. in general, studies with c. elegans are feasible due to the simplicity of the readouts that include lifespan, brood size, germ cell death, and visualization by epifluorescent microscopy, which can be set up in any laboratory. in c. elegans, two solute carrier transporters, the organic cation transporters oct-1 and its paralogue oct-2 have been partially characterized. oct-1 mutants display a significantly reduced lifespan and brood size, as well as exhibiting an increased susceptibility towards oxidative stress and a subset of dna damaging agents. these multiple phenotypes are directly linked to oct-1 depletion causing upregulation of oct-2, as rnai-mediated downregulation of oct-2 rescues the oct-1 mutant phenotypes. thus, in c. elegans oct-1 exerts control onto oct-2, and this latter transporter plays a predominant role in the uptake of various ligands. we first showed that oct-2 can efficiently mediate uptake of the widely used anticancer drug doxorubicin into the animals, but prevented uptake upon its downregulation. additional ligands of oct-2 including cisplatin and camptothecin were revealed by ligand-docking prediction studies. these analyses generated docking scores indicating that oct-2 can make robust contact with a number of therapeutics and anticancer drugs, as well as chemical compounds that possess the ability to target specific physiological pathways. several of the compounds displaying high docking scores with oct-2 were validated and indeed found to be substrates that oct-2 transported into the animals. this review provides an insight how the transporters oct-1 and oct-2 of a simple model organism c. elegans can be exploited to report on the cytotoxicities and genotoxicities of therapeutic agents, as well as trace amounts of undocumented toxic compounds with neoplastic potentials that are present in the environment. keywords drug uptake and resistance; c. elegans organic cation transporters; germ cells apoptosis; dna damage and repair pathways introduction the mainstay chemotherapeutics such as anthracyclines, cisplatin and radiomimetic drugs are often used to eliminate cancer cells and reduce the burden of tumors by inflicting damages to the genome [1-3]. these genotoxic agents are administered to cancer patients in combination, and in some cases with small http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:dindial.ramotar@umontreal.ca admet & dmpk 5(3) (2017) 146-158 oct-1 and oct-2 in caenorhabditis elegans doi: 10.5599/admet.5.3.394 147 molecule inhibitors to block dna repair pathways with the aim of sensitizing cancer cells towards the chemotherapeutic regimens [4-8]. although these molecules are hydrophilic in nature, it was commonly accepted that they diffuse into cells. this notion has been challenged as most chemotherapeutics are charged molecules and structurally unrelated, and thus cannot easily cross the plasma membrane barrier. moreover, a single chemotherapeutic agent does not show the same genotoxicity when given to different cell types, as would be expected in the case of simple drug diffusion. in fact, there is now growing evidence that these charged anticancer drugs depend on facilitated transport mechanisms to enter cells in order to effectively reach the target sites. in human cells, there are over 450 solute carrier transporters, which are believed to facilitate the uptake of nutrients and ions. however, these transporters also perform functions that are distinct as defects in 30 % of the transporters are linked to diseases that may be related to imbalances in the uptake of nutrients and ions at specific locations. for example, mutations in slc22a4 lead to crohn disease, which affects the gastrointestinal tract and mutations in slc12a3 lead to gitelman’s syndrome characterized by low blood pressure [9, 10]. besides the uptake of nutrients and ions, some of these transporters are actively involved in the influx of drugs for treating various illnesses. at least 12 of the transporters are specifically involved in the transport of drugs approved by the food and drug administration (fda), e.g, slc6a3 transports methylphenidate used for treating psychiatric disorders [11]. in addition, a few of the solute carriers have been shown to mediate uptake of anticancer drugs. for example, the slc35f2 transporter is completely responsible for the uptake into human tumour cells of the anticancer drug ym155, which is currently in clinical evaluation [12]. another recent study established that the organic cation transporter slc22a2 (oct2) is required for oxaliplatin uptake into renal cells and that downregulation of oct2 is a cause for oxaliplatin resistance in renal cell carcinoma [13]. however, these resistant renal cell carcinoma can be sensitized to oxaliplatin upon restoring oct2 expression [13]. in a final example, the functional level of another member of the organic cation transporter family, slc22a1 (oct1), was shown to be a critical marker for the responses of chronic-phase chronic myeloid leukemia towards the tyrosine kinase inhibitor imatinib [14]. patients with low activity levels of oct1 respond poorly to imatinib and exhibit lower overall survival as compared to those with high levels of oct1 [14]. this same transporter was also shown to transport the diabetes drug metformin, which decreases production of hepatic glucose, thus underscoring the possibility that many of the transporters may likely recognize multiple substrates [15]. these foregoing studies raise the possibility that although the transporters may have evolved to mediate uptake of common nutrients and ions, they have the ability to recognize distinct ligands. while it is difficult to determine the ligands for each transporter, it is anticipated that with a given substrate the transporter can be identified through biochemical and genetic analyses. since anthracyclines are widely used chemotherapeutic agents and initially believed to diffuse into cells, there has been little interest to determine whether a transporter exists for these drugs. we challenged the initial view and used anthracyclines as the ligands to search for a transporter. below highlights the key findings that establish the existence of a human transporter oct1, which mediates uptake of anthracyclines into cells and how this discovery was exploited to uncover similar transporters in the model organism caenorhabditis elegans (c. elegans). dindial ramotar admet & dmpk 5(3) (2017) 146-158 148 human oct1 transports anthracyclines anthracyclines such as doxorubicin and daunorubicin act by intercalating with the dna and blocking the function of dna topoisomerase leading to the accumulation of dna lesions and subsequent cell death [16]. monitoring anthracyclines entry into cells was made possible as these drugs autofluoresce at 640 nm, which provides an important physical property for the development of in vitro assays to quantitatively assess the intracellular uptake levels. these in vitro assays monitor drug uptake into the cells as a function of time using quantitative approaches such as fluorescence-activated cell sorting (facs) analysis and fluorometry, as well as a semi-quantitative method involving epifluorescence microscopy [17, 18]. using these assays and a panel of cancer cell lines such as the leukemia hl60 and the ovarian tov222g2 cell lines, it was shown that daunorubicin or doxorubicin at 5 µm enters these cells in a timeand a concentration-dependent manner and reaching saturation within 60 min [19]. these observations establish for the first time that there is an active transport mechanism for anthracyclines into cells, eliminating the possibility that diffusion accounts for the major accumulation of the drug into cells [19]. although earlier studies claimed that the active transport mechanism for doxorubicin uptake involved the human slc22a16 (hct2/oct6) and slc22a4 (octn1) transporters, subsequent findings proved that these were not the transporters [18, 20, 21]. however, it was studies from the budding yeast saccharomyces cerevisiae that provided clues to the actual transporter by directly linking anthracycline-mediated uptake to the family of oct transporters. in human cells, there are three oct homologues oct1, oct2 and oct3 that are members of one family of organic cation transporters all of which can transport choline with varying affinities. we therefore conducted competition experiments with the human cancer cell lines and documented that choline completely blocked the uptake of daunorubicin into the cells when added at nearly equimolar amounts [19]. since oct1 has a higher affinity for choline than oct2 and oct3 and can mediate uptake of several different ligands into cells, it is strongly suggested that it could be a suitable candidate to explore for a role in anthracycline uptake. the first direct evidence came from silencing oct1 expression by shrna in the ovarian cancer cells tov222g2, which sharply reduced the oct1 protein level, as revealed by anti-oct1 antibodies [19]. these oct1 downregulated cells showed nearly 85 % reduction in daunorubicin uptake when compared to the control shrna that showed no effect [19]. as a corollary experiment, overproduction of oct1 which is localized to the plasma membrane, stimulated daunorubicin uptake into the cancer cell lines [19]. collectively, these studies provide the first convincing evidence that oct1 functional level might be pivotal in determining the effectiveness of treatment with anthracyclines. since anthracyclines are used for treating leukemia as well as several other types of cancers such as lymphomas, breast, ovarian, and gastric, it is anticipated that patients who are refractory to anthracycline treatment are likely to be defective in the uptake transporter [22-26]. conversely, patients with induced levels or activated oct1 could be hypersensitive to anthracyclines. because the human oct1 transporter displays broad substrate specificity, it is imperative to investigate whether its function in drug transport can be influenced by the presence of additional drugs, nutrients, and or the status of other influx and efflux transporters. as such, it seems logical that a well characterized oct1 transporter would be beneficial in choosing the appropriate drug, without competition by other substrates, for targeted entry into cells. to rapidly address these issues, it would require a simple live whole model system that is readily tenable to genetic modifications and cost effectiveness such as caenorhabditis elegans. admet & dmpk 5(3) (2017) 146-158 oct-1 and oct-2 in caenorhabditis elegans doi: 10.5599/admet.5.3.394 149 c. elegans as a model to facilitate studies on drug uptake transporters c. elegans has been chosen as the whole model system as it offers a multitude of advantages over mammalian cells to rapidly study many highly conserved biological processes [27]. c. elegans developed from eggs to the adult stage within three days spanning four larval stages. genes can be readily downregulated by simply feeding the animals with bacteria expressing double stranded rna from a segment of the coding region of the targeted gene. c. elegans are transparent and facilitate easy visualization of germ and somatic cells by microscopy. during the last decade, c. elegans have been instrumental in several drug discovery programs that include the search for new antifungals and compounds to limit oxidative stress [27, 28], although the contributions of uptake transporters in these studies were not considered [29]. therefore, it is unclear how c. elegans uptake transporters would have influenced many of the high-throughput screens performed so far to identify novel small molecules that act, for example, as antimicrobials. one possibility is that the yield of bioactive compounds typically recovered from these screens, ranging from 0.03 to 1 %, could be substantially higher if there is greater influx of the molecules into the animal cells via elevated expression of uptake transporters [27, 30]. thus, defining the function and substrate specificities of uptake transporters in c. elegans will be advantageous towards improving the strategies employed to identify novel bioactive molecules. highlighted below are the current progresses regarding the roles of a previously and a recently discovered organic cation transporter, oct-1 and oct-2, respectively, in c. elegans and how these transporters can be exploited for drug testing. uptake transporters in c. elegans in comparison to the human genome, there are far fewer organic cation transporters in c. elegans following an analysis of the sequence information in the wormbase. these c. elegans transporters include the partially characterized oct-1 and oct-2 and at least six other putative transmembrane transporters pes-23, oat-1, svop-1, str-176, hmt-1.1/1.2 and gem-1. far less is known about these latter six putative transporters and will not be discussed further. the c. elegans oct-1 and oct-2 are counterparts of the human octs and little is still known about their roles and affinities towards distinct ligands. both oct-1 and oct-2 share 22.56 % identity at the amino acid level, but oct-2 differs from oct-1 as it harbours an extra 172 amino acid residues at the n-terminal, which has no homology to oct-1. in 1999, oct-1 was the first uptake transporter identified from c. elegans [31]. the oct-1 gene was designed to be expressed in mammalian cells and shown to mediate the transport of the organic cation tetraethylammonium, a prototypical substrate used for classifying octs, suggesting that oct-1 has the ability to function as an uptake transporter [31]. at the time, no extensive study was done to assess the detail substrate specificity of oct-1 following its expression in human cells. it was much later when the next study appeared in 2013 showing that c. elegans deleted for the oct-1 gene exhibited a shortened lifespan, as well as an increased susceptibility to oxidative stress [29]. these observations led to the proposition that oct-1 facilitates the import of antioxidants into the animals, which is required to protect oct-1 mutant animals from oxidative stress [29]. however, the level of the purported substrate of oct-1, the antioxidant ergothioneine, was not altered in oct-1 mutant animals as compared to the parent [29]. therefore, it seems plausible that an alternative explanation could account for the oct-1 mutant animal phenotypes. in an attempt to find substrates for oct-1, we expressed the c. elegans oct-1 gene in another heterologous model system, namely the budding yeast saccharomyces cerevisiae. since human oct1 can dindial ramotar admet & dmpk 5(3) (2017) 146-158 150 transport anthracyclines, we first tested these chemotherapeutic drugs. we observed that expression of c. elegans oct-1 can trigger the uptake of doxorubicin into yeast cells. the yeast strain we used was defective in agp2, which has been reported to be a plasma membrane amino acid transporter that also performs a regulatory role [32]. deletion of the agp2 gene in yeast blocked the expression of several target genes including the polyamine transporters dur3 and sam3 [18]. both dur3 and sam3 can independently transport doxorubicin when overexpressed in yeast cells lacking agp2 [18]. therefore, it is not clear whether the expression of c. elegans oct-1 in yeast substituted for the regulatory function of agp2 or directly for the roles of dur3 and sam3, since these transporters also mediate uptake of doxorubicin [18]. as such, it became necessary to test directly whether oct-1 might mediate the transport of doxorubicin into c. elegans. however, the existence of another related member of the organic cation transporter family, namely oct-2, in c. elegans, suggests that there needs to be a strategy to test the contribution of each transporter towards doxorubicin uptake. prior to our work, there was no function assigned to this putative transporter oct-2 [33]. moreover, there are no animals in the consortiums, usa or japan, carrying either partial or complete deletion of the oct-2 gene, as in the case of oct-1. thus, to examine whether oct-1 or oct-2 or both have roles in the uptake of doxorubicin in c. elegans, we used rna-interference (rnai) feeding bacteria to generate animals with different genotypes, in particular, either downregulated for one or both transporters [33]. using this approach, we uncovered several novel findings regarding the oct-1 and oct-2 transporters in c. elegans. first, unlike the oct-1 null animals, downregulation of oct-2 did not affect the lifespan of the animals. second, perhaps most interesting, the depletion of oct-2 rescued the shortened lifespan of animals already deleted for the oct-1 gene [33]. this striking observation prompted the possibility that oct-1 might influence oct-2 function, whereby oct-1 downregulation leads to oct-2 upregulation. this notion is based on the fact that oct1 knockout mice manifest upregulation of two related transporter genes, oct2 and oct3 [34]. and third, quantitative pcr analysis indeed revealed that downregulation of oct-1 led to the upregulation of oct-2 expression in several tissues of the animals, while several control genes, such as act-1 encoding actin, were unaffected [33]. as such, we proposed that oct-2 upregulation could be responsible for triggering uptake of toxic environmental compounds into the animals and thus account for the shortened lifespan displayed by the oct-1 animals. there is evidence to support that the upregulation of oct-2 leads to the import of toxic compounds from the environment into the animals. this was documented using a simple reporter animal that harbours within its genome the oxidative stress response gene gst-4 encoding glutathione s-transferase 4 fused to the green fluorescent protein (gfp) [35, 36]. in c. elegans skn-1, the counterpart of the mammalian nrf2, senses oxidative stress and turn on the expression of gst-4. we used the gst::gfp reporter animals to monitor oxidative stress caused by toxic compounds and showed that these animals expressed a basal level of gst-4::gfp in the intestine. however, the expression level of the gst-4::gfp was greatly stimulated following rnai downregulation of oct-1 that led to the upregulation of oct-2 [33]. consistent with this observation, downregulation of the oct-2 transporter gene by rnai lowered the basal levels of gst-4::gfp expression in the reporter strain [33]. thus, oct-2 upregulation, via oct-1 downregulation, allows entry of toxic compounds such as prooxidants into c. elegans. these prooxidants increased the oxidative stress of the animals that could cause oxidative damages to various macromolecules, such as dna, and leading to the shortened lifespan of the oct-1 null animals. admet & dmpk 5(3) (2017) 146-158 oct-1 and oct-2 in caenorhabditis elegans doi: 10.5599/admet.5.3.394 151 oct-2 mediates the accumulation of doxorubicin in c. elegans tissues equipped with the new knowledge on the regulation of oct-2 by oct-1, we checked if these transporters possess the ability to drive doxorubicin uptake and accumulation into c. elegans tissues. in the design of the experiment, the fluorescent property of doxorubicin, emitting fluorescence at wavelengths of λ ex 470 nm λ em 585 nm, became a convenient means to monitor its uptake in situ by imaging the pharynx, a very large and visible organ. on the basis of the previous studies in both human and yeast cells, we anticipated that c. elegans oct-1 would act as an uptake transporter and that oct-1 null animals would display very little doxorubicin uptake [18, 19]. unexpectedly, we observed that the c. elegans oct-1 null mutant exposed to doxorubicin depicted a markedly stimulated fluorescence intensity of the drug in the pharynx, as compared to the parent animals [33]. the simplest explanation for this finding is that deletion of the oct-1 gene resulted in the upregulation of oct-2, and it is oct-2 that is responsible for doxorubicin uptake. indeed, the downregulation of oct-2 via rnai feeding bacteria of the oct-1 null mutant completely blocked the uptake of doxorubicin, clearly indicating that oct-2 and not oct-1 has a predominant role in the drug uptake [33]. these observations in c. elegans contradict the previously reported role for oct-1 as an uptake transporter [18, 31]. it is possible that in the natural environment oct-1 may be modified to exert a distinct role other than functioning as an uptake transporter or alternatively exhibits specificity for an entirely different class of substrates. based on the present evidence, we believe that oct-2 might be the key transporter that mediates doxorubicin uptake into c. elegans. oct2 also has the ability to recognize and transport other cationic compounds. for example, choline can compete for the uptake of doxorubicin entry into c. elegans, suggesting that it is likely a substrate for oct2 [33, 37]. in fact, choline is a substrate for human oct1 and it impedes the uptake of daunorubicin into human cell lines [19]. it is reasonable to propose that any compound that competes for doxorubicin uptake into the c. elegans oct-1 null animals is likely a putative substrate for oct-2, thus providing a foundation to develop a screening tool to find therapeutic compounds that are transported by oct-2. in developing this assay further, it is noteworthy that other factors should be considered such as compounds, e.g., spermine that could quench the fluorescence of doxorubicin leading to false positives. nonetheless, the same assay can be extended to find compounds that could stimulate doxorubicin uptake revealing activators for oct-2. oct-1 and oct-2 play a pivotal role in controlling germ cell death in a manner similar to many stem cell systems, c. elegans has a self-renewing germ cell population [38]. these germ cells are very sensitive to genotoxic compounds and respond by using conserved dna repair mechanisms to faithfully maintain genomic stability [39]. germ cells with extensive dna damage undergo apoptosis and cannot form viable embryos, and which can be quantified in vivo by examining for corpses in the proximal zone of the gonad arm [38, 40, 41]. these corpses are easily seen by staining with the dna dye acridine orange [38, 40, 41]. thus, germ cell apoptosis is a suitable in vivo reporter system to determine whether natural compounds and those in the growth environment of the animals are genotoxic. we took advantage of this reporter system to assess whether oct-1 and oct-2 control the extent of germ cell death triggered by genotoxic agents, in particular, those compounds that cannot be tracked directly for their uptake kinetics due to the absence of a fluorescence signal or the availability of radioactively labeled form. in the wild type animals grown under normal conditions, there are typically two to four germ cell corpses [42]. however, when oct-1 is downregulated the animals depicted an average of five to eight apoptotic cells, perhaps resulting from an increase import of genotoxic prooxidants by the elevated levels of the oct-2 transporter leading to dead embryos [33]. interestingly, downregulation of oct-2 in the oct-1 null mutant reduced the germ cell death, and in some animals to undetectable level. it seems that oct-2 dindial ramotar admet & dmpk 5(3) (2017) 146-158 152 possesses the ability to recognize and transport toxic compounds from the growth environment into the animals which then induced germ cell death. at least, these toxic compounds can turn on the oxidative stress reporter gst-4::gfp, suggesting that they compose of prooxidants with the ability to cause genotoxicity in various tissues [33]. in fact, the oct-1 deficient animals produce an elevated level of dead embryos, which are believed to be the result of the accumulation of genotoxic dna lesions under normal growth conditions, and therefore account for the reduced lifespan observed in these animals. since doxorubicin acts by damaging the dna and previously shown to trigger germ cell apoptosis in c. elegans [43], then modulating oct-2 level should determine the extent of doxorubicin-induced germ cell death. indeed, oct-1 deletion animals showed substantial levels of germ cell deaths when treated with doxorubicin, as opposed to the wild-type. this phenomenon was greatly diminished following oct-2 downregulation in the doxorubicin treated oct-1 null animals. these observations are consistent with the model that oct-1 deletion led to oct-2 upregulation, which then mediates doxorubicin uptake that induces germ cell death. therefore, germ cell death is a genuine reporter for oct-2 dependent uptake of genotoxic agents. potential substrates of oct-2 while it is clear that additional substrates will be recognized by oct-2, it is an arduous task to test each potential substrate, for example, by competing for doxorubicin uptake. as such, we undertook an in silico approach by first making predictions of the structure of the transporter using the i-tasser protein structure prediction server [44]. the analysis predicted 3d structures not only for oct-2, but also oct-1, revealing that both possessed the entire 12 transmembrane domain helices and belonging to the solute carrier transporter family [33, 45]. the predicted 3d structures of the transporters were then modeled with drugs to look for interactions using the bsp-slim and coach algorithms. these programs approximate the amino acid residues of the transporters constituting the ligand–protein docking sites and thereby resulting in a docking score [46, 47]. a number of drugs were analyzed and these were mostly dna damaging agents including cisplatin and camptothecin and small molecules with the ability to inhibit dna repair such as olaparib and methoxyamine. of the 19 potential substrates, four resulted with a docking score of zero, whereas the other 15 revealed docking scores > 3.5 [33]. many showed higher docking scores with oct-2 than oct-1, favoring robust binding to oct-2 [33]. from the molecules tested, it seems that oct-2 may distinguish specific substrates to be transported into c. elegans tissues by recognizing an overall positive charge and perhaps structural features that include aromatic groups. it is noteworthy that there are compounds that can damage the dna and induce apoptotic corpses such as melphalan and methoxyamine. however, these agents do not dock onto oct-2 or oct-1, and may therefore require alternative transporters to enter the animals [33]. it is anticipated that protein-ligand modeling studies should reveal novel substrates for these uptake transporters, but these ligands must be experimentally validated as described below for the anticancer drug cisplatin and an inhibitor of dna repair. oct-2-dependent transport of cisplatin stimulates germ cell death in c. elegans mutants defective in dna repair cisplatin, a predicted ligand of oct-2, is a member of the platinum family of anticancer drugs that damages the dna by creating intraand inter-strand dna cross-links [48]. these dna lesions are repaired by the nucleotide excision dna repair pathway, which has the ability to recognize helix distortion followed admet & dmpk 5(3) (2017) 146-158 oct-1 and oct-2 in caenorhabditis elegans doi: 10.5599/admet.5.3.394 153 by incisions on either side of the lesion [49]. thus, defects in the nucleotide excision repair pathway cause hypersensitivity to cisplatin [49]. as cisplatin is not available in labeled forms, but can be monitored by liquid chromatography mass spectrometry, the germ cell apoptotic assay provides a remarkable alternative reporter for the uptake of the drug [39, 50]. the analysis revealed that cisplatin induced an increase level of germ cell death in the wild type animal, and which was further stimulated in the oct-1 null mutants [33]. downregulation of oct-2 in the oct-1 null mutants prevented the cisplatin-induced germ cell death indicating that oct-2 plays a central role in mediating the genotoxic effects of cisplatin [33]. thus, the ligand-docking studies described above serve as a useful prediction platform to screen for oct-2 substrates. one prediction from these findings is that the elevated uptake of genotoxic agents into animals defective in dna repair should result in lethal phenotypes. there are several mutants in the c. elegans consortium carrying defects in the major dna repair pathways and which are suitable to test whether oct2 upregulation can trigger sensitization of the animals to dna damage-induced germ cell death. for example, the xpa-1 deletion mutant lacking the xpa-1 protein needed for nucleotide excision repair of bulky dna lesions exhibited enhance germ cell death when exposed to cisplatin following downregulation of oct-1. in contrast, depletion of oct-2 by rnai in the xpa-1 mutant animals suppressed the high level of apoptotic cells observed in this mutant upon exposure to cisplatin. these data indicate that upregulation of oct-2 burdens the xpa-1 mutant animals with cisplatin-induced dna lesions leading to the enhancement of germ cell death [51] [52]. thus, the current studies suggest that by combining defects of various dna repair pathways with functional organic cation transporters such as oct-2, it is possible to determine whether an unknown compound has genotoxic effects and the type of lesions it may create. oct-2 transports the rad51 inhibitor b02 in addition to the genotoxic agents that docked onto oct-2, the analyses revealed that many chemical compounds such as 3-benzyl-2-[(e)-2-pyridin-3-ylethenyl]quinazolin-4-one (b02, cid: 5738263), referred to as b02, has the ability to robustly interact with oct-2, and not oct-1. in human cells, b02 has been shown to interfere with the dna strand exchange and the nuclear foci formation catalyzed by the dna repair protein rad51. rad51 is required to repair dna double strand breaks during meiosis and in response to dna damaging agents [53-55]. examination of the pharmacological effect of b02 in c. elegans revealed that this inhibitor caused sterility in the oct-1 null mutant. this is marked by the production of inviable embryos and a decrease in the number of viable animals, phenotypes that mimic those of the rad-51 homozygotes [56]. as predicted from the model, downregulation of oct-2 in the oct-1 null mutant prevented the b02 ligand from causing sterility [33]. under these conditions, the oct-1 null mutant produced viable embryos that resulted in nearly normal levels of viable animals [33]. from these data, it is clear that b02 can also inhibit rad-51 function in c. elegans and this depends on the function of the oct-2 uptake transporter [33]. thus, the docking scores of ligands and proper readouts from the targeted pathway provide valuable tools to monitor transporter-mediated drug uptake into c. elegans. perspectives there is no evidence yet that oct-1 can function as an uptake transporter in c. elegans, although when expressed in two different heterologous model system, human and yeast cells, it can stimulate uptake of tetraethylammonia and doxorubicin, respectively [18, 31]. while in human cells downregulation of oct1 caused resistance to anthracyclines, this was not the case in c. elegans devoid of oct-1, thus a direct role for oct-1 in the uptake of anthracyclines can be excluded. whether oct-1 acts as an uptake transporter dindial ramotar admet & dmpk 5(3) (2017) 146-158 154 for selective drugs remains a possibility. nonetheless, in c. elegans oct-1 has a role in exerting control on oct-2 expression, and it is oct-2 that is primarily involved in the uptake of doxorubicin and other substrates. we believe that oct-2 has a broad specificity and can transport various substrates, including toxic environmental compounds, from the pharynx into the intestine (fig. 1). the oct-2 in the intestine then transports the substrates into the vicinity of the oocytes and embryos, which are also equipped with the oct-2 transporter (fig. 1). thus, removal of oct-2 stands to protect the animal from germ cell death induced by environmental and chemical toxicants. the data obtained so far underscore the importance of uptake transporters in regulating the entry of drugs into cells and raise the possibility that (i) the drugresistance and -sensitivity exhibited by cancer patients could be governed at the level of the uptake transporters and (ii) the absence of one transporter could greatly impact the function of another. figure 1. illustration of key features of an adult c. elegans hermaphrodite. the blue squiggly line represents the putative location of oct-1 transporter, while the red squiggly line and short vertical red bar represent the putative locations of oct-2 transporter on the pharynx, intestine, oocytes and embryos. the studies to date also raised additional concern and that is whether the upregulated oct-2 transporter can import compounds from the growth media that cause dna damage and ultimately leading to genomic instability over time. if this were the case, then individuals with deregulated uptake transporters are likely at increased risk of being bombarded with genotoxic compounds that could lead to various diseases such as cancers. in fact, preliminary data revealed that the upregulation of oct-2, via rnai downregulation of oct-1, resulted in at least a two-fold increase in spontaneous mutations in a c. elegans strain designed to monitor frame shift mutations at the lacz reporter gene [57]. the downregulation of oct-1 leading to the upregulation of oct-2 was an unexpected finding, and provides a compelling argument that the animal has evolved tight regulation of oct-2 presumably to control uptake of unwanted compounds. oct-1 could perform a function as a surface sensor that acts as non-transporting transceptors by sensing the availability of nutrients and signaling the regulation of downstream plasma membrane transporters [32, 58]. thus, when nutrients are scarce, oct-1 might serve as a sensor to promote the upregulation of oct-2 to scavenge the limited resources, but sustain basal expression of oct-2 when nutrients are in surplus. this mode of regulation has been characterized in s. cerevisiae, drosophila melanogaster and homo sapiens [59-61]. thus, in view of the increasing number of sensors that are currently being identified in various organisms, it is plausible that c. elegans oct-1 may indeed act either as a non-transporting or transporting sensor leading to the regulation of oct-2 expression. if such a model is possible, other factors including kinases and transcriptional activators would be involved in conveying the signal from oct-1 to promote oct-2 expression as shown for the s. cerevisiae ssy1 amino acid sensing pathway [62]. we propose that exploiting oct-2 in c. elegans could have far reaching applications that supersede admet & dmpk 5(3) (2017) 146-158 oct-1 and oct-2 in caenorhabditis elegans doi: 10.5599/admet.5.3.394 155 other whole model systems in drug discovery programs with respect to cost and time. thus, maintaining the oct-2 transporter at optimal levels should be a useful step to incorporate into any high-throughput screens to improve the efficiency of identifying cationic bioactive molecules from chemical libraries. the overexpressed oct-2 is expected to significantly alter its kinetics such that it can transport chemicals that are present at lower concentrations in the libraries. we strongly suggest that the previous barriers posed by c. elegans to find bioactive molecules could be explained by the lack of an activated mechanism to efficiently take up compounds at lower concentrations. acknowledgments special thanks to taramatti harihar for her editorial comments. this work was funded by the research project grant 201610pjt (377317) to d.r. from the canadian institute of health research. references [1] a.c. begg, f.a. stewart, c. vens. strategies to improve radiotherapy with targeted drugs, nat rev cancer 11 (2011) 239-253. [2] c.j. lord, a. ashworth. the dna damage response and cancer therapy, nature 481 (2012) 287-294. [3] k. cheung-ong, g. giaever, c. nislow. dna-damaging agents in cancer chemotherapy: serendipity and chemical biology, chem biol 20 (2013) 648-659. [4] m.r. kelley, d. logsdon, m.l. fishel. targeting dna repair pathways for cancer treatment: what's new?, future oncol 10 (2014) 1215-1237. [5] w.g. kaelin, jr. synthetic lethality: a framework for the development of wiser cancer therapeutics, genome med 1 (2009) 99. [6] a. bapat, m.l. fishel, m.r. kelley. going ape as an approach to cancer therapeutics, antioxid redox signal 11 (2009) 651-668. [7] s.p. jackson, j. bartek. the dna-damage response in human biology and disease, nature 461 (2009) 1071-1078. [8] a. ciccia, s.j. elledge. the dna damage response: making it safe to play with knives, mol cell 40 (2010) 179-204. [9] v.d. peltekova, r.f. wintle, l.a. rubin, c.i. amos, q. huang, x. gu, b. newman, m. van oene, d. cescon, g. greenberg, a.m. griffiths, p.h. st george-hyslop, k.a. siminovitch. functional variants of octn cation transporter genes are associated with crohn disease, nat genet 36 (2004) 471-475. [10] s.t.d. consortium, a.l. williams, s.b. jacobs, h. moreno-macias, a. huerta-chagoya, c. churchhouse, c. marquez-luna, h. garcia-ortiz, m.j. gomez-vazquez, n.p. burtt, c.a. aguilarsalinas, c. gonzalez-villalpando, j.c. florez, l. orozco, c.a. haiman, t. tusie-luna, d. altshuler. sequence variants in slc16a11 are a common risk factor for type 2 diabetes in mexico, nature 506 (2014) 97-101. [11] a. cesar-razquin, b. snijder, t. frappier-brinton, r. isserlin, g. gyimesi, x. bai, r.a. reithmeier, d. hepworth, m.a. hediger, a.m. edwards, g. superti-furga. a call for systematic research on solute carriers, cell 162 (2015) 478-487. [12] g.e. winter, b. radic, c. mayor-ruiz, v.a. blomen, c. trefzer, r.k. kandasamy, k.v. huber, m. gridling, d. chen, t. klampfl, r. kralovics, s. kubicek, o. fernandez-capetillo, t.r. brummelkamp, g. superti-furga. the solute carrier slc35f2 enables ym155-mediated dna damage toxicity, nat chem biol 10 (2014) 768-773. [13] y. liu, x. zheng, q. yu, h. wang, f. tan, q. zhu, l. yuan, h. jiang, l. yu, s. zeng. epigenetic activation of the drug transporter oct2 sensitizes renal cell carcinoma to oxaliplatin, sci transl med 8 (2016) 348ra397. dindial ramotar admet & dmpk 5(3) (2017) 146-158 156 [14] d.l. white, v.a. saunders, p. dang, j. engler, t.p. hughes. oct-1 activity measurement provides a superior imatinib response predictor than screening for single-nucleotide polymorphisms of oct-1, leukemia 24 (2010) 1962-1965. [15] m.l. becker, l.e. visser, r.h. van schaik, a. hofman, a.g. uitterlinden, b.h. stricker. genetic variation in the organic cation transporter 1 is associated with metformin response in patients with diabetes mellitus, pharmacogenomics j 9 (2009) 242-247. [16] t.j. westmoreland, s.m. wickramasekara, a.y. guo, a.l. selim, t.s. winsor, a.l. greenleaf, k.l. blackwell, j.a. olson, jr., j.r. marks, c.b. bennett. comparative genome-wide screening identifies a conserved doxorubicin repair network that is diploid specific in saccharomyces cerevisiae, plos one 4 (2009) e5830. [17] k.k. karukstis, e.h. thompson, j.a. whiles, r.j. rosenfeld. deciphering the fluorescence signature of daunomycin and doxorubicin, biophys chem 73 (1998) 249-263. [18] n. brosseau, e. andreev, d. ramotar. complementation of the yeast model system reveals that caenorhabditis elegans oct-1 is a functional transporter of anthracyclines, plos one 10 (2015) e0133182. [19] e. andreev, n. brosseau, e. carmona, a.m. mes-masson, d. ramotar. the human organic cation transporter oct1 mediates high affinity uptake of the anticancer drug daunorubicin, sci rep 6 (2016) 20508. [20] m. okabe, m. unno, h. harigae, m. kaku, y. okitsu, t. sasaki, t. mizoi, k. shiiba, h. takanaga, t. terasaki, s. matsuno, i. sasaki, s. ito, t. abe. characterization of the organic cation transporter slc22a16: a doxorubicin importer, biochem biophys res commun 333 (2005) 754-762. [21] m. okabe, g. szakacs, m.a. reimers, t. suzuki, m.d. hall, t. abe, j.n. weinstein, m.m. gottesman. profiling slco and slc22 genes in the nci-60 cancer cell lines to identify drug uptake transporters, mol cancer ther 7 (2008) 3081-3091. [22] j.w. lown. anthracycline and anthraquinone anticancer agents: current status and recent developments, pharmacol ther 60 (1993) 185-214. [23] h. dohner, e.h. estey, s. amadori, f.r. appelbaum, t. buchner, a.k. burnett, h. dombret, p. fenaux, d. grimwade, r.a. larson, f. lo-coco, t. naoe, d. niederwieser, g.j. ossenkoppele, m.a. sanz, j. sierra, m.s. tallman, b. lowenberg, c.d. bloomfield. diagnosis and management of acute myeloid leukemia in adults: recommendations from an international expert panel, on behalf of the european leukemianet, blood 115 (2010) 453-474. [24] t.r. chauncey. drug resistance mechanisms in acute leukemia, curr opin oncol 13 (2001) 21-26. [25] j.p. marie. drug resistance in hematologic malignancies, curr opin oncol 13 (2001) 463-469. [26] j.c. byrd, k. mrozek, r.k. dodge, a.j. carroll, c.g. edwards, d.c. arthur, m.j. pettenati, s.r. patil, k.w. rao, m.s. watson, p.r. koduru, j.o. moore, r.m. stone, r.j. mayer, e.j. feldman, f.r. davey, c.a. schiffer, r.a. larson, c.d. bloomfield. pretreatment cytogenetic abnormalities are predictive of induction success, cumulative incidence of relapse, and overall survival in adult patients with de novo acute myeloid leukemia: results from cancer and leukemia group b (calgb 8461), blood 100 (2002) 4325-4336. [27] l.p. o'reilly, c.j. luke, d.h. perlmutter, g.a. silverman, s.c. pak. c. elegans in high-throughput drug discovery, adv drug deliv rev 69-70 (2014) 247-253. [28] t. kaletta, m.o. hengartner. finding function in novel targets: c. elegans as a model organism, nat rev drug discov 5 (2006) 387-398. [29] i.k. cheah, r.l. ong, j. gruber, t.s. yew, l.f. ng, c.b. chen, b. halliwell. knockout of a putative ergothioneine transporter in caenorhabditis elegans decreases lifespan and increases susceptibility to oxidative damage, free radic res 47 (2013) 1036-1045. [30] a.r. burns, i.m. wallace, j. wildenhain, m. tyers, g. giaever, g.d. bader, c. nislow, s.r. cutler, p.j. roy. a predictive model for drug bioaccumulation and bioactivity in caenorhabditis elegans, nat chem biol 6 (2010) 549-557. admet & dmpk 5(3) (2017) 146-158 oct-1 and oct-2 in caenorhabditis elegans doi: 10.5599/admet.5.3.394 157 [31] x. wu, y.j. fei, w. huang, c. chancy, f.h. leibach, v. ganapathy. identity of the f52f12.1 gene product in caenorhabditis elegans as an organic cation transporter, biochim biophys acta 1418 (1999) 239-244. [32] m. aouida, m. rubio-texeira, j.m. thevelein, r. poulin, d. ramotar. agp2, a member of the yeast amino acid permease family, positively regulates polyamine transport at the transcriptional level, plos one 8 (2013) e65717. [33] a. papaluca, d. ramotar. a novel approach using c. elegans dna damage-induced apoptosis to characterize the dynamics of uptake transporters for therapeutic drug discoveries, sci rep 6 (2016) 36026. [34] l. chen, y. shu, x. liang, e.c. chen, s.w. yee, a.a. zur, s. li, l. xu, k.r. keshari, m.j. lin, h.c. chien, y. zhang, k.m. morrissey, j. liu, j. ostrem, n.s. younger, j. kurhanewicz, k.m. shokat, k. ashrafi, k.m. giacomini. oct1 is a high-capacity thiamine transporter that regulates hepatic steatosis and is a target of metformin, proc natl acad sci u s a 111 (2014) 9983-9988. [35] m.l. fishel, x. wu, c.m. devlin, d.p. logsdon, y. jiang, m. luo, y. he, z. yu, y. tong, k.p. lipking, a. maitra, n.v. rajeshkumar, g. scandura, m.r. kelley, m. ivan. apurinic/apyrimidinic endonuclease/redox factor-1 (ape1/ref-1) redox function negatively regulates nrf2, j biol chem 290 (2015) 3057-3068. [36] f. cabreiro, c. au, k.y. leung, n. vergara-irigaray, h.m. cocheme, t. noori, d. weinkove, e. schuster, n.d. greene, d. gems. metformin retards aging in c. elegans by altering microbial folate and methionine metabolism, cell 153 (2013) 228-239. [37] k. hoffmann, f. grafe, w. wohlrab, r.h. neubert, m. brandsch. functional characterization of a high-affinity choline transport system in human keratinocytes, j invest dermatol 119 (2002) 118121. [38] a. gartner, p.r. boag, t.k. blackwell. germline survival and apoptosis, wormbook (2008) 1-20. [39] h. lans, w. vermeulen. tissue specific response to dna damage: c. elegans as role model, dna repair (amst) 32 (2015) 141-148. [40] h.r. horvitz. genetic control of programmed cell death in the nematode caenorhabditis elegans, cancer res 59 (1999) 1701s-1706s. [41] a.l. craig, s.c. moser, a.p. bailly, a. gartner. methods for studying the dna damage response in the caenorhabdatis elegans germ line, methods cell biol 107 (2012) 321-352. [42] t.l. gumienny, e. lambie, e. hartwieg, h.r. horvitz, m.o. hengartner. genetic control of programmed cell death in the caenorhabditis elegans hermaphrodite germline, development 126 (1999) 1011-1022. [43] a. sendoel, s. maida, x. zheng, y. teo, l. stergiou, c.a. rossi, d. subasic, s.m. pinto, j.m. kinchen, m. shi, s. boettcher, j.n. meyer, m.g. manz, d. bano, m.o. hengartner. depdc1/let-99 participates in an evolutionarily conserved pathway for anti-tubulin drug-induced apoptosis, nat cell biol 16 (2014) 812-820. [44] a. roy, a. kucukural, y. zhang. i-tasser: a unified platform for automated protein structure and function prediction, nat protoc 5 (2010) 725-738. [45] l. sun, x. zeng, c. yan, x. sun, x. gong, y. rao, n. yan. crystal structure of a bacterial homologue of glucose transporters glut1-4, nature 490 (2012) 361-366. [46] h.s. lee, y. zhang. bsp-slim: a blind low-resolution ligand-protein docking approach using predicted protein structures, proteins 80 (2012) 93-110. [47] j. yang, a. roy, y. zhang. protein-ligand binding site recognition using complementary bindingspecific substructure comparison and sequence profile alignment, bioinformatics 29 (2013) 25882595. [48] l. lin, s.w. yee, r.b. kim, k.m. giacomini. slc transporters as therapeutic targets: emerging opportunities, nat rev drug discov 14 (2015) 543-560. dindial ramotar admet & dmpk 5(3) (2017) 146-158 158 [49] e.c. friedberg, g.c. walker, w. siede, r.d. wood, r.a. schultz, t. ellenberger, dna repair and mutagenesis, asm press, washington, d.c., 2006. [50] a.n. shaik, d.a. altomare, l.j. lesko, m.n. trame. development and validation of a lc-ms/ms assay for quantification of cisplatin in rat plasma and urine, j chromatogr b analyt technol biomed life sci 1046 (2017) 243-249. [51] x. yang, j. fan, a.a. ishchenko, d. patel, m.k. saparbaev, d. ramotar. functional characterization of the caenorhabditis elegans dna repair enzyme apn-1, dna repair (amst) 11 (2012) 811-822. [52] g. minotti, p. menna, e. salvatorelli, g. cairo, l. gianni. anthracyclines: molecular advances and pharmacologic developments in antitumor activity and cardiotoxicity, pharmacol rev 56 (2004) 185-229. [53] f. huang, a.v. mazin. a small molecule inhibitor of human rad51 potentiates breast cancer cell killing by therapeutic agents in mouse xenografts, plos one 9 (2014) e100993. [54] f. huang, a.v. mazin. targeting the homologous recombination pathway by small molecule modulators, bioorg med chem lett 24 (2014) 3006-3013. [55] d.a. alagpulinsa, s. ayyadevara, r.j. shmookler reis. a small-molecule inhibitor of rad51 reduces homologous recombination and sensitizes multiple myeloma cells to doxorubicin, front oncol 4 (2014) 289. [56] a. alpi, p. pasierbek, a. gartner, j. loidl. genetic and cytological characterization of the recombination protein rad-51 in caenorhabditis elegans, chromosoma 112 (2003) 6-16. [57] c. zakaria, h. kassahun, x. yang, j.c. labbe, h. nilsen, d. ramotar. caenorhabditis elegans apn-1 plays a vital role in maintaining genome stability, dna repair (amst) 9 (2010) 169-176. [58] y. popova, p. thayumanavan, e. lonati, m. agrochao, j.m. thevelein. transport and signaling through the phosphate-binding site of the yeast pho84 phosphate transceptor, proc natl acad sci u s a 107 (2010) 2890-2895. [59] j. schothorst, h.n. kankipati, m. conrad, d.r. samyn, g. van zeebroeck, y. popova, m. rubiotexeira, b.l. persson, j.m. thevelein. yeast nutrient transceptors provide novel insight in the functionality of membrane transporters, curr genet 59 (2013) 197-206. [60] r.f. gaber, k. ottow, h.a. andersen, m.c. kielland-brandt. constitutive and hyperresponsive signaling by mutant forms of saccharomyces cerevisiae amino acid sensor ssy1, eukaryot cell 2 (2003) 922-929. [61] l. bianchi, a. diez-sampedro. a single amino acid change converts the sugar sensor sglt3 into a sugar transporter, plos one 5 (2010) e10241. [62] s. tumusiime, c. zhang, m.s. overstreet, z. liu. differential regulation of transcription factors stp1 and stp2 in the ssy1-ptr3-ssy5 amino acid sensing pathway, j biol chem 286 (2011) 4620-4631. ©2017 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ identify the radiotherapy-induced abnormal changes in the patients with nasopharyngeal carcinoma doi: 10.5599/admet.5.4.484 234 admet & dmpk 5(4) (2017) 234-241; doi: http://dx.doi.org/10.5599/admet.5.4.484 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper identify the radiotherapy-induced abnormal changes in the patients with nasopharyngeal carcinoma xiuxin wang 1+ , yi zhao 1+ , huiling zhang 1 , gang yin 2 , yangkun luo 2 , zixuan fan 2 , yin tian 1 * 1 bio-information college, chong qing university of posts and telecommunications, chong qing, china 2 department of radiotherapy, sichuan cancer hospital & institute, chengdu, china *corresponding author: e-mail: tiany20032003@163.com; tel.: +86-23-62460536; fax: +86-23-62460536 + the authors have equal contribution to this work received: december 02, 2017; revised: december 12, 2017; published: december 24, 2017 abstract radiotherapy (rt) is the standard treatment for nasopharyngeal carcinoma, which often causes inevitable brain injury in the process of treatment. the majority of patients has no abnormal signal or density change of the conventional magnetic resonance imaging (mri) and computed tomography (ct) examination in the long-term follow-up after radiation therapy. however, when there is a visible ct and conventional mr imaging changes, the damage often has been severe and lack of effective treatments, seriously influencing the prognosis of patients. therefore, the present study aimed to investigate the abnormal changes in nasopharyngeal carcinoma (npc) patients after rt. in the present study, we exploited the machine learning framework which contained two parts: feature extraction and classification to automatically detect the brain injury. our results showed that the method could effectively identify the abnormal regions reduced by radiotherapy. the highest classification accuracy was 82.5 % in the abnormal brain regions. the parahippocampal gyrus was the highest accuracy region, which suggested that the parahippocampal gyrus could be most sensitive to radiotherapy and involved in the pathogenesis of radiotherapy-induced brain injury in npc patients. keywords nasopharyngeal carcinoma (npc); radiotherapy (rt); injury; classification; structural mri (smri); introduction nasopharyngeal carcinoma (npc) is an endemic disease and rare in most regions of the world. however, it exists much more frequently in southeast asia and china. in southern china, npc is one of the most common malignant tumor and the incidence of npc is approximately 30–80 per 100,000 per year [1]. because of the special lesion location, nasopharyngeal carcinoma is not suitable for surgical treatment and radiotherapy plays a crucial part in npc therapy. however, brain injury caused by radiation therapy is a serious complication, which has a serious impact on the prognosis and quality of life of the patients. thus, it is necessary to understand the effect of radiation therapy on brain structure and find neural biomarkers to facilitate clinical diagnose, treatment, and prevention [2]. previous studies have found that radiation therapy for npc resulted in multiple regions abnormalities, http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:tiany20032003@163.com javascript:void(0); javascript:void(0); javascript:void(0); javascript:void(0); admet & dmpk 5(4) (2017) 234-241 radiotherapy-induced changes in patients with nasopharyngeal carcinoma doi: 10.5599/admet.5.4.484 235 such as temporal lobe necrosis [3, 4], precuneus, cuneus, lateral occipital cortex [5], the vermis, hippocampus, cerebellum lobule, middle occipital lobe, and insula [6, 7]. however, it was difficult for doctors to identify the abnormal changes after radiotherapy (rt) based on original image features. there was a lack of automatic classification method for npc patients before and after rt based on the big data of smri. therefore, we hoped that the machine learning framework with smri could improve the present situation on brain radiation therapy. materials and methods participants the total participants included 103 right-handed npc patients (80 males and 23 females, mean ±sd= 47.5 ± 10.8 years, age from 23 to 82 years). all the patients were recruited from the sichuan cancer hospital in chengdu, china, and they all had no other disease. each subject performed two mri scans, i.e. before (pre-rt) and after radiotherapy (post-rt). the research was approved by the ethics committee of sichuan cancer hospital & institute, chengdu, china. written informed consent was obtained from all subjects. mri acquisition the smri images of the whole brain were obtained by the rapid acquisition of t1 weighted sequences with a high-resolution three-dimensional magnetic spin echo using a siemens 1.5 mr scanner (tr = 750 ms, te = 11 ms, fov = 75x75 cm, flip angle = 150°, matrix size = 256x192, voxel size = 1x1x1 mm 3 , slice thickness = 3 mm, layer space = 3.6 mm, the continuous acquisition = 36 layers covering the whole brain. the voxel-based morphometry (vbm) toolbox of the statistical parametric mapping software (spm8) was used to perform the data-preprocessing. firstly, the structured images were registered into the standard mni space according to 12 radiation parameters. secondly, the spatially normalized images were segmented into gray matter, white matter and cerebrospinal fluid using a priori tissue probability maps (tpm) provided by international consortium for brain mapping (icbm). thirdly, the segmented gray images were further smoothed with a kernel of 8 mm to remove the influence of noise and increase the validity of statistical tests of the posterior parameters. methods the methods mainly included two parts: 1) feature extraction and feature selection with the principal component analysis (pca); 2) classification with support vector machine (svm). details were shown in figure1. a 3×3×3 cube was defined and vectorized. leave-one-out cross-validation (loocv) was used to evaluate the general classification performance. image data were divided into two sessions: one was a test session whose size were 1×27 was leaving out by turns in each round of loocv, and the other was a training session whose size were 205×27. then, pca was used to extract the features of the training session. liblinear was an open source library for large-scale linear classification that inherits many features of the popular svm library libsvm. in this paper, we used the liblinear toolbox [8] to perform the classification task. xiuxin wang et al. admet & dmpk 5(4) (2017) 234-241 236 figure1. the flow chart of classification. vi: voxel i; n was the number of patients; pca: principal component analysis. results and discussion as shown in figure 2, the final accuracy for distinguishing the pre-rt and post-rt was 82.5 %. the interested brain regions should be met: 1) the clusters must contain more than 50 continuous voxels. 2) the accuracy of all the voxels in the cluster must be greater than 70 %. the final classification accuracy was the greatest accuracy of the voxels in all the interested brain clusters, and the regions in the interested brain clusters were the most distinguishable brain regions. in this study, we identified some interested brain regions which mainly included the bilateral parahippocamps, the bilateral middle temporal gyrus, the right middle occipital gyrus, the left rolandic operculum, the bilateral cerebellum, the vermis, the right hippocampus, the right thalamus, the insula, the right precuneus, and the left cuneus. details were shown in the table 1. in this study, we used the machine learning framework based on feature extraction and classification to achieve the automatic classification between the pre-rt and post-rt. the method of pca+svm was quite helpful to improve the accuracy of the classification, and validly detected the abnormal brain regions, which widely contained the previous research achievement. as shown in table 1, our results were consistent with previous studies, suggesting that the method was in favour of identifying the injury caused by rt [5-7, 9]. the classification accuracy was 82.5 %, indicating that the method could well identify the abnormal changes induced by rt. the abnormal brain regions can be considered as biomarkers to detect the radiation-induced injury and help doctors diagnose and monitor the disease. some researchers revealed that the parahippocampal cortex and hippocampus were involved in memory [10-13]. the parahippocampal cortex played a role for spatial configuration of object in memory [14]. hippocampus was associated with attention processes and visuospatial working memory [15]. the temporal lobe necrosis was the most debilitating late-stage complication[16]. cui et al [17] found that the temporal lobe played critical roles in auditory and language processing, which might be associated with auditory verbal hallucinations (avhs) for schizophrenia patients. temporal regions might also be admet & dmpk 5(4) (2017) 234-241 radiotherapy-induced changes in patients with nasopharyngeal carcinoma doi: 10.5599/admet.5.4.484 237 figure 2. the classification accuracies of npc patients before rt and after rt based on the pca and svm table 1. the most distinguish force of gray matter volume brain regions for the pre-rt and post-rt groups peak coordinates clusters voxels x y z accuracy of the peak parahippocampal_l 575 -16.5 -19.5 -24 82.5 % temporal_mid_r 186 48 -54 13.5 80 % occipital_mid_r 996 28.5 -78 10.5 80 % rolandic_oper_l 496 -57 -6 9 77.5 % cerebellum_9_l 177 -12 -45 -51 77.5 % cerebellum_9_r 129 7.5 -51 -40.5 75 % vermis_1_2 716 1.5 -40.5 -22.5 80 % hippocampus_r 183 19.5 -10.5 -9 80 % parahippocampal_r 60 18 -24 -15 75 % thalamus_r 137 4.5 -13.5 4.5 75 % temporal_mid_l 480 -45 -70.5 18 77.5 % insula_r 723 30 -28.5 12 77.5 % precuneus_r 495 9 -51 13.5 77.5 % cuneus_l 61 -13.5 -57 18 75 % ps. parahippocampal_l: left parahippocampal gyrus; temporal_mid_r: right middle temporal gyrus; occipital_mid_r: right middle occipital gyrus; rolandic_oper_l: left rolandic operculum; cerebellum_9_l: left cerebellum; cerebellum_9_r: right cerebellum; hippocampus_r: right hippocampus gyrus; parahippocampal_r: right parahippocampal gyrus; thalamus_r: right thalam; temporal_mid_l: left middle temporal gyrus; insula_r: right insula; precuneus_r: right precuneus; cuneus_l: left cuneus. xiuxin wang et al. admet & dmpk 5(4) (2017) 234-241 238 involved in the human understanding insight which occupied an important position in understanding the clinical heterogeneity of obsessive-compulsive disorder (ocd) [18]. glasser and rilling [19] found that the temporal lobe, especially the left superior and middle temporal gyri, was a pivotal part of language pathway in the brain’s language network. the cerebellum was vital to the motor control, cognitive and affective regulation [7]. shen et al. [20] reported that the cerebral microbleeds (cmbs) observed in the cerebellum were related to cognitive dysfunction. the insula participated in numerous cognitive processing, including goal-directed cognition, conscious awareness, autonomic regulation, interoception and somatosensation [21]. the precuneus was found to be connected with a wide range of cognitive processes, involving reflective and self-related processing [22, 23], awareness and conscious information processing [24, 25], episodic memory [26-28], and visuospatial processing [29, 30]. the cuneus was significantly involved with visual processing [30, 31]. collignon et al. [32] demonstrated that auditory-spatial processing mainly recruited the right cuneus and the right middle occipital gyrus, and the specific occipital regions in visuospatial / motion processing for sighted individuals. the thalamus was believed to regulate and coordinate cortical activity both within and across functional regions, such as motor and visual cortices [33]. wu et al. [34] found that the npc patients after radiotherapy demonstrated significant changes in cognition which mainly referred to short term memory, delayed recall, language, attention, orientation, visuo-spatial and executive function. in addition, the patients showed a relatively high rate of general intelligence impairment, which might be related to a longer post-rt interval of mean 4.3 years and bigger total dosage of mean 70.7gy. tang et al [35] reported that npc patients with radiation-induced brain injury (ri) exhibited negative emotions, impaired cognitive function and quality of life (qol). the accumulated studies showed that the radiation-induced functional impairments included the disorder of short-term memory [36], personality changes [37] and motor abilities [38], a marked anterograde memory impairment for verbal material [39], neuropsychological impairments in recent memory, immediate and delayed verbal recall, and immediate visual recall [40]. hence, we inferred that these abnormal brain regions might be correlated with the dysfunction of patients with npc after radiotherapy, which might reveal a potential mechanism for radiation-induced impairment. furthermore, it was clear that the machine learning framework proposed in the current study could provide an effective reference for identifying the brain injury induced by radiotherapy in clinical diagnosis. conclusions in summary, this present study proposed an automatic classification method for pre-rt and post-rt. the experimental results showed that the method could effectively identify the abnormal changes between pre-rt and post-rt. furthermore, these changes might be considered as an underlying biomarker to detect the radiation-induced impairment. admet & dmpk 5(4) (2017) 234-241 radiotherapy-induced changes in patients with nasopharyngeal carcinoma doi: 10.5599/admet.5.4.484 239 acknowledgements: this research is supported by the chongqing research program of basic science and frontier technology (no. cstc2017jcyjbx0007; no. cstc2015jcyja10024). references [1] l.l. tang, r. guo, g. zhou, y. sun, l.z. liu, a.h. lin, h. mai, j. shao, l. li, j. ma. prognostic value and staging classification of retropharyngeal lymph node metastasis in nasopharyngeal carcinoma patients treated with intensity-modulated radiotherapy. plos one 9 (2014) e108375. [2] c. li, y. gang. an overview on the role of telomere , telomerase in degenerative diseases and cancer. admet & dmpk 3 (2015). [3] a.t. chan. nasopharyngeal carcinoma. annals of oncology official journal of the european society for medical oncology 21 suppl 7 (2010) vii308. [4] s. ying, g.q. zhou, z.y. qi, z. li, s.m. huang, l.z. liu, l. li, a.h. lin, j. ma. radiation-induced temporal lobe injury after intensity modulated radiotherapy in nasopharyngeal carcinoma patients: a dosevolume-outcome analysis. bmc cancer,13,1(2013-08-27) 13 (2013) 397-397. [5] j. lin, x. lv, m. niu, l. liu, j. chen, f. xie, m. zhong, s. qiu, l. li, r. huang. radiation-induced abnormal cortical thickness in patients with nasopharyngeal carcinoma after radiotherapy. neuroimage clinical 14 (2017) 610-621. [6] q. ma, d. wu, l.l. zeng, h. shen, d. hu, s. qiu. radiation-induced functional connectivity alterations in nasopharyngeal carcinoma patients with radiotherapy. medicine 95 (2016) e4275. [7] q. ma, l.l. zeng, j. qin, z. luo, j. su, d. wu, s. qiu, d. hu. radiation-induced cerebellar-cerebral functional connectivity alterations in nasopharyngeal carcinoma patients. neuroreport 28 (2017) 705. [8] r.e. fan, k.w. chang, c.j. hsieh, x.r. wang, c.j. lin. liblinear: a library for large linear classification. journal of machine learning research 9 (2008) 1871-1874. [9] x.f. lv, x.l. zheng, w.d. zhang, l.z. liu, y.m. zhang, m.y. chen, l. li. radiation-induced changes in normal-appearing gray matter in patients with nasopharyngeal carcinoma: a magnetic resonance imaging voxel-based morphometry study. neuroradiology 56 (2014) 423-430. [10] y. tian, h. zhang, w. xu, h. zhang, l. yang, s. zheng, y. shi. spectral entropy can predict changes of working memory performance reduced by short-time training in the delayed-match-to-sample task. frontiers in human neuroscience 11 (2017) 437. [11] y. tian, z. ding, k.y. tam, z. wang, h. zhang, d. zhao, y. zhao, w. xu, s. zheng. specific frequency bands of amplitude low-frequency fluctuations in memory-related cognitive impairment: predicting alzheimer’s disease. admet & dmpk 3 (2015) 659-668. [12] y. tian, l. yang, s. chen, d. guo, z. ding, k.y. tam, d. yao. causal interactions in resting-state networks predict perceived loneliness. plos one 12 (2017) e0177443. [13] y. tian, l. yang, w. xu, h. zhang. predictors for drug effects with brain disease: shed new light from eeg parameters to brain connectomics. european journal of pharmaceutical sciences 110 (2017) 2636. [14] v.d. bohbot, j.j.b. allen, a. dagher, s.o. dumoulin, a.c. evans, m. petrides, m. kalina, k. stepankova, l. nadel. role of the parahippocampal cortex in memory for the configuration but not the identity of objects: converging evidence from patients with selective thermal lesions and fmri. frontiers in human neuroscience 9 (2015) 431. [15] y. tian, s. liang, d. yao. attentional orienting and response inhibition: insights from spatial-temporal neuroimaging. neuroscience bulletin 30(1) (2014) 141-151. [16] c. jing, m. dassarath, z. yin, h. liu, k. yang, w. gang. radiation induced temporal lobe necrosis in patients with nasopharyngeal carcinoma: a review of new avenues in its management. radiation oncology 6 (2011) 1-8. xiuxin wang et al. admet & dmpk 5(4) (2017) 234-241 240 [17] y. cui, b. liu, m. song, d.m. lipnicki, j. li, s. xie, y. chen, p. li, l. lu, l. lv. auditory verbal hallucinations are related to cortical thinning in the left middle temporal gyrus of patients with schizophrenia. psychological medicine (2017) 1-8. [18] j. fan, m. zhong, j. gan, w. liu, c. niu, h. liao, h. zhang, c. tan, j. yi, x. zhu. spontaneous neural activity in the right superior temporal gyrus and left middle temporal gyrus is associated with insight level in obsessive-compulsive disorder. journal of affective disorders 207 (2017) 203-211. [19] m.f. glasser, j.k. rilling. dti tractography of the human brain's language pathways. cerebral cortex 18 (2008) 2471. [20] q. shen, f. lin, x. rong, w. yang, y. li, z. cai, p. xu, y. xu, y. tang. temporal cerebral microbleeds are associated with radiation necrosis and cognitive dysfunction in patients treated for nasopharyngeal carcinoma. international journal of radiation oncology biology physics 94 (2016) 1113. [21] d. borsook, r. veggeberg, n. erpelding, r. borra, c. linnman, r. burstein, l. becerra. the insula: a "hub of activity" in migraine. neuroscientist a review journal bringing neurobiology neurology & psychiatry 22 (2016) 632. [22] t.w. kjaer, m. nowak, k.w. kjaer, a.r. lou, h.c. lou. precuneus-prefrontal activity during awareness of visual verbal stimuli. consciousness & cognition 10 (2001) 356. [23] h.c. lou, b. luber, m. crupain, j.p. keenan, m. nowak, t.w. kjaer, h.a. sackeim, s.h. lisanby. parietal cortex and representation of the mental self. proceedings of the national academy of sciences of the united states of america 101 (2004) 6827. [24] t.w. kjaer, m. nowak, h.c. lou. reflective self-awareness and conscious states: pet evidence for a common midline parietofrontal core. neuroimage 17 (2002) 1080-1086. [25] b.a. vogt, s. laureys. posterior cingulate, precuneal & retrosplenial cortices: cytology & components of the neural network correlates of consciousness. progress in brain research 150 (2005) 205. [26] d. dörfel, a. werner, m. schaefer, r.v. kummer, a. karl. distinct brain networks in recognition memory share a defined region in the precuneus. european journal of neuroscience 47 (2009) 1947– 1959. [27] b.n. lundstrom, m. ingvar, k.m. petersson. the role of precuneus and left inferior frontal cortex during source memory episodic retrieval. neuroimage 27 (2005) 824-834. [28] b.n. lundstrom, k.j. petersson, m. johansson, p. fransson, m. ingvar. isolating the retrieval of imagined pictures during episodic memory: activation of the left precuneus and left prefrontal cortex. neuroimage 20 (2003) 1934-1943. [29] n. wenderoth, f. debaere, s. sunaert, s.p. swinnen. the role of anterior cingulate cortex and precuneus in the coordination of motor behaviour. european journal of neuroscience 22 (2005) 235. [30] y. tian, a.b. chica, p. xu, d. yao. differential consequences of orienting attention in parallel and serial search: an erp study. brain research 1391 (2011) 81. [31] j.s. lee, g. park, m.j. song, k.h. choi, s.h. lee. early visual processing for low spatial frequency fearful face is correlated with cortical volume in patients with schizophrenia. neuropsychiatric disease & treatment 12 (2016) 1-14. [32] o. collignon, g. vandewalle, p. voss, g. albouy, g. charbonneau, m. lassonde, f. lepore. functional specialization for auditory–spatial processing in the occipital cortex of congenitally blind humans. proceedings of the national academy of sciences of the united states of america 108 (2011) 44354440. [33] m. malekmohammadi, w.j. elias, n. pouratian. human thalamus regulates cortical activity via spatially specific and structurally constrained phase-amplitude coupling. cerebral cortex 25 (2015) 1618. admet & dmpk 5(4) (2017) 234-241 radiotherapy-induced changes in patients with nasopharyngeal carcinoma doi: 10.5599/admet.5.4.484 241 [34] x. wu, m. gu, g. zhou, x. xu, m. wu, h. huang. cognitive and neuropsychiatric impairment in cerebral radionecrosis patients after radiotherapy of nasopharyngeal carcinoma. bmc neurology 14 (2014) 1-6. [35] y. tang, d. luo, x. rong, x. shi, y. peng. psychological disorders, cognitive dysfunction and quality of life in nasopharyngeal carcinoma patients with radiation-induced brain injury. plos one 7 (2012) e36529. [36] k.y. hsiao, s.a. yeh, c.c. chang, p.c. tsai, j.m. wu, j.s. gau. cognitive function before and after intensity-modulated radiation therapy in patients with nasopharyngeal carcinoma: a prospective study. international journal of radiation oncology biology physics 77 (2010) 722-726. [37] d.o. oncology, q.v. hospital, candos, quatrebornes. báo cáo khoa học: "radiation induced temporal lobe necrosis in patients with nasopharyngeal carcinoma: a review of new avenues in its management" doc. gastrointestinal endoscopy 65 (2007) ab155. [38] m. cheung, a.s. chan, s.c. law, j.h. chan, v.k. tse. cognitive function of patients with nasopharyngeal carcinoma with and without temporal lobe radionecrosis. archives of neurology 57 (2000) 1347-1352. [39] a.j. parkin, n.m. hunkin. memory loss following radiotherapy for nasal pharyngeal carcinoma — an unusual presentation of amnesia. british journal of clinical psychology 30 (2011) 349-357. [40] m.s. hua, s.t. chen, l.m. tang, w.m. leung. neuropsychological function in patients with nasopharyngeal carcinoma after radiotherapy. journal of clinical & experimental neuropsychology 20 (1998) 684-693. ©2017 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ animal model investigation suggests betamethasone alters the pharmacokinetics of amikacin doi: 10.5599/admet.613 279 admet & dmpk 6(4) (2018) 279-283; doi: http://dx.doi.org/10.5599/admet.613 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index short communication animal model investigation suggests betamethasone alters the pharmacokinetics of amikacin seth kwabena amponsah 1 , kwabena frimpong-manso opuni 2 *, ama asiedua donkor 1 1 department of pharmacology and toxicology, school of pharmacy, university of ghana, accra, ghana 2 department of pharmaceutical chemistry, school of pharmacy, university of ghana, accra, ghana *corresponding author: e-mail: kfopuni@ug.edu.gh; telephone: +233208260595 received: august 30, 2018; revised: october 25, 2018; published: november 01, 2018 abstract corticosteroids, such as betamethasone, are sometimes administered to women who are at risk of pre-term delivery. these corticosteroids cross the placenta to the fetus and decrease respiratory distress syndrome in preterm newborns. preterm newborns are often susceptible to infections partly due to their immature immune system. amikacin is one of the aminoglycosides used in the treatment of newborn infections. there is, however, a dearth of information on the effect of prenatal corticosteroids on the disposition of aminoglycosides administered to newborns days later. we evaluated the effect of pre-administration of betamethasone on the disposition of amikacin, 72 h after last dose of betamethasone, using an animal model. the pharmacokinetic parameters of rats administered betamethasone followed by amikacin vis -avis rats administered saline followed by amikacin were as follows: cmax; 16.6 μmol l -1 vs. 31.4 μmol l -1 , auc0→8; 26.8 μmol h l -1 vs. 153.5 μmol h l -1 , ke; 0.26 h -1 vs. 0.18 h -1 , and t1/2; 2.6 h vs. 3.9 h, respectively. about a 1.5-fold increase in the elimination of amikacin was observed in the sprague-dawley rats pretreated with betamethasone compared with those pre-treated with saline. this ultimately led to differences in the other pharmacokinetic parameters amongst the two groups of rats. although an animal model investigation showed some level of interaction, a follow-up study in preterm newborns where possible interaction of the two drugs is studied later than day 1, is recommended. keywords excretion; interaction; preterm introduction according to the world health organization, about 15 million babies (more than 1 in 10 babies worldwide) are born prematurely annually [1]. in 2015, an estimated 1 million preterm deaths were recorded, while many survivors of preterm birth faced life-long morbidities such as learning disabilities and respiratory problems [1]. newborn care, which includes pharmacotherapy, has been one of the interventions to reduce the burden of morbidity and mortality in this cohort of infants. corticosteroids are usually administered to women who are at risk of imminent preterm birth. betamethasone, a corticosteroid of choice in some resource-poor settings, is known to decrease the occurrence of respiratory distress syndrome in preterm newborns. the main mechanism of betamethasone is to accelerate the development of types i and ii pneumocytes (alveolar cells) of the fetus after the drug http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:kfopuni@ug.edu.gh amponsah et al. admet & dmpk 6(4) (2018) 279-283 280 has crossed the placenta in its active form. usually, a single course of betamethasone is recommended between 24 and 33 weeks of gestation among pregnant women who are at risk of preterm delivery (often administered within 7 days). reports suggest that antenatal betamethasone increases glomerular filtration rate (gfr) on day 3 of the life of preterm newborns [2]. furthermore, newborns, especially those born prematurely are highly susceptible to infections due to their immature immune system [3]. the cornerstone of treatment of these infections is the use of antibiotics. amikacin is a commonly prescribed aminoglycoside for empirical treatment of gram-negative bacterial infections in neonatal intensive care units in most developing countries [4]. aminoglycosides are known to be excreted almost solely via the kidneys [5]. there is the likelihood, however, that the effect of corticosteroids on gfr may ultimately alter serum levels of aminoglycosides administered to newborns. a decrease in serum levels of aminoglycosides as a result of alteration in gfr by corticosteroids may lead to selection for resistant microorganism strains and infection treatment failure. a previous study has showed that administration of betamethasone to pregnant mothers had no significant effect on amikacin clearance on the first day of life of preterm newborns [6]. in that study, blood samples for therapeutic drug monitoring were drawn 40 min before and after completion of the second dose of amikacin [6]. hence, blood samples for pharmacokinetic analysis were taken between 23 – 43 h, according to various dosing schemes employed in that study. it is noteworthy, however, that another study concluded that prenatal exposure to betamethasone increased gfr; but this effect was observed on day 3 (72 h) of the life of preterm infants [2]. also considering the fact that amikacin treatment for newborns with infection could range from 5 – 7 days, there is the likelihood that day 1 blood samples for pharmacokinetic (clearance) estimations may not reflect the effect of prenatal administration of betamethasone on amikacin. we therefore sought to investigate the effect of pre-administration of betamethasone on the pharmacokinetics of amikacin using an animal model, where blood samples for pharmacokinetic estimates of amikacin were drawn 72 h after the last dose of betamethasone. an animal model will provide preliminary account of possible pharmacokinetic interaction of these two drugs, and serve as a basis for full clinical studies in humans. it is noteworthy that the initial experimental design tried to mimic exactly what pertains in humans: administration of betamethasone to pregnant rat, followed by administration of amikacin to litters, and serial drawing of blood from litters for pharmacokinetic estimation. however, this approach had ethical (drawing blood 5 times from rat litter) and operational implications (low blood volumes). therefore, the current study had to assess possibility of interaction between amikacin and betamethasone using young adult rats. experimental experimental animals the study was approved by the ethical and protocol review committee of the university of ghana medical school (protocol id: ms-et/m.8-p.5.3). the animals were handled using procedures and techniques in accordance with the guidelines published by the national institute of health for the care of laboratory animals [7]. twelve (12) healthy male sprague-dawley (sd) rats (7 8 weeks old, weighing between 200 – 250 g) were obtained and kept under suitable conditions throughout the experimental process. the animal model used only male rats to exclude sex as a confounding factor. a controlled breeding room of temperature 26 admet & dmpk 6(4) (2018) 279-283 betamethasone affects the disposition of amikacin doi: 10.5599/admet.613 281 ± 2 °c with a 12 h light/dark cycle was used. the sd rats were maintained on local animal feed and water ad libitum. the animals were put into two (2) groups of six (6) – group 1 and group 2. the animals in group 1 (test) were first administered with standard doses of betamethasone and thereafter, standard doses of amikacin. animals in group 2 (control) were treated in a similar manner except replacing betamethasone with normal saline. drug administration a dose of 0.15 mg kg -1 of betamethasone (diprofos*, schering-plough, brussels, belgium) was administered intraperitoneally to the test group every 24 h for 3 consecutive days. betamethasone dose was based on allometric scaling from humans (prenatal doses of 12 mg given every 24 h for 2 -3 days) to rats. after this, a dose of 10 mg kg -1 of amikacin (amikin®, bristol-myers pharmaceuticals, middlesex, england) was administered intraperitoneally to the rats every 24 h for 3 consecutive days. of note, similar amikacin dosing in rats has been reported elsewhere [8]. normal saline was administered intraperitoneally to rats in group 2, using same volume per weight administration as with betamethasone. the normal saline was administered to sd rats for 3 consecutive days. afterwards, a dose of 10 mg kg -1 of amikacin was administered via the intraperitoneal route every 24 h for 3 consecutive days. blood sample collection and estimation of serum amikacin level tail vein blood samples were taken following the fourth dose (72 h) of amikacin after 0.5, 1, 2, 4 and 8 h, for rats in groups 1 and 2. blood samples were collected into microtainer gel tubes and centrifuged at 1,500 rpm for 10 min to obtain serum. samples from sd rats were pooled as a result of low serum volume required for assay. therefore, one pooled-sample was obtained for each time point for groups 1 and 2. usually, challenges associated with low sample volume during pharmacokinetic studies can be circumvented by sample-pooling, which will not alter eventual pharmacokinetic estimation [9]. fluorescence polarization immunoassay (fpia), using cobas integra® 400 (roche diagnostic limited, basel, switzerland) as previously described [10] was the assay used in determining serum amikacin levels. the principle behind fpia is competition for antibody binding sites between drug in sample and drug labeled with an enzyme. the coefficient of variation over the entire calibration range during assay for amikacin was less than 4 %. pharmacokinetic analysis estimation of pharmacokinetic parameters of amikacin in the two groups was by a non-compartmental analytical approach. the linear trapezoidal rule was used in extrapolating area under the concentrationtime curves (aucs) for the two groups. aucs were from time 0 h till the last measurement point (8 h). elimination rate constant (ke) was assessed by linear regression of the terminal part of a log of the concentration and time curves. elimination half-life (t1/2) was calculated as; t1/2 = 0.693 ke -1 . results and discussion concentration-time curve of each group after amikacin samples were pooled at the various predetermined times in the sd rats is shown in figure 1. estimates of pharmacokinetic parameters of amikacin in sd rats (table 1) were obtained from concentration-time curves. elimination (ke) of amikacin was faster in sd rats pre-administered with amponsah et al. admet & dmpk 6(4) (2018) 279-283 282 betamethasone compared to sd rats pre-administered with normal saline, and this resulted in a shorter half-life for the former. the maximum amikacin concentration (cmax) for sd rats pre-administered with normal saline was about two-fold higher than sd rats pre-administered with betamethasone. total drug exposure from time 0 to 8 h (auc0→8) was also about four-fold greater for the sd rats pre-administered with normal saline. figure 1. concentration-time curves of amikacin in rats after pre-treatment with normal saline (red solid line; filled circles) and betamethasone (gray solid line; filled squares) table 1. summary of amikacin pharmacokinetic parameters among two groups of rats pk parameter (units) betamethasone + amikacin normal saline + amikacin cmax (μmol l -1 ) 16.60 31.40 auc0→8 (μmol l -1 h -1 ) 26.80 153.50 ke (h -1 ) 0.26 0.18 t1/2 (h) 2.62 3.96 possible causes of resistance to antibiotics include amongst others, drug-drug interactions. drug interactions leading to low serum levels of antibiotics can result in sub-therapeutic effect or emergence of resistant strains of microorganisms [11]. it is therefore imperative that potential drug interactions involving antibiotics are identified. in this study, we investigated the effect of pre-treatment of betamethasone on the disposition of amikacin, after 72 h, using an animal model. it is evident that amikacin was eliminated at a faster rate in the betamethasone pre-treated group compared to the normal saline pre-treated group. this led to a shorter half-life (2.62 h) in the betamethasone pre-treated group compared to the normal saline group (3.96 h). betamethasone is known to alter kidney function by increasing gfr [2], and this may have consequently led to the increase in renal elimination of amikacin in the betamethasone pre-treated group of this study. the outcome of this work suggests that pre-treatment with betamethasone could increase renal elimination of amikacin, possibly through increasing gfr 72 h after daily amikacin administration. if indeed this effect observed in the current study occurs in newborns, then there may be a need for dose modification in those who receive prenatal corticosteroids. admet & dmpk 6(4) (2018) 279-283 betamethasone affects the disposition of amikacin doi: 10.5599/admet.613 283 conclusions the current study showed that betamethasone altered the pharmacokinetic parameters of amikacin in sprague-dawley rats 72 h after daily amikacin administration. although this study was conducted in an animal model, this finding is noteworthy, as this may serve as a basis for future studies in humans where blood samples for pharmacokinetic evaluation of amikacin are taken after day 3 of the life of preterm newborns. acknowledgements the authors are grateful to perseus nana asare and kwabena antwi, who helped with animal handling. thanks also go to the technical staff of the department of pharmacology and toxicology, and department of pharmaceutical chemistry, school of pharmacy, university of ghana. references [1] care of the preterm and low-birth-weight newborn. world health organization, geneva, switzerland. [2] j.n. van den anker, w.c. hop, r. de groot, b.j. van der heijden, h.m. broerse, j. lindemans, p.j. sauer. effects of prenatal exposure to betamethasone and indomethacin on the glomerular filtration rate in the preterm infant. pediatr res 36 (1994) 578-581. [3] j.l. wynn, j. neu, l.l. moldawer, o. levy. potential of immunomodulatory agents for prevention and treatment of neonatal sepsis. j perinatol 29 (2009) 79-88. [4] a. siddiqi, d.a. khan, f.a. khan, a. razzaq. therapeutic drug monitoring of amikacin in preterm and term infants. singapore med j 50 (2009) 486-489. [5] k. bruce, r.w. hicks. perioperative pharmacology: a focus on aminoglycosides. aorn j 93 (2011) 464-468; quiz 469-471. [6] k. allegaert, v. cossey, m. rayyan, c. vanhole, h. devlieger. administration of betamethasone before birth has no effect on the amikacin pharmacokinetics in preterm infants at birth. paediatric and perinatal drug therapy 6 (2005) 149-154. [7] memorandum of understanding between the office of laboratory animal welfare natioanal institutes of health u.s. department of health and human services and the office of research oversight and the office of research and development veterans health administration u.s. department of veterans affairs concerning laboratory animal welfare. national institutes of health, 2007. [8] h.j. olguín, m.c. portugal, j.f. pérez, a.c. vieyra, c.f. pérez, a.a. rodríguez. effect of mannitol on the pharmacokinetics of amikacin in wistar rats. brazilian archives of biology and technology 52 (2009) 835-839. [9] l.e. riad, k.k. chan, r.j. sawchuk. determination of the relative formation and elimination clearance of two major carbamazepine metabolites in humans: a comparison between traditional and pooled sample analysis. pharmaceutical research 8 (1991) 541-543. [10] i. hemmila. fluoroimmunoassays and immunofluorometric assays. clin chem 31 (1985) 359-370. [11] j. van den anker. pharmacokinetics and renal function in preterm infants. acta paediatrica 85 (1996) 1393-1399. ©2018 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ manuscript doi: 10.5599/admet.5.1.375 9 admet & dmpk 5(1) (2017) 9-13; doi: 10.5599/admet.5.1.375 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index mini review biomembrane mimics and their roles in anti-bacterial drug discovery xiaohui hu and kin tam* drug development core, faculty of health sciences, university of macau, macau, china *corresponding author: e-mail: kintam@umac.mo; tel.: +853-8822-4988; fax: +853-8822-2314 received: february 17, 2017; revised: february 19, 2017; published: march 25, 2017 abstract cell membrane is at the frontline of the battle between pathogenic microbe and host. a thorough understanding of bacterial membrane is fundamental to tackle infection disease. membrane mimetic provides a powerful tool for mechanistic investigation of drug-membrane interaction. herein, we summarized major features of bacterial and mammalian cell in context of antibacterial therapy. many details were given to model membranes and their application in mechanistic studies. current challenge in antibacterial therapy and perspective of membrane mimics in antibacterial drug discovery were also provided. keywords drug-membrane interaction; mode of action; liposome; supported lipid bilayer; resistance. introduction cell membrane defines the physical boundary of a cell from outside environment. it contains lipids, proteins and carbohydrates, which usually arrange into structure unique to specific cell type and are critical to corresponding function. of all kinds of membrane components, lipids are relatively homogenous and better studied. lipids are amphipathic molecules with a polar or charged head group and a hydrophobic acyl chain tail. they tend to form head-by-head, tail-to-tail bilayer structure in cell membrane to extrude water, salt and other hydrophilic molecules and maintain low entropy state. this ordered bilayer structure serves as the membrane framework, on which proteins and carbohydrates carry out their designated biological functions. as the target of more than half of all modern drugs [1], membrane proteins interact with surrounding lipid through their transmembrane domain and are vital to variety of cell function, including signal transduction, transporting, enzymatic catalysis and cell adhesion [2]. carbohydrates in plasma membrane, predominantly glycoproteins but also with some glycolipids, are generally found on the extracellular side and play significant roles in many biological events such as cell adhesion and immunological response. as diverse as the membrane composition is, these components are not randomly or evenly distributed around the cell sphere. they tend to form distinct regions or microdomains, which contain specific type of lipids, glycolipids and proteins within fluid bilayer and perform unique biological function such as signal transduction [3,4]. http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:kintam@umac.mo hu and tam admet & dmpk 5(1) (2017) 9-13 10 studying biological membrane is always a challenging task because of its complicated composition and fluidic structure. these difficulties could be partially circumvented by using simplified lipids-based model systems, such as vesicles, micelles, bicelles, nanodiscs and bilayers. through control of the lipid content and buffer conditions (salt concentration, ph etc.), model systems can be utilized to explore interaction between membranes and exogenous objects (such as drugs) quantitatively. in this article, we will touch specifics about bacterial and mammalian cell membrane in context of antibacterial therapy. we will then discuss the development of membrane models and their application in studies pertaining to biological function of the membrane. we will close the review with current challenge in antibacterial therapy and perspective of membrane mimics. bacterial and mammalian membrane although lipids in plasma membrane can be schematized as bilayer for simplicity, its composition and structure could differ greatly among species, cells from different organisms and tissues. in general, grampositive bacteria contain a single cell membrane surrounded by a thick layer of peptidoglycan (cell wall), while gram-negative bacteria plasma membrane is made up of two layer membrane (innerand outer) encompassed by a thin layer of peptidoglycan. membrane of the model bacterium escherichia coli is made up of three major phospholipids: ~75 % phosphatidylethanolamine (pe), ~20 % phosphatidylglycerol (pg) and ~5 % cardiolipin (cl) [5]. but this cannot be generalized to all bacteria, as diversified lipids can be found in different microorganisms, or even in the same species in different environment [6]. in contrast to bacterial membrane, mammalian membrane is void of cell wall and exposed directly to surrounding environment, which thus demands more stringent regulation. mammalian membranes of different cell types or from distinct tissues could vary significantly in its fluidity, which is believed to be modulated at least partially by the content of cholesterol [4], an integral lipid unique to animal cell membrane. asymmetric distribution of lipid components on different sides of individual cell is another remarkable feature of multicellular organism cell membrane. structural and compositional difference between bacteria and mammalian cell membrane represents also a big opportunity in antibacterial therapy for patient. antimicrobial actions that are related to bacterial membranes pathogenic bacteria pose a major threaten to public health, especially in under-developed regions of the world. large scale production and use of antibiotic penicillin during the world war ii is a hallmark of modern medicine. since then, a plethora of antimicrobial agents have been developed and commercialized, playing significant roles in improvement of human health, highlighted by gradual increasing of average life span worldwide. depending on their structure and target sites within microorganism, different antibacterial drugs have distinct modes of action. modes of action related to bacterial membrane are inhibition or block of cell wall synthesis, inhibition of membrane protein or lipid synthesis and interference of cell membrane function. as aforementioned, bacterial membrane is surrounded by a layer of peptidoglycan cell wall. this unique feature rendered a great interest in searching for glycopeptide and lipoglycopeptide, which mimic bacterial cell wall synthesis substrates or bind to synthesis intermediates, inhibiting cell wall synthesis [7], while having little effect on host cells. some antibacterial peptides [8,9], peptidomimics [10] and small proteins [11] could also interact with or insert into cell membrane, where they interfere with membrane formation or alter local membrane structure, thus reducing bacteria growth or reproduction. interestingly, some admet & dmpk 5(1) (2017) 9-13 membrane mimics in anti-bacterial drug discovery doi: 10.5599/admet.5.1.375. 11 pathological microorganisms could also as a countermeasure secrete protein toxins which could attack host cells by forming pores on their membranes [12,13]. membrane is the combating field of the war between pathogens and their hosts, which also make mechanistic study on biological membrane fundamental to antibacterial therapy development. current biomimetic membrane models and applications the complexity of biological membrane makes it always difficult or impractical in a lot of case to study the membrane intact. a variety of membrane models has been developed over decades to overcome the obstacle. vesicles (liposome) are the most commonly used biological mimics as they are relatively easy to prepare and handle. vesicle sphere consists of self-assembled round lipid-bilayer(s) which mimic plasma membrane. enclosed aqueous compartment resembles cytoplasma can be used to incorporate diverse materials of interest (including drugs). depending on its size and structure, vesicle can be either smallunilamellar (suv), large-unilamellar (luv) or multilamellar (mlv). a variety of biophysical methods have been employed to study the interaction between drug and membrane mimicked by lipid vesicle. simple mathematic model of drug diffusion into or release from cell could be derived from those studies. in addition to its application in the mechanistic study of membrane, controlled inner environment of liposomes have been used to deliver charged drugs [14,15] through membrane [16]. micelles are aggregates of surfactant molecules dispersed in a liquid. the stability of colloid solution was established by the repulsion forces between hydrophobic surfactant tails and hydrophilic solvent molecules. lipid molecules can be used as surfactant to form micelle and to dissolve membrane protein in the structural studies [17–19]. however, the application of micelles as membrane mimics is limited by its small size and a single layer structure. in disc-shaped bicelles, lipid hydrophobic tails pack against each other as bilayer in the center and form an amphipathic, micelle-like assembly at the edge. bicelles are much smaller than vesicles, and are ideal in the experiments where the larger liposomes are not an option [20–22]. nanodiscs are patch of lipid bilayer whose edge is wrapped with an amphipathic coating protein [23]. compared to bicelles and micelles, nanodiscs are more stable and uniform in size (depending on the length of coating protein). membrane protein of interest (such as drug target) can be inserted into nanodiscs and examined both biophysically and functionally [24]. other than being freely mobile in aqueous solution, lipid-based membranes can also be studied while attached to device or to supporting materials. the first model developed of this kind is a black lipid membrane (blm) [25,26], where lipid bilayer is formed within an aperture on a thin layer of hydrophobic material such as teflon. black lipid membrane has been employed to study membrane properties (electrical potential [27], structure [28] and function [29]), lipid-protein interaction [30], peptides or protein diffusion across membrane [31]. application of micro-black lipid membrane technology had led to development of high throughput drug permeability assay system pampa (parallel artificial membrane permeation assay) [32]. lipid bilayer can also be attached to the surface of solid material to form a stable membrane mimic called supported lipid bilayer (slb). great stability and surface planarity of supported lipid bilayer make it especially useful for techniques requiring direct physical contact (e.g. atomic force microscopy (afm) [33] and electrode-supported lipid nanoassemblies [34]) or rigidly-supported planar surface (e.g. total internal reflection fluorescence microscopy (tirf) [35] and surface plasmon resonance (spr) [36]). supported lipid bilayer has another advantage in that, in light of the lipid rafts concept, lipids can be patterned into isolated regions through design of the surface [37]. hu and tam admet & dmpk 5(1) (2017) 9-13 12 beside lipids, biological membranes also contain variety of protein and carbohydrates. thus, extra care must be taken when interpreting the studies obtained from lipid-based membrane mimics. in more advanced model systems, membrane proteins can also be reconstituted into artificial lipid bilayer to simulate protein function in lipid environment [34]. challenge and perspective one of the biggest problems in the antibacterial chemotherapy field is development of pathogen resistance to antibiotics [38]. infectious pathogen could quickly develop resistance to drugs applied on them by either mutating into new strains to get around the drug blocking pathway or inducing counting mechanism to remove the drug [39]. on the other hand, development of novel antibiotics has been stagnant since a half century ago [40]. a thorough understanding of the mechanism of drug action, metabolism and transportation is critical to develop new drugs to cope with antibiotics resistance. model membrane study is fundamental to elucidating these mechanisms, upon which development of agents with novel [41-43] and/or multiple [39,44,45] mode(s) of action poses some most promising strategies to bring us new tools to combat microbial infection diseases. conclusions cell membranes are at the frontlines of the battle field between pathogenic microbe and host. mechanistic study of the membrane is crucial for a thorough understanding of host-pathogen interaction, which could lead to discovery of new drug for efficient infection control. biomembrane mimics are powerful tools to gain mechanistic insight of drug-membrane interaction by simplifying otherwise complicated membrane structure with controlled components. multiple type of biomimetic membrane have been developed to meet variety of needs in mechanistic studies. acknowledgements: we thank the financial support from the university of macau (grant no.: srg201300055-fhs). references [1] j.p. overington, b. al-lazikani, a.l. hopkins, nat. rev. drug discov. 5 (2006) 993–996. [2] m.s. almén, k.j.v. nordström, r. fredriksson, h.b. schiöth, bmc biol. 7 (2009) 50. [3] d. lingwood, k. simons, science 327 (2010) 46–50. [4] k. simons, d. toomre, nat. rev. mol. cell biol. 1 (2000) 31–39. [5] c. sohlenkamp, o. geiger, fems microbiol. rev. (f. narberhaus, ed.) 40 (2016) 133–159. [6] i.m. lópez-lara, o. geiger, biochim. biophys. acta mol. cell biol. lipids. (2016) doi: 10.1016/j.bbalip.2016.10.007. [7] d. kahne, c. leimkuhler, w. lu, c. walsh, chem. rev. 105 (2005) 425-448. [8] m. zasloff, nature 415 (2002) 389–395. [9] k.a. brogden, nat. rev. microbiol. 3 (2005) 238–250. [10] t. velkov, p.e. thompson, r.l. nation, j. li, j. med. chem. 53 (2010) 1898–1916. [11] s. mukherjee, h. zheng, m.g. derebe, k.m. callenberg, c.l. partch, d. rollins, d.c. propheter, j. rizo, m. grabe, q.-x. jiang, l.v. hooper, nature 505 (2014) 103–107. [12] c.j. rosado, s. kondos, t.e. bull, m.j. kuiper, r.h.p. law, a.m. buckle, i. voskoboinik, p.i. bird, j.a. trapani, j.c. whisstock, m.a. dunstone, cell. microbiol. 10 (2008) 1765–1774. admet & dmpk 5(1) (2017) 9-13 membrane mimics in anti-bacterial drug discovery doi: 10.5599/admet.5.1.375. 13 [13] r.h.p. law, n. lukoyanova, i. voskoboinik, t.t. caradoc-davies, k. baran, m.a. dunstone, m.e. d’angelo, e.v. orlova, f. coulibaly, s. verschoor, k.a. browne, a. ciccone, m.j. kuiper, p.i. bird, j.a. trapani, h.r. saibil, j.c. whisstock, nature 468 (2010) 447–451. [14] t.m. allen, p.r. cullis, adv. drug deliv. rev. 65 (2013) 36-48. [15] v.p. torchilin, nat. rev. drug discov. 4 (2005) 145–160. [16] a.s. ulrich, biosci. rep. 22 (2002) 129-150. [17] n. sanghera, t.j. pinheiro, protein sci. 9 (2000) 1194–1202. [18] c.m. franzin, p. teriete, f.m. marassi, j. am. chem. soc. 129 (2007) 8078–8079. [19] n.p. barrera, n. di bartolo, p.j. booth, c.v. robinson, science 321 (2008) 243–246. [20] s. faham, j.u. bowie, j. mol. biol. 316 (2002) 1–6. [21] s. faham, g.l. boulting, e.a. massey, s. yohannan, d. yang, j.u. bowie, protein sci. 14 (2005) 836– 840. [22] r. ujwal, j. abramson, j. vis. exp. 59 (2012) e3383. [23] a. nath, w.m. atkins, s.g. sligar, biochemistry 46 (2007) 2059-2069. [24] t.h. bayburt, s.g. sligar, febs lett. 584 (2010) 1721-1727. [25] p. mueller, d.o. rudin, h.t. tien, w.c. wescott, nature 194 (1962) 979–980. [26] h.t. tien, s. carbone, e.a. dawidowicz, nature 212 (1966) 718-719. [27] s. ozeki, h. kurashima, m. miyanaga, c. nozawa, langmuir 16 (2000) 1478–1480. [28] r.s. ries, h. choi, r. blunck, f. bezanilla, j.r. heath, j. phys. chem. b 108 (2004) 16040–16049. [29] p. van gelder, f. dumas, m. winterhalter, biophys. chem. 85 (2000) 153-167. [30] r. smekal, h.p. ting, l.g. augenstein, h.t. tien, science 168 (1970) 1108–1109. [31] k. wei, a. neef, q. van, s. kramer, i. gregor, j. enderlein, biophys. j. 105 (2013) 455–462. [32] m. kansy, f. senner, k. gubernator, j. med. chem. 41 (1998) 1007–1010. [33] o. domenech, y.f. dufrêne, f. van bambeke, p.m. tukens, m.-p. mingeot-leclercq, biochim. biophys. acta 1798 (2010) 1876–1885. [34] m. wang, w. zhan, acc. chem. res. 49 (2016) 2551–2559. [35] j.m. moran-mirabal, j.b. edel, g.d. meyer, d. throckmorton, a.k. singh, h.g. craighead, biophys. j. 89 (2005) 296–305. [36] m.a. cooper, d.h. williams, chem. biol. 6 (1999) 891–899. [37] j.t. groves, n. ulman, s.g. boxer, science 275 (1997) 651–653. [38] l.l. silver, clin. microbiol. rev. 24 (2011) 71–109. [39] f. guilhelmelli, n. vilela, p. albuquerque, l.s. derengowski, i. silva-pereira, c.m. kyaw, front. microbiol. 4 (2013) 353. [40] m.a. fischbach, c.t. walsh, science 325 (2009) 1089–1093. [41] s.c.p. williams, proc. natl. acad. sci. u. s. a. 111 (2014) 11227–11229. [42] d.g. brown, t. lister, t.l. may-dracka, bioorg. med. chem. lett. 24 (2014) 413–418. [43] l.l. ling, t. schneider, a.j. peoples, a.l. spoering, i. engels, b.p. conlon, a. mueller, t.f. schäberle, d.e. hughes, s. epstein, m. jones, l. lazarides, v.a. steadman, d.r. cohen, c.r. felix, k.a. fetterman, w.p. millett, a.g. nitti, a.m. zullo, c. chen, k. lewis, nature 517 (2015) 455–459. [44] l. otvos, c. snyder, b. condie, p. bulet, j.d. wade, int. j. pept. res. ther. 11 (2005) 29–42. [45] v. yarlagadda, s. samaddar, k. paramanandham, b.r. shome, j. haldar, angew. chem. int. ed. engl. 54 (2015) 13644–13649. ©2017 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ doi: 10.5599/admet.1.2.6 17 admet & dmpk 1(2) (2013) 17-18; doi: 10.5599/admet.1.2.6 open access : issn: 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index book review propelling drug research by physical chemistry vesna gabelica marković fidelta d.o.o. prilaz baruna filipovića 29, hr-10000 zagreb, croatia e-mail: vesna.gabelicamarkovic@glpg.com; tel.: +385-1-88886 308; fax: +385-1-8886 444 received: march xx, 2013; revised: april xx, 2013; published: april xx, 2013 mmk physico chemical methods in drug discovery and development. zoran mandić (editor), iapc publishing, 2012. physico-chemical methods in drug discovery and development is a compilation of the key methods and techniques in drug research and development. the book comprises nine chapters written by 15 authors from international academic and industrial laboratories. compliment to the editor for selecting the authors, who are more than excellent. the chapters cover a wide variety of topics and provide a broad introduction to any individual working on the physicochemical characterisation of compounds in drug research. the first chapter provides a comprehensive review of the theoretical background of physicochemical parameters, as well as methods for their determination. the second chapter of the book describes the application of hplc as a powerful technique for measuring physicochemical parameters. basic principles of chromatographic retention and hplc methodology are used to determine various physico-chemical properties. there is an extension to bio-mimetic chromatography, followed by the application of the data to predict human in vivo distribution. chapter 3 discusses the importance of and the different approaches to determining permeability. the main focus of this chapter is experimental models and procedures available in early drug discovery. the chapter that follows describes the protein-ligand interaction as an important method used in early drug discovery. there is a summary of all the biophysical methods currently in use, followed by a few brief case studies. chapter 5 provides a detailed and comprehensive review regarding the application of nmr spectroscopy in the study of ligand-receptor interactions. the main focus is on the nmr techniques applied to studying the bioactive molecules with their biological receptors. chapter 6 describes chiral separation for enantiomeric determination in the pharmaceutical industry and is focused on the types of molecular interactions. a significant list of selected compounds for the separation of their racemic mixtures is presented, as well as a large reference list. the next chapter presents the importance of choosing the optimal solid form of a drug, an unavoidable topic for this type of book. it includes the thermodynamics of http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:vesna.gabelicamarkovic@glpg.com admet & dmpk 1(2) (2013) 17-18 book review doi: 10.5599/admet.1.2.6 18 different forms, as well as form screening and selection. thermal analysis and calorimetric methods, which are basic methods currently used in the pharmaceutical industry, are detailed in chapter 8. this comprehensive chapter includes theory and instrumentations, followed by a few examples of application. the last chapter discusses the importance of near infrared spectroscopy in pharmaceutical research and development. in addition to a theoretical description, near infrared spectroscopy is presented as the technique of the greatest significance for process analytical technology. the strength of this book lies in its excellent review of numerous physicochemical methods. overall, the writing style is very easy to follow and the chapters are illustrated where appropriate. it is a very valuable book for anyone interested in physicochemical methods in drug discovery and development. the book is excellent for students and will become an essential reference for any individual working in the pharmaceutical industry. © 2013 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ manuscript doi: 10.5599/admet.2.4.42 221 admet & dmpk 2(4) (2014) 221-234; doi: 10.5599/admet.2.4.42 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper potential of imaging analysis in establishing skin concentrationdistance profiles for topically applied fitc-dextran 4 kda shosho kijima 1 , ryosuke masaki 1 , wesam r. kadhum 1 , hiroaki todo 1 , tomomi hatanaka 2 , kenji sugibayashi 1,  1 faculty of pharmaceutical sciences, josai university, 1-1 keyakidai, sakado, saitama 350-0295, japan 2 institute of innovative science and technology, medical science division, tokai university, 4-1-1 kitakaname, hiratsuka, kanagawa 259-1292, japan corresponding author: e-mail: sugib@josai.ac.jp; tel.: 049-271-7367; fax: 049-271-8137 received: may 26, 2014; revised: july 06, 2014; published: january 09, 2015 abstract quantitatively determining the skin concentration-distance profiles of topically applied drugs is important for evaluating their safety and efficacy. the aim of the present study was to quantitatively visu alize the distribution of hydrophilic drugs through the skin using confocal laser scanning microscopy (clsm) in order to obtain skin concentration-distance profiles. fitc-dextran with a molecular weight of approximately 4 kda (fd-4) was selected as the model fluorescent drug in the present study, and excised pig ear skin was used. the skin concentration of fd-4 at each depth of a skin section was assessed by imaging analysis of the intensity of fluorescence. the fd-4 skin concentration-distance profile obtained was analyzed using fick’s second law of diffusion, and was markedly similar to that using skin permeation parameters in the skin permeation study. these results suggest that the present clsm method may be a promising tool for quantitatively visualizing the concentration-distance profiles of drugs through the skin. keywords confocal microscopy; fitc-dextran, dermatopharmacokinetics, skin permeation introduction drug delivery through the skin remains an attractive and challenging area for research and development. most of the drugs contained in transdermal delivery systems are lipophilic with low molecular weights. however, various chemical and physical approaches have been used to overcome the barrier function of the stratum corneum so that even malabsorptive hydrophilic and high weight molecular compounds can diffuse into the deeper regions of the skin. an assessment of the chemical concentrations of drugs in the skin is essential for developing topical pharmaceutical products and functional cosmetics; therefore, drug disposition in the skin (stratum corneum, viable epidermis/dermis, and appendages) as well as the concentration-distance profiles of these compounds through the skin need to be elucidated in more detail [1-4]. since the skin, which is a structurally heterogeneous organ, is not uniformly distributed for topically applied drugs, it remains unclear how drugs permeate the skin layers from the surface to deeper regions http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:sugib@josai.ac.jp mailto:sugib@josai.ac.jp sugibayashi et.al. admet & dmpk 2(4) (2014) 221-234 222 over time. the skin is an extremely thin organ; thus, as a small amount of a drug may result in a high target concentration, a sensitive method is needed to determine the drug concentration in each layer of the skin after its topical application. several conventional methods, including the suction blister [5], punch and shave biopsy [6], and homogenization [7] methods, have been used to measure the concentration of a drug in the skin. however, these methods can only measure average drug concentrations in the total area of excised whole skin instead of those at each depth of the skin. furthermore, these methods cannot provide a visual observation of drug disposition from the surface to deeper regions of the skin, which is necessary for further understanding of the dermatopharmacokinetics of topically applied drugs. the tape-stripping method [8-12] has been used to measure drug concentrations at each depth of the skin, but only in the stratum corneum. a clearer understanding of drug disposition through the whole skin will be helpful for evaluating their safety and efficacy by specifying the skin tissue targeted by these drugs. however, the drug concentration at each depth of the skin has been predicted using various mathematical equations [13-18]. we previously reported that the cutaneous disposition of topically applied drugs could be predicted using the difference method based on fick’s 2 nd law of diffusion, and also that the average drug concentration in the stratum corneum and viable epidermis/dermis could be determined separately [19-22]. however, fick’s law-based approach requires a relatively advanced knowledge of mathematical equations in order to estimate the cutaneous disposition of topically applied drugs. figure 1. chemical structure of fd-4 therefore, the aim of the present study was to develop a quantitative, visual, simple, sensitive, and effective method to analysis concentration-distance profile of topically applied compounds using confocal laser scanning microscopy (clsm). the clsm could be used to quantitatively visualize disposition of fluorescent compounds or high molecular weight compounds with fluorescent labeling moiety through the skin from the viewpoint of depth. in the present study, the mathematical approach represented by the difference equation method based on fick’s law of diffusion to obtain a concentration-distance profile was also used for comparisons to the clsm technique. we applied fitc-dextran (fig. 1) having a molecular o h oh h o h ohh oh ch2 h o h o h o h ohh oh h ch2 cs nh oho oh o c o o h o h o h ohh oh ch2 h o h oh h o h ohh oh h ch2 cs nh o h oh h o h ohh oh ch2 h o h oh h o h ohh h ch2 o h oh h o h ohh oh ch2 h o h oh h o h ohh oh h ch2 n admet & dmpk 2(4) (2014) 221-234 imaging analysis of topically applied fitc-dextran doi: 10.5599/admet.2.4.42 223 weight of about 4 kda (fd-4) as a candidate to excised skin from the pig ear in which the stratum corneum had been removed by stripping. dermatodynamics could be closely related to drug concentration in viable epidermis and dermis not in the stratum corneum after topical application. thus, we selected the stripped skin to investigate the possibility of quantitative determination of fd-4 concentration in viable epidermis and dermis. concentration-distance profile of fd-4 through the skin was quantified by imaging analysis (clsm) over time. theoretical [22-25] stripped skin excised from the pig ear was assumed to be a homogeneous membrane [26]. the present chemical compound permeation study through the skin was performed at an infinite dose under a sink condition in the receiver compartment. figure 2 illustrates the typical concentration-distance profile of a chemical compound across stripped skin in the present study. membrane d c ・ c k d x = l x = 0 c donor receiver t1 steady state t2 t3 t5 figure 2. schematic diagram of a concentration-distance profile in stripped pig ear skin (one-layered diffusion model) the concentration of the chemical compound, cs, in the stripped skin at a position, x, and time, t, was expressed by fick’s 2 nd law of diffusion as follows: 2 2 s s s c c d t x      (1) where ds is the effective diffusion coefficient of the chemical compound in the stripped skin. the initial condition (i.c.) and boundary condition (b.c.) in the present study were as follows: i.c. t = 0 0 < x < l cs = 0 b.c. t > 0 x = 0 cs = kcv x = l cs=0 d s d s dc dc v ad dt dx  (2) sugibayashi et.al. admet & dmpk 2(4) (2014) 221-234 224 where l is the thickness of the stripped pig ear skin (1000 μm), k is the partition coefficient of the chemical compound from the donor solution to the stripped skin, cv is the chemical compound concentration in the donor solution, vd is the volume of the donor solution, and a is the effective permeation area of the skin. the left and right members in eq. (1) can be expressed by the following difference formulations [27]:   , , 1 , 1i j i j i j dc c c dt t    (3)   2 , 1, , 1,2 2 d 2 d i j s s i j i j i j c d d c c c x x      (4) where cij is the concentration of the chemical compound at the i-th position and j-th time. δx is xi+1-xi and δt is tj+1-tj. by substituting eqs.(3) and (4) into eq. (1), the following equation was obtained.   , 1 1, , 1, 1 2 i j i j i j i js s s s s s s c rd c rd c rd c        (5) where r shows ∆t/∆x 2 . the concentration of fd-4 in the stripped pig ear skin was obtained by a curve fitting the cumulative amount of fd-4 that permeated through the stripped skin to the theoretical values using a least squares method. the least squares calculation was performed using the spreadsheet software microsoftⓡ excel and ∆t and ∆x were set where dsr may become 0.5 or less. experimental materials and methods fd-4 was obtained from sigma-aldrich (st. louis, mo, u.s.a.). the cellophane tape (packaging tape, series 405) used for tape stripping was obtained from nichiban co., ltd. (tokyo, japan). other reagents and solvents were of reagent or hplc grade and used without further purification. experimental animals the edible ears of three yuan pigs were obtained from saitama experimental animals * (sugito, saitama, japan). animal experiments were approved by the institutional animal care and use committee of josai university (sakado, saitama, japan). skin extraction and pretreatment the skin used in the fd-4 permeation experiment was carefully washed and wiped well with distilled water, and the surface of the stratum corneum surface was tape-stripped 50 times using cellophane tape to remove the whole layer of the stratum corneum. one thousand-micrometer-thick stripped pig ear skin slices were prepared using a dermatome (acculan 3ti dermatome, aesculap-a b. braun company, melsungen, hessen, germany) to separate the upper epidermis from the dermis and subcutaneous tissues [28]. * now, all the business was entrusted to the tokyo laboratory animals (tokyo, japan). admet & dmpk 2(4) (2014) 221-234 imaging analysis of topically applied fitc-dextran doi: 10.5599/admet.2.4.42 225 preparation of pig ear skin pieces for the calibration of fd-4 concentrations to prepare a calibration curve for the quantitative analysis of fd-4, excised stripped pig ear skin pieces were cut into an appropriate size and completely immersed in solutions with different concentrations of fd-4 (0.05, 0.50, 1.00, 2.50, 4.00, and 5.00 mm). as a control, skin pieces were immersed into ph 7.4 phosphate-buffered saline (pbs). all skin pieces were removed after 24 hours, embedded in super cryoembedding medium (section-lab, hiroshima, japan), frozen in isopentane cooled with dry ice to -20 °c, and then stored at -30 °c before further use. skin section observations using a confocal laser scanning microscope (clsm) ten-micrometer-thick skin sections were prepared in a vertical direction by a cryostat (cm3050, leica, wetzler, hessen, germany) using the skin pieces obtained as described in (3.4). the intensity of fluorescence was observed in these skin sections by clsm (fluoview fv1000 and software: fv10-asw, olympus, tokyo, japan). the clsm conditions for fd-4 were as follows: wavelength, 473 nm; scan speed, 200 μs/pix; laser power, 1 %; high voltage, 236 v; gain, 1.125 x; and offset, 0. calculation of fluorescence intensity for fd-4 from clsm images analysis points (50 points) were selected to calculate the fluorescence intensity for fd-4 from one skin section, and an average value was obtained from these points. figure 3 shows an exemplified plot. the intensity of autofluorescence was obtained separately from an image of a skin section immersed in pbs alone and was then used to determine the actual fluorescence intensity for each concentration of fd-4. 200 μm figure 3. example of a clsm observation image for the calibration of fd-4 fluorescence intensity in stripped pig ear skin stratum corneum side dermis side sugibayashi et.al. admet & dmpk 2(4) (2014) 221-234 226 skin permeation experiments for fd-4 stripped skin samples were set in vertical-typed diffusion cells (effective diffusion area: 1.77 cm 2 ) with the epidermis side facing the donor compartment. pbs (ph 7.4, 1 and 6 ml) was applied to the donor and receiver cells, respectively, for 1 h prior to starting the skin permeation experiment in order not only to reduce the influence of autofluorescence leaking into the receiver side, but also to hydrate the skin. after recovering pbs from the epidermis side, the same volume of 5 mm fd-4 was added to start the permeation experiment. skin permeation experiments were performed at 32 °c over 5 h, and the receiver solution was continuously stirred with a star-head-type magnetic stirrer. at the predetermined times, an aliquot (0.5 ml) was withdrawn from the receiver cell and the same volume of pbs was added to the compartment to keep the volume constant. the concentration of fd-4 in the sample was measured by a fluorescence spectrophotometer (rf-5300, shimadzu corporation, kyoto, japan). determination of fd-4 the sample solution of fd-4 obtained from the permeation experiment was centrifuged (18,800 ×g, 4 ˚c, 5 min) and 300 μl of the supernatant was used for determination at excised and fluorescence wavelengths of 490 and 520 nm, respectively. skin treatment after the skin permeation experiment for fd-4 skin sectioning was only performed after the permeation experiment for fd-4. after the permeation experiment, the skin surface was washed three times using 1 ml of pbs on the franz-type diffusion cell. the skin sample taken from the diffusion cell was cut vertically using a razor blade, embedded in super cryoembedding medium, and frozen in isopentane cooled with dry ice. skin slices (10 μm in thickness) were made using a cryostat and observed under clsm. the clsm conditions have already been described above. image analysis of skin sections using clsm skin section images were obtained by clsm. six points under one line were plotted from the epidermis side (x = 0) onto these images to calculate the fluorescence intensity. four lines were then measured to determine the intensity of fluorescence. the average intensity of fluorescence was calculated from the same depth position of the plot. figure 4 shows a plotting example. the average fluorescence value at each depth was obtained by subtracting the average autofluorescence values. analysis of statistics the calibration curve for the concentration of fd-4 in skin sections using clsm was analyzed by pearson’s correlation coefficient (r2) to determine whether a linear relationship existed between the intensity of fluorescence and concentration of fd-4 in the skin sections. admet & dmpk 2(4) (2014) 221-234 imaging analysis of topically applied fitc-dextran doi: 10.5599/admet.2.4.42 227 figure 4. example of a clsm observation image for fd-4 fluorescence intensity in stripped pig ear skin after the permeation experiment. dashed line: boundary of the skin results clsm observations of pig ear skin sections immersed in solutions with various concentrations of fd-4 stripped pig ear skin was observed by clsm after the application of solutions with different concentrations of fd-4. figure 5 shows clsm images of pig ear skin sections after their immersion in solutions with various concentrations of fd-4. figure 5g shows the control (the immersed concentration of fd-4 is 0). elevations were observed in the intensity of fluorescence in the image with increases in the immersed concentration of fd-4 (fig. 5a→f). calibration curve of fd-4 in skin sections observed by clsm figure 6 shows a calibration curve of the average intensity of fluorescence due to fd-4 in stripped skin as obtained from clsm images (fig. 5). a strong correlation (r2=0.99636) was observed between the intensity of fluorescence in the skin and concentration of fd-4 applied. stratum corneum side dermis side sugibayashi et.al. admet & dmpk 2(4) (2014) 221-234 228 200 μm 200 μm 200 μm 200 μm 200 μm 200 μm 200 μm figure 5. clsm images for fd-4 fluorescence intensity in stripped pig ear skin after immersions in solutions with different concentrations of fd-4. a) 0.05 mm (0.18 mg/ml), b) 0.5 mm (1.83 mg/ml), c) 1 mm (3.67 μg/ml), d) 2.5 mm (9.17 mg/ml), e) 4 mm (14.68 mg/ml), f) 5 mm (18.35 mg/ml), g) 0 mm (control). dashed line: boundary of the skin y = 0.0745x + 5.9233 r² = 0.99636 0 400 800 1200 1600 0 5000 10000 15000 20000 f lu o r e sc e n c e i n te n si ty concentration (µg/ml) f lu o r e sc e n c e i n te n si ty o f c l s m figure 6. relationship between fd-4 fluorescence intensity observed by clsm and the concentration of fd-4 applied. each point shows the mean ± s.d. (n = 50) a) b) c) d) e) f) g) stratum corneum side dermis side admet & dmpk 2(4) (2014) 221-234 imaging analysis of topically applied fitc-dextran doi: 10.5599/admet.2.4.42 229 skin permeation and concentration-distance profiles under the infinite condition an in vitro skin permeation experiment was performed to further examine the permeation of fd-4 through stripped pig ear skin. skin permeation experiment was performed in three replicates. figure 7 shows the time course of the cumulative amount of three individual profiles after topical application of 5 mm fd-4 to the stripped pig ear skin. a typical lag time and subsequent steady state skin permeation were obtained. furthermore, the profiles obtained in three samples were similar. the concentration-distance profile in the steady state period (5 h after starting the permeation experiment) was determined by the permeation parameters of fd-4 from skin permeation profiles using the difference equation based on the fick’s 2nd law of diffusion as shown in the methods in the subsection theoretical (2). the calculated results are shown as dashed lines in figure 8. 0 100 200 300 400 500 0 1 2 3 4 5 time (h) c u m u la ti v e a m o u n t o f f d -4 ( µ g /c m 2 ) figure 7. time course of the cumulative amount of fd-4 that permeated through stripped pig ear skin. ●: sample a, : sample b, ▲: sample c two samples were selected from the time courses of the cumulative amounts of fd-4 that permeated through the skin in order to investigate the skin concentration-distance profile of fd-4 in a non-steady state (lag time period). dashed lines in figure 9 show the concentration-distance profile obtained in the non-steady state (0.5 h), and these were calculated using the skin permeation parameters obtained from the permeation profile results in figure 7. the results shown by the dashed lines in figures 8 and 9 suggested that the concentration-distance profiles of fd-4 in the stripped skin in the steady-state and before the steady-state could be determined from its permeation profiles using the difference equation based on the fick’s law. sugibayashi et.al. admet & dmpk 2(4) (2014) 221-234 230 0 2000 4000 6000 8000 10000 0 0.2 0.4 0.6 0.8 1 s k in c o n c e n tr a ti o n o f f d -4 ( µ g /m l ) distance ratio, x 0 2000 4000 6000 8000 10000 0 0.2 0.4 0.6 0.8 1 s k in c o n c e n tr a ti o n o f f d -4 ( µ g /m l ) distance ratio, x 0 2000 4000 6000 8000 10000 0 0.2 0.4 0.6 0.8 1 s k in c o n c e n tr a ti o n o f f d -4 ( µ g /m l ) distance ratio, x figure 8. concentration-distance profile of fd-4 in the steady state (5 h). symbols (○): observed values from image analysis, dashed line: calculated value by the difference equation based on fick’s law of diffusion from the skin permeation profile. each point shows the mean ± s.d. (n = 4). a), b) and c) correspond to samples a, b and c, respectively 0 2000 4000 6000 8000 10000 0 0.2 0.4 0.6 0.8 1 s k in c o n c e n tr a ti o n o f f d -4 ( µ g /m l ) distance ratio, x 0 2000 4000 6000 8000 10000 0 0.2 0.4 0.6 0.8 1 s k in c o n c e n tr a ti o n o f f d -4 ( µ g /m l ) distance ratio, x figure 9. concentration-distance profile of fd-4 in the non-steady state (0.5). symbols (□): observed value from image analysis, dashed line: calculated value by the difference equation based on fick’s law of diffusion from the skin permeation profile. each point shows the mean ± s.d. (n = 4). a) and b) correspond to samples a and b, respectively a) b) c) a) b) admet & dmpk 2(4) (2014) 221-234 imaging analysis of topically applied fitc-dextran doi: 10.5599/admet.2.4.42 231 imaging analysis by clsm and comparison of concentration-distance profiles between observed and calculated values the intensity of fluorescence in the skin was measured by clsm 0.5 and 5 h after starting the in vitro skin permeation experiment. figures 10 and 11 show the images obtained of skin sections for the distribution of fd-4 5 and 0.5 h after the skin permeation experiment (in the steady and non-steady states, respectively). the intensity of fluorescence for fd-4 on the applied surface of the stripped pig ear skin was markedly high, while that at the receiver surface was unstable; therefore, it was difficult to quantify the intensity of fluorescence in these two surfaces. these surface intensities were subsequently omitted from the fluorescence analysis. two images were randomly selected from the time courses of the cumulative amounts of fd-4. the concentration-distance profiles for fd-4 in the steady state (5 h) and non-steady state (0.5 h) were obtained from image analysis data by clsm. each point in figures 8 and 9 shows the concentration of fd-4 determined by clsm imaging. the dashed lines in these figures were determined by skin permeation data, as described in (4. 3). the results obtained by clsm were similar to those from the mathematical approach. these fluorescent gradients due to the permeation of fd-4 across the skin in both the steady state (5 h) and non-steady state (0.5 h) may be useful for visually understanding the concentration-distance profiles of chemical compounds across the skin. discussion we used clsm to quantitatively determine the skin concentration-distance profile of fd-4 at different depths through stripped skin. limitations have been associated with the conventional methods used to determine drug concentration-distance profiles in the skin. clsm could overcome these limitations because it generates high-resolution and sensitive images at multiple depths through the skin without the need for mechanical skin sections after the topical application of fluorescent drugs [29]. we applied a wellestablished mathematical approach represented by the difference method based on fick’s law of diffusion to obtain the skin concentration-distance profiles of topically applied drugs for comparisons in order to evaluate the present clsm technique. a strong correlation (r2=0.99636) was observed between the intensity of fluorescence on clsm images and the applied concentration of fd-4 (fig. 6), which indicated that the intensity of fluorescence at each depth of the skin could be quantitatively evaluated by clsm. although few studies have determined the drug concentration-distance profiles of topically applied drugs, even less attention has been given to understanding the disposition behavior of these drugs through the skin under both steady-state and non-steady state conditions. therefore, evaluating skin concentration-distance profiles is important under these conditions in order to address not only to efficacy, but also toxicological issues. sugibayashi et.al. admet & dmpk 2(4) (2014) 221-234 232 x = 0 200 μm x = ltot 200 μm 200 μm x = 0 x = ltot x = 0 x = ltot figure 10. clsm images of fd-4 fluorescence intensity in stripped pig ear skin in the steady state (5 h) after the permeation experiment. a), b) and c) show samples a, b and c, respectively. dashed line: boundary of the skin x = 0 x = ltot 200 μm x = 0 x = ltot 200 μm figure 11. clsm images of fd-4 fluorescence intensity stripped pig ear skin in the non-steady state (0.5 h) after the permeation experiment. a) and b) show samples a and b, respectively. dashed line: boundary of the skin theoretical and observed fd-4 skin concentration-distance profiles were determined in both the steady and non-steady states in the present study. the theoretical and observed fd-4 concentration-distance profiles were almost equal at the steady and non-steady states (figs. 8 and 9), which suggested that the concentration of the fluorescence compound may be precisely evaluated at multiple depths through skin using clsm. common artifacts include tearing, ripping, holes, and folding may occur during skin sectioning; stratum corneum side dermis side a) b) c) stratum corneum side dermis side x = 0 x = 0 x = ltot x = ltot admet & dmpk 2(4) (2014) 221-234 imaging analysis of topically applied fitc-dextran doi: 10.5599/admet.2.4.42 233 therefore, analysis points must be plotted to avoid such artifacts in order to calculate the fluorescent intensity for fd-4 from one skin section (fig. 3). this could explain why slightly different skin concentrationdistance values were obtained for fd-4 between clsm and the theoretical values (figs. 8 and 9). furthermore, plotted analysis points (25, 31, 37 and 43) (fig. 3) were not selected as reference points to determine skin concentration-distance profiles due to the high intensity of fd-4 at the epidermis side. plotted analysis points (30, 36, 42 and 48) (fig. 3) at the dermis side were also not selected as reference points to determine skin concentration-distance profiles because they overlapped with the receiver compartment, which may lead to unclear images by clsm. the present clsm method provided a visual observation of drug distribution at multiple depths through the skin (figs. 10 and 11) as an additional advantage over the mathematical approach. however, the intensity of fluorescence intensity was markedly high on the epidermis surface. thus, it is necessary to carefully wash the fluorescence compound that is adsorbed on the skin surface. further studies are needed to evaluate the distribution of fd-4 in skin sections cut laterally as well as through skin appendages such as hair follicles. we intend to use low molecular-weight fluorescence compounds in future studies. conclusions the results of the present study clearly demonstrated that clsm may be used to quantitatively determine the concentration-distance profiles of topically applied fluorescent drugs. these results provide a useful insight into dermatopharmacokinetics. references [1] a. naik, y.n. kalia, r.h. guy, pharm. sci. technol. today 3 (2001) 318-326. [2] h. wosicka, k. cal, j. dermatol. sci. 57 (2011) 83-89. [3] m.r. prausnitz, s. mitragotri, r. langer, drug discovery 3 (2004) 115-124. [4] t. oshizaka, h. todo, k. sugibayashi, yakugaku zasshi 132 (2012) 1237-1243. [5] g. cevc, u. vierl, j. control. release 141 (2010) 277-299. [6] o. sarheed, y. frum, int. j. pharm. 423 (2012) 179-183. [7] m. tsang, r.h. guy, br. j. dermatol. 165 (2010) 954-958. [8] u. kiistala, j. invest. dermatol. 50 (1968) 129-137. [9] c. surber, k.p. wilhelm, d. bermann, h.i. maibach, pharm. res. 10 (1993) 1291-1294. [10] l.k. pershing, b.s. silver, g.g. krueger, v.p. shah, j.p. skelley, pharm. res. 9 (1992) 45-51. [11] a. rougier, d. dupuis, c. lotte, r. roguet, h. schaefer, j. invest. dermatol. 81 (1983) 275-278. [12] v.p. shah, g.l. flynn, a. yacobi, h.i. maibach, c. bon, n.m. fleischer, t.j. franz, s.a. kaplan, j. kawamoto, l.j. lesko, j.p. marty, l.k. pershing, h. schaefer, j.a. sequeira, s.p. shrivastava, j. wilkin, r.l. williams, pharm. res. 15 (1998) 167-171. [13] m. lodén, l. ungerth, j. serup, acta derm. venereol. 87 (2007) 485-492. [14] b. n'dri-stempfer, w.c. navidi, r.h. guy, a.l. bunge, pharm. res. 26 (2009) 316-328. [15] j.e. rim, p.m. pinsky, w.w. van osdol, j. membr. sci. 20 (2007) 174-182. [16] b.d. weinberg, e. blanco, j. gao, j. pharm. sci. 97 (2008) 1681-1702. [17] y.g. anissimov, m.s. roberts, pharm. res. 28 (2011) 2119-2129. [18] l. hu, m.g. wientjes, j. li, j.l. au, aaps j. 12 (2010) 586-591. [19] k. sugibayashi, h. todo, t. oshizaka, y. owada, pharm. res. 27 (2010) 134-142. [20] h. ishii, h. todo, k. sugibayashi, chem. pharm. bull. 58 (2010) 556-561. [21] h. ishii, k. fujino, h. todo, k. sugibayashi, exp. anim. 61 (2012) 147-156. sugibayashi et.al. admet & dmpk 2(4) (2014) 221-234 234 [22] n. hada, t. hasegawa, h. takahashi, t. ishibashi, k. sugibayashi, j. control. release 108 (2005) 341350. [23] s. geinoz, s. rey, g. boss, a.l. bunge, r.h. guy, p.a. carrupt, m. reist, b. testa, pharm. res. 19 (2002) 1622-1629. [24] k. sato, n. mitsui, t. hasegawa, k. sugibayashi, y. morimoto, j. control. release 73 (2001) 269-277. [25] k. sugibayashi, t. hayashi, k. matsumoto, t. hasegawa, drug metab. pharmacokinet. 19 (2004) 352362. [26] r.j. scheuplein, i.h. blank, j. invest. dermatol. 60 (1973) 286-296. [27] k. sugibayashi, t. hayashi, t. hatanaka, m. ogihara, y. morimoto, pharm. res. 13 (1996) 855-860. [28] h. takeuchi, m. ishida, a. furuya, h. todo, h. urano, k. sugibayashi, biol. pharm. bull. 32 (2012) 192202. [29] r. alvarez-román, a. naik, y.n. kalia, h. fessia, r.h. guy, eur. j. pharm. biopharm. 58 (2004) 301316. ©2014 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ manuscript doi: 10.5599/admet.2.3.46 168 admet & dmpk 2(3) (2014) 168-178; doi: 10.5599/admet.2.3.46 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper voltammetric study of the partitioning of macrolide antibiotics at the water/nitrobenzene interface. relationship to the pharmacokinetic profiling of macrolides zoran mandić faculty of chemical engineering and technology, university of zagreb, hr-10000 zagreb, croatia corresponding author e-mail: zmandic@fkit.hr tel.: +385-1-4597-164; fax: +385-1-3733-640 received: july 20, 2014; revised: august 28, 2014; published: september 16, 2014 abstract the ion transfer of a series of ionised 14and 15-membered macrolide compounds across polarised water/nitrobenzene have been studied by cyclic voltammetry. fourteen-membered macrolides transfer as singly-charged molecules in the wide ph range, while 15-membered macrolides can transfer as both singly and doubly protonated cations, depending on the ph of the aqueous solution. the results showed that the formal transfer potentials ( 0'w o i ) of the investigated compounds varied to a great extent and that minor changes in macrolide structure caused significant change in the corresponding formal transfer potentials. among the investigated macrolides, azithromycin had the lowest formal transfer potential and its peculiar adme behaviour is ascribed to this specific property. keywords: azithromycin; erythromycin; oleandomycin; clarithromycin; liquid/liquid interface; cyclic voltammetry; pharmacokinetics; lipophilicity; adme introduction macrolides are a class of compounds that exert antibiotic activity against most gram-positive bacteria and several gram-negative respiratory and enteric pathogens [1-4]. the main building block common to all macrolides is a large and highly substituted 14-, 15or 16-membered macrolactone ring on which one to three sugar moieties such as cladinose and desosamine are usually attached. the mechanism of action of macrolides is the inhibition of rna-directed protein synthesis, usually achieved by binding to the 50s subunit of bacterial ribosomes [5,6]. erythromycin (1) is a naturally occurring compound first isolated from the saccharopolyspora erythraea, microorganism found in philippine soil (figure 1). it is the prototype 14-membered macrolide compound and almost all other 14and 15-membered macrolides are semi-synthetic derivatives of erythromycin. the expansion of the erythromycin ring by insertion of nitrogen into the 9a position leads to the 15-membered class of macrolides, widely known as azalides [7]. the most potent and widely prescribed antibiotic belonging to this class is azithromycin (2) (figure 1). in addition to its broad antibiotic spectrum against gram-positive bacteria, azithromycin has proven useful in the treatment of some gram-negative http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:zmandic@fkit.hr admet & dmpk 2(3) (2014) 168-178 standard transfer potential of macrolides doi: 10.5599/admet.2.3.46 169 respiratory pathogens such as haemophilus influenza, as well as enterobacteria such as salmonella [1,2,810]. macrolide pharmacokinetics differs from other antibiotics. they are able to penetrate a variety of cells and can achieve high intracellular concentrations. this includes increased accumulation in polymorphonuclear cells (pmn) and macrophages, which tend to concentrate at the infection sites, resulting in targeted drug delivery systems [4]. although all macrolides show similar adme properties and pharmacokinetics, azithromycin stands out due to its enhanced antimicrobial activity, a more favourable adverse reaction profile and improved pharmacokinetic properties. azithromycin concentrations are more than 200 times higher intracellularly than extracellularly [4], resulting in the extended exposure of phagocytes and other cells to the drug. it has been found that in the presence of bacteria in the extracellular medium, a rapid release of azithromycin from phagocytes occurs. as an additional advantage, azithromycin retains high and prolonged tissue levels, ensuring that pathogens are continuously exposed to high bactericidal effects [4]. as a result, azithromycin needs significantly shorter administration time and also shows better tolerance to treatment in patients [3]. such peculiar pharmacokinetics of macrolides, especially the high uptake of azithromycin in pmn and its fast release at the site of infection, is obviously connected with easy and unhindered membrane transport. in several cases, it has been proven that azithromycin transport was governed by ph differences on both sides of the membranes and that minor changes in the macrolide structure considerably altered their properties, as well as their in vivo availability [4]. the mechanism of macrolide transport is still unclear and the question remains whether the transport of macrolides through biological membranes is governed by active or passive process [4]. some evidence indicates the simultaneous coexistence of both mechanisms. both in vivo and in vitro results demonstrate unequal macrolide distribution at different biological membranes; therefore, it seems that the ph-partition hypothesis is not the only factor determining their pharmacokinetic profiles. almost all macrolides are weak bases due to the presence of tertiary amine nitrogen at the desosamine sugar (fig. 1). in addition, azalides possess another amine group as part of macrolactone ring. due to the pka values of these amine groups (pka1 ~ 8.7 and pka2 ~ 9.3), macrolides exist as charged species in the physiological and acidic ph ranges, either as mono-protonated (14-membered macrolides) or doubly protonated (azalides) forms. these properties may indicate that the uptake of macrolides is an electrochemical potential driven process, and that the electrical potential difference through biological membranes may be responsible for governing the overall mechanism of macrolide transport. if this is the case, the uptake of macrolides will depend on actual membrane potential, standard transfer potential at the membrane/water interface and the difference between them. in order to do in vitro testing and determine whether membrane electrical potential plays an important role in macrolide pharmacology, an appropriate interface model mimicking the in vivo membrane environment must be designed. it has already been proven that liquid/liquid interfaces formed between two immiscible liquids, such as water and an appropriate organic solvent, can be used as a model of biological membrane [11,12]. in such cases, a tested compound is dissolved in the system and the interface is electrochemically polarised, resulting in an ionic species transfer across the interface and its redistribution within the two phases. from such measurements, the transfer ability of a compound between two phases is represented by a physico-chemical parameter referred to as standard transfer potential,  0 , which can be easily determined using electrochemical measurements. the standard transfer potential of the particular compound is the interfacial potential difference, which will yield equal z. mandić admet & dmpk 2(3) (2014) 168-178 170 compound activities in both phases [13]. consequently, the distribution of charged species between two phases will conform to the nernst-like equation: 0 0'ln ln o o w w wi i o o i o iw w i i a crt rt zf zfa c                          (1) where is potential difference between two phases, and are standard transfer potential and formal transfer potential, respectively, of the ionised compound i at the organic solvent/water interface, a and c are activity and concentration of the ionised compound in the corresponding solvent, respectively, and r,t,z and f have their usual thermodynamic meaning. solvents having acceptable dielectric constants for dissolving electrolytes suitable for electrochemical measurements and which are usually used for mimicking biological membranes in electrochemistry are nitrobenzene and 1,2-dichloroethane. both solvents have properties similar to those of biological membranes and relatively high dielectric constants to enable sufficient dissolution of inert electrolytes, and to ensure electric conductivity suitable for performing the electrochemical measurements at their interface with aqueous phase. the interfaces nitrobenzene/water, as well as 1,2-dichloroethane/water are good models for biological membranes, since they enable the rapid and precise determination of the electrochemical properties of ionised biological molecules. both interfaces have already been used to probe the transfer of ionic drug compounds [14-18]. in the present work, we investigated the electrochemical properties of a series of macrolide compounds at the nitrobenzene/water interface. the structures of the investigated compounds are shown in the fig.1. the objectives were to measure the standard transfer potentials of macrolide compounds, to determine the mechanisms of their transport at the interface, and to provide some evidence on the importance of in vivo membrane electrical potential on the pharmacokinetics of macrolides. the value of standard transfer potential is correlated with the free energy of transfer of ionized species, δg=-zfδφ. lower the absolute value of standard transfer potential of the molecule becomes, higher is its negative free energy, and the molecule crosses the interface easier. experimental erythromycin (1), clarithromycin (3) and oleandomycin (9) were purchased from aldrich. azithromycin (2) was obtained from pliva d.d. as dihydrate, a 3-decladynosil derivative of azithromycin (4) 9a-aza-9-deoxo-9-dihydro-9a-homoerythromycin a (dhl) (5) and its 3-decladinosyl derivative (6), 3’-ndemetyl-n-methyl-9a-aza-9-deoxo-9-dihydro-9a-homoerythromycin a (7), 3’-n-demetyl-9a-aza-9-deoxo-9dihydro-9a-homoerythromycin a (8) were synthesised in pliva d.d. from azithromycin as a starting compound, according to the procedure described previously [7], and with a purity higher than 95 % as determined by hplc analysis. admet & dmpk 2(3) (2014) 168-178 standard transfer potential of macrolides doi: 10.5599/admet.2.3.46 171 o oh o ch 3 o o o o oh o o oh n o ch 3 ch 3 or o o n or 3 oh r1 oh oh o o oh n ch 3 r2 o oh o ch 3oh o o o o o o oh oh o n o ch 3 ch 3 ch 3 o 2: r1=r2=ch 3 ; r3=cld 4: r1=r2=ch3; r3=h 5: r1=h; r2=ch3; r3=cld 6: r1=h; r2=ch3; r3=h 7: r1=ch3; r2=h; r3=cld 8: r1=r2=h; r3=cld 1: r = h 3: r=ch 3 cld = 9 figure 1: the structures of antibiotics investigated in this work: erythromycin (1), azithromycin (2), clarithromycin (3), decladinosil derivative of azithromycin (4), dhl (5), decladinosil derivative of dhl (6), 3’-n-demethylated derivative of azithromycin (7), 3’-n-demethylated derivative of dhl (8), oleandomycin (9). the transfer of ionised antibiotics was studied using cyclic voltammetry at the water/nitrobenzene interface and with an interfacial area of 0.12 cm 2 . the investigated antibiotics were dissolved in a aqueous phase that consisted of a britton-robinson buffer of varying ph. the concentration of antibiotics was 2 mm. tetrabutylammonium tetraphenylborate (bu4nph4b) was used as a supporting electrolyte in the nitrobenzene phase. the cell can schematically be shown as: ag/agcl (0.1 m licl) 0.1 m br buffer, varying ph 0.05 m bu4nbph4 0.1 m bu4ncl agcl/ag (aqueous phase) (aqueous phase) (nitrobenzene) (aqueous phase) four electrode potentiostat (par m263a) with positive feedback ir compensation was used for performing electrochemical measurements. a platinum sheet and ag/agcl electrode served in each phase as counter and reference electrodes, respectively. the silver wire for the nitrobenzene solution was immersed in a 0.1 m aqueous solution of tetrabutylammonium chloride (bu4ncl). formal transfer potentials of the investigated antibiotics were determined by measuring their half-wave potentials, . the experimentally obtained values were transposed to the galvani potential scale by measuring the half-wave potential of tetramethylammonium ion (me4n + ), for which absolute galvani potential at water/nitrobenzene is known [19]. all potentials quoted in this paper are potentials of aqueous phase with respect to the nitrobenzene phase. z. mandić admet & dmpk 2(3) (2014) 168-178 172 ionization constants, pka, and log p values were determined by potentiometric titration method (sirius analytical, glpka). all measurements were carried out at constant temperature of 25 °c, constant ionic strength of 0.15 m kcl and with continuous flow of argon to prevent the absorption of co2 from the atmosphere. the ph-electrode was standardized using sirius four-plustm procedure. results and discussion electrochemical properties of the investigated macrolides and their distribution at water/nitrobenzene interface depend on their ph dependent speciation. fourteen-membered macrolides, such as erythromycin (1) and clarithromycin (3) possess a tertiary amine group with a pka value of 9.3 ± 0.05, making this class of macrolides weak bases existing as singly charged forms in a neutral and acidic ph range. macrocyclic amine groups of azalides introduce additional nitrogen into 15-membered macrolides with an ionisation constant measured to be around pka=8.7. since these ionisation constants present global values and no microionisation data were available in the literature, the exact speciation of the azalide antibiotics, especially in a slightly basic solution, is impossible to discern. transfer properties of the investigated macrolides, as well as their formal transfer potentials, were investigated by cyclic voltammetry as a function of aqueous ph. as was expected, the results could be grouped into two categories, depending on the number of basic nitrogen atoms in the molecule. -0.20 -0.15 -0.10 -0.05 0.00 0.05 0.10 0.15 0.20 -40 -20 0 20 40 60 i /    o w / mv azithromycin erythromycin supporting electrolyte figure 2: cyclic voltammograms of erythromycin and azithromycin at a ph of aqueous phase 5.8. scan rate: 20 mv/s. the comparison of the cyclic voltammograms of azithromycin and erythromycin taken at aqueous phase ph=5.8, where both compounds exist in fully protonated forms, is shown in fig. 2. only one reversible, diffusion-controlled wave was observed for each compound at this ph value. in an anodic potential direction, a transfer of protonated forms from the aqueous to nitrobenzene phase took place, giving rise to positive currents, while negative currents in the cathodic potential sweep represented the transfer of the same species back from the nitrobenzene to aqueous phase. the linearity of the current peak height with the square root of the scan rate revealed diffusion control of the transfer process. it was found that diffusion coefficients calculated using randles-ševčik equation depended on the actual ph of the aqueous phase. for azithromycin diffusion coefficients in aqueous phase ranged from 1.3 x 10 -5 cm 2 s -1 at ph 2 to 6.8 x 10 -6 cm 2 s -1 at ph 8.6. this can be explained by the complex equilibrium of ionized species at the admet & dmpk 2(3) (2014) 168-178 standard transfer potential of macrolides doi: 10.5599/admet.2.3.46 173 interface as already described by velický et al. [20]. no kinetically slow steps were detected in the cyclic voltammograms. anodic and cathodic current peak separations being close to 30 and 60 mv for azithromycin and erythromycin, respectively, confirmed doubly-protonated azithromycin and singlyprotonated erythromycin to be the predominant species at acidic phs and were involved in the interfacial transfer reactions: azh2 2+ (w)  azh2 2+ (nb) (2) erh + (w)  erh + (nb) (3) the half-wave potentials determined by these measurements were related to the formal transfer potentials of monoprotonated erythromycin molecules and diprotonated azithromycin molecules according to the following equation [21]: 0' 1/2 ln o w w i o o i w i drt zf d      (4) where and are diffusion coefficients of transferring species in nitrobenzene and aqueous phases, respectively. using equation (4), the formal transfer potentials of erythromycin and azithromycin were calculated assuming the proportionality of diffusion coefficients with the viscosities of water and nitrobenzene (walden rule). they amounted to 0'w o erh   = -58 mv and 2 2 0'w o azh   = -46 mv for erythromycin and azithromycin, respectively. no changes in the cyclic voltammograms of azithromycin and erythromycin were observed in the ph range 2-6. above ph 6, cyclic voltammograms for erythromycin shifted toward more positive potentials with a slope close to 60 mv/ph, while retaining its shape and reversibility. however, the electrochemical behaviour of azithromycin proved more complex. between ph 6-7, a second current wave appeared at more positive potentials, whereas the height of the first wave decreased (fig. 3a). above ph 7.5, the first current wave completely disappeared and the second wave shifted toward more positive potentials (fig. 3b), with the slope varying from 30 to 60 mv/ph (figure 4). -0.15 -0.10 -0.05 0.00 0.05 0.10 0.15 -40 -20 0 20 40 o w / v ph=6.9 ph=6.5 ph=6.0 ph=5.3 i /  a (a) figure 3. cyclic voltammograms of azithromycin at various aqueous phase ph. actual phs indicated on the graphs. scan rate = 20 mv/s. -0.15 -0.10 -0.05 0.00 0.05 0.10 0.15 0.20 -40 -20 0 20 40 i /  a  o w / v ph=7.9 ph=9.6 ph=8.9 ph=7.5 ph=6.9 (b) z. mandić admet & dmpk 2(3) (2014) 168-178 174 5 6 7 8 9 -0.06 -0.04 -0.02 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 ~60 mv/ph ~60 mv/ph   o w / v ph ~30 mv/ph figure 4: ph dependence of the transfer potentials of erythromycin (squares) and azithromycin (circles). these results indicate that in contrast to erythromycin molecules, the mechanism of azithromycin transfer changed when moving from an acidic towards a basic solution. it appears reasonable to assume that when aqueous solution ph approaches the neutral and slightly basic region, azithromycin will start to deprotonate and as neutral molecules redistribute between the two phases. in such a case, a neutral form of azithromycin will be transferred to nitrobenzene as a potentially independent process and will act as a protonophore, facilitating proton transfer from the aqueous to nitrobenzene phase according to the equation (5): az(nb) + h + (w)  azh + (nb) (5) the anodic and cathodic peak separation in basic solutions is around 60 mv indicating the transfer of singly protonated species. taking into account two protonation sites of azithromycin, with global pka values similar to one another (δpka ~ 0.5), it is possible for both monoprotonated azithromycin microspecies to transfer simultaneously from aqueous to nitrobenzene phase within the investigated ph range. lipophilicities of both protonated micro-forms are expected to be the same, since no evidence could be found in the literature of the potential conformational changes induced by the positive charge on either protonation site of the azithromycin molecule. consequently, the transfer potentials of these two monoprotonated species should not differ by a considerable amount and could not be discerned in the cyclic voltammetry measurements. formal transfer potential of the mono-protonated azithromycin molecule can be estimated using the extrapolation of 1/2 w o vs. ph data selected in the ph range 8-10 (see fig. 4) [22]:  0'1/2 2.303 logw w nb wo o az aazh rt p pk ph f         (6) admet & dmpk 2(3) (2014) 168-178 standard transfer potential of macrolides doi: 10.5599/admet.2.3.46 175 where log nb azp is the partition coefficient between water and nitrobenzene and w apk is the ionisation constant of mono-protonated azithromycin in water. from the experimental points indicated in fig. 4 and approximating log nb azp to the known partition coefficient of azithromycin in octanol, log oct azp , the value of the formal transfer potential of mono-protonated azithromycin was calculated to 0'wo azh  = -130 mv. the two macrolides described above, erythromycin and azithromycin, are prototype compounds representing 14and 15-membered macrolides. all other investigated macrolides showed electrochemical behaviour similar to either one of these two compounds, depending on which class of macrolides they belonged to: 14-membered macrolides with one protonation site or 15-membered macrolides with two protonation sites. thus, clarithromycin (3) and oleandomycin (9) exhibited one reversible current wave, representing the transfer of their protonated forms and the behaviour of other five azalide compounds, while 4-8 was consistent with the azithromycin transfer exhibiting two different types of transfer mechanism. however, the formal transfer potentials among these compounds vary, indicating that different energies are involved in the transfer processes. the values of formal transfer potentials, 0'w o i , for mono-protonated 14-membered macrolides and for both monoand di-protonated azalides are shown in table 1. table 1: formal transfer potentials of investigated compounds. macrolides log p 0'w o i / mv monoprotonated diprotonated 15membered azithromycin, 2 3.98 -130 * -46 dhl, 5 3.63 -90 * -4 3’-n-demethyl azithromycin, 7 3.62 -90 * -5 3’-n-demethyl dhl, 8 3.45 -60 * +24 3-decladinosil azithromycin, 4 3.11 -65 * +18 3-decladinosil dhl, 6 3.03 -35 * +45 14membered erythromycin, 1 3.02 -58 claritromycin, 3 3.24 < -100 oleandomycin, 9 2.76 -48 *calculated by equation (6) with the assumption of equal water/nitrobenzene and water/octanol partition coefficients and with pka=9 as the average value of two ionisation constants. macrolides' ability to pass through biological membranes and to accumulate in high concentrations intracellularly are of paramount importance for their pharmacokinetics, as proven by the significant amount of in vivo and in vitro data accumulated on the subject over the years [4]. in order to reach the site of the infection, macrolides have to cross biological membranes several times. upon administration, macrolides dissolve in the stomach due to the environment's favourable ph value (ph = 1-2) for the dissolution of weak bases. after passage of macrolides into the duodenum and further to the jejunum, where ph microclimates range from 5-7 [23], macrolides still predominantly remain in their charged state. from there, they are able to enter the bloodstream by passive diffusion, either transcellularly through the bilayer surface of epithelial cells or paracellularly. both neutral and charged forms can participate in z. mandić admet & dmpk 2(3) (2014) 168-178 176 passive diffusion. neutral forms can distribute according to the ph-partition hypothesis, while charged molecules can distribute according to the difference between the actual membrane potential and their formal transfer potential 0'wo i , as predicted by equation (1). it is assumed that weak bases such as macrolides are concentrated in the more acidic parts of membranes. this behaviour has already been experimentally observed several times in vivo [4] and was found to be the key process determining macrolide pharmacokinetics. however, it remains puzzling as to why azithromycin, compared to other macrolides, can achieve and retain very high (up to 200 times) intracellular concentrations, especially within polymorphonuclear cells such as neutrophils. being highly motile, neutrophils are quickly attracted to the site of an infection, where azithromycin is rapidly released. the whole system represents a very efficient drug delivery system. consequently, azithromycin is administered in patients in single daily doses during no more than two to three days. this is highly advantageous compared to treatment with other macrolides such as erythromycin and clarithromycin, which might achieve similar remedial effects only after a week or 10 days when administered in double or triple daily doses. high azithromycin accumulation might be partially the result of the involvement of active transmembrane transport, but it nonetheless seems plausible to assume that transmembrane electrical potential is the key factor in macrolide pharmacokinetics. indeed, when the data given in table 1 are analysed, it turns out that azithromycin has the lowest formal transfer potential among the investigated macrolide compounds. 0'w o i of a singly-charged molecule of azithromycin, azh + , is -130 mv and even 72 and 82 mv lower than 14-membered mono-protonated erythromycin and oleandomycin, respectively. the lowest 0'w o i means that charged azithromycin has the lowest free energy of transfer to and through the membrane. although the situation at the biological membrane is more complex due to the involvement of the two membrane interfaces with two membrane/water potential differences, the concept of the distribution of the charged molecules on both sides of the membrane remains valid and follows the equilibrium given by equation 1. depending on the difference between standard (or formal) transfer potential and the actual membrane potential, the distribution of different molecules at both sides of the membrane might vary to a significant extent. it is worth noting here that the treatment of neuthrophiles with ionophores such as gramicidin resulted in the reduction of azithromycin uptake in neutrophils [24]. ionophores increase the permeability of biological membranes to the ionic species dissipating their concentration gradients and consequently reducing the transmembrane electrical gradient. although decreased uptake of azithromycin in gramicidin treated neutrophils might be the result of the levelling off ph gradient across the membrane, it might point out also to the importance of the membrane potential on the pharmacokinetics of azithromycin [24]. a comparison of azithromycin to other investigated azalides highlights the importance of tertiary amine groups in the transfer process. the results were consistent with earlier observations that minor changes in the macrolide structure significantly affected their chemical and biological properties. the favourable influence of the two methyl groups and cladinose sugar moiety on the lipophilic and pharmacokinetic behaviour of azithromycin was demonstrated by the increase of the formal transfer potentials after their elimination from the macrolide scaffold. each methyl group on the amine contributed about 3.9 kj mol -1 and cladinose sugar even with 6.3 kj mol -1 in lowering free energy of transfer of azithromycin. this could not be explained by their more lipophilic contribution to the overall lipophilicity of admet & dmpk 2(3) (2014) 168-178 standard transfer potential of macrolides doi: 10.5599/admet.2.3.46 177 the azithromycin, at least, not in a significant extent. it is more likely that outstanding membrane transfer properties of azithromycin are connected to its ability to form two conformations, referred to as folded-in and folded-out conformations [25,26]. the fraction of each conformation depends on the polarity of its environment and it is possible that azithromycin undergoes conformational change after having been transferred from one phase into another. conclusions cyclic voltammetry at the water/nitrobenzene interface has been demonstrated as an important technique for evaluating the influence of the relevant physico-chemical property, 0'w o i , as it concerns the pharmacokinetics of macrolides. all macrolides studied in this paper transferred from the aqueous to the nitrobenzene phase at a sufficiently high interfacial potential. half-wave potential of the 14-membered macrolides shifted toward more positive potentials by 60 mv/ph, while 15-membered azalide compounds exhibited more complex behaviour. above ph 7, two mono-protonated species were involved in the transfer. azithromycin exhibited the lowest standard transfer potential among all macrolides, which is consistent with its outstanding in vivo behaviour and exceptional pharmacokinetics. these results show that cyclic voltammetry can be used in pharmaceutical research and development as a fast and simple method for screening out macrolide compounds with potentially good adme properties. acknowledgements: almost all work and results presented in this paper were done during my work in pliva d.d. i am very grateful to vesna gabelica marković, suzana žalac and ognjen čulić for their wholehearted support. i am also grateful to other colleagues and staff of pliva who helped in this work. references [1] g.m. bright, a.a. nagel, j. bordner, j. antibiot. 41 (1988) 1029-1047. [2] j. retsema, a. girard, w. schelkly, m. manousos, m. anderson, g. bright, r. borovoy, l. brennan, r. mason, antimicrob. agents chemother. 31 (1987) 1939-1947. [3] s. alvarez-elcoro, m.j. enzler, mayo clinic proceed., 74 (1999) 613-534. [4] w. schönfeld, h.a. kirst (eds.), macrolide antibiotics, birkhäser verlag, basel-boston berlin, 2002. [5] h.a. kirst, g.d. sides, antimicrob. agents chemother. 33 (1989) 1413-1418. [6] r.c. goldman, s.w. fesik, c.c. doran, antimicrob agents chemother. 34 (1990) 426-431. [7] s. djokić, g. kobrehel, g. lazarevski, n. lopotar, z. tamburašev, j. chem. soc. perkin trans. i, (1986) 1881-1890. [8] r.m. rakita, k. jacquespalaz, b.e. murray, antimicrob. agents chemother. 38 (1994) 1915-1921. [9] d. gur, y. kayaokay, g. hascelik, h.e. akalin, j. chemother. suppl. 5 (1993) 151-152. [10] m.e. gordillo, k.v. singh, b.e. murray, antimicrob. agents chemother. 37 (1993) 1203-1205. [11] l.e. morrison, g. weber, j. biophys. 52 (1987) 367-379. [12] a.g. volkov, d.w. deamer, liquid-liquid interfaces, theory and methods, crc press, 1996. [13] p. vanysek, electrochim. acta 40 (1995) 2841-2847. [14] m. velický, a. n.j. rodgers, r. a.w. dryfe and k. tam, admet & dmpk, 2 (2014) 143-156. [15] g. bouchard, a. pagliara, g. p. van balen, p. a. carrupt, b. testa, v. gobry, h. h. girault, g. caron, g. ermondi, and r. fruttero, helv. chim. acta 84 (2001) 375-387. z. mandić admet & dmpk 2(3) (2014) 168-178 178 [16] s. m. ulmeanu, h. jensen, g. bouchard, p.a. carrupt, h. girault, pharm. res., 20 (2003) 1317-1322. [17] r. gulaboski, f. borges, c.m. pereira, m.n.d.s. cordeiro, j. garrido, a.f. silva, comb. chem. high throughput screen., 10 (2007) 514-526. [18] l.m. yudi, a.m. baruzzi, v. solis, j. electroanal. chem., 360 (1993) 211-219. [19] t. wandlowski, v. mareček, z. samec, electrochim. acta, 35 (1990) 1173-1175. [20] m. velický, k.y. tam, r.a.w. dryfe, j. electroanal. chem. 683 (2012) 94-102. [21] h.h. girault, d.j. schiffrin, electroanal. chem., 15 (1989) 1-141. [22] s.m. ulemanu, h. jensen, g. bouchard, p.a. carrupt, h.h. girault, pharm. res., 20 (2003) 1317-1322. [23] a. avdeef, absorption and drug development: solubility, permeability, and charge state, 2 nd ed. john wiley & sons, inc., 2012. [24] unpublished results. [25] j. r. everett, j. w. tyler, j. chem. soc. perkin trans. 2, (1987) 1659. [26] p. novak, z. banić tomišić, p. tepeš, g. lazarevski, j. plavec, g. turkalj, org. biomol. chem. 3 (2005) 39-47. ©2014 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ supplementray_materialfinal supplementary material s. grandoni et al. “building in-house pbpk modelling tools for oral drug administration from literature information” system-specific parameters values values of the physiological parameters used in the pbpk model described in the paper are here summarized. these values refer to a typical subject of 250 g for rats, 10 kg for dogs and 70 kg for man. rat parameters fluxes [ml/min], [21,22] volumes [ml], [21,22,23] qbrain 1.79 vbrain 1.43 qgut 11.92 vgut 6.75 qspleen 0.8 vspleen 0.5 qliver 14.6 vliver 9.15 qmuscle 24.91 vmuscle 101.8 qadipose 6.27 vadipose 16.6 qheart 4.39 vheart 0.83 qkidney 12.64 vkidney 1.83 qrestofthebody 23.1 vrestofthebody 72.36 cardiac output 89.6 vlung 1.25 vven 10.12 vart 3.38 gastrointestinal absorption model parameters volumes of intestinal segments [ml], [25] vstomach 3 v1 0.6 v2 0.66 v3 0.66 v4 0.41 v5 0.41 v6 0.41 v7 0.41 vcolon 3 ph of intestinal segments, [25] phstomach 3 ph1 7.1 ph2 7.3 ph3 7.5 ph4 7.7 ph5 7.9 ph6 8 ph7 7.4 phcolon 7.6 mrt values stomach 10 min small intestine 88 min colon 228 min rat tissue composition to apply the poulin’s methods, [35] rat tissue composition, to apply the method of rodgers, [37] rat tissues volume fraction of phospholipids, vph volume fraction of neutral lipids, vnl volume fraction of water, vw volume fraction of interstitial space adipose 0.002 0.853 0.12 0.715 bone 0.0027 0.0273 0.446 0.42 brain 0.0533 0.0392 0.788 0.162 gut 0.0138 0.0292 0.749 0.39 heart 0.0118 0.014 0.779 0.156 kidney 0.0284 0.0123 0.771 0.346 liver 0.0303 0.0138 0.705 0.159 lung 0.014 0.0219 0.79 0.484 muscle 0.009 0.01 0.756 0.115 skin 0.018 0.0239 0.651 0.462 spleen 0.0136 0.0077 0.771 0.264 plasma 0.00083 0.00147 0.96 1 erythrocytes rat tissues neutral phospholipids neutral lipids extracellular water intracellular water tissue concentration of acidic phospholipids (mg/g) adipose 0.853 0.0016 0.135 0.017 0.40 bone 0.017 0.0017 0.100 0.346 0.67 brain 0.039 0.0015 0.162 0.620 0.40 gut 0.038 0.0125 0.282 0.475 2.41 heart 0.014 0.0111 0.320 0.475 2.25 kidney 0.012 0.0242 0.273 0.483 5.03 liver 0.014 0.0240 0.161 0.573 4.56 lung 0.022 0.0128 0.336 0.446 3.91 muscle 0.010 0.0072 0.118 0.630 1.53 pancreas 0.041 0.0093 0.120 0.664 1.67 skin 0.060 0.0044 0.382 0.291 1.32 tracheobronchial surface, stb: 81.75 cm 2 [14]. hematocrit to compute the distribution: 0.46 [22]. conversion factor to obtain the in vivo estimates of the hepatic clearance: mppgl: 45 mg/g [25], hpgl: 125*10^6 cells/g [25], liver weight: 9.15 g [21]. filtration parameter to model the renal clearance: gfr: 1.31 ml/min [22]. spleen 0.0077 0.0113 0.207 0.579 3.18 thymus 0.017 0.0092 0.150 0.626 2.30 dog parameters fluxes [ml/min], [21,22] volumes [ml], [21,22,23] qbrain 21 vbrain 78 qgut 216 vgut 368 qspleen 24 vspleen 27 qliver 288 vliver 329 qmuscle 227.9 vmuscle 456.5 qadipose 34 vadipose 1380 qheart 48.3 vheart 78 qkidney 181.65 vkidney 55 qrestofthebody 246.8 vrestofthebody 1538 cardiac output 21 vlung 82 vven 675 vart 225 gastrointestinal absorption model parameters volumes of intestinal segments [ml], [25] vstomach 14.54 v1 30.54 v2 32 v3 32 v4 20.1 v5 20.1 v6 20.1 v7 20.1 vcolon 290.9 ph of intestinal segments, [25] phstomach 1.5 ph1 6 ph2 6 ph3 6 ph4 6.2 ph5 6.2 ph6 6.2 ph7 7.4 phcolon 6.5 mrt values stomach 30 min small intestine 109 min colon 9.4 h tracheobronchial surface, stb =1176 cm 2, estimated with linear regression from the rat and man bw-stb data [14]. haematocrit: 0.42 [22]. conversion factor to obtain the in vivo estimates of the hepatic clearance: mppgl 43 mg/g [25], hpgl 120*10^6 cells/g [25], liver weight 329 g [21]. filtration parameter to model the renal clearance: gfr 61.3 ml/min [22]. human parameters fluxes [ml/min], [21,22] volumes [ml], [21,22,23] qbrain 745 vbrain 1400 qgut 1046 vgut 1155 qspleen 160 vspleen 182 qliver 1578 vliver 1799 qmuscle 1055 vmuscle 28000 qadipose 310 vadipose 14994 qheart 248 vheart 329 qkidney 1179 vkidney 308 qrestofthebody 1308 vrestofthebody 10801 cardiac output 6204 vlung 532 vven 3900 vart 1300 gastrointestinal absorption model parameters volumes of intestinal segments [ml], [25] vstomach 50 v1 105 v2 110 v3 110 v4 69 v5 69 v6 69 v7 69 vcolon 1000 ph of intestinal segments phstomach 2 ph1 6 ph2 6.2 ph3 6.6 ph4 6.8 ph5 7 ph6 7.2 ph7 7.4 phcolon 7 mrt values stomach 30 min small intestine 199.2 min colon 11 h information available on human tissue composition, [35] tracheobronchial surface, stb: 8990 cm 2 [14]. haematocrit: 0.44 [22]. conversion factor to obtain the in vivo estimates of the hepatic clearance: mgppgl: 32 mg/g [s1], hpgl: 99*10^6 cells/g [s1], liver weight: 1799 g [21]. filtration parameter to model the renal clearance: gfr 125 ml/min [22]. human tissues volume fraction of phospholipids, vph volume fraction of neutral lipids, vnl volume fraction of water, vw adipose 0.002 0.79 0.18 bone 0.0011 0.074 0.439 brain 0.0565 0.051 0.77 gut 0.0163 0.0487 0.718 heart 0.0166 0.0115 0.758 kidney 0.0162 0.0207 0.783 liver 0.0252 0.0348 0.751 lung 0.009 0.003 0.811 muscle 0.0072 0.0238 0.76 skin 0.0111 0.0284 0.718 spleen 0.0198 0.0201 0.788 plasma 0.00225 0.0035 0.945 erythrocytes drug-related parameters relationships in this section the equations to calculate the drug-specific parameters are reported. absorption the henderson-hasselbalch equations to calculate the solubility at a certain ph are here reported monoprotic acids csph=sint(1+10 (ph-pka1)) (s1) monoprotic bases csph=sint(1+10 (-ph+pka1)) (s2) diprotic acids csph=sint (1+10 (-ph+pka1)+10(2ph-pka1-pka2)) (s3) diprotic bases csph=sint (1+10 (-ph+pka1)+10(-2ph+pka1+pka2)) (s4) neutals csph=sint (s5) zwitterions csph=sint(1+10 (-ph+pkaa)+10(ph-pkab)) (s6) where pkaa is the acidic pka and pkab is the basic pka. partition coefficients this subsection contains the equations needed to calculate pt:b values with the method of poulin [35,36] and of rodgers [37,38]. for the latter the equations for each chemical species are reported. poulin’s method the fractional volumes of phospholipids (vph), neutral lipids (vnl) and water (vw), required to apply the method, are reported in the species-specific parameters section. in the following p indicates plasma and t tissue. pow=10logp dow=10logd fut=1/(1+(1-fup)/fup0.5) for non-adipose tissues pt:p=[(pow(vnlt+0.3vpht)+(vwt+0.7vpht)]/[pow(vnlp+0.3vphp)+(vwp+0.7vphp)](fup/fut) (s7) for adipose tissues pt:p=[(dow(vnlt+0.3vpht)+(vwt+0.7vpht)]/[dow(vnlp+0.3vphp)+(vwp+0.7vphp)]fup (s8) to obtain the values of pt:b from the pt:p, the tissue to plasma partition coefficient, can be applied the following equation: pt:b= pt:p/bp (s9) distribution rodger’s method the volumes related to tissues composition in terms of neutral lipids (nl), neutral phospholipids (nph), extracellular water (ew), intracellular water (iw), the ratios such as the lipoprotein ratio (lr), the albumin ratio (ar) and the tissue concentration of acidic phospholipids (ap) are reported in the species-specific parameters section. in the notation, t indicates the tissue and b the blood. the values of php, phiw and phbc are fixed, as reported by the authors, to 7.4, 7 and 7.22 respectively. the values for fnlp and fnpp are fixed as 0.0023 and 0.0013 respectively, as reported in the paper. for all tissues, except adipose ones, the value p in the subsequent equations is the n-octanol:water partition coefficient (here reported as p1); for the adipose tissues the vegetable oil:water partition coefficient was deemed more appropriate (here indicated as p2). to obtain the value of pt:b from the kpu (tissue to plasma unbound partition coefficient) the following equation can be applied: pt:b=kpu fup/bp (s10) p1=10logp (s11) logpveg=1.115 logp-1.35 (s12) p2=10logpveg (s13) acids x=1+10(phiw-pka) y=1+10(php-pka) kput=ewt+x iwt/y+((p nlt+(0.3 p+0.7) npht)/y)+(1/fup-1-(p fnlp+(0.3 p+0.7) fnpp)/y) art (s14) diprotic acids in this equations pka1<pka2 x=1+10(phiw-pka1)+10(-pka2-pka1+2phiw) y=1+10(php-pka1)+10(-pka2-pka1+2php) kput= ewt+x iwt/y +((p nlt+(0.3 p+0.7) npht)/y)+(1/fup-1-(p fnlp+(0.3 p+0.7) fnpp)/y) art (s15) bases x=1+10(pka-phiw) y=1+10(pka-php) x1=1+10(pka-phbc) y1=1+10(pka-php) x2=10(pka-phbc) kpubc=(bp-1+haematocrit)/haematocrit/fup kaap=(kpubc-(x1/y1 iwb)-(p nlb+(0.3 p+0.7) nphb)/y1) (y1/apb/x2) kput=ewt+x iwt/y+(p nlt+(0.3 p+0.7) npht)/y+(kaap apt (x-1))/y (s16) very week bases x=1+10(pka-phiw) y=1+10(pka-php) kput=ewt+x iwt/y+((p nlt+(0.3 p+0.7) npht)/y)+(1/fup-1-(p fnlp+(0.3 p+0.7) fnpp)/y) art (s17) diprotic bases in these equations pka1<pka2 x=1+10(pka2-phiw)+10(pka2+pka1-2 phiw) y=1+10(pka2-php)+10(pka2+pka1-2 php) x1=1+10(pka2-phbc)+10(pka2+pka1-2 phbc) x2=10(pka2-phbc)+10(pka2+pka1-2 phbc) y1=1+10(pka2-php)+10(pka2+pka1-2 php) kpubc=(bp-1+haematocrit)/haematocrit/fup kaap=(kpubc-(x1/y1 iwb)-((p nlb+(0.3 p+0.7) nphb)/y1)) (y1/apb/x2) kput=ewt+x iwt/y+(p nlt+(0.3 p+0.7) npht)/y+(kaap apt (x-1))/y (s18) very week diprotic bases in these equations pka1<pka2 x=1+10(pka2-phiw)+10(pka2+pka1-2 phiw) y=1+10(pka2-php)+10(pka2+pka1-2 php) kput=ewt+x iwt/y+((p nlt+(0.3 p+0.7) npht)/y)+(1/fup-1-(p fnlp+(0.3 p+0.7) fnpp)/y) art (s19) neutrals x=1 y=1 kput=x iwt/y+ewt+((p nlt+(0.3 p+0.7) npht)/y)+(1/fup-1-(p fnlp+(0.3 p+0.7) fnpp)/y) lrt (s20) zwitterions, with at least one basic pka>7 x=1+10(pkab-phiw)+10(phiw-pkaa) y=1+10(pkab-php)+10(php-pkaa) x1=1+10(pkab-phbc)+10(phbc-pkaa) y1=1+10(pkab-php)+10(php-pkaa) x2=10(pkab-phbc)+10(phbc-pkaa) kpubc=(bp-1+haematocrit)/haematocrit/fup; kaap=(kpubc-(x1/y1 iwb)-(p nlb+(0.3 p+0.7) nphb)/y1) (y1/apb/x2) kput=ewt+x iwt/y+(p nlt+(0.3 p+0.7) npht)/y+((kaap apt 10 (pkab-phiw))+10(phiw-pkaa))/y (s21) all other zwitterions x=1+10(pkab-phiw)+10(phiw-pkaa) y=1+10(pkab-php)+10(php-pkaa) kput=ewt+x iwt/y+((p nlt+(0.3 p+0.7) npht)/y)+(1/fup-1-(p fnlp+(0.3 p+0.7) fnpp)/y) art (s22) metabolism and elimination the equations to apply the “qgut” model [33], with the related scaling factors, to obtain fgut in humans from measurement of in vitro intrinsic clearance from hlm, for cyp3a metabolizers are here reported. the fraction of drug escaping the first pass metabolism can be calculated as follows: fgut=qvilli/(qvilli+fugutcluint,gut(1+qvilli/clperm)) (s23) where qvilli is the intestinal villi blood flow that for humans is 300 ml/min; fugut is the unbounded drug fraction in gut, if not available can be supposed equal to 1; cluint,gut is the net metabolic intrinsic clearance based on the unbound drug concentration, this last term can be obtained from the hlm as follows: cluint,gut =(cluint/pemp)newi (s24) where cluint is the unbound hepatic intrinsic clearance obtained from hlm and expressed in microliter/minute/milligram of protein, pemp is the picomol of cyp3a enzymes for milligram of protein that is 155 picomol/milligram of protein, newi is the value of nanomol of enzyme for the whole intestine that is 70.5 nanomol [33]. the value of clperm, can be obtained as: clperm=peffhuman a (s25) where a is the area of the intestine, for humans 6600 cm2 obtained supposing a radius of 1.75 cm and a length of 6 m [33]. additonal references [s1] zoe e. barter, martin k. bayliss, philip h. beaune, alan r. boobis, david j. carlile, robert j. edwards, j. brian houston, brian g. lake, john c. lipscomb, olavi r. pelkonen, geoffrey t. tucker1 and amin rostami-hodjegan. scaling factors for the extrapolation of in vivo metabolic drug clearance from in vitro data: reaching a consensus on values of human microsomal protein and hepatocellularity per gram of liver. current drug metabolism 8 (2007) 33-45. cellular membrane affinity chromatography (cmac) in drug discovery from complex natural matrices doi: 10.5599/admet.535 200 admet & dmpk 6(3) (2018) 200-214; doi: http://dx.doi.org/10.5599/admet.535 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index review cellular membrane affinity chromatography (cmac) in drug discovery from complex natural matrices cayman stephen 1 , abdelfatteh el omri 2 , lukasz ciesla* 1 1 department of biological sciences, the university of alabama, science and engineering complex 2329 tuscaloosa, al 35487, usa 2 genomics and biotechnology section and research group, department of biological sciences, faculty of science, king abdulaziz university (kau), jeddah 21589, box 80141, saudi arabia *corresponding author: e-mail: lmciesla@ua.edu; tel.: +1-205-348-1828; received: april 08, 2018; revised: may 06, 2018; published: may 21, 2018 abstract secondary plant metabolites are evolutionary-designed molecules that interact with multiple biological targets in human organisms. identification of pharmacologically active phytochemicals is usually a time consuming and costly process. cellular membrane affinity chromatography (cmac) allows the detection of secondary metabolites present in complex natural matrices, e.g. plant extracts and their interactions with the immobilized fully-functional transmembrane proteins. after the isolation process of the binding compounds, cmac columns can be used to study the binding process between the potential new ligands and the immobilized transmembrane protein target. the following parameters can be determined using cmac columns: binding affinity (kd), association rate constant (kon), dissociation rate constant (koff) and the equilibrium constant for complex formation (k). this review summarizes the preparation steps and the use of cmac columns in the drug discovery process of new potential drug leads present in complex natural matrices. keywords artificial membrane stationary phase; bioassay; chromatography; frontal affinity chromatography; iam.pc; immobilized protein; phytochemicals; plant extracts introduction for thousands of years, humankind has relied on remedies obtained from nature, mostly from plants. secondary plant metabolites can be considered as evolutionary designed drugs, that have evolved as a means of herbal communication with other organisms. plants use the array of phytochemicals that serve multiple purposes, for example: protection from abiotic stress, like excessive uv radiation (polyphenols, terpenes and terpenoids), attraction of pollinators (volatile terpenes), warding off herbivores (noxious phytochemicals, e.g. alkaloids), protection from bacterial or fungal infections (polyphenols, terpenoids, etc) [1-4]. herbivores evolved different forms of biochemical changes, that protect them from the deleterious effects of some of the phytochemicals, for example: cyp450 superfamily of enzymes, transporter proteins and stress-response cellular signalling pathways [1-4]. many of these mechanisms are highly conserved adaptive stress response systems that humans share with simple herbivores. numerous secondary http://www.pub.iapchem.org/ojs/index.php/admet/index tel:205-348-1828 admet & dmpk 6(3) (2018) 200-214 cell membrane affinity chromatography in drug discovery doi: 10.5599/admet.535 201 metabolites can be considered as already optimized drug leads targeting specific receptors, enzymes or biochemical processes [5]. the huge potential of plant-derived drugs is exemplified by numerous secondary plant metabolites currently used in clinics, for example, artemisinin, ivermectin, morphine, digitoxin, vincristine, podophyllotoxin, and many others [6]. unfortunately, plants do not reveal their secrets easily as the identification of pharmacologically active phytochemicals is usually a difficult and time-consuming process [7]. many pharmaceutical companies have abandoned drug discovery from natural products, considering it as a demanding high resource commitment and often leading to disappointing results (identification of known compounds and poor biological activity). technical problems also prevent now-commonly used high-throughput screening (hts) techniques from testing complex natural matrices for pharmacologically active compounds [8]. hts techniques are designed to test huge libraries of individual compounds. in cases of complex samples, only a simple answer would be delivered if the hts approach was used: the extract is active or not active. identification of pharmacologically active metabolites would have to follow the hts approach. in the traditional approach to drug discovery from natural samples, the first step usually involves the isolation of individual compounds. this step often leads to obtaining compounds which are easily isolated with reasonable quantities needed to run biological activity assays. the most abundant secondary metabolites are usually isolated and many of them have already well-established and known pharmacological activity profiles [7]. the use of bio-guided fractionation has provided some solutions by narrowing down the isolation process to subfractions that have been identified as having the highest activity. however, bio-guided fractionation still requires repetitive multiple steps to identify the subfractions that are rich in biologically active compounds. different attempts have been made to shorten the process of drug lead identification from plant extracts. many of these attempts focused on developing target-oriented assays enabling identification of compounds with desired activity. secondary metabolites act by interacting with numerous targets in the human body, for example, cytosolic proteins or transmembrane receptors [5]. these new assays are designed to screen complex matrices and identify compounds that selectively interact with the biological target (enzyme, receptor) [7]. development of target-oriented assays with immobilized functional enzymes has been more popular than the attempts to design tests aimed at identification of compounds targeting cell membrane proteins. cytosolic proteins are usually covalently bonded using their nor c-terminus groups forming the covalent bond with chemical moieties on the surface of the assay scaffold, e.g. the inside of the column or magnetic beads [9, 10]. many natural compounds target cell membrane proteins but the solutions developed for cytosolic proteins cannot be used to immobilize transmembrane proteins [7]. this review focuses on the development of novel bioassays with the immobilized fully functional cell membrane proteins and their use in drug leads discovery from complex natural matrices. one of the most important issues in developing assays targeting transmembrane receptors is to make sure functional protein is immobilized. introducing chemical changes in the immobilization process of enzymes may lead to irreversible modifications to protein structure and result in failure to develop functional assay [11]. it is important to keep boundary lipids surrounding the transmembrane protein intact as the subtle interactions between transmembrane domains and these lipids influences how the potential ligands bind to the protein. it must be remembered that in the process of transmembrane protein immobilization, the cell membrane fragments and not the pure proteins are immobilized [7]. stripping off the boundary lipids would not only lead to changing the receptor binding activity but also would probably result in the immobilization of protein aggregates, that would form during the cayman, omri & ciesla admet & dmpk 6(3) (2018) 200-214 202 immobilization process [11]. the aggregates would develop because nonpolar amino acid residues, forming the transmembrane domains would bury in the interior of the aggregates not to get exposed to an aqueous environment. the immobilization of cell membrane fragments with functional transmembrane receptors can be achieved in the process of cellular membrane affinity chromatography (cmac) column preparation [11]. successful immobilization of transmembrane fragments will result in obtaining an assay that enables the study of ligand-receptor interactions in close to physiological conditions (figure 1). cellular membrane affinity chromatography columns – a short historical perspective to the best of our knowledge, lundahl’s group was the first to successfully immobilize transmembrane protein, the glucose transporter, glut 1 [12]. this study stimulated wainer’s group to approach the immobilization process of nicotinic acetylcholine receptors (nachrs) on lipid monolayers [13]. interestingly α3β4 nachrs were also successfully immobilized on the immobilized artificial membrane (iam) stationary phase, developed by pidgeon et al. [14] and currently available from regis technologies, figure 2. figure 1. cellular membrane affinity chromatography column contains cell membrane fragments with fully functional transmembrane protein (tpm) targets and therefore allows for studying the interactions between small molecules and tmp in close to physiological conditions studies by the wainer group showed that cmac prepared with the use of iam particles were characterized by better reproducibility compared to liposomal columns [13]. since the successful immobilization of nachrs on the iam stationary phase, multiple transmembrane proteins have been used to prepare cmacs with iam.pc (immobilized artificial membrane phosphatidylcholine). the examples of transmembrane proteins immobilized on iam.pc particles include: ligand-gated ion channels (xy nachr x = 3,4,7; y = 2,3,4; gabaa, nmda, combination of nachr, nmda) [11, 13, 15-30], g-protein coupled receptors (opioid, -adrenergic, p2y1, histamine 1-4, cb1, cb2) [31-35] and drug transporters (abc: pgp, mrp1, mrp2, bcrp; slc: hoct and hoct snps, hoat) [36-39]. iam.pc mimics the cell membrane environment and keeps the experimental conditions close to the physiological state. the surface of the iam.pc stationary phase still contains residual silanol and amine groups on its surface, which may become charged during the analysis, resulting in changed retention parameters of the studied ligands. a new class of stationary phase has been developed by regis, iam.pc.dd2 particles with amine groups endcapped, figure 2. the endcapping leads to an increase of the stationary phase stability and its lipophilicity. to the best of our knowledge, there are no data available on the possible use of iam.pc.dd2 particles in the preparation of cmacs. future experiments should determine whether endcapping the residual amine and silanol groups would result in more reproducible data generated with cmacs. possible problems with the use of more lipophilic artificial membrane particles are discussed further in this review. admet & dmpk 6(3) (2018) 200-214 cell membrane affinity chromatography in drug discovery doi: 10.5599/admet.535 203 figure 2. differences in surface chemistry of iam.pc and iam.pc.dd2 particles. with permission from regis technologies another approach developed to study ligand-protein interactions involves the use of activated silica for the adsorption of cell membrane fragments and is called cell membrane chromatography (cmc) [40]. the use of silica introduces silica groups in the process of studying ligand-protein interactions, which does not closely resemble physiological conditions. however, numerous drug discovery approaches utilizing cmcs have been reported in the literature, for example [40-43]. this review purely focuses on the approach with the artificial membrane as the support for the adsorbing cell membrane fragments and therefore cmc is not further discussed in this paper. preparation steps of cellular membrane affinity chromatography columns the process of preparing cmac columns has been described in detail elsewhere by moaddel and wainer [11]. in this review, critical steps of column preparation are described and particular attention is paid to issues that have the most critical impact on the process of developing a functional cmac. the process of cmac development can be divided into 5 consecutive steps: (1) transmembrane protein source cell/tissue homogenization, (2) solubilization, (3) dialysis and immobilization, (4) packing and (5) column characterization (figure 3). homogenization the first step in preparing a cmac column is the choice of transmembrane protein source (cell line or tissue). if cmac is to be used in the drug discovery screening process of complex natural matrices then the use of transfected cell line overexpressing the transmembrane protein of choice is recommended. a cmac column prepared with a transfected cell line will result in a higher number of active sites increasing the capacity and selectivity of the new column [11]. however, if the experiments using cmac focus on the characterization of transmembrane protein and its interaction with particular ligands, the use of a native cell line is advised. the use of native cell line guarantees that the studied protein is surrounded by the boundary lipids that assure its full natural functionality. in native cell lines, transmembrane proteins also cayman, omri & ciesla admet & dmpk 6(3) (2018) 200-214 204 may undergo some specific post-translational modifications that may not take place in the case of a transfected cell line. figure 3. preparation steps of cell membrane affinity chromatography columns (cmac) the homogenization process is performed in a buffered environment (usually tris -hcl buffer, 50 mm, ph 7.4) with the addition of salts, to keep it close to physiological conditions and protect the transmembrane protein from degradation. the homogenization procedures are performed on ice and all centrifugation steps are executed at 4°c. glass-dounce homogenizers, polytron homogenizers or sonicators are commonly used in the process of cell homogenization [11]. the use of protease and peptidase inhibitors is required to immobilize functional transmembrane protein. the choice of proper inhibitors depends on the type of cell line used in the experiments. the number of cells used in the experiment depends on the level of protein expression and may vary from 1 million to 50 million cells per pellet. after rupturing the cell structure, the resulting homogenate is centrifuged at low speed to remove nuclei and the pellet is discarded. the remaining supernatant is subsequently centrifuged at high speed and the pellet containing cell membranes is collected. in some cases, the use of additional reagents is required for immobilization of functional protein. for example, the addition of glycerol enhances the stability of the protein in aqueous solutions and was found to be useful in some cmac preparation protocols [11]. as previously mentioned the interactions of protein transmembrane domains with boundary lipids influences the activity of the protein. the addition of some phospholipids in the buffer solutions may be required in some instances, as previously reported [11]. for more experimental details please refer to the manuscript by moaddel and wainer [11]. solubilization the protein isolation is completed using the solubilization step. as in the previous homogenization stage, all the experiments are performed at low temperature (4°c) in a buffered environment. compared to the homogenization buffer, used in step 1, solubilization buffer contains detergent, that solubilizes transmembrane proteins by forming micelles around them. several parameters need to be considered when choosing the detergent. again, it is of crucial importance to remember that the goal is to immobilize cell membrane fragments with transmembrane proteins and not the purified proteins. the use of strong detergents, like tween or triton x-100, should be avoided as it may lead to stripping off the boundary lipids resulting in the immobilization of nonactive proteins or protein aggregates. these detergents would also cause problems in the next step, that requires elimination of the detergent, as they are difficult to admet & dmpk 6(3) (2018) 200-214 cell membrane affinity chromatography in drug discovery doi: 10.5599/admet.535 205 remove using dialysis. previous studies have shown the use of the following detergents results in the immobilization of functional transmembrane proteins: sodium cholate (optimal for immobilization of ligand-gated ion channels), chaps (works well with gpcrs), and β-octyl glucopyranoside [11]. the pellet collected after the completion of the homogenization step is mixed with the solubilization buffer and left overnight on the orbital shaker at a temperature of 4°c. dialysis and immobilization after the completion of the overnight solubilization step, the mixture is centrifuged at high speed and the pellet discarded. at this point, the cell membrane fragments are in the supernatant “trapped” in the micelles formed by the detergent. the resulting supernatant is put in contact with iam.pc particles and left on the orbital shaker for 1 hour. to complete the process of immobilization of the cell membrane fragments on the iam particles, the mixture containing iam.pc stationary phase and micelles are placed in dialysis tubing and dialysed for 48 hours, to get rid of the detergent. disruption of micelles results in freeing the cell membrane fragments and the phospholipids forming cellular bilayer adsorb on iam particles to protect lipophilic carbon tails from their exposure to the aqueous environment. packing after completing the dialysis step the content of the dialysis tubing is centrifuged at low speed and the supernatant is discarded. the remaining pellet is washed three times with ammonium acetate buffer (10 mm, ph 7.4). ammonium acetate buffer is used also as running buffer, as it is compatible with mass spectrometers, commonly used as detectors in subsequent experiments involving the use of cmac. the use of tris-hcl buffer would result in an additional tris peak (m/z 122.1, positive ionization mode) in the chromatograms. following the three washes, the pellet is suspended in 1 ml of ammonium acetate buffer and packed in a glass column to produce a transmembrane protein iam.pc cmac column tmp-cmac (figure 4). the packing procedure is a low-pressure packing and therefore does not require any sophisticated equipment. the packing process resembles the slurry pack procedure used in classical column preparation for the isolation of secondary metabolites, but on a smaller scale. after packing the column should be washed with running buffer (ammonium acetate, 10 mm, ph 7.4) at the flow rate of approximately 0.2 ml/min. column characterization and usage after completing all the preparation steps, the moment of truth comes: testing the column to check if functional receptors are immobilized on the surface of the iam particles. this can be achieved through several experimental ways. firstly, the immobilization of transmembrane proteins onto iam particles can be confirmed using confocal microscopy [44]. iam with immobilized cell membrane fragments is incubated with a ligand known to bind to the target protein. the ligand used in the study should contain fluorochrome for the identification with the confocal microscope. iam particles should also be coincubated with a series of concentrations of another ligand, binding to the same site of the investigated receptor, to confirm the competitive displacement of the ligand containing fluorochrome. in another set of experiments confirming the immobilization of functional receptors onto iam particles an artificial mixture of known binders and non-binders is injected onto the column. known binders should be retained on the column with the immobilized receptors significantly longer compared to a negative control column cmac (-). the concept of negative control columns is fully explained in the next section of this article discussing the use of cmac columns in drug discovery from natural samples. cayman, omri & ciesla admet & dmpk 6(3) (2018) 200-214 206 figure 4. cell membrane affinity chromatography column after packing iam particles with the immobilized cell membrane fragments into a glass column finally, a series of concentrations of a known high-affinity ligand should be injected in consecutive runs onto the cmac column to determine the compound’s binding affinity (kd) [11]. the binding affinity value is obtained using frontal affinity chromatography (fac) and a detector sensitive enough to detect the marker ligand at the used concentration. mass spectrometers are most commonly used in this approach, but any other detector suitable for the detection of the marker ligand may be applied. the concept of frontal affinity chromatography was thoroughly reviewed in the paper by ng et al. [45] and will only be shortly discussed in this review. frontal affinity chromatography is a biophysical method to study the interactions between a ligand present in the mobile phase and the immobilized target molecule, e.g. a receptor [45]. a known marker ligand (for example epibatidine for nachrs) at constant concentration is run through the column with the immobilized transmembrane protein, producing a characteristic sigmoidal breakthrough curve (figure 5). at the beginning of the analysis, the number of available binding sites (immobilized transmembrane protein) is higher than the number of the titrating molecules of the marker ligand (the flat bottom portion of the curve). after all the binding sites are saturated with the ligand, the concentration of the ligand leaving the column gradually increases producing the breakthrough. at the end of the analysis, the concentration of the ligand entering and eluting from the column is identical as illustrated by the plateau portion of the curve. the inflexion point depends on the ligand concentration, the number of available binding sites and the compound’s binding affinity [45]. typical chromatograms using fac obtained for orthosteric (a) and allosteric (b) nachr binding sites are presented in figure 5 [30]. using frontal affinity chromatography with cmac columns was proved not only to be useful in characterization and studies involving the orthosteric but also with the allosteric binding sites of the immobilized transmembrane protein [30]. frontal displacement chromatography can be applied to determine the binding affinity of unknown ligand binding to the same site as the marker ligand. it is done by injecting constant concentration of the marker ligand with varying concentrations of the studied displacer and using the following equation [11]: -1 where: [dis] – concentration of the displacer; v – retention volume of the marker ligand, when its specific concentration is completely suppressed, e.g. by a high concentration of the displacer; kddis – dissociation constant of the displacer; p – product of the bmax (the number of the available binding sites) and (kddis/kdmarker), kdmarker – dissociation constant of the marker ligand [11]. admet & dmpk 6(3) (2018) 200-214 cell membrane affinity chromatography in drug discovery doi: 10.5599/admet.535 207 figure 5. frontal chromatograms of (a) increasing concentrations of [ 3 h]-epibatidine (a =240 pm, b=150 pm, c=100 pm, d=80 pm and e=60 pm) on the α3β4α5 nicotinic receptor column and (b) increasing concentrations of mecamylamine on α3β4 nicotinic receptor column. for details please refer to the paper by ciesla et al. [30]. with permission from [30] this approach was for example used to determine kd values of several alkaloids binding to 34 and 345 nicotinic receptors [30]. another chromatographic mode that can be used to study interactions between the adsorbed transmembrane receptors and ligands is nonlinear chromatography [11]. in this approach, the investigated small molecular compounds are injected onto a cmac column close to, or at target saturation. the observed peak asymmetry allows the study of the kinetics and thermodynamics of interaction between the ligands and immobilized protein target. cmac columns can be used as a first step to identify cayman, omri & ciesla admet & dmpk 6(3) (2018) 200-214 208 pharmacologically active secondary metabolites from natural mixtures and after isolation of these compounds, they can be applied to obtain pharmacokinetic and pharmacodynamic data. in zonal chromatography, a small amount of the ligand is loaded onto the column and this allows for ranking the affinity of the investigated compounds [11]. the use of cmacs in identification of pharmacologically active metabolites from natural mixtures in the early reports of cmac column usage, the columns were applied to the study of the binding characteristics of the immobilized transmembrane proteins [7]. however, cmac columns can be successfully applied in the process of identification of pharmacologically active compounds present in complex natural matrices targeting transmembrane receptors [7, 46]. there are several issues that need to be taken into consideration when applying cmacs in drug discovery from natural samples. the first decision that has to be made is in choosing extracts that will be screened using the newly developed bioassay. the samples to be run on the cmac columns should not be random, but carefully selected. plant extracts that previously were reported to exert desired pharmacological activity may be considered for the screening experiments. however, the observed biological effect may not be from the result of the interaction of secondary metabolites present in the extract and the targeted protein, immobilized on the artificial membrane support. another idea of narrowing down the possible screening extract candidates is to use a simple in vivo model (c. elegans or d. melanogaster) and test the extracts for biological activity. the extracts that exert the biological effect are further screened to identify phytochemicals responsible for this effect. again, there has to be a known, or at least a rightfully hypothesized connection between activation/inhibition of the targeted transmembrane protein and pharmacological activity tested in the in vivo model. the easiest way to check for the possible presence of phytochemicals interacting with the transmembrane receptors is to use frontal affinity chromatography and a marker ligand known to bind to the transmembrane receptor [7, 30]. the displacement observed after co-injecting the analysed extract and the marker ligand would indicate the presence of phytochemicals competing with the marker for the binding site. this approach can be applied to screen extracts for compounds interacting with both orthosteric and allosteric binding sites. figure 6 presents the displacement of mecamylamine, a known nachr allosteric inhibitor displaced by lycopodium clavatum l. extract on cmac columns with immobilized two types of nicotinic receptors α3β4 and α3β4α5 [30]. the identification process of individual secondary metabolites responsible for the observed displacement of the marker ligand is performed using the “missing peak chromatography” approach [46]. one of the greatest challenges in drug discovery from plant extracts, using cmac columns, is discerning specific interactions (phytochemicals interacting with the immobilized transmembrane target) from nonspecific interactions (compounds interacting with iam particles or other transmembrane proteins coimmobilized with the studied targeted receptor). lipophilic compounds have been found to be strongly retained on iam.pc particles and an alternative approach has been proposed to study such ligands: immobilization of cell membrane fragments onto the surface of an open tubular capillary (ot columns) [7]. for example, ot columns were successfully used to immobilize cell membrane fragments with cb1/cb2 receptors and further for the identification of endocannabinoid receptor ligands from the extract of zanthoxylum clava-herculis l. [34, 47]. the development and the use of ot columns remain beyond the scope of this review and interested readers are encouraged to refer to other publications [7]. coimmobilization of other proteins does not need to be considered as a cmac disadvantage, as the same column can be utilized to study compounds binding to multiple transmembrane proteins [28]. to admet & dmpk 6(3) (2018) 200-214 cell membrane affinity chromatography in drug discovery doi: 10.5599/admet.535 209 characterize multiple proteins co-immobilized on the iam particles, different marker ligands have to be used, that specifically bind to the receptors of interest. every receptor has to be characterized in separate runs. figure 6. frontal elution profile of 1 µm mecamylamine (i) and 1 µm mecamylamine + 5% lycopodium clavatum (ii) on (a) the α3β4 nicotinic receptor column and (b) α3β4α5 nicotinic receptor column. for details please refer to the paper by ciesla et al. [30]. with permission from [30]. the missing peak chromatography approach requires the use of a negative control cmac column, cmac(-). as previously stated, transfected cell lines stably overexpressing studied transmembrane proteins are recommended for the preparation of cmac columns. a negative control column (cmac(-)) is prepared using the same cell line, however not expressing the receptors overexpressed by the transfected line. for example, in the case of nachr cmac columns, hek cells overexpressing desired nachr subtypes were used cayman, omri & ciesla admet & dmpk 6(3) (2018) 200-214 210 to prepare cmac(+), while hek 293 cells were utilized to develop cmac(-) [30]. in the missing peak chromatography technique, plant extracts are chromatographed on cmac(+) and cmac(-) columns using ammonium acetate buffer (10 mm, ph 7.4) delivered at 0.2 ml/min as mobile phase [46]. fractions are collected based on the observed chromatograms. the obtained fractions are subsequently fingerprinted on c18 columns and the fingerprints obtained using cmac(+) and cmac(-) are compared. the difference between the cmac(+) and cmac(−) columns is in the availability of the immobilized protein target. the differences in the retention of a compound on the cmac(+) column relative to the cmac(−) column is due to specific interactions with the target [7, 46]. a peak that is missing in the fingerprint of a cmac(+) early fraction, but present in the control and corresponding cmac(−) early fraction, is considered to bind to the immobilized transmembrane protein target on the cmac(+) column [46]. elution of this particular compound in a later fraction of cmac(+), along with its absence in the corresponding fraction in cmac(−), is a further confirmation of the affinity of this compound for the targeted transmembrane receptor in cmac(+) [46]. compounds identified as specifically binding to the studied receptors can be subsequently isolated using analytical or semipreparative hplc, their structures elucidated and further tested in functional assays or in vivo models. cmac chromatography columns with immobilized cell membrane fragments on iam particles do not solve all of the problems in the drug discovery process from natural samples. this approach requires the use of negative control columns and the solvent choice is limited to buffered aqueous solvents (with only up to approx. 10% of organic modifier). cmac columns, however, do speed up the process of target identification, without the need to isolate individual compounds and test them separately using different assays. cmac is an assay, in which transmembrane proteins remain submerged in a natural lipid environment, in close to physiological conditions. cmac columns allow for the identification of compounds binding to receptor subtypes. after identification and isolation of pharmacologically active phytochemicals, cmac columns may be used to study their interaction with the immobilized protein target (obtaining pharmacodynamics and pharmacokinetic data). the drug discovery process with cmacs is a targetoriented technique, ideally suited for creating libraries of natural compounds interacting with a specific cell membrane target. future directions the potential of cell membrane affinity chromatography columns in the drug discovery process from natural samples has been underestimated. pharmaceutical companies have focused their discovery programs on automated high-throughput screening techniques (hts), which are not suitable for complex matrices. however, the introduction of hts techniques did not result in the predicted substantial increase of new classes of drugs entering the market [8]. secondary metabolites are evolutionary-designed compounds targeting specific, often highly-conserved signalling pathways and are already optimized to be pharmacologically active. one of the ways of unravelling the mysteries of natural compounds is to develop approaches enabling identification of these phytochemicals. cellular membrane affinity chromatography columns allow phytochemists to selectively identify compounds interacting with transmembrane proteins. we predict cmac columns have a potential to become the next generation of drug discovery tools in phytochemistry and pharmacology labs. these columns may be easily applied to build libraries of natural compounds that can move drug discovery projects forward to the next step. some advice is given: carefully choose your transmembrane target by closely studying the possible pathways involved in the disease. an example is given from our lab: we are interested in cellular signalling pathways that have evolved to protect neurons from different forms of biological stress (oxidative, metabolic, proteotoxic or admet & dmpk 6(3) (2018) 200-214 cell membrane affinity chromatography in drug discovery doi: 10.5599/admet.535 211 inflammatory stress). published data indicated that the activation of the bdnf (brain-derived neurotrophic factor) pathway is neuroprotective [48]. bdnf works by binding with transmembrane tropomyosin receptor kinase (trkb). a cmac column with the immobilized trkb receptors is to be used to screen selected plant extracts for natural compounds binding to this receptor and mimicking the neuroprotective effects of bdnf. preparation of cmac columns appears to be considered as a complicated approach and this may prevent it from common use in drug discovery or phytochemistry laboratories. the truth is, the preparation of cmac columns is possible in almost every lab focusing on natural product discovery. future direction may include the development of a kit that would allow interested labs to easily develop a column with the immobilized desired transmembrane protein. cmac has a potential to work well with other drug discovery approaches, for example, ligand fishing techniques. pharmacologically active compounds may be first identified using ligand fishing techniques followed by ligand-protein interaction studies with a cmac column. the advent of novel drug discovery approaches is needed in this new era when drug discovery from natural samples is again becoming popular. the use of old techniques, that were one of the reasons for abandoning natural product screening, need to be replaced by innovative assays. cellular membrane affinity chromatography together with ligand fishing techniques have everything that is required to become drug discovery techniques of the 21 st century. references [1] m.p. mattson. what doesn't kill you. sci am 313 (2015) 40-45. [2] m.p. mattson, a. cheng. neurohormetic phytochemicals: low-dose toxins that induce adaptive neuronal stress responses. trends neurosci 29 (2006) 632-639. [3] m.p. mattson, t.g. son, s. camandola. viewpoint: mechanisms of action and therapeutic potential of neurohormetic phytochemicals. dose response 5 (2007) 174-186. [4] v. murugaiyah, m.p. mattson. neurohormetic phytochemicals: an evolutionary-bioenergetic perspective. neurochem int 89 (2015) 271-280. [5] j. lee, d.g. jo, d. park, h.y. chung, m.p. mattson. adaptive cellular stress pathways as therapeutic targets of dietary phytochemicals: focus on the nervous system. pharmacol rev 66 (2014) 815-868. [6] b. shen. a new golden age of natural products drug discovery. cell 163 (2015) 1297-1300. [7] l. ciesla, r. moaddel. comparison of analytical techniques for the identification of bioactive compounds from natural products. nat prod rep 33 (2016) 1131-1145. [8] a.l. harvey, r. edrada-ebel, r.j. quinn. the re-emergence of natural products for drug discovery in the genomics era. nat rev drug discov 14 (2015) 111-129. [9] m. yasuda, d.r. wilson, s.d. fugmann, r. moaddel. synthesis and characterization of sirt6 protein coated magnetic beads: identification of a novel inhibitor of sirt6 deacetylase from medicinal plant extracts. anal chem 83 (2011) 7400-7407. [10] n. singh, s. ravichandran, d.d. norton, s.d. fugmann, r. moaddel. synthesis and characterization of a sirt6 open tubular column: predicting deacetylation activity using frontal chromatography. anal biochem 436 (2013) 78-83. [11] r. moaddel, i.w. wainer. the preparation and development of cellular membrane affinity chromatography columns. nat protoc 4 (2009) 197-205. [12] e. brekkan, a. lundqvist, p. lundahl. immobilized membrane vesicle or proteoliposome affinity chromatography. frontal analysis of interactions of cytochalasin b and d-glucose with the human red cell glucose transporter. biochemistry 35 (1996) 12141-12145. cayman, omri & ciesla admet & dmpk 6(3) (2018) 200-214 212 [13] y. zhang, y. xiao, k.j. kellar, i.w. wainer. immobilized nicotinic receptor stationary phase for on-line liquid chromatographic determination of drug-receptor affinities. anal biochem 264 (1998) 22-25. [14] s. ong, h. liu, x. qiu, g. bhat, c. pidgeon. membrane partition coefficients chromatographically measured using immobilized artificial membrane surfaces. anal chem 67 (1995) 755-762. [15] i.w. wainer, y. zhang, y. xiao, k.j. kellar. liquid chromatographic studies with immobilized neuronal nicotinic acetylcholine receptor stationary phases: effects of receptor subtypes, ph and ionic strength on drug-receptor interactions. j chromatogr b biomed sci appl 724 (1999) 65-72. [16] m.t. baynham, s. patel, r. moaddel, i.w. wainer. multidimensional on-line screening for ligands to the alpha3beta4 neuronal nicotinic acetylcholine receptor using an immobilized nicotinic receptor liquid chromatographic stationary phase. j chromatogr b analyt technol biomed life sci 772 (2002) 155-161. [17] k. jozwiak, j. haginaka, r. moaddel, i.w. wainer. displacement and nonlinear chromatographic techniques in the investigation of interaction of noncompetitive inhibitors with an immobilized alpha3beta4 nicotinic acetylcholine receptor liquid chromatographic stationary phase. anal chem 74 (2002) 4618-4624. [18] r. moaddel, i.w. wainer. immobilized nicotinic receptor stationary phases: going with the flow in high-throughput screening and pharmacological studies. j pharm biomed anal 30 (2003) 1715-1724. [19] k. jozwiak, s.c. hernandez, k.j. kellar, i.w. wainer. enantioselective interactions of dextromethorphan and levomethorphan with the alpha 3 beta 4-nicotinic acetylcholine receptor: comparison of chromatographic and functional data. j chromatogr b analyt technol biomed life sci 797 (2003) 373-379. [20] k. jozwiak, s. ravichandran, j.r. collins, i.w. wainer. interaction of noncompetitive inhibitors with an immobilized alpha3beta4 nicotinic acetylcholine receptor investigated by affinity chromatography, quantitative-structure activity relationship analysis, and molecular docking. j med chem 47 (2004) 4008-4021. [21] r. moaddel, k. jozwiak, r. yamaguchi, c. cobello, k. whittington, t.k. sarkar, s. basak, i.w. wainer. on-line screening of conformationally constrained nicotines and anabasines for agonist activity at the alpha3beta4and alpha4beta2-nicotinic acetylcholine receptors using immobilized receptorbased liquid chromatographic stationary phases. j chromatogr b analyt technol biomed life sci 813 (2004) 235-240. [22] r. moaddel, k. jozwiak, k. whittington, i.w. wainer. conformational mobility of immobilized alpha3beta2, alpha3beta4, alpha4beta2, and alpha4beta4 nicotinic acetylcholine receptors. anal chem 77 (2005) 895-901. [23] r. moaddel, k. jozwiak, r. yamaguchi, i.w. wainer. direct chromatographic determination of dissociation rate constants of ligand-receptor complexes: assessment of the interaction of noncompetitive inhibitors with an immobilized nicotinic acetylcholine receptor-based liquid chromatography stationary phase. anal chem 77 (2005) 5421-5426. [24] k. jozwiak, r. moaddel, r. yamaguchi, a. maciuk, i.w. wainer. non-competitive inhibitory activities of morphinan and morphine derivatives at the alpha 3 beta 4 neuronal nicotinic acetylcholine receptor determined using nonlinear chromatography and chemometric techniques. pharm res 23 (2006) 2175-2182. [25] r. moaddel, k. jozwiak, i.w. wainer. allosteric modifiers of neuronal nicotinic acetylcholine receptors: new methods, new opportunities. med res rev 27 (2007) 723-753. [26] k. jozwiak, s. ravichandran, j.r. collins, r. moaddel, i.w. wainer. interaction of noncompetitive inhibitors with the alpha3beta2 nicotinic acetylcholine receptor investigated by affinity chromatography and molecular docking. j med chem 50 (2007) 6279-6283. [27] r. moaddel, r.v. oliveira, t. kimura, p. hyppolite, m. juhaszova, y. xiao, k.j. kellar, m. bernier, i.w. wainer. initial synthesis and characterization of an alpha7 nicotinic receptor cellular membrane affinity chromatography column: effect of receptor subtype and cell type. anal chem 80 (2008) 4854. admet & dmpk 6(3) (2018) 200-214 cell membrane affinity chromatography in drug discovery doi: 10.5599/admet.535 213 [28] t. kitabatake, r. moaddel, r. cole, m. gandhari, c. frazier, j. hartenstein, a. rosenberg, m. bernier, i.w. wainer. characterization of a multiple ligand-gated ion channel cellular membrane affinity chromatography column and identification of endogenously expressed receptors in astrocytoma cell lines. anal chem 80 (2008) 8673-8680. [29] r. moaddel, j.f. cloix, g. ertem, i.w. wainer. multiple receptor liquid chromatographic stationary phases: the co-immobilization of nicotinic receptors, gamma-amino-butyric acid receptors, and nmethyl d-aspartate receptors. pharm res 19 (2002) 104-107. [30] l. ciesla, m. okine, a. rosenberg, k.s.s. dossou, l. toll, i.w. wainer, r. moaddel. development and characterization of the alpha3beta4alpha5 nicotinic receptor cellular membrane affinity chromatography column and its application for on line screening of plant extracts. j chromatogr a 1431 (2016) 138-144. [31] f. beigi, i.w. wainer. syntheses of immobilized g protein-coupled receptor chromatographic stationary phases: characterization of immobilized mu and kappa opioid receptors. anal chem 75 (2003) 4480-4485. [32] f. beigi, k. chakir, r.p. xiao, i.w. wainer. g-protein-coupled receptor chromatographic stationary phases. 2. ligand-induced conformational mobility in an immobilized beta2-adrenergic receptor. anal chem 76 (2004) 7187-7193. [33] r. moaddel, h.k. musyimi, m. sanghvi, c. bashore, c.r. frazier, m. khadeer, p. bhatia, i.w. wainer. synthesis and characterization of a cellular membrane affinity chromatography column containing histamine 1 and p2y(1) receptors: a multiple g-protein coupled receptor column. j pharm biomed anal 52 (2010) 416-419. [34] r. moaddel, a. rosenberg, k. spelman, j. frazier, c. frazier, s. nocerino, a. brizzi, c. mugnaini, i.w. wainer. development and characterization of immobilized cannabinoid receptor (cb1/cb2) open tubular column for on-line screening. anal biochem 412 (2011) 85-91. [35] i. gonzalez-mariscal, s.m. krzysik-walker, m.e. doyle, q.r. liu, r. cimbro, s. santa-cruz calvo, s. ghosh, l. ciesla, r. moaddel, o.d. carlson, r.p. witek, j.f. o'connell, j.m. egan. human cb1 receptor isoforms, present in hepatocytes and beta-cells, are involved in regulating metabolism. sci rep 6 (2016) 33302. [36] r. moaddel, l. lu, m. baynham, i.w. wainer. immobilized receptorand transporter-based liquid chromatographic phases for on-line pharmacological and biochemical studies: a mini-review. j chromatogr b analyt technol biomed life sci 768 (2002) 41-53. [37] r. moaddel, p.l. bullock, i.w. wainer. development and characterization of an open tubular column containing immobilized p-glycoprotein for rapid on-line screening for p-glycoprotein substrates. j chromatogr b analyt technol biomed life sci 799 (2004) 255-263. [38] r. moaddel, r. hamid, s. patel, p.l. bullock, i.w. wainer. identification of p-glycoprotein substrates using open tubular chromatography on an immobilized p-glycoprotein column: comparison of chromatographic results with caco-2 permeability. anal chim acta 578 (2006) 25-30. [39] k.l. habicht, n.s. singh, m.a. khadeer, r. shimmo, i.w. wainer, r. moaddel. characterization of a multiple endogenously expressed adenosine triphosphate-binding cassette transporters using nuclear and cellular membrane affinity chromatography columns. j chromatogr a 1339 (2014) 80-85. [40] s. wang, m. sun, y. zhang, h. du, l. he. a new a431/cell membrane chromatography and online high performance liquid chromatography/mass spectrometry method for screening epidermal growth factor receptor antagonists from radix sophorae flavescentis. j chromatogr a 1217 (2010) 52465252. [41] l.c. he, s.c. wang, g.d. yang, y.m. zhang, c.h. wang, b.x. yuan, x.f. hou. progress in cell membrane chromatography. drug discov ther 1 (2007) 104-107. [42] s. han, y. lv, f. wei, j. fu, q. hu, s. wang. screening of bioactive components from traditional chinese medicines using cell membrane chromatography coupled with mass spectrometry. phytochem anal (2018). cayman, omri & ciesla admet & dmpk 6(3) (2018) 200-214 214 [43] y. lin, c. wang, y. hou, w. sun, d. che, l. yang, t. zhang, m. sun, h. hehe, l. he. simultaneous identification of three pseudo-allergic components in danshen injection by using high-expression mas-related g protein coupled receptor x2 cell membrane chromatography coupled online to hplcesi-ms/ms. j sep sci (2018). [44] k.l. habicht, n.s. singh, f.e. indig, i.w. wainer, r. moaddel, r. shimmo. the development of mitochondrial membrane affinity chromatography columns for the study of mitochondrial transmembrane proteins. anal biochem 484 (2015) 154-161. [45] e.s. ng, n.w. chan, d.f. lewis, o. hindsgaul, d.c. schriemer. frontal affinity chromatography-mass spectrometry. nat protoc 2 (2007) 1907-1917. [46] a. maciuk, r. moaddel, j. haginaka, i.w. wainer. screening of tobacco smoke condensate for nicotinic acetylcholine receptor ligands using cellular membrane affinity chromatography columns and missing peak chromatography. j pharm biomed anal 48 (2008) 238-246. [47] k.s. dossou, k.p. devkota, c. morton, j.m. egan, g. lu, j.a. beutler, r. moaddel. identification of cb1/cb2 ligands from zanthoxylum bungeanum. j nat prod 76 (2013) 2060-2064. [48] m.p. mattson, k. moehl, n. ghena, m. schmaedick, a. cheng. intermittent metabolic switching, neuroplasticity and brain health. nat rev neurosci 19 (2018) 63-80. ©2018 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ estimation of skin permeation by liquid chromatography doi: 10.5599/admet.512 140 admet & dmpk 6(2) (2018) 140-152; doi: http://dx.doi.org/10.5599/admet.512 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper estimation of skin permeation by liquid chromatography sara soriano-meseguer 1 , elisabet fuguet 1,2 , adriana port 3 , martí rosés 1, * 1 departament de química analítica i institut de biomedicina, universitat de barcelona, martí i franquès 1-11, 08028 barcelona, spain 2 serra húnter programme, generalitat de catalunya, 08002 barcelona, spain 3 esteve, parc científic de barcelona, baldiri reixac 4-8, 08028 barcelona, spain *corresponding author: e-mail: marti.roses@ub.edu; tel.: +34-934039275; fax: +34-934021233 received: february 26, 2018; revised: april 24, 2018; available online: may 11, 2018 abstract dermal absorption is a key process in the drug delivery studies of the pharmaceutical and cosmetic industries, as well as in the fields of dermal toxicology, risk assessment, and the exposure of environmental pollutants. this process is typically described by the skin-water permeability coefficient. however, in vivo determination is laborious and expensive. thus, in the last few years, the development of prediction models from structure descriptors or subrogation through physico-chemical measurements has gained interest. in the present work, a previous subrogation model based on the chromatographic retention on a common c18 column has been tested for a wide set of drugs with very different chemical nature and having a wide range of permeability values. a total of 65 compounds have been used to establish the correlation between skin permeation and the hplc retention, corrected by the mcgowan volume of the drug. afterwards it was successfully validated in terms of robustness and prediction ability. finally, the permeability coefficient was estimated for a set of 29 new drugs, and results compared to the ones obtained by other estimation methods, as well as the available in vitro measured values, with very good agreement. keywords dermal absorption; high-performance liquid chromatography; estimation; skin permeation introduction the skin is the largest organ of the body, covering about 1.7 m 2 and comprising approximately 10% of the total body mass of an average person. the primary function of the skin is to provide a barrier between the body and the external environment (ultraviolet radiation, chemicals, allergens and microorganisms, and the loss of moisture and body nutrients). although the skin presents a barrier to the permeation of most compounds, it provides an ideal site for the administration of therapeutic compounds for local and systemic effects [1]. dermal absorption is a global term that refers to the transport of a chemical from the outer surface of the skin into the skin and the systemic circulation. the outermost layer of the skin, the stratum corneum, controls the dermal absorption process, since substances must initially permeate through it to be absorbed into the body. hence stratum corneum is considered the rate-determining barrier and the absorption process can only occur by passive diffusion since skin does not have any active transport mechanism [2,3]. http://www.pub.iapchem.org/ojs/index.php/admet/index admet & dmpk 6(2) (2018) 140-152 estimation of skin permeation by liquid chromatography doi: 10.5599/admet.512 141 dermal absorption is a key process in the drug delivery studies of the pharmaceutical and the cosmetic industries as well as in the fields of dermal toxicology, risk assessment and the exposure to environmental pollutants. it is typically described by the skin-water permeability coefficient (kp, units in cm s -1 ) and can be determined using experimental techniques, both in vivo and in vitro [3]. however, these methods are usually laborious, costly and ethically questionable. for this reason, there is a need for finding methods capable of predicting kp values in a simple, economic, fast and ethical way. to predict skin-water permeability coefficients and thus reduce the number of in vivo and in vitro assays, two types of models, empirical and theoretical, can be developed. empirical models relate the kp values of compounds to physicochemical descriptors, as for example descriptors of hydrophobicity, polarity, solubility, hydrogen bonding, or size [4]. in general, empirical models for skin permeation, as proposed by patel et al. [5], usually reveal the important roles of lipophilicity and molecular size. theoretical models are mechanistic approaches that attempt to simulate the dynamic processes that are involved in skin permeation. they consist of mathematical transport models that relate dermal absorption to parameters like diffusion coefficients and partition coefficients, taking into account factors such as possible routes of penetration and interactions between the permeating chemicals and the skin constituents [6-8]. linear free-energy relationships (lfers), a particular type of quantitative structure-activity relationship (qsar), are widely used to characterize chemical and biochemical processes. lfers assume that the free energy change associated with the property of interest is linearly related to solute molecular descriptors. the linear solvation energy relationships (lsers), proposed by kamlet et al. [9], are based on the change of free energy due to the solvation process of a solute into a solvent. among the variety of models based on these principles, the solvation parameter model developed by abraham [10] is one of the most widely used in order to achieve a better understanding of the type and the relative strength of the chemical interactions that control any solvation process of neutral compounds [11,12]. this model has been successfully applied to characterize many biological processes, including some toxicological and environmental ones, as well as a wide range of physicochemical processes ruled by the passive transport of solutes between two phases. the solvation parameter model [10,12] has been also applied to model the human skin permeation process [12-16]. abraham and martins [15] proposed the following equation for characterizing the permeation of neutral solutes from aqueous solution through the human skin, log kp = – 5.246 – 0.106e – 0.473s – 0.473a – 3.000b + 2.296v (1) n = 119 r 2 = 0.832 sd = 0.461 f= 112 this equation was obtained through the analysis of 119 compounds of different chemical type, which covered a wide range of log kp values. e, s, a, b, and v are independent variables and are the solute descriptors proposed by abraham: e represents the excess molar refraction, s is the solute dipolarity/polarizability, a and b are the solute’s effective hydrogen-bond acidity and hydrogen-bond basicity, respectively, and v is the mcgowan’s solute volume. eq. 1 shows that size is the only solute property that contributes to an increase in skin permeation (the only positive coefficient). over the last few years, the computational models based on calculated descriptors [17-19] have also become very useful and powerful approaches to provide estimations of kp values. however, skin permeation models based on experimental descriptors are generally preferred when the prediction of skinmarti rosés et al admet & dmpk 6(2) (2018) 140-152 142 water permeability of chemicals requires more accuracy and reliability. in a previous study [20] the ability of different chromatographic systems, basically liquid chromatography (hplc) and micellar electrokinetic chromatography (mekc), were assessed to subrogate the permeation of aqueous solutions of neutral solutes through human skin. it was observed that the correlations were improved after introducing the calculated mcgowan solute’s volume as an additional variable. with the volume correction, the best subrogation was obtained with an hplc system fitted with a c18 column according to the following equation, log kp= 4.76(±0.18) + 1.44(±0.12) log k 1.16(±0.10) v (2) n = 27 r 2 = 0.834 sd = 0.390 f = 126 this system was chosen to establish a methodology for the prediction of human skin permeability of untested compounds by chromatographic measurements. this methodology was developed only for neutral compounds at ph 7.0. this study has been expanded in the present work to compounds of pharmaceutical interest with a wide range of pka values, containing basic and acidic functions. additionally, a new methodology based on fast measurements by ms (mass spectrometry) has been developed and established to predict kp values of drugs of pharmaceutical interest. materials and methods instruments chromatographic measurements were performed with an agilent technologies 1200 series instrument equipped with a g1312b binary pump and a g1367d autoinjector. a g1315c dad was used at 254 nm for non-ionizable compounds and a uhd 6540 accurate-mass q-tof detector with electrospray ionization (esi) source was used for ionizable compounds. a 100 mm, 4.6 mm i.d, 2.6 µm octadecylsilica kinetex evo c18 analytical column provided by phenomenex with a core-shell technology was used for all determinations. this material is stable within the ph range 1-12. ph measurements were done with a microph 2001 ph-meter (crison) with a precision of ± 0.01 ph units. reagents acetonitrile lcms grade was purchased from fluka analytical vwr (west chester, pa, usa) and water was purified by a milli-q deionizing system from millipore (billerica, ma, usa) to a resistivity of 18.2 m. reagents used to prepare the buffer solutions were sodium phosphate monobasic monohydrate (sigmaaldrich, puriss pa ≥99.0%), formic acid (scharlau, eluent additive for lc-ms), acetic acid (fluka analytical, eluent additive for lc-ms), ethylendiamine (fluka analytical, puriss pa ≥ 99.5%) and ammonia solution 25 % w/w (sharlau, extrapur, pharmpur, ph eur). most drugs were purchased from sigma-aldrich (steinheim, germany), fluka analytical vwr (west chester, pa, usa), riedel-de haën (seelze, germany), merck (darmstadt, germany), carlo erba (milano, italy) and baker (center valley, pa, usa). other drugs were synthesized in esteve (barcelona, spain) only for the purpose of this study. admet & dmpk 6(2) (2018) 140-152 estimation of skin permeation by liquid chromatography doi: 10.5599/admet.512 143 procedure 94 solutes of different chemical nature were injected in the hplc system at 6 different ph values, between 2 and 11, to obtain the retention factors of the neutral species. the mobile phase composition was 40 % acetonitrile and 60 % aqueous buffer. the buffer solutions were selected according to their compatibility with the detection mode. volatile buffers were used for mass spectrometry detection. formic acid, acetic acid and ammonia solution were used at ph 3.0, 5.0 and 9.0, respectively. ethylendiamine was used at ph 7.0 and 11.0. the buffer concentrations at ph 3.0, 5.0, 7.0, 9.0 and 11.0 were 10 mm and were adjusted by addition of diluted acetic acid or diluted ammonia. the detection mode was esi+ or esifor ionizable compounds and uv for non-ionizable compounds. for mass spectrometry detection, the compounds were grouped for faster analysis. at ph 2.0, the detection was only performed by uv and a mixture of phosphoric acid and sodium dihydrogenphosphate was used as buffer, at a concentration of 50 mm. ph was adjusted with diluted hydrochloric acid. all experiments were done at 25 °c. stock solutions of the compounds at 5000 mg·l -1 were prepared by dissolving the appropriate weight or volume in methanol. more diluted solutions were prepared at 100 mg·l -1 by dissolving an aliquot of the previous stock solution in an acn-h2o mixture (40:60). isocratic conditions were used at a flow rate of 1 ml min -1 and the injection volume was 10 µl. the column hold-up time was measured by injections of an aqueous solution of potassium bromide (0.1 mg ml -1 ), whose detection was performed by uv at 200 nm. all results were the average of triplicate injections. the hplc retention factor (k) was calculated according to eq. 3, r 0 0 e t t k t t    (3) where tr corresponds to the solute retention time, t0 is the column hold-up time determined by an aqueous potassium bromide solution, and te is the extra column time determined by an analysis that excludes the chromatographic column, for each different detection system. calculation microsoft excel 2010 was used for all calculations and for multiple linear regression analysis. solvation parameter model descriptors of the compounds (e, s, a, b, v) were obtained from percepta [21]. matlab® (mathworks) was used for principal components analysis (pca). model establishment and validation data collection the evaluation of the consistency and quality of the experimental skin permeation data used is fundamental to the establishment of a good prediction model. for any qsar model, experimental data should be produced from standardized experimental procedures, and obtained for a set of chemicals that cover the domain of relevant chemical properties [22]. several studies that have focused on the determination of skin permeability coefficients [23-25] have pointed out that interand intra-variability within experimental kp values is mainly due to the thickness of the skin and also to the variability inherent to different human donors. in this work, the permeation data employed was selected from the abraham database [15], an extensive and carefully examined data set. 65 solutes that belong to different families, with different structural complexity and with varied experimental marti rosés et al admet & dmpk 6(2) (2018) 140-152 144 log kp values were selected to develop this model. because the permeation coefficients are from water through human skin, it might be expected that acids and bases would be partially ionized in solution depending on the ph of the solution. values of kp for ionized species are smaller than kp values for the neutral species. in these cases, abraham adjusted the literature values of the skin-water permeability coefficients for ionization in water and for a temperature of 37 °c [15]. a different set of 29 solutes was selected from other sources [5,26,27] to predict the kp values from our model’s equation and then compare them to the available literature experimental and/or estimated data. model establishment a correlation between the experimental log kp values and the parameters log k and the mcgowan solute’s volume (v) of the selected compounds was performed through a multiple linear regression analysis. log k values were the ones measured at a ph in which the solutes were in the neutral form. the mcgowan volume of the compounds can be easily calculated from the empirical formula and the number of single, double, and triple bounds in the molecule. compounds with a standard residual higher than 2.5 in absolute value were considered outliers. model validation the model required an internal validation to evaluate its robustness, and an external validation to evaluate its predictive ability. to this aim, the total data set of 65 compounds was divided into two sets, the training set for the internal validation and the test set for the external validation [28]. to construct the two subsets a representation of scores after principal components analysis (pca) was performed based on the abraham’s molecular descriptors of the compounds. the scores plot distributes the solutes according to their physicochemical properties, so it is easier to do a representative division of the global set of compounds into the two subsets. the training set, which is aimed at evaluating the robustness of the model, usually has around 50-70% of the selected compounds. the remaining compounds are used in the test set, which evaluates the predictive ability of the model. the correlation is repeated only with the training set compounds. the results of this second correlation must be compared to the global correlation. if the model is robust, the coefficients of both correlations should be similar. traditional statistical parameters such as the determination coefficient (r 2 ), the standard deviation (sd), and the f statistic were also considered to evaluate the robustness of the model. additionally, the leave-multiple-out cross-validation coefficient, q 2 lmo [29,30], was calculated. this coefficient is a parameter that assesses how the results of a statistical analysis will generalize to an independent data set. the test set was used to perform the external validation of the model, i.e. ensure its predictive ability. the external validation consists in predicting the biological property value of the test set compounds with the training set equation. then, predicted log kp values are plotted against the experimental values. to ensure that the model has a good predictive ability, the slope must be close to 1 and the origin ordinate close to 0. results and discussion establishment of a model to predict kp values table 1 shows the 65 compounds selected to establish the model, their experimental log kp values, chromatographic retention factors, log k values, and mcgowan volumes, v. a multiple linear regression admet & dmpk 6(2) (2018) 140-152 estimation of skin permeation by liquid chromatography doi: 10.5599/admet.512 145 analysis was done between the experimental log kp values and the two parameters log k and v. the set of compounds comprises a wide variety of chemical classes and structural complexity. this is important in order to make the model general enough to predict kp values of compounds of very different chemical nature. results are shown in equation 4 and figure 1. table 1. experimental log kp values, log k values of the neutral species and mcgowan solute’s volume of the compounds selected to establish the model compounds log kp log k mcgowans's volume (v) 2,4-dichlorophenol -4.30 0.556 1.0199 2-amino-4-nitrophenol -6.54 -0.089 1.0491 2-chlorophenol -4.56 0.237 0.8975 2-hydroxybenzoic acid (salycilic acid) -5.33 0.080 0.9904 2-isopropyl-5-methylphenol (thymol) -4.35 0.849 1.3387 2-naphtol (beta-naphthol) -4.65 0.419 1.1441 2-nitrophenol -4.08 0.343 0.9493 2-nitro-p-phenylenediamine -6.66 -0.187 1.0902 2-phenylethanol -5.20 0.003 1.0569 3-methylphenol (m-cresol) -4.89 0.172 0.9160 3-nitrophenol -5.33 0.145 0.9493 4-amino-2-nitrophenol -5.91 -0.010 1.0491 4-bromophenol -4.52 0.361 0.9501 4-chloro-3-methylphenol (4-chloro-m-cresol) -4.34 0.474 1.0384 4-chlorophenol -4.52 0.299 0.8975 4-ethylphenol -4.53 0.380 1.0569 4-hydroxybenzyl alcohol -6.26 -0.560 0.9747 4-hydroxy-methylphenylacetate -5.26 -0.177 1.2722 4-hydroxyphenylacetamide -6.89 -0.606 1.1724 4-hydroxy-phenylacetic acid -6.06 -0.539 1.1313 4-methylphenol (p-cresol) -4.83 0.169 0.9160 4-nitrophenol -5.33 0.091 0.9493 5,5-diethylbarbituric acid (barbital) -7.29 -0.377 1.3739 5-ethyl-5-phenylbarbituric acid (phenobarbital) -6.68 -0.059 1.6999 5-fluorouracil -6.82 -0.977 0.7693 8-methoxypsoralen -5.12 0.376 1.4504 aniline -4.73 0.018 0.8162 aspirin (acetylsalicylic acid) -5.69 -0.165 1.2879 atropine -7.15 0.551 2.2820 benzaldehyde -3.93 0.221 0.8730 benzene -4.27 0.587 0.7164 benzoic acid -5.15 -0.073 0.9317 benzyl alcohol -5.30 -0.113 0.9160 benzyl nicotinate -4.87 0.569 1.6393 caffeine -7.08 -0.487 1.3632 chloropheniramine -6.14 1.009 2.2098 marti rosés et al admet & dmpk 6(2) (2018) 140-152 146 table 1. continued… compounds log kp log k mcgowans's volume (v) codeine -7.20 0.116 2.2057 corticosterone -6.84 0.230 2.7389 dexamethasone -7.27 0.172 2.9132 diclofenac -5.30 0.995 2.0250 diethylcarbamazine -6.15 -0.049 1.7241 digitoxin -8.15 0.642 5.6938 ephedrine -5.49 0.314 1.4385 estradiol -5.61 0.534 2.1988 ethylbenzene -3.00 1.065 0.9982 fentanyl -4.89 1.144 2.8399 fluocinonide -6.33 0.904 3.4603 hydrocortisone -7.22 -0.055 2.7976 ibuprofen -5.30 1.007 1.7771 indomethacin -5.39 1.001 2.5299 isoquinoline -5.11 0.220 1.0443 lidocaine -5.42 0.764 2.0589 methyl 4-hydroxybenzoate -5.12 0.002 1.1313 methyl phenyl ether (anisole) -4.68 0.543 0.9160 naproxen -4.97 0.557 1.7821 nicotine -6.03 0.061 1.3710 o-phenylenediamine -6.70 -0.313 0.9160 phenol -5.27 -0.006 0.7751 p-phenylenediamine -6.98 -0.415 0.9160 progesterone -4.90 1.130 2.6215 resorcinol -6.70 -0.412 0.8338 scopolamine -7.90 -0.036 2.2321 sufentanil -4.84 1.379 3.1051 testosterone -5.54 0.557 2.3827 toluene -3.64 0.826 0.8573 log kp= 4.47(±0.12) + 1.98(±0.13) log k 1.02(±0.07) v (4) n = 63 r 2 = 0.826 sd = 0.461 f = 143 from the 65 substances, only atropine and chloropheniramine were excluded from the regression analysis because their standard residuals were greater than 2.5 in absolute value. both coefficients in equation 4 have statistical meaning at a 95 % level of confidence, and log k has the highest contribution to skin permeation. note that the log k coefficient is almost twice in absolute value the v coefficient. log k contributes in a positive way, which suggests that a high retention factor implies a better dermal absorption. the v coefficient applies a negative correction to the log kp value, more important for those compounds with higher volume. equation (4) is very similar to the previous equation (2). direct comparison of the parameters in both cases reveals almost the same values for the v coefficient and the intercept, and only a slight difference in the log k coefficient. this difference is mainly attributed to the number and variety of compounds studied. since the present work covers a larger number of compounds and they are more representative of a wide chemical space, it is considered that equation 4 might be more robust than eq. 2. admet & dmpk 6(2) (2018) 140-152 estimation of skin permeation by liquid chromatography doi: 10.5599/admet.512 147 figure 1. correlation between log kp experimental values and log k and v of the set of 65 compounds selected to establish the model as described previously, to validate the model, the global set of 63 compounds was divided into two sets: the training set (42 compounds) and the test set (21 compounds). a multiple linear regression between experimental log kp values and log k and v was performed with the 42 compounds of the training set. equation 5 and figure 2a show the results of the regression. log kp= 4.57(±0.13) + 2.06(±0.17)log k 0.98(±0.08) v (5) n = 42 r 2 = 0.832 sd = 0.450 f = 97 q 2 lmo=0.832 figure 2. (a) correlation between log kp experimental values and log k and v of the set of compounds selected in the training set. (b) estimated vs. experimental log kp values for the compounds of the test set comparison of the intercept and the coefficients of eq. 5 to those of eq. 4 demonstrate that nearly the same equation is obtained with less compounds, which means that the model is robust, i.e. with fewer points the physical meaning of the coefficients does not change. adequate determination coefficient, (a) (b) marti rosés et al admet & dmpk 6(2) (2018) 140-152 148 standard deviation and f values were obtained. q 2 lmo was higher than 0.6 and it can be ensured that none of the compounds present in the training set has had a big influence in the final model equation. to perform the external validation, the log kp values of the 21 compounds of the test set were predicted through the training set equation (eq. 5). agreement between experimental and predicted kp values is shown in figure 2b and through equation 6: log kp,pred = 1.08(±0.53) + 0.82(±0.09) log kp,exp (6) n = 21 r 2 = 0.807 sd = 0.469 f = 80 q 2 lmo=0.983 according to statistics, the model shows good prediction ability: the slope and intercept of the trend line are not significantly different from unity and zero, respectively, at 95% confidence level according to student’s t test. the determination coefficient is above 0.6, the correlation cross-validation coefficient is above 0.5 and fisher’s f parameter is significant. prediction of log kp values as a first step, skin permeability coefficients of the neutral species of a set of 20 solutes were determined through the model equation (eq. 4), after measuring their retention factors in the hplc system. table 2 shows the experimental log kp values [26, 27] for these compounds, the values predicted through eq. 4, and also the values predicted by the solvation parameter model (eq. 1). for the latter prediction the knowledge of solvation parameter model descriptors of the compounds is necessary. there is quite good agreement between the experimental log kp values and the values obtained by eq. 4. in most cases the error is around 10%, and is similar to the error presented by the abraham model (eq. 1). figure 3 represents the experimental values against the log kp obtained by both models. the solid line represents the theoretical line of slope equal to one and zero intercept. the plot clearly indicates that the accuracy of both models is very similar, with points equally distributed at both sides of the theoretical line, and with similar dispersion. only famotidine presents an anomalous low log kp value when predicted by the abraham model. in a second step, log kp values were predicted again through eq. 4 and then through eq. 1 for another set of 9 compounds, for which experimental log kp values were not known. only for five of them were the values calculated by patel et al. (5) available. as indicated in table 3, comparison of the data demonstrates that values obtained through chromatographic measurements (eq. 4) and through the abraham model (eq. 1) are in general consistent. however, patel’s predictions are systematically lower compared to the other two models. finally log kp values for four compounds of pharmaceutical interest (capsaicin, oxycodone, tramadol, and warfarin) are proposed for the first time. admet & dmpk 6(2) (2018) 140-152 estimation of skin permeation by liquid chromatography doi: 10.5599/admet.512 149 table 2. comparison of experimental (in vitro) and predicted log kp values compounds in vitro log kp values predicted log kp values [26] [27] this method [15] aminopyrine -6.55 -6.45 -6.97 antipyrine -7.74 -6.56 -7.31 atrazine -5.56 -5.26 -5.45 catechol -6.07 -5.88 -5.96 cortexolone -7.68 -6.72 -5.94 cortisone -7.38 -7.29 -6.74 estriol -7.95 -7.09 -6.67 estrone -6.00 -5.29 -5.11 famotidine -8.35 -7.63 -10.49 flurbiprofen -3.40 -4.60 -4.09 griseofulvin -6.44 -5.93 -5.92 hydroquinone -6.51 -6.32 -6.44 hydroxyprogesterone -6.78 -5.80 -5.08 ketoprofen -4.71 -5.43 -5.06 ketorolac -5.80 -5.79 -5.72 piroxicam -6.02 -6.19 -7.18 prednisolone -7.91 -7.46 -6.90 pregnenolone -6.38 -4.93 -4.70 propanolol -6.05 -4.96 -5.71 ranitidine -7.61 -7.17 -7.97 figure 3. experimental vs. predicted log kp values. for each compound, log kp is predicted through the chromatographic model (●) and through the solvation parameter model (▲) marti rosés et al admet & dmpk 6(2) (2018) 140-152 150 table 3. prediction of log kp values through different estimation models compounds this method [15] [5] 2-toluidine -5.09 -5.22 -5.83 3-xylene -3.37 -3.93 -4.41 cumene -3.11 -3.59 -6.14 n.n-dimethylaniline -4.20 -4.62 -5.29 pyridine -5.56 -5.90 -6.10 capsaicin -5.41 -4.19 oxycodone -5.89 -7.05 tramadol -5.16 -5.02 warfarin -5.37 -5.25 conclusions a methodology has been developed for the prediction of skin permeability of neutral species. the method is based on measurements on an hplc system fitted with a c18 column, coupled to uv-vis and ms detection. the model has been established with a set of 63 compounds with different chemical natures and structural complexity, which ensures the applicability of the model to a wide range of compounds. the chromatographic retention factor is the parameter that makes the highest contribution to the calculation of the permeation through skin, the kp value being greater as the retention increases. on the contrary, the mcgowan volume decreases the permeation value. the model has been validated, showing both good robustness and good prediction ability. the prediction ability has been further tested by comparison of the results to other prediction models (abraham and patel). results from this work are similar (both in accuracy and precision) to the results obtained through abraham’s model. however, the permeation coefficients obtained through patel’s model seem to be slightly lower. the main advantage of the present methodology is that kp can be directly obtained from simple chromatographic retention measurements and the mcgowan’s molecular volume of the compounds, without the need of additional molecular descriptors calculation and/or determination. acknowledgements: we thank magda bordas for her expert contribution to the analytical studies and with edmundo ortega for their excellent technical assistance. we also thank carmen almansa for her very helpful discussion and comments. financial support from the generalitat de catalunya (project di-2014 modality di-esp, resolution eco/1426/2014) and the ministerio de economia y competitividad from the spanish government (ctq2014-56253-p) is acknowledged. references [1] h.a.e. benson, a.c. watkinson. topical and transdermal drug delivery: principles and practice. wiley (2012). [2] k. hoang. dermal exposure assessment: principles and applications. washington: environmental protection agency, office of health and environmental assessment (1992). [3] j. kielhorn, s. melching-kollmuss, i. mangelsdorf. environmental health criteria 235: dermal absorption. geneva: world healt organization press (2006). [4] s. geinoz, r.h. guy, b. testa, p. carrupt. quantitative structure-permeation relationships (qspers) to predict skin permeation: a critical evaluation. pharmaceutical research 21 (2004) 83-92. admet & dmpk 6(2) (2018) 140-152 estimation of skin permeation by liquid chromatography doi: 10.5599/admet.512 151 [5] h. patel, w. ten berge, m.t.d. cronin. quantitative structure-activity relationships (qsars) for the prediction of skin permeation of exogenous chemicals. chemosphere 48(6) (2002) 603–613. [6] s. mitragotri, y.g. anissimov, a.l. bunge, h.f. frasch, r.h. guy, j. hadgraft, g.b. kasting, m.e. lane, m.s. roberts. mathematical models of skin permeability: an overview. international journal of pharmaceutics 418 (2011) 115-129. [7] a. tezel, s. mitragotri. on the origin of size-dependent tortuosity for permeation of hydrophilic solutes across the stratum corneum. journal of controlled release 86 (2003) 183-186. [8] a. wilschut, w. ten berge, p.j. robinson, t.e. mckone. estimating skin permeation. the validation of five mathematical skin permeation models. chemosphere 30 (1995) 1275-1296. [9] m.j. kamlet, r.m. doherty, j.l.m. abboud, m.h. abraham, r.w. taft. solubility: a new look. chemtech 16(9) (1986) 566-576 [10] m.h. abraham. scales of solute hydrogen-bonding: their construction and application to physicochemical and biochemical processes. chemical society reviews 22 (1993) 73-83. [11] m. vitha, p.w. carr. the chemical interpretation and practice of linear solvation energy relationships in chromatography. journal of chromatography a 1126 (2006) 143-194. [12] m.h. abraham, h.s. chadha, f. martins, r.c. mitchell, m.w. bradbury, j.a. gratton. hydrogen bonding part 46: a review of the correlation and prediction of transport properties by an lfer method: physicochemical properties, brain penetration and skin permeability. pest management science 55 (1999) 78-88. [13] m.h. abraham, chadha, h.s., mitchell, r.c. (1995b). the factors that influence skin penetration of solutes. journal of pharmacy and pharmacology, 47, 8-16. [14] m.h. abraham, f. martins, r.c. mitchell. algorithms for skin permeability using hydrogen bond descriptors: the problem of steroids. journal of pharmacy and pharmacology 49 (1997) 858-865. [15] m.h. abraham, f. martins. human skin permeation and partition: general linear free-energy relationship analyses. journal of pharmaceutical sciences 93 (2004) 1508-1523. [16] k. zhang, m. chen, g.k.e. scriba, m.h. abraham, a. fahr, x. liu. human skin permeation of neutral species and ionic species: extended linear free-energy relationship analyses. journal of pharmaceutical sciences 101 (2012) 2034-2044. [17] b.j. neely, s.v. madihally, r.l. robinson jr, k.a.m. gasem. nonlinear quantitative structureproperty relationship modeling of skin permeation coefficient. journal of pharmaceutical sciences, 98 (2009) 4069-4084. [18] p. buchwald, n. bodor. a simple, predictive, structure-based skin permeability model. journal of pharmacy and pharmacology 53 (2001) 1087-1098. [19] d. neumann, o. kohlbacher, c. merkwirth, t. lengauer. a fully computational model for predicting percutaneous drug absorption. journal of chemical information and modeling 46 (2006) 424-429. [20] m. hidalgo-rodríguez, s. soriano-meseguer, e. fuguet, c. ràfols, m. rosés. evaluation of the suitability of chromatographic systems to predict human skin permeation of neutral compounds. european journal of pharmaceutical sciences 50(5) (2013) 557–568. [21] acd/percepta, www.acdlabs.com, advanced chemistry development, inc., release (build 2726, 5 may 2014). [22] a.d. jones, i.p. dick, j.w. cherrie, m.t.d. cronin, j.j.m. van de sandt, d.j. esdaile, s. iyengar, w. ten berge, s.c. wilkinson, c.s. roper, s. semple, c. de heer, f.m. williams. cefic workshop on methods to determine dermal permeation for human risk assessment. european chemical industry council, december (2004) pp 1–86 [23] j.j. van de sandt, j.a. van burgsteden, s. cage, p.l. carmichael, i. dick, s. kenyon, g. korinth, f. larese, j.c. limasset, w.j. maas. in vitro predictions of skin absorption of caffeine, testosterone, and benzoic acid: a multi-centre comparison study. regulatory toxicology and pharmacology 39 (2004) 271–281. http://www.acdlabs.com/ marti rosés et al admet & dmpk 6(2) (2018) 140-152 152 [24] s.c. wilkinson, w.j.m. maas, j.b. nielsen, l.c. greaves, j.j.m. van de sandt, f.m. williams. interactions of skin thickness and physicochemical properties of test compounds in percutaneous penetration studies. international archives of occupational and environmental health 79 (2006) 405–413. [25] f. heuber, f. ouvrard-baraton, j.p. biesse, p. courtellemont, c.m. vincent, j.p. marty. pig ear skin as a model for in vitro percutaneous absorption studies: preliminary results of an inter-laboratory validation. in: k.r. brain, k.a. walters, editors. perspectives in percutaneous penetration, vol. 6a. cardiff: sts publishing ltd (1998). [26] b.e. vecchia, a.l. bunge. transdermal drug delivery (pp. 25–55). new york: marcel dekker (2003). [27] h. baba, j. takahara, h. mamitsuka. in silico predictions of human skin permeability using nonlinear quantitative structure-property relationships models. pharmaceutical research 32 (2015) 23602371. [28] k. roy. on some aspects of validation of predictive quantitative structure–activity relationship models. expert opinion on drug discovery 2(12) (2007) 1567–1577. [29] a. tropsha, p. gramatica, v. gombar. the importance of being earnest: validation is the absolute essential for successful application and interpretation of qspr models. qsar & combinatorial science 22(1) (2003) 69–77. ©2018 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ manuscript doi:10.5599/admet.2.3.50 157 admet & dmpk 2(3) (2014) 157-167; doi: 10.5599/admet.2.3.50 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index review prospects on the hyphenated electrochemistry and mass spectrometry as a practical analytical technique in the assessment of oxidative drug metabolism eslam nouri-nigjeh department of pharmacology, david geffen school of medicine, university of california at los angeles, ca 900956948, usa e-mail: enouri@mednet.ucla.edu, eslamnouri@gmail.com; tel.: +310-825-4321 received: september 07, 2014; revised: september 11, 2014; published: september 16, 2014 abstract for several years, electrochemistry in combination with mass spectrometry has been the subject of attention and development for studying the oxidative drug metabolism at early stages of new drug development. though the technique could successfully imitate the in vivo oxidative drug metabolism initiated by electron transfer, it lags in the imitation of reactions initiated by either hydrogen atom transfer or oxygen atom transfer. the prospect of using electrochemistry as a practical analytical technique in the imitation of oxidative drug metabolism, therefore, relies on the extension of its utility toward covering those reactions initiated by hydrogen atom transfer and oxygen atom transfer. in this brief critical review, i discuss potential electrochemical techniques that can benefit the application of electrochemistry beyond electron transfer reactions. keywords electrochemistry; mass spectrometry; oxidative drug metabolism 1. introduction due to the large number of drug candidates at the early stages of new drug development, there is a tangible need for a fast and facile analytical technique to assess the oxidative drug metabolism. hyphenated electrochemistry and mass spectrometry technique has been in the center of attention for the last several years to achieve the fast and facile assessment of oxidative drug metabolism at the early stages of new drug development [1,2]. this brief review will present the current status of the technique, and possible efforts to extend its application in a more practical sense in the assessment of oxidative drug metabolism. cytochrome p450s are the main family of monooxygenase enzymes that are involved in phase i in vivo oxidative drug metabolism pathways [3]. the reactive pocket of these enzymes include a prosthetic iron protoporphyrin ix group that is axially anchored with a cysteine residue in the protein backbone [3]. interaction of the substrate drug compound with a specific part of the reactive pocket triggers the activation of molecular oxygen in a chain of proton and electron transfer events to generate oxo-ferryl http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:enouri@mednet.ucla.edu mailto:eslamnouri@gmail.com e. nouri-nigjeh admet & dmpk 2(3) (2014) 157-167 158 radical cation species [4]. these species are considered the main reactive species produced during the catalytic activation of molecular oxygen with cytochrome p450s, and are highly reactive toward the oxidation of substrate drug compounds that are already located in the vicinity of the reactive pocket of the enzyme [5]. cytochrome p450s and their main reactive species, i.e. oxo-ferryl radical cations, promote different oxidative metabolism pathways; namely, those initiated by electron transfer reaction, such as ndealkylation of tertiary amine moieties, those initiated by hydrogen atom transfer, such as o-dealkylation pathways, and those initiated by oxygen atom transfer such as n-oxidation pathways. there are already a great number of studies directed toward exploration of reactive species generated during catalytic activation of molecular oxygen with cytochrome p450s, and their reactivities toward promotion of different oxidative metabolism pathways [6]. the products of oxidative drug metabolism, that are usually more hydrophilic than drug substrates, conjugate with glutathione to be secreted from the cells in second and third phases of oxidative drug metabolism [3]. direct electrochemical oxidations, unlike reactive species generated during catalytic activation of molecular oxygen by cytochrome p450s, can only promote oxidation reactions initiated by electron transfer reactions [7,8]. this means that a subset of in vivo metabolism pathways that are initiated by either hydrogen atom transfer, or oxygen insertion reactions might be difficult or impossible to be imitated merely using direct electrochemical oxidations [8]. among the important in vivo oxidative drug metabolism that are reportedly not mimicked with direct electrochemical oxidation are n-oxidation, o-dealkylation, and certain hydroxylation processes [8]. this issue highlights the immense need for development of electrochemical techniques that are versatile in promoting different oxidative metabolism pathways, and to have a practical tool for the fast and facile assessment of oxidative drug metabolism at early stages of new drug development. scheme 1. direct electrochemical oxidation of lidocaine (a test drug compound) resulting in the n-dealkylation of tertiary amine moiety, and para-hydroxylation reaction of the aromatic ring. n-dealkylation proceeds through generation of an imine intermediate and hydrolysis, while para-hydroxylation proceeds through the generation of a wheland-type intermediate and deprotonation. electrochemistry has shown already great promises in the mimicry of oxidative drug metabolism for a large number of drug models [9]. the reactions which are already imitated using electrochemistry are those reactions that are initiated by electron transfer. direct electrochemical oxidation of lidocaine, as a test drug compound, results in two major oxidative metabolites, n-dealkylation, and para-hydroxylation [10]. the detailed mechanisms for these reactions are shown in scheme 1. n-dealkylation of tertiary amine imine intermediate wheland-type intermediate intermediate admet & dmpk 2(3) (2014) 157-167 hyphenated electrochemistry & ms doi: 10.5599/admet.2.3.50 159 compounds have already been studied extensively [11,12]. n-dealkylation reaction proceeds through electron transfer from nitrogen, formation an imine intermediate, hydrolysis, intra-molecular rearrangement, and n-dealkylation [13]. para-hydroxylation pathway follows an ece mechanism with the first electron transfer, deprotonation, and second electron transfer that precede nucleophilic attack, and generating a wheland-type intermediate in which substituents determines the final regioselectivity of the nucleophilic attack [14]. electrochemical reactions are innately selective toward electron rich moieties, and a type of selectivity can be obtained by having controlled potential on the electrode surface, and a selectivity of reaction that can be obtained due to the onset potential of different reactions [7]. the electrochemical reactions can be integrated in microchips for a better application in industry [15]. moreover, the electrochemical reactions can be utilized to perform preparative synthesis of highly interesting metabolites. this critical review presents the advantages of using hyphenated electrochemistry and mass spectrometry in the assessment of oxidative drug metabolism and discusses different tools such as the use of electrochemical generated reactive oxygen species, and the electrode surfaces modified with prophyrins or enzymes as candidates to be used for this purpose. 2. electrochemistry combined with mass spectrometry although direct electrochemical oxidation confines the extension of using electrochemistry in the imitation of oxidative drug metabolism only to those reactions initiated by electron transfer reaction, it provides undisputable advantages to obtain insights in the mechanism of reactions. direct coupling of a flow-through electrochemical cell with a mass spectrometry ushers in the near-real-time analysis of the oxidation products generated on the electrode surface. this facilitates the impact study of different electrochemical parameters such as applied potential, electrode material and solvent type on the distribution of reaction products. reaction intermediates, and side products can be revealed and studied using hyphenated electrochemistry and mass spectrometry [12]. the addition of scavengers such as radical or cation scavengers would produce the evidence for the presence of certain short-living critical intermediates and their impact on the final product distribution [16]. the same approach can be valued for studying the reactivity of those intermediates with cell components [17], and their possible conjugated products with glutathione in phase ii drug metabolism [18]. in the case of electrochemical oxidation of acetaminophen, and its counterpart in vivo oxidation, a toxic intermediates, n-acetyl-p-benzoquinone imine (napqi), can be generated that should be detoxified by conjugation with glutathione [18]. interestingly, the detoxification of napqi with glutathione can follow two different nucleophilic pathways illustrated in scheme 2. the conjugated products have different retention times and can be resolved with reverse phase chromatography, figure 1a. the fragmentation patterns of the conjugated products are compared with the fragmentation of glutathione in figure 1b. the fragmentation of the conjugated products is distinctive in nature possibly due to the steric hindrances imposed during fragmentations, figure 1c, and figure 1d. this in overall shows the advantage of using hyphenated electrochemistry and mass spectrometry in the exploration of different types of conjugates that might be difficult to be studied during in vivo experiments. e. nouri-nigjeh admet & dmpk 2(3) (2014) 157-167 160 scheme 2. two different nucleophile reaction pathways for the glutathione and napqi which potentially result in the generation of two distinct conjugated products. 0 3 6 9 0,0 0,5 1,0 1,5 100 150 200 250 300 350 0,0 0,5 1,0 200 250 300 350 400 450 500 0,0 0,1 0,2 0,3 200 250 300 350 400 450 500 0,0 0,2 0,4 hooc h n h n cooh nh2 o sh o b2 y2 c1 z1 b1 y1 cooh nh n h cooh h2n o s o o hn o h cooh nh n h cooh h2n o s oho n o h d c ba b 2 y 2 308 233 179 162 b 1 y 1 c 1 z 1 io n i n te n s it y io n i n te n s it y m/z 457 382 328 b 2 y 2 b 1 y 1 c d time / min 328 382 457 m/z 311 b 2 y 2 b 1 y 1 c 1 z 1 figure 1. conjugated products of glutathione and napqi that are resolved on a reverse phase chromatography column (a), ms/ms fragmentation pattern of glutathione with the fragments resulted from the break downs on two "by" positions and one "cz" position (b), corresponding fragmentation patterns for the first and second napqi and glutathione conjugate products separated on the reverse chromatography column (c and d). admet & dmpk 2(3) (2014) 157-167 hyphenated electrochemistry & ms doi: 10.5599/admet.2.3.50 161 3. extension of electrochemistry to the reactions beyond electron transfer 3.1. electrochemical pulses electrochemical pulses consisting of repetitive oxidative and reductive steps have shown great promises in high yield and selective generation of drug metabolites [10]. for metallic electrodes such as gold and platinum, presence of an immediate reduction step would foster the reduction of oxide layer generated on the surface and removing the blocking reagents from electrode surface which in overall can result in higher oxidation yields. for the oxidation of lidocaine, hydroxylation pathway was enhanced by 50 times using electrochemical pulses with 1 hz frequency [10]. in addition, using electrochemical potential pulses introduces a new parameter that can be used to tune the selectivity of electrochemical reactions. interestingly, for lidocaine as a test drug compound, oxidations under higher frequencies favored ndealkylation pathway while for the longer pulse cycles this selectivity diverted to aromatic hydroxylation and n-oxidation products [10]. using electrochemical pulses not only can provide higher yields and selectivities, but also can promote reactions that were deemed impossible by constant potential oxidation. biotransformation of phenacetin to acetaminophen was achieved using this approach [16]. the reaction included the generation of a cation intermediate, hydrolysis and reduction of napqi on the electrode surface to generate acetaminophen. mechanistic studies revealed the generation of napqi in the presence of a strong nucleophile, and pquinone as the final product of direct oxidation that was visualized using mass spectrometry by adding glutathione. there is much to perform to use the technique for a wide range of drug compounds and under different conditions that would be the matter of time and effort to be invested. 3.2. electrochemically generated reactive oxygen species electrochemistry can be used to generate different reactive oxygen species on the surface. this type of reactions can be classified as indirect electrochemical reactions, in which electrochemistry generates reactive oxygen species to react with drug compounds. from this type of reactions, i will introduce the electrochemical formation of highly reactive hydroxyl radicals using electrochemically assisted fenton reaction [19], generation of hydrogen peroxide [20], and oxidation of water on boron-doped diamond electrode (bdd) [21]. hydroxyl radicals are highly reactive and easily promote the hydrogen atom transfer reactions; however, due to high reactivity, these species are not selective. a selective oxidation might be performed by electrocatalytic activation of hydrogen peroxide on platinum electrode surface to generate hypothetical platinum-oxo species that are capable of selective oxidation reactions. scheme 3 shows the type of reaction products that each of the reactive species presented in this section would produce when reacted with lidocaine as a test drug compound. electrochemically assisted fenton reaction: an earlier study indicated that the electrochemical reduction of fe 3+ to fe 2+ can promote the activation of hydrogen peroxide and the generation of hydroxyl radicals [19]. these radicals are highly reactive and less selective that lead to the generation of a wide range of product. a set of drug compounds were studied and it was explored that this type of reactions can imitate a wide range of oxidative drug metabolism [19]. in particular, reaction with hydroxyl radicals is capable of generating a wide range of hydroxylation products. a similar study for the oxidation of lidocaine also highlighted the fact that this type of reaction is capable of producing all types of in vivo metabolites of lidocaine (scheme 3). yet, due to the high reactivity of hydroxyl radicals, and their non-selective behavior, generation of a large quantity of unwanted side products is unavoidable. though reaction with hydroxyl e. nouri-nigjeh admet & dmpk 2(3) (2014) 157-167 162 radicals could generate all known metabolites of lidocaine, there is clear presence of reaction products with the radical recombination with solvent molecules. scheme 3. the reaction of different types of electrochemically generated reactive oxygen species with lidocaine as a test drug compound and the corresponding metabolites generated from each of the reactions electrochemical reduction of molecular oxygen: electrochemical reduction of molecular oxygen from aprotic solvents can result in superoxide anions, that in the presence of residual water a chain of reactions can follow up to hydrogen peroxide [22]. interestingly, superoxide anion can activate hydrogen peroxide in haber-weiss reaction type to produce highly reactive hydroxyl radicals. a past study to reduce molecular oxygen from acetonitrile in the presence of lidocaine generated a large quantity of n-oxidation metabolite, while hydroxylation products were not detectable [13]. this highlights the fact that the main reactive species generated from the reduction of molecular oxygen under mentioned condition was hydrogen peroxide. hydrogen peroxide reacts with the tertiary amine moiety of lidocaine and due to dehydrolysis noxidation metabolite, a highly reactive metabolite, can be formed. electrocatalytic activation of hydrogen peroxide: electrocatalytic behavior of platinum electrodes in the oxidation of hydrogen peroxide is a well-known phenomenon. yet, the generation of hypothetically reactive intermediates during electrocatalytic activation of hydrogen peroxide on platinum electrode to promote selective oxidation pathways is yet to be explored. in a recent study, we showed that oxidation of hydrogen peroxide on a platinum electrode can produce both meta and para isoforms of aromatic hydroxylation of lidocaine, in the absence of benzylic hydroxylation product, scheme 3 [20]. this helped us to suggest a more selective oxidation reaction in the absence of hydroxyl radicals due to generation of hypothetical platinum-oxo species that are capable of promoting selective oxygen insertion reactions. a admet & dmpk 2(3) (2014) 157-167 hyphenated electrochemistry & ms doi: 10.5599/admet.2.3.50 163 thorough mechanistic study needs to be performed, and the results of this relatively simple approach can widen the selective oxidation pathways of drug candidates. bdd electrodes: oxidation of water on bdd electrode can generate hydroxyl radicals capable of nonselective oxidation reactions. bdd electrode due to high polarization of molecular oxygen evolution has a wide range of applied potential window for the oxidation of water that results in the formation of hydroxyl radicals. this approach due to its simplicity has gained an ever increasing attention in the mimicry of oxidative drug metabolism, as well as electrochemical cleavage of peptides and proteins [21]. 3.3. electrochemical mimicry with porphyrins the prosthetic group of cytochrome p450s is an iron protoporphyrin, and hence metalloporphyrins can be in general good surrogates to mimic the catalytic behavior of cytochrome p450s [23]. on a mechanistic view, metallo-oxo species can be formed either with reaction with a mono-oxygen atom donor such as mcpba, as shown in scheme 4, or with hydrogen peroxide after dehydrolysis that are considered shunt pathways [3]. by using electrochemistry, it is possible to activate molecular oxygen in a two electron, two proton transfer reaction steps, as shown with the main catalytic cycle in scheme 4 [24]. the consecutive electron and proton transfers result in a heterolytic activation of oxygen-oxygen bond and generation of metallo-oxo species. the oxygen transfer reaction would re-generate the initial state of the porphyrin. in the case of electrochemistry, there is a need for a relatively strong acid such as acetic acid to perform the protonation, in addition to an axial coordinator that is usually 1-methyl imidazole [25]. the mechanism was verified by the cyclic voltammetry of hemin from dmso, 0.1 m tbap, under argon atmosphere, that shows a single reversible electron transfer, corresponding to the first electron transfer, shown in figure 2 [26]. interestingly, in the presence of dissolved molecular oxygen, this single redox peak changes to two distinct electron transfer peaks, supporting well the mechanism shown in scheme 4. the second cycle under oxygen atmosphere shows the absence of first electron transfer, this indicates that the species generated during the first cycle is stable and there is not fe 2+ available for re-reduction during the following cycle. in the presence of a strong acid, 100 mm acetic acid, instead of two separate peaks there is a single broad peak that is totally irreversible, and reproduces in the consecutive scans. this may indicate the generation of reactive species and their release. e. nouri-nigjeh admet & dmpk 2(3) (2014) 157-167 164 scheme 4. catalytic activation of molecular oxygen with hemin, as a mimetic surrogate for cytochrome p450s, that includes two electron transfer, and two proton transfer steps, resulting in the formation of highly reactive oxo ferryl radical cations. in addition, there are two shunt pathways by either reaction with hydrogen peroxide, or a mono oxygen atom donor such as m-cpba. figure 2. cyclic voltammograms of hemin on the glassy carbon electrode from dmso, 0.1 m tbap, under argon atmosphere (black), the first cycle under oxygen atmosphere and second cycle (red and green respectively), and the cycle under oxygen atmosphere and in the presence of 100 mm acetic acid (blue line) [26]. the product distribution depends on the type of metallic center and presence of electron-withdrawing substituent on the porphyrin ring. generally, having stronger electron-withdrawing substituents on the porphyrin ring makes the reacting metallo-oxo species more reactive and promotes more difficult reactions. a comparative study revealed that while manganese porphyrin in the presence of m-cpba only generates n-dealkyaltion product, under the same reaction condition, fluorinated manganese porphyrin admet & dmpk 2(3) (2014) 157-167 hyphenated electrochemistry & ms doi: 10.5599/admet.2.3.50 165 generates all the hydroxylation products, with a much larger quantity of n-dealkylation product. using asymmetric porphyrins can also result in stereospecific promotion of oxidative drug metabolism [27]. the metalloporphyrins can be immobilized on the electrode surface through electropolymerization [28,29], or through self-assembled monolayers [9]. figure 3a shows the hemin accumulated on the electrode surface, with a similar electrochemical behavior as hemin from solution. the significant difference is that there is a higher current density in the presence of a strong proton donor that might be due to the direct reduction of electrochemically generated reactive oxygen species. the main issue with the immobilized porphyrin electrochemistry is the instability of electrode surface that might be deteriorated after few cycles. figure 3b shows that the hemin from electrode surface is not stable and would dissolved to the solvent after few cycles of redox potentials. on the other hand using single atomic layer of porphyrins on the electrode surface has proven advantages for spectroscpoic purposes [30], but unfortunately the stability of surface during oxidation reaction is a perplexing challenge to overcome. given the wide variety of metallopoprphyrins with different metallic centers, substituents, and stereochemistry this can be a highly interesting topic of development for further mimicry studies of oxidative drug metabolism. figure 3. (a) cyclic voltammograms of hemin accumulated on a glassy carbon electrode under argon atmosphere (black), the first cycle under oxygen atmosphere and second cycle (red and green respectively), and the cycle under oxygen and in the presence of acetic acid (blue line). (b) repetitive cycles of hemin accumulated on the glassy carbon surface in dmso, 0.1 m tbap, 100 mm acetic acid [26]. 3.4. electrochemistry combined with electrode surfaces modified with enzymes a highly stereospecific and chemoselective mimicry can be achieved using synthetic heme peptides and enzymes, such as peroxidases and cytochrome p450s [9]. using an enzyme as a mimetic model is challenging due to the complications with charge transfer reactions, protein stability and inactivation due to immobilization. using isolated cytochrome p450s requires additional reductase cofactor. a protein engineering strategy might also help to obtain sufficient interaction of the enzyme with electrode surfaces using an enzyme reductase fused protein [31]. this is due to the fact that the enzyme immobilization should not inactivate or change the enzyme behavior. the possibility of using electrodes modified with enzymes in micro fluidic can add value to their industrial applications [32]. different strategies using enzymes in mimicry of oxidative drug metabolism are already discussed in an earlier review article [9]. e. nouri-nigjeh admet & dmpk 2(3) (2014) 157-167 166 4. conclusions hyphenated electrochemistry and mass spectrometry can gain a greater momentum in the practical senses for the assessment of oxidative drug metabolism. however, there is a great need to extend the application of electrochemistry to the reactions initiated by those other than electron transfer. this brief review presented a few promising electrochemical approaches that can help to achieve this goal. though electrochemically generated reactive species showed a great promise for the oxidation of lidocaine as a test drug compound, high reactivity of hydroxyl radicals and their non-selective behavior would refrain their use. electrochemically generated platinum-oxo species, though showed specific oxygen insertion reactions, need to be studied in the context of different compounds. the metalloporphyrins are interesting due to promoting highly selective reaction pathways with a wide selection of metallic centers and substituents; however, there is a need to develop stable electrode modifications. using enzymes as mimetic models can promote sterospecific and chemoselective reactions. yet, there are complications with electron transfer to enzymes, degradation, and deactivation due to immobilization. references [1] h. faber, m. vogel, u. karst, analytica chimica acta, 834 (2014) 9-21. [2] m. vogel, u. karst, analytical and bioanalytical chemistry, 403(2) (2012) 333-334. [3] p.r. ortiz de montellano, cytochrome p450: structure, mechanism, and biochemistry. third ed 2005, new york: kluwer academic/plenum publishers. [4] s. shaik, d. kumar, s.p. de visser, a. altun, w. thiel, chemical reviews, 105(6) (2005) 2279-2328. [5] s. shaik, s. cohen, y. wang, h. chen, d. kumar, w. thiel, chemical reviews, 110(2) (2010) 949-1017. [6] b. meunier, s.p. de visser, s. shaik, chemical reviews, 104(9) (2004) 3947-3980. [7] u. jurva, h.v. wikström, a.p. bruins, rapid commun. mass spectrom., 14(6) (2000) 529-533. [8] u. jurva, h.v. wikström, a.p. bruins, rapid commun. mass spectrom., 17(8) (2003) 800-810. [9] e. nouri-nigjeh, r. bischoff, a.p. bruins, h.p. permentier, current drug metabolism, 12(4) (2011) 359371. [10] e. nouri-nigjeh, h.p. permentier, r. bischoff, a.p. bruins, analytical chemistry, 83(14) (2011) 55195525. [11] l.r. hall, r.t. iwamoto, r.p. hanzlik, the journal of organic chemistry, 54(10) (1989) 2446-2451. [12] u. jurva, p. bissel, e.m. isin, k. igarashi, s. kuttab, n. jr castagnoli, journal of american chemical society, 127(35) (2005) 12368-12377. [13] e. nouri-nigjeh, h.p. permentier, r. bischoff, a.p. bruins, analytical chemistry, 82 (2010) 7625–7633. [14] l. eberson, k. nyberg, accounts of chemical research, 6(3) (1973) 106-112. [15] m. odijk, a. baumann, w. lohmann, f.t.g. van den brink, w. olthius, u. karst, lab on a chip, 9(12) (2009) 1687-1693. [16] e. nouri-nigjeh, r. bischoff, a.p. bruins, h.p. permentier, analyst, 136(23) (2011) 5064-5067. [17] w. lohmann, u. karst, analytical chemistry, 79(17) (2007) 6831-6839. [18] w. lohmann, u. karst, analytial and bioanalytical chemestry, 386(6) (2006) 1701-1708. [19] u. jurva, h.v. wikström, a.p. bruins, rapid communications in mass spectrometry, 16(20) (2002) 1934-1940. [20] e. nouri-nigjeh, a.p. bruins, r. bischoff, h.p. permentier, analyst, 137(20) (2012) 4698-4702. [21] j. roeser, n.f.a. alting, h.p. permentier, , a.p. bruins, r. bischoff, analytical chemistry, 85(14) (2013) 6626-6632. [22] j. wilshire, d.t. sawyer, accounts of chemical research, 12(3) (1979) 105-110. admet & dmpk 2(3) (2014) 157-167 hyphenated electrochemistry & ms doi: 10.5599/admet.2.3.50 167 [23] t. johansson, l. weidolf, u. jurva, rapid commun mass spectrom, 21(14) (2007) 2323-2331. [24] s.r. bell, j.t. groves, journal of the american chemical society, 131(28) (2009) 9640-9641. [25] s.e. creager, s.a. raybuck, r.w. murray, journal of the american chemical society, 108(14) (1986) 4225-4227. [26] e. nouri-nigjeh, h.p. permentier, r. bischoff, a.p. bruins, 59th annual meeting of the international society of electrochemistry, seville, spain (2008). [27] k. chen, l. que jr, chemical communications, (1999) (15): p. 1375-1376. [28] f. bedioui, j. devynck, c. bied-charreton, accounts of chemical research, 28(1) (1995) 30-36. [29] f. bedioui, j. devynck, c. bied-charreton, journal of molecular catalysis a: chemical, 113(1-2) (1996) 3-11. [30] a. bonifacio, d. millo, p. keizers, r. boegschoten, j. commandeur, n. vermeulen, c. gooijer, g. van der zwan, journal of biological inorganic chemistry, 13(1) (2008) 85-96. [31] s.j. sadeghi, r. meirinhos, g. catucci, v.r. dodhia, g.d. nardo, g. gilardi, journal of the american chemical society, 132(2) (2009) 458-459. [32] a. fantuzzi, e. capria, l.h. mak, v.r. dodhia, s.j. sadeghi, s. collins, g. somers, e. hug, g. gilardi, analytical chemistry, 82(24) (2010) 10222-10227. ©2014 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ in vitro biomimetic hplc and in vivo characterisation of gm6, an endogenous regulator peptide drug candidate for amyotrophic lateral sclerosis doi: 10.5599/admet.547 176 admet & dmpk 6(2) (2018) 176-189; doi: http://dx.doi.org/10.5599/admet.547 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper in vitro biomimetic hplc and in vivo characterisation of gm6, an endogenous regulator peptide drug candidate for amyotrophic lateral sclerosis klara valko* 1,2 , mark kindy 3,4 , james evans 5 , dorothy ko 6 1 bio-mimetic chromatography ltd, business & technology centre, stevenage, sg1 2dx united kingdom 2 department of biological and pharmaceutical chemistry, ucl school of pharmacy london wc1n 1ax united kingdom 3 department of pharmaceutical sciences, college of pharmacy, university of south florida, tampa, fl, usa. 4 james a. haley vamc, tampa, fl, usa. 5 phenovista, san diego, ca, usa 6 genervon biopharmaceuticals llc, pasadena, ca, usa *corresponding author e-mail: k.valko@ucl.ac.uk and klara_valko@bio-mimetic-chromatography.com; tel.: +44 7521 989558 received: may 08, 2018; revised: may 28, 2018; published: june 16, 2018 abstract amyotrophic lateral sclerosis (als) is an idiopathic, fatal neurodegenerative disease of the human motor system. subunits of the 33-amino acid containing motorneurontrophic factor (mntf) have been investigated and gm6 has been found as a potential peptide therapeutic for als. this linear peptide drug candidate has been characterized by hplc based physicochemical and biomimetic measurements to estimate its in vivo distribution behavior and to estimate its cell penetration and brain to plasma concentration ratio. the free tissue concentration vs time profile has been estimated using the measured physicochemical and biomimetic properties of the intact gm6 molecules and its microsomal stability. the in vitro and in vivo measurements supported the estimated in vivo distribution behavior of gm6. keywords motoneuron degeneration; peptide drug candidate; adme of peptide drugs; biomimetic properties; pk profile; free tissue concentration vs time pk introduction amyotrophic lateral sclerosis (als) is an idiopathic, fatal neurodegenerative disease of the human motor system. it is the most common motoneuron disease in human affecting approximately 2 people in 100,000 in both the us and europe and therefore it is considered as a rare orphan disease [1]. als manifests itself in the death of motoneurons which control the voluntary muscles. the cause of als is not known in 90 to 95 % of cases while familial inherited causes have been identified in 5 to 10 % of cases. there is no cure for als. the clinical onset of the disease usually starts at around age 60 and the survival time is around 2 to 4 years. it is very difficult to determine a specific date for the onset of the disease and the potentially long duration between the onset of pathological changes and the manifestation of the clinical disease. patients can present with bulbar-onset disease (about 25 %) or limb-onset disease (about http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:k.valko@ucl.ac.uk mailto:klara_valko@bio-mimetic-chromatography.com admet & dmpk 6(2) (2018) 176-189 in vitro and in vivo characterization of gm6 doi: 10.5599/admet.547 177 70 %) or with initial trunk or respiratory involvement (5 %). atypical modes of presentation can include weight loss, which is an indicator of a poor prognosis, cramps, emotional lability and frontal lobe-type cognitive dysfunction. the pathophysiological mechanisms underlying the development of als seem multifactorial with emerging evidence of a complex interaction between genetic and molecular pathways [2,3]. als is relentlessly progressive – 50 % of patients die within 30 months of symptom onset and about 20 % of patients survive between 5 10 years after symptom onset. some als subtypes tend to lead to a better prognosis; for example, flail-limb variants als and progressive muscular atrophy. both are predominantly lower motoneuron forms that are characterised by slower progression. advances in the understanding of the glutamate neurotransmitter system and the discovery of causal genes linked to the development of familial als has stimulated research interest. the survival in als is now understood to be dependent on several factors, such as clinical presentation, the rate of disease progression, early presence of respiratory failure and the nutritional status of the patients. the process by which motoneurons die remains unclear with several inter-related mechanisms suggested [1]. the neurodegeneration in als may result from any of the following: complex interactions of glutamate excitotoxicity, generation of free radicals, cytoplasmic protein aggregates, sod1 enzymes, combined with mitochondrial dysfunction, and disruption of axonal transport processes through the accumulation of neurofilament intracellular aggregates. riluzole has been approved by the fda for 20 years and inhibits the release of glutamic acid from cultured neurons, from brain slices, and from corticostriatal neurons in vivo. it increases the als patients’ life by two to three months by blocking glutamatergic neurotransmission in the cns. there is an unmet medical need to lengthen the life expectancy, slow down the disease progression and ultimately cure als patients. discovering a therapeutic agent that significantly increases the life expectancy of als patients or cures the disease would provide a significant breakthrough in this field. potential disease mechanisms protein misfolding and protein aggregation are known characteristics of various diseases, including major neurodegenerative disorders such as als, alzheimer’s and parkinson’s diseases. while the mechanistic link between protein aggregation and cell degeneration remains unclear, approaches to halt or reverse the aggregation process are a focus of current drug discovery approaches. insulin-like growth factor-i (igf-i) receptors are present in the spinal cord and, as members of the neurotrophin receptor family, igf-i receptors mediate signal transduction via the tyrosine kinase domain. igf-i was found to prevent the loss of choline acetyltransferase activity in embryonic spinal cord cultures, as well as reducing the programmed cell death of motor neurons in vivo during normal development or following axotomy [4]. over the past 30 years, glucose intolerance has been reported in a significant percentage of als patients. a reduction in glucose receptor space and insulin resistance has been postulated for a number of neuromuscular diseases [5]. the structurally related peptides, insulin and insulin-like growth factors (igf-i and igf-ii), have neurotrophic properties and potentially could be of therapeutic value in human neurodegenerative disorders [6]. the nerve growth factor (ngf) was discovered by levi-montalcini and cohen some 50 years ago [7] which supported the concept that secreted molecules produced by the target of a developing neuron are required for it to survive programmed cell death (pcd). this neurotrophic hypothesis provided the klara valko et al. admet & dmpk 6(2) (2018) 176-189 178 background for the identification of neurotrophic factors (ntfs). the motor neuron trophic factor (mntf) is a 33-amino acid peptide. it was identified, isolated and characterized by using monoclonal antibodies against a 35 kd rat peptide, isolated from 3-week-old rat muscle extract. mntf was subsequently derived by the cloning of a recombinant mntf1-f6 gene from a human retinoblastoma cdna expression library (clontech) and analyzed to determine its amino acid and cdna sequences [8]. gm6 is a 6 amino acid analogue of the mntf active site. gm6 is a cationic linear peptide drug consisting of 6 amino acids in their natural configurations (phe-ser-arg-tyr-ala-arg acetate salt) [9,10]. mntf 33 (1 µg/ml) and mntf 10 (1 µg/ml) induced differentiation of embryonic stem cells (es) that resulted in the expression of the mature motor neuron transcription factors hb9, chat and islet 1/2. mntf was found to have an affinity for the insulin receptor (ir) and the insulin-like growth factor (igf) of es cells that resulted in the autophosphorylation of tyr972 and tyr 1162/1163 markers of ir activation receptor (igf-1r) based on embryonic stem cell differentiation studies [10]. in this paper, the in vitro and in vivo properties of the mntf-33 derived gm6, a six-amino acid part of the endogenous motor neuron trophic factor is described, summarizing its potential as a new peptide therapeutic for als. peptide therapeutics have several advantages being highly selective and efficacious and are well tolerated with good safety profiles [11]. the challenges with peptide therapeutics are their impaired cell permeability and their stability in plasma that has to be improved [11]. therefore, the biomimetic properties of gm6 have been measured by hplc using chemically bonded protein and phospholipid stationary phases [12]. the models developed for small molecules [13] have been applied to estimate the in vivo pk profile of gm6 with special emphasis on the expected free tissue concentration. gm6 has been found to have good drug-like properties with an estimated good cellular accumulation and tissue partition as has been described previously [14]. discovery of gm6 genervon biopharmaceuticals llc has discovered and patented [9] several peptides. the hypothesis is that the master regulator for the differentiation of the human nervous system must be endogenous and the 3d structure should not be modified in order to be able to bind to a specific receptor. the natural motoneuron trophic factor (mntf) consists of 33 amino acids. it's smaller analogue gm6 which is a partial sequence of mntf and is a 6 amino acid cationic peptide. this potential drug modelled on an endogenous multi-target embryonic stage protein isolated from the nervous system. it was synthesized and then tested in the sod1/g93a transgenic mouse model for als, gm6 delayed disease onset (27 %) and significantly extended the survival (30 %) and improved motor function (4143 %) [15]. experimental physicochemical and biomimetic property measurements – estimating in vivo distribution of gm6 all high-performance liquid chromatography (hplc) measurement were performed on an agilent 100 equipped with an ultraviolet diode array detector (uv-dad). the reversed phase fast gradient retention times have been measured using a gemini nt c-18 column with the dimensions of 50 x 3 mm, particle size 5 µm with a 110 å pore size (phenomenex uk). the starting mobile phase was ph 2.6 0.01 m formic acid for low ph lipophilicity, 50 mm ammonium acetate buffers with ph adjusted to 7.4 and 10.5 by concentrated ammonia solution for neutral and basic ph lipophilicity measurements respectively. all reagents were hplc grade obtained from honeywell/sigma-aldrich, germany. a 1.00 ml/min flow rate was applied. the acetonitrile gradient was admet & dmpk 6(2) (2018) 176-189 in vitro and in vivo characterization of gm6 doi: 10.5599/admet.547 179 from 0 to 100 % from 0 to 3.5 min then 100 % acetonitrile was kept until 4.5 min and then returned to 0 % at 4.7 min. the run time was 6 min. hplc grade water and acetonitrile were obtained from rathburn chemicals ltd, walkerburn, uk. the gradient retention times were standardized using the valko test mixture and the chromatographic hydrophobicity index (chi) values as described previously [13,16]. the chi values were converted to the octanol/water log d scale using equation 1. chi log d = 0.0525 chi – 1.467 (1) in this way, the peptides lipophilicity can be compared to their octanol/water lipophilicity. measurements of membrane binding using immobilized artificial membrane (iam) chromatography for the measurements of gm6 interactions with phospholipids, the gradient retention times have been measured using an iam.pc.dd2 100x4.6 mm column with 10 µm diameter and 300 å pore size particles. the starting mobile phase was 50 mm ammonium acetate with the ph adjusted to 7.4. the acetonitrile gradient was applied from 0 to 90 % from 0 to 4.75 min and kept at 90 % until 5.25 min. from 5.25 to 5.5 min the acetonitrile concentration was dropped to 0 %. the mobile phase flow rate was 1.5 ml/min and the run time was 6 min. the retention times were standardized using the iam calibration mixture as described previously [17] (equation 2). log kiam = 0.29*exp(0.046 chi iam+0.42) + 0.70 (2) the log kiam values express the membrane partition comparable to the octanol/water lipophilicity. protein binding measurements using biomimetic protein stationary phases the peptides interactions with human serum albumin (hsa) and alpha-1-acid-glycoprotein (agp) have been measured using commercially available chemically bonded hsa (chiralpak-hsa) and agp (chiralpakagp) hplc columns with the dimensions of 50 x 3 mm with a 5 µm particle size obtained from hichrom ltd, reading, uk. the mobile phase was 50 mm ammonium acetate with the ph adjusted to 7.4. an isopropanol gradient (hplc grade, rathburn chemicals, walkerburn, uk) was used from 0 to 35 % from 0 to 3.5 min with a flow rate of 1.5 ml/min. the 35 % iso-propanol mobile phase was run till 4.5 min and then back to 0 % within 0.2 min. the run time was 6 min to allow re-equilibrate of the protein phase with the buffer. the retention times were standardized using the calibration set of compounds as described previously. using the slope and intercept values from the calibration line the logarithmic retention times were converted to log k values that can be converted to % binding values (% hsa and % agp) using equation 3 [18]. log log 101 10 % bound 1 10 k k    (3) in order to be able to estimate the in vivo distribution, cell penetration, cell membrane affinity and plasma protein binding biomimetic hplc characterization of gm6 were carried out using the methods described earlier [13,14,19]. the in vivo distribution behaviour of gm6 has been estimated on c-18, iam, hsa and agp stationary phases based on the measured generic gradient retention times described above. the model equations with the source references are listed in table 1. stability measurements in human plasma the plasma stability of gm6 has been investigated by using a standard hplc/ms/ms procedure by frontage laboratory inc (malvern, pa, usa). this bioanalytical laboratory measured the concentration of klara valko et al. admet & dmpk 6(2) (2018) 176-189 180 gm6 in human plasma using k2edta as the anticoagulant. the system was calibrated using gm6 in the concentration range of 250 ng/ml to 50 mg/ml. the mean gm6 concentration was measured at 10, 15, 30 and 60 min time points after incubation. a considerable difference was observed in the mean concentration of the peptide gm6 within each individual lot of whole human blood at times 10, 15, 30 and 60 min compared with the 1 minute time point, which indicated that in the absence of stabilizer, gm6 is not stable in fresh human blood. it is worth mentioning that in this study only the intact gm6 molecule concentration was monitored. table 1. the model equations used to derive estimated in vivo distribution (log vdss = logarithm of the volume of distribution, log vdu= logarithm of the unbound volume of distribution), drug efficiency (demax), brain tissue binding (log kbtb), fraction unbound in brain and plasma (fubtb and fuppb), brain to plasma ratio (kbb) and cell partition log kpcell). log kiam [17] = 0.29*e (0.026chi(iam)+0.42 ) + 0.7 log kiam [17] =0.046*chi(iam) + 0.42 log khsa [17] = e log k(hsa) log khsa [18] = log (%hsabound/(101%hsa bound)) estimated log vdss [17] = 0.44*log kiam -0.22*log khsa 0.62 estimated log vdu [20] = 0.23*log khsa +0.43*log kiam 0.72 demax [21] = 100/vdu log kbtb [22] =1.29*log kiam+1.03*log khsa 2.37 log kppb [22] =0.98∗log 𝑘𝐻𝑆𝐴+0.19∗log 𝑘𝐴𝐺𝑃+0.031∗chi log d7.4−0.20 %btb [22] =100*10 log kbtb /(1+10 log kbtb) %ppb [22] =100*10 log kppb/(1+10 log kppb) fu btb and ppb [22] =(100-%btb)/100 and (100-%ppb)/100 kbb [22] = fuppb/fubtb log kpcell [23] =1.1*log kiam 1.9 in vitro clearance and estimated pk profile of gm6 in order to estimate the expected pk profile of gm6 in human the in vitro microsomal stability has been measured. eurofins pharma discovery services cerep inc (france) carried out in vitro clearance measurements on gm6 using human liver microsomes. the in vitro clearance data so obtained can be used to estimate the in vivo clearance and half-life using a well-stirred model [24]. a comparative study was carried out using gm6 and riluzole, an fda approved drug for als. the purpose of the study was to test whether there was any interaction between gm6 and riluzole in the intrinsic clearance assay using human liver microsomes. a pool of 50 mixed gender human liver microsomes was used with a final microsomal concentration 0.1 mg/ml. the test compounds were dissolved to a 0.1 µm concentration in 0.01% dmso, 0.25 % acetonitrile and 0.25 % methanol. the test compounds were pre-incubated with the pooled liver microsomes in phosphate buffer and shaken for 5 min at 37 centigrade. the reaction was initiated by an nadphgenerating system and incubated for 0, 15, 30, 45 and 60 min. the reaction was then stopped by transferring the incubation mixture to an acetonitrile/methanol mixture. samples were then mixed and centrifuged and the supernatant analysed by hplc-ms/ms. the compound remaining after incubation was admet & dmpk 6(2) (2018) 176-189 in vitro and in vivo characterization of gm6 doi: 10.5599/admet.547 181 calculated from the peak area obtained for each sample time point relative to the zero-time point. the half-life was calculated from the slope of the initial linear range of the logarithmic curve of the compound remaining intact (%) versus time, assuming first-order kinetics. the intrinsic clearance was calculated using equation (4).  int 1/2 0.693 μl/min/mg protein protein conc. cl t   (4) quantitative immunofluorescence imaging phenovista biosciences llc (san diego, ca, usa) provided several phenotypic assays on four different cell types treated with gm6 in order to reveal the presence of, and reaction to, gm6 in these cells. the primary goal of the study was to identify antibody-based markers that could be used successfully in quantitative imaging-based analysis and to determine if the application of gm6 caused quantifiable effects on signalling pathways identified by a series of antibody markers. four human cell types were used: 1. ips-derived icell hepatocytes 2.0 (cdi phc-100-020-001) (cultured 10 days in manufacturers suggested media then treated with gm6), 2. human brain microvascular endothelial cells [hbmvecs] (ix cells 10hu-051) (cultured 2 days in manufacturers suggested media then treated with gm6), 3. ntera2 differentiated cells (atcc #crl-1973) (differentiated for 21 days in 10 µm retinoic acid (ra), 0.5 mm ibmx and 1 mm 8-br-camp then treated with gm6), 4. ips-derived icell gabaergic neurons (cdi nrc-100-010-001) (cultured 7 days in manufacturers suggested media then treated with gm6). all cells were cultured and treated in high density 384 well microplates before fixation, immunofluorescence staining and analysis using automated quantitative microscopy. the aim of the analysis was to determine uptake and downstream signalling effects of gm6 via assessment of the marker levels in each cell type under each treatment condition. due to the unknown distribution and response of the twelve unique markers selected, a general-purpose image analysis strategy was taken whereby changes in cellular intensity for each marker would be quantified. cells were seeded in 384 well plates (greiner) in culture media according to manufacturer’s guidelines and where applicable differentiated before treatment with gm6 for 15 minutes, 2 or 4 hours. treatments were performed via the addition of a 10x solution as 10 % of the final culture volume (e.g. 2.2 µl of 10 mm gm6 added to 20 µl culture volume for a final concentration of 1 mm gm6). after treatment the cells were fixed in 4 % formaldehyde for 15 minutes, washed in pbs containing 0.3 % triton x-100 for 5 minutes at room temperature then blocked in pbs containing either 5 % goat serum or 5 % donkey serum (depending on primary antibody host used), 1 % bsa, and 0.2 % fish gelatin overnight at 4 °c. primary antibody diluted in blocking buffer was incubated at 4 °c overnight followed by three washes in pbs then fluorescentlyconjugated secondary antibodies (1:500 dilution, and hoechst 33342 (10 µg/ml final) added in pbs and incubated at room temperature for 1 hour before washing in pbs three times and storage at 4 °c. twelve unique markers were assessed in combination with the nuclear marker hoechst 33342 in six staining sets (aka palettes); chicken anti-gm6 (genervon), rabbit anti-insulin receptor (abcam #ab131238), goat anti-sod1 (abcam #ab62800), mouse anti-insulin receptor beta (abcam #ab983), rabbit anti-phospho tau s356 (abcam #ab75603), mouse anti-pi3kp85 (abcam #ab86714), rabbit anti-tdp43 (thermoscientific #pa5-29949), rabbit anti igf1 receptor beta (cell signalling #9750), mouse anti-tau (synaptic systems klara valko et al. admet & dmpk 6(2) (2018) 176-189 182 #314011), rabbit anti-phospho tau t231 (abcam #ab151559), mouse anti-igf1 receptor (abcam ab16890) and rabbit anti-phospho insulin recptor beta y972 (abcam 5678). imaging was performed using a thermoscientific cx7 automated fluorescence microscope. images were collected using filter sets appropriate for the fluorescently-conjugated secondary antibodies and images analysed using the compartmental analysis (thermoscientific) algorithm as part of the hcs studio software package. data were processed using microsoft excel to calculate mean, standard deviation and strictly standardized mean difference (ssmd) scores to assess statistical significance versus control treatments. in vivo brain penetration studies c57bl6 mice were injected with a single bolus iv tail vein injection of gm6 at 0.2 and 2.0 mg/kg. at four hours, the animals were sacrificed, and half of the brain was frozen for elisa analysis. the detailed experimental process has been published previously [25]. results and discussion physicochemical and biomimetic properties of gm6 it is important to determine how gm6 behaves in vivo and what are its physicochemical and drug-like properties and can it’s in vivo distribution and pharmacokinetic behaviour be predicted. similar to other peptide therapeutics it was expected that gm6 would not be stable in plasma, so measurement of the traditionally used plasma vs. time pk profile would not be feasible. therefore, the protein and phospholipid binding of gm6 was assessed using biomimetic hplc experiments. the results of the measurements of protein and phospholipid binding of gm6 are listed in table 2, including the estimated in vivo distribution properties of gm6 using the model equations listed in table 1. it can be seen that due to the two arginine amino acids in the sequence gm6 is positively charged at physiological ph. therefore, it's binding to alpha-1-acid glycoprotein (agp) and immobilized artificial membrane (iam) is very strong. the total plasma protein binding is expected to be 31%. however, due to the strong iam binding, it is very likely that gm6 also partitions quickly into the blood cells. the volume of distribution was estimated from the difference of binding of gm6 to albumin and iam according to the model published for marketed drugs [17] that was 7.8 l/kg using the equation listed in table 1. the brain tissue binding of gm6 is estimated to be 58.3% which is higher than the plasma protein binding suggesting higher total brain concentration than total plasma concentration. the brain to plasma total concentration ratio is expected to be 1.65. due to the strong iam binding, the estimated intracellular concentration is approximately 5 times higher than the extracellular concentration using the models obtained with hela cell [23]. the drug efficiency which is defined as the free bio-phase concentration relative to the dose [26] is over 8% which is much above the expected 1% that is characteristic of the marketed drugs. from the biomimetic models only, the total brain to plasma concentration ratio can be estimated based on the difference between the brain tissue and the plasma protein binding as shown in table 1. figures 1 and 2 show that gm6 has good drug-like properties similar to marketed small molecule drugs. gm6 is well positioned in the middle of the marketed drugs based on the plot of the estimated brain to blood ratio (log kbb) and steady-state volume of distribution (log vdss). it was found that 70 % of gm6 was stable in human liver microsomes after 60 min. the addition of admet & dmpk 6(2) (2018) 176-189 in vitro and in vivo characterization of gm6 doi: 10.5599/admet.547 183 riluzole to the liver microsomes did not change the clearance of gm6 nor the riluzole clearance, suggesting no drug-drug interactions. further studies will be needed to estimate the renal clearance of peptides which is a more typical route of elimination [27]. table 2. the measured biomimetic properties of gm6 using published models for small molecule marketed drugs [22] property value explanation % bound hsa 22.90 albumin binding % bound agp 45.06 glycoprotein binding calc % ppb 31.40 plasma protein binding calc % blood 23.60 blood binding chi log d2 -3.38 lipophilicity at ph 2 chi log d7.4 -0.16 lipophilicity at ph 7.4 chi log d10.5 0.20 lipophilicity at ph 10.6 chi log p 0.20 neutral form lipophilicity iam 42.5 phospholipid binding cad-likeness 49.4 phospholipidotic potential est log bb 0.22 brain to blood ratio (log) brain to plasma 1.65 brain to plasma ratio log vd 0.89 log volume of distribution vd l/kg 7.83 volume of distribution log vdu 1.06 log unbound volume of distribution vdu 11.5 the unbound volume of distribution demax% 8.74 max drug efficiency mw 799.00 molecular weight log kbrain 0.15 brain tissue partition % btb 58.3 brain tissue binding log k lung tissue 0.51 lung tissue partition % lung tissue binding 72.3 lung tissue binding log kp skin -7.10 log skin partition log kp skin -7.69 skin partition log kp cell 0.71 log cell partition kp cell 5.2 cell partition it is well-known that plasma proteases can hydrolyse peptides rapidly and that special method should be applied to the plasma concentration determination [28,29]. gm6 showed 15 min half-life in vitro in human blood. it has been also reported that some therapeutic peptides could not be detected in the plasma but they still had a long-lasting pharmacological effect [30]. there are various methods for increasing the plasma half-life of peptides such as replacing l-amino acids to damino acids, blocking the n or c terminals, conjugations to macromolecules like peg and cyclisation [27]. other approaches such as increasing the albumin binding are also used to improve the physicochemical and in vivo admet properties of peptides [31] in order to open up new modalities in drug discovery. with gm6, the idea was not to change the peptide structure but to remain close to the endogenous mntf using the 6-amino acid sequence which was thought to be responsible for the neuroprotective effect. klara valko et al. admet & dmpk 6(2) (2018) 176-189 184 figure 1. boxplot of the average estimated log kbb values of marketed drugs, drugs that enter into the central nervous system (cns) and inhaled drugs, as well as gm6. the estimated brain to plasma total concentration ratio was calculated using the equations given in table 1. figure 2. the plot of the estimated brain to blood ratio (log kbb) and the steady-state volume of distribution (log vdss) of marketed drugs that enter into the cns. based on the measured physicochemical and biomimetic properties listed in table 2, it can be seen that gm6 is predicted to have high cell partition (kpcell = 5.2) which means 5 times higher concentration inside the cells that outside the cells. the model has been developed using 300 compounds data that were equilibrated with hela cells and measured the intracellular concentration [23]. hela cells do not contain active transporters, so it can be assumed that this is the passive partition of gm6 into cells. quantitative immunofluorescence imaging the results of the quantitative imaging experiments gave a very good indication that gm6 associates with cells well, especially to ips-derived neurons. gm6 demonstrated quantifiable labelling of cellular structures in all of the cell types tested with chicken anti-gm6 (genervon). while it is difficult to determine whether gm6 was restricted to the surface of the cells that it interacted with or was internalized or degraded, based on the time and dose-dependent effects on markers, it did suggest that gm6 was able to elicit cellular responses that could be a result of cell surface binding, via uptake or via a combination of admet & dmpk 6(2) (2018) 176-189 in vitro and in vivo characterization of gm6 doi: 10.5599/admet.547 185 both processes. it was found that an increase in gm6 abundance in all cell types could be quantified. the response of gm6 treatment on the gabaergic neurons was much more significant than on the hepatocytes, which suggested a specific affinity of gm6 to tissues of the nervous system and less so to the liver. based on the study we can hypothesize that this cationic peptide, gm6, crosses the cell membrane and enters the cytoplasm. figure 3 shows representative immunofluorescence images obtained in gaba neuron cells after treatment of gm6. figure 3. immunofluorescence images obtained after treating the gaba neuron cells with 1 mm gm6 or pbsonly control. (top row shows staining of human ips-derived gabaergic neurons treated with 1 mm gm6 for 15 minutes while the bottom row shows control pbs-treated neurons. leftmost panels show hoechst staining that identifies neuronal nuclei. panels in the middle column show detection of gm6 in the same cells using an anti-gm6 antibody while columns to the right show corresponding insulin receptor staining intensity, which has been included to show the location of the cell bodies and neurites. red arrows show gm6 in the cell body while yellow arrows show gm6 puncta localized to neurites. in some cases, staining of gm6 and insr is coincident.) results of the in vivo brain penetration studies in mice elisa assays with brain homogenate supernatant detected gm6 at statistically significant levels, at all doses, compared to the control (p=0.0001) as shown in figure 4. the vehicle group has some detection of gm6 (0.4050 µm). the 0.2 mg/kg group has 1.76 µm, the 2.0 mg/kg group has 12.92 µm. if we subtract the gm6 in the vehicle group, the net increase from injection of gm6 for the 0.2 mg/kg is 1.36 µm, for the 2.0 mg/kg is 12.52 µm. the net increase of gm6 detected is almost proportional. these results support the prediction that gm6 can partition into tissues based on its large volume of distribution and that has also a good brain penetration [25] (brain to plasma ratio was estimated as 1.65 as indicated in table 2). estimated pk profile of gm6 in order to estimate the dose we have calculated the minimum effective concentration of gm6 obtained from the in vivo mouse experiment described in more detail later. the in vivo results showed that klara valko et al. admet & dmpk 6(2) (2018) 176-189 186 a 20 mg/kg dose of gm6 was efficacious in the mouse. using the 12.3 factor from mouse to human according to the fda guidelines [32], the dose was estimated to be 1.6 mg/kg in human. as the drug efficiency of the compound is around 8 % it means that only 8 % of the administered dose is available as free. that is equivalent to 0.08·1.6 mg/kg, which is 0.13 mg/kg which equals 1.3·10 -4 g/kg. taking into account the 799 da molecular weight of gm6 it is equivalent to a 1.6·10 -7 m concentration. it means that the pic50 or in other words the minimum effective concentration (mec) of the peptide is estimated to be around 1.6·10 -7 m. the dose can be estimated from the pic50 and the drug efficiency as it has been described earlier [21]. this dose estimation refers to a steady state and does not take into account the elimination rate and dosing frequency. figure 4. detection gm6 in mouse brain after four hours of the treatment by elisa analysis. using the one-compartment model [33] and a 132.5 mg dose for an average 70 kg human, a hypothetical pk profile can be calculated using the parameters listed in table 3 based on equation (5).    24 mec exp 1 dose (mg/kg/day) dss elv k f      (5) where τ is the dosing interval, kel is the elimination rate calculated from the in vitro clearance (equation (4)), vdss is the steady-state volume of distribution, f is the bioavailability considered 1 and the absorption rate considered fast after iv administration. the intrinsic clearance (clint) for gm6 in human microsome assays was determined to be < 115.5 µg/ml/min/mg. to convert the clint data from 115 ug/min/mg human liver microsomes to ml/min/kg we used the human scaling factor of mg protein in liver: 115 x 39.7 = 4565.5 µg/min/g liver which is equivalent to 4565/1000 = 4.565 ml/min/g liver. this value should go to the well-stirred model (wsm) to estimate the human whole-body clearance (kel). using the wsm model it is equivalent to 15.5 ml/min/kg (which is equivalent to 86 % of the liver blood flow) as the clearance of gm6 which is relatively rapid. it is supposed that gm6 being an endogenous neuropeptide with two positive charges and relatively small (mw=799) the liver microsomes provide the major elimination route. it is also important to note that gm6 has a large estimated volume of distribution (over 7.8 l/kg) unlike other peptides and it binds to phospholipid very strongly (chi iam value is above 42). in a recent review [34] about the pharmacokinetic studies of protein, drugs raise the question whether the blood concentration is relevant or not to pharmacodynamics with the emphasis on the tissue concentration. admet & dmpk 6(2) (2018) 176-189 in vitro and in vivo characterization of gm6 doi: 10.5599/admet.547 187 table 3. the parameters used in estimating the pk profile of gm6 in human after 132 mg iv dose. body weight (kg) 70 dosing interval (τ) 1 dose (mg/day) 132.5 dose (mg/kg) 1.89 vd (l/kg) 7.8 cl (ml/min/kg) 15.5 cl (l/h/kg)) 0.93 cl (ml/min) 1085.0 cl/f (ml/min) 1085.0 cl/f (l/min) 1.085 fu blood (%) 68.60 tmax (h) 0.8 cmax (ng/ml) 222 ke (1/h) 0.12 t1/2 (h) 5.8 auc (ng/ml*h) 2037 vdu (l/kg) 11.4 drugeff max (%) 8.8 drugeff blood cmax u (%) 8.03 compound information gm6 molecular weight 799 pxc50 (based on in vivo efficacious concentration) 7 conc (nm) 100.00 total conc (nm) 145.77 conc (ng/ml) 79.90 total conc (ng/ml) 116.47 equation 5 describes the hypothetical plasma concentration vs time profile of drugs. as gm6 has a short half-life in plasma but it was detected in brain tissue after 4 hours, instead of the total plasma concentration vs time profile the free concentration-time profile was calculated that is referred to tissues. the estimated half-life in tissues including only the microsomal clearance was calculated as 5.8 h. this pk profile would have been very difficult to measure in vivo due to the technical difficulties of detecting gm6 in human blood. at this stage, the renal clearance of the compound had not been tested. even if the elimination rate is double (30 ml/min/kg) due to renal clearance of gm6 we can expect 3 hours half-life in the tissues when calculated using equation 5. conclusions in this paper, results have been shown and discussed for a 6-amino acid fragment (gm6) of the endogenous motor neuron trophic factor (mntf) containing 33 amino acids and further discussed as a potential peptide therapeutic for amyotrophic lateral sclerosis (als). the physicochemical properties of gm6 were measured using biomimetic hplc membrane and protein binding. gm6 showed very good druglike properties. due to the two positively charged arginine amino acids and based on the measured phospholipid and protein binding, gm6 has been predicted to have very strong tissue partition with an estimated volume of distribution of 7.8 l/kg. it was also predicted that it partitions into cells (kpcell = 5.2) and has a good brain to plasma ratio (1.65). studies in mice show penetration of the gm6 into the brain following a bolus iv injection with the drug still present after four hours. the concentration of gm6 klara valko et al. admet & dmpk 6(2) (2018) 176-189 188 decreased rapidly in blood suggesting a high affinity for target sites in the brain and/or active uptake into brain cells and gaba-ergic neuron cells as shown by immunofluorescence imaging studies. the estimated free tissue concentration – time profile based on only microsomal elimination showed approximately 6 hours half-life. the in vivo beneficial effects of gm6 are longer lasting. this would indicate that it probably provokes a cascade of biochemical events, after initial binding, and that influences multiple pathways that promote neuron regeneration. references [1] s. przedborski, h. mitsumoto, l.p. rowland. recent advances in amyotrophic lateral sclerosis research. curr. neurol. neurosci. rep. 3 (2003) 70–77. [2] m.c. kiernan, s. vučić, b.s. cheah, m.r. turner, a. eisen, o. hardiman, j.r. burrel, m.c. zoing. amyotrophic lateral sclerosis. lancet (london, england) 377 (2011) 942–955. [3] p. pasinelli, r.h. brown. molecular biology of amyotrophic lateral sclerosis: insights from genetics. nat. rev. neurosci. 7 (2006) 710–723. [4] r.e. ellis, j.y. yuan, h.r. horvitz. mechanisms and functions of cell death. annu. rev. cell biol. 7 (1991) 663–698. [5] e.t. reyes, o.h. perurena, b.w. festoff, r. jorgensen, w.v moore. insulin resistance in amyotrophic lateral sclerosis. j. neurol. sci. 63 (1984) 317–324. [6] s. cohen. purification of a nerve-growth promoting protein from the mouse salivary gland and its neuro-cytotxic antiserum. proc. natl. acad. sci. u. s. a. 46 (1960) 302–311. [7] i.a. hendry, k. stöckel, h. thoenen, l.l. iversen. the retrograde axonal transport of nerve growth factor. brain res. 68 (1974) 103–121. [8] r.m. wah. polynucleotides encoding motoneurontrophic factors. us patent 6309877. [9] r.m.w. chau, p.-y.d.ko. hong kong. mntf peptides and compositions and methods of use. us patent 7183373, (2007). [10] d.m. deshpande, d. a. kerr, m. ruxton, d.p.-y. ko. nmntf differentiation and growth of stem cellso title. us patent no 8986676, (2015). [11] k. fosgerau, t. hoffmann. peptide therapeutics: current status and future directions. drug discov. today 20 (2015) 122–128. [12] k.l. valkó. lipophilicity and biomimetic properties measured by hplc to support drug discovery. j. pharm. biomed. anal. 130 (2016) 35–54. [13] k.l. valko, s.p. teague, c. pidgeon. in vitro membrane binding and protein binding (iam mb/pb technology) to estimate in vivo distribution: applications in early drug discovery. admet dmpk 5 (2017) 14-38. [14] k. valko, g. ivanova-berndt, p. beswick, m. kindey, d. ko. application of biomimetic hplc to estimate lipophilicity, protein and phospholipid binding of potential peptide therapeutics. admet dmpk (2018) http://dx.doi.org/10.5599/admet.544. [15] m.k. d. ko. methods of treating neuronal disorders using mntf peptides and analogs. us patent 8673852, (2014). [16] k. valko, c. du, c. bevan, d. reynolds, m. abraham. rapid method for the estimation of octanol / water partition coefficient (log poct) from gradient rp-hplc retention and a hydrogen bond acidity term (sigma alpha2h). curr. med. chem. 8 (2001) 1137–1146. [17] f. hollósy, k. valkó, a. hersey, s. nunhuck, g. kéri, c. bevan. estimation of volume of distribution in humans from high throughput hplc-based measurements of human serum albumin binding and immobilized artificial membrane partitioning. j. med. chem. 49 (2006) 6958–6971. [18] k. valko, s. nunhuck, c. bevan, m.h. abraham, d.p. reynolds. fast gradient hplc method to determine compounds binding to human serum albumin. relationships with octanol/water and immobilized artificial membrane lipophilicity. j. pharm. sci. 92 (2003) 2236-2248. admet & dmpk 6(2) (2018) 176-189 in vitro and in vivo characterization of gm6 doi: 10.5599/admet.547 189 [19] k. valko. lipophilicity and biomimetic properties measured by hplc to support drug discovery. j. pharm. biomed. anal. 130 (2016) 35–54. [20] k.l. valkó, s.b. nunhuck, a.p. hill. estimating unbound volume of distribution and tissue binding by in vitro hplc-based human serum albumin and immobilized artificial artificial membrane-binding measurements. j. pharm. sci. 100 (2011) 849–862. [21] k. valko, e. chiarparin, s. nunhuck, d. montanari. in vitro measurement of drug efficiency index to aid early lead optimization. j. pharm. sci. 101 (2012) 4155-4169. [22] k. valko, physicochemical and biomimetic properties in drug discovery chromatographic techniques for lead optimization. wiley, hoboken nj (2014). [23] l.j. gordon, m. allen, p. artursson, m.m. hann, b.j. leavens, a. mateus, s. readshaw, k. valko, g.j. wayne, a. west. direct measurement of intracellular compound concentration by rapidfire mass spectrometry offers insights into cell permeability. j. biomol. screen. 21 (2016) 156-164. [24] a.b. ahmad, p.n. bennett, m. rowland. models of hepatic drug clearance: discrimination between the ‘well stirred’ and ‘parallel-tube’ models. j. pharm. pharmacol. 35 (1983) 219-224. [25] j. yu, h. zhu, d. ko, m.s. kindy. motoneuronotrophic factor analog gm6 reduces infarct volume and behavioral deficits following transient ischemia in the mouse. brain res. 1238 (2008) 143–153. [26] s. braggio, d. montanari, t. rossi, e. ratti. drug efficiency: a new concept to guide lead optimization programs towards the selection of better clinical candidates. expert opin. drug discov. 5 (2010) 609– 618. [27] l. di. strategic approaches to optimizing peptide adme properties. aaps j. 17 (2015) 134-143. [28] w. li, j. zhang, f.l. tse. strategies in quantitative lc‐ms/ms analysis of unstable small molecules in biological matrices. biomed. chromatogr. 25 (2011) 258–277. [29] l.t. nguyen, j.k. chau, n.a. perry, l. de boer, s.a. j. zaat, h.j. vogel. serum stabilities of short tryptophanand arginine-rich antimicrobial peptide analogs. plos one 5 (2010) 1–8. [30] h. hui, l. farilla, p. merkel, r. perfetti. the short half-life of glucagon-like peptide-1 in plasma does not reflect its long lasting beneficial effects. eur. j. endocrinol. 146 (2002) 863–869. [31] n. tsomaia. peptide therapeutics: targeting the undruggable space. eur. j. med. chem. 94 (2015) 459–470. [32] center for drug evaluation and research. guidance for industry: estimating the maximum safe starting dose in initial clinical trials for therapeutics in adult healthy volunteers. us dep. heal. hum. serv. (2005) 1–27. [33] d.f. mcginnity, j. collington, r.p. austin, r.j. riley. evaluation of human pharmacokinetics, therapeutic dose and exposure predictions using marketed oral drugs. curr. drug metabol. 8 (2007) 463-479. [34] e. ezan. pharmacokinetic studies of protein drugs: past, present and future. adv. drug deliv. rev. 65 (2013) 1065–1073. ©2018 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ the role of pharmacology in anticancer drug development doi: 10.5599/admet.6.1.496 4 admet & dmpk 6(1) (2018) 4-14; doi: http://dx.doi.org/10.5599/admet.6.1.496 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index review the role of pharmacology in anticancer drug development godefridus j. peters 1, *, anne-sophie govaerts 2 , hans r hendriks 3 , for the eortc pharmacology and molecular mechanism group 1 dept of medical oncology, vu university medical center, po box 7057, 1007 mb amsterdam, the netherlands 2 eortc, avenue e. mounier 83/11, 1200 brussels, belgium 3 hendriks pharmaceutical consulting, j. wagenaarstraat 67, 1443 lr purmerend, the netherlands *corresponding author: e-mail: gj.peters@vumc.nl received: january 11, 2018; revised: march 07, 2018; published: march 25, 2018 abstract drug development consists of many sequential and parallel steps; failure in one of the steps can lead to discontinuation of the process. the process is time-consuming and very expensive, especially the clinical phase. in order to enhance cancer drug development in the 1980s, the national cancer institute (nci) adopted a new screening system using 60 different tumour cell lines from various histologies. all standard drugs were tested in this panel and it is still open for testing of new chemical entities (n ce) of potential interest. the european nci compounds initiative, a collaborative programme of the nci, the cancer research campaign (crc; now cruk) and the pharmacology and molecular mechanism group (pamm) of the eortc (european organization on research and treatment of cancer), was initiated in 1993. the programme aimed to help the nci reducing its backlog of in vivo testing by further evaluation of interesting european compounds using a pharmacologically directed approach. considerable multidisciplinary expertise in drug development was combined by the crc and eortc-pamm: chemists, pharmacists, biologists, pharmacologists, oncologists. selection criteria for european nci compounds included novelty of the nce, in vitro activity, if available in vivo and hollow fibre activity, and compare negativity. over a period of more than 20 years 95 out of approximately 2,000 reviewed compounds were selected. these compounds were put through a series of stepwise pharmacological tests comprising solubility (suitable formulation to administer the nce to mice), feasibility to develop a simple analytical assay (usually hplc), limited toxicology and angiogenic properties. this paper provides examples to illustrate the rigorousness of the elimination process of the compounds and discusses the way to improve the process by inclusion of more physico-chemical parameters. keywords eortc-pamm; nci; cruk; in vitro screen; pharmacological approach; anticancer drug development; european nci compounds introduction anticancer drug development is a collaborative effort. when a compound has been synthesised chemically or isolated from a specific source, e.g. a natural product, it will be characterised chemically, physically, pharmacologically, pharmaceutically, toxicologically and biologically, requiring the input from various disciplines. when a compound passes all these steps successfully it continues to clinical development [1]. a tightly controlled process precedes the first administration to a patient. this process is http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:gj.peters@vumc.nl admet & dmpk 6(1) (2018) 4-14 the role of pharmacology in anticancer drug development doi: 10.5599/admet.6.1.496 5 subject to many regulations and requires approval by local hospital medical ethical review boards and national competent authorities. this first step is the phase i clinical study, in which the safety, side effects and best dose of a drug are determined. the latter can be the maximum tolerated dose (mtd) for cytotoxics or the maximum effective dose (usually lower than the mtd) for compounds affecting signaling pathways. this dose is used to determine the antitumor activity of the agent in a phase ii study [1-3]. in the classical drug development setting, an active compound (antitumor activity comparable to that of standard therapy), will proceed to a randomised phase iii study in which the new drug will be compared to the standard therapy or placebo. however, with several novel compounds in which a target is defined, a phase i study may be expanded to include only patients with predefined biomarkers [3]. drugs can obtain market approval from medicines agencies such as the ema and fda, based on an extensive risk/benefit ratio evaluation. an increasing understanding of cancer biology leads to a heterogeneity at multiple levels (genetic, functional, tumour components, signalling pathways), which translates into an increasing number of products in the pipeline of pharmaceutical companies. this high number of compounds resulted in more phase i, ii and iii clinical trials requiring even more patients, which led to extensively high costs for pharmaceutical industry and subsequently health care in society [4,5]. these numbers (costs and number of patients) are increasing in the last decades [6], while the rate of success in oncology drug development is limited, compared to other diseases [7]. in order to keep the costs as low as possible it is important that every aspect of the development process is critically reviewed, including optimisation of drug selection at an early stage. the european organization for research and treatment of cancer (eortc) is the largest european organization coordinating clinical studies in the development of new cancer therapies [8]. the national cancer institute (nci) in bethesda, md, usa, developed in the 2nd half of the 1980s a 60-cell line panel (called nci-60 panel throughout this paper), consisting of 60 cell lines representative for the 9 major tumour types, with the aim to select compounds that show selective potency against one of these malignancies [9]. this nci-60 panel was offered to the scientific community to test the potency of novel chemical entities (nces) and since that period hundreds of thousands of compounds have been submitted to the nci screen, but not all passed the initial nci selection procedure and were not tested. several tested compounds showed an interesting profile and were candidates for further development e.g. in vivo testing, but early in this process it was recognized that too many compounds fulfilled the criteria for in vivo testing. in order to speed up the backlog in this development and since many compounds came from europe, an international academic collaboration between both sides of the atlantic was initiated in 1993, between the eortc research branch, the cancer research campaign (crc, now cruk) and the nci, the so-called european nci compounds initiative, which focused on compounds synthesised in europe and showing interesting data in the nci screen. crc and eortc set up a review committee to select european compounds for further development [10,11]. about 2,000 compounds were evaluated by the crc-eortc review committee in >20 years and 95 were selected for further evaluation. further details on this process and the achievements have recently been summarised [10]. the purpose of this paper is to evaluate mechanistic details e.g. adme (absorption, distribution, metabolism, excretion) in this process by describing the development of several compounds that either failed or were successful. successful means moving into phase i or eventually phase ii clinical studies. failure may have been caused by obvious parameters, such as lack of potency, poor solubility, stability and other physico-chemical properties. godefridus peters et al. admet & dmpk 6(1) (2018) 4-14 6 the process of european nci compound evaluation in the early 1990s, the screening’s capacity of the nci-60 panel was 10,000 compounds annually. nces, submitted by scientists from all over the world, are selected for screening based on structural diversity and novelty of the compound. in the initial program selected nces were tested at 5 different concentrations (varying over 5 logs: 10 -4 –10 -8 or 10 -5 –10 -9 m) and cells were exposed for 48 hours to the drug. three parameters were determined for each drug: the gi50 (or ic50: 50 % growth inhibition), the tgi (total growth inhibition) and the lc50 (50 % cell kill or 50 % lethal concentration) (figure 1) [12]. the results of the screen are reported as growth inhibition curves as shown in figure 2 for compound nsc-625492. many current drug sensitivity tests often denote these curves as viability curves. this is not a correct terminology, because growth inhibition is measured. the nci data are summarized as mean graphs for gi50, tgi or lc50 values. these mean graphs depict whether a cell line is more or less sensitive than the gi50/tgi/lc50 mean of the test drug in the nci-60 panel. figure 3 shows the mean graphs of compound nsc-710305 [13], also known as phortress, and a prodrug of phortress 5f-203. these mean graphs demonstrate a tumour specific sensitivity for lung, ovarian, renal and breast cancer (bars to the right). subsequent mechanistic studies were performed according to the guidelines published recently [14]. figure 1. example growth inhibition curves after 48 hours exposure to a hypothetical drug tested at 0.1 to 1000 nm in three lung cancer cell lines. three parameters are shown: the gi50 (the concentration that gives 50 % growth inhibition, often also referred to as ic50), the tgi (the concentration that gives a total growth inhibition meaning that the number of cells at the end of the exposure is similar to that at the beginning) and the lc50 (the concentration that gives 50 % cell kill, meaning that at the end the number of cells is 50 % of the number of cells at the beginning). values are calculated according to the nci protocol: https://dtp.cancer.gov/discovery_development/nci-60/methodology.htm when positive, such as for phortress, compounds were further screened in a human tumour xenograft (htx) panel. however, the nci screen yielded too many potential compounds for htx testing, so that a fast low-cost preliminary in vivo screen was developed, the so-called hollow fibre assay (hfa) [15,16]. this assay combines advantages of in vitro screens (relatively fast, several tumour types together) and bypasses the disadvantages such as the poor predictive capacity for in vivo activity and lack of information on potential pharmacological properties. the latter are important to determine the amount of drug to be delivered [17]. the hfa appeared to predict for htx activity [18,19]. https://dtp.cancer.gov/discovery_development/nci-60/methodology.htm admet & dmpk 6(1) (2018) 4-14 the role of pharmacology in anticancer drug development doi: 10.5599/admet.6.1.496 7 figure 2. example of a typical nci-60 cell line growth inhibition report for compound nsc-625492. each panel represents cell lines derived from of a particular type of cancer (left to right, top to bottom): 5 breast cancer cell lines), 8 central nervous system (cns) cell lines, 9 colon cancer cell lines, 6 leukaemia cell lines, 11 melanoma cell lines, 11 non-small cell lung cancer (nsclc) cell lines, 7 ovarian cancer cell lines (including one with p-glycoprotein overexpression), 2 prostate cancer cell lines, 9 renal cancer cell lines). the two small cell lung cancer cell lines are not shown. the compound was tested at the concentration range of 10 -8 to 10 -4 m and cells were exposed for 48 hr. obtained from the nci dtp database https://dtp.cancer.gov/discovery_development/nci-60/methodology.htm. in addition to the potency of the drugs (e.g. figure 2), the selectivity for a certain tumour type in vitro (mean graph, figure 3), the hfa and if available the htx data, another important selection criterion for the european nci compounds were the compare data [20]. compare is a comparison tool developed by the nci based on the sensitivity profile of an nce compared to the whole database. compare-positive compounds (pearson correlation coefficient ≥0.8) often showed the same mechanism of action; these compounds could be structurally related but also unrelated compounds [9,12]. since compare-negative compounds (pearson correlation coefficient <0.8) usually did not have close matches in the compare programme, they might have a novel mechanism of action and were thought to be most interesting. for administration of drugs to animals often additional hurdles needed to be solved, such as the administration vehicle, which will be different from formulations used in vitro [14]. many compounds were poorly soluble and specific formulations were used, dependent on the compound, but keeping in mind what might be suitable for future administration to patients. hence dimethyl sulfoxide (dmso) might be suitable for in vitro and in vivo to mice, but not for patients (table 1). after this pre-formulation step, the potential bioavailability of a compound was investigated. after the development of a suitable hplc assay (or another analytical assay) and using a proper formulation, the mtd in mice was determined. the mtd was administered and plasma pharmacokinetics determined at various time points to ascertain whether the drug would reach plasma concentrations in the range of the in vitro gi50. subsequently, this dose and schedule were used to explore whether the compound would have suitable in vivo antitumor activity. if so, https://dtp.cancer.gov/discovery_development/nci-60/methodology.htm godefridus peters et al. admet & dmpk 6(1) (2018) 4-14 8 the development was continued and if not, stopped. in the next section, several examples of such a development will be presented. figure 3. example report of drug selectivity for 5f-203 and its prodrug phortress (source nci dtp). values are calculated relative to the mean gi50 for each compound. differences compared to the mean are given as a log 10 value, with bars to left representing less sensitive cell lines (10 3 to 0) and those to the right more sensitive cell lines (0 to 10 -3 ). cell lines from top to bottom represent 5 leukaemia cell lines, 9 nsclc, 7 colon cancer, 6 cns, 8 melanoma, 6 ovarian cancer, 8 renal cancer, 2 prostate cancer, 8 breast cancer (including one with p-glycoprotein overexpression). please note that the composition of this panel is different from figure. 2. the most sensitive cell lines are depicted in a larger font. the non-responsive in italics. examples of selected compounds in one of the early selected series of compounds, nsc-697786 (figure 4) showed an interesting profile in the compare analysis with a compare coefficient of 0.503 for methotrexate, 0.498 for dup785 (brequinar), 0.496 for n-phosphon-acetyl-l-aspartate (pala) and 0.474 for trimetrexate. this indicates some similarity with antifolates (methotrexate and trimetrexate) and pyrimidine de novo nucleotide inhibitors (dup785 and pala). the standard approach to exclude this is a comparison in cell lines with aberrations in folate metabolism and to perform rescue experiments with normal pyrimidine nucleosides, respectively. consequently, this compound was tested in colon cancer widr cells grown in standard culture medium (which has a high folate concentration) and widr cells that are adapted to a low physiological (comparable to human plasma; widr/f) folate concentration [21]. a similar ic50 was found in both lines (0.26 ± 0.06 and 0.25 ± 0.07 µm, respectively; means ± se of 4-7 experiments; 72-hour exposure). widr/f cells are more sensitive to antifolates such as methotrexate. addition of 1 or 5 mm uridine or cytidine did not protect the cells, which would have been indicative for a pyrimidine de novo inhibition [22]. fma3 wild type and fma3/tk(thymidine kinase deficient) [23] were similarly sensitive (ic50 values for nsc-697786 3.2 ± 0.2 and 2.0 ± 0.8 µm), which excluded a role for thymidine kinase. it was concluded that this compound is a not acting as an antimetabolite, and had possibly unique mechanism of action. however, it was discontinued due to lack of in vivo activity, possibly because of poor pharmacokinetics. admet & dmpk 6(1) (2018) 4-14 the role of pharmacology in anticancer drug development doi: 10.5599/admet.6.1.496 9 table 1. acceptable solvents for in vitro drug solutions or in vivo administration of drugs to mice or rats. solvent maximal concentration maximal dose species maximal final concentration in vivo in vitro culture medium dmem, rpmi (1 st choice) water 10% (final) 0.9% saline ph 7.4 0.5 ml mice; rats 2 nd choice saline ph 2.0 saline ph 11 0.5 ml 0.5 ml rats (dx5) rats (dx5) not acceptable not acceptable ethanol 10% 10 ml/kg mice (i.v.) 1% dimethylsulfoxide (dmso) 10% in water 1% in water 70% in saline 10 ml/kg 10 ml/kg 0.2 ml mice (i.v.; p.o.) mice (i.p.) rats (dx5) 0.25% (0.1% preferable) dimethylformamide 0.125 % (upper limit) tween 80 10% in water 25% in saline 10 ml/kg 0.2 ml mice (i.v.) rats (dx5) propylene glycol 50% 25% 10 ml/kg 10 ml/kg mice (i.v., slow) mice (i.p) 50%/25% 0.2 ml rats (dx5) n-methyl glucamine 1% in water 10 ml/kg mice (p.o.) carboxymethyl cellulose 1% in water 10 ml/kg mice (i.p) cremophor el 5% in water 10 ml/kg mice (i.v.) nahco3 5% in water 10 ml/kg mice (i.v.) n,n-dimethylacetamide (dma) 10% in water 10 ml/kg mice (i.p.) dma 10% in arachis oil 10 ml/kg mice (i.p.) guidelines developed by the joint formulation working party of the screening and pharmacology group (spg) which was continued as a pamm sub-committee and renamed into: drug discovery committee (ddc) [11] a completely different compound, nsc-625492 was selected for its unique molecular structure being a bismuth analogue (figure 4). the gi50 was low (0.3 µm) with activity in many tumour types (colon, ovarian, non-small cell lung cancer and leukaemia; figure 2). it was thought that it might form dna adducts, because of its similarity to other heavy metals (e.g. platinum). using this property, an assay using flameless atom absorption spectroscopy (faas) as used for cisplatin [24], was set up to determine its in vivo pharmacokinetics. the compound indeed showed interaction with dna, but was not cross-resistant against cisplatin resistant cell lines and was active in p-glycoprotein expressing cell lines. using a formulation of 10 % dma and arachis oil, the mtd was 30 mg/kg (8 % weight loss), which was used to determine its plasma pharmacokinetics (figure 5) and potential antitumor activity. nsc-625492 did not show a typical bell-shaped plasma profile curve, but the concentrations remained relatively constant over a 24-hr period, possibly indicating a lack of metabolism. although sufficiently high plasma concentration (> gi50) were found, a long retention and no protein binding (in contrast to cisplatin), there was no in vivo activity against godefridus peters et al. admet & dmpk 6(1) (2018) 4-14 10 various tumour models. the development of the compound was therefore discontinued. nsc-637828 also showed an interesting compare negative profile (correlation coefficients < 0.625) with rifamycin and the nitrogen mustard ccnu (lomustine; 1-(2-chloroethyl)-3-cyclohexyl-1-nitrosourea) being the highest. it was insoluble and a suspension could only be obtained in 10 % dma + arachis oil (but not in dmso or ethanol), which showed acceptable toxicity (< 10 %). due to these properties an adequate hplc method could not be established and drug sensitivity studies could not be completed. therefore, the development of this compound was stopped. n n s n n o o ohcl s n s bi 3+ s cl cl s nr nh 2 s nf n h nh 2 (ch 2 ) 4 nh 2 o n n o on n ome meo oo h h 2hcl nsc-697786 chemical formula: c14h9cln4o3s2 exact mass: 379.98 molecular weight: 380.83 nsc-625492 chemical formula: c6h10bicl2ns2 exact mass: 438.94 molecular weight: 440.17 df 203 (nsc 674495): r = h 5f 203 (nsc 703786): r = f chemical formula 5f 203: c 14 h 11f n 2 s exact mass: 258.06 molecular weight: 259.31 phortress (nsc 710305) chemical formula: c 20 h 23 fn 4 os exact mass: 386.16 molecular weight: 386.49 sjg-136 chemical formula: c 31 h 32 n 4 o 6 exact mass: 556.23 molecular weight: 556.61 figure 4. structural formulas and some physico-chemical properties of several of the investigated compounds. two examples of drugs with completely different mechanisms of action are the combretastatin a-4 analogues and nsc 734237, a paullone-like molecule. these two drugs were investigated in a collaborative effort of various pamm laboratories and the nci drug development and target analysis branches. combretastatin a-4 is a poorly soluble natural product and its more soluble metabolite combretastatin a-4 phosphate disrupts endothelial junctions of tumour blood vessels by preventing microtubule polymerisation in endothelial cells of tumour blood vessels; this resulted in tumour necrosis [25,26]. various analogues (nsc736359, nsc736992, and nsc733436) with a high potency (nm-µm range) were further analysed, but moderate antitumor activity was observed [27], possibly because under these conditions the density (size and number) of cd31 + blood vessels was not affected. the paullone-like molecule nsc 734237 is an isoindoloquinoxaline with low to sub µm ic50 values [28] for cns, colon, melanoma and breast cancer lines. however, the compound was quite insoluble (6 % active drug in a liposomal formulation), but showed antitumor activity against a doxorubicin-resistant, oestrogen admet & dmpk 6(1) (2018) 4-14 the role of pharmacology in anticancer drug development doi: 10.5599/admet.6.1.496 11 receptor negative human breast cancer xenograft mt-3/adr with a reduction in both cd31 + microvessel density and ki-67. hence the drug showed antivascular effects and cytotoxic activity, which are most likely due to inhibition of tubulin polymerisation and topoisomerase i, respectively. more water soluble analogues were investigated further, but did not improve the overall result. in the future, such compounds should optimally be investigated more extensively for their oral bioavailability and their formulation adapted. figure 5. concentration time curves of nsc-625492. mice were injected with 30 mg/kg nsc-625492 in 10% dma in arachis oil. for the pharmacokinetics study mice were sacrificed at 0.25, 3, 6 and 24 h and blood was collected by heparinization to prepare plasma. total bismuth was measured by mixing 1 volume of plasma with 5 volumes of triton/antifoam (1/1) and 4 volumes of 0.4 m hcl, which was directly introduced into the aas (varian spectraa 300 zeeman with certified bi-lamp; λ=306.8 nm). calibration lines were obtained with bismuth nitrate from 0.17 to 3 μm. free bismuth (non-protein bound) was measured in the supernatant after deproteinization (1 volume of plasma mixed with 2 volumes of ethanol and put at -20 °c for 2 hr followed by centrifugation for 10 min at 4000 g) however, several compounds that showed an interesting profile in the nci-screen, were also successful at further evaluation, such as drh-147 and phortress. drh-417, a pyrrolobenzodiazepine monomer acts by binding into the minor groove of dna and was very potent in the nci 60-panel (mean ic50 3 nm) and at subsequent testing by investigators of the european nci compounds group [29]. it showed favourable pharmacokinetics (plasma peak levels of 171 nm, far above the gi50 value) and interesting in vivo activity against several htx models (p<0.01). these data led to the development of sjg-136, a pyrrolobenzodiazepine dimer that went into a clinical phase i study and phase ii studies in ovarian cancer and leukaemia [30]. phortress (figure 3) is a water-soluble benzothiazole prodrug of 5f-203, that is active (low nanomolar gi50), and selective against breast, ovarian and lung cancer [31]. 5f-203 binds to cytoplasmic aryl hydrocarbon receptors, forming a complex that is translocated to the cellular nucleus, and subsequently binds to the cyp1a1 promoter [32,33]. this leads to an induction of the expression of cytochrome p450 1a1, responsible for metabolism of 5f-203 to a bio-reactive species that forms dna-adducts. this unique mechanism of action, together with a good pharmacokinetic profile, led to further evaluation in phase i clinical study [34], in which 14 out of 42 patients had stable disease. all together, these data show that the approach can lead to compounds with clinical activity. godefridus peters et al. admet & dmpk 6(1) (2018) 4-14 12 figure 6. flow chart showing the testing procedure and addition of novel parameters as outlined in the text. each step can lead to elimination when not full-filling the criteria. nce, new chemical entity; mtd, maximal tolerated dose; hfa, hollow fiber assay. the first three examples show that despite initial enthusiasm about the drug sensitivity profile and compare data, development had to be discontinued rapidly. unfavourable physico-chemical properties, such as insolubility, made it impossible to perform adequate in vitro and in vivo studies and to set up suitable analytical assays. in addition to this and lack of in vivo activity, other reasons for discontinuation included instability, impurity, unavailability of material and insufficient bioavailability. a careful analysis of the mechanism of action helps to understand the success (drh-147 and phortress) but also enabled a systematic development of novel analogues. this was due to a successful scientific collaboration between different groups that provides a model for future drug development. importantly the groups adhered to standard cellular pharmacology tools for testing [14]. conclusions the results of the pharmacologically directed approach clearly showed that a few simple procedures with small numbers of mice quickly allow discontinuing agents that were highly unlikely to have any antitumor activity. the process is outlined in the flow chart (figure 6). the chart also shows some streamlining to optimize the process and eliminate agents with unsuitable pharmacological parameters in less than three months. current approaches should put much more emphasis on physico-chemical properties such as solubility, pka values, log p and log d values, protein binding and drug distribution. this serves two purposes, whether to discontinue development rapidly or when promising how to design an analogue with better properties. regarding solubility, there are novel possibilities to use insoluble compounds with adequate solvents [35]. using biomimetic approaches one would be able to gain insight in elimination constants, and likelihood of brain uptake [36]. next to that, pharmacological evaluation should include phase i (oxidation) and ii (conjugation) drug metabolism which can currently easily being studied using isolated hepatocytes. inclusion of such novel approaches would more rapidly predict whether a drug will have suitable pharmacological properties to proceed to the next step in development. admet & dmpk 6(1) (2018) 4-14 the role of pharmacology in anticancer drug development doi: 10.5599/admet.6.1.496 13 acknowledgements: the authors thank all members of the european nci compounds group [10,11], consisting of investigators from the former eortc-spg and ptmg groups, the nci, crc and cruk and the current eortc-pamm group and its ddc group. references [1] k. tam. estimating the “first in human” dose – a revisit with particular emphases in oncology drugs. admet & dmpk 1 (2013) 63-75. [2] m.r. grever. accelerating safe drug development: an ideal approach to approval. hematology am soc hematol educ program. 2013 (2013) 24-9. [3] j.m. bullock, a. rahman, q. liu. lessons learned: dose selection of small molecule-targeted oncology drugs. clin cancer res 22 (2016) 2630-8. [4] j.w. scannell, a. blanckley, h. boldon, b. warrington. diagnosing the decline in pharmaceutical r&d efficiency. nat rev drug discov 11 (2012) 191-200. [5] d. buffery. the 2015 oncology drug pipeline: innovation drives the race to cure cancer. am health drug benefits 8 (2015) 216-22. [6] p.b. bach. limits on medicare's ability to control rising spending on cancer drugs. n engl j med. 360 (2009) 626-33. [7] i. kola, j. landis. can the pharmaceutical industry reduce attrition rates? nat rev drug discov. 3 (2004) 711-5. [8] d. lacombe, s. burock, f. meunier. academia-industry partnerships: are we ready for new models of partnership?: the point of view of the eortc, an academic clinical cancer research organisation. eur j cancer 49 (2013) 1-7. [9] m.r. grever, s.a. schepartz, b.a. chabner. the national cancer institute: cancer drug discovery and development program. semin oncol 19 (1992) 622-638. [10] h.r. hendriks, a.s. govaerts, i. fichtner, a.d. westwell, s. hurtles, g.j. peters, for the eortc ddc group. pharmacologically directed strategies in academic anticancer drug discovery based on the european nci compounds initiative. brit j cancer 117 (2017) 195-202. [11] g.j. peters, e. chatelut, a.k. larsen, n. zaffaroni, on behalf of pamm. eortc-related new drug discovery and development activities: role of the pharmacology and molecular mechanisms group. eur j cancer suppl 10 (2012) 128-140. [12] k.d. paull, r.h. shoemaker, l. hodes, a. monks, d.a. scudiero, l. rubinstein, j. plowman, m.r. boyd. display and analysis of patterns of differential activity of drugs against human tumour cell lines: development of mean graph and compare algorithm. j natl cancer inst 81 (1989) 1088-1092. [13] c.o. leong, m. suggitt, d.j. swaine, m.c. bibby, m.f.g. stevens, t.d. bradshaw. in vitro, in vivo, and in silico analyses of the antitumor activity of 2-(4-amino-3-methylpenyl)-5-fluorobenzothiazoles. mol cancer ther 3 (2004) 1565-1575. [14] p. perego, g. hempel, s. linder, t.d. bradshaw, a.k. larsen, g.j. peters, r.m. phillips on behalf of the eortc pamm group. cellular pharmacology studies of anticancer agents: recommendations from the eortc-pamm group. cancer chemother pharmacol 81 (2018) 427-441. [15] m.g. hollingshead, m.c. alley, r.f. camalier, b.j. abbott, j.g. mayo, l. malspeis, m.r. grever. in vivo cultivation of tumor cells in hollow fibers. life sci 57 (1995) 131-141. [16] m.g. hollingshead, j. plowman, m. alley, j. mayo, e. sausville. the hollow fiber assay. in contributions to oncology, vol 54, relevance of tumor models for anticancer drug development, (1999) h.h. fiebig, a.m. burger (eds),karger: basel pp. 109-120. [17] r.m. phillips, m.c. bibby, j.a. double. a critical appraisal of the predictive value of in vitro chemosensitivity testing. j natl cancer inst 82 (2008) 1457-1468. [18] i.v. bijnsdorp, g.j. peters. hollow fiber assay. in encyclopedia of cancer, (2008) m. schwab (ed). springer verlag: berlin, heidelberg, new york pp. 1404-1406. godefridus peters et al. admet & dmpk 6(1) (2018) 4-14 14 [19] j.i. johnson, s. decker, d. zaharevitz, l.v. rubinstein, j.m. venditti, s. schepartz, s. kalyandrug, m. christian, s. arbuck, m. hollingshead, e.a. sausville. relationships between drug activity in nci preclinical in vitro and in vivo models and early clinical trials. br j cancer 84 (2001) 1424-1431. [20] s.l. holbeck. update on nci in vitro drug screen utilities. eur j cancer 40 (2004) 785-793. [21] h.h.j. backus, h.m. pinedo, d. wouters, j.m. padron, n. molders, c.l. van der wilt, c.j. van groeningen, g. jansen, g.j. peters. folate depletion increases sensitivity of solid tumor cell lines to 5fluorouracil and antifolates. int. j. cancer 87 (2000) 771-778. [22] g.j. peters, s.l. sharma, e. laurensse, h.m. pinedo. inhibition of pyrimidine de novo synthesis by dup-785 (nsc 368390). invest. new drugs 5 (1987) 235-244. [23] i.v. bijnsdorp, e.m. comijn, j.m. padron, w.h. gmeiner, g.j. peters. mechanisms of action of fdump[10]; metabolite activation and thymidylate synthase inhibition. oncol. rep. 18 (2007) 287291. [24] c.j.a. van moorsel, h.m. pinedo, g. veerman, a.m. bergman, c.m. kuiper, j.b. vermorken, w.j.f. van der vijgh, g.j. peters. mechanisms of synergism between cisplatin and gemcitabine in ovarian and non-small cell lung cancer cell lines. brit. j. cancer 35 (1999) 808-814. [25] g.g. dark, s.a. hill, v.e. prise, g.m. tozer, g.r. pettit, d.j. chaplin. combretastatin a-4, an agent that displays potent and selective toxicity toward tumor vasculature. cancer res 57 (1997) 1829–1834. [26] g.m. tozer, c. kanthou, c.s. parkins, s.a. hill. the biology of the combretastatins as tumour vascular targeting agents. int j exp pathol 83 (2002) 21–38. [27] k. bonezzi, g. taraboletti, p. borsotti, f. bellina, r. rossi, r. giavazzi. vascular disrupting activity of tubulin-binding 1,5-diaryl-1h-imidazoles. j med chem 52 (2009) 7906–7910. [28] p. diana, a. martorana, p. barraja, a. montalbano, g. dattolo, g. cirrincione, f. dall'acqua, a. salvador, d. vedaldi, g. basso, g. viola. isoindolo[2,1-a]quinoxaline derivatives, novel potent antitumor agents with dual inhibition of tubulin polymerization and topoisomerase i. j med chem 51 (2008) 2387-2399. [29] a.m. burger, p.m. loadman, d.e. thurston, r. schultz, h.h. fiebig, m.c. bibby. preclinical pharmacology of the pyrrolobenzodiazepine (pbd) monomer drh-417 (nsc 709119). j chemother 19 (2007) 66-78. [30] j. mantaj, p.j.m. jackson, k. rahman, d.e. thurston. from anthramycin to pyrrolobenzodiazepines (pdb)containing antibody conjugates (adcs). angew chem int ed engl 56 (2017) 462-488. [31] i. fichtner, a. monks, c. hose, m.f. stevens, t.d. bradshaw. the experimental antitumor agents phortress and doxorubicin are equiactive against human-derived breast carcinoma xenograft models. breast cancer res treat 87 (2004) 97–107. [32] t.d. bradshaw, a.d. westwell. the development of the antitumour benzothiazole prodrug, phortress, as a clinical candidate. curr med chem 11 (2004) 1009-1021. [33] t.d. bradshaw. phortress: the smart antitumour agent which induces its own metabolism. pharm j 284 (2010) 23-24. [34] m. seckl, n. cresti, a. boddy, r. phillips, f. chapman, h. calvert, l. robson, r. plummer. a cancer research uk phase i trial of phortress (novel antitumour benzothiazole) given intravenously in consecutive 21 day cycles with treatment on day 1 of each cycle. in: proc 8 th ncri cancer conference, 4-7 november 2012, liverpool, uk. abstract lb79. [35] a. avdreef. suggested improvements for measurement of equilibrium solubility-ph of ionizable drugs. admet & dmpk 3 (2015) 84-109. [36] k. valko, s.p. teague, c. pidgeon. in vitro membrane binding and protein binding (iam mb/pb technology) to estimate in vivo distribution: applications in early drug discovery. admet & dmpk 5 (2017) 14-38. ©2018 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ development and validation of a novel method for simultaneous quantification of enzalutamide, darolutamide and their active metabolites in mice dried blood spots using lc-ms/ms: application to pharmacokinetic study in mice doi: 10.5599/admet.557 242 admet & dmpk 6(3) (2018) 242-257; doi: http://dx.doi.org/10.5599/admet.557 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper development and validation of a novel method for simultaneous quantification of enzalutamide, darolutamide and their active metabolites in mice dried blood spots using lcms/ms: application to pharmacokinetic study in mice neraj kumar saini, suresh p. sulochana, mohd zainuddin, ramesh mullangi* drug metabolism and pharmacokinetics, jubilant biosys, industrial suburb, yeshwanthpur, bangalore-560 022, india *corresponding author: e-mail: mullangi.ramesh@jubilantinnovation.com; tel.: +91-80-66628339 received: june 08, 2018; revised: july 05, 2018; available online: july 15, 2018 abstract a simple, sensitive and rapid assay method has been developed and validated for the estimation of enzalutamide, n-desmethylenzalutamide (active metabolite of enzalutamide), darolutamide and orm15341 (active metabolite of darolutamide) on mice dried blood spots (dbs) using liquid chromatography coupled to tandem mass spectrometry with electro spray ionization in the positive-ion mode. the method utilizes liquid extraction of enzalutamide, n-desmethylenzalutamide, darolutamide and orm-15341 from 3 mm punched disks from dbs cards (spiked or study samples). the extracted sample was chromatographed using an isocratic mobile phase (0.2 % formic acid : acetonitrile; 30:70, v/v) on an atlantis dc18 column. the total run time was 2.5 min. the ms/ms ion transitions monitored were m/z 465 → m/z 209, m/z 451 → m/z 195, m/z 399 → m/z 178, m/z 397 → m/z 194 and m/z 481 → m/z 453 for enzalutamide, n-desmethylenzalutamide, darolutamide, orm-15341 and the is (apalutamide-d3), respectively. method validation was performed as per regulatory guideline. the assay had a good linearity over the range of 0.93-2000 ng/ml. the intraand inter-batch accuracy and precision (%re & rsd) across quality controls met the acceptance criteria for all the analytes. stability studies showed that all the analytes were stable on dbs cards for one month. this novel method has been applied to analyze the dbs samples of enzalutamide, ndesmethylenzalutamide, darolutamide and orm-15341 obtained from a pharmacokinetic study in mice. keywords enzalutamide; n-desmethylenzalutamide; darolutamide; orm-15341; dried blood spot; lc-ms/ms; method validation; mice; pharmacokinetics introduction despite therapeutic advances, prostate adenocarcinoma is the highly prevalent cancer in men and responsible for 20 % of cancer related deaths in the western world [1]. initial approach to cure the localized prostate cancer is through radiation therapy or surgical castration or treatment with firstgeneration anti-androgens (flutamide, nilutamide, bicalutamide etc.) [2]. however, despite an initial response, most patients develop a most aggressive form of disease called metastatic castration-resistant prostate cancer (mcrpc) that is associated with tumor progression and survival less than 18-24 months [3]. consequently, in recent years, new therapy options for mcrpc with different mechanisms of action have http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:mullangi.ramesh@jubilantinnovation.com admet & dmpk 6(3) (2018) 242-257 lc-ms/ms determination of enzalutamide/darolutamide with metabolites on dbs doi: 10.5599/admet.557 243 become available like targeting androgen receptor signalling such as cyp17a1 inhibitor (abiraterone acetate) [4] and a second-generation androgen receptor antagonist (enzalutamide, apalutamide and darolutamide) [5-7]. treatment with these drugs showed improved overall survival and quality of life in mcrpc patients. enzalutamide (xtandi  ; fig. 1) was recently approved for the treatment of mcrpc [5,8]. enzalutamide is metabolized by cyp3a4 and 2c8 to produce an active circulatory metabolite, n-desmethylenzalutamide (fig. 1), which has similar in vitro potency to enzalutamide [9,10]. darolutamide (odm-201; fig. 1) is a novel, orally active, second-generation non-steroidal anti-androgen. it is an equi-mixture of two pharmacologically active diastereomers namely orm-16497 and orm-16555. darolutamide undergoes extensive phase-i metabolism and produces circulatory metabolite i.e., orm-15341 (fig. 1). orm-16497, orm-16555 and orm-15341 are fully antagonist against bicalutamide, hydroxyflutamide, enzalutamide and arn-509 mutants [11,12]. unlike enzalutamide and apalutamide, penetration of darolutamide and orm-15341 into brain is negligible (in preclinical species) suggests that low risk of causing seizures in patients [7]. in clinic, darolutamide was tolerated up to 1800 mg and no dose-limiting toxic effects were reported. it was reported that orm-15341 plasma concentrations were higher than sum of orm-16497 and orm-16555 plasma concentrations [13]. currently phase-iii clinical trials are being conducted with darolutamide in non-mcrpc patients globally [7]. n n cf 3 nc o s f o r n n cl nc n n h oh n h o enzalutamide, r=nhch3 n-desmethylenzalutamide, r=nh2 darolutamide n n cl nc n n h n h o o n n n cf 3 nc o s f n h cd 3 o orm-15341 apalutamide-d3 (is) figure 1. structural representation of enzalutamide, n-desmethylenzalutamide, darolutamide, orm-15341 and apalutamide-d3 (is) in contemporary literature few lc-ms/ms bioanalytical methods were reported for quantification of enzalutamide alone [14] or along with its metabolites [15,16] and other anti-androgens [17-19] or anticancer drugs [20] in various biological matrices (preclinical species plasma/tissues and human plasma). similarly darolutamide and orm-15341 quantification in mice plasma was reported post-administration of darolutamide [21] and along with other second-generation non-steroidal anti-androgens (enzalutamide, apalutamide) [20] using conventional achiral columns on lc-ms/ms. apart from these achiral methods, two chiral lc-ms/ms methods for quantification darolutamide diastereomers alone [22] or along with its metabolite [23] were reported in recent times. traditional bioanalysis of samples from preclinical species and clinical trials relies on collection of blood or its products (serum and plasma). collection of blood in a preclinical set up especially with rodents, pediatric cases and new-borns is a challenge and not possible to draw blood samples at multiple time r. mullangi et al. admet & dmpk 6(3) (2018) 242-257 244 points. this will force researcher to adopt sparse sampling, which limits the researcher to study the pharmacokinetics from individual animal/neonate. traditional samples are normally shipped on dry ice from one place to other, which incurs shipping and handling costs. over the wet sampling (blood/plasma/serum) techniques, dried blood spots (dbs) method offers several advantages like reduced blood/plasma/serum volume per time point, ease and safety in handling, potential cost saving in storage and shipment at ambient room temperature, less sample processing time and increase in throughput etc. [24]. to the best of our knowledge, simultaneous method for the determination of enzalutamide, ndesmethylenzalutamide, darolutamide and orm-15341 on mice dbs has not been reported in the literature. for this purpose, the present study is aimed to develop and validate a simple and rapid lcms/ms method for simultaneous quantification of enzalutamide, n-desmethylenzalutamide, darolutamide and orm-15341 on mice dbs without compromising on sensitivity. the validated method was successfully used in a mice pharmacokinetic study following intravenous administration of enzalutamide and darolutamide at 1.0 mg/kg dose. materials and methods materials enzalutamide (purity: >98.3 %) and n-desmethylenzalutamide (purity: >99.6 %) were purchased from bioorganics, bangalore, india. darolutamide (purity: 98 %) was procured from angene international limited, china. orm-15341 (purity: 99.4 %) and apalutamide-d3 (purity: 99.5 %; fig. 1) were synthesized by the medicinal chemistry group, jubilant biosys (bangalore, india) using literature information [25,26] and characterized using chromatographic (hplc, lc-ms/ms) and spectral techniques (ir, uv, mass, 1h and 13c-nmr) by the analytical research group, jubilant biosys. hplc-grade acetonitrile, methanol, methyl tert-butyl ether (tbme) and ethyl acetate were purchased from rankem, new delhi, india. whatman fta dmpk card c dbs cards were purchased from ge, bangalore. silica gel sachets and sealable plastic bags for the storage of dbs cards were purchased from local market. instrumentation and chromatographic conditions a shimadzu vp (shimadzu, japan) lc system equipped with degasser (g1379a), quaternary pump (10advp), column oven (cto-10asvp), auto-sampler (sil-htc) along with system controller (scl-10avp) was used to inject 5.0 µl aliquots of the processed samples on an atlantis dc18 column (50 x 4.6 mm, 3 µm; waters, milford, ma, usa), which was kept at ambient temperature (24 ± 1 °c). the isocratic mobile phase, a mixture of 0.2 % formic acid : acetonitrile (30:70, v/v) was filtered through a 0.45 µm membrane filter (xi5522050) (millipore, usa or equivalent) and then degassed ultrasonically for 5 min was delivered at a flow rate of 0.8 ml/min (with splitter 50 %) into the mass spectrometer electro spray ionization chamber. quantitation was achieved by ms/ms detection in positive ion mode for the analytes and the is using sciex (foster city, ca, usa) api-6500 mass spectrometer, equipped with a turboionspray interface at 500 °c and 5500 v ion spray voltage. the common parameters, i.e. curtain gas (cur), nebulizer gas (gs1), auxiliary gas (gs2) and collision gas (cad) were set at 55, 55, 65 and 10 psi, respectively. the compounds parameters, i.e. declustering potential (dp), collision energy (ce), collision exit potential (cxp) and entrance potential (ep) for the analytes and the is are shown in table 1. the dwell time was 100 msec. quantitation of enzalutamide, n-desmethylenzalutamide, darolutamide, orm-15341 and the is was achieved by admet & dmpk 6(3) (2018) 242-257 lc-ms/ms determination of enzalutamide/darolutamide with metabolites on dbs doi: 10.5599/admet.557 245 monitoring the precursor q1 → q3 ions at m/z 465→209, 451→195, 399→178, 397→194 and 481→453, respectively. both q1 and q3 quadruples were set on unit resolution. the analytical data was processed by analyst software (version 1.6.2). table 1. optimized mass spectrometry parameters for analytes and the is in volts analyte dp ep ce cxp enzalutamide 80 10 41 12 n-desmethylenzalutamide 80 10 41 12 darolutamide 91 10 31 30 orm-15341 60 10 27 12 apalutamide-d3 (is) 100 10 33 12 dp: declustering potential; ep: entrance potential; ce: collision energy and cxp: collision cell exit potential preparation of stock and standard solutions primary stock solutions of enzalutamide, n-desmethylenzalutamide, darolutamide and orm-15341 for preparation of calibration curve (cc) and quality control (qc) samples were prepared from separate weighing. individual primary stock solution of all the analytes at 200 µg/ml were prepared in dmso:methanol (0.2:99.8, v/v). similarly the primary stock solution of the is (200 µg/ml) was prepared in methanol. the primary stock solutions of enzalutamide, n-desmethylenzalutamide, darolutamide, orm15341 and the is were stored at -20 °c, which were found to be stable for one month. the primary stock solutions of enzalutamide, n-desmethylenzalutamide, darolutamide and orm-15341 were successively diluted in dmso:methanol (0.2:99.8, v/v) to prepare secondary stock and working stock solutions to prepare cc and qc samples. working stock solutions were stored approximately at 4 °c for 30 days and found to be stable. a working stock of the is solution (25 ng/ml) was prepared in methanol. samples for the determination of precision and accuracy were prepared by spiking control mice blood in bulk with mixed working stock solution of enzalutamide, n-desmethylenzalutamide, darolutamide and orm-15341 at appropriate concentrations 0.93 ng/ml (lower limit of quantitation quality control, lloq qc), 2.80 ng/ml (low quality control, lqc), 853 ng/ml (medium quality control, mqc) and 1493 ng/ml (high quality control, hqc). dbs cards were prepared (as mentioned in blood spotting section) at each qc level and were stored at -80 ± 10 °c until analysis. blood spotting the dbs cards were prepared from freshly drawn whole blood, harvested on the same day by spotting 25 µl of the respective spiked cc/qc or whole blood from the enzalutamide, n-desmethylenzalutamide, darolutamide and orm-15341 treated mouse onto sampling paper using a calibrated pipette. the samples were left to dry at room temperature least 3 h before storing at controlled room temperature (24 ± 1 °c) until analysis. recovery the efficiency of enzalutamide, n-desmethylenzalutamide, darolutamide, orm-15341 and the is extraction from dbs samples was determined by comparing the responses of the analyte extracted from replicate qc samples (n = 6) with the response of analyte from neat standards at equivalent concentrations by liquid extraction process. recovery of enzalutamide, n-desmethylenzalutamide, darolutamide and r. mullangi et al. admet & dmpk 6(3) (2018) 242-257 246 orm-15341 was determined at lqc (2.80 ng/ml) and hqc (1493 ng/ml) concentrations. recovery of the is was determined at a single concentration of 25 ng/ml. samples were prepared by punching a 3 mm diameter disk from the centre of the dbs using a hole puncher (harri-micro-punch  , 3 mm circle) into a 2.0 ml microcentrifuge tube using the 200 µl of extraction solvents viz., tbme, ethyl acetate, methanol, acetonitrile, methanol/acetonitrile:water (in different ratios). sample preparation to each dbs card (3 mm) in a glass tube, 200 µl of water was added vortex mixed for 10 min and sonicated for 10 min at room temperature. post sonication 200 µl of acetonitrile enriched with the is (25 ng/ml) was added vortex mixed for 10 min followed by centrifugation 10 min. clear supernatant (200 µl) was separated and transferred into a hplc vial for injection and 5.0 µl was injected onto lcms/ms system. method validation the validation experiments were performed in accordance with the us food and drug administration guideline [27]. selectivity selectivity of the method was determined by determining the presence of interfering peaks from six individual drug-free mice dbs samples at the retention times of enzalutamide, n-desmethylenzalutamide, darolutamide, orm-15341 and the is. limit of quantification and carry over the lloq was determined as the concentration that has a precision of <20 % of the relative standard deviation and accuracy between 80-120 % of the theoretical value. the auto-injector carryover was determined by injecting the highest calibration standard, followed by injection of blank samples. the response of the blanks was then compared to that of the lloq. calibration curve calibration samples (for all the analytes) were prepared on each validation day. peak area ratios of each analyte to that of is were used for all calculations. a least squares linear regression (1/x2 weighting factor) of eight non-zero samples was used to define the calibration curve. precision and accuracy the precision and accuracy of the method were evaluated by measuring the four qc samples (lloq qc, lqc, mqc and hqc), which were prepared on each validation day (n = 6 each). inter-day precision was assessed on three separate days. interand intra-day precisions were determined by calculating percent relative standard deviation (%rsd) that should be <15 % for all the qc levels except for lloq qc where it should be <20 %. the interand intra-day accuracy expressed as percent relative error (%re) was calculated by comparing the measured concentration with the nominal value and deviation was limited within ±15 % except for lloq qc where it should be <20 %. matrix effect matrix effect for enzalutamide, n-desmethylenzalutamide, darolutamide and orm-15341 was assessed by comparing each analyte mean peak areas at lqc and hqc concentration after with the mean peak areas for post extracted dbs samples spiked with analyte at equivalent concentrations. matrix effect for the is admet & dmpk 6(3) (2018) 242-257 lc-ms/ms determination of enzalutamide/darolutamide with metabolites on dbs doi: 10.5599/admet.557 247 was assessed at single concentration (25 ng/ml). stability the stability of enzalutamide, n-desmethylenzalutamide, darolutamide and orm-15341 was assessed at lqc and hqc levels in six replicates under all storage conditions. freeze-thaw stability was performed following three freeze-thaw cycles was evaluated (dbs cards were stored in -80 ± 10 °c between freeze/thaw cycles). short-term temperature stability was assessed by analyzing samples that had been kept at ambient temperature (25 ± 1 °c) for 7 days. long-term stability was performed by analyzing samples that had been stored at -80 ± 10 °c for 30 days. the stability of enzalutamide, n-desmethylenzalutamide, darolutamide, orm-15341 and the is in the injection solvent was determined periodically by injecting replicate preparations of processed dbs samples for up to 24 h (in the auto-sampler at 5 °c) after the initial injection. dilution effect to evaluate the effect of dilution over the calibration range, the accuracy and precision of dilution control samples at 10,000 ng/ml (n = 6; 5 times of the uloq) were assessed by performing a 10-fold dilution. influence of hematocrit hematocrit (hct) has a significant effect on the viscosity of the blood, which influences the flux and diffusion properties of the blood through dbs card used for sample collection. it can directly affect the accuracy of the analysis in dbs samples. hct is normally about 43-48 for balb/c mice (in house data). the effect of the hct on analytical performance was investigated by measuring lqcs and hqcs spiked in six different sources of blood adjusted with plasma to obtain values at 25, 35 and 45 % haematocrit that were analysed with calibrators prepared in blood at standard fixed 40 % hct. the relative error of ±15 % and precision of ≤15 % was considered acceptable. hematocrits were measured using mindray bc-5000vet. pharmacokinetic study all the animal experiments were approved by institutional animal ethical committee (iaec/jdc/2017/133). male balb/c mice (n = 12) purchased from vivo biotech, hyderabad were housed in jubilant biosys animal house facility in a temperature (22 ± 1 °c) and humidity (30-70 %) controlled room (15 air changes/h) with a 12:12 h light:dark cycles, had free access to rodent feed (altromin spezialfutter gmbh & co. kg., im seelenkamp 20, d-32791, lage, germany) and water for one week before using for experimental purpose. following ~4 h fast (during the fasting period animals had free access to water) mice (25-32 g) received enzalutamide and darolutamide intravenously [5 % dmso, 5 % solutol:absolute alcohol (1:1, v/v) and 90 % of normal saline; strength: 0.1 mg/ml; dose volume: 10 ml/kg] as a cassette dose at 1.0 mg/kg. post-dosing serial blood samples (25 µl, sparse sampling was done and at each time point three mice were used for blood sampling) were collected at regular intervals using micropipettes (microcaps®; catalogue number: 1-000-0500) through tail vein into polypropylene tubes containing k2.edta solution as an anti-coagulant 0.12, 0.25, 0.5, 1, 2, 4, 8, 24, 48, 72, 96, 120, 144 and 168 h and spotted onto the cards and air dried at room temperature for 3 h before storing at controlled room temperature (25 °c) until analysis. animals were allowed to eat feed 2 h post-dose of enzalutamide and darolutamide. the criteria for acceptance of the analytical runs encompassed the following: (i) 67 % of the qc samples accuracy must be within 85-115 % of the nominal concentration (ii) not less than 50 % at each qc concentration level must meet the acceptance criteria [27]. blood concentration versus time data of r. mullangi et al. admet & dmpk 6(3) (2018) 242-257 248 enzalutamide, darolutamide and orm-15341 was analyzed by non-compartmental method using phoenix winnonlin version 7.0. results ms/ms conditions and chromatography the optimum lc-ms/ms mobile phase was determined through several trials to achieve good resolution and symmetric peak shapes for the analytes and the is in short run time. to obtain good separation, resolution and peak shape for enzalutamide, n-desmethylenzalutamide, darolutamide, orm15341 and the is isocratic elution was used at a flow rate 0.8 ml/min. among the several commercial columns (inertsil, atlantis, kromasil, hypersil etc.) tested, an atlantis dc18 column (50 x 4.6 mm, 5 µm) was selected because it provided symmetric peaks and baseline separation of enzalutamide, n-desmethylenzalutamide, darolutamide, orm-15341 and the is within 2.5 min run time. the ms/ms ion transitions monitored were m/z 465→209, m/z 451→195, m/z 399→178, m/z 397→194 and m/z 481→453 for enzalutamide, n-desmethylenzalutamide, darolutamide, orm-15341 and the is, respectively for quantitation. the retention time of enzalutamide, n-desmethylenzalutamide, darolutamide, orm-15341 and the is was 1.32, 1.23, 1.01, 1.12 and 1.43 min, respectively. the total run time is 2.5 min (fig. 2a-d). method validation selectivity as shown in figures 2a-d no significant interferences in the blank dbs traces were found from endogenous components in drug-free mice blood at the retention times of the enzalutamide, ndesmethylenzalutamide, darolutamide, orm-15341 and the is indicating that the method is selective. sensitivity the lowest limit of reliable quantification for the analyte was set at the concentration of the lloq. the precision and accuracy at lloq concentration were found to be 7.00 and 103 % for enzalutamide; 3.70 and 98.0 % for n-desmethylenzalutamide; 5.25 and 95.2 % for darolutamide and 2.85 and 103 % for orm15341. there was no carry-over produced by the highest calibration sample on the following injected blank dbs extract sample. recovery the recovery (mean ± s.d) for enzalutamide, n-desmethylenzalutamide, darolutamide and orm-15341 (at lqc, mqc and hqc) is presented in table 2. recovery for the is was 85.4 ± 5.89 %. matrix effect six different lots of dbs samples, spiked with analytes concentration levels at lqc and hqc levels were analyzed. the results have shown that the precision and accuracy for analyzed samples were within acceptance range (table 2). overall it was found that there is no impact on the ionization of analytes and the is. matrix effect for the is was 105 ± 1.25 %. admet & dmpk 6(3) (2018) 242-257 lc-ms/ms determination of enzalutamide/darolutamide with metabolites on dbs doi: 10.5599/admet.557 249 a) b) c) d) figure 2.typical mrm chromatograms mice blank dbs (left panel), is (middle panel) and lloq (right panel) for (a) enzalutamide (b) n-desmethylenzalutamide (c) darolutamide and (d) orm-15341 r. mullangi et al. admet & dmpk 6(3) (2018) 242-257 250 table 2. recovery and matrix data for enzalutamide, n-desmethylenzalutamide, darolutamide and orm-15341 quality controls on dbs cards mean recovery, % mean ± sd (n=4) mean absolute matrix effect, % mean ± sd (n=4) lqc (2.80 ng/ml) enzalutamide n-desmethylenzalutamide darolutamide orm-15341 84.4 ± 11.0 78.0 ± 7.83 47.2 ± 4.55 79.7 ± 15.4 99.2 ± 5.85 101 ± 8.58 88.3 ± 2.34 95.8 ± 6.23 hqc (1493 ng/ml) enzalutamide n-desmethylenzalutamide darolutamide orm-15341 78.6 ± 4.50 76.7 ± 3.75 50.3 ± 1.99 77.1 ± 3.32 95.3 ± 12.0 108 ± 5.68 85.2 ± 1.35 98.6 ± 2.38 sd: standard deviation calibration curve the blood calibration curve was constructed using eight calibration standards (viz., 0.93-2000 ng/ml). the calibration standard curve had a reliable reproducibility over the standard concentrations across the calibration range. calibration curve was prepared by determining the best fit of peak-area ratios (peak area analyte/peak area is) versus concentration, and fitted to the y = mx + c using weighing factor (1/x 2 ). the average slope and intercept values were found to be 0.005358 and 0.002326; 0.005078 and -0.000221; 0.00682 and 0.000485; 0.00967 and 0.000272 for enzalutamide, n-desmethylenzalutamide, darolutamide and orm-15341, respectively. the average regression (n = 4) was found to be > 0.99 for all the analytes. the lowest concentration with the rsd <20 % was taken as lloq and was found to be 0.93 ng/ml. the % accuracy observed for the mean of back-calculated concentrations for four calibration curves was within 86.8-114; 85.0-108; 87.1-110 and 90.9-103; while the precision (% cv) values ranged from 4.42-10.3; 5.319.34; 1.60-11.5 and 1.12-9.62 for enzalutamide, n-desmethylenzalutamide, darolutamide and orm-15341, respectively. accuracy and precision accuracy and precision data for intraand inter-day dbs samples for enzalutamide, ndesmethylenzalutamide, darolutamide and orm-15341 are presented in table 3. the assay values on both the occasions (intraand inter-day) were found to be within the accepted variable limits. stability the measured concentrations for enzalutamide, n-desmethylenzalutamide, darolutamide and orm15341 at 2.80 and 1493 ng/ml were within 85-115 % of re and <15 % rsd in a battery of stability tests viz., in-injector (24 h), at room temperature for 7 days, repeated three freeze/thaw cycles and freezer stability at -80 ± 10 °c for at least for 30 days (table 4). dilution effect the accuracy of the nominal concentration of the diluted dbs samples were within 5.28 % and the precision was 8.56 % for all the analytes, which show the ability to dilute samples up to a dilution factor of ten in a linear fashion. admet & dmpk 6(3) (2018) 242-257 lc-ms/ms determination of enzalutamide/darolutamide with metabolites on dbs doi: 10.5599/admet.557 251 table 3. intraand inter-day precision and accuracy determination for determination of enzalutamide, ndesmethylenzalutamide, darolutamide and orm-15341 quality controls on dbs cards parameters enzalutamide n-desmethylenzalutamide darolutamide orm-15341 intra-day lloq qc (0.93 ng/ml) precision (% rsd) 6.09 2.08 9.36 2.68 accuracy (%re) 0.97 0.92 1.02 0.96 lqc (2.80 ng/ml) precision (% rsd) 3.24 0.18 10.2 11.7 accuracy (%re) 0.90 0.96 1.09 0.96 mqc (853 ng/ml) precision (% rsd) 6.53 3.45 5.17 4.73 accuracy (%re) 1.04 1.11 0.97 1.14 hqc (1493 ng/ml) precision (% rsd) 0.90 3.92 5.26 12.0 accuracy (%re) 0.99 1.00 0.86 1.13 inter-day lloq qc (0.93 ng/ml) precision (% rsd) 8.01 5.01 6.54 1.02 accuracy (%re) 1.00 1.07 1.05 1.06 lqc (2.80 ng/ml) precision (% rsd) 7.54 3.29 10.8 5.25 accuracy (%re) 0.96 1.04 0.96 1.02 mqc (853 ng/ml) precision (% rsd) 2.38 2.73 8.58 10.8 accuracy (%re) 1.02 1.12 1.12 1.14 hqc (1493 ng/ml) precision (% rsd) 2.18 1.58 11.2 9.28 accuracy (%re) 1.04 0.99 1.00 1.14 % rsd: relative standard deviation (sd x 100/mean) % re: relative error (measured value/actual value) hematocrit effect the measured enzalutamide, n-desmethylenzalutamide, darolutamide and orm-15341 concentrations were compared with the results obtained from dbs samples are given in table 5. the % difference was calculated by subtracting the % relative error of hct 25 and 45 % from % relative error of 40 % hct which was taken as standard hct value. no significant impact of hct on accuracy was observed for all the analytes. r. mullangi et al. admet & dmpk 6(3) (2018) 242-257 252 table 4. stability data of enzalutamide, n-desmethylenzalutamide, darolutamide and orm-15341 quality controls on dbs cards parameters enzalutamide n-desmethylenzalutamide darolutamide orm15341 room temperature (7 days) lqc (2.80 ng/ml) precision (% rsd) 2.58 7.58 5.98 7.68 accuracy (%re) 0.91 0.86 0.95 0.98 hqc (1493 ng/ml) precision (% rsd) 5.84 5.68 3.56 4.89 accuracy (%re) 1.02 0.95 0.98 0.86 in-injector (24 h) lqc (2.80 ng/ml) precision (% rsd) 6.78 0.73 7.04 4.60 accuracy (%re) 0.87 0.85 0.90 0.94 hqc (1493 ng/ml) precision (% rsd) 8.58 10.3 9.08 11.0 accuracy (%re) 0.95 0.86 1.00 0.87 3 f/t cycle lqc (2.80 ng/ml) precision (% rsd) 5.03 6.60 8.65 5.68 accuracy (%re) 1.10 0.99 0.95 0.88 hqc (1493 ng/ml) precision (% rsd) 6.47 11.6 5.28 11.7 accuracy (%re) 1.03 0.94 0.94 0.86 long-term stability at -80°c (30 days) lqc (2.80 ng/ml) precision (% rsd) 4.77 9.21 10.2 6.60 accuracy (%re) 1.01 1.02 0.91 0.89 hqc (1493 ng/ml) precision (% rsd) 4.03 8.81 10.2 4.85 accuracy (%re) 1.04 0.99 0.87 0.85 % rsd: relative standard deviation (sd x 100/mean) % re: relative error (measured value/actual value) pharmacokinetic study the sensitivity and specificity of the assay were found to be sufficient for accurately characterizing the pharmacokinetics of enzalutamide, n-desmethylenzalutamide (released from enzalutamide), darolutamide and orm-15341 (released from darolutamide) in blood following cassette dose intravenous administration at 1.0 mg/kg dose of enzalutamide and darolutamide to male balbc mice. fig. 3a shows mean blood concentration-time profile for enzalutamide and it was quantified till 168 h post-dosing by intravenous route. however the blood concentrations for n-desmethylenzalutamide were seen just above lloq at 4, 8, 24 and 48 h (ranged between 1.23-4.30 ng/ml) and below lloq at remaining time points hence admet & dmpk 6(3) (2018) 242-257 lc-ms/ms determination of enzalutamide/darolutamide with metabolites on dbs doi: 10.5599/admet.557 253 pharmacokinetic parameters are not calculated for n-desmethylenzalutamide by using this limited data. fig. 3b shows mean blood concentration-time profile for darolutamide and orm-15341, both were quantified till 24 and 8 h, respectively post-dosing by intravenous route. summary of the pharmacokinetic profiles of enzalutamide, darolutamide and orm-15341 after single dose intravenous administration in male balbc mice were shown in table 6. in order to validate this method the pharmacokinetics results of enzalutamide, darolutamide and orm-15341 in the present study were compared with earlier reported pharmacokinetic parameters determined using plasma as a matrix [19] and found that good correlation between both measurements. this shows that dbs can be used as a promising alternative suitable to predict exposure of enzalutamide, darolutamide and orm-15341. n-desmethylenzalutamide pharmacokinetic parameters were not calculated due to low exposure and in our earlier study we did not calculate n-desmethylenzalutamide pharmacokinetic parameters [19]. table 5. impact of hematocrit value of enzalutamide, n-desmethylenzalutamide, darolutamide and orm-15341 on the quality control sample concentration (n = 3) concentration spiked 2.80 ng/ml 1493 ng/ml concentration found (ng/ml) hematocrit (%) 25 35 45 25 35 45 enzalutamide 2.63 2.53 2.79 150 6 1561 1526 n-desmethylenzalutamide 2.96 3.04 2.78 155 4 1498 1585 darolutamide 2.80 2.85 2.86 153 7 1633 1583 orm-15341 3.07 3.06 2.56 146 6 1604 1498 accuracy (%re) enzalutamide 0.94 0.90 1.00 1.01 1.05 1.02 n-desmethylenzalutamide 1.06 1.09 0.99 1.04 1.00 1.06 darolutamide 1.00 1.02 1.02 1.03 1.09 1.06 orm-15341 1.10 1.09 0.91 0.98 1.07 1.00 precision (% rsd) enzalutamide 3.96 2.38 5.89 10.2 7.10 9.85 n-desmethylenzalutamide 7.12 7.33 8.65 5.87 10.6 8.36 darolutamide 9.21 7.12 2.35 11.2 12.3 10.9 orm-15341 8.18 5.55 9.38 12.4 2.95 12.3 table 6. pharmacokinetic parameters for enzaluatmide, daroluatmide and orm-15341 following cassette dose intravenous administration to mice at 1 mg/kg pk parameters enzalutamide darolutamide orm-15341 auc(0-) (ng  h/ml) 24332 1514 2214 c0 (ng/ml) 1180 3042 1251 t1/2 (h) 31.0 4.64 1.56 cl (ml/min/kg) 0.68 11.0 -- vd (l/kg) 1.84 4.42 -- r. mullangi et al. admet & dmpk 6(3) (2018) 242-257 254 a) 0 24 48 72 96 120 144 168 1 10 100 1000 10000 enzalutamide time (h) c o n c e n tr a ti o n i n b lo o d (n g /m l) 0 4 8 12 16 20 24 1 10 100 1000 10000 enzalutamide time (h) b) 0 4 8 12 16 20 24 1 10 100 1000 10000 darolutamide time (h) c o n c e n tr a ti o n i n b lo o d (n g /m l) orm-15341 figure 3. mean blood concentration-time profile of (a) enzalutamide in mice blood following intravenous cassette dosing of enzalutamide and darolutamide to mice at 1.0 mg/kg; (b) darolutamide and orm-15341 (released from darolutamide) in mice blood following intravenous cassette dosing of enzalutamide and darolutamide to mice at 1.0 mg/kg. discussion the objective of the present method was to develop and validate a new simple, rapid and sensitive dbs method for simultaneous quantification of enzalutamide, n-desemthylenzalutamide, darolutamide and orm-15341 in mice blood and apply to a pharmacokinetic study in mice. dbs [whatman dmpk fta or autocollect dbs (ahlstrom-munksjo) cards] is increasingly becoming an important tool in preclinical drug development, pharmacokinetic/toxicity studies, neonatal screening, clinical pharmacology, forensic toxicology, doping analysis and therapeutic drug monitoring. dbs method offers several advantages like reduced blood/plasma/serum volume per time point, minimally invasive, ease and safety in handling, reduce risk of infection when compared with conventional wet sampling. it offers potential cost saving in storage, shipment, sample processing time (usually done at room temperature) and increase in throughput etc [28,29]. admet & dmpk 6(3) (2018) 242-257 lc-ms/ms determination of enzalutamide/darolutamide with metabolites on dbs doi: 10.5599/admet.557 255 to achieve a dbs method for the first-time for second generation non-steroidal anti-androgens namely enzalutamide and darolutamide along with its active metabolite (n-desmethylenzalutamide and orm15341, respectively) during method development different options were evaluated to optimize dbs sample extraction and chromatography. recovery was very poor (<8 %) with organic solvents (tbme, ethyl acetate, methanol and acetonitrile) for all the analytes. then combination of methanol/acetonitrile:water in various ratios were systematically explored. the samples extracted with acetonitrile:water (80:20; 50:50 and 20:80, v/v) have shown low recovery (~12 %). at 80:20, 50:50 and 20:80 methanol:water v/v ratios the recoveries ranged between 2232 % but not reproducible for all the analytes. water as an extraction solvent gave good recovery for the analytes (~47-50 % for darolutamide and >77-84 % for rest of the analytes) and data was consistent and reproducible. in order to increase the assay precision, limit the variability between analyte(s) and the is, to mimic the analyte during ionization, extraction and chromatography we have used apalutamide-d3 as an is. to avoid the potentially co-eluting peaks from whole blood, which will influence the ionization efficiency of the analytes and the is, the mobile phase was optimized. the attained lloq (0.93 ng/ml) for each analyte (comparable lloq attained with plasma earlier reported by us) was sufficient to quantify pharmacokinetic parameters for enzalutamide, darolutamide and orm-15341 in a pharmacokinetic study in mice. the acceptable limit for both intraand inter-day accuracy and precision is ±15 % of the nominal values for all, except for lloq qc which should be within ±20%. in this method, both intraand inter-day accuracy and precision are well within this limit, indicating that the developed method is precise and accurate for the quantification of enzalutamide, n-desmethylenzalutamide, darolutamide and orm-15341. conclusions we developed and validated an analytical method for the determination of enzalutamide, ndesmethylenzalutamide, darolutamide and orm-15341 in mice dbs samples for these second generation non-steroidal anti-androgens. the method is selective, linear, accurate and precise in the range of 0.932000 ng/ml. the dbs samples were stable in the studied conditions and haematocrit did not influence enzalutamide, n-desmethylenzalutamide, darolutamide and orm-15341 quantification. the developed dbs method has several advantages viz., low blood volume (25 µl), reduce the large number of animal usage, allows serial blood sampling from mice, useful in toxicokinetic studies etc. further, blood samples having hematocrit values between 25 and 45 % demonstrated acceptable accuracy and precision in the quantitative measurements. the utility of this assay was demonstrated through its successful application to a mice pharmacokinetic study. conflict of interest: the authors wish to declare that there are no conflicts of interests in the contents of the manuscript. references [1] r.l. siegel, k.d. miller, a. jemal. cancer statistics 2017. ca: a cancer journal for clinicians 67 (2017) 7-30. [2] c.j. ryan, m.l. cheng. abiraterone acetate for the treatment of prostate cancer. expert opinion on pharmacotherapy 14 (2013) 91-96. r. mullangi et al. admet & dmpk 6(3) (2018) 242-257 256 [3] h.i. scher, c.l. sawyers. biology of progressive, castration-resistant prostate cancer: directed therapies targeting the androgen-receptor signaling axis. journal of clinical oncology 23 (2005) 8253-8261. [4] l.j. scott. abiraterone acetate: a review in metastatic castration-resistant prostate cancer. drugs 77 (2017) 1565-1576. [5] a. mullard. 2012 fda drug approvals. nature reviews drug discovery 12 (2013) 87-90. [6] z.t. al-salama. apalutamide: first global approval. drugs 78 (2018) 699-705. [7] a.m. moilanen, r. riikonen, r. oksala, l. ravanti, e. aho, g. wohlfahrt, p.s. nykänen, o.p. törmäkangas, j.j. palvimo, p.j. kallio. discovery of odm-201, a new-generation androgen receptor inhibitor targeting resistance mechanisms to androgen signaling-directed prostate cancer therapies. scientific reports 5 (2017) 12007. [8] j. semenas, n. dizeyi, j.l. persson. enzalutamide as a second generation antiandrogen for treatment of advanced prostate cancer. drug design development and therapy 7 (2013) 875-881. [9] j.a. gibbons, t. ouatas, w. krauwinkel, y. ohtsu, j.s. van der walt, v. beddo, m. de vries, j. mordenti. clinical pharmacokinetic studies of enzalutamide. clinical pharmacokinetics 54 (2015) 1043-1055. [10] j. weiss, j. kocher, c. mueller, s. rosenzweig, d. theile. impact of enzalutamide and its main metabolite n-desmethylenzalutamide on pharmacokinetically important drug metabolizing enzymes and drug transporters. biopharmaceutics & drug disposition 38 (2017) 517-525. [11] k. fizazi, c. massard, n.d. james. odm-201, a new generation androgen receptor inhibitor for castration-resistant prostate cancer: preclinical and phase i data. american society of clinical oncology journal 31 (suppl 6) (2013) abstract 65. [12] a. moilanen, r. riikonen, r. oksala. odm-201 new generation antiandrogen with excellent antiandrogenic and antitumor activity in nonclinical models of crpc. european journal of cancer 49 (suppl 2) (2013) abstract 685. [13] orion pharma clinical study code: 3104001/en3386-201 (accessed on 7 june 2018). [14] j.h. song, t.h. kim, j.w. jung, n. kim, s.h. ahn, s.o. hwang, n.s. kang, s.w. yoo, t.s. koo. quantitative determination of enzalutamide, an anti-prostate cancer drug, in rat plasma using liquid chromatography-tandem mass spectrometry, and its application to a pharmacokinetic study. biomedical chromatography 28 (2014) 1112-1117. [15] s. bennett, j.a. gibbons, r. mol, y. ohtsu, c. williard. validation of a method for quantifying enzalutamide and its major metabolites in human plasma by lc-ms/ms. bioanalysis 6 (2014) 737744. [16] y. ohtsu, d.r. thakker, j.a. gibbons, k. tsubota, s. otsuka, h. arai. determination of the androgen receptor inhibitor enzalutamide and its metabolites in animal plasma and brain homogenates using lc-ms/ms and its application to pharmacokinetic studies. chromatography 36 (2015) 115-122. [17] k.p. kim, r.a. parise, j.l. holleran, l.d. lewis, l. appleman, n. van erp, m.j. morris, j.h. beumer. simultaneous quantitation of abiraterone, enzalutamide, n-desmethylenzalutamide, and bicalutamide in human plasma by lc-ms/ms. journal of pharmaceutical and biomedical analysis 138 (2017) 197-205. [18] m. van nuland, m.j.x. hillebrand, h. rosing, j.h.m. schellens, j.h. beijnen. development and validation of an lc-ms/ms method for the simultaneous quantification of abiraterone, enzalutamide, and their major metabolites in human plasma. therapeutic drug monitoring 39 (2017) 243-251. [19] s.p. sulochana, n.k. saini, p. daram, p.s. babu, r. mullangi. validation of an lc-ms/ms method for simultaneous quantitation of enzalutamide, n-desmethylenzalutamide, apalutamide, darolutamide and orm-15341 in mice plasma and its application to a mice pharmacokinetic study. journal of pharmaceutical and biomedical analysis 156 (218) 170-178. admet & dmpk 6(3) (2018) 242-257 lc-ms/ms determination of enzalutamide/darolutamide with metabolites on dbs doi: 10.5599/admet.557 257 [20] m. herbrink, n. de vries, h. rosing, a.d.r. huitema, b. nuijen, j.h.m. schellens, j.h. beijnen. development and validation of a liquid chromatography-tandem mass spectrometry analytical method for the therapeutic drug monitoring of eight novel anticancer drugs. biomedical chromatography 32 (2017) e4147. [21] s. dittakavi, p.k.v.s.p. nagasuri, s.p. sulochana, s.m. saim, s.r. mallurwar, m. zainuddin, p. dewang, s. rajagopal, r. mullangi. lc-ms/ms-esi method for simultaneous quantification of darolutamide and its active metabolite, orm-15341 in mice plasma and its application to a pharmacokinetic study. journal of pharmaceutical and biomedical analysis 145 (2017) 454-461. [22] n. balaji, s.p. sulochana, n.k. saini, a.s. kumar, r. mullangi. validation of a chiral lc-ms/ms-esi method for the simultaneous quantification of darolutamide diastereomers in mice plasma and its application to a stereoselective pharmacokinetic study in mice. biomedical chromatography 32 (2018) e4173. [23] n. balaji, s.p. sulochana, n.k. saini, a.s. kumar, r. mullangi. validation of a chiral lc-ms/ms-esi method for the simultaneous quantification of darolutamide diastereomers and orm-15341 (active metabolite) in mice plasma: application to a pharmacokinetic study in mice. drug research (2018) in press. [24] y. xu, e.j. woolf, n.g. agrawal, p. kothare, v. pucci, k.p. bateman. merck's perspective on the implementation of dried blood spot technology in clinical drug development why, when and how. bioanalysis 5 (2013) 341-350. [25] r.k. rej, r.k. acharyya, s. nanda. asymmetric synthesis of dihydroartemisinic acid through intramolecular stetter reaction. tetrahedron 72 (2016) 4931-4937. [26] x. pang, y. wang, y. chen. design, synthesis, and biological evaluation of deuterated apalutamide with improved pharmacokinetic profiles. bioorganic & medicinal chemistry letters 27 (2017) 28032806. [27] us dhhs, fda, cder, cvm, guidance for industry: bioanalytical method validation, u.s. department of health and human services, food and drug administration, center for drug evaluation and research (cder), center for veterinary medicine (cv), (2018), rockville, md, usa. [28] p. a. demirev. dried blood spots: analysis and applications. analytical chemistry 85 (2013) 779-789. [29] j.d. freeman, l.m. rosman, j.d. ratcliff, p.t. strickland, d.r. graham, e.k. silbergeld. state of the science in dried blood spots. clinical chemistry 64 (2018) 656-679. ©2018 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ peptide retention time prediction for immobilized artificial membrane phosphatidylcholine stationary phase: method development and preliminary observations doi: 10.5599/admet.520 190 admet & dmpk 6(2) (2018) 190-199; doi: http://dx.doi.org/10.5599/admet.520 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper peptide retention time prediction for immobilized artificial membrane phosphatidylcholine stationary phase: method development and preliminary observations daniel gussakovsky 1 , haley neustaeter 1 , victor spicer 2 , oleg v. krokhin 2,3 * 1 department of chemistry, university of manitoba, 360 parker building, winnipeg, manitoba r3t 2n2, canada 2 manitoba centre for proteomics and systems biology, university of manitoba, 799 jbrc, 715 mcdermot avenue, winnipeg, manitoba r3e 3p4, canada 3 department of internal medicine, university of manitoba, 799 jbrc, 715 mcdermot avenue, winnipeg, manitoba r3e 3p4, canada *corresponding author: e-mail: oleg.krokhine@umanitoba.ca; tel.: +1-204-789-3283; fax: +1-204-480-1362 received: march 14, 2018; revised: may 16, 2018; available online: may 24, 2018 abstract development of the first peptide retention prediction model for immobilized artificial membrane phosphatidylcholine (iam.pc) stationary phase is reported. 2d liquid chromatography coupled to tandem mass spectrometry (2d lc-ms/ms) analysis of a whole cell lysate of s. cerevisiae yielded a retention dataset of ~29,500 tryptic peptides; sufficient for confident assignment of retention coefficients which determine the contribution of individual amino acids in peptide retention. retention data from the first dimension was used for the modelling: an iam.pc.dd2 column, with ph 7.4 ammonium bicarbonate, and a water/acetonitrile gradient. peptide separation using the iam.pc.dd2 phase was compared to a standard c18 phase (luna c18(2)). there was a significant reduction in peptide retention (~14 % acetonitrile on average), indicating that the phosphatidylcholine stationary phase is significantly more hydrophilic. in comparison to the c18 phase, a substantial increase was found in the relative retention contribution for the positively charged arg and lys, and the aromatic tyr, trp and his residues. a decrease in retention contribution was observed for the negatively charged asp and glu. this indicates an involvement of electrostatic interactions with the glycerophosphate functional groups, and possibly, delocalization effects from hydrogen bonds between the phosphate group and the aromatic side chains in the separation mechanism. keywords peptide retention modelling; immobilized artificial membrane chromatography introduction modern applications of chromatography have spread far beyond its original role as a method for preparative separations. years of development have established chromatography as a leading analytical technique covering virtually all fields of analytical chemistry. the contribution by high-performance liquid chromatography (hplc) into studying physicochemical fundamentals of interactions in heterogeneous biological systems is also well appreciated. establishing hydrophobicity scales of amino acids [1] to support original hydrophobicity measurement data obtained by x-ray crystallography [2] and studying interactions http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:oleg.krokhine@umanitoba.ca admet & dmpk 6(2) (2018) 190-199 peptide retention time prediction for iam.pc stationary phases doi: 10.5599/admet.520 191 of various substances in the systems mimicking real biological environments [3] represent some of the most interesting fundamental biochemical applications of chromatography. the latter, now known under the broad term of biomimetic chromatography, has found applications in developing fast assays to determine lipophilicity, protein binding, and phospholipid binding – all critically important parameters in drug design. measuring the retention properties of molecules on a c18 reversed-phase support provides information on the hydrophobicity of these molecules. similar measurements on phospholipid-modified phases such as immobilized artificial membrane phosphatidylcholine (iam.pc) provide more biologically relevant information on phospholipid binding, i.e. cell membrane permeability [4]. reversed-phase hplc has been pin-pointed as a powerful technique for elucidating the hydrophobicity contribution of individual amino acids [5] and the interacting domains between the peptide and hydrophobic surfaces [6]. houghten [6-8] and hodges’ [9-11] were the first to address the influence of amphipathic helicity in the stabilization of helical peptides upon contact with a hydrophobic surface – a key mechanism in antimicrobial peptides’ action [11]. studies such as this represent a perfect example of bridging the gap between hplc as a method of physicochemical study and drug development. all of these efforts have originated from the first attempts to model/predict peptide behaviour in reversed-phase liquid chromatography (rplc) [12,13]. the goal of the early modelling studies was to simplify method development for peptide analytical hplc with uv detection. the arrival of high-throughput proteomic technology led to the expansion of these applications into protein/peptide identification [14], development of quantitative lc-ms methods [15], and the guided design of multi-dimensional peptide separation systems [16]. proteomics has provided a significant increase in the size of peptide retention data, paving the way for the development of the first sequence-specific peptide retention prediction models [17-19]. the majority of retention modelling studies in the proteomics era targeted rplc separations with formic acid as the ion-pairing modifier. our sequence-specific retention calculator (ssrcalc) model has been a benchmark tool in this field since 2004 (available on-line at http://hs2.proteome.ca/ssrcalc/ssrcalcq.html), and followed the same trend but with the addition of models for trifluoroacetic acid [19] and high ph reversed-phase [20]. other peptide separation modes have largely excluded because of the poor compatibility of the eluents with esi-ms. however, the last few years have witnessed an expansion of prediction studies into other separation mechanisms. in 2017 our laboratory applied ssrcalc methodology to hydrophilic interaction liquid chromatography (hilic) [21], strong-cation exchange (scx) [22], strong-anion exchange (sax) (manuscript in preparation), and capillary zone electrophoresis (cze) [23]. all of the listed prediction models that we have generated are the most accurate models reported for the respective modes of peptide separation. application of 2d lc-ms/ms of the complex tryptic digests was the key innovation allowing collection of large retention datasets for peptide separations not possible with on-line esi-ms detection [20, 21]. standard rp (formic acid) lcms/ms was used in the second dimension as a “standard detection device”, while the first-dimension separation information was used as a modelling dataset: e.g. hilic-rplc-ms for hilic models, scx-rplcms for scx, etc. we are not aware of any peptide retention modelling studies on chromatographic supports for biomimetic applications. this would provide a significant advantage by expanding predictive approaches to peptide-based drug design. having extensive experience in peptide retention modelling in various separation systems, we concluded that the use of 2d iam.pc-rplc to study peptide separation in biomimetic applications would be the first step in this direction. the goal of this study was to establish a large-scale retention data collection protocol for peptide separation on the iam.pc.dd2 stationary phase http://hs2.proteome.ca/ssrcalc/ssrcalcq.html oleg krokhin et al. admet & dmpk 6(2) (2018) 190-199 192 and gain a first insight into peptide retention mechanism in this system. this was done by developing a peptide retention prediction model and assigning contributions of individual amino acids into peptide retention. experimental experimental procedures were identical to the previously reported modelling studies of hilic and scx [21, 22], except for the chromatographic parameters (columns and eluents) in the first-dimension separation. overall, the procedure (figure 1) included a tryptic digestion of the whole cell s. cerevisiae extract, first dimension separation in a reversed-phase mode using a luna c18(2) or and iam.pc.dd2 column, fraction collection, and lc-ms/ms analysis of the individual fractions followed by peptide identification using the x!tandem search algorithm. figure 1. overview of the experimental procedure for the large-scale peptide retention data collection using 2d lc-ms/ms. materials, chromatographic columns deionized water and hplc-grade acetonitrile were used for the preparation of the eluents. all chemicals were sourced from sigma chemicals (st. louis, mo) unless otherwise stated. amicon centrifugal filter units (15 ml) (merck millipore, ireland) and sequencing grade modified trypsin (promega, madison, wi) were used for the digestion. standard peptides p1-p6 [24] synthesized by biosynthesis inc. (lewisville, tx) were used for the preliminary experiments to establish the gradient separation conditions at ph 7.4. luna c18(2), 5 µm (phenomenex, torrance, ca) and iam.pc.dd2, 10 µm (regis technologies, grove, il) columns (1x100 mm) were packed in-house. first dimension separation an agilent 1100 series hplc system fitted with uv detector (214 nm), a 100 µl injection loop and operating at a 150 µl/min flow rate was used for separations of standard peptide mixtures and the complex digest. identical gradients of 1% acetonitrile per minute were used for both columns. eluent a consisted of 20 mm ammonium bicarbonate in water. a 200 mm ammonium bicarbonate stock solution was diluted 10 times and the ph was adjusted with formic acid to 7.4. eluent b consisted of 20 mm ammonium bicarbonate, ph 7.4, in 70/30 acetonitrile/water. the gradient program included the following steps: a linear increase from 0 to 71.4 % b in 50 min, 10 min wash with 90 % b and 30 min equilibration with 100 % a. one-minute fractions were collected within the expected interval of peptide elution. fractions were lyophilized and re-suspended in 30 l of 0.2% formic acid in water and spiked with ~200 fm of standard peptides p1-p6 for retention time alignment purposes. one-third of each collected fraction (10 µl) was injected in the second dimension. admet & dmpk 6(2) (2018) 190-199 peptide retention time prediction for iam.pc stationary phases doi: 10.5599/admet.520 193 second dimension lc-ms/ms second dimension lc-ms/ms was done using a standard data-dependent acquisition protocol using a 2d lc ultra system (eksigent, dublin, ca) and a tripletof5600 mass spectrometer (sciex, concord, on) as described [19]. lc settings featured a 100 µm x 200mm analytical column packed with 3 µm luna c18(2) (phenomenex, torrance, ca) and a 300 µm x 5 mm pepmap 100 trap-column (thermo fisher). a 500 nl/min flow rate was used with ~0.4 % acetonitrile per minute gradient. both buffers a (water) and b (acetonitrile) contained 0.1 % formic acid. the gradient program consisted of the following steps: a linear increase from 0.4 to 31 % buffer b (acetonitrile) in 77 minutes, 5 minutes at 80 % b and then 8 minutes at 0.4 % b for column equilibration (90 min total analysis time). data analysis and retention time assignment x!tandem’s search algorithm was used with the following parameters: 20 ppm and 50 ppm mass tolerance for parent and daughter ions, respectively; constant modification of cys with iodoacetamide. all identified tryptic non-modified peptides (log (e) < -1) were additionally filtered using retention time prediction in the second dimension. retention times in the first dimension were assigned as equal to the fraction number in which the peptide was found. when the peptide signal was distributed between two or more fractions, the intensity weighted average fraction number was used. results and discussion selection of peptide reversed-phase separation conditions on c18 and iam.pc.dd2 phases at ph 7.4. the majority of peptide retention time modelling studies have been performed using acidic eluent conditions (usually formic acid) – a standard setting in proteomic lc-ms. however, biomimetic separations usually use physiological ph to maximize the similarity between biological systems and the artificial biphasic separation environment. we decided to employ ammonium bicarbonate – based buffer at ph 7.4 and performed separations on both the iam.pc.dd2 phase and a standard c18 phase for comparison. figure 2 (a, b) shows the separation of a standard mixture of 6 peptides on these two columns. the p1p6 peptide mixture was designed to cover the entire hydrophobicity range for tryptic peptides: i.e. the elution window of the reversed-phase separations. at acidic eluent conditions (0.1% trifluoroacetic acid) they elute between 4 (p1) and 29.6 (p6) % acetonitrile (min) [24] – very similar to the ~26 % acetonitrile (min) retention window on c18 phase at ph 7.4 (figure 2a). peptide retention on the iam.pc is significantly lower (figure 2b). peptides p1-p3 are not retained, while the retention time decrease for p4-p6 was ~16.7 min (% acetonitrile) on average compared to c18. a significant decrease in separation efficiency for iam.pc.dd2 was also obvious and likely due to the introduction of mixed-mode interactions on the phosphatidylcholine phase and larger particle size. most tryptic peptides are expected to elute in the range between 5 and 45 min from the luna c18(2) column under the chromatographic conditions used. based on preliminary experiments with standard peptides, the elution window for iam.pc was expected to be smaller. noting this we performed separations of complex s. cerevisiae digests (~150 µg, figure 2(c,d)) and collected fractions up to 45 min for each run. as expected, both chromatograms showed no well-resolved chromatographic peaks due to the extremely high complexity of the mixture. at the same time, the overall retention profiles of tryptic peptides in these two systems were quite different. the majority of peptides were retained on the c18 column and eluted as a typical bell-shaped profile within a 5-40 min window (figure 2c). the separation on oleg krokhin et al. admet & dmpk 6(2) (2018) 190-199 194 the iam.pc.dd2 phase exhibited a significant “break-through” – a very high peak at the beginning of the chromatogram containing peptides, which were not retained under the starting gradient conditions – similar to p1-p3 in figure 1b. figure 2. separation of standard peptides and s. cerevisiae digest using luna c18(2) and iam.pc.dd2 columns in the reversed-phase separation mode. a and b – separation of standard peptides p1-p6 (lggggggdgsr, lggggggdfr, llggggdfr, lllggdfr, lllldfr, llllldfr, [24] injection of ~1 µg of each peptide) on luna c18(2) and iam.pc.dd2, respectively; c and d – separation of ~150 µg of s. cerevisiae digests. all chromatograms have been obtained using identical separation conditions with a 1 % acetonitrile per minute gradient at ph 7.4. identification outputs for both 2d lc-ms/ms runs identification outputs for both 2d lc-ms/ms runs are shown in table 1, indicating significantly higher redundancy in identification for the iam.pc.dd2 run. due to the lower separation efficiency in the first iam.pc.dd2 dimension, individual peptides were distributed through a larger number of fractions. this led to an acquisition of a larger number of ms/ms spectra, more identified spectra, but a lower number of unique peptide ids. figure 3a shows the correlation between retention time (fraction number) on the two columns. as expected, a significant portion of the peptides which show a moderate retention on luna c18, is not retained on iam.pc.dd2 and thus elute in the early fractions. table 1. identification output of 2d (rp-rp) lc-ms/ms and 2d (iam.pc-rp)-lc ms/ms for the analysis of whole cell yeast tryptic digest. separation mode number of fractions total lcms time (hr) amount injected (g) number of ms/ms number of identified peptides* number of nonredundant peptide ids* number of protein ids iam.pc-rp 40 60 ~40 382972 229565 40602 4225 rp-rp 45 67.5 ~45 312846 182435 43931 4295 * these numbers include ~10 % peptides with post-translational modifications (default settings of ptms for x!tandem was used (methionine oxidation, deamidation and n-terminal cyclization of cys and gln), which were excluded from the retention modelling. optimization of peptide retention prediction models the optimization of the peptide retention prediction models has been performed using the standard ssrcalc workflow [21,22]: 1) retention coefficients for individual amino acids were optimized to produce the best fit for experimental admet & dmpk 6(2) (2018) 190-199 peptide retention time prediction for iam.pc stationary phases doi: 10.5599/admet.520 195 vs. predicted retention values plot using an additive model with peptide length correction. 2) position-dependent retention coefficients have been introduced for four terminal positions from each terminus. 3) sequence-dependent corrections related to peptide helicity and presence of hydrophobic clusters were applied in an attempt to improve correlations. it should be noted, that modelling peptide helicity in reversed-phase separations still represents a major problem and has not been fully implemented in the ssrcalc model. in this work, we have used our helicity model developed for acidic c18 conditions and applied it directly to c18 at ph7.4. it's application to iam.pc.dd2 data (not shown here) did not a provide significant improvement. therefore, we applied its simplified version (counting i – i+3; i – i+4 interactions of hydrophobic residues) to the phospholipid phase data. figure 3. representation of the separation space of tryptic peptides on iam.pc.dd2 and c18, and their respective ssrcalc model accuracy. a – correlation between the retention times (fraction number) for the two chromatographic systems (26,594 peptides identified in both runs); b and c – the accuracy of the custom versions of the ssrcalc model for the luna c18(2) (40,105 peptides) and the iam.pc.dd2 (28,558), respectively. resulting correlation plots for luna c18(2) and iam.pc.dd2 are shown in figure 3b and 3c, respectively. the final model accuracy for the c18 packing material was found to be similar to other ssrcalc models for c18 separations (r 2 -value 0.96). it should be noted, that only peptide, with a retention of 3 min and higher were used for the iam.pc.dd2 model development. peptides with lower retention were considered unretained under the chromatographic conditions used and therefore their retention values could not be accurately assigned. therefore, out of 37,327 non-modified tryptic peptides identified only 28,558 were used for modelling as shown in figure 3c. the accuracy of the iam.pc.dd2 algorithm is lower than that for the c18 column due to a narrower range of peptide elution (~30 % acetonitrile vs. ~40 %) and the possible involvement of novel sequence-specific features of the retention on the phosphatidylcholine stationary phase, yet to be discovered. retention coefficients (rc) retention coefficients (rc) represent a measure of the participation of individual amino acids in the peptide retention on different chromatographic columns. ssrcalc models encode rc for individual residues in a position-dependent manner with four to five nand c-terminal rc’s, and internal rc’s. since tryptic peptides are fairly large, the latter represents the bulk of the residues and provide the most valuable information on the contribution of the residues. figure 4a shows the comparison of internal retention coefficients for c18 and iam.pc.dd2 separations at ph 7.4 and establishes the difference in retention contributions of amino acids. table 2 additionally compares these values to rp separations at acidic and basic conditions. when analyzing figure 4a, both hydrophobic character and charge state of the residues oleg krokhin et al. admet & dmpk 6(2) (2018) 190-199 196 should be taken into account. the amino acids usually considered to be hydrophobic are shown in red and hydrophilic in green. at ph 7.4 arg and lys are protonated, his is neutral, while asp and glu carry a negative charge. positively charged lys, arg and the aromatic tyr, his, and trp are among the residues, which showed an increase in interaction on iam.pc.dd2. at the same time negatively charged glu and asp exhibit reduced interaction. this suggests that additional electrostatic interactions with glycerophosphate groups play a major role in the separation through the attraction of lys and arg and repulsion of asp and glu. the increased interaction of aromatic residues on iam.pc.dd2 is most likely caused by the delocalization of the negative charge on the phosphate groups through the formation of a hydrogen bond. figure 4. retention contributions of individual residues in peptide retention on c18 and iam.pc.dd2 phases at ph 7.4. a – comparison of retention coefficients for the two columns; b and c – position-dependent retention coefficients for selected amino acids for luna c18(2) and iam.pc.dd2 columns, respectively. the contribution of the different amino acids to peptide retention is often found to be position dependent [17, 21, 22]. these effects occur due to various mechanisms such as ion-pairing at positively charged n-terminus [17] or the peptide orientation effect [22], characteristic in cation-exchange peptide separations. optimizing position dependent rc’s is a mandatory procedure for all ssrcalc models and was applied in this study. figure 4(b,c) shows nine position dependent rc values (four on each side plus internal) for selected residues: hydrophobic, negatively, and positively charged. iam.pc.dd2 does not exhibit significant position-dependent changes except for a small decrease of hydrophobic interactions from the nto c-terminus (figure 4c). respective plots for the c18 stationary phase show a substantial increase in the retention contribution for arg and lys and increased retention of the hydrophobic residues for the internal positions (figure 4b). comparing the retention contribution on the luna c18(2) and iam.pc.dd2 columns require understanding the differences in chemistry between the two stationary phases. iam.pc.dd2 is more hydrophilic because of its shorter aliphatic chain (c14 vs. c18), hydrophilic linkers and the presence of a zwitterionic head group. the positively charged choline group is located on the outside of the functional admet & dmpk 6(2) (2018) 190-199 peptide retention time prediction for iam.pc stationary phases doi: 10.5599/admet.520 197 layer, while the negatively charged glycerophosphate is positioned below it; separated by two additional methylene groups [4]. a peptide has to penetrate this zwitterionic bilayer to be partitioned into the hydrophobic environment. hydrophobic residues are found to have greater rc’s in figure 4a, suggesting that the majority of the separation is driven by hydrophobic interaction on both stationary phases. the differences in observed retention contributions between c18 and iam.pc.dd2 have to come from the interactions of the peptides with the zwitterionic head group. table 2. retention coefficients for ssrcalc peptide retention prediction models in different rp separation modes c18 ph 7.4 iam pc ph 7.4 c18 formic acid [20] c18 ph10 [20] w 10.73 12.75 35.74 13.34 f 9.89 9.66 32.99 11.33 l 8.52 8.09 30.57 10.52 i 7.56 7.00 27.75 9.20 m 6.31 6.34 21.32 8.23 y 5.15 7.80 16.69 5.75 v 5.00 4.58 18.38 6.28 k 2.54 8.07 -6.24 4.34 a 2.46 2.51 7.71 2.59 r 2.43 9.97 -1.66 5.35 t 1.62 1.65 4.25 1.93 p 1.35 1.07 5.36 1.26 h 1.32 3.68 -5.16 3.03 c 1.15 1.75 3.38 1.71 s 0.99 1.55 1.75 1.03 q 0.87 1.33 2.43 1.10 g 0.71 0.93 1.08 0.49 n 0.35 1.00 -0.07 0.55 e -2.36 -5.05 6.41 -5.00 d -2.54 -5.05 2.83 -6.04 the relative changes in retention for aromatic and charged residues suggest that the zwitterionic head group contributes to the separation mechanism. we observe that the negatively charged amino acids asp and glu have a decreased, and the positively charged arg and lys have an increased retention on the iam.pc.dd2 compared to c18 stationary phase. this suggests the involvement of electrostatic interactions (repulsion/attraction) with negatively charged glycerophosphate groups. considering the behaviour of aromatic amino acids (trp, tyr, his, and phe), all of which have larger rc values, except for phe, aromatic rings contain conjugated pi-systems, which possess electro-negative character. we have concluded that the reason these residues increase in retention is due to their interaction with the phosphate group. tyr, trp, and his all contain nitrogen or oxygen bonded to a hydrogen in their side chains connected to their aromatic rings, while phe does not. the partial positive charge of the hydrogen allows for a hydrogen bond to form between the side chain and the phosphate. this, in turn, delocalizes the negative charge of the phosphate across the entire aromatic ring. this is known to provide a stabilizing effect and thus would increase retention. the inability to form a hydrogen bond with phe is why phe behaves like the other hydrophobic residues and does not exhibit an increased contribution. the delocalization effect is further supported by the relatively large increase in retention of arg in comparison to lys. although lys and arg have nearly identical rc’s on the c18 phase, in the iam.pc.dd2 phase, the retention of arg is much greater than lys. this could be explained by the delocalized positive charge across the two amine groups interacting more strongly with the negatively charged phosphate than the single positively charged amine group of lys. overall, the major changes in retention between the columns can be explained by the interaction of the zwitterionic head groups in addition to the aliphatic chains that are similar to those on oleg krokhin et al. admet & dmpk 6(2) (2018) 190-199 198 the c18 stationary phase. conclusions a peptide retention prediction model for reversed-phase separation on immobilized artificial membrane phosphatidylcholine (iam.pc) stationary phase has been developed. chromatographic conditions for highthroughput measurements of peptide retention on an iam.pc.dd2 phase in reversed-phase separation mode at ph 7.4 have been established and compared to a standard c18 phase. iam.pc.dd2 was found to be more hydrophilic and to exhibit lower peptide retention (by ~14% acetonitrile on average, calculated for all s.cerevisiae peptides retained in both systems) compared to octadecyl-silica (c18). 2d lc-ms/ms analysis of a complex s. cerevisiae digest with iam.pc or c18 columns in the first dimension allowed the measurement of the retention properties of tens of thousands of peptides – sufficient for the confident assignment of retention coefficients and the development of a sequence-specific prediction algorithm. peptide retention on the iam.pc.dd2 phase is driven by hydrophobic interactions. however, we found a substantial increase in the relative retention contribution (compared to c18) for positively charged (arg and lys) and aromatic (tyr, trp and his) residues, and a decrease for negatively charged (asp and glu) residues compared to c18. this indicated the involvement of other types of interactions (electrostatic and electron delocalization), which results in a mixed-mode retention mechanism. due to the lower overall hydrophobicity of the iam.pc.dd2 stationary phase, the effect of amphipathic helicity on retention is less profound. at the same time, iam.pc.dd2 version of the ssrcalc algorithm showed a lower accuracy compared to c18 version, suggesting that additional sequence-specific features (yet to be discovered) play a role in peptide separation on the phosphatidylcholine stationary phase. acknowledgements: this work was supported by a grant from the natural sciences and engineering research council of canada (rgpin-2016-05963; o.v.k.). the authors also wish to thank scott anderson and melissa wilcox from regis technologies, inc. for providing the iam.pc.dd2 packing material. references [1] j.m. parker, d. guo, r.s. hodges. new hydrophilicity scale derived from high-performance liquid chromatography peptide retention data: correlation of predicted surface residues with antigenicity and x-ray-derived accessible sites. biochemistry 25 (1986) 5425-5432. [2] j. kyte, r.f. doolittle. a simple method for displaying the hydropathic character of a protein. j mol biol 157 (1982) 105-132. [3] k.l. valko. lipophilicity and biomimetic properties measured by hplc to support drug discovery. j pharm biomed anal 130 (2016) 35-54. [4] c. pidgeon, s. ong, h. liu, x. qiu, m. pidgeon, a.h. dantzig, j. munroe, w.j. hornback, j.s. kasher, l. glunz, et al. iam chromatography: an in vitro screen for predicting drug membrane permeability. j med chem 38 (1995) 590-594. [5] k.m. biswas, d.r. devido, j.g. dorsey. evaluation of methods for measuring amino acid hydrophobicities and interactions. j chromatogr a 1000 (2003) 637-655. [6] s.e. blondelle, e. perez-paya, g. allicotti, b. forood, r.a. houghten. peptide binding domains determined through chemical modification of the side-chain functional groups. biophys j 69 (1995) 604-611. [7] k. buttner, s.e. blondelle, j.m. ostresh, r.a. houghten. perturbation of peptide conformations induced in anisotropic environments. biopolymers 32 (1992) 575-583. [8] s.e. blondelle, j.m. ostresh, r.a. houghten, e. perez-paya. induced conformational states of amphipathic peptides in aqueous/lipid environments. biophys j 68 (1995) 351-359. admet & dmpk 6(2) (2018) 190-199 peptide retention time prediction for iam.pc stationary phases doi: 10.5599/admet.520 199 [9] t.j. sereda, c.t. mant, r.s. hodges. selectivity due to conformational differences between helical and non-helical peptides in reversed-phase chromatography. j chromatogr a 695 (1995) 205-221. [10] y. chen, c.t. mant, r.s. hodges. temperature selectivity effects in reversed-phase liquid chromatography due to conformation differences between helical and non-helical peptides. j chromatogr a 1010 (2003) 45-61. [11] z. jiang, b.j. kullberg, h. van der lee, a.i. vasil, j.d. hale, c.t. mant, r.e. hancock, m.l. vasil, m.g. netea, r.s. hodges. effects of hydrophobicity on the antifungal activity of alpha-helical antimicrobial peptides. chem biol drug des 72 (2008) 483-495. [12] j.l. meek. prediction of peptide retention times in high-pressure liquid chromatography on the basis of amino acid composition. proc natl acad sci u s a 77 (1980) 1632-1636. [13] d. guo, c.t. mant, a.k. taneja, j.m.r. parker, r.s. rodges. prediction of peptide retention times in reversed-phase high-performance liquid chromatography i. determination of retention coefficients of amino acid residues of model synthetic peptides. j chromatogr a 359 (1986) 499-518. [14] e.f. strittmatter, l.j. kangas, k. petritis, h.m. mottaz, g.a. anderson, y. shen, j.m. jacobs, d.g. camp, r.d. smith. application of peptide lc retention time information in a discriminant function for peptide identification by tandem mass spectrometry. j proteome res 3 (2004) 760-769. [15] r. kiyonami, a. schoen, a. prakash, s. peterman, v. zabrouskov, p. picotti, r. aebersold, a. huhmer, b. domon. increased selectivity, analytical precision, and throughput in targeted proteomics. mol cell proteomics 10 (2011) m110 002931. [16] v. spicer, p. ezzati, h. neustaeter, r.c. beavis, j.a. wilkins, o.v. krokhin. 3d hplc-ms with reversed-phase separation functionality in all three dimensions for large-scale bottom-up proteomics and peptide retention data collection. anal chem 88 (2016) 2847-2855. [17] o.v. krokhin, r. craig, v. spicer, w. ens, k.g. standing, r.c. beavis, j.a. wilkins. an improved model for prediction of retention times of tryptic peptides in ion pair reversed-phase hplc: its application to protein peptide mapping by off-line hplc-maldi ms. mol cell proteomics 3 (2004) 908-919. [18] k. petritis, l.j. kangas, b. yan, m.e. monroe, e.f. strittmatter, w.j. qian, j.n. adkins, r.j. moore, y. xu, m.s. lipton, d.g. camp, 2nd, r.d. smith. improved peptide elution time prediction for reversed-phase liquid chromatography-ms by incorporating peptide sequence information. anal chem 78 (2006) 5026-5039. [19] o.v. krokhin. sequence-specific retention calculator. algorithm for peptide retention prediction in ion-pair rp-hplc: application to 300and 100-a pore size c18 sorbents. anal chem 78 (2006) 77857795. [20] r.c. dwivedi, v. spicer, m. harder, m. antonovici, w. ens, k.g. standing, j.a. wilkins, o.v. krokhin. practical implementation of 2d hplc scheme with accurate peptide retention prediction in both dimensions for high-throughput bottom-up proteomics. anal chem 80 (2008) 7036-7042. [21] o.v. krokhin, p. ezzati, v. spicer. peptide retention time prediction in hydrophilic interaction liquid chromatography: data collection methods and features of additive and sequence-specific models. anal chem 89 (2017) 5526-5533. [22] d. gussakovsky, h. neustaeter, v. spicer, o.v. krokhin. sequence-specific model for peptide retention time prediction in strong cation exchange chromatography. anal chem 89 (2017) 11795-11802. [23] o.v. krokhin, g. anderson, v. spicer, l. sun, n.j. dovichi. predicting electrophoretic mobility of tryptic peptides for high-throughput cze-ms analysis. anal chem 89 (2017) 2000-2008. [24] o.v. krokhin, v. spicer. peptide retention standards and hydrophobicity indexes in reversed-phase high-performance liquid chromatography of peptides. anal chem 81 (2009) 9522-9530. ©2018 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ solubility-ph profiles of a free base and its salt: sibutramine as a case study 47 admet & dmpk 6(1) (2018) 47-54; doi: http://dx.doi.org/10.5599/admet.6.1.473 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper evaluation of the interactions between human serum albumin (hsa) and warfarin or diflunisal by using molecular fluorescence using two approaches susana amézqueta, anna maria bolioli, josé luis beltrán, clara ràfols* departament d’enginyeria química i química analítica and institute of biomedicine of the university of barcelona (ibub), universitat de barcelona, martí i franquès 1-11, 08028, barcelona, spain *corresponding author: e-mail: crafols@ub.edu; tel.: +34-934-034-874; fax: +34-934-021-233 received: november 03, 2017; revised: january 12, 2018; published: march 25, 2018 abstract serum albumin is the main drug transporter of the bloodstream and contains two main binding sites: sudlow i or acidic drug binding site, and sudlow ii or benzodiazepine binding site. warfarin, a well-known anticoagulant drug commonly used in the prevention of thrombosis and thromboembolism, binds to sudlow i site, whereas non-steroidal antiinflammatory drugs (nsaids) such as diflunisal bind preferentially to sudlow ii site. albumin is a fluorophore that modifies its fluorescence (quenching or enhancement effect) when it is bound to a drug. the application of the double logarithm stern -volmer equation allows the calculation of the stoichiometry and the binding constant of the process. this procedure does not consider the possible interferences coming from the fluorescence of the drug though. another strategy to evaluate the binding constants is to consider the whole spectrum, taking into account all the possible species in equilibrium; in this case we have used an extended version of the star program, which can deal with 300 spectra, each containing up to 300 data points. the aim of this work is to compare both approaches to evaluate the interaction between warfarin (sudlow i) and diflunisal (sudlow ii) and hsa: the double logarithm stern-volmer equation and the star program. keywords has; protein-albumin interaction; fluorescence; warfarin; nsaids introduction albumin, the most abundant protein in plasma and serum, is a water-soluble macromolecule with high biological significance. it is able to maintain the plasma oncotic pressure and modulate the fluid distribution among body compartments. native albumin, that is the one without ligands or bound molecules, is built up from three homologous domains (i, ii and iii), showing each one two distinct subdomains, named a and b. there are numerous binding sites on albumin, but drugs and other exogenous compounds bind, mainly, to two of them: sudlow i or acidic drug binding site, placed on subdomain iia, and sudlow ii or benzodiazepine binding site, located on subdomain iiia. thus, albumin plays a relevant role in binding and transport functions and, therefore, in the pharmacokinetics of drugs [1]. drug-albumin interactions can be evaluated using several complementary techniques such as http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:crafols@ub.edu clara ràfols et al. admet & dmpk 6(1) (2018) 47-54 48 isothermal titration calorimetry, fluorescence, frontal analysis capillary electrophoresis or equilibrium dialysis [2,3]. in the case of fluorescence, the albumin is considered as the main fluorophore, as it contains three fluorescent amino acids (tryptophan, tyrosine and phenylalanine). when the complex drug-protein is formed, the albumin fluorescence can be either quenched or enhanced. there are some equations to evaluate the fluorescence quenching/enhancement that allow calculating the binding parameters (binding constant and stoichiometry) [4]. these equations present several drawbacks, as they consider the albumin as the unique fluorophore, and assume that only one type of interaction is formed and that the concentration of the free drug is much higher than the bound fraction. however, there often exist other fluorophores in solution (such as the drug or a drug-albumin complex) that force to work under more selective but less sensitive conditions. another strategy to evaluate the binding constants is to consider the whole fluorescence spectrum, taking into account all the possible species in equilibrium. the aim of this work is to evaluate the interaction between warfarin, interacting into sudlow i, or diflunisal, interacting into sudlow ii with human serum albumin (hsa) by fluorescence, using both approaches to calculate the binding parameters and compare the results obtained. experimental hsa (99 %), warfarin (>99 %), and diflunisal (>98 %) were obtained from sigma-aldrich (st louis, mo, united states). hydrochloric acid (0.5 m), sodium dihydrogen phosphate monohydrate (>99 %), disodium hydrogen phosphate (>99 %), sodium chloride (>99 %), and potassium chloride (>99 %) were from merck (darmstadt, germany). water was purified by a milli-q plus system from millipore (bedford, ma, united states), with a resistivity of 18.2 mω cm. pbs (phosphate buffer solution) (10 mm, i=150 mm) consisted of an aqueous solution containing 0.137 m nacl, 0.0027 m kcl, 0.0015 mm kh2po4·3h2o and 0.0081 m na2hpo4·h2o. the solution ph was adjusted to 7.40 with 0.5 m hcl. drugs and albumin were dissolved in pbs at the desired concentration (300 1000 µm and 3 µm, respectively). quenching studies were carried out by titration over a hsa solution (2.7 ml) with an initial concentration of 3 µm. the albumin was placed in a fluorescence cuvette at 25 °c. the initial fluorescence spectrum was recorded. sixteen successive additions of the drug under study (warfarin, or diflunisal) were done to cover a drug concentration in the range 0.4 – 14 µm. after each addition, the solution in the cuvette was stirred to allow the temperature and reaction equilibration. then, the fluorescence spectrum was acquired using a cary eclipse fluorimeter from agilent technologies (santa clara, ma, united states) [4] using 5 nm slits in both, excitation and emission monochromators. these protein and drug concentrations allow measuring the specific binding of the drug to the protein and do not correspond to the plasmatic ones. in fact, the usual albumin concentration in plasma is about 600 µm and the drug:albumin ratio is usually as low as 1:150 for warfarin (30 mg/week) [5] and 1:2 for diflunisal (500 mg, single dose) [6]. data treatment fluorescence data were recorded in three ways: a standard fluorescence spectrum, and as synchronous fluorescence spectra at  = 15 nm and  = 60 nm for the selective measurement of the fluorophores tyrosine and tryptophan, respectively. to evaluate the stoichiometry and the binding constant, the double logarithm stern-volmer equation (dlsv) was considered in first instance (eq. 1) [4]. admet & dmpk 6(1) (2018) 47-54 warfarin-hsa and diflunisal-hsa interactions doi: 10.5599/admet.6.1.473 49  0 b flog log logh f f k n q f    (1) where f0 and f are the fluorescence intensity in the absence and the presence of a quencher, respectively; kb is the binding constant; nh is the hill coefficient; and [qf] is the free quencher (drug) concentration. usually, the free drug concentration [qf] is unknown and it is replaced by the total drug concentration [q] (it is assumed that the concentration of the macromolecule is negligible with respect to the drug one) (eq. 2). moreover, this equation only yields the binding stoichiometry in case of “infinite” cooperativity [7].  0 blog log logh f f k n q f    (2) the substitution of [qf] by [q] indicates that the dlsv equation should be applied only when the solution contains a large excess of drug. on the other hand, it is also assumed that the fluorescence intensity of the drug is negligible. therefore, the applicability of the dlsv equation is restricted to these scenarios. in order to overcome both problems we have applied an extended version of the non-linear least squares program star [8], which can handle the spectroscopic data from multiple-equilibria systems containing up to 300 solutions, measured at 300 data points. in this case, an equilibrium system can be described by a set of components (as albumin, quencher, hydrogen ion, etc.) which can form different species. these species are defined by their stoichiometric coefficients and the corresponding binding constants: s qss + qq s q      b s q s qs q k s q      . (3) star uses a chemical equilibrium model, which consists of the components of the system (e.g., hsa and drug), together with the guessed stoichiometry and binding constants of the species formed. moreover, the experimental data for each solution contains the total concentrations of components and the measured fluorescence spectrum. on the other hand, we can consider that the fluorescence spectrum of a given solution is equal to the sum of the fluorescence spectra of all species. if we have a set of n fluorescent species, star solves the mass balances of the components calculating the concentrations for all species for the given set of binding constants. then, the fluorescence intensity at a j wavelength can be calculated as: j i i,j 1 n i if c   , (4) where ci indicates the concentration of the i-species, and i,j the relative fluorescence of the i-species at the j-wavelength. the i,j values can be supplied as data input, or calculated by the program. the binding constants are refined by the gauss-newton iterative algorithm, until a minimum of the sum of squared differences (u) between the experimental and calculated fluorescence intensities is obtained:   2 i,j,exp , , 1 1 ns nw i j calc i i u if if      , (5) where ns and nw indicate the number of solutions and wavelengths, respectively; the subindexes exp and calc indicate experimental values and calculated by the program. the program output includes the standard deviation of the errors between experimental and calculated clara ràfols et al. admet & dmpk 6(1) (2018) 47-54 50 data, together with the estimated error in the binding constants. star allows the data treatment of the full fluorescence spectra (as it can handle up to 300 solutions, measured up to 300 data points in each spectrum), calculating the binding constants and the unitary spectrum of the species formed. in this way, the drawbacks owed to the application of the dlsv equation are easily overcome. results and discussion when a successive amount of the studied drugs is added to a fixed concentration of hsa, the albumin fluorescence is quenched; this is the case of the hsa fluorescence band at about 285 nm. as an example, figure 1 shows the changes in the synchronous fluorescence spectra (60 nm) for hsa after the addition of diflunisal. figure 1. quenching study of diflunisal-hsa at synchronous 60 nm the dlsv approach uses the fluorescence intensity value of the fluorophore at a fixed wavelength, usually the one that offers the maximum intensity. as the drugs used in the study may be fluorescent in the working conditions, an initial study of interferences has been performed in the three modes selected. hsa shows maximum fluorescence in the emission mode at 340-350 nm, and in the synchronous modes at 285288 nm. at these regions, diflunisal does not interfere with hsa, while warfarin (figure 2) interferes about 9 % in the emission mode, over 10% in synchronous  60 nm mode and does not interfere in synchronous  15 nm mode. although in some cases interferences exist, we have selected the wavelength with maximum fluorescence intensity for each drug and mode to evaluate the results. we assume that the interferences under 10% can be evaluated using the dlsv equation. the working conditions are indicated in table 1. first, the fluorescence data have been evaluated with the dlsv equation considering the range of linear response (0.4-5 or 0.4-14 m, depending on the drug and the fluorescence mode). the results obtained show that diflunisal and warfarin interact with hsa with a log kb value about 5 (table 2). for diflunisal, not significant differences in the log kb values have been observed when evaluating the data at the three modes of acquisition. in the case of warfarin, the log kb calculated considering the emission mode (where warfarin fluorescence interferes about 9 %) is slightly lower than the one obtained taking into account the synchronous =15 nm mode (without interferences). admet & dmpk 6(1) (2018) 47-54 warfarin-hsa and diflunisal-hsa interactions doi: 10.5599/admet.6.1.473 51 figure 2. comparison of fluorescence spectra in the study of warfarin-hsa: (a) emission, (b) synchronous 15 nm and (c) synchronous 60 nm modes table 1. optimum wavelengths found for the evaluation of the drug-hsa interaction for the three modes of interest. warfarin diflunisal emission, λ𝑒𝑥=285 nm 350 nm 348 nm synchronous ∆λ 15 nm 285 nm 286 nm synchronous ∆λ 60 nm 288 nm the binding constants have been also evaluated with the modified star program. the results are also shown in table 2. in the case of diflunisal, which does not present interferences at the working conditions, the results obtained with this program (restricting the model to a 1:1 binding equilibrium) are similar to the ones obtained by dlsv equation and are similar at the three modes evaluated. for warfarin, the results calculated with the star program agree with those obtained with the dlsv approach when there are not interferences (synchronous  15 nm). when interferences exist (emission mode), the star program is able to deal with them and offers log kb values similar to those obtained in the absence of interferences. moreover, the use of star program has allowed detecting the presence of two interactions between hsa and diflunisal when it is assumed the possible presence of more than one kind of interaction (table 2, last row). in fact, literature reports that diflunisal can bind to two different sites with different affinity [9]. for this drug, the first log kb value is slightly higher than the one obtained by dlsv equation, but when the equilibrium model in star contains only one interaction, the results agree with those obtained by dlsv equation. in this case, the calculated value by star for log kb is intermediate between log kb1 and log kb2, in order to accomplish the mass balances of the components. therefore, when multiple species are involved in the binding process, the dlsv approach allows calculating the predominant event, whereas the star program allows the study of all species in equilibrium. a) b) c) clara ràfols et al. admet & dmpk 6(1) (2018) 000-000 52 table 2. stoichiometry and binding constants obtained using two different approaches for data treatment (dlsv equation and star program) at three different fluorescence modes. emission synchronous  15 nm synchronous  60 nm n1 log kb1 n2 log kb2 n1 log kb1 n2 log kb2 n1 log kb1 n2 log kb2 hsawarfarin dlsv 0.98(0.02) 4.6(0.2) a 1.05(0.07) 5.0(0.3) a star 1 5.14(0.01) 1 4.97(0.02) hsadiflunisal dlsv 1.01(0.02) 5.2(0.1) a 1.09(0.02) 5.1(0.1) a 1.04(0.02) 5.3(0.1) a star b 1 5.18(0.01) 1 4.90(0.02) 1 5.34(0.01) star 1 5.8(0.2) 2 4.7(0.1) 1 5.44(0.04) 2 4.6(0.1) 1 5.7(0.1) 2 4.5(0.2) a : determination coefficient (r 2 ) ~ 0.98; b : the model is restricted to 1:1 binding equilibria n1 log kb1 n2 log kb2 ph t ( °c) buffer (concentration) ref. warfarin 1 6.15 7.3 25 pbs (5 mm) [10] 0.7 5.93 7.3 25 pbs (5 mm, i=50 mm) [10] 0.5 5.65 7.3 25 pbs (5 mm, i=200 mm) [10] 0.88 5.56 27 pbs (10 mm) + nacl (0.9%) [11] 5.38 7.4 25 pbs (9.5 mm, i=150 mm) [12] 5.36 25 pbs (50 mm, i=100 mm) [13] 5.45 4.15 7.4 25 pbs (67 mm) [14] 1.38 5.52 7.4 25 pbs (67 mm) [15] diflunisal 0.53 5.09 7 25 15 mm trisodium citrate (i=150 mm) [16] table 3. stoichiometry and binding constants reported in the literature for warfarin or diflunisal interactions with hsa. admet & dmpk 6(1) (2018) 000-000 warfarin-hsa and diflunisal-hsa interactions doi: 10.5599/admet.6.1.473 53 compared to the results reported by other authors (table 3) using fluorescence and different experimental conditions, the stoichiometry agrees with the reported data (n in the range 0.5-1.38 m). the log kb values obtained in the present work are slightly lower than those obtained by other authors. note that higher ionic strength implies lower log kb values and therefore the experimental conditions used can be meaningful. in the case of diflunisal, the results reported by davilas et al. [17] agree with the ones obtained in the present work when a unique species is considered. another work on diflunisal interaction with hsa [17], using the internal standard electrode method, reports two interactions (log kb1=5.7, log kb2=3.8) and agrees with the results obtained with the star program considering the presence of several species. conclusions hsa interaction with warfarin and diflunisal has been evaluated using fluorescence measurements. first, an interference study has been carried out showing that warfarin interferes under the working conditions, while diflunisal does not. next, the drug-albumin interaction has been studied using two different approaches. these drugs bind to the albumin with a kb in the order of 10 5 . in the absence of interferences or when a unique species is considered (stoichiometry 1:1), both approaches yield similar values. the modified star program is also able to deal with the interferences of fluorescent drugs and can consider multiple species. in the case of warfarin, the log kb values calculated agree in the absence and presence of interferences, and in the case of diflunisal it shows that there are two different processes involved, with stoichiometry 1:1 and 1:2. acknowledgements: we thank spanish government for the research grant (project ctq2014-56253-p). references [1] k. yamasaki, v.t.g. chuang, t. maruyama, m. otagiri. albumin–drug interaction and its clinical implication. biochimica et biophysica acta 1830 (2013) 5435-5443. [2] s.e. harding, b.z. chowdhry, protein-ligand interactions: hydrodynamics and calorimetry, oxford university press: new york 2011. [3] c. ràfols, s. zarza, e. bosch. molecular interactions between some non-steroidal anti-inflammatory drugs (nsaid׳ s) and bovine (bsa) or human (hsa) serum albumin estimated by means of isothermal titration calorimetry (itc) and frontal analysis capillary electrophoresis (fa/ce). talanta 130 (2014) 241-250. [4] j. yang, l. qu, w. y, y. huang, n. jiao, w. zhan, d. zhao, l. cui. interaction of hyperoside with human serum albumin and effect of glucose on the binding. journal of spectroscopy 386586 (2014) 1-9. [5] t. lomonaco, s. ghimenti, i. piga, d. biagini, m. onor, r. fuoco, a. paolicchi, l. ruocco, g. pellegrini, m.g. trivella, f. di francesco. monitoring of warfarin therapy: preliminary results from a longitudinal pilot study. microchem. j. 136 (2017) 170–176. [6] d.s. patel, n. sharma, m.c. patel, b.n. patel, p.s. shrivastav, m. sanyal. sensitive and selective determination of diflunisal in human plasma by lc – ms. j. chromatogr. sci. 51 (2013) 872–882. [7] g. li, b.s. liu, q. zhang, r. han. investigation on the effect of fluorescence quenching of bovine serum albumin by cefoxitin sodium using fluorescence spectroscopy and synchronous fluorescence spectroscopy. luminiscence 31 (2016) 1054-1062. [8] j.l. beltrán, r. codony, m.d. prat. evaluation of stability constants from multi-wavelength absorbance data: program star. analytica chimica acta 276 (1993) 441-454. [9] j. ghuman, p.a. zunszain, i. petitpas, a.a. bhattacharya, m. otagiri, s. curry. structural basis of the drug-binding specificity of human serum albumin. journal of molecular biology 353 (2005) 38-52. clara ràfols et al. admet & dmpk 6(1) (2018) 000-000 54 [10] p. bolel, s. datta, n. mahapatra, m. halder. exploration of ph-dependent behavior of the anion receptor pocket of subdomain iia of hsa: determination of effective pocket charge using the debye– hückel limiting law. journal of physical chemistry b 118 (2014) 26-36. [11] v. maes, y. engelborghs, j. hoebeke, y. maras, a. vercruysse. fluorimetric analysis of the binding of warfarin to human serum albumin. equilibrium and kinetic study. molecular pharmacology 21 (1982) 100-107. [12] m. poór, y. li, s. kunsági-máté, j. petrik, s. vladimir-knežević, t. kőszegi, t. koszegi. molecular displacement of warfarin from human serum albumin by flavonoid aglycones. journal of luminiscence 142 (2013) 122–127. [13] c. dufour, o. dangles. flavonoid-serum albumin complexation: determination of binding constants and binding sites by fluorescence spectroscopy. biochimica biophysica acta gen. subj. 1721 (2005) 164-173. [14] m. dockal, m. chang, d.c. carter, f. rüker. five recombinant fragments of human serum albumintools for the characterization of the warfarin binding site. protein science 9 (2000) 1455-1465. [15] r. abou-khalil, a. jraij, j. magdalou, n. ouaini, d. tome, h. greige-gerges. interaction of cucurbitacins with human serum albumin: thermodynamic characteristics and influence on the binding of site specific ligands. journal of photochemistry photobiology b 95 (2009) 189–195. [16] s. fountoulaki, f. perdih, i. turel, d.p. kessissoglou, g. psomas. non-steroidal anti-inflammatory drug diflunisal interacting with cu(ii). structure and biological features. journal of inorganic biochemistry 105 (2011) 1645–1655. [17] a. davilas, m. koupparis, p. macheras, g. valsami. in-vitro study on the competitive binding of diflunisal and uraemic toxins to serum albumin and human plasma using a potentiometric ion-probe technique. journal of pharmacy and pharmacology 58 (2006) 1467–1474. ©2018 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ bioanalysis of aminoglycosides using high-performance liquid chromatography doi: http://dx.doi.org/10.5599/admet.1183 27 admet & dmpk 10(1) (2022) 27-62; doi: https://doi.org/10.5599/admet.1183 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index review bioanalysis of aminoglycosides using high-performance liquid chromatography seth k. amponsah 1 , joseph a. boadu 2 , daniel k. dwamena 1 , kwabena f. m. opuni 2,* 1 department of medical pharmacology, university of ghana medical school, university of ghana, ghana 2 department of pharmaceutical chemistry, school of pharmacy, university of ghana, ghana *corresponding author: e-mail: kfopuni@ug.edu.gh; tel.: +233208260595 received: november 19, 2021; revised: december 30, 2021; published: january 11, 2022 abstract aminoglycosides are broad-spectrum antibiotics used in the treatment of gram-negative bacterial infections. due to their nephrotoxic and ototoxic potential (narrow therapeutic index), the use of aminoglycoside for clinical indications requires monitoring. the objective of this review was to identify relevant literature reporting liquid chromatographic methods for the bioanalysis of aminoglycosides in both preclinical and clinical settings/experiments. data on liquid chromatographic methods were collected from articles in an online academic database (pubmed, science direct, scopus, and google scholar). all 71 articles published from 1977 to 2020 were included in the review. reversed-phase liquid chromatography was the most used method for the bioanalysis of aminoglycosides. fluorescence or ultraviolet detection methods were mostly used from 1977 to 2002 (51 articles), while mass spectrometry was predominantly used as a detector from 2003 to 2020 (15 articles). sixty-seven articles reported calibration ranges, which varied significantly for the various drugs assayed: some in the range of 0.1-0.5 ng/ml and others 1250200000 ng/ml. also, 61 articles reported r 2 values (0.964-1.0) for almost all analytes under consideration. sixty-three articles reported percent recoveries mostly between 61.0 % to 114.0 %, with only two articles reporting recoveries of 4.9 % and 36 %. out of the 71 reviewed articles, 56 reported intermediate precision values ranging between 0.331 % to 19.76 %, which is within the acceptable limit of 20 %. this review will serve as a guide for research and/or routine clinical monitoring of aminoglycosides in biological matrices. ©2022 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords pk studies; therapeutic drug monitoring; biological matrices. introduction aminoglycosides are broad-spectrum antibiotics that are used in the treatment of gram-negative bacterial infections [1,2]. aminoglycosides elicit their pharmacological effect by binding to the 16s rrna ribosomal subunit of bacteria and blocking mrna translation, altering protein synthesis. structurally, this class of antibiotics has amino sugars in their core connected via glycosidic linkages to a dibasic aminocyclitol [3]. aminoglycosides are relatively hydrophilic, hence, rarely undergo biotransformation in vivo. aminoglycosides, to some extent, bind to plasma proteins and are excreted entirely unchanged in urine [4]. streptomycin, netilmicin, tobramycin, kanamycin, spectinomycin, gentamicin, neomycin, amikacin, and http://dx.doi.org/10.5599/admet.1183 https://doi.org/10.5599/admet.1183 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:kfopuni@ug.edu.gh http://creativecommons.org/licenses/by/4.0/ seth k. amponsah et al. admet & dmpk 10(1) (2022) 27-62 28 paromomycin are examples of aminoglycosides [5,6]. aminoglycosides are known to possess nephrotoxic and ototoxic potentials, limiting their clinical use [2]. in addition, due to their narrow therapeutic index, therapeutic monitoring is required for aminoglycosides. especially among patients with underlying renal problems, monitoring aminoglycosides in biological matrices ensures optimal therapy and reduces toxicity [7]. there are several validated bioanalytical methods used to quantitatively determine levels of aminoglycosides in biological matrices. microbiological assays, radioimmunoassay (ria), radioenzymatic assays, fluorescence polarization immunoassay (fpia), high-performance liquid chromatography (hplc), gas chromatography (gc) and mass spectrometric techniques are some common examples [8-12]. of these methods, hplc is the most preferred or routinely used [13]. indeed, several studies have used hplc for the bioanalysis of aminoglycosides [14-16]. although there are recently published reviews on aminoglycosides, the focus of these reviews have been a) pre-treatment and analysis methods of aminoglycosides in food [17], b) challenges in the development of analytical test procedures for aminoglycosides [18], and c) determination of kanamycin by high-performance liquid chromatography [19]. the current review focuses on liquid chromatographic methods employed for the bioanalysis of aminoglycosides. a total of 20 aminoglycosides were reported by 71 articles (total number reviewed). the aminoglycosides were gentamicin [20], netilmicin [21], amikacin [22], tobramycin [23], dibekacin [24], sisomicin [25], astromicin [25], micronomicin [25], kanamycin [26], streptomycin [27], neomycin [14], isepamicin [28], geneticin [29], dihydrostreptomycin [30], paromomycin [31], apramycin [5], hygromycin [5], etimicin [32], arbekacin [33], and spectinomycin [34]. the aforementioned aminoglycosides obtained from either natural products or semi-synthetic derivatives of soil actinomycetes notably streptomyces have suffix -mycin (examples are streptomycin, dihydrostreptomycin, kanamycin, apramycin, paromomycin, neomycin, tobramycin, spectinomycin, and hygromycin); and those obtained from other actinomycetes notably micromonospora have the suffix -micin (examples are gentamicin, netilmicin, isepamicin, sisomicin, etimicin, geneticin, astromicin, and micronomicin). there are other exceptions, such as amikacin, arbekacin, and dibekacin. the constitute structures of these 20 aminoglycosides are presented in figure 1. in the review, gentamicin was the most reported aminoglycoside (24 articles). this is not surprising since gentamicin is often used clinically because of its low cost and high efficacy against gram-negative aerobes. fifty (50) articles reported bioanalysis of at least one of the 20 aminoglycosides, whilst 21 articles reported analysis of more than one aminoglycoside. the highest number of aminoglycosides simultaneously assayed was 13 [5]. although the list of articles used in this review may not be exhaustive, suitable liquid chromatographic conditions used in assaying aminoglycosides in biological matrices have been identified. performance metrics of the various liquid chromatographic assays have also been appraised. also, highlights of current procedures, scope, characteristics, and limitations of chromatographic methods used in assaying aminoglycosides in biological matrices have been provided. although this is not a systematic review, it will serve as a comprehensive reference for subsequent related research that may involve the assay of aminoglycosides. methods this study reviewed relevant and accessible articles on liquid chromatographic assays of aminoglycosides in biological matrices from 1977 to 2020. articles were retrieved from journals in online academic databases (pubmed, science direct, scopus, and google scholar) and limited to only the english admet & dmpk 10(1) (2022) 27-62 bioanalysis of aminoglycozides by hplc doi: http://dx.doi.org/10.5599/admet.1183 29 language. keywords used during the search were aminoglycosides, assay, hplc, plasma, serum, milk, cerebrospinal fluid, and urine. the searched terms used were ‘‘chromatographic assay’’, ‘‘aminoglycosides’’ and ‘‘biological matrix’’. articles were excluded if they were not pertinent. figure 1. structures of the 20 aminoglycosides reported by the various articles reviewed aminoglycosides sixty-seven (67) out of the 71 articles reported the use of hplc in the bioanalysis of aminoglycosides, while four articles used ultra-performance liquid chromatography (uplc) [35]. the relevant aspect of liquid chromatographic conditions used in the various articles, such as matrix, sample preparation, flow rate, column selection, mobile phase, and detection, have been summarized (table 1). streptomycin dihydrostreptomycin kanamycin apramycin paromomycin neomycin tobramycin spectinomycin hygromycin gentamicin amikacin netilmicin isepamicin sisomicin geneticin arbekacin etimicindibekacin astromicin micronomicin http://dx.doi.org/10.5599/admet.1183 seth k. amponsah et al. admet & dmpk 10(1) (2022) 27-62 30 table 1. hplc conditions for bioanalysis of aminoglycosides. d e te ct io n m o d e f lu o re sc e n ce (e xc it a ti o n , 3 4 0 n m ; e m is si o n , k v 4 1 8 f il te r) f lu o re sc e n ce (e xc it a ti o n , 3 6 0 n m ; e m is si o n , 4 3 0 n m ) f lu o re sc e n ce (e xc it a ti o n , 2 2 0 n m ; e m is si o n , k v 4 7 0 f il te r) f lu o re sc e n ce (e xc it a ti o n , 3 4 0 n m ; e m is si o n , 4 1 8 n m ) f lu o re sc e n ce (e xc it a ti o n fi lt e rs , 7 -5 4 /7 6 0 ; e m is si o n fi lt e rs , 4 -7 6 /3 7 2 ) d e ri v a ti za ti o n a g e n t o -p h th a la ld e h y d e o -p h th a la ld e h y d e d a n sy l ch lo ri d e o -p h th a la ld e h y d e o -p h th a la ld e h y d e m o b il e p h a se ; e lu ti o n ; f lo w r a te ( m l/ m in ) 0 .2 m n a 2 s o 4 , 0 .0 2 m s o d iu m p e n ta n e s u lf o n a te , a n d 0 .1 % (v /v ) a ce ti c a ci d in a w a te r/ m e th a n o l (9 7 :3 , v /v ); i s o ; 2 .0 m e th a n o l/ w a te r (7 9 :2 1 , v /v ) co n ta in in g 2 g /l o f tr ip o ta ss iu m e d t a ; is o ; 2 .0 a ce to n it ri le /w a te r (9 5 :5 , v /v ); i s o ; 1 .0 0 .2 m o l o f n a 2 s o 4 , 0 .0 2 m o l o f so d iu m p e n ta n e su lf o n a te , a n d 1 7 .4 m m o l o f a ce ti c a ci d /m e th a n o l (9 7 :3 , v /v ); i s o ; 2 .0 0 .1 m o l o f n a 2 s o 4 , 0 .0 2 m o l o f so d iu m p e n ta n e su lf o n a te , a n d 1 7 .4 m m o l o f a ce ti c a ci d ; is o ; 2 .0 0 .1 m o l o f n a 2 s o 4 , 0 .0 2 m o l o f so d iu m p e n ta n e su lf o n a te , a n d 1 7 .4 m m o l o f a ce ti c a ci d ; is o ; 2 .0 0 .5 m o l/ l t ri s b u ff e r (p h 7 .9 )/ tr im e th y la m in e , re a d ju st e d p h 7 .9 w it h co n ce n tr a te d su lp h u ri c a ci d /m e th a n o l (2 5 0 :1 0 :7 4 0 , v /v /v ); i s o ; 2 .0 0 .5 m o l/ l t ri s b u ff e r (p h 7 .9 )/ tr im e th y la m in e , re a d ju st e d p h 7 .9 w it h co n ce n tr a te d su lp h u ri c a ci d /m e th a n o l (2 5 0 :1 0 :7 4 0 , v /v /v ); i s o ; 2 .0 m e th a n o l/ tr ip o ta ss iu m e th y le n e d in it ri lo te tr a a ce ta te ( 2 g /l ) (7 9 :2 1 , v /v ); i s o ; 2 .0 s ta ti o n a ry p h a se c 1 8 c 1 8 c 1 8 (a m b ie n t) c 1 8 c 1 8 s a m p le p re p a ra ti o n io n -e xc h a n g e g e l ch ro m a to g ra p h y (s p e ) s p e p p t ( a c n ) s p e p p t ( a c n ) m a tr ix h -s e ru m h -s e ru m d -s e ru m h -p la sm a h -s e ru m h -s e ru m in te rn a l s ta n d a rd n .i . n .i . n .i . n -a ce ty l g e n ta m ic in c 1 t o b ra m y ci n a m ik a ci n n .i . a n a ly te g e n ta m ic in g e n ta m ic in c 1 g e n ta m ic in c 1 a g e n ta m ic in c 2 n e ti lm ic in g e n ta m ic in a m ik a ci n t o b ra m y ci n n e ti lm ic in t o b ra m y ci n g e n ta m ic in r e f [2 0 ] [3 6 ] [2 1 ] [2 2 ] [2 3 ] admet & dmpk 10(1) (2022) 27-62 bioanalysis of aminoglycozides by hplc doi: http://dx.doi.org/10.5599/admet.1183 31 table 1. contined... d e te ct io n m o d e f lu o re sc e n ce (e xc it a ti o n , 3 4 0 n m ; e m is si o n , 4 1 8 n m ) u v ( 3 6 5 n m ) u v ( 2 3 0 n m ) f lu o re sc e n ce (e xc it a ti o n , 2 7 5 n m ; e m is si o n , 4 1 8 n m ) u v ( 3 6 0 n m ) f lu o re sc e n ce (e xc it a ti o n , 3 4 0 n m ; e m is si o n , 4 1 8 n m ) f lu o re sce n ce (e xcita tio n , 3 6 5 n m ; e m issio n , 4 4 0 n m ) f lu o re sc e n ce (e xc it a ti o n , 3 4 0 n m ; e m is si o n , 4 5 5 n m ) d e ri v a ti za ti o n a g e n t o -p h th a la ld e h y d e 1 -f lu o ro -2 ,4 d in it ro b e n ze n e b e n ze n e s u lp h o n y l ch lo ri d e f lu o re sc a m in e 1 -f lu o ro -2 ,4 d in it ro b e n ze n e o p h th a li cd ic a rb o xa ld e h y d e o -p h th a la ld e h y d e o -p h th a la ld e h y d e m o b il e p h a se ; e lu ti o n ; f lo w r a te ( m l/ m in ) m e th a n o l/ w a te r/ a ce to n it ri le ( 6 2 :3 5 .1 :2 .9 , v /v /v ) co n ta in in g 2 .5 g t ri p o ta ss iu m e th y le n e d ia m in e te tr a a ce ti c a ci d ; is o ; 1 .6 1 g /l t ri s( h y d ro xy m e th y l) a m in o m e th a n e a d ju st e d w it h h y d ro ch lo ri c a ci d t o p h 7 /a ce to n it ri le ( 3 0 :7 0 , v /v ); i s o ; 1 .5 a ce to n it ri le /m e th y le n e c h lo ri d e /w a te r/ m e th a n o l (8 0 :1 0 :8 :4 , v /v /v /v ); i s o ; 4 .0 a ce to n it ri le /p h o sp h o ri c a ci d ( 5 g /l ) (7 0 :3 0 ); i s o ; 2 .0 a ce to n it ri le /w a te r (6 8 :3 2 , v /v ); i s o ; 1 .0 a ce to n it ri le /t ri s( h y d ro xy m e th y l) a m in o m e th a n e ( 1 g /l ) a d ju st e d t o p h 3 .0 w it h 1 m h y d ro ch lo ri c a ci d (7 0 :3 0 , v /v ); i s o ; 1 .5 0 .1 m d is o d iu m 1 ,2 e th a n e d is u lf o n a te a n d 0 .0 0 5 m so d iu m o ct a n e su lf o n a te a d ju st e d t o a p h 3 .5 w it h a ce ti c a ci d /a ce to n it ri le ( 8 5 :1 5 , v /v ); i s o ; 0 .8 m e th a n o l/ w a te r/ e th y le n e d ia m in e te tr a -a ce ti c a ci d p h 7 .2 ( 8 0 :1 5 :5 , v /v /v ); i s o ; 1 .0 s ta ti o n a ry p h a se c n ( a m b ie n t) c 1 8 ( r t ) c 1 8 c a ti o n e xc h a n g e co lu m n s il ic a ; c 1 8 (6 0 -1 0 0 ° c ) c 8 ( 2 5 ° c ) c 1 8 ( 5 0 ° c ) c 1 8 ( 2 2 ° c ) s a m p le p re p a ra ti o n p p t ( a c n ) p p t ( a c n ) p p t ( a c n ) p p t ( a c n ) p p t ( m e o h ) p p t ( a c n ) s p e p p t s p e m a tr ix h -s e ru m h -u ri n e h -s e ru m h -s e ru m rp la sm a rb -u ri n e g -p la sm a h -p la sm a h -s e ru m h -u ri n e h -s e ru m h -s e ru m in te rn a l s ta n d a rd g e n ta m ic in c 2 n .i . n e ti lm ic in n .i . n .i . n e ti lm ic in n .i . t o b ra m y ci n t o b ra m y ci n a n a ly te t o b ra m y ci n g e n ta m ic in c 1 a g e n ta m ic in c 1 + c 2 g e n ta m ic in g e n ta m ic in a m ik a ci n g e n ta m ic in ( c 1 , c 1 a , a n d c 2 ) g e n ta m ic in g e n ta m ic in n e ti lm ic in r e f [3 7 ] [3 8 ] [3 9 ] [4 0 ] [4 1 ] [4 2 ] [4 3 ] [4 4 ] http://dx.doi.org/10.5599/admet.1183 seth k. amponsah et al. admet & dmpk 10(1) (2022) 27-62 32 table 1. contined... d e te ct io n m o d e u v ( 3 6 5 n m ) u v ( 3 4 0 n m ) f lu o re sc e n ce (e xc it a ti o n , u g 1 f il te r; e m is si o n , k v 4 1 8 f il te r) . u v ( 3 4 0 n m ) f lu o re sc e n ce (e xc it a ti o n , 3 6 0 n m ; e m is si o n , 4 5 0 n m ) d e ri v a ti za ti o n a g e n t 1 -f d n b 2 ,4 ,6 tr in it ro b e n ze n e su lf o n ic a ci d o -p h th a la ld e h y d e t n b s o -p h th a la ld e h y d e m o b il e p h a se ; e lu ti o n ; f lo w r a te ( m l/ m in ) a ce to n it ri le /w a te r/ a ce ti c a ci d ( 4 7 0 :5 3 0 :1 ); i s o ; 2 .5 a ce to n it ri le /5 0 m m o l/ l p h o sp h a te b u ff e r a d ju st e d to p h 3 .5 w it h p h o sp h o ri c a ci d ( 7 0 :3 0 , v /v ); i s o ; 3 .0 7 4 % m e th a n o lw a te r 8 0 % m e th a n o lw a te r 0 .1 m s o d iu m a ce ta te , p h 7 .4 ; g r a ; 1 .0 9 5 % m e th a n o lw a te r 0 .2 m s o d iu m a ce ta te , p h 5 .0 ; g r a ; 1 .0 8 0 % m e th a n o lw a te r; g r a ; 1 .0 0 .1 m s o d iu m a ce ta te , p h 7 .4 ; g r a ; 1 .0 a ce to n it ri le /p h o sp h a te b u ff e r (5 2 :4 8 , v /v ); i s o ; 2 .0 2 5 m m s o d iu m p -t o lu e n e su lp h o n a te ; so d iu m p e rc h lo ra te a n h g jy d ro u s; g r a ; 0 .8 ( c 8 ); 1 .5 ( c 1 8 ) s ta ti o n a ry p h a se c 1 8 c 1 8 ( 5 0 ° c ) c 1 8 ( 2 5 ° c ) c 8 ( 5 0 ° c ) c 8 ; c 1 8 ( 5 0 ° c ) s a m p le p re p a ra ti o n s p e s p e , p p t s p e p p t /s p e p p t , s p e m a tr ix h -s e ru m h -s e ru m h -s e ru m h -s e ru m h -s e ru m in te rn a l s ta n d a rd k a n a m y ci n s is o m ic in n e a m in e g e n ta m ic in c 1 a n .i . t o b ra m y ci n t o b ra m y ci n t o b ra m y ci n t o b ra m y ci n t o b ra m y ci n k a n a m y ci n t o b ra m y ci n a st ro m ic in n e ti lm ic in s is o m ic in a n a ly te a m ik a ci n t o b ra m y ci n a m ik a ci n t o b ra m y ci n n e ti lm ic in g e n ta m ic in c 1 g e n ta m ic in c 1 a g e n ta m ic in c 2 d ib e k a ci n s is o m ic in a m ik a ci n s is o m ic in n e ti lm ic in a st ro m ic in m ic ro n o m ic in r e f [4 5 ] [4 6 ] [2 4 ] [4 7 ] [2 5 ] admet & dmpk 10(1) (2022) 27-62 bioanalysis of aminoglycozides by hplc doi: http://dx.doi.org/10.5599/admet.1183 33 table 1. contined... d e te ct io n m o d e f lu o re sc e n ce (e xc it a ti o n , 3 6 5 n m ; e m is si o n , 4 4 0 n m ) f lu o re sc e n ce (e xc it a ti o n , 3 4 1 n m ; e m is si o n , 4 4 0 n m ) u v ( 1 9 5 n m ) f lu o re sc e n ce (e xc it a ti o n , 3 5 1 n m ; e m is si o n , 4 2 0 n m ) f lu o re sc e n ce (e xc it a ti o n , 3 4 5 n m ; e m is si o n , 4 3 3 n m f lu o re sc e n ce (e xc it a ti o n , 3 0 2 n m ; e m is si o n , 4 2 0 n m ) d e ri v a ti za ti o n a g e n t o -p h th a la ld e h y d e o -p h th a la ld e h y d e                           o -p h th a la ld e h y d e n in h y d ri n m o b il e p h a se ; e lu ti o n ; f lo w r a te ( m l/ m in ) 0 .1 m d is o d iu m 1 ,2 -e th a n e d is u lf o n a te a n d 0 .0 0 5 m so d iu m o ct a n e su lf o n a te i n w a te r/ m e th a n o l m ix tu re (6 4 :3 6 , v /v ), a d ju st e d t o p h 3 .5 w it h a ce ti c a ci d ; g r a ; 2 .0 2 2 m m d is o d iu m 1 , 2 -e th a n e d is u lf o n a te a n d 5 m m so d iu m o ct a n e s u lf o n a te i n a w a te r/ a ce to n it ri le m ix tu re ( 8 0 :2 0 , v / v ) a d ju st e d w it h a ce ti c a ci d t o a b o u t p h 3 .5 ; is o ; 1 .5 3 7 m m d is o d iu m 1 ,2 -e th a n e su lf o n a te a n d 5 m m so d iu m o ct a n e s u lf o n a te i n a w a te r/ a ce to n it ri le m ix tu re ( 8 0 :2 0 , v /v ) a d ju st e d w it h a ce ti c a ci d t o a b o u t p h 3 .5 ; is o ; 1 .5 3 .7 6 g o f so d iu m l -h e xa n e su lp h o n a te a n d 9 .5 0 g o f tr ib a si c so d iu m p h o sp h a te d o d e ca h y d ra te d is so lv e d in w a te r (1 l ), p h 3 .0 a d ju st e d w it h p h o sp h o ri c a ci d /a ce to n it ri le ( 9 2 :8 , v /v ); i s o ; 1 .0 2 0 m m d is o d iu m 1 ,2 -e th a n e d is u lf o n a te , 5 m m so d iu m o ct a n e su lf o n a te , a n d 0 .4 m m n q s i n a w a te r/ a ce to n it ri le ( 8 0 :2 0 , v /v ), a d ju st e d t o a b o u t p h 3 .3 w it h a ce ti c a ci d ; is o ; 1 .5 so lv e n t a : 1 0 m m s o d iu m s u lf a te , 8 m m s o d iu m p a n ta n e su lf o n a te a n d 2 0 m m a ce ti c a ci d ; g r a ; 0 .8 so lv e n t b : 6 0 m m s o d iu m s u lf a te , 8 m m s o d iu m p a n ta n e su lf o n a te a n d 2 0 m m a ce ti c a ci d ; g r a ; 0 .8 2 0 m m s o d iu m o ct a n e s u lf o n a te , a n d 5 m m n in h y d ri n i n a w a te r/ a ce to n it ri le ( 8 0 :2 0 , v /v ), a d ju st e d t o a b o u t p h 3 .3 w it h a ce ti c a ci d ; is o ; 1 .5 s ta ti o n a ry p h a se c 1 8 c 1 8 c 1 8 ( 5 0 ° c ) c 1 8 ( 6 5 ° c ) c 1 8 ( 5 0 ° c ) c 1 8 ( 5 0 -9 5 °c ) s a m p le p re p a ra ti o n p p t ( m e o h ) p p t ( 3 .5 % p e rc h lo ri c a ci d ) s p e p p t ( 3 .5 % p e rc h lo ri c a ci d ) s p e p p t ( 3 .5 % p e rc h lo ri c a ci d ) m a tr ix h -s e ru m h -s e ru m h -s e ru m h -s e ru m rb -s e ru m h -s e ru m in te rn a l s ta n d a rd s is o m ic in n e ti lm ic in s is o m ic in n .i . d ih y d ro st p t o m y ci n n .i . n .i . n .i . a n a ly te t o b ra m y ci n s is o m ic in n e ti lm ic in k a n a m y ci n d ib e k a ci n s tr e p to m y ci n s tr e p to m y ci n g e n ta m ic in c 1 g e n ta m ic in c 1 a g e n ta m ic in c 2 s tr e p to m y ci n r e f [4 8 ] [2 6 ] [2 7 ] [4 9 ] [5 0 ] [5 1 ] http://dx.doi.org/10.5599/admet.1183 seth k. amponsah et al. admet & dmpk 10(1) (2022) 27-62 34 table 1. contined... d e te ct io n m o d e f lu o re sc e n ce (e xc it a ti o n , 2 6 0 n m ; e m is si o n , 4 1 8 n m ) u v ( 3 6 5 n m ) f lu o re sc e n ce (e xc it a ti o n , 3 4 0 n m ; e m is si o n , 4 5 0 n m ) f lu o re sc e n ce (e xc it a ti o n , 3 4 0 n m ; e m is si o n , 4 5 5 n m ) f lu o re sc e n ce (e xc it a ti o n , 3 4 0 n m ; e m is si o n , k v 4 1 8 ) f lu o re sc e n ce (e xc it a ti o n , 3 3 8 n m ; e m is si o n , 4 1 8 n m ) f lu o re sce n ce (e xcita tio n , 3 4 0 n m ; e m issio n , 4 1 8 n m ) u v ( 3 4 0 n m ) f lu o re sc e n ce (e xc it a ti o n , 3 4 0 n m ; e m is si o n , 4 1 5 n m ) d e ri v a ti za ti o n a g e n t o -p h th a la ld e h y d e o -p h th a la ld e h y d e o -p h th a la ld e h y d e o -p h th a la ld e h y d e o -p h th a la ld e h y d e o -p h th a la ld e h y d e f d n b o -p h th a la ld e h y d e m o b il e p h a se ; e lu ti o n ; f lo w r a te ( m l/ m in ) 1 % t e a s o lu ti o n ( a d ju st e d t o p h 6 .2 ± 0 .1 w it h p h o sp h o ri c a ci d )/ m e th a n o l (7 9 :2 1 , v /v ); i s o ; 2 .0 w a te r/ a ce to n it ri le /a ce ti c a ci d ( 3 0 0 :7 0 0 :1 , v /v /v ); is o ; 2 .2 m e th a n o l/ so d iu m 1 -h e p ta n e s u lf o n a te ( 2 .5 g )/ a ce ti c a ci d /w a te r (8 0 0 :2 0 0 :4 2 :2 0 8 , v /v /v /v ); i s o ; 0 .9 3 0 % e th y le n e g ly co l in 0 .0 5 m p h o sp h a te b u ff e r (p h 7 ); i s o ; 0 .5 so lv e n t a : e th y le n e d ia m in e te tr a a ce ti c a ci d , tr ip o ta ss iu m s a lt ( 2 .0 g m ) d is so lv e d i n 1 l w a te r/ m e th a n o l, ( 3 0 0 :7 0 0 , v /v ) so lv e n t b : m e th a n o l; g r a ; 1 .7 m e th a n o l/ b u ff e r so lu ti o n c o n ta in in g 0 .0 1 m so d iu m h e xa n e su lp h o n a te , 0 .1 m s o d iu m s u lp h a te a n d 1 7 m m a ce ti c a ci d ( 1 0 :9 0 , v /v ); i s o ; 1 .1 0 .2 m s o d iu m s u lp h a te , 0 .0 2 m s o d iu m p e n ta n e su lp h a te a n d 1 m l a ce ti c a ci d i n 1 l d is ti ll e d w a te r; is o ; 1 .2 a ce to n it ri le /2 -m e th o xy e th a n o l/ te tr a h y d ro fu ra n g la ci a l a ce ti c a ci d /t ri s( h y d ro xy m e th y l) -a m in o e th a n e ( 1 % a q u e o u s so lu ti o n ) (4 1 :4 .5 2 :4 .2 4 :0 .2 1 :5 0 , v /v ); g r a ; p ro g ra m m e d f lo w ra te 0 .0 5 m n a 2 s o 4 a n d 0 .0 0 5 m s o d iu m o ct y ls u lf a te , p h 3 .5 a d ju st e d w it h g la ci a l a ce ti c a ci d /m e th a n o l (7 0 :3 0 , v /v ); i s o ; 1 .5 s ta ti o n a ry p h a se c 1 8 c 1 8 ( 2 5 º c ) c 1 8 c 1 8 s h p c 1 8 ( 2 5 º c ) c 1 8 c 1 8 ( 5 8 º c ) c 1 8 ( 4 5 º c ) s a m p le p re p a ra ti o n s p e p p t ( a c n ) s p e s p e s p e s p e s p e u lt ra fi lt ra ti o n p p t ( 1 0 % w /v t c a ) m a tr ix h -s e ru m h -s e ru m g -s e ru m h -d b s rb -s e ru m cm il k h -s e ru m h u ri n e h -s e ru m h -s e ru m h u ri n e d -p la sm a in te rn a l s ta n d a rd n .i . g e n ta m ic in c 1 a n .i . n .i . n .i . d ib e k a ci n t o b ra m y ci n n .i . n .i . a n a ly te g e n ta m ic in c 1 g e n ta m ic in c 1 a g e n ta m ic in c 2 n e ti lm ic in s is o m ic in s is o m ic in n e o m y ci n is e p a m ic in a m ik a ci n a m ik a ci n a m ik a ci n r e f [5 2 ] [6 ] [5 3 ] [5 4 ] [1 4 ] [2 8 ] [5 5 ] [5 6 ] [5 7 ] admet & dmpk 10(1) (2022) 27-62 bioanalysis of aminoglycozides by hplc doi: http://dx.doi.org/10.5599/admet.1183 35 table 1. contined... d e te ct io n m o d e u v ( 3 3 0 n m ) f lu o re sc e n ce (e xc it a ti o n , 3 3 7 n m ; e m is si o n , 4 3 7 n m ) f lu o re sc e n ce (e xc it a ti o n , 2 6 0 n m ; e m is si o n , 3 1 5 n m ) u v ( 3 4 0 n m ) f lu o re sc e n ce (e xc it a ti o n , 3 4 0 n m ; e m is si o n , 4 3 0 n m ) f lu o re sc e n ce (e xc it a ti o n , 3 3 8 n m ; e m is si o n , 4 1 7 n m ) f lu o re sc e n ce (e xc it a ti o n , 2 6 0 n m ; e m is si o n , 3 1 5 n m ) f lu o re sc e n ce (e xc it a ti o n , 2 6 3 n m ; e m is si o n , 4 3 5 n m ) f lu o re sc e n ce (e xc it a ti o n , 2 6 0 n m ; e m is si o n , 3 1 5 n m ) d e ri v a ti za ti o n a g e n t o -p h th a la ld e h y d e o -p h th a la ld e h y d e f m o c -c 1 d n f b o -p h th a la ld e h y d e o -p h th a la ld e h y d e f m o c -c 1 n .i . f m o c -c 1 m o b il e p h a se ; e lu ti o n ; f lo w r a te ( m l/ m in ) 1 -h e p ta n e su lf o n ic a ci d ( 5 g ) d is so lv e d i n a ce ti c a ci d /w a te r/ m e 0 h ( 5 0 :2 5 0 :7 0 0 ); i s o ; 0 .5 so lv e n t a : w a te r/ a ce ti c a ci d /h e p ta n e su lp h o n ic a ci d (0 .1 m ) (8 0 :1 0 :1 0 , v /v /v ) so lv e n t b : a ce to n it ri le ; g r a ; 2 .0 a ce to n it ri le /w a te r (9 0 :1 0 , v /v ); i s o ; 1 .0 so lv e n t a : a ce to n it ri le a n d w a te r (5 0 :5 0 ) so lv e n t b : a ce to n it ri le ; g r a ; 1 .0 0 .0 1 1 m p e n ta n e s u lf o n ic a ci d s o d iu m s a lt , 0 .0 0 5 6 m s o d iu m s u lp h a te a n d 0 .1 % a ce ti c a ci d i n w a te r/ m e th a n o l (8 2 :1 8 ); i s o ; 1 .5 so lv e n t a : 0 .0 1 m h e xa n e su lp h o n a te /0 .0 1 7 m ( 0 .1 % ) a ce ti c a ci d i n w a te r so lv e n t b : 0 .0 1 m h e xa n e su lp h o n a te /0 .0 1 7 m ./ a ce ti c a ci d /0 .1 0 m s o d iu m s u lp h a te /3 .5 3 m (1 5 % ) m e th a n o l; g r a ; 1 .1 a ce to n it ri le – w a te r (9 0 :1 0 , v /v ); i s o ; 1 .0 0 .8 g o f 1 ,2 -n a p h th o q u in o n e -4 -s u lf o n ic a ci d ( n q s ) in 0 .0 1 m s o d iu m h e xa n e -1 -s u lf o n ic a ci d /a ce to n it ri le ( 8 8 0 :1 2 0 , v /v ); i s o ; 0 .5 a ce to n it ri le /w a te r (8 4 .5 :1 5 .5 , v /v ); i s o ; 2 .5 s ta ti o n a ry p h a se c 1 8 c 1 8 c 1 8 c 1 8 ( 2 5 ° c ) c 1 8 ( 2 5 ° c ) c 1 8 c 1 8 ( 2 0 -2 5 °c ) c 8 ( 2 5 ° c ) c 1 8 ( 3 0 ° c ) s a m p le p re p a ra ti o n s p e ll e s p e p p t (m e o h /t c a ) s p e , p p t ( 3 0 % t c a ) p p t ( m e th y le n e ch lo ri d e ), s p e s p e s p e ll e m a tr ix cm il k h -s e ru m h -s e ru m m -p la sm a cm il k h -s e ru m h -u ri n e h -s e ru m cm il k h -s e ru m in te rn a l s ta n d a rd n e ti lm ic in g e n ta m ic in n .i . n .i . n .i . d ib e k a ci n n .i . n .i . n e o m y ci n a n a ly te g e n ta m ic in n e ti lm ic in g e n ta m ic in c 1 g e n ta m ic in c 1 a g e n ta m ic in c 2 g e n ta m ic in c 2 a g e n e ti ci n g e n ta m ic in is e p a m ic in n e o m y ci n n e ti lm ic in s is o m ic in s tr e p to m y ci n d ih y d ro st re p to m y ci n g e n ta m ic in r e f [5 8 ] [5 9 ] [6 0 ] [2 9 ] [6 1 ] [6 2 ] [6 3 ] [3 0 ] [1 6 ] http://dx.doi.org/10.5599/admet.1183 seth k. amponsah et al. admet & dmpk 10(1) (2022) 27-62 36 table 1. contined... d e te ct io n m o d e u v ( 3 6 5 n m ) u v ( 2 3 0 n m ) f lu o re sc e n ce (e xc it a ti o n , 3 4 0 n m ; e m is si o n , 4 1 8 n m ) u v ( 2 3 0 n m ) m s /e s i (+ ) m s /e s i (+ ) m s /e s i (+ ) e ls d c l d e ri v a ti za ti o n a g e n t f n d b n it c o -p h th a la ld e h y d e n it c m o b il e p h a se ; e lu ti o n ; f lo w r a te ( m l/ m in ) a ce to n it ri le /t ri s b u ff e r (8 .3 m m o l/ l, t it ra te d t o p h 7 .0 w it h h c l) ( 6 8 0 :3 2 0 , v /v ); i s o ; 1 .2 w a te r/ a ce to n it ri le ( 5 7 :4 3 , v /v ); i s o ; 0 .8 m e th a n o l/ g la ci a l a ce ti c a ci d /w a te r (8 0 0 :2 0 :1 8 0 , v /v /v ) co n ta in in g 0 .0 2 m s o d iu m h e p ta n e s u lf o n ic a ci d , p h 3 .4 ; is o ; 1 .0 w a te r/ a ce to n it ri le ( 5 0 :5 0 , v /v ); i s o ; 1 .3 so lv e n t a : w a te r co n ta in in g 2 m m a m m o n iu m a ce ta te , 0 .1 % ( v /v ) fo rm ic a ci d so lv e n t b : m e th a n o l co n ta in in g 2 m m a m m o n iu m a ce ta te , 0 .1 % ( v /v ) fo rm ic a ci d ) so lv e n t c : s o lv e n t a c o n ta in in g h e p ta fl u o ro b u ty ri c a ci d ( h b f a ) (1 0 m m )/ 2 0 % s o lv e n t b ; g r a ; 2 .5 so lv e n t a : a ce to n it ri le , 2 m m a m m o n iu m a ce ta te a n d f o rm ic a ci d ( 5 /9 5 /0 .2 , v /v /v ) so lv e n t b : a ce to n it ri le , 2 m m a m m o n iu m a ce ta te a n d f o rm ic ( 9 5 /5 /0 .2 , v /v /v ); g r a ; 0 .6 so lv e n t a : a ce to n it ri le /1 0 m m a m m o n iu m a ce ta te /f o rm ic a ci d ( 5 :9 5 :0 .2 , v /v /v ) so lv e n t b : a ce to n it ri le /1 0 m m a m m o n iu m a ce ta te /f o rm ic a ci d ( 9 5 :5 :0 .2 , v /v /v ); g r a ; 0 .6 w a te r/ a ce to n it ri le 5 5 :4 5 c o n ta in in g 1 .5 0 m l/ l h f b a (1 1 .6 m m ); i s o ; 1 .0 1 0 − 2 m o l/ l p o ta ss iu m h y d ro g e n p h th a la te a t p h 3 .3 5 , a d ju st e d w it h d il u te d s o d iu m h y d ro xi d e /a ce to n it ri le ( 9 0 :1 0 , v /v ); i s o ; 1 .0 s ta ti o n a ry p h a se c 1 8 ( 2 5 º c ) c 8 ( 7 0 º c ) c 1 8 c 1 8 c 1 8 ( 2 5 º c ) h il ic h il ic c 1 8 ( 4 5 º c ) c 1 8 s a m p le p re p a ra ti o n s p e ll e s p e p p t p p t ( a c n ) s p e s p e s p e p p t m a tr ix h -s e ru m d -s e ru m h -u ri n e d -u ri n e h -s e ru m h -u ri n e h -s e ru m h -s e ru m h -s e ru m h -s e ru m h -s e ru m h -u ri n e h -s e ru m h -u ri n e in te rn a l s ta n d a rd n .i . n a p h th a le n n e ti lm ic in a n th ra ce n e s is o m ic in n .i . n .i . n .i . n .i . a n a ly te g e n ta m ic in c 1 g e n ta m ic in c 1 a g e n ta m ic in c 2 a m ik a ci n g e n ta m ic in t o b ra m y ci n t o b ra m y ci n a m ik a ci n g e n ta m ic in k a n a m y ci n n e o m y ci n p a ro m o m yc in t o b ra m y ci n n e o m y ci n t o b ra m y ci n a m ik a ci n r e f [6 4 ] [6 5 ] [6 6 ] [6 7 ] [6 8 ] [3 1 ] [6 9 ] [7 0 ] [7 1 ] admet & dmpk 10(1) (2022) 27-62 bioanalysis of aminoglycozides by hplc doi: http://dx.doi.org/10.5599/admet.1183 37 table 1. contined... d e te ct io n m o d e m s /e s i (+ ) m s /e s i (+ ) f lu o re sc e n ce (e xc it a ti o n , 4 6 5 n m ; e m is si o n , 5 3 1 n m ) m s /e s i (+ ) p e d f lu o re sc e n ce (e xc it a ti o n , 2 5 0 n m ; e m is si o n , 3 9 5 n m ) f lu o re sc e n t (e xc it a ti o n , 4 9 0 n m ; e m is si o n , 5 1 8 n m ) m s /e s i (+ ) d e ri v a ti za ti o n a g e n t 7 -f lu o ro -4 n it ro b e n z2 -o xa -1 ,3 d ia zo le 6 -a m in o q u in o ly ln h y d ro xy s u cc in im id y l ca rb a m a te f it c m o b il e p h a se ; e lu ti o n ; f lo w r a te ( m l/ m in ) so lv e n t a : 2 0 m so lv e n t: b 2 0 m m p f p a i n w a te r/ a ce to n it ri le (5 0 /5 0 , v /v ); i s o ; 0 .3 m p f p a i n w a te r so lv e n t a : 2 0 m m f o rm ic a ci d a n d 1 0 m m n f p a i n w a te r so lv e n t b : 2 0 m m f o rm ic a ci d a n d 1 0 m m n f p a i n m e th a n o l; g r a ; 0 .5 0 .0 1 m s o d iu m a ce ta te p h 3 /a ce to n it ri le ( 8 5 :1 5 , v /v ); i s o ; 1 .2 2 m m a m m o n iu m a ce ta te ( p h 3 .2 )/ 5 % a ce to n it ri le (9 5 :5 , v /v ); i s o ; 0 .5 1 .5 g s o d iu m -1 -o ct a n e su lp h o n a te , 2 0 g a n h y d ro u s so d iu m s u lp h a te , 1 5 m l te tr a h y d ro fu ra n , 2 5 0 m l 0 .2 m p h o sp h a te b u ff e r p h 3 , w a te r u p t o 1 0 0 0 m l; is o ; 1 .0 2 0 m m k h 2 p o 4 c o n ta in in g 8 m m t e a ( p h 7 .0 )/ a ce to n it ri le ( 7 8 :2 2 , v /v ); i s o ; 1 .0 a ce to n it ri le /m e th a n o l/ g la ci a l a ce ti c a ci d /w a te r (4 2 0 :6 0 :5 :5 1 5 , v /v /v /v ); i s o ; 1 .0 p h a se a : a ce to n it ri le /w a te r (5 0 :9 5 0 , v /v ) co n ta in in g 2 0 m m h f b a p h a se b : a ce to n it ri le /w a te r (5 0 0 :5 0 0 , v /v ) co n ta in in g 2 0 m m h f b a ; g r a ; 0 .3 s ta ti o n a ry p h a se c 1 8 ( 3 0 ° c ) c 1 8 c 1 8 ( 6 0 ° c ) c 1 8 ( 4 0 ° c ) c 1 8 ( 4 0 ° c ) c 1 8 (a m b ie n t) c 1 8 ( 3 0 ° c ) s a m p le p re p a ra ti o n ll e , p p t ( t c a ), s p e p p t ( a c n ) p p t ( a c n ) p p t (a c n /p h o sp h a te b u ff e r) p p t ( a c n ) s p e s p e m a tr ix cm il k h -s e ru m r b se ru m h -s e ru m h -s e ru m h -c s f h -s e ru m h -u ri n e cm il k in te rn a l s ta n d a rd s tr e p to m y ci n k a n a m y ci n n .i . s is o m ic in n .i . a m ik a ci n n e o m y ci n n .i . a n a ly te d ih y d ro st re p to m y ci n (n e o m y ci n ) n e o 1 8 0 4 a 7 n e o 1 8 0 4 a 8 n e o 1 8 0 4 a 9 n e o 1 8 0 4 b 4 a m ik a ci n t o b ra m y ci n a m ik a ci n is e p a m ic in t o b ra m y ci n n e o m yc in s tr e p to m yc in d ih yd ro st re p to m yc in a m ik a ci n k a n a m yc in p a ro m o m yc in t o b ra m yc in s p e ct in o m yc in a p ra m yc in h yg ro m yc in g e n ta m ic in (c 1 ) g e n ta m ic in (c 1 a ) g e n ta m ic in ( c 2 0 r e f [7 2 ] [7 3 ] [7 4 ] [7 5 ] [1 5 ]* [7 6 ] [7 7 ] [5 ]* http://dx.doi.org/10.5599/admet.1183 seth k. amponsah et al. admet & dmpk 10(1) (2022) 27-62 38 table 1. contined... d e te ct io n m o d e f lu o re sc e n ce (e xc it a ti o n , 2 6 5 n m ; e m is si o n , 3 1 5 n m ) m s /e s i (+ ) r r s d (e xc it a ti o n /e m i ss io n , 3 7 0 n m ) m s /e s i (+ ) f lu o re sc e n ce (e xc it a ti o n , 3 9 0 n m ; e m is si o n , 4 8 0 n m ) m s /e s i (+ ) m s /e s i (+ ) d e ri v a ti za ti o n a g e n t f m o c -c 1 f lu o re sc a m in e m o b il e p h a se ; e lu ti o n ; f lo w r a te ( m l/ m in ) w a te r/ a ce to n it ri le ( 3 2 :6 8 , v /v ); i s o ; 1 .0 b u ff e r a ( w a te r/ fo rm ic a ci d 0 .0 5 % ) 0 .1 % t ri fl u o ro a ce ti c a ci d ( p h 2 .2 ); i s o ; 0 .4 w a te r/ a ce to n it ri le e a ch c o n ta in in g 0 .0 0 5 % (v /v ) tr if lu o ro a ce ti c a ci d a n d 0 .1 % ( v /v ) fo rm ic a ci d ; g r a ; 0 .7 m e th a n o l/ w a te r (6 0 :4 0 , v /v ); i s o ; 1 .0 so lv e n t a : p u ri fi e d w a te r so lv e n t b : a ce to n it ri le 1 0 0 % so lv e n t c : p e rf lu o ro p e n ta n o ic a ci d ( 2 0 0 m m )/ a m m o n iu m a ce ta te ( 1 3 0 m m ) in p u ri fi e d w a te r; g r a ; 0 .4 so lv e n t a : 1 0 m m n o n fl u o ro p e n ta n o ic a ci d ( n f p a ) so lv e n t b : a ce to n it ri le i n 1 0 m m n f p a ; g r a ; 0 .3 s ta ti o n a ry p h a se c 8 ( 3 0 ° c ) c 1 8 ( 3 5 ° c ) c 1 8 ( 3 0 ° c ) c 1 8 c 1 8 c 1 8 ( 3 0 ° c ) c 1 8 s a m p le p re p a ra ti o n p p t ( a c n ) p p t ( a c n ) p p t ( a c n ) p p t ( 0 .3 m p e rc h lo ri c a ci d ) p p t ( a c n ) p p t ( t c a ) p p t ( t c a ) m a tr ix rp la sm a h -s e ru m h -s e ru m h -u ri n e h -s e ru m h -p la sm a h -p la sm a b -m il k in te rn a l s ta n d a rd a m ik a ci n q u in o xa li n e n .i . d ib e k a ci n n .i . k a n a m y ci n b n .i . a n a ly te is e p a m ic in a m ik a ci n g e n ta m ic in a m ik a ci n n e ti lm ic in e ti m ic in a rb e k a ci n t o b ra m y ci n a m ik a ci n g e n ta m ic in c 1 g e n ta m ic in c 1 a g e n ta m ic in c 1 2 s p e ct in o m y ci n t o b ra m y ci n g e n ta m ic in k a n a m y ci n h y g ro m y ci n a p ra m y ci n s tr e p to m y ci n d ih y d ro st re p to m y ci n a m ik a ci n n e o m y ci n r e f [7 8 ] [7 9 ] [3 2 ] [3 3 ] [8 0 ] [8 1 ] [3 4 ] admet & dmpk 10(1) (2022) 27-62 bioanalysis of aminoglycozides by hplc doi: http://dx.doi.org/10.5599/admet.1183 39 table 1. contined... d e te ct io n m o d e f lu o re sc e n ce (e xc it a ti o n , 2 6 5 n m ; e m is si o n , 3 1 5 n m ) m s /e s i (+ ) m s /e s i (+ ) m s /e s i (+ ) m s -e s i (+ ) h -s e ru m , h u m a n s e ru m ; d -s e ru m , d o g s e ru m ; h -p la sm a , h u m a n p la sm a ; h -u ri n e , h u m a n u ri n e ; rp la sm a , ra t p la sm a ; rb -u ri n e , ra b b it u ri n e ; g -p la sm a , g u in e a p ig p la sm a ; rb -s e ru m , ra b b it s e ru m ; h -d b s , h u m a n d ri e d b lo o d s p o t; c -m ilk , co w m ilk ; m -p la sm a , m o u se p la sm a ; h -c s f, h u m a n c e re b ra l sp in a l fl u id ; m e o h , m e th a n o l; c n , cy a n o ; a c n , f a , fo rm ic a ci d ; a ce to n it ri le ; t c a , tr ic h lo ro a ce ti c a ci d ; r t , ro o m t e m p e ra tu re ; is o , is o cr a ti c; g r a , g ra d ie n t; s p e , so lid p h a se e xt ra ct io n ; p p t , p ro te in p re ci p it a ti o n ; ll e , liq u id -l iq u id e xt ra ct io n ; c l, c h e m ilu m in e sc e n ce d e te ct io n ; e d t a , e th yl e n e d ia m in e t e tr a a ce ti c a ci d ; e ls d , e va p o ra ti ve l ig h t sc a tt e ri n g d e te ct o r; e s i: e le ct ro sp ra y io n iz a ti o n ; h f b a : h e p ta fl u o ro b u ty ri c a ci d ; i. d ., i n te rn a l d ia m e te r; m s , m a ss s p e ct ro m e tr y; e s i, e le ct ro sp ra y io n iz a ti o n ; n a 2 s o 4 , so d iu m s u lp h a te ; n f p a , n o n a fl u o ro p e n ta n o ic a ci d ; n .i .: n o t in d ic a te d ; p e d : p u ls e e le ct ro ch e m ic a l d e te ct io n ; u h p lc , u lt ra -h ig h p e rf o rm a n ce l iq u id c h ro m a to g ra p h y; u v -v is : u lt ra vi o le tvi si b le l ig h t; f d n b , 1 -f lu o ro -2 ,4 -d in it ro b e n ze n e ; f m o c -c 1 , 9 -f lu o re n yl m e th yl ch lo ro fo rm a te ; n it c , 1 n a p h th yl is o th io cy a n a te ; h il ic , h yd ro p h ili c in te ra ct io n l iq u id c h ro m a to g ra p h y; r r s d , re so n a n ce r a y le ig h s ca tt e ri n g d e te ct io n ; * , m e th o d s th a t ca n b e o p ti m is e d f o r th e a n a ly si s o f a ll th e a m in o g ly co si d e s u n d e r re vi e w . d e ri v a ti za ti o n a g e n t f m o c -c 1 m o b il e p h a se ; e lu ti o n ; f lo w r a te ( m l/ m in ) a ce to n it ri le /w a te r (7 0 :3 0 , v /v ); i s o ; 0 .4 so lv e n t a : 0 .2 % ( v /v ) h f b a i n w a te r so lv e n t b : 0 .2 % ( v /v ) h f b a i n a c n ; g r a ; 0 .3 so lv e n t a : w a te r co n ta in in g 0 .1 % f o rm ic a ci d , 0 .0 1 % h f b a so lv e n t b : a ce to n it ri le c o n ta in in g 0 .1 % f o rm ic a ci d a n d 0 .0 1 % h f b a ; g r a ; 0 .4 so lv e n t: p u ri fi e d w a te r w it h 0 .1 % f o rm ic a ci d a n d 0 .0 1 % o f h f b a so lv e n t b : d a ce to n it ri le w it h 0 .1 0 .1 % f o rm ic a ci d a n d 0 .0 1 % o f h f b a ; g r a ; 0 .4 5 m m h f b a i n w a te r/ a ce to n it ri le ( 7 :3 , v /v ) m ix tu re (3 .5 m m h f b a i n t h e m ix tu re ); i s o ; 0 .4 s ta ti o n a ry p h a se c 1 8 ( 3 0 º c ) c 1 8 ( 4 0 º c ) c 1 8 ( 4 0 º c ) c 1 8 ( 4 0 º c ) c 1 8 ( 4 0 º c ) s a m p le p re p a ra ti o n p p t ( a c n ) s p e p p t ( 0 .1 % f a ) p p t ( 0 .1 % f a ) p p t ( 2 0 % w /v t c a ) m a tr ix h -s e ru m g p e ri ly m p h / c s f h -s e ru m h -p la sm a h -p la sm a in te rn a l s ta n d a rd n .i . t o b ra m y ci n k a n a m y ci n k a n a m y ci n d e u te ra te d p a ro m o m yc in a ce ti c a ci d a n a ly te a m ik a ci n n e o m y ci n a m ik a ci n a m ik a ci n p a ro m o m yc in r e f [8 2 ] [8 3 ] [8 4 ] [8 5 ] [3 5 ] http://dx.doi.org/10.5599/admet.1183 seth k. amponsah et al. admet & dmpk 10(1) (2022) 27-62 40 sample cleanup the matrix in which a drug is found can affect bioanalysis. also, matrices can compromise the sensitivity and selectivity of bioanalysis methods [86-89], which is deemed a “matrix effect”. the matrix effect could be due to endogenous or exogenous agents. some endogenous substances include salts, carbohydrates, amines, urea, lipids, peptides, and metabolites [90,91]. exogenous substances may include anticoagulants, blood preservatives, and mobile-phase additives such as buffer salts [92,93]. sample cleanup is required to isolate analyte(s) of interest from a matrix. sample cleanup tends to reduce or remove matrix components and concentrate the analyte(s). this process improves assay sensitivity and selectivity. an optimal sample cleanup system should be capable of minimizing matrix effect while ensuring reliable extraction recovery [94]. basically, there are three sample preparation methods that can be applied during hplc bioanalysis of aminoglycosides; protein precipitation (ppt), solid-phase extraction (spe), and liquid-liquid extraction (lle) [4]. all the articles reviewed provided sample preparation methods; 41 used ppt, 33 spe, and 4 lle. one article reported the use of ultrafiltration as a sample preparation method [56], 6 articles reported a combination of ppt and spe [25,42,46,47,61,62], and 1 article used a combination of ppt, spe, and lle [72]. solvents that were used for ppt included acetonitrile for the extraction of netilmicin, amikacin, tobramycin, gentamicin, etimicin, neomycin, and isepamicin [6,15,21,23,32,37-40,42,68,73-76,78,79,82,86]; trichloroacetic acid for the extraction of geneticin, spectinomycin, tobramycin, gentamicin, kanamycin, hygromycin, apramycin, streptomycin, dihydrostreptomycin, amikacin, neomycin, and paromomycin [29,34,35,57,61,72,81]; methanol for the extraction of geneticin, amikacin, tobramycin, sisomicin, and netilmicin [29,41,48]; perchloric acid for the extraction of kanamycin, dibekacin, arbekacin, and streptomycin [26,33,49,51]; formic acid for the extraction of amikacin [84,85]; and methylene chloride for the extraction of isepamicin [62]. ppt aids in reducing interference during derivatization [7], and this may have accounted for its use by 41 articles. spe is useful in isolating polar analytes such as aminoglycosides. spe has also been proven to be useful when the volume of a matrix is high [4]. preparation of matrices containing aminoglycosides using spe tends to give well reproducible recovery [95]. despite these merits, spe is relatively expensive compared to ppt. in this review, 33 of the articles reported the use of spe as a sample preparation technique for the extraction of gentamicin, amikacin, tobramycin, sisomicin, netilmicin, astromicin, micronomicin, streptomycin, neomycin, isepamicin, dihydrostreptomycin, kanamycin, paromomycin, spectinomycin, apramycin, and hygromycin in biological matrices. also, four articles adopted the lle as a sample preparation technique for the extraction of netilmicin, gentamicin, amikacin, and dihydrostreptomycin. aminoglycosides are hydrophilic in nature, and lle may not be ideal for extraction. chromatographic conditions all articles reviewed reported chromatographic conditions, and these are highlighted in this review. mode of chromatography the most-reported mode of chromatography used for bioanalysis of aminoglycosides was reversedphase hplc (66 articles). this may be due to the polarity of the analytes of interest [96]. other modes of chromatography for the separation of aminoglycosides used were hydrophilic interaction liquid chromatography (2 articles) for the analysis of amikacin, gentamicin, kanamycin, neomycin, paromomycin, admet & dmpk 10(1) (2022) 27-62 bioanalysis of aminoglycozides by hplc doi: http://dx.doi.org/10.5599/admet.1183 41 and tobramycin [31,69]; hydrophobic interaction chromatography (1 article) for the analysis of neomycin [14]; and ion-exchange chromatography (1 article) for the analysis of gentamicin [40]. one article reported the use of mixed-mode chromatography using both reversed-phase and normal chromatography for the analysis of amikacin [41]. stationary phase the selection of a chromatographic mode of separation is dependent on the choice of the column. in this review, 63 articles reported the use of c18 columns for the separation of any of the twenty aminoglycosides; while 3 other articles used c8 columns for the separation of streptomycin, dihydrostreptomycin, and amikacin [30,47,65]. also, ion exchange for the separation of gentamicin [40]; cyano for the separation of tobramycin [37]; shielded hydrophobic phase for the separation of neomycin [14]; silica for the separation of amikacin [41]; and hydrophilic interaction columns for the separation of amikacin, gentamicin, kanamycin, neomycin, paromomycin, and tobramycin [31,69] have been reported. this implies that c18 columns may be the most appropriate for liquid chromatographic bioanalysis of aminoglycosides. in this review, columns were kept at temperatures ranging from 22 °c to as high as 100 °c [41]. overall, the most widely used column temperature was 25 °c for the separation of netilmicin, amikacin, tobramycin, gentamicin, dibekacin, sisomicin, isepamicin, geneticin, streptomycin, dihydrostreptomycin, and neomycin [6,24,28-30,42,61,64,68,73]; followed by 30 °c for the separation of neomycin, streptomycin, dihydrostreptomycin, amikacin, kanamycin, paromomycin, tobramycin, spectinomycin, apramycin, hygromycin, gentamicin, netilmicin, etimicin, and isepamicin [5,16,32,72,78,81,82]; and 50 °c for the separation of sisomicin, netilmicin, astromicin, micronomicin, streptomycin, gentamicin, tobramycin, and amikacin [25,27,43,46,47,50]. internal standard internal standards are normally employed to offset injection volume errors and/or losses during sample extraction [97]. forty-four (44) articles used internal standards, which include n-acetyl gentamicin c1 [22], tobramycin [22,24,25,44,55,83], amikacin [22,76,78], gentamicin c2 [37], netilmicin [25,39,42,48,58,66], kanamycin [45,47,73,84,85], sisomicin [25,46,48,68,75], neamine [24], gentamicin c1a [6,24], astromicin [25], dihydrostreptomycin [27], dibekacin [28,33,62], gentamicin [59], neomycin [16,77], naphthalene [65], anthracene [67], streptomycin [72], quinoxaline [79], kanamycin b [81], and deuterated paromomycin acetic acid [6]. of the internal standards used, 13 were aminoglycosides, 3 were modified aminoglycosides [6,22,81], and 4 were not aminoglycosides or congeners of aminoglycosides [24,65,67,79]. mode of elution although the mobile phase is not directly responsible for chromatographic separation, it can affect chromatographic resolution, selectivity, and efficiency [96]. the selection of a suitable mobile phase for the hplc system is dependent on the physicochemical properties of the analyte [98]. since reversed-phase chromatography was reported by most articles, mobile phase solvents used consisted of an aqueous buffer and a non-ultraviolet active water-miscible organic solvent [96]. the solvents included acetonitrile for the separation of netilmicin, gentamicin, and amikacin [21,39,40,47,60]; methanol and water mixture for the separation of gentamicin, amikacin, tobramycin, netilmicin, dibekacin, and sisomicin [20,22,24,36,44]; and heptanesulphonic acid for separation of gentamicin, and netilmicin [58,59]. other solvents such as tetrahydrofuran for the separation of amikacin [15]; octanesulphonate for the separation of tobramycin, sisomicin, and netilmicin [48]; and ethylene glycol for the separation of sisomicin [54] were also reported. in the case where ionisable analytes were present, the ph of the mobile phase had a significant effect on the ionization state(s) of the analyte(s), which eventually affected resolution. thus, the buffer reported by http://dx.doi.org/10.5599/admet.1183 seth k. amponsah et al. admet & dmpk 10(1) (2022) 27-62 42 most articles were aqueous, which included tris buffer [23], sodium acetate [24], and potassium phosphate buffer [15]. basic analytes such as aminoglycosides are protonated at low ph when ionized. poor elution of analytes could contribute to peak broadening [96,99]. out of 71 articles reviewed, 50 employed isocratic elution for the separation of streptomycin, gentamicin, amikacin, dihydrostreptomycin, kanamycin, netilmicin, isepamicin, sisomicin, paromomycin, neomycin, tobramycin, etimicin, and dibekacin; whilst the remaining 21 used gradient elution for the separation of streptomycin, gentamicin, amikacin, dihydrostreptomycin, kanamycin, netilmicin, apramycin, isepamicin, sisomicin, paromomycin, geneticin, neomycin, tobramycin, arbekacin, spectinomycin, hygromycin, dibekacin, astromicin, and micronomicin. for complex multicomponent samples, gradient elution is used since all components cannot be eluted between the retention factor of 1 and 10. an isocratic mode is sufficient if the measured ratio is less than 0.25, however, if the ratio is greater than 0.25, then an elution gradient is deemed suitable [96,99]. derivatization and mode of detection of the 71 articles reviewed, fluorescence (36 articles), uv (15 articles), mass spectrometry (15 articles), resonance rayleigh scattering (1 article) [32], pulsed electrochemical (1 article) [15], chemiluminescence (1 article) [71], and evaporative light scattering (1 article) [70] were used in the detection of aminoglycosides. fluorescence (~50 %) was the most used method for the detection of streptomycin, gentamicin, amikacin, dihydrostreptomycin, kanamycin, netilmicin, isepamicin, sisomicin, neomycin, tobramycin, dibekacin, astromicin, and micronomicin. however, non-fluorescence drugs such as aminoglycosides are mostly difficult to detect by this mode due to the absence of a fluorophore. this shortcoming can be mitigated by derivatization [100]. additionally, the mobile phase used must be selected with care, as highly polar solvents or halide ions can quench fluorescence. it is noteworthy that fluorescence is mostly preferred to uv detection due to its high sensitivity and selectivity [101]. aminoglycosides assayed in various studies with fluorescence detectors were achieved at an excitation wavelength range of 220 nm [21] to 490 nm [77] and emission wavelength range of 415 nm [57] to 531 nm [74]. uv-visible detectors are not easily influenced by the mobile phase and surrounding temperature [102]. uv-visible detectors interact with compounds containing chromophores. since aminoglycosides do not have chromophores, derivatization of these compounds is necessary for their detection. uv-visible was used for the detection of gentamicin, amikacin, tobramycin, streptomycin, netilmicin, and geneticin [6,27,29,38,39,41,45,47,58,64,65,67]. in this review, the uv wavelength range used for the detection of aminoglycosides was 195 nm [27] to 365 nm [6,38,45,64]. in all the 51 articles that used fluorescence or uv detection, only 3 articles did not use derivatization for the detection of streptomycin, dihydrostreptomycin, and sisomicin [27,30,54]. the most common derivatizing agent used in the bioanalysis of aminoglycosides was o-phthalaldehyde; reported by 24 articles for the derivatization of gentamicin, amikacin, kanamycin, netilmicin, isepamicin, sisomicin, neomycin, tobramycin, dibekacin, astromicin, and micronomicin [6,14,20,22-26,28,36,37,43,44,48,50,52,53,55,5759,61,62,66]. dansyl chloride for the derivatization of netilmicin [21]; 1-fluoro-2,4-dinitrobenzene for the derivatization of geneticin, gentamicin, and amikacin [29,38,41,45,56,64]; 7-fluoro-4-nitrobenz-2-oxa-1,3diazole for the derivatization of amikacin [74]; benzene sulphonyl chloride for the derivatization of gentamicin [39]; 6-aminoquinolyl-n-hydroxysucciminidylcarbamate for the derivatization of isepamicin [76]; fluorescamine for the derivatization of gentamicin, and tobramycin [40,80]; 1-naphthyl isothiocyanate for the derivatization of amikacin, and tobramycin [65,67]; fluorescein isothiocyanate for the derivatization of tobramycin [77]; o-phthalicdicarboxaldehyde for the derivatization of gentamicin [42]; 2,4,6trinitrobenzene sulfonic acid for the derivatization of tobramycin, and amikacin [46,47]; -naphthoquinoneadmet & dmpk 10(1) (2022) 27-62 bioanalysis of aminoglycozides by hplc doi: http://dx.doi.org/10.5599/admet.1183 43 4-sulfonate for the derivatization of streptomycin [49]; ninhydrin for the derivatization of streptomycin [51]; and 9-fluorenylmethylchloroformate for the derivatization of gentamicin, neomycin, netilmicin, sisomicin, isepamicin, and amikacin [16,60,63,78,82] were other derivatizing agents used in the bioanalysis of aminoglycosides. challenges associated with uv and fluorescence detections (need for chromophore or fluorophore necessitating derivatization) can be circumvented using mass spectrometry. also, mass spectrometry can analyze small sample volumes with high precision, sensitivity, and selectivity [68,81]. in this review, 15 articles reported the use of mass spectrometry as a detector for the bioanalysis of neomycin, streptomycin, dihydrostreptomycin, amikacin, kanamycin, paromomycin, tobramycin, spectinomycin, apramycin, hygromycin, gentamicin, and arbekacin [5,31,33-35,68,69,72,73,75,79,81,83-85]. interestingly, articles published between 1977 [20] to 2002 [67] were dominated by fluorescence or uv detection methods. more importantly, mass spectrometry was the mostly used detection mode from 2014 to 2020 [34,35,81,83-85], except in one case [82]. this is not surprising as mass spectrometry appears to be an effective detection method [103]. in this review, hplc coupled with resonance rayleigh scattering detection was used in analyzing three aminoglycosides; amikacin, netilmicin, and etimicin [32]. an advantage of the resonance rayleigh scattering detector over other spectroscopic techniques is that the detection limit is lower by several orders of magnitude [104]. the pulsed electrochemical detector is mostly used to analyze carbohydrates and polyalcohol. they are also used in analyzing amines, amino acids, and sulphur-containing compounds [105, 106]. one (1) article reported the use of hplc coupled with the pulsed electrochemical detector in the bioanalysis of amikacin [15]. this approach was used to address the shortfall of using derivatization for the detection of aminoglycosides [15]. chemiluminescence allows the detection of analytes at ultra-high sensitivity. in this review, 1 article reported the use of chemiluminescence for the detection of amikacin, and the chemiluminescence reagent used was luminol in combination with hydrogen peroxide and cu 2+ [71]. hplc coupled with evaporative light scattering detector is rapidly becoming a quasi-universal detector, mitigating the need for derivatization of non-absorbing analytes. in this review, one article reported the use of an evaporative light scattering detector for the direct determination of tobramycin [70]. performance metrics performance metrics are quantifiable terms that indicate the quality of an analytical process. some performance metrics include specificity, sensitivity, linearity, the lower limit of quantification (lloq), limit of detection (lod), precision, accuracy, and calibration range. in this review, all the 71 articles reported some aspects of method performance characteristics. method performance characteristics reported by most articles included calibration range, linearity, recovery, repeatability, and intermediate precision (table 2). out of the 71 articles reviewed, only four reported on matrix effect [63,83-85]. resolution was reported by one article [75]. the calibration range is often obtained from a calibration curve [107]. sixty-seven (67) articles reported calibration ranges, whilst 4 did not [5,23,59,79] (table 2). these calibration ranges of hplc varied significantly for the various drugs assayed, with some in the range of 0.1-0.5 ng/ml [14] and others 1250200000 ng/ml [83]. of the 20 aminoglycosides reported, only eight had established a therapeutic reference range. the reported calibration range of some of the analytes reported was outside the established http://dx.doi.org/10.5599/admet.1183 seth k. amponsah et al. admet & dmpk 10(1) (2022) 27-62 44 therapeutic reference range [14,15,24,25,30,31,34,37,48,55,57,61,65,67,71,74,75,80-82], which invalidates the measurements obtained from the bioanalysis. the quality of a bioanalytical method is highly dependent on the linearity of the calibration curve [108]. the linearity of the calibration curve is usually expressed as a coefficient of correlation (r 2 ). the coefficient of correlation close to 1 (r 2 ≈1) is mostly considered ideal. in the current review, 61 articles reported r 2 values for almost all analytes under consideration, whilst 10 articles did not [23,28,33,39,45,47,50,55,56,84]. the reported coefficient of correlation was between 0.964 [75] and 1.0 [48,60,76,82]. additionally, y-intercept, the slope of the regression of line, and residual sum of squares can also be used in evaluating linearity [109]. out of 71 articles reviewed, 39 articles reported values for both the slope of regression and the y-intercept. accuracy of a bioanalytical method is normally expressed as the percent recovery by the assay of the known added amount of analyte. in this study, 63 articles reported percent recoveries that ranged mostly between 61.0 % [72] to 114.0 % [5]. two (2) articles reported recoveries of 4.9 % [28] and 36 % [76], respectively. lloq is defined as the amount of analyte in a biological matrix that can be quantitatively determined with suitable precision and accuracy [110]. thirty-six (36) articles reported lloq for aminoglycosides assayed between 1 ng/ml [72] to 2340 ng/ml [79]. forty-one (41) articles also reported lod values in the range 0.3 ng/ml [61] to 75000 ng/ml [56]. there were instances where the same article reported both the lloq and lod. repeatability expresses the closeness of results obtained with the same sample using the same procedure, operators, measuring system, operating conditions, and location over a short period of time. out of 71 articles, 58 reported repeatability values ranging from 0.28 % [69] to 36 % [30]. intermediate precision also refers to laboratory variations such as different days, instruments, and analyses. out of the 71 reviewed articles, 56 reported intermediate precision values ranging between 0.331 % [80] to 19.76 % [52], which is within the acceptable limit of 20 % [111]. admet & dmpk 10(1) (2022) 27-62 bioanalysis of aminoglycozides by hplc doi: http://dx.doi.org/10.5599/admet.1183 45 table 2. performance metrics of hplc method for bioanalysis of aminoglycosides. in te rm e d ia te p re ci si o n ( % ) n .i . 6 6 6 n .i . 3 .6 2 .8 3 .4 < 8 < 8 < 8 4 .8 3 -7 .9 1 (s e ru m ) 3 .0 0 -6 .0 0 (u ri n e ) n .i . n .i . < 2 .0 n .i . r e p e a ta b il it y (% ) n .i . 4 .2 -5 .6 4 .4 -7 .5 3 .9 -5 .1 n .i . n .i . 3 .2 2 .3 < 8 < 8 < 8 3 .7 9 -5 .6 0 (s e ru m ) n .i . (u ri n e ) n .i . n .i . 3 .5 n .i . ll o q (n g /m l) n .i . n .i . n .i . n .i . 5 0 0 1 0 0 0 5 0 0 2 0 0 (s e ru m ) 2 0 0 (u ri n e ) n .i . 2 0 0 n .i . n .i . lo d (n g /m l) n .i . n .i . 5 0 0 n .i . n .i . n .i . n .i . n .i . 1 0 0 0 1 0 0 0 r e co v e ry ( % ) > 9 5 8 0 -1 0 5 n .i . > 9 5 9 3 9 3 9 6 .7 -1 1 0 9 4 .1 -9 8 .3 9 1 .5 -9 1 .8 n .i . (s e ru m ) 8 3 .1 -9 4 .3 (u ri n e ) 8 3 .0 -8 4 .0 n .i . 9 3 n .i . y -i n te rc e p t 1 .3 6 0 .6 2 -0 .0 0 4 1 .3 6 0 .0 4 4 -0 .0 2 4 n .i . -0 .0 0 9 2 6 (s e ru m ) -0 .0 4 0 6 0 (u ri n e ) n .i . n .i . -0 .2 1 – 0 .7 3 -0 .1 0 – 1 .6 0 s lo p e 4 .7 2 0 .8 7 0 .5 6 3 4 .7 2 0 .1 3 6 0 .2 7 8 n .i . 0 .1 1 4 0 (s e ru m ) 0 .3 4 4 0 (u ri n e ) n .i . n .i . 2 .4 9 4 .2 6 1 .9 0 2 .3 3 r 2 0 .9 9 7 0 .9 9 0 .9 9 9 0 .9 9 7 0 .9 9 9 2 0 .9 9 9 7 n .i . 0 .9 9 9 0 (s e ru m ) 0 .9 9 9 0 (u ri n e ) n .i . n .i . 0 .9 9 7 7 0 .9 9 9 9 0 .9 9 6 3 0 .9 9 9 7 r e fe re n ce r a n g e (n g /m l) 5 0 0 0 -1 0 0 0 0 b ) 5 0 0 0 -1 0 0 0 0 b 5 0 0 0 -1 0 0 0 0 b 5 0 0 0 -1 0 0 0 0 b 1 0 0 0 0 -1 6 0 0 0 c) 5 0 0 0 -1 0 0 0 0 b ) 1 5 0 0 0 -2 5 0 0 0 b ) 5 0 0 0 -1 0 0 0 0 b 1 0 0 0 0 -1 6 0 0 0 c) 5 0 0 0 -1 0 0 0 0 b 5 0 0 0 -1 0 0 0 0 b 5 0 0 0 -1 0 0 0 0 b 5 0 0 0 -1 0 0 0 0 b 5 0 0 0 -1 0 0 0 0 b 5 0 0 0 -1 0 0 0 0 b 5 0 0 0 -1 0 0 0 0 b r a n g e ( n g /m l) 1 0 0 0 -1 0 0 0 0 0 -2 0 0 0 0 5 0 0 -1 0 0 0 0 1 0 0 0 -1 0 0 0 0 2 0 0 0 -3 2 0 0 0 2 0 0 0 -1 5 0 0 0 n .i . 1 0 0 0 -7 5 0 0 (s e ru m )a ) 1 0 0 0 -7 5 0 0 (u ri n e )a ) 1 0 0 0 -1 6 0 0 0 0 -1 0 0 0 0 0 -4 0 0 0 0 (p la sm a ) 0 -7 1 0 0 0 (u ri n e ) r e so lu ti o n n .i . n .i . n .i . n .i . n .i . n .i . n .i . n .i . n .i . m a tr ix e ff e ct n .i . n .i . n .i . n .i . n .i . n .i . n .i . n .i . n .i . a n a ly te g e n ta m ic in g e n ta m ic in c 1 g e n ta m ic in c 1 a g e n ta m ic in c 2 n e ti lm ic in g e n ta m ic in a m ik a ci n t o b ra m y ci n n e ti lm ic in t o b ra m y ci n g e n ta m ic in t o b ra m y ci n g e n ta m ic in c 1 a /g e n ta m ic in c 1 + c 2 g e n ta m ic in g e n ta m ic in r e f. [2 0 ] [3 6 ] [2 1 ] [2 2 ] [2 3 ] [3 7 ] [3 8 ] [3 9 ] [4 0 ] http://dx.doi.org/10.5599/admet.1183 seth k. amponsah et al. admet & dmpk 10(1) (2022) 27-62 46 table 2. continued... in te rm e d ia te p re ci si o n ( % ) n .i . < 8 ( s e ru m ) < 8 ( u ri n e ) < 3 .2 3 .7 -5 .8 3 .6 -6 .9 n .i . 4 .6 -5 .1 < 6 .0 < 6 .0 < 6 .0 < 6 .0 < 6 .0 < 6 .0 < 6 .0 < 6 .0 2 .8 -3 .1 r e p e a ta b il it y (% ) n .i . 2 .4 -1 0 .1 (s e ru m ) 2 .4 -1 0 .1 (u ri n e ) < 2 .5 n .i . 1 .5 -5 .3 4 .0 -4 .9 5 .9 n .i . n .i . 4 .9 4 .9 4 .9 n .i . n .i . 3 .5 -6 .0 ll o q (n g /m l) n .i . 5 0 0 (s e ru m ) 5 0 0 (u ri n e ) n .i . n .i . n .i . n .i . n .i . n .i . < 5 0 0 lo d (n g /m l) n .i . n .i . 5 0 0 5 0 0 5 0 0 n .i . < 2 0 0 n .i . n .i . r e co v e ry ( % ) n .i . > 8 5 (s e ru m ) > 8 5 (u ri n e ) 9 7 -1 0 3 8 0 -9 0 8 0 -9 0 7 2 9 4 -9 8 .6 n .i . 9 2 .8 -9 8 .4 y -i n te rc e p t 1 .9 2 n .i . n .i . -0 .1 2 6 n .i . 0 .0 2 n .i . n .i . n .i . n .i . -1 .6 -0 .7 0 .3 n .i . n .i . n .i . s lo p e 3 .2 1 n .i . n .i . 0 .7 9 2 n .i . 0 .9 2 n .i . n .i . n .i . n .i . 1 6 .6 9 .5 7 .2 n .i . n .i . n .i . r 2 0 .9 9 9 6 0 .9 9 9 0 (s e ru m ) 0 .9 9 9 0 (u ri n e ) 0 .9 9 9 0 .9 8 n .i . 0 .9 9 3 0 .9 5 n .i . n .i . n .i . 0 .9 9 9 8 0 .9 9 9 7 0 .9 9 9 8 n .i . n .i . n .i . r e fe re n ce r a n g e (n g /m l) 1 5 0 0 0 -2 5 0 0 0 b ) 5 0 0 0 -1 0 0 0 0 b 5 0 0 0 -1 0 0 0 0 b 5 0 0 0 -1 0 0 0 0 b 5 0 0 0 -1 0 0 0 0 b 1 0 0 0 0 -1 6 0 0 0 c) 1 5 0 0 0 -2 5 0 0 0 b ) 5 0 0 0 -1 0 0 0 0 b 1 5 0 0 0 -2 5 0 0 0 b ) 5 0 0 0 -1 0 0 0 0 b 1 0 0 0 0 -1 6 0 0 0 c) 5 0 0 0 -1 0 0 0 0 b 5 0 0 0 -1 0 0 0 0 b 5 0 0 0 -1 0 0 0 0 b n .a . n .a . 1 5 0 0 0 -2 5 0 0 0 b ) r a n g e ( n g /m l) 2 0 0 0 -6 4 0 0 0 5 0 0 -1 0 0 0 0 (s e ru m ) 5 0 0 -5 0 0 0 (u ri n e ) 2 7 0 0 -1 6 5 0 0 0 -2 0 0 0 0 0 -2 0 0 0 0 1 0 0 0 -6 4 0 0 0 1 0 0 0 -2 5 0 0 0 1 0 0 0 1 6 0 0 0 a ) 2 5 0 0 -5 0 0 0 0 r e so lu ti o n n .i . n .i . n .i . n .i . n .i . n .i . n .i . n .i . m a tr ix e ff e ct n .i . n .i . n .i . n .i . n .i . n .i . n .i . n .i . a n a ly te a m ik a ci n g e n ta m ic in ( c 1 , c 1 a , a n d c 2 ) g e n ta m ic in g e n ta m ic in n e ti lm ic in a m ik a ci n t o b ra m y ci n a m ik a ci n t o b ra m y ci n n e ti lm ic in g e n ta m ic in c 1 g e n ta m ic in c 1 a g e n ta m ic in c 2 d ib e k a ci n s is o m ic in a m ik a ci n r e f. [4 1 ] [4 2 ] [4 3 ] [4 4 ] [4 5 ] [4 6 ] [2 4 ] [4 7 ] admet & dmpk 10(1) (2022) 27-62 bioanalysis of aminoglycozides by hplc doi: http://dx.doi.org/10.5599/admet.1183 47 table 2. continued... in te rm e d ia te p re ci si o n ( % ) 2 .5 2 .8 3 .1 1 .9 2 .3 -2 .5 n .i . n .i . 2 .0 -5 .0 1 .9 -5 .6 < 6 3 .0 1 -3 .5 0 < 4 .0 2 .3 4 -2 .5 8 6 .4 9 -8 .8 7 7 .2 2 -1 9 .7 6 1 0 .3 5 -1 4 .8 2 r e p e a ta b il it y (% ) n .i . 2 .0 -2 .2 n .i . n .i . 0 .8 -2 .9 1 .6 -3 .1 < 6 2 .6 7 -3 .0 2 < 4 .0 1 .2 1 -2 .7 5 6 .0 5 -6 .7 6 5 .1 3 -1 1 .6 6 6 .3 6 -7 .0 2 ll o q (n g /m l) n .i . n .i . n .i . n .i . n .i . n .i . n .i . n .i . n .i . n .i . lo d (n g /m l) n .i . 3 0 0 3 0 0 3 0 0 n .i . 2 0 0 0 5 0 0 5 0 0 1 0 0 2 0 8 0 8 0 r e co v e ry ( % ) ~ 1 0 0 ~ 1 0 0 ~ 1 0 0 ~ 1 0 0 9 7 .5 -1 0 2 n .i . n .i . 9 9 9 8 .5 8 0 1 0 0 1 0 3 .9 9 9 .5 1 0 1 .1 1 0 0 n .i . y -i n te rc e p t 0 .0 0 2 3 0 .0 0 8 5 0 .0 0 7 7 5 0 .0 1 1 7 -0 .0 5 7 0 .0 0 2 0 .0 0 3 -0 .0 0 6 -0 .0 4 -0 .0 2 5 8 0 .0 7 2 n .i . 0 .0 8 6 1 2 .8 4 1 .5 4 .8 s lo p e 0 .0 3 4 0 .1 0 1 0 .2 4 4 0 .0 8 2 0 .1 0 1 0 .1 3 6 0 .1 0 .9 7 9 0 .9 9 8 0 .0 4 7 6 0 .3 2 9 n .i . 0 .1 3 6 3 1 1 .4 1 3 1 .6 1 2 4 .7 r 2 0 .9 9 9 7 0 .9 9 9 8 0 .9 9 9 7 0 .9 9 9 8 1 0 .9 9 9 0 .9 9 3 0 .9 9 9 0 .9 9 9 0 .9 9 9 7 0 .9 9 9 n .i . 0 .9 9 9 0 .9 9 1 0 .9 9 5 0 .9 9 6 r e fe re n ce r a n g e (n g /m l) n .a . 1 0 0 0 0 -1 6 0 0 0 c) n .a . n .a . 5 0 0 0 -1 0 0 0 0 b n .a . 1 0 0 0 0 -1 6 0 0 0 c) 1 5 0 0 0 -2 5 0 0 0 d ) n .a . 1 0 0 0 0 -3 5 0 0 0 e ) 1 0 0 0 0 -3 5 0 0 0 e ) 5 0 0 0 -1 0 0 0 0 b 5 0 0 0 -1 0 0 0 0 b 5 0 0 0 -1 0 0 0 0 b 1 0 0 0 0 -3 5 0 0 0 e ) 5 0 0 0 -1 0 0 0 0 b 5 0 0 0 -1 0 0 0 0 b 5 0 0 0 -1 0 0 0 0 b r a n g e ( n g /m l) 3 2 0 -2 2 8 0 0 1 7 0 -1 1 6 0 0 a ) 1 0 0 -6 3 0 0 1 0 0 0 -3 0 0 0 1 8 5 0 -1 4 8 3 0 1 6 6 0 -1 3 3 3 0 1 6 3 0 1 3 0 6 0 a ) 3 0 0 0 -5 0 0 0 0 5 0 0 -1 0 0 0 0 5 0 0 0 -5 0 0 0 0 5 0 0 0 -5 0 0 0 0 0 -5 0 0 0 0 5 0 0 0 -4 0 0 0 0 1 0 0 0 -2 0 0 0 0 r e so lu ti o n n .i . n .i . n .i . n .i . n .i . n .i . .n i n .i . m a tr ix e ff e ct n .i . n .i . n .i . n .i . n .i . n .i . n .i . n .i . a n a ly te s is o m ic in n e ti lm ic in a st ro m ic in m ic ro n o m ic in t o b ra m y ci n s is o m ic in n e ti lm ic in k a n a m y ci n d ib e k a ci n s tr e p to m y ci n s tr e p to m y ci n g e n ta m ic in c 1 g e n ta m ic in c 1 a g e n ta m ic in c 2 s tr e p to m y ci n g e n ta m ic in c 1 g e n ta m ic in c 1 a g e n ta m ic in c 2 r e f. [2 5 ] [4 8 ] [2 6 ] [2 7 ] [4 9 ] [5 0 ] [5 1 ] [5 2 ] http://dx.doi.org/10.5599/admet.1183 seth k. amponsah et al. admet & dmpk 10(1) (2022) 27-62 48 table 2. continued... in te rm e d ia te p re ci si o n ( % ) 4 .2 -4 .4 n .i . n .i . n .i . 2 .6 -5 .5 (p la sm a ) 2 .9 -7 .5 (u ri n e ) 3 .2 -2 5 .1 2 .8 -3 .8 0 .1 -6 .6 2 .9 -5 .6 n .i . 1 4 .2 3 6 .3 -7 .9 5 .7 -8 .9 5 .0 -5 .5 6 .0 -8 .9 r e p e a ta b il it y (% ) 1 .8 -1 0 .9 7 .9 n .i . n .i . 0 .5 -2 .5 (p la sm a ) 1 .0 -6 .0 (u ri n e ) 3 .9 -2 8 .4 1 .9 -3 .3 0 .1 -2 .1 1 .1 -4 .7 n .i . 9 .2 2 6 .4 -8 .6 5 .9 -6 .0 5 .8 -7 .0 4 .3 -5 .8 ll o q (n g /m l) n .i . n .i . n .i . n .i . 1 0 0 (p la sm a ) 5 0 (u ri n e ) 2 5 5 0 0 1 0 0 0 0 1 0 0 n .i . n .i . n .i . lo d (n g /m l) 5 0 0 < 5 0 6 2 .5 5 0 n .i . n .i . 5 0 0 7 5 0 0 0 2 5 6 0 0 1 0 0 < 5 0 r e co v e ry ( % ) 7 8 -8 1 9 1 .6 9 7 .5 9 4 -1 0 2 4 .9 ± 0 .1 n .i . 9 6 .9 9 2 .1 9 9 .5 -1 0 5 > 9 0 n .i . 9 6 .8 9 9 9 7 .8 9 3 .9 y -i n te rc e p t 0 .0 0 1 3 -1 .4 0 0 0 (d b s ) -0 .0 1 4 0 (w h o le b lo o d ) -0 .6 9 4 n .i . n .i . n .i . n .i . -4 2 4 0 4 n .i . n .i . -2 .5 -3 .7 -4 .7 -3 .3 s lo p e 0 .0 3 5 4 0 .6 0 0 0 (d b s ) 0 .6 6 9 0 (w h o le b lo o d ) 1 6 .7 5 3 n .i . n .i . n .i . n .i . 2 4 5 0 2 0 9 n .i . n .i . 1 1 8 .2 9 2 .9 1 2 0 .7 7 2 .7 r 2 0 .9 9 0 .9 9 5 0 (d b s ) 0 .9 9 5 0 (w h o le b lo o d ) 0 .9 9 9 6 0 .9 8 9 n .i . n .i . n .i . 0 .9 9 9 9 > 0 .9 9 0 0 .9 9 4 1 1 0 .9 9 9 9 0 .9 9 9 9 r e fe re n ce r a n g e (n g /m l) 1 0 0 0 0 -1 6 0 0 0 c) n .a . n .a . 5 1 6 1 -1 0 3 2 3 f) n .a . 1 5 0 0 0 -2 5 0 0 0 b ) 1 5 0 0 0 -2 5 0 0 0 b ) 1 5 0 0 0 -2 5 0 0 0 b ) 1 5 0 0 0 -2 5 0 0 0 b ) 5 0 0 0 -1 0 0 0 0 b 1 0 0 0 0 -1 6 0 0 0 c) 5 0 0 0 -1 0 0 0 0 b 5 0 0 0 -1 0 0 0 0 b 5 0 0 0 -1 0 0 0 0 b 5 0 0 0 -1 0 0 0 0 b r a n g e ( n g /m l) 5 0 0 -4 0 0 0 0 5 0 -5 0 0 0 (d b s ) 5 0 -5 0 0 0 (w h o le b lo o d ) 5 0 0 -5 0 0 0 1 0 0 -5 0 0 0 a ) 1 0 0 -1 0 0 0 0 0 2 5 -2 0 0 0 a ) 5 0 0 -7 5 0 0 0 1 0 0 0 0 5 0 0 0 0 0 1 0 0 -2 0 0 0 a ) 6 2 5 -8 0 0 0 0 n .i . 2 0 0 -5 0 0 0 0 r e so lu ti o n n .i . n .i . n .i . n .i . n .i . n .i . n .i . n .i . n .i . n .i . m a tr ix e ff e ct n .i . n .i . n .i . n .i . n .i . n .i . n .i . n .i . n .i . n .i . n .i . a n a ly te n e ti lm ic in s is o m ic in s is o m ic in n e o m y ci n is e p a m ic in a m ik a ci n a m ik a ci n a m ik a ci n g e n ta m ic in n e ti lm ic in g e n ta m ic in c 1 g e n ta m ic in c 1 a g e n ta m ic in c 2 g e n ta m ic in c 2 a r e f. [6 ] [5 3 ] [5 4 ] [1 4 ] [2 8 ] [5 5 ] [5 6 ] [5 7 ] [5 8 ] [5 9 ] [6 0 ] admet & dmpk 10(1) (2022) 27-62 bioanalysis of aminoglycozides by hplc doi: http://dx.doi.org/10.5599/admet.1183 49 table 2. continued... in te rm e d ia te p re ci si o n ( % ) 0 .0 9 n .i . 7 .4 3 .5 2 .4 2 .5 n .i . 2 .1 6 -3 .6 1 2 .0 -7 .7 (p la sm a ) 1 2 -1 6 (u ri n e ) 4 .8 .2 0 1 3 8 .7 -1 2 (u ri n e ) 6 .3 .2 0 1 2 1 4 -1 6 (u ri n e ) 2 .3 7 4 .2 5 -6 .3 2 r e p e a ta b il it y (% ) 0 .0 2 n .i . 1 .9 -1 5 .0 4 .5 3 .5 3 .5 5 .1 -3 6 .8 7 .3 -3 6 .9 2 .4 1 -2 .6 8 1 .1 -1 1 (p la sm a ) 7 .8 -8 .1 (u ri n e ) 2 .1 -7 .7 6 .1 -1 0 (u ri n e ) 4 .2 .2 0 1 0 3 .1 -4 .2 (u ri n e ) < 5 .8 2 .8 4 -5 .4 4 ll o q (n g /m l) n .i . 1 5 1 0 0 1 0 1 2 8 n .i . 7 0 1 0 0 1 0 0 2 5 0 lo d (n g /m l) 7 8 0 .3 0 .4 n .i . n .i . 8 1 2 1 4 n .i . n .i . n .i . 4 0 0 7 5 r e co v e ry ( % ) n .i . 7 8 -8 8 8 3 1 0 6 .7 9 5 .2 9 9 .1 7 7 .6 -9 6 .1 8 1 .6 -1 0 6 1 0 1 .1 1 0 5 .6 7 2 (p la sm a ) 9 8 ( u ri n e ) 7 2 (p la sm a ) 9 8 ( u ri n e ) 7 2 (p la sm a ) 9 8 ( u ri n e ) > 9 1 9 4 .3 y -i n te rc e p t n .i . n .i . n .i . -0 .3 7 .7 1 3 3 7 .8 1 7 .2 n .i . 0 (p la sm a ) 0 ( u ri n e ) 0 (p la sm a ) 0 ( u ri n e ) 0 (p la sm a ) 0 ( u ri n e ) -0 .0 4 9 8 ± 0 .0 0 5 5 0 .2 4 6 5 s lo p e n .i . n .i . n .i . 1 9 1 .8 2 2 9 .5 1 6 7 .1 0 .0 0 0 0 9 5 0 .0 0 0 3 4 n .i . 3 4 0 (p la sm a ) 4 7 0 (u ri n e ) 1 8 3 (p la sm a ) 2 4 5 (u ri n e ) 1 6 9 (p la sm a ) 2 3 5 (u ri n e ) 0 .0 1 8 0 ± 0 .0 0 0 2 0 .4 9 8 6 r 2 0 .9 9 9 > 0 .9 9 5 0 0 .9 9 0 .9 9 8 6 0 .9 9 8 7 9 9 6 5 0 .9 9 5 5 0 .9 9 0 5 0 .9 9 7 5 0 .9 9 9 0 (p la sm a ) 0 .9 9 9 0 (u ri n e ) 0 .9 9 6 0 (p la sm a ) 0 .9 9 8 0 (u ri n e ) 0 .9 9 8 0 (p la sm a ) 0 .9 9 8 0 (u ri n e ) 0 .9 9 8 0 .9 9 8 r e fe re n ce r a n g e (n g /m l) n .a . 5 0 0 0 -1 0 0 0 0 b n .a . 5 1 6 1 -1 0 3 2 3 f 1 0 0 0 0 -1 6 0 0 0 c) n .a . 1 0 0 0 0 -3 5 0 0 0 e ) n .a . 5 0 0 0 -1 0 0 0 0 b 5 0 0 0 -1 0 0 0 0 b 5 0 0 0 -1 0 0 0 0 b 5 0 0 0 -1 0 0 0 0 b 1 5 0 0 0 -2 5 0 0 0 b ) 5 0 0 0 -1 0 0 0 0 b r a n g e ( n g /m l) 7 8 -1 0 0 0 0 1 5 -6 0 a ) 1 0 0 -1 0 0 0 0 0 1 0 0 -1 0 0 0 0 -2 0 0 0 a ) 2 0 0 -2 0 0 0 0 0 -5 0 0 0 0 0 -5 0 0 0 0 0 -5 0 0 0 0 4 0 0 0 2 0 8 0 0 a ) 5 0 0 -1 0 0 0 0 r e so lu ti o n n .i . n .i . n .i . n .i . n .i . n .i . n .i . n .i . n .i . m a tr ix e ff e ct n .i . n .i . n .i . n .i . n .i . n .i . n .i . n .i . n .i . a n a ly te g e n e ti ci n g e n ta m ic in is e p a m ic in n e o m y ci n n e ti lm ic in s is o m ic in s tr e p to m y ci n d ih y d ro st re p to m y ci n g e n ta m ic in g e n ta m ic in c 1 g e n ta m ic in c 1 a g e n ta m ic in c 2 a m ik a ci n g e n ta m ic in r e f. [2 9 ] [6 1 ] [6 2 ] [6 3 ] [3 0 ] [1 6 ] [6 4 ] [6 5 ] [6 6 ] http://dx.doi.org/10.5599/admet.1183 seth k. amponsah et al. admet & dmpk 10(1) (2022) 27-62 50 table 2. continued... in te rm e d ia te p re ci si o n ( % ) < 5 .2 4 .0 -6 .0 n .i . 6 .6 -7 .6 1 .1 < 9 6 .0 -1 5 .7 4 .6 6 -8 .9 9 r e p e a ta b il it y (% ) < 2 .1 2 .7 -5 .8 0 .5 -9 .4 6 .1 -1 1 .4 0 .5 -1 2 .7 0 .3 -1 1 .6 0 .3 -1 0 .2 0 .4 -9 .9 0 .2 8 -7 .9 4 1 < 9 3 .7 -1 6 .0 n .i . ll o q (n g /m l) n .i . 1 5 0 1 0 0 1 0 0 n .i . 2 0 0 0 1 2 0 0 lo d (n g /m l) 2 3 0 1 0 0 n .i . n .i . 3 0 0 5 0 0 .6 n .i . r e co v e ry ( % ) > 9 9 9 3 -1 0 5 1 0 0 9 4 .4 1 0 2 .3 9 2 .4 1 0 4 .8 9 4 .5 1 0 1 .7 9 4 .5 1 0 4 .1 9 6 .0 1 0 1 .6 1 0 0 8 5 .5 (p la sm a ) 9 0 .9 (u ri n e ) > 9 2 6 1 6 6 .6 y -i n te rc e p t -0 .0 2 1 5 ± 0 .0 0 4 0 0 .0 0 5 n .i . n .i . n .i . n .i . 0 .0 8 3 0 3 0 .0 0 0 7 3 2 0 .0 0 0 2 2 8 0 .0 0 0 5 4 s lo p e 0 .1 0 2 7 ± 0 .0 0 1 1 0 .2 1 n .i . n .i . n .i . n .i . 0 .0 0 3 1 0 1 0 .0 3 7 6 0 .0 3 8 5 0 .0 4 0 3 r 2 0 .9 9 9 9 0 .9 9 8 6 > 0 .9 9 3 0 0 .9 9 8 5 > 0 .9 9 9 2 (p la sm a ) > 0 .9 9 2 (u ri n e ) ≤ 0 .9 9 7 7 0 .9 9 9 5 0 .9 9 6 4 0 .9 9 6 1 0 .9 9 8 2 r e fe re n ce r a n g e (n g /m l) 5 0 0 0 -1 0 0 0 0 b 5 0 0 0 -1 0 0 0 0 b 1 5 0 0 0 -2 5 0 0 0 b ) 5 0 0 0 -1 0 0 0 0 b 1 5 0 0 0 -2 5 0 0 0 d ) 5 1 6 1 -1 0 3 2 3 f n .a . 5 0 0 0 -1 0 0 0 0 b 5 1 6 1 -1 0 3 2 3 f 5 0 0 0 -1 0 0 0 0 b 1 5 0 0 0 -2 5 0 0 0 b ) n .a . 5 1 6 1 -1 0 3 2 3 f r a n g e ( n g /m l) 9 3 0 -9 3 4 0 a ) 0 -5 0 0 0 0 1 0 0 -5 0 0 0 a ) 1 0 0 -5 0 0 0 0 1 0 0 0 -3 8 0 0 0 1 5 0 -2 0 0 0 0 a ) 0 -2 0 0 2 0 0 -5 0 0 0 0 r e so lu ti o n n .i . n .i . n .i . n .i . n .i . n .i . n .i . n .i . m a tr ix e ff e ct n .i . n .i . n .i . n .i . n .i . n .i . n .i . n .i . a n a ly te t o b ra m y ci n t o b ra m y ci n a m ik a ci n g e n ta m ic in k a n a m y ci n n e o m y ci n p a ro m o m yc in t o b ra m y ci n n e o m y ci n t o b ra m y ci n a m ik a ci n d ih y d ro st re p to m y ci n n e o m y ci n (n e o 1 8 0 4 a 7 ) (n e o 1 8 0 4 a 8 ) (n e o 1 8 0 4 a 9 ) r e f. [6 7 ] [6 8 ] [3 1 ] [6 9 ] [7 0 ] [7 1 ] [7 2 ] [7 3 ] admet & dmpk 10(1) (2022) 27-62 bioanalysis of aminoglycozides by hplc doi: http://dx.doi.org/10.5599/admet.1183 51 table 2. continued... in te rm e d ia te p re ci si o n ( % ) n .i . 8 .3 -1 1 .1 n .i . 1 .7 -1 3 .0     5 -1 0 7 -8 6 -9 1 3 -1 7 6 -1 0 1 2 -1 7 1 0 -1 6 8 -1 2 1 0 -1 4 r e p e a ta b il it y (% ) n .i . 5 .6 -1 1 .4 0 .9 7 .5 -1 0 .8     3 -6 4 -8 5 -9 8 -1 1 4 -8 7 -1 4 1 1 -1 4 6 -9 7 -1 3 ll o q (n g /m l) 1 0 0 5 0 5 0 0 3 5 0 6 0 5 0 0 2 5 0 n .i . lo d (n g /m l) 5 0 n .i . 2 6 0 1 0 0 n .i . n .i . 7 0 n .i . r e co v e ry (% ) 9 5 .1 5 9 3 -1 0 5 1 0 0 .3 2 9 9 .9 8 ≤ 1 0 0 3 6 .1 ± 6 .1 > 9 9 7 6 -9 7 7 2 -1 1 3 8 1 -1 0 7 6 9 -9 7 7 8 -1 0 3 7 0 -9 4 6 2 -8 9 6 7 -9 2 7 1 -9 8 y -i n te rc e p t 0 .0 0 0 6 4 9 n .i . -0 .7 7 7 4 3 0 3 -0 .0 6 4 0 .0 0 1 2 ± 0 .0 8 7 3 n .i . s lo p e 0 .0 3 0 3 n .i . 7 .1 8 6 8 .1 1 2 6 3 5 7 0 .0 8 0 .0 6 7 7 ± 0 .0 1 1 8 n .i . r 2 0 .9 9 5 9 0 .9 9 1 2 0 .9 6 4 0 -0 .9 9 0 0 .9 9 9 8 0 .9 9 9 4 0 .9 9 1 0 .9 9 8 9 0 .9 9 9 8 0 .9 9 9 4 0 .9 9 9 0 .9 9 1 6 0 .9 9 8 5 0 .9 9 8 2 0 .9 9 8 1 0 .9 9 5 3 0 .9 9 0 6 r e fe re n ce r a n g e (n g /m l) 1 5 0 0 0 -2 5 0 0 0 b ) 5 0 0 0 -1 0 0 0 0 b 5 0 0 0 -1 0 0 0 0 b 5 1 6 1 -1 0 3 2 3 f 1 5 0 0 0 -2 5 0 0 0 b ) n .a . 5 0 0 0 -1 0 0 0 0 b 5 1 6 1 -1 0 3 2 3 f 1 0 0 0 0 -3 5 0 0 0 e ) n .a . 1 5 0 0 0 -2 5 0 0 0 b ) 1 5 0 0 0 -2 5 0 0 0 d ) n .a . 5 0 0 0 -1 0 0 0 0 b 1 4 4 0 0 0 2 1 0 0 0 0 h ) n .a . r a n g e ( n g /m l) 5 0 -1 0 0 0 0 a ) 5 0 -1 0 0 0 a ) 8 0 0 -4 0 0 0 a ) 5 0 0 -1 0 0 0 0 a ) 6 0 -4 0 0 0 a ) 5 0 0 -5 0 0 0 0 2 5 0 -2 0 0 0 0 n .i . r e so lu ti o n n .i . n .i . 1 .7 1 .4 n .i . n .i . n .i . n .i . m a tr ix e ff e ct n .i . n .i . n .i . n .i . n .i . n .i . a n a ly te (n e o 1 8 0 4 b 4 ) a m ik a ci n t o b ra m y ci n g e n ta m ic in n e o m y ci n a m ik a ci n is e p a m ic in t o b ra m y ci n n e o m y ci n s tr e p to m y ci n d ih y d ro st re p to m y ci n a m ik a ci n k a n a m y ci n p a ro n o m y ci n t o b ra m y ci n s p e ct in o m y ci n a p ra m y ci n r e f. [7 4 ] [7 5 ] [1 5 ] [7 6 ] [7 7 ] [5 ] http://dx.doi.org/10.5599/admet.1183 seth k. amponsah et al. admet & dmpk 10(1) (2022) 27-62 52 table 2. continued... in te rm e d ia te p re ci si o n ( % ) 9 -1 4 7 -1 0 8 -1 1 7 -1 0 < 5 .0 < 1 2 .1 0 < 3 .0 < 3 .0 < 3 .0 n .i . 0 .3 3 1 -0 .7 8 4 1 .8 -1 2 .5 2 .0 -1 1 .0 2 .6 -1 0 .8 2 .8 -9 .3 0 .9 -8 .3 r e p e a ta b il it y (% ) 9 -1 2 7 -9 4 -1 0 5 -8 < 5 .0 < 1 1 .5 6 < 4 .8 < 4 .8 < 4 .8 n .i . 0 .5 7 6 0 .8 0 0 2 .0 -2 .8 1 .4 -3 .2 1 .6 -8 .5 1 .4 -4 .6 3 .2 -6 .9 ll o q (n g /m l) 4 5 0 2 3 4 0 6 3 0 n .i . 1 0 0 1 6 .3 n .i . 5 0 lo d (n g /m l) 1 0 0 5 9 0 3 2 0 1 8 2 1 5 5 n .i . 5 .3 4 n .i . 2 0 r e co v e ry (% ) 7 8 -9 8 8 2 1 0 7 7 6 -1 1 4 7 0 -1 0 5 9 9 .2 0 1 0 3 .1 7 8 5 .2 8 3 .6 9 9 .2 1 0 0 .3 9 9 .4 1 0 1 .2 9 9 .6 1 0 2 .4 9 1 .8 1 0 3 .6 9 8 .3 3 1 0 1 .7 4 1 0 5 .5 1 0 2 .9 1 0 0 1 0 5 .4 8 8 y -i n te rc e p t -0 .2 8 1 5 n .i . 1 .3 0 2 3 .0 2 6 1 .0 9 4 0 .0 0 7 5 1 1 .0 6 5 2 n .i . n .i . s lo p e 1 .1 8 9 6 n .i . 1 5 8 4 .0 6 2 1 8 5 .5 5 2 1 4 5 .9 2 0 .9 7 3 0 .0 2 4 7 ± 0 .0 0 3 n .i . n .i . r 2 0 .9 9 1 8 0 .9 9 9 5 0 .9 9 9 1 0 .9 9 6 7 0 .9 9 9 7 0 .9 9 8 0 .9 9 8 0 .9 9 9 8 0 .9 9 9 7 0 .9 9 9 7 n .i . 0 .9 9 9 0 .9 9 9 9 0 .9 9 8 8 0 .9 9 9 2 0 .9 9 9 4 < 0 .9 9 9 0 r e fe re n ce r a n g e (n g /m l) n .a . 5 0 0 0 -1 0 0 0 0 b 5 0 0 0 -1 0 0 0 0 b 5 0 0 0 -1 0 0 0 0 b n .a . 1 5 0 0 0 -2 5 0 0 0 b ) 5 0 0 0 -1 0 0 0 0 b 1 5 0 0 0 -2 5 0 0 0 b ) 1 0 0 0 0 -1 6 0 0 0 c) n .a . 1 5 0 0 0 -2 0 0 0 0 g ) 5 0 0 0 -1 0 0 0 0 b 1 5 0 0 0 -2 5 0 0 0 b ) 5 0 0 0 -1 0 0 0 0 b 5 0 0 0 -1 0 0 0 0 b 5 0 0 0 -1 0 0 0 0 b 1 4 4 0 0 0 2 1 0 0 0 0 h ) r a n g e ( n g /m l) 6 2 5 -1 5 0 0 0 n .i . 2 5 -8 0 0 0 3 0 -7 4 0 0 0 5 0 -6 3 0 0 0 1 0 0 -4 5 9 0 0 2 0 -2 0 0 a ) 3 0 0 -5 0 0 0 a ) 1 0 0 0 1 0 0 0 0 0 1 0 0 0 1 0 0 0 0 0 1 0 0 0 1 0 0 0 0 0 5 -5 0 0 a ) r e so lu ti o n n .i . n .i . n .i . n .i . n .i . n .i . n .i . m a tr ix e ff e ct n .i . n .i . n .i . _ n .i . n .i . n .i . a n a ly te h y g ro m y ci n g e n ta m ic in ( c 1 ) g e n ta m ic in ( c 1 a ) g e n ta m ic in ( c 2 ) is e p a m ic in a m ik a ci n g e n ta m ic in a m ik a ci n n e ti lm ic in e ti m ic in a rb e k a ci n t o b ra m y ci n a m ik a ci n g e n ta m ic in c 1 g e n ta m ic in c 1 a g e n ta m ic in c 2 s p e ct in o m y ci n r e f. [7 8 ] [7 9 ] [3 2 ] [3 3 ] [8 0 ] [8 1 ] [3 4 ] admet & dmpk 10(1) (2022) 27-62 bioanalysis of aminoglycozides by hplc doi: http://dx.doi.org/10.5599/admet.1183 53 table 2. continued... in te rm e d ia te p re ci si o n ( % ) 4 .3 -7 .5 1 .4 -1 1 .1 1 .2 -6 .7 1 .6 -6 .2 1 .8 -4 .7 1 .3 -6 .2 1 .4 -6 .5 4 .3 -5 .8 4 .7 -8 .3 2 .6 4 -5 .8 0 7 .0 -1 0 .4 3 .8 -5 .6 3 .8 -4 .9 ≤ 2 .3 r 2 , co e ff ic ie n t o f co rr e la ti o n ; lo d , lim it o f d e te ct io n ; ll o q , lo w e r lim it o f q u a n ti fi ca ti o n ; n .i ., n o t in d ic a te d ; n .a ., n o t a v a ila b le ; a ) c a lib ra ti o n r a n g e d o e s n o t co ve r th e t h e ra p e u ti c re fe re n ce r a n g e ; b ) s o u rc e , [1 1 2 ]; c) s o u rc e , [1 1 3 ]; d ) s o u rc e , [1 1 4 ]; e ) s o u rc e , [1 1 5 ]; f) so u rc e , [1 1 6 ] ; g ) s o u rc e , [1 1 7 ]; h ) s o u rc e , [1 1 8 ]. r e p e a ta b il it y (% ) 4 .1 -9 .9 6 .0 -6 .6 4 .0 -5 .0 4 .8 -5 .7 3 .7 -6 .8 4 .3 -5 .1 4 .2 -6 .1 3 .8 -7 .6 2 .7 -6 .0 2 .1 2 -5 .0 7 5 .5 0 -1 1 .9 4 .7 .2 0 0 6 3 .6 -6 .6 ≤ 4 .3 ll o q (n g /m l) 1 2 5 2 5 3 7 .5 1 2 5 1 2 5 5 0 5 0 1 2 5 1 2 5 n .i . 1 2 5 0 5 0 0 5 0 0 5 lo d (n g /m l) 5 0 1 5 1 5 5 0 5 0 3 0 2 0 5 0 5 0 5 0 6 2 5 n .i . n .i . n .i . r e co v e ry (% ) 9 5 8 7 9 2 8 7 9 4 8 9 9 1 9 1 9 3 8 8 .0 2 1 0 2 .5 6 9 8 .9 1 1 3 .7 8 3 .1 -8 9 .7 8 2 .7 1 0 0 y -i n te rc e p t 1 8 9 7 5 2 8 n .i . n .i . n .i . n .i . s lo p e 1 0 9 7 9 3 1 n .i . n .i . n .i . n .i . r 2 1 0 .9 9 0 9 n /i 0 .9 9 0 .9 9 7 r e fe re n ce r a n g e (n g /m l) 5 0 0 0 -1 0 0 0 0 b 5 0 0 0 -1 0 0 0 0 b 1 5 0 0 0 -2 5 0 0 0 d ) n .a . n .a . 1 0 0 0 0 -3 5 0 0 0 e ) n .a . 1 5 0 0 0 -2 5 0 0 0 b ) 5 1 6 1 -1 0 3 2 3 f 1 5 0 0 0 -2 5 0 0 0 b ) 5 1 6 1 -1 0 3 2 3 f 1 5 0 0 0 -2 5 0 0 0 b ) 1 5 0 0 0 -2 5 0 0 0 b ) n .a . r a n g e ( n g /m l) 5 0 0 -1 0 0 0 0 a ) 1 2 5 0 2 0 0 0 0 0 5 0 0 -1 0 0 0 0 0 5 0 0 -1 0 0 0 0 0 5 -1 0 0 0 r e so lu ti o n n .i . n .i . n .i . n .i . n .i . m a tr ix e ff e ct n .i . -0 .1 0 – 1 .3 3 -1 .1 -7 .6 8 .8 ≤ 1 2 a n a ly te t o b ra m y ci n g e n ta m ic in k a n a m y ci n h y g ro m y ci n a p ra m y ci n s tr e p to m y ci n d ih y d ro st re p to m y ci n a m ik a ci n n e o m y ci n a m ik a ci n n e o m y ci n a m ik a ci n a m ik a ci n p a ro m o m yc in r e f. [8 2 ] [8 3 ] [8 4 ] [8 5 ] [3 5 ] http://dx.doi.org/10.5599/admet.1183 seth k. amponsah et al. admet & dmpk 10(1) (2022) 27-62 54 conclusion and outlook despite reported nephrotoxic and ototoxic potentials of aminoglycosides, their use in clinical settings remains relevant. the current study sought to review bioanalytical methods (specifically liquid chromatography) used in the assay of aminoglycosides in biological matrices. in all, 71 articles were reviewed, and 66 of these articles reported the use of reversed-phase liquid chromatography as a bioanalytical method [7,119]. the commonest sample treatment procedures adopted in the analysis of aminoglycosides using hplc were protein precipitation (50 %) and solid phase extraction (39 %). surprisingly, none of the current sample preparation methods was used by any of the articles reported in this review. it will be interesting if recent sample preparation methods such as solid-phase microextraction, micro-solid-phase extraction, dispersive micro-solid-phase extraction, magnetic solid-phase extraction, microextraction by packed sorbent, stir bar sorptive extraction, spin column extraction, liquid-phase microextraction, single-drop microextraction, hollow fiber liquid-phase microextraction, dispersive liquid-liquid microextraction, molecularly imprinted solid-phase extraction, and molecularly imprinted solid-phase micro-extraction [120122] could be applied for bioanalysis of aminoglycosides with the potential of improving method sensitivity and selectivity. fluorescence (50 %), uv (24 %), and mass spectrometry (21 %) were the most adopted mode of detection in the assay of aminoglycosides, according to this review. since mass spectrometry has been established as the detection mode of choice for bioanalysis of aminoglycoside using liquid chromatography in recent years, it is strongly recommended for use except in resource-challenged countries where fluorescence or uv detection methods can be applied after derivatization. there is the need to establish a therapeutic reference range for all the clinically reported 20 aminoglycosides since the calibration range of analytical methods for the bioanalysis of aminoglycosides should cover such a range. it was quite surprising that some of the calibration range was outside the established therapeutic reference range. it is recommended that future liquid chromatography methods for the analysis of aminoglycosides should have calibration ranges covering established reference therapeutic ranges. although this review is not a systematic one, the information provided is intended to serve as a comprehensive reference for related research that may involve the assay of aminoglycosides (pharmacokinetic or drug monitoring studies). conflict of interest: the authors declare no conflict of interest. references [1] r. hari, s. taherunnisa, s.y. raut, s. mutalik, k.b. koteshwara. challenges in the development of analytical test procedure for aminoglycosides: a critical review. j. appl. pharm. sci 9 (2019) 145-152. https://doi.org/10.7324/japs.2019.91118. [2] k. chan, w. wang, k.r. ledesma, t. yin, v.h. tam. a robust lc–ms/ms method for amikacin: application to cellular uptake and pharmacokinetic studies. bioanalysis 12 (2020) 445-454. https://doi.org/10.4155/bio-2020-0007. [3] m.-p. mingeot-leclercq, y. glupczynski, p.m. tulkens. aminoglycosides: activity and resistance. antimicrobial agents chemotherapy 43 (1999) 727-737. https://doi.org/10.1128/aac.43.4.727. [4] l. šoltés. aminoglycoside antibiotics-two decades of their hplc bioanalysis. biomedical chromatography 13 (1999) 3-10. https://doi.org/10.1002/(sici)1099-0801(199902)13:1<3::aidbmc811>3.0.co;2-t. https://doi.org/10.7324/japs.2019.91118 https://doi.org/10.4155/bio-2020-0007 https://doi.org/10.1128/aac.43.4.727 https://doi.org/10.1002/(sici)1099-0801(199902)13:1%3c3::aid-bmc811%3e3.0.co;2-t https://doi.org/10.1002/(sici)1099-0801(199902)13:1%3c3::aid-bmc811%3e3.0.co;2-t admet & dmpk 10(1) (2022) 27-62 bioanalysis of aminoglycozides by hplc doi: http://dx.doi.org/10.5599/admet.1183 55 [5] w.-x. zhu, j.-z. yang, w. wei, y.-f. liu, s.-s. zhang. simultaneous determination of 13 aminoglycoside residues in foods of animal origin by liquid chromatography–electrospray ionization tandem mass spectrometry with two consecutive solid-phase extraction steps. journal of chromatography a 1207 (2008) 29-37. https://doi.org/10.1016/j.chroma.2008.08.033. [6] s. dionisotti, f. bamonte, m. gamba, e. ongini. high-performance liquid chromatographic determination of netilmicin in guinea-pig and human serum by fluorodinitrobenzene derivatization with spectrophotometric detection. journal of chromatography b: biomedical sciences 434 (1988) 169-176. https://doi.org/10.1016/0378-4347(88)80071-9. [7] a. cabanes, y. cajal, i. haro, j.g. anton, f. reig, m. arboix. gentamicin determination in biological fluids by hplc, using tobramycin as internal standard. journal of liquid chromatography 14 (1991) 1989-2010. https://doi.org/10.1080/01483919108049669. [8] s.k. amponsah, g.o. adjei, c. enweronu-laryea, k.a. bugyei, k. hadji-popovski, j.a.l. kurtzhals, k. kristensen. population pharmacokinetic characteristics of amikacin in suspected cases of neonatal sepsis in a low-resource african setting: a prospective nonrandomized single-site study. curr ther res clin exp 84 (2017) e1-e6. https://doi.org/10.1016/j.curtheres.2017.01.001. [9] s.k. amponsah, k.f. opuni, k.a. antwi, v.p. kunkpeh. effect of aminophylline on the pharmacokinetics of amikacin in sprague-dawley rats. j infect dev ctries 13 (2019) 251-254. https://doi.org/10.3855/jidc.10514. [10] s.k. amponsah, k.f.m. opuni, a.a. donkor. animal model investigation suggests betamethasone alters the pharmacokinetics of amikacin. admet and dmpk 6 (2018) 279-283. http://dx.doi.org/10.5599/admet.613. [11] s.k. banerjee, a. wells, a. dasgupta. operator error: a critical determinant of false amikacin and tobramycin concentrations using fluorescence polarization immunoassay kits and tdx analyzer. therapeutic drug monitoring 21 (1999) 540. https://doi.org/10.1097/00007691-199910000-00009. [12] s.k. maitra, t.t. yoshikawa, c.m. steyn, l.b. guze, m.c. schotz. amikacin assay in serum by highperformance liquid chromatography. antimicrobial agents chemotherapy 14 (1978) 880-885. https://doi.org/10.1128/aac.14.6.880. [13] a. marzo, l. dal bo. chromatography as an analytical tool for selected antibiotic classes: a reappraisal addressed to pharmacokinetic applications. journal of chromatography a 812 (1998) 17-34. https://doi.org/10.1016/s0021-9673(98)00282-9. [14] v.k. agarwal. high performance liquid chromatographic determination of neomycin in milk using a hisep column. journal of liquid chromatography 13 (1990) 2475-2487. https://doi.org/10.1080/01483919008049047. [15] g. brajanoski, j. hoogmartens, k. allegaert, e. adams. determination of amikacin in cerebrospinal fluid by high-performance liquid chromatography with pulsed electrochemical detection. j chromatogr b analyt technol biomed life sci 867 (2008) 149-152. https://doi.org/10.1016/j.jchromb.2008.03.016. [16] a. yusuf, s. al-rawithi, d. raines, h. frayha, a. toonsi, i. al-mohsen, a. el-yazigi. simplified highperformance liquid chromatographic method for the determination of gentamicin sulfate in a microsample of plasma: comparison with fluorescence polarization immunoassay. therapeutic drug monitoring 21 (1999) 647. https://doi.org/10.1097/00007691-199912000-00012. [17] y. al-shamalii, y.m. ali, r.a. al-shamalii, m. al-melahi, f.r. al-shammari, a. alsaber, w. al-adsani. don't close the book on tocilizumab for the treatment of severe covid-19 pneumoniathe jury is still outthe kuwait experience. plos one 16 (2021). https://doi.org/10.1371/journal.pone.0254379. [18] r. hari, s. taherunnisa, s.y. raut, s. mutalik, k.b. koteshwara. challenges in the development of analytical test procedure for aminoglycosides: a critical review. j. appl. pharm. sci 9 (2019) 145-152. https://doi.org/10.7324/japs.2019.91118. [19] x. zhang, j. wang, q. wu, l. li, y. wang, h. yang. determination of kanamycin by high performance liquid chromatography. molecules 24 (2019). https://doi.org/10.3390/molecules24101902. http://dx.doi.org/10.5599/admet.1183 https://doi.org/10.1016/j.chroma.2008.08.033 https://doi.org/10.1016/0378-4347(88)80071-9 https://doi.org/10.1080/01483919108049669 https://doi.org/10.1016/j.curtheres.2017.01.001 https://doi.org/10.3855/jidc.10514 http://dx.doi.org/10.5599/admet.613 http://dx.doi.org/10.5599/admet.613 https://doi.org/10.1097/00007691-199910000-00009 https://doi.org/10.1128/aac.14.6.880 https://doi.org/10.1016/s0021-9673(98)00282-9 https://doi.org/10.1080/01483919008049047 https://doi.org/10.1080/01483919008049047 https://doi.org/10.1016/j.jchromb.2008.03.016 https://doi.org/10.1016/j.jchromb.2008.03.016 https://doi.org/10.1097/00007691-199912000-00012 https://doi.org/10.1371/journal.pone.0254379 https://doi.org/10.7324/japs.2019.91118 https://doi.org/10.3390/molecules24101902 seth k. amponsah et al. admet & dmpk 10(1) (2022) 27-62 56 [20] j.p. anhalt. assay of gentamicin in serum by high-pressure liquid chromatography. antimicrobial agents chemotherapy 11 (1977) 651-655. https://doi.org/10.1128/aac.11.4.651. [21] g.w. peng, g.g. jackson, w.l. chiou. high-pressure liquid chromatographic assay of netilmicin in plasma. antimicrobial agents chemotherapy 12 (1977) 707-709. https://doi.org/10.1128/aac.12.6.707. [22] j.p. anhalt, s.d. brown. high-performance liquid-chromatographic assay of aminoglycoside antibiotics in serum. clinical chemistry 24 (1978) 1940-1947. https://doi.org/10.1093/clinchem/24.11.1940. [23] s.-e. bäck, i. nilsson-ehle, p. nilsson-ehle. chemical assay, involving liquid chromatography, for aminoglycoside antibiotics in serum. clinical chemistry 25 (1979) 1222-1225. https://doi.org/10.1093/clinchem/25.7.1222. [24] l. essers. an automated high-performance liquid chromatographic method for the determination of aminoglycosides in serum using pre-column sample cleanup and derivatization. journal of chromatography b: biomedical sciences 305 (1984) 345-352. https://doi.org/10.1016/s03784347(00)83348-4. [25] t. kawamoto, i. mashimo, s. yamauchi, m. watanabe. determination of sisomicin, netilmicin, astromicin and micronomicin in serum by high-performance liquid chromatography. journal of chromatography b: biomedical sciences 305 (1984) 373-379. https://doi.org/10.1016/s03784347(00)83351-4. [26] h. kubo, y. kobayashi, t. nishikawa. rapid method for determination of kanamycin and dibekacin in serum by use of high-pressure liquid chromatography. antimicrobial agents chemotherapy 28 (1985) 521-523. https://doi.org/10.1128/aac.28.4.521. [27] n. kurosawa, s. kuribayashi, e. owada, k. ito, m. nioka, m. arakawa, r.j.j.o.c.b.b.s. fukuda, applications. determination of streptomycin in serum by high-performance liquid chromatography. journal of chromatography b: biomedical sciences 343 (1985) 379-385. https://doi.org/10.1016/s0378-4347(00)84606-x. [28] j.a. maloney, w.m. awni. high-performance liquid chromatographic determination of isepamicin in plasma, urine and dialysate. journal of chromatography b: biomedical sciences 526 (1990) 487-496. https://doi.org/10.1016/s0378-4347(00)82530-x. [29] c. bethune, t. bui, m.l. liu, m.a. kay, r.j. ho. development of a high-performance liquid chromatographic assay for g418 sulfate (geneticin). antimicrob agents chemother 41 (1997) 661664. https://doi.org/10.1128/aac.41.3.661. [30] g. suhren, k. knappstein. detection of incurred dihydrostreptomycin residues in milk by liquid chromatography and preliminary confirmation methods. analyst 123 (1998) 2797-2801. https://doi.org/10.1039/a805050c. [31] r. oertel, v. neumeister, w. kirch. hydrophilic interaction chromatography combined with tandemmass spectrometry to determine six aminoglycosides in serum. journal of chromatography a 1058 (2004) 197-201. https://doi.org/10.1016/j.chroma.2004.08.158. [32] l. zhang, j. peng, j. tang, b. yuan, r. he, y. xiao. description and validation of coupling high performance liquid chromatography with resonance rayleigh scattering in aminoglycosides determination. analytica chimica acta 706 (2011) 199-204. https://doi.org/10.1016/j.aca.2011.09.025. [33] a.r. breaud, c.l. henemyre-harris, s. schools, n. emezienna, w. clarke. rapid quantification of the aminoglycoside arbekacin in serum using high performance liquid chromatography–tandem mass spectrometry. j clinica chimica acta 418 (2013) 102-106. https://doi.org/10.1016/j.cca.2013.01.004. [34] j.b. arsand, l. jank, m.t. martins, r.b. hoff, f. barreto, t.m. pizzolato, c. sirtori. determination of aminoglycoside residues in milk and muscle based on a simple and fast extraction procedure followed by liquid chromatography coupled to tandem mass spectrometry and time of flight mass spectrometry. talanta 154 (2016) 38-45. https://doi.org/10.1016/j.talanta.2016.03.045. https://doi.org/10.1128/aac.11.4.651 https://doi.org/10.1128/aac.12.6.707 https://doi.org/10.1128/aac.12.6.707 https://doi.org/10.1093/clinchem/24.11.1940 https://doi.org/10.1093/clinchem/24.11.1940 https://doi.org/10.1093/clinchem/25.7.1222 https://doi.org/10.1093/clinchem/25.7.1222 https://doi.org/10.1016/s0378-4347(00)83348-4 https://doi.org/10.1016/s0378-4347(00)83348-4 https://doi.org/10.1016/s0378-4347(00)83351-4 https://doi.org/10.1016/s0378-4347(00)83351-4 https://doi.org/10.1128/aac.28.4.521 https://doi.org/10.1016/s0378-4347(00)84606-x https://doi.org/10.1016/s0378-4347(00)84606-x https://doi.org/10.1016/s0378-4347(00)82530-x https://doi.org/10.1128/aac.41.3.661 https://doi.org/10.1039/a805050c https://doi.org/10.1016/j.chroma.2004.08.158 https://doi.org/10.1016/j.aca.2011.09.025 https://doi.org/10.1016/j.aca.2011.09.025 https://doi.org/10.1016/j.cca.2013.01.004 https://doi.org/10.1016/j.talanta.2016.03.045 admet & dmpk 10(1) (2022) 27-62 bioanalysis of aminoglycozides by hplc doi: http://dx.doi.org/10.5599/admet.1183 57 [35] i.c. roseboom, b. thijssen, h. rosing, j. mbui, j.h. beijnen, t.p. dorlo. highly sensitive uplc-ms/ms method for the quantification of paromomycin in human plasma. journal of pharmaceutical biomedical analysis 185 (2020) 113245. https://doi.org/10.1016/j.jpba.2020.113245. [36] s.k. maitra, t.t. yoshikawa, j.-l. hansen, i. nilsson-ehle, w. palin, m.-c. schotz, l. guze. serum gentamicin assay by high-performance liquid chromatography. j clinical chemistry 23 (1977) 22752278. https://doi.org/10.1093/clinchem/23.12.2275. [37] d.b. haughey, d.m. janicke, m. adelman, j.j. schentag. high-pressure liquid chromatography analysis and single-dose disposition of tobramycin in human volunteers. antimicrob agents chemother 17 (1980) 649-653. https://doi.org/10.1128/aac.17.4.649. [38] d.m. barends, j.s. van der sandt, a. hulshoff. micro determination of gentamicin in serum by highperformance liquid chromatography with ultraviolet detection. j chromatogr 182 (1980) 201-210. https://doi.org/10.1016/s0378-4347(00)81624-2. [39] n.-e. larsen, k. marinelli, a.m. heilesen. determination of gentamicin in serum using liquid column chromatography. journal of chromatography b: biomedical sciences 221 (1980) 182-187. https://doi.org/10.1016/s0378-4347(00)81023-3. [40] s. walker, p. coates, applications. high-performance liquid chromatographic method for determination of gentamicin in biological fluids. journal of chromatography b: biomedical sciences 223 (1981) 131-138. https://doi.org/10.1016/s0378-4347(00)80075-4. [41] l.t. wong, a.r. beaubien, a.p. pakuts, applications. determination of amikacin in microlitre quantities of biological fluids by high-performance liquid chromatography using 1-fluoro-2, 4dinitrobenzene derivatization. journal of chromatography b: biomedical sciences 231 (1982) 145154. https://doi.org/10.1016/s0378-4347(00)80518-6. [42] j. d'souza, r.i. ogilvie. determination of gentamicin components c1a, c2 and c1 in plasma and urine by high-performance liquid chromatography. j chromatogr 232 (1982) 212-218. https://doi.org/10.1016/s0378-4347(00)86029-6. [43] h. kubo, t. kinoshita, y. kobayashi, k. tokunaga. micro determination of gentamicin in serum by high-performance liquid chromatography. journal of chromatography b: biomedical sciences 227 (1982) 244-248. https://doi.org/10.1016/s0378-4347(00)81624-2. [44] j. marples, m. oates. serum gentamicin, netilmicin and tobramycin assays by high performance liquid chromatography. journal of antimicrobial chemotherapy 10 (1982) 311-318. https://doi.org/10.1093/jac/10.4.311. [45] d. barends, j. blauw, m. smits, a. hulshoff. determination of amikacin in serum by high-performance liquid chromatography with ultraviolet detection. journal of chromatography b: biomedical sciences applications 276 (1983) 385-394. https://doi.org/10.1016/s0378-4347(00)85105-1. [46] p.m. kabra, p.k. bhatnagar, m.a. nelson, j.h. wall, l.j. marton. liquid-chromatographic determination of tobramycin in serum with spectrophotometric detection. j clinical chemistry 29 (1983) 672-674. https://doi.org/10.1093/jat/7.6.283. [47] p.m. kabra, p.k. bhatnager, m.a. nelson. liquid chromatographic determination of amikacin in serum with spectrophotometric detection. journal of chromatography b: biomedical sciences 307 (1984) 224-229. https://doi.org/10.1016/s0378-4347(00)84093-1. [48] h. kubo, t. kinoshita, y. kobayashi, k. tokunaga. micro-scale method for determination of tobramycin in serum using high-performance liquid chromatography. journal of liquid chromatography 7 (1984) 2219-2228. https://doi.org/10.1080/01483918408068871. [49] h. kubo, y. kobayashi, t. kinoshita. fluorescence determination of streptomycin in serum by reversed-phase ion-pairing liquid chromatography. analytical chemistry insights 58 (1986) 26532655. https://doi.org/10.1021/ac00126a017. [50] g. tamai, h. imai, h. yoshida. on-line deproteinization of serum sample for hplc analysis of hydrophilic compounds and its application to gentamicin. chromatographia 21 (1986) 519-522. https://doi.org/10.1007/bf02310539. http://dx.doi.org/10.5599/admet.1183 https://doi.org/10.1016/j.jpba.2020.113245 https://doi.org/10.1093/clinchem/23.12.2275 https://doi.org/10.1128/aac.17.4.649 https://doi.org/10.1016/s0378-4347(00)81624-2 https://doi.org/10.1016/s0378-4347(00)81023-3 https://doi.org/10.1016/s0378-4347(00)80075-4 https://doi.org/10.1016/s0378-4347(00)80518-6 https://doi.org/10.1016/s0378-4347(00)86029-6 https://doi.org/10.1016/s0378-4347(00)86029-6 https://doi.org/10.1016/s0378-4347(00)81624-2 https://doi.org/10.1093/jac/10.4.311 https://doi.org/10.1093/jac/10.4.311 https://doi.org/10.1016/s0378-4347(00)85105-1 https://doi.org/10.1093/jat/7.6.283 https://doi.org/10.1016/s0378-4347(00)84093-1 https://doi.org/10.1080/01483918408068871 https://doi.org/10.1021/ac00126a017 https://doi.org/10.1007/bf02310539 seth k. amponsah et al. admet & dmpk 10(1) (2022) 27-62 58 [51] h. kubo, h. li, y. kobayashi, t. kinoshita. fluorometric determination of streptomycin in serum by high-performance liquid chromatography using mobile phase containing fluorogenic reagent. analytical biochemistry 162 (1987) 219-223. https://doi.org/10.1016/0003-2697(87)90030-3. [52] r. rumble, m. roberts. high-performance liquid chromatographic assay of the major components of gentamicin in serum. journal of chromatography b: biomedical sciences 419 (1987) 408-413. https://doi.org/10.1016/0378-4347(87)80308-0. [53] t. fujimoto, r. tawa, s. hirose. fluorometric determination of sisomicin, an aminoglycoside antibiotic, in dried blood spots on filter paper by reversed-phase high-performance liquid chromatography with pre-column derivatization. chemical pharmaceutical bulletin 36 (1988) 15711574. https://doi.org/10.1248/cpb.36.1571. [54] h. matsunaga, t. fujimoto, r. tawa, s. hirose. an on-line cleanup procedure for large sample volume analysis of serum aminoglycoside antibiotics by reversed-phase high-performance liquid chromatography. chemical j pharmaceutical methods 36 (1988) 1565-1570. https://doi.org/10.1016/s0021-9673(98)00342-2. [55] b. wichert, h. schreier, h. derendorf. sensitive liquid chromatography assay for the determination of amikacin in human plasma. journal of pharmaceutical biomedical analysis 9 (1991) 251-254. https://doi.org/10.1016/0731-7085(91)80154-2. [56] e.a. papp, c.a. knupp, r.h. barbhaiya. high-performance liquid chromatographic assays for the quantification of amikacin in human plasma and urine. journal of chromatography b: biomedical sciences 574 (1992) 93-99. https://doi.org/10.1016/0378-4347(92)80102-v. [57] f. sar, p. leroy, a. nicolas, p. archimbault, g. ambroggi. determination of amikacin in dog plasma by reversed-phase ion-pairing liquid chromatography with post-column derivatization. analytical letters 25 (1992) 1235-1250. https://doi.org/10.1080/00032719208016125. [58] m.a. fennell, c.e. uboh, r.w. sweeney, l.r. soma. gentamicin in tissue and whole milk: an improved method for extraction and cleanup of samples for quantitation on hplc. journal of agricultural food chemistry 43 (1995) 1849-1852. https://doi.org/10.1021/jf00055a018. [59] m. santos, e. garcia, f. lópez, j. lanao, a. dominguez-gil. determination of netilmicin in plasma by hplc. journal of pharmaceutical biomedical analysis 13 (1995) 1059-1062. https://doi.org/10.1016/0731-7085(95)01331-e. [60] d. stead, r. richards. sensitive fluorimetric determination of gentamicin sulfate in biological matrices using solid-phase extraction, pre-column derivatization with 9-fluorenylmethyl chloroformate and reversed-phase high-performance liquid chromatography. journal of chromatography b: biomedical sciences 675 (1996) 295-302. https://doi.org/10.1016/0378-4347(95)00355-x. [61] p.j. kijak, j. jackson, b. shaikh. determination of gentamicin in bovine milk using liquid chromatography with post-column derivatization and fluorescence detection. journal of chromatography b: biomedical sciences 691 (1997) 377-382. https://doi.org/10.1016/s03784347(96)00445-8. [62] c.-c. lin, e. radwanski, c. korduba, m. affrime, m.n. cayen. pharmacokinetics of intramuscularly administered isepamicin in man. chemotherapy 43 (1997) 86-93. https://doi.org/10.1159/000239541. [63] d. stead, r. richards, applications. sensitive high-performance liquid chromatographic assay for aminoglycosides in biological matrices enables the direct estimation of bacterial drug uptake. journal of chromatography b: biomedical sciences 693 (1997) 415-421. https://doi.org/10.1016/s03784347(97)00032-7. [64] n. isoherranen, s. soback. determination of gentamicins c1, c1a, and c2 in plasma and urine by hplc. j clinical chemistry 46 (2000) 837-842. https://doi.org/10.1093/clinchem/46.6.837. [65] c. feng, s. lin, h. wu, s. chen. trace analysis of amikacin in human plasma by high-performance liquid chromatography. chromatographia 53 (2001) s213-s217. https://doi.org/10.1007/bf02490330. https://doi.org/10.1016/0003-2697(87)90030-3 https://doi.org/10.1016/0378-4347(87)80308-0 https://doi.org/10.1248/cpb.36.1571 https://doi.org/10.1016/s0021-9673(98)00342-2 https://doi.org/10.1016/s0021-9673(98)00342-2 https://doi.org/10.1016/0731-7085(91)80154-2 https://doi.org/10.1016/0378-4347(92)80102-v https://doi.org/10.1080/00032719208016125 https://doi.org/10.1021/jf00055a018 https://doi.org/10.1016/0731-7085(95)01331-e https://doi.org/10.1016/0731-7085(95)01331-e https://doi.org/10.1016/0378-4347(95)00355-x https://doi.org/10.1016/s0378-4347(96)00445-8 https://doi.org/10.1016/s0378-4347(96)00445-8 https://doi.org/10.1159/000239541 https://doi.org/10.1159/000239541 https://doi.org/10.1016/s0378-4347(97)00032-7 https://doi.org/10.1016/s0378-4347(97)00032-7 https://doi.org/10.1093/clinchem/46.6.837 https://doi.org/10.1007/bf02490330 https://doi.org/10.1007/bf02490330 admet & dmpk 10(1) (2022) 27-62 bioanalysis of aminoglycozides by hplc doi: http://dx.doi.org/10.5599/admet.1183 59 [66] a.i. al-amoud, b.j. clark, h. chrystyn. determination of gentamicin in urine samples after inhalation by reversed-phase high-performance liquid chromatography using pre-column derivatisation with ophthalaldehyde. journal of chromatography b 769 (2002) 89-95. https://doi.org/10.1016/s15700232(01)00636-5. [67] c.-h. feng, s.-j. lin, h.-l. wu, s.-h. chen. trace analysis of tobramycin in human plasma by derivatization and high-performance liquid chromatography with ultraviolet detection. journal of chromatography b 780 (2002) 349-354. https://doi.org/10.1016/s1570-0232(02)00544-5. [68] b.g. keevil, s.j. lockhart, d.p. cooper. determination of tobramycin in serum using liquid chromatography–tandem mass spectrometry and comparison with a fluorescence polarisation assay. journal of chromatography b 794 (2003) 329-335. https://doi.org/10.1016/s1570-0232(03)00492-6. [69] r. oertel, u. renner, w. kirch. determination of neomycin by lc–tandem mass spectrometry using hydrophilic interaction chromatography. journal of pharmaceutical biomedical analysis 35 (2004) 633-638. https://doi.org/10.1016/j.jpba.2004.01.018. [70] n.c. megoulas, m.a. koupparis. development and validation of a novel hplc/elsd method for the direct determination of tobramycin in pharmaceuticals, plasma, and urine. anal bioanal chem 382 (2005) 290-296. https://doi.org/10.1007/s00216-004-2948-8. [71] j.m. serrano, m. silva. determination of amikacin in body fluid by high-performance liquidchromatography with chemiluminescence detection. journal of chromatography b 843 (2006) 20-24. https://doi.org/10.1016/j.jchromb.2006.05.016. [72] m. cherlet, s. de baere, p. de backer. quantitative determination of dihydrostreptomycin in bovine tissues and milk by liquid chromatography‐electrospray ionization‐tandem mass spectrometry. journal of mass spectrometry 42 (2007) 647-656. https://doi.org/10.1002/jms.1194. [73] d.g. mascher, c.p. unger, h. mascher. determination of neomycin and bacitracin in human or rabbit serum by hplc–ms/ms. journal of pharmaceutical biomedical analysis 43 (2007) 691-700. https://doi.org/10.1016/j.jpba.2006.08.008. [74] a.a. al-majed. a new lc method for determination of some aminoglycoside antibiotics in dosage forms and human plasma using 7-fluoro-4-nitrobenz-2-oxa-1, 3-diazole as a fluorogenic pre-column label. chromatographia 68 (2008) 927-934. https://doi.org/10.1365/s10337-008-0792-2. [75] m.e. attema-de jonge, j.m. bekkers, h.m. oudemans-van straaten, r.w. sparidans, e.j. franssen. simple and sensitive method for quantification of low tobramycin concentrations in human plasma using hplc–ms/ms. journal of chromatography b 862 (2008) 257-262. https://doi.org/10.1016/j.jchromb.2007.12.008. [76] s. hosokawa, k. nakamura, y. fujita, r. horiuchi, k. yamamoto. determination of isepamicin in human plasma by hplc with fluorescence detection after derivatization using 6-aminoquinolyl-nhydroxysuccinimidyl-carbamate. pharmaceutical methods 31 (2008) 1866-1869. https://doi.org/10.1248/bpb.31.1866. [77] m. mashat, h. chrystyn, b. clark, k. assi. development and validation of hplc method for the determination of tobramycin in urine samples post-inhalation using pre-column derivatisation with fluorescein isothiocyanate. journal of chromatography b 869 (2008) 59-66. https://doi.org/10.1016/j.jchromb.2008.05.012. [78] x. chang, j. peng. lc analysis of isepamicin in plasma samples post-inhalation with fluorescence detection and its application to a pharmacokinetic study. chromatographia 70 (2009) 1429-1433. https://doi.org/10.1365/s10337-009-1331-5. [79] l. baietto, a. d’avolio, f.g. de rosa, s. garazzino, m. michelazzo, g. ventimiglia, m. siccardi, m. simiele, m. sciandra, g. di perri. development and validation of a simultaneous extraction procedure for hplc-ms quantification of daptomycin, amikacin, gentamicin, and rifampicin in human plasma. analytical bioanalytical chemistry 396 (2010) 791-798. https://doi.org/10.1007/s00216-009-3263-1. [80] a.a. el-zaher, m.a. mahrouse. utility of experimental design in pre-column derivatization for the analysis of tobramycin by hplc-fluorescence detection: application to ophthalmic solution and human plasma. analytical chemistry insights 8 (2013) aci. s11612. https://doi.org/10.4137/aci.s11612. http://dx.doi.org/10.5599/admet.1183 https://doi.org/10.1016/s1570-0232(01)00636-5 https://doi.org/10.1016/s1570-0232(01)00636-5 https://doi.org/10.1016/s1570-0232(02)00544-5 https://doi.org/10.1016/s1570-0232(03)00492-6 https://doi.org/10.1016/j.jpba.2004.01.018 https://doi.org/10.1007/s00216-004-2948-8 https://doi.org/10.1016/j.jchromb.2006.05.016 https://doi.org/10.1002/jms.1194 https://doi.org/10.1016/j.jpba.2006.08.008 https://doi.org/10.1365/s10337-008-0792-2 https://doi.org/10.1016/j.jchromb.2007.12.008 https://doi.org/10.1016/j.jchromb.2007.12.008 https://doi.org/10.1248/bpb.31.1866 https://doi.org/10.1248/bpb.31.1866 https://doi.org/10.1016/j.jchromb.2008.05.012 https://doi.org/10.1016/j.jchromb.2008.05.012 https://doi.org/10.1365/s10337-009-1331-5 https://doi.org/10.1007/s00216-009-3263-1 https://doi.org/10.4137/aci.s11612 https://doi.org/10.4137/aci.s11612 seth k. amponsah et al. admet & dmpk 10(1) (2022) 27-62 60 [81] y. bijleveld, t. de haan, j. toersche, s. jorjani, j. van der lee, f. groenendaal, p. dijk, a. van heijst, a.w. gavilanes, r. de jonge. a simple quantitative method analysing amikacin, gentamicin, and vancomycin levels in human newborn plasma using ion-pair liquid chromatography/tandem mass spectrometry and its applicability to a clinical study. journal of chromatography b 951 (2014) 110118. https://doi.org/10.1016/j.jchromb.2014.01.035. [82] b. chauhan, s. jalalpure. analysis of amikacin in human serum by uhplc with fluorescence detector using chloro-formate reagent with glycine. pharmaceutical methods 7 (2016) 99-103. https://doi.org/10.5530/phm.2016.7.15. [83] m. zu, j. jiang, h. zhao, s. zhang, y. yan, s. qiu, s. yuan, j. han, y. zhang, w. guo. rapid analysis of neomycin in cochlear perilymph of guinea pigs using disposable spe cartridges and high performance liquid chromatography-tandem mass spectrometry. journal of chromatography b 1093 (2018) 52-59. https://doi.org/10.1016/j.jchromb.2018.06.055. [84] a.c.c. da silva, l.d.l.f. lizot, m.f. bastiani, m.v. antunes, n. brucker, r. linden. ready for tdm: simultaneous quantification of amikacin, vancomycin and creatinine in human plasma employing ultra-performance liquid chromatography-tandem mass spectrometry. clinical biochemistry 70 (2019) 39-45. https://doi.org/10.1016/j.clinbiochem.2019.06.011. [85] a.c.c. da silva, l.d.l.f. lizot, m.f. bastiani, m.v. antunes, n. brucker, r. linden. dried plasma spots for therapeutic monitoring of amikacin: validation of an uhplc-ms/ms assay and pharmacokinetic application. journal of pharmaceutical biomedical analysis 184 (2020) 113201. https://doi.org/10.1016/j.jpba.2020.113201. [86] j.-p. antignac, k. de wasch, f. monteau, h. de brabander, f. andre, b. le bizec. the ion suppression phenomenon in liquid chromatography–mass spectrometry and its consequences in the field of residue analysis. analytica chimica acta 529 (2005) 129-136. https://doi.org/10.1016/j.aca.2004.08.055. [87] c. ghosh, c.p. shinde, b.s. chakraborty. influence of ionization source design on matrix effects during lc–esi-ms/ms analysis. journal of chromatography b 893 (2012) 193-200. https://doi.org/10.1016/j.jchromb.2012.03.012. [88] b. matuszewski, m. constanzer, c. chavez-eng. strategies for the assessment of matrix effect in quantitative bioanalytical methods based on hplc− ms/ms. analytical chemistry insights 75 (2003) 3019-3030. https://doi.org/10.1021/ac020361s. [89] p.j. taylor. matrix effects: the achilles heel of quantitative high-performance liquid chromatography– electrospray–tandem mass spectrometry. clinical biochemistry 38 (2005) 328-334. https://doi.org/10.1016/j.clinbiochem.2004.11.007. [90] o.a. ismaiel, t. zhang, r.g. jenkins, h.t. karnes. investigation of endogenous blood plasma phospholipids, cholesterol and glycerides that contribute to matrix effects in bioanalysis by liquid chromatography/mass spectrometry. journal of chromatography b 878 (2010) 3303-3316. https://doi.org/10.1016/j.jchromb.2010.10.012. [91] j.l. little, m.f. wempe, c.m. buchanan. liquid chromatography–mass spectrometry/mass spectrometry method development for drug metabolism studies: examining lipid matrix ionization effects in plasma. journal of chromatography b 833 (2006) 219-230. https://doi.org/10.1016/j.jchromb.2006.02.011. [92] m. garcia. the effect of the mobile phase additives on sensitivity in the analysis of peptides and proteins by high-performance liquid chromatography–electrospray mass spectrometry. journal of chromatography b 825 (2005) 111-123. https://doi.org/10.1016/j.jchromb.2005.03.041. [93] h. mei, y. hsieh, c. nardo, x. xu, s. wang, k. ng, w.a. korfmacher. investigation of matrix effects in bioanalytical high‐performance liquid chromatography/tandem mass spectrometric assays: application to drug discovery. rapid communications in mass spectrometry 17 (2003) 97-103. https://doi.org/10.1002/rcm.876. [94] w. li, w. jian, y. fu. basic sample preparation techniques in lc‐ms bioanalysis: protein precipitation, liquid–liquid extraction, and solid‐phase extraction. sample preparation in lc‐ms bioanalysis (2019) 1-30. https://doi.org/10.1002/9781119274315.ch1. https://doi.org/10.1016/j.jchromb.2014.01.035 https://doi.org/10.5530/phm.2016.7.15 https://doi.org/10.1016/j.jchromb.2018.06.055 https://doi.org/10.1016/j.clinbiochem.2019.06.011 https://doi.org/10.1016/j.jpba.2020.113201 https://doi.org/10.1016/j.jpba.2020.113201 https://doi.org/10.1016/j.aca.2004.08.055 https://doi.org/10.1016/j.aca.2004.08.055 https://doi.org/10.1016/j.jchromb.2012.03.012 https://doi.org/10.1016/j.jchromb.2012.03.012 https://doi.org/10.1021/ac020361s https://doi.org/10.1016/j.clinbiochem.2004.11.007 https://doi.org/10.1016/j.clinbiochem.2004.11.007 https://doi.org/10.1016/j.jchromb.2010.10.012 https://doi.org/10.1016/j.jchromb.2006.02.011 https://doi.org/10.1016/j.jchromb.2006.02.011 https://doi.org/10.1016/j.jchromb.2005.03.041 https://doi.org/10.1002/rcm.876 https://doi.org/10.1002/9781119274315.ch1 admet & dmpk 10(1) (2022) 27-62 bioanalysis of aminoglycozides by hplc doi: http://dx.doi.org/10.5599/admet.1183 61 [95] a. posyniak, j. zmudzki, j. niedzielska. sample preparation for residue determination of gentamicin and neomycin by liquid chromatography. journal of chromatography a 914 (2001) 59-66. https://doi.org/10.1016/s0021-9673(00)00980-8. [96] v. gupta, a.d.k.j. jain, n. gill, k. guptan. development and validation of hplc method-a review. international research journal of pharmaceutical applied sciences 2 (2012) 17-25. [97] k.f.m. opuni, j.a. boadu, s.k. amponsah, c.a. okai. high performance liquid chromatography: a versatile tool for assaying antiepileptic drugs in biological matrices. journal of chromatography b 1179 (2021) 122750. https://doi.org/10.1016/j.jchromb.2021.122750. [98] m. charde, a. welankiwar, j. kumar. method development by liquid chromatography with validation. international journal of pharmaceutical chemistry 4 (2014) 57-61. https://doi.org/10.7439/ijpc.v4i1.65. [99] y. vidushi, b. meenakshi. a review on hplc method development and validation. res j life sci, bioinform, pharm chem sci 2 (2017) 178. https://doi.org/10.26479/2017.0206.12. [100] d. kowalczuk, r. pietraś, b. paw, a. czerkies. applying liquid chromatography with fluorescence detection to determine gentamicin. pol. j. environ. stud 19 (2010) 587-591. [101] h. itagaki. fluorescence spectroscopy. experimental methods in polymer science, academic press, san diego, ca (2000) 155-260. [102] r. lobrutto, y. kazakevich, reversed-phase hplc, hplc for pharmaceutical scientists, wiley online library2007, pp. 139-195. [103] k.f.m. opuni, m. al-majdoub, y. yefremova, r.f. el-kased, c. koy, m.o. glocker. mass spectrometric epitope mapping. mass spectrom rev 37 (2018) 229-241. https://doi.org/10.1002/mas.21516. [104] m. chen, h.h. cai, f. yang, d. lin, p.h. yang, j. cai. highly sensitive detection of chromium (iii) ions by resonance rayleigh scattering enhanced by gold nanoparticles. spectrochim acta a mol biomol spectrosc 118 (2014) 776-781. https://doi.org/10.1016/j.saa.2013.09.058. [105] w.r. lacourse, s.j. modi. microelectrode applications of pulsed electrochemical detection. electroanalysis: an international journal devoted to fundamental practical aspects of electroanalysis 17 (2005) 1141-1152. https://doi.org/10.1002/elan.200403233. [106] p.j. vandeberg, d.c. johnson. pulsed electrochemical detection of cysteine, cystine, methionine, and glutathione at gold electrodes following their separation by liquid chromatography. analytical chemistry 65 (1993) 2713-2718. https://doi.org/10.1021/ac00068a002. [107] s.m. moosavi, s. ghassabian. linearity of calibration curves for analytical methods: a review of criteria for assessment of method reliability. calibration validation of analytical methods—a sampling of current approaches, intechopen ltd.: london, uk (2018) 109-127. [108] a.m.d. almeida, m.m. castel-branco, a. falcao. linear regression for calibration lines revisited: weighting schemes for bioanalytical methods. journal of chromatography b 774 (2002) 215-222. https://doi.org/10.1016/s1570-0232(02)00244-1. [109] l.c. rodríguez, a.m.g. campa ta, c.j. linares, m.r. ceba. estimation of performance characteristics of an analytical method using the data set of the calibration experiment. analytical letters 26 (1993) 1243-1258. https://doi.org/10.1080/00032719308019900. [110] g. smith. european medicines agency guideline on bioanalytical method validation: what more is there to say? bioanalysis 4 (2012) 865-868. https://doi.org/10.4155/bio.12.44. [111] l.v. sonawane, b.n. poul, s.v. usnale, p.v. waghmare, l.h. surwase. bioanalytical method validation and its pharmaceutical application-a review. pharm anal acta 5 (2014) 2. https://doi.org/10.4172/2153-2435.1000288. [112] a. dasgupta. advances in antibiotic measurement. adv clin chem 56 (2012) 75-104. https://doi.org/10.1016/b978-0-12-394317-0.00013-3. [113] p. noone. sisomicin, netilmicin and dibekacin. a review of their antibacterial activity and therapeutic use. drugs 27 (1984) 548-578. https://doi.org/10.2165/00003495-198427060-00003. http://dx.doi.org/10.5599/admet.1183 https://doi.org/10.1016/s0021-9673(00)00980-8 https://doi.org/10.1016/j.jchromb.2021.122750 https://doi.org/10.7439/ijpc.v4i1.65 https://doi.org/10.7439/ijpc.v4i1.65 https://doi.org/10.26479/2017.0206.12 https://doi.org/10.1002/mas.21516 https://doi.org/10.1016/j.saa.2013.09.058 https://doi.org/10.1002/elan.200403233 https://doi.org/10.1021/ac00068a002 https://doi.org/10.1016/s1570-0232(02)00244-1 https://doi.org/10.1080/00032719308019900 https://doi.org/10.4155/bio.12.44 https://doi.org/10.4172/2153-2435.1000288 https://doi.org/10.1016/b978-0-12-394317-0.00013-3 https://doi.org/10.2165/00003495-198427060-00003 seth k. amponsah et al. admet & dmpk 10(1) (2022) 27-62 62 [114] j.p. hieber, j.d. nelson. reevaluation of kanamycin dosage in infants and children. antimicrob agents chemother 9 (1976) 899-902. https://doi.org/10.1128/aac.9.6.899. [115] drugs.com. streptomycin dosage., 2021. [116] g.g. duncan, c.f. clancy, j.r. wolgamot, b. beidleman. neomycin: results of clinical use in ten cases. journal of the american medical association 145 (1951) 75-80. https://doi.org/10.1001/jama.1951.02920200015005. [117] m. urata, y. narita, d. kadowaki, k. tanoue, i. tashiro, m. fukunaga, a. shibata, y. yoshida, k. soejima, s. hirata. interaction of arbekacin with dialysis membrane. renal replacement therapy 2 (2016) 35. https://doi.org/10.1186/s41100-016-0045-z. [118] e. novak, c.a. schlagel, l.a. lezotte, r.t. pfeifer. the tolerance of high dose intravenous spectinomycin therapy in man. j clin pharmacol 14 (1974) 442-447. https://doi.org/10.1002/j.15524604.1974.tb02326.x. [119] k.m. krause, a.w. serio, t.r. kane, l.e. connolly. aminoglycosides: an overview. cold spring harbor perspectives in medicine 6 (2016) a027029. https://doi.org/10.1101/cshperspect.a027029. [120] v. alampanos, v. samanidou. current trends in green sample preparation before liquid chromatographic bioanalysis. current opinion in green and sustainable chemistry 31 (2021) 100499. https://doi.org/10.1016/j.cogsc.2021.100499. [121] v. alampanos, v. samanidou, i. papadoyannis. trends in sample preparation for the hplc determination of penicillins in biofluids. journal of applied bioanalysis 5 (2019) 9-17. https://doi.org/10.17145/jab.19.003. [122] k.d. clark, c. zhang, j.l. anderson. sample preparation for bioanalytical and pharmaceutical analysis. anal chem 88 (2016) 11262-11270. https://doi.org/10.1021/acs.analchem.6b02935. ©2022 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) https://doi.org/10.1128/aac.9.6.899 https://doi.org/10.1001/jama.1951.02920200015005 https://doi.org/10.1186/s41100-016-0045-z https://doi.org/10.1002/j.1552-4604.1974.tb02326.x https://doi.org/10.1002/j.1552-4604.1974.tb02326.x https://doi.org/10.1101/cshperspect.a027029 https://doi.org/10.1016/j.cogsc.2021.100499 https://doi.org/10.17145/jab.19.003 https://doi.org/10.1021/acs.analchem.6b02935 http://creativecommons.org/licenses/by/3.0/ preclinical assessment of ulixertinib, a novel erk1/2 inhibitor doi: 10.5599/admet.5.4.437 212 admet & dmpk 5(4) (2017) 212-223; doi: http://dx.doi.org/10.5599/admet.5.4.437 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper preclinical assessment of ulixertinib, a novel erk1/2 inhibitor vinod a balakrishna 1 , anitha police 1 , rakesh a hiremath 2 , anusha raj 2 , suresh p sulochana 1 , devaraj v chandrasekhar 1 , mohd zainuddin 1 , ravi kanth bhamidipati 1 and ramesh mullangi 1 * 1 drug metabolism and pharmacokinetics, jubilant biosys, industrial suburb, yeshwanthpur, bangalore-560 022, india 2 j.s.s college of pharmacy, mysore, india *corresponding author: e-mail: mullangi.ramesh@jubilantinnovation.com; tel.: 9611333488 received: september 08, 2017; revised: october 03, 2017; published: december 24, 2017 abstract ulixertinib (bvd-523) is a novel and selective reversible inhibitor of erk1/erk2. in xenograft studies it inhibited tumor growth in braf-mutant melanoma and colorectal xenografts as well as kras-mutant colorectal and pancreatic models. ulixertinib is currently in phase i clinical development for the treatment of advance solid tumors. the objective of the study is to assess the metabolic stability (in various preclinical and human liver microsomes/hepatocytes), permeability, protein binding, cyp inhibition, cyp induction and cyp phenotyping of ulixertinib. we have also studied the oral and intravenous pharmacokinetics of ulixertinib in mice, rats and dogs. ulixertinib was found to be moderately to highly stable in various liver microsomes/hepatocytes tested. it is a medium permeable (2.67 x 10 -6 cm /sec) drug and a substrate for efflux (efflux ratio: 3.02) in caco-2 model. ulixertinib was highly bound to plasma proteins. cyps involved in its limited metabolism and it is cyp inhibition ic50 ranged between 10-20 µm. post oral administration ulixertinib exhibited early tmax (0.50-0.75 h) in mice and rats indicating absorption was rapid, however in dogs tmax attained at 2 h. the half-life (t½) of ulixertinib by intravenous and oral routes ranged between 1.0-2.5 h across the species. clearance and volume of distribution by intravenous route for ulixertinib were found to be 6.24 ml/min/kg and 0.56 l/kg; 1.67 ml/min/kg and 0.36 l/kg and 15.5 ml/min/kg and 1.61 l/kg in mice, rats and dogs, respectively. absolute oral bioavailability in mice and rats was > 92 %, however in dogs it was 34 %. keywords ulixertinib; erk inhibitor; in vitro studies; pharmacokinetics; mice; rats; dogs introduction the extracellular-signal-regulated kinases, erk1 and erk2 (erk1/2), play a critical role in the ras/raf/mek signaling pathway, that controls several fundamental cellular processes, driving proliferation, differentiation and cell survival [1]. ras/raf/mek signaling pathway is frequently activated by mutations in upstream targets such as braf, ras and receptor tyrosine kinases [2]. most of the resistance mechanisms to braf and mek inhibitors ultimately lead to increase in phosphorylation of erk1/2 suggesting the importance of this node in the ras/raf/mek pathway even in the resistance setting [3]. ras activating mutations have been reported in about 90 % of pancreatic carcinomas, followed by colon carcinomas (50 %), lung cancers and myeloid leukemia cases (30 % each) [4]. therefore, inhibition of http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:mullangi.ramesh@jubilantinnovation.com admet & dmpk 5(4) (2017) 212-223 preclinical assessment of ulixertinib doi: 10.5599/admet.5.4.437 213 erk1/2 offers a promising strategy to address both innate and acquired resistance to braf and mek inhibitors in various solid tumors. ulixertinib (bvd-523, vrt752271; figure 1) is a first-in-class novel small molecule, which potently and selectively inhibits erk1/2 kinases in a reversible, atp-competitive fashion. inhibition of erk1/2 leads to proliferation inhibition (by decreasing the expression of vegfa and vegfr2 at mrna and protein level), cell cycle arrest and apoptosis in lymphoma cell lines [5]. ulixertinib inhibits tumor growth in vivo in braf-mutant melanoma and colorectal xenografts as well as in kras-mutant colorectal and pancreatic models [6]. in clinical studies, ulixertinib was well tolerated by patients with advanced solid tumors. in an oral phase-i dose escalation study (having 9 doses) with an end point to determine the dose limited toxicities (dlt), maximum tolerated dose (mtd) along with pharmacokinetic profile and preliminary efficacy assessment it was administered in a dose range of 10-900 mg in a b.i.d regimen. ulixertinib showed linear pharmacokinetics up to 600 mg (b.i.d), this was found to be mtd [7]. figure 1. structural representation of ulixertinib to date there is no publication on preclinical pharmacokinetics of erk1/2 inhibitors, hence we felt that preliminary data on various in vitro adme studies and in vivo pharmacokinetics in mice, rats and dogs of ulixertinib will help researchers in this field to provide the differentiating features in this class. this paper describes various in vitro experiments done to (a) assess the metabolic stability of ulixertinib in various preclinical species and human liver microsomes and hepatocytes (b) determine the permeability in caco-2 cells (c) determine the unbound fraction in preclinical species and human plasma (d) cyp inhibition (e) cyp induction and (f) cyp phenotyping of ulixertinib. moreover, a series of in vivo experiments in balb/c mice, sprague dawley rats and beagle dogs were performed to characterize and understand pharmacokinetic parameters and absolute oral bioavailability of ulixertinib across the species. materials and methods materials ulixertinib (purity >99 %) was purchased from medkoo biosciences, inc., nc, usa. hplc grade acetonitrile, formic acid and methanol were purchased from rankem, ranbaxy fine chemicals limited, new delhi, india. preclinical species and human liver microsomes and hepatocytes were purchased from life technologies (gibco®), inchinnan, uk. substrates and inhibitors of cyp isoforms like midazolam, quinidine, diclofenac, bufuralol, omeprazole, verapamil, rifampicin, warfarin, tolbutamide, phenacetin, ketoconazole, sulfaphenazole, nootkatone, furafylline and digoxin were purchased from sigma-aldrich chemie gmbh, steinheim, germany. all other chemicals/reagents were of research grade and used without further purification. r. mullangi et al. admet & dmpk 5(4) (2017) 212-223 214 in vitro studies metabolic stability in liver microsomes metabolic stability is defined as the percentage of parent compound lost over time in the presence of a metabolically active test system. the incubation mixtures consisted of liver microsomes (0.5 mg microsomal protein/ml), ulixertinib (1 μm) or positive control (verapamil, 1 μm). the reactions were initiated by adding 20 μl of 10 mm nadph [8]. reactions without nadph (0 and 30 min) were also incubated to rule out non-nadph metabolism or chemical instability in the incubation buffer [8]. all reactions were terminated using 200 μl of ice-cold acetonitrile containing internal standard (200 ng/ml of phenacetin) at 0, 5, 15 and 30 min. the vials were centrifuged at 3000 rpm (eppendorf 5424r, germany) for 15 min. the supernatants thus obtained were analyzed on lc-ms/ms to monitor the disappearance of ulixertinib. metabolic stability in hepatocytes after checking the viability of cryopreserved hepatocytes, they were distributed 1 million viable cells per ml. an aliquot of 100 µl of thawed hepatocytes cell suspension was aliquoted into different vials labeled as 0, 30, 60, 120 and 180 min (mouse, rat, dog or human hepatocytes) and ulixertinib (3 μm). imipramine or naloxone (3 μm) was used as a positive control [9]. the vials were pre-incubated in a 37 °c water bath for 5 min. all reactions were terminated using 200 μl of ice-cold acetonitrile containing internal standard (200 ng/ml of phenacetin) at specified time points. the vials (having reaction mixture terminated contents) were centrifuged at 3000 rpm (eppendorf 5424r, germany) for 15 min. the supernatants thus obtained were analyzed on lc-ms/ms to monitor the disappearance of ulixertinib. clearance rate and halflife were calculated using the obtained data. caco-2 permeability caco-2 human intestinal epithelial cells were plated in 24-transwell ® dual chamber plates (millipore, billerica, ma, usa) (cell density of 80,000 cells/cm 2 on day-1). the permeability studies were conducted with the monolayers cultured for 21 to 22 days. the integrity of each caco-2 cell monolayer was certified by trans epithelial electrical resistance (teer) test (pre-experiment) and by determining the permeability of reference compound i.e., lucifer yellow. caco-2 cell monolayers with teer values greater than 500 ω cm 2 were considered for experimentation. digoxin (5 µm) was used as a positive control for p-gp substrate [10]. the concentration of ulixertinib used in the assay was 10 µm. hbss buffer was used as the medium for the transport assay and the final concentration of dmso in spiking solution was 0.05 %. the bidirectional permeability study was initiated by adding an appropriate volume of hbss buffer containing ulixertinib to respective apical and basolateral chambers (n=2) [11]. an aliquot of sample (100 µl) was taken from both chambers at 0 and 60 min of the incubation period and to this equal volume of acetonitrile with internal standard (200 ng/ml of phenacetin) was added, mixed gently and centrifuged at 4000 rpm (eppendorf 5424r, germany) for 10 min. an aliquot of 100 µl was subsequently transferred to the auto-sampler and injected for analysis on lc-ms/ms. plasma protein binding to evaluate the ability of ulixertinib to bind the plasma proteins, the most common approach of plasma protein binding using equilibrium dialysis was used [12]. ulixertinib was tested at a final concentration of 3 µm in mouse, rat and dog plasma. an aliquot of 150 μl plasma containing ulixertinib was added in first half (plasma side) of the well of 96-well micro-equilibrium dialysis device. an aliquot of 150 μl of 100 mm admet & dmpk 5(4) (2017) 212-223 preclinical assessment of ulixertinib doi: 10.5599/admet.5.4.437 215 sodium phosphate buffer ph 7.4 was added in the second half (buffer side) of the well of 96-well ht equilibrium dialysis device. the plate containing plasma and buffer was equilibrated at 37 ± 1 °c for 4.5 h, with constant shaking at 120 rpm on an orbital shaker. samples were collected from respective halves after the completion of incubation time. the proteins were precipitated using organic solvents. the samples were subjected to centrifugation and the supernatants were analyzed analysis on lc-ms/ms. cyp inhibition cyp inhibition potential of ulixertinib was assessed in human liver microsomes [life technologies (gibco®), inchinnan, uk] against cyp3a4, 2d6, 1a2, 2c9 and 2c19 in the following sequential steps. standard reaction mixture (final volume 300 μl) contained 66.7 m potassium phosphate buffer (ph 7.4), protein [0.1 (for cyp3a4), 0.25 (for cyp2d6), 0.5 mg/ml (for cyp1a2, 2c9 and 2c19)] and ulixertinib (at 2.0, 10 and 20 µm, added as 0.9 µl dmso solution with a final dmso concentration of 0.1 %). the mixtures were pre-incubated at 37 ± 1 °c for 5 min. the reaction (in duplicate) was initiated by addition of 30 µl of nadph (10 mm) [13]. reaction was terminated at 10 min by adding 300 µl of ice cold acetonitrile with internal standard. the activity of liver microsomes was confirmed with positive controls i.e., monitoring the hydroxylation of midazolam, bufuralol, diclofenac and omeprazole for cyp3a4, 2d6, 2c9 and 2c19, respectively and deethylation of phenacetin for cyp1a2 (data not shown). the reaction mixtures (obtained from the above studies) were extracted, processed, analyzed on lc-ms/ms. rhcyp metabolism to evaluate the metabolism of ulixertinib, the most common approach of reaction phenotyping by using cdna expressed enzyme system was used. ulixertinib was incubated with a panel of individually-expressed recombinant human cyp enzymes (cyp1a2, 2b6, 2c8, 2c9, 2c19, 2d6 and 3a4, purchased from life technologies (gibco®), inchinnan, uk expressed in baculovirus-infected insect cell membranes. the incubation mixture contained ulixertinib at a final concentration of 0.1 μm, expressed cyp enzyme (50 pmol/ml), phosphate buffer (66.7 mm, ph 7.4) in a total volume of 1 ml. the reaction was initiated by the addition of 20 µl of nadph after pre-incubation of the aliquots (180 µl) marked 0, 5, 15, 30 and 60 min followed by incubation at 37 ± 1 °c. the reaction was terminated by the addition of equal volumes of ice cold acetonitrile with internal standard (200 ng/ml of phenacetin). the samples were centrifuged at 3000 rpm (eppendorf 5424r, germany) at 4 °c for 20 min to precipitate proteins. supernatants were then transferred to clean vials and stored at -20 °c until analysis. incubations without nadph were used as negative controls. control incubations using probe substrates for individual recombinant humans cyps were included to check for appropriate incubation conditions and as positive control for activity [15]. the samples were analyzed on lc-ms/ms. disappearance of ulixertinib was compared to that of control treatment (no cyp enzymes present) in order to assess the contribution of a given cyp in ulixertinib metabolism. in vivo studies formulations for oral administration a suspension formulation was prepared using tween-80 and 0.5 % methyl cellulose. weighed amount of ulixertinib was taken into a mortar and grinded into a fine powder with the help of a pestle. to this tween-80 was added drop wise to wet the entire powder. then slowly 0.5 % methyl cellulose was added with constant stirring to get a uniform suspension. the final strength of the suspension formulation was 10 mg/ml and it was administered to animal species at 10 ml/kg body weight. for intravenous (i.v.) administration a solution formulation was used (prepared using 10 % dmso, 10 % r. mullangi et al. admet & dmpk 5(4) (2017) 212-223 216 solutol:absolute ethyl alcohol (1:1, v/v) and 80 % normal saline). the volume of administration for intravenous was 10 ml/kg for mice and 2 ml/kg for rats and dogs. animal experiments details institutional animal ethical committee (iaec) of jubilant biosys (iaec/jdc/2017/121) nominated by cpcsea (committee for the purpose of control and supervision of experiments on animals) approved the rodents pharmacokinetic experiments. male sprague dawley rats (~7-8 weeks old) and balbc mice (~6-10 weeks old) were purchased from vivo biotech, hyderabad, india. animals were quarantined in jubilant biosys animal house for a period of 7 days with a 12:12 h light:dark cycles, had free access to rodent food (altromin spezialfutter gmbh & co. kg., im seelenkamp 20, d-32791, lage, germany) and water ad libitum. for all the experimental work animals were kept for fasting (4 h for balbc mice and 12 h for sprague dawley rats) and during this time they were allowed to take water ad libitum. food was provided 2 h post-dose and water was allowed ad libitum. iaec of palamur biosciences, telangana, india (1312/po/rcbibt/s/l/09/cpcsea) approved studies conducted in dogs. male beagle dogs (~1.0-1.5 year old) were housed in palamur biosciences private limited animal house facility in a temperature (18-28 °c) and humidity (30-70 %) controlled room and fed with pedigree standard pellet food and water ad libitum for one week before using for experimental purpose. for all the experimental work animals were kept for 12 h overnight fasting and during this time they were allowed to take water ad libitum. food was provided 4 h post-dose and water was allowed ad libitum. pharmacokinetic studies oral bioavailability of ulixertinib was evaluated in male balb/c mice, sprague dawley rats and beagle dogs. fasted mice (~4 h) and rats (overnight ~12 h) were divided into two groups (mice: 12/group; rats: 4/group), however in cases of dogs (n=2; fasted for ~12 h overnight) same dogs were dose after one week wash-out period (cross over design). in each species group-1 animals received ulixertinib through intravenous route at a dose of 1 mg/kg; whereas group 2 received ulixertinib through oral gavage at 10 mg/kg dose. serial blood samples [100 µl in case of mice (sparse sampling; n=3 at each time point) and rats; 500 µl in dogs)] were collected from retro-orbital plexus at 0.12, 0.25, 0.5, 1, 2, 4, 8, and 24 h (intravenous administration) or at 0.25, 0.5, 1, 2, 4, 8, 10 and 24 h (oral administration) after drug administration. blood samples were collected in tubes containing k2.edta as the anticoagulant and centrifuged for 5 min at 14000 rpm in a refrigerated centrifuge (biofuge, heraeus, germany) maintained at 4 °c for plasma separation and stored frozen at -80 ± 10 °c until analysis. sample processing in vitro studies samples analysis a shimadzu (shimadzu, kyoto, japan) sil series lc system equipped with a degasser (dgu-20a3), isopump (lc-20 ad) and column oven (cto-10as) along with an auto-sampler (sil-htc) was used to inject 10 µl aliquots of the processed samples on a atlantis c18 column (50 x 4.6 mm, 3 µm; waters, ireland, uk), which was maintained at 40 °c. the mobile phase system consisted of reservoir a (acetonitrile) and reservoir b (0.2 % formic acid in water) were run as per gradient program (0-0.1 min: 10 % a and 90 % b; 0.1-2.8 min: 100 % a; 2.9-4.0 min 10 % a and 90 % b). a flow rate of 0.8 ml/min with a 50 % splitter was used throughout the analytical run. quantitation was achieved by ms/ms detection in positive ion mode for ulixertinib using an api-4000 q trap mass spectrometer (mds sciex, toronto, canada) equipped with a admet & dmpk 5(4) (2017) 212-223 preclinical assessment of ulixertinib doi: 10.5599/admet.5.4.437 217 turboionspray interface at 450 °c temperature and 5500 v ion spray voltage. the source parameters viz., curtain gas, gs1, gs2 and cad were set at 30, 35, 40 and 6 psi. the compound parameters viz., declustering potential (dp), entrance potential (ep), collision energy (ce) and collision cell exit potential (cxp) were 81, 10, 49 and 15 v for ulixertinib and 80, 10, 29 and 14 v for the is. detection of the ions was performed in the multiple reaction monitoring (mrm) mode, monitoring the transition of the m/z 433 precursor ion to the m/z 262 product ion for ulixertinib and m/z 180 precursor ion to the m/z 110 product ion for the is. quadrupole q1 and q3 were set on unit resolution. the dwell time was 150 msec. plasma samples processing and analysis an aliquot of 50 µl plasma sample was precipitated with 200 µl of acetonitrile:methanol (1:1, v/v) enriched with internal standard (200 ng/ml of phenacetin) and centrifuged at 14,000 rpm for 5 min (eppendorf 5424r, germany) at 5 °c. clear supernatant (125 µl) was transferred into vials and 5 µl of supernatant was injected onto lc-ms/ms system for analysis using a validated method [16]. the linearity range was 1.05-2096 ng/ml. in-study quality control (qc) samples, supplemented with concentrations of 3.14, 1048 and 1747 ng/ml of ulixertinib, were analysed with the unknowns. plasma samples showing high concentration above the high calibration standard were diluted with appropriate animal species blank plasma to bring the concentration within linearity range. for plasma samples analysis the criteria for acceptance of the analytical runs encompassed the following: (i) 67 % of the qc samples accuracy must be within 85-115 % of the nominal concentration (ii) not less than 50 % at each qc concentration level must meet the acceptance criteria. following completion of the analysis both the linearity and quality control samples values were found to be within the accepted variable limits. pharmacokinetic analysis pharmacokinetic parameters were calculated by a non-compartmental method using phoenix winnonlin 7.0 software (pharsight, mountain view, ca, usa). absolute oral bioavailability (f) was calculated using the relationship, f = [dose (i.v.) x auc(0-)oral / dose (oral) x auc(0-)i.v.] x 100. results in vitro studies metabolic stability in liver microsomes metabolic stability of ulixertinib in different species of liver microsomes is presented in table 1. ulixertinib was found to be moderately stable in mice and dog liver microsomes and highly stable rat and human liver microsomes. metabolic stability in hepatocytes hepatocyte stability of ulixertinib in different species is presented in table 2. ulixertinib was found to be highly stable in rat and human hepatocytes but moderately stable in mice and dog hepatocytes. the in vivo intrinsic clearance (clint) was relatively low to moderate in rat and human hepatocytes but high in mice and dog hepatocytes. caco-2 permeability table 3 shows the rate of transport (papp) of ulixertinib and digoxin from a → b and b → a along with efflux ratio values. the findings suggested that ulixertinib is a medium permeable compound and may be a r. mullangi et al. admet & dmpk 5(4) (2017) 212-223 218 substrate of an active efflux transporter. table 1. metabolic stability data of ulixertinib at 1 µm in liver microsomes of various preclinical species species percent metabolized t½ (min) mclint (µl/min/mg) clint (ml/min/kg) mice 44.5 36 19.3 27.7 rat 16.4 122 5.7 7.6 dog 35.5 52 13.3 9.1 human 12.8 142 9.8 5.4 table 2. hepatocyte stability data of ulixertinib at 3 µm in liver hepatocytes of various preclinical species species percent metabolized t½ (min) hclint (ml/min/106 cells) clint scaled (ml/min/kg) mice 79 78 8.9 41.3 rat 27 410 1.7 7.1 dog 76 73 9.1 16.4 human 37 306 2.3 5.0 table 3. permeability data of ulixertinib in caco-2 cell monolayers % recovery permeability (x 10 -6 cm/sec) efflux ratio compound concentration (µm) b  a a  b b  a a  b b  a/ a  b ulixertinib 10 79 65 8.09 2.67 3.02 digoxin 5.0 98 75 10.5 1.00 10.5 plasma protein binding ulixertinib had a very high binding in rat plasma followed by mouse and dog plasma with fraction unbound of 0.001, 0.003 and 0.050, respectively. the stability and recovery of ulixertinib in plasma was good across the species tested. cyp inhibition the predicted ic50 values of ulixertinib were determined for cyp-specific hydroxylation of midazolam, bufuralol, diclofenac and omeprazole for cyp 3a4, 2d6, 2c9 and 2c19, respectively and deethylation of cyp1a2. ulixertinib did not show notable inhibition against these enzymes and the predicted ic50 values of ulixertinib were found to be ~10 and 20 µm for cyp 3a4 and 2d6, respectively indicating it is a weak inhibitor against these two cyps. furthermore, its inhibitory effect was much weaker against cyp1a2, 2c9 and 2c19 as its ic50 was > 20 µm. thus, when compared with known inhibitors, ulixertinib shows less significant inhibition of cyp enzyme activity in vitro (table 4). cyp induction admet & dmpk 5(4) (2017) 212-223 preclinical assessment of ulixertinib doi: 10.5599/admet.5.4.437 219 ulixertinib was tested at two different concentrations (1 and 10 µm). the readings suggested no or very less percent activation compared against rifampicin, a known cyp3a4 inducer. table 4. ic50 values for cyp450 inhibition by ulixertinib and other known cyp inhibitors ic50 (µm) inhibitor cyp3a4 cyp2d6 cyp2c9 cyp2c19 cyp1a2 ulixertinib ~10 ~20 >20 >20 >20 ketoconazole 0.08 nt nt nt nt quinidine nt 0.12 nt nt nt sulphaphenazole nt nt 0.72 nt nt nootkatone nt nt nt 1.02 nt furafylline nt nt nt nt 0.58 rhcyp metabolism data following incubations with different cyp isozymes (cyp1a2, 2b6, 2c8, 2c9, 2c19, 2d6 and 3a4) considered in the panel suggested very minimal contribution of cyp1a2, 2c9, 2c19, 2d6 and 3a4 towards the metabolism of ulixertinib involving oxidation and dealkylation. this data is corroborating with the work reported by germann et al [6]. pharmacokinetic studies the profiles of plasma concentration for ulixertinib following single oral and intravenous administration to mice, rats and dogs are shown in figure 2a and 2b, respectively. the estimates of pharmacokinetic parameters in these species are summarized in table 5. 0 4 8 12 16 20 24 1 10 100 1000 10000 100000 balb/c mice plasma conce ntration ve rsus time profile of ulixe rtinib by oral route at 10 mg/kg time (h) c o n c e n tr a ti o n i n p la s m a (n g /m l ) m e a n ± s .d . sd rats dogs 0 4 8 12 16 20 24 1 10 100 1000 10000 balb/c mice plasma conce ntration ve rsus time profile of ulixe rtinib by intravenous route at 1 mg/kg time (h) c o n c e n tr a ti o n i n p la s m a (n g /m l ) m e a n ± s .d . sd rats dogs figure 2. plasma concentration-time profiles of ulixertinib after (a) oral (10 mg/kg) and (b) intravenous (1 mg/kg) administration to male balb/c mice (mean  s.d, n = 12), sprague dawley rats (mean  s.d, n = 4) and beagle dogs (mean, n = 2). a b r. mullangi et al. admet & dmpk 5(4) (2017) 212-223 220 in the mice, plasma concentrations of ulixertinib decreased mono-exponentially manner after 1 mg/kg intravenous administration. the clearance was 6.24 ml/min/kg [7 % hepatic blood flow (hbf)]. the in vivo clearance was bit over predicted by the in vitro microsomes (~4-fold) and hepatocytes (~6-fold) scaled clearance. ulixertinib had a high volume of distribution of 0.56 l/kg in mice, which is 38 times higher than that of total body water (tbw) of 0.015 l/kg [17]. the terminal t1/2 was found to be 1.04 h. post oral administration maximum plasma concentrations (cmax: 7768 ng/ml) attained at 0.5 h indicating rapid absorption from gastrointestinal tract. the terminal t1/2 (2.06 h) determined after oral administration was longer than that after intravenous administration (1.04 h), which may indicate multiple sites absorption. the auc0- attained post oral dose was 24460 ng*h/ml. the oral bioavailability in mice at 10 mg/kg was 92 % (table 5). table 5. pharmacokinetic parameters of ulixertinib after intravenous and oral administration to mice, rats and dogs at 1 and 10 mg/kg, respectively strain route of adm. dose (mg/kg) auc0-∞ (ng h/ml) cmax/co (ng/ml) tmax (h) t½ (h) cl (ml/min/kg) vd (l/kg) f (%) balbc mice intravenous 1.0 2672 1918 --1.04 6.24 0.56 -- oral 10 24460 7768 0.50 2.06 ----91.6 sprague dawley rats intravenous 1.0 10179 ± 1528 6644 ± 1812 -- 2.52 ± 0.07 1.67 ± 0.25 0.36 ± 0.06 -- oral 10 98421 ± 14005 15026 ± 2098 0.75 ± 0.29 2.02 ± 0.17 ----97.7 beagle dogs intravenous 1.0 1091 1033 --1.21 15.5 1.61 -- oral 10 3687 1442 2.00 1.29 ----33.8 in the rats, plasma concentrations of ulixertinib decreased mono-exponentially manner after 1 mg/kg intravenous administration. the clearance was found to be 1.67 ml/min/kg (3 % of hbf), which is in good agreement with in vitro (microsomes and hepatocytes) scaled clearance. the volume of distribution (0.36 l/kg) of ulixertinib in rat is approximated to tbw. the terminal t1/2 was found to be 2.52 h. in rats also post oral administration early tmax of 0.75 h suggesting that ulixertinib has a rapid absorption from gastrointestinal tract. the auc0- and cmax attained post oral dose was 98421 ng*h/ml and 15026 ng/ml, respectively. the terminal t1/2 determined after oral administration is 2.02 h, which is comparable to intravenous route half-life. the oral bioavailability in rat at 10 mg/kg was 98 % (table 5). in the dogs also plasma concentrations of ulixertinib decreased mono-exponentially manner after 1 mg/kg intravenous administration. the clearance was 15.5 ml/min/kg (~50 % hbf). the in vivo clearance was in good agreement with in vitro (microsomes and hepatocytes) predicted scaled clearance. the apparent volume of distribution 1.61 l/kg of ulixertinib in dogs is approximated to tbw. the terminal t1/2 was found to be 1.21 h. in dogs post oral administration, maximum plasma concentrations (1442 ng/ml) of ulixertinib observed at 2.00 h, indicating it is a delayed tmax (when compared with tmax value of rodents). the auc0- was 3687 ng h/ml. the terminal t1/2 determined after oral administration was 1.29 h. the oral bioavailability in dogs at 10 mg/kg was 34 % (table 5). admet & dmpk 5(4) (2017) 212-223 preclinical assessment of ulixertinib doi: 10.5599/admet.5.4.437 221 discussion ras/raf/mek signaling pathway is frequently activated in many cancers [2] and clinical efficacy of braf and mek inhibitors confirms targeting ras/raf/mek pathway has therapeutic potential and great promise [18]. most of the resistance mechanisms to braf and mek inhibitors ultimately lead to increase in phosphorylation of erk1/2 [3]. therefore, inhibition of erk1/2 offers a promising strategy to address both innate and acquired resistance to braf and mek inhibitors in various solid tumors. ulixertinib is a novel compound, which selectively and potently inhibits erk1/2 kinases in a reversible, atp-competitive fashion. as a single-agent it inhibits tumor growth in vivo in braf-mutant melanoma and colorectal xenografts as well as in kras-mutant colorectal and pancreatic models [6]. in phase-1 clinical studies, ulixertinib was well tolerated by patients with advanced solid tumors. ulixertinib showed linear pharmacokinetics up to 600 mg (b.i.d), this was found to be mtd [7]. to the best of our knowledge there is no preclinical pharmacokinetic data reported for erk1/2 inhibitors in the literature. preclinical pharmacokinetics of a novel target class drug has great influence on the development and investigation of potential candidates for the further drug design, which necessitates a through exploration and understanding of pharmacokinetic disposition. this is important since in vivo pharmacokinetic behavior and pharmacokinetic-pharmacodynamic correlation act as surrogates for clinical effectiveness of potential drug candidates. in this paper we report the metabolic stability, permeability, protein binding, cyp inhibition, cyp induction, cyp phenotyping and pharmacokinetics in mice, rats and dogs for ulixertinib. the metabolic stability data suggested that ulixertinib is moderately to highly stable under in vitro conditions using liver microsomal and hepatocytes preparations of various pre-clinical species and humans. the caco-2 cell monolayer transport assay is well established test system for estimating drug absorption [19]. this caco-2 cell monolayer system generated papp value in absorptive direction for ulixertinib was 2.67 x 10 -6 cm/sec, indicating it is a moderately permeable compound, however the papp value in secretive direction appeared to much higher (8.09 x 10 -6 cm/sec), suggesting involvement of efflux transporters in the caco-2 monolayer. our results indicate that ulixertinib weakly inhibits the major cyps tested and ic50 values are greater than 20 µm (for most of cyps tested). ulixertinib did not show propensity to induce cyp34. so, cyp inhibition or induction is likely to represent a clinically significant risk of drug-drug interactions. ulixertinib oral absorption in mice and rats was not limited by efflux as bioavailability in these two species was > 92 %. however in dogs the oral bioavailability was 34 % and this may be due to high unbound concentrations (fu 0.05 in dog plasma vs 0.001 and 0.003 in mice and rat plasma, respectively) in plasma and extensive metabolism in dogs thus ulixertinib had relatively high plasma clearance. in addition, higher volume of distribution of mice indicates that ulixertinib is highly distributed in mice compared to rats and dogs. ulixertinib demonstrated high plasma clearance in dogs compared to rats and mice, which is 50, 3.0 and 7.0 % of hbf, respectively and was very much predicted by the in vitro experiment, indicating the difference in hepatic metabolism. conclusions in conclusion, the preclinical data gathered in this work provided evidence that ulixertinib has excellent oral bioavailability in rodents and acceptable bioavailability in dogs, adequate metabolic stability and devoid of drug-drug interactions liability and different mechanism of action may provide differentiating features from other oncology compounds in this class. overall, looking at the ulixertinib preclinical data presented by us and its current clinical data provide a good platform for researches to develop novel erk1/2 inhibitors with potential benefit as a single or concomitant solid tumor therapy. r. mullangi et al. admet & dmpk 5(4) (2017) 212-223 222 conflict of interest: the authors have no conflict of interest references [1] a. akinleye, m. furqan, n. mukhi, p. ravella, d. liu. mek and the inhibitors: from bench to bedside. journal of hematology & oncology 6 (2013) 27. [2] d. mattallanas, p. crespo. new druggable targets in the ras pathway? current opinion in molecular therapeutics 12 (2010) 674-683. [3] m. warthaka, c.h. adelmann, t.s. kaoud, r. edupuganti, c, yan, w.h. johnson jr, s. ferguson, c.d. tavares, l.j. pence, e.v. anslyn, p. ren, k.y. tsai, k.n. dalby. quantification of a pharmacodynamic erk end point in melanoma cell lysates: toward personalized precision medicine. acs medicinal chemistry letters 5 (2015) 47-52. [4] j.l. bos. ras oncogenes in human cancer: a review. cancer research 49 (1989) 4682-4689. [5] w.j. ding, w.p. tao, t. zeng, j.j. deng, h.b. lei, w. ge. selective erk inhibitor inhibits proliferation and induces apoptosis in lymphoma cell lines. international journal of clinical experimental medicine 9 (2016) 10955-10962. [6] u. germann, b. furey, j. roix, w. markland, r. hoover, a. aronov, m. hale, g. chen, g. martinezbotella, r. alargova, b. fan, d. sorrell, k. meshaw, p. shapiro, m. j. wick, c. benes, m. garnett, g. decrescenzo, m. namchuk, s. saha, d.j. welsch. the selective erk inhibitor bvd-523 is active in models of mapk pathway-dependent cancers, including those with intrinsic and acquired drug resistance. cancer research 75 (2015) abstract no. 4693. [7] j.r. infante, f. janku, a.w. tolcher, m.r. patel, r.j. sullivan, k. flaherty, r.d. carvajal, a.m. varghese, d.j. lee wong, m. sznol, j.a. sosman, a. wang-gillam, h.a. burris, a. ribas, s.p. patel, d.j. welsch, s. saha. dose escalation stage of a first-in-class phase i study of the novel oral erk 1/2 kinase inhibitor bvd-523 (ulixertinib) in patients with advanced solid tumors. journal of clinical oncology 33 (2015) abstract no. 2506. [8] r.s. obach. prediction of human clearance of twenty-nine drugs from hepatic microsomal intrinsic clearance data: an examination of in vitro half-life approach and nonspecific binding to microsomes. drug metabolism and disposition 27 (1999) 1350-1359. [9] d.f. mcginnity, m.g. soars, r.a. urbanowicz, r.j. riley. evaluation of fresh and cryopreserved hepatocytes as in vitro drug metabolism tools for the prediction of metabolic clearance. drug metabolism and disposition 32 (2004) 1247-1253. [10] z. wang, c.e. hop, k.h. leung, j. pang. determination of in vitro permeability of drug candidates through a caco-2 cell monolayer by liquid chromatography/tandem mass spectrometry. journal of mass spectrometry 35 (2000) 71-76. [11] p.v. balimane, y.h. han, s. chong. current industrial practices of assessing permeability and pglycoprotein interaction. journal of aaps 8 (2006) e1-13. [12] i. kariv, h. cao, k.r. oldenburg. development of a high throughput equilibrium dialysis method. journal of pharmaceutical sciences 90 (2001) 580-587. [13] r.s obach, r.l. walsky, k. venkatakrishnan, e.a. gaman, j.b. houston, l.m. tremaine. the utility of in vitro cytochrome p450 inhibition data in the prediction of drug-drug interactions. journal of pharmacology and experimental therapeutics 316 (2006) 336-348. [14] v. chu, h.j. einolf, r. evers, g. kumar, d. moore, s. ripp, j. silva, v. sinha, m. sinz, a. skerjanec . in vitro and in vivo induction of cytochrome p450: a survey of the current practices and recommendations: a pharmaceutical research and manufacturers of america perspective. drug metabolism and disposition 37 (2009) 1350-1359. [15] t.w. harper, p.j. brassil. reaction phenotyping: current industry efforts to identify enzymes responsible for metabolizing drug candidates. journal of aaps 10 (2008) 200-207. admet & dmpk 5(4) (2017) 212-223 preclinical assessment of ulixertinib doi: 10.5599/admet.5.4.437 223 [16] r. kumar, p.s. suresh, g. rudresh, m. zainuddin, p. dewang, r.r. kethiri, s. rajagopal, r. mullangi. determination of ulixertinib in mice plasma by lc-ms/ms and its application to a pharmacokinetic study in mice. journal of pharmaceutical biomedical analysis 125 (2016) 140-144. [17] b. davies, t. morris. physiological parameters in laboratory animals and humans. pharmaceutical research 10 (1993) 1093-1094. [18] k.t. flaherty, j.r. infante, a. daud, r. gonzalez, r.f. kefford, j. sosman, o. hamid, l. schuchter, j. cebon, n. ibrahim, r. kudchadkar, h.a. burris 3rd, g. falchook, a. algazi, k. lewis, g.v. long, i. puzanov, p. lebowitz, a. singh, s. little, p. sun, a. allred, d. ouellet, k.b. kim, k. patel, j. weber. combined braf and mek inhibition in melanoma with braf v600 mutations. new england journal medicine 367 (2012) 1694-1703. [19] l.m. chan, s. lowes, b.h. hirst. the abcs of drug transport in intestine and liver:efflux proteins limiting drug absorption and bioavailability. european journal of pharmaceutical sciences 21 (2004) 25-51. ©2017 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ manuscript doi: 10.5599/admet.2.4.48 235 admet & dmpk 2(4) (2014) 235-247; doi: 10.5599/admet.2.4.48 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper mucoadhesive polymer hyaluronan as biodegradable cationic/zwitterionic-drug delivery vehicle francisco torrens and gloria castellano 1 institut universitari de ciència molecular, universitat de valència, edifici d’instituts de paterna, p. o. box 22085, e-46071 valència, spain 1 facultad de veterinaria y ciencias experimentales, universidad católica de valencia san vicente mártir, guillem de castro 94, e-46001 valència, spain corresponding author: e-mail: torrens@uv.es; tel.: +34-963-544-431; fax: +34-963-543-274 received: july 23, 2014; revised: august 28, 2014; published: january 09, 2015 abstract mucoadhesive polymers in pharmaceutical formulations release drugs in mucosal areas. they interact and fix to mucus through molecular interpenetration, etc., which increase drug bioavailability. polymers physicochemical properties affect formulation mucoadhesion, rheological behaviour and drug absorption. hyaluronan (ha) is selected as a mucoadhesive and biodegradable polymer. geometric, topo logical and fractal analyses are carried out with program topo. reference calculations are performed with algorithm gepol. procedure topo underestimates molecular volume by 0.7 %. the error results 5 % in surface area and derived topological indices. the solvent-accessible surface is undercalculated by 3 %: from hexamer ha to ha·3ca and hydrate, the hydrophobic term rises by 42% and decays by 26 %, as well as the hydrophilic part drops by 14% and rises by 58 %, in agreement with the number of h-bonds. the accessibility rises by 9 % and decays by 8 %. the fractal dimension is underevaluated by 1 % and, for ha, it results 1.566; on going to ha·3ca and hydrate, it rises by 2 % and 1 %. the nonburied-atoms dimension increases by 11 %: for ha, it results 1.725. when going to ha·3ca and hydrate, it augments by 4 % and 0.3%. ongoing from ha to ha·3ca and hydrate, the external minus molecular dimension enlarges by 20 % and decays by 9 %. the hydrate globularity is lower than for water, ca 2+ and averages of o-atoms in ha. the rugosity of ca 2+ is smaller than for hydrate, averages of o-atoms in ha and water. the accessibilities of ca 2+ and water are greater than for hydrate. as cations exchange in ha·3ca requires ca 2+ alteration, rises of drug zwitterionic character and acidic ph increase absorption. keywords medicine absorption; medicine delivery; dipole moment; fractal dimension; metal hyaluronate; mucosa. introduction hyaluronic acid (ha) is a high-molecular-weight (mw) (hmw) polysaccharide present in the extracellular matrix of most vertebrate tissues [1]. its functions vary from maintaining constant volume of interstitial fluid to organizing extracellular matrix and immunosuppression [2]. its presence on plasma membranes and concentration variation in pericellular spaces are associated with cell aggregation during morphogenesis and metastasis formation during malignant transformation and tumours invasion [3–5]. it is an anionic, nonsulphated, linear, hmw polyglycosaminoglycan (cf. figure 1) consisting of repeating units http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:torrens@uv.es mailto:torrens@uv.es francisco torrens and gloria castellano admet & dmpk 2(4) (2014) 235-247 236 of disaccharide d-glucuronic acid (gcu)-(3)-n-acetyl-d-glucosamine (nag)-(4). an ha molecule consists of 10 4 gcu/nag. its anionic charge under physiological conditions is caused by gcu coo – , which m z+ interaction contributes to global supramolecular structure [6]. other factors include: ph, temperature, hydration and, especially, m z+ [7,8]. structural and literature data for transition metal complexes with ha are limited to aqueous co-ordination complexes with ca 2+ , ag + , cd 2+ , pb 2+ and fe 3+ [9–11]. x-ray fibre diffraction solved solid-state structure of m z+ ha where the formation of 2and 4-fold-helices was reported [12–15]. polyanion conformation is stabilized by h-bonds across glycosidic linkages between ha monomers. adjacent antiparallel chains are held together through –coo – –ca 2+ – – ooc– bridges and six hbonded water molecules. the polymer secondary structure is similar to ca 2+ ha for other m 2+ . amorphous m 2+ ha (m = cu, ni, mn, co) was prepared at ph 5.5 precipitating aqueous solutions with cold ethanol. local structure around m 2+ was determined by extended x-ray absorption fine structure (exafs) and x-ray absorption near edge structure (xanes) [16,17]. co-ordination polyhedron around cu 2+ is a distorted octahedron: four o-atoms at a distance of 1.95 å occupy planar equatorial sites; at axial places o-atoms are present at 2.46 å. though o-atoms are preferred at axial locations, n-atoms from nag cannot be excluded. o o o oh ho o o–o nh o oh ho n figure 1. disaccharide repeating unit of ha comprising gcu/nag. the mw ranges in 10 4 –10 7 g·mol –1 . by using quantum chemical methods, the basic gcu/nag unit was studied. semi-empirical and ab initio molecular orbital (mo) calculations showed optimized geometries in agreement with crystallographic data [18]. the ca/cu 2+ ha are amorphous materials. a combined quantum mechanical/molecular mechanics (qm/mm) approach [19] to enzymology and m z+ –protein/ha binding allowed structural elucidation [20– 28]. transition-m z+ binding was studied by using density-functional theory (dft) [29–33]. an ha:ca/cu 2+ complexation qm/mm was performed [34]. the ha plays a structural role in cartilage and other tissues. aggrecan, the cartilage proteoglycan, is bound to ha chains, the bond being stabilized by link proteins [35]. the formed aggregates of mw ~ 10 8 g·mol –1 are deposited within the collagen framework, without which interaction the proteoglycans would not be retained in cartilage. mucoadhesive polymers were used in pharmaceutical formulations to release drugs in mucosas [gastrointestinal (gi)/vaginal tracts, ocular mucosa, bucal/nasal cavity] [36]. they fix to mucus through mechanisms (molecular interpenetration, van der waals forces, hydrophobic interactions, electrostatic forces, h-bonds) increasing drug bioavailability [37–40]. in earlier publications, program topo was applied to the valence topological charge-transfer indices for the molecular dipole moment [41,42], as well as fractal dimension of percutaneous absorption enhancer phenyl alcohols [43] and 4-alkylanilines [44]. lysozyme showed hydrolytic activity vs. peptidoglycans [45– 48]. a new tool was described for interrogating macromolecular structure [49]. in the present report, ha geometric, topological and fractal analyses are performed with topo. the aim of this study is to find properties distinguishing ha·3ca·9h2o. the goal is to validate the indices by using ha differentiation. the ultimate reason for modelling is to improve pharmaceutical formulations to release drugs in mucosal areas as ha properties affect mucoadhesion, rheological behaviour and drug absorption with the goal of increasing biodisponibility and decreasing toxicity. the following section describes the experimental in admet & dmpk 2(4) (2014) 235-247 mucoadhesive polymer hyaluronan for drug delivery doi: 10.5599/admet.2.4.48 237 silico computational methods. next, two sections illustrate and discuss the results. finally, the last section summarizes our conclusions. experimental in our program topo for the theoretical simulation of crystal-fragment shape [50], structure surface is represented by the external surface of a set of overlapping spheres with appropriate radii, centred on the atomic nuclei [51,52]. a crystal fragment is treated as a solid in space defined by tracing spheres about atomic nuclei. it is computationally enclosed in a graduated rectangular box, and the geometric descriptors are evaluated by counting the points within the solid or close to chosen surfaces. the crystal-fragment volume is approximated as v = p·grid 3 , where p is the number of points within the fragment volume (within a distance rx of any atomic nucleus x) and grid is the mesh-grid size. as a first approximation, the crystal-fragment bare surface area was calculated as s = q·grid 2 , where q is the number of points close to the bare surface area (within a distance between rx and rx + grid of any atomic nucleus x). however, the estimate was improved: if a point falls exactly on the surface of one of the atomic spheres it accounts indeed for grid 2 units of area on fragment bare surface, which is because total surface of atom x can accommodate 4rx 2 /grid 2 points. when a point falls beyond the surface it represents grid 2 units of area on the surface of a sphere of radius r > rx, not on the surface of atom x on which it accounts only for a fraction of this quantity: grid 2 (rx/r) 2 . the total bare surface area is calculated as s = f·grid 2 , where f is the sum of elements af = rx 2 /r 2 (i) for those points close enough to the surface of any atom x. the rx 2 is the square radius of atom x and r 2 (i) is the square distance of point i from atomic nucleus x. two topological indices of crystal-fragment shape are calculated: g = se/s (se = surface of an equivalent sphere) stands for the fragment globularity and g’ = s/v denotes the fragment rugosity. the hydrated-system properties are related to the contact surface between solute and water molecules. another crystal-fragment geometric descriptor was proposed: the solvent-accessible surface area as [53], which is defined by using a probe sphere that is allowed to roll on the outside while maintaining contact with the bare surface [54]. it is calculated in the same way as the bare surface area by using pseudo-atoms whose van der waals radii [55] were increased by the probe radius r [56]. the accessibility is a dimensionless quantity ranged in 0–1 representing the ratio of solvent-accessible surface area in a particular structure to solvent-accessible surface area of the same atom when isolated from the crystal-fragment. the fractal dimension, d, results [57]: d 2 d(logas) d(logr) (1) it provides a quantitative indication of the degree of surface accessibility towards different solvents [58]. program topo allows an atom-to-atom analysis of d on every atom i to obtain an atomic dimension index di from the atomic contributions to asi. the di is weight averaged to obtain a new crystal-fragment dimension index d’ = (iasidi)/as, where asi are used as weights for di. if an asi = 0 for any probe, di cannot be calculated for atom i and this does not contribute to d’, which represents a d averaged for atoms nonburied to any solvent-accessible surface in the range of probe spheres. a version of topo was implemented in algorithms amyr [59], gepol [60] and surmo2 [61]. procedure amyr performs the theoretical simulation of molecular associations and chemical reactions. software gepol performs a francisco torrens and gloria castellano admet & dmpk 2(4) (2014) 235-247 238 triangular tessellation of the crystal-fragment surface. it is used for reference calculations. codes topo and gepol recognize the cavity-like spaces in crystal-fragments and are adequate to study intercalation compounds; however, surmo2 does not distinguish internal cavities. combination surmo2 with topo or gepol allows characterizing the crystal-fragment surface of the cavities. our version of surmo2 was corrected for deviation from the spherical shape, by dividing the contribution of every point by the cosine of the angle formed by the semi-axis and the corresponding normal vector to the surface at this point. the volume and surface areas of the crystal fragments with cavities were corrected by maximizing in every angular orientation the distance of the most distant atom in each semi-axis. calculation results doubly crosslinked networks (dxns) were engineered by embedding gelatine conjugated ha hydrogel particles (hgps) (ghgps) in a secondary net established by ha glycidyl methacrylate (gma) (hagma) (cf. figure 2) [62]. r= 2. 1. 3. o o o nh nh 2 o h figure 2. chemical modifications in synthetic procedures for the fabrication of ha hgps and dxns. for drug administration on mucous membranes and tissues, as well as skin, the active principle is prepared in matrix systems with a hydrophilic polymer {carbopol ® [poly(acrylic acid) (paa), cf. figure 3a], ha} and alcohol [propylene glycol (pg, figure 3b), poly(ethylene glycol) (peg, figure 3c)]. ch o oh ch 2 n h o o h n ho oh a b c figure 3. (a) poly(acrylic acid) (paa); (b) propylene glycol (pg); (c) poly(ethylene glycol) (peg). fourier-transform infrared (ftir) spectroscopy showed that the carboxylic acid groups cooh of paa react completely with the alcohol groups oh in the matrix [63]: ~o–h + h–o–c(=o)~  ~o–c(=o)~ + h2o (2) in gel ca alginate used for wound dressing, ca 2+ releasing helps in healing and exchanges exuded na + . stable cationic colloids were elaborated by using chitosan (cs):ha polyelectrolyte complexation [64]. ag + /ag3po4 nanoparticles:ha/cs complexes (cf. figure 4) resulted antimicrobial [65,66]. hydrogels are used in medical applications, e.g., implants, tissue engineering and contact lenses. admet & dmpk 2(4) (2014) 235-247 mucoadhesive polymer hyaluronan for drug delivery doi: 10.5599/admet.2.4.48 239 o o o oh ho o o–o nh o oh ho agno3 o o o oh ho o oo nh o oh ho ag n n figure 4. suggested structure of complex of hyaluronic acid (ha) and silver ag + . the ha was selected as a mucoadhesive and biodegradable polymer. polar molecules (cf. table 1) administered with ha show high dipole moment. notice the low mw of metronidazole, etc. and the low human skin permeability of caffeine used as transdermal anticellulite. some small, lipid-soluble drugs cross the blood–brain barrier (bbb) simply by diffusion through the cell membrane and others, e.g., caffeine, enter successfully through specialized transporter proteins. a model showed that the decimal logarithm of 1-octanol–water partition coefficient (log p) positively correlated and mw negatively associated with log kp [67]: logkp 6.30.71logp 0.0061mw, n 93 r  0.82 (3) where n is the number of points and r, correlation coefficient. table 2 shows the administration routes of drug metronidazole for bacterial, fungal and protozoal vaginitis [68,69]. table 1. molecular dipole moments of water, theophylline and polar molecules administered with ha. molecule μ (d) a ref. b mw c pka d logp e logkp f uses water 1.861 1.85 18 15.74 -1.38 -6.71 solvent and metabolism eugenol 2.378 – 164 10.19 2.27 – local antiseptic and anaesthetic theophylline 3.262 – 180 8.81 -0.773 – respiratory diseases caffeine 3.567 3.64 194 10.4 -0.07 -7.14 cellulite metronidazole 3.681 – 171 2.62 -0.02 – bacterial/fungal/protozoal vaginitis terconazole 4.098 – 532 <1.5 3.51 – antifungal minoxidil 4.292 – 209 4.61 1.24 – androgenic alopecia betamethasone 5.346 – 392 12.42 1.94 – anti-inflammatory and immunosupressor a μ: molecular dipole moment (debye) calculated with mopac-am1 [70], b μ: calculated molecular dipole moment (debye), c mw: molecular weight (g·mol –1 ), d ka: acid dissociation constant (mol·l –1 ) , e p: 1-octanol–water partition coefficient, f kp: human skin permeability (cm·s –1 ). table 2. administration routes of drug metronidazole for bacterial, fungal and protozoal vaginitis. route onset peak (hr) duration (hr) oral (po) rapid 1–3 8 po (extended) rapid unknown up to 24 intravenous (iv) rapid end of infusion 6–8 topical unknown 6–12 unknown vaginal unknown 6–12 12 an ha comparative analysis was performed in three forms: hexamer (trimer of heterodimers) ha (cf. figure 5), ha·3ca and ha·3ca·9h2o. the geometry of the systems was taken from the protein data bank code 4hya (x-ray fibre diffraction). francisco torrens and gloria castellano admet & dmpk 2(4) (2014) 235-247 240 figure 5. wire-frame molecular structure of ha (perspective view). in 4hya, nine crystallization water molecules are included most strongly bound in the first hydration shell (cf. figure 6). figure 6. hydrogen-suppressed ball-and-stick structure of hyaluronic acid hexamer ha·3ca·9h2o. the 4hya consists of six saccharide residues (126 atoms, cf. table 3) with mw = 1123g·mol –1 . the glycosaminoglycan hexamer co-ordinates three ca 2+ ions also surrounded by at least nine water molecules. the 11 n/o–h…o h-bonds include one n–h…o and 10 o–h…o, e.g., three pairs of h-shared h-bonds. many h-bonds are formed because of the polyanionic nature of ha main chain. table 3. number of atoms, hydrogen bonds and molecular weight of hyaluronic acid (ha). molecule no. of atoms mw (g·mol –1 ) a number of n/o–h…o hydrogen bonds ha 126 1123 1 n–h…o ha·3ca 129 1243 1 n–h…o ha·3ca·9h2o 152 1401 11 (1 n–h…o + 10 o–h…o, 6 shared) a mw: molecular weight (g·mol –1 ). figure 7 illustrates four out of 11 n/o–h…o h-bonds in ha·3ca·9h2o: one n–h…o and three out of 10 o–h…o. admet & dmpk 2(4) (2014) 235-247 mucoadhesive polymer hyaluronan for drug delivery doi: 10.5599/admet.2.4.48 241 figure 7. four of the n/o–h…o hydrogen bonds of hyaluronic acid hexamer ha·3ca·9h2o. van der waals molecular surface of ha·3ca·9h2o (cf. figure 8) shows some small open cavities. figure 8. van der waals molecular surface of hyaluronic acid hexamer ha·3ca·9h2o. francisco torrens and gloria castellano admet & dmpk 2(4) (2014) 235-247 242 geometric and topological analyses were performed with our program topo (cf. table 4). reference calculations were performed with our version of gepol. algorithm topo underestimated the molecular volumes v and surface areas s by 0.7% and 5 %, respectively. surface-derived molecular globularities g were overestimated by 5% and rugosities g’ resulted underestimated by 5 %. table 4. geometric descriptors and topological indices of hyaluronic acid (ha). molecule v a v ref. b s c s ref. b g d g ref. b g’ e g’ ref. b ha 816.4 822.8 942.78 999.24 0.448 0.425 1.155 1.214 ha·3ca 913.3 919.9 1069.45 1128.66 0.426 0.405 1.171 1.227 ha·3ca·9h2o 1070.0 1077.8 1247.91 1320.73 0.405 0.385 1.166 1.225 a molecular volume (å 3 ), b reference calculation carried out with program gepol, c molecular surface area (å 2 ), d molecular globularity, e molecular rugosity (å –1 ). figure 9a shows the wire-frame molecular structure of ha. figure 9b reveals the ball-and-stick (b&s) molecular model of ha·3ca·9h2o. the corey-pauling-koltun (cpk) molecular structure of ha·3ca·9h2o (figure 9c) defines the van der waals molecular surface area s. notice some small open cavities on cpk surface in agreement with figure 8. the water-accessible surface cpk + of ha·3ca·9h2o (figure 9d) delimits the solvent-accessible surface area as. most open cavities on cpk (figure 9c) close on cpk + surface meaning that they are not accessible to a water molecule (radius r = 1.25 å). the cpk ++ of ha·3ca·9h2o (figure 9e) denotes the surface accessible by a protein side chain (r = 3.50 å) and is used to calculate the fractal dimension d. again, most open cavities on cpk (figure 9c) close on cpk ++ , which are not accessed by a protein side chain. structures comparison suggests that ha·3ca·9h2o reaction with a molecule or drug begins with water and, even, ca 2+ exchanges in the co-ordination compound. as many drugs are either cationic or zwitterionic, e.g., 65 % of orally available drugs are charged or zwitterionic, ca 2+ exchange is favoured by electrostatics at acidic ph; e.g., drugs metronidazole and minoxidil are ionic pairs. figure 9. ha·3ca·9h2o molecular structures: (a) wire frame; (b) b&s; (c) cpk; (d) cpk + ; (e) cpk ++ . admet & dmpk 2(4) (2014) 235-247 mucoadhesive polymer hyaluronan for drug delivery doi: 10.5599/admet.2.4.48 243 surface cpk + of ha (cf. figure 10) shows more open cavities than ha·3ca·9h2o (figure 9d), which could be accessed by a water molecule. figure 10. cpk + molecular surface of ha. program topo allows the atom-to-atom partition analysis of all geometric descriptors. water solventaccessible surface analysis (cf. table 5) shows that the solvent-accessible surface (as) area was underestimated by 3 %. when going from hexamer ha to ha·3ca and ha·3ca·9h2o the hydrophobic solvent-accessible surface (hbas) area rose by 42% and decayed by 26 %, respectively. on the other hand, the hydrophilic solvent-accessible surface (hlas) area dropped by 14 % and increased by 58 %. it resulted the most sensitive geometric descriptor for the presence of ca 2+ and, especially, water, in agreement with the highest number of h-bonds in the hydrated co-ordination complex (table 3). ongoing from ha to ha·3ca and ha·3ca·9h2o, the accessibility enlarged by 9% and decreased by 8%. the fractal dimension d was underevaluated by 1 %. the fractal dimension of ha turned out to be 1.566. when going to ha·3ca and ha·3ca·9h2o, it rose by 2 % and 1 %. the low accessibility and relatively low dimension of ha·3ca·9h2o suggest that its reaction requires water or ca 2+ exchange, in agreement with the closed character of cpk +/++ surfaces (figure 9). the fractal dimension averaged for nonburied atoms d’ increased by 11% with regard to the molecular dimension. in particular, for ha the external dimension resulted 1.725. ongoing to ha·3ca and ha·3ca·9h2o it augmented by 4% and 0.3%. when going from ha to ha·3ca and ha·3ca·9h2o, the difference nonburied minus molecular dimensions (d’–d) enlarged by 20 % and decayed by 9 %. it resulted greatly sensitive to ca 2+ and water incidences. table 5. geometric descriptors and fractal dimensions of the solvent-accessible surface of ha. molecule as a as ref. b hbas c hlas d accessibility e d f d ref. b d’ g d’–d ha 1255.03 1293.95 589.72 665.31 11.14 1.566 1.582 1.725 0.159 ha·3ca 1409.77 1453.85 834.57 575.20 12.09 1.601 1.619 1.792 0.191 ha·3ca·9h2o 1530.07 1578.40 620.88 909.19 11.15 1.625 1.644 1.798 0.173 a water solvent-accessible surface area (å 2 ), b reference calculation carried out with program gepol, c hydrophobic solventaccessible surface area (å 2 ), d hydrophilic solvent-accessible surface area (å 2 ), e accessibility (%), f molecular fractal dimension, g molecular fractal dimension averaged for nonburied atoms. atom-to-atom partition analyses of ha·3ca·9h2o topological indices and fractal dimension, cf. table 6, show that globularity g of hydrated co-ordination complex was lower than for o-atoms in water, ca 2+ and averages of o-atoms in ha in 2.0–2.4. rugosity g’ of ca 2+ was smaller than for hydrate, o-atom averages in ha in 1.2–1.3å –1 and o-atoms in water. hydrate accessibility lay in the range of o-atom averages in ha (5.9–11.9%); however, accessibilities of ca 2+ and o-atoms in water (34%) were much greater. on the other hand, the fractal dimensions of hydrate, ca 2+ and o-atoms in water lay in the range of averages of o-atoms in ha (1.1–3.6). in particular, maximum fractal dimension corresponds to gcu o-atoms for which the utmost reactivity is expected, in agreement with m z+ -interchange and ag + reactions (figure 4). francisco torrens and gloria castellano admet & dmpk 2(4) (2014) 235-247 244 table 6. atom-atom partition analyses of ha·3ca·9h2o topological indices and fractal dimension. fragment g a g’ b accessibility (%) d c average of o atoms in nag1 2.415 1.177 11.94 1.104 average of o atoms in gcu1 2.038 1.295 8.95 3.597 average of o atoms in nag2 2.067 1.290 8.11 1.710 average of o atoms in gcu2 2.069 1.270 5.93 1.358 average of o atoms in nag3 2.332 1.215 9.36 2.245 average of o atoms in gcu3 1.975 1.293 10.66 1.801 average of o atoms in nag 2.271 1.227 9.80 1.686 average of o atoms in gcu 2.027 1.286 8.51 2.252 average of o atoms in ha 2.138 1.259 9.10 1.995 average of ca 2+ cations 1.430 1.087 33.57 1.492 average of o atoms in h2o 1.258 1.733 33.54 1.700 ha·3ca·9h2o 0.405 1.166 11.15 1.625 a molecular globularity, b molecular rugosity (å –1 ), c molecular fractal dimension. discussion natural polysaccharide ha gelifies at ph < 4. it is characterized by its regenerative properties on different mucous zones, e.g., gi, oral, vaginal, etc. the ha is mainly used in medical and cosmetic applications. among the many benefits hmw-ha holds in medicine are: (1) the maintenance of tissue space for surgery, (2) protection of cells and tissue and (3) therapeutic effectiveness. healing efficiency of ha depends critically on mw: the higher the mw, the longer its benefit. the use of ha in cosmetics presents an impact on public health because of the risk of ingredient penetration into the deeper skin layers, which are cleared by systemic circulation (systemic absorption is more likely to occur in injured skin, e.g., sunburnt, atopic, eczematous, psoriatic skin). the problem is important in the case of anticellulite creams as the tissue target is on the dermis. caffeine, coleus forskohlii and nelumbo nucifera extracts are efficacious ingredients of anticellulite cosmetics. in order to develop formulations, it is interesting to enhance their low skin permeabilities (table 1). however, care should be taken concerning toxicity. finally, it is proposed that (1) polymer coohs react completely with matrix ohs to give esters, (2) polymer reaction with drugs begins with water or, even, ca 2+ exchange, (3) matrix rheological properties, e.g., phdependent gelification, etc., are important and (4) caffeine is a potential lead structure in drug design. conclusions from the present results and discussion the following conclusions can be drawn: 1. the advantage of program topo with respect to gepol is that the former allows an atom-to-atom partition analysis of the geometric, topological, fractal indices and their combination: hydrophobic and hydrophilic solvent-accessible surface areas. the latter was detected as the most sensitive geometric descriptor for the presence of ca 2+ and, especially, water, in agreement with the number of h-bonds, which were calculated with topo. ongoing from hyaluronan to ca 2+ co-ordinated to hydrate, accessibility rose by 9 % and decayed by 8 %. the relative error with program topo is 0.7 % in volume, as well as 5 % in surface area and its derived indices. 2. the fractal dimension was underevaluated by 1 %. ongoing to ca 2+ co-ordinated and hydrate, fractal dimension of hyaluronan (1.566) rose by 2 % and 1 %, respectively. the fractal dimensions of nonburied atoms increased by 11 % with regard to molecular fractal indices. ongoing to ca 2+ co-ordinated to hydrate, fractal dimension of external atoms (1.725) augmented by 4 % and 0.3 %. nonburied minus molecular fractal dimension enlarged by 20 % and decayed by 9 %. it was sensitive to ca 2+ and water occurrences. admet & dmpk 2(4) (2014) 235-247 mucoadhesive polymer hyaluronan for drug delivery doi: 10.5599/admet.2.4.48 245 3. the hydrate globularity was lower than for o-atoms in water, ca 2+ and o-atom average in hyaluronan. the rugosity of ca 2+ was smaller than for hydrate, o-atoms average in hyaluronan and o(water). the accessibility of ca 2+ and o(water) was much greater than for hydrate. the maximum fractal dimension corresponds to d-glucuronic-acid o-atoms for which the utmost reactivity is expected, in agreement with m z+ interchange and ag + reactions. 4. as cations exchange in hyaluronan co-ordination compound and action requires ca 2+ alteration, the reinforcement of drug cationic/zwitterionic character and acidic ph increases absorption. hyaluronan is an articular-cartilage component where it is present as coat around chondrocytes. when aggrecan monomers bind to it in link-protein presence, large, anionic aggregates form, which are responsible for cartilage resilience. its water interaction, represented by 10–15 h-bonds per disaccharide repeat, indicates binding capacity and importance for cartilage viscoelastic properties. in mucosas, a need for the elimination of pathogens and prophylaxis exists. the hyaluronan rheological behaviour in ph 4–7 did not differ. however, ph < 4 generated gels because of hydrophobic interactions and h-bonds, which resulted promising for drug administration on mucous membranes and skin. further work will analyze polymer reactions with alcohols in the matrix/drugs (eugenol, metronidazole, betamethasone, etc.), matrix rheological properties and caffeine as a potential lead structure in drug design. caffeine, etc. are efficacious ingredients of anticellulite cosmetics. however, care should be taken concerning toxicity, especially for injured skin. work is in progress on the prediction of chromatographic retention times and molecular classification of (1) tea compounds and (2) methylxanthines (caffeine and its metabolites) and nicotine-metabolite cotinine in human plasma. acknowledgements: f. t. thanks support from the spanish ministerio de economía y competitividad (project no. bfu2013-41648-p) and eu erdf. references [1] e.a. balazs, ed., the chemistry and molecular biology of the intercellular matrix, academic, new york, ny, 1970 [2] b. alberts, a. johnson, j. lewis, m. raff, k. roberts, p. walter, molecular biology of the cell, garland, new york, ny, 2002 [3] t.c. laurent, j.r.e. fraser, faseb j. 6 (1992) 2397-2404 [4] k.l. goa, p. benfield, drugs 47 (1994) 536-566 [5] e. zimmerman, b. geiger, l. addadi, biophys. j. 82 (2002) 1848-1857 [6] m. moulabbi, h. broch, l. rober, d. vasilescu, j. mol. struct. (theochem) 395-396 (1997) 477-508 [7] p. sipos, m. veber, k. burger, j. illes, g. machula, acta chim. hung. models chem. 129 (1992) 671683 [8] e.r. morris, d.a. rees, e.j. welsh, j. mol. biol. 138 (1980) 383-400 [9] n. figueroa, b. nagy, b. chakrabarti, biochem. biophys. res. commun. 74 (1977) 460-465 [10] h. sterk, m. braun, o. schmut, h. feichtinger, carbohydr. res. 145 (1985) 1-11 [11] l. lapcik, c. dammer, m. valko, colloid polym. sci. 201 (1992) 1049-1052 [12] j.k. sheehan, e.d. atkins, int. j. biol. macromol. 5 (1983) 215-221 [13] e.d. atkins, j.k. sheehan, c.f. phelps, biochem. j. 128 (1972) 1255-1263 [14] a.k. mitra, s. raghunathan, j.k. sheehan, s. arnott, j. mol. biol. 169 (1983) 829-859 [15] w.t. winter, s. arnott, j. mol. biol. 117 (1977) 761-784 francisco torrens and gloria castellano admet & dmpk 2(4) (2014) 235-247 246 [16] e. tratar pirc, i. arcon, p. bukovec, a. kodre, carbohydr. res. 324 (2000) 275-282 [17] e. tratar pirc, i. arcon, a. kodre, p. bukovec, carbohydr. res. 339 (2004) 2549-2554 [18] r. eklund, g. widmalm, carbohydr. res. 338 (2003) 393-398 [19] a. warshel, m. levitt, j. mol. biol. 103 (1976) 227-249 [20] b.r. brooks, r.e. bruccoleri, b.d. olafson, d.j. states, s. swaminatham, m. karplus, j. comput. chem. 4 (1983) 187-217 [21] a.d. becke, phys. rev. a 38 (1988) 3098-3100 [22] c.t. lee, w.t. yang, g. parr, phys. rev. b 37 (1988) 785-789 [23] a.d. mackerell jr., d. bashford, m. bellott, r.l. dunbrack jr., j.d. evanseck, m.j. field, s. fischer, j. gao, h. guo, s. ha, d. joseph-mccarthy, l. kuchnir, k. kuczera, f.t.k. lau, c. mattos, s. michnick, t. ngo, d.t. nguyen, b. prodhom, w.e. reiher iii, b. roux, m. schlenkrich, j.c. smith, r. stote, j. straub, m. watanabe, j. wiórkiewicz-kuczera, d. yin, m. karplus, j. phys. chem. b 102 (1998) 35863616 [24] e. tratar pirc, j. zidar, p. bukovec, m. hodoscek, carbohydr. res. 340 (2005) 2064-2069 [25] l. masgrau, a. roujeinikova, l.o. johannissen, p. hothi, j. basran, k.e. ranaghan, a.j. mulholland, m.j. sutcliffe, n.s. scrutton, d. leys, science 312 (2006) 237-241 [26] j. zidar, e. tratar pirc, m. hodoscek, p. bukonec, j. chem. model. 48 (2008) 283-287 [27] y. zhao, d. truhlar, theor. chem. acc. 120 (2008) 215-241 [28] b.r. brooks, c.l. brooks iii, a.d. mackerell jr., l. nilsson, r.j. petrella, b. roux, y. won, g. archontis, c. bartels, s. boresch, a. caflisch, l. caves, q. cui, a.r. dinner, m. feig, m. fischer, j. gao, m. hodoscek, w. im, k. kuczera, t. lazaridis, j. ma, v. ovchinnikov, e. paci, r.w. pastor, c.b. post, j.z. pu, m. schaefer, b. tidor, r.m. venable, h.l. woodcock, x. wu, w. yang, d.m. york, m. karplus, j. comput. chem. 30 (2009) 1545-1614 [29] t.e. hofstetter, c. howder, g. berden, j. oomens, p.b. armentrout, j. phys. chem. b 115 (2011) 12648-12661 [30] m. remko, d. fitz, r. broer, b.m. rode, j. mol. model. 17 (2011) 3117-3128 [31] m.h. khodabandeh, r. hamid, k. zare, m. zahedi, int. j. mass spectrom. 313 (2012) 47-57 [32] p.e. siegbahn, j. biol. inorg. chem. 11 (2006) 695-701 [33] y. minenkov, å. singstad, g. occhipinti, v.r. jensen, dalton trans. 41 (2012) 5526-5541 [34] e. tratar pirc, j. zidar, p. bukovec, int. j. mol. sci. 13 (2012) 12036-12045 [35] d. heinegård, å. oldberg, faseb j. 3 (1989) 2042-2051 [36] g.s. asane, s.a. nirmal, k.b. rasal, a.a. naik, m.s. mahadik, y.m. rao, drug dev. ind. pharm. 34 (2008) 1246-1266 [37] n.a. peppas, p.a. buri, j. control release 2 (1985) 257-275 [38] j.m. gu, j.r. robinson, s.h. leung, crit. rev. ther. drug carrier syst. 5 (1988) 21-67 [39] j. woodley, clin. pharmacokinet. 40 (2001) 77-84 [40] j.d. smart, adv. drug deliv. rev. 57 (2005) 1556-1568 [41] f. torrens, molecules 8 (2003) 169-185 [42] f. torrens, molecules 9 (2004) 1222-1235 [43] f. torrens, leb. sci. j. 5(1) (2004) 61-70 [44] f. torrens, g. castellano, j. liq. chromatogr. relat. technol. 31 (2008) 2337-2347 [45] f. torrens, g. castellano, a. campos, c. abad, j. mol. struct. 924-926 (2009) 274-284 [46] f. torrens, g. castellano, anal. chim. acta 654 (2009) 2-10 [47] f. torrens, g. castellano, j. life sci. 5 (2011) 167-181 [48] f. torrens, g. castellano, asia pacific j. life sci. 5 (2013) 89-121 [49] f. torrens, g. castellano, j. mater. sci. eng. b 4(2) (2014) 55-63 admet & dmpk 2(4) (2014) 235-247 mucoadhesive polymer hyaluronan for drug delivery doi: 10.5599/admet.2.4.48 247 [50] f. torrens, e. ortí, j. sánchez-marín, j. chim. phys. phys.-chim. biol. 88 (1991) 2435-2441 [51] a.y. meyer. j. chem. soc., perkin trans. 2 1985 (1985) 1161-1169 [52] a.y. meyer, j. comput. chem. 9 (1988) 18-24 [53] b. lee, f.m. richards, j. mol. biol. 55 (1971) 379-400 [54] r.b. hermann, j. phys. chem. 76 (1972) 2754-2759 [55] a. bondi, j. phys. chem. 68 (1964) 441-451 [56] s.j. wodak, j. janin, proc. natl. acad. sci. u. s. a. 77 (1980) 1736-1740 [57] m. lewis, d.c. rees, science 230 (1985) 1163-1165 [58] f. torrens, j. sánchez-marín, i. nebot-gil, j. comput. chem. 22 (2001) 477-487 [59] f. torrens, m. rubio, j. sánchez-marín, comput. phys. commun. 115 (1998) 87-89 [60] j.l. pascual-ahuir, e. silla, j. tomasi, r. bonaccorsi, j. comput. chem. 8 (1987) 778-787 [61] b. terryn, j. barriol, j. chim. phys. phys.-chim. biol. 78 (1981) 207-212 [62] x. xu, a.k. jha, d.a. harrington, m.c. farach-carson, x. jia, soft matter 8 (2012) 3280-3294 [63] s.a.a. al-abdulla, n.a.h. al-assady, s.g. syyah, j. thi-qar sci. 3 (2012) 129-140 [64] r.c. polexe, t. delair, molecules 18 (2013) 8563-8578 [65] a.m. abdel-mohsen, r. hrdina, l. burget, g. krylova, r.m. abdel-rahman, a. krejcova, m. steinhart, l. benes, carbohydr. polym. 89 (2012) 411-422 [66] d. chudobova, l. nejdl, j. gamulec, o. krystofova, m.a. merlos rodrigo, j. kynicky, b. ruttkaynedecky, p. kopel, p. babula, v. adam, r. kizek, int. j. mol. sci. 14 (2013) 13592-13614 [67] r.o. potts, r.h. guy, pharm. res. 9 (1992) 663-669 [68] c.d. freeman, n.e. klutman, k.c. lamp, drugs 54 (1997) 679-708 [69] s. löfmark, c. edlund, c.e. nord, clin. infect. dis. 50 (2010) s16-s23 [70] m.j.s. dewar, e.g. zoebisch, e.f. healy, j.j.p. stewart, j. am. chem. soc. 107 (1985) 3902-3909 ©2014 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ bioenhancing effects of naringin on atorvastatin doi: 10.5599/admet.5.4.647 174 admet & dmpk 7(3) (2019) 174-182; doi: http://dx.doi.org/10.5599/admet.647 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper bioenhancing effects of naringin on atorvastatin venkatesh sama 1 *, balaraju pagilla 1 , rajeswari chiluka 1 , ravi alvala 1 , ravi kumar pola 1 , ramesh mullangi 2 1 g. pulla reddy college of pharmacy, mehdipatnam, hyderabad-500028, india 2 jubilantbiosys, industrial suburb, yeshwanthpur, bangalore-566022, india *corresponding author: e-mail: venkateshsama@hotmail.com; tel.: +91-40-2351 7222 received: december 15, 2018; revised: march 18, 2019; published online: june 6, 2019 abstract naringin (cas no: 10236-47-2) is a flavonone glycoside obtained from citrus paradisi (grapefruit), a natural bioenhancer and reported to enhance the bioavailability of drugs by inhibiting cytochrome p450 and p glycoprotein (p-gp). the aim of the present study was to investigate the effect of naringin on antihyperlipidemic properties of atorvastatin (ast) in tyloxapol induced hyperlipidemic rats and the effects were supported with measurement of plasma concentrations of ast by hplc method. animals received ast along with naringin (15 and 30 mg/kg) shown higher percent reduction in both cholesterol and triglycerides levels, when compared to animals received ast alone at dose of 25 and 50 mg/kg and it was found that the higher percent reduction in cholesterol and triglycerides was proportional to increase in plasma concentration of ast. from the results it is evident that the co-administration of naringin along with ast increased the plasma concentration of ast. the findings of the present study confirmed that naringin could be used as bioenhancer. the co-administration of ast and the diet with naringin (grapefruit) to the patients may potentiate the therapeutic efficacy of ast. keywords naringin; bioenhancer; atorvastatin; tyloxapol; plasma concentrations; rp-hplc introduction recent advancement in bioavailability enhancement of drugs by compounds of herbal origin has produced a revolutionary shift in the way of therapeutics. poorly bioavailable drugs remain sub-therapeutic because a major portion of a dose never reaches the plasma or exerts its pharmacological effect unless and until very large doses are given, which may lead to serious side effects. any significant improvement in bioavailability will result in lowering the dose or the dose frequency of that particular drug. the global focus is now on methods aimed at reducing drug dosage, and thus drug treatment cost [1]. bioenhancers of herbal origin are reported to enhance the bioavailability and bioefficacy of a therapeutic drugs or nutrients with which it is combined, without any typical pharmacological activity of its own at the dose used. the concept of bioenhancers of herbal origin can be tracked back from the ancient knowledge of indian system of medicine (ayurveda). bose (1929) reported an enhanced antiasthmatic effect of an ayurvedic formula containing vasaka (adhatoda vasica) when administered with long pepper [2]. bioenhancers act through several mechanisms. they affect the absorption process, drug metabolism or on http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:venkateshsama@hotmail.com admet & dmpk 7(3) (2019) 174-182 effect of naringin on atorvastatin doi: 10.5599/admet.647 175 drug target. among the known reasons for the use of bioenhancers is its non-toxicity, effective at low concentration levels and simple formulation processes. bioenhancers are effective when administered with other drug classes such as antibiotics, antituberculosis, antiviral, antifungal and anticancer drugs. bioenhancers also improve oral absorption of a wide range of nutrients such as vitamins, minerals and amino acids etc. [3]. naringin (7-[[2-o-(6-deoxy-α-l-mannopyranosyl)-β-d-glucopyranosyl]oxy]-2,3-dihydro-5-hydroxy-2-(4hydroxyphenyl)-4h-1-benzopyran-4-one), a flavanone glycoside (rutinoside), occurs naturally in the pericarp of citrus fruits and grapefruits (citrus paradisi; family: rutaceae) [4]. naringin possess the ability of inhibition of the p-gp efflux pump [5,6] and also inhibit cytochrome p450 in particular cyp3a4 [7-9]. literature search revealed that naringin has antiviral 10], anticancer [11], hepatoprotective [12], antiinflammatory [13], anti-ulcer [14,15] and antioxidant activities [16]. naringin has an inhibitory effect against cyp3a4 activity in human liver microsomes [9] and promotes ascorbic acid induced lipid peroxidation [17]. naringin has a potent acyl coa-cholesterol-o-acyltransferase (acat-promotes the esterification of cholesterol in blood) inhibitory activity, macrophage-lipid complex accumulation inhibitory activity and preventive or treating activity on the hepatic diseases [18]. previous studies have reported that pretreatment of naringin appeared to be effective to alter the pharmacokinetics of drugs (verapamil, diltiazem, paclitaxel, tamoxifen) that are substrates of p-gp and/or cyp3a over the dose range of 3-30 mg/kg in rats/rabbits [19-22]. atorvastatin (ast) chemically is a calcium salt of (βr, δr)-2-(4-fluorophenyl)-β,δ-dihydroxy-5-(1methylethyl)–3-phenyl-4-[(phenylamino)-carbonyl]-1h-pyrrole-1-heptanoic acid (2:1) trihydrate. ast (lipitor®) is a 3-hydroxy-3-methyl-glutaryl coenzyme a (hmg-coa) reductase inhibitor and used for treating various dyslipidemic disorders. ast is rapidly absorbed from gastrointestinal tract. the extent of absorption increases in proportion to the dose and the absolute bioavailability is 12 %. protein binding is very high (≥98 %) and undergoes extensive hepatic or extra hepatic metabolism. pharmacokinetic studies of ast revealed the formation of ortho and parahydroxylated derivatives as primary metabolites. the cytochrome p450 family in particular cyp3a4 is potential to catalyse the formation of ortho and parahydroxylated derivatives [23,24]. the wide range of doses ranging from 10-80 mg of ast is in use as per the lipid lowering profile requirement. due to low bioavailability of ast high dosing and/or repeated administration is required to achieve steady state of cmax. this research paper aims at reporting the effect of naringin on the antihyperlipidemic activity of ast in tyloxapol induced hyperlipidemia in wistar rats. the observed activity was correlated with plasma concentration of ast in experimental animals. materials and methods chemicals and reagents a gift sample of ast (purity >99.5 %) is obtained from m/s, biocon pharmaceuticals, bangalore, india. naringin procured from sigma-aldrich, mumbai, india. tyloxapol (isooctyl-polyoxyethylene phenol/triton wr 1339) was purchased from himedia (b. no: 25301-02-4) mumbai, india. hplc grade water, acetonitrile, methanol and orthophosphoric acid were purchased from sd fine-chem limited and sodium dihydrogen orthophosphate from otto chemicals, mumbai, india. disodium salt of edta is procured from sd finechem limited, mumbai, india. anesthetic ether is obtained from tkm pharma, hyderabad, india. diagnostic kits for total cholesterol and triglycerides were obtained from span diagnostics ltd, surat, r. mullangi et al. admet & dmpk 7(3) (2019) 174-182 176 gujarat, india. cellulose acetate filters, pore size 0.2 µm obtained from sartorius stedim biotech, germany. all aqueous solutions including the buffer for mobile phase was prepared with hplc grade water. control wistar rat plasma was obtained from department of pharmacology, g. pulla reddy college of pharmacy, hyderabad, india. experimental animals male wistar rats weighing (~3 months age, weighing 180-200 g) were procured from national institute of nutrition, hyderabad, india. animals were maintained in standard cages under controlled laboratory conditions. the animals had free access to feed (national institute of nutrition, hyderabad) and water ad libitium during quarantine period of seven days. animals were fasted ~12 h before experiment but had been allowed free access to water. the institutional animal ethics committee of g. pulla reddy college of pharmacy, hyderabad, india has approved the animal experimental protocols. effect of naringin on ast in tyloxapol induced hyperlipidemia the method tyloxapol induced hyperlipidemia in rats was performed as described by vogel [25]. hyperlipidemia was induced by single intraperitonial injection of 15 % w/v tyloxapol in sterile normal saline at a dose of 400 mg/kg. after 30 h, rats with marked hypercholesterol were separated and divided into 10 groups of 6 animals each. all the test substances were administered orally as a fine aqueous suspension of 0.5 % w/v carboxy methyl cellulose (cmc). group 1 served as disease control and received vehicle. group 2 and 3 received naringin at a dose of 15 and 30 mg/kg, respectively. animals of group 4 received ast at a dose of 25 mg/kg; group 5 and 6 received ast along with naringin at 25 & 15 and 25 & 30 mg/kg, respectively. group 7 received ast 50 mg/kg; animals of group 8 and 9 received ast along with naringin at 50 & 15 and 50 & 30 mg/kg, respectively. group 10 animals served as normal control. ast was administered 30 min after oral dose of naringin. blood samples (~200 µl) have been collected through retro-orbital puncture under light ether anesthesia just prior to and at 1 st , 2 nd and 4 th h after ast administration in vials containing disodium edta as an anticoagulant. plasma was separated and divided into two aliquots. one aliquot was immediately used to estimate total cholesterol and triglycerides levels using commercially available kits at 510 nm. the other aliquot was stored at -20 ± 5 °c until analysis for the quantification of ast. the percentage variation of total cholesterol and triglycerides were calculated for each group using the following formula: % variation in total cholesterol and triglycerides = {ci-ct} ×100/ ci where, ci is the concentration of disease control and ct is the concentration of ast at 1 st , 2 nd and 4 th hour. quantitation of ast in rat plasma samples by rp-hplc quantitation of ast in rat plasma samples were carried out by rp-hplc using earlier reported method with minor modifications [26]. hplc (shimadzu, japan) is equipped with lc-20 at vp system controller, lc20 at pump, spd-20a uv detector and a phenomenex-c18 column (250 × 2.6 mm, 5 µm). the data were acquired and processed using lc solutions (version 3.1) software. to an aliquot of 100 µl of rat plasma, 200 µl of methanol was added. the mixture was vortexed for 20 min followed by centrifugation at 3000 rpm for 10 min at 4 °c. the organic layer was separated and filtered through a 0.2 µm cellulose acetate filter. the chromatographic resolution of ast was achieved by using the admet & dmpk 7(3) (2019) 174-182 effect of naringin on atorvastatin doi: 10.5599/admet.647 177 isocratic mobile phase consists of acetonitrile : 0.05m sodium phosphate buffer (ph 4; 65:35, v/v) delivered at a flow rate of 1.0 ml/min and the eluent was monitored by uv detector set at 236 nm. an aliquot of 20 µl of organic layer was injected onto column and corresponding peak areas were noted. the calibration curve (y= 86.89x + 3832) was found to be linear from 100 to 2000 ng/ml (r 2 =0.9987). statistical analysis all the values were expressed as a mean ± sem. results were analyzed statistically using one-way analysis of variance (anova) following dunnett test. values of p<0.05 were considered statistically significant. results effect of naringin on ast in tyloxapol induced hyperlipidemia effect on total cholesterol administration of tyloxapol has significantly increased the total cholesterol after 30 h. it was observed that cholesterol levels were raised by 2.5 folds. the administration of ast alone at two test dose levels caused a statistical significant (p<0.001) dose dependent decrease in total cholesterol levels with maximum reduction observed at 2 nd h after drug administration (table 1). among the two test dose levels, ast with 50 mg/kg has produced maximum protection, with % reduction of total cholesterol was 19.09, whereas a percentage reduction of total cholesterol in animals received 25 mg/kg was 12.70 (table 1). the effect of naringin alone on hyperlipidemia was not significant. the percent reduction in total cholesterol was 1.89 and 4.93 with 15 and 30 mg/kg, respectively at 2 h post administration of naringin (table 1). the coadministration of naringin (15 and 30 mg/kg) with ast at two test dose levels (25 and 50 mg/kg) has produced significant reduction in total cholesterol levels and is quantitatively high at 2nd and 4th h after treatment when compared to ast alone treated animals (table 1). however, the maximum percent reduction was noted at 2 nd h. the reduction in total cholesterol with ast (50 mg/kg) when co-administered with naringin (15 and 30 mg/kg) was 33.00 and 41.46 %, respectively where the total cholesterol is reduced by 19.09 % in animals received 50 mg/kg ast alone. similarly, the reduction in total cholesterol was 24.02 and 31.30 % in animals received the ast (25 mg/kg) along with naringin (15 and 30 mg/kg). the reduction in total cholesterol is 12.70 % when treated with ast (25 mg/kg) alone (table 1). effect on triglycerides administration of tyloxapol has increased the triglycerides levels to 9 to 9.5-folds from normal levels. the increased triglycerides levels were in the range of 650 to 695 mg/dl, when compared to initial levels in the range of 60-80 mg/dl. there is no significant fall in triglyceride levels in control rats during the course of experiment. the effect of naringin alone on triglyceride levels was not significant and the fall in the triglyceride levels in naringin treated rats is 2.04 and 4.08 % with 15 and 30 mg/kg doses, respectively, when compared to control rats (table 2). the administration of ast alone has produced a significant dose dependent fall in triglyceride levels. the maximum reduction was observed at 2 nd h of the experiment with percentage reduction of 13.40 and 17.15 at dose levels of 25 and 50 mg/kg, respectively (table 2). the coadministration of naringin (15 and 30 mg/kg) along with ast at two test dose levels statistically significant increase in the fall in triglyceride levels when compared to ast alone treated animals. the percent reduction in animals treated with ast along with naringin (25 + 15; 25 + 30 mg/kg) was 20.03 and 26.77, respectively. whereas the percent reduction was 28.10 and 38.98 with test dose of ast and naringin at 50 r. mullangi et al. admet & dmpk 7(3) (2019) 174-182 178 + 15 and 50 + 30 mg/kg, respectively after 2 h post-treatment. the results pertaining to effect on triglyceride levels were shown in table 2. table 1. effect of naringin on atorvastatin (ast) on total cholesterol levels induced by tyloxapol in male wistar rats values are expressed as mean ± sem, **p<0.01, ***p<0.001; n=6 figures in parenthesis indicate the percentage decrease in total cholesterol levels quantitation of ast in rat plasma samples by rp-hplc the plasma concentration of ast was determined in hyperlipidemic rats at 1, 2 and 4th h after oral administration of ast alone and ast along with naringin. hplc chromatograms of (a) rat blank plasma (b) rat plasma spiked with ast (300 ng/ml) and (c) 1 h in vivo rat plasma sample obtained after oral administration of ast at 50 mg/kg along with naringin 15 mg/kg (group 9) were shown in figure 1. the maximum plasma concentration of ast was observed at 2nd h after oral administration of ast and it was found to be 43.33 and 198.92 ng/ml at dose levels of 25 and 50 mg/kg, respectively in ast alone treated animals. table 3 the plasma concentration of ast. the concentration of ast is increased proportionally with increase in naringin dose. the plasma concentration of ast has increased with the co-administration of naringin at dose levels of 25 + 15, 25 + 30, 50 + 15 and 50 + 30 mg/kg was found to be 288.5, 663.81, 793.03 and 1233.72 ng/ml, respectively. group treatment dose (mg/kg) plasma concentration of total cholesterol (mg/dl) 0 min 1 h 2 h 4 h i disease control 132.2 ± 3.07 127.8 ± 2.28 137.8 ± 2.27 124.0 ± 2.44 ii naringin 15 132.2 ± 1.99 126.5 ± 2.94 (1.02) 135.2 ± 2.35 (1.89) 122.2 ± 2.30 (1.45) iii naringin 30 129.7 ± 2.26 125.3 ± 2.10 (1.96) 131.0 ± 2.78 (4.93) 119.2 ± 2.02 (3.87) iv ast 25 135.3 ± 3.08 118.8 ± 2.72 (7.04) 120.3 ± 2.92*** (12.70) 110.0 ± 2.01*** (11.29) v ast + naringin 25 + 15 134.2 ± 2.34 111.3 ± 2.51*** (12.91) 104.7 ± 2.07*** (24.02) 98.83 ± 2.70*** (20.30) vi ast + naringin 25 + 30 132.0 ± 3.07 108.8 ± 3.19*** (14.87) 94.67 ± 2.60*** (31.30) 94.17 ± 2.71*** (24.06) vii ast 50 132.8 ± 2.33 115.7 ± 2.96* (9.47) 111.5 ± 2.95*** (19.09) 106.2 ± 2.30*** (14.35) viii ast + naringin 50 + 15 134.7 ± 2.98 110.7 ± 2.81*** (13.38) 92.33 ± 2.02*** (33.00) 95.00 ± 2.60*** (23.39) ix ast + naringin 50 + 30 135.8 ± 2.41 106.2 ± 3.04*** (16.90) 80.67 ± 1.94*** (41.46) 88.17 ± 2.41*** (28.90) x normal 54.0 ± 1.91 54.30 ± 1.82 54.50 ± 1.45 53.50 ± 1.94 admet & dmpk 7(3) (2019) 174-182 effect of naringin on atorvastatin doi: 10.5599/admet.647 179 table 2. effect of naringin on atorvastatin (ast) on triglyceride levels induced by tyloxapol in male wistar rats values are expressed as mean ± sem, **p<0.01, ***p<0.001; n=6 figures in parenthesis indicate the percentage decrease in total triglycreide levels table 3. effect of naringin on pharmacokinetic profile of atorvastatin (ast) in tyloxapol induced hyperlipidemia male wistar rats group treatment dose (mg/kg) plasma concentration of total triglycerides (mg/dl) 0 min 1 h 2 h 4 h i disease control 652.5 ± 2.74 670.8 ± 1.83 694.0 ± 3.07 662.2 ± 2.63 ii naringin 15 666.8 ± 2.30 661.3 ± 2.27 (1.42) 674.2 ± 2.27*** (2.85) 648.7 ± 2.45** (2.04) iii naringin 30 663.8 ± 1.90 650.5 ± 2.47*** (3.03) 659.8 ± 2.35*** (4.93) 635.2 ± 2.02*** (4.08) iv ast 25 664.0 ± 2.87 640.2 ± 2.25*** (4.56) 601.0 ± 2.49*** (13.40) 591.2 ± 2.27*** (10.72) v ast + naringin 25 + 15 660.3 ± 2.36 615.0 ± 2.78*** (8.32) 555.0 ± 1.93*** (20.03) 567.7 ± 2.99*** (14.27) vi ast + naringin 25 + 30 668.0 ± 2.54 594.2 ± 2.89*** (11.4) 508.2 ± 2.75*** (26.77) 522.0 ± 2.12*** (21.17) vii ast 50 660.5 ± 1.97 630.5 ± 2.71*** (6.01) 575.0 ± 1.75*** (17.15) 575.7 ± 2.10*** (13.06) viii ast + naringin 50 + 15 670.2 ± 2.86 589.8 ± 3.16*** (12.08) 499.0 ± 2.06*** (28.10) 522.3 ± 1.76*** (21.13) ix ast + naringin 50 + 30 670.3 ± 2.43 563.8 ± 2.64*** (15.95) 423.5 ± 2.32*** (38.98) 482.3 ± 2.43*** (27.17) x normal 71.50 ± 2.46 73.50 ± 1.74 74.00 ± 1.62 73.17 ± 2.02 group treatment dose (mg/kg) concentration (ng/ml) 1 st hour 2 nd hour 4 th hour iv ast 25 20.53 43.33 39.48 v ast + naringin 25 + 15 144.7 288.5 226.3 vi ast + naringin 25 + 30 484.8 663.8 596.4 vii ast 50 82.86 198.9 159.2 viii ast + naringin 50 + 15 301.9 793.0 650.7 ix ast + naringin 50 + 30 508.9 1233.7 951.5 r. mullangi et al. admet & dmpk 7(3) (2019) 174-182 180 figure 1. hplc chromatograms of (a) rat blank plasma (b) rat plasma spiked with ast (300 ng/ml) and (c) 1 h in vivo rat plasma sample obtained after oral administration of ast at 50 mg/kg along with naringin 15 mg/kg (group 9). discussion a bioenhancer is an agent capable of enhancing the bioavailability/bioefficacy of a particular drug with which it is co-administered. naringin is a flavanone glycoside obtained from citrus fruits. naringin possess diverse pharmacological properties. previously, it was reported that supplementation with naringin for 3 weeks did not exhibit a hypolipidemic effect, however it reported to show beneficial effects of lowering hepatic cholesterol biosynthesis and levels of plasma lipids when supplemented for 6 weeks in a high fructose and high cholesterol fed rat model [27]. naringin is known to enhancement of bioavailability of various structurally and therapeutically diverse drugs (verapamil, diltiazem, paclitaxel, tamoxifen) in rats by inhibition of cyp mediated metabolism and/or p-gp mediated permeability [19-22]. ast is an antihyperlipidemic agent belongs to statin family. it acts by blocking the enzyme responsible for cholesterol biosynthesis known as hmg co-a reductase. ast is a potent dyslipidemic agent having significant use in lowering the blood circulation lipids such as low density lipoprotein (ldl), totalcholesterol (tc), triglycerides (tg) and apolipoproteins. ast is metabolized by cyp3a4. in the present study, the effect of naringin on antihyperlipidemic properties of ast was investigated with coadministration of naringin in tyloxapol induced hyperlipidemia in wistar rats. the observed activity was correlated with plasma concentration of ast. conclusions the findings of the present study suggest that increased in plasma concentration of ast may be due to inhibition of cyps and p-gp by naringin, which may be responsible for higher percent reduction in both cholesterol and triglycerides levels by ast. the treatment of hyperlipidemic patients with ast along with naringin rich diet may potentiate the therapeutic efficacy of ast and could be exploited to achieve better therapeutic control and patient compliance. admet & dmpk 7(3) (2019) 174-182 effect of naringin on atorvastatin doi: 10.5599/admet.647 181 acknowledgments the authors wish to thank the aicte for providing financial support and college management for encouragement and providing facilities. conflict of interest: ramesh mullangi is vice president of jubilantbiosys. references [1] n. atal, k.l. bedi. bioenhancers: revolutionary concept to market. journal of ayurveda and integrative medicine 1 (2010) 96-99. [2] k.g. bose. 1929. pharmacographia indica, bose laboratories, calcutta, india. [3] r. singh, s. devi, j.h. patel. indian herbal bioenhancers: a review. pharmacognosy reviews 3 (2009) 90-92. [4] r.f. albach, g.h. redman. composition and inheritance of flavanones in citrus fruit. phytochemistry 8 (1969) 127-143. [5] g. scambia, f.o. ranelletti, p.b. panici, d.r. vincenzo, g. bonanno, g. frrandina, m. paiantelle, s. bussa, c. rumi, m. ciantriglia. quercetin potentiates the effect of adriamycin in a multidurgresistant mcf-7 human breast-cancer cell line: p-glycoprotein as a possible target. cancer chemotherapy and pharmacology 36 (1995) 448-450. [6] d.g. bailey, j.m. arnold, j.d. munoz spence. grapefruit juice-felodipine interaction: mechanism, predictability, and effect of naringin. clinical pharmacology & therapeutics 53 (1993) 637-642. [7] h. peter, t. pavel, s. marie. flavoniods-potent and versatile biologically active compounds interacting with cytochrome p450. chemico-biological interactions 139 (2002) 1-12. [8] h. doostdar, m.d. burke, r.t. mayer. bioflavonoids: selective substrates and inhibitors for cytochrome p450 (cyp1a and cyp1b1). toxicology 144 (2000) 31-38. [9] f. fukuda, t. ohta, y. yamazoe. grapefruit component interacting with rat and human p450 cyp3a: possible involvement of non-flavonoid components in drug interaction. biological and pharmaceutical bulletin 20 (1997) 560-564. [10] a.a. gammal, r.m. mansour. antimicrobial activities of some flavonoid compound. zentralbl mikrobiol[ta] 141 (1986) 561-565. [11] v.s. aboobaker, a.d. balgi, r.k. bhattacharya. in vivo effect of dietary factors on the molecular action of aflatoxin b1: role of non-nutrient phenolic compounds on the catalytic activity of liver fraction. in vivo 8 (1994) 1095-1098. [12] p.b. gordon, i. holen, p.o. seglen. protection by naringin and some other flavonoids of hepatocytic autophagy and endocytosis against inhibition by okadaic acid. journal of biological chemistry 270 (1995) 5830-5838. [13] j.g. liu, y.m. xie, w.l. deng. experimental study on anti-inflammation effect of assemble flavonoids of rhizomadrynariae. chinese journal of natural medicines 2 (2004) 232-234. [14] g. ekambaram, p. rajendran, v. magesh, d. sakthisekaran. naringenin reduces tumor size and weight lost in n-methyl-n′-nitro-n-nitrosoguanidine–induced gastric carcinogenesis in rats. nutrition research 28 (2008) 106-112. [15] y.c. hsiao, w.h. kuo, p.n. chen, h.r. chang, t.h lin, w.e. yang. flavanone and 2′-oh flavanone inhibit metastasis of lung cancer cells via down-regulation of proteinases activities and mapk pathway. chemico biological interactions 167 (2007) 193-206. [16] d. ozyurt, b. demirata, r. apak. determination of total antioxidant capacity by a new spectrophotometric method based on ce (iv) reducing capacity measurement. talanta 71 (2007) 1155-1165. r. mullangi et al. admet & dmpk 7(3) (2019) 174-182 182 [17] a.k. ratty, n.f. das. effects of flavonoids on nonenzymatic lipid peroxidation: structure–activity relationship. biochemical medicine and metabolic biology 39 (1988) 69-79. [18] s.h. bok, t.s. jeong, k.h. bae, y.b. park, m.s. choi, s.s. moon. naringin and naringenin as inhibitors of acyl coa-cholesterol-o-acyltransferase. united states patent, patent number us006165984a (2000). [19] h.j. kim, j.s. choi. effects of naringin on the pharmacokinetics of verapamil and one of its metabolites, norverapamil in rabbit. biopharmaceutics and drug disposition 26 (2005) 295-300. [20] j.s. choi, k.h. han. enhanced oral exposure of diltiazem by the concomitant use of naringin in rats. international journal of pharmaceutics 305 (2005) 122-128. [21] j.s. choi, s.c. shin. enhanced paclitaxel bioavailability after oral co-administration of paclitaxel prodrug with naringin to rats. international journal of pharmaceutics 292 (2005) 149-56. [22] j.s. choi, k.w. kang. enhanced tamoxifen bioavailability after oral administration of tamoxifen in rats pretreated with naringin. archives of pharmcal research 31 (2008) 1631-1636. [23] lipitor® (atorvastatin). data sheet. pfizer new zealand ltd, new zealand. 2012, p. 1-24. [24] product monograph, gd-atorvastatin (atorvastatin calcium tablets). lipid metabolism regulator, gen med, a division of pfizer canada inc. 2012, p. 1-54. [25] h. g. vogel. drug discovery and evaluation. pharmacological assays. ed(2). springer, 2002, p. 11061107. [26] a. mohammadi, n. rezanour, m. ansari dogaheh, f. ghorbani bidkorbeh, m. hashem, r. walker. a stability–indicating high performance liquid chromatographic (hplc) assay for the simultaneous determination of atorvastatin and amlodipine in commercial tablets. journal of chromatography b 846 (2007) 215-221. [27] s.y. kim, h.j. kim, m.k. lee, s.m. jeon, g.m. do, e.y. kwon, y.y. cho, d.j. kim, k.s. jeong, y.b. park, t.y. ha, m.s. choi. naringin time-dependently lowers hepatic cholesterol biosynthesis and plasma cholesterol in rats fed high-fat and high-cholesterol diet. journal of medicinal food 9 (2006) 582586. [28] t. ishikawa, n. fidge. changes in the concentration of plasma lipoproteins and apo proteins following the administration of triton wr 1339 to rats. journal of lipid research 20 (1979) 254264. [29] m. kuroda, k. tanzawa, y. tsujita, a. endo. mechanism for elevation of hepatic cholesterol synthesis and serum cholesterol levels in triton wr-1339-induced hyperlipidemia. biochimica et biophysica acta 489 (1977) 119-125. [30] a. scanu, p. oriente. triton hyperlipidemia in dogs: in vitro effects of the detergent on serum lipoproteins and chylomicrons. journal of experimental medicine 113 (1961) 735-757. [31] p.e. schurr, j.r. schultz, t.m. parkinson. triton induced hyperlipidaemia in rats as an animal model for screening hypolipideamic drugs. lipids 7 (1972) 68-74. ©2019 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ block relevance (br) analysis and polarity descriptors in property-based drug design doi: 10.5599/admet.532 215 admet & dmpk 6(3) (2018) 215-224; doi: http://dx.doi.org/10.5599/admet.532 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index review block relevance (br) analysis and polarity descriptors in property-based drug design giuseppe ermondi and giulia caron* molecular biotechnology and health sciences dept., università degli studi di torino, via quarello 15, 10135 torino, italy *corresponding author: e-mail: giulia.caron@unito.it; tel.: +39-011-6708337; received: april 06, 2018; revised: may 23, 2018; published: may 29, 2018 abstract block relevance (br) analysis is a tool to interpret qspr/pls models which can provide the information content of any physicochemical determinant used in property-based drug discovery; its application for the characterization of experimental polarity descriptors is discussed. keywords adme; hydrogen bonding; iam; qspr; pls; psa; vip property-based drug design: the need for the information content of physicochemical descriptors the pool of physicochemical properties including size, lipophilicity, solubility, permeability, polarity, and hydrogen bond (hb) capacity defines the drug-like profile of potential candidates. thus, during the lead optimization process, medicinal chemists control the absorption, distribution, metabolism and excretion (adme) fate of a candidate by monitoring the variation in the physicochemical profile following chemical modifications [1]. any molecular property can be quantified by a few physicochemical descriptors. for instance, lipophilicity can be described by the log of the distribution coefficient in the octanol/water system (i.e. log doct) but also by the log of the distribution coefficient in the toluene/water system (i.e. log dtol). log doct and log dtol encode diverse information [2] and thus are not interchangeable. therefore, it could be relevant for a medicinal chemist to know how a structural change impacts both log doct and log dtol. more generally, to select the most relevant set of descriptors for any drug discovery program, it is crucial to know the information content of any descriptor. abraham’s solvation equations (their review is beyond the aim of this paper) are the most known quantitative structure-property relationships (qsprs) tools that provide the information content of any molecular property [3,4]. briefly, by measuring a linear free-energy related solute property for a set of at least 25 compounds with known 2d molecular descriptors (e, s, a, b and v, optimised for small monoor bifunctional organic molecules in unionized form) the solvation equation can be built using a multiple linear regression (mlr). to provide an alternative tool to abraham’s equations, we have designed and implemented a http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:giulia.caron@unito.it ermondi and caron admet & dmpk 6(3) (2018) 215-224 216 chemoinformatic strategy named block relevance (br) analysis [5] which affords an interpretation “at a glance” of qsprs models based on a selected pool of descriptors and a partial least square (pls) algorithm. to fully understand br analysis and thus its relevance in property-based drug design, we have briefly reviewed [6,7] some basic concepts of qspr. qspr for modelling physicochemical properties quantitative structure-property relationships can be formulated using the eq. 1 [8]: 𝑃𝑟𝑜𝑝𝑒𝑟𝑡𝑦 = 𝑓(𝐷𝑒𝑠𝑐𝑟𝑖𝑝𝑡𝑜𝑟𝑠) (1) where a property is in principle any physicochemical property, the descriptors are numerical values that are related to the structure of the chemical compounds and, f is a tailored function able to quantitatively correlate the property and the descriptors. in a context of properties-based drug design, the endpoint is modelling the pool of physicochemical properties which define drug-like profile of the candidate compounds, e.g. lipophilicity, solubility, permeability, polarity and hydrogen bond (hb) capacity. in this paper we will use property to indicate what we want to model i.e. experimental descriptors like log poct which quantifies the molecular property of lipophilicity. although a plethora of descriptors has been reported in the literature [9] there is still no agreement on what constitutes the “best” pool for molecular design [10] since it depends on the property that is under study. the subset of volsurf+ (vs+) descriptors (v. 1.1.2, www.moldiscovery.com) based on 3d molecular fields (mifs, see below) are a reasonable choice for modelling physicochemical properties. shortly, a 3d mif is calculated using the grid force field [11,12] which may be viewed as a 3d matrix, with attractive and repulsive energy values between a chemical probe and a target molecule (e.g. a drug candidate). by using four probes and various cut-off values of energy, the capacity of the drug to form intermolecular interactions can be quantified and converted into molecular descriptors. to model physicochemical properties five blocks of vs+ descriptors corresponding to five types of intermolecular interactions are essential: 1) the dry block: the dry probe is used to calculate descriptors related to the hydrophobic interaction taking into account both the entropic and enthalpic effects [13]; 2) the oh2 block: the oh2 probe mimics water interaction with the target surface and thus is used to calculate descriptors related to hydrophilic properties of the target; 3) the nh block: the hydrogen bond acceptor (hba) properties of the target are evaluated using a hydrogen bond donor (nh) probe 4) the o block: the hydrogen bond donor (hbd) properties of the target are evaluated using an hydrogen bond acceptor (o) probe, 5) the others block: additional molecular descriptors are obtained using a combination of information derived from different probes to represent the unbalance of hydrophilic and hydrophobic regions on the surface target. finally, in line with the evidence that molecular dimensions are crucial to characterize drug-like candidates [14], a sixth block of size and shape descriptors (the size block) cannot be neglected. overall the six blocks of vs+ descriptors listed above define the capacity of the drug to form intermolecular interactions and thus can be reasonably used in qspr for modelling physicochemical properties. in some cases, the properties are related to chemical descriptors through physicochemical theories, (e.g. absorbance vs concentration, lambert-beer equation); more often, properties and descriptors relations can be found through a statistical treatment. when a statistical approach is required, a mechanistic interpretation of the validated model is not always straightforward since the presence of a correlation between two variables does not imply causation. therefore, to obtain a mechanistic interpretation of the model (this is an essential step for property-based drug discovery purposes), it is necessary to weight the contribution of each descriptor to the final model. many algorithms could be used to correlate properties http://www.moldiscovery.com/ admet & dmpk 6(3) (2018) 215-224 block relevance analysis in property-based drug design doi: 10.5599/admet.532 217 with descriptors, showing different capacities to accurately describe the contribution of the descriptors to the model [15]. multilinear regression (mlr) is the simplest tool to obtain and to interpret models but its application is appropriate only with a few models and not intercorrelated descriptors are needed. the partial least squares (pls) method data analysis is also widely used in medicinal chemistry. pls manages a large number of potential intercorrelated descriptors by projecting the data into a low dimensional and not intercorrelated “latent variable” space. components in pls are constructed to maximize the covariance between the dependent variable y and the original independent variables x [8]. to extract the contribution of any descriptor from pls models, the analysis of vips and coefficients plots is needed (see below). recently, nonlinear approaches, such as artificial neural networks (ann) or support vector machines (svm), have attracted the attention of many practitioners. these methods can “catch” hidden nonlinearities between properties and descriptors providing better predictors than the linear models. however, these nonlinear methods are not as easy to interpret as the mlr or pls models [8], since they are by definition “black boxes” preventing any interpretation of underlying mechanisms. block relevance (br) analysis as a tool to interpret qspr/pls models as mentioned above pls is a widely used algorithm to generate qspr models which, however, are often not easy to interpret. it is described below how br analysis can overcome this limit. theory variable importance in projection (vip) plots are often used to find the relative contribution of each descriptor to the final pls model. the vip for the descriptor j, vipj is defined according to wold et al.[16]: 𝑉𝐼𝑃𝑗 = { ∑ 𝑤𝑗𝑓 2 𝐹 𝑓=1 ⋅𝑆𝑆𝑌𝑓⋅𝐽 𝑆𝑆𝑌𝑡𝑜𝑡𝑎𝑙∙𝐹 } 1 2 (2) where wjf is the weight value for variable j component f, ssyf is the sum of squares of explained variance for the f th component and j the number of variables. ssytotal is the total sum of squares explained by the dependent variable, and f is the total number of components. the weights in a pls model reflect the covariance between the independent and dependent variables and the inclusion of the weights is what allows vip to reflect not only how well the dependent variable is described but also how important that information is for the model of the independent variables. note that the sum of squared vip values is equal to the number of descriptors (j): ∑ 𝑉𝐼𝑃ⅈ=1 2𝐹 𝑗=1 = 𝐽 (3) to make easier pls interpretations we assume that descriptors, if conveniently chosen, can be grouped in blocks. the block relevance (br) of each block of descriptors was defined as the ratio of the sum of the squared vip values of a given block of descriptors to the number of those descriptors. 𝐵𝑅𝑖 = ∑ 𝑉𝐼𝑃𝑗 2 𝑁𝑖 𝑗=1 𝑁𝑖 (4) where i is the number of blocks (6), n is the number of descriptors for any block, vipj is the value of vip for each predictor fitting the pls model. br shows the relevance of a certain block of descriptors in the model; the higher the value of br, the more important is that block. ermondi and caron admet & dmpk 6(3) (2018) 215-224 218 depending on the sign of the pls coefficient, br was broken down into br (+) and br (−), eq. (5). brⅈ = brⅈ(+) + brⅈ(−) (5) the current br analysis version according to its definition, br analysis requires a supervised grouping of ad hoc selected descriptors. in the current version we implement 82 of the original 128 vs+ descriptors since, as discussed above, they define the capacity of the drug to participate in intermolecular interactions. the 82 descriptors are grouped in six blocks (fig. 1a) and each block contains information about the propensity of the compound to be involved in a specific intermolecular interaction, e.g. the oh2 block reveals the propensity of the molecule to interact with water. br analysis workflow is shown in fig. 1b. in the first step, the dataset is set-up. this consists in collecting physicochemical data (e.g. log doct) and obtaining the corresponding compounds structures with standard procedures. secondly, the dataset is submitted to vs+ to calculate the 82 descriptors and to obtain and validate the pls models (internal and external validations are always performed as reported in the literature [17]). finally, vips and coefficients values are exported to excel, elaborated and submitted to br analysis which produces an easy-to-interpret graphical output. figure 1. (a) selected volsurf+ descriptors organized in the six blocks as implemented in the current version of br analysis (see text for more details) (b) br analysis workflow. our recent modelling study of immobilized artificial membrane (iam) chromatography [2] will be used to illustrate details of the br approach. in this case, br analysis was applied to extract the relative contribution of the intermolecular forces governing the iam.pc.dd2 log kw iam parameter from pls models. the model was built using the whole dataset of compounds used in [2]: (n°(training) = 189, n°(test) = 64, r 2 = 0.82, q 2 = 0.71). in fig. 2, the vips are shown (fig. 2a) and the coefficients plots (fig. 2b) which should be combined to interpret the pls model. in fact, the vips plot reveals the relevance of a descriptor in the model, whereas the coefficients plot establishes if the increase of variables corresponds to an increase dataset set-up data selection structure input descriptors calculation pls creation and validation of the model vips and coefficients calculation br analysis grouping of descriptors in blocks elaboration of vips and coefficients -3.0 -2.0 -1.0 0.0 1.0 2.0 3.0 size oh2 dry o n1 others graphical output • v,s,r,g size (4) • w1-w8, iw1-iw4, cw1-cw8; psa, psar oh2 (22) • d1-d8, dd1-dd8, id1-id4, cd1-cd8, hsa, phsar dry (30) • wo1-wo6 o (6) • wn1-wn6 n1 (6) • hl1, hl2, a, cp, drdrdr, drdrac, drdrdo, dracac, dracdo, drdodo, acacac, acacdo, acdodo, dododo others (14) a) b) admet & dmpk 6(3) (2018) 215-224 block relevance analysis in property-based drug design doi: 10.5599/admet.532 219 (positive weights) or a decrease (negative weights) of the property. using fig. 2a and fig. 2b to interpret the pls model is far from being user-friendly. figure 2. log kw iam model [2]: (a) the pls vip plot (standard output), (b) the pls coefficients plot (standard output), (c) the pls vip plot (output coloured by blocks), the lines correspond to three reasonable threshold values (1.21, 1.00, 0.83), (d) the pls coefficients plot (output coloured by blocks) (e) br graphical output: the size block (green) is related to the influence of size and shape; the water block (cyan) reflects the role of the hydrophilicity whereas the dry block (yellow) the role of hydrophobicity; red and blue blocks are related to the solutes’ hbd and hba properties, respectively; and, finally, the others block (grey) is related to the unbalance of hydrophilic and hydrophobic regions. the two plots, fig. 2c (vips) and 2d (coefficients) have been represented by using the colour codes of br blocks. fig. 2c and 2d are clearer than fig. 2a and 2b but the identification of the most relevant vips remains a crucial step for pls interpretation since several threshold criteria have been reported in the literature [18]. a low threshold could result in the selection of too many descriptors and in a problematic interpretation, whereas a high threshold could result in the loss of important descriptors and in an inaccurate interpretation. for instance, in fig. 2c three reasonable threshold values (1.21, 1.00 and 0.83) are reported which allows the identification of 18, 36 and 43 descriptors, respectively. if the 18 descriptors obtained with the largest threshold value are considered, it can be verified that most (but not all) of them belong to the size and dry blocks. size descriptors coefficients are all positive (fig. 2d), thus the larger the compound, the larger log kw iam . the behavior of hydrophobic descriptors is less clear since strong hydrophobic interactions (d5-d8) increase log kw iam values, whereas weak interactions (d1-d4) are not favourable (fig. 2d). overall, the vips and coefficients plot together do not allow a clear and unique pls model interpretation. br analysis graphical output for the same model is shown in fig. 2e. for a correct interpretation of the br analysis graphical results the following rules are applied: a) blocks with positive weighting (e.g., the green block) show how much the property increases log kw iam , whereas those with negative weighting indicate how much the property decreases log kw iam ; b) according to our experience, block significance was ermondi and caron admet & dmpk 6(3) (2018) 215-224 220 set at 0.2 and thus the blue and red blocks in fig. 2e are not significant and c) a block with comparable positive and negative contributions (e.g. oh2 and others in fig. 2e), indicates that it is poorly relevant in the description of the investigated property. overall, br analysis shows that log kw iam is essentially related to the dimensions of the molecules. table 1. molecular properties analysed with the br analysis descriptor type descriptor reference biomimetic chromatographic index log kw iam [2] biomimetic chromatographic index log k mucin [19] biomimetic combined chromatographic index δlog kw iam [2] biophysical method log kd (spr) [20] chromatographic index log k’ c18 [5] chromatographic index log k’ ph [5] chromatographic index log k’ cn-rp [5] chromatographic index log k’ oh-rp [5] chromatographic index log k’ cn-rp [5] chromatographic index log k’ oh-rp [5] chromatographic index log k’ nh2-np [5] chromatographic index log k’ hilic [5] chromatographic index log k’ epsa [21] chromatographic index log k’80 plrp-s [22] chromatographic index elogd [22] combined partition coefficient δlog poct-tol [6] computed descriptor tpsa [21] computed distribution coefficient calc log doct [23] computed partition coefficient calc log ptol [22] distribution coefficient log dlip [20] distribution coefficient log d 7.4 oct [2] partition coefficient log poct [5] partition coefficient log poct [7] partition coefficient elogp [7] partition coefficient log poct [6] partition coefficient log ptol [6] partition coefficient log poct [2] permeability rate log papp(rrck) [21] permeability rate log papp(caco-2) [21] permeability rate log papp (6.5/7.4) [23] permeability rate log papp (7.4/7.4) [23] as expected, the vips/coefficients plots and the br analysis provide a similar picture, but the br analysis immediately focuses on the main features that drive the model. it is noteworthy to underline that the two approaches can be combined; using the br analysis to obtain a general idea of the main interaction governing the property under study and then the vips/coefficients plots for a more detailed analysis of the results. admet & dmpk 6(3) (2018) 215-224 block relevance analysis in property-based drug design doi: 10.5599/admet.532 221 up to now br analysis has been applied to various properties related to the drug discovery process (table 1) to deconvolute the different balance of intermolecular forces governing the properties under investigation. overall, we verified the interchangeability of descriptors obtained with different methods and claimed, or not to provide the same information. this was done for lipophilicity, permeability and polarity descriptors. the next section describes application of br analysis to the identification of experimental polarity descriptors. br analysis and polarity descriptors polarity is a molecular property of remarkable interest in the characterization of the propensity of compounds to form an intramolecular hydrogen bond (imhb) [24] and thus in the prediction of permeability [25]. it is often estimated using the polar surface area (psa) which can be calculated either using the 2d structure, topological polar surface area (tpsa), or the 3d structure, the molecular polar surface area (mpsa). both descriptors have severe limitations since a) hbd and hba strength and hb directionality are not considered b) hba and hbd contributions are not distinguished and c) some atoms with partial charge different from zero are not considered to be polar [26]. moreover, as discussed in a previous paper, increasing the size of the molecules, the psa takes into account steric effects that tpsa does not reproduce [19]. these limitations suggest that measuring rather than predicting polarity could improve property-based drug discovery. to verify whether an experimental physicochemical descriptor is a clean polarity determinant, we can apply br analysis and check if the considered descriptor is mainly driven by the three blocks related to polarity (light blue, red and blue blocks, see fig. 3a for a concise colour code scheme). below we describe how br analysis provides insight into three descriptors that in principle can be considered polarity descriptors. δlog poct-tol is the difference between log poct (the logarithm of the partition coefficient p in the octanol/water system) and log ptol (the logarithm of the partition coefficient p in the toluene/water system). br analysis (fig. 3b) indicates that δlog poct-tol is a clean descriptor of exposed hbd properties [6]. notably, the solutes’ hba properties do not significantly contribute to δlog poct-tol which therefore cannot be strictly considered a polarity descriptor. epsa is an exposed polarity measurement which assesses polarity by retention time using controlled supercritical fluid chromatography (sfc) conditions [24,27,28]. br analysis was applied to understand the balance of intermolecular forces governing epsa (fig. 3c). results showed that epsa is governed by the solutes’ hbd (red, positive) and hba (blue, negative). because hbd and hba blocks have the opposite sign, br analysis gives evidence that the presence of hba groups could be considered as an interference and therefore the hbd properties of the molecule can be underestimated by epsa when the structure includes more hba groups. [21] these findings suggest that also epsa cannot be strictly considered a polarity descriptor. according to grumetto et al. [29] log kw iam is the difference between the logarithm of the experimental chromatographic retention factor (log kw iam ) and the value here named clog kw iam calculated using eq. 7. 𝑐𝑙𝑜𝑔 𝐾𝑤 𝐼𝐴𝑀 = 0.92 ∗ log 𝑃𝑜𝑐𝑡 𝑁 − 1.03 (7) br analysis showed that δlog kw iam is a clean polarity descriptor since it is due to both hbd and hba properties (fig. 3d). [2] ermondi and caron admet & dmpk 6(3) (2018) 215-224 222 figure 3. (a) blocks color codes and significance to help br graphical output interpretation. br analysis graphical output for (b) δlog poct-tol.[6] (c) epsa [19], (d) δlog kw iam [2] since the determination of experimental polarity is crucial for large and flexible structures, δlog poct-tol, epsa and log kw iam data was collected for indinavir and nelfinavir (structures and data in fig. 4). in a previous study it was shown that nelfinavir has a larger propensity to form imhbs than indinavir [30] and thus it was expected that δlog poct-tol, epsa and log kw iam of nelfinavir are smaller than the correspondent values determined for indinavir. this was verified for δlog poct-tol and log kw iam but not for epsa, which fails in ranking the polarity of the two protease inhibitors; probably because of the high and different number of hba groups present in the structures. figure 4. indinavir and nelfinavir: comparison of their physicochemical profile (in bold the polarity descriptors discussed in the text) admet & dmpk 6(3) (2018) 215-224 block relevance analysis in property-based drug design doi: 10.5599/admet.532 223 conclusions the design of drug candidates with the right balance of physicochemical properties (property-based drug design) calls for a tool like the block relevance (br) analysis which can interpret qspr/pls models and provide an easy-to-read map of the information content of any experimental molecular descriptor. up to now br analysis has been applied to about 30 molecular descriptors mostly related to lipophilicity, polarity and permeability and some of these applications were the outcome of a collaboration with scientists in pfizer (groton, ct, usa). in this publication, after reviewing some basic concepts of qspr and the br analysis methodology, we have showen how br analysis can identify the different features of three polarity descriptors (log poct-tol, epsa and log kw iam ) and thus help to define their applicability domain. in the near future, we plan to apply br analysis to molecular determinants obtained through biomimetic chromatography and biophysical methods for adme-related protein-ligand interactions. to further improve the potential of br analysis, the implementation of new sets of descriptors are in course of study in our laboratories. acknowledgements: we thank gilles goetz (pfizer, groton, ct, usa) for providing epsa values for indinavir and nelfinavir. references [1] k. valko, physicochemical and biomimetic properties in drug discovery, john wiley & sons, hoboken,n.j.,usa, 2014. [2] g. ermondi, m. vallaro, g. caron, learning how to use iam chromatography for predicting permeability. eur. j. pharm. sci. 114 (2018) 385–390. [3] m.h. abraham, hydrogen bonding. 31. construction of a scale of solute effective or summation hydrogen-bond basicity. j. phys. org. chem. 6 (1993) 660–684. [4] m.h. abraham, h.s. chadha, r. leitao, r.c. mitchell, w.j. lambert, r. kaliszan, a. nasal, p. haber, determination of solute lipophilicity, as log p(octanol) and log p(alkane) using poly(styrene– divinylbenzene) and immobilised artificial membrane stationary phases in reversed-phase highperformance liquid chromatography. j. chromatogr. a 766 (1997) 35–47. [5] g. ermondi, g. caron, molecular interaction fields based descriptors to interpret and compare chromatographic indexes. j. chromatogr. a 1252 (2012) 84-89. [6] g. ermondi, a. visconti, r. esposito, g. caron, the block relevance (br) analysis supports the dominating effect of solutes hydrogen bond acidity on δlogp(oct-tol). eur. j. pharm. sci. 53 (2014) 50–54. [7] g. caron, m. vallaro, g. ermondi, the block relevance (br) analysis to aid medicinal chemists to determine and interpret lipophilicity. med. chem. commun. 4 (2013) 1376-1381. [8] a.r. katritzky, m. kuanar, s. slavov, c.d. hall, m. karelson, i. kahn, d.a. dobchev, quantitative correlation of physical and chemical properties with chemical structure: utility for prediction, chem. rev. 110 (2010) 5714–5789. [9] r. todeschini, v. consonni, handbook of molecular descriptors, wiley-vch verlag gmbh & co. kgaa, weinheim, 2008. [10] d.j. livingstone, the characterization of chemical structures using molecular properties. a survey. j. chem. inf. comput. sci. 40 (2000) 195–209. [11] g. cruciani, p. crivori, p.a. carrupt, b. testa, molecular fields in quantitative structure-permeation relationships: the volsurf approach. j. mol. struct. (theochem) 503 (2000) 17–30. [12] p. crivori, g. cruciani, p.-a. carrupt, b. testa, predicting blood−brain barrier permeation from threedimensional molecular structure, j. med. chem. 110 (2000) 2204–2216. ermondi and caron admet & dmpk 6(3) (2018) 215-224 224 [13] r.c. wade, calculation and application of molecular interaction fields, in: g. cruciani (ed.), mol. interact. fields, wiley-vch verlag gmbh & co. kgaa, weinheim, germany, 2006: pp. 27–42. [14] c.a. lipinski, f. lombardo, b.w. dominy, p.j. feeney, experimental and computational approaches to estimate solubility and permeability in drug discovery and development settings. adv. drug deliv. rev. 23 (1997) 3–26. [15] l. michielan, s. moro, pharmaceutical perspectives of nonlinear qsar strategies. j. chem. inf. model. 50 (2010) 961–978. [16] s. wold, e. johansson, m. cocchi, pls: partial least squares projections to latent structures, in: h. kubinyi (ed.), 3d qsar drug des., kluwer academic publisher, dordrecht (the netherlands), 2000: pp. 523–550. [17] a. tropsha, p. gramatica, v. gombar, the importance of being earnest: validation is the absolute essential for successful application and interpretation of qspr models. qsar comb. sci. 22 (2003) 69–77. [18] t. mehmood, k.h. liland, l. snipen, s. sæbø, a review of variable selection methods in partial least squares regression. chemom. intell. lab. syst. 118 (2012) 62–69. [19] g. caron, s. visentin, c. pontremoli, g. ermondi, profile of the intermolecular forces governing the interaction of drugs with mucin. int. j. pharm. 488 (2015) 67–69. [20] g. ermondi, g. caron, do surface-based match solution-based techniques? the case of drug-liposome interaction. int. j. pharm. 508 (2016) 123-127. [21] g.h. goetz, m. shalaeva, g. caron, g. ermondi, l. philippe, relationship between passive permeability and molecular polarity using block relevance analysis. mol. pharm. 14 (2017) 386-395. [22] g. caron, m. vallaro, g. ermondi, g.h. goetz, y.a. abramov, l. philippe, m. shalaeva, a fast chromatographic method for estimating lipophilicity and ionization in nonpolar membrane-like environment. mol. pharm. 13 (2016) 1100–1110. [23] t. potter, g. ermondi, g. newbury, g. caron, relating caco-2 permeability to molecular properties using block relevance analysis. med. chem. commun. 6 (2015) 626-629. [24] g.h. goetz, w. farrell, m. shalaeva, s. sciabola, d. anderson, j. yan, l. philippe, m.j. shapiro, high throughput method for the indirect detection of intramolecular hydrogen bonding. j. med. chem. 57 (2014) 2920–2929. [25] c.r.w. guimarães, a.m. mathiowetz, m. shalaeva, g. goetz, s. liras, use of 3d properties to characterize beyond rule-of-5 property space for passive permeation. j. chem. inf. model. 52 (2012) 882–890. [26] g. caron, g. ermondi, molecular descriptors for polarity: the need of going beyond psa. fut.med.chem. 8 (2016), 2013-2016. [27] g.h. goetz, l. philippe, m.j. shapiro, epsa: a novel supercritical fluid chromatography technique enabling the design of permeable cyclic peptides. acs med. chem. lett. 5 (2014) 1167–1172. [28] s. sciabola, g.h. goetz, g. bai, b.n. rogers, d.l. gray, a. duplantier, k.r. fonseca, m.a. vanasefrawley, n.m. kablaoui, systematic n-methylation of oxytocin: impact on pharmacology and intramolecular hydrogen bonding network. bioorganic med. chem. 24 (2016) 3513–3520. [29] l. grumetto, g. russo, f. barbato, polar interactions drug/phospholipids estimated by iam-hplc vs cultured cell line passage data: their relationships and comparison of their effectiveness in predicting drug human intestinal absorption. int. j. pharm. 500 (2016) 275–290. [30] g. caron, m. vallaro, g. ermondi, log p as a tool in intramolecular hydrogen bond considerations. drug discov. today: technol. (2018) doi.org/10.1016/j.ddtec.2018.03.001. ©2018 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ physiologically-based pharmacokinetic simulations in pharmacotherapy: selection of the optimal administration route for exogenous melatonin doi: 10.5599/admet.625 44 admet & dmpk 7(1) (2019) 44-59; doi: http://dx.doi.org/10.5599/admet.625 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper physiologically-based pharmacokinetic simulations in pharmacotherapy: selection of the optimal administration route for exogenous melatonin adriana savoca, davide manca* pse-lab, process systems engineering laboratory, dipartimento di chimica, materiali e ingegneria chimica “giulio natta”, politecnico di milano, piazza leonardo da vinci 32, 20133 milano, italy *corresponding author: e-mail: davide.manca@polimi.it; tel.: +39 02 23993271; fax: +39 02 23993280 received: october 09, 2018; revised: january 09, 2019; published: february 23, 2019 abstract the benefits of melatonin on human body are drawing increasing attention from several researchers in different fields. while its role as cure for sleep disturbances (e.g., jet lag, insomnia) is well documented and established, new functions in physiological and pathophysiological processes are emerging. to investigate these effects, there is need for the characterization of melatonin transport processes in the body and resulting pharmacokinetics. although recent works propose physiologically-based pharmacokinetic modelling of melatonin, no work has yet highlighted the potential of pbpk simulations to shed light on melatonin pharmacokinetic aspects and discrimination among administration routes. this paper presents, validates, and discusses a versatile pbpk model featuring different ways of administration and compares the resulting pharmacokinetic profiles of intravenous, oral, and transdermal administration, with the goal of understanding which is the optimal route to achieve either physiological and/or supraphysiological melatonin levels. keywords melatonin; pbpk; administration route; transdermal; controlled release; simulation; clinical efficacy. introduction in recent years, physiologically-based pharmacokinetic (pbpk) models have become widely-used and accepted tools to study, simulate, and predict drugs concentration in the body as well as provide insight on their pharmacological effects via combination with pharmacodynamic models. pbpk models are currently applied throughout the phases of drug discovery and development with various goals, e.g., inter-species extrapolation, analysis of chemical toxicity or efficiency, investigation of different routes of administration, and study of inter-individual variability [1-5]. indeed, pbpk simulations are extremely useful to study the pharmacokinetic differences among individuals, from pediatric patients [6] to healthy adults to special subjects, with particular conditions (e.g., pregnancy [7]) or specific diseases with high probability of affecting drugs pharmacokinetics (e.g., renal insufficiency or liver diseases). the reason is that pbpk models (first theorized in 1937 by teorell [8]) incorporate the anatomy and physiology of the patients’ body into the mathematical description of drugs absorption, distribution, metabolism, and elimination (adme) processes. their recent success is also related to the current availability of modern tools to solve complex http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:davide.manca@polimi.it admet & dmpk 7(1) (2019) 44-59 pbpk simulation in pharmacotherapy doi: 10.5599/admet.625 45 mathematical problems, such as systems of ordinary differential equations (odes) with a large number of parameters. in silico simulations are appealing because of the possibility to carry out “free” and fast experiments [9], compared to the actual clinical trials, whose costs and duration have increased over the past 20 years [10]. not only drug discovery and development [11], but also the clinical practice may take advantage from simulation via pbpk models, as it tackles the problem of selecting the optimal dose that maximizes therapeutic efficacy while minimizing adverse effects. on one hand, inter-individual variability and medication errors are significant obstacles in this decision, on the other hand, the choice of the optimal administration route and dosing regimen are crucial degrees of freedom of this problem. in this respect, melatonin is a useful and interesting case-study. the pleiotropic functions of melatonin in the human body are catalyzing the attention of several researchers in different fields, and its exogenous administration can follow different pathways. although melatonin is particularly popular as a cure for sleep disturbances (i.e. jet-lag, insomnia), a number of other physiological and pathophysiological functions have been investigated and are still emerging. for instance, receptor-mediated actions include regulatory functions, e.g., immune response, homeostasis, and blood pressure regulation [12-16]. indeed, melatonin receptors are distributed in the whole body. besides, non-receptor mediated actions are of great interest, especially the potency of its antioxidant, antiproliferative, and anti-inflammatory action via radical scavenging [17]. the application in chemotherapy in combination with other substances improves both the chances of survival and quality of the patients’ life [18]. in healthy people, melatonin is endogenously produced by the pineal gland. the production rhythm is entrained with the day-night cycle, with darkness causing the onset around 9-10 pm, peak between 2-4 am (with cmax range 60-100 pg/ml), and baseline low values during the day at about 5-10 pg/ml [19]. however, this endogenous rhythm may be subject to either disruption or levels reduction, and medical doctors think that this has negative impact on the patients’ health status, especially in critically ills [20-22]. in order to identify the optimal melatonin dosage, a detailed characterization of exogenous melatonin adme processes within the human body is recommended. through the years, several authors have carried out pharmacokinetic studies to identify the most suitable dosage and route of administration to produce physiological and supraphysiological melatonin levels in different populations [23-26]. although some recent works exist on the pbpk modelling of melatonin in the human body (e.g., [27]), our aim is not only to provide a valuable pbpk model but also to compare melatonin levels that result from different routes of administration, i.e. intravenous (iv), oral (per os, po), and transdermal (td). the first goal is to understand which route has the highest potential to reproduce the endogenous profile of healthy patients, with the purpose of restoring melatonin physiological roles. the second goal is to identify the routes that allow achieving higher levels, with the purpose of producing pharmacological effects (for instance strong antioxidative action for icu, intensive care unit, patients). despite high inter-individual variability that is typical of melatonin pharmacokinetics (e.g., related to different physical characteristics, genetic factors, and presence of impairments/diseases), we intend to show that in silico simulations can provide guidance and advice in selecting the optimal routes of administration and dosage, once the reliability of the employed model is verified. indeed, model simulations constitute a powerful tool for optimal pharmacotherapy, especially in combination with experimental studies. savoca and manca admet & dmpk 7(1) (2019) 44-59 46 methods in general, the pbpk approach combines anatomical and physiological aspects with mathematical modeling, by assuming that the organs and tissues of the human body can be represented by compartments with homogeneous concentration. the reference model of this work (from [28]) considers 8 compartments in the description of the human body: plasma, gastric lumen (gl), small intestinal lumen (sil), large intestinal lumen (lil), liver, gastro-intestinal circulatory system (gics), poorly perfused tissues (pt), and highly perfused organs (ho). actually, some compartments represent single organs while other compartments represent lumped parts so to reduce the number of model parameters. in fact, a too high number of parameters may lead to mathematical predicaments of over-parameterization and model identification (see [10] for an exhaustive discussion on this topic). the ho compartment stands for organs that are highly perfused by blood, i.e. kidneys, lungs, spleen, and heart. the pt compartment lumps tissues that are poorly reached by blood vessels, e.g., adipose tissue, skin, and muscles (specifically in ill/treated patients). the gics compartment lumps the portal vein, the mesenteric artery, and the microcirculatory blood vessels of the gastrointestinal system. we applied some modifications to this basic structure of the model to adapt it to melatonin pharmacokinetic features. particularly, we added (i) the pineal gland, and (ii) the salivary glands. pineal gland is the source of endogenous melatonin. within our pbpk model, the material balance on the pineal gland accounts for the production of endogenous melatonin with a term that exhibits a 24-h periodicity (see [2]). several authors evaluate melatonin endogenous and exogenous amount by measuring either saliva and plasma or only saliva concentrations [29-32]. thus, we found more correct (from a physiological point of view) to add the salivary glands to the model compartments. the drug material balances, in the form of an ode system, describe the concentration dynamics of melatonin in each compartment. finally, an additional equation allows accounting for the dynamics of melatonin main metabolite 6-sulfatoxymelatonin (amt6s). in case of iv route, the drug directly inputs the plasma compartment. conversely, in case of po administration, the drug enters the gl and moves through sil and lil to be absorbed through the intestinal walls and conveyed to liver via the portal vein. this results into the so called “first-pass metabolism effect”. after that, it is drained from the liver to reach the systemic circulation and distributes to the other organs and tissues via the bloodstream. it is worth stressing that the model structure takes into consideration gl, sil, and lil only in case of po administration. in fact, in other cases, we assume that the drug counterdiffusion from gics to sil and lil is negligible, and therefore it is possible to neglect such compartments, along with gl and reduce significantly the number of odes. we do not report here the complete mathematical description of the model, as it is extensively detailed in [10,28]. while in case of iv and po routes, the skin is incorporated into the pt compartment, in case of td administration the skin becomes the mean for drug absorption and therefore calls for a specific and detailed description. in particular, melatonin evolution has to be considered not only in time but also along the skin depth coordinate. thus, the homogenous approach (based on the perfectly mixed hypothesis) to compartment modeling is replaced and the resulting skin mathematical description involves partial differential equations (pdes) with suitable boundary conditions [2]. particularly, three skin layers are considered: (i) stratum corneum that is the most external and thinnest but also the main barrier, (ii) viable epidermis that may constitute a metabolism site, and (iii) dermis, from which the drug is supposed to reach the systemic circulation via the contained blood vessels, and then distribute to the rest of the body. in case of td administration, the pdes describing the skin and the odes describing the rest of the body admet & dmpk 7(1) (2019) 44-59 pbpk simulation in pharmacotherapy doi: 10.5599/admet.625 47 are combined via the finite differences method. in fact, the pdes are discretized respect to the spatial coordinate (i.e. skin depth) and therefore converted to odes [2]. independently of the administration pathway, the model parameters can be divided into three categories: (i) individualized, (ii) assigned, and (iii) regressed. individualized parameters (e.g., volumes of compartments and flowrates among them) are calculated according to empirical correlations that are available in the literature and depend on the patients’ physical characteristics. we consider as specific features the sex, body weight, and height. assigned parameters are some drug physicochemical properties whose value can be determined from the scientific literature (e.g., protein binding). some parameters, strictly related to the transport properties, can be neither found in the literature nor calculated by empirical correlations (e.g., diffusivity, transfer coefficients, metabolic constants), thus they are obtained via a nonlinear regression procedure respect to experimental data from the literature. indeed, although the value of some transfer coefficients might be determined from in vitro studies, such experiments would not account for the interactions among organs and tissues in the living organism, and therefore would affect the reliability of the mathematical model and consistency/validity of the simulated results. once the model transfer coefficients and metabolic constants are identified (with data from [33] for iv route, [34] and [29] for po, and [35] for td), a model validation with additional experimental pharmacokinetic data allows assessing its prediction capability. to do so, we chose (i) the median squared error (mese) (eq. (1)) over the mean squared error (mse) [27] for robustness reasons, (ii) the difference between the experimental area under the curve aucexp and the model prediction aucmod ([28], eq. (2)), and (iii) the difference between the observed and predicted values of cmax. comments on the difference between the observed and predicted values of tmax are also present. the auc is calculated via trapezoidal rule over the nm measured concentration values. we consider satisfactory mese values below 0.1 [ng/ml] and %δauc values below 30 %.    2 1, , exp mod i i mese median c c i nm   (1) % 100 exp mod auc exp auc auc auc     (2) once the prediction capability of the pbpk model is evaluated, it is interesting to assess in silico the optimal administration route. in particular, we investigated three distinct administration routes: (i) iv continuous infusion over 7 h, (ii) po in both the immediate and controlled release (cr) formulations (the last one with a release time of 7 h), and (iii) td with a standard patch of 10 cm 2 . the po (cr) tablet release is modeled according to the dissolution characteristics elucidated in [36] and employed in [35]. results from all the administration routes are compared for an assigned dose range between 0.75-12 mg, grounding on the state-of-art pharmacokinetic studies on exogenous melatonin administration. for this preliminary study, we do not consider high doses [37]. our virtual subject is a healthy adult male of 80 kg and 185 cm. to provide a quantitative comparison of the pharmacokinetics resulting from the three administration routes, we calculate and compare the auc and the maximum concentration cmax. finally, we comment on the concentration dynamics in the different compartments of the body that result by simulating the administration of melatonin 3 mg via iv, po, and td routes to the same in silico patient. for an unbiased comparison of melatonin adme, we intentionally neglected the endogenous melatonin production. savoca and manca admet & dmpk 7(1) (2019) 44-59 48 results and discussion we computed the prediction performance with data coming from melatonin pharmacokinetic studies. the validation cases for each route of administration are proposed and discussed. iv validation case figure 1 shows the model curve resulting from the simulation of 20 µg iv infusion over 5 h to 6 healthy subjects (a-f panels) and 1 individual subjected to pinealectomy 2 years earlier (g panel), as in [39]. experimental data (red diamonds) show the individual pharmacokinetic profiles. the model performance (i.e. the blue line) is acceptable, but for a panel of figure 1. nevertheless, the values of the performance indexes (see table 1) remain quite satisfactory, as even the auc value of that patient (a panel) is only slightly higher than 30 %. it is worth observing that the experimental inter-individual variability of melatonin levels is reduced in case of iv administration if compared to other routes (see also figure 2 and figure 3). as a result, also confidence intervals of the iv model parameters are narrower (see values tables a-c reported in appendix). figure 1 (g panel) shows the experimental trend for the pinealectomized patient where the model performance is as good as for the others. the iv model was further tested with experimental results of additional patients from the same study (subjected to bolus injection) and supplementary validation cases [34,39], for which the results of the %δauc, %δcmax, and mese (not reported) are adequate as well. in this case, we do not calculate the experimental/predicted tmax and the relative error, because in case of iv constant rate of infusion the tmax corresponds to the infusion duration (i.e. in this case study, equal to 5 h). table 1 performance indexes %δauc , %δcmax, and mese values for the iv validation case. patient % auc % maxc mese a 31.2 32.1 2.4e-4 b 18.8 24.5 2.6e-4 c 14.2 13.4 4.6e-5 d 20.6 10.1 8.3e-5 e 10.5 10.4 2.0e-5 f 16.3 31.8 4.5e-5 g 0.7 6.4 5.1e-5 po validation case figure 2 shows experimental results on melatonin concentration in case of po administration. the blue line in figure 2 (a and b panels) simulates the pharmacokinetics after administration of 2 and 4 mg respectively, to 12 healthy volunteers [33]. the experimental data (red diamonds) show mean concentration values of the volunteers group. figure 2 (c panel) shows both individual (red dots) and median (black diamonds) concentration profiles of 5 healthy subjects administered with 50 mg [25]. finally, figure 2 (d panel) shows the simulation (blue line) of the averaged profile of 5 healthy subjects administered with 2 mg [40]. in all these cases, only a single curve is displayed, because the literature data report only averaged demographic and/or pharmacokinetic data. despite the literature differences in features and dosages, the model performance is acceptable as the simulation curve is near to the average experimental profile in all the cases. in fact, table 2 lists low values of mese, except for figure 2 (c panel). admet & dmpk 7(1) (2019) 44-59 pbpk simulation in pharmacotherapy doi: 10.5599/admet.625 49 figure 1 experimental data (red diamonds, [37]) represent the pharmacokinetics of 6 healthy subjects (a-f panels) and 1 pinealectomized patient (g panel) who received iv 20 µg infused over 5 h. the blue continuous line is the model-simulated pharmacokinetic profile. savoca and manca admet & dmpk 7(1) (2019) 44-59 50 it is worth observing that the simulated profile (i.e. blue line) anticipates the experimental data (see figure 2 (a, b, and c panels)). the difference between the observed and predicted tmax is about 30 min. this may be related to digestion features and to the patients’ condition (e.g., fed or fasting). future work should adapt the po model to such issues. however, the observed tmax depends also on the experimental protocol, and in particular, on the blood sampling time. in addition, this parameter is affected by a certain degree of experimental error. the %δauc in table 2 is always below 15 % while the relative error between the observed and predicted cmax is below 15 % except for case a. it is fair to acknowledge that the pharmacokinetics resulting from the po route features a higher degree of interand intra-individual variability compared to the iv route, because of several interacting factors that affect absorption (e.g., ph, stomach emptying time, intestinal transit times, and variation of blood supply to stomach and intestine) and metabolism (e.g., genetic factors and presence of diseases). in fact, mese results for the iv validation cases are at least one order of magnitude lower and, consistently, confidence intervals of the model parameters are larger (table b in appendix). additional case studies are employed for validation, with similar results in terms of performance assessment [39]. figure 2 a-b panels: experimental data (red diamonds, [33]), represent the average pharmacokinetics of 12 subjects administered with melatonin po 2 and 4 mg. c panel: experimental individual (red dots) and median (black diamonds) pharmacokinetics after melatonin po 50 mg [25]. d panel: experimental data (red diamonds) averaged over 5 healthy subjects from [40] (melatonin po 2 mg). the blue continuous line is the model-simulated pharmacokinetic profile. td validation case figure 3 shows a validation case from [41] for the td route. in the study, melatonin was administered 2.1 mg/12 cm 2 as td patch. since the reported demographic data consist only of averaged measures over the subjects’ group, the model curve (blue line) is the pharmacokinetics of an averaged individual, while red diamonds represent the experimental values of melatonin concentration of the individuals who took part to admet & dmpk 7(1) (2019) 44-59 pbpk simulation in pharmacotherapy doi: 10.5599/admet.625 51 the study, connected with red lines for the sake of clarity. the model prediction is quite near to three out of four individual trends. the most distant individual trend shows atypical pharmacological levels, which may be related to either differences in the skin features of that specific subject and/or melatonin dermal deposition [41]. this results into a nonsensical value of %δauc (table 3). in general, td pharmacokinetic data show high inter-individual variability having to do with the process of transdermal absorption [41]. this aspect is also reflected in the confidence intervals of the model parameters (table c in appendix) and in the variability of observed cmax and tmax values, although it should be remarked that blood sampling occurred every hour. thus, it is not guaranteed that the real experimental maximum value corresponds to the observed cmax. in any case, the %δcmax is around 25-30 % for the first three individuals. as far as the tmax is concerned, it is worth observing that the model seems to predict a slower absorption compared to the experimental trend. despite the high values of %δauc, it is likely that the reliability of the model for this route might further improve by relying on a higher number of experimental data sets for the identification of the parameters. table 2 performance indexes %δauc and mese values for the po validation case. panel % auc % maxc mese a 2.6 40.1 0.009 b 7.3 13.5 0.103 c 14.3 10.7 5.68 d 8.4 5.0 0.004 figure 3 experimental data (red diamonds, [41]) are the individual pharmacokinetic trends, resulting from melatonin td 2.1 mg over 12 cm 2 patch administration. the blue continuous line is the model-simulated pharmacokinetic profile of the averaged subject. table 3 performance indexes %δauc, %δcmax, and mese values for the td validation case. patient % auc % maxc mese 1 39.1 34.9 8.2e-4 2 38.1 27.7 2.1e-4 3 26.2 23.6 5.4e-5 4 >100 >100 4.0e-3 savoca and manca admet & dmpk 7(1) (2019) 44-59 52 for all the considered routes (i.e. iv, po, and td), the results are acceptable enough to continue with the analysis of melatonin adme in the body as a function of the different administration pathways. in silico simulations for optimal dose selection a number of pharmacokinetic studies focuses on selecting the optimal dose that produces either physiological (e.g., [38]) or supraphysiological levels. in fact, while physiological levels can improve sleep maintenance and resynchronize circadian rhythms [17], supraphysiological levels may produce strong antioxidant action [25] and analgesic effects [37]. to investigate melatonin pharmacokinetic properties, a few studies compare the in vivo results of different routes of administration [34], and/or specific populations (e.g., elderly [26], critically ills [24], patients suffering from severe oxidative stress [25]). to prove the efficiency of in silico simulations within this context, we compare the pharmacokinetic profile resulting from three different routes, with doses ranging from 0.75 to 12 mg. the selected range is considered safe as it has been covered by a number of pharmacokinetic studies. figure 4 shows the results of the simulations, along with comparison to experimental data of endogenous profile in healthy adult volunteers from [32]. as expected, smoother and more sustained levels are achieved via td and po (cr) formulations. the slow absorption phase, which is characteristic of td release, proves particularly suitable for mimicking the endogenous levels produced by the pineal gland. equally, the po (cr) solution provides sustained levels as well (tmax about 4 h), coupled with a steeper absorption (see especially the case of 0.75 mg). this difference in the velocity of absorption has to be considered in the choice of the administration time, as this will affect the onset time of melatonin effects. in case of po (cr) 0.75 mg administration, shifting the time of administration would allow quite a close imitation of the endogenous profile. the same consideration holds for the case of td 6 mg administration. thus, not only dosing, but also the time of administration is a key degree of freedom in the problem of melatonin delivery optimization to restore/produce physiological levels. failing in considering this aspect would likely result into unsatisfactory outcomes in terms of pharmacodynamic effects. in this sense, pbpk simulations can be used as a tool for therapy design, to determine the time of administration that more likely leads to the desired effects. it should also be noted that, although the td route produces sustained levels over 24 h, it is unlikely that the subject will manifest adverse effects, for instance related to sleep. firstly, levels are quite similar to the endogenous pattern (see the black circles), and after about 10 h, they start decreasing towards the daily baseline (black dotted line). secondly, doses up to 3500 mg (po) have been administered without any acute adverse effects and the scientific literature does not report any toxic threshold for melatonin dose [37]. for instance, in [37] there is no evidence of sedative effects for doses up to 100 mg (iv), which would produce more than 3-order-ofmagnitude higher levels than those shown in figure 4, case 12 mg via td route (according to our simulations and consistently with experimental results reported in the study). predictably, even low doses of continuous iv infusion produce the highest levels and bioavailability, thus it is probably the most appropriate mean to reach prompt pharmacological (i.e. supraphysiological) levels. in fact, even for the lowest dose considered (i.e. 0.75 mg), the resulting plasma concentration is an order of magnitude higher than the endogenous one (see black circles compared to the blue line). on the contrary, td administration should be excluded for the purpose of producing pharmacological levels (see highest doses 12 mg in figure 4 and cmax value in figure 5). figure 4 also shows that in case of melatonin, po immediate release formulation is not able to produce sustained levels. however, for doses higher than 5 mg, this administration route can be considered to reach pharmacological levels, alternatively to iv infusion. all of these considerations are confirmed by the values of the pharmacokinetic parameters auc and cmax, compared in figure 5. the highest auc is in fact associated with the iv continuous infusion route, whereas admet & dmpk 7(1) (2019) 44-59 pbpk simulation in pharmacotherapy doi: 10.5599/admet.625 53 the other routes of administration produce lower values. another possibility to be explored is the combination of oral immediate release and cr formulations. figure 4 dynamics of the melatonin plasma concentration over 24 h after iv, po, po (cr), and td administration of 1.5 to 12 mg to a virtual subject (male, adult, 80 kg, 185 cm). black circles describe the endogenous profile in healthy adult volunteers [32]. black horizontal dashed lines indicate the range of endogenous cmax in healthy subjects (60-100 pg/ml). black dotted line marks the average value of daily melatonin baseline in plasma. depending on the treatment goal, for instance the attainment of either endogenous or pharmacological levels, it is possible to explore additional contributing factors, other than the route, dose, and time of administration. in fact, different dissolution curves can be employed in the po (cr) formulations and can be combined with the pbpk model to study the resulting adme processes. the same approach can be applied in case of td route, since both the features, position, and application extent of the patch are degrees of freedom for the medical doctor. the main degree of freedom of the iv infusion route is its duration. once the main goal is assigned (in terms of ideal pharmacokinetic profile for a specific application), an optimization can be carried out to identify the optimal dose and dosing regimen by considering those additional degrees of freedom. savoca and manca admet & dmpk 7(1) (2019) 44-59 54 figure 5 comparison of pharmacokinetic parameters cmax (left panel) and auc (right panel) resulting from the three routes of administration. the x-axis reports the simulated dose range, i.e. 0.75 to 12 mg. since melatonin roles affect several organs and tissues, with cerebral, immune, gastrointestinal, cardiovascular, renal, and endocrine functions [12], and melatonin receptors are distributed in the whole body, model compartment levels should be visualized and discussed, as well. figure 6 shows the simulation in different compartments for 3 mg administered iv, po, po (cr), and td. the slow drug absorption, typical of td administration, is reflected in the slow distribution to the organs/tissues of the body. conversely, iv infusion induces higher levels in all the compartments (see highly perfused organs and liver compartments in figure 6). thus, in case the goal of melatonin administration is a diffused anti-oxidant action in the patient body via radicals scavenging, this route should be preferred. the same can be stated in case of immune system enhancement (hence with potential beneficial effects in terms of cancer cells detection and elimination). in addition, when target organs are the highly perfused ones (e.g., pancreas, liver, and kidneys), this route should be definitely considered. when a more localized target action is required, it should be considered that in case of po administration, higher levels are expected in the liver and gastrointestinal tract, as confirmed by the model simulation. according to [14], melatonin is gastroprotective at endogenous levels, whereas pharmacological levels of melatonin (in combination with other drugs) contribute to healing of gastroduodenal ulcers. the difference of goal (i.e. gastro-protection vs healing of local ulcer) will be the discriminating factor for selecting the most suitable dose to produce either endogenous or higher levels. the velocity of excretion via the kidneys is comparable for both the iv and po routes, and is faster for these routes when compared to td. concluding, anatomical and physiological considerations can be converted into quantitative data to be carefully assessed, analyzed, and visualized via pbpk model simulations. this kind of information is not only useful when several routes of administration are viable, but also especially important when the drug target site is not plasma. admet & dmpk 7(1) (2019) 44-59 pbpk simulation in pharmacotherapy doi: 10.5599/admet.625 55 figure 6 simulation of the melatonin concentration dynamics in the compartments saliva, liver, highly and poorly perfused organs/tissues, and the kidneys-excreted amount after administration of 3 mg via iv, po, po (cr), and td. conclusions while pbpk simulators allow evaluating melatonin levels in plasma and the rest of the body, further practical considerations should support the pharmacokinetic investigation with the aim of achieving optimal clinical efficacy. in fact, while iv route may hold the advantage of the highest bioavailability and fastest distribution to organs and tissues, most of the patients may find it distressing. therefore, it may result suitable only for specific categories such as critically ill patients, who usually receive continuous infusion of different drugs and enteral nutrition for quite long periods. on one hand, po route is easy and simple but it is subject to first-pass hepatic metabolism, which implies a certain degree of inter-individual variability related to different metabolism characteristics, and different patients’ features (e.g., savoca and manca admet & dmpk 7(1) (2019) 44-59 56 gastrointestinal ph, temperature, and other previously mentioned factors). as well as po (cr) option, td route allows obtaining sustained levels and avoids first-pass hepatic metabolism. on the other hand, it is subject to slow absorption through skin, possible metabolism within viable epidermis, and high interand intra-individual variability related to different skin features. in this work, we introduced and discussed a case-study to compare pharmacokinetic levels resulting from different doses and three administration routes, i.e. iv, po, and td, also considering both oral immediate release and cr formulations. pbpk simulations are particularly interesting for their intrinsic nature and structure, because they provide quantitative information on the drug adme processes in the body. besides, coupling with intelligent drug design and in vitro experiments enhances the potential to maximize their efficacy. as far as melatonin is concerned and with reference to both practical and pharmacokinetic aspects, it is possible to conclude that po (cr) and td routes represent the best options in case of disruption of the endogenous rhythm (e.g., in people suffering from either insomnia or jet-lag and critically ills). equally, po (with doses significantly higher than 3 mg) and iv infusion are preferable when higher concentration levels are required for other goals, for instance to contrast severe oxidative stress and possibly cancer, and target specific organs as sites of pharmacological action. this work can be extended and improved by focusing on one administration route and running a numerical optimization of the melatonin dose respect to a target trajectory, also considering a number of degrees of freedom depending on the selected route. it is also worth stressing the transferability of the presented approach to any other drugs that are versatile from the point of view of the administration routes. such investigations may become especially interesting in case of drugs with narrow therapeutic windows, such as chemotherapy drugs whose pharmacokinetics quantification is essential and critical. acknowledgements: the authors acknowledge the precious discussions and suggestions provided by dr. giovanni mistraletti and prof. franco fraschini. conflict of interest: none. references [1] h.m. jones, i.b. gardner, w.t. collard, p.j. stanley, p. oxley, n.a. hosea, d. plowchalk, s. gernhardt, j. lin, m. dickins, s.r. rahavendran, b.c. jones, k.j. watson, h. pertinez, v. kumar, s. cole. simulation of human intravenous and oral pharmacokinetics of 21 diverse compounds using physiologically based pharmacokinetic modelling. clin. pharmacokinet. 50 (2011) 331-347. [2] a. savoca, g. mistraletti, d. manca. a physiologically-based diffusion-compartment model for transdermal administration – the melatonin case study. comput. chem. eng. 113 (2018) 115-124. [3] c. hall, e. lueshen, a. mošat', a.a. linninger. interspecies scaling in pharmacokinetics: a novel whole-body physiologically based modeling framework to discover drug biodistribution mechanisms in vivo. j. pharm. sci. 101 (2012) 1221-1241. [4] f.y. bois, m. jamei, h.j. clewell. pbpk modelling of inter-individual variability in the pharmacokinetics of environmental chemicals. toxicology 278 (2010) 256-267. [5] z. lin, n.a. monteiro-riviere, r. kannan, j.e. riviere. a computational framework for interspecies pharmacokinetics, exposure and toxicity assessment of gold nanoparticles. nanomedicine 11 (2016) 107-119. [6] a. ke, s. nallani, p. zhao, a. rostami-hodjegan, j. unadkat. a pbpk model to predict disposition of cyp3a-metabolized drugs in pregnant women: verification and discerning the site of cyp3a induction. cpt: pharmacometrics & systems pharmacology 1 (2012) 3. admet & dmpk 7(1) (2019) 44-59 pbpk simulation in pharmacotherapy doi: 10.5599/admet.625 57 [7] k. abduljalil, m. jamei, a. rostami-hodjegan, t.n. johnson. changes in individual drug-independent system parameters during virtual paediatric pharmacokinetic trials: introducing time-varying physiology into a paediatric pbpk model. the aaps journal 16 (2014) 568-576. [8] t. teorell. kinetics of distribution of substances administered to the body, i : the extravascular modes of administration. archives internationales de pharmacodynamie et de therapie 57 (1937) 205-225. [9] j.e. sager, j. yu, i. raguenau-majlessi, n. isoherranen. physiologically based pharmacokinetic (pbpk) modeling and simulation approaches: a systematic review of published models, applications and model verification. drug metab. dispos. 43 (2015) 1823-1837. [10] r.a. abbiati, a. savoca, d. manca, chapter 2 an engineering oriented approach to physiologically based pharmacokinetic and pharmacodynamic modeling, in: d. manca (ed.) computer aided chemical engineering, elsevier (2018), pp. 37-63. [11] r. del cont, m. abrami, d. hasa, b. perissutti, d. voinovich, a. barba, g. lamberti, g. grassi, i. colombo, d. manca, m. grassi. a physiologically-oriented mathematical model for the description of in vivo drug release and absorption. admet and dmpk 2 (2014) 80-97. [12] j.r. calvo, c. gonzález-yanes, m.d. maldonado. the role of melatonin in the cells of the innate immunity: a review. j. pineal res. 55 (2013) 103-120. [13] d. tan, r.j. reiter, l.c. manchester, m. yan, m. el-sawi, r.m. sainz, j.c. mayo, r. kohen, m. allegra, r. hardeland. chemical and physical properties and potential mechanisms: melatonin as a broad spectrum antioxidant and free radical scavenger. curr. top. med. chem. 2 (2002) 181-197. [14] k. celinski, s. j. konturek, p.c. konturek, t. brzozowski, h. cichoz‐lach, m. slomka, p. malgorzata, w. bielanski, r.j. reiter. melatonin or l‐tryptophan accelerates healing of gastroduodenal ulcers in patients treated with omeprazole. j. pineal res. 50 (2011) 389-394. [15] s. r. pandi-perumal, i. trakht, v. srinivasan, d. w. spence, g. j. maestroni, n. zisapel, d. p. cardinali. physiological effects of melatonin: role of melatonin receptors and signal transduction pathways. prog. neurobiol. 85 (2008) 335-353. [16] r.m. slominski, r.j. reiter, n. schlabritz-loutsevitch, r.s. ostrom, a.t. slominski. melatonin membrane receptors in peripheral tissues: distribution and functions. mol. cell. endocrinol. 351 (2012) 152-166. [17] r.j. reiter, d.x. tan, a. galano. melatonin: exceeding expectations. physiology (bethesda) 29 (2014) 325-333. [18] p.f. innominato, a.s. lim, o. palesh, m. clemons, m. trudeau, a. eisen, c. wang, a. kiss, k.i. pritchard, g.a. bjarnason. the effect of melatonin on sleep and quality of life in patients with advanced breast cancer. support care cancer 24 (2016) 1097-1105. [19] a. brzezinski. melatonin in humans. n. engl. j. med. 336 (1997) 186-195. [20] g. mistraletti, g. sabbatini, m. taverna, m.a. figini, m. umbrello, p. magni, m. ruscica, e. dozio, r. esposti, g. demartini, f. fraschini, r. rezzani, r.j. reiter, g. iapichino. pharmacokinetics of orally administered melatonin in critically ill patients. j. pineal res. 48 (2010) 142-147. [21] r.t. dauchy, s. xiang, l. mao, s. brimer, m.a. wren, l. yuan, m. anbalagan, a. hauch, t. frasch, b.g. rowan, d.e. blask, s.m. hill. circadian and melatonin disruption by exposure to light at night drives intrinsic resistance to tamoxifen therapy in breast cancer. cancer res. 74 (2014) 4099-4110. [22] a. damasceno, a.s. moraes, a. farias, b.p. damasceno, l.m. dos santos, f. cendes. disruption of melatonin circadian rhythm production is related to multiple sclerosis severity: a preliminary study. j. neurol. sci. 353 (2015) 166-168. [23] s.l. markantonis, e. tsakalozou, a. paraskeva, c. staikou, a. fassoulaki. melatonin pharmacokinetics in premenopausal and postmenopausal healthy female volunteers. j. clin. pharmacol. 48 (2008) 240-245. [24] g. mistraletti, r. paroni, m. umbrello, l. d'amato, g. sabbatini, m. taverna, p. formenti, e. finati, g. favero, f. bonomini, r. rezzani, r.j. reiter, g. iapichino. melatonin pharmacological blood levels increase total antioxidant capacity in critically ill patients. int. j. mol. sci. 18 (2017) e759. savoca and manca admet & dmpk 7(1) (2019) 44-59 58 [25] h.f. galley, d.a. lowes, l. allen, g. cameron, l.s. aucott, n.r. webster. melatonin as a potential therapy for sepsis: a phase i dose escalation study and an ex vivo whole blood model under conditions of sepsis. j. pineal res. 56 (2014) 427-438. [26] n.s. gooneratne, a.y. edwards, c. zhou, n. cuellar, m.a. grandner, j.s. barrett. melatonin pharmacokinetics following two different oral surge-sustained release doses in older adults. j. pineal res. 52 (2012) 437-445. [27] h.t. peng, f. bouak, o. vartanian, b. cheung. a physiologically based pharmacokinetics model for melatonin--effects of light and routes of administration. int. j. pharm. 458 (2013) 156-168. [28] r.a. abbiati, g. lamberti, m. grassi, f. trotta, d. manca. definition and validation of a patientindividualized physiologically-based pharmacokinetic model. comput. chem. eng. 84 (2016) 394408. [29] s.i. shirakawa, s. tsuchiya, y. tsutsumi, t. kotorii, n. uchimura, t. sakamoto, s. yamada. time course of saliva and serum melatonin levels after ingestion of melatonin. psychiatry and clinical neurosciences 52 (1998) 266-267. [30] m.l. laakso, t. hätönen, d. stenberg, a. alila, s. smith. one-hour exposure to moderate illuminance (500 lux) shifts the human melatonin rhythm. j. pineal res. 15 (1993) 21-26. [31] s. benloucif, h.j. burgess, e.b. klerman, a.j. lewy, b. middleton, p.j. murphy, b.l. parry, v.l. revell. measuring melatonin in humans. j. clin. sleep med. 4 (2008) 66-69. [32] a. voultsios, d.j. kennaway, d. dawson. salivary melatonin as a circadian phase marker: validation and comparison to plasma melatonin. j. biol. rhythms 12 (1997) 457-466. [33] r.l. demuro, a.n. nafziger, d.e. blask, a.m. menhinick, j.s. bertino, jr. the absolute bioavailability of oral melatonin. j. clin. pharmacol. 40 (2000) 781-784. [34] l.p.h. andersen, m.u. werner, m.m. rosenkilde, n.g. harpsøe, h. fuglsang, j. rosenberg, i. gögenur. pharmacokinetics of oral and intravenous melatonin in healthy volunteers. bmc pharmacol. toxicol. 17 (2016) 8. [35] l. benes, b. claustrat, f. horriere, m. geoffriau, j. konsil, k.a. parrott, g. degrande, r.l. mcquinn, j.w. ayres. transmucosal, oral controlled-release, and transdermal drug administration in human subjects: a crossover study with melatonin. j. pharm. sci. 86 (1997) 1115-1119. [36] b.j. lee, k.a. parrott, j.w. ayres, r.l. sack. development and characterization of an oral controlled-release delivery system for melatonin. drug dev. ind. pharm. 22 (1996) 269-274. [37] l.p. andersen, m.u. werner, m.m. rosenkilde, a.q. fenger, m.c. petersen, j. rosenberg, i. gogenur. pharmacokinetics of high-dose intravenous melatonin in humans. j. clin. pharmacol. 56 (2016) 324-329. [38] c. mallo, r. zaidan, g. galy, e. vermeulen, j. brun, g. chazot, b. claustrat. pharmacokinetics of melatonin in man after intravenous infusion and bolus injection. eur. j. clin. pharmacol. 38 (1990) 297-301. [39] j.b. fourtillan, a.m. brisson, p. gobin, i. ingrand, j.p. decourt, j. girault. bioavailability of melatonin in humans after day-time administration of d(7) melatonin. biopharm. drug dispos. 21 (2000) 1522. [40] m. aldhous, c. franey, j. wright, j. arendt. plasma concentrations of melatonin in man following oral absorption of different preparations. br. j. clin. pharmacol. 19 (1985) 517-521. [41] d. aeschbach, b.j. lockyer, d.j. dijk, s.w. lockley, e.s. nuwayser, l.d. nichols, c.a. czeisler. use of transdermal melatonin delivery to improve sleep maintenance during daytime. clin. pharmacol. ther. 86 (2009) 378-382. admet & dmpk 7(1) (2019) 44-59 pbpk simulation in pharmacotherapy doi: 10.5599/admet.625 59 appendix table a. – key model parameters in case of the iv route, correlated by 90 % confidence intervals. parameters description regressed values 90 % 𝑪𝑰𝒍𝒃 90 % 𝑪𝑰𝒖𝒃 𝒌𝑻−𝑷 [ 𝒎𝒊𝒏 −𝟏] plasma-poor tissues transfer coefficient 0.3955 0.373 0.418 𝒌𝑷−𝑻 [ 𝒎𝒊𝒏 −𝟏] poor-tissues plasma transfer coefficient 0.8 0.774 0.826 𝒌𝑯𝑷−𝑷 [ 𝒎𝒊𝒏 −𝟏] highly perfused organs plasma transfer coefficient 0.047 0.045 0.05 𝒌𝑷−𝑯𝑷 [ 𝒎𝒊𝒏 −𝟏] plasma-highly perfused organs transfer coefficient 1.48 1.416 1.544 table b. – key model parameters in case of the po route, correlated by 90 % confidence intervals. parameters description regressed values 90 % 𝑪𝑰𝒍𝒃 90 % 𝑪𝑰𝒖𝒃 𝒌𝑨,𝑺𝑰𝑳 [ 𝒎𝒊𝒏 −𝟏] 𝑆𝐼𝐿 absorption constant 2.205 0.245 5.655 𝒌𝑪𝑨,𝑺𝑰𝑳 [ 𝒎𝒊𝒏 −𝟏] 𝑆𝐼𝐿 counter-diffusion constant 2.920 0.158 5.683 𝒌𝑨,𝑳𝑰𝑳 [ 𝒎𝒊𝒏 −𝟏] 𝐿𝐼𝐿 absorption constant 0.167 0.003 0.337 𝒌𝑪𝑨,𝑳𝑰𝑳 [ 𝒎𝒊𝒏 −𝟏] 𝐿𝐼𝐿 counter-diffusion constant 0.455 0.059 0.851 𝑬𝒇𝒇𝑯 [ −] hepatic metabolism efficiency 0.467 0.235 0.699 𝑬𝒇𝒇𝑲 [ −] kidneys excretion efficiency 0.053 0. 007 0.113 table c. – key model parameters in case of the td route, correlated by 90 % confidence intervals. parameters description regressed values 90 % 𝑪𝑰𝒍𝒃 90 % 𝑪𝑰𝒖𝒃 𝕯𝑺𝑪 [ 𝒄𝒎𝟐 𝒎𝒊𝒏 ⁄ ] 𝑆𝐶 diffusivity 3.352 ∗ 10 −5 1.269 ∗ 10−5 5.431 ∗ 10−5 𝕯𝑽𝑬 [ 𝒄𝒎𝟐 𝒎𝒊𝒏 ⁄ ] 𝑉𝐸 diffusivity 5.943 ∗ 10 −3 2.122 ∗ 10−3 9.763 ∗ 10−3 𝕯𝑫𝑬 [ 𝒄𝒎𝟐 𝒎𝒊𝒏 ⁄ ] 𝐷𝐸 diffusivity 2.776 ∗ 10 −3 1.395 ∗ 10−3 6.946 ∗ 10−3 𝒌𝒑𝒂𝒓𝒕𝟏 [−] 𝑆𝐶/𝑉𝐸 partition coefficient 1.763 1.209 2.317 ©2019 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ computational screens can speed up the discovery of pharmaceutical cocrystals doi: 10.5599/admet.641 284 admet & dmpk 6(4) (2018) 284-287; doi: http://dx.doi.org/10.5599/admet.641 open access: issn: 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index communication computational screens can speed up the discovery of pharmaceutical cocrystals rafel prohens a and christopher a. hunter b a unitat de polimorfisme i calorimetria, centres científics i tecnològics, universitat de barcelona, baldiri reixac 10, 08028 barcelona, spain b department of chemistry, university of cambridge, lensfield road, cambridge cb2 1ew, uk e-mails: rafel@ccit.ub.edu (prohens); herchelsmith.orgchem@ch.cam.ac.uk (hunter) received: november 26, 2018; accepted: november 27, 2018; published: december 27, 2018 abstract the calculation of surface site interaction points for cocrystal computational screens in combination w ith efficient experimental cocrystallization techniques has been applied successfully to several drug compounds. the basics of this combined approach are briefly reviewed in this communication. keywords cocrystals; virtual screen; h-bond parameters the search for new cocrystals of apis can be very expensive in terms of time and economical resources since the number of potential coformers available with suitable toxicity profiles is huge. thus, there is a real need in the pharmaceutical industry to optimize screening methodologies, because cocrystals offer solutions to some of the drawbacks inherent to many drug candidates, such as low solubility and bioavailability. historically, the supramolecular synthon approach has been the paradigm of crystal engineering for the design of new multicomponent crystals based on strong intermolecular interactions [1]. although very limited in the scope for efficiently locating new coformers, supramolecular synthons are the basis for understanding the formation of a cocrystal. more recently other approaches have been developed for predicting the outcome of cocrystallization experiments. the hansen solubility parameter [2] has been investigated to guide cocrystal screening. statistical analyses of the properties of compounds known to form cocrystals show promise. for example, fabián showed that cocrystals are usually formed by molecules of similar shapes and polarities [3], and the hydrogen bond propensity tool developed by the cambridge crystallographic data centre has been used to predict the cocrystallization of several apis [4]. computational methods have also been applied to the cocrystal problem. a high-level dft study of 350 organic cocrystals concluded that the formulation of general rules to guide cocrystal formation remains difficult [5]. we have developed a fast computational method to screen large libraries of coformers [6]. the approach uses hunter's hydrogen bond parameters ( and ) [7] to predict the formation of intermolecular interactions in crystals and is based on the principle that the crystal structure of an organic compound is http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:rafel@ccit.ub.edu mailto:herchelsmith.orgchem@ch.cam.ac.uk admet & dmpk 6(4) (2018) 284-287 computational screens of pharmaceutical cocrystals doi: 10.5599/admet.641 285 controlled by a hierarchical organisation of functional group interactions. thus, the energy of a solid form (e) can be calculated by pairing the best h-bond donor with the best h-bond acceptor, the second best hbond donor with the second best h-bond acceptor, and so on, until all of the interaction sites (i,j) are satisfied (eq. 1). i j ij e    (1) the difference between the interaction site pairing energies of the cocrystal (ecc) and the two components (e1 and e2) is related to the cocrystallization probability (eq. 2). 1 2cce e ne me    (2) where n and m refer to the cocrystal stoichiometry. the method is based on assumptions that ignore all geometrical aspects of crystal packing: (a) all potential interactions in the crystalline solid are established (b) all h-bond sites on the surface of a molecule are independent and free to interact with any other site in the solid, (c) steric constraints on contacts are not considered, (d) cooperativity between sites is ignored. the consequence of these approximations is that the method offers a solution to the problem of rapid virtual screening of large compound libraries. the method was initially calibrated by performing calculations on two apis with the highest number of reported experimental cocrystals (caffeine and carbamazepine) and a list of nearly 1,000 potential coformers, providing a validation of the relationship between the value of δe and the probability of cocrystal formation (fig.1). figure 1. recall plots for the prediction of 1:1 cocrystals of caffeine (a) and carbamazepine (b) with 849 and 860 potential coformers respectively. h is the fraction of total hits found plotted as function of the fraction of compounds screened (n). the line (r) represents probability of finding a hit as the result of random chance. (c) relationship between the probability of obtaining a cocrystal, p, and the calculated cocrystal energy, . in addition, a set of validations was conducted with cocrystal screens of nine apis from the literature (diclofenac, piracetam, pyrazine carboxamide, acetazolamide, indomethacin, furosemide, nalidixic acid, prohens and hunter admet & dmpk 6(4) (2018) 284-287 286 paracetamol and a drug candidate) that had reported both successful and unsuccessful coformers, and in most cases, calculations reproduced the experimental results [8]. the method has been further applied in combination with experimental screening techniques to optimize the search for new cocrystals with improved physicochemical properties of apis, such as nalidixic acid [9] spironolactone, griseofulvin [10], zafirlukast [11], sildenafil [12,13] and a novel inhaled jak-stat inhibitor [14]. figure 2. drug compounds tested in combined virtual/experimental cocrystal screens virtual cocrystal screening for coformer selection is just part of the more complicated procedure of cocrystal investigation in the pharmaceutical industry. thus, other pre-screening techniques must be applied prior to the experimental search for new cocrystals in order to reduce time and optimize resources. for instance, solubility prediction can also be conducted if an increase/decrease in solubility is a requirement for the formulation [15], and the safety/toxicity profile of every potential coformer must be assessed using databases or in silico prediction [16]. finally, once all the computational work is done, information about the solubility of the api and the chosen coformers in the organic solvents used for crystallisation and about the presence of polymorphism/solvatomorphism is necessary. this information is essential for selection of the experimental conditions which will allow the most extensive and efficient exploration of the cocrystal landscape. in summary, the use of computational approaches together with experimental screening methodologies, if combined efficiently, can reduce the time and cost of the development of a new cocrystal formulation. references [1] g.r. desiraju. supramolecular synthons in crystal engineering a new organic synthesis. angew. chem. int. ed. 34 (1995) 2311-2327. [2] m.a. mohammad, a. alhalaweh, s.p. velaga. hansen solubility parameter as a tool to predict cocrystal formation. int. j. pharm. 407 (2011) 63-71. [3] l. fabián, cambridge structural database analysis of molecular complementarity in cocrystals. cryst. growth des. 9 (2009) 1436–1443. [4] a. delori, p.t.a. galek , e. pidcock , m. patni, w. jones. knowledge-based hydrogen bond prediction and the synthesis of salts and cocrystals of the anti-malarial drug pyrimethamine with various drug and gras molecules. crystengcomm 15 (2013) 2916-2928. admet & dmpk 6(4) (2018) 284-287 computational screens of pharmaceutical cocrystals doi: 10.5599/admet.641 287 [5] c.r. taylor, g.m. day. evaluating the energetic driving force for cocrystal formation. cryst. growth des. 18 (2018) 892-904. [6] d. musumeci, c.a. hunter, r. prohens, s. scuderi, j.f. mccabe. virtual cocrystal screening. chem. sci. 2 (2011) 883-890. [7] c.a. hunter. quantifying intermolecular interactions: guidelines for the molecular recognition toolbox. angew. chem. int. ed. 43 (2004) 5310–5324. [8] t. grecu, c.a. hunter, e.j. gardiner, j.f. mccabe. validation of a computational cocrystal prediction tool: comparison of virtual and experimental cocrystal screening results. cryst. growth des. 14 (2014) 165–171. [9] t. grecu, h. adams, c.a. hunter, j.f. mccabe, a. portell, r. prohens. virtual screening identifies new cocrystals of nalidixic acid. cryst. growth des. 14 (2014) 1749–1755. [10] t. grecu, r. prohens, j.f. mccabe, e.j. carrington, j.s. wright, l. brammer, c.a. hunter. cocrystals of spironolactone and griseofulvin based on an in silico screening method. crystengcomm 19 (2017) 3592-3599. [11] a. llinas, r. barbas, m. font-bardia, m. j. quayle, s. velaga, r. prohens. two new polymorphic cocrystals of zafirlukast: preparation, crystal structure, and stability relations. cryst. growth des. 15 (2015) 4162-4169. [12] p. lawton, h. l. jones, n.m. kanji, r. barbas, r. prohens. pct/gb2017/050921. [13] r. barbas, m. font-bardia, a. paradkar, c.a. hunter, r. prohens. combined virtual/experimental multicomponent solid forms screening of sildenafil: new salts, cocrystals, and hybrid salt– cocrystals. cryst. growth des. (2018) in press: doi: 10.1021/acs.cgd.8b01413. [14] a. llinas, r. barbas, m. font-bardia, a. smailagic, r. prohens. derisking development by cocrystallisation screen of a novel selective inhaled jak-stat inhibitor. cryst. growth des. (2018) in press: doi: 10.1021/acs.cgd.8b01492. [15] c.a. hunter, r. prohens. solid form and solubility. crystengcomm 19 (2017) 23-26. [16] p.a. corner, d. berry, j.f. mccabe, r. barbas, r. prohens, h. du, h. zhou, a. llinas. property prediction and pharmacokinetic evaluation of mixed stoichiometry cocrystals of zafirlukast, a drug delivery case study. crystengcomm 20 (2018) 1346-1351. ©2018 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ effects of amount of excess solid, the type of stirring and sedimentation time on solubility of sodium phenytoin and lamotrigine doi: 10.5599/admet.621 269 admet & dmpk 6(4) (2018) 269-278; doi: http://dx.doi.org/10.5599/admet.621 open access: issn: 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper effects of amount of excess solid, the type of stirring and sedimentation time on solubility of sodium phenytoin and lamotrigine shahrzad moattar mohammadi 1,2 , ali shayanfar 3 *, shahram emami 3 , abolghasem jouyban 4 1 biotechnolgy research center, tabriz university of medical sciences, tabriz, iran 2 student research committee and faculty of pharmacy, tabriz university of medical sciences, tabriz, iran 3 drug applied research center and faculty of pharmacy, tabriz university of medical sciences, tabriz, iran 4 pharmaceutical analysis research center and faculty of pharmacy, tabriz university of medical science, tabriz, iran *corresponding author: e-mail: shayanfara@tbzmed.ac.ir; tel.: +984133341315; fax: +1-111-111-112 received: september 27, 2018; revised: october 22, 2018; published: november 03, 2018 abstract solubility is the maximum quantity of a drug dissolved in a given volume of solvent at a specific temperature. several factors affect equilibrium solubility. therefore, different solubility data have been reported for a solute in a certain solvent and temperature in the literature. these variations in solubility are one of the possible reasons for unsuccessful attempts of medicinal chemists for developing models as well as deviation of experimental works for solubility prediction in aqueous, non-aqueous and solvent mixtures. the present research aim is to investigate the effect of the amount of excess solid and the t ype of stirring on the solubility of drugs. the solubility of two antiepileptic drugs, namely sodium phenytoin and lamotrigine was determined in water, ethanol, and hcl 0.1 m at 37 °c. different excess amounts of drugs were added to the constant volume of solvent. additionally, different stirring methods such as magnetic stirrer and shake-flask and sedimentation time were investigated on the solubility values. saturation solubility of drugs after dilution with water was measured using a spectrophotometer, and the concentration was calculated according to the calibration curve. amount of excess solid, especially when the drug is in ionized form, and sedimentation time after 24 h have a significant effect on solubility values. keywords antiepileptic drugs; equilibrium solubility; determination introduction solubility of drugs is an important aspect from the earliest stages of drug discovery to the latest stage of the drug formulation [1,2]. it is defined as the maximum quantity of a drug dissolved in a given volume of a solvent at a certain temperature [3]. solubility data for a solute, especially for pharmaceuticals, is a significant physicochemical property in crystallization, extraction of an analyte from different matrices, evaluation of oral bioavailability and preparation of liquid and semi-solid dosage forms. furthermore, solubility in water is needed to make a solution of the drug or drug-like molecule to be tested for its pharmacological/toxicological activities. despite simple methods developed for experimental http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:shayanfara@tbzmed.ac.ir mailto:shayanfara@tbzmed.ac.ir ali shayanfar et al. admet & dmpk 6(4) (2018) 269-278 270 determination of the solubility of drugs [4], various parameters can affect accuracy of the obtained data [5]. several different sets of solubility data have been reported for a drug at a certain solvent and temperature in the literature [6-8]. these solubility variations are one of the possible reasons for unsuccessful attempts of medicinal chemist in developing models for solubility prediction in water, non-aqueous solvents and solvent mixtures with deviation in agreement with experimental works. for example, the best model for aqueous solubility prediction has a prediction error more than 100 % [9], and estimation error for solubility in the solvent mixture is higher than 25 % for pharmaceuticals [6]. there are various methods for solubility determination of pharmaceuticals [10,11]. the shake-flask method is a common method for solubility determination which an excess amount of drug is added to a certain solvent and after appropriate shaking at a constant temperature to reach the equilibrium condition, and the saturated concentration is determined by a valid analytical method [11]. in some cases, variation of solubility data originates from the nature of studied compounds, such as polymorphism and formations of drug aggregates, micelles, drug-buffer complexes, stability, solute and solvent impurities; they can affect the solubility value. another problem in solubility determination is invalidity of the applied method for quantification of drug concentration. other sources of variations, which can affect solubility values related to the solubility determination process such as time of equilibrium, incomplete dissolution over the equilibration time and inappropriate phase separation solution handling and adsorption to untreated surfaces, amount of residual solid, speed and stirring or shaking patterns of the suspension [5,7,12]. although, various sets of solubility data for chemical and pharmaceutical materials in water and organic solvents are annually collected by different research groups and reported in the literature, there are few investigations into the effect of different parameters on solubility determination of drugs. baka et al. [8] investigated the effect of various parameters such as amount of solid excess, stirring time, temperature, and sedimentation time on solubility of hydrochlorothiazide as a model drug. some reports have been investigated the effect of excess solid on solubility of drugs [13,14]; and the reported results are discussed in review articles [11,15]. lamotrigine is a basic drug (pka=5.7) [16] existing in unionized form in water, while it is in ionized form in 0.1 m hcl. sodium phenytoin is a salt form of phenytoin (an acidic compound with pka=8.3 [16]) being partially ionized form in water. in addition, it changes to the parent drug (un-ionized form) in acidic medium. in this study, the effects of the amount of solid excess, shaking patterns of solution and sedimentation time were investigated on solubility of two antiepileptic drugs, namely sodium phenytoin and lamotrigine in water, ethanol and acidic medium (hcl 0.1 m). experimental samples and chemicals sodium phenytoin (99 %) and lamotrigine (99 %) were purchased from alhavi pharmaceutical company (tehran, iran). ethanol (99 %) was provided from scharlau (barcelona, spain) and hcl (37 %) was supplied from merck (darmstadt, germany), and double distilled water was used as received in the lab. apparatus a shaker incubator (heidolph, germany) or a magnetic stirrer (heidolph mr hei-tec, germany) was used to mix two phases and maintain the temperature of samples at 37±0.1 °c during the solubility measurements. samples were filtered using a hydrophilic membrane (450 nm pore size). the concentration of the saturated solution was measured using a uv–vis spectrophotometer (shimadzu, admet & dmpk 6(4) (2018) 269-278 solubility of sodium phenytoin and lamotrigine doi: 10.5599/admet.621 271 japan). the ph of the solutions was measured by a ph meter (metrohm, switzerland). measurement of solubility and investigation of the effect of excess amount of solute and type of stirring on solubility excess amount of each drug was added to a small volume (e.g. 5 ml) of the studied solvent in an erlenmeyer flask. the amount of excess was added according to the preliminary experiments or reported solubility of drugs in the studied solvent from the literature (solubility of phenytoin in water and ethanol are 86 mg/ml and 19.9 mg/ml at 35 °c [17], respectively and lamotrigine solubility in water and hcl 0.1 m are 0.19 mg/ml [18] and 2.22 mg/ml [19] at 25 °c, respectively). therefore, in this study, the volume of solvent was considered 5 ml in an erlenmeyer flask, being similar to the routine experiment of solubility determination in the pharmaceutical and chemical laboratories. for highly soluble compounds, e.g. sodium phenytoin in water 10 %, 25 %, 100 %, 150 % and 200 % of saturated solubility of the drug was applied as an excess solid amount. while for low solubility values, e.g. lamotrigine in water, the minimum amount of drug could be weighted by a digital balance (the minimum amount of weighting 10-30 mg), was selected as the lowest concentration of suspension for solubility determination and other solutions up to 5-folds of this amount were prepared to simulate a routine solubility determination experiment. then, the samples were placed in an incubator equipped with a temperature-controlling system at 37 °c in 150 rpm for 48 hours. afterward, the saturated solutions of the drugs were filtered using hydrophilic filters (0.45 µm). in addition, magnetic stirrer and shake-flask approaches, separately, were used to saturate sodium phenytoin and lamotrigine in water at 37 °c and 150 rpm for 48 hours. furthermore, the concentration of samples was measured after 24-hour sedimentation. the concentration of samples was measured by a uv spectrophotometer. each datum was obtained from at least three replicate independent measurements, and the maximum relative standard deviation (rsd) of each experiment was 10 %. results the effect of the excess solid amount on solubility of sodium phenytoin in water, ethanol and hcl (0.1 m) figure 1 shows the solubility of sodium phenytoin in water at 37 °c. the experiment was investigated in various excess amounts of phenytoin sodium. a linear relationship exists between excess solid amount of drug and solubility. the percentage change between the maximum and minimum values of solubility is 116 %; however, the excess amount has a low effect on solubility of phenytoin in hcl 0.1 m (figure 2) in comparison to solubility in water (percentage change was less than 30 %). moreover, to confirm the effect of excess amount in ionized form, phenytoin solubility was performed in ethanol (dielectric constant of ethanol is lower than water and the ionization in ethanol is less than water) and the results were illustrated in figure 3. a similar pattern to hcl (percentage difference was 21 %) was obtained for solubility of sodium phenytoin in ethanol. the effect of the excess solid amount on solubility of lamotrigine in water, ethanol and hcl (0.1 m) figure 4 illustrates solubility of lamotrigine in water at 37 °c in different amounts of excess solid, and the obtained data revealed that excess amount had no significant effect on solubility value. however, it has a considerable effect on solubility value in hcl 0.1 m (figure 5, the percentage difference between the maximum and minimum amount of solubility value was higher than 40 %), which lamotrigine is in an ionized form similar to the data obtained for phenytoin sodium. ali shayanfar et al. admet & dmpk 6(4) (2018) 269-278 272 the effect of type of stirring and sedimentation time on solubility of sodium phenytoin and lamotrigine in water magnetic stirrer and shake-flask approaches were used to saturate sodium phenytoin and lamotrigine in water at 37 °c. figure 6 illustrates the effect of type of stirring on the obtained solubility value, and shown, the type of stirring for solubility determination has no significant effect on the saturated concentration (solubility) of phenytoin sodium. moreover, sedimentation time after 24-hour (figures 7 and 8) was investigated on the solubility values, and a significant change was observed in solubility of sodium phenytoin in both stirring instruments. similar results were observed for solubility determination of lamotrigine in water (figure 9). the data indicate that sedimentation time seems to have a significant effect on solubility value, particularly for stirring of the solution with magnet stirrer. discussion the effect of excess amount on solubility of sodium phenytoin and lamotrigine the study results show that excess solid amount of the studied drugs has a considerable effect on solubility, especially in ionized form i.e. sodium phenytoin in water and lamotrigine in hcl (0.1 m). one of the reasons for change in solubility in the presence of different values of excess solid is change in ph of saturated solution [20]. in this study, solubility was determined in aqueous solution at own ph. however, ph determination of final solutions (various excess amounts) showed no substantial change (ph: 11.6±0.05). a differential scanning calorimetry (dsc) study of the remaining solid of lamotrigine confirms no possible change in crystalline phase due to a sharp peak corresponding to lamotrigine melting at 220°c, being in agreement with the literature [19]. however, dsc thermogram sodium phenytoin in water exhibited some new peaks resultant to phase transformation of drug to the hexahydrate form similar to the reported study by rubino [21], and it is not possible to judge about the polymorphic change or dimer formation whenever the amount of excess solid is changed. kawakami et al. [13] have reported that solubility of indomethacin increased with an increase in the solid amount in ionizable form (ph=6.5-7) against indomethacin solubility in non-ionizable form (ph=5). they discussed the anomalous behavior in dissolution and crystallization rates at near equilibrium conditions as the possible mechanism. however, one of the potential reasons based on avdeef’s comment on that study is the possible partitioning (adsorption) of the charged form of indomethacin onto the excess solid present in the suspensions, and some ionizable compounds may be surface-active. however, further mechanisms such as dimer formation are possible, and the results are not interpretable without analysis of the solid form [15]. nevertheless, a similar research study about solubility measurement of hydrochlorothiazide reported no significant difference for solubility values in water when different excess amounts of the drug were used [8]. hydrochlorothiazide as an acidic compound converts to non-ionized form in saturated aqueous solution. another study by mosharraf et al. [22] reported that solubility of chemical compounds (in salt form) was affected by the amount of excess solid. increasing the amount of excess solid of barium sulphate or calcium carbonate can affect solubility value due to the disordering particle surface and existence of a peripheral disordered layer. according to the obtained study, solubility of chemical compounds can change by the amount of excess solid, and the mutual issue between the reported studies is ionization and salt form of solute. change in the diffusion layer and disordering particle surface due to adsorption of the charged form of the drug and admet & dmpk 6(4) (2018) 269-278 solubility of sodium phenytoin and lamotrigine doi: 10.5599/admet.621 273 surface activity of solute in ionized form could be an acceptable interpretation for increasing solubility values in the higher amount of excess solids. similar phenomena have been proposed for increase in saturation solubility of nanocrystals (size of particle less than 1000 nm) by the change in layer of unstirred layer surrounding the drug particle (diffusion layer) [23]. moreover, when drug salts are added to solution, their solubility is expected to vary according to the amount of solid added, since the solubility product may not be satisfied below a certain amount of added salt. the partial conversion back of sodium phenytoin to free acid form during solubility testing reported by chaing and wang [24] could be reason for variability in the solubility values. the effect of type of stirring and sedimentation time on solubility of lamotrigine and phenytoin sodium type of stirring is a critical parameter in evaluating dissolution of pharmaceuticals [24]. however, it has been not yet considered as an effective factor in solubility determination of solute. magnetic stirrer and shake-flask are common approaches in the literature to attain the solution to equilibrium. type of stirring could be considered as a possible reason for the variation of solubility values because of supersaturation phenomenon. it is a state of a solution that contains more of the dissolved material than could be dissolved in the solvent [7]. the results in figures 6-9 show that type of stirring has a slight effect on obtained solubility for lamotrigine and phenytoin sodium. however, sedimentation time after 24-hour, in both types of stirring, has a considerable effect on solubility values. in this regard, baka et al. [8] conducted the only study about the effect of sedimentation time for solubility of hydrochlorothiazide; similar to the results of this study, sedimentation time had a considerable effect on the aqueous solubility of drug. the results of this study indicate that although solubility determination of drugs seems a simple experiment, the numerical values of solubility of a compound are affected by different factors. amount of excess solid and sedimentation time could have a significant effect on solubility values. figure 1. the effect of excess solid amount on solubility of sodium phenytoin in water at 37 °c (added weight of drug associated with each of the points in 5 ml of solution is 450, 500, 750, 1000, 1250 mg, respectively). ali shayanfar et al. admet & dmpk 6(4) (2018) 269-278 274 figure 2. the effect of the excess solid amount on solubility of sodium phenytoin in hcl 0.1 m at 3 7 °c. (added weight of drug associated with each of the points in 5 ml of solution is 15, 25, 50 and 75 mg, respectively). figure 3. the effect of the excess solid amount on solubility of sodium phenytoin in ethanol at 37 °c. (added weight of drug associated with each of the points in 5 ml of solution is 125, 150, 175, 200, 225, 350 and 450 mg, respectively). figure 4. the effect of the excess solid amount on solubility of lamotrigine in water at 37 °c. (added weight of drug associated with each of the points in 5 ml of solution is 25, 75, 175, 250, 375, 500 mg, respectively). admet & dmpk 6(4) (2018) 269-278 solubility of sodium phenytoin and lamotrigine doi: 10.5599/admet.621 275 figure 5. the effect of the excess solid amount on solubility of lamotrigine in hcl (0.1m) at 37 °c. (added weight of drug associated with each of the points in 5 ml of solution is 25, 30, 35, 40, 50 and 75 mg, respectively). figure 6. the effect of type of stirring on sodium phenytoin solubility in water at 37 °c. (added weight of drug associated with each of the points in 5 ml of solution is 450 mg and 1250 mg). figure 7. the effect of sedimentation on sodium phenytoin solubility in water at 37 °c (the suspensions were stirred with magnetic-stirrer). (added weight of drug associated with each of the points in 5 ml of solution is 450 mg and 1250 mg, respectively). ali shayanfar et al. admet & dmpk 6(4) (2018) 269-278 276 figure 8. the effect of sedimentation on sodium phenytoin solubility in water at 37 °c (the amount of excess solid was 450 mg/5 ml). figure 9. the effect of sedimentation on lamotrigine solubility in water at 37 °c (the amount of excess solid was 175 mg/5 ml). conclusion solubility is a classic and common physicochemical property of pharmaceuticals; however, its values are affected by various parameters. the findings of this study indicated that the amount of excess solid, especially in ionized form and sedimentation time after appropriate stirring, had significant effects on the obtained solubility of sodium phenytoin and lamotrigine. the findings show that these factors are possible reasons for variation in the reported solubility data in the literature and relatively unsuccessful attempts to develop solubility prediction models. it seems that the proper reporting of experimental details is necessary in solubility studies. acknowledgements this article is a part of the results of s.m’s pharm.d thesis no. 85 registered at faculty of pharmacy, tabriz university of medical sciences, tabriz, iran. a.s. thanks the ministry of health and medical education (grant for young assistant professors), tehran, iran, for financial support. admet & dmpk 6(4) (2018) 269-278 solubility of sodium phenytoin and lamotrigine doi: 10.5599/admet.621 277 references [1] l. di, p.v. fish, t. mano. bridging solubility between drug discovery and development. drug discov.today 17 (2012) 486-495. [2] f. martínez, a. jouyban, w.e. acree. pharmaceuticals solubility is still nowadays widely studied everywhere. pharm. sci. 23 (2017) 1-2. [3] a.t. florence, d. attwood, physicochemical principles of pharmacy, pharmaceutical press, london, 2006. [4] a. jouyban, m.a.a. fakhree, experimental and computational methods pertaining to drug solubility, in: w.e. acree, jr. (ed.) toxicity and drug testing, intech publisher: new york, 2012. [5] j. alsenz, m. kansy. high throughput solubility measurement in drug discovery and development. adv. drug deliv. rev. 59 (2007) 546-567. [6] a. jouyban. review of the cosolvency models for predicting solubility of drugs in water-cosolvent mixtures. j. pharm. pharm. sci. 11 (2008) 32-58. [7] a. avdeef, e. fuguet, a. llinàs, c. ràfols, e. bosch, g. völgyi, t. verbić, e. boldyreva, k. takács-novák. equilibrium solubility measurement of ionizable drugs–consensus recommendations for improving data quality. admet & dmpk 4 (2016) 117-178. [8] e. baka, j.e.a. comer, k. takács-novák. study of equilibrium solubility measurement by saturation shake-flask method using hydrochlorothiazide as model compound. j. pharm. biomed. anal. 46 (2008) 335-341. [9] a. shayanfar, m.a.a. fakhree, a. jouyban. a simple qspr model to predict aqueous solubility of drugs. j. drug deliv. sci. technol. 20 (2010) 467-476. [10] a. ghanbari, y. sarbaz, v. jouyban-gharamaleki, k. jouyban-gharamaleki, j. soleymani, a. jouyban. an improved automated setup for solubility determination of drugs. pharm. sci. 22 (2016) 210-214. [11] s.b. murdande, m.j. pikal, r.m. shanker, r.h. bogner. aqueous solubility of crystalline and amorphous drugs: challenges in measurement. pharm. dev. technol. 16 (2011) 187-200. [12] a. avdeef. suggested improvements for measurement of equilibrium solubility-ph of ionizable drugs. admet & dmpk 3 (2015) 84-109. [13] k. kawakami, k. miyoshi, y. ida. impact of the amount of excess solids on apparent solubility. pharm. res. 22 (2005) 1537-1543. [14] z. wang, l.s. burrell, w.j. lambert. solubility of e2050 at various ph: a case in which apparent solubility is affected by the amount of excess solid. j. pharm. sci. 91 (2002) 1445-1455. [15] a. avdeef. solubility of sparingly-soluble ionizable drugs. adv. drug deliv. rev. 59 (2007) 568-590. [16] a.c. moffat, m.d. osselton, b. widdop, j. watts, clarke's analysis of drugs and poisons, pharmaceutical press, london, 2011. [17] a. mabhoot, a. jouyban. solubility of sodium phenytoin in ethanol + water mixtures at various temperatures. chem. eng. technol. 203 (2016) 1009-1012. [18] a. shayanfar, m.a.a. fakhree, w.e. acree jr, a. jouyban. solubility of lamotrigine, diazepam, and clonazepam in ethanol + water mixtures at 298.15 k. j. chem. eng. data 54 (2009) 1107-1109. [19] k. beattie, g. phadke, j. novakovic, chapter 6 lamotrigine, in: h.g. brittain (ed.) profiles of drug substances, excipients and related methodology, academic press, 2012, pp. 245-285. [20] a. avdeef, absorption and drug development: solubility, permeability, and charge state, john wiley & sons, hoboken, new jersey, 2012. [21] j.t. rubino. solubilities and solid state properties of the sodium salts of drugs. j. pharm. sci. 78 (1989) 485-489. [22] m. mosharraf, t. sebhatu, c. nyström. the effects of disordered structure on the solubility and dissolution rates of some hydrophilic, sparingly soluble drugs. int. j. pharm. 177 (1999) 29-51. ali shayanfar et al. admet & dmpk 6(4) (2018) 269-278 278 [23] f. nielloud, pharmaceutical emulsions and suspensions: revised and expanded, crc press, boca raton, 2000. [24] j. bevernage, j. brouwers, m.e. brewster, p. augustijns. evaluation of gastrointestinal drug supersaturation and precipitation: strategies and issues. int. j. pharm. 453 (2013) 25-35. ©2018 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ the function and regulation of pd-l1 in immunotherapy doi: 10.5599/admet.5.3.442 159 admet & dmpk 5(3) (2017) 159-172; doi: http://dx.doi.org/10.5599/admet.5.3.442 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index review the function and regulation of pd-l1 in immunotherapy libin guo 1 , yao lin 2 and hang fai kwok 1 * 1 faculty of health sciences, university of macau, macau sar 2 college of life sciences, fujian normal university, fuzhou, fujian province, china *corresponding author: e-mail: hfkwok@umac.mo; tel.: +853 8822 4991; fax: +853 8822 4991 received: september 18, 2017; revised: september 27, 2017; published: september 29, 2017 abstract pd-l1, also known as b7-h1, is a type i transmembrane protein, which is expressed in different kinds of tumor cells. it is correlated with poor clinical outcome of patients with various types of tumors. pd -l1 can regulate tumor microenvironment or tumor related immune response through suppressing t cell or nk cell mediated immune response. pd-l1 expression is regulated by various cytokines, such as lps, gm-csf, il-4, tgf-β, tnf-α. pd-1 and pd-l1 are the members of b7 and cd28 superfamily, respectively. the b7/cd28 interaction plays a central role in immune tolerance. pd-l1 can bind to pd-1, which leads to the suppression of lymphocyte activation and apoptosis of lymphocytes. anti-pd-l1 therapy is one of the immunotherapies to treat cancer (especially solid tumor). pd-l1 expression may be associated with efficacy of anti pd-1/pd-l1 therapy. in this review, we will focus on the regulation mechanism of pd-l1 expression, and describe the role of pd-1/pd-l1 binding on the anti-pd-1/pd-l1 therapy. keywords immune checkpoint; pd-l1 expression; signal pathway; regulation mechanism. introduction cancer is the first leading cause of death in china and second leading cause of death in united states in recent years [1, 2]. although, the cancer death rates have declined over 2 decades, the death rates caused by cancers of uterine corpus and liver are still increasing. it is estimated that by 2030 about 13–17 million people will die from cancer [3]. cancer is a major public health issue, and the search for advanced detection methods and novel treatments are important. cancer treatments include various types of therapies, such as surgery, radiation, chemotherapy, biological therapy and targeted therapy [4-8]. among these therapies, immunotherapy for cancer has now become a new standard strategy to treat cancer. with the approval of rituximab and trastuzumab, immunotherapy the field of cancer therapy changed considerably [9]. over recent decades, immunotherapy has been used for solid and hematological malignances treatment [10]. there are several types of immunotherapy, including vaccines, monoclonal antibodies, t cell therapy, oncolytic virus therapy, non-specific immunotherapies [11-15]. as one of approaches of immunotherapy, antibody-based cancer therapy employs monoclonal antibodies targeting receptor tyrosine kinase (rtks), immune checkpoint inhibitor, or other membrane antigens. antibodies targeting rtks (such as human epidermal growth factor receptor 2 (her2) and extracellular signal regulated kinase (erk)) are often limited by resistance. for example, colorectal cancer patients are http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:hfkwok@umac.mo guo, lin and kwok admet & dmpk 5(3) (2017) 159-172 160 resistant to cetuximab and panitumumab (antiepidermal growth factor receptor (egfr) antibodies) when the tumor has a mutated ras-protein [16, 17]. bispecific antibodies (bsab) and homo-combinations of antibodies targeting rtks are employed to improve sensitivity of patients to antibodies[18]. immune checkpoint inhibitor, such as anti-programmed cell death protein 1 (pd-1) antibody, anti-programmed death ligand 1 (pd-l1) antibody and anti-cytotoxic t-lymphocyte-associated protein 4 (ctla-4) antibody, can enhance the ability of cytotoxic t cells to recognize tumor without limitation by resistance [19]. pd-l1 is known as b7 homolog 1 (b7-h1) or cluster of differentiation (cd274) and is recognized as a cell surface protein belonging to b7 superfamily. many studies have reported that pd-l1 can suppress t lymphocyte activation through interacting with some pd-l1 receptor, such as pd-1. pd-l1 directed therapy of cancer many studies showed that pd-l1 expression is present in various kinds of tumors, which suggested that pd-l1 may be closely related with cancer. liu et al. demonstrated that pd-l1 expression was found in multiple myeloma plasma cells, but not in cells which were isolated from monoclonal gammopathy or healthy donors [20]. pd-l1 expression has not only been found on glial tumor cells, but also on various brain metastases [21, 22]. 66 % of freshly isolated head and neck squamous cell carcinoma (hnscc) showed constitutive expression of pd-l1 [23]. except for cancers mentioned above, upregulation of pd-l1 has also been found in other common cancers, such as melanoma, ovarian cancer, lung cancer, urothelial carcinoma, esophageal cancer, cervical cancer, pancreatic cancer and wilms tumor [24-31]. recent clinical data showed that increased expression of pd-l1 was closely related with a poor prognosis in chronic hepatitis b virus (hbv) infected hepatocellular carcinoma patients [32]. in breast cancers it was shown that in this disease a higher expression of pd-l1 had a strong link with several characteristics, such as tumor size, american joint committee on cancer (ajcc) primary tumor classification, tumor grade, lymph node status, and high ki-67 expression [33]. meanwhile, pd-l1 was also shown to be related to poor prognosis in human breast cancer [33]. upregulation of pd-l1 was closely related with late stage clinical development and decreased rate of disease-free survival [34]. sarah et al. reported that pd-l1 was expressed in tumor cells of patients with epstein-barr virus (ebv) infected gastric cancer (gc), but not in other gc cancer cells [35]. besides, preclinical experiments also showed an effect of pd-l1 on tumor treatment. in an acute myeloid leukemia (aml) model, blockade of pd-l1 by antibodies enhanced immune response against leukemia cells (c1498.gfp cells) in c57bl/6 mice [36]. meanwhile, cd8+ t cell receiving anti-pd-l1 antibody was shown to increase in livers from c57bl/6 mice compared to control without treatment with antibodies [36]. iwai et al. showed that in vitro p815 cells transfected with papuroxs-pd-l1 stably expressed pd-l1, and was susceptible to t cell antigen receptor (tcr)-mediated lysis induced by cytotoxic t cells [37]. most importantly, tumorigenesis and invasiveness of these tumor cells was enhanced compared to parental cells which didn’t express pd-l1 [37]. the promoting effect of pd-l1 on cancer was also shown by suppression of anti-pd-l1 antibody on the growth of myeloma cells [37]. pd-l1expression in immune diseases as a target of immune checkpoint inhibitor, pd-l1 is not only associated with anti-cancer treatment, but also with other immune diseases. pd-l1 has a strong link with autoimmune disease. in non-obese diabetic (nod) prediabetic murine model, pd-l1 expression was significantly upregulated on infiltrating mononuclear cells in the islet [38]. mohammed et al. also reported that expression of pd-l1 was found in the inflamed islets of nod mice, suggesting a regulatory role of pd-l1 on the progression of autoimmune admet & dmpk 5(3) (2017) 159-172 the function and regulation of pd-l1 in immunotherapy doi: 10.5599/admet.5.3.442 161 diabetes [39]. pd-l1 has been reported to contribute to etiopathogenetic of autoimmune diseases in mouse models. mohammed et al. reported that blockade of pd-l1 by monoclonal antibodies (mabs) against pd-l1 led to diabetes in female nod mice [39]. the nod mouse model is useful for studying diabetes caused by autoimmune process [40]. blockade of pd-1/pd-l1 interaction didn’t induce diabetes in nor mice, showing the specific islet cell toxicity by mabs treatment [39]. bing et al. reported that in balb/c mice, blockade of pd-l1 remarkably increased incidence of experimental autoimmune encephalomyelitis (eae) in the presence of myelin oligodendrocyte glycoprotein (mog) peptide 35–55 [41]. furthermore, in myelin proteolipid protein (plp) peptide 139–151 immunized b10.s mice, blockade of pd-l1 also significantly enhanced severity of eae [41]. pd-l1 deficiency mice were completely susceptible to eae after e2 (known as 17β-estradiol, an estrogen) treatment [42]. this study also reported that the protective effect against eae mediated by e2 was significantly reduced in recipient mice transfected with pd-l1−/− b-cell [42]. therefore pd-l1 in b cells may be crucial for the protective effect of estrogen against eae. enhanced expression of pd-l1 was present on ductal and acinar epithelial cells in the salivary glands of patients with sjogren’s syndrome [43]. increased expression of pd-l1 has also been found on t cells, b cells as well as monocytes of patients with systemic lupus erythematosus (sle), implying that pd-l1 might be correlated with sle disease [44]. the expression of pd-l1 increased on mononuclear cells in the lamina propria of inflammatory bowel disease (ibd) patients, as well as on mononuclear cells in the lamina propria in experimental colitic mice [45]. pd-l1 in virus associated disease besides, pd-l1 is also associated with chronic viral infection and chronic inflammatory diseases. upregulated expression of pd-l1 was found in livers of patients infected with hbv compared with healthy donors [46]. previous studies showed that most of merkel cell carcinoma (mcc) were linked to merkel cell polyomavirus (mcpyv) infection [47]. enhanced expression of pd-l1 was found in mcpyv infected mcc cells compared to mcpyv-negative mcc cells [48]. it was reported that interferon (ifn)-γ had an effect on increasing the number of cd8+ cytotoxic t lymphocytes (ctls) when infected by the virus [49]. blockade of pd-l1 enhanced the ability of cd8+ ctls to produce ifn-γ through an increase in the number of ifn-γproducing hbv-specific ctls [50]. daniel et al. reported that in lymphocytic choriomeningitis virus (lcmv) infected mice blockade of pd-l1 and pd-1 interaction restored the ability of exhausted cd8 t cell to kill infected cells [51]. the pd-l1/pd-1 pathway has also been found to mediate the function of virus-specific exhausted cd8+ t cell in human immunodeficiency virus (hiv) infected patients in several studies [51]. dai et al. showed that blockade of interaction of pd-1/pd-l1 by anti-pd-l1 restored immune response generated by cd8+ t cell after being infected by hiv [52]. anti-pd-l1 antibody also enhanced proliferation and cytokine production of hepatitis c virus (hcv)-specific cd8 cells during chronic hcv infection in the study reported by simona et al. [53]. the result of these studies suggested that pd-l1 can inhibit the immune response during chronic viral infection [54]. pd-l1 was involved in intestinal mucosal inflammation, such as ulcerative colitis. in severe combined immune-deficient (scid) mice, which were reconstituted with cd45rbhigh cd4+ t cells, blockade of pd-l1 by mabs against b7-h1 suppressed wasting disease with colitis [45]. role of pd-l1 in efficacy of treatment pre-clinical data showed that the anticancer effect of anti-pd-l1 antibody is different kinds of cancer cells. meanwhile, clinical activity of anti-pd-l1 antibodies has been observed in various malignancies, such as melanoma, non-small cell lung cancer, squamous head and neck cancer, microsatellite-unstable guo, lin and kwok admet & dmpk 5(3) (2017) 159-172 162 colorectal cancer, and other types of cancers [55-58]. a predictive role of pd-l1 expression and til (tumor infiltrating lymphocytes) has been found in lung cancer patients receiving anti-pd-1/pd-l1 immunotherapy and could be used to improve clinical interpretations [59]. however, pd-l1 expression can be constitutive and inducible. induced pd-l1 upregulation is regulated by various other cytokines, such as ifn, tumor necrosis factor (tnf)-α and toll-like receptor (tlr) etc. [60]. most importantly, pd-l1 upregulation can also improve survival of mice with lung tumor, which received anti-pd-l1 treatment [61]. this review will discuss the basic structure of pd-l1, which is important for pd-l1 regulation and therapy. besides, this review will also discuss regulation and signaling pathway of pd-l1, which can influence anti-pd-l1 immunotherapy. structure of pd-l1 pd-l1 was found by searching the expressed sequence tag database (generated from dendritic cells and activated macrophages) for molecules which contain homology to b7-1 and b7-2 [62]. pd-l1 is encoded by the cd274 gene, which is situated on mouse chromosome 19 and human chromosome 9 at a band p24. pd-l1 is a 40kda protein containing 290 amino acid. human pd-l1 shares 70 % homology in amino acids with mouse pd-l1 [63]. mazanet et al. reported that the promoter region of cd274 gene consisted several elements responded to ifn-γ, which was necessary for upregulation of pd-l1 expression mediated by ifn-γ [64]. pd-l1 exerts a role as a type i transmembrane protein and contains four domains [56]. they are ig (immunoglobulin) v-like domain, ig c-like domain, and hydrophobic transmembrane domain, as well as cytoplasmic domains, encoded by single exon sequences [62]. the schematic diagram and structure of pdl1 is shown in figure 1. the ig v-like domain and ig c-like domain are two anti-parallel β sandwich immunoglobulin superfamily (ig sf) domains, which are related to domains of immunoglobulins [65, 66]. the ig v-like domain is formed by bed and agfcc’c’’ β sheets, and necessary for interaction of pd-l1 and b7-1 in murine [67]. what’s more, the pd-l1:pd-1 binding interface is also on its ig v-like domain. the ig c domain of pd-l1 has c1-set domains with β-strands forming abed and cfg sheets [66]. the cytoplasmic domain which is encoded by the last exon is about 30 amino acids, and highly conserved in all species reported [23]. while a potential site of pd-l1 which could be phosphorylated by protein kinase c (pkc) is on its intracellular domain [23]. regulation of pd-l1 pd-l1 expression is present on antigen-presenting cells (apcs), such as human monocytes, as well as activated human and murine dendritic cells. moreover, pd-l1 is also expressed in nonlymphoid tissues such as heart and lung, thymus, kidney. pd-l1 is easily induced by pro-inflammatory cytokines on different kinds of cells, which is shown in figure 2. pd-l1 can be induced by ifn-γ and tlr ligands in mm plasma cells [20]. this study revealed that blockade of mitogen-activated protein kinase (mapk)/erk pathway and inhibition of signal transducer and activator of transcription 1（stat1）suppressed the enhanced expression of pd-l1 induced by ifn-γ [20]. meanwhile, inhibition of the myeloid differentiation primary response gene 88（myd88）and tnf receptor-associated factor 6（traf6）has been reported to inhibit pd-l1 expression either induced by tlr ligands or by ifn-γ [20]. in human lung cancer cells interferon regulatory factor-1 (irf-1) was necessary for induction of pd-l1 expression mediated by ifn-γ in stat1-manner [68]. besides, a recent study also showed that enhanced expression of pd-l1 induced by tlr also was dependent on il-, il-10 and stat3 in apcs [69]. in addition, ifn-α has also been reported to induce upregulated mrna expression and protein admet & dmpk 5(3) (2017) 159-172 the function and regulation of pd-l1 in immunotherapy doi: 10.5599/admet.5.3.442 163 expression of pd-l1 on microvascular endothelial cells (ecs) [70]. moreover the enhanced expression of pd-l1 protein and mrna was induced by ifn-β on monocytes and mature dendritic cells of healthy donors and multiple sclerosis (ms) patients [71]. figure 1. schematic diagram and structure of pd-l1 protein. a. schematic diagram of pd-l1 protein. sp: signal peptide; ecd: extracellular domain; tm: transmembrane domain; icd; intracellular domain. b. structure of pd-l1 protein. this structure shows two main anti-parallel β sandwich immunoglobulin superfamily (ig sf) domains of pd=l1 protein. inflammatory macrophages upregulate expression of pd-l1, which was also induced by lipopolysaccharide (lps) [72]. enhanced expression of pd-l1 was present on immature dendritic cells (idcs) after being treated with granulocyte-macrophage colony-stimulating factor (gm-csf) [73]. ou et al. reported that in sle monocytes opposing actions of tnf-α and transforming growth factor (tgf)-β regulated the expression of pd-l1 [74]. pd-l1 expression could be revived by exogenous tnf-α on lupus monocytes [74]. tgf-β has shown to inhibit pd-l1 expression on monocytes [74]. induction of pd-l1 expression is also associated with janus kinase (jak)/stat signaling and activator protein 1 (ap1) activity [75]. michael et al. showed that in classical hodgkin lymphoma reed–sternberg cells constitutive activation of ap-1 resulted in the binding of ap-1 components to enhancer of cd274 gene, which enhanced the activity of pd-l1 promoter [75]. the induction of pd-l1 expression by jak2 was also observed in this study, which was associated with jak/stat-dependent promoter in cd274 gene [75]. besides, the increase of pdl1 expression protein was induced by inhibition of phosphatase and tensin homolog (pten) in glioma cancer and colorectal cancer [22]. recent studies reported that pd-l1 expression was inhibited by p53 through regulating mir-34a and mir-200 [76, 77]. the il family also was indicated to be important for induction of pd-l1 expression. for instance, il-27 enhanced the expression of cd274 gene which is dependent on stat1 on naïve t cells [78]. further study guo, lin and kwok admet & dmpk 5(3) (2017) 159-172 164 showed that pd-l1 expression was increased which was mediated by il-4 in nuclear factor-kappa b (nf-κb) p50–/– p65+/– dendritic cell (dc) [38]. the pd-l1 expression was not only induced by multiple pro-inflammatory molecules, but also influenced by t cells. for example, enhanced expression of pd-l1 induced by type 1 t helper (th1) cells was also found on different macrophages [72]. activation of human t cells by anti-cd3 and anti-cd28 may lead to increased expression of pd-l1 mrna [79]. figure 2. pd-l1 expression is regulated by several cytokines. a. upregulation of pd-l1 is induced by various pro-inflammatory molecules. pd-l1 could be induced by ifn-γ and tlr ligands via regulation of downstream adaptor proteins of the tlr signaling pathway. ifn-γ also induced pd-l1 expression through jak-stat1 signaling pathway. except that, some other cytokines, like gm-csf, tnf-alpha, il , can increase pd-l1 expression. b. inhibition of pd-l1 expression is mediated by p53, pten and tgf-β. pd-l1/pd-1 signaling pathway several studies reported that pd-l1 can interact with pd-1, and this interaction delivered an inhibitory signals to regulate immune tolerance and immunopathology [80]. most importantly, pd-l1/pd-1 signaling may exert its inhibitory effect on immune response through signaling pathway mediated by various types of cytokines, such as shp-1, tcr, phosphoinositide 3-kinase (pi3k). figure 3 shows pd-l1/pd-1 signaling pathway regulating t cell survival. the effector t cells interaction of pd-l1 and pd-1 blocked tcr signal transduction, leading to inhibition of t cell cytotoxic activity [81]. the src homology region 2 domain containing phosphatase-1 (shp1) could regulate activation of cd8+ t cell and inhibition of shp-1 by sodium stibogluconate enhanced the function of t cells [82]. the cytoplasmic domain of pd-1 consists of two motifs, including the tyrosine-based inhibitory motif (itim) and immunoreceptor tyrosine-based switch motif (itsm) [83]. the pd-1 and pd-l1 interaction can phosphorylate a tyrosine in the itsm motif of pd-1, and then recruit the shp-1 and shp2 to the itsm motif [84]. after recruiting, this signaling pathway can inhibit stop signals, and block the interaction of t cells and dendritic cells [85]. finally, blockade of the tcr signal transduction caused inhibition of the pi3k/akt (akt is known as protein kinase b) and mapk signaling [86]. most importantly, inhibition of pi3k activation suppresses the expression of b cell lymphoma extra-large (bcl-xl) and activation of akt, which furtherly leads to increased apoptosis of t cells [12]. pd-l1 not only inhibits function of activated t cell via inhibition of pi3k/akt pathway and ras/mek/erk pathway, but also through inhibition of transcription factors necessary for t cell survival. pd-l1 and pd-1 interaction was reported to inhibit expression of gata-3 and t-bet [87]. gata-3 has been reported to be a admet & dmpk 5(3) (2017) 159-172 the function and regulation of pd-l1 in immunotherapy doi: 10.5599/admet.5.3.442 165 transcription factor which is critical for differentiation of t helper 2 (th2) cells [21]. meanwhile, t-bet, which is known as t-box transcription factor, can contribute to t-cell development [88]. pd-l1/pd-1 interaction is also involved in signal pathway mediated by tgf-β. loise et al. reported that the function of tgf-β in induced t cell regulatory (itreg) cell development could be reduced by the loss of pd-l1 [89]. this study also reported that during the conversion of itreg cells from mature t cells, pd-l1/pd1 signaling reduced the phosphorylation of akt and its downstream substrates mtor and s6 [89]. barber et al. reported that in exhausted cd8+ t cell blockade of pd-l1 and pd-1 signaling pathway enhanced t cell ability to secrete cytokine and kill infected cells with lcmv, suggesting that the pd-l1/pd-1 signaling pathway affected activation of downstream molecules of the t cell activation [51]. the study reported by carter et al. showed that pd-l1 and pd-1 interaction inhibited il-2 production in cd4+ and cd8+ t cell, which further inhibited lymphoproliferation [90]. in endothelial cells mouse anti-pd-l1 blocked the interaction of pd-l1 and pd-l, and stimulated ifn-γ production secreted by cd8+ t cell, resulting in t cell activation [91]. in hodgkin lymphoma -infiltrating t cells, inhibition of the activation of pd-1 and pd-l1 signaling pathway led to inhibition of shp-2 phosphorylation and increased the production of ifn-γ [92]. in the liver pd-l1 and pd-1 interaction led to reduced ifn-γ production secreted by cd8+ t lymphocytes(clt) [50]. figure 3. the pd-l1/pd-1 signaling pathway in t cells. pd-l1/pd1 signaling pathway could mediate t cell survival through regulation of different signal transductions. pd-1 and pd-l1 interaction leads to pd-1 phosphorylation, and recruits shp-1 and shp2 to the itsm motif of pd-1. after recruiting, this signaling pathway blocks the interaction of t cells. blockade of tcr signal transduction inhibited the pi3k/akt and mapk signaling. inhibition of pi3k activation suppressed the expression of b cell lymphoma extra-large (bcl-xl) and activation of akt, which leads to increased apoptosis of t cells. pd-l1 and pd-1 interaction also inhibits expression of gata-3 and t-bet. gata-3 is critical for differentiation of th2 cells. t-bet contributes to t-cell development. conclusions this review describes the regulation of the pd-l1/pd-1 signaling pathway. pd-l1 is a transmembrane protein, and inhibits the function of t cell through binding to its receptor pd-1. the interaction of pd-l1 and pd-1 inhibits t cell activation and proliferation, which inhibits the function of t cells to produce cytokines and kill targeting tumor cells. blockade of pd-l1 signaling by anti-pd-l1 antibodies not only guo, lin and kwok admet & dmpk 5(3) (2017) 159-172 166 inhibits tumor growth, but also leads to etiopathogenetic of autoimmune diseases (such as diabetes). there are several kinds of antibodies in development against pd-l1 in cancer immunotherapy, such as atezolizumab, avelumab, durvalumab, bms-936559. the safety study with bms-936659 showed that in 39% patient side effects such as rash, hypothyroidism, diabetes mellitus, myasthenia gravis and other disease occurred [93]. this makes pd-l1 an attractive target for anti-cancer therapy. however, knowledge on the side effects of anti-pd-l1 therapy is still limiting and, more studies are needed to get a better in the regulation of pd-l1 on cancer. acknowledgements: this work was supported by the science and technology development fund of macau sar (fdct) [023/2015/amj] and the international s&t cooperation program of china [2016yfe0121900]. miss libin guo was in receipt of phd studentship from the faculty health sciences university of macau and fdct. references [1] z. lian, y. xie, y. lu, d. huang, h. shi. trends in the major causes of death in china, 1982-2010, chin med j (engl) 127 (2014) 777-781. [2] r.l. siegel, k.d. miller, a. jemal. cancer statistics, 2017, ca cancer j clin 67 (2017) 7-30. [3] l. yang, l. wang, y. zhang. immunotherapy for lung cancer: advances and prospects, am j clin exp immunol 5 (2016) 1-20. [4] g. lezoche, a.m. paganini, r. campagnacci, r. ghiselli, m. pelloni, a. rombini, m. guerrieri. treatment of rectal cancer by transanal endoscopic microsurgery: review of the literature, minerva chir 68 (2013) 1-9. [5] a. munshi, j.p. agarwal, k.c. pandey. cancer patients with cardiac pacemakers needing radiation treatment: a systematic review, j cancer res ther 9 (2013) 193-198. [6] m. johnson. chemotherapy treatment decision making by professionals and older patients with cancer: a narrative review of the literature, eur j cancer care (engl) 21 (2012) 3-9. [7] a.m. manganoni, c. zane, l. pavoni, c. farisoglio, e. sereni, p. calzavara-pinton. cutaneous melanoma in patients in treatment with biological therapy: review of the literature and case report, dermatol online j 17 (2011) 12. [8] m. kudo. targeted therapy for liver cancer: updated review in 2012, curr cancer drug targets 12 (2012) 1062-1072. [9] d.g. maloney, a.j. grillolopez, c.a. white, d. bodkin, r.j. schilder, j.a. neidhart, n. janakiraman, k.a. foon, t.m. liles, b.k. dallaire, k. wey, i. royston, t. davis, r. levy. idec-c2b8 (rituximab) anticd20 monoclonal antibody therapy in patients with relapsed low-grade non-hodgkin's lymphoma, blood 90 (1997) 2188-2195. [10] i. mellman, g. coukos, g. dranoff. cancer immunotherapy comes of age, nature 480 (2011) 480489. [11] c. guo, m.h. manjili, j.r. subjeck, d. sarkar, p.b. fisher, x.y. wang. therapeutic cancer vaccines: past, present, and future, adv cancer res 119 (2013) 421-475. [12] m. themeli, i. riviere, m. sadelain. new cell sources for t cell engineering and adoptive immunotherapy, cell stem cell 16 (2015) 357-366. [13] s.j. russell, k.w. peng, j.c. bell. oncolytic virotherapy, nat biotechnol 30 (2012) 658-670. [14] l. galluzzi, e. vacchelli, j.m. bravo-san pedro, a. buque, l. senovilla, e.e. baracco, n. bloy, f. castoldi, j.p. abastado, p. agostinis, r.n. apte, f. aranda, m. ayyoub, p. beckhove, j.y. blay, l. bracci, a. caignard, c. castelli, f. cavallo, e. celis, v. cerundolo, a. clayton, m.p. colombo, l. coussens, m.v. dhodapkar, a.m. eggermont, d.t. fearon, w.h. fridman, j. fucikova, d.i. gabrilovich, j. galon, a. garg, f. ghiringhelli, g. giaccone, e. gilboa, s. gnjatic, a. hoos, a. admet & dmpk 5(3) (2017) 159-172 the function and regulation of pd-l1 in immunotherapy doi: 10.5599/admet.5.3.442 167 hosmalin, d. jager, p. kalinski, k. karre, o. kepp, r. kiessling, j.m. kirkwood, e. klein, a. knuth, c.e. lewis, r. liblau, m.t. lotze, e. lugli, j.p. mach, f. mattei, d. mavilio, i. melero, c.j. melief, e.a. mittendorf, l. moretta, a. odunsi, h. okada, a.k. palucka, m.e. peter, k.j. pienta, a. porgador, g.c. prendergast, g.a. rabinovich, n.p. restifo, n. rizvi, c. sautes-fridman, h. schreiber, b. seliger, h. shiku, b. silva-santos, m.j. smyth, d.e. speiser, r. spisek, p.k. srivastava, j.e. talmadge, e. tartour, s.h. van der burg, b.j. van den eynde, r. vile, h. wagner, j.s. weber, t.l. whiteside, j.d. wolchok, l. zitvogel, w. zou, g. kroemer. classification of current anticancer immunotherapies, oncotarget 5 (2014) 12472-12508. [15] s. carvalho, f. levi-schaffer, m. sela, y. yarden. immunotherapy of cancer: from monoclonal to oligoclonal cocktails of anti-cancer antibodies: iuphar review 18, brit j pharmacol 173 (2016) 1407-1424. [16] r.g. amado, m. wolf, m. peeters, e. van cutsem, s. siena, d.j. freeman, t. juan, r. sikorski, s. suggs, r. radinsky, s.d. patterson, d.d. chang. wild-type kras is required for panitumumab efficacy in patients with metastatic colorectal cancer, j clin oncol 26 (2008) 1626-1634. [17] j.r. thompson, s.p. menon, g.k. dy. tyrosine kinase inhibitors for egfr-and alk-mutated nonsmall cell lung cancer, admet and dmpk 4 (2016) 186-211. [18] s.j. demarest, k. hariharan, j. dong. emerging antibody combinations in oncology, mabs 3 (2011) 338-351. [19] e.a. hawkes, a. grigg, g. chong. programmed cell death-1 inhibition in lymphoma, lancet oncol 16 (2015) e234-e245. [20] j. liu, a. hamrouni, d. wolowiec, v. coiteux, k. kuliczkowski, d. hetuin, a. saudemont, b. quesnel. plasma cells from multiple myeloma patients express b7-h1 (pd-l1) and increase expression after stimulation with ifn-{gamma} and tlr ligands via a myd88-, traf6-, and mek-dependent pathway, blood 110 (2007) 296-304. [21] j.f. jacobs, a.j. idema, k.f. bol, s. nierkens, o.m. grauer, p. wesseling, j.a. grotenhuis, p.m. hoogerbrugge, i.j. de vries, g.j. adema. regulatory t cells and the pd-l1/pd-1 pathway mediate immune suppression in malignant human brain tumors, neuro oncol 11 (2009) 394-402. [22] a.t. parsa, j.s. waldron, a. panner, c.a. crane, i.f. parney, j.j. barry, k.e. cachola, j.c. murray, t. tihan, m.c. jensen, p.s. mischel, d. stokoe, r.o. pieper. loss of tumor suppressor pten function increases b7-h1 expression and immunoresistance in glioma, nat med 13 (2007) 84-88. [23] s.e. strome, h. dong, h. tamura, s.g. voss, d.b. flies, k. tamada, d. salomao, j. cheville, f. hirano, w. lin, j.l. kasperbauer, k.v. ballman, l. chen. b7-h1 blockade augments adoptive t-cell immunotherapy for squamous cell carcinoma, cancer res 63 (2003) 6501-6505. [24] j.c. routh, r.a. ashley, t.j. sebo, c.m. lohse, d.a. husmann, s.a. kramer, e.d. kwon. b7-h1 expression in wilms tumor: correlation with tumor biology and disease recurrence, j urology 179 (2008) 1954-1959. [25] r. hino, k. kabashima, y. kato, h. yagi, m. nakamura, t. honjo, t. okazaki, y. tokura. tumor cell expression of programmed cell death-1 ligand 1 is a prognostic factor for malignant melanoma, cancer-am cancer soc 116 (2010) 1757-1766. [26] j. hamanishi, m. mandai, m. iwasaki, t. okazaki, y. tanaka, k. yamaguchi, t. higuchi, h. yagi, k. takakura, n. minato, t. honjo, s. fujii. programmed cell death 1 ligand 1 and tumor-infiltrating cd8+ t lymphocytes are prognostic factors of human ovarian cancer, proc natl acad sci u s a 104 (2007) 3360-3365. [27] j. konishi, k. yamazaki, m. azuma, i. kinoshita, h. dosaka-akita, m. nishimura. b7-h1 expression on non-small cell lung cancer cells and its relationship with tumor-infiltrating lymphocytes and their pd-1 expression, clin cancer res 10 (2004) 5094-5100. [28] j. nakanishi, y. wada, k. matsumoto, m. azuma, k. kikuchi, s. ueda. overexpression of b7-h1 (pdl1) significantly associates with tumor grade and postoperative prognosis in human urothelial cancers, cancer immunol immun 56 (2007) 1173-1182. guo, lin and kwok admet & dmpk 5(3) (2017) 159-172 168 [29] y. ohigashi, m. sho, y. yamada, y. tsurui, k. hamada, n. ikeda, t. mizuno, r. yoriki, h. kashizuka, k. yane, f. tsushima, n. otsuki, h. yagita, m. azuma, y. nakajima. clinical significance of programmed death-1 ligand-1 and programmed death-1 ligand-2 expression in human esophageal cancer, clin cancer res 11 (2005) 2947-2953. [30] r. karim, e.s. jordanova, s.j. piersma, g.g. kenter, l. chen, j.m. boer, c.j. melief, s.h. van der burg. tumor-expressed b7-h1 and b7-dc in relation to pd-1+ t-cell infiltration and survival of patients with cervical carcinoma, clin cancer res 15 (2009) 6341-6347. [31] t. nomi, m. sho, t. akahori, k. hamada, a. kubo, h. kanehiro, s. nakamura, k. enomoto, h. yagita, m. azuma, y. nakajima. clinical significance and therapeutic potential of the programmed death-1 ligand/programmed death-1 pathway in human pancreatic cancer, clin cancer res 13 (2007) 21512157. [32] z. zeng, f. shi, l. zhou, m.n. zhang, y. chen, x.j. chang, y.y. lu, w.l. bai, j.h. qu, c.p. wang, h. wang, m. lou, f.s. wang, j.y. lv, y.p. yang. upregulation of circulating pd-l1/pd-1 is associated with poor post-cryoablation prognosis in patients with hbv-related hepatocellular carcinoma, plos one 6 (2011). [33] s. muenst, a.r. schaerli, f. gao, s. daster, e. trella, r.a. droeser, m.g. muraro, p. zajac, r. zanetti, w.e. gillanders, w.p. weber, s.d. soysal. expression of programmed death ligand 1 (pd-l1) is associated with poor prognosis in human breast cancer, breast cancer res tr 146 (2014) 15-24. [34] c.t. wu, w.c. chen, y.h. chang, w.y. lin, m.f. chen. the role of pd-l1 in the radiation response and clinical outcome for bladder cancer, sci rep 6 (2016) 19740. [35] s. derks, x.y. liao, a.m. chiaravalli, x.s. xu, m.c. camargo, e. solcia, f. sessa, t. fleitas, g.j. freeman, s.j. rodig, c.s. rabkin, a.j. bass. abundant pd-l1 expression in epstein-barr virusinfected gastric cancers, oncotarget 7 (2016) 32925-32932. [36] l. zhang, t.f. gajewski, j. kline. pd-1/pd-l1 interactions inhibit antitumor immune responses in a murine acute myeloid leukemia model, blood 114 (2009) 1545-1552. [37] y. iwai, m. ishida, y. tanaka, t. okazaki, t. honjo, n. minato. involvement of pd-l1 on tumor cells in the escape from host immune system and tumor immunotherapy by pd-l1 blockade, proc natl acad sci u s a 99 (2002) 12293-12297. [38] s.c. liang, y.e. latchman, j.e. buhlmann, m.f. tomczak, b.h. horwitz, g.j. freeman, a.h. sharpe. regulation of pd-1, pd-l1, and pd-l2 expression during normal and autoimmune responses, eur j immunol 33 (2003) 2706-2716. [39] m.j. ansari, a.d. salama, t. chitnis, r.n. smith, h. yagita, h. akiba, t. yamazaki, m. azuma, h. iwai, s.j. khoury, h. auchincloss, jr., m.h. sayegh. the programmed death-1 (pd-1) pathway regulates autoimmune diabetes in nonobese diabetic (nod) mice, j exp med 198 (2003) 63-69. [40] t.l. delovitch, b. singh. the nonobese diabetic mouse as a model of autoimmune diabetes: immune dysregulation gets the nod, immunity 7 (1997) 727-738. [41] b. zhu, i. guleria, a. khosroshahi, t. chitnis, j. imitola, m. azuma, h. yagita, m.h. sayegh, s.j. khoury. differential role of programmed death-ligand 1 [corrected] and programmed death-ligand 2 [corrected] in regulating the susceptibility and chronic progression of experimental autoimmune encephalomyelitis, j immunol 176 (2006) 3480-3489. [42] s. bodhankar, d. galipeau, a.a. vandenbark, h. offner. pd-1 interaction with pd-l1 but not pd-l2 on b-cells mediates protective effects of estrogen against eae, j clin cell immunol 4 (2013) 143. [43] m. kobayashi, s. kawano, s. hatachi, c. kurimoto, t. okazaki, y. iwai, t. honjo, y. tanaka, n. minato, t. komori, s. maeda, s. kumagai. enhanced expression of programmed death-1 (pd-1)/pdl1 in salivary glands of patients with sjogren's syndrome, j rheumatol 32 (2005) 2156-2163. [44] m.f. liu, c.t. weng, m.y. weng. variable increased expression of program death-1 and program death-1 ligands on peripheral mononuclear cells is not impaired in patients with systemic lupus erythematosus, j biomed biotechnol (2009). [45] t. kanai, t. totsuka, k. uraushihara, s. makita, t. nakamura, k. koganei, t. fukushima, h. akiba, h. yagita, k. okumura, t. machida, h. iwai, m. azuma, l.p. chen, m. watanabe. blockade of b7-h1 admet & dmpk 5(3) (2017) 159-172 the function and regulation of pd-l1 in immunotherapy doi: 10.5599/admet.5.3.442 169 suppresses intestinal inflammation the development of chronic, journal of immunology 171 (2003) 4156-4163. [46] z. xie, y. chen, s. zhao, z. yang, x. yao, s. guo, c. yang, l. fei, x. zeng, b. ni, y. wu. intrahepatic pd1/pd-l1 up-regulation closely correlates with inflammation and virus replication in patients with chronic hbv infection, immunol invest 38 (2009) 624-638. [47] a. kassem, a. schopflin, c. diaz, w. weyers, e. stickeler, m. werner, a. zur hausen. frequent detection of merkel cell polyomavirus in human merkel cell carcinomas and identification of a unique deletion in the vp1 gene, cancer res 68 (2008) 5009-5013. [48] e.j. lipson, j.g. vincent, m. loyo, l.t. kagohara, b.s. luber, h. wang, h. xu, s.k. nayar, t.s. wang, d. sidransky, r.a. anders, s.l. topalian, j.m. taube. pd-l1 expression in the merkel cell carcinoma microenvironment: association with inflammation, merkel cell polyomavirus and overall survival, cancer immunol res 1 (2013) 54-63. [49] j.k. whitmire, j.t. tan, j.l. whitton. interferon-gamma acts directly on cd8(+) t cells to increase their abundance during virus infection, journal of experimental medicine 201 (2005) 1053-1059. [50] h. maier, m. isogawa, g.j. freeman, f.v. chisari. pd-1:pd-l1 interactions contribute to the functional suppression of virus-specific cd8+ t lymphocytes in the liver, j immunol 178 (2007) 2714-2720. [51] d.l. barber, e.j. wherry, d. masopust, b. zhu, j.p. allison, a.h. sharpe, g.j. freeman, r. ahmed. restoring function in exhausted cd8 t cells during chronic viral infection, nature 439 (2006) 682687. [52] l. trautmann, l. janbazian, n. chomont, e.a. said, s. gimmig, b. bessette, m.r. boulassel, e. delwart, h. sepulveda, r.s. balderas, j.p. routy, e.k. haddad, r.p. sekaly. upregulation of pd-1 expression on hiv-specific cd8(+) t cells leads to reversible immune dysfunction (vol 12, 1198, 2006), nature medicine 12 (2006) 1329-1329. [53] s. urbani, b. amadei, d. tola, g. pedrazzi, l. sacchelli, m.c. cavallo, a. orlandini, g. missale, c. ferrari. restoration of hcv-specific t cell functions by pd-1/pd-l1 blockade in hcv infection: effect of viremia levels and antiviral treatment, j hepatol 48 (2008) 548-558. [54] b.b. dai, l. xiao, p.d. bryson, j.x. fang, p. wang. pd-1/pd-l1 blockade can enhance hiv-1 gagspecific t cell immunity elicited by dendritic cell-directed lentiviral vaccines, mol ther 20 (2012) 1800-1809. [55] s.y. liu, y.l. wu. ongoing clinical trials of pd-1 and pd-l1 inhibitors for lung cancer in china, j hematol oncol 10 (2017) 136. [56] a.i. daud, k. loo, m.l. pauli, r. sanchez-rodriguez, p.m. sandoval, k. taravati, k. tsai, a. nosrati, l. nardo, m.d. alvarado, a.p. algazi, m.h. pampaloni, i.v. lobach, j. hwang, r.h. pierce, i.k. gratz, m.f. krummel, m.d. rosenblum. tumor immune profiling predicts response to anti-pd-1 therapy in human melanoma, j clin invest 126 (2016) 3447-3452. [57] n.h. segal, s.h.i. ou, a.s. balmanoukian, m.g. fury, e. massarelli, j.r. brahmer, j. weiss, p. schoffski, s.j. antonia, c. massard, d.p. zandberg, s. khleif, x. li, m. rebelatto, k. steele, p.b. robbins, j.a. blake-haskins, m.o. butler. safety and efficacy of medi4736, an anti-pd-l1 antibody, in patients from a squamous cell carcinoma of the head and neck (scchn) expansion cohort., j clin oncol 33 (2015). [58] n.j. llosa, m. cruise, a. tam, e.c. wicks, e.m. hechenbleikner, j.m. taube, r.l. blosser, h.n. fan, h. wang, b.s. luber, m. zhang, n. papadopoulos, k.w. kinzler, b. vogelstein, c.l. sears, r.a. anders, d.m. pardoll, f. housseau. the vigorous immune microenvironment of microsatellite instable colon cancer is balanced by multiple counter-inhibitory checkpoints, cancer discov 5 (2015) 4351. [59] d. liu, s. wang, w. bindeman. clinical applications of pd-l1 bioassays for cancer immunotherapy, j hematol oncol 10 (2017) 110. [60] j. chen, c.c. jiang, l. jin, x.d. zhang. regulation of pd-l1: a novel role of pro-survival signalling in cancer, ann oncol 27 (2016) 409-416. guo, lin and kwok admet & dmpk 5(3) (2017) 159-172 170 [61] n. chen, w.f. fang, j.h. zhan, s.d. hong, y.n. tang, s.y. kang, y.x. zhang, x.b. he, t. zhou, t. qin, y. huang, x.p. yi, l. zhang. upregulation of pd-l1 by egfr activation mediates the immune escape in egfr-driven nsclc implication for optional immune targeted therapy for nsclc patients with egfr mutation, j thorac oncol 10 (2015) 910-923. [62] h. dong, g. zhu, k. tamada, l. chen. b7-h1, a third member of the b7 family, co-stimulates t-cell proliferation and interleukin-10 secretion, nat med 5 (1999) 1365-1369. [63] g.j. freeman, a.j. long, y. iwai, k. bourque, t. chernova, h. nishimura, l.j. fitz, n. malenkovich, t. okazaki, m.c. byrne, h.f. horton, l. fouser, l. carter, v. ling, m.r. bowman, b.m. carreno, m. collins, c.r. wood, t. honjo. engagement of the pd-1 immunoinhibitory receptor by a novel b7 family member leads to negative regulation of lymphocyte activation, j exp med 192 (2000) 10271034. [64] m.m. mazanet, c.c.w. hughes. b7-h1 is expressed by human endothelial cells and suppresses t cell cytokine synthesis, journal of immunology 169 (2002) 3581-3588. [65] m. collins, v. ling, b.m. carreno. the b7 family of immune-regulatory ligands, genome biol 6 (2005) 223. [66] y. chen, p. liu, f. gao, h. cheng, j. qi, g.f. gao. a dimeric structure of pd-l1: functional units or evolutionary relics?, protein cell 1 (2010) 153-160. [67] m.j. butte, m.e. keir, t.b. phamduy, a.h. sharpe, g.j. freeman. programmed death-1 ligand 1 interacts specifically with the b7-1 costimulatory molecule to inhibit t cell responses, immunity 27 (2007) 111-122. [68] s.j. lee, b.c. jang, s.w. lee, y.i. yang, s.i. suh, y.m. park, s. oh, j.g. shin, s. yao, l.p. chen, i.h. choi. interferon regulatory factor-1 is prerequisite to the constitutive expression and ifn-gammainduced upregulation of b7-h1 (cd274), febs lett 580 (2006) 755-762. [69] s.j. wolfle, j. strebovsky, h. bartz, a. sahr, c. arnold, c. kaiser, a.h. dalpke, k. heeg. pd-l1 expression on tolerogenic apcs is controlled by stat-3, european journal of immunology 41 (2011) 413-424. [70] m.j. eppihimer, j. gunn, g.j. freeman, e.a. greenfield, t. chernova, j. erickson, j.p. leonard. expression and regulation of the pd-l1 immunoinhibitory molecule on microvascular endothelial cells, microcirculation 9 (2002) 133-145. [71] b. schreiner, m. mitsdoerffer, b.c. kieseier, l. chen, h.p. hartung, m. weller, h. wiendl. interferonbeta enhances monocyte and dendritic cell expression of b7-h1 (pd-l1), a strong inhibitor of autologous t-cell activation: relevance for the immune modulatory effect in multiple sclerosis, j neuroimmunol 155 (2004) 172-182. [72] p. loke, j.p. allison. pd-l1 and pd-l2 are differentially regulated by th1 and th2 cells, proc natl acad sci u s a 100 (2003) 5336-5341. [73] j.a. brown, d.m. dorfman, f.r. ma, e.l. sullivan, o. munoz, c.r. wood, e.a. greenfield, g.j. freeman. blockade of programmed death-1 ligands on dendritic cells enhances t cell activation and cytokine production, j immunol 170 (2003) 1257-1266. [74] j.n. ou, a.e. wiedeman, a.m. stevens. tnf-alpha and tgf-beta counter-regulate pd-l1 expression on monocytes in systemic lupus erythematosus, sci rep 2 (2012) 295. [75] m.r. green, s. rodig, p. juszczynski, j. ouyang, p. sinha, e. o'donnel. constitutive ap-1 activity and ebv infection induce pd-l1 in hodgkin lymphomas and posttransplant lymphoproliferative disorders: implications for targeted therapy (vol 18, pg 1611, 2012), clin cancer res 18 (2012) 2117-2117. [76] t. kim, a. veronese, f. pichiorri, t.j. lee, y.j. jeon, s. volinia, p. pineau, a. marchio, j. palatini, s.s. suh, h. alder, c.g. liu, a. dejean, c.m. croce. p53 regulates epithelial-mesenchymal transition through micrornas targeting zeb1 and zeb2, journal of experimental medicine 208 (2011) 875883. admet & dmpk 5(3) (2017) 159-172 the function and regulation of pd-l1 in immunotherapy doi: 10.5599/admet.5.3.442 171 [77] x. wang, j.g. li, k. dong, f. lin, m. long, y.r. ouyang, j.x. wei, x. chen, y.y. weng, t. he, h.z. zhang. tumor suppressor mir-34a targets pd-l1 and functions as a potential immunotherapeutic target in acute myeloid leukemia, cell signal 27 (2015) 443-452. [78] k. hirahara, k. ghoreschi, x.p. yang, h. takahashi, a. laurence, g. vahedi, g. sciume, a.o. hall, c.d. dupont, l.m. francisco, q. chen, m. tanaka, y. kanno, h.w. sun, a.h. sharpe, c.a. hunter, j.j. o'shea. interleukin-27 priming of t cells controls il-17 production in trans via induction of the ligand pd-l1, immunity 36 (2012) 1017-1030. [79] f. bennett, d. luxenberg, v. ling, i.m. wang, k. marquette, d. lowe, n. khan, g. veldman, k.a. jacobs, v.e. valge-archer, m. collins, b.m. carreno. program death-1 engagement upon tcr activation has distinct effects on costimulation and cytokine-driven proliferation: attenuation of icos, il-4, and il-21, but not cd28, il-7, and il-15 responses, journal of immunology 170 (2003) 711-718. [80] m.e. keir, m.j. butte, g.j. freeman, a.h. sharpel. pd-1 and its ligands in tolerance and immunity, annu rev immunol 26 (2008) 677-704. [81] k. karwacz, f. arce, c. bricogne, g. kochan, d. escors. pd-l1 co-stimulation, ligand-induced tcr down-modulation and anti-tumor immunotherapy, oncoimmunology 1 (2012) 86-88. [82] m. hebeisen, l. baitsch, d. presotto, p. baumgaertner, p. romero, o. michielin, d.e. speiser, n. rufer. shp-1 phosphatase activity counteracts increased t cell receptor affinity, journal of clinical investigation 123 (2013) 1044-1056. [83] l.m. shlapatska, s.v. mikhalap, a.g. berdova, o.m. zelensky, t.j. yun, k.e. nichols, e.a. clark, s.p. sidorenko. cd150 association with either the sh2-containing inositol phosphatase or the sh2containing protein tyrosine phosphatase is regulated by the adaptor protein sh2d1a, journal of immunology 166 (2001) 5480-5487. [84] j.m. chemnitz, r.v. parry, k.e. nichols, c.h. june, j.l. riley. shp-1 and shp-2 associate with immunoreceptor tyrosine-based switch motif of programmed death 1 upon primary human t cell stimulation, but only receptor ligation prevents t cell activation, journal of immunology 173 (2004) 945-954. [85] b.t. fife, k.e. pauken, t.n. eagar, t. obu, j. wu, q.z. tang, m. azuma, m.f. krummel, j.a. bluestone. interactions between pd-1 and pd-l1 promote tolerance by blocking the tcr-induced stop signal, nat immunol 10 (2009) 1185-u1170. [86] n. patsoukis, j. brown, v. petkova, f. liu, l. li, v.a. boussiotis. selective effects of pd-1 on akt and ras pathways regulate molecular components of the cell cycle and inhibit t cell proliferation, sci signal 5 (2012) ra46. [87] r. nurieva, s. thomas, t. nguyen, n. martin-orozco, y. wang, m.k. kaja, x.z. yu, c. dong. t-cell tolerance or function is determined by combinatorial costimulatory signals, embo j 25 (2006) 2623-2633. [88] d.m. dorfman, e.s. hwang, a. shahsafaei, l.h. glimcher. t-bet, a t cell-associated transcription factor, is expressed in hodgkin's lymphoma, hum pathol 36 (2005) 10-15. [89] l.m. francisco, v.h. salinas, k.e. brown, v.k. vanguri, g.j. freeman, v.k. kuchroo, a.h. sharpe. pdl1 regulates the development, maintenance, and function of induced regulatory t cells, journal of experimental medicine 206 (2009) 3015-3029. [90] l.l. carter, l.a. fouser, j. jussif, l. fitz, b. deng, c.r. wood, m. collins, t. honjo, g.j. freeman, b.m. carreno. pd-1 : pd-l inhibitory pathway affects both cd4(+)and cd8(+) t cells and is overcome by il-2, european journal of immunology 32 (2002) 634-643. [91] n. rodig, t. ryan, j.a. allen, h. pang, n. grabie, t. chernova, e.a. greenfield, s.c. liang, a.h. sharpe, a.h. lichtman, g.j. freeman. endothelial expression of pd-l1 and pd-l2 down-regulates cd8+ t cell activation and cytolysis, eur j immunol 33 (2003) 3117-3126. [92] r. yamamoto, m. nishikori, t. kitawaki, t. sakai, m. hishizawa, m. tashima, t. kondo, k. ohmori, m. kurata, t. hayashi, t. uchiyama. pd-1-pd-1 ligand interaction contributes to immunosuppressive microenvironment of hodgkin lymphoma, blood 111 (2008) 3220-3224. guo, lin and kwok admet & dmpk 5(3) (2017) 159-172 172 [93] j.r. brahmer, s.s. tykodi, l.q. chow, w.j. hwu, s.l. topalian, p. hwu, c.g. drake, l.h. camacho, j. kauh, k. odunsi, h.c. pitot, o. hamid, s. bhatia, r. martins, k. eaton, s. chen, t.m. salay, s. alaparthy, j.f. grosso, a.j. korman, s.m. parker, s. agrawal, s.m. goldberg, d.m. pardoll, a. gupta, j.m. wigginton. safety and activity of anti-pd-l1 antibody in patients with advanced cancer, n engl j med 366 (2012) 2455-2465. ©2017 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ leveraging chromatography based physicochemical properties for efficient drug design doi: 10.5599/admet.529 85 admet & dmpk 6(2) (2018) 85-104; doi: http://dx.doi.org/10.5599/admet.529 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index review leveraging chromatography based physicochemical properties for efficient drug design gilles h. goetz*, marina shalaeva molecular properties group, worldwide medicinal chemistry, pfizer worldwide research & development, pfizer, inc., groton, connecticut 06340, united states *corresponding author e-mail: gilles.h.goetz@pfizer.com; tel.: +1-860-715-6311 received: march 28, 2018; revised: may 26, 2018; available online: june 04, 2018 abstract applications of chromatography derived lipophilicity, polarity, and 3d concepts such as conformational states, exposed polarity and intramolecular hydrogen bonds (imhb), are discussed along with recently developed methods for incorporating these concepts into drug design strategies. in addition, the drug design process is described with examples and practices used at pfizer, as well as experimental and computed parameters used for parallel optimization of properties leading to drug candidate nominations. keywords elogd, epsa; lipophilicity, exposed polarity introduction an active molecule with favorable pharmacokinetic/pharmacodynamic (pk/pd) properties is the goal of drug design. this means a biologically active molecule can be developed into a drug only if it fits strict pharmacological profile requirements. tablets are the preferred form of administration due to convenience, patient compliance and ease of manufacturing. therefore most drug design projects are focused on creating drugs suitable for oral dosage. in order to be absorbed in the gastrointestinal (gi) tract, a solid formulation must be ingested, solubilized. only then can the solvated drug be absorbed into the blood stream. here drug distribution, metabolism and excretion affect its clearance, defining the pharmacokinetic parameters, half-life and therapeutic index (ti). balancing absorption, distribution, metabolism, excretion and toxicity (admet) properties is a major challenge in drug discovery and development. failure to achieve the right admet balance was cited in 1997 as the reason for the attrition of 40% of drug candidates during development [1]. here we present and discuss recent advances and approaches, mostly developed at pfizer, using chromatographically determined physicochemical properties for the optimization of admet properties. lipophilicity lipophilicity (as defined by iupac) represents the affinity of a molecule or a moiety for a lipophilic environment. it is commonly measured by its distribution behavior in a biphasic system, either liquid-liquid http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:gilles.h.goetz@pfizer.com g.h. goetz and m. shalaeva admet & dmpk 6(2) (2018) 85-104 86 (partition coefficient between 1-octanol and water) or solid-liquid (retention on reversed phase high performance liquid chromatography (rp hplc)) [2]: lipophilicity may be expressed as log p for neutral molecules and log d for ionized molecules at a particular ph (log dph), where permeation p and distribution d are defined by the neutral and ionized species concentration ratios:     octanol water neutral neutral p      octanol water ionized + neutral ionized + neutral d  . successful drug designs require careful balancing of in vivo and in vitro properties. lipophilicity is a descriptor intrinsic to many of these properties. computational models of admet properties generally include log p as a significant descriptor of the modelled parameter. an extensive analysis of the literature linking admet and lipophilicity in drug discovery was published by waring [3]. it details important, but obscure aspects of the log d parameter. for example, many studies rely on compilations of log p/log d values from a variety of sources with poorly described experimental details. these non-homogeneous datasets lead to discrepancies in analyses. there is also confusion between log p and log d; terms often used interchangeably. for drug compounds ionized at physiological ph, log dph is more relevant than log p, however, it deviates because of its dependence on pka. this is especially flagrant, in case of predicted log dph , due to unreliable pka predictions. the widely cited publication by mannhold et al. [4] compares several computed log p models to over 96,000 measured values. unfortunately, actual experimental details were not provided. many research organizations have developed their own methods for lipophilicity measurements. they use their particular values for drug design and optimization, i.e. chromatographic hydrophobicity index (chi) at glaxosmithkline [5], and elog d at pfizer [6]. early stage drug design strives to create the most active molecule for a particular therapeutic target. the ultimate goal is the discovery of a molecule possessing “drug-like” properties. the reliance on lipophilicity to understand molecules' drug-like properties was defined in 1997 lipinski's “rule of five” (ro5) [7]. now it is ingrained into drug design that oral drugs require a clogp < 5. rules, depending mostly on log d, help guide designs towards optimal drug-like chemistry space. log d is used to calculate ligand efficiency (le) [8,9], metabolic efficiency (mete) [3,10,11], central nervous system multiparameter optimization (cns mpo) [12,13], and toxicity scores [14-16]. recent developments are reviewed in the application of log d for balancing potency and adme properties. lipophilicity and ligand efficiency the review by hopkins et al. [17] concluded that recently marketed oral drugs “frequently have highly optimized ligand efficiency values for their targets”. early stage design efforts often lead to series of hundreds of newly synthesized compounds. these structures need to be rapidly evaluated for projects to proceed. promising leads may be prioritized using lipophilic ligand efficiency (lle) [9] or lipe [10]: lle (or lipe) = pic50 – clogp . the lle/lipe concept has been developed following the introduction of ligand efficiency (le) [8]. the le parameter allows direct comparison of molecules through their target binding energy. le can be expressed admet & dmpk 6(2) (2018) 85-104 leveraging chromatography physchem properties in drug design doi: 10.5599/admet.529 87 by the following simplified equation in which n is the number of non-hydrogen atoms: le = 1.37(-log ic50)/n . project progress may be tracked using both le and lipe; lipe is expected to reach values of 6 or 7 for a potential drug candidate to be chosen for further development [18-20]. target binding increases with the addition of a lipophilic fragment, at the expense of potential for offtarget binding and decreased solubility. an efficient drug design requires potency to be achieved with a minimal increase in mw or lipophilicity [21-24]. lipophilicity and oral absorption oral administration, the most desirable route of medicine delivery, requires a drug to be absorbed while in the gi tract, with cell membrane permeation being the major determinant of absorption. there are several processes involved in crossing the intestinal mucosa, including active transport [15], paracellular transport, and drug efflux. passive membrane diffusion is considered to be the major contributing process [25]. recent analyses continue to demonstrate the significance of the non-specific permeation mechanisms governed by molecular properties (namely molecular size and polarity) [16]. the favorable property space for an orally administered drugs is defined by the afore mentioned ro5 [26,27]. in order to achieve absorption, an oral drug needs to have a clogp < 5; mw < 500; hydrogen bond donor (hbd) count (nh+oh) < 5; and a hydrogen bond acceptor (hba) count (n+o) < 10. more specific rules were developed for central nervous system (cns) projects, where crossing the blood-brain barrier is challenging due to tighter junctions and higher efflux. the cns mpo score introduced by wager et al. [12] was derived by analysis of drugs and candidates with regard to their ability to cross the blood-brain barrier. the cns mpo includes ionization constant (pka), log p, and log d7.4. this emphasizes the importance of ionization and lipophilicity adjusted for ionization [13,23,28,29]. recently z. ranković further refined the cns desirability criteria in cns mpov.2 [30]. lipophilicity measurements also help to estimate solubility in discovery settings [31]; solubility decreases with log p > 3. current trend towards and beyond ro5 compounds recently, the introduction of innovative “flat” targets and the potential for higher potency and selectivity provided by larger molecules, has led to a renewed interest in peptide-based drugs. these fill the gap between small molecules and biotherapeutics, as illustrated in figure 1 [32,33]. the exploration of new targets required new design concepts including “an update” to ro5. bunnage et al. defined "four pillars of target validation”, the first pillar being “exposure at the site of action”, defined largely by permeability, where tactics such as increasing lipophilicity could help improve both potency and permeability [34]. failing one or more of the ro5 may decrease oral bioavailability, due to low solubility, and or poor membrane permeability, in addition to metabolic instability, synthetic complexity and other disadvantages of large drug molecules. the study by guimarães et al. [35] highlighted the importance of 3-d properties for passive permeation of beyond ro5 (bro5) molecules, where taking molecular conformation into account, as well as size and polarity assessments, are needed to aid design in bro5 space. g.h. goetz and m. shalaeva admet & dmpk 6(2) (2018) 85-104 88 figure 1. schematic illustration of the molecular weight (mw) gap between conventional small molecule drugs (< 500) and biologics (> 5000). reprinted from [32] copyright (2013), with permission from wiley. the ro5 defines properties of drug-like molecules, essentially, via molecular size and polarity, where mw describes the size, hbd and hba counts describe polarity and log d describes both. notably, mw and nh, oh, o and n are just simple counts and only roughly describe interactions with the environment, especially for molecules of any complexity. for example, intramolecular hydrogen bonding (imhb) in large, flexible molecules could lead to shielding of hbd and hba and, consequently, to changes in molecular shape and size, which are not accounted for by mw or hbd and hba counts. these simple counts, including polar surface area (psa), are indifferent to changes in molecular shape and size, the major determinants of passive diffusion [35]. experimental log d, on the other hand, accounts for molecular size and polarity in solution and for 3d conformational changes adopted by the molecule. therefore, the “extension” of the ro5 to large molecules depends even more on the “effectiveness” of the log d parameter. while the accuracy of log d prediction is important for large, flexible molecules, it is also much more difficult to predict, as it deals with the assessment of multiple conformations and their populations. measured log d values are necessary to build and test such computational models for bro5 compounds. however, log d values > 3 are challenging to obtain via conventional shake-flask log d (sflog d), due to limits in quantitation for the concentration of above 1000 between phases. kihlberg et al. analyzed orally available drugs in bro5 space [36] and observed significant extension, almost doubling in values of ro5 parameters (figure 2). they proposed an “extended ro5 space” (mw < 700, 0 < clogp < 7.5, hbd < 5, psa < 200 å 2 , and the number of rotatable bonds (nrotb) < 20. they also proposed a substantially larger ‘‘possible to be oral” bro5 space with the limits for oral bioavailability extended to approximately mw < 1000, -2 < clogp < 10, hbd < 6, hba < 15, psa < 250 å 2 , nrotb < 20 (dashed box). notably, hbd has hardly changed (increased by only 1 count) highlighting the significance of limiting hbd exposure for oral absorption of large molecules. admet & dmpk 6(2) (2018) 85-104 leveraging chromatography physchem properties in drug design doi: 10.5599/admet.529 89 figure 2. physicochemical property space of drugs and clinical candidates with mw > 500. solid box marks “extended ro5” space. dashed box marks “bro5” space. reprinted from [36] copyright (2014), with permission from elsevier. wang et al. [37] observed a correlation of rp hplc log k’ with passive permeability determined by human colorectal adenocarcinoma cells (caco-2) and the parallel artificial membrane permeability assay (pampa) for a set of macrocyclic peptides. bockus et al. used elog d (5.5 to 7.5) to study macrocycles. the lipophilicity values correlated with cell-based permeability values by madin-darby canine kidney (mdck)low efflux cells (mdck-le) [38]. lokey et al. investigated ro5 and bro5 compounds with mw > 1000 and demonstrated a sharp decline in apparent passive permeability for compounds with molecular size above 750 å [39]. they also concluded that bulk physical properties contributing to passive permeability could be approximated by lipophilicity and molecular size (log khydrocarbon/water and mw respectively). molecular size relates to passive permeability with pka and aqueous solubility also very important properties to consider. the lipophilicity parameter used by lokey et al. (log khydrocarbon/water) is lipophilicity measured in a non-polar environment and is different from log doctanol/water. it has been shown to correlate well with passive permeability measured by pampa and mdck-le. the use of δlog p = log palkane/water – log poctanol/water was first introduced by seiler [40] and applied to improve absorption [41], as a guide in the design of novel brain-penetrating h2 antagonists [42], or as a measure of hbd acidity [43,44]. hydrogen bonds characteristically feature binding energies and contact distances that can lead to large variations even for a single donor-acceptor pair [45]. the 96-well plate-based shake-flask log ptoluene/water method (ph 1 to 11) was developed along with the imhb interpretation scheme based on δlog p [46]. this method allows verification of molecular conformations predicted by cosmo-rs software [47] which describes the geometry of virtual molecules' interactions in both polar and non-polar environments. furthermore, an rp hplc method using a polystyrene-divinylbenzene stationary phase (plrp-s) was developed to simulate a non-polar lipidic membrane environment [48], and to obtain experimental lipophilicity values. kihlberg et al described errors in clogp; log poct predictions that were far from the experimentally determined values [36]. posaconazole has a clogp of 5.4, while the experimentally determined log p is 2.4 [49]. however, in this particular case, the measured value may be incorrect. these studies reiterate the need for an extended range of lipophilicity measurements for bro5 compounds that would accommodate the large lipophilic molecules designed for current targets. they highlight the importance of lipophilicity as an approximation of molecular size and conformations in specific environments. g.h. goetz and m. shalaeva admet & dmpk 6(2) (2018) 85-104 90 lipophilicity and clearance avoiding highly lipophilic compounds is a design principle used to improve clearance [50]. the analysis of matching pairs conducted by stepan et al. [11] confirmed that changes in metabolic stability largely come from changes in lipophilicity (figure 3). to determine how structural changes affect the relative clearance between analogues, the lipophilic metabolism efficiency parameter (lipmete) was introduced. lipmete relates lipophilicity to the in vitro metabolic clearance measured by the human liver microsomes (hlm) assay. lipmete and unbound clearance (clint,u) are defined as follows: lipmete = log d7.4 − log (clint,u) , clint,u = clint,app/fu,mic . figure 3. the plot of unbound clearance, clint,u, versus experimental log d, elog d. the 45° lines represent different values of lipmete. compounds that parallel the same 45° line offer the same ratio of metabolic clearance to lipophilicity. reprinted with permission from [11]. copyright (2013) american chemical society. stepan et al. [11] suggested choosing high lipmete compounds as a starting point for optimization due to their wider logd range. thus, potency and permeability could be improved while maintaining low clearance. lipe and lipmete have become complementary opposites (“yin and yang”) in medicinal chemistry decision making at pfizer and the authors provide specific guidance on simultaneously optimizing these lipophilicity driven parameters. optimization of adme properties, such as cell membrane permeability and metabolic stability, often comes to an act of balancing these orthogonal parameters. gleeson [14] describes the contribution of mw, ionization state, and clogp to in vivo clearance. log clearance (log cl) was largest for a clogp > 5, less pronounced for clogp 3–5, and most favorable for a clogp < 3. johnson et al. [16] analyzed permeability and clearance data for 47,018 pfizer compounds and observed broad trends for favorable clearance and permeability compared to mw and log d. using lipophilicity determination, they identified a “golden triangle” of mw and log d values for optimal permeability and metabolic stability. figure 4 shows permeable and stable compounds in blue and compounds failing permeability or metabolic stability in grey against mw and log d. compounds with favorable clearance and permeability properties are clustered within this “golden triangle” of mw and log d. admet & dmpk 6(2) (2018) 85-104 leveraging chromatography physchem properties in drug design doi: 10.5599/admet.529 91 figure 4. combined in vitro permeability and clearance trends across mw and log d. reprinted from [16] copyright (2009), with permission from elsevier. lipophilicity and toxicity avoiding high lipophilicity is also critical for reducing the probability of adverse safety findings [3, 51]. hughes et al. [52] demonstrated that lipophilic compounds (clogp > 3) with a low polar surface area (psa < 75 å 2 ) have a 6-fold greater risk of toxicity findings in preclinical toxicology studies [53]. an alternative to using log p, which could oversimplify in vivo behavior of compounds, was proposed by wenlock [54]. this criterion is based on the amount of compound in the body at a steady state. the relationship of this criterion at an acceptable human dose of 0.5 mg/ kg for 242 oral drugs with different in vivo plasma clearances was established with regard to their safety profiles. experimental log d determinations the optimal range of log d to satisfy admet properties centers on 2 (+/-1) [3], but expands far beyond that range for some marketed drugs [36]. the wide range of lipophilicity employed in modern drug design requires experimental methods and computational models to support exploitation of these chemical spaces. the chromatography based elog d method [6, 55] has been developed at pfizer to alleviate limitations the of the classical shake-flask technique and to extend to compounds bro5. since its introduction, it has been used successfully in a traditional chemistry space [22, 56-60] as well as in the bro5 programs [38]. the simultaneous optimization of potency and clearance using both measured and calculated elog d has been described for the takeda-g-protein-receptor-5 (tgr5) program seeking an orally available compound for improved glycemic control via glucose-dependent insulin secretion. this has led to significantly improved clearance (hlm < 100 ), log d = 2, and equivalent potency compared to competitor compounds with high clearance hlm > 200 and log d > 2 [50]. lipophilicity (elog d) has also been utilized in clinical studies to understand the specific and non-specific binding of an active ingredient to the beta-amyloid plaques in alzheimer’s disease patients which has been observed with high contrast positron emission tomography (pet) imaging [61]. at pfizer, log d at ph 7.4 is routinely measured for most compounds using either sflog d or elog d or both methods; each method has limitations. while elog d cannot be applied to acids, sflog d has a limitation in measuring values above 3.5 due to the increasing errors in quantification of concentration g.h. goetz and m. shalaeva admet & dmpk 6(2) (2018) 85-104 92 extremes (indeed, log d = 4 implies that the concentration in 1-octanol is 10,000 times higher than in the aqueous phase). additionally, an accurate prediction of log d at ph = 7.4 is still an issue for novel structures and therefore comparisons of sflog d and elog d values are often useful. with project x, compounds were evaluated using sflog d (and occasionally elog d), but significant discrepancies were observed for “low” sflog d samples (figure 5. (a)). however, if elog d – sflog d differences (residuals) were plotted against elog d it became apparent that the largest discrepancies were in the range above 3.0. (figure 5(b)). this illustrates how the sflog d procedure may underestimate lipophilicity values above 3. figure 5. (a) comparison of sflog d and elog d values for project x; (b) sflog d elog d residuals vs. elog d in summary, if analyzed by elog d, both sflog d and elog d give, as expected, very similar values (within +0.5) in the elog d < 3 range. however, in the elog d > 3 range, many sflog d values are under-evaluated (by up to 2 log units). therefore, it is important to recognize and respect the experimental methods applicability domain when using log d for sar or building computational models based on such data. computed log d day-to-day experience in measuring log d demonstrated that despite great efforts to improve log d calculations, there were still instances where log d predictions are inaccurate; especially if ionization is involved and an accurate pka is needed to adjust the calculated log p of a neutral form to a log d at a particular ph, usually 7.4. the cpflogd model built at pfizer takes into account both sflog d and elog d experimental values while respecting the limits of each method, i.e. sflog d < 3 and elog d for bases and neutrals only. consequently, the model gives preference to sflog d values in the log d range below 3 and to elog d values in the log d range above 3. the cpflogd model is built using the cubist non-linear in-silico regression methodology and it is regularly updated with the latest experimental data, thus providing improved cpflogd predictions for the next design cycle. performance of the cpflogd model is demonstrated in figure 6 on a rather challenging subset of diverse “natural-product” like compounds. this subset of about 300 molecules is more complicated than encountered in most small molecule projects. it encompasses mw from 200 to 2000, has compounds violating up to four ro5, with high nrotb, hbd and hba. for many compounds in the log d > 4 range a sflog d value could not be acquired due to extremely low concentration in the aqueous phase and therefore elog d values were measured on all compounds, excluding acids. admet & dmpk 6(2) (2018) 85-104 leveraging chromatography physchem properties in drug design doi: 10.5599/admet.529 93 figure 6. (a) measured elog d vs. cpflogd; (b) measured sflog d values vs. cpflogd. a diverse subset of “natural product”-like library, about 300 compounds. the size of the markers represents molecular weight (193 to 2013 da). the colours represent violations of the ro5. as expected, cpflogd predicted sflog d more accurately in the < 2 range and, elog d more accurately in the > 2 range. several outliers in the upper left corner of the elog d vs. cpflogd figure 6(a) represent the underestimated sflogd results used by the cpflogd model leading to underestimated prediction values. comparison of cpflogd to commercially available acdlabs log d7.4 predictions for the same subset are shown in figure 7, where many compounds violating three ro5 criteria (yellow markers) fell far outside cpflogd, especially in the low lipophilicity range. figure 7. lipophilicity calculations using log d7.4 by acd labs vs. cpflogd for a subset of a “natural product”like library. it should be emphasized that project teams usually design around an active lead molecule and work on a few series, where “diversity” in a general chemistry sense is rather limited. therefore, measured log d values determined on a few compounds in a series introduced into the model often allow significant improvement in accuracy of cpflogd predictions in that chemical space. polarity in the otherwise comprehensive iupac compendium of chemical terminology, also known as the gold book, (unlike lipophilicity, hydrophobicity and hydrophilicity), the compound polarity is not defined [2]. the only reference to polarity in the gold book is coupled with the concept of solvent rather than solute (see below). g.h. goetz and m. shalaeva admet & dmpk 6(2) (2018) 85-104 94 when applied to solvents, the term polarity covers their overall solvation capability (solvation power) for solutes (i.e. in chemical equilibria: reactants and products, in reaction rates: reactants and activated complex, in light absorptions: ions or molecules in the ground and excited state). a solvent’s solvation power depends on the action of all possible intermolecular interactions between solute ions or molecules and solvent molecules. those interactions can be both nonspecific and specific but exclude the interactions leading to definite chemical alterations of the ions or molecules of the solute. tpsa in the context of drug design, the polarity of a molecule has been redefined as its polar surface area (psa). originally computed using molecular mechanics calculations, the fragments of the dynamic van der waals’ surface area associated with oxygen, nitrogen and their attached hydrogen atoms were considered as the polar portions of the molecular surface area. this showed correlation to cell permeability in caco-2 cells for a homologous series of beta-adrenoreceptor antagonists [62]. the popularity of psa has recently increased significantly, due to its performance in predicting intestinal adsorption as well as blood-brain barrier penetration [63-70]. in 2000 ertl et al. introduced the topological polar surface area (tpsa)[71], an index based on the addition of tabulated surface contributions of polar fragments. based on ~35.000 drug-like molecules, the surface contributions of 43 fragments centered on polar oxygen, nitrogen, phosphorus and sulfur were determined by the least square fit and tabulated. this methodology proved 2-3 orders of magnitude less computationally intensive than the previous psa calculations and gave similar results [71]. the use of psa in medicinal research has since been reviewed by ertl [72]. in addition to dynamic, molecular and topological polar surface areas, quantum mechanical polar surface area (qmpsa) was recently introduced and showed a good correlation with the fraction absorbed (fa) after oral administration for a set of 18 drugs when carboxyl groups were deprotonated, suggesting adsorption to be strongly related to polar interactions of molecules in water solution [73]. it should be noted that significant increases in computing resources are needed to generate qmpsa, compared to the table-entry recalling tpsa. while mostly used for barrier crossing prediction in an adme context, tpsa has also been recently investigated as a descriptor for 2d-qsar for diverse pharmacological activity data [74], as well as for active drug transport by multidrug resistance associated protein 1 (mrp1) [75]. the first study pointing to a hard psa limit dates back to 1997, in which palm et al. [65] showed on a diverse set of 20 model drugs, that fully absorbed drugs (fa > 90 %) had a psa ≤ 60 å 2 while drugs that are less than 10 % absorbed had a psa > 140 å 2 . later, veber at all [76] showed that a psa ≤ 140 å 2 and the number of rotatable bonds ≤ 10 is as efficient and selective a criterion as the lipinski’s ro5 for selecting oral bioavailability of > 20-40 %. as far as the blood-brain barrier was concerned, it was found that the upper limit for psa for a molecule to penetrate the brain was around 90 å 2 [69,77]. tpsa in mpo in 2010 wager at al introduced the cns mpo (multiparameter optimization) desirability tool, which incorporated tpsa along with 5 other fundamental physicochemical properties: clogp, clogd, mw, pka, and hbd. a monotonic decreasing function was used for clogp, clogd, mw, pka, and hbd, and a hump function was used to define tpsa (figure 8). the cns mpo desirability method is quite simple, with parameters derived from medicinal chemistry best practices, and it is able to balance multiple variables admet & dmpk 6(2) (2018) 85-104 leveraging chromatography physchem properties in drug design doi: 10.5599/admet.529 95 while avoiding hard cut-offs. it demonstrated a good correlation between a high score (> 4 out of a possible max of 6) and good in vitro adme properties [12,78]. figure 8. components of cns multiparameter optimization desirability tool (mpo) epsa having access to tpsa, a simple, inexpensive and efficient way to assess polarity from a 2d structure, which correlates reasonably well with cell permeability, the need for a polarity measurement was never really felt by drug design professionals. while historically relevant for the targets pursued in the late 1990s, criteria such as the ro5 or the veber psa-rotatable bonds rule [76], tend to lose traction due to the rarity of easy targets, resulting in drug discovery programs having to foray into more challenging chemical spaces. in jürgen drews’ words “one truth is there for all to see: many ‘easy’ targets or molecules have been found and developed” [79]. over the time, simplification of the polar surface area calculations down to tpsa may have led to the false equivalency of certain considerations such as the actual position of a polar group in a molecule with regard to its immediate and adjacent environment. for example, the tpsa values for ortho-, metaand paraanalogues of any aromatic polar compound are identical, and similarly, the tpsa values of any regioisomers are also identical, due to the nature of the tpsa calculation, solely based on adding fragment contributions to overall polarity. we contend here that each polar fragment does not contribute equally to the polar surface area of a molecule; in essence advocating a return to a more “3d-psa”-like descriptor [35,62]. utilizing separation sciences expertise, a team of pfizer scientists developed an assay which outputs a polarity readout termed epsa (not an acronym) [80]. the main driver behind epsa was to have access to a fast, robust method capable of identifying the potential for a compound to form intramolecular hydrogen bonds (imhb) enabling it to hide polarity that might otherwise reduce its passive permeability. indeed, the identification of compounds likely to form imhbs is an important drug design consideration given the correlation of intramolecular hydrogen bonding with increased membrane permeability [81-85]. the epsa method provides, under controlled supercritical fluid chromatography (sfc) conditions, an exposed/hidden polarity readout that is derived from the retention time of a compound on a specific column. the stationary phase, (phenomenex chirex 3014) was selected for its balance of lipophilic and polar attributes and its capacity to separate compounds with wide polarity differences. sfc, which essentially uses normal phase-like conditions, provides an environment with a low dielectric constant conducive to imhb formation. polar compounds are retained more under these conditions, and a low-slope gradient of methanol achieves elution based on the increasing polarity of the mobile phase. in the cases that have been studied, using matched molecular pairs with or without imhbs, compounds with imhb resulted in a significant reduction in polarity and eluted earlier in the chromatogram. results are normalized through the use of calibration standards generating a linear relationship between retention times and epsa values [80]. this is possible to confirm or disprove the presence of an intramolecular hydrogen bond (imhb) in small g.h. goetz and m. shalaeva admet & dmpk 6(2) (2018) 85-104 96 molecules or peptides using epsa and pairwise analysis. such imhbs are a critical factor affecting the conformation of molecules and can have a direct impact on their absorption into the human body, their permeability into target cells, and even their potency against therapeutic targets. this method has been successfully used in a small molecule medicinal chemistry project to assess exposed polarities of their lead compounds. cheng et al. [86] determined epsa for their mechanistic target of rapamycin (mtor), phosphatidylinositol-3 kinase (pi3k) dual inhibitors series and found no direct correlation between epsa and tpsa. however, they showed that the formation of imhb can significantly reduce the effective polar surface area. suggesting that for designs with high tpsa, the introduction of an imhb can help achieve good permeability and cellular potency by modulating the epsa to a desirable range. indeed, several compounds, with tpsa greater than 135 å 2 , exhibit epsa values between 82 and 101, which is a range capable for this series of yielding permeability and cellular potency. wakenhut et al. [87] determined epsa for non-structural protein 5a (ns5a) inhibitors and notably compared two analogues: 12 and 13. compound 12 has less hbd and hba, less rotatable bonds, lower tpsa (105 å 2 ) and, by all standard medicinal chemistry principles, 12 was expected to be more permeable than 13 (tpsa 175 å 2 ). it turned out that the opposite was true; compound 12 was not permeable while 13 was permeable. compound 12 had an exposed polarity value (epsa) of 128, while the epsa value of 13 was 103, significantly lower than for 12. these facts can be explained by the peripheral functionality capable of forming imhbs in 13 within a membrane environment, thereby masking h-bonding (hbd and hba) character and contributing to passive permeability, as compared with the more extended conformations available to compounds such as 12 incapables of forming imhbs (figure 9). the team based its subsequent designs on enabling imhb formation within the analogues being developed. figure 9. representation of compound 12 and 13. compound 13 initially in solution phase (6 hba, 4 hbd, psa = 175 å 2 ) and as it traverses a hydrophobic membrane (4 hba, 2 hbd, 3d-psa~120 å 2 ) as a possible explanation of the good membrane permeability of 13. adapted from [87] copyright (2014), with permission from wiley. epsa and peptides once epsa was established as a viable polarity monitoring tool for small molecules, its remit was expanded to peptides, notably cyclic peptides, which have generally demonstrated insufficient permeability to be used as oral drugs. guiding the design of such peptides without a reliable permeability monitoring method is a challenge. since for peptides, the main obstacle to permeability is polarity, the epsa experiment was a perfect fit. admet & dmpk 6(2) (2018) 85-104 leveraging chromatography physchem properties in drug design doi: 10.5599/admet.529 97 goetz at al. [88] have reported a project team design strategy illustrating the suitability of epsa as a surrogate for cyclic peptide permeability. while working towards improving permeability of their lead compounds, the team was faced with a flat structure-permeability relationship (no significant improvement in permeability as measured by rrck cells was detected with any modification). since the lead compound had an epsa value in the high 160s, reducing measured polarity became the surrogate objective. epsa measurements drove each new design. plans that increased epsa were abandoned in favor of those decreasing epsa. with each design cycle it was observed that a gradual decrease in polarity was obtained but without any significant improvement in permeability as measured by rrck, until a threshold in epsa was reached (epsa < 100) upon which permeability became measurable and increased to reach acceptable levels (epsa < 80). at pfizer, and across multiple academic collaborations [38,89-91], project teams routinely use epsa data to inform peptide design as a predictor of permeability and as an indicator of imhb patterns in cyclic peptides. furthermore, epsa is now used in a prospective way after the development of an in-house computational model [92]. it was built using the following procedure. compounds undergo a thorough conformational analysis (macro model) followed by 3d descriptors calculation (volsurf+) based on molecular interaction fields (mifs). a partial least squares (pls) regression algorithm is then used to establish a quantitative relationship between the matrix of descriptors and the epsa endpoint. implementation of the computational epsa prediction model, combined with other physicochemical properties, has enabled virtual compound design, selection of better candidates for synthesis and has accelerated efficiency of design cycles. figure 10. illustration of polarity based design cycle enabled by the pfizer computational epsa prediction model in the particular example depicted in figure 10, the computational epsa prediction model was successfully implemented by the project team, first designing virtual compound libraries, and then testing their epsa predictions, as a permeability surrogate. only compounds with predicted epsa values < 95 were prioritized for synthesis. compounds with reasonable potency and measured epsa < 100 were prioritized for cell-based passive permeability evaluation and hlm stability assessment. learnings from this data the next design cycle was refined [93]. the epsa technique and its application are not reserved or limited to pfizer and its collaborators, indeed, epsa has been implemented across the pharmaceutical industry, i.e. at abbvie [94], novartis [95], and merck [96], epsa services are available through an established analytical cro [97]. additionally, peter wipf in his january 2017 editorial [98] selected epsa as one of the 6 technical innovations published in acs medicinal chemistry letters over the past 3 years that facilitate the practice of medicinal chemistry and are likely to become essential items in the medicinal chemist’s toolbox. g.h. goetz and m. shalaeva admet & dmpk 6(2) (2018) 85-104 98 conclusions across the pharmaceutical industry, each r&d department has its own philosophy. based on historical approaches, resource allocations and leadership advocacy, drug design strategies essentially follow the same guidelines with regards to lead optimization: exquisite potency along with favorable admet profiles in the least amount of time and at the lowest possible cost. guiding those design strategies through chromatography based physicochemical properties (i.e. elog d and epsa) is one of the more successful approaches developed in order to accelerate the process. acknowledgements: laurence philippe for her continuous support and mentorship, patrick mullins and jason ramsay for critical editing and justin montgommery for figures 6 and 7. references [1] t. kennedy. managing the drug discovery/development interface. drug discov today 2 (1997) 436444. [2] a.d.w. mcnaught, m. nic, j. jirat, b. kosata, a. jenkins. iupac. compendium of chemical terminology, 2nd ed. iupac the "gold book", blackwell scientific publications, oxford 2006. [3] m.j. waring. lipophilicity in drug discovery. expert opinion on drug discovery 5 (2010) 235-248. [4] r. mannhold, g.i. poda, c. ostermann, i.v. tetko. calculation of molecular lipophilicity: state-ofthe-art and comparison of log p methods on more than 96,000 compounds. j pharm sci 98 (2009) 861-893. [5] k. valkó, c. bevan, d. reynolds. chromatographic hydrophobicity index by fast-gradient rp-hplc: a high-throughput alternative to log p/log d. analytical chemistry 69 (1997) 2022-2029. [6] f. lombardo, m.y. shalaeva, k.a. tupper, f. gao. elogdoct: a tool for lipophilicity determination in drug discovery. 2. basic and neutral compounds. journal of medicinal chemistry 44 (2001) 24902497. [7] c.a. lipinski, f. lombardo, b.w. dominy, p.j. feeney. experimental and computational approaches to estimate solubility and permeability in drug discovery and development settings1pii of original article: s0169-409x(96)00423-1. the article was originally published in advanced drug delivery reviews 23 (1997) 3–25.1. advanced drug delivery reviews 46 (2001) 3-26. [8] a.l. hopkins, c.r. groom, a. alex. ligand efficiency: a useful metric for lead selection. drug discov today 9 (2004) 430-431. [9] p.d. leeson, b. springthorpe. the influence of drug-like concepts on decision-making in medicinal chemistry. nat rev drug discov 6 (2007) 881-890. [10] t. ryckmans, m.p. edwards, v.a. horne, a.m. correia, d.r. owen, l.r. thompson, i. tran, m.f. tutt, t. young. rapid assessment of a novel series of selective cb(2) agonists using parallel synthesis protocols: a lipophilic efficiency (lipe) analysis. bioorg med chem lett 19 (2009) 4406-4409. [11] a.f. stepan, g.w. kauffman, c.e. keefer, p.r. verhoest, m. edwards. evaluating the differences in cycloalkyl ether metabolism using the design parameter “lipophilic metabolism efficiency” (lipmete) and a matched molecular pairs analysis. journal of medicinal chemistry 56 (2013) 69856990. [12] t.t. wager, x. hou, p.r. verhoest, a. villalobos. moving beyond rules: the development of a central nervous system multiparameter optimization (cns mpo) approach to enable alignment of druglike properties. acs chemical neuroscience 1 (2010) 435-449. [13] m.m. claffey, c.j. helal, p.r. verhoest, z. kang, k.s. fors, s. jung, j. zhong, m.w. bundesmann, x. hou, s. lui, r.j. kleiman, m. vanase-frawley, a.w. schmidt, f. menniti, c.j. schmidt, w.e. hoffman, m. hajos, l. mcdowell, r.e. o’connor, m. macdougall-murphy, k.r. fonseca, s.l. becker, f.r. nelson, s. liras. application of structure-based drug design and parallel chemistry to identify admet & dmpk 6(2) (2018) 85-104 leveraging chromatography physchem properties in drug design doi: 10.5599/admet.529 99 selective, brain penetrant, in vivo active phosphodiesterase 9a inhibitors. journal of medicinal chemistry 55 (2012) 9055-9068. [14] m.p. gleeson. generation of a set of simple, interpretable admet rules of thumb. journal of medicinal chemistry 51 (2008) 817-834. [15] m.v. varma, r.s. obach, c. rotter, h.r. miller, g. chang, s.j. steyn, a. el-kattan, m.d. troutman. physicochemical space for optimum oral bioavailability: contribution of human intestinal absorption and first-pass elimination. j med chem 53 (2010) 1098-1108. [16] t.w. johnson, k.r. dress, m. edwards. using the golden triangle to optimize clearance and oral absorption. bioorganic & medicinal chemistry letters 19 (2009) 5560-5564. [17] a.l. hopkins, g.m. keserü, p.d. leeson, d.c. rees, c.h. reynolds. the role of ligand efficiency metrics in drug discovery. nature reviews drug discovery 13 (2014) 105. [18] j.s. scott, m.j. waring. practical application of ligand efficiency metrics in lead optimisation. bioorganic & medicinal chemistry (2018). [19] r.j. young, p.d. leeson. mapping the efficiency and physicochemical trajectories of successful optimizations. journal of medicinal chemistry (2018). [20] t.w. johnson, r.a. gallego, m.p. edwards. lipophilic efficiency as an important metric in drug design. journal of medicinal chemistry (2018). [21] c.r. butler, e.m. beck, a. harris, z. huang, l.a. mcallister, c.w. am ende, k. fennell, t.l. foley, k. fonseca, s.j. hawrylik, d.s. johnson, j.d. knafels, s. mente, g.s. noell, j. pandit, t.b. phillips, j.r. piro, b.n. rogers, t.a. samad, j. wang, s. wan, m.a. brodney. azetidine and piperidine carbamates as efficient, covalent inhibitors of monoacylglycerol lipase. journal of medicinal chemistry 60 (2017) 9860-9873. [22] m. ammirati, s.w. bagley, s.k. bhattacharya, l. buckbinder, a.a. carlo, r. conrad, c. cortes, r.l. dow, m.s. dowling, a. el-kattan, k. ford, c.r.w. guimarães, d. hepworth, w. jiao, j. laperle, s. liu, a. londregan, p.m. loria, a.m. mathiowetz, m. munchhof, s.t.m. orr, d.n. petersen, d.a. price, a. skoura, a.c. smith, j. wang. discovery of an in vivo tool to establish proof-of-concept for map4k4based antidiabetic treatment. acs medicinal chemistry letters 6 (2015) 1128-1133. [23] m. pettersson, d.s. johnson, c. subramanyam, k.r. bales, c.w. am ende, b.a. fish, m.e. green, g.w. kauffman, p.b. mullins, t. navaratnam, s.m. sakya, c.m. stiff, t.p. tran, l. xie, l. zhang, l.r. pustilnik, b.c. vetelino, k.m. wood, n. pozdnyakov, p.r. verhoest, c.j. o’donnell. design, synthesis, and pharmacological evaluation of a novel series of pyridopyrazine-1,6-dione γ-secretase modulators. journal of medicinal chemistry 57 (2014) 1046-1062. [24] c. hansch, j.p. bjorkroth, a. leo. hydrophobicity and central nervous system agents: on the principle of minimal hydrophobicity in drug design. j pharm sci 76 (1987) 663-687. [25] l. di, p. artursson, a. avdeef, g.f. ecker, b. faller, h. fischer, j.b. houston, m. kansy, e.h. kerns, s.d. krämer. evidence-based approach to assess passive diffusion and carrier-mediated drug transport. drug discov today 17 (2012) 905-912. [26] c.a. lipinski, f. lombardo, b.w. dominy, p.j. feeney. experimental and computational approaches to estimate solubility and permeability in drug discovery and development settings1. advanced drug delivery reviews 23 (1997) 3-25. [27] c.a. lipinski, f. lombardo, b.w. dominy, p.j. feeney. experimental and computational approaches to estimate solubility and permeability in drug discovery and development settings. advanced drug delivery reviews 46 (2001) 3-26. [28] m.e. kavanagh, m.r. doddareddy, m. kassiou. the development of cns-active lrrk2 inhibitors using property-directed optimisation. bioorganic & medicinal chemistry letters 23 (2013) 36903696. [29] m.p. deninno, m. andrews, a.s. bell, y. chen, c. eller-zarbo, n. eshelby, j.b. etienne, d.e. moore, m.j. palmer, m.s. visser, l.j. yu, w.j. zavadoski, e. michael gibbs. the discovery of potent, selective, and orally bioavailable pde9 inhibitors as potential hypoglycemic agents. bioorg med chem lett 19 (2009) 2537-2541. g.h. goetz and m. shalaeva admet & dmpk 6(2) (2018) 85-104 100 [30] z. rankovic. cns physicochemical property space shaped by a diverse set of molecules with experimentally determined exposure in the mouse brain. j med chem 60 (2017) 5943-5954. [31] n.m. ahmad. solubility-driven lead optimisation: recent examples and personal perspectives. bioorganic & medicinal chemistry letters 26 (2016) 2975-2979. [32] d.j. craik, d.p. fairlie, s. liras, d. price. the future of peptide-based drugs. chem biol drug des 81 (2013) 136-147. [33] k. fosgerau, t. hoffmann. peptide therapeutics: current status and future directions. drug discovery today 20 (2015) 122-128. [34] m.e. bunnage, e.l.p. chekler, l.h. jones. target validation using chemical probes. nature chemical biology 9 (2013) 195. [35] c.r. guimaraes, a.m. mathiowetz, m. shalaeva, g. goetz, s. liras. use of 3d properties to characterize beyond rule-of-5 property space for passive permeation. journal of chemical information and modeling 52 (2012) 882-890. [36] b.c. doak, b. over, f. giordanetto, j. kihlberg. oral druggable space beyond the rule of 5: insights from drugs and clinical candidates. chem biol 21 (2014) 1115-1142. [37] c.k. wang, s.e. northfield, j.e. swedberg, b. colless, s. chaousis, d.a. price, s. liras, d.j. craik. exploring experimental and computational markers of cyclic peptides: charting islands of permeability. eur j med chem 97 (2015) 202-213. [38] a.t. bockus, k.w. lexa, c.r. pye, a.s. kalgutkar, j.w. gardner, k.c.r. hund, w.m. hewitt, j.a. schwochert, e. glassey, d.a. price, a.m. mathiowetz, s. liras, m.p. jacobson, r.s. lokey. probing the physicochemical boundaries of cell permeability and oral bioavailability in lipophilic macrocycles inspired by natural products. journal of medicinal chemistry 58 (2015) 4581-4589. [39] c.r. pye, w.m. hewitt, j.a. schwochert, t.d. haddad, c.e. townsend, l. etienne, y. lao, c. limberakis, a. furukawa, a.m. mathiowetz, d.a. price, s. liras, r.s. lokey. non-classical size dependence of permeation defines bounds for passive adsorption of large drug molecules. journal of medicinal chemistry (2017). [40] p. seiler. interconversion of lipophilicities from hydrocarbon/water systems into the octanol/water system. eur j med chem 39 (1974) 473-479. [41] t.w. von geldern, d.j. hoffman, j.a. kester, h.n. nellans, b.d. dayton, s.v. calzadilla, k.c. marsh, l. hernandez, w. chiou, d.b. dixon, j.r. wu-wong, t.j. opgenorth. azole endothelin antagonists. 3. using δ log p as a tool to improve absorption. journal of medicinal chemistry 39 (1996) 982-991. [42] r.c. young, r.c. mitchell, t.h. brown, c.r. ganellin, r. griffiths, m. jones, k.k. rana, d. saunders, i.r. smith. development of a new physicochemical model for brain penetration and its application to the design of centrally acting h2 receptor histamine antagonists. journal of medicinal chemistry 31 (1988) 656-671. [43] j.t. goodwin, r.a. conradi, n.f. ho, p.s. burton. physicochemical determinants of passive membrane permeability: role of solute hydrogen-bonding potential and volume. j med chem 44 (2001) 3721-3729. [44] m.h. abraham, r.e. smith, r. luchtefeld, a.j. boorem, r. luo, w.e. acree, jr. prediction of solubility of drugs and other compounds in organic solvents. j pharm sci 99 (2010) 1500-1515. [45] g. gilli, p. gilli, the nature of the hydrogen bond: outline of a comprehensive hydrogen bond theory oxford university press, new york2009. [46] m. shalaeva, g. caron, y.a. abramov, t.n. o’connell, m.s. plummer, g. yalamanchi, k.a. farley, g.h. goetz, l. philippe, m.j. shapiro. integrating intramolecular hydrogen bonding (imhb) considerations in drug discovery using δlogp as a tool. journal of medicinal chemistry 56 (2013) 4870-4879. [47] c.k. wittekindt, a. cosmo-rs as a predictive tool for lipophilicity. qsar comb. sci. 28 (2009) 874−877. admet & dmpk 6(2) (2018) 85-104 leveraging chromatography physchem properties in drug design doi: 10.5599/admet.529 101 [48] g. caron, m. vallaro, g. ermondi, g.h. goetz, y.a. abramov, l. philippe, m. shalaeva. a fast chromatographic method for estimating lipophilicity and ionization in nonpolar membrane-like environment. molecular pharmaceutics 13 (2016) 1100-1110. [49] p. saha, j.h. kou. effect of solubilizing excipients on permeation of poorly water-soluble compounds across caco-2 cell monolayers. european journal of pharmaceutics and biopharmaceutics 50 (2000) 403-411. [50] k. futatsugi, k.b. bahnck, m.b. brenner, j. buxton, j.e. chin, s.b. coffey, j. dubins, d. flynn, d. gautreau, a. guzman-perez, j.r. hadcock, d. hepworth, m. herr, t. hinchey, a.m. janssen, s.m. jennings, w. jiao, s.y. lavergne, b. li, m. li, m.j. munchhof, s.t.m. orr, d.w. piotrowski, n.s. roush, m. sammons, b.d. stevens, g. storer, j. wang, j.s. warmus, l. wei, a.c. wolford. optimization of triazole-based tgr5 agonists towards orally available agents. medchemcomm 4 (2013) 205-210. [51] m.j. waring, j. arrowsmith, a.r. leach, p.d. leeson, s. mandrell, r.m. owen, g. pairaudeau, w.d. pennie, s.d. pickett, j. wang, o. wallace, a. weir. an analysis of the attrition of drug candidates from four major pharmaceutical companies. nature reviews drug discovery 14 (2015) 475. [52] j.d. hughes, j. blagg, d.a. price, s. bailey, g.a. decrescenzo, r.v. devraj, e. ellsworth, y.m. fobian, m.e. gibbs, r.w. gilles, n. greene, e. huang, t. krieger-burke, j. loesel, t. wager, l. whiteley, y. zhang. physiochemical drug properties associated with in vivo toxicological outcomes. bioorg med chem lett 18 (2008) 4872-4875. [53] d.a. price, j. blagg, l. jones, n. greene, t. wager. physicochemical drug properties associated with in vivo toxicological outcomes: a review. expert opin drug metab toxicol 5 (2009) 921-931. [54] m.c. wenlock. designing safer oral drugs. medchemcomm 8 (2017) 571-577. [55] f. lombardo, m.y. shalaeva, k.a. tupper, f. gao, m.h. abraham. elogpoct: a tool for lipophilicity determination in drug discovery. journal of medicinal chemistry 43 (2000) 2922-2928. [56] y. gravenfors, j. viklund, j. blid, t. ginman, s. karlström, j. kihlström, k. kolmodin, j. lindström, s. von berg, f. von kieseritzky, c. slivo, b.-m. swahn, l.-l. olsson, p. johansson, s. eketjäll, j. fälting, f. jeppsson, k. strömberg, j. janson, f. rahm. new aminoimidazoles as β-secretase (bace-1) inhibitors showing amyloid-β (aβ) lowering in brain. journal of medicinal chemistry 55 (2012) 9297-9311. [57] p. mukherjee, m. pettersson, j.k. dutra, l. xie, c.w. am ende. trifluoromethyl oxetanes: synthesis and evaluation as a tert-butyl isostere. chemmedchem 12 (2017) 1574-1577. [58] a.f. stepan, k. karki, w.s. mcdonald, p.h. dorff, j.k. dutra, k.j. dirico, a. won, c. subramanyam, i.v. efremov, c.j. o'donnell, c.e. nolan, s.l. becker, l.r. pustilnik, b. sneed, h. sun, y. lu, a.e. robshaw, d. riddell, t.j. o'sullivan, e. sibley, s. capetta, k. atchison, a.j. hallgren, e. miller, a. wood, r.s. obach. metabolism-directed design of oxetane-containing arylsulfonamide derivatives as gamma-secretase inhibitors. j med chem 54 (2011) 7772-7783. [59] d.a. griffith, r.l. dow, k. huard, d.j. edmonds, s.w. bagley, j. polivkova, d. zeng, c.n. garciairizarry, j.a. southers, w. esler, p. amor, k. loomis, k. mcpherson, k.b. bahnck, c. preville, t. banks, d.e. moore, a.m. mathiowetz, e. menhaji-klotz, a.c. smith, s.d. doran, d.a. beebe, m.f. dunn. spirolactam-based acetyl-coa carboxylase inhibitors: toward improved metabolic stability of a chromanone lead structure. j med chem 56 (2013) 7110-7119. [60] j.e. davoren, c.w. lee, m. garnsey, m.a. brodney, j. cordes, k. dlugolenski, j.r. edgerton, a.r. harris, c.j. helal, s. jenkinson, g.w. kauffman, t.p. kenakin, j.t. lazzaro, s.m. lotarski, y. mao, d.m. nason, c. northcott, l. nottebaum, s.v. o'neil, b. pettersen, m. popiolek, v. reinhart, r. salomon-ferrer, s.j. steyn, d. webb, l. zhang, s. grimwood. discovery of the potent and selective m1 pam-agonist n-[(3r,4s)-3-hydroxytetrahydro-2h-pyran-4-yl]-5-methyl-4-[4-(1,3-thiazol-4yl)ben zyl]pyridine-2-carboxamide (pf-06767832): evaluation of efficacy and cholinergic side effects. j med chem 59 (2016) 6313-6328. [61] a. forsberg, a. juréus, z. cselényi, m. eriksdotter, y. freund-levi, f. jeppsson, b.-m. swahn, j. sandell, p. julin, m. schou, j. andersson, p. johnström, k. varnäs, c. halldin, l. farde, s. svensson. low background and high contrast pet imaging of amyloid-β with [(11)c]azd2995 and g.h. goetz and m. shalaeva admet & dmpk 6(2) (2018) 85-104 102 [(11)c]azd2184 in alzheimer’s disease patients. european journal of nuclear medicine and molecular imaging 40 (2013) 580-593. [62] k. palm, k. luthman, a.l. ungell, g. strandlund, p. artursson. correlation of drug absorption with molecular surface properties. j pharm sci 85 (1996) 32-39. [63] h. van de waterbeemd, g. camenisch, g. folkers, o.a. raevsky. estimation of caco-2 cell permeability using calculated molecular descriptors. quantitative structure-activity relationships 15 (1996) 480-490. [64] y.h. zhao, m.h. abraham, a. ibrahim, p.v. fish, s. cole, m.l. lewis, m.j. de groot, d.p. reynolds. predicting penetration across the blood-brain barrier from simple descriptors and fragmentation schemes. j chem inf model 47 (2007) 170-175. [65] k. palm, p. stenberg, k. luthman, p. artursson1. polar molecular surface properties predict the intestinal absorption of drugs in humans. pharm res 14 (1997) 568-571. [66] d.e. clark. rapid calculation of polar molecular surface area and its application to the prediction of transport phenomena. 1. prediction of intestinal absorption. j pharm sci 88 (1999) 807-814. [67] d.e. clark. rapid calculation of polar molecular surface area and its application to the prediction of transport phenomena. 2. prediction of blood-brain barrier penetration. j pharm sci 88 (1999) 815821. [68] s. winiwarter, n.m. bonham, f. ax, a. hallberg, h. lennernas, a. karlen. correlation of human jejunal permeability (in vivo) of drugs with experimentally and theoretically derived parameters. a multivariate data analysis approach. j med chem 41 (1998) 4939-4949. [69] j. kelder, p.d. grootenhuis, d.m. bayada, l.p. delbressine, j.p. ploemen. polar molecular surface as a dominating determinant for oral absorption and brain penetration of drugs. pharm res 16 (1999) 1514-1519. [70] l.h. krarup, i.t. christensen, l. hovgaard, s. frokjaer. predicting drug absorption from molecular surface properties based on molecular dynamics simulations. pharm res 15 (1998) 972-978. [71] p. ertl, b. rohde, p. selzer. fast calculation of molecular polar surface area as a sum of fragmentbased contributions and its application to the prediction of drug transport properties. j med chem 43 (2000) 3714-3717. [72] p. ertl, polar surface area, molecular drug properties, wiley-vch verlag gmbh & co. kgaa2008, pp. 111-126. [73] g. schaftenaar, j. de vlieg. quantum mechanical polar surface area. journal of computer-aided molecular design 26 (2012) 311-318. [74] s. prasanna, r.j. doerksen. topological polar surface area: a useful descriptor in 2d-qsar. curr med chem 16 (2009) 21-41. [75] j. fernandes, c.r. gattass. topological polar surface area defines substrate transport by multidrug resistance associated protein 1 (mrp1/abcc1). j med chem 52 (2009) 1214-1218. [76] d.f. veber, s.r. johnson, h.y. cheng, b.r. smith, k.w. ward, k.d. kopple. molecular properties that influence the oral bioavailability of drug candidates. j med chem 45 (2002) 2615-2623. [77] h. van de waterbeemd, g. camenisch, g. folkers, j.r. chretien, o.a. raevsky. estimation of bloodbrain barrier crossing of drugs using molecular size and shape, and h-bonding descriptors. j drug target 6 (1998) 151-165. [78] t.t. wager, x. hou, p.r. verhoest, a. villalobos. central nervous system multiparameter optimization desirability: application in drug discovery. acs chem neurosci 7 (2016) 767-775. [79] j. drews. stategic trends in the drug industry. drug discov today 8 (2003) 411-420. [80] g.h. goetz, w. farrell, m. shalaeva, s. sciabola, d. anderson, j. yan, l. philippe, m.j. shapiro. high throughput method for the indirect detection of intramolecular hydrogen bonding. journal of medicinal chemistry 57 (2014) 2920-2929. admet & dmpk 6(2) (2018) 85-104 leveraging chromatography physchem properties in drug design doi: 10.5599/admet.529 103 [81] a. alex, d.s. millan, m. perez, f. wakenhut, g.a. whitlock. intramolecular hydrogen bonding to improve membrane permeability and absorption in beyond rule of five chemical space. medchemcomm 2 (2011) 669-674. [82] b. kuhn, p. mohr, m. stahl. intramolecular hydrogen bonding in medicinal chemistry. journal of medicinal chemistry 53 (2010) 2601-2611. [83] s.b. rafi, b.r. hearn, p. vedantham, m.p. jacobson, a.r. renslo. predicting and improving the membrane permeability of peptidic small molecules. journal of medicinal chemistry 55 (2012) 3163-3169. [84] t. rezai, j.e. bock, m.v. zhou, c. kalyanaraman, r.s. lokey, m.p. jacobson. conformational flexibility, internal hydrogen bonding, and passive membrane permeability:  successful in silico prediction of the relative permeabilities of cyclic peptides. journal of the american chemical society 128 (2006) 14073-14080. [85] t. rezai, b. yu, g.l. millhauser, m.p. jacobson, r.s. lokey. testing the conformational hypothesis of passive membrane permeability using synthetic cyclic peptide diastereomers. j am chem soc 128 (2006) 2510-2511. [86] h. cheng, j.e. hoffman, p.t. le, m. pairish, r. kania, w. farrell, s. bagrodia, j. yuan, s. sun, e. zhang, c. xiang, d. dalvie, s.v. rahavendran. structure-based design, sar analysis and antitumor activity of pi3k/mtor dual inhibitors from 4-methylpyridopyrimidinone series. bioorganic & medicinal chemistry letters 23 (2013) 2787-2792. [87] f. wakenhut, t.d. tran, c. pickford, s. shaw, m. westby, c. smith-burchnell, l. watson, m. paradowski, j. milbank, d. stonehouse, k. cheung, r. wybrow, f. daverio, s. crook, k. statham, d. leese, d. stead, f. adam, d. hay, l.r. roberts, j.y. chiva, c. nichols, d.c. blakemore, g.h. goetz, y. che, i. gardner, s. dayal, a. pike, r. webster, d.c. pryde. the discovery of potent nonstructural protein 5a (ns5a) inhibitors with a unique resistance profile-part 2. chemmedchem 9 (2014) 13871396. [88] g.h. goetz, l. philippe, m.j. shapiro. epsa: a novel supercritical fluid chromatography technique enabling the design of permeable cyclic peptides. acs medicinal chemistry letters 5 (2014) 11671172. [89] d.s. nielsen, h.n. hoang, r.j. lohman, t.a. hill, a.j. lucke, d.j. craik, d.j. edmonds, d.a. griffith, c.j. rotter, r.b. ruggeri. improving on nature: making a cyclic heptapeptide orally bioavailable. angewandte chemie international edition 53 (2014) 12059-12063. [90] j. schwochert, r. turner, m. thang, r.f. berkeley, a.r. ponkey, k.m. rodriguez, s.s. leung, b. khunte, g. goetz, c. limberakis. peptide to peptoid substitutions increase cell permeability in cyclic hexapeptides. organic letters 17 (2015) 2928-2931. [91] s. sciabola, g.h. goetz, g. bai, b.n. rogers, d.l. gray, a. duplantier, k.r. fonseca, m.a. vanasefrawley, n.m. kablaoui. systematic n-methylation of oxytocin: impact on pharmacology and intramolecular hydrogen bonding network. bioorganic & medicinal chemistry 24 (2016) 3513-3520. [92] s.g. sciabola, gilles ; farrell, william high-throughput method for the indirect detection of imhb. ssc14 14th scandinavian symposium on chemometrics 2015sardinia, italy, 2015. [93] g. goetz. epsa: a novel supercritical fluid chromatography technique enabling the design of permeable cyclic peptides. world preclinical congressboston, 2016. [94] d.a. degoey, h.-j. chen, p.b. cox, m.d. wendt. beyond the rule of 5: lessons learned from abbvie’s drugs and compound collection: miniperspective. journal of medicinal chemistry (2017). [95] t. vorherr, i. lewis, j. berghausen, s. desrayaud, m. schaefer. modulation of oral bioavailability and metabolism for closely related cyclic hexapeptides. international journal of peptide research and therapeutics (2017) 1-14. [96] s.c. ha, seth optimizing permeability of peptides. computational drug developmentboston, 2015. [97] j.p.l. kiplinger, p. m.; rego, m. j.; rouse, e. s. . epsa: a new bioavailability / permeability assay. in: http://www.avericadiscovery.com/blog/epsa/ (ed.), 2015. http://www.avericadiscovery.com/blog/epsa/ g.h. goetz and m. shalaeva admet & dmpk 6(2) (2018) 85-104 104 [98] p. wipf. acs medicinal chemistry letters: technology notes. acs medicinal chemistry letters 8 (2017) 1-2. ©2018 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ pkquest: pbpk modeling of highly lipid soluble and extracellular solutes doi: 10.5599/admet.579 60 admet & dmpk 7(1) (2019) 60-75; doi: http://dx.doi.org/10.5599/admet.579 open access: issn: 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper pkquest: pbpk modeling of highly lipid soluble and extracellular solutes david g. levitt department of integrative biology and physiology, university of minnesota, 6-125 jackson hall, 321 church st. s. e., minneapolis, mn 55455, usa e-mail: levit001@umn.edu received: august 12, 2018; revised: november 16, 2018; published: november 27, 2018 abstract one of the primary objectives of physiologically based pharmacokinetics (pbpk) is the prediction of a drug’s pharmacokinetics just from knowledge of its physicochemical structure. unfortunately, at present, the accuracy of this prediction is limited for most drugs because of uncertainty about the drug’s organ/blood partition coefficient (k). however, there are two classes of solutes which are exceptions to this: 1) the highly lipid soluble (hls) solutes, and 2) the extracellular (ecs) solutes. since the hls drugs (eg, volatile anesthetics, propofol, cannabinol) have lipid/water partition coefficients (pl/w) of 100 or greater, their k is dominated by the tissue fat fraction and one can accurately predict k just from in vitro measurements of pl/w along with prior anatomic measurements of the fat fraction of the organs in the pbpk model. since the ecs drugs, such as most antibiotics, cannot penetrate cells, they are not subject to the intracellular binding that complicates the prediction of k for the weak bases and acids. the ecs k is determined primarily by plasma and interstitial albumin binding and can be predicted from in vitro measurements of plasma albumin binding along with prior measurements of interstitial tissue volume and albumin concentrations. this review provides an in depth discussion of the pbpk modeling of these two drug classes along with many specific clinical examples illustrating the good pbpk predictions possible with just zero (volatile anesthetics) or 1 (the clearance) adjustable parameter. the pbpk analysis uses pkquest, a freely distributed, general purpose pharmacokinetic program. pkquest is designed so that application to the hls and ecs solute classes is especially easy. the user only needs to enter the specific parameters that are required to characterize the drug (eg, pl/w for hls or plasma albumin binding for ecs) with all the other pbpk parameters (organ blood flow, fat fraction, extracellular volumes, etc.) are set by default. keywords pharmacokinetics; anesthetics; interstitial; extracellular; adipose introduction the standard well-stirred, flow-limited pbpk differential equation describing the solute balance of organ i is deceptively simple [1]: i i i a id ( ) [c ( ) ( )] d c t v f t c t t   (1) where f i is the organ blood flow, c a (t) is the arterial blood concentration and c i (t) is the venous blood concentration leaving the organ and it is assumed that the free, unbound well-mixed tissue concentration http://www.pub.iapchem.org/ojs/index.php/admet/index admet & dmpk 7(1) (2019) 60-75 lipid soluble and extracellular solutes doi: 10.5599/admet.579 61 is equal to the free concentration in the venous blood of the organ. all the complexity is in the parameter v i which is the organ “volume of distribution” defined by: i i i i i i i i i i b t t b t i i i i b t totalamountofsolute inorgan i [ ] v v c v c v c c v k v v k v         (2) where vb i and vt i are the anatomic blood and extravascular organ volumes, respectively, ct i is the tissue concentration and k i (= ct i /c i ) is the tissue/blood partition coefficient. the organ blood flow (f i ) and the blood (vb i ) and tissue (vt i ) volumes can be accurately estimated from prior human physiological and anatomic measurements. however, k i , which depends on the specific physicochemical properties of the solute and varies from values of 0.2 or less to greater than 100, dominates the kinetics and is the major uncertainty and limitation of pbpk analysis. in general, for an accurate pbpk analysis it is necessary to determine the k i for the solute of interest from complex animal (usually rat) tissue measurements and then assume (hope) that the rat values can be extrapolated to the human. ideally, one would like to be able to predict k i just based on its chemical structure, allowing one to determine a drug’s detailed pk a priori. unfortunately, for the majority of drugs that are weak acids or bases, it is not, as yet, possible to predict k i accurately enough to meet this ideal. the current state of the art in predicting k i is illustrated by the algorithm developed by poulin and colleagues over many years of analysis [2]. about 50 % of the predictions are accurate to within a factor of 2, with about 15 % of the predictions off by a factor of greater than 3 fold. although this level of accuracy is useful for, eg, predicting preliminary dosage levels for a trial drug, it severely limits the value of pbpk predictions because they are unlikely to be superior to those made using just a simple 1-compartment model based just on the standard pharmacokinetic (pk) measurements of clearance and volume of distribution. however, there are two major exceptions to this limitation: the highly lipid soluble (hls) and the extracellular drugs (ecs). the hls drugs (volatile anesthetics, propofol, cannabinol, etc.) have lipid/water partition coefficients (pl/w) of 100 or greater and their k i is dominated by the tissue fat fraction [1]. this means that one can predict the k i just from in vitro measurements of pl/w along with previous measurements of the fat fraction of all the tissues in the pbpk model (see fig. 1). the ecs drugs, such as most antibiotics, cannot penetrate cells, and therefore are not subject to the intracellular binding that complicates the prediction of k i for the weak bases and acids. the k i of most ecs drugs is determined primarily by the volume and albumin concentration in the interstitial space and the albumin binding affinity. thus, the ecs k i can be predicted from prior measurements of tissue interstitial volumes and albumin concentrations and in vitro measurements of albumin binding affinity. pkquest is a freely distributed (www.pkquest.com), java based, pk and pbpk program. the major effort in its design has been to provide a user interface that is both simple enough to be used by students as an adjunct in pk courses and general enough to be applicable to most pk and pbpk applications. since its introduction in 2002, pkquest has been applied to hundreds of solutes described in more than 11 publications [3-13]. recently, a freely distributed textbook (“computer assisted human pharmacokinetics”) has been developed that covers most pk topics and is closely integrated with pkquest[1]. pkquest implements the standard pbpk approach. the body is divided into 14 “tissue” compartments (fig. 1), each of which is described by eqs. (1) and (2) and three “tissue” parameters (vt i , f i , and k i ). for a given solute input (oral, iv, etc.) the set of equations is solved numerically, and a plot of the time http://www.pkquest.com/ david levitt admet & dmpk 7(1) (2019)60-75 62 dependent solute concentration for each of the organs is plotted. usually, of particular interest, is the time dependence of the concentration in the “vein” compartment. figure 1. pbpk model used in pkquest. the body is modeled by 12 tissue regions, arranged in parallel (except for the “portal” and “lung” tissue which are in series) connected by the “vein” and “artery” compartments. the tissue “portal” refers to the stomach, small and large intestine, spleen and pancreatic organs. the tissue parameters are listed in table 1. in general, pbpk modeling requires input of a standard set of the 28 tissue volume (vti) and blood flow (fi) parameters. in pkquest, a “standard human” set of these parameters has been refined by application of pkquest to hundreds of solutes with varying properties. for example, the muscle blood flow was determined from the pk of d2o and the adipose blood flows from the pk of the volatile anesthetics (see below). these “standard human” values are listed in table 1 (in pkquest, the default units are minutes, liters and kilograms). since these are the default background values, the user does not need to be concerned with supplying them. of course, pkquest allows these values to be varied in the more general case. one of the novel features of pkquest is that it is designed so that the pbpk modeling of hls and ecs solutes can be run by entering just the minimum set of adjustable parameters (body fat fraction, pl/w, albumin binding, etc.) with all the other required human parameters as standard defaults. in addition to potentially predicting the pk of unknown solutes, this pbpk application is of heuristic value in, for example, teaching nurse anesthetists how the pk of volatile anesthetics depends on, e.g. respiratory rate. the following two sections will review the pk of these two solute classes along with clinical examples of the application of pkquest to specific solutes. highly lipid soluble solutes (hls) for an ideal hls solute, the tissue/blood partition k i is determined entirely by the lipid fraction of the tissue (fl i ) and blood (fl b ) and the lipid/water partition fraction (pl/w) [1]: admet & dmpk 7(1) (2019) 60-75 lipid soluble and extracellular solutes doi: 10.5599/admet.579 63 1 1 i i l l/w li b b l l/w l ( ) ( ) f p f k f p f      (3) equation (3) assumes that the blood hls binding is due entirely to the blood lipid fraction (fl b ). in humans, the directly measured fl b is about 0.005 [12]. however, using this value often underestimates the blood hls binding because there is additional plasma hls binding, primarily by albumin. the procedure used in pkquest is to input the in vitro measurement of the “free plasma fraction” if it is known and then use this value to estimate the apparent fl b . if it is not known, the default fl b of 0.015 (table 1) provides a rough starting estimate and it is then treated as an adjustable parameter. table 1. organ pbpk parameters for the standard 20 % fat, 70 kg human. the organ “portal” refers to all the organs drained by the portal vein (eg, gi tract, spleen, pancreas). the organs “tendon” and “other” represent poorly characterized, low flow connective tissues. the adipose and adipose 2 organ weights (assumed equal) are adjusted for the individual’s body fat fraction. the ecf fraction is the fraction of the non-lipid organ weight that is extracellular. kalb is the ratio of the albumin concentration in the edta interstitial space relative to the plasma albumin. organ weight (kg) perfusion (l/min/kg) lipid fraction ecf fraction kalb vein 4.29 1.0 0.015 0.595 na artery 1.21 0.0 0.015 0.595 na liver 1.8 0.25 0.02 0.23 0.5 portal 1.5 0.75 0.016 0.3 0.35 kidney 0.31 4.0 0.0136 0.165 0.35 brain 1.4 0.56 0.0176 0 0.1 heart 0.33 0.8 0.0136 0.264 0.5 muscle 26.0 0.0225 0.0136 0.15 0.5 skin 2.6 0.1 0.0136 0.6 0.25 lung 0.536 -1.0 0.0136 0.2 0.35 tendon 3.0 0.01 0.0136 1 0.25 other 5.522 0.02 0.0136 0.2 0.25 adipose 8.758 0.074 0.8 1 0.35 adipose 2 8.758 0.01408 0.8 1 0.35 bone 4.0 0.0 0 0 0 the volatile anesthetics represent the ideal hls because they are relatively inert, have negligible specific tissue binding and, in most cases, are not metabolized with their clearance determined solely by the respiration rate and blood/gas partition. thus, their pk can be predicted using the hls pbpk model along with in vitro measurements of oil/water, blood/water and air/water partition coefficients with no adjustable individual subject parameters. for the volatile anesthetics there are detailed measurements of the oil/water, air/water, blood/gas and various tissue partition coefficients (k i ), all of which are qualitatively consistent with eq. (3). the pbpk parameter that dominates the pk of the hls is the adipose tissue volume and blood flow. it is essential that the magnitude of the heterogeneity (if any) of adipose blood flow is included in the pbpk model. the time constant (t) for adipose tissue equilibration is described by (assuming a well-mixed, flow limited tissue): david levitt admet & dmpk 7(1) (2019)60-75 64 adiposeweight (adipose/bloodpartition) adiposebloodflow t  (4) the units are kg for adipose weight, and kg/min for blood flow while the adipose/blood partition is dimensionless. for the volatile anesthetics t ranges from about 500 min to 3 days. at early times (several hours) before the lipid becomes saturated, the adipose tissue behaves like an infinite sink and the pk is not dependent on the heterogeneity. it is only at times that are long relative to t that the heterogeneity becomes apparent. thus, in order to determine adipose perfusion heterogeneity it is necessary to have pk measurements that extend to days. probably the best available set of measurements of this kind is the remarkable series of studies of the pk of the volatile anesthetics isoflurane, sevoflurane and desflurane by eger and colleagues [14, 15]. they measured the ventilation rate and the inspired, mixed and end tidal gas concentration for 6 days following a 30 minute uptake in normal volunteers. dr. eger kindly provided this data which was then modeled using pkquest to estimate the adipose perfusion heterogeneity. the analysis indicated that at least two equal volume adipose compartments with perfusion rates of 0.074 and 0.014 l/kg/min were needed to accurately model the long time data[10] and these are the default values in pkquest (table 1). figure 2. pkquest interface. the top panel lists all the input required to specify the pbpk pharmacokinetics for isoflurane (see text for details). figure 2 shows the interactive pkquest interface for isoflurane. the isoflurane kinetics are completely characterized by the entries in the top panel. the check box “volatile” turns on the pbpk for volatile solutes (which are, presumably, also hls). setting kbair, kwair, kfwat characterizes isoflurane. the parameters vent (alveolar ventilation) and vol (alveolar volume) are the standard 70 kg human values. the entries in the other panels describe the details of the dose regimen and the variables that should be admet & dmpk 7(1) (2019) 60-75 lipid soluble and extracellular solutes doi: 10.5599/admet.579 65 plotted in the output. figures 3 and 4 show semi-log plots of the pkquest pbpk prediction of the short and long time expired partial pressure (pexpired) relative to the pinspired during the 30 minute uptake period for isoflurane, sevoflurane and desflurane. it can be seen that, although these 3 anesthetics have significantly different properties (blood/air and oil/air partition), the human pk can be accurately fit using the same default pbpk model with no adjustable parameters. pbpk analysis is ideally suited for modeling the volatile anesthetics and has clear advantages over the 5 compartment mammillary model with 10 adjustable parameters that is classically used for these solutes [14, 15]. figure 3. pkquest pbpk predictions (black line) of short time pharmacokinetics of isoflurane, sevoflurane and desflurane using the identical basic pbpk model with two equal weight adipose organs with blood perfusions of 0.014 and 0.074 l/kg/min. isofluranes and isofluraned refer to two different sets of data. this same set of pbpk organ parameters is also applicable to non-volatile hls. figure 5 shows the pkquest interface top panel for cannabinol. the “fat/water partition” check box turns on the hls option. cannabinol has a very high oil/water partition of about 200,000 [10] (input in the “kfwat” box.). only two adjustable parameters were required to provide a good fit to the pk data of johansson et. al. [16]: the fractional liver clearance (0.651) and the blood fat fraction (0.0075). figure 6 is the pkquest output showing the pkquest fit to this data. david levitt admet & dmpk 7(1) (2019)60-75 66 figure 4. pkquest pbpk predictions (black line) of long time pharmacokinetics of isoflurane, sevoflurane and desflurane using the identical basic pbpk model with two equal weight adipose organs with blood perfusions of 0.014 and 0.074 l/kg/min. isofluranes and isofluraned refer to two different sets of data. figure 5. top panel of the pkquest interface for cannabinol (see text for details) an important clinical application of the pbpk hls model is in predicting the pk of anesthetics in obese subjects. the prediction is obtained simply by changing the “fat fr” entry in the pkquest interface (figs. 2 and 5). figures 7 and 8 show two examples obtained from the pkquest analysis of propofol [17]. servin et. al. [18] had previously shown that the “eye opening” time following a 180 minute propofol infusion was shorter in obese subjects (10.3 min) versus normal weight subjects (18.4 min). figure 7 shows that the pbpk model provides a similar prediction. propofol is routinely used for long term sedation. figure 8 shows admet & dmpk 7(1) (2019) 60-75 lipid soluble and extracellular solutes doi: 10.5599/admet.579 67 the arterial concentration during the washout period following a 10 day constant infusion. it can be seen that in obese subjects (red) the washout is markedly slower with a greatly prolonged waking time. in determining the pk for the obese subjects, it is assumed that the standard human tissue parameters, including those for “adipose1” and “adipose2” (table 1) can be extrapolated to the obese subjects. this is a major limitation because a careful evaluation of these parameters is not available for these subjects. figure 6. pkquest pbpk predictions (red line) of pharmacokinetics of cannabinol figure 7. plot of arterial blood propofol concentration in normal weight (black) and morbidly obese subjects (red) following a 180 minute propofol infusion. at a time of 180 minutes, the constant propofol infusion is terminated. the time for the arterial level to drop below the “eye opening concentration” is about 5 minutes for the obese and 13 minutes for the normal subjects david levitt admet & dmpk 7(1) (2019)60-75 68 figure 8. comparison of arterial propofol concentration in normal (black) and obese (red) subjects during washout following a 10 day (14,400 minutes) constant infusion. the constant infusion rate has been adjusted so that the concentrations at the end of the 10 day period are identical for the normal and obese subjects. the extreme limit of hls solutes are the persistent organic pollutants (pop), such as dioxins and polychlorinated biphenyls, which are characterized by human life times of several years and have lipid/water (pl/w) partition ranging from 10 5 to greater than 10 7 [12]. because of these long lifetimes and the near impossibility of obtaining accurate experimental human pk data, pop modeling and prediction has become one of the most important pbpk applications. there is a common misperception that the long lifetimes results, primarily, from their very high lipid partition and resultant slow adipose tissue washout [19]. what is not commonly recognized is that, as described in eq. (3), for pl/w greater than about 1000, k adipose reaches a maximum value equal to fl adipose /fl blood , and further increases in pl/w do not increase it above this maximum value described by:: l/w adipose adipose blood l l1000 / 0.8/ 0.005 160 p k f f     (5) substituting this relation for the maximum value of k adipose into eq. (4), the longest possible time constant for flow-limited adipose washout is: l/w adipose 1 adipose adipose 1000 / 160/ 0.01min 16,000min 11days p t k f       (6) where fadipose is the adipose perfusion rate. thus, high lipid partition can only account for washout time constants of about 11 days, much shorter than the year or more that is observed for many pops. one possible explanation is that eq. (6) is incorrect because it assumed that the adipose/blood exchange is flow limited, while, as clearly shown by levitt [12], it becomes diffusion limited for some pops. this diffusion limitation arises because the high pl/w of the pop produces such a low free water concentration that diffusion through the capillary wall becomes rate limiting. however, even taking admet & dmpk 7(1) (2019) 60-75 lipid soluble and extracellular solutes doi: 10.5599/admet.579 69 account of this diffusion limitation, the adipose exchange time constant is at least 10 fold shorter than the pop lifetimes and cannot be responsible for the observed human pop persistence [12]. the explanation of this discrepancy between the adipose pop time constant and the whole body human time constant is simply that the pk of the pops are limited by their extremely low rate of metabolism and excretion, not by the adipose/blood exchange. as the lipid partition increases, the free water concentration in the blood decreases and, presumably, the rate of liver metabolism decreases proportionally. if the metabolic time constant for excretion is long compared to the time constant for adipose and other tissue exchange, then at long times, one can ignore the pbpk model details and the plasma concentration can be described by a simple 1-compartment model characterized by its clearance (cl) and volume of distribution (v): /( / ) ( ) ( / )e ( / )e /    cl t tcl v t cl c t d v d v t v cl (7) where d is the dose and tcl is the metabolic (ie, liver) time constant. although the detailed pbpk model is required to describe the short time pk of the pops, at long times the pk depends only on the liver clearance and is independent of the kinetics of solute exchange in the peripheral tissues. this is quantitatively illustrated using the pbpk cannabinol model described above [12]. figure 9 shows a comparison of the plasma cannabinol concentration for the pbpk (black line) versus the 1-compartment model (red line) as the metabolic time constant (tcl) increases from 1.25 days (top panel) to 125 days (bottom panel). it can be seen (bottom panel) that when the metabolic excretion rate becomes rate limiting (tcl = 125 days >> tadipose =11 days), the one compartment model provides a good prediction of the pk after the initial ≈11 day transient period when the adipose tissue exchange is limiting. extracellular solutes (ecs) because the ecs, by definition, do not enter cells, they are not subject to the poorly characterized intracellular binding that confounds predictions of the k i for the typical weak acid or base. since the ecs solutes are confined to the plasma and interstitial space, the basic pbpk equations described above (eqs. (1) and (2)) are modified as follows: i i i a ip ec p p p i i i i ec p ec ect d ( ) [c ( ) ( )] c t v f t c t dt v v k v     (8) where fp i is the organ plasma flow, cp a is the plasma arterial concentration, and cp i is the venous plasma concentration. the ecs volume of distribution (vec i ) is defined in terms of the anatomical plasma (vp i ) and interstitial (vect i ) volumes and the ecs interstitial/plasma partition coefficient (kec i ). because both the plasma and interstitial ecs binding is dominated by the plasma and interstitial albumin, it can be shown [7] that the ecs partition coefficient (kec i ) can be described by: i i i ec alb p alb(1 )k k f k   (9) where kalb i is ratio of interstitial/plasma albumin concentration in organ i and fp is the fraction of the ecs solute that is free (unbound) in plasma and characterizes the ecs albumin binding affinity. if fp = 1 (no albumin binding), kec i = 1. if fp = 0 (very high affinity binding), kec i = kalb i . levitt [7] has carried out a detailed review of the literature and determined an optimal set of values for kalb i (the ratio of interstitial/plasma albumin concentration) and the fraction of each organ that is david levitt admet & dmpk 7(1) (2019)60-75 70 extracellular (ecf fraction) with vect i (the anatomic interstitial volume for each organ i) equal to the ecf fraction times the total organ volume. these parameters are listed in table 1 and are the default parameters in pkquest. using these predetermined values, the only parameter that is solute specific is fp (the free solute fraction in plasma) and this can be determined from in vitro measurements. thus, all of the parameters in eqs. (8) and (9) can be determined a priori, providing a complete description of the ecs pbpk. figure 9. comparison of 1-compartment (red) vs pbpk (black) model for hls (eg, cannabinol) with an adipose time constant of about 11 days. as the metabolic time constant (ie, liver clearance) increases from 1.25 days (top), to 12.5 (middle) to 125 days (bottom), the pbpk model approaches the 1-compartment model at times longer than 11 days. for most ecs, there is only one adjustable parameter, the clearance (usually renal). figure 10 shows the top panel of the pkquest interface for amoxicillin. the “extracellular” check box turns on the ecs pbpk model. the “freepl” is the fraction free in plasma (= fp), which characterizes both the plasma and interstitial albumin binding. the only adjustable parameter is the “renal clr” which is set to 0.353 (the fraction of admet & dmpk 7(1) (2019) 60-75 lipid soluble and extracellular solutes doi: 10.5599/admet.579 71 renal plasma cleared). note also that the “cap perm” is set to 1.0, indicating infinite capillary permeability (see below). figure 11 shows the good fit of the pkquest pbpk prediction to the experimental data of arancibia et. al.[20] following a bolus iv input. for some solutes the plasma albumin binding (= freepl) is poorly characterized or variable, and one treats freepl as another adjustable model parameter. using this identical pbpk model, similarly good fits were obtained to the experimental pk measurements for the ecs solutes edta, dtpa, morphine-6-glucuronide, morphine-3-glucuronide, mannitol and for the β-lactam antibiotics amoxicillin, piperacillin and cefatrizine [7]. figure 10. top panel of pkquest interface showing the parameters that characterize amoxicillin pbpk pharmacokinetics (see text for details). for all the solutes discussed above, the standard pbpk “flow-limited, well-stirred” model has been assumed for the tissue pk. this assumes that the free, unbound concentration in the tissue space is uniform for the entire tissue and is equal to the free plasma concentration that leaves the organ in the vein. this assumes, in essence, that the capillary permeability is infinite. this assumption may be incorrect because the ecs solutes cannot penetrate cell membranes and must traverse the capillary wall through the potentially limiting interendothelial slits [21]. figure 11. semi-log plot of comparison of experimental antecubital amoxicillin concentration (blue circles) and pkquest pbpk predicted antecubital concentration (red line) following a bolus iv input. there is no question that some ecs solutes have a capillary permeability limitation. the classical example is inulin, whose capillary permeability has been determined in animal models using a variety of david levitt admet & dmpk 7(1) (2019)60-75 72 experimental approaches [21]. pkquest has been modified to allow for a capillary permeability limitation characterized by the parameter “flcr”, the fraction of plasma solute that equilibrates with the interstitial space in one pass through the capillary. it is defined by: i i a v i a t fclr c c c c    (10) where ca, cv i and ct i are the free unbound concentrations in the artery, vein and tissue, respectively. if the capillary permeability is infinite, cv i = ct i and fclr = 1. if the permeability is 0, ca =cv and fclr = 0. in pkquest, the input parameter “cap perm” (fig. 9) is the fclrmuscle. the flcr of the other tissues are then set equal to proportional values determined from a literature review [7]. figure 12 shows a comparison of the permeability limited (top panel) versus infinite permeability (bottom panel) pkquest pbpk fit to the inulin data of odeh et al. [22]. it can be seen that although the permeability limitation only affects the early time data and is a relatively small effect, the addition of the limitation clearly improves the fit. the best fit was obtained with the fclrmuslc of 0.45 which corresponds to a capillary permeability (ps, eq. (11)) of skeletal muscle of 0.61 ml/min/100 gm, similar to that determined directly in animal studies [21]. this pbpk approach represents a new way to study capillary permeability and is the first human measurement of muscle capillary permeability for these solutes. figure 12.pkquest pbpk pharmacokinetics (red line) for inulin following a 5 minute constant iv infusion. the top panel is with a capillary permeability limitation (fclr muscle = 0.45) and the bottom panel is for an infinite permeability (fclr muscle = 1). admet & dmpk 7(1) (2019) 60-75 lipid soluble and extracellular solutes doi: 10.5599/admet.579 73 it had not been previously recognized that, as the albumin binding affinity increases, the capillary permeability should decrease. one can show that the parameter fclr i is described by the relation [3,4,7]: i i ifclr 1 exp( / )pf ps f   (11) where fp is the fraction free in plasma, p is the capillary permeability of the free, unbound solute, and s i and f i are the capillary surface area and flow rate for tissue i. as p approaches infinity, fclr approaches 1 and the solute becomes flow limited. it can be seen that solutes that have a high “intrinsic” p if they are unbound, may become capillary limited as the albumin binding affinity increases and fp approaches zero. this fp dependence of the capillary permeability was tested using the experimental pk of a series of βlactam antibiotics with fp varying from 0.8 for amoxicillin, to 0.52 for piperacillin, to 0.07 for flucloxacillin and 0.03 for dicloxacillin [7]. the quality of the pkquest pbpk fits to the experimental data with and without a capillary permeability limitation was compared. for flucloxacillin and dicloxacillin, the two solutes with the highest albumin binding, the fit clearly improved when a permeability limitation was introduced. figure 13 shows a comparison of the pkquest output for dicloxacillin with a permeability limitation (top panel, fclr muscle = 0.3) versus infinite permeability (bottom panel, fclr muscle = 1). again it can be seen that although the permeability limitation only affects the early time data and is a relatively small effect, the addition of the limitation clearly improves the fit. the estimated pbpk human ps values are similar to those of edta and mannitol, ecs solutes of similar size, measured directly in the cat [7]. figure 13.pkquest pbpk pharmacokinetics (red line) for dicloxacillin following a 30 minute constant iv infusion. the top panel is with a capillary permeability limitation (fclr muscle = 0.3) and the bottom panel is for an infinite permeability (fclr muscle = 1). david levitt admet & dmpk 7(1) (2019)60-75 74 the possibility of a capillary permeability limitation complicates the pk of the ecs solutes. as shown above, for solutes the size of the β-lactam antibiotics and fp greater than 0.3, it can be assumed that the permeability is infinite. for larger solutes (eg, inulin) or higher albumin binding affinity (eg, dicloxacillin), there may be significant permeability limitation that can be modeled by treating the pkquest capillary permeability parameter “fclr” as an additional adjustable parameter. conclusion it is clear from the above discussion that the hls and ecs represent the ideal solutes for pbpk analysis. the pk of these solutes can usually be accurately predicted using only zero (volatile anesthetics) or one adjustable parameter (the clearance). not only can one predict the blood or plasma pk, but having so few adjustable parameters increases ones confidence in the pbpk model, allowing the determination of the time dependent concentration in the different tissues and how the pk varies with, eg, exercise or body fat fraction. unfortunately, the hls and ecs solutes classes are limited, with the examples discussed above nearly exhausting the important drug classes. in contrast, for the much more common weak acid or base solutes, an accurate pbpk model requires detailed animal experimental measurement of the individual organ k i , and then an uncertain extrapolation to the human. these factors have significantly limited the confidence in the use of pbpk models for, eg, toxicological applications [23]. it does not seem to be widely recognized that the hls and ecs solutes are important exceptions to the usual problems associated with pbpk analysis and it is hoped that this discussion will remedy this. this discussion has focused on only one aspect of pkquest: the pbpk analysis of the hls and ecs solutes. pkquest also offers several other novel pbpk features, including: 1) the first incorporation of the use of antecubital vein sampling in a pbpk model [9]; and 2) the first ethanol pbpk model with a rigorous definition of the non-linear bioavailability [6]. in addition to the pbpk module, pkquest also has a number of other pk features. there is major emphasis on deconvolution which is a powerful and underutilized modeling approach. there are 6 different deconvolution routines available, each with its strengths and limitations. one novel deconvolution application is a general approach to determine the intestinal permeability during normal human drug absorption, with application of this approach to 90 different drugs [13]. all of these features are discussed with detailed worked examples in the freely distributed textbook “computer assisted human pharmacokinetics” [1]. it is hoped that the interested reader will download the free software and textbook from www.pkquest.com and try out these and other pkquest pk features. references [1] d.g. levitt. computer assisted human pharmacokinetics: non-compartmental, deconvolution, physiologically based, intestinal absorption, non-linear; www.pkquest.com. 2017. [2] p. poulin. a paradigm shift in pharmacokinetic-pharmacodynamic (pkpd) modeling: rule of thumb for estimating free drug level in tissue compared with plasma to guide drug design. j pharm sci 104 (2015) 2359-2368. [3] d.g. levitt. pkquest: a general physiologically based pharmacokinetic model. introduction and application to propranolol. bmc clin pharmacol 2 (2002) 5. [4] d.g. levitt. pkquest: capillary permeability limitation and plasma protein binding application to human inulin, dicloxacillin and ceftriaxone pharmacokinetics. bmc clin pharmacol 2 (2002) 7. [5] d.g. levitt. pkquest: volatile solutes application to enflurane, nitrous oxide, halothane, methoxyflurane and toluene pharmacokinetics. bmc anesthesiol 2 (2002) 5. [6] d.g. levitt. pkquest: measurement of intestinal absorption and first pass metabolism application to human ethanol pharmacokinetics. bmc clin pharmacol 2 (2002) 4. http://www.pkquest.com/ http://www.pkquest.com/ admet & dmpk 7(1) (2019) 60-75 lipid soluble and extracellular solutes doi: 10.5599/admet.579 75 [7] d.g. levitt. the pharmacokinetics of the interstitial space in humans. bmc clin pharmacol 3 (2003) 3. [8] d.g. levitt. the use of a physiologically based pharmacokinetic model to evaluate deconvolution measurements of systemic absorption. bmc clin pharmacol 3 (2003) 1. [9] d.g. levitt. physiologically based pharmacokinetic modeling of arterial antecubital vein concentration difference. bmc clin pharmacol 4 (2004) 2. [10] d.g. levitt. heterogeneity of human adipose blood flow. bmc clin pharmacol 7 (2007) 1. [11] d.g. levitt. pkquest_java: free, interactive physiologically based pharmacokinetic software package and tutorial. bmc res notes 2 (2009) 158. [12] d.g. levitt. quantitative relationship between the octanol/water partition coefficient and the diffusion limitation of the exchange between adipose and blood. bmc clin pharmacol 10 (2010) 1. [13] d.g. levitt. quantitation of small intestinal permeability during normal human drug absorption. bmc pharmacol toxicol 14 (2013) 34. [14] n. yasuda, s.h. lockhart, e.i. eger, 2nd, r.b. weiskopf, b.h. johnson, b.a. freire, a. fassoulaki. kinetics of desflurane, isoflurane, and halothane in humans. anesthesiology 74 (1991) 489-498. [15] n. yasuda, s.h. lockhart, e.i. eger, 2nd, r.b. weiskopf, j. liu, m. laster, s. taheri, n.a. peterson. comparison of kinetics of sevoflurane and isoflurane in humans. anesth analg 72 (1991) 316-324. [16] e. johansson, a. ohlsson, j.e. lindgren, s. agurell, h. gillespie, l.e. hollister. single-dose kinetics of deuterium-labelled cannabinol in man after intravenous administration and smoking. biomed environ mass spectrom 14 (1987) 495-499. [17] d.g. levitt, t.w. schnider. human physiologically based pharmacokinetic model for propofol. bmc anesthesiol 5 (2005) 4. [18] f. servin, r. farinotti, j.p. haberer, j.m. desmonts. propofol infusion for maintenance of anesthesia in morbidly obese patients receiving nitrous oxide. a clinical and pharmacokinetic study. anesthesiology 78 (1993) 657-665. [19] d. mackay, a. fraser. kenneth mellanby review award. bioaccumulation of persistent organic chemicals: mechanisms and models. environ pollut 110 (2000) 375-391. [20] a. arancibia, j. guttmann, g. gonzalez, c. gonzalez. absorption and disposition kinetics of amoxicillin in normal human subjects. antimicrob agents chemother 17 (1980) 199-202. [21] c. crone, d.g. levitt. capillary permeability to small solutes. in: e.m. renkin, c.c. michell (eds.) the microcirculation, handbook of physiology, american physiological society, bethesda, md., 1984, pp. 411-466. [22] y.k. odeh, z. wang, t.i. ruo, t. wang, m.c. frederiksen, p.a. pospisil, a.j. atkinson, jr. simultaneous analysis of inulin and 15n2-urea kinetics in humans. clin pharmacol ther 53 (1993) 419-425. [23] y.m. tan, r.r. worley, j.a. leonard, j.w. fisher. challenges associated with applying physiologically based pharmacokinetic modeling for public health decision-making. toxicol sci 162 (2018) 341-348. ©2018 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ manuscript doi: 10.5599/admet.4.4.338 335 admet & dmpk 4(4) (2016) 335-360; doi: 10.5599/admet.4.4.338 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper investigation of biopharmaceutical and physicochemical drug properties suitable for orally disintegrating tablets asami ono*, takumi tomono 1 , takuo ogihara 1 , katsuhide terada 2,a , and kiyohiko sugano 2 asahi kasei pharma, 632-1 mifuku, izunokuni, shizuoka 410-2321, japan. 1 laboratory of clinical pharmacokinetics, graduate school of pharmaceutical sciences, takasaki university of health and welfare, 60 nakaorui takasaki, gunma 370-0033, japan. 2 department of pharmaceutics, faculty of pharmaceutical sciences, toho university, 2-2-1 miyama, funabashi, chiba 274-8510, japan. a present address: laboratory of molecular pharmaceutics and technology, faculty of pharmacy, takasaki university of health and welfare, 60 nakaorui takasaki, gunma 370-0033, japan. *corresponding author: e-mail: ono.ar@om.asahi-kasei.co.jp; tel.: +81-558-76-7061; fax: +81-558-76-7137 received: september 14, 2016; revised: november 18, 2016; published: december 26, 2016 abstract the purpose of this study was to evaluate the biopharmaceutical and physicochemical drug properties suitable for orally disintegrating tablets (odts). the molecular weight (mw), polar surface area (psa), hydrogen bond donor (hbd) and acceptor (hba) numbers, net charge at ph 7.4, log d6.5, the highest dose strength, solubility in water, dose number, and elimination t1/2 of 57 odt drugs and 113 drugs of immediate-release (ir) formulations were compared. these drugs were classified according to the biopharmaceutical classification system (bcs). a lower dose strength and a longer elimination t1/2 have been observed as characteristic properties of odts. the proportion of basic drugs was higher in the odts than in the ir formulations. a significant difference was not observed between the odt and the ir formulation for mw, psa, hbd, hba, log d6.5, solubility in water, and dose number. the distributions of the odts and ir formulations among each bcs class were similar, suggesting that an odt can be developed regardless of the bcs class of a drug. keywords orally disintegrating tablet (odt); bioequivalence; biopharmaceutics classification system (bcs). introduction an orally disintegrating tablet (odt) improves patient compliance because it can be taken without water, does not cause dysphagia, and can prevent patients from spitting out their medication [1,2]. many pharmaceutical companies are working on the development of odts. an odt should be bioequivalent to a corresponding standard formulation, e.g. an immediate-release (ir) formulation. an odt disintegrates and dissolves rapidly in the oral cavity within 30 seconds. therefore, the dissolution rate may differ significantly between odts and ir formulations. when the dissolution rates are different between the odt and the ir formulation, the risk to fail in a bioequivalence (be) study is high. therefore, during the development of an odt, it is important to assess the risk of failing in a clinical be study. it is preferable to reduce the risk of failing a clinical be study because such studies are expensive, time consuming, and a burden to healthy http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:ono.ar@om.asahi-kasei.co.jp asami ono et al. admet & dmpk 4(4) (2016) 335-360 336 volunteers. however, there has been no research investigating the biopharmaceutical properties of drugs suitable to be an odt. in the present study, several physicochemical and biopharmaceutical properties were selected and a survey was performed to compare the properties between the two types of oral dose formulations. methods drug list the physicochemical and biopharmaceutical properties of drugs marketed as odts and ir formulations were compared in this study. currently, the number of drugs developed as odt is largest in japan. in addition, all the odt formulations approved in japan were proved to be bioequivalent to the ir formulation with and without water intake. therefore, the odt formulations in japanese market were selected in this study. the list of the odt drugs was obtained from the pharmaceuticals and medicinal devices agency website (http://www.info.pmda.go.jp/psearch/html/menu_tenpu_base.html) (supplement table 1). the list of the ir formulations were selected from the top 200 pharmaceutical products in japan (supplement table 2) [3]. finally, 57 compounds for the odt and 113 compounds for the ir formulation were selected and analysed. the 25 compounds were overlapping between the lists of the odts and ir formulations. biopharmaceutical drug properties the physicochemical and biopharmaceutical properties of drugs related to the formulation design and oral bioavailability were selected [4,5]. the molecular weight (mw), polar surface area (psa), hydrogen bond donor (hbd) and acceptor (hba) numbers, dissociation constants (pka), and n-octanol/water distribution coefficients at ph6.5 (log d6.5) were calculated using acd percepta (acd/labs software v 14.0.0 (http://www.acdlabs.com/products/percepta/predictors.php)). the calculated pka and log d were used for all drugs, as the experimental values were not available for some drugs. net charge (nc) at ph 7.4 was represented as the weighted sum of the charge of each species. nc = (±0) ∙ 𝑓0 + (+1) ∙ 𝑓+ + (−1) ∙ 𝑓− + (+2) ∙ 𝑓++ + ⋯ (1) (𝑓0 + 𝑓+ + 𝑓− + 𝑓++ + ⋯ = 1) where f0 is the fraction of the undissociated species, f+ is that of +1 charged species, etc. each fraction was calculated by the pka and the ph (set to be 7.4 in this study) using the henderson-hasselbalch equation [6]. the nc equations for acids and bases containing up to three ionization centers are summarized in supplement table 3. when a drug was more than 50 % dissociated at ph 7.4, it was classified as an acid (nc < -0.5) or as a base (nc > 0.5). log d was calculated at ph 6.5 to estimate the permeability, whereas nc was calculated at ph 7.4 to discuss the pharmacokinetics after the absorption. the solubility in water, the highest dose strength, and the elimination t1/2 were obtained from the prescription information. when a reliable solubility figure was unavailable, a solubility value was assigned based on the solubility category defined by the japanese pharmacopeia (supplement table 4). provisional classification according to the biopharmaceutical classification system according to the biopharmaceutical classification system (bcs), drugs can be categorized into the four classes, i.e. high solubility/high permeability (class i), low solubility/high permeability (class ii), high solubility/low permeability (class iii), and low solubility/low permeability (class iv). moreover, bcs class ii drugs can be sub-classified into acid (class iia), base (class iib), and undissociated drugs (class iic) [7]. admet & dmpk 4(4) (2016) 335-360 biopharmaceutical and physchem properties for odt doi: 10.5599/admet.4.4.331 337 337 in the previous studies, solubility in water and calculated log d6.5 were used as the surrogates of solubility and permeability data to provide provisional bcs class [3,8]. the same approach was taken in this study. according to the official bcs guidance, the equilibrium solubility of a drug at the physiological gastrointestinal ph range (namely, ph 1.2 to ph 6.8 or 7.4) is required. however, the ph solubility profile data were not available for many drugs. therefore, solubility in water reported in the prescription information was used in this study. the dose number (do) is a dimensionless number expressed by the ratio of the dose and the maximum dissolved amount in the intestine. do was calculated by using eq. 2 [5]: 𝐷𝑜 = 𝑀 𝑆 × 𝑉 (2) where s is the solubility of a drug in water, v is the intestinal fluid volume (set to be 250 ml in this study), and m is the highest dose strength. drugs were defined as highly soluble when the do was ≤ 1. according to the official bcs guidance, to classify the permeability category, the fraction of a dose absorbed (fa%) in humans or caco-2 permeability data is required. however, due to the limited availability of these data, the permeability was classified based on the calculated log d6.5 in this study. the log d6.5 value of metoprolol was chosen as the criteria for high permeability [fa% in human, 95 % [9]; human effective permeability, 1.26 × 10 -4 cm/s [9]; log d6.5, -0.92 (acd percepta)]. statistical analysis a student’s t-test was used to evaluate the significance of difference between the odts and ir formulations in mw, psa, hbd, hba, nc at ph 7.4, log d6.5, the highest dose strength, solubility in water, do, and elimination t1/2. one-way analysis of variance was used to evaluate the significance of difference between acids, bases, and undissociated drugs in elimination t1/2 for the odts and ir formulations, respectively. a minimum p value of 0.05 was used as the significance level for all tests. microsoft excel 2010 (microsoft) was used for statistical analysis. results mw of free form, psa, hbd, hba, nc at ph 7.4, log d6.5, the highest dose strength, solubility in water, do, and elimination t1/2 are shown in table 1. some calculated pka and log d6.5 might have a margin of error about 1 log unit (e.g. pkas of cetirizine (6.7 (b); 7.7 (b) [10]), domperidone (9.0 (b); 7.1 (b) [10]), glimepiride (5.1 (a); 6.2 (a) [11]), log d6.5 of famotidine (-2.14; -1.3 [11]), glimepiride (1.51; 3.0 [11]) (calculated values; experimental values). nc at ph 7.4 was significantly higher in the odt than that for the ir formulation (p = 0.02) (figure 1e). the percentages of acid, base, and undissociated drugs for the odts were 13, 41 and 45 %, respectively (2 % unclassifiable). the corresponding percentages in the ir formulations were 30, 30 and 38 %, respectively (2 % unclassifiable). the highest dose strength of the odts was significantly lower than that of the ir formulations (p = 0.01) (figure 1g). the medians of the highest dose strength in the odts and ir formulations were 10 mg and 30 mg, respectively. the maximum values of the highest dose strength in the odts and ir formulations were 200 mg and 900 mg, respectively (table 2). the elimination t1/2 of the odts tended to be longer than that of the ir formulations (p = 0.07) (figure 1j). the medians of t1/2 in the odts and ir formulations were 6.7 h, and 3.3 h, respectively (table 2). a significant difference was not observed between the odts and the ir formulations for mw, psa, hbd, hba, log d6.5, solubility in water, and do (figure 1a, b, c, d, f, h, i). the distribution of the odts and ir formulations among each bcs class were similar (figure 2). the percentages of bcs class i, class ii, class iii, and class iv for the odt were 52, 29, 20 and 0 %, respectively asami ono et al. admet & dmpk 4(4) (2016) 335-360 338 (2 % unclassifiable). the corresponding percentages for the ir formulation were 48, 30, 16 and 4 %, respectively (2 % unclassifiable). the distribution of bcs ii subclass was also similar (figure 2). the percentages of bcs class iia, class iib, and class iic for the odt were 7, 12 and 9 %, respectively, whereas the corresponding percentages for the ir formulation were 10, 10 and 11 %, respectively. figure 1. box and whisker plots of biopharmaceutical properties of odt and ir formulation drugs ; the bottom and top of the box are the first and third quartiles, the band inside the box is the median, and the ends of the whiskers are the minimum and maximum. (a) mw of free form. (b) polar surface area. (c) hydrogen bond donor number. (d) hydrogen bond acceptor number. (e) net charge at ph 7.4. (f) log d6.5. (g) the highest dose strength. (h) solubility in water. (i) dose number. (j) elimination t1/2. *p < 0.05. 0 200 400 600 800 1000 1200 1400 1600 odt ir formulation m w o f fr e e f o rm 0 100 200 300 400 500 600 odt ir formulation p s a 0 5 10 15 20 odt ir formulation h b d -4 -3 -2 -1 0 1 2 3 odt ir formulation n c a t p h 7 .4 -10 -5 0 5 10 15 odt ir formulation lo g d 6 .5 0.001 0.01 0.1 1 10 100 1000 odt ir formulation d o se ( m g ) 0.001 0.01 0.1 1 10 100 1000 10000 odt ir formulation s o lu b il it y ( m g / m l) -8 -6 -4 -2 0 2 odt ir formulation lo g d o 0 20 40 60 80 100 120 140 odt ir formulation t 1 /2 (h ) * * (a) (b) (c) (e) (f) (g) (h) (i) (j) 0 5 10 15 20 25 30 odt ir formulation h b a (d) admet & dmpk 4(4) (2016) 335-360 biopharmaceutical and physchem properties for odt doi: 10.5599/admet.4.4.338 339 table 1. mw of free form, psa, hbd, hba, nc at ph 7.4, log d6.5, the highest dose strength, solubility in water, do, elimination t1/2, and bcs class of drugs used in this study. drug mw (free form) psa b hbd b hba b nc a ph 7.4 pka b log d6.5 b highest dose strength (mg) c solubility in water (mg/ml) c do t1/2 (h) c bcs class acarbose d,e 646 321 14 19 0.0 – -4.88 100 1429 0.00028 3.2 3 acetaminophen 151 49 2 3 0.0 – 0.40 500 10 0.2 2.9 1 acetylsalicylic acid 180 64 1 4 -1.0 3.5 (a) -1.39 100 3.33 0.12 0.4 3 acyclovir 225 115 4 8 0.0 – -1.23 400 1 1.6 2.4 4 alendronate sodium hydrate 249 181 7 8 -2.0 1.7 (a), 2.7 (a), 6.1 (a), 10.3 (b) -8.44 35 25.6 0.00546 1.5 3 alfacalcidol 401 40 2 2 0.0 – 7.58 0.03 0.01 0.0012 17.6 1 allopurinol 136 66 2 5 0.0 – -0.70 100 0.1 4 1.6 2 ambroxol hydrochloride 378 58 4 3 1.0 8.7 (b) 0.48 45 26.8 0.00672 11.2 1 amlodipine besylate 409 100 3 7 1.0 9.0 (b) 1.09 10 2.22 0.018 36.2 1 aripiprazole 448 45 1 5 0.6 7.7 (b) 4.42 24 0.01 9.6 59.6 2 atenolol 266 85 4 5 1.0 9.4 (b) -2.53 50 1 0.2 10.8 3 atorvastatin calcium 559 112 4 7 -1.0 4.3 (a) 2.06 10 0.145 0.276 10.8 1 azithromycin hydrate 749 180 5 14 1.8 8.2 (b), 8.6 (b) -0.19 600 0.01 240 61.9 2 azulene sulfonate sodium hydrate 278 63 1 3 -1.0 1.7 (a) -2.03 2 10 0.0008 2.45 3 benidipine hydrochloride 506 114 1 9 0.8 8.0 (b) 3.54 8 0.01 3.2 1.0 2 bepotastine besylate 389 63 1 5 0.0 4.4 (a), 8.9 (b) 0.86 10 23.3 0.00172 2.5 1 beraprost sodium 398 87 3 5 -1.0 4.8 (a) 1.42 0.04 833 1.92 × 10 -7 1.1 1 bicalutamide 430 116 2 6 0.0 – 2.53 80 0.01 32 4.9 2 bisoprolol fumarate 325 60 2 5 1.0 9.4 (b) -0.55 5 1250 1.6 × 10 -5 8.6 1 brotizolam 394 71 0 4 0.0 – 2.80 0.25 0.01 0.1 7 1 cabergoline 452 72 2 7 1.0 9.4 (b) 0.40 1 0.01 0.4 43 1 camostat mesylate 398 137 4 9 1.0 9.1 (b) -0.53 100 45. 5 0.0088 1.7 1 candesartan cilexetil 440 119 2 9 -2.0 2.1 (a), 4.2 (a) 0.29 12 0.01 4.8 2.2 2 asami ono et al. admet & dmpk 4(4) (2016) 335-360 340 table 1. (continued) drug mw (free form) psa b hbd b hba b nc a ph 7.4 pka b log d6.5 b highest dose strength (mg) c solubility in water (mg/ml) c do t1/2 (h) c bcs class carvedilol 406 76 3 6 0.9 8.2 (b) 2.20 20 0.01 8 7.7 2 cefcapene pivoxil hydrochloride 594 257 3 16 0.0 5.9 (a), 9.0 (b) -1.59 100 1.88 0.213 1.1 3 cefdinir 395 212 5 10 -1.0 2.8 (a) -3.92 100 0.381 1.05 1.6 4 cefditoren pivoxil 621 257 3 13 -0.2 8.1 (a) 2.45 100 0.01 40 1.1 2 cefotiam hexetil hydrochloride 696 276 3 16 0.8 8.0 (b) -0.14 200 1000 0.0008 0.8 1 cetirizine hydrochloride 389 53 1 5 -0.8 3.5 (a), 6.7 (b) -0.09 10 1000 4.0 × 10 -5 7.5 1 cilostazol 369 82 1 7 0.0 – 3.01 100 0.01 40 10.1 2 clarithromycin 748 183 4 14 0.9 8.2 (b) 1.56 200 0.01 80 4.4 2 cyclosporin a 1203 279 5 23 0.0 – 1.80 50 3.38 0.0592 1.6 1 desmopressin acetate hydrate 1069 486 18 26 0.0 – -7.34 0.24 33 2.9 × 10 -5 2 3 diclofenac sodium 296 49 2 3 -1.0 4.2 (a) 2.17 25 17.2 0.0058 1.2 1 dienogest 311 61 1 3 0.0 – 2.64 1 0.01 0.4 8.0 1 domperidone 426 68 2 7 1.0 9.0 (b) 2.26 10 0.01 4 0.9 2 donepezil hydrochloride 379 39 0 4 1.0 8.8 (b) 1.97 10 10 0.004 70.7 1 doxazosin mesylate 451 112 2 10 0.1 6.5 (b) 1.20 4 1 0.016 11.8 1 doxifluridine 246 99 3 7 -0.4 7.6 (a) -1.13 200 33 0.0242 0.8 f 3 droxidopa 213 124 6 6 -0.1 2.1 (a), 8.3 (b) -3.57 200 2.4 0.333 2 3 ebastine 470 30 0 3 0.9 8.2 (b) 5.49 10 0.01 4 17.6 2 enalapril maleate 376 96 2 7 -1.0 3.1 (a) -1.06 10 21 0.00191 6.1 3 epalrestat 319 115 1 4 -1.0 3.6 (a) -1.04 50 0.009 22.2 1.8 4 eperisone hydrochloride 259 20 0 2 0.9 8.5 (b) 2.02 50 200 0.001 1.6 1 epinastine hydrochloride 249 42 2 3 1.0 12.0 (b) 0.77 20 133 0.0006 9.2 1 ethyl icosapentate 331 26 0 2 0.0 – 6.65 900 0.01 360 58.9 2 etizolam 343 71 0 4 0.0 – 2.87 1 0.01 0.4 6.3 1 admet & dmpk 4(4) (2016) 335-360 biopharmaceutical and physchem properties for odt doi: 10.5599/admet.4.4.331 341 table 1. (continued) drug mw (free form) psa b hbd b hba b nc a ph 7.4 pka b log d6.5 b highest dose strength (mg) c solubility in water (mg/ml) c do t1/2 (h) c bcs class famotidine 337 238 8 9 0.8 7.9 (b) -2.14 20 0.741 0.108 3.1 3 fexofenadine hydrochloride 502 81 3 5 0.0 4.4 (a), 9.4 (b) 2.43 60 1.43 0.168 19 1 fluconazole 306 82 1 7 0.0 – 0.70 400 1 1.6 30 2 flurbiprofen 244 37 1 2 -1.0 4.1 (a) 1.48 40 0.01 16 2.7 2 fluvastatin sodium 411 83 3 5 -1.0 4.3 (a) 1.33 30 82.0 0.00146 1.3 1 fluvoxamine maleate 318 57 2 4 1.0 9.4 (b) 0.36 75 14 0.0214 14.1 1 fursultiamine 399 152 3 7 0.0 – 2.05 50 200 0.001 14.7 1 galantamine hydrobromide 287 42 1 4 0.8 7.9 (b) 0.12 12 33 0.00145 6.8 1 gefitinib 447 69 1 7 0.3 7.0 (b) 3.07 250 0.01 100 30.1 2 gimeracil 146 53 2 3 0.1 6.5 (a), 12.6 (b) -2.58 7.25 1.74 0.0167 3 3 glimepiride 491 133 3 9 -1.0 5.1 (a) 1.51 3 0.01 1.2 5.8 2 granisetron hydrochloride 312 50 1 5 1.0 10.5 (b) -0.95 2 588 1.36 × 10 -5 5.3 3 hydrochlorothiazide 298 135 4 7 0.0 – 0.01 12.5 0.1 0.5 9.1 1 imatinib mesylate 494 86 2 8 0.6 7.6 (b) 2.00 200 1300 0.000615 15.9 1 imidafenacin 319 61 2 4 0.6 7.6 (b) 1.68 0.1 0.01 0.04 3.1 1 imidapril hydrochloride 405 116 2 9 -1.0 2.4 (a) -2.51 10 49.3 0.000812 1.7 3 irsogladine maleate 359 96 1 6 0.0 – 1.81 4 0.01 1.6 128 2 itraconazole 706 101 0 12 0.1 6.5 (b) 4.67 200 0.01 80 27.9 2 ketoprofen 254 54 1 3 -1.0 4.2 (a) 0.87 75 0.01 30 1.6 2 ketotifen fumarate 309 49 0 2 1.0 8.8 (b) 1.79 1 1 0.004 6.7 1 l-carbocysteine 179 126 4 5 -1.0 2.1 (a), 3.8 (a), 8.8 (b) -3.80 500 0.1 20 1.6 4 lafutidine 432 104 1 7 0.4 7.2 (b) 0.56 10 0.01 4 1.6 2 lansoprazole 369 87 1 5 0.0 – 2.40 30 0.0323 3.72 1.4 2 asami ono et al. admet & dmpk 4(4) (2016) 335-360 342 table 1. (continued) drug mw (free form) psa b hbd b hba b nc a ph 7.4 pka b log d6.5 b highest dose strength (mg) c solubility in water (mg/ml) c do t1/2 (h) c bcs class levofloxacin 361 73 1 7 -0.5 5.2 (a), 7.4 (b) -1.76 500 16. 7 0.12 7.9 3 limaprost alfadex 366 95 3 5 -1.0 4.6 (a) 0.85 0.005 370 5.4 × 10 -8 0.5 1 loratadine 383 42 0 4 0.0 – 5.32 10 0.00303 13.2 20.1 2 losartan potassium 423 93 2 7 -1.0 4.2 (a) 1.77 100 1000 0.0004 1.8 1 loxoprofen sodium hydrate 246 54 1 3 -1.0 4.4 (a) 0.37 60 1000 0.00024 1.2 1 manidipine hydrochloride 611 117 1 10 0.1 6.1 (b) 5.29 20 3.88 0.0206 1.5 1 mecobalamin 1344 nd nd nd – nd nd 0.5 12.5 0.00016 12.5 uc meloxicam 351 136 2 7 -1.0 4.5 (a) 0.29 10 0.01 4 27.6 2 memantine hydrochloride 179 26 2 1 1.0 10.8 (b) 0.40 20 33 0.00242 53.6 1 menatetrenone 445 34 0 2 0.0 – 9.55 15 0.01 6 3.9 2 mesalazine 153 84 4 4 -1.0 1.9 (a) -1.85 500 1 2 6.4 4 methylmethionine sulfonium chloride 164 nd nd nd – nd nd 25 1000 0.0001 3.2 uc mexiletine hydrochloride 179 35 2 2 0.9 8.6 (b) 0.39 100 833 0.00048 9.4 1 midodrine hydrochloride 254 94 4 6 0.7 7.8 (b) -1.24 2 138 5.80 × 10 -5 2.4 3 miglitol 207 104 5 6 0.1 6.5 (b) -2.27 75 791 0.000379 2.0 3 montelukast sodium 586 96 2 4 -1.0 4.8 (a) 5.82 10 200 0.0002 4.3 1 mosapride citrate hydrate 422 77 3 6 0.1 6.2 (b) 2.89 5 0.01 2 2 2 naftopizil 393 45 1 5 0.2 6.9 (b) 3.65 75 0.01 30 11.2 2 nicardipine hydrochloride 480 114 1 9 0.4 7.3 (b) 3.94 20 6 0.0133 1.5 1 nicergoline 484 57 0 6 0.1 6.3 (b) 4.29 5 0.01 2 3.3 2 nicorandil 211 97 1 7 0.0 – 0.93 5 10 0.002 0.8 1 nifedipine 346 110 1 8 0.0 – 3.45 10 0.01 4 1.0 2 nilvadipine 385 134 1 9 0.0 – 3.23 4 0.01 1.6 10.7 2 nizatidine 331 140 2 7 0.5 7.3 (b) -0.61 150 17.3 0.0347 1.7 1 admet & dmpk 4(4) (2016) 335-360 biopharmaceutical and physchem properties for odt doi: 10.5599/admet.4.4.331 343 table 1. (continued) drug mw (free form) psa b hbd b hba b nc a ph 7.4 pka b log d6.5 b highest dose strength (mg) c solubility in water (mg/ml) c do t1/2 (h) c bcs class olanzapine 312 56 1 4 1.0 10.4 (b) -0.35 10 0.01 4 30.6 2 olopatadine hydrochloride 337 50 1 4 0.0 4.3 (a), 9.2 (b) 1.54 5 10 0.002 0.8 1 ondansetron 293 40 0 4 0.6 7.5 (b) 1.76 4 0.01 1.6 4.5 2 oseltamivir phosphate 312 91 3 6 1.0 8.8 (b) -0.19 75 521 0.000576 7 1 oteracil potassium 157 108 3 7 -1.0 -3.1 (a) -5.54 24.5 7.42 0.0132 3 3 paroxetine hydrochloride hydrate 329 40 1 4 1.0 9.7 (b) 0.84 20 2 0.04 13.6 1 pergolide mesylate 314 44 1 2 0.8 8.0 (b) 2.39 0.25 5 0.0002 8.8 1 perindopril erbumine 368 96 2 7 -1.0 3.2 (a) 0.07 4 500 3.2 × 10 -5 0.8 1 pilsicainide hydrochloride hydrate 272 32 1 3 1.0 10.4 (b) -0.12 50 100 0.002 4.4 1 pioglitazone hydrochloride 356 94 1 5 -0.9 6.3 (a) 2.16 30 0.01 12 6.7 2 pitavastatin calcium 421 91 3 5 -1.0 4.2 (a) 1.17 4 0.1 0.16 9.3 1 polaprezinc nd nd nd nd – nd nd 75 0.01 30 2.2 uc pramipexole hydrochloride hydrate 211 79 3 3 1.0 9.5(b) -1.03 0.5 1000 2.0 × 10 -6 8.4 3 pranlukast hydrate 482 119 2 9 -1.0 5.3 (a) 2.74 112.5 0.0012 375 1.2 2 pravastatin sodium hydrate 425 124 4 7 -1.0 4.3 (a) 0.06 10 100 0.0004 2.7 1 procaterol hydrochloride 290 82 4 5 1.0 9.4 (b) -1.47 0.05 50 4.0 × 10 -6 3.8 3 propiverine hydrochloride 367 39 0 4 0.7 7.8 (b) 4.10 20 124 0.000645 10.7 1 quetiapine fumarate 384 74 1 5 0.2 6.7 (b) 1.92 200 3.38 0.2364 3.5 1 ramosetron hydrochloride 279 51 1 4 0.4 7.3 (b) 1.80 0.005 575 3.48 × 10 -8 7 1 ranitidine hydrochloride 314 112 2 7 0.9 8.4 (b) -1.47 150 1429 0.00042 2.7 3 rebamipide 371 96 3 6 -1.0 3.4 (a) -0.75 100 0.006 66.7 1.9 2 risedronate sodium hydrate 283 168 5 8 -3.0 1.4 (a), 2.4 (a), 6.0 (a) -8.38 75 33 0.00909 1.5 3 risperidone 410 62 0 6 0.8 8.1 (b) 0.99 2 0.01 0.8 0.3 1 asami ono et al. admet & dmpk 4(4) (2016) 335-360 344 table 1. (continued) drug mw (free form) psa b hbd b hba b nc a ph 7.4 pka b log d6.5 b highest dose strength (mg) c solubility in water (mg/ml) c do t1/2 (h) c bcs class rizatriptan benzoate 269 50 1 5 1.0 9.5 (b) -0.70 10 34 0.00118 1.7 1 sarpogrelate hydrochloride 430 85 1 7 -0.2 4.3 (a), 8.1 (b) 1.37 100 1 0.4 0.8 1 selegiline hydrochloride 187 3 0 1 0.6 7.5 (b) 1.81 2.5 1667 6.0 × 10 -6 5.3 1 sertraline hydrochloride 306 12 1 1 1.0 9.5 (b) 2.45 100 2.70 0.148 25.1 1 sildenafil citrate 475 118 1 10 0.0 – 1.85 50 1 0.2 1.6 1 simvastatin 419 73 1 5 0.0 – 4.60 20 0.01 8 2.3 2 solifenacin succinate 362 171 5 12 1.0 9.0 (b) 1.36 5 610 3.28 × 10 -5 46.5 1 sultamicillin tosilate 595 216 3 13 0.2 6.8 (b) -0.46 375 0.1 15 1.0 2 tacrolimus hydrate 804 178 3 13 0.0 – 4.10 5 0.01 2 32.0 2 taltirelin hydrate 405 171 5 12 0.2 6.7 (b) -2.73 5 208 9.6 × 10 -5 2 3 tamoxifen citrate 372 12 0 2 1.0 8.7 (b) 4.68 20 0.09 0.889 27.2 1 tamsulosin hydrochloride 409 108 3 7 1.0 8.8 (b) -0.06 0.2 11.8 6.76 × 10 -5 11.7 1 tegafur 200 59 1 5 -0.4 7.6 (a) -0.43 25 16.8 0.00595 1.9 1 temocapril hydrochloride 477 149 2 7 -1.0 3.7 (a) 0.62 4 0.719 0.0223 0.2 1 teprenone 331 17 0 1 0.0 – 7.40 50 8.32 0.0240 1.8 1 terbinafine hydrochloride 291 3 0 1 0.2 6.9 (b) 5.45 125 5.07 0.0987 6.2 1 ticlopidine hydrochloride 264 31 0 1 0.3 7.1 (b) 3.41 100 58.8 0.0068 1.6 1 tocopherol nicotinate 536 48 0 4 0.0 – 11.14 200 0.01 80 4.3 2 tramadol hydrochloride 263 33 1 3 1.0 9.6 (b) -0.18 50 100 0.002 5.7 1 tulobuterol hydrochloride 228 32 2 2 1.0 9.6 (b) -0.17 1 714 5.6 × 10 -6 3.2 1 ursodeoxycholic acid 393 78 3 4 -1.0 4.8 (a) 2.01 100 0.01 40 1.1 2 valacyclovir hydrochloride 324 147 5 10 0.7 7.8 (b) -1.78 500 100 0.02 3.0 3 valproate sodium 144 37 1 2 -1.0 4.8 (a) 0.97 200 1000 0.0008 9.5 1 valsartan 436 112 2 8 -2.0 3.6 (a), 4.2 (a) -0.71 160 0.17 3.76 7.7 2 admet & dmpk 4(4) (2016) 335-360 biopharmaceutical and physchem properties for odt doi: 10.5599/admet.4.4.331 345 table 1. (continued) drug mw (free form) psa b hbd b hba b nc a ph 7.4 pka b log d6.5 b highest dose strength (mg) c solubility in water (mg/ml) c do t1/2 (h) c bcs class voglibose 267 154 8 8 0.2 6.8 (b) -3.57 0.3 1000 1.2 × 10 -6 5.3 3 zolmitriptan 287 57 2 5 1.0 9.5 (b) -0.29 2.5 0.1 0.1 2.9 1 zolpidem tartrate 307 38 0 4 0.2 6.8 (b) 2.78 10 8.9 0.00449 1.9 1 zonizamide 212 95 2 5 0.0 – 0.45 25 0.270 0.37 119.1 1 nd = no data. uc = unclassifiable. a net charge at ph 7.4 calculated by eq. 1. b calculated value (acd/labs software v 14.0.0). c data from prescription information (supplement tables 1 and 2) otherwise noted. d underscored drugs correspond to odts. e drugs in italics correspond to ir formulations. f reference [12] admet & dmpk 4(4) (2016) 335-360 biopharmaceutical and physchem properties for odt doi: 10.5599/admet.4.4.338 346 table 2. minimum, maximum, and median of biopharmaceutical properties of odt and ir formulation drugs. biopharmaceutical properties odt ir formulation mw of free form n 56 113 median 359 371 min. – max. 146 – 1069 136 – 1344 psa n 56 111 median 70 84 min. – max. 3 – 486 3 – 321 hbd n 56 111 median 1 2 min. – max. 0 – 18 0 – 14 hba n 56 111 median 5 6 min. – max. 1 – 26 1 – 23 nc a ph 7.4 n 56 111 median 0.2 0.0 min. – max. -2.0 – 1.0 -3.0 – 1.8 log d6.5 n 56 111 median 1.02 0.97 min. – max. -7.34 – 5.49 -8.44 – 11.14 the highest dose strength (mg) n 57 113 median 10 30 min. – max. 0.005 – 200 0.003 – 900 solubility in water (mg/ml) n 57 113 median 1.43 3.33 min. – max. 0.003 – 1667 0.0012 – 1429 do n 57 113 median 0.1 0.04 min. – max. 3.48 × 10 -8 – 66.7 5.4 × 10 -8 – 375 elimination t1/2 (h) n 57 113 median 6.7 3.3 min. – max. 0.9 – 128 0.2 – 70.7 a net charge at ph 7.4 calculated by eq. 1. figure 2. provisional bcs classification of drugs in odts and ir formulations. 0 10 20 30 40 50 60 p e rc e n ta g e o f d ru g s w it h in b c s c la ss odt ir formulation admet & dmpk 4(4) (2016) 335-360 biopharmaceutical and physchem properties for odt doi: 10.5599/admet.4.4.338 347 discussion a significant difference was seen in the highest dose strength between the odt and ir formulation. the odt would have to have a feasible tablet size and drug loading [13,14]. therefore, a drug with a high dose strength (>200 mg) would be less suitable an odt. there was no significant difference in the solubility in water and do between the odt and ir formulation. this result suggests that many drugs can be developed as an odt regardless of their solubility and do (e.g., bicalutamide: s = 0.01 mg/ml, do = 32, cilostazole: s = 0.01 mg/ml, do = 40). furthermore, an odt shows be with a corresponding ir formulation with and without water intake. therefore, even though it is counterintuitive, water intake may have little effect on the dissolution and oral absorption of low solubility drugs. previously, sumesen et al. reported that the oral absorption of danazol was not significantly altered when administered together with 1000 ml of water compared to when administered with 200 ml [15]. danazol, which has poor water solubility (0.2 g/ml) and high permeability (log d6.5 = 4.5), is a typical bcs class ii drug [16]. the drugs that have been developed as odts tended to have a longer elimination t1/2. previously, we reported that the elimination t1/2 of drugs influence the be of cmax [17]. for the drugs with high permeability and short elimination t1/2, be of cmax between two formulations with different dissolution rates would become more difficult to prove. this point has been suggested by several articles [18-21]. a drug with a long elimination t1/2 might be suitable for an odt. the proportion of bases was larger in the odt than in the ir formulation. it is well known that the basic lipophilic drugs have a large distribution volume, and a long elimination t1/2 due to wide tissue distribution [22]. a significant difference was seen in elimination t1/2 between acids, bases, and undissociated drugs in the ir formulation (p = 0.04) (figure 3). figure 3. box and whisker plot of elimination t1/2 of acid, base, and undissociated drugs the biowaiver schemes (bws) has been discussed based on bcs proposed by amidon et al. in 1995 [5]. in 2000, the us food and drug administration (fda) adopted the bcs-bws [23]. the guideline allows bcs class i drugs which show rapid dissolution (>85 % dissolution in 30 min) to waive clinical be studies. the world health organization (who) and other regulatory agencies followed the bcs-bws [24-27]. however, odt ir formulation 0 20 40 60 80 100 120 140 t 1 /2 (h ) asami ono et al. admet & dmpk 4(4) (2016) 335-360 348 there are some differences among these guidelines, e.g. solubility ph range, criteria for high permeability, and definition of the dose used for the do calculation [28]. one of the most significant differences is about the biowaiver for bcs class iii drugs. who, european union, and canada accept biowaiver for bcs class iii drugs which show very rapid dissolution (>85 % dissolution in 15 min), while only bcs class i drugs are eligible for biowaiver in the us fda and korea fda guideline. however, there are several computer simulation and experimental studies on the biowaiver for bcs class iii drugs, suggesting that bcs class iii drugs are suitable for biowaiver [20,21,29-33]. in addition, it has been pointed out that many bcs class iii drugs show be even when the dissolution profiles are different between the test and reference drugs, for example famotidine, hydrochlorothiazide, and cimetidine [33,34]. moreover, who adopts the possibility of biowaiver for bcs class iia drugs with low solubility at acidic ph and high solubility at neutral ph that are absorbed completely. however, it was reported that, in the case of ibuprofen, the typically bcs class iia drug, the be of cmax is more sensitive to the difference of dissolution rates [35,36]. the bcs class distribution of odts and ir formulations may reflect their non-be risk due to the difference of the dissolution rates. based on the bcs-bws, be is most easily established for bcs class i drugs. however, no difference in the distribution of the bcs classes, including subclasses, between the odts and ir formulations was observed in this study. this result may suggest that a bcs class i drug would not necessarily be suitable to show be. ramirez et al. reported that in the 124 clinical be studies there is no difference in the number of subjects in the be study and the interand intra-subject variability for cmax or auc between four bcs classes [37]. all the bcs classes drugs have the risk of non-be. their results also showed that all of the bioinequivalent parameters in bcs class i drugs was cmax, but not auc. the results of the present study suggest that the elimination t1/2 would affect the success rate of a be study more significantly than the bcs class of a drug (e.g. ibuprofen). in conclusion, drugs with a lower dose strength (<10 mg) and a longer elimination t1/2 (>6.5 h) were suggested to be more suitable for an odt. however, the distributions of the odt and ir formulation among each bcs class were similar, suggesting that the bcs classes are irrelevant to the development risk of odts. references [1] s. v. sastry, j. r. nyshadham, j. a. fix. recent technological advances in oral drug delivery – a review. pharm. sci. technolo. today 3 (2000) 138-145. [2] d. hobbs, j. karagianis, t. treuer, j. raskin. an in vitro analysis of disintegration times of different formulations of olanzapine orodispersible tablet: a preliminary report. drugs r d 13 (2013) 281-288. [3] t. takagi, c. ramachandran, m. bermejo, s. yamashita, l. x. yu, g. l. amidon. a provisional biopharmaceutical classification of the top 200 oral drug products in the united states, great britain, spain, and japan. mol. pharm. 3 (2006) 631-643. [4] c. a. lipinski, f. lombardo, b. w. dominy, p. j. feeney. experimental and computational approaches to estimate solubility and permeability in drug discovery and development settings. advanced drug delivery reviews 64 (2012) 4-17. [5] g. amidon, h. lennernäs, v. shah, j. crison. a theoretical basis for a biopharmaceutic drug classification: the correlation of in vitro drug product dissolution and in vivo bioavailability. pharm. res. 12 (1995) 413-420. [6] s. k. poole, s. patel, k. dehring, h. workman, c. f. poole. determination of acid dissociation constants by capillary electrophoresis. j. chromatogr. a 1037 (2004) 445-454. [7] y. tsume, d. m. mudie, p. langguth, g. e. amidon, g. l. amidon. the biopharmaceutics classification system: subclasses for in vivo predictive dissolution (ipd) methodology and ivivc. eur. j. pharm. sci. 57 (2014) 152-163. admet & dmpk 4(4) (2016) 335-360 biopharmaceutical and physchem properties for odt doi: 10.5599/admet.4.4.338 349 [8] n. a. kasim, m. whitehouse, c. ramachandran, m. bermejo, h. lennernäs, a. s. hussain, h. e. junginger, s. a. stavchansky, k. k. midha, v. p. shah, g. l. amidon. molecular properties of who essential drugs and provisional biopharmaceutical classification. mol. pharm. 1 (2003) 85-96. [9] h. lennernäs. intestinal permeability and its relevance for absorption and elimination. xenobiotica 37 (2007) 1015-1051. [10] a. avdeef. absorption and drug development, second edition. john wiley & sons, inc (2012) [11] k. sugano. biopharmaceutics modeling and simulations. wiley (2012) [12] b. reigner, j. verweij, l. dirix, j. cassidy, c. twelves, d. allman, e. weidekamm, b. roos, l. banken, m. utoh. effect of food on the pharmacokinetics of capecitabine and its metabolites following oral administration in cancer patients. clin. cancer res. 4 (1998) 941-948. [13] j. j. hirani, d. a. rathod, k. r. vadalia. orally disintegrating tablets: a review. trop. j. pharm. res. 8 (2009) 161-172. [14] t. k. ghosh, w. r. pfiste. drug delivery to the oral cavity; chapter 14. quick-dissolving oral dosage forms taylor & francis group, llc (2005) 337-356. [15] v. h. sunesen, r. vedelsdal, h. g. kristensen, l. christrup, a. mullertz. effect of liquid volume and food intake on the absolute bioavailability of danazol, a poorly soluble drug. eur. j. pharm. sci. 24 (2005) 297-303. [16] k. sugano. fraction of a dose absorbed estimation for structurally diverse low solubility compounds. int. j. pharm. 405 (2011) 79-89. [17] k. sugano, k. terada. rateand extent-limiting factors of oral drug absorption: theory and applications. j. pharm. sci. 104 (2015) 2777-2788. [18] s. colon-useche, i. gonzalez-alvarez, v. mangas-sanjuan, m. gonzalez-alvarez, p. pastoriza, i. molina-martinez, m. bermejo, a. garcia-arieta. investigating the discriminatory power of bcsbiowaiver in vitro methodology to detect bioavailability differences between immediate release products containing a class i drug. mol. pharm. 12 (2015) 3167-3174. [19] a. garcia-arieta, j. gordon, l. gwaza, v. mangas-sanjuan, c. alvarez, j. j. torrado. agitation rate and time for complete dissolution in bcs biowaivers based on investigation of a bcs biowaiver for dexketoprofen tablets. mol. pharm. 12 (2015) 3194-3201. [20] u. fagerholm. evaluation and suggested improvements of the biopharmaceutics classification system (bcs). j. pharm. pharmacol. 59 (2007) 751-757. [21] h. kortejarvi, a. urtti, m. yliperttula. pharmacokinetic simulation of biowaiver criteria: the effects of gastric emptying, dissolution, absorption and elimination rates. eur. j. pharm. sci. 30 (2007) 155-166. [22] g. a. siebert, d. y. hung, p. chang, m. s. roberts. ion-trapping, microsomal binding, and unbound drug distribution in the hepatic retention of basic drugs. j. pharmacol. exp. ther. 308 (2004) 228-235. [23] fda. waiver of in vivo bioavailability and bioequivalence studies for immediate-release solid oral dosage forms based on a biopharmaceutics classification system. fda guidance for industry (2000) [24] who. proposal to waive in vivo bioequivalence requirements for who model list of essential medicines immediate-release, solid oral dosage forms. who technical report series, no. 937, annex 8 (2006) [25] ema. guideline on the investigation of bioequivalence. cpmp/ewp/qwp/1401/98 (2010) [26] h. canada. biopharmaceutics classification system based biowaiver. guidance document (2013) [27] k. fda. bioequivalence study. guidance document (2008) [28] b. davit, a. c. braddy, d. p. conner, l. x. yu. international guidelines for bioequivalence of systemically available orally administered generic drug products: a survey of similarities and differences. aaps j 15 (2013) 974-990. [29] l. yu, g. amidon, j. polli, h. zhao, m. mehta, d. conner, v. shah, l. lesko, m.-l. chen, v. l. lee, a. hussain. biopharmaceutics classification system: the scientific basis for biowaiver extensions. pharm. res. 19 (2002) 921-925. asami ono et al. admet & dmpk 4(4) (2016) 335-360 350 [30] c. l. cheng, l. x. yu, h. l. lee, c. y. yang, c. s. lue, c. h. chou. biowaiver extension potential to bcs class iii high solubility-low permeability drugs: bridging evidence for metformin immediate-release tablet. eur. j. pharm. sci. 22 (2004) 297-304. [31] e. jantratid, s. prakongpan, g. l. amidon, j. b. dressman. feasibility of biowaiver extension to biopharmaceutics classification system class iii drug products. clin. pharmacokinet. 45 (2006) 385399. [32] y. tsume, g. l. amidon. the biowaiver extension for bcs class iii drugs: the effect of dissolution rate on the bioequivalence of bcs class iii immediate-release drugs predicted by computer simulation. mol. pharm. 7 (2010) 1235-1243. [33] a. ono, k. sugano. application of the bcs biowaiver approach to assessing bioequivalence of orally disintegrating tablets with immediate release formulations. eur. j. pharm. sci. 64 (2014) 37-43. [34] e. jantratid, s. prakongpan, j. b. dressman, g. l. amidon, h. e. junginger, k. k. midha, d. m. barends. biowaiver monographs for immediate release solid oral dosage forms: cimetidine. j. pharm. sci. 95 (2006) 974-984. [35] c. alvarez, i. nunez, j. j. torrado, j. gordon, h. potthast, a. garcia-arieta. investigation on the possibility of biowaivers for ibuprofen. j. pharm. sci. 100 (2011) 2343-2349. [36] a. garcia-arieta, j. gordon, h. potthast. on the biopharmaceutics classification system biowaiver of ibuprofen. j. pharm. sci. 104 (2015) 2429-2432. [37] e. ramirez, o. laosa, p. guerra, b. duque, b. mosquera, a. m. borobia, s. h. lei, a. j. carcas, j. frias. acceptability and characteristics of 124 human bioequivalence studies with active substances classified according to the biopharmaceutic classification system. br. j. clin. pharmacol. 70 (2010) 694-702. ©2016 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ admet & dmpk 4(4) (2016) 335-360 biopharmaceutical and physchem properties for odt doi: 10.5599/admet.4.4.338 351 supplement table 1. odt drug list used in this study drug prescription information acarbose http://www.info.pmda.go.jp/go/interview/1/630004_3969003f3037_1_005_1f ambroxol hydrochloride http://www.info.pmda.go.jp/go/interview/1/300119_2239001g1039_1_061_1f amlodipine besylate http://www.info.pmda.go.jp/go/interview/2/671450_2171022f1029_2_1f http://www.info.pmda.go.jp/go/interview/3/400093_2171022f1045_3_017_1f aripiprazole http://www.info.pmda.go.jp/go/interview/1/180078_1179045f4022_1_012_1f bepotastine besylate http://www.info.pmda.go.jp/go/interview/1/400315_4490022f3022_1_090_1f bicalutamide http://www.info.pmda.go.jp/go/interview/1/670227_4291009f1039_1_171_1f brotizolam http://www.info.pmda.go.jp/go/interview/1/650168_1124009f2025_1_14x_1f cetirizine hydrochloride http://www.info.pmda.go.jp/go/interview/1/300119_4490020f1225_1_051_1f cilostazol http://www.info.pmda.go.jp/go/interview/1/180078_3399002f3020_1_005_1f desmopressin acetate hydrate http://www.info.pmda.go.jp/go/interview/1/670666_2419001f1023_2_1f dienogest http://www.info.pmda.go.jp/go/interview/1/790005_2499010f1023_1_m02_1f domperidone http://www.info.pmda.go.jp/go/interview/1/230124_2399005f3020_1_004_1f donepezil hydrochloride http://www.info.pmda.go.jp/go/interview/1/170033_1190012f3029_1_028_1f doxazosin mesylate http://www.info.pmda.go.jp/go/interview/2/671450_2149026f1026_2_1f droxidopa http://www.info.pmda.go.jp/go/interview/2/400093_1169006c1039_2_009_1f ebastine http://www.info.pmda.go.jp/go/interview/1/400093_4490019f1028_1_018_1f famotidine http://www.info.pmda.go.jp/go/interview/1/800126_2325003f1024_1_1f fexofenadine hydrochloride http://www.info.pmda.go.jp/go/interview/1/780069_4490023f1024_1_021_1f galantamine hydrobromide http://www.info.pmda.go.jp/go/interview/1/800155_1190019f1028_1_005_1f gimeracil http://www.info.pmda.go.jp/go/interview/1/400107_4229101f1026_1_10e_1f glimepiride http://www.info.pmda.go.jp/go/interview/1/780069_3961008f4070_1_020_1f hydrochlorothiazide http://www.info.pmda.go.jp/go/interview/1/480235_2132004f1103_1_002_1f imidafenacin http://www.info.pmda.go.jp/go/interview/1/230109_2590013f1027_1_a13_1f http://www.info.pmda.go.jp/go/interview/1/180188_2590013f1035_1_009_1f irsogladine maleate http://www.info.pmda.go.jp/go/interview/1/530263_2329020f3020_1_05f_1f http://www.info.pmda.go.jp/go/interview/1/630004_3969003f3037_1_005_1f http://www.info.pmda.go.jp/go/interview/1/300119_2239001g1039_1_061_1f http://www.info.pmda.go.jp/go/interview/2/671450_2171022f1029_2_1f http://www.info.pmda.go.jp/go/interview/1/180078_1179045f4022_1_012_1f http://www.info.pmda.go.jp/go/interview/1/400315_4490022f3022_1_090_1f http://www.info.pmda.go.jp/go/interview/1/670227_4291009f1039_1_171_1f http://www.info.pmda.go.jp/go/interview/1/650168_1124009f2025_1_14x_1f http://www.info.pmda.go.jp/go/interview/1/300119_4490020f1225_1_051_1f http://www.info.pmda.go.jp/go/interview/1/180078_3399002f3020_1_005_1f http://www.info.pmda.go.jp/go/interview/1/670666_2419001f1023_2_1f http://www.info.pmda.go.jp/go/interview/1/790005_2499010f1023_1_m02_1f http://www.info.pmda.go.jp/go/interview/1/230124_2399005f3020_1_004_1f http://www.info.pmda.go.jp/go/interview/2/671450_2149026f1026_2_1f http://www.info.pmda.go.jp/go/interview/2/400093_1169006c1039_2_009_1f http://www.info.pmda.go.jp/go/interview/1/400093_4490019f1028_1_018_1f http://www.info.pmda.go.jp/go/interview/1/780069_4490023f1024_1_021_1f http://www.info.pmda.go.jp/go/interview/1/800155_1190019f1028_1_005_1f http://www.info.pmda.go.jp/go/interview/1/400107_4229101f1026_1_10e_1f http://www.info.pmda.go.jp/go/interview/1/780069_3961008f4070_1_020_1f http://www.info.pmda.go.jp/go/interview/1/480235_2132004f1103_1_002_1f http://www.info.pmda.go.jp/go/interview/1/230109_2590013f1027_1_a13_1f http://www.info.pmda.go.jp/go/interview/1/180188_2590013f1035_1_009_1f http://www.info.pmda.go.jp/go/interview/1/530263_2329020f3020_1_05f_1f asami ono et al. admet & dmpk 4(4) (2016) 335-360 352 lafutidine http://www.info.pmda.go.jp/go/interview/1/400107_2325006f3020_1_04e_1f lansoprazole http://www.info.pmda.go.jp/go/interview/1/400256_2329023f1020_1_006_1f loratadine http://www.info.pmda.go.jp/go/interview/1/170050_4490027f1022_1_015_1f memantine hydrochloride http://www.info.pmda.go.jp/go/interview/1/430574_1190018f1023_1_m09_1f midodrine hydrochloride http://www.info.pmda.go.jp/go/interview/1/400059_2160002f2024_1_004_1f miglitol http://www.info.pmda.go.jp/go/interview/1/300297_39690a0f1026_1_001_1f naftopizil http://www.info.pmda.go.jp/go/interview/1/100898_2590009f4020_1_1f olanzapine http://www.info.pmda.go.jp/go/interview/1/530471_1179044f4028_1_18f_1f olopatadine hydrochloride http://www.info.pmda.go.jp/go/interview/1/230124_4490025f3026_1_006_1f ondansetron hydrochloride http://www.info.pmda.go.jp/go/interview/1/340278_2391006f1023_1_1f oteracil potassium http://www.info.pmda.go.jp/go/interview/1/400107_4229101f1026_1_10e_1f paroxetine hydrochloride hydrate http://www.info.pmda.go.jp/go/interview/1/480235_1179041f1254_1_005_1f pioglitazone hydrochloride http://www.info.pmda.go.jp/go/interview/1/400256_3969007f3027_1_007_1f pitavastatin calcium http://www.info.pmda.go.jp/go/interview/1/270072_2189016f4027_1_005_1f polaprezinc http://www.pmda.go.jp/pmdasearch/iyakudetail/resultdatasetpdf/380077_2329027f1029_1_04 pramipexole hydrochloride hydrate http://www.info.pmda.go.jp/go/interview/1/480235_1169012f3025_1_004_1f ramosetron hydrochloride http://www.info.pmda.go.jp/go/interview/1/800126_2399014f3029_1_1f http://www.info.pmda.go.jp/go/interview/1/800126_2391004f1024_1_1f rebamipide http://www.info.pmda.go.jp/go/interview/1/530113_2329021f2028_1_002_1f risperidone http://www.info.pmda.go.jp/go/interview/1/800155_1179038f5029_1_005_1f rizatriptan benzoate http://www.info.pmda.go.jp/go/interview/2/230109_2160006f1026_2_011_1f selegiline hydrochloride http://www.info.pmda.go.jp/go/interview/1/170654_1169010f2020_1_s03_1f sertraline hydrochloride http://www.info.pmda.go.jp/go/interview/2/671450_1179046f1028_2_1f sildenafil citrate http://www.info.pmda.go.jp/go/interview/1/480235_259000af3027_1_002_1f solifenacin succinate http://www.info.pmda.go.jp/go/interview/1/800126_2590011f3020_1_1f taltirelin hydrate http://www.info.pmda.go.jp/go/interview/1/400315_1190014f2021_1_050_1f tamsulosin hydrochloride http://www.info.pmda.go.jp/go/interview/1/800126_2590008f1026_1_1f tegafur http://www.info.pmda.go.jp/go/interview/1/400107_4229101f1026_1_10e_1f http://www.info.pmda.go.jp/go/interview/1/400107_2325006f3020_1_04e_1f http://www.info.pmda.go.jp/go/interview/1/400256_2329023f1020_1_006_1f http://www.info.pmda.go.jp/go/interview/1/170050_4490027f1022_1_015_1f http://www.info.pmda.go.jp/go/interview/1/430574_1190018f1023_1_m09_1f http://www.info.pmda.go.jp/go/interview/1/400059_2160002f2024_1_004_1f http://www.info.pmda.go.jp/go/interview/1/300297_39690a0f1026_1_001_1f http://www.info.pmda.go.jp/go/interview/1/100898_2590009f4020_1_1f http://www.info.pmda.go.jp/go/interview/1/530471_1179044f4028_1_18f_1f http://www.info.pmda.go.jp/go/interview/1/230124_4490025f3026_1_006_1f http://www.info.pmda.go.jp/go/interview/1/340278_2391006f1023_1_1f http://www.info.pmda.go.jp/go/interview/1/400107_4229101f1026_1_10e_1f http://www.info.pmda.go.jp/go/interview/1/400256_3969007f3027_1_007_1f http://www.info.pmda.go.jp/go/interview/1/270072_2189016f4027_1_005_1f http://www.pmda.go.jp/pmdasearch/iyakudetail/resultdatasetpdf/380077_2329027f1029_1_04 http://www.info.pmda.go.jp/go/interview/1/800126_2399014f3029_1_1f http://www.info.pmda.go.jp/go/interview/1/800155_1179038f5029_1_005_1f http://www.info.pmda.go.jp/go/interview/2/230109_2160006f1026_2_011_1f http://www.info.pmda.go.jp/go/interview/1/170654_1169010f2020_1_s03_1f http://www.info.pmda.go.jp/go/interview/2/671450_1179046f1028_2_1f http://www.info.pmda.go.jp/go/interview/1/800126_2590011f3020_1_1f http://www.info.pmda.go.jp/go/interview/1/400315_1190014f2021_1_050_1f http://www.info.pmda.go.jp/go/interview/1/400107_4229101f1026_1_10e_1f admet & dmpk 4(4) (2016) 335-360 biopharmaceutical and physchem properties for odt doi: 10.5599/admet.4.4.338 353 tramadol hydrochloride http://www.info.pmda.go.jp/go/interview/1/530263_1149038f1024_1_01f_1f valsartan http://www.info.pmda.go.jp/go/interview/1/300242_2149041f5026_5_dio_1f voglibose http://www.info.pmda.go.jp/go/interview/1/400256_3969004f3023_1_006_1f zolmitriptan http://www.info.pmda.go.jp/go/interview/1/670227_2160004f2023_1_101_1f zolpidem tartrate http://www.info.pmda.go.jp/go/interview/3/800126_1129009f1025_3_1f zonizamide http://www.info.pmda.go.jp/go/interview/1/400093_1169015f2022_1_004_1f http://www.info.pmda.go.jp/go/interview/1/530263_1149038f1024_1_01f_1f http://www.info.pmda.go.jp/go/interview/1/300242_2149041f5026_5_dio_1f http://www.info.pmda.go.jp/go/interview/1/400256_3969004f3023_1_006_1f http://www.info.pmda.go.jp/go/interview/1/670227_2160004f2023_1_101_1f http://www.info.pmda.go.jp/go/interview/3/800126_1129009f1025_3_1f http://www.info.pmda.go.jp/go/interview/1/400093_1169015f2022_1_004_1f asami ono et al. admet & dmpk 4(4) (2016) 335-360 354 supplement table 2. ir formulation drug list used in this study drug prescription information acarbose http://www.info.pmda.go.jp/go/interview/1/630004_3969003f3037_1_005_1f acetaminophen http://www.info.pmda.go.jp/go/interview/3/671610_1141007f1063_3_15l_1f acetylsalicylic acid http://www.info.pmda.go.jp/go/interview/1/630004_3399007h1021_1_002_1f acyclovir http://www.info.pmda.go.jp/go/interview/w/340278_6250002f1025_1_001_1f alendronate sodium hydrate http://www.info.pmda.go.jp/go/interview/2/170050_3999018f1021_2_018_1f alfacalcidol http://www.info.pmda.go.jp/go/interview/1/450045_3112001m1046_1_007_1f allopurinol http://www.info.pmda.go.jp/go/interview/w/340278_3943001f1314_1_1f ambroxol hydrochloride http://www.info.pmda.go.jp/go/interview/1/650168_2239001n1135_1_148_1f amlodipine besylate http://www.info.pmda.go.jp/go/interview/2/671450_2171022f1029_2_1f atenolol http://www.info.pmda.go.jp/go/interview/3/670227_2123011f1155_1_131_1f atorvastatin calcium http://www.info.pmda.go.jp/go/interview/1/800126_2189015f1023_1_1f azithromycin hydrate http://www.info.pmda.go.jp/go/interview/2/671450_6149004f1028_2_1f azulene sulfonate http://www.info.pmda.go.jp/go/interview/1/530263_2323001f1225_1_04f_1f benidipine hydrochloride http://www.info.pmda.go.jp/go/interview/1/230124_2171021f1024_1_001_1f beraprost sodium http://www.info.pmda.go.jp/go/interview/1/480220_3399005f1021_1_a06_1f bicalutamide http://www.info.pmda.go.jp/go/interview/1/670227_4291009f1039_1_171_1f bisoprolol fumarate http://www.info.pmda.go.jp/go/interview/1/400315_2123016f1107_1_120_1f brotizolam http://www.info.pmda.go.jp/go/interview/1/650168_1124009f2025_1_14x_1f cabergoline http://www.info.pmda.go.jp/go/interview/4/671450_1169011f1028_4_1f camostat mesylate http://www.info.pmda.go.jp/go/interview/1/180188_3999003f1297_1_005_1f candesartan cilexetil http://www.info.pmda.go.jp/go/interview/1/400256_2149040f1026_1_004_1f carvedilol http://www.info.pmda.go.jp/go/interview/2/430574_2149032f1021_2_a13_1f cefcapene pivoxil hydrochloride http://www.info.pmda.go.jp/go/interview/1/340018_6132016f1023_1_011_1f cefdinir http://www.info.pmda.go.jp/go/interview/3/800126_6132013m1029_3_1f cefditoren pivoxil http://www.info.pmda.go.jp/go/interview/1/780009_6132015f1037_1_01a_1f cefotiam hexetil hydrochloride http://www.pmda.go.jp/pmdasearch/iyakudetail/resultdatasetpdf/400256_6132012f1025_1_12 http://www.info.pmda.go.jp/go/interview/1/630004_3969003f3037_1_005_1f http://www.info.pmda.go.jp/go/interview/3/671610_1141007f1063_3_15l_1f http://www.info.pmda.go.jp/go/interview/1/630004_3399007h1021_1_002_1f http://www.info.pmda.go.jp/go/interview/w/340278_6250002f1025_1_001_1f http://www.info.pmda.go.jp/go/interview/2/170050_3999018f1021_2_018_1f http://www.info.pmda.go.jp/go/interview/1/450045_3112001m1046_1_007_1f http://www.info.pmda.go.jp/go/interview/w/340278_3943001f1314_1_1f http://www.info.pmda.go.jp/go/interview/1/650168_2239001n1135_1_148_1f http://www.info.pmda.go.jp/go/interview/2/671450_2171022f1029_2_1f http://www.info.pmda.go.jp/go/interview/3/670227_2123011f1155_1_131_1f http://www.info.pmda.go.jp/go/interview/1/800126_2189015f1023_1_1f http://www.info.pmda.go.jp/go/interview/2/671450_6149004f1028_2_1f http://www.info.pmda.go.jp/go/interview/1/530263_2323001f1225_1_04f_1f http://www.info.pmda.go.jp/go/interview/1/230124_2171021f1024_1_001_1f http://www.info.pmda.go.jp/go/interview/1/480220_3399005f1021_1_a06_1f http://www.info.pmda.go.jp/go/interview/1/670227_4291009f1039_1_171_1f http://www.info.pmda.go.jp/go/interview/1/400315_2123016f1107_1_120_1f http://www.info.pmda.go.jp/go/interview/1/650168_1124009f2025_1_14x_1f http://www.info.pmda.go.jp/go/interview/4/671450_1169011f1028_4_1f http://www.info.pmda.go.jp/go/interview/1/180188_3999003f1297_1_005_1f http://www.info.pmda.go.jp/go/interview/1/400256_2149040f1026_1_004_1f http://www.info.pmda.go.jp/go/interview/2/430574_2149032f1021_2_a13_1f http://www.info.pmda.go.jp/go/interview/1/340018_6132016f1023_1_011_1f http://www.info.pmda.go.jp/go/interview/3/800126_6132013m1029_3_1f http://www.info.pmda.go.jp/go/interview/1/780009_6132015f1037_1_01a_1f http://www.pmda.go.jp/pmdasearch/iyakudetail/resultdatasetpdf/400256_6132012f1025_1_12 admet & dmpk 4(4) (2016) 335-360 biopharmaceutical and physchem properties for odt doi: 10.5599/admet.4.4.338 355 cetirizine hydrochloride http://www.info.pmda.go.jp/go/interview/2/820110_4490020f1020_2_013_1f cilostazol http://www.info.pmda.go.jp/go/interview/1/180078_3399002f3020_1_005_1f clarithromycin http://www.info.pmda.go.jp/go/interview/1/400059_6149003f2038_1_028_1f cyclosporin a http://www.info.pmda.go.jp/go/interview/2/300242_3999004m3021_2_neo_1f diclofenac sodium http://www.info.pmda.go.jp/go/interview/1/300242_1147002f1560_1_vol_1f donepezil hydrochloride http://www.info.pmda.go.jp/go/interview/1/170033_1190012f3029_1_028_1f doxazosin mesylate http://www.info.pmda.go.jp/go/interview/2/671450_2149026f1026_2_1f doxifluridine http://www.info.pmda.go.jp/go/interview/2/450045_4223004m1027_2_004_1f ebastine http://www.info.pmda.go.jp/go/interview/1/400093_4490019f1028_1_018_1f enalapril maleate http://www.info.pmda.go.jp/go/interview/2/170050_2144002f1024_2_016_1f epalrestat http://www.info.pmda.go.jp/go/interview/1/180188_3999013f1231_1_004_1f eperisone hydrochloride http://www.info.pmda.go.jp/go/interview/1/170033_1249009d1030_1_007_1f epinastine hydrochloride http://www.info.pmda.go.jp/go/interview/1/650168_4490014f1025_1_119_1f ethyl icosapentate http://www.info.pmda.go.jp/go/interview/1/790005_3399004m2022_1_m03_1f etizolam http://www.info.pmda.go.jp/go/interview/2/400315_1179025c1054_2_150_1f famotidine http://www.info.pmda.go.jp/go/interview/1/800126_2325003f1024_1_1f fexofenadine hydrochloride http://www.info.pmda.go.jp/go/interview/1/780069_4490023f1024_1_021_1f fluconazole http://www.info.pmda.go.jp/go/interview/2/671450_6290002m1020_2_1f flurbiprofen http://www.info.pmda.go.jp/go/interview/1/200022_1149011d1032_1_080_1f fluvastatin sodium http://www.info.pmda.go.jp/go/interview/3/300242_2189012f1020_3_loc_1f fluvoxamine maleate http://www.info.pmda.go.jp/go/interview/1/780009_1179039f1028_1_1f fursultiamine http://www.pmda.go.jp/pmdasearch/iyakudetail/resultdatasetpdf/400256_3122007f2039_1_10 gefitinib http://www.info.pmda.go.jp/go/interview/1/670227_4291013f1027_1_192_1f glimepiride http://www.info.pmda.go.jp/go/interview/1/780069_3961008f4070_1_020_1f granisetron hydrochloride http://www.info.pmda.go.jp/go/interview/1/450045_2391002c1037_1_010_1f imatinib mesylate http://www.info.pmda.go.jp/go/interview/1/300242_4291011f1028_1_gli_1f imidapril hydrochloride http://www.info.pmda.go.jp/go/interview/1/400315_2144008f1021_1_090_1f itraconazole http://www.info.pmda.go.jp/go/interview/1/800155_6290004m1029_2_008_1f http://www.info.pmda.go.jp/go/interview/2/820110_4490020f1020_2_013_1f http://www.info.pmda.go.jp/go/interview/1/180078_3399002f3020_1_005_1f http://www.info.pmda.go.jp/go/interview/1/400059_6149003f2038_1_028_1f http://www.info.pmda.go.jp/go/interview/2/300242_3999004m3021_2_neo_1f http://www.info.pmda.go.jp/go/interview/1/300242_1147002f1560_1_vol_1f http://www.info.pmda.go.jp/go/interview/1/170033_1190012f3029_1_028_1f http://www.info.pmda.go.jp/go/interview/2/671450_2149026f1026_2_1f http://www.info.pmda.go.jp/go/interview/2/450045_4223004m1027_2_004_1f http://www.info.pmda.go.jp/go/interview/1/400093_4490019f1028_1_018_1f http://www.info.pmda.go.jp/go/interview/2/170050_2144002f1024_2_016_1f http://www.info.pmda.go.jp/go/interview/1/180188_3999013f1231_1_004_1f http://www.info.pmda.go.jp/go/interview/1/170033_1249009d1030_1_007_1f http://www.info.pmda.go.jp/go/interview/1/650168_4490014f1025_1_119_1f http://www.info.pmda.go.jp/go/interview/1/790005_3399004m2022_1_m03_1f http://www.info.pmda.go.jp/go/interview/2/400315_1179025c1054_2_150_1f http://www.info.pmda.go.jp/go/interview/1/800126_2325003f1024_1_1f http://www.info.pmda.go.jp/go/interview/1/780069_4490023f1024_1_021_1f http://www.info.pmda.go.jp/go/interview/2/671450_6290002m1020_2_1f http://www.info.pmda.go.jp/go/interview/1/200022_1149011d1032_1_080_1f http://www.info.pmda.go.jp/go/interview/3/300242_2189012f1020_3_loc_1f http://www.info.pmda.go.jp/go/interview/1/780009_1179039f1028_1_1f http://www.pmda.go.jp/pmdasearch/iyakudetail/resultdatasetpdf/400256_3122007f2039_1_10 http://www.info.pmda.go.jp/go/interview/1/670227_4291013f1027_1_192_1f http://www.info.pmda.go.jp/go/interview/1/780069_3961008f4070_1_020_1f http://www.info.pmda.go.jp/go/interview/1/450045_2391002c1037_1_010_1f http://www.info.pmda.go.jp/go/interview/1/300242_4291011f1028_1_gli_1f http://www.info.pmda.go.jp/go/interview/1/400315_2144008f1021_1_090_1f http://www.info.pmda.go.jp/go/interview/1/800155_6290004m1029_2_008_1f asami ono et al. admet & dmpk 4(4) (2016) 335-360 356 ketoprofen a http://www.info.pmda.go.jp/go/interview/2/650208_1149700j1035_2_1f ketotifen fumarate http://www.info.pmda.go.jp/go/interview/1/300242_4490003m1263_1_z-c_1f l-carbocysteine http://www.info.pmda.go.jp/go/interview/1/230109_2233002f1174_1_001_1f levofloxacin http://www.info.pmda.go.jp/go/interview/1/430574_6241013c2024_1_c11_1f limaprost alfadex http://www.info.pmda.go.jp/go/interview/1/180188_3399003f1073_1_010_1f loratadine http://www.info.pmda.go.jp/go/interview/1/170050_4490027f1022_1_015_1f losartan potassium http://www.info.pmda.go.jp/go/interview/2/170050_2149039f1031_2_018_1f loxoprofen sodium hydrate http://www.info.pmda.go.jp/go/interview/1/430574_1149019c1149_1_lo8_1f manidipine hydrochloride http://www.pmda.go.jp/pmdasearch/iyakudetail/resultdatasetpdf/400256_2149027f1020_1_09 mecobalamin http://www.info.pmda.go.jp/go/interview/1/170033_3136004c1038_1_006_1f meloxicam http://www.info.pmda.go.jp/go/interview/1/650168_1149035f1020_1_156_1f menatetrenone http://www.info.pmda.go.jp/go/interview/1/170033_3160002m2028_1_008_1f mesalazine http://www.info.pmda.go.jp/go/interview/1/230109_2399009f1149_1_015_1f methylmethionine sulfonium chloride http://www.info.pmda.go.jp/go/interview/1/270072_2321001f1042_1_002_1f mexiletine hydrochloride http://www.info.pmda.go.jp/go/interview/1/650168_2129003m1021_1_14y_1f montelukast sodium http://www.info.pmda.go.jp/go/interview/1/230109_4490026f2040_1_035_1f. mosapride citrate hydrate http://www.info.pmda.go.jp/go/interview/1/400093_2399010b1034_1_020_1f nicardipine hydrochloride http://www.info.pmda.go.jp/go/interview/1/300119_2149019f1085_1_041_1f nicergoline http://www.info.pmda.go.jp/go/interview/1/400315_2190021b1095_1_100_1f nicorandil http://www.info.pmda.go.jp/go/interview/1/450045_2171017f1028_1_009_1f nifedipine http://www.info.pmda.go.jp/go/interview/1/630004_2171014m1104_1_001_1f nilvadipine http://www.info.pmda.go.jp/go/interview/3/800126_2149022f1028_3_1f nizatidine http://www.pmda.go.jp/pmdasearch/iyakudetail/resultdatasetpdf/380077_2325005f1031_1_06 olanzapine http://www.info.pmda.go.jp/go/interview/1/530471_1179044f4028_1_18f_1f olopatadine hydrochloride http://www.info.pmda.go.jp/go/interview/1/230124_4490025f3026_1_006_1f oseltamivir phosphate http://www.info.pmda.go.jp/go/interview/1/450045_6250021m1027_1_026_1f paroxetine hydrochloride hydrate http://www.info.pmda.go.jp/go/interview/2/340278_1179041f1025_1_018_1f http://www.info.pmda.go.jp/go/interview/2/650208_1149700j1035_2_1f http://www.info.pmda.go.jp/go/interview/1/300242_4490003m1263_1_z-c_1f http://www.info.pmda.go.jp/go/interview/1/230109_2233002f1174_1_001_1f http://www.info.pmda.go.jp/go/interview/1/430574_6241013c2024_1_c11_1f http://www.info.pmda.go.jp/go/interview/1/180188_3399003f1073_1_010_1f http://www.info.pmda.go.jp/go/interview/1/170050_4490027f1022_1_015_1f http://www.info.pmda.go.jp/go/interview/2/170050_2149039f1031_2_018_1f http://www.info.pmda.go.jp/go/interview/1/430574_1149019c1149_1_lo8_1f http://www.pmda.go.jp/pmdasearch/iyakudetail/resultdatasetpdf/400256_2149027f1020_1_09 http://www.info.pmda.go.jp/go/interview/1/170033_3136004c1038_1_006_1f http://www.info.pmda.go.jp/go/interview/1/650168_1149035f1020_1_156_1f http://www.info.pmda.go.jp/go/interview/1/170033_3160002m2028_1_008_1f http://www.info.pmda.go.jp/go/interview/1/230109_2399009f1149_1_015_1f http://www.info.pmda.go.jp/go/interview/1/270072_2321001f1042_1_002_1f http://www.info.pmda.go.jp/go/interview/1/650168_2129003m1021_1_14y_1f http://www.info.pmda.go.jp/go/interview/1/230109_4490026f2040_1_035_1f. http://www.info.pmda.go.jp/go/interview/1/400093_2399010b1034_1_020_1f http://www.info.pmda.go.jp/go/interview/1/300119_2149019f1085_1_041_1f http://www.info.pmda.go.jp/go/interview/1/400315_2190021b1095_1_100_1f http://www.info.pmda.go.jp/go/interview/1/450045_2171017f1028_1_009_1f http://www.info.pmda.go.jp/go/interview/1/630004_2171014m1104_1_001_1f http://www.info.pmda.go.jp/go/interview/3/800126_2149022f1028_3_1f http://www.pmda.go.jp/pmdasearch/iyakudetail/resultdatasetpdf/380077_2325005f1031_1_06 http://www.info.pmda.go.jp/go/interview/1/530471_1179044f4028_1_18f_1f http://www.info.pmda.go.jp/go/interview/1/230124_4490025f3026_1_006_1f http://www.info.pmda.go.jp/go/interview/1/450045_6250021m1027_1_026_1f http://www.info.pmda.go.jp/go/interview/2/340278_1179041f1025_1_018_1f admet & dmpk 4(4) (2016) 335-360 biopharmaceutical and physchem properties for odt doi: 10.5599/admet.4.4.338 357 pergolide mesylate http://www.info.pmda.go.jp/go/interview/2/230124_1169008f1026_2_004_1f perindopril erbumine http://www.info.pmda.go.jp/go/interview/3/230124_2144012f1028_3_001_1f pilsicainide hydrochloride hydrate http://www.info.pmda.go.jp/go/interview/3/430574_2129008m1024_3_s10_1f pioglitazone hydrochloride http://www.info.pmda.go.jp/go/interview/1/400256_3969007f3027_1_007_1f pranlukast hydrate http://www.info.pmda.go.jp/go/interview/1/180188_4490017m1036_1_009_1f pravastatin sodium http://www.info.pmda.go.jp/go/interview/2/430574_2189010c1032_2_m10_1f procaterol hydrochloride hydrate http://www.info.pmda.go.jp/go/interview/1/180078_2259004f2168_1_005_1f propiverine hydrochloride http://www.info.pmda.go.jp/go/interview/1/400107_2590007c1025_1_04l_1f quetiapine fumarate http://www.info.pmda.go.jp/go/interview/2/800126_1179042c1023_2_1f ranitidine hydrochloride http://www.info.pmda.go.jp/go/interview/1/340278_2325002f1194_1_009_1f rebamipide http://www.info.pmda.go.jp/go/interview/1/180078_2329021d1020_1_013_1f risedronate sodium hydrate http://www.info.pmda.go.jp/go/interview/2/111890_3999019f1026_2_016_1f risperidone http://www.info.pmda.go.jp/go/interview/1/800155_1179038f5029_1_005_1f sarpogrelate hydrochloride http://www.info.pmda.go.jp/go/interview/3/400315_3399006c1020_3_15a_1f simvastatin http://www.info.pmda.go.jp/go/interview/2/170050_2189011f1025_2_022_1f sultamicillin tosilate hydrate http://www.info.pmda.go.jp/go/interview/2/671450_6131008f1030_2_1f tacrolimus hydrate http://www.info.pmda.go.jp/go/interview/3/800126_3999014d1022_3_1f taltirelin hydrate http://www.info.pmda.go.jp/go/interview/1/400315_1190014f2021_1_050_1f tamoxifen citrate http://www.info.pmda.go.jp/go/interview/1/670227_4291003f1163_1_012_1f tegafur http://www.info.pmda.go.jp/go/interview/1/400107_4229101f1026_1_10e_1f temocapril hydrochloride http://www.info.pmda.go.jp/go/interview/3/430574_2144009f1026_3_c10_1f teprenone http://www.info.pmda.go.jp/go/interview/1/170033_2329012c1026_1_009_1f terbinafine hydrochloride http://www.info.pmda.go.jp/go/interview/1/730012_6290005f1032_1_006_1f ticlopidine hydrochloride http://www.info.pmda.go.jp/go/interview/4/780069_3399001c1027_4_011_1f tocopherol nicotinate http://www.info.pmda.go.jp/go/interview/1/170033_2190006c1037_1_008_1f tulobuterol hydrochloride http://www.info.pmda.go.jp/go/interview/3/100159_2259002f1065_3_004_1f ursodeoxycholic acid http://www.info.pmda.go.jp/go/interview/2/400315_2362001f1088_2_11a_1f valacyclovir hydrochloride http://www.info.pmda.go.jp/go/interview/1/340278_6250019f1020_1_013_1f http://www.info.pmda.go.jp/go/interview/2/230124_1169008f1026_2_004_1f http://www.info.pmda.go.jp/go/interview/3/230124_2144012f1028_3_001_1f http://www.info.pmda.go.jp/go/interview/3/430574_2129008m1024_3_s10_1f http://www.info.pmda.go.jp/go/interview/1/400256_3969007f3027_1_007_1f http://www.info.pmda.go.jp/go/interview/1/180188_4490017m1036_1_009_1f http://www.info.pmda.go.jp/go/interview/2/430574_2189010c1032_2_m10_1f http://www.info.pmda.go.jp/go/interview/1/180078_2259004f2168_1_005_1f http://www.info.pmda.go.jp/go/interview/1/400107_2590007c1025_1_04l_1f http://www.info.pmda.go.jp/go/interview/2/800126_1179042c1023_2_1f http://www.info.pmda.go.jp/go/interview/1/340278_2325002f1194_1_009_1f http://www.info.pmda.go.jp/go/interview/1/180078_2329021d1020_1_013_1f http://www.info.pmda.go.jp/go/interview/2/111890_3999019f1026_2_016_1f http://www.info.pmda.go.jp/go/interview/1/800155_1179038f5029_1_005_1f http://www.info.pmda.go.jp/go/interview/3/400315_3399006c1020_3_15a_1f http://www.info.pmda.go.jp/go/interview/2/170050_2189011f1025_2_022_1f http://www.info.pmda.go.jp/go/interview/2/671450_6131008f1030_2_1f http://www.info.pmda.go.jp/go/interview/3/800126_3999014d1022_3_1f http://www.info.pmda.go.jp/go/interview/1/400315_1190014f2021_1_050_1f http://www.info.pmda.go.jp/go/interview/1/670227_4291003f1163_1_012_1f http://www.info.pmda.go.jp/go/interview/1/400107_4229101f1026_1_10e_1f http://www.info.pmda.go.jp/go/interview/3/430574_2144009f1026_3_c10_1f http://www.info.pmda.go.jp/go/interview/1/170033_2329012c1026_1_009_1f http://www.info.pmda.go.jp/go/interview/1/730012_6290005f1032_1_006_1f http://www.info.pmda.go.jp/go/interview/4/780069_3399001c1027_4_011_1f http://www.info.pmda.go.jp/go/interview/1/170033_2190006c1037_1_008_1f http://www.info.pmda.go.jp/go/interview/3/100159_2259002f1065_3_004_1f http://www.info.pmda.go.jp/go/interview/2/400315_2362001f1088_2_11a_1f http://www.info.pmda.go.jp/go/interview/1/340278_6250019f1020_1_013_1f asami ono et al. admet & dmpk 4(4) (2016) 335-360 358 valproate sodium http://www.info.pmda.go.jp/go/interview/1/230124_1139004f1096_1_005_1f valsartan http://www.info.pmda.go.jp/go/interview/1/300242_2149041f5026_5_dio_1f voglibose http://www.info.pmda.go.jp/go/interview/1/400256_3969004f3023_1_006_1f zolpidem tartrate http://www.info.pmda.go.jp/go/interview/3/800126_1129009f1025_3_1f a ir formulations of ketoprofen are no longer in production in japan. the dose strengths of the ir formulations were from the f da website (http://www.accessdata.fda.gov/scripts/cder/drugsatfda/index.cfm). http://www.info.pmda.go.jp/go/interview/1/230124_1139004f1096_1_005_1f http://www.info.pmda.go.jp/go/interview/1/300242_2149041f5026_5_dio_1f http://www.info.pmda.go.jp/go/interview/1/400256_3969004f3023_1_006_1f http://www.info.pmda.go.jp/go/interview/3/800126_1129009f1025_3_1f http://www.accessdata.fda.gov/scripts/cder/drugsatfda/index.cfm admet & dmpk 4(4) (2016) 335-360 biopharmaceutical and physchem properties for odt doi: 10.5599/admet.4.4.338 359 supplement table 3. model equation for nc calculation by using the henderson-hasselbalch equation. compound type model equation mono acid nc = (−1) × 10−𝑝𝐾𝑎 10−𝑝𝐾𝑎 + 10−𝑝𝐻 mono base nc = 10−𝑝𝐻 10−𝑝𝐾𝑎 + 10−𝑝𝐻 di acid nc = (−1) × 10−𝑝𝐾𝑎1 × 10−𝑝𝐻 + (−2) × 10−𝑝𝐾𝑎1 × 10−𝑝𝐾𝑎2 [10−𝑝𝐻]2 + 10−𝑝𝐾𝑎1 × 10−𝑝𝐻 +10−𝑝𝐾𝑎1 × 10−𝑝𝐾𝑎2 di base nc = 2 × [10−𝑝𝐻 ]2 + 10−𝑝𝐾𝑎1 × 10−𝑝𝐻 [10−𝑝𝐻]2 + 10−𝑝𝐾𝑎1 × 10−𝑝𝐻 +10−𝑝𝐾𝑎1 × 10−𝑝𝐾𝑎2 mono acid/mono base nc = [10−𝑝𝐻 ]2 + (−1) × 10−𝑝𝐾𝑎1 × 10−𝑝𝐾𝑎2 [10−𝑝𝐻]2 + 10−𝑝𝐾𝑎1 × 10−𝑝𝐻 +10−𝑝𝐾𝑎1 × 10−𝑝𝐾𝑎2 tri acid nc = (−1) × 10−𝑝𝐾𝑎1 × [10−𝑝𝐻]2 + (−2) × 10−𝑝𝐾𝑎1 × 10−𝑝𝐾𝑎2 × 10−𝑝𝐻 + (−3) × 10−𝑝𝐾𝑎1 × 10−𝑝𝐾𝑎2 × 10−𝑝𝐾𝑎3 [10−𝑝𝐻]3 + 10−𝑝𝐾𝑎1 × [10−𝑝𝐻 ]2+10−𝑝𝐾𝑎1 × 10−𝑝𝐾𝑎2 × 10−𝑝𝐻+10−𝑝𝐾𝑎1 × 10−𝑝𝐾𝑎2 × 10−𝑝𝐾𝑎3 tri base nc = 3 × [10−𝑝𝐻 ]3 + 2 × 10−𝑝𝐾𝑎1 × [10−𝑝𝐻 ]2 + 10−𝑝𝐾𝑎1 × 10−𝑝𝐾𝑎2 × 10−𝑝𝐻 [10−𝑝𝐻]3 + 10−𝑝𝐾𝑎1 × [10−𝑝𝐻 ]2+10−𝑝𝐾𝑎1 × 10−𝑝𝐾𝑎2 × 10−𝑝𝐻+10−𝑝𝐾𝑎1 × 10−𝑝𝐾𝑎2 × 10−𝑝𝐾𝑎3 di acid/mono base nc = [10−𝑝𝐻]3 + (−1) × 10−𝑝𝐾𝑎1 × 10−𝑝𝐾𝑎2 × 10−𝑝𝐻 +(−2) × 10−𝑝𝐾𝑎1 × 10−𝑝𝐾𝑎2 × 10−𝑝𝐾𝑎3 [10−𝑝𝐻]3 + 10−𝑝𝐾𝑎1 × [10−𝑝𝐻 ]2+10−𝑝𝐾𝑎1 × 10−𝑝𝐾𝑎2 × 10−𝑝𝐻+10−𝑝𝐾𝑎1 × 10−𝑝𝐾𝑎2 × 10−𝑝𝐾𝑎3 mono acid/di base nc = 2 × [10−𝑝𝐻]3 + 10−𝑝𝐾𝑎1 × [10−𝑝𝐻 ]2 + (−1) × 10−𝑝𝐾𝑎1 × 10−𝑝𝐾𝑎2 × 10−𝑝𝐾𝑎3 [10−𝑝𝐻]3 + 10−𝑝𝐾𝑎1 × [10−𝑝𝐻 ]2+10−𝑝𝐾𝑎1 × 10−𝑝𝐾𝑎2 × 10−𝑝𝐻+10−𝑝𝐾𝑎1 × 10−𝑝𝐾𝑎2 × 10−𝑝𝐾𝑎3 doi: 10.5599/admet.4.4.338 360 supplement table 4 solubility definitions solubility definition parts of solvent required for 1 part of solute solubility range (mg/ml) solubility assigned (mg/ml) very soluble <1 ≥1000 1000 freely soluble 1‐10 100‐1000 100 soluble 10‐30 33‐100 33 sparingly soluble 30‐100 10‐33 10 slightly soluble 100‐1000 1‐10 1 very slightly soluble 1000‐10000 0.1-1 0.1 practically insoluble ≥10000 <0.1 0.01 manuscript doi: 10.5599/admet.2.4.64 254 admet & dmpk 2(4) (2014) 254-271; doi: 10.5599/admet.2.4.64 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper emulgel based topical delivery system for loratadine vijay kumar 1 , sheefali mahant 2, , rekha rao 3 , sanju nanda 2 1 mm college of pharmacy, mm university, mullana-ambala, india133207 2 department of pharmaceutical sciences, maharshi dayanand university, rohtak, india124001 3 department of pharmaceutical sciences, guru jambeshwar university of science & technology, hisar, india125001  corresponding author: sheefali mahant e-mail: sheefali.m@gmail.com; tel.: +91-9416327336 received: august 27, 2014; revised: november 27, 2014; published: january 09, 2015 abstract the present study was taken up with the objective of formulating emulgels to facilitate topical delivery of loratadine in the treatment of localized skin allergy. it was attempted to prepare the emulgels using three different types of surfactants, i.e. cationic, non-ionic and anionic. further, the aim was to compare the formulations for their drug release and stability. loratadine was incorporated into an o/w system, which was, subsequently, gellified using carbopol 940. the resulting emulgels were subject to tests for physical characteristics. thereafter, ex vivo drug release study, skin irritation test and in vivo test for anti-allergic activity were also carried out. a considerably higher drug release was recorded from the emulgel formulated with the cationic surfactant, cetrimide. drug release kinetics was investigated by fitting the drug release data into various models. the drug release from all the formulations was found to follow zero-order kinetics. in addition, the prepared emulgels exhibited satisfactory physical characteristics and good stability. besides being non-irritant to the skin, they were effective in alleviating symptoms of skin allergy. keywords gellified emulsion; cetrimide; allergy; surfactant; carbopol; sodium lauryl sulphate; tween; span introduction allergy (hypersensitivity) may be defined as untoward immunological reaction to an environmental immunogen, called allergen. allergy mainly results due to the release of inflammatory mediators such as histamine, leukotrienes, etc. from the mast cells. skin allergy is characterized by symptoms such as rash, angioedema, inflammation, pain, irritation, pruritis, urticaria and wheals [1-3]. according to the american academy of allergy, asthma and immunology, there has been a continuous rise in the prevalence of allergy in industrialized nations for more than fifty years. worldwide, urticaria is reported to have a lifetime prevalence of more than 20 % [4]. in 2010, 9.4 million cases of skin allergy were reported in children in us [5]. http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:sheefali.m@gmail.com admet & dmpk 2(4) (2014) 254-271 loratadine emulgels doi: 10.5599/admet.2.4.64 255 the pharmacotherapy for skin allergy consists of administration of antihistamines, topical corticosteroids and systemic corticosteroids [6]. antihistamines form the first-line treatment for skin allergy. loratadine is a second generation h1-antihistaminic agent [7]. it is a bcs class ii drug, having low molecular weight and high partition coefficient. it is typically administered by oral route [8]. oral administration of the drug is associated with adverse effects like headache, nausea and fatigue. moreover, the oral bioavailability of loratadine is poor [7]. in light of these facts, adopting the topical route for the delivery of loratadine would be useful in treating skin conditions characterized by localized allergic reaction. topical drug delivery can be defined as the application of a drug formulation to the skin to directly treat cutaneous disorders [9]. the topical agents use cutaneous delivery to specifically target the site of application. the sites of action for topical agents are the soft tissues and peripheral nerves underlying the site of application. in topical drug delivery systems, serum levels generally remain relatively low, as a result of which, the likelihood of systemic side effects or drug-drug interactions is abated. the vehicle in which the active ingredient(s) are delivered plays a significant role as it affects the depth of skin penetration and rate of absorption into the epidermis. ideally, a topical agent should have a low molecular weight (<500 da) and possess both, hydrophobic as well as hydrophilic features, in order to traverse the stratum corneum and to penetrate the aqueous epidermis, respectively [10]. emulgels or gellified emulsions are the topical formulations comprising of emulsion and gel, hence, possessing properties contributed by both. the oil phase, gelling agent and emulsifying agent constitute the major components of an emulgel system. their concentrations significantly affect the rate and extent of drug release from the formulation [11]. emulgels for dermatological use have several favorable properties such as being thixotropic, greaseless, easily spreadable, easily removable, emollient, non-staining, water soluble, more stable, bio-friendly, transparent and pleasant appearance. the advantages of emulgels include easy incorporation of hydrophobic drug into gel using oil-in-water emulsion system, increased stability, better loading capacity, and controlled release [11-14].owing to the merits of emulgels over the conventional dermatological formulations, many drugs have been incorporated into them. emulgels have been formulated for varied drug categories, such as non steroidal anti-inflammatory drugs, anti-fungal agents, anti-viral drugs, antibacterial drugs and local anaesthetics. topically, loratadine can be used in the therapy of disease or disorder characterized by hypersensititvity, such as urticaria, allergy based dermatoses (allergic skin reactions), atopical eczema, itching, redness, sunburn and insect bites [15]. the present study was taken up with the view to develop emulgels for dermal delivery of loratadine. further, the aim was to formulate loratadine emulgels using three different types of emulsifying agents (cationic, anionic and non-ionic surfactants) and compare them for their drug release and stability. the formulation objectives were accomplished in two steps: the preliminary studies directed towards the formulation of stable emulsion systems and, the subsequent gellification stage, whereby stable emulsions were converted to emulgels. the emulgels so developed were subject to physicochemical characterization, ex vivo and in vivo evaluation and, investigation of drug release kinetics. a study of the literature revealed that an emulgel formulation of loratadine has not been developed so far. although, loratadine gel formulations have been developed [15], an emulgel has its own unique advantages. loratadine, being a highly lipophilic drug, is an appropriate candidate to be incorporated in an mahant et al. admet & dmpk 2(4) (2014) 254-271 256 oil-in-water emulgel. from the therapeutic standpoint, an emulgel would be more suitable for the treatment of skin allergies because of its emollient properties, attributed to the presence of an oily phase in the formulation. the emollience imparted by the oily phase would soothe the irritated skin. additionally, the gel phase present in the formulation would allow prolonged release of the drug, apart from rendering elegance to the product. experimental the materials used in the course of the present study are enlisted as follows: loratadine, carbopol 940 (qualikem fine chem ltd., vadodara), sodium lauryl sulfate (qualikem fine chem ltd., vadodara), cetrimide (nice chem pvt. ltd., kochi), propylene glycol (qualikem fine chem ltd., vadodara), light liquid paraffin (qualikem fine chem ltd., vadodara), tween 20 (qualikem fine chem ltd., new delhi), methyl paraben (qualikem fine chem ltd. new delhi), span 20 (sdfcl, mumbai), cetyl alcohol (sdfcl, mumbai), triethanolamine (nice chem pvt. ltd, kochi), dichloromethane (nice chem pvt. ltd, kochi), glyceryl monostearate (loba chemicals pvt. ltd, mumbai), methanol (rfcl ltd, new delhi), ethanol (changshu yangyuan chemicals, china). to achieve the goal of the present study, different emulsifying agents were employed, based upon their ionic nature. cetylpyridinium bromide (cetrimide) was taken as cationic surfactant, sodium lauryl sulfate as the anionic surfactant and, span 20 and tween 20 were used in the non-ionic category. in order to enhance product stability, glyceryl monostearate, cetyl alcohol and cetostearyl alcohol were incorporated as auxiliary emulsifying agents, in varying proportions. carbopol 940, which is a hydrophilic polyacrylate polymer, was employed as the gelling agent. light liquid paraffin formed the oil phase of the emulsion. in the formulation of emulgel, ethanol was used as a cosolvent for the drug, while propylene glycol was used as a cosolvent as well as humectant. butylated hydroxytoluene (bht) and ethylenediamine tetra acetic acid (edta) were added as the antioxidant and chelating agent, respectively. for their preservative action, methyl paraben and propyl paraben were added. triethanolamine was used to adjust the ph of the formulation between 6.3 and 6.5. drug-excipient compatibility studies in the preparation of emulgel formulation, the drug and excipients may interact as they are in close contact with each other, which could result in instability of drug or the formulation. for this reason, preformulation studies regarding drug-excipient interaction are very critical in selecting appropriate excipients. fourier-transform infra red (ftir) spectroscopy was employed to ascertain the compatibility of the drug with various excipients used. ftir spectra of the individual drug and drug with each excipient were obtained using sodium chloride discs. the samples were mixed with heavy liquid paraffin and the applied between the sodium chloride discs. by using these discs in the ftir spectrophotometer, the graphs were obtained, which were further studied for any change in peak i.e. shifting or disappearance of any peak with respect to peaks of the pure drug. the results of the study did not reveal any sign of incompatibility between the drug and the excipients used. analytical method uv spectrophotometry was used to analyze the drug concentration in the formulations. the absorption maximum of the drug was found to be ph dependent. the absorption maxima of the drug were admet & dmpk 2(4) (2014) 254-271 loratadine emulgels doi: 10.5599/admet.2.4.64 257 determined by dissolving the drug in different media and spectra were obtained using uv spectrophotometer. the absorption maxima (λmax) of loratadine in the selected media are reported in table 1. table 1. absorption maxima of loratadine in selected media formulation and preparation of emulgels preliminary studies. prior to the preparation of emulgel formulations, preliminary experimental batches were formulated wherein their respective emulsions were prepared without the drug and, checked for their homogeneity and stability. the formulae for the preliminary study batches are reported in tables 2-4. batches c1-c7 represent the emulsions prepared with cetrimide, while, s1-s7 designate the emulsions containing sodium lauryl sulphate. likewise, n1-n4 stand for emulsions prepared with span 20 and tween 20. the emulsion batches were formulated using the surfactants and auxiliary emulsifying agents in varying proportions, so as to achieve optimum product stability and consistency. the following paragraph describes the procedure adopted for the preparation of the emulsions, employing different quantities of cetrimide, sls and, span 20 and tween 20. the oil phase of emulsion was prepared by dissolving the oil soluble surfactant and butylated hydroxy toluene (bht) in light liquid paraffin. the aqueous phase was prepared by dissolving the water soluble surfactant in distilled water. methyl paraben and propyl paraben were dissolved in propylene glycol. propylene glycol solution was, then, dissolved in aqueous phase with constant stirring. both, the oil phase and the aqueous phase, were separately heated to 60-70°. next, the oil phase was added to the aqueous phase with constant homogenization. thereafter, the emulsions were allowed to cool to room temperature. table 2. composition of preliminary emulsion batches using cetrimide s. no. ingredients c1 (g) c2 (g) c3 (g) c4 (g) c5 (g) c6 (g) c7 (g) 1. liquid paraffin(light) 7.5 7.5 7.5 7.5 7.5 7.5 7.5 2. propylene glycol 5 5 5 5 5 5 5 3. cetrimide 0.15 0.15 0.15 0.15 0.15 0.5 0.3 4. glyceryl monostearate 1.5 1.5 1.5 1.5 5. cetyl alcohol 1.25 0.5 0.75 0.5 0.5 0.5 6. cetostearyl alcohol 1.25 0.75 7. methyl paraben 0.03 0.03 0.03 0.03 0.03 0.03 0.03 8. propyl paraben 0.01 0.01 0.01 0.01 0.01 0.01 0.01 9. distilled water (q.s.) 50 50 50 50 50 50 50 s. no. media absorption maxima (λmax) 1. 0.1n hcl 277 nm 2. citro phosphate buffer ph 5.0 267 nm 3. phosphate buffer ph 7.4 248 nm mahant et al. admet & dmpk 2(4) (2014) 254-271 258 table 3. composition of preliminary emulsion batches using sodium lauryl sulfate s. no. ingredients s1 (g) s2 (g) s3 (g) s4 (g) s5 (g) s6 (g) s7 (g) 1. liquid paraffin(light) 6.34 6.34 6.34 7.5 7.5 7.5 7.5 2. propylene glycol 5.64 5.64 5.64 5 5 5 5 3. sodium lauryl sulfate 1 1 1 1 1 1 1 4. glyceryl monostearate 2.5 2 1 1.5 1.75 5. cetyl alcohol 0.5 2.3 0.5 6. cetostearyl alcohol 1.25 1.25 0.75 7. methyl paraben 0.03 0.03 0.03 0.03 0.03 0.03 0.03 8. propyl paraben 0.01 0.01 0.01 0.01 0.01 0.01 0.01 9. distilled water (q.s.) 50 50 50 50 50 50 50 table 4. composition of preliminary emulsion batches using span 20 & tween 20 s. no. ingredients n1 (g) n2 (g) n3 (g) n4 (g) 1. liquid paraffin(light) 7.5 7.5 7.5 7.5 2. propylene glycol 5 5 5 5 3. span 20 1.35 2.1 1.35 2.1 4. tween 20 1.65 0.9 1.65 0.9 5. cetyl alcohol 0.5 0.5 6. edta 0.02 0.02 0.02 0.02 7. bht 0.02 0.02 0.02 0.02 8. methyl paraben 0.03 0.03 0.03 0.03 9. propyl paraben 0.01 0.0 1 0.01 0.01 10. distilled water (q.s.) 50 50 50 50 preparation of loratadine emulgels. the preliminary formulation studies were followed by the preparation of the final emulgel batches. prior to this, the formulae for emulgel batches were selected on the basis of homogeneity and stability testing of the preliminary emulsion batches. as a result, formulations c7, s7 and n4 were selected owing to their good consistency and adequate stability. henceforth, auxiliary emulsifying agents, glyceryl monostearate (1.75 %) and cetyl alcohol (0.5 %) were used in the formulation of emulgel with sls (as shown in table 5). glyceryl monostearate (1.5 %) and cetyl alcohol (0.5 %) were used in the formulation of emulgel with cetrimide (refer to table 6). similarly, for the preparation of emulgel with non-ionic surfactants, span 20 and tween 20 were employed in the concentrations of 2.1 % and 0.9 %, respectively (given in table 7). the emulgels were prepared by the following procedure: to begin with, the emulsion part of the emulgel was prepared. the oil phase of emulsion was prepared by dissolving oil soluble surfactant and bht in light liquid paraffin. in order to prepare the aqueous phase, edta was dissolved in distilled water. subsequently, the water soluble surfactant was added to the solution and dissolved. loratadine was dissolved in ethanol while methyl paraben and propyl paraben were dissolved in propylene glycol. thereafter, ethanol and propylene glycol solutions were dissolved in aqueous phase with constant stirring. both, the oil phase and aqueous phase were separately heated to 60-70 ° . oil admet & dmpk 2(4) (2014) 254-271 loratadine emulgels doi: 10.5599/admet.2.4.64 259 phase was added to aqueous phase with constant homogenization, until it cooled to room temperature. the gel phase was prepared by dispersing carbopol 940 in distilled water, which was kept overnight, followed by stirring at a moderate and constant speed. the ph of gel base was adjusted to 6-7 using triethanolamine. the emulsion prepared by this method was gellified in gel base in the ratio 1:1, with gentle stirring. in this manner, emulgels were obtained. table 5. composition of emulgels formulated with sodium lauryl sulfate s. no. ingredients s8 (g) s9 (g) s10 (g) 1. loratadine 1 1 1 2. ethanol 10 10 20 3. liquid paraffin (light) 5 5 5 4. propylene glycol 5 5 5 5. sodium lauryl sulfate 1 1 1 6. glyceryl monostearate 1.75 1.75 1.75 7. cetyl alcohol 0.5 0.5 0.5 8. methyl paraben 0.03 0.03 0.03 9. propyl paraben 0.01 0.01 0.01 10. bht 0.02 0.02 0.02 11. edta 0.02 0.02 0.02 12. triethanolamine q.s. q.s. q.s. 13. carbopol 940 0.25 0.5 0.75 14. distilled water (q.s.) 100 100 100 table 6. composition of emulgels formulated with cetrimide s. no. ingredients c8 (g) c9 (g) c10 (g) 1. loratadine 1 1 1 2. ethanol 10 20 20 3. liquid paraffin(light) 5 5 5 4. propylene glycol 5 5 5 5. glycerol 3 3 3 6. cetrimide 0.3 0.3 0.3 7. glyceryl monostearate 1.75 1.75 1.75 8. cetyl alcohol 0.5 0.5 0.5 9. methyl paraben 0.03 0.03 0.03 10. propyl paraben 0.01 0.01 0.01 11. bht 0.02 0.02 0.02 12. edta 0.02 0.02 0.02 13. triethanolamine q.s. q.s. q.s. 14. carbopol 940 0.25 0.5 0.75 15. distilled water (q.s.) 100 100 100 mahant et al. admet & dmpk 2(4) (2014) 254-271 260 table 7. composition of emulgel formulated with span 20 & tween 20 s. no. ingredients n5 (g) 1. loratadine 1 2. ethanol 10 3. liquid paraffin(light) 5 4. propylene glycol 5 5. span 20 2.1 6. tween 20 0.9 7. methyl paraben 0.03 8. propyl paraben 0.01 9. bht 0.02 10. edta 0.02 11. carbopol 940 0.5 12. triethanolamine q.s. 13. distilled water (q.s.) 100 characterization of emulgels the prepared emulgels were evaluated for the relevant physical parameters as described below. homogeneity study. the emulgels were visually inspected for color, homogeneity and consistency. determination of ph. emulgels (1 g) were accurately weighed and dispersed in 100 ml of distilled water. the ph of the dispersion was measured using ph meter (max instruments chandigarh, india), which was calibrated before each reading, with buffered solution at ph 4.0 and 7.0. rheological studies. in the context of semisolid dosage forms, viscosity is of paramount importance. hence, all the emulgels were checked for their viscosity using dv-1 prime (brookfield engineering laboratories, usa) viscometer, at 25 0 , at 12 rpm, using spindle s64. centrifugation study. centrifugation study is a useful means to check the stability of the prepared emulgels. these studies were performed one week after the preparation of emulgels. centrifugation was carried out using minicentrifuge at 3000 rpm for duration of 30 minutes. microscopic evaluation. globule size of the prepared emulgels was determined by optical microscopy. a compound microscope was used for this purpose and the globules were observed under 40x magnification. prior to observation, the eye-piece micrometer was calibrated with a stage micrometer and calibration factor was obtained. subsequently, mean globule size was calculated [16]. temperature swing test. temperature swing test gives the formulator an idea about the stability of the formulation in extreme temperature conditions. therefore, this test is frequently employed in the evaluation of topical semisolid dosage forms. for the purpose of this test, the formulations were subjected to freeze and thaw cycles. one cycle comprised of 8 h storage at -4 0 and another was carried out for 16 h at 40 0 . this was performed for 2 days. the formulations were visually inspected at the end of the test to ascertain their stability [13]. drug content determination. in order to determine the drug content of the emulgels, two different methods were developed and validated in the laboratory: one for the emulgels containing cationic and anionic surfactant and, another for that formulated with non-ionic surfactants. the emulgel formulated with span 20 and tween 20 was analyzed using 0.1 m hcl as the solvent. accurately weighed, 1 g of admet & dmpk 2(4) (2014) 254-271 loratadine emulgels doi: 10.5599/admet.2.4.64 261 prepared emulgel was dispersed in 0.1 m hcl solution. then, this dispersion was further diluted with 0.1 m hcl and analyzed using uv spectrophotometer. for drug content determination of emulgels prepared with sls and cetrimide, the following method was employed. in the first step, accurately weighed 500 mg emulgel was dissolved in dichloromethane and shaken for 2 h. the solution was filtered, and transferred to watch glass. by heating the solution on watch glass, dichloromethane was evaporated, leaving behind the residue. to the residue, 10 ml of methanol was added and, this was further diluted with phosphate buffer ph 7.4 and analyzed using uv spectrophotometer. stability studies. stability studies of the prepared emulgels were carried out at two different temperature conditions, that is, at 4° ± 2° and 25° ± 2° for 3 months. to perform these studies, the formulations were packed in 15 ml glass vials. the selected formulations were analyzed for consistency, drug content and ph, at fortnightly intervals. spreadability study. spreadability is an important criterion for evaluating a topical formulation, since it determines the ease of its application. spreadability is the term expressed to denote the extent of area over which the formulation readily spreads upon application to skin or in the affected part. thus, it also affects the therapeutic benefit derived from the formulation. the spreadability of the emulgels was measured by an apparatus suggested by mutimer [17]. it consists of wooden block, attached to pulley at one end. this property is determined upon the basis of ‘slip’ and ‘drag’ characteristics of the emulgel. spreadability is expressed in terms of time in seconds taken by two slides to slip off from emulgel, placed in between the slides under the direction of certain load. the lesser the time taken for separation of two slides, better the spreadability. it is calculated by using the formula given in equation 1: s = m x l/t (1) where m = weight tied to upper slide (g), l = length of glass slides (cm), t = time taken to separate the slides (s) [18]. an excess of emulgel (about 2 g) was applied on the ground slide. then, the emulgel was sandwiched between the ground slide and the second slide (upper slide), having the same dimensions as that of fixed ground slide. the upper slide was provided with a hook. some weight was placed on the top of two slides for 5 minutes to expel air and to provide uniform film of the emulgel between the two slides. a known weight was placed in the pan attached to the pulley with the help of hook. the time (in seconds) taken by the upper slide to cover a distance 5 cm was noted [12]. ex vivo drug release study. ex vivo drug release study of the prepared emulgels (c10, s10 and n5) was performed using male albino rats, weighing 180-200 g. the rats were sacrificed by cervical dislocation and their skin was removed and shaved. skin explants with a contact area of 1.53 cm 2 were mounted on receptor compartment of the franz diffusion cells. the receptor compartment was filled with citro phosphate buffer (ph 5.0). thereafter, the emulgel was applied on the outer skin surface through donor compartment. to maintain sink conditions, the medium in receptor compartment was stirred using magnetic stirrer at 600 rpm. samples (3 ml) were withdrawn from the receptor compartment and replenished with an equal volume of fresh medium at hourly intervals, up to 8 hours. the amount of drug released was determined by analyzing the samples spectrophotometrically. the study was performed in triplicate for each formulation. thereafter, the mean and standard deviation values were calculated. mahant et al. admet & dmpk 2(4) (2014) 254-271 262 drug release kinetics studies. drug release kinetics was investigated by fitting the data obtained from ex vivo drug release studies into various drug release kinetics models. for the purpose of this study, zero order, first order and higuchi plots were prepared. the zero order plots were prepared by plotting cumulative percent drug permeated on vertical axis against time (in hours) on horizontal axis. in order to study the formulations for first order kinetics, log cumulative percent drug remaining was plotted versus time (in hours). furthermore, cumulative percent drug permeated taken on y-axis versus square root of time (in hours) on x-axis gave the higuchi plots. the most appropriate model was selected based upon goodness-of fit test. the plot which resulted in maximum value of correlation coefficient, r 2 , was considered to be the most appropriate for the given formulation [19,20]. skin irritation study. skin irritation study is a prerequisite for any topical formulation as it enables to determine its skin irritation potential. prior to conducting the animal studies, approval was obtained from the institutional animal ethical committee, mmu, mullana-ambala (mmcp/iec/11/10). all the animals were treated according to the guidelines provided by the institutional animal ethical committee. the skin irritation studies were performed on guinea pigs. the dorsal skin region of the guinea pigs was shaved using clipper one day before study. the animals were treated with test formulations for 3 days and the treated skin was examined visually for erythema and edema. the observations were compared with that of the untreated skin, which was taken as control [19]. in vivo anti-allergic activity. the test for anti-allergic activity was performed on male albino rats. 500 mg of the selected emulgels was applied to the dorsal skin of the animals and the formulation was evenly spread until it was completely absorbed. the animals (n = 6) were divided into different groups and were given the following treatment: group i was treated with formulation s10, group ii was treated with formulation c10, group iii with formulation n5, group iv was treated with histamine alone, while group v served as the control and was, therefore, left untreated. hence, groups i, ii and iii received treatment with the prepared emulgels. subsequent to the application of the formulations, 50 µl of histamine hydrochloride solution (0.1 µg/ml in a solution of water/glycerol: 50/50 v/v) was injected into the upper layers of the dorsal skin of each of group i, ii, iii and iv. the effect of histamine-induced cutaneous reactions was noted by visual inspection at definite intervals, i.e. 15, 30 and 60 minutes. the observed responses were evaluated with respect to those of groups iv and v [20]. statistical analysis. statistical analysis was performed using student’s paired t-test using graphpad version 2.01, san diego, ca. the data was considered significant at (p < 0.05). results and discussion physical evaluation of emulgels homogeneity of a topical formulation is gives the formulator an idea about the consistency, stability and elegance of the final product. incorporation of propylene glycol as a humectant and carbopol 940 as the gelling agent contributed towards the homogeneity of the emulgels. the findings of this study formed the basis for selecting the composition of the final study batches. the results of homogeneity study performed on preliminary emulsion batches are shown in table 8-10. the compositions which gave stable and homogeneous emulsions were taken up for formulation into final emulgels. further, the emulgels so prepared were also tested for their homogeneity. the results are admet & dmpk 2(4) (2014) 254-271 loratadine emulgels doi: 10.5599/admet.2.4.64 263 reported in table 11. all the formulations were white in color. it is evident from the observations of this study that c10, s10 and n5 exhibited adequate homogeneity and were, therefore, subject to further evaluation. table 8. observation table for preliminary batches of emulsion with cetrimide s. no. test formulation code c1 c2 c3 c4 c5 c6 c7 1. homogeneity good consistency foamy appearance, good consistency foamy appearance, good consistency too viscous, clumping observed good appearance but too viscous good consistency and elegant appearance good consistency and elegant appearance 2. stability* stable stable stable _ _ stable stable *performed using minicentrifuge at 7000 rpm for 5 min table 9. observation table for preliminary batches of emulsion with sls s. no. test formulation code s1 s2 s3 s4 s5 s6 s7 1. homogeneity very thick good consistency undesirable due to excessive foaming phase separation observed poor consistency cracking observed good consistency 2. stability* stable stable unstable stable unstable stable *performed using minicentrifuge at 7000 rpm for 5 min table 10. observation table for preliminary batches of emulsion with span 20 and tween 20 s. no. test formulation code n1 n2 n3 n4 1. homogeneity phase separation within 1h of preparation poor consistency too viscous good consistency 2. stability* unstable unstable stable stable *performed using minicentrifuge at 7000 rpm for 5 min table 11. observation table for selected batches of emulgels observation formulation code c8 c9 c10 s8 s9 s10 n5 homogeneity study unstable, precipitation of drug observed stable but poor consistency good consistency very low viscosity low viscosity good consistency good consistency thus, formulations c10, s10 and n5 were tested for their physical properties like ph, spreadability, drug content, average globule size and effect of centrifugation. the ph was determined at room temperature using digital ph meter. these values were well within the ph range of the human skin. this was achieved by the incorporation of triethanolamine [21,22]. when subject to centrifugation, the emulgels did not show any sign of phase separation or change in the appearance. this implies that the prepared formulations are stable to mechanical stress. as reported in the literature too, formulations showing no phase separation are considered to be thermodynamically stable systems [23,24]. temperature swing test. temperature swing test was performed in b.o.d incubator by storing the formulations, initially, at -4 °c for 8 h and, later, at 40 °c for the next 16 h. the emulgels did not show any mahant et al. admet & dmpk 2(4) (2014) 254-271 264 sign of inconsistency at the end of the temperature swing test. this further strengthens the outcome of centrifugation study, suggesting optimum stability of the selected formulations [23]. the drug content of all the formulations was found to be within the desired range. globule size/average droplet size is also considered to be an important factor in context of physical stability of the product [25,26]. therefore, mean globule size was calculated for the prepared emulgels. mean globule size is affected by a number of process and formulation variables. since, our study focused on the effect of formulation variables, the findings of average globule size determination may be correlated with the surfactants used in the formulations. as given in literature, incorporation of non-ionic surfactants tends to lower the droplet size in the emulsions stabilized by acrylate polymers [22]. formulation n5, prepared with non-ionic surfactants, had the lowest average globule diameter, which supports the findings cited in the literature. the results are tabulated in the table 12. table 12. observation table for physical evaluation of selected emulgels *c10: emulgel containing cetrimide; s10: emulgel containing sls; n5: emulgel containing tween 20 & span 20 rheological study. viscosity has profound significance with respect to the performance of topical products. product characteristics, such as spreadability, ease of application, drug release and stability are closely linked to the viscosity of the formulation. formulation variables, viz. gelling agent, surfactant, auxiliary emulsifiying agent, cosolvent and oil phase influence the rheological properties of the formulation. the results of viscosity determination are reported in table 13. as evident, the viscosity of the prepared emulgels followed the order: n5>c10>s10. table 13. observation table for viscosity measurement *c10: emulgel containing cetrimide; s10: emulgel containing sls; n5: emulgel containing tween 20 & span 20 **mean of triplicate reading generally, the viscosity of a formulation increases with the increase in the concentration of the gelling agent [13]. conversely, n5 was found to be the most viscous formulation, despite the low concentration of carbopol 940. on the other hand, el-megrab et al. report that the presence of alcohol in a formulation decreases its viscosity [27]. formulations c10 and s10 contain twice the amount of alcohol as compared to n5. s. no. formulation code* ph spreadability (g.cm/s) drug content (%) average globule size (µm) centrifugation study 1. c10 6.63±0.015 13.32±0.759 102.62±1.63 2.32 stable 2. s10 6.67±0.03 18.84±1.58 97.61±1.24 1.82 stable 3. n5 6.81±0.02 20.25±1.403 95.42±0.56 0.875 stable s. no. formulation code* spindle used rpm viscosity** (cps) 1. c10 s64 12 43404±13.5 2. s10 s64 12 42391±56.66 3. n5 s64 12 44443±42.55 admet & dmpk 2(4) (2014) 254-271 loratadine emulgels doi: 10.5599/admet.2.4.64 265 moreover, it has also been mentioned in the literature that non-ionic surfactants increase the yield stress and viscosity of the emulsions stabilized using acrylates [22]. in view of the foregoing facts, the presence of non-ionic surfactants in n5 and a greater percentage of alcohol in c10 and s10 explains the observed order of viscosity in the emulgels. it may also be added that the inclusion of glycerol (3 %) in c10 led to a more viscous formulation than s10. stability study. stability study was performed with the view to ascertain that the prepared formulations retain their physicochemical and therapeutic attributes. no significant change was observed in the physical appearance, consistency, percentage drug content and ph of any emulgel, throughout the period of storage. thence, it was concluded that all the three types of surfactants, namely, cetrimide, sls and, span 20 and tween 20 resulted in equally stable formulations. the observations for percent drug content and ph determination are tabulated in tables 14(a) and (b), respectively. table 14(a). observation table for stability study (percent drug content) s. no. temperature condition (°c) time period (days) c10 s10 n5 1. 4 15 102.04±1.11 97.24±0.50 95.39±1.26 2. 4 30 101.88±1.63 97.27±0.81 95.09±1.80 3. 4 45 102.28±0.99 97.2±0.58 95.22±1.08 4. 4 60 102.20±1.61 96.70±1.41 94.89±0.785 5. 4 75 102.04±0.99 97.27±0.63 95.03±0.93 6. 4 90 100.81±1.44 97.34±1.43 95.02±0.924 7. 25 15 102.61±0.75 97.23±0.82 95.61±1.08 8. 25 30 102.2±0.89 97.41±0.47 95.35±1.08 9. 25 45 102.37±0.38 97.27±0.63 95.61±1.33 10. 25 60 101.06±0.62 97.23±1.05 95.55±1.57 11. 25 75 102.20±1.61 97.34±0.75 95.42±0.89 12. 25 90 100.81±1.44 97.16±0.96 95.22±0.986 *mean of triplicate reading table 14(b). observation table for stability study (ph determination) s. no. temperature condition (°c) time period (days) c10 s10 n5 1. 4 15 6.62±0.01 6.66±0.005 6.82±0.01 2. 4 30 6.62±0.01 6.66±0.01 6.82±0.02 3. 4 45 6.62±0.015 6.67±0.005 6.82±0.015 4. 4 60 6.64±0.01 6.67±0.01 6.85±0.015 5. 4 75 6.63±0.015 6.65±0.01 6.84±0.015 6. 4 90 6.64±0.015 6.65±0.01 6.82±0.017 7. 25 15 6.64±0.005 6.68±0.005 6.81±0.005 8. 25 30 6.63±0.005 6.67±0.011 6.82±0.1 9. 25 45 6.64±0.005 6.67±0.01 6.84±0.015 10. 25 60 6.63±0.005 6.68±0.01 6.83±0.1 11. 25 75 6.61±0.01 6.69±0.005 6.83±0.02 12. 25 90 6.62±0.01 6.68±0.02 6.83±0.005 *mean of triplicate reading spreadability study. conventionally, spreadability of a formulation is inversely related to its viscosity [28]. however, in our study the outcomes of the spreadability testing differ from the general argument. this observation may be explained by the globule size of the respective formulations. it was found that mahant et al. admet & dmpk 2(4) (2014) 254-271 266 emulgel n5 had the least average globule size, followed by s10 and c10. it has been reported that lower globule size of emulsion phase results in better spreadability of the product [22]. hence, spreadability of the prepared emulgels decreased in the order: n5>s10>c10. ex vivo drug release study. the ex vivo drug release study was performed on male albino rats. with the view to analyze the drug release from the three emulgels, cumulative amount of drug released was plotted against time. figure 1 depicts the drug release obtained from the emulgels for a period of 8 hours. amount of drug release was found to decrease in the following order: c10>s10>n5. the results obtained for ex vivo drug release study can be explained by taking into account the various factors that are known to influence the availability of the drug from a topical formulation. these factors include viscosity, mean droplet size, gelling agent, auxiliary emulsifying agent, presence of cosolvents and surfactants. it is well reported in literature that viscosity of a formulation strongly affects the release of drug [29]. as stated earlier, an increase in viscosity decreases the drug release. however, this was not the case in our study, indicating that other factors contribute more strongly towards drug release. mahdi et al. have reported the effect of globule size on drug release. their findings indicate that lower globule size results in higher release [30]. on the contrary, in our study, formulation c10 with the largest globule size showed maximum drug release. capakova et al. emphasized the importance of selecting appropriate gelling agent and its concentration for optimum drug availability. their group evaluated the effect of varying concentration of carbopol 980 upon the release of loratadine from hydrogels. as per their report, increasing the concentration of carbopol 980 led to lower drug release [29]. when compared to our findings, formulation c10, containing greater amount of gelling agent gave maximum release of the drug. this suggests that in case of emulgels, gelling agent concentration does not play a dominant role in controlling drug release. in relation to auxiliary emulsifying agents, literature reports that the presence of glyceryl monostearate in a formulation, increases drug release significantly [31]. in our study too, drug release was found to be highest from c10, followed by s10, both containing glyceryl monostearate as the auxiliary emulsifying agent. from the perspective of cosolvents, ethanol and propylene glycol have been used to enhance solubilization of loratadine. song et al state that propylene glycol promotes the uptake of the drug into the skin by modifying the driving force for its diffusion [32]. on the other hand, the solubilizing property of ethanol facilitates drug release from the formulation [27]. further, ethanol also decreases the viscosity of carbopol gel, thereby, affecting drug release [21]. in the present study, formulation c10 and s10, contain higher concentration of ethanol, as compared to n5. this accounts for greater drug release from the former two emulgels. form the discussion, it may be inferred that ethanol and glyceryl monostearate were responsible for greater drug release from c10 and s10. since, both the ingredients were present in equal amounts in the two formulations, they can be assumed to have contributed equally towards release of the drug. accordingly, the above argument leads to the conclusion that the type of surfactant plays a major role in promoting drug release. in the present investigation, effect of different types of surfactants i.e. cationic, anionic and non-ionic, has been compared. admet & dmpk 2(4) (2014) 254-271 loratadine emulgels doi: 10.5599/admet.2.4.64 267 a higher drug release was obtained from formulation c10 as compared to s10, signifying cetrimide’s ability to produce greater reduction in interfacial tension between the oily phase of the emulsion and aqueous layer of the dispersion medium, thereby, facilitating the penetration of the dissolution medium into the emulgels. this permits greater drug release from the formulation [33]. drug release kinetics study. the data obtained from ex vivo drug release study was fitted into various kinetic models and it was processed to obtain the relevant plots i.e. zero order, first order and higuchi plot. the r 2 values obtained for the three emulgels for the various plots are listed in table 15. the r 2 values clearly suggest that the drug release from the all the formulations is concentration independent on account of the highest linearity for zero order plots. figure 1. graph showing comparison between cumulative amount released (µg) versus time (h) for different formulations table 15. r 2 values obtained for various plots *c10: emulgel containing cetrimide; s10: emulgel containing sls; n5: emulgel containing tween 20 & span 20 skin irritation test. the results of skin irritation test were assessed on the basis of draize-fhsa (federal hazardous substance act) scoring system [11] and the scores obtained for each formulation are given in table 16. since, there was no sign of erythema or edema on the skin of the guinea pigs treated with the test formulations, a score of zero was assigned to each, with regard to skin irritation potential. hence, all the emulgel formulations passed the skin irritation test. s. no. formulation code* r 2 values zero order first order higuchi plot 1. c10 0.983 0.9753 0.8356 2. s10 0.9833 0.9787 0.837 3. n5 0.9776 0.9725 0.8221 mahant et al. admet & dmpk 2(4) (2014) 254-271 268 table 16. scores for erythema and edema produced in guinea pigs s. no. formulation code* score 24 h 72 h 1. c10 0 0 2. s10 0 0 3. n5 0 0 *c10: emulgel containing cetrimide; s10: emulgel containing sls; n5: emulgel containing tween 20 & span 20 in vivo antiallergic activity. skin testing for immediate type 1 hypersensitivity, prick and intracutaneous techniques were performed. this test consists of introduction of allergen extract or histamine into the skin resulting in ige mediated allergic response, characterized by immediate wheal and flare reaction, due to activation of mast cells releasing vasoactive agents, which cause both, extravasation and vasodilation. the reactions are assessed by the degree of redness and swelling and the size of the wheal produced. to evaluate the efficacy of the tested formulations, the potential irritation due to the cutaneous reactions induced by histamine injection on the dorsal surface of the rats was estimated. the visual examination of skin, 60 minutes after the administration of the developed formulations, revealed that the symptoms of hypersensitivity i.e. redness and itching, were absent in groups i, ii and iii. this indicates that loratadine emulgels were effective in alleviating the symptoms of skin allergy. conclusions topical drug delivery constitutes a major part of the therapy for skin diseases or disorders. of late, topical drug delivery systems have assumed even greater importance owing to their localized action and lesser plausibility of systemic toxicity. in the recent years, emulgels have been explored as an alternative to conventional topical formulations to deliver several topical agents. in the present study, it was attempted to incorporate loratadine into emulgel bases, using anionic, cationic and non-ionic surface-active agents. the emulgels were successfully prepared and exhibited optimum physical characteristics. all the three surfactants were found to produce equally stable formulations. loratadine release from emulgels was found to be influenced by the interactions of the drug with surfactants, auxiliary emulsifying agents and cosolvents used in the formulation. these interactions affect the partitioning of the drug between the aqueous and oily phase of the formulation. the ex vivo drug release was found to be greatest for the emulgel formulated with cetrimide. moreover, drug release from the formulations was independent of drug concentration. the formulations were non-irritant to the skin and effective in alleviating the symptoms of skin allergy. acknowledgements: the authors are highly grateful to dr. anroop nair for his guidance and comments which were useful in designing the study. admet & dmpk 2(4) (2014) 254-271 loratadine emulgels doi: 10.5599/admet.2.4.64 269 references [1] p.s. gelone, remington: science and practice of pharmacy, wolters kluwer pvt. ltd, new delhi, india, 2006, p. 1615. [2] a. pentz, c. els, o. coetzee, r.j. green, current allergy and clinical immunology 27 (2014) 9-13. http://www.allergysa.org/journals/march2014/anidealchildrenschest.pdf [3] p. small, canadian family physician 35 (1989) 1859-1862. http://www.ncbi.nlm.nih.gov/pmc/articles/pmc2280873/pdf/canfamphys00163-0157.pdf [4] allergy statistics, http://www.aaaai.org/about-the-aaaai/newsroom/allergy-statistics.aspx (7th aug, 2014). [5] allergies statistics and facts, http://www.healthline.com/health/allergies/statistics (7th aug, 2014). [6] disease summaries: contact dermatitis, http://www.worldallergy.org/public/allergic_diseases_center/contactdermatitis/ (7th aug, 2014). [7] c. motala, current allergy and clinical immunology 22 (2009) 71-74. http://reference.sabinet.co.za/webx/access/electronic_journals/caci/caci_v22_n2_a6.pdf [8] p. singh, m. bajpai, s. srivastava, journal of pharmaceutical science and research 3(2011) 11701175. http://www.jpsr.pharmainfo.in/documents/volumes/vol3issue04/jpsr%2003110407.pdf [9] v. singla , s. saini , b. joshi, ac rana, international journal of pharma and biosciences 3(2012) 485-498. http://www.ijpsr.com/v5i5/8%20vol.%205,%20issue%205,%20may%202014,ijpsr,%20re%2011 98,%20paper%208.pdf [10] s.p. stanos, journal of pain and symptom management 33(2007) 342-355. http://www.jpsmjournal.com/article/s0885-3924%2807%2900003-6/fulltext [11] m. i. mohamed, the aaps journal 6(2004) 1-7. http://www.ncbi.nlm.nih.gov/pmc/articles/pmc2751251/pdf/12248_2008_article_6381.pdf [12] r. khullar, d. kumar, n. seth, s. saini, saudi pharmaceutical journal 20(2012) 63-67. http://www.ncbi.nlm.nih.gov/pmc/articles/pmc3745000/pdf/main.pdf [13] m. shahin, s. a. hady, m. hammad , n. mortada, aaps pharmsci tech 12 (2012) 239-247. http://www.ncbi.nlm.nih.gov/pmc/articles/pmc3066381/pdf/12249_2011_article_9583.pdf [14] r. khullar, s. saini, n. seth, a.c. rana, international journal of pharmaceutical and biological sciences 1(2011) 117-127. http://www.ijpbs.com/ijpbsadmin/upload/ijpbs_50c82835a2df7.pdf [15] hatto walch, us20030129209a1 (2003). http://www.google.com/patents/us20030129209 [16] a. pasupathi, p. palanisamy, b. jayakar, r. m. chandira, b. s. venkateswarlu, indian journal of research in pharmacy and biotechnology 1(2013) 95-105. http://www.ijrpb.com/current%20issues/1/v1/ijrpb%201_20%20page%2095-105.pdf [17] m. n. mutimer, c. riffkin, j. a. hill, m. e. glickman, g. n. cyr, journal of the american pharmaceutical association 45(1956) 212–218. [18] a. jain, p. deveda , n. vyas, j. chauhan, h. khambete, s. jain, international journal of pharmaceutical research and development 2(2011) 18-25. http://www.ijprd.com/article_no_257_feb_11_03.html [19] b. j. saqib, m. i. howard, handbook of cosmetic science and technology, informa healthcare, new york, usa, 2009, p. 471-474. [20] d. l. m. moretti, a. t. peana, v. sanna, international journal of pharm tech research 2(2010) 863-869. http://sphinxsai.com/sphinxsaivol_2no.1/pharmtech_vol_2no.1/pharmtech_vol_2no.1pdf/pt=1 31%20%28863-869%29.pdf http://www.allergysa.org/journals/march2014/anidealchildrenschest.pdf http://www.ncbi.nlm.nih.gov/pmc/articles/pmc2280873/pdf/canfamphys00163-0157.pdf http://www.aaaai.org/about-the-aaaai/newsroom/allergy-statistics.aspx http://www.healthline.com/health/allergies/statistics http://www.worldallergy.org/public/allergic_diseases_center/contactdermatitis/ http://reference.sabinet.co.za/webx/access/electronic_journals/caci/caci_v22_n2_a6.pdf http://www.jpsr.pharmainfo.in/documents/volumes/vol3issue04/jpsr%2003110407.pdf http://www.ijpsr.com/v5i5/8%20vol.%205,%20issue%205,%20may%202014,ijpsr,%20re%201198,%20paper%208.pdf http://www.ijpsr.com/v5i5/8%20vol.%205,%20issue%205,%20may%202014,ijpsr,%20re%201198,%20paper%208.pdf http://www.jpsmjournal.com/article/s0885-3924%2807%2900003-6/fulltext http://www.ncbi.nlm.nih.gov/pmc/articles/pmc2751251/pdf/12248_2008_article_6381.pdf http://www.ncbi.nlm.nih.gov/pmc/articles/pmc3745000/pdf/main.pdf http://www.ncbi.nlm.nih.gov/pmc/articles/pmc3066381/pdf/12249_2011_article_9583.pdf http://www.ijpbs.com/ijpbsadmin/upload/ijpbs_50c82835a2df7.pdf http://www.google.com/patents/us20030129209 http://www.ijrpb.com/current%20issues/1/v1/ijrpb%201_20%20page%2095-105.pdf http://www.ijprd.com/article_no_257_feb_11_03.html http://sphinxsai.com/sphinxsaivol_2no.1/pharmtech_vol_2no.1/pharmtech_vol_2no.1pdf/pt=131%20%28863-869%29.pdf http://sphinxsai.com/sphinxsaivol_2no.1/pharmtech_vol_2no.1/pharmtech_vol_2no.1pdf/pt=131%20%28863-869%29.pdf mahant et al. admet & dmpk 2(4) (2014) 254-271 270 [21] c. m. setty, s. r. babubhai, i. b. pathan, international journal of pharmacy and pharmaceutical sciences 2(2010) 70-73. http://www.ijppsjournal.com/vol2suppl1/337.pdf [22] e kettler, c.b. muller, r. klemp, m hloucha, t. doring, w. von rybinski, w. richtering, progress in colloid and polymer science: surface and interfacial forcesfrom fundamentals to applications, springer-verlag berlin heidelberg, 2008, 91. [23] m. s. alam, m.s. ali, n. alam, m. i. alam, t. anwer, f. imam, m. d. ali, m. r. siddiqui, m. shamim, journal of applied pharmaceutical science 2(2012) 148-158. http://www.japsonline.com/admin/php/uploads/684_pdf.pdf [24] j. d. modi, j. k. patel, international journal of pharmacy and pharmaceutical science research 1(2011) 6-12. http://urpjournals.com/tocjnls/24_2.pdf [25] t. mahmood, n. akhtar, s. manickam, journal of nanobiotechnology 12(2014) 12-20. http://www.jnanobiotechnology.com/content/12/1/20 [26] t. mahmood, n. akhtar, the scientific world journal 2013(2013) 1-7. http://www.hindawi.com/journals/tswj/2013/153695/ [27] n. a. el-megrab, h. m. el-nahas, g. f. balata, saudi pharmaceutical journal, 14(2006) 155-162. [28] m. rao, g. sukre, s. aghav, m. kumar, journal of pharmaceutics, 2013(2013) 1-9. http://www.hindawi.com/journals/jphar/2013/501082/ [29] z. capkova, z. vitkova, m. subova, acta facultatis pharmaceuticae universitatis comenianae 52(2005) 73-78. https://www.fpharm.uniba.sk/fileadmin/user_upload/admin/acta_facultatis/tomus_lii/09__capkova__z._-_vitkova__z._-_subova__m.pdf [30] e. s. mahdi, a. m. noor, m. a. sattar, international journal of nanomedicine 6(2011) 2499-2512. http://www.ncbi.nlm.nih.gov/pmc/articles/pmc3439863/ [31] a.a.m. lira, e. a. sester, a. l. m. carvalho, r. r. strattmann, m. m. albuquerque, a. g. wanderley and d. p. santana, aaps pharmsci tech 9(2008) 163-168. http://www.ufpi.br/subsitefiles/ppgcf/arquivos/files/1.pdf [32] j. h. song, s. c. chin, drug development and industrial pharmacy 35(2009)897-903. [33] d.a. el-rhman, m. a. nabarawi, s.a. abdel halim, d. a. helal, indian streams research journal 4(2014)1-14. http://www.isrj.net/uploadeddata/4929.pdf http://www.ijppsjournal.com/vol2suppl1/337.pdf http://www.japsonline.com/admin/php/uploads/684_pdf.pdf http://urpjournals.com/tocjnls/24_2.pdf http://www.jnanobiotechnology.com/content/12/1/20 http://www.hindawi.com/journals/tswj/2013/153695/ http://www.hindawi.com/journals/jphar/2013/501082/ https://www.fpharm.uniba.sk/fileadmin/user_upload/admin/acta_facultatis/tomus_lii/09_-_capkova__z._-_vitkova__z._-_subova__m.pdf https://www.fpharm.uniba.sk/fileadmin/user_upload/admin/acta_facultatis/tomus_lii/09_-_capkova__z._-_vitkova__z._-_subova__m.pdf http://www.ncbi.nlm.nih.gov/pmc/articles/pmc3439863/ http://www.ufpi.br/subsitefiles/ppgcf/arquivos/files/1.pdf http://www.isrj.net/uploadeddata/4929.pdf admet & dmpk 2(4) (2014) 254-271 loratadine emulgels doi: 10.5599/admet.2.4.64 271 appendix table a1. comparison of ftir peaks of pure drug versus different excipients s. no. wavelength (cm -1 ) functional groups interference pure drug drug + cetrimide drug+ cetyl alcohol drug + glyceryl monostearate drug + sodium lauryl sulfate 1. 2960, 2846 2909, 2774 2963, 2845 2968, 2850 2953, 2864 (c-h) aromatic stretch no 2. 1690 1692 1695 1710 1688 c=o (ester group) no 3. 1446 1576 1456 1448 1437 c=c (aromatic) no 4. 1226 1223 1234 1240 1225 c-n (amines) no ©2014 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ chromatographic technique to support admet&dmpk in early drug discovery 71 admet & dmpk 6(2) (2018) 71-73; doi: 10.5599/admet.559 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index editorial chromatographic technique to support admet&dmpk in early drug discovery klara valko 1,2 associate editor admet & dmpk 1 bio-mimetic chromatography ltd, business & technology centre, stevenage, sg1 2dx united kingdom 2 department of biological and pharmaceutical chemistry, ucl school of pharmacy london wc1n 1ax united kingdom e-mail: k.valko@ucl.ac.uk and klara_valko@bio-mimetic-chromatography.com; tel.: +44 7521 989558 the drug discovery process usually involves the discovery of a potent new chemical entity which has shown some activity on a target that is believed to be relevant in a certain type of disease. however, the structure and property of these putative drug molecules may have to be modified in order to ensure that they are able to exert the desired in vivo pharmacological effect. the candidate molecules usually have to be absorbed from the gastrointestinal system and therefore they must show appropriate solubility and permeability to be able to reach the target enzyme in vivo with a sufficient free biophase concentration at the site of action. this means that the physicochemical properties and the protein and phospholipid binding has to be optimized in order to achieve the desired in vivo admet and dmpk profile. the adme studies usually require animal experiments or the use of biological samples, for example tissues which is expensive and involves time-consuming procedures. new technologies and approaches can accelerate this process. hplc methodology is well established and used throughout the drug discovery process for analytical measurements of a compound’s purity and concentration in various biological samples such as blood, plasma and tissues. during the hplc separation process, the compounds distribute between the mobile and the stationary phases. as different compounds have a different interaction with the stationary phase that is modified by the mobile phase, they move at a different speed through the chromatographic system resulting in different retention times. as these retention times are characteristic of the compound and the hplc system and can be related to a compound’s dynamic distribution coefficient, they can be used to describe the properties of the compounds. when biomimetic stationary phases are used during chromatography, such as albumin, glycoprotein, and phospholipids, the calibrated retention times can be converted to binding constants representing these body components. these data can then be used to estimate the compound’s in vivo distribution behavior. in view of the important roles of hplc methodology, admet & dmpk devoted a special issue to cover chromatographic technique in early drug discovery. contributions were invited and received from several researchers who are well respected in this field to enlighten the readers of the more recent developments and applications in chromatography and in separation science as applied in support of drug discovery. this is presented as both original scientific papers describing the original contributions or review articles. the application of chromatography to predict skin penetration of compounds by marti rosés and his http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:k.valko@ucl.ac.uk mailto:klara_valko@bio-mimetic-chromatography.com 72 team at university of barcelona, spain is described [1]. the application of pressure assisted microemulsion electro-kinetic chromatography for lipophilicity determination has also been described by the same research group in a second contribution [2]. automated measurements using biomimetic chromatography in support of drug discovery programme research at gsk are reviewed by shenaz bunally and rob young [3]. similar efforts in pfizer has been reviewed by gilles goetz and marina shalajeva in their contributions including their recent application of supercritical chromatography to measure the embedded polar surface area (epsa) of molecules that could be related to their permeability [4]. klara valko and co-workers from genervon biopharmaceuticals llc (california, usa) and bicycle therapeutics (cambridge, united kingdom) have investigated the application of biomimetic chromatography to estimate lipophilicity, protein and phospholipid binding of new modalities such as linear and cyclic peptides as potential therapeutics [5]. the in vivo distribution behavior, cellular accumulation and brain penetration have been estimated using biomimetic hplc measurements for gm6, a linear peptide discovered by genervon biopharmaceuticals, that has shown therapeutic potential in vivo for amyotrophic lateral sclerosis (als) [6]. oleg krokhine and his research group at the university of manitoba, canada have demonstrated the possibility to predict peptide interaction with immobilized artificial membranes using over 30 thousand peptides from tryptic digest to estimate their phospholipid binding and have compared these interactions with traditional reversed-phase chromatography that is most widely used for lipophilicity measurements [7]. several papers belonging to the chromatography thematic issue will be published in the next issue of admet and dmpk including the papers of ciesla et al. [8], tsopelas et al. [9] and caron et al. [10]. lukasz ciesla and his colleagues at the university of alabama, usa have presented a new and innovative application of the immobilized artificial membrane stationary phase to bind membrane receptors on the surface called cellular membrane affinity chromatography (cmac) that can enable the “fishing out” of potent molecules from complex mixtures [8]. fotios tsopelas, chrysanthos stergiopoulos and anna tsantili-kakoulidou have presented a comprehensive review article on the application of immobilized artificial membrane chromatography (iam) in support of the drug design and discovery process [9]. giulia caron and her colleagues at the university of torino, italy have presented a review article about the block relevance approach which they developed to identify structure-property correlations that significantly enhanced the design of molecules with the right admet and dmpk properties [10]. the guest editor would like to thank all the contributors for their tremendous efforts made in presenting their work so lucidly. it is hoped that readers will be pleased to read these contributions and find them useful in their search for new drugs that will help to cure diseases and improve the quality of life of humans. references [1] s. soriano-meseguer, e. fuguet, a. port, martí rosés. estimation of skin permeation by liquid chromatography. admet and dmpk 6 (2018) 140-152. [2] x. subirats, l. redón, m. rosés. lipophilicity determination of acidic compounds: meekc as a reliable high-throughput methodology. admet and dmpk 6 (2018) 153-161. 73 [3] s. bunally, r.j. young, the role and impact of high throughput biomimetic measurements in drug discovery. admet and dmpk 6 (2018) 74-84. [4] g.h. goetz, m. shalaeva. leveraging chromatography based physicochemical properties for efficient drug design. admet and dmpk 6 (2018) 85-104. [5] k. valko, g. ivanova-berndt, p. beswick, m. kindey, d. ko, application of biomimetic hplc to estimate lipophilicity, protein and phospholipid binding of potential peptide therapeutics. admet and dmpk 6 (2018) 162-175. [6] k. valko, m. kindy, j. evans, d. ko, in vitro biomimetic hplc and in vivo characterisation of gm6, an endogenous regulator peptide drug candidate for amyotrophic lateral sclerosis. admet and dmpk 6 (2018) 176-189. [7] d. gussakovsky, h. neustaeter, v. spicer, o.v. krokhin. peptide retention time prediction for immobilized artificial membrane phosphatidylcholine stationary phase: method development and preliminary observations. admet and dmpk 6 (2018) 190-199. [8] c. stephen, a. el omri, l. ciesla. cellular membrane affinity chromatography (cmac) in drug discovery from complex natural matrices. admet and dmpk 6 (2018) doi: http://dx.doi.org/10.5599/admet.535. [9] f. tsopelas, c. stergiopoulos, a. tsantili-kakoulidou. immobilized artificial membrane chromatography: from medicinal chemistry to environmental sciences. admet and dmpk 6 (2018) to be published. [10] g. ermondi, g. caron. block relevance (br) analysis and polarity descriptors in property-based drug design. admet and dmpk 6 (2018) http://dx.doi.org/10.5599/admet.532. ©2018 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://dx.doi.org/10.5599/admet.535 http://dx.doi.org/10.5599/admet.532 http://creativecommons.org/licenses/by/3.0/ prediction of adme-tox properties and toxicological endpoints of triazole fungicides used for cereals protection doi: 10.5599/admet.668 161 admet & dmpk 7(3) (2019) 161-173; doi: http://dx.doi.org/10.5599/admet.668 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper prediction of adme-tox properties and toxicological endpoints of triazole fungicides used for cereals protection ionuț mădălin gridan 1 , alecu aurel ciorsac 2 and adriana isvoran 1,* 1 department of biology-chemistry and advanced environmental research laboratories, west university of timișoara, timișoara, romania 2 department of physical education and sport, university politehnica timișoara, timișoara, romania; *corresponding author: e-mail: adriana.isvoran@e-uvt.ro; tel.: +40-256-592-634; fax: +40-256-592-620 received: february 25, 2019; revised: may 02, 2019; published: may 21, 2019 abstract within this study we have considered 9 triazole fungicides that are approved to be used in european union for protecting cereals: cyproconazole, epoxiconazole, flutriafol, metconazole, paclobutrazole, tebuconazole, tetraconazole, triadimenol and triticonazole. we have summarized the few available data that support their effects on humans and used various computational tools to obtain a widely view concerning their possible harmful effects on humans. the results of our predictive study reflect that all triazole fungicides considered in this study reveal good oral bioavailability, are envisaged as being able to penetrate the blood brain barrier and to interact with p-glycoprotein and with hepatic cytochromes. the predictions concerning the toxicological endpoints for the investigated triazole fungicides reveal that they. reflect potential of skin sensitization, of blockage of the herg k+ channels and of endocrine disruption, that they have not mutagenic potential and their carcinogenic potential is not clear. epoxiconazole and triadimenol are predicted to have the highest potentials of producing numerous harmful effects on humans and their use should be avoided or limited. keywords oral bioavailability; endocrine disruption; carcinogenicity; mutagenicity; cardiotoxicity introduction pesticides are a broad class of chemical compounds that are deliberately used on a planetary basis since several decades in the control of pests and vectors of diseases. in spite of the benefits of their use, their continued application and their release into various ecosystems has become a matter of concern for both humans and environment [1]. the general population is exposed to low dose of pesticides repeated over time through the food, domestic use and environment [2]. a significant risk of exposure to pesticides is registered by the people engaged in agriculture [3] and for those people living close to a workplace that uses pesticides [4]. literature data suggest the relationship between pesticide exposure and several human pathologies: cancer, endocrine disruption, diabetes, neurologic and reproductive disorders [5,6]. there are more than 1000 active ingredients on various types of pesticides used worldwide [7]. furthermore, the formulation of new pesticides is growing due to the appearance of resistant pests, growing global population and regulation of pesticides. the use of some pesticides such as carbamates and http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:name@address.domain gridan, ciorsac, isvoran admet & dmpk 7(3) (2019) 161-173 162 organophosphates decreased since 2007, but the use of triazoles have significantly increased after 2007 [8]. triazole agents are heterocyclic compounds having a five-membered ring of two carbon and three nitrogen atoms that are widely used as antifungal agents in both medicine and in agriculture [9]. they represent around a third of the fungicides used for the protection of crop yields [10] and an important fungicide category used in the prevention of cereals diseases [11]. the residues of triazole agricultural fungicides have been frequently detected in numerous human and environmental media and there are reported toxicity data that have led them as compounds with health concern [12]. within this study we focus our attention on the triazole fungicides that are approved to be used in european union for protecting cereals [13]: cyproconazole, epoxiconazole, flutriafol, metconazole, paclobutrazole, tebuconazole, tetraconazole, triadimenol and triticonazole. the structural formulas of these compounds are illustrated in the figure 1. the routes of exposure for these chemicals usually are inhalation and dermal contact, but the oral route can be also considered as they could be found in food [4,6]. oh n n n cl o f n n n cl n n n f f oh n n n ch 3 ch 3 oh cl n n n cl oh ch 3 ch 3 ch 3 n n n cl oh n n n cl o f f f f cl n n n cl o o n n n cl oh cyproconazole 2-(4-chlorophenyl)-3-cyclopropyl-1(1,2,4-triazol-1-yl)butan-2-ol epoxiconazole 1-[[3-(2-chlorophenyl)-2-(4fluorophenyl)oxiran-2-yl]methyl]1,2,4-triazole flutriafol 1-(2-fluorophenyl)-1-(4fluorophenyl)-2-(1,2,4-triazol-1 yl)ethanol metconazole 5-[(4-chlorophenyl)methyl]-2,2dimethyl-1-(1,2,4-triazol-1ylmethyl)cyclopentan-1-ol paclobutrazole 1-(4-chlorophenyl)-4,4-dimethyl-2(1,2,4-triazol-1-yl)pentan-3-ol tebuconazole 1-(4-chlorophenyl)-4,4-dimethyl-3(1,2,4-triazol-1-ylmethyl)pentan-3-ol tetraconazole 1-[2-(2,4-dichlorophenyl)-3-(1,1,2,2tetrafluoroethoxy)propyl]-1,2,4-triazole triadimenol 1-(4-chlorophenyl)-3,3-dimethyl1-(1,2,4-triazol-1-yl)butan-2-ol triticonazole (5e)-5-[(4chlorophenyl)methylidene-2,2dimethyl-1-(1,2,4-triazol-1ylmethyl)cyclopentan-1-ol figure 1. structural formulas of triazole fungicides considered in the present study, their common and iupac names regulatory agencies establish standard approaches to risk assessment of chemicals. taking into account the ethics principles of conducting large-scale, long-term and highly invasive toxicity tests in humans, the risk assessment tests are usually based on toxicity studies conducted in animals. there also are important admet & dmpk 7(3) (2019) 161-173 admet of triazole fungicides doi: 10.5599/admet.668 163 data resulting from accidental human exposures to chemicals (occupational exposure, industrial accidents, unintentional environmental releases) that are considered when setting regulatory standards to protect public health. both scientific community and large society is divided in supporters and critics, and there are arguments for and against the intentional human dosing studies [14]. transfer of data obtained through animal tests to humans usually depend on the ability of measuring the same endpoints in animals and data obtained from studies on humans subjects are considered more relevant for assessing toxicity than data obtained from tests on animals. considering the costs and the ethical concerns on using both humans and animals for testing purposes, the role of computational approaches in hazard assessment become recognized. the amount and diversity of data obtained through experimental toxicity studies allowed the building of truthful computational models for toxicology assessment. it also conducted to the development of various computational tools that can be used for hazard assessment. they are recognized by the organization of economic and cooperation development (oecd) [15] and european food safety association [16] and are constantly used nowadays in assessing the toxicological effects of various chemicals on humans [17-21]. besides the large applicability of these tools for human hazard assessment, there are some limitations mostly related to the predictability of training models and to the fact that they do not take into account the dose. we use some of these computational tools in our study and taking into account that they are meant to predict the effects on humans, for validation purposes we have only collected data obtained from human exposure on triazole fungicides. specific literature data contains some information concerning the effects of these triazole fungicides on the human health. we have collected data from the specific literature and toxicity databases: toxnet/hazardous substances data bank (hsdb) [https://toxnet.nlm.nih.gov], occupational safety and health administration (osha) [https://www.osha.gov], national institute of occupational safety and health (niosh) / pocket guide to chemical hazards [http://www.aresok.org/npg/nioshdbs/npg/default.html], pan pesticide database (http://www.pesticideinfo.org), pubchem (https://pubchem.ncbi.nlm.nih.gov), and pesticides properties database (ppdb, https://sitem.herts.ac.uk/aeru/ppdb/en) available on january-february 2019. available information is presented in table 1. niosh and osha databases do not contain data concerning the toxic effects of investigated triazole fungicides. also, there is no information concerning their cardiotoxicity. besides the information presented in table 1, there are other few known effects of triazole fungicides: flutriafol is known to produce anaemia (ppdb), paclobutrazole is considered harmful if inhaled or swallowed (ppdb), tebuconazole is considered harmful if swallowed (pubchem) and triadimenol may cause nausea, headache, sneezing and vomiting (toxnet). data in table 1 illustrate that toxicological data from human exposure to pesticides are limited and sometimes controversial. the aim of this study is to use various computational algorithms to predict the adme-tox profiles and toxicological endpoints of the triazole fungicides used for cereals protection in european union. method specific literature is abundant in computational tools available for predicting admet profiles and biological effects of chemicals (for more information concerning available tools, please visit http://www.vls3d.com/index.php/links/chemoinformatics/admet/admet-and-physchem-predictions-andrelated-tools). we have selected for our study a few computational tools that are free accessible online, are continuously updated, are robust and their accuracy of predictions is higher than 70 %. https://toxnet.nlm.nih.gov/ https://www.osha.gov/ http://www.aresok.org/npg/nioshdbs/npg/default.html http://www.aresok.org/npg/nioshdbs/npg/default.html http://www.pesticideinfo.org/ https://pubchem.ncbi.nlm.nih.gov/ https://pubchem.ncbi.nlm.nih.gov/ http://www.vls3d.com/index.php/links/chemoinformatics/admet/admet-and-physchem-predictions-and-related-tools http://www.vls3d.com/index.php/links/chemoinformatics/admet/admet-and-physchem-predictions-and-related-tools gridan, ciorsac, isvoran admet & dmpk 7(3) (2019) 161-173 164 table 1. known human health effects of triazole fungicides used for cereals protection fafdrugs (now arrived to version 4) is an online tool that can be used for the computational prediction of the admet profile based on filtering rules that take into account some physicochemical descriptors (molecular weight, polar surface area, log p, number of hydrogen bonds donors and acceptors, or number of rigid or rotatable bonds, etc.) [22]. it requires as input the structural data file (sdf) of the investigated compound and outputs a summary result page and a detailed result page for every analysed compound. the summary result page contains a brief statistical summary of the filtering process with graphical representations of the distributions of numerous properties, illustrating all computed values. the detailed triazole fungicide skin/eye irritations respiratory tract irritant carcinogenicity /mutagenicity endocrine disruption potential reproductive toxicity hepatotoxicity neuro toxicity cyproconazole mild eye irritant (toxnet), skin and eye irritant (ppdb) yes (ppdb) probable carcinogen (toxnet, pan) /no data found. inhibition of aromatase activity, decrease of estrogens production (ppdb) no data found. possible liver toxicant (ppdb) no (ppdb) epoxiconazole no (ppdb) no (ppdb) probable carcinogen (ppdb, pubchem, pan)/ no data found. inhibition of aromatase activity, decrease of estrogen production (ppdb), suspect to produce endocrine disruption (pan) no data found. liver toxicant (ppdb) no (ppdb) flutriafol no (ppdb) yes (ppdb) no (ppdb, pan)/ no data found. weak estrogen inhibition (ppdb), suspect to produce endocrine disruption (pan) no data found. possible liver toxicant (ppdb) no (ppdb) metconazole no (ppdb) yes (pubchem) yes (ppdb) no (ppdb, pan)/ no data found. no data found. suspect of damaging fertility or the unborn child. (pubchem) possible liver toxicant (ppdb) no (ppdb) paclobutrazole yes (ppdb, pubchem) no data found. no (ppdb)/ no data found. no data found. suspect of damaging fertility or the unborn child. (pubchem) no data found. no (ppdb) tebuconazole eye irritant (ppdb) no (ppdb) probable carcinogen (toxnet, pan)/ no (ppdb) no (ppdb), suspect to produce endocrine disruption (pan) yes (ppdb, pubchem) no data found. no (ppdb) tetraconazole eye irritant (toxnet), no effects (ppdb) no (toxnet, ppdb) probable carcinogen (toxnet, ppdb, pan)/ no (ppdb) no (ppdb) no data found. liver toxicant (ppdb) no (ppdb) triadimenol yes (toxnet, ppdb) yes (toxnet, ppdb) probable carcinogen (ppdb, pan)/ no data found. yes (toxnet), estrogenic effect (ppdb), suspect to produce endocrine disruption (pan) yes (ppdb, pubchem) liver toxicant (ppdb) no data found. triticonazole no (ppdb) no (ppdb) probable carcinogen (toxnet), noncarcinogen (ppdb)/ no data found. no data found. no data found. no data found. no data found. admet & dmpk 7(3) (2019) 161-173 admet of triazole fungicides doi: 10.5599/admet.668 165 result page contains a list of all encountered problems (if it is the case), radar plots illustrating how the compound’s properties fit into the defined physicochemical filter, a principal component analysis mapping the analysed compound into the oral chemical space of drugs, the oral bioavailability assessment considering lipinski, veber, egan and bayer rules and a drug safety profiling based on the gsk 4/400 rule, pfizer 3/75 rule, estimation of phospholipidosis inducing and the lilly medchem rules with a minimum of 70 % of accuracy. when computing the lilly medchem rules [23], the regular demerit level has been applied. swissadme [24] computational facility has been used to predict pharmacokinetics of investigated triazole fungicides. for every investigated compound, swissadme outputs predictions concerning the: passive human gastrointestinal absorption (gi), blood-brain barrier (bbb) permeation, skin penetration coefficient, substrate or non-substrate of the permeability glycoprotein (p-gp), interaction of molecules with five major isoforms of the human cytochromes p450 (cyp1a2, cyp2c19, cyp2c9, cyp2d6, cyp3a4) known to be involved in the metabolism of numerous endogenous and exogenous compounds [24]. the statistical performance of the classification models used by swissadme for predicting pharmacokinetics profile is between 72 % and 94 %. as dermal exposure to triazole fungicides is not negligible, we have used predskin [25] computational tool to assess the skin sensitization potential of investigated compounds. this application is based on binary qsar models of skin sensitization potential using human and murine local lymph node assay data (llna) (performance ranging between 70 and 84 %) and a multiclass skin sensitization potential model based on llna data (performance being about 73 %). the application also outputs, for every compound, a probability map illustrating the predicted contribution to skin sensitization potential of chemical fragments [25]. pred-herg is a web tool based on quantitative structure-activity relationship (qsar) models built on the curated dataset of 5,984 compounds and used to predict the herg k+ channel blockage. three outcomes are available when using pred-herg: predictions using a binary model (accuracy 80 %), predictions using a multi-class model (accuracy 70 %) and a probability map of atomic contribution to herg k(+) channels blockage [26]. carcinopred-el (carcinogenicity prediction using ensemble learning methods) is a web server allowing the classification of chemical compounds as carcinogens or non-carcinogens starting from their twodimensional structures [27]. this web server incorporates three ensemble learning models (ensemble xgboost, ensemble svm and ensemble rf) for predicting the carcinogenicity of chemicals. ensemble xgboost model gives the highest accuracy of predictions (70.1 %) [27]. mutagenicity of considered triazole fungicides has been predicted using ames test implemented under the open source toxtree software [28]. this is an open source application that estimates mutagenicity by applying a decision tree approach based on benigni and bosa rules [29] with an accuracy of about 79 %. endocrine disruption potential of investigated fungicides has been evaluated using endocrine disruptome computational tool [30]. it uses the molecular docking approach based on autodock vina algorithm to predict the interactions between the investigated chemicals and human nuclear receptors: androgen receptor, estrogen receptors α and β, glucocorticoid receptor, liver x receptors α and β, peroxisome proliferator activated receptors α, β/δ and γ, retinoid x receptor α and thyroid receptors α and β [24]. in the results section, three thresholds, calculated from the docking score and the validation experiments and expressed as sensitivity (se) parameter, were set for every structure allowing a division gridan, ciorsac, isvoran admet & dmpk 7(3) (2019) 161-173 166 into 4 probability binding classes: compounds with high probability of binding (se<0.25), compounds with intermediate probability of binding (0.25<se<0.50), compounds with moderate probability of binding, (0.50<se< 0.75) and compounds with low probability of binding (se>0.75) [30]. results and discussion admet profiles of investigated fungicides have been obtained using fafdrugs4 and respectively swissadme computational facilities and are illustrated in tables 2 and 3. data presented in tables 2 and 3 illustrate that all investigated triazole fungicides have good oral bioavailability, can be easily absorbed in the gastrointestinal tract and, consequently, are able to reach the systemic blood circulation and to produce various biologic effects. every investigated compound is predicted to have at least a mean toxicity when applying pfizer rule, the toxicity being related to their content in halogen atoms and, in the case of epoxiconazole, the epoxide group also reflect toxicity. epoxiconazole and triadimenol also do not pass lilly medchem rules meaning that these two pesticides may illustrate activities that damage proteins [23]. none of investigated compounds is predicted to induce phospholipidoses (data not shown). table 2. prediction of the admet profiles obtained using the fafdrugs4 computational tool for the triazole fungicides considered in this study: green boxes correspond to rules that are respected, orange boxes correspond to rules that are partially violated and red grey boxes correspond to rules that are not respected. triazole fungicide oral bioavailability drug safety profiling lipinski’s rule veber’s rule eagan’s rule bayer rule gsk 4/400 rule pfeizer 3/75 rule lilly medchem rules cyproconazole epoxiconazole flutriafol metconazole paclobutrazole tebuconazole triticonazole triadimenol tetraconazole table 3. prediction of the pharmacokinetics profiles of the triazole fungicides using swissadme computational tool: gi gastrointestinal absorption, bbbblood brain barrier permeant, p-gp – p-glycoprotein, cyp – human cytochrome, log kp skin penetration coefficient in logarithmic scale. all investigated fungicides are predicted as being able to penetrate the blood brain barrier and it illustrates their potential to affect the central nervous system, and/or the transport of nutrients, drugs and waste products into and out of the brain. available information illustrates that most of investigated triazoled fungicides do not produce human neurotoxicity (see table 1), but there are published data exemplifying neurotoxic effects of these compounds on rats [31-32]. epoxyconazole and triticonazole are active substance gi absorption bbb permeant p-gp substrate cyp1a2 inhibitor cyp2c19 inhibitor cyp2c9 inhibitor cyp2d6 inhibitor cyp3a4 inhibitor log kp cyproconazole high yes no no yes no no no -6.02 epoxiconazole high yes yes no yes yes yes no -5.87 flutriafol high yes no no yes no yes no -6.50 metconazole high yes no no yes no yes no -5.52 paclobutrazole high yes no no yes no no no -5.82 tebuconazole high yes no yes yes no yes no -5.55 triticonazole high yes yes no yes yes no no -5.87 triadimenol high yes no no yes no no no -5.92 tetraconazole high yes no yes yes yes no no -6.04 admet & dmpk 7(3) (2019) 161-173 admet of triazole fungicides doi: 10.5599/admet.668 167 predicted as p-gp substrates and it illustrates their possible active efflux both from the gastrointestinal tract to the lumen and from the brain, being known that one of the roles of p-gp is to protect the central nervous system from xenobiotics. the predictions are in good agreement with literature data revealing that the triazoles are substrates and/or inhibitors of transport proteins in the atp-binding cassette transporter protein family [33]. furthers studies are necessary for assessing the risk of neurotoxicity of trizole fungicides on humans. all investigated fungicides are predicted to be able to inhibit cyp2c19 and some of them are also able to inhibit cyp1a2, cyp2c9 and/or cyp2d6. these predictions are in good agreement with published data concerning the interactions of triazole antifungal human drugs with hepatic cyps. literature data reveal that triazole antifungal agents as human drugs are substrates and inhibitors of various cyps [34], including cyp3a4, cyp2c19 and cyp2c9 that are involved in catalysing the triazole biotransformation [35]. it is possible that investigated triazole fungicides to be metabolized by cyps. in this case, the polymorphisms within these enzymes may conduct in their accumulation into the human organism. the inhibition of human cyps involved in the metabolism of many endogenous compounds and of xenobiotics by the investigated triazole fungicides may conduct to pharmacokinetics-related drug-xenobiotics interactions leading to toxic and/or adverse effects. it underlines the importance of predictions of which isoforms of cyps are affected by triazole fungicides. the predicted values of the skin penetration coefficients in logarithmic values (log kp) of investigated triazole fungicides were compared with those of diclofenac (a compound with a good skin penetration, log kp=-4.96), and of oubain (a compound that is not able to penetrate skin, log kp = -11.07) [22]. the more negative value of computed log kp means lower skin permeability for the investigated compound. this comparison illustrates the medium ability of triazole fungicides to penetrate skin and it is in good correlation with data proving that some of the investigated fungicides produce skin irritations (see table 1). we have also evaluated the skin sensitization potential of investigated triazole fungicides using predskin computational tool and the results are illustrated in table 4. the probability map illustrating the predicted contribution to skin sensitization potential of chemical fragments of cyproconazole when using the binary prediction based on human skin sensitization model is shown in figure 2 a. figure 2. the redicted probability map of the predicted contribution of chemical fragments of cyproconazole to: (a) skin sensitization potential; (b) blockage of the herg k+ channels. green color corresponds to atoms illustrating an increase in skin sensitization (a) or towards blockage of herg (b), pink color corresponds to atoms emphasizing a decrease in skin sensitization (a) or in herg blockage (b) and gray lines delimit the region of split between the positive and negative contributors toward skin sensitization (a) or herg blockage (b). more continuous contour lines mean an increased pozitive contribution and more dashed contour lines mean an increased negative contribution of an atom to skin sensitization (a) or herg blockage (b) respectively. gridan, ciorsac, isvoran admet & dmpk 7(3) (2019) 161-173 168 this picture illustrates the positive contribution to skin sensitization of the chlorophenyl and cyclopropyl groups, but not of triazole ring. table 4 illustrates that all investigated triazole fungicides are predicted as skin sensitizers. available information concerning the ability of some of the investigated fungicides to produce skin irritations corresponds to an acute exposure. our predictions reveal that exposure (especially when prolonged) to these compounds may conduct to skin sensitization, a complex immunological disease that have an important impact on quality of life and on working ability. table 4. prediction of skin sensitization potential of the considered triazole fungicides: llna murine local lymph node assay, drpa direct peptide reactivity assay, h-clat human cell line activation test. green boxes correspond to predictions of non-sensitizer potential, orange boxes to predictions of moderate skin sensitizer potential and red boxes illustrate predictions of skin sensitizer potential. numbers in parenthesis correspond to the accuracy of every prediction. table 5 contains predictions of the following toxicological endpoints of the investigated triazole fungicides: cardiotoxicity (blocking potential of h-erg k+ channels) obtained using pred-herg tool, mutagenicity obtained using toxtree software and carcinogenicity using both toxtree and carcinopred-el computational facilities. the two models used by pred-herg tool give contradictory predictions. for all the investigated tiazole fungicides, the use of multiclass model do not predict the blockage of the herg k+ channels potential, but the use of the binary model reveal their potential to block the herg k+ channels. taking into account that the binary model has a higher accuracy of predictions, we consider that investigated triazole fungicides reflect potential to block the herg k+ channels. figure 2b shows the probability map of the predicted contribution of atoms and/or fragments of cyproconazole toward blockage of the herg k+ channels. it illustrates that triazole fragment and the halogen atom reveal potential of the blockage of the herg k+ channels. besides the triazole ring, there is at list one fragment containing a halogen atom attached to the phenyl group in every of investigated fungicides (see figure 1) and it reflects the herg k+ channel-blocking potential of these pesticides. our findings are in good correlation with published data revealing that compounds containing a polar group (such as a halogen atom) attached to the phenyl ring at one end of the molecule reveal herg k+ channel-blocking potential [36,37]. furthermore, the class of antifungal triazoles used as human drugs [38] and other compounds fungicide human skin sensitization llna dpra h-clat) keratinosens binary prediction binary prediction multiclass prediction binary prediction binary prediction binary prediction cyproconazole sensitizer (70%) non-sensitizer (50%) moderate (70%) sensitizer (90%) sensitizer (80%) sensitizer (80%) epoxiconazole sensitizer (80%) non-sensitizer (60%) moderate (60%) sensitizer (90%) sensitizer (80%) sensitizer (80%) flutriafol sensitizer (80%) non-sensitizer (50%) moderate (80%) sensitizer (90%) sensitizer (80%) sensitizer (80%) metconazole sensitizer (80%) non-sensitizer (50%) moderate (80%) sensitizer (90%) sensitizer (80%) sensitizer (80%) paclobutrazol sensitizer (60%) non-sensitizer (50%) moderate (80%) sensitizer (90%) sensitizer (80%) sensitizer (80%) tebuconazole sensitizer (70%) sensitizer (50%) moderate (90%) sensitizer (90%) sensitizer (80%) sensitizer (80%) tetraconazole sensitizer (60%) non-sensitizer (50%) moderate (80%) sensitizer (90%) sensitizer (80%) sensitizer (90%) triticonazole sensitizer (60%) non-sensitizer (50%) moderate (70%) sensitizer (90%) sensitizer (80%) sensitizer (90%) triadimenol nonsensitizer (60%) non-sensitizer (60%) moderate (70%) sensitizer (90%) sensitizer (80%) sensitizer (80%) admet & dmpk 7(3) (2019) 161-173 admet of triazole fungicides doi: 10.5599/admet.668 169 containing the triazole ring [39] are recognized as being able to cause cardiac dysrhythmias by blocking herg k+ channels in the heart. table 5. predictions concerning the following toxicological endpoints: cardiotoxicity, carcinogenicity and mutagenicity of investigated triazole fungicides. the outcomes of carcinopred-el computational tool reveal carcinogenic potential for flutriafol, paclobutrazole, triadimenol and triticonazole when using xgboost ensemble learning model, this model giving the highest accuracy of predictions. only flutriafol is predicted as carcinogen by all models considered by carcinopred-el tool. toxtree outcomes are quite different by comparison with those of cracinopred-el and reveal both genotoxic and non-genotoxic carcinogenicity for epoxiconazole and, excepting flutriafol, non-genotoxic carcinogenicity for all the other investigated triazole fungicides. literature data show that most of the investigated triazole fungicides are considered as probable human carcinogen. usually, the predictions obtained with the two computational tools are not in agreement each other or with known information concerning the carcinogenic effects of triazole fungicides. it underlines the need of further investigations concerning the carcinogenic potential of triazole fungicides on humans. predictions obtained using endocrine disruptome computational tool reflected that investigated triazole fungicides are able to influence the activity of some of the human nuclear receptors, as presented in table 6. red boxes correspond to the high probability of binding, orange boxes to intermediate probability of binding, yellow boxes to moderate probability of binding and green boxes correspond to low probability of binding of triazole fungicides to these receptors. all investigated triazole fungicides are predicted as revealing antagonistic effects on the androgen receptor. epoxiconazole is the compound predicted as reflecting the highest endocrine disruption potential as it interacts with numerous human nuclear receptors. the predictions of endocrine disruptome tool reveal that the most affected nuclear receptors by the analysed triazole fungicides are: ar, erα, erβ, gr, trα and trβ. this result is in good agreement with published data revealing the ability of triazole fungicides to interact with nuclear receptors, especially with the androgen and estrogen receptors [12, 40-33]. fungicide h_erg channel blockage predictions carcinopred-el predictions for carcinogenicity toxtree predictions binary model multiclass model xgboost model rf model svm model genotoxic carcinogenicity non genotoxic carcinogenicity ames mutagenicity cyproconazole blocker (70 %) non-blocker (70 %) no no no no yes no epoxiconazole blocker (70 %) non-blocker (60 %) no no no yes yes no flutriafol blocker (60 %) non-blocker (70 %) yes yes yes no no no metconazole blocker (80 %) non-blocker (70 %) no no no no yes no paclobutrazole blocker (70 %) non-blocker (60 %) yes no no no yes no tebuconazole blocker (70 %) non-blocker (60 %) no no no no yes no tetraconazole blocker (70 %) non-blocker (70 %) no no no no no no triadimenol blocker (60 %) non-blocker (70 %) yes no no no yes no triticonazole blocker (70 %) non-blocker (60 %) yes no no no yes no gridan, ciorsac, isvoran admet & dmpk 7(3) (2019) 161-173 170 table 6. prediction of the endocrine disruption potential of investigated triazole fungicides: red boxes correspond to the high probability of binding, orange boxes correspond to intermediate probability of binding, yellow boxes correspond to moderate probability of binding and green boxes correspond to low probability of binding of triazole fungicides to these receptors. fungicide ar ar an erα erα an erβ erβ an gr gr an lxrα lxr β ppar α ppar β ppar γ rx rα trα trβ cyproconazole -6.5 -7.4 -7.8 -7.9 -7.6 -7.5 -7.3 -6.9 -8.4 -8.2 -6.7 -7.0 -6.7 -7.6 -7.5 -7.7 flutriafol -6.3 -8.3 -8.7 -8.7 -8.7 -8.6 -7.9 -7.8 -9.1 -8.8 -7.9 -7.7 -7.4 -8.3 -7.9 -8.8 tebuconazole -7.3 -7.5 -7.8 -7.1 -7.5 -7.2 -7.2 -6.6 -8.6 -8.3 -7.1 -7.2 -7.7 -8.5 -6.9 -7.8 triticonazole -5.3 -7.3 -8.5 -7.8 -8.1 -7.7 -8.3 -7.2 -8.7 -9.3 -7.1 -7.3 -8.4 -9.9 -6.2 -8.0 epoxyconazole -8.3 -9.2 -9.2 -9.3 -8.8 -8.9 -8.5 -7.9 -10.0 -9.4 -8.4 -8.0 -8.0 -9.0 -8.0 -8.6 tetraconazole -7.2 -7.9 -8.3 -8.1 -8.2 -7.7 -7.5 -7.2 -8.9 -8.6 -7.5 -7.9 -7.8 -8.4 -7.9 -8.4 triadimenol -6.8 -7.7 -7.8 -7.6 -7.8 -7.4 -7.1 -6.8 -8.4 -8.1 -7.3 -7.1 -7.8 -7.9 -7.1 -8.3 prothioconazole -7.3 -7.6 -8.3 -8.0 -8.2 -7.4 -8.2 -7.0 -8.5 -8.4 -7.7 -7.3 -7.4 -8.0 -8.3 -7.9 metconazole -6.5 -7.4 -7.8 -8.1 -8.6 -8.2 -7.8 -7.1 -9.1 -9.5 -7.9 -7.8 -8.1 -9.0 -6.7 -8.0 paclobutrazol -6.9 -7.7 -7.9 -7.7 -7.6 -7.4 -7.3 -6.6 -8.2 -8.3 -7.1 -7.1 -8.0 -8.4 -6.8 -7.8 conclusions within this study we have considered 9 triazole fungicides commonly used for cereal crops protection. we have summarized the available data that support their effects on humans and we have also used various computational tools to predict their harmful effects. available data are incomplete and the outcomes of this predictive study are meant to obtain a widely view concerning the harmful effects of these pesticides. the predictions that we have obtained using various tools usually are in good agreement with each other and with available data, this accord increasing their relevance. triazole fungicides considered in this study are predicted as revealing high oral bioavailability, being able to penetrate the blood brain barrier, and to interact with p-glycoprotein and with hepatic cyps. the predicted toxicological endpoints of studied triazole fungicides are: (i) skin sensitising potential; (ii) blockage of the herg k+ channels and (iii) endocrine disruption potential, the most affected human nuclear receptor being the androgen receptor, estrogen receptors α and β, glucocorticoid receptor and thyroid receptors α and β. triazole fungicides considered in this study are predicted to have not mutagenic potential and their carcinogenic potential is not clear. of the investigated fungicides, our predictions show that epoxiconazole and triadimenol seem to have the highest potentials of producing harmful effects on humans and their use should be avoid or limited. as the use of triazole fungicides is increasing, the outcomes of our study are important for both population and professional exposure. conflict of interest: the authors declare no conflict of interest. admet & dmpk 7(3) (2019) 161-173 admet of triazole fungicides doi: 10.5599/admet.668 171 references [1] m.i. pinto, h.d. burrows, g. sontag, c. vale, j.p. noronha. priority pesticides in sediments of european coastal lagoons: a review. marine pollution bulletin 112 (2016) 6-16. (https://www.sciencedirect.com/science/article/pii/s0025326x16305069?via%3dihub) [2] n.i. rousis, e. zuccato, s. castiglioni. monitoring population exposure to pesticides based on liquid chromatography-tandem mass spectrometry measurement of their urinary metabolites in urban wastewater: a novel biomonitoring approach. science of the total environment 571 (2016) 13491357. (https://www.sciencedirect.com/science/article/pii/s0048969716314814) [3] v. yusa, m. millet, c. coscolla, o. pardo, m. roca. occurrence of biomarkers of pesticide exposure in non-invasive human specimens. chemosphere 139 (2015) 91-108. (https://www.sciencedirect.com/science/article/pii/s0045653515005639?via%3dihub) [4] c.a. damalas, i.g. eleftherohorinos. pesticide exposure, safety issues, and risk assessment indicators. international journal of environmental research and public health 8 (2011) 1402-1419. (https://www.ncbi.nlm.nih.gov/pmc/articles/pmc3108117/) [5] v. yusa, m. millet, c. coscolla, m. roca, analytical methods for human biomonitoring of pesticides. a review. analytica chimica acta 891 (2015) 15-31. (https://www.sciencedirect.com/science/article/pii/s0003267015007552?via%3dihub) [6] k.h. kim, e. kabir, s.a. jahan. exposure to pesticides and the associated human health effects. science of the total environment 575 (2017) 525-535. (https://www.sciencedirect.com/science/article/pii/s004896971631926x?via%3dihub) [7] s.c. machado, i. martins. risk assessment of occupational pesticide exposure: use of endpoints and surrogates. regulatory toxicology and pharmacology 98 (2018) 276-283. (https://www.sciencedirect.com/science/article/pii/s0273230018302198?via%3dihub) [8] w.j. zhang. global pesticide use: profile, trend, cost / benefit and more. proceedings of the international academy of ecology and environmental sciences 8 (2018) 1-27. (http://www.iaees.org/publications/journals/piaees/articles/2011-1(2)/global-pesticide-consumptionpollution.pdf) [9] e. giavini, e. menegola. are azole fungicides a teratogenic risk for human conceptus? toxicology letters 198 (2010) 106-111. (https://www.sciencedirect.com/science/article/pii/s0378427410015870?via%3dihub) [10] a. chowdhary, s. kathuria, h.s. randhawa, s.n. gaur, c.h. klaassen. isolation of multiple-triazoleresistant aspergillus fumigatus strains carrying the tr/l98h mutations in thecyp51a gene in india. journal of antimicrobial chemotherapy 67 (2012) 362–366. (https://academic.oup.com/jac/article/67/2/362/700010) [11] f.j. machado, f.m. santana, d. lau, e.m.d. ponte. quantitative review of the effects of triazole and benzimidazole fungicides on fusarium head blight and wheat yield in brazil. plant disease 101 (2017) 1633-1641. (https://apsjournals.apsnet.org/doi/10.1094/pdis-03-17-0340-re) [12] x. lv, l. pan, j. wang, l. lu, w. yan, y. zhu, y. xu, m. guo, s. zhuang. effects of triazole fungicides on androgenic disruption and cyp3a4 enzyme activity. environmental pollution 22 (2017) 504-512. (https://www.sciencedirect.com/science/article/pii/s0269749116308612?via%3dihub) [13] eu pesticides database european commission, http://ec.europa.eu/food/plant/pesticides/eupesticides-database/public/?event=homepage&language=en (accessed between december 2018 and february 2019). [14] national research council. 2004. intentional human dosing studies for epa regulatory purposes: scientific and ethical issues. washington, dc: the national academies press. https://www.nap.edu/read/10927/chapter/1 (accessed on 25 january 2019). [15] guidance document on the validation of (quantitative)structure-activity relationships [(q)sar] models. oecd 69 (2007) (https://read.oecd-ilibrary.org/environment/guidance-document-on-thevalidation-of-quantitative-structure-activity-relationship-q-sar-models_9789264085442-en#page1 – accessed on 18 avril 2019). https://www.sciencedirect.com/science/article/pii/s0025326x16305069?via%3dihub https://www.sciencedirect.com/science/article/pii/s0048969716314814 https://www.sciencedirect.com/science/article/pii/s0045653515005639?via%3dihub https://www.sciencedirect.com/science/article/pii/s0045653515005639?via%3dihub https://www.ncbi.nlm.nih.gov/pmc/articles/pmc3108117/ https://www.sciencedirect.com/science/article/pii/s0003267015007552?via%3dihub https://www.sciencedirect.com/science/article/pii/s0003267015007552?via%3dihub https://www.sciencedirect.com/science/article/pii/s004896971631926x?via%3dihub https://www.sciencedirect.com/science/article/pii/s004896971631926x?via%3dihub https://www.sciencedirect.com/science/article/pii/s0273230018302198?via%3dihub https://www.sciencedirect.com/science/article/pii/s0273230018302198?via%3dihub https://www.sciencedirect.com/science/article/pii/s0378427410015870?via%3dihub https://www.sciencedirect.com/science/article/pii/s0378427410015870?via%3dihub https://academic.oup.com/jac/article/67/2/362/700010 https://academic.oup.com/jac/article/67/2/362/700010 https://apsjournals.apsnet.org/doi/10.1094/pdis-03-17-0340-re https://www.sciencedirect.com/science/article/pii/s0269749116308612?via%3dihub http://ec.europa.eu/food/plant/pesticides/eu-pesticides-database/public/?event=homepage&language=en http://ec.europa.eu/food/plant/pesticides/eu-pesticides-database/public/?event=homepage&language=en https://www.nap.edu/read/10927/chapter/1 https://read.oecd-ilibrary.org/environment/guidance-document-on-the-validation-of-quantitative-structure-activity-relationship-q-sar-models_9789264085442-en#page1 https://read.oecd-ilibrary.org/environment/guidance-document-on-the-validation-of-quantitative-structure-activity-relationship-q-sar-models_9789264085442-en#page1 gridan, ciorsac, isvoran admet & dmpk 7(3) (2019) 161-173 172 [16] european food safety authority, modern methodologies and tools for human hazard assessment of chemicals, efsa journal 12 (2014) 3638. (https://efsa.onlinelibrary.wiley.com/doi/pdf/10.2903/j.efsa.2014.3638 – accessed on 18 avril 2019). [17] d. craciun, d. modra, a. isvoran. adme-tox profiles of some food additives and pesticides, aip conference proceedings 1694 (2015) 040007. (https://doi.org/10.1063/1.4937259) [18] a. isvoran, a. ciorsac, v. ostafe. adme-tox profiling of some low molecular weight water soluble chitosan derivatives, admet and dmpk 5 (2017) 192-200. (http://dx.doi.org/10.5599/admet.5.3.423) [19] m. roman, d.l. roman, v. ostafe, a. ciorsac. computational assessment of pharmacokinetics and biological effects of some anabolic and androgenic steroids. pharmaceutical research 35 (2018) 41. (doi: http://dx.doi.org/10.1007/s11095-018-2353-1) [20] m. roman, d.l. roman, v. ostafe, a. isvoran. computational assessment of biological effects of methyl-,ethyl-, propyland butyl-parabens. journal of bioinformatics, genomics, proteomics 3 (2018) 1029. [21] v.m. alves, e.n. muratov, a. zakharov, n.n. muratov, c.h. andrade, a. tropsha. chemical toxicity prediction for major classes of industrial chemicals:is it possible to develop universal models covering cosmetics, drugs, and pesticides?. food and chemical toxicology 112 (2018) 526-534. (https://www.ncbi.nlm.nih.gov/pmc/articles/pmc5638676/) [22] d. lagorce, l. bouslama, j. becot, m.a. miteva, b.o. villoutreix. faf-drugs4: free adme-tox filtering computations for chemical biology and early stages drug discovery. bioinformatics 33 (2017) 36583660. (https://academic.oup.com/bioinformatics/article/33/22/3658/4056067) [23] r.f. bruns, i.a. watson. rules for identifying potentially reactive or promiscuous compounds. journal of medicinal chemistry 55 (2012) 9763−9772. (https://pubs.acs.org/doi/10.1021/jm301008n) [24] a. daina, o. michielin, v. zoete. swissadme: a free web tool to evaluate pharmacokinetics, druglikeness and medicinal chemistry friendliness of small molecules. scientific reports 7 (2017) 42717. (https://www.ncbi.nlm.nih.gov/pmc/articles/pmc5335600/) [25] r.c. braga, v.m. alves, e.n. muratov, j. strickland, n. kleinstreuer, a. trospsha, c.h. andrade. predskin: a fast and reliable web application to assess skin sensitization effect of chemicals. journal of chemical information and modeling 57 (2017) 1013-1017. (https://pubs.acs.org/doi/10.1021/acs.jcim.7b00194) [26] r.c. braga, v.m. alves, m.f. silva, e. muratov, d. fourches, l.m. lião, a. tropsha, c.h. andrade. pred-herg: a novel web-accessible computational tool for predicting cardiac toxicity. molecular informatics 34 (2015) 698-701. (https://www.ncbi.nlm.nih.gov/pmc/articles/pmc5720373/) [27] l. zhang, h. ai, w. chen, z. yin, h. hu, j. zhu, j. zhao, q. zhao, h. liu. carcinopred-el: novel models for predicting the carcinogenicity of chemicals using molecular fingerprints and ensemble learning methods. scientific reports 7 (2017) 2118. (https://www.ncbi.nlm.nih.gov/pmc/articles/pmc5437031/) [28] g. patlewicz , n. jeliazkova , r.j. safford , a.p. worth, b. aleksiev. an evaluation of the implementation of the cramer classification scheme in the toxtree software, sar and qsar in environmental research 19 (2008) 495-524. (https://www.tandfonline.com/doi/abs/10.1080/10629360802083871) [29] r. benigni, c. bossa, t. netzeva, a. rodomonte, i. tsakovska. mechanistic qsar of aromatic amines: new models for discriminating between mutagens and nonmutagens, and validation of models for carcinogen. environmental and molecular mutagenesis 48 (2007) 754-771. (https://onlinelibrary.wiley.com/doi/abs/10.1002/em.20355) [30] k. kolsek, j. mavri, m. sollnerdolenc, s. gobec, s. turk. endocrine disruptome an open source prediction tool for assessing endocrine disruption potential through nuclear receptor binding. journal of chemical information and modeling 54 (2014) 1254-1267. (https://pubs.acs.org/doi/10.1021/ci400649p) https://efsa.onlinelibrary.wiley.com/doi/pdf/10.2903/j.efsa.2014.3638 https://efsa.onlinelibrary.wiley.com/doi/pdf/10.2903/j.efsa.2014.3638 https://doi.org/10.1063/1.4937259 http://dx.doi.org/10.5599/admet.5.3.423 http://dx.doi.org/10.5599/admet.5.3.423 http://dx.doi.org/10.1007/s11095-018-2353-1 https://www.ncbi.nlm.nih.gov/pmc/articles/pmc5638676/ https://academic.oup.com/bioinformatics/article/33/22/3658/4056067 https://pubs.acs.org/doi/10.1021/jm301008n https://pubs.acs.org/doi/10.1021/jm301008n https://www.ncbi.nlm.nih.gov/pmc/articles/pmc5335600/ https://pubs.acs.org/doi/10.1021/acs.jcim.7b00194 https://pubs.acs.org/doi/10.1021/acs.jcim.7b00194 https://www.ncbi.nlm.nih.gov/pmc/articles/pmc5720373/ https://www.ncbi.nlm.nih.gov/pmc/articles/pmc5437031/ https://www.ncbi.nlm.nih.gov/pmc/articles/pmc5437031/ https://www.tandfonline.com/doi/abs/10.1080/10629360802083871 https://www.tandfonline.com/doi/abs/10.1080/10629360802083871 https://onlinelibrary.wiley.com/doi/abs/10.1002/em.20355 https://pubs.acs.org/doi/10.1021/ci400649p https://pubs.acs.org/doi/10.1021/ci400649p admet & dmpk 7(3) (2019) 161-173 admet of triazole fungicides doi: 10.5599/admet.668 173 [31] x. shen, f. chen, l. chen, y. su, p. huang, r.-s. ge. effects of fungicides on rat’s neurosteroid synthetic enzymes, biomed research international 2017 (2017) 5829756. [32] n.m. filipov, d.a. lawrence. developmental toxicity of a triazole fungicide: consideration of interorgan communication. toxicological sciences 62 (2001) 185–186. (https://academic.oup.com/toxsci/article/62/2/185/1663615) [33] m. chaâbane, i. ghorbel, a. elwej, h. mnif, t. boudawara, s. ellouze chaâbouni, n. zeghal, n. soudani. penconazole alters redox status, cholinergic function, and membrane-bound atpases in the cerebrum and cerebellum of adult rats. human & experimental toxicology 36 (2016) 854-866. (https://journals.sagepub.com/doi/abs/10.1177/0960327116672911) [34] v.j. lempers, j.j. van den heuvel, f.g. russel, r.e. aarnoutse, d.m. burger, r.j. brüggemann, j.b. koenderink. inhibitory potential of antifungal drugs on atp-binding cassette transporters pglycoprotein, mrp1 to mrp5, bcrp, and bsep. antimicrobial agents and chemotherapy 60 (2016) 3372–3379. (https://www.ncbi.nlm.nih.gov/pmc/articles/pmc4879353/) [35] j.r. amsden, p.o. gubbins. pharmacogenomics of triazole antifungal agents: implications for safety, tolerability and efficacy. expert opinion on drug metabolism&toxicology 13 (2017) 1135-1146. (https://www.tandfonline.com/doi/abs/10.1080/17425255.2017.1391213?src=recsys&journalcod e=iemt20) [36] p.o. gubbins. triazole antifungal agents drug--drug interactions involving hepatic cytochrome p450. expert opinion on drug metabolism&toxicology 7 (2011) 1411–1429. (https://www.tandfonline.com/doi/abs/10.1517/17425255.2011.627854?journalcode=iemt20) [37] a. cavalli, e. poluzzi, f. de ponti, m. recanatini. toward a pharmacophore for drugs inducing the long qt syndrome: insights from a comfa study of herg k_ channel blockers. journal of medicinal chemistry 45 (2002) 3844–385. (https://pubs.acs.org/doi/abs/10.1021/jm0208875) [38] m. perry, m.j. de groot, r. helliwell, d. leishman, m. tristani-firouzi, m.c. sanguinetti, j. mitcheson. structural determinants of herg channel block by clofilium and ibutilide. molecular pharmacology 66 (2004) 240–249. (http://molpharm.aspetjournals.org/content/66/2/240.long) [39] k. arunrungvichian, p. khongkaew, s. panyasang, j. sarasamkan. preliminary in vitro permeability, cytotoxicity and cardiotoxicity evaluation of triazole-quinuclidine t6. pharmaceutical sciences asia 46 (2019) 156-162. (https://www.pharmacy.mahidol.ac.th/journal/_files/2019-46-3_156-162.pdf) [40] d. schrey, t.j. walsh, a.h. groll. antifungal drugs. side effects of drugs annual 33 (2011) 541-566. (https://www.sciencedirect.com/science/article/pii/b9780444537416000271?via%3dihub) [41] m.b. kjaerstad, c. taxvig, c. nellemann, a.m. vinggaard, h.r. andersen. endocrine disrupting effects in vitro of conazole antifungals used as pesticides and pharmaceuticals. reproductive toxicology 30 (2010) 573–582. (https://www.sciencedirect.com/science/article/pii/s0890623810002625?via%3dihub) [42] t. mankame, r. hokanson, r. chowdhary, d. busbee. altered gene expression in human cells induced by the agricultural chemicals enable. toxicology and industrial health 20 (2004) 89–102. (https://journals.sagepub.com/doi/abs/10.1191/0748233704th198oa) [43] s. rieke, s. koehn, k. hirsch-ernst, r. pfeil, c. kneuer, p. marx-stoelting. combination effects of (tri)azole fungicides on hormone production and xenobiotic metabolism in a human placental cell line. international journal of environmental research and public health 11 (2014) 9660-9679. (https://www.ncbi.nlm.nih.gov/pmc/articles/pmc4199042/pdf/ijerph-11-09660.pdf) ©2019 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) https://academic.oup.com/toxsci/article/62/2/185/1663615 https://academic.oup.com/toxsci/article/62/2/185/1663615 https://journals.sagepub.com/doi/abs/10.1177/0960327116672911 https://www.ncbi.nlm.nih.gov/pmc/articles/pmc4879353/ https://www.tandfonline.com/doi/abs/10.1080/17425255.2017.1391213?src=recsys&journalcode=iemt20 https://www.tandfonline.com/doi/abs/10.1080/17425255.2017.1391213?src=recsys&journalcode=iemt20 https://www.tandfonline.com/doi/abs/10.1517/17425255.2011.627854?journalcode=iemt20 https://pubs.acs.org/doi/abs/10.1021/jm0208875 http://molpharm.aspetjournals.org/content/66/2/240.long https://www.sciencedirect.com/science/article/pii/b9780444537416000271?via%3dihub https://www.sciencedirect.com/science/article/pii/s0890623810002625?via%3dihub https://www.sciencedirect.com/science/article/pii/s0890623810002625?via%3dihub https://journals.sagepub.com/doi/abs/10.1191/0748233704th198oa https://www.ncbi.nlm.nih.gov/pmc/articles/pmc4199042/pdf/ijerph-11-09660.pdf http://creativecommons.org/licenses/by/3.0/ nanogels as nocarriers for drug delivery: a review doi: http://dx.doi.org/10.5599/admet.724 1 admet & dmpk 8(1) (2020) 1-15; doi: http://dx.doi.org/10.5599/admet.724 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index review nanogels as nanocarriers for drug delivery: a review rahul rajput, jitendra narkhede, jitendra naik* university institute of chemical technology, kbc north maharashtra university, jalgaon, 425001, india *corresponding author: e-mail: jbnaik@nmu.ac.in received: september 16, 2019; revised: december 12, 2019; available online: december 21, 2019 abstract nanogels are submicron-size aqueous dispersions of water-swollen particles, composed of nano-sized three-dimensional highly cross-linked networks of hydrophilic polymers. an active pharmaceutical agent or therapeutic agent with high or low molecular weight can be easily encapsulated into nanogels that can be delivered to the site of action via various routes, including oral, pulmonary, nasal, parenteral and intraocular routes, among others. therapeutic agents encapsulated into nanogels improve the therapeutic activity in the biological environment. the application of different nanogels in drug delivery and recent clinical trial studies has been described concisely in this review. ©2020 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords drug; polymers; nanogels; drug delivery; applications introduction nanogels are commonly used in sensing, diagnostics, and bioengineering, but they are also often used in drug delivery [1,2]. nanogels have benefits over conventional and macro-sized delivery systems because of their higher drug loading capacity, high stability, and improved contact time with the skin surface, which makes it more convenient as a transdermal drug delivery system. water-soluble non-ionic polymers like hydroxylpropyl methylcellulose as well as ethylcellulose are commonly used to stabilise nanogel dispersions [3-7]. phase separation of drug-loaded nanogels could occur due to interactions (electrostatic, hydrophobic, van der waals) between the polymeric matrix and the active agent, which could be prevented by dispersing hydrophilic polymers. the dispersed hydrophilic polymer becomes exposed to the skin surface by forming a protective layer around the nanogel, allowing drug particles to remain dispersed in the gel matrix [8-10]. modified natural biopolymers possess a high degree of functional groups with additional functional crosslinkers used for the formulation of biopolymer-based nanogels. innovative techniques such as photopolymerisation, chemical cross-linking, click chemistry-based cross-linking etc., are used to derive the self-assembly and cross-linking of hydrophilic block copolymers. between internal and external layers of nanogels, block polymers permit the control of drug release from a polymer matrix [11-13]. for targetspecific or cell-specific drug delivery, nanogels are modified with ligands to permit receptormediated drug release at the site of action [14-15]. drugor biologically-loaded nanogels cross biological barriers and release the therapeutic agent inside cells [16-18]. in recent years, nanogels were effectively http://dx.doi.org/10.5599/admet.724 http://dx.doi.org/10.5599/admet.724 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:jbnaik@nmu.ac.in http://creativecommons.org/licenses/by/4.0/ jitendra naik et al. admet & dmpk 8(1) (2020) 1-15 2 utilised in the field of biotechnology to deal with genetics, enzyme immobilisation and protein synthesis, thus providing an efficient tool to cater for novel therapeutic systems in medicine (table 1). a novel coreshell magnetic nanogel was prepared using poly(acrylamide) for cancer therapy. pluronic poly(ethyleneimine) was used to prepare a thermoresponsive nanogel for the transdermal delivery of an active agent. nanogel-based drug delivery formulations increase the effectiveness and safety of certain anti-cancer drugs as well as many other drugs due to their chemical composition, which has been confirmed by in vivo studies in animal models. nanogels are a favourable and innovative drug delivery system that can play a vital role by addressing the problems associated with old and modern therapeutics such as nonspecific effects and poor stability [19-23]. table 1. applications and types of nanogel in drug delivery [21-23] polymer type of nanogel uses pullulan/folate-pheophorbide self-quenching polysaccharide-based minimal phototoxicity of pheophorbide cross-linked branched network of poly(ethyleneimine) and peg polyplex nanogel elevated activity and reduced cytotoxicity of fludarabine acetylated chondroitin sulphate self-organising nanogel doxorubicin loaded heparin pluronic nanogel self-assembled nanogel rna enzyme delivery internalized in cells poly(ethyleneimine) nanogels size-dependent property nanogel suicide gene htert –cd-tk delivered for lung cancer poly(n-isopropylacrylamide) and chitosan thermosensitive magnetically modalised hyperthermia cancer treatment and targeted drug delivery poly(acrylamide) novel core shell magnetic nanogel radiopharmaceutical carrier for cancer radiotherapy methylacrylic acid and n,n’methylene-bis-(acrylamide) supermagnetic nanogel functionalised with carboxyl group α-chymotrypsin immobilized on aminated nanogel methylacrylic acid and n,n’methylene-bis-(acrylamide) magnetic nanogel hydrophilic polymers α-chymotrypsin immobilized on carboxyl group poly(ethyleneimine) nanogels size-dependent property nanogel suicide gene htert –cd-tk delivered for lung cancer acylate group modified cholesterol bearing pullulan cross-linked raspberry-like assembly nanogel efficient interleukin-12 encapsulation and plasma levels poly(n-isopropylacrylamide-coacrylamide) in situ gelatinised thermosensitive nanogel drug loading capacity, bovine serum albumin glycol chitosan grafted with 3diethylaminopropyl groups ph-responsive doxorubicin uptake accelerated acetylated hylauronic acid specific targeting nanogel doxorubicin loaded nanogel pluronic poly(ethyleneimine) temperature responsive and volume transition nanogels thermo responsive endosomal rupture by nanogel and drug release methods used for the preparation of nanogels i) preparation of nanogel using polymeric precursors ii) synthesis of nanogel network by heterogeneous polymerisation of monomers polymers and monomers having different nanoscopic structures formed by amphiphilic copolymers are used for the preparation of nanogels. amphiphilic copolymers are able to self-assemble in solution, hence admet & dmpk 8(1) (2020) 1-15 nanogels as nocarriers for drug delivery doi: http://dx.doi.org/10.5599/admet.724 3 they form nanogels [24-25]. i) preparation of nanogels from polymeric precursors disulphide-based cross-linking disulphide linkages were prepared by inverse mini-emulsion atom transfer radical polymerisation (atrp). in this process, water-soluble monomer oligo(ethylene glycol) monomethyl ether methacrylate (oeoma) with different molecular weights were cross-linked in the atrp reaction with the disulphidefunctionalised cross-linker. the nanogels formed are considered to have a uniformly cross-linked network, which is supposed to improve aqueous solubility and control the release of encapsulated agents; the nanogels are shown to be biodegradable into water-soluble polymers in the presence of a biocompatible glutathione tripeptide which is commonly found in cells. the biodegradation of nanogels could trigger the release of drugs from the nanogels. amphiphilic random copolymers are used to prepare a nanogel system by self-cross-linking [26-27]. a nanosized product in aqueous solution is formed by hydrophilic poly(ethylene glycol) and pyridyl disulphide. thiol disulfide exchange reaction is depending primarily on the concentration of thiol exchangers like dithiothreitol (dtt). the size of the nanogel would be reduced by using cross-linking monomer or polymer chains. a lower critical solution temperature (lcst) of polymers also affects the size of the nanogel. lipoic acid-encapsulated dextran was prepared by thiol-exchange using the same method. a catalytic amount of dtt freely cross-linked with doxorubicin was synthesised from the assembly of the polymer [28-29]. poly(ε-caprolactone) (pcl) and hydrophilic poly(ethylethylenephosphate) (peep) were used as drug carriers for the development of a micellar nanoparticle system for intracellular drug release which is triggered by glutathione in tumour cells. tang et al. synthesised a disulphide-linked di-block copolymer of pcl and peep (pcl-ss-peep), which forms biocompatible micelles in aqueous solution and detaches the shell material under glutathione stimulus, resulting in rapid drug release with the destruction of the micellar structure shown in figure 1 [30]. figure 1. synthesis pathway of the disulphide-linked pcl-ss-peep and schematic illustration of intracellular drug release (reprinted with the permission from [30]. copyright (2016) american chemical society). http://dx.doi.org/10.5599/admet.724 jitendra naik et al. admet & dmpk 8(1) (2020) 1-15 4 amine group is more common in amine-based cross-linked nanogel development because the amine group has higher reactivity against carboxylic acids, isocyanates, and iodides. cross-linked knedelresembling structures using amine cross-linkers were prepared by the wooley group. hydrophilic, amphiphilic di-block copolymers were prepared by reversible addition fragmentation chain transfer polymerisation. amidation of carboxyl group-containing self-assembled block copolymers with a diamine cross-linking agent leads to the cross-linking of micellar assemblies; the remaining carboxylic group was altered for orthogonal surface modifications in the form of other functional moieties to form a cross-linked nanogel. furthermore, reaction with isocyanate carriers is an alternative cross-linking approach to prepare nanogels. ph-responsive cross-linked micelles were prepared by the addition of 1,8-diaminooctane to a micellar combination of 3-isopropenyl-α,α-dimethyl benzyl isocyanate bearing copolymers [31-32]. we prepared poly(acrylic acid) (paa) and sodium carboxymethylcellulose (nacmc)-based luliconazole loaded nanogels. luliconazole was encapsulated in paa and na-cmc by free radical polymerisation. luliconazole is an azole antifungal that works by preventing the growth of the fungus [33] and is used to treat skin infections such as athlete's foot, jock itch, and ringworm. figure 2 shows particle size analysis for the optimised nanogel containing luliconazole. the average particle size of nanogel was 259 nm at the count rate of 360 kcps with polydispersity index (pdi) of 0.2 showed narrow particle size distribution. figure 2. average particle size of luliconazole loaded poly(acrylic acid) nanogel the formation of nanogel (nacmc-g-paa) from nacmc and acrylic acid/sodium acrylate in the presence of n,n’-methylene bisacrylamide (mba) was carried out using potassium persulphate as a free radical initiator [34-35]. figure 3 shows the fesem micrograph of the optimised nanogel formulation. from the figure, it can be observed that the formed nanogel is spherical in nature. click chemistry-based cross-linking this method is discovered by wooley and hawker group for the synthesis of nanogels [36]. in this method, alkynyl groups were restrained to the corona of assembled micelles prepared from amphiphilic diblock copolymers of poly(acrylic acid)-b-polystyrene via the amidation of acrylic acid groups. azido dendrimers and micelles are generally prepared by click reactions. covalently cross-linked micelles are entrapped into the nanogel assemblies. cross-linked polyion nanogel micelles were prepared by liu et al. using a click chemistry approach. core cross-linked polyion complex micelles had thermosensitive coronas with high stability against salt and ph [36-37]. admet & dmpk 8(1) (2020) 1-15 nanogels as nocarriers for drug delivery doi: http://dx.doi.org/10.5599/admet.724 5 figure 3. fesem image of luliconazole loaded poly(acrylic acid) nanogel thermoresponsive poly(n-isopropylacrylamide) (pnipam) represents an attractive candidate to introduce physical cross-linking via the association of hydrophobic domains because it has a gelling temperature below body temperature and good biocompatibility. pnipam is a non-biodegradable, thermoreversible hyaluronan-poly (n-isopropylacrylamide) (ha-pnipam) hydrogel with a well-defined molecular architecture and properties; this hydrogel can be synthesised through reversible addition fragmentation chain transfer (raft) polymerisation and “click” chemistry polymerisation method [38]. the effect of pnipam grafting length and density on ha-pnipam properties was evaluated by methods relevant for cell therapy. it was found that the reversibility of the pnipam gelling process was improved in the presence of ha. the efficiency of the “click” reaction facilitates the control of the ds of pnipam chains. raft polymerisation allows the preparation of pnipam of controlled molecular weight and low pdi. this control of the critical parameters of pnipam molecular weight and grafting density allowed the gel to be figure 4. n-(3-dimethylaminopropyl)-n-ethylcarbodiimide hydrochloride (edc), n-hydroxy-succinimide (nhs) synthesis of hyaluronanpropargylamide (hapa), followed by copper-catalysed azidealkyne cycloaddition of hapa with azido-terminated poly(n-isopropylacrylamide) n3-pnipam (reprinted with the permission from [38]. copyright (2010) american chemical society). http://dx.doi.org/10.5599/admet.724 jitendra naik et al. admet & dmpk 8(1) (2020) 1-15 6 optimised for regenerative medicinal applications. the two synthetic steps of ha-pnipam are shown in figure 4. the edc/nhs mediated coupling of ppa to carboxylic acid groups on the hyaluronan salt and the copper-catalysed azide-alkyne cycloaddition of the n3-pnipam to the hyaluronan-propargylamide [38-40]. photo-induced cross-linking this is an alternative method for nanogel preparations. in this method, the polymer chain is stabilised using cross-linking and is functionalised with dimerisable or polymerisable units. double hydrophilic block copolymers can be encapsulated into the coumarin dimer. coumarin dimers, when cured with uv light >310 nm, are assembled into micelles and then photo-cross-linked to form nanogels (figure 5). figure 5. scheme for uv photopolymerisation this nanogel shows interaction between lower critical solution temperature (lcst) and upper critical solution temperature (ucst) behaviours. ucst is the critical temperature above which the contents of a mixture are miscible. the light penetrating capability has been incorporated into dendrimer structures to increase drug release in response to light stimulation. dendrimers of coumarin act as alternative crosslinkers to control the accessibility of substrate in the nanogel networks. when the solution of coumarin was cured in uv light, the ester groups were confined in the interior assembly of nanogel. enzymatic degradation of the substrate was very much inhibited. curing the cross-linked assembly by uv light improves the enzymatic action due to the de-cross-linking of the coumarin dimer, which exposes the substrate to enzymes [41-43]. a stimuli-responsive nanogel prepared in water by a core cross-linking technique using diblock copolymer micelles by photo-cross-linking of the micelle core [44]. the preparation of poly(ethylene glycol)-b-poly(2-(diethylamino) ethyl methacrylate-co-2-cinnamoyloxyethyl acrylate) (pegb-p(deaema/cea)), a ph-responsive block copolymer, by reversible addition-fragmentation chain transfer (raft)-controlled radical polymerisation using poly(ethylene glycol)-based chain transfer agent (peg-cta). the poly(ethylene glycol) (peg) block in peg-b-p(deaema/cea) is soluble in water, independent of ph, while the solubility of deaema depends on ph (figure 6) [44-45]. we have formulated poly(acrylamide) (paa) nanogel using uv polymerisation. paa gels are mainly polymerised using catalysts such as tetramethylenediamine (temed) and ammonium persulphate (aps), which are highly toxic and lead to slow polymerisation, which is time-consuming and takes 45 min to 1 h for lower gel precursor concentrations. photo-crosslinking with various photoinitiators, such as irgacure 2959, has been more recently used for the synthesis of paa hydrogels with a stiffness gradient and used for the quick preparation of large paa hydrogel arrays for applications such as drug screening. photo-crosslinking circumvent the use of toxic catalysts and is characteristically much faster than chemical cross-linking method. final properties of uv-polymerized gels are depending on the uv wavelength, consistency, light intensity, and interaction time [46]. figure 7 shows the average particle size of thymol loaded paa nanogel and figure 8 shows the fesem micrograph of thymol loaded paa. it is clear from the figure that the nanogel admet & dmpk 8(1) (2020) 1-15 nanogels as nocarriers for drug delivery doi: http://dx.doi.org/10.5599/admet.724 7 prepared by uv-polymerisation is spherical in nature. figure 6. synthetic route for poly (ethylene glycol)-b-poly(2-(diethylamino) ethyl methacrylate-co-2cinnamoyloxyethyl acrylate) (reprinted with permission from [45]. copyright (2009) american chemical society). figure 7. average particle size of thymol loaded polyacrylamide nanogel prepared by photopolymerisation physical cross-linking physically cross-linked gels are also known as pseudo gels which have weaker van der waals linkages, hydrogen bonding, hydrophobic or electrostatic interactions that are involved in the synthesis of pseudo gels. the physicochemical properties of gels depend on properties of the polymer, temperature, ionic strength, concentration of polymer and the cross-linking agent. combination of amphiphilic block copolymers and complexation of oppositely charged polymeric chains is used for the formulation of pseudo gels [43,47]. http://dx.doi.org/10.5599/admet.724 jitendra naik et al. admet & dmpk 8(1) (2020) 1-15 8 figure 8. fesem image of thymol loaded poly(acrylamide) acid nanogel prepared by photopolymerisation ii) synthesis of nanogel networks by heterogeneous polymerisation of monomers bi-functional monomers are chemically entrapped into nanogels. heterogeneous colloidal systems are responsible for the activation of polymerisation. emulsion polymerisation and atrp are used for the preparation of biodegradable nanogel. disulphide-linked bi-functional monomers are used in the stimulation of polymerisation. protein nanogel hybrids using atrp in water/oil mini emulsions or inverse mini-emulsion are useful for the entrapment of covalently bonded proteins into nanogel. in the inverse mini-emulsion, a co-initiator was used to initiate the polymerisation of monomers which are firmly dispersed in the system [48-49]. drug release mechanism of nanogel ph-responsive mechanism ph-responsive, nanosized nanogels have received significant attention because of their biological relevance and due to their potential applications in drug delivery systems. drug release is affected by the different ph values throughout the human body physiological conditions. ph-responsive block copolymer micelles are suitable for controlled delivery applications. in such applications, however, the polymer micelles may experience dilutions below the critical micelle concentration (cmc), leading to dissociation into monomers. in contrast, nanogels with a cross-linked structure are robust at a diluted concentration. insoluble 3d structures and staying alive at low ph are the main characteristics of methacrylic acid ethyl acrylate. the polymeric chain repulsions begin and lead to the precise release profile in procaine hydrochloride due to the cumulative ph ranges of acidic group ionisation. suitable ph at the site of action helps with the diffusion of nanogels. ph-responsive monomers play an important role in the preparation of nanogels; these are commonly ph-responsive functional groups that deionise in the polymeric assemblies. a nanogel containing platinum nanoparticles exhibited on and off catalytic activity for shifting reactive oxygen types. in the acidic ph of skin, the protonation of pendant amine of cross-linked poly(2-(n,ndiethylamino) methacrylate) core as well as peg group in the polymer greatly enhances the solubilisation of drug [43,50-51]. admet & dmpk 8(1) (2020) 1-15 nanogels as nocarriers for drug delivery doi: http://dx.doi.org/10.5599/admet.724 9 thermosensitive and volume transition mechanism variations in the capacity of nanogels according to temperature are known as the volume phase transition temperature (vptt). polymers become quenched and hydrated when the surrounding medium is below the vptt. a shrunken and hydrated polymer swells and releases the loaded therapeutic agent. thermo-responsive nanogels rupture in cells and the biological environment when they swell and rise in volume. n-isopropyl acrylamide synthesised nanogels have thermoresponsive properties. these nanogels have important characteristics, such as rapid contraction in gel volume and the efflux of indomethacin due to the maintenance of heat beyond the lower critical solution temperature (lcst). the poly(nisopropylacrylamide-co-acrylamide)-loaded 5-fluorouracil gel has been tested on rats in ex vivo studies. the loading of the therapeutic agent at lower temperatures and the release from nanogels at body temperature makes this suitable for drug delivery. pluronic acid-modified thermoresponsive poly(ethyleneimine) nanogels were effectively used as gene delivery systems. thermoresponsive nanogels with pnipam have very exciting and promising applications in the biomedical field, such as the treatment of certain cancers through hyperthermia. they can be loaded with an anticancer drug and, at the target location, by moderately increasing the temperature above the lcst, the nanogel can change with volume and the drug release can be increased [52]. photoisomerisation and photochemical internalization stimulation of photosensitiser-loaded nanogels leads to the synthesis of singlet oxygen and reactive oxygen species which causes oxidation of cellular compartment walls such as endosomal barrier walls; this affects the release of therapeutics into the cytoplasm. an azo dextran nanogel loaded with aspirin showed the e-configuration of the azole group rather than the z-configuration at 365 nm; cis-trans isomerisation of azobenzene by photo-regulation in an azo-dextran nanogel loaded with aspirin as a model drug exhibited that the e-configuration of the azo group leads to a better release profile of the drug than the zconfiguration at 365 nm radiation [12,43,53]. miscellaneous examples degradation of disulphide linkages in cross-linked hyaluronic acid nanogels causes the degradation of the nanogel assembly due to the action of reducing agents; in this way, doxorubicin is released by the simple diffusive process. the size of the nanogel increases and the layer by layer release of an active ingredient is possible without a rapid burst of the drug. the release can be sustained by simple diffusion and controlled following initial release mediated by a coating with anionic and cationic polyelectrolytes [54]. the application of nanogels nanogel-based drug delivery formulations improve the effectiveness and safety of anti-cancer drugs, antifungal drugs, and anti-diabetic drugs, due to their physicochemical properties, as well as improving the ease of administration, as confirmed by in vivo studies. nanogels have minimum toxicity to nearby tissues and high healing effects in cancer treatment at the site of action [55]. transdermal drug delivery of an antipyretic drug the nanosized dispersion of aceclofenac was formulated by emulsion-solvent diffusion methods and then incorporated into a carbopol 940. the formulation showed optimal permeability properties and stability, and achieved a sustained drug release. a nanogel formulation containing diclofenac sodium was prepared by the emulsion-solvent diffusion method and then incorporated into a carbopol 940 [56]. http://dx.doi.org/10.5599/admet.724 jitendra naik et al. admet & dmpk 8(1) (2020) 1-15 10 ophthalmic applications curcumin-loaded cationic nanostructured lipid carriers (cnlc) were prepared by film-ultrasonic techniques and thermosensitive gelling agents were used to improve pre-ocular retention and the ocular permeation capacity of curcumin. muscone has maximum drug loading in the hydrogel, and the rheology results showed that the phase transition temperature was 34°c. blinking of eyes was resisted due to the thixotropy; the recovery time indicated that hydrogel was effective [43,57]. diabetic applications in diabetic patients, insulin is injected into muscles every day, which is a very painful process. to overcome this problem lee et al. (2012) developed a chitosan-loaded inhalable deoxycholic acid altered glycol chitosan (doca-gc) nanogel. nanogels are self-assembled due to hydrophobic attractions with deoxycholic acid; these nanogels formed constant hypoglycaemia over a period of 2 days comparatively at the low dose [58]. carrier for antifungal agents in fungal infections, physicians and patients prefer the transdermal route. a fluconazole-chitin nanogel was formulated by using regeneration chemistry and the wet milling method. chitin nanogels were redeveloped from chitin solution. fluconazole-chitin has a controlled release pattern which is perfect for the continuous availability of fluconazole over a longer period for effective fungal treatment [59]. the synthesis of a vitamin e nanoemulsion-based nanogel consisting of the high molecular weight active agent amphotericin b has been effectively used for cutaneous fungal infections; the nanogel showed a nearly 4fold higher skin deposition through porcine ear skin [60]. nanogels in diagnostics and imaging nanogels have properties like structural flexibility, high water content, fluid-like transport, biodegradability, and biocompatibility. gadolinium-assembled nanogels were synthesised by the crosslinking of branched polyethyleneimines with gadolinium ions. inverse microemulsion followed by surfacefunctionalisation with polyethylene glycol chains was performed to increase the blood circulation time [61]. properties of nanogels high water content/swellability nanogels have rapid swelling and de-swelling properties. water-soluble nanogels show the benefits of hydrogels with certain advantages that are necessary for their nanoscale size. like microgels, nanogels can contain and protect drugs and regulate their release by integrating high-affinity functional groups containing polymers [43,62]. softness the softness of nanogels is a very important parameter in the biomedical field and alters their biodistribution properties. softness can be adjusted by changing the chemical structure of the nanogel [63]. colloidal stability the surface charge of polymers inhibits the development of aggregates in the bloodstream, along with their associated problems. this can be altered by increasing the zeta potential that results in higher repulsive forces between particles which electrostatically stabilise nanogels. another method includes the integration of surfactants like polyethylene glycol which produce a steric effect and hydration forces to give a stable nanosuspension [43,64]. admet & dmpk 8(1) (2020) 1-15 nanogels as nocarriers for drug delivery doi: http://dx.doi.org/10.5599/admet.724 11 biocompatibility and degradability natural or synthetic polymers are used to synthesise nanogels. these are biocompatible and biodegradable, thereby preventing their accumulation in the systemic circulation. chitosan, poly-acrylic acid, methyl cellulose, sodium alginate, and several polysaccharide-based polymers like dextran, pullulan, and cyclodextrin can be used to formulate nanogels. polysaccharides are typically carbohydrate-based polymers formed of repeating monosaccharide units linked by glycosidic bonds. these polymers are stable, non-toxic, hydrophilic and biodegradable in nature [65]. particle size nanogels are able to diffuse through the skin, tissues or compromised areas of the endothelium and in some cases through a specific transport system. some routes of administration face the challenge of crossing the blood-brain barrier (bbb) due to their particle size. so, to overcome this issue, nanogels were developed which have a size in the diameter range from 20-200 nm. nanogels have smaller sizes, so cross the bbb while inhibiting rapid clearance mechanisms at the same time [43,66]. concluding remarks nanogels are advanced pharmaceutical nanocarriers for pharmaceutical agents as well as therapeutic agents. nanogel systems could be easily prepared with biomacromolecules with the maximum entrapment ability and stability of the resulting formulation in dispersion. nanogel systems control pharmaceuticallyactive compounds with different drug structures. biopolymers and low molecular mass hydrophobes can also be encapsulated in nanogels. the discovery of a new polymeric system is very important for the development of nanogels. advanced polymerisation or cross-linking approaches have a promising role in therapies. this is a new approach in the synthesis of nanogel assemblies. hence, we can expect that these advanced nanocarrier systems will be focused upon in future pharmaceutical developments. acknowledgments the authors would like to thank technical education quality improvement program (teqip-iii), world bank and mhrd, new delhi for providing financial support to carry out this work. conflicts of interest: the authors have no conflicts of interest. references [1] j.k. oh, d.i. lee, j.m. park. biopolymer-based microgels/nanogels for drug delivery applications. progress in polymer science 34(12) (2009) 1261-1282. [2] j.k. oh, r. drumright, d.j. siegwart, k. matyjaszewski. the development of microgels/nanogels for drug delivery applications. progress in polymer science 33(4) (2008) 448-477. [3] r.t. chacko, j. ventura, j. zhuang, s. thayumanavan. polymer nanogels: a versatile nanoscopic drug delivery platform. advanced drug delivery reviews 64(9) (2012) 836-851. [4] m. hamidi, a. azadi, p. rafiei. hydrogel nanoparticles in drug delivery. advanced drug delivery reviews 60(15) (2008) 1638-1649. [5] s.a. ferreira, f.m. gama, m. vilanova. polymeric nanogels as vaccine delivery systems. nanomedicine: nanotechnology, biology and medicine 9(2) (2013) 159-173. [6] t. hoare, d. sivakumaran, c.f. stefanescu, m.w. lawlor, d.s. kohane. nanogel scavengers for drugs: local anesthetic uptake by thermoresponsive nanogels. acta biomaterialia 8(4) (2012) 1450-1458. http://dx.doi.org/10.5599/admet.724 jitendra naik et al. admet & dmpk 8(1) (2020) 1-15 12 [7] h. ayame, n. morimoto, k. akiyoshi. self-assembled cationic nanogels for intracellular protein delivery. bioconjugate chemistry 19(4) (2008) 882-890. [8] j. ramos, a. imaz, j. forcada. temperature-sensitive nanogels: poly(n-vinylcaprolactam) versus poly(n-isopropylacrylamide). polymer chemistry 3(4) (2012) 852-856. [9] w. lv, s. liu, w. feng, j. qi, g. zhang, f. zhang, x. fan. temperature‐and redox‐directed multiple self assembly of poly (n‐isopropylacrylamide) grafted dextran nanogels. macromolecular rapid communications 32(14) (2011) 1101-1107. [10] f. cavalieri, a. postma, l. lee, f. caruso. assembly and functionalization of dna− polymer microcapsules. acs nano 3(1) (2009) 234-240. [11] m. molinos, v. carvalho, d.m. silva, f.m. gama. development of a hybrid dextrin hydrogel encapsulating dextrin nanogel as protein delivery system. biomacromolecules 13(2) (2012) 517-527. [12] f. schmitt, l. lagopoulos, p. käuper, n. rossi, n. busso, j. barge, g. wagnières, c. laue, c. wandrey, l. juillerat-jeanneret. chitosan-based nanogels for selective delivery of photosensitizers to macrophages and improved retention in and therapy of articular joints. journal of controlled release 144(2) (2010) 242-250. [13] m.c. ferrer, s. dastgheyb, n.j. hickok, d.m. eckmann, r.j. composto. designing nanogel carriers for antibacterial applications. acta biomaterialia 10(5) (2014) 2105-2111. [14] e.a. murphy, b.k. majeti, r. mukthavaram, l.m. acevedo, l.a. barnes, d.a. cheresh. targeted nanogels: a versatile platform for drug delivery to tumors. molecular cancer therapeutics 10(6) (2011) 972-982. [15] s.a. ferreira, c. oslakovic, r. cukalevski, b. frohm, b. dahlbäck, s. linse, f.m. gama, t. cedervall. biocompatibility of mannan nanogel—safe interaction with plasma proteins. biochimica et biophysica acta (bba)-general subjects 1820(7) (2012) 1043-1051. [16] l. he, h. liang, l. lin, b.r. shah, y. li, y. chen, b. li. green-step assembly of low-density lipoprotein/sodium carboxymethyl cellulose nanogels for facile loading and ph-dependent release of doxorubicin. colloids and surfaces b: biointerfaces 126 (2015) 288-296. [17] n. li, j. wang, x. yang, l. li. novel nanogels as drug delivery systems for poorly soluble anticancer drugs. colloids and surfaces b: biointerfaces 83(2) (2011) 237-244. [18] k. akiyoshi, n. morimoto, h. ayame. self-assembled cationic nanogels for intracellular protein delivery. bioconjugate chemistry 19(4) (2008) 882-890. [19] s. honary, f. zahir. effect of zetz potential on the properties of nano-drug delivery systemsa review. tropical journal of pharmaceutical research 12(2) (2013) 255-264. [20] c. zhao, q. chen, k. patel, l. li, x. li, q. wang, g. zhang, j. zheng. synthesis and characterization of ph-sensitive poly(n-2-hydroxyethyl acrylamide)–acrylic acid (poly (heaa/aa)) nanogels with antifouling protection for controlled release. soft matter 8(30) (2012) 7848-7857. [21] p. lemieux, s. vinogradov, c. gebhart, n. guerin, g. paradis, h.k. nguyen, b. ochietti, y. suzdaltseva, e. bartakova, t. bronich, y. st-pierre. block and graft copolymers and nanogel™ copolymer networks for dna delivery into cell. journal of drug targeting 8(2) (2000) 91-105. [22] y. lee, s.y. park, c. kim, t.g. park. thermally triggered intracellular explosion of volume transition nanogels for necrotic cell death. journal of controlled release 135(1) (2009) 89-95. [23] y. sasaki, k. akiyoshi. nanogel engineering for new nanobiomaterials: from chaperoning engineering to biomedical applications. the chemical record 10(6) (2010) 366-376. [24] y. xia, c. cheng, r wang, h. qin, y. zhang, l. ma, h. tan, z. gu, c. zhao. surface-engineered nanogel assemblies with integrated blood compatibility, cell proliferation and antibacterial property: towards multifunctional biomedical membranes. polymer chemistry 5(20) (2014) 5906-5919. [25] n. morimoto, t. endo, m. ohtomi, y. iwasaki, k. akiyoshi. hybrid nanogels with physical and chemical cross‐linking structures as nanocarriers. macromolecular bioscience 5(8) (2005) 710-716. admet & dmpk 8(1) (2020) 1-15 nanogels as nocarriers for drug delivery doi: http://dx.doi.org/10.5599/admet.724 13 [26] z.y. qiao, r. zhang, f.s. du, d.h. liang, z.c. li. multi-responsive nanogels containing motifs of ortho ester, oligo (ethylene glycol) and disulfide linkage as carriers of hydrophobic anti-cancer drugs. journal of controlled release 152(1) (2011) 57-66. [27] h. lee, h. mok, s. lee, y.k. oh, t.g. park. target-specific intracellular delivery of sirna using degradable hyaluronic acid nanogels. journal of controlled release 119(2) (2007) 245-252. [28] j. ramos, a. imaz, j. forcada. temperature-sensitive nanogels: poly(n-vinylcaprolactam) versus poly(n-isopropylacrylamide). polymer chemistry 3(4) (2012) 852-856. [29] s. patnaik, a.k. sharma, b.s. garg, r.p. gandhi, k.c. gupta. photoregulation of drug release in azodextran nanogels. international journal of pharmaceutics 342(1) (2007) 184-193. [30] l.y. tang, y.c. wang, y. li, j.z. du, j. wang. shell-detachable micelles based on disulfide-linked block copolymer as potential carrier for intracellular drug delivery. bioconjugate chemistry 20(6) (2009) 1095-1099. [31] d. basak, s. ghosh. ph-regulated controlled swelling and sustained release from the core functionalized amphiphilic block copolymer micelle. acs macro letters 2(9) 2013 ) 799-804. [32] o. borisova, l. billon, m. zaremski, b. grassl, z. bakaeva, a. lapp, p. stepanek, o. borisov. synthesis and ph-and salinity-controlled self-assembly of novel amphiphilic block-gradient copolymers of styrene and acrylic acid. soft matter 8(29) (2012) 7649-7659. [33] p. klinpituksa, p. kosaiyakanon. superabsorbent polymer based on sodium carboxymethyl cellulose grafted polyacrylic acid by inverse suspension polymerization. international journal of polymer science (2017) 3476921. [34] c. duan, d. zhang, f. wang, d. zheng, l. jia, f. feng, y. liu, y. wang, k. tian, f. wang, q. zhang. chitosan-g-poly(n-isopropylacrylamide) based nanogels for tumor extracellular targeting. international journal of pharmaceutics 409(1-2) (2011) 252-259. [35] a. sharma, t. garg, a. aman, k. panchal, r. sharma, s. kumar, t. markandeywar. nanogel—an advanced drug delivery tool: current and future. artificial cells, nanomedicine, and biotechnology 44(1) (2016) 165-177. [36] r.k. iha, k.l. wooley, a.m. nyström, d.j. burke, m.j. kade, c.j. hawker. applications of orthogonal “click” chemistries in the synthesis of functional soft materials. chemical reviews 109(11) (2009) 5620-5686. [37] d.p. chan, s.c. owen, m.s. shoichet. double click: dual functionalized polymeric micelles with antibodies and peptides. bioconjugate chemistry 24(1) (2013) 105-113. [38] d. mortisen, m. peroglio, m. alini, d. eglin. tailoring thermoreversible hyaluronan hydrogels by “click” chemistry and raft polymerization for cell and drug therapy. biomacromolecules 11(5) (2010) 1261-1272. [39] a. sanchez-sanchez, i. pérez-baena, j. pomposo. advances in click chemistry for single-chain nanoparticle construction. molecules 18(3) (2013) 3339-3355. [40] r.t. chacko, j. ventura, j. zhuang, s. thayumanavan. polymer nanogels: a versatile nanoscopic drug delivery platform. advanced drug delivery reviews 64(9) (2012) 836-851. [41] j.m. spruell, m. wolffs, f.a. leibfarth, b.c. stahl, j. heo, l.a. connal, j. hu, c.j. hawker. reactive, multifunctional polymer films through thermal cross-linking of orthogonal click groups. journal of the american chemical society 133(41) (2011) 16698-16706. [42] a. leitner, t. walzthoeni, r. aebersold. lysine-specific chemical cross-linking of protein complexes and identification of cross-linking sites using lc-ms/ms and the xquest/xprophet software pipeline. nature protocols 9(1) (2014) 120-137. [43] f. sultana, manirujjaman, m. imran-ul-haque, m. arafat, s. sharmin. an overview of nanogel drug delivery system. journal of applied pharmaceeutical science 3(8) (2013) s95-s105. [44] k.w. chun, j.b. lee, s.h. kim, t.g. park. controlled release of plasmid dna from photo-cross-linked pluronic hydrogels. biomaterials. 26(16) (2005) 3319-3326. http://dx.doi.org/10.5599/admet.724 jitendra naik et al. admet & dmpk 8(1) (2020) 1-15 14 [45] s.i. yusa, m. sugahara, t. endo, y. morishima. preparation and characterization of a ph-responsive nanogel based on a photo-cross-linked micelle formed from block copolymers with controlled structure. langmuir 25(9) (2009) 5258-5265. [46] s. sheth, e. jain, a. karadaghy, s. syed, h. stevenson, s.p. zustiak. uv dose governs uv-polymerized polyacrylamide hydrogel modulus. international journal of polymer science (2017) 5147482. [47] m.a. pujana, l. pérez-álvarez, l.c. iturbe, i. katime. biodegradable chitosan nanogels crosslinked with genipin. carbohydrate polymers 94(2) (2013) 836-842. [48] t.g. van thienen, k. raemdonck, j. demeester, s.c. de smedt. protein release from biodegradable dextran nanogels. langmuir 23(19) (2007) 9794-9801. [49] w.c. lee, y.c. li, i. chu. amphiphilic poly(d,l‐lactic acid)/poly(ethylene glycol)/poly(d,l‐lactic acid) nanogels for controlled release of hydrophobic drugs. macromolecular bioscience 6(10) (2006) 846854. [50] a. tamura, m. oishi, y. nagasaki. enhanced cytoplasmic delivery of sirna using a stabilized polyion complex based on pegylated nanogels with a cross-linked polyamine structure. biomacromolecules 10(7) (2009) 1818-1827. [51] g. tamura, y. shinohara, i. akiba, a. tamura, m. oishi, y. nagasaki, k. sakurai, y. amemiya. phresponsive structural change of pegylated amine-bearing nanogel explored by small angle x-ray scattering. journal of physics: conference series 272 (2011) 012018. [52] h. tokuyama, y. kato. preparation of poly (n-isopropylacrylamide) emulsion gels and their drug release behaviors. colloids and surfaces b: biointerfaces 67(1) (2008) 92-98. [53] r. cheng, f. feng, f. meng, c. deng, j. feijen, z. zhong. glutathione-responsive nano-vehicles as a promising platform for targeted intracellular drug and gene delivery. journal of controlled release 152 (2011) 2–12. [54] p.p. constantinides, m.v. chaubal, r. shorr. advances in lipid nanodispersions for parenteral drug delivery and targeting. advanced drug delivery reviews 60 (2008) 757–767. [55] h.s. abandansari, m.r. nabid, s.j. rezaei, h. niknejad. ph-sensitive nanogels based on boltorn® h40 and poly (vinylpyridine) using mini-emulsion polymerization for delivery of hydrophobic anticancer drugs. polymer 55(16) (2014) 3579-3590. [56] s. talele, p. nikam, b. ghosh, c. deore, a. jaybhave, a. jadhav. a research article on nanogel as topical promising drug delivery for diclofenac sodium. indian journal of pharmaceutical education and research 51 (4s) (2017) 5680-5685. [57] r. liu, l. sun, s. fang, s. wang, j. chen, x. xiao, c. liu. thermosensitive in situ nanogel as ophthalmic delivery system of curcumin: development, characterization, in vitro permeation and in vivo pharmacokinetic studies. pharmaceutical development and technology 21(5) (2016) 576-582. [58] g. wang, q. nie, c. zang, b. zhang, q. zhu, g. luo, s. wang. self-assembled thermoresponsive nanogels prepared by reverse micelle→ positive micelle method for ophthalmic delivery of muscone, a poorly water-soluble drug. journal of pharmaceutical sciences 105(9) (2016) 2752-2759. [59] j. lee, c. lee, t.h. kim, e.s. lee, b.s. shin, s.c. chi, e.s. park, k.c. lee, y.s. youn. self-assembled glycol chitosan nanogels containing palmityl-acylated exendin-4 peptide as a long-acting anti-diabetic inhalation system." journal of controlled release 161(3) (2012) 728-734. [60] m. nishil, s.n. rejinold, s. mangalathillam, r. biswas, s. nair, r. jayakumar. fluconazole loaded chitin nanogels as a topical ocular drug delivery agent for corneal fungal infections. journal of biomedical nanotechnology 9(9) (2013) 1521-1531. [61] l. kaur, s.k. jain, k. singh. vitamin e tpgs based nanogel for the skin targeting of high molecular weight anti-fungal drug: development and in vitro and in vivo assessment. rsc advances, 5 (2015) 53671-53686. [62] l.c. keun, a. singh, j. heo, d. kim, k.e. lee, h. jeon, j. koh, i.c. kwon, s. kim. gadolinium-coordinated elastic nanogels for in vivo tumor targeting and imaging. biomaterials 34(28) (2013) 6846-6852. [63] i. neamtu, a.g. rusu, a. diaconu, l.e. nita, a.p. chiriac. basic concepts and recent advances in nanogels as carriers for medical applications. drug delivery 24(1) (2017) 539-557. admet & dmpk 8(1) (2020) 1-15 nanogels as nocarriers for drug delivery doi: http://dx.doi.org/10.5599/admet.724 15 [64] m.d. moya-ortega, c. alvarez-lorenzo, a. concheiro, and t. loftsson. cyclodextrin-based nanogels for pharmaceutical and biomedical applications. international journal of pharmaceutics 428(1-2) (2012) 152-163. [65] k.s. soni, s.s. desale, t.k. bronich. nanogels: an overview of properties, biomedical applications and obstacles to clinical translation. journal of controlled release 240 (2016) 109-126. [66] h.k.s yadav, n.a. al halabi, g.a. alsalloum. nanogels as novel drug delivery systems-a review. journal of pharmacy and pharmaceutical research 1 (2017) 5. ©2020 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://dx.doi.org/10.5599/admet.724 http://creativecommons.org/licenses/by/3.0/ validated hplc method for quantification of copanlisib in mice plasma: application to a pharmacokinetic study doi: http://dx.doi.org/10.5599/admet.782 113 admet & dmpk 8(1) (2020) 113-121; doi: http://dx.doi.org/10.5599/admet.782 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper validated hplc method for quantification of copanlisib in mice plasma: application to a pharmacokinetic study ashok zakkula 1 , pavan kumar kurakula 2 , sreekanth dittakavi 1 , prasanthi daram 1 , ram murthi bestha 1 , mohd zainuddin 1 , ravi kumar trivedi 1 , ramesh mullangi 1 * 1 drug metabolism and pharmacokinetics, jubilant biosys ltd, industrial suburb, yeshwanthpur, bangalore -560 022, india 2 department of pharmacology, raghavendra institute of pharmaceutical education and research, anantapur-515721, a.p, india *corresponding author: e-mail: mullangi.ramesh@jubilantinnovation.com; tel.: +91-80-66628339 received: january 29, 2020; revised: february 18, 2020; published online: march 04, 2020 abstract copanlisib is a pan phosphatidylinositol 3-kinase (pi3k) inhibitor approved for follicular lymphoma. in this paper, we present the data of development and validation of a high-performance liquid chromatography (hplc) method for the quantitation of copanlisib in mice plasma as per the fda regulatory guideline. the method involves the extraction of copanlisib along with internal standard (is, enasidenib) from mice plasma (100 µl) using ethyl acetate as an extraction solvent. the chromatographic resolution of copanlisib and the is was achieved on a hypersil gold c18 column maintained at 40 °c using a binary gradient mobile phase [10 mm ammonium formate (ph 4.0) and acetonitrile]. the flow-rate was 0.8 ml/min. for the detection of copanlisib and the is, the photo-diode array detector was set at λmax 310 nm. copanlisib and the is eluted at 6.60 and 7.80 min, respectively with a total run time of 10 min. the calibration curve was linear over a concentration range of 50 to 5000 ng/ml for copanlisib (r 2  0.998). the results of intraand inter-day accuracy and precision studies were within the acceptable limits. copanlisib was stable on bench-top, in auto-sampler, up to three freeze/thaw cycle and long-term storage at -80 °c. the application of the validated method was shown in a mice pharmacokinetic study. ©2020 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords copanlisib; hplc; method validation; mice plasma; pharmacokinetics introduction phosphoinositide 3-kinases (pi3ks) are lipid kinases, which play a pivotal role in cell cycle, cellular metabolism, apoptosis etc. along with α-serine/threonine-protein kinase (akt)/mammalian target of rapamycin (mtor) pathway [1]. abnormal activation of pi3k pathway has been shown to drive tumorigenesis [2]. among three classes of the pi3k enzyme, class i pi3k is linked with malignance and it exists in four isoforms: α, β, γ and δ [3,4]. pi3k inhibitors represent a novel class of targeted therapies for the treatment of human malignancies [5]. preclinical and clinical studies demonstrated that inhibition of pi3k is an effective therapeutic strategy for lymphomas treatment [6]. copanlisib (fig. 1; bay 80-6946), is a http://dx.doi.org/10.5599/admet.782 http://dx.doi.org/10.5599/admet.782 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:mullangi.ramesh@jubilantinnovation.com http://creativecommons.org/licenses/by/4.0/ r. mullangi et al. admet & dmpk 8(1) (2020) 113-121 114 pan-pi3k inhibitor with predominant activity against pi3k-α and pi3k-δ forms, which are expressed in malignant b-cells. the ic50 (half-maximal inhibitory concentration) against pi3k α, β, γ and δ isoforms was 0.5, 3.7, 6.4 and 0.7 nmol/l, respectively [7]. it induces tumor cell death by apoptosis and inhibition of proliferation of primary malignant b cell lines. copanlisib was approved for treatment of follicular lymphoma as an intravenous infusion for adults. in clinic, at 0.8 mg/kg (intravenous infusion), copanlisib showed promising efficacy in patients with solid tumors and hematological malignancies. across the dose range of 0.1-1.2 mg/kg, it has shown dose proportional increase in cmax (maximum concentration in plasma) and auc0-t (area under curve from time zero to last measurable time point). typically, cmax attained between 0.5-1.0 h (tmax). the human plasma protein binding of copanlisib is ~84 %. copanlisib was majorly metabolized by cyp3a4 (>90 % metabolism) and to some extent by cyp1a1 (~10 % of metabolism) to yield m1, an active metabolite, which possess similar potency as copanlisib. the terminal half-life and clearance for copanlisib at 0.8 mg/kg were 39.1 h and 18.9 l/h. copanlisib has high volume of distribution (871 l). up to 50 % of unchanged drug and remaining amount in the form of metabolites is excreted in humans [8]. figure 1. structural representation of copanlisib and enasidenib (internal standard, is) until date, only one bioanalytical method was published for quantification of copanlisib [9]. in the reported lc-ms/ms (liquid chromatography coupled with tandem mass spectrometry) method, authors used one step liquid-liquid extraction method for mice plasma samples processing. the linearity range was 3.59-3588 ng/ml. using an isocratic mobile phase, copanlisib and the internal standard were resolved on a hypurity c18 column having a total run time of 3.0 min. although lc-ms/ms is a powerful tool for the quantitation of drugs in various biological matrices with higher sensitivity in short run time, its high cost and maintenance limit its availability for most of the hospitals, academic institutes and research laboratories. the lower cost and affordability of the hplc instrument compared to that of lc-ms/ms renders hplc methods more eligible for wide routine use. in clinic, at efficacy dose of 0.8 mg/kg, copanlisib plasma concentrations were ~50 ng/ml at 8 h (both on day-1 and day-15) in patients with non-hodgkin’s lymphoma and advanced solid tumors [10,11]. by achieving 50 ng/ml sensitivity for copanlisib on hplc-uv, we believe our present method can be used in hospitals for routine therapeutic drug monitoring of copanlisib. besides, the proposed method can also be in research laboratories for routine pharmacokinetic and/or toxicokinetic studies samples analysis. in order to ensure the reliability, reproducibility and sensitivity of the method, the developed analytical method was validated for various parameters in accordance with fda guideline. the validated method was applied to investigate the pharmacokinetics of copanlisib post intravenous administration to mice. admet & dmpk 8(1) (2020) 113-121 copanlisib quantification by hplc in mice plasma doi: http://dx.doi.org/10.5599/admet.782 115 materials and methods chemicals and reagents copanlisib (purity: 99.7 %) was obtained from beijing yibai biotechnology co., ltd, beijing, china. enasidenib (purity: 98 %) used as an internal standard (is) was purchased from aaron, shanghai, china. solutol, ethanol and dimethyl sulfoxide (dmso) were purchased from sigma-aldrich, st. louis, mo, usa. hplc grade acetonitrile and methanol were purchased from j.t. baker avantor, pa, usa. analytical grade hydrochloric acid and ammonium formate were purchased from s.d. fine chemicals, mumbai, india. all other chemicals and reagents were of analytical grade and used without further purification. the control mice k2.edta plasma was procured from animal house, jubilant biosys, bangalore. hplc operating conditions analysis of copanlisib in plasma samples was performed on a waters 2695 alliance hplc system (waters, milford, usa) equipped with performance plus inline degasser along with an auto-sampler, column oven and photo diode array (pda) detector set at max 310 nm. chromatographic resolution of copanlisib and the internal standard (is) was achieved by injecting 25 µl of the processed sample on a hypersil gold c18 column (250  4.0 mm, 5 µm; thermo scientific, usa) maintained at 40±1 °c using a binary gradient mobile phase consisting 10 mm ammonium formate, ph: 4.0 (adjusted with formic acid) (solvent a) and acetonitrile (solvent b) delivered at a flow-rate of 0.8 ml/min. initial eluent composition was 90 % a, maintained for 5.0 min, and followed by a linear 0.5 min ramp to 10 % a, which was maintained for until 5.5 min. the mobile phase composition returned to 90 % a at 8.0 min. equilibration time was 2.0 min. preparation of stock solutions for copanlisib and the is for the preparation of calibration curve (cc) and quality control (qc) samples, two primary stock solutions of copanlisib were made at 1.0 mg/ml in 0.1 n hcl:dmso (2:98, v/v). due to copanlisib limited solubility in organic solvents we have used 0.1 n hcl. the primary stock solution of the is (1.0 mg/ml) was prepared in dmso and subsequently diluted with 80% methanol to get work stock at 0.60 μg/ml. the primary stock solutions of copanlisib and the is were stored at -20±5 °c, which were found to be stable for 50 days. preparation of calibration curve standards and quality control samples the first set of primary stock solution of copanlisib was diluted appropriately and subsequently used to prepare a calibration curve (cc) standards. the calibration standard samples were made by spiking the blank mice plasma (90 µl) with each corresponding working solution of copanlisib (10 µl) thereby yielding final concentrations of 50, 100, 500, 750, 1250, 2500, 3750 and 5000 ng/ml. for the determination of precision and accuracy, samples were prepared by spiking blank mice plasma in bulk with the second working stock solution of copanlisib at appropriate concentrations and 100 l aliquots were distributed into different tubes. the qcs prepared were: 50 ng/ml (lower limit of quantification quality control; lloq qc), 150 ng/ml (low quality control; lqc), 2250 ng/ml (medium quality control; mqc) and 3500 ng/ml (high quality control; hqc). all the qcs were stored together at -80±10 °c until analysis. http://dx.doi.org/10.5599/admet.782 r. mullangi et al. admet & dmpk 8(1) (2020) 113-121 116 sample preparation to an aliquot of 100 µl mice plasma sample, 1.0 ml of ethyl acetate was added and vortex mixed for 3 min; followed by centrifugation for 5 min at 14,000 rpm in a refrigerated centrifuge (eppendorf 5424r) maintained at 5 °c. the organic layer (800 µl) was separated and evaporated to dryness at 50 °c using a gentle stream of nitrogen (turbovap®, zymark®, kopkinton, ma, usa). the residue was reconstituted in 100 µl of the is solution (600 ng/ml) and 25 µl was injected onto hplc system for analysis. validation procedures a full validation according to the us fda guidance was performed for the quantitation of copanlisib in mice plasma [12]. the selectivity of the proposed method was assessed by evaluating the presence of interfering the peaks at the retention times of copanlisib and the is in six different batches of blank mice plasma samples. the auto-injector carry over was determined by injecting the highest calibration standard (5.0 µg/ml) followed by injection of mice plasma blank samples. recovery of copanlisib was determined by comparing their respective response from qcs (lqc and hqc) after the extraction process against their non-extracted samples. recovery of the is was determined at 600 ng/ml. intraand inter-day accuracy and precision were determined at four qc levels [lloq qc (50 ng/ml), lqc (150 ng/ml), mqc (2250 ng/ml) and hqc (3500 ng/ml)] along with calibration curve (0.05-5.00 μg/ml). stability (auto-sampler, bench-top, freezethaw and long-term) studies, dilution effect and incurred sample reanalysis (isr) were also evaluated as per regulatory guideline requirement [12]. pharmacokinetic study in mice twelve male balb/c mice (weight range: 26-28 g) were procured from vivo biotech, hyderabad, india. animal study protocol used in this study was approved by the institutional animal ethics committee, jubilant biosys (iaec/jdc/2019/188r). mice were housed for a period of seven days having free access to feed and water before the pharmacokinetic study. mice received copanlisib intravenously [2 % 0.1 n hcl, 10 % dmso, 10 % solutol:absolute alcohol (1:1, v/v) and 78 % normal saline; strength: 0.5 mg/ml; dose volume: 10 ml/kg] at 5.0 mg/kg as a bolus dose. blood samples (100 µl) were collected at pre-determined time points (0.12, 0.25, 0.5, 1, 2, 4, 8 and 24 h) through retro-orbital plexus (using micropipettes, drummond scientific, pa, usa; catalogue number: 1-000-0500) into polypropylene tubes (having k2.edta as an anti-coagulant). sparse sampling technique (three mice per time point) was adopted during blood collection so that blood loss from each mouse was kept less than 10 % of total blood volume. plasma was harvested by centrifuging the blood using biofuge (hereaus, germany) at 1760 g for 5 min and stored frozen at -80±10 °c until analysis. mice were allowed to access feed 2 h post-dosing. pharmacokinetic analysis pharmacokinetic parameters were calculated by a non-compartmental method using phoenix winnonlin 8.1 software (pharsight, mountain view, ca, usa). key pharmacokinetic parameters like extrapolated plasma drug concentration at time zero following intravenous bolus injection (c0), area under the curve from time zero to infinity (auc0-), volume of distribution (vd), total body clearance (cl) and halflife (t½) were determined for copanlisib. admet & dmpk 8(1) (2020) 113-121 copanlisib quantification by hplc in mice plasma doi: http://dx.doi.org/10.5599/admet.782 117 results and discussion method development and optimization several trials were taken with various columns, mobile phase compositions to select the chromatographic conditions, which will give a good resolution of copanlisib and the is from the endogenous matrix substances within a suitable run time. several mobile phases were tried by changing the combination of different organic solvents (acetonitrile and methanol) and buffers (eg: formic acid, ammonium acetate, phosphate buffer etc.) with altered flow-rates (in the range of 0.60-1.20 ml/min). to choose a stationary phase a variety of columns namely x-terra phenyl, atlantis, hypersil gold c18 were evaluated. our trials revealed that mobile phase comprising 10 mm ammonium formate (ph 4.0):acetonitrile delivered in a gradient program at a flow-rate of 0.8 ml/min on hypersil gold c18 column gave a stable base line with good resolution between copanlisib and the is with a total run time of 10 min with no interference of endogenous plasma peaks. the uv detector was set at λmax 310 nm. for the optimized conditions, enasidenib was found to be a suitable internal standard as it exhibited good resolution, retention time and uv absorbance intensity (uv λmax 287 nm) at the same wave length of copanlisib. method validation with protein precipitation technique the recovery of copanlisib and is was very poor (<40 %). liquidliquid extraction with ethyl acetate gave best results in terms of extraction recovery, reproducibility and cleaner samples. the mean ± s.d recovery of copanlisib at lqc and hqc was 83.8 ± 4.51 and 83.9 ± 1.17 %, respectively. the recovery of the is was 99.5 ± 1.33 %. as shown in fig. 2, both copanlisib and the is peaks were well resolved and no interference at the retention times of copanlisib and the is from the endogenous components of mice. the retention time of copanlisib and the is was 6.60 and 7.80 min, respectively. the calibration curves (n=4) for copanlisib were observed to be linear in the range of 50-5000 ng/ml. a representative equation for the calibration curves is as follows: y = 0.0005 x + 0.003. a regression equation with a weighting factor of 1/x 2 of each drug to the is concentration was found to produce the best fit for the concentration-detector response relationship. the correlation coefficients (r 2 ) were more than 0.999, indicating an acceptable linearity of our method. the accuracy observed for the mean of backcalculated concentrations for four calibration curves was within 89.3-109 %; while the precision (%re) values ranged from 0.95-1.05 %. we did not observe any carry-over produced by the highest calibration sample on the following injected mice blank plasma extracted sample for copanlisib. figure 2. hplc chromatograms of a 25 µl injection of (a) blank mice plasma (b) blank mice plasma spiked with copanlisib (lloq: 50 ng/ml) along with the is (c) a 0.5 h plasma sample showing the peak of copanlisib (concentration: 3500 ng/ml) following intravenous administration of copanlisib to mice at 5.0 mg/kg. http://dx.doi.org/10.5599/admet.782 r. mullangi et al. admet & dmpk 8(1) (2020) 113-121 118 table 1. intraand inter-day precision and accuracy determination of copanlisib quality controls in mice plasma accuracy and precision data for intraand inter-day mice plasma samples determined for copanlisib (from four different batches) at lloq qc (50 ng/ml), lqc (150 ng/ml), mqc (2250 ng/ml) and hqc (3500 ng/ml) are presented in table 1. the intraand inter-day precisions (rsd) were within 7.59 %, and accuracy (re) ranged between 0.97-1.07 %. the assay values on both the occasions (intraand interday) were found to be within the accepted variable limits indicating that the present method is reproducible, accurate and precise. table 2 summarizes the results of stability studies conducted for copanlisib in mice plasma. the measured concentrations for copanlisib at lqc (150 ng/ml) and hqc (3500 ng/ml) deviated within ±15 % of the nominal concentrations in a battery of stability tests namely in-injector (16 h), benchtop (6 h), repeated three freeze/thaw cycles and freezer stability at -8010 °c for 30 days (table 2) supported the stability of copanlisib at various stability conditions. the dilution integrity was confirmed for qc samples that exceeded the upper limit of the quantitation (uloq) of calibration curve (up to 35000 ng/ml). the mean accuracy and precision for 15 times diluted samples were found to be less than 7.62 and 6.25 %, respectively, which show the ability to dilute samples up to a dilution factor of ten in a linear fashion (table 2). all the samples selected for isr met the acceptance criteria. the back-calculated accuracy values ranged between 89.2-110 % from the initial assay results as shown in fig. 3 using a bland-altman plot. pharmacokinetic study plasma samples collected during pharmacokinetic study were thawed at room temperature and processed as mentioned in sample preparation section. along with plasma samples, lqc, mqc and hqc samples (made in blank plasma) were assayed in duplicate and were distributed among unknown samples in the analytical run. plasma samples showed high concentration above the high calibration standard (5.00 µg/ml) were diluted appropriately with mice blank plasma to bring the concentration within linearity range. the criteria for acceptance of the analytical runs encompassed the following: (i) ≥67 % of qc theoretical concentration (ng/ml) batch measured concentration (ng/ml) mean ± sd rsd accuracy, % intra-day (six replicates at each concentration) 50.0 1 49.4 ± 4.76 9.63 98.9 2 50.8 ± 3.89 7.65 101 3 54.3 ± 1.35 2.49 108 4 51.0 ± 3.26 6.53 103 150 1 139 ± 1.95 1.39 93.2 2 151 ± 5.77 3.80 101 3 146 ± 2.88 1.98 97.3 4 145 ± 3.46 2.36 97.3 2250 1 2155 ± 46.5 2.16 95.8 2 2120 ± 78.9 3.72 94.2 3 2297 ± 38.1 1.66 102 4 2190 ± 54.1 2.50 97.4 3500 1 3460 ± 81.2 2.32 112 2 3507 ± 70.0 2.00 104 3 3545 ± 53.3 1.50 105 4 3516 ± 68.1 1.94 107 inter-day (twenty four replicates at each concentration) 50.0 51.2 ± 3.91 7.59 103 150 146 ± 6.22 4.28 97.3 2250 2190 ± 95.1 4.35 96.9 3500 3504 ± 72.0 2.06 99.9 rsd: relative standard deviation (sd  100/mean) re: relative error (measured value/actual value) sd: standard deviation admet & dmpk 8(1) (2020) 113-121 copanlisib quantification by hplc in mice plasma doi: http://dx.doi.org/10.5599/admet.782 119 samples should be ±15 % of the nominal concentration values (ii) ≥ 50 % of qc samples per level should be ±15 % of their nominal concentration values [12]. table 2. stability and dilution integrity data for copanlisib in mice plasma experiment spiked concentration (ng/ml) measured concentration (ng/ml) mean ± sd a (n = 6) accuracy (%) b precision (% cv) bench-top (6 h) stability 150 150 ± 2.72 99.8 1.80 3500 3548 ± 66.6 101 1.88 in-injector (16 h) stability 150 146 ± 4.77 98.0 3.25 3500 3808 ± 34.0 107 0.89 freeze-thaw (3 cycles) stability 150 159 ± 12.3 106 7.70 3500 3774 ± 107 107 2.84 long-term stability at -80°c (30 days) 150 146 ± 3.44 97.3 2.35 3500 3758 ± 23.0 106 0.61 *dilution integrity 2333 2522 ± 403 108 6.25 rsd: relative standard deviation (sd x 100/mean) re: relative error (measured value/actual value) *plasma samples prepared at 35,000 ng/ml (10-fold above hqc) diluted with blank plasma by 15-fold and analyzed figure 3. bland-altman plot showing the incurred sample reanalysis (isr) data for copanlisib in the mice, plasma concentrations of copanlisib decreased mono-exponentially after intravenous administration. fig. 4 depicts the mean ± s.d plasma concentrations versus time for copanlisib following intravenous administration to mice at 5.0 mg/kg. copanlisib was quantifiable up to 8.0 h post intravenous administration to mice. post intravenous administration, the cl and vd were found to be 42.1 ml/min/kg and 27.6 l/kg, respectively. the auc0-∞ attained post intravenous administration was 1978 (ng h)/ml. the t½ was 7.58 h. in summary the validated method was sensitive enough to calculate the pharmacokinetic parameters of copanlisib. patnaik et al. [10] and kim et al. [11] reported the plasma concentrations of copanlisib post administration at efficacy dose (0.8 mg/kg as 1 h intravenous infusion) to patients with advanced solid tumors and non-hodgkin’s lymphoma. plasma samples collected in these studies on day-1 and day-15 were analyzed using an lc-ms/ms method. we have digitalized the reported plasma concentrations versus time plots of copanlisib reported by patnaik et al. [10] and kim et al. [11] using digitizeit software [13] and found http://dx.doi.org/10.5599/admet.782 r. mullangi et al. admet & dmpk 8(1) (2020) 113-121 120 that in both studies copanlisib showed ~50 ng/ml till 8 h. by achieving 50 ng/ml sensitivity for copanlisib on hplc-uv in the present method, we believe our present method can be reliably used in hospitals for routine therapeutic drug monitoring of copanlisib. by increasing the plasma volume for sample processing and injection volume for hplc analysis, there is a great possibility that our validated hplc-uv can be used to quantify the copanlisib plasma concentration at terminal time points at therapeutic doses. figure 4. mean plasma concentration versus time profile for copanlisib in mice plasma following intravenous administration to mice. conclusions a simple reversed-phase hplc method for determination of copanlisib in mice plasma has been developed and validated. the proposed method is highly specific, accurate, precise and reproducible. all the validation parameters were within the acceptable limits for a bioanalytical method as per regulatory guideline. this method has been successfully applied to a pharmacokinetic study in mice. conflict of interest: the authors are scientists at jubilant biosys ltd. references [1] l.c. cantley. the phosphoinositide 3-kinase pathway. science 296 (2002) 1655-1657. [2] t.l. yuan, l.c. cantley. pi3k pathway alterations in cancer: variations on a theme. oncogene 27 (2008) 5497-5510. [3] k.d. courtney, r.b. corcoran, j.a. engelman. the pi3k pathway as drug target in human cancer. journal of clinical oncology 28 (2010) 1075-1083. [4] j.e. kurtz, i. ray-coquard. pi3 kinase inhibitors in the clinic: an update. anticancer research 32 (2012) 2463-2470. [5] b. markman, j.j. tao, m. scaltriti. pi3k pathway inhibitors: better not left alone. current pharmaceutical design 19 (2013) 895-906. [6] b.l. lampson, j.r. brown. pi3kδ-selective and pi3kα/δ-combinatorial inhibitors in clinical development for b-cell non-hodgkin lymphoma. expert opinion on investigational drugs 26 (2017) 1267-1279. [7] n. liu, b.r. rowley, c.o. bull, c. schneider, a. haegebarth, c.a. schatz, p.r. fracasso, d.p. wilkie, m. hentemann, s.m. wilhelm, w.j. scott, d. mumberg, k. ziegelbauer. bay 80-6946 is a highly selective intravenous pi3k inhibitor with potent p110α and p110δ activities in tumor cell lines and xenograft models. molecular cancer therapeutics 12 (2013) 2319-2330. [8] aliqopa tm . aliqopa (copanlisib) injection [prescribing information]. whippany, nj: bayer healthcare pharmaceuticals inc.; september 2017. admet & dmpk 8(1) (2020) 113-121 copanlisib quantification by hplc in mice plasma doi: http://dx.doi.org/10.5599/admet.782 121 [9] s. dittakavi, r. mullangi. lc-esi-ms/ms determination of copanlisib, a novel pi3k inhibitor in mouse plasma and its application to a pharmacokinetic study in mice. biomedical chromatography 33 (2019) e4460. [10] a. patnaik, l.j. appleman, a.w. tolcher, k.p. papadopoulos, m. beeram, d.w. rasco, g.j. weiss, j.c. sachdev, m. chadha, m. fulk, s. ejadi, j.m. mountz, m.t. lotze, f.g.s. toledo, e. chu, m. jeffers, c. peña, c. xia, s. reif, i. genvresse, r.k. ramanathan. first-in-human phase i study of copanlisib (bay 80-6946), an intravenous pan-class i phosphatidylinositol 3-kinase inhibitor, in patients with advanced solid tumors and non-hodgkin’s lymphomas. annals of oncology 27 (2016) 1928-1940. [11] r.d. kim, s.r. alberts, c. peña, i. genvresse, a. ajavon-hartman4, c. xia, a. kelly, j.e. grilley-olson. phase i dose-escalation study of copanlisib in combination with gemcitabine or cisplatin plus gemcitabine in patients with advanced cancer. british journal of cancer 118 (2018) 462-470. [12] us dhhs, fda, cder, cvm, guidance for industry: bioanalytical method validation, u.s. department of health and human services, food and drug administration, center for drug evaluation and research (cder), center for veterinary medicine (cv), (2018), rockville, md, usa. [13] https://www.digitizeit.de version 2.0.0; accessed on 10 january 2020. ©2020 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://dx.doi.org/10.5599/admet.782 https://www.digitizeit.de/ http://creativecommons.org/licenses/by/3.0/ perspectives in solubility measurement and interpretation doi: 10.5599/admet.686 88 admet & dmpk 7(2) (2019) 88-105; doi: http://dx.doi.org/10.5599/admet.686 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index review perspectives in solubility measurement and interpretation christel a.s. bergström 1 and alex avdeef 2, * 1 department of pharmacy, uppsala university, bmc p.o. box 580, se-751 23 uppsala, sweden 2 in-adme research, 1732 first avenue, #102, new york, ny 10128, usa *corresponding author: e-mail: alex@in-adme.com; tel.: +1-646-678-5713 orcid id: christel a.s. bergström: 0000-0002-8917-2612; alex avdeef: 0000-0002-3139-5442 received: march 12, 2019; revised: march 26, 2019; published: april 05,2019. abstract several key topics in solubility measurement and interpretation are briefly summarized and illustrated with case studies drawing on published solubility determinations as a function of ph. featured are examples of ionizable molecules that exhibit solubility-ph curve distortion from that predicted by the traditionally used henderson-hasselbalch equation and possible interpretations for these distortions are provided. the scope is not exhaustive; rather it is focused on detailed descriptions of a few cases. topics discussed are limitations of kinetic solubility, ‘brick-dust and grease-balls,’ applications of simulated and human intestinal fluids, supersaturation and the relevance of pre-nucleation clusters and sub-micellar aggregates in the formation of solids, drug-buffer/excipient complexation, hydrotropic solubilization, acid-base ‘supersolubilization,’ cocrystal route to supersaturation, as well as data quality assessment and solubility prediction. the goal is to highlight principles of solution equilibria – graphically more than mathematically – that could invite better assay design, to result in improved quality of measurements, and to impart a deeper understanding of the underlying solution chemistry in suspensions of drug solids. the value of solid state characterizations is stressed but not covered explicitly in this mini-review. keywords solubility-ph; shake-flask solubility; intrinsic solubility; thermodynamic solubility; henderson-hasselbalch equation; supersaturation; pre-nucleation clusters, drug aggregates; drug salts; pharmaceutical cocrystals introduction for many oral drugs, the rate-limiting step in intestinal absorption is dictated by the kinetics of dissolution of the active pharmaceutical ingredient from the solid form. the dissolution rate is a function of a number of factors, foremost being the aqueous solubility of the drug, which, for ionizable molecules, can be affected by the ph of the medium [1]. also, depending on the drug that has been introduced, the dissolution may overshoot the saturation concentration and produce a supersaturated solution. the supersaturation is typically achieved when the api is introduced in its amorphous form or in a charged form as a salt. in such a supersaturated solution, drug molecules may self-associate as sub-micellar aggregates/clusters, particularly if they are surface-active or polarizable. for bases introduced as drug salts into an alkaline-ph solution, the charged drug in the supersaturated solution can disproportionate into oil or undergo precipitation into a new amorphous solid, with which charged water-soluble aggregates may coexist. given enough time, the oil or amorphous solid (and any other type of aggregates) are expected to undergo phase transformation into a thermodynamically more stable crystalline solid, the steps governed http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:alex@in-adme.com admet & dmpk 7(2) (2019) 88-105 perspectives in solubility measurements and interpretation doi: 10.5599/admet.686 89 by kinetics of nucleation and crystal growth. for drugs that are practically insoluble in water, dissolution and re-crystallization can be particularly complicated kinetic processes occurring at different interfaces of, e.g., undissolved api or colloidal structures composed of excipients. aggregation and recrystallization are to date not fully understood and hence may be difficult to predict. about two-thirds of the compounds in the drug discovery pipeline are very poorly soluble in water, and therefore show low and/or erratic intestinal absorption. to increase solubility (and thus absorption), oral drugs are often formulated as salts. however, there are numerous other strategies which have been successfully implemented [2,3]. several aspects of the above complexity are summarized in this mini-review. it is intended to serve as prologue or accompaniment to an upcoming session on solubility at the iapc-8 meeting in split, croatia, 911 september 2019, to be co-chaired by the authors. since 2009, the international association of physical chemists (iapc, www.iapchem.org) series of symposia maintained extensive coverage of the topic of solubility measurement, both from solid state and solution perspectives. at the 2015 iapc-4 meeting, the “thermodynamic solubility measurement of practically insoluble ionizable drugs case studies & suggested method improvements” special session resulted in a ‘white paper’ publication drawing on expert consensus thoughts of researchers from six countries (hungary, russia, serbia, spain, sweden, united states) [4]. at the 2017 iapc-6 meeting, the “pharmaceutical cocrystals -physicochemical properties and formulations” special session included presentations from internationally-recognized experts in cocrystal solubility-ph measurement. several publications on the topic emerged out of the meeting. it is anticipated that future sessions will continue to cover solubility methods and strategies, to critically address the different needs in pharmaceutical research, spanning from drug discovery to drug development. brief survey of topics in solubility several key topics are briefly summarized below and are illustrated with case studies drawing on solubility determinations as a function of ph. significant effort was made to find useful examples from the published literature which illustrated principles of solution equilibria. the coverage is not exhaustive, but is considerable in some cases, to address issues of measurement and interpretation of ph dependency in solubility profiles. the value of solid state characterizations is stressed but not covered explicitly in this mini-review. kinetic or thermodynamic? it’s been more than two decades since lipinski’s rule of 5 was published [5], addressing the problems of poorly soluble compounds. high-throughput microtitre plate methods for determining ‘kinetic’ solubility based on turbidity detection became very popular. usually, a dmso solution of the drug is added gradually to an aqueous buffer, until the first signs of solid formation are detected. the resultant precipitate is most often amorphous, up to a hundred times more soluble than that measured by conventional ‘thermodynamic’ (slow-to-medium throughput) methods. pharmaceutical company databases mostly contain kinetic values. gradually, it became evident that placing heavy focus on using kinetic data could drive the lead-to-candidate optimization process in the wrong direction: towards making more amorphoustending compounds, generating problems later on in development [6-8]. limitations of these early methods can lead to misconceptions and promote the lack of understanding of solubility data [7]. one way to mitigate some of the development-stage risks is to implement ‘miniaturized’ solubility methods (which can produce values closer to those of shake-flask thermodynamic measurements), brought forward alongside potency assessments [9]. this is not necessarily easy to do due to the inherently higher costs of measurement, particularly when directed at large compound libraries [10]. also, such implementation requires some understanding of solubility-ph relationships. http://www.iapchem.org/ bergström and avdeef admet & dmpk 7(2) (2019) 88-105 90 brick dust and grease balls low-soluble compounds are often colloquially classified as ‘brick dust’ or ‘grease balls’ [11]. the solubility of the former type is mainly limited by a strong crystal lattice, making it difficult for the compound to dissociate from the solid form. such ‘brick dusts’ are high melting point crystalline solids (tm > 200 °c) in which strong intermolecular bonds within the crystal lattice overcome the competing pull of solvation by water. in contrast, the ‘grease ball’ molecules are limited in their solubility by poor hydration. they typically have high octanol-water partition coefficients (log p > 3), and are hydrophobic compounds that form relatively low melting point solids. these compounds are unable to form strong bonds with the water molecules, and hence exhibit high log p [12]. contemporary medicinal chemistry synthesizes compounds as insoluble grease balls rather than brick dust [8,12,13]. traditional formulation strategies (excipients, solubilizing agents, cyclodextrins, amorphous dispersions, etc.) are likely to improve oral absorption of grease balls, more so than that of brick dust [12]. one may also speculate that many of the ‘brick dust’ molecules never reach the development stage because they are too problematic to work with in assays (e.g., pharmacological potency, toxicity and adme properties). hence, they are screened out early during the discovery stage. solubility in simulated and human intestinal fluids a drug’s solubility and dissolution behavior can be investigated in human intestinal fluid (hif), but such media are difficult to obtain since they require in vivo sampling of the fluid. hif composition is also highly variable [14], both intraand inter-individually, and it is relatively difficult to work with on the lab bench due to its low buffer capacity. while the variability is important to capture to understand and predict potential population effects, this complexity also makes it more suitable for later stage studies. the development of multiple simulated intestinal fluid (sif) recipes (fassif, fessif, others) applicable to the early drug development stage has therefore evolved [15]. a recent study by khadra et al. [16] applied a mathematically-rigorous design-of-experiment (doe) statistical technique to investigate how equilibrium drug solubility can be influenced by seven typical sif components (sodium taurocholate, lecithin, sodium phosphate, sodium chloride, ph, pancreatin and sodium oleate) within concentration ranges found in hif. poorly-soluble acids (naproxen, indomethacin, phenytoin, and piroxicam), bases (aprepitant, carvedilol, zafirlukast, tadalafil), and uncharged drugs (fenofibrate, griseofulvin, felodipine, probucol) were tested. with the exception of pancreatin, all of the factors individually had a significant influence on equilibrium solubility. differences between acids, bases, and neutral drugs were evident; however, the study also pointed towards (as many other studies also have done) that responses to changes in a solvent, e.g., micellar fraction and composition, salt and ph, are often compound specific and there are no generalizing rules or one-size-fits-all solutions. supersaturation, liquid-liquid phase separation, amorphous solid, oil-to-crystal transformations buffers can affect the formation of drug aggregates (i.e., molecular clusters). as an example: citrate, succinate, phthalate increase surface activity of phenothiazines; acetate decreases it [17]. certain surface active bases produce oil when the drug hydrochloride is converted to the free base in situ at ph > phmax. for example, a 30 μm chlorpromazine hydrochloride solution at ph 9 becomes supersaturated, which then partly separates as oil, which is more soluble than the crystalline free base. in contrast, a 2-5 μm alkaline solution (ph 9-12) prepared from chlorpromazine free base showed no surface activity [17]. figure 1 shows log solubility-ph profiles of haloperidol free base at different temperatures and in different buffer media reported by several groups [18-21]. the dashed curves refer to the profiles calculated from the simple henderson-hasselbalch (hh) equation (not corrected for ionic strength or admet & dmpk 7(2) (2019) 88-105 perspectives in solubility measurements and interpretation doi: 10.5599/admet.686 91 dilution effects), using the independently-determined pka values appropriate for the temperature. the solid curves are calculated using the refined constants from fitting the measured log s-ph data (filled circles) to the proposed equilibrium model, using the pdisol-x program (in-adme research; www.inadme.com/pdisol_x.html). s0 refers to the thermodynamic intrinsic solubility in pure water of the uncharged form of the dissolved substance in equilibrium with the crystalline solid. its value is indicated by the lowest value in the dashed curve, in the highly alkaline region. the supersaturation in the phmax region noted in the case of chlorpromazine hydrochloride also occurs for the haloperidol mesylate salt (as illustrated in figs. 3c,d in ref. [4]). however, it does not occur for the haloperidol free base. figures 1a,d do show deviations from the theoretical dashed curve (hh) in the alkaline region (ph > 7), where the uncharged species appears to possess elevated solubility. figure 1. haloperidol (free base) log s-ph profiles, illustrating alkaline-region distortions from the shape expected from the henderson-hasselbalch equation (dashed blue line). the solid red lines are the best fit to the measured data, using the regression analysis program pdisol-x [22]. see text for further elaboration. several phenomena can account for such a behavior: (i) the neutral (but not the charged) species forms a supersaturated solution; (ii) stable water-soluble uncharged aggregates of haloperidol form; (iii) the added crystalline free base, soon after dissolution, re-precipitate into an amorphous solid phase; (iv) liquidliquid phase separation (llps) takes place, where oil forms as the second liquid phase, in addition to the aqueous phase; and (v) the free-base (but not the charged-form) drug forms a complex with a component of the buffer, thus elevating solubility. overlapping contributions from the above effects may also play a role. without a specific molecular model for the phenomenon of ‘supersaturation’ (i), it is not clear why the extent of distortions is different in each of the frames of figure 1 (and also in figs. 3b,c,e in ref. [4]). case http://www.in-adme.com/pdisol_x.html http://www.in-adme.com/pdisol_x.html bergström and avdeef admet & dmpk 7(2) (2019) 88-105 92 (ii) – thermodynamically-stable aggregation reaction – is not supported by the data, since the distortion is different in each of the frames. cases (iii) and (iv) – amorphous solid and oil, resp. – are possible explanations for the distortion. figure 1c, based on a 48-h equilibration time, shows no distortion in the alkaline region. it could be that the amorphous solid or oil have essentially converted into the thermodynamically most stable crystalline form by then, but not entirely during the 24-h equilibration in figures 1a,d. the extent of distortion in figure 1a, compared to figure 1d, could be indicative of a possible role for drug-phosphate complex formation (v), since 0.15 m phosphate constituted the medium in figure 1a, but no buffers were in added in figure 1d. it is evident from the above examples that solid state characterization would be needed to help decide on the most likely mechanism behind the observed distortions in figure 1. changing buffer types in the medium and selecting different equilibration-time approaches [4] would add further insight into the mechanistic basis for the shape distortions in figure 1. it is worth noting that a comprehensive analysis of each of the log s-ph sets in figure 1 (e.g., mass action model using all relevant chemical equilibria [22]) yields very similar log s0 values, showing systematic temperature dependence. based on van’t hoff analysis (e.g., eq. (3) in ref. [23]), the intercept value of log s0 25 at 25 °c is -5.78 ±0.02 and the slope value (enthalpy of solution), δhsol o = 29 ± 6 kj/mol. based on the analysis of ten different published studies, the interlaboratory mean value for haloperidol is log s0 25 = -5.71 ± 0.17 (0.73 μg/ml) for haloperidol [24]. the above 0.17 log unit standard deviation is precisely the value obtained in the averaging of 870 molecules whose solubility was reported in two or more studies, as described below. a common practice in the literature is to report the measured solubility near ph = pka +/2 (for bases/acids, resp.) as the intrinsic value. the above examples suggests that caution is warranted for such a procedure: from figures 1a,c,d, the mean value would be -5.07 ±0.82 (3.2 ±16 μg/ml), which would not represent the realistic accuracy of the underlying data. a common perception is that reproducibility in solubility measurements is typically not better than 0.7 log unit [25]. the haloperidol example suggests that better reproducibility can be achieved if published data are critically analyzed. the diclofenac example below will further elaborate this. figure 2. time sequence of liquid-liquid phase boundary transformation to crystalline phase at 0, 12, and 22 h [26]. reprinted with permission from s. veesler, l. lafferrère, e. garcia, c. hoff, phase transitions in supersaturated drug solution. org. proc. res. dev. 7 (2003) 983-989. copyright 2003 american chemical society. the haloperidol discussion suggests that oil or amorphous solid forms in alkaline solutions. generally, a supersaturated solution forms on initial dissolution. such a solution can then spontaneously form two liquid phases (aqueous and oil), followed soon after by the appearance of crystals, as shown in the time sequence in figure 2. solute is transported from oil drops to crystal via the continuous phase in a solvent-mediated process in which the drops disappear as the crystal grows. after heating and cooling the crystals, llps is not regenerated, which further confirms that the llps is not a thermodynamically stable phase [26,27]. admet & dmpk 7(2) (2019) 88-105 perspectives in solubility measurements and interpretation doi: 10.5599/admet.686 93 amorphous and crystalline solids high-throughput solubility (hts) methods, where dmso solutions of the sample are added to buffer media, generally indicate a measure of the solubility of amorphous solids, with values greater than that indicated by crystalline solids. amorphous phases are not thermodynamically stable and will undergo crystallization over time. two amorphous phases may co-exist, or a transient oil phase may appear. an example of this is shown in figure 3a for the case of terfenadine measurement using the μsol method (pion inc.), with 1 %v/v dmso in each of the 96-well buffer solutions [28]. the unfilled circles represent solubility values collected at the end of 23-h equilibration, and may suggest the presence of oil plus amorphous solid. (alternatively, the ‘oil’ can be a more energetic amorphous solid co-precipitated with a more stable amorphous form.) the filled circles represent data collected at 68-h equilibration time, and may suggest the sole presence of the more stable amorphous solid. the apparent solubility at ph 11 is 1.9 μg/ml for the shorter time. this value appears to be an order of magnitude greater than the analyzed value of 0.16 μg/ml, corresponding to the apparent intrinsic solubility of the more stable amorphous solid. solid state characterizations were not reported for the above hts measurements. figure 3. terfenadine examples of oil/amorphous vs. crystalline phases solubility-ph profiles. (a) highthroughput measurement [29]. dotted curve is best-fit of 23-h equilibration data; solid curve is best-fit of 68h data. as a visual guide, the dashed curve is the hh curve calculated for crystalline terfenadine, based on the analysis of the data from al omari et al. (2007) [29], shown in the frame on the right. (b) shake-flask thermodynamic measurements [29]. see text. the dashed curve in figure 3a represents the thermodynamic solubility of the crystalline solid, 0.01 μg/ml, based on the reported citrate-buffer data of al omari et al. [29]. the detailed analysis of the latter is shown in figure 3b. the authors studied the solubility characteristics in 50 mm phosphate (unfilled circles) as well as citrate (filled circles) buffer. the 30 °c shake-flask data at 48-h equilibration appears to be of good quality. the phosphate-containing medium seems to indicate drug-buffer complexation for ph > phmax (not seen in the citrate medium), as evidenced by the elevation in solubility above the dashed hh curve (ph 7-8). the analysis suggests the equilibrium constant k = [tf.tfh + .hpo4 2]/[tf][tfh + ][hpo4 2] = 10 +8.96 m -2 for the monoanionic complex [24]. between phmax 6.7 and ph 8.3, the complex appears to shed one proton with increasing ph. in both the citrate and phosphate media, at ph below phmax, two salts form between the drug and each buffer, consistent with the solid and dotted curves being below the hh dashed curve. the salt solubilities are quite similar for the citrate and phosphate systems. figure 4 shows a simpler example of a single amorphous-phase log s-ph profile of clotrimazole, measured at about 1-h equilibration time, compared to the corresponding crystalline phase profile, measured at 48-h equilibration [30]. in both cases, the curves nearly follow the ph dependence calculated by the henderson-hasselbalch equation (with a constant vertical displacement between the two), although bergström and avdeef admet & dmpk 7(2) (2019) 88-105 94 the amorphous curve shows a slight amount liquid-liquid separation above ph 7, as indicated by a small difference between the solid and dashed upper curves in neutral solution (which is absent in the lower crystalline curves). figure 4. clotrimazole example of amorphous vs. crystalline phase solubility-ph profiles [30]. see text for details. pre-nucleation clusters (i.e. self-aggregates, sub-micelles, drug-buffer complexes) although classical nucleation theory describes drug crystallization from a supersaturated aqueous solution, the understanding of the molecular basis underlying the nucleation steps is still incomplete [31]. the role of pre-nucleation clusters as solute precursors to nucleation and the link between the chemical speciation of homogeneous solutions and the process of phase separation of such clusters was reviewed by gebauer et al. [31]. sub-micellar pre-nucleation clusters form in many supersaturated solutions: calcium phosphate, amino acids – particularly glycine (see below), and many surface-active drug molecules, such as in phosphate-buffered solutions of chlorpromazine, acetylpromazine, verapamil, orphenadrine, hydralazine, pramoxine, and amiodarone, at ph slightly above that of phmax [21,32]. the buffers in solution critically influence the stability of such clusters. the link between the clusters and the kinetics of crystallization is still work in progress, and could turn out to be a vibrant area for future investigations. it should also be noted that the loss of supersaturation may potentially start at the surface of the dissolving material that is provided in excess and not always from the supersaturated solution [33]. solubility-ph profiles can indicate self-aggregation/cluster formations. in the ph region where the drug is charged and self-aggregation takes place, the slope in the log s-ph plot can reveal the order of aggregation [28]. figure 5 shows three examples indicating formation of charged aggregates. in figure 5a, brequinar appears to form a negatively-charged dimer at ph above 7, as indicated by a slope of +2 in the log s-ph plot; the corresponding sodium-drug salt precipitates above ph 9 [34]. figure 5b shows the surface active molecule, 3-(4-heptylbenzoyl)-benzoic acid, apparently forming a negatively-charged octamer above ph 7, as indicated by the slope of +8 in the plot [35]. the authors proposed that anionic micelles formed, with a critical micelle concentration of about 0.3 mm. figure 5c shows ciprofloxacin cation appearing to form a pentameric aggregate below ph 5 (slope = -5), overlapping with the formation of a phosphate salt at lower ph [36]. (the possibility of the formation of a soluble drug-phosphate complex near ph 5 has not been explored.) most often, when such sub-micellar charged aggregates form, the distortion in the log s-ph curve is most prominent right below the phmax for acids and right above the phmax for bases, as indicated by the examples in figure 5. admet & dmpk 7(2) (2019) 88-105 perspectives in solubility measurements and interpretation doi: 10.5599/admet.686 95 figure 5. examples of charged-species aggregation. (a) brequinar appears to form a negatively-charged dimer above ph 7, as indicated by a slope of +2 in the log s-ph plot. sodium-drug salt precipitates above ph 9 [34]. (b) 3-(4-heptyl-benzoyl)-benzoic acid appears to form an 8-monomer aggregate above ph 7, as indicated by the slope of +8 in the plot [35]. (c) ciprofloxacin cation appears to form a pentameric aggregate below ph 6, as well as a phosphate salt for ph < phmax [36]. the possibility of drug-phosphate complexation has not been ruled out. as discussed by gebauer et al. [31], amino acids, particularly glycine, form sub-micellar pre-nucleation clusters. figure 6 displays log s-ph profiles of six amino acids in buffer-free solutions, where ph was modified with hcl/naoh [37-39]. amino acids are highly soluble in water. consequently, the ionic strength reaches nearly 5 m at the low-ph end of the profile in three of the cases. the ph < 3 and ph > 9 electrode readings may be significantly affected. the symmetrically u-shaped dashed curves in the figures were calculated using the simple hh equation, which does not factor in the effects of ionic strength. the solid curves, which are less symmetrical than the hh curves, are the results of regression analysis where effects of ionic strength were factored in [22]. the apparent log s0 values listed in the figures were calculated using only the data in the ph 4-8 interval, with the rest of the points assigned zero weights during the regression analysis. it can be seen that for ph < 3 and ph > 9, the measured points are notably below the solid curves. this is extremely so in the case of glycine in figure 6b. ordinarily, such distortion would suggest that the zwitterionic form of the molecules self-associate, thus raising the solubility [28]. however, it was not possible to fit the glycine data with an aggregation model that contained a single zwitterionic stoichiometry. at least several overlapping oligomers appear to be necessary to rationalize the nearly flat log s-ph profile of polymeric glycine. it may be that glycine forms a high-molecular weight polymer strands in saturated solutions, which persist across a broad swathe of ph. hydrotropic solubilization via cluster formation in 1916, neuberg [40] coined the term ‘hydrotrope’ for compounds that increase the solubility of sparingly-soluble organic substances, by interacting as ‘co-solubilizers.’ usually, high concentrations of hydrotropes are necessary to increase solubility of the sparingly-soluble substance. hydrotropes are thought to form aggregates in solution. typical examples of hydrotropes include urea, nicotinamide, bergström and avdeef admet & dmpk 7(2) (2019) 88-105 96 sodium salicylate, and benzoates with a short alkyl chains, as these molecules can self-associate into small, sub-micellar clusters of 3-4 molecules [41]. a poorly-soluble molecule can be associated with these clusters and become solubilized in the process. figure 7 shows an example of the hydrotropic solubilization of medazepam by salicylate anions [42]. it would be interesting to explore whether phosphate at high concentrations acts as a hydrotrope. figure 6. the solubility-ph profiles of amino acids showing departures from the ideal henderson-hasselbalch shape. the distortion is particularly extreme in the case of glycine. since the data for ph < 4 and ph > 8 were not used in the regression analysis, the listed log s0 values are ‘apparent.’ the actual values may be lower, especially for glycine. figure 7. hydrotropic solubilization. plot of the total concentration of medazepam in a saturated solution as a function of added sodium salicylate. the solubility data of badwan et al. [42] were fit (using pdisol-x) to an equilibrium model consisting of the drug forming a complex with four units of salicylate. ‘cac’ is the abbreviation for critical aggregate concentration. admet & dmpk 7(2) (2019) 88-105 perspectives in solubility measurements and interpretation doi: 10.5599/admet.686 97 supersolubilization and amorphous dispersions parikh and serajuddin [43] have developed a novel method of preparing solid dispersion by using acid-base interaction in water. haloperidol was dissolved in concentrated aqueous solutions of a weak acid (e.g., malic, tartaric, or citric acid). each weak acid was selected such that it would not form salts with the drug. the solubility of haloperidol exceeded 300 mg/g aqueous solution of malic acid. this was an extremely high aqueous solubility for haloperidol, considering that its intrinsic aqueous solubility in water as the free base is only 0.7 μg/ml, and the solubility of its salt forms (e.g., hcl, phosphate, maleate) is only about 1-5 mg/ml. similar solubility enhancements were observed with itraconazole by the investigators. the extraordinary enhancement by acid-base interaction has been called ‘supersolubilization.’ on drying, the drug-containing material is a semisolid and sticky amorphous solid dispersion. free-flowing powders that can be compressed into tablets are obtained by e.g., adsorbing drug solid dispersions with acids onto porous silica [44]. cocrystals enhancement of solubility xanthines (e.g., caffeine, theophylline, theobromine) are known to form complexes with many pharmaceutical compounds. in a series of papers (1952-1954) higuchi, zuck and lach [45] studied the xanthine molecular associations formed with simple organic molecules in saturated solutions. weak caffeine-containing complexes were determined to form, in the decreasing order of stability: phydroxybenzoic acid > benzocaine > butylparaben > p-aminobenzoic acid > salicylic acid > benzoic acid > aspirin > o-phthalic acid > benzoate > sulfathiazole > picric acid > sulfadiazine > suberic acid. such complexes raised the solubility of the xanthine. these classic studies may bear relevance to the selection of ‘coformers’ in cocrystal formation. a recent and still evolving formulation strategy to increase intestinal absorption is based on using cocrystals [46]. cocrystals are materials composed of two or more different uncharged molecules (e.g., drug and ‘coformer’) within the same crystal lattice, which are associated by nonionic and noncovalent bonds. the altered solid state properties in cocrystals are generally expected to elevate the drug concentration in a similar way that drug salts do, as the drug is released in a dissolution process. the initially elevated concentration of the drug could lead to increased oral absorption of the drug. the analyses of solubility of salts and cocrystals are similar, since both are multi-component solids. drug salts are characterized by a solubility product, ksp salt = [drug][counterion], where the drug is charged. cocrystals are similarly characterized by an equilibrium solubility product, ksp cc = [drug][coformer], but in contrast to salts, both the drug and the coformer are uncharged in the cocrystal. a method to predict ksp cc and the solubility enhancement of cocrystals, using an approach based on measured drug and coformer intrinsic solubility (s0 drug , s0 cof ), combined with in silico h-bond descriptors, has been recently described [47,48]. the latter publications arose out of the iapc-6 special session on cocrystal solubility. figure 8 shows an example of a solubility profile as a function of ph for the indomethacin:saccharin cocrystal [49]. the points represent the measured total concentrations of the coformer saccharin (blue squares) and the drug indomethacin (red circles) under equilibrium conditions poised at 72-96 h, where both the cocrystal and the free-acid crystalline drug are indefinitely stable co-precipitates below ph 3.7. as the cocrystal partially dissolves below ph 3.7, it releases the solvated free drug at the concentration indicated by the scc thick dashed line. this is the initial supersaturated concentration of the drug. it may not persist long, but while the drug is in the elevated concentration, its intestinal absorption may be significantly enhanced. the henderson-hasselbalch curve for indomethacin is indicated by the dashed red curve in the bottom of the drawing. the equilibrium concentration of indomethacin (solid red curve) is bergström and avdeef admet & dmpk 7(2) (2019) 88-105 98 above the dashed curve, with the difference peaking at about ph 3.7. this can be explained by the complexation of uncharged drug by the anionic saccharin. the effect diminishes above ph 5, since the concentration of the free drug diminishes due to ionization [47]. figure 8. plot of the total equilibrium concentrations of saccharin (upper solid line, blue), indomethacin (lower solid line, red), and of the maximum possible concentration of the solubilized indomethacin released from the cocrystal (thick dashed curve, scc), as a function of ph. scc is termed the ‘cocrystal solubility.’ the elevation of the scc dashed curve above that of the indomethacin equilibrium solid red curve is viewed as the cocrystal ‘solubility advantage.’ the solubility product is no longer satisfied above ph 3.7, so the cocrystal no longer forms above that ph. data quality a comprehensive study of equilibrium solubility data quality based on the review of over 800 publications was reported in 2015 [50]. since then, more studies have been examined and a more robust estimate of the true interlaboratory reproducibility can be offered here. improved estimates [4,50] of interlaboratory reproducibility are better revealed when solubility data are normalized for ph (to produce intrinsic solubility, s0, derived from water solubility, sw, or multiple-ph measurements) and temperature (by transforming measurements performed in the range 10-50 °c to the standard value of 25 °c [23]). by way of an addendum to the 2015 study, the data-quality study now has 6355 intrinsic solubility entries in the wiki-ps0 database (in-adme research), for 3014 different pharmaceutically-relevant molecules (solids at room temperature), drawing on the review of 1325 publications. of all the entries, 2144 are singletons, generally reported as sw, but transformed to s0, and temperature adjusted (if needed) to 25 °c. these single-source results offer no direct indication of interlaboratory reproducibility. that indication comes from the 870 molecules for which solubility was reported from at least two different sources. some molecules have been studied in surprisingly large number of different laboratories. for example, there were 34 different reports of the solubility of diclofenac found to date. seventeen of these were measured at several different ph values, and are displayed in figure 9. the next most-frequently studied molecules appear to be phenytoin, barbital, and ketoprofen with 30, 26, and 24 interlaboratory determinations, respectively. the average interlaboratory reproducibility, based on the curated 870 replicated studies, has been determined to be 0.17 log unit, significantly lower than the interlaboratory reproducibility (~0.7 log unit) suggested in past studies [25]. many factors can lead to the perception of poor quality of data. comparing “raw” water solubilities (sw) determined in unbuffered solutions can be problematic. for example, using inaccurate values of pka to interpret the log s-ph data can lead to erroneous intrinsic solubility values. ambient levels of carbon dioxide can significantly affect measured sw values in the cases of sparingly soluble bases. so can unsuspected buffer-drug interactions, aggregate/micelle formations, presence of amorphous solids, supersaturated solutions, differences in the measurement temperature, and other experimental conditions. when an effort is made to critically factor in the possible sources of systematic error, the interlaboratory reproducibility is actually quite good. still, admet & dmpk 7(2) (2019) 88-105 perspectives in solubility measurements and interpretation doi: 10.5599/admet.686 99 for some drug molecules, the intrinsic solubility is very uncertain, with standard deviation (sd) values exceeding 0.5 log unit. table 1. averaged intrinsic solubility of 32 drugs at 25 °c with the poorest interlaboratory reproducibility compound log s0 25 (m) sd a n b compound log s0 25 (m) sd n clofazimine -9.05 0.93 5 amiodarone -10.41 0.59 4 telmisartan -6.73 0.84 5 saquinavir -5.92 0.58 3 buprenorphine -6.07 0.83 3 quinine -3.06 0.57 7 mifepristone -5.22 0.75 4 diflunisal -4.99 0.56 11 tamoxifen -7.52 0.72 7 raloxifene -6.82 0.56 6 terfenadine -7.74 0.71 11 diphenhydramine -3.21 0.55 4 sulfadimethoxine -3.74 0.70 3 etoxadrol -1.96 0.55 3 curcumin -5.36 0.68 3 didanosine -1.24 0.54 3 rifabutin -4.09 0.66 3 danazol -6.10 0.52 10 iopanoic_acid -5.49 0.66 3 ezetimibe -4.94 0.51 4 pioglitazone -6.20 0.66 4 chlorprothixene -5.99 0.51 6 amodiaquine -5.49 0.65 3 bromocriptine -5.50 0.51 5 fentiazac -5.84 0.65 4 miconazole -5.82 0.50 6 itraconazole -8.98 0.61 3 omeprazole -3.70 0.50 3 procaine -2.30 0.60 3 amantadine -2.19 0.50 3 bisoprolol -2.09 0.59 3 thiabendazole -3.97 0.50 4 a calculated interlaboratory standard deviation. b number of reported studies from different laboratories. in addressing data quality, it needs to be emphasized that standard conditions are essential: intrinsic solubility values are to be compared at the designated reference temperature of 25 °c. mixing different types of solubility and not normalizing for temperature increase the apparent variance between different laboratory results, as noted in the haloperidol example (cf., fig. 1). although the realistic average interlaboratory reproducibility is relatively good, some low-soluble molecules are still problematic. table 1 lists the 32 compounds with the poorest reproducibility, drawing on sources where at least three different studies were reported. at the top of the list is clofazimine. based on five reported values, the standard deviation is 0.93 log unit. similarly, even though terfenadine solubility has been reported in eleven studies, its interlaboratory reproducibility at 0.71 is still high. still, 32 out of 3014 molecules is a small fraction. the rest of the molecules are much better determined. absent from the list in table 1 is diclofenac, because its intrinsic solubility has a tight consensus value. its averaged log s0 25 = -5.34 ±0.18, based on 34 published studies. the standard deviation is almost identical to the estimated interlaboratory reproducibility of 0.17. thus, the diclofenac data from different laboratories offer an opportunity to examine issues that affect the quality of solubility measurements as a whole. that is, many of the possible challenges in the interpretation of data in the assessment of interlaboratory reproducibility can be revealed in the examination of diclofenac. it is evident that several different patterns of distortion occur in the diclofenac log s-ph profiles in figure 9 which cannot be predicted by the henderson-hasselbalch equation. one case, in figure 9g, shows distortion similar to those shown in figure 5, suggestive of charge-species aggregation reaction. the data can be rationalized by the formation of the hexameric diclofenac anion complex associated with three protonated 1-(2-hydroxyethyl)pyrrolidine cations. the points with the highest solubility (above the solid curve) appear to be indicative of supersaturation, and could not be explained by a simple equilibrium model. distortions involving the free acid are evident in figures 9f-o, which show elevated solubility in low-ph region, above that indicated by the dashed hh-calculated curve. these may indicate amorphous phase solubility (cf., figure 1a,d). in some cases, the effect may be due to the limit of detection in the analytical technique used to measure the concentration of the practically-insoluble drug. bergström and avdeef admet & dmpk 7(2) (2019) 88-105 100 different solubility methods were used to collect the data in the figures. figure 9a is based on hts data (μsol method) – there are many points in the plot, but they show considerable scatter. the rest of the examples are based on shake-flask measurement. also, the temperature is not the same in the figure. two of the studies were done at 22 °c, four at 23 °c (‘room temperature’), ten at 25 °c, and one at 37 °c. figure 9. the diclofenac log s-ph measurements from 17 publications [28,51-64] reveal a plethora of effects that might lead to poor reproducibility in the determination of the intrinsic solubility. see text for details. admet & dmpk 7(2) (2019) 88-105 perspectives in solubility measurements and interpretation doi: 10.5599/admet.686 101 the rigorous regression analysis of the data in the figures, taking into account all relevant reactions in the equilibrium model [22], produced the intrinsic solubility values indicated in each frame of figure 9. when these values are further transformed to a common reference temperature of 25 °c [23], the resultant mean value of the 17 cases turns out to be log s0 25 = -5.30 ±0.17 (1.6 ± 0.6 μg/ml). this mean is essentially the same as the one obtained from the averaging all 34 cited values, as noted above. the above standard deviation of the mean is identical to the interlaboratory reproducibility obtained from the analysis of 870 different molecules, and is quite a bit lower than the frequently cited uncertainty of ~0.7 log unit in solubility measurement, when comparing results from different laboratories [25]. how does such an impression of poor reproducibility in solubility measurement arise? as mentioned earlier, it is not uncommon to assume that the solubility 2 ph units below the pka (for acids) or above the pka (for bases) can be equated with the intrinsic solubility. clearly, such an approximation would not be accurate in cases of the distortions found in figures 9f-o (and those in fig. 1a,d). the consequences of the above approximation can be illustrated by selecting the lowest solubility in each of the 17 cases and calculating the average: log s0 25 (approx.) = -5.01 ±0.59 (6.5 ± 8.5 μg/ml). the increased apparent standard deviation, 0.59, seems to be consistent with the ~0.7 value. prediction of intrinsic solubility many computational methods for predicting solubility have been described. these serve the important need to predict solubility risks during structural design in the early stages of drug discovery. the earlier methods used relatively small training sets (<1000 compounds), consisting of molecules (including liquids and gases) such as alcohols, pesticides, and herbicides. this limited the accuracy of the prediction of the solubility of drug solids. as more drug data became available for training the prediction models, methods improved. the ‘solubility challenge’ posed by llinàs et al. [65,66] spurred fresh discussion about prediction efficacy and the quality of available data. the meta-analysis described in the preceding section examined factors related to data quality [50]. the study suggested two ways to improve quality of legacy measurement of equilibrium solubility, as discussed above. methods based on shake-flask measurement as a function of ph (‘gold standard’), and two potentiometric methods (psol & cheqsol) can be recommended [4]. predictions of solubility in simulated and aspirated intestinal fluids are exciting new topics in current research [67]. summary and conclusions this review, albeit brief, points towards the complexity in understanding and predicting solubilityrelated processes. these include dissolution, solubilization/aggregation, ph-, counterionand buffer strength effects on solubility, and the impact of salt and cocrystal formation on dissolution, solubility and supersaturation. the debate around how consistent data we can achieve from solubility assays and to what extent the data can be used to model and predict, e.g., ph-dependence or solubilization tendencies in different solvents will continue. a platform for such a vibrant discussion which will spur new ideas is the upcoming iapc-8 meeting where these topics will be further explored. conflict of interest: none. bergström and avdeef admet & dmpk 7(2) (2019) 88-105 102 references [1] d. hörter, j.b. dressman. influence of physicochemical properties on dissolution of drugs in the gastrointestinal tract. adv. drug deliv. rev. 25 (1997) 3-14. [2] n.g. solankin, k. lam, m. tahsin, s.g. gumaste, a.v. shah, a.t.m. serajuddin. effects of surfactants on itraconazole-hpmcas solid dispersion prepared by hot melt extrusion. i: miscibility and drug release. j. pharm. sci. (2018) in press. doi: 10.1016/j.xphs.2018.10.058. [3] c.a.s. bergström, r. holm, s.a. jørgensen, s.b.e. andersson, p. artursson, s. beato, a. borde, k. box, m. brewster, j. dressman, k.-i. feng, g. halbert, e. kostewicz, m. mcallister, u. muenster, j. thinnes, r. taylor, a. mullertz. early pharmaceutical profiling to predict oral drug absorption: current status and unmet needs. eur. j. pharm. sci. 57 (2014) 173-199. [4] a. avdeef, e. fuguet, a. llinàs, c. ràfols, e. bosch, g. völgyi, t. verbić, e. boldyreva, k. takács-novák. equilibrium solubility measurement of ionizable drugs – consensus recommendations for improving data quality. admet & dmpk 4 (2016) 117-178. [5] c.a. lipinski, f. lombardo, b.w. dominy, p.j. feeney. experimental and computational approaches to estimate solubility and permeability in drug discovery and development settings. adv. drug delivery rev. 23 (1997) 3-25. [6] c. saal, a.c. petereit. optimizing solubility: kinetic versus thermodynamic solubility temptations and risks. eur. j. pharm. sci. 47 (2012) 589-595. [7] n.m. ahmad. solubility-driven lead optimization: recent examples and personal perspectives. bioorg. med. chem. lett. 26 (2016) 2975-2979. [8] l. di, p.v. fish, t. mano. bridging solubility between drug discovery and development. drug disc. today 17 (2012) 486-495. [9] t. sou, c.a.s. bergström. automated assays for thermodynamic (equilibrium) solubility determination. drug discov. today: technol. 2 (2018) 11-19. [10] á. könczöl, g. dargó. brief overview of solubility methods: recent trends in equilibrium solubility measurement and predictive models. drug discov. today: technol. 27 (2018) 3-10. [11] c.a.s bergström, w.n. charman, c.j.h. porter. computational prediction of formulation strategies for beyond-rule-of-5 compounds. adv. drug deliv. rev. 101 (2016) 6-21. [12] c.a. bergström, c.m. wassvik, k. johansson, i. hubatsch. poorly soluble marketed drugs display solvation limited solubility. j. med. chem. 50 (2007) 5858-5862. [13] a.p. hill, r.j. young. getting physical in drug discovery: a contemporary perspective on solubility and hydrophobicity. drug disc. today 15 (2010) 648-655. [14] d. riethorst, j. brouwers, j. motmans, p. augustijns. human intestinal fluid factors affecting intestinal drug permeation in vitro. eur. j. pharm. sci. 121 (2018) 338-346. [15] e. galia, e. nicolaides, d. horter, r. lobenberg, c. reppas, j.b. dressman. evaluation of various dissolution media for predicting in vivo performance of class i and ii drugs. pharm. res. 15 (1998) 698-705. [16] i. khadra, z. zhou, c. dunn, c.g. wilson, g. halbert. statistical investigation of simulated intestinal fluid composition on the equilibrium solubility of biopharmaceutics classification system class ii drugs. eur. j. pharm. sci. 67 (2015) 65-75. [17] g. zografi, i. zarenda. the surface activity of phenothiazine derivatives at the air-solution interface. biochem. pharmacol. 15 (1966) 591-598. [18] h.l. lee, j.m. vasoya, m de l. cirqueira, k.l. yeh, t. lee, a.t.m. serajuddin. continuous preparation of 1:1 haloperidol-maleic acid salt by a novel solvent-free method using a twin screw melt extruder. mol. pharmaceutics 14 (2017) 1278-1291. [19] s. li, s. wong, s. sethia, h. almoazen, y.m. joshi, a.t.m. serajuddin. investigation of solubility and dissolution of a free base and two different salt forms as a function of ph. pharm. res. 22 (2005) 628635. admet & dmpk 7(2) (2019) 88-105 perspectives in solubility measurements and interpretation doi: 10.5599/admet.686 103 [20] m.m. al omari, m.b. zughul, j.e.d. davies, a.a. badwan. a study of haloperidol inclusion complexes with β-cyclodextrin using phase solubility, nmr spectroscopy and molecular modeling techniques. j. solution. chem. 38 (2009) 669-683. [21] c.a.s. bergström, k. luthman, p. artursson. accuracy of calculated ph-dependent aqueous drug solubility. eur. j. pharm. sci. 22 (2004) 387-398. [22] g. völgyi, a. marosi, k. takács-novák, a. avdeef. salt solubility products of diprenorphine hydrochloride, codeine and lidocaine hydrochlorides and phosphates – novel method of data analysis not dependent on explicit solubility equations. admet & dmpk 1 (2013) 48-62. [23] a. avdeef. solubility temperature dependence predicted from 2d structure. admet & dmpk 3 (2015) 298-344. [24] a. avdeef. unpublished results (2018). [25] d.s. palmer, j.b.o. mitchell. is experimental data quality the limiting factor in predicting the aqueous solubility of druglike molecules?. mol. pharmaceutics 11 (2014) 2962-2972. [26] s. veesler, l. lafferrère, e. garcia, c. hoff. phase transitions in supersaturated drug solution. org. proc. res. dev. 7 (2003) 983-989. [27] p.e. bonnett, k.j. carpenter, s. dawson, r.j. davey. solution crystallization via a submerged liquid– liquid phase boundary: oiling out. chem. commun. (2003) 698–699. [28] a. avdeef. absorption and drug development second edition, wiley-interscience, hoboken nj, 2012. [29] m.m. al omari, m.b. zughul, j.e.d. davies, a.a. badwan. effect of buffer species on the complexation of basic drug terfenadine with β-cyclodextrin. j. inclus. phenom. macrocycl. chem. 58 (2007) 227-235. [30] a.s. indulkar, k.j. box, r. taylor, r. ruiz, l.s. taylor. ph-dependent liquid−liquid phase separation of highly supersaturated solutions of weakly basic drugs. mol. pharmaceutics 12 (2015) 2365-2377. [31] d. gebauer, m. kellermeier, j.d. gale, l. bergström, h. cölfen. pre-nucleation clusters as solute precursors in crystallization. chem. soc. rev., 43 (2014) 2348-2371. [32] a. pobudkowska, c. ràfols, x. subirats, e. bosch, a. avdeef. phenothiazines solution complexity – determination of pka and solubility-ph profiles exhibiting sub-micellar aggregation at 25 and 37 °c. eur. j. pharm. sci. 93 (2016) 163-176. [33] k. edueng, d. mahlin, p. larsson, c.a.s. bergström. mechanism-based selection of stabilisation strategy for amorphous formulations: insights into crystallization pathways. j. control. rel. 256 (2017) 193-202. [34] s.-y. p. king, a.m. basista, g. oros. self-association and solubility behaviors of a novel anticancer agent, brequinar sodium. j. pharm. sci. 78 (1989) 95-100. [35] n.k. pandit, j.m. strykowski. effect of temperature on the solubility of a surface active carboxylic acid. j. pharm. sci. 78 (1989) 767-770. [36] d.l. ross, c.m. riley. aqueous solubilities of some variously substituted quinolone antimicrobials. int. j. pharm. 63 (1990) 237-250. [37] h.-c. tseng, c.-y. lee, w.-l. weng, i.-m. shiah. solubilities of amino acids in water at various ph values under 298.15 k. fluid phase equilib. 285 (2009) 90-95. [38] x. yang, x. wang, c.b. ching. solubility of form alpha and form gamma of glycine in aqueous solutions. j. chem. eng. data 53 (2008) 1133-1137. [39] t.e. jr. needham, a.n. paruta, r.j. gerraughty. solubility of amino acids in pure solvent systems. j. pharm. sci. 60 (1971) 565-567. [40] c. neuberg. hydrotropic phenomena. biochem. z. 76 (1916) 107-176. [41] w. kunz, k. holmberg, t. zemb. hydrotropes. curr. opin. coll. interf. sci. 22 (2016) 99-107. [42] a.a. badwan, l.k. el-khordagui, a.m. saleh, s.a. khalil. the solubility of benzodiazepines in sodium salicylate solution and a proposed mechanism for hydrotropic solubilization. int. j. pharm. 13 (1983) 67-74. bergström and avdeef admet & dmpk 7(2) (2019) 88-105 104 [43] t. parikh, a.t.m. serajuddin. development of fast-dissolving amorphous solid dispersion of itraconazole by melt extrusion of its mixture with weak organic carboxylic acid and polymer. pharm. res. 35 (2018) 127. in press. doi: 10.1007/s11095-018-2407-4. [44] a. shah, a.t.m. serajuddin. conversion of solid dispersion prepared by acid–base interaction into free-flowing and tabletable powder by using neusilin® us2. int. j. pharm. 484 (2015) 172-180. [45] t. higuchi, j.l. lach. investigation of some complexes formed in solution by caffeine vi. comparison of complexing behaviors of methylated xanthines with p-aminobenzoic acid, salicylic acid, acetylsalicylic acid, and p-hydroxybenzoic acid. j. amer. pharm. assoc. 43 (1954) 527-530. [46] s.l. childs, n. rodríguez-hornedo, l.s. reddy, a. jayasankar, c. maheshwari, l. mccausland, r. shipplett, b.c. stahly. screening strategies based on solubility and solution composition generate pharmaceutically acceptable cocrystals of carbamazepine. cryst. eng. comm. 10 (2008) 856-864. [47] a. avdeef. cocrystal solubility product analysis – dual concentration-ph mass action model not dependent on explicit solubility equations. eur. j. pharm. sci. 110 (2017) 2-18. [48] a. avdeef. cocrystal solubility product prediction using an in combo model and simulations to improve design of experiments. pharm. res. (2018) 35:40. https://doi.org/10.1007/s11095-018-23433. [49] a. alhalaweh, l. roy, n. rodríguez-hornedo, s.p. velaga. ph-dependent solubility of indomethacinsaccharin and carbamazepine-saccharin cocrystals in aqueous media. mol. pharmaceutics 9 (2012) 2605-2612. [50] a. avdeef. suggested improvements for measurement of equilibrium solubility-ph of ionizable drugs. admet & dmpk 3 (2015) 84-109. [51] b. bard, s. martel, p.-a. carrupt. high throughput uv method for the estimation of thermodynamic solubility and the determination of the solubility in biorelevant media. eur. j. pharm. sci. 33 (2008) 230-240. [52] a. chiarini, a. tartarini, a. fini. ph-solubility relationship and partition coefficients for some antiinflammatory arylaliphatic acids. arch. pharm. (weinheim) 317 (1984) 268-273. [53] a. avdeef, c.m. berger, c. brownell. ph-metric solubility. 2. correlation between the acid-base titration and the saturation shake-flask solubility-ph methods. pharm. res. 17 (2000) 85-89. [54] a. fini, g. fazio, g. feroci, solubility and solubilization properties of non-steroidal antiinflammatory drugs. int. j. pharm. 126 (1995) 95-102. [55] m.t. ledwidge, o.i. corrigan. effects of surface active characteristics and solid state forms on the ph solubility profiles of drug-salt systems. int. j. pharm. 174 (1998) 187-200. [56] p. luger, k. daneck, w. engel, g. trummlitz. k. wagner. structure and physicochemical properties of meloxicam, a new nsaid. eur. j. pharm. sci. 4 (1996) 175-187. [57] l. zhou, l. yang, s. tilton, j. wang. development of a high throughput equilibrium solubility assay using miniaturized shake-flask method in early drug discovery. j. pharm. sci. 96 (2007) 3052-3071. [58] m.l. manca, m. zaru, g. ennas, d. valenti, c. sinico, g. loy, a.m. fadda. diclofenac-beta-cyclodextrin binary systems: physicochemical characterization and in vitro dissolution and diffusion studies. aaps pharmscitech 6 (2005) e464-e472. [59] b. chuasuwan, v. binjesoh, j.e. polli, h. zhang, g.l. amidon, h.e. junginger, k.k. midha, v.p. shah, s. stavchansky, j.b. dressman, d.m. barends. biowaiver monographs for immediate release solid oral dosage forms: diclofenac sodium and diclofenac potassium. j. pharm. sci. 98 (2009) 1206-1219. [60] c.d. herzfeldt, r. kuemmel. dissociation constants, solubilities and dissolution rates of some selected nonsteroidal antiinflammatories. drug dev. ind. pharm. 9 (1983) 767-793. [61] y. maitani, m. nakagaki, t. nagai. determination of the acid dissociation constants in ethanol-water mixtures and partition coefficients for diclofenac. int. j. pharm. 74 (1991) 105-116. [62] a. pobudkowska, u. domańska, study of ph-dependent drugs solubility in water. chem. ind. chem. eng. quart. 20 (2014) 115-126. admet & dmpk 7(2) (2019) 88-105 perspectives in solubility measurements and interpretation doi: 10.5599/admet.686 105 [63] w. riess, h. stierlin, p. degen, j.w. faigle, a. gerardin, j. moppert, w. theobald, j. wagner. pharmacokinetics and metabolism of the anti-inflammatory agent voltaren. scand. j. rheumato. suppl. 22 (1978) 17-29. [64] g. völgyi, d. csicsák, k. takács-novák. right filter-selection for phase separation in equilibrium solubility measurement. eur. j. pharm. sci. 123 (2018) 98-105. [65] a. llinàs, r.c. glen, j.m. goodman, solubility challenge: can you predict solubilities of 32 molecules using a database of 100 reliable measurements?. j. chem. inf. model. 48 (2008) 1289-1303. [66] a.j. hopfinger, e.x. esposito, a. llinàs, r.c. glen, j.m. goodman. findings of the challenge to predict aqueous solubility. j. chem. inf. model. 49 (2009) 1-5. [67] j.h. fagerberg, e. karlsson, j. ulander, g. hanisch, c.a.s. bergström. computational prediction of drug solubility in fasted simulated and aspirated human intestinal fluid. pharm. res. 32 (2015) 578589. ©2019 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ revisiting the application of immobilized artificial membrane (iam) chromatography to estimate in vivo distribution properties of drug discovery compounds based on the model of marketed drugs doi: http://dx.doi.org/10.5599/admet.757 78 admet & dmpk 8(1) (2020) 78-97; doi: http://dx.doi.org/10.5599/admet.757 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper revisiting the application of immobilized artificial membrane (iam) chromatography to estimate in vivo distribution properties of drug discovery compounds based on the model of marketed drugs klara valko*, silvia rava 1 , shenaz bunally 2 and scott anderson 3 bio-mimetic chromatography ltd. business and technology centre, bessemer drive, stevenage, sg1 2dx, united kingdom 1 university of pavia, italy, erasmus internship at bio-mimetic chromatography ltd. business and technology centre, bessemer drive, stevenage, sg1 2dx, united kingdom 2 physicochemical group, gsk, gunnels wood road, stevenage, sg1 2ny united kingdom 3 regis technologies inc. 8210 austin avenue morton grove, il us 60053 *corresponding author: e-mail: klara_valko@bio-mimetic-chromatography.com; received: november 27, 2019; revised: january 06, 2020; available online: january 31, 2020 abstract immobilized artificial membrane (iam) chromatography columns have been used to model the in vivo distribution of drug discovery compounds. regis technologies inc., the manufacturer, had to replace the silica support and consequently introduced a new iam.pc.dd2 column that shows slightly different selectivity towards acidic and basic compounds. the application of the new iam.pc.dd2 columns has been evaluated and the in vivo distribution models have been compared with the previous batches of columns. it was found that due to the improved endcapping of the silica, some of the positively charged drug molecules showed shorter retention than previously published. therefore, the column system suitability data have been updated. however, these differences do not significantly affect the previously published models for the volume of distribution, brain tissue binding and drug efficiency. therefore, the published models can be used with the new iam.pc.dd2 columns. ©2020 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords phospholipid binding, hplc, volume of distribution, tissue binding, drug efficiency. introduction pidgeon and venkataran [1] patented a method for immobilizing phospholipids on hplc-grade silica stationary phases. the immobilized phospholipids mimic the lipid environment of a fluid cell membrane on a solid matrix. the retention factors (k) of compounds obtained on the iam stationary phase are proportional to their affinity (partition coefficient) (k) to phospholipids according to the equation (1): log k(iam) = log k(iam) + log (vs/vm) (1) where k is the retention factor obtained from the retention time (tr) and the dead time (t0) according to http://dx.doi.org/10.5599/admet.757 http://dx.doi.org/10.5599/admet.757 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:klara_valko@bio-mimetic-chromatography.com http://creativecommons.org/licenses/by/4.0/ admet & dmpk 8(1) (2020) 78-97 revisiting the iam chromatography doi: http://dx.doi.org/10.5599/admet.757 79 equation (2) k = (tr-t0)/t0 (2) and vs/vm is the volume ratio of the stationary and mobile phases, respectively. many drug molecules bind strongly to phospholipids as it is a pre-requisite for membrane permeability. to reduce the retention times various concentrations of acetonitrile can be used in the mobile phase. there is a linear relationship between the log k values and the percentage of acetonitrile in the mobile phase that allows the extrapolation of the retention factor to the zero percentage of acetonitrile, i.e. to the pure aqueous buffer (log k0) as shown by equation 3. log k = slope ∙ acn% + log k0 (3) equation 3 enables the calculation of the acetonitrile concentration that is necessary to achieve log k = 0, when the retention time is exactly double that of the dead time, which means an equal distribution of the compound in the mobile and the iam stationary phase [2]. this acetonitrile concentration expressed in volume percentage is called the chromatographic hydrophobicity index (chi (iam)) [3]. it has also been shown that the chi (iam) values have linear relationships with the gradient retention times of the compounds obtained on the iam stationary phases. by using an acetonitrile gradient of 0 % to a maximum of 85 % the obtained gradient retention times can be calibrated from the chi (iam) values by using a set of standards, eliminating the need to carry out isocratic measurements. the so obtained chi (iam) values show a good correlation to the log k0 (iam) values, that are the extrapolated logarithmic retention factors to 0 % acetonitrile concentration in the isocratic mode. the chromatographic determination of phospholipid-binding is based on measuring gradient retention times. it is independent of the amount of compound injected onto the column, so there is no need for quantitative analysis. the time measurement is very reproducible, and the calibrated (relative) gradient retention data are suitable for inter-laboratory comparison and application in published models of in vivo drug distribution. the data so obtained can be compiled in databases and is suitable for use in establishing quantitative structure retention relationships, thus enabling chemists to design compounds with the appropriate iam binding. the application of iam chromatography as an aid in drug discovery has been reviewed extensively in recent years [4-8]. the molecular factors that influence iam retention have also been discussed [9,10]. it has been found that iam columns retain strongly positively charged compounds as the iam columns are negatively charged on the surface similar to cell membranes [11-13]. several applications of iam chromatography have also published [14] and this is summarized below. intestinal absorption and drug distribution depend on a compound’s partitioning into phospholipids. compounds have to possess a certain degree of affinity to membranes in order to permeate through the biological phospholipid bilayers. therefore, iam chromatography can provide an insight into the potential intestinal absorption of compounds [15]. as tissues represent a more non-polar lipid environment relative to the plasma, the human volume of distribution can be modelled by the binding differences between iam and human serum albumin (hsa) [13,16], as shown by equation (4): log vdss= 0.44 log k (iam) – 0.22 log k (hsa) – 0.66 (4) the sum of the albumin and phospholipid-binding inversely correlates to the unbound fraction of the compounds in tissues according to equation (5) [17]. log fut = -0.66 log k (hsa) – 0.52 log k (iam) + 0.55 (5) http://dx.doi.org/10.5599/admet.757 klara valko et al. admet & dmpk 8(1) (2020) 78-97 80 the in vivo drug efficiency can be modelled by the sum of the iam and hsa binding of compounds [18] according to equation 6. the in vitro drugeff shows an excellent correlation with the in vivo drugeff. drug efficiency and potency enables early dose estimation as drug efficiency approximates to the proportion of free drug concentration and dose [19]. log drugeff= 2 (0.23 log k (hsa) + 0.43 log k (iam) 0.72) (6) it has been found that basic compounds that have chi (iam) value greater than 50 have phospholipidosis potential [20] and also show promiscuous binding, interacting with several targets, causing potential toxicity and side effects [21]. at present regis technologies inc. (morton grove, il, usa) is the only manufacturer of the immobilized artificial membrane (iam) stationary phase. the stationary phase contains only a monolayer of the phosphatidylcholine (pc) with the aliphatic part chemically bonded to the silica while the choline polar head group faces towards the mobile phase. initially, “type a” silica was used in the bonding procedure. the pc molecule has only one alkyl chain available with the second alkyl chain directly bonded to the silica surface. in this way, a denser bonding could be achieved which was closer to the natural density of pc molecules on the cell surface. however, in 2018 the company had to change the base silica material to a newer “type b” as the old supply was no longer available. “type b” silica has been chosen as the closest in properties to “type a” regarding the specific surface area, pore size and particle size. when testing the retention selectivity of the new iam.pc.dd2 columns which were otherwise manufactured in the same way, differences were observed towards acidic and basic molecules. this was somewhat surprising as the new silica support had identical physical specifications to the previously used silica (10 µm particle size, 300 å pore size and nearly identical surface area). however, given the newer silica is of a much higher purity with only single ppm metal content, these observations are understandable as the new silica provides a much more inert surface and likely contributes to less non-specific binding interactions. the difference in selectivity can also be attributed to the different arrangements of the positive and negative charges of the choline headgroup on the stationary phase surface, again somewhat affected by the higher purity silica during the bonding process. the new batches of columns did not pass the original system suitability test criteria. hence the aim was to establish new chi (iam) values for the system suitability test compounds that could be reproduced on the new batches of iam columns. in this paper, the investigation of the selectivity difference is reported using over 70 marketed drug molecules and the in vivo distribution models are compared for the old and the new batches of iam.pc.dd2 stationary phases. the estimated volume of distribution data was compared with the data from the previously published models. the differences of the models for the estimated brain tissue binding and drug efficiency values of the known drug molecules are investigated and compared with the published data obtained using the previous batches of iam.pc.dd2 columns. experimental all high-performance liquid chromatography measurements were performed on an agilent 1100 series hplc instrument equipped with an ultraviolet diode array detector (dad). measurements of membrane binding using immobilized artificial membrane (iam) chromatography the iam binding was measured using iam.pc.dd2 hplc columns (regis technologies, inc., il, usa) with dimensions of 100 x 4.6 mm, particle size 10 µm with 300 å pore size. two batches of iam.pc.dd2 columns were studied (r20511-014-3 and p20511-014-3) that had been admet & dmpk 8(1) (2020) 78-97 revisiting the iam chromatography doi: http://dx.doi.org/10.5599/admet.757 81 prepared using the new batches of silica support. the selectivity and the models obtained were compared with the published average iam data obtained on the columns manufactured between 2004 and 2012 [14]. mobile phase a was 50 mm ammonium acetate adjusted to ph 7.4 and mobile phase b was 100 % acetonitrile. the mobile phase flow rate was 1.5 ml/min and the run time was 6 min. the acetonitrile gradient was applied from 0 to 85 % from 0 to 4.75 min and kept at 85 % until 5.15 min. from 5.15 to 5.25 min the acetonitrile concentration dropped back to 0 %. the retention times were standardized using the iam calibration mixture, containing octanophenone, heptanophenone, hexanophenone, valerophenone, butyrophenone, propiophenone, acetophenone, acetanilide and paracetamol. the retention times of the compounds were plotted against the isocratically determined and predefined chi (iam) values [4] listed in table 1. the slope and intercept of the obtained straight line was used to calculate the chi iam values of new compounds. the chi iam values were converted to the octanol/water lipophilicity scale and expressed as log k (iam) values using equation (7). log k (iam) = 0.29 exp(0.026 chi (iam) + 0.70 (7) the so-obtained log kiam values express the membrane partition comparable to the octanol/water lipophilicity as described in reference [22]. a typical chromatogram is shown in figure 1. figure 1. a typical chromatogram obtained on iam.pc.dd2 100 x 4.6 mm hplc column. flow rate: 1.5 ml/min; mobile phase a: 50 mm ammonium acetate adjusted to ph 7.4, mobile phaseb: acetonitrile. run time 6 min; gradient: 0 to 4.75 min 0 to 85 % b, 4.75 to 5.15 min 85 % b, 5.15 to 5.25 min 0 % b; detection wavelength: 250 nm the natural state of the phosphatidylcholine headgroup on the iam stationary phase provides the column selectivity that was tested using the system suitability test compounds containing acidic, basic and neutral compounds. protein binding measurements using biomimetic protein stationary phases the interaction of the compounds listed in table 2 with human serum albumin (hsa) has been measured using commercially available chemically bonded hsa columns (chiralpak-hsa from hichrom, uk) with the dimensions of 50 x 3 mm and a particle size of 5 μm. mobile phase a was 50 mm ammonium acetate adjusted to ph 7.4 and mobile phase b was 2-propanol. the flow rate was 1.5 ml/min and the run time was 6 min. the 2-propanol gradient was applied from 0 to http://dx.doi.org/10.5599/admet.757 klara valko et al. admet & dmpk 8(1) (2020) 78-97 82 35 % from 0 to 3.00 min and kept at 35 % until 4.00 min. from 4.00 to 4.25 min the 2-propanol concentration was dropped back to 0 %. although column manufacturers suggest using ph 7 for chiral separations using the hsa column, in this case, the ph 7.4 mobile phase is used to mimic the plasma ph for the plasma protein binding measurements. table 1. the constant chi (iam) and log k (iam) values for the calibration set of compounds with typical retention times. the retention times were standardized using the calibration set of compounds described previously [23]. compounds studied in this study, 72 commercial drugs from diverse therapeutic areas have been analysed. 50 µl of 10 mm dmso solutions of the drugs were diluted down to 150 µl and 10 µl of the diluted solution was injected onto the hplc columns. the clinical volume of distribution data for all of the drugs is available in the literature [13]. table 2 shows the marketed drug molecules and their clinical volume of distribution data obtained from reference [13] and the previously measured chi (iam) data [14] using an iam column manufactured by regis technologies between 2004 and 2012. data analysis the statistical analysis was performed using jmp 13 (sas institute, usa). results and discussion the selectivity and reproducibility of iam.pc.dd2 columns have been extensively studied. first, the chi (iam) values of the system suitability test compounds (see table 2) were scrutinized. the chi (iam) values of the test compounds were then compared with the chi (iam) values obtained on the new batches of iam.pc.dd2 columns. it was found, as shown in table 3, that carbamazepine, warfarin, indomethacin, nicardipine, ketoprofen, haloperidol and budesonide had very similar chi (iam) values. however, imipramine and chlorpromazine showed greater than the 5 chi (iam) unit deviations which would be within the experimental error and therefore they did not pass the system suitability criteria. these compounds are strong bases and positively charged at ph 7.4. most probably the new iam.pc.dd2 columns have less negative charge on the surface and therefore reduce the retention times of positively charged compounds. it is also possible that more of the free silanol groups were end-capped which would contribute to a shorter retention of basic compounds. therefore, the system suitability criteria were changed for basic compounds. the “new” chi (iam) values are listed in table 3. the in vivo distribution models were re-examined using the data from the new iam.pc.dd2 columns. because of the selectivity differences, it was essential to investigate the chi (iam) values for a larger set of compounds and check how the previously described in vivo distribution models performed using the data from the new columns manufactured with the new silica material. the 72 drug molecules from the original set of marketed drugs for which the models were developed [13] were analysed on the new iam.pc.dd2 columns and the hsa column. table 4 shows the chi (iam) values and the log k hsa binding data obtained from reference [13] and remeasured in this study. compound chi (iam) log k (iam) tr 7.4 (min) octanophenone 49.4 4.97 4.37 heptanophenone 45.7 4.30 4.17 hexanophenone 41.8 3.71 3.93 valerophenone 37.3 3.16 3.66 butyrophenone 32.0 2.62 3.32 propiophenone 25.9 2.15 2.90 acetophenone 17.2 1.67 2.34 acetanilide 11.5 1.45 2.04 paracetamol 2.90 1.20 1.58 admet & dmpk 8(1) (2020) 78-97 revisiting the iam chromatography doi: http://dx.doi.org/10.5599/admet.757 83 table 2. the name of the marketed drugs, their cas number, their previously published chi (iam) data and the logarithm of their clinical volume of distribution. compound cas no acid/base character chi (iam) [14] log vdss [13] acecainide 32795-44-1 basic 23.3 0.10 acetanilide 103-84-4 neutral 10.7 -0.13 acetazolamide 59-66-5 neutral 1.7 -0.46 alclofenac 22131-79-9 acidic 17.8 -1.10 amoxapine 14028-44-5 basic 56.5 1.67 bamethan 3703-79-5 basic 18.6 0.01 betamethasone 378-44-9 neutral 31.7 0.18 carbamazepine 298-46-4 neutral 26.5 -0.09 chlorpromazine 50-53-3 basic 61.9 2.15 cinoxacin 28657-80-9 acidic 1.6 -0.42 colchicine 64-86-8 neutral 23.7 0.03 cytarabine 147-94-4 weak base -15.0 -0.31 diazoxide 364-98-7 weak acid 24.7 -0.16 diclofenac 15307-86-5 acidic 33.5 -0.67 diprophylline 479-18-5 neutral -4.0 -0.24 ethinyl estradiol 57-63-6 neutral 46.8 0.39 famotidine 76824-35-6 weak base 15.7 -0.06 felbamate 25451-15-4 neutral 19.1 -0.64 felodipine 72509-76-3 neutral 46.1 0.35 fenoprofen 31879-05-7 acidic 17.2 -1.27 finasteride 98319-26-7 neutral 38.9 0.30 floxacillin 5250-39-5 acidic 23.4 -0.56 flumazenil 78755-81-4 neutral 18.4 -0.04 flurbiprofen 5104-49-4 acidic 26.8 -1.56 furosemide 54-31-9 acidic 21.4 -0.71 gemfibrozil 25812-30-0 acidic 32.9 -1.34 glipizide 29094-61-9 acidic 21.1 -0.73 griseofulvin 126-07-8 neutral 33.1 0.12 hydrochlorothiazide 58-93-5 weak acid 15.9 -0.13 hydrocortisone 50-23-7 neutral 27.9 0.12 imipramine 50-49-7 basic 51.6 1.16 indomethacin 53-86-1 acidic 25.3 -0.96 isradipine 75695-93-1 neutral 40.0 0.13 ketoconazole 65277-42-1 weak base 42.9 0.32 ketoprofen 22071-15-4 acidic 21.9 -1.01 labetalol 36894-69-6 amphoteric 43.9 0.71 lignocaine 137-58-6 basic 32.6 0.40 methylprednisolone 83-43-2 neutral 32.1 0.16 metronidazole 443-48-1 weak base -3.3 -0.25 minoxidil 38304-91-5 weak base 19.0 -0.04 nabumetone 42924-53-8 neutral 38.4 0.03 nadolol 42200-33-9 basic 20.2 -0.33 nicardipine 55985-32-5 weak base 45.9 0.47 nifedipine 21829-25-4 neutral 29.0 -0.04 nitrendipine 39562-70-4 neutral 40.5 0.16 http://dx.doi.org/10.5599/admet.757 klara valko et al. admet & dmpk 8(1) (2020) 78-97 84 papaverine 58-74-2 weak base 34.4 0.00 pentoxifylline 6493-05-6. neutral 12.0 -0.10 perphenazine 58-39-9 basic 56.3 1.35 phenytoin 57-41-0 weak acid 31.6 0.03 pindolol 13523-86-9 basic 42.0 0.85 prazosin 19216-56-9 weak base 31.6 -0.05 prednisolone 50-24-8 neutral 28.0 0.14 prednisone 53-03-2 neutral 25.9 0.05 primidone 125-33-7 neutral 8.9 -0.15 probenecid 57-66-9 acidic 20.1 -0.75 procainamide 614-39-1 basic 19.9 0.02 propanolol 525-66-6 basic 50.8 1.29 propylthiouracil 51-52-5 weak acid 3.9 -0.34 proxyphylline 603-00-9 neutral 1.1 -0.85 sulfachlorpyridazine 80-32-0 acidic 6.0 -0.76 sulfameter 651-06-9 acidic 3.7 -0.61 sulfamethoxypyridazine 80-35-3 weak acid 9.6 -0.59 sulfinpyrazone 57-96-5 acidic 26.1 -0.65 sulfisoxazole 127-69-5 acidic 2.9 -0.72 sulphadimidine 57-68-1 amphoteric 34.6 0.22 sulpiride 15676-16-1 basic 25.5 0.12 tamoxifen 10540-29-1 basic 58.7 1.48 theobromine 83-67-0 weak acid -4.1 -0.25 tolfenamic acid 13710-19-5 acidic 36.6 -0.76 trazodone 19794-93-5 weak base 36.3 0.01 trimethoprim 738-70-5 weak base 20.8 -0.06 warfarin 81-81-2 acidic 19.9 -0.78 table 3. the chi (iam) obtained from the literature [13] (chi (iam) reference), the chi (iam) values obtained on the new iam.pc.dd2 columns and the acid/base character of the iam calibration compounds compound acid/base character chi (iam) chi (iam) “new” carbamazepine neutral 28.9 27.0 colchicine neutral 23.7 23.0 warfarin acidic 19.9 24.0 indomethacin acidic 32.5 30.0 nicardipine weak base 45.9 45.0 propranolol basic 45.1 42.0 imipramine basic 54.1 45.0 ketoprofen strong acid 21.9 20.0 chlorpromazine basic 61.9 54.0 haloperidol basic 44.3 48 budesonide neutral 38.7 38 figure 2 shows the plot of the experimental chi (iam) values on iam.pc.dd2 columns p20511-014-3 and r2055-014-3. the chi (iam) values obtained on the columns are very similar indicating good batch to batch reproducibility of the new columns. admet & dmpk 8(1) (2020) 78-97 revisiting the iam chromatography doi: http://dx.doi.org/10.5599/admet.757 85 table 4. the experimental chi (iam) values of the investigated compounds obtained on the new iam.pc.dd2 columns and the chi (iam) obtained from the literature [13] (reference chi (iam)) compound chi (iam) from ref [13] r20511-014-3 chi (iam) p20511-014-3 chi (iam) log k hsa from ref [13] log k hsa acecainide 23.30 17.72 17.57 0.69 0.52 acetanilide 10.70 11.55 11.50 0.62 0.43 acetazolamide 1.66 0.99 0.74 1.47 1.44 alclofenac 17.80 20.47 21.63 5.11 5.40 amoxapine 56.50 45.41 44.74 2.15 2.70 bamethan 18.60 18.53 18.03 0.49 0.42 betamethasone 31.70 30.18 30.05 1.09 1.39 carbamazepine 26.50 27.18 27.07 1.99 1.81 chlorpromazine 61.90 51.54 50.94 1.18 1.27 cinoxacin 1.60 5.31 6.67 0.79 0.91 colchicine 23.70 23.45 23.18 0.69 0.24 cytarabine -15.00 -6.76 -6.99 1.59 1.87 diazoxide 24.70 23.39 23.44 5.42 5.58 diclofenac 33.50 35.18 35.85 0.63 0.24 diprophylline -4.00 -1.49 -1.98 4.1 4.24 ethinyl estradiol 46.80 45.30 45.23 0.46 0.50 famotidine 15.70 16.82 16.08 1.39 3.43 felbamate 19.07 39.16 39.21 3.46 4.15 felodipine 46.10 44.37 44.45 5.91 6.11 fenoprofen 17.20 27.17 28.26 1.58 2.32 finasteride 38.90 37.15 37.17 3.62 3.55 floxacillin 23.40 23.76 24.59 0.56 0.65 flumazenil 18.40 18.27 18.27 7.81 8.50 flurbiprofen 26.80 30.23 30.89 1.26 4.01 furosemide 21.41 21.96 22.69 3.64 8.50 gemfibrozil 32.90 33.70 34.53 3.32 4.05 glipizide 21.10 22.82 23.22 1.49 1.92 griseofulvin 33.10 31.06 30.98 0.77 0.80 hydrochlorothiazide 15.90 13.04 13.08 0.92 1.05 hydrocortisone 27.90 27.23 27.02 1.95 2.59 imipramine 51.60 45.58 44.81 6.17 5.59 indomethacin 25.30 32.49 33.16 2.86 3.36 isradipine 40.00 39.11 39.16 2.9 3.30 ketoconazole 42.90 39.50 39.26 4.8 5.40 ketoprofen 21.90 22.93 23.89 1.28 1.80 labetalol 43.90 38.29 37.48 0.69 0.52 lignocaine 32.60 24.89 25.53 1.35 1.53 methylprednisolone 32.10 30.91 30.77 0.76 0.28 metronidazole -3.30 -2.71 -3.24 0.66 0.72 minoxidil 19.00 15.93 16.08 3.16 3.41 nabumetone 38.40 35.90 35.96 0.49 2.14 nadolol 20.20 19.89 19.23 2.93 3.55 nicardipine 45.90 43.82 43.73 1.41 1.95 nifedipine 29.00 28.11 28.16 3.13 3.37 http://dx.doi.org/10.5599/admet.757 klara valko et al. admet & dmpk 8(1) (2020) 78-97 86 nitrendipine 40.50 39.06 39.19 2.35 2.70 papaverine 34.40 31.22 31.15 0.63 0.39 pentoxifylline 12.00 14.04 13.87 3.27 4.03 perphenazine 56.30 49.57 48.97 1.6 2.02 phenytoin 31.60 30.18 30.20 0.54 0.63 pindolol 42.00 25.69 25.31 1.81 2.40 prazosin 31.60 29.82 29.35 0.89 0.95 prednisolone 28.00 27.31 27.20 0.8 1.08 prednisone 25.90 25.25 25.13 0.62 0.42 primidone 8.90 10.31 10.49 3.44 4.01 probenecid 20.10 23.45 24.38 0.76 0.49 procainamide 19.90 11.73 11.89 0.73 1.67 propranolol 50.80 42.24 41.46 0.63 1.02 propylthiouracil 3.90 6.56 6.52 2.67 3.21 proxyphylline 1.10 10.05 10.89 1.97 3.02 sulfachlorpyridazine 6.00 10.08 10.83 2.04 2.20 sulfameter 3.70 6.92 7.09 4.09 2.44 sulfamethoxypyridazine 9.60 11.11 11.16 2.33 4.27 sulfinpyrazone 26.10 26.19 26.89 3.17 2.66 sulfisoxazole 2.90 6.48 7.37 1.54 3.68 sulphadimidine 34.55 10.89 11.09 0.75 1.72 sulpiride 25.50 15.80 16.15 4.84 0.71 tamoxifen 58.70 54.23 53.78 1.29 4.82 theobromine -4.10 -3.16 -3.71 4.68 0.27 tolfenamic acid 36.60 40.72 41.53 2.46 6.60 trazodone 36.30 34.09 33.86 0.8 3.06 trimethoprim 20.80 20.99 20.52 4.45 0.95 warfarin 19.90 23.47 24.18 0.7 4.16 figure 3 shows the plot of the chi (iam) from the literature [13], reference chi (iam), with the experimental chi (iam) values using the iam.pc.dd2 column r2055-014-3. the chi (iam) values of procainamide, sulpiride, zolmitriptan and pindolol were outliers from the correlation with lower chi (iam) values on the new batches of iam columns. compounds are considered outliers when the difference between the measured and predicted value is greater than the double of the standard error of the estimate, in this case 7.2 chi iam units. the slope of the regression line (red) is less than one (0.81) and the regression line is different from the line of unity (green). when these 4 strong basic compounds are left out from the correlation, the correlation coefficient increases to 0.96 and the slope increases to 0.82 as shown in figure 4. the results shown in figure 4 suggests that the rank order of the phospholipid binding will be very similar using the chi iam data obtained on the new iam.pc.dd2 column and the old iam.pc.dd2 column. however, some basic compounds may show significantly weaker phospholipid binding. the absolute chi iam values maybe lower when the new batches of iam columns are used. admet & dmpk 8(1) (2020) 78-97 revisiting the iam chromatography doi: http://dx.doi.org/10.5599/admet.757 87 figure 2. the plot of the experimental chi (iam) values on iam.pc.dd2 columns p20511-014-3 and r20511-014-3 figure 3. the plot of the chi (iam) values from the literature [13] (reference chi (iam)) with the chi (iam) values using iam.pc.dd2 column r25011-014-3. the green line is the line of unity. figure 4. comparison of the experimental chi (iam) values obtained on the iam.pc.dd2 (r20511-014-3) column with the reference chi (iam) values without the 4 strong basic compounds. the green line is the line of unity http://dx.doi.org/10.5599/admet.757 klara valko et al. admet & dmpk 8(1) (2020) 78-97 88 the human clinical volume of distribution and protein binding data of the investigated marketed drugs the volume of distribution model is based on two properties of the compounds, the iam binding and the human serum albumin (hsa) binding. therefore, the iam binding data obtained using the new iam.pc.dd2 columns were examined in the models of the human clinical volume of distribution together with the re-measured protein binding data. figure 5 shows the reproducibility of the hsa columns by plotting the newly measured log k (hsa) values as a function of the “reference log k (hsa) “ values published previously [13]. it can be seen that the slope is very close to 1 and the intercept is very close to 0. there are slightly greater discrepancies between the reference and the re-measured log k (hsa) values for strongly bound compounds. this shows that the specific binding sites of the human serum albumin may vary only slightly from column to column. with this excellent agreement between the log k (hsa) values obtained on various chiral hsa columns, it can be assumed that they would not affect the volume of distribution model which is based on two binding properties for the compounds, namely the iam and hsa binding. figure 5 the plot of the log k (hsa) values from the literature [13] and the re-measured log k (hsa) for this study estimating the volume of distribution using the old and new iam.pc.dd2 columns the logarithmic values of the hsa and iam binding data have been used to estimate the human clinical volume of distribution. using the published equation (equation 4), the estimated values are compared using the iam data obtained on columns r20511-014-3 and p20511-014-3 and the iam data from the literature [12]. the estimated log vdss values have been calculated using equation 4 and the measured log k (hsa) and log k (hsa) data obtained on the new iam.pc.dd2 columns. the estimated log vdss values are listed in table 5 together with the acid/base character of the compounds. figure 6 shows the plot of the estimated log vdss values using the chi (iam) values obtained on the iam.pc.dd2 column p20511-014-3 and those derived from the reference chi (iam) values. a good correlation was found between the estimated log vdss values. in this study equation 4 was used with the same log k (hsa) values and different log k (iam) values (depending on which iam column was used). admet & dmpk 8(1) (2020) 78-97 revisiting the iam chromatography doi: http://dx.doi.org/10.5599/admet.757 89 table 5. the estimated log vdss values of the investigated compounds using equation 4 and their acid/base character drug name acid/base character r20511-014-3 log vdss p20511-014-3 log vdss reference log vdss acecainide basic -0.03 -0.03 0.1 acetanilide neutral -0.12 -0.12 -0.13 acetazolamide neutral -0.47 -0.47 -0.46 alclofenac acidic -1.04 -1.02 -1.1 amoxapine basic 0.62 0.57 1.67 bamethan basic 0.01 0 0.01 betamethasone neutral 0.12 0.12 0.18 carbamazepine neutral -0.07 -0.07 -0.09 chlorpromazine basic 0.88 0.82 2.15 cinoxacin acidic -0.38 -0.37 -0.42 colchicine neutral 0.02 0.01 0.03 cytarabine weak base -0.26 -0.26 -0.31 diazoxide weak acid -0.19 -0.19 -0.16 diclofenac acidic -0.6 -0.57 -0.67 diprophylline neutral -0.22 -0.23 -0.24 ethinyl estradiol neutral 0.28 0.27 0.39 famotidine weak base -0.04 -0.05 -0.06 felbamate neutral 0.07 0.07 -0.64 felodipine neutral 0.23 0.23 0.35 fenoprofen acidic -1.02 -0.98 -1.27 finasteride neutral 0.21 0.21 0.3 floxacillin acidic -0.55 -0.53 -0.56 flumazenil neutral -0.04 -0.04 -0.04 flurbiprofen acidic -1.44 -1.42 -1.56 furosemide acidic -0.7 -0.68 -0.71 gemfibrozil acidic -1.31 -1.27 -1.34 glipizide acidic -0.69 -0.68 -0.73 griseofulvin neutral 0.04 0.03 0.12 hydrochlorothiazide weak acid -0.18 -0.17 -0.13 hydrocortisone neutral 0.1 0.09 0.12 imipramine basic 0.66 0.6 1.16 indomethacin acidic -0.72 -0.69 -0.96 isradipine neutral 0.08 0.08 0.13 ketoconazole weak base 0.12 0.1 0.32 ketoprofen acidic -0.98 -0.96 -1.01 labetalol amphoteric 0.38 0.34 0.71 lignocaine basic 0.14 0.16 0.4 methylprednisolone neutral 0.12 0.11 0.16 metronidazole weak base -0.24 -0.25 -0.25 minoxidil weak base -0.11 -0.1 -0.04 nabumetone neutral -0.09 -0.09 0.03 nadolol basic -0.34 -0.35 -0.33 nicardipine weak base 0.32 0.32 0.47 nifedipine neutral -0.07 -0.07 -0.04 nitrendipine neutral 0.08 0.08 0.16 http://dx.doi.org/10.5599/admet.757 klara valko et al. admet & dmpk 8(1) (2020) 78-97 90 papaverine weak base -0.13 -0.13 0 pentoxifylline neutral -0.07 -0.07 -0.1 perphenazine basic 0.66 0.61 1.35 phenytoin weak acid -0.02 -0.02 0.03 pindolol basic 0.14 0.13 0.85 prazosin weak base -0.11 -0.13 -0.05 prednisolone neutral 0.12 0.12 0.14 prednisone neutral 0.03 0.03 0.05 primidone neutral -0.13 -0.13 -0.15 probenecid acidic -0.66 -0.64 -0.75 procainamide basic -0.13 -0.12 0.02 propanolol basic 0.64 0.59 1.29 propylthiouracil weak acid -0.31 -0.31 -0.34 proxyphylline neutral -0.75 -0.74 -0.85 sulfachlorpyridazine acidic -0.71 -0.7 -0.76 sulfameter acidic -0.57 -0.57 -0.61 sulfamethoxypyridazine weak acid -0.57 -0.56 -0.59 sulfinpyrazone acidic -0.64 -0.62 -0.65 sulfisoxazole acidic -0.68 -0.67 -0.72 sulphadimidine amphoteric -0.41 -0.41 0.22 sulpiride basic -0.11 -0.1 0.12 tamoxifen basic 0.94 0.89 1.48 theobromine weak acid -0.24 -0.25 -0.25 tolfenamic acid acidic -0.54 -0.49 -0.76 trazodone weak base -0.09 -0.1 0.01 trimethoprim weak base -0.05 -0.06 -0.06 warfarin acidic -0.69 -0.68 -0.78 figure 6. the plot of the estimated log vdss values using the chi (iam) values obtained on the iam.pc.dd2 column p20511-014-3 and from the reference [13]. the green line is the line of unity. figure 7 shows the plot of the estimated log vdss values using the chi (iam) values obtained on the iam.pd.dd2 column p20511-014-3 and those obtained from the reference but without the 4 strong basic compounds (procainamide, sulpiride, zolmitriptan and pindolol). it can be seen that the agreement is slightly improved with an r 2 value of 0.96, however, the slope is still less than one, while the intercept is admet & dmpk 8(1) (2020) 78-97 revisiting the iam chromatography doi: http://dx.doi.org/10.5599/admet.757 91 close to zero. this means that the rank order of the volume of distribution data would be the same when using the new batches of iam.pc.dd2 columns as it was when using the old batches of iam columns. however, absolute values may be slightly different for the neutral, acidic and basic compounds and greater discrepancies can be expected for some strong basic compounds which are similar to those left out from the plot. figure 7. the plot of the estimated log vdss values using the chi (iam) values obtained on column p20511-0143 and from the reference values [12] without the four strong basic compounds. the green line is the line of unity. it was important to investigate how the estimated volume of distribution using the chi (iam) data from the new batches of the iam.pc.dd2 columns correlated with the actual clinical volume of distribution data. figures 8 and 9 show the correlation of the estimated log vdss values using the same coefficients as in equation 4 with the clinical log vdss values. figure 9 shows that the chi (iam) values obtained on the new batches of iam.pc.dd2 column can be used to estimate the human clinical volume of distribution data by using equation 4. there is better agreement with the clinical data when using the log vdss obtained on the new batch of the iam.pc.dd2 column. however, the absolute values may be slightly different and a correction factor (slope 0.62) will be required. in a drug discovery setting for ranking and differentiating compounds using an estimated log vdss, these values can be used with confidence. however, it is important to acknowledge that certain strong basic compounds may appear to have a lower volume of distribution than expected. it was also important to investigate how the chi (iam) and log k (iam) values perform in other in vivo distribution models. estimating brain tissue binding using the old and new batches of iam.pc.dd2 columns the brain tissue binding can be estimated from the sum of the albumin and the phospholipid binding as described by equation 8. the model was built on 135 drug discovery compounds from several neuroscience projects using rat brain tissue binding data obtained by equilibrium dialysis and measured hsa and iam binding data as described in reference [12]. log k (btb)=1.29 (±0.10) log k (iam) + 1.03(±0.10) log k (hsa) 0.237 (8) n=135 r 2 =0.76 s=0.35 f=212 where n is the number of compounds, r 2 is the correlation coefficient, s is the standard error of the http://dx.doi.org/10.5599/admet.757 klara valko et al. admet & dmpk 8(1) (2020) 78-97 92 estimate and f is the fisher-test value. figure 8. the plot of the estimated and the measured clinical volume of distribution data using the old batches of iam columns (reference [13] reference log vdss) using equation 4. figure 9. the plot of the estimated and measured clinical volumes of distribution using the new batch r20511-014-3 iam.pc.dd2 column versus the reference log vdss [13]. figure 10 shows the comparison of the estimated brain tissue binding (%btb) when the old (reference [13]) and the new column p20511-014-3 iam.pc.dd2 data was used in the model. the agreement is very good with an r 2 value of 0.95, the slope is close to 1 and the intercept is close to zero using the chi (iam) values obtained on p20511-014-3 from the latest batch of iam.pc.dd2 columns. however, the four basic compounds that showed significantly lower chi (iam) values are significant outliers resulting in these compounds being estimated to have lower brain tissue binding than their actual brain tissue binding. estimating the unbound volume of distribution using the old and new batches of iam.pc.dd2 columns the agreement between the estimated unbound volume of distribution data using the reference [13] chi (iam) values and the chi iam values obtained on column r25011-014-3 are shown in figure 11. admet & dmpk 8(1) (2020) 78-97 revisiting the iam chromatography doi: http://dx.doi.org/10.5599/admet.757 93 figure 10. the plot of the estimated brain tissue binding (%btb) obtained using the iam data obtained from the reference [13] and that obtained using iam.pc.dd2 column p20511-014-3 figure 11. a plot of the estimated unbound volume of distribution using the iam data from the reference (reference [12]) and on the iam.pc.dd2 r20511-014-3 column it can be seen that the estimated unbound volume of distribution using the chi (iam) values obtained on column r20511-014-3 shows good agreement with the logvdu from the reference chi (iam) values. it can be observed from the plot that basic compounds tend to show a lower unbound volume of distribution using the chi (iam) values obtained on the new column r20511-014-3. estimating drug efficiency using the old and new batches of iam.pc.dd2 columns figure 12 shows the comparison of the estimated drug efficiency values using the old and a new batch (r20511-014-3) of iam.pc.dd2 column. there is good agreement between the drug efficiency values obtained on the new batch of the iam column and the literature data. however, the basic compounds showed greater drug efficiency than expected when using the new batch (r20511-014-3) of iam.pc.dd2 column. http://dx.doi.org/10.5599/admet.757 klara valko et al. admet & dmpk 8(1) (2020) 78-97 94 figure 12. the plot of the drug efficiency values for the investigated drugs using the iam data obtained from the reference (reference [12]) and by r20511-014-3 column in conclusion, some differences between the old and the latest batches of iam.pc.dd2 columns were observed when estimating the in vivo distribution of compounds. the discrepancies are more significant when the models are based on the differences between the measured iam and hsa binding (volume of distribution). when the models are based on the sum of the two types of binding, the correlations between the estimated values (brain tissue binding, the unbound volume of distribution and drug efficiency) are much stronger. however, systematic differences could be observed for basic compounds in the estimated drug efficiency values. polar basic (log p less than 2) compounds with over 6 rotatable bonds were estimated to have greater dug efficiency values than previously estimated using the old batches of iam.pc.dd2 columns. these discrepancies are due to the weaker interaction of the new batches of iam.pc.dd2 stationary phases with positively charged basic compounds. this could be the result of better endcapping procedures to block the free silanol groups and it could also be due to the higher purity of the silica providing less overall negatively charged silanols. based on these observations it was decided to investigate the potential benefit of developing new models using the data obtained on the new silica based iam stationary phase and to compare the new models with the existing model equation (equation 4). developing new distribution models using the new batches of iam.pc.dd2 columns the correlations between the old and the new batches of iam retention data have been investigated. it was found that the reference chi (iam) data can be estimated using the chi (iam) data obtained on the new batches (p20511-014-3 and r20511-014-3) of iam.pc.dd2 columns when the retention difference on c-18 columns obtained at ph 10.6 and ph 2.6 are included in the equation (equation (9)). this second variable significantly increased the explained variance of the reference chi (iam). reference chi (iam) = 1.06(±0.03) chi (iam) (r-20511-01403) + 0.008(±0.01)δchi (ph10.5 ph2.6) – 1.06 (±0.73) (9) n= 67 r 2 =0.97 s=2.7 f=1100 the correlation was slightly improved (from 0.92 to 0.97) between the reference chi (iam) values and the chi (iam) values obtained on the new batch of iam.pc.dd2 columns. however, in order to be able to use equation 9, the chi measurements on c-18 columns should be carried out at high and low ph. this admet & dmpk 8(1) (2020) 78-97 revisiting the iam chromatography doi: http://dx.doi.org/10.5599/admet.757 95 requires two additional measurements for each compound. when resources are limited, the calculated molecular descriptor can be used. the following parameters have been calculated for the investigated drug molecules: log p, molecular weight, number of h-bond donor and acceptor groups, total polar surface area, flexibility and the number of rotatable bonds (nrb). it was found that the only significant variables among the calculated parameters that improved the correlation between the chi (iam) values obtained using the new and old batches of iam.pc.dd2 columns were clogp and the total polar surface area (tpsa) as shown in equation 10. however, the correlation coefficient improved only slightly from 0.92 to 0.96. reference chi (iam)=1.38(±0.07) chi (iam)(r20511 – 014 -3) – 2.42(±0.57) clogp -0.03(±0.01) tpsa 1.28 (10) n=68 r 2 =0.96 s=3.3 f=530 the other option was to build a new model for the estimation of the volume of distribution using the new batches of iam.pc.dd2 columns and the clinical volume of distribution data. the new equation for modelling clinical steady-state volume of distribution data is shown in equation 11. the mathematical-statistical parameters are very similar to the original equation (equation 4); however, the coefficients of the iam and hsa binding data are very different. log vdss clinical =0.54(±0.04) log k (iam)(r20511-01403) 0.24(±0.02) log k (hsa) -0.7 (11) n=72 r 2 =0.78 s=0.30 the plot of the estimated and measured clinical steady-state volume of distribution data are shown in figure 13. when comparing the coefficients of equation 4 (the published vdss model equation) and equation 10, it can be seen that the coefficients for the hsa binding do not differ significantly (0.24 and 0.22 with an error of 0.02). however, the coefficients of log k (iam) are different at the 95 % confidence interval and agree only within a 90 % confidence interval: 0.54 ± 0.04 and 0.44 ± 0.02. figure 13. the plot of the calculated log vdss values using the new model with the chi (iam) values obtained on column (r20511-014-3) as a function of the clinical log vdss values. conclusions the use of a new silica support provided a slightly different selectivity for the new iam.pc.dd2 columns regardless of the bonding procedure used. therefore, new chi (iam) values have been proposed for the system suitability test compounds. new iam.pc.dd2 columns should be evaluated using the new chi (iam) values (table 4.). the reproducibility of the new batches of iam.pc.dd2 columns are very good and are http://dx.doi.org/10.5599/admet.757 klara valko et al. admet & dmpk 8(1) (2020) 78-97 96 within 2 chi (iam) units. the effect of the selectivity change has been investigated on previously developed in vivo models for drug distribution that are based on chi (iam) values. a set of 72 marketed drugs with clinical distribution data (volume of distribution, unbound volume of distribution and drug efficiency) have been analyzed on the new batches of iam columns. the chi (iam) values were compared with the values obtained from the previous batches of iam columns and it was observed that small polar basic compounds showed significantly shorter retention times on the new iam.pc.dd2 columns. the correlation coefficient improved significantly when the data from these compounds are removed from the analysis. the possible reason for these discrepancies may be due to the better endcapping procedure on the new iam columns which is blocking the silanol effects that increase retention of basic compounds. it is also likely that the newer “type b” silica is higher purity silica with a more inert surface and creates fewer negatively charged silanols decreasing the amount of non-specific binding. it was found that using the new chi (iam) data in the previously published model equations, a good correlation was obtained with the original values. however, the absolute values were different but the rank order for the in vivo distribution values remain the same using the old and the new columns. new models have also been developed for estimating the clinical volume of distribution of compounds using the data from the new iam columns. the mathematical statistical characteristics of the new models are similar, but the coefficients of the chi (iam) and log k (hsa) parameters are slightly different. since the correlation between the estimated and clinical volume of distribution data was the same or slightly better with the data obtained from the new batches of iam columns, it is expected that the in vivo estimates can be as accurate with the data obtained from the new batches of iam columns as with the older batches of iam columns. the newly derived equation (eqation 11) for the estimation of the volume of distribution is suggested when someone begins to use biomimetic chromatographic measurements with the new columns. however, reference can still be made to the existing data previously published using the older columns. conflict of interest: klara valko is the founder of bio-mimetic chromatography ltd. scott anderson is director at regis technologis inc, while shenaz bunally is scientist at physicochemical group at gsk. references [1] c. pidgeon, u.v. venkataram. immobilized artificial membrane chromatography: supports composed of membrane lipids. anal. biochem. 176 (1989) 36–47. [2] k. valkó, p. slégel. new chromatographic hydrophobicity index (φ0) based on the slope and the intercept of the log k′ versus organic phase concentration plot. j. chromatogr. a, 631 (1993) 49–61. [3] k. valko, c.m. du, c.d. bevan, d.p. reynolds, m.h. abraham. rapid-gradient hplc method for measuring drug interactions with immobilized artificial membrane: comparison with other lipophilicity measures. j. pharm. sci. 89 (2000) 1085–1096. [4] k. valko, s. teague, c. pidgeon. in vitro membrane binding and protein binding (iam mb/pb technology) to estimate in vivo distribution: applications in early drug discovery. admet dmpk 5 (2017) 14-38. [5] z. jiang, j. reilly. chromatography approaches for early screening of the phospholipidosis-inducing potential of pharmaceuticals. j. pharm. biomed. anal. 61 (2012) 184–190. [6] f. tsopelas, c. giaginis, a. tsantili-kakoulidou. lipophilicity and biomimetic properties to support drug discovery. expert opin. drug discov. 12 (2017) 885–896. [7] f. tsopelas, t. vallianatou, a. tsantili-kakoulidou. advances in immobilized artificial membrane (iam) chromatography for novel drug discovery. expert opin. drug discov. 11 (2016) 473-478. [8] s. bunally, r.j. young. the role and impact of high throughput biomimetic measurements in drug discovery, admet and dmpk 6 (2018) 74–84. admet & dmpk 8(1) (2020) 78-97 revisiting the iam chromatography doi: http://dx.doi.org/10.5599/admet.757 97 [9] a. taillardat-bertschinger, c.a. martinet, p.a. carrupt, m. reist, g. caron, r. fruttero, b. testa. molecular factors influencing retention on immobilized artifical membranes (iam) compared to partitioning in liposomes and n-octanol. pharm. res. 19 (2002) 729–737. [10] e. lázaro, c. ràfols, m.h. abraham, m. rosés. chromatographic estimation of drug disposition properties by means of immobilized artificial membranes (iam) and c18 columns. j. med. chem. 49 (2006) 4861–4870. [11] f. barbato, g. di martino, l. grumetto, m.i. la rotonda. prediction of drug-membrane interactions by iam-hplc: effects of different phospholipid stationary phases on the partition of bases. eur. j. pharm. sci. 22 (2004) 261–269. [12] l. grumetto, g. russo, f. barbato. immobilized artificial membrane hplc derived parameters vs pampa-bbb data in estimating in situ measured blood-brain barrier permeation of drugs. mol. pharm. 13 (2016) 2808–2816. [13] f. hollosy, k. valko, a. hersey, s. nunhuck, g. keri, c. bevan. estimation of volume of distribution in humans from hplc measurements of human serum albumin binding and immobilized artificial membrane partitioning. j. med. chem. 49 (2006) 6958–6971. [14] k. valko. physicochemical and biomimetic properties in drug discovery: chromatographic techniques for lead optimization. wiley hoboken nj, 2014. [15] f. tsopelas, t. vallianatou, a. tsantili-kakoulidou. the potential of immobilized artificial membrane chromatography to predict human oral absorption. eur. j. pharm. sci. 81 (2016) 82–93. [16] x. sui, j. sun, h. li, y. wang, j. liu, x. liu, w. zhang, l. chen, z.he. prediction of volume of distribution values in human using immobilized artificial membrane partitioning coefficients, the fraction of compound ionized and plasma protein binding data. eur. j. med. chem. 44 (2009) 4455– 4460. [17] k.l. valkó, s.b. nunhuck, a.p. hill. estimating unbound volume of distribution and tissue binding by in vitro hplc-based human serum albumin and immobilized artificial artificial membranebinding measurements. j. pharm. sci. 100 (2011) 849–862. [18] k. valko, e. chiarparin, s. nunhuck, d. montanari. in vitro measurement of drug efficiency index to aid early lead optimization. j. pharm. sci. 101 (2012) 4155-4169. [19] s. teague, k. valko. how to identify and eliminate compounds with a risk of high clinical dose during the early phase of lead optimisation in drug discovery. eur. j. pharm. sci. 110 (2017) 37-50. [20] a. casartelli, m. bonato, p. cristofori, f. crivellente, g. dal negro, i. masotto, c. mutinelli, k. valko, v. bonfante. a cell-based approach for the early assessment of the phospholipidogenic potential in pharmaceutical research and drug development. cell biol. toxicol. 19 (2003) 161-176. [21] e. lounkine, m.j. keiser, s. whitebread, d. mikhailov, j. hamon, j. jenkins, p. lavan, e. weber, a.k. doak, s. côté, b.k. shoichet, l. urban. large scale prediction and testing of drug activity on side effect targets,” nature, 486 (2012) 361–367. [22] f. hollosy, k. valko, a. hersey, s. nunhuck, g. keri, c. bevan. estimation of volume of distribution in humans from hplc measurements of human serum albumin binding and immobilized artificial membrane partitioning. j. med. chem. 49 (2006) 6958–6971. [23] k. valko, s. nunhuck, c. bevan, m. abraham, d. reynolds. fast gradient hplc method to determine compounds binding to human serum albumin. relationship wth octanol water and immobilized artificial membrane lipophilicity. j. pharm. sci. 92 (2003) 2236–2248. ©2020 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://dx.doi.org/10.5599/admet.757 http://creativecommons.org/licenses/by/3.0/ harmonizing solubility measurement to lower inter-laboratory variance – progress of consortium of biopharmaceutical tools (cobito) in japan doi: 10.5599/admet.704 183 admet & dmpk 7(3) (2019)183-195; doi: http://dx.doi.org/10.5599/admet.704 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper harmonizing solubility measurement to lower inter-laboratory variance – progress of consortium of biopharmaceutical tools (cobito) in japan asami ono 1, *, naoya matsumura 2 , takahiro kimoto 3 , yoshiyuki akiyama 3 , satoko funaki 4 , naomi tamura 4 , shun hayashi 5 , yukiko kojima 6 , masahiro fushimi 6 , hiroshi sudaki 7 , risa aihara 7 , yuka haruna 8 , maiko jiko 8 , masaru iwasaki 9 , takuya fujita 10 and kiyohiko sugano 10 1 asahi kasei pharma corporation, 632-1 mifuku, izunokuni, shizuoka 410-2321, japan 2 ono pharmaceutical co., ltd., 3-3-1, sakurai, shimamoto-cho, mishima-gun,osaka 618-8585, japan 3 central pharmaceutical research institute, japan tobacco inc, 1-1 murasaki-cho, takatsuki,osaka 569-1125, japan 4 shionogi & co., ltd., 3-1-1, futaba-cho, toyonaka-shi,osaka 561-0825, japan 5 sumitomo dainippon pharma co., ltd., 3-1-98 kasugadenaka, konohana-ku, osaka 554-0022, japan 6 sawai pharmaceutical co., ltd., 5-2-30, miyahara, yodogawa-ku, osaka 532-0003, japan 7 nippon boehringer ingelheim co. ltd., 6-7-5 minatojima-minamimachi, chuo-ku, kobe, hyogo 650-0047, japan 8 towa pharmaceutical co., ltd., 134 chudoji minami-machi, shimogyo-ku, kyoto 600-8813, japan 9 daiichi sankyo rd novare co., ltd, 1-16-13 kitakasai, edogawa-ku, tokyo 134-8630, japan 10 college of pharmaceutical sciences, ritsumeikan university, 1-1-1 noji-higashi, kusatsu, shiga 525-8577, japan *corresponding author: e-mail: ono.ar@om.asahi-kasei.co.jp; tel.: +81-558-76-7061; fax: +81-558-76-7137 received: june 12, 2019; revised: july 19, 2019; available online: august 05, 2019 abstract the purpose of the present study was to harmonize the protocol of equilibrium solubility measurements for poorly water-soluble drugs to lower inter-laboratory variance. the “mandatory” and “recommended” procedures for the shake-flask method were harmonized based on the knowledge and experiences of each company and information from the literature. the solubility of model drugs was measured by the harmonized protocol (hp) and the non-harmonized proprietary protocol of each company (nonhp). albendazole, griseofulvin, dipyridamole, and glibenclamide were used as model drugs. when using the nonhp, the solubility values showed large inter-laboratory variance. in contrast, inter-laboratory variance was markedly reduced when using the hp. keywords shake-flask solubility; equilibrium solubility; poorly water-soluble drugs. introduction in recent years, drug candidates have tended to be poorly water-soluble [1-3]. the oral absorption of a drug is determined by the solubility, dissolution rate, and membrane permeability of the drug in the gastrointestinal tract [4-7]. poor water-solubility often causes poor and variable oral absorption [8-11]. therefore, solubility measurements are routinely performed in drug discovery and development [12-15]. the quality of solubility data is critically important for drug design [16-19], oral absorption prediction [20mailto:ono.ar@om.asahi-kasei.co.jp asami ono et. al admet & dmpk 7(3) (2019) 183-195 184 23], and formulation development [24-26]. however, the solubility values of poorly water-soluble drugs vary depending on the experimental procedures [27]. recently, avdeef et al. reported consensus recommendations on the experimental procedures of equilibrium solubility measurements by the shake-flask method [28]. they discussed various aspects of equilibrium solubility measurements, such as incubation temperature, incubation time, the separation method of solids from solution, and ph measurements. as a part of the consortium of biopharmaceutical tools (cobito) project, we conducted a questionnaire concerning the solubility measurements of each company. it was revealed that each company employs different experimental procedures (table 1), suggesting that there would be large inter-laboratory variance in solubility data. table 1. summary of solubility measurement protocols of each company before harmonization. company a b c d e f h vigorous agitation yes yes no yes yes yes no temperature control yes (25 °c) yes (37 °c) yes (37 °c) yes (37 °c) yes (37 °c) yes (37 °c) yes (37 °c) equilibration for > 24 h no (18 h) no (1 h) yes (24 h) yes (24 h) yes (24 h) yes (24 h) yes (24 h) pre-saturation of filter no yes no yes no a yes no preventing adsorption no yes no yes no yes no solid state characterization no no no yes no yes no a separated by centrifugation. the purpose of the present study was to harmonize the experimental procedures of solubility measurements by the shake-flask method for low solubility drugs to lower inter-laboratory variance. albendazole, griseofulvin, dipyridamole, and glibenclamide were used as model drugs (table 2, figures 1 and 2). figure 1. chemical structures of model drugs. admet & dmpk 7(3) (2019) 183-195 harmonizing solubility measurement to lower inter-laboratory variance doi: 10.5599/admet.704 185 table 2. physicochemical properties of model drugs. compound mw pka a log poct b lit. solubility calculated solubility at ph6.8 c , μg/ml ref. solubility (37 °c), μg/ml ph albendazole 265 4.2 (base) 3.1 0.20 (25 °c) 7.4 0.20 31 (25 °c) 0.95 7 0.95 32 0.74 (intrinsic solubility) 0.74 33 0.85 d 6.5 0.85 34 1.1 d 5 0.95 34 0.80 d 6.5 0.79 35 griseofulvin 353 n/a e 2.5 5.3 (25 °c) 7.4 n/a 31 15 d 6.5 n/a 35 15 6.5 n/a 36 16 6.5 n/a 37 dipyridamole 505 4.9 (base) 3.9 4.9 7 4.9 32 (37 °c) 6.6 d 6.5 6.5 35 5.4 6.5 5.3 36 6 6.5 5.9 37 glibenclamide 494 5.2 (acid) 3.1 4.4 (25 °c) 7.4 1.1 31 (37 °c) 5.6 7 3.6 32 4.5 d 6.5 8.8 34 0.3 d 5 7.5 34 4.4 d 6.5 8.7 35 3.0 6.5 5.8 38 0.30 5 7.5 38 a reference [29]; b reference [30]; c calculated by the henderson-hasselbalch equation from the reported solubility value, ph and pka; d measured by μdiss profiler; e not applicable experimental materials albendazole was purchased from tokyo chemical industry co., ltd (tokyo, japan). griseofulvin, dipyridamole, and glibenclamide were purchased from fujifilm wako pure chemical corporation (osaka, japan). an identical batch was used by all companies. the japanese pharmacopeia (jp) 2 nd fluid for dissolution testing (12.5 mm kh2po4 and 12.5 mm na2hpo4, ph 6.8, buffer capacity 10 mm/ ∆ph) was used for the solubility measurement. asami ono et. al admet & dmpk 7(3) (2019) 183-195 186 figure 2. polarized light microscopy images of model drugs (eclipse ti (nikon, japan)). equilibrium solubility measurement by the shake-flask method the equilibrium solubility of the model drugs was measured by the harmonized protocol (hp), and the non-harmonized proprietary protocol of each company (nonhp). the protocol outlined below was presented to each company as the hp. to be viable at each company, the details of the hp were not fully specified. the solubility of the model drug in the jp 2 nd fluid at 37 °c was measured. the experiments were performed in triplicate. 1) add an excess of a drug to a vial, then add a buffer solution (e.g. 5 mg/ 5 ml in a 15 ml vial). 2) after stirring for 1 min with a vortex mixer, shake the sample vigorously at 37 °c for 24 h with light shielding. if possible, measure the concentration-time profile 1 . 3) let stand for a few minutes to sediment solids. 4) filter the supernatant, discarding the first few portions (e.g., 0.1 ml for 4 mm diameter filter). select an appropriate filter material. if possible, pre-heat the filter and syringe to 37 °c before use. 5) dilute the filtrate as necessary. if possible, to avoid adsorption, pre-treat a tip with the filtrate, add a surfactant (a few microliters) to the filtrate, and/or rinse the tip with an organic solvent. 6) measure the final ph. 7) measure the concentration of a drug by an appropriate method (e.g. uv-vis, hplc/uv-vis, or lcms). 8) if possible, analyse the initial and residual solids by powder x-ray diffraction (pxrd) 1 . 1 either a pxrd measurement or a concentration-time profile measurement is required. admet & dmpk 7(3) (2019) 183-195 harmonizing solubility measurement to lower inter-laboratory variance doi: 10.5599/admet.704 187 solubility measurement by μdiss profiler the solubility of the model drugs was also measured with a µdiss profiler™ instrument (pion inc) by company f. this instrument employs eight fibre-optic dip probes, each with its own dedicated photodiode array spectrometer [39]. an excess of a model drug and 20 ml of a buffer solution pre-heated to 37 °c were added to the vessels and stirred at 37 °c for 24 h. the concentration of the model drug was measured by in situ uv probes (wave length 270-280 nm for albendazole, griseofulvin, dipyridamole, and 250-265 nm for glibenclamide). the experiments were performed in duplicate. results and discussion harmonizing the protocol for equilibrium solubility measurements by the shake-flask method consensus recommendations about the shake flask method have been summarized by avdeef et al. [28]. however, since every company has its own situation, it may not be practical to pay equal attention to every recommendation. therefore, we first divided the recommendations into two categories, “mandatory” and “recommended”, including a few additional recommendations based on their possible impact on solubility data. we then embodied these mandatory and recommended items as a harmonized protocol. mandatory agitation. sufficient agitation should be applied to firmly wet and suspend drug particles. poor wettability can often be a problem for poorly water-soluble drugs. sonication is not recommended as it can promote aggregation of solids [28]. incubation temperature. the temperature must be controlled. solubility can vary depending on the drug, by a factor of 0.4 (e.g., erythromycin) to 6.5 (e.g., triflupromazine) between 25 °c and 37 °c [40]. incubation time. the incubation time should be set to > 24 h to achieve equilibrium [28,41-43]. if possible, the concentration should be measured over time to ensure achieving equilibrium, for example at 1, 6, and 24 h [24]. filtration. lipophilic drugs tend to adsorb to the filter material [44,45,46]. a hydrophilic type filter (e.g., hydrophilic pvdf and pes) is recommended for unionized species [28]. the first few portions of a filtrate must be discarded to pre-saturate the filter. pipetting. lipophilic drugs tend to adsorb to a pipet tip and vial materials [47,48]. to avoid adsorption, pipet tips can be pre-treated by the filtrate solution. addition of a drop of surfactant is also effective for preventing adsorption. after pipetting, the tip can be rinsed with an organic solvent. final ph. the ph value can change during solubility measurements for ionizable drugs, especially when a large excess of an ionizable drug is added in a salt form. solid state characterization. the initial form may transform to a more stable form, a free form, a hydrate, etc. during solubility measurements. the solid forms of the initial solids and the residual solids after solubility measurements (equilibrium maker) should be characterized by pxrd, differential scanning calorimetry, raman spectra, etc. recommended filtration temperature. to maintain the temperature of an equilibrated solution during filtration, it is asami ono et. al admet & dmpk 7(3) (2019) 183-195 188 recommended to pre-warm filters and syringes before filtration [49]. light shielding. it is recommended to shade the sample during the solubility measurements. sedimentation after agitation. it is recommended to settle the residual solids after agitation to facilitate filtration. sedimentation can also prevent the formation of a supersaturated solution by intensive stirring [28, 43]. the abovementioned mandatory and recommended items were embodied as a hp, considering practicability at each company (see experimental section). some flexibility was left in the hp so that it could be practically applied to each company. in the previous consensus recommendation paper, sedimentation was recommended as the first choice to separate solids from solution [28]. however, direct sampling from the supernatant requires some technical skills, especially when using small vials or microtubes. fine powders may be re-suspended even by slight stimulation. therefore, we used filtration in the harmonized protocol. equilibrium solubility measured by nonhp and hp eight companies participated in this study (a to h). the equilibrium solubility of four model drugs in the jp 2 nd fluid (ph 6.8) was measured by the nonhp and hp (tables 3 and 4, respectively). the comparisons of the solubility values between laboratories are shown in figure 3. table 3. equilibrium solubility of model drugs in the jp 2 nd fluid (ph 6.8) obtained by the proprietary protocols of each company before harmonization. compound solubility (sd) (µg/ ml) cv% a a b c d e f h average albendazole 0.29 (0.02) 0.67 (0.05) 0.25 (0.03) 0.86 (0.02) 3.74 (1.44) 1.10 (0.01) 0.25 (0.03) 1.02 (1.15) 113 griseofulvin 5.7 (0.3) 12.0 (0.2) 11.0 (0.1) 11.7 (0.1) 12.2 (0.2) 12.1 (0.0) 11.1 (0.1) 10.8 (2.12) 20 dipyridamole 1.24 (0.11) 4.08 (0.29) 0.66 (0.02) 4.63 (0.02) 7.69 (2.88) 5.22 (0.14) 0.66 (0.02) 3.45 (2.49) 72 glibenclamide 1.1 (0.0) 3.4 (0.1) 3.0 (0.0) 4.6 (0.0) 5.4 (0.4) 3.2 (0.2) 3.0 (0.0) 3.4 (1.3) 38 a coefficient of variation of solubility values obtained from 7 companies. table 4. equilibrium solubility of model drugs in the jp 2 nd fluid (ph 6.8) obtained by the harmonized protocol. compound solubility (sd) (µg/ ml) cv% a a b c d e f g h average albendazole 1.11 (0.14) 0.73 (0.08) 0.71 (0.03) 0.86 (0.02) 0.79 (0.02) 0.96 (0.04) 1.01 (0.02) 0.70 (0.03) 0.86 (0.14) 17 griseofulvin 12.0 (0.5) 10.5 (0.1) 11.4 (0.2) 12.3 (0.1) 11.7 (0.1) 12.7 (0.3) 13.0 (0.1) 11.5 (0.2) 11.9 (0.75) 6 dipyridamole 5.15 (0.02) 3.27 (0.06) 3.69 (0.37) 5.65 (0.09) 5.67 (0.50) 5.07 (0.09) 5.62 (0.25) 3.70 (0.37) 4.73 (0.94) 20 glibenclamide 5.2 (0.4) 3.7 (0.2) 4.8 (0.0) 4.4 (0.5) 5.2 (0.3) 5.7 (0.2) 5.6 (0.1) 4.8 (0.0) 4.9 (0.6) 13 a coefficient of variation of solubility values obtained from 8 companies. the solubility values measured by the nonhps showed marked inter-laboratory variance for albendazole, dipyridamole, and glibenclamide (cv = 113, 72, and 38 %, respectively). for griseofulvin, the admet & dmpk 7(3) (2019) 183-195 harmonizing solubility measurement to lower inter-laboratory variance doi: 10.5599/admet.704 189 inter-laboratory variance was smaller (cv = 20 %). the solubility of griseofulvin would be less susceptible to adsorption to a filter and plastic materials because its lipophilicity is lower than the other model drugs. the solubility values measured by a, c and h were lower than the others probably due to adsorption to filters and tips. on the other hand, the solubility measured by e was higher than the others probably because centrifugation was used to separate solids from a solution. figure 3. solubility of model drugs measured by the non-harmonized proprietary protocols (nonhps) and the harmonized protocol (hp). in contrast to the nonhps, the solubility values measured by the hp showed smaller inter-laboratory variance for all model drugs (cv < 20 %). furthermore, the solubility values were in good agreement with the literature values (table 5). the final ph values were within ± 0.03 of the initial ph values for all model drugs. the solid phases of the samples were evaluated by pxrd before and after solubility measurements (by a, b, c, f, and h). as shown in figure 4, no solid form transformation occurred during the solubility measurements for all model drugs. the concentration-time profiles were measured to confirm achieving equilibrium by d, e, f, and g (data not shown). the solubility measurements were also performed by the µdiss profiler. the dissolution profiles of the model drugs were shown in figure 5. the solubility values of albendazole, griseofulvin, dipyridamole, and glibenclamide were 0.69, 13, 5.2, 5.5 µg/ml, respectively. these solubility values were similar to the values measured by the hp (table 5 and figure 6). asami ono et. al admet & dmpk 7(3) (2019) 183-195 190 figure 4. powder x-ray diffraction patterns of samples obtained before (red lines) and after (blue lines) solubility assay (measured by company c) (rigaku smartlab diffractometer, rigaku, japan). the diffraction patterns were collected from 5° to 40° at a scan rate of 5°/min (cu kα radiation (40 kv and 30 ma), rigaku d/tex ultra-highspeed position-sensitive detector). figure 5 dissolution profiles of model drugs by μdiss profiler admet & dmpk 7(3) (2019) 183-195 harmonizing solubility measurement to lower inter-laboratory variance doi: 10.5599/admet.704 191 table 5. average values of equilibrium solubility of model drugs obtained by different methods at ph 6.8, 37 °c. solubility (g/ml) (log, m) compound nonhps hp diss literature nonhps hp diss literature albendazole 1.02 0.86 0.69 0.86 a -5.41 -5.49 -5.58 -5.49 griseofulvin 11 12 13 15 b -4.51 -4.47 -4.43 -4.36 dipyridamole 3.5 4.7 5.2 5.7 c -5.17 -5.03 -4.99 -4.95 glibenclamide 3.4 4.9 5.5 7.0 d -5.16 -5.00 -4.95 -4.85 a reference [32-35] b reference [35-37] c reference [35-37] d reference [32, 34, 35, 38] figure 6 comparison of equilibrium solubility at ph 6.8, 37 °c of model drugs obtained by different methods. in this study, the ph value did not change during incubation, because the drug concentration was set to about 1 mg/ml (as of added solids) and the buffer capacity was set to be sufficient. however, in the case when a large amount of a drug and/or a buffer with limited buffer capacity are used, the final ph after incubation would be different from the initial ph. it should be noted that even when starting with a salt form, at the neutral ph region, the residual solids become a free form after achieving equilibrium if the ph is below (for acids) or above (for bases) the phmax (the ph where the system changes from the phcontrolled region to the common-ion-effect-controlled region). therefore, the observed solubility becomes the same regardless of the starting material being a salt or a free form if the solution is well-buffered to maintain the ph (unless there is a difference in the solid forms of residual free form solids). this point is often misunderstood by pharmacokinetic scientists who perform computational oral absorption simulation for a salt form drug substance. the dissolution of a salt and the supersaturation and phase separation of a free form must be taken into account to simulate the oral absorption of the drug. the ksp (the drug-salt solubility product) value is at least required to simulate the dissolution of a salt. however, the dissolution mechanisms of a salt are not fully understood, especially the chemical reactions in the unstirred water layer [50, 51]. it is highly recommended to analyse the solid forms of both initial and residual solids. in drug discovery, little information about polymorphs and pseudo-polymorphs is available. therefore, the information about asami ono et. al admet & dmpk 7(3) (2019) 183-195 192 the solid form obtained along with the solubility data is important for further solid form characterization. for the assessment of drug solubility in physiological gastrointestinal conditions, an hcl solution of ph 1 to 2 is often used. in an hcl solution, the solubility of a drug is often controlled by the common ionic effect of the chloride ion. in this case, the residual solid after the solubility measurement becomes an hcl salt. in addition, attention should be paid to the formation of soluble drug-buffer complexes or insoluble drugbuffer precipitates in a phosphate buffer for basic drugs [52, 53]. the analysis of the residual solid after solubility measurements can clarify the formation of insoluble drug-buffer precipitates. the harmonized protocol would be suitable for solubility measurements at the preclinical stages, e.g., the late lead optimization and pre-formulation stages. we successfully predicted the oral absorption of proprietary drug candidates in dogs using the solubility data measured by the hp [54]. it should be noted that the solid form may change during the preparation of a suspension formulation for a preclinical in vivo study. conclusions in conclusion, this study demonstrated that the inter-laboratory variance was successfully reduced by applying the harmonized solubility protocol. the harmonized protocol proposed in this study is practically applicable to drug discovery and development. acknowledgements: this study was conducted as part of the consortium for biopharmaceutics tools (cobito) in research center for drug discovery and pharmaceutical development science of ritsumeikan university. kiyohiko sugano is the founder of bioavailabilitydesign llc. conflict of interest: there is no conflict of interest in the work presented. references [1] c. a. lipinski. drug-like properties and the causes of poor solubility and poor permeability. j. pharmacol. toxicol. methods 44 (2000) 235-249. [2] m. m. hann, g. m. keserü. finding the sweet spot: the role of nature and nurture in medicinal chemistry. nature reviews drug discovery 11 (2012) 355. [3] m. j. waring, j. arrowsmith, a. r. leach, p. d. leeson, s. mandrell, r. m. owen, g. pairaudeau, w. d. pennie, s. d. pickett, j. wang, o. wallace, a. weir. an analysis of the attrition of drug candidates from four major pharmaceutical companies. nature reviews drug discovery 14 (2015) 475. [4] j. dressman, g. amidon, d. fleisher. absorption potential: estimating the fraction absorbed for orally administered compounds. j. pharm. sci. 74 (1985) 588-589. [5] g. amidon, h. lennernäs, v. shah, j. crison. a theoretical basis for a biopharmaceutic drug classification: the correlation of in vitro drug product dissolution and in vivo bioavailability. pharm. res. 12 (1995) 413-420. [6] l. yu. an integrated model for determining causes of poor oral drug absorption. pharm. res. 16 (1999) 1883-1887. [7] k. sugano. introduction to computational oral absorption simulation. expert opin. drug metab. toxicol. 5 (2009) 259-293. [8] s. venkatesh, r. a. lipper. role of the development scientist in compound lead selection and optimization. j. pharm. sci. 89 (2000) 145-154. [9] v. h. thomas, s. bhattachar, l. hitchingham, p. zocharski, m. naath, n. surendran, c. l. stoner, a. elkattan. the road map to oral bioavailability: an industrial perspective. expert opin. drug metab. toxicol. 2 (2006) 591-608. [10] s. stegemann, f. leveiller, d. franchi, h. de jong, h. linden. when poor solubility becomes an issue: from early stage to proof of concept. eur. j. pharm. sci. 31 (2007) 249-261. admet & dmpk 7(3) (2019) 183-195 harmonizing solubility measurement to lower inter-laboratory variance doi: 10.5599/admet.704 193 [11] p. van hoogevest, x. liu, a. fahr. drug delivery strategies for poorly water-soluble drugs: the industrial perspective. expert opin drug deliv 8 (2011) 1481-1500. [12] k. sugano, a. okazaki, s. sugimoto, s. tavornvipas, a. omura. solubility and dissolution profile assessment in drug discovery. drug metab. pharmacokinet. 22 (2007) 225-254. [13] j. alsenz, m. kansy. high throughput solubility measurement in drug discovery and development. adv drug deliv rev 59 (2007) 546-567. [14] l. di, p. v. fish, t. mano. bridging solubility between drug discovery and development. drug discov today 17 (2012) 486-495. [15] a. konczol, g. dargo. brief overview of solubility methods: recent trends in equilibrium solubility measurement and predictive models. drug discovery today. technologies 27 (2018) 3-10. [16] l. di, e. h. kerns, g. t. carter. drug-like property concepts in pharmaceutical design. curr. pharm. des. 15 (2009) 2184-2194. [17] c. a. lipinski, f. lombardo, b. w. dominy, p. j. feeney. experimental and computational approaches to estimate solubility and permeability in drug discovery and development settings. advanced drug delivery reviews 64 (2012) 4-17. [18] g. papadatos, m. alkarouri, v. j. gillet, p. willett, v. kadirkamanathan, c. n. luscombe, g. bravi, n. j. richmond, s. d. pickett, j. hussain. lead optimization using matched molecular pairs: inclusion of contextual information for enhanced prediction of herg inhibition, solubility, and lipophilicity. j. chem. inf. model. 50 (2010) 1872-1886. [19] c. le manach, a. t. nchinda, t. paquet, d. gonzalez cabrera, y. younis, z. han, s. bashyam, m. zabiulla, d. taylor, n. lawrence, k. l. white, s. a. charman, d. waterson, m. j. witty, s. wittlin, m. e. botha, s. h. nondaba, j. reader, l. m. birkholtz, m. b. jimenez-diaz, m. s. martinez, s. ferrer, i. angulo-barturen, s. meister, y. antonova-koch, e. a. winzeler, l. j. street, k. chibale. identification of a potential antimalarial drug candidate from a series of 2-aminopyrazines by optimization of aqueous solubility and potency across the parasite life cycle. j. med. chem. 59 (2016) 9890-9905. [20] r. takano, k. furumoto, k. shiraki, n. takata, y. hayashi, y. aso, s. yamashita. rate-limiting steps of oral absorption for poorly water-soluble drugs in dogs; prediction from a miniscale dissolution test and a physiologically-based computer simulation. pharm. res. 25 (2008) 2334-2344. [21] k. sugano. fraction of a dose absorbed estimation for structurally diverse low solubility compounds. int. j. pharm. 405 (2011) 79-89. [22] n. matsumura, y. yamaura, j. katagi, s. ono, s. kim, s. yamashita, k. sugano. evaluation of using dogs to predict fraction of oral dose absorbed in humans for poorly water-soluble drugs. j. pharm. sci. 107 (2018) 2489-2496. [23] c. a. bergstrom, r. holm, s. a. jorgensen, s. b. andersson, p. artursson, s. beato, a. borde, k. box, m. brewster, j. dressman, k. i. feng, g. halbert, e. kostewicz, m. mcallister, u. muenster, j. thinnes, r. taylor, a. mullertz. early pharmaceutical profiling to predict oral drug absorption: current status and unmet needs. eur. j. pharm. sci. 57 (2014) 173-199. [24] v. saxena, r. panicucci, y. joshi, s. garad. developability assessment in pharmaceutical industry: an integrated group approach for selecting developable candidates. j. pharm. sci. 98 (2009) 1962-1979. [25] y. kawabata, k. wada, m. nakatani, s. yamada, s. onoue. formulation design for poorly watersoluble drugs based on biopharmaceutics classification system: basic approaches and practical applications. int. j. pharm. 420 (2011) 1-10. [26] a. dahan, a. beig, d. lindley, j. m. miller. the solubility-permeability interplay and oral drug formulation design: two heads are better than one. adv drug deliv rev 101 (2016) 99-107. [27] d. s. palmer, j. b. mitchell. is experimental data quality the limiting factor in predicting the aqueous solubility of druglike molecules? mol. pharm. 11 (2014) 2962-2972. [28] a. avdeef, e. fuguet, a. llinàs, c. ràfols, e. bosch, g. völgyi, t. verbić, e. boldyreva, k. takács-novák. equilibrium solubility measurement of ionizable drugs – consensus recommendations for improving data quality. admet and dmpk 4 (2016) 117. [29] a. avdeef. absorption and drug development, second edition. john wiley & sons, inc (2012) [30] k. sugano. biopharmaceutics modeling and simulations. wiley (2012) [31] y. w. alelyunas, r. liu, l. pelosi-kilby, c. shen. application of a dried-dmso rapid throughput 24-h equilibrium solubility in advancing discovery candidates. eur. j. pharm. sci. 37 (2009) 172-182. asami ono et. al admet & dmpk 7(3) (2019) 183-195 194 [32] a. glomme, j. marz, j. b. dressman. comparison of a miniaturized shake-flask solubility method with automated potentiometric acid/base titrations and calculated solubilities. j. pharm. sci. 94 (2005) 116. [33] k. sugano, t. kato, k. suzuki, k. keiko, t. sujaku, t. mano. high throughput solubility measurement with automated polarized light microscopy analysis. j. pharm. sci. 95 (2006) 2115-2122. [34] j. h. fagerberg, o. tsinman, n. sun, k. tsinman, a. avdeef, c. a. bergström. dissolution rate and apparent solubility of poorly soluble drugs in biorelevant dissolution media. mol. pharm. 7 (2010) 1419-1430. [35] j. h. fagerberg, y. al-tikriti, g. ragnarsson, c. a. bergstrom. ethanol effects on apparent solubility of poorly soluble drugs in simulated intestinal fluid. mol. pharm. 9 (2012) 1942-1952. [36] e. söderlind, e. karlsson, a. carlsson, r. kong, a. lenz, s. lindborg, j. j. sheng. simulating fasted human intestinal fluids: understanding the roles of lecithin and bile acids. mol. pharm. 7 (2010) 1498-1507. [37] r. takano, k. sugano, a. higashida, y. hayashi, m. machida, y. aso, s. yamashita. oral absorption of poorly water-soluble drugs: computer simulation of fraction absorbed in humans from a miniscale dissolution test. pharm. res. 23 (2006) 1144-1156. [38] s. clarysse, j. brouwers, j. tack, p. annaert, p. augustijns. intestinal drug solubility estimation based on simulated intestinal fluids: comparison with solubility in human intestinal fluids. eur. j. pharm. sci. 43 (2011) 260-269. [39] a. avdeef, o. tsinman. miniaturized rotating disk intrinsic dissolution rate measurement: effects of buffer capacity in comparisons to traditional wood’s apparatus. pharm. res. 25 (2008) 2613-2627. [40] a. avdeef. solubility temperature dependence predicted from 2d structure. admet & dmpk 3 (2015) [41] k. sugano, h. hamada, m. machida, h. ushio. high throughput prediction of oral absorption: improvement of the composition of the lipid solution used in parallel artificial membrane permeation assay. journal of biomolecular screening 6 (2001) 189-196. [42] k. kawakami, k. miyoshi, y. ida. impact of the amount of excess solids on apparent solubility. pharm. res. 22 (2005) 1537-1543. [43] e. baka, j. e. comer, k. takacs-novak. study of equilibrium solubility measurement by saturation shake-flask method using hydrochlorothiazide as model compound. j. pharm. biomed. anal. 46 (2008) 335-341. [44] l. pan, q. ho, k. tsutsui, l. takahashi. comparison of chromatographic and spectroscopic methods used to rank compounds for aqueous solubility. j. pharm. sci. 90 (2001) 521-529. [45] m. lindenberg, c. wiegand, j. b. dressman. comparison of the adsorption of several drugs to typical filter materials. dissolution technologies 12 (2005) 22-25. [46] g. volgyi, d. csicsak and k. takacs-novak. right filter-selection for phase separation in equilibrium solubility measurement. eur. j. pharm. sci. 123 (2018) 98-105. [47] j. j. palmgren, j. monkkonen, t. korjamo, a. hassinen, s. auriola. drug adsorption to plastic containers and retention of drugs in cultured cells under in vitro conditions. eur. j. pharm. biopharm. 64 (2006) 369-378. [48] t. fukazawa, y. yamazaki, y. miyamoto. reduction of non-specific adsorption of drugs to plastic containers used in bioassays or analyses. j. pharmacol. toxicol. methods 61 (2010) 329-333. [49] t. tokumura. a screening system of solubility for drug design and discovery. pharm tech japan 16 (2000) 19-27. [50] j. oki, d. watanabe, t. uekusa, k. sugano. mechanism of supersaturation suppression in dissolution process of acidic drug salt. mol. pharm. 16 (2019) 1669-1677. [51] t. uekusa, k. sugano. precipitation behavior of pioglitazone on the particle surface of hydrochloride salt in biorelevant media. j. pharm. biomed. anal. 161 (2018) 45-50. [52] e. shoghi, e. fuguet, e. bosch and c. rafols. solubility-ph profiles of some acidic, basic and amphoteric drugs. eur. j. pharm. sci. 48 (2013) 291-300. [53] o. s. markovic, m. p. pesic, a. v. shah, a. t. m. serajuddin, t. z. verbic and a. avdeef. solubility-ph profile of desipramine hydrochloride in saline phosphate buffer: enhanced solubility due to drugbuffer aggregates. eur. j. pharm. sci. 133 (2019) 264-274. admet & dmpk 7(3) (2019) 183-195 harmonizing solubility measurement to lower inter-laboratory variance doi: 10.5599/admet.704 195 [54] y. akiyama, t. kimoto, h. mukumoto, s. miyake, s. ito, t. taniguchi, y. nomura, n. matsumura, t. fujita, k. sugano. prediction accuracy of mechanism-based oral absorption model for dogs. j. pharm. sci. (2019). ©2019 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ permeation characteristics of tetracyclines in parallel artificial membrane permeation assay ii: effect of divalent metal ions and mucin doi: http://dx.doi.org/10.5599/admet.797 129 admet & dmpk 8(2) (2020) 129-138; doi: http://dx.doi.org/10.5599/admet.797 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper permeation characteristics of tetracyclines in parallel artificial membrane permeation assay ii: effect of divalent metal ions and mucin sachika yamauchi, daisuke inoue, kiyohiko sugano * molecular pharmaceutics lab., college of pharmaceutical sciences, ritsumeikan university, 1-1-1, noji-higashi, kusatsu, shiga 525-8577, japan *corresponding author: e-mail: suganok@fc.ritsumei.ac.jp; tel.: +81-77-561-2773 received: february 27, 2020; revised: april 05, 2020; published: may 01, 2020 abstract the bioavailability of tetracyclines is markedly decreased when co-administered with antacids, milk, or food containing ca 2+ . previously, it was suggested that the effective intestinal permeation of tetracycline (tc) was decreased due to ca 2+ linked mucin binding in the mucosal side. in the present study, we investigated the effect of ca 2+ , mg 2+ , and mucin on the membrane permeation of six tetracyclines (tc, oxytetracycline (otc), minocycline (mino), doxycycline (doxy), demeclocycline (dmctc), and chlortetracycline (ctc)). the membrane permeability values (pe) of tetracyclines were measured by the parallel artificial membrane permeation assay (pampa) using soybean lecithin – decane (sl–pampa) and octanol (oct–pampa) membranes. in sl–pampa, ca 2+ markedly decreased the pe values of all tetracyclines. in oct–pampa, ca 2+ increased the pe values of tc, ctc, and dmctc, but not doxy, otc, and mino. mg 2+ decreased the pe values of all tetracyclines in both sl–pampa and oct–pampa (except for ctc in oct–pampa). the addition of mucin had little or no effect in all cases. in contrast to the previously suggested mechanism, the results of the present study suggested that ca 2+ chelate formation decreased the membrane permeation of tetracyclines, irrespective of ca 2+ linked mucin binding. molecular speciation analysis suggested that the permeation of tc – metal chelates was negligibly small in sl-pampa. ©2020 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords artificial membrane; permeability; phospholipid; tetracycline; metal; cation; mucin introduction co-administration of multivalent metal ions reduces the bioavailability of various drugs, such as tetracyclines, fluoroquinolones, hiv-integrase inhibitors, and platelet-stimulating agents [1–4]. for example, the bioavailability of tetracyclines is markedly decreased when co-administered with antacids, milk, and food containing ca 2+ [4–10]. it is generally accepted that chelate formation between tetracycline (tc) and ca 2+ is behind the observed decrease in the bioavailability of tetracyclines [11–16]. chelate formation of tetracyclines has been extensively investigated (ref. [17] and references therein). however, the exact mechanism of the ca 2+ effect on the bioavailability of tetracyclines has not been clear. http://dx.doi.org/10.5599/admet.797 http://dx.doi.org/10.5599/admet.797 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:suganok@fc.ritsumei.ac.jp http://creativecommons.org/licenses/by/4.0/ kiyohiko sugano et al. admet & dmpk 8(2) (2020) 129-138 130 several ex-vivo and in-situ studies have shown that divalent metal ions, such as ca 2+ , mg 2+ , and fe 2+ , reduce the intestinal wall permeation of tetracyclines [18–22]. in 1968, kakemi et al. investigated the effect of ca 2+ on the effective intestinal wall permeation of tetracycline (tc) using the rat small intestine [21,22]. they also measured the isopentanol – buffer partition coefficient as a surrogate of passive transcellular membrane permeability without the interference from mucin. they found that ca 2+ decreased the effective intestinal wall permeation of tc in the rat ex-vivo experiment, but increased the partition coefficient of tc. they also found that tc bound to the intestinal mucin layer in the presence of ca 2+ . based on these observations, they suggested that ca 2+ linked mucin binding decreased the tc concentration available for membrane permeation, resulting in a decrease in the effective intestinal wall permeation. schumacher and linn also reported that ca 2+ increased the transfer rate of tc from the aqueous phase to the octanol phase [23]. however, it is questionable whether these alcohol systems could be a good surrogate model for investigating the effect of divalent metal ions on the membrane permeation of drugs. since divalent metal ions may affect cellular integrity, mucin-free cell-based systems such as caco-2 have rarely been used to examine the effects of multivalent metal ions on membrane permeation [24,25]. the parallel artificial membrane permeation assay (pampa) has been widely used to assess the passive membrane permeation of a drug [26–29]. phospholipid-based artificial membranes are most commonly used with pampa. pampa permeability correlates with the in vivo and cellular permeation of drugs better than the octanol-buffer partition coefficient [27,28]. recently, we reported the permeation characteristics of tetracyclines in a phospholipid-based pampa [30]. only a weak correlation was observed between the pampa permeability (pe) and the octanol-buffer partition coefficients (log doct) for tetracyclines, suggesting that chemoselectivity differs between these systems. however, the effects of divalent metal ions on the pampa permeation of tetracyclines have been unknown. the purpose of the present study was to investigate the effect of ca 2+ , mg 2+ , and mucin on the phospholipid-based pampa permeation of tetracyclines. six tetracycline derivatives were used in this study (figure 1). the physicochemical properties of these tetracyclines have been summarized in table 1 [31–33]. figure 1. chemical structures of tetracyclines experimental materials tetracycline hydrochloride (tc), decane, calcium dichloride, magnesium dichloride, octanol, pig stomach mucin, and 8 m naoh were purchased from wako pure chemical industries, ltd (osaka, japan). admet & dmpk 8(2) (2020) 129-138 effect of metal ions on pampa permeation of tetracyclines doi: http://dx.doi.org/10.5599/admet.797 131 oxytetracycline hydrochloride (otc), minocycline hydrochloride (mino), and doxycycline hyclate (doxy) were purchased from tci (tokyo, japan). 2-morpholinoethanesulfonic acid (mes) was purchased from dojindo laboratories (tokyo, japan). demeclocycline hydrochloride (dmctc) and chlortetracycline hydrochloride (ctc) were purchased from lkt labs, inc (mn, usa). soybean lecithin (slp-white) was provided by tsuji oil mills co., ltd (mie, japan). table 1. physicochemical properties of tetracyclines mw pka log doct (ph 6.5) a pka ref. chlortetracycline 479 3.3, 7.6, 9.3 -0.88 [31] b 3.25, 6.72, 8.84 [32] c demeclocycline 465 3.4, 7.4, 9.4 -0.67 [31] b doxycycline 444 3.0, 8.0, 9.2 -0.08 [31] b 3.50, 7.25, 9.58 [32] c minocycline 457 2.8, 5.0, 7.8, 9.5 0.20 [33] d oxytetracycline 460 3.2, 7.5, 8.9 -0.96 [31] b 3.53, 7.25, 9.58 [32] c tetracycline 444 3.3, 7.8, 9.6 -1.09 [31] b 3.35, 7.29, 9.88 [32] c a measured by a shake-flask method. ref. [30]. b potentiometry (23 °c), ionic strength = 0.01 or 0.05 m. c potentiometry (25 °c), ionic strength = 0.1 m. d method not described in the literature. pampa assay the pampa sandwich was consisted of a 96 well filter plate (hydrophobic pvdf, 0.45 μm) and a pampa acceptor plate (merck millipore, ma, usa). before forming the pampa sandwich, the bottom (acceptor) plate was filled with 300 μl of a 50 mm mes buffer (ph 6.5). the filter of the top (donor) compartment was coated with 5 μl of a 10 % soybean lecithin (sl) – decane solution or octanol. a drug solution (0.5 mm, 200 μl) with or without a divalent metal ion (5 mm) and/or mucin (1 %) in the same buffer was added to the donor compartment. the pampa sandwich was then incubated for 3 h at 37 °c. after incubation, 150 μl of both the donor and acceptor solutions were transferred to a uv plate. the concentrations of tetracyclines were measured at 360 nm. the pampa permeability (pe) was calculated by the following equation [34]. d v d e v v d 2.303 1 ( )1 log 1 1 (0) v r c t p r at r r c                    (1) d a d v d ( ) ( )1 1 (0) (0) c t c t r c r c    (2) d v a v r v  (3) where pe is the effective permeation coefficient (cm/s), a is the filter surface area (0.266 cm 2 ), vd and va are the volumes (ml) in the donor and acceptor phase, t is the incubation time, cd(t) is the concentration of a drug in the donor phase at time t, r is the membrane retention factor, and rv is the volume ratio. we confirmed that the phospholipid decane membrane is stable during the experimental period (no leakage of impermeable substrate) (data not shown). http://dx.doi.org/10.5599/admet.797 kiyohiko sugano et al. admet & dmpk 8(2) (2020) 129-138 132 results and discussion previously, we reported that the pe value of tc was markedly affected by the composition of phospholipids in pampa [30]. in this study, a soybean lecithin (sl, 10 %) – decane membrane (sl–pampa) was used because it most likely mimics the intestinal membrane [34]. the soybean lecithin contained phosphatidylcholine (24-32 %), phosphatidylethanolamine (20-28 %), phosphatidylinositol (12-20 %), phosphatidic acid (8-15 %), and lysophosphatidylcholines (1–5 %) (based on the product information provided by the manufacturer). in addition, an octanol membrane (oct–pampa) [35,36] was also used because ca 2+ was reported to increase the log doct of tc [23]. since ca 2+ interacts with phosphate and citrate ions, mes buffer was used in this study. the concentration of ca 2+ was set to 5 mm based on the standard of the daily intake in food (600 mg) [37] and the gastrointestinal fluid volume in the fed state [38]. we previously reported that the pe value of tc in sl-pampa was not affected by the ionic strength up to 2 mol/l (adjusted by nacl) [34]. in sl–pampa, ca 2+ and mg 2+ markedly decreased the pe values of all tetracyclines investigated in this study, whereas mucin showed little or no effect (figure 2). these results suggest that, in contrast to the previous suggestion based on the alcohol–water partition coefficient [21–23], ca 2+ chelate formation decrease the membrane permeation of tetracyclines, irrespective of ca 2+ linked mucin binding. figure 2. effect of additives on sl–pampa permeation of tetracyclines (mean ± sd, n = 3 6). molecular speciation analysis was performed to elucidate the effect of ca 2+ and mg 2+ on the slpampa permeation of tetracyclines. the details of molecular speciation analysis have been reported by werner et al [39]. tetracyclines and divalent metal ions can form a chelate with various stoichiometries (2:1, 1:1, 1:2), depending on the ionization state of tetracyclines and metal ion species [17,40–44]. in this analysis, macro pka and major molecular species are considered [45,46]. the fraction of each molecular species (l 0 , l -1 , l -2 , m 2+ l -1 , m 2+ l -2 : l = tetracyclines, m = metal) (figure 3) was calculated from the pka values and the metal ion admet & dmpk 8(2) (2020) 129-138 effect of metal ions on pampa permeation of tetracyclines doi: http://dx.doi.org/10.5599/admet.797 133 association constants (kml = [l z m 2+ ]/([m 2+ ][l z ], z = -1, -2) of tetracyclines (tables 1 and 2) [17,39,47]. in the neutral ph region, tetracyclines mainly exist as an equilibrium between a charge–neutral form (l 0 ), and negatively charged forms (l -1 , l -2 ) (figure 3) [31,40,47]. even though l 0 does not bind to the metal ions [47], ca 2+ and mg 2+ reduce the fraction of l 0 (fl0) at ph 6.5 by shifting the equilibrium (table 3). the reduction of fl0 corresponded to that of pe, except for the ca 2+ effect on oxy permeability, suggesting that the slpampa membrane is impermeable to m 2+ l -1 . the ph pe relationship in our previous study [30] suggested that tc 0 , but not tc -1 , predominantly permeates the sl-pampa membrane. however, further investigation is needed to better understand the effect of metal ions on tetracycline membrane permeation. the kml values reported in the literature show large variation [17,39,47]. the fl0 value is especially sensitive to the kml value of m 2+ l -1 . in addition, m 2+ l -1 chelates may have different stoichiometry (1: 1 or 1: 2) [17,39,47]. micro speciation with micro pka values is required to decouple the contributions of uncharged and zwitterionic forms in l 0 [45]. oh n + oh o o ch3 ch3 ch3 o nh2 ooh h h oh h h h – oh n + o o o ch3 ch3 ch3 o nh2 ooh h h oh h h h – – tc 0 tc -1 pk a = 7.8 oh n o o o ch3 ch3 ch3 o nh2 ooh h h oh h h – – pk a = 9.6 tc -2 figure 3 ionization states of tetracycline (tc) at the neutral ph region. the macro pka value and major molecular species are shown in this figure. table 2. association constants of ca 2+ and mg 2+ with tetracyclines (l = tc, ctc, or oxy) reactions log kml tc ctc oxy ca 2+ + l -1 ⇄ ca 2+ l -1 3.4 a , 3.0 b 3.8 c , 2.9 b , 2.9 b ca 2+ + l -2 ⇄ ca 2+ l -2 5.8 a , 4.0 b 5.9 c , 3.9 b , 3.8 b , 4.9 c mg 2+ + l -1 ⇄ mg 2+ l -1 3.9 a , 3.5 b 3.3 c , 3.2 b 3.3 b mg 2+ + l -2 ⇄ mg 2+ l -2 4.1 a , 4.2 b 4.7 c , 4.1 b 4.3 b , 5.2 c a ref. [39] b ref. [47] c ref. [17] in oct–pampa, ca 2+ increased the pe values of tc, ctc, and dmctc (figure 4). this result is in good agreement with the previous studies investigating the ca 2+ effect on the alcohol – water partition coefficient of tc (octanol and isopentanol) [21–23]. interestingly, ca 2+ affected sl–pampa and oct– pampa in the opposite direction for tc, ctc, and dmctc, but in the same direction for doxy, otc, and mino. on the other hand, mg 2+ decreased the pe values of all tetracyclines in oct–pampa except for ctc (no effect). these results suggest that it could be inappropriate to use octanol as a surrogate of a phospholipid membrane for investigating the effect of divalent metal ions. in oct-pampa, the octanol phase could contain water molecules in reverse micelle structures [48]. this may facilitate the permeation of charged species, such as the tc – metal chelates. in similar to sl–pampa, the addition of mucin did not affect the pe values in oct–pampa, suggesting that there is no interaction between tetracyclines and mucin. as expected, there is a good correlation between log doct and log pe in oct–pampa (figure 5) [35,36]. http://dx.doi.org/10.5599/admet.797 kiyohiko sugano et al. admet & dmpk 8(2) (2020) 129-138 134 table 3. fraction of molecular species at ph 6.5 a tetracyclines (l) metal ions (m) fraction of molecular species b reduction % l 0 l -1 l -2 m 2+ l -1 m 2+ l -2 fl0 pe tc none 0.95 0.05 < 0.01 c ca 2+ d,e 0.55 0.03 < 0.01 0.35 0.07 42 35 mg 2+ d,e 0.33 0.02 < 0.01 0.65 < 0.01 65 67 ctc none 0.93 0.07 < 0.01 ca 2+ d,e 0.24 0.02 < 0.01 0.61 0.12 74 64 mg 2+ d,e 0.53 0.04 < 0.01 0.42 0.02 43 56 oxy none 0.91 0.09 < 0.01 ca 2+ d,e 0.66 0.07 < 0.01 0.26 < 0.01 27 61 mg 2+ d,e 0.47 0.05 < 0.01 0.47 0.02 49 64 a activity coefficients were assumed to be 1. see ref. [39] for details; b l = tc, ctc, or oxy. m = ca or mg; c not applicable; d 5.0×10 -3 mol/l; e the kml values were from ref. [39], [17], [47] for tc, ctc, and oyx, respectively. figure 4. effect of additives on oct–pampa permeation of tetracyclines (mean ± sd, n = 3 6). admet & dmpk 8(2) (2020) 129-138 effect of metal ions on pampa permeation of tetracyclines doi: http://dx.doi.org/10.5599/admet.797 135 figure 5. correlation between log doct and log pe of tetracyclines in oct–pampa. the log doct values were taken from the literatrue [30]. clinically, co-administration of food and milk has been reported to decrease the bioavailability of tetracyclines (table 4) [4–10]. the effect of food and milk on bioavailability is greater for tc and oxy, but relatively small for doxy and mino [4]. however, in the present study, the percent reduction of pe by ca 2+ was smallest for tc (mino (87 %) > dmctc (67 %) ≈ ctc (64 %) ≈ doxy (62 %) ≈ oxy (61 %) > tc (35 %)). the pe values of mino and doxy are higher than that of the other tetracyclines. in addition, after oral administration, mino and doxy are almost completely absorbed, whereas tc, oxy, ctc and dmctc are incompletely absorbed [4]. therefore, the reduction of pe by ca 2+ may have less impact on the bioavailability of mino and doxy. barza et al. reported that, after the administration of tetracyclines with milk into the ileal loop in dogs, the remaining fraction in the luminal contents is doxy >> oxy ≈ mino ≈ tc [49]. lipophilicity may play a role in food and milk binding. the balance of metal ion chelating, food/ milk binding, and membrane permeation may determine the extent of food and milk effects. interestingly, the effects of metal ions and ph [30] on the bioavailability of tc to e.coli are similar to that on sl-pampa permeation [50]. table 4. summary of food and milk effects on bioavailability of tetracyclines drugs percentage absorption, % a bioavailability reduction, % b references for food and milk effect food milk ctc 25–30 45 c na d [9] dmctc 66 na d 70 [6] doxy 95 26 (3-49) 30 (9-53) [10] mino 95-100 14 (2-51) 27 (8-61) [6,9,10] oxy 58 41 (4-77), 0 c, e 83 (45-96) [6,9,10] tc 77–88 46 (13-73%), 72 c 65 [5,8–10] a ref. [4]; b in humans unless otherwise noted; c in pigs; d data not available in the literature; e low bioavailability (3 % in both fasted and fed pigs) we could not find any plausible chemical structural elucidation for the differences among tetracyclines regarding the effects of ca 2+ and mg 2+ . tetracyclines can easily modify their tautomerism in response to various chemical environments [44]. metal binding to anionic phospholipids in the sl-pampa membrane may be another possible mechanism to reduce the permeation of tetracyclines. in our previous study, the addition of an anionic lipid neutralizer (tetrahexylammonium) did not affect the permeation of tc in slpampa, suggesting that the ionic interaction with anionic phospholipids do not facilitate the permeation of http://dx.doi.org/10.5599/admet.797 kiyohiko sugano et al. admet & dmpk 8(2) (2020) 129-138 136 tc [30]. further investigation is required to clarify the interactions among tetracyclines, metal ions, and phospholipids. we are currently investigating the effects of metal ions on the sl-pampa permeation of structurally diverse drugs. conclusion in contrast to the previously suggested mechanism [21,22], in this study, ca 2+ chelate formation decreased the membrane permeation of tetracyclines, irrespective of ca 2+ linked mucin binding. ca 2+ affected the pe values in sl–pampa and oct–pampa in the opposite direction for some tetracyclines. sl– pampa can be a simple tool to qualitatively evaluate the effect of multivalent metal ions on the membrane permeation of drugs. conflict of interest: the authors declare no conflict of interest. references [1] b.m. lomaestro, g.r. bailie. absorption interactions with fluoroquinolones. drug saf. 12 (1995) 314– 333. [2] r. krishna, l. east, p. larson, c. valiathan, k. butterfield, y. teng, m. hernandez-illas. effect of metalcation antacids on the pharmacokinetics of 1200 mg raltegravir. j. pharm. pharmacol. 68 (2016) 1359–1365. [3] d.d. williams, b. peng, c.k. bailey, m.b. wire, y. deng, j.w. park, d.a. collins, s.g. kapsi, j.m. jenkins. effects of food and antacids on the pharmacokinetics of eltrombopag in healthy adult subjects: two single-dose, open-label, randomized-sequence, crossover studies. clin. ther. 31 (2009) 764–776. [4] k.n. agwuh, a. macgowan. pharmacokinetics and pharmacodynamics of the tetracyclines including glycylcyclines. j. antimicrob. chemother. 58 (2006) 256–265. [5] h. jung, a.a. peregrina, j.m. rodriguez, r. moreno-esparza. the influence of coffee with milk and tea with milk on the bioavailability of tetracycline. biopharm. drug dispos. 18 (1997) 459–463. [6] j.j. leyden. absorption of minocycline hydrochloride and tetracycline hydrochloride: effect of food, milk, and iron. j. am. acad. dermatol. 12 (1985) 308–312. [7] bioavailability of tetracycline and doxycycline in fasted and nonfasted subjects. antimicrob. agents chemother. 11 (1977) 462–469. [8] h.j. cook, c.r. mundo, l. fonseca. influence of the diet on bioavailability of tetracycline. biopharm. drug dispos. 14 (1993) 549–553. [9] p. nielsen, n. gyrd-hansen. bioavailability of oxytetracycline, tetracycline and chlortetracycline after oral administration to fed and fasted pigs. j. vet. pharmacol. ther. 19 (1996) 305–311. [10] f.p. meyer. minocycline for acne. food reduces minocycline’s bioavailability. bmj br. med. j. 312 (1996) 1101. [11] p.j. neuvonen. interactions with the absorption of tetracyclines. drugs. 11 (1976) 45–54. [12] a.h. caswell, j.d. hutchison. selectivity of cation chelation to tetracyclines: evidence for special conformation of calcium chelate. biochem. biophys. res. commun. 43 (1971) 625–630. [13] k.w. kohn. mediation of divalent metal ions in the binding of tetracycline to macromolecules. nature 191 (1961) 1156–1158. [14] l. lambs, b. decock-le reverend, h. kozlowski, g. berthon. metal ion-tetracycline interactions in biological fluids. 9. circular dichroism spectra of calcium and magnesium complexes with tetracycline, oxytetracycline, doxycycline, and chlortetracycline and discussion of their binding modes. inorg. chem. 27 (1988) 3001–3012. [15] j.t. doluisio, a.n. martin. metal complexation of the tetracycline hydrochlorides. j. med. chem. 6 (1963) 16–20. admet & dmpk 8(2) (2020) 129-138 effect of metal ions on pampa permeation of tetracyclines doi: http://dx.doi.org/10.5599/admet.797 137 [16] m.o. schmitt, s. schneider. spectroscopic investigation of complexation between various tetracyclines and mg 2+ or ca 2+. physchemcomm. 3 (2000) 42–55. [17] r. pulicharla, k. hegde, s.k. brar, r.y. surampalli. tetracyclines metal complexation: significance and fate of mutual existence in the environment. environ. pollut. 221 (2017) 1–14. [18] s. banerjee, k. chakrabarti. the transport of tetracyclines across the mouse ileum in vitro: the effect of cations and other agents. j. pharm. pharmacol. 28 (1976) 133–138. [19] h. poiger, c. schlatter. interaction of cations and chelators with the intestinal absorption of tetracycline. naunyn. schmiedebergs. arch. pharmacol. 306 (1979) 89–92. [20] l. v allen jr, r.s. levinson, c. robinson, a. lau. effect of surfactant on tetracycline absorption across everted rat intestine. j. pharm. sci. 70 (1981) 269–271. [21] k. kakemi, h. sezaki, h. ogata, t. nadai. absorption and excretion of drugs. xxxvi. effect of ca2+ on the absorption of tetracycline from the small intestine (1). chem. pharm. bull. 16 (1968) 2200–2205. [22] k. kakemi, h. sezaki, m. hayashi, t. nadai. absorption and excretion of drugs. xxxvii. effect of ca2+ on the absorption of tetracycline from the small intestine (2). chem. pharm. bull. 16 (1968) 2206– 2212. [23] g.e. schumacher, e.e. linn. kinetic and thermodynamic aspects of in vitro interphase transfer of tetracyclines ii: influence of divalent metal salts. j. pharm. sci. 67 (1978) 1717–1720. [24] p.l. nicklin, w.j. irwin, i.f. hassan, m. mackay. development of a minimum-calcium caco-2 monolayer model: calcium and magnesium ions retard the transport of pamidronate. int. j. pharm. 123 (1995) 187–197. [25] d. el-sabawi, r. abu-dahab, w.a. zalloum, f. ijbara, i.i. hamdan. the effect of ferrous ions, calcium ions and citric acid on absorption of ciprofloxacin across caco-2 cells: practical and structural approach. drug dev. ind. pharm. 45 (2019) 292–303. [26] m. kansy, f. senner, k. gubernator. physicochemical high throughput screening: parallel artificial membrane permeation assay in the description of passive absorption processes. j. med. chem. 41 (1998) 1007–1010. [27] k. sugano, h. hamada, m. machida, h. ushio. high throughput prediction of oral absorption: improvement of the composition of the lipid solution used in parallel artificial membrane permeation assay. j. biomol. screen. 6 (2001) 189-196. [28] k. sugano, m. kansy, p. artursson, a. avdeef, s. bendels, l. di, g.f. ecker, b. faller, h. fischer, g. gerebtzoff, h. lennernaes, f. senner. coexistence of passive and carrier-mediated processes in drug transport. nat. rev. drug discov. 9 (2010) 597-614. [29] a. avdeef, s. bendels, l. di, b. faller, m. kansy, k. sugano, y. yamauchi. pampa critical factors for better predictions of absorption. j. pharm. sci. 96 (2007) 2893-2909. [30] s. yamauchi, k. sugano. permeation characteristics of tetracyclines in parallel artificial membrane permeation assay. admet dmpk 7 (2019) 151–160. [31] s. babić, a.j.m. horvat, d. mutavdžić pavlović, m. kaštelan-macan. determination of pka values of active pharmaceutical ingredients. trac trends anal. chem. 26 (2007) 1043–1061. [32] s. şanli, n. şanli, g. alsancak. determination of protonation constants of some tetracycline antibiotics by potentiometry and lc methods in water and acetonitrile-water binary mixtures. j. braz. chem. soc. 20 (2009) 939–946. [33] pmda. minocycline hydrochloride drug product information. (2011). https://www.info.pmda.go.jp/igo/pack/6152005f1052_2_11/. [34] a. avdeef. absorption and drug development: solubility, permeability, and charge state, second edition, john wiley & sons, inc., 2012. [35] b. faller, h.p. grimm, f. loeuillet-ritzler, s. arnold, x. briand. high-throughput lipophilicity measurement with immobilized artificial membranes. j. med. chem. 48 (2005) 2571–2576. [36] z. song, k. terada, k. sugano. reversed phase parallel artificial membrane permeation assay for log p measurement. admet dmpk 4 (2016) 54-59. http://dx.doi.org/10.5599/admet.797 kiyohiko sugano et al. admet & dmpk 8(2) (2020) 129-138 138 [37] mhlw. overview of dietary reference intakes for japanese (2015), 2015. https://www.mhlw.go.jp/file/06-seisakujouhou-10900000-kenkoukyoku/overview.pdf. [38] c. schiller, c.p. fröhlich, t. giessmann, w. siegmund, h. mönnikes, n. hosten, w. weitschies. intestinal fluid volumes and transit of dosage forms as assessed by magnetic resonance imaging. aliment. pharmacol. ther. 22 (2005) 971–979. [39] j.j. werner, w.a. arnold, k. mcneill. water hardness as a photochemical parameter: tetracycline photolysis as a function of calcium concentration, magnesium concentration, and ph. environ. sci. technol. 40 (2006) 7236–7241. [40] l. jin, x. amaya-mazo, m.e. apel, s.s. sankisa, e. johnson, m.a. zbyszynska, a. han. ca2+ and mg2+ bind tetracycline with distinct stoichiometries and linked deprotonation. biophys. chem. 128 (2007) 185–196. [41] d.c. maxwell, p.j.a. smith, s.p. wilford. stabilities of some alkaline earth chelates of tetracycline. nature 198 (1963) 577–578. [42] l.z. benet, j.e. goyan. thermodynamics of chelation by tetracyclines. j. pharm. sci. 55 (1966) 1184– 1190. [43] m. novák-pékli, m.e.-h. mesbah, g. pethö. equilibrium studies on tetracycline—metal ion systems. j. pharm. biomed. anal. 14 (1996) 1025–1029. [44] h.a. duarte, s. carvalho, e.b. paniago, a.m. simas. importance of tautomers in the chemical behavior of tetracyclines. j. pharm. sci. 88 (1999) 111–120. [45] a. pagliara, p.a. carrupt, g. caron, p. gaillard, b. testa. lipophilicity profiles of ampholytes. chem. rev. 97 (1997) 3385–3400. [46] h. terada, t. inagi. proposed partition mechanism of tetracycline. chem. pharm. bull. 23 (1975) 1960–1968. [47] s.r. martin. equilibrium and kinetic studies on the interaction of tetracyclines with calcium and magnesium. biophys. chem. 10 (1979) 319–326. [48] n.p. franks, m.h. abraham, w.r. lieb. molecular organization of liquid n-octanol: an x-ray diffraction analysis. j. pharm. sci. 82 (1993) 466–470. [49] m. barza, r.b. brown, c. shanks, c. gamble, l. weinstein. relation between lipophilicity and pharmacological behavior of minocycline, doxycycline, tetracycline, and oxytetracycline in dogs. antimicrob. agents chemother. 8 (1975) 713–720. [50] y. zhang, s.a. boyd, b.j. teppen, j.m. tiedje, h. li. role of tetracycline speciation in the bioavailability to escherichia coli for uptake and expression of antibiotic resistance. environ. sci. technol. 48 (2014) 4893–4900. ©2020 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) https://www.mhlw.go.jp/file/06-seisakujouhou-10900000-kenkoukyoku/overview.pdf https://www.mhlw.go.jp/file/06-seisakujouhou-10900000-kenkoukyoku/overview.pdf http://creativecommons.org/licenses/by/3.0/ a physiologically-based pharmacokinetic model of oseltamivir phosphate and its carboxylate metabolite for rats and humans doi: 10.5599/admet.628 22 admet & dmpk 7(1) (2019) 22-43; doi: http://dx.doi.org/10.5599/admet.628 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper a physiologically-based pharmacokinetic model of oseltamivir phosphate and its carboxylate metabolite for rats and humans guanghua gao 1,3 *, francis law 2 , ricky ngok shun wong 3 , nai ki mak 3 , mildred sze ming yang 3 1 center for drug evaluation，china food and drug administration，beijing 100022，china; 2 department of biological sciences, simon fraser university, burnaby, b.c. canada v3n 4w8; 3 department of biology, hong kong baptist university, kowloon, hong kong. saz. china. *corresponding author: e-mail: gaogh@cde.org.cn; tel.: 010-85243575 received: october 30, 2018; revised: january 31, 2019; published: february 23, 2019 abstract oseltamivir phosphate (op, tamiflu®) is a widely used prodrug for the treatment of influenza viral infections. orally administered op is rapidly hydrolyzed by the carboxylesterases in animals to oseltamivir carboxylate (oc), a potent influenza virus neuraminidase inhibitor. the goals of this study were to develop and validate a physiologically-based pharmacokinetic (pbpk) model of op/oc in rats and humans, and to predict the internal tissue doses for op and oc in humans after receiving op orally. to this end, a pbpk model of op/oc was first developed in the rat, which was then scaled up to humans by replacing the physiological and biochemical parameters with human-specific values. the proposed pbpk model consisted of an op and an oc sub-models each containing nine first-order, flow-limited tissue/organ compartments. op metabolism to oc was assumed to carry out mainly by hepatic carboxylesterases although extrahepatic metabolism also occurred especially in the plasma. the pbpk model was developed and validated by experimental data from our laboratories and from the literature. the proposed pbpk model accurately predicted the pharmacokinetic behavior of op and oc in humans and rats after receiving a single or multiple doses of op orally or an oc dose i.v. the pbpk model was used to predict the internal tissue doses of op and oc in a hypothetical human after receiving the recommended dose of 75 mg/kg op b.i.d. for 6 days. steady-state oc concentrations in the plasma and major organs such as the lung and the brain were higher than the minimum in vitro ic50 reported for h1n1 influenza virus neuraminidase, confirming op is an effective, anti-viral agent. op side-effects in the gastrointestinal tract and brain of humans were explainable by the tissue doses found in these organs. the pbpk model provides a quantitative tool to evaluate the relationship between an externally applied dose of op and the internal tissue doses in humans. as such the model can be used to adjust the dose regimens for adult patients in disease states e.g., renal failure and liver damage. keywords oseltamivir; physiologically-based pharmacokinetic modelling; tissue concentration; dose metrics introduction oseltamivir phosphate (op, tamiflu®) is an effective prodrug for the prevention and treatment of influenza viral infection [1]. after oral dosing, op is rapidly absorbed into the systemic circulation of humans and hydrolyzed by the carboxylesterase enzymes in different tissues and organs to oseltamivir carboxylate (oc), a potent inhibitor of the neuraminidase enzyme [2] (fig. 1). because influenza virus needs http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:gaogh@cde.org.cn admet & dmpk 7(1) (2019) 22-43 a physiologically-based pharmacokinetic model for oseltamivir doi: 10.5599/admet.628 23 neuraminidase to grow and duplicate, oc is able to reduce the duration and severity of influenza viral infection in humans [3]. op is a drug with a wide margin of safety. however, adverse events such as abdominal pain, diarrhea, nausea, and skin reaction have been reported during early medication periods especially if the oral dose is >200 mg op b.i.d. [3,4]. thus neurological symptoms such as headache, vertigo, somnolence, insomnia, numbness, hyper-excitation and an increase in suicidal tendency also have been reported in japanese children and young adolescents [5,6]. the mechanism(s) by which op causes neuropsychiatric effects remains unclear: toovey et al. [7] have suggested that since op and oc are p-glycoprotein substrates [8], little or no op/oc is able to penetrate the blood-brain barrier of humans and cause neuropsychiatric effects [9]. conversely, op has been shown to block ganglioside-regulated excitatory opioid receptormediated hyperalgesia and increase opioid analgesia in mice [10,11]. thus, op or oc may alter the neuronal excitability of the brain by increasing the synchronization of hippocampus network [12]. it is also possible that op/oc change human behaviors if the blood-brain barrier is immature or damaged as in the case of young children. op metabolic disposition has been studied in rats [13], ferrets [14] and mice [15]. the oral bioavailability of op in these animals ranges from 30-73 % [16] but in humans, it may reach 80 %. therapeutic levels of oc have been found in the lungs, trachea, nasal mucosa of ferrets [14], and the middle ear and sinuses of humans [17]. however, little or no information is available on the dose metrics of op and oc in the internal tissues/organs of humans. andersen [18] has suggested using the physiologically based pharmacokinetic (pbpk) model to predict the pharmacokinetics and tissue doses of environmental chemicals in humans. moreover, by determining the relationship between the internal tissue dose (instead of the external dose) and the pharmacological or toxicological effects of xenobiotics, we can greatly improve the accuracy of the dose-response relationship of prescription drugs and chemicals [19]. o nh 2 o n h o o oh nh 2 o n h o o esterase oseltamivir phosphate (op) oseltamivir carboxylate (oc) figure 1. hydrolysis of oseltamivir phosphate to oseltamivir carboxylate hitherto, at least two different op pbpk models have been reported for neonates, infants and adult humans [20,21]. none of these models have been used to predict an internal tissue dose of op/oc in humans. parrott et al. [20] first develop an oc pbpk model in marmoset monkeys and then extrapolate the model to young children. the pbpk model was developed using a multi-compartment permeability-limited liver. due to a lack of tissue distribution data in monkeys and humans, the pbpk model needs refinement and improvement. the aims of this study were to develop and validate a pbpk model for op/oc in rats, to scale up the rat model to humans, and to predict the internal tissue doses of op/oc in rats and humans using peak concentration (cmax) and/or area under the curve (aucplasma) of concentration-time curve as the dose g. gao et al. admet & dmpk 7(1) (2019) 22-43 24 surrogates. attempts were made to explain the therapeutic and/or side effects of op in humans using the model-predicted dose metrics. since our op/oc pbpk model is based on the actual anatomy and physiology of rats (or humans) and the physicochemical properties of op/oc, it should be useful in predicting the internal tissue doses of op/oc across the external doses, administration routes, and animal species [22]. materials and methods metabolic disposition of op in the rat chemicals. op/oc standards were kindly donated by hoffmann-la roche ltd., switzerland. chemical purity of both chemicals exceeded 99 %. dichlorvos was purchased from supelco (bellfonte, pa, us). analytical grade diethyl ether and hplc grade acetonitrile were purchased from labscan asia co. ltd (bangkok, thailand). animals. male sprague-dawley rats weighing about 250 g were purchased from the chinese university of hong kong, hong kong. the animals were maintained on a constant light-dark cycle with light from 07:00 to 19:00 h and darkness from 19:00 to 07:00 h. tap water and food were provided ad libitum. the rats were used after a 7-day acclimation period. the procedure associated with animal care and experimentation was carried out according to the hong kong council on animal care guidelines and with formal approval of the animal care committee of the hong kong baptist university. tissue distribution studies. thirty rats were used in the tissue distribution studies. the rats were assigned randomly into two groups of 15 rats each. after fasting overnight, one group of rats were given a single dose of 10 mg/kg op by gavage. the other group of rats received 50 mg/kg op. three rats from each group were selected randomly and euthanized by diethyl ether at the following time points: 0.5, 2, 3, 6, and 12 h post-dosing for the 10 mg/kg dose group; 0.5, 2, 3, 6 and 10 h post-dosing for the 50 mg/kg dose group. the abdominal wall was opened and a blood sample was withdrawn by cardiac puncture. the blood samples were centrifuged immediately at 3000 rpm for 10 min to collect the plasma. the plasma was put in a centrifuge tube containing a mixture of sodium heparin and dichlorvos which was used to inhibit op hydrolysis in the sample [23]. the liver, spleen, lung, and kidney were also removed from the rats. the tissue samples were rinsed briefly with distilled water, wiped dry with kimwipes, and stored at -80 °c until analysis. extraction of op and oc from plasma and tissue samples. one gram of tissue/organ (0.5 g for spleen) was weighed accurately in a test tube. after adding distilled water (3 ml) and dichlorvos (600 µg), the tissue/organ was homogenized in a kinemetica gmbh homogenizer (pcu-2-110, switzerland) for about 2 min. a 0.2 ml aliquot of the tissue homogenate (or plasma) was added to 0.8 ml of acetonitrile/water (3:1, v/v) in an eppendorf tube. the content in the eppendorf tube was mixed by vortex for 1 min and then centrifuged at 13000 rpm for 10 min. the supernatant was removed and stored at -80 °c until analysis. lc/ms/ms analysis of op and oc. frozen tissue extracts were thawed, diluted with the mobile phase, and injected 10.0 µl of the final solution into the hplc. op/oc in the extracts was quantified using a modified lc/ms/ms procedure reported by lindegardh et al. [24]. briefly, op/oc separation was performed on a c18 analytical column (ods-2, 150mm, 5u, waters corp. milford, ma) in a hp 1100 system (hewlett packard, santa clara, california) preceded by a guard column (2.1 by 5 mm; particle size, 1.7 µm) (waters corp. milford, ma). the column was maintained at 26 °c; it was eluted isocratically at a flow rate of 0.200 ml/min. the hplc mobile phase consisted of a solution with an equal volume of ammonium admet & dmpk 7(1) (2019) 22-43 a physiologically-based pharmacokinetic model for oseltamivir doi: 10.5599/admet.628 25 formate (8.0 mm) and methanol. the solution was adjusted to ph 3.50 with formic acid. the analysis employed an api 3200tm triple quadrupole mass spectrometer (ap scienx) equipped with a turbo vtm ion source controlled by the linac® collision cell technology for output scanning. the optimal transitions were 313.14 m/z to 224.92 m/z for op, 285.2 m/z to 197.0 m/z for oc. the limits of detection for op and oc were 20 and 30 pg/ml, respectively. the limits of quantification ranged from 0.4-0.6 ng/ml for plasma and 1.6-2.4 ng/g for tissues. recoveries were achieved by comparing op/oc concentration in the extract to the theoretical concentration; they ranged from 96 % to 101 % in the plasma/tissue samples. no obvious matrix effects were detected in the tissues within the op/oc concentration ranges used in this study. blood to plasma ratio. about 2 ml of freshly drawn rat blood was mixed with a solution containing of 100 µg/ml op (or oc) standard and 200 µg/ml of dichlorvos. preliminary studies indicated that equilibrium between the plasma and blood for op/oc was established within 20 min. after incubating the whole blood at 37 °c for 30 min, an aliquot was removed and centrifuged at 3000 rpm for 10 min to obtain the plasma. op and oc concentrations in the plasma were measured by lc/ms/ms as described above. op and oc concentrations in the blood were assumed to be the theoretical concentrations. blood/plasma concentration ratio (blplr) was calculated by dividing the concentration of op (or oc) in the whole blood with that in the plasma. the blplr of op and oc was about 1.0 which indicated op and oc had similar protein binding capacities to blood and plasma. op metabolism by rat plasma in vitro. in vitro op metabolism was studied by incubating different concentrations of op solution (0.1-0.5 mg/ml) with 30 µl rat plasma and tris-hcl buffer (ph 7.4) in a shaking water bath maintained at 37 °c. the final incubation volume was 200 µl. control incubation was carried out in a similar manner but without rat plasma. at 1 h after incubation, ice-cold acetonitrile (400 µl) was used to terminate the hydrolysis reaction. the mixture was centrifuged at 13200 rpm in a 4 °c centrifuge (model 5415r, eppendorf ag, barkhausenweg 122339 hamburg, germany) for 10 min to separate into layers. the supernatant was transferred to an hplc vial and analyzed by a hewlett packard 1100 hplc equipped with a uv detector. op and oc were separated on an hplc column (ods-2, 150 mm, 5 µm) using gradient elution. the mobile phase consisted of a solution of 0.4 % phosphoric acid (ph 3.0) and acetonitrile (80:20, v/v). acetonitrile in the mobile phase increased linearly from 20 % to 40 % in 10 min. the flow rate of the mobile phase was 1 ml/min. the detection wavelength was set at 215 nm. op metabolism rate was expressed as µg oc formed/ml plasma/h. the vmax and km of op metabolism were determined using the lineweaver-burke plot and were found to be 61.2 mg/h and 300 mg/l, respectively. pbpk modeling of op/oc in rats and humans model structure. fig. 2 shows a schematic diagram of the op/oc pbpk model in rats after oral administration. the pbpk model actually consisted of an op sub-model and an oc sub-model. each submodel was composed of 9 first-order flow-limited tissue compartments: the lung, kidney, muscle, brain, liver, spleen, gut, blood and rest of body. the brain was lumped with the fat and the lung was lumped with the heart because of possible similar tissue/plasma partition coefficients. the rest of body compartment included all other tissues which had not been identified in the model i.e., the skin, bone, eye, prostate gland, etc. the lung and the brain are modeled as separate compartments because h5n1 influenza virus mainly affected the respiratory tract and the central nervous system of japanese children after oral administration of op [5,6]. absorption of op was assumed to be a linear, first-order input process. after absorption, op was hydrolyzed to oc mainly by the carboxylesterases in the plasma and the liver. for the purpose of model development, op metabolism was assumed to occur only in the liver which linked the op and oc sub-models together (fig. 2). also, one mole of op was converted by the carboxylesterase g. gao et al. admet & dmpk 7(1) (2019) 22-43 26 enzymes to one mole of oc. both op and oc were excreted into the urine by the kidney. figure 2. schematic diagram of the physiologically-based pharmacokinetic model for rats or humans after op administration. c represents op or oc concentrations (ng/g or ng/ml); q represents plasma flow rates (l/h). the subscripts (p or c) under specific tissue concentration c in the sub-models refer to op or oc, respectively physiological parameters. the physiological parameters of the rat pbpk model were parameterized a priori (table 1) as follows: tissue volume and blood flow were taken from luttringer et al. [25] and from davies and morris [26]; these parameter values were expressed as the percentage of total body volume or cardiac output (co). total blood volume was divided into a two-thirds venous pool and one-third arterial pool. gut content volumes was assumed to be 0.014 l for a 0.25 kg rat [27]. pharmacokinetic parameters. absorption rate constant (ka) of op was determined to be 0.76 h -1 which was obtained by fitting the oc concentration-time curve in plasma following oral administration of 10 mg/kg of op to a classical, one-compartment pharmacokinetic model. considering the uncertainty of the absorption related parameters, f and ka were allowed to be optimized to a reasonable extent during the modelling process based on the individual experimental datasets (table 2). the oral bioavailability of op was reported to be 0.35 calculated by aucp.o. (op)/auci.v.(oc) ratio, it should be noticed that in the pbpk model f represents the oral bioavailability factor which was derived from the integrative modeling process of the pbpk model for all organs [28] and might be different from the experimental absolute bioavailability (aucp.o./auci.v.). the clearance (clc) of oc from rats was reported to be 0.38 l/h after injecting a single dose of 10 mg/kg intravenously [15,16]. in vitro km (300 mg/l) and vmax (61.2 mg/h) values were obtained by incubating op with rat plasma. fecal excretion rate constant (kf) was assumed to be 1/transit time of the small intestine; the transit time of the small intestine was taken from davies and morris [26]. admet & dmpk 7(1) (2019) 22-43 a physiologically-based pharmacokinetic model for oseltamivir doi: 10.5599/admet.628 27 table 1. physiological parameters of the op/oc pbpk model in rats and humans * parameter organ weight (% of body weight) regional blood flow (% of cardiac output) partition coefficients tissue rat human rat human op oc adipose 7.6 12 7 5 brain 0.57 0.2 2 12 lumped brain 8.13 12.2 9 17 0.5 0.25 heart 0.33 0.47 lung 0.5 0.76 lumped lung 0.83 1.23 2.5 0.5 kidney 0.73 0.44 14.1 19 5 8 liver 3.66 2.57 32.6 25 3 5 gi tract 2.7 1.7 13.1 17 10 1 muscle 40.4 40 27.8 17 1.4 0.5 spleen 0.2 0.26 2 2 1.5 0.5 the rest of body ** 43.4 41.6 16.5 22 3 1 total 100 100 100 100 * mean data on tissue blood flow rate and volume were adapted from luttringer et al. [25]. organ weight was based on the mean bw of a 0.26 kg rat and a 70 kg human; cardiac output for a 0.25 kg rat is 4.99 l/h; it is calculated from the equation, 14.1*bw 0.75 (l/h). cardiac output for a 70 kg human is 390 l/h; it is calculated from the equation, 16.1*bw 0.75 (l/h). ** the brain was lumped into the adipose tissues to form the “brain compartment”; the lung was lumped into the heart to form the “lung compartment”; the remaining organs/tissues including the bone, skin, eye, etc., were grouped together to form “the rest of body”. table 2. pharmacokinetic parameters for the op pbpk model of rats and humans chemical specific parameters molecular weight oseltamivir phosphate salt (op):410.4 oseltamivir carboxylate (oc): 284.35 used in model unbound fraction in plasma op: 53.7 % (rat); 57 (human); oc: no data (rat); 97 % (human) not used in model, for reference pk parameters rat human parameter experimental optimized experimental optimized f 0.35 0.8-0.86 0.75 0.75 tlag (h) na 0.32-0.44 na 0.44 ka (1/h) 0.76 12-15 na 0.35 (0.14-0.28) *** blplrp 1 1 na 1 blplrc 1 1 (0.7) ** na 1 (0.6)** clc (l/h) 0.38 0.38 (0.5) ** 18 18(30)** clp (l/h) na 0.4 20 100 km (mg/l) 300 300 76.0**** 76.0**** vmax (mg/h)* 61.2 61.2 2.6 x 10 5 **** 2.6 x 10 5 **** *vmax was based on a 0.25 kg rat and a 70 kg human **bracketed parameter values were used for the i.v. route of administration ***bracketed parameter values were used for multiple oral administration; they were obtained by optimizing the plasma concentration-time data from he et al. [14]. ****shi et al [2] , km =187 µm and vmax=114 nmol/mg/min in this paper, the final vmax was calculated with 70 kg bw, 1.8 kg of liver weight and 52.5 mg microsomes/gram of liver for a standard human adult. na: no result. g. gao et al. admet & dmpk 7(1) (2019) 22-43 28 tissue/plasma partition coefficients. these were estimated by applying the area method of gallo et al. [29] to the experimental data of rats after receiving a single oral dose of 10 mg/kg op (table 1). briefly, the experimental op/oc concentration-time curves in the plasma and organ/tissue were analyzed separately with the non-compartmental approach (winnonlin®, scientific consulting, inc. version 1.5) to obtain the area under the concentration-time curve (auc). the in vivo tissue/plasma partition coefficient for a specific organ was estimated from the auctissue/aucplasma ratio, table 1, lists the final tissue/plasma partitioning coefficients which had been adjusted to available experimental data. in rat, plasma protein binding for op was reported to be 53.7 % [30], similar binding capacity in blood as in plasma was found in our study for both op and oc. the tissue/plasma partition coefficients basically agreed with the lipid contents of the tissues and the lipid solubility of op or oc. model simulation. the differential and algebraic equations describing the movement of op and oc through the rat or human were formulated as a computer program. mass balance differential equations for the model and the definition of the algebraic terms are given as parts of the model structure (fig. 1 and appendix). after incorporating the parameter values (tables 1 and 2) into the model (fig. 1), the differential and algebraic equations were solved numerically with the aid of acslxtreme®2.4.2.1 (aegis technologies group, inc., huntsville, al). an i.v. injection function instead of oral absorption was used to drive the model when op/oc was administered intravenously [31]. scaling up of rat pbpk models to humans. the rat model (fig. 2) was scaled up to humans by replacing the physiological parameters and pharmacokinetic parameters in the rat model with human specific values (table 1). the human pbpk model was parameterized as follows: (a) tissue volumes and blood flows to tissues were taken from the literature [25,26], (b) tissue/plasma partitioning coefficients were assumed to be the same as those of rats, (c) f was set to 75 % based on the absolute bioavailability reported in humans [14,32], (d) plasma protein binding of oc and op in humans was 3 % and 43 %, respectively [4], (d) gut content volumes was scaled based on the body weights of rats and humans; the gut content of a 70 kg human was determined to be 0.16 l [26], (e) clearance of oc (clc) in urine was reported to be 18 l/h [14,33]. clearance value for op (clp) was scaled from the clearance of rats using the allometric equation, clbc = clb/(bw) 0.66 [34] initially, it was reported to be 20 l/h while cl/f was 0.438 l/min [14], considering the uncertainty of absorption parameters and the difference between classical model and pbpk model, clp was allowed to be optimized using the plasma concentration-time curve reported by he et al. [14], (f) in vitro km and vmax of op metabolism in human liver microsomal incubates were reported to be 187 µm and 114 nmol/mg/min, respectively [2]; these translated into in vivo km and vmax values of 76 mg/l and about 2.6x10 5 mg/h, respectively for humans based on an average microsomal yield of 52.5 mg microsomal proteins per g liver [14,31,35], and (g) ka and kf rate constants (table 2) were scaled up from rats using the equation, kc=ka/(bw) -0.3 , where ka represented the rate constant, kc represented the scaling coefficient and bw was the average body weight of human volunteers [36]. model parameters that could not be parameterized a priori were optimized by fitting the pbpk model to available experimental data (see data fitting below). final parameter values used to implement the human pbpk model are listed in table 3. data fitting and parameter optimization. the parameters that could not be obtained a priori or determined accurately were optimized by fitting the pbpk model to available data. optimization was accomplished by minimizing the squared differences between model prediction and the experimental data using the maximized log likelihood function of the acslxtreme® optstat program (aegis technologies group, inc., huntsville, al). only the parameters that needed to be parameterized were varied, the values of other parameters in tables 1 and 2 were kept constant. admet & dmpk 7(1) (2019) 22-43 a physiologically-based pharmacokinetic model for oseltamivir doi: 10.5599/admet.628 29 table 3. normalized sensitivity coefficients for oc concentrations in the lung and the plasma based on the simulated data of administering 10 mg/kg op p.o. to the rat* pulmonary oc concentration-time data as the model output surrogate time (h) parameter 0.5 1 2 4 6 8 10 12 high (>0.5) f 0.03 0.67 0.44 0.11 0.02 -** tlag -0.59 -0.03 0.12 0.04 0.01 medium vmax 0.03 0.50 0.25 0.03 km -0.03 -0.49 -0.25 -0.03 low (<0.2) clc -0.09 -0.18 -0.09 -0.02 -0.01 clp -0.02 -0.06 -0.04 -0.02 ka 0.03 0.03 -0.02 -0.01 blplrc 0*** 0 0 0 0 0 0 0 blplrp 0 0 0 0 0 0 0 0 plasma oc concentration-time data as the model output surrogate high (>0.5) blplrc -0.03 -0.68 -0.45 -0.11 -0.02 f 0.03 0.67 0.44 0.11 0.02 tlag -0.59 -0.03 0.12 0.04 0.01 medium vmax 0.03 0.50 0.25 0.03 km -0.03 -0.49 -0.25 -0.03 low (<0.2) clc -0.09 -0.18 -0.09 -0.02 -0.01 clp -0.02 -0.06 -0.04 -0.01 ka 0.03 0.03 -0.03 -0.01 blplrp 0 0 0 0 0 0 0 0 * the sensitivity coefficients were determined using the parameter values from rats dosed with a single oral dose of 10 mg/kg op (table 2) **absolute value ≤0.001 ***: 0 means a change in parameter value did not result in any significant effect on model outputs validation and calibration of pbpk model. the pbpk models of rats and humans were validated or calibrated by comparing model-simulated results with experimental data not used in model development. the empirical data were obtained either from our own laboratories or from the literature. because we were unable to obtain the original concentration-time data from the literature, they were read digitally from the publications using digimatic® (windows version 2.2c, feb software, chesterfield, virginia). the pbpk model was assumed to be validated if model simulation described closely the experimental data. thus, the rat pbpk model was validated using the experimental data from rats dosed with a single oral dose of 50 mg/kg op. the rat model was further validated using the plasma concentration-time data from rats after receiving a single oral dose of 30 mg/kg op or an i.v. infusion of 30 mg/kg oc [37]. similarly, the human pbpk model was calibrated using the datasets reported by he et al. [14] for human volunteers dosed with a single oral dose of 150 mg op or an i.v. infusion of 150 mg oc. the human model was further validated based on the experimental data of humans receiving multiple oral doses of 50, 100, 200 or 500 mg of op b.i.d. for one week. statistical and sensitivity analyses. mean absolute prediction error (mape) was used as a measure of good fit between model-predicted concentration (cpredi) and experimental concentration (cexpti). it was g. gao et al. admet & dmpk 7(1) (2019) 22-43 30 calculated using the following equation: exp exp 1 mape (%) (100 / ) / n ti predi ti i n c c c         (1) where, i represents individual data points; n is the total number of data points. a deviation within a factor of two between predicted and experimental concentration data (i.e., mape<50 %) was used as the criteria for goodness of fit [38]. log-normalized sensitivity parameter (lsp) was used to identify key model parameters that had significant impacts on model prediction. lsp was defined as the partial derivatives of model response to the corresponding model parameter [39]: lsp = ∂ln r/∂ln x (2) where r is the model output and x is the parameter for which the sensitivity is assessed. the sensitivity analysis was conducted by acslxtreme 2.0.1.2 (aegis technologies group, inc., huntsville, al) using the central difference method. the lsp represented the percentage change in an output value associated with the percentage change in the input parameter. the plasma and lung oc concentrations were chosen as the target tissues for model responses following oral administration of 10 mg/kg of op to rats. model parameters were changed individually by 0.1 % to assess their impact on model prediction of plasma and lung concentrations. for each model response, magnitude, time dependency, and the mathematical sign of each sensitivity coefficient were examined. the ranking of effects of input parameters on the model prediction was based on the absolute magnitude of the sensitivity coefficient at the time [40]. generally, the parameters that have large absolute sensitivity coefficients often cause sensitive model outputs in response to a small change in their values because they have a high probability for prediction error to occur due to inaccuracy of the parameter values. results development and validation of rat pbpk model table 2 summarizes the final parameter values used to implement the rat and human pbpk models. the oral bioavailability factor (f) of op was about 80 % for both rats and humans. plasma op clearances (clp) for humans and rats were 100 l/h and 0.4 l/h, respectively. if the clp values were normalized by their respective body weights, the clp of humans (1.4 l/kg/h) and rats (1.6 l/kg/h) were nearly identical. in contrast, plasma oc clearances (clc) of rats and humans were very different even after they were normalized by the body weight; the clc were 18 l/h for 70kg bw human adult (0.26 l/kg/h) and 0.38 l/h for 0.25 kg bw rat (1.5 l/kg/h), respectively (table 2). the km and vmax values of op biotransformation were 61.2 mg/h (245 mg/kg/h) and 300 mg/l for rats, and 2.6x10 5 mg/h (3714 mg/kg/h for 70 kg bw) and 76 mg/l for humans. op/oc concentration-time profiles in plasma/tissues of rats op was metabolized primarily to oc in rats. thus, mostly oc and occasionally low concentrations of op were found in the plasma (figs. 3 and 4) although both op and oc were found in the major organs of rats. the cmax of oc in the plasma were 0.80 and 5.63 µg/ml, respectively after receiving 10 mg/kg and 50 mg/kg of op by the rat. it decreased in the order of liver > kidney > lung > plasma (figs. 3 and 4). although op was found at a higher concentration than oc in the lung (figs. 3 and 4) and spleen (data not shown), op concentration was lower than oc concentration in kidney, liver (figs. 3 and 4), and muscle (data not admet & dmpk 7(1) (2019) 22-43 a physiologically-based pharmacokinetic model for oseltamivir doi: 10.5599/admet.628 31 shown). the auc0-4h of oc in the lungs were 5.2 and 26.1 µg·h/ml, respectively when rats were treated with an oral dose of 10 mg/kg and 50 mg/kg op. these results showed that the dose of oc delivered to the lungs increased in proportion to the op dose applied externally, a finding which was consistent with the assumption of linear kinetics in the pbpk model. (a) liver (b) kidney (c) lung (d) plasma (e) brain figure 3. time course of op and oc concentrations in the plasma and tissues of rats after administering a single oral dose of 10 mg/kg op. (a) liver, (b) kidney, (c) lung, (d) plasma, (e) brain. data points, means + s.d. of the experimental plasma/tissue concentrations from three different rats; (___) model-simulated op concentrations; (— —) model-simulated oc concentrations g. gao et al. admet & dmpk 7(1) (2019) 22-43 32 (a) liver (b) kidney (c) lung (d) plasma (e) brain figure 4. time course of op and oc concentrations in the plasma and tissues of rats after administering a single oral dose of 50 mg/kg op. (a) liver, (b) kidney, (c) lung, (d) plasma, (e) brain. data points, means + s.d. of the experimental plasma/tissue concentrations from three different rats; (___) model-simulated op concentrations; (— —) model-simulated oc concentrations the rat pbpk model was developed based on the experimental data of the 10 mg/kg treated rats (fig. 3). it was subsequently validated using the experimental data of the 50 mg/kg treated rats (fig. 4). in order to fit the experimental data well, however, the ka and f values of the pbpk model for the 50 mg/kg treated rats had to be higher than those of the 10 mg/kg treated rats, suggesting that op absorption was more rapid and extensive in the 50 mg/kg treated rats than the 10 mg/kg treated rats (table 2). the rat pbpk model was also validated using the plasma oc concentration-time data reported by eisenberg et al. [37]. the rat pbpk model was able to predict accurately the experimental data (fig. 5) and the tissue distribution data of the 10 mg/kg-treated (fig. 3) and 50 mg/kg-treated rats (fig. 4). the pbpk model also was used to simulate the kinetic profiles of op/oc in the rat’s brain (figs. 3e and 4e). for rats dosed with 10 mg/kg op, admet & dmpk 7(1) (2019) 22-43 a physiologically-based pharmacokinetic model for oseltamivir doi: 10.5599/admet.628 33 the predicted cmax for op and oc in the brain were 1.25 µg/g and 0.34 µg/g, respectively. for rats dosed with 50 mg/kg op, the predicted cmax for op and oc were 6.64 µg/g and 1.65 µg/g, respectively. these results were in agreement with the other rat tissue/organs in which op/oc tissue doses were found to increase linearly with the applied op dose (figs. 3 and 4). we also performed route-to-route extrapolation by comparing model simulation with the experimental data reported by eisenberg et al. [37]. the pbpk model was able to describe closely the plasma op/oc concentration-time profiles of rats receiving either a single dose of 30 mg/kg op orally or an i.v. infusion of 30 mg/kg oc (fig. 5). sensitivity analysis of rat pbpk model the sensitivity analysis provided a quantitative means to evaluate how the various input parameters might affect the oc dose metrics in the lungs and plasma when the rats were treated with a single dose of 10 mg/kg op orally. table 3 shows a summary of the sensitivity analysis results. the sensitivity coefficients were a mix of positive and negative values which varied with time over the 12 h post-dosing period. we arbitrarily grouped the parameters into 3 different categories according to their absolute sensitivity coefficients: high (>0.5), medium (0.2-0.5), and low (<0.2). based on these groupings, f and tlag were classified as high impact parameters, km and vmax were classified as medium impact parameters, and clc, clp, ka and blplrp were classified as low impact parameters for the simulated oc dose metrics in the plasma and lungs. interestingly, blplrc had a high impact on the plasma dose metric but a low impact on the pulmonary dose metric. because the f value was positive over the 12 h post-dosing period, oc concentrations in the lung and plasma would increase with an increasing f value. because both of the clc and clp sensitivity coefficients had negative values, oc concentrations in the lung and plasma would decrease with increasing clc and clp values. (a) p.o. administration (b) i.v. injection figure 5. plasma oc concentration-time profiles in rats after administering (a) a single oral dose of 30 mg/g op, (b) an i.v. injection of 30 mg/kg oc. data points, means + s.d. of the experimental plasma/tissue concentrations. (— —) model-simulated oc concentrations. experimental data were adapted from eisenberg et al. [37] calibration of the human pbpk model the human pbpk model was calibrated using the plasma concentration-time data from humans receiving either a single oral dose of 150 mg op or an i.v. infusion of 150 mg oc [14]. fig. 6 shows the observed concentrations together with the simulated concentrations for humans after p.o. administration of op. fig. 7 shows the model simulation and the experimental concentration-time profiles of plasma oc in humans during and after a 3-hour i.v. infusion of 150 mg oc. the pbpk model was able to describe the experimental data for both administration routes reasonably well. the optimized f value for human model g. gao et al. admet & dmpk 7(1) (2019) 22-43 34 was 0.75 (table 2) which showed extensive absorption of orally administered op. simulated cmax for op and oc in the plasma was 0.08 and 0.44 µg/ml, respectively, after dosing humans with 150 mg op orally (fig.6). the time to reach peak plasma concentration (tmax) for op and oc was 1 and 4 h, respectively. these results indicated that oc could remain in the plasma for a longer period than op. following oral administration of op, plasma oc concentrations in humans were always >40 ng/ml after the initial absorption phase (fig. 6), which was much higher than the minimum ic50 of neuraminidase inhibition levels (3-10 ng/ml) reported previously by davies [26]. figure 6. typical op and oc concentration-time profiles in the plasma of human volunteers receiving an oral dose of 150 mg op. data points represent means + s.d. of the experimental data which are adapted from he et al., [14]. (___) model-simulated op concentrations; (— —) model-simulated oc concentrations figure 7. a typical oc concentration-time profile in the plasma of human volunteers receiving an i.v. infusion of 150 mg oc. data points represent means + s.d. of the experimental data which are adapted from he et al., [14]. (— —) modelsimulated oc concentrations prediction of tissue/plasma op/oc concentrations in humans receiving multiple op dose regimens the pbpk model was used to simulate the kinetics of plasma oc in humans after receiving multiple oral dose regimens of op [14]. in their studies, the human volunteers were dosed with 50, 100, 200 and 500 mg op b.i.d. for 6 days. the model-predicted cmax/cmin ratios for oc in the plasma were 3.5 (218 ng/ml : 62 ng/ml), 2.5 (400 ng/ml : 158 ng/ml), 3.0 (838 ng/ml : 279 ng/ml) and 1.9 (1846 ng/ml : 959 ng/ml), respectively (fig. 8). the cmax/cmin ratio decreases with an increasing external op dose except for the 100 mg b.i.d dose, which indicates less oc concentration fluctuation in the plasma at a higher op dose. in addition, the pbpk model was used to predict the internal tissue dose of oc in the brain and the lung of these studies. the steady-state oc doses/concentrations predicted for the lungs and brain were found to be lower than that of the plasma. when the dose metrics of oc were expressed in auc0-6day, they were 21.4, 42.5, 85.3 and 210.2 µg/ml·h, respectively, for the 50, 100, 200, and 500 mg doses. the corresponding oc dose metrics for the lungs were 10.7, 21.2, 42.6, 105.0 µg/ml·h, respectively. for the brain, they were 5.3, 10.6, 21.3, 52.5 µg/ml·h, respectively. thus, the dose metrics of oc in these organs/tissues generally decrease in the order of plasma > lung > brain although they were all linearly related to the externally applied op dose. admet & dmpk 7(1) (2019) 22-43 a physiologically-based pharmacokinetic model for oseltamivir doi: 10.5599/admet.628 35 figure 8. plasma oc concentration-time profiles in human volunteers receiving multiple oral dose regimens of 50, 100, 200 or 500 mg op b.i.d. for 6 days. data points represent means + s.d. of the experimental plasma oc concentrations which are adapted from he et al, (17). (___) model-simulated oc concentrations we have also used the human pbpk model to simulate oc (fig. 9) and op (fig. 10) dose metrics in a hypothetical human after receiving the recommended dose regimen of 75 mg op b.i.d. for 6 days. fig. 9 shows the simulated oc concentrations in the plasma and major organs, which decreases in the order of liver > kidney > plasma ≈ the rest of body > lung > brain ≈ fat. the mean oc concentrations predicted for the lung and brain were ≥0.052 µg/g and ≥ 0.026µg/g, respectively. fig 10 shows the simulated op concentrations in the human plasma/organs which decreased in the order of liver > lung > kidney > the rest of body > muscle ≈ spleen > plasma > brain ≈ fat. the predicted mean op concentrations in the gut was high (3.72 µg/g) indicating a significant amount of op was not absorbed. figure 9. simulated oc concentration-time profiles in the plasma and tissues of a hypothetical 70-kg human receiving 75 mg op b.i.d. for 6 days. (___) model-simulated oc concentrations figure 10. simulated op concentration-time profiles in the plasma and tissues of a hypothetical 70-kg human receiving 75 mg op b.i.d. for 6 days(___) model-simulated op concentrations discussion previous studies have shown rats and humans metabolize op differently under in vitro incubation conditions; op is hydrolyzed primarily to oc by the carboxylesterases in rat plasma but is not hydrolysed by the carboxylesterases in human plasma [23]. also, op is metabolized by the hepatic cyp450 system of rats to minor metabolites [13]. in contrast, although op is metabolized to oc by the hepatic carboxylesterases in humans, the cyp450 system does not seem to involve in its biotransformation [14].thus, although op g. gao et al. admet & dmpk 7(1) (2019) 22-43 36 and oc are found in the plasma of humans after administering op orally [14,32], only oc and occasionally op are detected in the plasma of rats (figs. 3 and 4). the species difference in op metabolism may be related to the different types of carboxylesterases in rats and humans [41]. as a result, not all systemic op in humans are converted to oc by the hepatic carboxylesterases [14,32]. results of our study show the kinetics of in vitro op biotransformation in rat plasma can be adequately described by the michaelis-menten equation (table 2). because carboxylesterase activity varies among different animal species and organ sites [41], op biotransformation is assumed to take place only in the liver of rats during pbpk model development (fig. 2) although the km (300 mg/l) and vmax (61.2 mg/h) of the rat pbpk model have been derived from rat plasma incubation studies. in contrast, the km (187 µm, 76mg/l) and vmax (114 nmol/mg/min, 2.6 x 10 5 mg/h) of the human pbpk model have been obtained from in vitro human liver microsomal studies of shi et al., [2] (table 2). our studies also show that the cmax of oc in the organs/tissues of rats decreases in the order of liver > kidney > lung > plasma > brain. similar orders of radioactivity distribution were observed in the ferrets [14] and mice [15] after oral administration of radiolabeled op. however, because the radioactivity is comprised of a mixture of op, oc and other metabolites, it is not possible to quantify op and oc individually based on the results of the tissue distribution studies. our pbpk model provides a means to integrate relevant in vitro and in vivo data into a coherent description of op/oc pharmacokinetics for the whole animal. the model is comprised of an op and an oc sub-models (fig. 2). model performance is evaluated by comparing simulation results with experimental data. the proposed pbpk model is quite complex but can be collapsed into a less complex model by consolidating the various tissues/organs into rich and poor perfusion compartments. we have chosen a more complex model for the present study because it possesses the structural details to address the pharmacological and toxicological effects of op in humans. in addition, although it is more difficult to fit available experimental data to a complex model, the derived parameter values from a complex model are more accurate than those derived from a less complex model [19]. more than 80 parameter values are required to implement the pbpk model of op/oc. there is always uncertainty regarding whether they actually represent true parameter values. we have used the analytical sensitivity analysis [39] to identify the model parameters that significantly affect the oc dose metrics in the lungs and plasma. as shown in table 3, none of the parameter has an absolute sensitivity coefficient significantly greater than 1.0. this indicates that there is no amplification of errors from the input parameters in the prediction of oc dose metrics in these target organs. in other words, the oc dose metrics in the lungs and plasma are predicted with a certain degree of confidence even though some of the model parameter values may have been determined inaccurately. the pbpk model in rat was developed majorly with the dataset following orally administered 10 mg/kg, among all the parameters, distribution, metabolism and elimination parameters were basically from experimental data while absorption related parameters were allowed to be optimized to a certain extent, the model could predict the tissue distributions satisfactorily for 10 mg/kg and two published rat plasma datasets [15,16]. while the developed model was used to simulate the tissue exposure in rat following orally administered 50 mg/kg for validation, the simulations seemed to have the tendency of overestimating cmax and underestimating auc for both op and oc in kidney and lung, the discrepancy might be related to the higher optimized parameters related to absorption (ka, f) and renal clearance of op (clp) compared to the reported values. considering the satisfactory simulations for exposure at 10 mg/kg admet & dmpk 7(1) (2019) 22-43 a physiologically-based pharmacokinetic model for oseltamivir doi: 10.5599/admet.628 37 and the uncertainty of absorption processes, this leaves space for further refinement or validation. on the other hand, the current pbpk model will be definitely improved with further specific parameters for metabolism considering different enzymes and happening sites between human and rat, absorption/distribution parameters related to p-gp transporters (brain/gi tract) [42-44] and elimination parameters related to organic anion transporters (oat) in kidney [45]. the therapeutic effects of op in humans are related to the inhibition of neuraminidase activity of the influenza virus. in other words, inhibition of neuraminidase activity by oc will result in a decrease or even stoppage of progeny virion release and propagation in the respiratory tract. thus op efficacy may be evaluated by comparing the dosimetry of oc in the lung and the ic50 of oc of neuraminidases. he et al., [14] showed that the ic50 of oc for neuraminidases range from 0.3-2.0 nm (0.08-0.28 ng/ml) in an enzymatic assay to 0.6-155 nm (0.17-32.8 ng/ml) in a viral replication inhibition assay. as such the ic50 of oc differs among the subtype of influenza virus neuraminidases; it may range from 0.01-3.3 nm and 0.72.2 nm for h2n3 and h1n1 influenza viruses, respectively [32]. as shown in fig. 9, the steady-state concentrations of oc in the plasma/organs of a hypothetical human receiving the recommended op dose regimen are much higher than the minimum of in vitro ic50 of h1n1 virus neuraminidase (3-10 ng/ml) [14]. also, the trough (cmin) of the oc concentration-time curve in the lung is more than 10 fold higher than the ic50 value (fig. 9). as such, op is an effective anti-viral drug for h1n1 virus and perhaps, for other influenza virus strains as well [32]. the dosimetry of oc in the lung also explains why a dose regimen higher than 75 mg op b.i.d. for one week would not increase the efficacy of op significantly [3, 4]. as shown above, pulmonary oc concentration already is much higher than the minimum of in vitro ic50. thus, a dose regimen higher than 75 mg op b.i.d. is unlikely to have additional inhibitory effects on neuraminidase activity. the side effects of op/oc also are explainable by the dose metrics of op and oc in the brain and gastrointestinal tract; no threshold doses of toxicity have been reported in these organs. as shown in fig. 8, modeled oc aucbrain/aucplasma ratios for humans dosed with 50, 100, 200, and 500 mg op are 5.34/21.4, 10.6/42.5, 21.3/85.3, and 52.5/210.2, respectively, (fig.8) yielding a mean oc aucbrain/aucplasma ratio of about 0.25. jhee et al [42] have reported that the mean spinal fluid/plasma concentration ratio for oc in humans is 2.93 + 4.06 which is about 10-fold higher than the modeled aucbrain/aucplasma ratio observed in the present study. other study reported the brain-to-plasma exposure ratios approximately 0.2 for op and 0.01 for oc [43, 44], these result suggested significant variation between different studies. in our model, the brain/ plasma partition coefficients for op and oc were 0.5 and 0.25 respectively. it is assumed that only a small amount of oc seems able to penetrate the blood-brain barrier to reach the brain tissue considering oc with high polarity and op as the substrate of efflux transporter-p-glycoprotein. as such, op/oc probably is not the major cause of the observed neuropsychiatric effects in humans. as for the gastrointestinal side effects (e.g., abdominal pain, diarrhea, and nausea) [3,4], they may be related to the high levels of op in the digestive system and contact site toxicity since op concentrations in the gastrointestinal tract and liver may be as high as 3.72 µg/g and 2.00 µg/g, respectively (fig. 8). this is the first study in which a pbpk model is used to predict the internal tissue doses of op/oc in humans. the human and rat pbpk models are able to predict the pharmacokinetic behaviors and internal tissue doses of op and oc accurately. thus, the pbpk model is a valuable, quantitative tool for routerelated (fig. 5), species-related (fig. 6), and dose-related (fig. 8) extrapolation. the proposed pbpk model has many practical applications: it can be used to assess the dosimetry of op/oc in the internal tissues/organs of humans providing a more realistic and accurate account of the dose-response g. gao et al. admet & dmpk 7(1) (2019) 22-43 38 relationship of op/oc than the externally applied dose. the pbpk model also is useful in adjusting the dose regimens for patients in disease states such as renal failure and liver damage [22]. conflict of interest statement: there is no conflict of interest for this study. list of abbreviations pbpk model: physiologically-based pharmacokinetic model; op: oseltamivir phosphate; oc: oseltamivir carboxylate; f: fraction of absorption or empirical bioavailability factor ka: absorption rate constant; tlag: lag time for absorption cl: whole body clearance; km: michaelis constant; vmax: maximum reaction rate; kf: the fecal excretion rate constant qxy: blood flow to the organ; cxy: op/oc concentration in the organ; vxy: volume of the organ; rxy: tissue/plasma partition coefficient for the tissue; blplr: blood and plasma concentration ratio for op or oc auc: area under the time-concentration curve bw: the average body weight of humans and rats. acknowledgements: the authors thank the rfcid/hhsrf/hsrf/hcpf in hong kong for financial support of this project, and hoffman la-roche for providing tamiflu (both the pro-drug oseltamivir phosphate and the active metabolite oseltamivir carboxylate) for the project. the authors also thank prof. hei wun leung and the cma laboratory for their help in developing analytical methods for tamiflu and its metabolite, and dr. j.a. raymond for reviewing the manuscript. references [1] j.r. smith, c.r. rayner, b. donner, m. wollenhaupt, k. klumpp, r. dutkowski. oseltamivir in seasonal, pandemic, and avian influenza: a comprehensive review of 10-years clinical experience. adv. ther. 28(11) (2011) 927-959. [2] d. shi, j. yang, d. yang, e.l. lecluyse, c. black, l.you, f. akhlaghi, b.yan. anti-influenza prodrug oseltamivir is activated by carboxylesterase human carboxylesterase 1, and the activation is inhibited by antiplatelet agent clopidogrel. j. pharmacol. exp. ther. 319 (2006) 1477-1484. [3] f.g. hayden, j.j. treanor, r.s. fritz, m. lobo betts, r.f. miller, m.n. kinnersley, r.g. mills, p. ward, s.e. straus. use of the oral neuraminidase inhibitor oseltamivir in experimental human influenza: randomized controlled trials for prevention and treatment. jama 282 (1999) 1240-1246. [4] k.g. nicholson, f.y. aoki, a.d. osterhaus, s. trottier, o. carewicz, c.h. mercier, a. rode, n. kinnersley, p ward. efficacy and safety of oseltamivir in treatment of acute influenza: a randomized controlled trial. lancet 355 (2000) 1845-1850. [5] fda, http://www.fda.gov/medwatch/safety/, safety alerts for human medical products (drugs, biologics, medical devices, special nutritionals, and cosmetics). 2008/safety08.htm# tamiflu. posted on march 4 [accessed 06.11.08]. http://www.fda.gov/medwatch/safety/ admet & dmpk 7(1) (2019) 22-43 a physiologically-based pharmacokinetic model for oseltamivir doi: 10.5599/admet.628 39 [6] n. goshima, t. nakano, m. nagao, t. ihara. a clinical study of abnormal behaviors in patients with influenza. infect. immun child 18 (2006) 371-376. [7] s. toovey, c. rayner, e. prinssen, t. chu, b. donner, b. thakrar, r. dutkowski, g. hoffmann, a. breidenbach, l. lindemann, e. carey, l. boak, r. gieschke, s. sacks, j. solsky, j. small, d. reddy. assessment of neuropsychiatric adverse events in influenza patients treated with oseltamivir: a comprehensive review. drug safety 31 (2008) 1097-1114. [8] k. morimoto, m. nakakariya, y. shirasaka, c. kakinuma, t. fujita, l. tamai, t. ogihara. oseltamivir (tamiflu) efflux transport at the blood-brain barrier via p-glycoprotein. drug metab. dispos. 36 (2008) 6-9. [9] g. hoffmann, c. funk, s. fowler, m.b. otteneder, a. breidenbach, c.r. rayner, t. chu, e.p. prinssen. nonclinical pharmacokinetics of oseltamivir and oseltamivir carboxylate in the central nervous system, antimicrob. agents chemother. 53 (2009) 753-4761 [10] s.m. crain, k.f. shen. neuraminidase inhibitor ,oseltamivir blocks gm1 ganglioside-regulated excitatotory opioid receptor-mediated hyperalgesia, enhances opioid analgesia and attenuates tolerance in mice. brain res. 995 (2004) 260-266. [11] h. ono, y. nagano, n. matsunami, s. sugiyama, s. yamamoto, m. tanabe. oseltamivir, an antiinfluenza virus drug, produces hypothermia in mice. biol. pharm. bull. 31 (2008) 638-642. [12] y. izumi, k. tokuda, k.a. o’dell, c.f. zorumski, t. narahashi. neuroexcitatory actions of tamiflu and its carboxylate metabolite. neurosci. lett. 426 (2007) 54-58. [13] d. sweeny, g. lynch, a.m. bidgood, w. lew, k.y.wang, k.c. cundy. metabolism of the influenza neurominidase inhibitor prodrugoseltamivir in the rat. drug metab. dispos. 28 (2000) 737-774. [14] g. he, j. massarella, p. ward. clinical pharmacokinetics of the prodrug oseltamivir and its active metabolic ro 64-0802, clin. pharmacokinet. 37 (1999) 471-484. [15] a. hatori, t. arai, k. yanamoto, t. yamasaki, y. kawamura, j. yui, f. konno, r. nakao, k. suzuki, m.r. zhang. biodistribution and metabolism of the anti-influenza drug [11c]oseltamivir and its active metabolite [11c]ro 64-0802 in mice. nucl. med. biol. 36 (2009) 47-55. [16] w. li, p.a. escarpe, e.j. eisenberg, k.c. cundy, c. sweet, k.j. jakeman, j. merson, w. lew, m. williams, l. zhang, c.u. kim, n. bischofberger, m.s. chen, d.b. mendel. identification of gs 4104 as an orally bioavailable prodrug of the influenza virus neurominidase inhibitor gs 4071. antimicrob. agents chemother. 42 (1998) 647-653. [17] m. kurowski, c. oo, h. wiltshire. oseltamivir distributes to influenza virus replication sites in the middle ear and sinuses. clin. drug invest. 24 (2004) 49-53. [18] m.e. andersen. tissue dosimetry in risk assessment, or what’s the problem here anyway? in: pharmacokinetics in risk assessment, drinking water and health. pp 8-23 (1987), national academy press, washington dc. [19] who, international programme on chemical safety, characterization and application of physiologically based pharmacokinetic models in risk assessment. harmonization project document, no. 9 (2010), who press, geneva 27, switzerland. [20] n. parrott, b. davies, g. hoffmann, a. koerner, t. lave, e. prinssen, e. theogaraj, t. singer. development of a physiologically based model for oseltamivir and simulation of pharmacokinetics in neonates and infants. clin. pharmacokinet. 50(9) (2011) 613-623. [21] m. boberg, m.vrana, a. mehrotra, r. e. pearce, a. gaedigk, d.k.bhatt, j. s.leeder, b. prasad. agedependent absolute abundance of hepatic carboxylesterases (ces1 and ces2) by lc-ms/ms proteomics: application to pbpk modeling of oseltamivirin vivo pharmacokinetics in infants, drug metab. dispos. 45 (2017) 216–223. [22] i. nestorov. whole body pharmacokinetic models. clin. pharmacokinet. 42 (2003) 883-908. [23] q. chang, m.s.s chow, z. zuo. studies on the influence of esterase inhibitor to the pharmacokinetic profiles of oseltamivir and oseltamivir carboxylate in rats using an improved lc/ms/ms method. biomed. chromatogr. 23 (2009) 852-857. g. gao et al. admet & dmpk 7(1) (2019) 22-43 40 [24] n. lindegardh, w. hanpithakpong, y. wattanagoon, p. singhasivanon, n.j. white, n.p.j. day. development and validation of a liquid chromatographic-tandem mass spectrometric method for determination of oseltamivir and its metabolite oseltamivir carboxylate in plasma, saliva and urine. j. chromatogr. b 859 (2007) 74-83. [25] o. luttringer, f.p. theil, p. poulin, a.h. schmitt-hoffmann, t.w. guentert, t. lave. physiologicallybased pharmacokinetic (pbpk) modeling of disposition of epiroprim in humans. j. pharm. sci. 92 (2003) 1990-2007. [26] b. davies, t. morris. physiological parameters in laboratory animals and humans. pharm. res. 10 (1993) 1093-1095. [27] m.j. angelo, a.b. pritchard. route-to-route extrapolation of dichloromethane exposure using a physiological pharmacokinetic model. in: pharmacokinetics and in risk assessment: drinking water and health. national academy press, 1987 pp 254-264, washington, dc. [28] m.j. anderton, m.m. manson, r. verschoyle, a. gescher, w.p. steward, m.l.williams, d.e. mager. physiological modeling of formulated and crystalline 3,3’-diindolylmethane pharmacokinetics following oral administration in mice. drug metab. dispos. 32 (2004) 632–638. [29] j.m. gallo, f.c. lam, d.g. perrier. area method for the estimation of partition coefficients for physiological pharmacokinetic models. j. pharmacokinet. biopharm. 15 (1987) 271-280. [30] z. ye, h. gao, approaches to measure protein binding of enzymatically unstable compounds in plasma, bioanalysis 10 (2018) 451-459. [31] k.b. bischoff, r.l. dedrick, d.s. zaharko, j.a. longstreth. methotrexate pharmacokinetics. j. pharm. sci. 60 (1971) 1128-1133. [32] b.e. davies. pharmacokinetics of oseltamivir: an oral antiviral for the treatment and prophylaxis of influenza in diverse populations. j. antimicrob. chemother. 65 (suppl. 2) (2010) ii5-ii10. [33] s. lennon, j. barrett, c. kirkpatrick, c. rayner. oseltamivir oral suspension and capsules are bioequivalent for the active metabolite in healthy adult volunteers. int. j. clin. pharmacol. ther. 47 (2009) 539-549. [34] w.l. chiou, g. robbie, s.m. chung, t.c. wu, c. ma. correlation of plasma clearance of 54 extensively metabolized drugs between humans and rats: mean allometric coefficient of 0.66. pharm. res. 15 (1998) 1474-1479. [35] t. iwatsubo, n. hirota, t. ooie, h. suzuki, n. shimada, k. chiba, t. ishizaki, c.e. green, c.a. tyson, y. sugiyama. prediction of in vivo drug metabolism in the human liver from in vitro metabolism data. pharmacol. ther. 73 (1997) 147-171. [36] c.c. travis. interspecies and dose-route extrapolations, in: pharmacokinetics in risk assessment: drinking water and health, vol. 8. national academy press, 1987, washington, d.c. [37] f.j. eisenberg, a. bidgood, k.c. cundy. penetration of gs 4071, a novel influenza neuraminidase inhibitor into rat bronchoalveolar lining fluid following oral administration of the prodrug gs 4104. antimicrob. agents chemother. 41 (1997) 1949-1952. [38] s. bjorkman, d.r. wada, d.r. stanski, w.f. ebling. comparative physiological pharmacokinetics of fentanyl and alfentanil in rats and humans based on parametric single-tissue models. j. pharmacokinet. biopharmaceut. 22 (1994) 381410. [39] h.j. clewell, t.s. lee, r.l. carpenter. sensitivity of physiologically based pharmacokinetic models to variation in model parameters: methylene chloride. risk anal. 14 (1994) 521-531. [40] g. gao, f.c.p. law. physiologically-based pharmacokinetics of matrine in the rat after oral administration of pure chemical and acapha. drug metab. dispos. 37 (2009) 884-91. [41] c.y. quon, k. mai, g. patil, h.f. stampfli. species differences in the stereoselective hydrolysis of esmolol by blood esterases. drug metab. dispos. 16 (1988) 425-428. [42] s.s. jhee, m. yen, l. ereshefsk. low penetration of oseltamivir and its carbonylate into cerebrospinal fluid in healthy japanese and caucasian volunteers. antimicrob. agents chemother. 52 (2008) 36873693. http://www.ncbi.nlm.nih.gov/pubmed?term=iwatsubo%20t%5bauthor%5d&cauthor=true&cauthor_uid=9131722 http://www.ncbi.nlm.nih.gov/pubmed?term=hirota%20n%5bauthor%5d&cauthor=true&cauthor_uid=9131722 http://www.ncbi.nlm.nih.gov/pubmed?term=ooie%20t%5bauthor%5d&cauthor=true&cauthor_uid=9131722 http://www.ncbi.nlm.nih.gov/pubmed?term=suzuki%20h%5bauthor%5d&cauthor=true&cauthor_uid=9131722 http://www.ncbi.nlm.nih.gov/pubmed?term=shimada%20n%5bauthor%5d&cauthor=true&cauthor_uid=9131722 http://www.ncbi.nlm.nih.gov/pubmed?term=ishizaki%20t%5bauthor%5d&cauthor=true&cauthor_uid=9131722 admet & dmpk 7(1) (2019) 22-43 a physiologically-based pharmacokinetic model for oseltamivir doi: 10.5599/admet.628 41 [43] a. ose, h. kusuhara, k. yamatsugu, m. kanai, m. shibasaki, t. fujita, a.yamamoto, y. sugiyama. pglycoprotein restricts the penetration of oseltamivir across the blood–brain barrier. drug metab. dispos. 2 (2008) 427–434. [44] g. hoffmann, c. funk, s. fowler, m.b. otteneder, a. breidenbach, c.r. rayner, t. chu, e.p. prinssen. nonclinical pharmacokinetics of oseltamivir and oseltamivir carboxylate in the central nervous system. antimicrob. agents chemother. 53 (2009) 4753-4761. [45] c.h. hsueh, v. hsu, p. zhao, l. zhang, k.m. giacomini. s.m. huang, pbpk modeling of the effect of reduced kidney function on the pharmacokinetics of drugs excreted renally by organic anion transporters, clin. pharmacol. 103(3) (2018) 485-492. appendix the mass balance equations for non-eliminating organs and tissues such as lung, brain, spleen, muscle, and the rest of body are as follows: xyp xyp xy xy bap xyp d d c c v q c t r        (1a) xyc xyc xy xy bap xyc d d c c v q c t r        (1b) where xy represents the lung, brain, spleen, muscle, and the rest of body. the terms qxy, vxy and rxy represent tissue blood flow, volume and tissue/plasma partition coefficient, respectively; subscripts p and c represent op and oc, respectively. brain brp brp br br bap brp d d c c v q c t r        (2a) brc brc br br bap d d cbrc c c v q c t r        (2b) lung lgp lgp lg lg bvp lgp d d c c v q c t r        (3a) lgc lgc lg lg bvc d d clgc c c v q c t r        (3b) gut contents  cgcpgc d d p gcp gt c v kf c v t   (4a)  cgccgc d d c gcc gt c v kf c v t   (4b) kidney kdp kdp kd kd bap p bap kdp d cl d c c v q c c t r         (5a) g. gao et al. admet & dmpk 7(1) (2019) 22-43 42 kdc kdc kd kd bac c bac kdc d cl d c c v q c c t r         (5b) venous blood cbvp mspftp hrp kdp lvp bv ft hr kd ms lv ftp hrp kdp msp lvp brp rbp br rb c bvp brp rbp d d c cc c c c v q q q q q t r r r r r c c q q q c r r          (6a) cbvc ftc hrc kdc msc lvc bv ft hr kd ms lv ftc hrc kdc msc lvc brc rbc br rb c bvc brc rbc d d c c c c c c v q q q q q t r r r r r c c q q q c r r          (6b) mixed venous plasma bvp pvp pblplr c c  (7a) bvc pvc cblplr c c  (7b) cpvp and cpvc are the concentration in mixed venous plasma of op and oc, cbvp and cbvc are the concentration in mixed venous blood, blplr is the concentration ratio in blood and plasma. arterial blood lgpbap ba c bap lgp d d cc v q c t r        (8a) bac lgc ba c bac lgc d d c c v q c t r        (8b) liver   gup spplvplv lv gu sp bap gu gup spp lvp lvp lv max,p lvp lvp lvp m,p lvp d d c cc v q q q c q qsp t r r c c q v r c r k r               (9a)  lvc guc spclv lv gu sp bac gu sp guc spc lvc lvp lv max,p lvc lvp lvp m,p lvp d d c c c v q q q c q q t r r c c q v r r c r k r               (9b) admet & dmpk 7(1) (2019) 22-43 a physiologically-based pharmacokinetic model for oseltamivir doi: 10.5599/admet.628 43 gut tissue gtp gtp gt gt bap gtp d rao d c c v q c t r         (10a) gtc gtc gt gt bac gtc d d c c v q c t r        (10b) where, kf is the fecal excretion rate constant; vmax and km are the maximum metabolic rate for op and the constant respectively; r = mwoc /mwop, mw is molecular weight; rao is op input rate into the blood from the gut, rao = ka(f)(dose)e -ka(t-tlag) , where ka represents the absorption rate constant of op, dose represents op dosage, f represents the apparent or empirical bioavailability factor, and tlag represents the lag times for absorption; kf is the fecal excretion rate constant; clp and clc are the renal clearance for op and oc, respectively. ©2019 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ exploring the impact of crohn’s disease on the intragastric environment of fasted adults doi: http://dx.doi.org/10.5599/admet.846 122 admet & dmpk 8(2) (2020) 122-128; doi: http://dx.doi.org/10.5599/admet.846 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper exploring the impact of crohn’s disease on the intragastric environment of fasted adults maria vertzoni 1 , christina koulouri 1 , androniki poulou 2 , konstantinos goumas 2 , christos reppas 1, * 1 department of pharmacy, national and kapodistrian university of athens, zografou, greece 2 department of gastroenterology, red cross hospital of athens, athens, greece *corresponding author: e-mail: reppas@pharm.uoa.gr; tel.: +30 210 727 4678; fax: +30 210 727 4027 received: may 02, 2020; revised: may 27, 2020; published: june 15, 2020 abstract we explored the potential impact of crohn’s disease on the intragastric environment of fasted adults with a view to potential effects on intragastric performance of orally administered drugs in the fasted state. data were collected from 15 healthy individuals and 15 patients with crohn’s disease. all subjects remained fasted for at least 12h prior to gastroscopy. intragastric resting volume and ph were measured upon aspiration. osmolality, surface tension, pepsin activity, and content of six bile acids were measured within 4 months upon sample collection. unlike intragastric volumes, intragastric osmolality was significantly increased by crohn’s disease. however, mean osmolality value in patients was only slightly higher than in healthy individuals (293 vs. 257 mosmol/kg, respectively), therefore, unlikely to affect intragastric drug product performance. primarily due to the high variability of data in healthy individuals, the potential effects on intragastric ph and surface activity could not be evaluated on a statistical basis. however, based on average (mean and median) values, even if they are statistically significant, it seems unlikely to be of clinical significance. inter-subject variability of pepsin activity, and total bile acids content was high in both the healthy and the patients’ groups. statistical investigation of the potential impact of crohn’s disease on these parameters requires prior designation of the minimum differences to be detected; such differences will determine the minimum sample size required of relevant investigations. ©2020 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords intragastric volume; ph; osmolality; surface tension; pepsin activity; bile salts introduction during the last two decades, substantial progress has been made in characterizing physicochemically the gastrointestinal (gi) contents of healthy adults [1] resulting to the development of novel in vitro and in silico methodologies for evaluating the luminal fate of orally administered drug products [2]. however, relevant luminal data in patients with a gi disease are very limited [3]. crohn’s disease is a heterogeneous disorder with a multifactorial etiology, including genetic factors, environmental insults and intestinal microbiota, characterized by chronic, segmental and transmural inflammation that affects the gastrointestinal tract and may involve any segment of the oral cavity up to http://dx.doi.org/10.5599/admet.846 http://dx.doi.org/10.5599/admet.846 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:reppas@pharm.uoa.gr http://creativecommons.org/licenses/by/4.0/ admet & dmpk 8(2) (2020) 122-128 impact of crohn’s disease on the intragastric environment doi: http://dx.doi.org/10.5599/admet.846 123 the anus. the involvement of the upper gi lumen is less known, as a routine upper endoscopy is not commonly indicated [4]. to date, apart from few data on the ph along the gi lumen of patients with inflammatory bowel disease [5], data on the physicochemical characteristics of luminal contents have not been investigated [2]. gastric contents, in addition to gastric acid, contain a variety of components that are either secreted by the stomach (e.g. enzymes, electrolytes, mucus), swallowed/ingested (e.g. saliva and, in the fed state, food components) or refluxed from the duodenum in the stomach (e.g. bile components) which can affect oral dosage form performance [1,6]. the purpose of the present exploratory study was to evaluate whether the fasting gastric environment of patients could be affected by the crohn’s disease. experimental collection and treatment of aspirates human aspirates were collected at the red cross hospital of athens, after receiving approval by the scientific and the executive committee of the hospital (ap 152/160307). fifteen healthy adults (21-60 years old, 8 males, 7 females) and fifteen adults with crohn’s disease (1860 years old, 8 males and 7 females) participated in the study during a routine gastroscopic procedure for diagnostic purposes at the gastroenterology department of red cross hospital of athens (table 1). diagnosis of crohn’s disease was based on histologic examination. fourteen had the disease in the distal and one in the proximal region of the digestive tract. all subjects gave their informed consent for performing gastroscopy and collecting gastric contents. table 1. demographics of individuals participated in the present investigation. healthy adults adults with crohn’s disease gender age (y) bmi (kg/m 2 ) gender age (y) bmi (kg/m 2 ) duration (y) f 26 21.19 f 18 19.05 < 0.5 m 50 31.14 f 26 19.38 1 m 34 22.22 m 40 24.31 0.5 m 50 25.17 m 33 19.05 6 f 52 23.44 f 27 19.53 10 m 21 22.84 f 48 23.44 1 f 58 26.22 f 52 25.71 not confirmed m 60 29.41 m 60 23.96 1 m 38 30.45 f 25 20.76 0.5 f 38 31.02 m 35 29.22 4 m 52 26.85 m 31 18.38 14 f 25 19.53 m 19 21.37 1 m 49 25.91 f 29 23.63 not confirmed f 59 24.52 m 18 21.45 < 0.5 f 59 22.04 m 19 21.14 not confirmed mean (sd) 44(13) 25.5(3.7) 32(13) 22.0(3.0) median (min-max) 50 (21-60) 25.2 (19.5-31.1) 29 (18-60) 21.4 (18.4-29.2) subjects remained fasted for at least 12 h prior to gastroscopy. following the standard protocol, gastrointestinal fluids were aspirated during the procedure to facilitate the diagnostic process i.e., improve visibility of the gastric wall. normally, these fluids are directed to the waste upon aspiration. complete http://dx.doi.org/10.5599/admet.846 m. vertzoni et al. admet & dmpk 8(2) (2020) 122-128 124 aspiration of gastric contents was performed to determine volume of resting gastric fluid. upon aspiration, each sample was kept for 2 h in the freezer, its ph was measured, and it was then transferred to -70 °c. in all cases, to avoid multiple thaw-freezing cycles, prior to freezing the material to be stored was distributed to five smaller vials (figure 1). osmolality, surface tension, pepsin activity and content of six bile acids were measured within 4 months upon sample collection. materials pepsin from porcine gastric mucosa was from sigma chemical co., st. louis, mo, usa. all other chemicals used were of analytical grade and all solvents were of lc-ms grade and were purchased from sigma aldrich chemie gmbh. analysis of samples ph values were measured by a ph electrode (schott, modelcg842, mainz, germany). surface tension was measured at 37 °c using the denouy ring method (sigma 70 tensiometer, ksv instruments ltd, helsinki, finland). osmolality was measured by using the freezing point depression technique (semimicro osmometer typ dig l; knauer, berlin, germany). pepsin activity was measured by a modification of the method described by anson, and quantification was based on pepsin from porcine gastric as a standard [7,8]. the quantification limit was calculated to be 424 usp units/ml. individual bile salts, that is taurocholates (tc), glycocholates (gc), taurochenodeoxycholates (tcdc), ursodeoxycholates (udc), glycochenodeoxycholates (gcdc) and glycochenocholates (gdc)] were analyzed using hplc-cad as described previously [9]. figure 1. measurements performed in samples aspirated in this study. a double lined box implies a specific vial that was stored at -70 °c prior to the relevant measurement. data treatment raw data are presented as box-whisker plots showing the median, 10 th , 25 th , 75 th , and 90 th percentiles and the individual outlying data points. within each box, horizontal solid lines indicate median values and horizontal dotted lines indicate mean values. where possible, i.e. the power of the test allowed and normality and equal variance tests passed, the difference between the two groups was evaluated with the unpaired t-test. type i error was set to 0.05. all statistical comparisons were performed using sigmastat 3.5 (spss science inc., new york, usa). results and discussion mean (sd) values for the resting gastric volumes were 5.8 (3.6) ml and 6.1 (4.5) ml for crohn’s disease patients and healthy subjects, respectively (figure 2a). these values are smaller than the resting gastric admet & dmpk 8(2) (2020) 122-128 impact of crohn’s disease on the intragastric environment doi: http://dx.doi.org/10.5599/admet.846 125 fluid volumes reported previously, based on data collected with a water-sensitive magnetic resonance imaging technique, i.e. individual volumes ranged from 13 ml to 72 ml [10]. figure 2. resting volume of gastric contents (a), ph (b), surface tension (c) and pepsin activity (d) of the contents of stomach of healthy adults (white boxes) and crohn’s disease patients (yellow boxes) measured in the fasted state. n is the number of subjects contributed to the construction of box plots. asterisk (*) indicates that the difference is significant. intragastric ph values of patients with crohn’s disease were less variable than in healthy adults. median (range) ph values of gastric contents in crohn’s patients and healthy adults were 2.0 (range: 1.2-8.1) and 1.7 (range: 1.0-8.1), respectively (figure 2b). similar ph values were observed in the literature. median intragastric ph values estimated from data collected in 12 patients with crohn’s disease and in 12 healthy adults using a radiotelemetry capsule were 2 and 1.55, respectively [11]. interestingly, the difference was significant whereas no differences between patients in remission and with active disease were observed [11]. in contrast, using a free-floating ph-sensitive telemetric capsule, ewe et al. observed no significant differences in intragastric ph values between crohn’s disease patients with the disease active at the ileocecal regions (n=15) and healthy adults (n=15) [12]. surface tension of gastric contents was less variable in crohn’s disease patients (figure 2c). median (range) values in the crohn’s disease patients and in healthy adults were 46.9 (43.4-52.0) and 47.6 (37.054.1), respectively. mean(sd) osmolality values of gastric contents were 293 (35) mosmol/kg and 257 (39) mosmol/kg in the crohn’s disease and healthy adults, respectively (figure 2d). intragastric osmolality was slightly but significantly different in adults with crohn’s disease (p= 0.034). http://dx.doi.org/10.5599/admet.846 m. vertzoni et al. admet & dmpk 8(2) (2020) 122-128 126 high inter-subject pepsin activity was observed in both the crohn’s disease patients and the healthy individuals groups. coefficients of variation were 101 % and 94 %, respectively. median (range) pepsin activity values of crohn’s disease gastric contents were 131,300 (range: < 424 595,300) usp units/ml and 137,000 (range: < 424 333,600) usp units/ml for the crohn’s disease and healthy adults, respectively. generally, these data are in agreement with a study in which the effect of malnutrition and subsequent refeeding on digestive function in crohn’s disease patients was investigated [13]. gastric acid secretion was stimulated by a one-hour infusion of pentagastrin and continuous aspiration through the gastric port enabled measurement of gastric acid secretion. the malnourished crohn’s patients (bmi <17 kg/m2) had a mean basal acid output of 0.64 ± 0.33 meq/h compared with 3.85 ± 0.93 meq/h in healthy adults (p< 0.01). following the period of nutritional support, the basal acid output reached levels of 2.12 ± 0.88 meq/h, not statistically different to the levels observed in healthy adults (p=0.06). in our study, crohn’s patients were within the normal weight apart from one patient who had the disease for more than 14 years and his bmi was borderline to be underweight. it should be noted that a statistically significant decrease of bmi in crohn’s disease patient compared to healthy adults was observed [mean (sd) bmi 22.0 (3.0) vs 25.5 (3.7), p= 0.009] (table 1). individual bile acid concentrations were highly variable in both study groups; concentrations were below the quantification limit (6 μm) in 11 out of 15 patients and in 7 out of 15 healthy individuals. coefficients of variation of individual bile salt concentrations in crohn’s patients and in healthy individuals ranged between 290 % and 387 % and between 206 % and 304 %, respectively. the predominant bile salts were glycoconjugates and the relative mean luminal concentrations were gc > gcdc > gdc ∼tc ∼ tcdc > udc (table 2). mean (median) values for total bile salts concentration in the gastric contents of patients with crohn’s disease and healthy individuals were 141(12) μμ and 99(10) μμ, respectively (figure 3). based on pepsin activity data and bile acids data, investigation of the impact of potential differences between crohn’s patients and healthy individuals requires prior designation of the minimum differences to be detected, as these will determine the minimum number of samples sizes required for reaching statistically relevant conclusions. table 2. total and individual bile salts content of gastric contents of healthy and crohn’s disease adults. in all cases, normality test failed; data are median (range); significance of difference between the two groups was tested with mann-whitney test; differences were not significant. healthy crohn’s disease total bile salts (μm) 10 (0 656) 12 (0 1828) glycocholates 8 (<1 327) 7 (<1 490) glycochenodeoxycholates <2 (<2 718) <2 (<2 227) taurocholates 1 (<1 107) <1 (<1 277) taurochenodeoxycholates 0 (0 81) 0 (0 342) glycodeoxycholates <2 (<2 49) <2 (<2 45) ursodeoxycholates <1 (<1 10) <1 (<1 15) admet & dmpk 8(2) (2020) 122-128 impact of crohn’s disease on the intragastric environment doi: http://dx.doi.org/10.5599/admet.846 127 figure 3. total bile salts concentration of gastric contents of healthy adults (white boxes) and crohn’s disease patients (yellow boxes) measured in the fasted state. n is the number of subjects contributed to the construction of box plots. conclusions compared to healthy individuals, intragastric osmolality was significantly different but only slightly higher in crohn’s patients; this difference is unlikely to affect the performance of drugs administered orally in the fasted state. based on this exploratory study, average values of all other characteristics also suggest that differences on the physicochemical characteristics of gastric contents between adults with crohn’s disease and healthy adults if any, are unlikely to affect the performance of orally administered drugs in the fasted state. however, especially for the most highly variable parameters, i.e. pepsin activity and bile acids concentrations, designation of the minimum differences to be detected is required in order to reach statistically relevant conclusions. acknowledgements: the authors would like to thank ms. maria koursari for her excellent technical assistance during the study day. part of this work was presented at aaps 2012 annual meeting and exposition october 14-18, 2012, chicago, illinois, usa. conflict of interest: the authors declare that they have no conflicts of interest to disclose. references [1] m. vertzoni, p. augustijns, m. grimm, m. koziolek, g. lemmens, n.j. parrott, c. pentafragka, c. reppas, j. rubbens, j. van den αbeele, t. vanuytsel, w. weitschies, c.g. wilson. impact of regional differences along the gastrointestinal tract of healthy adults on oral drug absorption: an ungap review. european journal of pharmaceutical sciences 134 (2019) 153-175. [2] j. butler, b. hens, m. vertzoni, j. brouwers, p. berben, j. dressman, c.j. andreas, j.k. schaefer, j. mann, m. mcallister, m. jamei, e. kostewicz, f. kesisoglou, p. langguth, m. minekus, a. müllertz, r. schilderink, m. koziolek, p. jedamzik, w. weitschies, c. reppas, p. augustijns. in vitro models for the prediction of in vivo performance of oral dosage forms: recent progress from partnership through the imi orbito collaboration. european journal of pharmaceutics and biopharmaceutics 136 (2019) 70-83. [3] c. stillhart , k. vučićević, p. augustijns, a. basit, h. batchelor, t. flanagan, i. gesquiere, r. greupink, d. keszthelyi, m. koskinen, c.madla, c. matthys, g. miljuš, m. mooij, n. parrott, a-l ungell, s. de wildt, m. orlu, s. klein, a. müllertz. impact of gastrointestinal physiology on drug absorption in http://dx.doi.org/10.5599/admet.846 m. vertzoni et al. admet & dmpk 8(2) (2020) 122-128 128 special populations--an ungap review. european journal of pharmaceutical sciences 147 (2020) 105280. [4] a.m. pimentel, r. rocha, g.o. santana. crohn's disease of esophagus, stomach and duodenum. world journal of gastrointestinal pharmacology and therapeutics 10 (2019) 35–49. [5] s.g. nugent, d. kumar, d.s. rampton, d.f. evans. intestinal luminal ph in inflammatory bowel disease: possible determinants and implications for therapy with aminosalicylates and other drugs. gut 48 (2001) 571–577. [6] m. koziolek, s. alcaro, p. augustijns, a.w. basit, m. grimm, b. hens, c.l. hoad, p. jedamzik., c.m. madla, m. maliepaard, l. marciani, a. maruca, n. parrott, p. pávek, c.j.h. porter, c. reppas, d. van riet-nales, j. rubbens, m. statelova, n.l. trevaskis, k. valentová, m. vertzoni, d.v. čepo, m. corsetti. the mechanisms of pharmacokinetic food-drug interactions a perspective from the ungap group. european journal of pharmaceutical sciences 134 (2019) 31-59. [7] m. l. anson. the estimation of pepsin, trypsin, papain and cathepsin with hemoglobin. the journal of general physiology 22 (1938) 79-89. [8] l. kalantzi, k. goumas, v. kalioras, b. abrahamsson, j.b. dressman, c. reppas. characterization of the human upper gastrointestinal contents under conditions simulating bioavailability/bioequivalence studies. pharmaceutical research 23 (2006) 165–176. [9] m. vertzoni, h. archontaki, c. reppas. determination of intralumenal individual bile acids by hplc with charged aerosol detection. journal of lipid research 49 (2008) 2690-2695. [10] c. schiller, c.p. fröhlich, t. giessmann, w. siegmund, h. mönnikes, n. hosten, w. weitschies. intestinal fluid volumes and transit of dosage forms as assessed by magnetic resonance imaging. alimentary pharmacology and therapeutics 22 (2005) 971-979. [11] a.g. press, i.a. hauptmann, l. hauptmann, b fuchs, m. fuchs, k. ewe, g. ramadori. gastrointestinal ph profiles in patients with inflammatory bowel disease. alimentary pharmacology and therapeutics 12 (1998) 673–678. [12] k. ewe, s. schwartz, s. petersen, a.g press. inflammation does not decrease intraluminal ph in chronic inflammatory bowel disease. digestive diseases and sciences 44 (1999) 1434–1439. [13] t.a. winter, s.j. o'keefe, m. callanan, t. marks. impaired gastric acid and pancreatic enzyme secretion in patients with crohn's disease may be a consequenece of a poor nutritional state. inflammatory bowel diseases 10 (2004) 618‐625. ©2020 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/) http://creativecommons.org/licenses/by/4.0/ mechanistic modeling of gastrointestinal motility with integrated dissolution for simulating drug absorption doi: http://dx.doi.org/10.5599/admet.829 314 admet & dmpk 8(3) (2020) 314-324; doi: http://dx.doi.org/10.5599/admet.829 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper mechanistic modeling of gastrointestinal motility with integrated dissolution for simulating drug absorption kevin c. johnson intellipharm, llc, 47 magazine street, cambridge, ma 02139 e-mail: kjohnson@intellipharm.com received: april 17, 2020; revised: may 31, 2020; published: june 09, 2020 abstract a new computational method ̶ the multiple moving plug (mmp) model ̶ is described to simulate the effect of gastrointestinal motility and dissolution on the pharmacokinetic profile of any given drug. the method is physiologically more consistent with the experimental evidence that fluid exists in discrete plugs in the gastrointestinal tract, and therefore is more realistic than modeling the gastrointestinal tract as a series of compartments with first-order transfer. the number of plugs used in simulations, their gastric emptying times and volumes, and their residence times in the small intestine can be matched with experimental data on motility. in sample simulations, drug absorption from a series of fluid plugs emptied from the stomach at evenly spaced time intervals showed lower cmax and higher tmax than an equivalent dose emptied immediately as a single plug. to the extent that new techniques can establish typical ranges for the volumes of fluid emptied from the stomach and their respective timing, the mmp model may be able to predict the effect of gastric emptying on the variability seen in pharmacokinetic profiles. this could lead to an expanded safe space for the regulatory acceptance of formulations based on dissolution data. ©2020 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords pbpk modeling; gastric emptying; drug particle size; multiple moving plug model introduction several reports have expressed the view that gastrointestinal (gi) transit is more realistically modeled as discrete plugs of fluid emptying from the stomach and moving through the small intestine rather than as a first-order process between static compartments [1-3]. experimental evidence comes from magnetic resonance imaging showing isolated pockets of fluid in the gi tract [4-6]. two of the reports [1,2] describe stochastic models for gastric emptying, but these models omit the effects of drug dissolution, absorption, and metabolism on plasma drug concentrations. by contrast, the multiple moving plug (mmp) model [3] is a natural extension of a well-established dissolution model [7-9], the default model in gastroplus [10], and can simulate the effects of both the gastric emptying of plugs and the dissolution within the plugs on plasma drug concentrations. the mmp model is capable of utilizing either experimental data or stochastic simulations of the volumes and timing of plugs emptying from the stomach. the computational method of the mmp model is described here in more rigorous mathematical detail than previously discussed [3]. http://dx.doi.org/10.5599/admet.829 http://dx.doi.org/10.5599/admet.829 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:kjohnson@intellipharm.com http://creativecommons.org/licenses/by/4.0/ admet & dmpk 8(3) (2020) 314-324 mechanistic modeling of gastrointestinal motility doi: http://dx.doi.org/10.5599/admet.829 315 sample simulations show the effects of gastric emptying on the pharmacokinetics of a hypothetical drug. computational method a system of coupled differential equations describing drug dissolution, absorption, and metabolism was solved numerically using the fourth-order runge-kutta method. development of this system has been described previously as well as the derivation of the noyes-whitney equation assuming spherical drug particle geometry [3,7,8]. the equations used are summarized in table 1 including the definitions of variables and parameters. new is the addition of the index j to create and track discrete plugs that move independently through the gi tract as illustrated in figure 1. it should be noted that figure 1 was drawn to show a reasonably realistic length-to-diameter ratio for the small intestine and was untangled to facilitate this perspective [11]. in reality, the small intestine must be much more convoluted to fit into the small space of the abdominal cavity. independence is achieved by making plug-specific initial conditions and time-/position-dependent changes to variables associated with each plug. increasing the number of plugs allows more refined simulation of localized gi events and the resulting variability in drug plasma concentrations. a specific example of the microsoft visual basic code used for the simulations is available at https://intellipharm.com/wp-content/uploads/2020/04/mmp.txt. reference 3 provides a general explanation of the coding techniques applied to the mmp model. figure 1. drawing of the gi tract illustrating multiple moving plugs of fluid. in the simulations for figures 2-4, the first plug (j=1) to empty from the stomach at time zero is shown in red, the second (j=2) at 15 minutes is shown in blue, and the third (j=3) at 30 minutes is shown in green. a step size of 0.001 minute was used for the runge-kutta method. the iterative nature of this method allows any parameter to be changed at a frequency equal to the step size. equations 1-4 describe how drug mass changes in various locations, but other parameters can be changed to match the physical characteristics of the location. for example, it has already been described how hij changes with a change in drug particle radius as particles dissolve [3,7,8]. because the number of parameter changes that could be simulated is unlimited, only kaj was altered to explore the potential effect of gastric emptying on the drug plasma concentration profile. table 2 lists the other parameters used in the simulations that were held constant. they are similar to the values reported for nifedipine [12]. http://dx.doi.org/10.5599/admet.829 https://intellipharm.com/wp-content/uploads/2020/04/mmp.txt kevin c. johnson admet & dmpk 8(3) (2020) 314-324 316 table 1. differential equations and symbol definitions for the multiple moving plug model. quantity being tracked equation solid drug mass (mg) in the gi tract in plug j from particle size fraction i 2 33 1 3 tij ij ij j j i ds o s s j o j d d i dx x xx c t h r v (1) dissolved drug mass (mg) in the gi tract in plug j from particle size fraction i 1 2 3 33 tij ij j j j ij i do s s a d ij o j d d ijd dx x xx c k x t h r v (2) drug mass (mg) in the plasma compartment originating from particle size fraction i in plug j 1 1 21 2 ij j ij ij ija j d 12 d d d y cl k f x k y k y t v (3) drug mass (mg) in the tissue compartment originating from particle size fraction i in plug j 2 12 1 21 2 ij ij ij d d y k y k y t (4) total mass of dissolved drug (mg) in plug j from all particle size fractions 1 t ijj i d d i n x x (5) total drug mass (mg) in the plasma compartment from all plugs and all particle size fractions 1 1 1 1 p t ij i j nn y y (6) drug concentration (mg/ml) in the plasma compartment 1t central d y c v or 1t d b y v m (7) symbol definition cl clearance (ml/min or ml/min/mg) csj time-dependent drug solubility (mg/ml) for drug in plug j d drug diffusion coefficient (cm 2 /min) fj bioavailability factor for presystemic metabolism in plug j hij time-dependent diffusion layer thickness (cm) for particles in plug j from particle size fraction i kaj time-dependent absorption rate constant for plug j k12 rate constant for transfer of drug from the plasma to the tissue compartment k21 rate constant for transfer of drug from the tissue to the plasma compartment mb body mass when vd is given in units of volume per unit of body mass n number of drug particle size fractions np number of plugs ρ drug density (mg/cm 3 ) io r initial particle radius (cm) for particles in particle size fraction i t time (min) vd volume of distribution (ml or ml/mg) vj time-dependent volume (ml) of plug j xoij initial solid drug mass (mg) in the gi tract in plug j from particle size fraction i in keeping with the generally accepted view that the main site of drug absorption is the small intestine and not the stomach [13], kaj was assumed to be zero while plug j was in the stomach. the number of plugs exiting the stomach could be any number np, and the volume of each plug vj could be any fraction of the total volume containing any fraction of the total dose. for simplicity, each vj was equal to 240 ml divided by np (240 ml/3 = 80 ml). admet & dmpk 8(3) (2020) 314-324 mechanistic modeling of gastrointestinal motility doi: http://dx.doi.org/10.5599/admet.829 317 table 2. parameters used for simulations. to simulate gastric emptying, kaj was transitioned from its value of zero in the stomach to its value of 0.07 min -1 in the duodenum. the transition time was assumed to be 1 minute. table 3 shows the equation and parameters used to simulate a smooth sigmoidal transition for the value of kaj, but any appropriate function could be used to accomplish the transition. a simulation using immediate changes for kaj instead of smooth transitions was essentially the same as the original simulation. however, smooth transitions are recommended as abrupt changes can lead to instabilities in the numerical method. table 3. equation and symbol definitions used to simulate a smooth transition for the absorption rate constant. quantity being tracked equation time dependent absorption rate constant for plug j 2 2 a si a stj j j j j j j j a si a stj j k ka si a st a si a st a a st k k [ ( )] cos t t tk k k k k k t t (8) symbol definition kajsi absorption rate constant for plug j in the small intestine kajst absorption rate constant for plug j in the stomach tkajsi time at which kaj completes its transition to kajsi from kajst tkajst time at which kaj begins its transition from kajst to kajsi the function described by equation 8 in table 3 is bounded by two pair of points: (tkajst, kajst) and (tkajsi, kajsi). the number of transitions can be increased by using equation 8 and common points as the start and end points of transitions. the time course of transitions taking place within a given plug can be made to be completely independent of that in any other plug. if the initial conditions and transitions that define each plug are exactly the same, then the simulation would be equivalent to a single plug of the equivalent total dose and volume. modeled as a series of discrete fluid plugs exiting the stomach, each containing drug particles in the process of dissolving, the effect of gastric emptying and drug particle size on drug pharmacokinetics was simulated for three evenly-spaced plugs. the total dose and fluid volume were held constant, but the drug particle size and dose in individual plugs were changed between simulations as shown in table 4. in each simulation, the first plug began to transition from the stomach to the small intestine immediately, the second beginning at 15 minutes, and the third beginning at 30 minutes. table 4. plug specific parameters used in figure 2-4 simulations. figure plug # gastric emptying onset time (min) dose in plug (mg) drug particle size (µm) 2 1 0 (red) 0.8 1 2 15 (blue) 0.8 1 3 30 (green) 0.8 1 3 1 0 (red) 0.8 10 2 15 (blue) 0.8 10 3 30 (green) 0.8 10 4 1 0 (red) 0 1 2 15 (blue) 1.2 1 3 30 (green) 1.2 1 parameter value drug solubility (mg/ml) 0.01 total dissolution volume (ml) 240 drug density (g/cm 3 ) 1.3 diffusion coefficient (cm 2 /min) 0.0003 body mass (kg) 70 bioavailability factor 0.5 clearance (ml/min/kg) 4 volume of distribution (ml/kg) 600 k12 (min -1 ) 0.03 k21 (min -1 ) 0.01 http://dx.doi.org/10.5599/admet.829 kevin c. johnson admet & dmpk 8(3) (2020) 314-324 318 results and discussion figures 2-4 show a series of simulations using the multiple moving plug model. the main goal was to provide a visual and conceptual understanding of the model and its capabilities within a limited number of simulations. with this in mind, the number of plugs (3), their individual volumes (80 ml), and the timing of their gastric emptying (0, 15, and 30 min) were held constant. the spacing of plug emptying times was intentionally exaggerated to allow visual distinction between the plugs. to make more accurate simulations, the number, volumes, and emptying times of plugs should be selected to agree with emerging experimental measurements. figures 2-4 show the individual plasma compartment drug concentrations arising from drug absorption from individual plugs as well as the total drug concentration from all plugs. the solid lines represent plasma concentrations using the left axis and dashed lines represent absorption rate constants using the right axis. for the absorption rate constant versus time profiles, the sigmoidal shape of the transition representing gastric emptying is difficult to discern, but it should be noted that it is not an abrupt change. in figure 2, the concentration profiles from individual plugs appear to have the same shape. this is because dissolution is essentially complete for all plugs when they are simulated to empty from the stomach. in comparison, the individual profiles in figure 3 are clearly different due to the slower dissolution rate from larger drug particles. in figure 3, the first plug (red) has significantly less drug dissolved than the last plug (green) when they empty from the stomach. this results in a slower rate of drug absorption from the first plug compared to the last plug and illustrates the ability of the mmp model to model plugs independently. figure 4 illustrates the ability of the mmp model to simulate inhomogeneity in the stomach by assigning different initial conditions to various plugs. the situation depicted in figure 4 might arise if a dosage form does not disintegrate immediately after ingestion with water. in this case, the first plug of fluid to empty from the stomach does not contain any drug. instead, the 0.8 mg of drug in the first plug in figure 2 was divided evenly between the second and third plugs. as expected, there was a delay in the onset of drug absorption in figure 4 relative to figure 2. the simulation in figure 4 also resulted in a higher cmax for the total drug concentration compared to figure 2. figure 2. simulated plasma drug concentration (left axis) from 3 individual plugs (red, blue, and green solid lines) as well as the summation of all plugs (black solid line). absorption from individual plugs begins at the onset of the absorption rate constant transitioning from 0 to 0.07 min -1 shown as dashed lines and scaled using the right axis. table 4 lists figure specific simulation parameters. admet & dmpk 8(3) (2020) 314-324 mechanistic modeling of gastrointestinal motility doi: http://dx.doi.org/10.5599/admet.829 319 figure 3. simulated plasma drug concentration (left axis) from 3 individual plugs (red, blue, and green solid lines) as well as the summation of all plugs (black solid line). absorption from individual plugs begins at the onset of the absorption rate constant transitioning from 0 to 0.07 min -1 shown as dashed lines and scaled using the right axis. table 4 lists figure specific simulation parameters. figure 3 uses a larger drug particle size compared to figure 2. figure 5 compares the summation plasma concentration profiles in figures 2-4 as well as a new simulation for absorption of the same total dose from a single plug to serve as a reference point. in general, simulating gastric emptying as a sequential movement of a series of independent plugs of fluid from the stomach to the intestine resulted in a reduced cmax and increased tmax compared to a single plug. figure 4. simulated plasma drug concentration (left axis) from 3 individual plugs (red, blue, and green solid lines) as well as the summation of all plugs (black solid line). absorption from individual plugs begins at the onset of the absorption rate constant transitioning from 0 to 0.07 min -1 shown as dashed lines and scaled using the right axis. table 4 lists figure specific simulation parameters. figure 4 removes the amount of drug in the first plug and splits it evenly between the second and third plugs compared to figure 2. http://dx.doi.org/10.5599/admet.829 kevin c. johnson admet & dmpk 8(3) (2020) 314-324 320 figure 5. simulated plasma drug concentration (summation only) from figure 2 (gray line), figure 3 (dashed line), and figure 4 (solid thinner line) compared to a one-plug simulation for reference (solid heavier line) that emptied from the stomach using the absorption rate constant transition for the first plug in figures 2-4. the mmp model invokes the need for real time data on the number, timing, and volume of fluid plugs emptying from the stomach and moving through the intestine. schiller et al. [6] reported a mean small intestinal fluid volume of 105 ml contained in a mean number of separated fluid pockets of 4. however, this assessment was a one-time observation made 1 hour after the last dose in a sequence of nondisintegrating marker capsules was given in the fasting state. mudie et al. [4] tracked the number, volume, and approximate location of fluid packets at fixed time points. after dosing 240 ml of water in the fasted state, the number of small intestinal water pockets increased from about 8 initially to about 16 before decreasing gradually over the 2-hour time period studied. the mean volume of the pockets was in the 3-7 ml range. mudie et al. [4] acknowledge that it is possible that a significant portion of the dose could be confined to just a few pockets. this suggests that the focus should be on the fluid that is likely to contain drug at the time of dosing and on its movement thereafter. it may also be possible to group fluid that empties from the stomach over a certain time range to reduce the number of fluid plugs while retaining the characteristics of the mmp model. the advantage in doing so is to reduce the computational time required to run a simulation. the same strategy has been applied successfully for simulating the dissolution of polydisperse drug powders by grouping drug particles within a certain size range into a single size group that is representative of the range. figure 6 compares the simulation in figure 2 with a simulation where the only differences are an increase in the number of plugs from 3 to 7 and a decrease in the dose in each plug from one third of the total dose to one seventh. the total plasma concentration profiles from the two simulations were not substantially different, indicating that there is potential to reduce the number of plugs without changing the simulation significantly. the number of plugs used in the mmp should be no more than the number observed experimentally and potentially only a fraction thereof. the summation technique of the mmp model is similar to the superposition principle, except that the summation occurs at each step of the numerical method, as opposed to summing the contributions after all plugs have been simulated. this treatment allows the mmp model to handle non-linear pharmacokinetics. different metabolic rates can be tied to the continuously updated drug mass in different admet & dmpk 8(3) (2020) 314-324 mechanistic modeling of gastrointestinal motility doi: http://dx.doi.org/10.5599/admet.829 321 locations: the intestine for gut wall metabolism, the portal vein for first-pass hepatic metabolism, or the plasma compartment for renal clearance. the pharmacokinetic model described in table 1 would have to be expanded to allow for the simulation of portal vein drug concentrations. figure 6. simulated plasma drug concentration from figure 2 (red) compared to a new simulation (black) except that the new simulation used 7 plugs instead of 3, and the total dose was divided evenly between 7 plugs instead of 3. the heavier lines represent summations from the lighter individual plug lines. the advantage of the mmp model is that it provides a more realistic computational approach to simulating gastric emptying and gi motility compared to the advanced compartmental absorption and transit (acat) model. historically, the acat model has assumed that both solid and dissolved drug move from compartment to compartment in a first-order process without any concomitant movement of compartmental fluid, which remains static in terms of both position and volume [14]. these assumptions are inconsistent with physical observations and cannot be replicated experimentally. indeed, the development of dissolution testing apparatus involving more than one dissolution vessels to mimic the stomach and segments of the small intestine rely on the concomitant movement of water and solid and dissolved drug [15]. ehrlein and schemann [16] have created a website with links to videos using fluoroscopy that provide visual confirmation of the true nature of gi motility. the other advantage of the mmp model is that all processes are modeled from the perspective of isolated plugs of moving fluid, making experimental validation of the model components possible. for example, it has already been mentioned that the dissolution component of the mmp model has been established [7,8] and independently verified [9] for a constant volume of fluid. however, in vivo, the volume of the dissolution fluid will change; increasing due to secretion of fluids or decreasing due to water absorption. testing of the dissolution model to simulate secretion can be accomplished by adding fluid to the dissolution vessel at a known rate. on the other hand, removing fluid to simulate water absorption is more difficult. evaporation is probably not fast enough to mimic the rate of water absorption from the gi tract, but could be used to test the dissolution model over a narrow range. there is no reason to believe that the principles of the mechanistic dissolution model should fail in extrapolating to higher rates of water removal. moreover, there is a need to simulate the effect of water absorption on drug dissolution and absorption as the reduction in water volume in the gi tract after dosing has been clearly shown by mudie et al. [4] in either case, water secretion or absorption, the volume change in a dissolution vessel can be measured and corrected at each step of the numerical simulation. the dissolution fluid could also be http://dx.doi.org/10.5599/admet.829 kevin c. johnson admet & dmpk 8(3) (2020) 314-324 322 titrated with physiological buffers, dynamically adjusting the ph-dependent solubility and volume change simultaneously. the mmp model is completely flexible; any number of plugs containing dissolving drug can empty from the stomach, each having a volume and emptying time to match experimental measurements. the critical physiological parameters are the volume of the plugs, the timing of their emptying, and their residence time in various segments of the intestinal tract. regional differences in permeability and intestinal wall metabolism can be adjusted to reflect the current position of individual plugs. although the simulations shown in figures 2-4 were run with equally spaced plugs of the same volume to demonstrate the model, using real gastric emptying data will enable more accurate simulations. the mmp model assumes that suspended and dissolved drug are homogeneously mixed within the plug, and that only dissolved drug can leave the plug through permeation of the intestinal membrane. there is no mass transfer between plugs, but plugs could be made to converge or diverge by manipulating the time-dependent parameters. it would also be possible to break up a large volume of fluid into several smaller plugs in order to simulate a stepwise concentration gradient axially from the leading to the trailing end of the plug series. it has long been recognized that gastric emptying can affect the rate of drug absorption [17,18]. this rate, characterized by cmax and tmax, is used to establish bioequivalence when considering a formulation change or when comparing a generic dosage form to an innovator’s product. simulations show that cmax and tmax are sensitive to gastric emptying as modeled by the mmp model. if the bioequivalence criteria are too tight, a dosage form could fail bioequivalence testing due to physiological factors beyond the control of the formulation. to address this problem, dickinson et al. [19] have discussed the concept of a safe space to allow for the effect of physiological factors like gastric emptying on the bioequivalence criteria. physiologically based pharmacokinetic (pbpk) modeling holds promise for establishing this safe space and for providing confidence in making predictions. the hope is that a simple dissolution test could serve to establish bioequivalence in conjunction with pbpk modeling. the theoretical advantage of the mmp model lies in its mechanistically-based mathematical integration of dissolution, gastric emptying, and intestinal motility in a way that is closer to how these processes occur in the gi tract. conclusions in conclusion, the mmp model provides a more mechanistically realistic computational tool to study the effects of dissolution, gastric emptying, and intestinal motility on the drug plasma concentration profile. the ability to establish the impact of dissolution and gi motility on bioequivalence criteria may expand safe space for regulatory approval of pharmaceutical products based on limited dissolution data and mechanistically-based simulations. acknowledgements: the author would like to thank michael camilleri, md of the mayo clinic for his helpful discussion. conflict of interest: the author is the president of intellipharm llc. references [1] a. talattof, g.l. amidon. pulse packet stochastic model for gastric emptying in the fasted state: a physiological approach. mol. pharm. 15 (2018) 2107-2115. admet & dmpk 8(3) (2020) 314-324 mechanistic modeling of gastrointestinal motility doi: http://dx.doi.org/10.5599/admet.829 323 [2] j. yokrattanasak, a. de gaetano, s. panunzi, p. satiracoo, w.m. lawton, y. lenbury. a simple, realistic stochastic model of gastric emptying. plos one 11 (2016) e0153297. [3] k.c. johnson, computer simulation for pharmaceutical scientists, intellipharm press, niantic, connecticut, usa, 2018, p 53. [4] d.m. mudie, k. murray, c.l. hoad, s.e. pritchard, m.c. garnett, g.l. amidon, p.a. gowland, r.c. spiller, g.e. amidon, l. marciani. quantification of gastrointestinal liquid volumes and distribution following a 240 ml dose of water in the fasted state. mol. pharm. 11 (2014) 3039-3047. [5] k. murray, c.l. hoad, d.m. mudie, j. wright, k. heissam, n. abrehart, s.e. pritchard, s. al atwah, p.a. gowland, m.c. garnett, g.e. amidon, r.c. spiller, g.l. amidon, l. marciani. magnetic resonance imaging quantification of fasted state colonic liquid pockets in healthy humans. mol. pharm. 14 (2017) 2629-2638. [6] c. schiller, c.p. fröhlich, t. giessmann, w. siegmund, h. mönnikes, n. hosten, w. weitschies. intestinal fluid volumes and transit of dosage forms as assessed by magnetic resonance imaging. aliment. pharmacol. ther. 22 (2005) 971-979. [7] r.j. hintz, k.c. johnson. the effect of particle size distribution on dissolution rate and oral absorption. int. j. pharm. 51 (1989) 9-17. [8] a.t. lu, m.e. frisella, k.c. johnson. dissolution modeling: factors affecting the dissolution rates of polydisperse powders. pharm. res. 10 (1993) 1308-1314. [9] j. jinno, n. kamada, m. miyake, k. yamada, t. mukai, m. odomi, h. toguchi, g.g. liversidge, k. higaki, t. kimura. effect of particle size reduction on dissolution and oral absorption of a poorly water-soluble drug, cilostazol, in beagle dogs. j. control. release 111 (2006) 56-64. [10] t.s. samant, s. dhuria, y. lu, m. laisney, s. yang, a. grandeury, m. mueller-zsigmondy, k. umehara, f. huth, m. miller, c. germa, m. elmeliegy. ribociclib bioavailability is not affected by gastric ph changes or food intake: in silico and clinical evaluations. clin. pharmacol. ther. 104 (2018) 374-383. [11] d.m. mudie, g.l. amidon, g.e. amidon. physiological parameters for oral delivery and in vitro testing. mol. pharm. 7 (2010) 1388-1405. [12] k.c. johnson. dissolution and absorption modeling: model expansion to simulate the effects of precipitation, water absorption, longitudinally changing intestinal permeability, and controlled release on drug absorption. drug dev. ind. pharm. 29 (2003) 833-842. [13] j. hirtz. the gastrointestinal absorption of drugs in man: a review of current concepts and methods of investigation. br. j. clin. pharmacol. 19 suppl.2 (1985) 77s-83s. [14] b. hens, m.b. bolger. application of a dynamic fluid and ph model to simulate intraluminal and systemic concentrations of a weak base in gastroplus™. j. pharm. sci. 108 (2019) 305-315. [15] p.j. o'dwyer, c. litou, k.j. box, j.b. dressman, e.s. kostewicz, m. kuentz, c. reppas. in vitro methods to assess drug precipitation in the fasted small intestine a pearrl review. j. pharm. pharmacol. 71 (2019) 536-556. [16] the moving gut by hans jörg ehrlein and michael schemann, https://www.humanbiology.wzw.tum.de/index.php?id=22&l=1 (accessed march 11, 2020). [17] r.c. heading, j. nimmo, l.f. prescott, p. tothill. the dependence of paracetamol absorption on the rate of gastric emptying. br. j. pharmacol. 47 (1973) 415-421. [18] w.s. nimmo, r.c. heading, j. wilson, p. tothill, l.f. prescott. inhigition of gastric emptying and drug absorption by narcotic analgesics. br. j. clin. pharmacol. 2 (1975) 509-513. [19] p.a. dickinson, w.w. lee, p.w. stott, a.i. townsend, j.p. smart, p. ghahramani, t. hammett, l. billett, s. behn, r.c. gibb, b. abrahamsson. clinical relevance of dissolution testing in quality by design. aaps j. 10 (2008) 380-390. ©2020 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://dx.doi.org/10.5599/admet.829 https://www.humanbiology.wzw.tum.de/index.php?id=22&l=1 https://www.humanbiology.wzw.tum.de/index.php?id=22&l=1 http://creativecommons.org/licenses/by/3.0/ strategies of solubility enhancement and perspectives in solubility measurements of pharmaceutical compounds doi: https://doi.org/10.5599/admet.910 176 admet & dmpk 8(3) (2020) 176-179; doi: http://dx.doi.org/10.5599/admet.910 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index editorial strategies of solubility enhancement and perspectives in solubility measurements of pharmaceutical compounds christel a.s. bergström 1,2 * and antonio llinas 3 * 1 department of pharmacy, uppsala university, uppsala biomedical center, p.o. box 580,se-751 23 uppsala, sweden 2 the swedish drug delivery center, department of pharmacy, uppsala university, uppsala biomedical center, p.o. box 580, se-751 23 uppsala, sweden 3 respiratory, inflammation and autoimmunity imed biotech unit, astrazeneca, gothenburg, sweden *corresponding authors emails: christel.bergstrom@farmaci.uu.se; antonio.llinas@astrazeneca.com. received: september 19, 2020; revised: september 25, 2020; published: september 27, 2020 the genesis of this theme issue was a conversation we had during the iapc-8 meeting in split, croatia in september 2019. in it, we started the interesting and important discussions surrounding the special session dedicated to ‘state-of-the-art solubility in drug development’. the session was introduced in the preliminary mini-review paper intended to serve as prologue or accompaniment to an upcoming session on solubility at the iapc-8 meeting in split, croatia, 9-11 september 2019 [1]. the session had received a high number of abstracts and hence, was extended from a half-day session to a session spanning over two days with many engaged speakers on topics around solubility, dissolution and enabling formulation strategies. in response to this successful session, professor mandic and professor tam in their role as editors suggested us to take lead for a special issue dedicated to these topics as they had listened in to the scientific discussions. we were both delighted to take on the role as guest editors. we have been overwhelmed by the response and number of high quality articles submitted for this special issue. indeed, it was during the course of this work decided to split the special issue into two consecutive issues due to the large number of papers being accepted. with this short editorial, we thank all authors for their excellent contribution to this special issue on solubility enhancement and measurements. in the first issue of admet&dmpk focus is set on computational tools and mathematical modelling useful to understand solubility and dissolution, with avdeef and kansy [2] as well as caron and colleagues [3] focusing their work on compounds in the beyond rule-of-5 (bro5) chemical space. avdeef and kansy explored to what extent computational models established on training sets from the small molecular chemical space were applicable to the bro5 chemical space. they concluded that random forest (rf) regression methods predicted the solubility slightly better for these compounds than the general solubility equation (gse) and the abraham solvation equation (absolv). however, the gse performed better (for the dataset explored) when the log p coefficient was changed from -1 to -0.4, providing less weight to lipophilicity in the predictions. they also concluded that the increased flexibility of the larger molecules in the bro5 space is one of the underlying reasons for less accurate predictions of traditionally developed models. the latter is also the conclusion from the study by caron and coworkers [3]. they emphasise the https://doi.org/10.5599/admet.910 http://dx.doi.org/10.5599/admet.910 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:christel.bergstrom@farmaci.uu.se mailto:antonio.llinas@astrazeneca.com admet & dmpk 8(3) (2020) 176-179 editorial, solubility issue doi: https://doi.org/10.5599/admet.910 177 role of flexibility and its importance for chameleonic behaviour, i.e. the ability of the molecule to adopt to the environment it is exposed to. they conclude that the current models are not accurate enough for predictions of this type of compounds and that the solubility models need to be developed for this chemical space. in relation to generation of more predictive solubility models, mitchell contributes with a paper using three different machine learning models to predict the ‘2019 solubility challenge dataset’ [4]. it should be noted that this dataset is more traditional, and biased towards the small molecule space, i.e. not the bro5 space. all three models were based on tree-like classifiers; rf, the extra trees algorithm, and a consensus model of the two. it was found the extra trees provided the best classifier whereas both rf and the consensus model were slightly better for an explored test set. the author reports rmses of 0.95 (tight dataset; extra trees) and 1.49 (loose dataset; rf/consensus) indicating the dependency of the chemical space explored. falcón-cano et al. [5] also explored rf approaches and consensus modelling to predict solubility, with focus on generating better algorithms and improved work-flow during the modelling for data curation and variable selection. in their study the reported rmse of the generated model is 0.93 log units for the test set taken from the solubility challenge dataset. the authors also contrast their models to other solubility models, providing an updated view on the standing of this field. improved modelling approaches are further studied by mecklenfeld and raabe [6]. they embarked on a study to optimise the force field parameters used for solvation energy calculations. they developed and validated general amber force field (gaff) parameters by taking into account the lennard-jones parameters in combination with implicit polarized charges. the validation was performed by making use of 142 free energies and 100 densities of binary mixtures and resulted in improved models of relative solubilities and fluid phase behaviour. the final two papers in the first special issue focus on mathematical models of dissolution of particles and gastrointestinal motility. grassi and coworkers [7] described dissolution of poly-dispersed particles in a finite liquid environment, and developed two different equations that take into consideration size reduction, polydispersity and particle geometry (spherical, cylindrical and parallelepided). these mathematical models are useful to, among others, early evaluate the importance of wettability for the dissolution process. johnson also models dissolution by coupling the dissolution process to the gastrointestinal motility [8]. in this work, the multiple moving plug model was described and used to simulate the effect of the gi motility and particle dissolution on the resulting pharmacokinetic profile of the compound. this work builds on the experimental evidence that the gi fluid exists in water pockets, and hence, the model provides a more physiologically relevant approach to explore the dissolution process in vivo. in the second issue focus is shifted towards experimental assessment and technologies for profiling and enabling absorption of poorly-water soluble compounds. a series of extensive reviews are provided in this special issue, describing the current standing of electrospun nanofibers for drug delivery and optimised dissolution profiles [9], supercritical fluid technology for controlled production of nanoand microparticles of poorly soluble compounds [10], nanoparticles and their in vivo performance [11] and solubility aspects for proteins [12]. in addition to these timely reviews on important topics for solubility, dissolution and enabling formualations, papers describing original research on assessment and drug delivery of poorly soluble compounds are included. štukelj et al. [13] describes a new methodology to determine dissolution of amorphous powders based on image analysis. in this work, image-based single particle analysis was used to study dissolution of amorphous indomethacin over time and the results were compared to powder xrd analysis. the single particle analysis was useful to study local recrystallization occurring on the particle surface, providing crucial information to understand surface chemistry effects during storage and dissolution. dissolution was also studied by gigante et al. [14] who set out to perform a global evaluation of https://doi.org/10.5599/admet.910 c. bergström and a. llinas admet & dmpk 8(3) (2020) 176-179 178 the who harmonised protocol for equilibrium solubility experiments. the study focused on measurement of 16 compounds for which the biopharmaceutics classification had previously been performed. they successfully validated the current protocol in this extensive collaboration involving 11 laboratories worldwide. the study also indicates that some compounds may have been sorted into the wrong bcs class in previous studies, meriting further exploration. dissolution evaluation of furosemide products was performed by medina-lopéz and coworkers [15]. they studied dissolution for furosemide reference tablets in usp apparatus 1, 2 and 4, under different experimental conditions. they observed statistically significant differences between usp apparatus 1 and 2, whereas conditions producing similar dissolution profiles for usp apparatus 2 and 4 were identified. in the paper authored by isvoran et al. [16] solubility of short oligomers of lactic acid is studied. polylactic acid has potential to be used in medical applications, and hence the solubility of this polymer is important. the authors conclude that solubility decreases linearly with increased molecular weight and that the short oligomers have limited toxicology but potentially may interact with organic anion transportrs (oatp1b1 and oatp1b3). the final contribution to this special issue is a paper by prestidge et al. [17]. they study a system composed of a lipid-silica hybrid to deliver highly lipophilic antipsychotic compounds. while lurasidone performed as expected in this system, with a 23-fold increase in solubilisation being observed for the developed hybrid system, risperidone instead showed a 2.2-fold reduction when encapsulated in this hybrid. it was shown that differences in ionization between these two compounds resulted in the risperidone adsorbed stronger to the silica than the lurasidone, emphasising the interplay of the drug delivery system and the drug itself for optimal performance. references [1] c.a.s. bergström, a. avdeef. perspectives in solubility measurement and interpretation. admet and dmpk 7 (2019) 88-105. [2] a. avdeef, m. kansy. can small drugs predict the intrinsic aqueous solubility of ‘beyond rule of 5’ big drugs?. admet and dmpk 8(3) (2020) 180-206. [3] g. ermondi, v. poongavanam, m. vallaro, j. kihlberg, g. caron. solubility prediction in the bro5 chemical space: where are we right now?. admet and dmpk 8(3) (2020) 207-214. [4] j. mitchell. three machine learning models for the 2019 solubility challenge. admet and dmpk 8(3) (2020) 215-250. [5] g. falcón-cano, c. molina, m.a. cabrera-pérez. adme prediction with knime: in silico aqueous solubility consensus model based on supervised recursive random forest approaches. admet and dmpk 8(3) (2020) 251-273. [6] a. mecklenfeld, g. raabe. gaff/ipolq-mod+lj-fit: optimized force field parameters for solvation free energy predictions. admet and dmpk 8(3) (2020) 274-296. [7] m. abrami, l. grassi, r. di vittorio, d. hasa, b. perissutti, d. voinovich, g. grassi, i. colombo, m. grassi. dissolution of an ensemble of differently shaped poly-dispersed drug particles undergoing solubility reduction: mathematical modelling. admet and dmpk 8(3) (2020) 297-313. [8] k.c. johnson. mechanistic modeling of gastrointestinal motility with integrated dissolution for simulating drug absorption. admet and dmpk 8(3) (2020) 314-324. [9] l. castillo-henríquez, r. vargas-zúñiga, j. pacheco-molina, j. vega-baudrit. electrospun nanofibers: a nanotechnological approach for drug delivery and dissolution optimization in poorly water-soluble drugs. admet and dmpk 8(4) (2020) 325-353. [10] s.k. misra, k. pathak. supercritical fluid technology for solubilization of poorly water soluble drugs via microand naonosized particle generation. admet and dmpk 8(4) (2020) 355-374. [11] d. dahlgren, e. sjögren, h. lennernäs. intestinal absorption of bcs class ii drugs administered as nanoparticles: a review based on in vivo data from intestinal perfusion models. admet and dmpk 8(4) (2020) 375-390. admet & dmpk 8(3) (2020) 176-179 editorial, solubility issue doi: https://doi.org/10.5599/admet.910 179 [12] m. vihinen. solubility of proteins. admet and dmpk 8(4) (2020) 391-399. [13] j. štukelj, m. agopov, j. yliruusi, c.j. strachan, s. svanbäck. image-based dissolution analysis for tracking the surface stability of amorphous powders. admet and dmpk 8(4) (2020) 401-409. [14] v. gigante, g.m. pauletti, s. kopp, m. xu, i. gonzalez-alvarez, v. merino, m.p. mcintosh, a. wessels, b.-j. lee, k.r. rezende, g.k.e. scriba, g.p.s. jadaun, m. bermejo. admet and dmpk doi: https://doi.org/10.5599/admet.850. [15] r. medina-lópez, s. guillén-moedano, m. hurtado. in vitro release studies of furosemide reference tablets: influence of agitation rate, usp apparatus and dissolution media. admet and dmpk 8(4) (2020) 411-423. [16] d. dascălu, d.l. roman, m. filip, a.a. ciorsac, v. ostafe, a. isvoran. solubility and admet profiles of short oligomers of lactic acid. admet and dmpk 8(4) (2020) 425-436. [17] t.r. meola, k. paxton, p. joyce, h.b. schultz, c.a. prestidge. the effect of drug ionization on lipidbased formulations for the oral delivery of anti-psychotics. admet and dmpk 8(4) (2020) 437-451. ©2020 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) https://doi.org/10.5599/admet.910 https://doi.org/10.5599/admet.850 http://creativecommons.org/licenses/by/3.0/ in vitro release studies of furosemide reference tablets: influence of agitation rate, usp apparatus, and dissolution media doi: http://dx.doi.org/10.5599/admet.801 411 admet & dmpk 8(4) (2020) 411-423; doi: http://dx.doi.org/10.5599/admet.801 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper in vitro release studies of furosemide reference tablets: influence of agitation rate, usp apparatus, and dissolution media raúl medina-lópez *, sergio guillén-moedano and marcela hurtado departamento sistemas biológicos, universidad autónoma metropolitana-xochimilco, mexico city, mexico *corresponding author: e-mail: rmlopez@correo.xoc.uam.mx; tel.: + 52 55 54837000 ext. 3445; fax: + 52 55 55947929 received: march 06, 2020; revised: june 20, 2020; published: june 29, 2020 abstract furosemide is a diuretic drug widely used in chronic renal failure. the drug has low solubility and permeability, which cause clinical problems. studying the in vitro release performance elucidates the rate and extent of drug dissolved from dosage forms under different conditions. furosemide reference tablets were tested using usp apparatuses 1 and 2 as well as the flow-through cell method (usp apparatus 4), a dissolution apparatus that simulates the human gastrointestinal tract better than the other methods. dissolution profiles were created with usp apparatuses 1 and 2 at 25, 50, and 75 rpm and 900 ml of 0.1 m hydrochloric acid, acetate buffer (ph 4.5), and phosphate buffer (ph 6.8). usp apparatus 4 with a laminar flow of 16 ml/min and 22.6 mm cells was used. drug dissolution was quantified at 274 nm for 60 min. mean dissolution time, dissolution efficiency, time to 50 % dissolution, and time to 80 % dissolution data were used to compare dissolution profiles. additionally, zero-order, first-order, higuchi, hixson-crowell, makoid-banakar, and weibull models were used to adjust furosemide dissolution data. between usp apparatus 1 and 2, significant differences were observed in almost all parameters at 50 and 75 rpm (p < 0.05). a similar dissolution profile (f2 > 50) with a pharmacopoeial dissolution method (usp apparatus 2 at 50 rpm and 900 ml of phosphate buffer ph 5.8) and usp apparatus 4 (laminar flow of 16 ml/min, 22.6 mm cells, and ph 6.8) was observed. the weibull function was the best mathematical model to describe the in vitro release performance of furosemide in the three usp dissolution apparatuses. these results could be used to manufacture better furosemide dosage forms and decrease the negative clinical impact of current furosemide formulations. ©2020 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords flow-through cell method; furosemide; lasix® drug product; usp basket and paddle apparatus introduction furosemide is a diuretic drug widely used in the treatment of oedematous states associated with cardiac, chronic renal failure, hypertension, congestive heart failure, and cirrhosis [1]. furosemide is a weak acid (pka = 3.8) with low solubility and permeability [2]. according to the biopharmaceutical classification system, drugs with these characteristics belong to class iv [3]. the chemical structure of furosemide is shown in figure 1. http://dx.doi.org/10.5599/admet.801 http://dx.doi.org/10.5599/admet.801 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:rmlopez@correo.xoc.uam.mx http://creativecommons.org/licenses/by/4.0/ medina-lópez et al. admet & dmpk 8(4) (2020) 411-423 412 factors that affect drug bioavailability related to the pharmaceutical dosage form and manufacturing process have been described [4-5]. techniques and strategies for the development of class iv drug formulations have also been discussed [6]. this information can support the design of better oral dosage forms of furosemide than those currently available. the oral bioavailability of furosemide has been reported to be 37–51 % [7] and 60–70 %, with a variable and erratic absorption [8]. for the reference product lasix®, an absolute bioavailability of 56 % has been observed [9]. this limited bioavailability could be associated with significant differences in the dissolution behaviour of furosemide commercial formulations shown by several authors. kaojarern et al. [10] reported an in vitro dissolution study of 13 brands of furosemide tablets (40 mg), of which only four fulfil the q pharmacopoeial specification. stüber et al. [11] studied the bioavailability of four furosemide drug products, of which three formulations have bioavailabilities of 81–83 % and lower in vitro dissolution performance than the reference in each of following conditions: ph 7.8/paddle at 25 rpm, ph 7.8/paddle at 50 rpm, ph 5.3/paddle at 50 rpm, and flow-through cell (100 ml/h)/ph 7.8. the difference in in vitro dissolution is more pronounced at ph 5.3/paddle at 50 rpm. currently, the official dissolution test for furosemide tablets is the usp apparatus 2 at 50 rpm with 900 ml of phosphate buffer ph 5.8 and no less than 80 % should be dissolved at 60 min [12]. a biowaiver monograph has been reported to waiver the in vivo bioequivalence of furosemide solid oral dosage forms by in vitro dissolution studies; however, given the available data, granero et al. [3] concluded that a biowaiver procedure for this drug cannot be justified. low solubility and permeability are problematic characteristics of class iv drugs; thus, the determination of in vitro release performance using different agitation rates, dissolution media, and dissolution apparatuses provides important information for improving the manufacture and evaluation of generic formulations. despite the wide use of usp basket and paddle apparatuses (usp apparatus 1 and 2, respectively) to monitor the physical quality of tablets and capsules, several investigations about the hydrodynamic environment that surrounds oral formulations have reported that these usp apparatuses do not adequately reproduce the natural environment of the gastrointestinal tract [13-15]; thus, it is necessary to document the in vitro release performance of poorly soluble drugs under different conditions to establish, in the best possible way, the environment in which the solid dosage forms will be within the gastrointestinal tract. further, alternative apparatuses must be developed to achieve this goal. the flowthrough cell method (usp apparatus 4) has been introduced as a dissolution apparatus to elucidate the rate and extent of the dissolution of drugs with low solubility under sink conditions [16]. usp apparatus 4 is more reliable, reproducible, and discriminative than the other methods [17], and it generates a hydrodynamic environment similar to that inside the gastrointestinal tract [18]. furthermore, an in vitro/in vivo correlation (ivivc) has been established between the in vitro data generated with this apparatus and in vivo parameters [19,20]. the aim of this study was to evaluate the in vitro release performance of furosemide reference tablets in the hydrodynamic environments generated by different usp apparatuses and dissolution media of physiological relevance to identify the rate and extent of furosemide dissolution under these conditions. this information will reflect the characteristics of the reference drug product that can be considered in the preparation of a generic formulation. lasix® tablets were tested with 0.1 m hydrochloric acid, acetate buffer (ph 4.5), and phosphate buffer (ph 6.8) using usp apparatuses 1 and 2 (at different agitation rates) and usp apparatus 4. figure 1. chemical structure of furosemide. admet & dmpk 8(4) (2020) 411-423 in vitro release studies of furosemide doi: http://dx.doi.org/10.5599/admet.801 413 experimental materials lasix® furosemide tablets (sanofi-aventis de mexico, s.a. de c.v., mexico city, mexico) were used. mexican health authorities have designated this formulation as a reference product to be used in dissolution and bioequivalence studies [21]. reagents hydrochloric acid, sodium acetate, acetic acid, and phosphate salts were purchased from j.t. bakermexico (xalostoc, mexico). furosemide reference standard was purchased from sigma-aldrich co. (st. louis, mo, usa). all samples were filtered through 0.45 µm nitrocellulose filters (millipore, ireland). standard solutions were prepared by serial dilutions of the stock solutions of furosemide (1 mg/ml) to achieve concentrations of 1.25–20 µg/ml. the dissolution media comprised 0.1 n hydrochloric acid, acetate buffer (ph 4.5), and phosphate buffer (ph 6.8). content uniformity and assay content uniformity and assay tests were performed with the drug product according to the procedures described in the usp [12]. usp basket and paddle apparatus dissolution profiles of furosemide were obtained using usp apparatuses 1 and 2 (model at-7 smart, sotax, basel, switzerland) with 25, 50, and 75 rpm agitation rates. additionally, pharmacopoeial dissolution conditions (usp apparatus 2 at 50 rpm with phosphate buffer (ph 5.8)) were tested [12]. an ultravioletvisible (uv/vis) spectrophotometer (model lambda 35, perkin elmer, usa) with 1 mm flow cells was used. the equipment was controlled by specific software designed by sotax. furosemide tablets were sprinkled on 900 ml of 0.1 n hydrochloric acid, acetate buffer (ph 4.5), and phosphate buffer (ph 6.8) at 37.0 ± 0.5 °c. samples were taken automatically every 5 min for 60 min. dissolved furosemide was quantified with standard calibration curves, in each dissolution medium, at 274 nm. flow-through cell method dissolution profiles of furosemide were obtained with usp apparatus 4 (model ce6, sotax ag, basel, switzerland) with 22.6 mm cells (i.d.). the laminar flow (originated with 6 g of glass beads) of 16 ml/min was tested. the dissolution media also comprised 0.1 m hydrochloric acid, acetate buffer (ph 4.5), and phosphate buffer (ph 6.8) at 37.0 ± 0.5 °c. samples were taken automatically every 5 min for 60 min. dissolved furosemide was quantified in a uv/vis spectrophotometer (model lambda 10, perkin elmer, usa) with 1 mm cells at 274 nm. for every trial, and depending on the schedule work, a standard calibration curve in 0.1 n hydrochloric acid, acetate buffer (ph 4.5) or phosphate buffer (ph 6.8) was prepared. dissolution data analysis dissolution profiles were compared with model-independent and model-dependent approaches. for model-independent comparisons, mean dissolution time (mdt) and dissolution efficiency (de) were calculated. mdt is the time to dissolve 63.2 % of the drug and was calculated according to the statistical moment’s theory [22,23]. other authors have indicated the mdt to be 62–64 % [24]. de is the area under the dissolution curve up to a certain time, t, expressed as a percentage of the area of the rectangle described by 100 % dissolution in the same period [25]. both parameters were calculated with the excel http://dx.doi.org/10.5599/admet.801 medina-lópez et al. admet & dmpk 8(4) (2020) 411-423 414 add-in ddsolver program [26]. for model-dependent comparisons, dissolution data were adjusted to a hyperbole equation (y = ax/b+x) and, using a and b parameters, time to 50 % dissolution (t50%) and time to 80 % dissolution (t80%) were calculated. the fit was calculated using sigmaplot software (version 11.0). for a complete comparison of dissolution data by a model-dependent approach, dissolution data were fitted to zero-order, first-order, higuchi, hixson-crowell, makoid-banakar, and weibull models. the model with the highest adjusted determination coefficient (r 2 adjusted) and lowest akaike information criterion (aic) is the best-fit model [27]. data analysis was carried out using excel add-in ddsolver program [28]. all statistical comparisons were carried out with student’s t-tests with significant differences at p < 0.05. results content uniformity and assay the drug product used met the content uniformity and assay tests specified in the usp. the percentage of furosemide in the content uniformity test ranged from 97.4–100.31 % (pharmacopoeial criteria, 85–115 %) and that in the assay test was 100.79 % (criteria, 90–110 %) [12]. dissolution profiles with usp basket and paddle apparatus dissolution profiles of furosemide obtained with the usp apparatuses 1 and 2 at different agitation rates and in different dissolution media are shown in figure 2. a limited amount of furosemide dissolved in 0.1 m hydrochloric acid at 60 min with both usp dissolution apparatuses (< 20 %). almost 60 % of the drug dissolved using usp apparatus 2 at 75 rpm with acetate buffer (ph 4.5) as the dissolution medium. a complete release of the drug was achieved using phosphate buffer (ph 6.8) at 50 and 75 rpm, independent of the usp apparatus used. to compare dissolution profiles between usp apparatuses 1 and 2, model-independent and -dependent parameters, at ph 6.8, were calculated, the results of which are shown in table 1. at 25 rpm and with both usp apparatuses, less than 65 % of furosemide dissolved; thus, the t80% was not calculated. at this agitation rate, significant differences in mdt values were observed (p < 0.05). at 50 rpm, significant differences in all calculated parameters were observed (p < 0.05). these results suggest a complete non-equivalence in the dissolution performance of furosemide between usp apparatuses 1 and 2 at these conditions (ph 6.8 and 50 rpm). at 75 rpm, significant differences in mdt, de, t50%, and t80% values were observed (p < 0.05). higher values of mdt, t50%, and t80% were observed with usp apparatus 1 than with usp apparatus 2, at 50 and 75 rpm; these findings could be attributed to slower in vitro dissolution rates in usp apparatus 1. dissolution data at ph 6.8, adjusted with different mathematical models, are shown in table 2. admet & dmpk 8(4) (2020) 411-423 in vitro release studies of furosemide doi: http://dx.doi.org/10.5599/admet.801 415 figure 2. dissolution profiles of furosemide reference tablets using usp apparatuses 1 and 2 with dissolution media in physiological ph range. mean, n = 6. table 1. model-independent and -dependent parameters of furosemide at ph 6.8. mean ± sem, n = 6. agitation rate (rpm) parameter usp apparatus 1 usp apparatus 2 25 diss. at 60 min (%) 64.61 ± 1.64 62.98 ± 5.87 mdt (min) 18.94 ± 0.16 17.11 ± 0.42* de (%) 44.42 ± 1.18 45.15 ± 4.49 t50% (min) 31.09 ± 1.89 33.07 ± 8.47 t80% (min) † † 50 diss. at 60 min (%) 93.90 ± 2.23 101.76 ± 0.54* mdt (min) 15.00 ± 0.81 4.99 ± 0.14* de (%) 70.56 ± 2.85 93.30 ± 0.43* t50% (min) 11.15 ± 1.35 2.07 ± 0.11* t80% (min) 32.27 ± 4.02 6.75 ± 0.30* 75 diss. at 60 min (%) 102.43 ± 0.30 102.02 ± 0.48 mdt (min) 7.49 ± 0.41 3.42 ± 0.01* de (%) 89.64 ± 0.77 96.20 ± 0.45* t50% (min) 4.13 ± 0.30 0.72 ± 0.02* t80% (min) 11.95 ± 0.75 2.58 ± 0.08* *p < 0.05. †data not calculated http://dx.doi.org/10.5599/admet.801 medina-lópez et al. admet & dmpk 8(4) (2020) 411-423 416 table 2. criteria used for the selection of the best-fit model at ph 6.8. mean, n = 6. parameter agitation rate (rpm) zeroorder firstorder higuchi hixsoncrowell makoidbanakar weibull usp apparatus 1 r 2 adjusted 25 0.5474 0.8935 0.9731 0.8155 0.9938 0.9997 50 -0.2118 0.9725 0.8774 0.8929 0.9932 0.9996 75 -3.99 0.9492 -0.2872 0.6144 0.9281 0.9994 aic 25 87.76 69.95 53.49 76.81 37.07 1.85 50 104.11 56.38 74.14 74.83 42.38 8.99 75 117.66 60.73 101.10 85.25 66.30 0.64 usp apparatus 2 r 2 adjusted 25 0.3232 0.7889 0.9521 0.6872 0.9952 0.9976 50 -17.85 0.9611 0.8774 -1.85 0.9932 0.9999 75 -144.29 0.7176 -50.97 -31.37 0.9016 0.9998 aic 25 90.15 73.14 55.70 79.54 27.37 17.52 50 120.96 36.84 74.14 98.05 42.38 -34.36 75 123.27 42.04 110.92 105.23 37.35 -40.46 considering the established criteria to choose the best-fit model (higher r 2 adjusted and lower aic values), the weibull function was the best mathematical equation to describe all dissolution data at ph 6.8. the expression of this function is as follows [26]: 𝐹 = 𝐹max [𝑒 − (𝑡−𝑇𝑖)𝛽 𝛼 ] , (1) where f is the percent of the drug that dissolved vs. t time, fmax is the maximum percent of the drug that dissolved at infinite time, α is the scale factor of the process, β is the shape factor, and ti is a location parameter of time in which the drug begins to dissolve. the furosemide dissolution data of both usp apparatuses adjusted to the weibull model dissolution profiles were statistically compared with td values derived from fitting to this equation. the td value represents the time interval necessary to dissolve or release 63.2 % of the drug present in the pharmaceutical dosage form [25] and coincides with mdt if the dissolution rate-time curve can be approximated by a monoexponential equation [22]. the mean values of α, β, ti, fmax, and td are shown in table 3. significant differences were observed in all comparisons (p < 0.05). table 3. weibull parameters and td values at ph 6.8. mean, n = 6. agitation rate (rpm) α β ti fmax td ± sem (min) usp apparatus 1 25 10.25 0.63 3.12 95.48 50.15 ± 9.38 50 5.81 0.59 2.70 111.03 22.78 ± 3.93 75 13.30 1.15 0.72 102.43 7.64 ± 0.43 usp apparatus 2 25 30.04 0.94 -0.11 70.50 25.63 ± 3.51* 50 2.75 0.84 1.53 101.76 4.72 ± 0.18* 75 1.87 0.77 -0.39 102.03 1.80 ± 0.12* *p < 0.05. model-dependent comparisons (by comparing td values) indicated that usp apparatuses 1 and 2 at 50 and 75 rpm generated different dissolution profiles. as both usp dissolution apparatuses and agitation rates created different hydrodynamic environments, these results were expected. dissolution profiles with flow-through cell method dissolution profiles of furosemide reference tablets using usp apparatus 4 with 0.1 m hydrochloric acid, acetate buffer (ph 4.5), phosphate buffer (ph 6.8), and laminar flow of 16 ml/min are shown in figure 3. admet & dmpk 8(4) (2020) 411-423 in vitro release studies of furosemide doi: http://dx.doi.org/10.5599/admet.801 417 figure 3. dissolution profiles of furosemide reference tablets using usp apparatus 4 with dissolution media in physiological ph range. mean ± sd, n = 12. with the flow-through cell method, less than 20 % of furosemide dissolved when 0.1 m hydrochloric acid and acetate buffer (ph 4.5) were used, whereas almost 90 % of the drug was released with phosphate buffer (ph 6.8). slower dissolution rates with usp apparatus 4 than with usp apparatuses 1 and 2 have been found; however, in this case, a dissolution medium with low ph was an important factor for the low dissolution of furosemide. as more than 80 % of the drug dissolved at 60 min at ph 6.8 and, for comparative purposes, the dissolution profile of furosemide tablets was obtained using a pharmacopoeial method. a test was carried out with usp apparatus 2 at 50 rpm with 900 ml of phosphate buffer (ph 5.8) (q = 80 % at 60 min), the results of which are shown in figure 4. figure 4. dissolution profiles of furosemide reference tablets using the pharmacopoeial conditions (usp 2) and flow-through cell method (usp 4) with different dissolution media. mean ± sd, n = 12. under official conditions, the furosemide tablets met the q pharmacopoeial specification (> 80 % dissolved at 60 min). the dissolution profiles of usp apparatuses 2 and 4 were similar (f2 > 50). this result suggested that a pharmacopoeial method (usp apparatus 2) can produce a similar dissolution profile to that obtained with equipment (usp apparatus 4) that generates a hydrodynamic environment similar to that inside the gastrointestinal tract and for which a correlation with in vivo data has been shown [19,20]. for a complete comparison between the profiles, model-independent and -dependent parameters were http://dx.doi.org/10.5599/admet.801 medina-lópez et al. admet & dmpk 8(4) (2020) 411-423 418 calculated and statistically compared, the results of which are shown in table 4. table 4. model-independent and -dependent parameters of furosemide. mean ± sem, n = 12. parameter usp apparatus 2 (ph 5.8) usp apparatus 4 (ph 6.8) diss. at 60 min (%) 102.55 ± 0.61 89.49 ± 1.50* mdt (min) 15.61 ± 0.68 15.28 ± 1.06 de (%) 75.93 ± 1.52 66.57 ± 1.44* t50% (min) 10.86 ± 0.79 13.35 ± 1.10 t80% (min) 25.99 ± 1.50 36.64 ± 2.34* *p < 0.05. significant differences were observed in percent dissolved at 60 min, de, and t80% values (p < 0.05), whereas there was no difference in mdt and t50%, between usp apparatuses 2 and 4. if we consider mdt and t50% as parameters that reflect the in vitro dissolution rate, this is similar between the conditions (usp apparatus 2 at 50 rpm and phosphate buffer (ph 5.8)/usp apparatus 4 with a laminar flow of 16 ml/min and phosphate buffer (ph 6.8)), at least until the time at which 63.2 % of the drug is dissolved. furosemide dissolution data obtained with pharmacopoeial conditions and the flow-through cell method, adjusted to mathematical models, are shown in table 5. table 5. criteria used for the selection of the best-fit model. mean, n = 12. parameter usp apparatus zero-order first-order higuchi hixsoncrowell makoidbanakar weibull r 2 adjusted 2 (ph 5.8) 0.1571 0.9788 0.9152 0.9882 0.9770 0.9994 4 (ph 6.8) 0.0115 0.9205 0.8048 0.8488 0.9743 0.9991 aic 2 (ph 5.8) 104.08 58.28 75.30 52.62 30.04 8.29 4 (ph 6.8) 101.87 66.11 81.14 74.67 57.47 15.79 the weibull function was the best-fit model to describe the in vitro release performance of furosemide reference tablets in these two dissolution apparatuses. weibull parameters and td values are shown in table 6. no significant differences were observed in the td data between the apparatuses (p > 0.05). table 6. weibull parameters and td values. mean, n = 12. usp apparatus α β ti fmax td ± sem (min) 2 (ph 5.8) 626.71 1.13 -1.99 110.39 20.89 ± 3.93 4 (ph 6.8) 12.22 0.89 3.12 92.57 16.63 ± 1.43 *p < 0.05 the dissolution of furosemide reference tablets using usp apparatus 4 exceeded the q pharmacopoeial criterion (only at ph 6.8) set for certain conditions in the usp apparatus 2. when the dissolution data of usp apparatus 4 were compared with those of usp apparatus 2 (50 rpm and phosphate buffer ph 5.8), an equivalent in vitro release performance was achieved based on f2, mdt, t50%, and td comparisons. under all conditions used, furosemide dissolution profiles were well described by weibull function. the shape factor of the weibull function characterises the dissolution profile as exponential (β = 1); sigmoidal, with upward curvature followed by a turning point (β > 1); or parabolic, with steeper initial slope consistent with exponential (β < 1) [28]. in this case, furosemide tablets evaluated with usp apparatus 1 at 75 rpm and ph 6.8, as well as with usp apparatus 2 at 50 rpm and ph 5.8, generated β values > 1, meaning sigmoidal profiles. admet & dmpk 8(4) (2020) 411-423 in vitro release studies of furosemide doi: http://dx.doi.org/10.5599/admet.801 419 discussion several authors have studied the effect of the hydrodynamic environment on the tablet dissolution rate. wu et al. [29] studied the rate process that underlies tablet dissolution to understand the role of external hydrodynamics on mass transfer rate and film thickness during dissolution. shah et al. [30] stated that the proper medium and appropriate rotational speed of the basket and paddle are of great importance to assure that the test procedure used is useful and discriminatory. additionally, levy et al. [31] concluded that the in vitro dissolution rate correlates with the in vivo absorption rate only at a low agitation rate (55 rpm). these results support the search for better dissolution schemes, especially with drugs with poor solubility. poor permeability is a problem in the design of solid oral dosage forms with good bioavailability; thus, it is important to understand the influence of agitation rate, usp apparatus, and dissolution media on the in vitro release performance of class iv drugs. the search for adequate dissolution conditions is not limited to poorly soluble drugs or the use of usp apparatuses 1 and 2. shabir [32] indicated that the rate of in vitro release of a hydrosoluble drug can be accurately controlled through the usp apparatus. shabir worked with atenolol (class iii drug) generic tablets using usp basket and paddle apparatuses. although these dissolution apparatuses are currently the most popular methods, gao [33] explains that both methods are operated under closed finite sink conditions and cannot mimic the conditions in the gastrointestinal tract. the flow-through cell method has gained recent acceptance into the dissolution field for its versatility in the testing of novel dosage forms where traditional dissolution apparatuses and methods have failed [34]. usp apparatus 4 has several advantages: 1. sink conditions can be maintained for poorly soluble drugs throughout the dissolution run; 2. it is easy to change media (suggested media is physiological ph range) and modify flow rate to simulate in vivo conditions; 3. it simulates intraluminal hydrodynamics efficiently; 4. it can be modified for different dosage forms; and 5. it measures the in vitro release rate profile as an output that is similar to the shape of an in vivo profile [35]. american and european pharmacopoeias suggest three flow rates for testing with usp apparatus 4 (4, 8, and 16 ml/min) [36]. in this in vitro release study of furosemide reference tablets, we used only the flow rate of 16 ml/min as the sotax equipment model ce6 has a working flow range of 10‒50 ml/min. based on these characteristics, it is possible to establish a meaningful ivivc with usp apparatus 4. when a meaningful ivivc has been established, it can be used as a surrogate for and to minimize the number of bioequivalence studies during drug product development [37]. some authors have reported a better estimate of the absorption rate of cilostazol [19] and diclofenac sodium [20] formulations, both drugs with solubility problems, with the flow-through cell method. some drugs can be dissolved using dissolution media in the physiological ph range (ph 1.2, 4.5, and 6.8) or biorelevant media, such as fassif and fessif (media that simulate the absence or presence of food, respectively) [38], to document the release performance of oral dosage forms through the gastrointestinal tract. this is especially important with poorly soluble drugs, such as furosemide. based on the physicochemical characteristics of this drug, dissolution at low ph is not physiologically relevant. for a complete dissolution scheme in a physiological ph range, the food and drug administration recommends dissolution tests with a 0.1 n hydrochloric acid or simulated gastric fluid usp without enzymes, ph 4.5 or ph 6.8 buffer, or simulated intestinal fluid usp without enzymes [39]. the dissolution of furosemide reference tablets was carried out under these conditions, with the exception of simulated fluids, and the best results were observed with phosphate buffers at ph 5.8 and ph 6.8. more than 80 % of the drug dissolved at 60 min and time parameters, such as mdt, t50%, and t80% were calculated. at 50 rpm, significant differences in http://dx.doi.org/10.5599/admet.801 medina-lópez et al. admet & dmpk 8(4) (2020) 411-423 420 all calculated parameters (usp apparatus 1 vs. 2) were observed. differences between both usp apparatuses were expected owing to the different hydrodynamics of each apparatus; however, it is necessary to understand at what agitation rate this difference is more evident. despite the wide use of usp apparatuses 1 and 2 at different agitation rates, they have still been evaluated by several authors in terms of the surrounding hydrodynamic environments, which do not adequately reproduce the natural environment of the gastrointestinal tract [13-15,32,40]. the fitting of dissolution data to mathematical models was carried out without any physiological significance to discover the best equation to explain the in vitro release performance of furosemide reference tablets. these models were used to facilitate the analysis and interpretation of dissolution data because they describe the dissolution profiles as a function of the few parameters that can be statistically compared [41]. han et al. [42] documented the first-order kinetics as the best-fit model to adjust furosemide dissolution data from commercial tablets; however, an incomplete fit scheme was created as only zero-order and first-order kinetics were used to adjust their dissolution data. in our study, after testing several equations (including common dissolution kinetics), weibull function was the best mathematical model to explain the dissolution performance of furosemide. the development of more discriminative methods than using pharmacopoeial conditions to evaluate the biopharmaceutical quality of generic formulations has been documented. studies with class ii drugs, such as carbamazepine [43], meloxicam [44], and naproxen sodium [45], tested with usp apparatuses 2 and 4 have shown that usp apparatus 2 may not reflect the dissolution performance of generic formulations and references. the choice of the hydrodynamic environment for the drug release is key to identify a meaningful ivivc [35]. for class ii drugs, ivivcs have been identified and, with in vitro studies, they provide a good estimate of the absorption rate of class ii drugs. several authors have shown this important association only with usp apparatus 4 [19,20,35]. however, similar dissolution profiles for usp apparatuses 2 and 4 have been reported for naproxen sodium tablets [45] and ibuprofen suspensions [46]. these results are important where no flow-through cell method is available, and an equivalent hydrodynamic environment is required to test solid dosage forms. conclusions the in vitro release performance of furosemide reference tablets was determined using usp apparatuses 1 and 2 at different agitation rates and dissolution media of physiological relevance. a limited amount of furosemide dissolved with both at ph 1.2 and 4.5. better results were obtained with a dissolution medium of ph 6.8. although usp basket and paddle apparatuses are the most widely used, it is important to take advantage of usp apparatus 4 for the evaluation of furosemide tablets under the hydrodynamic environment that this equipment generates. all information collected is important to reduce the negative clinical impact that this class iv drug presents. this furosemide reference product is the comparative formulation for generic drug products; thus, it is important to understand the in vitro release performance under all possible schemes for the design of better commercial formulations. acknowledgements: authors thanks to qfb alexander domínguez reyes for his technical assistance. conflict of interest: declared none admet & dmpk 8(4) (2020) 411-423 in vitro release studies of furosemide doi: http://dx.doi.org/10.5599/admet.801 421 references [1] h.s. abou-auda, m.j. al-yamani, a.m. morad, s.a. bawazir, s.z. khan, k.i. al-khamis. highperformance liquid chromatographic determination of furosemide in plasma and urine and its use in bioavailability studies. journal of chromatography b 710 (1998) 121-128. [2] a. berthod. hydrophobicity of ionizable compounds. a theoretical study and measurements of diuretic octanol-water partition coefficients by countercurrent chromatography. analytical chemistry 71 (1999) 879-888. [3] g.e. granero, m.r. longhi, m.j. mora, h.e. junginger, k.k. midha, v.p. shah, s. stavchansky, j.b. dressman, d.m. barends. biowaiver monographs for immediate-release solid oral dosage forms: furosemide. journal of pharmaceutical sciences 99 (2010) 2544-2556. [4] g. tambosi, p.f. coelho, l. soares, i.c. schmücker lenschow, m. zétola, h.k. stulzer, b. ramos pezzini. challenges to improve the biopharmaceutical properties of poorly water-soluble drugs and the application of the solid dispersion technology. revista matéria 23 (2018) e-12224. [5] a. garcía-arieta. interactions between active pharmaceutical ingredients and excipients affecting bioavailability: impact on bioequivalence. european journal of pharmaceutical sciences 65 (2014) 89-97. [6] r. ghadi, n. dand. bcs class iv drugs: highly notorious candidates for formulation development. journal of controlled release 248 (2017) 71-95. [7] m.r. kelly, r.e. cutler, a.w. forrey, b.m. kimpel. pharmacokinetics of orally administered furosemide. clinical pharmacology and therapeutics 15 (1973) 178˗186. [8] s. sweetman. martindale: the complete drug reference. (electronic version) pharmaceutical press. thomson/micromedex, london, uk/greenwood village, colorado. 2009. [9] a. grahnén, m. hammarlund, t. lundqvist. implications of intraindividual variability in bioavailability studies of furosemide. european journal of clinical pharmacology 27 (1984) 595˗602. [10] s. kaojarern, o. poobrasert, a. utiswannakul, u. kositchaiwat. bioavailability and pharmacokinetics of furosemide marketed in thailand. journal of the medical association of thailand 73 (1990) 191˗197. [11] w. stüber, e. mutschler, d. steinbach. the pharmaceutical and biological availability of commercial preparations of furosemide. arzneimittelforschung drug research 32 (1982) 693-697. [12] united states pharmacopeia and national formulary usp 42-nf 37; united states pharmacopeial convention, inc: rockville md; 2019. [13] m. morihara, n. aoyagi, n. kaniwa, n. katori, s. kojim. hydrodynamic flows around tablets in different pharmacopeial dissolution tests. drug development and industrial pharmacy 28 (2002) 655-662. [14] k. greco, t.l. bergman, r. bogner. design and characterization of laminar flow-through dissolution apparatus: comparison of hydrodynamic conditions to those of common dissolution techniques. pharmaceutical development and technology 16 (2011) 75-87. [15] v. todaro, t. persoons, g. grove, m.a. healy, d.m. d’arcy. characterization and simulation of hydrodynamics in the paddle, basket and flow-through dissolution testing apparatuses a review. dissolution technologies 24 (2017) 24-36. [16] s.n. bhattachar, j.a. wesley, a. fioritto, p.j. martin, s.r. babu. dissolution testing of a poorly soluble compound using the flow-through cell dissolution apparatus. international journal of pharmaceutics 236 (2002) 135-143. [17] e. beyssac, j. lavigne. dissolution study of active pharmaceutical ingredients using the flow through apparatus usp 4. dissolution technologies 12 (2005) 23-25. [18] n. fotaki, c. reppas. the flow through cell methodology in the evaluation of intralumenal drug release characteristics. dissolution technologies 12 (2005) 17-21. http://dx.doi.org/10.5599/admet.801 medina-lópez et al. admet & dmpk 8(4) (2020) 411-423 422 [19] j. jinno, n. kamada, m. miyake, k. yamada, t. mukai, m. odomi, h. toguchi, g.g. liversidge, k. higaki, t. kimura. in vitro-in vivo correlation for wet-milled tablet of poorly water-soluble cilostazol. journal of controlled release 130 (2008) 29-37. [20] e. jantratid, v. de maio, e. ronda, v. mattavelli, m. vertzoni, j.b. dressman. application of biorelevant dissolution tests to the prediction of in vivo performance of diclofenac sodium from an oral modified-release pellet dosage form. european journal of pharmaceutical sciences 37 (2009) 434-441. [21] listado actualizado de medicamentos de referencia 2019/01, https://www.gob.mx/cms/uploads/attachment/file/441323/listado_de_medicamentos_de_referencia_.pdf, accessed on 04 march 2020. [22] y. tanigawara, k. yamaoka, t. nakagawa, t. unzo. new method for the evaluation of in vitro dissolution time and disintegration time. chemical and pharmaceutical bulletin 30 (1982) 10881090. [23] p.s. collier. the interpretation of in vivo mean dissolution time data. in: benet l.z., levy g., ferraiolo b.l. (eds) pharmacokinetics. springer-verlang. boston, usa. 1984. [24] f. podczeck. comparison of in vitro dissolution profiles by calculating mean dissolution time (mdt) or mean residence time (mrt). international journal of pharmaceutics 97 (1993) 93-100. [25] p. costa, j.m. sousa lobo. modeling and comparison of dissolution profiles. european journal of pharmaceutical sciences 13 (2001) 123-133. [26] y. zhang, m. huo, j. zhou, a. zou, w. li, c. yao, s. xie. dd solver: an add-in program for modeling and comparison of drug dissolution profiles. the aaps journal 12 (2010) 263˗271. [27] n. yuksel, a.e. kanik, t. baykara. comparison of in vitro dissolution profiles by anova-based, model-dependent and independent-methods. international journal of pharmaceutics 209 (2000) 57-67. [28] f. langenbucher. linearization of dissolution curves by the weibull distribution. journal of pharmacy and pharmacology 24 (1972) 979-981. [29] y. wu, d.o. kildsing, e.s. ghaly. effect of hydrodynamic environment on tablet dissolution rate. pharmaceutical development and technology 9 (2004) 25-37. [30] v.p. shah, m. gurbarg, a. noory, s. dighe, j.p. skelly. influence of higher rates of agitation on release patterns of immediate-release drug products. journal of pharmaceutical sciences 81 (1992) 500-503. [31] g. levy, b.a. hayes. physicochemical basis of the buffered acetylsalicylic acid controversy. new england journal of medicine 262 (1960) 1053-1058. [32] g.a. shabir. evaluation of usp basket and paddle dissolution methods using different generic atenolol tablets. turkish journal of pharmaceutical sciences 8 (2011) 253-260. [33] z. gao. in vitro dissolution testing with flow-through method: a technical note. aaps pharmscitech 10 (2009) 1401-1405. [34] i. singh, h.y. aboul-eneim. advantages of usp apparatus iv (flow-through cell apparatus) in dissolution studies. journal of the iranian chemical society 3 (2006) 220-222. [35] j.b. fang, v.k. robertson, a. rawat, t. flick, z.j. tang, n.s. cauchon, j.s. mcelvain. development and application of a biorelevant dissolution method using usp apparatus 4 in early phase formulation development. molecular pharmaceutics 7 (2010) 1466-1477. [36] b. steffansen, b. brodin, c. und nielsen. molecular biopharmaceutics: aspects of drug characterization, drug delivery and dosage form evaluation. ulla pharmacy series. pharmaceutical press. london, united kingdom. 2009. [37] c. larsen, s.w. larsen, h. jensen, a. yaghmur, j. østergaard. role of in vitro release models in formulation development and quality control of parenteral depots. expert opinion on drug delivery 6 (2009) 1283-1295. https://www.gob.mx/cms/uploads/attachment/file/441323/listado_de_medicamentos_de_referencia_.pdf https://www.gob.mx/cms/uploads/attachment/file/441323/listado_de_medicamentos_de_referencia_.pdf admet & dmpk 8(4) (2020) 411-423 in vitro release studies of furosemide doi: http://dx.doi.org/10.5599/admet.801 423 [38] j.h. fagerberg, o. tsinman, n. sun, k. tsinman, a. avdeef, c.a.s. bergström. dissolution rate and apparent solubility of poorly soluble drugs in biorelevant dissolution media. molecular pharmaceutics 7 (2010) 1419-1430. [39] us dhhs, fda, cder, guidance for industry: waiver if in vivo bioavailability and bioequivalence studies for immediate-release solid oral dosage forms based on a biopharmaceutics classification system, u.s. department of health and human services, food and drug administration, center for drug evaluation and research (cder) (2017), rockville, md, usa. [40] s. missaghi, r. fassihi. release characterization of dimenhydrinate from an eroding and swelling matrix: selection of appropriate dissolution apparatus. international journal of pharmaceutics 29 (2005) 35-42. [41] e. adams, d. coomans, j. smeyers-verbeke, d.l. massart. non-linear mixed effects models for the evaluation of dissolution profiles. international journal of pharmaceutics 240 (2002) 37-53. [42] y.k. han, l.d. simionato, r.g. calvo, m.b. mattei, a.i. segall. comparison of dissolution profiles of furosemide tablets available in the argentinian market. journal of applied solution chemistry and modeling 3 (2014) 186-193. [43] j.r. medina, d.k. salazar, m. hurtado, a.r. cortés, a.m. domínguez-ramírez. comparative in vitro dissolution study of carbamazepine immediate-release products using the usp paddles method and the flow-through cell system. saudi pharmaceutical journal 22 (2014) 141-147. [44] j.r. medina-lópez, j.a. orozco-juárez, m. hurtado. dissolution performance of meloxicam formulations under hydrodynamics of usp paddle apparatus and flow-through cell method. international journal of applied pharmaceutics 11 (2019) 182-188. [45] j.r. medina, a. uribe, m. hurtado, a.m. domínguez-ramírez. in vitro equivalence study of generic naproxen tablets using the usp paddle apparatus and the flow-through cell method. international journal of pharmacy and pharmaceutical sciences 7 (2015) 348-354. [46] j.r. medina, m. cortes, e. romo. comparison of the usp apparatus 2 and 4 for testing the in vitro release performance of ibuprofen generic suspension. international journal of applied pharmaceutics 9 (2017) 90-95. ©2020 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://dx.doi.org/10.5599/admet.801 http://creativecommons.org/licenses/by/3.0/ do you know your r2? doi: https://doi.org/10.5599/admet.888 1 admet& dmpk 9(1) (2021) 69-74; doi: https://doi.org/10.5599/admet.888 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index commentary do you know your r 2 ? alex avdeef in-adme research, 1732 first avenue #102, new york, ny 10128 usa corresponding author: e-mail: alex@in-adme.com; tel.: +1-646-678-5713, orcid id: alex avdeef: orchid.org/0000-0002-3139-5442 received: july 22, 2020; revised: august 10, 2020; published: august 30, 2020 abstract the prediction of solubility of drugs usually calls on the use of several open-source/commercially-available computer programs in the various calculation steps. popular statistics to indicate the strength of the prediction model include the coefficient of determination (r 2 ), pearson’s linear correlation coefficient (rpearson), and the root-mean-square error (rmse), among many others. when a program calculates these statistics, slightly different definitions may be used. this commentary briefly reviews the definitions of three types of r 2 and rmse statistics (model validation, bias compensation, and pearson) and how systematic errors due to shortcomings in solubility prediction models can be differently indicated by the choice of statistical indices. the indices we have employed in recently published papers on the prediction of solubility of druglike molecules were unclear, especially in cases of drugs from ‘beyond the rule of 5’ chemical space, as simple prediction models showed distinctive ‘bias-tilt’ systematic type scatter. ©2021 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords coefficient of determination, linear correction coefficient, root-mean-square error, linear regression. introduction the ubiquitous coefficient of determination (r 2 ) and root-mean-square error (rmse) are statistics which enumerate the strength of a physical property prediction model [1-4]. yet their estimated values depend conditionally not only on random errors in the observed data but also on systematic errors generated as a result of limitations in a particular prediction model. when comparing the strength of prediction from different studies based on different models, it is vital to ensure that the same kinds of statistics are invoked. here, the commentary confines the discussion to statistics derived by linear regression of scatter plots of log s0 obs vs. log s0 calc (log s0 = logarithm of aqueous intrinsic solubility), with observed values treated as dependent variables (y-axis) and calculated values treated as independent variables (x-axis) [3]. three types of r 2 and rmse statistics are considered here: ❶ model validation (r 2 val, rmseval), ❷ validation with ‘bias’ compensation (r 2 bias, rmsebias), and ❸ validation with ‘bias-tilt’ compensation, i.e., pearson’s approach [4] (r 2 peason, rmsepearson). whether r 2 or rmse is a better statistic to use is beyond the scope of this commentary. https://doi.org/10.5599/admet.888 https://doi.org/10.5599/admet.888 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:alex@in-adme.com http://creativecommons.org/licenses/by/4.0/ alex avdeef admet & dmpk 9(1) (2021) 69-74 70 the precise definitions of r 2 and rmse are especially pertinent to prediction competitions, for ranking performances consistently. the second ‘solubility challenge’ (sc-2) has been described recently [5], modeled after the first competition (sc-1) which took place in 2008 [6]. in sc-2, two test sets of highlycurated aqueous intrinsic solubility data were presented to the computational community to challenge participants to predict the solubility values of the druglike molecules. concomitant to the sc-2 competition, we also published predictions [7] of the two test sets in sc-2, as well as the test set in sc-1. figure 1. correlation plots (ps0 = -log s0) – three distinct definitions of coefficients of determination (val = model validation, bias = bias compensation, and pearson), illustrated by simulated data (squares) containing random and systematic errors. the statistics arising from case ❸ place the prediction in the most favorable light (with rmse referring to the experimental random error scatter about the green dash-dot curves). those of case ❶ refer to model validation (with rmse referring to the data scatter about the solid black ‘identity’ diagonal lines). the dashed red lines correspond to the intermediate case ❷. here, we calculated the three types of statistics in order to clarify and put into context the statistics we have employed in our recent studies [7,8], so as to allow consistent comparison of the strengths of our prediction models to those of others [5,6]. method figure 1 illustrates the three definitions of the coefficient of determination and the corresponding rmse, with the aid of simulated data. the ‘observed’ data contain random errors of ±(0.24-0.57). the ‘calculated’ data either have no errors (frame a) or have systematic errors (frames b-d). frame (a) depicts a scatter plot based on a strong prediction model, where the statistics are mainly indications of the random ‘experimental’ errors. the data in frame (b) have a superimposed negative bias, but there is no distortion to the slope in the scatter plot (i.e., no ‘tilt’ to the data trend). frame (c) has no added bias, but there is a substantial tilt to the data trend (or negative bias without tilt). frame (d) contains both a positive bias and a admet & dmpk 9(1) (2021) 69-74 do you know your r 2 ? doi: https://doi.org/10.5599/admet.888 71 tilt added to the random errors (or positive bias without tilt). the simulated prediction model is assumed to have been ‘trained’ using a large diverse data set. the strength of the prediction can be determined by a randomly-selected smaller set of ‘test’ compounds not used in the training. three types of statistics may be of interest in the analyzed scatter plot for the test compounds: ❶ r 2 val and rmseval may be used to assess how effectively the training-set derived model predicts the test set (i.e., model validation), as indicated by the dispersion of data about the ‘identity’ line (y = x). ❷ r 2 bias and rmsebias may be used when the prediction model generates a constant bias (a) in the scatter plot, as indicated by the dispersion of data about the unit-slope regression line, displaced from the identity line by the extent of bias (y = a + x). ❸ r 2 pearson and rmsepearson pearson’s statistics [4] are based on regression analysis (y = a + bx) of a scatter plot showing both bias (intercept, a) and ‘tilt’ (slope, b). the statistics depend on the dispersions about the (non-unit slope and non-zero intercept) regression line. the above considerations suggest three constraints for linear regression, y = a + bx: ❶ constrained a = 0 (no bias) and b = 1 (no tilt), ❷ constrained b = 1 (no tilt) and determined a (bias), ❸ both a and b determined (without constraints). the statistics which are calculated in these three cases can be quite different, depending on the type and extent of systematic errors. for case ❶, the explicit equations for the two statistics are: 𝑟val 2 = 1 − ∑ 𝑖 (𝑦i obs −𝑦i calc ) 2 ∑ 𝑖 (𝑦i obs− <𝑦obs>) 2 (1) rmseval = √ ∑ i (𝑦𝑖 obs−𝑦i calc) 2 𝑛 (2) where y = log s0 and <y obs > is the mean of log s0 values. the r 2 val in eq. (1) is often called the ‘coefficient of determination,’ or simply, ‘r-squared.’ according to eq. (1), if all the calculated log s0 values match the observed values (‘perfect fit’), then r 2 val = 1. inappropriate/poor models can lead to r 2 val < 0. for case ❷ statistics, the bias (a) is incorporated into the expressions: 𝑟bias 2 = 1 − ∑ i (𝑦i obs − 𝑎 − 𝑦i calc ) 2 ∑ i (𝑦i obs− <𝑦obs>) 2 (3) rmsebias = √ ∑ i (𝑦i obs− 𝑎− 𝑦i calc) 2 𝑛 −1 (4) for case ❸ statistics, both the bias (a) and the slope factor (b) are incorporated into the expressions: 𝑟pearson 2 = 1 − ∑ i (𝑦i obs − 𝑎 − 𝑏𝑦i calc ) 2 ∑ 𝑖 (𝑦i obs− <𝑦obs>) 2 (5) rmsepearson = √ ∑ i (𝑦i obs− 𝑎− 𝑏𝑦i calc) 2 𝑛 −2 (6) pearson’s r is more explicitly calculated as [4] https://doi.org/10.5599/admet.888 alex avdeef admet & dmpk 9(1) (2021) 69-74 72 𝑟pearson = ∑ i (𝑦i obs − <𝑦obs>) 2 · (𝑦i calc − <𝑦calc>) 2 √∑ j (yj obs− <yobs>) 2 · ∑ k (yk calc− <ycalc>) 2 (7) in the absence of systematic errors (fig. 1a), it does not matter which of the three definitions is used. the statistics take on the same values. however, if there is bias (without tilt) in the fit (fig. 1b), then cases ❷ and ❸ produce comparable statistics, which are ‘better’ than then those of case ❶. when there is a tilt in the trend or when there is a combined tilt and bias, then the three sets of statistic produce different values, as illustrated in figures 1c,d. for such cases, r 2 pearson > r 2 bias > r 2 val, while rmsepearson < rmsebias < rmseval. the greater the systematic distortion, the greater the difference between the three sets of metrics. if the source of random errors is solely from the data, then rmsepearson may be a good indicator of effective measurement errors; rmseval is the better indicator of overall solubility prediction. both eq. (1) and eq. (7) are popularly used. but in many publications it is not clear which was actually applied. also, it may not be readily apparent which r 2 is calculated in some open-source/commercial programs from the provided documentation. this can lead to some confusion when comparing statistics between independent predictions of solubility coming from different laboratories, using different methods and programs. results and discussion in our previous publications [7,8] we listed r 2 bias and rmsebias in our scatter plots without the subscript designations, thus inadvertently ascribing them to eqs. (1) and (2) definitions. in most cases, the differences between the two types of statistics are negligible, but not in all cases. for example, the general solubility equation (gse) and the abraham solvation equation (absolv) models used to predict the solubility of drugs from ‘beyond the rule of 5’ chemical space showed (e.g., figs. 4b, 5b in ref. [8]) distinctive bias-tilt type scatter, with different degrees of systematic aberrations introduced by the limitations in the models when applied to such large molecules (similar to what is shown in fig. 1d here). in contrast, the random forest regression (rfr) model (e.g., fig. 13c in ref. [7] and fig. 6c in ref. [8]) was relatively free of such systematic distortions (similar to what is shown in fig. 1a here), and consequently the three sets of statistics are nearly the same in the rfr examples (cf., tables below). sample calculations and possible confusion in ref. [7], the gse was used to predict the 28 intrinsic solubility values taken from the sc-1 competition [6]. since the gse requires no ‘training,’ we expected to see some bias and tilt in the resulting scatter plots. fig. 11b in ref. [7] shows a log s0 obs vs. log s0 calc scatter plot (cf., table 1 below). the statistics listed in that figure are r 2 bias = 0.26 and rmsebias = 1.23. we used sigmaplot to construct publication-quality figures. in the accompanying statistics calculation, the bias was determined by fitting the function: log s0 obs = a + blog s0 cal , where the b regression coefficient was constrained to be 1.0, so the determined bias = a. in the above fig. 11b example, the calculated bias = 0.61 log unit. sigmaplot calculated the values ‘rsqr’ = 0.26 and ‘standard error of estimate’ = 1.23, which we listed in the plot. this is consistent with the calculations of eqs. (3) and (4). however, eqs. (1) and (2) produce r 2 val = 0.07 and rmseval = 1.34. furthermore, for the same example, the open-source default cor(x,y) function [9] calculated ‘r-squared’ = 0.45 and the sample script function defined by walters [2] calculated ‘rmserror’ = 1.07. this is consistent with the calculations of eqs. (5) and (6) – pearson’s equations. so, the three ‘r-squared’ statistics were calculated as 0.07, 0.26, and 0.45 and the corresponding ‘rmse’ admet & dmpk 9(1) (2021) 69-74 do you know your r 2 ? doi: https://doi.org/10.5599/admet.888 73 values were 1.34, 1.23, and 1.07, respectively. this can be confusing when comparing prediction models. it’s not that any of these values is wrong – it’s just that different equations/assumption are used/implied. generally, the appropriate definition of the coefficient of determination is according to eq. (1) and the rmse is according to eq. (2), since these focus on the actual strength of the model in linking prediction to measurement. table 1. recalculated statistics for the scatter plots in ref. [7] type a fig. in ref. [7] r 2 pearson eq. (5) r 2 bias eq. (3) b r 2 val eq. (1) rmsepearson eq. (6) rmsebias eq. (4) b rmseval eq. (2) bias eq. (3) gse, acids 6a 0.62 0.61 0.58 1.21 1.24 1.27 -0.29 gse, bases 6b 0.60 0.57 0.56 1.16 1.21 1.21 -0.14 gse, neutrals 6c 0.61 0.54 0.54 1.05 1.15 1.18 -0.30 gse, zwitterions 6d 0.24 0.07 0.02 1.38 1.54 1.57 0.34 absolv, acids 7a 0.66 0.66 0.65 1.14 1.15 1.16 -0.15 absolv, bases 7b 0.64 0.64 0.62 1.10 1.10 1.13 -0.28 absolv, neutrals 7c 0.61 0.61 0.61 1.05 1.05 1.05 -0.11 absolv, zwitterions 7d 0.68 0.68 0.67 0.90 0.90 0.92 -0.20 rfr 8a 0.98 0.98 0.98 0.28 0.28 0.28 0.00 rfr 8b 0.90 0.89 0.90 0.60 0.60 0.60 -0.02 rfr, zwitterions 8b -inset 0.91 0.91 0.91 0.45 0.45 0.45 0.01 gse, test set 1 11a 0.78 0.78 0.73 0.97 0.97 1.01 -0.41 gse, test set 2 11b 0.45 0.26 0.07 1.07 1.23 1.34 -0.61 gse, test set 3 11c 0.46 0.26 0.20 0.94 1.10 1.13 -0.31 gse, test set 4 11d 0.69 0.69 0.68 1.23 1.24 1.25 -0.08 absolv, test set 1 12a 0.77 0.69 0.58 0.98 1.15 1.27 -0.65 absolv, test set 2 12b 0.55 0.55 0.35 0.98 0.98 1.13 -0.62 absolv, test set 3 12c 0.47 0.36 0.26 0.94 1.02 1.10 -0.41 absolv, test set 4 12d 0.72 0.72 0.70 1.18 1.18 1.18 -0.29 rfr, test set 1 13a 0.90 0.83 0.82 0.66 0.84 0.83 -0.23 rfr, test set 2 13b 0.66 0.66 0.57 0.85 0.85 0.92 -0.41 rfr, test set 3 13c 0.66 0.66 0.64 0.74 0.75 0.76 -0.18 rfr, test set 4 13d 0.82 0.77 0.71 0.95 1.05 1.15 -0.54 gse, test set 1 14 0.91 0.90 0.89 0.62 0.66 0.66 0.02 a gse = general solubility equation; absolv = abraham solvation equation; rfr = random forest regression. b statistics reported in ref. [7]. table 2. recalculated statistics for the scatter plots in ref. [8] type fig. in ref. [8] r 2 pearson eq. (5) r 2 bias eq. (3) a r 2 val eq. (1) rmsepearson eq. (6) rmsebias eq. (4) a rmseval eq. (2) bias eq. (3) gse, small molecules 4a 0.62 0.59 0.57 1.17 1.21 1.23 -0.22 gse, large molecules 4b 0.48 -3.8 -3.82 1.00 3.05 2.95 0.16 gse, modified 4c 0.48 0.34 0.33 1.00 1.13 1.1 0.04 absolv, small molecules 5a 0.67 0.67 0.66 1.08 1.08 1.1 -0.2 absolv, large molecules 5b 0.13 -1.39 -5.24 1.30 2.15 3.36 -2.64 absolv, modified 5c 0.48 -0.91 2.07 1.01 1.92 2.07 0.92 rfr, training set 6a 0.98 0.98 0.98 0.26 0.27 0.27 0.00 rfr, internal validation 6b 0.89 0.89 0.89 0.64 0.64 0.64 0.02 rfr, large molecules 6c 0.45 0.42 0.37 1.03 1.06 1.07 0.30 a statistics reported in ref. [8]. https://doi.org/10.5599/admet.888 alex avdeef admet & dmpk 9(1) (2021) 69-74 74 recalculation of the statistics for our previous studies tables 1 and 2 list three types of ‘r-squared’ and root-mean-square errors for the scatter plots in refs. [7] and [8]. in these two studies, we used the bias-compensated statistics originating from the sigmaplot calculation, but inadvertently ascribed them to eqs. (1) and (2). as can be seen in cases where the bias is negligible, the three sets of statistics are nearly the same (e.g., fig. 8 [7] or fig. 6 [8] rfr results in tables 1, 2). in many of the scatter plots, the differences between the different sets of statistics are very small. conclusion statistics from ready-made programs may be easily verified (e.g., spreadsheet calculation using eqs. (1)-(6)), so that the intended values are reported. the expanded calculations of statistics (tables 1 and 2) applied for our recent prediction studies [7,8] should now allow for valid comparisons between the strength of our predictions of solubility to those reported by others: e.g., in ‘solubility challenges’ sc-2 [5] and sc-1 [6]. acknowledgements we are grateful to dr. mare oja and prof. uko maran (univ. of tartu) for pointing out to us that there was an inconsistency between the formulas for r 2 and rmse in ref. [7] and the values listed in the various scatter plots in that paper. also, their thoughtful comments regarding this manuscript are greatly appreciated. references [1] n. chirico, p. gramatica. real external predictivity of qsar models: how to evaluate it? comparison of different validation criteria and proposal of using the concordance correlation coefficient. j. chem. inf. model. 51 (2011) 2320-2335; doi: https://doi.org/10.1021/ci200211n. [2] w.p. walters. what are our models really telling us? a practical tutorial on avoiding common mistakes when building predictive models. in: j. bajorath (ed.). chemoinformatics for drug discovery. john wiley & sons, hoboken, nj, 2014, pp. 1-31. [3] g. piñeiro, s. perelman, j.p. guerschman, j.m. paruelo. how to evaluate models: observed vs. predicted or predicted vs. observed? ecol. modelling 216 (2008) 316-322. [4] wikipedia: pearson correlation coefficient. https://en.wikipedia.org/wiki/pearson_correlation_coefficient. accessed 16 july 2020. [5] a. llinas, a. avdeef. solubility challenge revisited after ten years, with multi-lab shake-flask data, using tight (sd ∼ 0.17 log) and loose (sd ∼ 0.62 log) test sets. j. chem. inf. model. 59 (2019) 30363040. doi: https://doi.org/10.1021/acs.jcim.9b00345. [6] a.j. hopfinger, e.x. esposito, a. llinàs, r.c. glen, j.m. goodman. findings of the challenge to predict aqueous solubility. j. chem. inf. model. 49 (2009) 1-5. [7] a. avdeef. prediction of aqueous intrinsic solubility of druglike molecules using random forest regression trained with wiki-ps0 database. admet & dmpk 8 (2020) 29-77; doi: https://dx.doi.org/10.5599/admet.766 . [8] a. avdeef, m. kansy. can small drugs predict the intrinsic aqueous solubility of ‘beyond rule of 5’ big drugs? admet & dmpk 8 (2020). doi: https://doi.org/10.5599/admet.794. [9] r open-source package documentation: cor function. https://www.rdocumentation.org/packages/stats/versions/3.6.2/topics/cor. ©2020 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) https://doi.org/10.1021/ci200211n https://en.wikipedia.org/wiki/pearson_correlation_coefficient https://en.wikipedia.org/wiki/pearson_correlation_coefficient https://doi.org/10.1021/acs.jcim.9b00345 https://dx.doi.org/10.5599/admet.766 https://dx.doi.org/10.5599/admet.766 https://doi.org/10.5599/admet.794 https://www.rdocumentation.org/packages/stats/versions/3.6.2/topics/cor https://www.rdocumentation.org/packages/stats/versions/3.6.2/topics/cor http://creativecommons.org/licenses/by/3.0/ solubility of proteins doi: http://dx.doi.org/10.5599/admet.831 391 admet & dmpk 8(4) (2020) 391-399; doi: http://dx.doi.org/10.5599/admet.831 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index feature article solubility of proteins mauno vihinen * protein structure and bioinformatics, department of experimental medical science, bmc b13, se-22184 lund, sweden *corresponding author: e-mail: mauno.vihinen@med.lu.se; tel.: +46-72-5260022 received: april 24, 2020; revised: june 18, 2020; published: june 28, 2020 abstract solubility is a fundamental protein property that has important connotations for therapeutics and use in diagnosis. solubility of many proteins is low and affect heterologous overexpression of proteins, formulation of products and their stability. two processes are related to soluble and solid phase relations. solubility refers to the process where proteins have correctly folded structure, whereas aggregation is related to the formation of fibrils, oligomers or amorphous particles. both processes are related to some diseases. amyloid fibril formation is one of the characteristic features in several neurodegenerative diseases, but it is related to many other diseases, including cancers. severe complex v deficiency and cataract are examples of diseases due to reduced protein solubility. methods and approaches are described for prediction of protein solubility and aggregation, as well as predictions of consequences of amino acid substitutions. finally, protein engineering solutions are discussed. protein solubility can be increased, although such alterations are relatively rare and can lead to trade-off with some other properties. the aggregation prediction methods mainly aim to detect aggregation-prone sequence patches and then making them more soluble. the solubility predictors utilize a wide spectrum of features. ©2020 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords aggregation; protein engineering; biologic; solubility prediction introduction solubility is an important property for all drugs, including biologics. many proteins and polypeptides are poorly soluble and those trespassing through cellular membranes, membrane proteins, are not in traditional sense soluble at all. proteome-wide analysis of solubility in caenorhabditis elegans indicated that about 75 % of proteins appear in cells in abundances close to their solubility limits [1]. we can distinguish two phenomena in relation to protein behaviour in solution. solubility and aggregation are related but have different meanings. solubility is defined here and in many other publications as the concentration in which intact protein is in equilibrium with solid phase [2-4]. in the case of aggregation, protein molecules bind together, often due to irreversibly altered conformation, and form insoluble high molecular weight forms (see figure 1) [5]. unlike aggregated forms, precipitated solubilizable protein in solid phase can be made soluble by dilution. aggregated protein is in solid phase and typically undergoes irreversible structural changes, thus aggregated protein cannot be returned back to soluble form and original structure. proteins in these http://dx.doi.org/10.5599/admet.831 http://dx.doi.org/10.5599/admet.831 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:mauno.vihinen@med.lu.se http://creativecommons.org/licenses/by/4.0/ mauno vihinen admet & dmpk 8(4) (2020) 391-399 392 amyloid fibrils have extensive β-strand secondary structures. when intrinsically disordered proteins aggregate they form amorphous deposits, whereas native-like structures aggregate to native-like deposits (figure 1) [5]. both solubility and aggregation have biological and biotechnological consequences. reduced solubility as well as increased aggregation are related to several diseases. plaques formed by aggregated proteins are common in neurodegenerative diseases. altered conformations of prion proteins can “infect” other proteins and cause their aggregation. β-sheet formation is often related to aggregation and prion formation. these secondary structural elements can stack intraand intermolecularly to form large insoluble aggregates. even protein crystallization is related to solubility. ordered protein crystals are needed for x-ray crystallography to reveal protein structures. these crystals are grown slowly and contain typically large amounts of solvent. the goal of the crystallization is to keep the proteins in their native conformation, however certain packing effects affecting local conformation are common. figure 1. relationships of protein solubility, aggregation and degradation. insoluble protein is correctly folded, whereas the structure of aggregated protein is altered. most irreversibly misfolded proteins are degraded as part of normal protein turnover. intrinsically disordered proteins and native-like structures have their specific aggregation mechanisms. many factors affect solubility and aggregation, including intrinsic properties of the protein, solvent and additives as well as physical conditions. list of relevant protein factors is long and could be started with amino acid sequence and composition, three-dimensional structure and exposure of residues, intramolecular interactions within the protein (salt bridges, hydrogen bonds, electrostatic and hydrophobic interactions, multimeric status etc.), and protonation status. solvent properties (polarity, bond and interaction forming ability, density etc.) and constituents (such as excipients, salts or organic solvents) and their concentrations have significant contributions to solubility. further, factors like ph and temperature affect both the protein and solvent. proteins are large and relatively fragile molecules, thus not necessarily ideal as drugs. lots of research has been devoted to find smaller structures having (some) functions of the biological proteins. miniproteins or protein scaffolds have been tested as biologics [6]. some functions can be retained even in short peptides as highlighted by 2018 nobel prize in chemistry for frances h. arnold, george p. smith and sir gregory p. winter, the last two ones "for the phage display of peptides and antibodies" by developing small binding molecules. in this respect, cyclic peptides are of interest as they have more conformational restraints than linear molecules. admet & dmpk 8(4) (2020) 391-399 solubility of proteins doi: http://dx.doi.org/10.5599/admet.831 393 quite successful solubility prediction methods have been developed for small molecules that are most common as drugs (see chapters in this volume). these methods, however, do not work for proteins. there are several reasons. proteins are much larger, they have lots of groups and protein folding affects the solvent accessibility of the groups. further, the protein structures are flexible and have a large number of slightly different conformations. therefore, different approaches are needed for protein solubility prediction. here, methods for protein solubility and aggregation prediction are introduced. further, methods to investigate the effects on solubility due to amino acid substitutions (variations) are discussed. this topic is important for the design of changes for protein engineering e.g. to increase protein solubility, production etc. protein solubility is discussed here from the perspective of predictions for solubility, aggregation and for variation effects. the principles of the methods are discussed along with performances of the tools. it is apparent that the performances of the methods are not very high, which is due to the complexity of the phenomenon and thereby difficulty in finding good features that would reliably distinguish between the solubility states. one limiting factor has been the small number of experimentally verified cases. however, the field is making progress and there are some rather good methods available. solubility prediction during the last decade, several computational methods have been developed to predict protein solubility, especially in the context of heterologous protein overexpression. these methods utilize different approaches, often in the field of machine learning. solubility is a complex phenomenon and good predictive features are difficult to find, however, there are some trends such as residue-residue interactions [7] and structural flexibility [8]. based on examples with known solubility, computer tools have been trained (see e.g. [9]). protein related parameters such as hydropathy scales and amino acid compositions have been used as features. the best methods claim accuracy of over 80 % for two-state predictions of soluble/insoluble. methods in this category include ccsol omics [10], deepsol [11], parsnip [12], protein-sol [13], soda [14], solart [15], solpro [16], swi [8] and others. table 1. protein solubility predictors method url trained with data from structural biology knowledgebase ccsol omics http://s.tartaglialab.com/update_submission/45568/57e42bea38 deepsol https://zenodo.org/record/1162886#.xoxw_egxvey parsnip https://github.com/redarawi/parsnip swi https://tisigner.com/sodope trained with data from esol proteinsol https://protein-sol.manchester.ac.uk/ solart http://babylone.ulb.ac.be/solart/ solpro http://scratch.proteomics.ics.uci.edu/ trained with data from pon-sol soda http://old.protein.bio.unipd.it/soda/ most methods are trained on one of two major datasets. protein structure initiative was a large project to determine protein structures en masse. their structural biology knowledgebase [17] contains information for crystallization trials and for (heterologous) protein production and has been used for training many of the tools including ccsol omics, parsnip, different versions of deepsol and swi. esol http://dx.doi.org/10.5599/admet.831 http://s.tartaglialab.com/update_submission/45568/57e42bea38 https://zenodo.org/record/1162886#.xoxw_egxvey https://github.com/redarawi/parsnip https://tisigner.com/sodope https://protein-sol.manchester.ac.uk/ http://babylone.ulb.ac.be/solart/ http://scratch.proteomics.ics.uci.edu/ http://old.protein.bio.unipd.it/soda/ mauno vihinen admet & dmpk 8(4) (2020) 391-399 394 solubility database (http://tanpaku.org/tp-esol/index.php?lang=en) for almost all esherichia coli proteins [18] has been used for some other methods, such as protein-sol, solart along with saccharomyces cerevisiae protein solubility details [19]. the remaining methods are trained with pon-sol data [4] or not trained at all. although many proteins are poorly soluble their solubility is biologically sufficient as many proteins have very low abundance in cells [1]. data extracted by deepsol developers from structural biology knowledgebase indicated that 45.2 % out of 129.643 tested inherent and heterologously expressed proteins in e. coli were soluble [11]. deepsol is a deep learning method, ccsol omics is a support vector machine solution, parsnip utilizes gradient boosting, solart random forest, swi and protein-sol are based on weighted scores. details for the algorithm used in soda have not been released. as with any prediction task, the choice of the tool(s) is important. the best comparisons are independent benchmark studies [20, 21], however such studies are not available for any of the prediction tasks discussed in here. instead, there are some comparisons of methods along with predictor description, including comparisons of four [14], seven [12] and eight [11] tools. the first of these studies indicated soda to be the best, accuracy 59.2, the second parsnip (accuracy 74.11 and matthews correlation coefficient (mcc) 0.48) and now defunct proso ii (accuracy 64.35, mcc 0.31) [22], and in the third deepsol, parsnip and proso ii were the best ones with accuracies and mcc values of 0.77/0.55, 0.74/0.48 and 0.64/0.34, respectively. analysis of predictions for 57 udp-dependent glycosyltransferases with 11 predictors [23] is interesting, however, may be biased due to containing proteins just from a single family. the best performing methods in their analysis were soluprot, deepsol versions 3, 1 and 2, solpro and parsnip. unfortunately, the results are provided only in a format of a figure, thus no numbers are available. aggregation prediction aggregation is largely mediated by short sequence stretches of consecutive residues, these regions are typically 15 residues or longer [24]. almost all human proteins can form self-complementary and thus aggregation-prone segments [25]. however, many proteins do not aggregate and during evolution have adapted to prevent amyloid formation [26]. chaperones that assist proteins to fold are in central role [27]. aggregated proteins form either amyloid fibrils, amorphous or native-like deposits. cpad [28], amyload [29] and amypro [30] are databases dedicated for information about protein aggregation. cpad includes data for amyloid peptides, aggregation prone peptides and aggregation rates while amyload contains amylodogenic sequence information. proteins containing experimentally verified amyloidogenic regions are collected to amypro. aggregation predictors can be grouped to two major categories: sequence-based and three-dimensional structure-based. the structure-based methods utilize calculations of free energy difference between solution and aggregation phases, β-structure formation propensity, residue exposure and so on. since protein 3d structures are not always available, sequence-based methods are needed, as well. machine learning methods use features such as amino acid composition and proportions of certain amino acid types for training. aggregation predictors include for example aggrescan [31], aggrescan3d [32], amylpred [33], archcandy [34], foldamyloid [35], metamyl [36], pasta 2.0 [37], tango [38], and waltz [39]. aggrescan3d and archcandy are based on three dimensional structures, some of the others use some structural features, as well. admet & dmpk 8(4) (2020) 391-399 solubility of proteins doi: http://dx.doi.org/10.5599/admet.831 395 table 2. protein aggregation predictors aggrescan http://bioinf.uab.es/aggrescan/ aggrescan3d https://bitbucket.org/lcbio/aggrescan3d/src/master/ amylpred2 http://aias.biol.uoa.gr/amylpred2/ archcandy https://bioinfo.crbm.cnrs.fr/index.php?route=tools&tool=7&lang=uk foldamyloid http://bioinfo.protres.ru/fold-amyloid/ metamyl http://metamyl.genouest.org/e107_plugins/metamyl_aggregation/db_prediction_meta.php pasta 2.0 http://protein.bio.unipd.it/pasta2/help.html rfamyloid http://server.malab.cn/rfamyloid/ tango http://tango.crg.es/ waltz https://waltz.switchlab.org/ amylpred and metamyl are metapredictors, i.e. use predictions from other tools. rfamyloid is machine learning-based method, waltz has a position specific scoring matrix, the others are based on physicochemical propensities and other features including information about secondary structural elements, amino acid composition, structural features, physicochemical propensities of amino acids, packing density, hydrogen bonds etc. neural network tool for amyloid aggregation rate prediction is a related application [40]. some of the methods can be used both for solubility and aggregation prediction, such as solart [15]. soda, a solubility predictor, utilizes aggregation and disorder propensities [14]. systematic independent method performance assessments are missing. some recent studies contain comparisons of several methods. pasta 2.0 was compared to eight methods, showing the best performance (mcc 0.24) along with amylpred 2 (0.22) and metamyl (0.19) [37]. however, the performances are not very high, specificity being clearly better than sensitivity for all the tools, the best matthews correlation coefficient being only 0.24. developers of archcandy saw really big differences in performances for six predictors [41], their own tool (error rate for amyloid prediction 1.4 %) along with pasta (4.2 %), tango (5.0 %) and waltz (11.4 %) being the best. this test contained only soluble segments of 15 residues or longer. they were collected from three-dimensional protein structures determined with nmr, thus the proteins are soluble, however contained highly flexible regions without ordered structures, typically in the termini. comprehensive performance assessment would require both positive and negative cases [20, 21], thereby we do not know if the best methods are overpredicting soluble regions and underpredicting aggregation-prone segments. prediction of solubility and aggregation affecting variants the tools described above are for predictions on entire proteins or polypeptides. much less effort has been put on predicting the effect of variations on solubility or aggregation. we are looking at tools predicting consequences of amino acid substitutions as there is not enough data for other types of variations. single amino acid alterations can have profound effects of solubility and lead to diseases, including severe complex v deficiency [42] and cataract [43]. protein solubility and aggregation are mechanisms in diseases, including cancers [44, 45]. predictors for solubility affecting variants include camsol [46], optsolmut [47], pon-sol [4], soda [14] and solubis [48]. camsol uses residue-specific solubility profile. the method is not available as a tool, only the algorithm has been described. optsolmut has been trained with a small dataset that contains also aggregation cases. weights for scoring function were optimized with linear programming for 137 cases of single and multiple variants. pon-sol is a random forest-based machine learning method. it was trained and tested on 406 single amino acid substitutions for which solubility effect have been experimentally http://dx.doi.org/10.5599/admet.831 http://bioinf.uab.es/aggrescan/ https://bitbucket.org/lcbio/aggrescan3d/src/master/ http://aias.biol.uoa.gr/amylpred2/ https://bioinfo.crbm.cnrs.fr/index.php?route=tools&tool=7&lang=uk http://bioinfo.protres.ru/fold-amyloid/ http://metamyl.genouest.org/e107_plugins/metamyl_aggregation/db_prediction_meta.php http://protein.bio.unipd.it/pasta2/help.html http://server.malab.cn/rfamyloid/ http://tango.crg.es/ https://waltz.switchlab.org/ mauno vihinen admet & dmpk 8(4) (2020) 391-399 396 determined. it predicts variants into three classes: solubility decreasing and increasing variants and those not affecting solubility. this is a more realistic scenario than binary prediction. in the three-state prediction pon-sol had correct prediction ratio of 0.597 on cross validation and 0.488 for independent test set (note that random prediction has a score of 0.33). thus, there is still place for substantial improvements. the method development has been hampered by small number of known solubility-affecting variants. pon-sol is no more available; however, a new extended and improved version will be released soon. soda has been recommended to predict variants decreasing solubility [14]. it was developed with pon-sol data. these methods can be used for numerous purposes including identification of disease related amino acid substitutions, predictions of solubility of heterologous recombinant protein expression and enhanced crystallizability. of the aggregation prediction tools, pasta 2.0 can predict also effects of amino acid substitutions. solubis has a somewhat different application, optimization of multiple variants to increase protein solubility [48]. it detects aggregation prone segments and then suggests variants to modify them. so, actually it is an aggregation prevention predictor. it combines predictions from interaction analysis tool foldx [49], aggregation predictor tango [38] and structural analysis with yasara [50]. predictions for protein engineering protein properties have complex relations. several approaches have been tried to improve solubility or prevent aggregation. electrostatic interactions have been a starting point for one approach [51], stability and aggregation for another [52], and surface patches to avoid aggregation for a third one [53]. solubis tries to reduce aggregation propensity [48]. structural changes designed by four tools have been reviewed in relation to structure-based predictions [54]. the discussed tools included aggrescan3d, camsol, spatial aggregation propensity (sap) and solubis. sap was a proposal to apply molecular dynamics simulations, not an implemented tool [55]. systematic performance assessments have not been made to these methods. massively parallel reporter assay (mpra) of two proteins, tem-1 β-lactamase, a common antibiotic resistance protein in gram positive bacteria, and e. coli levoglucosan kinase, indicated trade-offs between fitness and solubility [56]. solubility in this paper was defined as properly folded protein. they used two analysis methods, which revealed whether the protein was folded. in yeast surface display screen the investigated protein had to be folded otherwise the fusion protein was degraded. in twin-arginine translocation-selective export the protein was exported to bacterial periplasm only if correctly folded. they generated >93 % of all possible single amino acid variants in the two proteins. solubility increasing variants were rare, only 4 to 5 % had this effect. many solubility increasing variants affected also some other property. comparisons to fitness increasing variants revealed that these two features co-occurred very rarely and there were trade-offs between them. conclusions computational methods available for the prediction of protein solubility and aggregation were discussed along with tools for engineering solubility or aggregation by introducing amino acid substitutions. many features and different algorithms have been applied to the available solutions. although systematic performance assessments have not been performed, it is evident that the methods have widely varying performances. solubility and avoidance of aggregation are crucial properties for any protein to be used for diagnosis or therapy. admet & dmpk 8(4) (2020) 391-399 solubility of proteins doi: http://dx.doi.org/10.5599/admet.831 397 acknowledgements: financial support from vetenskapsrådet, swedish cancer society and alfred österlunds stiftelse is gratefully acknowledged. conflict of interest: none. references [1] g. vecchi, p. sormanni, b. mannini, a. vandelli, g.g. tartaglia, c.m. dobson, f.u. hartl, m. vendruscolo. proteome-wide observation of the phenomenon of life on the edge of solubility. proc. natl. acad. sci. u. s. a. 117 (2020) 1015-1020. [2] t. arakawa, s.n. timasheff. theory of protein solubility. methods enzymol. 114 (1985) 49-77. [3] p. garidel. protein solubility from biochemical, physicochmical and colloidal perspective. am. pharm. rev. december 13 (2013). [4] y. yang, a. niroula, b. shen, m. vihinen. pon-sol: prediction of effects of amino acid substitutions on protein solubility. bioinformatics 32 (2016) 2032-2034. [5] f. chiti, c.m. dobson. protein misfolding, amyloid formation, and human disease: a summary of progress over the last decade. annu. rev. biochem. 86 (2017) 27-68. [6] z.r. crook, n.w. nairn, j.m. olson. miniproteins as a powerful modality in drug development. trends biochem. sci. 45 (2020) 332-346. [7] q. hou, r. bourgeas, f. pucci, m. rooman. computational analysis of the amino acid interactions that promote or decrease protein solubility. sci. rep. 8 (2018) 14661. [8] b.k. bhandari, p.p. gardner, c.s. lim. solubility-weighted index: fast and accurate prediction of protein solubility. (2020). https://doi.org/10.1093/bioinformatics/btaa578. [9] n. habibi, s.z. mohd hashim, a. norouzi, m.r. samian. a review of machine learning methods to predict the solubility of overexpressed recombinant proteins in escherichia coli. bmc bioinformatics 15 (2014) 134. [10] f. agostini, d. cirillo, c.m. livi, r. delli ponti, g.g. tartaglia. ccsol omics: a webserver for solubility prediction of endogenous and heterologous expression in escherichia coli. bioinformatics 30 (2014) 2975-2977. [11] s. khurana, r. rawi, k. kunji, g.y. chuang, h. bensmail, r. mall. deepsol: a deep learning framework for sequence-based protein solubility prediction. bioinformatics 34 (2018) 2605-2613. [12] r. rawi, r. mall, k. kunji, c.h. shen, p.d. kwong, g.y. chuang. parsnip: sequence-based protein solubility prediction using gradient boosting machine. bioinformatics 34 (2018) 1092-1098. [13] m. hebditch, m.a. carballo-amador, s. charonis, r. curtis, j. warwicker. protein-sol: a web tool for predicting protein solubility from sequence. bioinformatics 33 (2017) 3098-3100. [14] l. paladin, d. piovesan, s.c.e. tosatto. soda: prediction of protein solubility from disorder and aggregation propensity. nucleic acids res. 45 (2017) w236-w240. [15] q. hou, j.m. kwasigroch, m. rooman, f. pucci. solart: a structure-based method to predict protein solubility and aggregation. bioinformatics 36 (2020) 1445-1452. [16] c.n. magnan, a. randall, p. baldi. solpro: accurate sequence-based prediction of protein solubility. bioinformatics 25 (2009) 2200-2207. [17] m.j. gabanyi, p.d. adams, k. arnold, l. bordoli, l.g. carter, j. flippen-andersen, l. gifford, j. haas, a. kouranov, w.a. mclaughlin, d.i. micallef, w. minor, r. shah, t. schwede, y.p. tao, j.d. westbrook, m. zimmerman, h.m. berman. the structural biology knowledgebase: a portal to protein structures, sequences, functions, and methods. j. struct. funct. genomics 12 (2011) 45-54. [18] t. niwa, b.w. ying, k. saito, w. jin, s. takada, t. ueda, h. taguchi. bimodal protein solubility distribution revealed by an aggregation analysis of the entire ensemble of escherichia coli proteins. proc. natl. acad. sci. u. s. a. 106 (2009) 4201-4206. http://dx.doi.org/10.5599/admet.831 https://doi.org/10.1093/bioinformatics/btaa578 mauno vihinen admet & dmpk 8(4) (2020) 391-399 398 [19] e. uemura, t. niwa, s. minami, k. takemoto, s. fukuchi, k. machida, h. imataka, t. ueda, m. ota, h. taguchi. large-scale aggregation analysis of eukaryotic proteins reveals an involvement of intrinsically disordered regions in protein folding. sci. rep. 8 (2018) 678. [20] m. vihinen. how to evaluate performance of prediction methods? measures and their interpretation in variation effect analysis. bmc genomics 13 suppl 4 (2012) s2. [21] m. vihinen. guidelines for reporting and using prediction tools for genetic variation analysis. hum. mutat. 34 (2013) 275-282. [22] p. smialowski, g. doose, p. torkler, s. kaufmann, d. frishman. proso ii--a new method for protein solubility prediction. febs j. 279 (2012) 2192-2200. [23] f.a. ghomi, t. kittilä, d.h. welner. a benchmark of protein solubility prediction methods on udpdependent glycosyltransferases. (2020). doi: https://doi.org/10.1101/2020.02.28.962894. [24] a.b. ahmed, a.v. kajava. breaking the amyloidogenicity code: methods to predict amyloids from amino acid sequence. febs lett. 587 (2013) 1089-1095. [25] l. goldschmidt, p.k. teng, r. riek, d. eisenberg. identifying the amylome, proteins capable of forming amyloid-like fibrils. proc. natl. acad. sci. u. s. a. 107 (2010) 3487-3492. [26] e. monsellier, f. chiti. prevention of amyloid-like aggregation as a driving force of protein evolution. embo rep. 8 (2007) 737-742. [27] n.b. nillegoda, a.s. wentink, b. bukau. protein disaggregation in multicellular organisms. trends biochem. sci. 43 (2018) 285-300. [28] p. rawat, r. prabakaran, r. sakthivel, a. mary thangakani, s. kumar, m.m. gromiha. cpad 2.0: a repository of curated experimental data on aggregating proteins and peptides. amyloid (2020) 1-6. [29] p.p. wozniak, m. kotulska. amyload: website dedicated to amyloidogenic protein fragments. bioinformatics 31 (2015) 3395-3397. [30] m. varadi, g. de baets, w.f. vranken, p. tompa, r. pancsa. amypro: a database of proteins with validated amyloidogenic regions. nucleic acids res. 46 (2018) d387-d392. [31] o. conchillo-sole, n.s. de groot, f.x. aviles, j. vendrell, x. daura, s. ventura. aggrescan: a server for the prediction and evaluation of "hot spots" of aggregation in polypeptides. bmc bioinformatics 8 (2007) 65. [32] r. zambrano, m. jamroz, a. szczasiuk, j. pujols, s. kmiecik, s. ventura. aggrescan3d (a3d): server for prediction of aggregation properties of protein structures. nucleic acids res. 43 (2015) w306w313. [33] k.k. frousios, v.a. iconomidou, c.m. karletidi, s.j. hamodrakas. amyloidogenic determinants are usually not buried. bmc struct biol. (2009) jul 9; 9:44. doi: https://doi.org/10.1186/1472-6807-9-44. [34] a.b. ahmed, n. znassi, m.t. chateau, a.v. kajava. a structure-based approach to predict predisposition to amyloidosis. alzheimers dement. 11 (2015) 681-690. [35] s.o. garbuzynskiy, m.y. lobanov, o.v. galzitskaya. foldamyloid: a method of prediction of amyloidogenic regions from protein sequence. bioinformatics. 26(3) (2010) 326-332. doi: https://doi.org/10.1093/bioinformatics/btp691. [36] m. emily, a. talvas, c. delamarche. metamyl: a meta-predictor for amyloid proteins. plos one. 8(11) (2013) e79722. doi: https://doi.org/10.1371/journal.pone.0079722. [37] i. walsh, f. seno, s.c. tosatto, a. trovato. pasta 2.0: an improved server for protein aggregation prediction. nucleic acids res 42 (2014) w301-w307. [38] a.m. fernandez-escamilla, f. rousseau, j. schymkowitz, l. serrano. prediction of sequencedependent and mutational effects on the aggregation of peptides and proteins. nat. biotechnol. 22 (2004) 1302-1306. [39] k.l. morris, a. rodger, m.r. hicks, m. debulpaep, j. schymkowitz, f. rousseau, l.c. serpell. exploring the sequence-structure relationship for amyloid peptides. biochem j. 450(2) (2013) 275-283. doi: https://doi.org/10.1042/bj20121773. https://doi.org/10.1101/2020.02.28.962894 https://doi.org/10.1186/1472-6807-9-44 https://doi.org/10.1093/bioinformatics/btp691 https://doi.org/10.1371/journal.pone.0079722 admet & dmpk 8(4) (2020) 391-399 solubility of proteins doi: http://dx.doi.org/10.5599/admet.831 399 [40] w. yang, p. tan, x. fu, l. hong. prediction of amyloid aggregation rates by machine learning and feature selection. j. chem. phys. 151 (2019) 084106. [41] d.b. roche, e. villain, a.v. kajava. usage of a dataset of nmr resolved protein structures to test aggregation versus solubility prediction algorithms. protein sci. 26 (2017) 1864-1869. [42] a. meulemans, s. seneca, t. pribyl, j. smet, v. alderweirldt, a. waeytens, w. lissens, r. van coster, l. de meirleir, j.p. di rago, d.l. gatti, s.h. ackerman. defining the pathogenesis of the human atp12p w94r mutation using a saccharomyces cerevisiae yeast model. j. biol. chem. 285 (2010) 4099-4109. [43] u.p. andley, m.a. reilly. in vivo lens deficiency of the r49c alphaa-crystallin mutant. exp. eye res. 90 (2010) 699-702. [44] g.a.p. de oliveira, y. cordeiro, j.l. silva, t. vieira. liquid-liquid phase transitions and amyloid aggregation in proteins related to cancer and neurodegenerative diseases. adv. protein chem. struct. biol. 118 (2019) 289-331. [45] m. kanapathipillai. treating p53 mutant aggregation-associated cancer. cancers (basel) 10 (2018) 154. [46] p. sormanni, f.a. aprile, m. vendruscolo. the camsol method of rational design of protein mutants with enhanced solubility. j. mol. biol. 427 (2014) 478-490. [47] y. tian, c. deutsch, b. krishnamoorthy. scoring function to predict solubility mutagenesis. algorithms mol. biol. 5 (2010) 33. [48] j. van durme, g. de baets, r. van der kant, m. ramakers, a. ganesan, h. wilkinson, r. gallardo, f. rousseau, j. schymkowitz. solubis: a webserver to reduce protein aggregation through mutation. protein eng. des. sel. 29 (2016) 285-289. [49] r. guerois, j.e. nielsen, l. serrano. predicting changes in the stability of proteins and protein complexes: a study of more than 1000 mutations. j. mol. biol. 320 (2002) 369-387. [50] h. land, m.s. humble. yasara: a tool to obtain structural guidance in biocatalytic investigations. methods mol. biol. 1685 (2018) 43-67. [51] c.j. o'brien, m.a. blanco, j.a. costanzo, m. enterline, e.j. fernandez, a.s. robinson, c.j. roberts. modulating non-native aggregation and electrostatic protein-protein interactions with computationally designed single-point mutations. protein eng. des. sel. 29 (2016) 231-243. [52] m. gil-garcia, m. bano-polo, n. varejao, m. jamroz, a. kuriata, m. diaz-caballero, j. lascorz, b. morel, s. navarro, d. reverter, s. kmiecik, s. ventura. combining structural aggregation propensity and stability predictions to redesign protein solubility. mol. pharm. 15 (2018) 3846-3859. [53] m.a. carballo-amador, e.a. mckenzie, a.j. dickson, j. warwicker. surface patches on recombinant erythropoietin predict protein solubility: engineering proteins to minimise aggregation. bmc biotechnol. 19 (2019) 26. [54] s. navarro, s. ventura. computational re-design of protein structures to improve solubility. expert opin. drug discov. 14 (2019) 1077-1088. [55] n. chennamsetty, v. voynov, v. kayser, b. helk, b.l. trout. design of therapeutic proteins with enhanced stability. proc. natl. acad. sci. u. s. a. 106 (2009) 11937-11942. [56] j.r. klesmith, j.p. bacik, e.e. wrenbeck, r. michalczyk, t.a. whitehead. trade-offs between enzyme fitness and solubility illuminated by deep mutational scanning. proc. natl. acad. sci. u. s. a. 114 (2017) 2265-2270. ©2020 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://dx.doi.org/10.5599/admet.831 http://creativecommons.org/licenses/by/3.0/ investigation of possible solubility and dissolution advantages of cocrystals, i: aqueous solubility and dissolution rates of ketoconazole and its cocrystals as functions of ph doi: 10.5599/admet.661 106 admet & dmpk 7(2) (2019) 106-130; doi: http://dx.doi.org/10.5599/admet.661 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper investigation of possible solubility and dissolution advantages of cocrystals, i: aqueous solubility and dissolution rates of ketoconazole and its cocrystals as functions of ph jaydip m. vasoya, 1 ankita v. shah, 1,2 abu t.m. serajuddin 1, * 1 department of pharmaceutical sciences, college of pharmacy and health sciences, st. john’s university, 8000 utopia parkway, queens, ny 11439, usa (jaydip m. vasoya: jaydip.vasoya15@stjohns.edu) 2 current address: freund-vector corporation, 675 44 th street, marion, ia 52302, usa (ankita v shah: ankita.shah@freund-vector.com) * corresponding author: dr. abu t.m. serajuddin, email: serajuda@stjohns.edu, office phone: 718-990-7822 received: february 04, 2019; revised: april 01, 2019; published: april 05, 2019 abstract since there are conflicting reports in the literature on solubility and dissolution advantages of cocrystals over free forms, we systematically studied solubility and intrinsic dissolution rates of a weakly basic drug, ketoconazole, and its cocrystals with fumaric acid and succinic acid as functions of ph to determine what advantages cocrystals provide. ph-solubility profiles were determined in two different ways: one by lowering ph of ketoconazole aqueous suspensions using hcl, fumaric acid and succinic acid, and the other by adjusting ph of cocrystal suspensions using respective coformer acids or naoh. similar ph-solubility profiles were obtained whether free base or cocrystals were used as starting materials to determine solubility. with the addition of fumaric and succinic acids to aqueous suspensions of free base to lower ph, the maximum solubility (phmax) was reached at ph ~3.5-4.0, below which the solubility decreased and cocrystals formed. the solubility, however, continued increasing when hcl was added to ketoconazole suspension as no cocrystal or salt was formed. during determination of cocrystal solubility, a conversion to free base was observed when ph was raised above phmax. thus, ph-solubility profiles of cocrystals resembled solubility profiles commonly encountered with salts. above phmax, both free base and cocrystal had similar solubility under identical ph conditions; the solubility of cocrystal was higher only if the ph differed. in contrast, intrinsic dissolution rates of cocrystals at ph>phmax under identical bulk ph were much higher than that of free ketoconazole since cocrystals had lower microenvironmental ph at the dissolving surface, where the solubility was high. thus, cocrystals of basic drugs can potentially provide higher dissolution rates under intestinal ph conditions. keywords ketoconazole; cocrystal; solubility; ph effect; intrinsic dissolution rate; solid phase; phase conversion introduction crystalline forms of acidic and basic drugs or their salts have traditionally been used as active pharmaceutical ingredients (api) [1]. during the past two decades, cocrystal, which is another class of crystalline drug substance, has gained much interest in the pharmaceutical field for potential development as api [2-5]. the us food and drug administration (fda) defined cocrystals as “crystalline materials composed of two or more different molecules, typically active pharmaceutical ingredient (api) and cohttp://www.pub.iapchem.org/ojs/index.php/admet/index mailto:jaydip.vasoya15@stjohns.edu mailto:serajuda@stjohns.edu admet & dmpk 7(2) (2019) 106-130 solubility and dissolution of cocrystals doi: 10.5599/admet.661 107 crystal formers (‘coformers’), in the same crystal lattice” [6]. to exclude solvate, hydrate, salt, etc., from such a definition, an international panel of experts defined cocrystals as “solids that are crystalline singlephase materials composed of two or more different molecular and ionic compounds generally in a stoichiometric ratio, which are neither solvates nor simple salts” [7]. the cocrystals differ from salts as no acid-base proton transfers are involved in cocrystal formation. they are usually formed by hydrogen bonding or other electrostatic interactions like π-π interaction, halogen bonding, etc., between two species, and, thus, cocrystals may form even between neutral molecules [8]. it is believed that cocrystals dissociate into individual molecular species when they dissolve in aqueous media. according to kale et al. [9], the development of cocrystals into viable drug products could be challenging because of such factors as ‘lower than expected supersaturation solubility, safety of coformer, difficulties in the manufacturing of high-dose drugs, polymorphism, atypical behavior of cocrystal in the formulation, difficulty in ivivc, and so forth. despite such challenges, there is active research on the development of cocrystals as potential drug products, and to the best of our knowledge, three of them have already received regulatory approvals. the first is ipragliflozin l-proline, a 1:1 cocrystal of the selective sglt2 inhibitor ipragliflozin with l-proline, which received approval in japan in 2014 for the treatment of type 2 diabetes (suglat® tablet; astellas) [10,11]; to the best of our knowledge, it is not yet available in the usa or european countries. the second is sacubitril/valsartan, which is a 1:1 cocrystal between sacubitril sodium, a neprilysin (nep) inhibitor, and valsartan sodium, an angiotensin receptor 1 (at1) blocker (entresto® tablet; novartis); it was approved by the fda and the european medicines agency (ema) in 2015 for the treatment of long-term heart failure in adult population. the third is etrugliflozin l-pga, which is a 1:1 cocrystal between etrugliflozin, another sglt2 inhibitor, and l-pyroglutamic acid; it received fda and ema approvals in, respectively, 2017 and 2018 for the treatment of type ii diabetes (steglatro® tablets; merck/pfizer). there is not much information in the published literature on physicochemical properties of these cocrystals and what advantages they provide over those of their parent molecules. there is the potential that a cocrystal may have higher solubility than that of the parent chemical entity. however, the solubility improvement was possibly not the primary reasons why the three marketed cocrystals were developed. in case of ipragliflozin l-proline (suglat®), the chemical structure of ipragliflozin indicates that, by itself, it could be water-soluble since it is a glycoside with a -f and several –oh groups [10], and, therefore, the cocrystal was possibly developed for certain other advantages than the solubility improvement. in case of sacubitril/valsartan (entresto®), both components were active moieties, and it was not a typical cocrystal between drug and therapeutically inactive coformer. also, since both components of the cocrystal were water-soluble sodium salts, the primary advantage of cocrystal formation in this case appears to be the improvement of physicochemical properties of materials by crystallizing two otherwise amorphous or less crystalline api together as one highly crystalline drug substance [12]. etrugliflozin is a biopharmaceutics classification i (bcs i) drug having good solubility and good permeability throughout the gastrointestinal ph range [13] and, therefore, its conversion to etrugliflozin l-pga cocrystal (steglatro®) could also be for reasons other than the solubility enhancement. although, as mentioned above, there may be different reasons behind the development of cocrystals, much of the recent interest in this area appears to be the potential advantage of using cocrystals to improve solubility and dissolution rate of poorly water-soluble drugs [14,15]. however, there are mixed reports in the literature on the actual increase in aqueous solubility that may be attained by cocrystal formation. kozak et al. [16] reported that the solubility of a neutral compound ethenzamide in water increased from around 0.80-0.85 µg/ml to around 1.20-1.30 µg/ml by cocrystallization with glutaric acid, malonic acid and maleic acid. although this could be 1.4-1.6 time increase in solubility, considering the vasoya et al. admet & dmpk 7(2) (2019) 106-130 108 extremely low aqueous solubility of ethenzamide and the cocrystals, the increase may not be of biopharmaceutical significance. indeed, the cocrystallization did not increase dissolution rates of ethenzamide, and, rather, cocrystals exhibited lower dissolution rates in water than the parent compound. although possibly not suitable for neutral compounds, it could be possible that the cocrystal formation may be especially advantageous to basic and acidic compounds in cases where they do not normally form salts. stanton and bak [17] determined comparative solubility of ten cocrystals of an experimental basic drug, amg 517 (pka 0.68), with ten acids (pka 3.07 to 4.20) in simulated fasted intestinal fluid (fasif; 5 mm sodium taurocholate and 1.5 mm lecithin in ph 6.8 phosphate buffer) and observed that, as compared to the free base solubility of 5 µg/ml, the solubility of the cocrystals ranged from 1 to 21 µg/ml. thus, the solubility of the compound either decreased or only marginally increased. in another study [18], the solubility of a basic drug, lamotrigine, in water was 0.154 mg/ml, while the solubility of its cocrystals with glutaric acid, sorbic acid, propionic acid and acetic acid increased to 1.10, 1.12, 2.17 and 2.55 mg/ml, respectively, and, thus, there were 7 to 17 times increase in aqueous solubility. however, at the same time, ph of the aqueous slurries during the determination of solubility decreased from 6.8 for the free base to, respectively, 5.2, 5.1, 5.0 and 4.9 for the four cocrystals. since lamotrigine has a pka value of 5.7, the increase in its aqueous solubility by the formation of cocrystals may be attributed partly or fully to the ph change and the resultant protonation of the compound rather than cocrystal formation. similarly, gao et al. [19] observed a four-fold increase in solubility of adefovir dipivoxil, a basic drug, in water from 0.79 mmol/l to 3.33 mmol/l due to cocrystal formation with saccharin that may also be attributed to the decrease in ph of saturated solutions from 6.7 to 3.1. cheney et al. [20] found that when there was no difference in ph of saturated solutions, there was also no significant difference among solubilities of free base and cocrystals forms of drug. in this case, solubilities of lamotrigine-methyl paraben, lamotriginenicotinamide and lamotrigine-nicotinamide (monohydrate) cocrystals at 1:1 ratios and that of the lamotrigine free base in water, all having ph of ~5.5, were, respectively, 0.21, 0.30, 0.23 and 0.28 mg/ml, indicating no significant difference in solubility due to cocrystal formation. in another study, ràfols et al. [21] compared dissolution rates of a zwitterionic compound, ciprofloxacin, and its cocrystal with resorcinol under different ph conditions and observed that the parent compound dissolved faster at ph 2, while the cocrystal had somewhat higher dissolution rates in the intermediate ph conditions of 4, 5 and 5.5. however, at the intestinal ph condition of 6.5, both ciprofloxacin and the cocrystal had much lower and essentially similar solubility and dissolution rates irrespective of whether the studies were conducted in buffer or fasted state simulated intestinal fluid (fassif), indicating that there were no solubility and dissolution advantages of cocrystal formation under intestinal ph conditions. the above review of published studies does not show any major difference in equilibrium solubility between cocrystal and neutral chemical entity or between cocrystal and free base in aqueous media, and where a difference was observed, it could possibly be explained by the change in ph of solutions. there are, however, several reports where differences between kinetic solubility of cocrystals and free forms of drug were observed when dissolution rates of powders were studied. it was observed that there were initial increases in drug concentrations, which decreased with time as the free base forms precipitated out [17,22,23]. this is analogous to the supersaturation of amorphous drugs prior to their conversion to crystalline forms [24]; when the solubility reaches equilibrium, any difference between solubility of cocrystal and free drug form diminishes or disappears. in contrast to the above-mentioned experimental findings, several investigators have demonstrated through a combination of experimental studies and theoretical calculations that cocrystals can indeed have major solubility advantages over their respective free forms, sometimes by several orders in magnitude admet & dmpk 7(2) (2019) 106-130 solubility and dissolution of cocrystals doi: 10.5599/admet.661 109 [22,23,25,26]. the higher solubility of cocrystals as the function of ph in these studies were obtained by applying a new 3-step process: first, drug and coformer concentrations in equilibria with cocrystals were measured at ph values of interest; second, cocrystal solubility and solubility product (ksp) were determined from the measured equilibrium concentrations; and, third, the ph dependence of cocrystal solubility is theoretically calculated from ksp of cocrystal, pka values of drug and coformer, and corresponding solubility equations [22]. based on such theoretical calculations, chen et al. [22] observed much higher solubility for 1:1 cocrystals of ketoconazole, a basic drug (pka 2.94 and 6.51) [27], with fumaric acid, succinic acid and adipic acid in the ph range of 4 to 7 as compared to that of the free base in the same ph range. martin et al. [28] also observed 75 to 100 times higher solubility of ketoconazole cocrystals with fumaric acid, succinic acid and adipic acid in water than that of ketoconazole free base. however, in the latter case, the ph values of the slurries of cocrystals in water decreased from the initial 5.8 to the range of 3.4 to 4.1, where the solubility of the drug could be much higher and thus be responsible for the observed difference between cocrystal and drug solubility. the inconsistencies between above-mentioned experimental and theoretical findings indicate that much additional work is necessary to elucidate what advantages cocrystals may have over their free forms with respect to solubility and dissolution rate. therefore, the present investigation has been undertaken to systematically determine solubility and dissolution rates of a basic drug, ketoconazole, and two of its cocrystals with fumaric acid and succinic acid as functions of ph. although not many studies on ph versus solubility of cocrystals have been reported in the literature, there are numerous reports on the effect of ph on solubility of acids, bases and their salt forms available [29-32]. there are also numerous reports on higher dissolution rates of salts as compared to their respective free base or acid forms due to change in ph and increased drug solubility in diffusion layers of solids during dissolution testing [29,32,33]. it was, therefore, of interest to determine whether cocrystals could influence drug solubility and dissolution rate analogous to salts and, if any differences in performance of salts and cocrystals existed, what were those differences. based on these considerations, primary objectives of the present study with ketoconazole and its cocrystals were the following: (a) determine equilibrium solubility of free base and cocrystals as functions of ph to ascertain what differences exist between the solubility of two forms, (b) determine solid phases in equilibria with solutions to study whether any conversion from cocrystal to free base or vice versa occur during solubility vs. ph determination, (c) determine possible similarity and dissimilarity between ph vs. solubility of cocrystal with ph-solubility profiles of salts reported in the literature and, in particular, whether any phmax (ph of maximum solubility) exists during the interconversion of cocrystal and free base like that with salts, and (d) determine comparative intrinsic dissolution rates of cocrystals and the free base at different ph conditions to ascertain what role cocrystal formation plays in dissolution and whether there are any impacts of microenvironmental ph at the dissolving solid surface on the dissolution rate. experimental materials ketoconazole was purchased from alfa aesar, a thermo fisher scientific company (ward hill, ma, usa) and used as received. succinic acid and fumaric acid were purchased, respectively, from vwr life sciences (radnor, pa, usa) and fluka analytical (fisher scientific, pittsburg, usa). all other solvents and reagents used in this investigation were of analytical grade or better. distilled water was used, as necessary, for all experiments. vasoya et al. admet & dmpk 7(2) (2019) 106-130 110 preparation of ketoconazole cocrystals cocrystals were prepared by modifying solvent evaporation methods reported earlier [28]. in a typical experiment for the preparation of ketoconazole-succinic acid cocrystal, 2 millimoles of ketoconazole (1063 mg) and 2.2 millimoles of succinic acid (260 mg) were dissolved in 4 ml of 3:1 v/v chloroform-methanol solvent mixture by stirring on a hot plate magnetic stirrer at 60-65 °c. the solution was then kept under vacuum for solvent evaporation and thereby cocrystal precipitation. after complete evaporation of solvent, the dried material was washed with 5 ml of methanol to remove any unconverted free drug and coformer. the washed cocrystals were dried in a vacuum oven at 50 °c to remove any residual methanol. dried cocrystals were analyzed using differential scanning calorimetry (dsc), powder x-ray diffraction (pxrd) and raman spectroscopy. the purity of cocrystals or the presence of any unconverted ketoconazole or succinic acid with cocrystals were assessed by dsc, pxrd and raman spectroscopy patterns. a similar method was also used for the preparation of ketoconazole-fumaric acid cocrystal where 2 millimoles of ketoconazole (1063 mg) and 2.2 millimoles of fumaric acid (255.4 mg) were dissolved in 5 ml of 3:2 v/v chloroformmethanol mixture. multiple batches of the two cocrystals were prepared for use in the course of the present investigation and the quality of materials was found to be consistently similar. differential scanning calorimetry (dsc) the dsc scans were recorded using the q200 differential scanning calorimeter (ta instruments, wilmington, de, usa). accurately weighed samples (4-6 mg each) were taken into tzero aluminum pans and sealed with hermetic lids with pin holes. samples were heated at the rate of 10 °c/min from 5 °c to 350 °c under nitrogen flush at the rate of 50 ml/min. data acquisition was performed using ta analysis software. powder x-ray diffraction (pxrd) powder x-ray diffraction analysis was conducted using shimadzu xrd-6000 diffractometer (shimadzu, kyoto, japan), equipped with ni filter and monochromatic cu-ka radiation source. the diffractometer was operated with a copper anode tube at the generator voltage and the current of 40 kv and 30 ma, respectively. x-ray patterns of all solid samples were recorded between 5° to 50° two-theta angles at the rate of 2° per minute. all pxrd patterns were analyzed for characteristic peaks of ketoconazole and its cocrystals. raman spectroscopy solid samples as well as the excess solids present in suspensions during the determination of aqueous solubility were analyzed by raman spectroscopy. raman spectra were generated with the rxn1 system (kaiser optics, ann arbor, mi, usa) using fiber optic probe emitting 785 nm laser with the total power of 400 mw. data acquisition was performed using ic raman 4.1 software from mettler toledo (columbus, oh, usa). for in situ analysis of solids suspended in aqueous media, aliquots of suspensions were transferred into small aluminum cups and then exposed to laser beam through non-contact raman phat probe (kaiser optics) in dark. after acquisition of the spectrum, the sample was poured back into the vial and the shaking was continued. hplc analysis drug concentration was analyzed by reverse phase high pressure liquid chromatography (hplc) (hp 1100 series, agilent technologies, wilmington, de). the mobile phase consisted of 3:1 ratio of acetonitrile: acetate buffer, where, for the preparation of acetate buffer, 0.04 m sodium acetate trihydrate solution was prepared in distilled water, ph was adjusted to 4.5 using glacial acetic acid, and to 1000 ml of this solution, admet & dmpk 7(2) (2019) 106-130 solubility and dissolution of cocrystals doi: 10.5599/admet.661 111 2 ml of triethyl amine was added. for hplc analysis, c18 column (agilent zorbax eclipse xdb columns, agilent technologies, santa clara, ca) was used with 0.75 ml/min flow rate, 230 nm wavelength, and 25°c temperature. determination of ph vs solubility profiles in a ‘white paper’ published by an international panel of experts, avdeef et al. [34] reported that a systematic determination of solubility by taking into consideration such factors as equilibration time, stirring rate, equilibrium ph, temperature, compositions of aqueous media (buffers, ionic strength, etc.), characteristics of solid phase, and so forth, is critically important for accurate measurement of solubility of drug substances and especially when the drug substances are ionizable. the composition of solid phases in equilibria with aqueous media may also change depending on the buffering agents used [35]. for these reasons, we have undertaken systematic determination of aqueous solubility of ketoconazole, an ionizable drug, under different experimental conditions, as described below, by taking into considerations the recommendations made in these publications. ketoconazole solubility by ph adjustment with hcl the ph-solubility profile of ketoconazole was generated by adjusting ph of its suspension in water using different concentrations of hcl solutions (0.001 to 1 m), as necessary, to obtain the desired ph. in a typical experiment, 5 ml of distilled water was taken in a 20-ml glass vial with a small amount of ketoconazole (5-10 mg), and the vial was shaken in a water bath shaker (innova 3100, new brunswick scientific co., edison, nj) at 25 °c. multiple vials were prepared using different concentration of hcl solutions (0.001m to 1m) so that different ph values could be attained, and, when needed, more amount of free base was added to maintain excess solid in the vials. ketoconazole solubility was measured at enough number of ph points by repeating above mentioned procedure as exact replicates of ph point is experimentally difficult. the ph of the suspension was measured at 4 and 24 h and concentrations of ketoconazole at these time points were determined by filtering 0.5 to 1 ml of aliquots through poly(vinylidene fluoride) (pvdf) filter with 0.45 micron pore size and analyzing filtrate by hplc. additional hcl solution and excess ketoconazole were added to the suspension in vial and shaken for another ph point, and the process was repeated until enough points on the ph scale were obtained. more than one suspension in different vials were used to generate a complete profile. cocrystal solubility by adjusting ph with coformer acids and naoh cocrystal solubility at different ph values was determined using solutions of coformer acid (succinic acid or fumaric acid) to lower ph of aqueous suspensions and by adding naoh solutions to adjust higher ph. excess amount of cocrystal was dispersed in 5 ml of distilled water in a 20 ml glass vial, and when the cocrystal suspension reached a specific ph after equilibration in a water bath at 25 °c by shaking, coformer solutions (4 mg/ml fumaric acid or 50 mg/ml succinic acid) were added to respective cocrystal suspensions to lower ph, and the vials were shaken for additional 96 h at 25 °c in water bath. after equilibration, the concentration of ketoconazole at the attained ph was determined by withdrawing 0.5-1 ml of aliquot for filtration through 0.45 micron filter, followed by hplc analysis. for higher ph, different volumes of 0.1 m naoh solution was added to cocrystal suspensions and shaken for 96 h, and the drug concentration was determined according to the method described above. the solid phase in equilibrium with a suspension was analyzed in situ intermittently by raman spectroscopy and by dsc at the end of the experiment for any conversion of cocrystal to free base. vasoya et al. admet & dmpk 7(2) (2019) 106-130 112 cocrystal solubility in buffered/non-buffered media cocrystal solubilities were determined at ph 1 solution (0.1 m hcl), ph 2 hcl buffer), ph 5 acetate buffer and ph 8 phosphate buffer. buffers were prepared according to the united states pharmacopeia (usp) procedures. for the determination of solubility, 5 ml of each buffer was equilibrated with excess cocrystal in a 20 ml glass vial and shaken for 96 h in a water bath shaker at 25 °c, 0.5-1 ml of samples were withdrawn at 24 and 96 h, and ph of the solution and concentration of ketoconazole were measured. aliquots were filtered by 0.45-micron pvdf filter and analyzed by hplc. the solid phase was analyzed by raman spectroscopy. ketoconazole solubility by adjusting ph with fumaric acid and succinic acid ph-solubility profiles of ketoconazole were also generated using the two coformer acids, e.g., succinic acid and fumaric acid. a series of solutions with different concentrations of acids were prepared and 3 ml quantities of these solutions were distributed in separate glass vials. for the determination of ketoconazole solubility versus ph using succinic acid, the concentration of acid ranged from 100 µg/ml to 50 mg/ml, while, in case of fumaric acid, the concentration of acid ranged 50 µg/ml to 2 mg/ml; the difference in acid concentration was due to difference in aqueous solubility of acids, fumaric acid being much less soluble. to each of these solutions, excess amounts of ketoconazole (~10-15 mg more than the expected solubility at a specific ph) was added and kept for shaking in a water bath shaker at 25 °c for up to 96 h. aliquots of suspensions (0.5-0.7 ml each) were, however, withdrawn from each vial after 4, 24 and 96 h of shaking for analysis of drug concentration by hplc according to the method described earlier. for the analysis of solid phase, contents of the vials were transferred to 3-ml eppendorf tubes and centrifuged using beckman gs 6r centrifuge at 2000 rpm for 5 min. after centrifugation, the supernatants were removed and solid contents were dried in a vacuum oven at 50 °c, which were then analyzed by dsc and raman spectroscopy. solubility of 1:1 physical mixtures of ketoconazole with different coformers aqueous solubility was also determined for 1:1 molar physical mixtures of ketoconazole with each of the coformers (succinic acid and fumaric acid). in this study, different quantities of the mixtures (12 mg to 350 mg) were added to different 20 ml glass vials, and then 5 ml of water was added to each vial. thus, the volume of water in each vial remained the same and only the amount of material added differed. the vials were shaken in the water bath shaker at 25 °c and 250 rpm to achieve equilibrium solubility. ph of the suspensions differed because there were different amounts of acids added in the same volume of water, and all the acids dissolved in water. samples were withdrawn from each vial at 4, 24 and 96 h to measure drug concentration, as described earlier. the solid phase of the vials was analyzed using raman spectroscopy to determine the identity of equilibrium species. from some of the vials, solid contents were separated by centrifugation, dried in a vacuum oven, and then analyzed by dsc. intrinsic dissolution rate intrinsic dissolution profiles of ketoconazole free base and its cocrystals were determined using an intrinsic dissolution apparatus (agilent technologies, santa clara, ca) with 0.5 cm 2 exposed surface area at 25 °c. for each determination, 100 mg of powder was compressed into pellet by carver press at 1 ton pressure for 10 seconds. the apparatus with the exposed surface of the material was immersed into 250 ml of dissolution medium and rotated at 200 rpm. ph of dissolution media during the intrinsic dissolution testing were maintained by two different methods: in one case, unbuffered media were used where either 0.1 m hcl or 0.1 m naoh were added to keep ph constant at 3, 4 or 5 (±0.15), and in the second case, ph 3, 4 and 5 phthalate buffers prepared according to the usp were used as dissolution media to keep ph admet & dmpk 7(2) (2019) 106-130 solubility and dissolution of cocrystals doi: 10.5599/admet.661 113 constant. two ml of aliquot was withdrawn at each of 5, 15, 30, 60, 90 and 120 min time points, filtered by 0.45-micron pvdf filter, and analyzed by hplc for drug concentration. results and discussion ketoconazole cocrystals ketoconazole-fumaric acid and ketoconazole-succinic acid cocrystals prepared by the solvent evaporation method in the present investigation were analyzed by differential scanning calorimetry (dsc) and powder x-ray diffraction (pxrd) to assess their crystallinity and purity. dsc scans of two cocrystals as well as those of parent compounds (ketoconazole, fumaric acid and succinic acid) are shown in fig. 1, and the corresponding pxrd patterns are given in fig. 2. ketoconazole showed a melting endotherm at 150 °c, while ketoconazole-fumaric acid and ketoconazole-succinic acid cocrystals showed melting endotherms at 170 °c and 167 °c, respectively; these values are in agreement with those reported elsewhere in the literature [28]. the relatively sharp endotherms indicated that the cocrystals were highly crystalline, and the absence of any separate peaks for ketoconazole alone or free acids in dsc scans confirmed the absence of any impurity due to the presence of unconverted ketoconazole or coformers in cocrystals. the pxrd patterns in fig. 2 showed characteristic peaks for ketoconazole cocrystals and their parent materials, indicating that the cocrystals are highly crystalline. raman spectra in fig. 3 also demonstrated that all the materials have distinct spectral patterns that are different from each other. ketoconazole is a weakly basic drug having two pka values of 2.94 and 6.51 [27]. it formed cocrystals with weak organic acids such as succinic acid and fumaric acid. succinic acid has pka values of 3.93 and 5.30 for its first and second ionization, respectively, while fumaric acid has pka values of 2.74 and 4.17 for its first and second ionization, respectively. childs et al. [36] postulated that a pka difference of greater than 3 between a basic drug and its conjugate acid (∆pka >3) favors salt formation rather than cocrystals, while the cocrystal formation is favored when ∆pka is <3. the ∆pka for ketoconazole and succinic acid is <3 and, therefore, the formation of cocrystal was expected. although the pka difference between ketoconazole and fumaric acid is >3, a cocrystal, and no salt, was formed. the cocrystal formation could be due to hydrogen bonding between acetyl group of ketoconazole and one of the acid groups in succinic acid or fumaric acid. the second carboxyl groups in the acids could also form hydrogen bonds with imidazole n-atom of ketoconazole [28]. as reported by martin et al. [28] for ketoconazole-fumaric acid pair, hydrogen bonding is favored instead of proton transfer due to spatial crystal environment in ketoconazole-fumaric acid cocrystals. they also observed that the structure of ketoconazole-succinic acid cocrystal completely resembles that of ketoconazole-fumaric acid cocrystal. raman spectra shown in fig. 3 also indicate cocrystal formation. ketoconazole has a characteristic peak at 1640 cm -1 for -c=o stretch while raman spectra of succinic acid shows a peak at 1650 cm -1 for -c=0 (carbonyl group) of acid [37,38]. in the ketoconazole-succinic acid cocrystal spectrum, the -c=o peak shifts to 1724 cm -1 because of hydrogen bond formation between one of the carboxylic acid groups of succinic acid and the -c=o group of ketoconazole. this change in raman spectrum of ketoconazole-succinic acid cocrystal shows the formation of hydrogen bond. in this cocrystal, one more hydrogen bond formation occurs between the second carboxyl group of succinic acid and nitrogen of imidazole group of ketoconazole. due to the participation of nitrogen of imidazole (ketoconazole) in hydrogen bond formation, a shift in n=c—h stretch from 3113 cm -1 to 3131 cm -1 could be observed in the cocrystal as compared to the ketoconazole spectrum. in the ketoconazole-succinic acid cocrystal, the four-member ring formation involving two ketoconazole and two succinic acid moieties was observed [28], which is supported by two hydrogen bonds linking all four molecules in a circuit network. vasoya et al. admet & dmpk 7(2) (2019) 106-130 114 figure 1. comparative dsc scans of ketoconazole, succinic acid, fumaric acid, and ketoconazole cocrystals with fumaric acid and succinic acid. figure 2. comparative powder x-ray patterns of ketoconazole, succinic acid, fumaric acid, and ketoconazole cocrystals with fumaric acid and succinic acid. similarly, raman spectrum of fumaric acid shows peak at 1682 cm -1 for the carboxylic acid group that shifts towards a higher wave number of 1718 cm -1 due to hydrogen bonding with -c=o group of ketoconazole in the cocrystal. the ketoconazole -c=o peak shifts from 1640 cm -1 to 1665 cm -1 due to hydrogen bond formation with the cocrystal. also, a hydrogen bond is formed between the second carboxylic acid group of fumaric acid and nitrogen of imidazole group in ketoconazole. thus, ketoconazoleadmet & dmpk 7(2) (2019) 106-130 solubility and dissolution of cocrystals doi: 10.5599/admet.661 115 fumaric acid cocrystal forms a similar 4-member ring structure as that of the ketoconazole-succinic acid cocrystal. figure 3. comparative raman spectral scans of ketoconazole, succinic acid, fumaric acid, and ketoconazole cocrystals with fumaric acid and succinic acid. ph versus solubility of ketoconazole base for a comparison of solubility of the basic drug, ketoconazole, with that of cocrystals, the ph-solubility profile of ketoconazole was determined using hcl to adjust ph. the results are tabulated in table 1, and they are also plotted in fig. 4 for visual observation of the nature of change in solubility as a function of ph. the intrinsic solubility of the ketoconazole free base, i.e., the solubility of the nonprotonated species, at ph higher than 8 was found to be around 2 µg/ml, and the solubility increased exponentially as the ph was decreased gradually in the range of approximately 5.5 and 3, as indicated by the linearity of the semilogarithmic plot in fig. 4. the highest solubility of 148.3 mg/ml (not shown in figure) for ketoconazole at ph 3.1 was measured in the present study, when the solution turned orange. any further reduction in ph to increase solubility by still keeping the solution in equilibrium with a solid phase was not possible; the added ketoconazole dissolved if the ph was decreased by adding concentrated hcl and the ph went back to around 3 when more solid ketoconazole was added to the solution. although there is a report that the preparation of ketoconazole dihydrochloride salt having very high solubility and the melting point of 229 °c is feasible [39], neither the monohydrochloride salt nor the dihydrochloride salt precipitated out during the ph-solubility study in the present investigation. attempts to crystallize ketoconazole hydrochloride salts from anhydrous methanolic solutions in separate studies were also negative as they yielded amorphous materials. vasoya et al. admet & dmpk 7(2) (2019) 106-130 116 table 1 equilibrium solubility of ketoconazole base at different ph at 25 °c, where ph of aqueous suspensions was lowered using hcl solutions. ph and corresponding solubilities from individual vials, including repeating observations, are shown. ph of suspension solubility (μg/ml) ph of suspension solubility (μg/ml) ph of suspension solubility (μg/ml) 3.1 148320 4.28 729 7.23 4 3.35 5457 4.48 345 7.23 4 3.35 5429 4.56 354 7.27 3 3.37 7100 4.56 354 7.72 2 3.58 3425 4.95 178 7.92 3 3.67 3031 4.99 175 8.49 2 3.71 2782 4.99 175 3.81 1931 7.12 5 as mentioned earlier, ketoconazole has two pka values of 2.94 and 6.51, and the increase in solubility of the compound plotted in fig. 4 is obviously due to the effect of the pka of 6.51, since any significant effect of the pka of 2.94 may not be observed in the ph range studied. therefore, the solubility data are plotted in the figure according to the equation (1), where st is the total solubility at a particular ph and [b]s is the intrinsic solubility of the free base [31], which in case of ketoconazole was taken as 2 µg/ml. 𝑆𝑇,𝑏𝑎𝑠𝑒 (𝑝𝐻 > 𝑝𝐻𝑚𝑎𝑥 ) = [𝐵]𝑠 (1 + [𝐻3𝑂 +] 𝐾𝑎 ) = [𝐵]𝑠 (1 + 10 𝑝𝐾𝑎−𝑝𝐻 ) (1) the linearity of the semi-logarithmic plot of solubility as a function of ph in the ph range of ~6.0 to ~3.5 showed that the increase in solubility follows the classic henderson-hasselbalch equation. there was no deviation from the linearity of the graph until the ph decreased below 3.5 since there was no effect of the second pka value on the solubility of ketoconazole at higher ph. the nonlinear increase in solubility around ph 3 in the semi-logarithmic plot could be due to second protonation of the molecule because of the effect of the second pka value. ph versus solubility of ketoconazole cocrystals the solubility of ketoconazole cocrystal as a function of ph was determined by adjusting the ph of a saturated solution of cocrystal to a lower ph by adding aqueous solutions of respective coformer and to a higher ph by adding naoh aqueous solutions. the results are plotted in fig. 5. when excess amounts of fumaric acid and succinic acid cocrystals were dissolved in water, ph values of the saturated solutions were, respectively, 4.1 and 4.2, and the solubilities of cocrystals were, respectively, 2.6 and 1.5 mg/ml (as ketoconazole equivalents). as examined by dsc and pxrd analyses, the solid phases in equilibria with these cocrystal solutions in water were mixtures of cocrystals and free base, the fraction of free base being relatively very small in the total solid. as shown in fig. 5, when ph values of cocrystal solutions in water were lowered by adding respective acidic coformers, there were initial decreases in solubility, which then increased with continued addition of coformer solutions to further decrease ph. the solid phases in equilibria with these solutions were only cocrystals since no conversion to free base occurred when excess coformer was added. the mechanism of the decrease in solubility that was followed by the increase has admet & dmpk 7(2) (2019) 106-130 solubility and dissolution of cocrystals doi: 10.5599/admet.661 117 not been elucidated. it is possible that some forms of complexes may exist between ketoconazole and coformers in solution in this ph range and the solubility profiles reflect the change in nature and solubility of such complexes. figure 4. ketoconazole ph-solubility profile generated by using hydrochloric acid to adjust ph. when ph values of the saturated solutions of ketoconazole cocrystals with fumaric acid and succinic acid were increased by adding naoh solutions, the solubility decreased for both cocrystals. the ph versus solubility of ketoconazole free base, which was generated by the adjustment of ph using hcl or naoh, is also plotted in fig. 5. a comparison of the three plots show that the solubility of the free base and the two cocrystals were identical at ph>4.1. solid phases in equilibria with solutions during the determination of ph-solubility profiles under these ph conditions were mixtures of ketoconazole free base and cocrystals, and only at ph>5, the full conversion of the solid phase to the free base form was observed. it is apparent that when alkali is added to adjust ph, it neutralizes the acidic coformers for partial conversion of the solid phase to the free base, and when sufficient alkali is added and the ph rises above 5, all the excess cocrystals convert to the free base. it is possible that initially formed free base coated the surface of cocrystal particles and thereby only the free base was in equilibrium with the solution. these results demonstrate that the solubility of cocrystals follows the general ph-solubility profile of free base above a certain critical ph value, where the solubility decreases with the increase in ph. this is analogous to the decrease in solubility of pharmaceutical salts above the phmax, i.e., the ph of maximum solubility, as a function of ph [32]. vasoya et al. admet & dmpk 7(2) (2019) 106-130 118 solubility of ketoconazole cocrystals at different ph in buffered solutions in the above section, fig. 5 shows a decrease in solubility of ketoconazole cocrystals with the increase in ph. however, as mentioned earlier, there are reports in the literature where increased solubility of cocrystals under high ph conditions was predicted based on theoretical calculations only [20]. since, for the development of drug products, it is necessary to experimentally ascertain solubility and dissolution rate in the gastrointestinal ph range, the solubility of ketoconazole cocrystals was also studied in buffered solutions and compared with that of the free base. as shown in fig. 5, there are certain limits how much lowering of ph may be achieved by adding acidic coformer solutions. for this reason, the solubility of ketoconazole cocrystals was also determined in acidic buffers to mimic gastric ph conditions. the results of these studies are tabulated in table 2. it may be noted in this table that there were certain shifts in ph of buffers after equilibration with solids, which may be attributed to the liberation of either fumaric acid or succinic acid from cocrystals. the solubility of free base and cocrystals somewhat differed due to the ph shift. difference in buffer compositions and ionic strength may also play roles in the observed difference in solubility in table 2. however, overall, table 2 shows the solubility of ketoconazole cocrystals indeed decreases at relatively higher ph conditions and the solubility of free base and cocrystals under such ph conditions are essentially similar. such similarity in phsolubility profiles of ketoconazole cocrystals and free base is analogous to the similarity in ph-solubility profiles of salt and free base irrespective of whether one or the other is used as starting material to determine solubility [30, 33]. figure 5. ph versus solubility profiles of ketoconazole-fumaric acid and ketoconazole-succinic acid cocrystals determined by using respective acidic coformers at ph lower than the ph of saturated solutions of cocrystals in water and by using naoh to obtain higher ph. the solubility of the ketoconazole free base determined by adjusting ph with hcl is also plotted for reference. admet & dmpk 7(2) (2019) 106-130 solubility and dissolution of cocrystals doi: 10.5599/admet.661 119 table 2. solubility of ketoconazole (ktz) base, ketoconazole-fumaric acid cocrystal and ketoconazole-succinic acid cocrystal in different buffer solutions along with initial ph of buffers and ph after equilibration with solids. solubility represent average of three determinations (± standard deviation). initial ph and buffers used ketoconazole (ktz) ktz-fumaric acid cocrystal ktz-succinic acid cocrystal ph after equilibrat ion solubility (mg/ml) ph after equilibration solubility (mg/ml) ph after equilibration solubility (mg/ml) ph 1.1 (0.1 m hcl) 2.9* 27* 2.03± 0.01 21.54± 0.55 2.52± 0.05 23.99± 0.30 ph 2.1 hcl/kcl buffer 3.6* 3.4* 2.89 ± 0.02 3.39± 0.02 3.38 ± 0.00 3.69± 0.07 ph 5.0 acetate buffer 4.95 0.13 4.31 ± 0.02 0.60± 0.02 4.60± 0.02 0.38± 0.03 ph 7.9 phosphate buffer 7.82±0.01 0.001±0.000 7.17 ± 0.01 0.003± 0.002 7.14 ± 0.01 0.006± 0.001 *approximate ph and solubility. as more solid materials are added, ph shifts and solubility changes. ph versus solubility profiles of ketoconazole determined by adding fumaric acid and succinic acid it has been shown earlier in fig. 5 that the solubility of cocrystals was influenced by ph; when ph was raised above ~ph 4.1, the cocrystals converted gradually to free base and the solubility decreased. on the other hand, when ph of the solutions was reduced by adding respective acids (coformers), the solid phase in equilibria with solutions remained unchanged and the solubility remained high. it was, therefore, of interest to determine whether a similar ph versus solubility relationship would exist when ketoconazole free base was used as the starting material and the ph was adjusted using fumaric acid and succinic acid. in particular, any possible conversion of ketoconazole to cocrystals by the reduction of ph during the determination of ph-solubility profile was studied. figs. 6 and 7 give ph-solubility profiles of ketoconazole with respect to fumaric acid and succinic acid. in these studies, the solutions were equilibrated for 96 h. both figs. 6 and 7 were generated by dissolving different amounts of 1:1 molar mixtures of ketoconazole with, respectively, fumaric acid and succinic acid; the ph shift occurred due to the difference in amounts of acids present in added solids. separately, solubility studies were also conducted using pre-dissolved solutions of fumaric acid and succinic acid and adding excess ketoconazole base. the results were the same whether physical mixtures were used or the free base was added to acid solutions, which indicate that acids from physical mixtures first dissolved in water and then the free base equilibrated with the solution. therefore, we have plotted only data from physical mixtures in figs. 6 and 7. in both of these figures, the solubility of ketoconazole increased gradually with the decrease in ph, and, after reaching peak solubilities in the ph range of 3.5 to 4.0, the solubility decreased with further addition of ketoconazole-coformer mixture or the coformer alone. thus, it is apparent that, similar to phmax (the ph of maximum solubility) generally observed in the ph-solubility profile of a free base and its salt form, the phmax also exists for the cocrystal formation in aqueous media. it was confirmed by dsc analysis and in situ raman analysis that, above such phmax, solutions in both figs. 6 and 7 were in equilibria with the solid free base. below phmax, solid phases in equilibria with solutions were mixtures of cocrystals and free base. this is because when the maximum solubility was reached at phmax due the lowering of ph, the equilibrium solid phase was the free base, and, although the cocrystals crystallized out of solutions upon further lowering of ph, the undissolved free base did not readily convert to cocrystals and formed physical mixtures with cocrystals. however, upon further addition of fumaric acid or succinic acid to decrease ph, the free base in the solid phase gradually decreased and the solid converted fully to cocrystals. the ph versus solubility profile of ketoconazole as determined by lowering the ph of free base with hcl, as shown earlier in fig. 4, is also partially superimposed in figs. 6 and 7 for vasoya et al. admet & dmpk 7(2) (2019) 106-130 120 comparison. it is evident that until the onset of cocrystal formation (i.e, above the phmax), ph-solubility profiles were similar whether they were determined by adjusting ph with fumaric acid, succinic acid or hcl. the solubility is also similar to that shown in fig. 5 by increasing ph of cocrystal solutions with naoh. thus, the solubility above phmax was the same for free base and cocrystal and there was no special influence of cocrystal formation. figure 6. the ph-solubility profile of ketoconazole determined by using different amounts of 1:1 molar physical mixtures of ketoconazole and fumaric acid and equilibrating for 96 h. the ph-solubility profile of ketoconazole determined by ph adjustment with hcl is also superimposed for comparison. the solubility profiles from a narrow ph range (boxed region) is shown in expanded scale in the inset to indicate decrease in ketoconazole solubility after conversion to cocrystal during dissolution of the physical mixture. asterisk symbols in the figure indicates solubility where both ketoconazole and its cocrystal with fumaric acid coexist in the solid phase. figure 7. the ph-solubility profile of ketoconazole determined by using different amounts of 1:1 molar physical mixtures of ketoconazole and succinic acid and equilibrating for 96 h. the ph-solubility profile of ketoconazole determined by ph adjustment with hcl is also superimposed for comparison. the solubility profiles from a narrow ph range (boxed region) is shown in expanded scale in the inset to indicate decrease in ketoconazole solubility after conversion to cocrystal during dissolution of the physical mixture. asterisk symbols in the figure indicates solubility where both ketoconazole and ketoconazole-succinic acid cocrystals coexist in the solid phase. effect of equilibration time on ketoconazole solubility and its conversion to cocrystals the conversion of ketoconazole to a cocrystal in the solid phase during the determination of phsolubility profile was a slow process. as mentioned in the experimental section, the solubility was determined by equilibrating solutions for 4, 24 and 96 h by shaking aqueous media with excess solids, and when suspensions with excess 1:1 molar mixture of ketoconazole and acids (coformers) at or below phmax were equilibrated for 4h, the solid phase was the free base and no conversion to cocrystal was observed. at 24 h, the solid phase in some of the suspensions was the mixture of free base and cocrystal, and the saturation solubility was not consistently the same when the experiments were repeated, indicating that admet & dmpk 7(2) (2019) 106-130 solubility and dissolution of cocrystals doi: 10.5599/admet.661 121 the systems did not reach equilibria. the conversion of free base to cocrystal was also influenced by how much excess solid was present in the system; the conversion to cocrystal was faster as larger amounts of excess solids were added during the determination of solubility. for these reasons, the equilibration period was increased to 96 h, when consistent results with respect to both solubility and conversion to cocrystals were obtained. it may be mentioned here that, in separate experiments, we observed that 1:1 physical mixtures of ketoconazole with coformers can be converted immediately and completely to cocrystals by cogrinding with the addition of a small amount of water. thus, the conversion of the solid mixtures to cocrystals appears to be dependent on the proximity of ketoconazole and the coformer and how intimately they are mixed. figure 8. ph-solubility profiles of ketoconazole-fumaric acid physical mixtures after equilibration for 4 and 96 h. asterisk symbols in the figure indicates solubility where both ketoconazole and ketoconazole-fumaric acid cocrystal coexist in the solid phase. to demonstrate the effect of equilibration time on the solubility of ketoconazole, ph-solubility profiles of ketoconazole after equilibration with fumaric acid for 4 and 96 h were determined and the results are given in fig. 8. in this figure, the solubility profile of ketoconazole after equilibration with fumaric acid at each data point for 4 h is compared with the solubility profile after equilibration for 96 h that has been shown earlier in fig. 7. the solubility of ketoconazole increased gradually and the ph decreased when increasing amounts of 1:1 ketoconazole-fumaric acid mixture were equilibrated for 4 h at each time point. the experiment was discontinued when the solubility of ketoconazole reached about 7000 µg/ml (7 mg/ml). there was no conversion of the solid phase to cocrystals at any of the data points. in contrast, the maximum solubility of ketoconazole base observed after equilibration for 96 h was about 3000 µg/ml (3 mg/ml) at ph~3.7 and after that the solubility decreased due to the crystallization of ketoconazole-fumaric acid cocrystal. it may also be noticed in fig. 8 that there was a small increase in ph by about 0.2 unit in the solubility profile determined by equilibration for 96 h as compared to that for 4 h. the mechanism of this effect was not investigated; however, it could be due to complexation of fumaric acid before or after nucleation into cocrystal and thus being unable to act as the free acid form. vasoya et al. admet & dmpk 7(2) (2019) 106-130 122 essentially similar effects of equilibration time were observed when succinic acid was used to adjust ph (data not shown). these results demonstrate that equilibrium solubilities of free base and their cocrystals as well as the equilibrium solid phases may differ depending on how long the systems are equilibrated. therefore, care must be taken in interpreting any solubility differences between free bases and their cocrystal forms by taking the equilibration time into consideration. intrinsic dissolution of ketoconazole free base and cocrystals at different ph fig. 9 gives the dissolution profiles of ketoconazole and its cocrystals with fumaric acid and succinic acid at ph 3, 4 and 5 from a constant surface area of 0.5 cm 2 . these ph conditions were selected because ph 3 was below the phmax, ph 4 was at around the phmax and ph 5 was above the phmax. additionally, ph-solubility profiles in figs. 5, 6 and 7 showed a very large decrease in solubility of ketoconazole at ph 5 as compared to those at ph 3 and 4, and it was, therefore, of interest to study whether the difference is also reflected in the dissolution rate. there are two sets of dissolution profiles in fig. 9: the first was in unbuffered media (left hand column), where the ph of dissolution media was kept constant by adding naoh or hcl, as necessary, and the second was in buffered media (right hand column), where no ph adjustment was necessary since the ph did not change due to the dissolution of drug. the dissolution profiles at ph 3, 4 and 5 under ph-stat conditions in unbuffered media in fig. 9 demonstrate that there was a great impact of cocrystal formation on the dissolution rate of ketoconazole. at all three ph conditions, dissolutions rates of cocrystals were much higher than those of the free base. previously, comparative dissolution studies of free base and salt forms of drugs were extensively reported in the literature [30-33], where the higher dissolution rates of salts than that of their respective free base forms were attributed to the difference in microenvironmental ph conditions. essentially, the microenvironmental ph represents the ph at the surface of solid when the diffusion layer thickness, h, approaches zero, i.e., h=0, and the solubility of drug at such a ph, and not at the bulk ph of the dissolution medium, dictates the dissolution rate of drug. the microenvironmental ph at the surface of dissolving solid is often represented as phh=0. the difference in dissolution rates of ketoconazole and its cocrystals may also be explained by possible differences in their phh=0. to confirm the relationship between microenvironmental ph and superior dissolution rates of cocrystals as compared to the free base, dissolution studies of cocrystals and the free base at ph 3, 4 and 5 were also conducted in buffered media, and the results are given in fig. 9 in the right hand side along with those in the left hand side for unbuffered media. table 3 gives their phh=0 values under buffered bulk ph conditions of 3, 4 and 5 as determined by the slurry ph method described earlier in the literature [30, 40]. essentially, in the present investigation, about 0.5 ml of buffer was added to 200 mg of powder in a small vial and mixed by vortexing, and then ph was recorded when an equilibrium was reached after 5-10 min of mixing. although no dissolution testing was done at ph 7, phh=0 values at this bulk ph condition are also given in table 2 for a comparison with those at ph 5. it may be observed that phh=0 values of ketoconazole cocrystals during dissolution in ph 3, 4 and 5 buffers were in the range of 3.13 and 4.38. the dissolution profiles in fig. 9 are in general agreement with the ketoconazole solubility in this microenvironmental ph range, as shown in fig. 5. for example, it may be noted in table 3 that phh=0 values of the free base and the two cocrystals in the ph 4 buffer ranged from 4.08 to 4.38, and due to the similarity in solubility in this ph range, their dissolution profiles in fig. 9 were also similar. when the free base by itself is considered, its phh=0 values at ph 3 and 4 buffers were, respectively, 4.17 and 4.38 (table 3) and, therefore, its dissolution rates under ph 3 and 4 bulk ph conditions were also relatively high as compared to that in the ph 5 buffer (fig. 9), where the microenvironmental ph did not change significantly and the solubility would be low. admet & dmpk 7(2) (2019) 106-130 solubility and dissolution of cocrystals doi: 10.5599/admet.661 123 it may be noticed in table 3 that throughout the wide bulk ph range of 3 to 7, phh=0 of the succinic acid cocrystal remained within a narrow range of 4.2 to 4.7. it is apparent that this is due to the buffering action of succinic acid present in the cocrystal, since succinic acid has two pka values of 4.2 and 5.6 and a buffer would be formed around the lower pka value. the fumaric acid cocrystals exhibited somewhat lower phh=0 value since it would possibly buffer to a lower ph due to lower pka values (3.0 and 4.4). thus, the acidity of coformers can have certain effects on the dissolution of formed cocrystals as different coformers can modulate microenvironmental ph differently. figure 9. intrinsic dissolution profiles of ketoconazole base (ktz) and its cocrystals ketoconazole-fumaric acid (ktz fum) and ketoconazole-succinic acid (ktz suc) from a constant surface area of 0.5 cm 2 in unbuffered media where ph was kept constant by adding hcl or naoh, as necessary (left hand column) and in buffers (right hand column). vasoya et al. admet & dmpk 7(2) (2019) 106-130 124 in addition to the modulation of microenvironmental ph, the dissolution rates of cocrystals and free base may also be influenced by the supersaturation of drugs in the diffusion layer. such supersaturation was observed earlier during the dissolution of salts [33] and free bases [30], which greatly influenced their dissolution rates. it is possible that a similar supersaturation occurred in diffusion layers during the dissolution of ketoconazole cocrystals at relatively high ph (e.g., ph 5) due to the acid-base interaction; the dissolution of free base at relatively lower ph (e.g., ph 3 and 4) would also be influenced by such an interaction. therefore, a combination of the lowering of surface ph (phh=0) and the supersaturation of drug in the diffusion layer could be responsible for high dissolution rates of cocrystals observed in ph 5 buffered and unbuffered media. in addition to phh=0 values for ph 3, 4 and 5 buffers, such values for the ph 7 buffer are also given in table 3. since phh=0 values for fumaric acid and succinic acid cocrystals at this ph were, respectively, 4.53 and 4.67, where the drug solubility is expected to be much higher than that at the phh=0 of free base (ph 7.05), it is expected that dissolution rates of cocrystals at ph 7 would also be much higher than that of the free base. table 3. microenvironmental ph (phh=0 ) at the surface of ketoconazole (ktz) free base, ketoconazole-fumaric acid (ktz-fum) cocrystal and ketoconazole-succinic acid (ktz-suc) cocrystal during dissolution testing in different buffers as determined by the slurry ph method. each ph value is the average of two determinations buffers ph h=0 (slurry ph) composition initial ph (bulk ph) ktz base ktz-fum cocrystal ktz-suc cocrystal ph 3 phthalate buffer 2.93 4.17 3.13 4.38 ph 4 phthalate buffer 3.96 4.38 4.08 4.19 ph 5 neutralized phthalate buffer 4.97 5.00 4.23 4.40 ph 7 phosphate buffer 6.97 7.05 4.53 4.67 surface ph or phh=0 during the dissolution of cocrystals and free base in unbuffered media could not be accurately predicted by the slurry ph method. unlike dissolution testing, where hcl or naoh were continuously added to keep the ph of bulk media constant, no such addition of acid or alkali was possible during the determination of slurry ph. since the media did not have any buffer capacity, the ph in presence of excess solids in these media were essentially similar to those in water. however, the dissolution profiles in unbuffered media in fig. 9 indicated that they were also influenced by the microenvironmental ph at the surface of dissolving solids. table 4. intrinsic dissolution rates of ketoconazole (ktz) free base and its cocrystals with fumaric acid (ktz-fum) and succinic acid (ktz-suc) in unbuffered and buffered media at different ph. ph dissolution rate constant (µg.min -1 .cm -2 ) unbuffered media* buffers ktz-fum ktz-suc ktz base ktz-fum ktz-suc ktz base ph 3 164 185 106 108 91 108 ph 4 124 105 16 71 65 45 ph 5 93 62 -** 51 48 -** *ph of bulk media kept constant by adding hcl or naoh solutions, as necessary **extremely low dissolution rate (practically zero) that could not be calculated with accuracy. intrinsic dissolution rates of ketoconazole-fumaric acid cocrystal, ketoconazole-succinic acid cocrystal and ketoconazole free base at ph 3, 4 and 5, as calculated from the graphs in fig. 9, are given in table 4, which essentially confirm high dissolution rates of both cocrystals at ph 3, 4 and 5 mentioned above. in contrast, dissolution rates of the free base decreased with the increase in ph from 3 to 4, and at ph 5, the admet & dmpk 7(2) (2019) 106-130 solubility and dissolution of cocrystals doi: 10.5599/admet.661 125 dissolution rate decreased to practically negligible level. although individual values for dissolution rate constants at each ph may vary due to such factors as microenvironmental ph, supersaturation, etc., mentioned earlier, the results demonstrate that although there may not be significant difference in equilibrium solubility between ketoconazole free base and cocrystals at any particular ph, the cocrystal formation can greatly increase dissolution rates of basic drugs at relatively high ph conditions by modulating microenvironmental ph during dissolution testing. consequently, for certain basic drugs like ketoconazole, dissolution rates of cocrystals could be higher than their respective free base forms. the mechanism why dissolution rates of cocrystals at ph 3, 4 and 5 in table 4 in unbuffered media were higher than those in buffered media has not been investigated. summary and conclusions to address different issues with respect to solubility and dissolution rate of cocrystal and their free base form, we have outlined in introduction of the paper several objectives for the present investigation. our findings and conclusions against those objectives are summarized below: a) comparative solubility of free base and cocrystals as a function of ph – ph-solubility profiles of ketoconazole cocrystals with fumaric acid and succinic showed phmax (ph of maximum solubility) at ph ~ 3.5-4.0. under ph conditions above phmax, the solubility profiles were identical whether fumaric acid, succinic acid, hcl or naoh were used to adjust ph, and the profiles were also similar whether the free base or cocrystals were used as starting materials to determine solubility. at ph<phmax, the solubility decreased in case of the addition of fumaric acid and succinic acid due to the formation of cocrystals, while the solubility continued to increase when hcl was added to lower ph as no cocrystal or salt were formed. thus, above phmax (in this case ph~4), there is no solubility advantage of cocrystal formation if the ph is maintained the same, and the solubility may differ only if ph of saturated solutions of free base and cocrystals also differ. usually, ph values of the suspensions of cocrystals in water are lower than those of the free base in water, and, as a result, cocrystals usually exhibit higher solubility. b) solid phases in equilibria with saturated solutions – during the determination of ph-solubility profiles of ketoconazole by adding fumaric acid, succinic acid or hcl, the solubility continued to increase according to the henderson-hasselbalch equation and the solid phase in equilibria with solutions was only the free base. after reaching a relatively high solubility at a certain ph during the addition fumaric acid or succinic acid, which may be referred to as phmax or the ph of maximum solubility, the cocrystal precipitated out. initially, the precipitate was a mixture of cocrystal and free base as the conversion of free base to cocrystal in the excess solid present was relatively slow. when the ph-solubility profile was determined by adding hcl, instead of the addition of fumaric acid or succinic acid, free ketoconazole base was always the equilibrium species as no cocrystals or salts were formed. when cocrystals were used as starting materials, instead of the free base, for the determination of phsolubility profile, the excess solids at phmax and lower were only cocrystals, and mixtures of cocrystals and free base were present when the ph was raised above phmax. thus, the conversion from free base to cocrystal and vice versa occur, respectively, when the ph is decreased below or increased above the phmax. c) similarity in ph vs solubility profiles of cocrystals and salt – ph-solubility profiles of ketoconazolefumaric acid and ketoconazole-succinic acid cocrystals obtained in the present investigation appeared to be similar to ph-solubility profiles commonly encountered with salts. in both cases, there are phmax (the ph of maximum solubility) and the free base is the equilibrium species at ph>phmax. the only difference between the solubility profiles of cocrystals and salts is the nature of solid phases at vasoya et al. admet & dmpk 7(2) (2019) 106-130 126 ph<phmax; it is cocrystal in the ph-solubility profile of cocrystal, while salt is the equilibrium species in the ph-solubility profile of salt. considering the phenomenon in another way, the two solid phases at ph<phmax have certain similarities as both are complexes, the cocrystal being an acid-base nonionic complex and salt being an acid-base ionic complex. accordingly, since both cocrystal and salt are solid complexes with definite stoichiometric ratios between basic drugs with acidic conformer or counter ion, the cocrystal may be considered as a new chemical entity similar to a salt. d) cocrystal vs free base intrinsic dissolution rates – although the solubility of cocrystals and the free base were identical at ph>phmax, their dissolution rates differed. dissolution rates of cocrystals were found to be much higher than that of the free base at such ph due to the modulation of microenvironmental ph at the dissolving surface of cocrystals. for example, at ph 5, the dissolution rate of the cocrystal was higher than that of the free base, because the ph at the surface of cocrystal was lower than 5, where the solubility of ketoconazole was much higher. it was also predicted by determining microenvironmental ph that the dissolution rates of cocrystals at ph 7 would also be much higher than that of the free base. thus, under intestinal ph conditions, the cocrystals can have better dissolution rates and, possibly, better bioavailability. it is hoped that the results presented in this report will clarify some of the basic concepts with respect to solubility and dissolution rates of cocrystals and will provide a systematic approach to identify and select cocrystals for dosage form development. it should, however, be mentioned here that the results may be particularly applicable to basic drugs that demonstrate ph-dependent profiles in the gastrointestinal ph range. the concepts may also be applicable to cocrystals of acidic compound. however, they may not be generalized for cocrystals of neutral compounds where no significant change in solubility as a function of ph is observed. further studies with cocrystals of relatively neutral compounds are currently underway in our laboratory. acknowledgement: this work was presented, in part, at the special session on ‘pharmaceutical cocrystals – physicochemical properties and formulations’ held at the 6 th world conference on physico-chemical methods in drug discovery and development that was organized by international association of physical chemists (iapc-6), at zagreb, croatia, in september 2017. further studies were conducted based on feedback received from the participants of the session. the author abu serajuddin thanks dr. alex avdeef (in-adme research, new york) and prof. zoran mandić (university of zagreb, croatia) for invitation to the conference, and he is also grateful to prof. mandić for continued encouragement to write this paper. conflict of interest: authors declare no conflict of interest associated with the present research work. references [1] j. aaltonen, m. alleso, s. mirza, v. koradia, k.c. gordon, j. rantanen. solid form screening--a review. european journal of pharmaceutics and biopharmaceutics 71 (2009) 23-37. [2] p. vishweshwar, j.a. mcmahon, j.a. bis, m.j. zaworotko. pharmaceutical co-crystals. journal of pharmaceutical sciences 95 (2006) 499-516. [3] n. schultheiss, a. newman. pharmaceutical cocrystals and their physicochemical properties. crystal growth & design 9 (2009) 2950-2967. admet & dmpk 7(2) (2019) 106-130 solubility and dissolution of cocrystals doi: 10.5599/admet.661 127 [4] p. cerreia vioglio, m.r. chierotti, r. gobetto. pharmaceutical aspects of salt and cocrystal forms of apis and characterization challenges. advanced drug delivery reviews 117 (2017) 86-110. [5] d. douroumis, s.a. ross, a. nokhodchi. advanced methodologies for cocrystal synthesis. advanced drug delivery reviews 117 (2017) 178-195. [6] fda: regulatory classification of pharmaceutical co-crystals guidance for industry. https://www.fda.gov/downloads/drugs/guidances/ucm281764.pdf. published in february 2018. last accessed august 17, 2018. [7] s. aitipamula, r. banerjee, a.k. bansal, k. biradha, m.l. cheney, a.r. choudhury, g.r. desiraju, a.g. dikundwar, r. dubey, n. duggirala, p.p. ghogale, s. ghosh, p.k. goswami, n.r. goud, r.r.k.r. jetti, p. karpinski, p. kaushik, d. kumar, v. kumar, b. moulton, a. mukherjee, g. mukherjee, a.s. myerson, v. puri, a. ramanan, t. rajamannar, c.m. reddy, n. rodriguez-hornedo, r.d. rogers, t.n.g. row, p. sanphui, n. shan, g. shete, a. singh, c.c. sun, j.a. swift, r. thaimattam, t.s. thakur, r. kumar thaper, s.p. thomas, s. tothadi, v.r. vangala, n. variankaval, p. vishweshwar, d.r. weyna, m.j. zaworotko. polymorphs, salts, and cocrystals: what’s in a name? crystal growth & design 12 (2012) 2147-2152. [8] d.j. berry, j.w. steed. pharmaceutical cocrystals, salts and multicomponent systems; intermolecular interactions and property based design. advanced drug delivery reviews 117 (2017) 3-24. [9] d.p. kale, s.s. zode, a.k. bansal. challenges in translational development of pharmaceutical cocrystals. journal of pharmaceutical sciences 106 (2017) 457-470. [10] r.m. poole, r.t. dungo. ipragliflozin: first global approval. drugs 74 (2014) 611-617. [11] a. kumar, s. kumar, a. nanda. a review about regulatory status and recent patents of pharmaceutical co-crystals. advanced pharmaceutical bulletin 8 (2018) 355-363. [12] l. feng, p.h. karpinski, p. sutton, y. liu, d.f. hook, b. hu, t.j. blacklock, p. e. fanwick, m. prashad, s. godtfredsen, c. ziltener. lcz696: a dual-acting sodium supramolecular complex. tetrahedron letters 53 (2012) 275-276. [13] assessment report: steglatro, european medicines agency, 25 january, 2018. https://www.ema.europa.eu/documents/assessment-report/steglatro-epar-public-assessmentreport_en.pdf. last accessed january 12, 2019. [14] r. thakuria, a. delori, w. jones, m.p. lipert, l. roy, n. rodríguez-hornedo. pharmaceutical cocrystals and poorly soluble drugs. international journal of pharmaceutics 453 (2013) 101-125. [15] d.j. good, n. rodríguez-hornedo. solubility advantage of pharmaceutical cocrystals. crystal growth & design 9 (2009) 2252-2264. [16] a. kozak, p.h. marek, e. pindelska. structural characterization and pharmaceutical properties of three novel cocrystals of ethenzamide with aliphatic dicarboxylic acids. journal of pharmaceutical sciences 108 (2019) 1476-1485. (https://doi.org/10.1016/j.xphs.2018.10.060) [17] m.k. stanton, a. bak. physicochemical properties of pharmaceutical co-crystals: a case study of ten amg 517 co-crystals. crystal growth & design 8 (2008) 3856-3862. [18] r. chadha, a. saini, s. khullar, d.s. jain, s.k. mandal, t.n. guru row. crystal structures and physicochemical properties of four new lamotrigine multicomponent forms. crystal growth & design 13 (2013) 858-870. [19] y. gao, j. gao, z. liu, h. kan, h. zu, w. sun, j. zhang, s. qian. coformer selection based on degradation pathway of drugs: a case study of adefovir dipivoxil-saccharin and adefovir dipivoxilnicotinamide cocrystals. international journal of pharmaceutics 438 (2012) 327-335. [20] m.l. cheney, n. shan, e.r. healey, m. hanna, l. wojtas, m.j. zaworotko, v. sava, s. song, j.r. sanchez-ramos. effects of crystal form on solubility and pharmacokinetics: a crystal engineering case study of lamotrigine. crystal growth & design 10 (2010) 394-405. [21] c. ràfols, h. fael, e. fuguet, b. outhwaite, s. lee, r. ruiz. dissolution rates of ciprofloxacin and its cocrystal with resorcinol. admet and dmpk 6 (2018) 61-70. http://www.fda.gov/downloads/drugs/guidances/ucm281764.pdf http://www.ema.europa.eu/documents/assessment-report/steglatro-epar-public-assessment-report_en.pdf http://www.ema.europa.eu/documents/assessment-report/steglatro-epar-public-assessment-report_en.pdf vasoya et al. admet & dmpk 7(2) (2019) 106-130 128 [22] y.m. chen, n. rodríguez-hornedo. cocrystals mitigate negative effects of high ph on solubility and dissolution of a basic drug. crystal growth & design 18 (2018) 1358-1366. [23] m.r. shimpi, a. alhayali, k.l. cavanagh, n. rodríguez-hornedo, s.p. velaga. tadalafil–malonic acid cocrystal: physicochemical characterization, ph-solubility, and supersaturation studies. crystal growth & design 18 (2018) 4378-4387. [24] n.j. babu, a. nangia. solubility advantage of amorphous drugs and pharmaceutical cocrystals. crystal growth & design 11 (2011) 2662-2679. [25] k.l. cavanagh, c. maheshwari, n. rodríguez-hornedo. understanding the differences between cocrystal and salt aqueous solubilities. journal of pharmaceutical sciences 107 (2018) 113-120. [26] a. avdeef. cocrystal solubility product prediction using an in combo model and simulations to improve design of experiments. pharmaceutical research 35 (2018) 40. [27] m.t. esclusa-diaz, m. guimaraens-méndez, m.b. pérez-marcos, j.l. vila-jato, j.j. torres-labandeira. characterization and in vitro dissolution behaviour of ketoconazole/βand 2-hydroxypropyl-βcyclodextrin inclusion compounds. international journal of pharmaceutics 143 (1996) 203-210. [28] f.a. martin, m.m. pop, g. borodi, x. filip, i. kacso. ketoconazole salt and co-crystals with enhanced aqueous solubility. crystal growth & design 13 (2013) 4295-4304. [29] a.t.m. serajuddin, c.i. jarowski. effect of diffusion layer ph and solubility on the dissolution rate of pharmaceutical acids and their sodium salts. ii: salicylic acid, theophylline, and benzoic acid. journal of pharmaceutical sciences 74 (1985) 148-154. [30] a.t.m. serajuddin, c.i. jarowski. effect of diffusion layer ph and solubility on the dissolution rate of pharmaceutical bases and their hydrochloride salts i: phenazopyridine. journal of pharmaceutical sciences 74 (1985) 142-147. [31] a.t.m. serajuddin. m. pudipeddi, d. j. w. grant, p. h. stahl. solubility and dissolution of weak acids, bases, and salts. in handbook of pharmaceutical salts: properties, selection, and use. , wiley-vch, zurich, switzerland, 2002, pp. 19-39. [32] a.t. serajuddin. salt formation to improve drug solubility. advanced drug delivery reviews 59 (2007) 603-616. [33] s. li, s. wong, s. sethia, h. almoazen, y.m. joshi, a.t. serajuddin. investigation of solubility and dissolution of a free base and two different salt forms as a function of ph. pharmaceutical research 22 (2005) 628-635. [34] a. avdeef, e. fuguet, a. llinàs, c. ràfols, e. bosch, g. völgyi, t. verbić, e. boldyreva, k. takács-novák. equilibrium solubility measurement of ionizable drugs – consensus recommendations for improving data quality. admet & dmpk 4 (2016) 117-178. (http://dx.doi.org/10.5599/admet.4.2.292) [35] o.s. marković, m.p. pešić, a.v. shah, a.t.m. serajuddin, t.ž. verbić, a. avdeef. solubility-ph profile of desipramine hydrochloride in saline phosphate buffer: enhanced solubility due to drug-buffer aggregates. european journal of pharmaceutical sciences (2019). (https://doi.org/10.1016/j.ejps.2019.03.014) [36] s.l. childs, g.p. stahly, a. park. the salt−cocrystal continuum: the influence of crystal structure on ionization state. molecular pharmaceutics 4 (2007) 323-338. [37] m.m. al-shalalfeh, a.t. onawole, t.a. saleh, a.a. al-saadi. spherical silver nanoparticles as substrates in surface-enhanced raman spectroscopy for enhanced characterization of ketoconazole. materials science & engineering. c, materials for biological applications 76 (2017) 356-364. [38] m. suzuki, t. shimanouchi. infrared and raman spectra of succinic acid crystal. journal of molecular spectroscopy 28 (1968) 394-410. [39] a.h. hosmani, y.s. thorat. synthesis and evaluation of nanostructured particles of salt of ketoconazole for solubility enhancement. digest journal of nanomaterials and biostructures 6 (2011) 1411-1418. admet & dmpk 7(2) (2019) 106-130 solubility and dissolution of cocrystals doi: 10.5599/admet.661 129 [40] m. pudipeddi, e.a. zannou, m. vasanthavada, a. dontabhaktuni, a.e. royce, y.m. joshi, a.t.m. serajuddin. measurement of surface ph of pharmaceutical solids: a critical evaluation of indicator dye-sorption method and its comparison with slurry ph method. journal of pharmaceutical sciences 97 (2008) 1831-1842. ©2019 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ application of biomimetic hplc to estimate lipophilicity, protein and phospholipid binding of potential peptide therapeutics doi: 10.5599/admet.544 162 admet & dmpk 6(2) (2018) 162-175; doi: http://dx.doi.org/10.5599/admet.544 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper application of biomimetic hplc to estimate lipophilicity, protein and phospholipid binding of potential peptide therapeutics klara valko* 1,2 , gabriela ivanova-berndt 3 , paul beswick 3 , mark kindey 4,5 , dorothy ko 6 1 bio-mimetic chromatography ltd, business & technology centre, stevenage, sg1 2dx united kingdom 2 department of biological and pharmaceutical chemistry, ucl school of pharmacy london wc1n 1ax united kingdom 3 bicycle therapeutics ltd, cambridge, united kingdom 4 department of pharmaceutical sciences, college of pharmacy, university of south florida, tampa, fl, usa 5 james a. haley vamc, tampa, fl, usa 6 genervon biopharmaceuticals llc, pasadena, ca, usa *corresponding author: e-mail: klara_valko@bio-mimetic-chromatography.com; tel.: +447521 989558 received: april 27, 2018; revised: may 28, 2018; published: june 06, 2018 abstract peptide therapeutics are new modalities offering several challenges to drug discovery. they are generally less stable and permeable in vivo. the characterization of their lipophilicity cannot be carried out using the traditional in silico or wet octanol/water partition coefficients. the prediction of their in vivo distribution and permeability is also challenging. in this paper, it is demonstrated that the biomimetic properties such as lipophilicity, protein and phospholipid binding can be easily assessed by hplc using chemically bonded protein and immobilized artificial membrane (iam) stationary phases. the obtained properties for a set of potential therapeutic peptides with 3 to 33 amino acids have been analysed and it was found that similar characteristics of the properties could be observed as for small molecule drugs. the albumin binding showed correlation with their measured lipophilicity on the c-18 stationary phase with acidic peptides showing stronger than expected albumin binding. the (iam) chromatography revealed peptide membrane affinity, which was stronger for positively charged peptides (containing arginine) and showed correlation to the alpha-1-acid glycoprotein (agp) binding, which was also stronger for positively charged compounds. the in vivo volume of distribution and drug efficiency of the peptides have been estimated using the models developed for small molecules. one of the candidate linear peptides has been assessed in various cellular and in vivo assays and the results have confirmed the estimated cell partition and brain to plasma ratio. it can be demonstrated, that up to 21 amino acids, the peaks of the peptides obtained on the protein phase were symmetrical and narrow. the interaction of larger peptides with the protein stationary phases resulted in wide peaks showing multiple equilibrium processes with slow kinetics during chromatography. the larger peptides showed narrow and symmetrical peaks on the iam column enabling the quantification of peptide cell membrane interactions. keywords human serum albumin binding; immobilized artificial membrane; chromatography, peptides, tissue binding, brain to plasma ratio http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:klara_valko@bio-mimetic-chromatography.com admet & dmpk 6(2) (2018) 162-175 biomimetic hplc in the estimation of physchem properties of peptides doi: 10.5599/admet.544 163 introduction pharmaceutical and drug discovery companies are searching for new modalities outside the traditional small molecular drug space. peptides, defined as having less than 50 amino acids, have been long recognized as potential therapeutics [1]. more than 140 peptides are now in clinical trials and more than 60 peptide drugs have been approved by fda [2]. the therapeutic area ranges from antimicrobial to anticancer, but it is continuously expanding to new areas such as irritable bowel syndrome (linaclotide), cushing’s syndrome (pasireotide), myeloma (carfilzomib) and a few new peptides in phase iii clinical trials for osteoporosis and type ii diabetes. it is worth mentioning that 6 peptides therapeutics out of 14 which are in phase iii clinical trials [1] are targeting central nervous system (cns). peptide therapeutics are able to inhibit protein-protein interactions (ppi) and are opening up numerous ppi targets for peptides and macrocyclic large molecules that are outside the traditional drug property space as defined by the lipinski rule of five [3]. in general, peptides are selective and efficacious signalling molecules such as hormones, neurotransmitters, growth factors and ion channel ligands. however, there are several issues with peptide therapeutics that have to be resolved. in spite of usually being water soluble, their oral absorption is compromised because of peptide hydrolysis occurring in the gastrointestinal tract. proteolytic enzymes can decompose peptides quickly resulting in short plasma halflife even when the peptides are administered intravenously [4]. approaches to extend the plasma half-life include substitution of cleavage sites by other amino acids, or to change the folding through enhancement of the secondary structure of the peptide. it is common to use lactam bridges, cyclisation and acylation. another approach is to increase the serum albumin binding of peptide therapeutics to potentially increase the plasma half-life [4]. therefore, measurements of albumin binding of potential peptide therapeutics are very important. it is possible that pharmacologically active peptide therapeutics do not necessarily possess long plasma half-lives. for example, studies on sgx93 (soligenix inc) [5], a short peptide drug candidate that exhibited potential in oral mucositis in clinical trials showed a rapid degradation in plasma. due to their polar nature, peptides often have limited cell permeability. cyclisation, reduction of the hbond donor groups and other methods have been tried to identify cell-permeable peptides (cpp) [6-7]. there are three possible mechanisms by which peptides can enter into the cell: 1. the inverted micelle model; 2. direct penetration and 3. endocytic uptake [8]. when the target is intracellular it is important to design cell penetrating peptide therapeutics [9-10]. various mechanisms on how peptides can enter into the cells have been reviewed by madani et al. [11]. all of the aforementioned mechanisms require the peptide to have good affinity to the negatively charged membrane surface. thus including arginine in the sequence could offer the potential to increases the cell permeability of the peptides [12]. chromatography using immobilized artificial membrane (iam) as a stationary phase [13-14] is a promising tool to measure the affinity of the peptide to the phospholipid membrane and relate it to cell permeability. another aspect of the rational design of peptide therapeutics is to improve their physicochemical properties, such as solubility and stability and to avoid aggregation. the stability of peptides can be improved by stabilizing α-helixes, salt bridges or lactam bridges. therefore, several efforts have been published [6,15–19], for the synthesis of rigid or flexible cyclic peptides in relation to their cell permeability. due to the low oral bioavailability of peptides, various formulation strategies have been developed such as liquid formulations. the search for a non-conventional administration route such as inhaled, topical or intravenous is a potential option to improve bioavailability. klara valko et al. admet & dmpk 6(2) (2018) 162-175 164 therefore, any predictive tool that can be used to estimate the in vivo distribution of potential peptide therapeutics would be very valuable. lipophilicity of molecules, expressed as an octanol/water partition coefficient (log d), has been used traditionally to model biological distribution processes [20]. it was found that the traditional octanol/water partition coefficient measurements were not suitable for the lipophilicity determination of peptides as peptides are charged and very polar and do not partition into octanol at all [21]. moreover it is likely peptides could be concentrated on the interface between the octanol and the water. for instance, surface active peptides may help to form foams and emulsions that make the log d determination very difficult. there is a study that describes a possible calculation method using 3d descriptors for the lipophilicity of the peptides in the octanol/water system [22]. in silico calculation software for log d provides very low calculated lipophilicities (below -5 and -10) in most cases. for example, in this study the calculated log d values using acdlabs (advanced chemical development inc. toronto, canada) software for lp5 was -9.27 and for lp7 it was -9.84. on the other hand, these two compounds showed interactions with the iam stationary phase and also had retention on the c-18 stationary phase. ermondi et al. [23] demonstrated that reversed phase chromatography could be used to assess and compare the lipophilicity of peptides. others have also reported [24-25] that the lipophilicity of peptides is an important physicochemical property and that reversed phase chromatography can provide a tool to assess lipophilicity via retention time measurements. biswas et al. [25] applied reversed-phase chromatography for lipophilicity assessment of amino acids and established an additivity rule of lipophilicity of peptides up to 15 amino acids. however, the additivity rule appeared to break down in larger peptides due to the secondary structure formation. in silico predictions of reversed phase retention may provide the possibility to assess the lipophilicity of peptides based on the retention coefficients for amino acids [26-28]. gussakovsky et al. [29] analysed over 29 thousand tryptic peptides on the iam.pc.dd2 stationary phase and compared the retention contribution of the amino acids with those from the c-18 columns. they found that amino acids containing positive charges such as arginine and lysine retain stronger on the iam phase than on the c18 phase. they also determined that the amino acid contributions depended on the location of the amino acid in the sequence. in this paper, we have investigated the retention behaviour of 20 potential peptide therapeutics, (one of them has been tested in vivo) including linear and bicyclic peptides on reversed-phase and biomimetic hplc stationary phases such as human serum albumin (hsa), alpha-1-acid-glycoprotein (agp) and iam phases [14,30,32]. the measured properties have been analysed in comparison to marketed drugs’ properties. the models developed for small molecules have been applied to the peptides to predict in vivo distribution and cell penetration. experimental all chromatographic measurements were carried out on an agilent 1100 series hplc equipped with a uv-diode array detector. lipophilicity measurements using c-18 stationary phase with acidic, neutral and basic phs the reversed phase fast gradient retention times have been measured using a gemini nt c-18 column with the dimensions of 50 x 3 mm, particle size 5 µm with a 110 å pore size (phenomenex uk). the starting mobile phase was 0.01 m formic acid at ph 2.6 for low ph lipophilicity, 50 mm ammonium acetate buffers with ph adjusted to 7.4 and 10.5 by concentrated ammonia solution for neutral and basic ph lipophilicity admet & dmpk 6(2) (2018) 162-175 biomimetic hplc in the estimation of physchem properties of peptides doi: 10.5599/admet.544 165 measurements, respectively. all reagents were hplc grade obtained from honeywell/sigma-aldrich, germany. a 1.00 ml/min flow rate was applied. the acetonitrile gradient was from 0 to 100 % from 0 to 3.5 min then 100 % acetonitrile was kept until 4.5 min and then returned to 0 % at 4.7 min. the run time was 6 min. hplc grade water and acetonitrile were obtained from rathburn chemicals ltd, walkerburn, uk. the gradient retention times were standardized using the valko test mixture and the chromatographic hydrophobicity index (chi) values as described previously [14,32]. the chi values were converted to the octanol/water log d scale using equation 1. chi log d = 0.0525 chi – 1.467 (1) in this way, the peptides lipophilicity can be compared to their octanol/water lipophilicity. measurements of membrane binding using immobilized artificial membrane (iam) chromatography for the measurements of the peptides interactions with phospholipids, the gradient retention times of the peptides have been measured using iam.pc.dd2 100 x 4.6 mm column with 10 µm diameter and 300 å pore size particles. the starting mobile phase was 50 mm ammonium acetate with the ph adjusted to 7.4. the acetonitrile gradient was applied from 0 to 90 % from 0 to 4.75 min and kept at 90 % until 5.25 min. from 5.25 to 5.5 min the acetonitrile concentration was dropped to 0. the mobile phase flow rate was 1.5 ml/min and the run time was 6 min. the retention times were standardized using the iam calibration mixture as described previously [33]. the natural state of the phosphatidyl choline headgroup on the iam stationary phase was tested using the six system suitability test compounds listed in table 1. the chi iam values were converted to the octanol/water scale using equation 2 [33]. log kiam = 0.29*exp(0.046 chi iam+0.42) + 0.70 (2) the log kiam values express the membrane partition comparable to the octanol/water lipophilicity. table 1. the iam system suitability test compounds and their expected chi iam values [14]. compound acid/base typical retention time (min) chi iam carbamazepine neutral 2.98 22.3 colchicin neutral 3.31 28.0 warfarin acidic 2.62 16.4 indomethacin acidic 3.15 25.3 nicardipine basic 4.38 46.1 propranolol basic 4.72 51.9 protein binding measurements using biomimetic protein stationary phases the peptides interactions with hsa and agp have been measured using commercially available chemically bonded hsa (chiralpak-hsa) and agp (chiralpak-agp) hplc columns with the dimensions of 50 x 3 mm with 5 µm particle size obtained from hichrom ltd, reading, uk. the mobile phase was 50 mm ammonium acetate with the ph adjusted to 7.4. an iso-propanol gradient (hplc grade, rathburn chemicals, walkerburn, uk) was used from 0 to 35 % from 0 to 3.5 min with a flow rate of 1.5 ml/min. the 35 % iso-propanol mobile phase was run till 4.5 min and then back to 0 % within 0.2 min. the run time was 6 min to allow re-equilibrate of the protein phase with the buffer. the retention times were standardized using the calibration set of compounds as described previously. using the slope and intercept values from the calibration line the logarithmic retention times were converted to log k values that can be converted to % binding values (% hsa and % agp) using equation 3 [34]. klara valko et al. admet & dmpk 6(2) (2018) 162-175 166 log log 101 10 % bound 1 10 k k    (3) compounds studied the linear peptides were obtained from genervon biopharmaceuticals llc (california, usa). these peptides were discovered as endogenous motor neuron trophic peptides that have a great potential to cure amyotrophic lateral sclerosis (als). the peptide with 6 amino acids (gm6) has already been tested in animal models of als and in phase ii human clinical trials. the analogues are marked as linear peptides (lp) with a serial number. the bicyclic peptides were obtained from bicycle tx, cambridge, uk and are marked as bcp with a serial number. these are in the early stages of the drug discovery process. data analysis the calculation of the biomimetic binding properties from the measured retention time values and estimating the in vivo distribution characteristics and cellular concentration [35] are based on the published models [36] and have been carried out using the bio-mimetic chromatography ltd propriety excel macros. the statistical analysis was carried out using jmp 13 (sas institute, usa) and the visualizations have been made by stardrop software (optibrium ltd. cambridge, uk). results and discussion the chromatographic peaks were symmetrical and narrow on the c18 and iam stationary phases. a few examples are shown in figure 1. however, larger peptides (containing more than 10 amino acids) showed wide peaks on the hsa column with a narrow peak at the dead time showing a size exclusion effect (figure 2). size exclusion occurs when large molecules cannot get into the pores of the stationary phases and elute with short retention times. it is worth mentioning that the gemini nt c-18 column particle pore size is 110 å and the iam pc.dd2 column pore size is 300 å and that would not produce a size exclusion effect for peptides below 3000 da molecular weight. the good peak shape also suggests single interaction points between the peptide and the stationary phase on the c-18 and on the iam stationary phases. the broad peak observed for lp10 indicates multiple binding interactions with a slow kinetics between the peptide and the protein. the bicyclic peptides also showed broad peaks both for compounds with shorter and longer retention times. the peak shape of the linear peptides was narrow and symmetrical up to 21 amino acids. table 2 shows the measured data for 10 linear and 10 bicyclic peptides. the lipophilicity difference at acidic, neutral and basic phs reveal the overall charges on the molecules at physiological phs and whether the peptides are more positively or negatively charged. it has been found that the ph dependence of the gradient retention times shows a similar profile to the ph dependence of the octanol/water distribution coefficients (log d) [37]. if the lipophilicity increased at a basic ph relative to the neutral ph lipophilicity, it would suggest that the peptide has more positive than negative charge. if the lipophilicity increases at an acidic ph, the peptide has more negative charge. it can be seen that the linear peptides are positively charged at physiological ph, while bcp1, bcp2 and bcp10 have more negative charge at physiological phs. when plotting the chi log d at ph 7.4 as a function of the molecular weight, we can see some trends for the linear peptides as shown in figure 3. admet & dmpk 6(2) (2018) 162-175 biomimetic hplc in the estimation of physchem properties of peptides doi: 10.5599/admet.544 167 table 2. the measured protein binding (%hsa, log k hsa, %agp, log k agp), phospholipid binding (chi iam, log k iam) and lipophilicity at three phs (chi log d 2.6, chi log d 7.4, chi log d 10.5) and phospholipid binding of the investigated linear (lp) and bicyclic peptides (bcp). the data of the known macrocyclic drug, cyclosporine is marked as cp. peptide % bound hsa log k hsa % bound agp log k agp chi iam log k iam chi log d 2.6 chi log d 7.4 chi log d 10.5 lp1 7.85 -1.07 3.70 -1.42 5.47 1.27 -1.94 -0.54 -0.50 lp2 13.58 -0.81 6.85 -1.14 15.32 1.59 -0.75 -0.26 -0.23 lp3 13.15 -0.82 6.82 -1.14 17.46 1.69 -0.71 -0.14 -0.11 lp4 24.87 -0.49 61.24 0.19 43.04 3.90 -3.39 -0.14 0.22 lp5 20.73 -0.59 37.97 -0.22 42.25 3.78 -3.37 -0.17 0.19 lp6 77.54 0.52 49.58 -0.02 25.82 2.15 0.87 1.01 1.05 lp7 52.66 0.04 69.83 0.35 43.29 3.93 -0.15 0.77 1.07 lp8 95.75 1.26 74.92 0.46 37.47 3.17 0.18 0.83 0.82 lp9 98.97 1.69 89.84 0.91 43.80 4.01 1.18 1.59 0.61 lp10 99.07 1.71 87.83 0.82 37.66 3.20 1.17 1.60 0.61 bcp1 71.25 0.38 63.54 0.23 16.51 1.64 0.98 0.59 0.28 bcp2 88.58 0.85 1.31 -1.88 17.84 1.70 1.11 0.75 0.36 bcp3 19.03 -0.63 1.30 -1.89 16.10 1.63 0.24 0.54 0.25 bcp4 18.24 -0.66 35.68 -0.26 32.48 2.67 0.02 0.89 0.38 bcp5 92.92 1.06 73.34 0.42 39.98 3.48 0.89 1.79 1.54 bcp6 26.00 -0.46 2.69 -1.56 21.77 1.90 0.75 0.84 0.56 bcp7 9.24 -1.00 1.58 -1.80 18.08 1.71 0.57 0.65 0.36 bcp8 94.41 1.16 83.76 0.69 44.90 4.18 1.02 1.79 1.70 bcp9 82.29 0.64 32.81 -0.32 25.91 2.15 1.00 1.20 0.93 bcp10 96.28 1.31 49.46 -0.02 21.50 1.89 2.27 1.23 0.92 cp 91.00 0.97 60.9 0.18 51.33 5.38 5.31 figure 1. chromatograms of lp2 and lp8 obtained on gemini nx c-18 and iam.pc.dd2 columns at ph 7.4. a) lp2 c-18 column, b) lp8 on c-18 column, c) lp2 on iam.pc.dd2 column, d) lp8 on iam.pc.dd2 column klara valko et al. admet & dmpk 6(2) (2018) 162-175 168 figure 2. the chromatogram obtained on chiralpak-hsa column for lp4 (a) and lp10 (b). chromatographic conditions as described in the experimental section. figure 3. the chi log d7.4 values of the investigated peptides as a function of their molecular weight. figure 4. the chi iam values of the investigated peptides as a function of molecular weight. admet & dmpk 6(2) (2018) 162-175 biomimetic hplc in the estimation of physchem properties of peptides doi: 10.5599/admet.544 169 figure 5. the human serum albumin binding (log khsa) of the investigated peptides as a function of molecular weight. from lp1 to lp10 the lipophilicity increases with size, while lp6 is slightly more lipophilic. with the linear peptides (lp8 and lp10) the lipophilicity increased but not proportionally with their larger molecular weight. this is in agreement with the finding that peptides with more than 16 amino acids start to fold and have secondary structure that may hinder the effect of certain amino acids on their lipophilicity [25]. similar trends have been observed on the iam.pc.dd2 stationary phase. the interactions with the phospholipid increased with the increasing number of amino acids as shown in figure 4. the trend breaks for larger amino acids. interestingly, the candidate peptide drug lp4 has stronger interactions with the phospholipids than the trend would show with the molecular weight due to its positive charge having two arginine in its sequence. figure 5 shows a similar plot for the human serum albumin binding in the function of molecular weight. for the linear peptides, as the amino acid number increase, the albumin binding increases up to lp8. although the albumin binding is stronger for lp8 and lp10, it does not increase proportionally with the molecular weight. some of the bicyclic peptides showed much weaker albumin binding, probably because the cyclization hindered several amino acids’ interactions with the protein. almost all of the bicyclic peptides showed wide peaks on the protein phases indicating multiple equilibria and sluggish kinetics. based on the measured biomimetic properties we have estimated the in vivo distribution behaviour of the investigated compounds. the model equations with the source references are listed in table 3. table 4 contains the calculated total plasma protein binding (log k ppb), brain to plasma total concentration ratio (log k bb, brain to plasma ratio), volume of distribution (log vdss and vdss l/kg), unbound volume of distribution (log vdu, vdu), drug efficiency (demax%) and estimated cell partition (kpcell) data. the brain to plasma total concentration ratio has been calculated using the plasma protein binding (%ppb) and brain tissue binding (%btb) models shown in table 3 and described in [35]. from the %ppb and %btb we can calculate the unbound fraction of the peptides (fu) in plasma and brain tissue using the equations listed in table 3. according to the free drug hypothesis, the unbound concentration in plasma (cu,plasma) and brain (cu, brain ) should be equal. the unbound concentrations can be expressed as fu klara valko et al. admet & dmpk 6(2) (2018) 162-175 170 times ctotal which leads to equation 4. table 3. the model equations used to derive estimated in vivo distribution (log vdss, log vdu), drug efficiency (demax), brain tissue binding (log kbtb), fraction unbound in brain and plasma (fu btb and fu ppb), brain to plasma ratio (kbb) and cell partition log kpcell). log kiam [33] = 0.29*e (0.026chi(iam)+0.42 ) + 0.7 log kiam [33] =0.046*chi(iam) + 0.42 log khsa [33] = e log k(hsa) log khsa [34] = log (%hsabound/(101%hsa bound)) estimated log vdss [33] = 0.44*log kiam -0.22*log khsa 0.62 estimated log vdu [38] = 0.23*log khsa +0.43*log kiam 0.72 demax [39] = 100/vdu log kbtb [36] =1.29*log kiam+1.03*log khsa 2.37 log kppb [36] =0.98∗log 𝑘𝐻𝑆𝐴+0.19∗log 𝑘𝐴𝐺𝑃+0.031∗chi log d7.4−0.20 %btb [36] =100*10^log kbtb/(1+10^log kbtb) %ppb [36] =100*10^log kppb/(1+10^log kppb) fu btb and ppb [36] =(100-%btb)/100 and (100-%ppb)/100 kbb [36] = fuppb/fubtb log kpcell [35] =1.1*log kiam 1.9 fu, plasma x ct, plasma = fu, brain x ct, brain (4) expressing the ratio of the total concentrations from equation 4, the equation for kbb can be obtained and is listed in table 3. drugs that are able to go to the central nervous system (cns) have significantly higher kbb values than drugs that do not partition into the cns. when kbb is larger than 1 it means that the total concentration of the compound in the brain is higher than in the plasma. therefore it is very likely that the compounds partition into the brain tissue (unless active transporters efflux the compound back to the plasma). as the peptides easily decompose in plasma or blood under in vitro experimental conditions, it is very useful to estimate the peptides in vivo distribution behaviour considering them as intact molecules. the data in table 4 reveals that there is a wide range of estimated total plasma protein binding in spite of the polar character of the peptides. the brain to plasma distribution ratio is generally low but some of the derivatives especially those that have positively charged amino acids, particularly with an arginine in the sequence, bind more strongly to phospholipids than to albumins thus increasing the expected brain to plasma total concentration ratio. the linear peptides are designed as candidate therapeutics for amyotrophic lateral sclerosis (als). therefore brain penetration and cellular penetration are desired for these peptides. we can see from the measured data that lp4 (gm6 patented by genervon) is predicted to have a good brain to plasma ratio. the volume of distribution is also quite large which means that it partitions into tissues extensively. peptides that are more lipophilic at high ph (chi log d at ph 10.5 is greater than chi log d at ph 7.4) are predicted to have good cell penetration and the brain to plasma ratio is predicted to be greater than 1. this is because the phospholipid binding of positively charged peptides is stronger than that of the neutral molecules. admet & dmpk 6(2) (2018) 162-175 biomimetic hplc in the estimation of physchem properties of peptides doi: 10.5599/admet.544 171 table 4. estimated in vivo distribution and tissue binding properties of the investigated peptides. peptide % bound ppb log k ppb log k bb brain to plasma log vd vd l/kg log vdu vdu % demax kpcell lp1 6.85 -1.14 -0.03 0.93 -0.18 0.66 -0.10 0.80 124.83 0.07 lp2 13.36 -0.82 -0.05 0.88 -0.06 0.88 0.07 1.17 85.62 0.22 lp3 13.03 -0.83 -0.05 0.89 -0.01 0.97 0.11 1.27 78.44 0.28 lp4 36.55 -0.25 0.23 1.70 0.92 8.29 1.10 12.49 8.00 5.49 lp5 27.63 -0.42 0.20 1.58 0.88 7.62 1.03 10.82 9.25 5.01 lp6 85.20 0.73 -0.39 0.40 -0.08 0.82 0.59 3.89 25.70 0.74 lp7 68.47 0.32 0.34 2.21 0.84 6.96 1.21 16.23 6.16 5.66 lp8 98.17 1.54 -0.04 0.92 -0.04 0.91 1.46 28.62 3.49 2.87 lp9 99.99 2.00 -1.49 0.03 -0.08 0.82 2.25 176.76 0.57 6.00 lp10 99.99 2.00 -1.83 0.01 -0.47 0.34 1.92 84.13 1.19 2.93 bcp1 81.77 0.63 -0.25 0.25 -0.26 0.55 0.32 2.10 47.56 0.25 bcp2 84.04 0.70 0.19 0.36 -0.43 0.37 0.55 3.56 28.06 0.29 bcp3 14.84 -0.76 -0.01 0.88 -0.06 0.87 0.10 1.26 79.24 0.24 bcp4 25.70 -0.47 -0.03 0.94 0.40 2.51 0.55 3.51 28.47 1.60 bcp5 96.85 1.37 0.52 1.40 0.23 1.72 1.44 27.51 3.64 3.85 bcp6 23.29 -0.52 0.00 0.84 0.04 1.09 0.24 1.75 57.17 0.46 bcp7 7.46 -1.10 -0.02 0.94 0.01 1.03 0.10 1.26 79.11 0.30 bcp8 97.98 1.51 0.63 2.17 0.48 3.03 1.81 64.42 1.55 6.82 bcp9 87.21 0.80 0.11 0.43 -0.13 0.74 0.64 4.40 22.72 0.75 bcp10 97.97 1.51 -0.38 0.15 -0.64 0.23 0.94 8.77 11.40 0.45 cp 100 2.00 0.78 5.63 1.13 13.39 2.20 158 0.68 14.43 it is well known that cell penetration of peptides is usually impaired, as they are very hydrophilic and cannot go through the cell membrane with the same mechanism as small molecules [8]. however, endocytosis or micelle formation with phospholipids are possible cell penetration mechanisms which require strong interactions between the peptides and the phospholipid membrane. the model for cell partition has been investigated using small molecule partition into hela cells [35] and it was found that the iam partition correlated better with intracellular concentration than other lipophilicity values. therefore, the cell partition (kp cell) has been estimated using the correlation with the iam retention. although the distribution models have been developed using the properties of small molecule marketed drugs based on their clinical data, it was supposed that the same models could be applied to peptides. diaz-eufracio et al [40] have compared the physicochemical property space of marketed peptide therapeutics with other marketed drugs and they found that they occupy similar property space. this study suggests that physicochemical properties of the peptides are not significantly different from small molecule marketed drugs and the drug-like properties can be described by the same physicochemical properties. they used calculated physicochemical properties. in this study, the measured physicochemical and biomimetic properties are compared with marketed drugs [36] and macrolides [41]. the marketed drugs were divided into inhaled and cns group. a cyclic peptide natural product drug cyclosporine (marked as cp) was also included in the analysis. figure 6a shows the lipophilicity distribution and figure 6b shows the iam binding distribution of the peptides and the marketed drugs. it can be seen that the peptides are less lipophilic but their membrane partition (iam binding) is within the range of other drug molecules. klara valko et al. admet & dmpk 6(2) (2018) 162-175 172 a) b) figure 6. the lipophilicity (chi log d7.4) (a) and membrane partition (chi iam) (b), of the peptides in comparison to macrocycle (mc), small molecule drugs (drug), inhaled drugs and drugs that can go to the central nervous system (cns). cp stands for the natural cyclic peptide drug cyclosporin. the estimated brain to blood ratio (log k bb) and cell partition (log kpcell) have been also compared based on the measured hplc based biomimetic properties for the peptides, for the known drugs and for the macrolides as shown in figure 7. a) b) figure 7. distribution of the estimated brain to blood ratio (log k bb) (a) and cell partition (kpcell) (b), for the peptides in comparison to macrocycle (mc), small molecule drugs (drug), inhaled drugs and drugs that can go to the central nervous system (cns). cp stands for the natural cyclic peptide drug cyclosporin. (the macrocycles (mc) did not have all the data for estimating the log k bb therefore they are not shown in (a) the interrelationship between the measured lipophilicity and biomimetic properties of the peptides have been investigated and a comparison made to the marketed drugs. in every respect, very similar relationships were observed. figure 8 shows a good trend between the agp and iam binding of the compounds, as both bindings are influenced by the presence of the positive charge and general lipophilicity of the compounds. both the linear and the cyclic peptides follow the same trend as the marketed drugs and macrolides. one of the investigated compounds, genervon’s gm6 has been evaluated in vivo in a mouse model and also in humans. it has shown an effect in the sod1 mouse model for als and in als patients [42] . as predicted, the peptide has been detected in the cells, especially in the motoneuron cells and was strongly associated with the cell membrane as confirmed using immune imaging technology. gm6 (lp4) has also been detected in the cerebrospinal fluid in mice after 4 hours. these preliminary results encourage further studies of new modalities, such as peptides, cyclic peptides and macrocycles using biomimetic chromatographic methods and models. admet & dmpk 6(2) (2018) 162-175 biomimetic hplc in the estimation of physchem properties of peptides doi: 10.5599/admet.544 173 figure 8. the trend between the alpha-1-acid glycoprotein binding (log k agp) and phospholipid binding (log k iam) for peptides, marketed drugs and macrolides. (colour code is the same as in figure 6 and 7. conclusions the hplc based physicochemical and biomimetic property measurements have been carried out on linear and bicyclic peptides. the measured properties, such as lipophilicity expressed as chi log d, phospholipid binding expressed as chi iam (log k iam) and protein binding (%hsa and %agp) of selected linear and bicyclic peptides have been compared to the similar measured properties of marketed drugs. the lipophilicity, phospholipid binding and protein binding have been compared to the molecular weight of the peptides. it was found that larger molecular weight peptides had greater lipophilicity, phospholipid and protein binding; however, the trend was broken for bicyclic peptides and linear peptides above 16 amino acids. it suggests that the effect of the amino acids on the interactions with lipids and proteins are shielded for larger or cyclic peptides. the in vivo distribution models developed for small molecules have been applied for the first time to peptides. the estimation of good cell and brain penetration for lp4 (gm6), a linear peptide with 6 amino acids, has been validated in in vitro assays. biomimetic hplc measurements can, therefore, contribute to the design of peptide therapeutics with good drug-like properties. references [1] a.a. kaspar, j.m. reichert. future directions for peptide therapeutics development. drug discov. today 18 (2013) 807–817. [2] k. fosgerau, t. hoffmann. peptide therapeutics: current status and future directions. drug discov. today 20 (2015) 122–128. [3] c.a. lipinski, f. lombardo, b.w. dominy, p.j. feeney. experimental and computational approaches to estimate solubility and permeability in drug discovery and develop ment settings. adv. drug deliv. rev. 23 (1997) 3–25. [4] l.t. nguyen, j.k. chau, n.a. perry, l. de boer, s.a.j. zaat, h.j. vogel. serum stabilities of short tryptophanand arginine-rich antimicrobial peptide analogs. plos one 5 (2010) 1–8. [5] j.r. north, s. takenaka, a. rozek, a. kielczewska, s. opal, l.a. morici, b.b. finlay, c.j. schaber, r. straube, o. donini. a novel approach for emerging and antibiotic resistant infections: innate defense regulators as an agnostic therapy. j. biotechnol. 226 (2016) 24–34. [6] z. qian, t. liu, y.y. liu, r. briesewitz, a.m. barrios, s.m. jhiang, d. pei. efficient delivery of cyclic peptides into mammalian cells with short sequence motifs. acs chem. biol. 8 (2013) 423–31. klara valko et al. admet & dmpk 6(2) (2018) 162-175 174 [7] v. sebbage. cell-penetrating peptides and their therapeutic applications. biosci. horizons, 2 (2009) 64–72. [8] r. trehin, h.p. merkle. chances and pitfalls of cell penetrating peptides for cellular drug delivery. eur. j. pharm. biopharm. 58 (2004) 209–223. [9] j.d. ramsey, n.h. flynn. cell-penetrating peptides transport therapeutics into cells. pharmacol. ther. 154 (2015) 78–86. [10] f. wang, y wang, x. zhang, w. zhang, s. guo, f. jin. recent progress of cell-penetrating peptides as new carriers for intracellular cargo delivery. j. control. release 174 (2014) 78–86. [11] a. gräslund, f. madani, s. lindberg, ü. langel, s. futaki. mechanisms of cellular uptake of cellpenetrating peptides. j. biophys., 2011 (2011) 414729. [12] n. schmidt, a. mishra, g.h. lai, g.c.l. wong. arginine-rich cell-penetrating peptides. febs lett. 584 (2010) 1806–1813. [13] k. valko, c. m. du, c. bevan, d. reynolds, m.h. abraham. rapid-gradient hplc method for measuring drug interactions with immobilised artificial membrane: comparison with other lipophilicity measures. j. pharm. sci. 89 (2000) 1085–1095. [14] k.l. valko, s.p. teague, c. pidgeon. in vitro membrane binding and protein binding (iam mb/pb technology) to estimate in vivo distribution: applications in early drug discovery. admet dmpk 5 (2017) 14-38. [15] s.h. joo. cyclic peptides as therapeutic agents and biochemical tools. biomol. ther. 20 (2012) 19–26. [16] p. thansandote, r.m. harris, h.l. dexter, g.l. simpson, s. pal, r.j. upton, k. valko. improving the passive permeability of macrocyclic peptides: balancing permeability with other physicochemical properties. bioorganic med. chem. 23 (2015) 322–327. [17] j.g. beck, j. chatterjee, b. laufer, m.u. kiran, a.o. frank, s. neubauer, o. ovadia, s. greenberg, c. gilon, a. hoffman, h. kessler. intestinal permeability of cyclic peptides: common key backbone motifs identified. j. am. chem. soc. 134 (2012) 12125–12133. [18] f. giordanetto, j. kihlberg. macrocyclic drugs and clinical candidates: what can medicinal chemists learn from their properties?. j. med. chem. 57 (2014) 278–295. [19] n.c. tan, p. yu, y.u. kwon, t. kodadek. high-throughput evaluation of relative cell permeability between peptoids and peptides. bioorganic med. chem. 16 (2008) 5853–5861. [20] b. testa, p. crivori, m. reist, p.a. carrupt. the influence of lipophilicity on the pharmacokinetic behavior of drugs: concepts and examples. perspect. drug discov. des. 19 (2000) 179–211. [21] r. gulaboski, f. scholz. lipophilicity of peptide anions: an experimental data set for lipophilicity calculations. j. phys. chem. b (2003) 5650–5657. [22] a. visconti, g. ermondi, g. caron, r. esposito. prediction and interpretation of the lipophilicity of small peptides. j. comput. aided. mol. des. 29 (2015) 361–370. [23] g. ermondi, f. catalano, m. vallaro, i. ermondi, m.p. leal, l. rinaldi, s. visentin, g. caron. lipophilicity of amyloid β-peptide 12-28 and 25-35 to unravel their ability to promote hydrophobic and electrostatic interactions. int. j. pharm. 495 (2015) 179–185. [24] j.l. meek. prediction of peptide retention times in high-pressure liquid chromatography on the basis of amino acid composition. proc. natl. acad. sci. 77 (1980) 1632–1636. [25] k.m. biswas, d.r. devido, j.g. dorsey. evaluation of methods for measuring amino acid hydrophobicities and interactions. j. chromatogr. a 1000 (2003) 637–655. [26] c.t. mant, r.s. hodges. reversed-phase liquid chromatography as a tool in the determination of the hydrophilicity/hydrophobicity of amino acid side-chains at a ligand-receptor interface in the presence of different aqueous environments: ii. effect of varying peptide ligand hydro. j. chromatogr. a 972 (2002) 61–75. [27] c. vraka, s. mijailović, v. fröhlich, m. mitterhauser. expanding logp: present possibilities. nucl. med. biol. 58 (2013) 20–32. admet & dmpk 6(2) (2018) 162-175 biomimetic hplc in the estimation of physchem properties of peptides doi: 10.5599/admet.544 175 [28] k.m. biswas, d.r. devido, j.g. dorsey. evaluation of methods for measuring amino acids hydrophobicities and interactions. j. chromatogr. a 1000 (2003) 637-655. [29] d. gussakovsky, h. neustaeter, v. spicer, o. krokhine. peptide retention time prediction for immobilized artificial k.m. biswas, d.r. devido, j.g. dorsey. evaluation of methods for measuring amino acids hydrophobicities and interactions. j. chromatogr. a 1000 (2003) 637-50655membrane phosphatydilcholine stationary phase: method development and preliminary observations. admet dmpk (2018) doi: http://dx.doi.org/10.5599/admet.520. [30] k. valko. lipophilicity and biomimetic properties measured by hplc to support drug discovery. j. pharm. biomed. anal. 130 (2016) 35–54. [31] f. tsopelas, c. giaginis, a. tsantili-kakoulidou. lipophilicity and biomimetic properties to support drug discovery. expert opinion on drug discovery 12 (2017) 885-896. [32] k. valko, c. du, c. bevan, d. reynolds, m. abraham. rapid method for the estimation of octanol / water partition coefficient (log poct) from gradient rp-hplc retention and a hydrogen bond acidity term (sigma alpha2h). curr. med. chem. 8 (2001) 1137–1146. [33] f. hollosy, k. valkó, a. hersey, s. nunhuck, g. kéri, c. bevan. estimation of volume of distribution in humans from hplc measurements of human serum albumin binding and immobilized artificial membrane partitioning. j. med. chem. 49 (2006) 6958–6971. [34] k. valko, s. nunhuck, c. bevan, m.h. abraham, d.p. reynolds. fast gradient hplc method to determine compounds binding to human serum albumin. relationships with octanol/water and immobilized artificial membrane lipophilicity. j. pharm. sci. 92 (2003) 2236-2248. [35] l.j. gordon. m. allen, p. artursson, m.m. hann, b.j. leavens, a. mateus, s. readshaw, k. valko, g.j. wayne, a. west. direct measurement of intracellular compound concentration by rapidfire mass spectrometry offers insights into cell permeability. j. biomol. screen. 21 (2016) 156-164. [36] k. valko, physicochemical and biomimetic properties in drug discovery chromatographic techniques for lead optimization. wiley, hoboken, nj (2014). [37] i. canals, k. valkó, e. bosch, a.p. hill, m. rosés. retention of ionizable compounds on hplc. 8. influence of mobile-phase ph change on the chromatographic retention of acids and bases during gradient elution. anal. chem. 73 (2001) 4937–4945. [38] k.l. valkó, s.b. nunhuck, a.p. hill. estimating unbound volume of distribution and tissue binding by in vitro hplc-based human serum albumin and immobilized artificial artificial membrane-binding measurements. j. pharm. sci. 100 (2011) 849–862. [39] k. valko, e. chiarparin, s. nunhuck, d. montanari. in vitro measurement of drug efficiency index to aid early lead optimization. j. pharm. sci. 101 (2012) 4155-4169. [40] b. i. díaz-eufracio, o. palomino-hernández, r. a. houghten, j.l. medina-franco. exploring the chemical space of peptides for drug discovery: a focus on linear and cyclic penta-peptides. mol. divers. (2018) doi: 10.1007/s11030-018-9812-9. [41] v. stepanić, d. žiher, v. gabelica-marković, d. jelić, s. nunhuck, k. valko, s. koštrun.. physicochemical profile of macrolides and their comparison with small molecules. eur. j. med. chem. 47 (2012) 462–472. [42] k. valko, m. kindy, j. evans, d. ko. in vitro and in vivo characterization of gm6 an endogenous regulator peptide drug candidate for amyotrophic lateral sclerosis (als). admet dmpk (2018) doi: http://dx.doi.org/10.5599/admet.547. ©2018 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ biophysical methods in early drug discovery doi: 10.5599/admet.733 222 admet & dmpk 7(4) (2019) 222-241; doi: http://dx.doi.org/10.5599/admet.733 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index review biophysical methods in early drug discovery geoffrey holdgate* 1 , kevin embrey 2 , alexander milbradt 2 and gareth davies 1 1 hit discovery, discovery sciences, r&d, astrazeneca, alderley park, uk 2 structure, biophysics and fblg, discovery sciences, r&d, astrazeneca, cambridge, uk *corresponding author: e-mail:geoff.holdgate@astrazeneca.com; tel.: +44 (0)1625 234218 received: september 30, 2019; revised: november 19, 2019; published: december 11, 2019 abstract biophysical methods such as mass spectrometry, surface plasmon resonance, nuclear magnetic resonance, and both differential scanning isothermal titration calorimetry are now well established as key components of the early drug discovery process. these approaches are used successfully for a range of activities, including assay development, primary screening, hit confirmation and detailed mechanistic characterisation of compound binding. matching the speed, sensitivity and information content of the various techniques to the generation of critical data and information required at each phase of the drug discovery process has been key. this review describes the framework by which these methods have been applied in the drug discovery process and provides case studies to exemplify the impact. keywords affinity selection mass spectrometry; surface plasmon resonance; nuclear magnetic resonance; differential scanning fluorimetry; isothermal titration calorimetry; microscale thermophoresis introduction biophysical methods, which can encompass a wide range of techniques focussed on measuring the structure, properties, dynamics or function of biomolecules, have been increasingly employed in the drug discovery process since their first introduction in the early 1990s. it was during this time that commercial instruments such as the first isothermal titration calorimeter (itc) [1] and the first surface plasmon resonance (spr) [2] instruments became available and their use exemplified using biochemical systems. alongside this, existing biophysical approaches were used in novel ways to identify and characterise protein-ligand interactions, for example the first report of the use of affinity selection, coupled to detection by mass spectrometry, for the identification of molecules binding to a macromolecule [3], and the use of nuclear magnetic resonance (nmr) to identify fragments that could subsequently be optimised and linked to form more potent compounds [4]. the development of these biophysical methods coincided with the advent of high-throughput screening (hts), leading from natural product screening of a few hundred compounds each week in the late 1980s, through to hts hits being responsible for starting matter for almost half of drug companies’ portfolios in the mid-1990s [5,6]. this allowed the valuable combination of biophysics with hts to contribute to the establishment of high quality, high-throughput assays through the characterisation of protein and tool ligands, as well as the evaluation of hts output through orthogonal application of biophysical methods to screen for true target engagement [7]. more recently, biophysical http://dx.doi.org/10.5599/admet.733 http://www.pub.iapchem.org/ojs/index.php/admet/index admet & dmpk 7(4) (2019) 222-241 biophysical methods in early drug discovery doi: 10.5599/admet.733 223 methods have matured, throughput has increased, and sensitivity improved, such that some of these methods can also now be employed in primary screening, not only for fragments, which typically screen lower numbers of low molecular weight compounds [8], but for screening libraries comprising many thousands of compounds [9]. the high-fidelity nature of many biophysical methods, coupled to the specific information content they can access, has also meant that they are used increasingly to characterise the mode of action of hits and leads. some methods provide access to kinetic data [10], whilst others enable thermodynamic characterisation of ligand binding [11]. others may allow structural insights into binding mode and binding site [12,13]. our understanding of the strengths and weaknesses of the methods, as well as a better appreciation of the impact that results from appropriate positioning within the drug discovery process has contributed significantly to the increased use of biophysical methods observed today. this review will provide an overview of the benefit that can be gained from the incorporation of biophysical methods within areas of the drug discovery process and will provide case studies to exemplify their impact. assay development the capability to design, build and implement assays that are specific, robust and sensitive enough to identify and characterise potential new drug molecules is fundamental to drug discovery. in developing a new assay there are several factors that must be considered. these include: the nature of the reagents, such as their identity, purity, concentration, functionality and stability; the features of the detection system, such as the sensitivity, dynamic range, potential for interference and reproducibility; the analysis of the data; and the subsequent statistical examination of assay performance. the application of biophysical methods early during assay development can help to understand some of these features and to ensure that high-fidelity, fit for purpose assays are developed. of course, this extra resource needs to be considered during project planning, but the benefits of applying this early outweighs the resource that would need to be applied to rescue projects which have been misled or have failed due to spurious activity from poorly characterised reagents or assays. reagent quality control early application of biophysical methods often focusses on understanding the quality of the reagents available, both in terms of the suitability of the target protein and of the behaviour of the known tool compounds. for a protein to be deemed suitable to be used in subsequent drug discovery assays it should fit criteria associated with the group of characteristics highlighted in the paragraph above. the protein should be the right protein, so its identity should be confirmed, otherwise invalid and misleading results may be obtained. the purity of the protein is also important as impurities may have similar activities or bind test compounds. the concentration of the protein should be measured so that considerations around concentration dependent effects can be understood. the functionality of the protein should also be investigated. this may require measurements of ligand binding, the catalytic activity (in the case of enzymes) or the ability to carry out other functions or interactions. finally, the stability of the protein should be assessed, both in terms of its thermodynamic and kinetic stability, as well as its ability to withstand certain conditions such as freeze-thawing. it also may be necessary to characterise the binding of other molecules such as substrates, cofactors, binding partners or competing probe molecules, to understand their requirement and concentration range desired and permitted within the assay methodologies under consideration. these attributes of the target protein may be investigated using a variety of approaches and biophysical methods can help to provide confidence in the target protein, by being employed alongside or instead of standard biochemical methods for protein characterisation. table 1 shows several possible methods that may be employed in such characterisation and it highlights the holdgate et al. admet & dmpk 7(4) (2019) 222-241 224 information that is generated to allow effective decision making to ensure that only protein of sufficient quality is used in hit-finding and subsequent optimisation assays. some of these methods may also be used to characterise the binding of ligands to the target protein, to evaluate their purity and concentration. table 1. potential biochemical and biophysical approaches for protein quality control checks. group methods (biochemical and biophysical) example information required for acceptable quality control 1. identity amino acid analysis & sequencing exact, correct sequence identified lc-ms (liquid chromatography-mass spectrometry) correct relative molecular mass (mr) within instrument error peptide mapping to identify post translational modifications (ptms) (eg phosphorylation) number & sites of phosphorylation; extent of phosphorylation 2. purity sds-page (sodium dodecyl sulfate – polyacrylamide electrophoresis) / native page single band on a gel; still a single band at high loading dynamic laser light scattering (dls) monodisperse, mr ± 20 % expected analytical gel filtration defined single gaussian peak for a monomer analytical ultra centrifugation (auc) indicates homogeneity & correct mr 3. concentration ultraviolet (uv) spectrum peak at 280 nm; peak at 205 nm; no peaks above ~ 340 nm; test for light scattering (look into ratio at different wavelengths eg a280/a230); concentration calculated using ε bradford assay linearity with bsa standards 4. functionality functional assay functional activity observed with expected parameters (eg kcat, km, kcat/km) isothermal calorimetry (itc) with known tool ligand: n  15 % of expected; kd within 2-fold of reference value; h within 1 kcal/mol surface plasmon resonance (spr) direct binding assay (dba): kd within 2-fold of reference value; expected theoretical rmax; inhibition in solution assay (isa): [protein] within ±15 % of two different concentration measures (bradford & a280); competition observed between target definition compound (tdc) and tdc in solution functional comparison between protein batches compare kd, h, stoichiometry, km, kcat, kcat/km (usually > 10 6 s -1 m -1 ), ki; single phase kinetics validity of construct compare kd, km, ki, h with full-length protein; compare structure-activity relationship (sar) 5. stability differential scanning calorimetry (dsc) good pre-transition baseline; visible tm (above 37 °c); good post-transition baseline differential scanning fluorimetry (dsf) good pre-transition baseline; visible tm (above 37 °c); good post-transition baseline selwyn’s test overlay of plots of [p] vs [e].t for different combinations of [e] and t where lc-ms is liquid chromatography mass spectrometry, ptm is post translational modification, uv is ultraviolet, mr is relative molecular mass, bsa is bovine serum albumin, ε is the molar extinction coefficient, kcat is turnover number, km is michaelis constant, n is stoichiometry, kd is equilibrium dissociation constant, h is binding enthalpy, dba is direct binding assay, isa is inhibition in solution assay, rmax is maximum response, tdc is target definition compound, sar is structure activity relationship, tm is the the melting temperature, [p] is product concentration, [e] is enzyme concentration, t is time. clearly, it is essential to understand the quality of the protein and its behaviour before significant time is admet & dmpk 7(4) (2019) 222-241 biophysical methods in early drug discovery doi: 10.5599/admet.733 225 spent developing assays and certainly before bulk amounts of protein are made for activities such as highthroughput screening and x-ray crystallography. in this respect it is useful to remember that the target protein goes into every well of the experiment, and so issues with the performance of this reagent has the potential to compromise the entire experiment. this can be contrasted with efforts to ensure individual compound purity and activity. although individual compound integrity is important to ensure that the identification of useful chemical start points from any active hit compound is not missed, it does not have the same propensity to compromise the whole hit identification process, as would be the case for a protein of questionable integrity. protein integrity, and its resulting ability to bind to test compounds which may then modulate the biological function, can be compromised in several ways, some examples of which are illustrated in table 2. table 2. examples of factors that have compromised protein integrity for use in drug discovery projects, resulting observations and actions taken to overcome the issues. protein target quality control issue biophysical methods employed observations actions taken lactate dehydrogenase cofactor present in protein preparation itc, spr tool compounds and added cofactor binding more weakly than expected new purification method established atad2 protein aggregation nmr, itc, tsa protein showing poor spectrum, negative shifts with compounds in tsa, no binding of tool compounds new construct designed acper reduced binding functionality itc low stoichiometry and enthalpy for cofactor binding new batch of protein prepared mapkapk2 no binding to p38 nmr, itc short construct used for nmr did not show binding to p38 and differences in compound affinity observed for long and short constructs in phosphorylation assays longer construct, containing putative site for p38a binding, used for activity and mechanistic assays where atad2 is atpase family aaa domain-containing protein 2, acper is acyl carrier protein enoyl reductase, mapkapk2 is mapk activated protein kinase 2. the application of biophysics to characterise reagent quality may not be a large, resource-intensive effort. often, a single but decisive experiment can be extremely informative, and often critical in understanding the behaviour of the target protein. additionally, since these methods are often relatively generic, requiring little assay development, their use at this early stage is not prohibitive, and is often influential and impactful. even if methods do require additional development time, for example spr assays may take longer to establish than a simple itc experiment, this is usually time well invested, as the methods are frequently used again for hit evaluation, and so the development time is in effect just positioned earlier in the workflow than it otherwise may have been. in the case of spr, having an assay in place to characterise hits post hts is very valuable, and will be discussed further below. assay quality control in addition to being able to characterise reagents, biophysical methods are also an invaluable tool applied to evaluating assays for their ability to identify compounds that engage with the target protein to bring about the desired effect. any biochemical assay has the potential for artefacts to arise due to nondesired mechanisms. these may be specific mechanisms that are unwanted in a drug-like compound, such holdgate et al. admet & dmpk 7(4) (2019) 222-241 226 as reactivity (for example thiol reactivity), redox cycling, colloidal aggregation, heavy metal contamination, protein unfolding, protein denaturation (so called pan assay interference compounds or pains [14]) or biological system or technology interferences, such as coupled enzyme inhibitors, fluorescent compounds or quenchers. understanding the liability of a biochemical assay to these types of compounds helps to understand the potential output from high-throughput screens (hts) utilising that assay approach. at astrazeneca, hts development includes testing a small library of around 1000 compounds with unwanted mechanism of inhibition (the umoi set) and around 7000 compounds (the validation set), which are meant to represent the diversity present in the full screening set of around 2 million compounds. this allows the assay to be assessed in terms of its susceptibility to pains, as well as providing an assessment of reproducibility, likely hit rate and to highlight potential artefacts and propensity for false positives and negatives. biophysical methods such as spr or nmr are employed following this early screening activity to characterise the hits, so that the knowledge arising from an understanding of the reasons behind false positives may be used to further optimise the assay to avoid these types of hits in the full screen. table 3 exemplifies several assays, across different target classes, where biophysical characterisation post validation set testing, influenced the subsequent hit identification strategy or tactics. this workflow therefore allows decisions to be made based on any specific issues that arise due to the nature of the assay, that may be mitigated or avoided by modifying the screening cascade. for example, knowing that heavy metals contaminants or redox cycling compounds may be hits in the biochemical assay, but the compounds are not true binders to the target protein in a high-fidelity biophysical method allows the primary screen to be modified to reduce the liability to such effects. in these examples, the use of metal chelating agents and investigating different reducing agents may lead to changes to the assay protocol. alternatively, post screen triage may involve introducing additional assays that allow compounds functioning by these undesirable mechanisms to be identified and deprioritised. thus, the use of biophysical methods in this way represents a valuable investment to increase the probability that highquality hits will be identified during the primary screen. as can be seen from table 3, in some cases, the impact of the biophysical testing was taking a decision not to run the primary screen at all, but to pursue alternative approaches, such as fragment-based lead generation (fblg). these decisions, although difficult at the time, due to the previously committed investment, ultimately may result in substantial cost savings, firstly from not committing to hts (at an average cost of around $90k) and secondly from not following up spurious hits. figure 1 shows the interplay that is required between the assay methodology, the use of high-quality reagents and the role that biophysics has in helping to characterise these aspects to facilitate the implementation of valid screening assays. primary screening there are few biophysical methods that can be applied to primary screening as usually defined in the high-throughput setting, which often refers to the testing of 1 million or more compounds, and usually these methods are more frequently applied for secondary hit evaluation. the predominate reasons are the amount of protein required and the throughput required to achieve primary screening in a reasonable timeframe. to address these issues compound mixtures are often used to facilitate reductions in reagents and time, since many compounds may be screened from a single well. often, mixtures have no adverse effects on the protein, or on the ability to detect binding, but sometimes problems are experienced if there are compound-compound interactions or if the compound organic load or the compound solvent concentration [dmso] is too high. admet & dmpk 7(4) (2019) 222-241 biophysical methods in early drug discovery doi: 10.5599/admet.733 227 table 3. impact of biophysical evaluation of hits identified during hts development. protein target assay methodology number of compounds tested biophysical methods employed observations actions taken keap1 htrf 180 nmr, spr no genuine hits identified hts was stopped and fblg approach used instead malt1 fluorescence intensity following proteolytic cleavage 60 nmr 17% of hits showed specific binding. 38% showed redox cycling behaviour. 27% were not soluble incorporation of a redox-artefact assay in the cascade reduced the number of redox-active compounds reaching the nmr assay from 38% to 5% errγ fret 180 nmr fret assay suggested that hit rate would be low. nmr suggested that 90% were false positives hts in this format was not run ttbk1 adp-glo 250 spr 69 verified hits, then profiled versus phosphorylated and non-phosphorylated protein large number of reactive compounds identified adp-glo assay was not run acper fluorescence intensity following substrate turnover tool compounds and 630 fragments spr, itc characterisation demonstrated that compounds showing several different mechanisms of inhibition could be found project view on needing a cofactor competitive inhibitor was changed and assays configured to find all mechanisms ltc4s htrf, rapidfire total of 50 selected from actives in one or both assays nmr, spr 77% of the total hits shown to bind and also to displace tool ligand. confirmation rate was 90% for rapidfire hits, 40% for htrf hits rapidfire assay prioritised for full hts apc 4 different assays: 1) chromogenic cleavage 2) peptide cleavage coupled assay 3) peptide cleavage rapidfire 4) fibrin clot assay total of 250 identified from the assays as follows: 1) 90 2) 50 3) 10 4) 100 190 selected for biophysical testing (90 from assays 1-3 and total output from 4 spr, nmr numbers of confirmed hits originating from each assay approach: 1) 8 2) 16 3) 9 4) 13 fibrin clot assay was shown to identify compounds binding at a site distal from the active site fibrin clot assay was selected for hts, based on the ability to identify novel, exo site binders. where keap1 is kelch like ech associated protein 1, malt1 is mucosa-associated lymphoid tissue lymphoma translocation protein 1, errγ is estrogen-related receptor gamma, ttbk1 is tau tubulin kinase 1, acper is acyl carrier protein enoyl reductase, ltc4s is leukotriene c4 synthase and apc is activated protein c. two biophysical methods that are often capable of providing throughputs approaching that of traditional hts are affinity selection mass spectrometry (as-ms) [15] and thermal shift assays (tsa) [16]. these methods utilise mass detection of ligands bound to a target protein and the ligand-induced increase in thermal stability respectively to identify hits. the benefit of these approaches in primary screening is that the assay is generic and can rapidly be optimised for the target of interest. the disadvantage in biophysical primary screening is that identified hits may bind to the target protein but may not have the desired biological effect. an alternative use of biophysical methods in primary screening is in testing a much smaller library of low molecular weight compounds in so-called fragment-based screening [17]. this was first described holdgate et al. admet & dmpk 7(4) (2019) 222-241 228 employing nmr screening to identify fragments that could subsequently be linked and optimised to make higher affinity compounds [4]. more recently, fragment screening has been described by other biophysical methods including both spr [18] and tsa [19]. figure 1. interplay between reagents, assay methodology and biophysical quality control in the development of valid assays. high-throughput screening – affinity selection mass spectrometry (as-ms) the application of mass spectrometry (ms) in the drug discovery process has been well established for many years. it has been applied both in the characterisation of target proteins, where it has been primarily focussed on quality control as well as target identification and validation. for small molecules, the technique has been used for metabolism and pharmacokinetics studies as compound identification. more recently, ms approaches have been developed to detect and characterise protein-small molecule complexes. affinity selection mass spectrometry (as-ms) is now the predominant biophysical method used for primary screening. it was first introduced in the late 1990s [20] and various formats have been described, evaluated and implemented since then [21]. however, the basic premise of all these methods is the detection of compounds that bind to the target protein using mass detection of the bound ligand. this simple binding assay has the benefit of reduced interference, since there is no requirement for functional activity which often entails more complex assays and therefore often increased probability of artefacts arising due to effects on the read-out, rather than from true target engagement. another advantage is that this screening approach can be applied to orphan genomic targets and targets for which functional assays cannot be developed. the major disadvantage, alongside the requirement, mentioned above, for assessing functional activity post-screen, is that these methods are not commercially available and systems, expertise and sometimes even the software solutions required for data analysis have to be developed in house. although there are subtle differences in the way as-ms methods may be employed, the principle of the different approaches is essentially the same. it involves incubating a library of small molecule compound mixtures with the target protein, separating the bound small molecules from non-binders and detecting those binders using mass spectrometry, figure 2. the relatively slow ms detection still limits the throughput of these methods and even with the development of acoustic mist ionisation approaches [22], which have increased the speed of mass spectrometry detection, large compound mixtures are still required with mixture sizes ranging from around admet & dmpk 7(4) (2019) 222-241 biophysical methods in early drug discovery doi: 10.5599/admet.733 229 100 to almost 3000 compound per well. consideration of the composition of these mixtures is important as these mixtures have potential for introducing solubility issues, compound-compound reactivity and instability. smaller mixtures are clearly less disposed to these types of problems but may impose restrictions on the size of the compound library that be screened. whilst this method is not dependent upon protein function, ms has a low tolerance to detergents present in biological buffers. this is primarily due to the propensity for large aggregates to form which may interfere with binding and detection. thus, the application is limited to screening soluble proteins as the detergents required for membrane protein preparations are not compatible. figure 2. schematic representation of the workflow for as-ms primary screening. mixtures of test compound are incubated with the target protein before separation of bound ligands from free ligands by size plate-based exclusion chromatography. bound ligands, which are eluted with the target protein, are subsequently detected by mass spectrometry. even with these limitations, as-ms still represents an important addition to the high-throughput screening toolbox. the sensitivity means that relatively low amounts of target protein are required, of course there is no requirement for labelling of reagents and the process can be automated, all features that contribute to the quality, speed and cost considerations required when deciding upon a primary screening approach. high-throughput screening – thermal shift assay (tsa) the thermal shift assay (tsa) also known as thermofluor or differential scanning fluorimetry has been used for several years to study protein stability [23]. it is a rapid and simple method that allows the melting temperature, tm (temperature at which 50 % of the protein is unfolded), to be determined under different conditions. it has extensively been used to optimise buffer conditions for x-ray crystallographic studies on soluble proteins, such that those conditions yielding the highest stability may be used as start points for crystallisation trials. the addition of compounds that bind to the target protein, thermodynamically stabilise the protein relative to controls in the absence of ligand, and this stabilisation can be used to identify binders from non-binders. the tsa experiment involves incubating the target protein with test compounds in the presence of a dye that binds to hydrophobic regions of the protein. the temperature is then increased uniformly, and the fluorescence of dye monitored with temperature. as the protein unfolds more hydrophobic regions are holdgate et al. admet & dmpk 7(4) (2019) 222-241 230 exposed, there is increased dye binding and the fluorescence intensity increases. compounds that increase the thermal stability of the protein may be identified as those giving an increased tm, resulting from a shift in the unfolding curve to higher temperatures, figure 3(a). 0 10 20 30 40 50 60 70 0 1000 2000 3000 4000 5000 temperature °c r f u 0 20 40 60 80 -100 0 100 200 300 400 temperature °c  r f u / t figure 3. (a) typical unfolding curve in a tsa. squares / solid line show the protein unfolding in the absence of ligand. triangles / dotted line show unfolding in the presence of a compound that stabilises by 5 °c. the red arrow indicates the shift in tm caused by the addition of compound; (b) first derivative of the data in (a). the advantages of tsa as a primary screening method include the simplicity of the approach, the cost effectiveness and the potential to access a wide range of binding affinities. the method requires little assay development, which is in effect limited to adjusting [protein] and [dye] to give a suitable signal. the reagent requirements are protein and dye only, meaning that the costs associated with expensive biochemical reagents are avoided. the method does not require specialised instrumentation and is carried out using standard thermocyclers often used for real-time polymerase chain reactions (rt-pcr). protein quantities can potentially limit the method, since screening 500,000 single wells can require around 0.5 g of a 40 kda target protein. whilst there is no direct correlation between tm and kd [24], due to differences in enthalpic and entropic contributions to binding affinity having a differing effect on tm, it is possible to observe stabilisation conferred by ligands covering a wide range of affinities from mm to very tight-binding. analysis of tsa data typically involves determining the tm for each curve, which can be achieved by fitting an appropriate equation to the data, or for simple 2 state transitions, simply by taking the first derivative of the fluorescence versus temperature data, figure 3 (b). for high-throughput analysis genedata screener® (genedata, basel, switzerland) includes a module that can robustly and efficiently analyse the many temperature curves created, which significantly reduces the analysis time. fragment-based screening fragment-based drug discovery (fbdd) is now a well-established approach with fbdd having delivered 2 marketed medicines (vemurafenib and venetoclax) and around 35 compounds currently in clinical trials [25]. identifying fragment hits is a challenge in traditional biochemical assays, since they are likely to bind weakly to the target protein, and the apparent affinity may be weaker still, due to competition with substrates. biophysical methods are well suited to fragment-based screening [25], as they are sufficiently sensitive to detect weak interactions, and reliable enough to avoid false hits. historically, techniques including nmr, spr and x-ray crystallography were applied to fragment screening, but more recently methods such as microscale thermophoresis (mst) [26] and tsa have also been used. compared to traditional hts, the throughput of fragment screening is generally lower, and the (a) (b) admet & dmpk 7(4) (2019) 222-241 biophysical methods in early drug discovery doi: 10.5599/admet.733 231 biophysical methods employed often require larger quantities of protein. fortunately, this issue is overcome by the ability to screen fewer molecules, but to cover a much larger proportion of chemical space using fragment screening libraries. for example, screening 1 million compounds out of the estimated 10 30 compounds that could potentially be synthesised with 36 heavy atoms (around 500 da) is 10 19 -fold less efficient than screening 1000 compounds from the 10 8 potential compounds with 12 heavy atoms (around 160 da). another way of considering this is to reframe the coverage of chemical space by considering how many compounds need to be screened in order to obtain a sufficient number of hits against the target of interest. hann, et al [27] found that hit rates decreased as the complexity of ligands increased. of course, smaller ligands will generally bind less tightly and so the apparent hit rate depends upon the sensitivity of the detection method. thus, the probability of detecting binding for ligands of different sizes is expected to be low for very small ligands (due to detection sensitivity) high for small ligands and decreases with size for larger ligands (due to the increasing probability of steric clashes between the ligand and protein). this suggests that focusing on fragments with lower heavy atom counts with the most sensitive biophysical screening methods provides the highest probability of success. nmr is well-suited for fragment screening, as it can detect binding for fragments having millimolar kd values. two general approaches may be applied for proton nmr-based screening, which monitor either differences in the spectra of the small molecules or the protein. ligand-based screening methods are often used for medium-sized proteins but work better with larger target proteins. no isotope labelling is needed, and the quantity of protein required is relatively small. another advantage is that when ligand-observed screens are undertaken, knowledge of the chemical shift pattern for each ligand avoids the necessity for deconvolution. disadvantages of direct-detection ligandbased screens are that ligand-based screening does not provide information on the binding site, and often false positive rates may be greater than with protein-observed methods, as it is sometimes difficult to discriminate between promiscuous, non-specific binding due to compound aggregation and site-specific binding. finally, and almost uniquely for biophysical methods, ligand-based nmr screening becomes challenging when binding is too tight. these issues can be overcome by using a reporter or “spy-molecule” in the nmr experiment. however, this requires that such a ligand is available when the assay is developed and has the disadvantage that compounds binding at a non-overlapping site may be missed. probably the most robust fragment screening method is protein-observed nmr, where changes in chemical shift for an isotopically labelled protein are monitored. the advantage of using this method is that not only can hits be detected, but affinities can be determined, and binding sites identified if the protein signals have been assigned. this method is suitable for proteins of around 10 – 50 kda providing that uniformly 15 n-labeled protein can be obtained. a detailed description of the application of different ligand and protein observed nmr methods in fragment-based screening is beyond the scope of this article, but a valuable overview is given by harner et al [28]. since fragments usually have weak binding affinities, they are almost always in the fast exchange regime, and kd values can be calculated from changes in chemical shifts with increasing fragment concentration. however, if fragment binding affinity is higher or fragment optimisation leads to compounds that have improved affinity (for example kds of the order of 10 50 μm) then intermediate exchange of resonances can become a problem, with resonances broadening and disappearing and nmr is no longer useful for kd determinations. at this point, other biophysical techniques such as itc and spr are often used to measure affinity. holdgate et al. admet & dmpk 7(4) (2019) 222-241 232 spr can also be applied to primary fragment screening, and many of the technical practical challenges including working with low-molecular weights compounds with limited solubility and showing low-affinity interactions in high refractive index solvents such as dmso have been addressed through appropriate assay design and control experiments. improvements in instrumentation and data analysis procedures have also helped to position spr alongside nmr in the consideration of methods for delivering fragment screening hits [29]. immobilisation of the target protein in a functional manner is still a key factor, and several different strategies may need to be explored. our experience suggests that use of an avitag tm (avidity llc) [30] with coupling via biotinylation of the tagged protein and capture on a streptavidin chip often produces a suitable surface for fragment screening, and subsequent characterisation. issues with potential fragment binding to the streptavidin are usually overcome by blocking with suitable biotin analogues. other tagging approaches for immobilisation have been used successfully to capture membrane proteins, providing the possibility of using fbdd versus members of this important class of drug targets, such as g-protein coupled receptors (gpcrs) [31]. as mentioned previously, fragment screening using tsa has also proved to be a useful approach, and has been exemplified as a primary fragment screen, with detailed biophysical follow up, for identifying fragment hits with the potential for disrupting protein-protein interactions (ppis) [32]. this method, whilst not being suitable for membrane proteins, has the advantage of not requiring immobilisation or labelling of the protein, and so may be more widely applicable to proteins that may be difficult to work with in nmr or spr. microscale thermophoresis (mst) [33] is a developing technique that has also been applied for fragment screening [34]. mst detects the movement of fluorescent molecules in a microscopic temperature gradient created by focusing an infra-red laser beam on a section of a microliter-volume capillary. binding of ligands typically changes the size, charge, and/or hydration shell of the target protein, producing a change in the thermophoretic movement of the protein. mst requires that the position of one binding partner can be fluorescently monitored, so for screening purposes this is most likely the target protein. whilst this can potentially introduce artefacts, and the low protein requirement the absence of a need for immobilisation can position this method as useful approach if nmr or spr cannot be used. fragment screening is clearly amenable to a range of biophysical methods, due in large part to the reduced numbers of compounds that are routinely tested. at astrazeneca the fragment library consists of a soluble set of around 14000 compounds. within that is contained a core set of 3456 compounds which are routinely used for biophysical screening. this library is further divided into a soluble set used for spr screening of 3072 compounds and a simple set of 1152 compounds, which are often screened by nmr. there are 768 compounds that are common to both the soluble and simple sets. the design and usage principles behind these sets were that all compounds should have a heavy atom count of less than 20, have molecular weight less than 275, a calculated log p of less than 3 and have a predicted aqueous solubility of greater than 100 µm. additionally, the soluble set was designed for spr screening, so contained compounds that were larger (potentially important for the indirect mass detection), more complex, and had previously been checked for issues of aggregation / interaction with the dextran matrix by ‘cleanscreening’ [35]. the overlapping 768 compounds formed a ‘ligandability’ set that were of intermediate complexity, soluble and ’clean’ and could be used to assess ligandability by any appropriate biophysical method, figure 4. using these sets of compounds enables an efficient workflow for both initial ligandability assessment and primary fragment screening by nmr and spr. admet & dmpk 7(4) (2019) 222-241 biophysical methods in early drug discovery doi: 10.5599/admet.733 233 figure 4. description of the astrazeneca core fragment library. hit confirmation to ensure that hits identified via hts are valid before subsequent time and resource are invested in optimisation activities several steps are applied in the workflow shown in figure 5. this involves re-testing actives in the same assay format as the primary screen to confirm that the activity is reproducible. concentration response assays are then employed to measure the concentration of test compound that brings about 50 % of the maximal effect (ec50). a range of counter screens may be used to determine whether potential hit compounds demonstrate artefactual behaviour, if they possess unwanted mechanisms of action or whether they also affect a number of related targets whose modulation should be avoided for selectivity reasons. typically, compounds are then grouped into clusters, representing the active compounds from each interesting chemical series. this allows a small number of representative compounds from each cluster to be tested in a suitable biophysical technique. positioning biophysical testing early in this way allows focus on those compounds that are confirmed by a biophysical method and de-prioritising those that may likely be false positives. in this context, biophysical methods provide a means of selecting compounds based on positive selection criteria such as target engagement, mode of action, and for some methods even functional activity (ms, nmr and even itc can for instance, be used to monitor inhibition of substrate conversion for enzymatic reactions), rather than the negative criteria associated with filtering compounds for unwanted mechanism of action. figure 5. the combination of approaches that are used to triage from primary screening actives to validated hits. holdgate et al. admet & dmpk 7(4) (2019) 222-241 234 as with other application areas of biophysical methods, each has its own strengths and weaknesses in hit confirmation and the choice of method will depend upon the throughput required, the amount of protein available, the affinity range expected and the information content desired, table 4. if possible, the hit confirmation strategy may include several techniques to provide increased confidence and deliver a deeper understanding of the binding interaction. positioning one or more biophysical methods will depend upon what is known about the target protein, what is required in terms of setting up each biophysical method and the information desired. an important consideration is how biophysics might be used in the confirmation of hits resulting from cell-based screening. of course, for isolated protein-based biochemical assays, the same or similar protein constructs may be used for the biophysical triage step, and so the physiological relevance (or lack thereof) of each is at least consistent. this is different for the case of cell screens, where the biophysical approach may be considered less physiologically relevant than the cell assay. biophysical methods still have value in confirming binding and providing additional information, but extra caution needs to be applied in comparing the results and in making decisions about fate of compounds based on similarities or differences. in this situation, there is additional value in the use of tool compounds which may have similar modes of action to hts hits. confidence can be gained for those hts hits that show similar behaviour to the tool compounds during the biophysics confirmation stage. it may also be prudent to explore additional (biophysical) methods that allow interrogation of target engagement in cells such as cetsa (cellular thermal shift assay) [36]. table 4. comparison of the some of the most common methods used for hit confirmation. technique specific requirements protein consumption throughput dynamic range information content nmr 15 n labelling for 2d protein observed nmr high medium mm µm high (binding site) spr suitable immobilisation low high mm pm high (kinetics) itc protein and ligand in identical buffer high low mm nm high (thermodynamics) mst fluorescent labelling of one partner low medium mm pm medium (affinity) mechanistic characterisation biophysical methods are extremely valuable in helping to carry out in depth characterisation of proteinligand binding interactions. they often provide a simpler way of complementing biochemical approaches in providing kinetic, thermodynamic and mode of action information. for example, kinetic binding information can be obtained directly using spr, whereas traditional enzyme kinetics experiments are required to access rate constants for slow-binding interactions. this often entails establishing time-courses under suitable concentrations of substrate(s) for which the control is linear and then observing the slow decrease in enzymatic rate as the test compound equilibrates with the target protein. these types of experiments are time-consuming and often can be difficult to analyse to extract the relevant rate constants controlling ligand binding. a further example is discerning order of addition of substrates to the enzyme in the reaction mechanism. this can be assessed using traditional enzyme kinetic experiments where the rate is measured whilst varying one substrate concentration in the presence of a fixed concentration of the other substrate. this can also be extended to test compounds. however, it can be complemented relatively straightforwardly by using itc to determine whether the presence of one ligand (for example substrate) is required or competes with binding of another (a second substrate, or analogue, or a test compound). admet & dmpk 7(4) (2019) 222-241 biophysical methods in early drug discovery doi: 10.5599/admet.733 235 itc can also provide the values of the thermodynamic contributions (enthalpy, h and entropy, s) to the binding affinity directly, and from the temperature dependence of the kd values, van’t hoff values can be calculated from any of the other biophysical methods. the same temperature dependent studies for the association and dissociation rate constants measured by spr lead to transition state energies for the association and dissociation steps of the binding reaction which enables construction of detailed thermodynamic reaction pathway models for protein-ligand binding. these approaches were used to provide structural and dynamic insights into the binding of different compounds to fgfr1 kinase, to understand the energetics required for movement of the activation loop [37]. although, it can be difficult to predict these values or to use this type of thermodynamic data directly to design new compounds during lead optimisation, having the ability to dissect the contributions to binding and transition state free energies can provide valuable understanding that will ultimately lead to a more thorough appreciation of the features that contribute to high affinity binding interactions. more interest, perhaps rightly, seems to have been focussed on kinetic parameters compared to thermodynamic data and the value of combining knowledge of target-compound residence times with pharmacokinetic (pk) and pharmacodynamic (pd) data seems to be gaining awareness [38]. the ability to combine kinetic data from spr on isolated proteins with data from cell washout experiments provides an extra level of information during lead optimisation. for example, the utility of pk/pd modelling, which attempts to describe the kinetics of the effects of compound following administration, is likely to be enhanced by direct measurements of the kinetics of compound binding and target turnover. case study 1 biophysical characterisation of phgdh fragment hits the enzyme 3-phosphohydroxyglycerate dehydrogenase (phgdh) utilises oxidised ninotinamide dinucleotide (nad+) and catalyses the conversion of 3-phosphohydroxyglycerate to 3phosphohydroxypyruvate in humans. the conversion of 3-phosphohydroxyglycerate to 3phosphohydroxypyruvate is the first, and rate-limiting step, in synthesis of the amino acid serine. studies have linked phgdh to the in vivo tumourigenesis in aggressive breast tumours and functional genomics reveal that the serine synthesis pathway is essential in breast cancer [39]. thus, the druggability of the target was explored with small molecules. to this end a fragment screen of 384 fragments was undertaken using crystal cocktail soaking (mixtures of 4) against the nad binding domain of phgdh and 34 hits were identified. the binding affinities of these hits were determined in a 2d nmr binding assay using 15nlabeled nad binding domain, figure 6a. these data, in conjunction with x-ray crystallography data, figure 6b, were used to identify fragments with the greatest potential for development, scheme 1. a simple analogue of an initial crystallography hit (compound 1), the 5-fluoroindole-2-carboxamide (compound 2) bound at a resolution of 1.95 å in the adenine pocket of the nad binding site (figure 6b). the indole nh forms a water-bridged hydrogen bond network with ser-211, and the 5-fluoro substituent fits nicely into a lipophilic cleft in the binding pocket. in addition, the carboxamide of compound 2 makes a water-bridged interaction with asp-174 and the crystal structure shows favourable vectors to grow the fragment into the phosphate binding site. in this endeavour, compounds were made that should displace the water mediating the hydrogen bond interaction between the carboxamide of compound 2 and asp-174. the synthesis of the α-hydroxymethylbenzylamide, compound 4, led to improved affinity to the single digit µm. a crystal structure of compound 4 (figure 6c) showed clear opportunities to further explore the phosphate binding pocket by installing additional functionality on the phenyl ring of the benzyl amide and this led to the synthesis of compound 5 with a further 10-fold improvement in affinity. holdgate et al. admet & dmpk 7(4) (2019) 222-241 236 figure 6 (a) 1h15n-trosyhsqc spectra illustrating binding of compound 2, compound concentrations used are shown as is the plot of the chemical shift change versus compound concentration used to determine the affinity (kd = 980 ± 60 µm); (b) crystal structure of compound 2 bound at a resolution of 1.95 å in the adenine pocket of the nad binding site showing a water-bridged interaction from the carboxamide to asp-174 (c) crystal structure of compound 4 shows displacement of the water mediated interaction with asp-174 and opportunities to add additional functionality on the phenyl ring of the benzyl amide, which led to the synthesis of compound 5. despite the abundant 2d nmr and crystallography data clearly demonstrating that the nad binding domain of phgdh was a folded and stable protein capable of binding ligands, further biophysical data comparing the function of this domain with the full-length protein highlighted differences. a combination of itc and spr measurements, figure 7a-c, clearly demonstrate that nadh binds approximately 20-fold more tightly to the full-length protein. considering this knowledge, a 1d nmr competition assay with nad+ as the reference compound was developed. this screen employed the more physiologically relevant full length phgdh protein, to provide continuity for affinity measurement during the subsequent analoguing phase. it is noteworthy that fragments and many of the synthesised analogues had no measurable activity in the human phgdh protein nad fluorescence intensity biochemical assay. whilst the biophysical methods allowed a wide dynamic range of weak mm to sub µm affinities be determined, the biochemical assay could only be used to provide affinities that could drive sar once those affinities were sub micromolar. in summary, a fragment crystal cocktail screen, supported with biophysical affinity measurements, provided the first known small molecule known nanomolar inhibitors of phgdh and valuable tools for interrogating the biology of this target. scheme 1. initially identified fragment binder (compound 1) and key compounds (2 – 5) that were made in the search for a cell active compound to probe the biology associated with phgdh as a drug target. admet & dmpk 7(4) (2019) 222-241 biophysical methods in early drug discovery doi: 10.5599/admet.733 237 figure 7 (a) itc titration data for the binding of nadh to full length phgdh, kd = 0.33 ± 0.08 µm, n = 0.78 ± 0.01. (b) spr titration data for nadh binding to full length phgdh, kd = 0.4 ± 0.03 µm determined from a steady state fit, with the kinetic fit curve shown in the inset. (c) spr titration data for nadh binding to nad binding domain of phgdh, kd = 6.7 ± 0.5 µm determined from a steady state fit, with the kinetic fit curve shown in the inset. case study 2 biophysical characterisation of hits from dna encoded library (del) screening dna-encoded library (del) technologies [40] offer an alternative to traditional high-throughput screening and have the unique ability to interrogate very large compound libraries (of around 10 9 –10 11 small molecules). whereas traditional hts methods often rely on activity-based isolated protein or cellbased assays, the del screening method is based on affinity selection versus an immobilised protein target, is rapid and requires only microgram quantities of protein. targets can be screened under multiple different experimental conditions in parallel. for example, different protein concentrations, addition of different cofactors or competing inhibitors, including proteins which represent important selectivity targets could all represent selection conditions that represent useful ways for identifying novel start points, with potentially differentiated modes of action. in a del screen versus the acyl carrier protein enoyl reductase inha, a pool of 11 dna-encoded libraries consisting of more than 65 billion on-dna compounds were tested and subsequent selections made utilising different forms of inha including the apo protein, inha.nad + , and inha.nadh complexes. analysis of output highlighted four general profiles of hits: (i) enriched only versus apo inha, (ii) enriched only versus inha.nad + complex, (iii) enriched only versus inha: nadh complex but not in the presence of an added inhibitory small molecule, and (iv) enriched in the presence of inha.nad + and inha.nadh but not in the presence of an added inhibitory small molecule. compounds, which were subsequently synthesised off-dna were tested in in vitro enzyme assays as well as profiled biophysically using spr. compounds were injected either alone, in the presence of 2 mm nad + , or in the presence of 100 μm nadh to monitor binding to different forms of the protein. where required, excess cofactor was included in the compound injections to ensure that the protein remained saturated with cofactor. compounds selected from these different conditions could then be easily profiled against apo, nadh or nad + -bound forms of inha and their respective affinities determined from spr experiments, table 5, allowing potentially differentiated profiles to be investigated. the del technology allowed the identification of multiple classes of inha inhibitors, some of which had cell-based activity directly from the primary screen. compounds were identified as cofactor-specific binders of inha with often with higher affinity for the nadh bound form. compounds similar to compound 7, figure 8, were demonstrated to inhibit bacterial growth in mycobacterium tuberculosis minimum inhibitory concentration (mic) assays and holdgate et al. admet & dmpk 7(4) (2019) 222-241 238 to kill mycobacterium tuberculosis infected human thp-1 cells. table 5. example of spr profiling of del screening hits compound kd µm apo inha inha.nad + inha.nadh 6 >100 13.4 ± 4.3 0.3 ± 0.2 7 >100 46.7 ± 11.6 0.09 ± 0.06 8 >100 >100 0.3 ± 0.1 9 >100 >100 36.8 ± 3.3 10 >100 49.0 ± 2.9 0.06 ± 0.03 11 12.4 ± 1.4 0.3 ± 0.1 5.0 ± 1.2 12 >100 >100 6.3 ± 2.6 s o n n n n o o o n n n n n o o o n n n o f n n n o n n n n n n o o o n n n n n n o o o o n n n f f f nn br oo o n n n n o o o compound 6 compound 7 compound 8 compound 9 compound 10 compound 11 compound 12 figure 8. compound structures identified via del screening and profiled using spr. discussion there are a range of biophysical methods that are available for application at many points during early drug discovery. each method has its own advantages and disadvantages that lead to different applications dependent upon the reagents available and the information content desired. for example, asms is most suited to higher throughput screening, whereas spr is positioned to deliver kinetic data. nmr can provide structural insights and is the preferred approach for primary fragment screening and itc provides a rapid thermodynamic characterisation of the binding event. the decision about which method or combination of methods to employ can be complex and subject to change depending upon particular projects, even if they are following similar hit identification and lead generation processes. the choice should ultimate focus on what information is required, the numbers of compounds required from which data needs to be generated to provide this information and the timescale required to deliver this data. of course, consideration of the target itself and the nature of the binding site(s) being targeted may also influence this decision. ultimately, practical limitations around protein availability, and costs will also need to be taken into account. however, projects often benefit from the impact that a combination of biophysical and biochemical assay derived information can provide, and biophysical methods should be an integral part of any protein centric drug discovery project. admet & dmpk 7(4) (2019) 222-241 biophysical methods in early drug discovery doi: 10.5599/admet.733 239 conclusions biophysical methods have evolved from being employed to address isolated issues with tool compounds or target proteins, experienced in some early stage projects [41], to now being an essential and integral part of the workflow positioned to establish and pursue hit identification and characterisation across the whole portfolio. over the last 5 years or so, this progression has been driven by the impact these methods have had, an increased throughput for some methods, and the recognition that a better understanding of the reagents, tools, assays and that mechanistic characterisation and differentiation of hits yields a more efficient early stage drug discovery process. this has led to a more focussed use of biophysics alongside the more traditional approaches, such as enzyme and cell-based assays, which has increased the quality of these early hit-finding assays. biophysical methods also are increasingly used as primary hit finding approaches, and no longer just for small, fragment-based screens, but also for screening increasingly large compound libraries. the recognition that compound binding, can be placed ahead of activity-based screening, in an orthogonal dyad has been embraced by screening groups and it may be especially useful for novel targets, whose functional activity is unknown or difficult to assay. as a result, biophysical methods will remain a key facet for increasing drug discovery success. conflict of interest: the authors are all employees of astrazeneca pharmaceuticals references [1] t. wiseman, s. williston, j.f. brandts, l.n. lin. rapid measurement of binding constants and heats of binding using a new titration calorimeter. anal. biochem. 179 (1989) 131-137. [2] l.g. fägerstam, a. frostell, r. karlsson, m. kullman, a. larsson, m. malmqvist, h. butt. detection of antigen-antibody interactions by surface plasmon resonance. application to epitope mapping. j. mol. recognit. 3 (1990) 208-214. [3] r.n. zuckermann, j.m. kerr, m.a. siani, s.c. banville, d.v. santi. identification of highest-affinity ligands by affinity selection from equimolar peptide mixtures generated by robotic synthesis. proc. natl. acad. sci. u.s.a. 89 (1992) 4505-4509. [4] s.b. shuker, p.j. hajduk, r.p. meadows, s.w. fesik. discovering high-affinity ligands for proteins: sar by nmr. science 274 (1996) 1531-1534. [5] r.m. burch, d.j. kyle. mass receptor screening for new drugs. pharm. res. 8 (1991) 141-147. [6] d.a. pereira, j.a. williams. origin and evolution of high throughput screening. br. j. pharmacol. 152 (2007) 53-61. [7] j.r. miller, v. thanabal, m.m. melnick, m. lall, c. donovan, r.w. sarver, d.y. lee, j. ohren, d. emerson. the use of biochemical and biophysical tools for triage of high-throughput screening hits a case study with escherichia coli phosphopantetheine adenylyltransferase. chem. biol. drug. des. 75 (2010) 444-454. [8] s.j. pfaff, m.s. chimenti, m.j. kelly, m.r. arkin. biophysical methods for identifying fragment-based inhibitors of protein-protein interactions. methods mol. biol. 1278 (2015) 587-613. [9] t.n. o'connell, j. ramsay, s.f. rieth, m.j. shapiro, j.g. stroh. solution-based indirect affinity selection mass spectrometry--a general tool for high-throughput screening of pharmaceutical compound libraries. anal. chem. 86 (2014) 7413-7420. [10] p.j. tonge. drug-target kinetics in drug discovery. acs chem. neurosci. 9 (2018) 29-39. [11] g. klebe. applying thermodynamic profiling in lead finding and optimization. nat. rev. drug. discov. 14 (2015) 95-110. [12] m.g. carneiro, e. ab, s. theisgen, g. siegal. nmr in structure-based drug design. essays biochem. 61 (2017) 485-493. holdgate et al. admet & dmpk 7(4) (2019) 222-241 240 [13] g.r. masson, m.l. jenkins, j.e. burke. an overview of hydrogen deuterium exchange mass spectrometry (hdx-ms) in drug discovery. expert opin. drug discov. 12 (2017) 981-994. [14] j.baell, m. a. walters. chemical con artists foil drug discovery. nature 513 (2014) 481-483. [15] m.w. pantoliano, e.c. petrella, j.d. kwasnoski, v.s. lobanov, j. myslik, e. graf, t. carver, e. asel, b.a. springer, p. lane, f.r. salemme. high-density miniaturized thermal shift assays as a general strategy for drug discovery. j. biomol. screen. 6 (2001) 429-440. [16] t.n. o'connell, j. ramsay, s.f. rieth, m.j. shapiro, j.g. stroh. solution-based indirect affinity selection mass spectrometry--a general tool for high-throughput screening of pharmaceutical compound libraries. anal. chem. 86 (2014) 7413-7420. [17] b. lamoree, r.e. hubbard. current perspectives in fragment-based lead discovery (fbld). essays biochem. 61 (2017) 453-464. [18] i. navratilova, a.l. hopkins. fragment screening by surface plasmon resonance. acs med. chem. lett. 1 (2010) 44-48. [19] j.k. kranz, c. schalk-hihi. protein thermal shifts to identify low molecular weight fragments. meth. enzymol. 493 (2011) 277-298. [20] r.b. van breemen, c.r. huang, d. nikolic, c.p. woodbury, y.z. zhao, d.l. venton. pulsed ultrafiltration mass spectrometry: a new method for screening combinatorial libraries. anal. chem. 69 (1997) 21592164. [21] t.n. o'connell, j. ramsay, s.f. rieth, m.j. shapiro, j.g. stroh. solution-based indirect affinity selection mass spectrometry--a general tool for high-throughput screening of pharmaceutical compound libraries. anal. chem. 86 (2014) 7413-7420. [22] i. sinclair, m. bachman, d. addison, m. rohman, d.c. murray, g. davies, e. mouchet, m.e. tonge, r.g. stearns, l. ghislain, s.s. datwani, l. majlof, e. hall, g.r. jones, e. hoyes, j. olechno, r.n. ellson, p.e. barran, s.d. pringle, m.r. morris, j. wingfield. acoustic mist ionization platform for direct and contactless ultrahigh-throughput mass spectrometry analysis of liquid samples. anal. chem. 91 (2019) 3790-3794. [23] t.e. carver, b. bordeau, m.d. cummings, e.c. petrella, m.j. pucci, l.e. zawadzke, b.a. dougherty, j.a. tredup, j.w. bryson, j. yanchunas, m.l. doyle, m.r. witmer, m.i. nelen, r.l. desjarlais, e.p. jaeger, h. devine, e.d. asel, b.a. springer, r. bone, f.r. salemme, m.j. todd. decrypting the biochemical function of an essential gene from streptococcus pneumoniae using thermofluor technology. j. biol. chem. 280 (2005) 11704-11712. [24] m. redhead, r. satchell, c. mccarthy, s. pollack, j. unitt. thermal shift as an entropy-driven effect. biochemistry 56 (2017) 6187-6199. [25] d.a. erlanson, s.w. fesik, r.e. hubbard, w. jahnke, h. jhoti. twenty years on: the impact of fragments on drug discovery. nat. rev. drug discov. 15 (2016) 605-619. [26] p. linke, k. amaning, m. maschberger, f. vallee, v. steier, p. baaske, s. duhr, d. breitsprecher, a. rak. an automated microscale thermophoresis screening approach for fragment-based lead discovery. j. biomol. screen. 21 (2016) 414-421. [27] a.r. leach, m.m. hann. molecular complexity and fragment-based drug discovery: ten years on. curr. opin. chem. biol. 15 (2011) 489-496. [28] m.j. harner, a.o. frank, s.w. fesik. fragment-based drug discovery using nmr spectroscopy. j. biomol. nmr 56 (2013) 65-75. [29] a.m. giannetti. from experimental design to validated hits a comprehensive walk-through of fragment lead identification using surface plasmon resonance. meth. enzymol. 493 (2011) 169-218. [30] d. beckett, e. kovaleva, p.j. schatz. a minimal peptide substrate in biotin holoenzyme synthetasecatalyzed biotinylation. protein sci. 8 (1999) 921-929. [31] c.a. shepherd, a.l. hopkins, i. navratilova. fragment screening by spr and advanced application to gpcrs. prog. biophys. mol. biol. 116 (2014) 113-123. admet & dmpk 7(4) (2019) 222-241 biophysical methods in early drug discovery doi: 10.5599/admet.733 241 [32] d.e. scott, m.t. ehebauer, t. pukala, m. marsh, t.l. blundell, a.r. venkitaraman, c. abell, m. hyvönen. using a fragment-based approach to target protein-protein interactions. chembiochem 14 (2013) 332-342. [33] m. jerabek-willemsen, c.j. wienken, d. braun, p. baaske, s. duhr. molecular interaction studies using microscale thermophoresis. assay drug dev. technol. 9 (2011) 342-353. [34] p. linke, k. amaning, m. maschberger, f. vallee, v. steier, p. baaske, s. duhr, d. breitsprecher, a. rak. an automated microscale thermophoresis screening approach for fragment-based lead discovery. j biomol. screen. 21 (2016) 414-421. [35] c.c. genick, d. barlier, d. monna, r. brunner, c. bé, c. scheufler, j. ottl. applications of biophysics in high-throughput screening hit validation. j. biomol. screen. 19 (2014) 707-714. [36] d. martinez molina, p. nordlund. the cellular thermal shift assay: a novel biophysical assay for in situ drug target engagement and mechanistic biomarker studies. annu. rev. pharmacol. toxicol. 56 (2016) 141-161. [37] t. klein, n. vajpai, j.j. phillips, g. davies, g.a. holdgate, c. phillips, j.a. tucker, r.a. norman, a.d. scott, d.r. higazi, d. lowe, g.s. thompson, a.l. breeze. structural and dynamic insights into the energetics of activation loop rearrangement in fgfr1 kinase. nat. commun. 6 (2015) 7877. [38] p.j. tonge. drug-target kinetics in drug discovery. acs chem. neurosci. 9 (2018) 29-39. [39] r. possemato, k.m. marks, y.d. shaul, m.e. pacold, d. kim, k. birsoy, s. sethumadhavan, h.k. woo, h.g. jang, a.k. jha, w.w. chen, f.g. barrett, n. stransky, z.y. tsun, g.s. cowley, j. barretina, n.y. kalaany, p.p. hsu, k. ottina, a.m. chan, b. yuan, l.a. garraway, d.e. root, m. mino-kenudson, e.f. brachtel, e.m. driggers, d.m. sabatini. functional genomics reveal that the serine synthesis pathway is essential in breast cancer. nature 476 (2011) 346-350. [40] d. neri, r.a. lerner. dna-encoded chemical libraries: a selection system based on endowing organic compounds with amplifiable information. annu. rev. biochem. 87 (2018) 479-502. [41] g. holdgate, s. geschwindner, a. breeze, g. davies, n. colclough, d. temesi, l. ward. biophysical methods in drug discovery from small molecule to pharmaceutical. methods mol. biol. 1008 (2013) 327-355. ©2019 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ metformin effects on zonulin level in polycystic ovarian women doi: http://doi.org/10.5599/admet.905 49 admet & dmpk 9(1) (2021) 49-55; doi: http://dx.doi.org/10.5599/admet.905 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper metformin effects on zonulin level in polycystic ovarian women manal ibrahim 1* , mutaz ahmeid 2 1 master of clinical biochemistry, college of pharmacy /university of mosul, iraq 2 ph.d. clinical biochemistry, college of medicine /university of tikrit, iraq *corresponding author: e-mail: alfarhamanal@gmail.com. received: september 04, 2020; revised: october 22, 2020; published: december 09, 2020 abstract zonulin protein is a haptoglobin precursor and functions to modulate the permeability of tight junctions between enterocytes. local inflammation or systemic inflammation can trigger zonulin expression. while the increased zonulin level causes an increase of intestinal permeability and entrance of foreign antigens, the latter can increase insulin resistance and inflammation. polycystic ovarian syndrome affects women during their reproductive age characterized by hyperinsulinemia and/or hyperandrogenemia and associated with infertility problems. changes in gut permeability such as irritable bowel syndrome are often found in pcos patients. while metformin increases insulin mediates glucose uptake and, acts as an insulin-sensitizing drug used to treat pcos patients is recently discovered to reshape intestinal bacteria and hence may affect intestinal action. this study was designed to find any association between zonulin level and other parameters in pcos patients and to find metformin treatment effects on zonulin in pcos patients. thirty-one newly diagnosed pcos women agree to take metformin 850 mg twice daily for three months and, and to give fasting serum samples to measure zonulin, fsh, lh, total testosterone, free testosterone, shbg, fasting insulin, and fasting serum glucose before and after treatment. the free testosterone and zonulin are determined by the elisa technique while the other parameters are determined by the cobas technique. according to patients’ homeostatic model assessment (homa-ir), the patients were divided into two sub-groups: the first group consisting of those with initial homa-ir less than two and the second group was those of an initial homa-ir of between two to four. whereas the first group showed no significant response to treatment, the second group showed a better response to metformin treatment, as demonstrated by their lh, total testosterone, free testosterone, fasting insulin, zonulin, and glucose levels. these parameters markedly improved after metformin treatment with pvalues of 0.08, 0.09, 0.07. 0.04, 0.01 and 0.06, respectively. ©2021 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords homa-ir; homeostatic model assessment for insulin resistance; pcos; polycystic ovarian syndrome; fsh; follicle stimulating hormone; lh; luteinizing hormone, shbg; sex hormone binding globulin introduction the polycystic ovarian syndrome can affect women during their reproductive age at an incidence rate of 3-10 % and characterized by sex hormone imbalance [1,2]. irregular menstrual cycle, polycystic ovary morphology, and hirsutism are the most frequent symptoms, with features of infertility and insulin resistance [3,4]. insulin resistance and obesity play roles in pcos pathophysiology [5,6]. insulin resistance http://dx.doi.org/10.5599/admet.905 http://dx.doi.org/10.5599/admet.905 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:alfarhamanal@gmail.com http://creativecommons.org/licenses/by/4.0/ m. ibrahim, m. ahmeid admet & dmpk 9(1) (2021) 49-55 50 gives rise to hyperinsulinemia [7], which directs ovarian steroidogenic enzymes to produce more androgen and abolish ovulation [8,9]. therefore, insulin-resistant pcos patients show more metabolic lipogenic changes, and abnormal steroidogenesis because of a decrease in insulin signaling, while non-insulin resistant pcos patients show abnormality in steroidogenesis only [10,11]. the homa-ir is a technique used to assess the β-cell function and insulin resistance from basal glucose and insulin and it equalizes ((fasting insulin multiplied by fasting glucose) divided by 22.5), normal homa ir is 1 [12,13]. insulin resistance occurs when fasting insulin is homa-ir>2 [14]. zonulin is a haptoglobin precursor that has a molecular-weight of 47 kda and modulates the permeability of tight junction (tj) between cells of the digestive tract [15,16]. an increase in zonulin is associated with an increase in gut permeability. zonulin expression is triggered by either local inflammation or systemic sub-clinical inflammation [17] and is suggested to have a role in the production of metabolic syndrome and insulin resistance [18]. in pcos women, zonulin is correlated with insulin resistance and the severity of the menstrual disorder [19]. moreover, irritable bowel syndrome and chronic fatigue syndrome, two conditions common in pcos are linked to an increase in gut permeability [20,21]. metformin drug used in pcos treatment can decrease hepatic gluconeogenesis by different mechanisms [22]: delay intestinal glucose absorption [23], reshape microbiota in the intestine [24,25], increase insulin-mediated glucose uptake in the liver, visceral fat, and muscle [26], and stimulates insulin signaling pathway in human granulosa cells and other cells of the body [27]. this insulin-sensitizing drug, promises pcos patients, may play a role in zonulin level and gut permeability. therefore, the study is designed to find the metformin effects on zonulin level in pcos patients and its association with other parameters. experimental thirty-one newly diagnosed pcos women according to rotterdam criteria are included in this study. patients with adrenal dysfunction, thyroid dysfunction, diabetics, and those on oral contraceptives are excluded from the study. fasting insulin, fasting glucose, fsh, lh, testosterone, free testosterone, shbg, and serum zonulin are measured before and after intake of metformin (bigomet®), 850 mg twice daily for three months. the research plan was approved by the scientific committee of the college of medicine, tikrit university, on 24/october/ 2018. the parameters of fsh, lh, total testosterone, shbg, and insulin are determined by cobas technique (cobas e 411 systems); fasting serum glucose was determined by cobas c 311analyzers. serum zonulin has been determined according to “human zonulin elisa kit cat no. mbs2607261 “supplied by mybiosource/usa company. free testosterone is determined by the elisa technique. statistical analysis performed by spss and the significant changes was considered when p-value less 0.05. the patients are divided into two subgroups according to their initial homeostatic model assessment (homa-ir) that is determined before starting metformin treatment. first group sixteen patients have initial homa-ir less than two, eleven of them have primary infertility and only five women have secondary infertility, and their mean age was 28.81 years. second group fifteen patients have initial homa-ir two to four, twelve of them have primary infertility and only three of them have secondary infertility, and their mean age was 26.93 years. results and discussion first group, those of initial homa-ir <2, did not show any significant difference in any of the parameters admet & dmpk 9(1) (2021) 49-55 metformin effects on zonulin level in polycystic ovarian women doi: http://doi.org/10.5599/admet.905 51 that measured including bmi, fsh, lh, testosterone, free testosterone, shbg, fasting insulin, fasting glucose, homa ir, and zonulin as seen in table 1. table 1. statistical analysis of parameters and zonulin before metformin therapy in pcos with initial (before treatment) homa-ir <2 initial homa-ir<2 mean +sd p-value bmi before 28.26+5.93 0.4 after 26.6+4.13 fsh(uiu/ml) before 6.16+1.69 0.28 after 7.66+4.15 lh(uiu/ml) before 10.26+5.04 0.25 after 7.83+5.47 testosterone(ng/ml) before 0.55+0.4 0.13 after 0.36+0.26 free testosterone (pg/ml) before 2.86+0.74 0.12 after 2.4+0.72 shbg (nmol/l) before 34.01+33.82 0.9 after 35.26+15.12 fasting insulin (uiu/ml) before 4.29+2.72 0.15 after 9.13+10.11 fasting glucose (mg/dl) before 89.63+11.86 0.27 after 95.33+13.66 homa-ir before 0.96+0.62 0.13 after 2.43+2.93 serum zonulin (ng/ml) before 20.83+15.46 0.52 after 16.90+15.21 table 2. statistical analysis of parameters and zonulin before and after metformin therapy in pcos with initial (before treatment) homa-ir = 2-4 initial homa-ir=2 – 4 mean +sd p-value bmi before 28.95+8.55 0.3 after 27.04+4.48 fsh (uiu/ml) before 5.94+2.39 0.11 after 4.68+1.89 lh (uiu/ml) before 12.11+8.08 0.08 after 7.24+6.01 testosterone(ng/ml) before 0.65+0.48 0.09 after 0.39+0.28 free testosterone (pg/ml) before 2.84+0.92 0.07 after 2.38+0.6 shbg (nmol/l) before 27.27+18.04 0.2 after 38.12+22.64 fasting insulin(uiu/ml) before 12.74+4.29 0.04* after 8.7+5.33 fasting glucose (mg/dl) before 89.73+24.2 0.06 after 82.91+8.37 homa-ir before 2.8+0.93 0.02* after 1.79+1.17 serum zonulin (ng/ml) before 16.26+10.10 0.01* after 7.92+5.09 *p-value less than 0.05 http://dx.doi.org/10.5599/admet.905 m. ibrahim, m. ahmeid admet & dmpk 9(1) (2021) 49-55 52 the second group those of initial homa-ir 2-4 (insulin resistance group), showed better response to metformin treatment (table 2). this group exhibits a marked decrease in lh, total testosterone, free testosterone, fasting glucose with a p-value of 0.08, 0.09, 0.07 and 0.06 respectively, and a significant decrease in fasting insulin, homa-ir, and zonulin with p-values of 0.04, 0.02 and 0.01 respectively. the study showed no change in body mass index (bmi) (tables 1 and 2), which is in agreement with sharma et al 2019 study [28]. it is well known that most pcos women are obese, and they have an increase in blood volume and the volume of distribution of any drug [29]. therefore, a higher dose of any drug is required to do equivalent action in non-obese patients [30]. the first group, those of initial homa-ir less than two (table 1), exhibited no significant decrease in their homa-ir and fasting glucose after metformin treatment, while the second group, those of initial homa-ir two to four (table 2) exhibited a significant decrease in homa-ir after the end of three months therapy with a p-value of 0.04, and a marked decrease in fasting glucose, with a p-value of 0.06. this result agrees with the study of nawrocka and starczewski, who showed that metformin can significantly decrease homa-ir only in insulin-resistant pcos patients [31]. moreover, the first group (table 1) showed no significant changes in zonulin value after treatment, while the second group (table 2), showed a significant decrease in zonulin with a p-value of 0.01, suggesting that metformin can decrease zonulin in insulinresistant patients with no menstrual disturbances. ilhan and yildizhan study suggested that zonulin is a pathogenic novel biomarker in pcos women and there is a strong correlation between homa-ir and zonulin in pcos patients [32]. also, zhang et al suggested that zonulin plays a role in pcos pathogenesis and can be used as a marker for detection of the severity of disease and in the prognosis of pcos treatment [19]. this is obvious, especially that increase zonulin leads to alteration in gut permeability, decrease the integrity of the intestinal barrier allowing an excess of the infectious agent and dietary antigen to facilitate mucosal immune elements such as il-6 which can lead to insulin resistance [33]. we proposed that the metformin effect on zonulin is related to the antibiotic effect of metformin on the gut and modulation of intestinal microbiota. however, the first group of those of initial (homa-ir <2) exhibited no changes in zonulin value as there is no difference in gut permeability and microbiota composition before and after treatment, as non-insulin resistance associated with no alteration in intestinal permeability. this agrees to a certain degree with cetin’s study conclusion, who included non-obese, normal homa-ir pcos patients in his study and suggested that insulin resistance is not triggered if there is integrity in the gut permeability [34]. it seems that this study is the first study to show the effect of metformin on zonulin in pcos patients. we found that metformin can decrease zonulin in insulin-resistant pcos patients with no menstrual irregularities and not prone to have diabetes, and we suggest that metformin effects on zonulin are not related to insulin-sensitizing activity but may be due to the modulation of intestinal bacteria. testosterone and free testosterone showed no significant decrease in the non-insulin resistant group (table 1) p-value 0.13 and 0.12 respectively, while the second group showed a marked but insignificant decrease in total testosterone p-value 0.09 and free testosterone p-value 0.07. this result agrees with baqer et al study in 2018 when they administer metformin 500 mg three times for three months and showed an insignificant decrease in total testosterone [35]. metformin can suppress androgen production by its action on the ovary and adrenal gland and increase shbg production by the liver [23]. different studies revealed that metformin causes a reduction in androgen, an increase of shbg, and a restoration of the normal menstrual cycle [36-40] but this study reveals a slight insignificant increase in shbg in both groups, which agree with singh et al. study who show no significant decrease in serum testosterone, fasting plasma glucose and bmi with a significant decrease in homa-ir, after intake of metformin 500 mg twice daily for 3 months [41]. on the other hand, upadhyaya and rehan study has reported that women of pcos admet & dmpk 9(1) (2021) 49-55 metformin effects on zonulin level in polycystic ovarian women doi: http://doi.org/10.5599/admet.905 53 with hyperinsulinemia show a significant reduction in testosterone in compare to the normo-insulinemic group after metformin treatment and they suggested that hyperinsulinemic, normo-androgenic pcos is a favorable condition for metformin response [42], also moghetti et al reveal that metformin responders usually had higher fasting plasma insulin [43]. concerning the level of lh, the second group (table 2) showed a decrease in lh with a p-value of 0.08, while the first group of less insulin-resistant patients (table 1) exhibited a p-value of 0.25. this is attributed to that the hyperinsulinemia increased androgen production by augmenting the lh effect for androgen production, while metformin can reverse this by decrease insulin level [44]. conclusions metformin alone does not affect body mass index within three months of treatment and has no effect on insulin level in non-insulin resistant pcos patients, but can decrease insulin in insulin-resistant pcos patients. concerning zonulin, the significant decrease in zonulin in women with higher insulin resistance may be attributed to the metformin effect on intestinal bacterial distribution. these results give rise to conclude that patients with higher homa-ir value before treatment (insulin resistant) are better metformin responders. conflict of interest: authors declare no conflict of interest. references [1] r.p. kauffman, v.m. baker, p. dimarino, t. gimpel, v.d. castracane. polycystic ovarian syndrome and insulin resistance in white and mexican american women: a comparison of two distinct populations. am. j. obstet. gynecol. 187 (2002) 1362-1369. [2] e.m. okoroh, w.c. hooper, h.k. atrash, h.r. yusuf, s.l. boulet. prevalence of polycystic ovary syndrome among the privately insured, united states, 2003-2008. am. j. obstet. gynecol. 207 (2012) 299.e1-7. [3] w.a. march, v.m. moore, k.j. willson, d.i. phillips, r.j. norman, m.j. davies. the prevalence of polycystic ovary syndrome in a community sample assessed under contrasting diagnostic criteria. hum. reprod. 25 (2010) 544-551. [4] r. azziz, e. carmina, d. dewailly, e. diamanti-kandarakis, h.f. escobar-morreale, w. futterweit, o.e. janssen, r.s. legro, r.j. norman, a.e. taylor, s.f. witchel. the androgen excess and pcos society criteria for the polycystic ovary syndrome: the complete task force report. fertil. steril. 91 (2009) 456-488. [5] s. franks, l.j. webber, m. goh, a. valentine, d.m. white, g.s. conway, s. wiltshire, m.i. mccarthy. ovarian morphology is a marker of heritable biochemical traits in sisters with polycystic ovaries. j. clin. endocrinol. metab. 93 (2008) 3396-3402. [6] c.m. povel, j.m. boer, e.j. feskens. shared genetic variance between the features of the metabolic syndrome: heritability studies. mol. genet. metab. 104 (2011) 666-669. [7] j. holte. disturbances in insulin secretion and sensitivity in women with the polycystic ovary syndrome. baillieres. clin. endocrinol. metab. 10(2) (1996) 221–247. [8] z.t. bloomgarden. world congress on the insulin resistance syndrome: insulin resistance mechanisms, the brain, and insulin resistance in youth and in the polycystic ovary syndrome. diabetes care 33(10) (2010) e124-e130. [9] r.s. legro, c.l. gnatuk, a.r. kunselman, a. dunaif. changes in glucose tolerance over time in women with polycystic ovary syndrome: a controlled study. j. clin. endocrinol. metab. 90 (2005) 3236-3242. [10] j.c. marshall, a. dunaif. should all women with pcos be treated for insulin resistance?. fertil. steril. 97 (2012) 18-22. http://dx.doi.org/10.5599/admet.905 m. ibrahim, m. ahmeid admet & dmpk 9(1) (2021) 49-55 54 [11] m. belani, a. deo, p. shah, m. banker, p. singal, s. gupta. differential insulin and steroidogenic signaling in insulin resistant and non-insulin resistant human luteinized granulosa cells-a study in pcos patients. j. steroid biochem. mol. biol. 178 (2018) 283-292. [12] t.m. wallace, j.c. levy, d.r. matthews. use and abuse of homa modeling. diabetes care 27 (2004) 1487-1495. [13] d.r. matthews, j.p. hosker, a.s. rudenski, b.a. naylor, d.f. treacher, r.c. turner. homeostasis model assessment: insulin resistance and beta-cell function from fasting plasma glucose and insulin concentrations in man. diabetologia 28 (1985) 412-419. [14] b. hedblad, p. nilsson, l. janzon, g. berglund. relation between insulin resistance and carotid intimamedia thickness and stenosis in non-diabetic subjects. results from a cross-sectional study in malmö, sweden. diabet. med. 17 (2000) 299-307. [15] d. rittirsch, m.a. flierl, b.a. nadeau, d.e. day, m.s. huber-lang, j.j. grailer, f.s. zetoune, a.v. andjelković, a. fasano, p.a. ward. zonulin as prehaptoglobin2 regulates lung permeability and activates the complement system. am. j. physiol. lung cell. mol. physiol. 304 (2013) l863-l872. [16] l. pacifico, e. bonci, l. marandola, s. romaggioli, s. bascetta, c. chiesa. increased circulating zonulin in children with biopsy-proven nonalcoholic fatty liver disease. world j. gastroenterol. 20 (2014) 17107-17114. [17] t. küme, s. acar, h. tuhan, g. çatlı, a. anık, ö. gürsoy çalan, e. böber, a. abacı. the relationship between serum zonulin level and clinical and laboratory parameters of childhood obesity. j. clin. res. pediatr. endocrinol. 9 (2017) 31-38. [18] j.m. moreno-navarrete, m. sabater, f. ortega, w. ricart, j.m. fernández-real. circulating zonulin, a marker of intestinal permeability, is increased in association with obesity-associated insulin resistance. plos one 7 (2012) e37160. [19] d. zhang, l. zhang, f. yue, y. zheng, r. russell. serum zonulin is elevated in women with polycystic ovary syndrome and correlates with insulin resistance and severity of anovulation. eur. j. endocrinol. 172 (2015) 29-36. [20] r. mathur, a. ko, l.j. hwang, k. low, r. azziz, m. pimentel. polycystic ovary syndrome is associated with an increased prevalence of irritable bowel syndrome. dig. dis. sci. 55 (2010) 1085-1089. [21] b.l. harlow, l.b. signorello, j.e. hall, c. dailey, a.l. komaroff. reproductive correlates of chronic fatigue syndrome. am. j. med. 105 (1998) 94s-99s. [22] g. rena, d.g. hardie, e.r. pearson. the mechanisms of action of metformin. diabetologia 60 (2017) 1577-1585. [23] c.j. bailey, r.c. turner. metformin. n. engl. j. med. 334 (1996) 574-579. [24] h. lee, g. ko. effect of metformin on metabolic improvement and gut microbiota. appl. environ. microbiol. 80 (2014) 5935-5943. [25] f. cabreiro, c. au, k.y. leung, n. vergara-irigaray, h.m. cochemé, t. noori, d. weinkove, e. schuster, n.d. greene, d. gems. metformin retards aging in c. elegans by altering microbial folate and methionine metabolism. cell 153 (2013) 228-239. [26] e. diamanti-kandarakis, c.d. christakou, e. kandaraki, f.n. economou. metformin: an old medication of new fashion: evolving new molecular mechanisms and clinical implications in polycystic ovary syndrome. eur. j. endocrinol. 162 (2010) 193-212. [27] h. xu, y. zhou, y. liu, j. ping, q. shou, f. chen, r. ruo. metformin improves hepatic irs2/pi3k/akt signaling in insulin-resistant rats of nash and cirrhosis. j. endocrinol. 229 (2016) 133-144. [28] n. sharma, siriesha, y. lugani, a. kaur, v. k. ahuja. effect of metformin on insulin levels, blood sugar, and body mass index in polycystic ovarian syndrome cases. j. fam. med. prim. care 8 (2019) 2691– 2695. [29] f. akpinar, b. dilbaz, d.a. cırık, s. yilmaz, s. kiykac, e. karahanoglu, l. mollamahmutoglu. the significance of anthropometric and endocrine parameters in ovulation induction with clomiphene citrate in women with polycystic ovary syndrome. saudi med. j. 37 (2016) 1272-1275. admet & dmpk 9(1) (2021) 49-55 metformin effects on zonulin level in polycystic ovarian women doi: http://doi.org/10.5599/admet.905 55 [30] g. sachdeva, s. gainder, v. suri, n. sachdeva, s. chopra. comparison of clinical, metabolic, hormonal, and ultrasound parameters among the clomiphene citrate-resistant and clomiphene citrate-sensitive polycystic ovary syndrome women. j. hum. reprod. sci. 12 (2019) 216-223. [31] j. nawrocka, a. starczewski. effects of metformin treatment in women with polycystic ovary syndrome depends on insulin resistance. gynecol endocrinol 23 (2007) 231-237. [32] g. a. ilhan and b. yildizhan. evaluation of zonulin, a marker of intestinal permeability as a novel biomarker in polycystic ovary syndrome. fertility and sterility 110(2018) :e117 [33] s.e. shoelson, j. lee, a.b. goldfine. inflammation and insulin resistance. j. clin. invest. 116 (2006) 1793-1801. [34] z. cetin, a. kosem, b. can, o. baser, m. catak, t. turhan, d. berker. serum zonulin level is not elevated in patients with polycystic ovary syndrome without metabolic syndrome. arch. gynecol. obstet. 300 (2019) 1785-1790. [35] l.s. baqer, m.s. ahmeid, a.h. al-obaidi. evaluation the effect of metformin on hormones serum levels in women with polycystic ovary syndrome. tikrit journal of pure science. 22 (2018) 1-5. [36] l.c. morin-papunen, i. vauhkonen, r.m. koivunen, a. ruokonen, h.k. martikainen, j.s. tapanainen. endocrine and metabolic effects of metformin versus ethinyl estradiol-cyproterone acetate in obese women with polycystic ovary syndrome: a randomized study. j. clin. endocrinol. metab. 85 (2000) 3161-3168. [37] r. fleming, z.e. hopkinson, a.m. wallace, i.a. greer, n. sattar. ovarian function and metabolic factors in women with oligomenorrhea treated with metformin in a randomized double blind placebo-controlled trial. j. clin. endocrinol. metab. 87 (2002) 569-574. [38] r. al-fadhli, t. tulandi. laparoscopic treatment of polycystic ovaries: is its place diminishing. curr. opin. obstet. gynecol. 16 (2004) 295-298. [39] c.j. glueck, n. goldenberg, j. pranikoff, m. loftspring, l. sieve, p. wang. height, weight, and motorsocial development during the first 18 months of life in 126 infants born to 109 mothers with polycystic ovary syndrome who conceived on and continued metformin through pregnancy. hum. reprod. 19 (2004) 1323-1330. [40] e.m. velazquez, s. mendoza, t. hamer, f. sosa, c.j. glueck. metformin therapy in polycystic ovary syndrome reduces hyperinsulinemia, insulin resistance, hyperandrogenemia, and systolic blood pressure, while facilitating normal menses and pregnancy. metabolism 43 (1994) 647-654. [41] b. singh, s. panda, r. nanda, s. pati, m. mangaraj, p.k. sahu, p.c. mohapatra. effect of metformin on hormonal and biochemical profile in pcos before and after therapy. indian j. clin. biochem. 25 (2010) 367-370. [42] p. upadhyaya, h. s. rehan. a subgroup analysis of effect of metformin in indian women with pcos. jk science 13(2011):19-22 [43] p. moghetti, r. castello, c. negri, f. tosi, f. perrone, m. caputo, e. zanolin, m. muggeo. metformin effects on clinical features, endocrine and metabolic profiles, and insulin sensitivity in polycystic ovary syndrome: a randomized, double-blind, placebo-controlled 6-month trial, followed by open, long-term clinical evaluation. j. clin. endocrinol. metab. 85 (2000) 139-146. [44] m. zahra, m. shah, a. ali, r. rahim. effects of metformin on endocrine and metabolic parameters in patients with polycystic ovary syndrome. horm. metab. res. 49 (2017) 103-108. ©2021 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/) http://dx.doi.org/10.5599/admet.905 http://creativecommons.org/licenses/by/4.0/ image-based dissolution analysis for tracking the surface stability of amorphous powders doi: https://dx.doi.org/10.5599/admet.839 401 admet & dmpk 8(4) (2020) 401-409; doi: https://doi.org/10.5599/admet.839 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper image-based dissolution analysis for tracking the surface stability of amorphous powders jernej štukelj* 1,2 , mikael agopov 2 , jouko yliruusi 2 , clare j. strachan 1 and sami svanbäck 2 1 division of pharmaceutical chemistry and technology, university of helsinki, viikinkaari 5e, 00790 helsinki, finland 2 the solubility company oy, viikinkaari 4, 00790 helsinki, finland *corresponding author: e-mail: jernej.stukelj@helsinki.fi; tel.: +358-29-415-9204 received: may 01, 2020; revised: july 10, 2020; published: july 13, 2020 abstract poor solubility of crystalline drugs can be overcome by amorphization – the production of high-energy disordered solid with improved solubility. however, the improved solubility comes at a cost of reduced stability; amorphous drugs are prone to recrystallization. because of recrystallization, the initial solubility enhancement is eventually lost. therefore, it is important to understand the recrystallization process during storage of amorphous materials and its impact on dissolution/solubility. here, we demonstrate the use of image-based single-particle analysis (spa) to consistently monitor the solubility of an amorphous indomethacin sample over time. the results are compared to the xrpd signal of the same sample. for the sample stored at 22 °c/23 % relative humidity (rh), full crystallinity as indicated by xrpd was reached around day 40, whereas a solubility corresponding to that of the γ crystalline form was measured with spa at day 25. for the sample stored at 22 °c/75 % rh, the xrpd signal indicated a rapid initial phase of crystallization. however, the sample failed to fully crystallize in 80 days. with spa, solubility slightly above that of the crystalline γ form was measured already on the second day. to conclude, the solubility measured with spa directly reflects the solid-state changes occurring on the particle surface. therefore, it can provide vital information – in a straightforward manner while requiring only minuscule sample amounts – for understanding the effect of storage conditions on the dissolution/solubility of amorphous materials, especially important in pharmaceutical science. ©2020 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords solubility; stability; crystallization; amorphous; single-particle analysis; spa introduction in pharmaceutical industry, solid-state form selection of an active pharmaceutical ingredient (api) is an important step in the development of a solid dosage form. the solid-state form governs the processability, stability and bioavailability of the final product. thus, poorly soluble crystalline apis can be replaced by their respective, more soluble, amorphous forms [1-3]. however, an increase in apparent solubility, and consequently bioavailability, comes at a price of instability. therefore, monitoring and understanding of amorphous form instability – recrystallization – is crucial for optimal product design. http://dx.doi.org/10.5599/admet.839 https://doi.org/10.5599/admet.839 http://www.pub.iapchem.org/ojs/index.php/admet/index http://creativecommons.org/licenses/by/4.0/ jernej štukelj et al. admet & dmpk 8(4) (2020) 401-409 402 the non-steroidal-anti-inflammatory drug indomethacin has been widely used as model drug to study recrystallization behaviour with the first studies dating all the way back to the 1970s [4,5]. stability studies looking at amorphous indomethacin can be divided in two groups: amorphous stability during storage and amorphous stability during dissolution testing. the behaviour of amorphous indomethacin during storage is governed by temperature and humidity [6– 8]. the two parameters determine the direction of crystallization of amorphous indomethacin prepared by cooling the melt; at 30 °c, the metastable α form predominantly crystallizes at high humidity (>56 % relative humidity (rh)) and the thermodynamically stable γ form predominantly crystallizes at low humidity (<43 % rh) [6]. water acts as a plasticizer and thus, with an increase in humidity, the glass transition (tg) of amorphous indomethacin decreases. temperatures below the tg favour formation of the γ form and temperatures above tg favour formation of the α form. furthermore, crystallization may also be considered from a surface versus bulk perspective: for quench-cooled amorphous films, v. andronis et al. reported 30 °c/11 % rh as a limit below which crystallization was surface-initiated and 30 °c/21 % rh as a limit above which crystallization was bulk-initiated [6]. in addition, wu et al. reported surface crystallization being two orders of magnitude higher compared to bulk crystallization for amorphous films stored at 40 °c [8]. when exposed to dissolution medium, the amorphous solid undergoes solid-state transformation via two pathways: a direct recrystallization of the solid and/or a solution mediated crystallization from the supersaturated solution [9]. d. alonzo et al. reported that crystallization from supersaturated solution was the main mechanism reducing the solubility of amorphous indomethacin. however, the adsorption of water onto the exposed amorphous surface increases molecular mobility and thus potentially promotes the surface crystallization rate [6,10]. the dissolution studies in flow-through chambers coupled with raman probes enabled the monitoring of dissolution rate and crystallization, revealing the dependency of dissolution rate on processing history of the amorphous solid [11,12]. priemel et al. studied amorphous tablets stored at 30 °c/23 % rh with differential scanning calorimetry (dsc), x-ray powder diffraction (xrpd) and fourier-transform infrared (ftir) spectroscopy and correlated the results of the analytical techniques to the dissolution performance of tablets after five days of storage [13]. the dissolution is a purely surface phenomenon, and thus correlating it with the results provided by the bulk analytic techniques can be challenging. novakovic et al. employed nonlinear optics to study surface crystallization and dissolution of indomethacin tablets (300 mg) stored at 30 °c/23 % rh and 30 °c/75 % rh for 1, 2, 7 and 22 days [14]. a trend towards decreased dissolution rate with increased storage time was observed. furthermore, novakovic et al. reported a delay in crystallization onset during storage and multiple factors affecting dissolution of amorphous indomethacin tablets (500 mg) as a result of the polymer coating [15]. by now, one should note the amount of research conducted with amorphous indomethacin is quite considerable. however, the studies that have investigated the dissolution/solubility enhancement of amorphous indomethacin – or any other amorphous drug – throughout the storage time are scarce. in our opinion, this is due to the following factors: substantial amount of drug compound needed and challenges for analytical methods to detect the changes in dissolution/solubility of the originally amorphous sample [16]. nevertheless, these studies are highly important as surfaces dictate the dissolution behaviour of pharmaceutical compounds and by understanding the processes occurring on the surfaces we can better predict their in vivo performance. with the emergence of the image-based single-particle analysis (spa) for determining the solubility (figure 1a), these two limiting factors above can be easily tackled [17]. the spa methodology requires minuscule sample amounts (<100 µg), short equilibration time and it has already proven successful in admet & dmpk 8(4) (2020) 401-409 amorphous solubility over time doi: https://dx.doi.org/10.5599/admet.839 403 directly measuring the amorphous solubility, also referred to as apparent solubility, of several compounds [18]. in this study, we continuously monitor the xrpd signal and the solubility, employing spa, of amorphous indomethacin stored at two different ambient conditions for 80 days. experimental materials indomethacin (γ form) was received from orion pharma, espoo, finland. the α form was prepared by dissolving 3 g of indomethacin in 20 ml of etoh and heated up to 70 °c. upon complete dissolution, the system was kept at the set temperature for an additional 5 min and then rapidly cooled down in an ice bath. the formed slurry was filtered and dried in a vacuum oven at 70 °c for 1 h. amorphous indomethacin was prepared by melting the γ form on a hot plate at 165 °c. after visual confirmation of complete melting, the melt was kept on a hot plate for additional 30 seconds. after 30 seconds, the melt was immediately cooled down on an aluminium heat sink. it has been shown previously that this preparative technique does not cause significant sample degradation [19–21]. the quench cooled melt was then stored over phosphorus pentoxide at room temperature for one hour. after one hour, the amorphous indomethacin was gently ground with a mortar and pestle. a small amount of the sample was taken for the dsc analysis and the remaining amount was halved and placed into two xrpd holders with a low-scatter background (figure 1b). figure 1. (a) schematic of the single-particle analysis (spa) setup. (b) schematic of the low-scatterbackground holder and sampling protocol used in the study. the middle area (red) was scanned, on the measuring days, with the xrpd and was not altered throughout the course of the study. the spa sampling area was between the xrpd scan area and the flush aluminium surface of the holder (grey). on the measuring day, three samples from this area were taken with a spatula from opposite sides of the holder starting with day zero at the three dashed lines and then moving clockwise with each consecutive measuring day. in that way, the sampling area distribution was maximized and no spot was sampled more than once. storage conditions one of the two sample holders with the amorphous indomethacin was stored at 23 % rh and the other at 75 % rh. both samples were kept at 22±0.5 °c. http://dx.doi.org/10.5599/admet.839 jernej štukelj et al. admet & dmpk 8(4) (2020) 401-409 404 differential scanning calorimetry (dsc) dsc measurements were conducted to characterize the solid-state of crystalline and amorphous samples. samples (2 to 5 mg) were packed into standard aluminium crucibles (40 µl) with pierced lids. the samples were analysed with a dsc823e instrument (mettler-toledo, greifensee, switzerland) equipped with a cooling system (julabo ft 900, seelbach, germany). nitrogen (50 ml/min) was used as a purge gas. equilibration at 25 °c for 3 minutes was followed by linear heating with a heating rate of 10 °c/min. measurements were made in triplicate. thermal events were analysed using the stare software (mettlertoledo, greifensee, switzerland). x-ray powder diffraction (xrpd) xrpd diffractograms were recorded to characterize the solid-state of the crystalline and amorphous samples. moreover, xrpd was used to follow the solid-state changes of the two amorphous samples over time (figure 1b). an aeris diffractometer (malvern panalytical b.v., almelo, netherlands) using cu kα radiation (λ = 1. 1.540598 å) and a divergence slit of 0.76 mm was used. samples were placed on a lowscatter-background holder and measured with a step size of 0.0108664° at 40 kv and 8.0 ma from 5° to 35° (2θ). the xrpd measurements of the two amorphous samples were carried out on the following days: 0, 2, 4, 7, 10, 14, 17, 21, 25, 29, 32, 38, 44, 53, 60 and 80. the height of the peaks at 11.6°, 16.7°, 21.8° and 26.6° of the 2θ angle was used to quantify the solid-state change that occurred. the height of each peak was divided by the height of the silicon peak (28.4°) measured on the same day to correct the response for the fluctuation of the x-ray beam. the silicon peak was measured from the silicon plate provided by malvern panalytical. the four corrected peak height values were then averaged in order to obtain a single quantifying value. single-particle analysis (spa) sampling of the two amorphous samples stored at 23 % and 75 % rh was carried out as presented in figure 1. on each of the measuring days (0, 2, 4, 7, 10, 14, 17, 21, 25, 29, 32, 38, 44, 53, 60 and 80) three samples were extracted from opposite sides of the xrpd sample holder and never from the same location twice. the xrpd scan area was left intact in order for sampling not to affect the xrpd signal. an aqueous solution of hcl ph 2.0 was used as a solvent. separately, crystalline samples of α and γ indomethacin were also measured with the spa method. the spa methodology has already been previously described by svanbäck et al. and štukelj et al. [17,18,22]. shortly, the spa is comprised of two prime components – the flow-through setup and the custom-made analysis software (figure 1a). the flow-through setup is capable of immobilizing the drug particles under steady flow. immobilized particles are imaged as they dissolve – reduce in size. in this study, each measurement took roughly 5 minutes. captured images of the particles are analysed using the custom-made software, which tracks the reduction in particle morphology throughout the course of a measurement. the decrease in mass over time or dissolution rate (dm/dt) under sink conditions (when the bulk concentration (cb) can be assumed to be zero) is, according to noyes and whitney (eq. 1), directly proportional to the equilibrium solubility (cs) of a compound. the setup-dependent-transport-rate constant (k) is determined as described by svanbäck et al. [23]. furthermore, if the compound is in its most stable crystalline form, the determined equilibrium solubility corresponds to the thermodynamic equilibrium solubility of a compound. on the other hand, for a purely amorphous sample, in the absence of crystallisation during the spa measurement, the measured solubility is actually the amorphous solubility – the maximum drug concentration in solution upon dissolution of an amorphous solid. the approach of measuring amorphous solubility using the spa technology was recently validated by štukelj et al. (2019), admet & dmpk 8(4) (2020) 401-409 amorphous solubility over time doi: https://dx.doi.org/10.5599/admet.839 405 with the amorphous solubility of five diverse compounds being assessed by spa and two orthogonal methods: standardized supersaturation and precipitation, and theoretical estimation based on thermal analysis [18].  s b s d d d d m m k c c kc t t       (1) results and discussion characterization of the initial solid-state form the dsc thermograms of the amorphous and two crystalline indomethacin forms are presented in figure 2a. the amorphous thermogram exhibits a glass transition mid-point at 44.2 ± 1.0 °c, a crystallization onset at 99.5 ± 0.2 °c, a melting peak onset corresponding to the α form at 153.2 ± 0.1 °c and a melting peak onset corresponding to the γ form at 159.8 ± 0.1 °c. crystallization of freshly-prepared melted and cooled amorphous indomethacin during heating to both these forms has been observed earlier [13,24]. the thermogram of the γ form exhibits a melting peak at 159.8 ± 0.2 °c and the thermogram of the α form exhibits a melting peak at 152.7 ± 0.2 °c. the xrpd diffractogram of the amorphous sample exhibits a distinctive amorphous halo and is devoid of any peaks that would indicate crystallinity (figure 2b) [13]. the diffractograms of the α and γ crystalline samples match the predicted diffractograms of the respective forms (indmet02 and indmet01) from the cambridge structural database. figure 2. (a) thermograms of α, γ and amorphous indomethacin samples. (b) diffractograms of α, γ and amorphous indomethacin samples. solubility of α and γ indomethacin the intrinsic solubility of α and γ indomethacin measured using the spa method is compared to the intrinsic solubility of these two forms as assessed using the orthogonal methods in table 1. the intrinsic solubility of the α form was measured previously by surwase et al. using a shake-flask method [20]. the intrinsic solubility of the γ form was measured by štukelj et al. using a µdiss profiler™ [18]. the obtained values are in good agreement and within the generally accepted average uncertainty in experimental solubility measurements of 0.6 log units [25]. table 1. intrinsic solubilities of α and γ indomethacin measured with the spa method compared to previously determined intrinsic solubility values using the shake-flask (α form) and µdiss profiler™ (γ form) methods. intrinsic solubility (µg/ml) shake-flask/µdiss profiler™ spa α form 2.4±0.2 a 3.5±0.6 γ form 1.3±0.3 b 1.5±0.2 a reference [20]. b reference [18] http://dx.doi.org/10.5599/admet.839 jernej štukelj et al. admet & dmpk 8(4) (2020) 401-409 406 tracking the solid-state changes and amorphous solubility over time the xrpd diffractograms recorded over time of the two amorphous samples stored at 22 °c/23 % rh and 22 °c/75 % rh are presented in figure 3. crystallization started sooner at 75 % rh than at 23 % rh, as peaks of crystallinity can already be seen on the second day (figure 3b). in contrast, at 23 % rh the first peaks of crystallinity can be only observed at day 7 (figure 3a). both samples predominantly crystallized into the γ form; the differences in relative peak intensities for the stored samples versus the reference γ form diffractogram most likely arise due to crystal habit (or possibly preferred crystal orientation) differences associated with the different crystal growth conditions. the presence of a minor amount of the α form is indicated by a small peak appearing at 8.5°. a study with amorphous indomethacin tablets prepared via the melt and stored at 30 °c/23 % rh and 30 °c/75 % rh was performed by novakovic et al. [26]. the researchers stored tablets for 22 days and observed transformation into the γ form at 23 % rh and the α form at 75 % rh. the transformation, however, was not exclusive, as regions that could be associated with the α form were also detected at 23 % rh. in contrast, patterson et al. obtained a predominantly γ sample with some α form present when they stored quench cooled amorphous indomethacin at 30 °c/75 % rh [21]. this ostensibly stochastic event could have been the result of other factors. besides temperature and humidity, preparation method and parameters can also influence the resulting polymorphic form [24]. figure 3. xrpd diffractograms of the amorphous samples stored at (a) 22 °c/23 % rh and (b) 22 °c/75 % rh. dashed lines indicate the γ form xrpd peaks used for quantification: 11.6°, 16.7°, 21.8° and 26.6°. experimentally recorded diffractograms of α and γ indomethacin are plotted for comparison. to quantify the solid-state changes of the amorphous samples, four distinctive peaks characteristic of the γ form at 11.6°, 16.7°, 21.8° and 26.6° 2θ were selected and are indicated by the dashed lines in figure 3. the resulting values were averaged and the log value of the average plotted against time (figure 4). in 80 days, the amorphous sample stored at 23 % rh almost fully converted into the γ form (figure 4a). in contrast, full conversion of the sample stored at 75 % rh, as measured by the averaged peak intensity, did not occur despite the crystallization starting sooner than for the sample stored at 23 % rh. the presence of the elevated xrpd baseline, indicating amorphousness, at the end of the study (figure 3b) could be explained by one, or a combination, of two possibilities: i) rapid nucleation at higher humidity resulted in the formation of nanocrystals, which due to their small size and many defects still appeared partially amorphous, and ii) the presence and morphology of crystals at the particle surface may have acted as a mechanical barrier for further crystal growth in the particle core. when interpreting the xrpd results, one must also keep in mind that the penetration depth of x-rays is in the order of several hundred µm and that results may significantly vary if analysed with a different technique [13]. admet & dmpk 8(4) (2020) 401-409 amorphous solubility over time doi: https://dx.doi.org/10.5599/admet.839 407 sampling for the spa measurements was conducted as depicted in figure 1b. three measurements of the amorphous solubility of each sample conducted on the same day were averaged and plotted against time (figure 4). for the sample stored at 23 % rh, the amorphous solubility gradually decreased and reached that of the γ crystalline form (1.69 ± 0.58 µg/ml) at day 25. in contrast, the xrpd signal reached a plateau only around day 40 (figure 4a). thus, even prior to the sample being fully crystalline based on the xrpd measurements, its dissolution rate and solubility were already fully governed by the crystalline form. this can be explained by the surface crystallizing faster than the bulk [8]. priemel et al. have observed that only the surface of freshly prepared amorphous tablets crystallized during dissolution testing [13]. in contrast, the whole tablet crystallized during dissolution when it was stored for 5 days at 30 °c/23 % rh. the researchers explained the observation by a denser crystalline layer forming during dissolution, which acted as a mechanical barrier for further solution-mediated crystallization, versus a less dense crystalline layer forming during storage, which eventually promoted the migration of crystallization further into the tablet core. consequently, the dissolution rate of amorphous indomethacin tablets stored for five days matched the dissolution rate of the γ form. figure 4. (a) plot of the xrpd quantified signal and amorphous solubility of the sample stored at 22 °c/23 % rh. dotted line represents the average plateau solubility (1.69 ± 0.58 µg/ml). (b) plot of the xrpd quantified signal and amorphous solubility of the sample stored at 22 °c/75 % rh. dotted line represents the average plateau solubility (2.04 ± 0.79 µg/ml). for the sample stored at 75 % rh, the solubility sharply drops already on the second day and then fluctuates around 2.04 µg/ml – slightly above the solubility of the reference γ form, but below the solubility of the α form measured in this study (3.5 ± 0.6 µg/ml). we have earlier reported the solubility of the α indomethacin form being roughly 3 times higher compared to the solubility of the γ form [22]. however, the higher solubility of the amorphous sample stored at 75 % rh measured over time in this study is most likely the result of one or both of the following phenomena: i) the presence of minor amounts of the α form in the sample, and ii) differences in surface morphology, including crystal habit, size and crystal defects, due to nanocrystals forming during storage, as observed by priemel et al. [13], versus single micro-scale crystals of the reference γ form. the sharp drop in solubility on the second day, on the other hand, may be the result of two coinciding factors: i) grinding resulted in the formation of crystallization nuclei on the surface of the particles, which were undetectable with the xrpd [21], and ii) during storage, adsorbed water on the surface increased the molecular mobility and decreased the tg of the amorphous indomethacin thus further facilitating the surface crystallization [6,10]. once exposed to the dissolution medium, even if the surface was not yet fully crystalline, it was certainly very prone to rapid crystallization. http://dx.doi.org/10.5599/admet.839 jernej štukelj et al. admet & dmpk 8(4) (2020) 401-409 408 conclusions in this study, we have demonstrated the suitability of the spa technology for tracking the stability/solubility of an initially amorphous sample during storage. the short equilibration time and low sample consumption of the method enabled – for the first time – continuous observation of solubility of a particulate amorphous sample as opposed to tablets which have conventionally been required. the results for the amorphous indomethacin stored at 22 °c/23 % rh indicate a gradual decrease in solubility towards the intrinsic solubility of the γ form. the y form solubility was reached at day 25, which was roughly 15 days prior to xrpd detecting a fully crystalline sample. for the sample stored at 22 °c/75 % rh, the xrpd signal initially indicated fast crystallization. however, the sample did not fully crystallize in 80 days. in contrast, on the second day, spa detected a rapid decrease in solubility towards a value slightly above the solubility of the γ form. the slightly higher solubility was most likely measured due to the presence of crystals with different morphology, nanocrystals, and some associated residual disorder on the surface compared to single crystals of the reference γ form. the stability of an amorphous material approached from the solubility perspective, as enabled with spa, directly reflects the changes occurring on particle surfaces that dictate the dissolution behaviour. this straight-forward and sensitive approach could provide crucial information for better understanding of the link between stability of amorphous, and also other, materials, their solubility and in vivo performance, which is especially important in pharmaceutical science. acknowledgements: j.š. acknowledges doctoral programme in drug research (dpdr) funding. conflict of interest: j.š., m.a., j.y. and s.s. are shareholders of the solubility company oy that owns the intellectual property rights to the spa method. references [1] b.c. hancock, m. parks. what is the true solubility advantage for amorphous pharmaceuticals?. pharm. res. 17 (2000) 397–404. [2] s.b. murdande, m.j. pikal, r.m. shanker, r.h. bogner. solubility advantage of amorphous pharmaceuticals: i. a thermodynamic analysis. j. pharm. sci. 99 (2010) 1255–1263. [3] s.b. murdande, m.j. pikal, r.m. shanker, r.h. bogner. solubility advantage of amorphous pharmaceuticals: ii. application of quantitative thermodynamic relationships for prediction of solubility enhancement in structurally diverse insoluble pharmaceuticals. pharm. res. 27 (2010) 2704–2714. [4] l. borka. the polymorphism of indomethacine. new modifications, their melting behavior and solubility. acta pharm. suec. 11 (1974) 295–303. [5] h. imaizumi, n. nambu, t. nagai. stability and several physical properties of amorphous and crystalline forms of indomethacin. chem. pharm. bull. 28 (1979) 2565–2569. [6] v. andronis, m. yoshioka, g. zografi. effects of sorbed water on the crystallization of indomethacin from the amorphous state. j. pharm. sci. 86 (1997) 346–357. [7] v. andronis, g. zografi. crystal nucleation and growth of indomethacin polymorphs from the amorphous state. j. non. cryst. solids. 271 (2000) 236–248. [8] t. wu, l. yu. surface crystallization of indomethacin below tg. pharm. res. 23 (2006) 2350–2355. [9] d.e. alonzo, g.g. z. zhang, d. zhou, y. gao, l.s. taylor. understanding the behavior of amorphous pharmaceutical systems during dissolution. pharm. res. 27 (2010) 608–618. [10] v. andronis, g. zografi. the molecular mobility of supercooled amorphous indomethacin as a function of temperature and relative humidity. pharm. res. 15 (1998) 835–842. admet & dmpk 8(4) (2020) 401-409 amorphous solubility over time doi: https://dx.doi.org/10.5599/admet.839 409 [11] m. savolainen k. kogermann, a. heinz, j. aaltonen, l. peltonen, c. strachan, j. yliruusi. better understanding of dissolution behaviour of amorphous drugs by in situ solid-state analysis using raman spectroscopy. eur. j. pharm. biopharm. 71 (2009) 71–79. [12] k. greco, r. bogner. crystallization of amorphous indomethacin during dissolution: effect of processing and annealing. mol. pharm. 7 (2010) 1406–1418. [13] p.a. priemel, h. grohganz, k.c. gordon, t. rades, c.j. strachan. the impact of surfaceand nanocrystallisation on the detected amorphous content and the dissolution behaviour of amorphous indomethacin. eur. j. pharm. biopharm. 82 (2012) 187–193. [14] d. novakovic, a. isomäki, b. pleunis, s.j. fraser-miller, l. peltonen, t. laaksonen, c.j. strachan. understanding dissolution and crystallization with imaging : a surface point of view. mol. pharm. 15 (2018) 5361–5373. [15] d. novakovic, l. peltonen, a. isomäki, s. j. fraser-miller, l. h. nielsen, t. laaksonen, c. j. strachan. surface stabilization and dissolution rate improvement of amorphous compacts with thin polymer coatings: can we have it all?. mol. pharm. 17 (2020) 1248-1260. [16] s.b. murdande, m.j. pikal, r.m. shanker, r.h. bogner. aqueous solubility of crystalline and amorphous drugs: challenges in measurement. pharm. dev. technol. 16 (2011) 187–200. [17] s. svanbäck, h. ehlers, o. antikainen, j. yliruusi. on-chip optofluidic single-particle method for rapid microscale equilibrium solubility screening of biologically active substances. anal. chem. 87 (2015) 5041–5045. [18] j. štukelj, s. svanbäck, m. agopov, k. löbmann, c.j. strachan, t. rades, j. yliruusi. direct measurement of amorphous solubility. anal. chem. 91 (2019) 7411–7417. [19] m. savolainen, a. heinz, c. strachan, k. c. gordon, j. yliruusi, t. rades, n. sandler. screening for differences in the amorphous state of indomethacin using multivariate visualization. eur. j. pharm. sci., vol. 30 (2007) 113–123. [20] s.a. surwase, j.p. boetker, d. saville, b.j. boyd, k.c. gordon, l. peltonen, c.j. strachan. indomethacin: new polymorphs of an old drug. mol. pharm. 10 (2013) 4472–4480. [21] j.e. patterson, m.b. james, a.h. forster, r.w. lancaster, j.m. butler, t. rades. the influence of thermal and mechanical preparative techniques on the amorphous state of four poorly soluble compounds. j. pharm. sci. 94 (2005) 1998–2012. [22] j. štukelj, s. svanbäck, j. kristl, c.j. strachan, j. yliruusi. image-based investigation: biorelevant solubility of α and γ indomethacin. anal. chem. 91 (2019) 3997–4003. [23] s. svanbäck. toward accurate high-throughput physicochemical profiling using image-based singleparticle analysis. ph.d. dissertation. university of helsinki, finland, 2016. [24] p. karmwar, k. graeser, k.c. gordon, c.j. strachan. investigation of properties and recrystallisation behaviour of amorphous indomethacin samples prepared by different methods. int. j. pharm. 417 (2011) 94–100. [25] w.l. jorgensen, e.m. duffy. prediction of drug solubility from structure. adv. drug deliv. rev. 54 (2002) 355–366. [26] d. novakovic, j. saarinen, t. rojalin, o. antikainen, s. j. fraser-miller, t. laaksonen, l. peltonen, a. isomäki, c. j. strachan. multimodal non-linear optical imaging for sensitive detection of multiple pharmaceutical solid-state forms and surface transformations. anal. chem. 89 (2017) 11460–11467. ©2020 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://dx.doi.org/10.5599/admet.839 http://creativecommons.org/licenses/by/3.0/ validated lc-esi-ms/ms method for the determination of ivosidenib in 10 µl mice plasma: application to a pharmacokinetic study doi: 10.5599/admet.648 131 admet & dmpk 7(2) (2019) 131-139; doi: http://dx.doi.org/10.5599/admet.648 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper validated lc-esi-ms/ms method for the determination of ivosidenib in 10 µl mice plasma: application to a pharmacokinetic study sreekanth dittakavi 1,2 , rakesh kumar jat 2 , sadanand rangnathrao mallurwar 1 , ravi kumar jairam 1 , ramesh mullangi 1 * 1 drug metabolism and pharmacokinetics, jubilant biosys, industrial suburb, yeshwanthpur, bangalore-560 022, india 2 institute of pharmacy, shri jagdish prasad jhabarmal tibrewala university, jhunjhunu-churu road, chudela, jhunjhunu-333001, rajasthan *corresponding author: e-mail: mullangi.ramesh@jubilantinnovation.com; tel.: +91-80-66628339 received: december 18, 2018; revised: february 04, 2019; available online: march 05, 2019 abstract a simple, selective and rapid lc-esi-ms/ms method has been developed and validated for the quantification of ivosidenib in mice plasma using warfarin as an internal standard (i.s.) as per regulatory guideline. sample preparation was accomplished through a simple protein precipitation process. chromatography of ivosidenib and the i.s. was achieved on an atlantis dc18 column using an isocratic mobile phase comprising 0.2 % formic acid in water and acetonitrile (25:75, v/v) delivered at a flow rate of 1.0 ml/min. lc-ms/ms was operated under the multiple reaction-monitoring mode (mrm) using the electrospray ionization technique in positive ion mode and the transitions of m/z 583.1→186.1 and m/z 309.2→251.3 were used to quantitate ivosidenib and the i.s, respectively. the total chromatographic run time was 2.0 min. linearity was established in the concentration range of 1.10-3293 ng/ml (r 2 >0.99). the intraand inter-day accuracy and precision for ivosidenib in mice plasma were in the range of 5.72-9.91 and 5.90-10.7 %, respectively. ivosidenib was found to be stable on bench-top for 6 h, up to three freeze-thaw cycles, in in-injector for 24 h and for one month at -80 °c. the applicability of the validated method has been demonstrated in a mice pharmacokinetic study. following intravenous (2 mg/kg) and oral (5 mg/kg) administration of ivosidenib to mice, concentrations were quantifiable up to 24 and 48 h, respectively. the bioavailability was 61 %. keywords ivosidenib; lc-ms/ms; method validation; mice plasma; pharmacokinetics; bioavailability introduction isocitrate dehydrogenase isoforms (idh1 and idh2/3 located in cytoplasm and mitochondria, respectively) play a key role in normal cell growth and regulation. all three isoforms are responsible for the production of nicotinamide adenine dinucleotide phosphate (nadph) from nadp + by catalyzing the oxidative decarboxylation of isocitrate to α-ketoglutarate (α-kg) [1]. recent investigations discovered that idh1/2 mutations in 20 % of acute myeloid leukemia (aml) population and rarely in breast, lung and prostate cancers [2]. targeting mutant idh1/2 proved to be an option http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:mullangi.ramesh@jubilantinnovation.com r. mullangi et al. admet & dmpk 7(2) (2019) 131-139 132 for the treatment of aml as a targeted therapy alone or in combination with other antileukemic agents [3]. ivosidenib (ag-120, tibsovo®; fig. 1), is a novel, first-in-class, oral small molecule, which selectively and reversibly inhibits midh1 (ic50: 12 nm) without any idh2 inhibition (at micromolar concentration) along with good cellular potency. it showed tumor regression in mice xenograft model of idh-mutated aml at 50 and 150 mg/kg doses [4]. recently, fda approved ivosidenib as an orphan drug for relapsed or refractory idh1 mutant aml patients as detected by an fda-approved test [5]. in humans as a monotherapy it is administered at a daily dose of 500 mg. following oral administration cmax (maximum concentration in plasma) attained at ~3 h (tmax). it has shown less than dose proportional increase in auc (area under concentration time curve) and cmax from 200-1200 mg. unchanged ivosidenib is majorly (92 %) seen in plasma. ivosidenib is 92-96 % bound to plasma proteins. cyp3a4 mainly involved in its metabolism (n-dealkylation and hydrolysis). it is eliminated as parent up to 67 and 10 % in the feces and urine, respectively. the oral bioavailability was found to be 57 % [5]. lc-ms/ms is a powerful tool for rapid quantitation of drugs in various biological matrices with great accuracy and sensitivity with a shorter run time. preclinical pharmacokinetics has a great influence on the development and investigation of potential preclinical candidates with better advice for the further drug design. pharmacokinetic studies in preclinical species can both provide toxicological and clinical information and direct optimization of drug candidates; as a result, they play necessary parts in drug discovery and development. to date there is no lc-ms/ms method reported for quantification of ivosidenib in any biological matrix. in this paper, we are reporting the development and validation of a sensitive, selective and rapid lc-esi-ms/ms method for quantitation of ivosidenib in mice plasma. the validated method was successfully used in a mice pharmacokinetic study following intravenous and oral administration of ivosidenib at 2.0 and 5.0 mg/kg dose, respectively. materials and methods materials ivosidenib (purity: 98 %) was purchased from angene international limited, england, uk. warfarin (i.s; purity: 98.7 %) was purchased from sigma-aldrich, st. louis, usa. lc-ms-grade acetonitrile and methanol were purchased from jt baker, avantor performance materials, pa, usa. formic acid was purchased from avra synthesis pvt. ltd, hyderabad, india. the control balb/c mice k2.edta plasma sample was procured from animal house, jubilant biosys. instrumentation and chromatographic conditions a shimadzu hplc prominence (shimadzu, japan) coupled with sciex 6500 triple quadrupole (sciex, redwood city, ca, usa) mass spectrometer was used for all analyses. the instrument was controlled using analyst software (version 1.6.2). ivosidenib and the i.s. were eluted using an isocratic mobile phase, which is a mixture of 0.2 % formic acid in water and acetonitrile (25:75, v/v) and chromatographed on an atlantis dc18 column (50  4.6 mm, 3 m) maintained at 40 ± 1 °c. the flowrate was 1.0 ml/min. the mass spectrometer was operated in the multiple reaction mode (mrm) with positive electrospray ionization for the quantitation of ivosidenib and the i.s. ionization was conducted by applying a voltage of 5500 v and source temperature was set at 500 °c. for analyte and the is the optimized source parameters viz., curtain gas, gs1, gs2 and cad were set at 55, 55, 65 and admet & dmpk 7(2) (2018) 131-139 quantitation of ivosidenib in mice plasma by lc-esi-ms/ms doi: 10.5599/admet.648 133 10 psi. the compound parameters viz., declustering potential (dp), entrance potential (ep), collision energy (ce) and collision cell exit potential (cxp) 78, 10, 36, and 5 v for ivosidenib and 61, 10, 20, and 18 v for the is. the mass transition m/z (precursor ion  product ion) 583.1186.1 and 309.2251.3 were monitored for ivosidenib and the i.s, respectively. quadrupole q1 and q3 were set on unit resolution. the dwell time was 100 msec. preparation of stock and standard solutions primary stock solutions of ivosidenib for preparation of calibration curve (cc) and quality control (qc) samples were prepared from separate weighing. the primary stock solution of ivosidenib (1144 µg/ml) and the i.s. (1000 µg/ml) were made in methanol:water (80:20, v/v) and dmso, respectively. the primary stock solution of ivosidenib and the i.s. were stored at -20°c, which were found to be stable for 45 days. ivosidenib primary stock solution was successively diluted in methanol:water (80:20, v/v) to prepare secondary stock and working stock solutions, which were used to prepare ccs and qcs. working stock solutions were stored approximately at 4 °c for ten days. a working stock of the i.s. solution (100 ng/ml) was prepared in methanol. samples for the determination of precision and accuracy were prepared by spiking control mice plasma in bulk with ivosidenib at appropriate concentrations: 1.10 ng/ml (lower limit of quantitation quality control, lloq qc), 3.29 ng/ml (low quality control, lqc), 1715 ng/ml (medium quality control, mqc) and 2607 ng/ml (high quality control, hqc) and were stored at -80 ± 10 °c until analysis. sample preparation to an aliquot of 10 µl plasma, 500 µl of methanol containing 100 ng/ml i.s. was added and vortex mixed for 2 min. thereafter the contents were centrifuged for 5 min at 14,000 rpm in a refrigerated centrifuge (eppendorf 5424r) maintained at 5 °c. clear supernatant (200 µl) was transferred into vials and 2.0 µl was injected onto lc-ms/ms system for analysis. validation procedures the validation experiments were performed in accordance with the us food and drug administration guideline [6]. the selectivity of the method was determined by the presence of interfering peaks from six individual drug-free mice plasma samples at the retention times of ivosidenib and the i.s. the autoinjector carryover was determined by injecting the highest calibration standard, followed by injection of blank plasma samples. the lloq was determined as the concentration that has a precision of <20 % of the relative standard deviation and accuracy between 80-120 % of the theoretical value. the response of the blanks was then compared to that of the lloq. recovery was determined at lqc and hqc. recovery for ivosidenib was calculated by comparing the mean peak response of pre-extraction spiked samples (spiked before extraction; n=6) to that of neat samples (n=6) at each qc level. matrix effect was assessed by comparing the analyte mean peak areas at lqc and hqc concentration after extraction with the mean peak areas for post extracted blank plasma samples spiked with analyte at equivalent concentrations. recovery and matrix effect for the i.s. was assessed at single concentration (100 ng/ml). the precision and accuracy of the method were evaluated by measuring the four qc samples r. mullangi et al. admet & dmpk 7(2) (2019) 131-139 134 (lloq qc, lqc, mqc and hqc), which were prepared on each validation day (n=6 each). inter-day precision was assessed on four separate days. interand intra-day precisions were determined by calculating percent relative standard deviation (%rsd) that should be 15 % for all qcs except for lloq qc where it should be 20 %. the interand intra-day accuracy expressed as percent relative error (%re) was calculated by comparing the measured concentration with the nominal value and deviation was limited within ±15 % except for lloq qc where it should be 20 %. freeze-thaw stability following three freeze-thaw cycles was evaluated (one day duration between each freeze-thaw cycle and ivosidenib spiked plasma samples were stored in -80 ± 10 °c between freeze/thaw cycles). the plasma samples were thawed at room temperature for 1 h and returned to the freezer. bench-top stability was assessed by analyzing samples that had been kept at ambient temperature (25 ± 1 °c) for 6 h. long-term stability was performed by analyzing samples that had been stored at -80 ± 10 °c for 30 days. the stability of ivosidenib and the i.s. in the injection solvent was determined periodically by injecting replicate preparations of processed plasma samples for up to 24 h (in the auto-sampler at 5 °c) after the initial injection. the peak-areas of the analyte and the i.s. obtained at initial cycle were used as the reference to determine the stability at subsequent points. these stability samples were processed and quantified against freshly spiked calibration curve standards along with freshly spiked qc samples. samples were considered to be stable if assay values were within the acceptable limits of accuracy (±15 % re) and precision (20 % rsd). to evaluate the effect of dilution over the calibration range, the accuracy and precision of dilution control samples at 9881 ng/ml (n=6; 3 times of the uloq) were assessed by performing a 5and 10fold dilution. pharmacokinetic study in mice all the animal experiments were approved by institutional animal ethical committee (iaec/jdc/2017/133). male balb/c mice (n=24) were procured from vivo biotech, hyderabad, india. the animals were housed in jubilant biosys animal house facility in a temperature (22 ± 2 °c) and humidity (30-70 %) controlled room (15 air changes/hour) with a 12:12 h light:dark cycles, had free access to rodent feed (altromin spezialfutter gmbh & co. kg., im seelenkamp 20, d-32791, lage, germany) and water for one week before using for experimental purpose. following 4 h fast (during the fasting period animals had free access to water) animals were divided into two groups (n=12/group). group i animals (26-30 g) received ivosidenib orally at 5.0 mg/kg (suspension formulation comprising 0.5 % tween-80 and methyl cellulose (0.5 % in water); strength: 0.5 mg/ml; dose volume: 10 ml/kg), whereas group ii animals (27-31 g) received ivosidenib intravenously [5 % dmso, 5 % solutol:absolute alcohol (1:1, v/v) and 90 % of normal saline; strength: 0.2 mg/ml; dose volume: 10 ml/kg] at 2.0 mg/kg dose. post-dosing serial blood samples (40 µl, sparse sampling was done and at each time point three mice were used for blood sampling) were collected using micropipettes (microcaps®; catalogue number: 1-000-0500) through tail vein into polypropylene tubes containing k2edta solution as an anti-coagulant at 0.25, 0.5, 1, 2, 4, 8, 10, 12, 24, 30, 36 and 48 h (for oral study) and 0.12, 0.25, 0.5, 1, 2, 4, 8, 10, 12, 24 and 36 h (for intravenous study). plasma was harvested by centrifuging the blood using biofuge (hereaus, germany) at 1760 g for 5 min and stored frozen at -80 ± 10 °c until analysis. animals were allowed to access feed 2 h post-dosing. admet & dmpk 7(2) (2018) 131-139 quantitation of ivosidenib in mice plasma by lc-esi-ms/ms doi: 10.5599/admet.648 135 results and discussion liquid chromatography and ms/ms conditions to avoid the potentially co-eluting peaks from blank plasma, which will influence the ionization efficiency of ivosidenib and the i.s, the mobile phase was optimized along with chromatographic conditions. in order to get chromatograms with good separation, peak shape and resolution for ivosidenib and the i.s, feasibility of various mixture(s) of solvents such as acetonitrile and methanol using different buffers such as ammonium acetate, ammonium formate and formic acid along with altered flow-rates (in the range of 0.8-1.2 ml/min) were tested. the resolution of ivosidenib and the i.s. was achieved with 0.2 % formic acid:acetonitrile (25:75, v/v) at a flow rate of 1.0 ml/min. atlantis dc18 column (100  4.6 mm, 3 µm) was best suited to provide the symmetric peaks and baseline separation of ivosidenib and the i.s. with the retention time of 0.96 and 1.02 min, respectively. the total chromatographic run time was 2.0 min. to obtain optimum ionization and sensitivity 100 ng/ml ivosidenib solution was directly injected into mass spectrometer and electrospray ionization (esi) full scans were carried out both in positive and negative ion detection modes. positive esi mode was chosen as it provided good intensity signals and formed protonated [m+h] + at m/z 583.1. this precursor ion (q1) was selected for fragmentation in ms/ms mode to obtain most intense and consistent product ion (q3: m/z 186.1) by optimizing the declustering potential, collision energy and collision cell exit potential. subsequently, various gases like nebulizer gas (gs1), auxiliary gas (gs2), collision gas and source temperature were optimized to obtain adequate and reproducible response. mrm mode was used to obtain better selectivity, with dwell time of 100 ms for each transition. fragment ions at m/z 476, 214, 186 and 158 were produced as prominent product ions for ivosidenib due to sequential loss of m/z 107, 262, 28 and 28 from ivosidenib. the postulated fragmentation pattern of ivosidenib is shown in figure 1. due to nonavailability of deuterated ivosidenib to use it as an i.s, we have used warfarin as an i.s. and found to be the best for present purpose based on the chromatographic elution, ionization and reproducible and good extraction efficiency. for quantitation purpose m/z 583.1 precursor ion to the m/z 186.1 and m/z 309.2 precursor ion to the m/z 251.3 was used for ivosidenib and the i.s, respectively. figure 1. fragmentation pattern of ivosidenib r. mullangi et al. admet & dmpk 7(2) (2019) 131-139 136 validation results selectivity among the tested drug-free mice blank plasma samples no sample shown any interference at the retention times of the ivosidenib and the i.s. indicating that the method is selective (figure 2). (a) mice blank plasma (b) i.s. (c) ivosidenib lloq (d) 0.25 h plasma sample figure 2. typical mrm chromatograms of (a) mice blank sample (b) mice blank plasma spiked with i.s. (c) mice blank plasma spiked with ivosidenib at lloq (1.10 ng/ml) and (d) 0.25 h plasma sample showing ivosidenib (1953 ng/ml) peak obtained following 5.0 mg/kg oral dose of ivosidenib to mice. sensitivity and carry over the lowest limit of reliable quantification for the analyte was set at the concentration of the lloq. the precision and accuracy at lloq concentration were found to be 6.12 and 91.0 % for ivosidenib. there was no carry-over produced by the highest calibration sample on the following injected blank plasma extract sample. recovery the mean ± s.d recovery at lqc and hqc was found to be 66.5 ± 8.97 and 76.1 ± 1.43 %, respectively. the recovery of the i.s. was 100 ± 8.99 %. matrix effect six different lots of plasma samples, spiked with analyte concentration levels at lqc and hqc levels were analyzed. matrix effect for ivosidenib at lqc and hqc was 0.99 ± 6.82 and 1.04 ± 1.70 %, respectively. the matrix effect for the i.s. was 1.09 ± 4.13%. the results have shown that the precision and accuracy for analyzed samples were within acceptance range. overall it was found that there is no impact on the ionization of analyte and the i.s. calibration curve the plasma calibration curve was constructed using eight calibration standards (viz., 1.10, 2.20, 21.9, 109, 548, 1097, 2264 and 3293 ng/ml). under the analytical conditions used in the present study, the calibration curves of ivosidenib were found to be linear. calibration curve was prepared by determining the best fit of peak-area ratios (peak area analyte/peak area is) versus concentration, and fitted to the y = mx + c using weighing factor (1/x 2 ). the average slope and intercept values were admet & dmpk 7(2) (2018) 131-139 quantitation of ivosidenib in mice plasma by lc-esi-ms/ms doi: 10.5599/admet.648 137 found to be 0.001173 and 0.00014, respectively. the average regression (n=4) was found to be 0.995. the lowest concentration with the rsd < 20 % was taken as lloq and was found to be 1.10 ng/ml. the % accuracy observed for the mean of back-calculated concentrations for four calibration curves for ivosidenib was within 90.8-110; while the precision (%rsd) values ranged from 2.08-9.42. accuracy and precision a summary of the accuracy and precision data for intraand inter-day is listed in table 1. the assay values on both the occasions (intraand inter-day) were found to be within the accepted variable limits. table 1. precision, accuracy, recovery and matrix effect for determination of ivosidenib quality controls in mice plasma lloq qc (1.10 ng/ml) lqc (3.29 ng/ml) mqc (1715 ng/ml) hqc (2607 ng/ml) intra-day (n=6) mean ± s.d 1.09 ± 0.11 3.20 ± 0.28 1631 ± 96.2 2582 ± 188 precision (%rsd) 9.91 8.86 5.72 7.31 accuracy (%re) 0.99 0.97 0.98 0.99 inter-day (n=24) mean ± s.d 1.10 ± 0.12 3.19 ± 0.27 1634 ± 96.4 2588 ± 188 precision (%rsd) 10.7 8.48 5.90 7.25 accuracy (%re) 1.00 0.97 0.95 0.99 %re: relative error (measured value/actual value ×100) rsd: relative standard deviation (sd × 100/mean) sd: standard deviation stability the predicted concentrations for ivosidenib at 3.29 ng/ml (lqc) and 2607 ng/ml (hqc) samples in a battery of stability tests namely in-injector (24 h), bench-top (6 h), repeated three freeze/thaw cycles and freezer stability at -80  10 °c for 30 days are shown in table 2. the results were found to be within the assay variability limits during the entire process. table 2. stability data of ivosidenib quality controls in mice plasma nominal concentration (ng/ml) stability mean ± sd a n = 6 (ng/ml) accuracy (%) b precision (% rsd) 3.29 0 h 6 h (bench-top) 24 h (in-injector) freeze-thaw 30 day at -80°c 3.21 ± 0.39 3.13 ± 0.40 2.98 ± 0.20 3.39 ± 0.50 3.32 ± 0.33 na 97.6 92.9 106 103 12.1 12.6 6.87 14.8 10.1 2607 0 h 6 h (bench-top) 24 h (in-injector) freeze-thaw 30 day at -80°c 2687 ±156 2709 ±76.4 2357 ±108 2632 ± 110 2609 ± 170 na 101 87.7 98.0 97.1 4.78 2.82 4.57 4.18 6.50 a back-calculated plasma concentrations b (mean assayed concentration / mean assayed concentration at 0 h) × 100 %re: relative error (measured value/actual value × 100) rsd: relative standard deviation (sd × 100/mean) dilution effect the accuracy and precision of the nominal concentration of the diluted plasma samples were within 0.79 % and 3.34 %, which show the ability to dilute samples up to a dilution factor of 10-fold in a linear fashion. r. mullangi et al. admet & dmpk 7(2) (2019) 131-139 138 pharmacokinetic study the sensitivity and specificity of the validated assay was found to be sufficient for accurately characterizing the pharmacokinetics of ivosidenib in mice plasma following intravenous and oral administration. the estimates of the pharmacokinetic parameters are summarized in table 3. the mean plasma concentration versus time profiles for ivosidenib following single oral and intravenous route is depicted in figures 3a and 3b, respectively. following intravenous administration, the plasma concentrations decreased in a mono-exponential manner. ivosidenib was quantifiable up to 24 h postdosing by intravenous route. the auc0- (area under curve from time zero to infinity) was found to be 4967 ng  h/ml. the in vivo clearance (cl) and volume of distribution (vd) was found to be 6.85 ml/min/kg and 1.70 l/kg, indicating that ivosidenib has low clearance and moderate volume of distribution. after oral administration of ivosidenib to mice, maximum concentration in plasma (cmax: 1668 ng/ml) attained at 0.50 h (tmax) indicating rapid absorption from gastrointestinal tract. ivosidenib was quantifiable up to 36 h post oral dosing. the auc(0-) was found to be 7462 ng  h/ml. the terminal half-life was 2.87 and 4.06 h, by intravenous and oral routes, respectively. the absolute oral bioavailability was 61 %. table 3. pharmacokinetic parameters of ivosidenib in mice parameter intravenous oral dose (mg/kg) 2.0 5.0 auc0-∞ (ng × h/ml) 4967 7462 c0/cmax (ng/ml) 6085 1668 tmax (h) --0.50 t1/2 (h) 2.87 4.06 cl (ml/min/kg) 6.85 -- vd (l) 1.70 -- a) b) time (h) m e a n  s .d ( n g /m l ) c o n c e n tr a ti o n o f iv o si d e n ib i n m ic e p la sm a 0 4 8 12 16 20 24 1 10 100 1000 10000 time (h) m e a n  s .d ( n g /m l ) c o n c e n tr a ti o n o f iv o si d e n ib i n m ic e p la sm a 0 4 8 12 16 20 24 28 32 36 40 44 48 1 10 100 1000 10000 figure 3. mean plasma concentration-time profile of ivosidenib following (a) intravenous and (b) oral dosing to mice conclusions a rapid, simple, sensitive and selective lc-esi-ms/ms method was developed and validated for the determination of ivosidenib using 10 µl mice plasma. this method demonstrates good accuracy and precision and meets us fda validation criteria. to our knowledge, this is the first report of an lcadmet & dmpk 7(2) (2018) 131-139 quantitation of ivosidenib in mice plasma by lc-esi-ms/ms doi: 10.5599/admet.648 139 ms/ms method for quantification of ivosidenib in plasma and successfully applied for a pharmacokinetic study in mice. conflict of interest: the authors are scientists at jubilantbiosys. references [1] z.j. reitman, h. yan h. isocitrate dehydrogenase 1 and 2 mutations in cancer: alterations at a crossroads of cellular metabolism. journal of national cancer institute 102 (2010) 932-941. [2] e. stein, k. yen. targeted differentiation therapy with mutant idh inhibitors: early experiences and parallels with other differentiation agents. annual reviews in cancer biology 1 (2017) 379401. [3] m.j. buege, a.j. dipippo, c.d. dinardo. evolving treatment strategies for elder leukemia patients with idh mutations. cancers 10 (2018) e187. [4] j. popovici-muller, r.m. lemieux, e. artin, j.o. saunders, f.g. salituro, j. travins et al. discovery of ag-120 (ivosidenib): a first-in-class mutant idh1 inhibitor for the treatment of idh1 mutant cancers. acs medicinal chemistry letters 9 (2018) 300-305. [5] tibsovo (ivosidenib) prescribing information, agios pharmaceuticals, inc, july 2018. available at www.accessdata.fda.gov. accessed november 9, 2018. [6] us dhhs, fda, cder, cvm. guidance for industry: bioanalytical method validation. u.s. department of health and human services, food and drug administration, center for drug evaluation and research (cder), center for veterinary medicine (cvm), (2018), rockville, md, usa. ©2019 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://www.accessdata.fda.gov/ http://creativecommons.org/licenses/by/3.0/ nicotinic acid in nanocontainers. encapsulation and release from ion exchangers doi: 10.5599/admet.626 76 admet & dmpk 7(1) (2019) 76-87; doi: http://dx.doi.org/10.5599/admet.626 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper nicotinic acid in nanocontainers. encapsulation and release from ion exchangers heinrich altshuler 1 *, elena ostapova 1 , olga altshuler 1,2 , galina shkurenko 1 , natalya malyshenko 1 , sergey lyrschikov 1 , roman parshkov 1 1 laboratory of supramolecular polymer chemistry, institute of coal chemistry and material science, federal research center of coal and coal chemistry, siberian branch of russian academy of sciences, kemerovo, 650000, russian federation; 2 institute of basic sciences, kemerovo state university, kemerovo, 650000, russian federation *corresponding author: e-mail: altshulerh@gmail.com ; tel.: +7-905-966-6685 received: october 16, 2018; revised: december 06, 2018; published: december 26, 2018 abstract the paper is devoted to the study of the ion exchange encapsulation of nicotinic acid in nanocontainers on polymer matrices. dowex-50 cation exchanger, sulphonated polymer based on metacyclophanoctol, polymer zirconium phosphate, and strongly basic dowex-1 anion exchanger are used as polymer matrices. it was confirmed that commercial ion exchangers can encapsulate up to 0.64 g of nicotinic acid per gram of polymer. the high elution rate of nicotinic acid from nanocontainers via the ion exchange mechanism makes it possible to achieve the desired pharmacokinetics of drug release in vivo. keywords nicotinic acid; encapsulation; ion exchange polymers; pharmacokinetics introduction currently, research is being conducted on the design of polymeric nanocontainers for drug encapsulation [1]. polymeric nanocontainers facilitate establishment of the desired pharmacokinetics, i.e., a given drug release timeframe, decreased frequency of administration and dose, and targeted delivery of molecules to the disease site. the possibilities for preserving and storing dosage forms are practically unlimited. the therapeutic agent is used as drug delivery vector to the disease-site molecular target. thus, the drugs encapsulation in nanocontainers creates unlimited possibilities for storing dosage forms and the change in their pharmacokinetic properties. it is important to consider the processes of encapsulation and the process of releasing substances from molecular containers immobilized on matrices of network ion exchange polymers. dowex type ion exchangers based on polystyrene matrices are widely used in the medical industry for the preparation of drinking and pyrogen-free water, in the production of antibiotics, in a clinical setting for regulating of human water-salt balance and in the treatment of hyperkalemia by oral administrating [2-5]. earlier we investigated the encapsulation of benzocaine and the kinetics of its desorption from sulphonated polycalixarene and cu-23 30/100 macroporous sulphocathionite [6,7]. it was suggested to use the kinetic characteristics studied to model the drug-release pharmacokinetics from a http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:altshulerh@gmail.com admet & dmpk 7(1) (2019) 76-87 encapsulation and release of nicotinic acid doi: 10.5599/admet.626 77 nanocontainer upon oral administration [7]. nicotinic acid (3-pyridinecarboxylic acid, the chemical formula c6h5o2n, synonyms  niacin, vitamin b3, vitamin pp [8]) plays an important role in the human metabolism. 3-pyridinecarboxylic acid is a prosthetic group of redox coenzymes codohydrase i (nicotine adenine dinucleotide) and codohydrase ii (nicotinamide adenine dinucleotide phosphate) which are key components to cellular metabolic reactions in biological systems [8,9]. nicotinic acid as a drug is prescribed for the prevention and treatment of pellagra, with spasms of limb vessels and brain, ulcers, the neuritis of the facial nerve, infectious and gastrointestinal diseases [8]. the world demand for nicotinic acid and its derivatives continuously grows. the forecast for the year of 2020 is 100000 tons [10]. free nicotinic acid is readily absorbed in all parts of gastrointestinal track. we assume that its encapsulation in ion exchangers will allow targeted delivery of the drug substance only to the stomach or intestines at oral administration. the behavior of nicotinic acid in vivo during passage through the gastrointestinal tract as well as by its ion exchange encapsulation and release from molecular containers is mainly determined by the acid-base equilibria (scheme 1). see below acid-base equilibria in aqueous solutions of nicotinic acid. nicotinic acid ionization n coon h+ coon h+ cooh l hl h2l k1 k2 +h+ +h+ scheme 1. acid-base equilibria in aqueous solutions of nicotinic acid the molar fractions of various ionic forms of nicotinic acid vs ph of the solution calculated from the equilibrium constants (log k1 = 4.81 and log k2 = 2.07 [11]) are shown in figure 1. figure 1. molar fractions of ionic forms of nicotinic acid xl (1), xhl (2), xh 2 l (3) vs ph of the solution. heinrich altshuler et al. admet & dmpk 7(1) (2019) 76-87 78 free hydrochloric acid is secreted in the stomach of a living organism. the ph of the medium is usually equal to 2. as can be seen from figure 1 the concentration of nicotinic acid cations (xh2l) is predominant at ph  2. nicotinic acid is represented by the anionic form (l) in the intestine where ph > 6. it is desirable to encapsulate and release the nicotinic acid from nanocontainer in the form of cations or anions for drug targeted delivery in vivo. the aim of this paper is to study the encapsulation and the release of nicotinic acid in ionized forms proceeding in molecular containers immobilized on matrices of network ion exchange polymers. encapsulation of nicotinic acid in a cation form is carried out in nanocontainers on matrices of crosslinked polystyrene sulphonate (dowex-50 cation exchanger), sulphonated polymer based on metacyclophanctol and polymer zirconium phosphate by the cation exchange reaction: ,   hсoohpyhсoohpyhh (1) h + py  cooh  cation of nicotinic acid (cation of protonated 3-pyridinecarboxylic acid, h2l), the line indicates the polymer phase. encapsulation of nicotinic acid in anion form is carried out on the matrix of cross-linked strong basic dowex-1 anion exchanger by the anion exchange reaction .   clсoopyсoopyсl (2) here py  coo   anion of nicotinic acid (anion of 3-pyridinecarboxylic acid, l). the process (1) of the sorption of nicotinic acid cations (h2l) on the cation exchanger by the ion exchange mechanism has been studied at the concentration of h2l cations exceeding the concentration of the l anions and the hl molecules in the solution, i.e. at ph < 2 (figure 1). the process (2) of the chloride anion exchange of by the nicotinic acid anion on the anion exchanger is studied at ph > 6, when the concentration of l anions in solution exceeded the concentration of h2l cations and hl molecules (figure 1). experimental materials nicotinic acid (obtained from j.s.c. «organica», russian federation) contains 99.0 % 3-pyridinecarboxylic acid and meets the requirements of the international pharmacopoeia [12]. dowex-50 cation exchanger and strongly basic dowex-1 anion exchanger were purchased from sigma-aldrich. dowex-50, cation exchanging sulphonated copolymer of styrene with 8 % divinylbenzene, contains only one type of ionogenic groups, viz. sulpho group (so3h). the total dynamic ion exchange capacity is 5.2 mequiv/g of the h-form of a dry polymer. a sulphonated polymer based on metacyclophanoctol was prepared according to the procedure described previously [13]. it has a gel structure and contains two types of ionogenic groups: phenolic -oh and -so3h groups. the total dynamic ion exchange capacity of sulphonated polycalixarene is 5.65 mequiv./g. its capacity with respect to sulpho groups is 2.45 mequiv./g [13]. polymer zirconium phosphate is prepared according to the procedure described previously [14]. the dynamic ion exchange capacity with respect to 0.1 m nacl is 1.05 mequiv./g. dowex-1 is a strong basic anion exchanger. it contains benzyltrimethylammonium groups at the matrix of the styrene with 8 % divinylbenzene copolymer. the total dynamic ion exchange capacity of the anion exchanger with respect to 0.1 м кон is equal to 2.7 mequiv. per 1 g of the cl-form of the dry polymer. admet & dmpk 7(1) (2019) 76-87 encapsulation and release of nicotinic acid doi: 10.5599/admet.626 79 methods the encapsulation of nicotinic acid in nanocontainers was carried out by the ion exchange sorption. to encapsulate nicotinic acid into nanocontainers on the matrices of the cation exchangers and to determine the dynamic capacity of the cation exchangers with respect to nicotinic acid, 0.01 m solution of nicotinic acid in 0.01 m hydrochloric acid was passed through an ion exchange column containing cation exchanger in h-form until the solution composition at the inlet and outlet from the column became equal. to encapsulate nicotinic acid into a nanocontainer on the anion exchanger, 0.01-0.05 m solution of the potassium salt of nicotinic acid was passed through an ion exchange column containing dowex-1 in the clform until the nicotinic acid concentration at the inlet and outlet from the column became equal. the dynamic ion exchange capacity with respect to nicotinic acid was calculated as the arithmetic mean of 7 measurements on each polymer. to elute nicotinic acid from nanocontainers on the cation exchanger matrices, a 0.01 m hcl solution was passed through the cation exchanger layer containing nicotinic acid until the nicotinic acid in the eluate disappeared. when the nicotinic acid was eluted from the dowex-1 anion exchanger through the ion exchanger layer containing nicotinic acid, 0.01-0.1 m aqueous nacl solutions was passed. the nicotinic acid was crystallized from the obtained eluates by their evaporation and precipitation at ph 3.3-3.6 via the procedure [15]. the selection of spherical granules and their size determination for kinetic studies was carried out on the imc 100 × 50, a microscope. the size distribution of polymer granules is described by the gaussian function. the radius of spherical particles calculated as the arithmetic mean of 1000 granules sizes was equal to (1.8±1.2)∙10 -4 m for dowex-50 cation exchanger and (2.2±0.6)∙10 -4 m for dowex-1 anion exchanger. the errors were calculated with 0.95 confidence level. the kinetics of the nicotinic acid release (elution) from nanocontainers was studied at 298 k using the dynamic thin-layer method [16]. the fact that the solution flowed through a thin layer of the ion exchanger at high speed is a special feature of the method [16]. the infinite volume of 0.01 m aqueous hcl solution or pure water were passed through the layer of dowex-50 cation exchanger in the form of nicotinic acid cations. the infinite volume of 0.1 m aqueous nacl solution was passed through the layer of dowex-1 anion exchanger in the form of nicotinic acid anions. after a certain period of time the concentrations of nicotinic acid were determined using an sf-46 spectrophotometer at λ = 262.7 nm in a buffer solution with ph 6.9. the degree of conversion was calculated using the formula f = mt/m∞, where mt is the amount of nicotinic acid desorbed up to time t and m∞ is the amount of nicotinic acid desorbed up to infinite time. 13 c nmr solid-state spectra were obtained on a bruker avance ii+ 300wb instrument at the operating frequencies 75.48 mhz ( 13 c). fourier ir spectra of nicotinic acid in tablets with kbr were obtained on the "infralum ft-801" spectrometer. molecular design and calculations of the formation energies of nanocontainer structures containing encapsulated nicotinic acid were carried out by the pm7 method in the framework of the mopac2016 program taken from the web site: www.openmopac.net [17] on a computer based on the intel (r) core (tm) processor i5-2310 cpu @ 2.9 ghz 2.9 ghz. http://www.openmopac.net/ heinrich altshuler et al. admet & dmpk 7(1) (2019) 76-87 80 results and discussion nanocontainers elementary units of network polymers, which act as nanocontainers for nicotinic acid are shown in figure 2. figure 2. the structures of nanocontainers containing nicotinic acid in (a) dowex-50 cation exchanger, (b) sulphonated polymer based on metacyclophanoctol, (c) polymer zirconium phosphate, (d) dowex-1 anion exchanger. the structures have a minimum of internal energy within the mopac2016 program. elementary units of dowex-50 cation exchanger (figure 2(a)) and dowex-1 anion exchanger (figure 2(d)) contain hydrophobic baskets consisting of alkylaromatic chains with ionogenic sulphoor tetraalkylammonium group, respectively. the elementary unit of sulphonated polymer based on metacyclophanoctol (figure 2(b)) contains a macrocyclic cavity and two hydrophilic rims. the lower rim includes eight hydroxyl groups; the upper rim consists of four ionogenic sulpho groups (one so3h group per benzene ring). polymer zirconium phosphate is a matrix with a layered structure, which is capable of locating small guest molecules in the interlayer space. the elementary unit of this polymer is shown in figure 2(c). as shown via quantum-chemical calculations, the elementary units of the investigated polymers complementarily interact with ionized nicotinic acid molecules (cations of protonated 3pyridinecarboxylic acid or anions 3-pyridinecarboxylate) by host-guest type. encapsulation the experimental values of the dynamic exchange capacity of polymers at the encapsulation of nicotinic acid are given in the table 1. it is found out that 0.33 and 0.64 grams of nicotinic acid are encapsulated in one gram of dowex-1 and admet & dmpk 7(1) (2019) 76-87 encapsulation and release of nicotinic acid doi: 10.5599/admet.626 81 dowex-50, respectively. the exchange capacities of the dowex-50 cation exchanger and the sulphonated polymer based on metacyclophanoctol with respect to nicotinic acid (table 1) correspond to the contents of strongly acidic sulpho groups in polymers. the capacity of the dowex-1 anion exchanger with respect to nicotinic acid corresponds to the content of benzyltrimethylammonium ionogenic groups in the polymer. table 1. capacities of ion exchange polymers with respect to nicotinic acid (mean ± error, 0.95 confidence level). ion exchange polymer equilibrium solution dynamic exchange capacity (content of nicotinic acid per one gram of dry polymer), mmol/g dowex-50 cation exchanger 0.01 m nicotinic acid in 0.01 m hcl 5.2 ± 0.1 sulphonated polymer based on metacyclophanoctol 0.01 m nicotinic acid in 0.01 m hcl 2.45 ± 0.05 polymer zirconium phosphate 0.01 m nicotinic acid in 0.01 m hcl 0.18 ± 0.02 dowex-1 anion exchanger potassium nicotinate in water, 0.01 m 2.70 ± 0.05 13 c nmr spectra of the solid samples of nicotinic acid sulfate (sulfate of protonated 3-pyridinecarboxylic acid), sulphonated polymer based on metacyclophanoctol and dowex-50 cation exchanger are given in figure 3. 13 c nmr spectra of the solid samples of the potassium salt of the nicotinic acid (potassium 3pyridinecarboxylate), dowex-1 anion exchanger are given in figure 4. as can be seen from figure 3, there is a resonance line corresponding to the chemical shift of  = 165 ppm (c = o [18]) in the spectrum (1) of nicotinic acid sulfate and in the spectra (2), (4) of cation exchangers containing encapsulated nicotinic acid. in the 13 c nmr spectrum (figure 4) of the potassium salt of the nicotinic acid as well as in the spectrum of the dowex-1 filled with nicotinic acid anions, there is a resonance line next to  = 170 ppm (chemical shift of carboxylate anions [18]). there is no such line in spectrum of the dowex-1, free of nicotinic acid. thus, it follows from the 13 c spectra that the encapsulated nicotinic acid is actually contained in ion exchange polymers in an ionized forms as protonated 3-pyridinecarboxylic acid cations or as anions of 3pyridinecarboxylate. release water solutions simulating the electrolyte composition of the human gastrointestinal tract are used for the release (elution) of the nicotinic acid from nanocontainers. the release of nicotinic acid from cation exchangers is carried out using hcl solution (at ph  2). this simulates the electrolyte composition of the stomach. the release of the nicotinic acid from the anion exchanger is carried out using aqueous solutions of nacl (at ph > 6) which simulates the electrolyte composition of the intestine. it has been found that the dynamic exchange capacities during encapsulation and the release of nicotinic acid are equal to each other. it is clear that the encapsulation of the nicotinic acid proceeds in accordance with the stoichiometric ion exchange reactions (1) and (2). the release of nicotinic acid from nanocontainers on matrices of ion exchange polymers can be described by the ion exchange reactions (3) and (4). ,сoohpyhhhсoohpyh   (3) .   сoopyсlclсoopy (4) nicotinic acid that is precipitated from eluates contains 99 % of the 3-pyridinecarboxylic acid. the melting point of 235–236 с corresponds to these data [19]. elemental analyses (%) of the precipitated nicotinic acid are c, 57.9 ± 0.5; h, 4.1 ± 0.1; n, 11.2 ± 0.5; o, 25.0 ± 1.0 which corresponds within the error of determination to the calculated c7h5no4 formula (c, 58.30; h, 4.08; n, 11.34; o, 25.91). thus, the heinrich altshuler et al. admet & dmpk 7(1) (2019) 76-87 82 composition of the product released from the encapsulated state coincides with the elemental composition of the nicotinic acid. figure 3. 13 c nmr spectra of solid samples of (1) nicotinic acid sulfate, (2) sulphonated polymer based on metacyclophanoctol containing nicotinic acid, (3) sulphonated polymer based on metacyclophanoctol, (4) dowex-50 cation exchanger containing nicotinic acid, (5) dowex-50 cation exchanger. figure 4. 13 c nmr spectra of solid samples of (1) of potassium salt of the nicotinic acid (potassium 3pyridinecarboxylate); (2) dowex-1 anion exchanger containing nicotinic acid anions; (3) dowex-1 anion exchanger. admet & dmpk 7(1) (2019) 76-87 encapsulation and release of nicotinic acid doi: 10.5599/admet.626 83 the fourier ir spectrum of the product precipitated from the eluate corresponds (figure 5) to the spectrum [20] of nicotinic acid. it contains the intense vibration bands at 1712 сm – 1 and 1113 сm – 1 belonging to the carbonyl and carboxyl group of the 3-pyridinecarboxylic acid [20,21]. the 13 с nmr spectrum of the product obtained from the eluate coincided with the nmr spectrum of the nicotinic acid. the nicotinic acid is encapsulated, and then completely released from the researched polymers. this is resulted from the material balance of the ion exchange processes (1), (2) of encapsulation and (3), (4) release, nmr and ir spectra, elemental analysis, and melting point of nicotinic acid. nicotinic acid in nanocontainers exists in the cation or anion forms. figure 5. fourier ir spectrum of nicotinic acid precipitated from eluate in tablets with kbr. it is established that dowex-50 cation exchanger, sulphonated polymer based on metacyclophanoctol, polymeric zirconium phosphate as well as a strongly basic dowex-1 anion exchanger were not destroyed, during the process of encapsulating and releasing of the nicotinic acid. the performance characteristics of commercial polymers dowex-1 and dowex-50 were not changed in five cycles of encapsulation and during the release of the nicotinic acid. taking into account the above, the nicotinic acid encapsulation in nanocontainers based on commercial polymers: dowex-50 cation exchanger and dowex-1 anion exchanger expands the possibilities of obtaining prolonged forms of the active substance. kinetics the release vs time profile of nicotinic acid from commercial ion exchangers is shown in figure 6. as can be seen, the best nicotinic acid release profile is achieved on the dowex 50 ion exchanger using 0.01 m hcl as eluent. the research of the kinetics and mechanism of the ion exchange processes (3) and (4) of the nicotinic acid release from nanocontainers is of current interest. the dependences form of the conversion degree on time and the passage of lines through the origin of coordinates (figure 6(a)) in accordance with the known criteria [22] indicate that the mechanism of the nicotinic acid release from ion exchangers using hcl or nacl is controlled by diffusion of exchangeable ions into the polymer; i.e., gel diffusion kinetics of ion exchange occurs. the gel diffusion kinetics of ion exchange in the case of constant diffusion coefficient and spherical symmetry is described [23] by the differential equation: heinrich altshuler et al. admet & dmpk 7(1) (2019) 76-87 84 , 2 2 2                 r c rr c d t с (5) where d is the diffusion coefficient of the component, c is the current concentration of the component in the polymer, and r is the value of radius vector. figure 6. release-time profile of nicotinic acid from ion exchangers. (a) red triangles represent the experimental data of nicotinic acid release from dowex-50 by 0.01 м hcl, green circles represent the experimental data of nicotinic acid release from dowex-1 by 0.1 м nacl. (b) dark blue triangles represent the experimental data of nicotinic acid release from dowex-50 by h2o. error bars show deviation from mean values, n=3. in a monofunctional ion exchanger, the diffusion kinetics of ion exchange in a spherical polymer particle contacting with a solution of invariable composition and infinite volume is described by the known dependence [24]: admet & dmpk 7(1) (2019) 76-87 encapsulation and release of nicotinic acid doi: 10.5599/admet.626 85  2 2 2w 02 2 1 6 1 1 exp / , n f d n t r n        (6) where dw is the interdiffusion coefficient of exchangeable ions in the polymer, r0 is the radius of a polymer particle. the kinetic characteristics of the processes (3) and (4) of nicotinic acid release are shown in figure 7. figure 7. kinetics of the processes of nicotinic acid release from ion exchangers. the solid and dotted lines correspond to calculations using equation (6); the numbers near the lines show the interdiffusion coefficients expressed in m 2 /s. the red triangles represent the experimental data of nicotinic acid release from dowex-50 and green circles represent the experimental data of nicotinic acid release from dowex-1. a comparison of the time dependences of the conversion degree and the experimental data (at 0 < f < 0.5) shows that the kinetics of processes (3) and (4) are determined by a slow diffusion of the components in the polymer phase. this is confirmed by the agreement between the experimental kinetic characteristics and theoretical equations as well as low values of the diffusion coefficients characteristic of the polymer phase. as one can see from figure 7, the experimental data on the release of the protonated nicotinic acid from polymer are approximated well by equation (6). the interdiffusion coefficient of organic cations is (0.9 ± 0.1)∙10 −11 m 2 /s in dowex-50 while the interdiffusion coefficient of organic anions in dowex-1 is (1.3 ± 0.4)∙10 -12 m 2 /s. for the processes involving ionized pyridinecarboxylic acid the interdiffusion coefficients are significantly lower than the interdiffusion coefficient for the ion exchange h na  in amberlite ir-1 (3.7∙10 -10 m 2 /s [24]). for the processes (3) and (4), the interdiffusion coefficients within the selected interval of compositions of the ion exchangers are definitely controlled by the mobility of the organic ions in the polymers. for f < 0.5 the interdiffusion coefficient errors were calculated with a confidence equal to 0.95 using the least squares method. the half-conversion time of nicotinic acid release is obtained from experimental data (figure 6 ) or is calculated by equation (6) using the values of interdiffusion coefficients. the half-conversion time of nicotinic acid release by 0.01 m solution of hcl from dowex-50 is 140 s, which equals to 10,000 s when nicotinic acid is desorbed by water. the halfconversion time of nicotinic acid release by 0.1 m solution of heinrich altshuler et al. admet & dmpk 7(1) (2019) 76-87 86 nacl from dowex-1 is 1200 s that equals to more than one year when nicotinic acid is desorbed by water. conclusions the nicotinic acid encapsulation in nanocontainers based on commercial polymers: dowex-50 cation exchanger and dowex-1 anion exchanger expands the possibilities of obtaining prolonged forms of the active substance. nicotinic acid is released from the polymers by the ion exchange mechanism using a strong electrolyte as the eluent. this should be taken into account when predicting the pharmacokinetics of the release of a drug substance in vivo. it is likely that nicotinic acid encapsulated in dowex-50 upon oral administration remains immobilized in the esophagus and it is rapidly released at ph 2 upon reaching the stomach where hydrochloric acid is secreted. nicotinic acid encapsulated in dowex-1 can be targeted to the intestine in which it is easily released at ph 6. acknowledgements: the work is carried out within the framework of the state assignment of the institute of coal chemistry and material science, federal research center of coal and coal chemistry, siberian branch, russian academy of sciences (project aaaa-a17-117041910146-5) using the equipment of the collective-use center, federal research center of coal and coal chemistry. references [1] d.j. cram, s. karbach, h.e. kim, c.b. knobler, e.f. maverick, j.l. ericson, r.c. helgeson. host-guest complexation. 46. cavitands as open molecular vessels form solvates. journal of the american chemical society 110 (1988) 2229-2237. [2] ion exchange resins in medicine and biological research, annals of new york academy of sciences, v. 57, art. 3, 1953, pp. 61–324. [3] e. hagan, c.a. farrington, g.c. wall, m.m. belz. sodium polystyrene sulfonate for the treatment of acute hyperkalemia: a retrospective study. clinical nephrology 85 (2016) 38-43. [4] p. georgianos, i. liampas, a. kyriakou, v. vaios, v. raptis, n. savvidis, a. sioulis, v. liakopoulos, e. balaskas, p. zebekakis. evaluation of the tolerability and efficacy of sodium polystyrene sulfonate for long-term management of hyperkalemia in patients with chronic kidney disease. international urology and nephrology 49 (2017) 2217-2221. [5] g.h. kim, m.y. yu, j.s. park, c.h.lee, j.h. yeo. long-term efficacy of oral calcium polystyrene sulfonate for treating hyperkalemia in ckd patients. nephrology dialysis transplantation 31, issue suppl_1 (2016) i4, https://academic.oup.com/ndt/article/31/suppl_1/i4/2223735. [6] h.n. altshuler, g.yu. skurenko, s.yu. lyrschikov, a.a. gorlov, o.h. altshuler. solid nanoreactor. part 5. polymer nanocontainers for benzocaine. butlerov communications 44 (2015) 69-72. [7] o.h. altshuler, g.yu. shkurenko, h.n. altshuler. weak base diffusion in strong acid cation exchangers. solvent extraction and ion exchange 34(5) (2016) 502-508. [8] m.d. mashkovskiy, lekarstvennyye sredstva, new wave, moscow, russian federation, 2012, p. 1216. [9] k.e. smith, m.p. callahan, p.a. gerakines, j.p. dworkin, c.h. house. investigation of pyridine carboxylic acids in cm2 carbonaceous chondrites: potential precursor molecules for ancient coenzymes. geochimica et cosmochimica acta 136 (2014) 1-12. [10] e.m. goncalves, c.e. bernardes, h.p. diogo, m.e. minas da piedade. energetics and structure of nicotinic acid (niacin). the journal of physical chemistry b 114 (16) (2010) 5475–5485. [11] j.j. christensen, r.m. izatt, d.p. wrathall, l.d. hansen. thermodynamics of proton ionization in dilute aqueous solution. part xi. pk, δh°, and δs° values for proton ionization from protonated amines at 25°. journal of the chemical society a: inorganic, physical, theoretical (1969) 1212-1223. https://www.ncbi.nlm.nih.gov/pubmed/29027620 https://www.ncbi.nlm.nih.gov/pubmed/29027620 https://academic.oup.com/ndt/article/31/suppl_1/i4/2223735 admet & dmpk 7(1) (2019) 76-87 encapsulation and release of nicotinic acid doi: 10.5599/admet.626 87 [12] the international pharmacopoeia. seventh edition. 2017, http://apps.who.int/phint/en/p/about/ (date accessed 11.10.2018). [13] h.n. altshuler, l.p. abramova, o.h. altshuler (kemerovo state university), ru 2291171 (2007). [14] g.n. altshuler, n.v. malyshenko, a.n. popova. the ion-exchange properties of polymeric zirconium phosphate and zirconium dioxide. inorganic materials: applied research 9 (4) (2018) 746-750. [15] f. wang, k.a. berglund. monitoring ph swing crystallization of nicotinic acid by the use of attenuated total reflection fourier transform infrared spectrometry. industrial & engineering chemistry research 39 (2000) 2101-2104. [16] o.n. fedoseyeva, e.p. cherneva, n.n. tunitskii, zhurnal fizicheskoy khimii 33(4) (1959) 936–942 (in russian). [17] mopac, http://openmopac.net/ (date accessed 11.10.2018). [18] e. pretsch, p. buhlmann, m. badertscher, structure determination of organic compounds. tables of spectral data, springer-verlag, berlin heidelberg, germany, 2009, p. 433. [19] a. habibi-yangjeh, e. pourbasheer, m. danandeh-jenagharad. prediction of melting point for druglike compounds using principal component-genetic algorithm-artificial neural network. bulletin of the korean chemical society 29 (4) (2008) 833-841. [20] a. lal, n. shukla, v.b. singh, d. kumar singh. theoretical and experimental studies of vibrational spectra of nicotinic acid. journal of chemical and pharmaceutical research, 8 (4) (2016) 136-142. [21] p. singh, n.p. singh, r.a. yadav. quantum mechanical studies of conformers, molecular structures and vibrational characteristics of hetero-cyclic organics: nicotinic acid and 2-fluoronicotinic acid. journal of chemical and pharmaceutical research 3 (1) (2011) 737-755. [22] f. helfferich, ion exchange, dover publications, new york, usa, 1995, p. 624. [23] j. crank, the mathematics of diffusion, clarendon press, oxford, united kindom, 1975, p. 414. [24] g.e. boyd, a.w. adamson, l.s. myers. the exchange adsorption of ions from aqueous solutions by organic zeolites. ii. kinetics. journal of the american chemical society 69 (11) (1947) 2836-2848. ©2018 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://apps.who.int/phint/en/p/about/ http://openmopac.net/ http://creativecommons.org/licenses/by/3.0/ solubility and admet profiles of short oligomers of lactic acid doi: http://dx.doi.org/10.5599/admet.843 425 admet & dmpk 8(4) (2020) 425-436; doi: http://dx.doi.org/10.5599/admet.843 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper solubility and admet profiles of short oligomers of lactic acid daniela dascălu 1# , diana larisa roman 1# , madalina filip 1 , alecu ciorsac 2 , vasile ostafe 1 , adriana isvoran 1* 1 department of biology-chemistry and advanced environmental research laboratories, west university of timișoara, timișoara, romania 2 department of physical education and sport, university politehnica timișoara, timișoara, romania; *corresponding author: e-mail: adriana.isvoran@e-uvt.ro; tel.: +40-256-592-634; fax: +40-256-592-620 # the two authors have an equal contribution to this study. received: may 01, 2020; revised: june 21, 2020; published: june 28, 2020 abstract polylactic acid (pla) is a polymer with an increased potential to be used in different medical applications, including tissue engineering and drug-carries. the use of pla in medical applications implies the evaluation of the human organism's response to the polymer inserting and to its degradation products. consequently, within this study, we have investigated the solubility and admet profiles of the short oligomers (having the molecular weight lower than 3000 da) resulting in degradation products of pla. there is a linear decrease of the molar solubility of investigated oligomers with molecular weight. the results that are obtained also reveal that short oligomers of pla have promising pharmacological profiles and limited toxicological effects on humans. these oligomers are predicted as potential inhibitors of the organic anion transporting peptides oatp1b1 and oatp1b3, they present minor probability to affect the androgen and glucocorticoid receptors, have a weak potential of hepatotoxicity, and may produce eye injuries. these outcomes may be used to guide or to supplement in vitro and/or in vivo toxicity tests such as to enhance the biodegradation properties of the biopolymer. ©2020 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords pharmacokinetics; toxicological endpoints; biodegradation products of pla introduction for a few decades, biopolymers are extensively in use as food additives, cosmetics, medical materials, water treatment chemicals, packaging, etc. because of their intrinsic properties, especially non-cytotoxicity, biocompatibility and biodegradation, biopolymers are a class of materials that provides a wide range of applications in medicine [1]. they are considered outstanding candidates to be used for the preparation of many medical and body implants and also as controlled drug delivery systems. polylactic acid (pla), with the chemical formula (c3h4o2)n and illustrated in figure 1, is one of the most commercially competitive polymers used for medical applications acting as biologically inert supporting materials as scaffolds or drug-carriers [2,3]. pla is bio absorbable being transformed in the human organism by simple hydrolysis to products that can be further metabolized or excreted [4,5]. the http://dx.doi.org/10.5599/admet.843 http://dx.doi.org/10.5599/admet.843 http://www.pub.iapchem.org/ojs/index.php/admet/index file:///y:/users/adriana/appdata/local/temp/adriana.isvoran@e-uvt.ro http://creativecommons.org/licenses/by/4.0/ d. dascălu et al. admet & dmpk 8(4) (2020) 425-436 426 biomedical applications of the polylactic acid (pla) are drug delivery systems, suture threads, bone fixation screws [3]. pla is degraded in situ through hydrolysis with the production of oligomers which are more water-soluble. depending on the extent of hydrolysis, the products can be short oligomers (olas) and lactic acid (2-hydroxypropanoic acid the monomer) [3,5]. lactic acid is a nontoxic compound and a biochemical intermediate in carbohydrate metabolism [6]. the oligomers produced from hydrolysis may act as catalysts and they can diffuse from the sample toward the surfaces, these phenomena influencing the polymer degradation [3]. the counter-diffusion of the larger oligomers can be the determining phase for biodegradation as the oligomers having large dimensions cannot diffuse fast enough. increased solubility of oligomers may facilitate diffusion and consequently, the biodegradability of the polymer. it is already known that the degradation of biopolymers is important for controlled drug delivery systems and for various types of implants. biopolymers can be degraded by random chain scission to oligomers with lower molecular weight [7]. the degradation products, the oligomers, may also affect the human organism, the effects they produce being dependent on absorption, distribution, metabolism, excretion and toxicity (admet) profile [8]. the solubility in aqueous media is one of the most important physicochemical characteristics influencing the admet properties [9,10]. the soluble degradation products are either metabolized or transported through the lymphatic system to the kidney to be excreted from the organism [11]. pla is known as a biodegradable, biocompatible, non-toxic and eco-friendly polymer [12] but there is a recent in vitro study revealing some toxic effects of pla [13]: pla inhibited bioluminescence (an endpoint illustrating cytotoxicity in mammalian cells) with high efficiency, induced an oxidative stress response and it contains estrogenic compounds. to the best of our knowledge, there are no available studies concerning the biological effects of olas. taking into account this information, the aim of this study is to predict the aqueous solubility and admet profiles for short oligomers (containing up to 40 monomeric units and having the molecular weight up to 3000 da) resulted from the degradation of pla. our methodology is based on the fact that controlling agencies in the field of drug and medical devices development have confidence in the results obtained by the use of computational tools for the prediction of the biological effects. method within this study, we have considered short pla oligomers (olas) containing from 1 to 40 lactic acid units. we used the swissadme computational facility [14] to predict the values of their decimal logarithm of the molar solubility in water (log s) and their admet properties. three methods are included in swissadme to predict aqueous solubility: the estimate solubility (esol) model [15], a method adapted from ali and co-workers [16], and a method based on a system of 16 fragmental contributions modulated by the squared root of molecular weight implemented under filter-it software, version 1.0.2 (http://silicos-it.be.s3-website-eu-west-1.amazonaws.com/software/filter-it/1.0.2/filter-it.html). esol is a method used for estimating the aqueous solubility starting from its molecular topology [15] and is based on 4 molecular descriptors: the computed partition coefficient (clogp), the molecular weight (mw), the number of rotatable bonds (rb) and the proportion of heavy atoms in aromatic systems (ap) [15]. the equation allowing the computation of logs taking into account these descriptors is: figure 1. the structural formula of poly-lactic acid http://silicos-it.be.s3-website-eu-west-1.amazonaws.com/software/filter-it/1.0.2/filter-it.html admet & dmpk 8(4) (2020) 425-436 admet profiles of oligomers of lactic acid doi: http://dx.doi.org/10.5599/admet.843 427 wlog 0.16 0.63 log 0.062 0.066 0.74s c p m rb ap     . (1) the model introduced by ali and co-workers is based on two molecular descriptors: log p and topological polar surface area (tpsa) [16]: log 1.0377log 0.021 0.4488s p tpsa    . (2) besides the contribution of the other physicochemical parameters, these equations consider a linear decrease of the logarithm of the molar solubility coefficient with mw and log p, respectively. the calculation of the log s in filter-it software is based on the equation: w i ilog 0.898 0.104s m w c    , (3) where wi and ci are the respective weights and counts for the fragment i, and mw is the molecular weight of the compound (http://silicos-it.be.s3-website-eu-west-1.amazonaws.com/software/filter-it/1.0.2/filterit.html). swissadme computational tool uses for lipophilicity descriptor (log p) in equations (1) and (2) the value computed using xlogp3 method [17]. the input for the swissadme tool is the structure of a chemical compound in the simplified molecular-input line-entry system (smiles) format. admet profiles, organ and genomic toxicity of olas have been predicted using admetsar2.0 [18,19], pred-h-erg, pred-skin [20,21], endocrine disruptome [22], toxtree [23] and carcino-predel [24]. we have selected these computational tools among the numerous available facilities, as they have the accuracy of prediction usually higher than 70 % and friendly interfaces and tutorials that are available for free (online or open-source). a short description of the considered computational tools is given in table 1. these computational tools have been used for assessing the admet profiles and toxicological endpoints for numerous classes of chemicals: chito-oligomers [8], synthetic steroids [25], cosmetic ingredients [26,27], pesticides [26,28], water-soluble derivatives of chitosan [29]. it demonstrates their wide-ranging applicability. all the considered computational tools use as inputs smiles (simplified molecular-input line-entry system) formulas of the oligomers under investigation. these formulas have been obtained using acd/chemsketch utility (https://chemicalize.com accessed – accessed in march 2019). as it may be noticed from table 1, none of these methods considers the concentration of the investigated compound when making predictions. this is one of the limitations of computational assessment of biological effects of chemicals based on expert rules and/or qsar methods. for scientific graphing and data analysis we have used origin 8.0 software. results and discussions pla is a hydrophobic polymer, its aqueous solubility decreasing as its molecular weight increases [5]. within this study, we have assessed the dependence of the logarithm of the molar solubility in water (log s) of olas on the molecular weight and log p, respectively. the swissadme tool allowed to compute the values of the logs using the three methods described above. the obtained logs values are plotted against the molecular weight (figures 2) and log p (figures 3) and they were fitted with appropriate curves such as to obtain the highest values of the coefficients of determination (r squared). the equations obtained by data analysis are presented in table 2. for values computed using both esol and ali methods, there is a linear decrease in the logs values of olas with the mw and logp respectively, the decrease described by ali method being more pronounced. the silicos-it method conducts to a polynomial fit of order 3 for the variation of the logs with mw and log p, respectively (table 2). http://dx.doi.org/10.5599/admet.843 http://silicos-it.be.s3-website-eu-west-1.amazonaws.com/software/filter-it/1.0.2/filter-it.html http://silicos-it.be.s3-website-eu-west-1.amazonaws.com/software/filter-it/1.0.2/filter-it.html https://chemicalize.com/ d. dascălu et al. admet & dmpk 8(4) (2020) 425-436 428 table 1. short presentation of the computational tools that were used in the present study. qsar – quantitative structure-activity relationship tool method output accuracy of predictions, % references swissadme expert-rules based 2d qsar drug likeness, pharmacokinetic profile specifying yes or no for every investigated biological action 72-94 [14] admetsar2.0 2d qsar pharmacokinetic profiles, organ (eye, heart, liver) and genomic toxicity specifying the probability of the presence or absence of a biological action. 72-77 [18,19] pred-herg 2d qsar ability of a chemical compound to inhibit the human ether-à-go-go related gene (herg)k+ channels using both a binary and a multiclass model. 70-89 [21,30,31] pred-skin 2d-qsar skin sensitization potential based on multiple qsar models: prediction by binary model using human data, binary and multiclass predictions of murine skin sensitization potential based on animal data, and binary predictions based on non-animal data, i.e direct peptide reactivity assay (dpra), keratinosens, and the human cell line activation test (h-clat)] 70-84 [20,21,31] endocrine disruptome molecular docking probability of binding to nuclear receptors. 70-90 [22] toxtree expert-rules based carcinogenic and mutagenic potential expressed by yes or no. 70 [23] carcinopredel 2d qsar carcinogenic potential expressed by yes or no. 70 [24] figure 2. solubility in water dependence on the molecular weight of the small oligomers of pla: log s values are computed using esol and ali methods (a) and silicos-it method (b). red lines correspond to the fitting of data, linear fit for the values obtained using esol and ali methods (a) and polynomial fit of order 3 for the values obtained using silicos-it method (b). admet & dmpk 8(4) (2020) 425-436 admet profiles of oligomers of lactic acid doi: http://dx.doi.org/10.5599/admet.843 429 figure 3. solubility in water dependence on the log p values for the small oligomers of pla: log s values are computed using esol and ali methods (a) and silicos-it method (b). red lines correspond to the fitting of data, linear fit for the values obtained using esol and ali methods (a) and polynomial fit of order 3 for the values obtained using silicos-it method (b). table 2. equations corresponding to data analysis of the molar solubility in water (log s) values plotted against molecular weight (mw) and partition coefficient (log p) respectively for the three used methods to compute log s values esol, ali and silicos-it. method/ parameter mw log p equation r 2 equation r 2 esol log s =0.61-0.006 mw 0.9996 log s =-0.82-2.51log p 0.9987 ali log s =0.88-0.012 mw 0.9997 log s =-1.92-4.89 log p 0.9989 silicos it log s =1.02-0.002 mw + 1.15.10 -6 mw 2 -1.8.10 -10 mw 3 0.9767 log s =0.66-0.51 log p + 0.17 log p 2 -0.012 log p 3 0.9923 excepting the lactic acid, no experimental data on the molar solubility of the other olas were available for comparison and for evaluating these mathematical models. taking into account that the aqueous solubility of pla decreases with increasing molecular weight [5] and the r 2 values in table 2, ali method seems to be appropriate to be used to predict aqueous solubility for olas. the discrepancy of the three prediction may come from the data sets used to derive the models for predicting the solubility values. the data set used to derive the esol model (2874 compounds, predictive ability r 2 =0.72) [15] is much larger than the data set (1265 compounds, predictive ability r 2 = 0.869) being used in ali model. the ali model uses experimentally determined logp values and esol model uses calculated logp values (clogp). ali and his co-worker tested the esol model for 489 compounds with measured values for logp, solubility and melting points, and there was an improvement of the model [16]. furthermore, ali method explicitly accounts for the effect of polar and polarizable atoms on aqueous solubility. the discrepancy of the prediction models also underlines the limitations of the in silico studies and the necessity of further experimental measurements such as to improve our understanding of olas solubility. we have also assessed the admet properties of short olas. data obtained using admetsar2.0 and swissadme computational tools are illustrated in tables 3 and 4. numerical data represent the values of the probabilities that olas have (positive values) or have not (negative values) a certain biological action. usually, there is a good correlation between the predictions obtained using the two computational tools. furthermore, many of the predictions that we have obtained for olas are in good agreement with experimental literature data concerning the biological actions of pla: estrogenic effects [13], good gastrointestinal absorption [32], weak potential of p-gp inhibition [33,34], binding to human serum albumin [35,36], penetration of the blood brain barrier [37]. http://dx.doi.org/10.5599/admet.843 d. dascălu et al. admet & dmpk 8(4) (2020) 425-436 430 table 3. admet properties of olas: gi – gastrointestinal absorption, bbbp – blood-brain barrier penetration, p-gpssubstrate of the glycoprotein p, p-gpiinhibitor of the glycoprotein p, ppb – plasma proteins binding. the numerical values in this table represent the probabilities that investigated oligomers have (positive values) or have not (negative values) a certain biological action. the notation ”u” refers to the number of the monomers in the oligomer. oligomer/ biological action and tool gi bbbp p-gps p-gpi ppb admet sar swiss adme admet sar swiss adme admet sar admet sar swiss adme admet sar ola 1u 0.86 high 0.95 no -1.00 -0.99 no 0.63 ola 2u 0.90 high 0.98 no -0.99 -0.98 no 0.63 ola 3u 0.90 high 0.98 no -0.99 -0.94 no 0.58 ola 4u 0.90 low 0.98 no -0.99 -0.85 no 0.55 ola 5u 0.90 low 0.98 no -0.99 -0.66 no 0.53 ola 6u 0.90 low 0.98 no -0.99 -0.49 no 0.52 ola 7u 0.90 low 0.98 no -0.99 0.62 no 0.52 ola 8u 0.90 low 0.98 no -0.99 0.68 no 0.51 ola 10u 0.90 low 0.98 no -0.99 0.73 no 0.50 ola 12u 0.90 low 0.98 no -0.99 0.74 no 0.50 ola 14u 0.90 low 0.98 no -0.99 0.74 no 0.49 ola 16uola 40u 0.90 low 0.98 no -0.99 0.74 no 0.49 data presented in table 3 illustrate that pla short oligomers are predicted to have good gastrointestinal absorption and they are not considered substrates of the p-glycoprotein. the good gastrointestinal absorption of pla has been also observed by fernandez et al (2017) [32] by modeling the pla nanoparticles absorption under gastrointestinal conditions. their study revealed that the intestinal absorption of grape seed and skin extracts encapsulated in pla nanoparticles was significantly increased. admetsar tool reveals that oligomers containing between 7 (mw = 523 da) and 40 (mw = 2900 da) lactic acid units illustrate a mean probability to be inhibitors of p-glycoprotein, this probability increases with the chain length. this prediction is in good agreement with published data by li and his co-workers (2013) [33]. the study of li and co-workers revealed that copolymers of mpeg-pla could inhibit p-gp mediated efflux and that concentration played a major role in the p-gp inhibition activity. polyethylene glycol (peg) is a known p-gp inhibitor [34] and the fact that the inhibitory effect on p-gp efflux of the copolymers mpeg-pla having the same peg chain length is depended on the pla chain length, underlines that pla also has an inhibitory effect on p-gp activity. the copolymer with pla chain length of 4802 da was the most efficient pgp activity inhibitor and copolymers with longer or shorter pla chain lengths, gradually showed weaker inhibitory potential on p-gp function [33]. there is a disparity between the predictions made by the two computational tools concerning the ability of olas to penetrate the blood-brain barrier, admetsar predicts high probabilities of penetration of bloodbrain barrier and swissadme outcomes reveal that olas are not able to penetrate this barrier. swissadme tool uses for computing blood-brain–barrier permeation a model containing 260 molecules (156 permeant and 104 non-permeant) with reliable measurements of blood–brain partition and has a classification accuracy of 90 % [38]. admetsar tool uses for predicting blood barrier penetration a binary model containing a higher number of compounds (1839 with 1438 permeant and 401 no-permeant), has an accuracy of prediction of 90.7 % and a specificity of 86.2 % [18]. the in vivo experiments proved that the pla nanoparticles are able to penetrate bbb using transcytosis by microvascular endothelial cells [34]. admetsar2.0 tool seems to better predict the ability of the blood barrier permeation for olas. the admet & dmpk 8(4) (2020) 425-436 admet profiles of oligomers of lactic acid doi: http://dx.doi.org/10.5599/admet.843 431 disagreement of the predictions made by the two computational tolls also underlines the limitations of the computational evaluation of the biological effects of chemicals and the necessity to perform experimental studies to assess this property for olas. data from table 3 also illustrate that the probability of olas to bind to plasma proteins is reduced and it decreases with increasing the chain length. earlier studies revealed that lactic acid is able to bind to bovine serum albumin [35] and adsorption/desorption of human serum albumin at the surface of pla nanoparticles (nps) that decreased with increasing the diameter of nps has been observed [36]. table 4. probabilities that olas are substrates (s) and inhibitors (i) of the human cytochromes (cyp) involved in the metabolism of xenobiotics. the values in this table represent the probabilities that investigated oligomers are (positive values) or are not (negative values) substrates or inhibitors of cytochromes. the notation ”u” refers to the number of the monomers in the oligomer. tool/ cytochrom/ oligomer ola 1u ola 2u ola 3u ola 40u admetsar cyp3a4s -0.84 -0.71 -0.69 cyp2c9s 0.60 0.60 0.60 cypsd6s -0.87 -0.88 -0.88 cyp3a4i -0.98 -0.97 -0.95 cyp2c9i -0.88 -0.91 -0.93 cyp2c19i -0.98 -0.96 -0.97 cyp2d6i -0.98 -0.96 -0.95 cyp1a2i -0.97 -0.98 -0.98 swiss adme cyp3a4i no no no cyp2c9i no no no cyp2c19i no no no cyp2d6i no no no cyp1a2i no no no data presented in table 4 illustrate that olas are not considered substrates and inhibitors of the human cytochromes involved in the metabolism of xenobiotics and their presence in the human organism do not interfere with other compounds that are metabolized by these enzymes. we were not able to find scientific literature mentioning the effects of olas or pla on the human cytochromes. the probabilities corresponding to the ability of investigated olas to inhibit the organic anions and/or cations transporters are presented in figure 4. figure 4 illustrates that olas are able to inhibit the liverspecific organic anion transporters oatp1b1 and oatp1b3. this possibility should be addressed in experimental studies as these transporters are of particular importance for hepatic pharmacokinetics and elimination of xenobiotics, and the inhibition of these transporters may result in drug–drug interactions [39]. the potential of endocrine disruption of small oligomers of olas has been obtained using endocrine disruptome computational tool and is presented in figure 5. oligomers containing more than 10 lactic acid units were too big to accommodate in the active sites of the considered nuclear receptors and the computations have been aborted. investigated oligomers containing up to 7 lactic acid units present a small binding capacity to the androgen receptor in the antagonistic conformation, and they may produce reproductive dysfunctions. oligomers containing from 6 to 10 lactic acid units may affect the glucocorticoid receptor. these results are in good agreement with published data, zimmerman et al (2019) [13] illustrated that pla may have estrogenic effects. http://dx.doi.org/10.5599/admet.843 d. dascălu et al. admet & dmpk 8(4) (2020) 425-436 432 figure 4. probabilities of investigated olas to inhibit (positive values) or not inhibit (negative values) the organic anions and/or cations transporters (oatp/oct) obtained using admetsar tool. the notation ”u” refers to the number of the monomers in the oligomer. figure 5. predictions obtained using endocrine disruptome computational tool to assess the endocrine disruption potential of investigated olas: ar androgen receptor, erα and erβ oestrogen receptors α and β, gr glucocorticoid receptor, lxrα and lrxβ liver x receptors α and β, ppraα, ppraβ and ppraγ peroxisome proliferator activated receptors α, β/δ and γ, rxrα retinoid x receptor α, trα and trβ thyroid receptors α and β. both agonistic and antagonistic (an) effects for the nuclear receptors ar, erα, erβ and gr are predicted. the notation ”u” refers to the number of the monomers in the oligomer. oligomers containing more than 10 lactic acid units were too big to accommodate in the active sites of the considered nuclear receptors and the computations have been aborted. the results that we have obtained for the organ and genomic toxicity of small olas using all considered computational tools mentioned above are presented in figure 6. none of the investigated compounds reflects skin sensitization potential, cardiotoxicity, carcinogenicity and mutagenicity. all investigated oligomers are considered to be able to produce eye corrosion and smaller oligomers (from 1 to 3 lactic acid units) may also produce eye irritations and olas containing at least 5 units of lactic acids emphasize a weak potential of hepatotoxicity. literature data reveal one case of blindness and ophthalmoplegia following a treatment of the left periorbital region with the subcutaneous filler of pla [40]. in vitro and in vivo animal data illustrated that lactic acid has skin and eye irritation potential (pubchem, accessed in 24 of april, 2020). admet & dmpk 8(4) (2020) 425-436 admet profiles of oligomers of lactic acid doi: http://dx.doi.org/10.5599/admet.843 433 figure 6. predictions of organ and genomic toxicity of small olas: green cells illustrate non-toxicity and yellow cells illustrate possible toxicological effects. the notation ”u” refers to the number of the monomers in the oligomer and crg means carcinogenicty. also, literature data mention non-carcinogenicity [41], non-mutagenicity [42], non-hepatotoxicity and non-skin sensitization potential [43] of pla. we were not able to find information concerning the cardiotoxicity of this polymer and of the investigated oligomers. conclusions within this study, we have predicted the solubility and admet profiles of short oligomers of poly-lactic acid (containing from 1 to 40 lactic acid units, olas) that may be released during the degradation in the human organisms. the outcomes of our computational study reveal a linear decrease of the solubility of olas with molecular weight and logp respectively. the low aqueous solubility of oals is an important factor limiting the hydrolysis processes and the linear decreases of logs with the molecular weight may be used as an evaluating tool for predicting the olas behaviour in aqueous environments. although the oligomers of the biodegradable polymers are considered to be easily excreted through common metabolic pathways, their possible toxic effects need to be deeply understood and detailed studies must be conducted. admet profile assessment for investigated olas reveals their favourable pharmacological profiles and limited toxicological effects on humans. according to the outcomes of the computational tools that we have used, these oligomers may inhibit the organic anion transporting peptides oatp1b1 and/or oatp1b3, they illustrate a minor probability of affecting the androgen and glucocorticoid receptors, have a weak potential of hepatotoxicity, and may produce eye injuries. these outcomes may be used to define further experimental measurements needed to improve our understanding of the effects of the product of pla degradation. acknowledgements: this work was supported by the grant pniii-p3-285, polymeric nanobiomaterials for drug delivery: developing and implementation of safe-by-design concept enabling safe healthcare solutions. conflict of interest: the authors declare no conflict of interest. http://dx.doi.org/10.5599/admet.843 d. dascălu et al. admet & dmpk 8(4) (2020) 425-436 434 references [1] r. rebelo, m. fernandes, r. fangueiro. biopolymers in medical implants: a brief review. procedia engineering 200 (2017) 236–243. [2] i. manavitehrani, a. fathi, h. badr, s. daly, a. negahi shirazi, f. dehghani. biomedical applications of biodegradable polyesters. polymers (basel) 8 (2016) 20. doi: https://doi.org/10.3390/polym8010020. [3] g. gorrasi, r. pantani. hydrolysis and biodegradation of poly (lactic acid). advances in polymer science 279 (2018) 119-151. doi: https://doi.org/10.1007/12_2016_12. [4] m. savioli lopes, a.l. jardini, r. maciel filho. poly (lactic acid) production for tissue engineering applications. procedia engineering 42 (2012) 1402 – 1413. doi: https://doi.org/10.1016/j.proeng.2012.07.534. [5] t. casalini, f. rossi, a. castrovinci, g. perale. a perspective on polylactic acid-based polymers use for nanoparticles synthesis and applications. frontiers in bioengineering and biotechnology 7 (219) 259. doi: https://doi.org/10.3389/fbioe.2019.00259. [6] s. shakeri, m. ashrafizadeh, a. zarrabi, r. roghanian, e.g. afshar, a. pardakhty, r. mohammadinejad, a. kumar, v.k. thakur. multifunctional polymeric nanoplatforms for brain diseases diagnosis, therapy and theranostics. biomedicines 8 (2020) 13. doi: https://doi.org/10.3390/biomedicines8010013. [7] m.a. elsawy, k.h. kim, j.w. park, a. deep. hydrolytic degradation of polylactic acid (pla) and its composites. renewable and sustainable energy reviews 79 (2017) 1346-1352. https://doi.org/10.1016/j.rser.2017.05.143. [8] d.l. roman, m. roman, c. som, m. schmutz, e. hernandez, p. wick, t. casalini, g. perale, v. ostafe, a. isvoran. computational assessment of the pharmacological profiles of degradation products of chitosan. frontiers in bioengineering and biotechnology 7 (2019) 214. doi: https://doi.org/10.3389/fbioe.2019.00214. [9] d.f. veber, s.r. johnson, h.y. cheng, b.r. smith, k.w. ward, k.d. kopple. molecular properties that influence the oral bioavailability of drug candidates. journal of medicinal chemistry 45 (2002) 26152623. doi: https://doi.org/10.1021/jm020017n. [10] k.t. savjani, a.k. gajjar, j.k. savjani. drug solubility: importance and enhancement techniques. international scholarly research notices (2012) article id 195727. doi: https://doi.org/10.5402/2012/195727. [11] d.j. de boer, c. van blitterswijk, p. thomsen, j. hubbell, r. cancedda, j.d. de bruijn, a. lindahl, j. sohier, d.f. williams. tissue engineering, academic press, london, uk 2008, p. 217. [12] r.p. pawar, s.u. tekale, s.u. shisodia, j.t. totre, a.j. domb. biomedical applications of poly (lactic acid). recent patents on regenerative medicine 4 (2014) 40-51. doi: https://doi.org/10.2174/2210296504666140402235024. [13] l. zimmermann, g. dierkes, t.a. ternes, c. völker, m. wagner. benchmarking the in vitro toxicity and chemical composition of plastic consumer products. environmental science & technology 53 (2019) 11467−11477. doi: https://doi.org/10.1021/acs.est.9b02293. [14] a. daina, o. michielin, v. zoete. swissadme: a free web tool to evaluate pharmacokinetics, druglikeness and medicinal chemistry friendliness of small molecules. scientific reports 7 (2017) 42717. doi: https://doi.org/10.1038/srep4271. [15] j.s. delaney. esol: estimating aqueous solubility directly from molecular structure. journal of chemical information and computer sciences 44 (2004) 1000−1005. doi: https://doi.org/10.1021/ci034243x. [16] j. ali, p. camilleri, m.b. brown, a.j. hutt, s.b. kirton. revisiting the general solubility equation: in silico prediction of aqueous solubility incorporating the effect of topographical polar surface area. journal of chemical information and modeling 52 (2012) 420−428. doi: https://doi.org/10.1021/ci200387c. [17] t. cheng, y. zhao, x. li, f. lin, y. xu, x. zhang, y. li, r. wang, l. lai. computation of octanol− water partition coefficients by guiding an additive model with knowledge. journal of chemical information and modeling 47 (2017) 2140-2148. doi: https://doi.org/10.1021/ci700257y. https://doi.org/10.3390/polym8010020 https://doi.org/10.1007/12_2016_12 https://doi.org/10.1016/j.proeng.2012.07.534 https://doi.org/10.1016/j.proeng.2012.07.534 https://doi.org/10.3389/fbioe.2019.00259 https://doi.org/10.3390/biomedicines8010013 https://doi.org/10.1016/j.rser.2017.05.143 https://doi.org/10.1016/j.rser.2017.05.143 https://doi.org/10.3389/fbioe.2019.00214 https://doi.org/10.3389/fbioe.2019.00214 https://doi.org/10.1021/jm020017n https://doi.org/10.5402/2012/195727 https://doi.org/10.5402/2012/195727 https://doi.org/10.2174/2210296504666140402235024 https://doi.org/10.2174/2210296504666140402235024 https://doi.org/10.1021/acs.est.9b02293 https://doi.org/10.1038/srep4271 https://doi.org/10.1021/ci034243x https://doi.org/10.1021/ci034243x https://doi.org/10.1021/ci200387c https://doi.org/10.1021/ci700257y admet & dmpk 8(4) (2020) 425-436 admet profiles of oligomers of lactic acid doi: http://dx.doi.org/10.5599/admet.843 435 [18] f. cheng, w. li, y. zhou, j. shen, z. wu, g. liu, p.w. lee, y. tang. admetsar: a comprehensive source and free tool for evaluating chemical admet properties. journal of chemical information and modeling 52 (2012) 3099-3105. doi: https://doi.org/10.1021/ci300367a. [19] h. yang, c. lou, l. sun, j. li, y. cai, z. wang, w. li, g. liu, y. tang. admetsar 2.0: web-service for prediction and optimization of chemical admet properties. bioinformatics 35 (2019) 1067–1069. doi: https://doi.org/10.1093/bioinformatics/bty707. [20] r.c. braga, v.m. alves, e.n. muratov, j. strickland, n. kleinstreuer, a. trospsha, c.h. andrade. predskin: a fast and reliable web application to assess skin sensitization effect of chemicals. journal of chemical information and modeling 57 (2017) 1013-1017. doi: https://doi.org/10.1021/acs.jcim.7b00194. [21] v.m. alves, s.j. capuzzi, r.c. braga, j.v.b. borba, a.c. silva, t. luechtefeld, t. hartung, c.h. andrade, e.n. muratov, a. tropsha. a perspective and a new integrated computational strategy for skin sensitization assessment. acs sustainable chemistry & engineering 6 (2018) 2845-2859. https://doi.org/10.1021/acssuschemeng.7b04220. [22] k. kolsek, j. mavri, m. sollner dolenc, s. gobec, s. turk. endocrine disruptome an open source prediction tool for assessing endocrine disruption potential through nuclear receptor binding. journal of chemical information and modeling 54 (2014) 1254-1267. doi: https://doi.org/10.1021/ci400649p. [23] g. patlewicz, n. jeliazkova, r.j. safford, a.p. worth, b. aleksiev. an evaluation of the implementation of the cramer classification scheme in the toxtree software. sar nd qsar in environmental research 19 (2008) 495-524. doi: https://doi.org/10.1080/10629360802083871. [24] l. zhang, h. ai, w. chen, z. yin, h. hu, j. zhu, j. zhao, q. zhao, h. liu. carcinopred-el: novel models for predicting the carcinogenicity of chemicals using molecular fingerprints and ensemble learning methods. scientific reports 7 (2017) 2118. doi: https://doi.org/10.1038/s41598-017-02365-0. [25] m. roman, d.l. roman, v. ostafe, a. ciorsac, a. isvoran. computational assessment of pharmacokinetics and biological effects of some anabolic and androgenic steroids. pharmaceutical research 35 (2018) 41. doi: https://doi.org/10.1007/s11095-018-2353-1. [26] v.m. alves, e.n. muratov, a. zakharov, n.n. muratov, c.h. andrade, a. tropsha. chemical toxicity prediction for major classes of industrial chemicals: is it possible to develop universal models covering cosmetics, drugs, and pesticides?. food and chemical toxicology 112 (2018) 526-534. doi: https://doi.org/10.1016/j.fct.2017.04.008. [27] m. roman, d.l. roman, v. ostafe, a. isvoran. computational assessment of biological effects of methyl-,ethyl-, propyland butyl-parabens. jsm bioinformatics, genomics and proteomics 3 (2018) 1029. [28] i.m. gridan, a.a. ciorsac, a. isvoran. prediction of adme-tox properties and toxicological endpoints of triazole fungicides used for cereals protection. admet&dmpk 7 (2019) 161-173. https://dx.doi.org/10.5599/admet.668. [29] a. isvoran, a. ciorsac, v. ostafe. adme-tox profiling of some low molecular weight water soluble chitosan derivatives. admet & dmpk 5 (2017) 192-200. https://doi.org/10.5599/admet.5.3.423. [30] r.c. braga, v.m. alves, m.f. silva, e. muratov, d. fourches, l.m. lião, a. tropsha, c.h. andrade. predherg: a novel web-accessible computational tool for predicting cardiac toxicity. molecular informatics 34 (2015) 698-701. doi: https://doi.org/10.1002/minf.201500040. [31] v.m. alves, r.c. braga, e. muratov, c.h. andrade. development of web and mobile applications for chemical toxicity prediction. journal of the brazilian chemical society 29 (2018b) 982-988. https://doi.org/10.21577/0103-5053.20180013. [32] k. fernández, m. roeckel, e. canales, j. dumont. modeling of the nanoparticles absorption under a gastrointestinal simulated ambient condition. aaps pharmscitech 18 (2017) 2691-2701. doi: https://doi.org/10.1208/s12249-017-0751-z. [33] w. li, x. li, , y. gao, , y. zhou, s. ma, y. zhao, j. li, y. liu, x. wang, d. yin. inhibition mechanism of pglycoprotein mediated efflux by mpeg-pla and influence of pla chain length on p-glycoprotein http://dx.doi.org/10.5599/admet.843 https://doi.org/10.1021/ci300367a https://doi.org/10.1093/bioinformatics/bty707 https://doi.org/10.1021/acs.jcim.7b00194 https://doi.org/10.1021/acs.jcim.7b00194 https://doi.org/10.1021/acssuschemeng.7b04220 https://doi.org/10.1021/acssuschemeng.7b04220 https://doi.org/10.1021/ci400649p https://doi.org/10.1080/10629360802083871 https://doi.org/10.1038/s41598-017-02365-0 https://doi.org/10.1007/s11095-018-2353-1 https://doi.org/10.1016/j.fct.2017.04.008 https://dx.doi.org/10.5599/admet.668 https://dx.doi.org/10.5599/admet.668 https://doi.org/10.5599/admet.5.3.423 https://doi.org/10.1002/minf.201500040 https://doi.org/10.21577/0103-5053.20180013 https://doi.org/10.1208/s12249-017-0751-z d. dascălu et al. admet & dmpk 8(4) (2020) 425-436 436 inhibition activity. molecular pharmaceutics 11 (2013) 71–80. doi: https://doi.org/10.1021/mp4004223. [34] m. mokhtar, p. gosselin, f. lacasse, p. hildgen. design of peg-grafted-pla nanoparticles as oral permeability enhancer for p-gp substrate drug model famotidine. journal of microencapsulation 34(1) (2017) 91–103. doi: https://doi.org/10.1080/02652048.2017.1290155. [35] j.c. chatham, j.r. forder. lactic acid and protein interactions: implications for the nmr visibility of lactate in biological systems. biochim biophys acta 1426 (1999) 177-184. doi: https://doi.org/10.1016/s0304-4165(98)00154-8. [36] t. verrecchia, p. huve, d. bazile, m. veillard, g. spenlehauer, p. couvreur. adsorption/desorption of human serum albumin at the surface of poly (1actic acid) nanoparticles prepared by a solvent evaporation process. journal of biomedical materials research 27 (1993) 1019-1028. https://doi.org/10.1002/jbm.820270807. [37] wang hf, hu y, sun wq, xie cs. polylactic acid nanoparticles across the brain-blood barrier observed with analytical electron microscopy. chinese journal of biotechnology 20(5) (2004) 790794. https://pubmed.ncbi.nlm.nih.gov/15974011/. [38] a. daina, v. zoete. a boiled-egg to predict gastrointestinal absorption and brain penetration of small molecules. chemmedchem 11 (2016) 1117–1121. doi: https://doi.org/10.1002/cmdc.201600182. [39] n.f. smith, w.d. figg, a. sparreboom. role of the liver-specific transporters oatp1b1 and oatp1b3 in governing drug elimination. expert opinion on drug metabolism & toxicology 1 (2005) 429-445. https://doi.org/10.1517/17425255.1.3.429. [40] s.a. roberts, b.p. arthurs. severe visual loss and orbital infarction following periorbital aesthetic poly-(l)-lactic acid (plla) injection. ophthalmic plastic and reconstructive surgery 28 (2012) 68-70. doi: https://doi.org/10.1097/iop.0b013e3182288e4d. [41] r.a. abd alsaheb, a. aladdin, n.z. othman, r.a. malek, o.m. leng, r. aziz, h.a. el enshasy. recent applications of polylactic acid in pharmaceutical and medical industries. journal of chemical and pharmaceutical research 7 (2015) 51-63. http://www.jocpr.com/articles/recent-applications-ofpolylactic-acid-in-pharmaceutical-and-medical-industries.pdf. [42] s. maisanaba, s. pichardo, m. jordá-beneyto, s. aucejo, a.m. cameán, á. josa. cytotoxicity and mutagenicity studies on migration extracts from nanocomposites with potential use in food packaging. food and chemical toxicology 66 (2014) 366-372. https://doi.org/10.1016/j.fct.2014.02.011. [43] r.y. li, z.g. liu, h.q. liu, l. chen, j.f. liu, y.h. pan. evaluation of biocompatibility and toxicity of biodegradable poly (dl-lactic acid) films. american journal of translocation research 7 (2015) 13571370. https://www.ncbi.nlm.nih.gov/pubmed/26396667. ©2020 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) https://doi.org/10.1021/mp4004223 https://doi.org/10.1021/mp4004223 https://doi.org/10.1080/02652048.2017.1290155 https://doi.org/10.1016/s0304-4165(98)00154-8 https://doi.org/10.1016/s0304-4165(98)00154-8 https://doi.org/10.1002/jbm.820270807 https://pubmed.ncbi.nlm.nih.gov/15974011/ https://doi.org/10.1002/cmdc.201600182 https://doi.org/10.1002/cmdc.201600182 https://doi.org/10.1517/17425255.1.3.429 https://doi.org/10.1097/iop.0b013e3182288e4d http://www.jocpr.com/articles/recent-applications-of-polylactic-acid-in-pharmaceutical-and-medical-industries.pdf http://www.jocpr.com/articles/recent-applications-of-polylactic-acid-in-pharmaceutical-and-medical-industries.pdf https://doi.org/10.1016/j.fct.2014.02.011 https://doi.org/10.1016/j.fct.2014.02.011 https://www.ncbi.nlm.nih.gov/pubmed/26396667 http://creativecommons.org/licenses/by/3.0/ permeation characteristics of tetracyclines in parallel artificial membrane permeation assay doi: 10.5599/admet.657 151 admet & dmpk 7(3) (2019) 151-160; doi: http://dx.doi.org/10.5599/admet.657 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper permeation characteristics of tetracyclines in parallel artificial membrane permeation assay sachika yamauchi, kiyohiko sugano molecular pharmaceutics lab., college of pharmaceutical sciences, ritsumeikan university, 1-1-1, noji-higashi, kusatsu, shiga 525-8577, japan *corresponding author: e-mail: suganok@fc.ritsumei.ac.jp; tel.: +81-77-561-2773 received: january 17, 2019; revised: march 06, 2019; available online: may 8, 2019 abstract the purpose of the present study was to characterize the passive permeation of tetracyclines in the parallel artificial membrane permeation assay (pampa). tetracyclines exist as zwitterion at physiological ph. the pampa membrane was prepared by impregnating a phospholipid/decane solution to a filter support. the permeation coefficient (pe) of tetracycline (tc) was markedly affected by the lipid composition of the pampa membrane. no permeation was observed when phospholipid was not added (pure decane membrane, pe < 0.05 × 10 -6 cm/sec). with the addition of 2 % pc, little or no increase in pe was observed. the addition of 1 % pe increased the pe value more than tenfold. the addition of 2 % soybean lecithin containing phosphatidylinositol (pi) and phosphatidic acid (pa) increased the p e value to above 4 × 10 -6 cm/sec. the pe value was further increased to 15 × 10 -6 cm/sec by increasing the concentration of soybean lecithin from 2 to 10 %. the pe value showed ph and temperature dependence, whereas it was not affected by the ionic strength, tc concentration, and ion-pair transport inhibitors. a weak correlation was observed between the pe values and octanol-buffer distribution coefficients of tetracyclines. these results suggest that inter-molecular interactions between tc and pe, pi and/or pa facilitate the passive diffusion of tc across the pampa membrane. keywords zwitterion; artificial membrane; permeability; phospholipid introduction zwitterionic drugs are an important chemical class as they exhibit unique physicochemical and pharmacokinetic properties [1, 2]. a zwitterionic drug possesses at least one acidic group and at least one basic group (acidic pka < basic pka). typical examples of zwitterionic drugs are antibacterials, antiallergics, and diuretics. zwitterion drugs are less liable to human ether-a-go-go related gene (herg) and phospholipidosis compared to hydrophobic bases [3, 4]. according to the ph-partition theory, the passive lipid bilayer permeation of a zwitterionic drug is expected to be negligible. however, many zwitterionic drugs such as tetracyclines and fluoroquinolones show moderate to high passive permeability in vitro [5] and good bioavailability in vivo [6]. the parallel artificial membrane permeation assay (pampa) has been widely used as a high throughput assay for passive membrane permeation [7-10]. interestingly, several zwitterionic drugs showed moderate http://www.pub.iapchem.org/ojs/index.php/admet/index yamauchi and sugano admet & dmpk 7(3) (2019) 151-160 152 to high permeability in pampa [6, 11]. in the case of cationic drugs, ion pair formation with an anionic phospholipid enhances the passive permeation of a drug [12,13]. however, the permeation mechanism of zwitterionic drugs has not been investigated. the purpose of the present study was to characterize the passive permeation of zwitterion drugs in pampa. tetracyclines were used as model zwitterionic drugs in this study. tetracyclines have three dissociative functional groups (figure 1) [14-16]. at the physiological ph in the small intestine, they predominantly exist as zwitterion [14]. the octanol buffer partition coefficients (log d) of tetracyclines are below 0.2 (table 1) [17]. in addition, they violate the lipinski’s rule of five in the hydrogen bond number [18]. however, tetracyclines show good oral absorption in vivo [19]. interestingly, tetracycline showed moderate permeability in the biomimetic pampa [9]. figure 1. chemical structures of tetracyclines experimental materials tetracycline hydrochloride, l-leucine, decane, sodium dihydrogen phosphate, sodium chloride, phosphatidylethanolamine (pe), and 8n naoh were purchased from wako pure chemical industries, ltd (osaka, japan). oxytetracycline hydrochloride, minocycline hydrochloride, doxycycline hyclate, and 2aminooctanoic acid were purchased from tci (tokyo, japan). demeclocycline hydrochloride and chlortetracycline hydrochloride were purchased from lkt labs, inc (mn, usa). phosphatidylcholine (pc) was purchased from nof corporation (tokyo, japan). tetrahexylamine bromide (tha) was purchased from sigma-aldrich co. llc (mo, usa). procainamide hydrochloride was purchased from combi-blocks inc (ca, admet & dmpk 7(3) (2019) 151-160 permeation characteristics of tetracyclines in pampa doi: 10.5599/admet.657 153 usa). soy bean lecithins (slp-pc 70, slp-white, slp-pi grades) were provided by tsuji oil mills co., ltd (mie, japan). table 1. physicochemical properties of tetracyclines mw pka log d (ph 6.5) a pka ref. chlortetracycline 479 3.3, 7.6, 9.3 -0.88 (14) b 3.25, 6.72, 8.84 (16) c demeclocycline 465 3.4, 7.4, 9.4 -0.67 (14) b doxycycline 444 3.0, 8.0, 9.2 -0.08 (14) b 3.50, 7.25, 9.58 (16) c minocycline 457 2.8, 5.0, 7.8, 9.5 0.20 (15) d oxytetracycline 460 3.2, 7.5, 8.9 -0.96 (14) b 3.53, 7.25, 9.58 (16) c tetracycline 444 3.3, 7.8, 9.6 -1.09 (14) b 3.35, 7.29, 9.88 (16) c a measured in this study. b potentiometry (23 °c), ionic strength = 0.01 or 0.05 m. c potentiometry (25 °c), ionic strength = 0.1 m. d method not described in the literature. table 2. lipid composition of soy bean lecithin a phospholipid slp-pc70 (%) slp-white (%) slp-pi (%) phosphatidylcholine (pc) 65 – 75 24 – 32 15 – 22 phosphatidylethanolamine (pe) 10 – 15 20 – 28 25 – 32 phosphatidylinositol (pi) 0 – 1 12 – 20 18 – 25 phosphatidic acid (pa) 1 – 3 8 – 15 8 – 15 lysophosphatidylcholines (lpc) 1 – 5 1 – 5 1 – 5 a taken from the product information provided by the manufacturer. methods pampa assay the pampa sandwich was consisted of a 96 well filter plate (hydrophobic pvdf, 0.45 μm) and a pampa acceptor plate (merck millipore, ma, usa). before forming the pampa sandwich, the bottom (acceptor) plate was filled with 300 μl of a 50 mm sodium phosphate buffer. the filter of the top (donor) compartment was coated with 5 μl of a phospholipid – decane solution. the compositions of soy bean lecithins were shown in table 2. the buffer conditions were the same for both donor and acceptor compartments (iso-ph and iso-ionic strength condition). the pampa sandwich was placed in a plastic container containing a small amount of water on the bottom and incubated for 3 h (at 25 and 37 °c) or 18 h (at 15 °c). after incubation, 100 μl of both the donor and acceptor solutions were transferred to uv plates. the concentrations of tetracyclines and procainamide were measured at 360 and 280 nm, respectively. the pampa permeability was calculated by the following equation [19]: d v d v v d 2.303 1 ( )1 log 1 1 (0) e v r c t p r a t r r c                      (1) d a d v d ( ) ( )1 1 (0) (0) c t c t r c r c    (2) yamauchi and sugano admet & dmpk 7(3) (2019) 151-160 154 d v a v r v  (3) where pe is the effective permeation coefficient (cm/s), a is the filter surface area (0.266 cm 2 ), vd and va are the volumes in the donor and acceptor phase (0.15 and 0.3 ml, respectively), t is the incubation time, cd(t) is the concentration of a drug in the donor phase at time t, r is the membrane retention factor, and rv is the volume ratio. octanol-buffer distribution coefficient the octanol-buffer distribution coefficients (log d) of tetracyclines were determined by a shake-flask method at ph 6.5 (50 mm sodium – phosphate buffer). the octanol and buffer phases were mutually presaturated before use. a buffer solution of a model drug (1.0 mm, 0.5 ml) and octanol (2.5 ml) were added to a 15 ml tube. the sample was vigorously shaken for 60 min at room temperature. the concentrations of tetracyclines were determined by uv spectroscopy at 360 nm. results effect of membrane composition the pampa permeation of tetracycline (tc) was markedly affected by the lipid composition of the pampa membrane (figure 2). no permeation was observed when phospholipid was not added (pure decane membrane, pe < 0.05 × 10 -6 cm/sec). with the addition of 2 % pc, little or no increase in pe was observed. the addition of 1 % pe increased the pe value more than tenfold. the addition of 2 % soybean lecithin containing phosphatidylinositol (pi) and phosphatidic acid (pa) increased the pe value to above 4∙10 -6 cm/sec. the pe value was further increased to 15∙10 -6 cm/sec by increasing the concentration of soybean lecithin from 2 to 10 %. the 10% soybean lecithin (slp white) decane membrane was used in the following studies. figure 2. effect of membrane composition on pampa permeation of tetracycline (mean ± sd, n = 3 9). assay conditions: tc 0.5 mm, 10 % slp-white/decane, ph 6.5 50 mm sodium phosphate buffer, 37 °c. admet & dmpk 7(3) (2019) 151-160 permeation characteristics of tetracyclines in pampa doi: 10.5599/admet.657 155 ph and ionic strength dependency the effect of ph on the pampa permeation of tc is shown in figure 3. the pe values decreased as ph was increased above 7.0. the effect of the ionic strength (i) is shown in figure 4. the ionic strength showed little or no effect in the range of i = 0.15 to 2.0 m. figure 3. effect of ph on pampa permeation of tetracycline (mean ± sd, n = 3). assay conditions: tc 0.5 mm, 10 % spl-white / decane, 50 mm sodium phosphate buffer, at 37 °c. figure 4. effect of ionic strength on pampa permeation of tetracycline (mean ± sd, n = 3). assay conditions: tc 0.5 mm, 10 % spl-white / decane, ph 6.5, 37 °c. the ionic strength of the medium was adjusted by nacl. yamauchi and sugano admet & dmpk 7(3) (2019) 151-160 156 temperature dependence the effect of temperature on the pampa permeation of tc is shown in figure 5. as the temperature was increased from 15 °c to 37 °c, the pe value was increased 6.5 fold. figure 5. effect of temperature on pampa permeation of tetracycline (mean ± sd, n = 3). assay conditions: tc 0.5 mm, 10 % spl-white / decane, ph 6.5 50 mm sodium phosphate buffer. concentration dependence the effect of tc concentration on the pampa permeation is shown in figure 6. the pe value of tc was not affected by the tc concentration up to 0.5 mm. due to the solubility of tc in the medium, the concentration of tc could not be increased above 0.5 mm. figure 6. effect of tetracycline concentration on pampa permeation (mean ± sd, n = 3). assay conditions: 10 % spl-white / decane, ph 6.5 50 mm sodium phosphate buffer, 37 °c. admet & dmpk 7(3) (2019) 151-160 permeation characteristics of tetracyclines in pampa doi: 10.5599/admet.657 157 effect of additives in the donor media the effect of possible inhibitors on the pampa permeation of tc is shown in figure 7. procainamide was used as a control of the ion-pair transport (a cation drug and an anion phospholipid). the permeation of procainamide was inhibited by tetrahexylammonium (tha), however not by l-leucine and 2-amino octanoic acid (aoa). the permeation of tc was not inhibited by the inhibitors employed in this study. figure 7. effect of additives on pampa permeation of tetracycline and procainamide (mean ± sd, n = 3). assay conditions: substrates 0.5 mm, additives 10 mm, 10 % spl-white / decane, ph 6.5 50 mm sodium phosphate buffer, 37 °c. log pe – log d relationship the log pe – log d relationship for tetracyclines is shown in table 3 and figure 8. a weak correlation was observed between log pe and log d. table 3. pampa permeability for tetracyclines a pe (10 -6 cm/sec, mean ± sd, n = 6) chlortetracycline 41 ± 3 demeclocycline 33 ± 1 doxycycline 64 ± 5 minocycline 71 ± 6 oxytetracycline 9.2 ± 0.7 tetracycline 15 ± 1 a tetracyclines 0.5 mm, 10% spl-white / decane, ph 6.5 50 mm sodium phosphate buffer, 37 °c. discussion we first investigated the effect of the membrane composition on the pampa permeation of tc. the results of the lipid composition dependency study suggest that intermolecular interactions between tc and pe, ps and/or pi facilitate the passive diffusion of tc across the pampa membrane [20]. the results of this study are in good agreement with the previous finding that the lipid composition is critically important for the pampa assay [9,21]. the pe value of tc in the 10 % soybean lecithin/decane membrane was similar to that observed in the biomimetic pampa in which 1,7-octadiene was used as an organic solvent. even though the biomimetic pampa showed promising predictability for in vivo oral drug absorption, as 1,7octadiene is irritant, it is not suitable for routing use. yamauchi and sugano admet & dmpk 7(3) (2019) 151-160 158 figure 8. correlation between log d and pampa permeability for tetracyclines (mean ± sd, n = 6). pampa conditions: tetracyclines 0.5 mm, 10 % spl-white / decane, ph 6.5 50 mm sodium phosphate buffer, 37 °c. tc predominantly exists as zwitterion form (net zero charge) in the range of ph 4.0 to 7.0. in this ph range, the pe value remained constant. above ph 7.0, the pe value decreased as tc becomes negatively charged (two anions and one cation) due to the dissociation of the phenolic diketone part above ph 7.0 (figure 1) [22]. however, the inflection ph point was below the pka of tc. in addition, the slope of the ph log pe line above ph 7.0 was about -0.5. these deviate from the theoretical ph – log pe curve based on the ph partition theory. further investigation is required to clarify the reasons for these deviations. ionic strength had no effect on the pe value, suggesting that tc did not form an ion pair with the inorganic ions in the buffer at ph 6.5. the pampa permeation of tc showed marked temperature dependence in this study. the ratio of pe values between 15 °c and 37 °c was 6.5. this ratio was greater than previously reported by vizserálek et al. for undissociable, mono-acid, and mono-base drugs using the 2 % pc/1 % cholesterol/dodecane membrane (1.1 to 3.7) [23]. to investigate the permeation mechanism of tc, the saturation and inhibition of tc permeation were investigated. neither saturation nor inhibition was observed. this result was different from the results for hydrophilic basic drugs, for which both saturation and inhibition have been reported [12]. for hydrophilic basic drugs, ion pair formation between a drug cation and an anionic phospholipid was suggested as the permeation mechanism [12, 13]. since the addition of pe, pi and/or pa significantly enhanced the pe value of tc, there might exist some nonspecific inter-molecular interaction between tc and these phospholipids other than ion-pair formation [20]. only a weak correlation was observed between log d and log pe, suggesting that the inter-molecular interaction between phospholipids and tetracyclines might be different from that between octanol and tetracyclines. admet & dmpk 7(3) (2019) 151-160 permeation characteristics of tetracyclines in pampa doi: 10.5599/admet.657 159 conclusions in conclusion, in this study, the permeation characteristics of tc were investigated in detail. the phospholipid composition and incubation temperature showed marked effects on the permeation of tc, whereas the ionic strength of the media, the concentration of tc, and the addition of ion pair inhibitors showed little or no effect. conflict of interest: none. references [1] a. pagliara, p. carrupt, g. caron, p. gaillard, b. testa. lipophilicity profiles of ampholytes. chem. rev. 97 (1997) 3385-3400. [2] c. chen. some pharmacokinetic aspects of the lipophilic terfenadine and zwitterionic fexofenadine in humans. drugs in r & d. 8 (2007) 301-314. [3] c. jamieson, e.m. moir, z. rankovic, g. wishart. medicinal chemistry of herg optimizations: highlights and hang-ups. j. med. chem. 49 (2006) 5029-5046. [4] k. tomizawa, k. sugano, h. yamada. physicochemical and cell-based approach for early screening of phospholipidosis-inducing potential. j. toxicol. sci. 31 (2006) 315-324. [5] k.f. hermann, c.s. neuhaus, v. micallef, b. wagner, m. hatibović, h.e. aschmann, f. paech, r. alvarez-sanchez, s.d. krämer, s. belli. kinetics of lipid bilayer permeation of a series of ionisable drugs and their correlation with human transporter-independent intestinal permeability. eur. j. pharm. sci. 104 (2017) 150-161. [6] k.y. tam, a. avdeef, o. tsinman, n. sun. the permeation of amphoteric drugs through artificial membranes− an in combo absorption model based on paracellular and transmembrane permeability. j. med. chem. 53 (2009) 392-401. [7] m. kansy, f. senner, k. gubernator. physicochemical high throughput screening: parallel artificial membrane permeation assay in the description of passive absorption processes. j. med. chem. 41 (1998) 1007-1010. [8] f. wohnsland, b. faller. high-throughput permeability ph profile and high-throughput alkane/water log p with artificial membranes. j. med. chem. 44 (2001) 923-930. [9] k. sugano, h. hamada, m. machida, h. ushio. high throughput prediction of oral absorption: improvement of the composition of the lipid solution used in parallel artificial membrane permeation assay. j. biomol. screen. 6 (2001) 189-196. [10] l. di, e.h. kerns, k. fan, o.j. mcconnell, g.t. carter. high throughput artificial membrane permeability assay for blood–brain barrier. eur. j. med. chem. 38 (2003) 223-232. [11] m. bermejo, a. avdeef, a. ruiz, r. nalda, j.a. ruell, o. tsinman, i. gonzález, c. fernández, g. sánchez, t.m. garrigues. pampa—a drug absorption in vitro model: 7. comparing rat in situ, caco-2, and pampa permeability of fluoroquinolones. eur. j. pharm. sci. 21 (2004) 429-441. [12] k. sugano, y. nabuchi, m. machida, y. asoh. permeation characteristics of a hydrophilic basic compound across a bio-mimetic artificial membrane. int. j. pharm. 275 (2004) 271-278. [13] z.s. teksin, k. hom, a. balakrishnan, j.e. polli. ion pair-mediated transport of metoprolol across a three lipid-component pampa system. j. controlled release. 116 (2006) 50-57. [14] s. babić, a.j. horvat, d.m. pavlović, m. kaštelan-macan. determination of pka values of active pharmaceutical ingredients. trac-trend anal. chem. 26 (2007) 1043-1061. [15] minocycline drug information, www.info.pmda.go.jp/go/interview/2/671450_6152005m1043_2_1f. 2019. http://www.info.pmda.go.jp/go/interview/2/671450_6152005m1043_2_1f yamauchi and sugano admet & dmpk 7(3) (2019) 151-160 160 [16] s. şanli, n. şanli, g. alsancak. determination of protonation constants of some tetracycline antibiotics by potentiometry and lc methods in water and acetonitrile-water binary mixtures. j. braz. chem. soc. 20 (2009) 939-946. [17] j.l. colaizzi, p.r. klink. ph‐partition behavior of tetracyclines. j. pharm. sci. 58 (1969) 1184-1189. [18] c.a. lipinski, f. lombardo, b.w. dominy, p.j. feeney. experimental and computational approaches to estimate solubility and permeability in drug discovery and development settings. adv. drug deliv. rev. 23 (1997) 3-25. [19] k.n. agwuh, a. macgowan. pharmacokinetics and pharmacodynamics of the tetracyclines including glycylcyclines. j. antimicrob. chemother. 58 (2006) 256-265. [20] f. assmus, a. ross, h. fischer, j. seelig, a. seelig. 31p and 1h nmr studies of the molecular organization of lipids in the parallel artificial membrane permeability assay. mol. pharm. 14 (2016) 284-295. [21] a. avdeef. absorption and drug development: solubility, permeability, and charge state, john wiley & sons, 2012. [22] a. sigler, p. schubert, w. hillen, m. niederweis. permeation of tetracyclines through membranes of liposomes and escherichia coli. eur. j. biochem. 267 (2000) 527-534. [23] g. vizserálek, t. balogh, k. takács-novák, b. sinkó. pampa study of the temperature effect on permeability. eur. j. pharm. sci. 53 (2014) 45-49. ©2019 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ admet & dmpk x(y) (20xx) pp-pp; doi: 10 doi: 10.5599/admet.1.3.8 45 admet & dmpk 1(3) (2013) 45; doi: 10.5599/admet.1.3.8 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index editorial board member: brief biography hong wan dr. hong wan is currently executive director and head of dmpk/tox of shanghai hengrui pharmaceutical co ltd (shhrp). prior to joining shhrp in may 2013, he was executive director of dmpk & bioanalysis department at a cro in china, crown bioscience inc (taicang), for 3 years. before moving back to china in 2010, he worked for astrazeneca rd, mölndal, sweden for 10 years, from discovery dmpk & bioanalytical chemistry to lead generation dmpk and physical chemistry, and as a principal scientist since 2007. while in this role, he contributed to a number of drug hunting projects as well as led in developing high-throughput dmpk & physicochemical assays. he received his phd from the department of analytical chemistry, stockholm university in 1996, and was associate professor in the department of chemistry, karlstad university in 2001. he completed his bs degree in 1984 and his ms degree in 1987 in the analytical chemistry department of lanzhou university and the lanzhou institute of chemical physics, chinese academia of science, respectively. dr. wan has authored and co-authored over 50 publications in peer-reviewed journals and has received invitations to peer review. he is also an editorial board member for expert opinion on drug metabolism and toxicology, the open spectroscopy journal, and admet/dmpk. ©2013 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution licence (http://creativecommons.org/licenses/by/3.0/) http://www.pub.iapchem.org/ojs/index.php/admet/index http://creativecommons.org/licenses/by/3.0/ adme prediction with knime: a retrospective contribution to the second “solubility challenge” doi: http://dx.doi.org/10.5599/admet.979 209 admet & dmpk 9(3) (2021) 209-218; doi: https://doi.org/10.5599/admet.979 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper adme prediction with knime: a retrospective contribution to the second “solubility challenge” gabriela falcón-cano 1 , christophe molina* 2 , and miguel ángel cabrera-pérez* 1,3,4 1 unit of modelling and experimental biopharmaceutics. centro de bioactivos químicos. universidad central “marta abreu” de las villas. santa clara 54830, villa clara, cuba 2 pikaïros s.a., 31650 saint orens de gameville, france 3 department of pharmacy and pharmaceutical technology, university of valencia, burjassot 46100, valencia, spain 4 department of engineering, area of pharmacy and pharmaceutical technology, miguel hernández universit y, 03550 sant joan d'alacant, alicante, spain *corresponding authors: e-mail: macabreraster@gmail.com; tel.: +53-42-281473; fax: +53-42-281130; e-mail: christophe.molina@pikairos.com. received: march 09, 2021; revised: june 21, 2021; available online: july 12, 2021 abstract computational models for predicting aqueous solubility from the molecular structure represent a promising strategy from the perspective of drug design and discovery. since the first “solubility challenge”, these initiatives have marked the state-of-art of the modelling algorithms used to predict drug solubility. in this regard, the quality of the input experimental data and its influence on model performance has been frequently discussed. in our previous study, we developed a computational model for aqueous solubility based on recursive random forest approaches. the aim of the current commentary is to analyse the performance of this already trained predictive model on the molecules of the second “solubility challenge”. even when our training set has inconsistencies related to the ph, solid form and temperature conditions of the solubility measurements, the model was able to predict the two sets from the second “solubility challenge” with statistics comparable to those of the top ranked models. finally, we provided a knime automated workflow to predict aqueous solubility of new drug candidates, during the early stages of drug discovery and development, for ensuring the applicability and reproducibility of our model. ©2021 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords second solubility challenge; quantitative structure-property relationship (qspr); knime; aqueous solubility; adme; machine learning; random forest; supervised recursive variable selection introduction pharmacokinetic parameters are usually influenced by a combination of different physicochemical properties. among these, solubility has occupied a very important role due to its influence on the absorption process. the need to balance solubility, avoiding excess or insufficiency, is a challenge from the perspective of drug discovery. in this regard, several research efforts have been made to provide accurate prediction of aqueous solubility through quantitative structure-property relationship (qspr) approaches. undoubtedly, the first http://dx.doi.org/10.5599/admet.979 https://doi.org/10.5599/admet.979 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:macabreraster@gmail.com mailto:christophe.molina@pikairos.com http://creativecommons.org/licenses/by/4.0/ falcón-cano et al. admet & dmpk 9(3) (2021) 209-218 210 and second “solubility challenges” proposed by llinas et al. have been a very effective indicator of the progress and state-of-art of solubility estimation [1,2]. recently, llinas et al. have reviewed the results of the second “solubility challenge” to analyse the evolution of the computational methods used in this prediction task and the influence of data quality on the results [3]. in our previous publication, we presented a new method based on recursive random forest approaches to predict aqueous solubility values of drug and drug-like molecules [4]. it was based on the development of two novel recursive machine-learning approaches used for data cleaning and variable selection, and a consensus model generated by the combination of regression and classification algorithms. this model was able to provide good solubility prediction compared to many of the models described in the literature. considering that our model was developed from a database of aqueous solubility values with limited information on the experimental conditions of the solubility assay, could our model successfully predict the intrinsic solubility values of the two sets of drugs used in the second "solubility challenge"? the present study describes the performance of our model with the molecules of the second “solubility challenge” and the comparison of the results with those obtained with the best performing models of the competition. it is necessary to clarify that, for this task, the model was not trained, retrained or optimized based on the molecules of the challenge tests, i.e., the model parameters or hyper-parameters remained exactly the same as those set in previously published work [4]. materials and methods challenge sets the second “solubility challenge” consisted of evaluating the intrinsic solubility estimation of two sets of drugs. the first set is composed of 100 drugs with an average inter-laboratory standard deviation estimated of ~0.17 log units. the second test set consists of 32 “difficult” drugs, characterized by poor inter-laboratory reproducibility: standard deviation ~0.62 log units. a detailed list of these molecules have been shown in a previous paper [3]. software the konstanz information miner (knime) is a free and public software tool that has become one of the main analytical platforms for innovation, data mining and machine learning. the flexibility of workflows developed in knime to include different tools allows users to read, create, edit, train and test machine learning models, greatly facilitating the automation of predictions and application by any user [5,6]. in this study, we used the open source software knime analytical platform version 4.0.2 [7] and its free complementary extensions for transformation, analysis, modelling, data visualization and data prediction. for the generation of molecular descriptors from structures, the “descriptor” node from “alvadesc” extension [8] and the “rdkit descriptor” node [9] were employed. modelling dataset to predict the molecules of the second “solubility challenge”, we used as the training set the curated set of aqueous solubility published in our previous paper. this set consists of two large aqueous solubility databases [10,11]. for each molecule, taking the smiles (simplified molecular input line entry specification) code as input format, a structure cleaning, standardization, and duplicate removal protocol was developed. the inchi (iupac international chemical identifier) code was used for duplicate identification and the standard deviation among experimental measurements was computed. a detailed description of this procedure has been shown in our previous article [4]. although the hypothesis that -the admet & dmpk 9(3) (2021) 209-218 a knime retrospective contribution to solubility doi: http://dx.doi.org/10.5599/admet.979 211 quality of the experimental data is the main limiting factor in predicting aqueous solubilityhas been challenged [12], any variability in the experimental protocol is always “noise” for in silico modelling purposes. in this sense, our model had several challenges such as: 1) the ph value for the solubility measurement of the collected compounds was not stated, 2) the solid form of the molecule (polymorphs, hydrates, solvates, amorphous) was not characterized in the reported solubility measurements, 3) it was not possible to verify the type of solubility measurement (kinetic or thermodynamic) and 4) the experimental measurement method was not specified. modelling algorithm due to the uncertainty of the database, we considered the importance of a rigorous protocol for data selection in the development of the original model, in order to discriminate those molecules with potential unreliability. as a first step, we selected a reliable test set, consisting of molecules with more than one reported measurement and with inter-source standard deviation greater than 0 and less than 1 logarithmic unit. we used beyond 1 logarithmic unit as a threshold to discriminate unreliable samples. this reliable test set was used for model optimization. from the qspr perspective, it is necessary to select a set of descriptors that leads to the most predictive model and facilitates model interpretation. to this end, we developed a recursive variable selection algorithm based on regression random forest (rrf). rrf is a widely used ensemble method that assembles multiple decision trees and outputs the consensus predictions from individual trees [13]. it is recognized for its ability to select “important” descriptors. based on this ability, we use the number of occurrences of a variable in the rrf as a measure of the descriptor's importance, combined with a correlation analysis between variables to avoid collinearity. each numerical descriptor was injected in the rrf in two ways: nonshuffled and shuffled. once the individual decision trees were trained and extracted from the ensemble, the total number of occurrences of each variable was calculated. only variables with a number of occurrences greater than a marginal threshold of 110 were retained. among those, variables were discarded if the non-shuffled variable had a number of occurrences lower than the number of occurrences of its homologous shuffled variable. all shuffled variables were eventually discarded too. the final set of variables was selected recursively by initially computing the linear correlation between variables, and then keeping only those with the highest number of occurrences among variables with a correlation coefficient greater than a threshold of 0.51 between them. in an attempt to reduce the uncertainty of the data, independent of any external set, a cleaning procedure based on an rrf approach was developed. this procedure uses the prediction variance (pv) of the rrf as a metric to discriminate unreliable samples. the pv is an rrf score that highlights the variability of each individual prediction with respect to the mean. a high pv can be a sign of anomalous behaviour or uncertainty. this procedure was applied to the unreliable set, i.e. molecules with aqueous solubility standard deviation between sources equal to 0 or greater than 1. to set the parameters of this algorithm, the minimization of the root mean squared error (rmse) of the reliable test was used as the objective function. first, the unreliable set was randomly divided into two sets of 50 % and 50 % cardinal. a regression random forest was trained on one of the two sets and used to predict the aqueous solubility and pv of the other set. in addition, the pv of the out-of-bag samples was also calculated. recursively, molecules were classified as within the pv threshold (clean data) or alternatively as beyond the pv threshold (unclean data), until no molecules changed from clean to unclean labelled set or vice versa. using the clean set, a gradient boosting model (gbm) was trained for classification using logs = -2 as the cut-off to label molecules into highly soluble or soluble and slightly soluble or insoluble. two independent rrf models were developed based on these two subsets of labelled molecules and one more http://dx.doi.org/10.5599/admet.979 falcón-cano et al. admet & dmpk 9(3) (2021) 209-218 212 rrf model was trained on all clean data. finally, the average prediction among the three gbm models was assumed as the final prediction value. the parameters of all models were optimized based on the rmse minimization of the reliable test set. full details on our developed algorithm are given in previous published paper [4]. second “solubility challenge” prediction first, we ensured that all test set molecules found in the initial source set used as the training set were removed. since the model was previously validated using the reliable test set and by 5-fold crossvalidation, we used the entire database (including the reliable test set) to predict the test challenge samples. to analyse the performance of the solubility regression models, two types of coefficient of determination (r 2 ), root mean squared error (rmse), mean absolute error (mae), bias and the percent of molecules with an absolute error less than 0.5 logarithmic units (% 0.5 log) were calculated. results and discussion model performance the statistics obtained for both sets (test set 1 = 100 molecules and test set 2 = 32 molecules) are shown in table 1 and figure 1. to demonstrate model robustness, the results are reported as mean and standard deviation (std). table 1. performance of the final consensus model for the molecules of the second “solubility challenge” test r 2 (validation) r 2 (pearson) rmse (validation) mae (validation) bias % 0.5 log mean std mean std mean std mean std mean std mean std test set 1 0.458 0.01 0.58 0.01 0.925 0.03 0.74 0.03 -0.234 0.01 40 1 (n = 100) test set 2 0.777 0.02 0.78 0.01 1.019 0.1 0.77 0.1 -0.278 0.02 40 6 (n = 32) figure 1. plot of log s (predicted) vs log s0 (experimental) for both test sets. molecules with residual values higher than 0.5 (logarithm units) are highlighted in red. figure 2 compares our results with the top-rank models of the second “solubility challenge”. according to the mean rmse value, our consensus model ranks ninth among the top-ranked models for the prediction of test set 1 and first for the prediction of test set 2. admet & dmpk 9(3) (2021) 209-218 a knime retrospective contribution to solubility doi: http://dx.doi.org/10.5599/admet.979 213 figure 2. comparison between the top-rank models of the second solubility challenge and our results (according to rmse) although there are no significant differences in terms of prediction performance, the training set we have used contains aqueous solubility measurements under non-specified experimental conditions (ph, method and solid form), without information on their type of solubility (aqueous or intrinsic). it is known that the presence of acidic and basic groups in a molecule and the ph of the medium affect the solubility value. intrinsic solubility corresponds to the solubility of the uncharged molecular species, whereas aqueous solubility depends on the ph used for measurements. therefore, not all the values in the training set are true intrinsic solubility values, which influences the model prediction of the external test set with intrinsic solubility measurements, leading in some cases to higher uncertainty for samples contained in the training set. we analysed the overlap of our source set with the molecules from the second "solubility challenge", resulting on two overlaps of 88 and 21 molecules, 1 st and 2 nd test respectively. only for the case of these 109 overlapping molecules, a correlation analysis was performed between the intrinsic solubility values reported in the second "solubility challenge" and the aqueous solubility values reported in our initial source set. the overlapping molecules were eliminated from the training set for modelling purposes. this analysis is shown in figure 3. considering the lack of real intrinsic solubility values in the training set, the most problematic molecules in the second "solubility challenge" should be the ionizable compounds. the analysis of residuals showed that amiodarone (ts2), cisapride (ts1) and folic acid (ts1) are response outliers. all of them contain at least one acidic or basic functional group and are practically insoluble compounds. for these molecules, the aqueous solubility value (log sw) is different from the intrinsic solubility value, since not enough solute is dissolved to modify the ph in order to maintain a near-neutral species in the poorly buffered medium. table 2 describes the values of log s0 (second "solubility challenge"), log sw (initial data source), log sw (reported in other sources) and log sw (predicted). http://dx.doi.org/10.5599/admet.979 falcón-cano et al. admet & dmpk 9(3) (2021) 209-218 214 figure 3. overlapping log s0 against log sw analysis between the molecules of the second “solubility challenge” and the training set. for modelling purposes, these overlapping molecules were eliminated from the training set. table 2. summary of solubility values for the outliers structure name log s0 a log sw b (initial source set) log sw (predicted) log sw c (other sources) o o n i o i amiodarone -10.4 -9.35 -7.54 -7.17 [14] nh n o o o o f nh 2 cl cisapride -6.78 -5.23 -4.27 -4.7 [15] folic acid -5.96 -5.44 -3.12 > -2.87 [15] a intrinsic aqueous solubility reported in the second “solubility challenge”, b aqueous solubility reported for the three outliers in the initial source set, c aqueous solubility reported in other sources to assess whether the method was able to deal with the uncertainty in the data, a simple experiment was performed. as shown in figure 3, 88 molecules from the first test set of the challenge overlapped with our initial source set. a correlation analysis between the two solubility values reported by each overlapping molecule showed a root mean squared error of 0.568 log units. we assume that the value reported in the challenge refers to a curated and reliable measurement, whereas the value reported in our initial source set could be of potential uncertainty. there is a significant difference between the two sets of values for the 88 admet & dmpk 9(3) (2021) 209-218 a knime retrospective contribution to solubility doi: http://dx.doi.org/10.5599/admet.979 215 molecules (confidence interval (ci): 95 %; p = 2.9e-5). next, a paired-sample t-test was developed for comparing the performance of two models based on two different training sets: (a) the literature solubility data reported in our initial source set and (b) the reliable intrinsic solubility measurements reported in the first set of the challenge. both models were evaluated on the second challenge test. there was no significant difference (ci: 95%, p = 0.58) between the root mean squared errors achieved on the second challenge test using one or the other training sets. however, if a single random forest regression without recursive selection of data and variables and without applying a consensus model is used as the modelling algorithm, the t-test highlights a significant difference (ci: 95%; p = 3.3 e-6). the influence of data quality on model performance depends on the modelling procedure used. thus, data quality was not the determinant factor when an appropriate modelling approach was designed to address data uncertainty by selecting the most important variables and using a consensus model of combined single model predictions. table 3 shows a review of the results. table 3. mean with std statistics based on two training sets when predicting the second test of the second “solubility challenge” using our method (recursive random forest (consensus)) versus a single rrf: reliable solubility measurements (data challenge) and literature solubility data. *the results are reported as mean (std). the std was computed by repeating 10-times the modelling procedure. automated system for aqueous solubility prediction we trust there is a need to make publicly available a reliable and diverse data set of intrinsic solubility measurements for a rigorous comparison between modelling algorithms, due to the relative influence of data quality on the performance of a model. furthermore, applicability and reproducibility of solubility qspr models should be a priority for data to be findable, accessible, interoperable and reusable (fair) [16–18]. in this regard, the final purpose of the current commentary is to make publicly available an automated system for in silico aqueous solubility assessment. our model has been successfully validated in a previous published study and has been blind tested with the second “solubility challenge”, showing an adequate performance. the knime workflow published with the paper contains the results of our model on the second “solubility challenge” and allows the prediction of new sets. the user can download the workflow and follow the instructions it contains from https://pikairos.eu/download/aqueous _solubility_prediction/. we developed a version based on rdkit and alvadesc descriptors, calculated using the “descriptor” node contained in the “alvadesc” extension. alvadesc 1.0.16 is available with academic or commercial licenses, which can be obtained by requesting a quote online (registration required) or by contacting them directly by email (chm@kode-solutions.net). only the smiles codes of the structures are needed for aqueous solubility prediction, as the model does not require any experimentally determined value for solubility calculation. the model is characterized by its simplicity since it is only based on 0-2d descriptors. in addition, the model is implemented in the open-source analytics platform knime, which is a user-friendly software suitable for further data analysis and visualization. test reliable solubility measurements (data challenge) n (training) = 88 literature solubility data (reported in initial data source) n (training) = 88 r 2 (validation)* rmse (validation)* r 2 (validation)* rmse (validation)* recursive random forest (consensus) 0.30 (0.05) 1.79 (0.06) 0.29 (0.05) 1.80 (0.05) single random forest regression 0.19 (0.01) 1.93 (0.02) 0.14 (0.06) 1.98 (0.06) http://dx.doi.org/10.5599/admet.979 https://pikairos.eu/download/aqueous_solubility_prediction/ https://pikairos.eu/download/aqueous_solubility_prediction/ mailto:chm@kode-solutions.net falcón-cano et al. admet & dmpk 9(3) (2021) 209-218 216 conclusions the results obtained with the evaluation of the second “solubility challenge” reinforce the idea that data quality is not the major limiting factor for obtaining adequate solubility predictions if the implemented modelling methodology can cope with data uncertainty. in our case, the developed algorithm was able to overcome data variability to obtain acceptable aqueous solubility prediction results. the results published here are a blind prediction, since the experimental aqueous solubility values of the challenge test set were not accessible at the time of our model development and training. although the achieved performance is comparable to those reported in the review of the second solubility challenge, our model is only based on public data compared to some of the best models of the second solubility challenge, which were based on the huge aqueous solubility databases available from pharmaceutical companies. furthermore, the algorithm of our model is global, as demonstrated by the use of generic data without the bias of "training close to the test data". the automation of the proposed methodology and its possible application on larger databases, collected under more homogeneous conditions, could be a step forward to improve solubility prediction during drug discovery and development stages. in attention to the importance of sharing data and methods to ensure reproducibility and applicability of qspr models, we made the data publicly available along with our predictive model based on the knime analytical platform as a new free tool for the assessment of aqueous solubility of drug candidates. abbreviations adme: absorption-distribution-metabolism-excretion qspr: quantitative structure-property relationship knime: konstanz information miner rf: random forest rrf: regression random forest std: standard deviation yiobs: experimental intrinsic solubility value yicalc: predicted aqueous solubility value (model) r2 (val): the square of the correlation coefficient of regression (validation). r2 (val) = r2 = 1 σi (yiobs yicalc)2 / σi (yiobs <yobs>)2, where yiobs is the experimental log s0 and <yobs> is the mean value of the experimental log s0 values. r2 (pearson): the square of the correlation coefficient of regression (pearson). r2 (pearson) = r2 = 1 σi (yiobs a byicalc)2 / σi (yiobs <yobs>)2, where yiobs is the experimental log s0, <yobs> is the mean value of the experimental log s0 values, a is the intercept and b the slope. rmse: the root mean squared error. rmse = 1/n σi (yiobs – yicalc) 2]1/2, where yobs/ ycalc = observed/calculated value of logs0, n = number of samples. mae: mean absolute error. mae = 1/n σi |yiobs – yicalc|, where yobs/ ycalc = observed/calculated value of log s0, n = number of samples. bias = 1/n σi (yiobs – yicalc), where yobs/ ycalc = observed/calculated value of log s0, n = number of samples. ts: test set pvs: prediction solubility variance ci: confidence interval fair: findable, accessible, interoperable and reusable admet & dmpk 9(3) (2021) 209-218 a knime retrospective contribution to solubility doi: http://dx.doi.org/10.5599/admet.979 217 acknowledgements: all the authors acknowledge knime and its many contributors for making the knime data-mining environment available free of charge, as well as alvascience for the academic licence of alvadesc. conflict of interest: the authors declare no conflict of interest. references [1] a. llinàs, r. c. glen, j. m. goodman. solubility challenge : can you predict solubilities of 32 molecules using a database of 100 reliable measurements?. j. chem. inf. model. 48 (2008) 1289–1303. https://doi.org/10.1021/ci800058v. [2] a. llinas, a. avdeef. solubility challenge revisited after ten years, with multilab shake-flask data, using tight (sd ∼ 0.17 log) and loose (sd ∼ 0.62 log) test sets. j. chem. inf. model. 59 (2019) 3036– 3040. https://doi.org/10.1021/acs.jcim.9b00345. [3] a. llinas, i. oprisiu, a. avdeef. findings of the second challenge to predict aqueous solubility. j. chem. inf. model. 60, (2020) 4791–4803. https://doi.org/10.1021/acs.jcim.0c00701. [4] g. falcón-cano, c. molina, m. á. cabrera-pérez. adme prediction with knime: in silico aqueous solubility consensus model based on supervised recursive random forest approaches. admet dmpk 8 (2020) 1–23. https://doi.org/10.5599/admet.852. [5] p.m. mazanetz, j.r. marmon, b.t.c. reisser, i. morao. drug discovery applications for knime: an open source data mining platform. curr. top. med. chem. 12 (2012) 1965–1979. https:/doi.org/10.2174/156802612804910331. [6] m.-a. trapotsi. development and evaluation of adme models using proprietary and opensource data. university of hertfordshire, 2017. https://doi.org/10.18745/th.19719. [7] “knime analytics platform 4.0.2.” [online]. available: https://www.knime.com/download-previousversions. [accessed: 17-mar-2021]. [8] a. mauri, “alvadesc: a tool to calculate and analyze molecular descriptors and fingerprints,” in ecotoxicological qsars. methods in pharmacology and toxicology, k. roy, ed. humana press inc., 2020, pp. 801–820. [9] “rdkit knime integration.” [online]. available: https://www.knime.com/rdkit. [accessed: 19-jun2020]. [10] m.c. sorkun, a. khetan, s. er. aqsoldb, a curated reference set of aqueous solubility and 2d descriptors for a diverse set of compounds. sci. data 6 (2019) 1–8, dec. 2019. https://doi.org/10.1038/s41597-019-0151-1. [11] q. cui, s. lu, b. ni, x. zeng, y. tan, y.d. chen, h. zhao. improved prediction of aqueous solubility of novel compounds by going deeper with deep learning. front. oncol. 10 (2017) 1–9. https://doi.org/10.3389/fonc.2020.00121. [12] d.s. palmer, j.b.o. mitchell. is experimental data quality the limiting factor in predicting the aqueous solubility of druglike molecules?. mol. pharm. 11 (2014) 2962–2972. https://doi.org/10.1021/mp500103r. [13] v. svetnik, a. liaw, c. tong, j.c. culberson, r.p. sheridan, b.p. feuston. random forest: a classification and regression tool for compound classification and qsar modeling. j. chem. inf. comput. sci. 43 (2003) 1947–1958. https://doi.org/10.1021/ci034160g. [14] m. salahinejad, t.c. le, d.a. winkler. aqueous solubility prediction: do crystal lattice interactions help?. mol. pharm. 10 (2013) 2757–2766. https://doi.org/10.1021/mp4001958. [15] s.h. yalkowsky, y. he, p. jain. handbook of aqueous solubility data, second. 6000 broken sound parkway nw, suite 300 boca raton, fl 33487-2742, usa: crc press taylor & francis group, 2010. [16] m. d. wilkinson, m. dumontier, i.j. aalbersberg et al. comment: the fair guiding principles for scientific data management and stewardship. sci. data 3 (2016) 1–9. https://doi.org/10.1038/ sdata.2016.18. http://dx.doi.org/10.5599/admet.979 https://doi.org/10.1021/ci800058v https://doi.org/10.1021/acs.jcim.9b00345 https://doi.org/10.1021/acs.jcim.0c00701 https://doi.org/10.5599/admet.852 https://doi.org/10.2174/156802612804910331 https://doi.org/10.18745/th.19719 https://www.knime.com/download-previous-versions https://www.knime.com/download-previous-versions https://www.knime.com/rdkit https://doi.org/10.1038/s41597-019-0151-1 https://doi.org/10.3389/fonc.2020.00121 https://doi.org/10.1021/ci034160g https://doi.org/10.1021/mp4001958 https://doi.org/10.1038/sdata.2016.18 https://doi.org/10.1038/sdata.2016.18 falcón-cano et al. admet & dmpk 9(3) (2021) 209-218 218 [17] j. wise, a.g. de barron, a. splendiani et al. implementation and relevance of fair data principles in biopharmaceutical r&d. drug discovery today 24, (2019) 933–938. https://doi.org/10.1016/j.drudis. 2019.01.008. [18] k.m. merz, r. amaro, z. cournia, m. rarey, t. soares, a. tropsha, h.a. wahab, r. wang. editorial: method and data sharing and reproducibility of scientific results. j. chem. inf. model. 60 (2020) 5868–5869. https://doi.org/10.1021/acs.jcim.0c01389. ©2021 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) https://doi.org/10.1016/j.drudis.2019.01.008 https://doi.org/10.1016/j.drudis.2019.01.008 https://doi.org/10.1021/acs.jcim.0c01389 http://creativecommons.org/licenses/by/3.0/ immobilized artificial membrane chromatography: from medicinal chemistry to environmental sciences doi: 10.5599/admet.553 225 admet & dmpk 6(3) (2018) 225-241; doi: http://dx.doi.org/10.5599/admet.553 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index review immobilized artificial membrane chromatography: from medicinal chemistry to environmental sciences fotios tsopelas 1* , chrysanthos stergiopoulos 1 and anna tsantili-kakoulidou 2 1 laboratory of inorganic and analytical chemistry, school of chemical engineering, national tech nical university of athens, iroon polytechniou 9, 157 80 athens, greece. 2 laboratory of pharmaceutical chemistry, school of pharmacy, university of athens, panepistimiopolis, zografou, 157 71 athens, greece. *corresponding author: e-mail: ftsop@central.ntua.gr phone: +30 210 7724022, +30 210 7723210 received: may 21, 2018; revised: august 15, 2018; available online: september 22, 2018 abstract immobilized artificial membrane (iam) chromatography constitutes a valuable tool for medicinal chemists to prioritize drug candidates in early drug development. the retention on iam stationary phases encodes lipophilicity, electrostatic and other secondary interactions contrary to traditional octanol-water partitioning. an increasing number of publications in recent years have suggested that iam indices, including isocratic log k(iam) or extrapolated log kw(iam) retention factors, chromatographic hydrophobicity index chi(iam) or the polarity parameter δlog kw(iam) can successfully model the passage of xeniobiotics through biological membranes and barriers and predict pharmacokinetic properties, often in combination with additional descriptors. examples referring to the modeling of human oral absorption, blood-brain penetration and skin partition are described. more recently, iam chromatography has been applied to estimate toxicological endpoints in regard to drug safety, such as phospholipidosis potential, or in regard to chemical risk hazards including the bioconcentration factor and aquatic organisms’ toxicity. the promising results in both medicinal chemistry and in environmental science in combination with the speed, reproducibility and low analyte consumption suggest that a broader application of iam chromatography in the early drug discovery process and in ecotoxicity may save time and money in initial drug candidate selection and will contribute to a reduced risk hazard of chemicals. keywords iam chromatography; lipophilicity; phospholipophilicity; membrane permeability; human oral absorption; ecotoxicity; aquatic toxicity; bioconcentration factor (bcf) introduction to iam chromatography high-performance liquid chromatography (hplc) is a powerful analytical technique based on the differentiation in the elution of different compounds in a given sample when passed through a chromatographic column (stationary phase) by the flow of a mobile phase. thus, dissolved in the mobile phase chemicals participate in a dynamic equilibrium between the mobile and stationary phase and their elution expresses their distribution coefficients between the two phases, depending on various primary and secondary interactions. as several pharmacokinetic and ecotoxicological properties involve a dynamic distribution of xenobiotics between general circulation and tissues, or aqueous environment and tissues of http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:ftsop@central.ntua.gr f. tsopelas, c. stergiopoulos, a. tsantili-kakoulidou admet & dmpk 6(3) (2018) 225-241 226 aquatic organisms, hplc can be used in order to predict pharmacokinetic properties and environmental risk indices. the strategy and the key advantages of using hplc as a tool to achieve biomimetic conditions have been addressed by valko [1,2]. the use of a chromatographic index (e.g. retention factor) as a measure of a biological endpoint (e.g. pharmacokinetic properties, toxicity and ecotoxicological profile) has generated quantitative retention activity relationships (qrar) [3]. the main advantage of qrar is that it is not necessary to know the values of the molecular descriptors of the compound under investigation in order to estimate its pharmacokinetic or ecotoxicological profile as in classic quantitative structure-activity relationship (qsar) studies. this is important in the case of new compounds. the first attempt of introducing hplc to the modelling of biological processes with the employment of reversed-phase liquid chromatography as an alternative for lipophilicity assessment was made in order to overcome certain disadvantages of direct partitioning experiments [4]. lipophilicity was known as a predominant determinant in drug transport across membranes as proved by the traditional hansch analysis [5]. the development of biomimetic chromatographic columns provided a substantial advance for the experimental evaluation of admet properties in the early stages of drug/chemical development, combining the simulation of biological processes with rapid measurements [6-8]. in pharmaceutical and environmental sciences three types of biomimetic chromatography are used; immobilized artificial membrane (iam), immobilized plasma protein chromatography as well as biopartitioning micellar chromatography (bmc). the first two types are based on biomimetic stationary phases, iam and immobilized plasma protein (i.e. human serum albumin and alpha-1 acid glycoprotein), respectively. the third uses reversed phase columns and the biological environment is mimicked by micellar mobile phases obtained upon addition of a surfactant above its critical micellar concentration [9,10]. the retention under biomimetic conditions is referred to as biomimetic properties which may play an essential role in the drug design process of selecting the appropriate compound with the right pharmacokinetic/ toxicological profile to become an effective drug [11]. biomimetic chromatography has the advantage of combining simulation of the biological environment with high speed, is economic and user friendly and requires only a small amount of the analyte without the necessity to be pure [12,13]. of particular importance is the modelling of biological membranes which constitute barriers, which bioactive compounds have to cross in order to reach their molecular target and exert their pharmacological actions or their potential adverse effects. membrane permeability can be modeled by a range of approaches, starting with the traditional octanol-water partitioning, or liposomes partitioning and proceeding to the development of different cell cultures or artificial membrane assays, iam chromatography however, constitutes a powerful alternative due to its superior performance in terms of the simplicity of the experiments, high throughput, reproducibility and the possibility of automation. in this present review, the latest advances in iam chromatography in both the pharmaceutical and the environmental sciences are presented and discussed. types of iam stationary phases and their interrelation iam stationary phases are prepared by immobilization of phospholipids (e.g. phosphatidylcholine) on propylamino-silica skeleton. at present, three types of iam columns are commercially available; the single chain iam.pc (pc stands for “phosphatidylcholine”); the double-chain iam.pc.mg (mg stands for “methylglycolate”) and the iam.pc.dd2 (dd stands for “drug discovery”). according to its producer iam.pc and iam.pc.mg are suggested for “membrane protein purification”, while iam.pc.dd2 is recommended admet & dmpk 6(3) (2018) 225-241 iam chromatography in medicinal chemistry and environment doi: 10.5599/admet.553 227 for drug discovery purposes. iam.pc.dd2 is available in 3, 10 and 15 cm long columns, iam.pc in 15 and 30 cm columns, while iam.pc.mg is produced in 3 and 15 cm length columns. all types of the columns have an i.d. of 4.6 mm and a stationary phase of 10 μm particle size. an iam fastscreen mini column (1 cm x 3mm) is also available for rapid estimations of drug permeability in highthroughput screening programs for the characterization of large libraries of compounds [14]. according to the literature, not only the iam.pc.dd2 column but also the iam.pc.mg column are widely employed in pharmaceutical [15-20] and environmental research [21]. the difference between the two columns is in the end-capping of the free (unreacted) propylamino residues, performed either by methylglycolate or decyl/propyl anhydride as described in detail in the literature [6,7,22-24]. previous comparative investigations on the two iam stationary phases using a structurally diverse set of compounds [16,18,25-27] showed a similar underlying elution mechanism with an almost 1:1 correlation between the retention factors measured in a pure aqueous phase, log kw(iam). a representative interrelationship based on 161 structurally diverse compounds is expressed by eq.(1) [27]: log kw(iam.dd2) = 0.1437 + 1.0757 · log kw(iam.pc.mg) (1) (n=161, r 2 =0.93, s=0.314) where n is the number of compounds, r 2 is the correlation coefficient and s is the standard deviation. the similarity between the two columns permits the combination of retention factors measured on both iam stationary phases in order to model biological processes. iam indices and conditions used the most common approach for the employment of iam chromatography is the simple determination of the retention factor of the solute, according to eq. (2): 0 0 r ( ) t t k iam t  (2) where tr is the retention of the test compound and t0 is the elution time of an unretained compound, such as sodium citrate, lcysteine or ammonium oxalate [28]. the iam retention factor, k(iam), is proportional to the partition coefficient, k(iam), between the iam stationary phase and the mobile phase according to equation (3): s= × (iam) (iam) m v k k v (3) where vs and vm are the volumes of stationary and mobile phase, respectively, whose ratio for a given column is constant. for iam.pc.dd2, vs/vm is equal to 0.053 [29] and, therefore, the k(iam)/k(iam) relation is given according to eq. (4): 1 r 0 = × = 18.9 × (iam) (iam)0.053 0 t t k k t (4) in modelling biological processes or ecotoxicity, retention factors are usually used in their logarithmic form in order to achieve linear relationships with free energy. iam measurements are performed at ambient temperature [17,20,26,30-34] contrary to biopartitioning micellar chromatography which is usually employed at 36.5-37.0 °c [9, 10]. mobile phases consist of a buffer, typically phosphate-buffered saline (pbs) in order to mimic physiological conditions. a limitation in the preparation of mobile phases is f. tsopelas, c. stergiopoulos, a. tsantili-kakoulidou admet & dmpk 6(3) (2018) 225-241 228 their ph value, which should be adjusted to between 2.5 and 7.4. although it is common to add an organic modifier to the mobile phases, iam stationary phases can also be used with a pure aqueous phase [26,28,31,34,35]. in such cases, the measured log k(iam) values correspond to the actual log kw(iam) values, where the symbol “w” denotes the pure aqueous phase. in the case of compounds exhibiting strong affinity with the iam stationary phase, acetonitrile can be added up to concentrations of 30 % and log kw(iam) values can be obtained by linear extrapolation of isocratic log k(iam) values, measured in the presence of at least four different acetonitrile fractions (φ) in the mobile phase, according to eq.(5): log = log × (iam) w(iam)k k s φ (5) where s is the slope of the log k vs φ relationship, calculated by linear regression. comparative studies between actual and extrapolated log kw(iam) values showed only small differences [17,26]. such findings permit the combined use of actual log kw(iam) values with extrapolated values in the case of lipophilic compounds. however, the use of mobile phases with high concentration of acetonitrile should be avoided as the structure of water layer on the stationary phase surface can be disrupted [36]. the use of methanol as an organic modifier should also be avoided as it can cause instability of the stationary phase due to methanolysis of the phospholipids [37,38]. to this point, the high interand intra-laboratory reproducibility of actual or extrapolated log kw(iam) values under standard conditions (buffer pbs and acetonitrile as organic modifier) should be highlighted [7]. as an alternative to using iam retention factors, valko has proposed the chromatographic hydrophobicity index for iam chromatography, chi(iam) [1,39,40], which, in recent literature, has been also termed as “membrane binding index” (iam mb) [41]. chi(iam) corresponds to the percentage of acetonitrile required for the equal partitioning of the solute between the mobile and the stationary phase (e.i. log k=0), which according to eq.(5) is equal to φ0 = log kw(iam)/ s. the rapid determination of chi(iam) [1,39] can be easily performed by the measurement of the gradient elution times (tg) measured under the appropriate gradient conditions. the obtained tg for each analyte is converted to its chi(iam) value by using a linear calibration equation regression line chi(iam)= f(tg) obtained by using 9 standard compounds, namely octanophenone, heptanophenone, hexanophenone, valerophenone, butyrophenone, propiophenone, acetophenone, acetanilide and paracetamol [39]. in such experiments usually, ammonium acetate is the buffer, which is compatible with mass spectrometry, while acetonitrile is the organic modifier [39]. chi(iam) can be further converted to log kw(iam) for the neutral form of the molecule by means of eq.(6) which was established from 86 diverse drug molecules [11]. log kw(iam)=0.046 chi(iam) +0.42 (6) (n=86, r 2 =0.94, s=0.29) relationship of iam retention with lipophilicity and other molecular factors in order to unravel the potential applications of iam chromatography in pharmaceutical and environmental sciences, it is crucial to investigate the elution mechanism, governing the iam retention mechanism. there is general agreement that iam retention is governed by lipophilicity, involving hydrophobicity, polarity and the additional factors, of electrostatic interactions in the case of ionizable compounds (e.g. deprotonated acids, protonated bases and zwitterions) with the charged centres of the phospholipids and the consequent formation of ionic bonds [17,21,26,31,34,38]. in such cases, iam retention may be considered as a border case between passive diffusion and binding as was highlighted by van balen et al. [42]. the specific contribution of each parameter to the so-called “phospholipophilicity” admet & dmpk 6(3) (2018) 225-241 iam chromatography in medicinal chemistry and environment doi: 10.5599/admet.553 229 can be expressed by eq. (7) and has been discussed in a previous review [7,8]: phospholipophilicity = hydrophobicity ± polarity + electrostatic interactions (7) the sign of polarity in eq. (7) can be negative reducing phospholipophilicity in an analogous manner to its effect on lipophilicity, or positive as a result of hydrogen bonding to the glycerol ester group of the phospholipids. electrostatic interactions lead overall to enhanced retention, although repulsive forces between solute anions and iam phosphate groups may also occur. the crucial role of hydrophobicity to iam retention can be highlighted by considering log kw(iam)/ log p relationships, which are generally characterized by good statistics. in the case of ionizable compounds log kw(iam) may still correlate better with log p than log d as a result of the partial compensation of ionization via electrostatic interactions [26]. however, the relationship of log kw(iam) with log d can be considerably improved by introducing the positively (f + ) or negatively charged (f ) molecular fraction of the analytes [17,21,26,31,34] as additional parameters. positive signs (higher for f + than for f ) are generated, as shown in eq. (8) at ph=7.4, supporting the overall positive contribution of electrostatic interactions. + -log = 0.62 × log + 1.14 × f + 0.48 × f + 0.327.4w(iam)k d (8) (n= 56, r 2 = 0.938, s= 0.413) the roustness of equation (8) has been sucessfully validated by other sets of structurally diverse drugs [17,26]. other authors have confirmed the electrostatic interactions in iam retention using linear free energy relationships (lfer) upon introduction of the ionization correction factors (δ-correction or j + and j for ionic species) as additional parameters [43,44]. a representative equation obtained at ph=7.0 is shown in eq. (9) [44]: log kw(iam)= -0.607 + 0.881 · e – 0.562 · s – 0.348 · a1.794 · b + 2.250 · v -0.245 · j + + 2.010 · j (9) (n=68, r 2 = 0.896, s=0.431) in eq. (9), the solute descriptors are as follows: v is the mcgowan’s characteristic volume, e is the excess molar refraction, s is the dipolarity/polarizability, and a and b are the hydrogen-bond acidity and basicity, respectively. the additional terms, δ-correction or j + and j have positive signs (higher for cations than for anions). a possible explanation for the stronger electrostatic interactions of cations is the location of the iam phosphate groups close to the hydrophobic part of phospholipids, in contrast to the choline nitrogen which is more exposed to the solvent effect at the outer terminal of the iam surface [21,26]. thus, phosphate anions enhance the partitioning of cations into the iam stationary phase by attraction forces, while as already commented, repulsive forces may occur with anions [21,26]. however, in the case of very hydrophilic anions, partitioning into the hydrophobic core may not be the determinant factor and attraction by the positively charged choline nitrogen becomes similarly important. in such cases, retention is mainly governed by electrostatic interactions [17]. comparative block relevance analysis of pls models based on volsurf descriptors further supports the similar information content of log kw(iam) and log p for neutral compounds. as for bases and acids the analysis of log kw(iam) (and also of log p) generates a positive sign for the block of h-bond donor (hbd) properties in accordance with the enhanced affinity to the iam stationary phase , favoured for cations over anions , while the opposite sign (negative) is observed in the log d model [45]. f. tsopelas, c. stergiopoulos, a. tsantili-kakoulidou admet & dmpk 6(3) (2018) 225-241 230 the contribution of hydrogen bonding has also been reported [7,25,31,34] and barbato [46] has discussed the role of glycerol as a hydrogen acceptor in solute retention. this may be the reason for only the moderate interrelation between log kw(iam) on the two columns, iam.pc.dd2 and iam.pc.mg obtained for a set of 41 flavonoids, mainly as a result of their different content in alcoholic hydroxyl groups [20]. the polar/electrostatic interactions can be expressed as (δlog kw(iam)), which is the difference between the experimental log kw(iam) value and the log kw(iam) estimated by a calibration equation of log kw(iam) versus log p, which is obtained for neutral compounds, interacting with the iam stationary phase by a purely hydrophobic mechanism. δlog kw(iam) is higher for poorly lipophilic compounds, while high lipophilicity produces negative values [16,18,19,25]. as confirmed also by block relevance analysis at least for neutral compounds, δlog kw(iam) is considered as a polarity index, which can be used in qspr studies replacing other polarity parameters such as polar surface area (psa) or δlog poct-alk [45]. recently, the contribution of secondary interactions due to residual silanol and propylamine groups to the elution mechanism on iam stationary phases has been investigated. such secondary interactions are attributed to the change of the net charge of the iam surface as a function of ph. at ph > 5 there is an overall negative charge on the iam surface due to the predominance of ionized silanol groups, while at lower ph protonated propylamine groups prevail resulting in a positive charge. hence, at physiological conditions, an increased retention capacity of protonated bases can be anticipated and therefore, an overestimation of the affinity of such compounds to phospholipids. the resulting electrostatic effect is more pronounced in low salinities of eluents [47]. the role of silanophilic interactions in iam retention has also been reported by other authors [30,34,38]. the identification of the molecular factors expressed in iam retention has provided tools for in silico prediction of phospholipophilicity. ledbetter et al. [48] developed an algorithm for the prediction of log kw(iam) values of compounds based on fragments and correction factors using a data set which consisted of 22 aliphatic and 42 aromatic compounds. recently, russo et al. [27] proposed a model for prediction of log kw(iam) values, involving four in silico calculated parameters; lipophilicity, hydrophilic/ lipophilic balance, molecular size and molecular flexibility. the work was based on their measurements on 205 experimental values made by them on iam.pc.dd2 and iam.pc.mg column types, to minimize inter-laboratory variability, leading to a good degree of accuracy (r 2 = 0.85). the applications of iam chromatography to predict admet properties in pharmaceutical sciences interactions between drugs and immobilized phospholipids, expressed in iam retention, can serve as important indices for various processes, such as drug permeability and absorption (considering passive diffusion as the main process), tissue and subcellular distribution as well as toxicity due to functional and morphological changes in cells as for instance phospholipidosis. comparative studies have shown that iam retention factors correlate with caco-2 permeability [49-51] or with permeability through mdck cell lines, two standard protocols which used to measure permeability in vitro [17]. of significant importance are the applications of iam chromatography to model gastrointestinal absorption/ bioavailability. it is known that passive permeation routes for drugs involve diffusion through the aqueous pores at the tight junctions, the so-called paracellular absorption and transcellular absorption through the enterocyte itself [52]. published investigations have reported mainly the establishment of non linear models for transcellular based human oral absorption (hoa) combining iam retention with other admet & dmpk 6(3) (2018) 225-241 iam chromatography in medicinal chemistry and environment doi: 10.5599/admet.553 231 physicochemical properties, such as molecular weight (mw) or polarity terms. yoon et al. reported a sigmoidal model for a data set of 40 structurally diverse drugs [53] using isocratic retention factors corrected with mw to the nth power. the best fit was obtained for k(iam) with n=2.5. the same group continued their study combining iam retention with hepatic metabolic clearance to model the oral absorption potential for a limited set of 15 drugs [54]. yen et al. reported a reciprocal correlation between the negative value of the iam retention factor (-1/k(iam)) and human intestinal absorption for 52 drugs (r=0.64). this correlation was further improved upon the inclusion of molecular descriptors expressing molecular size and shape, solubility and polarity (r=0.83) [55]. kotecha et al. suggested that as intestinal absorption occurs along the gastrointestinal tract and mainly in the small intestine with a ph gradient, the maximum log kw(iam), measured in a certain ph range (e.g. 4.5-7.4) and defined as best w(iam) log k should be considered [32,33]. the general model proposed by kotecha et al. is a sigmoidal function of %hoa, which can be improved by additional physicochemical parameters, leading to the general eq. (10):  α + α × x + α × x + …+ α xn n1 10 2 2 100 % hoa = 1 + 10 (10) where x1 stands for best w(iam) log k , x2, x3, ..xn are additional physicochemical properties and α0, α1, α2, .., αn are regression coefficients obtained by non-linear fitting. the best model for 27 structurally diverse drugs was obtained using best w(iam) log k in combination with the polar surface area [33]. tsopelas et al. used an extended data set of 85 drugs to fit best w(iam) log k in eq. (10). they obtained successfully validated models by including mw, the abraham’s hydrogen-bond acidity parameter (a), both with a negative sign, and the negatively charged molecular fraction with a positive sign. if zwitterions were excluded from the data set the model will include mw and positively charged molecular fraction (f + ), both with a negative sign. the negative sign of f + indicates that electrostatic interactions with the phosphate anions on the iam surface are over expressed and may be associated with the binding component, reflecting drug-membrane interactions besides membrane permeability. analogous models in terms of statistics and predictive performance were obtained by replacing the iam retention factors with octanol-water distribution coefficients. the main difference between the two models relies on the opposite sign of the f + parameter (positive in the log d best model, negative in the iam model) attraction forces to the negatively charged inner membrane, which do play a role in absorption but are not expressed in the octanol-water system. in addition, the hydrogen bond acidity a is statistically non-significant in the log d best model [17]. iam retention factors have also served to model blood-brain barrier (bbb) penetration in combination with other physicochemical parameters such as molecular volume and structural indices. in an early study, the performance of iam to model bbb penetration (log bb) was compared with octanol-water partitioning. the log bb/ log kw correlations were slightly inferior in terms of statistics when compared to the corresponding log bb/ log p data. molecular volume was statistically significant with a negative contribution in both models [56]. reichel and begley found a significant correlation between the logarithm of brain uptake indices divided by the square root of the molecular weight (log bui) of six biogenic amines with iam retention (r 2 =0,747), although the same indices for six steroids correlated better with computational clogp than with log kw(iam) (r 2 = 0.809 and r 2 = 0.729, respectively) [57]. pehourcq et al. reported a parabolic relationship for the diffusion of cerebrospinal fluid (csf) for a set of eight amyl propionate non-steroidal anti-inflammatory drugs. the inclusion of molecular weight led to an improved model [58]. grumetto et al. established also parabolic relationships for the bbb passage for a set of 14 f. tsopelas, c. stergiopoulos, a. tsantili-kakoulidou admet & dmpk 6(3) (2018) 225-241 232 basic drugs with retention on either iam.pc.mg or iam.pc.dd2 stationary phases. however, the better model obtained with log p directed the authors to use the δlog kw(iam) parameter, assuming that the electrostatic interactions were encoded in the iam retention as a factor that hinders drugs from transmembrane penetration [59]. they established significant linear negative log bb/ δlog kw(iam), which were confirmed for an enlarged set of 40 structurally diverse drugs [25]. more recently, they have proposed a partial least square (pls) model based on isocratic iam retention factors in the presence of 30 % meoh and in silico calculated physicochemical and topological indices. the pls model constructed for 79 structurally unrelated analytes had however slightly inferior statistics to that obtained with retention factors on biopartitioning micellar chromatography in the presence of sodium dodecyl sulfate (sds) [10]. grumetto et al. [19] have found a parabolic relationship between log bb and log kw(iam) or δlog kw(iam) , measured on an iamdd2 column, with an opposite sign for the linear term (r 2 = 0.825 and r 2 = 0.826 respectively). results were inferior when an iam.mg column was used. better parabolic relationships, however, were obtained with log p of the neutral form (r 2 =0.880) [59]. the same research group performed a comparative study between iam retention indices (log kw(iam) and δlog kw(iam)) and pampa using a porcine brain lipid extract for the prediction of in situ bbb permeability where they found the δlog kw(iam) parameter to be more efficient than pampa [19]. yoon et al. [60] developed a classification model based on iam retention divided by mw to the n th power (kw(iam)/mw n ). the best model was obtained using kw(iam)/mw 4 measured at ph values of 5.5 and 7.0, with cutoff values equal to 1.01 and 0.85, respectively. efforts have also been made for the employment of iam chromatography as a tool to model the transdermal transport of drugs. in a comparative investigation, lazaro et al. explored the potential of iam and c18 stationary phases. they suggested that iam can mimic skin partition but that skin permeation is closer to elution on reversed-phase columns [30]. barbato et al. investigated a limited set of 12 drugs and used both the log kw(iam) and the δlog kw(iam) parameters with no correlation of skin permeability have been observed with log kw(iam), but a negative correlation with δlog kw(iam) was evident, indicating that polar and electrostatic interactions hinder molecules from crossing the skin layer, possibly due to repulsive forces [61]. a point of major importance lies in early drug safety assessment. considering the binding component of the iam retention mechanism, a high affinity to iam stationary phases in the case of basic amphiphilic drugs may be considered as an indication of drug-induced phospholipidosis, a situation associated with excessive accumulation of phospholipids within lysosomes [62,63]. investigations have shown good correlation between iam-chi indexes and phospholipidosis risk [64] or phospholipidogenic potential [65]. in terms of chi(iam) a value > 50 may be related to phospholipidosis [41]. in a comparative study of 36 drugs, jiang and reilly developed a simple and rapid physicochemical screening approach based on chromatographic methods for predicting the potential of compounds to induce phospholipidosis [66]. they found an equally satisfactory correlation with the retention on electrokinetic chromatography based on docusate sodium salt (r 2 =0.87 and r 2 =0.86 respectively) or with the logarithm of the volume of distribution log vd calculated from iam chromatography (chi-iam) and human serum albumin chromatography, according to εq. (11) as suggested by hollosy et al [40]. log vd = 0.44 · log k(iam) – 0.22 log k(hsa) -0.66 (11) (n= 179, r 2 = 0.76, s= 0.33) an analogous relationship, resulting from a combination of retention on both iam and hsa stationary admet & dmpk 6(3) (2018) 225-241 iam chromatography in medicinal chemistry and environment doi: 10.5599/admet.553 233 phases was reported by valko et al. [67] to predict the unbound volume of distribution (vdu), defined as the volume of distribution divided by the unbound fraction (fu) of the drug in plasma: vdu=vd/fu. log vdu = 0.43 log k(iam) + 0.23 log k(hsa) 0.72 (12) (n=70, r 2 = 0.84, s=0.32) the unbound volume of distribution is an important pharmacokinetic parameter since it is in principle the reciprocal value of the maximum drug efficiency (demax) defined by braggio et al. [68]. cell accumulation and retention may be very important for drug action in particular for drugs which are intended to penetrate to intracellular targets or intracellular pathogens, as well as to maintain a prolonged presence in the inflammatory cells [69]. encouraging results have been reported by gordon et al. [70] when plotting the log cell concentration, determined by the high-throughput screening assay rapidfire, versus log k(iam) a positive trend was observed with a correlation coefficient r=0.69, while the corresponding correlation with clogp showed a correlation coefficient r=0.42. the use of iam chromatography for the prediction of ecotoxicity the environment is continuously exposed to a great variety of chemical substances coming from industrial, agricultural, municipal and natural physicochemical and biological processes. new compounds are synthesized and tested as drugs, ingredients of cosmetics, food additives, pesticides, plasticizers or for general use in technical and industrial purposes, while others can be produced as a result of natural biological processes [71]. responsible product design should take into consideration the possible risks in human health and to ecosystems. the evaluation of ecotoxicity indices is involved in the reach regulation “no data no market” [72]. among other eco-toxicological endpoints, median toxicity, i.e. lethal concentration lc50 in aquatic organisms and lethal dose, ld50, for terrestrial organisms as well as bioaccumulation are of major importance. experimental measurements on biological organisms for the thousands of compounds for various uses are simply not possible. in silico systems can be employed for ecotoxicological investigations, in which the logarithm of the octanol-water partition coefficient, symbolized in such studies as log kow, is a principal parameter [73]. several authors have proposed linear solvation energy relationships (lser) for ecological risk assessment [71,74-81]. however, failure in the selection of the appropriate model can result in a 1000-fold error in the estimated indices [82]. a popular predictive tool for environmental investigations is the estimation program interface (epi suite) software, freely available from the environmental protection agency (epa) [83]. module bcfbaf, based on the work of meylan et al. [84], calculates fish the bioconcentration factor and ecosar module, based on the work of russon et al. [82], and estimates the aquatic toxicity of compounds. chromatographic techniques offer an experimental alternative for the rapid evaluation of ecotoxicological endpoints. bio-partitioning liquid chromatography (blc), as named by the escuder-gilabert group [85] was the first biomimetic liquid chromatographic approach implemented for ecotoxicological purposes. blc uses micellar mobile phases formed mainly by the addition of the non-ionic surfactant polyoxyethylene(23)lauryl ether (brij35). it has been employed for the predictions of bioconcentration factors [86,87], skin permeability [87] and toxicity (plc50 in fathead minnow) [88] of xenobiotics. in recent years, iam technology has unfolded new perspectives in the field of ecotoxicology. as iam retention factors reflect permeability of xenobiotics through cell membranes governed by passive diffusion as well as membrane accumulation, it is plausible that it can be used to provide ecotoxicological indices to f. tsopelas, c. stergiopoulos, a. tsantili-kakoulidou admet & dmpk 6(3) (2018) 225-241 234 express median toxicity and bioconcentration or to construct a lipophilicity scale. stepnowski and storoniak [89] compared iam and reversed phase retention as well as octanol-water partitioning as lipophilicity indices of hydrophilic imidazolium ionic liquids. the extrapolated log kw(iam) values were higher than the corresponding logkw on reversedphase stationary phases and log p, due to the development of electrostatic interactions of the positively charged nitrogen of the solutes with the phosphate anions of the iam surface [89]. an analogous investigation was performed by janicka [90], who compared iam with reversedphase, cholesterol phase and micellar chromatography. the authors used retention under the different chromatographic conditions as lipophilicity indices for 15 phenoxyacetic and carbamic acid derivatives with potential use as herbicides. comparison with in silico calculated biological endpoints showed that all of the chromatographic indices were promising measures for the partitioning of chemicals in waterplant cuticle, waterhuman serum albumin and waterskin systems [90]. cimpean et al. correlated iam retention factors in the presence of 10% methanol with the nonspecific toxicity (-log lc50) of the fathead minnow fish (pimephales promelas) [91]. hidalgorodriguez et al. [92,93] and fernandez pumarega et al. [94] measured retention on several chromatographic and electrophoretic systems in order to find the best conditions to emulate non-specific aquatic toxicity and soil sorption. as aquatic toxicity endpoints, the lethal concentration (lc50) of fathead minnow fish [92] and the tadpole narcosis concentration ( nar 1 log c       ) [94] were used, while soil sorption was also considered, expressed by the logarithm of the organic carbon normalized sorption coefficient, log koc [93]. chromatographic and ecotoxicological data were submitted to lser analysis. comparison of the solvation parameter models (spm) was based on the distance parameter d (the difference between the coefficients of the respective spms), on the precision of biological-chromatographic correlations, on principal component analysis (pca) of the coefficients of the compared systems and on dendrogram plots. finally, the chromatographic indices selected according to the above tools were directly correlated with the ecotoxicological data. for aquatic toxicity, iam chromatography along with micellar electrokinetic chromatography based on sodium taurocholate showed the best predictive performance [92,94]. furthermore, iam and electrokinetic chromatography based either on sodium dodecyl sulfate or sodium taurocholate provided the most precise correlation models for soil sorption. especially, the iam model showed the lowest experimental error as reflected in the standard deviation between the observed and calculated data [92]. the potential of iam chromatography to predict bioconcentration of pharmaceutical compounds in aquatic organisms was recently investigated by our group [21]. since very limited experimental bioconcentration factors (bcf) of pharmaceutical compounds are available, bcf values predicted by epi suite software were used to compile a data set of 125 structurally diverse drugs. epi suite software employs log kow (log p calculated according to meylan et al. [84]) as a major descriptor for bcf calculations, while for ionic and highly hydrophilic compounds an arbitrary value of log bcf = 0.50 is assigned by the software. therefore, such drugs were excluded from the analysis and were posteriori predicted by the iam model. highly significant linear log bcf/ log kw(iam) values were established. the constructed model was improved by the inclusion of the parameter biowin5, which expresses the decrease in the bioaccumulation tendency as a result of the degradation potential of the compounds. biowin5 is predicted by the epi suite software and is included also in the epi-bcf model. the log bcf/ log kw(iam) relationship is shown in eq.(13) [21]: log bcf = 0.85 · log 7.4w(iam)k 0.47 · biowin5 + 0.08 (13) (n= 77, r 2 = 0.761, s= 0.437) admet & dmpk 6(3) (2018) 225-241 iam chromatography in medicinal chemistry and environment doi: 10.5599/admet.553 235 validation of the constructed models was based on a test set of drugs with experimentally determined bcf values reported in the literature. predictions of the test set were compared with those provided by the epi software and in most cases, they were in favour of the iam model. this can be explained on the basis that log kw(iam) values involved electrostatic interactions due to ionization of the species at a ph close to those of a marine/ freshwater environment, while log kow is refers only to a neutral species expressing only a lipiddriven bioconcentration [95]. in recent work by russo et al. [96], the iam phospholipophilicity of seven bisphenol a analogues was compared with toxicity investigated on four different cell cultures. results showed that the ranking of toxicity according to the cell cultures was partly consistent with the iam retention, indicating that toxicity increases with increasing membrane affinity [96]. unpublished results from our research group have revealed the ability of iam chromatography to model median toxicity of a series of structurally diverse pesticides in aquatic organisms, involving fish species (e.g. fathead minnow and rainbow trout), eastern oyster and water flea as well as in honey bee. in comparison with octanol-water partitioning, models based on log kw(iam) demonstrate equal or superior statistics and, in some cases, they do not require the use of any additional physicochemical or topological parameters [97]. conclusions the nature of iam retention is between passive diffusion and binding underlines its potential for numerous applications in the field of pharmaceutical and environmental sciences. an increasing number of publications suggest that iam indices can be an equal or even a more effective tool than traditional octanol-water partitioning to express the passage of xenobiotics through biological membranes and barriers as well as drug-membrane interactions. the great advantages are speed, reproducibility and low analyte consumption, as well as the flexibility to provide more than one index (different isocratic log k(iam), extrapolated log kw(iam), chi(iam) or δlog kw(iam)) which can be alternatively used in correlation with biological endpoints. in addition, pharmacokinetic parameters like the volume of distribution or the fraction of unbound in tissue can be easily estimated on the basis of iam chromatographic indices, while they may also serve for drug classification according to certain properties, such as cns penetration or phospholipidosis potential. less investigated has been the performance of iam chromatography in predicting the toxicity of candidate drugs and ecotoxicity of xenobiotics as well as the combined use of iam with other biomimetic chromatographic indices. such studies need to be extended to further toxicity endpoints (e.g. cardiotoxicity) and ecotoxicity indices on terrestrial organisms (e.g. honey bee). the promising results obtained mainly in modelling pharmacokinetic and drug efficacy related properties and, more recently, also in ecotoxicology suggest that a broader application of iam chromatography in early drug discovery processes and in environmental science may save time and money in initial drug candidate selection and contribute to a reduced risk hazard of chemicals. the high intralaboratory reproducibility of logkw(iam) and chi(iam) values under analogous conditions permits their incorporation in to databases which could then serve as a tool for classification purposes, as well as to expand the druglike concept, by incorporating limit values or ranges for iam chromatographic indices. f. tsopelas, c. stergiopoulos, a. tsantili-kakoulidou admet & dmpk 6(3) (2018) 225-241 236 acknowledgements: under the ‘‘special grant and support program for scholars’ association members” (grant no. r zn 004-1/2017-2018). references [1] k. valko. application of highperformance liquid chromatography based measurements of lipophilicity to model biological distribution. journal of chromatography a 1037 (2004) 299-310. [2] k. valko. physicochemical and biomimetic properties in drug discovery: chromatographic techniques for lead optimization. wiley, hoboken, nj, 2014. [3] r. kaliszan. chromatography and capillary electrophoresis in modeling the basic process of drug action. tractrends in analytical chemistry 18 (1999) 400-410. [4] j.g. dorsey, m.g. khaledi. hydrophobicity estimations by reversed-phase liquid chromatography. implications for biological partitioning processes. journal of chromatography a 656 (1993) 485-499. [5] a. leo, c. hansch, d. elkins. partition coefficients and their uses. chemical reviews 71 (1971) 525616. [6] m. chrysanthakopoulos, f. tsopelas, a. tsantilikakoulidou. biomimetic chromatography, a useful tool in the drug discovery process. in: e. grunshka, n. grinberg (eds.). advances in chromatography, vol. 51 (2014), pp. 91-125. new york, crc press. [7] f. tsopelas, t. vallianatou, a. tsantilikakoulidou. advances in immobilized artificial membrane (iam) chromatography for novel drug discovery. expert opinion on drug discovery 11 (2016) 473488. [8] f. tsopelas, c. giaginis, a. tsantilikakoulidou. lipophilicity and biomimetic properties to support drug discovery. expert opinion in drug discovery 12 (2017) 885-896. [9] m. salary, m. hadjmohammadi. human serum albuminmimetic chromatography based hexadecyltrimethylammonium bromide as a novel direct probe for protein binding of acidic drugs. journal of pharmaceutical and biomedical analysis 114 (2015) 1-7. [10] g. russo, l. grumetto, r. szucs, f. barbato, f. lynen. determination of in vitro and in silico indexes for the modeling of blood-brain barrier partitioning of drugs via micellar and immobilized artificial membrane liquid chromatography. journal of medicinal chemistry 60 (2017) 3739-3754. [11] k. valko. lipophilicity and biomimetic properties measured by hplc to support drug discovery. journal of pharmaceutical and biomedical analysis 130 (2016) 35-54. [12] c.j. van leeuwen, j.l.m. hermens, risk assessment of chemicals. an introduction. kluwer academic, dordrecht, 1995. [13] v. pliska, b. testa, h. van de waterbeemd (eds.), lipophilicity in drug action and toxicology, vch, weinheim, 1996. [14] regies technologies, inc, 2016. available at http://www.registech.com/ (retrieved at 8-5-2018). [15] f. tsopelas, a. tsantilikakoulidou, m. ochsenkühnpetropoulou. biomimetic chromatographic analysis of selenium species: application for the estimation of their pharmacokinetic properties. analytical and bioanalytical chemistry 397 (2010) 2171-2180. [16] l. grumetto, c. carpentiero, p. di vaio, f. frecentese, f. barbato. lipophilic and polar interaction forces between acidic drugs and membrane phospholipids encoded in iamhplc indices: their role in membrane partition and relationships with bbb permeation data. journal of pharmaceutical and biomedical analysis 75 (2013) 165-172. admet & dmpk 6(3) (2018) 225-241 iam chromatography in medicinal chemistry and environment doi: 10.5599/admet.553 237 [17] f. tsopelas, t. vallianatou, a. tsantilikakoulidou. the potential of immobilized artificial membrane chromatography to predict human oral absorption. european journal of pharmaceutical sciences 81 (2016) 82-93. [18] l. grumetto, g. russo, f. barbato. polar interactions drug/ phospholipids and comparison of their effectiveness in predicting drug human intestinal absorption. international journal of pharmacy 500 (2016) 275-290. [19] l. grumetto, g. russo, f. barbato. immobilized artificial membrane hplc derived parameters vs pampabbb data in estimating in situ measured blood-brain barrier permeation of drugs. molecular pharmaceutics 13 (2016) 2808-2816. [20] f. tsopelas, m. tsagkrasouli, p. poursanidis, m. pitsaki, g. vasios, p. danias, i. panderi, a. tsantili kakoulidou. retention behavior of flavonoids on immobilized artificial membrane chromatography and correlation with cell-based permeability. biomedical chromatography 32 (2018) e4108. [21] f. tsopelas, c. stergiopoulos, l.a. tsakanika, m. ochsenkühnpetropoulou, a. tsantilikakoulidou. the use of immobilized artificial membrane chromatography to predict bioconcentration of pharmaceutical compounds. ecotoxicology and environmental safety 139 (2017) 150-157. [22] h. liu, s. ong, l. glunz, c. pidgeon. predicting drugmembrane interactions by hplc: structural requirements of chromatographic surfaces. analytical chemistry 67 (1995) 355-3557. [23] f. barbato, g. di martino, l. grumetto, m.i. la rotonda. can protonated β-blockers interact with biomembranes stronger than neutral isolipophilic compounds? a chromatographic study on three different phospholipid stationary phases (iamhplc). european journal of pharmaceutical sciences 25 (2005) 379-386. [24] c. giaginis, a. tsantilikakoulidou. alternative measures of lipophilicity: from octanol-water partitioning to iam retention. journal of pharmaceutical sciences 97 (2008) 2984-3004. [25] l. grumetto, g. russo, f. barbato. indexes of polar interactions between ionizable drugs and membrane phospholipids measured by iamhplc: their relationships with data of blood-brain barrier passage. european journal of pharmaceutical sciences 65 (2014) 139-146. [26] f. tsopelas, n. malaki, t. vallianatou, m. chrysanthakopoulos, d. vrakas, m. ochsenkühn petropoulou, a. tsantili-kakoulidou. insight into the retention mechanism on immobilized artificial membrane chromatography using two stationary phases. journal of chromatography a 1396 (2015) 25-33. [27] g. russo, l. grumetto, f. barbato, g. vistoli, a. pedretti. prediction and mechanism elucidation of analyte retention on phospholipid stationary phases (iamhplc) by in silico calculated physicochemical descriptors. european journal of pharmaceutical sciences 99 (2017) 173-184. [28] f. tsopelas, m. ochsenkühnpetropoulou, a. tsantilikakoulidou. void volume markers in reversed phase and biomimetic liquid chromatography. journal of chromatography a 1217 (2010) 2847-2854. [29] s. ong, c. pidgeon. thermodynamics of solute partitioning into immobilized artificial membranes. analytical chemistry 67 (1995) 2119-2128. [30] e. lazaro, c. rafols, m.h. abraham, m. roses. chromatographic estimation of drug disposition properties by means of immobilized artificial membranes (iam) and c18 columns. journal of medicinal chemistry 49 (2006) 4861-4870. [31] d. vrakas, c. giaginis, a. tsantilikakoulidou. the different behaviour of structurally diverse basic and neutral drugs in the immobilized artificial membrane and reversed-phase high-performance liquid chromatography: comparison with octanol-water partitioning. journal of chromatography a 1116 (2006) 158-164. [32] j. kotecha, s. shah, i. rathod, g. subbaiah. relationships between immobilized artificial membrane chromatographic retention and human oral absorption of structurally diverse drugs. international journal of pharmaceutics 333 (2007) 127-135. [33] j. kotecha, s. shah, i. rathod, g. subbaiah. prediction of oral absorption in humans by experimental immobilized artificial membrane chromatography indices and physicochemical descriptors. international journal of pharmaceutics 360 (2008) 96-106. f. tsopelas, c. stergiopoulos, a. tsantili-kakoulidou admet & dmpk 6(3) (2018) 225-241 238 [34] d. vrakas, c. giaginis, a. tsantilikakoulidou. electrostatic interactions and ionization effect in immobilized artificial membrane retention. a comparative study with octanol-water partitioning. journal of chromatography a 1187 (2008) 67-78. [35] f. tsopelas, a. tsantilikakoulidou, m. ochsenkühnpetropoulou. lipophilicity, biomimetic retention profile and antioxidant activity of selenium species. microchemical journal 110 (2013) 711-718. [36] y. marcus, y. migron. polarity, hydrogen bonding, and structure of mixtures of water and cyanomethane. journal of physical chemistry 95 (1991) 400-406. [37] k. morse, c. pidgeon. importance of mobile phase in immobilized artificial membrane chromatography. in: b. testa, h. van de waterbeemd, g. folkers, r. guy (eds.), pharmacokinetic optimization in drug research, verlag helvetica chimica acta, zurich and wileyvch, weinheim, 2001, pp. 351-381. [38] a. taillardatbertschinger, p.a. carrupt, f. barbato, b. testa. immobilized artificial membrane hplc in drug research. journal of medicinal chemistry 46 (2003) 655-665. [39] k. valko, c.m. du, c.d. bevan, d.p. reynolds, m.h. abraham. rapidgradient hplc method for measuring drug interactions with immobilized artificial membrane: comparison with other lipophilicity measures. journal of pharmaceutical sciences 89 (2000) 1085-1096. [40] f. hollosy, k. valko, a. hersey, s. nunhuck, g. keri, c. bevan. estimation of volume of distribution in humans from high throughput hplcbased measurements of human serum albumin binding and immobilized artificial membrane partitioning. journal of medicinal chemistry 49 (2006) 6958-6971. [41] k. valko, s. teague, c. pidgeon. in vitro membrane binding and protein binding (iam mb/pb technology) to estimate in vivo distribution: applications in early drug discovery. admet & dmpk 5(1) (2017) 14-38. [42] g.p. van balen, c.a.m. martinet, g. caron, g. bouchard, m. reist, p.a. carrupt, r. fruttero, a. gasco, b. testa. liposome/ water lipophilicity: methods, information content, and pharmaceutical applications. medicinal research reviews 24 (2004) 299-324. [43] j. li, j. sun, z. he. quantitative structure-retention relationship studies using immobilized artificial membrane chromatography i: amended linear solvation energy relationships with the introduction of a molecular electronic factor. journal of chromatography a 1132 (2006) 174-182. [44] m.h. abraham, w.e. acree, a. fahr, x. liu. analysis of immobilized artificial membrane retention factors for both neutral and ionic species. journal of chromatography a 1298 (2013) 44-49. [45] g. ermondi, m. vallaro, g. caron. learning how to use iam chromatography for predicting permeability. european journal of pharmaceutical sciences 114 (2018) 385-390. [46] f. barbato. the use of immobilized artificial membrane (iam) chromatography for determination of lipophilicity. current computer-aided drug design 2 (2006) 341-352. [47] s.t.j. droge. dealing with confounding ph-dependent surface charges in immobilized artificial hplc columns. analytical chemistry 88 (2016) 960-967. [48] m.r. ledbetter, s. gutsell, g. hodges, s. o’connor, j.c. madden, p.h. rowe, m.t.d. cronin. prediction of immobilized artificial membrane chromatography retention factors using theoretical molecular fragments and structural features. sar and qsar in environmental research 24 (2013) 661-678. [49] b.h. stewart, f.y. chung, b. tait, c.j. blankley, o.h. chan. hydrophobicity of hiv protease inhibitors by immobilized artificial membrane chromatography: application and significance to drug transport. pharmaceutical research 15 (1998) 1401-1406. [50] t. osterberg, m. svensson, p. lundahl. chromatographic retention of drug molecules on immobilized liposomes prepared from egg phospholipids and from chemically pure phospholipids. european journal of pharmaceutical sciences 12 (2001) 427-439. [51] e.c. chan, w.l. tan, p.c. ho, l.j. fang. modelling caco-2 permeability of drugs using immobilized artificial membrane chromatography and physicochemical descriptors. journal of chromatography a 1072 (2005) 159-168. admet & dmpk 6(3) (2018) 225-241 iam chromatography in medicinal chemistry and environment doi: 10.5599/admet.553 239 [52] j. keldenich. measurement and prediction of oral absorption. chemistry and biodiversity 6 (2009) 2000-2013. [53] c.h. yoon, b.s. shin, h.s. chang, l.s. kwon, h.y. kim, s.e. yoo, s.d. yoo. rapid screening of drug absorption potential using the immobilized artificial membrane phosphatidylcholine column and molar volume. chromatographia 60 (2004) 399-404. [54] b.s. shin, c.h. yoon, j.p. balthasar, b.y. choi, s.h. hong, h.j. kim, j.b. lee, s.w. hwang, s.d. yoo. prediction of drug bioavailability in humans using immobilized artificial membrane phosphatidylcholine column chromatography and in vitro hepatic metabolic clearance. biomedical chromatography 23 (2009) 764-769. [55] t.e. yen, s. agatonovickustrin, a.m. evans, r.l. nation, j. ryand. prediction of drug absorption based on immobilized artificial membrane (iam) chromatography separation and calculated molecular descriptors. journal of pharmaceutical and biomedical analysis 38 (2005) 472-478. [56] t. salminen, a. pulli, j. taskinen. relationship between immobilized artificial membrane chromatographic retention and the brain penetration of structurally diverse drugs. journal of pharmaceutical and biomedical analysis 15 (1997) 469-477. [57] a. reichel, d.j. begley. potential of immobilized artificial membranes for predicting drug penetration across the blood-brain barrier. pharmaceutical research 15 (1998) 1270-1274. [58] f. pehourcq, m. matoga, b. bannwarth. diffusion of arylpropionate nonsteroidal anti-inflammatory drugs into the cerebrospinal fluid: a quantitative structure-activity relationship approach. fundamental and clinical pharmacology 18 (2004) 65-70. [59] l. grumetto, c. carpentiero, f. barbato. lipophilic and electrostatic forces encoded in iam-hplc indexes of basic drugs: their role in membrane partition and their relationships with bbb passage data. european journal of pharmaceutical sciences 45 (2012) 685-692. [60] c.h. yoon, s.j. kim, b.s. shin, k.c. lee, s.d. yo. rapid screening of blood-brain barrier penetration of drugs using the immobilized artificial membrane phosphatidylcholine column chromatography. journal of biomolecular screening 11 (2006) 13-20. [61] f. barbato, b. cappello, a. miro, m.i. la rotonda, f. quaglia. chromatographic indexes on immobilized artificial membranes for the prediction of transdermal transport of drugs. farmaco 53 (1998) 655-661. [62] b.r. berridge, l.a. chatman, m. odin, a.e. schultze, p.e. losco, j.t. meehan, t. peters, s.l. vonderfecht. phospholipidosis in nonclinical toxicity studies. toxicologic pathology 35 (2007) 325. [63] u.m. hanumegowda, g. wenke, a. regueiroren, r. yordanova, j.p. corradi, s.p. adams. phospholipidosis as a function of basicity, lipophilicity and volume of distribution of compounds. chemical research in toxicology 23 (2010) 749-755. [64] j.m. alakoskela, p. vitovic, p.k.j. kinnunen. screening for the drugphospholipid interaction: correlation to phospholipidosis. chemmedchem 4 (2009) 1224-1251. [65] a. casartelli, m. bonato, p. cristofori, f. crivellente, g. dal negro, i. masotto, c. mutinelli, k. valko, v. bonfante. a cell-based approach for the early assessment of the phospholipidogenic potential in pharmaceutical research and drug development. cell biology and toxicology 19 (2003) 161-176. [66] z. jiang, j. reilly. chromatography approaches for early screening of the phospholipidosisinducing potential of pharmaceuticals. journal of pharmaceutical and biomedical analysis 61 (2012) 184-190. [67] k. valko, s.b. nunhuck, a.p. hill. estimating unbound volume of distribution and tissue binding by in vitro hplcbased human serum albumin and immobilized artificial membranebinding measurements. journal of pharmaceutical sciences 100 (2011) 849-862. [68] s. braggio, d. montanari, t. rossi, e. ratti. application of drug efficiency index in drug discovery: a strategy towards low therapeutic dose. expert opinion in drug discovery 5 (2010) 609–618. [69] g.w. amsden. advanced-generation macrolides: tissuedirected antibiotics. international journal of antimicrobial agents 18 (supplement 1) (2001) 11-15. f. tsopelas, c. stergiopoulos, a. tsantili-kakoulidou admet & dmpk 6(3) (2018) 225-241 240 [70] l.j. gordon, m. allen, p. artursson, m.m. hann, b.j. leavens, a. mateus, s. readshaw, k. valko, g.j. wayne, a. west . direct measurement of intracellular compound concentration by rapidfire mass spectrometry offers insights into cell permeability. journal of biomolecular screening 21 (2016) 156-164. [71] k.r. hoover, w.e. acree, m.h. abraham. chemical toxicity correlations for several fish species based on the abraham solvation parameter model. chemical research in toxicology 18 (2005) 1497-1505. [72] regulation (ec) no 1907/ 2006 of the european parliament and the council of 18 december 2006 concerning the registration, evaluation, authorisation and restriction of chemicals (reach). available at http://eur-lex.europa.eu/legal-content/en/txt/?uri=celex:32006r1907 (retrieved at 53-2018). [73] european commission, joint research centre, european union reference laboratory for alternatives to animal testing, 2018. retrieved from https://eurl-ecvam.jrc.ec.europa.eu/validation-regulatoryacceptance/environmental-toxicity-fate/environmental-toxicity-bioaccumulation on 23-4-2018. [74] m.h. abraham, c. rafols. factors that influence tadpole narcosis. an lfer analysis. journal of the chemical society, perkin transactions 2 10 (1995) 1843-1851. [75] k.r.bowen, k.b. flanagan, w.e. acree, m.h. abraham, c. rafols. correlation of the toxicity of organic compounds to tadpoles using the abraham model. the science of the total environment 371 (2006) 99-109. [76] h. laue, h. gfeller, k.j. jenner, j.w. nichols, s. kern, a. natsch. predicting the bioconcentration of fragrance ingredients by rainbow trout using measured rates of in vitro intrinsic clearance. environmental science and technology 48 (2014) 9486-9495. [77] n. kluver, c. vogs, r. altenburger, b.i. escher, s. scholz. development of a general baseline toxicity qsar model for the fish embryo acute toxicity test. chemosphere 164 (2016) 164-173. [78] e. perales, j.i. garcia, e. pires, l. aldea, l. lomba, b. giner. ecotoxicity and qsar studies of glycerol ethers in daphnia magna. chemosphere 183 (2017) 277-285. [79] s. bakire, x. yang, g. ma, x. wei, h. yu, j. chen, h. lin. developing predictive models for toxicity of organic chemicals to green algae based on mode of action. chemosphere 190 (2018) 463-470. [80] p. gramatica, e. papa, a. sangion. qsar modeling of cumulative environmental endpoints for the prioritization of hazardous chemicals. environmental science: processes and impacts 20 (2018) 3847. [81] j. he, t. peng, x. yang, h. liu. development of qsar models for predicting the binding affinity of endocrine disrupting chemicals to eight fish estrogen receptor. ecotoxicology and environmental safety 148 (2018) 211-219. [82] c. l. russom, s.p. bradbury, s.j. broderius, d.e. hammermeister, r.a. drummond. predicting modes of toxic action from chemical structure: acute toxicity in the fathead minnow (pimephales promelas). environmental toxicology and chemistry 16 (1997) 948-967. [83] u.s., environmental protection agency (epa). estimation program interface (epi) suite, version 4.11., 2012. u.s. environmental protection agency, office of pollution, prevention and toxics: washington. available at https://www.epa.gov/tsca-screening-tools/epi-suitetm-estimation-program-interface (retrieved at 21-4-2018). [84] w.m. meylan, p.h. howard, r.s. boethling, d. aronson, h. printu, s. gouchie. improved method for estimating bioconcentration/ bioaccumulation factor from octanol/ water partition coefficient. environmental toxicology and chemistry 18 (1999) 664-672. [85] l. escudergilabert, s. sagrado, r.m. villanuevacamanas, m.j. medinahernandez. quantitative retentionstructureactivity relationship studies of local anesthetics by micellar liquid chromatography, analytical chemistry 70 (1998) 28-34. [86] j.m. bermudezsaldana, l. escudergilabert, m.j. medinahernandez, r.m. villanuevacamanas, s. sagrado. modelling bioconcentration of pesticides in fish using biopartitioning micellar chromatography. journal of chromatography a 1063 (2005) 153-160. http://eur-lex.europa.eu/legal-content/en/txt/?uri=celex:32006r1907 https://eurl-ecvam.jrc.ec.europa.eu/validation-regulatory-acceptance/environmental-toxicity-fate/environmental-toxicity-bioaccumulation%20on%2023-4-2018 https://eurl-ecvam.jrc.ec.europa.eu/validation-regulatory-acceptance/environmental-toxicity-fate/environmental-toxicity-bioaccumulation%20on%2023-4-2018 https://www.epa.gov/tsca-screening-tools/epi-suitetm-estimation-program-interface admet & dmpk 6(3) (2018) 225-241 iam chromatography in medicinal chemistry and environment doi: 10.5599/admet.553 241 [87] c-r. yin, l-y. ma, j-g. huang, l. xu, z-g. shi. fast profiling ecotoxicity and skin permeability of benzophenone ultraviolet filters using biopartitioning micellar chromatography based on penetrable silica spheres. analytica chimica acta 804 (2013) 321-327. [88] j.m. bermudezsaldana, l. escudergilabert, m.j. medinahernandez, r.m. villanuevacamanas, s. sagrado. modelling bioconcentration of pesticides in fish using biopartitioning micellar chromatography. journal of chromatography b 852 (2007) 353-361. [89] p. stepnowski, p. storoniak. lipophilicity and metabolic route prediction of imidazolium ionic liquids. environmental science and pollution research 12 (2005) 199-204. [90] m. janicka. correlations between chromatographic parameters and bioactivity predictors of potential herbicides. journal of chromatographic science 52 (2014) 676-684. [91] cimpean d.m., poole c.f. systematic search for surrogate chromatographic models of biopartitioning processes. analyst 127 (2002) 724-729. [92] m. hidalgorodriguez, e. fuguet, c. rafols, m. roses. modelling nonspecific toxicity of organic compounds to the fathead minnow fish by means of chromatographic systems. analytical chemistry 84 (2012) 3446-3452. [93] m. hidalgorodriguez, e. fuguet, c. rafols, m. roses. performance of chromatographic systems to model soilwater sorption. journal of chromatography a 1252 (2012) 136-145. [94] a. fernandezpumarega, s. amezqueta, e. fuguet, m. roses. tadpole toxicity prediction using chromatographic systems. journal of chromatography a 1418 (2015) 167-176. [95] f. grisoni, v. consonni, s. villa, m. vighi, r. todeschini. qsar models for bioconcentration: is the increase in the complexity justified by more accurate predictions? chemosphere 127 (2015) 171-179. [96] g. russo, a. capuozzo, f. barbato, c. irace, r. santamaria, l. grumetto. cytotoxicity of seven bisphenol analogues compared to bisphenol a and relationships with membrane affinity data. chemosphere 201 (2018) 432-440. [97] c. stergiopoulos, d. makarouni, m. ochsenkuhnpetropoulou, a. tsantilikakoulidou, f. tsopelas. immobilized artificial membrane chromatography as a tool for the prediction of ecotoxicity of pesticides. 2018. (unpublished results). ©2018 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ in vitro ph dependent passive transport of ketoprofen and metformin doi: http://dx.doi.org/10.5599/admet.916 57 admet & dmpk 9(1) (2021) 57-68; doi: https://doi.org/10.5599/admet.916 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper in vitro ph dependent passive transport of ketoprofen and metformin alisa elezović* 1 , amina marić 1 , amila biščević 1 , jasmina hadžiabdić 1 , selma škrbo 2 , selma špirtović-halilović 3 , ognjenka rahić 1 , edina vranić 1 , amar elezović 4 1 department of pharmaceutical technology, faculty of pharmacy, university of sarajevo, zmaja od bosne 8, 71000 sarajevo, bosnia and herzegovina 2 department of clinical pharmacy, faculty of pharmacy, university of sarajevo, zmaja od bosne 8, 71000 sarajevo 3 department of pharmaceutical chemistry, faculty of pharmacy, university of sarajevo, zmaja od bosne 8, 71000 sarajevo, bosnia and herzegovina 4 control laboratory of the agency for medicinal products and medical devices, titova 9, 71000 sarajevo *corresponding author: e-mail: alisa.elezovic@ffsa.unsa.ba; tel.: +387-33-586-173; fax: +387-33-586-180 received: october 16, 2020; revised: december 02, 2020; published: december 09, 2020 abstract the kinetics of passive transport of ketoprofen and metformin, as model substances for high and low permeability, respectively, across the artificial membrane under the influence of the ph of donor solution was investigated. there was an upward trend in the apparent permeation coefficient (papp) of ketoprofen with the decrease in ph to a value close to pka. at the ph value below pka the permeation coefficient had lower value, due to the higher retention of ketoprofen in the artificial membrane. metformin is a low permeable compound, and the highest permeation values were recorded at ph 7.4. two dissociation constants determine that metformin at physiological ph exists as a hydrophilic cationic molecule, i.e. predominantly in ionized form. at ph values below 2.8, metformin mainly exists in diprotonated form, and it was, thus, very poorly permeable. the highest retention, i.e. affinity of both ketoprofen and metformin to the membrane, was at the lowest ph values, which is explained by different mechanisms. at higher ph values of donor compartment the substances showed significantly less affinity to the membrane. the obtained values of apparent permeation coefficients at studied ph values showed good correlation with the obtained experimental values by other in vitro methods. ©2021 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords dynamic diffusion cell model; artificial membrane; in vitro permeation; bcs ii; bcs iii introduction the oral route is the most preferred route of drug administration. it is considered safe, effective, and readily available, with minimal discomfort to the patient compared to other routes [1]. the absorption of orally administered drugs is a function of its dissolution, solubility and permeability of the substance along the length of the gastrointestinal tract (git) [2]. the absorption of active substances varies depending on the ph along the gastrointestinal tract [3]. in general, the absorption of active substances mainly takes place in the small intestine, i.e. in the jejunum which has the largest absorption surface [4]. http://dx.doi.org/10.5599/admet.916 https://doi.org/10.5599/admet.916 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:alisa.elezovic@ffsa.unsa.ba http://creativecommons.org/licenses/by/4.0/ alisa elezović et al. admet & dmpk 9(1) (2021) 57-68 58 the biopharmaceutical classification system (bcs) categorizes drugs into four classes according to their water solubility and permeability across intestinal membrane. based on the available data on low solubility and high permeability of ketoprofen, it belongs to class ii (bcs ii) [5], while metformin as a highly soluble and low permeable substance belongs to bsc class iii [6]. ketoprofen is a non-steroidal anti-inflammatory analgesic. within the ph of the gastrointestinal tract, ketoprofen is an ionized substance (weak acid). the pka value of ketoprofen is approximately 4.45 at 25 °c [7]. metformin is often described as an insulin stabilizer because, due to its positive effects on insulin acceptor and tyrosine kinase activity, it leads to a decrease in plasma insulin levels and insulin resistance [8]. metformin has two different values of the dissociation constant, namely pka 2.8 and 11.5, and as a result of which at physiological ph it exists as a hydrophilic cationic molecule. at ph values below 2.8, it mainly exists in diprotonated form, and thus difficult to permeate. at ph values higher than 11.5 it is found in non-ionized form which is the most favorable form for passive transport across membranes [9], however this ph is not found physiologically. models for determining permeability are used in the early stages of drug discovery and development and are a major strategy to prevent drug rejections due to poor pharmacokinetics in the later, more expensive phases of clinical trials [10]. the most reliable way to determine the permeation of active substances is in vivo testing of human intestinal perfusion. however, due to simpler, faster and much less expensive performance compared to the in vivo tests, in situ, ex vivo and in vitro models are mainly used for permeation assessment [11]. in vitro methods are the most commonly used alternative for testing the permeation of active substances [12]. depending on the biopharmaceutical class of the active substance, they may have a good correlation with the in vivo results [13]. most active substances are absorbed by passive transport, so the application of in vitro methods using artificial membranes that mimic passive transport is considered an effective and reproducible method for assessing permeability [12], especially in screening stages of drug discovery. absorption, distribution, metabolism, excretion and toxicity are greatly influenced by the charge of the active substances at different ph conditions. consideration of ph values in relation to other molecular properties is of great importance and has the potential to be used to further improve the efficacy of the application of active substances [14]. the method of permeation testing on artificial membranes, which could reliably predict the passive transport of the active substance as a function of ph, would significantly facilitate the characterization of new candidate substances in the initial stages of drug development. in this work, the kinetics of passive transport of ketoprofen (bcs ii) and metformin (bcs iii), as model substances for high and low permeability, respectively, across the artificial membrane under the influence of the ph value of the donor solution was investigated. the results were compared with the available literature data obtained by different permeation methods of these two substances. experimental materials deionized water was obtained on arium 611, sartorius (sartorius lab instruments gmbh & co. kg, göttingen, germany). the materials used in this experimental work include: disodium hydrogen phosphate admet & dmpk 9(1) (2021) 57-68 passive transport of ketoprofen and metformin doi: http://dx.doi.org/10.5599/admet.916 59 dihydrate (kemika, zagreb, croatia), cholesterol (sarajevolijek, sarajevo, bosnia and herzegovina), potassium dihydrogen phosphate (semikem, sarajevo, bosnia and herzegovina), disodium hydrogen phosphate (kemika, zagreb, croatia), ketoprofen, working standard (lek, slovenia), metformin hydrochloride, reference standard (merck, darmstadt, germany), sodium chloride (merck, darmstadt, germany), n-octanol (semikem, sarajevo, bosnia and herzegovina), hydrochloric acid (lach-ner, sro neratovice, czech republic), 85 % phosphoric acid (mallinckrodt baker bv, deventer, the netherlands). egg lecithin (lipoid® e80) was a kind gift from lipoid gmbh, ludwigshafen, germany. methods in the experimental work, the instructions for use of the apparatus absorption simulator sartorius model sm 16750 (sartorius membranfilter gmbh, germany) and the method of the authors corti et al. [1,15] and our study group [16] were used. cellulose nitrate filter (mixed cellulose ester), pore size 0.45 µm (sartorius ag, göttingen, germany) was impregnated with 2 ml solution of 2.10 % cholesterol and 1.70 % egg lecithin (lipoid® e 80) in n-octanol for 10 minutes. the excess impregnation solution was removed by pressing a filter located between the two absorbent papers. the mass of the filter before and after impregnation was determined using an analytical balance (ag 400, metler toledo gmbh, switzerland) from which the percentage of impregnation solution retained on the filter was calculated as filter mass increase. the filter prepared in this way was placed in a diffusion cell of absorption simulator connected to the donor and acceptor compartments. the donor compartment contained a solution of 500 µmol l -1 of active substance (ketoprofen or metformin) in 100 ml of phosphate buffer of different ph values. the ph values of the buffer in the donor compartment were: 7.4, 6.8, 4.5 and 3.0 or 2.0 for ketoprofen or metformin, respectively. phosphate buffer of ph 7.4 was present in the acceptor compartment in all experiments. the temperature in both compartments was measured and maintained at 37 ± 0.5 °c. under the action of the peristaltic pump, the contents of the donor and acceptor compartment were constantly circulating through the diffusion cell on both sides of the membrane (figure 1). at 30, 60, 90 and 120 minutes, sampling was performed from the acceptor compartment and the concentration of active substances (ketoprofen or metformin) was determined spectrophotometrically. spectrophotometric measurements were performed on uv-vis spectrophotometer uv-1601 (shimadzu, kyoto, japan). the method was validated prior to the permeation experiments according to ich guideline q2-r1 [17]. after reading the results on the spectrophotometer, the samples were returned to the acceptor compartment. at the end (120 minutes), the samples were also taken from donor compartment, and concentration of ketoprofen or metformin was determined spectrophotometrically. this data was used to calculate the recovery value (r). each experiment was performed in triplicate. from the samples taken from the acceptor compartment, the apparent permeability coefficients of ketoprofen and metformin were calculated based on equation 1: 0 1 app dq p dt ac  , (1) where papp apparent permeation coefficient, cm s -1 ; dq/dt mass of the substance that passed through the membrane per unit time, mg s -1 ; a effective surface of the artificial membrane through which the permeation was examined, cm 2 ; c0 initial concentration of active substance in donor solution, mg ml -1 . http://dx.doi.org/10.5599/admet.916 alisa elezović et al. admet & dmpk 9(1) (2021) 57-68 60 figure 1. absorption simulator sartorius model sm 16750 recovery of ketoprofen and metformin (as a percentage of the substance not retained in the membrane) was determined according to equation 2: , , ,0 % x 100 a end a d end d d d c v c v r c v   , (2) where ca, end and cd, end concentration of test substance measured at the end of the experiment in acceptor and donor solution, mg ml -1 ; cd,0 initial concentration of test substance in donor solution, mg ml -1 ; va and vd volumes of acceptor and donor solution, ml. results and discussion permeations of ketoprofen and metformin were monitored in this experimental work at different ph values of the donor compartment, while the ph of the acceptor compartment was 7.4 in all cases. in this way, it was possible to examine the permeation of the substances from the donor compartment at different ph values, which simulated git conditions across an artificial membrane embedded in a diffusion cell in the acceptor compartment that reflected the ph conditions of the systemic circulation. a membrane filter with pore size of 0.45 µm was used in this work. as we mentioned in the earlier work [16] the pore size of the membrane filter does not affect the permeability of active substances across the artificial membrane, since in corti et al. [1] there was no clear causal relationship between papp and pore sizes established. according to the instructions for use of the apparatus (sartorius model sm 16750), the increase in filter weight after impregnation should be 100 ± 5 %. the mean value of the percentage of impregnation of the membrane, i.e. the average increase in filter mass was 158.67 ± 5.59 %. these values are similar to the values of experimental studies conducted by corti et al. [1] and our study group [16]. according to both studies, the structure of a membrane filter that was not originally intended for use in this experiment is responsible for the larger increase in filter mass after impregnation. admet & dmpk 9(1) (2021) 57-68 passive transport of ketoprofen and metformin doi: http://dx.doi.org/10.5599/admet.916 61 permeation of ketoprofen dependent on the ph of the donor solution the ketoprofen permeation profiles' parameters at different ph of the donor compartment are presented in table 1 and figure 2. table 1. mean values (n = 3) of the apparent permeation coefficients (papp) of ketoprofen, the recovery of the substance not retained by the membrane (r) and percentage of calculated (non)ionized species (ha/a ) at different ph values of donor compartment. ph papp (× 10 -5 cm s -1 ) ± sd r(%) ± sd ha/a (%) 7.4 6.95 ± 0.51 97.95 ± 2.42 0.11 / 99.89 6.8 16.8 ± 1.10 87.83 ± 2.77 0.44 / 99.56 4.5 41.6 ± 1.67 83.52 ± 1.68 47.17 / 52.83 3.0 12.9 ± 0.28 51.91 ± 1.79 96.61 / 3.39 figure 2. overall permeated amount of ketoprofen through the artificial membrane at different ph values of donor compartment. large values of the apparent permeation coefficients confirm that it is a highly permeable substance. the low standard deviation of the apparent permeation coefficients shows good reproducibility of the method for highly permeable substances. the recovery values indicate whether ketoprofen was retained in the membrane, and this retention varied significantly depending on the ph value. in general, ionized substances are better soluble, and non-ionized ones pass through the cell membrane better. the calculated percentage of non-ionized ketoprofen decreased from 96.61 % for ph = 3.0 to 0.11 % at ph = 7.4 (table 1). ketoprofen is a weak acid and its maximum permeation is expected in areas where the ph value is close to the pka of the substance, which in the case of ketoprofen is 4.45, where there is 47.17 % non-ionized ketoprofen. these obtained calculated values of the proportions of ionized and nonionized species explain the values of the apparent permeation coefficients and recovery factors at individual ph, discussed below. it can be seen from figure 2 that the equilibrium permeation concentrations were not established at any ph during the 120-minute test period. http://dx.doi.org/10.5599/admet.916 alisa elezović et al. admet & dmpk 9(1) (2021) 57-68 62 in general the values of papp in all ph media were high (table 1). such results indicate high permeability of ketoprofen, which is in line with the literature data showing that ketoprofen belongs to bcs class ii. the recovery value increased with increasing ph (table 1), and for ph = 7.4 the recovery value was 97.95 %. such a high value indicates that a very small percentage of the substance was retained by the membrane. the recovery value decreased with decreasing ph, and the lowest percentage of the recovered substance was measured at ph = 3 and it was r = 51.93 %. in case of weak acids, such as ketoprofen, the proportion of non-ionized substance is increased at ph values below pka, i.e. ph of 3.0 in this case. the increased portion of non-ionized substance makes this substance more lipophilic and thus has a higher affinity for the lipophilic membrane. when analysing the obtained values of apparent permeation coefficients from table 1, it can be noticed that unlike recovery values, papp did not have a linear growth with ph values. the lowest mean value of the apparent permeation coefficient (papp = 6.95 ± 0.51 × 10 -5 cm s -1 ) was at ph = 7.4. regardless of the fact that the recovery value at this ph shows the highest percentage, and papp is the lowest, such results are consistent with the results obtained in the original method. in the original papp method, the value for ketoprofen at ph = 7.4 was 4.27 ± 0.08 × 10 -5 cm s -1 and r = 99.1 % [15]. the apparent permeation coefficient showed the highest values at ph = 4.5, which was to be expected given that the maximum permeation of ketoprofen is expected at ph values that are closest to the pka value of the substance, which for ketoprofen has a value of 4.45. this is due to the fact that at this ph value there is a higher ratio of non-ionized substance, which permeates, compared to higher ph values, and, on the other hand, there is less retention in the membrane compared to lower ph value that is bellow pka value, probably due to the larger portion of the ionized form. the apparent permeation coefficient was 41.6±1.67 × 10 -5 cm s -1 . this result confirmed the high permeability of ketoprofen, and is in line with the fact that ketoprofen is a weak acid, the absorption of which is expected at a weakly acidic to neutral ph values. measurement with phosphate buffer in the donor compartment having a ph of 6.8 resulted in a papp value of 16.8±1.1 × 10 -5 cm s -1 . these values correlate with all the above facts related to the permeation properties of ketoprofen. at ph = 3.0, the mean papp value obtained was 12.9±0.28 × 10 -5 cm s -1 . these results were expected, and it is assumed that if there was no increased affinity of the substance and the membrane, it would be even higher at this ph value. table 2. experimentally obtained apparent permeation coefficients of ketoprofen (cm s -1 ) at different ph values and literature data obtained by other methods. ph papp ± sd (×10 -5 ) fa % a papp ± sd (×10 -5 ) b papp ± sd (×10 -5 ) c peff ± sd (×10 -5 ) d log d e 3.0 12.9 ± 0.28 / / / / / 4.5 41.6 ± 1.67 99.9 6.02 ± 2.4** / / 1.8** 6.8 16.8 ± 1.1 100 1.86 ± 1.5*** 26.0 ± 13.4*** 29.2 ± 0.7*** 0.8 7.4 6.95 ± 0.5 100 0.24 ± 0.7 15.8 ± .24 23.5±0.7 0.1 a values of simulated fa (dose fraction absorbed in humans) [18]; b bml-pampa (biomimetic lipid parallel artificial membrane permeability assay) papp values [19]; c papp values obtained by ussing chambers' model [20]; d spip peff (effective permeation coefficient) values [20]; e log d (distribution coefficient) values [20,21]; * values at ph = 2; **values at ph = 5.5; ***values at ph = 6.5 numerous methods for testing absorption/permeation have been described in the literature. table 2 shows the literature accounts of experimentally obtained values of other methods for testing the permeation of ketoprofen at certain ph values, as well as the literature values of the in silico simulated dose fraction absorbed in humans (fa), and log d. the correlations between the obtained results and these admet & dmpk 9(1) (2021) 57-68 passive transport of ketoprofen and metformin doi: http://dx.doi.org/10.5599/admet.916 63 literature data were examined. in general, this type of testing referring to ph dependence is scarce and lacking. due to the lack of experimental data on the permeation coefficients of ketoprofen at the exact same ph values, the values at approximate ph were taken for certain parameters, as indicated in the legend of table 2. dose fraction absorbed in humans of ketoprofen is 90 %, and it is mainly absorbed in the small intestine [5]. however, due to the lack of ph dependent data on its in vivo absorption, the obtained results cannot be compared or correlated. the results were compared with the available results of in silico simulated fraction of the dose absorbed in humans at different ph values and gave the correlation factor of r 2 = 0.9239. the comparisons of the results of in situ and ex vivo models [20] were possible only at two ph values, namely ph = 6.8 and ph = 7.4, so it was not purposeful to calculate the correlation coefficient. however, the trend of increasing permeation with decreasing ph of the environment in both cases was evident (table 2). the same trend was present for the bml-pampa assay and log d results, even taking into consideration that the ph was different near pka of ketoprofen in the two experiments. comparison of the obtained results with the results of corti et al. [15] and our earlier results [16] was not possible due to the lack of data on apparent permeation coefficients for all ph values. permeation of metformin dependent on the ph of the donor solution the metformin permeation profiles' parameters at different ph of the donor compartment are presented in table 3 and figure 3. the pka value of 11.5 makes metformin a strong base found in non-ionized form in the percentage of only 0.008 % at ph 7.4. given the physicochemical properties of metformin (pka values, log p) that determine its high solubility and poor permeability, metformin will prevail at physiological ph in ionized form, which will favor its solubility, but will not favor passive transport. due to its properties, metformin is mainly absorbed in vivo by active transport via number of transporters located in intestinal membranes [22]. otherwise, for drugs belonging to bcs class iii, it would be possible to establish an in vivo in vitro correlation if the drug is highly soluble (such as metformin), if it is not subject to the first pass effect and is not metabolized (such as metformin) and if it is not transported by carriers. since it does not meet the third criterion, it is not possible to establish an in vitro in vivo correlation for metformin, but it is possible to examine the extent to which passive transport is involved in its permeation and the effects of ph on this type of permeation. table 3. mean values (n = 3) of the apparent permeation coefficients (papp) of metformin, the recovery of the substance not retained by the membrane (r) and percentage of calculated (non)ionized species (bh2 2+ /bh + /b) at different ph values of donor compartment. ph papp (× 10 -8 cm s -1 ) ± sd r(%) ± sd bh2 2+ /bh + /b (%) 7.4 29.82 ± 1.095 96.19 ± 4.92 0.0025 / 99.99 / 0.008 6.8 9.83 ± 0.199 90.00 ± 2.73 0.010 / 99.99 / 0.002 4.5 9.79 ± 0.234 90.37 ± 3.20 1.96 / 98.04 / 9.8 × 10 -6 2.0 5.53 ± 0.556 85.31 ± 4.45 86.32 / 13.68 / 4.33 × 10 -9 http://dx.doi.org/10.5599/admet.916 alisa elezović et al. admet & dmpk 9(1) (2021) 57-68 64 figure 3. overall permeated amount of metformin across the artificial membrane at different ph values of donor compartment. at ph 2.0 (which corresponds to the gastric ph) most of metformin (86.32 %) is in diprotonated form, i.e. it is in the form of a double positively charged ion. in the protonated form, positively charged metformin at this ph value is present at 13.68 %, while the non-ionized form is represented by only 4.33 × 10 -9 %. these data point to the fact that metformin at the ph value of the stomach is very difficult to passively permeate through physiological membranes, which the obtained results confirm. table 3 shows that the papp value was 5.53 × 10 -8 cm s -1 , very low due to the low permeability of the substance at ph 2.0. metformin is at gastric ph, a highly ionized molecule and as such cannot passively permeate through physiological membranes. the recovery value was 85.31 %, indicating a higher retention of metformin in the artificial membrane than at other ph values of the donor compartment, probably due to the affinity of metformin for membrane components made of cellulose nitrate and phospholipids from the impregnation solution. at ph 4.5, the share of the diprotonated form of metformin is 1.95 %, while the single positively charged form dominates (98.04 %). non-ionized form of the molecule is present in a percentage of 9.8 x 10 -6 %. although present in extremely low percentage, the non-ionized form of metformin is still insufficient for unobstructed passive permeation, and the papp value was 9.79 ×10 -8 cm/s at this ph. the ph value of 4.5 is present at the level of the duodenum from where in vivo metformin absorption will start to take place. at ph values of 6.8, metformin is found in the non-ionized form in portion of 0.0019 %. passive permeation in this case is more pronounced than at lower ph values. generally, in vivo metformin absorption mainly takes place from git regions where the ph value is 6.8. however, in vivo absorption is aided by a number of protein carriers and transporters located along the length of the intestinal mucosa that facilitate and enable cellular uptake of metformin from the apical to the basolateral membrane of enterocytes. therefore, an in vitro model based exclusively on passive transport could not adequately map events at physiological ph 6.8, but only the passive component of membrane transport. since the nonionized content is lower than at ph 7.4, the permeation rate is lower at this ph of the environment. the mean papp value shown in table 3 was 9.83 ×10 -8 cm s -1 . admet & dmpk 9(1) (2021) 57-68 passive transport of ketoprofen and metformin doi: http://dx.doi.org/10.5599/admet.916 65 at ph 7.4 metformin is present in the non-ionized form 0.008 %. although the concentration of the uncharged form is extremely low, it is the highest at the studied ph values and as such shows the best permeation rate and extent through the artificial membrane. the ph value of the donor compartment was the furthest from the lower pka value for metformin (2.8), which favored its existence in non-ionized form, and enabled its passive permeation across the artificial membrane. from the results shown in table 3, it can be seen that the mean papp at ph 7.4 was 2.98 × 10 -7 cm s -1 . although this permeation rate is low compared to substances that show good permeability, this is the highest value of permeation obtained in this experimental work due to the fact that at these ph values the highest percentage of non-ionized form is present compared to other studied ph values. thus, metformin is a low permeable compound whose absorption is more favored by alkaline ph values (figure 3). the recovery value of r = 96.19 % (table 3) indicates that metformin mostly uninterruptedly permeated. in summary, metformin is highly ionized throughout the length of the git, which explains its low permeability by passive transport. it also explains the fact that metformin is in bcs class iii (high solubility low permeability). its permeability, as mentioned earlier, is significantly higher in vivo, since it is suggested to be absorbed by active transport through supported carriers [22]. in figure 3, it can be seen that as the ph value increases, the slopes of the metformin permeation profile increase, so the highest papp values were observed at ph 7.4. at this ph value, metformin has good solubility, low lipophilicity, and the highest concentration of non-ionized substance compared to the other studied ph values, thus it more easily permeates into the acceptor compartment with very small percentage of retention in the artificial membrane. the log d value for metformin is -1.43, which indicates low solubility in lipids, and thus low affinity for the membrane, which is lipid in nature. as can be seen from figure 3, after the initial jump in the concentration of permeated metformin after 30 minutes at ph 2.0, almost a plateau is reached, which can be explained in that the saturation of metformin in the membrane is reached and diprotonated molecules repel each other impeding further permeation through the membrane. this is similar, although less pronounced, at higher ph values. the behavior of metformin at ph values of 6.8 and 4.5 corresponding to the small intestine was shown to be very similar (figure 3). even though low permeation rates were observed at both ph values, permeation of metformin at ph 4.5 physiologically present in the duodenum showed a slight, although insignificant, advantage. in humans, metformin absorption is incomplete and bioavailability shows interindividual variability at 55 ± 16 %. absorption ceases 6 10 hours after administration, regardless of the amount of metformin administered, which is approximately the time of passage of the drug through the stomach and small intestine [27]. human data are lacking in ph dependent transport across the git membranes, and thus cannot be further discussed. lassoued et al. [23] in their comparative study of permeation of metformin by the sartorius sm 16750 absorption simulator model and inverted sac technique showed a good correlation in the determination of metformin permeability. the results of the study conducted by song et al. [24] concluded that metformin is transported via saturable carriers. the results were not correlated with dynamic diffusion cell model due to the difference in the used methods. http://dx.doi.org/10.5599/admet.916 alisa elezović et al. admet & dmpk 9(1) (2021) 57-68 66 table 4. experimentally obtained apparent permeation coefficients (cm s -1 ) of metformin at different ph values and literature data obtained by other methods. ph papp ± sd (×10 -6 ) papp± sd a (×10 -6 ) papp ± sd a (×10 -6 ) papp ± sd b (×10 -6 ) papp ± sd b (×10 -6 ) peff ± sd c (×10 -6 ) log d d papp d (×10 -6 ) papp ± sd e (×10 -6 ) 2.0 0.055±0.006 / / / / / / / / 4.5 0.098±0.002 / / / / 45.1±10.8 (duodenum) -1.41 / / 6.8 0.098±0.002 v 0.74±0.15 s 0.61±0.1 / / 32.6±7.3 (jejunum) 29.6 ± 3.6 (ileum) -1.30 -1.31 / 0.50±0.057 7.4 0.298±0.011 / / a 8.6±1.4 b 10.9±1.9 / -1.23 5.5 / a values papp obtained by inverted sac and sartorius model [23]; b values papp obtained on the same sartorius model with the same experimental conditions with impregnation solution variation [16]; c values papp obtained on isolated duodenum, jejunum and ileum segments in caco-2 cell model [24]; d values log d and papp obtained by caco-2 cell model [25]; e values papp obtained by caco-2 cell model [26] however, in the study conducted by nicklin et al. using the caco-2 model it was concluded that metformin is transported in the caco-2 monolayer by passive, unsaturable transport [25]. the hydrophilicity and molecular weight of metformin determined the papp value of 5.5 × 10 -6 cm s -1 , which is comparable with the results of other hydrophilic high molecular weight compounds tested by the caco-2 model [28], but also with the results obtained in this experimental work at the same ph values (table 4). on the other hand, in the caco-2 cell model study conducted by proctor et al. [26] the transport and absorption of metformin took place via saturable transport mechanisms. the relative contribution of transcellular and paracellular transport of metformin at a concentration of 0.05 mm was estimated at 9 % and 91 %, respectively [26]. log d values at the studied ph were below 0.5 which implies poor permeability of metformin within the physiological ph range prevailing at the git level, which is in line with the obtained results. however, the correlation of the results with the available log d values was very poor (r 2 = 0.63). conclusions ketoprofen is a highly permeable acid, while metformin is dibasic substance with low permeability. ph dependent passive transport through an artificial biomimetic membrane of both substances could be explained by their ionization properties at studied ph values of donor compartments. log d values of ketoprofen correlated well, while those of metformin did not. the literature accounts of ph dependent permeation/absorption are scarce. however, when considering those available, the results correlated well with the other in vitro studies. the dynamic diffusion cell model has proven to be an efficient, reproducible and simple method for predicting the permeation of ketoprofen. this model examines passive transport across the membrane, and for substances that are absorbed by other forms of transport, such as metformin, this method can be used to predict only its passive transport component. the dynamic diffusion cell model is a good foundation for future research as a valuable resources-saving screening tool for the new drug candidates at the beginning stages of their development, as well as for known substances permeation elucidation, especially considering its potential for testing at wide range of ph conditions, which is very difficult by other methods. admet & dmpk 9(1) (2021) 57-68 passive transport of ketoprofen and metformin doi: http://dx.doi.org/10.5599/admet.916 67 conflict of interest: the authors declare no conflict of interest. references [1] g. corti, f. maestrelli, m. cirri, s. furlanetto, p. mura. development and evaluation of an in vitro method for prediction of human drug absorption assessment of artificial membrane composition, european journal of pharmaceutical sciences 27 (2006) 346-353. https://doi.org/10.1016/ j.ejps.2005.11.004. [2] g.l. amidon, h. lennernas, v.p. shah, j.r. crison. a theoretical basis for a biopharmaceutic drug classification: the correlation of in vitro drug product dissolution and in vivo bioavailability, pharmaceutical research 12 (1995) 413-420. https://doi.org/10.1023/a:1016212804288. [3] a. avdeef. physicochemical profiling (solubility, permeability and charge state), current topics in medicinal chemistry 1 (2001) 277-351. https://doi.org/10.2174/1568026013395100. [4] h. lennernas. intestinal permeability and its relevance for absorption and elimination, xenobiotica 37 (2007) 1015–1051. https://doi.org/10.1080/00498250701704819. [5] i. shohin, j. kulinich, g. ramenskaya, b. abrahamsson, s. kopp, p. langguth, j. polli, v. shah, d. groot, d. barends, j. dressman. biowaiver monographs for immediate-release solid oral dosage forms: ketoprofen, journal of pharmaceutical sciences 101 (2012) 3593-3603. https://doi.org/10.1002/jps.23233. [6] c.-l. cheng, l. x. yu, h.-l. lee, c.-y. yang, c.-s. lue, c.-h. chou. biowaiver extension potential to bcs class iii high solubility-low permeability drugs: bridging evidence for metformin immediate-release tablet, european journal of pharmaceutical sciences 22 (2004) 297-304. https://doi.org/10.1016/j.ejps.2004.03.016. [7] j. sheng, n. a. kasim, r. chandrasekharan, g. l. amidon. solubilization and dissolution of insoluble weak acid, ketoprofen: effects of ph combined with surfactant, european journal of pharmaceutical sciiences 29 (2006) 306–314. https://doi.org/10.1016/j.ejps.2006.06.006. [8] j.e. gunton, p.j. delhanty, s. takahashi, r.c. baxter. metformin rapidly increases insulin acceptor activation in human liver and signals preferentially through insulin-acceptor substrate-2, journal of clinical endocrinology and metabolism 88 (2003) 1323–1332. https://doi.org/10.1210/jc.2002021394 [9] g.s. clarke, a.e. bretnall, metformin hydrochloride, in analytical profiles of drug substances and excipients, h.g. brittain, ed., academic press, san diego, ca, usa, 1998, 243-284 isbn: 9780122608254. [10] p. berben, a. bauer-brandl, m. brandl, b. faller, g.e. flaten, a.c. jacobsen, j. brouwersa, p. augustijns. drug permeability profiling using cell-free permeation tools: overview and applications, european journal of pharmaceutical sciences 119 (2018) 219–233. https://doi.org/ 10.1016/j.ejps.2018.04.016. [11] d.a. volpe, drug permeability studies in regulatory biowaiver applications, in drug absorption studies, biotechnology: pharmaceutical aspects, c. ehrhardt, k.j. kim, eds., springer, boston, ma, usa, 2008, 665–680. https://doi.org/10.1007/978-0-387-74901-3_28. [12] a. avdeef, solubility, permeability, and charge state, in absorption and drug development, 2 nd ed., a. avdeef, ed., john wiley & sons, inc., new jersey, usa, 2012, 116-246. https://doi. .org/10.1002/9781118286067. [13] r. joubert, j.d. steyn, h.j. heystek, j.h. steenekamp, j.l. du preez, j.h. hamman. in vitro oral drug permeation models: the importance of taking physiological and physico-chemical factors into consideration, expert opinion on drug delivery 14 (2016) 179–187 https://doi.org/ 10.1080/17425247.2016.1211639. [14] d. manallack, r. prankerd, g. nassta, o. ursu, t. oprea, d. chalmers. a chemogenomic analysis of ionization constants-implications for drug discovery, chemmedchem 8 (2013) 242-255. https://doi.org/10.1002/cmdc.201200507. http://dx.doi.org/10.5599/admet.916 https://doi.org/10.1016/j.ejps.2005.11.004 https://doi.org/10.1016/j.ejps.2005.11.004 https://doi.org/10.1023/a:1016212804288 https://doi.org/10.2174/1568026013395100 https://doi.org/10.1080/00498250701704819 https://doi.org/10.1002/jps.23233 https://doi.org/10.1016/j.ejps.2004.03.016 https://doi.org/10.1016/j.ejps.2006.06.006 https://doi.org/10.1210/jc.2002-021394 https://doi.org/10.1210/jc.2002-021394 https://doi.org/10.1016/j.ejps.2018.04.016 https://doi.org/10.1016/j.ejps.2018.04.016 https://doi.org/10.1007/978-0-387-74901-3_28 https://doi.org/10.1080/17425247.2016.1211639 https://dx.doi.org/10.1002%2fcmdc.201200507 alisa elezović et al. admet & dmpk 9(1) (2021) 57-68 68 [15] g. corti, f. maestrelli, m. cirri, n. zerrouk, p. mura. development and evaluation of an in vitro method for prediction of human drug absorption ii demonstration of the method suitability, european journal of pharmaceutical sciences 27 (2006) 354-362. https://doi.org/10.1016/ j.ejps.2005.11.005. [16] a. elezović, j. hadžiabdić, e. vranić, o. rahić, a. elezović (2019): pilot study on variation of dynamic diffusion cell artificial membrane method for the assessment of active substances permeability, research journal of pharmaceutical, biological and chemical sciences 10 (2019) 64-75. https://doi.org/10.33887/rjpbcs/2019.10.3.9. [17] international council for harmonization of technical requirements for pharmaceuticals for human use: validation of analytical procedures: text and methodology q2(r1) (2005). [18] y. tsume, p. langguth, a. garcia-arieta, g.l. amidon. in silico prediction of drug dissolution and absorption with variation in intestinal ph for bcs class ii weak acid drugs: ibuprofen and ketoprofen, biopharmaceutical drug disposition 33 (2012) 366-377. https://doi.org/10.1002/bdd.1800. [19] k. sugano, h. hamada, m. machida, h. ushio, k. saitoh, k. terada. optimized conditions of biomimetic artificial membrane permeation assay, international journal of pharmaceutics 228 (2001) 181-188. https://doi.org/10.1016/s0378-5173(01)00845-6. [20] c. roos, d. dahlgren, e. sjögren, c. tannergren, b. abrahamsson, h. lennernäs. regional intestinal permeability in rats: a comparison of methods, molecular pharmaceutics 14 (2017) 4252-4261. https://doi.org/10.1021/acs.molpharmaceut.7b00279. [21] y.e. arnold, j. thorens, s. bernard, y.n. kalia. drug transport across porcine intestine using an ussing chamber system: regional differences and the effect of p-glycoprotein and cyp3a4 activity on drug absorption, pharmaceutics 11 (2019) 139. https://doi.org/10.3390/pharmaceutics11030139. [22] a.j. scheen. clinical pharmacokinetics of metformin, clinical pharmacokinetics 30 (1996) 359-371. https://doi.org/10.2165/00003088-199630050-00003. [23] m.a. lassoued, f. khemiss, s. sfar. comparative study of two in vitro methods for assessing drug absorption: sartorius sm 16750 apparatus versus everted gut sac, journal of pharmacy and pharmaceutical sciences 14 (2011) 117 – 127. https://doi.org/10.18433/j3gc7r. [24] n.n. song, q.s. li, c.x. liu. intestinal permeability of metformin using single-pass intestinal perfusion in rats, world journal of gastroenterology 12 (2006) 4064-4070. https://doi.org/10.3748/wjg. v12.i25.4064. [25] p. nicklin, a.c. keates, t. page, c.j. bailey. transfer of metformin across monolayers of human intestinal caco-2 cells and across rat intestine, international journal of pharmaceutics 128 (1996) 155–162. https://doi.org/10.1016/0378-5173(95)04259-8. [26] w.r. proctor, d.l. bourdet, d.r. thakker. mechanisms underlying saturable intestinal absorption of metformin, drug metabolism and disposition 36 (2008) 1650-1658. https://doi. org/10.1124/dmd.107.020180. [27] g.g. graham, j. punt, m. arora, r.o. day, m.p. doogue, j.k. duong, t.j. furlong, j.r. greenfield, l.c. greenup, c.m. kirkpatrick, j.e. ray, p. timmins, k.m. williams. clinical pharmacokinetics of metformin, clinical pharmacokinetics 50 (2011) 81–98. https://doi.org/10.2165/11534750000000000-00000. [28] j. karlsson, p. artursson. a method for the determination of cellular permeability coefficients and aqueous boundary layer thickness in monolayers of intestinal epithelial (caco-2) cells grown in permeable filter chambers, international journal of pharmaceutics 71 (1991) 55-64. https://doi.org/10.1016/0378-5173(91)90067-x. ©2021 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) https://doi.org/10.1016/j.ejps.2005.11.005 https://doi.org/10.1016/j.ejps.2005.11.005 https://doi.org/10.33887/rjpbcs/2019.10.3.9 https://doi.org/10.1002/bdd.1800 https://doi.org/10.1016/s0378-5173(01)00845-6 https://doi.org/10.1021/acs.molpharmaceut.7b00279 https://doi.org/10.3390/pharmaceutics11030139 https://doi.org/10.2165/00003088-199630050-00003 https://doi.org/10.18433/j3gc7r https://dx.doi.org/10.3748%2fwjg.v12.i25.4064 https://dx.doi.org/10.3748%2fwjg.v12.i25.4064 https://doi.org/10.1016/0378-5173(95)04259-8 https://doi.org/10.1124/dmd.107.020180 https://doi.org/10.2165/11534750-000000000-00000 https://doi.org/10.2165/11534750-000000000-00000 https://doi.org/10.1016/0378-5173(91)90067-x http://creativecommons.org/licenses/by/3.0/ validated dbs method for filgotinib quantitation in rat dried blood spots and its application to a pharmacokinetic study in rats doi: http://dx.doi.org/10.5599/admet.796 139 admet & dmpk 8(2) (2020) 139-148; doi: http://dx.doi.org/10.5599/admet.796 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper validated dbs method for filgotinib quantitation in rat dried blood spots and its application to a pharmacokinetic study in rats abhishek dixit, vinay kiran, bhavesh babulal gabani, ramesh mullangi* drug metabolism and pharmacokinetics, jubilant biosys ltd, industrial suburb, yeshwanthpur, bangalore -560 022, india *corresponding author: e-mail: mullangi.ramesh@jubilantinnovation.com; tel.: +91-80-66628339 received: february 24, 2020; revised: april 23, 2020; published online: may 06, 2020 abstract filgotinib is a selective jak1 (janus kinase) inhibitor, showed efficacy in patients suffering from moderateto-severe rheumatoid arthritis. in this paper, we present the data on the development and validation of a sensitive, selective and high-throughput lc-ms/ms (liquid chromatography with tandem mass spectrometry) method for the quantitation of filgotinib from rat dried blood spot (dbs) cards. to the dbs disc cards, 0.2 % formic acid enriched with internal standard (is) was added and sonicated. thereafter the extraction of filgotinib and the is (tofacitinib) was accomplished using ethyl acetate as an extraction solvent. the resolution of filgotinib and the is was achieved on a gemini c18 column with an isocratic mobile phase, which is a mixture of 0.2 % formic acid:acetonitrile (20:80, v/v) at a flow-rate of 0.9 ml/min. the total run time was 2.90 min and the retention time of filgotinib and the is was ~1.31 and 0.89 min, respectively. filgotinib and the is were analyzed using positive ion scan mode and parent-daughter mass to charge ion (m/z) transition of 426.3291.3 and m/z 313.2149.2, respectively, for quantitation. the calibration range was 1.37-1937 ng/ml. no matrix effect and carry over were observed. all the validation parameters met the acceptance criteria. the validated method has been applied to a pharmacokinetic study in rats. a good correlation between dbs and plasma concentrations for filgotinib was observed. ©2020 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords filgotinib; lc-ms/ms; method validation; rat blood; dbs; pharmacokinetics introduction rheumatoid arthritis is a chronic autoimmune inflammatory disease with a prevalence of 1-2 % of the world population. rheumatoid arthritis primarily affects peripheral joints and subsequently damages the synovial tissue and cartilage damage [1]. although disease-modifying anti-rheumatic drugs (dmards: methotrexate, sulfasalazine, leflunomide etc) and biologic dmards (etanercept, adalimumab etc.) are available as first-line drugs due to their low therapeutic benefit and severe side effects they cannot be used for long-treatment [2]. to overcome these drawbacks, janus kinase (jak)/signal transducer and activator of the transcription (stat) signal pathway has been identified as one of the new therapeutic targets to treat rheumatoid arthritis. jak-stat pathway plays a critical role in the downstream signaling of cytokines. the http://dx.doi.org/10.5599/admet.796 http://dx.doi.org/10.5599/admet.796 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:mullangi.ramesh@jubilantinnovation.com http://creativecommons.org/licenses/by/4.0/ r. mullangi et al. admet & dmpk 8(2) (2020) 139-148 140 inhibition of jaks is an attractive therapeutic target to treat rheumatoid arthritis [3]. tofacitinib and baricitinib are the pan-jak inhibitors (jak1/jak3) approved for the treatment of moderate to severe rheumatoid arthritis but they showed dose-related toxicities [4]. recent findings suggest that selective inhibition of jak1 might reduce the toxicity without a significant detriment to efficacy [2,3]. filgotinib (figure 1; glpg0634), is a selective jak1 inhibitor (ic50: 629 nm) with 30-fold selectivity over jak2 [5]. it has shown dose-dependent reduction of disease progression in a collagen-induced arthritis model post oral administration in rodents [5]. in phase-3 clinical trials, filgotinib was well tolerated, showed efficacy and was found to be safe in rheumatoid arthritis patients, when it was administered as monotherapy at 100 or 200 mg, once daily or with methotrexate [6]. filgotinib is currently being filed in the us and japan to treat the patients suffering from moderate-to-severe rheumatoid arthritis [7,8]. filgotinib tofacitinib (is; internal standard) figure 1. structural representation of filgotinib and tofacitinib (is). to date, two lc-ms/ms (liquid chromatography with tandem mass spectrometry) methods were reported for the quantification of filgotinib. the first method, reported by namuor et al. (2015) employed solid-phase extraction method for the quantitation of filgotinib and its active metabolite from phase-i study plasma samples [100 µl plasma sample was spiked with 20 µl of deuterated filgotinib (125 ng/ml) as an internal standard]. the reported lower limit of quantification (lloq) was 3.00 ng/ml. however, full details on chromatography, mass spectrometer conditions and validation parameters were not reported [9]. very recently, dixit et al. (2020) reported a validated lc-ms/ms method for the quantification of filgotinib. the authors have attained an lloq of 0.78 ng/ml with 50 µl rat plasma. plasma samples were processed using ethyl acetate as an extraction solvent [10]. dried blood spot (dbs) methodology has seen significant progress during recent times in the quantitative analysis of various drugs [11-16]. dbs offers several advantages over traditional sampling (plasma/blood/serum) techniques such as reduction of commercial costs for laboratory equipment, convenience in collection, reduction in collection of blood volume, no requirements for trained phlebotomist, ease of sample handling/storage/shipping, safety in handling, less time in processing and increase in throughput etc. it is anticipated that much clinical development and therapeutic drug monitoring programs in the future may switch to dbs technique to characterize the pharmacokinetic data. to the best of our knowledge, there is no dbs method reported for the quantitation of filgotinib. in this paper, we report the development and validation of an lc-ms/ms method for the quantitation of filgotinib on rat dbs. the applicability of the validated method was shown in a rat pharmacokinetic study. excellent correlation was observed between dbs versus plasma filgotinib concentrations, indicating that the dbs method can be used as an alternative for plasma sampling for pharmacokinetic analysis. admet & dmpk 8(2) (2020) 139-148 dbs method for filgotinib quantitation doi: http://dx.doi.org/10.5599/admet.796 141 materials and methods materials filgotinib (purity: >95 %) was purchased from angene international limited, tsuen wan, hong kong. tofacitinib (is; purity: 98 %) was purchased from sigma-aldrich (st. louis, usa). hplc grade acetonitrile and methanol were purchased from j.t. baker, pa, usa. analytical grade formic acid was purchased from s.d fine chemicals, mumbai, india. all other chemicals and reagents were of analytical grade and used without further purification. the control sprague dawley rat blood was procured from animal house, jubilant biosys, bangalore. liquid chromatography and mass spectrometry conditions we used the similar chromatographic conditions that we previously reported [10]. in brief, a gemini c18 (100 × 4.6 mm, 3 µm) column maintained at ambient room temperature along with an isocratic mobile phase (0.2 % formic acid in milli-q water:acetonitrile, 20:80, v/v) at a flow-rate of 0.9 ml/min was used for separation of filgotinib and the is. the injection volume was 5.0 µl. under these optimized conditions the retention times were 1.31 and 0.89 min, for filgotinib and the is, respectively with a total run time of 2.90 min. the lc was coupled to a sciex 4000 mass spectrometer controlled by sciex analyst 1.6.2 software. the ms was equipped with electro-spray ionization and operated in the multiple reaction monitoring (mrm) mode for the quantitation. ionization was conducted by applying a voltage of 5500 v, and the source temperature was set at 550 °c. the gas settings were as follows curtain gas: 35 psi, gs1: 50 psi, gs2: 55 psi and cad: 8.0 psi. the compound parameters declustering potential, entrance potential, collision energy, and collision cell exit potential were set at 91, 10, 40, and 10 v for filgotinib and 100, 10, 41, and 12 v for the is. the mass transitions m/z (precursor ionproduct ion) 426.3291.3 and 313.2149.2 were monitored for filgotinib and the is, respectively. quadrupole q1 and q3 were set on the unit resolution. dwell time was 150 ms. preparation of stock solutions and standard samples two separate primary stock solutions of filgotinib were prepared at 1670 g/ml in dmso. appropriate secondary and working stocks of filgotinib were prepared from primary stock by successive dilution of primary stock with methanol:water (80:20, v/v) to prepare the calibration curve (cc) and quality controls (qcs). the is primary stock solution was made in dmso at a concentration of 1000 µg/ml, which was diluted with methanol to 500 ng/ml and subsequently used as is working stock solution. the primary stock solutions of filgotinib and the is were stored at -20 °c, which were found to be stable for 50 days. working stock solutions were stored at 4 °c for 25 days. blood spotting freshly drawn rat blood was used to prepare the dbs cards. with the help of a calibrated pipette, 25 μl of the respective spiked cc/qc blood or whole blood collected from the pharmacokinetic study (post administration of filgotinib by intravenous and oral routes) was sampled on dbs card. spiked cards were allowed to dry at ambient room temperature for 3 h and stored appropriately in a sealed bag with desiccant in a desiccator. dbs homogeneity the spot homogeneity was evaluated by punching out the disc from the periphery of the dbs. blood spots at qc low and qc high level were prepared in triplicate. the obtained dbs discs were processed and http://dx.doi.org/10.5599/admet.796 r. mullangi et al. admet & dmpk 8(2) (2020) 139-148 142 analyzed as described in the sample preparation section. dbs sample preparations using a hole puncher (harri-micro-punch®), 3 mm diameter disc was punched from the center of each dbs (fta dmpk-c card) card directly into micro-centrifuge tubes. to each micro-centrifuge tube, 200 µl of 0.2 % formic acid enriched with 100 ng/ml of is was added, thereafter the contents were vortex mixed for 5 min (thermomixer ® , eppendorf) and sonicated (elmasonic s300 h) for 15 min at ambient room temperature. after sonication, to the same micro-centrifuge, 1.0 ml of ethyl acetate was added and the mixtures were vortexed further for 3 min. the samples were centrifuged at 14,000 rpm for 3 min. clear organic layer (800 µl) was pipetted out after centrifugation and dried under gentle stream of nitrogen (turbovap ® , zymark ® ,kopkinton, ma, usa) at 50 °c. the residue was reconstituted with 200 µl of 0.2 % formic acid in acetonitrile and 150 µl clear supernatant was aliquoted into a hplc vial and 5.0 µl was injected onto the column for lc-ms/ms analysis. validation procedure the validation experiments were performed in accordance with the us food and drug administration guideline [17]. various parameters covered under validation are: selectivity, carry over (repetitive punching and auto-injector), recovery, matrix effect, linearity, precision, accuracy, stability (in-injector, 6 h at room temperature and 30-day long-term at -80 ± 10 °c) and incurred sample reanalysis. influence of hematocrit two different hematocrit (hct) blood samples at 30 and 50 % were prepared at lqc and hqc by adding or removing the plasma to the rat blood. these samples (in quadruplicate at each qc) were analysed with calibrators prepared in blood at standard fixed 40 % hct. the relative error of ±15 % and precision of ±15 % was considered acceptable. hct value in rats ranged between 38-46 % (in-house data; measured using mindray bc-5000vet) and it is in line with reported values [18]. pharmacokinetic study in rats all the animal experiments were approved by institutional animal ethical committee (iaec/jdc/2019/189r). male sprague dawley rats (n=8) were procured from vivo biotech, hyderabad, india. the animals were housed in jubilant biosys animal house facility in a temperature (22 ± 2 °c) and humidity (30-70 %) controlled room (15 air changes/hour) with a 12:12 h light:dark cycles, had free access to rodent feed (altromin spezialfutter gmbh & co. kg., im seelenkamp 20, d-32791, lage, germany) and water for one week before using for experimental purpose. following ~12 h fast (during the fasting period animals had free access to water) animals were divided into two groups having four rats in each group. group i animals (205-210) received filgotinib orally as a suspension formulation (prepared using tween-80 and 0.5 % methyl cellulose) at 10 mg/kg (strength: 1.0 mg/ml; dose volume: 10 ml/kg), whereas group ii animals (215-225 g) received filgotinib intravenously (5 % dmso, 5 % solutol:absolute alcohol (1:1, v/v) and 90 % of normal saline; strength: 0.5 mg/ml; dose volume: 4 ml/kg) at 2.0 mg/kg dose. post-dosing serial blood samples (50 µl) were collected through retro-orbital plexus into polypropylene tubes containing k2.edta solution as an anti-coagulant at 0.25, 0.5, 1, 2, 4, 8, 10, 12 and 24 h (for oral study) and 0.083, 0.25, 0.5, 1, 2, 4, 8 and 24 h (for intravenous study). animals were allowed to access feed 2 h post-dosing. pharmacokinetic analysis blood concentration-time data of filgotinib was analyzed by non-compartmental method and the admet & dmpk 8(2) (2020) 139-148 dbs method for filgotinib quantitation doi: http://dx.doi.org/10.5599/admet.796 143 relevant pharmacokinetic parameters namely auc0- (area under the blood concentration-time curve from time zero to infinite time), c0 (extrapolated blood concentration at time zero), cmax (maximum blood concentration), tmax (time to reach cmax), vd (volume of distribution), cl (clearance) and t½ (terminal halflife) were calculated using phoenix winnonlin software (version 8.1; pharsight corporation, mountain view, ca). absolute oral bioavailability (f) was calculated using this formula [dose (i.v.) × auc(0-)oral / dose (oral) × auc(0-)i.v.] × 100. results and discussion method development the objective of this work was to develop a sensitive and rugged dbs method for the quantification of filgotinib in rat blood and subsequently show the correlation with previously developed method [10] in a pharmacokinetics study so that dbs method can be used as an alternative method to measure the filgotinib concentrations. previously reported lc-ms/ms conditions were utilized for this method [10]. since chromatography and mass spectrometric conditions were optimized, to obtain better sensitivity extraction optimization was done with various extraction solvents namely acetonitrile, methanol, ethyl acetate, acidified methanol/acetonitrile, mixture of water and methanol/acetonitrile, pre-treatment of dbs discs with formic acid followed by extraction with methanol/acetonitrile/ethyl acetate. method validation parameters recovery none of the single organic solvents and pure water gave good recovery of filgotinib from the dbs discs. the recovery with ethyl acetate, water and methanol/acetonitrile was 5, 39 and 43 %, respectively. with methanol/acetonitrile:water 50:50 and 20:80, the recovery was found to be 30 and 43 %, respectively. in order to improve the recovery, the dbs discs were pre-treated with 0.2 % formic acid and extracted with ethyl acetate. this process helped to attain the recovery up to 62 %. the mean ± s.d recovery of filgotinib at lqc and hqc was found to be 62.4 ± 8.39 and 64.2 ± 4.26 %, respectively. the recovery of the is was 95.2 ± 4.85 %. matrix effect the mean absolute matrix effect for filgotinib on rat dbs cards at lqc and hqc was found to be 0.81 ± 0.08 and 0.83 ± 0.08 %, respectively. the matrix effect for the is was 0.97 ± 0.05 % (at 100 ng/ml). these results indicate that the minimal matrix effect did not obscure the quantification of on filgotinib and the is from rat dbs cards. selectivity and carry over figures 2a,b,c show chromatograms for the blank rat dbs cards (free of analyte and the is; figure 2a), blank rat dbs cards spiked with filgotinib at lloq and the is (figure 2b) and an in vivo blood sample obtained at 0.25 h after oral administration of filgotinib along with the is (figure 2c). there was no carryover produced by the highest calibration sample on the following injected blank dbs extract sample. additionally, no dbs-specific device-oriented carry-over was noted. http://dx.doi.org/10.5599/admet.796 r. mullangi et al. admet & dmpk 8(2) (2020) 139-148 144 a) b) c) figure 2. typical mrm chromatograms of filgotinib (left panel) and the is (right panel) in (a) rat blank dbs card (b) rat blank dbs card spiked with filgotinib at lloq (1.37 ng/ml) and the is (c) a 0.25 h in vivo plasma sample showing filgotinib peak obtained following oral administration to rat along with the is. d if fe r e n c e ( % ) 1 10 100 1000 10000 -20 -10 0 10 20 mean val ue (ng/ml )-l og scal e figure 3. bland-altman plot showing the incurred sample re-analysis data for filgotinib on dbs. calibration curve the calibration curve was constructed in the linear range using eight calibrators 1.34, 2.67, 14.7, 26.7, 400, 801, 1603 and 1937 ng/ml. the typical regression equation for calibration curve was y = 0.000849 x + 0.00127. the correlation coefficient (r) average regression (n=4) was found to be >0.9966 for filgotinib. the lowest concentration with the rsd <20 % was taken as lloq and was found to be 1.37 ng/ml. the accuracy observed for the mean of back-calculated concentrations for four calibration curves for filgotinib was within 95.9-110 %; while the precision (cv) values ranged from 3.87-7.94 %. admet & dmpk 8(2) (2020) 139-148 dbs method for filgotinib quantitation doi: http://dx.doi.org/10.5599/admet.796 145 accuracy and precision accuracy and precision data for intraand inter-day dbs samples for filgotinib is presented in table 1. the assay values on both the occasions (intraand inter-day) were found to be within the accepted variable limits. stability the predicted concentrations for filgotinib at 4.01 and 1736 ng/ml samples deviated within ±15 % of the fresh sample concentrations in a battery of stability tests: bench-top (6 h), in-injector (21 h) and freezer stability at -80  10 °c for at least for 30 days (table 2). the results were found to be within the assay variability limits during the entire process. table 1. precision and accuracy determination of filgotinib quality controls in rat plasma lloq qc (1.37 ng/ml) lqc (4.01 ng/ml) mqc (1068 ng/ml) hqc (1736 ng/ml) intra-day (n=6) mean ± s.d 1.34 ± 0.10 4.30 ± 0.45 1136 ± 73.3 1715 ± 124 precision (%rsd) 7.38 10.4 6.46 7.23 accuracy (%re) 1.00 1.07 1.06 0.99 inter-day (n=24) mean ± s.d 1.33 ± 0.15 3.94 ± 0.44 1064 ± 96.8 1663 ± 128 precision (%rsd) 11.0 11.2 9.09 7.71 accuracy (%re) 1.00 0.98 1.00 0.96 rsd: relative standard deviation (sd 100/mean) re: relative error (measured value/actual value) table 2. stability data of filgotinib quality controls in rat plasma rsd: relative standard deviation (sd 100/mean) re: relative error (measured value/actual value) incurred samples reanalysis (isr) as per guidance [17] around 10 % of the study samples should be selected for isr if the total sample size is less than 1000. in this validation a total of 20 samples were chosen for isr from the oral and intravenous rat pharmacokinetic studies. from oral arm samples near cmax (0.5 h) and elimination phase (4 h and 24 h) were selected. however, for intravenous arm representative samples at 0.083, 2 and 8 h time points were selected. figure 3 shows the comparison of isr values versus original values using blandaltman plot suggesting that all the isr samples were within ±20 % of the original values. concentration spiked (ng/ml) bench-top for 6 h (n=6) long-term 30 days at -80 °c (n=6) in-injector for 21 h (n=6) % re %rsd % re %rsd % re %rsd 4.01 0.96 11.1 0.97 15.1 0.91 5.99 1736 0.88 4.09 0.96 6.04 0.88 2.39 http://dx.doi.org/10.5599/admet.796 r. mullangi et al. admet & dmpk 8(2) (2020) 139-148 146 hematocrit effect hematocrit (hct) has a significant effect on the viscosity of the blood, which influences the flux and diffusion properties of the blood through dbs card used for sample collection. it can directly affect the accuracy of the analysis in dbs samples. the reported blood-to-plasma ratio for filgotinib was 1.22 [19] indicating an almost equal plasma and red blood cell distribution. the measured filgotinib concentrations were compared with the results obtained from dbs samples are given in table 3, indicating that hct had no significant influence on the filgotinib concentration. table 3. haematocrit effect analysis concentration spiked (ng/ml) hematocrit (%) concentration found (ng/ml) accuracy (%re) precision (% rsd) 4.01 30 3.95 ± 0.19 0.98 4.81 50 3.90 ± 0.24 0.97 6.29 1736 30 1664 ± 80.4 0.95 4.83 50 1786 ± 88.7 1.02 4.96 0 2 4 6 8 10 12 14 16 18 20 22 24 1 10 100 1000 10000 time (h) m e a n ± s .d (n g /m l ) c o n c e n tr a ti o n o f fi lg o ti n ib iv 2 mg/kg plasma iv 2 mg/kg dbs 0 2 4 6 8 10 12 14 16 18 20 22 24 1 10 100 1000 10000 time (h) m e a n ± s .d (n g /m l ) c o n c e n tr a ti o n o f fi lg o ti n ib po 10 mg/kg plasma po 10 mg/kg dbs figure 4. a) mean plasma concentration-time profiles of filgotinib in rat blood and plasma following intravenous administration of filgotinib; b) mean plasma concentration-time profiles of filgotinib in rat blood and plasma following oral administration of filgotinib. pharmacokinetic study the sensitivity of the present dbs method was found to be sufficient for accurately characterizing the pharmacokinetics of filgotinib by oral and intravenous routes in rats. to assure acceptance of study sample analytical runs, at least two-thirds of the qc samples had to be within ±15 % accuracy, with at least half of the qc samples at each concentration meeting these criteria. results indicated that qcs met the acceptance criteria. figures 4a and 4b show the mean blood concentration versus time for filgotinib post administration of intravenous and oral route, respectively in rats. the pharmacokinetic parameters are summarized in table 4 along with previously reported pharmacokinetic parameters by dixit et al. [10]. filgotinib was quantifiable up to 24 h post intravenous and oral administration to rats. following intravenous administration at 2.0 mg/kg, the plasma concentrations decreased mono-exponentially. filgotinib exhibited moderate clearance (cl) of 18.7 ± 1.86 ml/min/kg, which is ~3-fold lower than hepatic blood flow (55 ml/min/kg) and volume of distribution (vd) was 4.17 ± 0.33 l/kg in rats. the terminal halflife was found to be 3.52 ± 0.23 h. following oral administration filgotinib maximum plasma concentrations (cmax: 804 ± 78.4 ng/ml) attained at 0.50 ± 0.00 h (tmax) in all rats suggesting that filgotinib has a rapid a) b) admet & dmpk 8(2) (2020) 139-148 dbs method for filgotinib quantitation doi: http://dx.doi.org/10.5599/admet.796 147 absorption from the gastrointestinal tract. the auc0- (area under the plasma concentration-time curve from time zero to infinity) was found to be 3722 ± 329 and 1801 ± 188 ngh/ml, by oral and intravenous routes, respectively. the terminal half-life (t½) determined after oral administration was 4.87 ± 0.45 h. the absolute oral bioavailability for filgotinib in rats at 10 mg/kg was 42.5 %. table 4. pharmacokinetic parameters for filgotinib in rats following oral and intravenous administration pk parameters oral intravenous dose (mg/kg) 10 2.0 auc(0-) (ng  h/ml) 3722 ± 329 (3475 ± 547) 1801 ± 188 (1912 ± 123) cmax/c0 (ng/ml) 804 ± 78.4 (802 ± 149) 1602 ± 55.5 (1767 ± 128) tmax (h) 0.50 ± 0.00 (0.50 ± 0.00) -- t1/2 (h) 4.87 ± 0.45 (4.72 ± 0.24) 3.52 ± 0.23 (3.56 ± 0.30) cl (ml/min/kg) -- 18.7 ± 1.86 (17.4 ± 1.08) vd (l/kg) -- 4.17 ± 0.33 (5.41 ± 0.77) f (%) 42.5 (36.3) -- the values in parentheses are reported by dixit et al. (2020) [10]. in figures 4a and 4b the blood versus time concentration profiles of filgotinib obtained from intravenous and oral routes in rats were presented along with plasma versus time concentration profiles reported earlier by us [10] showed excellent correlation (r 2 > 0.99). this indicates that dbs can be used as an alternative strategy for determination of filgotinib circulatory concentration in a pharmacokinetic study. to the best of our knowledge, to date there is no comparison between plasma and whole blood for filgotinib. our study demonstrated that dbs is a promising alternative to plasma sampling for filgotinib in rheumatoid arthritis patients especially with elderly patients in remote or resource-limited settings or immobile patients. conclusions in summary, a simple and rapid method using lc-ms/ms for the determination of filgotinib in rat blood using dbs cards was developed and validated as per us fda regulatory guideline. the validated method suitability was shown in a rat pharmacokinetic study. conflict of interest: the authors are scientists at jubilant biosys ltd. references [1] j.k. buer. a history of the term dmard. inflammopharmacology 23 (2015) 163-171. [2] c. chough, s. lee, m. joung, j. lee, j.h. kim, b.m. kim. design, synthesis and evaluation of (r)-3-(7(methyl(7h-pyrrolo[2,3-d]pyrimidin-4-yl)amino)-5-aza-spiro[2.4]heptan-5-yl)-3-oxopropanenitrile as a jak1-selective inhibitor. medchemcomm 9 (2018) 477. [3] j.j. o’shea, m. gadina. (2019). selective janus kinase inhibitors come of age. nature reviews rheumatology 15 (2019) 74-75. [4] k. yamaoka. janus kinase inhibitors for rheumatoid arthritis. current opinion in chemical biology 32 (2016) 29-33. http://dx.doi.org/10.5599/admet.796 r. mullangi et al. admet & dmpk 8(2) (2020) 139-148 148 [5] l. van rompaey, r. galien, e. m. van der aar, p. clement-lacroix, l. nelles, b. smets, l. lepescheux, t. christophe, k. conrath, n. vandeghinste, b. vayssiere, s. de vos, s. fletcher, r. brys, g. van't klooster, j. h. m. feyen and c.menet. preclinical characterization of glpg0634, a selective inhibitor of jak1, for the treatment of inflammatory diseases. journal of immunology 191 (2013) 3568-3577. [6] m. genovese, r. westhovens, l. meuleners, a. van der aa, p. harrison, c. tasset, a. kavanaugh. effect of filgotinib, a selective jak 1 inhibitor, with and without methotrexate in patients with rheumatoid arthritis: patient-reported outcomes. arthritis research & therapy 20 (2018) 57. [7] pharma japan (2019). https://pj.jiho.jp/article/240785. [8] https://www.gilead.com/news-and-press/press-room/press-releases (accessed may 01, 2020.) [9] f. namour, p.m. diderichsen, e. cox, b. vayssie`re, a.v. aa, c. tasset, v.g. klooster. pharmacokinetics and pharmacokinetic/pharmacodynamic modeling of filgotinib (glpg0634), a selective jak1 inhibitor, in support of phase iib dose selection. clinical pharmacokinetics 54 (2015) 859-874. [10] a. dixit, v. kiran, b.b. gabani, z. mohd, r.r. trivedi, r. mullangi. validated lc-ms/ms method for quantitation of a selective jak1 inhibitor, filgotinib in rat plasma and its application to a pharmacokinetics study in rats. biomedical chromatography (2020) accepted. [11] sulochana sp, daram p, srinivas nr, mullangi r. review of dbs methods as a quantitative tool for anticancer drugs. biomedical chromatography 33 (2019) e4445. [12] k.i. foerster, a. huppertz, a.d. meid, o.j. müller, t. rizos, l. tilemann, w.e. haefeli, j. burhenne. dried blood spot technique to monitor direct oral anticoagulants: clinical validation of an uplc/ms/ms based assay. analytical chemistry 90 (2018) 337-341. [13] h.m. kim, j. park, n.p. long, d.d. kim, s.w. kwon. simultaneous determination of cardiovascular drugs in dried blood spot by liquid chromatography-tandem mass spectrometry. journal of food and drug analysis 27 (2019) 906-914. [14] j. gallaya, s. prod’hom, t. mercierc, c. bardinet, d. spaggiari, e. pothina, t. buclinc, b. gentona, l.a. decosterd. lc-ms/ms method for the simultaneous analysis of seven antimalarials and two active metabolites in dried blood spots for applications in field trials: analytical and clinical validation. journal of pharmaceutical and biomedical analysis 154 (2018) 263-277. [15] m. moretti, f. freni, b. valentini, c. vignali, a. groppi, s.d visonà, a.m.m. osculati, l. morini. determination of antidepressants and antipsychotics in dried blood spots (dbss) collected from post-mortem samples and evaluation of the stability over a three-month period. molecules 24 (2019) pii: e3636. [16] s. velghe, s. deprez, c.p. stove. fully automated therapeutic drug monitoring of anti-epileptic drugs making use of dried blood spots. journal of chromatography a 1601 (2019) 95-103. [17] dhhs, fda, cder, & cvm (2018). guidance for industry: bioanalytical method validation. rockville, md: us department of health and human services, food and drug administration, center for drug evaluation and research and center for veterinary medicine. [18] m. o’mara, b. hudson-curtis, k. olson, y. yueh, j. dunn, n. spooner. the effect of hematocrit and punch location on assay bias during quantitative bioanalysis of dried blood spot samples. bioanalysis 20 (2011) 2335-2347. [19] m.e. dowty, t.h. lin, m.i. jesson, m. hegen, d.a. martin, v. katkade, s. menon, j.b. telliez. janus kinase inhibitors for the treatment of rheumatoid arthritis demonstrate similar profiles of in vitro cytokine receptor inhibition. pharmacology research & perspectives (2019) e00537. ©2020 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) https://pj.jiho.jp/article/240785 https://www.gilead.com/news-and-press/press-room/press-releases http://creativecommons.org/licenses/by/3.0/ physiologically based pharmacokinetic (pbpk) modeling and simulation in drug discovery and development doi: 10.5599/admet.667 1 admet & dmpk 7(1) (2019) 1-3; doi: 10.5599/admet.667 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index editorial physiologically based pharmacokinetic (pbpk) modeling and simulation in drug discovery and development abdul naveed shaik 1* , ansar ali khan 2 guest editors; admet & dmpk 1 simulations plus inc., lancaster, ca. usa 2 gvk biosciences. hyderabad, ts. india *corresponding author: e-mail: naveedshaik@gmail.com; received: february 19, 2019; published: february 23, 2019 physiologically based pharmacokinetic (pbpk) modeling is a mechanistic or physiology based mathematical modeling technique which integrates the knowledge from both drug-based properties including physiochemical and biopharmaceutical properties and system based or physiological properties to generate a model for predicting the absorption, distribution, metabolism and excretion (adme) of a drug as well as pharmacokinetic behavior of a drug in preclinical species and humans. even though pbpk models can provide mechanistic insights, it was seldom used for long time due to the lack of highperformance computing systems. with the advent of high-performance computing facilities, multiple differential equations can be solved at a time. development of commercial pbpk software such as pk-sim (bayer technology services), gastroplus (simulations plus inc.), simcyp (certara), simbiology (mathworks) etc. has resulted in increased use of pbpk in drug development. pbpk models can be used in different phases of drug development including the first in human (fih) dose prediction from preclinical data. in this phase data generated from in vitro experiments as well as data from in silico predictions can be integrated to develop model which can then be used to predict fih dose range. this is referred as the bottoms up approach. on the other hand, in the top down approach, clinical data can be used to develop and refine the pbpk model. moreover, pbpk model can be used to evaluate drug safety and toxicity, as well as the prediction of drug exposure in different conditions or populations studying the effect of age, gender, ethnicity disease state etc. these models are also commonly used in the evaluation of drug-drug interaction (ddi) to provide insights into any undesirable issues of the investigative new drug (ind) with other marketed drugs. recent advancement in pbpk modeling and the understanding of metabolizing enzymes, drug transporters and genetic make-up of different races have led to an increase in the use of pbpk models in drug discovery and development. these models can provide a simple in vitro in vivo correlation (ivivc) based on the thorough in vitro data from drug metabolism [1-3] and drug transporter studies [2, 4]. pbpk is very useful in modelling of: a) absorption where the model takes into consideration the solubility, dissolution, and permeability both passive and active with the involvement of transporters, gut extraction http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:naveedshaik@gmail.com shaik and kahn admet & dmpk 7(1) (2019) 1-3 2 through metabolism; b) distribution where physiological parameters and compound related parameters are taken into consideration and various equations can be applied to calculate the tissue partition coefficient (kp) for perfusion rate limited or permeability limited situations; c) metabolism where data from in vitro sources including recombinant enzymes, liver microsomes and hepatocytes and transporter data from hepatocytes or over expressed cell lines can be used to inform the rate and extend of in vivo metabolism; d) excretion through biliary, renal routes can be easily modelled using pbpk models. there is an increase in the acceptability of pbpk models in drug submission package by usfda, ema and pmda as evident by the fact that the number of ind applications as well as submissions with ddis and in special populations including paediatrics that involved the use of pbpk have increased significantly in the recent years by regulatory agencies. in particular, these regulatory agencies have issued guidance for application and use of pbpk models at various stages of drug development including, investigational new drug applications (inds), new drug applications (ndas), biologics license applications (blas), or abbreviated new drug applications (andas) etc. with the increased use of pharmacodynamics (pd) with pbpk models (pbpk-pd) where simple pd models can be linked to pbpk model, it is anticipated that pbpk model-based drug development is going to be very popular in pharmaceutical research. in the foreseeable future, we anticipate that more pharmaceutical companies and academicians are going to adopt pbpk modeling in their research and drug development projects. with the increase in the number of research publication utilizing pbpk from industry, academia and regulatory authorities alike and with the success of prior special issues focused on various stages of drug development we devoted a special issue on “physiologically based pharmacokinetic (pbpk) modeling and simulation in drug discovery and development”. we received four original articles and a review article for this special issue. these articles covered varied areas including “building “in-house” pbpk modeling tools for oral drug administration from literature information” [5] where the researchers have shown how to develop in-house pbpk tools from species-related physiological information available in the literature and a limited number of drug specific parameters and validated the in-house pbpk tools with 25 compounds. another research article entitled “a physiologically-based pharmacokinetic model of oseltamivir phosphate and its carboxylate metabolite for rats and humans” [6] where the model was built with both the parent oseltamivir phosphate and its carboxylate metabolite. the model built in healthy subjects has been applied in disease-subjects as well as in subjects with liver or renal impairment. one other research article was titled "physiologically-based pharmacokinetic simulations in pharmacotherapy: selection of the optimal administration route for exogenous melatonin" [7] where the authors showed the significance of pbpk model in dose optimization and selection of route of administration. one article of a great interest was a review entitled “pbpk modelling of highly lipid soluble and extracellular solutes” [8], which extensively discussed two classes of drugs, namely the highly lipid soluble (hls) solutes, and the extracellular (ecs) solutes. these submissions clearly indicate the utilization and significance of pbpk in various stages of drug development. references [1] a.n. shaik. changing trends in use of hepatocytes and microsomes for evaluating metabolism studies in drug discovery. admet and dmpk 4 (2016) 60-61. [2] s. basu, a.n. shaik. is there a paradigm shift in use of microsomes and hepatocytes in drug discovery and development? admet and dmpk 4 (2016) 114-116. admet & dmpk 7(1) (2019) 000-000 running title doi: 10.5599/admet.667 3 [3] a.n. shaik, t. bohnert, d.a. williams, l.l. gan, b.w. leduc. mechanism of drug-drug interactions between warfarin and statins. journal of pharmaceutical sciences 105 (2016) 1976-1986. [4] s. basu, a.n. shaik. role of drug transporters in drug development: a qualitative and quantitative approach. admet and dmpk 5 (2017) 57-58. [5] s. grandoni, n. cesari, g. brogin, p. puccini, p. magni. building in-house pbpk modelling tools for oral drug administration from literature information. admet and dmpk 7 (2019) 4-21. [6] g. gao, f.c.p. law, r.n.s. wong, n.k. mak, m.s. yang. a physiologically based pharmacokinetic model of oseltamivir phosphate and its carboxylate metabolite for rats and humans. admet and dmpk 7 (2019) 22-43. [7] a. savoca, d. manca. physiologically-based pharmacokinetic simulations in pharmacotherapy: selection of the optimal administration route for exogenous melatonin. admet and dmpk 7 (2019) 44-59. [8] d.g. levitt. pkquest: pbpk modelling of highly lipid soluble and extracellular solutes. admet and dmpk 7 (2019) 60-75. ©2019 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ microsoft word 9.doc doi: 10.5599/admet.1.3.9  19  admet & dmpk 1(3) (2013) 19‐28; doi: 10.5599/admet.1.3.9    open access : issn : 1848‐7718      perspective  what adme tests should be conducted for preclinical studies?  hong wan  shanghai hengrui pharmaceutical co., ltd.    dept. of dmpk/tox, shanghai, p.r. china  e‐mail: wanh@shhrp.com  tel: +86(0)21‐54759066‐1313; fax: +86(0)21‐54759072; mobile: +86‐18036618308  received: june 06th, 2013; revised: july 01st, 2013; published: july 03rd, 2013      abstract  the pharmaceutical industry has been evolving in recent years, which made numerous cros and biotech companies  conducting drug discovery and development programs and services in china. as a very comprehensive and important  technology  platform  bridging  efficacy  and  safety,  the  dmpk  has  become  a  mature  discovery  function  to  optimise  adme properties in drug design and screening, and dramatically mitgate attrition rates during the last decades. in this  article, the author addresses several frequent questions associated with adme/dmpk studies, e.g., what adme tests  should be conducted for preclinical studies? what should a typical investigational new drug (ind) enabling package  cover? which adme/dmpk studies require good laboratory practice (glp) or non‐glp? what does a good pk profile  look like?    the author presents a straightforward overview of these basic questions from his many years’ experience  in both pharmaceutical research and cro in supporting drug discovery projects and ind filing.  keywords  dmpk, pk, toxicity, safety, screening cascade, metabolite, drug‐drug interaction, investigational new drug    1. dmpk/adme: definition and study purpose  dmpk, or drug metabolism and pharmacokinetics, is an important part of studies often referred to as  adme (absorption, distribution, metabolism, and elimination).    absorption (how much and how fast, often referred to as the absorbed fraction or bioavailability)  distribution (where the drug is distributed, how fast and how extensive)  metabolism (how fast, what mechanism/route, what metabolite is formed, and whether they are  active or toxic)  elimination (how fast, which route).  in the drug discovery process, early  in vitro adme screening and  in vivo pk profiling provide a basis for  choosing new molecular entities (nmes) and lead compounds that have desirable drug metabolism, pk or  safety  profiles,  necessary  for  drug  candidate  selection  (cs)  and  late‐stage  preclinical  and  clinical  development.  the  adme  properties  of  a  drug  allow  the  drug  developer  to  understand  the  safety  and  efficacy data required for regulatory approval.    wan    admet & dmpk 1(3) (2013) 19‐28  20    today, dmpk studies performed in vitro and in vivo have become more standardised procedures across the  pharmaceutical  industry.  a  couple  of  typical  examples  include  the  most  commonly  used  in  vitro  adme  studies, such as liver microsomes and the whole hepatocyte models for in vitro metabolism. both models  contain  major  metabolism  enzymes,  such  as  cyp450  for  phase  i  metabolism  and  udp‐glucuronosyltransferase (ugt) for phase ii metabolism. these relevant in vitro models are also applied  to cyp  inhibition/induction studies as well as drug metabolite  identification and profiling across species.  other  key  in  vitro  assays,  such  as  caco‐2  or  mdck  cell‐based  models  are  often  used  for  intestinal  permeability evaluation. for in vivo pk studies, association for assessment and accreditation of laboratory  animal care (aaalac)‐accredited animals, such as mice, rats, dogs, and non‐human primates are employed  to generate in vivo pk data like drug clearance (cl), bioavailability (f%), exposure (auc), half life (t1/2), and  distribution volume (l). currently, almost all dmpk‐related assays are carried out by available automated  technology  platforms  combined  with  high‐throughput  liquid  chromatography‐mass  spectrometry  (lc/ms/ms) bioanalysis, which has considerably speeded up the adme data generation for decision‐making  during the drug discovery and development process.    2. what adme tests should be conducted for preclinical studies?      potency <100 nm enzyme/cell based assays solubility 7.4 > 1-10 µm (pbs buffer) metabolic stability clint (h, r, d, m) caco-2 permeability > 0.3x 1e-6cm/s, efflux logd7.4 (1.5-4.5) in vivo pk (m, r, d, m) f>20%, cl<50% bf, dose<100mg ppb (r, h, d, m) fu> 0.1-1% cyp inhibition (ic50) major cyp>1-10 µm reactive metabolite (gsh) met id, cyp id metabolism route pk/pd dose response solubility 6.5 >10 µm (fassif) hepatocytes (r, h) blood/plasma ratio (r, h, d, m) non-glp tox (rodent, dog) glp tox (safety package) cs (candidate selection) cyp induction h it-to-lead lead-optim isation c andidate transporter, p-gp etc substrate/inhibition mdck pamap herg ic50 ind enabling (candidate) clinical candidate (glp) mtd (rodent) human pk dose, pk/pd f clce(ss) dose τ•• =   figure 1: typical adme/pk screening cascade    figure 1 highlights a typical adme/dmpk screening cascade to support drug discovery programs from early  hit‐to‐lead, to  lead optimisation and candidate evaluations. as depicted  in figure 1, a number of dmpk  studies  are  required  before  cs,  followed  by  non‐glp  and  glp  toxicology  studies  for  the  ind‐enabling  package  that  is  necessary  for  clinical  candidate  development.  various  studies  and  assays  should  be  conducted at different stages to address adme issues and meet specific criteria to enable the project to  admet & dmpk 1(3) (2013) 19‐28  adme tests for preclinical studies  doi: 10.5599/admet.1.3.9  21  move forward. it should be noted that these criteria, as exemplified in this cascade, are not fixed numbers  or  filters,  and  depend  on  other  parameters  as  well  as  the  specific  therapeutic  targets  of  different  drug  discovery  programs.  this  cascade  just  demonstrates  a  traditional  downstream  of  adme/pk  tests  for  a  typical drug discovery programs before the ind‐enabling stage. often, at the early stage of hit‐to‐lead, key  in vitro assays and studies, such as solubility,  liver microsomal stability, cyp  inhibition, and permeability  should  be  conducted  to  provide  key  data  for  chemists  to  select  the  more  metabolically  stable  and  active/potent compounds for further in vivo pk studies. solubility and log d data can serve as a basis to run  various in vitro assays as well as for drug stability assessment and selection of suitable formulations. also, it  is  well  known  that  log  d/log  p  a  crucial  factor  governing  passive  membrane  partitioning,  influencing  permeability opposite to its effect on solubility. moreover, if the in vitro metabolic stability data from phase  i  metabolism  cannot  predict  in  vivo  hepatic  plasma  clearance,  the  phase  ii  in  vitro  metabolism  in  hepatocytes and even hepatic transporters such as oatp1b1 are then necessary to find the relevant in vitro  models for in vivo metabolism prediction to drive chemists to design and make more potent and selective  compounds  with  improved  pk  profiles.  even  at  the  very  early  stage  of  hit‐to‐lead,  selective  in  vivo  pk  studies are valuable to confirm whether the applied in vitro assays (in vitro metabolism and absorption) can  serve as good predictive models for  in vivo pk  in terms of plasma clearance and bioavailability. in some  circumstances,  if  both  in  vitro  liver  microsomes  and  hepatocyte  models  might  not  predict  well  in  vivo  metabolism, more extensive in vivo pk studies have to be performed as a result of poor in vitro and in vivo  correlations based on evaluations by a simplified well‐stirred model. a clear structure‐activity relationship  (sar)/adme certainly facilitates the design of new compounds. for the drug absorption evaluations, there  are three commonly used models, such as pampa, and caco‐2 and mdck permeability. from the cost point  of view, the pampa membrane permeability assay  is the most cost‐effective among these three  in vitro  models,  with  a  fast  turnaround,  while  the  caco‐2  and  mdck‐mdr1  models  can  offer  more  accurate  absorption  prediction  as  well  as  allow  efflux  and  transporter  studies  for  substrate  and  inhibition  identification.      traditionally, the cyp inhibition study is conducted at a late stage, e.g., from lead optimisation to cs. this  has  been  now  moved  to  earlier  stages,  even  hit‐to‐lead.  table  1  shows  a  general  decision  tree  on  cyp  inhibition based on drug‐drug interaction (ddi) assessment, with an interpretation of potential for in vivo  inhibition (derived from fda guidance for industry: drug interaction studies – study design, data analysis,  implications for dosing, and labeling recommendations) [1]. an estimated [i]/ki ratio of greater than 0.1 is  considered  positive  and  a  follow‐up  in  vivo  evaluation  is  recommended.  the  likelihood  of  an  in  vivo  interaction is projected based on the [i]/ki ratio where [i] represents the mean steady‐state cmax value for  total drug (bound plus unbound) following administration of the highest proposed clinical dose. as the ratio  increases, the likelihood of an interaction increases. also, for cyp3a inhibition, two structurally unrelated  substrates should be evaluated in vitro. if one of the two evaluations suggests a potential interaction (i.e.,  [i]/ki of more than 0.1), an in vivo evaluation should be carried out. at the earlier screening stage, ic50 by a  single‐point screening assay, which is a more cost‐effective approach with acceptably good correlation with  full  curve  ic50 assay, can  be  used for  initial ddi evaluation.  a full  curve  ic50 or ki determination can be  utilised for ddi assessment at a late phase for candidates such as nmes. although quantitative predictions  of in vivo ddi from in vitro studies are not possible, rank order across the different cyp enzymes for the  same drug may help prioritise in vivo ddi assessment.    wan    admet & dmpk 1(3) (2013) 19‐28  22      table 1: prediction of clinical relevance of competitive cyp inhibition  likelihood  ic50 (μm) [i]/ki [1].  likely  < 1 > 1 possible  1 < ic50 <10 0.1 <[i] /ki <1  unlikely  > 10 < 0.1 [i] is the concentration of inhibitor exposed to the active site of the enzyme in vivo,    ki is the inhibition  constant, which can be obtained by a full curve ic50 assay, or roughly estimated from the ic50, i.e.,  ki≈ic50/2    similar studies associated with transporters are needed to identify whether the nmes could be a potential  substrate or inhibitor for p‐gp or other important transporters for ddi as early as possible. more complex  decision‐making trees and recommended models for transporter‐mediated ddi assessment are outlined in  the fda’s new guidance on drug interaction studies. it should be pointed out that the fda has extended  ddi guidance from the 2006 draft, with 52 pages, to the current 75 pages (issued in 2012), reflecting the  importance and complexity of drug interaction studies among the dmpk [1,2]. this is attributed to the high  percentage of drug failures caused by ddi.    another important aspect includes the drug metabolite identification and profiling as well as metabolism  pathways in connection with cyp metabolism enzyme identification (the so‐called cyp id). nowadays, the  rapid metabolite identification (metid) at a very early stage has also become relatively easy, such as with  metabolynx‐assisted  structure  identification  software,  which  is  especially  necessary  for  metabolically  unstable compounds, whose biotransformation can help chemists understand the site of metabolic liability,  for further structure modification for a desirable metabolism profile. with the issuance of the fda’s final  guidance for industry on safety testing of drug metabolites (mist), there is increased concern for obtaining  metabolite  data  as  early  as  possible  in  preclinical  studies.  the  discovery  of  disproportionate  drug  metabolites late in drug development can potentially cause development and market delays. as guided in  mist published in 2008 as well as in the ich m3 (r2) published in 2009 [3,4], if a metabolite is a unique  human metabolite, or more commonly, if a circulating metabolite is present at disproportionately higher  levels  in  humans  than  in  the  animal  species  used  in  toxicology  studies,  additional  non‐clinical  safety  assessment studies may be required. as exemplified in figure 2, any metabolites found to be >10% (based  on systemic exposure at steady state) of total drug‐related material should then be monitored in repeating  dose animal toxicology studies to determine if they are present at higher  levels in other species as well.  exceptions might be possible for lower‐risk metabolites, such as most phase ii conjugated metabolites. if  the major metabolite exposure  in at  least one species  is greater than  in the human, then no additional  safety studies of the metabolite are needed. it should be noted that the metabolite decision tree shown in  figure  2  is  only  a  suggestion  or  recommendation,  with  an  approximate  10%  cut‐off,  which  does  not  necessarily  mean  a  metabolite  testing  safety  guarantee.  for  instance,  if  a  drug  shows  a  low  predicted  efficacy dose (10 mg or lower), a metabolite found to be over 10% of total drug‐related material might not  require further safety tests. in contrast, for a highly dosed compound, even if a metabolite is lower than  10% of total drug‐related material, it could be toxic, especially with accumulation of the metabolite. also, a  reactive metabolite might be toxic at a very low concentration. overall, a metabolite safety testing decision  should consider many factors, such as in vitro‐in vivo correlations and especially in vivo animal toxicology  findings, e.g., abnormalities or even animal deaths with repeat doses on animal studies. in this case, it is  essential to identify the biotransformed metabolites to understand whether the observed toxicity is due to  the accumulation of parent exposure or toxic metabolites, as even metabolites at a level lower than 10%  admet & dmpk 1(3) (2013) 19‐28  adme tests for preclinical studies  doi: 10.5599/admet.1.3.9  23  can  accumulate.  such  additional  toxicity  testing  of  the  major  metabolite  can  be  informative  to  confirm  whether the cause of toxicity is the metabolite or not. in this case, for rodent or acute rat toxicology, the  identification and profiling of drug metabolites using toxicokinetic plasma samples can be used. in short,  relevant studies must be carried out to determine if human metabolites are adequately evaluated during  non‐clinical safety studies.                                                          figure 2: fda suggested metabolite decision tree flow diagram    in addition to the cyp inhibition/induction and transporter studies associated with ddi, other safety‐related  studies,  such  as  reactive  metabolism  and  herg  should  be  performed  as  early  as  possible.  the  herg  inhibition  test  for  cardiotoxicity  has  been  shown  to  predict  increased  risk  of  arrhythmia,  which  can  be  initially evaluated under non‐glp condition by in vitro approaches that are cost‐effective and have a fast  turnaround, and later confirmed using qt interval prolongation measurements according to fda guidelines  [5]. another important in vivo study is basic safety tests that are most commonly performed on rats and  dogs, e.g., non‐glp acute toxicology (7‐day or 14‐day rat and dog repeat dose) for oral drugs, which usually  include body weight, clinical chemistries, haematology, and histopathology (high dose and control only if no  disproportionate drug metabolite >10% parent  systemic exposure  �auc� ≤10% parent  systemic exposure  �auc� formed in any animal test  no  further  testing  needed  yes how much?  no exposure in animal studies  does  approach  human exposure  exposure in animal studies  does  not  approach  human exposure  no further testing  needed  nonclinical testing of the drug metabolite  wan    admet & dmpk 1(3) (2013) 19‐28  24    pathology  is  seen  at  the  high  dose).  this  study  should  be  designed  to  establish  a  dose  that  induces  a  minimal toxic effect, or alternatively, to establish a safety margin (e.g., 10–100x depending on the different  therapeutic  targets).  the  selection  of  dose  levels  for  repeat  oral  dose  in  toxicology  studies  is  generally  based  on  the  minimum  efficacy  dose  (med)  and  the  maximum  tolerated  dose  (mtd).  based  on  the  measured  auc  (exposures)  and  toxicokinetic  data,  an  approximate  safety  margin  and  compound  accumulation characteristics can be known, which are paramount for the decision‐making on whether to  follow up a glp‐tox design or not. such non‐glp studies help to identify more accurate dose level design for  the subsequent glp studies.    in  summary,  a  comprehensive  nonclinical  adme/pk  package,  including  key  toxicity  data,  should  be  generally complete by the time of cs. the ultimate goals of all adme/pk/tox studies before the cs stage are  to better understand the compound’s metabolite‐mediated toxicity and safety profile to make a concrete  decision for the purpose of enabling ind, which is a crucial part of the drug approval process.  3. ind‐enabling package for glp or non‐glp?  what preclinical studies should be conducted to enable an ind? should all studies be performed  in glp  laboratories for ind filing? some researchers are mis‐communicating that at ind stage, all studies should be  conducted in glp laboratories. actually, the fda requires glp documentation only for safety‐related in vivo  studies like glp‐tox. in other words, in vitro adme and in vivo pk studies do not need to be performed in  glp labs. in fact, from a bioanalytical data quality point of view, there is no apparent discrepancy between  non‐glp  and  glp  laboratories  regarding  bioanalytical  procedures,  because  most  non‐glp  bioanalysis  methods are performed according to fda bioanalysis guidance [6]. the major difference can be in the study  design, e.g., a typical non‐glp tox study utilises a small group of animals (three or four per group), while a  glp‐tox  study  usually  utilises  more  than  10  animals  per  group  with  half  males  and  half  females  in  one  study. in addition, more strict documentation must be recorded for the purpose of regulatory compliance.  obviously, the glp study has much higher costs than a similar non‐glp study, which is not recommended  before cs for the ind‐enabling package. key components of the ind data package include pharmacology,  toxicology and safety pharmacology, adme and chemistry, and the manufacturing and control sections of  the  submission.  a  typical  ind‐enabling  adme  package  contains  data  from  the  following  studies:  bioanalytical  method  validation  in  one  rodent  and  one  or  more  non‐rodent  species;  single‐  and  multiple‐dose  pk;  dose  proportionality  and  absolute  bioavailability  in  one  rodent  and  one  or  more  non‐rodent species; and  in vitro cyp  inhibition/induction  in human  liver microsomes for ddi assessment  (including likely transporter studies). other data, such as mass balance and routes of excretion in rodents,  including bilary excretion in rats, metabolite profiling and identification in rodents and non‐rodents, tissue  distribution using radio‐labelled compound by quantitative whole‐body autoradiography (qwba) form part  of  the  surrogate  end‐points,  although  these  studies  are  not  required  for  submission.  overall,  complete  adme and toxicology data can make the submission package more compelling to move a compound into  phase i clinical studies. knowing the objectives, expectations, and processes of assembling and filing an ind  is the key factor not only for a successful filing, but it can also accelerate a promising clinical development  path forward. more detailed  information on the general requirements and strategies for successful  ind  filings can be found in fda’s guidance [7], which is beyond the scope of the current main topic. in order to  reduce the potential development cost, exploratory ind (eind) preclinical studies are recommended by the  fda [8]. such eind studies are conducted prior to the traditional dose escalation, safety, and tolerance  studies that ordinarily initiate a clinical development program, which is less extensive than for traditional  ind  studies.  overall,  eind  studies  provide  the  opportunity  to  already  obtain  human  data  in  the  drug  discovery phase, such as important information on pk, pk/pd, and basic toxicology data for a key decision  admet & dmpk 1(3) (2013) 19‐28  adme tests for preclinical studies  doi: 10.5599/admet.1.3.9  25  for selecting  the most promising  lead  product. even a microdose approach (e.g., <100 μg, or <1/100 of  med)  in  a  human  study,  using  an  advanced  technology  such  as  accelerator  mass  spectrometry  (ams)  combined with radio‐labelled compound can be very beneficial to gain human pk information as early as  possible for ind enabling [9].  4. what is a good pk profile?  another frequent question asked by drug discovery scientists is ‘is this compound’s pk good?’ ideally, the  desirable pk profiles of a preclinical candidate that is being considered for taking into development should:  • have acceptable solubility for development;  • be completely absorbed, preferably via passive absorption;  • have high bioavailability (e.g., f>50%) for oral drug;  • have  a  low  plasma  clearance  cl  (e.g,  <30%  blood  flow),  long  half‐life  (t1/2)  (e.g.,  >6  hrs),  and  acceptable distribution volume;  • have a linear kinetics, i.e., exposure proportional to dose and a clear pk/pd correlation;  • be eliminated by several pathways, i.e., renal excretion and hepatic metabolism, also metabolised  preferably by more than one enzyme for de‐risking ddi;  • have a simple metabolite profile, with no reactive metabolite;  • have no obvious cyp and major transporters like p‐gp inhibition or induction or low ddi potential;  and  • have a sufficient or at least acceptable safety margin (safety margin >10x, depending on different  therapeutic targets).    in  reality,  we  can  readily  rank  which  compound’s  pk  is  better  than  others  from  the  same  series  of  compounds during the screening. however, we should not rule out a compound by only looking at its pk  data, especially for those with super potent nmes. as long as their potencies and mechanism are attractive,  they should be tested in in vivo efficacy models despite poor pk profiles. indeed, a number of marketed  drugs have shown poor pk profiles but have still become the top ever selling drug – lipitor (atorvastatin),  with absolute bioavailability of 5% and 14% in rat and human, respectively [10,11]. as illustrated in figure 3,  in an analysis of 600 market drugs [12], more than 30% of them have relatively low bioavailability (f <30%),  and  22%  of  the  drugs  have  even  lower  bioavailability  (f  <10  %).  nearly  50%  of  the  drugs  have  moderate‐to‐high clearance, and 17% show high clearance, or extremely high clearance (even over blood  flow).  as  mentioned  above,  some  of  these  top  selling  drugs  have  many  undesirable  properties  in  terms  of  physicochemical  properties,  such  as  solubility,  log  d,  in  vitro  microsomal  stability  (clint),  and  permeability/efflux data, as well as high in vivo clearance or low bioavailability. if we use only pk data to  screen  the  compounds,  some  potential  drug‐like  compounds  might  be  discarded.  in  other  words,  the  atorvastatin‐like compounds would never have become good drugs from the sole view of pk profiles. these  examples  suggest  that  a  better  understanding  of  in  vitro‐in  vivo  correlation  and  frontloading  pk/pd  is  crucial for compound selection even at an early stage. obviously, a super pd can save a poor pk as long as  the pd effect is notable.    wan    admet & dmpk 1(3) (2013) 19‐28  26        figure 3: distribution of human clearance and bioavailability of market drugs      5. concluding remarks  drug discovery and development are a high‐risk, high‐cost but highly rewarding business. many resources  are invested, and thus wasted on, candidate products that are subsequently found to have unacceptable  profiles when evaluated in human – only less than 10% of inds from nmes progress beyond the submission  of new drug application (nda) according to the fda’s report [8]. as a very comprehensive and important  science and technology platform bridging efficacy and safety, dmpk has become a discovery function to  optimise adme properties in drug design, to reduce attrition rates. recent studies have demonstrated that  it  is  viable  to  implement  high‐throughput  cutting‐edge  technologies  and  cost‐effective  adme/dmpk  screening  assays,  and  design  relevant  studies  to  address  various  adme  issues  at  different  stages  for  de‐risking the investment and avoiding development mistakes early.    apparently, at drug discovery stage, a better understanding of sar and  in vitro‐in vivo correlations, and  more input on drug design can be anticipated. moreover, frontloading pk/pd for a better understanding of  the  efficacy‐exposure  relationship  and  dose  prediction  as  early  as  possible  can  benefit  projects.  the  integration of in vitro and in vivo drug metabolism data with physicochemical properties as well as pk and  pd  data  is  highly  recommended  for  safety  and  toxicity  liability  evaluation,  e.g.,  are  there  any  drug  accumulations? are there any potential toxic metabolites? are there any likely ddis? are estimated safety  margins enough? these safety‐related questions have to be answered as early and clearly as possible to  de‐risk the drug discovery projects. another concern is risk assessment associated with herg, cyp inhibition  and  induction,  metabolite  profiling  and  safety,  transporter  substrate  and  inhibition,  safety  margin,  etc.,  which  is of ultimate  importance for cs and decision‐making for the ind‐enabling package. future adme  studies will continue to focus on p450 enzymes and transporters’ impact on safe drug delivery as well as  accurate prediction of human pk, pk/pd modeling and optimum clinical dosage.      on the other hand, from the point of view of regulatory review, the fda has been becoming increasingly  conservative than ever, except for oncology and life‐threatening diseases where a risk‐benefit assessment  can be considered. if a drug is toxic at clinically relevant doses, the ind may never go forward. therefore, a  complete study/data package including as much detailed toxicity information as possible can facilitate ind  filing  and  increase  the  clinical  trial  success  rate,  which  should  cover  pharmacology  (efficacy  and  animal  admet & dmpk 1(3) (2013) 19‐28  adme tests for preclinical studies  doi: 10.5599/admet.1.3.9  27  proof of principle) studies, a rodent general toxicology study of at least a 14‐day duration, a non‐rodent  general toxicology study of at least a 14‐day duration, genotoxicity studies, safety pharmacology studies,  adme/pk  as  well  as  bioanalytical  validations,  etc.  in  the  end,  we  must  clearly  answer  these  two  key  questions:    (1) is the preclinical candidate safe enough to take into humans?    (2) is the preclinical candidate effective in patients?    acknowledgements: the author would like to thank dr. kin tam for proofreading this article with valuable  comments.  declaration:  this  article  reflects  the  author’s  personal  opinions.  the  author  declares  that  he  has  no  competing financial interests.    references  [1] fda: draft guidance for industry: drug interaction studies – study design, data analysis,  implications for dosing, and labeling recommendation, (sept. 2006)  [2] fda: guidance for industry: drug interaction studies – study design, data analysis, implications for  dosing, and labeling recommendation (february 2012)  [3] fda: guidance for industry: safety testing of drug metabolites (february 2008)  [4] ich m3(r2): guidance on nonclinical safety studies for the conduct of human clinical trials and  marketing authorization for pharmaceuticals (june 2009)  [5] fda: guidance for industry s7b nonclinical evaluation of the potential for delayed ventricular  repolarization (qt interval prolongation) by human pharmaceuticals (2005)  [6] fda: guidance for industry bioanalytical method validation (2001)  [7] fda: guidance for industry content and format of investigational new drug applications (inds) for  phase 1 studies of drugs, including well‐characterized, therapeutic, biotechnology‐derived  products (1997)  [8] fda: guidance for industry, investigators, and reviewers ‐ exploratory ind studies (jan 2006)  [9] g. lappin, m. seymour, bioanalysis 2 (2010) 1315‐1324.  [10] y.y. lau, h. okochi, y. huang and l.z. benet, drug metabolism and disposition 34 (2006) 1175‐1181.  [11] h. lennernas, clinical pharmacokinetics 42 (2003) 1141‐1160.  [12] l.s. goodman, a. gilman, the pharmacological basis of therapeutics, mcgraw‐hill publishers, new  york, usa, 2006.    abbreviations  adme      absorption distribution metabolism excretion  cro       contact research organization  dmpk      drug metabolism and pharmacokinetics  pk        pharmacokinetics  wan    admet & dmpk 1(3) (2013) 19‐28  28    cyp450      cytochrome p450  mdck      madin‐darby canine kidney cells  ppb       plasma protein binding  herg      human ether‐a‐go‐go related gene  metid      metabolite identification  pd        pharmacodynamics  p‐gp       p‐glycoprotein  mtd      maximum tolerated dose  ind        investigational new drug  glp        good lab practice  cl        clearance  f        absolute bioavailability  ce(ss)      effective plasma concentration at steady stage  τ        dose interval  cs        candidate selection  pmapa      parallel artificial membrane permeability assay  nmes      new molecular entities      ©2013 by the authors; licensee iapc, zagreb, croatia. this article is an open‐access article distributed under the terms and  conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/)        quantitative structure-activity relationship to elucidate human cyp2a6 inhibition by organosulfur compounds doi: 10.5599/admet.678 196 admet & dmpk 7(3) (2019) 196-209; doi: http://dx.doi.org/10.5599/admet.678 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper quantitative structure-activity relationship to elucidate human cyp2a6 inhibition by organosulfur compounds daniela a. ramirez 1 , eduardo j. marchevsky 2 , juan m. luco 2,* and alejandra b. camargo 1,* 1 instituto de biología agrícola de mendoza (ibam), consejo nacional de investigaciones científicas y técnicas (conicet). almirante brown 500, chacras de coria, mendoza, argentina 2 área de química analítica, facultad de química, bioquímica y farmacia, universidad nacional de san luis, chacabuco y pedernera 5700 san luis, argentina * corresponding authors: dr. alejandra carmago, email: alebcamargo@gmail.com; dr. juan luco, email: jmluco@unsl.edu.ar. received: march 06, 2019; revised: july 12, 2019; available online: august 05, 2019 abstract cyp2a6 is a human enzyme responsible for the metabolic elimination of nicotine, and it is also involved in the activation of procarcinogenic nitrosamines, especially those present in tobacco smoke. several investigations have reported that reducing this enzyme activity may contribute to anti-smoking therapy as well as reducing the risk of promutagens in the body. for these reasons, several authors investigate selective inhibitors molecules toward this enzyme. the aim of this study was to evaluate the interactions between a set of organosulfur compounds and the cyp2a6 enzyme by a quantitative structure-activity relationship (qsar) analysis. the present work provides a better understanding of the mechanisms involved, with the final goal of providing information for the future design of cyp2a6 inhibitors based on dietary compounds. the reported activity data were modeled by means of multiple regression analysis (mlr) and partial least-squares (pls) techniques. the results indicate that hydrophobic and steric factors govern the union, while electronic factors are strongly involved in the case of monosulfides. keywords quantitative structure-activity relationship (qsar); cyp2a6; inhibitors; organosulfur compounds introduction cytochrome p450s (cyps) comprise a superfamily of heme-containing enzymes responsible for the metabolism of numerous compounds, therefore they are important objects of study mainly in the areas of pharmacology and toxicology [23]. cyps are known for the diversity of reactions they catalyze, as well as, the range of different molecules with whom they interact. cytochrome p450s are responsible for the majority of phase i metabolism of exogenous molecules such as drugs, xenobiotics, environmental pollutants, dietary compounds, and endogenous molecules such as steroids, fatty acids, and prostaglandins [3,5,10]. the effects of these transformations can be manifested in drug bioavailability and toxicity, adverse drugs interactions, activation of procarcinogenic compounds, and biotransformation of molecules for subsequent elimination [8,9]. http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:alebcamargo@gmail.com mailto:jmluco@unsl.edu.ar admet & dmpk 7(3) (2019) 196-209 organosulfur – cyp2a6 qsar study doi: 10.5599/admet.678 197 among the cyp enzymes, cyp2a6 is the only member of the 2a subfamily that is expressed in humans and it catalyzes the coumarin 7-hydroxylation reaction (marker substrate)[5,6,23]. the current interest in cyp2a6, its polymorphisms and alleles, is mostly due to its effects on tobacco addiction, and the subsequent impact on lung cancer. accordingly, two important features are noteworthy: cyp2a6 is the major catalyst of metabolic elimination of nicotine, and it is also involved in the activation of procarcinogenic nitrosamines, especially those present in tobacco smoke [14]. on the other side, nicotine is the essential component that causes tobacco dependence. the pharmacological, as well as, physiological effects of smoking are due to this alkaloid. pulmonary absorption of nicotine is extremely rapid, and once it is absorbed, it is rapidly and extensively metabolized and eliminated through urine. in general, 70-80 % of its biotransformation consists mainly in the oxidation of (-)-nicotine to form (-)-cotinine mediated by cyp2a6. nicotine is also metabolized to nornicotine, via ndemethylation. in humans, 2-3 % of nicotine is excreted as nornicotine in 24 h urine [5,20,21]. several nicotine preparations have been developed as a medication to assist in smoking cessation; also other therapies and treatments have been proposed to reduce tobacco consumption, among them, we can name the use of nicotine patches, dopamine uptake inhibitor (bupropion), consumption of tricyclic antidepressants and buspirone anxiolytic. however, the previous therapies are of limited use because of the secondary effects that come with them and due to a low proven success (below 40%), for this reasons alternative treatments are encouraged [22]. with this goal, several investigations have shown that the inhibition of cyp2a6 by different chemical compounds may represent a potential supplement to antismoking therapy [12]. the efficiency of cyp2a6-mediated biotransformation of nicotine is associated with nicotine blood levels for keeping the addiction liability. potent and specific inhibitors of cyp2a6 might increase nicotine half-life elimination time. consequently, the inhibition of cyp2a6 results in a diminished desire to smoke and a reduced consumption of toxic or carcinogenic components of cigarette smoke [17]. moreover, the administration of chemicals that strongly and specifically inhibit the cyp2a6 activity might also result in a reduction of mutagens in the body, since the n-nitrosamines activation does not occur when this enzyme is altered [5]. numerous compounds have been proposed and tested as cyp2a6 inhibitors, among them methoxsalen, (r)-(+)-menthofuran, and tranylcypromine can be mentioned. even though they have shown strong inhibitory effects, methoxsalen and tranylcypromine also inhibited other forms of cyp [5]. on the other hand, in silico studies, such as qsar models, serve to predict and explain molecular properties and/or biological activities from the structure of chemical compounds reducing the number of tests, thereby resulting in a quick and simple alternative for designing new drugs with a desirable effect. in this sense, some authors have carried out in silico studies to predict selective cyp2a6 inhibitors based on a series of nicotine derivatives [17], naphthalene and non-naphthalene derivatives [11,15], or based on virtual screening of a large number of compounds [13]. to date, all the studied in silico inhibitors are from synthetic origin. however, it has been proved that some sulfide and disulfide compounds present in allium vegetables, inhibit cyp activity. in this sense, the anticarcinogenic and antitumorigenic effects of garlic in rodents have been attributed to modulation of cyp activity for some oscs such as diallyl sulfide (das), diallyl disulfide (dads) and allyl methyl sulfide (ams) [7,16]. fujita and kamataki, (2001) [5] studied the inhibition of cyp2a6 by determining the ability of 22 organosulfur compounds to inhibit coumarin 7hydroxylase activity. based on these experimental values, we carried out a quantitative structure-activity relationship (qsar) analysis to evaluate these inhibitors-enzyme interactions in order to obtain a wider vision into the mechanisms involved. in turn, a potential selective inhibitor of cyp2a6 based on naturally occurring compounds can be proposed. the reported activity data were modeled by means of multiple d.a. ramirez et al. admet & dmpk 7(3) (2019) 196-209 198 regression analysis (mlr) and partial least-squares (pls) techniques. for the quantitative description of the analyzed structures, different quantum chemical indices (am1) and physicochemical parameters were calculated. non-empirical descriptors, such as topological and geometrical, were also used as descriptive variables. finally, the obtained results provided an explanation about which are the main features and parameters of oscs (lipophilicity, steric and electronic features) that contribute to an optimal inhibitory activity on cyp2a6. materials and methods biological and chemical data the chemical structures along with observed activity data of the compounds used in this study are shown in table 1. the inhibitory activity data on cyp2a6 were obtained from fujita and kamataki, (2001) [5], and it was expressed as percentage of residual activity (ra%) when using 10 µm of each inhibitor and 2.5 µm of coumarin. in order to evaluate whether the biological data comes from a normal distribution, the standardized skewness and kurtosis values were calculated for ra% (n=22). structural descriptors a set of molecular descriptors were calculated to characterize the compounds under study. as for indicators of molecular size, the following were considered: molar volume (vm), molecular weight (mw), and molar refractivity (mr). to explain lipophilicity effects, several calculated partition coefficients were obtained by using alogps 2.1 software (log pexp, alogps, ia log p, cosmofrag, milog p, kowwin and xlogp). another group of structural descriptors included quantum chemical indices obtained by hyperchem package (release 7.5 for windows). three-dimensional molecular structures were built using the mm+ molecular mechanics potential-energy function. in a follow-up procedure, complete optimization of the geometrical parameters was carried out by using the am1 method, implemented in the standard version of mopac 6.0. the following indices obtained from molecular orbital calculations were considered: total energy (etotal), heat of formation (¢hf), energy of highest occupied molecular orbital (homo), energy of lowest unoccupied molecular orbital (lumo), dipole moment (µ), absolute total charge (qtotal), the most positive and the most negative absolute charges (qp,max, qn,max), and the positive and negative relative charge (rncg, rpcg). the last group of descriptors considered in this study included various geometrical and topological indices: the wiener index, the valence and connectivity molecular indices, the kappa shape indices, and several geometrical indices calculated from the optimized distance matrix am1 by using dragon software (v 3.0). a stepwise multiple regression procedure, based on the algorithms forwardselection and backward-elimination, was used for the inclusion or rejection of descriptors in the screened models. statistical methods partial least squares projections in latent variables (pls) together with multiple regression analysis (mlr) were the selected methods to search for relationships between the biological activity data and the structural descriptors. the determination of the significant number of pls components was made by crossvalidation [1,19]. pls analysis was carried out using the simca-p 7.01 software obtained from umetri ab, umea, sweden, and mlr analysis was performed using the 7.0 version of statgraphics plus software. prior conducting the pls-1 analysis for each one of the classes, all the variables were auto-scaled to zero mean and unit variance. this transformation assures data standardization and allows to give each descriptor equal importance in the pls analysis. to avoid overestimations or misunderstanding interpretation of the resulting models, pairs of variables with r ≥ 0.75 were classified as intercorrelating admet & dmpk 7(3) (2019) 196-209 organosulfur – cyp2a6 qsar study doi: 10.5599/admet.678 199 ones, and only one of these was included in the screened model. the predictive ability of the model was evaluated by the crossvalidation coefficient (q 2 ), which is based on the prediction error sum of squares (press). the press statistic is computed as the squared differences between observed and predicted values when the observations are kept out of the derived model. this procedure is repeated several times until every observation has been kept out once, and only once. results multiple regression analysis mlr was performed on the 22 oscs described in table 1, whereas table 2 shows the molecular descriptors included in the selected models. table 1. cyp2a6 inhibitory activity data of organosulfur compounds n° name chemical structure residual activity % 1 dimethyl sulfide ch3 s ch3 98.4 2 dimethyl disulfide ch3 s s ch3 94.1 3 diethyl sulfide s ch3ch3 89.4 4 diethyl disulfide s s ch3 ch3 79.6 5 di-n-propyl sulfide s ch3ch3 83.4 6 di-n-propyl disulfide s sch3 ch3 39.5 7 di-n-butyl sulfide sch3 ch3 82.0 8 di-n-butyl disulfide s s ch3 ch3 43.9 9 di-n-amyl sulfide s ch3 ch3 88.8 10 di-n-amyl disulfide s s ch3 ch3 66.3 11 diallyl sulfide s ch2ch2 91.3 12 diallyl disulfide s sch2 ch2 46.2 13 allyl methyl sulfide s ch3ch2 97.6 14 allyl n-propyl sulfide s ch2 ch3 94.2 15 allyl phenyl sulfide ch2 s 32.8 16 diphenyl sulfide s 46.1 17 diphenyl disulfide s s 65.7 d.a. ramirez et al. admet & dmpk 7(3) (2019) 196-209 200 n° name chemical structure residual activity % 18 phenyl ciclopropyl sulfide s 42.1 19 difurfuryl disulfide s s o o 69.1 20 4,4’-dipyridyl disulfide n s s n 97.0 21 4,4’-dipyridyl sulfide n s n 3.22 22 2,2’-dipyridyl disulfide n s s n 2.26 table 2. molecular descriptors a for the compounds described in table 1 n° ra% clogp ehomo elumo µ ldip rpcg rncg e1u e3u spam asp qmean 1 98.400 0.845 -8.882 0.457 1.961 0.251 0.200 0.484 0.707 0.444 0.517 0.237 0.093 2 94.100 0.845 -9.399 -1.978 2.569 0.249 0.193 0.473 0.724 0.449 0.545 0.235 0.095 3 89.400 1.817 -8.856 0.362 1.941 0.154 0.118 0.282 0.654 0.533 0.488 0.665 0.076 4 79.600 1.817 -9.352 -1.976 2.646 0.160 0.132 0.289 0.676 0.484 0.504 0.521 0.078 5 83.400 2.604 -8.857 0.353 1.910 0.146 0.090 0.216 0.632 0.571 0.480 0.765 0.074 6 39.500 2.604 -9.351 -1.996 2.660 0.151 0.103 0.230 0.677 0.358 0.489 0.547 0.075 7 82.000 3.291 -8.861 0.342 1.886 0.134 0.071 0.170 0.612 0.593 0.475 0.875 0.073 8 43.900 3.291 -9.351 -1.999 2.676 0.139 0.083 0.183 0.642 0.367 0.477 0.659 0.074 9 88.800 3.911 -8.865 0.337 1.881 0.129 0.059 0.141 0.605 0.606 0.472 0.900 0.072 10 66.300 3.911 -9.354 -2.002 2.677 0.133 0.069 0.153 0.643 0.374 0.473 0.681 0.073 11 91.300 2.352 -8.901 -0.068 2.004 0.153 0.133 0.169 0.594 0.348 0.525 0.763 0.100 12 46.200 2.352 -9.313 -2.010 2.650 0.157 0.127 0.171 0.626 0.294 0.515 0.346 0.102 13 97.600 1.672 -8.894 0.146 1.987 0.187 0.178 0.314 0.590 0.310 0.527 0.604 0.098 14 94.200 2.459 -8.880 0.108 1.958 0.150 0.139 0.209 0.594 0.353 0.518 0.778 0.086 15 32.800 3.437 -8.566 -0.076 1.649 0.140 0.113 0.161 0.574 0.227 0.514 0.642 0.105 16 46.100 4.412 -8.371 -0.142 1.265 0.127 0.130 0.127 0.613 0.000 0.474 0.615 0.110 17 65.700 4.412 -8.960 -2.164 2.495 0.139 0.096 0.188 0.624 0.331 0.477 0.484 0.105 18 42.100 3.527 -8.557 0.086 1.627 0.158 0.105 0.217 0.618 0.172 0.467 0.524 0.112 19 69.100 1.182 -9.143 -2.170 2.191 0.126 0.123 0.152 0.627 0.376 0.519 0.749 0.101 20 97.000 1.406 -9.320 -2.365 5.911 0.140 0.108 0.158 0.602 0.248 0.498 0.471 0.105 21 3.220 1.406 -9.007 -0.667 0.321 0.135 0.160 0.133 0.620 0.000 0.487 0.630 0.113 22 2.260 1.406 -9.573 -2.627 0.150 0.121 0.111 0.160 0.602 0.248 0.500 0.534 0.108 a for an explanation of the symbols, refer to the text first, in order to evaluate the data set distribution the standardized skewness and kurtosis values were calculated for ra% (n=22). skewness is a measure of symmetry, or more precisely, the lack of symmetry; admet & dmpk 7(3) (2019) 196-209 organosulfur – cyp2a6 qsar study doi: 10.5599/admet.678 201 and kurtosis indicates whether the data distribution is heavy-tailed or light-tailed relative to a normal distribution. in other words, data sets with high kurtosis tend to have heavy tails, or outliers. the obtained results considering ra% were -1.5517 and -0.2634 for the standardized skewness and kurtosis values, respectively. values outside -2 to +2 range indicate significant departures from normality. in this case, the obtained values are within the expected range for data presenting normal distribution. on the other hand, because of the large number of structural descriptors considered in this study, the vip (variable importance for the projection) parameter [18] was used to unravel which descriptors were relevant to explain the enzyme activity. vip is based on the pls projection method, which constitutes a robust method for variables selection. considering the arguments presented by yano et al. (2006) [22] along with other background qsar studies about cyp2a6 enzyme, which showed that the nature of the enzyme active site is strongly hydrophobic, it was decided first, to study the relationship between activity and lipophilicity. after using several calculated log p parameters, the obtained equations showed no significant correlations in this direction. on the other hand, since lipophilicity can be described by the contribution of the molecular size and molecular polarity, models that incorporated both molecular properties were investigated. the best regression equation that includes the parameter clogp, was: ra% = -193.05 (30.01) + 11.60 (3,13) clogp + 17.76 (2.50) µ (1) (0.000) (0.002) (0.000) + 126.33 (16.40) e3u + 823.4 (113.5) rpcg + 81.78 (18,63) asp (0.000) (0.000) (0.000) r 2 = 0.90; r 2 (cv) = 0.673; s = 10.83; n = 22; f = 28.77; wclogp = 0.441; wµ = 0.646; we3u = 0.693; wrpcg = 1.041; wasp = 0.489 in equation (1) and the following equations, n is the number of compounds, s is the standard deviation, r 2 is the squared correlation coefficient, r 2 (cv) is the squared crossvalidation coefficient, and f is the fisher f statistic. values in parentheses correspond to the standard deviations and p-values of coefficients, and the term w, represents the standardized regression coefficient when variables are scaled to the same numerical range (0-1). the statistical quality of the model is suitable. although the relationship between numbers of observations versus the number of variables is high, the good quality in fit and predictive ability, as expressed by r 2 and r 2 (cv) coefficients, suggests that the obtained model is valid. other models were also investigated, in which the relationship between observations / variables were lower in relation to the previously derived equation. ra% = 301.77 (82.54) + 39.177 (9.379) ehomo + 130.41 (18.08) e3u (2) (0.002) (0.001) (0.000) + 15.155 (2.744) µ + 333.23 (76.60) rpcg (0.000) (0.000) r 2 = 0.853; r 2 (cv) = 0.783; s = 12.71; n = 22; f = 24.71; wehomo = 0.415; we3u = 0.715; wµ = 0.551; wrpcg = 0.421 d.a. ramirez et al. admet & dmpk 7(3) (2019) 196-209 202 this equation is highly significant, and the molecular descriptors here used, did not show significant intercorrelations. figure 1 shows the relationships between the experimental and calculated residual activity (ra%) for equations (1) and (2), respectively. an interesting point to highlight in this equation is an improvement in the crossvalidation coefficient value, which suggests a better predictive ability. figure 1. relationship between observed and calculated values for (a) equations (1) and (b) equation (2) an alternative way to evaluate the reliability and robustness of a qsar model consists in applying different statistical approaches, but using the same descriptors that have been included in the original model. accordingly, a linear discriminant analysis (lda) was carried out based on the molecular descriptors used in equation (2) to maintain an adequate balance between the numbers of observations vs. the number of independent variables. taking into account the information from the inhibitory activity data, the compounds were separated into three groups, namely highly active (ra% 1-10), active and weakly active (ra% 11-70) and weakly active or inactive (ra% 71-100). the discriminant ability was assessed by the correct classifications percentage, and the discriminant function quality was evaluated using the wilks parameter, λ, which was obtained by a multivariate analysis of variance that tests the equality of group means for the variable in the discriminant model. the first standardized discriminating function was: df1 = 1.842 ehomo + 1.699 µ + 1.884 e3u + 1.063 rpcg (3) f = 132.14; wilks’ lambda (λ) = 0.0627; p(value) = 0.0000; n = 22 this equation is highly significant as p < 0.00001, and amongst the 22 observations used to fit the model, 22 (or 100.0 %) were correctly classified. in order to check an overestimation of the obtained model, the rate of correct classification was evaluated by using crossvalidation and a 95.5% was obtained (21 to 22). from the relative magnitude of the obtained coefficients in equation (3), it can be appreciated the contribution of a single variable to discriminate amongst the groups. figure 2 plots each observation in the space of the two linear discriminant functions (df1, df2). admet & dmpk 7(3) (2019) 196-209 organosulfur – cyp2a6 qsar study doi: 10.5599/admet.678 203 figure 2. 2-d plot of discriminant functions df1 and df2. the fact that similar conclusions are obtained by using both kinds of statistical methods; that is, mlr and lda, provides more evidence about the model robustness here reported. the data set, includes both monosulfides as disulfides. in the case of monosulfides, these can act as strong nucleophiles (nu θ ) or reducing agents depending on the nature of the electrophile(e⊕), the substituent group (r-s), and the medium in which the reaction occurs. such property is related to the ehomo energy of monosulfides, thus, a greater ehomo implies a higher ionization potential and a superior ability to donate electrons. moreover, disulfides act as electrophilic oxidants, i.e. tend to accept electrons, which is reflected by their low lumo energy orbital. considering the above, the set of compounds under study was divided among monoand disulfides, which were analyzed independently. the relationship between monosulfides and their corresponding ehomo energy showed a parabolic correlation as presented in equation (4). ra% = 30195 (5506) + 7083 (1272) ehomo + 415.83 (73.42) ehomo 2 (4) (0.001) (0.001) (0.000) r 2 = 0.968; r 2 (cv) = 0.909; s = 4.82; n = 11; f = 122.28; wehomo = 0.535; wehomo 2 = 0.544 the quadratic term ehomo is statistically significant (p = 0.000), although this term depends very strongly on compound 16. thus, only the analysis of a data serie including a larger amount of monosulfides would help to validate the parabolic relationship obtained. another aspect to highlight, is that compound 21 (4,4' dipyridyl sulfide) was excluded from the analysis because it presented an outlier behavior, given its polarity and structural nature, that differs from the other monosulfides (this may be due to the fact that compound 21 is the only pyridyl monosulfide compound). in the case of the 10 disulfides under study, the best-obtained regression equation was: ra% = -171.4 (35.97) + 16.871 (2.114) µ + 294 (46.64) e3u + 907.1 (243.2) qmean (5) (0.003) (0.000) (0.001) (0.010) r 2 = 0.938; r 2 (cv) = 0.863; s = 8.653; n = 10; f = 30.34; wµ = 0.819; we3u = 0.798; wqmean = 0.470 discriminant function 1 d is c r im in a n t f u n c t io n 2 5.02.50.0-2.5-5.0-7.5-10.0 3 2 1 0 -1 -2 class 1 2 3 c lass 1 : range 7 0 -1 0 0 ra % c lass 2 : range 1 0 -7 0 ra % c lass 3 : range 1 -1 0 ra % d.a. ramirez et al. admet & dmpk 7(3) (2019) 196-209 204 in figure 3, it is possible to observe an agreement between the experimental and calculated values of ra%, as well as the normal distribution of residual values. the values of the correlation coefficient (r) between pairs of variables used in this equation showed an acceptable intercorrelation: µ/e3u -0.099, µ/qmean -0.028, and e3u/qmean -0.586. figure 3. relationship between observed and calculated values for equation (4). partial least square-1 regression analysis a preliminary analysis of different pls models performed on the whole series of compounds showed that the most important variables were those that had been useful in the regression analysis. finally, the analysis of molecular descriptors was performed using pls-pseudoregression coefficients, and the statistical significance of all qsar-pls derivatives models was evaluated by means of the following statistical parameters: variance of the matrices x and y (r2x and r2y), the correlation coefficient (r), standard deviation (rmss) and the statistical f. the pls model was built using 21 of the total set of oscs; compound 15 (allyl phenyl sulfide) was not included in the model because of its outlier behavior, as observed in mlr-qsar models. the pls-1 analysis resulted in two statistically significant components (cpls), which statistical parameters are detailed in table 3. table 3: statistical parameters and pls model comp. r2x(cum) r2y(cum) q 2 (cum) rmss f statistical cpls-1 0.400 0.587 0.418 9.1347 122.380 cpls-2 0.566 0.914 0.817 (r=0.965) (q=0.904) admet & dmpk 7(3) (2019) 196-209 organosulfur – cyp2a6 qsar study doi: 10.5599/admet.678 205 figure 4. relationship between observed and calculated values by the pls-1 model derived. figure 4, shows the relationship between ra% experimental values and the corresponding calculated values derived from the pls-1 models. as can be observed in figure 5, the normal probability plot of the residuals is approximately linear, which indicates that the error terms are normally distributed. moreover, the corresponding pseudoregression coefficients are shown in figure 6. from these values, it can be observed how much a single variable contributes to ra% modelling. figure 5. n-probability plot of residuals from the pls-1 model. figure 6. bar graph of the standardized coefficients of the pls-1 model. d.a. ramirez et al. admet & dmpk 7(3) (2019) 196-209 206 validation of pls-1 model the actual predictive ability of a qsar model can be judged using compounds not included originally in the calibration series, or alternatively, by using a permutation method [4]. thus, to show that the developed pls-1 model was not a result of chance, both validation methodologies were used. ra values [%], predicted for 4 unused compounds are shown in table 4. as can be seen, predictions are within the range of experimental error, so they are quite acceptable. it should be mentioned that the classification model using lda, as previously shown in equation (3), correctly predicts the activity of the four inhibitors shown in table 4. table 4. predictions from the derived qsar model known inhibitors activity (observed) a activity (predicted) b pmodx (ps) c nicotine 98.1 91.614 0.03391 tranylcypromine 12.9 29.952 0.31252 cotinine 96.2 99.324 0.00343 sm-12502 85.9 84.476 0.01204 a activities expressed as ra% (source: fujita and kamataki, 2001) b predicted activity values (ra%) using the developed pls-1 model c probability of belonging to the descriptor space used in the model. in this case, tranylcypromine and sm-12502 they belong, while nicotine and cotinine are borderline. as previously mentioned, the validity of the pls-1 model was additionally tested by a permutation test. models were recalculated for randomly reordered response data (ra%). these permuted ra% values were related to intact predictor data by refitting the model and including crossvalidation. when r2 and q2 were plotted as a function of the correlation coefficient between the original values and the predicted values, the interception point with the y axis expressed how much these values rely on chance. in figure 7, the corresponding plot of ra% response permutation test is presented. the vertical axis gives the r2 and q2 values of each test. the horizontal axis represents the correlation coefficient between the observed and the permuted ra% values (200 permutations of the ra% dependent variable under study). figure 7. permutation graph of the pls-1 derived model. the intersection of the two regression lines (for r 2 and q 2 ) in the figure indicates the degree of overshooting and predictability. in general, the valid model limits include values of r 2 <0.30 and q 2 <0.30. this condition is fulfilled, indicating that the pls-1 model obtained is suitable both, for its adjustability as for its predictive ability. admet & dmpk 7(3) (2019) 196-209 organosulfur – cyp2a6 qsar study doi: 10.5599/admet.678 207 discussion the qsar models here proposed showed the main factors prevailing the osc-cyp2a6 binding. several statistical analyses were carried out for this purpose. four equations resulted from the multiple regression analysis. the presence of clogp in equation (1) suggests that the molecular lipophilicity plays an essential role in the inhibition of cyp2a6. this fact was experimentally demonstrated by yano et al. (2006) [22] and corresponds with the hydrophobic nature of the active site. on the other hand, equation (1) shows a high influence of the molecular shape in the inhibition process, as expressed by the asp and e3u descriptors. the presence of the asp (asphericity) descriptor, which is a measure of the spherical shape deviation of the molecule, allows us to infer that at higher molecular elongation, a more significant inhibitory activity can be observed. regarding the e3u geometric descriptor, which belongs to the directional whim descriptors encoding the planar characteristics of the molecules, it will acquire values tending to zero for molecules ideally planar, and values tending to one for molecules that deviate from this property. thus, considering equation (1) for the compounds under study, it was observed that compounds having lower e3u descriptor value, generally exhibit a higher inhibitory capacity. furthermore, from the analysis of equation (1), it was observed that the electronic factors are also involved in the inhibition of cyp2a6, as expressed by µ and rpcg descriptors. hence, these descriptors reflect the ability of the compounds to participate in polar interactions, suggesting that higher molecular polarity will correspond to lower activity. an interesting feature to consider from equations (2) and (3) is the highly significant contribution of the electronic parameter ehomo. the cavity in the substrate-binding site of cyp2a6 is strongly hydrophobic and this characteristic is due to the presence of several amino acid residues of phenylalanine. thus, ehomo parameter suggests a π-π and / or n-π interaction between oscs and phenylalanine residues. according to previous studies concerning the catalytic versatility of cytochrome p450 [2], regarding the functional active oxygen of heme groups, it has been postulated that this group shows nucleophilic and/or electrophilic properties catalyzing different type of reactions such as aldehyde deformilations, epoxidations and/or hydroxylation. on another note, the analysis of equation (4), which only considers monosulfides, suggests that the electrons donor ability of these compounds, expressed through ehomo parameter, plays a critical role in the interaction with cyp2a6. thus, at higher ehomo energy of monosulfides, a greater inhibitory capacity can be observed. in short, from the analysis of the equation, it is possible to infer that the monosulfides could act as nucleophiles and interact with hydroperoxo-iron forms and/or oxenoid-iron from heme group. however, the interaction efficiency between monosulfides-cyp2a6 also depends on the hydrophobicity, or alternatively on the lower polarity that these compounds present, as reflected by equations (1) and (2). an important aspect to mention is that the elumo parameter was not significant as a descriptor variable in any of the models investigated. this fact suggests that in the case of the disulfides, the ability to accept electrons was not relevant to the cyp2a6-disulfide interaction. thus, it is possible to postulate that the inhibitory potency of the disulfides depends mainly on geometric and electronic factors. to further investigate the oscs-cyp2a6 binding, a pls-1 analysis was carried out. analyzing the model coefficients (figure 6), it was observed that the geometric terms e1u, e3u, and spam play an essential role in the model fit, in agreement with the mlr technique results. this aspect indicates the crucial role of steric factors (including non-specific van der waals interactions) in the inhibitory activity exerted by such compounds. the other molecular parameters showed the importance of the electronic aspects of the oscd.a. ramirez et al. admet & dmpk 7(3) (2019) 196-209 208 cyp2a6 interaction. the presence of rncg, µ and ldip coefficients in the model indicates that an increase in the molecular polarity plays a negative effect on the inhibitory activity of the oscs; which fully corresponds with the current knowledge on the hydrophobic nature of the active site. finally, the model showed that in the case of monosulfides, ehomo electronic parameter, or the ability to donate electrons, is a significant factor in the enzyme-inhibitor binding. conclusions this work allows obtaining a better understanding of the mechanisms involved between the enzyme cyp2a6 and oscs. the obtained results indicate that hydrophobic and steric factors govern the union, while electronic factors via the electrons donor ability are strongly involved in the case of monosulfides, which could act as nucleophiles and interact with hydroperoxo-iron forms and/or iron-oxenoid from heme groups. it was also possible to conclude that the inhibitory effect of the disulfides depends mainly on geometric and electronic factors. furthermore, it was determined that an increase in the molecular polarity exerts a negative effect on the inhibitory activity of the tested compounds, which corresponds with previous studies that indicate a high hydrophobicity of the enzyme active site. on the other side, this study provides evidence of the enormous potential of whim descriptors for the development of qsar models. authors contributions the manuscript was written through contributions of all authors. all authors have given approval to the final version of the manuscript. funding sources the authors would like to thank the financial support of the consejo nacional de investigaciones científica y técnicas (conicet), the agencia nacional de promoción científica y tecnológica (anpcyt), the universidad nacional de cuyo, and the universidad nacional de san luis. conflict of interests: none. references [1] a.b. camargo, e. marchevsky, j.m. luco. qsar study for the soybean 15-lipoxygenase inhibitory activity of organosulfur compounds derived from the essential oil of garlic. j. agric. food chem. 55 (2007) 3096–103. [2] m.j. coon. multiple oxidants and multiple mechanisms in cytochrome p450 catalysis. biochem. biophys. res. commun. 312 (2003) 163–168. [3] p.b. danielson. the cytochrome p450 superfamily: biochemistry, evolution and drug metabolism in humans. curr. drug metab. 3 (2002) 561–597. [4] l. eriksson, e. johansson, m. miiller, s. wold. cluster-based design in environmental qsar. quant. struct. relationships. 16 (1997) 383–390. [5] k.-i. fujita, t. kamataki. screening of organosulfur compounds as inhibitors of human cyp2a6. drug metab. dispos. 29 (2001) 983–989. [6] a. le gal, y. dréano, d. lucas, f. berthou. diversity of selective environmental substrates for human cytochrome p450 2a6: alkoxyethers, nicotine, coumarin, n-nitrosodiethylamine, and nnitrosobenzylmethylamine. toxicol. lett. 144 (2003) 77–91. admet & dmpk 7(3) (2019) 196-209 organosulfur – cyp2a6 qsar study doi: 10.5599/admet.678 209 [7] m.b. genter, h. liang, j. gu, x. ding, m. negishi, r.a. mckinnon, et al., role of cyp2a5 and 2g1 in acetaminophen metabolism and toxicity in the olfactory mucosa of the cyp1a2 ( -/) mouse. biochem. pharmacol. 55 (1998) 1819–1826. [8] f.p. guengerich. common and uncommon cytochrome p450 reactions related to metabolism and chemical toxicity. chem. res. toxicol. 14 (2001) 611–650. [9] f.p. guengerich. mechanisms of cytochrome p450 substrate oxidation: minireview. j. biochem. mol. toxicol. 21 (2007) 163–168. [10] d.r. nelson, l. koymans, t. kamataki, j. stegeman, r. feyereisen, d. waxman, et al. p450 superfamily: update on new sequences, gene mapping, accession numbers and nomenclature. pharmacogenetics 6 (1996) 1–42. [11] m. rahnasto, h. raunio, a. poso, c. wittekindt, r.o. juvonen. quantitative structure activity relationship analysis of inhibitors of the nicotine metabolizing cyp2a6 enzyme. j. med. chem. 48 (2005) 440–449. [12] m. rahnasto, c. wittekindt, r.o. juvonen, m. turpeinen, a. petsalo, o. pelkonen, et al. identification of inhibitors of the nicotine metabolising cyp2a6 enzyme--an in silico approach. pharmacogenomics j. 8 (2008) 328–38. [13] m.k. rahnasto, h. a raunio, c. wittekindt, k. a salminen, j. leppänen, r.o. juvonen, et al. identification of novel cyp2a6 inhibitors by virtual screening. bioorg. med. chem. 19 (2011) 7186– 93. [14] h. raunio, m. rahnasto-rilla. cyp2a6: genetics, structure, regulation, and function. drug metabol. drug interact. 27 (2012) 73–88. [15] k. roy, p.p. roy. exploring qsar and qaar for inhibitors of cytochrome p450 2a6 and 2a5 enzymes using gfa and g/pls techniques. eur. j. med. chem. 44 (2009) 1941–51. [16] s. takahashi, k. hakoi, h. yada. enhancing effects of diallyl sulfide on hepatocarcinogenesis and inhibitory actions of the related diallyl disulfide on colon and renal carcinogenesis in rats. carcinogenesis. 13 (1992) 1513–1518. [17] y. wang, y. li, b. wang. an in silico method for screening nicotine derivatives as cytochrome p450 2a6 selective inhibitors based on kernel partial least squares. int. j. mol. sci. 8 (2007) 166–179. [18] h. van de waterbeemd. chemometric methods in molecular design. weinheim, frg, 1995. [19] s. wold, m. sjostrom, l. eriksson. pls-regression: a basic tool of chemometrics. chemom. intell. lab. syst. 58 (2001) 109–130. [20] y. yamaguchi, i. akimoto, k. motegi, t. yoshimura, k. wada, n. nishizono, et al. synthetic models related to methoxsalen and menthofuran – cytochrome p450 ( cyp ) 2a6 interactions. benzofuran and coumarin derivatives as potent and selective inhibitors of cyp2a6. chem pharm bull (tokyo) 61 (2013) 997–1001. [21] h. yamanaka, m. nakajima, t. fukami, h. sakai, a. nakamura, m. katoh, et al. cyp2a6 and cyp2b6 are involved in nornicotine formation from nicotine in humans: interindividual differences in these contributions. drug metab. dispos. 33 (2005) 1811–1818. [22] j.k. yano, t.t. denton, m.a. cerny, x. zhang, e.f. johnson, j.r. cashman. synthetic inhibitors of cytochrome p-450 2a6 : inhibitory activity, difference spectra, mechanism of inhibition and protein cocrystallization. j. med. chem. 49 (2006) 6987–7001. [23] u.m. zanger, m. schwab. cytochrome p450 enzymes in drug metabolism: regulation of gene expression, enzyme activities, and impact of genetic variation. pharmacol. ther. 138 (2013) 103– 141. ©2019 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ multi-lab intrinsic solubility measurement reproducibility in cheqsol and shake-flask methods doi: 10.5599/admet.698 210 admet & dmpk 7(3) (2019)210-219 doi: http://dx.doi.org/10.5599/admet.698 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index short communication multi-lab intrinsic solubility measurement reproducibility in cheqsol and shake-flask methods alex avdeef in-adme research, 1732 first avenue, #102, new york, ny 10128, usa corresponding author e-mail: alex@in-adme.com; tel.: +1-646-678-5713, orcid id: alex avdeef: 0000-0002-31395442 received: may 13, 2019; revised: may 27, 2019; published online: june 05, 2019 abstract this commentary compares 233 cheqsol intrinsic solubility values (log s0) reported in the wiki-ps0 database for 145 different druglike molecules to the 838 log s0 values determined mostly by the saturation shakeflask (ssf) method for 124 of the molecules from the cheqsol set. the range of log s0 spans from -1.0 to -10.6 (log molar units), averaging at -3.8. the correlation plot between the two methods indicates r 2 = 0.96, rmse = 0.34 log unit, and a slight bias of -0.07 log unit. the average interlaboratory standard deviation (sdi) is slightly better for the cheqsol set than that of the ssf set: sdi cs = 0.15 and sdi ssf = 0.24. the intralaboratory errors reported in the cheqsol method (0.05 log) need to be multiplied by a factor of 3 to match the expected interlaboratory errors for the method. the scale factor, in part, relates to the hidden systematic errors in the single-lab values. it is expected that improved standardizations in the ‘gold standard’ ssf method, as suggested in the recent ‘white paper’ on solubility measurement methodology, should make the sdi of both methods be about ~0.15 log unit. the multi-lab averaged log s0 (and the corresponding sdi) values could be helpful additions to existing training-set molecules used to predict the intrinsic solubility of drugs and druglike molecules. keywords interlaboratory solubility measurement errors; aqueous intrinsic solubility; shake-flask solubility; potentio metric solubility; thermodynamic solubility; henderson-hasselbalch equation introduction this commentary considers the interlaboratory reproducibility of published aqueous intrinsic solubility data (log s0) for 124 drugs, determined both by the potentiometric cheqsol (cs) method (233 reported log s0 cs values) and mostly by the ‘gold standard’ saturation shake-flask (ssf) method (838 log s0 ssf ). for each drug, its method-dependent interlaboratory measurement standard deviations (sdi cs and sdi ssf ) are estimated, by comparing solubility values for a given drug determined in different laboratories. the multilab averaged log s0 (and the corresponding sdi values) could be helpful additions to existing training-set molecules used to predict the intrinsic solubility of drugs and druglike molecules. the present contribution is the third in a series of papers aiming to address contemporary issues of solubility measurement, interpretation and prediction [1,2]. these are intended to serve as http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:alex@in-adme.com admet & dmpk 7(3) (2019) 210-219 multi-lab log s0 measurement reproducibility in cheqsol and shake-flask methods doi: 10.5599/admet.698 211 prologue/accompaniment to an upcoming session on solubility at the iapc-8 meeting in split, croatia, 9-11 september 2019. since 2009, the international association of physical chemists (iapc, www.iapchem.org) series of symposia maintained extensive coverage of the topic of solubility measurement, both from solid state and solution perspectives. at the 2015 iapc-4 meeting, the special session on solubility measurement resulted in a widely-circulated ‘white paper’ drawing on expert consensus thoughts of scientists from six countries (hungary, russia, serbia, spain, sweden, united states) [3]. it is expected that future sessions will continue to cover solubility methods and strategies, to critically address the different needs in pharmaceutical research, spanning from drug discovery to drug development. background most of the small-molecule research compounds in today’s drug discovery projects are ionizable and poorly soluble in water, and are thus prone to show low and/or erratic in vivo intestinal absorption [4]. in discovery, high-throughput microtitre plate methods are used to estimate solubility, where small volumes of 10 mm dmso solutions of library compounds are added to buffer solutions to induce formation of solid suspensions in the wells. such estimates of solubility are needed, in part, to anticipate whether compounds would precipitate in bioassays (and thus indicate false positives). in parallel, methods to predict solubility of molecules play an important role in drug discovery, since virtual screening of compound libraries could prioritize molecules for testing in early in vitro screens [2,5,6]. saturation shake-flask (ssf) and potentiometric (cheqsol) methods in more advanced stages of drug research, solubility measurement necessarily becomes more rigorous, where ‘solubility’ refers to the concentration of a solute in a saturated solution, where the dissolved molecule is in a thermodynamic equilibrium with its crystalline form suspended in the solution (of a known ph, composition, and temperature). accurate solubility measurement of druglike molecules can be very difficult to do well, although to an untrained eye it ought to be as easy as measuring the concentration of a molecule in water. in development projects, thermodynamic solubility measurements are usually done using some variant of the ssf method [3]. as an alternative, a potentiometric procedure, called the dissolution template titration (dtt) method was introduced in 1998 [7] and validated two years later [8]. the methodology is capable of producing highly precise (intralaboratory sd ~0.07 log unit) measurements of aqueous intrinsic solubility, log s0, i.e., the solubility of the uncharged form of an ionizable molecule. a much faster variant of the ph-titration method, called cheqsol, was described in 2005 [9]. instruments implementing the potentiometric methods have been used in several universities and pharmaceutical companies. all of the potentiometric methods require that the molecule be ionizable and that the accuratelymeasured pka be provided. the molecule cannot be too soluble, since the method depends on the ph difference between a saturated and an unsaturated solution in the titration where the molecule is half ionized. so, it is ideally suited for low-soluble molecules, since these molecules display large ph differences. furthermore, to calculate the cheqsol log s0, it is assumed that solubility as a function of ph follows the curve predicted by the henderson-hasselbalch equation. the molecule needs to be stable to hydrolysis when repeatedly exposed to ph conditions far from neutral. means to recognize hydrolytic decomposition are important to incorporate into the measurement. sometimes multiple polymorphs may form in the cheqsol method, which requires solid state characterization to identify. in the traditional ssf method, most of the time, the thermodynamically most-stable form of the solid is the one associated with the measured solubility [3]. equilibration times are selected to be long (24-168 h) to ensure this expectation. http://www.iapchem.org/ alex avdeef admet & dmpk 7(3) (2019) 210-219 212 challenging measurement what can make solubility measurement of an ionizable low-soluble molecule so difficult? there are two sides to consider for the reaction at equilibrium: (a) the solid state and (b) the solution. temperature needs to be regulated and specified. to keep the following examples simple, let’s assume that the solvent is distilled water or an aqueous buffer, and that the crystalline form of the test compound is a free-acid/base or a salt. multiple circumstances may arise, some making the interpretation of the measurement challenging: (i) the simplest suspension is the one where no drug ionization takes place on equilibration. one needs to measure the concentration of the compound in the saturated solution, and to confirm that the solid state form is unchanged – simple. (ii) however, if the solid introduced is not the thermodynamically most-stable polymorph (or is amorphous), then it is possible that the measured solubility would correspond to a different solid form. that’s important to know. (iii) complication can arise if a low-soluble weak base is added to water (usually saturated with co2 from the air). the ph will change, depending on the pka of the molecule. the ambient co2 may act as a buffer, so the final ph of the saturated solution needs to be carefully measured. otherwise, the calculated log s0 could be quite erroneous. (iv) if a solid salt of the compound is added to water, a supersaturated solution may form. on equilibration, there could be two precipitates in the suspension: the original compound salt and the neutral free-acid/base form of the solid (at the ph called ‘phmax’). solid state characterization of the solid(s) and the measurement of ph would be highly beneficial. (v) when a buffer medium is used, the analysis of the solution and solid states can be complicated, as water-soluble drug-buffer complexes or aggregates may form [10]. in a supersaturated solution, drug molecules may self-associate as sub-micellar aggregates, particularly if they are surface-active [11,12]. for bases introduced as drug salts into a high-ph solution, the charged drug in the supersaturated solution can disproportionate into oil or undergo precipitation into an amorphous solid, along with which charged water-soluble aggregates may co-exist [10-13]. given enough time, the multiple phases are expected to undergo transformation into a thermodynamically most-stable crystalline solid. good understanding of solution chemistry and solid state characterization is essential for correctly interpreting the results of solubility measurements, so that high-quality intrinsic solubility data can be reported [3]. the need for high-quality data in accurate solubility prediction accurate prediction of the intrinsic solubility of druglike molecules [2] requires that (a) the log s0 values used to train the prediction method are of high quality (with water solubility values, log sw, or values measured at a particular ph, log sph, properly corrected for ionization [14] and with all solubility values referring to the same temperature [15]), and (b) the compounds in the training set cover the druglike chemical space of the test set of compounds. these two important notions became the focus of a number of studies since 2008, spurred by the publications of llinàs et al. [16] and hopfinger et al. [17]. these authors introduced the ‘solubility challenge,’ a competition to probe the limits of prediction methods. the cheqsol method was used to measure the log s0 of 132 structurally diverse drugs. the log s0 values of 100 molecules were offered as the training set for the prediction of an external test set of 32 molecules (not found in the training set), whose values were not revealed before the completion of the competition. in a number of earlier studies, it was suggested that the typical error in measured aqueous solubility is ~0.6 log unit or higher, when the solubility values were collected from many published sources [18]. this suggested that the quality of prediction methods was approaching the experimental limit. however, in the admet & dmpk 7(3) (2019) 210-219 multi-lab log s0 measurement reproducibility in cheqsol and shake-flask methods doi: 10.5599/admet.698 213 solubility challenge competition all of the values came from one laboratory, and the intralaboratory precision (repetitive measurements of the same sample by the same chemist, using the same instrument) of the cheqsol data was reported to be sd = 0.05 log unit. it was not known what the expected interlaboratory precision would be, given the unknown systematic errors that might affect the accuracy of results. for example, when 125 published cheqsol values were compared in 2015 to those obtained by the ssf method, it was reported that r 2 = 0.90, prediction root-mean-square-error, rmse = 0.52 log unit, and there was a slight bias of -0.13 log unit [19]. the values in the comparison came from the wiki-ps0 database (in-adme research), which at that time contained 4557 log s0 entries. the database now has 6355 entries, with many newly added cheqsol and ssf values. it was thus of interest to update and better characterize the comparison of data quality between the cheqsol and the ‘gold standard’ saturation shake-flask methods. in parallel, armed with new curated data, the second solubility challenge has just been announced [2], with the prediction submission deadline set to 8 september 2019, the day before the iapc8 conference starts. the excel submission form in supporting info at https://pubs.acs.org/doi/suppl/10.1021/acs.jcim.9b00345 is freely downloadable for those interested to participate. the data described below could be a useful addition to other druglike training sets currently in circulation. method data source: wiki-ps0 database the ongoing wiki-ps0 database project [2,3,15,19], which started in 2011, now has 6355 log s0 entries for 3014 different drug-relevant molecules (solids at room temperature), drawing on the study of 1325 publications. the overall interlaboratory standard deviation, sdi all = 0.17 log unit, has been estimated from the 870 molecules for which solubility was reported from two or more different sources (comprising 4209 individual s0 values), by taking the average of the individual 870 sd values. the sdi all , being lower than the older estimate of solubility measurement error (~0.6 log unit [18]), indicates that (i) when legacy data are subjected to critical analysis, as recommended in [3,15,19], improvements in the quality of the extracted log s0 data can be achieved, and (ii) there is room for further improvement to the current prediction methods. alongside the database, the pdisol-x program (in-adme research) was designed to interpret solubility data and make temperature corrections, to produce a reliable estimate of the underpinning log s0 [10,11,20]. results there are 233 reported cheqsol log s0 values in the wiki-ps0 database for 145 different druglike molecules. for 124 of the molecules, there are 838 reported log s0 determined mostly by the ssf method. of the 838 entries, 298 (36 %) were log s0 values calculated from log s vs. ph data (using pdisol-x), based on a total of 2925 individual log sph measurments. (for 21 of the 145 molecules, ssf data have not been located in the literature.) table 1 lists the solubility values for the 124 overlapping molecules measured by the cheqsol and ssf methods. the range of log s0 spans from -1.0 to -10.6 (log molar units), averaging at -3.8. note that indomethacin is not listed in the table. the cheqsol value was not accepted in the wiki-ps0 database due to the hydrolytic decomposition encountered during the cheqsol assay [30]. on the other hand, the ph-metric ddt method indicated log s0 dtt = -5.33 [31], in good agreement with the average of 20 interlaboratory ssf measurements: log s0 ssf = -5.49, sdi = 0.23. https://pubs.acs.org/doi/%1fsuppl/10.1021/acs.jcim.9b00345 https://pubs.acs.org/doi/%1fsuppl/10.1021/acs.jcim.9b00345 alex avdeef admet & dmpk 7(3) (2019) 210-219 214 table 1. averaged saturation shake-flask (ssf) and cheqsol (cs) intrinsic solubility (log molar) a compound avg. 25 °c log s0 ssf sd ssf n avg. 25 °c log s0 cs sd cs n ref. (cheqsol) acebutolol -2.50 0.42 2 -2.68 0.31 1 [17] acetaminophen -0.97 0.10 18 -1.03 0.05 2 [16,22] acetazolamide -2.38 0.18 10 -2.44 0.04 1 [16] alprenolol -2.83 0.16 1 -2.66 0.04 2 [22,28] amantadine -2.76 0.50 1 -1.90 0.07 2 [22,29] amiodarone -10.58 0.35 4 -9.68 0.15 1 [22] amitriptyline -4.66 0.40 5 -4.55 0.15 1 [22] amodiaquine -4.74 0.35 1 -5.87 0.10 2 [16,22] amoxicillin -2.13 0.06 10 -1.97 0.08 1 [17] aripiprazole -6.74 0.16 2 -6.43 0.30 1 [28] atenolol -1.20 0.06 7 -1.25 0.06 4 [22,23,27,28] atropine -2.01 0.29 4 -2.00 0.07 1 [16] barbital,hexo-2.79 0.15 5 -2.67 0.07 1 [16] barbital,pheno-2.30 0.08 25 -2.29 0.07 1 [22] bendroflumethiazide -4.35 0.34 4 -4.19 0.15 2 [24,29] benzocaine -2.18 0.12 13 -2.33 0.31 1 [17] benzoic_acid -1.59 0.05 12 -1.61 0.15 1 [22] benzoic_acid,4-hydroxy-1.38 0.10 5 -1.46 0.04 1 [22] benzthiazide -4.84 0.28 4 -4.86 0.04 2 [22,24] bifonazole -6.27 0.25 4 -6.01 0.15 1 [25] bisoprolol -2.40 0.34 2 -1.46 0.26 1 [28] bupivacaine -3.47 0.18 9 -3.08 0.21 2 [16,24] carprofen -4.63 0.05 1 -4.70 0.22 1 [22] carvedilol -5.46 0.36 12 -4.46 0.27 1 [28] cephalothin -3.40 0.32 1 -2.94 0.04 1 [16] chlorpheniramine -2.65 0.10 1 -2.60 0.12 3 [21,22] chlorpromazine -5.45 0.31 11 -5.55 0.04 1 [22] chlorpropamide -3.15 0.17 5 -3.21 0.05 2 [16,28] chlorprothixene -5.66 0.38 3 -6.31 0.44 3 [16,22] chlorzoxazone -2.78 0.11 3 -2.64 0.04 2 [16,29] cimetidine -1.50 0.22 7 -1.69 0.04 1 [16] ciprofloxacin -3.57 0.19 19 -3.60 0.18 1 [22] cyproheptadine -5.02 0.46 4 -5.00 0.15 1 [24] diazoxide -3.46 0.26 3 -3.36 0.07 1 [16] dibucaine -3.70 0.35 1 -4.20 0.26 2 [17,24] diclofenac -5.33 0.19 27 -5.40 0.14 6 [9,21,22,24,26] diethylstilbestrol -4.38 0.38 6 -4.43 0.48 1 [17] difloxacin -3.94 0.10 2 -3.60 0.04 1 [16] diflunisal -5.08 0.40 5 -4.92 0.70 6 [17,21] diltiazem -2.95 0.20 1 -3.05 0.16 2 [22] diphenhydramine -3.29 0.64 3 -2.95 0.04 1 [22] dipyridamole -5.13 0.13 10 -5.16 0.02 1 [17] enrofloxacin -3.16 0.13 7 -3.18 0.18 1 [16] famotidine -2.67 0.27 5 -2.65 0.04 1 [22] admet & dmpk 7(3) (2019) 210-219 multi-lab log s0 measurement reproducibility in cheqsol and shake-flask methods doi: 10.5599/admet.698 215 compound avg. 25 °c log s0 ssf sd ssf n avg. 25 °c log s0 cs sd cs n ref. (cheqsol) fenoprofen -3.92 0.20 1 -3.70 0.26 1 [16] flufenamic_acid -5.27 0.28 10 -5.35 0.04 1 [22] flumequine -4.10 0.08 1 -3.80 0.11 2 [16,22] flurbiprofen -4.36 0.19 21 -4.05 0.15 2 [16,22] fluvastatin -3.78 0.06 1 -3.87 0.16 1 [28] furosemide -4.51 0.20 19 -4.18 0.12 3 [21,22] glibenclamide -6.63 0.45 10 -6.41 0.07 2 [27,28] gliclazide -4.27 0.40 14 -4.20 0.15 3 [23,27,28] glimepiride -7.22 0.42 5 -6.44 0.15 1 [27] glipizide -5.66 0.24 6 -5.51 0.10 3 [27-29] guanine -4.09 0.07 2 -4.43 0.37 1 [16] haloperidol -5.76 0.14 8 -5.52 0.19 2 [22,24] hydroflumethiazide -2.72 0.11 15 -2.70 0.03 3 [21,29] hydrochlorothiazide -3.08 0.08 5 -2.97 0.44 1 [16] ibuprofen -3.80 0.22 19 -3.97 0.23 4 [21-23] imipramine -4.39 0.28 7 -4.15 0.13 4 [17,21,22] ketoprofen -3.45 0.23 20 -3.23 0.02 4 [17,21,24] lidocaine -1.82 0.08 19 -1.87 0.05 1 [22] loperamide -6.80 0.32 3 -7.10 0.04 2 [16,22] maprotiline -4.54 0.26 3 -4.75 0.08 2 [22,24] meclizine -5.55 0.66 1 -6.48 0.15 1 [29] meclofenamic_acid -6.88 0.08 2 -6.56 0.42 2 [17,22] mefenamic_acid -6.36 0.42 6 -6.54 0.28 2 [16,22] metoclopramide -3.18 0.28 1 -3.58 0.07 1 [29] metoprolol -1.31 0.16 2 -1.22 0.01 2 [22,28] metronidazole -1.25 0.06 17 -1.22 0.04 1 [16] miconazole -6.04 0.41 4 -5.38 0.45 2 [16,22] nadolol -1.29 0.40 2 -1.63 0.08 2 [22,28] nalidixic_acid -3.76 0.26 8 -3.61 0.04 1 [16] naphthoic_acid,2-3.82 0.28 5 -3.77 0.15 1 [22] naphthol,1-2.09 0.10 6 -1.98 0.07 1 [22] naproxen -4.21 0.14 15 -4.41 0.25 2 [16,22] niflumic_acid -4.18 0.13 6 -4.11 0.08 2 [16,22] nitrofurantoin -3.34 0.11 11 -3.29 0.06 2 [16,22] norfloxacin -2.88 0.16 18 -2.75 0.15 1 [22] nortriptyline -3.94 0.23 3 -3.93 0.02 2 [16,22] ofloxacin -1.37 0.14 6 -1.27 0.04 1 [16] olanzapine -4.47 0.17 1 -4.23 0.18 1 [24] orphenadrine -3.71 0.35 1 -3.17 0.15 1 [22] paliperidone -4.56 0.18 1 -4.31 0.22 1 [28] papaverine -4.40 0.14 8 -4.21 0.24 4 [16,21,22,24] phenazopyridine -3.99 0.16 6 -4.19 0.07 1 [16] phenol,4-iodo-1.83 0.07 1 -1.72 0.04 1 [22] phenylbutazone -4.51 0.36 10 -4.39 0.04 1 [22] phenytoin -4.08 0.13 29 -3.86 0.18 1 [16] phthalic_acid,2-1.46 0.02 10 -1.55 0.08 2 [16,22] alex avdeef admet & dmpk 7(3) (2019) 210-219 216 compound avg. 25 °c log s0 ssf sd ssf n avg. 25 °c log s0 cs sd cs n ref. (cheqsol) pindolol -3.81 0.14 6 -3.64 0.13 3 [22,24,28] pioglitazone -6.21 0.81 3 -6.16 0.46 1 [28] piroxicam -4.46 0.26 16 -4.73 0.06 4 [16,21,22] pramoxine -3.46 0.39 1 -3.02 0.15 1 [22] probenecid -4.82 0.24 3 -4.86 0.04 1 [17] procaine -2.59 0.46 2 -1.72 0.07 1 [22] prochlorperazine -4.38 0.26 1 -4.75 0.15 1 [22] promethazine -4.39 0.19 10 -4.26 0.15 1 [22] propranolol -3.82 0.30 4 -3.49 0.06 6 [21-24,27,28] pyrimethamine -3.96 0.57 3 -4.11 0.22 1 [22] quinine -3.16 0.67 5 -2.80 0.01 2 [16,22] ranitidine -1.70 0.57 1 -2.50 0.26 1 [16] rosiglitazone -5.28 0.50 3 -5.22 0.61 1 [28] salicylic_acid -1.88 0.08 20 -1.93 0.04 1 [17] sertraline -5.41 0.28 2 -4.83 0.15 1 [22] sulfacetamide -1.50 0.05 5 -1.52 0.04 1 [16] sulfamerazine -3.10 0.07 6 -3.12 0.05 1 [22] sulfamethazine -2.63 0.27 6 -2.73 0.04 1 [16] sulfamethizole -2.77 0.14 5 -2.78 0.14 1 [17] sulfasalazine -6.47 0.08 7 -6.21 0.10 2 [16,22] sulfathiazole -2.61 0.23 8 -2.69 0.15 1 [22] terfenadine -7.68 0.71 10 -8.40 0.15 1 [29] tetracaine -3.22 0.11 1 -3.05 0.06 2 [16,24] tetracycline -3.29 0.05 6 -3.00 0.09 2 [16,29] tetracycline,oxy-3.27 0.07 4 -3.09 0.33 1 [16] thiabendazole -4.13 0.46 3 -3.48 0.10 1 [17] thymol -2.18 0.05 3 -2.19 0.04 1 [22] tolbutamide -3.55 0.11 5 -3.50 0.05 2 [17,28] tolmetin -4.05 0.15 1 -4.11 0.03 2 [16,22] trazodone -3.31 0.09 4 -3.19 0.40 2 [17] trichlormethiazide -3.64 0.12 3 -3.38 0.18 3 [16,22] trimethoprim -2.64 0.39 9 -2.95 0.15 1 [16] verapamil -4.36 0.28 9 -4.07 0.22 3 [21,22] warfarin -4.77 0.22 9 -4.80 0.14 2 [22,24] a sd refers to the standard deviation of reported values from n different sourcs. the average interlaboratory sdi values for for the two sets of data are: sdi ssf = 0.24 and sdi cs = 0.15 log unit. figure 1 shows the correlation plot between the two types of measurements, with each point showing both method interlaboratory error bars. the statistics have improved slightly over the 2015 comparison [19], with current values being r 2 = 0.96, rmse = 0.34 log unit, with a lower slight bias of -0.07 log unit. the average interlaboratory standard deviation is slightly lower for the cheqsol set over that of the ssf set: sdi cs = 0.15 and sdi ssf = 0.24, which probably highlights the benefit of using a highly standardized method (cheqsol) over an ‘open’ method (ssf). it should be kept in mind that the above comparison sets are small. for 870 of these sorts of comparisons, sdi all = 0.17. the intralaboratory comparison between the methods by one group of researchers performing both the ssf and cheqsol measurements (both highlystandardized) [32] produced the statistics r 2 = 0.96, rmse = 0.20 for 15 compounds, comparable to sdi all . admet & dmpk 7(3) (2019) 210-219 multi-lab log s0 measurement reproducibility in cheqsol and shake-flask methods doi: 10.5599/admet.698 217 the latter is the target for computational methods to aim at, provided that the training sets are of high and consistent quality. figure 1. the correlation plot between published cheqsol and saturation shake-flask (ssf) intrinsic solubility values (log molar) at 25 °c. the solid diagonal line is the identity line. the dashed lines are displacements by ±0.5 log. conclusions this brief commentary reasserts that the quality of the standardized cheqsol measurements is comparable to that of the ‘gold standard’ saturation shake-flask measurements. measurement errors are much lower than commonly acknowledged in the computational prediction community. the intralaboratory (single instrument) errors reported in the cheqsol method (0.05 log) need to be multiplied by a factor of 3 to match the expected interlaboratory errors for the method (0.15 log). the scale factor, in part, relates to the hidden systematic errors in the single-lab values. it is expected that better standardizations in the ‘open’ ssf methods, as recommended in the ‘white paper’ [3], may equalize the sdi of both methods at about ~0.15 log unit. when solubility prediction methods indicate rmse below 0.15, ‘overfitting’ is probably taking place, overlapping noise with information. acknowledgement the commentary sprang out of a suggestion made by a reviewer of the earlier paper in the series [2], that it would be helpful if a training set of data were provided for the new solubility challenge [2]. we are grateful for this good thought. conflict of interest. none. alex avdeef admet & dmpk 7(3) (2019) 210-219 218 references [1] c.a.s. bergström, a. avdeef. perspectives in solubility measurement and interpretation. admet & dmpk 7 (2019) 88-105. [2] a. llinas, a. avdeef. solubility challenge revisited after 10 years, with multi-lab shake-flask data, using tight (sd ~0.17 log) and loose (sd ~0.62 log) test sets, j. chem. inf. model. (2019). doi: 10.1021/acs.jcim.9b00345. [3] a. avdeef, e. fuguet, a. llinàs, c. ràfols, e. bosch, g. völgyi, t. verbić, e. boldyreva, k. takács-novák. equilibrium solubility measurement of ionizable drugs – consensus recommendations for improving data quality. admet & dmpk 4 (2016) 117-178. [4] c.a.s. bergström, r. holm, s.a. jørgensen, s.b.e. andersson, p. artursson, s. beato, a. borde, k. box, m. brewster, j. dressman, k.-i. feng, g. halbert, e. kostewicz, m. mcallister, u. muenster, j. thinnes, r. taylor, a. mullertz. early pharmaceutical profiling to predict oral drug absorption: current status and unmet needs. eur. j. pharm. sci. 57 (2014) 173-199. [5] j.c. dearden. in silico prediction of aqueous solubility. expert opin. drug discov. 1 (2006) 31–52. [6] j. taskinen, u. norinder. in silico prediction of solubility. in: b. testa, h. van de waterbeemd (eds.). comprehensive medicinal chemistry ii, elsevier: oxford, uk, 2007, pp. 627-648. [7] a. avdeef. ph-metric solubility. 1. solubility-ph profiles from bjerrum plots. gibbs buffer and pka in the solid state. pharm. pharmacol. commun. 4 (1998) 165-178. [8] a. avdeef, c.m. berger, c. brownell. ph-metric solubility. 2. correlation between the acid-base titration and the saturation shake-flask solubility-ph methods. pharm. res. 17 (2000) 85-89. [9] m. stuart, k. box. chasing equilibrium: measuring the intrinsic solubility of weak acids and bases. anal. chem. 77 (2005) 983-990. [10] o.s. marković, m.p. pešić, a.v. shah, a.t.m. serajuddin, t.z. verbić, a. avdeef. solubility-ph profile of desipramine hydrochloride in saline phosphate buffer: enhanced solubility due to drug-buffer aggregates. eur. j. pharm. sci. 133 (2019) 264-274. [11] a. pobudkowska, c. ràfols, x. subirats, e. bosch, a. avdeef. phenothiazines solution complexity – determination of pka and solubility-ph profiles exhibiting sub-micellar aggregation at 25 and 37 o c. eur. j. pharm. sci. 93 (2016) 163-176. [12] e. shoghi, e. fuguet, e. bosch, c. ràfols. solubility-ph profile of some acidic, basic and amphoteric drugs. eur. j. pharm. sci. 48 (2012) 290-300. [13] g. zografi, i. zarenda. the surface activity of phenothiazine derivatives at the air-solution interface. biochem. pharmacol. 15 (1966), 591–598. [14] a. avdeef. absorption and drug development second edition, wiley-interscience, hoboken nj, 2012, ch. 6. [15] a. avdeef. solubility temperature dependence predicted from 2d structure. admet & dmpk 3 (2015) 298-344. [16] a. llinàs, r.c. glen, j.m. goodman. solubility challenge: can you predict solubilities of 32 molecules using a database of 100 reliable measurements? j. chem. inf. model. 48 (2008) 1289-1303. [17] a.j. hopfinger, e.x. esposito, a. llinàs, r.c. glen, j.m. goodman. findings of the challenge to predict aqueous solubility. j. chem. inf. model. 49 (2009) 1-5. [18] a.r. katritzky, y. wang, s. sild, t. tamm, m. karelson. qspr studies on vapor pressure, aqueous solubility, and the prediction of water-air partition coefficients. j. chem. inf. model. 38 (1998) 720725. [19] a. avdeef. suggested improvements for measurement of equilibrium solubility-ph of ionizable drugs. admet & dmpk 3 (2015) 84-109. [20] g. völgyi, a. marosi, k. takács-novák, a. avdeef. salt solubility products of diprenorphine hydrochloride, codeine and lidocaine hydrochlorides and phosphates – novel method of data analysis not dependent on explicit solubility equations. admet & dmpk 1 (2013) 48-62. admet & dmpk 7(3) (2019) 210-219 multi-lab log s0 measurement reproducibility in cheqsol and shake-flask methods doi: 10.5599/admet.698 219 [21] k. baek, s.b. jeon, b.k. kim, n.s. kang. method validation for equilibrium solubility and determination of temperature effect on the ionization constant and intrinsic solubility of drugs. j. pharm. sci. emerg. drugs 6 (2018) 1-6. [22] k.j. box, j.e.a. comer. using measured pka, log p and solubility to investigate supersaturation and predict bcs class. curr. drug metab. 9 (2008) 869-878. [23] k. etherson, g. halbert, m. elliott. determination of excipient based solubility increases using the cheqsol method. int. j. pharm. 465 (2014) 202-209. [24] e. fornells, e. fuguet, m. mañéa, r. ruiz, k. box, e. bosch, c. ràfols. effect of vinylpyrrolidone polymers on the solubility and supersaturation of drugs; a study using the cheqsol method. eur. j. pharm. sci. 117 (2018) 227-235. [25] y.-l. hsieh, g.a. ilevbare, b. van eerdenbrugh, k.j. box, m.v. sanchez-felix, l.s. taylor. ph-induced precipitation behavior of weakly basic compounds: determination of extent and duration of supersaturation using potentiometric titration and correlation to solid state properties. pharm. res. 29 (2012) 2738-2753. [26] a. llinàs, j.c. burley, k.j. box, r.c. glen, j.m. goodman. diclofenac solubility: independent determination of the intrinsic solubility of three crystal forms. j. med. chem. 50 (2007) 979-983. [27] l.y.s. narasimham, v.d. barhate. kinetic and intrinsic solubility determination of some β-blockers and antidiabetics by potentiometry. j. pharmacy res. 4 (2011) 532-536. [28] d. schönherr, u. wollatz, d. haznar-garbacz, u. hanke, k.j. box, r. taylor, r. ruiz, s. beato, d. becker, w. weitschies. characterisation of selected active agents regarding pka values, solubility concentrations and ph profiles by siriust3. eur. j. pharm. biopharm. 92 (2015) 155-170. [29] c. sköld, s. winiwarter, j. wernevik, f. bergström, l. engström, r. allen, k. box, j. comer, j. mole, a. hallberg, h. lennernäs, t. lundstedt, a.-l. ungell, a. karlén. presentation of a structurally diverse and commercially available drug data set for correlation and benchmarking studies. j. med. chem. 49 (2006) 6660-6671. [30] j. comer, s. judge, d. matthews, l. towes, b. falcone, j. goodman, j. dearden. the intrinsic aqueous solubility of indomethacin. admet & dmpk 2 (2014) 18-32. [31] c.a.s. bergström, m strafford, l. lazarova, a. avdeef, k. luthman, p. artursson. absorption classification of oral drugs based on molecular surface properties. j. med. chem. 46 (2003) 558-570. [32] k.j. box, g. völgyi, e. baka, m. stuart, k. takács-novák, j.e.a. comer. equilibrium versus kinetic measurements of aqueous solubility, and the ability of compounds to supersaturate in solutiona validation study. j. pharm. sci. 95 (2006) 1298-1307. ©2019 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ intestinal absorption of bcs class ii drugs administered as nanoparticles: a review based on in vivo data from intestinal perfusion models doi: https://dx.doi.org/10.5599/admet.881 375 admet & dmpk 8(4) (2020) 375-390; doi: https://dx.doi.org/10.5599/admet.881 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index review intestinal absorption of bcs class ii drugs administered as nanoparticles: a review based on in vivo data from intestinal perfusion models david dahlgren, erik sjögren and hans lennernäs* department of pharmaceutical biosciences, translational drug discovery and development, uppsala university, sweden *corresponding author: e-mail: hans.lennernas@farmbio.uu.se; tel.: +46 – 18 471 4317; fax: +46 – 18 471 4223 received: june 26, 2020; revised: september 10, 2020; published: september 17, 2020 abstract an established pharmaceutical strategy to increase oral drug absorption of low solubility–high permeability drugs is to create nanoparticles of them. reducing the size of the solid-state particles increases their dissolution and transport rate across the mucus barrier and the aqueous boundary layer. suspensions of nanoparticles also sometimes behave differently than those of larger particles in the fed state. this review compares the absorption mechanisms of nanoand larger particles in the lumen at different prandial states, with an emphasis on data derived from in vivo models. four bsc class ii drugs— aprepitant, cyclosporine, danazol and fenofibrate—are discussed in detail based on information from preclinical intestinal perfusion models. ©2020 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords nano particle; nano drug delivery; nanomedicines; intestinal perfusion; pharmaceutical development. introduction rate and extent of drug absorption and bioavailability are critical pharmacokinetic (pk) parameters for oral pharmaceutical products. these parameters are determined and characterized in drug discovery and in preclinical and clinical development of a drug product. successful design and development of any oral drug product requires understanding the complex interplay between the pharmacokinetic parameters, manufacturing methods, pharmaceutical excipients, and biopharmaceutics, as well as potency (the major driver for drug discovery), pharmacology, safety, and toxicity. gastrointestinal (gi) absorption and bioavailability are both affected by a number of pharmaceutical, biopharmaceutical and physiological processes. despite comprehensive knowledge about drug-product related and physiological properties in gi absorption, many drug candidates and drugs have suboptimal biopharmaceutical properties for oral dosing [1,2]. several challenging drug candidates are listed in the biopharmaceutical classification system (bcs) as class ii, iii, or iv, for which absorption is limited by solubility properties in gi luminal fluids and/or intestinal permeability [3]. limitations in any one of the main bcs parameters may reduce the rate and fraction dose https://dx.doi.org/10.5599/admet.881 https://dx.doi.org/10.5599/admet.881 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:hans.lennernas@farmbio.uu.se mailto:hans.lennernas@farmbio.uu.se http://creativecommons.org/licenses/by/4.0/ h. lennernäs et al. admet & dmpk 8(4) (2020) 375-390 376 absorbed (fabs) from conventional, oral dosage forms intended for immediate release in the gi lumen. drugs with low solubility and/or low intestinal permeability also tend to have highly variable plasma exposure profiles within and between individuals [4,5]. in the design and development of oral drug dosage forms, a pk assessment is fundamental for distinguishing between fabs and bioavailability (f or ba) [6]. in general pharmacology, f represents the fraction of an administered dose of a drug that reaches the systemic circulation in intact form (parent drug) and it is one of the primary pk properties of any drug product. it is assumed that the bioavailability for any drug product given intravenously is 100 % and the bioavailability of any orally administered dosage form is measured relative to this value, calculated as the ratio of dose-normalized plasma exposure. many of the drug candidates currently generated in early development are lipophilic. this is probably a result of the drug discovery lead-finding techniques applied, which rely on in silico and in vitro screening of receptor-ligand interactions [7-9]. these molecular screening tools propose candidates with a high potential for interacting with designated target receptors. as these receptors are usually associated with lipophilic drugs, the tools invariably select lipophilic candidates [10-12]. while lipophilic active pharmaceutical ingredients (apis) tend to have more than sufficiently high effective permeability across the apical membrane of the enterocytes, they suffer from limited solubility and dissolution in the gi lumen [13]. the dissolution process of a solid api particle can briefly be described in two steps. first, the drug molecules are released from the particle surface to the surrounding dissolution media, which creates a saturated, stagnant layer adjacent to the solid surface of the particle. thereafter, the released drug diffuses into the bulk of the solvent from regions of high, to regions of low, drug concentration (schematically displayed in figure 1). figure 1. dissolution of a solid api releases the drug molecule, shown here for an unrestricted volume. the rate of drug dissolution (dm/dt) is traditionally described by the noyes-whitney/nernst-brunner equation 1 [14-18]: s( )d d td a c cm t h     (1) where dm/dt is the dissolution rate (i.e., the change in dissolved drug concentration with time), a is the surface area (4πr 2 ) of the particle available for dissolution, d is the diffusion rate constant, h is the thickness admet & dmpk 8(4) (2020) 375-390 intestinal absorption of bcs class ii drugs nanoparticles doi: https://dx.doi.org/10.5599/admet.881 377 of an aqueous boundary layer (abl) with limited convection surrounding the particle, cs is the saturation solubility of the drug at the particle-dissolution media interface, and ct the dissolved drug concentration in the bulk at time t. the equation assumes that the diffusion of an api monomer through the abl is the slowest (i.e., rate-limiting) step in this series of events. this assumption allows the incorporation of fick’s law of diffusion to quantify the dissolution rate of a solid in its own solution (figure 1) [19]. the noyes-whitney/nernst-brunner equation (equation 1) predicts the dissolution of particles greater than a few micrometres; however, this equation may not be optimal for nanoparticles [16]. in addition to increasing total surface area, it should also be mentioned that particle size reduction (down to about 1 µm in diameter) will affect the abl surrounding each particle, both according to the prandtl equation (equation 2) but also by a reduced influence of convection [20,21]. 3 ( ) l h k v  (2) in equation (2) h stands for the hydrodynamic boundary layer thickness, k is a constant, l is the length of the particle surface in the flow direction, and v is the relative velocity of the liquid surrounding the particle. theoretically, a reduced particle size also increases the saturation solubility, according to the kelvin effect [20]. the theorem originates from the description of vapour pressure over a droplet, which increases with droplet curvature. this theorem originates from the description of vapour pressure over a droplet, which increases with droplet curvature. the true implication of this effect is however difficult to assess. first, as the effect apply to particles below 10 nm only a fraction (<1 %) of the original mass of a conventional nanosized (~100 nm) system remains to be dissolved. secondly, in polydisperse particles systems, the gained dissolution of small particles may be countered by growth of larger particles, according to the ostwald ripening principles. dissolution is typically the key parameter for the assessment of the onset of action of an oral dosage form with bcs class ii drugs. the dissolution rate of any oral drug product is determined by the physicochemical properties of the substance, dosage form composition and functions, as well as by the physiological gi conditions (table 1). the latter varies within and between subjects as well as between the fasted and fed state. table 1. overview of the pharmaceutical, physicochemical and the physiological properties that influence intestinal drug dissolution. properties affected by physicochemical parameters pharmaceutical parameters physiological parameters drug particle surface area wettability, particle size, aggregation, surfactnts deaggregation by surfactants in gastric juice and intestinal fluid (from bile) drug diffusion molecular size (stokes–einstein equation) viscosity of gi luminal contents (fasted/fed) (stokes–einstein equation) diffusion layer surface energy particle size motility patterns and luminal flow rate drug solubility lipophilicity, pka, melting point crystal form, solubilisation, precipitation, pharmaceutical excipients, impurities luminal ph, buffer capacity, bile and food composition amount of drug dissolved see above see above intestinal permeability volume of solvent available gi secretion, fluid absorption, coadministered fluids https://dx.doi.org/10.5599/admet.881 h. lennernäs et al. admet & dmpk 8(4) (2020) 375-390 378 a range of formulation strategies to increase the dissolution rate have been invented and established for oral drug products with apis of low solubility [16]. for acidic and basic apis, a common approach is to form a salt with the charged form of the api. use of a counter-ion increases the solubility and dissolution rate in the diffusion layer around the solid drug particles [22]. three other strategies are to: (i) predissolve the drug in a lipid-based formulation for circumventing the slow dissolution step; (ii) change the solid state properties of the apis from stable crystalline state(s) to metastable amorphous structures with lower intermolecular cohesive forces; or (iii) reduce the particle size, which increases the surface area of the api and consequently the dissolution rate [23-25]. two established pharmaceutical strategies to reduce the size of the solid state particles are to create microor nanosuspensions of them [26]. particle size reduction (i.e., micronization) enhance the dissolution rate (equation 1) and may consequently increase the intestinal absorption and bioavailability of low solubility drug compounds. however, for bcs ii apis, nanosuspensions appear to offer additional advantages over microsuspensions; the enhancement of the in vivo dissolution rate required to increase the gi drug absorption and bioavailability for low solubility drug candidates is often greater than what a nanosuspension can achieve [27]. this is shown for a nanosuspension of a test drug with low molecular mass 450, acid pka 4.7, log p of 5 and a solubility of 2 μm at ph 6.8 [28]. interestingly, with these nanotechnologies for solid-state particles, it is apparent that additional intraluminal mechanism(s) than dissolution must be operating [29]. despite the advantages with nanoparticles, a number of hurdles need to be resolved with these oral drug delivery systems. for instance, instability related to aggregation of nanoparticles and/or to a change in the solid-state properties of the api may reduce their applicability in pharmaceutical development [30]. still, in oral product development, nanotechnologies have been successful to develop products to a clinical stage based on solubility, but mainly by enhancing dissolution. however, there is a need to fully understand the contribution from enhancement of solubility and/or dissolution as well as additional absorption mechanisms, such as particle drifting in the mucus layer [31]. there are even strategies in which nanoparticles may increase absorption of low permeability compounds through increased transepithelial transport [32]. obviously, more formulation research and development beyond the established pharmaceutical strategies are needed, to better understand the mechanisms, physiological possibilities, and limitations in the in vivo gi absorption of orally administered nanoparticles. this review focuses on nanoparticle-based oral drug delivery systems and their in vivo performance based on experimental studies in single-pass intestinal perfusion (spip) models. in particular, the intestinal absorption mechanisms will be discussed. these data may contribute to a future framework for enabling oral formulation strategies for mainly bcs class ii drugs. this review looks at the role of nanoparticles of four low solubility-high permeability drugs: aprepitant, fenofibrate, cyclosporine, and danazol. a summary of data from both fasted and fed simulated conditions is included, because prandial state has a strong effect on rate and extent of intestinal drug absorption for some drug products. prandial state also illustrates some of the different intraluminal dissolution and absorption mechanisms that come into play. intraluminal transport of nanoparticles and drugs in the intestine pharmaceutical development of oral drug delivery systems for low solubility drugs frequently uses particle size reduction as a formulation method. development of particle sizes in the nano-range (diameter <1000 nm) for oral products has increased as a result of improvements in pharmaceutical processing and manufacturing [33]. as discussed above, oral nanosuspensions dissolve faster than microsuspensions, and accordingly increase intestinal absorption and bioavailability of bcs ii compounds [28,34-37]. theoretically, admet & dmpk 8(4) (2020) 375-390 intestinal absorption of bcs class ii drugs nanoparticles doi: https://dx.doi.org/10.5599/admet.881 379 api particle size reduction reaches a point at which dissolution is no the longer the rate-limiting step in in vivo absorption. at that point, further size reduction is not expected to increase intestinal absorption as it has become solubility controlled [8]. however, experimental observations indicate otherwise. absorption rates continue to increase with decreasing particle size, which suggests that additional absorptionpromoting intestinal processes are taking place simultaneously [29,38,39]. it has been proposed that nanoparticles are crossing the epithelial mucus layer to a greater extent than larger particles [40]. mucus is predominately comprised of the glycoprotein mucin and water. this gives it hydrophilic and hydrophobic domains, a negative net charge, and high porosity and pore interconnectivity. in addition, mucus is a dynamic, semipermeable transport barrier that is continuously secreted, shed, and digested. mucus thickness is about 120 μm, 480 μm and 830 μm in the jejunum, ileum and colon, respectively, with a turnover time of a few hours [41,42]. the small intestinal mucus contains a single mucus layer that is not a rate-limiting step in the absorption, not even for lipophilic high-permeability drugs and/or high permeability drugs transported by an efficient carrier-mediated (cm) transport route [43-45]. based on in vivo jejunal perfusion at different rates in humans, the resistance of the mucus layer to the intestinal absorption of highly permeable solutes is markedly overestimated. instead, intestinal absorption in humans seems to be membrane controlled for both lowand high-permeability compounds, irrespective of transport mechanism [46]. it is well-recognized that nanoparticles interact with mucus by different mechanisms and that nanoparticles may acquire different biological and physicochemical properties as a consequence of adsorption of intraluminal gi biomolecules to their surface (i.e., corona formation). commercially available oral nanoparticles are typically around 100–300 nm, which should allow them to pass faster than larger api particles across the gel-like, mucin-based mesh structure composing the mucus layer. a continuous mucus layer exists throughout the small intestine, and is highly stratified adjacent to the epithelium. experiments show that mucin pore size is consistent with free diffusion of 100 nm particles, but limited diffusion of 500 nm particles [40,47]. this is due to increased diffusivity for a spherical particle with smaller radius, which then establishes a high local drug concentration available for permeation adjacent to the apical enterocyte membrane [31,48]. the primary mechanism of transport through this layer is diffusion, as opposed to the intestinal lumen, where the primary transport mechanism is convection [49]. the combined effects of increased dissolution, mucus penetration, and mucus layer diffusivity for small particles seem to explain the increased intestinal absorption of nanoparticles compared to larger ones. method of in vivo absorption: single-pass intestinal perfusion model and absorption calculations this review discusses mainly data from the rat and pig single-pass intestinal perfusion (spip) models in relation to other relevant in vivo data [50-53]. the spip model has been extensively used in preclinical studies to investigate and determine epithelial effective intestinal permeability (peff), as well as other absorption parameters for drugs and drug delivery systems, such as suspensions and lipid-based ones (figure 2). intestinal peff, absorption flux (japp), and fraction dose absorbed (fabs) are the three key biopharmaceutical variables that describe the absorption and transport properties of a drug across the intestinal barrier in the spip model [3,54]. this model is used to investigate oral drug delivery and characterize pre-formulations, as part of the development process of a pharmaceutical product. the spip model is also useful for investigating regional intestinal permeability and transport mechanisms involved during in vivo–relevant conditions, and for determining the bcs classification of any api [54]. this model is also used to investigate the interplay between in vivo dissolution of various dosage forms, such as nanoformulations and various pharmaceutical excipients, and intestinal permeability, during both fed and fasted conditions [55]. https://dx.doi.org/10.5599/admet.881 h. lennernäs et al. admet & dmpk 8(4) (2020) 375-390 380 figure 2. graphical illustration of a single-pass intestinal perfusion (spip) in rat with the intestinal segment placed on the outside of the abdomen. the different numbers show: 1) perfusion solution/suspension under constant stirring on a heating table, and perfused using a peristaltic pump; 2) a syringe pump allows coadministration of e.g., enzymes; 3) perfusate is collected after passing through an intestinal segment; 4) blood sampling from the femoral artery; 5) body temperature is monitored using a rectal probe, connected to a heating pad; 6) blood pressure and heart rate are monitored in the femoral artery. in the spip model, absorption parameters can be calculated in several ways. one of the most common is to calculate effective permeability (peff), which is based determining the disappearance of an api from the intestinal lumen during single-pass perfusion of a well-defined intestinal segment (equation 3):  in out inln / eff q c c p a   (3) where qin is the perfusate flow rate, cout (corrected for water flux) and cin are the concentrations of api leaving and entering the intestinal segment, and a is the surface area of the perfused intestinal segment, which is assumed to be a smooth cylinder [56]. the cout is corrected for water flux using an unabsorbable marker and by weighing the luminal perfusates leaving the segment [57]. the disappearance flux (jdisapp) of an api can also be calculated from perfusate data leaving the intestinal segment by correcting for water flux as previously described (equation 4):  in in out disapp q c c j a   (4) alternatively, absorption flux (japp) can be calculated in the spip model by monitoring plasma drug appearance using the deconvolution method (equation 5). however, this requires a separate intravenous injection of the study drug for calculating the disposition pk parameters in the deconvolution. app input rate j a  (5) the pig data in this review are based on the same perfusion setup as the loc-i-gut model, for which it has been clearly shown that a well-stirred model describes the hydrodynamics during single-pass experiment [56,58]. the following calculations from the perfusion experiments in the pig spip model were made from steady-state concentrations of the outlet jejunal perfusate [55]. the fraction of the drug absorbed in the segment during the perfusion (fabs) was calculated from equation 6: admet & dmpk 8(4) (2020) 375-390 intestinal absorption of bcs class ii drugs nanoparticles doi: https://dx.doi.org/10.5599/admet.881 381 in abs out 1 out in c peg f c peg        (6) where cin and cout are the concentrations of the study drug, and pegin and pegout are the non-absorbable volume correction marker ( [ 14 c]-peg 4000), entering and leaving the jejunal segment, respectively. the jejunal peff of each drug is calculated according to a well-mixed tank model, as shown in equation 7 [59]: in outin eff in a c cq p c        (7) where a is the cylindrical area of the perfused jejunal segment (a). in vivo absorption of four model drugs, under different luminal conditions this review focuses on the in vivo, small intestine absorption of four bcs class ii drugs: aprepitant, cyclosporine, danazol, and fenofibrate. a range of physicochemical properties of these drugs are presented in table 2. they have been investigated in the small intestine as nano-formulations under different luminal conditions. table 2. physicochemical descriptors of danazol, cyclosporine, aprepitant, and fenofibrate. danazol [55,60] cyclosporine [55] aprepitant [60] fenofibrate [60] molecular mass (g/mol) 337 1202 535 361 water solubility at 37 °c (µg/ml) 0.5 7 0.37 0.25 log p 3.7 3 4.7 6.9 papp (·10 -6 cm/s) 14.15 2.6 170 220 aprepitant is an orally administered neurokinin nk-1 receptor antagonist, used clinically to prevent acute and delayed chemotherapy-induced nausea and vomiting [61]. the molecular mass is 534 da. it has a basic pka of 2.4, an acidic pka of 9.2, and is primarily uncharged at a jejunal ph of 6.5. the aqueous solubility is poor (0.37 μg/ml) and it has a high apparent permeability (papp) in caco-2 cells (1.7·10 −4 cm/s), which predicts a high peff in both small and large intestine [62]. in fact, the rat small intestinal peff is reported to be 1.7·10 −4 cm/s [63]. aprepitant is mainly metabolized by gut and hepatic cyp3a4, which means that plasma pharmacokinetics (i.e. f) can be affected by differences in small intestinal absorption rate [64]. the oral product with aprepitant (emend; merck & co., inc., nj) is based on nanoparticles of the api and has an average particle diameter below 200 nm. these particles are coated onto larger cellulose beads and encapsulated [65]. cyclosporine is an immunosuppressant used to prevent organ transplant rejection and danazol is a steroid used to treat endometriosis. as evident by their bcs class ii classification, both cyclosporine and danazol have high intestinal permeability values and low aqueous solubility. both cyclosporine and danazol are metabolized in the small intestine and liver by cyp3a4, where the gut wall and liver first-pass extraction might be affected by differences in small intestinal absorption [66-68]. the size of the nanoparticles of cyclosporine and danazol used in the experiments discussed below was 650 and 150 nm, respectively. to ensure that the nanosuspension was homogenous and stabilized, cyclosporine and danazol were added to an aqueous solution containing small amounts of water-soluble polymer and surfactant as stabilizers/dispersants (peg-4000 0.13 mg/ml, monoolein 4 mg/ml, sodium taurocholate 20 mg/ml, oleic acid 20 mg/ml, and phosphatidylcholine 6 mg/ml). both model drugs were single-passed https://dx.doi.org/10.5599/admet.881 h. lennernäs et al. admet & dmpk 8(4) (2020) 375-390 382 perfused through pig jejunum in isotonic fluid alone with and without a p-gp inhibitor, and with dietary and endogenous lipids. to determine the effect of food on the in vivo dissolution of cyclosporine and danazol, saturated drug solutions in isotonic fluid containing lipids were also perfused in pig. fenofibrate is mainly used for primary hypercholesterolemia or mixed dyslipidemia. fenofibrate is a low molecular mass (360.4 da) bcs class ii drug that is very lipophilic (log p = 5.24), with an aqueous solubility of <0.1 mg/ml. fenofibrate is an ester of fenofibric acid. after oral administration in humans, it is completely converted to its active metabolite, fenofibric acid. fenofibric acid and fenofibric acid glucuronide are then excreted into urine (60 % of the dose) and feces (25 %) [69]. comparison between microand nanosuspensions of bcs class ii drugs experimental data and simulations show that small intestinal absorption of larger particles of danazol (226 µm), griseofulvin (118 µm), and aprepitant (26 µm), is dissolution rate-controlled [70]. takano et al. also show that reduction of particle size increases in vitro dissolution and improves intestinal absorption in dogs. increasing the dissolution rate by using of smaller particles of these three selected model drugs does not improve their small intestinal absorption because their absorption is limited by their solubility during non-sink luminal conditions [70]. however, for some drug formulated as nanoparticles in the range of 50– 200 nm, the absorption and plasma exposure increased. the increased absorption rate is often explained by enhanced drug dissolution rate [28,34,71]. these differences in experimental in vivo data and their interpretation will be discussed below. the intestinal absorption rate of aprepitant increases by increasing the concentration of nanoparticles (200 µm versus 20 µm) in the perfusion suspension passed along the small intestine segment in the rat spip model [5]. aprepitant permeates the apical membrane as a monomer very swiftly and has a small intestinal peff of 1.7 · 10 -4 cm/s. both the drifting and deposition of nanoparticles with aprepitant in the abl, as well as colloidal structures, may contribute to an increase in intestinal absorption. absorption of nanoand microsuspensions has been compared in the rat spip model when the small intestinal segment was perfused with buffer, fassif, and fessif. in with single-pass perfusions of jejunum, the plasma concentration time–curves clearly show that fed conditions, i.e. fessif, increase the intestinal absorption of aprepitant as microsuspensions, but not as nanosuspensions. this difference is in line with previous data from both dog and human, in which nanosuspensions seem to prevent variation in plasma exposure between different prandial states, while microsuspensions show significant food-effects [72-75]. the mucus layer also influences absorption of drugs delivered orally as nanoparticles [76]. for example, the hydrophobicity, electrostatic properties, and steric hindrance, of mucus are key feature that prevents the hydrophilic pancreatic proteases from acting on, and thereby injuring, the intestinal epithelium [77]. to quantify the jejunal flux of aprepitant across the enterocytes, japp was calculated according to eq. 4 (table 3). there were no significant differences for microand nanosuspensions, but a large, albeit non-significant one, for the microsuspensions in fessif compared to buffer and fassif (64 and 4.8 times higher, respectively). for the nanosuspensions, the japp were all within a 1.6-fold range, reiterating the effective elimination of a luminal food effect by the nanoformulations. the high variability of the data in this paper demonstrates that studies with more animals are needed to verify these observed trends for nanoand microsuspensions (table 3). admet & dmpk 8(4) (2020) 375-390 intestinal absorption of bcs class ii drugs nanoparticles doi: https://dx.doi.org/10.5599/admet.881 383 table 3. historical in vivo absorption flux (japp) (sd) for aprepitant nanoand microsuspensions determined in the jejunum of the single-pass perfused rat reported by roos et al. 2018 [78]. media flux (× 10 -3 µmol / (hr × cm 2 )) nanosuspension microsuspension buffer 2.91 (1.02) 0.104 (0.081) fassif 3.57 (1.97) 1.58 (2.10) fessif 2.25 (1.69) 6.63 (7.32) in vivo absorption of bcs class ii drugs in nanoparticles at different luminal conditions danazol and cyclosporine intestinal absorption mechanisms have been investigated in fed and fasted states in a pig spip model for low solubility apis as monomers or nanoparticles. nanoparticle suspensions of two bcs class ii model drugs, danazol and cyclosporine (cya), were single-pass perfused through the pig jejunum in isotonic buffer alone (control) and isotonic buffer with verapamil, a p-glycoprotein (p-gp) inhibitor, or in isotonic buffer containing dietary and intestinal lipids. co-administration of cya and verapamil increased the jejunal peff of cya from 0.63 ± 0.05 · 10 -4 cm/s to 1.01 ± 0.09 · 10 -4 cm/s [55]. the human jejunal peff for cya was 1.65 ± 0.53 · 10 -4 cm/s, when determined with the perfusion method, which classifies cya as a low solubility-high permeability class ii bcs drug. cya is well-recognized to have highly variable absorption from approved oral formulations because of poor dissolution [54,79,80]. when dietary lipids were introduced to the perfused segment in the spip model, they reduced the jejunal absorption of a cya as a saturated solution and nanosuspension by approximately 50 % and 83 %, respectively (table 4). in contrast, dietary lipids increased the jejunal absorption of danazol nanosuspensions more than two-fold. however, fabs was reduced by 60 % when danazol was administered as a solution in the same media (table 4). table 4. mean ( s.d) of effective jejunal permeability (peff) and fraction absorbed (fabs) for cyclosporine and danazol during a single-pass perfusion of nanoparticles, nanoparticles in the fed state, and as a saturated solution. parameter nanoparticles nanoparticles – fed state saturated solution cyclosporine peff (10 -4 cm/s) 0.63  0.05 0.11  0.14 0.34  0.08 fabs (%) 29  2 5  6 13  8 danazol peff (10 -4 cm/s) 0.9  0.4 2.3  0.4 0.33  0.09 fabs (%) 36  10 58  4 17  4 this difference for nanosuspensions and monomer solutions of danazol is most likely due to increased dissolution and/or a more efficient transport of the danazol nanoparticles (150 nm) than the cya (650 nm) through the particle abl. the difference in the effect of intraluminal lipids may also be explained by different distribution of any of the study drugs to the colloidal structures present in the media. these structures affect the convective and diffusion rates of the drugs differently in the perfused segment. cya has a higher partitioning (2.5 times) to the colloidal structures than danazol. based on previous observations, danazol probably partitions primarily to the mixed intraluminal micelles, and to a lesser extent to the lipid vesicles [55]. despite a lower log p value, cya most likely partitions into larger lipid vesicles to a greater extent than danazol. it has a larger molecular size, which results in a reduced thermodynamic activity (i.e., lower free concentration available for permeation). furthermore, there would be an overall slower diffusion and convection in the intestinal lumen. another plausible mechanism is that the slow and incomplete https://dx.doi.org/10.5599/admet.881 h. lennernäs et al. admet & dmpk 8(4) (2020) 375-390 384 partitioning from the vesicles to the aqueous-based chime leads to a subsequent rapid intestinal absorption of cya. a temporarily unsaturated aqueous phase might form along the perfused jejunal segment and reduce thermodynamic activity. it is also clear that drug-solubilization is a more important factor in fed state for cya than p-gp inhibition for food–drug interaction. drug partitioning into different luminal cs, and their fate in the lumen and across the abl, are important processes to consider in designing nano-based oral formulations of poorly soluble drugs. both the peff and fabs were higher for danazol when administered as a nanosuspension in lipid-containing media than in the control buffer. however, due to the rapid dissolution of the nanoparticles, the difference in absorption did not correspond to previously observed increases in bioavailability of danazol administered with food. administration of danazol in the same media, but as a solution, decreased both peff and fabs. this shows the importance of the rapid dissolution of the drug nanoparticles. when danazol is administered as a solution, thermodynamic activity may decrease because absorption of the drug might create a state of undersaturation. aprepitant the solubility of aprepitant in biorelevant media, e.g., fasted-state simulated intestinal fluid (fassif), is approximately 60 times higher than in buffer alone [60]. this indicates that aprepitant readily partitions to luminal colloidal structures. the influence of these colloidal structures on the in vivo absorption of aprepitant was investigated in the rat spip model. when colloidal structures were added in the perfusion medium, the absorption was indeed affected as the observed small intestinal absorption rate increased. however, the experimentally determined increase in small intestinal absorption rate cannot be explained only by an increased luminal dissolution rate. rather, the increased absorption rate may be a consequence of changes to the total effective diffusion from the bulk to the epithelial membrane of aprepitant, i.e., transport as free monomers and monomers partitioned to colloidal structures, and as particles [63]. total effective diffusion may be an appropriate parameter to describe these experimental observations and to translate the spip data on nanoparticle formulations to in vivo performance. these specific data have been used to evaluate a mathematical model simulating the absorption-promoting effects of nanoparticle formulations. this model included interlinked descriptions of hydrodynamics, particle dissolution, and particle diffusion for both the api and luminal colloidal structures. this mathematical, mechanistic model adequately describes the above discussed in vivo absorption data of nano-formulated aprepitant. it also supports the proposed mechanism by which contributing effects of the diffusion across the mucus layer of both aprepitant nanoparticles and colloidal structures into which the drug had partitioned [81]. this exemplifies a model in which representation of physiology and description of mechanistic formulationphysiology relationships was pivotal to establish the role of formulation for effective drug absorption. the need of such physiologically based biopharmaceutics models (pbbm) for predictions and assessments of formulations in vivo performance has recently been highlighted with an increased scientific activity in the area [1,82]. fenofibrate and other drugs in contrast to these absorption-promoting results for nanoparticles aprepitant and danazol in both fasted and fed conditions in the spip model, only the fed state reduces the intestinal absorption of nanosized drugs such as fenofibrate, csa, pafenolol [37,55,83,84]. in a regional absorption study in humans with a site-specific delivery system (enterion capsule), fenofibrate were given as single, equimolar doses into the stomach, proximal small bowel, distal small bowel, and colon. the bioavailability (related to an intravenous dose of fenofibric acid) in the stomach, proximal small bowel, distal small bowel, and colon was 69 %, 73 %, admet & dmpk 8(4) (2020) 375-390 intestinal absorption of bcs class ii drugs nanoparticles doi: https://dx.doi.org/10.5599/admet.881 385 66 %, and 22 %, respectively [85]. this clearly shows that fenofibrate is absorbed from the small intestine. however, the solubility and dissolution conditions in the colon—less fluid and limited amounts of bile acids and colloidal structures—prevent efficient fenofibrate absorption from this region. an in vivo study in rats compared in vitro dissolution, solubility and bioavailability of three different oral nanosystems with fenofibrate powder. the three systems included pvp nanospheres, hp-β-cd nanocapsules, and gelatin nanocapsules [86]. the most improved apparent solubility and oral bioavailability in fasted rats was with fenofibrate nanoencapsualted with gelatin at a ratio of 1:8 (w/w). the authors of the study proposed that the improved dissolution rate and subsequent intestinal absorption was due to: (i) a higher solubility of fenofibrate as it converted into the amorphous form or nanocrystalline state; (ii) a large surface area for dissolution of drug; (iii) improved wetting of fenofibrate by the polymeric matrix in the formulation; and (iv) reduction in the crystalline intensity. a method to investigate in vivo-relevant mechanisms of drug absorption in humans is by gi fluid sampling [52,87]. gi intubation can explore directly the dynamic interplay of drug release, dissolution, precipitation, and absorption of the drug from a dosage form. the dissolved drug and solid-state concentration–time profiles from different segments in the gi tract can be obtained. in addition, direct infusion of a suspension into the duodenum allows patient control of the therapeutic system [88,89]. one study compared nanoand microsuspensions of fenofibrate administered orally in fed state in humans. the nanosuspensions gave higher local concentrations than the microsuspensions in the proximal small intestinal lumen and resulted in higher absorption and plasma exposure–time profile of the drug [37]. in fed conditions, the duodenal concentrations of fenofibrate were higher for both oral formulations, but there was no increase in plasma exposure of fenofibric acid. micellar encapsulation of the fenofibrate in the lumen may have limited potential to permeate from colloidal structures in fed-state intestinal fluids. absorption can also decrease in the presence of micelle-forming lipids due to lower thermodynamic activity [39]. the absorption of bcs class ii drugs does not necessarily increase when administered with food, as proposed by others [90]. taken together, these in vivo data demonstrate that increasing the intestinal luminal concentration of bcs class ii drugs in complex intestinal fluids does not always result in more rapid and extensive intestinal absorption. especially in fed state, the encapsulation of a lipophilic drug in micelles and vesicles in the intestinal lumen may reduce the potential for intestinal flux. conclusion the use of oral, nanosystems for drug delivery can increase intestinal absorption and bioavailability, reduce the risk for food–drug interactions, and ph-dependent intestinal absorption. however, for some drugs, the interaction with colloidal structures in the lumen may prevent absorption even if the oral nanosystems enhance dissolution in vitro compared to the standard oral formulation. nevertheless, nanoparticle formulations offer an important strategy in the development of new drugs and for currently difficult poorly soluble drugs. conflict of interest: conflict of interest does not exist. references [1] b. abrahamsson, m. mcallister, p. augustijns, p. zane, j. butler, r. holm, p. langguth, a. lindahl, a. müllertz, x. pepin. six years of progress in the oral biopharmaceutics area–a summary from the imi orbito project. european journal of pharmaceutics and biopharmaceutics (2020). https://dx.doi.org/10.5599/admet.881 h. lennernäs et al. admet & dmpk 8(4) (2020) 375-390 386 [2] e. sjögren, b. abrahamsson, p. augustijns, d. becker, m.b. bolger, m. brewster, j. brouwers, t. flanagan, m. harwood, c. heinen, r. holm, h. juretschke, m. kubbinga, a. lindahl, v. lukacova, u. münster, s. neuhoff, m. nguyen, a. peer, c. reppas, a. hodjegan, c. tannergren, w. weitschies, c. wilson, p. zane, h. lennernäs, p. langguth. in vivo methods for drug absorption–comparative physiologies, model selection, correlations with in vitro methods (ivivc), and applications for formulation/api/excipient characterization including food effects. european journal of pharmaceutical sciences 57 (2014) 99-151. [3] g.l. amidon, h. lennernas, v.p. shah, j.r. crison. a theoretical basis for a biopharmaceutic drug classification: the correlation of in vitro drug product dissolution and in vivo bioavailability. pharmaceutical research 12 (1995) 413-420. [4] e.t. hellriegel, t.d. bjornsson. interpatient variability in bioavailability is related to the extent of absorption: implications for bioavailability and bioequivalence studies. clinical pharmacology and therapeutics 60 (1996) 601-607. [5] m. sugihara, s. takeuchi, m. sugita, k. higaki, m. kataoka, s. yamashita. analysis of intra-and intersubject variability in oral drug absorption in human bioequivalence studies of 113 generic products. molecular pharmaceutics 12 (2015) 4405-4413. [6] c.y. wu, l.z. benet, m.f. hebert, s.k. gupta, m. rowland, d.y. gomez, v.j. wacher. differentiation of absorption and first‐pass gut and hepatic metabolism in humans: studies with cyclosporine. clinical pharmacology and therapeutics 58 (1995) 492-497. [7] h.h. refsgaard, b.f. jensen, p.b. brockhoff, s.b. padkjær, m. guldbrandt, m.s. christensen. in silico prediction of membrane permeability from calculated molecular parameters. j med chem 48 (2005) 805-811. [8] e. sjögren, j. westergren, i. grant, g. hanisch, l. lindfors, h. lennernäs, b. abrahamsson, c. tannergren. in silico predictions of gastrointestinal drug absorption in pharmaceutical product development: application of the mechanistic absorption model gi-sim. european journal of pharmaceutical sciences 49 (2013) 679-698. [9] h.d. williams, n.l. trevaskis, s.a. charman, r.m. shanker, w.n. charman, c.w. pouton, c.j. porter. strategies to address low drug solubility in discovery and development. pharmacological reviews 65 (2013) 315-499. [10] c.a. lipinski. drug-like properties and the causes of poor solubility and poor permeability. journal of pharmacological and toxicological methods 44 (2000) 235-249. [11] p.w. kenny. the nature of ligand efficiency. journal of cheminformatics 11 (2019) 1-18. [12] m.m. hann, a.r. leach, g. harper. molecular complexity and its impact on the probability of finding leads for drug discovery. journal of chemical information and computer sciences 41 (2001) 856-864. [13] c. lipinski. poor aqueous solubility—an industry wide problem in drug discovery. am pharm rev 5 (2002) 82-85. [14] a.a. noyes, w.r. whitney. the rate of solution of solid substances in their own solutions. journal of the american chemical society 19 (1897) 930-934. [15] e. brunner. reaktionsgeschwindigkeit in heterogenen systemen. zeitschrift für physikalische chemie 47 (1904) 56-102. [16] d.a. shah, s.b. murdande, r.h. dave. a review: pharmaceutical and pharmacokinetic aspect of nanocrystalline suspensions. journal of pharmaceutical sciences 105 (2016) 10-24. [17] j.b. dressman, g.l. amidon, c. reppas, v.p. shah. dissolution testing as a prognostic tool for oral drug absorption: immediate release dosage forms. pharmaceutical research 15 (1998) 11-22. [18] b. shekunov, e.r. montgomery. theoretical analysis of drug dissolution: i. solubility and intrinsic dissolution rate. journal of pharmaceutical sciences 105 (2016) 2685-2697. [19] j. siepmann, f. siepmann. mathematical modeling of drug dissolution. international journal of pharmaceutics 453 (2013) 12-24. [20] v.b. junyaprasert, b. morakul. nanocrystals for enhancement of oral bioavailability of poorly watersoluble drugs. asian journal of pharmaceutical sciences 10 (2015) 13-23. admet & dmpk 8(4) (2020) 375-390 intestinal absorption of bcs class ii drugs nanoparticles doi: https://dx.doi.org/10.5599/admet.881 387 [21] r.h. müller, k. peters. nanosuspensions for the formulation of poorly soluble drugs: i. preparation by a size-reduction technique. international journal of pharmaceutics 160 (1998) 229-237. [22] a.t. serajuddin. salt formation to improve drug solubility. adv drug deliv rev 59 (2007) 603-616. [23] p. kanaujia, p. poovizhi, w. ng, r. tan. amorphous formulations for dissolution and bioavailability enhancement of poorly soluble apis. powder technology 285 (2015) 2-15. [24] g.g. liversidge, k.c. cundy. particle size reduction for improvement of oral bioavailability of hydrophobic drugs: i. absolute oral bioavailability of nanocrystalline danazol in beagle dogs. international journal of pharmaceutics 125 (1995) 91-97. [25] l.c. alskär, c.j. porter, c.a. bergström. tools for early prediction of drug loading in lipid-based formulations. molecular pharmaceutics 13 (2016) 251-261. [26] b. van eerdenbrugh, g. van den mooter, p. augustijns. top-down production of drug nanocrystals: nanosuspension stabilization, miniaturization and transformation into solid products. international journal of pharmaceutics 364 (2008) 64-75. [27] e.r. cooper. nanoparticles: a personal experience for formulating poorly water soluble drugs. journal of controlled release 141 (2010) 300-302. [28] k. sigfridsson, a.j. lundqvist, m. strimfors. particle size reduction for improvement of oral absorption of the poorly soluble drug ug558 in rats during early development. drug development and industrial pharmacy 35 (2009) 1479-1486. [29] k. sugano. possible reduction of effective thickness of intestinal unstirred water layer by particle drifting effect. int j pharm 387 (2010) 103-109. [30] s.b. murdande, d.a. shah, r.h. dave. impact of nanosizing on solubility and dissolution rate of poorly soluble pharmaceuticals. journal of pharmaceutical sciences 104 (2015) 2094-2102. [31] f. kesisoglou, m. wang, k. galipeau, p. harmon, g. okoh, w. xu. effect of amorphous nanoparticle size on bioavailability of anacetrapib in dogs. journal of pharmaceutical sciences 108 (2019) 29172925. [32] d. babadi, s. dadashzadeh, m. osouli, m.s. daryabari, a. haeri. nanoformulation strategies for improving intestinal permeability of drugs: a more precise look at permeability assessment methods and pharmacokinetic properties changes. journal of controlled release (2020). [33] s. colombo, m. brisander, j. haglöf, p. sjövall, p. andersson, j. østergaard, m. malmsten. matrix effects in nilotinib formulations with ph-responsive polymer produced by carbon dioxide-mediated precipitation. international journal of pharmaceutics 494 (2015) 205-217. [34] j.-i. jinno, n. kamada, m. miyake, k. yamada, t. mukai, m. odomi, h. toguchi, g.g. liversidge, k. higaki, t. kimura. effect of particle size reduction on dissolution and oral absorption of a poorly water-soluble drug, cilostazol, in beagle dogs. journal of controlled release 111 (2006) 56-64. [35] e. merisko-liversidge, g.g. liversidge, e.r. cooper. nanosizing: a formulation approach for poorlywater-soluble compounds. european journal of pharmaceutical sciences 18 (2003) 113-120. [36] r. müller, c. jacobs, o. kayser. nanosuspensions as particulate drug formulations in therapy: rationale for development and what we can expect for the future. adv drug deliv rev 47 (2001) 3-19. [37] b. hens, j. brouwers, m. corsetti, p. augustijns. gastrointestinal behavior of nano-and microsized fenofibrate: in vivo evaluation in man and in vitro simulation by assessment of the permeation potential. european journal of pharmaceutical sciences 77 (2015) 40-47. [38] k. sugano. aqueous boundary layers related to oral absorption of a drug: from dissolution of a drug to carrier mediated transport and intestinal wall metabolism. molecular pharmaceutics 7 (2010) 1362-1373. [39] g.e. amidon, w.i. higuchi, n.f. ho. theoretical and experimental studies of transport of micelle‐ solubilized solutes. j pharm sci 71 (1982) 77-84. [40] s.k. lai, d.e. o'hanlon, s. harrold, s.t. man, y.-y. wang, r. cone, j. hanes. rapid transport of large polymeric nanoparticles in fresh undiluted human mucus. proceedings of the national academy of sciences 104 (2007) 1482-1487. https://dx.doi.org/10.5599/admet.881 h. lennernäs et al. admet & dmpk 8(4) (2020) 375-390 388 [41] k. netsomboon, a. bernkop-schnürch. mucoadhesive vs. mucopenetrating particulate drug delivery. european journal of pharmaceutics and biopharmaceutics 98 (2016) 76-89. [42] j. leal, h.d. smyth, d. ghosh. physicochemical properties of mucus and their impact on transmucosal drug delivery. international journal of pharmaceutics 532 (2017) 555-572. [43] m.d. levitt, a. strocchi, d.g. levitt. human jejunal unstirred layer: evidence for extremely efficient luminal stirring. american journal of physiology-gastrointestinal and liver physiology 262 (1992) g593-g596. [44] m.d. levitt, j.k. furne. shaking of the intact angulation diminish layer. gastroenterology 103 (1992) 1460-1466. [45] d. dahlgren, m. sjöblom, h. lennernäs. intestinal absorption-modifying excipients: a current update on preclinical in vivo evaluations. european journal of pharmaceutics and biopharmaceutics 142 (2019) 411-420. [46] u. fagerholm, h. lennernäs. experimental estimation of the effective unstirred water layer thickness in the human jejunum, and its importance in oral drug absorption. european journal of pharmaceutical sciences 3 (1995) 247-253. [47] b.h. bajka, n.m. rigby, k.l. cross, a. macierzanka, a.r. mackie. the influence of small intestinal mucus structure on particle transport ex vivo. colloids and surfaces b: biointerfaces 135 (2015) 7380. [48] a. einstein. on the motion of small particles suspended in liquids at rest required by the molecularkinetic theory of heat. annalen der physik 17 (1905) 549-560. [49] y. wang, j.g. brasseur. three-dimensional mechanisms of macro-to-micro-scale transport and absorption enhancement by gut villi motions. physical review e 95 (2017) 062412. [50] d. dahlgren, c. roos, p. johansson, c. tannergren, a. lundqvist, p. langguth, m. sjöblom, e. sjögren, h. lennernas. the effects of three absorption-modifying critical excipients on the in vivo intestinal absorption of six model compounds in rats and dogs. international journal of pharmaceutics 547 (2018) 158-168. [51] e. sjödin, h. fritsch, u.g. eriksson, u. logren, a. nordgren, p. forsell, l. knutson, h. lennernäs. intestinal and hepatobiliary transport of ximelagatran and its metabolites in pigs. drug metabolism and disposition 36 (2008) 1519-1528. [52] l. bønløkke, l. hovgaard, h.g. kristensen, l. knutson, h. lennernäs. direct estimation of the in vivo dissolution of spironolactone, in two particle size ranges, using the single-pass perfusion technique (loc-i-gut®) in humans. european journal of pharmaceutical sciences 12 (2001) 239-250. [53] l. bønløkke, f.n. christensen, l. knutson, h.g. kristensen, h. lennernäs. a new approach for direct in vivo dissolution studies of poorly soluble drugs. pharmaceutical research 14 (1997) 1490-1492. [54] d. dahlgren, c. roos, e. sjögren, h. lennernäs. direct in vivo human intestinal permeability (peff) determined with different clinical perfusion and intubation methods. journal of pharmaceutical sciences 104 (2014) 2702-2726. [55] e.m. persson, a. nordgren, p. forsell, l. knutson, c. öhgren, s. forssén, h. lennernäs, b. abrahamsson. improved understanding of the effect of food on drug absorption and bioavailability for lipophilic compounds using an intestinal pig perfusion model. european journal of pharmaceutical sciences 34 (2008) 22-29. [56] h. lennernäs, ö. ahrenstedt, r. hällgren, l. knutson, m. ryde, l.k. paalzow. regional jejunal perfusion, a new in vivo approach to study oral drug absorption in man. pharmaceutical research 9 (1992) 1243-1251. [57] s.c. sutton, m. rinaldi, k. vukovinsky. comparison of the gravimetric, phenol red, and 14 c-peg-3350 methods to determine water absorption in the rat single-pass intestinal perfusion model. aaps pharmsci 3 (2001) 93. [58] h. lennernäs, i.d. lee, u. fagerholm, g.l. amidon. a residence‐time distribution analysis of the hydrodynamics within the intestine in man during a regional single‐pass perfusion with loc‐i‐gut: in‐ vivo permeability estimation. journal of pharmacy and pharmacology 49 (1997) 682-686. admet & dmpk 8(4) (2020) 375-390 intestinal absorption of bcs class ii drugs nanoparticles doi: https://dx.doi.org/10.5599/admet.881 389 [59] h. lennernas. human jejunal effective permeability and its correlation with preclinical drug absorption models. journal of pharmacy and pharmacology 49 (1997) 627-638. [60] e. sjögren, j. westergren, i. grant, g. hanisch, l. lindfors, h. lennernäs, b. abrahamsson, c. tannergren. in silico predictions of gastrointestinal drug absorption in pharmaceutical product development: application of the mechanistic absorption model gi-sim. european journal of pharmaceutical sciences (2013). [61] t.m. dando, c.m. perry. aprepitant. drugs 64 (2004) 777-794. [62] c. tannergren, a. bergendal, h. lennernäs, b. abrahamsson. toward an increased understanding of the barriers to colonic drug absorption in humans: implications for early controlled release candidate assessment. molecular pharmaceutics 6 (2009) 60-73. [63] c. roos, d. dahlgren, s. berg, j. westergren, b. abrahamsson, c. tannergren, e. sjögren, h. lennernäs. in vivo mechanisms of intestinal drug absorption from aprepitant nanoformulations. molecular pharmaceutics 14 (2017) 4233-4242. [64] r.i. sanchez, r.w. wang, d.j. newton, r. bakhtiar, p. lu, s.-h.l. chiu, d.c. evans, s.-e.w. huskey. cytochrome p450 3a4 is the major enzyme involved in the metabolism of the substance p receptor antagonist aprepitant. drug metabolism and disposition 32 (2004) 1287-1292. [65] r. hargreaves, j.c.a. ferreira, d. hughes, j. brands, j. hale, b. mattson, s. mills. development of aprepitant, the first neurokinin‐1 receptor antagonist for the prevention of chemotherapy‐induced nausea and vomiting. annals of the new york academy of sciences 1222 (2011) 40-48. [66] p.f. augustijns, t.p. bradshaw, l.-s.l. gan, r.w. hendren, d.r. thakker. evidence for a polarized efflux system in caco-2 cells capable of modulating cyclosporine a transport. biochemical and biophysical research communications 197 (1993) 360-365. [67] t. saeki, k. ueda, y. tanigawara, r. hori, t. komano. human p-glycoprotein transports cyclosporin a and fk506. journal of biological chemistry 268 (1993) 6077-6080. [68] t. kronbach, v. fischer, u.a. meyer. cyclosporine metabolism in human liver: identification of a cytochrome p‐450iii gene family as the major cyclosporine‐metabolizing enzyme explains interactions of cyclosporine with other drugs. clinical pharmacology and therapeutics 43 (1988) 630635. [69] a. munoz, j. guichard, p. reginault. micronised fenofibrate. atherosclerosis 110 (1994) s45-s48. [70] r. takano, k. furumoto, k. shiraki, n. takata, y. hayashi, y. aso, s. yamashita. rate-limiting steps of oral absorption for poorly water-soluble drugs in dogs; prediction from a miniscale dissolution test and a physiologically-based computer simulation. pharmaceutical research 25 (2008) 2334-2344. [71] l. jia, h. wong, c. cerna, s.d. weitman. effect of nanonization on absorption of 301029: ex vivo and in vivo pharmacokinetic correlations determined by liquid chromatography/mass spectrometry. pharmaceutical research 19 (2002) 1091-1096. [72] y. shono, e. jantratid, f. kesisoglou, c. reppas, j.b. dressman. forecasting in vivo oral absorption and food effect of micronized and nanosized aprepitant formulations in humans. european journal of pharmaceutics and biopharmaceutics 76 (2010) 95-104. [73] y. wu, a. loper, e. landis, l. hettrick, l. novak, k. lynn, c. chen, k. thompson, r. higgins, u. batra. the role of biopharmaceutics in the development of a clinical nanoparticle formulation of mk-0869: a beagle dog model predicts improved bioavailability and diminished food effect on absorption in human. int j pharm 285 (2004) 135-146. [74] a.k. majumdar, l. howard, m.r. goldberg, l. hickey, m. constanzer, p.l. rothenberg, t.m. crumley, d. panebianco, t.e. bradstreet, a.j. bergman. pharmacokinetics of aprepitant after single and multiple oral doses in healthy volunteers. the journal of clinical pharmacology 46 (2006) 291-300. [75] f. kesisoglou, s. panmai, y. wu. nanosizing—oral formulation development and biopharmaceutical evaluation. adv drug deliv rev 59 (2007) 631-644. [76] p. lundquist, p. artursson. oral absorption of peptides and nanoparticles across the human intestine: opportunities, limitations and studies in human tissues. advanced drug delivery reviews 106 (2016) 256-276. https://dx.doi.org/10.5599/admet.881 h. lennernäs et al. admet & dmpk 8(4) (2020) 375-390 390 [77] s.m. sharpe, x. qin, q. lu, e. feketeova, d.c. palange, w. dong, s.u. sheth, m.a. lee, d. reino, d.-z. xu. loss of the intestinal mucus layer in the normal rat causes gut injury, but not toxic mesenteric lymph nor lung injury. shock (augusta, ga.) 34 (2010) 475. [78] c. roos, d. dahlgren, e. sjögren, m. sjöblom, m. hedeland, h. lennernäs. jejunal absorption of aprepitant from nanosuspensions: role of particle size, prandial state and mucus layer. european journal of pharmaceutics and biopharmaceutics 132 (2018) 222-230. [79] y.-y. chiu, k. higaki, b.l. neudeck, j.l. barnett, l.s. welage, g.l. amidon. human jejunal permeability of cyclosporin a: influence of surfactants on p-glycoprotein efflux in caco-2 cells. pharmaceutical research 20 (2003) 749-756. [80] h. lennernas. intestinal permeability and its relevance for absorption and elimination. xenobiotica 37 (2007) 1015-1051. [81] c. roos, j. westergren, d. dahlgren, h. lennernäs, e. sjögren. mechanistic modelling of intestinal drug absorption–the in vivo effects of nanoparticles, hydrodynamics, and colloidal structures. european journal of pharmaceutics and biopharmaceutics 133 (2018) 70-76. [82] t. heimbach, s. suarez-sharp, m. kakhi, n. holmstock, a. olivares-morales, x. pepin, e. sjögren, e. tsakalozou, p. seo, m. li. dissolution and translational modeling strategies toward establishing an in vitro-in vivo link—a workshop summary report. springer, 2019. [83] h. lennernäs, c.-g. regårdh. evidence for an interaction between the β-blocker pafenolol and bile salts in the intestinal lumen of the rat leading to dose-dependent oral absorption and double peaks in the plasma concentration–time profile. pharmaceutical research 10 (1993) 879-883. [84] c. regårdh, a. heggelund, k. kylberg‐hanssen, p. lundborg. pharmacokinetics of pafenolol after iv and oral administration of three separate doses of different strength to man. biopharmaceutics and drug disposition 11 (1990) 607-617. [85] t. zhu, j.c. ansquer, m.t. kelly, d.j. sleep, r.s. pradhan. comparison of the gastrointestinal absorption and bioavailability of fenofibrate and fenofibric acid in humans. the journal of clinical pharmacology 50 (2010) 914-921. [86] a.m. yousaf, d.w. kim, y.-k. oh, c.s. yong, j.o. kim, h.-g. choi. enhanced oral bioavailability of fenofibrate using polymeric nanoparticulated systems: physicochemical characterization and in vivo investigation. international journal of nanomedicine 10 (2015) 1819. [87] b. borgström. studies on intestinal cholesterol absorption in the human. the journal of clinical investigation 39 (1960) 809-815. [88] d. nyholm, a. johansson, s.-m. aquilonius, e. hellquist, h. lennernäs, h. askmark. complexity of motor response to different doses of duodenal levodopa infusion in parkinson disease. clinical neuropharmacology 35 (2012) 6-14. [89] d. nyholm, h. askmark, c. gomes–trolin, t. knutson, h. lennernäs, c. nyström, s.-m. aquilonius. optimizing levodopa pharmacokinetics: intestinal infusion versus oral sustained-release tablets. clinical neuropharmacology 26 (2003) 156-163. [90] d. fleisher, c. li, y. zhou, l.-h. pao, a. karim. drug, meal and formulation interactions influencing drug absorption after oral administration. clinical pharmacokinetics 36 (1999) 233-254. ©2020 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ evaluation of pharmacokinetics of warfarin from validated pharmacokinetic-pharmacodynamic model doi: https://dx.doi.org/10.5599/admet.909 143 admet & dmpk 9(2) (2021) 143-149; doi: https://doi.org/10.5599/admet.909 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper evaluation of pharmacokinetics of warfarin from validated pharmacokinetic-pharmacodynamic model kannan sridharan 1* , rashed al banna 2 , aysha husain 2,3 1 department of pharmacology & therapeutics, college of medicine & medical sciences, arabian gulf university, manama, kingdom of bahrain. 2 department of cardiology, salmaniya medical hospital, ministry of health, manama, kingdom of bahrain. 3 rcsi-mub, manama, kingdom of bahrain full affiliation, address *corresponding author: e-mail: skannandr@gmail.com; orcid-id: 0000-0003-3811-6503 received: september 21, 2020; revised: january 06, 2021; published online: january 18, 2021 abstract background: pharmacokinetics of warfarin has not been described in our population. we derived the pharmacokinetic parameters from a validated pharmacokinetic-pharmacodynamic model. methods: patients receiving warfarin for at least 6 months were recruited and their demographic characteristics, prothrombin time international normalized ratio (pt-inr), warfarin doses and concomitant drugs were collected. using a validated pharmacokinetic-pharmacodynamic model, we predicted maximum plasma concentration (cmax), total clearance (cl), volume of distribution (vd) and elimination rate (k). warfarin sensitive index (wsi) and warfarin composite measures (wcm) were estimated from the dose and inr values. liver weight was predicted using validated formula. results: two-hundred and twenty patients were recruited. the following were the predicted pharmacokinetic parameters: cmax (mg/l) was 5.8 (0.4); k (l/day) was 1 (0.1); cl (l/day) was 2.1 (0.2); and vd (l) was 7.6 (0.2). patients with cmax and elimination rate outside the mean+1.96 sd had significantly lower wsi and higher wcm. significant correlations were observed between cmax with cl, vd, and k of warfarin. significant correlations were also observed between cl and vd of warfarin with liver weight of the study participants. conclusion: we predicted pharmacokinetic parameters of warfarin from the validated pharmacokinetic-pharmacodynamic model in our population. more studies are needed exploring the relationship between various pharmacodynamic indices of warfarin and pharmacokinetic parameters of warfarin. ©2021 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords pharmacokinetics; warfarin; anticoagulant introduction warfarin, an oral anticoagulant, exhibits narrow therapeutic window and poses clinical challenges in maintaining appropriate therapeutic effect. warfarin is metabolized in liver and its anticoagulant effect is considered optimum when the prothrombin time international normalized ratio (pt-inr) is to be maintained between 2.5 and 3.5 for those undergoing heart valve replacements and between 2 and 3 for others [1]. warfarin is completely absorbed with time to maximum concentration of 2-6 hours; volume of distribution (vd) of 10 l; clearance (cl) of 0.2 https://dx.doi.org/10.5599/admet.909 https://doi.org/10.5599/admet.909 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:skannandr@gmail.com http://creativecommons.org/licenses/by/4.0/ k. sridharan et al. admet & dmpk 9(2) (2021) 143-149 144 l/hour; and elimination half-life of 35 hours [2]. warfarin is a racemic mixture of sand r-warfarin posing challenges in measurements [3]. pharmacokinetic-pharmacodynamic (pk-pd) relationship of warfarin has been well elucidated and therapeutic drug monitoring is carried out by pt-inr measurement [4]. warfarin pharmacokinetics has not been described in our population. hence, we carried out the present study to predict the pharmacokinetic parameters of warfarin from the pharmacodynamic variables using a validated pd-pd model. experimental study ethics and design the study was retrospective carried out as a part of warfarin pharmacogenomics study after obtaining approval from institutional ethics committee and consent from study participants. the present study complies with the latest world medical association declaration of helsinki guidelines. study procedure participants receiving warfarin for at least 6 months were recruited in the present study. their demographics, pt-inr values, dose and frequency of warfarin, and concomitant drugs were obtained. we assessed cha₂ds₂-vasc, hasbled and same-tt2r2 scores for the study participants [5,6]. drugs/drug classes such as statins, proton pump inhibitors, carbamazepine and amiodarone were categorized as drugs with potential interaction with warfarin [7]. we followed national institute for health and care excellence (nice) guidelines for classifying anticoagulation control into good (ttr ≥ 65 %) and poor (ttr < 65 %) [8]. we estimated warfarin sensitive index (wsi) based on the ratio of pt-inr upon the last dose of warfarin [9]. we assessed the difference in pt-inr between the two consequent visits that was squared and divided by the time interval in days. this cross product with the average of pt-inr on the visits followed by square root of this product provided the inr variability. we log-transformed it to base 10 to obtain log-inr variability [10]. standardized z scores were obtained for ttr and log-inr variability and warfarin composite measure (wcm) was estimated [11]. liver weight was estimated by using the validated formula: liver weight (g) = 218 + body weight (kg) * 12.3 + gender * 51 (male-1; female-0) [12]. predicted pharmacokinetic parameters we used jpkd software © for estimating pharmacokinetic parameters of warfarin from the validated warfarin pd-pd inbuilt model [13]. the following equations were input in the model: cs = ((1/(-(m*cl/v)/(k^2)*(1-(k*tau/24)/(1-exp(-k*tau/24)))-(m/k)*ln((dose/v)/(cmax*(1-exp((cl/v)*(tau/24))))))+3.36)/4.368)^(1/0.383) inr = ((1/(-(m*cl/v)/(k^2)*(1-(k*tau/24)/(1-exp(-k*tau/24)))-(m/k)*ln((dose/v)/(cmax*(1-exp((cl/v)*(tau/24))))))+3.36)/4.368)^(1/0.383) dose = exp((1+((m*cl/v)/k^2)*(1-(k*tau/24)/(1-exp(-k*tau/24)))*(4.368*inr^(0.383)-3.36))/((m/k)*(4.368*inr^(0.383)-3.36)))* cmax *(1-exp((-cl/v)*tau/24))*v statistical analysis we represented the demographic details using descriptive statistics. chi-square test for association was used for assessing the categorical variables. numerical variables following the assessment of their distributions were tested using mann-whitney u test. correlations between the variables were tested using pearson correlation tests. a p-value of < 0.05 was considered significant. spss version 26 (ibm corp. admet & dmpk 9(2) (2021) 143-149 warfarin pharmacokinetics doi: https://dx.doi.org/10.5599/admet.909 145 released 2019. ibm spss statistics for windows, version 26.0. armonk, ny: ibm corp.) was used for statistical analysis. normal distributions were created for each of the key pharmacokinetic parameters and characteristics were compared between those with values within mean+1.96 sd and those outside this range. results demographics two-hundred and twenty patients were enrolled in this study. their demographic details are listed in table 1. the concomitant diagnoses in the study participants were as follows: systemic hypertension (n=130), diabetes mellitus (n=98), atrial fibrillation (n=85), ischemic heart disease (n=55), thyroid disorders (n=24), congestive cardiac failure (n=9) and bronchial asthma (n=15). table 1. demographic details of the study participants (n=220) pharmacokinetic parameters mean (sd) cmax (mg/l) amongst the study participants was 5.8 (0.4); elimination rate (l/day) was 1 (0.1); cl (l/day) was 2.1 (0.2); and vd (l) was 7.6 (0.2). distributions of the pharmacokinetic parameters are depicted in figure 1. nineteen patients had their cmax outside the mean+1.96 sd and 14 had the same for elimination rate. such patients were observed with significantly lower wsi and greater wcm (table 2). for cl, two had their values lower than mean–1.96 sd; and for vd, two had lower than mean–2 sd, and four more than mean+2 sd. due to the number constraints, we could not analyze the differences in any of the variables for cl and vd of warfarin. no significant differences were observed between the predicted pharmacokinetic parameters across the age groups (figure 2). similar values were obtained for males [cmax (mg/l – 5.8 (0.4); k (per day) – 1 (0.1); cl (l/day) – 2.1 (0.2); and vd (l) – 7.6 (0.2)] and females [cmax (mg/l – 5.8 (0.4); k (per day) – 1 (0.05); cl (l/day) – 2.1 (0.2); and vd (l) – 7.7 (0.2)]. significant correlations were observed between cmax and cl (r = 0.4; p=0.0001), vd (r = 0.28; p = 0.004) and k (r = 0.8; p = 0.0001) (figure 3). association between liver weights and predicted pharmacokinetic parameters scatterplot revealed significant correlations between warfarin cl (r=0.2; p=0.04) and vd (r= -0.2; p=0.03) with weight of the liver (figure 3). variables values age (years) $ 66.6 (13.4) male : female 1 : 1 body weight (kg) $ 75.2 (18.1) predicted liver weight (g) $ 1148 (265) cha2ds2-vasc score $ 3.7 (1.5) hasbled score $ 2.3 (1.1) same-tt2r2 score $ 1.4 (0.6) duration of warfarin (days) $ 1027 (572.2) warfarin sensitive index $ 0.7 (0.4) log inr variability $ -0.84 (0.2) warfarin composite measure score $ -1.3 (1.2) number of patients with drugs with potential interaction [n (%)] 136 (61.8) $ represented in mean (sd) https://dx.doi.org/10.5599/admet.909 k. sridharan et al. admet & dmpk 9(2) (2021) 143-149 146 figure 1. distribution of pharmacokinetic parameters in the study population. (a) distribution of cmax (mg/l); (b) elimination rate (per day); (c) clearance (l/day); (d) volume of distribution (l) in the study population table 2. comparison of characteristics between those within and outside 95% of normal distribution curves in the study population variables cmax (mg/l) elimination rate (per day) within 95% of nd (n= 201) outside 95% of nd (n= 19) p-values within 95% of nd (n=206) outside 95% of nd (n=14) p-values age (years) $ 67.1 (13.3) 61.7 (13.7) 0.7 68.6 (12.7) 64.2 (14.5) 0.6 body weight (kg) $ 75 (18.4) 77.6 (14.1) 0.4 72.8 (19) 81.6 (12) 0.2 ttr (%) $ 65.6 (17.2) 64.7 (16.4) 0.6 65.7 (16.1) 65 (21) 1 wsi $ 0.7 (0.4) 0.3 (0.1) 0.0001* 0.8 (0.4) 0.3 (0.1) 0.0001* log inr variability $ -0.8 (0.2) -0.9 (0.2) 0.06 -0.8 (0.2) -0.9 (0.2) 0.06 wcm $ -1.2 (1.2) -1.8 (0.9) 0.004* -1.1 (1.2) -1.7 (1.1) 0.08 potentially interacting drugs [n (%)] 123 (61.2) 12 (63.2) 0.9 135 (65.5) 8 (57.1) 0.5 y-young; m-middle-aged & older adults; e – elderly; ve – very elderly; ttr – time spent in therapeutic range; wsi – warfarin sensitive index; wcm – warfarin composite measure; nd – normal distribution; $ represented in mean (sd); * statistically significant. (a) (b) (c) (d) admet & dmpk 9(2) (2021) 143-149 warfarin pharmacokinetics doi: https://dx.doi.org/10.5599/admet.909 147 figure 2. predicted pharmacokinetic parameters of warfarin in different age groups. (a) predicted cmax (mg/l); (b) predicted elimination (l/day); (c) predicted clearance (l/day); (d) predicted vd (l) in different age groups. figure 3. scatter plot matrix between the pharmacokinetic parameters and the liver weight. warfarin clearance and volume of distribution were significantly correlated with the weight of the liver. clclearance; k – elimination rate; vd – volume of distribution (a) (b) (c) (d) https://dx.doi.org/10.5599/admet.909 k. sridharan et al. admet & dmpk 9(2) (2021) 143-149 148 discussion we evaluated the pharmacokinetic parameters of warfarin using validated pk-pd model in our population. two-hundred and twenty patients were recruited. the following were the predicted pharmacokinetic parameters: cmax (mg/l) was 5.8 (0.4); elimination rate (l/day) was 1 (0.1); cl (l/day) was 2.1 (0.2); and vd (l) was 7.6 (0.2). patients with cmax and elimination rate outside the mean+1.96 sd had significantly lower wsi and higher wcm. significant correlations were observed between cmax with cl, vd, and k of warfarin. significant correlations were also observed between cl and vd of warfarin with liver weight of the study participants. the pharmacokinetic parameters of racemic warfarin observed in the present study is like other populations [14]. although it would be better to measure sand rwarfarin concentrations, the clinical relevance of estimating enantiomers is limited. ours is the first study exploring the relationship between pharmacokinetic parameters and wsi and wcm (the pharmacodynamic parameters). lower wsi is corroborated with warfarin resistance. genotyping of vitamin k epoxide reductase complex in our population might provide a clue towards the prevalence and depth of warfarin resistance. the present study is also the first in exploring the relationship between liver weight and pharmacokinetic parameters of warfarin. though, liver weight has been argued to be a better corroborator of pharmacokinetic parameters in children, [15] its relationship in adults has not been studied. the study is limited in not validating the pharmacokinetic parameters using real-time measurements of plasma warfarin concentrations. to conclude, we predicted pharmacokinetic parameters of warfarin from the validated pharmacokinetic-pharmacodynamic model in our population. more studies are needed exploring the relationship between various pharmacodynamic indices of warfarin and pharmacokinetic parameters of warfarin. acknowledgements: we thank research technical support team, ministry of health, kingdom of bahrain for their ethics approval of this study. we are grateful to agu rcsi steering committee for their support in carrying out this research study. conflict of interest: none funding: the study was carried out as a part of the warfarin pharmacogenomics study that was funded by agu rcsi joint grant with grant number 2019-02. references [1] k. harter, m. levine, s.o. henderson. anticoagulation drug therapy: a review. west. j. emerg. med. 16(1) (2015) 11-17. [2] c. lv, c. liu, z. yao et al. the clinical pharmacokinetics and pharmacodynamics of warfarin when combined with compound danshen: a case study for combined treatment of coronary heart diseases with atrial fibrillation. front pharmacol. 8 (2017) 826. [3] d.a. barnette, b.p. johnson, d.l. pouncey et al. stereospecific metabolism of rand s-warfarin by human hepatic cytosolic reductases. drug metab. dispos. 45(9) (2017) 1000-1007. [4] j. bubp, r. coleman, d. mungall, g. raskob, r. hull. comparison of warfarin pharmacodynamic modelpredictions using manchester comparative reagent vs simplastin reagent. clin. pharmacol. ther. 47 (1990) 157. [5] d.a. lane, g.y. lip. use of the cha(2)ds(2)-vasc and has-bled scores to aid decision making for thromboprophylaxis in nonvalvular atrial fibrillation. circulation 126 (2012) 860-865. [6] g.y. lip. atrial fibrillation and stroke prevention: brief observations on the last decade. expert rev. cardiovasc. ther. 12 (2014) 403-406. admet & dmpk 9(2) (2021) 143-149 warfarin pharmacokinetics doi: https://dx.doi.org/10.5599/admet.909 149 [7] c. de fátima colet, a.t. amador, i. heineck. drug interactions and adverse events in a cohort of warfarin users attending public health clinics. international journal of cardiovascular sciences 32 (2019) 110-117. [8] atrial fibrillation. anticoagulation control. national institute for health and care excellence (nice). available at https://www.nice.org.uk/guidance/qs93/chapter/quality-statement-4-anticoagulationcontrol (accessed on 15 july 2020). [9] k. sridharan, r. al banna, a.m. qader, a hussain. evaluation of inter-patient variability in the pharmacodynamic indices of warfarin. expert rev. cardiovasc. ther. 18(11) (2020) 835-840. doi: https://doi.org/10.1080/14779072.2020.1814144. [10] s.d. fihn, a.a. gadisseur, e. pasterkamp, f.j. van der meer, w.g. breukink-engbers, l.m. gevenboere et al. comparison of control and stability of oral anticoagulant therapy using acenocoumarol versus phenprocoumon. thromb haemost. 90(2) (2003) 260-266. [11] z. razouki, j.f. burgess jr, a. ozonoff, s. zhao, d. berlowitz, a.j. rose. improving anticoagulation measurement novel warfarin composite measure. circ. cardiovasc. qual. outcomes 8(6) (2015) 600607. [12] s.c. chan, c.l. liu, c.m. lo, b.k. lam, e.w. lee, y. wong, s.t. fan. estimating liver weight of adults by body weight and gender. world j. gastroenterol. 12(14) (2006) 2217-2222. [13] java pk for desktop software. available at: http://pkpd.kmu.edu.tw/jpkd/index.php?pagenum_rs_news=1&totalrows_rs_news=23 (accessed on 21 sep 20). [14] a. desai, t. yamazaki, a.j. dietz, d. kowalski, c. lademacher, h. pearlman, s. akhtar, r. townsend. pharmacokinetic and pharmacodynamic evaluation of the drug-drug interaction between isavuconazole and warfarin in healthy subjects. clin. pharmacol. drug dev. 6(1) (2017) 86-92. [15] h. takahashi, s. ishikawa, s. nomoto, y. nishigaki, f. ando, t. kashima, s. kimura, m. kanamori, h. echizen. developmental changes in pharmacokinetics and pharmacodynamics of warfarin enantiomers in japanese children. clin. pharmacol. ther. 68(5) (2000) 541-555. ©2021 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) https://dx.doi.org/10.5599/admet.909 https://www.nice.org.uk/guidance/qs93/chapter/quality-statement-4-anticoagulation-control https://www.nice.org.uk/guidance/qs93/chapter/quality-statement-4-anticoagulation-control https://doi.org/10.1080/14779072.2020.1814144 http://pkpd.kmu.edu.tw/jpkd/index.php?pagenum_rs_news=1&totalrows_rs_news=23 http://pkpd.kmu.edu.tw/jpkd/index.php?pagenum_rs_news=1&totalrows_rs_news=23 http://creativecommons.org/licenses/by/3.0/ induced fit for cytochrome p450 3a4 based on molecular dynamics doi: 10.5599/admet.729 252 admet & dmpk 7(4) (2019) 252-266; doi: http://dx.doi.org/10.5599/admet.729 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper induced fit for cytochrome p450 3a4 based on molecular dynamics israel quiroga 1 , thomas scior 1, * 1 faculty of chemical sciences, benemérita universidad autónoma de puebla, puebla, pue., mexico *corresponding author: dr. thomas scior, e-mail: tscior@gmail.com received: september 20, 2019; revised: november 19, 2019; published: december 11, 2019 abstract the present study aims at numerically describing to what extent substrate enzyme complexes in solution may change over time as a natural process of conformational changes for a liganded enzyme in comparison to those movements which occur independently from substrate interaction, i.e. without a ligand. to this end, we selected structurally known pairs of liganded / unliganded cyp450 3a4 enzymes with different geometries hinting at induced fit events. we carried out molecular dynamics simulations (md) comparing the trajectories in a “cross-over” protocol: (i) we added the ligand to the unliganded crystal form which should adopt geometries similar to the known geometry of the liganded crystal structure during md, and – conversely – (ii) we removed the bound ligand form the known liganded complex to test if a geometry similar to the known unliganded (apo-) form can be adopted during md. to compare continues changes we measured root means square deviations and frequencies. results for case (i) hint at larger conformational changes required for accepting the substrate during its approach to final position – in contrast to case (ii) when mobility is fairly reduced by ligand binding (strain energy). in conclusion, a larger conformational sampling prior to ligand binding and the freezing-in (rigidity) of conformations for bound ligands can be interpreted as two conditions linked to induced-fit. keywords drug metabolism; cytochrome p450; cyp3a4; conformational selection; substrate selectivity; docking; rmsd, rmsf introduction two major tenets have been established in the field of admet research worldwide. on the one hand, experimental work ever since has laid the groundwork and much of the extant literature evidence in the field has been gathered by either clinical studies or in vivo animal tests. on the other hand, during the last three decades efforts have been undertaken to reduce the need for animal tests, applying ex vivo or in vitro tests (isolated enzymes, cell-based, or living tissues, etc.) in tandem with in silico applications of biophysical techniques in pharmacokinetics and drug toxicology. nowadays both tenets coexist in a complementary way all of which is exemplified by two recent drug inhibition studies of cyp targets: (i) animal tests with repaglinide and atazanavir [1], and (ii) computed inhibition models with organosulfur compounds [2]. cytochrome p450 (cyp450 or cyp for short) belongs to a superfamily of monooxygenase enzymes that http://dx.doi.org/10.5599/admet.729 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:tscior@gmail.com admet & dmpk 7(4) (2019) 252-266 induced fit of cyp3a4 doi: 10.5599/admet.729 253 participate in the pathway of elimination of xenobiotics, including drugs in humans [3]. certain cyp enzymes are capable of metabolizing a large variety of substrates of different sizes. they belong to a group of proteins with structural flexibility in addition to a huge cavity at the active site [4-8]. these enzymes are capable of adopting different conformations, both in their main chain and in their side chains, which allows the metabolic activity to be carried out on their various substrates; a phenomenon that in enzymes is known as induced fit. prior to our present work, we laid the knowledge bases for induced fit and molecular flexibility of cyp and drugs in a survey article [9]. induced fit describes the structural changes upon ligand receptor binding. the ligand induces these changes at the binding site which normally constitutes a cavity in form of a deep cleft or a shallow surface depression. in our study the cyp enzyme binds its substrate for biotransformation, here a drug. induced fit also means that the geometry of the entrance to the binding site has a direct impact on the local shape (so-called active conformations) of the substrate. this matters in drug design when only three-dimensional models from free ligands (unbound drugs) and receptors (without ligands) have been reported, but the exact atom positions of the ligand receptor complex are necessary to understand the molecular mechanism of action to design new drugs. to simulate the effect that the induced fit has on the cyp activities in computational studies is a challenge. albeit, some biophysical techniques have been proposed that would overcome the difficulties to simulate induced fit: (i) "soft-docking", is the deliberate reduction of repulsive forces; (ii) the management of side chain rotations using rotamer libraries; (iii) multiple structure modeling that considers the use of several crystal structures with different conformations of the same protein or related isoforms, or (iv) molecular dynamics (md) calculations and the subsequent use of docking in a frame of the md simulation that shows a favorable conformation [10-12]. md simulates the structural flexibility of proteins as well as the characteristics that kept them stable. however, these techniques are not fool-proof and do not guarantee success. in some cases, neither crystal structures are at hand nor do rotamer libraries provide solutions to model the experimental data. in the context of our study hydroxylation patterns, cyp enzyme substrate binding specificities, flexible protein segments or thermodynamic properties are amenable to computer-aided simulations [13-17]. skopalik et al. compared the flexibility patterns of cyp3a4, cyp2c9 and cyp2a6 [14]. the researchers identified flexible loops adjacent to the active sites, e.g. b-c loop in cyp2c9 or f-g loop in cyp3a4 (figure 1). moreover, chang et al. found that salt bridges, aliphatic and aromatic interactions between secondary structures of cyps maintain their initial conformations [18]. in 2009 seifert and pleiss published a study on substrate specificities of cyp [5]. they reasoned that the sites of metabolism (som), i.e. which ligand atoms could be hydroxylated, correspond tightly to the conformational changes at the active sites. as a most valuable asset, predicting products and som for drugs by cyp enzymes was found amenable to in-silico approaches [15]. in this context we also explored the som of certain cyp substrates applying ligand docking and molecular dynamics techniques [19,20]. crystallography does not always provide evidence for main or side chain movements since structure elucidation is carried out under solid phase package forces at very low temperatures and under destructive x-ray exposure of the crystal probes. other model limitations are known as follows: (i) substrate size and adaptive volume of its cavity; (ii) resolution of the crystallographic image; (iii) crystallographic structure optimization and modeling to predict geometries for unresolved (highly flexible) areas; or (iv) 3d templates for homology modeling of unknown target structures [21-23]. quiroga and scior admet & dmpk 7(4) (2019) 252-266 254 and even point mutations further away from the active sites could alter the overall three-dimensional geometries – and subsequently affect the way of how cyp enzymes interact with their substrates [24,25]. figure 1. display of human enzyme cyp3a4. the three-dimensional model is centered on the active site with the heme group with central atom fe (brown). the segments are labeled by letters from a to l following the sequence from nto c-term. the red color marks the f-g loop with its f’-f’’ segment. it was reported to undergo conformational changes upon ligand binding [14]. on the other hand, lampe and colleagues of the group of ortiz de montellano studied the conformational changes of liganded and unliganded cyp119 combining nmr and md techniques [26]. the group found that weak binders to the active site allow the complex to move more than when stronger binders occupied the cavity at the active site. the authors conclude that substrate recognition and binding could be more a matter of conformational selection than a defined induced fit which could be described as the induction of a distinct conformational change in the receptor in the presence of a ligand. in the case of substrate cyp complexes the substrate occupies the cavity at the heme site in a favorable orientation for its subsequent biotransformation. recently, we examined the details of the p450 catalytic cycle elsewhere [9]. it is noteworthy to underscore that the aforementioned reports [4-8,10-17] do not undertake numeric measurements to reproduce the displacement of initial atom positions during md production runs. root means square deviations of atom positions in space (rmsd for short) can be used to numerically reflect atom movements during a period of time in an md simulation when following selected atoms. a larger rmsd value (typically 4 to 6 or 7) indicates that protein structures have moved more freely and have suffered larger geometrical shifts [27]. rmsd comparison for structural deviations against a template, however, is not fool-proof and other distance measurements have been proposed [28]. in particular, global rmsd measurements may reflect changes in irrelevant parts and not just at the center of interest; for instance, the movements may occur more around the active site. to circumvent such drawbacks, any movement was under scrutiny, such that not only the heme site mattered but also all other protein parts did because any geometrical changes in liganded and unliganded proteins could hint at induced fit phenomena. we averaged the positions of each atom along the md trajectory as a reference to calculate rmsd. this way, we obtained a trustworthy framework of the average positions of each atom of the protein because any larger magnitude of movement would be admet & dmpk 7(4) (2019) 252-266 induced fit of cyp3a4 doi: 10.5599/admet.729 255 represented by higher rmsd values (table 5 in the results section). besides rmsd, so-called heat maps of root mean square fluctuation were also analyzed. given the reported conundrum about geometrical changes in general and in the context of induced fit for cyp enzymes [9,21,26-28], we decided to study pairs of unliganded and liganded crystal complexes of the same cyp enzyme. this way, geometrical differences are known before md simulation thanks to experimental determination (pdb data). moreover, after relaxation to thermal equilibrium any changes in the observed geometries during md simulation could be related to induced fit [29-33]. the silently made assumption here was that the structural differences between the liganded and unliganded crystal structures could no longer be due to crystal package forces after force field treatment to relax both protein structures (cf. aforementioned model limitations). according to our literature survey [9], no publication considered the pairwise study of cyp enzymes in presence and absence of ligands and how this could affect their geometries. based on what is known to date and not being explicitly documented in the extant literature, the need arises to systematically study the ligand effect on initial atom positions by computational md and evaluate the numerical results in the context of induced fit. experimental structural input data molecular dynamics techniques were used to evaluate the conformational changes in the structure of the protein backbone. the md results were evaluated based on the atoms’ start positions which were taken from the crystal structures. selection of input structures after inspecting pairs of crystal structures of human cyps with and without drug substrates in their respective cavities in the protein database (pdb) [34], it was decided to use two pdb entries for human cyp3a4: 1tqn and 3nxu [29,30]. the former presents a single chain biological unit without substrate. in turn, 3nxu presents a homodimer with the antiviral drug ritonavir (rit) as its substrate. rit occupies the active site. for this study, we removed chain b from 3nxu. in 1tqn four amino acids (k282, e283, t284 and e285) are missing in a loop (h-i) not belonging to the active site while 3nxu shows unresolved residues in positions similar to that of 1tqn. this pair was chosen based on the following five reasons: (i) both structures were chosen because of the geometrical changes which can readily be ascribed to the presence or absence of ligand at the active site: in 1tqn the f-g loop is displaced inwards into the catalytic cavity, while in 3nxu the backbone and side chain geometries of this loop are relocated due to the presence of ritonavir (figure 2); (ii) they are crystallographic works of different research teams; (iii) they have a similar resolution of high quality; (iv) 3nxu uses molecular replacement as a method to determine its structure, i.e. 1tqn was taken as a crystallographic reconstruction template; (v) atoms are not missing at their active sites and displacement of their main and side chains is clearly observable. we measured the binding site volume to confirm the generally accepted view that hcyp3a4 can bind many drugs of different shapes and sizes thanks to its huge and variable cavity between 1440 and 2130 [å 3 ], cf. human cyp2c9, for example, has a smaller volume: up to 1400 [å 3 ] (pdb code: 1r9o) [20]. quiroga and scior admet & dmpk 7(4) (2019) 252-266 256 figure 2. superposition of a pair of liganded and unliganded enzyme pair. the two crystal structures of human cyp3a4 are displayed by backbone lines indicating helical and loop parts in gray color. on top two colors mark conformational differences in the f-g loops of unliganded 1tqn (magenta) and liganded 3nxu (bluish). the ligand is ritonavir (atom sticks in the center, blue c, dark blue n, red o). the catalytic heme group (brown sticks with o atoms in red and fe in violet/purple). all h atoms omitted. four combinations of enzyme and substrate structures as input for molecular dynamics the following section describes how four input models for md were created from a pair of pdb entries. the two crystal structures were taken as start geometries for two input model without changes: (i) the chain a of the originally unliganded 1tqn [29] and (ii) the originally liganded 3nxu [30]. in addition, two more input models were generated to complete the cross-over design to study all four possible combinations systematically: (iii) an unliganded model which was generated from the originally liganded 3nxu by removing the ligand and leaving all spatial coordinates of the protein unchanged. furthermore, another liganded 1tqn model (iv) was created from the originally unliganded 1tqn. here the ligand’s start position was placed into the cavity near its observed pose (from 3nxu) near the heme group in a way which again left the cartesian coordinates of the enzyme unchanged (table 1). table 1. three-dimensional molecular models created to study induced fit for human cyp3a4. the label “+rit” symbolizes the presence of ritonavir at the active site, regardless whether the position of rit was taken from crystal positions or computed. the “-rit” (minus ligand) was used in case of unliganded enzymes, i.e. the ligand is absence. complexed enzyme uncomplexed enzyme observed (x-ray) systems pdb code: 3nxu, liganded receptor, the ligand is ritonavir: (i) 3nxu+rit pdb code: 1tqn, unliganded receptor: (ii) 1tqn-rit modeled systems (iii) 1tqn+rit (iv) 3nxu-rit descriptors and chemometrics for induced fit evaluation md simulations studied whether the geometries of the liganded or unliganded complexes will change during simulation time. whether or not the model conserves or changes its overall structure can be measured as rmsd comparing the crystals geometries to the md production geometries for the liganded and unliganded complexes. the rmsd is defined as: 2 1 atomsn i i atoms rmsd n    , admet & dmpk 7(4) (2019) 252-266 induced fit of cyp3a4 doi: 10.5599/admet.729 257 where δi is the distance between the atom i and the reference atom. using existing computational tools, it is possible to measure descriptors of the active site that allow us to correlate the induced fit effects with its biological activities. the vmd software was used to calculate the measurements between the model complexes [35]. parameterization and molecular dynamics namd [36] software was used with charmm27 force field [37] and displayed by vmd [35] to carry out the md runs and the equilibration described above. the four input models described in table 1 were solvated in a cubic water box with the water model tip3p (1nm). the periodic boundary conditions were applied in all directions. the parameters for heme central atom (fe) were set to coordinate oxygen moieties at 2.5 å [38-40], and thereby maintaining a neutral charge between fe and o with q= +2 and -2, respectively. of note heme group atoms and reactive oxygen as a distal coordinated moiety were fixed in their experimentally observed positions as a positional reference to analyze md runs. there were 100000 steps of energy minimization for all atomic coordinates. to balance each model the temperature was increased from 0 k to 310 k gradually for 0.25 ns. in the next stage, the balanced systems were subjected to 60 ns of production time at a time step of 2 fs at room temperature (310 k) and normal pressure (1 atm). analysis of the results once the md simulations were performed, it was expected that the md geometries of the liganded and unliganded input models will merge into similar geometries of the crystal structures 3nxu or 1tqn. in turn, it was also expected that both liganded models undergo restricted movements of their side chains, seen as smaller rmsd values. the two models without substrate would be similar to crystal structure 1tqn. but with smaller rmsd values for the unliganded model based on the originally unliganded 1tqn, and conversely the need of larger rearrangements in space (more mobility) with fairly higher rmsd values of the unliganded model stemming from the originally liganded 3nxu (table 2). table 2. the four md simulations and their expected outcome in terms of end geometries. the starting positions are related with table 1. comparison of the protein structures of the four systems after the md run with the original protein structures of the crystals reported in pdb 1tqn and 3nxu. type of 3d-model start geometry (system) enzyme geometric similarity with pdb interpretation liganded enzyme (iii) 1tqn+rit 3nxu induced fit liganded enzyme (iii) 1tqn+rit 1tqn no induced fit liganded enzyme (i) 3nxu+rit 1tqn no induced fit liganded enzyme (i) 3nxu+rit 3nxu induced fit unliganded enzyme (iv) 3nxu-rit 1tqn induced fit unliganded enzyme (iv) 3nxu-rit 3nxu no induced fit unliganded enzyme (ii) 1tqn-rit 3nxu no induced fit unliganded enzyme (ii) 1tqn-rit 1tqn induced fit so far, our analysis does not account for trajectory differences. to tackle this issue, too, we compared rmsd in a pair-wise manner between the four md models to detect some portion of induced fit. after all, the two molecular systems with ritonavir had to show higher similarities compared the two models without it. quiroga and scior admet & dmpk 7(4) (2019) 252-266 258 the cross-over study design and the pairwise rmsd comparison helped assessing the mobility or rigidity effect that the substrate had on human cyp3a4. as a most valuable asset, the approach allows not only evaluating the effect of initial positions of the crystal atoms in the md studies but also testing the extent of conformational changes with the subsequent interpretation as induced fit for the enzymatic activities of cyp targets. results and discussion the spatial difference between a pair of corresponding crystal structures (pdb codes: 1tqn and 3nxu) was measured in a first step. effectively, the pair constitutes one of the rare cases where the same cyp enzyme is available in liganded and unliganded states such that their geometrical difference can be interpreted as induced fit after reducing crystal package strain energies by relaxation to equilibrium state under the charmm force field (cf. details about assumptions or limitations). of note, two rmsd types were calculated: (i) with only main chain atoms, or (ii) with all atoms except hydrogen atoms. both types of calculations were carried out on either the entire protein or at the active site. the latter comprises the literature-known segments from residue h54 to f60, s100 to e122, t207 to p218, i238 to e244, v296 to s311 and f367 to r375. then, all possible pairs of rmsd were compared and special attention drawn to the overall mobility and the f-g loop displacements for substrate recognition (table 3). the geometrical changes at the active sites of 1tqn and 3nxu were documented, too (table3). both represent the same enzyme, but their major structural differences could be pinpointed to only those residues which interacted with ritonavir. compared region rmsd value (å) the whole protein 1.6 active site 3.8 table 3. the computed value of the root-mean square distance (rmsd, in å) between the original crystal structures of pdb 1tqn and 3nxu comparing the entire protein and only the region corresponding to the active site. the calculation of both regions was taking into account the atoms corresponding to the main chain and later to the atoms that were not hydrogens. following table 2, we expected the unliganded (liganded) models to merge into geometries similar to 1tqn (3nxu) during md simulations. intriguingly, the md results in table 4 fell short of expectation. the absence of rit allows cyp3a4 to explore the conformational space much more than the liganded enzyme which resembles more the crystal structure 3nxu (table 4). table 4. computed rmsd values [å] for 1tqn and 3nxu. the averaged values represent a movement corresponding to an area around the active site compared against all positions during the entire simulation. the main chain and all non-h atoms in the area were compared. compared region : whole protein active site reference pdb crystal 1tqn 3nxu 1tqn 3nxu rmsd of system (iii) 1tqn+rit 1.5 2.1 1.5 3.8 rmsd of system (ii) 1tqn-rit (orig.) 2.2 2.8 3.9 5.1 rmsd of system (i) 3nxu+rit (orig.) 2.2 1.2 4.0 1.1 rmsd of system (iv) 3nxu-rit 2.2 1.7 3.8 2.4 playback analysis by vmd showed that the liganded models follow trajectories in a much shorter range of near-site atomic displacements, i.e. their average geometries lie closer to the starting geometries. in more general terms, ligand receptor complexes are less flexible by nature, due to ligand-mediated noncovalent bonds at the active site (figure 3). admet & dmpk 7(4) (2019) 252-266 induced fit of cyp3a4 doi: 10.5599/admet.729 259 figure 3. intermolecular interactions between ritonavir substrate and cyp3a4 enzyme (3nxu). the ligand’s favourable intermolecular interactions at the active site stabilized the complexes (1tqn+rit and 3nxu+rit). the noncovalent bonds (intermittent lines) between ligand and heme or amino acids decrease significantly the protein flexibility. the presence of rit at the active site had already induced a larger geometrical shift (figure 4 and table 5). this conclusion is in total keeping with the experimentally observed huge rmsd differences between ligand-bound and unbounded crystal structures: rmsd 3.8 at the active site or rmsd 1.6 for the entire complex. the two moving f-g loops from 3nxu with or without ligand were superimposed to display the difference in mobility throughout the entire dynamic simulation (figure 4). again, with rit in the binding cleft smaller displacements were noticed (figure 3). the presence of rit exercises a twofold effect: (i) preventing the f-g loop from occupying positions close to the heme group, and (ii) intermolecular interactions stiffening the loop architecture. table 5. calculated rmsd values [å] for unliganded 1tqn, 1tqn+rit, 3nxu-rit and liganded 3nxu. the averaged values represent a movement corresponding to an area around the active site compared against all positions during the entire simulation. the columns sd, min and max show the values of standard deviation sd, rmsd minima and maxima. hcyp3a4 model rmsd [å] sd min [å] max [å] 1tqn+rit 1.2 0.36 0.32 3.59 1tqn-rit (orig.) 2.0 0.66 0.36 5.07 3nxu+rit (orig.) 1.2 0.38 0.20 2.74 3nxu-rit 2.2 0.84 0.29 5.35 in table 5 the differential behavior of rit-bound versus unbound systems becomes evident with less flexible main chains for the former and high flexibilities for the latter. it is, however, possible to measure the overall geometric similarities without the dynamic contributions after assessing corresponding differences between the four molecular systems in table 1. to get rid of rmsd bias due to movement restrains of more stable liganded complexes, the rmsd of any two models (now called system1 or system2, respectively) was calculated taking as reference the average value of corresponding atom positions from both systems. the calculated rmsd of system1 with respect to the average positions of both systems taken together is called a1, and the calculated rmsd of system2 respect to the average quiroga and scior admet & dmpk 7(4) (2019) 252-266 260 positions of both systems is called a2. rmsd is amenable to recalculation thanks to the fact that all systems have the same total number of atoms in their data files. therefore, even ever higher a values (figure 5) were met in a system, the greater the dynamic and structural differences would be. the equation (a1m1)+(a2-m2), where m represents the rmsd for the unbiased movement of a system (table 5), allows recalculating the structural similarities between two systems in a way which is independent of their intrinsic (own) movements (figure 5 ). this way, it was possible to evaluate similarities for all model pairs: (i) the higher the values, the less structural similarities during simulations; (ii) conversely, the lower the values, the higher the outcome of similar conformations during md. figure 4. superposition of md conformations with the f-g loop near the active site. model 3nxu-rit (left) and liganded 3nxu (center). both models were aligned in space (right). liganded 3nxu (center) had a much smaller mobility range than 3nxu-rit (left) during dynamics simulation. figure 5. averaged rmsd calculation with the equation (a1-m1) + (a2-m2). m is the rmsd that expresses the movement of a system (table 5). a is the rmsd of a system when compared to the average positions of the atoms of both systems. the higher the calculated value, the more structural differences two systems have in an md run. the results of the recalculated rmsd were also summarized (table 6). a higher rmsd in columns 5 and 6 of table 6 expressed which of the two systems moved more relative to the other one. if both systems moved in a similar way they would have comparable rmsd values. as indicator of changes only atoms belonging to the active site and the main chain atoms were considered. upon calculating the difference between a1 and a2 (figure 5), it becomes evident that the lowest dynamic differences belong to the pairs 1tqn+rit and liganded 3nxu, and unliganded 1tqn and 3nxurit. another bias (unfair treatment of numeric data) was detected, namely the initial atom position admet & dmpk 7(4) (2019) 252-266 induced fit of cyp3a4 doi: 10.5599/admet.729 261 dependence for the developing trajectories as shown by the values in column 7 in table 6. those md data were the ones that had the greatest structural similarities when generated from the same input model. precisely, this finding paved the way to work-around the initial position bias by recalculating rmsd of system1 and system2 (the liganded and unliganded models). table 6. structural and dynamic differences in the modeled systems for this study (1tqn-rit, 1tqn + rit, 3nxu-rit and 3nxu + rit) measured as rmsd (in å). the comparison between the four systems is made among themselves taking into account the six possible combinations of table 2. m1 and m2 are the values of rmsd for the calculated movement of systems 1 and 2 respectively in table 5. a1 and a2 represent the rmsd of systems 1 and 2 respectively, having as reference the average positions of the atoms of both systems. (m1-a1) + (m2-a2) measures the differences between systems 1 and 2. system1 system2 m1 m2 a1 a2 (a1-m1)+(a2-m2) 3nxu-rit 3nxu+rit 2.2 1.2 1.5 1.2 -0.6 1tqn-rit 1tqn+rit 2.0 1.2 1.7 1.5 0.0 1tqn+rit 3nxu-rit 1.2 2.2 1.8 2.0 0.4 1tqn-rit 3nxu-rit 2.0 2.2 2.5 2.5 0.8 1tqn+rit 3nxu+rit 1.2 1.2 2.0 2.0 1.6 1tqn-rit 3nxu+rit 2.0 1.2 2.7 2.6 2.2 figure 6 shows illustrations of the model dynamics and representative (averaged) structures thereof. in particular, it displays the superpositions with the f-g loop. systems with ritonavir have a smaller range of movement than those without ritonavir in the superposed region and elsewhere occupy different locations (midsection in figure 6). figure 6. superposition of md conformations for the the f-g loop. the f-g loops of our four models are compared. they lies in close proximity to the heme group at the active site. first row to the left: the 3nxu-rit systems (red) superimposed with 3nxu+rit (orange). second row to the left: 1tqn-rit (blue) superimposed with 1tqn+rit (black). bottom row to the left: 1tqn+rit (black) superimposed with 3nxu-rit (red). first row to the right: 3nxu-rit (red) superimposed with unliganded 1tqn (blue). second row to the right: 1tqn+rit (black) superimposed with 3nxu+rit (orange). bottom row to the right: 1tqn-rit (blue) superimposed with 3nxu+rit (orange). upon superposition of 3nxu-rit with 3nxu+rit (right side of figures 4 and 6), the occupied and unoccupied spaces at the active sites can be distinguished, as it is the particular case with the f-g loop of 3nxu+rit (figure 7, orange). it is located nearby the heme group and ligand rit, while the f-g loop of the quiroga and scior admet & dmpk 7(4) (2019) 252-266 262 3nxu-rit system (figure 7, red) is displaced to a distinct location above the f-g loop of 3nxu+rit. the md simulation allows us to infer that this f-g loop displacement of human cyp3a4 into an energetically more favorable conformation fairly reduces strain energy by steric hindrance. the underlying geometrical changes in side chains and backbone reflect the enzyme’s capability to metabolize substrates with an even larger volume than rit – all of which leads to the well-known flexibility ascribed to cyp enzymes in general. as a direct result, our md simulation provides computed evidence for induced-fit between receptor and ligand (here enzyme and substrate). figure 7. left panel: superposition of md conformations (left) with the f-g loop near the active site of 3nxurit (red) or liganded 3nxu (orange). the liganded 3nxu (orange) lies in closer proximity to heme than ritonavir during md. right panel: root mean square distance over time plot (rmsd) for the distance between ritonavir and f-g loop during md. of note, lacking substrate the coordinates, the unliganded models (here: 3nxu-rit) the corresponding atoms of the crystal structure were used as reference to calculate rmsd. the minimum distances between references coordinates and protein of both systems (3nxu-rit and liganded 3nxu) were assessed during 60 ns. comparing the heights in the plot of the root mean square fluctuation (rmsf) reflects changing flexibilities within the protein sequences (figure 8). the lower fluctuations stem from 1tqn+rit and 3nxu+rit. more rigid regions can be spotted between them, too. the highest peaks are similar sequence regions, between residue id 240, 270, 282, 95 and 44. in good keeping with the expected binding relevance of the f-g loop, a mayor geometrical shift had occurred when placing the drug into the unbound enzyme structure (1tqn+rit). this constitutes a huge fluctuation near residue id 200. the unbound start conformation had to adopt geometries (binding conformations) to accommodate the ligand. intriguingly, unresolved segments in the crystal structure (here regions at residue id 270 to 282) reflect low complexity areas and coincide with higher fluctuations all of which nicely explain why they were missing in the crystal structure. in a more general view, md studies simulate molecular movements in water solution under thermodynamic conditions which are relevant in biological systems, but other aspects may become important too: solvent entropies from bulk water and water-exclusion zones in lipophilic pockets or shallow surfaces, enthalpies from destabilizing hydrogen bonding, etc. in sight of such model limitations, trajectory interpretation was undertaken with utmost caution. moreover, the physical and chemical microconditions (temperature and molecular neighborhood effects) at the moment of crystallogenesis or crystallography are fairly different from the physiological ones in living cells all of which can cause artifacts admet & dmpk 7(4) (2019) 252-266 induced fit of cyp3a4 doi: 10.5599/admet.729 263 of (starting) geometries [21]. in our case, the average positions of the f-g loop of cyp3a4 of the unliganded 1tqn system could not approach the central heme group, giving way for a larger cavity and yielding higher geometric similarities with the crystal structure of cyp3a4 (3nxu-rit). this is in keeping with published findings [26] because our rmsd data showed that the movements of enzyme-substrate complexes become more restricted than their unliganded counterparts. figure 8. root mean square fluctuation plot (rmsf). the x axis represents the residues of the primary sequence. the y axis quantifies the corresponding movement of each residue measured in [å]. the higher the y value, the higher the mobility of adjacent residues in segment of the protein. the four combinations of the liganded / unliganded pair are color coded (inlay panel). conclusions the induced fit or conformational selection process between drug and cytochrome p450 enzymes (cyp) was studied by molecular dynamics simulations. the study demonstrates that the smallest dynamic differences are met in pairs which share either the presence or the absence of bound substrates. indeed, our pairwise study design unraveled that substrate alone does provide an effect on the geometric changes of the enzyme and that they can be observed and measured in molecular dynamics studies. the computed trajectories were parameterized as rmsd reflecting spatial rearrangements of the start geometries which were taken from crystal structures of substrate-bound and unbound states of identical proteins. to this end, crystal structures of known liganded/unliganded pairs had been gathered prior to simulation. precisely, comparing averaged rmsd data (originally liganded 3nxu: 0.9 < liganded 1tqn < unliganded 3nxu: 1.55 < originally unliganded 1tqn: 1.9) lent insight into hidden driven forces leading to geometrical changes for substrate recognition of human cyp3a4 proteins. for a worst-case scenario, we would state that both unliganded models did not follow a trajectory with f-g loop rearrangements to merge into the observed structure of the originally unliganded protein (1tqn). this finding, however, attests to the wellknown conformational flexibility of the cyp family in general, reflecting the need for flexibility for successful ligand binding. hence, the not merging trajectories into crystal geometries simulate a natural (intrinsic) behavior here. conversely, ligand binding fairly reduced the flexibility of the complex. in a natural situation the active site has evolved to strengthen the substrate attachment until its biotransformation quiroga and scior admet & dmpk 7(4) (2019) 252-266 264 took place (oxidation, hydroxylation). interpreting in this way our results, the simulations come close – if not – reflect properly natural behavior, all of which is our best-case scenario where md is capable of reproducing induced fit of cyps. all told, it seems not far-fetched to assume that averaged rmsd captures the essence of conformational changes of induced fit. md estimates the necessary changes for substrate recognition and regioselectivity. the results on theoretical grounds further the awareness for induced fit mechanisms upon ligand binding. our work contributes to describe induced fit which has been frequently cited in the extant literature body but poorly described at an atomic scale. authors’ contributions: md was carried out by iq. the manuscript was written through contributions of both authors. all authors have given approval to the final version of the manuscript. acknowledgements: the authors would like to thank the financial support of consejo nacional de ciencia y tecnologia (conacyt) for providing iqm’s doctoral scholarship. conflict of interest: none. references [1] t.e. goud, s. maddi, n. devanna, t.r. prasad. evaluation of pharmacokinetic and pharmacodynamic interaction between repaglinide and atazanavir in healthy, diabetic and hepatic impaired rats: possible inhibition of cyp3a, oatp, and p-glycoprotein transporters. admet & dmpk 4 (2016) 269279. [2] d.a. ramirez, e. marchevsky, j.m. luco, a. camargo. quantitative structure-activity relationship to elucidate human cyp2a6 inhibition by organosulfur compounds. admet & dmpk 7 (2019) 196-209. [3] f.j. gonzalez, r.h. tukey. the pharmacological basis of therapeutics, mcgraw-hill,new york, 2011, p. 71-91. [4] l. perić-hassler, e. stjernschantz, c. oostenbrink, d.p. geerke. cyp 2d6 binding affinity predictions using multiple ligand and protein conformations. int. j. mol. sci. 14 (2013) 24514-24530. [5] a. seifert, j. pleiss. identification of selectivity‐determining residues in cytochrome p450 monooxygenases: a systematic analysis of the substrate recognition site 5. proteins: struct. funct. bioinf. 74 (2009) 1028-1035. [6] a. seifert, s. tatzel, r.d. schmid, j. pleiss. multiple molecular dynamics simulations of human p450 monooxygenase cyp2c9: the molecular basis of substrate binding and regioselectivity toward warfarin. proteins: struct. funct. bioinf. 64 (2006) 147-155. [7] c. sotriffer. accounting for induced-fit effects in docking: what is possible and what is not?. curr. top. med. chem. 11 (2011) 179-191. [8] x. yu, v. cojocaru, r.c. wade. conformational diversity and ligand tunnels of mammalian cytochrome p450s. biotechnol. appl. biochem. 60 (2013) 134-145. [9] t. scior, i. quiroga, b. kammerer. inquiry of literature evidence for induced fit and regioselectivity of cytochrome p450 enzyme cyp2c9: a critical review. sciol biotechnology 1 (2018) 30-48. [10] s.b. nabuurs, m. wagener, j. de vlieg. a flexible approach to induced fit docking. j. med. chem. 50 (2007) 6507-6518. [11] e. sano, w. li, h. yuki, x. liu, v. furihata, k. kobayashi, k. chiba, s. neya, t. hoshino. mechanism of the decrease in catalytic activity of human cytochrome p450 2c9 polymorphic variants investigated by computational analysis. j. comput. chem. 15 (2010) 2746-2758. [12] y. sheng, y. chen, l. wang, g. liu, w. li, y. tang. effects of protein flexibility on the site of metabolism prediction for cyp2a6 substrates. j. mol. graph. model. 54 (2014) 90-99. admet & dmpk 7(4) (2019) 252-266 induced fit of cyp3a4 doi: 10.5599/admet.729 265 [13] n. krishnamoorthy, p. gajendrarao, s. thangapandian, y. lee, k. w. lee. probing possible egress channels for multiple ligands in human cyp3a4: a molecular modeling study. j. mol. model. 16 (2010) 607-614. [14] j. skopalík, p. anzenbacher, m. otyepka. flexibility of human cytochromes p450: molecular dynamics reveals differences between cyps 3a4, 2c9, and 2a6, which correlate with their substrate preferences. j. phys. chem. b. 112 (2008) 8165-8173. [15] j. kirchmair, m.j. williamson, j.d. tyzack, l. tan, p.j. bond, a. bender, r.c glen. computational prediction of metabolism: sites, products, sar, p450 enzyme dynamics, and mechanisms. j. chem. inf. model. 52 (2012) 617-648. [16] e. asciutto, m. dang, s. pochapsky, j. madura, t. pochapsky. experimentally restrained molecular dynamics simulations for characterizing the open states of cytochrome p450cam. biochemistry 50 (2011) 1664-1671. [17] a. roberts, m. cheesman, a. primak, m. bowman, w. atkins, a. rettie. intramolecular heme ligation of the cytochrome p450 2c9 r108h mutant demonstrates pronounced conformational flexibility of the b−c loop region: implication for substrate binding. biochemistry 49 (2010) 8700-8708. [18] y. chang, g. loew. homology modeling, molecular dynamics simulations, and analysis of cyp119, a p450 enzyme from extreme acidothermophilic archaeon sulfolobussolfataricus. biochemistry 39 (2000) 2484-2498. [19] i. quiroga, t. scior. structure-function analysis of the cytochromes p450, responsible for phenprocoumon metabolism. j. mex. chem. soc. 61 (2017) 349-360. [20] i. quiroga, f. meléndez, k. atonal-salvador, b. kammerer, t. scior. identification a new site of metabolism for phenprocoumon by modeling it’s cyp2c9 hydroxylation pattern. saj pharm. pharmacol. 5 (2018) 1-12. [21] m. depristo, p. de bakker, t. blundell. heterogeneity and inaccuracy in protein structures solved by x-ray crystallography. structure 12 (2004) 831-838. [22] m. andrec, d.a. snyder, z. zhou, j. young, g.t. montelione, r.m. levy. a large data set comparison of protein structures determined by crystallography and nmr: statistical test for structural differences and the effect of crystal packing. proteins 69 (2007) 449-465. [23] h.p. hersleth, k.k. andersson. how different oxidation states of crystalline myoglobin are influenced by x-rays. biochim biophys acta 1814 (2011) 785-796. [24] k.h. tiong, n.a.m. yunus, b.c. yiap, e.l. tan, r. ismail, c. e. ong. inhibitory potency of 8methoxypsoralen on cytochrome p450 2a6 (cyp2a6) allelic variants cyp2a6* 15, cyp2a6* 16, cyp2a6* 21 and cyp2a6* 22: differential susceptibility due to different sequence locations of the mutations. plos one 9 (2014) e86230. [25] c.r. lee, j.a. goldstein, j.a. pieper. cytochrome p450 2c9 polymorphisms: a comprehensive review of the in-vitro and human data. pharmacogenet. genomics 12 (2002) 251-263. [26] j. lampe, r. brandman, s. sivaramakrishnan, p.o. de montellano. two-dimensional nmr and allatom molecular dynamics of cytochrome p450 cyp119 reveal hidden conformational substates. j. biol. chem. 13 (2010) 9594-9603. [27] v.n. maiorov, g.m. crippen. significance of root-mean-square deviation in comparing threedimensional structures of globular proteins. j. mol. biol. 2 (1994) 625-634. [28] i. kufareva, r. abagyan. methods of protein structure comparison. methods mol. biol. 857 (2012) 231-257. [29] j.k. yano, m.r. wester, g.a. schoch, k.j. griffin, c.d. stout, e.f. johnson. the structure of human microsomal cytochrome p450 3a4 determined by x-ray crystallography to 2.05-å resolution. j. biol. chem. 279 (2004) 38091-38094. [30] i.f. sevrioukova, t.l. poulos. interaction of human cytochrome p4503a4 with ritonavir analogs. arch. biochem. biophys. 520 (2012) 108-116. quiroga and scior admet & dmpk 7(4) (2019) 252-266 266 [31] m. ekroos, t. sjögren. structural basis for ligand promiscuity in cytochrome p450 3a4. proc. natl. acad. sci. 103 (2006) 13682-13687. [32] e. scott, y. he, m. wester, m. white, c. chin, j. halpert, e. johnson, c. stout. an open conformation of mammalian cytochrome p450 2b4 at 1.6 a resolution. proc. natl. acad. sci. 23 (2003) 1319613201. [33] p. wilderman, m. shah, t. liu, s. li, s. hsu, a. roberts, d. goodlett, q. zhang, v. woods, c. stout, j. halpert. plasticity of cytochrome p450 2b4 as investigated by hydrogen-deuterium exchange mass spectrometry and x-ray crystallography. j. biol. chem. 49 (2010) 38602-38611. [34] f.c. bernstein, t.f. koetzle, g.j. williams, e.f. meyer, m.d. brice, j.r. rodgers, m. tasumi. the protein data bank. eur. j. biochem. 80 (1977) 319-324. [35] w. humphrey, a. dalke, k. schulten. vmd: visual molecular dynamics. j. mol. graph. 14 (1996) 3338. [36] j.c. phillips, r. braun, w. wang, j. gumbart, e. tajkhorshid, e. villa, c. chipot, r.d. skeel, l. kale, k. schulten. scalable molecular dynamics with namd. j. comput. chem. 26 (2005) 1781-1802. [37] b.r. brooks, c.l. brooks, iii, a.d. mackerell, jr., l. nilsson, r.j. petrella, b. roux, y. won, g. archontis, c. bartels, s. boresch, a. caflisch, l. caves, q. cui, a.r. dinner, m. feig, s. fischer, j. gao, m. hodoscek, w. im, k. kuczera, t. lazaridis, j. ma, v. ovchinnikov, e. paci, r.w. pastor, c.b. post, j.z. pu, m. schaefer, b. tidor, r. m. venable, h. l. woodcock, x. wu, w. yang, d.m. york, m. karplus. charmm: the biomolecular simulation program. j comput chem. 30 (2010) 1545–1614. [38] m.r. wester, j.k. yano, g.a. schoch, c. yang, k.j. griffin, c.d. stout, e.f. johnson. the structure of human cytochrome p450 2c9 complexed with flurbiprofen at 2.0-å resolution. j. biol. chem. 279 (2004) 35630-35637. [39] r. raag, t.l. poulos. crystal structure of the carbon monoxide-substrate-cytochrome p-450cam ternary complex. biochemistry 28 (1989) 7586-7592. [40] t.l. poulos, b.c. finzel, a.j. howard. high-resolution crystal structure of cytochrome p450cam. j. mol. biol. 195 (1987) 687-700. ©2019 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ biomimetic properties and estimated in vivo distribution of chloroquine and hydroxy-chloroquine enantiomers doi: https://doi.org/10.5599/admet.929 151 admet & dmpk 9(2) (2021) 151-165; doi: https://dx.doi.org/10.5599/admet.929 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper biomimetic properties and estimated in vivo distribution of chloroquine and hydroxy-chloroquine enantiomers klara l valko* 1 and tong zhang 2 1bio-mimetic chromatography ltd, business & technology centre, bessemer drive, stevenage, herts sg1 2dx uk 2 chiral technologies europe, parc d'innovation 160, bd gonthier d'andernach cs 80140 67404 illkirch cedex france *corresponding author: e-mail: klara_valko@bio-mimetic-chromatography.com; tel.: +44-7521-989558; received: november 16, 2020; revised: january 10, 2021; published online: january 25, 2020 abstract chloroquine and hydroxy-chloroquine already established as anti-malarial and lupus drugs have recently gained renewed attention in the fight against the covid-19 pandemic. bio-mimetic hplc methods have been used to measure the protein and phospholipid binding of the racemic mixtures of the drugs. the tissue binding and volume of distribution of the enantiomers have been estimated. the enantiomers can be separated using chiralpak agp hplc columns. from the α-1-acid-glycoprotein (agp) binding, the lung tissue binding can be estimated for the enantiomers. the drugs have a large volume of distribution, showed strong and stereoselective glycoprotein binding, medium-strong phospholipid-binding indicating only moderate phospholipidotic potential, herg inhibition and promiscuous binding. the drug efficiency of the compounds was estimated to be greater than 2 % which indicates a high level of free biophase concentration relative to dose. the biomimetic properties of the compounds support the well-known tolerability of the drugs. ©2021 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords protein binding; chiral separation; chloroquine and hydroxy-chloroquine; tissue binding; the volume of distribution introduction chloroquine (cq) and hydroxy-chloroquine (hcq) are well established for the treatment of malaria, rheumatoid arthritis, and lupus. cq was first synthesised in 1934 by hans andersag at bayer in elberfeld germany, by modifying the chemical structure of quinine an anti-malarial natural alkaloid [1]. during the second world war, when millions of soldiers were taking cq for malaria prevention, its beneficial effects were discovered in the treatment of rheumatoid arthritis (ra) and systemic lupus erythematosus (sle). when the resistance of cq against malaria plasmids became apparent, the search for analogues resulted in the discovery of hcq in 1945 [1]. the activity of cq against viruses was first reported in 1972. by shimizu et al. [2]. more than ten years ago, cq was proposed as a broad-spectrum antimicrobial agent against bacteria, fungal and viral infections [3]. in december 2019, an outbreak of pneumonia of unknown aetiology was reported in wuhan, china [4,5]. this disease problem was then declared as a pandemic by https://doi.org/10.5599/admet.929 https://dx.doi.org/10.5599/admet.929 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:klara_valko@bio-mimetic-chromatography.com http://creativecommons.org/licenses/by/4.0/ klara valko and tong zhang admet & dmpk 9(2) (2021) 151-165 152 the world health organization (who) on the 11 th march 2020. the coronavirus was identified as the cause of this respiratory syndrome, and it was named sars-cov-2. the disease that it caused was named covid19. since then, attention has been given to the use of cq and hcq as potential treatments for covid-19. the various mechanisms by which cq and hcq may be involved in their antiviral effects have been reviewed recently [6] as a result of increased interest due to the covid-19 pandemic. potential mechanisms include prevention of the virus entering the cells, altering virus replication, inhibiting autophagy and modulating the immune response. an interesting antiviral mechanism of cq and hcq is that they allow zn ions to enter the cell and act as zinc ionophores [7]. zinc ions exhibit anticancer and antiviral properties by altering the lysosome membrane permeability [8] and the viral rna-dependent rna polymerase enzyme and thus inhibit the virus replication process inside the cells. figure 1 shows the chemical structures of cq and hcq. hcq has an additional hydroxyl group that makes the molecule less lipophilic, reducing the severity of side effects and toxicity. chloroquine (cq) hydroxy-chloroquine (hcq) figure 1. the chemical structure of chloroquine (cq) and hydroxy-chloroquine (hcq) both compounds are racemic mixtures, but it has been observed that the enantiomers behave differently in vivo regarding their pharmacokinetics [9] and metabolism [10]. it has been reported that the desethyl cq excreted by men after treatment with racemic cq is optically active [11] and that the renal clearance of the racemic cq drug is a stereoselective process [9,12,13]. in vitro efficacy studies did not show any difference between the efficacy of the enantiomers [14], but in vivo (rat) studies have shown a greater efficacy for the (+)-(s)-cq than the (-)-r-enantiomer [13,15,16]. more recent article [17] suggests chiral switches of enantiomers as more potent forms against covid-19. it may be explained by the stereoselective disposition of cq and hcq in vivo. several studies have already been published on the stereoselective blood and plasma protein binding of the two compounds using pure enantiomers [18-20] added to plasma and blood. the application of chemically bonded chiral stationary phases (chiralpak hsa and chiralpak agp) is a powerful tool for separating racemic mixtures [20]. it was found that the cq enantiomers differ in their binding to plasma. the (+)-(s)-cq has a higher binding to human serum albumin (hsa) than the (-)-(r)enantiomer, whereas the latter bound more strongly to α-1-acid-glycoprotein (agp) than the (+)-(s)-cq. the in vitro human plasma protein binding of hcq was also enantioselective: the binding of (+)-(s)-hcq was almost 2-fold higher than (-)-(r)-hcq (64 % against 37 %) [21]. chromatographic separations of cq and hcq and their metabolites have been reported using chiralpak agp and chiralpak ad hplc columns [18,22], [23]. other chromatographic methods have been reported for analytical and preparative separations of cq and hcq enantiomers [24,25]. biomimetic hplc properties can be used to estimate the in vivo distribution (volume of distribution, vd, the unbound volume of distribution, vdu), drug efficiency and various tissue binding properties of admet & dmpk 9(2) (2021) 151-165 biomimetic properties of chlorquine enantiomers doi: https://doi.org/10.5599/admet.929 153 compounds [26-28]. the hsa and agp binding can be measured using chiralpak hsa and chiralpak agp hplc columns (chiral technologies europe, france). in this way, a different estimation of the in vivo distribution and tissue binding for the different enantiomers can be obtained supposing they provided two requisite peaks from the chromatography, enabling property measurements. the aim of this study was to measure the biomimetic properties of cq and hcq with special emphasis on the different properties of their enantiomers. the other aim was to rationalise the differences in the in vivo potency of the enantiomers that were not detected in vitro. a comparison can then be made of the estimated tissue binding and distribution data with the published in vivo distribution data. the significance of the differences in properties between the (r) and (s) enantiomers can then be evaluated. experimental racemic cq and hcq of the highest purity were purchased from merck (sigma-aldrich) and were dissolved in dmso at a 1 mm concentration. the 1 mm solutions were diluted further to 0.1 mm concentration from which 10 µl was injected onto the different hplc columns. an agilent 1100 series hplc system equipped with an autosampler and diode array detector was used for the measurements. measurements of lipophilicity at three phs gemini nx c-18(2) 5 um columns with dimensions of 3 x 50 mm (phenomenex ltd. macclesfield, uk) were used for the determination of the chromatographic hydrophobicity index (chi) and the chilogd values [26]. the flow rate was 1.0 ml/min, with starting mobile phases of 0.01 m formic acid (ph 2.6), 50 mm ammonium acetate adjusted to ph 7.4, and 50 mm ammonium acetate with ph adjusted to 10.5 to determine the lipophilicity of the compounds at acidic, neutral and alkaline phs, respectively. an acetonitrile linear gradient was used from 0 to 100 %. the acetonitrile concentration reached 100 % in 3.5 min. the 100 % acetonitrile mobile phase was maintained for an additional 1 min before it was returned to 0 % at 4.7 min. the cycle time of the gradient run was 6 min, with an additional equilibration time of 1 min before the next injection. the error in the retention time measurements is ±0.005 min in general after repeated injections. the set of calibration compounds gave an excellent straight line when their retention time data was plotted against the predefined chi values listed in table 1. table 1. the chi values of the calibration set of compounds at three phs [29]. these values were obtained by fitting the isocratically determined chi values and the gradient retention time values. the standard error ranged from 0.1 to 0.8 chi values. the chromatographic hydrophobicity index (chi) approximates the acetonitrile concentration when the compound elutes. it can be converted to the octanol/water log d scale using chilogd = 0.0525*chi -1.467. compound chi at ph 2.6 chi at ph 7.4 chi at ph 10.5 theophylline 17.9 18.4 5.0 phenyl tetrazole 42.2 23.6 16.0 benzimidazole 6.3 34.3 30.6 colchicine 43.9 45.0 43.9 phenyl theophylline 51.7 51.2 51.3 acetophenone 64.1 65.1 64.1 indole 72.1 71.5 72.1 propiophenone 77.4 77.4 77.4 butyrophenone 87.3 87.5 87.3 valerophenone 96.4 96.20 96.36 https://doi.org/10.5599/admet.929 klara valko and tong zhang admet & dmpk 9(2) (2021) 151-165 154 measurements of phospholipid-binding at ph 7.4 using an immobilized artificial membrane (iam) column the phospholipid-binding was measured using an iam pc.dd2 column with dimensions of 100 x 4.6 mm (regis technologies inc., morton grove, il, usa). the gradient retention times of the compounds were measured using a 50 mm ammonium acetate starting mobile phase with the ph adjusted to 7.4. the mobile phase flow rate was 1.5 ml/min. the acetonitrile gradient was applied from 0 min to 4.75 min to reach 90 %. the 90 % acetonitrile concentration was maintained for an additional 0.5 min (to 5.25 min), then returned to 0 % by 5.5 min. the cycle time was 6 min plus an additional 1 min equilibration time was applied while the injector prepared for the next injection. the gradient retention times were calibrated with the acetophenone homologues for which the chi iam values have been established using isocratic measurements [26]. the chromatographic hydrophobicity index on the iam column [chi(iam)] approximates to the acetonitrile concentration in the mobile phase when the compound elutes. chi(iam) values above 45 indicate strong phospholipid binding. repeating the retention time measurements provided the error only in the third digit of the minute (±0.005). measurements of plasma protein binding using chiralpak hsa and agp columns the protein binding measurements were carried out on chiralpak hsa and chiralpak agp columns with the dimensions of 3 x 50 mm with 5 um particle size stationary phase (chiril technologies europe, france). the mobile phase was 50 mm ammonium acetate adjusted to ph 7.4 with a flow rate of 1.2 ml/min. the standard isopropanol (ipa) gradient up to 35 % was achieved in 3.5 min and then maintained for 1 min, before returning to 0 % at 4.7 min. the cycle time was 6 min with an additional 1 min re-equilibration time. the racemic warfarin showed separation of its enantiomers at retention times of 3.58 and 3.77 mins. the reproducibility of the retention time measurements was within ±0.01 min. the calibration set of compounds and their literature % binding data that were also converted to log k data are shown in table 2. the racemic cq and hcq, however, did not show a separation of their enantiomers under such chromatographic conditions and therefore, other gradients and isocratic mobile phases were tested. figures 2a and figure 2b show the calibration lines obtained for the hsa and agp binding, respectively. the 8 min, “slow gradient” method on the chiral hsa and agp columns to be able to separate the cq and hcq enantiomers, the ipa gradient was modified. from 0 to 4 min, the ipa concentration was raised from 0 to 15 %, and from 4 to 5 min the 15 % ipa was kept constant before it was reduced to 0 % at 5.2 min. the cycle time was 8 min plus an additional 1 min re-equilibration time. the same compounds and data that are listed in table 2 were used to calibrate the retention times; however, warfarin, diclofenac and isopropanol could not be eluted under this condition. table 2. the protein binding data of marketed drug molecules that were used to calibrate the retention times obtained on the chiral protein columns (chiralpak hsa and chiralpak agp). the % binding data obtained by equilibrium dialysis were converted to log k data using log k = log (%binding/(101-%binding). compound name % hsa log k (hsa) retention times hsa (tr) % agp log k (agp) retention times agp (tr) warfarin 97.9 1.5 3.6 83.2 0.7 2.6 paracetamol 14.0 -0.8 0.4 3.2 -1.5 0.3 nizatidine 20.4 -0.6 0.5 37.1 -0.2 1.3 trimethoprim 37.6 -0.2 1.0 46.2 -0.1 1.5 propranolol 66.6 0.3 1.4 86 0.8 2.8 carbamazepine 75.0 0.5 1.6 65 0.3 2.1 nicardipine 95.0 1.2 2.9 87 0.8 2.9 indomethacin 99.5 1.8 4.5 56 0.1 1.8 diclofenac 99.8 1.9 4.3 60 0.2 1.9 admet & dmpk 9(2) (2021) 151-165 biomimetic properties of chlorquine enantiomers doi: https://doi.org/10.5599/admet.929 155 figure 2. the calibration line obtained on the chiralpak hsa (a) and chiralpak agp (b) columns by plotting the logarithmic value of the obtained retention times and the log k values derived from the literature % binding data [30,31] isocratic measurements of cq and hq on chiralpak hsa column as the cq and hcq enantiomers were not separated using the longer ipa gradient, isocratic analyses were also carried out using a 1.2 ml/min flow rate with 5, 8 and 10 % ipa, respectively. using 5 % ipa in the mobile phase, only paracetamol and propranolol eluted from the calibration set of compounds. when 10 % ipa was used warfarin, propranolol and nicardipine were eluted and their respective retention times were used to create a calibration curve for the calculation of the binding of cq and hcq from their obtained retention times using the same hplc conditions. measurements of human plasma protein binding by rapid equilibrium dialysis (red) the test compounds were prepared in 100 % pooled human plasma. the measurement was performed using equilibrium dialysis with the two compartments separated by a semi-permeable membrane. the plasma solution was added to one side of the membrane while buffer (ph 7.4) was added to the other side. the system was allowed to reach equilibrium at 37 °c. compound concentration on both sides of the membrane was measured by lc-ms/ms, and the fraction of unbound compound calculated. calibration standards were prepared in plasma and buffer. test compound incubations were performed in triplicate. a control compound was included in each experiment. the solutions for each batch of compounds were combined into two groups (protein-free and protein-containing). cassette analysed by lc-ms/ms using two sets of calibration standards for protein-free solutions (7 points) and protein-containing solutions (7 points). the experiments were carried out at cyprotex discovery ltd (uk) and repeated at sygnature discovery ltd (uk) laboratories using their generic lc/ms method. the unbound fraction (fu) is calculated from the concentration of the test compound in the protein-containing side cp and the protein-free side cf based on the formula: fu= 1 – (cp-cf)/cp . results and discussion the lipophilicity of compounds is an important physicochemical property that may explain the protein binding, the volume of distribution and other distribution properties of drug molecules. it is usually characterised by octanol/water partition (log p) and distribution (log d) coefficients. the log p refers to the neutral form of the molecule, while log d refers to the compound’s distribution between octanol and water at physiological ph (ph 7.4). the calculated and measured physicochemical properties of cq and hcq are listed in table 3. (a) (b) https://doi.org/10.5599/admet.929 klara valko and tong zhang admet & dmpk 9(2) (2021) 151-165 156 table 3. literature and calculated data for cq and hcq it can be seen in table 3 that the measured and calculated lipophilicity values vary a lot. in almost all cases, hcq is less lipophilic due to the hydrophilic hydroxy group that also affects the basicity of this compound. comparing the log d and log p values, it is apparent that a high percentage of both molecules are ionised (positively charged) at ph 7.4. the measured chi log p may well be underestimating the lipophilicity as the pka values are around 10. therefore, the mobile phase applied ph 10.5 may not suppress the ionisation of both molecules completely, and the actual lipophilicity of the neutral form of these molecules can be much greater. table 4 shows the measured phospholipid and protein binding data of the two compounds obtained using the standard 6 min protocol [26]. it can be seen from table 4 that hcq has marginally weaker binding to phospholipids and plasma proteins. the enantiomers could not be separated on the chiral hsa and agp stationary phases using the standard high-throughput protocol for biomimetic property measurements. the phospholipid-binding is in a medium-strong range. it has been reported that compounds that have chi iam values greater than 50 have the potential for promiscuous binding and causing phospholipidosis [34]. the herg inhibition, which is an indicator for causing arrhythmia and heart condition, is also associated with lipophilic and positively charged compounds [35]. positively charged compounds usually have stronger binding to iam than isolipophilic compounds without the presence of positive charge [36]. however, the chi iam values of cq and hcq are below 50. the measured protein binding was greater than has been published [19], ranging from 46 to 74 % in patients with rheumatoid arthritis. strong agp binding was also observed, being approximately the same as the albumin binding. table 4. the measured phospholipid binding (chi iam), protein binding and estimated total plasma protein binding (ppb) based on the % hsa and % agp binding and the lipophilicity of the compounds. the error of the values depends on the model error and the error obtained from fitting the calibration lines. the chi iam values are reproducible with ±3 chi unit, the protein binding % error is larger at the middle range (± 10 %) and much smaller above 90 % binding (±0.1 %), the estimation of the total plasma protein binding includes the model error of ± 0.36 in log k ppb values. property cq hcq chi (iam) 47 45.4 log k (iam) 2.6 2.5 % hsa bound 87.6 84.9 log k hsa 0.8 0.7 % agp bound 88.8 89.0 log k agp 0.9 0.9 estimated % plasma protein binding (ppb) 94.8 93.4 log k ppb 1.2 1.1 however, the agp concentration in the plasma ranges from 1 to 10 % while the albumin concentration in the plasma is around 60 %, being the most abundant protein. it is also well established that agp prefers to bind to positively charged compounds since it has negatively charged sialic acids at its binding site [37]. the lysosomes inside the cell also contain glycoproteins that are negatively charged and attract positively property cq hcq acd log p (calculated) 4.69 3.77 chemaxon log p (calculated) 3.93 2.89 acd log d (calculated) 1.74 1.96 chemaxon log d (calculated) 0.88 0.33 measured log i (chemspider [32, 33]) 2.68 -4.63 1.55 – 3.85 chi log d (ph 7.4) 0.79 0.66 chi log p 2.68 2.13 chemaxon pks (calculated) 10.32 9.76 admet & dmpk 9(2) (2021) 151-165 biomimetic properties of chlorquine enantiomers doi: https://doi.org/10.5599/admet.929 157 charged compounds. the strong agp binding of cq and hcq may explain the large variability of the plasma protein binding data as the agp concentration depends on the disease state and it can increase significantly in inflammation, cancer and with age [38]. table 5 shows the equations of the models that were used to make estimations of the in vivo properties of the drugs from the measured biomimetic hplc data. note, that the log k hsa and log k iam data had to be transformed to log k hsa and log k iam data as the measured log k data for acetophenone homologues up to octanophenone did not show linear correlation with their octanol water partition coefficients (log p) [39]. for the mechanistic model of the volume of distribution, we converted both the measured albumin and phospholipid binding data so that they showed linear correlation with the linear free-energy related log p values. thus, the regression coefficients in linear regression analysis are meaningful and reflect to the physiological proportion of the albumin and phospholipids in the body. the estimated distribution and tissue binding data for cq and hcq were evaluated using the standard 6-min hplc based measurements of biomimetic properties (chilogd at three phs, log k hsa, log k agp and log k iam). they are based on the calibrated retention times obtained on c-18, iam, chiralpak hsa and chiralpak agp columns. the total plasma protein binding, lung tissue binding, drug efficiency, brain tissue binding and brain to blood ratio, cell partition coefficient and lung tissue binding have been calculated using the equations listed in table 5 and are shown in table 6. table 5. the published and validated model equations that are used together with the measured properties to estimate the in vivo distribution behaviour of drugs. (abbreviations: iam = immobilized artificial membrane; hsa = human serum albumin; vd = volume of distribution; vdu = unbound volume of distribution, demax = maximum drug efficiency; btb = brain tissue binding; ppb = plasma protein binding, fu = unbound fraction; kbb = brain to blood partition coefficient; kpcell = cell partition coefficient; ltb = lung tissue binding) log k iam [43] = 0.29*e(0.026chi(iam)+0.42) +0.7 log k iam [43] = 0.046*chi(iam)+0.42 phospholipidotic potential (cadlikeness) = chi (iam) + (chi at ph 10.5 – chi at ph 7.4) log k hsa [43] = elog k(hsa) log k hsa [30] = log (%hsabound/(101%hsa bound)) estimated log vd [43,44] = 0.44*log k iam -0.22*log k hsa – 0.62 estimated log vdu[27] = 0.23*log k hsa +0.43*log k iam -0.72 demax [45] = 100/vdu log k btb [31] = 1.29*log k iam+1.03*log k hsa-2.37 log k (ppb)[31] = 0.98∗log 𝑘 𝐻𝑆𝐴+0.19∗log 𝑘 𝐴𝐺𝑃+0.031∗chi log d 7.4−0.20 %btb [31] = 100*10log k btb/(1+10log k btb) %ppb [31] = 100*10log k ppb/(1+10log k ppb) fu btb and ppb [31] = (100-%btb)/100 and (100 %-%ppb)/100 kbb [31] = fuppb/fubtb log kpcell [46] = 1.1log k iam-1.9 log k ltb [31] = 0.49*log k ppb+0.34chilog d-0.069 % ltb [31] = 100*10 logkltb /(1+10 logkltb ) it can be seen from table 6 that both compounds have cad-likeness values greater than 50, which indicates that both have phospholipidotic potential. it is known that chloroquine causes phospholipidosis both in animals and in human [40-42]. however, the phospholipidotic potential of hcq is much less than cq. although the phospholipid-bindings (chi iam) were not very different, the greater difference in cadlikeness values is due to the larger percentage of the positively charged form of cq than in hcq. it is obvious from the lower acid dissociation constant of hcq and the smaller difference between the chilogp https://doi.org/10.5599/admet.929 klara valko and tong zhang admet & dmpk 9(2) (2021) 151-165 158 and chilogd values (see table 3). it also implies that hcq is less likely to cause herg inhibition and promiscuous binding and in general, fewer side effects. based on the measured data, the expected brain tissue binding did not show a significant difference between cq and hcq. on the contrary, the brain to blood ratio is smaller in hcq than cq due to the slightly stronger agp binding of hcq that keeps the compound in the plasma compartment. table 6. the estimated biomimetic properties of cq and hcq interestingly the expected volume of distribution based on the measured biomimetic properties is 6.9 and 5.8 l/kg for cq and hcq, respectively, which is not as high as over 100 l/kg as has been described in the literature [16,4749]. the vd of cq was 116 to 285 l/kg in healthy volunteers while hcq was reported to have a much higher volume of distribution at above 4000 l/kg [48]. tett et al. [50], reported a range of steady-state vd values for hcq ranging from 2402 to 8346 l that are equivalent to 34.3 to119 l/kg. the volume of distribution was 5 to 7 times greater when the plasma concentration was considered instead of the blood concentration as the blood/plasma ratio was reported to be 5 to 7. the reported plasma protein binding value that was determined in 1986 was 66.6 % [20]; enantioselective plasma protein binding, albumin and agp binding were also reported. the human plasma protein binding measurements were repeated using the rapid equilibrium dialysis method and the obtained average values were 52.6 % (±6.2 %) and 30.7 % (±11.3 %) for cq and hcq, respectively. the estimated plasma protein binding based on the biomimetic hplc measurements was much stronger (95 % and 93 % for cq and hcq, respectively) than that published using in vivo data. at this stage, there is no explanation for the difference. this is the cause of the much lower estimated volume of distribution. the agp binding using equilibrium dialysis method was reported to be 48.4 % (59.9 and 34.9 % of the r and s enantiomers, respectively) for cq and 34 % (29 and 41 % for the r and s enantiomer, respectively) for hcq [51]. the albumin binding was only 38.6 and 40 % for cq and hcq, respectively, by the same author [51]. these values were quoted from earlier publications where the actual measurements were originally reported [52-54]. it was observed that under the standard ipa gradient conditions, both cq and hcq eluted between carbamazepine and nicardipine, being closer to nicardipine, which binds to plasma proteins at 75 and 95 %, respectively. this explains why the hsa binding was calculated as being 87 and 83 % (see table 3). the demax values reveal the free plasma concentration relative to the dose. most marketed drug molecules have demax values of around 1 %. it can be seen that both cq and hcq have greater than 1 % drug efficiency, which indicates that a high free concentration can be achieved with a relatively low dose. the estimated lung tissue binding for both cq and hcq was weaker than their brain tissue binding. as mucus binding correlates very well with the agp binding, strong mucus binding can be expected for both compounds. the cellular partition coefficients of both compounds are high (kpcell above 7), which means a 7 to 8 times greater concentration than the intracellular concentration can be expected for both compounds. unfortunately, our standard fast gradient protocols used for the hsa and agp binding measurements did not allow the separation of the enantiomers of cq or hcq. this is also an indication that there are no significant differences in the biomimetic properties of each respective enantiomer. therefore, new hplc property cq hcq phospholipidotic potential (cad-likeness) 82.8 73.6 %btb 98.4 97.6 est log k bb 0.9 0.8 brain to plasma 3.4 2.8 vd (l/kg) 6.9 5.1 vdu (l/kg) 56.1 38.8 demax% 1.8 2.6 % ltb 96.3 93.9 kpcell 8.7 7.3 admet & dmpk 9(2) (2021) 151-165 biomimetic properties of chlorquine enantiomers doi: https://doi.org/10.5599/admet.929 159 conditions were evaluated on the chiralpak hsa and chiralpak agp column by increasing the gradient analysis time and reducing the maximum ipa concentration in the mobile phase. slowing down the gradient to reach 10 and 20 % ipa did not result in the separation of the enantiomers on the hsa column, and both cq and hcq eluted after carbamazepine (75 % plasma protein binding). the calculated hsa binding was 82.2 and 81.9 % for cq and hcq, respectively. isocratic measurements using 5, 8 and 15 % ipa in the mobile phase resulted in 80 % binding when the relative retention times obtained using a partial set of calibration compounds were taken into consideration. no separation of the enantiomers was observed. figure 3 shows the chromatograms of cq and hcq obtained on a chiralpak hsa column using %ipa in isocratic mode. however, no separation of enantiomers was observed. figure 3. the chromatograms of cq (a) and hcq (b) on chiralpak hsa column using 0 % ipa in the mobile phase in isocratic mode. when similar changes in the mobile phases were applied using the chiralpak agp column, excellent separations of both the cq and hcq enantiomers were observed albeit with a very short column size (l=50 mm) and fast flow rate (1.2 ml/min). it can be seen in figure 4 that the separation of enantiomers occurred under the shallower gradient conditions. still, no significant differences between the retention times were observed for cq and hcq confirming that the agp binding of the two molecules is quite similar. the chiral separations were also obtained in the isocratic mode. however, the difference in binding between the enantiomers was only 3 to 4 %. based on these measurements, the agp binding is much stronger for both cq and hcq than previously reported. when the agp binding is above 80 %, the compounds are expected to have variable efficacy in the clinic. it is because the agp concentration changes in various disease states and it can alter the total plasma protein binding significantly, and that may cause a reduction of the free therapeutic concentration of the drugs. figure 4. the chromatogram of racemic chloroquine (cq) (a) and hydroxychloroquine (hcq) (b) obtained using the shallower gradient method (“slow gradient”) with 0 to 15 % ipa over 4.0 min and then maintained at 15 % for 1 min before returning to 0 % ipa on a chiralpak agp column (50 x 3 mm) with a 1.2 ml/min flow rate. the enantiomers are identified based on ref [51] that reported cq (r) and hcq (s) as the stronger binders to agp. (a) (b) (a) (b) https://doi.org/10.5599/admet.929 klara valko and tong zhang admet & dmpk 9(2) (2021) 151-165 160 table 7 shows the % binding data obtained using a slow gradient (up to 15 % ipa). the protein binding measurements were repeated using isocratic conditions with 5, 8 and 15 % ipa in the mobile phase. under such conditions, stronger binding compounds from the calibration mixture (warfarin, indomethacin and diclofenac) could not be eluted. the retention times of cq and hcq obtained when using 5, 8 and 15 % ipa on chiralpak hsa and chiralpak agp columns are listed in table 8. table 7. the measured and estimated biomimetic properties of cq and hcq based on the data obtained using the slower and shallower gradient (up to 15 % ipa). property (+)-s cq (-)-r cq (+)-s hcq (-)-r hcq % bound hsa 82.2 82.2 81.9 81.9 % bound agp 85.0 88.4 88.5 84.3 calc%ppb 90.6 91.1 91.2 91.7 chi log d2 -1.2 -1.2 -1.3 -1.3 chi log d7.4 0.8 0.8 0.7 0.7 chi log d10.5 2.7 2.7 2.1 2.1 chilog p 2.7 2.7 2.1 2.1 chi iam 46.9 46.9 45.4 45.4 cad-likeness 82.9 82.9 73.6 7367 %btb 97.6 97.6 97.1 97.0 est log k bb 0.9 0.9 0.9 0.8 brain to plasma 4.0 3.8 3.0 2.8 vd l/kg 8.2 8.2 6.4 6.3 vdu 45.7 45.8 35.3 35.1 demax% 2.2 2.2 2.8 2.9 % lung tissue binding 95.5 95.6 93.1 93.3 table 8. the retention times obtained for cq and hcq (enantiomers) using a 1.2 ml/min flow rate and 5, 8 and 15 % ipa on chiralpak hsa and chiralpak agp columns. the calculated % binding data are also given in brackets. the retention times of propranolol and nicardipine, the two basic compounds from the calibration set, are also shown. the last two rows in the table are the data from reference [51] where the values were obtained using an equilibrium dialysis method (n/a means not applicable) compound/cond. cq (r) cq (s) hcq (r) hcq (s) propranolol (66 % ppb) nicardipine (96 % ppb) hsa 5 % ipa (min) 0.89 (60 %) 1.15 (68 %) 0.94 (61 %) 1.4 (73 %) 0.76 (54 %) 6 (96 %) hsa 8 % ipa (min) 0.87 (64 %) 0.87 (64 %) 0.91 (66 %) 0.91 (66 %) 0.73 (57.6 %) 3.39 (95 %) hsa 15 % ipa (min) 0.68 (80 %) 0.68 (80 %) 0.69 (80 %) 0.69 (80 %) 0.58 (66 %) 1.05 (95 %) agp 5 % ipa (min) 3.84 (86 %) 5.24 (90 %) 3.84 (86 %) 5.4 (90 %) 2.78 (85 %) >8 min (>98 %) agp 8 % ipa (min) 2.53 (84 %) 2.55 (84 %) 2.95 (85 %) 3.85 (87 %) 1.73 (68 %) 5.63 (98 %) agp 15 % ipa (min) 2.20 (82 %) 2.00 (81 %) 1.98 (81 %) 2.23 (83 %) 1.58 (67 %) 5.0 (96 %) % hsa bound [51] 30 % 54 % 50 % 29 % n/a n//a % agp bound [51] 59 % 35 % 29 % 34 % n/a n/a it can be seen in table 8 that the estimated % binding data using the retention time measured for the partial calibration set of compounds changes depending on the experimental conditions. the hsa binding ranged from 60 to 80 % but generally was much greater than the published albumin (hsa) and alpha-1-acid glycoprotein (agp) binding data. this is because the retention times dependency on the ipa concentration was greater for the calibration set of compounds than for the cq and hcq. table 9 shows the measured admet & dmpk 9(2) (2021) 151-165 biomimetic properties of chlorquine enantiomers doi: https://doi.org/10.5599/admet.929 161 and estimated biomimetic properties of cq and hcq based on the data obtained when using the isocratic method on the chiralpak hsa and chiralpak agp column (ipa concentration 15 %). probably, cq and hcq do not bind to the hydrophobic site on the albumin but could bind to a specific area that is not affected strongly by the ipa concentration. the agp binding was also much stronger using the biomimetic methodology that converts the % binding to the total plasma protein concentration where the agp concentration is only 1 to 3 %. the difference can be explained by the different scaling used when expressing the % of the compound bound in plasma or to a single protein. however, when using the lowest ipa concentration in the mobile phase (5 %), the data obtained was between 77 and 87 % for the enantiomers of cq and hcq. the in vivo pk data also showed large individual variations in humans which suggests that these compounds may bind to a particular type of negatively charged proteins in plasma and in tissues that may have large concentration variations. table 9. the estimated in vivo properties of cq and hcq enantiomers obtained from the data when the protein binding was obtained with the 15 % ipa isocratic conditions. property (+)s cq (-)r cq (+)s hcq (-)-r hcq % bound hsa 79.9 79.9 79.9 79.9 % bound agp 82.0 81.0 85.1 87.3 calc% ppb 89.9 89.7 89.0 88.9 chi logd2 -1.2 -1.2 -1.3 -1.3 chi logd7.4 0.8 0.8 0.7 0.7 chi logd10.5 2.7 2.7 2.1 2.1 chilogp 2.68 2.7 2.1 2.1 chi iam 46.92 46.9 45.4 45.4 cad-likeness 82.89 82.9 73.6 73.6 %btb 97.26 97.3 96.7 96.7 est log k bb 0.96 1.0 0.9 0.9 brain to plasma 4.06 4.0 3.1 3.0 log vd 0.94 0.9 0.8 0.8 vd l/kg 8.65 8.7 6.7 6.6 log vdu 1.63 1.6 1.5 1.5 vdu 43.01 43.0 33.5 33.3 demax% 2.33 2.3 23.0 3.0 % lung tissue binding 96.05 95.1 92.6 92.7 it can be seen that the estimated distribution data shown in table 7 and table 9 are very similar, and do not differ significantly despite the large differences in the % protein binding data. it is important to note that the % binding data are not linearly related to the protein binding partition coefficients. still, they have a sigmoidal relationship with the log k and logarithmic retention time data. this means that small variations in the log k hsa and log k agp values result in a large variation when expressed as percentage binding in the range of 30 to 80 %. in the stronger binding range, large variations in retention times will result in only a small per cent of change. the biomimetic protein binding data have been validated for differentiating between strong binders and are more accurate in that region [30]. conflict of interest: klara valko is the founder of biomimetic chromatography ltd and tong chang is a scientist at chiral technologies europe. conclusions cq and hcq have gained renewed interest as potential cures for covid-19 when used in combination with antibiotics and zinc. biomimetic properties derived from biomimetic hplc stationary phases have https://doi.org/10.5599/admet.929 klara valko and tong zhang admet & dmpk 9(2) (2021) 151-165 162 never been measured for these compounds. cq and hcq did not show extreme phospholipid-binding that would give the compounds a negative toxicity profile. cq and hcq bind to agp very strongly and stereoselectively, though there is no significant difference between cq and hcq in this regard. the stereoselective binding to agp did not result in significant differences in the estimated total plasma protein binding, lung tissue binding and volume of distribution. the retention time measurements are very reproducible on the biomimetic columns, but the absolute values of the calculated binding data originate from the error of the calibration line fit. however, the order of the retention data is very reproducible. the in vivo distribution models include 0.5 to 1 log unit errors but again the rank order of the compounds is reproducible. the comparison between the estimated distribution properties and the properties published in the literature show great variations, however the biomimetic protein binding data were greater than the measured plasma protein binding data obtained by equilibrium dialysis. the aim of the study was to differentiate the binding of the various enantiomers and estimate the in vivo distribution of each enantiomer. it was found that the small differences in the protein binding data did not manifest itself in significant differences in the predicted in vivo distribution data, so this cannot explain the observed differences in the in vivo activity of the enantiomers. it was found that the drugs bind to glycoproteins that are present in humans in variable amounts from various disease states, which may explain the big variations observed in the in vivo distribution and protein binding data. references [1] k. krafts, e. hempelmann, a. skórska-stania. from methylene blue to chloroquine: a brief review of the development of an antimalarial therapy. parasitol. res. 111 (2012) 1–6. doi: https://doi.org/10.1007/s00436-012-2886-x. [2] y. shimizu, s. yamamoto, m. homma, n. ishida. effect of chloroquine on the growth of animal viruses. arch. gesamte virusforsch. 36 (1972) 93–104. doi: https://doi.org/10.1007/bf01250299. [3] j.m. rolain, p. colson, d. raoult. recycling of chloroquine and its hydroxyl analogue to face bacterial, fungal and viral infections in the 21st century. int. j. antimicrob. agents 30 (2007) 297–308. doi: https://doi.org/10.1016/j.ijantimicag.2007.05.015. [4] h. lu, c.w. stratton, y.w. tang. outbreak of pneumonia of unknown etiology in wuhan, china: the mystery and the miracle. j. med. virol. 92 (2020) 401–402. doi: https://doi.org/10.1002/jmv.25678. [5] r. lu, x. zhao, j. li, p. niu, b. yang, h. wu, w. wang, h. song, b. huang, n. zhu, y. bi, x. ma, f. zhan, l. wang, t. hu, h. zhou, z. hu, w. zhou, l. zhao, j. chen, y. meng, j. wang, y. lin, j. yuan, z. xie, j. ma, w.j. liu, d. wang, w. xu, e.c. holmes, g.f. gao, g. wu, w. chen, w. shi, w. tan. genomic characterisation and epidemiology of 2019 novel coronavirus: implications for virus origins and receptor binding. lancet 395 (2020) 565–574. doi: https://doi.org/10.1016/s0140-6736(20)30251-8. [6] t.j. oscanoa, r. romero-ortuno, a. carvajal, a. savarino. a pharmacological perspective of chloroquine in sars-cov-2 infection. int. j. antimicrob. agents 56 (2020) 106078. doi: https://doi.org/10.1016/j.ijantimicag.2020.106078. [7] j. xue, a. moyer, b. peng, j. wu, b.n. hannafon, w.q. ding. chloroquine is a zinc ionophore. plos one 9 (2014) 1–6. doi: https://doi.org/10.1371/journal.pone.0109180. [8] j. zheng, x.x. zhang, h. yu, j.e. taggart, w.q. ding. zinc at cytotoxic concentrations affects posttranscriptional events of gene expression in cancer cells. cell. physiol. biochem. 29 (2012) 181– 188. doi: https://doi.org/10.1159/000337599. [9] h. fieger, j. iredale, i.w. wainer. enantioselective determination of hydroxychloroquine and its major metabolites in urine and the observation of a reversal in the (+)/(−)‐hydroxychloroquine ratio. chirality 5 (1993) 65–70. doi: https://doi.org/10.1002/chir.530050205. [10] y. wei, g.a. nygard, s.l. ellertson, s.k.w. khalil. stereoselective disposition of hydroxychloroquine and its metabolites in rats. chirality 7 (1995) 598–604. doi: https://doi.org/10.1002/chir.530070807. https://doi.org/10.1007/s00436-012-2886-x https://doi.org/10.1007/s00436-012-2886-x https://doi.org/10.1007/bf01250299 https://doi.org/10.1016/j.ijantimicag.2007.05.015 https://doi.org/10.1002/jmv.25678 https://doi.org/10.1016/s0140-6736(20)30251-8 https://doi.org/10.1016/j.ijantimicag.2020.106078 https://doi.org/10.1371/journal.pone.0109180 https://doi.org/10.1159/000337599 https://doi.org/10.1002/chir.530050205 https://doi.org/10.1002/chir.530070807 admet & dmpk 9(2) (2021) 151-165 biomimetic properties of chlorquine enantiomers doi: https://doi.org/10.5599/admet.929 163 [11] c.m. eldefield e.o. titus, l.c. craig, c. golumbic, h. r. highton. identification by distribution studies; application to metabolic studies of 4-aminoquinoline antimalarial. j. org. chem. 13 (1948) 39–62. [12] h.s.d. ofori-adjei, o. ericson, b. lindström, j. hermansson, k. adjepon-yamoah. enantioslective analysis of chloroquine and desethylchloroquine after oral administration of racemic chloroquine. ther. drug monit. 8 (1986) 457–461. https://doi.org/10.1097/00007691-198612000-00014. [13] j. ducharme, h. fieger, m. ducharme, s. khalil, i. wainer. enantioselective disposition of hydroxychloroquine after a single oral dose of the racemate to healthy subjects. br. j. clin. pharmacol. 40 (1995) 127–133. doi: https://doi.org/10.1111/j.1365-2125.1995.tb05768.x. [14] s. fu, a. björkman, d. ofori-adjei, b. wählin, o. ericsson, f. sjöqvist. in vitro activity of chloroquine, two enantiomers of chloroquine, desethylchloroquine and pyronaridine against plasmodium falciparum, br. j. clin. pharmac. 22 (1986) 93-96. [15] g.b.a. haberkorn, h.p. kraft, g. blaschke. antimalarial activity in animals of the optical isomers of chloroquine diphosphate. tropenmed parasitol. 30 (1979) 308–312. [16] r.f.j. ducharme. clinical pharmacokinetics and metabolism of chloroquine. clin. pharmacokinet. 31 (2012) 2570274. [17] i. d’acquarica, i. agranat. chiral switches of chloroquine and hydroxychloroquine: potential drugs to treat covid-19. drug discov. today 25 (2020) 1121–1123. doi: https://doi.org/10.1016/j.drudis.2020.04.021. [18] j. iredale, i.w. wainer. determination of hydroxychloroquine and its major metabolites in plasma using sequential achiral-chiral high-performance liquid chromatography. j. chromatogr. b biomed. sci. appl. 573 (1992) 253–258. doi: https://doi.org/10.1016/0378-4347(92)80126-b. [19] o. walker, d. birkett, g. alvan, l. gustafsson, f. sjoqvist. characterization of chloroquine plasma protein binding in man. br. j. clin. pharmacol. 15 (1983) 375–377. doi: https://doi.org/10.1111/j.1365-2125.1983.tb01513.x. [20] d. ofori‐adjei, o. ericsson, b. lindstrom, f. sjoqvist. protein binding of chloroquine enantiomers and desethylchloroquine. br. j. clin. pharmacol. 22 (1986) 356–358. doi: https://doi.org/10.1111/j.13652125.1986.tb02900.x. [21] a. mclachlan, s. tett, d. cutler, r. day. disposition of the enantiomers of hydroxychloroquine in patients with rheumatoid arthritis following multiple doses of the racemate. br. j. clin. pharmacol. 36 (1993) 78–81. doi: https://doi.org/10.1111/j.1365-2125.1993.tb05897.x. [22] k.e. ibrahim, a.f. fell. separation of chloroquine enantiomers by high-performance liquid chromatography. j. pharm. biomed. anal. 8 (1990) 449–452. doi: https://doi.org/10.1016/07317085(90)80074-y. [23] a. aubry, f. gimenez, r. farinotti, i.w. wainer. enantioselective chromatography of the antimalarial agents chloroquine, mefloquine, and enpiroline on a α1-acid glycoprotein chiral stationary phase. chirality 35 (1992) 3–8. doi: https://doi.org/10.1002/chir.530040108. [24] g. li, j. sun, y.y. huang, y. li, y. shi, z. li, x. li, f.h. yang, j. zhao, h.b. luo, t. zhang, x. zhang. enantiomers of chloroquine and hydroxychloroquine exhibit different activities against sars-cov-2 in vitro , evidencing s -hydroxychloroquine as a potentially superior drug for covid-19. (2020). doi: https://doi.org/10.1101/2020.05.26.114033. [25] daicel chiral application, n.d. https://search.daicelchiral.com. [26] k.l. valko, s.p. teague, c. pidgeon. in vitro membrane binding and protein binding (iam mb/pb technology) to estimate in vivo distribution: applications in early drug discovery. admet dmpk 5 (2017) 14. doi: https://doi.org/10.5599/admet.5.1.373. [27] k.l. valkó, s.b. nunhuck, a.p. hill. estimating unbound volume of distribution and tissue binding by in vitro hplc-based human serum albumin and immobilized artificial membrane-binding measurements. j. pharm. sci. 100 (2011) 849–862. doi: https://doi.org/10.1002/jps.22323. [28] k. valko. lipophilicity and biomimetic properties measured by hplc to support drug discovery. j. pharm. biomed. anal. 130 (2016) 35–54. doi: https://doi.org/10.1016/j.jpba.2016.04.009. https://doi.org/10.5599/admet.929 https://doi.org/10.1097/00007691-198612000-00014 https://doi.org/10.1111/j.1365-2125.1995.tb05768.x https://doi.org/10.1016/j.drudis.2020.04.021 https://doi.org/10.1016/j.drudis.2020.04.021 https://doi.org/10.1016/0378-4347(92)80126-b https://doi.org/10.1111/j.1365-2125.1983.tb01513.x https://doi.org/10.1111/j.1365-2125.1983.tb01513.x https://doi.org/10.1111/j.1365-2125.1986.tb02900.x https://doi.org/10.1111/j.1365-2125.1986.tb02900.x https://doi.org/10.1111/j.1365-2125.1993.tb05897.x https://doi.org/10.1016/0731-7085(90)80074-y https://doi.org/10.1016/0731-7085(90)80074-y https://doi.org/10.1002/chir.530040108 https://doi.org/10.1101/2020.05.26.114033 https://search.daicelchiral.com/ https://doi.org/10.5599/admet.5.1.373 https://doi.org/10.1002/jps.22323 https://doi.org/10.1016/j.jpba.2016.04.009 klara valko and tong zhang admet & dmpk 9(2) (2021) 151-165 164 [29] k. valko, c. du, c. bevan, d. reynolds, m. abraham. rapid method for the estimation of octanol / water partition coefficient (log poct) from gradient rp-hplc retention and a hydrogen bond acidity term (sigma alpha2h). curr. med. chem. 8 (2001) 1137–1146. doi: https://doi.org/10.2174/0929867013372643. [30] k. valko, s. nunhuck, c. bevan, m. abraham, d. reynolds. fast gradient hplc method to determine compounds binding to human serum albumin. relationship wth octanol water and immobilized artificial membrane lipophilicity. j. pharm. sci. 92 (2003) 2236–2248. doi: https://doi.org/10.1002/jps.10494. [31] k. valko. physicochemical and biomimetic properties in drug discovery chromatographic techniques for lead optimization. wiley, hoboken, nj, usa, 2014. [32] c. hansch, a. leo. exploring qsar: hydrophobic, electronic, and steric constants. american chemical society, washington, dc., 1995. [33] d.c. warhurst, j.c.p. steele, i.s. adagu, j.c. craig, c. cullander. hydroxychloroquine is much less active than chloroquine against chloroquine-resistant plasmodium falciparum, in agreement with its physicochemical properties. j. antimicrob. chemother. 52 (2003) 188–193. doi: https://doi.org/10.1093/jac/dkg319. [34] z. jiang, j. reilly. chromatography approaches for early screening of the phospholipidosis-inducing potential of pharmaceuticals. j. pharm. biomed. anal. 61 (2012) 184–190. doi: https://doi.org/10.1016/j.jpba.2011.11.033. [35] z. yu, a.p. ijzerman, l.h. heitman. kv11.1 (herg)-induced cardiotoxicity: a molecular insight from a binding kinetics study of prototypical kv11.1 (herg) inhibitors. br. j. pharmacol. 172 (2015) 940–955. doi: https://doi.org/10.1111/bph.12967. [36] f. barbato, v. cirocco, l. grumetto, m. immacolata la rotonda. comparison between immobilized artificial membrane (iam) hplc data and lipophilicity in n-octanol for quinolone antibacterial agents. eur. j. pharm. sci. 31 (2007) 288–297. doi: https://doi.org/10.1016/j.ejps.2007.04.003. [37] r. kaliszan, a. nasal, m. turowski. binding-site for basic drugs on alpha(1)-acid glycoprotein as revealed by chemometric analysis of biochromatographic data. biomed. chromatogr. 9 (1995) 211– 215. doi: https:/doi.org/10.1002/bmc.1130090504. [38] z.h. israili, p.g. dayton. human alpha 1-acid glycoprotein and its interactions with drugs. drug metab. rev. 33 (2001) 161–235. doi: https:/doi.org/10.1081/dmr-100104402. [39] f. hollosy, k. valko, a. hersey, s. nunhuck, g. keri, c. bevan. estimation of volume of distribution in humans from hplc measurements of human serum albumin binding and immobilized artificial membrane partitioning. j. med. chem. 49 (2006) 6958–6971. doi: https://doi.org/10.1021/jm050957i. [40] j.a. shayman, a. abe. drug induced phospholipidosis: an acquired lysosomal storage disorder. biochim. biophys. acta 1831 (2013) 602–611. doi: https://doi.org/10.1016/j.bbalip.2012.08.013. [41] j. müller-höcker, h. schmid, m. weiss, u. dendorfer, g.s. braun. chloroquine-induced phospholipidosis of the kidney mimicking fabry’s disease: case report and review of the literature. hum. pathol. 34 (2003) 285–289. doi: https://doi.org/10.1053/hupa.2003.36. [42] j.-p.h.t.m. ploemen, j. kelder, t. hafmans, h. van de sandt, j.a. van burgsteden, p.j.m. salemink, e. van esch. use of physicochemical calculation of pka and clogp to predict phospholipidosis-inducing potential. exp. toxicol. pathol. 55 (2004) 347–355. doi: https://doi.org/10.1078/0940-2993-00338. [43] f. hollósy, k. valkó, a. hersey, s. nunhuck, g. kéri, c. bevan. estimation of volume of distribution in humans from high throughput hplc-based measurements of human serum albumin binding and immobilized artificial membrane partitioning. j. med. chem. 49 (2006) 6958–71. doi: https://doi.org/10.1021/jm050957i. [44] k. valko, s. rava, s. bunally, s. anderson. revisiting the application of immobilized artificial membrane (iam) chromatography to estimate in vivo distribution properties of drug discovery compounds based on the model of marketed drugs. admet dmpk 8 (2020) 78–97. doi: https://doi.org/10.5599/admet.757. https://doi.org/10.2174/0929867013372643 https://doi.org/10.2174/0929867013372643 https://doi.org/10.1002/jps.10494 https://doi.org/10.1002/jps.10494 https://doi.org/10.1093/jac/dkg319 https://doi.org/10.1093/jac/dkg319 https://doi.org/10.1016/j.jpba.2011.11.033 https://doi.org/10.1016/j.jpba.2011.11.033 https://doi.org/10.1111/bph.12967 https://doi.org/10.1016/j.ejps.2007.04.003 https://doi.org/10.1002/bmc.1130090504 https://doi.org/10.1081/dmr-100104402 https://doi.org/10.1021/jm050957i https://doi.org/10.1021/jm050957i https://doi.org/10.1016/j.bbalip.2012.08.013 https://doi.org/10.1053/hupa.2003.36 https://doi.org/10.1078/0940-2993-00338 https://doi.org/10.1021/jm050957i https://doi.org/10.1021/jm050957i https://doi.org/10.5599/admet.757 admet & dmpk 9(2) (2021) 151-165 biomimetic properties of chlorquine enantiomers doi: https://doi.org/10.5599/admet.929 165 [45] k. valko, e. chiarparin, s. nunhuck, d. montanari. in vitro measurement of drug efficiency index to aid early lead optimization. j. pharm. sci. 101 (2012). doi: https://doi.org/10.1002/jps.23305. [46] l.j. gordon, m. allen, p.e.r. artursson, m.m. hann, b.j. leavens, a. mateus, s.i.m.o.n. readshaw, k. valko, g.j. wayne, a.n.d.y. west. direct measurement of intracellular compound concentration by rapidfire mass spectrometry offers insights into cell permeability. j. biomol. screen. 21 (2016). doi: https://doi.org/10.1177/1087057115604141. [47] s. krishna, n.j. white. pharmacokinetics of quinine, chloroquine and amodiaquine. clinical implications. clin. pharmacokinet. 30 (1996) 263–299. doi: https://doi.org/10.2165/00003088199630040-00002. [48] h.s. lim, j.s. im, j.y. cho, k.s. bae, t.a. klein, j.s. yeom, t.s. kim, j.s. choi, i.j. jang, j.w. park. pharmacokinetics of hydroxychloroquine and its clinical implications in chemoprophylaxis against malaria caused by plasmodium vivax. antimicrob. agents chemother. 53 (2009) 1468–1475. doi: https://doi.org/10.1128/aac.00339-08. [49] l. gustafsson, o. walker, g. alvan, b. beermann, f. estevez, l. gleisner, b. lindstrom, f. sjoqvist. disposition of chloroquine in man after single intravenous and oral doses. br. j. clin. pharmacol. 15 (1983) 471–479. doi: https://doi.org/10.1111/j.1365-2125.1983.tb01532.x. [50] s. tett, d. cutler, r. day, k. brown. a dose‐ranging study of the pharmacokinetics of hydroxychloroquine following intravenous administration to healthy volunteers. br. j. clin. pharmacol. 26 (1988) 303–313. doi: https://doi.org/10.1111/j.1365-2125.1988.tb05281.x. [51] d.e. furst. pharmacokinetics of hydroxychloroquine and chloroquine during treatment of rheumatic diseases. lupus 5 (1996) 11–15. doi: https://doi.org/10.1177/0961203396005001041. [52] a.j. mclachlan, s.e. tett, d.j. cutler, r.o. day. disposition and absorption of hydroxychloroquine enantiomers following a single dose of the racemate. chirality 6 (1994) 360–364. doi: https://doi.org/10.1002/chir.530060421. [53] i.w. wainer, j.c. chen, h. parenteau, s. abdullah, j. ducharme, h. fieger, j. iredale. distribution of the enantiomers of hydroxychloroquine and its metabolites in ocular tissues of the rabbit after oral administration of racemic‐hydroxychloroquine. chirality 6 (1994) 347–354. doi: https://doi.org/10.1002/chir.530060419. [54] p. augustijns, n. verbeke. stereoselective pharmacokinetic properties of chloroquine and de-ethylchloroquine in humans. clin. pharmacokinet. 24 (1993) 259–269. doi: https://doi.org/10.2165/00003088-199324030-00007. ©2021 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/) https://doi.org/10.5599/admet.929 https://doi.org/10.1002/jps.23305 https://doi.org/10.1177/1087057115604141 https://doi.org/10.2165/00003088-199630040-00002 https://doi.org/10.2165/00003088-199630040-00002 https://doi.org/10.1128/aac.00339-08 https://doi.org/10.1111/j.1365-2125.1983.tb01532.x https://doi.org/10.1111/j.1365-2125.1988.tb05281.x https://doi.org/10.1177/0961203396005001041 https://doi.org/10.1002/chir.530060421 https://doi.org/10.1002/chir.530060421 https://doi.org/10.1002/chir.530060419 https://doi.org/10.1002/chir.530060419 https://doi.org/10.2165/00003088-199324030-00007 https://doi.org/10.2165/00003088-199324030-00007 http://creativecommons.org/licenses/by/4.0/ solubility prediction in the bro5 chemical space: where are we right now? doi: https://dx.doi.org/10.5599/admet.834 207 admet & dmpk 8(3) (2020) 207-214; doi: https://doi.org/10.5599/admet.834 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper solubility prediction in the bro5 chemical space: where are we right now? giuseppe ermondi 1 , vasanthanathan poongavanam 2 , maura vallaro 1 , jan kihlberg 2 * and giulia caron 1 * 1 department of molecular biotechnology and health sciences, university of torino, quarello 15, 10135, torino, italy 2 department of chemistry bmc, uppsala university, se-75123, uppsala, sweden *corresponding authors: e-mail: giulia.caron@unito.it; tel.: +39 (0)11 6708337; e-mail: jan.kihlberg@unito.it; tel.: +46 (0)18 4713801 received: april 28, 2020; revised: july 05, 2020; published: july 08, 2020 abstract modelling the solubility of compounds in the “beyond rule of 5” (bro5) chemical space is in its infancy and to date only a few studies have been reported in the literature. based on our own results, and those already published, we conclude that consideration of conformational flexibility and chameleon like behaviour is important, but quantitative models that account for these properties remain to be developed. inclusion of 3d information appears to be somewhat less important than for cell permeability and extremely challenging due to the difficulties of accurate conformational sampling in the bro5 space. currently, methods for modelling of solubility will have to be tailored to the set of investigated compounds. ©2020 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords chameleonicity, flexibility, ionization; lipophilicity; polarity. solubility, and solubility in the bro5 chemical space aqueous solubility plays a crucial role in filtering lead compounds and drug candidates in early stages of drug discovery and development [1]. although dependent on the research program, gsk on the basis of the general solubility equation (gse, see below) recently suggested that solubility is satisfactory (high) when >200 μm, while 30–200 μm was considered as intermediate and <30 μm as poor [2]. application of in silico methods is one of the most appealing strategies to overcome solubility issues in drug discovery projects [3]. however, predicting solubility is not an easy task mainly because of the high uncertainty affecting the experimental data [4], with typical interlaboratory measurement reproducibilities of 0.6 log s units (with s in mol/l) [5]. the main approaches for prediction of solubility have recently been reviewed by bergström and larsson [6] and abramov et al. [7]. in summary, solubility can be predicted using either of two methods: quantitative structure property relationships (qspr), which includes the gse, and physics-based methods based on modelling of the thermodynamic cycle. the gse, physics-based https://dx.doi.org/10.5599/admet.834 https://doi.org/10.5599/admet.834 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:giulia.caron@unito.itn mailto:jan.kihlberg@unito.it http://creativecommons.org/licenses/by/4.0/ giulia caron et al. admet & dmpk 8(3) (2020) 207-214 208 methods and a few qsprs build on the three key steps taken when a molecule transitions from the crystalline state into an aqueous solution [6], i.e. a) dissociation from the crystal lattice [main descriptor: the melting point (mp)], b) preparation of a solvent cavity for incorporation of the molecule [main descriptor: the molecular weight (mw)] and c) insertion of the molecule in the solvent cavity and interactions with water [main descriptor: the logarithm of the partition coefficient between octanol and water (log p)]. in general, higher values for each of these three descriptors result in lower solubilities, but they are not by themselves providing perfect explanations of the individual steps in the solubilization process. for example, cavity formation could be better described by the molecular volume, which in turn is closely correlated to molecular weight. it was recently demonstrated that the solid-state contribution is the limiting factor in the accuracy and predictive power of models of solubility [8]. therefore, if the solubility of a series of compounds is mainly controlled by their crystal packing, it is difficult to obtain an accurate prediction [3]. in such situations the application of a quantum mechanical (qm)-based thermodynamic cycle approach has been suggested [7]. however, if the solubility of a compound is mainly governed by lipophilicity, it is easier to predict its solubility with good accuracy. drug discovery for difficult-to-drug targets often results in ligands that are large, highly lipophilic and semi-flexible compounds, i.e. compounds residing in the “beyond rule of 5” (bro5) chemical space [9]. development of such compounds is associated with high pharmacokinetic risks, low solubility being one of them [10]. in addition to the aforementioned difficulties, additional issues in the prediction of solubility are encountered when dealing with bro5 compounds. first, the low amount of experimental data available in the public domain limits the generation of models. second, many drugs in this space display chameleonlike behaviour (i.e. they adapt their properties to the environment) which originates from their semiflexibility and results in dynamic formation of intra-molecular hydrogen bonds (imhbs) [11] and/or shielding of polar surface area [12,13]. this introduces an additional level of complexity that should be taken into account in any modelling procedure. to provide an update about the current status of solubility modelling in the bro5 chemical space we first review the few bro5 solubility models described in the literature. then we report some computational strategies we applied to model the solubility of a dataset of ten bro5 drugs and drug candidates, and to a second larger dataset of natural product inspired macrocycles. lastly, we have summarized some key findings and attempted to set up preliminary guidelines for how to obtain reliable solubility models for drug discovery in the bro5 chemical space. recent developments we recently investigated the solubility of a structurally diverse set of 11 drugs residing far into the bro5 chemical space [12]. the selected drugs consisted of erythronolides and rifamycin antibacterial agents, as well as hiv-1 and hcv ns3/4a protease inhibitors. as determined by x-ray crystallography each drug populated >2 different conformations (rmsd >1.4 å). due to the difficulties in predicting the relevant conformational space for bro5 drugs [14], these experimentally determined conformations were used to assess the impact of using 3d descriptors when modelling solubility. solubility determined at ph 7.4, where seven of the drugs were ionized, was used in the solubility models, i.e. s (solubility at a given ph where the molecules can be fully or partially ionized) and not s0 (solubility of the neutral species) was used. we found that aqueous solubility was explained to some extent by the 2d descriptor of polarity, i.e. tpsa (r 2 = 0.53), but that the correlation improved substantially when descriptors calculated from the 3d structures of the drugs [15] were used. the best model was based on the conformation of each drug that had the maximum admet & dmpk 8(3) (2020) 207-214 solubility prediction in the bro5 chemical space doi: https://dx.doi.org/10.5599/admet.834 209 molecular 3d psa (max m 3d-psa, r 2 = 0.83), while use of solvent accessible 3d psa provided inferior models. the positive slopes of the correlations support, as expected, that the larger the psa, the more soluble the drug. notably, only a small difference in the quality of the regression model was obtained when the minimum molecular 3d-psa was used instead of the max m 3d-psa. this finding, together with the observation that use of solvent accessible 3d psa provided inferior models, suggests that the overall polarity of the molecule, originating from sampling of multiple aqueous conformations, is the most predictive for solubility. solubility was also very well modelled by experimental lipophilicity (i.e. log d7.4) since the correlation between log s and log d7.4 had r 2 = 0.82. however, the relationship between solubility and calculated lipophilicity was not sought. very recently avdeef and kansy investigated to what extent the solubilities of small, ro5-compliant molecules can be used to predict the intrinsic aqueous solubility of large molecules in the bro5 chemical space [16]. three solubility models published for ro5 compliant molecules [4] were used to predict the solubility of a set of 31 bro5 compounds, for which carefully curated log s0 values have been reported. the gse and the abraham solvation equation failed to predict the solubility of the larger compounds in bro5 space, whereas the random forest regression (rfr) method provided better results. the three methods differ in the applied algorithm, but also in the number of descriptors. three were used in the gse, five or six in the abraham solvation equation and about 200 in rfr. notably, 3d structural information was not used, but the authors suggest that the use of 3d descriptors (e.g. lipophilicity) could significantly improve predictions, since flexibility and conformational preferences can be expected to be more important for big than for small molecules. cyclic peptides are useful model systems for mapping solubility in the bro5 chemical space, and also of major interest as leads on drug discovery projects. qualitative structure-solubility relationships have been reported for cyclic peptides, but to the best of our ability we have not found any specific quantitative models. for instance, lokey and co-workers reported that small variations in the side chains of synthetic analogues of the cyclic peptide natural product sanguinamide a had a large effect on their aqueous solubility [17]. interestingly, in depth studies of three of the cyclic peptides revealed that the one that displayed conformational flexibility had chameleon-like behaviour resulting in high solubility and cell permeability, where two rigid analogues had low solubility but retained the high permeability. another paper from the same group further exemplified the importance of conformational flexibility and chameleon-like behaviour for conveying high solubility and permeability of cyclic peptides from the phepropeptin and epiphepropeptin series [18]. overall, these studies suggest that flexibility and conformational preferences should be included in the prediction of the solubility of cyclic peptides, but a more general approach on how to do this in practice still remains to be described. solubility models for a small set of bro5 drugs we investigated additional methods to model solubility for 10 of the 11 drugs studied earlier by us (rifampicin was excluded because of its zwitterionic nature) [12]. as lipophilicity is one of the three major determinants of solubility according to the gse, and as the size of these 10 drugs does not show a large variation (mw 671-837 da), we focused the modelling on log p and log d calculated by different methods (figure 1, all data are in table s1). mlogp, the 2d lipophilicity descriptor implemented in the lipinski’s ro5, provided a moderate correlation with log s (figure 1a). as expected, significantly better correlations were obtained with log d7.4 calculated by moka (www.moldiscovery.com) and log d7.5 from volsurf+ (vs+, www.moldiscovery.com), highlighting the importance of incorporating the charge of the drug in the https://dx.doi.org/10.5599/admet.834 http://www.moldiscovery.com/ http://www.moldiscovery.com/ giulia caron et al. admet & dmpk 8(3) (2020) 207-214 210 models (figures 1b and 1c). it is worth to note that inclusion of 3d structural information [log d7.5 (vs+)] did not significantly improve the statistical significance of the regression model found with 2d log d [log d7.4 (moka)] for this set of bro5 drugs. figure 1. solubility models for a dataset of 10 drugs in bro5 space. experimentally determined solubility at ph 7.4 (log s) and its correlation to (a) mlogp, implemented in the ro5, (b) log d7.4 calculated in moka, (c) log d7.5 calculated using volsurf+, and (d) log s7.5 calculated using volsurf+. volsurf+ also allows calculation of log s0 and log s at different ph values. we predicted log s7.5 (vs+) based on 3d descriptors derived from an average conformation produced by the software from the smiles code of each of the 10 drugs. however, the correlation between log s7.5 (vs+) and the experimentally determined solubility (figure 1d) was of lower quality than those obtained with log d descriptors (figures 1b and 1c). besides volsurf+ some other in silico tools apply 2d models to the prediction of solubility from smiles codes. most if not all of them have been set-up using datasets of small molecules having solubility values curated at different levels of quality. nevertheless, considering the free availability and user-friendly interfaces we decided to evaluate them in the bro5 chemical space. the smiles codes of the 10 drugs were therefore submitted to admetsar (http://lmmd.ecust.edu.cn/admetsar2/), admetlab (http://admet.scbdd.com/calcpre/index/), pkcsm (http://biosig.unimelb.edu.au/pkcsm) and marvin sketch (https://chemaxon.com/products/marvin). notably, only log s calculated with marvin sketch provided a good linear relationship with the experimental data (r 2 =0.81 with log s, 0.59 with log s0; all the data are in table s2). however, the slope and the intercept were significantly different from 1 and 0 (0.42 and -2.61, respectively) and thus the predicted values are not close to the experimental ones. a solubility classification model for a set of structurally complex macrocycles previously, some of us determined the aqueous solubility, lipophilicity (log d) and permeability across caco-2 cell monolayers for a set of 200 non-peptidic, de novo–designed macrocycles, the structures of which were inspired by natural products [19]. in-depth analysis of this dataset revealed that stereoand http://lmmd.ecust.edu.cn/admetsar2/ http://admet.scbdd.com/calcpre/index/ http://biosig.unimelb.edu.au/pkcsm https://chemaxon.com/products/marvin admet & dmpk 8(3) (2020) 207-214 solubility prediction in the bro5 chemical space doi: https://dx.doi.org/10.5599/admet.834 211 regiochemistry can have a large influence on passive permeability and cellular efflux, whereas their impact on solubility appeared to be lower. moreover, an appropriate conformational flexibility was concluded to be a highly desirable property that may provide compounds in bro5 space with chameleonic properties that allow them to display both high aqueous solubility and high cell permeability. however, as structuresolubility relationships were not investigated for this set of macrocycles we now selected a subset consisting of 65 of the macrocycles for more detailed studies. the macrocycles in this set were selected by having no or very low efflux across caco-2 cells (er <2), i.e. by possessing one important property favourable for development of drugs in bro5 space. the solubility distribution of the subset suggested that a classification rather than a regression strategy should be applied (figure 2), and the threshold to distinguish soluble (47) from poorly soluble (18) compounds was fixed at 50 m. this is a slightly more liberal cut-off than that proposed by gsk (30 m) to distinguish compounds having poor from those having intermediate solubility [2]. figure 2. distribution of the aqueous solubility for the de-novo designed macrocycles. in contrast to the 10 bro5 drugs discussed above, descriptors for lipophilicity (log p or log d) failed to provide models for solubility for this set of de novo-designed macrocycles (data not shown) and therefore more complex methods were investigated. the corina (https://www.mn-am.com/online_demos/corina_demo) conformer of the charged and neutral forms of the selected macrocycles was obtained and conformational sampling was performed for both forms using omega (https://www.eyesopen.com/omega). then descriptors were calculated for (1) the 2d structure (which provided the 2d dataset), (2) the corina conformer (named 3d) and (3) the minimum energy conformer from omega (named mec). a pool of 2d descriptors were calculated for the first (2d) dataset, while both 2d and 3d descriptors were calculated for the 3d and mec datasets. random forest (rf) classification models were built for data matrixes using weka v3.8 (https://www.cs.waikato.ac.nz/ml/weka/) and their quality was evaluated using the matthews correlation coefficient (mcc), which takes into account true positives and negatives and returns a value between -1 and +1. a perfect prediction is characterized by a coefficient of +1, a random prediction by 0, while a completely incorrect prediction has an mcc of -1. in general, models having mcc values greater than 0.4 are considered to be predictive. all models for this set of macrocycles were of good or high quality (table 1, leave-5-out crossvalidation was used), but those obtained for the charged species, (mcc: 0.84 – 1.00) were superior to those for the neutral species (mcc = 0.43 – 0.92). this finding agrees well with the fact that most macrocycles in the dataset are predicted to be positively charged at ph 7.4. notably, the best model (mcc = 1) was obtained for the charged forms using only 2d descriptors, while slightly inferior models were obtained when 3d descriptors were incorporated. the classification models were further assessed using an internal test set obtained by splitting the dataset into a training (50 macrocycles) and a test set (15 macrocycles). again, the https://dx.doi.org/10.5599/admet.834 https://www.mn-am.com/online_demos/%1fcorina_demo https://www.mn-am.com/online_demos/%1fcorina_demo https://www.eyesopen.com/%1fomega https://www.eyesopen.com/%1fomega https://www.cs.waikato.ac.nz/ml/weka/ giulia caron et al. admet & dmpk 8(3) (2020) 207-214 212 2d-based rf model performed better than the models that included 3d information from conformational sampling (mcc = 0.71, table s3). eleven descriptors were found to have a high impact on the rf classification models, among which those describing lipophilicity, charge and surface area descriptors are the most important (table s4). table 1. summary of solubility classification models for the dos macrocyclic dataset (five-fold cross validation, #descriptors = number of descriptors of the model; tp = true positives, fn = false negatives; tn = true negatives; fp = false positives, mcc = matthews correlation coefficient) dataset #descriptors tp fn tn fp mcc n e u tr a l 2d 3 39 8 11 7 0.43 3d 12 46 1 17 1 0.92 mec 3 46 1 16 2 0.88 c h a rg e d 2d 11 47 0 18 0 1.00 3d 5 46 0 18 1 0.96 mec 10 46 1 15 3 0.84 general considerations on solubility in the bro5 chemical space the previous sections clearly support that different bro5 datasets can require different strategies for modelling of their solubility. for instance, the solubility of the 10 drugs in bro5 space showed an excellent correlation with calculated log d only, whereas the solubility of the de-novo designed set of macrocycles required development of an advanced rf model. overall, these observations highlight that models developed for specific, small datasets usually cannot be transferred to other datasets. another key observation is that the impact of ionization on solubility cannot be neglected when dealing with bro5 compounds, just as for ro5 compliant small molecules. therefore, the pka of the investigated compounds should be accurately predicted before modelling solubility, which is a far from an easy and trivial task. moreover, lipophilicity and polarity descriptors are needed to model solubility, but they should be specifically designed and validated for large and flexible compounds. in principle, a 3d description of compounds in bro5 space should be important for modelling solubility since conformational changes that expose surfaces with different properties could be required when a molecule dissociates from the crystal lattice and moves into solution. however, the examples discussed herein seem to suggest that the impact of the 3rd dimension on solubility is less important than for cell permeability. in fact, inclusion of 3d descriptors failed to improve the solubility models both for the 10 bro5 drugs [12] and for the de-novo designed macrocycles [19]. in contrast, avdeef and kansy suggested that inclusion of 3d information might be important [16], and this is also observed for the cyclic peptides studied by lokey and co-workers [17,18]. thus, it remains to be established if modelling of aqueous solubility is facilitated by methods for prediction of the conformations adopted in aqueous solution. however, we recently showed that reproducing experimental conformations by tools designed for conformational sampling of large and macrocyclic compounds is far from being an easy task [14]. overall, more experimental data is needed to draw general conclusions about what the best approaches are for modelling the solubility of large and flexible compounds. this data is likely to be available in the pharmaceutical industry and partnering with academic researchers could be the preferred way to further analysis. we hope that this weakness can be overcome so that more reliable methods for prediction of the solubility of compounds lying in the bro5 chemical space can be developed. admet & dmpk 8(3) (2020) 207-214 solubility prediction in the bro5 chemical space doi: https://dx.doi.org/10.5599/admet.834 213 acknowledgements: yuriy abramov and gilles goetz are kindly acknowledged for sharing their expertise in the field. we thank openeye scientific software and chemaxon for providing free academic licenses. conflict of interest: none. references [1] e.h. kerns, l. di. pharmaceutical profiling in drug discovery. drug discovery today 8(7) (2003) 316– 23. https://doi.org/10.1016/s1359-6446(03)02649-7. [2] s.b. bunally, c.n. luscombe, r.j. young. using physicochemical measurements to influence better compound design. slas discovery 24(8) (2019) 791–801. https://doi.org/10.1177/2472555219859845. [3] l. di, p. v. fish, t. mano. bridging solubility between drug discovery and development. drug discovery today 17(9-10) (2012) 486–495. https://doi.org/10.1016/j.drudis.2011.11.007. [4] a. avdeef. prediction of aqueous intrinsic solubility of druglike molecules using random forest regression trained with wiki-ps 0 database. admet & dmpk 8(1) (2020) 29-77. http://dx.doi.org/10.5599/admet.766 [5] c.a.s. bergström, r. holm, s.a. jørgensen, s.b.e. andersson, p. artursson, s. beato, a. borde, k. box, m. brewster, j. dressman, k.i. feng, g. halbert, e. kostewicz, m. mcallister, u. muenster, j. thinnes, r. taylor, a. mullertz. early pharmaceutical profiling to predict oral drug absorption: current status and unmet needs. european journal of pharmaceutical sciences 57 (2014) 173–199. https://doi.org/10.1016/j.ejps.2013.10.015. [6] c.a.s. bergström, p. larsson. computational prediction of drug solubility in water-based systems: qualitative and quantitative approaches used in the current drug discovery and development setting. international journal of pharmaceutics 540(1-2) (2018) 185–193. https://doi.org/10.1016/j.ijpharm.2018.01.044. [7] y.a. abramov, g. sun, q. zeng, q. zeng, m. yang. guiding lead optimization for solubility improvement with physics-based modeling. molecular pharmaceutics 17(2) (2020) 666–673. https://doi.org/10.1021/acs.molpharmaceut.9b01138. [8] y.a. abramov. major source of error in qspr prediction of intrinsic thermodynamic solubility of drugs: solid vs nonsolid state contributions? molecular pharmaceutics 12(6) (2015) 2126–2141. https://doi.org/10.1021/acs.molpharmaceut.5b00119. [9] b.c. doak, j. zheng, d. dobritzsch, j. kihlberg. how beyond rule of 5 drugs and clinical candidates bind to their targets. journal of medicinal chemistry 59(6) (2016) 2312–2327. https://doi.org/10.1021/acs.jmedchem.5b01286. [10] p. matsson, b.c. doak, b. over, j. kihlberg. cell permeability beyond the rule of 5. advanced drug delivery reviews 101 (2016) 42-61. https://doi.org/10.1016/j.addr.2016.03.013 [11] g. caron, j. kihlberg, g. ermondi. intramolecular hydrogen bonding: an opportunity for improved design in medicinal chemistry. medicinal research reviews 39(5) (2020) 1707-1729. https://doi.org/10.1002/med.21562. [12] m. rossi sebastiano, b.c doak, m. backlund, v. poongavanam, b. over, g. ermondi, g. caron, p. matsson, j. kihlberg. impact of dynamically exposed polarity on permeability and solubility of chameleonic drugs beyond the rule of 5. journal of medicinal chemistry 61(9) (2018) 4189–4202. https://doi.org/10.1021/acs.jmedchem.8b00347. [13] e. danelius, v. poongavanam, s. peintner, l.h.e. wieske, m. erdelyi, j. kihlberg. solution conformations explain the chameleonic behavior of macrocyclic drugs. chemistry 26(23) (2020) 5231-5244. https://doi.org/10.1002/chem.201905599. [14] v. poongavanam, e. danelius, s. peintner, l. alcaraz, g. caron, m.d. cummings, s. wlodek, m. erdelyi, p.c.d. hawkins, g. ermondi, j. kihlberg. conformational sampling of macrocyclic drugs in different environments – can we find the relevant conformations? acs omega 3(9) (2018) 1174211757. https://doi.org/10.1021/acsomega.8b01379 https://dx.doi.org/10.5599/admet.834 https://doi.org/10.1016/s1359-6446(03)02649-7 https://doi.org/10.1177/%1f2472555219859845 https://doi.org/10.1177/%1f2472555219859845 https://doi.org/10.1016/j.drudis.2011.11.007 http://dx.doi.org/10.5599/admet.766 http://dx.doi.org/10.5599/admet.766 https://doi.org/%1f10.1016/j.ejps.2013.10.015 https://doi.org/%1f10.1016/j.ejps.2013.10.015 https://doi.org/10.1016/j.ijpharm.2018.01.044 https://doi.org/10.1016/j.ijpharm.2018.01.044 https://doi.org/10.1021/acs.molpharmaceut.5b00119 https://doi.org/10.1021/acs.jmedchem.5b01286 https://doi.org/10.1021/acs.jmedchem.5b01286 https://doi.org/10.1016/j.addr.2016.03.013 https://doi.org/10.1021/acs.jmedchem.8b00347 https://doi.org/10.1002/chem.201905599 https://doi.org/10.1021/acsomega.8b01379 giulia caron et al. admet & dmpk 8(3) (2020) 207-214 214 [15] g. caron, g. ermondi. molecular descriptors for polarity: the need of going beyond psa. future med.chem. 8(17) (2016) 2013-2016. https://doi.org/10.4155/fmc-2016-0165. [16] a. avdeef, m. kansy. can small drugs predict the intrinsic aqueous solubility of ‘beyond rule of 5’ big drugs. admet & dmpk (2020). doi: https://dx.doi.org/10.5599/admet.794. [17] a.t. bockus, j.a. schwochert, c.r. pye, c.e. townsend, v. sok, m.a. bednarek, r.s. lokey. going out on a limb: delineating the effects of -branching, n-methylation, and side chain size on the passive permeability, solubility, and flexibility of sanguinamide a analogues. journal of medicinal chemistry 58(18) (2015) 7409–7418. https://doi.org/10.1021/acs.jmedchem.5b00919. [18] j. schwochert, y. lao, c.r. pye, m.r. naylor, p. v. desai, i.c. gonzalez valcarcel, j.a. barrett, g. sawada, m.j. blanco, r.s. lokey. stereochemistry balances cell permeability and solubility in the naturally derived phepropeptin cyclic peptides. acs medicinal chemistry letters 7(8) (2016) 757– 761. https://doi.org/10.1021/acsmedchemlett.6b00100. [19] b. over, p. matsson, c. tyrchan, p. artursson, b.c. doak, m.a. foley, c. hilgendorf, s.e. johnston, m.d. lee, r.j. lewis, p. mccarren, g. muncipinto, u. norinder, m.w.d. perry, j.r. duvall, j. kihlberg. structural and conformational determinants of macrocycle cell permeability. nature chemical biology 12 (2016) 1065–1074. https://doi.org/10.1038/nchembio.2203. ©2020 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) https://doi.org/10.4155/fmc-2016-0165 https://dx.doi.org/10.5599/admet.794 https://doi.org/10.1021/acs.jmedchem.5b00919 https://doi.org/10.1021/acsmedchemlett.6b00100 https://doi.org/10.1038/nchembio.2203 http://creativecommons.org/licenses/by/3.0/ nmr spectroscopy in drug discovery and development: evaluation of physico-chemical properties doi: 10.5599/admet.737 242 admet & dmpk 7(4) (2019) 242-251; doi: http://dx.doi.org/10.5599/admet.737 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index review nmr spectroscopy in drug discovery and development: evaluation of physico-chemical properties mire zloh ucl school of pharmacy, university college london, 29/39 brunswick square, london, wc1n 1ax, uk faculty of pharmacy, university business academy, trg mladenaca 5, 21000, novi sad, serbia *corresponding author: e-mail: zloh@live.co.uk; tel.: +44-777-00-23411 received: october 18, 2019; revised: november 19, 2019; published: december 11, 2019 abstract determination of physico-chemical properties of compounds is one of the corner-stones in selection of hit molecules for further progression into lead development in the modern drug discovery process. the speed of traditional analyses and limited quantities of hit molecules are obstacles for efficient acquisition of experimental data. herein, the range of applications of quantitative nuclear magnetic resonance (nmr) spectroscopy as a fast technique used to evaluate solubility, log p and pka are discussed. keywords logp; qnmr; drug solubility; dosy; pharmaceutical formulations. introduction the process of modern drug discovery and development is highly complex process that aims to create new medicines that meet stringent requirements set by governmental bodies, such as food and drug administration (fda) in the usa, european medicines agency (ema) and others around the world [1]. efforts of the industry have already heavily utilized most of the easiest disease targets, which reduces the chances of serendipity in discovery of new therapeutic agents. this requires systematic approaches in identifying new targets or paradigm shift in finding novel lead molecules, resulting in the need for iterative, tedious and quite often expensive multidisciplinary processes. although the record breaking number of 59 drugs, including 17 biologics, were approved by fda in 2018 [2], the high failure rates of clinical candidates frustrates pharmaceutical industry and academic research groups. it is shown that safety issues are linked to off-target interactions of these compounds [3] or their physico-chemical properties [4]. therefore, pharmaceutical industry was focused on the drug likeness properties of drug candidates in recent years. absorption, distribution, metabolism, excretion, and toxicity (admet) properties are generally included as filters during process of identification of hit molecules or a guide in the design or modifications of lead molecules. a significant progress was achieved in prediction of admet properties using in silico methods and these approaches are generally cheaper and greener when compared to their experimental determination [5]. although these methods can be used in virtual high throughput screening as filtering steps, they are often limited by their applicability domains, as prediction of admet molecules with novel http://dx.doi.org/10.5599/admet.737 http://www.pub.iapchem.org/ojs/index.php/admet/index admet & dmpk 7(4) (2019) 242-251 nmr spectroscopy in drug discovery and development doi: 10.5599/admet.737 243 scaffolds outside of chemical space of training sets may not be reliable. most relevant assessments of compounds properties can be achieved by employing in vivo assays, often preceded by in vitro experiments as indicators of their adme fate [6]. such two-tier approaches cannot always be employed for large number of compounds or in a high throughput manner, which demands utilization of biophysical techniques and development of novel assays to evaluate physico-chemical characteristics in relation to admet properties. significant progress has been achieved in utilization of chromatography to measure lipophilicity of compounds and their biomimetic properties [7,8], as well as to evaluate their polarity, conformational states and intramolecular hydrogen bonds [9]. the development of great variety of chromatographic methods is the result of possibility to utilize a range of stationary phases and to vary mobile phases to represent different parts of biological systems [8], while the automation increases the throughput of analysis. on the contrary, nuclear magnetic spectroscopy (nmr) spectroscopy is not readily associated with measuring admet properties. up to date, the scopus search using keyword “nmr” in conjunction with “adme” results in 125 records, which is significantly lower when compared to 456 when the “chromatography” keyword is used. this may be a consequence that nmr requires expensive instrumentation and samples are generally dissolved in deuterated solvents at concentrations that may be considerably higher than those observed in physiological conditions. therefore, this spectroscopic technique is mainly utilized to acquire qualitative information on molecular structure of organic compounds as well as biopolymers, or for the assessment of the sample purity. however, nmr can also provide information on three dimensional (3d) structures at atomic level resolution of pure samples and mixtures, with a possibility to detect and even quantify intermolecular interactions as dynamic processes, making this techniques versatile and applicable to different stages in the search for new therapeutic agents [10]. therefore, despite its disadvantages, nmr spectroscopy plays a significant role in the drug discovery and development processes. some of these applications include: protein structure determination in solution [11] or target structure validation in presence of membranes [12], enabling structure-based drug discovery in solution [13,14] or in-cell [15], identification of hit compounds [16] and determination of their conformations [17], development of biopharmaceutical formulations [18], and evaluation of drug metabolism and toxicity in patients [19,20] that can even play part in post-market surveillance [21,22]. it appears that the above applications may not be generally implemented as an integral part of research programs in pharmaceutical industry [10]. therefore, the information that can be readily obtained by nmr should be further integrated in the drug discovery process and wider utilized for the evaluation of the physico-chemical properties. herein, simpler applications of hydrogen nuclear magnetic spectroscopy ( 1 h nmr) spectroscopy, not often reported as a tool for evaluating some admet properties, are explored. nmr quantitative nmr (qnmr) most of the adme properties rely on concentration measurements, where quantitative nmr (qnmr) can play a significant role in the pharmaceutical industry. one of the major advantages of the nmr spectroscopy is its inherent quantitative nature of the signal, as the measured signals can be directly proportional to the number of investigated nuclei in the same environment. these signals can be obtained for nuclei most relevant to pharmaceutical industry, such as 1 h, 13 c, 19 f and 31 p. although the acquisition of 19 f and 31 p nmr signals results in high specificity analysis, the 1 h signal is most commonly used in qnmr mire zloh admet & dmpk 7(4) (2019) 242-251 244 spectroscopy due to its presence in most drugs, short relaxation times and higher sensitivity when compared to other nuclei. the adequate setup of the experiment (pulse sequence that allows nuclei to fully relax and to minimize nuclear overhauser effect) will result in an one dimensional (1d) nmr spectrum with areas under peaks proportional to concentrations of sample components. this phenomenon is observed for all molecules that contain investigated nuclei, as the nmr does not depend on compound-specific properties, such as presence of a chromophore, chemical structure or volatility. furthermore, if a sample is fully dissolved, concentrations of all compounds with nmr active nuclei can be determined using a single standard, without reliance on specific reference standards for each compound [23]. the qnmr can be used as a relative quantification method by evaluating molar ratio between two compounds (analyte and referent compound) in a solution (eq. 1), or it can be used to determine absolute content of the analyte of a mixture if the concentration of the referent compound is known (eq. 2): an refan ref ref an i nn n i n  (1) an ref an an ref ref an ref i n m m m i n m  (2) where n is the number of mols, i is the area under the peak that correspond to a signal, n number of nuclei contributing to the signal, m is molar mass and m is the gravimetric mass. the above approach does not require calibration curves or isolation of compounds. multicomponent mixtures can be subjected to direct and simultaneous determination of concentrations of several molecules as long as it is possible to find a characteristic signal for each analyte. for example, in a case of quantification of caffeine, formic acid, trigonelline and 5-hmf in soluble coffees, singlet peaks for molecule were identified that were not overlapping with the signals from the other components present in the sample. that allowed quick and reliable measurement of their concentrations with the limits of the detection for each component that were between 0.3 mg to 1.3 mg per gram of solid product [24]. in pharmaceutical industry setting, impurities (including toxic compounds) [25] or components of formulations can be quantified at the same time with measurement of the concentration of the active pharmaceutical ingredient. the quantification of memantine hydrochloride from tablets dissolved in d2o shown a good separation between a selected signal from signal (three hydrogens giving a singlet at 0.75 ppm) and caffeine signal used as a referent standard (three hydrogens giving a singlet at 3.13 ppm) (figure 1). the method was validated according to ich guideline q2 (r1) with demonstrated recovery of 99.26 % and rsd of 0.38 % [25]. the advantages resulted in the of use of qnmr not only in analyses of samples of simple mixtures of small organic molecules, but also in the evaluation of complex mixtures found in the field of metabolomics and studies of physiological pathways. presence of many compounds in such mixtures results in a difficulty to identify an isolated signal of the referent compound to ensure reliable quantification. to facilitate the optimal choice of referent compounds for a range of samples, a set of 15 molecules were tested for their properties and suitability for the use in qnmr [26]. in cases where the addition of an additional components can affect the sample, the referent compound can be positioned externally or an artificial referent signal can be generated by an nmr instrument such as eretic (electronic referencing to access in vivo concentrations), pig (pulse into gradient), artsi (amplitude‐corrected referencing through signal injection), pulcon (pulse length‐based concentration measurements) and quantas (quantification by artificial signal) [27]. admet & dmpk 7(4) (2019) 242-251 nmr spectroscopy in drug discovery and development doi: 10.5599/admet.737 245 the reliability and accuracy of concentration determination using qnmr depend on the experiment parameters setup and consideration of sample conditions, which is extensively covered elsewhere [28,29]. in reality, the experimental setup can be optimized to reduce the quantitative inaccuracy of qnmr below 2.0 %, an acceptable limit for precise and accurate quantification[30]. furthermore, the nmr concentration determination can be achieved from low micromolar concentrations 4 μm even if the solvent water peak is used as the concentration reference [31]. this approach expands the applicability domain of the qnmr as the effects of deuterated water on pka, log p and other physico-chemical parameters are removed. figure 1. overlay of 1 h nmr spectra of memantine hydrochloride from tablets in solution (a), drug standard (b), internal standard (c), and blank (d). all spectra were acquired with same parameters and processed similarly. taken with permission from [25]. copyright (2016) with permission from wiley. drug solubility drug solubility is one of the key properties that influences the permeability and concentration of the active pharmaceutical ingredient in the physiological conditions after administration. while the hplc measurements are generally used to determine solubility of drugs, as one of the most reliable and sensitive methods, it may be relatively slow process as often requires method development and creation of a calibration curve for quantitation. the need for faster measurements of large number of compounds, often with limited amounts, qnmr can be used to develop faster methods for solubility measurement, even in complex matrices that may affect drug solubility. as the nmr is fast method with relatively simple sample preparation procedure, it can be used to determine drug solubility in simple formulations, by calculation the integration ratio of the drug signal to the internal water soluble standard 3-(trimethylsilyl)propionic2,2,3,3-d4 acid, sodium salt signal [32]. the further advantage is that the concentrations of all hydrogen containing components of the formulation could be determined simultaneously, along with detecting possible impurities. the new peaks that appear in samples, possibly during the degradation studies, can be quantified by nmr and further characterized by a combination of nmr spectroscopy and mass spectrometry (ms) protocol (figure 2) [33]. additionally, qnmr also allows determination of drug concentrations in presence of excipients without separation and at different temperatures [34]. mire zloh admet & dmpk 7(4) (2019) 242-251 246 figure 2. workflow of the mass spectrometric-nuclear magnetic resonance (ms-nmr) approach. taken from [33]. copyright (2019) from mdpi under creative commons attribution (cc by) license. often, quantification of the concentration of an active pharmaceutical ingredient in complex mixtures is impeded in one-dimensional hydrogen (1d 1 h) nmr spectra due to the signal overlap. the addition in signal separation along the second dimension in two-dimensional (2d) nmr spectra generally leads to reduction of signal overlap. the 2d correlation spectroscopy (cosy) nmr spectra can be used for quantification purposes by integrating the volumes of crosspeaks of metabolites, that are proportional to their concentrations [35]. there is a plethora of different homoand heteronuclear experiments that can be used to obtain quantitative information about sample components. although this relationship between the peak volume and the concentration is not as simple as in 1d nmr spectroscopy, the volume of the signal depends on various peak relevant parameters, including relaxation times, coupling between different nuclei and pulse sequence parameters, there is a range of strategies that can be employed to achieve absolute quantification [36,37]. however, the need for integration of the peak volumes and careful validation of 2d methods employed[38], are limiting factors that may affect applicability of 2d qnmr spectroscopy as a fast and simple method. alternatively, diffusion-ordered nmr spectroscopy (dosy) can be used as 2d method that detects not only association between different components that may include small compounds and polymers [39], but it can be used for quantification too. diffusion coefficients, determined by dosy, can be used to evaluate solubilisation capacity of micelles of sodium dodecyl sulphate (sds) for poorly soluble molecules such as artemisinin and curcumin. the solubility of these molecules can be evaluated, individually and in combination of different components by plotting selected chemical shifts against the diffusion coefficients determined by 2d dosy spectra [40]. furthermore, beside the information on solubility of investigated compounds in fasted state simulated intestinal fluid, the dosy spectra provide information on diffusion properties of micelles formed by bile acids and poorly soluble drugs [41]. admet & dmpk 7(4) (2019) 242-251 nmr spectroscopy in drug discovery and development doi: 10.5599/admet.737 247 log p determination the partition coefficient (p) reflects the different solubility of a molecule in two immiscible phases. the logarithm of p (log p) determined for water and octanol is often correlated with lipophilicity and of the molecules and indirectly with its permeability through biological barriers. the log p determination requires measuring equilibration concentrations of the molecule in water and octanol. the above-mentioned advantages of nmr, direct detection of analyte without reliance on presence of a chromophore, simple data interpretation, and minimal sample preparation, can be used in determination of log p. if a drug, or a drug candidate, has the aqueous solubility higher than the limit of quantification by qnmr, its concentration in water and octanol can be determined directly after separation and without the need for calibration curve [42]. this approach can be applied after compound partitioning between water and 1‐octanol using standard methods by taking aliquots from two layers. the signal amplitudes extracted the a compound from each can be analyzed using the craft software, that directly calculates both log p and log d7.4 values [43]. the similar experiment can be conducted directly in the nmr tube by mini shake-flask approach, where the initial determination of equilibrium concentrations of an analyte in water is followed by addition of the octanol layer and equilibration of the sample (figure 3). the measurement the concentrations of the analyte a in the water phase before and after equilibration. the concentration of the analyte can be calculated as the difference between the amount of analyte in water before and after equilibration step [44]. this approach could easily be implemented as a follow up step to solubility determination of drug candidates and be possibly incorporated in automation assays that use nmr spectroscopy. measurement of log p for fluorinated compounds can employ 19 f nmr spectroscopy in addition to 1 h nmr spectroscopy. using similar procedures to previously described procedures, there is no need for calibration curves providing a straightforward method for the measurement of lipophilicity of fluorinated compounds [45,46]. figure 3. schematic diagram of the analyte partition between the water and 1-octanol layers upon equilibration. taken with permission from [44] (https://doi.org/10.1021/acsomega.7b01102) and further requests for permission should be directed to acs. . pka determination the acid-base dissociation constant (pka) of a drug is also a physicochemical parameter that affects many of its biophysical characteristics., including solubility, lipophilicity, binding to proteins and crossing biological barriers. the pka determination involves monitoring the ionization state of a molecule with changing ph. traditional methods, such as potentiometric titrations or following the change of uv absorption as a function of ph, involve complex experimental setup and analyses of results. nmr spectroscopy is also based on analysis of protonation degree of a molecule with changing ph. the protonation degree is monitored either by the chemical shift changes close to the acidic or basic regions of molecule [47] or via changes of the integration values of peaks that arise from ionizable groups (figure 4) [48]. https://doi.org/10.1021/acsomega.7b01102 mire zloh admet & dmpk 7(4) (2019) 242-251 248 figure 4. different ph value 1 h nmr spectrums obtained by the 10 mm tetrabutylammonium hydroxide solution (tbah) titration of 1‐[2‐(4‐methoxyphenyl)ethyl]‐3‐methylbenzimidazolium iodide (2) compound in 10 mm solution and the sigmoidal curve that was acquired with the passing of integral values obtained from these spectrums to the against ph values. taken with permission from [48]. copyright (2019) with permission from wiley. conclusions the nmr spectroscopy is extensively used in the various drug discovery and development process, however further applications of qnmr can have a significant impact on the early stages of selection of lead compounds with adequate adme properties. furthermore, the determination of physico-chemical properties is not affected by the presence of other components, moreover it may provide significant information on behaviour of molecules in environments that are more similar to conditions in in vitro and possibly in vivo experiments. despite first nmr signals discovered over 70 years ago, the development of new approaches in solving of real-life problems is still undergoing and it has a potential to play even greater role in the future drug discovery programmes. conflict of interest: the author declares no conflict of interest. references [1] j.l. koh, d. gottfried, d.r. lee, s. navarrete, understanding and addressing regulatory concerns in research, basic methods handbook for clinical orthopaedic research, springer2019, pp. 515-531. [2] b. g de la torre, f. albericio. the pharmaceutical industry in 2018. an analysis of fda drug approvals from the perspective of molecules. molecules 24 (2019) e809. [3] s. bendels, c. bissantz, b. fasching, g. gerebtzoff, w. guba, m. kansy, j. migeon, s. mohr, j.-u. peters, f. tillier, r. wyler, c. lerner, c. kramer, h. richter, s. roberts. safety screening in early drug discovery: an optimized assay panel. journal of pharmacological and toxicological methods 99 (2019) 106609. admet & dmpk 7(4) (2019) 242-251 nmr spectroscopy in drug discovery and development doi: 10.5599/admet.737 249 [4] m.j. waring, j. arrowsmith, a.r. leach, p.d. leeson, s. mandrell, r.m. owen, g. pairaudeau, w.d. pennie, s.d. pickett, j. wang, o. wallace, a. weir. an analysis of the attrition of drug candidates from four major pharmaceutical companies. nat rev drug discov 14 (2015) 475-486. [5] l.l. ferreira, a.d. andricopulo. admet modeling approaches in drug discovery. drug discovery today 24 (2019) 1157-1165. [6] t.d. chung, d.b. terry, l.h. smith, in vitro and in vivo assessment of adme and pk properties during lead selection and lead optimization–guidelines, benchmarks and rules of thumb, assay guidance manual [internet], eli lilly & company and the national center for advancing translational sciences2015. [7] s. bunally, r.j. young. the role and impact of high throughput biomimetic measurements in drug discovery. admet and dmpk 6 (2018) 74-84. [8] k.l. valkó. lipophilicity and biomimetic properties measured by hplc to support drug discovery. journal of pharmaceutical and biomedical analysis 130 (2016) 35-54. [9] g.h. goetz, m. shalaeva. leveraging chromatography based physicochemical properties for efficient drug design. admet and dmpk 6 (2018) 85-104. [10] w. jahnke. perspectives of biomolecular nmr in drug discovery: the blessing and curse of versatility. journal of biomolecular nmr 39 (2007) 87-90. [11] d. wishart. nmr spectroscopy and protein structure determination: applications to drug discovery and development. current pharmaceutical biotechnology 6 (2005) 105-120. [12] y. miao, h. qin, r. fu, m. sharma, t.v. can, i. hung, s. luca, p.l. gor'kov, w.w. brey, t.a. cross. m2 proton channel structural validation from full-length protein samples in synthetic bilayers and e. coli membranes. angewandte chemie international edition 51 (2012) 8383-8386. [13] t. sugiki, k. furuita, t. fujiwara, c. kojima. current nmr techniques for structure-based drug discovery. molecules 23 (2018) e148. [14] y. li, c. kang. solution nmr spectroscopy in target-based drug discovery. molecules 22 (2017) e1399. [15] c. kang. applications of in-cell nmr in structural biology and drug discovery. international journal of molecular sciences 20 (2019) e139. [16] a.d. gossert, w. jahnke. nmr in drug discovery: a practical guide to identification and validation of ligands interacting with biological macromolecules. progress in nuclear magnetic resonance spectroscopy 97 (2016) 82-125. [17] e. chiarparin, m.j. packer, d.m. wilson. experimental free ligand conformations: a missing link in structure-based drug discovery. future science, 2019, pp. 79-82. [18] a. pandya, m.j. howard, m. zloh, p.a. dalby. an evaluation of the potential of nmr spectroscopy and computational modelling methods to inform biopharmaceutical formulations. pharmaceutics 10 (2018) e165. [19] a. vignoli, v. ghini, g. meoni, c. licari, p.g. takis, l. tenori, p. turano, c. luchinat. high‐throughput metabolomics by 1d nmr. angewandte chemie international edition 58 (2019) 968-994. [20] a. guleria, a. kumar, u. kumar, r. raj, d. kumar. nmr based metabolomics: an exquisite and facile method for evaluating therapeutic efficacy and screening drug toxicity. current topics in medicinal chemistry 18 (2018) 1827-1849. [21] p. keizers, a. van drongelen, w. de jong, c. van oostrom, b. roszek, b. venhuis, c. de vries, r. geertsma, r. janssen. silicone breast implants in the netherlands: a market surveillance study. (2016). [22] p.-c. wu, y.-c. yang, d.-y. wang, h.-f. cheng. scientific and regulatory perspective on monoclonal antibody biosimilars. biopharmaceuticals (2018) 111. [23] s.k. bharti, r. roy. quantitative 1 h nmr spectroscopy. trac trends in analytical chemistry 35 (2012) 5-26. mire zloh admet & dmpk 7(4) (2019) 242-251 250 [24] g. del campo, i. berregi, r. caracena, j. zuriarrain. quantitative determination of caffeine, formic acid, trigonelline and 5-(hydroxymethyl)furfural in soluble coffees by 1 h nmr spectrometry. talanta 81 (2010) 367-371. [25] a. sahu, m. narayanam, m. kurmi, m.k. ladumor, s. singh. quantitation of memantine hydrochloride bulk drug and its tablet formulation using proton nuclear magnetic resonance spectrometry. magnetic resonance in chemistry 54 (2016) 632-636. [26] t. rundlöf, m. mathiasson, s. bekiroglu, b. hakkarainen, t. bowden, t. arvidsson. survey and qualification of internal standards for quantification by 1 h nmr spectroscopy. journal of pharmaceutical and biomedical analysis 52 (2010) 645-651. [27] c. sterling, r. crouch, d.j. russell, a.i. calderón. 1 h‐nmr quantification of major saccharides in açaí raw materials: a comparison of the internal standard methodology with the absolute intensity qnmr method. phytochemical analysis 24 (2013) 631-637. [28] y.b. monakhova, b.w. diehl. practical guide for selection of 1 h qnmr acquisition and processing parameters confirmed by automated spectra evaluation. magnetic resonance in chemistry 55 (2017) 996-1005. [29] s. singh, r. roy. the application of absolute quantitative 1 h nmr spectroscopy in drug discovery and development. expert opinion on drug discovery 11 (2016) 695-706. [30] f. malz, h. jancke. validation of quantitative nmr. journal of pharmaceutical and biomedical analysis 38 (2005) 813-823. [31] h. mo, d. raftery. solvent signal as an nmr concentration reference. analytical chemistry 80 (2008) 9835-9839. [32] m. lin, m. tesconi, m. tischler. use of 1 h nmr to facilitate solubility measurement for drug discovery compounds. international journal of pharmaceutics 369 (2009) 47-52. [33] c. palaric, r. molinié, d. cailleu, j.-x. fontaine, d. mathiron, f. mesnard, y. gut, t. renaud, a. petit, s. pilard. a deeper investigation of drug degradation mixtures using a combination of ms and nmr data: application to indapamide. molecules 24 (2019) e1764. [34] p. saveyn, e. cocquyt, w. zhu, d. sinnaeve, k. haustraete, j.c. martins, p. van der meeren. solubilization of flurbiprofen within non-ionic tween 20 surfactant micelles: a 19 f and 1 h nmr study. physical chemistry chemical physics 11 (2009) 5462-5468. [35] j. alonso, c. arús, w.m. westler, j.l. markley. two-dimensional correlated spectroscopy (cosy) of intact frog muscle: spectral pattern characterization and lactate quantitation. magnetic resonance in medicine 11 (1989) 316-330. [36] p. giraudeau. quantitative 2d liquid-state nmr. magnetic resonance in chemistry 52 (2014) 259272. [37] x. li, k. hu, quantitative nmr studies of multiple compound mixtures, annual reports on nmr spectroscopy, elsevier2017, pp. 85-143. [38] g. dufour, b. evrard, p. de tullio. 2d-cosy nmr spectroscopy as a quantitative tool in biological matrix: application to cyclodextrins. the aaps journal 17 (2015) 1501-1510. [39] p. hodge, p. monvisade, g.a. morris, i. preece. a novel nmr method for screening soluble compound libraries. chemical communications (2001) 239-240. [40] s. lapenna, a.r. bilia, g.a. morris, m. nilsson. novel artemisinin and curcumin micellar formulations: drug solubility studies by nmr spectroscopy. journal of pharmaceutical sciences 98 (2009) 3666-3675. [41] m. vogtherr, a. marx, a.-c. mieden, c. saal. investigation of solubilising effects of bile salts on an active pharmaceutical ingredient with unusual ph dependent solubility by nmr spectroscopy. european journal of pharmaceutics and biopharmaceutics 92 (2015) 32-41. [42] h. mo, k.m. balko, d.a. colby. a practical deuterium-free nmr method for the rapid determination of 1-octanol/water partition coefficients of pharmaceutical agents. bioorg med chem lett 20 (2010) 6712-6715. admet & dmpk 7(4) (2019) 242-251 nmr spectroscopy in drug discovery and development doi: 10.5599/admet.737 251 [43] d. soulsby, j.a.m. chica. determination of partition coefficients using 1 h nmr spectroscopy and time domain complete reduction to amplitude-frequency table (craft) analysis. magnetic resonance in chemistry 55 (2017) 724-729. [44] h. cumming, c. rücker. octanol–water partition coefficient measurement by a simple 1 h nmr method. acs omega 2 (2017) 6244-6249. [45] b. linclau, z. wang, g. compain, v. paumelle, c.q. fontenelle, n. wells, a. weymouth-wilson. investigating the influence of (deoxy)fluorination on the lipophilicity of non-uv-active fluorinated alkanols and carbohydrates by a new log p determination method. angewandte chemie international edition 55 (2016) 674-678. [46] z. wang, b.f. jeffries, h.r. felstead, n.j. wells, e. chiarparin, b. linclau. a new straightforward method for lipophilicity (logp) measurement using 19f nmr spectroscopy. j vis exp (2019). [47] j. bezençon, m.b. wittwer, b. cutting, m. smieško, b. wagner, m. kansy, b. ernst. pka determination by 1 h nmr spectroscopy–an old methodology revisited. journal of pharmaceutical and biomedical analysis 93 (2014) 147-155. [48] a. mumcu, h. küçükbay. determination of pka values of some novel benzimidazole salts by using a new approach with 1 h nmr spectroscopy. magnetic resonance in chemistry 53 (2015) 1024-1030. ©2019 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ estimation of the octanol-water distribution coefficient of basic compounds by a cationic microemulsion electrokinetic chromatography system doi: http://dx.doi.org/10.5599/admet.760 98 admet & dmpk 8(1) (2020) 98-112; doi: http://dx.doi.org/10.5599/admet.760 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper estimation of the octanol-water distribution coefficient of basic compounds by a cationic microemulsion electrokinetic chromatography system alejandro fernández-pumarega 1 , belén martín-sanz 1 , susana amézqueta 1, *, elisabet fuguet 1,2 , martí rosés 1,§ 1 departament d’enginyeria química i química analítica and institut de biomedicina (ibub), facultat de química, universitat de barcelona, martí i franquès 1-11, 08028, barcelona, spain. 2 serra húnter programme. generalitat de catalunya. spain § editorial board member *corresponding author: e-mail: samezqueta@ub.edu; tel.: (+34) 934021277; fax: (+34) 934021233 received: november 29, 2019; revised: february 14, 2020; available online: march 04, 2020 abstract the octanol-water partition coefficient (po/w), or the octanol-water distribution coefficient (do/w) for ionized compounds, is a key parameter in the drug development process. in a previous work, this parameter was estimated through the retention factor measurements in a sodium dodecyl sulfate (sds) microemulsion electrokinetic chromatography (meekc) system for acidic compounds. nonetheless, when ionized basic compounds were analyzed, undesirable ion pairs were formed with the anionic surfactant and avoided a good estimation of log do/w. for this reason, an alternative meekc system based on a cationic surfactant has been evaluated to estimate po/w or do/w of neutral compounds and ionized bases. to this end, it has been characterized through the solvation parameter model (spm) and compared to the octanol -water partition system. results pointed out that both systems show a similar partition behavior. hence, the log po/w of a set of neutral compounds has been successfully correlated against the logarithm of the retention factor (log k) determined in this meekc system. then, the log do/w of 6 model bases have been estimated at different ph values and they have been compared to data from the literature, determined by the reference shake-flask and potentiometric methods. good agreement has been observed between the literature and the estimated values when the base is neutral or partially ionized (up to 99 % of ionization). keywords lipophilicity; log po/w; log do/w; meekc; ionized bases; solvation parameter model; microemulsion; system surrogation. introduction chromatographic systems based on different techniques such as high performance liquid chromatography (hplc) and electrokinetic chromatography (ekc) have been widely used to determine biopartitioning properties [1–6]. in these systems, the compounds experience a partition between an aqueous phase and a stationary phase, for hplc systems, or a pseudoestationary phase, for ekc systems, similar to the partition that compounds experience in a biological system. therefore, if the partition in the chromatographic system and in the biological one is similar enough, it is possible to estimate the biological property through a correlation like eq. 1: http://dx.doi.org/10.5599/admet.760 http://dx.doi.org/10.5599/admet.760 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:samezqueta@ub.edu admet & dmpk 8(1) (2020) 98-112 estimation of do/w by meekc doi: http://dx.doi.org/10.5599/admet.760 99 log 𝑆𝑃bio = q + p log 𝑆𝑃chrom (1) where spbio is the solute biological property, spchrom the solute physicochemical property (generally the retention factor, k), and “q” and “p” are the intercept and the slope of the resulting correlation between both parameters, respectively. the lipophilicity of compounds is a property of high interest in biological processes due to its direct relationship with membrane permeability. lipophilicity is usually evaluated through the octanol-water partition coefficient (po/w). the direct method to determine this parameter is the shake-flask method. however, it is a tedious procedure and it is not fully automated. for this reason, the estimation of po/w through high-throughput systems such as chromatographic ones is of great interest. in the previous works [3–6], eq. 1 was applied to the determination of the octanol-water partition coefficient of neutral solutes and the distribution coefficient of partially ionized acidic compounds (do/w) from microemulsion electrokinetic chromatography (meekc) measurements. a meekc system based on a negatively charged microemulsion (me) constituted of heptane, 1-butanol, and sodium dodecyl sulfate (sds) was used. the aim of the present study is to evaluate the applicability of meekc measurements to estimate the log po/w and log do/w of ionized basic compounds, which are positively charged in their ionized form. however, the sds-meekc system cannot be used for basic compounds since the formation of ion pairs between the surfactant, negatively charged, and the protonated base, positively charged has been observed. thus, the retention factor would be altered leading to wrong log do/w estimated values. in this work, the replacement of the anionic surfactant (sds) by a cationic one (tetradecyltrimethylammonium bromide, ttab) is tested to solve this problem. so, as the only role of the surfactant is to stabilize the oil droplets, we think that the ttab-meekc system should be able to estimate the po/w or do/w of basic solutes as well as the sds-meekc does for acidic solutes. actually, ishihama et al. [7] evaluated 3 different surfactants and showed that neither the ionic groups nor the hydrocarbon chain lengths affected the selectivity of the me for neutral compounds. thus, the aim of the present work is to test this potential ability of the ttab-meekc system. theory the retention factor of an ionized base can be calculated through eq. 2 [8]: 𝑘 = 𝜇−( 𝜇 𝜇0 ) ephedrine cation ·𝜇0 𝜇me−𝜇 (2) where, μ and μ0 are the electrophoretic mobility of the compound in meekc and in plain buffer (capillary zone electrophoresis, cze), respectively. μme is the electrophoretic mobility of the me marker in the meekc analysis, and (µ/µ0)ephedrine cation is the viscosity correction factor (whose value is 0.84). this correction needs to be introduced since two solutions with quite different viscosities (ƞ) are employed in k calculation, and µ is inversely related to ƞ [9]. therefore, the mobility in cze is corrected to make it equivalent to the one measured in meekc. ephedrine has been selected because it is a small and polar compound that is not retained in the chromatographic system when it is fully ionized. so, (µ/µ0)ephedrine cation results from the ratio of mobilities of the ephedrine cation measured in meekc and cze modes. μ values can be obtained from: 𝜇 = [ 1 𝑡r − 1 𝑡0 ] · [ 𝐿t𝐿d 𝑉 ] (3) http://dx.doi.org/10.5599/admet.760 susana amézqueta et al. admet & dmpk 8(1) (2020) 98-112 100 where, tr and t0 are the elution times of the compound and the electroosmotic flow marker, respectively. lt and ld are the total and the effective length of the capillary, respectively, and v is the voltage applied during the separation. similarly to acidic solutes [10], the retention factor of a monoprotic basic compound in a meekc system varies with the ph of the media through: 𝑘 = 𝑘 (bh+) +𝑘(b)·10 ph−p𝐾a ′ 1+10ph−p𝐾a ′ (4) where k(b) and k(bh+) are the retention factors of the neutral and the fully ionized base, respectively, and the pka’ is the apparent acidity constant of the base. experimental equipment all the analysis were performed with a capillary electrophoresis (ce) 7100 equipped with a uv diode array detector from agilent technologies (santa clara, ca, usa). the fused-silica capillaries employed were from polymicro technologies (phoenix, az, usa). the capillaries used had a 50 μm internal diameter. water was purified by a milli-q plus system from millipore (burlington, ma, usa) with a resistivity of 18.2 mω·cm. to determine the ph of the solutions a ph-meter glp 22 from crison (barcelona, spain) was used. reagents hydrochloric acid (1n tritisol tm ), and sodium hydroxide (0.5n tritisol tm ) were acquired from merck (darmstadt, germany). methanol (hplc-grade) was obtained from thermo fisher scientific (waltham, ma, usa). heptane (99 %), dodecanophenone (98 %), ttab (>99 %), 1-butanol (≥99.7 %), 2-[bis(2-hydroxyethyl)amino]-2-(hydroxymethyl)propane-1,3diol (bis-tris) (>99%), and sodium phosphate dodecahydrate (>98 %) were purchased from sigma-aldrich (st. louis, mo, usa). disodium hydrogen phosphate (99.5%) was from baker (phillipsburg, nj, usa). test compounds with high purities were supplied from different manufacturers: sigma-aldrich, merck, baker, carlo erba (milan, italy), fluka (st. louis, mo, usa), riedel-de häen (seelze, germany), and scharlab (barcelona, spain). analysis conditions buffer preparation to prepare the buffers at ph 5.4 and ph 7.0, the ph of an aliquot of a 0.2 m bis-tris solution previously protonated with hcl was adjusted with naoh, while to prepare the buffer at ph 11.4 a mixture of a 0.2 m disodium hydrogen phosphate solution and a 0.2 m sodium phosphate dodecahydrate solution was used. all the buffers were prepared maintaining the ionic strength at 0.05 m. me preparation to prepare the me 1.70, g of ttab were dissolved in around 70 ml of the corresponding buffer solution. then, 8.15 ml of 1-butanol were added, finishing by the addition of 1.15 ml of heptane. the co-surfactant and the oil were added under continuous magnetic stirring, and if the solution remained turbid, it was sonicated until clarification [4]. finally, buffer was added up to a total volume of 100 ml. the concentrations of each component with respect the total volume of the me were: 1.70 % (w/v) ttab, admet & dmpk 8(1) (2020) 98-112 estimation of do/w by meekc doi: http://dx.doi.org/10.5599/admet.760 101 8.15 % (v/v) 1-butanol, and 1.15 % (v/v) heptane. instrumental parameters all the measurements were performed keeping the temperature at 25 °c and using a fused-silica capillary with an effective length of 30 cm and a total length of 38.5 cm. the voltage applied in meekc measurements was negative and between -11.5 and -14 kv. regarding the internal pressure applied, it ranged between 0 and 25 mbar. the analysis conditions were different, depending on the ph, to obtain the best electrophoretic window possible. in the case of cze measurements, the analyses were performed employing positive voltages. most compounds were dissolved at 200 mg·l -1 in a 9:1 me:methanol mixture in the meekc analysis and in a 9:1 buffer:methanol mixture in the cze analysis. solutes with low uv absorbance were analyzed at a concentration of 10 % (v/v). the injection of the compounds was performed applying during 5 s an internal pressure of 50 mbar. the detection of the solutes was performed at λ=200, 214 or 254 nm (depending on the absorbance profile of the compounds). the me marker used was dodecanophenone (200 mg·l -1 , detected at λ = 254 nm), while the electroosmotic flow marker was methanol (10 % v/v, detected at λ = 200 nm) [11]. comparison of systems characterized by the solvation parameter model the ttab-meekc system has been characterized through the solvation parameter model (spm) to check its suitability to surrogate the octanol-water partition system [12]. spm relates a free energy solvation property (in this work the log po/w or the log k) to five different solute descriptors. the equation resulting from the characterization of a system through the spm model is shown below: log 𝑆𝑃 = c + 𝑒𝐸 + 𝑠𝑆 + 𝑎𝐴 + 𝑏𝐵 + 𝑣𝑉 (5) where sp is the solute property in a given partitioning system; e, s, a, b and v are the abraham solute descriptors (which are specific for each compound), and the coefficients e, s, a, b and v provide the characteristics of the partition system. e is the excess molar refraction, s the solute dipolarity/polarizability, a and b are the solute hydrogen-bond acidity and basicity descriptors, respectively, and v the mcgowan’s volume of the molecule. each of the system coefficients are complementary to their respective solute descriptor. the system coefficients can be calculated by a multiple linear regression between the depending solute property of a set of representative neutral compounds and their solute descriptors [13]. the similarity between different systems can be determined by comparing the coefficients resulting from the characterization of the systems by the spm. lazaro, et al. [14] proposed the distance parameter d to compare two different systems (i and j) through their normalized spm coefficients (eqs. 6-10). 𝑒u = 𝑒 𝑙 (6) 𝑠u = 𝑠 𝑙 (7) 𝑎u = 𝑎 𝑙 (8) 𝑏u = 𝑏 𝑙 (9) 𝑣u = 𝑣 𝑙 (10) where eu, su, au, bu, vu are the normalized coefficients (or coefficients of the unitary vector) and l is the length of the coefficients vector; which is calculated as follows: http://dx.doi.org/10.5599/admet.760 susana amézqueta et al. admet & dmpk 8(1) (2020) 98-112 102 𝑙 = √𝑒 2 + 𝑠2 + 𝑎2 + 𝑏2 + 𝑣2 (11) then, the d parameter is calculated by eq. 12. 𝑑 = √(𝑒ui − 𝑒uj) 2 + (𝑠ui − 𝑠uj) 2 + (𝑎ui − 𝑎uj ) 2 + (𝑏ui − 𝑏uj) 2 + (𝑣ui − 𝑣uj ) 2 (12) the smaller the d parameter, the more similar will be the systems compared. as a practical convention, if the distance between the systems compared is below 0.25, it can be concluded that these systems are analogous [14]. moreover, the precision of the correlation (eq. 13) between log po/w and log k in the ttab-meekc system has been estimated following the approach published elsewhere [15]. the overall precision of the correlation (sdcorr 2 ) can be estimated by the sum of three different contributions: 𝑆𝐷corr 2 = 𝑆𝐷𝑃o/w 2 + (p · 𝑆𝐷meekc ) 2 + 𝑆𝐷d 2 (13) sdpo/w and sdmeekc are the standard deviations (sd) obtained from the characterization through the spm of the octanol-water partition system and the ttab-meekc system, respectively. sdd and “p” are the sd and the slope obtained from the correlation of the log po/w and log k values of a set of representative solutes. the values of the parameters were calculated using the equations resulting from the spm characterization and the descriptors of the solutes. by using these equations, sdpo/w and sdmeekc are equal to 0. so, the sd of this correlation can be attributed uniquely to the dissimilarity between the systems. data analysis table curve 2d from systat software inc. (san jose, ca, usa) was employed to obtain the log do/w-ph profiles of the basic compounds. data calculations were performed using excel 2010 from microsoft (redmond, wa, usa). results and discussion suitability of the ttab-meekc system to estimate log po/w values characterization of the meekc system by the spm the ttab-meekc system has been characterized through the spm in order to later evaluate its similarity with the octanol-water partition system. to obtain accurate solvation coefficients of the model, representative analytes have been selected to characterize the system. in 2001, fuguet et al. [13] proposed 71 compounds from a 2975 solute data base, selected according to their solute-solvent interactions, as a good set to characterize chromatographic systems by the spm. the retention factors of the proposed set of compounds in the ttab-meekc system have been measured at ph 7.0, where all compounds are neutral. during the characterization, some of the compounds have been discarded since they are not lipophilic enough and elute with the eof marker (propan-1,3-diol, butan-1,4-diol, and pentan-1,5-diol), they are too lipophilic and they co-elute with the me marker (butylbenzene, and α-pinene), or they present other experimental troubles in the ttab-meekc system (myrcene, pentan-1-ol and pentan-3-ol). the retention factor of neutral compounds can be calculated using eq. 2, being µ0 equal to 0 as they are uncharged. table 1 lists the compounds used in the characterization together with their solute descriptors, and the log k in the ttab-meekc system. moreover, the log po/w values of each compound reported in the literature are also listed [16–30]. admet & dmpk 8(1) (2020) 98-112 estimation of do/w by meekc doi: http://dx.doi.org/10.5599/admet.760 103 table 1. abraham solute descriptors, log k and log po/w of the chosen neutral analytes compound e s a b v log k log po/w a) propan-1-ol 0.236 0.42 0.37 0.48 0.5900 -0.71 0.30 propan-2-ol 0.212 0.36 0.33 0.56 0.5900 -0.79 0.05 butan-1-ol 0.224 0.42 0.37 0.48 0.7309 0.08 0.88 pentan-1-ol 0.219 0.42 0.37 0.48 0.8718 1.56 pentan-3-ol 0.218 0.36 0.33 0.56 0.8718 1.21 propan-1,3-diol 0.397 0.91 0.77 0.85 0.6487 -1.04 butan-1,4-diol 0.395 0.93 0.72 0.90 0.7860 -0.83 pentan-1,5-diol 0.388 0.95 0.72 0.91 0.9305 -0.43 thiourea 0.840 0.82 0.77 0.87 0.5696 -1.06 -1.02 benzene 0.610 0.52 0.00 0.14 0.7164 0.36 2.13 toluene 0.601 0.52 0.00 0.14 0.8573 0.77 2.69 ethylbenzene 0.613 0.51 0.00 0.15 0.9982 1.10 3.15 propylbenzene 0.604 0.50 0.00 0.15 1.1391 1.53 3.68 butylbenzene 0.600 0.51 0.00 0.15 1.2800 4.38 p-xylene 0.613 0.52 0.00 0.16 0.9982 1.12 3.15 naphthalene 1.340 0.92 0.00 0.20 1.0854 1.21 3.37 chlorobenzene 0.718 0.65 0.00 0.07 0.8388 0.85 2.90 bromobenzene 0.882 0.73 0.00 0.09 0.8914 0.99 2.99 anisole 0.708 0.75 0.00 0.29 0.9160 0.28 2.11 benzaldehyde 0.820 1.00 0.00 0.39 0.8730 -0.10 1.47 acetophenone 0.818 1.01 0.00 0.48 1.0139 -0.03 1.58 propiophenone 0.804 0.95 0.00 0.51 1.1548 0.35 2.24 butyrophenone 0.797 0.95 0.00 0.51 1.2957 0.68 2.65 valerophenone 0.795 0.95 0.00 0.50 1.4366 1.05 3.40 heptanophenone 0.720 0.95 0.00 0.50 1.7184 1.88 4.41 benzophenone 1.447 1.50 0.00 0.50 1.4808 1.02 3.18 methyl benzoate 0.733 0.85 0.00 0.46 1.0726 0.35 2.12 benzyl benzoate 1.264 1.42 0.00 0.51 1.6804 1.49 3.97 benzonitrile 0.742 1.11 0.00 0.33 0.8711 -0.06 1.56 aniline 0.955 0.96 0.26 0.50 0.8162 -0.28 0.90 o-toluidine 0.970 0.90 0.23 0.59 0.9751 -0.04 1.32 3-chloroaniline 1.050 1.10 0.30 0.36 0.9390 0.47 1.88 4-chloroaniline 1.060 1.10 0.30 0.35 0.9390 0.41 1.84 2-nitroaniline 1.180 1.37 0.30 0.36 0.9904 0.25 1.79 3-nitroaniline 1.200 1.71 0.40 0.35 0.9904 0.07 1.32 4-nitroaniline 1.220 1.91 0.42 0.38 0.9904 0.11 1.39 nitrobenzene 0.871 1.11 0.00 0.28 0.8906 0.16 1.85 2-nitroanisole 0.965 1.34 0.00 0.38 1.0902 0.17 1.73 benzamide 0.990 1.50 0.49 0.67 0.9728 -0.39 0.64 4-aminobenzamide 1.340 1.94 0.80 0.94 1.0726 -0.86 -0.41 acetanilide 0.870 1.36 0.46 0.69 1.1137 -0.17 1.19 4-chloroacetanilide 0.980 1.50 0.64 0.51 1.2357 0.51 2.12 phenol 0.805 0.89 0.60 0.30 0.7751 0.05 1.48 3-methylphenol 0.822 0.88 0.57 0.34 0.9160 0.38 2.02 http://dx.doi.org/10.5599/admet.760 susana amézqueta et al. admet & dmpk 8(1) (2020) 98-112 104 table 1. continued… compound e s a b v log k log po/w a) 2,3-dimethylphenol 0.850 0.90 0.52 0.36 1.0569 0.64 2,4-dimethylphenol 0.840 0.80 0.53 0.39 1.0569 0.71 2.42 thymol 0.822 0.79 0.52 0.44 1.3387 1.20 3.34 4-chlorophenol 0.915 1.08 0.67 0.20 0.8975 0.79 2.39 catechol 0.970 1.10 0.88 0.47 0.8338 -0.09 0.88 resorcinol 0.980 1.00 1.10 0.58 0.8338 -0.18 0.80 hydroquinone 1.000 1.00 1.16 0.60 0.8338 -0.43 0.58 2-naphthol 1.520 1.08 0.61 0.40 1.1441 1.04 2.84 1,2,3-trihydroxybenzene 1.165 1.35 1.35 0.62 0.8925 -0.29 furan 0.369 0.53 0.00 0.13 0.5363 -0.31 1.34 2,3-benzofuran 0.888 0.83 0.00 0.15 0.9053 0.71 2.67 quinoline 1.268 0.97 0.00 0.51 1.0443 0.20 2.15 pyrrole 0.613 0.73 0.41 0.29 0.5774 -0.44 0.75 pyrimidine 0.606 1.00 0.00 0.65 0.6342 -1.03 -0.40 antipyrine 1.320 1.50 0.00 1.48 1.5502 -0.67 0.56 caffeine 1.500 1.60 0.00 1.33 1.3632 -0.77 -0.01 corticosterone 1.860 3.43 0.40 1.63 2.7389 0.65 1.90 cortisone 1.960 3.50 0.36 1.87 2.7546 0.23 1.47 hydrocortisone 2.030 3.49 0.71 1.90 2.7975 0.39 1.53 estradiol 1.800 3.30 0.88 0.95 2.1988 1.35 4.01 estratriol 2.000 3.36 1.40 1.22 2.2575 0.67 2.45 monuron 1.140 1.50 0.47 0.78 1.4768 0.32 1.94 myrcene 0.483 0.29 0.00 0.21 1.3886 4.17 α-pinene 0.446 0.14 0.00 0.12 1.2574 4.83 geraniol 0.513 0.63 0.39 0.66 1.4903 1.07 a) from references [16-30]. to obtain the spm coefficients of the ttab-meekc system, butan-1-ol, and thiourea were considered as outliers and discarded, as they presented standard residues higher than 2.5. estradiol had a standard residue slightly greater than 2.5. however, it was not considered as outlier because discarding it had no big influence on the result. the parameters and statistics (f, fisher’s f parameter; sd, standard deviation; r 2 , determination coefficient; n, number of compounds) resulting from the system characterization are: log 𝑘 = − 0.96(±0.05) + 0.44(±0.08)𝐸 − 0.63(±0.06)𝑆 + 0.20(±0.05)𝐴 − 2.13(±0.07)𝐵 + 2.36(±0.06)𝑉 (14) r 2 = 0.972, sd = 0.12, f = 368, n = 59. the large values of the coefficients b and v show that the hydrogen bond basicity and the volume of the solute are the more important parameters. the v coefficient is positive, meaning that it is easier to form a cavity to place the solute in the me phase than in the aqueous one. a negative b coefficient is obtained indicating a higher hydrogen bonding acidity of the aqueous phase with respect to the me phase. the e coefficient is positive, so the me is more polarizable than the aqueous phase. the negative value of the s coefficient indicates that the me system is less dipolar than the aqueous phase. the coefficient a is close to zero, showing that hydrogen bonding basicities of me and aqueous phases are similar, and the variation of solute hydrogen bond acidity has a small effect on the system. admet & dmpk 8(1) (2020) 98-112 estimation of do/w by meekc doi: http://dx.doi.org/10.5599/admet.760 105 model comparison the ttab-meekc system has been compared to the octanol-water partition system and the sds-meekc system. the last system was used in other research studies to estimate the log po/w of ionizable acids [5,6]. the normalized coefficients obtained from the characterization of the ttab-meekc, the octanol-water partition [31], and the sds-meekc [3] (calculated using eqs. 6-11), and the d distance parameter (eq. 12) are summarized in table 2. table 2. normalized coefficients and d distances of the compared systems. system eu su au bu vu n sd r 2 d ref. ttab meekc 0.13 -0.19 0.06 -0.65 0.72 59 0.12 0.972 this work octanol-water 0.11 -0.20 0.01 -0.65 0.72 613 0.12 0.994 0.06 [31] sds meekc 0.07 -0.16 -0.01 -0.67 0.72 53 0.09 0.988 0.11 [3] the normalized coefficients of the three systems are very similar, being for all of them the b and v coefficients the most important ones. furthermore, the d parameter between ttab-meekc and octanolwater partition is very small (0.06), meaning that the ttab-meekc system is a good approximation to surrogate the octanol-water partition one, and consequently, lipophilicity. the d parameter obtained between ttab-meekc and sds-meekc is also small (0.11), therefore ttab is a good substitute of sds for log po/w determination. the chromatographic precision of the correlation between log k in the ttab-me and the octanol-water partition has also been calculated. as explained before and elsewhere in the literature [15], the overall precision results can be estimated from a sum of three different contributions (sdpo/w 2 , (p·sdmeekc) 2 , sdd 2 ). here, the largest contribution to the overall precision comes from the chromatographic data ((p·sdmeekc) 2 = 0.034) and the dissimilarity between the two systems (sdd 2 = 0.013) is similar than the variance of the biopartitioning data (sdpo/w 2 = 0.013). however, the calculated overall precision (sdcorr 2 = 0.061) is low. therefore, the correlation obtained between the ttab-meekc and octanol-water systems should be good, meaning that, the ttab-meekc system is a good candidate to estimate satisfactorily the octanol/water lipophilicity of neutral compounds. correlation between log po/w and log k a linear correlation between the log po/w and the log k of 58 solutes presented in table 1 has been established. the correlation between log po/w and log k is plotted in figure 1. the equation and statistics obtained are as follows: log 𝑃o/w = 1.68(±0.05) · log 𝑘 + 1.35(±0.04) (15) r 2 =0.954; sd=0.25; f=1174; n=58; which is not significantly different from the log po/w – log k correlation already reported for the sds-meekc system employing a set of neutral compounds (eq. 16) [6]. log 𝑃o/w = 1.60(±0.11) · log 𝑘 + 1.51(±0.08) (16) r 2 = 0.916; sd = 0.33; f = 196; n = 20 good statistics have been obtained for the correlation: r 2 value close to 1, and a small sd value of the regression, whereby the ttab-meekc system has been proved to provide a good estimation of the log po/w of neutral compounds, as already pointed out through the model comparison tools. therefore, eq. 15 is going to be used as calibration curve for the estimation of do/w of ionized bases in the following section. http://dx.doi.org/10.5599/admet.760 susana amézqueta et al. admet & dmpk 8(1) (2020) 98-112 106 figure 1. log po/w vs log k correlation with the compounds from table 1 (eq. 15). log do/w estimation of ionized basic compounds. with the aim to extent the applicability of the method to ionizable bases, the log do/w of six basic compounds (alprenolol, nadolol, oxprenolol, penbutolol, pindolol, and propranolol) have been estimated at different ph values, thus at different ionization degrees and compared to literature experimental values. all these compounds are used as pharmaceutical drugs and present a basic group with a pka inside the electrophoretic ph working range (2-12) [27]. similarly to k (eq. 4), the log do/w of ionized bases depends on the pka’ of the base and the ph of the me through: log 𝐷o/w = log ( 10 log 𝑃 o/w(bh+) +10 log𝑃o/w(b) ·10ph−p𝐾𝑎 ′ 1+10ph−p𝐾𝑎 ′ ) (17) where log po/w(bh+) and log po/w(b) are the logarithms of the octanol–water partition coefficient of the fully ionized and neutral base, respectively. the log do/w – ph profiles for each of the model compounds have been calculated fitting eq. 17 to the log do/w values determined at different values of ph taken from the literature [32–37]. the log do/w values compiled have been determined using the reference shake-flask, and the ph-metric titration methods. the profiles, the parameters and the statistics resulting from these fittings can be found in figure 2 and table 3. table 3. parameters resulting from the fit of eq. 17 to log do/w vs ph data. k(bh+) and k(b) have been experimentally determined in the ttab-meekc system. compound pka’ (sd) log po/w(bh+) (sd) log po/w(b) (sd) r 2 sd f k(bh+) k(b) alprenolol 9.63 (0.28) -0.46 (0.16) 3.20 (0.23) 0.939 0.36 116 0.67 11.09 nadolol 9.31 (0.21) -2.05 (0.11) 0.83 (0.15) 0.975 0.21 156 0.05 0.88 oxprenolol 9.38 (0.16) -1.43 (0.10) 2.14 (0.13) 0.985 0.18 328 0.14 2.82 penbutolol 9.37 (0.31) 1.23 (0.24) 4.06 (0.24) 0.987 0.24 37 3.20 61.64 pindolol 9.32 (0.17) -1.39 (0.10) 1.79 (0.13) 0.970 0.23 243 0.16 2.43 propranolol 9.24 (0.26) -0.17 (0.15) 2.99 (0.21) 0.908 0.37 94 1.19 16.51 then, the k(bh+) and k(b) of the bases have been determined in the ttab-meekc system at ph 5.4 and 11.4, respectively (table 3). the retention factors and the pka’ of each of the bases have been used in eq. 4 to estimate k values at the ph of interest. then, log do/w has been estimated through eq. 15. the log do/w from the literature [32–37] and the estimated values are summarized in table 4. the ionization degree -2.0 -1.0 0.0 1.0 2.0 3.0 4.0 5.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 2.0 2.5 lo g p o /w l og k admet & dmpk 8(1) (2020) 98-112 estimation of do/w by meekc doi: http://dx.doi.org/10.5599/admet.760 107 (α(bh+)) of the bases at each ph value (calculated using eq. 18) are also provided. 𝛼(bh+) = 10p𝐾a ′ −ph 1+10p𝐾a ′ −ph (18) figure 2. log do/w – ph profiles obtained adjusting eq. 17 to the data from the literature (•): a) alprenolol, b) nadolol, c) oxprenolol, d) penbutolol, e) pindolol, f) propranolol. figure 3 compares the estimated and the experimental data from table 4 for bases neutral or fully ionized (figure 3a) or partially ionized, from 1 to 99 % (figure 3b). the same figure also represents a line with a slope of 1 and an intercept of 0, and two extra lines showing the 95 % confidence interval (±2 sd) of the calibration curve for neutral compounds. analyzing the results from figure 3 and table 4, it can be concluded that, generally, accurate estimated log do/w values are obtained especially when the compound is in its neutral or partially ionized forms. the error obtained is similar to the one obtained for neutral compounds, as indicated by the error bars. nonetheless, when the bases are highly or totally ionized (α ≈ 1) an overestimation of the parameter is observed. note that when the compound is completely ionized the results are less comparable due to the high dependence of the log po/w(bh+) value with the solution medium (nature of counter-ions and concentrations as usually ion pairs are formed). -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 1.0 4.0 7.0 10.0 13.0 lo g d o /w ph a) -2.5 -2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 1.0 4.0 7.0 10.0 13.0 lo g d o /w ph b) -2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 2.0 2.5 1.0 4.0 7.0 10.0 13.0 lo g d o /w ph c) 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 1.0 4.0 7.0 10.0 13.0 lo g d o /w ph d) -2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 2.0 2.5 1.0 4.0 7.0 10.0 13.0 lo g d o /w ph e) -1.0 -0.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 1.0 4.0 7.0 10.0 13.0 lo g d o /w ph f) http://dx.doi.org/10.5599/admet.760 susana amézqueta et al. admet & dmpk 8(1) (2020) 98-112 108 the results obtained in this work agree with the ones published before for acidic compounds when a sds-meekc system was employed [6]. table 4. differences between literature (log dlit.) and estimated (log dest.) values of the bases at different ph values through eq. 15. compound ph α log dlit. a) log dest. log dlit. – log dest alprenolol 2.0 1.00 0.13 1.06 -0.93 2.0 1.00 -0.89 1.06 -1.95 3.6 1.00 -0.80 1.06 -1.86 4.5 1.00 0.14 1.06 -0.92 4.5 1.00 -0.77 1.06 -1.83 5.6 1.00 -0.60 1.06 -1.66 6.6 1.00 0.68 1.07 -0.39 6.7 1.00 0.73 1.07 -0.34 7.0 1.00 0.51 1.08 -0.57 7.4 0.99 1.09 1.12 -0.03 7.4 0.99 1.08 1.12 -0.04 8.0 0.98 1.44 1.28 0.16 8.2 0.96 1.60 1.38 0.22 8.5 0.93 1.63 1.59 0.04 9.2 0.73 2.59 2.26 0.33 10.2 0.21 3.20 2.94 0.26 11.0 0.04 2.84 3.08 -0.24 13.0 0.00 3.50 3.11 0.39 nadolol 2.0 1.00 -2.20 -0.84 -1.36 3.6 1.00 -2.10 -0.84 -1.26 4.5 1.00 -2.00 -0.84 -1.16 5.6 1.00 -1.80 -0.83 -0.97 6.7 1.00 -1.60 -0.81 -0.79 7.4 0.99 -1.30 -0.70 -0.60 7.65 0.98 -0.82 -0.61 -0.21 8.2 0.93 0.08 -0.26 0.34 9.2 0.56 0.22 0.70 -0.48 10.2 0.11 0.76 1.17 -0.41 13.0 0.00 0.93 1.26 -0.33 oxprenolol 2.0 1.00 -1.70 -0.08 -1.62 3.6 1.00 -1.40 -0.08 -1.32 4.5 1.00 -1.30 -0.08 -1.22 5.6 1.00 -1.04 -0.08 -0.96 6.7 1.00 -0.30 -0.06 -0.24 7.0 1.00 -0.37 -0.03 -0.34 7.4 0.99 0.13 0.05 0.08 7.65 0.98 0.69 0.13 0.56 8.0 0.96 0.61 0.33 0.28 8.2 0.94 0.76 0.49 0.27 9.2 0.60 1.67 1.49 0.18 10.2 0.13 2.10 2.01 0.09 13.0 0.00 2.16 2.11 0.05 admet & dmpk 8(1) (2020) 98-112 estimation of do/w by meekc doi: http://dx.doi.org/10.5599/admet.760 109 table 4. continued… compound ph α log dlit. a) log dest. log dlit. – log dest. penbutolol 1.2 1.00 1.24 2.20 -0.96 7.4 0.99 1.97 2.33 -0.36 7.65 0.98 2.53 2.41 0.12 13.0 0.00 4.06 4.36 -0.30 pindolol 2.0 1.00 -1.17 0.01 -1.18 2.0 1.00 -1.60 0.01 -1.61 3.6 1.00 -1.55 0.01 -1.56 4.5 1.00 -1.40 0.01 -1.41 4.5 1.00 -1.22 0.01 -1.23 5.6 1.00 -1.35 0.01 -1.36 6.6 1.00 -0.56 0.03 -0.59 6.7 1.00 -0.62 0.04 -0.66 7.0 1.00 -0.92 0.06 -0.98 7.4 0.99 0.29 0.13 0.16 7.4 0.99 -0.10 0.13 -0.23 8.0 0.95 0.19 0.38 -0.19 8.2 0.93 0.46 0.52 -0.06 8.5 0.87 1.18 0.78 0.40 9.2 0.57 1.29 1.45 -0.16 10.2 0.12 1.72 1.91 -0.19 11.0 0.02 1.82 1.98 -0.16 13.0 0.00 1.80 2.00 -0.20 propranolol 2.0 1.00 0.20 1.48 -1.28 2.0 1.00 -0.58 1.48 -2.06 3.6 1.00 -0.55 1.48 -2.03 4.5 1.00 0.18 1.48 -1.30 4.5 1.00 -0.59 1.48 -2.07 5.5 1.00 0.7 1.5 -0.8 5.6 1.00 -0.46 1.48 -1.94 6.5 1.00 0.9 1.50 -0.6 6.6 1.00 0.68 1.50 -0.82 6.7 1.00 0.54 1.50 -0.96 7.0 0.99 0.73 1.53 -0.80 7.4 0.99 1.4 1.6 -0.2 7.4 0.99 1.20 1.60 -0.40 7.4 0.99 1.07 1.60 -0.53 7.65 0.97 1.62 1.68 -0.06 8.0 0.95 1.72 1.86 -0.14 8.2 0.92 1.67 2.01 -0.34 8.5 0.85 1.60 2.27 -0.67 9.2 0.52 2.32 2.91 -0.59 10.2 0.10 2.98 3.33 -0.35 11.0 0.02 2.94 3.38 -0.44 13.0 0.00 3.28 3.40 -0.12 a) from references [32–37]. http://dx.doi.org/10.5599/admet.760 susana amézqueta et al. admet & dmpk 8(1) (2020) 98-112 110 figure 3. correlation between the log do/w from the literature (log dlit.) and the log do/w estimated through the present method (log dest.) at different degrees of ionization. alprenolol (▲), nadolol (■), oxprenolol (●), penbutolol (♦), pindolol ( ), and propranolol ( ). a) comparison of the data when the bases are in their neutral form (0-1 % of ionization, blue full symbol) or fully ionized (≥99 % of ionization, red empty symbol). b) comparison of the data when the bases are partially ionized (1-99 % of ionization). in addition, a line with a slope of 1 and an intercept of 0, and two extra lines corresponding to ±2 sd of the calibration curve (eq. 15) (which corresponds to the 95 % confidence interval) are also shown. conclusions the ttab-meekc system characterized through the solvation parameter model shows great similarity with the octanol-water partition and the sds-meekc systems. thus, both me-based systems can be used for log po/w estimation of neutral solutes. a good correlation has been obtained between log po/w data and log k measured in the ttab-meekc system for neutral compounds. moreover, the applicability of the method has been widen to the estimation of log do/w of partially ionized bases. accurate estimated values have been obtained when the compound is neutral or partially ionized (up to a 99 % of ionization). acknowledgements: financial support from the ministerio de economía y competitividad from the spanish government (ctq2017-88179-p) and the catalan government (2017sgr1074) is acknowledged. afp wishes to thank the university of barcelona for his apif phd fellowship. conflict of interest: the authors declare no conflict of interest. references [1] a. fernández-pumarega, s. amézqueta, s. farré, l. muñoz-pascual, m.h. abraham, e. fuguet, m. rosés. modeling aquatic toxicity through chromatographic systems. analytical chemistry 89 (2017) 7996–8003. [2] s. soriano-meseguer, e. fuguet, a. port, m. rosés. estimation of skin permeation by liquid chromatography. admet & dmpk 6 (2018) 140–152. [3] m.h. abraham, c. treiner, m. roses, c. ràfols, y. ishihama. linear free energy relationship analysis of microemulsion electrokinetic chromatographic determination of lipophilicity. journal of chromatography a 752 (1996) 243–249. -4 -3 -2 -1 0 1 2 3 4 5 6 -3 -2 -1 0 1 2 3 4 5 6 lo g d li t. log dest. a) -4 -3 -2 -1 0 1 2 3 4 5 6 -3 -2 -1 0 1 2 3 4 5 6 lo g d li t. log dest. b) admet & dmpk 8(1) (2020) 98-112 estimation of do/w by meekc doi: http://dx.doi.org/10.5599/admet.760 111 [4] x. subirats, h.p. yuan, v. chaves, n. marzal, m. rosés. microemulsion electrokinetic chromatography as a suitable tool for lipophilicity determination of acidic, neutral, and basic compounds. electrophoresis 37 (2016) 2010–2016. [5] a. fernández-pumarega, s. amézqueta, e. fuguet, m. rosés. feasibility of the estimation of octanolwater distribution coefficients of acidic drugs by microemulsion electrokinetic chromatography. admet & dmpk 6 (2018) 55–60. [6] a. fernández-pumarega, s. amézqueta, e. fuguet, m. rosés. estimation of the octanol-water distribution coefficient of acidic compounds by microemulsion electrokinetic chromatography. journal of pharmaceutical and biomedical analysis 179 (2020) 112981. [7] y. ishihama, y. oda, n. asakawa. hydrophobicity of cationic solutes measured by electrokinetic chromatography with cationic microemulsions. analytical chemistry 68 (1996) 4281–4284. [8] a. fernández-pumarega, s. amézqueta, e. fuguet, m. rosés. determination of the retention factor of ionizable compounds in microemulsion electrokinetic chromatography. analytica chimica acta 1078 (2019) 221–230. [9] d.r. baker, capillary electrophoresis, wiley, hoboken, nj, usa, 1995, p. 39. [10] m.g. khaledi, s.c. smith, j.k. strasters. micellar electrokinetic capillary chromatography of acidic solutes: migration behavior and optimization strategies. analytical chemistry 63 (1991) 1820–1830. [11] e. fuguet, c. ràfols, e. bosch, m. rosés. solute-solvent interactions in micellar electrokinetic chromatography: iv. characterization of electroosmotic flow and micellar markers. electrophoresis 23 (2002) 56–66. [12] m.h. abraham. scales of solute hydrogen-bonding: their construction and application to physicochemical and biochemical processes. chemical society reviews 22 (1993) 73–83. [13] e. fuguet, c. ràfols, e. bosch, m.h. abraham, m. rosés. solute-solvent interactions in micellar electrokinetic chromatography: iii. characterization of the selectivity of micellar electrokinetic chromatography systems. journal of chromatography a 942 (2002) 237–248. [14] e. lázaro, c. ràfols, m.h. abraham, m. rosés. chromatographic estimation of drug disposition properties by means of immobilized artificial membranes (iam) and c18 columns. journal of medicinal chemistry 49 (2006) 4861–4870. [15] m. hidalgo-rodríguez, e. fuguet, c. ràfols, m. rosés. estimation of biological properties by means of chromatographic systems: evaluation of the factors that contribute to the variance of biologicalchromatographic correlations. analytical chemistry 82 (2010) 10236–10245. [16] e.o. dillingham, r.w. mast, g.e. bass, j. autian. toxicity of methyl‐ and halogen‐substituted alcohols in tissue culture relative to structure—activity models and acute toxicity in mice. journal of pharmaceutical sciences 62 (1973) 22–30. [17] bio−loom, biobyte corp. (claremont, ca, usa), http://www.biobyte.com version 1.7. (accessed 25 october 2019). [18] t. sotomatsu, m. shigemura, y. murata, t. fujita. octanol/water partition coefficient of ortho‐ substituted aromatic solutes. journal of pharmaceutical sciences 82 (1993) 776–781. [19] s. banerjee, s.h. yalkowsky, c. valvani. water solubility and octanol/water partition coefficients of organics. limitations of the solubility-partition coefficient correlation. environmental science and technology 14 (1980) 1227–1229. [20] j. li, e.m. perdue, s.g. pavlostathis, r. araujo. physicochemical properties of selected monoterpenes. environment international 24 (1998) 353–358. [21] p.r. rich, r. harper. partition coefficients of quinones and hydroquinones and their relation to biochemical reactivity. febs letters 269 (1990) 139–144. [22] d.e. leszczynski, r.m. schafer. nonspecific and metabolic interactions between steroid hormones and human plasma lipoproteins. lipids 25 (1990) 711–718. [23] d.m. miller. evidence that interfacial transport is rate-limiting during passive cell membrane permeation. biochimica et biophysica acta (bba) biomembranes 1065 (1991) 75–81. http://dx.doi.org/10.5599/admet.760 susana amézqueta et al. admet & dmpk 8(1) (2020) 98-112 112 [24] t. fujita, j. iwasa, c. hansch. a new substituent constant, π, derived from partition coefficients. journal of the american chemical society 86 (1964) 5175–5180. [25] j. iwasa, t. fujita, c. hansch. substituent constants for aliphatic functions obtained from partition coefficients. journal of medicinal chemistry 8 (1965) 150–153. [26] j. de bruijn, f. busser, w. seinen, j. hermens. determination of octanol/water partition coefficients for hydrophobic organic chemicals with the “slow‐stirring” method. environmental toxicology and chemistry 8 (1989) 499–512. [27] a. avdeef. absorption and drug development: solubility, permeability, and charge state, 2nd edition, wiley, hoboken, nj, usa, 2012, p. 108-126. [28] a. andrés, m. rosés, c. ràfols, e. bosch, s. espinosa, v. segarra, j.m. huerta. setup and validation of shake-flask procedures for the determination of partition coefficients (log d) from low drug amounts. european journal of pharmaceutical sciences 76 (2015) 181–191. [29] l. hitzel, a.p. watt, k.l. locker. an increased throughput method for the determination of partition coefficients. pharmaceutical research 17 (2000) 1389–1395. [30] c. hansch, s.m. anderson. the effect of intramolecular hydrophobic bonding on partition coefficients. journal of organic chemistry 32 (1967) 2583–2586. [31] m.h. abraham, h.s. chadha, g.s. whiting, r.c. mitchell. hydrogen bonding. 32. an analysis of water‐ octanol and water‐alkane partitioning and the δlog p parameter of seiler. journal of pharmaceutical sciences 83 (1994) 1085–1100. [32] s. winiwarter, n.m. bonham, f. ax, a. hallberg, h. lennernäs, a. karlén. correlation of human jejunal permeability (in vivo) of drugs with experimentally and theoretically derived parameters. a multivariate data analysis approach. journal of medicinal chemistry 41 (1998) 4939–4949. [33] f. barbato, g. caliendo, m.i. la rotonda, p. morrica, c. silipo, a. vittoria. relationships between octanol-water partition data, chromatographic indices and their dependence on ph in a set of betaadrenoceptor blocking agents. farmaco 45 (1990) 647–663. [34] m. recanatini. partition and distribution coefficients of aryloxypropanolamine β‐adrenoceptor antagonists. journal of pharmacy and pharmacology 44 (1992) 68–70. [35] h.s. huang, r.d. schoenwald, j.l. lach. corneal penetration behavior of β-blocking agents ii: assessment of barrier contributions. journal of pharmaceutical sciences 72 (1983) 1272–1279. [36] d. hellenbrecht, b. lemmer, g. wiethold, h. grobecker. measurement of hydrophobicity, surface activity, local anaesthesia, and myocardial conduction velocity as quantitative parameters of the nonspecific membrane affinity of nine β-adrenergic blocking agents. naunyn-schmiedeberg’s archives of pharmacology 277 (1973) 211–226. [37] n. gulyaeva, a. zaslavsky, p. lechner, m. chlenov, a. chait, b. zaslavsky. relative hydrophobicity and lipophilicity of β-blockers and related compounds as measured by aqueous two-phase partitioning, octanol – buffer partitioning, and hplc. european journal of pharmaceutical sciences 17 (2002) 81– 93. ©2020 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ the effect of drug ionization on lipid-lased formulations for the oral delivery of anti-psychotics doi: https://dx.doi.org/10.5599/admet.830 437 admet & dmpk 8(4) (2020) 437-451; doi: https://doi.org/10.5599/admet.830 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper the effect of drug ionization on lipid-based formulations for the oral delivery of anti-psychotics tahlia r meola # , kara paxton # , paul joyce, hayley b schultz, and clive a prestidge* 1 unisa: clinical and health sciences, university of south australia, city west campus, adelaide, south australia 5000, australia 2 arc centre of excellence in convergent bio-nano science & technology, university of south australia, city west campus, adelaide, south australia 5000, australia # tahlia r meola and kara paxton contributed equally to this experimental research. *corresponding author: e-mail: clive.prestidge@unisa.edu.au ; tel.: +61 8 830 22438; fax: +61 8 8302 2389 received: april 21, 2020; revised: july 15, 2020; published: july 17, 2020 abstract lipid-based formulations (lbfs) are well-known to improve the oral bioavailability of poorly water-soluble drugs (pwsds) by presenting the drug to the gastrointestinal environment in a molecularly dispersed state, thus avoiding the rate-limiting dissolution step. risperidone and lurasidone are antipsychotics drugs which experience erratic and variable absorption, leading to a low oral bioavailability. the aim of this research was to develop and investigate the performance of risperidone and lurasidone when formulated as an emulsion and silica-lipid hybrid (slh). lurasidone and risperidone were dissolved in capmul® mcm at 100% and 80% their equilibrium solubility, respectively, prior to forming a sub-micron emulsion. slh microparticles were fabricated by spray-drying a silica stabilised sub-micron emulsion to form a solid powder. the performances of the formulations were evaluated in simulated intestinal media under digesting conditions, where the emulsion and slh provided a 17-fold and 23-fold increase in lur solubilisation, respectively. however, the performance of ris was reduced by 2.2-fold when encapsulated within slh compared to pure drug. owing to its pka, ris adsorbed to the silica and thus, dissolution was significantly hindered. the results reveal that lbfs may not overcome the challenges of all pwsds and physiochemical properties must be carefully considered when predicting drug performance. ©2020 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords silica; nanoparticles; lurasidone; risperidone; pka; precipitation; interactions; dissolution; lipolysis introduction adherence to antipsychotic medication is a significant challenge for patients, whereby research suggests at least half of patients are non-compliant, leading to illness relapse, hospital remittance or suicide [1]. due to the chronic nature of schizophrenia, the oral route of administration is preferred. however, as a class, antipsychotic drugs are generally poorly water-soluble, resulting in limited absorption and sub-optimal oral bioavailability when administered as a conventional tablet [2, 3]. the ability of lipids to improve the oral absorption and bioavailability of poorly water-soluble drugs is https://dx.doi.org/10.5599/admet.830 https://doi.org/10.5599/admet.830 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:clive.prestidge@unisa.edu.au http://creativecommons.org/licenses/by/4.0/ admet & dmpk 8(4) (2020) 437-451 drug ionisation and lipid-based formulations doi: https://dx.doi.org/10.5599/admet.830 438 well-established in literature [4, 5]. the enhanced performance provided by lipid-based formulations (lbfs) can be attributed to the following proposed mechanisms: (i) lipids in the gastrointestinal tract hinder gastric emptying from the stomach, therefore increase transit time of the drug in the small intestine, allowing greater absorption to occur [6, 7]; (ii) drug can be encapsulated and pre-solubilised in lipid and presented to the gastrointestinal tract in its molecularly dispersed state, therefore drug absorption is no longer limited by its dissolution rate [8, 9], and (iii) exogenous lipids stimulate the release of bile and pancreatic lipase from the gall bladder and pancreas, respectively, which triggers the production of solubilising species that can further solubilise the drug and be transported across the epithelium via enterocytes [6, 7]. despite the magnitude of advantages, lbfs account for only 2-4 % of commercialised pharmaceutical formulations [10, 11]. this is likely attributed to disadvantages associated with their liquid-state, such as poor physical stability, drug leakage, oxidation of lipid excipients and drug precipitation [12, 13]. silica-lipid hybrid (slh) microparticles are a solid-state formulation which encapsulates liquid-lipid within the porous matrix. slh were developed to overcome liquid-state limitations by combining the solubilising effects of lipids with the stabilising effects of silica nanoparticles, to provide enhanced physical stability and superior performance [14]. this lbf has proven successful for numerous poorly water-soluble compounds such as ibuprofen, celecoxib, abiraterone acetate and simvastatin [15-18]. slh microparticles are fabricated by spray-drying a silica nanoparticle-stabilised pickering emulsion to form a solid nanostructured powder. owing to a highly porous matrix, lipase enzymes can readily digest the lipid, thus promoting drug release and absorption [19]. lurasidone (lur) and risperidone (ris) are two orally administered poorly water-soluble antipsychotic drugs which are stipulated to benefit from reformulation within lbfs. lur and ris are weak monoprotic and diprotic basic compounds, respectively, with ph dependent solubility. the pka of lur is 7.6, whereas ris has two pka values of 3.1 and 8.2 [20, 21]. therefore, at neutral ph, simulating the environment within the small intestine, lur is approximately 40 % positively ionised, whereas ris is approximately 80 % positively ionised. subsequently, only 9-19 % of the administrative oral lur dose (ranging between 20-160 mg daily) is absorbed, when co-administered with a meal [22, 23]. in comparison, a daily oral ris dose of 2-16 mg has an oral bioavailability of approximately 70%, with erratic and variable absorption of up to 25 % between patients [24]. thus, the reliance of patient compliance for co-administering medication with food, along with the unpredictable bioavailability, is highly problematic when dosing anti-psychotics, which highlights the critical need for reformulation of lur and ris in delivery systems that enhance absorption while reducing variability. in this study, lur and ris will be hosted within the lipid component of slh microparticles in an attempt to improve their solubilisation in simulated intestinal conditions. the solubilising effects of slh formulations will be compared to a conventional drug in lipid emulsion, and unformulated pure drug. in doing so, the impact of (i) drug ionisation and (ii) formulation physicochemical properties, on in vitro dissolution and solubilisation under non-digesting and digesting conditions will be critically evaluated. it was hypothesised that fabrication of ris and lur within lbfs, specifically slh microparticles, will maximise drug solubilisation, under simulated intestinal conditions in the fasted state, due to the ability for lipid species to maintain the drug in a molecular form. by investigating the impact of drug ionisation on solubilisation behaviour, new insights will be derived regarding the key physicochemical properties of solidstate lbfs required for improving oral delivery performance. https://dx.doi.org/10.5599/admet.830 admet & dmpk 8(4) (2020) 437-451 drug ionisation and lipid-based formulations doi: https://dx.doi.org/10.5599/admet.830 439 experimental materials risperidone (ris), lurasidone (lur) and ziprasidone powders were purchased from hangzhou dayangchem co. ltd (hangzhou, china). capmul® mcm (glyceryl monoand dicaprylate) was a gift from abitec corporation (wisconsin, usa). hydrophilic fumed silica nanoparticles (aerosil® 300) were supplied by evonik degussa (essen, germany). soybean lecithin was purchased from bdh merck (sydney, australia). sodium dihydrogen phosphate (nah2po4), orthophosphoric acid, potassium dihydrogen phosphate (kh2po4), sodium hydroxide pellets (naoh), tween 80, and trizma® maleate, sodium chloride (nacl), calcium chloride dehydrate (cacl2.2h2o), egg lecithin (consisting of 60 % phosphatidylcholine from dried egg yolk), sodium taurodeoxycholate (natdc), 4-bromophenylboronic acid (4-bba) and myristic acid were purchased from sigma aldrich (sydney, australia) and porcine pancreatin extracts from mp biomedicals (sydney, australia). high purity (milli-q) water and hplc grade solvents were used during the study. high-performance liquid chromatography (hplc) the concentrations of ris and lur were quantified using high performance liquid chromatography (hplc) analysis on a shimadzu prominence ultra-fast liquid chromatograph (uflc xr) system, equipped with a phenomenex luna 5 µm c18 100 å (250 mm × 4.6 mm) column, maintained at 40 °c. an isocratic elution method was employed at a flow rate of 1 ml/min, using a mobile phase comprising of 0.05 m nah2po4 buffer (adjusted to ph 3 using orthophosphoric acid) and methanol (40:60 and 25:75 v/v for ris and lur, respectively). ris was detected at a wavelength of 280 nm at a retention time of 3.8 min, whereas lur was detected at 231 nm with an average retention time of 5.8 min. linear calibration curves were generated for both compounds using chromatographic peak area against standard concentrations over a range of 0.02-20 µg/ml. quality control standards were analysed with every sample batch. the precision and accuracy of interand intra-day analyses were within an acceptable range of less than 10 %. liquid chromatography-mass spectrometry (lcms) the concentration of lur for drug solubilisation studies was determined using liquid chromatographymass spectrometry (lcms) on a shimadzu 8030 mass spectrometer (kyoto, japan). a 10 µl sample was injected into a phenomenex kinetex 2.6 µm c18 100 å (50 mm × 3 mm) column, maintained at 40 °c. a gradient elution method consisting of (a) milli-q water + 0.1% formic acid and (b) acetonitrile + 0.1% formic acid was employed at a flow rate of 0.4 ml/min. multiple mode monitoring was used to quantify the internal standard ziprasidone precursor at 413.2 m/z and product at 194.0 m/z. lur precursor was detected at 493.2 m/z and product at 166.1 m/z. linear calibration curves were generated over concentrations ranging from 0.2-1 µg/ml by plotting peak area ratio of analyte: internal standard against known concentrations. preparation of drug loaded lipid emulsions the equilibrium solubilities of lur and ris in miglyol® 812, capmul® mcm and soybean oil were determined by adding excess drug to a known quantity of lipid. the drug-lipid suspensions were sonicated for 10 min to facilitate dissolution, prior to rotating for 72 h at room temperature, protected from light. aliquots were analysed at 24, 48 and 72 h to determine whether equilibrium had been reached. samples were centrifuged at 29066 × g for 20 min at 24 °c to separate any undissolved drug. solvent extraction was performed to extract the drug from the lipid supernatant by diluting the lipid supernatant with methanol (100-fold dilution) and sonicating for 10 min, prior to further centrifugation at 29066 × g for 10 min at 24 °c for phase separation. the resulting supernatant was diluted appropriately and analysed via hplc. https://dx.doi.org/10.5599/admet.830 admet & dmpk 8(4) (2020) 437-451 drug ionisation and lipid-based formulations doi: https://dx.doi.org/10.5599/admet.830 440 submicron lipid emulsions with drug loaded into lipid at 80 % and 100 % the equilibrium solubility for ris and lur, respectively, were formulated. soybean lecithin (6 % w/w) was dissolved in capmul® mcm with the aid of sonication. drug was dissolved in the lipid prior to the addition of milli-q water. the coarse emulsions were sonicated for 1 h to form submicron emulsions. fabrication of drug loaded silica-lipid hybrid microparticles slh microparticles were fabricated following a two-stage process modified from tan et al. [14]. a silica dispersion (5 % w/v) was added to the drug-lipid emulsions to achieve a lipid to silica ratio of 1:1 and 1:2 for ris and lur, respectively. the silica-stabilised pickering emulsions were stirred overnight at room temperature prior to spray-drying to form solid slh microparticles (mini sprayer b-290, büchi labortechnik ag, switzerland). processing parameters were as follows: inlet temperature of 160 °c, aspirator setting at 100%, pump setting at 20% and an outlet temperature of approximately 65 °c. physicochemical characterisation the drug content of the slh formulations were determined by a solvent extraction method. approximately 10 mg of powder was dispersed in 10 ml of methanol. the dispersion was sonicated for 45 min and centrifuged at 29066 × g for 20 min at 24 °c, prior to dilution for hplc analysis. the particle size of the slh formulations were analysed using a malvern mastersizer 3000 hydro lv (malvern, united kingdom). formulations were added to the sample vessel until a laser obscuration of 1020 % was achieved. a particle refractive index of 1.54 was utilised. results are expressed as volume diameter (µm) at 50 % cumulative volume (d (v.0.5)). the surface morphologies of the formulations were analysed through high-resolution analytical scanning electron microscopy (sem) (carl ziess microscopy merlin with a gemini ii column, germany) at an accelerating voltage of 8 kv and spot size of 3-10 μm. samples were attached to double-faced adhesive tape and sputter coated with gold (5-10 nm thickness) prior to imaging. the crystallinity of encapsulated drug was analysed via differential scanning calorimetry (dsc q100, ta instruments, australia). approximately 2 mg of slh was hermetically sealed in aluminium pans and heated at a rate of 10 °c/min from 25 °c to 170 °c for lur and 200 °c for ris, under nitrogen (80 ml/min). in vitro dissolution and release studies in vitro dissolution and release studies were performed under non-sink conditions using a usp type ii apparatus operating at 50 rpm. formulations equivalent to 10 mg ris and 5 mg lur were added to ph 7.4 phosphate buffer containing 1% w/v tween 80, maintained at 37 °c. aliquots were taken at pre-determined time points and replaced with an equal volume of fresh media. samples were immediately centrifuged at 29066 × g for 10 min at 24 °c. the supernatant was diluted appropriately with mobile phase for hplc analysis. to study the interaction between ris and silica, excess ris was added to the ph 7.4 phosphate buffer containing 1 % w/v tween 80 in a glass vial and sonicated for 1 h to facilitate dissolution. temperature was maintained at 37 °c by manual adjustment. vials were rotated at 37 °c for an additional 20 min and an aliquot was taken. aerosil® 300 silica nanoparticles (equivalent mass to slh content) were added to the vial and further aliquots were taken at pre-determined time points. samples were centrifuged at 29066 × g for 10 min at 24 °c, and the resulting supernatant was diluted prior to hplc analysis. https://dx.doi.org/10.5599/admet.830 admet & dmpk 8(4) (2020) 437-451 drug ionisation and lipid-based formulations doi: https://dx.doi.org/10.5599/admet.830 441 solubilisation studies under digesting conditions during in vitro lipolysis simulated fasted-state intestinal media was prepared using a method previously developed by sek et al. [25]. mixed fasted-state micelles with a bile salt: phosphatidylcholine concentration of 5 mm: 1.25 mm were prepared via the following procedure. egg lecithin was dissolved in chloroform prior to evaporation to form a thin, dry lecithin film. natdc and digestion buffer (consisting of 50 mm trizma® maleate, 150 mm sodium chloride and 5 mm calcium chloride) were added, and the mixture was left to stir overnight to produce a clear micellar solution. in vitro lipolysis was performed using a 902 titrando ph-stat titration apparatus (metrohm, switzerland). formulations equivalent to 5 mg lur or ris (357 µl for lur emulsion, 667 mg for lur slh, 160 ul for ris emulsion and 305 mg for ris slh) were added to 18 ml of fasted-state micelles, stirred and maintained at 37 °c in a jacketed vessel. lipolysis was initiated by the addition of 2 ml of pancreatic lipase (1000 tbu). fatty acids produced during lipolysis were automatically titrated with 0.6 m naoh to maintain a constant ph of 6.5 in the digestion medium. the cumulative volume of naoh was converted into percentage lipolysis using equation 1, where v is volume, c is concentration, mw is molecular weight and m is the moles of fatty acids produced. lipolysis (%) = 𝑉𝑁𝑎𝑂𝐻 × 𝐶𝑁𝑎𝑂𝐻 𝑀𝑊𝑙𝑖𝑝𝑖𝑑× 𝑚𝐹𝑎𝑡𝑡𝑦 𝑎𝑐𝑖𝑑 (1) to monitor drug solubilisation, aliquots were taken at pre-determined time points up to 60 min and added to centrifuge tubes containing 0.5 m 4-bba to inhibit lipase action. samples were centrifuged at 35170 × g for 1 h at 37 °c to separate the solubilised aqueous phase and precipitated pellet phase. the aqueous phase was then extracted with acetonitrile, sonicated for 10 min and centrifuged again at 29066 × g for 10 min at 24 °c. samples were diluted appropriately for analysis via hplc and lcms for ris and lur, respectively. zeta potential measurements of silica: drug: fatty acid complexes zeta potential measurements were performed using phase analysis light scattering (pals; zetasizer nano zs, malvern instruments, worcestershire, uk), to investigate possible interactions between the porous silica particles, drug molecules and fatty acids. firstly, the zeta potential of spray dried fumed silica particles, dispersed in lipolysis media at a concentration of 50 µg/ml, was assessed. then, drug dissolved in 50: 50 methanol: water mix was added to this dispersion at a final concentration of 5 µg/ml and the zeta potential was reanalysed. finally, the silica: drug complex was spiked with myristic acid (medium chain, c12 fatty acid; 50 µg/ml) and the zeta potential was again reanalysed. zeta potential measurements are presented as mean ± standard deviation of 3 replicate measurements. statistical analysis experimental data were statistically analysed using an unpaired student’s t-test in graphpad. data was considered statistically significance when p < 0.05. results and discussion fabrication and physicochemical characterisation of ris and lur formulations the drug loading levels of lbfs is dependent on the solubilising capacity of the active pharmaceutical ingredient in the lipid reservoir. therefore, the equilibrium solubilities (seq) of lur and ris, at room temperature, were investigated in miglyol® 812, capmul® mcm and soybean oil, since each lipid comprises of varying chain lengths and compositions (figure 1). capmul® mcm, a mixture of medium chain monoand https://dx.doi.org/10.5599/admet.830 admet & dmpk 8(4) (2020) 437-451 drug ionisation and lipid-based formulations doi: https://dx.doi.org/10.5599/admet.830 442 di-glyceride medium chain fatty acids, provided the highest solubilisation for both compounds, reaching a solubilisation capacity of 39.7 ± 2.8 mg/g and 14.0 ± 0.6 mg/g for ris and lur, respectively. miglyol® 812, a mixture of medium chain triglycerides, revealed a higher solubilisation for lur, reaching a seq of 12.8 ± 0.3 mg/g, compared to 3.13 ± 0.1 mg/g for ris. no significant difference was found between ris and lur in soybean oil, which is a mixture of long chain triglycerides (8.53 ± 0.2 mg/g for ris and 7.99 ± 0.6 mg/g for lur). therefore, capmul® mcm was chosen for further formulation development since it provided the highest solubilisation capacity for both ris and lur and allows for the highest drug loading levels within each lbf. figure 1. equilibrium solubilities of ris and lur in various lipids. each value represents the mean ± sd, n = 3. the clinical oral daily dose of lur ranges from 20 – 160 mg/day, compared to 2 – 16 mg/day for ris [23, 24]. consequently, due to different dosing requirements, submicron emulsions were fabricated by dissolving lur and ris in capmul® mcm at 100 % and 80 % their seq, respectively. table 1 summarises the drug loading levels and loading efficiencies of the investigated formulations. the submicron emulsions for lur and ris will be referred to as lur emulsion and ris emulsion, respectively, herein. the drug loading level of the ris emulsion was 3.12 ± 0.0 % w/w, compared to 1.4 ± 0.0 % w/w for the lur emulsion. table 1. formulation drug loading levels and encapsulation efficiencies (each value represents the mean ± sd, n = 3). slh formulations were manufactured by spray-drying the submicron emulsion with a silica nanoparticle dispersion to form solid slh microparticles. ris-slh particles comprised of a 1: 1 lipid: silica ratio and obtained a final drug load of 1.64 ± 0.2 % w/w. in attempt to increase the formulation drug load of lur-slh to meet the higher dosing requirements, a 2:1 ratio of lipid:silica was utilised, achieving a final drug loading level of 0.75 % w/w. both slh formulations formed free-flowing powders following the spray drying procedure. particle size analysis after a 5 min dispersion period in aqueous media revealed the particle size of ris-slh to be slightly larger than lur-slh (d[v,0.5] of 9.39 ± 1.3 µm and 7.46 ± 1.5 µm for ris-slh and lur-slh, respectively). as demonstrated by the sem images in figure 2, ris-slh consisted of spherical porous microparticles, compared to lur-slh which appeared to have a collapsed porous structure. as the spray-drying parameters were consistent for both formulations, the different shape may be attributed to the different lipid:silica composition, i.e. a greater amount of lipid resulted in the collapse of the microparticles. formulation drug loading (% w/w) encapsulation efficiency (%) ris emulsion 3.12 ± 0.0 100 ± 0.0 ris-slh 1.64 ± 0.2 99.1 ± 1.3 lur emulsion 1.4 ± 0.0 100 ± 0.0 lur-slh 0.75 ± 0.0 81.4 ± 0.0 https://dx.doi.org/10.5599/admet.830 admet & dmpk 8(4) (2020) 437-451 drug ionisation and lipid-based formulations doi: https://dx.doi.org/10.5599/admet.830 443 figure 2. sem images depicting the surface morphologies of (a) ris-slh, and (b) lur-slh. given lbfs are known for their ability to maintain poorly water-soluble drugs in a non-crystalline state, the degree of crystallinity of lur and ris encapsulated in the slh formulations was investigated using dsc (figure 3). the thermograms of pure ris and lur powders displayed strong endothermic peaks at approximately 173 °c and 150 °c, respectively, corresponding to the drug’s melting points and thus, indicating crystallinity. due to the low drug loading levels for slh formulations, physical mixtures containing 0.1 % w/w lur in silica and 1 % w/w ris in silica were analysed to determine whether dsc provided adequate sensitivity to detect the drug within the formulations. the characteristic peaks were present for the physical mixtures; however, the ris endothermic peak in the physical mixture shifted to approximately 177 °c due to potential molecular interaction with the silica. the characteristic peaks were absent for ris-slh and lur-slh, confirming that the compounds were encapsulated within the silica-lipid matrix in a non-crystalline state. although dsc was not performed on the pre-cursor submicron emulsions, it is hypothesised that both ris and lur within the emulsions would be non-crystalline due to being loaded at or below their equilibrium solubilities in the lipid. figure 3. dsc thermograms of (a) lur formulations and (b) ris formulations. in vitro dissolution and release studies impact of lur reformulation on dissolution and release in vitro dissolution and release studies were conducted in simulated intestinal media under non-sink conditions. pure lur exhibited the lowest rate and extent of dissolution reaching a maximum of 16.0 ± 0.5 % (1.2 µg/ml) after 2 h, reflecting the drug’s poor aqueous solubility and crystalline nature (figure 4). when formulated as a submicron emulsion, the release of lur was enhanced 3.4-fold, compared to pure lur (52.9 ± 0.8 % release, 3.9 µg/ml). it is postulated that this is due to the pre-solubilisation of the drug in https://dx.doi.org/10.5599/admet.830 admet & dmpk 8(4) (2020) 437-451 drug ionisation and lipid-based formulations doi: https://dx.doi.org/10.5599/admet.830 444 the lipid, therefore avoiding the critical rate-limiting dissolution step. lur-slh displayed a rapid release of lur, achieving complete drug release after a 10 min period. similar to the submicron emulsion, the rapid release rate of lur-slh can be explained by the encapsulation of lur in the non-crystalline state. however, it is apparent that the 1.9-fold enhancement in drug release, compared to the submicron emulsion, is due to the presence of hydrophilic silica and porous structure, aiding in drug wettability and dissolution [16]. figure 4. in vitro dissolution profiles of pure lur [▼], lur emulsion [●], and lur-slh [■]. each value represents the mean ± sd, n = 3. impact of ris reformulation on dissolution and release the opposite trend in performance was evident for ris formulations, whereby the dissolution of pure ris was superior to both lbfs, reaching 28.0 ± 1.0 % (156.1 µg/ml) after a 2 h period (figure 5). during the initial 5 min, the rate of ris release from the ris emulsion mimicked the release of pure ris, prior to plateauing at 17.0 ± 0.5 % (83.7 µg/ml) for the remaining experimental period. ris-slh exhibited the lowest extent of drug release reaching 7.7 ± 0.0 % (38.5 µg/ml) after 120 min, a 4-fold reduction compared to pure drug dissolution. to investigate the reason for this poor performance, a study was conducted to determine whether any interaction between silica and ris was present, by introducing aerosil® silica to dissolution media containing dissolved ris (figure 5b). results revealed that the addition of silica triggered a significant reduction in the concentration of ris in solution from 28.0 % to 15.2 ± 0.5 %. this was hypothesised to be due to the ionic interactions and adsorption of the positively charged ris to the negatively charged silica. nevertheless, drug release from the ris emulsion was inferior to pure ris, even though silica was absent. therefore, it is reasonable to suggest that due to the lipophilic nature of ris (log p = 3.04) [21], ris favours capmul® mcm, resulting in less drug partitioning into the aqueous phase and reducing the overall extent of drug release. it is recognised that a limitation exists when analysing lbfs during in vitro release studies as no enzyme is present to digest the lipid to release the drug and thus, the formulation performance may be underestimated. hence, in vitro solubilisation studies under digesting conditions were subsequently performed. in vitro solubilisation studies under digesting conditions impact of lur reformulation on lipolysis and drug solubilisation similar to the previous dissolution study, lbfs were successful in significantly enhancing the solubilisation of lur under digesting conditions, when compared to pure drug (figure 6a). after the 60 min experimental period, lur-slh achieved the highest drug solubilisation of 2.7 ± 0.2 % (6.6 µg/ml), compared to the lur emulsion which reached 1.2 ± 0.0 % (4.7 µg/ml) solubilisation. only 0.1 ± 0.0 % percent of the pure drug was solubilised in fasted state digesting conditions. the enhancement in lur solubilisation from both lbfs is attributable to the digestion of the lipid component, which triggers the formation of colloidal https://dx.doi.org/10.5599/admet.830 admet & dmpk 8(4) (2020) 437-451 drug ionisation and lipid-based formulations doi: https://dx.doi.org/10.5599/admet.830 445 species, composed primarily of monoand di-glycerides and fatty acids. these colloidal lipid species, such as micelles, mixed micelles and vesicles, have demonstrated the ability to improve the dissolution of numerous poorly water-soluble drugs [7, 26], and the improved drug solubilisation during lipid digestion indicates that lur is readily incorporated within these species. figure 5. in vitro dissolution profiles of (a) pure ris [▲], ris emulsion [◆], and ris-slh [✕], and (b) pure ris solution followed by the addition of silica at t = 0 min. each value represents the mean ± sd, n = 3. figure 6. (a) in vitro solubilisation profiles under digesting conditions, and (b) lipolysis data for pure lur [▼], lur emulsion [●], and lur-slh [■]. each value represents the mean ± sd, n = 3. in vitro lipase-mediated digestion of lur lbfs, under simulated intestinal fasted state conditions, revealed that lur-slh enhanced lipolysis, compared to the lur emulsion, by digesting 33.0 ± 0.3 % of lipid hosted within the porous silica matrix (figure 6b). in contrast, only 24.1 ± 0.2 % of the lur emulsion was digested over the course of 60 min lipolysis period. this is in accordance with previous studies that have highlighted the increased surface area of the porous silica matrix affords enhanced adsorption of lipase molecules in their active conformation, triggering an increase in catalytic activity [27]. furthermore, owing to the negatively charged aerosil® silica, the digestion of slh has shown to trigger the release and repel the negatively charged free fatty acids from the nanostructured matrix into the aqueous phase [19]. the ability for slh to promote lipid digestion and increase the concentration of fatty acid-rich solubilising vesicles within the aqueous phase are therefore considered the driving force for improved lur solubilisation, when compared the lur emulsion. drug precipitation is commonly observed with lbfs [28, 29]. upon digestion of the exogenous lipid, the https://dx.doi.org/10.5599/admet.830 admet & dmpk 8(4) (2020) 437-451 drug ionisation and lipid-based formulations doi: https://dx.doi.org/10.5599/admet.830 446 solubilisation capacity for the drug is enhanced and thus supersaturated drug concentrations may be generated [30]. the current solubilisation study was conducted in non-sink conditions and thus the lack of solubilising capacity of the aqueous phase resulted in rapid drug precipitation from lbfs, as evident within the initial 10 min of solubilisation (figure 6a). however, precipitation during in vitro studies does not necessarily correlate with a reduction in bioavailability in vivo, as many studies suggests that this model over predicts the extent of drug precipitation due to the absence of an absorption compartment [31, 32]. impact of ris reformulation on lipolysis and drug solubilisation in contrast to lur, pure ris displayed the greatest level of drug solubilisation under digesting conditions, reaching a maximum of 93.4 ± 5.3 % (254 µg/ml) after 60 min, followed by the ris emulsion and ris-slh (figure 7a). the drug solubilisation observed for pure ris is indicative of the high solubility of the ionised drug in neutral conditions. a drug solubilisation level of 80.5 ± 0.8 % (209 µg/ml), was observed for the ris emulsion, likely owing to incomplete digestion of the lipid and thus, only partial drug partitioning to the aqueous phase. however, it is noteworthy that the performance of the ris emulsion, relative to pure drug, was significantly greater during solubilisation studies under digestion conditions compared to non-digesting conditions, indicating the importance of lipase-provoked drug release. however, for ris-slh, a significant 2.2-fold reduction in solubilisation was evident, leading to a ris solubilisation of only 43.2 ± 2.2 % (109 µg/ml) after 60 min. figure 7. (a) in vitro solubilisation profiles under digesting conditions, and (b) lipolysis data for pure ris [▲], ris emulsion [◆], and ris-slh [✕]. each value represents the mean ± sd, n = 3. the reduced ability for ris-slh to promote solubilisation under digesting conditions can be attributed a reduction in lipolysis kinetics and the interaction between the ionised drug and the negatively charged silica surfaces and fatty acids produced during digestion. in contrary to the lipid digestion observed for lur lbfs, the rate and extent of lipolysis was greatest for the ris emulsion, with 34.9 ± 0.3 % lipid being digested over the 60 min period, compared to only 30.2 ± 0.4 % for ris-slh (figure 7b). the difference in lipolysis behaviour between the two drugs, when encapsulated within slh, is due to the relative degrees of ionisation, where ris is mostly present in its ionised, positively charged form. since the porous silica matrix and the fatty acids produced from lipolysis are both negatively charged, complexation occurs between ionised ris and both the silica surface and lipid digestion products, as depicted in figure 8. this interaction was confirmed by analysing the zeta potential of the various silica: drug: fatty acid complexes that are predicted to form during lipase-mediated digestion of both lurand ris-slh (table 2). when porous silica particles were spiked with dissolved ris, the zeta potential increased from -19.1 ± 2.7 mv to -12.3 ± 0.8 mv, highlighting the capacity for ionised ris to adsorb to the negatively charged silica surface. however, once https://dx.doi.org/10.5599/admet.830 admet & dmpk 8(4) (2020) 437-451 drug ionisation and lipid-based formulations doi: https://dx.doi.org/10.5599/admet.830 447 the silica: ris complex was exposed to fatty acids, representative of those present during in vitro lipolysis (i.e. medium chain length fatty acids), the zeta potential again decreased to -15.0 ± 0.3 mv, indicating the propensity for negatively charged fatty acids to adsorb onto the ris-coated silica surface through electrostatic interactions. in contrast, zeta potential variations when porous silica particles were exposed to lur were statistically insignificant; thus, highlighting the ability for lur and negatively charged fatty acids to be repelled from the bare silica surface (table 2). figure 8. schematic representation showing the difference in lur and ris solubilisation when encapsulated within slh microparticles. at neutral ph, ris is predominately ionised and carries a positive charge. as the silica matrix and fatty acids which are liberated upon lipase-mediated digestion of the lipid are negatively charged, complexation occurs between the ionised ris and the silica surface, interfering with the availability for lipase-lipid interaction, and thus hindering lipid digestion and ris solubilisation. conversely, lur is predominately unionised. therefore, lipase can readily adsorb and digest the lipid, facilitating the production of negatively charged and highly solubilising lur micelles which can be expelled into the aqueous phase due to repulsion from the negatively charged silica surface. table 2. zeta potential analysis of the various complexes formed between porous silica particles, drug molecules and fatty acids (each value represents the mean ± sd, n = 3). drug zeta potential (mv) porous silica particles (psp) psp: drug complex psp: drug complex: fa complex lurasidone -20.3 ± 3.3 -19.8 ± 1.6 -22.6 ± 4.2 risperidone -19.1 ± 2.7 -12.3 ± 0.8 -15.0 ± 0.3 the adsorption of positively charged ris molecules on the silica surface is therefore hypothesised to interfere with the ability for lipase to adsorb to the exposed silica surface in its active form, removing the potential to form a substrate-enzyme complex that enhances hydrolytic activity [33]. the digestion of lipid hosted within porous silica typically leads to the rapid expulsion of fatty acids from the silica pores, into the aqueous environment, due to an electrostatic repulsion interaction between the negatively charged silica surface and fatty acids. this process further enhances lipid digestion within slh due to the absence of surface active, amphiphilic fatty acids at the lipid-in-water interface, which competitively inhibit lipase adsorption [27]. however, since ionised ris can complex with fatty acids, their ability to be expelled into the aqueous phase is inhibited, and thus, they are likely to ‘spoil’ the lipid-in-water interface by retarding lipase adsorption (figure 8), leading to reduced lipolysis kinetics. dening et al. recently characterised such drug-carrier and drug-fatty acid interactions when encapsulating ionisable drugs within solid-state lbfs composed of the smectite clay, montmorillonite, and https://dx.doi.org/10.5599/admet.830 admet & dmpk 8(4) (2020) 437-451 drug ionisation and lipid-based formulations doi: https://dx.doi.org/10.5599/admet.830 448 medium chain triglycerides [34]. it was revealed that drug solubilisation under simulated intestinal digesting conditions for the model weak base, blonanserin, were reduced 3-fold compared to the pure drug, due to the electrostatic-driven interactions between ionised drug and montmorillonite and fatty acids. inferior in vitro solubilisation performance correlated well with in vivo pharmacokinetics, whereby blonanserin bioavailability was reduced in montmorillonite-lipid hybrids, compared to the pure drug [35]. thus, based on these findings, it may be expected that oral ris bioavailability will also be reduced in vivo, when formulated with slh. however, a greater dynamic flow is observed in vivo, and transcellular transport may be facilitated by uptake of the ris-silica complex by intestinal cells; therefore, further analysis in animal models is necessary to determine whether performance in vivo will correlate with in vitro results for ris-slh. the importance of physicochemical properties in formulation design a summary of the impact of lbf type on ris and lur in vitro dissolution and solubilisation, relative to pure drug, is shown in figure 9. a constant reduction in ris performance is observed when ris was formulated as a lbf, especially when fabricated as slh. as described previously, this is a result from interactions between ionised ris and negatively charged silica, thus hindering drug release. in contrast, such strong interactions were not present for lur due to existing predominately in an unionised form. figure 9. formulation performance for ris and lur, emulsions and slh, relative to pure drug, during (a) in vitro dissolution and (b) in vitro solubilisation under digestion conditions. the conflicting performance observed for ris and lur when formulated as a lbf demonstrates the importance of drug pka and ionisation when developing a novel delivery system. a blanket statement has been used in literature surrounding the application of lbfs to overcome poor water-solubility and improve dissolution kinetics [4, 36, 37], however the performance of ris diverges from this statement whereby a lbf hindered drug performance. predictive criteria for determining suitable formulation approaches for poorly water-soluble compounds from current knoweldge is scarce, yet the present study reveals the need for a greater understanding of physicochemical parameters, such as pka, in relation in performance when developing novel drug delivery systems. conclusions the poorly water-soluble antipsychotic drugs, ris and lur, have been successfully formulated and physicochemically characterised as lipid emulsions and slh microparticles. the lbfs significantly improved the performance of lur up to 23-fold during in vitro solubilisation studies in comparison to the unformulated drug, however a 2.2-fold reduction in solubilisation was observed for ris due to its strong https://dx.doi.org/10.5599/admet.830 admet & dmpk 8(4) (2020) 437-451 drug ionisation and lipid-based formulations doi: https://dx.doi.org/10.5599/admet.830 449 interaction with silica, demonstrating that the application of lbfs and the resulting performance may not be readily predicted. the results reveal that the pka values of a compound must be considered during formulation development, and a general statement that lbfs may overcome the challenges of all poorly water-soluble drugs is not necessarily correct. therefore, further investigation is required to develop predictive criteria to evaluate the performance of lbfs for poorly water-soluble compounds. acknowledgements: the australian government research training program is acknowledged for the phd scholarships of tahlia r. meola and hayley b. schultz. the australian research council’s centre of excellence in convergent bio-nano science and technology are acknowledged for research funding and support. this work was performed (in part) at the south australian node of the australian national fabrication facility under the national collaborative research infrastructure strategy. conflict of interest: none. references [1] d.o. perkins. predictors of noncompliance in patients with schizophrenia. journal of clinical psychiatry 63 (2002) 1121-1128. [2] j. dunbar-jacob, m.k. mortimer-stephens. treatment adherence in chronic disease. journal of clinical epidemiology 54 (2001) s57-s60. [3] t.j. dening, s. rao, n. thomas, c.a. prestidge. oral nanomedicine approaches for the treatment of psychiatric illnesses. journal of controlled release 223 (2016) 137-156. [4] o.m. feeney, m.f. crum, c.l. mcevoy, n.l. trevaskis, h.d. williams, c.w. pouton, w.n. charman, c.a.s. bergström, c.j.h. porter. 50 years of oral lipid-based formulations: provenance, progress and future perspectives. advanced drug delivery reviews 101 (2016) 167-194. [5] w.n. charman, c.j.h. porter, s. mithani, j.b. dressman. physicochemical and physiological mechanisms for the effects of food on drug absorption: the role of lipids and ph. journal of pharmaceutical sciences 86 (1997) 269-282. [6] v. jannin, j. musakhanian, d. marchaud. approaches for the development of solid and semi-solid lipid-based formulations. advanced drug delivery reviews 60 (2008) 734-746. [7] c.j.h. porter, c.w. pouton, j.f. cuine, w.n. charman. enhancing intestinal drug solubilisation using lipid-based delivery systems. advanced drug delivery reviews 60 (2008) 673-691. [8] c.j.h. porter, n.l. trevaskis, w.n. charman. lipids and lipid-based formulations: optimizing the oral delivery of lipophilic drugs. nature reviews drug discovery 6 (2007) 231-248. [9] j.f. cuiné, c.l. mcevoy, w.n. charman, c.w. pouton, g.a. edwards, h. benameur, c.j.h. porter. evaluation of the impact of surfactant digestion on the bioavailability of danazol after oral administration of lipidic self-emulsifying formulations to dogs. journal of pharmaceutical sciences 97 (2008) 995-1012. [10] d.j. hauss. oral lipid-based formulations. advanced drug delivery reviews 59 (2007) 667-676. [11] r.g. strickley, currently marketed oral lipid-based dosage forms: drug products and excipients, in oral lipid-based formulations: enhancing the bioavailability of poorly water-soluble drugs, informa healthcare, new york, 2007 [12] s. dokania, a.k. joshi. self-microemulsifying drug delivery system (smedds)--challenges and road ahead. drug delivery 22 (2015) 675-690. [13] c. washington. stability of lipid emulsions for drug delivery. advanced drug delivery reviews 20 (1996) 131-145. [14] a. tan, s. simovic, a.k. davey, t. rades, c.a. prestidge. silica-lipid hybrid (slh) microcapsules: a novel oral delivery system for poorly soluble drugs. journal of controlled release 134 (2009) 62-70. https://dx.doi.org/10.5599/admet.830 admet & dmpk 8(4) (2020) 437-451 drug ionisation and lipid-based formulations doi: https://dx.doi.org/10.5599/admet.830 450 [15] h.b. schultz, n. thomas, s. rao, c.a. prestidge. supersaturated silica-lipid hybrids (super-slh): an improved solid-state lipid-based oral drug delivery system with enhanced drug loading. european journal of pharmaceutics and biopharmaceutics 125 (2018) 13-20. [16] a. tan, a.k. davey, c.a. prestidge. silica-lipid hybrid (slh) versus non-lipid formulations for optimising the dose-dependent oral absorption of celecoxib. pharmaceutical research 28 (2011) 2273-2287. [17] h.b. schultz, p. joyce, n. thomas, c.a. prestidge. supersaturated-silica lipid hybrids improve in vitro solubilization of abiraterone acetate. pharmaceutical research 37 (2020) 77. [18] t.r. meola, h.b. schultz, k.f. peressin, c.a. prestidge. enhancing the oral bioavailability of simvastatin with silica-lipid hybrid particles: the effect of supersaturation and silica geometry. european journal of pharmaceutical sciences 150 (2020) 105357. [19] p. joyce, t.j. barnes, b.j. boyd, c.a. prestidge. porous nanostructure controls kinetics, disposition and self-assembly structure of lipid digestion products. rsc advances 6 (2016) 78385-78395. [20] s. shah, b. parmar, m. soniwala, j. chavda. design, optimization, and evaluation of lurasidone hydrochloride nanocrystals. aaps pharmscitech 17 (2016) 1150-1158. [21] g. mannens, w. meuldermans, e. snoeck, j. heykants. plasma protein binding of risperidone and its distribution in blood. psychopharmacology 114 (1994) 566-572. [22] l. citrome. lurasidone for schizophrenia: a review of the efficacy and safety profile for this newly approved second-generation antipsychotic. international journal of clinical practice 65 (2011) 189210. [23] l. citrome. using oral ziprasidone effectively: the food effect and dose-response. adv ther 26 (2009) 739-748. [24] v. madaan, d.p. bestha, v. kolli, s. jauhari, r.c. burket. clinical utility of the risperidone formulations in the management of schizophrenia. neuropsychiatric disease and treatment 7 (2011) 611-620. [25] l. sek, c.j.h. porter, w.n. charman. characterisation and quantification of medium chain and long chain triglycerides and their in vitro digestion products, by hptlc coupled with in situ densitometric analysis. journal of pharmaceutical and biomedical analysis 25 (2001) 651-661. [26] t.s. wiedmann, l. kamel. examination of the solubilization of drugs by bile salt micelles. journal of pharmaceutical sciences 91 (2002) 1743-1764. [27] p. joyce, h. gustafsson, c.a. prestidge. engineering intelligent particle-lipid composites that control lipase-mediated digestion. advances in colloid and interface science 260 (2018) 1-23. [28] n. thomas, t. rades, a. müllertz. recent developments in oral lipid-based drug delivery. journal of drug delivery science and technology 23 (2013) 375-382. [29] s. chakraborty, d. shukla, b. mishra, s. singh. lipid – an emerging platform for oral delivery of drugs with poor bioavailability. european journal of pharmaceutics and biopharmaceutics 73 (2009) 1-15. [30] a.t. larsen, a.g. ohlsson, b. polentarutti, r.a. barker, a.r. phillips, r. abu-rmaileh, p.a. dickinson, b. abrahamsson, j. østergaard, a. müllertz. oral bioavailability of cinnarizine in dogs: relation to snedds droplet size, drug solubility and in vitro precipitation. european journal of pharmaceutical sciences 48 (2013) 339-350. [31] p.j. sassene, m.h. michaelsen, m.d. mosgaard, m.k. jensen, e. van den broek, k.m. wasan, h. mu, t. rades, a. müllertz. in vivo precipitation of poorly soluble drugs from lipid-based drug delivery systems. molecular pharmaceutics 13 (2016) 3417-3426. [32] n. thomas, k. richter, t.b. pedersen, r. holm, a. müllertz, t. rades. in vitro lipolysis data does not adequately predict the in vivo performance of lipid-based drug delivery systems containing fenofibrate. the aaps journal 16 (2014) 539-549. [33] p. joyce, c.p. whitby, c.a. prestidge. nanostructuring biomaterials with specific activities towards digestive enzymes for controlled gastrointestinal absorption of lipophilic bioactive molecules. advances in colloid and interface science 237 (2016) 52-75. https://dx.doi.org/10.5599/admet.830 admet & dmpk 8(4) (2020) 437-451 drug ionisation and lipid-based formulations doi: https://dx.doi.org/10.5599/admet.830 451 [34] t.j. dening, s. rao, n. thomas, c.a. prestidge. montmorillonite-lipid hybrid carriers for ionizable and neutral poorly water-soluble drugs: formulation, characterization and in vitro lipolysis studies. international journal of pharmaceutics 526 (2017) 95-105. [35] t.j. dening, n. thomas, s. rao, c. van looveren, f. cuyckens, r. holm, c.a. prestidge. montmorillonite and laponite clay materials for the solidification of lipid-based formulations for the basic drug blonanserin: in vitro and in vivo investigations. molecular pharmaceutics 15 (2018) 41484160. [36] h. mu, r. holm, a. müllertz. lipid-based formulations for oral administration of poorly water-soluble drugs. international journal of pharmaceutics 453 (2013) 215-224. [37] c.w. pouton. formulation of poorly water-soluble drugs for oral administration: physicochemical and physiological issues and the lipid formulation classification system. european journal of pharmaceutical sciences 29 (2006) 278-287. ©2020 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) https://dx.doi.org/10.5599/admet.830 http://creativecommons.org/licenses/by/3.0/ molecular docking studies of salubrinal and its analogs as inhibitors of the gadd34:pp1 enzyme doi: 10.5599/admet.632 140 admet & dmpk 7(2) (2019) 140-150; doi: http://dx.doi.org/10.5599/admet.632 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper molecular docking studies of salubrinal and its analogs as inhibitors of the gadd34:pp1 enzyme pavlo v. zadorozhnii*, ihor o. pokotylo, vadym v. kiselev, oxana v. okhtina, aleksandr v. kharchenko department of organic substances and pharmaceutical preparations, ukrainian state university of chemical technology, gagarin ave., 8, dnipro 49005, ukraine *corresponding author: e-mail: torfp@i.ua received: october 31, 2018; revised: february 25, 2019; available online: april 05, 2019 abstract the phenomenon of the endoplasmic reticulum (er) stress as a molecular pathophysiological process underlies diseases as cancer, diabetes mellitus, myocardial infarction, neurodegenerative disorders, diseases of the urinary system, disorders associated with bone integrity, etc. to prevent er stress, salubrinal, which is a phosphatase inhibitor of the eukaryotic translation initiation factor gadd34:pp1, is currently being intensively studied. the aim of this work is to search for new analogues of this drug using molecular docking methods. optimization of the geometry of the studied structures and molecular docking was carried out using the arguslab 4.0.1 software package. the three-dimensional crystal structure of the gadd34: pp1 enzyme (pdb id: 4xpn) was loaded in the pdb format from the protein molecule data bank. the model of the binding site was created on the basis of the phosphoric acid residue (403 po4). the dimensions of the binding site were set manually and were 40.000 å along the x-axis, 40.000 å the y-axis and 40.000 å the z-axis. the docking was done with a flexible ligand, and the semi-empirical ascore function was used for the scoring procedure. it was shown that for the salubrinal molecule the most favorable was the conformation stabilized by the intramolecular hydrogen bond formed between the hydrogen atom of the thiourea fragment and the oxygen atom of the amide fragment. according to molecular docking data, six compounds from the fifty-four analyzed analogues of salubrinal exceed it in the stability of the complex formed with gadd34:pp1. the results of this work can be used to create new phosphatase inhibitors of the eukaryotic translation initiation factor gadd34:pp1. keywords salubrinal; molecular docking; gadd34:pp1; rmsd; endoplasmic reticulum stress introduction endoplasmic reticulum (er) is an intracellular membrane organelle that is extremely sensitive to changes in homeostasis. the membrane er is integrated with the cell nucleus membrane. the internal er space opens directly into the perinuclear space, which accompanies the contact of the er signalling device with the genetic material. there are granular (rough) er and agranular (smooth) er. smooth er is located on the periphery of the organelle and is responsible for the synthesis of lipids, steroids, the metabolism of carbohydrates, medicines and other exogenous products [1,2]. http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:torfp@i.ua admet & dmpk 7(2) (2019) 140-150 molecular docking studies of salubrinal doi: 10.5599/admet.632 141 rough er is an extension of the cell nucleus membrane. on its cytosolic surface, ribosomes are deposited, which provide for the translation of the protein directly into the er cavity through the system of transmembrane channels. inside the granular er, "immature" protein molecules are foldable, i.e. take a correct spatial conformation. all unfolded or incorrectly folded proteins are caught and necessarily destroyed. accumulation of misfolded protein molecules results in a functional overload of er. this phenomenon is called er stress and it leads to disorders in the normal functioning of the cell and threatens it with death [1,2]. figure 1. schematic simplified image of the main signal-sensory systems of er stress over the past 15 years [3], molecular mechanisms of er stress have been intensively studied as a fundamental phenomenon of cell protection from the action of various factors and as a molecular pathophysiological process leading to many severe diseases [3-21]. figure 1 schematically depicts the response of the cell to er stress, which is necessary for the cell to find ways to escape from the state of stress caused by the accumulation of unfolded or misfolded proteins, and which is mediated by three signal-sensory systems that begin in the er lumen and terminate in the cytoplasm and nucleus [22-26]. induction of er stress stops the penetration of synthesized proteins into it and accompanies both the proper folding of proteins, which are already in it and the degradation of misfolded ones. this is necessary for the survival of the cell under the conditions of the factors that induce this stress, or death of the cell through the apoptosis system associated with er [27,28]. the main signal-sensory er stress systems (perk, atf6 and ire1), which originate in its lumen under the conditions of accumulation of unfolded or incorrectly folded proteins in it, initiate total repression of translation initiation by phosphorylation of eukaryotic translation initiation factor 2α (eif2α), and activation of the transcription of stress dependent genes by the formation of an active form of transcription factors atf4 and atf6, as well as an alternative splice variant of the transcription factor xbp1 (x-box protein-1), which controls the expression of the cell genes [19,26]. p.v. zadorozhnii et al. admet & dmpk 7(2) (2019) 140-150 142 in this way, the response of cells to er stress, which is mediated by the three signal-sensory systems, is necessary for the cell to find possible ways out of the state of stress caused by the accumulation of unfolded or incorrectly folded proteins in the er lumen. eif2α is a key participant in protein translation because it is responsible for binding the 40s ribosomal subunit to trnaimet (initiation of methionine trna), which recognizes the mrna start codon and starts the synthesis of the peptide chain [19]. perk phosphorylates eif2α translating it into an inactive eif2αp form. however, the holoenzyme complex gadd34: pp1 dephosphorylates eif2αp, again translating it into an active eif2α form. in 2005, m. boyce and colleagues reported that salubrinal (fig. 2) acted as a phosphatase inhibitor gadd34:pp1, selective for eukaryotic translation initiation factor 2α (eif2α) [29]. thus, salubrinal weakens the synthesis of unfolded or misfolded proteins contributing to the preservation of homeostasis in er and saving cells from apoptosis. figure 2. structure of the salubrinal molecule since the beginning of intensive studies of salubrinal, its protective effect has been confirmed in a number of studies [30]. although salubrinal is currently under development, we can already say with certainty about its prospects in the treatment of diabetes [31], myocardial infarction [32], neurodegenerative disorders [33,34], oncological diseases [35], diseases of the genitourinary system [36] and disorders related to the integrity of bone tissue [37,38]. work is underway to study its toxicity and the development of analogues [39]. in this paper, using the methods of molecular docking [40,41], we have established the binding site of the salubrinal preparation with holoenzyme gadd34:pp1 and searched for analogues of this drug. materials and methods computer specification all calculations were carried out on a toshiba personal computer, the satellite l650d model, amd phenom(tm) ii p820 triple-core processor. a 64-bit operating system was used. ligand preparation the search for structures for research was conducted in the scifinder database (https://scifinder.cas.org) (see supporting information). prior to molecular docking, the structures of all the compounds studied were optimized within the semiempirical pm3 method [42] using the arguslab 4.0.1 software package [43-47]. the calculation of the electron density distribution in the static salubrinal molecule was carried out with the zindo approximation method [48] in the same software package. protein preparation the three-dimensional crystal structure of the gadd34:pp1 enzyme (4xpn) was loaded in the pdb format from the protein molecules data bank (http://www.rcsb.org). prior to docking, the molecules https://scifinder.cas.org/ http://www.rcsb.org/ admet & dmpk 7(2) (2019) 140-150 molecular docking studies of salubrinal doi: 10.5599/admet.632 143 of all the non-proteinaceous components, except for one phosphoric acid residue, having the code in co-crystallisate 403 po4, were removed. hydrogen atoms were added throughout the protein structure before molecular docking. molecular docking procedure based on the phosphoric acid residue (403 po4), a ligand group was created with the given name ligand_x-ray. based on this group, a three-dimensional model of a binding site was created, the dimensions of which were set manually and amounted along the x-axis – 40.000 å, the y-axis – 40.000 å and the z-axis – 40.000 å. docking was performed with a flexible ligand. for the scoring procedure, the semi-empirical function ascore was used created on the basis of the xscore function [49]. the resolution of the cell was set at 0.250 å. the calculation type was dock; docking engine arguslab. visualization of the results was carried out using the program pymol [50]. results and discussion the results of ligand geometry optimization according to the results of optimization of the geometry of the salubrinal molecule, the most stable is the conformation stabilized by the intramolecular hydrogen bond formed between the hydrogen atom of the thiourea fragment and the oxygen atom of the amide fragment (fig. 3a). the length of the nh...o=c bond is 1.891 å. that is, the salubrinal molecule exists as a pseudo 1,3,5oxadiazine ring with an angle h...o=c 108.33 °. according to x-ray diffraction data for 1,3,5-oxadiazine cycles, this angle is somewhat larger and lies within the range of 114.76-120.00° [51-53]. the appearance of an intramolecular hydrogen bond is obviously associated with a large difference in the static charges on the oxygen atom of the amide fragment and the hydrogen atom of the thiourea fragment. according to calculations of the electron density distribution in the static salubrinal molecule (the zindo approximation method), on the oxygen atom δ lies within -0.0409→-0.0500, in turn, on the hydrogen atom δ + is 0.0500→0.0409 (fig. 3b). the presence of an intramolecular hydrogen bond is characteristic of all the salubrinal analogues studied (see supporting information, tables s1 and s2). figure 3. a) the calculated structure of the salubrinal molecule (pm3 method), visualization in pymol; b) distribution of electron density in the static salubrinal molecule. colors: a) □ 0.0500 → 0.0409; b) ■ 0.0409 → 0.0318; c) ■ 0.0318 → 0.0227; d) ■ 0.0227 → 0.0136; e) ■ 0.0136 → 0.0045; f) ■ 0.0045 → 0.0045; g) ■ -0.0045 → -0.0136; h) ■ -0.0136 → -0.0227; i) ■ -0.0227 → -0.0318; j) ■ -0.0318 → -0.0409; k) ■ -0.0409 → -0.0500 the results of molecular docking the active center of selective dephosphatase of the eukaryotic translation initiation factor (fig. 4a) contains the phosphoric acid residue, and two mg 2+ ions (not shown in the figure). in the active site of the gadd34:pp1, it is possible to distinguish three sites, one hydrophilic located approximately in its center, and two lipophilic ones located on the periphery. therefore, the interactions of the salubrinal p.v. zadorozhnii et al. admet & dmpk 7(2) (2019) 140-150 144 molecule with the active site of the gadd34:pp1 enzyme are represented by both polar contacts (fig. 4b) and lipophilic interactions between the cinnamic acid residue, the quinoline ring and the lipophilic regions of the active site. the molecule of salubrinal effectively interacts with the gadd34:pp1 enzyme closing access to the active site. the energy of the complex gadd34:pp1-salubrinal forms 12.2489 kcal/mol. the salubrinal molecule is additionally fixed in the active center of the enzyme due to the formation of an intermolecular hydrogen bond involving the amino acid tyr 272 (fig. 4b). a hydrogen bond arises between the nitrogen atom of the pyridine type of the quinoline ring and the hydroxyl group of tyr 272 (the length of the n...ho bond is 3.432 å). the salubrinal molecule is also fixed due to the shortened intermolecular polar contacts: 1) between the oxygen atom of the amide fragment and the hydroxyl group tyr 134; 2) between the sulfur atom of the thiourea fragment and the guanidine fragment arg 221. figure 4. a) the structure of the gadd34:pp1 holoenzyme. protein phosphatase 1 (pp1) is represented in white color, and gadd34 pink light. in the active site of the holoenzyme, there is a phosphoric acid residue, depicted in the form of spheres; b) the orientation of the salubrinal molecule in the active site of the gadd34:pp1 holoenzyme according to molecular docking data. to determine the effect of the residue of cinnamic acid and quinoline cycle in the salubrinal preparation on the ability to bind to the active site of the gadd34:pp1 enzyme, we periodically replaced one of the fragments with other groups. according to the molecular docking data, six compounds out of fifty-four analyzed salubrinal analogues (see supporting information table s3 and s4) exceeded it in the stability of the complex formed with gadd34:pp1 (fig. 5). the most stable complex with gadd34:pp1 is formed by (e)-3-(thiophen-2-yl)-n-(2,2,2-trichloro-1(3-(quinolin-8-yl)thioureido)ethyl)acrylamide (s1) (fig. 6a), the energy of the complex with gadd34:pp1 is -12.8833 kcal/mol, rmsd 1.4 å. the molecule of the compound (s1) is additionally fixed in the enzyme active site due to the intermolecular hydrogen bond formed between the nitrogen atom of the thiourea fragment and the -oh group of tyr 272, the length of the hn...ho bond is 3.605 å). it is also fixed due to the formation of shortened intermolecular polar contacts: 1) between the oxygen atom of the amide fragment and the hydroxyl group of tyr 134; 2) between the sulfur atom of the thiourea fragment and the guanidine fragment arg 221. the energy of the complex n-(2,2,2-trichloro-1-(3-(2-chlorophenyl)thioureido)ethyl)cinnamamide (s2) with gadd34:pp1 forms -12.5738 kcal/mol, rmsd 2.3 å (fig. 6b). the molecule of the compound (s2) is additionally fixed in the active site of the enzyme due to the formation of two intermolecular hydrogen bonds involving amino acids his 125 and asn 124 (fig. 6b). both hydrogen bonds are formed by the oxygen atom of the amide fragment. in the first case, the hydrogen bond is with the pyrrole atom of nitrogen of the imidazole ring his 125 (the c=o...hn bond length is 2.133 å), and in the second case with the amide fragment of amino acid asn 125 (the c=o...h2nc(o) bond length is 2.884 å). the molecule of the compound (s2) is also fixed due to the shortened intermolecular polar contact between the sulfur atom of the thiourea fragment and the guanidine fragment arg 221. admet & dmpk 7(2) (2019) 140-150 molecular docking studies of salubrinal doi: 10.5599/admet.632 145 figure 5. structures of salubrinal analogues, surpassing it in the strength of the salubrinal preparation superior to the strength of the formed complex with the gadd34:pp1 holoenzyme. n-(2,2,2-trichloro-1-(3-(naphthalen-2-yl)thioureido)ethyl)cinnamamide (s3) forms the complex with the gadd34:pp1 enzyme having the energy of -12.4218 kcal/mol, rmsd 1.4 å (fig. 6c). the molecule of the compound (s3) is additionally stabilized in the enzyme active center due to the intermolecular hydrogen bond formed by the oxygen atom of the oxygen amide fragment and the hn group of tyr arg 221, the c=o...hn bond length is 2.481 å. the molecule of the compound (s3) is also stabilized due to the shortened intermolecular polar contact between the sulfur atom of the thiourea fragment and the guanidine fragment arg 221. the energy of the complex n-(2,2,2-trichloro-1-(3(naphthalene-1-yl)thioureido)ethyl)cinnamamide (s4) with gadd34:pp1 forms -12.4195 kcal/mol, rmsd 4.8 å (fig. 6d). the molecule of the compound (s4) is additionally stabilized in the enzyme active center due to the formation of intermolecular hydrogen bonds: 1) between the oxygen atom of the amide fragment and the -oh group of tyr 134, the c=o...ho bond length is 2.758 å; 2) between the nitrogen atom of the thiourea fragment and the oh group of tyr 272, the hn...ho bond length is 2.999 å. moreover, stabilization occurs due to the shortened intermolecular polar contact between the sulfur atom of the thiourea fragment and the guanidine fragment arg 221. (e)-3-(thiophen-3-yl)-n-(2,2,2-trichloro-1-(3-(quinolin-8-yl)thioureido)ethyl)acrylamide (s5) and n(2,2,2-trichloro-1-(3-(4-chlorophenyl)thioureido)ethyl)cinnamamide (s6) form complexes with the gadd34:pp1 enzyme having the energy of -12.3286 kcal/mol (rmsd 1.3 å) and -12.3140 kcal/mol (rmsd 5.4 å), respectively (fig. 6e, 6f). the compounds (s5) and (s6) do not form intermolecular hydrogen bonds in the active center of gadd34:pp1, their interaction with amino acids forming the active site is obviously hydrophobic in nature. in this case, the formation of weak shortened intermolecular polar contacts is possible, for example, between the sulfur of the thiourea fragment and the guanidine fragment arg 221. it should be noted, that the molecule of the compound (s6) in the active center of gadd34:pp1 is rotated 180° as compared to the salubrinal molecule and the remaining compounds hits. figure 7 shows that for both quinoline derivatives and cinnamic acid derivatives, the energy of the complex formed is clearly related to gadd34:pp1 from the rmsd value. the quinoline derivatives interact closely with the active site of the enzyme, the rmsd value does not exceed 3.5 å. while for the cinnamic acid derivatives, rmsd can vary from 1.5 to 12.0 å. this is due to the fact that there are p.v. zadorozhnii et al. admet & dmpk 7(2) (2019) 140-150 146 two lipophilic sites in the active center of the enzyme. the quinoline cycle, due to spatial difficulties, is clearly fixed only in one of them, and the cinnamic acid residue can interact with both. this can lead to a reversal of the inhibitor molecule located in the active center by 180° relative to the salubrinal molecule, which is observed, for example, for the compound (s6). figure 6. position of the molecules of the compounds (s1)-(s6) in the active center of the gadd34:pp1 holoenzyme. figure 7. energy dependence of the complex gadd34:pp1-inhibitor on the rmsd value. color: a) ■ quinoline derivatives; ■ cinnamic acid derivatives; ■ salubrinal, taken into account when constructing a linear regression line in both cases. based on our findings, when searching for the gadd34:pp1 inhibitors, other than cinnamic acid and quinoline derivatives, special attention should be paid to the compounds containing a naphthalene and isoquinoline ring, heterocyclic analogs of cinnamic acid, and compounds containing admet & dmpk 7(2) (2019) 140-150 molecular docking studies of salubrinal doi: 10.5599/admet.632 147 chlorine atoms in the aromatic ring. the results of our work are in good agreement with the already published experimental data on establishing the dependence of the structure-activity of salubrinal analogues [39,54]. for example, the low activity of 2-amino-pyridine derivatives, for which the ec50 lies in the range of 28-72 μm [39], as compared to the derivatives of 8-aminoquinoline (ec50 = 15-16 μm) [39,54], can be explained by the high energy of the complex that they form with gadd34:pp1. the lower energy of the gadd34:pp1-inhibitor complex can also explain the high efficiency of (e)-3-(thiophen-2-yl)acrylamide derivatives (ec50 = 4-43 μm) compared with cinnamamide derivatives (ec50 = 6-57 μm) [54]. conclusions in this paper, the search for new analogues of salubrinal has been carried out by molecular modeling. we have shown that the most stable conformation of the salubrinal molecule and its analogues contains the intramolecular hydrogen bond between the hydrogen atom of the thiourea fragment and the oxygen atom of the amide fragment. the binding site of salubrinal to the active site of the enzyme has been established. we have found the compounds, which form stronger complexes with the enzyme than salubrinal itself. the results of this work can be used to create new phosphatase inhibitors of the eukaryotic translation initiation factor gadd34:pp1. conflict of interest: the authors declare no conflicts of interests. references [1] d.s. schwarz, m.d. blower. the endoplasmic reticulum: structure, function and response to cellular signaling. cell mol. life sci. 73 (2016) 79-4. doi:10.1007/s00018-015-2052-6. [2] a.r. english, g.k. voeltz. endoplasmic reticulum structure and interconnections with other organelles. cold. spring. harb. perspect. biol. 5 (2013) a013227. doi: 10.1101/cshperspect.a013227. [3] a. blumental-perry. endoplasmic reticulum stress response, the future of cancer research and a new designated journal. endoplasmic reticulum stress in cancers 1 (2013) 1-3. doi:10.2478/ersc-2012-0001. [4] a.h. schönthal. endoplasmic reticulum stress: its role in disease and novel prospects for therapy. scientifica (2012) 26 pages. doi: 10.6064/2012/857516. [5] c. hetz. the unfolded protein response: controlling cell fate decisions under er stress and beyond. nat. rev. mol. cell biol. 13 (2012) 89-102. doi:10.1038/nrm3270. [6] s.a. oakes, f.r. papa. the role of endoplasmic reticulum stress in human pathology. annu. rev. pathol. 10 (2015) 173-194. doi:10.1146/annurev-pathol-012513-104649. [7] l. ozcan, i. tabas. role of endoplasmic reticulum stress in metabolic disease and other disorders. annu. rev. med. 63 (2012) 317-328. doi:10.1146/annurev-med-043010-144749. [8] t. hosoi, k. ozawa. endoplasmic reticulum stress in disease: mechanisms and therapeutic opportunities. clin. sci. (lond) 118 (2009) 19-29. doi:10.1042/cs20080680. [9] b.m. gardner, d. pincus, k. gotthardt, c.m. gallagher, p. walter. endoplasmic reticulum stress sensing in the unfolded protein response. cold. spring. harb. perspect. biol. 5 (2013) a013169. doi:10.1101/cshperspect.a013169. [10] b. bhandary, a. marahatta, h.-r. kim, h.-j. chae. an involvement of oxidative stress in endoplasmic reticulum stress and its associated diseases. int. j. mol. sci. 14 (2013) 434-456. doi:10.3390/ijms14010434. p.v. zadorozhnii et al. admet & dmpk 7(2) (2019) 140-150 148 [11] m.j. pagliassotti. endoplasmic reticulum stress in nonalcoholic fatty liver disease. annu. rev. nutr. 32 (2012) 17-33. doi:10.1146/annurev-nutr-071811-150644. [12] a.v. cybulsky. endoplasmic reticulum stress, the unfolded protein response and autophagy in kidney diseases. nat. rev. nephrol. 13 (2017) 681-696. doi:10.1038/nrneph.2017.129. [13] f. navid, r.a. colbert. causes and consequences of endoplasmic reticulum stress in rheumatic disease. nat. rev. rheumatol. 13 (2017) 25-40. doi:10.1038/nrrheum.2016.192. [14] h. urra, e. dufey, t. avril, e. chevet, c. hetz. endoplasmic reticulum stress and the hallmarks of cancer. trends. cancer. 2 (2016) 252-262. doi:10.1016/j.trecan.2016.03.007. [15] s.m. colgan, a.a. al-hashimi, r.c. austin. endoplasmic reticulum stress and lipid dysregulation. expert. rev. mol. med. 13 (2011) e4. doi:10.1017/s1462399410001742. [16] n.t. sprenkle, s.g. sims, c.l. sánchez, g.p. meares. endoplasmic reticulum stress and inflammation in the central nervous system. mol. neurodegener. 12 (2017) 42. doi: 10.1186/s13024-017-0183-y. [17] t. omura, m. kaneko, y. okuma, k. matsubara, y. nomura. endoplasmic reticulum stress and parkinson’s disease: the role of hrd1 in averting apoptosis in neurodegenerative disease. oxid. med. cell longev. (2013) 7 pages. doi:10.1155/2013/239854. [18] d.l. eizirik, m. miani, a.k. cardozo. signalling danger: endoplasmic reticulum stress and the unfolded protein response in pancreatic islet inflammation. diabetologia 56 (2013) 234-241. doi:10.1007/s00125-012-2762-3. [19] d.a. chistiakov, i.a. sobenin, a.n. orekhov, y.v. bobryshev. role of endoplasmic reticulum stress in atherosclerosis and diabetic macrovascular complications. biomed. res. int. (2014) 14 pages. doi:10.1155/2014/610140. [20] k. meyerovich, f. ortis, f. allagnat, a.k. cardozo. endoplasmic reticulum stress and the unfolded protein response in pancreatic islet inflammation. j. mol. endocrinol. 57 (2016) r1-r17. doi:10.1530/jme-15-0306. [21] c. xiang, y. wang, h. zhang, f. han. the role of endoplasmic reticulum stress in neurodegenerative disease. apoptosis 22 (2017) 1-26. doi:10.1007/s10495-016-1296-4 [22] s.j. marciniak, d. ron. endoplasmic reticulum stress signaling in disease. physiol. rev. 86 (2006) 1133-1149. doi:10.1152/physrev.00015.2006. [23] a. mcquiston, j.a. diehl. recent insights into perk-dependent signaling from the stressed endoplasmic reticulum. f1000res. 6 (2017) 11 pages. doi:10.12688/f1000research.12138.1. [24] d.s. schwarz, m.d. blower. the endoplasmic reticulum: structure, function and response to cellular signaling. cell mol. life sci. 73 (2016) 79-94. doi:10.1007/s00018-015-2052-6. [25] o. pluquet, a. pourtier, c. abbadie. the unfolded protein response and cellular senescence. a review in the theme: cellular mechanisms of endoplasmic reticulum stress signaling in health and disease. am. j. physiol. cell physiol. 308 (2015) c415-425. doi:10.1152/ajpcell.00334.2014. [26] h. kadowaki, h. nishitoh. signaling pathways from the endoplasmic reticulum and their roles in disease. genes (basel) 4 (2013) 306-333. doi:10.3390/genes4030306. [27] r. sano, j.c. reed. er stress-induced cell death mechanisms. biochim. biophys. acta 1833 (2013) 3460-3470. doi:10.1016/j.bbamcr.2013.06.028. [28] r. iurlaro, c. muñoz-pinedo. cell death induced by endoplasmic reticulum stress. febs j. 283 (2016) 2640-2652. doi:10.1111/febs.13598. [29] m. boyce, k.f. bryant, c. jousse, k. long, h.p. harding, d. scheuner, r.j. kaufman, d. ma, d.m. coen, d. ron, j. yuan. a selective inhibitor of eif2α dephosphorylation protects cells from er stress. science 307 (2005) 935-939. doi:10.1126/science.1101902. [30] m. matsuoka, y. komoike. experimental evidence shows salubrinal, an eif2α dephosphorylation inhibitor, reduces xenotoxicant-induced cellular damage. int. j. mol. sci. 16 (2015) 16275-16287. doi:10.3390/ijms160716275. admet & dmpk 7(2) (2019) 140-150 molecular docking studies of salubrinal doi: 10.5599/admet.632 149 [31] m. cnop, l. ladriere, p. hekerman, f. ortis, a.k. cardozo, z. dogusan, d. flamez, m. boyce, j. yuan, d.l. eizirik. selective inhibition of eukaryotic translation initiation factor 2alpha dephosphorylation potentiates fatty acid-induced endoplasmic reticulum stress and causes pancreatic beta-cell dysfunction and apoptosis. j. biol. chem. 282 (2007) 3989-3997. doi:10.1074/jbc.m607627200. [32] y. liu, j. wang, s.y. qi, l.s. ru, c. ding, h.j. wang, j.s. zhao, j.j. li, a.y. li, d.m. wang. reduced endoplasmic reticulum stress might alter the course of heart failure via caspase-12 and jnk pathways. can. j. cardiol. 30 (2014) 368-375. doi:10.1016/j.cjca.2013.11.001. [33] a.l. sokka, n. putkonen, g. mudo, e. pryazhnikov, s. reijonen, l. khiroug, n. belluardo, d. lindholm, l. korhonen. endoplasmic reticulum stress inhibition protects against excitotoxic neuronal injury in the rat brain. j. neurosci. 27 (2007) 901-908. doi:10.1523/jneurosci.428906.2007. [34] y. zhu, p. fenik, g. zhan, b. sanfillipo-cohn, n. naidoo, s.c. veasey. eif-2a protects brainstem motoneurons in a murine model of sleep apnea. j. neurosci. 28 (2008) 2168-2178. doi:10.1523/jneurosci.5232-07.2008. [35] s.k. lee, y.s. kim. phosphorylation of eif2α attenuates statin-induced apoptosis by inhibiting the stabilization and translocation of p53 to the mitochondria. int. j. oncol. 42 (2013) 810-816. doi: 10.3892/ijo.2013.1792. [36] e. katsoulieris, j.g. mabley, m. samai, i.c. green, p.k. chatterjee. α-linolenic acid protects renal cells against palmitic acid lipotoxicity via inhibition of endoplasmic reticulum stress. eur. j. pharmacol. 623 (2009) 107-112. doi:10.1016/j.ejphar.2009.09.015. [37] l. he, j. lee, j.h. jang, k. sakchaisri, j. hwang, h.j. cha-molstad, k.a. kim, i.j. ryoo, h.g. lee, s.o. kim, n.k. soung, k.s. lee, y.t. kwon, r.l. erikson, j.s. ahn, b.y kima. osteoporosis regulation by salubrinal through eif2α mediated differentiation of osteoclast and osteoblast. cell. signal. 25 (2013) 252-260. doi:10.1016/j.cellsig.2012.11.015. [38] k. hamamura, a. nishimura, t. iino, s. takigawa, a. sudo, h. yokota. chondroprotective effects of salubrinal in a mouse model of osteoarthritis. bone joint. res. 4 (2015) 84-92. doi:10.1302/2046-3758.45.2000378. [39] j. liu, k.l. he, x. li, r.j. li, c.l. liu, w. zhong, s. li. sar, cardiac myocytes protection activity and 3d-qsar studies of salubrinal and its potent derivatives. curr. med. chem. 19 (2012) 60726079. doi:10.2174/0929867311209066072. [40] d.c. young. computational drug design. john wiley & sons, inc, new jersey, usa 2009. [41] h.-d. holtje, w. sippl, d. rognan, r. folkers. molecular modeling. basic principles and applications. wiley-vch verlag gmbh & co. kgaa, weinheim, germany, 2008. [42] w. thiel. semiempirical quantum-chemical methods. wiley interdisciplinary reviews: computational molecular science 4 (2014) 145-157. doi:10.1002/wcms.1161. [43] m.a. thompson, m.c. zerner. a theoretical examination of the electronic structure and spectroscopy of the photosynthetic reaction center from rhodopseudomonas viridis. j. am. chem. soc. 113 (1991) 8210-8215. doi:10.1021/ja00022a003. [44] m.a. thompson, e.d. glendening, d. feller. the nature of k + /crown ether interactions: a hybrid quantum mechanical-molecular mechanical study. j. phys. chem. 98 (1994) 10465-10476. doi:10.1021/j100092a015. [45] m.a. thompson, g.k schenter. excited states of the bacteriochlorophyll b dimer of rhodopseudomonas viridis: a qm/mm study of the photosynthetic reaction center that includes mm polarization. j. phys. chem. 99 (1995) 6374-6386. doi:10.1021/j100017a017. [46] m.a. thompson. qm/mmpol: a consistent model for solute/solvent polarization. application to the aqueous solvation and spectroscopy of formaldehyde, acetaldehyde, and acetone. j. phys. chem. 100 (1996) 14492-14507. doi:10.1021/jp960690m. [47] m. thompson. arguslab 4.0.1. planaria software llc, seattle, wash, usa, 2004. http://www.arguslab.com. http://www.arguslab.com/ p.v. zadorozhnii et al. admet & dmpk 7(2) (2019) 140-150 150 [48] d.c. young. computational chemistry: a practical guide for applying techniques to real world problems. john wiley & sons, inc, new york, usa, 2001. doi:10.1002/0471220655. [49] r. wang, l. lai, s. wang. further development and validation of empirical scoring functions for structure-based binding affinity prediction. j. comput. aided. mol. des. 16 (2002) 11-26. [50] w.l. delano. the pymol molecular graphics system; delano scientific: palo alto, ca, 2003; http://www.pymol.org . [51] p.p. onys’ko, a.a. sinitsa, v.v. pirozhenko, a.n. chernega. synthesis of phosphorylated 1,3,5oxadiazines via n-acyltrifluoroacetimidoilphosphonates. heteroatom. chem. 13 (2002) 22-26. doi:10.1002/hc.1102. [52] k.k. kennard, k.a. byriel, t.ch. woon, d.p. fairlie. structure of a novel protonated oxadiazine: an unusual heterocycle from the cycloaddition of a ketone with nitriles. chem. commun. 15 (1996) 1731-1732. doi:10.1039/cc9960001731. [53] p.v. zadorozhnii, v.v. kiselev, i.o. pokotylo, a.v. kharchenko. a new method for the synthesis of 4h-1,3,5-oxadiazine derivatives. heterocycl. commun. 23 (2017) 369-374. doi:10.1515/hc-20170083. [54] k. long, m. boyce, h. lin, j. yuan, d. ma. structure-activity relationship studies of salubrinal lead to its active biotinylated derivative. bioorg. med. chem. lett. 15 (2005) 3849-52. doi:10.1016/j.bmcl.2005.05.120. ©2019 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://www.pymol.org/ http://creativecommons.org/licenses/by/3.0/ admet & dmpk 7(2) (2019) s1-s15 molecular docking studies of salubrinal: supporting information doi: 10.5599/admet.632 s1 supporting information for the paper: molecular docking studies of salubrinal and its analogs as inhibitors of the gadd34:pp1 enzyme table of contents table s1. the results of geometry optimization of salubrinal analogues containing cinnamic acid residue s2 table s2. the results of geometry optimization of salubrinal analogues containing quinoline ring s7 table s3. the results of molecular docking of salubrinal analogues containing cinnamic acid residue s9 table s4. the results of molecular docking of salubrinal analogues containing quinoline ring s14 p.v. zadorozhnii et al. admet & dmpk 7(2) (2019) s1-s15 s2 table s1. the results of geometry optimization of salubrinal analogues containing cinnamic acid residue entry r cas no e, kcal/mol the length of the nh...o=c bond, å angle value с=o…h, degrees time, s 1 405060-59-9 -111285.9025 1.891 108.33 347 2 863036-22-0 -111285.2829 1.880 108.88 348 3 294654-78-7 -110631.9704 2.273 105.11 344 4 324769-18-8 -110634.7191 1.863 108.63 365 5 405060-99-3 -98255.3837 1.919 109.47 192 admet & dmpk 7(2) (2019) s1-s15 molecular docking studies of salubrinal: supplementary information doi: 10.5599/admet.632 s3 6 405060-96-0 -104370.9745 1.883 108.96 207 7 863036-35-5 -129157.9187 1.943 108.52 736 8 301359-85-3 -99001.5764 1.891 108.31 232 9 301359-95-5 -102453.0936 1.863 108.48 272 10 301359-86-4 -102454.5158 1.864 108.24 240 11 301359-87-5 -102454.6059 1.868 108.08 249 12 1429483-71-5 -105904.3082 1.863 108.55 315 13 301359-93-3 -105905.3171 1.865 108.40 288 14 301359-94-4 -105905.2465 1.861 108.55 287 p.v. zadorozhnii et al. admet & dmpk 7(2) (2019) s1-s15 s4 15 294657-79-7 -105777.6924 1.858 107.98 247 16 3037775-31-7 -105778.1081 1.890 108.48 271 17 405060-94-8 -103291.5220 1.863 108.80 246 18 301359-88-6 -109214.6095 1.874 108.24 288 19 1346508-38-0 -109215.6380 1.886 108.42 298 20 303775-35-5 -115287.0750 1.858 108.47 295 21 294655-14-4 -115289.4152 1.870 108.75 318 22 294655-12-2 -118722.8463 1.858 108.12 359 23 294654-81-2 -115870.4210 1.871 108.65 291 admet & dmpk 7(2) (2019) s1-s15 molecular docking studies of salubrinal: supplementary information doi: 10.5599/admet.632 s5 24 294654-82-3 -115870.3348 1.868 108.55 301 25 301359-89-7 -105946.6228 1.864 108.36 241 26 301359-90-0 -105947.3592 1.860 108.50 224 27 301359-91-1 -105947.5373 1.859 108.49 234 28 301359-92-2 -106797.8334 1.860 108.54 237 29 301359-97-7 -112891.5613 1.862 108.63 254 30 301359-98-8 -112891.2835 1.863 108.60 277 31 301359-96-6 -138810.5173 1.859 108.45 355 32 301815-13-4 -116057.8543 1.863 108.53 375 p.v. zadorozhnii et al. admet & dmpk 7(2) (2019) s1-s15 s6 33 294653-17-1 -124258.2307 1.896 108.75 723 34 294654-77-6 -120926.7177 1.862 108.67 409 35 863036-23-1 -99655.3253 1.857 108.48 215 36 1349267-41-9 -110899.9614 1.873 108.85 294 37 1349267-41-9 -112093.4153 2.339 104.89 272 38 1346508-37-9 -97718.9301 2.507 105.22 233 39 1346508-36-8 -101035.0119 2.521 105.03 225 40 1429483-72-6 -101809.4188 2.493 105.51 275 admet & dmpk 7(2) (2019) s1-s15 molecular docking studies of salubrinal: supplementary information doi: 10.5599/admet.632 s7 table s2. the results of geometry optimization of salubrinal analogues containing quinoline ring entry r cas no e, kcal/mol the length of the nh...o=c bond, å angle value с=o…h, degrees time, seconds 1 ch3294658-37-0 -90772.1333 1.835 109.78 141 2 324769-75-5 -97666.2121 1.849 109.08 207 3 294658-28-9 -101115.4287 1.832 109.23 261 4 (ch3)3c412962-51-7 -101113.4907 1.843 109.50 247 5 305856-11-5 -108568.4619 2.370 101.87 319 6 294646-80-3 -105112.9884 2.387 101.67 291 7 324017-95-0 -108568.5668 2.460 101.65 323 8 330684-99-6 -108574.0942 2.261 105.45 330 p.v. zadorozhnii et al. admet & dmpk 7(2) (2019) s1-s15 s8 9 294658-30-3 -108574.2167 2.418 105.03 313 10 330567-60-7 -112916.8710 2.288 104.53 308 11 294658-45-0 -115333.8079 2.382 102.98 363 12 405060-98-2 -116754.1849 2.262 104.94 437 13 324017-57-4 -105701.1794 2.347 102.79 246 14 1346508-21-1 -109395.6421 1.901 108.14 336 15 1346508-22-2 -109398.3840 1.913 107.96 313 admet & dmpk 7(2) (2019) s1-s15 molecular docking studies of salubrinal: supplementary information doi: 10.5599/admet.632 s9 table s3. the results of molecular docking of salubrinal analogues containing cinnamic acid residue entry r cas no e, kcal/mol rmsd, å time, seconds 1 405060-59-9 -12.2489 68 2 863036-22-0 -11.4506 9.8 64 3 294654-78-7 -12.4195 4.8 66 4 324769-18-8 -12.4218 1.4 68 5 405060-99-3 -10.4440 8.6 222 p.v. zadorozhnii et al. admet & dmpk 7(2) (2019) s1-s15 s10 6 405060-96-0 -10.9403 8.7 712 7 863036-35-5 -11.0821 6.3 25055 8 301359-85-3 -11.0494 2.4 91 9 301359-95-5 -11.8415 2.5 78 10 301359-86-4 -11.7071 3.4 81 11 301359-87-5 -10.9713 9.5 82 12 1429483-71-5 -11.4002 3.5 70 13 301359-93-3 -11.0643 2.6 74 14 301359-94-4 -10.7865 6.86 70 admet & dmpk 7(2) (2019) s1-s15 molecular docking studies of salubrinal: supplementary information doi: 10.5599/admet.632 s11 15 294657-79-7 -11.9634 2.8 76 16 3037775-31-7 -10.8881 6.7 80 17 405060-94-8 -11.2845 2.0 75 18 301359-88-6 -11.3522 3.6 152 19 1346508-38-0 -9.8336 5.5 182 20 303775-35-5 -11.7954 3.4 176 21 294655-14-4 -10.0874 3.4 202 22 294655-12-2 -11.1046 3.7 501 23 294654-81-2 -9.9688 4.2 201 p.v. zadorozhnii et al. admet & dmpk 7(2) (2019) s1-s15 s12 24 294654-82-3 -10.2628 11.0 195 25 301359-89-7 -12.5738 2.3 86 26 301359-90-0 -10.2843 3.2 88 27 301359-91-1 -12.3140 5.4 88 28 301359-92-2 -10.4781 10.6 87 29 301359-97-7 -10.5242 9.6 82 30 301359-98-8 -11.0901 9.8 77 31 301359-96-6 -10.5026 8.7 140 32 301815-13-4 -10.9633 6.9 178 admet & dmpk 7(2) (2019) s1-s15 molecular docking studies of salubrinal: supplementary information doi: 10.5599/admet.632 s13 33 294653-17-1 -11.6777 6.8 45 34 294654-77-6 -10.2446 10.0 108 35 863036-23-1 -9.8478 11.1 87 36 1349267-41-9 -10.0146 11.9 60 37 1349267-41-9 -9.5065 6.9 60 38 1346508-37-9 -10.1080 4.1 35 39 1346508-36-8 -10.0909 6.1 35 40 1429483-72-6 -10.3209 5.5 31 p.v. zadorozhnii et al. admet & dmpk 7(2) (2019) s1-s15 s14 table s4. the results of molecular docking of salubrinal analogues containing quinoline ring entry r cas no e, kcal/mol rmsd, å time, seconds 1 ch3294658-37-0 -9.5079 3.4 30 2 324769-75-5 -9.5363 3.3 37 3 294658-28-9 -10.1110 3.4 48 4 (ch3)3c412962-51-7 -10.5443 2.4 35 5 305856-11-5 -11.0971 3.5 55 6 294646-80-3 -12.0377 1.3 38 7 324017-95-0 -12.0111 3.3 34 8 330684-99-6 -10.2993 2.6 37 admet & dmpk 7(2) (2019) s1-s15 molecular docking studies of salubrinal: supplementary information doi: 10.5599/admet.632 s15 9 294658-30-3 -10.9900 1.9 36 10 330567-60-7 -11.9627 3.4 36 11 294658-45-0 -10.4252 2.4 54 12 405060-98-2 -11.1993 2.4 35 13 324017-57-4 -9.8686 1.4 38 14 1346508-21-1 -12.8833 1.4 74 15 1346508-22-2 -12.3286 1.3 75 doi: 10.5599/admet.1.3.7 29 admet & dmpk 1(3) (2013) 29-44; doi: 10.5599/admet.1.3.7 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index mini review the roles of doxorubicin in hepatocellular carcinoma kin tam faculty of health science, university of macau, macau, china e-mail: kintam@umac.mo; tel.: +853-8397-8501; fax: +853-8397-8500 received: june 01 st ,2013; revised: july 03 rd , 2013, published: july 03 rd , 2013 abstract liver cancer is the third most common cause of death from cancer worldwide. hepatocellular carcinoma (hcc) accounts for 80 – 90 % of all primary liver cancers. the mortality of hcc is high, and the treatment options are limited. doxorubicin exhibits a broad spectrum of anti-tumour activity. when administrated via the hepatic artery, doxorubicin showed anti-tumour effects and partial response in hcc patients. in this mini review, we will first outline the treatment options for hcc. then, we will briefly discuss the pharmacological, toxicological and pharmacokinetic aspects of doxorubicin. finally, we will review the new developments and future directions of using doxorubicin in the therapy of hcc. keywords primary liver cancer; anticancer; cytotoxic drug; transarterial chemoembolization; drug eluting beads 1. introduction liver cancer is the third most common cause of death from cancer worldwide. globally, over 600,000 people develop liver cancer and the fatality rate is high [1]. the highest incidence of liver cancer occurs mostly in developing countries, such as in sub-saharan africa and eastern asia, and especially in men: the overall male: female sex ratio is about 2.4 [1]. regions with low incidence include north and south america, australia, and northern europe. hepatocellular carcinoma (hcc) accounts for 80 – 90 % of all primary liver cancers [2]. patients with cirrhosis exhibit a high probability of developing hcc [3]. many factors can lead to the development of cirrhosis. these include: infection with the hepatitis b or hepatitis c viruses, alcoholism, inherited metabolic diseases, diabetes and smoking as well as exposure to aflatoxins, which is a group of mycotoxins produced by the aspergillus fungus in foodstuffs like corn and peanuts during storage in warm, damp conditions [4-7]. doxorubicin was one of the early anthracyclines, isolated from streptomyces peucetius almost four decades ago [8]. it has been regarded as one of the effective chemotherapy drugs for cancer treatment [9]. unfortunately, the clinical uses of this drug are hampered by cumulative cardiotoxicity, with the risk becoming ever greater at cumulative doses ≥ 550 mg/m 2 [10]. in hcc chemotherapy, clinical studies showed that systemic administration of doxorubicin showed limited clinical benefits. however, when the drug was given by the hepatic artery route, tumour shrinkage and partial responses were seen in 30 – 70 % of patients, usually associated with some form of hepatic artery occluding agent [11]. http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:kintam@umac.mo tam admet & dmpk 1(3) (2013) 29-44 30 there are ongoing intensive preclinical and clinical efforts to search for better drugs and treatment regimens for hcc. while the research continues, it is valuable to pursue a deeper understanding of old drugs like doxorubicin, with the broad aims of identifying and/or devising new strategies for its better use in patients. in this mini review, we will first outline the treatment options for hcc. then, we will briefly discuss the pharmacological, toxicological and pharmacokinetic aspects of doxorubicin. finally, we will review the development and future directions of using doxorubicin in therapy for hcc. 2. treatment options for hepatocellular carcinoma hcc is often diagnosed late in its course, partly because of the lack of distinctive pathological symptoms [12]. as a result, the prognosis is poor, with a median survival of 1 – 8 months and a 5-year survival rate of only 3 % [13]. despite the progress in early diagnosis [14], the treatment options are limited. hepatic resection and liver transplantation represent the curative treatment options, offering the prospect of cure with 5-year survival rates of up to 50 [15,16] and 70 % [17,18], respectively. however, only 10 – 20 % of patients are suitable for surgical resection due to the burden of hepatic tumour, the presence of extrahepatic spread, or the extent of underlying liver disease. liver transplantation is the only surgical option for hcc patients with impaired liver function. in view of the severe shortage of liver donors, hepatic resection is still regarded as the mainstay of therapy in patients with preserved hepatic function. for non-surgical patients with early stage hcc, ablative techniques could be beneficial. commonly used liver-directed therapies include percutaneous ethanol injection and radiofrequency ablation (see also section 5.3). other investigative techniques include percutaneous acetic acid injection [19], microwave coagulation [20], cryoablation [21], and laser thermotherapy [22]. chemotherapy is often used to slow the progress of advanced liver cancer [23]. typically, this involves systemic administration (either intravenous or oral) of anti-cancer drugs such as doxorubicin, 5-fluorouracil, or cisplatin. even these drugs shrink only a small portion of tumours and the responses often do not last long. in particular, the side effects and cardiotoxicity of doxorubicin can limit its routine use at an efficacious dose. targeted chemotherapy, represented by newer drugs such as sorafenib, has been shown to slow the growth of advanced liver cancer and exhibit less severe side effects [24]. sorafenib is a potent multikinase inhibitor with antiangiogenic and antiproliferative properties. it targets the serine-threonine kinases raf-1 and b-raf, the tyrosine kinases of vascular endothelial growth factor receptors (vegfr-1/2/3) as well as the platelet-derived growth factor receptor β (pdgfr-β). cellular signalling in connection with raf-1 and vegf plays an important role in the development of hcc [25]. with this new treatment, median survival and time to radiologic progression were extended by about 3 months for patients with advanced hcc [25]. over the years, clinical studies have been carried out to broaden the treatment options. these include radioembolotherapy, adjuvant therapy, and intra-arterial approaches [26,27]. palliative therapies via transarterial chemoembolization (tace) are used for liver cancer not amenable to surgical procedures [28,29]. in the next section (2.1), we will briefly discuss the tace procedure. 2.1. transarterial chemoembolization (tace) figure 1 shows a schematic diagram of a tace procedure, which involves combined intra-arterial administration of cytotoxic anticancer drug(s) and embolization agent(s) into a liver tumour. admet & dmpk 1(3) (2013) 29-44 doxorubicin and hcc doi: 10.5599/admet.1.3.7 31 figure 1. schematic diagrams show: (a) targeted intra-arterial infusion of anti-cancer drugs leading to high local concentration at the cancerous mass, (b) the infusion of embolization particles leading to ischemia of the cancerous mass, and (c) the intra-arterial infusion procedure which involves placement of a catheter through the femoral artery into a branch of the hepatic artery (anti-cancer drugs, embolization particles, or drug eluting beads can be delivered to the cancerous mass). a multidisciplinary team consisting of a surgeon, clinical oncologist and radiologist in a hospital usually conduct the procedure. there are variants on these transarterial techniques, which are dependent on the experience of the clinical specialists and the particular treatment protocols of the hospital. the simplest form of transarterial chemoembolization (tace) involves a 2-step process, namely, (1) intra-arterial administration of chemotherapeutic agents into the tumour-feeding artery via an intra-arterially inserted catheter, and (2) selective embolization of the tumour-feeding artery. some investigators refer to this approach as conventional tace. the liver has a unique dual blood supply from both the portal vein and the hepatic artery. the normal parenchyma of the liver receives two-thirds of its necessary blood supply from the portal vein and receives the remaining one-third from the hepatic artery. hypervascular tumours such as hepatocellular carcinomas (hccs) receive their blood supply mainly from the hepatic artery [30,31]. thus, intra-arterial infusion of a cytotoxic anticancer drug (see section 2.2) with a viscous oily material, lipiodol, followed by embolization of the artery with embolic agent(s) which themselves will occlude the arterial blood supply to the tumour, cause an infarct and subsequent necrosis of the tumour. embolic agents commonly used in tace procedure include gelatin sponge, polyvinyl alcohol (pva) particles, and degradable starch microspheres. tace treatment for liver malignance can induce extensive tumour necrosis in most patients. however, the ability of tace to induce tumour necrosis does not necessarily mean longer survival rates for hcc patients. almost 70 – 80 % of patients treated with tace die due to tumour progression during follow-up because of the eventual re-growth of residual tumour cells after regaining a vascular supply or remote recurrence of tumours in other areas of the liver [32]. tace is not indicated for curative therapies. in 2002, two randomized controlled trials (rcts) from hong kong [33] and spain [34] showed the survival benefits of tace compared to the best conservative treatment. these rcts were followed by cumulative metatam admet & dmpk 1(3) (2013) 29-44 32 analyses that included all published rcts [35,36], showing that tace significantly reduced the overall 2-year mortality rate compared to control patients who received conservative or inactive treatments. 2.2. cytotoxic anticancer drugs in tace, the most common cytotoxic drugs are doxorubicin, followed by cisplatin, epi/doxorubicin, mitoxantrone and mitomycin c. the dose of the cytotoxic drug given depends on the size of the tumour, the position of the catheter, the patient’s liver function, and the response to previous courses of tace, if any. however, the criteria for determining the dosages of chemotherapeutic agents are variable and not standardized: some clinicians prefer to factor in the patient's body surface area, weight, tumour burden or bilirubin level, while others use a fixed dose. in some early studies, 40 – 100 mg of doxorubicin [37] or 10 – 70 mg of cisplatin [38] per session was used. a few randomized controlled trials (rcts) failed to show significant differences in survival between the use of doxorubicin and other drugs such as cisplatin or epirubicin and, to date, there is no evidence of the superiority of any single chemotherapeutic agent over other drugs or for mono-drug chemotherapy versus combination chemotherapy. in the united states, the most common combination of chemotherapeutic agents includes a mixture of 100 mg cisplatin, 50 mg doxorubicin and 10 mg mitomycin c. it should be noted that some of the commonly used cytotoxic drugs in tace, such as doxorubicin, could be infused directly into the systemic circulation for cancer chemotherapy but patients who receive this treatment may suffer from serious side effects (e.g. cardiac toxicity) [10]. moreover, efficacy could be compromised by suboptimal drug concentration at the liver, since the dose distribution is mainly achieved via the systemic circulation. indeed, the local drug concentration could be further limited by the fact that doxorubicin cannot diffuse very far through tissue that is necrotic, and the drug is not carried to the site when the neighbouring microvasculature is destroyed [39]. 2.3. drug eluting beads more recently, advances have been made in materials design, such that the embolic agent itself can also be a drug carrier, which appears to be a more convenient and efficient procedure. specifically, these drugloaded carriers are directly injected intra-arterially for the treatment of liver cancer in a single operation. this kind of drug carrier is referred to as a drug-eluting bead (deb) in the literature. a typical example includes the dc bead (biocompatibles, surrey, uk), which has undergone clinical investigations [40,41]. dc beads are indicated for the treatment of a variety of malignant hypervascularised tumours, including hcc. it is a pva based microspherical embolization agent, prepared from n-acrylamidoacetaldehyde derivatized polyvinyl alcohol copolymerized with 2-acrylamido-2-methylpropane sulfonate. the presence of the anionic sulphonate group enables the sequestering of positively charged drugs, such as doxorubicin, epirubicin or irinotecan, by coulomb charge interactions. the drug is slowly released from the beads at the site of administration. in particular, the interventional procedure is greatly simplified as the drug (e.g. doxorubicin) and the embolizing device (the sulfonate-modified pva bead) are delivered at the same time, which greatly facilitates the local and sustained delivery of the drug. dc beads are generally supplied in a hydrated form in a saline solution of suitable ionic strength in a vial. before performing embolization, the saline solution is removed and the drug solution at the suggested concentration is mixed with the beads and left for an appropriate time, depending on the drug, the loading solution concentration and the bead size. typically, loading will require a minimum of 20 minutes for the smallest bead size and up to 120 minutes for the largest. the beads are available in different size ranges: 100 – 300 μm, 300 – 500 μm, 500 – 700 μm, and 700 – 900 μm. different drug loadings, varying from 5 mg/ml to 45 mg/ml of hydrated beads, have been reported [42]. the recommended maximum dose of doxorubicin administered per treatment is 150 mg (75 mg/m 2 ), with a maximum recommended lifetime admet & dmpk 1(3) (2013) 29-44 doxorubicin and hcc doi: 10.5599/admet.1.3.7 33 dose of 450 mg/m 2 in light of cardiac toxicity induced by the drug when it is introduced systemically [43]. drug-loaded beads can be aspirated directly for the catheter delivery procedure. patients may receive three or four chemoembolization treatments within 6 months. it has been demonstrated that dc bead spheres can easily be loaded with doxorubicin to a recommended level of 25 mg/ml hydrated beads [42]. lammer et al. reported a clinical study comparing deb-tace (intra-arterial injection of dc beads loaded with doxorubicin) with conventional tace (intra-arterial injection of doxorubicin in lipiodol followed by particle embolization with an embolic agent, such as pva particles) for the treatment of cirrhotic patients with hcc [44]. the patient group treated with deb-tace showed higher response rates than the group treated with conventional tace. despite the hypothesis of superiority not being met, the clinical data suggested that the drug eluting bead treatment offered improved tolerance, with a significant reduction in serious liver toxicity, and a significantly lower rate of doxorubicin-related side effects [44]. in a more recent clinical study [45], malagari reported the meta-analysis on the 5-year survival of hcc patients treated with deb-tace. it has demonstrated good responses, with overall survival rates of 93.6, 62, and 22.5 % at 1, 3, and 5 years after sequential sessions of deb-tace in hcc patients not amenable to curative treatments. thus, deb-tace appears to be safe and effective in the treatment of hcc and could offer a benefit to patients. regardless of the means of implementation, tace is able to offer highly concentrated doses of chemotherapy for effective delivery to the tumour bed, while sparing the surrounding hepatic parenchyma [46]. moreover, particle embolization of the tumour-feeding arteries sustains the effects of the chemotherapy by reducing its washout, allowing prolonged contact with the tumour. as a result, selective arterial obstruction-induced ischemic tumour necrosis is achievable while minimizing any damage to normal liver tissue, as the blood supply to the normal liver tissue is still maintained by the dominant blood flow from the portal vein. this combination of embolotherapy and regional chemotherapy has synergistic, anti-tumour effects, resulting in a high objective response rate. in next two sections, we will turn our attention to the pharmacology, toxicity and pharmacokinetics of doxorubicin. 3. pharmacology and toxicity of doxorubicin doxorubicin is a cytotoxic anthracycline antibiotic isolated from cultures of streptomyces peucetius. it consists of a naphthacenequinone linked through a glycosidic bond at ring atom 7 to an amino sugar, daunosamine, with the structural formula shown in fig. 2. o o o o hooh oh oh o o oh nh2 1 2 3 4 5 6 7 8 9 101112 13 14 abcd figure 2. structure of doxorubicin. tam admet & dmpk 1(3) (2013) 29-44 34 the tetracyclic ring system is lipophilic, but the saturated end of the ring system contains several groups and the amino sugar, which forms a hydrophilic centre. the molecule itself is amphoteric, containing acidic functions in the phenolic groups and a basic function in the sugar amino group. doxorubicin is now widely used in the treatment of several of the most commonly diagnosed malignancies including leukaemia, lymphoma, bronchogenic, and traditional cell bladder carcinoma, as well as in the treatment of wilms’ tumour, neuroblastoma, sarcoma, and carcinoma of the breast, ovary, and stomach [47]. liposomal doxorubicin formulations (doxil® or lipodox®) have been developed to improve the therapeutic index of doxorubicin for use in conventional chemotherapy (systemic administration) while maintaining its anti-tumour activity. these liposomal formulations have been used for the treatment of metastatic breast/ovary cancers and of aids related kaposi’s sarcoma [48]. doxorubicin in a heat sensitive liposome formulation has been given an orphan designation (eu/3/10/833) in europe for the treatment of hepatocellular carcinoma [49]. despite extensive clinical utilization, the exact mechanisms of action of doxorubicin in cancer cells remain a matter of controversy. for instance, doxorubicin has been reported to have at least seven different means of producing cellular dysfunction or death [50]. it is generally believed that the cytotoxic effects of doxorubicin on cancer cells are related to nucleotide base intercalation. in particular, the drug molecule stabilizes the topoisomerase ii complex after it has broken the dna chain for replication, preventing the dna double helix from being resealed and thereby stopping the process of replication. crystallographic data [51] suggest that the planar tetracyclic portion of the molecule intercalates between two base pairs of the dna, while the sugar amino group sits in the minor groove and interacts with flanking base pairs immediately adjacent to the intercalation site. moreover, it has been suggested that doxorubicin activates p53-dna binding, leading to induced apoptosis [52,53]. doxorubicin-induced apoptosis may result in therapeutic effects and/or toxicities. doxorubicin may undergo a cascade of free radical processes in normal and cancer cells. typically this involves interactions of the molecule with the cell's electron transport chain, specifically the cytochrome p450 reductase (p450r) [54,55]. as shown in fig. 3, the quinone moiety readily accepts the transfer of a single electron to form the semiquinone free radical, which could be directly cytotoxic in hypoxic environments via covalent modification of cellular macromolecules. under aerobic conditions, the semiquinone radicals can shuttle electrons to molecular oxygen, giving rise to superoxide anion radicals [56-59]. these toxic intracellular radicals can be further converted to hydrogen peroxide and hydroxyl radicals, which can directly damage dna, rna, lipids, and proteins [60,61]. this oxidative stress mechanism appears to be a contributor to doxorubicin’s ability to cause cell death, as well as its dose-limiting cardiotoxicity. in the process of electron shuttling, the doxorubicin ring system is restored to its original state and can be available for additional redox cycling reactions. admet & dmpk 1(3) (2013) 29-44 doxorubicin and hcc doi: 10.5599/admet.1.3.7 35 o o 5 12 c o o 5 12 c nadph nadp+ p450r o2 o2 h2o2 oh dna & protein damage . . . 1e1/2 1/2 1e figure 3. simplified schematic of the redox cycling of doxorubicin. the major doxorubicin metabolite in humans is doxorubicinol, which is produced via the cytosolic carbonyl reductase-catalyzed reduction of the ketone at c-13 on the parent drug (fig. 2) [62,63]. reductive deglycosylation at c-7 by p450r leads to formation of the 7-deoxy-doxorubicinol metabolite. pharmacokinetic studies in humans revealed that doxorubicinol and 7-deoxy-doxorubicinol were the metabolites detected in vivo [64]. doxorubicinol also exhibits antiproliferative and antineoplastic properties. because doxorubicinol is more polar than the parent drug molecule, it is less likely to traverse through the cell membrane back to the extracellular space. hence, it is more likely to stay inside cells, increasing its cytotoxic effects [65,66]. 7-deoxy-doxorubicinol is considerably less potent in inhibiting tumour growth in a mouse lymphocytic leukaemia cell line [67]. it has been suggested that doxorubicinol could accumulate in the heart and contribute significantly to the chronic cumulative cardiotoxicity of doxorubicin therapy [68]. three types of toxicities are recognized for doxorubicin, namely acute, chronic, and local. the acute toxicities including nausea, vomiting, blood count suppression, and stomatitis are usually dose limiting. also, alopecia occurs in more than half of patients receiving standard intermittent doses [9]. congestive heart failure and dilated cardiomyopathy are chronic toxicities of doxorubicin that have received considerable attention [10] and generated much worry, especially when patients have received large cumulative dosages. pathologic findings from patients’ endomyocardial biopsies revealed loss of myofibrils, dilation of the sarcoplasmic reticulum, vascular degeneration, swelling of mitochondria, and increased numbers of lysosomes. this morphologic pattern has also been seen in rodents (e.g. mice and rats) dosed with doxorubicin, suggesting the existence of a species-independent pathway leading to these morphologic changes [69]. an early clinical study demonstrated a 10 % incidence of clinical cardiomyopathy at a cumulative doxorubicin dosage of 550 mg/m 2 , with a sharp increase in the incidence curve at progressively higher dosages [10]. however, it has been reported that the probability of clinical congestive heart failure is never zero with any dose of doxorubicin, with some patients developing this complication even after a single dose [70]. long-term follow-up is usually necessary because congestive heart failure may develop several years after therapy is completed. although the mechanisms leading to the cardiotoxicity are not tam admet & dmpk 1(3) (2013) 29-44 36 fully understood, these are thought to related to the activities of toxic metabolites, particularly doxorubicinol, and to the generation of reactive oxygen species within the cardiomyocytes (see fig. 3). the most common hypothesis of doxorubicin-induced cardiotoxicity is the oxidative stress theory, where the principal toxic mechanisms involve iron and redox reactions [71]. it has been reported that when cardiomyocytes are exposed to doxorubicin, activation of the nuclear transcription factor nfkb and apoptosis occurred after redox cycling and formation of superoxide anion radicals and hydrogen peroxides [72]. moreover, it has been shown that iron (fe(ii)) plays a key role in lipid peroxidation [73]. because cells have very little or no free iron available to catalyse free radical reactions [74], it is conceivable that doxorubicin and doxorubicinol mediate iron release from ferritin and other cellular stores [73,75]. toxic effects on a number of mitochondrial functions have been reported for metabolites of doxorubicin [76,77]. two local toxicities of doxorubicin with potentially devastating consequences are extravasation skin injury (necrosis) [78] and the radiation recall reaction [79]. the development of extravasation skin injury in connection with doxorubicin is largely related to the therapy being given by intravenous infusion. with this route of administration, tissue damage can occur in close proximity to the infusion site, probably due to the drug is being absorbed by local cells in the tissue leading to cell death. the radiation recall reaction, which remains a poorly understood phenomenon, occurred in patients who had had prior irradiation for neoplastic disease. the observed reactions were painful, erythematous, and warm dermatitides located precisely in the previous field of irradiation. 4. pharmacokinetics of doxorubicin 4.1. systemic administration doxorubicin exhibits linear pharmacokinetics after intravenous administration. it is widely distributed in the plasma and tissues, with a volume of distribution exceeding 500 l/m 2 . it undergoes triphasic plasma decay, with an initial half-life (t1⁄2α) of 4.8 min, a t1⁄2β of 2.6 hours and a terminal t1⁄2γ of about 48 hours [80]. in patients with normal liver function, hepatic clearance is high and more than 50 % of the drug is excreted in the bile within 7 days after treatment. due to the important role of hepatic metabolism and biliary excretion, patients with liver dysfunction show a rather different pattern in drug distribution. in particular, the half-life and auc were found to be increased almost by 3-fold (and clearance decreased by almost 3fold) compared with patients with normal hepatic function [81]. on the other hand, the pegylated liposomal formulations of doxorubicin showed somewhat different pharmacokinetic performance. the presence of the outer polyethylene glycol polymer layer greatly reduces uptake of the drug by the reticulo-endothelial system, resulting in prolonged circulation in the plasma, relatively little tissue distribution and reduced volume of distribution. the pharmacokinetics are reported to be biphasic, with half-lives of 1 – 3 hours and 30 – 90 hours, respectively. the auc after a dose of 50 mg/m 2 is approximately 300-fold greater than that with free drug [82]. clearance is drastically reduced (250-fold). tumour neovasculature has been reported to permit penetration of liposomal doxorubicin into tumour tissue [82]. it has been shown that in a preclinical model, the hepatic distribution of liposomal doxorubicin was restricted to the vascular space and the liposomes were unable to pass through the fenestrations in the sinusoidal endothelium of the liver [83]. this observation may offer insight into the reduced hepatic metabolism of liposomal doxorubicin, since the drug is unable to gain access to the hepatocytes. in hcc patients, it has been shown that when doxorubicin is dosed intra-arterially (without any subsequent embolization procedure), the pharmacokinetics is very similar to that of the intravenously admet & dmpk 1(3) (2013) 29-44 doxorubicin and hcc doi: 10.5599/admet.1.3.7 37 dosed drug [84]. in other words, the drug disappearance profiles obtained from intra-arterial and intravenous administrations were similar. clearly, intravenous or intra-arterial administration of doxorubicin alone may not necessarily be able to deliver a sufficient drug concentration to a hepatic tumour. 4.2. deb and locoregional administration next, we turn our attention to deb, namely the dc beads loaded with doxorubicin. the in vitro elution kinetics of doxorubicin from dc beads has been assessed using a t-cell apparatus, with a schematic depicted in fig. 4 [42]. figure 4. schematic diagram of the t-cell apparatus (citation see text). it has been shown that the doxorubicin did not release from the beads when the elution medium was pure water. when the elution medium contained ions and phosphate-buffered saline solution, reproducible and sustained release profiles were achievable. with a drug load of 25 mg/ml bead, it has been shown that the rate of drug release from 700 – 900 μm beads was slower than that from 100 – 300 μm beads, with half-lives of 1,730 and 150 hours, respectively [42]. these half-life data translate to less than 1 % and 20 % of drug being released over 24 hours from the total available drug loaded on the 700 – 900 μm beads and 100 – 300 μm beads, respectively. these in vitro data nicely demonstrated a sustained and controlled delivery of the doxorubicin without any burst effect, which may otherwise lead to acute toxicity in vivo. in a subsequent study [85], it was shown that the loading and release of doxorubicin followed a dose-response curve. using 500 – 700 μm beads, it was found that the half-live increased from 381 hours to 3658 hours as the concentration of the doxorubicin load increased from 6.25 mg/ml to 37.5 mg/ml. for a fixed drug load of 37.5 mg/ml, the half-life was only weakly dependent on bead size, with a minimum of 1505 hours for the 100 – 300 μm beads. interestingly, the in vitro elution data of doxorubicin has been shown to correlate well with the areas under the curve (i.e. the doxorubicin concentration in human plasma measured as a function of time) of 15 patients treated with dc beads loaded with doxorubicin in the precision clinical study [85]. this covered all doses used in the study: 6.25 mg/ml, 12.5 mg/ml, 18.75 mg/ml, 25 mg/ml, and 37.5 mg/ml in 24 hours. as far as we are aware, this is the first report available in open literature to tam admet & dmpk 1(3) (2013) 29-44 38 demonstrate a good in-vivo in-vitro correlation (ivivc) for dc beads. undoubtedly, this kind of good ivivc should help predict the local dose that should be administered before systemic effects (e.g. acute toxicity) occur. however, the amount of drug released in situ into tumour tissue in humans is difficult to determine and model by this kind of in vitro measure, because of practical and ethical issues. studies on pre-clinical species using a vx-2 tumour rabbit model have been performed, and confirmed a high level of doxorubicin in the tumour over the entire period of study (14 days) associated with widespread necrosis of the tumour tissue [86]. the pharmacokinetics of doxorubicin in hcc patients have been determined as part of two phase i/ii studies to evaluate the use of dc beads in the deb-tace procedure. one of the studies compared the pharmacokinetic profiles of doxorubicin after injection of dc beads against those seen after intra-arterial injection [87]. the average auc after dc bead injection was found to be about 100-fold lower than intraarterial injection, despite a higher drug load being used in the dc beads. the other study compared the pharmacokinetic profile of deb-tace against conventional tace [88]. the peak drug concentration (cmax) was reached within 5 min after injection, but it was about 10-fold higher in conventional tace patients. the auc was also about 3-fold higher in the conventional tace patients. it is evident that deb-tace is effective in limiting the systemic exposure to doxorubicin, resulting in reduced systemic toxicity. the hepatic distribution of doxorubicin in hcc patients has recently been reported [89]. in this reported study, 6 hcc patients were given the same deb-tace procedure followed by liver transplantation after embolization from 8 hours to 36 days. from liver explant samples, it was determined that the mean doxorubicin concentration surrounding the debs peaked at 5 μm at 8 hours, and then diminished to 2.10 μm at 2 weeks, and 0.65 μm at 32 – 36 days. most of the dc beads (100 – 300 μm) were confined within 10 mm of the tumour boundary. necrosis of tissues surrounding debs has been observed in 9 – 36 day liver explants, suggesting deb may lead to antiproliferative and cytotoxic effects on the tumour cells surrounding the beads. 5. the new developments and future directions hcc mortality is generally high and the available treatment options are limited. chemotherapy using anti-tumour cytotoxic agents, such as doxorubicin, appears to be one of the non-curative therapeutic approaches to help hcc patients. systemic chemotherapies have been evaluated for many years [11]. the response rate is poor (below 30 %) and the survival advantage is minimal. the uses of adjuvant or neoadjuvant systemic chemotherapy for liver resection also failed to show clinical benefits [11]. on the other hand, locoregional administration of doxorubicin via deb-tace produced relatively high response rates and survival advantage [45]. it would be particularly valuable to explore the best use of this approach in combination with other treatment options in the therapy of hcc. 5.1. standardization of tace procedure one of the current issues with tace concerns the technique of chemoembolization. despite deb offering some advantage, conventional tace is still being used widely. while recommendations have been published by a panel of experts on the use of deb-tace [90], there is still no consensus on how to perform conventional tace. for instance, there are variants on how the anticancer drug is formulated in lipiodol. there is also a wide range of choices for embolizing agents [91,92]. consequently, implementation could vary depending on the preference of the investigators. the evolution from conventional tace towards debtace will hopefully help to harmonize the implementation procedure. admet & dmpk 1(3) (2013) 29-44 doxorubicin and hcc doi: 10.5599/admet.1.3.7 39 5.2. combination with targeted chemotherapy targeted chemotherapy such as sorafenib represents a different mechanism of action against hcc by acting on raf-1/vegf cell signalling pathways. used as a single agent, sorafenib can produce tumour shrinkage and offer a survival advantage of about 3 months in advanced hcc patients [25]. given the different modes of actions between the kinase inhibitor (sorafenib) and the cytotoxic agent (doxorubicin), it is perfectly reasonable to evaluate the combination of these two classes of drug and assess the clinical benefits that may bring to hcc patients. excitingly, a recent phase ii study revealed the combined use of sorafenib and doxorubicin (administrated as conventional tace) in patients from the asia-pacific region with intermediate hcc [93]. the disease control rate and overall response rate were reported to be 91.2 % and 52.4 %, respectively. thus, concurrent sorafenib and doxorubicin-tace therapy appears to be effective. similar clinical studies are currently being conducted globally [94,95]. it is hoped that the findings from these studies corroborate the positive results obtained from the asia-pacific region [93]. other molecularly targeted agents, such as bevacizumab/erlotinib, are currently under clinical investigation for therapy for hcc [96]. bevacizumab is a humanized monoclonal antibody that inhibits vascular endothelial growth factor a, while erlotinib is a reversible tyrosine kinase inhibitor that acts on the epidermal growth factor receptor [97]. as in the case of sorafenib, these drug actions are based on different mechanisms as compared with doxorubicin. it would be of great interest to evaluate their synergistic effects with doxorubicin and see if these might result in tumour shrinkage and enhanced survival. 5.3. combination with other treatment options doxorubicin-tace could offer the possibility of extending hcc patients’ lives while waiting for curative treatment such as a liver transplant. it has been pointed out that the patient drop-off (e.g. death) risk while awaiting liver transplantation for hcc is 22 % [98]. doxorubicin-tace treatment significantly reduced the drop-off risk, resulting in longer survival [98]. the combination of tace with radiofrequency ablation (rfa) is a promising approach to treating hcc. rfa represents an attractive alternative to liver resection for patients with early-stage hcc. rfa uses radiofrequency energy-generated heat to destroy the tumour. the technique works well with small hcc (< 3 cm), with an impressive 4-year survival rate of 66 – 82 % as reported in japan [99]. as the size of tumour grows beyond 5 cm, the efficacy of rfa is generally reduced [100], probably due to increased blood flow leading to heat loss and/or incomplete ablation. it has been reported that by performing hepatic intraarterial embolization ahead of rfa treatment (tace-rfa), the heat sink effect mediated by hepatic arterial flow should be minimized [101]. giving tace before rfa administration offers additional locoregional chemotherapy, which should result in better response and improved survival. clinical studies have been conducted to evaluate the tace-rfa procedure in hcc patients. cheng et al. were among the first to report a clinical trial involving tace-rfa in patients with hcc tumour sizes from 3 to 7.5 cm [102]. it has been shown that patients treated with tace-rfa had better overall survival than those treated with tace alone or rfa alone. a more recent study also revealed a similar picture [103]. with a median follow-up of 36 months, the 1-, 3-, and 4-year overall survival for the tace-rfa treated patients and the rfa treated patients was 92.6 % and 66.6 %, 61.8 % and 85.3 %, and 59 % and 45.0 %, respectively. no significant difference in toxicity profiles was observed between the tace-rfa treated patients and the rfa treated patients. tam admet & dmpk 1(3) (2013) 29-44 40 6. conclusions hcc mortality is high, and the treatment options are limited. surgical procedures such as liver resection or liver transplantation are still regarded as the only curative treatments currently available. in general, less than 20 % of hcc patient are deemed suitable candidates to perform curative resection. tace is emerging as a promising palliative treatment for hcc patients that are not suitable for surgical procedures. the recently developed deb offers some operational convenience in performing the tace procedure. doxorubicin is one of the most commonly used cytotoxic drugs in tace. despite doxorubicin exhibiting a broad spectrum of anti-tumour activity, systemic administration can often lead to fatal cardiotoxicity and adverse side effects. as a liver directed, locoregional procedure, tace considerably reduces the systemic exposure to doxorubicin, resulting in improved toxicity profiles, and has been shown to be useful in preventing tumour progression and in prolonging patients’ lives. tace in combination with targeted chemotherapy, such as sorafenib, has already demonstrated clinical advantages in the asia-pacific region. moreover, concurrent use of tace and rfa has shown promising survival benefits in several clinical studies. it is envisaged that doxorubicin administration via tace could be one more useful tool in combatting hcc as it offers the exciting possibility of enhancing the currently available therapies for hcc patients. obviously, more clinical studies are required to fully evaluate the clinical benefits of these combined approaches. references [1] globocan 2008. international agency for cancer research liver cancer fact sheet, http://globocan.iarc.fr/ (accessed 16 may, 2013) [2] hepatocellular carcinoma: esmo clinical practice guidelines for diagnosis, treatment and follow-up, european society for medical oncology (2012). [3] a. forner, j.m. llovet, j. bruix, lancet 31 (2012) 1245-1255. [4] f.x. bosch, j. ribes, m. díaz, r. cléries, gastroenterology 127 (2004) s5-s16. [5] e.e. calle, c. rodriguez, k. walker-thurmond k, m.j. thun, the new england journal of medicine 348 (2003) 1625-1638. [6] j. bruix, j.m. llovet, lancet 373 (2009) 614-616. [7] s.c. chuang, c. la vecchia, p. boffetta , cancer letters 286 (2009) 9-14. [8] k. yoshikawa, h. kitaoka, the japanese journal of surgery 1 (1971) 256-262. [9] r.b. weiss, seminars in oncology 19 (1992) 670–686. [10] d.d. von hoff dd, m.w. layard, p. basa, h.l. davis, a.l. von hoff, m. rozencweig, f.m. muggia, annals of internal medicine 91 (1979) 710–717. [11] b.i. carr, s. nagalla, in hepatocellular carcinoma: diagnosis and treatment, brain carr (ed)., humana press, new york, usa, 2010, p. 527. [12] m.c. kew, h.a. dos santos, s. sherlock, british medical journal 4 (1971) 408–411. [13] a. forner, a.j. hessheimer, m. isabel real m, j. bruix, critical reviews in oncology/hematology 60 (2006) 89–98. [14] b.h. zhang, b.h., yang, z.y. tang, journal of cancer research and clinical oncology 130 (2004) 417– 422. [15] t. shuto, k. hirohashi, s. kubo, h. tanaka, t. tsukamoto, t. yamamoto, t. ikebe, h. kinoshita, surgery today 28 (1998) 1124–1129 http://globocan.iarc.fr/ admet & dmpk 1(3) (2013) 29-44 doxorubicin and hcc doi: 10.5599/admet.1.3.7 41 [16] x.d. zhou, z.y. tang, b.h. yang bh, z.y. lin, z.c. ma, s.l. ye, z.q. wu, j. fan, l.x. qin, b.h. zheng, cancer 91:1479–1486 [17] v. mazzaferro, e. regalia, r. doci, s. andreola, a. pulvirenti, f. bozzetti, f. montalto, m. ammatuna, a. morabito, l. gennari, the new england journal of medicine 334 (1996) 693–699 [18] j.m. llovet, j. bruix, j. fuster, j.c. garcia-valdecasas, l. grande, a. franca, c. brú, m. navasa, m.c. ayuso, m. solé, m.i. real, r. vilana, a. rimola, j.visa, j. rodés, hepatology 27 (1998) 1572–1577 [19] k. ohnishi, h. yoshioka, s. ito, k. fujiwara, hepatology 27 (1998) 67-72. [20] t. shibata, y. iimuro, y. yamamoto, y. maetani, f. ametani, k. itoh, j. konishi, radiology 223 (2002) 331-337. [21] r. adam, e.j. hagopian, m. linhares, j. krissat, e. savier, d. azoulay, f. kunstlinger, d. castaing, h. bismuth, archives of surgery 137 (2002) 1332-1340. [22] c.m. pacella, g. bizzarri, f. magnolfi, p. cecconi, b. caspani, v. anelli, a. bianchini, d. valle, s. pacella, g. manenti, z. rossi, radiology 221 (2001) 712-720. [23] p. simmonds, british medical journal 321 (2000) 531-535. [24] r.k. kelley, a.p. venook, journal of clinical oncology, 26 (2008) 5845-5848. [25] j.m. llovet, s. ricci, v. mazzaferro, p. hilgard, e. gane, j.f. blanc, a.c. de oliveira, a. santoro, j.l. raoul, a. forner, m. schwartz, c. porta, s. zeuzem, l. bolondi, t.f. greten, p.r. galle, j.f. seitz, i. borbath, d. häussinger, t. giannaris, m. shan, m. moscovici, d. voliotis, j. bruix, the new england journal of medicine 359 (2008) 378-390. [26] a. kennedy, s. nag, r. salem, r. murthy, a.j. mcewan, c. nutting, a. benson, j. espat, j.i. bilbao, r.a. sharma, j.p. thomas, d. coldwell, international journal of radiation oncology, biology and physics 68 (2007) 13–23. [27] d.l. bartlett, j. berlin, g.y. lauwers, w.a. messersmith, n.j. petrelli, a.p. venook, annals of surgical oncology 13 (2006) 1284-1292. [28] j.w. chung, hepatogastroenterology 45 suppl 3 (1998), 1236-1241. [29] e. liapi, k.h. lee, c.c. georgiades, k. hong, j.f.h. geschwind, techniques in vascular & interventional radiology 10 (2007) 261-269. [30] c. breedis, g. young, the american journal of pathology 30 (1954) 969-977. [31] t. nakashima, m. kojiro, seminars in liver disease 6 (1986) 259-266. [32] s.w. shin, korean journal of radiology 10 (2009) 425-434. [33] c.m. lo, h. ngan, w.k. tso, c.l. liu, c.m. lam, r.t. poon, s.t. fan, j. wong, hepatology 35 (2002) 1164-1171. [34] j.m. llovet, m.i. real, x. montana, r. planas, s. coll, j. aponte, c. ayuso, m. sala, j. muchart, r. sola, j. rodes, j. bruix, lancet 359 (2002) 1734-1739. [35] c. camma, f. schepis, a. orlando, m. albanese, l. shahied, f. trevisani, p. andreone, a. craxi, m. cottone, radiology 224 (2002) 47-54. [36] j.m. llovet, j. bruix, hepatology 37 (2003) 429-442. [37] h. nakamura, t. hashimoto, h. oi, s. sawada, radiology 170 (1989) 783-786. [38] a.o. chan, m.f. yuen, c.k. hui, w.k. tso, c.l. lai, cancer 94 (2002) 1747-1752. [39] j. gao, f. qian, a. szymanski-exner, n. stowe, j. haaga, journal of biomedical materials research 62 (2002) 308-314. [40] k. malagari, expert review of anticancer therapy 8 (2008) 1643-1650. [41] m. sadick, s. haas, m. loehr, m. elshwi, m.v. singer, j. brade, s.o. schoenberg, s.j. diehl, onkologie 33 (2010) 31-37. [42] a.l. lewis, m.w. gonzalez, a.w. lloyd, b. hall, y.q. tang, s.l. willis, s.w. leppard, l.c. wolfenden, r.r. palmer, p.w. stratford, journal of vascular and interventional radiology 17 (2006) 335-342. [43] j. kettenbach, a. stadler, i.v. katzler, r. schernthaner, m. blum, j. lammer, t. rand, cardiovascular and interventional radiology 31 (2008) 468-476. tam admet & dmpk 1(3) (2013) 29-44 42 [44] j. lammer, k. malagari, t. vogl, f. pilleul, a. denys, a. watkinson, m. pitton, g. sergent, t. pfammatter, s. terraz, y. benhamou, y. avajon, t. gruenberger, m. pomoni, h. langenberger, m. schuchmann, j. dumortier, c. mueller, p. chevallier, r. lencioni, p.v. investigators, cardiovascular and interventional radiology 33 (2010) 41-52. [45] k. malagari, m. pomoni, h. moschouris, e. bouma, j. koskinas, a. stefaniotou, a. marinis, a. kelekis, e. alexopoulou, a. chatziioannou, k. chatzimichael, s. dourakis, n. kelekis, s. rizos, d. kelekis, cardiovascular and interventional radiology 35 (2012) 1119–1128. [46] r. yamada, m. sato, m. kawabata, h. nakatsuka, k. nakamura, s. takashima, radiology 148 (1983) 397-401. [47] doxorubicin official fda information, http://www.drugs.com/pro/doxorubicin.html (accessed 28 may, 2013). [48] lipodox official fda information, http://www.drugs.com/pro/lipodox.html (accessed 28 may, 2013). [49] pso 096-10 doxorubicin hydrochloride (in heat sensitive liposomes), http://www.ema.europa.eu/docs/en_gb/document_library/orphan_designation/2011/03/wc500 102686.pdf (accessed 28 may, 2013). [50] g. powis, pharmacology & therapeutics 35 (1987) 57–162. [51] c.a. frederick, l.d. williams, g. ughetto, g.a. van der marel, j.h. van boom, a rich, a.h. wang, biochemistry 29 (1990) 2538–2549. [52] f. penault-llorca, a. cayre, f. bouchet mishellany, s. amat, v. feillel, g. le bouedec, j.p. ferriere, m. de latour, p. chollet, international journal of oncology 22 (2003) 1319–1325; [53] v. stearns, b. singh, t. tsangaris, j.g. crawford, a. novielli, m.j. ellis, c. isaacs, m. pennanen, c. tibery, a. farhad, r. slack, d.f. hayes, clinical cancer research 9 (2003) 124–133. [54] n.r. bachur, s.l. gordon, m.v. gee, molecular pharmacology 13 (1977) 901–910. [55] n.r. bachur, m. gee, journal of pharmacology and experimental therapeutics 197 (1976) 681–686. [56] j. cummings, l. anderson, n willmott, j.f. smyth, european journal of cancer 27 (1991) 532–535. [57] j.h. doroshow, cancer research 43 (1983) 4543–4551. [58] d.w. reif, free radical biology and medicine 12 (1992) 417–427. [59] d.a. gewirtz, biochemical pharmacology 57 (1999) 727–741 [60] p.m. kanter, h.s. schwartz, leukemia research 3 (1979) 277–283. [61] c.e. myers, w.p. mcguire, r.h. liss, i. ifrim, k. grotzinger, r.c. young, science 197 (1977) 165–167. [62] d.h. huffman, n.r. bachur, cancer research 32 (1972) 600–605. [63] n.k. ahmed, r.l. felsted, n.r bachur, journal of pharmacology and experimental therapeutics 209 (1979) 12–19. [64] k. mross, p. maessen, w.j. van der vijgh, h. gall, e. boven, h.m. pinedo, journal of clinical oncology 6 (1988) 517-526. [65] w.j. pigram, w. fuller, l.d. hamilton, nature 235 (1972) 17–19. [66] e.n. dessypris, d.e. brenner, m.r. baer, k.r. hande, cancer research 48 (1988) 503-506. [67] d.w. yesair, p.s. thayer, s. mcnitt, k. teague, european journal of cancer 16 (1980) 901-907. [68] r.d. olson, p.s. mushlin, d.e. brenner, s. fleischer, b.j. cusack, b.k. chang, r.j. boucek, proceedings of the national academy of sciences 85 (1988) 3585-3589. [69] p.k. singal, t. li, d. kumar, i danelisen, n. iliskovic, molecular and cellular biochemistry 207 (2000) 77–85. [70] m.r. bristow, p.d. thompson, r.p. martin, j.m. mason, m.e. billingham, d.c. harrison, american journal of medicine 65 (1978) 823–832. [71] k.j. schimmel, d.j. richel, r.b. van den brink, h.j. guchelaar, cancer treatment review 30 (2004) 181–191. http://www.drugs.com/pro/doxorubicin.html http://www.drugs.com/pro/lipodox.html http://www.ema.europa.eu/docs/en_gb/document_library/orphan_designation/2011/03/wc500102686.pdf http://www.ema.europa.eu/docs/en_gb/document_library/orphan_designation/2011/03/wc500102686.pdf admet & dmpk 1(3) (2013) 29-44 doxorubicin and hcc doi: 10.5599/admet.1.3.7 43 [72] s. wang, s. kotamraju, e. konorev, s. kalivendi, j. joseph, b. kalyanaraman, biochemical journal 367 (2002) 729–740. [73] g. minotti, chemical research in toxicology 6 (1993) 134–146. [74] g. cairo, s. recalcati, a. pietrangelo, g. minotti, free radical biology and medicine 32 (2002) 1237– 1243. [75] g. minotti, a.f. cavaliere, a. mordente, m. rossi, r. schiavello, r. zamparelli, g.f. possati, journal of clinical investigation 95 (1995) 1595–1605. [76] p.m. sokolove, m.b. kester, j. haynes, biochemical pharmacology, 46 (1993) 691-697. [77] p.m. sokolove, r.g. shinaberry, biochemical pharmacology, 37 (1988) 803-812. [78] t.v. goolsby, f.a. lombardo, seminars in oncology 33 (2006) 139-143. [79] h.a. burris, j. hurtig, the oncologist 15 (2010) 1227-1237. [80] c.m. camaggi, r. comparsi, e. strocchi, f. testoni, b. angelelli, f. pannuti, cancer chemotherapy and pharmacology 21 (1988) 221-228. [81] s.c. piscitelli, k.a. rodvold, d.a. rushing, d.a. tewksbury, clinical pharmacology & therapeutics 53 (1993) 555-561. [82] a. gabizon, h. shmeeda, y. barenholz, clinical pharmacokinetics 42 (2003) 419-436. [83] s.n. hilmer, v.c. cogger, m. muller, d.g. le couteur, drug metabolism and disposition 32 (2004) 794–799. [84] p.j. johnson, c. kalayci, n. dobbs, n. raby, e.m. metivier, l. summers, p. harper, r. williams, journal of hepatology 13 (1991) 120-127. [85] m.v. gonzalez, y. tang, g.j. phillips, a.w. lloyd, b. hall, p.w. stratford, a.l. lewis, journal of materials science materials in medicine 19 (2008) 767-775. [86] k. hong, a. khwaja, e. liapi, m.s. torbenson, c.s. georgiades, j.f. geschwind, clinical cancer research 12 (2006) 2563-2567. [87] r.t.p. poon, w.k. tso, r.w.c. pang, k.k.c. ng, r. woo, k.s. tai, s.t. fan, clinical gastroenterology and hepatology 5 (2007) 1100–1108. [88] m. varela, m.i. real, m. burrel, a. forner, m. sala, m. brunet, c. ayuso, l. castells, x. montana, j.m. llovet, j. bruix, journal of hepatology 46 (2007) 474–481. [89] j. namur, s.j. citron, m.t. sellers, m.h. dupuis, m. wassef, m. manfait, a. laurent, journal of hepatology 55 (2011) 1332–1338. [90] r. lencioni, t. de baere, m. burrel, j.g. caridi, j. lammer, k. malagari, r.c.g. martin, e. o’grady, m.i. real, t.j. vogl, a. watkinson, j.f.h. geschwind, cardiovascular and interventional radiology 35 (2012) 980–985. [91] a. fohlen, j.p. pelage, in transcatheter embolization and therapy, d. o. kessel, c.e. ray, ed(s)., springer, london, uk, 2010, p. 29. [92] m. brinckman, in transcatheter embolization and therapy, d. o. kessel, c.e. ray, ed(s)., springer, london, uk, 2010, p. 41. [93] y.h. chung, g. han, j.h. yoon, j. yang, j. wang, g.l. shao, b.i. kim, t.y. lee, y. chao, international journal of cancer 132 (2013) 2448–2458. [94] chemoembolization with or without sorafenib tosylate in treating patients with liver cancer that cannot be removed by surgery, http://clinicaltrials.gov/ct2/show/nct01004978 (accessed 30 may, 2013). [95] transarterial chemoembolization using doxorubicin beads with or without sorafenib tosylate in treating patients with liver cancer that cannot be removed by surgery, http://clinicaltrials.gov/ct2/show/nct01324076 (accessed 30 may, 2013). [96] erlotinib plus bevacizumab in hepatocellular carcinoma (hcc) as second-line therapy, http://clinicaltrials.gov/ct2/show/nct01180959 (accessed 30 may, 2013). http://clinicaltrials.gov/ct2/show/nct01004978 http://clinicaltrials.gov/ct2/show/nct01324076 http://clinicaltrials.gov/ct2/show/nct01180959 tam admet & dmpk 1(3) (2013) 29-44 44 [97] s.c. sweetman, (ed)., martindale: the complete drug reference (35 th edition ed.), pharmaceutical press, london, uk, electronic version, 2006. [98] c. frangakis, j.f. geschwind, d. kim, y. chen, a. koteish, k. hong, e. liapi, c.s. georgiades, cardiovascular and interventional radiology 34 (2011) 1254–1261. [99] y. minami, m. kudo, international journal of hepatology article id 104685 (2011) 9 pages. [100] j. huang, l. yan, z. cheng, h. wu, l. du, j. wang, y. xu, y. zeng, annals of surgery 252 (2010) 903– 912. [101] t. yamasaki, f. kurokawa, h. shirahashi, n. kusano, k. hironaka, k. okita, cancer 95 (2002) 23532360. [102] b.q. cheng, c.q. jia, c.t. liu, w. fan, q.l. wang, z.l. zhang, c.h. yi, journal of the american medical association 299 (2008) 1669–1677. [103] z.w. peng, y.j. zhang, m.s. chen, l. xu, h.h. liang, x.j. lin, r.p. guo, y.q. zhang, w.y. lau, journal of clinical oncology 31 (2013) 426-432. ©2013 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ global testing of a consensus solubility assessment to enhance robustness of the who biopharmaceutical classification system doi: https://dx.doi.org/10.5599/admet.850 23 admet & dmpk 9(1) (2021) 23-39; doi: https://dx.doi.org/10.5599/admet.850 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific article global testing of a consensus solubility assessment to enhance robustness of the who biopharmaceutical classification system valeria gigante 1, *, giovanni m. pauletti 2 , sabine kopp 1 , minghze xu 3 , isabel gonzalez-alvarez 10 , virginia merino 4 , michelle p. mcintosh 5 , anita wessels 6 , beomjin lee 7 , kênnia rocha rezende 8 , gerhard k.e. scriba 9 , gaurav p. s.jadaun 11 , marival bermejo 10 1 norms and standards for pharmaceuticals, world health organization, geneva, switzerland; 2 department of pharmaceutical and administrative sciences, st. louis college of pharmacy, st. louis, missouri, usa 3 institute for chemical drug control, china national institutes for food and drug control, beijing, china 4 department of pharmaceutics and pharmaceutical technology and parasitology, university of valencia, valencia, spain 5 drug delivery disposition and dynamics, monash institute of pharmaceutical sciences, monash university, parkville, australia 6 north_west university, school of pharmacy, potchefstroom, south africa 7 college of pharmacy and institute of pharmaceutical science and technology, ajou university, suwon, republic of korea 8 faculty of pharmacy, federal university of goiás, brazil 9 department of pharmaceutical chemistry, friedrich schiller-university, jena, germany 11 indian pharmacopoeia commission, ministry of health & family welfare, govt. of india, ghaziabad, india 10 department of engineering: pharmacy section, universidad miguel hernández de elche, alicante, spain *corresponding author: e-mail: gigantev@who.int; tel.: +41(0)22-791 4589 received: may 11, 2020; revised: september 29, 2020; published: october 07, 2020 abstract the who biopharmaceutical classification system (bcs) is a practical tool to identify active pharmaceutical ingredients (apis) that scientifically qualify for a waiver of in vivo bioequivalence studies. the focus of this study was to engage a global network of laboratories to experimentally quantify the ph-dependent solubility of the highest therapeutic dose of 16 apis using a harmonized protocol. intra-laboratory variability was ≤5 %, and no apparent association of inter-laboratory variability with api solubility was discovered. final classification “low solubility” vs “high solubility” was consistent among laboratories. in comparison to the literature-based provisional 2006 who bcs classification, three compounds were reclassified from “high” to “low-solubility”. to estimate the consequences of these experimental solubility results on bcs classification, dose-adjusted in silico predictions of the fraction absorbed in humans were performed using gastroplus®. further expansion of these experimental efforts to qualified apis from the who essential medicines list is anticipated to empower regulatory authorities across the globe to issue scientifically-supported guidance regarding the necessity of performing in vivo bioequivalence studies. ultimately, this will improve access to affordable generic products, which is a critical prerequisite to reach universal health coverage. ©2021 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords biowaiver; multisource products; essential medicines; permeability; regulatory guidance https://dx.doi.org/10.5599/admet.850 https://dx.doi.org/10.5599/admet.850 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:gigantev@who.int http://creativecommons.org/licenses/by/4.0/ valeria gigante et al. admet & dmpk 9(1) (2021) 23-39 24 introduction in a world where one third of the population lacks access to essential medicines [1] and medicines represent the largest family expenditure after food, generic competition and differential pricing are major drivers to sustain a healthy population, especially in low-income countries [2]. therefore, the world health organization (who) has identified access to essential medicines as one of its key priorities to meet the millennium sustainable development goals and reach universal health coverage. in general, regulatory agencies carry the public health responsibility to carefully balance access to and safety of medicines within their respective authorities. successful accomplishment of this objective requires risk management strategies that are based on scientifically valid data. the biopharmaceutical classification system (bcs) introduced in 1995 [3] recognizes the fundamental contribution of aqueous solubility of an active pharmaceutical ingredient (api) within the physiologically relevant gastrointestinal ph range and its mucosal permeability in governing oral absorption in vivo. consequently, various regulatory agencies around the world have adopted guidances specifying the use of the bcs as a scientific framework to grant a waiver of mandatory human in vivo bioequivalence studies for immediate-release solid oral dosage. hence, qualifications defined for a biowaiver directly impact access to and affordability of a finished pharmaceutical product while assuring the quality of the medicine. the who recognizes the possibility of waiving in vivo bioequivalence studies (i.e., biowaiver) as an effective regulatory strategy to accelerate development of generic products for eligible apis in order to improve access and save lives. therefore, as part of its 2006 guidance on waiving of bioequivalence requirements for immediate-release oral solid dosage forms, the who published a provisional classification of apis based on the 14 th who essential medicines list (eml) with regards to their eligibility for a biowaiver. as this initial list was largely based on literature data, which can vary widely due to different experimental conditions employed, the who expert committee on specifications for pharmaceutical preparations recommended in 2016 that the who revises this list to the current eml [4] using experimentally determined laboratory data according to a globally harmonized protocol in order to facilitate effective generic development, including regulatory review by national regulatory authorities. the primary objective of the who biowaiver guidance is to establish a scientific framework whereby the risk of bioinequivalence is reduced to an acceptable level. leveraging an api’s main physicochemical characteristics that influence intestinal absorption, its clinical safety and efficacy profile, and comparative consideration of the finished pharmaceutical product, it is accepted that an api’s solubility, permeability, and comparative dissolution data of the finished pharmaceutical product represent adequate scientific evidence to enable an informed decision whether or not a biowaiver could be safely recommended for a specific api. eligible substances are those assigned to class i (highly soluble and highly permeable) or class iii (highly soluble and low permeable) according to the bsc framework [5].the bcs is widely accepted by various regulatory authority, including ema and fda, and highlights the pivotal role of solubility underpinning the biowaiver approach. according to the who, an api is considered highly soluble when the highest single therapeutic dose (e.g., the maximum dose administered orally at once) as specified in the approved product label of the originator is soluble in 250 ml or less of aqueous media over the ph range of 1.2–6.8 [6]. comparative dissolution data are expected to be generated by pharmaceutical manufactures for eligible finished pharmaceutical product containing the apis classified within this research project. regulators will assess the similarity in the dissolution profiles generated in order to evaluate a potential waiver from in vivo be studies. following a recommendation from the who expert committee on specifications for pharmaceutical admet & dmpk 9(1) (2021) 23-39 enhancing robustness of the who biopharmaceutical classification system doi: https://dx.doi.org/10.5599/admet.850 25 preparations, the who norms and standards for pharmaceuticals team initiated in 2017 the who biowaiver project aimed at revising its 2006 bcs [6]. to support revision of this provisional classification list, which was largely based on secondary aqueous solubility references and a diverse array of permeability data published in the literature, a global multi-center research team was assembled and tasked to study solubility profiles of selected apis using a harmonized experimental protocol. the project began in 2018 as a small-scale pilot (cycle i) focusing on the development of a universal, scientifically sound methodology to consistently assess an api’s solubility within the desired gastrointestinal ph range of 1.2–6.8. preliminary validation of the refined equilibrium solubility protocol was performed by three independent laboratories and details of the procedure were published as the “who protocol to conduct equilibrium solubility experiments for the purpose of biopharmaceutics classification system-based classification of active pharmaceutical ingredients for biowaiver”, hereinafter referred as the protocol. to prioritize apis for the second phase of the project (cycle ii), which was intended to compare inter-laboratory variability on api solubility assessment and classification using the globally harmonized protocol, the following selection criteria were defined [7]: i) api included in medicines listed on the eml; ii) intended to be formulated as immediate-release, solid oral dosage forms; iii) relevant to therapeutic areas of major public interest; and iv) exhibiting well-characterized physicochemical properties. pharmaceuticals of primary interest to the who are those associated with who programs focusing on non-communicable diseases, neglected tropical diseases, and maternal-child health, including tuberculosis, malaria, and hiv/aids. candidate apis for cycle ii of the who biowaiver project were identified through a prioritization exercise in collaboration with the who prequalification of medicines assessment team. final selection of the api training set for this study was performed after public consultation according to stakeholders’ needs and priorities. table 1 summarizes the main properties of the selected apis that were distributed to 10 different international laboratories located in who member states representing the african region, region of the americas, south-east asian region, european region, and western pacific region, respectively. experimental materials all drug substances used in this study were donated by various suppliers and compliant with compendial specifications as outlined in the international pharmacopoeia [8] and other national or regional pharmacopoeias. each api was provided to multiple laboratories according to a risk-based distribution plan that assured experimental solubility assessment by three independent facilities. controlled substances, which are difficult to ship across borders, were measured in two facilities. table 1 summarizes api-specific information regarding therapeutic indication, selected physicochemical properties, and the provisional bcs classification assigned by the who in 2006 based on solubility and permeability data obtained from the primary literature. all chemicals selected for experimental solubility assessments were of high purity or analytical grade and were used as received. solubility assessment solubility of the highest therapeutic dose according to the approved label or summary of product characteristics of the originator was determined for each api at 37 ± 1 °c in aqueous buffer solutions at ph 1.2, 4.5, and 6.8 using a globally harmonized protocol [9]. across the various laboratories engaged in this study, equilibrium solubility was quantified by the “shake flask” method supported by a validated, stabilityindicating analytical methodology such as high-performance liquid chromatography. https://dx.doi.org/10.5599/admet.850 valeria gigante et al. admet & dmpk 9(1) (2021) 23-39 26 table 1. active pharmaceutical ingredients prioritized for who bcs classification in cycle ii api (solid form listed in eml) therapeutic area indication a highest therapeutic dose [mg] b highest strength [mg] 2006 who provisiona l bcs c logp d mw e [g/mol] number of suppliers aciclovir antiviral medicines antiherpes medicines 800 200 iii -1.40 225.2 1 amoxicillin (trihydrate) antibacterials antibiotics 3000 500 i -1.94 419.5 2 azithromycin (dihydrate) antibacterials antibiotics 2000 500 iv/ii 3.39 785.0 1 cefixime (trihydrate) antibacterials antibiotics 400 400 iv -1.23 507.5 1 codeine (phosphate hemihydrate) (2:2:1) medicines for pain and palliative care opioid analgesics 60 30 iii 1.61 812.7 1 daclatasvir (dihydrochloride) antiviral medicines medicines for hepatitis c 60 60 not classified 4.06 811.8 2 darunavir (ethanolate) antiviral medicines antiretrovirals (hiv) 800 800 not classified 1.93 593.7 2 dolutegravir antiviral medicines antiretrovirals (hiv) 50 50 not classified 1.20 419.4 1 efavirenz antiviral medicines antiretrovirals (hiv) 600 600 iv/ii 4.19 315.7 3 ethionamide antibacterials antituberculosi s medicines 500–1000 250 iii/i 1.26 166.3 1 furosemide cardiovascular medicines medicines used in heart failure 80 40 iv/ii 2.07 330.7 2 primaquine (phosphate) (1:2) antiprotozoal medicines antimalarial medicines (curative treatment of p.vivax and p.ovale infections) 15 15 i 2.86 455.3 2 pyrimethamine antiprotozoal medicines antimalarial medicines 75 25 iv/iii 2.53 248.7 3 raltegravir (potassium) antiviral medicines antiretrovirals (hiv in pregnant women and in second-line) 400 400 not classified 1.97 482.5 2 rifampicin antibacterials antituberculosi s/antileprosy medicines 750 300 ii 2.53 822.9 3 tenofovir disoproxil fumarate (1:1) antiviral medicines antiretrovirals (hiv) 300 300 not classified -1.34 635.5 1 a according to the 21 st who model list of essential medicines (2019) b according to summary of product characteristics from who-pq or national/regional regulatory authority c proposal to waive in vivo bioequivalence requirements for who model list of essential medicines immediate-release, solid oral dosage forms. in: who expert committee on specifications for pharmaceutical preparations: fortieth report. geneva: world health organization; 2006: annex 8 (who technical report series, no. 937; https://www.who.int/medicines/areas/quality_safety/quality_assurance/proposalwaivevivobioequivalencerequirementsmodellistessentialmedicinesimmediatereleasesolidoraldosageformstrs937annex8.pdf?ua=1, accessed 4 may 2020). updated requirement in: multisource (generic) pharmaceutical products: guidelines on registration requirements to establish interchangeability. in: who expert committee on specifications for pharmaceutical preparations: fifty-first report. geneva: world health organization; 2017: annex 6 (who technical report series, no. 1003; https://www.who.int/medicines/areas/quality_safety/quality_assurance/trs1003_annex6.pdf?ua=1, accessed 2 october 2020). d in silico log p values predicted using admet predictor® 9.5 (simulations plus, inc., lancaster, ca) e the molecular weight corresponds to the solid form of the api used for the solubility experiments as indicated in column 1 of this table. https://www.who.int/medicines/areas/quality_safety/quality_assurance/proposalwaivevivobioequivalencerequirementsmodellistessentialmedicinesimmediatereleasesolidoraldosageformstrs937annex8.pdf?ua=1 https://www.who.int/medicines/areas/quality_safety/quality_assurance/proposalwaivevivobioequivalencerequirementsmodellistessentialmedicinesimmediatereleasesolidoraldosageformstrs937annex8.pdf?ua=1 https://www.who.int/medicines/areas/quality_safety/quality_assurance/trs1003_annex6.pdf?ua=1 admet & dmpk 9(1) (2021) 23-39 enhancing robustness of the who biopharmaceutical classification system doi: https://dx.doi.org/10.5599/admet.850 27 the time required to reach equilibrium was experimentally defined during a preliminary experiment performed at the ph value where lowest api solubility was predicted. subsequently, pivotal experiments were initiated in triplicate at ph 1.2, 4.5, and 6.8, respectively, using approximately a 10-50 % excess amount of the api estimated to meet the who “high solubility” class boundary (see below). undissolved solid was separated by either filtration or centrifugation prior to api quantitation employing a validated, stability-indicating analytical methodology. individual solubility was reported in mg/ml. as specified in the who technical report series no. 1003 annex 7 entitled “multisource (generic) pharmaceutical products: guidelines on registration requirements to establish interchangeability” [5], an api is considered “highly soluble” when the dose/solubility volume (dsv) representing the volume of liquid necessary to completely dissolve the highest single therapeutic dose of the api as recommended by the approved label or summary of product characteristics of the originator product is consistently smaller than 250 ml over the entire phrange of 1.2 – 6.8. permeability data to estimate the consequences of the experimental solubility results generated by this global laboratory consortium using the harmonized who protocol on bcs classification of each api, it was necessary to obtain api-specific permeability data indicative of the fraction absorbed in humans (fa). to limit variability and selection bias associated with different experimental conditions used to define api permeability as shown by larregieu and benet for caco-2 cell permeability assessment [10], a consistent computational simulation approach was implemented using gastroplus®9.7 (simulations plus, inc., lancaster, ca) under a free academic license. gastroplus®9.7 is a bottom-up, whole-body physiologically-based pharmacokinetic (pbpk) model that uses the advanced compartmental absorption transit (acat) mechanistic absorption model to mimic the human intestinal absorption of oral formulations from the gi tract [11]. simulations were conducted for the highest therapeutic dose as well as the highest strength of each api reported in table 1 using the human fasted physiological model for an individual with an average bodyweight of 70 kg without enterohepatic recirculation following administration of an immediate release tablet associated with a default gastric transit time of 15 min. gastrointestinal absorption rate in the acat model is estimated over a physiological ph range of 1.3 – 6.8 employing an absorption scale factor logd model, which considers regional permeability according to predicted physiological changes in transand paracellular transport, anatomical changes in surface area due to the presence of villi and microvilli, and gastrointestinal ph gradient. a summary of general input parameters used for this pbpk modeling approach is provided in table 2. api-specific input parameters were generated in silico using the admet predictor © v9.5.0.0 (simulations plus, inc., lancaster, ca). to compare in silico predicted fa values for gastrointestinal absorption in humans to clinically determined oral bioavailability data indicative of the fraction absorbed in vivo, a literature search was performed in the pubmed database (www.ncbi.nlm.nih.gov) using the international nonproprietary name of the selected api in combination with one or more of the following search terms: absorption, bcs, bioavailability, fraction absorbed, gastrointestinal, isotope, mass balance, oral, and pharmacokinetics. moreover, publicly available documents associated with regulatory approvals of originator and generic drug products for the respective api were consulted to collect pertinent in vivo pharmacokinetic information from human trials. statistical analysis all experiments were carried out at least in three independent laboratories, and results are reported as mean ± standard deviation (sd). coefficients of variation (%cv) were calculated as 100×sd/mean. statistically significant differences (p<0.05) between groups were evaluated using unpaired student’s t-test https://dx.doi.org/10.5599/admet.850 http://www.ncbi.nlm.nih.gov/ valeria gigante et al. admet & dmpk 9(1) (2021) 23-39 28 or one-way analysis of variance (anova) where appropriate (microsoft excel 2010, microsoft, redmond, wa). table 2. general formulation and pbpk simulation parameters used as model input variables in gastroplus® input parameter value dosage form dose dose volume [ml] mean precipitation time [s] drug particle density [g/ml] mean particle radius [µm] number of bins shape factor dissolution model bile salt effect pbpk model small intestine transit time [h] asf model paracellular model biliary clearance fraction simulation mode simulation time [h] immediate release tablet highest therapeutic dose/formulation strength 250 900 1.2 25 1 1 johnson off human fasted (male, 30 years, 70 kg) 3.2 optlogd model sa/v 6.1 zhimin 0 single 24 results and discussion complexities of solubility measurements a distinct feature of the who bcs when compared to other bcs classification systems as defined by various regional regulatory authorities is that the who classification centers on the highest single oral therapeutic dose of each apis (as recommended by the approved label/summary of product characteristics of the originator) to estimate the dsv. the rationale for this definition is that bcs classification is mimicking more closely real life conditions and corresponds to the “worst case scenario” when patients are taking the therapeutic oral dose prescribed. moreover, a focus on the highest single oral dose provides the necessary flexibility to consider different strengths available on the market when regulatory decisions regarding a biowaiver are discussed. overall, these measures are consistent with the who’s priority on facilitating access to affordable medicines globally. in general, samples are expected to reach equilibrium within 24 hours. however, the time required to arrive at equilibrium between solid and dissolved api may be influenced by various experimental factors ranging from intrinsic api characteristics, excess of solid included, to the type and speed of agitation method used [12]. this may leads to long dissolution times until saturation is reached that carry the risk of introducing structural changes in the drug substance when in contact with buffers solutions, including transition into a different polymorphic state, salt or solvate. in solution, salt forms are more likely to recrystallize resulting in highly stable molecular configurations (i.e., free forms) with significantly altered dissolution properties. for ionized molecules, in addition, the propensity for self-aggregation and/or complexation is increased while in contact with buffer components. the conventional shake-flask method represents the “gold standard” for solubility assessment as it is simple and easy to perform and only requires inexpensive instrumentation. when executed according to a well-designed protocol, it generally results in high quality data, with standard deviation ≤5 %. however, the shake-flask method is also timeand labor-intensive and may consume significant amounts of drug substance in order to measure thermodynamic solubility of a highly soluble api in the presence of its solid admet & dmpk 9(1) (2021) 23-39 enhancing robustness of the who biopharmaceutical classification system doi: https://dx.doi.org/10.5599/admet.850 29 form. to minimize such constraints, the harmonized protocol defines the highest api concentration for pivotal solubility determination in this study to a concentration representing at least twice the highest therapeutic dose over 250 ml of buffer. consequently, the experimentally determined concentration value does not represent a true “equilibrium” solubility but rather assures that dsv<1, which is compliant with the definition of “high solubility”. the results of this global solubility assessment for each api are summarized in table 3. it is important to re-emphasize that individual laboratories did not attempt to quantify the thermodynamic equilibrium solubility (i.e., saturation concentration in the presence of excess solid). therefore, for apis with high solubility, recorded values were measured after complete dissolution of the solid. consequently, greater variability in absolute api concentrations recorded in table 3 under specified ph conditions is the result of different api amounts used for the solubility experiments rather than a demonstration of highly variable techniques implemented by the different laboratories. among the 10 apis that were previously included in the 2006 who bcs, which was largely based on secondary aqueous solubility references, experimental values reported from three independent laboratories unambiguously justified re-classification of aciclovir, amoxicillin, and ethionamide from a “high solubility” (hs) to a “low solubility” (ls) drug. the change in classification of amoxicillin and aciclovir was related to the solubility assessment performed at the highest therapeutic dose (i.e., 3,000 mg for amoxicillin and 800 mg for aciclovir) instead of the highest strength as both drugs exhibit dose-dependent solubility profiles [13,14]. table 3. experimentally determined ph-dependent api solubility using a globally harmonized protocol api ph cs mean mg/ml a sd cv% dsv solubility class aciclovir 1.2 4.14 4.71 113.87 193.44 ls 4.5 1.25 1.11 88.96 642.57 6.8 1.27 1.10 86.45 628.93 amoxicillin (trihydrate) 1.2 9.68 13.84 142.96 206.64 ls 4.5 2.78 1.64 58.96 718.17 6.8 2.84 2.51 88.38 703.69 azithromycin (dihydrate) 1.2 9.51 na na 210.39 ls 4.5 6.32 4.83 76.42 316.53 6.8 6.43 0.08 1.25 310.90 cefixime (trihydrate) 1.2 0.72 0.76 106.07 555.56 ls 4.5 4.44 4.31 97.12 90.15 6.8 7.45 4.43 59.46 53.66 codeine (phosphate hemihydrate) 1.2 60.68 42.18 69.52 0.99 hs 4.5 66.86 51.12 76.46 0.90 6.8 70.38 54.91 78.02 0.85 daclatasvir (dihydrochloride) 1.2 88.17 125.07 141.86 0.68 ls 4.5 1.78 3.48 195.74 33.72 6.8 0.10 0.14 142.09 624.09 https://dx.doi.org/10.5599/admet.850 valeria gigante et al. admet & dmpk 9(1) (2021) 23-39 30 table 3. continued… api ph cs mean mg/ml a sd cv% dsv solubility class darunavir (ethanolate) 1.2 0.57 0.32 55.56 1402.40 ls 4.5 0.19 0.04 23.30 4202.05 6.8 0.17 0.03 20.14 4776.12 dolutegravir 1.2 0.04 0.03 68.10 1252.09 ls 4.5 0.04 0.02 61.99 1273.34 6.8 0.05 0.02 44.62 943.99 efavirenz 1.2 0.18 0.20 109.87 3331.65 ls 4.5 0.20 0.24 115.60 2947.17 6.8 0.21 0.24 114.51 2857.39 ethionamide 1.2 10.23 12.82 125.24 24.43 ls 4.5 0.96 0.16 16.71 259.97 6.8 0.51 0.08 15.29 488.28 furosemide 1.2 0.01 0.02 137.59 5818.26 ls 4.5 0.14 0.05 36.66 558.95 6.8 3.74 2.04 54.39 21.37 primaquine (phosphate) 1.2 50.51 62.22 123.19 0.30 hs 4.5 51.38 63.55 123.67 0.29 6.8 48.16 59.06 122.63 0.31 pyrimethamine 1.2 1.70 0.85 50.17 44.24 ls 4.5 4.91 3.98 80.98 15.27 6.8 0.27 0.35 128.09 277.50 raltegravir (potassium) 1.2 0.16 0.18 110.85 2513.14 ls 4.5 0.17 0.15 91.02 2363.23 6.8 0.48 0.56 115.85 826.45 rifampicin 1.2 41.36 39.37 95.18 14.51 ls 4.5 0.70 0.39 55.54 851.79 6.8 1.15 0.61 52.93 521.92 tenofovirdisoproxil fumarate 1.2 9.24 12.79 138.44 32.48 hs 4.5 3.94 3.46 87.78 76.14 6.8 4.35 4.53 104.12 68.97 a experimental data represent mean values of three individual experiments inter-laboratory reproducibility to assess the robustness of the globally harmonized solubility protocol, each api was analyzed by at least three different laboratories. the only exception was codeine, which was analyzed by two different laboratories due to its controlled substance status that limited distribution across borders. figure 1 illustrates the variability in ph-dependent solubility data reported for darunavir, as a representative example, which was analyzed by six different laboratories. laboratory-specific interpretation of experimental parameters that were deliberately kept flexible in the globally harmonized protocol to enable wide spread adoption across various resource settings [15] introduces unavoidable variability in absolute solubility data, ranging between 20-50 %. however, it is important to note that intra-laboratory variability was always ≤5 % and the final solubility classification of darunavir as a “low solubility” drug was consistent among all laboratories. similar consistencies were observed for all other apis (see supplemental information). admet & dmpk 9(1) (2021) 23-39 enhancing robustness of the who biopharmaceutical classification system doi: https://dx.doi.org/10.5599/admet.850 31 figure 1. experimental ph-dependent solubility values reported for darunavir by six different laboratories. dotted line represents the class boundary between “high solubility” and “low solubility” for the highest therapeutic dose of darunavir (i.e., 800 mg) avdeef and co-workers examined over 800 publications describing equilibrium solubility assessment of sparingly soluble ionizable drug-like molecules by the shake-flask and related methods [16]. this comprehensive analysis identified many factors that affect the quality of the experimental outcome. among all studies, the reported standard deviation among laboratories varied from 0.17 to 0.58 [17] log units with reported values up to 1.48 log units for some sparingly soluble compounds. it is predicted that a standardized protocol can control some of this inter-laboratory variability. however, compound-specific factors such as formation of polymorphs, hydrates, solvates, amorphous solids, and the impact of stereoisomers remain sources for experimental variability despite the best intention for consistent experimental execution of a protocol. in contrast, the overall standard deviation among all reported solubility values from the different laboratories in this study was 0.83 log units, but 50 % of the api data fell within a narrower range of 0.7 log units, which is consistent with previous literature comparisons. figure 2 attempts to visualize this inter-laboratory variability by representing the distribution of solubility data as coefficient of variation (cv%) in comparison to the average observe value (i.e., 100 %). the absence of a defined association between inter-laboratory variability and respective api solubility supports the robustness of the globally harmonized protocol. most importantly, the overall solubility classification based on individual solubility data was always consistent among all laboratories indicating that key determinants of the experimental approach in this globally harmonized protocol were adequately controlled to allow consistent solubility assessment for the purpose of bcs classification. to explore the dependence of these solubility assessments on the api source material used for the experimental approach, various laboratories received the same api that was provided by two or three different suppliers. figure 3 summarizes the results for rifampicin, which was obtained from three different manufacturers and analyzed by two different laboratories. the consistent solubility data obtained for all three batches suggest that the physical properties of the api provided by the different suppliers were highly similar. this conclusion is consistent with the requirement of this study that all apis must comply with compendial specifications that include physical properties. it is noted that absolute solubility values obtained for the different batches of rifampicin at ph 1.2 exhibit greater variability than measured at ph 4.5 and ph 6.8, respectively. the reason for the higher variability observed at ph 1.2 was probably due to chemical instability that was reported by most laboratories analyzing this api. https://dx.doi.org/10.5599/admet.850 valeria gigante et al. admet & dmpk 9(1) (2021) 23-39 32 figure 2. observed inter-laboratory variability versus average solubility values across experiments figure 3. comparative ph-dependent solubility profile of rifampicin obtained from three different suppliers and analyzed by two different laboratories intestinal permeability harmonization across established biowaiver guidance documents consistently defines the permeability class boundary for “high permeability” as the fraction absorbed (fa) ≥85 %. however, different regulatory authorities identify either the highest therapeutic dose or the highest marketed formulation strength as the relevant dose that should be considered for the purpose of a biowaiver pathway [18]. to explore the sensitivity of the bcs classification with respect to the permeability of the oral dose administered, computational simulations using gastroplus® were performed estimating fa for the highest therapeutic dose and highest formulation strength of each api. the results from this in silico modeling approach that utilized a consistent set of model input variables as outlined in table 2 are summarized in figure 4. the almost perfect overlay between the two symbols representing the different dose levels for most of the selected apis in this training set suggests that biopharmaceutical properties remain quite constant, irrespectively whether the highest therapeutic dose or the highest formulation strength is modeled. in the context of the bcs, this result implies that admet & dmpk 9(1) (2021) 23-39 enhancing robustness of the who biopharmaceutical classification system doi: https://dx.doi.org/10.5599/admet.850 33 gastrointestinal solubility and permeability remain unchanged within the two different dose levels compared. mechanistically, this dose proportionality also supports the hypothesis that gastrointestinal absorption for most of these apis is predominantly driven by passive diffusion. interestingly, the visual representation of these computational simulations identifies two notable exceptions. the predictions for amoxicillin estimate a fa = 43.9 % for the 500 mg strength, whereas the predicted fa for a therapeutic oral dose of 3,000 mg is only 40.4 %. although this small reduction in the predicted fa for 3,000 mg dose may need confirmation using a broader set of simulation parameters, it seems consistent with results from human pharmacokinetic studies that identified nonlinear absorption kinetics for this penicillin-type antibiotic after oral administration [19]. mechanistic evaluation of intestinal permeation pathways contributing to oral absorption of amoxicillin revealed a saturable, capacity-limited component mediated by the intestinal oligopeptide transporter, pept1 [20]. the second api in this training set predicted to exhibit dose-dependent absorption kinetics is rifampicin. for the 300 mg strength, the estimated fa = 65.6 % but dramatically decreases to fa= 37.9 for a therapeutic oral dose of 750 mg. mariappan and singh assessed the mechanisms underlying variable gastrointestinal absorption of rifampicin using a rat model [21]. the results from this preclinical study provided evidence for limited transfer of this antituberculosis agent across the mucosal barrier in a regio-specific manner (jejunum > ileum) due to significant affinity for intestinal efflux systems such as p-glycoprotein. moreover, it was discovered that transepithelial flux of rifampicin in the duodenum increased up to limit of 300 µg/ml and becoming constant thereafter, which is indicative of a saturable absorption component such as an influx transporter. since the predicted fa for either the highest therapeutic dose or highest formulation strength of amoxicillin and rifampicin were <85 %, both apis were qualified as “low permeability” drugs when using the computationally predicted permeability parameter for the purpose of bcs classification. figure 4. intestinal permeability classification. fraction absorbed in humans after oral administration of an immediate-release tablet was predicted for the highest daily dose (red squares) and the highest formulation strength (green circles) using gastroplus®. these simulated predictions were compared with results from clinical oral bioavailability studies performed in humans (see table 3 for references). the “high permeability” class boundary at the fraction absorbed (fa) = 85 % is represents by the dashed line the results from these exploratory in silico permeability predictions using default simulation algorithms of gastroplus © identified seven apis from this training set, namely codeine, daclatasvir, dolutegravir, ethionamide, primaquine, pyrimethamine, and raltegravir as “high permeability” drugs according to the 0 2 0 4 0 6 0 8 0 1 0 0 1 2 0 0 2 0 4 0 6 0 8 0 1 0 0 1 2 0 o r a l b i o a v a i l a b i l i t y i n h u m a n s m e a s u r e d [ % ] f r a c t i o n a b s o r b e d i n h u m a n s p r e d i c t e d [ % ] oral bioavailability in humans measured, % f ra ct io n a b so rb e d in h u m a n s p re d ic te d , % https://dx.doi.org/10.5599/admet.850 valeria gigante et al. admet & dmpk 9(1) (2021) 23-39 34 who bcs threshold of fa ≥85 %. in comparison to clinically determined human bioavailability (f) data for these “high permeability” drugs identified from the primary literature, the variability in the fractions of the drug reaching the systemic circulation in vivo (30.0 % ≤ f ≤ 96.0 %) is significantly greater than the computationally predicted fraction absorbed using gastroplus © (95.8 % ≤ fa ≤ 99.8 %). this is an important reminder of the critical role of model input data that are used to perform pbpk simulations. gastroplus© was used in this study as an example to illustrate the potential of a bottom-up pbpk simulation tool when exploring bcs classification of apis for which no clinical bioavailability data are available. nevertheless, default simulation parameters as outlined in table 2 may not adequately represent the specific conditions encountered in vivo. this may include important volumetric terms of gastrointestinal segments that can affect drug solubility and/or precipitation rate of a free salt form and a solvate. alternatively, physiological contribution of inand efflux transporters that have the ability to significantly modulate the extent and rate at which a drug substance becomes available in the systemic circulation after oral administration (= bioavailability) may not be adequately represented by default model parameters. similarly, gene polymorphism responsible for different expression levels of major metabolism enzymes (see supplemental information) that can lead to variable pre-systemic elimination of the api at the brush-border membrane may not be appropriately considered when using default model parameters. therefore, various research groups have already initiated comprehensive research studies to identify critical model input parameters that must be defined in order to enhance the in vivo relationship and bioequivalence prediction using pbpk absorption models [22,23]. api assignment to who biopharmaceutical classification system the main objective of this study was to validate provisional api classification performed earlier using literature-based data (i.e., 2006 who bcs) with an updated experimental solubility assessment that was based on a globally harmonized protocol. table 4 summarizes the results from this study and compares provisional api assignments to the who bcs. completion of equilibrium solubility experiments by a consortium of 10 international laboratories located in different who regions provided the scientific basis to change the solubility designation from “high solubility” to “low solubility” for aciclovir, amoxicillin, and ethionamide. this result may have been the consequence of updated highest therapeutic dose values when compared to the provisional 2006 who bcs classification. in addition, the results from this study provided qualified solubility data for five apis (i.e., daclatasvir, darunavir, dolutegravir, raltegravir, and tenofovir disoproxil fumarate) that were not considered during the previous who bcs assignment in 2006. for the remaining six apis, experimental solubility assessment confirmed the earlier solubility classification based on literature data. however, due the experimental difficulty to directly assess fa in humans, the research community is faced with widespread uncertainty regarding the most appropriate methodology to consistently categorize apis as “high permeability” or “low permeability” drugs [39]. provisional classification of apis according to the who bcs in 2006 [6] was exclusively based on literature data. ambiguity regarding the permeability classification was noted in 28.2 % of all 131 apis examined. admet & dmpk 9(1) (2021) 23-39 enhancing robustness of the who biopharmaceutical classification system doi: https://dx.doi.org/10.5599/admet.850 35 table 4. comparison of provisional api assignment to who bcs api highest therapeutic dose [mg] a provisional 2006 who bcs b experimentally assigned solubility class gastroplus © predicted fraction absorbed in humans oral bioavailability in humans [%] c provisional 2020 who bcs aciclovir 800 iii ls lp *20.0 [24] iv amoxicillin trihydrate 3000 i ls lp 45.0 [19] iv azithromycin dihydrate 2000 iv/ii ls lp *37.0 [25] iv cefixime trihydrate 400 iv ls lp 40.2 [26] iv codeine phosphate semihydrate 60 iii hs hp 54.8 [27] i/iii daclatasvir dihydrochloride 60 not classified ls hp 67.0 [28] ii/iv darunavir ethanolate 800 not classified ls lp *37.0 [29] iv dolutegravir 50 not classified ls hp 31.0 [30] ii/iv efavirenz 600 ii/iv ls lp 19.5 [31] iv ethionamide 500–1000 iii/i ls hp *83.3 [32] ii/iv furosemide 80 iv/ii ls lp *42.8 [33] iv primaquine phosphate 15 i hs hp 96.0 [34] i pyrimethamine 75 iv/iii ls hp *50.0 [35] ii/iv raltegravir potassium 400 not classified ls hp 30.0 [36] ii/iv rifampicin 750 ii ls lp *93 [37] ii/iv tenofovir disoproxil fumarate 300 not classified hs lp 25 [38] iii a according to summary of product characteristics from who-pq or national/regional regulatory authority. b proposal to waive in vivo bioequivalence requirements for who model list of essential medicines immediate-release, solid oral dosage forms. in: who expert committee on specifications for pharmaceutical preparations: fortieth report. geneva: world health organization; 2006: annex 8 (who technical report series, no. 937; https://www.who.int/medicines/areas/quality_safety/quality_assurance/proposalwaivevivobioequivalencerequirementsmodellistessentialmedici nesimmediatereleasesolidoraldosageformstrs937annex8.pdf?ua=1, accessed 4 may 2020). updated requirement in: multisource (generic) pharmaceutical products: guidelines on registration requirements to establish interchangeability. in: who expert committee on specifications for pharmaceutical preparations: fifty-first report. geneva: world health organization; 2017: annex 6 (who technical report series, no. 1003; https://www.who.int/medicines/areas/quality_safety/quality_assurance/trs1003_annex6.pdf?ua=1, accessed 2 oct 2020). c oral bioavailability data were obtained from the primary literature. ls = low solubility; hs = high solubility; lp = low permeability; hp = high permeability * clinical bioavailability data using an oral dose different than the highest therapeutic dose listed in table 1 to explore alternative options that may overcome limitations of literature-based permeability classification, various research groups have reported the use of computational pbpk absorption models for bioequivalence predictions [22,23]. in this study, the gastroplus © platform was used as an example to illustrate the potential of a bottom-up pbpk simulation tool when exploring bcs classification of apis, particularly in situations when clinical bioavailability data are not available. the results from this in silico exercise are consistent with previous literature-based permeability assignments for aciclovir, furosemide, and primaquine. furthermore, adoption of this a priori modeling approach using default gastroplus © input parameters (see table 2), combined with clinical bioavailability data generated for different oral doses, https://dx.doi.org/10.5599/admet.850 https://www.who.int/medicines/areas/quality_safety/quality_assurance/proposalwaivevivobioequivalencerequirementsmodellistessentialmedicinesimmediatereleasesolidoraldosageformstrs937annex8.pdf?ua=1 https://www.who.int/medicines/areas/quality_safety/quality_assurance/proposalwaivevivobioequivalencerequirementsmodellistessentialmedicinesimmediatereleasesolidoraldosageformstrs937annex8.pdf?ua=1 https://www.who.int/medicines/areas/quality_safety/quality_assurance/trs1003_annex6.pdf?ua=1 valeria gigante et al. admet & dmpk 9(1) (2021) 23-39 36 made it feasible to define with high confidence permeability for darunavir and tenofovir disoproxil, which were not considered during the previous who bcs assignment in 2006. similarly, using the same consensus approach that compares the in silico predicted fa in humans and clinically determined bioavailability data for the same api but measured at different doses, the ambiguous bcs classification of azithromycin, efavirenz, and furosemide were clarified. however, the results summarized in table 4 clearly underline significant discrepancies between pbpk-predicted human absorption data and clinically measured bioavailability values, specifically for codeine, dolutegravir, ethionamide, pyrimethamine, raltegravir, and rifampicin. for some of those apis, such as rifampicin, which has been described to experience regiospecific absorption (jejunum > ileum) due to significant affinity for intestinal efflux systems such as pglycoprotein [21], the literature may provide clues for a reasonable scientific hypothesis that could explain these inconsistent results. for other apis, scientific explanations are less clear and may require further exploration using carefully designed experimental approaches to assess permeability. it is generally assumed that physiologically-based in silico absorption models are promising supplemental tools to traditional in vitro assays and preclinical in vivo studies. however, as recently reported by sjögren and colleagues [40], different modeling platforms are associated with inconsistent overand under-prediction of the fraction absorbed. consequently, a broader data mining approach will be necessary to maximize confidence in model performance. conclusions the results from this study demonstrate successful validation of a globally harmonized solubility protocol across different international laboratories to support who bcs classification. implementation of a consensus approach that compares in silico predicted fa in humans using the gastroplus © pbpk simulation platform and clinically determined bioavailability data for the same api but measured at different doses enabled unambiguous assignment of 10 out of the 16 apis selected for this pilot study to one of the four bcs classes. further expansion of these experimental efforts to qualified apis from the who essential medicines list is predicted to provide regulatory authorities across the globe with scientifically validated data to support decisions regarding the need for in vivo bioequivalence studies. ultimately, this will improve access to affordable generic products, which is a critical prerequisite to reach universal health coverage. acknowledgements equilibrium solubility experiments were conducted by universities, official national control laboratories, pharmacopoeia laboratory, and who collaborating centres. the authors warmly thank: liezl badenhorst, north_west university, potchefstroom, south africa; jeronimo r. de oliveira neto, faculty of pharmacy, federal university of goiás, goiânia, goiás, brazil; elena soler, universidad miguel hernández de elche, alicante, spain; matilde merino-sanjuán, university of valencia, valencia, spain; nilesh m. meghani, college of pharmacy and institute of pharmaceutical science and technology, ajou university, suwon, republic of korea; tri-hung nguyen, medicines manufacturing innovation centre, monash institute of pharmaceutical sciences, monash university, parkville, australia; raju kanumuri, university of cincinnati, cincinnati, ohio, usa; satish agrawal, st. louis college of pharmacy, st, louis, missouri, usa; jing xiong, national institutes for food and drug control (nifdc), beijing, china; ramesh kumar jha, indian pharmacopoeia commission, ministry of health & family welfare, govt. of india, ghaziabad, india; for the precious collaboration in producing experimental data. all apis studied in cycles i and ii were received as in-kind donations from pharmaceutical manufacturers admet & dmpk 9(1) (2021) 23-39 enhancing robustness of the who biopharmaceutical classification system doi: https://dx.doi.org/10.5599/admet.850 37 supporting who in this scientific work: acs dobfar; alcaliber s.a.u.; aurobindo; cipla ltd; guilin pharmaceutical co. ltd; ipca laboratories ltd; laurus labs limited; lupin ltd.; macleods pharmaceuticals ltd; mangalam drugs & organics ltd; mylan laboratories ltd; sandoz; sanofi; shanghai desano chemical pharmaceutical co. ltd; shenyang antibiotic; pfizer; zhejiang jiangbei pharmaceutical co. ltd. conflict of interest: the authors report no conflict of interest. disclaimer: this manuscript represents the personal opinion of the authors and does not necessarily represent the views or policy of their corresponding institutions. references [1] world health organization, access to medicines: making market forces serve the poor. ten years in public health 2007–2017., (2017). https://www.who.int/publications/10-year-review/en/ (accessed may 6, 2020). [2] the world medicines situation: chapter 7. access to essential medicines, (1999). http://digicollection.org/hss/en/d/js6160e/9.html (accessed may 6, 2020). [3] g.l. amidon, h. lennernäs, v.p. shah, and j.r. crison, a theoretical basis for a biopharmaceutic drug classification: the correlation of in vitro drug product dissolution and in vivo bioavailability, pharm. res. an off. j. am. assoc. pharm. sci. 12 (1995) 413–420. doi: https://doi.org/10.1023/a:1016212804288. [4] world health organization, the selection and use of essential medicines the selection and use of essential medicines who technical report series, geneva, 2019. http://www.who.int/ (accessed may 6, 2020). [5] world health organization, multisource (generic) pharmaceutical products: guidelines on registration requirements to establish interchangeability, in: who expert comm. specif. pharm. prep., geneva, 2017. https://www.who.int/medicines/areas/quality_safety/quality_assurance/expert_committee/who_t rs_1003_full-version.pdf?ua=1 (accessed may 6, 2020). [6] world health organization, proposal to waive in vivo bioequivalence requirements for who model list of essential medicines immediate-release, solid oral dosage forms, 2006. https://www.who.int/medicines/areas/quality_safety/quality_assurance/proposalwaivevivobioequi valencerequirementsmodellistessentialmedicinesimmediatereleasesolidoraldosageformstrs937a nnex8.pdf?ua=1 (accessed may 6, 2020). [7] world health organization, who expert committee on specifications for pharmaceutical preparations the expert committee on specifications for pharmaceutical, world health organization, geneva, 2020. https://apps.who.int/iris/handle/10665/331814 (accessed may 6, 2020). [8] world health organization, the international pharmacopoeia, ninth edition, (2019). https://apps.who.int/phint/en/p/docf/ (accessed may 6, 2020). [9] world health organization, annex 4 protocol to conduct equilibrium solubility experiments for the purpose of biopharmaceutics classification system-based classification of active pharmaceutical ingredients for biowaiver, in: who expert comm. specif. pharm. prep. fifty-third rep., 2019. https://www.who.int/medicines/areas/quality_safety/quality_assurance/who_trs_1019_annex4.p df?ua=1 (accessed may 6, 2020). [10] c.a. larregieu, and l.z. benet. drug discovery and regulatory considerations for improving in silico and in vitro predictions that use caco-2 as a surrogate for human intestinal permeability measurements. aaps j. 15 (2013) 483–497. doi: https://doi.org/10.1208/s12248-013-9456-8. [11] b. agoram, w.s. woltosz, and m.b. bolger. predicting the impact of physiological and biochemical processes on oral drug bioavailability. adv. drug deliv. rev. 50 (2001). doi: https://doi.org/10.1016/s0169-409x(01)00179-x. https://dx.doi.org/10.5599/admet.850 https://www.who.int/publications/10-year-review/en/ http://digicollection.org/hss/en/d/js6160e/9.html https://doi.org/10.1023/a:1016212804288 http://www.who.int/ https://www.who.int/medicines/areas/quality_safety/quality_assurance/expert_committee/who_trs_1003_full-version.pdf?ua=1 https://www.who.int/medicines/areas/quality_safety/quality_assurance/expert_committee/who_trs_1003_full-version.pdf?ua=1 https://www.who.int/medicines/areas/quality_safety/quality_assurance/proposalwaivevivobioequivalencerequirementsmodellistessentialmedicinesimmediatereleasesolidoraldosageformstrs937annex8.pdf?ua=1 https://www.who.int/medicines/areas/quality_safety/quality_assurance/proposalwaivevivobioequivalencerequirementsmodellistessentialmedicinesimmediatereleasesolidoraldosageformstrs937annex8.pdf?ua=1 https://www.who.int/medicines/areas/quality_safety/quality_assurance/proposalwaivevivobioequivalencerequirementsmodellistessentialmedicinesimmediatereleasesolidoraldosageformstrs937annex8.pdf?ua=1 https://apps.who.int/iris/handle/10665/331814 https://apps.who.int/phint/en/p/docf/ https://www.who.int/medicines/areas/quality_safety/quality_assurance/who_trs_1019_annex4.pdf?ua=1 https://www.who.int/medicines/areas/quality_safety/quality_assurance/who_trs_1019_annex4.pdf?ua=1 https://doi.org/10.1208/s12248-013-9456-8 https://doi.org/10.1016/s0169-409x(01)00179-x valeria gigante et al. admet & dmpk 9(1) (2021) 23-39 38 [12] a. avdeef, e. fuguet, a. llinàs, c. ràfols, e. bosch, g. völgyi, t. verbic, e. boldyreva, and k. takácsnovák. equilibrium solubility measurement of ionizable drugs consensus recommendations for improving data quality. admet dmpk 4 (2016) 117–178. doi: https://doi.org/10.5599/admet.4.2.292. [13] d. thambavita, p. galappatthy, u. mannapperuma, l. jayakody, r. cristofoletti, b. abrahamsson, d.w. groot, p. langguth, m. mehta, a. parr, j.e. polli, v.p. shah, and j. dressman. biowaiver monograph for immediate-release solid oral dosage forms: amoxicillin trihydrate. j. pharm. sci. 106 (2017) 2930–2945. doi: https://doi.org/10.1016/j.xphs.2017.04.068. [14] j. arnal, i. gonzalez-alvarez, m. bermejo, g.l. amidon, h.e. junginger, s. kopp, k.k. midha, v.p. shah, s. stavchansky, j.b. dressman, and d.m. barends. biowaiver monographs for immediate release solid oral dosage forms: aciclovir. j. pharm. sci. 97 (2008) 5061–5073. doi: https://doi.org/10.1002/jps.21392. [15] a. ono, n. matsumura, t. kimoto, y. akiyama, s. funaki, n. tamura, s. hayashi, y. kojima, m. fushimi, h. sudaki, r. aihara, y. haruna, m. jiko, m. iwasaki, t. fujita, and k. sugano. harmonizing solubility measurement to lower inter-laboratory variance progress of consortium of biopharmaceutical tools (cobito) in japan. admet dmpk 7 (2019) 183–195. doi: https://doi.org/10.5599/admet.704. [16] a. avdeef. suggested improvements for measurement of equilibrium solubility-ph of ionizable drugs. admet dmpk 3 (2015) 84–109. doi: https://doi.org/10.5599/admet.3.2.193. [17] a. katritzky, y. wang, s. sild, t. tamm, and m. karelson. qspr studies on vapor pressure, aqueous solubility, and the prediction of water−air partition coefficients. j. chem. inf. comput. sci. 38 (1998) 720–725. doi: https://doi.org/10.1021/ci980022t. [18] m.a. hofsäss, and j.b. dressman. the discriminatory power of the bcs-based biowaiver: a retrospective with focus on essential medicines. j. pharm. sci. 108 (2019) 2824–2837. doi: https://doi.org/10.1016/j.xphs.2019.04.030. [19] g. paintaud, g. alván, m.l. dahl, a. grahnén, j. sjövall, and j.o. svensson. nonlinearity of amoxicillin absorption kinetics in human. eur. j. clin. pharmacol. 43 (1992) 283–288. doi: https://doi.org/10.1007/bf02333024. [20] p. luckner, and m. brandsch. interaction of 31 β-lactam antibiotics with the h +/peptide symporter pept2: analysis of affinity constants and comparison with pept1. eur. j. pharm. biopharm. 59 (2005) 17–24. doi: https://doi.org/10.1016/j.ejpb.2004.07.008. [21] t.t. mariappan, and s. singh. evidence of efflux-mediated and saturable absorption of rifampicin in rat intestine using the ligated loop and everted gut sac techniques. mol. pharm. 1 (2004) 363–367. doi: https://doi.org/10.1021/mp049937n. [22] m. bermejo, b. hens, j. dickens, d. mudie, p. paixão, y. tsume, k. shedden, and g.l. amidon. a mechanistic physiologically-based biopharmaceutics modeling (pbbm) approach to assess the in vivo performance of an orally administered drug product: from ivivc to ivivp. pharmaceutics 12 (2020). doi: https://doi.org/10.3390/pharmaceutics12010074. [23] r. jereb, j. opara, i. legen, b. petek, and d. grabnar-peklar. in vitro–in vivo relationship and bioequivalence prediction for modified-release capsules based on a pbpk absorption model. aaps pharmscitech 21 (2020). doi: https://doi.org/10.1208/s12249-019-1566-x. [24] p. de miranda, and m. blum. pharmacokinetics of acyclovir after intravenous and oral administration. j. antimicrob. chemother. 12 (1983) 29–37. doi: https://doi.org/10.1093/jac/12.suppl_b.29. [25] g. foulds, r. shepard, and r. johnson. the pharmacokinetics of azithromycin in human serum and tissues. pubmed ncbi. j. antimicrob. chemother. 25 (1990) 73–82. doi: https://doi.org/10.1093/jac/25.suppl_a.73. [26] r.d. faulkner, p. fernandez, g. lawrence, l.l. sia, a.j. falkowski, a.i. weiss, a. yacobi, and b.m. silber. absolute bioavailability of cefixime in man. j. clin. pharmacol. 28 (1988) 700–706. doi: https://doi.org/10.1002/j.1552-4604.1988.tb03203.x. https://doi.org/10.5599/admet.4.2.292 https://doi.org/10.1016/j.xphs.2017.04.068 https://doi.org/10.1002/jps.21392 https://doi.org/10.5599/admet.704 https://doi.org/10.5599/admet.3.2.193 https://doi.org/10.1021/ci980022t https://doi.org/10.1016/j.xphs.2019.04.030 https://doi.org/10.1007/bf02333024 https://doi.org/10.1016/j.ejpb.2004.07.008 https://doi.org/10.1021/mp049937n https://doi.org/10.3390/pharmaceutics12010074 https://doi.org/10.1208/s12249-019-1566-x https://doi.org/10.1093/jac/12.suppl_b.29 https://doi.org/10.1093/jac/25.suppl_a.73 https://doi.org/10.1002/j.1552-4604.1988.tb03203.x admet & dmpk 9(1) (2021) 23-39 enhancing robustness of the who biopharmaceutical classification system doi: https://dx.doi.org/10.5599/admet.850 39 [27] j.h. hull, j.w.a. findlay, j.f. rogers, r.m. welch, r.f. butz, and j.a. bustrack. an evaluation of the effects of smoking on codeine pharmacokinetics and bioavailability in normal human volunteers. drug intell. clin. pharm. 16 (1982) 849–854. doi: https://doi.org/10.1177/106002808201601107. [28] y. gandhi, t. eley, a. fura, w. li, r.j. bertz, and t. garimella. daclatasvir: a review of preclinical and clinical pharmacokinetics. clin. pharmacokinet. 57 (2018) 911–928. doi: https://doi.org/10.1007/s40262-017-0624-3. [29] c. wolfe, and c. hicks. profile of darunavir in the management of treatment-experienced hiv patients. hiv/aids res. palliat. care. (2009) 13. doi: https://doi.org/10.2147/hiv.s4842. [30] fda, dolutegravir. decisional review for nda 204790., 2013. https://www.accessdata.fda.gov/scripts/cder/daf/index.cfm?event=overview.process&varapplno=2 04790 (accessed may 6, 2020). [31] fda, sustiva approval history: application number 20-360., 2011. https://www.accessdata.fda.gov/drugsatfda_docs/nda/98/20972biopharm_review.pdf. [32] h.b. nyberg, h.r. draper, a.j. garcia-prats, s. thee, a. bekker, h.j. zar, a.c. hooker, h. simon schaaf, h. mcilleron, a.c. hesseling, and p. denti., population pharmacokinetics and dosing of ethionamide in children with tuberculosis. antimicrob. agents chemother. 64 (2020). doi: https://doi.org/10.1128/aac.01984-19. [33] d. smith, e. lin, and l. benet. absorption and disposition of furosemide in healthy volunteers, measured with a metabolite-specific assay. drug metab. dispos. 8 (1980) 337–342. https://www.ncbi.nlm.nih.gov/pubmed/?term=6107232 (accessed may 6, 2020). [34] g. mihaly, s. ward, g. edwards, d. nicholl, m. orme, and a. breckenridge. pharmacokinetics of primaquine in man. i. studies of the absolute bioavailability and effects of dose size. br. j. clin. pharmacol. 19 (1985) 745–750. doi: https://doi.org/10.1111/j.1365-2125.1985.tb02709.x. [35] d.s. almond, i.s.f. szwandt, g. edwards, m.g. lee, and p.a. winstanley. disposition of intravenous pyrimethamine in healthy volunteers. antimicrob. agents chemother. 44 (2000) 1691–1693. doi: https://doi.org/10.1128/aac.44.6.1691-1693.2000. [36] d.m. brainard, l.a. wenning, j.a. stone, j.a. wagner, and m. iwamoto. clinical pharmacology profile of raltegravir, an hiv-1 integrase strand transfer inhibitor. j. clin. pharmacol. 51 (2011) 1376–402. doi: https://doi.org/10.1177/0091270010387428. [37] u. loos, e. musch, j.c. jensen, g. mikus, h.k. schwabe, and m. eichelbaum. pharmacokinetics of oral and intravenous rifampicin during chronic administration. klin. wochenschr. 63 (1985) 1205–1211. doi: https://doi.org/10.1007/bf01733779. [38] p. barditch-crovo, s.g. deeks, a. collier, s. safrin, d.f. coakley, m. miller, b.p. kearney, r.l. coleman, p.d. lamy, j.o. kahn, i. mcgowan, and p.s. lietman. phase i/ii trial of the pharmacokinetics, safety, and antiretroviral activity of tenofovir disoproxil fumarate in human immunodeficiency virus-infected adults. antimicrob. agents chemother. 45 (2001) 2733–2739. doi: https://doi.org/10.1128/aac.45.10.2733-2739.2001. [39] p. bransford, j. cook, m. gupta, s. haertter, h. he, r. ju, j. kanodia, h. lennernäs, d. lindley, j.e. polli, l. wenning, and y. wu. ich m9 guideline in development on biopharmaceutics classification system-based biowaivers: an industrial perspective from the iq consortium. mol. pharm. 17 (2020) 361–372. doi: https://doi.org/ 10.1021/acs.molpharmaceut.9b01062. [40] e. sjögren, h. thörn, and c. tannergren. in silico modeling of gastrointestinal drug absorption: predictive performance of three physiologically based absorption models. mol. pharm. 13 (2016) 1763–1778. doi: https://doi.org/10.1021/acs.molpharmaceut.5b00861. ©2021 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) https://dx.doi.org/10.5599/admet.850 https://doi.org/10.1177/106002808201601107 https://doi.org/10.1007/s40262-017-0624-3 https://doi.org/10.2147/hiv.s4842 https://www.accessdata.fda.gov/scripts/cder/daf/index.cfm?event=overview.process&varapplno=204790 https://www.accessdata.fda.gov/scripts/cder/daf/index.cfm?event=overview.process&varapplno=204790 https://www.accessdata.fda.gov/drugsatfda_docs/nda/98/20972biopharm_review.pdf https://doi.org/10.1128/aac.01984-19 https://doi.org/10.1111/j.1365-2125.1985.tb02709.x https://doi.org/10.1128/aac.44.6.1691-1693.2000 https://doi.org/10.1177/0091270010387428 https://doi.org/10.1007/bf01733779 https://doi.org/10.1128/aac.45.10.2733-2739.2001 https://doi.org/%2010.1021/acs.molpharmaceut.9b01062 https://doi.org/10.1021/acs.molpharmaceut.5b00861 http://creativecommons.org/licenses/by/3.0/ permeability evaluation of gemcitabine-cpp6 conjugates in caco-2 cells doi: http://dx.doi.org/10.5599/admet.882 41 admet & dmpk 9(1) (2021) 41-48; doi: https://dx.doi.org/10.5599/admet.882 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper permeability evaluation of gemcitabine-cpp6 conjugates in caco-2 cells abigail ferreira 1,2 , sara moreira 1,3 , rui lapa 2 and nuno vale 1,4* 1 oncopharma research group, center for health technology and services research (cintesis), rua dr. plácido da costa, 4200-450 porto, portugal 2 laqv/requimte, laboratory of applied chemistry, department of chemical sciences, faculty of pharmacy, university of porto, rua de jorge viterbo ferreira, 228, 4050-313 porto, portugal 3 faculty of sciences, university of porto, rua do campo alegre, 687, 4169-007 porto, portugal 4 faculty of medicine, university of porto, al. prof. hernâni monteiro, 4200-319 porto, portugal *corresponding author: e-mail: nunovale@med.up.pt; tel.: +351-225513600; fax: +351-225513601 received: july 04, 2020; revised: october 13, 2020; published: october 26, 2020 abstract cancer is one of the most alarming diseases due to its high mortality and still increasing incidence rate. currently available treatments for this condition present several shortcomings and new options are continuously being developed and evaluated, aiming at increasing the overall treatment efficiency and reducing associated adverse side effects. gemcitabine has proven activity and is used in chemotherapy. however, its therapeutic efficiency is limited by its low bioavailability as a result of rapid enzymatic inactivation. additionally, tumor cells often develop drug resistance after initial tumor regression related to transporter deficiency. we have previously developed three gemcitabine conjugates with cellpenetrating hexapeptides (cpp6) to facilitate intracellular delivery of this drug while also preventing enzymatic deamination. the bioactivity of these new prodrugs was evaluated in different cell lines and showed promising results. here, we assessed the absorption and permeability across caco-2 monolayers of these conjugates in comparison with gemcitabine and the respective isolated cell-penetrating peptides (cpps). cpp6-2 (klpvmw) and respective gem-cpp6-2 conjugate showed the highest permeability in caco2 cells. ©2021 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords cell-penetrating peptides; permeability; solubility; caco-2; dmpk. introduction cancer is the second leading cause of death worldwide. there were 18 million new cases and 9.6 million cancer related deaths in 2018. globally, about one in six deaths is due to cancer. lung and prostate cancers are the most prevalent in men, while breast and colorectum cancers affect women the most [1]. drug resistance and overall treatment inefficacy are responsible for the high mortality of this multifactorial disease. difficulty in recognizing altered cells and distinguishing them from normal cells as well as reaching metastases are the main hurdles of cancer therapy. research in this field strives to maximize efficacy while reducing adverse side effects. http://dx.doi.org/10.5599/admet.882 https://dx.doi.org/10.5599/admet.882 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:nunovale@med.up.pt http://creativecommons.org/licenses/by/4.0/ n. vale et al. admet & dmpk 9(1) (2021) 41-48 42 gemcitabine (gem, dfdc or 2′,2′-difluoro-2′-deoxycytidine, scheme 1) is a cytotoxic nucleoside analogue used as a chemotherapeutic drug. it is effective against an extensive range of solid tumors, such as pancreatic, non-small cell lung, breast and ovarian cancers, and is often a first-line treatment in clinical settings [2]. despite being administered intravenously, treatment with gemcitabine has limited efficacy largely due to its short half-life (8-17 min), since gemcitabine is rapidly inactivated metabolically in the serum through deamination of its 4-n amine by cytidine deaminase (cda), present in high levels in both human plasma and liver. as such, much higher doses are required to reach an effective plasma concentration, increasing toxicity and the risk of adverse side effects. another major obstacle is the drug resistance related to nucleoside transporter deficiency, which is developed by some tumor cells after initial tumor regression. as a hydrophilic drug, gemcitabine cellular uptake is primarily facilitated by the human equilibrative nucleoside transporter 1 (hent1). thus, the expression of this transporter plays a key role in gemcitabine intracellular uptake [3]. there have been many efforts to improve gemcitabine clinical efficacy and numerous derivatives and prodrugs have been designed to alter some of the unfavorable physicochemical properties of gemcitabine and ideally improve its oral bioavailability [4]. our research group has previously developed gemcitabine conjugates with cell-penetrating peptides (cpps) to facilitate intracellular delivery of this drug [5-6]. in three of these new prodrugs, a cell-penetrating hexapeptide (cpp6) was covalently conjugated to the aniline moiety of gemcitabine through suitable bio-/chemoreversible bonds. two cpp5 (klpvm and vpmlk) with reported high percentage of cell-penetration were selected. one tryptophan (trp) residue was added to the terminal of these cpp5 to further improve their capacity of cell-penetration, since trp has high propensity to be inserted into membranes [7-8], yielding three novel hexapeptides: cpp6-1 (wklpvm), cpp6-2 (klpvmw), and cpp6-3 (wvpmlk). these conjugates were designed to facilitate intracellular delivery of this drug and even overcome the problem of transporter deficiency related drug resistance, as cpps are non-cytotoxic vectors and drug delivery vehicles. additionally, gemcitabine is protected from cda enzymatic deamination since its 4-n amine was modified. the bioactivity of these new prodrugs was previously evaluated in different cell lines and showed promising results [5]. in this work, we have evaluated the absorption and permeability of gemcitabine, three cpp6 and the respective gem-cpp6 conjugates across caco-2 monolayers. the single layer of epithelial cells that covers the inner intestinal wall forms the rate-limiting barrier to the absorption of dissolved compounds administered orally. consequently, proper reconstitution of a human differentiated epithelial cell monolayer in vitro can be used to predict the absorption and permeability coefficients of orally administered drugs. the human colon carcinoma cell line caco-2 has been successfully used to fulfill this purpose and monolayers grown on permeable filters have become the in vitro golden standard for these predictions [9]. this method is recognized by the american food and drug administration (fda) [10]. experimental cell culture human colon adenocarcinoma cell line caco-2 (passage 25-47, kindly provided by the department of biomedicine of the faculty of medicine of university of porto, and previously acquired via atcc) was routinely maintained in dulbecco’s modified eagle’s medium (dmem) supplemented with 10 % fetal bovine serum (fbs), l-glutamine and antibiotics penicillin and streptomycin. cells were cultured at 37 °c in a 5 % co2 humidified atmosphere. culturing medium was replaced every 2-3 days and cell subculture was conducted once a week by trypsinization (0.25 % trypsin-edta wt/vol, 5 min, 37 °c). admet & dmpk 9(1) (2021) 41-48 permeability of gemcitabine-cpp6 conjugates in caco-2 cells doi: http://dx.doi.org/10.5599/admet.882 43 cytotoxicity assay the cell growth inhibitory activity of gemcitabine in caco-2 cells was assessed with the mtt assay. briefly, a growth curve was traced to determine the best cell density for the assay. cells were seeded in 96well plates (growth surface 0.322 cm 2 , from tpp®, product no. 92696) with an initial cell density of 9.3 x 10 3 cells/ml (200 μl per well). cells were allowed to attach for 24 h and were then either left untreated (culture medium was replaced for fresh medium) or treated with gemcitabine (0.1, 1, 5, 10, 20, 50, 100, 1000, 10000 and 100000 µm). following 72 h incubation, cell medium was removed and 100 μl of mtt (3(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) solution were added per well. cells were then incubated for another 4 h protected from light. finally, mtt solution was removed, dmso (100 μl per well) was added to solubilize the formazan crystals formed by viable cells and absorbance was measured at 570 nm in an automated microplate reader (sinergy ht, biotek instruments inc, vermont, usa). all conditions were performed in triplicate. gemcitabine cytotoxicity results were compared with the untreated control (mean of values was set to 100 %) and expressed as mean ± sem. the statistical significance between different gemcitabine concentrations was analyzed in graphpad prism 7 (san diego, eua) using one-way anova (p<0.05). scheme 1. chemical structure of gemcitabine (before and after modification and conjugation with cpp6) and of the three cpp6 studied, cpp6-1 (wklpvm), cpp6-2 (klpvmw) and cpp6-3 (wvpmlk). preparation of cpps and gemcitabine-cpp6 conjugates all the compounds evaluated in this work (except for gemcitabine, that was purchase from sigmaaldrich as gemcitabine hydrochloride, g6423) had been previously synthesized by our research group [5]. briefly, cpp6-1 (wklpvm), cpp6-2 (klpvmw) and cpp6-3 (wvptlk) were manually synthesized using gemcitabine gemcitabine gemcitabine http://dx.doi.org/10.5599/admet.882 n. vale et al. admet & dmpk 9(1) (2021) 41-48 44 standard solid phase peptide synthesis (spps) and fmoc chemistry. gemcitabine was modified to allow conjugation with cpp6. first, the two hydroxyl groups of gemcitabine were selectively protected with boc groups; then, succinic anhydride was linked to the 4-n amine and lastly a cpp6 was conjugated to this anticancer drug. some properties of these new conjugates are presented in table 1. permeability assay a monolayer of caco-2 cells was established in a 12-well plate [11]. each well contains a permeable filter insert, a transparent collagen treated (equimolar mixture of types i and iii collagen) polytetrafluoroethylene (ptfe) membrane, with 12 mm diameter and 0.4 µm pore (corning inc, corning transwell col collagen coated membrane inserts, ny, usa, cat. no. 3493). firstly, filters were pre-wet with 0.1 ml of culture medium for 2 minutes. cells were seeded in the apical side (0.5 ml of cells per well, cell density of 4.0 x 10 5 cells/ml). the basolateral compartment was filled with 1.5 ml of culture medium and plates were incubated for 6 h (5 % co2 humidified atmosphere at 37 °c). then, to remove non-adherent cells and reduce the risk of multilayer formation, the medium in the apical side was removed and replaced with fresh medium. cells were maintained for 29 days, replacing the culture medium from both compartments every other day. transport across caco-2 monolayer assay all the solutions used were pre-warmed to 37 °c. culture medium was replaced with fresh medium 24 h prior the beginning of the experiment. next, this culture medium was removed first from the basolateral and then from the apical compartment. the apical compartment was carefully washed and then filled with 0.5 ml of hank’s balanced salt solution (hbss, ph 7.4) and the basolateral compartment was also filled with 1.5 ml of hbss. the plates were incubated for 17 min at 37 °c under gentle shaking (190 rpm, gfl® orbital shaker 3015). all tested compounds were prepared in hbbs (60 µm) and added to the apical compartment. throughout this assay, the final volume was 0.4 ml in the apical compartment and 1.2 ml in the basolateral chamber. at t = 0 min, 0.45 ml of the donor solution was added to the apical compartment and a sample (0.05 ml) was immediately taken. the plate was incubated (lid-covered) at 37 °c under gentle shaking (190 rpm). every 30 min for the next 2 h, a sample of 0.6 ml was taken from the basolateral compartment and replaced with the same volume of hbss. after 120 min, a sample of 0.05 ml was taken from the apical side. hplc quantification the concentrations of the evaluated compound in the basolateral and apical compartments were determined by high-performance liquid chromatography (hplc) (vwr international lcc, lachrom ultra, pennsylvania, usa). elution was performed with a variable gradient of acetonitrile (acn) in water containing 0.05 % trifluoroacetic acid (tfa), at a 0.7 ml/min flow and detection at variable wavelength (243 nm for gemcitabine and 220 nm for peptides and conjugates). all chemicals were either analytical or hplc grade. results and discussion gemcitabine cytotoxicity every concentration of gemcitabine tested caused a statistically significant inhibition of cell growth in caco-2 cells compared to untreated cells (figure 1). the maximum inhibition was observed after treatment with the highest concentration of gemcitabine evaluated (100000 µm). the ic50 was calculated as 52.4 µm. admet & dmpk 9(1) (2021) 41-48 permeability of gemcitabine-cpp6 conjugates in caco-2 cells doi: http://dx.doi.org/10.5599/admet.882 45 figure 1. concentration-dependent cytotoxicity of gemcitabine in caco-2 cells (abs 570 nm). results are presented as (a) the percentage of inhibition compared to control untreated cells and (b) plotted as the log [gemcitabine] against the percentage of control untreated cells. results are expressed as mean ± sem (n=3; p<0.05, one-way anova). caco-2 monolayer and permeability of gemcitabine microscopic analysis of the filter revealed a uniformly formed monolayer, with no detection of anomalies or areas without cells. a caco-2 monolayer was successfully established and this is a reliable indicator for proceeding with the permeability study. after 30 minutes, just 3 % of gemcitabine was found in the basolateral side. at the 90 min time point, a plateau was reached and only an additional 1 % of the drug crossed the membrane until the end of assay (120 min), with a total of 18 % of the amount of gemcitabine initially applied in the apical side having crossed into the basolateral compartment. permeability of cpp6 comparable amounts of the 3 cpp6 crossed the monolayer of caco-2 cells after 120 minutes. still, cpp62 exhibited the highest permeability, with 40 % being recovered in the basolateral chamber, followed by cpp6-3 (37 %) and cpp6-1 (33 %). during the first 30 minutes, only 4 to 5 % of all cpps reached the basolateral side; the rate of permeation was highest between 30 and 90 minutes. cpp6-2 and cpp6-3 registered a decrease in this rate for the last 30 minutes of this assay. throughout the experiment, a slower rate of permeation was observed for cpp6-1 (figure 2). by the end of the experiment, the sum of the cpp absorbed from the apical side and found in the basolateral chamber was 84 % for cpp6-2, 75 % for cpp6-3 and 50 % for cpp6-1. this indicates that some http://dx.doi.org/10.5599/admet.882 n. vale et al. admet & dmpk 9(1) (2021) 41-48 46 peptide is retained inside the cells or was degraded. figure 2. concentration of cpp6 in the basolateral compartment after incubation with 60 µm in the apical compartment and fitted curves. permeability of gem-cpp6 conjugates as for the gem-cpp6 conjugates, the conjugate of gemcitabine with cpp6-2 clearly stands out as the most permeable across the caco-2 cell monolayer. the extent of permeation of this conjugate was about 3fold the observed for the gem-cpp6-1 and gem-cpp6-3 conjugates. approximately 31 % of the gem-cpp62 conjugate was found in the basolateral chamber, while only 9 % of gem-cpp6-1 and gem-cpp6-3 conjugates were able to cross the monolayer of caco-2 cells and was quantified in this compartment (figure 3). the higher permeability of the gem-cpp6-2 conjugate is in agreement with the previous evaluation of the permeability of the isolated cpp6, with cpp6-2 being the most permeable peptide. of the three studied cpp6, cpp6-2 is the only one that has a trp residue in its c-terminal (cpp6-1 and cpp6-3 have a trp residue in the n-terminal position). additionally, the only difference between cpp6-1 and cpp6-2 is the position of this amino acid residue (n-terminal versus c-terminal). tryptophan can be considered hydrophobic due to its uncharged side chain. given its aromatic character, it can form hydrogen bonds and strongly interact with cellular membranes, disrupting the stability of the membrane lipidic chains. these properties grant this amino acid residue a great tendency to insert into membranes [12-15]. it has been shown that the replacement of the 2 trp residues of the known cpp penetratin for phenylalanine residues completely eliminates the penetration ability of this peptide [16]. furthermore, rydberg et al. have reported that the number of trp residues in a cpp is relevant, with a higher number of trp residues in a cpp corresponding to greater penetration ability [12]. these authors have also demonstrated that the position of this residue is another factor influencing permeability; in this case, cpps with trp residues in the n-terminal position were less absorbed. some properties of these new gem-cpp6 conjugates were predicted using in silico tools and are presented in table 1. gem-cpp6-2 has the highest calculated papp and is also predicted as the most permeable conjugate by gastroplus™, a mechanistically based pharmacokinetics and pharmacodynamics simulation software. admet & dmpk 9(1) (2021) 41-48 permeability of gemcitabine-cpp6 conjugates in caco-2 cells doi: http://dx.doi.org/10.5599/admet.882 47 figure 3. permeability of the gem-cpp6 conjugates, expressed as the percentage recovered in the basolateral compartment. table 1. physicochemical and in silico pharmacokinetic data of cpp6-gemcitabine conjugates. peptide/conjugate mw (g/mol) psa a (å 2 ) hbd peff (cm/s) b (x10 -6 ) peff (cm/s) c (x10 -9 ) papp (cm/s) (x10 -8 ) gem-cpp6-1 1117.29 247.63 13 3.19 1.31 1.326 gem-cpp6-2 1117.29 247.63 13 3.82 1.31 4.568 gem-cpp6-3 1087.20 267.86 14 2.47 0.412 1.326 a polar surface area, predicted by medchem designer (version 3.1.0.30, simulations plus, inc., lancaster, california, usa); b effective permeability, predicted by gastroplus™ (version 9.5, simulations plus, inc., lancaster, california, usa); c effective permeability, calculated from winiwarter equation: log peff = (-2.546) – (0.011 x psa) – (0.278 x hbd); hbd: hydrogen bond donors; [17]. conclusions the monolayer of caco-2 cells was a suitable method to study the permeability of the anticancer drug gemcitabine, the cell-penetrating hexapeptides (cpp6) and the new prodrug conjugates gem-cpp6. every cpp6 and gem-cpp6 conjugate was able to cross this monolayer, suggesting their potential as drug delivery systems. in agreement with previous reports, the position of the trp residue was a determinant factor influencing the permeability of the cpp6 and gem-cpp6 conjugates. amongst the evaluated compounds, cpp6-2 and the respective gem-cpp6-2 conjugate showed the highest permeability, crossing the cell monolayer to a greater extent. this data puts forward this conjugate as a lead for further studies and as the most promising for potential clinical application. acknowledgements this work was financed by feder – fundo europeu de desenvolvimento regional through the compete 2020 – operational programme for competitiveness and internationalization (poci), portugal 2020, and by portuguese funds through fct – fundação para a ciência e a tecnologia, in a framework of cintesis, r&d unit (reference uidb/4255/2020). n.v. also thanks support fct and feder (european union), award number if/00092/2014/cp1255/ct0004. rl thanks fct through grant uid/qui/50006/2019 (laqvrequimte). af thanks fct for a doctoral fellowship (pd/bd/135120/2017). the contents of this article are solely the responsibility of the authors and do not necessarily represent the official view of the fct. conflict of interest: on behalf of all authors, the corresponding author states that there is no conflict of interest. http://dx.doi.org/10.5599/admet.882 n. vale et al. admet & dmpk 9(1) (2021) 41-48 48 references [1] global cancer burden in 2018 (globocan 2018 database, international agency for research on cancer). https://gco.iarc.fr/today/data/factsheets/populations/900-world-fact-sheets.pdf (accessed june 10, 2020). [2] a. wong, r.a. soo, w.-p. yong, f. innocenti. clinical pharmacology and pharmacogenetics of gemcitabine. drug metabolism reviews 41 (2009) 77-88. [3] e. mini, s. nobili, b. caciagli, i. landini, t. mazzei. cellular pharmacology of gemcitabine. annals of oncology 17 (2006) 7-12. [4] e. moysan, g. bastiat, j.-p. benoit. gemcitabine versus modified gemcitabine: a review of several promising chemical modifications. molecular pharmaceutics 10 (2013) 430-444. [5] c. correia, c.p.r. xavier, d. duarte, a. ferreira, s. moreira, m.h. vasconcelos, n. vale. development of potent cpp6-gemcitabine conjugates against human prostate cancer cell line (pc-3). rsc medicinal chemistry 11 (2020) 268-273. [6] n. vale, a. ferreira, i. fernandes, c. alves, m.j. araujo, n. mateus, p. gomes. gemcitabine antiproliferative activity significantly enhanced upon conjugation with cell-penetrating peptides. bioorganic and medicinal chemistry 27 (2017) 2898-2901. [7] k. kachel, e. asuncion-punzalan, e. london. anchoring of tryptophan and tyrosine analogs at the hydrocarbon-polar boundary in model membrane vesicles. biochemistry 34 (1995) 15475-15479. [8] d.j. schibli, r.f. epand, h.j. vogel, r.m. epand. tryptophan-rich antimicrobial peptides: comparative properties and membrane interactions. biochemistry cell biology 80 (2002) 667-677. [9] i.d. angelis, l. turco. caco‐2 cells as a model for intestinal absorption. current protocols in toxicology 47 (2011) 20.6.1-20.6.15. [10] u.s. department of health and human services, food and drug administration, center for drug evaluation and research (cder). waiver of in vivo bioavailability and bioequivalence studies for immediate-release solid oral dosage forms based on a biopharmaceutics classification system. december 2017, biopharmaceutics. [11] i. hubatsch, e.g. ragnarsson, p. artursson. determination of drug permeability and prediction of drug absorption in caco-2 monolayers. nature protocols 2 (2007) 2111-2119. [12] h.a. rydberg, m. matson, h.l. åmand, e.k. esbjorner, b. nordén. effects of tryptophan content and backbone spacing on the uptake efficiency of cell-penetrating peptides. biochemistry 51 (2012) 55315539. [13] d.i. chan, e.j. prenner, h.j. vogel. tryptophan-and arginine-rich antimicrobial peptides: structures and mechanisms of action. biochimica et biophysica acta (bba)-biomembranes 1758 (2006) 11841202. [14] g. trinquier, y.-h. sanejouand. which effective property of amino acids is best preserved by the genetic code? protein engineering 11 (1998) 153-169. [15] w.-m. yau, w.c. wimley, k. gawrisch, s.h. white. the preference of tryptophan for membrane interfaces. biochemistry 37 (1998) 14713-14718. [16] m.-l. jobin, m. blanchet, s. henry, s. chaignepain, c. manigand, s. castano, s. lecomte, f. burlina, s. sagan, i.d. alves. the role of tryptophans on the cellular uptake and membrane interaction of arginine-rich cell-penetrating peptides. biochimica et biophysica acta (bba)-biomembranes 1848 (2015) 593-602. [17] s. winiwarter, n.m. bonham, f. ax, a. hallberg, h. lennernäs, a. karlén. correlation of human jejunal permeability (in vivo) of drugs with experimentally and theoretically derived parameters. a multivariate data analysis approach. journal of medicinal chemistry 41 (1998) 4939-4949. ©2020 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) https://gco.iarc.fr/today/data/factsheets/populations/900-world-fact-sheets.pdf http://creativecommons.org/licenses/by/3.0/ supercritical fluid technology for solubilization of poorly water soluble drugs via microand naonosized particle generation doi: http://dx.doi.org/10.5599/admet.811 355 admet & dmpk 8(4) (2020) 355-374; doi: http://dx.doi.org/10.5599/admet.811 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index review supercritical fluid technology for solubilization of poorly water soluble drugs via microand naonosized particle generation shashi kiran misra 1 , kamla pathak 2 * 1 university institute of pharmacy, csjm university, kanpur, 208026, india, email: shashisarthak@gmail.com 2 faculty of pharmacy, uttar pradesh university of medical sciences, saifai, etawah 206130, india corresponding author e-mail: kamlapathak5@gmail.com received: march 15, 2020; revised: june 18, 2020; published: june 29, 2016 abstract approximately two-third of the compounds in the pharmaceutical industry were developed through combinatorial chemistry and high throughput screening of particulate solids. poor solubility and bioavailability of these pharmaceuticals are challenging attributes confronted by a formulator during product development. hence, substantial efforts have been directed into the research on particle generation techniques. although the conventional methods, such as crushing or milling and crystallization or precipitation, are still used; supercritical fluid technology introduced in the mid-1980s presents a new method of particle generation. supercritical fluid processes not only produce microand nanoparticles with a narrow size distribution, they are also employed for the microencapsulation, cocrystallization, and surface coating with polymer. recognized as a green technology, it has emerged as successful variants chiefly as rapid expansion of supercritical solutions (ress), supercritical anti-solvent (sas) and particles from gas saturated solution (pgss) depending upon type of solvent, solute, antisolvent and nebulization techniques. being economical and eco-friendly, supercritical fluid technolgy has garnered considerable interest both in academia and industry for modification of physicochemical properties such as particle size, shape, density and ultimately solubility. the current manuscript is a comprehensive update on different supercritical fluid processes used for particle generation with the purpose of solubility enhancement of drugs and hence bioavailability. ©2020 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords supercritical fluid techniques; micronization; nanonization; solubility; bioavailability; composite particles introduction in the last decade, supercritical fluid (scf) processes have been extensively utilized for pursuing chemical reactions, extraction, crystallization, precipitation, purification, and development of microand nanoparticles. considerable research efforts are being made to modify solubility and improvement of bioavailability of poorly soluble drugs through scf. as a well known fact the bioavailability of drugs depends on the absorption through gastrointestinal tract that is in turn governed by their solubility and dissolution. thus the particle size is of utmost importance. conventional methods such as crushing, milling, crystallization and precipitation are commonly utilized for particle development in the pharmaceutical industry. each of these methods has their own set of limitations. scf technology presents an innovative http://dx.doi.org/10.5599/admet.811 http://dx.doi.org/10.5599/admet.811 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:shashisarthak@gmail.com mailto:kamlapathak5@gmail.com http://creativecommons.org/licenses/by/4.0/ s.k. misra and k. pathak admet & dmpk 8(4) (2020) 355-374 356 approach for particle formation that evades most of the downsides related to the traditional methods [1]. hence, the technology has firmly positioned itself in the pharmaceutical arena. scf finds a vital application in the development of dry powder inhalers (mean particle size of 2 to 5 µm) that accurately deliver precise dose to the lungs. furthermore, scf technology can be explored for the development of sustained and controlled release systems via microencapsulation [2,3], coating and formation of composite particles [4]. this technology is eco-friendly, green process that generates less waste during operation and produces fine product at minimum cost. the accuracy presented by scf processes permits micronization of drug particles, often to submicron stage. the scf processes can generate nanosuspensions of particles to the tune of 5-2,000 nm [5]. the technology acts as a re‐precipitation aid for rapid and uniform nucleation of the solute in all its variants of fine particle formation. the performance efficiency of this technology is based on proper solvent selection and by adjusting critical parameters (temperature and pressure) during operations (fig. 1). as documented in the literature, rapid expansion of supercritical solutions (ress), supercritical anti solvent (sas) and particles from gas saturated solutions (pgss) are the frequently employed methods not only for the fabrication of monodisperse fine powders, but also to control crystal polymorphism. it is also established that ress can be used for co2 soluble molecules, while sas can process non-soluble molecules. however, the selection is not so simple and a good knowledge of operating conditions and phase equilibrium thermodynamics is required. scf methods not only produce microand nanoparticles of uniform size distribution, but are extensively applied for microencapsulation and polymeric surface coating on drug crystals, cocrystallization with excipients and development of soluble complexes with cyclodextrin. nijlen et al. [6] demonstrated significant reduction in particle size of artemisinin and improved dissolution rate when processed with scf technology. table 1 lists various scf techniques that can be used for solubility enhancement based on particle generation. scf processes find vital applications in almost all drug delivery routes, such as oral, intravenous, ophthalmic, pulmonary, transdermal, and implants. revercheon et al. [7] emphasized the construction of various nanostructures i.e. nanofibers, nanotubes, nanowires, nanoparticles and other nano-constructions using supercritical fluid-based techniques. byrappa et al. [8] explored the adaptive properties of scf for the synthesis of advanced nanomaterials including carbon nanotubes, fullerenes, magnetic particles, quantum dots, phosphors, nanocomposites (peptide/hydroxyapatite), and gold nanoshells for drug delivery and other biomedical applications such as imaging, sensing, and cancer theranostics. the review elaborates the applications of rapid expansion of supercritical solutions (ress), supercritical anti solvent (sas) and particles from gas saturated solutions (pgss) for the generation of microand nano sized drug particles and composite particles. the write up also entails other scf technologies that have been used for solubility enhancement of poorly water soluble drugs. figure 1. scheme of the scf technology process admet & dmpk 8(4) (2020) 355-374 scf technology for solubility enhancement doi: http://dx.doi.org/10.5599/admet.811 357 table 1. protagonists of supercritical fluid technology processing component process/ acronym solvent rapid expansion of supercritical solution (ress) rapid expansion of supercritical solution into a liquid solvent ( resolv) rapid expansion of supercritical solution into an aqueous solution (ress-as) rapid expansion of supercritical solution with a non-solvent (ress-n) anti-solvent gas anti-solvent (gas) supercritical anti-solvent (sas) aerosol solvent extraction system (ases) particles by compressed anti-solvent (pca) solution enhanced dispersion by supercritical fluids (seds) solute particles from gas-saturated solutions (pgss) depressurization of an expanded liquid organic solution (delos) ress for particle generation ress is composed of two steps, first to dissolve the solid compound in a supercritical fluid, and the second step results in the formation of particles by the virtue of supersaturation. the supercritical fluid (co2) is allowed to pump at required pressure and temperature to the solid substance contained in the extraction chamber. the supercritical fluid expands adiabatically in the vessel, triggering a downfall of temperature and pressure, leading to the formation of fine particles [9,10]. hasty expansion of the supercritical solution causes reduction in the density and hence particle precipitation with minimum residual solvent occurs (fig. 2). figure 2. ress process depicted diagrammatically for solubility enhancement of a hypothetical drug charoenchaitrakool et al. [11] aimed at micronization of racemic ibuprofen and examined the dissolution rate of the micronized product in a buffered solution. the solubility data at various temperatures was modeled using peng-robinson equations of state with vander waals mixing rules. the solubility of (s)-ibuprofen exhibited solubility in co2 similar to (r)-form. the aggregated particles were easily dispersed by ultrasonication in water. the degree of crystallinity was slightly decreased and the intrinsic dissolution rate was higher than the original form. likewise the amorphous nanoparticles of cefuroxime axetil were produced directly by ress technology, without any additive. the nanoparticles obtained were between 158 and 513 nm. more than 90 % of the nanoparticles dissolved in 3 min and complete dissolution occurred within 20 min, while the commercial drug achieved only about 50 % dissolution in 60 min [12]. bioavailability of the pharmaceutical substances is very important for their activity. in case of necessity, bioavailability can be improved by reducing the particle size of the drugs. hezave and esmaeilzadeh [13] aimed to manufacture fine particles of diclofenac and optimize the ress conditions for generating uniform particles. micronization resulted in the average particle size between 10.92 and 1.33 μm. the morphology http://dx.doi.org/10.5599/admet.811 s.k. misra and k. pathak admet & dmpk 8(4) (2020) 355-374 358 of the processed particles changed to quasi-spherical while the virgin particles of diclofenac were irregular. similarly, the reduction in particle size of digitoxin was achieved by atila et al., [14] and response surface method was used to optimize the process parameters. the particle size of digitoxin was decreased from 0.2-8 µm to 68-458 nm by ress technique and 97 % of the particles were below 200 nm depending on the different experimental conditions. the solubility of the drug substance in supercritical co2 has a major effect on the average diameter of the particles prepared by ress process. this was proved by kim et al. when they used ress for the preparation of ultra-fine drug particles using supercritical co2, with no organic solvent. three drug substances (lidocaine, griseofulvin, benzoic acid) with different solubility in supercritical co2 were used, and orifice disks and capillary tubes were adapted as an expansion device. the solubilities of drug substances in supercritical co2 and the effects of various operating parameters on the characteristics of the particles prepared by ress process were experimentally investigated. the solubility of the drug substance in supercritical co2 had a major effect on the average diameter of the particle prepared by ress process, and the particle diameter decreased with the solubility for all the drugs and operating conditions [15]. in another report, the particle size of raloxifene was reduced from 45 μm to 19 nm by ress process, the smallest of which was 18.93 nm. in addition, dissolution rate study indicated that a 7-fold increase in dissolution rate could be obtained by particle size reduction of raloxifene using ress. response surface methodology was used for the optimization of the results and it was demonstrated that the smallest particle size could be achieved at a temperature of 50 °c, pressure of 17.7 mpa and a spray distance of 10 cm [16]. in addition to particle size reduction, ress can also be used to achieve microencapsulation and surface coating of an active substance particle with a polymer or co-crystallization with excipients or host molecules like cyclodextrins. kim et al. [17] used ress process to produce polymeric microparticles loaded with naproxen for drug delivery applications. poly (l‐lactic acid) (l‐pla), naproxen, and a mixture of naproxen and l‐pla were dissolved in supercritical co2 and precipitated by the ress process. composite particles appeared as a naproxen core encapsulated in a polymer coating. mishima et al. [18] reported a new method — rapid expansion from supercritical solution with a non-solvent (ress-n) for forming polymer microparticles containing proteins such as lysozyme and lipase. a suspension of protein in co2 was made that contained a cosolvent and a dissolved polymer amongst poly (ethylene glycols), poly (methyl methacrylate), poly (l-lactic acid), poly (dl-lactide-co-glycolide) and peg–poly(propylene glycol) (ppg)–peg triblock copolymer. the solubilities of these polymers in co2 increased significantly with low-molecularweight alcohols as cosolvents. the wide applications of ress namely micronization and nanonization of apis and generation of composite particles is further summarized in table 2. sas process for particle generation the sas process is a highly useful for the micronization and nanonization of synthetic drugs and natural compounds. this method refers to the precipitation of compounds in a provided supercritical fluid. the selected supercritical fluid should be essentially miscible with solvent whereas the solute should be insoluble in supercritical anti-solvents. to process sas, selection of solvent depends upon two types of requisites, first is good miscibility with co2 i.e. ethanol, acetone, toluene; and second is the solubility of solute to be precipitated. indeed, the solvent must usually belong to class 3 (non-toxic) of the pharmaceutical guidelines. in any case, the amount of residual solvent in the crystallized powder must not exceed 5000 ppm. admet & dmpk 8(4) (2020) 355-374 scf technology for solubility enhancement doi: http://dx.doi.org/10.5599/admet.811 359 table 2 a cross section of micronization and nanonization of some drugs using ress technology api objective research highlight ref coumarin coumarin nanoparticles were prepared and the effect of temperature, pressure, spray distance and nozzle diameter on particle size and solubility was assessed. a considerable decline in particle size was observed from 40.35 μm to 21.37 nm thereby affecting solubility. quadrupolar cubic plus association theory and perturbedchain polar statistical associating fluid theory were applied to interpret the solubility data. [19] ibuprofen, aspirin and griseofulvin ress was used to reduce particle size. the solubility study of poorly soluble drugs was performed employing five cubic equation of state (eos) with two mixing rules. based on the calculated solubilities, two of the most accurate equations of state are pr-vdw and pr-km with less absolute percent deviation than the other eos for all systems [20] ipriflavone the solubility of ipriflavone was enhanced through the ress process using supercritical co2. results outlined improved solubility of ipriflavone in supercritical co2. additionally, the particle size was reduced to 4.4 μm from the original 30.9 μm utilizing ress process. [21] letrozole ress with solid cosolvent (resssc) was employed to precipitate nanoparticles of letrozole obtained findings suggested enhanced solubility of letrozole (7.1 times) in the ternary phase with solid cosolvent application in ress process. the average particle size was reduced 30 nm to 19 nm. [22] aprepitant effect of parameters i.e. pressure, temperature, spraying distance and nozzle diameter was studied on the nanoparticles morphology. significant reduction in the particle size (micrometer to nanometer) was observed for the nanoparticles developed through ress-sc method. the dissolution rate of aprepitant was increased by 8.2 times, suggesting improved solubility of the drug. [23] tbtpp (5, 10, 15,20-tetrakis (3, 5-bistri fluoro methyl phenyl porphyrin ress process was investigated applying numerical modeling for particle formation of tbtpp. peng-robinson eos with kwak-mansoori mixing rules were applied after the optimization of pressure and temperature. the improved solubility was measured by numerical modeling and the results were compared with experimental data. [24] progesterone fine progesterone particles were produced and the solubility was analyzed by varying temperature and pressure and compared with a well known model. the solubility studies were correlated with empirical density-based models and the peng-robinson equation of state model. it was found to be improved range of 5.3 × 10 − 5 –8.9 × 10 − 4 , with submicron size. [25] paracetamol a novel fluidized-bed coating process using ress was described for the coating of fine particles. microspheroidal catalyst particles (average particle size 56 μm) were used as the core particles. supercritical co2 solution of paraffin was expanded through the nozzle into the bed that was fluidized by air. the coating mass and coating rates were measured by a sampling method. a stable coating of fine particles was achieved without the formation of agglomerates at room temperature [26] ibuprofen and nicotinamide ress was used as a means of simultaneous micronization and cocrystallization of drugs with poor aqueous solubility. 1:1 cocrystals of ibuprofen and nicotinamide with high purity were produced. the specific surface area of ress cocrystals was increased by almost tenfold in comparison to cocrystals produced by slow solvent evaporation and the mean dissolution time of ibuprofen from ress cocrystals was decreased. for drugs with dissolution limited bioavailability, ress cocrystallization may be a superior approach in comparison to established cocrystallization techniques. [27] many pharmaceuticals have been processed using sas and derived processes (seds, pca, ases). a very broad range of molecules can be used namely antibiotics, proteins, biopolymers, paracetamol, salbutamol, naproxen, ascorbic acid etc [28]. kordikowski et al. [29] worked with sulfathiazole, a compound that http://dx.doi.org/10.5599/admet.811 s.k. misra and k. pathak admet & dmpk 8(4) (2020) 355-374 360 exhibits five different polymorphs. using a semi-continuous sas process with methanol and co2, the researchers were able to produce pure polymorph by controlling the flow rate of methanol and the temperature. three pure polymorphs i, iii and iv could be obtained by choosing the right temperature while the flow rate, and the ratio of methanol: co2 had less influence on the polymorphs. the method is widely studied today because of its potential industrial applications. a semi-continuous sas precipitation has been used to produce rifampicin microand nanoparticles with controlled particle size and particle size distribution; using different liquid solvents. the best micronization results were obtained using dimethyl sulfoxide at 400 °c. the nanoparticles with mean diameter ranging from 0.4 to 1 µm were obtained at a pressure of ≥ 120 bars, while microparticles with mean diameter ranging from 2.5 to 5 µm were obtained at pressures between 90 and 110 bars. the morphology of rifampicin precipitates was different too. nanoparticles connected in small aggregates were obtained at pressures higher than 120 bars, whereas, spherical single microparticles were obtained at lower operating pressures [30]. in a research work, a swirl mixer was employed to produce the micronized curcumin with polyvinyl pyrrolidone (pvp) by the sas process to improve the bioavailability of curcumin. the effects of operating parameters such as curcumin: pvp ratio, feed concentration, temperature, pressure, and co2 flow rate were investigated. the result showed that the optimal condition for the production of curcumin-pvp particles were at curcumin:pvp ratio of 1:30, feed concentration of 5 mg/ml, temperature of 40 °c, pressure of 15 mpa, and co2 flow rate of 15 ml/min. curcumin-pvp particles (< 150 nm) were completely soluble in aqueous solution to form a clear yellow solution unlike poorly soluble raw curcumin. the solubility of curcumin-pvp particles was 2.34 µg/ml whereas that of raw curcumin was 0.006 µg/ml after 12 h. the reason attributable was that the water-soluble polymer (pvp) can modify the surface properties of the particle and thus enhance the solubility of curcumin in aqueous solution [31]. sas process was used for telmisartan (bcs class ii drug) in a variety of ways including micronization, amorphization and solid dispersion. solid dispersions were prepared using hpmc and pvp at 1:0.5, 1:1, and 1:2 weight ratios of drug to polymer, and pure telmisartan was also treated using the sas process. after the sas process, all samples were converted to the amorphous form and were hundreds nm in size. solubility and dissolution rate were increased compared to the raw material. though the drug’s solubility increased with increase in the amount of polymer used; the dissolution rate decreased with increasing polymer concentration. processed pure telmisartan showed higher drug release than its original form, even though it had lower solubility compared to other solid dispersions. the authors concluded that that after controlling the formulation of solid dispersion, the sas process could be a promising approach for improving the solubility and dissolution rate of telmisartan [32]. the sas process was used to modify the solid state characteristics of fluconazole by preparing its polymorphs by varying the temperature, pressure and solvents. fluconazole anhydrate form i was obtained at low temperature (40 °c) and anhydrate form ii was obtained at high temperature (80 °c). not much difference was found in solubility of the polymorphs [33]. the same research group improved dissolution rate of poorly water-soluble drug, cilostazol, using sas process. in particular, the mean particle size and distribution were markedly influenced by drug solution concentration during sas process. moreover, the drug did not change its crystal form and the operating parameters probably controlled the 'crystal texture'. the micronized particles exhibited a 6.5 fold increase in dissolution rate compared to the unprocessed cilostazol [34]. likewise, amorphous sas treated nanoparticles of atorvastatin calcium showed increased bioavailability of atorvastatin owing to their nano-dimensional size that offered high admet & dmpk 8(4) (2020) 355-374 scf technology for solubility enhancement doi: http://dx.doi.org/10.5599/admet.811 361 solubility and increased dissolution rate. the oral absorption of amorphous atorvastatin calcium nanoparticles in rats was obviously higher when compared with the crystalline atorvastatin calcium after a single dose of 25 mg/kg. the auc0-12h of the amorphous atorvastatin calcium nanoparticles (179 nm) was 2.1 times that of unprocessed drug. thus, the sas process for generation of amorphous atorvastatin calcium nanoparticles is a promising method for enhancing their bioavailability [35]. the sas process can also be utilized to establish co-precipitation of two different compounds and to form beta cyclodextrin complexes of poorly soluble drugs. the inclusion complex of apigeninhydroxypropyl -beta-cyclodextrin was prepared through a sas process and its bioavailability was evaluated. the inclusion complex exhibited improved wettability and the dissolution of apigenin inclusion complex was significantly enhanced. the inclusion complex demonstrated an enhanced oral bioavailability of approximately 6 fold when compared to pure apigenin, in rats [36]. table 3 summarizes few additional applications of sas process in improving solubility, dissolution or drug release of certain poorly soluble pharmaceuticals. table 3. a compilation of reports on solubility enhancement of poorly soluble drugs affected by sas technology api objective outcome ref tolfenamic acid sas parameters were evaluated for solid state property modification and improvement of dissolution profile of tolfenemic acid. sas technology was efficient in modifying the solidstate. it produced microparticles with improved dissolution behavior. [37] n-acetyl-cysteine the study aimed to micronize nacetylcysteine by the anti-solvent seds technique. micronized n-acetylcysteine presented prominent biological activity (100 times) depicted by lower minimum inhibitory concentration compared to non-micronized n-acetylcysteine [38] curcumin curcumin based dye extract was developed employing sas. eudragit® l100, pluronic® 127 and tween 20 were added to improve the aqueous solubility and stability. formulation of a soluble curcumin was carried out for food application. highest aqueous stability and solubility was observed at ph 4. the mean diameter and zeta potential of the amorphous curcumin particles was 5667.4 nm and 11.21 mv respectively. [39] warfarin to determine solubility of warfarin in supercritical co2 using sas regular crystals of warfarin with a mean particle size of 6.6 μm were produced [40] irbesartan to improve the dissolution of irbesartan through solid dispersions using sas concept. the crystalline state of the drug was transformed into the amorphous state. the dissolution was enhanced after formation of irbesartan solid dispersions [41] azithromycin solid dispersions of azithromycin were developed utilizing variable amounts of peg 6000, sorbitol, sls and poloxamer 188, the amorphous solid dispersions of azithromycin demonstrated enhanced solubility with peg 6000 and sls. [42] the sas technology offers reasonable morphology, high drug loading and improved bioavailability but occasionally leads to particle aggregation that can be resolved by use of ultrasonic processor [43]. among the various sas based micronization techniques reported, the seds (solution enhanced dispersion by supercritical fluids) is an efficient process for the development of uniform sized nanoparticles [44]. york and hanna developed modified sas or seds process in 1996. this revised technique curtails the drawbacks and limitations of sas and produces completely dried, uniform, narrow sized particles with low residual organic solvent [45]. in this modified sas system, initially the drug and excipients are allowed to dissolve in an appropriate organic solvent followed by rigorous agitation. thereafter, the blended components come in contact with the supercritical fluid. especially designed coaxial nozzle in seds process efficiently sprays http://dx.doi.org/10.5599/admet.811 s.k. misra and k. pathak admet & dmpk 8(4) (2020) 355-374 362 the components at high pressure that results in micronized uniform particles (fig. 3). seds has been employed to enhance dissolution of baicalein via micronization of baicalein. seds reduced the particle size from 25.335 µm to 0.614 µm and the particle morphology was transformed to flakes in comparison to original rod shaped crystals. the in vitro release studies demonstrated 50 % of the drug release in approximately 20 min, and over 80 % of the baicalein was released in 60 min. however, the original powder exhibited a much slower dissolution rate, and less than 25 % of the baicalein was released in 60 min. the faster dissolution rate of baicalein microparticles was attributed to the reduced particle size of the baicalein and the extremely large specific surface area [46]. likewise, the micronization of resveratrol via seds enhanced its solubility by approximately 2.8 times and the dissolution rate by 1.8 fold. the antioxidant efficacy of the resultant product was also enhanced significantly [47]. seds has been investigated to enhance solubility of bixin [48] and the dissolution rate of aescin by 5.5. fold [49]. figure 3. operational design of seds for the formation of micronized particles the process can also be utilized for development of solid dispersion of poorly water soluble drugs. water-soluble pvp and astaxanthin nanocoprecipitates were successfully prepared by seds precipitation. it was found that the operating pressure, temperature, pvp:astaxanthin ratio, and z-isomer content of astaxanthin affected the particle size and the astaxanthin content in the coprecipitates. the researchers selectively used z-isomers of astaxanthin as it has higher bioavailability and antioxidant capacity, than the e-isomer. the authors predicted that use of pvp-z-isomers of astaxanthin coprecipitates would improve its functionality [50]. lee et al. [51] investigated the application of seds to improve aqueous solubility of and rographolide through particle engineering. the sticky crude andrographis paniculata extract was precipitated into powder from co2-acetone system and co2-acetone:ethanol (1:1 v/v) system. the modification of aqueous solubility of andrographolide was then attempted by manipulating its precipitation process. a. paniculata powder precipitated from co2-acetone system at 150 bar and 40 °c consisted of large, irregularly shaped, less crystalline particles with the highest andrographolide aqueous solubility (two fold increment compared to crude extract). complete dissolution of andrographolide from a. paniculata powder admet & dmpk 8(4) (2020) 355-374 scf technology for solubility enhancement doi: http://dx.doi.org/10.5599/admet.811 363 precipitated from co2-acetone system was achieved within 90 min. based on the higher aqueous solubility and dissolution of andrographolide, and different morphology observed from the less crystalline a. paniculata powder precipitated from co2-acetone system, it was concluded that fewer impurities could have co-precipitated with andrographolide. conclusively, the seds process offers many advantages over conventional sas technique such as requirement of less solvent, applicability for thermosensitive materials and less concentration of residual solvent in the product. pgss for particle generation pgss is a favorable technique for polymeric encapsulation of drugs, proteins and peptides as fine particles without employing organic solvents. in this process, carriers such as polymers are melted with the dissolved or suspended active pharmaceutical ingredient (api) contained within them (fig. 4). the resulting product is then equilibrated with supercritical-co2 and expanded through a nozzle in an expansion chamber in order to form fine and porous composite particles [52]. several carriers and apis have been micronized using a pgss process [53,54]. for example, ibuprofen has been successfully micronized with different carrier materials such as polyethylene glycol (peg) 6000 [55], poloxamers, gelucire1 and glyceryl monostearate [56]. peg 4000 has also been used as a carrier for the micronization of poorly water-soluble drugs [57]. another particle generation process close to pgss has been described by sievers et al. [58] that consists of production of a dense fine droplet aerosol plume followed by a drying step. the aim of this process was to obtain fine particles usable in a dry powder inhaler form. this patented process has been used with lactose, for developing dry powder inhaler of anti asthamatic drugs: albuterol sulfate and cromolyn sodium. the product comprised of fine spherical particles in the range of 0.1– 3 µm making them suitable for inhalation. figure 4. schematic illustration of working of pgss (particles from gas saturated solution) in the experiments carried out in a pre-expansion pressure range from 100-200 bar, the particle size of nifedipine was reduced from 50 to 15 µm. at higher pressures smaller particles were formed. with the http://dx.doi.org/10.5599/admet.811 s.k. misra and k. pathak admet & dmpk 8(4) (2020) 355-374 364 particle size reduction the dissolution rate increased to some extent, but the anticipated effective surface area was probably reduced by the drug's hydrophobicity and agglomeration of the particles during micronization [59]. pestieau et al. [60] optimized pgss process for a fenofibrate lipid-based solid dispersion formulation. the researchers aimed to develop a formulation containing fenofibrate and gelucire1 50/13 (gattefossé, france) in order to improve the oral bioavailability of the drug. the pgss process was optimized according to the in vitro drug dissolution profile obtained using a biphasic dissolution test. based on the fact that the propensity of drug precipitation during in vitro testing of lipid-based formulations can serve as a potential indicator for the in vivo performance of a drug, the authors predicted less time to reach cmax and its sustainment for longer period for pgss derived solid dispersion than conventional solid dispersions. the authors also deduced that these formulations may avoid the precipitation of this poorly watersoluble drug in vivo. furthermore, an increase in apparent solubility induced by the formulation used can lead to an enhancement of a drug's permeability through biological membranes. thus, improvement of the oral bioavailability of fenofibrate should be more pronounced with the pgss formulation as a result of supersaturation being maintained for longer period. fenofibrate solid dispersions were also investigated by krananja et al. [61]. the authors used pgss process for the carrier materials: brij 5100 and polyethylene glycol peg 4000, for the incorporation of the insoluble drugs nimodipine, fenofibrate, and o-vanillin with the purpose of improving their bioavailability and dissolution rate. the authors however, reported primarily the influence of processing parameters of pgss on process yield, particle size distribution, loading efficiency and dissolution rates. the general conclusion was that with increasing pre-expansion pressure, the mean particle size of nimodipine-brij s100, vanillin-brij s100, and vanillin-peg 4000 solid dispersion(s) decreased. in the case of a mixture of fenofibrate-brij s100, the anticipated effective surface area was slightly reduced with pressure as a result of agglomeration. the loading efficiency of drugs in carriers was high and the particles obtained were irregular in shape. the authors deduced that a combination of factors, including particle size reduction and interaction between drug(s) and hydrophilic carrier(s), contributed to enhancing the dissolution rates of precipitated solid particles. on average, a 3.5-fold greater amount of nimodipine was dissolved in 1 h from solid dispersions, compared to unprocessed nimodipine. dissolution profiles were compared with a different f1 factor and a similarity factor f2. it was confirmed that the dissolution character of processed ovanillin and fenofibrate by pgss was different from that of unprocessed o-vanillin and fenofibrate. the pgss process is not limited to production of solid dispersions; the technique also finds applications in production of a variety of composite particles namely solid lipid particles, microparticles and microcapsules that have potential to modulate drug release and thereby bioavailability. table 4 presents a cross section of such research reports. pgss process is simpler in operation than other techniques as the therapeutic substance need not be necessarily solubilized in the supercritical fluid (co2). it requires low solvent gas supplies and pressure for operative purposes as compared to other processes. pgss process can be employed to develop inorganic powders into multifunctional pharmaceutical compounds. however, precautions need to be taken while processing thermolabile substances. admet & dmpk 8(4) (2020) 355-374 scf technology for solubility enhancement doi: http://dx.doi.org/10.5599/admet.811 365 table 4. a cross section of composite particles of drugs prepared using pgss process api and excipients strategy result highlight (s) ref curcumin, tristearin, soyphosphatidyl-holine, dmso curcumin embedded solid lipid particles were developed via pgss with less quantity of organic solvent. the use of helium in the process of pgss, for designing of lipid mixture exhibited improvement in the biopharmaceutical properties and therapeutic value of curcumin. [62] s-(+)-ibuprofen, poloxamers, gelucire, glyceryl monostearate(gms) pgss was employed for the enhancement of solubility of (+)-ibuprofen using hydrophilic or hydrophobic carrier. spherical and porous particles (50-200 µm) with 90 % encapsulation efficiency were produced. the solubility of ibuprofen was significantly enhanced with poloxamer in the gastro-intestinal fluids; gelucire and gms did not enhance the solubility of ibuprofen. [63] ketoprofen, glyceryl monooleate (gmo), gelucire 43/01™, geleol™ and gelucire 50/13™ for production of structured lipid carriers a liquid glycerolipid (gmo), was incorporated into three solid glycerolipids with hydrophilic-lipophilic balance ranging from 1 to 13 and compared with solid lipid particles. irregular porous microparticles with a wide particle size distribution were obtained. the drug loading capacity of the structured lipid carriers increased as the gmo content in the particles increased, achieving a maximum encapsulation efficiency of 97 % for the 3:1 mass ratio. the structured lipid carriers presented an immediate release of ketoprofen from its matrix with higher permeation through a mucous-membrane model, while solid lipid particles presented controlled release of the drug with less permeation capacity. [64] quercetin, soybean, lecithin, and pluronic l64 to modify bioavailability of quercetin through microencapsulation more homogenous lyophilized less crystalline encapsulated quercetin particles were reported with enhanced bioavailability [65] β-carotene, poly-(εcaprolactone) βcarotene was encapsulated in poly caprolactones, and precipitated out using pgss process. small, regular, uniform, microencapsulated β-carotene particles in the size range of 100 –600 μm were obtained, that demonstrated enhanced solubility. [66] 1,3-diphenyl-2propenone (chalcone) microparticles of chalcone alone and with lipid carriers were developed via pgss and the solubility was analyzed. the lipid carriers influenced the solubility of trans-chalcone in simulated gastric and intestinal fluids, without addition of enzymes. [67] omega-3 pufa-rich salmon oil and astaxanthin microparticles of omega-3 pufa-rich salmon oil in peg6000 were developed through pgss concept. developed microparticles showed significant thermogravimetric stability up to 350 °c. moreover, in vitro release of oil in fluids stimulating gastric conditions was faster than in distilled water. [68] other supercritical fluid techniques for enhancement of solubility rapid expansion of a supercritical solution into a liquid solvent (resolv) when the traditional ress is modified by expanding the supercritical solution into a liquid solvent instead of ambient air it is termed as rapid expansion of a supercritical solution into a liquid solvent (resolv). this technique can be used for the production of nanoscale particles (less than 50 nm in average diameter) from a co2 soluble polymer [69]. the development of versatile methods for the preparation of homogeneously distributed nanoscale drug particles and their stable aqueous suspension is still a major challenge, despite the extensive effort based on traditional techniques. resolv was employed for the production of drug nanoparticles of two model drugs: ibuprofen and naproxen, which are somewhat soluble in supercritical co2 and practically insoluble in water. the resolv process yielded aqueous suspensions of homogeneously distributed ibuprofen (average size of 40 nm in diameter and a size distribution standard deviation of 8.5 nm) and naproxen nanoparticles of average size of 64 nm in diameter and a size distribution standard deviation of 10 nm. the nanoparticles agglomerated http://dx.doi.org/10.5599/admet.811 s.k. misra and k. pathak admet & dmpk 8(4) (2020) 355-374 366 to form larger aggregates on a longer time scale. the agglomeration can be minimized by the presence of a stabilization agent, e. g. poly (n-vinyl-2-pyrrolidone) in the aqueous suspension. the technique may serve as a “clean” way for nanosizing the drug particles and the preparation of stable suspensions thereof for formulation and other delivery related requirements, specifically addressing the bioavailability issues [70]. depressurization of an expanded liquid organic solution (delos) polymorphism is very common in pharmaceutical drug substances since their solubility and bioavailability are determined by the crystalline structure adopted by the solid drugs [71]. in addition, the drug substance will in most cases be handled as a solid in some stages of the manufacturing process, and its handling and stability properties may depend on the solid phase. consequently, the control of the production of a given solid polymorph is of the utmost importance in such commercial applications and industries. in a study by sala et al. [72] it was observed that the precipitation of pure monoclinic e form of stearic acid was favored by delos process, a kinetically controlled crystallization in which high supersaturation levels are rapidly achieved. remarkably, the delos process for the first time provided a pure polymorphic monoclinic e form, without the presence of traces from other polymorphs that always are present when using other kinetically driven methods, such as the conventional fast cooling. on the contrary, the c polymorph was preferentially obtained by the thermodynamically controlled gas technique in which the increase of the solution supersaturation is slow or low supersaturation levels are attained. delos can be used as a route to obtain nutraceutical products that might show enhanced functional properties. phytosterols are absorbed to a much smaller extent in the body compared to cholesterol and they interfere with the intestinal absorption of cholesterol. delos methodology has the potential to process phytosterols into micrometer or submicrometer particles, which is not possible with conventional technologies. moreno-calvo et al. [73] processed β-sitosterol through delos thereby reducing the crystals from 188 μm to <6.5 μm, with a narrow size distribution at all processing conditions employed. the new phase showed higher chemical purity and higher crystallinity than the native mixture. the authors recommend further studies to confirm the expected enhancement of absorption and bioavailability of βsitosterol. aerosol solvent extraction system (ases) ases methodology has been used extensively for processing pharmaceuticals and biopolymers and is capable of producing micrometer sized or nanosized particles with low levels of residual solvent [74]. furthermore, in the ases process the particle size and morphology can be easily modulated by the optimization of the processing parameters. the therapeutic applications of silybin, an antihepatotoxic polyphenolic substance, are strongly limited by its poor solubility and low bioavailability. the issues can be addressed by designing nanodrug via ases. in the process, water soluble and biocompatible pvp was used to improve the dissolution rate and bioavailability of silybin. the size of the silybin pvp nanodrug was to the tune of 100 to 300 nm. compared with raw silybin, the nanodrug had low crystallinity and hence showed solubility enhancement by more than 8-fold and hence in dissolution too [74]. copper-indomethacin is a non-steroidal anti-inflammatory drug currently available for veterinary use. its application is limited to oral formulations because of its poor solubility in biocompatible solvents. meure et al. [75] prepared microspheres of pvp and copper-indomethacin that ranged in size from 50 nm to 4 μm. a coprecipitate containing 10 %wt copper-indomethacin and 90 %wt pvp was found to be at least 93 times more soluble in ethanol than factory-grade copper-indomethacin. the significance of these results is that the coprecipitate of copper-indomethacin may be used for parenteral applications [75]. admet & dmpk 8(4) (2020) 355-374 scf technology for solubility enhancement doi: http://dx.doi.org/10.5599/admet.811 367 rao et al. [76] suggested enhancement of the dissolution rate, apparent solubility and oral bioavailability of tadalafil by nanosized amorphous particles prepared by using antisolvent precipitation. optimization of processing parameters yielded amorphous tadalafil solid dispersion of approximately 510 μm. the solid dispersion obtained using the optimized process conditions exhibited 8.5 times faster dissolution rates in the first minute of dissolution, 22 times greater apparent solubility at 10 min and a 3.67-fold increase in oral bioavailability than the unprocessed tadalafil. supercritical solvent impregnation (ssi) among the different approaches that employ supercritical fluids for pharmaceutical purposes, polymer impregnation techniques have been recently used in literature to achieve the impregnation of many kinds of polymers namely, poly(lactic-co-glycolic acid), ethyl/methyl cellulose, poly(dimethylsiloxane), poly(methyl methacrylate, etc. with a drug [77]. in particular, reports where pvp was impregnated with some crystalline drugs (i.e. carbamazepine, ibuprofen, ketoprofen) can be found [78-80]. in these processes, the crystalline drug is dissolved by supercritical co2 and thus conveyed through the swollen polymeric matrix until the partition equilibrium takes place between the phases. the encapsulated particles display extensive solubility, better diffusion and substantial dissolution profile owing to the supercritical fluid that plays a role of cosolvent. in a study by banchero et al. [81] successful impregnation was reported for all the pvp k15-piroxicam systems at 300 bar and 100 °c. good results in terms of acceleration in the drug release were obtained with the pvp k15-piroxicam system. the best result was obtained for the impregnated sample containing a piroxicam amount equal to 11.3 %, which released 94.7 % of the drug after 10 min, with respect to 7.8 % released by the corresponding physical mixture after the same period of time. gas antisolvent process (gas) in the gas processes, a solute is dissolved in solvent and loaded into a crystallizer. the solution is then expanded by injecting carbon dioxide into the crystallizer. a sharp reduction of solute solubility in liquid phase is observed and subsequently particle precipitation occurs. this technique is used for drugs with low solubility in the supercritical fluid. the mean particle size and particle size distribution are controlled by gas variables. various drugs have been micronized by gas process namely, microparticle production of carbamazepine [82], controlled crystallization of β-carotene [83], caffeine [84], phenanthrene [85] and paracetamol [86]. control of gas processing variables resulted in a decrease in ampicillin particle size from 359 to 260 nm. the mean particle size was 425 nm for the lowest pressure (9 mpa). when the pressure was increased, a smaller mean particle size (220 nm) was obtained. the smaller mean particle size was observed at the lower temperature and low solute concentration [87]. all the detailed scf techniques are virtuous alternatives for micronization and nanonization of drugs that require particle engineering for modification and improvement of solubility. scf processes are frequently utilized to formulate readily solubilized drug carrier systems i.e. microparticles, nanoparticles, inclusion complexes, solid dispersions, macromolecular powders and microporous foams. some of these have been elaborated in table 5. http://dx.doi.org/10.5599/admet.811 s.k. misra and k. pathak admet & dmpk 8(4) (2020) 355-374 368 table 5. a cross-section of research reports on micronized particles using discrete scf technology scf process api purpose highlight ref delos ibuprofen and naproxen micronization and determination of solubility ibuprofen showed same solubility profile, both in co2-expanded ethanol and co2–expanded acetone mixtures; whereas the naproxen solubility was greatly dependent on the nature of the solvent i.e. high in co2–expanded ethanol. [88] resolv gambogic acid nanoparticles of gambogic acid were prepared to improve solubility. results outlined successfully preparation of nanosuspension of gambogic acid. extended solubility data was correlated with density-based models that suggested enhanced bioavailability and antineoplastic efficacy of nanosized gambogic acid. [89] delos phytosterol nanonization and decrease crystallinity of phytosterol to modify solubility phytosterol nanoparticles were formulated through fast cooling, the crystallinity of the impregnated phytosterols was found to decrease in comparison to the pure phytosterol that modified water solubility. [90] resolv poly (l-lactide) (plla) nanoparticles loaded with retinyl palmitate nanoparticles of plla retinyl palmitate were developed with pluronic f127, f68 and sodium dodecyl sulfate spherical pllaretinyl palmitate nanoparticles were prepared that possessed mean size of 40–110 nm with improved solubility and great retinyl palmitate loading. [91] resolv fenofibrate precipitation and stabilization as ultrafine particles of fenofibrate the mean particle size was approximately 3 μm, which suggested improved solubility. the particles were found to be stable for 24 h. [92] gas resveratrol and isoniazid, nicotinamide co-crystals of resveratrol were prepared with isoniazid and nicotinamide using co2 antisolvent the developed co-crystals exhibited enhanced bioavailability when compared to original resveratrol. [93] gas 5-fluorouracil halloysite clay nanocarrier was developed to obtain high drug loading of 5fluorouracil prepared nanocarrier loaded with 5-fluorouracil released significantly high drug release at ph 7.4 owing to improve solubility. [94] ssi quercetin quercetin was impregnated on silica to enhance solubility several parameters i.e. temperature, time, pressure, and different cosolvents in the supercritical impregnation process were reported influential for quercetin solubility. [95] ssi promogran promogran was embedded on a spilanthol-enriched extract to modify solubility jambu extract that is completely soluble in fluid phase, was used to demonstrate enhanced solubility of promogran. a 4 h processing period was used for complete dissolution of the extract in scf. [96] ases irbesartan development of irbesartan micro-particles and its composite microparticles results highlighted modified solubility (7.5 times) and dissolution rate of irbesatan microparticles compared to pure drug. [97] ases β-sitosterol preparation of submicroparticles of βsitosterol powdered submicroparticles of β-sitosterol exhibited improved solubility. [98] ases itraconazole preparation of solidinclusion complexes of itraconazole with hp-βcd ases-processed itr-hp-β-cd inclusion complex solid powder showed 90 % drug dissolution within 10 minutes. [99] admet & dmpk 8(4) (2020) 355-374 scf technology for solubility enhancement doi: http://dx.doi.org/10.5599/admet.811 369 conclusions supercritical fluid techniques for micronization and solubility enhancement have been progressively applied in pharmaceutical industry. the characteristic advantages of scf technology include non-toxicity, eco-friendly and flexibility that make it suitable for green chemistry. scf processes are proven promising strategies to develop and design drug delivery system of those drugs whose solubility and bioavailability is significantly low. moreover, scf technologies are also utilized for formulating drug carrier owing to unique solvent features that can be readily modified by altering operating temperature and pressure. several issues still remain i.e. the influence of operating parameters on the characteristics of the particle produced (size, morphology, polymorphism), the comprehension of the fluid dynamics, the nucleation phenomenon, the crystal growth under provided conditions etc. whatsoever, the technology has arrived and is promising green option for pharmaceutical development. conflict of interest: the author declares no conflict of interest. references [1] a. baldelli, m.a. boraey, d.s. nobes, r. vehring. analysis of the particle formation process of structured microparticles. molecular pharmaceutics 12 (2015) 2562-2573. [2] d.l. tomasko, h.li, d. liu, x. han, m.j. wingert, l. j. lee, k.w. koelling. a review of co2 applications in the processing of polymers. industrial and engineering chemical research 42 (2003) 6431-6456. [3] k. mishima. biodegradable particle formation for drug and gene delivery using supercritical fluid and dense gas. advance drug delivery reviews 60 (2008) 411-432. [4] y.c. tien, c.s. su, l.h. lien, yp. chen. recrystallization of erlotinib hydrochloride and fulvestrant using supercritical antisolvent process. the journal of supercritical fluids b (2010) 292-299. [5] t. yasuji, h. takeuchi, y. kawashima. particle design of poorly water-soluble drug substances using supercritical fluid technologies. advance drug delivery reviews 60(3) (2008) 388-398. [6] t.v. nijlen, k. brennan, g. v. mooter, n. blaton, r. kinget, p. augustijns. improvement of the dissolution rate of artemisinin by means of supercritical fluid technology and solid dispersions. international journal of pharmaceutics 254 (2003) 173-181. [7] e. reverchon, r. adami. nanomaterials and supercritical fluids. the journal of supercritical fluids 37 (2006) 1-22. [8] k. byrappa, s. ohara, t. adschiri. nanoparticles synthesis using supercritical fluid technology-towards biomedical applications. advanced drug delivery reviews 60 (2008) 299-327. [9] m. sauceau, j. fages, j.j. letourneau, d. richon. a novel apparatus for accurate measurements of solid solubilities in supercritical phases. industrial engineering chemical research 39 (2000 )4609– 4614. [10] m. sauceau, j.j. letourneau, d. richon, j. fages. solid compound solubilities in sc–co2 and cosolvent: development of enhanced density-based models. fluid phase equilibrium 208 (2003) 99– 113. [11] m. charoenchaitrakool, f. dehghani, n.r. foster, h.k. chan. micronization by rapid expansion of supercritical solutions to enhance the dissolution rates of poorly water soluble pharmaceuticals. industrial engineering chemical research 39 (2000) 4794–4802. [12] j. varshosaz, f. hassanzadeh, m. mahmoudzadeh, a. sadeghi. preparation of cefuroxime axetil nanoparticles by rapid expansion of supercritical fluid technology. powder technology 189 (2009) 97102. [13] a.z. hezave, f. esmaeilzadeh. the effects of ress parameters on the diclofenac particle size. advanced powder technology 22 (2011) 587-595. http://dx.doi.org/10.5599/admet.811 s.k. misra and k. pathak admet & dmpk 8(4) (2020) 355-374 370 [14] c. atila, n. yildiz, a. caliml. particle size design of digitoxin in supercritical fluids. the journal of supercritical fluids 51 (2010) 404-411. [15] j-t. kim, h-l. kim, c-s. ju. micronization and characterization of drug substances by ress with supercritical co2. korean journal of chemical engineering 27 (2010) 1139-1144. [16] a. keshavarz, j. karimi-sabet, a. fattahi, a.a. golzary, m. rafiee-tehrani, f.a. dorkoosh. preparation and characterization of raloxifene nanoparticles using rapid expansion of supercritical solution (ress). the journal of supercritical fluids 63 (2012) 169-179. [17] j.h. kim, t.e. paxton, d.l. tomasko. microencapsulation of naproxen using rapid expansion of supercritical solutions. biotechnology progress 12 (1996) 650– 661. [18] k. mishima, k. matsuyama, d. tanabe, s. yamauchi, t.j. young, k.p. johnston. microencapsulation of proteins by rapid expansion of supercritical solution with a non-solvent. aiche journal 46 (2000) 857– 865. [19] g. sodeifian, n.s. ardestani, s.a. sajaadian, h.s. panah, experimental measurements and thermodynamic modeling of coumarin-7 solid solubility in supercritical carbon dioxide: production of nanoparticles via ress method. fluid phase equilibria 483 (2019) 122-143. [20] h. bagheri, g.a. mansoori, h. hashemipour. a novel approach to predict drugs solubility in supercritical solvents for ress process using various cubic eos-mixing rule. journal of molecular liquids 261 (2018) 174-188. [21] b. wang, c. shaansu. solid solubility measurement of ipriflavone in supercritical carbon dioxide and microparticle production through the rapid expansion of supercritical solutions process. journal of co2 utilization 37 (2020) 285-294. [22] g. sodeifian, s.a. sajadian. solubility measurement and preparation of nanoparticles of an anticancer drug (letrozole) using rapid expansion of supercritical solutions with solid cosolvent (ress-sc). the journal of supercritical fluids 133 (2018) 239-252. [23] g.s. seyed, a.s. s. daneshyan. preparation of aprepitant nanoparticles (efficient drug for coping with the effects of cancer treatment) by rapid expansion of supercritical solution with solid cosolvent (ress-sc). the journal of supercritical fluids 140 (2018) 72-84. [24] h. bagheri, h.h. mipour, m. mirzaie. investigation on hydrodynamic and formation of nano particle by ress process: the numerical study. journal of molecular liquids 281 (2019) 490-505. [25] z. huang, y-h. guo, h. miao, l-j. teng. solubility of progesterone in supercritical carbon dioxide and its micronization through ress. powder technology 258 (2014) 66-77. [26] a. tsutsumi, s. nakamoto, t. mineo, k. a. yoshida. a novel fluidized-bed coating of fine particles by rapid expansion of supercritical fluid solution. powder technology 85 (1995) 275-278. [27] k.c. mullers, m. paisana, m.a. wahl. simultaneous formation and micronization of pharmaceutical cocrystals by rapid expansion of supercritical solutions (ress). pharmaceutical research 32 (2015) 702-713. [28] j. jung, m. perrut. particle design using supercritical fluids: literature and patent survey. the journal of supercritical fluids 20 (2001) 179 – 219. [29] a. kordikowski, b. shekunov, p. york. crystallisation of sulfathiazole polymorphs using co2, international society for the advancement of supercritical fluids 1 (2000) 117– 122. [30] e. reverchon, i. marco, g. porta. rifampicin microparticles production by supercritical antisolvent precipitation, international journal of pharmaceutics 243 (2002 )83–91. [31] k. chhouk, w.h. kanda, s-i. kawasaki, m. goto. micronization of curcumin with biodegradable polymer by supercritical anti-solvent using micro swirl mixer. frontiers of chemical science and engineering 12 (2018) 184-193. [32] j. park, w. cho, k.h. cha, j. ahn, k. han, s. j. hwang. solubilization of the poorly water soluble drug, telmisartan, using supercritical anti-solvent (sas) process. international journal of pharmaceutics 441 (2013) 1-2. https://www.sciencedirect.com/science/article/pii/s0896844618302419#! https://www.sciencedirect.com/science/article/abs/pii/003259109503021x#! https://www.sciencedirect.com/science/article/abs/pii/003259109503021x#! https://www.sciencedirect.com/science/article/abs/pii/003259109503021x#! admet & dmpk 8(4) (2020) 355-374 scf technology for solubility enhancement doi: http://dx.doi.org/10.5599/admet.811 371 [33] h.j. park, m.s. kim, s. lee, j.s. kim, j. s. woo, j.s. park, s.j. hwang. recrystallization of fluconazole using the supercritical antisolvent (sas) process. international journal of pharmaceutics 328 (2007) 152–160. [34] m.s. kim, s. lee, j.s. park, j.s. woo, s.j. hwang. micronization of cilostazol using supercritical antisolvent (sas) process: effect of process parameters. powder technology 177 (2007) 64–70. [35] m.s. kim, s.j. jin, j.s. kim, h.j. park, h.s. song, r.h. neubert, s.j. hwang. preparation, characterization and in vivo evaluation of amorphous atorvastatin calcium nanoparticles using supercritical antisolvent (sas) process. european journal of pharmaceutics and biopharmaceutics 69 (2008) 454–465. [36] y. huang, y. zu, x. zhao, m. wu, z. feng, y. deng, c. zu, l. wang. preparation of inclusion complex of apigenin-hydroxypropyl-beta-cyclodextrin by using supercritical antisolvent process for dissolution and bioavailability enhancement. international journal of pharmaceutics 511 (2016) 921–930. [37] h.h. chen, c.s. su, j.j. liu, m.t. sheu. solid-state property modification and dissolution rate enhancement of tolfenamic acid by supercritical antisolvent process. the journal of supercritical fluids 101 (2015) 17–23. [38] g.p.s. aguiar, m. marcon, r. mocelin, a.p. herrmann, l.m.p.c. chaves, a. l. piato, m. lanza, v. oliveira. micronization of n-acetylcysteine by supercritical fluid: evaluation of in vitro and in vivo biological activity. the journal of supercritical fluids 130 (2017) 282-291. [39] a.a. ruiz, a. martin. m.j. cosero, c. jimenez, j. londono. encapsulation of curcumin using supercritical antisolvent (sas) technology to improve its stability and solubility in water. food chemistry 258 (2018) 156-163. [40] j-m. ciou, b-c. wang, c-s, su, j-j. liu, m-t. sheu. measurement of solid solubility of warfarin in supercritical carbon dioxide and recrystallization study using supercritical antisolvent process. advanced powder technology 29 (2018) 479-487. [41] e. adeli. the use of supercritical anti-solvent (sas) technique for preparation of irbesartan-pluronic® f-127 nanoparticles to improve the drug dissolution. powder technology 298 (2016) 65-72. [42] e. adeli. a comparative evaluation between utilizing sas supercritical fluid technique and solvent evaporation method in preparation of azithromycin solid dispersions for dissolution rate enhancement. the journal of supercritical fluids 87 (2014) 9-21. [43] r. campardelli, r. adami, g. d. porta, e. reverchon. nanoparticle precipitation by supercritical assisted injection in a liquid antisolvent. the chemical engineering journal 192 (2012) 246–251. [44] s. li, y. zhao. preparation of zein nanoparticles by using solution-enhanced dispersion with supercritical co2 and elucidation with computational fluid dynamics. international journal of nanomedicine 12 (2017) 3485–3494. [45] r.k. kankala, b-q. chen, c.g. liu, h-x. tang, s-b. wang, a-z. chen. solution-enhanced dispersion by supercritical fluids: an ecofriendly nanonization approach for processing biomaterials and pharmaceutical compounds. international journal of nanomedicine 13 (2018) 4227–4245. [46] t. yan, y.c. zhixiang, w.d. huang, h. miao, y. zhang. preparation and characterization of baicalein powder micronized by the seds process. the journal of supercritical fluids 104 (2015 )177-189. [47] g.p.s. aguiar, b.d. arcari, m.p.c. chaves, c.d. magro, d. l. boschetto, a.l. piato, m. lanza, j. v. oliveira. micronization of trans-resveratrol by supercritical fluid: dissolution, solubility and in vitro antioxidant activity. industrial crops and product 112 (2018) 1-5. [48] q.l. suo, w.z. he, y. c. huang, c. huang, c. pingli, h.l. hong, y.x. li, m.d. zhu. micronization of the natural pigment-bixin by the seds process through prefilming atomization. powder technology 154 (2005) 110-115. [49] j. wang, k. zhang, d. zhou, f. ge, y. zhao. aescin nanoparticles prepared using seds: composition stability and dissolution enhancement. the journal of supercritical fluids 130 (2017) 267-272. http://dx.doi.org/10.5599/admet.811 s.k. misra and k. pathak admet & dmpk 8(4) (2020) 355-374 372 [50] k. kaga, m. honda, t. adachi, m. honjo, m. goto. nanoparticle formation of pvp-astaxanthin inclusion complex by solution-enhanced dispersion by supercritical fluids (seds): effect of pvp and astaxanthin z-isomer content. the journal of supercritical fluids 136 (2018) 44-51. [51] s.y. lee, l.c. abdullah, r.a. rahman, f. abas, w.k. tan, g.h. chong. solution enhanced dispersion by supercritical fluids (seds): an approach in particle engineering to modify aqueous solubility of andrographolide from andrographis paniculata. chemical engineering research and design 138 (2018) 176-189. [52] y. hakuta, h. hayashi, k. arai. fine particle formation using supercritical fluids. current opinion on solid state material science 7 (2003) 341–351. [53] j. jung, m. perrut. particle design using supercritical fluids: literature and patent survey. the journal of supercritical fluids 20 (2001) 179–219. [54] z. knez, m. škerget, h.m. knez, d. cucek. particle formation using suband supercritical fluids. in: v. a. fan (ed.), supercritical fluid technology for energy and environmental applications. elsevier, boston, 2014, pp. 31–67 [55] w. chen, x. hu, y. hong, y. su, h. wang, j. li. ibuprofen nanoparticles prepared by a pgsstm-based method. powder technology 245 (2013) 241–250. [56] m. fraile, y. martín, d. deodato, s. rodriguez-rojo, i.d. nogueira, a.l. simplício, m. j. cocero, c.m.m. duarte. production of new hybrid systems for drug delivery by pgss (particles from gas saturated solutions) process. the journal of supercritical fluids 81 (2013) 226–235. [57] e. weidner, r. steiner, z. knez. powder generation from polyethylene glycols with compressible fluids, in: p.r. von rohr, c. trepp, (eds.) process technology proceedings 12, high pressure chemical engineering. netherlands, elsevier, 1996, pp. 223-228. [58] r.e. sievers, p.d. milewski, s.p. sellers, b.a. miules, b.j. korte, k.d. kusek, g.s. clack, b. mioskowski, j.a. villa. supercritical and near-critical carbon dioxide assisted low-temperature bubble drying. industrial and engineering chemical research 39 (2000) 4831–4836. [59] p.s. božić, s. srčič, z. knez, j. kerc. improvement of nifedipine dissolution characteristics using supercritical co2. international journal of pharmaceutics 148 (1997) 123-130. [60] a. pestieau, f. krier, p. lebrun, a. brouwers, b. streel, b. evrard. optimization of a pgss process for a fenofibrate lipid-based solid dispersion formulation. international journal of pharmaceutics 485 (2015) 295-305. [61] g. krananja, z. knez, p. kotnik, b. ljubec, m. knez. formulation of nimodipine, fenofibrate, and ovanillin with brij s100 and peg 4000 using the pgss™ process. the journal of supercritical fluids 135 (2018) 245-253. [62] a.s. pedro, s.d. villa, p. calicet, s.a.b.v. melo, e.c. albuquerque, a. bertucco, s. salmaso. curcuminloaded solid lipid particles by pgss technology. the journal of supercritical fluids 107 (2016) 534541. [63] m. fraile, a. martin, d. deodato, s. rodriguez-rojo, i.d. nogueira, a.l. simplicio, m.t.j. cocero, c.m.m. duarte. production of new hybrid systems for drug delivery by pgss. the journal of supercritical fluids 81 (2013) 226-235. [64] v. s.s. gonçalves, a. a. matias, s. rodríguez-rojo, i. d. nogueira, c.m.m. duarte. supercritical fluid precipitation of ketoprofen in novel structured lipid carriers for enhanced mucosal delivery-a comparison with solid lipid particles. international journal of pharmaceutics 495 (2015) 302-311. [65] g. levai, a. martin, a. moro, a.a. matias, v.s.s. goncalves, m.r. bronze. production of encapsulated quercetin particles using supercritical fluid technologies. powder technology 317 (2017) 142-153. [66] e. paz, a. martin, m.m. duarte, m.j. cocero. formulation of β-carotene with poly-(ε-caprolactones) pgss process. powder technology 217 (2012) 77-83. [67] a.r.s. sousa, r. silva, f.h. tay, a.l. simplicio, s.g. kazarian, c.m.m. duarte. solubility enhancement of trans-chalcone using lipid carriers and supercritical co2 processing. the journal of supercritical fluids 48 (2009) 120-125. https://www.sciencedirect.com/science/article/pii/s0263876218304295#! admet & dmpk 8(4) (2020) 355-374 scf technology for solubility enhancement doi: http://dx.doi.org/10.5599/admet.811 373 [68] m. haq, b.s. chun. microencapsulation of omega-3 polyunsaturated fatty acids and astaxanthin-rich salmon oil using particles from gas saturated solutions (pgss) process. lwt-food science and technology 92 (2018) 523-530. [69] m. j. meziani, p. pathak, r. hurezeanu, m. c. thies, r.m. enick, y.p. sun. supercritical fluid processing technique for nanoscale polymer particles. angewandte chemie 43 (2004) 704-707. [70] p. pathak, m.j. meziani, t. desai, y.p. sun. nanosizing drug particles in supercritical fluid processing. journal of american chemical society (2004) 10842-10843. [71] c.p. price, a. l. grzesiak, a. j. matzger. crystalline polymorph selection and discovery with polymer heteronuclei. journal of american chemical society 127 (2005) 5512-5517. [72] s. sala, e. elizondo, e. moreno, t. calvet, m. a. cuevas-diarte, n. ventosa, j. veciana. kinetically driven crystallization of a pure polymorphic phase of stearic acid from co2-expanded solutions. crystal growth and design 10 (2010) 1226-1232. [73] e. moreno-calvo, f. temelli, a. cordoba, n. masciocchi, j. veciana, n. ventosa. a new microcrystalline phytosterol polymorph generated using co2-expanded solvents. crystal growth and design 14 (2014) 58−68. [74] w. teng, j. wang, n. r. foster, n. wen, j. zhang. preparation of silybin-poly (vinyl pyrrolidone) nanodrugs by using the aerosol solvent extraction system for improving drug solubility. industry and engineering chemistry research 53 (2014) 10519−10524. [75] l. a. meure, b. warwik, f. dehgnani, h.l. regtop, n.r. foster. increasing copper indomethacin solubility by coprecipitation with poly (vinylpyrrolidone) using the aerosol solvent extraction system, industrial engineering in chemical research 43 (2004) 1103-1112. [76] q. rao, z. qiu, t. zhenyu, j. zhang, d. luo, p. pan, l. zhang, y. liu, s. guan, z. qingguol. enhancement of the apparent solubility and bioavailability of tadalafil nanoparticles via antisolvent precipitation. european journal of pharmaceutical sciences 128 (2019) 222-231. [77] o. guney, a. akgerman. synthesis of controlled-release products in supercritical medium, aiche journal 48 (2002) 856–866. [78] s. sethia, e. squillante. solid dispersion of carbamazepine in pvp k30 by conventional solvent evaporation and supercritical methods. international journal of pharmaceutics 272 (2004) 1–10. [79] s.g. kazarian, g.g. martirosyan. spectroscopy of polymer drug formulations processed with supercritical fluids: in situ atr-ir and raman study of impregnation of ibuprofen into pvp. international journal of pharmaceutics 232 (2002) 81–90 [80] l. manna, m. banchero, d. sola, a. ferri, s. ronchetti, s. sicardi. impregnation of pvp microparticles with ketoprofen in the presence of supercritical co2. the journal of supercritical fluids 42 (2007) 378–384 [81] m. banchero, l. manna, s. ronchetti, p. campanelli, a. ferri. supercritical solvent impregnation of piroxicam on pvp at various polymer molecular weights. the journal of supercritical fluids 49 (2009) 271–278. [82] m. moneghini, i. kikic, d. voinovich, b. perissutti, j. filipovic-grcic. processing of carbamazepine-peg 4000 solid dispersions with supercritical carbon dioxide: preparation, characterization and in vitro dissolution. international journal of pharmaceutics 222 (2001) 129–138. [83] m.j. cocero, s. ferrero. crystallization of β-carotene by a gas process in batch effect of operating conditions. the journal of supercritical fluids 22 (2002) 237–245. [84] s. park, s. yeo. recrystallization of caffeine using gas antisolvent process. the journal of supercritical fluids 47 (2008) 85–92. [85] y. bakhbakhi, s. rohani, p.a. charpentier. micronization of phenanthrene using the gas antisolvent process. 1. experimental study and use of ftir. industrial engineering in chemical research 44 (2005) 7337–7344. [86] f. fusaro, m. mazzotti, g. muhrer. gas antisolvent recrystallization of paracetamol from acetone using compressed carbon dioxide as antisolvent. crystal growth and design 4 (2004) 881–889. http://dx.doi.org/10.5599/admet.811 https://www.sciencedirect.com/science/journal/09280987/128/supp/c s.k. misra and k. pathak admet & dmpk 8(4) (2020) 355-374 374 [87] n. esfandiari, s. m. ghoreishi. ampicillin nanoparticles production via supercritical co2 gas antisolvent process. aapspharmscitech 16 (2015) 1263-1269. [88] m. munto, n. ventosa, s. sala, j. veciana. solubility behaviors of ibuprofen and naproxen drugs in liquid “co2–organic solvent” mixtures. the journal of supercritical fluids 47 (2008) 147-153. [89] s.t. xiang, r.k. kankala, s.b. wang, a.z. chen. solubility measurement and resolv-assisted nanonization of gambogic acid in supercritical carbon dioxide for cancer therapy. the journal of supercritical fluids 150(2019)147-155. [90] a. ubeyitogullari, o.n. ciftci. generating phytosterol nanoparticles in nanoporous bioaerogels via supercritical carbon dioxide impregnation: effect of impregnation conditions. journal of food engineering 207 (2017) 99-107. [91] a. sane, j. limtrakul. formation of retinyl palmitate-loaded poly (l-lactide) nanoparticles using rapid expansion of supercritical solutions into liquid solvents (resolv). the journal of supercritical fluids 51 (2009) 230-237. [92] s.v. dalvi, m.a. azad, r. dave. precipitation and stabilization of ultrafine particles of fenofibrate in aqueous suspensions by resolv. powder technology 236 (2013) 75-84. [93] a.s. pessoa, g. pablo, s. aguiar, j. vladimir, o. adailton, j.b ortoluzzi, a. paulino, m. lanza. precipitation of resveratrol-isoniazid and resveratrol-nicotinamide cocrystals by gas antisolvents. the journal of supercritical fluids 145 (2019) 93-102. [94] s. harikrishnan, r. sedev, c. beh, c. priest. loading of 5-fluorouracil onto halloysite nanotubes for targeted drug delivery using a subcritical gas antisolvent process (gas), the journal of supercritical fluids 159 (2020) 104756. https://doi.org/10.1016/j.supflu.2020.104756. [95] i.g. casas, c. crampon, a. montes, c. pereyra, e.j. martinez. supercritical co2 impregnation of silica microparticles with quercetin. the journal of supercritical fluids 143 (2019) 157-161. [96] c.v. silva, v.j. pereira, g.m.n. costa, e.c.m.c. albuquerque, s.a.b.v. melo, h.c. sousa, a.m.a. dias, m.e.m. braga. supercritical solvent impregnation/deposition of spilanthol-enriched extracts into a commercial collagen/cellulose-based wound dressing, the journal of supercritical fluids 133 (2018) 503-511. [97] t. y. yan, y. zhang, m. ji, z. wang, t. yan. preparation of irbesartan composite microparticles by supercritical aerosol solvent extraction system for dissolution enhancement. the journal of supercritical fluids 153 (2019) 104594. https://doi.org/10.1016/j.supflu.2019.104594. [98] w. yu, f. xia, h. jin, c. lin, y. zhao, s. jiang, l. he. production of submicroparticles of β-sitosterol using an aerosol solvent extraction system. chinese journal of chemical engineering 16 (2008) 956960. [99] s.y. lee, i. jung, j.k. kim, g.b. lim, j.h. ryu. preparation of itraconazole/hp-β-cd inclusion complexes using supercritical aerosol solvent extraction system and their dissolution characteristics. the journal of supercritical fluids 44 (2008) 400-408. ©2020 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) https://doi.org/10.1016/j.supflu.2020.104756 https://doi.org/10.1016/j.supflu.2019.104594 http://creativecommons.org/licenses/by/3.0/ dissolution of an ensemble of differently shaped poly-dispersed drug particles undergoing solubility reduction: mathematical modelling doi: http://dx.doi.org/10.5599/admet.841 297 admet & dmpk 8(3) (2020) 297-313; doi: http://dx.doi.org/10.5599/admet.841 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper dissolution of an ensemble of differently shaped poly-dispersed drug particles undergoing solubility reduction: mathematical modelling michela abrami 1 , lucia grassi 2 , rosario di vittorio 1 , dritan hasa 3 , beatrice perissutti 3 , dario voinovich 3 , gabriele grassi 4 , italo colombo 1 and mario grassi 1* 1 dept. of engineering and architecture, trieste university, via alfonso valerio, 6/a, trieste, i-34127 italy 2 liceo scientifico g. galilei, trieste, via mameli 4, i-34139 italy 3 dept. of chemical and pharmaceutical sciences, trieste university, piazzale europa 1, trieste, i-34127, italy 4 dept. of life sciences, cattinara university hospital, trieste university, strada di fiume 447, trieste, i-34149 italy *corresponding author: e-mail: mario.grassi@dia.units.it; tel.: +39-040-558-3435; fax: +39-040-569823 received: april 30, 2020; revised: june 18, 2020; published: july 14, 2020 abstract the aim of this theoretical paper is to develop a mathematical model for describing the dissolution process, in a finite liquid environment, of an ensemble of poly-dispersed drug particles, in form of sphere, cylinder and parallelepiped that can undergo solubility reduction due to phase transition induced by dissolution. the main result of this work consists in its simplicity as, whatever the particular particles size distribution, only two ordinary differential equations are needed to describe the dissolution process. this, in turn, reflects in a very powerful and agile theoretical tool that can be easily implemented in electronic sheets, a widespread tool among the research community. another model advantage lies on the possibility of determining its parameters by means of common independent techniques thus enabling the evaluation of the importance of solid wettability on the dissolution process. keywords recrystallization; mathematical modelling; dissolution; particles; poly-dispersion; bioavailability introduction the great variety of controlled drug delivery systems is designed in reason of the drug to be delivered and the clinical target that, in turn, determine the choice of administration routes, among which the most important are the oral, injectable, inhaled, transmucosal, transdermal and the implantable one [1]. despite the specific delivery system and the administration route considered, the in vivo drug fate is usually characterized by some common steps such as the release from the delivery system and the subsequent tissue absorption and distribution, metabolism and elimination (l-adme processes) [2]. specifically to the first step, the release kinetics can be ruled by different phenomena including swelling, erosion, drug dissolution, drug transport (due to diffusion and/or convection), drug interaction with the delivery system, initial drug distribution inside the delivery system and some delivery system geometrical characteristics [3,4]. independently from the administration route and drug, a key factor for the success and reliability of every delivery system is drug bioavailability, defined as the rate and extent to which the active drug is http://dx.doi.org/10.5599/admet.841 http://dx.doi.org/10.5599/admet.841 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:mario.grassi@dia.units.it mario grassi et al. admet & dmpk 8(3) (2020) 297-313 298 absorbed from a pharmaceutical form thus becoming available at the site of drug action [5,6]. in turn, bioavailability depends strongly on drug permeability through cells membrane and drug solubilisation in physiological fluids. this is clearly pointed out by the bio-pharmaceutic classification according to which drugs can be subdivided into four classes: class i (high solubility and high permeability), class ii (low solubility, high permeability), class iii (high solubility, low permeability) and class iv (low solubility and low permeability) [7]. while permeation implies drug partitioning between a polar aqueous phase and an apolar phase (cellular membranes), unless active mechanisms rule drug permeation, solubilisation implies the drug dissolution process. this phenomenon must not be confused with the release process as dissolution is defined as “the mixing of two phases with the formation of a new homogeneous phase (i.e. the solution)” [8]. dissolution, in particular, assumes relevant importance for class ii drugs that, interestingly, represent about 40 % of the marketed drugs [9,10] and 70-90 % of the new chemical entities [9–12]. indeed, most of the drugs are optimized solely on the basis of their pharmacological activity and not for what concerns bioavailability. examples of commonly marketed drugs that are poorly soluble in water (less than 100 g/cm 3 [13]) include non-steroidal anti-inflammatory drugs (nsaids), anticholesterol, antimycotics, antibiotics, anticonvulsants, chemotherapeutics, antivirals, β-blockers, calcium channel blockers and immunosuppressants [4,14-18]. the above considerations make clear why the drug dissolution process is actively studied in the pharmaceutical field from both an experimental and a theoretical viewpoint [4,19]. experimentalists, typically, prefer to perform dissolution rate test (drt) [20,21], implying the dissolution of an ensemble of poly-dispersed drug particles in water or in a physiological fluid. in contrast, theorists propend for intrinsic dissolution rate test (idr), implying the dissolution from the flat surface of a drug tablet fixed to a rotating shaft [4]. while drt is much closer to the in vivo drug performance, idr is much simpler to be modelled. indeed, idr takes place in a much more controlled frame as fluid hydrodynamics around the flat rotating surface is well understood thanks to the elegant approach of levich [22]. however, nowadays, the necessity of designing more and more sophisticated delivery systems stimulated theorists to move towards experimentalists. therefore, interesting studies about the hydrodynamic conditions taking place in drt [2324] and the direct observation of particles size reduction during drt [19] have been undertaken. additionally, due to the relevance in the pharmaceutical field, theorists started considering also possible variation of solubility occurring upon dissolution, whose kinetics, as later discussed, essentially depends on the surface available for dissolution, on mass transport resistance occurring at, and around, the solid/liquid interface and on drug solubility in the liquid phase. indeed, it is well known that the contact between the drug and the liquid environment can lead to polymorphic transformations, as in the case of anhydrous theophylline that becomes hydrated [25] or nicergoline [26], or to amorphous drugs that recrystallize in the most stable crystalline form as it occurs for temazepam [14], nimesulide [27] and posaconazole [28]. typically, phase transformations imply a reduction of solubility that, unavoidably, reflects in a reduction of drug bioavailability. thus, these evidences underline the important role played by solubility on dissolution kinetics and increase the complexity of the already complex problem regarding solubility determination [29–31]. historically, the first fundamental approach aimed at the description of drt was that of hixson and crowell [32-34]. for the first time, they accounted for surface reduction upon dissolution of spherical particles and build up the famous cubic law. then, the elegant model of pedersen and co-workers accounted also for spherical particles poly-dispersion [35-38]. interestingly, this model reduces to the hixson-crowell one in the case of monodispersed spherical particles. since then, many other models were built up. among others, we can remember the one of thormann and co-workers that considers also the admet & dmpk 8(y) (2020) 297-313 mathematical modelling of dissolution rate doi: http://dx.doi.org/10.5599/admet.841 299 possible drug degradation after dissolution [39], the model of hirai and co-workers that does not account explicitly for the shape of particles but focuses the attention on a law able to describe the time dependence of the dissolution surface [40] and the model of guo and co-workers focusing on polymorphic transformation induced by dissolution (rifampicin, from form ii to i) [41]. the aim of this paper is to build up a mathematical model able to merge a reasonably accurate description of the physical phenomena occurring in drt with the practical need of experimentalists who demand theoretical tools able to guide experiments design and to allow a reliable data interpretation. in particular, the model will account for drug solubility reduction upon dissolution, particle poly-dispersion, particle geometry (spherical, cylindrical and parallelepiped) and the presence of a finite release environment. mathematical modelling dissolution can be considered as a consecutive process made up by five steps [8,42,43]: 1) contact of the solvent with the solid surface (wetting), which implies the production of a solid/liquid interface starting from solid/vapor one, 2) breakdown of intermolecular bonds in the solid phase (fusion), 3) molecules transfer from the solid phase to the solid/liquid interface (solvation), 4) diffusion of the solvated molecules through the unstirred boundary layer surrounding the solid surface (diffusion), 5) convective transport of solvated drug molecules into the well stirred bulk solution (convection). the first four steps, that can be viewed as the sum of four energy steps, represent the total resistance that the drug molecules have to overcome in order to move from the solid phase to the solution one (dissolution). obviously, the higher the dissolution energy required (i.e. the higher the mass transfer resistance) the lower the dissolution kinetics is. in order to connect the four steps, we need to know the time evolution of the drug concentration profile in the unstirred boundary layer surrounding the solid surface, whose presence is unavoidable and whose thickness depends on the relative velocity among particles and external fluid, on fluid kinematic viscosity, on particles dimension and on drug diffusion coefficient in the unstirred layer [24]. for this purpose, recourse can be made to the fick’s second equation:   c d c t      (1) where c is drug concentration, t is time, d is the drug diffusion coefficient in the unstirred layer and  is the gradient vector whose components analytical expression depends on the particular reference system chosen (cartesian, spherical, cylindrical). the first model assumption relies on the hypothesis that mass transport inside the unstirred layer is one dimensional (the direction is that perpendicular to the solid surface) while the second one consists of the rapid attainment of pseudo-stationary conditions in the unstirred layer (this hypothesis is supported by the numerical solution of eq. (1) assuming usual values for d (~ 10 –10 m 2 s –1 ) [4] and stagnant layer thickness ≤ 20 m). thus, eq. (1) becomes:   0d c   (2) eq. (2) has to be solved considering the following initial and boundary conditions: initial: c(ξ) = 0 ξmin < ξ ≤ ξmax (3) boundary:     min min( )m sd c k c c        n (4) c(ξmax) = cb (5) http://dx.doi.org/10.5599/admet.841 mario grassi et al. admet & dmpk 8(3) (2020) 297-313 300 where  is the one dimensional spatial coordinate, (max -min) is the thickness of the unstirred layer (), n is the surface normal versor, cs is drug solubility in the dissolution liquid while km is the interface mass transfer coefficient mainly depending on the dissolution surface wetting properties and representing dissolution step 1. eq. (3) affirms that, at the beginning, the unstirred layer does not contain drug molecules while eq. (4) states that the drug flux leaving the solid surface depends on km and on the difference between drug solubility and drug concentration at the solid-liquid interface on the liquid side. finally, eq. (5) imposes that drug concentration equates cb at the unstirred layer – bulk liquid interface. eq. (2) solution, in cartesian, cylindrical and spherical coordinates reads, respectively:  s b b d m ( ) 1 1 c c c c k k              cartesian (6)   max maxm mb s b d min d min ( ) ln / ln k k c c c c k k                         cylindrical (7)   2 maxm mb s b min max min d d c( ) = 1 / k k c c c k k                          spherical (8) where cb is the drug concentration in the bulk liquid at time t, cs is drug solubility in the liquid phase and kd is the mass transport coefficient (= d/) accounting for the fourth dissolution step (drug diffusion through the unstirred liquid layer). eqs. (6) – (8) make clear that while in the cartesian case (flat dissolution surface) a linear concentration profile develops in the unstirred layer, not linear concentration profiles take place when the dissolution surface is not flat. in addition, eqs. (6) – (8) allow to evaluate the drug concentration c0 in  = min, i.e., at the solid-liquid interface (liquid side): m m 0 s b s d d ( ) / (1 ) k k c c c c k k    cartesian (9)   max maxm m0 b s b d min min d min ln / ln k k c c c c k k                         cylindrical (10)   m m0 b s b min max min d d c = / k k c c c k k            spherical (11) eqs. (9) – (11) affirm that at t = 0, when cb = 0, c0 is a fraction of cs, while c0 = cs after a very long time when cb = cs, i.e. c0 increases with time up to cs. when no wettability problems occur (km → ∞), c0 is always equal to cs while it is equal to cb = 0 when km → 0 (very poorly wettable solids). by means of eqs. (6) – (8), it is possible writing the ordinary differential equation accounting for drug concentration increase in the liquid phase: v dcb dt =-s (d ∂c ∂ξ )| ξ=ξmax = sk(cs-cb) , (12) where v is the liquid volume and k is an overall mass transport coefficient assuming different analytical expressions depending on the particular coordinate system considered: cartesian cylindrical spherical m d 1 k = 1 / 1 /k k min max max min m d max min / k = ln( / )1 1 ( / 1)k k         in ax max d m min / k = 1 1 ( ( )) m m k k      (13) admet & dmpk 8(y) (2020) 297-313 mathematical modelling of dissolution rate doi: http://dx.doi.org/10.5599/admet.841 301 eq. (12) states that the increase of the drug mass in the fluid phase is equal to the drug flow leaving the solid surface available for dissolution. in addition, eq. (13) says that when no wettability problems arise (km → ∞) k coincides with kd, provided that the thickness of the unstirred layer  is very small (max ≈ min) in the case of cylindrical and spherical geometry (obviously, when km → 0, k → 0). although k varies with particle dimension as kd depends on particle dimension [24], in order to simplify the scenario, the third hypothesis of the present model considers the k independence on particle dimensions so that an average value has to be considered. as a matter of fact, the great advantage of this usual hypothesis consists in an easy description of the dissolution process characterizing an ensemble of poly-dispersed particles of different shape (parallelepiped, cylinder, sphere). indeed, it allows assuming that the solid amount dissolved per unit time and surface is the same whatever the surface delimiting the solid particle. accordingly, in the case of a parallelepiped of dimensions xi, yi, zi, we have: p i i i i i i i i i i i i i d d( ) d( ) d( ) d( ) d d d d d m xy z x y z y z x z xy t t t t t        , (14)    i i i ii i i i s b s b d( ) d( ) d( ) d( ) ρy    -2     ---       -2  =       d d d d x x y zk z y z k c c c c t t t t        , (15) where mi p and  are the parallelepiped mass and density, respectively. as the displacement  of the dissolution front is the same for each one of the parallelepiped surface, from eq. (15) we have:  i i i0i s b d( ) d( ) d( )d( ) 1 1 1 -2 -- d 2 d 2 d 2 d i x y z k x x c c t t t t           (16) in the case of a cylinder of dimensions ri, li, we have: c 2 2i i i i i i i i d d( ) d( ) d( ) 2π π d d d d m r l r l lr r t t t t       (17)    i ii i i i s b s b d( ) d( ) 2π  -2π      --      d d r r k lr rlk c c c c t t        (18)    2 2i ii i s b s b d( ) d( ) π    -2π           --      2 d d l l k r r k c c c c t t        (19) where mi c is the cylinder mass. consequently, it follows:  i ii 0i i 0i s b d( ) d( )d( ) 1 ; -2 -- = d d 2 d r l k r r l l c c t t t           (20) finally, for a sphere of radius ri, we have:     3 s i 2 2i i i i i s b s b 4 d π d d( ) d( )3 4π    -4π  -    d d d d r m r r k k r r c c c c t t t t                 (21) where mi s is the sphere mass. consequently, it follows:  ii 0i s b d( )d( ) -- = d d r k r r c c t t        . (22) the most important message of eqs. (14) – (22) relies on the possibility of determining particles dimensions upon dissolution, regardless of the geometry, by the evaluation of just the dissolution front displacement . http://dx.doi.org/10.5599/admet.841 mario grassi et al. admet & dmpk 8(3) (2020) 297-313 302 in order to complete the model, it is necessary evaluating both cs and cb. indeed, as discussed in the introduction, it is quite common that, upon dissolution, the drug undergoes a phase transformation (polymorphic or amorphous – crystalline) implying a solubility reduction. typically, this phenomenon is described by a first order reaction [44] occurring at the solid-liquid interface and leading to the following expression for the cs temporal reduction: cs = csf+(cs-in-csf)e (-krt) , (23) where csf and cs-in are, respectively, the final and initial values of solubility while kr is the recrystallization constant and t is time. as a matter of fact, eq. (23) accounts for the dissolution step 2 as solubility is directly connected with the crystal network breakdown attitude that is quantified by its melting temperature and enthalpy [45]. in order to account also for a possible recrystallization in the bulk phase (occurring when cb(t) > cs(t)), it is necessary considering the following equation: dmc dt = krb v(cs(t) cb(t)) , (24) where mc is the amount of recrystallized solid and krb is the bulk recrystallization constant that can differ from kr. obviously, while eq. (24) works only when cb(t) exceeds cs(t), the initial value for mc is set to zero. the determination of cb relies on a global mass balance ensuring that the initial solid mass (m0) must be equal, at any time, to the sum of the undissolved mass, the solubilized drug present in the bulk solution and mc (the very small thickness of the boundary layer renders negligible the drug amount contained in it):     i=n i=n 0 pi i ci=1 0 pi pi b c b i=1 ( ) = ( ) - m n v m t m n v c t v m t c t v          , (25) where npi represents the number of particles characterized by a volume vpi and n is the number of classes into which the particle size distribution is subdivided in. it is interesting to notice that eq. (25), de facto, accounts for the presence of a finite volume of the liquid phase (v). indeed, when v → ∞, cb will be always zero. for its mathematical attitude in describing particle size distributions, the weibull equation was considered to describe the initial particles size distribution [46]: i min(-(2 ) ) i i 0 =1v w e v      , (26) where v0 is the total volume of the solid drug, vi is the volume occupied by all the particles sharing the same characteristic dimension i (radius in the case of spheres and cylinders, x dimension in the case of parallelepipeds), min is the minimum value of i while  and  are two model parameters. once particle geometry is fixed, vi enables the determination of npi by a simple geometric relation. obviously, when the generic i goes to zero due to dissolution, npi is set to zero as this particle class has disappeared and it no longer contributes to the dissolution phenomenon. thus, eq. (26) serves only to evaluate the initial number of particles belonging to class i th . in conclusion, the proposed model, establishing a connection among the dissolution steps, assumes that the dissolution kinetics can be essentially affected by steps 1, 2 and 4, attributing to step 3 a negligible role. in addition, whatever the number of classes constituting the particles size distribution, this model needs only two differential equations: one among the differential equations appearing in eq. (16), (20) and (22), aimed at the prediction of the time evolution of the dissolution front position (), plus eq. (24) accounting for the time evolution of the recrystallized drug amount from the solution. indeed, once  is known, it is admet & dmpk 8(y) (2020) 297-313 mathematical modelling of dissolution rate doi: http://dx.doi.org/10.5599/admet.841 303 possible evaluating the dimension i for all the particle class particles resorting to the algebraic relation linking i to  for the different geometries (see the first relation appearing in eqs. (16), (20) and (22)). of course, when i ~ 0 due to the dissolution process, the particle class particles is no longer considered letting i = npi = 0. as this model does not lead to an analytical solution, its iterative numerical solution (relaxation method, relaxation parameter  = 1, relative tolerance = 10 -3 [47]) was performed discretizing the two differential equations by means of the implicit euler method. in order to ensure the numerical solution accuracy and stability, the time step was set to 0.25 s and the particles size distribution was subdivided into n = 200 classes. results and discussion as many parameters affect model output, it would be impossible discussing the effects of all of them and how they interact to affect the dissolution kinetics. accordingly, we decided to focus on three aspects, namely particle shape (parallelepiped, cylinder, sphere), particle size distribution and the ratio (v + ) between the liquid phase volume (v) and the initial solid drug volume (m0/). indeed, especially the last two, represent the most important parameters available for the designing of the experimental set up. in order to properly evaluate the effect of particles shape on dissolution kinetics, the parameters of the weibull distribution, whose differential expression is: i min(-(2 ) ) (α-1)i 0 i min di d( / ) = (2 )(2 ) d v v w e           , (27) were fixed to get similar profile of eq. (27) for parallelepiped, cylinder and sphere when plotted versus the volume competing to particles identified by the characteristic dimension i. in the case of sphere i = ri and the corresponding volume is (4/3)ri 3 , in the case of cylinder i = ri and the corresponding volume is ri 3 fr, while, in the case of parallelepiped, i = xi so that the corresponding volume is xi 3 fyxfzx. for the sake of simplicity, in the case of cylinders and parallelepipeds, it was assumed that the ratio fr between cylinder radius and length, and the ratios fyx = yi/xi and fzx = zi/xi for parallelepiped were the same for all particles. accordingly, fr, fyx and fzx can be considered as shape factors characterizing, respectively, cylinders and parallelepipeds. figure 1a shows the trend of wdi for spheres, parallelepipeds and cylinders vs vi assuming fr = 2 and fyx = fzx = 1 (cubical cylinder and cube). although the three wdi do not share exactly the same wideness in vi, they are characterized by the same peak amplitude and position. in addition, as all of them are very narrow (10 m ≤ ri ≤ 11 m, 16 m ≤ xi ≤ 17 m and 9 m ≤ ri ≤ 10 m for spheres, parallelepipeds and cylinders, respectively) they can be thought to be representative of approximately mono-dispersed size distributions. relaying on these particle size distributions and assuming typical values for the other model parameters [4], figure 1b shows model predictions considering different values of the ratio v + between liquid volume (v) and initial particles volume (v0=m0/) in the case of a drug undergoing recrystallization upon dissolution. it is clear that whatever v + , particle shape does not seem important as the dimensionless drug concentration in the liquid phase cb + (= cb/cf; left vertical axis) is very similar for spheres (thick line), parallelepiped-cubes (thin line) and cubical cylinders (dashed line). in addition, when v + is high (≥ 1500), the classical oversaturation peak does not appear (the over saturation condition is never met as cb + (t) is always lower that than cs + (t)) and only for smaller v + it occurs, becoming evident for v + ≤ 270. finally, it is interesting to underline that the cb + peak always occurs when the time dependent dimensionless drug solubility cs + (= cs(t)/cf; right vertical axis) crosses the cb + trend. indeed, from now on, the solution is in over saturation conditions and only after a very long time cb + equates cs + . http://dx.doi.org/10.5599/admet.841 mario grassi et al. admet & dmpk 8(3) (2020) 297-313 304 figure 1. (a) weibull differential distribution wdi vs particles volume vi for spheres, parallelepipeds and cylinders. wdi parameters read: sphere –  = 2,  = 0.8 m, rmin = 10 m -, parallelepiped  = 2.35,  = 0.9 m, xmin = 16 m, fyx = fzx = 1 (cubes) -, cylinder  = 2,  = 0.8 m, rmin = 9 m, fr = 2 (cubical cylinders). (b) model predictions relative to the particle size distribution shown in figure 1a. cb + (left vertical axis) and cs + (right vertical axis) are, respectively, the dimensionless drug concentration and solubility in the liquid phase (normalized with respect of the drug final solubility cf) while t + is a dimensionless time defined by t+= (tk) √v0 3⁄ , being k the dissolution constant and v0 the particles volume before dissolution. v + is the ratio between liquid volume (v) and v0. other model parameters, set according to [4], read:  = 1.5 g/cm 3 , cs-in = 2*10 -2 g/cm 3 , csf = 10 -3 g/cm 3 , kr = krb = 10 -2 s -1 , k = 10 -3 cm/s. in order to appreciate the effect of particle shape factor on dissolution kinetics, we considered parallelepipeds characterized by three different values of the shape factor fzx (0.25, 1 and 10), the same value for fyx = 1 (parallelepipeds with square basis and different heights) and the same xi size distribution adopted in figure 1a (parallelepiped case). the three distributions considered, depicted in figure 2a, clearly show that all other parameters being equal, the decrease of fzx implies smaller particles and this, in turn, reflects in an increased dissolution surface. practically speaking, we are comparing almost monodispersed particles size distributions composed by thin platelets (fzx = 0.25), cubes (fzx = 1) and long rods (fzx = 10). figure 2. (a) weibull differential distribution wdi vs. particles volume vi for parallelepipeds characterized by three values of the fzx shape factor (0.25, 1, 10), the same shape factor fyx = 1 and the parameters adopted in figure 1a (parallelepiped case):  = 2.35,  = 0.9 m, xmin = 16 m. (b) model predictions relative to the particle size distribution shown in figure 2a. cb + is the dimensionless drug concentration in the liquid phase (normalized with respect of the drug final solubility cf) while t + is a dimensionless time defined by t+= (tk) √v0 3⁄ , being k the dissolution constant and v0 the particles volume before dissolution. v + (= 150) is the ratio between liquid volume (v) and v0. other model parameters, set according to [4], read:  = 1.5 g/cm 3 , cs-in = 2*10 -2 g/cm 3 , csf = 10 -3 g/cm 3 , kr = krb = 10 -2 s -1 , k = 10 -3 cm/s. 0 5000 10000 15000 20000 25000 4.0e-09 4.5e-09 5.0e-09 5.5e-09 6.0e-09 w d i (c m -1 ) vi (cm 3) sphere parallelepiped cylinder a 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 0.2 0.4 0.6 0.8 cs +cb + t+ sphere parallelepiped cylinder c+b v + = 150 c+s c+b v + = 270 c+b v + = 750 c+b v + = 1500 b 0 5000 10000 15000 20000 25000 1.0e-09 1.0e-08 1.0e-07 w d i( cm -1 ) vi(cm 3) fzx = 1 fzx = 10 fzx = 0.25 a 0 2 4 6 8 10 12 0 0.2 0.4 0.6 0.8 cb + t+ fzx = 10 fzx = 0.25 fzx = 1 b admet & dmpk 8(y) (2020) 297-313 mathematical modelling of dissolution rate doi: http://dx.doi.org/10.5599/admet.841 305 assuming v + = 150 and the same model parameters adopted in figure 1b, model output is shown in figure 2b. it is clear that, in this case, the effect of particle shape is no longer negligible as cb + profile is undoubtedly affected by the increasing surface area available for dissolution induced by the different values of the shape factor fzx. thus, particle shape factor becomes important as soon as it reflects in a significant variation of the area available for dissolution. in addition, it turns out that, in the case of drugs undergoing recrystallization, the characteristics of the oversaturation peak can also depend on particles shape factor. with the aim of exploring also the effect of particles poly-dispersion, figure 3a considers parallelepipeds in form of platelets (fzx = 0.25, fyx = 1), cubes (fzx = 1, fyx = 1) and rods (fzx = 10, fyx = 1) characterized by the same poly-dispersion of the xi dimension. with respect to the distributions shown in figure 2a, the wideness of the three distributions is now increased as xi spans from about 1 m up to 100 m while in figure 2a we have 16 m ≤ xi ≤ 17 m. figure 3b witnesses that also in this case, despite poly-dispersion, a clear difference in the cb + kinetics emerges as, again, platelets are characterized by the smallest particles and rods correspond to the biggest one while cubes put in the middle as shown by the position of each distribution peak. obviously, the cb + temporal evolution is a little bit depressed with respect to that reported in figure 2b as poly-dispersion implies a reduction of the surface area available for dissolution. again, the importance of poly-dispersion is strictly related to the effects it implies on the surface area available for dissolution. figure 3. (a) weibull differential distribution wdi vs particles volume vi for parallelepipeds characterized by three values of the fzx shape factor (0.25, 1, 10), the same shape factor fyx = 1 and the following parameters:  = 2.35,  = 90 m, xmin = 1 m. (b) model predictions relative to the particle size distribution shown in figure 3a. cb + is the dimensionless drug concentration in the liquid phase (normalized with respect of the drug final solubility cf) while t + is a dimensionless time defined by t+= (tk) √v0 3⁄ , being k the dissolution constant and v0 the particles volume before dissolution. v + (= 150) is the ratio between liquid volume (v) and v0. other model parameters, set according to [4], read:  = 1.5 g/cm 3 , cs-in = 2*10 -2 g/cm 3 , csf = 10 -3 g/cm 3 , kr = krb = 10 -2 s -1 , k = 10 -3 cm/s. case studies after having explored some important model characteristics, with the aim to come closer to experimentalist need, in the following sections the attention will be focused on three different drugs undergoing solubility reduction upon dissolution. the first drug considered is anhydrous theophylline (c7h8n4o2, mw = 180.2; essentially neutral compound; it is a bronco-dilatator indicated mainly for asthma, bronchospasm, and copd) that, upon dissolution in water, transforms into the more stable monohydrate form (c7h8n4o2●h2o, mw = 198.2,  = 1.49 g/cm 3 helium pycnometry) [25]. this polymorphic transformation implies a solubility reduction (t = 25 °c and ph = 7) from cs-in = 11.6 mg/cm 3 (anhydrous) to csf = 6.1 mg/cm 3 (monohydrate). in addition, the 0 50 100 150 200 250 1.00e-12 1.00e-10 1.00e-08 1.00e-06 w d i( cm -1 ) vi(cm 3) a 0 2 4 6 8 10 12 0 0.2 0.4 0.6 0.8 cb + t+ fzx = 10 fzx = 0.25 fzx = 1 b http://dx.doi.org/10.5599/admet.841 mario grassi et al. admet & dmpk 8(3) (2020) 297-313 306 analysis of idr experiments reveals that k = 3.4*10 -3 cm/s and kr = 6*10 -3 s -1 [4]. as one of the key factors ruling dissolution kinetics is the surface area of theophylline particles, particular care has to be devoted to the determination of this parameter. first of all, the inspection of figure 4a suggests that rods (~ fyx = fzx = 0.25) could approximate the shape of anhydrous theophylline particles. then, on the basis of this consideration, the parameters ruling the weibull distribution, eq. (27), are chosen in order to meet the experimentally determined theophylline specific area (0.331 m 2 /g, mercury porosimetry, unpublished data). the following values come out from this procedure:  = 2,  = 115 m and 1 m ≤ xi ≤ 150 m. although this is not the unique choice of eq. (27) parameters leading to the experimentally determined specific surface area, it surely represents a physically sound choice. relying on this set of parameters, and assuming that the recrystallisation constant in the bulk liquid, krb, equates that on particles surface (kr), model predictions about anhydrous theophylline dissolution are performed assuming different values of v + as this is one of the most important and easy parameter to act on in order to properly design the experimental set up. figure 4b indicates that the over saturation peak appears for v + ≤ 194 and its amplitude increases with v + reduction up to about 2, i.e. the value of the ratio cs-in/csf. the decrease of the dimensionless solubility cs + , indicated by the dotted line in figure 4b, always intersects the cb + trend in its maximum and it is a measure of how fast the polymorphic transition takes place. in addition, figure 4c allows to evaluate, for different v + values, the temporal evolution of the dimensionless recrystallized drug amount (mc + ) jointly with the dimensionless drug amount that has not yet undergone dissolution (ms + ). while ms + shows a faster decrease for increasing v + , less regular is the behaviour of mc + due to its sigmoidal shape. the second drug considered is griseofulvin (c17h17clo6, mw = 352.8;  = 1.49 g/cm 3 helium pycnometry; neutral compound; it is an antifungal drug used to treat a number of types of dermatophytoses), a typical class ii drug that, upon milling, rapidly transforms in its amorphous state. the analysis of idr tests reveals that, at 37 °c and ph = 7.5, the solubility of amorphous griseofulvin in water is 235 g/cm 3 while crystalline solubility is 60 g/cm 3 . it is worth mentioning that 60 g/cm 3 exceeds the true griseofulvin solubility, i.e. that competing to griseofulvin crystals obtained by double recrystallization of native drug from acetonitrile, that is 12 g/cm 3 . the reason for this discrepancy is due to the presence of many defects in the crystal structure of the commercially available griseofulvin so that it results to be a mixture of macrocrystals, nanocrystals and amorphous phase [4]. as it is well known that solubility increases with the reduction of crystal dimension [45], this discrepancy is not surprising. however, as, typically, drt tests are performed on commercially available drugs, the solubility value of 60 g/cm 3 will be adopted for the following model simulations. furtherly, the analysis of idr test provides the following values for the dissolution k = 3.4*10 -3 cm/s and the recrystallization kr = 9*10 -4 s -1 constants [4]. again, the determination of the griseofulvin size distribution is performed by looking at the morphology of griseofulvin particles (figure 5a) and, then, by choosing a reasonable set of weibull parameters able to lead to the experimentally measured specific surface area (0.954 m 2 /g [48]). looking at figure 5a, we can roughly assume that griseofulvin particles are cubes (fyx = fzx = 1) while a reasonable choice for the weibull parameters is:  = 1.49,  = 13.6 m and 0.1 m ≤ xi ≤ 100 m. relaying on this set of parameters, and assuming that the recrystallisation constant in the bulk liquid, krb, equates that on particles surface (kr), model predictions about griseofulvin dissolution are performed assuming different values of v + as this is one of the most useful parameters aimed at the proper experimental set up design. figure 5b underlines that, due to relatively small kr (= krb) value (at least in comparison with that competing to anhydrous theophylline), the over saturation peak is scarcely visible, although present (see in figure 5a the intersection of the various cb + trends with cs + trend), whatever v + . in this case, indeed, we admet & dmpk 8(y) (2020) 297-313 mathematical modelling of dissolution rate doi: http://dx.doi.org/10.5599/admet.841 307 should more properly speak about an “over saturation plateau” whose entity, obviously, increases with v + decrease. a direct consequence of this behaviour is 1) the fast reduction of the drug amount that has not yet undergone dissolution (ms + ) at time t, whose reduction kinetics increases with v + (figure 5c) and 2) the limited amount of the dimensionless recrystallized drug amount (mc + ) that never exceeds the 10 % of m0 even for the smallest v + considered (figure 5c). the last drug studied is nimesulide (c13h12n2o5s, mw = 308.5;  = 1.49 g/cm 3 helium pycnometry; weak acid compound characterized by pka = 6.5 [49]; it is a non-steroidal anti-inflammatory drug), a typical class ii drug that, upon milling, can be transformed in nanocrystals and amorphous state. figure 4. (a) picture of anhydrous theophylline particles. (b) model predictions about theophylline drt assuming different values of the ratio v + between liquid (v) and particles (v0) volume. cb + (left vertical axis) and cs + (right vertical axis) are, respectively, the dimensionless drug concentration and solubility in the liquid phase (normalized with respect to the drug final solubility cf). (c) dimensionless amount of drug (ms + ) still solid after time t and amount of recrystallized drug (mc + ) in the liquid phase. both amounts are normalized with respect of the initial drug particles mass m0. nimesulide is, among the three drugs considered, the less soluble in water being its solubility at 37 °c and ph ≤ 6 equal to 9 g/cm 3 and around 100 g/cm 3 at the same temperature but ph = 7.5 [50]. fixing the attention on the ph ≤ 6 range, nimesulide solubility increases after co-grinding (in presence of a stabilizer) due to the transformation of macrocrystals into nanocrystals. although solubility depends on nanocrystals dimension [45], we can set for nanocrystals solubility the value of 28 g/cm 3 that corresponds to four hours co-grinding as described in [27]. figure 6a inspection reveals that nimesulide crystals look like needles so that their approximation by long cylinder (fr ~ 15) seems the best choice. in order to match the a 300 m 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 0 100 200 300 400 500 cs + cr + t(s) cr +, v+ = 75 cr +, v+ = 150 cr +, v+ = 194 cr +, v+ = 270 cs + b 0 10 20 30 40 50 60 70 80 90 100 0 5 10 15 20 25 0 100 200 300 400 500 1 0 0 m s+ 1 0 0 m c + t(s) ms +, v+ = 75 ms +, v+ = 150 ms +, v+ = 270 c http://dx.doi.org/10.5599/admet.841 mario grassi et al. admet & dmpk 8(3) (2020) 297-313 308 experimentally determined specific area of nimesulide powder after grinding (2.7 m 2 /g, mercury porosimetry; unpublished data), a reasonable choice for the weibull distribution is:  = 4,  = 0.54 m and 0.3 m ≤ ri ≤ 0.8 m. finally, the analysis of the release test performed in [27] let to conclude that the dissolution k is equal to 1.8*10 -5 cm/s and that the recrystallization constant kr is equal to 1.3*10 -2 s -1 . if the small k value witnesses the very low nimesulide propensity to dissolve, this being due to its low wettability (static water contact angle ~ 70° [51]), the high kr value indicates that nanocrystals and the amorphous phase are very instable and tend to rapidly recrystallize in the more stable crystalline form (this is the reason why nimesulide has to be co-ground together with a stabilizer, typically polyvinylpyrrolidone). relaying on the above parameters set, and assuming that the re-crystallisation constant in the bulk liquid, krb, equates that on particles surface (kr), model predictions about nimesulide dissolution are performed assuming different values of v + . figure 6b shows the appearance of more and more pronounced oversaturation peak for reducing v + . in addition, the high kr value (in comparison to those competing to theophylline and griseofulvin), makes peak shape very evident as the re-crystallization process is fast. clearly, when v + is sufficiently high, the peak disappears. the presence of a clearly identifiable peak, however, does not mean that a considerable amount of nimesulide goes in solution, as shown in figure 6c. indeed, dimensionless undissolved drug amount (ms + ) is always close to 100 % m0 and the dimensionless re-crystallized drug amount (mc + ) is always lower that 0.025% m0, whatever v + considered in the simulations. figure 5. (a) sem picture of griseofulvin particles. (b) model predictions about griseofulvin drt assuming different values of the ratio v + between liquid (v) and particles (v0) volume. cb + (left vertical axis) and cs + (right vertical axis) are, respectively, the dimensionless drug concentration and solubility in the liquid phase (normalized with respect to the drug final solubility cf). (c) dimensionless amount of drug (ms + ) still solid after time t and amount of recrystallized drug (mc + ) in the liquid phase. both amounts are normalized with respect of the initial drug particles mass m0. a 2 m 0 0.5 1 1.5 2 2.5 3 3.5 4 0 0.5 1 1.5 2 2.5 3 3.5 4 0 100 200 300 400 500 cs + cb + t(s) cb +, v+ = 745 cr +, v+ = 2980cs + b 0 2 4 6 8 10 12 0 10 20 30 40 50 60 70 80 90 100 0 100 200 300 400 500 1 0 0 m s+ 1 0 0 m c + t(s) ms +, v+ = 2980 ms +, v+ = 745c admet & dmpk 8(y) (2020) 297-313 mathematical modelling of dissolution rate doi: http://dx.doi.org/10.5599/admet.841 309 figure 6. (a) sem picture of nimesulide particles (needles) co-ground with polyvinylpyrrolidone (bigger particles). (b) model predictions about nimesulide drt assuming different values of the ratio v + between liquid (v) and particles (v0) volume. cb + (left vertical axis) and cs + (right vertical axis) are, respectively, the dimensionless drug concentration and solubility in the liquid phase (normalized with respect to the drug final solubility cf). (c) dimensionless amount of drug (ms + ) still solid after time t and amount of recrystallized drug (mc + ) in the liquid phase. both amounts are normalized with respect of the initial drug particles mass m0. conclusions assuming that: 1) mass transport inside the unstirred layer surrounding each particle is one dimensional, 2) that a rapid attainment of pseudo-stationary conditions in it takes place and that 3) the dissolution constant k is independent on particle dimensions, the proposed model provides a description of the dissolution kinetics (drt) from an ensemble of solid drug particles, eventually undergoing re-crystallization in a finite release environment, by means of only two differential equations, regardless of the number of the classes into which the continuous particle size distribution is subdivided. obviously, this represents a considerable advantage in computational and practical terms as it allows to easily implement the model also by means of electronic sheets that are widespread in the research community. in addition, this model allows to select particle shape among spheres, cylinder (characterized by different length/radius ratios) and parallelepiped (characterized by different y/x and z/x ratios), this driving the model close to the real situation. despite this model requires a considerable number of parameters, the majority of them can be properly determined by means of common independent experiments (typically idr test, mercury porosimetry, gas adsorption analysis (b.e.t.) and helium pycnometry) and only the dissolution constant k should be a 1 m 0 0.5 1 1.5 2 2.5 0 0.5 1 1.5 2 2.5 0 100 200 300 400 500 cs + cb + t(s) cs + b 0 20 40 60 80 100 0 0.005 0.01 0.015 0.02 0.025 0 100 200 300 400 500 1 0 0 m s+ 1 0 0 m c + t(s) ms +, v+ = 30, 60, 150, 745 c http://dx.doi.org/10.5599/admet.841 mario grassi et al. admet & dmpk 8(3) (2020) 297-313 310 determined by data fitting. indeed, as it depends on the relative velocity among particles and the surrounding liquid phase, its theoretical determination would require a complex analysis of the hydrodynamic conditions taking place in the liquid phase. in addition, we have to remember that k depends also on the mass transfer coefficient connected to the wettability of the solid surface (km) (1/k is equal to the sum of the mass transfer resistances due to poor wettability (1/km) and due to the presence of the unstirred layer surrounding each particle (1/kd)). consequently, the k experimental determination is, in our opinion, the simplest way to proceed also because a proper experimental set up, implying drt tests performed at increasing values of the reynolds number (re) (increasing with the stirring velocity imposed in the liquid phase), allows to separately evaluate kd and km. indeed, for high re, k should be almost independent on re as kd is proportional to re 0.5 [24] and, thus, dkd/dre → 0 for high re. consequently, k should be mainly dependent on solid wettability (lim𝑅𝑒→∞(1 𝐾⁄ ) = lim 𝑅𝑒→∞ (1 𝑘𝑑⁄ + 1 𝑘𝑚⁄ ) ≈ 1 𝑘𝑚⁄ ). once km is known, kd could be determined in correspondence of the different re considered in order to know the function kd(re) as km is re independent. finally, km could be related to the work of solid immersion in the liquid phase. in conclusion, this model allows to study and design drt experiments permitting to enucleate the ratedetermining steps ruling the entire phenomenon. acknowledgements: none conflict of interest: the authors declare no conflict of interest. references [1] handbook of pharmaceutical controlled release technology. d. l. wise editor. marcel dekker inc; new york, ny, usa: 2000. [2] n. bertrand n, j.c. leroux. the journey of a drug-carrier in the body: an anatomophysiological perspective. journal of controlled release 161 (2012) 152 – 163. [3] m. grassi, g. grassi. application of mathematical modeling in sustained release delivery systems. expert opinion on drug delivery 11 (2014) 1299 – 1301. [4] m. grassi, g. grassi, r. lapasin, i. colombo. understanding drug release and absorption mechanisms: a physical and mathematical approach, crc press, boca raton, usa, 2007, 371-479. [5] k. c. kwan, oral bioavailability and first pass effects. drug metabolism and disposition 5 (1997) 1329 – 1336. [6] r. naidu, k.t. semple, m. megharaj, a.l. juhasz, n.s. bolan, s.k. gupta, b.e. clothier and r. schulin. bioavailability: definition, assessment and implications for risk assessment. developments in soil science 32 (2008) 39 – 51. [7] g.l. amidon, h. lennernäs, v.p. shah, j.r. crison. a theoretical basis for a biopharmaceutics drug classification: the correlation of in vitro drug product dissolution and in vivo bioavailability. pharmaceutical research 12 (1995) 413 – 420. [8] j. siepmann, f. siepmann. mathematical modelling of drug dissolution. international journal of pharmaceutics 453 (2013) 12 – 24. [9] t. loftsson, m.e. brewster. pharmaceutical applications of cyclodextrins: basic science and product development. journal of pharmacy and pharmacology 62 (2010) 1607 – 1621. [10] m. davis, g. walker. recent strategies in spray drying for the enhanced bioavailability of poorly water-soluble drugs. journal of controlled release 269 (2018) 110 – 127. admet & dmpk 8(y) (2020) 297-313 mathematical modelling of dissolution rate doi: http://dx.doi.org/10.5599/admet.841 311 [11] s. bertoni, b. albertini, n. passerini. spray congealing: an emerging technology to prepare solid dispersions with enhanced oral bioavailability of poorly water soluble drugs. molecules 24 (2019) 3471 1-21. [12] m.r. gigliobianco, c. casadidio, r. censi, p. di martino. nanocrystals of poorly soluble drugs: drug bioavailability and physicochemical stability. pharmaceutics 10 (2018) 134 2-29. [13] k.t. savjani, a.k. gajjar, j.k. savjani. drug solubility: importance and enhancement techniques. isrn pharmaceutics 2012 (2012) 195727 1 – 10. [14] m. grassi, i. colombo, r. lapasin. drug release from an ensemble of swellable crosslinked polymer particles. journal of controlled release 68 (2000) 97–113. [15] d. hasa, d. voinovich, b. perissutti, m. grassi, a. bonifacio, v. sergo, c. cepek, m. r. chierotti, r. gobetto, s. dall’acqua, s. invernizzi. enhanced oral bioavailability of vinpocetine through mechanochemical salt formation: physico-chemical characterization and in vivo studies. pharmaceutical research 28 (2011) 1870-1880. [16] y. tsume, d. m. mudie, p. langguth, g. e. amidon, g. l. amidon. the biopharmaceutics classification system: subclasses for in vivo predictive dissolution (ipd) methodology and ivinc. european journal of pharmaceutical sciences 57 (2014) 152 – 163. [17] s. baghel, h. cathcart, n. j. o’reilly. polymeric amorphous solid dispersions: a review of amorphization, crystallization, stabilization, solid-state characterization, and aqueous solubilization of biopharmaceutical classification system class ii drugs. journal of pharmaceutical sciences 105 (2016) 2527 2544. [18] s. deshmukh, a. avachat, a. garkal, n. khurana, j. m. cardot. optimization of a dissolution method in early development based on ivivc using small animals: application to a bcs class ii drug. dissolution technologies november (2016) 25 – 41. [19] r. laitinen, j. lahtinen, p. silfsten, e. vartiainen, p. jarho, j. ketolainen. an optical method for continuous monitoring of the dissolution rate of pharmaceutical powders. journal of pharmaceutical and biomedical analysis 52 (2010) 181 – 189. [20] z. li, x he, s. tian, g. feng, c. huang, m. xun, z. wu, y. wang. simultaneous evaluation of dissolution and permeation of oral drug solid formulations for predicting absorption rate–limiting factors and in vitro–in vivo correlations: case study using a poorly soluble weakly basic drug. aaps pharmscitech 20 (2019) 321 1-13. [21] s. saboo,n. a. mugheirbi, d. y. zemlyanov, u. s. kestur, l. s. taylor. congruent release of drug and polymer: a “sweetspot” in the dissolution of amorphous solid dispersions. journal of controlled release 298 (2019) 68–82. [22] v. g. levich. physicochemical hydrodynamics, prentice hall, englewood cliffs, n. j. 1962, pages 60 72. [23] s. kindge, h. wachtel, b. abrahamsson, p. langguth. computational fluid dynamics simulation of hydrodynamics and stresses in the pheur/usp disintegration tester under fed and fasted fluid characteristics. journal of pharmaceutical sciences 104 (2015) 2956 – 2968. [24] d.m. d’arcy, t. persoons. mechanistic modelling and mechanistic monitoring: simulation and shadowgraph imaging of particulate dissolution in the flow-through apparatus. journal of pharmaceutical science 100 (2011) 1102 – 1115. [25] j.h. de smidth, j.g. fokkens, h. grijseels, d.j.a. crommelin. dissolution of theophylline monohydrate and anhydrous theophylline in buffer solutions. journal of pharmaceutical science 75 (1986) 497 – 501. [26] r. censi, m. r. gigliobianco, c. casadidio, p. di martino. changes in the solid state of nicergoline, a poorly soluble drug, under different grinding and environmental conditions: effect on polymorphism and dissolution. journal of pharmaceutical science 108 (2019) 929 – 948. [27] n. coceani, l. magarotto, d. ceschia, i. colombo m. grassi. theoretical and experimental analysis of drug release from an ensemble of polymeric particles containing amorphous and nano-crystalline drug. chemical engineering science 71 (2012) 345 – 355. http://dx.doi.org/10.5599/admet.841 mario grassi et al. admet & dmpk 8(3) (2020) 297-313 312 [28] a. elkhabaz, s. sarkar, g.j. simpson, l.s. taylor. characterization of phase transformations for amorphous solid dispersions of a weakly basic drug upon dissolution in biorelevant media. pharmaceutical research 36 (2019) 174 1 17. [29] a. avdeef. solubility temperature dependence predicted from 2d structure. admet & dmpk 3 (2015) 298 – 344. [30] a. avdeef, e. fuguet, a. llinàs, c. ràfols, e. bosch, g. völgyi, t. verbić, e. boldyreva, k. takács-novák. equilibrium solubility measurement of ionizable drugs – consensus recommendations for improving data quality. admet & dmpk 4 (2016) 117 – 178. [31] c.a.s. bergström, a. avdeef. perspectives in solubility measurement and interpretation. admet & dmpk 7 (2019) 88 105. [32] a.w. hixson, j. h. crowell. dependence of reaction velocity upon surface and agitation i. theoretical consideration. industrial and engineering chemistry 23 (1931) 923 – 931. [33] a.w. hixson, j.h. crowell. dependence of reaction velocity upon surface and agitation ii. experimental procedure in study surface. industrial and engineering chemistry 23 (1931) 1002 – 1009. [34] a.w. hixson, j.h. crowell. dependence of reaction velocity upon surface and agitation iii. experimental procedure in study of agitation. industrial and engineering chemistry 23 (1931) 1160 – 1169. [35] v.p. pedersen, k.f. brown. dissolution profile in relation to initial particle distribution. journal of pharmaceutical science 64 (1975) 1192 – 1195. [36] v.p. pedersen, k.f. brown. size distribution effects in multiparticulate dissolution. journal of pharmaceutical science 64 (1977) 1981 – 1986. [37] v.p. pedersen, k.f. brown. general class of multiparticulate dissolution models. journal of pharmaceutical science 66 (1977) 1435 – 1438. [38] v.p. pedersen, j.w. myrick. versatile kinetic approach to analysis of dissolution data. journal of pharmaceutical science 67 (1978) 1450 – 1455. [39] u. thormann, m. de mieri, m. neuburger, s. verjee, p. altmann, m. hamburger, g. imanidis. mechanism of chemical degradation and determination of solubility by kinetic modeling of the highly unstable sesquiterpene lactone nobilin in different media. journal of pharmaceutical sciences 103 (2014) 3139 – 3152. [40] d. hirai, y. iwao, s. i. kimura, s. noguchi, s. itai. mathematical model to analyze the dissolution behaviour of metastable crustals or amophous druig accompanied with solid-liquid interface reaction. international journal of pharmaceutics 522 (2017) 58 – 65. [41] n. guo, b. hou, n. wang, y. xiao, j. hiang, y. guo, s. zong, h. hao. in situ monitoring and modeling of the solution-mediated polymorphic transformation of rifampicin: from form ii to form i. journal of pharmaceutical sciences 107 (2018) 344 – 352. [42] d. hasa, b. perissutti, d. voinovich, m. abrami, r. farra, s. m. fiorentino, g. grassi, m. grassi. drug nanocrystals: theoretical background of solubility increase and dissolution rate enhancement. chemical and biochemical engineering quarterly 28 (2014) 247 – 258. [43] a. parmar, s. sharma. engineering design and mechanistic mathematical models: standpoint on cutting edge drug delivery. trends in analytical chemistry 100 (2018) 15 – 35. [44] h. nogami, t. nagai, t. yotsuyanag. dissolution phenomena of organic medicinals involving simultaneous phase changes. chemical pharmaceutical bulletin 17 (1969) 499 – 509. [45] g. chiarappa, a. piccolo, i. colombo, d. hasa, d. voinovich, m. moneghini, g. grassi, r. farra, m. abrami, p. posocco, s. pricl, m. grassi. exploring the shape influence on melting temperature, enthalpy, and solubility of organic drug nanocrystals by a thermodynamic model. crystal growth design 17 (2017) 4072 – 4083. [46] b. g. tenchov, t. k. yanev. weibull distribution of particle sizes obtained by uniform random fragmentation. journal of colloid interface science 111, (1986) 1 – 7. admet & dmpk 8(y) (2020) 297-313 mathematical modelling of dissolution rate doi: http://dx.doi.org/10.5599/admet.841 313 [47] s.c. chapra, r.p. canale. numerical methods for engineers 3 rd edition, mcgraw-hill, boston, 1998, pages 233 263. [48] k.y. chow, d.j.w. grant. surface analysis of griseofulvin powders by krypton adsorption: evaluation of specific surface area, bet constant c and polanyi adsorption potential. powder technology, 56 (1988) 209 – 223. [49] s. singh, n. sharda, l. mahajan. spectrophotometric determination of pka of nimesulide. international journal of pharmaceutics 176 (1999) 261–264. [50] m. grassi, n. coceani, l. magarotto. modelling partitioning of sparingly soluble drugs in a two-phase liquid system. international journal of pharmaceutics 239 (2002) 157–169. [51] i. de simone, n. coceani, r. farra, s. m. fiorentino, g. grassi, r. lapasin, d. hasa, b. perissutti, m. grassi, d. voinovich. study on polymer-surfactant interactions for the improvement of drug delivery systems wettability. chemical and biochemical engineering quarterly 26 (2012) 405–415. ©2020 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://dx.doi.org/10.5599/admet.841 http://creativecommons.org/licenses/by/3.0/ prediction of herg inhibition of drug discovery compounds using biomimetic hplc measurements doi: http://dx.doi.org/10.5599/admet.995 191 admet & dmpk 9(3) (2021) 191-207; doi: https://doi.org/10.5599/admet.995 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper prediction of herg inhibition of drug discovery compounds using biomimetic hplc measurements chrysanthos stergiopoulos 1* , fotios tsopelas 1 and klara valko 2 1 laboratory of inorganic and analytical chemistry, school of chemical engineering, national technical university of athens 2 bio-mimetic chromatography ltd. stevenage, herts, united kingdom *corresponding author: chrysanthos stergiopoulos; e-mail: chrisxp3@hotmail.com; tel.: +302107724022 received: april 06, 2021; revised: may 20, 2021; available online: june 06, 2021 abstract the major causes of failure of drug discovery compounds in clinics are the lack of efficacy and toxicity. to reduce late-stage failures in the drug discovery process, it is essential to estimate early the probability of adverse effects and potential toxicity. cardiotoxicity is one of the most often observed problems related to a compound's inhibition of the herg channel responsible for the potassium cation flux. biomimetic hplc methods can be used for the early screening of a compound's lipophilicity, protein binding and phospholipid partition. based on the published herg pic50 data of 90 marketed drugs and their measured biomimetic properties, a model has been developed to predict the herg inhibition using the measured binding of compounds to alpha-1-acid-glycoprotein (agp) and immobilised artificial membrane (iam). a representative test set of 16 compounds was carefully selected. the training set, involving the remaining compounds, served to establish the linear model. the mechanistic model supports the hypothesis that compounds have to traverse the cell membrane and bind to the herg ion channel to cause the inhibition. the agp and the herg ion channel show structural similarity, as both bind positively charged compounds with strong shape selectivity. in contrast, a good iam partition is a prerequisite for cell membrane traversal. for reasons of comparison, a corresponding model was derived by replacing the measured biomimetic properties with calculated physicochemical properties. the model established with the measured biomimetic binding properties proved to be superior and can explain over 70% of the variance of the herg pic50 values. ©2021 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords cardiotoxicity; proarrhythmia; qt prolongation; torsades de pointes; herg inhibition; iam binding, agp binding introduction it has been recognised that the physicochemical properties of drug candidates can be related to the latestage attrition of compounds in the drug development process. the early problems with bioavailability and absorption have been successfully improved by optimising solubility and permeability [1]. recently, toxicity and the lack of efficacy have been identified as the major cause of compound attrition in clinics. together, preclinical toxicity and adverse events account for approximately one-third of all attrition cases [2]. http://dx.doi.org/10.5599/admet.995 https://doi.org/10.5599/admet.995 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:chrisxp3@hotmail.com http://creativecommons.org/licenses/by/4.0/ stergiopoulos, tsopelas and valko admet & dmpk 9(3) (2021) 191-207 192 cardiotoxicity is one of the major causes of concern during clinical trials together with liver and central nervous system (cns) toxicity [3]. it accounts for approximately 27 % of drug development failures, and it does not seem to be restricted to specific high-risk therapeutic areas [4]. one particular focus of cardiovascular adverse effects has been drug-induced arrhythmia or "proarrhythmia" as a consequence of an increased recognition of a relationship between drug-induced qt interval prolongation and torsades de pointes (tdp) [5]. tdp is a dangerous type of proarrhythmia, described as a rare ventricular tachycardia with potential sudden cardiac death, which has led to approximately one-third of all drug withdrawals between 1990 and 2006 [4]. furthermore, 15 % of drugs still on the market can cause qt prolongation, and 4 % are associated with tdp arrhythmia risk. therefore, it is important to recognize a compound’s cardiotoxicity potential early in the drug discovery process, not only because of the associated loss of human life or health, but also because of the enormous financial loss in investment and future revenue potential [6]. the cardiac action potential is regulated by the electrical current flows of ions across cardiomyocyte membranes. many drugs can bind to ion channels, block ionic flow and disrupt the regulation of the action potential [7]. upon blockade, the action potential will rest longer, which results in an increased duration of the relative qt interval that can be observed in electrocardiograph (ecg) traces. disturbing the qt interval may lead to instability in the heart rhythm [8]. patients with long qt syndrome (lqts) exhibit a significant predisposition for the tdp type's cardiac arrhythmia [9]. a prolongation of the cardiac action potential and the qt interval has been associated with loss of function or drug-trapping inside the central cavity of the kv11.1 [10] potassium channel, which is encoded by herg (human ether-a-go-go related gene) and carries the rapid delayed rectifier potassium current (ikr) [7,11]. this channel has a tetrameric structure formed by co-assembly of four identical subunits, each composed of six helical transmembrane domains (denoted s1– s6). the s4 domain contains six positive charges, typical for voltage-gated k + channels [12]. the channel pore is asymmetrical, and its dimensions change depending on its state (open-closed-inactivated). the herg channel has been shown to interact with a wide range of drugs owing to the unique shape of the ligand-binding site, its hydrophobic character and the large vestibule of the channel [13,14]. the risk tolerance for qt prolongation may vary significantly depending on the dose and indication of the drug. documented herg-blocking activity reduces the value of a molecule, as it increases the risk of clinical failure. it has also been estimated that about 60 % of drugs in development exhibit herg block [11]. various attempts have been made to predict the herg inhibition potential of drugs in silico to avoid the synthesis of risky molecules [15]. when studying therapeutic areas and the safety margins regarding the free therapeutic plasma concentration of drugs [16], it was found that a wide variety of drugs, including antiarrhythmic, antibacterial, antipsychotic and pain-killer drugs showed potential risk. as toxicity, just like potency, is dose-dependent, it is essential to relate the herg inhibitory concentration to the drugs' free therapeutic plasma concentration. it was found that a less than 30-fold difference between the therapeutic and inhibitory concentration indicates a high risk. redfern et al. [16] also investigated the relative value of preclinical cardiac electrophysiology data (in vitro and in vivo) for predicting the risk of tdp in drug clinical use. in vivo, telemetry experiments in non-rodents (typically dogs) are the ultimate preclinical test for cardiotoxicity. however, its high cost severely limits its use at the earlier discovery stage [17]. in vitro voltage-clamp techniques are widely used to provide real-time mechanistic information on ion channels [18]. the experiments are performed in mammalian cells transfected with the gene for herg. the overwhelming majority of predictive herg models have been built using mammalian patch clamp data. techniques such as fluorescence-based assays with cells transfected with herg and radioligand (typically dofetilide or mk-499) displacement assays [17] have also been successfully used. since the success of any admet & dmpk 9(3) (2021) 191-207 prediction of herg inhibition by biomimetic measurements doi: http://dx.doi.org/10.5599/admet.995 193 model building depends on the quality of the biological data, it was important to carefully select reliable and informative cardiotoxicity data for a wide variety of drugs in order to develop a continuous model. as the determination of the half-maximal inhibitory concentration (ic50) value requires measurements of inhibitory activities at multiple concentrations, the ic50 information was considered more reliable, and was selected over the inhibition type entries for positive/negative classification. therefore, ic50 of drugs and their log unit values (pic50) in response to herg were collected from the literature. certain physicochemical properties of molecules have been recognized as early indicators of potential problems with early drug discovery compounds [19]. besides lipophilicity [20], solubility [21] and permeability, biomimetic properties such as protein [22] and phospholipid binding [23] can be measured at the early stages of the drug discovery process [24]. the chromatographic technique provides an automated, high throughput and reliable measurement of important properties of the drug discovery compounds [25] that can be used to estimate later stage in vivo properties of compounds such as the volume of distribution, the unbound volume of distribution [26] and the drug efficiency [27]. measurements can also estimate cell membrane partition and skin penetration of compounds based on chromatographic principles [28,29]. various toxicity indicators have already been related to a compound’s physicochemical properties, including herg inhibition and hepatotoxicity [30]. the toxicity potential of compounds has been studied using the immobilised artificial membrane (iam) chromatography [31]. in this work, several chromatography-based techniques were investigated to search for the properties of the compounds that could be used to predict their toxicity, with special emphasis on cardiotoxicity. in this study, herg pic50 data from a set of 90 diverse marketed drugs from a wide range of therapeutic areas and with different physicochemical properties were correlated with their measured biomimetic properties. the measurement of the biomimetic properties of the available drugs was conducted in our laboratories. generic gradient hplc methods were used to determine the chromatographic hydrophobicity index (chi) [32,33] using mobile phases at three different ph values. the protein binding of the compounds was measured using immobilised human serum albumin (hsa) [22], and alpha-1-acid-glycoprotein (agp) stationary phases [34]. the phospholipid-binding was measured using the immobilised artificial membrane (iam) stationary phase [23]. the aim was to establish relationships between the cardiotoxicity potential and the biomimetic binding properties of the drugs and to evaluate their predictive performance. experimental the drugs were obtained from sigma-aldrich (merck) and dissolved in dimethylsulfoxide (dmso) at 10 mm concentration. the 10 µl stock solutions were diluted down to 100 µl before injecting them onto an agilent 1100 hplc system. chi lipophilicity measurements the chromatographic hydrophobicity index (chi) was measured using the compounds' calibrated gradient retention times obtained from an agilent 1100 hplc fitted with a gemini nx-c-18 column (phenomenex ltd macclesfield, uk) with dimensions of 50 x 3 mm and 5 µm particle size. the mobile phase a was either 0.01 m formic acid (ph 2.6), a 50 mm ammonium acetate buffer with an adjusted ph of 7.4 or a 50 mm ammonium acetate buffer with an adjusted ph of 10.5. the mobile phase b was 100 % acetonitrile. the flow rate was 1.0 ml/min, with starting mobile phases of 0.01m formic acid (ph 2.6), 50 mm ammonium acetate adjusted to ph 7.4, and 50 mm ammonium acetate adjusted to ph 10.5 to determine the lipophilicity of the compounds at acidic, neutral and alkaline phs, respectively. an http://dx.doi.org/10.5599/admet.0000 stergiopoulos, tsopelas and valko admet & dmpk 9(3) (2021) 191-207 194 acetonitrile linear gradient was used from 0 to 100 %. the acetonitrile concentration reached 100 % in 3.5 min. the 100 % acetonitrile mobile phase was maintained for an additional 1 min before it was returned to 0 % at 4.7 min. the gradient run cycle time was 6 min, with an additional equilibration time of 1 min before the next injection. the standard deviation in the retention time measurements is ±0.005 min from repeated injections. the retention time values for a standard set of compounds listed in table1 were used to convert the drug retention times to chi values. table 1. the chi values of the calibration set of compounds at three phs [33]. these values were obtained by fitting the isocratically determined chi values and the gradient retention time values. the standard error ranged from 0.1 to 0.8 chi values. chi approximates to the acetonitrile concentration when the compound elutes and can be converted to the octanol/water log d scale using chi log d = 0.0525*chi -1.467 [35]. compound chi at ph 2.6 chi at ph 7.4 chi at ph 10.5 theophylline 17.9 18.4 5.0 phenyl tetrazole 42.2 23.6 16.0 benzimidazole 6.3 34.3 30.6 colchicine 43.9 45.0 43.9 phenyl theophylline 51.7 51.2 51.3 acetophenone 64.1 65.1 64.1 indole 72.1 71.5 72.1 propiophenone 77.4 77.4 77.4 butyrophenone 87.3 87.5 87.3 valerophenone 96.4 96.2 96.4 measurements of plasma protein binding using chiralpak hsa and agp columns the protein binding measurements were carried out on chiralpak hsa and chiralpak agp columns with dimensions of 3 x 50 mm and 5 μm particle size (chiral technologies europe, france). the mobile phase was 50 mm ammonium acetate adjusted to ph 7.4, with a 1.2 ml/min flow rate. the standard isopropanol (ipa) gradient reached 35 % in 3.5 min, which was maintained for 1 min, before returning to 0 % at 4.7 min. the cycle time was 6 min with an additional 1 min re-equilibration time. the racemic warfarin showed separation of its enantiomers at retention times of 3.58 and 3.77 min. the precision of the retention time measurements was within ±0.01 min. the calibration set of compounds and their literature % binding data which were also converted to log k data are shown in table 2. table 2. the protein binding data of the marketed drug molecules that were used to calibrate the retention times obtained on the chiral protein columns (chiralpak hsa and chiralpak agp). the % binding data obtained by equilibrium dialysis were converted to log k data using log k = log (%binding/(101-%binding)). compound name %hsa log k hsa % agp log k agp warfarin 97.9 1.5 83.2 0.7 paracetamol 14.0 -0.8 3.2 -1.5 nizatidine 20.4 -0.6 37.1 -0.2 trimethoprim 37.6 -0.2 46.2 -0.1 propranolol 66.6 0.3 86.0 0.8 carbamazepine 75.0 0.5 65.0 0.3 nicardipine 95.0 1.2 87.0 0.8 indomethacin 99.5 1.8 56.0 0.1 diclofenac 99.8 1.9 60.0 0.2 admet & dmpk 9(3) (2021) 191-207 prediction of herg inhibition by biomimetic measurements doi: http://dx.doi.org/10.5599/admet.995 195 measurements of phospholipid-binding at ph 7.4 using an iam column the phospholipid-binding was measured using an iam pc.dd2 column with dimensions of 100 x 4.6 mm (regis technologies inc., morton grove, il, usa). the gradient retention times were measured using a 50 mm ammonium acetate mobile phase with the ph adjusted to 7.4. the mobile phase flow rate was 1.5 ml/min. the acetonitrile gradient was applied to reach 90 % in 4.75 min. the 90 % acetonitrile concentration was maintained for an additional 0.5 min (to 5.25 min) and returned to 0 % by 5.5 min. the cycle time was 6 min, plus an additional 1 min equilibration time was applied while the injector prepared for the next injection. the gradient retention times were calibrated with the acetophenone homologues for which the chi iam values have been established using isocratic measurements [34]. table 3 shows the calibration set of compounds and their predetermined chi iam values. the chi index on the iam column (chi iam) approximates the acetonitrile concentration in the mobile phase when the compound elutes. chi iam values above 45 indicate strong phospholipid binding. the chi iam values have been converted to log k iam values derived from the chi iam values using equation 1. it represents the equivalent value derived from several isocratic measurements with extrapolated log retention factors to 100 % aqueous mobile phase [23]. the log k iam values can be converted to log k (iam) values and show linear relationships with the octanol/water partition coefficients [26]. equation 2 shows the conversion: log k iam = 0.045* chi iam + 0.42 (1) log k iam = 0.29 e (0.045chi iam +0.42) +0.7 (2) repeating the retention time measurements provided a standard deviation of ±0.005 min. table 3. the calibration set of compounds used on the iam.pc.dd2 hplc column and their predetermined chi iam values. database search for pic50 values assessing the risk of a blockade of the human ether à-go-go related gene potassium channels could greatly facilitate the development of therapeutic compounds and the withdrawal of hazardous marketed drugs. the development of highthroughput automated patch clamp assays has increased the amount of herg-associated data available in public databases [17]. integrated databases are now available using the chembl and pubchem public databases. a large integrated database created by sato et al. [36] has been used in this study. this database curates herg-related data from in vitro assays, such as binding assays (radioligand replacement assay) and electrostatic assays (automated patch-clamp assays), in chembl, pubchem, gostar, nih chemical genomics center (ncgc) and hergcentral and integrates them into the largest database about herg inhibition. ic50 values of the compounds and their pic50 values expressed in molar concentrations were carefully searched and collected from this database, which is freely available at https://drugdesign.riken.jp/hergdb/. data entries using inequality signs, null values and value ranges were excluded. in cases of differences in the reported data for the same compound, mean values were calculated and considered for the model building while outlier values were omitted when the deviation in the results was significant (data points not falling within three standard deviations of the mean). compound chi iam octanophenone 49.4 heptanophenone 45.7 hexanophenone 41.8 valerophenone 37.3 butyrophenone 32.0 propiophenone 25.9 acetophenone 17.2 acetanilide 11.5 paracetamol 2.9 http://dx.doi.org/10.5599/admet.0000 https://drugdesign.riken.jp/hergdb/ stergiopoulos, tsopelas and valko admet & dmpk 9(3) (2021) 191-207 196 calculated physicochemical properties adme boxes v.3.0 software (pharma algorithm) was used to calculate various physicochemical parameters of the investigated compounds such as octanol-water partition (log p) and distribution (log d) coefficients at the ph values of 7.4, hydrogen bond donor (hbd) and acceptor (hba) groups, abraham's hydrogen bond acidity (a) and basicity (b), total polar surface area (tpsa), molecular weight (mw), as well as the molecular fractions of positively charged (f+), negatively charged (f-) and zwitterionic species (fz) at ph =7.4. statistical and visualisation software jmp v13.0 (sas institute inc) and spss 23.0 (ibm spss statistics) were used for the statistical calculations and the stepwise regression analysis. for visualisation, stardrop (optibrium ltd) chemically aware visualisation tools were used to create the plots. results and discussion table 4 contains the collected and quality checked pic50 data of the investigated 90 drug molecules with their generic names and the measured biomimetic hplc data. the drugs used in the training set and test set are listed separately in alphabetical order. table 4. the investigated marketed drugs, their herg pic50 values and the measured biomimetic lipophilicity chi log d at ph 7.4, chi log p, protein binding (log k hsa and log k agp) and phospholipid partition (log k iam). the test set of compounds are listed in the second part of the table in bold. drug pic50 chi log d7.4 chi log p log k hsa log k agp chi iam log k iam charge amitriptyline 5.10 2.65 5.21 0.91 1.05 55.27 2.96 basic apomorphine 5.59 1.17 2.66 1.04 0.83 39.67 2.24 weak base astemizole 6.69 2.68 3.52 1.17 1.12 51.07 2.77 basic atenolol 3.00 -0.40 0.64 -1.08 -1.83 15.75 1.14 basic bepridil 6.42 4.40 6.62 1.27 1.14 57.73 3.08 basic brompheniramine 5.61 1.74 4.27 0.68 0.68 51.69 2.8 basic bupivacaine 5.51 3.06 3.97 -0.10 1.35 41.66 2.34 weak base carbamazepine 3.98 1.22 1.33 0.07 -0.85 23.98 1.53 neutral cetirizine 4.65 1.63 2.04 0.85 0.23 40.56 2.29 zwitterionic chloroquine 5.03 1.18 2.68 0.70 1.12 43.99 2.40 basic chlorpromazine 5.65 -0.58 0.97 1.35 1.79 48.16 2.64 basic ciprofloxacin 3.02 -0.09 0.05 -0.54 -1.19 25.39 1.59 zwitterionic cisapride 6.88 2.65 3.17 1.11 0.84 41.85 2.35 basic citalopram 5.27 1.59 4.32 0.40 0.44 48.80 2.66 basic clarithromycin 4.32 1.9 4.55 -0.24 0.20 49.68 2.70 basic clemastine 6.69 3.18 3.18 1.25 1.18 60.12 3.19 basic clozapine 6.5 2.6 3.05 1.04 0.99 52.42 2.83 weak base desipramine 5.36 1.84 3.49 0.76 0.94 53.10 2.86 basic diltiazem 4.81 2.76 3.12 0.36 0.41 41.95 2.35 basic dolasetron 4.65 1.46 1.46 0.60 0.69 33.20 1.95 weak base domperidone 6.79 1.37 2.13 1.02 0.92 43.47 2.42 weak base doxazosin 6.03 2.09 2.24 1.04 0.72 37.80 2.16 weak base droperidol 6.89 2.26 2.68 0.98 1.69 39.31 2.23 weak base admet & dmpk 9(3) (2021) 191-207 prediction of herg inhibition by biomimetic measurements doi: http://dx.doi.org/10.5599/admet.995 197 table 4. cont’d ebastine 6.27 4.96 6.54 1.62 1.34 58.65 3.12 basic erythromycin 4.43 1.31 2.60 -0.22 0.32 38.79 2.20 basic fexofenadine 4.61 1.28 1.91 0.45 0.02 32.24 1.90 zwitterionic fluvoxamine 5.2 2.01 3.34 0.4 0.59 50.22 2.73 basic glibenclamide 4.29 2.07 3.17 1.52 0.39 32.63 1.93 acidic glimepiride 4.13 2.42 3.25 1.50 0.40 30.76 1.83 basic granisetron 4.65 0.79 2.50 0.34 0.52 45.95 2.53 basic imipramine 5.40 2.05 4.60 0.95 0.96 39.44 2.23 basic isradipine 5.71 3.25 3.29 1.19 1.32 38.76 2.21 neutral ketoconazole 5.49 2.66 2.84 1.18 0.75 37.86 2.16 weak base levobupivacaine 5.48 3.25 4.20 0.13 1.25 41.83 2.34 weak base levofloxacin 2.93 0.54 0.54 -0.34 -0.74 28.55 1.73 zwitterionic lidocaine 3.58 2.65 3.01 -0.71 0.17 32.38 1.91 weak base lomefloxacin 3.93 -0.14 -0.14 -0.10 -1.01 37.89 2.16 zwitterionic loratadine 4.90 3.86 4.25 1.38 1.34 44.30 2.46 weak base lovastatin 5.16 4.09 4.24 1.24 0.87 44.38 2.46 neutral maprotiline 5.17 2.03 4.79 0.77 0.99 56.73 3.03 basic mefloquine 5.45 2.18 4.22 1.38 1.11 41.65 2.29 basic metoclopramide 5.27 0.5 1.91 0.15 0.23 40.38 2.28 basic mibefradil 5.88 2.96 3.99 1.10 1.11 55.18 2.96 basic miconazole 5.68 4.64 4.70 1.54 1.26 53.72 2.89 weak base moxifloxacin 3.83 0.69 0.81 0.54 -0.46 31.61 1.84 zwitterionic nicotine 3.61 0.12 1.20 -0.03 -0.98 13.79 1.05 weak base nifedipine 3.96 2.59 2.71 0.67 0.27 25.12 1.58 basic nitrendipine 5.10 3.15 3.19 1.17 0.69 39.36 2.23 neutral ofloxacin 2.93 0.53 0.53 -0.40 -0.81 26.17 1.62 zwitterionic olanzapine 5.06 1.96 2.76 0.76 0.72 49.74 2.71 weak base ondansetron 5.73 1.33 1.66 0.54 1.30 39.01 2.22 weak base pergolide 6.52 2.47 3.81 0.73 1.04 52.73 2.85 basic perphenazine 5.88 2.92 3.50 1.39 1.13 47.64 2.61 weak base phenytoin 3.71 1.73 1.85 0.69 0.15 29.66 1.78 weak acid pimozide 6.43 3.03 3.89 1.41 1.47 52.31 2.83 weak base prazosin 5.22 1.09 1.29 0.81 0.53 26.32 1.63 weak base procainamide 3.86 -0.59 0.95 -0.71 -0.54 19.30 1.29 basic propafenone 5.96 2.07 3.47 0.84 0.92 46.04 2.54 basic propiverine 5.22 3.84 5.05 1.01 0.94 58.60 3.12 basic propranolol 5.03 1.63 3.03 0.47 0.84 42.08 2.36 basic pyrilamine 5.18 1.72 3.32 0.44 0.49 46.62 2.56 basic quetiapine 5.21 2.55 2.69 0.92 0.97 39.38 2.23 weak base quinidine 5.51 1.25 2.38 0.54 0.68 49.78 2.71 basic risperidone 6.00 1.46 2.18 0.55 0.63 36.31 2.09 weak base ritonavir 5.09 3.25 3.41 1.22 0.64 38.64 2.2 neutral roxithromycin 4.44 1.95 3.70 -0.07 0.60 51.33 2.78 basic saquinavir 5.82 3.31 3.39 1.22 1.59 42.17 2.36 weak base sotalol 3.57 -0.39 -0.35 -0.75 -1.41 21.90 1.43 basic http://dx.doi.org/10.5599/admet.0000 stergiopoulos, tsopelas and valko admet & dmpk 9(3) (2021) 191-207 198 table 4. cont’d sulfamethoxazole 2.66 -0.42 1.03 0.4 -1.57 13.19 1.01 acidic tamsulosin 4.89 1.43 2.01 0.36 0.60 34.32 2.00 basic thioridazine 6.30 3.02 5.46 1.33 1.30 70.24 3.65 basic tolterodine 6.23 2.00 4.95 0.55 0.79 52.45 2.83 basic trifluoperazine 5.72 3.8 5.14 1.52 1.19 68.00 3.55 basic verapamil 5.95 2.66 3.82 1.00 0.58 44.82 2.48 basic alfuzosin 4.28 0.79 1.31 0.08 0.32 34.60 1.98 weak base desloratadine 5.75 1.32 4.03 0.85 1.05 54.56 2.93 basic diphenhydramine 4.94 1.59 3.61 0.33 0.60 35.63 2.06 basic dofetilide 5.72 1.18 1.12 -0.05 0.59 33.10 1.94 weak base flecainide 5.25 1.68 3.01 -0.05 0.16 41.51 2.33 basic fluoxetine 5.73 2.15 3.80 0.99 0.99 54.86 2.94 basic indomethacin 3.72 1.47 3.31 1.69 0.18 30.11 1.81 acidic irbesartan 4.71 1.39 2.00 1.30 1.10 26.14 1.62 zwitterionic lamotrigine 3.55 0.86 0.81 0.16 -0.33 23.25 1.49 weak base metoprolol 3.84 0.77 1.86 -0.72 -0.27 35.38 2.05 basic protriptyline 5.30 1.72 4.70 0.69 1.03 51.59 2.79 weak base sildenafil 4.61 2.59 2.62 0.87 0.42 35.00 2.03 weak base spironolactone 4.64 2.76 2.97 0.78 0.50 36.93 2.12 neutral trazodone 5.30 2.51 2.59 1.09 0.46 32.7 1.93 weak base trimethoprim 3.62 -1.60 0.42 -0.05 -0.07 12.83 1.01 basic ziprasidone 6.44 2.98 2.96 1.42 1.09 47.92 2.62 weak base table 5 contains the calculated physicochemical properties of the investigated compounds. the test set listed separately in alphabetical order in the last part of the table. table 5. the calculated physicochemical properties of the compounds. mw is the molecular weight, hbd and hba are the numbers of h-bond donor and acceptor groups, respectively, tpsa is the topological polar surface area, log p and log d are the logarithm of the calculated octanol/water partition coefficient of the neutral form and the combined ionised form of the molecules at ph 7.4, f+, fand fz are the calculated fractions of the positive, negative and zwitterionic charges at physiological ph (ph 7.4), a and b are the abraham h-bond acidity and bbond basicity parameters. drug mw hbd hba tpsa log p log d f+ ffz a b amitriptyline 277.41 0 1 3.2 5.04 3.70 0.98 0.00 0.00 0.00 1.00 apomorphine 267.32 2 3 43.7 2.49 2.16 0.62 0.00 0.02 0.77 1.10 astemizole 458.57 1 5 42.3 5.70 4.54 0.98 0.00 0.00 0.13 1.64 atenolol 266.34 4 5 84.6 0.16 -1.89 0.99 0.00 0.00 0.69 2.00 bepridil 366.54 0 3 15.7 6.31 4.9 0.97 0.00 0.00 0.00 1.32 brompheniramine 319.24 0 2 16.1 2.88 1.64 0.99 0.00 0.00 0.00 1.02 bupivacaine 288.43 1 3 32.3 3.41 3.73 0.78 0.00 0.00 0.26 1.19 carbamazepine 236.27 2 3 46.3 2.30 2.58 0.00 0.00 0.00 0.53 1.10 cetirizine 388.89 1 5 53.0 1.70 0.34 0.00 0.22 0.78 0.57 1.76 chloroquine 319.87 1 3 28.2 4.63 2.60 1.00 0.00 0.00 0.13 1.29 chlorpromazine 318.86 0 2 31.8 5.35 3.34 0.99 0.00 0.00 0.00 0.94 ciprofloxacin 331.34 2 6 72.9 -1.08 -3.03 0.05 0.03 0.91 0.73 1.85 cisapride 465.94 3 7 86.1 4.20 2.49 0.98 0.00 0.00 0.50 2.17 admet & dmpk 9(3) (2021) 191-207 prediction of herg inhibition by biomimetic measurements doi: http://dx.doi.org/10.5599/admet.995 199 table 5. cont’d citalopram 324.39 0 3 36.3 3.76 0.99 0.99 0.00 0.00 0.00 1.08 clarithromycin 747.95 4 14 182.9 3.16 1.16 0.95 0.00 0.00 0.80 4.49 clemastine 343.89 0 2 12.5 5.79 3.63 0.99 0.00 0.00 0.00 0.97 clozapine 326.82 1 4 30.9 3.32 4.68 0.58 0.00 0.00 0.18 1.44 desipramine 266.38 1 2 15.3 4.90 1.57 1.00 0.00 0.00 0.09 0.91 diltiazem 414.52 0 6 84.4 2.70 1.97 1.00 0.00 0.00 0.00 2.12 dolasetron 324.37 1 5 62.4 2.70 2.66 0.10 0.00 0.00 0.31 1.52 domperidone 425.92 2 7 67.9 4.05 3.49 0.85 0.00 0.00 0.72 1.83 doxazosin 451.47 2 10 112.3 2.07 1.97 0.22 0.00 0.00 0.23 2.60 droperidol 379.43 1 5 52.7 3.50 2.61 0.85 0.00 0.00 0.33 1.67 ebastine 469.66 0 3 29.5 7.55 6.14 0.97 0.00 0.00 0.00 1.41 erythromycin 733.92 5 14 193.9 2.54 0.65 0.95 0.00 0.00 1.05 4.63 fexofenadine 501.65 3 5 81.0 4.35 1.83 0.00 0.04 0.97 1.20 2.12 fluvoxamine 318.33 2 4 56.8 3.63 2.32 0.96 0.00 0.00 0.23 1.14 glibenclamide 494.00 3 8 122.0 4.02 1.83 0.00 0.99 0.00 0.85 2.01 glimepiride 490.62 3 9 133.1 4.25 2.05 0.99 0.00 0.00 0.75 2.15 granisetron 312.41 1 5 50.2 0.79 -1.31 0.99 0.00 0.00 0.26 1.56 imipramine 280.41 0 2 4.8 4.28 2.61 0.99 0.00 0.00 0.00 1.15 isradipine 371.39 1 8 103.6 4.18 1.48 0.00 0.00 0.00 0.13 1.79 ketoconazole 531.43 0 8 69.1 4.34 3.98 0.15 0.00 0.00 0.00 2.22 levobupivacaine 288.43 1 3 32.3 4.35 3.73 0.78 0.00 0.00 0.26 1.19 levofloxacin 361.37 1 7 73.3 -0.24 -2.34 0.05 0.10 0.85 0.57 2.05 lidocaine 234.34 1 3 32.3 2.26 2.44 0.78 0.00 0.00 0.12 1.21 lomefloxacin 351.35 2 6 72.9 -0.80 -3.43 0.05 0.03 0.91 0.73 1.81 loratadine 382.88 0 4 42.4 5.20 4.94 0.00 0.00 0.00 0.00 1.14 lovastatin 404.54 1 5 72.8 4.26 4.40 0.00 0.00 0.00 0.31 1.44 maprotiline 277.41 1 1 12.0 4.85 1.48 1.00 0.00 0.00 0.13 0.68 mefloquine 378.31 2 3 45.2 3.28 1.50 0.99 0.00 0.00 0.38 1.22 metoclopramide 299.80 3 5 67.6 1.40 0.36 0.99 0.00 0.00 0.50 1.63 mibefradil 495.63 1 6 67.5 4.97 2.37 1.00 0.00 0.00 0.35 1.80 miconazole 416.13 0 3 27.1 5.34 5.45 0.15 0.00 0.00 0.00 0.79 moxifloxacin 401.43 2 7 82.1 -0.08 -2.87 0.05 0.01 0.94 0.72 2.04 nicotine 162.23 0 2 16.1 0.75 -0.12 0.88 0.00 0.00 0.00 0.91 nifedipine 346.33 1 8 113.4 3.27 1.12 1.00 0.00 0.00 0.23 1.45 nitrendipine 360.36 1 8 113.5 4.15 1.60 0.00 0.00 0.00 0.13 1.54 ofloxacin 361.37 1 7 73.3 -0.39 -2.34 0.05 0.10 0.85 0.57 2.05 olanzapine 312.44 1 4 59.1 3.00 4.05 0.58 0.00 0.00 0.13 1.45 ondansetron 293.36 0 4 39.8 1.96 1.09 0.88 0.00 0.00 0.00 1.53 pergolide 314.49 1 2 44.3 4.01 2.97 0.92 0.00 0.00 0.31 1.01 perphenazine 403.97 1 4 55.3 3.69 3.86 0.74 0.00 0.00 0.23 1.84 phenytoin 252.27 2 4 58.2 2.15 1.90 0.00 0.10 0.00 0.85 1.00 pimozide 461.54 1 4 35.6 6.30 5.49 0.85 0.00 0.00 0.33 1.44 prazosin 383.40 2 9 107.0 0.45 0.35 0.22 0.00 0.00 0.23 2.17 procainamide 235.32 2 4 58.4 0.88 -0.83 0.99 0.00 0.00 0.50 1.45 propafenone 341.44 2 4 58.6 3.41 1.26 0.99 0.00 0.00 0.29 1.67 http://dx.doi.org/10.5599/admet.0000 stergiopoulos, tsopelas and valko admet & dmpk 9(3) (2021) 191-207 200 table 5. cont’d propiverine 367.48 0 4 38.8 4.01 1.53 1.00 0.00 0.00 0.00 1.31 propranolol 259.34 2 3 41.5 3.09 0.89 0.99 0.00 0.00 0.17 1.42 pyrilamine 285.38 0 4 28.6 3.27 1.63 0.97 0.00 0.00 0.00 1.59 quetiapine 383.51 1 5 73.6 2.27 1.91 0.58 0.00 0.00 0.23 2.01 quinidine 324.42 1 4 45.6 2.64 1.06 0.95 0.00 0.00 0.23 1.81 risperidone 410.48 0 6 61.9 3.04 1.09 0.88 0.00 0.00 0.00 1.70 ritonavir 720.94 4 11 202.3 5.64 5.64 0.00 0.00 0.00 0.88 3.14 roxithromycin 837.04 5 17 216.9 3.79 1.00 0.95 0.00 0.00 1.05 5.12 saquinavir 670.84 6 11 166.8 3.77 3.67 0.22 0.00 0.00 1.46 3.89 sotalol 272.36 3 5 86.8 0.24 -1.35 0.91 0.00 0.08 0.74 1.75 sulfamethoxazole 253.28 3 6 106.6 0.89 -0.85 0.00 0.98 0.00 0.59 1.21 tamsulosin 408.51 2 7 119.3 2.38 1.07 0.95 0.00 0.00 0.59 2.11 thioridazine 370.58 0 2 57.1 5.90 4.17 0.99 0.00 0.00 0.00 1.13 tolterodine 339.51 1 2 23.5 5.98 3.49 1.00 0.00 0.00 0.50 1.08 trifluoperazine 407.50 0 3 35.0 5.03 4.37 0.85 0.00 0.00 0.00 1.50 verapamil 454.60 0 6 64.0 3.83 3.17 0.98 0.00 0.00 0.00 1.89 alfuzosin 389.45 3 9 111.8 -0.23 -0.85 0.78 0.00 0.00 0.84 2.24 desloratadine 310.83 1 2 24.9 4.13 2.42 0.98 0.00 0.00 0.13 0.99 diphenhydramine 255.35 0 2 12.5 3.40 1.82 0.97 0.00 0.00 0.00 0.95 dofetilide 441.57 2 8 121.6 2.41 1.06 0.88 0.00 0.08 0.72 2.16 flecainide 414.34 2 5 59.6 3.78 0.32 0.99 0.00 0.00 0.41 1.32 fluoxetine 309.33 1 2 21.3 4.50 2.44 1.00 0.00 0.00 0.13 0.78 indomethacin 357.79 1 5 68.5 4.27 0.71 0.00 1.00 0.00 0.57 1.57 irbesartan 428.53 1 7 87.1 4.72 2.22 0.00 0.00 1.00 0.56 1.78 lamotrigine 256.09 4 5 90.7 2.63 2.62 0.01 0.00 0.00 0.35 0.96 metoprolol 267.36 2 4 50.7 1.88 -0.43 0.99 0.00 0.00 0.17 1.76 protriptyline 263.38 1 1 12.0 4.91 4.04 0.88 0.00 0.00 0.13 0.73 sildenafil 474.58 1 10 117.5 2.73 2.47 0.37 0.06 0.04 0.26 2.68 spironolactone 416.57 0 4 85.7 2.52 2.52 0.00 0.00 0.00 0.00 1.63 trazodone 371.86 0 6 42.4 3.80 4.60 0.22 0.00 0.00 0.00 1.92 trimethoprim 290.32 4 7 105.5 1.28 0.71 0.47 0.00 0.00 0.28 1.62 ziprasidone 412.94 1 5 46.7 4.6 4.14 0.64 0.00 0.00 0.48 1.65 selection of the test set of compounds constructed toxicity models require external validation to prove their predictive ability. hence, a test set, usually consisting of about 20 % of the entire set, is necessary to evaluate the established models in terms of their predictive performance [37]. for that reason, a principal component analysis using the calculated physicochemical properties of the compounds was performed. by plotting the first two principal components (figure 1), four compounds from each quadrant were selected by taking into account the compounds' therapeutic areas to ensure the test set's diversity. the remaining compounds were used for modelling as the training set. table 6 shows the therapeutic areas of the compounds selected as the test set. admet & dmpk 9(3) (2021) 191-207 prediction of herg inhibition by biomimetic measurements doi: http://dx.doi.org/10.5599/admet.995 201 figure 1 the score plot from the principal component analysis on the calculated properties. compounds served as the test set are marked in red, as follows: (1) dofetilide, (2) sildenafil, (3) irbesartan, (4) flecainide, (5) ziprasidone, (6) trazodone, (7) fluoxetine, (8) protryptiline, (9) diphenhydramine, (10) desloratanide, (11) spironolactone, (12) metoprolol, (13) lamotrigine, (14) indomethacin, (15) trimethoprim, (16) alfuzosin. table 6. compounds selected for the test set and their indications. pharmaceutical drug class alfuzosin alpha blocker desloratadine tricyclic antihistamine diphenhydramine antihistamine dofetilide class iii antiarrhythmic agent flecainide class i antiarrhythmic agent fluoxetine selective serotonin reuptake inhibitor (ssri) antidepressant indomethacin non-steroidal anti-inflammatory (nsaid) irbesartan angiotensin receptor blocker (arb) lamotrigine antiepileptic/anticonvulsant metoprolol class ii antiarrhythmic agent protriptyline tricyclic antidepressant sildenafil phosphodiesterase (pde) inhibitor spironolactone aldosterone antagonist trazodone serotonin antagonist and reuptake inhibitor (sari) antidepressant trimethoprim antibacterial ziprasidone atypical antipsychotic using stepwise regression analysis on the training set, a model was built using only the measured properties listed in table 4. the best model can be described by equation 3: http://dx.doi.org/10.5599/admet.0000 stergiopoulos, tsopelas and valko admet & dmpk 9(3) (2021) 191-207 202 pic 50 = 3.63(±0.37) + 0.88(±0.12)*log k agp + 0.42(±0.18)*log k iam (3) (r=0.850, r 2 =0.722, r 2 adj=0.714, n=74, s=0.553, f=92.1) r is the correlation coefficient, n is the number of compounds, s is the standard error of the estimate, f is the fisher test value. the iam binding and agp binding variables are highly significant and showed a week intercorrelation (r 2 =0.42). therefore, the equation was recalculated using partial least squares regression (pls) and the same coefficients and intercept were obtained. the best model using only in silico calculated properties, the molecular weight (mw), the number of hbond donors (hbd) and acceptor groups (hba), the polar surface (tpsa), the logarithm of the calculated octanol/water partition coefficient of the neutral form and the combined ionised form of the molecules at ph 7.4 (log p and log d), calculated fractions of the positive, negative and zwitterionic charges at physiological ph (f+, f-, fz), and the abraham h-bond acidity and b-bond basicity parameters (a, b) can be described by equation 4. pic50 = 3.94(±0.17) + 0.38(±0.05)*log p 1.49(±0.51)*f(4) (r = 0.740, r 2 = 0.547, r 2 ad j= 0.534, n = 74, s = 0.706, f = 42.9) it was found that log p is correlated better than log d with pic50 and only negatively charged molecular fraction fstands as statistically significant additional physicochemical parameter next to log p. the statistical insignificance of f+ may be attributed to the fact that its positive sign due to ionization counterbalanced with its positive influence to pic50. the models' statistical parameters are much worse when compared to the model using measured agp binding (log k agp) data and membrane partition (log k iam) data. the estimated pic50 values of the test set have been calculated using equation 3 and plotted in figure 2. the test set compounds are marked with larger circles. figure 2 shows the literature herg pic50 data and the back-calculated pic50 data using equation 3. blue circles mark positively charged compounds at ph 7.4; red circles mark negatively charged compounds at ph7.4. the green circles mark neutral compounds at ph 7.4; green circle's shade reflects the presence of weak acidic (lighter green) or weak basic (darker green) groups in the molecules; purple circles indicate compounds with zwitterionic character at ph 7.4 admet & dmpk 9(3) (2021) 191-207 prediction of herg inhibition by biomimetic measurements doi: http://dx.doi.org/10.5599/admet.995 203 it can be seen in figure 2 that the majority of the positively charged compounds show a pic50 value greater than 5 log units in the in vitro herg experiments. acidic and zwitterionic compounds show only weak herg inhibition. it is also interesting to note that the agp binding data showed a strong correlation with the compounds binding to herg channel receptors. the explanation for this may lie in the similarity of the two proteins. it was found [38] that the agp binding site can be featured as a funnel-like structure. the side of the funnel represents a lipophilic region. the funnel's top width provides a steric hindrance for molecules wider than the funnel, while at the narrow end of the funnel are the negatively charged sialic acid residues that bind the positive charges if they fit into the deep end of the funnel. the structure is illustrated in figure 3. the iam binding, which shows the compound's membrane partition, was also significant in the model, which is not surprising as the ion channel receptor is in the membrane. the compound needs to have high membrane affinity to be able to approach the channel. the positive charge also promotes the binding to the negatively charged surface. both the agp and iam stationary phases show strong shape selectivity, which is also essential to herg inhibition. although a wide variety of molecules show high pic50 values the shape of the molecule is very important because of the channel opening's well-defined size. this fact reduces the power of the in silico models if only 2d descriptors are used in the model building. as a result, the 3d description of the molecules would probably enhance the success of in silico models. agp binding herg k + ion-channel blockage figure 3. the similarity of the binding region of agp and herg potassium ion channels. the steric structure and the negatively charged surface of agp and the herg ion channel suggest strong similarities. compounds that block the channel have to penetrate the cell and have a relatively high concentration in the cell membrane where the potassium ion channel can be found [39]. this explains the importance of the membrane-binding properties in the model, as shown in figure 4. validation of both models was performed by predictions of the 16 compounds included in the test set. the results are presented in table 7. it can be seen that the prediction of the test set was superior in the case of the model derived with the measured properties, and the residuals did not exceed double the model error (0.693). on the other hand, predictions from the model derived with the calculated properties exhibited much worse residuals in most cases, with the pic50 predictions of irbesartan and lamotrigine exceeding 1 log unit. http://dx.doi.org/10.5599/admet.0000 stergiopoulos, tsopelas and valko admet & dmpk 9(3) (2021) 191-207 204 figure 4. drug trapping within the k1 channel vestibule. table 7. the experimental and the predicted pic50 values of the test set of compounds using the model described by equation 3. the residuals show the difference between the measured and predicted values. drug model derived with measured properties (log k agp and log k iam) model derived with calculated properties (log p and f-) experimental pic50 predicted pic50 residual predicted pic50 residual alfuzosin 4.28 4.74 0.46 3.85 0.42 desloratadine 5.75 5.78 0.04 5.51 0.24 diphenhydramine 4.94 5.02 0.08 5.23 0.29 dofetilide 5.72 4.96 0.76 4.85 0.87 flecainide 5.25 4.75 0.50 5.38 0.13 fluoxetine 5.73 5.74 0.01 5.65 0.08 indomethacin 3.72 4.55 0.83 4.08 0.36 irbesartan 4.71 5.28 0.57 5.73 1.02 lamotrigine 3.55 3.97 0.42 4.94 1.39 metoprolol 3.84 4.25 0.41 4.65 0.82 protriptyline 5.30 5.71 0.40 5.81 0.50 sildenafil 4.61 4.85 0.24 4.89 0.28 spironolactone 4.64 4.96 0.32 4.90 0.26 trazodone 5.30 4.84 0.46 5.38 0.08 trimethoprim 3.62 3.99 0.37 4.43 0.81 ziprasidone 6.44 5.69 0.75 5.69 0.75 admet & dmpk 9(3) (2021) 191-207 prediction of herg inhibition by biomimetic measurements doi: http://dx.doi.org/10.5599/admet.995 205 conclusions the herg channel inhibition properties of drugs and drug discovery compounds are an important attribute as compounds with strong binding can cause cardiotoxic side effects during clinical trials. early recognition of a compound’s herg inhibition potential is important to avoid the progression of compounds that fail later because of cardiotoxicity. it has been demonstrated that two biomimetic hplc property measurements can be used for screening molecules for herg inhibition potential at an early stage of the drug discovery process. the model is based on the strong similarity between the agp and herg channel structures. both attract positively charged compounds with a significant degree of lipophilicity. both exhibit steric hindrance depending on the size and shape of the molecule being investigated. the membrane partition is also an important parameter as it reveals the membrane affinity of the compounds where the ion channel receptor is located. it has also been shown that two-dimensional physicochemical descriptors cannot provide an acceptable model for estimating the herg pic50 of molecules. conflict of interest: klara valko is the founder of biomimetic chromatography ltd. references [1] i. kola, j. landis. can the pharmaceutical industry reduce attrition rates?, nat. rev. drug discov. 3 (2004) 1–5. https://doi.org/10.1038/nrd1470. [2] m.j. waring, j. arrowsmith, a.r. leach, p.d. leeson, s. mandrell, r.m. owen, g. pairaudeau, w.d. pennie, s.d. pickett, j. wang, o. wallace, a. weir. an analysis of the attrition of drug candidates from four major pharmaceutical companies. nat. rev. drug discov. 14 (2015) 475–486. https://doi.org/10.1038/nrd4609. [3] f. peter guengerich. mechanisms of drug toxicity and relevance to pharmaceutical development. drug metab. pharmacokinet. 26 (2011) 3–14. https://doi.org/10.2133/dmpk.dmpk-10-rv-062. [4] a. casartelli, a. lanzoni, r. comelli, f. crivellente, r. defazio, r. dorigatti, n. fasdelli, i. faustinelli, s. pagliarusco, m. tontodonati, p. cristofori. a novel and integrated approach for the identification and characterization of drug-induced cardiac toxicity in the dog. toxicol. pathol. 39 (2011) 361–371. https://doi.org/10.1177/0192623310390704. [5] n. ferri, p. siegl, a. corsini, j. herrmann, a. lerman, r. benghozi. drug attrition during pre-clinical and clinical development: understanding and managing drug-induced cardiotoxicity. pharmacol. ther. 138 (2013) 470–484. https://doi.org/10.1016/j.pharmthera.2013.03.005. [6] p.j. o’brien. high-content analysis in toxicology: screening substances for human toxicity potential, elucidating subcellular mechanisms and in vivo use as translational safety biomarkers. basic clin. pharmacol. toxicol. 115 (2014) 4–17. https://doi.org/10.1111/bcpt.12227. [7] b.j. ridder, d.j. leishman, m. bridgland-taylor, et al. a systematic strategy for estimating herg block potency and its implications in a new cardiac safety paradigm. toxicol. appl. pharmacol. 394 (2020) 114961. https://doi.org/10.1016/j.taap.2020.114961. [8] c. kramer, b. beck, j.m. kriegl, t. clark. a composite model for herg blockade, chemmedchem. 3 (2008) 254–265. https://doi.org/10.1002/cmdc.200700221. [9] a.o. aptula, m.t.d. cronin. prediction of herg k+ blocking potency: application of structural knowledge, sar qsar environ. res. 15 (2004) 399–411. https://doi.org/10.1080/10629360412331297353. [10] e. raschi, v. vasina, e. poluzzi, f. de ponti. the herg k+ channel: target and antitarget strategies in drug development. pharmacol. res. 57 (2008) 181–195. https://doi.org/10.1016/j.phrs.2008.01.009. http://dx.doi.org/10.5599/admet.0000 https://doi.org/10.1038/nrd1470 https://doi.org/10.1038/nrd4609 https://doi.org/10.1038/nrd4609 https://doi.org/10.2133/dmpk.dmpk-10-rv-062 https://doi.org/10.1177/0192623310390704 https://doi.org/10.1016/j.pharmthera.2013.03.005 https://doi.org/10.1111/bcpt.12227 https://doi.org/10.1016/j.taap.2020.114961 https://doi.org/10.1002/cmdc.200700221 https://doi.org/10.1016/j.phrs.2008.01.009 stergiopoulos, tsopelas and valko admet & dmpk 9(3) (2021) 191-207 206 [11] s. munawar, m.j. windley, e.g. tse, m.h. todd, a.p. hill, j.i. vandenberg, i. jabeen. experimentally validated pharmacoinformatics approach to predict herg inhibition potential of new chemical entities. front. pharmacol. 9 (2018) 1–20. https://doi.org/10.3389/fphar.2018.01035. [12] j.w. warmke, b. ganetzky. a family of potassium channel genes related to eag in drosophila and mammals, proc. natl. acad. sci. u. s. a. 91 (1994) 3438–3442. https://doi.org/10.1073/pnas.91.8.3438. [13] l. du, m. li, q. you. the interactions between herg potassium channel and blockers. curr. top. med. chem. 9 (2012) 330–338. https://doi.org/10.2174/156802609788317829. [14] j.t. milnes, o. crociani, a. arcangeli, j.c. hancox, h.j. witchel. blockade of herg potassium currents by fluvoxamine: incomplete attenuation by s6 mutations at f656 or y652. br. j. pharmacol. 139 (2003) 887–898. https://doi.org/10.1038/sj.bjp.0705335. [15] a.m. aronov. predictive in silico modeling for herg channel blockers. drug discov. today 10 (2005) 149–155. https://doi.org/10.1016/s1359-6446(04)03278-7. [16] w.s. redfern, l. carlsson, a.s. davis, w.g. lynch, i. mackenzie, s. palethorpe, p.k.s. siegl, i. strang, a.t. sullivan, r. wallis, a.j. camm, t.g. hammond. relationships between preclinical cardiac electrophysiology, clinical qt interval prolongation and torsade de pointes for a broad range of drugs: evidence for a provisional safety margin in drug development. cardiovasc. res. 58 (2003) 32– 45. https://doi.org/10.1016/s0008-6363(02)00846-5. [17] r. netzer, a. ebneth, u. bischoff, o. pongs. screening lead compounds for qt interval prolongation. drug discov. today 6 (2001) 78–84. https://doi.org/10.1016/s1359-6446(00)01602-0. [18] c. wood, c. williams, g.j. waldron. patch clamping by numbers. drug discov. today 9 (2004) 434– 441. https://doi.org/10.1016/s1359-6446(04)03064-8. [19] r.j. young, d.v.s. green, c.n. luscombe, a.p. hill. getting physical in drug discovery ii : the impact of chromatographic hydrophobicity measurements and aromaticity. drug discov. today 16 (2011) 822– 830. https://doi.org/10.1016/j.drudis.2011.06.001. [20] a. fernández-pumarega, b. martín-sanz, s. amézqueta, e. fuguet, m. rosés. estimation of the octanol-water distribution coefficient of basic compounds by a cationic microemulsion electrokinetic chromatography system. admet dmpk 8 (2020) 98–112. https://doi.org/10.5599/admet.760. [21] c.a.s. bergstrӧm, a. avdeef. perspectives in solubility measurement and interpretation. admet dmpk 7 (2019) 88–105. https://doi.org/10.5599/admet.686. [22] k. valko, s. nunhuck, c. bevan, m.h. abraham, d.p. reynolds. fast gradient hplc method to determine compounds binding to human serum albumin. relationships with octanol/water and immobilized artificial membrane lipophilicity. j. pharm. sci. 92 (2003) 2236-48. https://doi.org/10.1002/jps.10494. [23] k. valko, c.m. du, c.d. bevan, d.p. reynolds, m.h. abraham. rapid-gradient hplc method for measuring drug interactions with immobilized artificial membrane: comparison with other lipophilicity measures. j. pharm. sci. 89 (2000) 1085-96. https://doi.org/10.1002/15206017(200008)89:8<1085::aid-jps13>3.0.co;2-n. [24] s. bunally, r.j. young. the role and impact of high throughput biomimetic measurements in drug discovery. admet dmpk 6 (2018) 74–84. https://doi.org/10.5599/admet.530. [25] g.h. goetz, m. shalaeva. leveraging chromatography based physicochemical properties for efficient drug design. admet dmpk 6 (2018) 71-73. https://doi.org/10.5599/admet.529. [26] f. hollosy, k. valko, a. hersey, s. nunhuck, g. keri, c. bevan. estimation of volume of distribution in humans from hplc measurements of human serum albumin binding and immobilized artificial membrane partitioning. j. med. chem. 49 (2006) 6958–6971. https://doi.org/10.1021/jm050957i. [27] k. valko, e. chiarparin, s. nunhuck, d. montanari. in vitro measurement of drug efficiency index to aid early lead optimization. j. pharm. sci. 101 (2012) 4155–69. https://doi.org/10.1002/jps.23305. [28] s. scheler, a. fahr, x. liu. linear combination methods for prediction of drug skin permeation. admet dmpk 2 (2015) 199–220. https://doi.org/10.5599/admet.2.4.147. https://doi.org/10.3389/fphar.2018.01035 https://doi.org/10.2174/156802609788317829 https://doi.org/10.1038/sj.bjp.0705335 https://doi.org/10.1016/s1359-6446(04)03278-7 https://doi.org/10.1016/s0008-6363(02)00846-5 https://doi.org/10.1016/s1359-6446(00)01602-0 https://doi.org/10.1016/s1359-6446(04)03064-8 https://doi.org/10.1016/j.drudis.2011.06.001 https://doi.org/10.5599/admet.760 https://doi.org/10.5599/admet.686 https://doi.org/10.1002/jps.10494 https://doi.org/10.1002/jps.10494 https://doi.org/10.1002/1520-6017(200008)89:8%3c1085::aid-jps13%3e3.0.co;2-n https://doi.org/10.1002/1520-6017(200008)89:8%3c1085::aid-jps13%3e3.0.co;2-n https://doi.org/10.5599/admet.530 https://doi.org/10.5599/admet.529 https://doi.org/10.1021/jm050957i https://doi.org/10.1002/jps.23305 https://doi.org/10.5599/admet.2.4.147 admet & dmpk 9(3) (2021) 191-207 prediction of herg inhibition by biomimetic measurements doi: http://dx.doi.org/10.5599/admet.995 207 [29] c. stephen, a. el omri, l. ciesla, cellular membrane affinity chromatography (cmac) in drug discovery from complex natural matrices. admet dmpk 6 (2018) 200–214. https://doi.org/10.5599/admet.535. [30] p.c. ray, m. huggett, p.a. turner, m. taylor, l.a.t. cleghorn et al. spirocycle mmpl3 inhibitors with improved herg and cytotoxicity profiles as inhibitors of mycobacterium tuberculosis growth. acs omega 6 (2021) 2284-2311. https://doi.org/10.1021/acsomega.0c05589. [31] f. tsopelas, c. stergiopoulos, a. tsantili-kakoulidou. immobilized artificial membrane chromatography: from medicinal chemistry to environmental sciences. admet dmpk 6 (2018) 225– 241. https://doi.org/10.5599/admet.553. [32] c.m. du, k. valko, c. bevan, d. reynolds, m.h. abraham. rapid gradient rp-hplc method for lipophilicity determination: a solvation equation based comparison with isocratic methods. anal. chem. 70 (1998) 4228-4234. https://doi.org/10.1021/ac980435t. [33] k. valko, c.m. du, c. bevan, d.p. reynolds, m.h. abraham, rapid method for the estimation of octanol/water partition coefficient (log poct) from gradient rp-hplc retention and a hydrogen bond acidity term (∑α2 h). curr. med. chem. 8 (2001) 1137–1146. https://doi.org/10.2174/0929867013372643. [34] k.l. valko, s.p. teague, c. pidgeon. in vitro membrane binding and protein binding (iam mb/pb technology) to estimate in vivo distribution: applications in early drug discovery. admet dmpk 5 (2017) 14-38. https://doi.org/10.5599/admet.5.1.373. [35] k. valko. handbook of analytical separations volume 8, separation methods in drug synthesis and purification. 2nd ed., elsevier 2020 p.687. isbn: 9780444640703. [36] t. sato, h. yuki, k. ogura, t. honma. construction of an integrated database for herg blocking small molecules. plos one 13 (2018) 1–18. https://doi.org/10.1371/journal.pone.0199348. [37] c. stergiopoulos, d. makarouni, a. tsantili-kakoulidou, m. ochsenkühn-petropoulou, f. tsopelas. immobilized artificial membrane chromatography as a tool for the prediction of ecotoxicity of pesticides. chemosphere 224 (2019) 128–139. https://doi.org/10.1016/j.chemosphere.2019.02.075. [38] r. kaliszan, a. nasal, m. turowski. quantitative structure-retention relationships in the examination of the topography of the binding site of antihistamine drugs on α1-acid glycoprotein. j. chromatogr. a 722 (1996) 25–32. https://doi.org/10.1016/0021-9673(95)00523-4. [39] m. tristani-firouzi, j. chen, j.s. mitcheson, m.c. sanguinetti. molecular biology of k+ channels and their role in cardiac arrhythmias. am. j. med. 110 (2001) 50–59. https://doi.org/10.1016/s00029343(00)00623-9. ©2021 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://dx.doi.org/10.5599/admet.0000 https://doi.org/10.5599/admet.535 https://doi.org/10.5599/admet.535 https://doi.org/10.1021/acsomega.0c05589 https://doi.org/10.5599/admet.553 https://doi.org/10.1021/ac980435t https://doi.org/10.2174/0929867013372643 https://doi.org/10.2174/0929867013372643 https://doi.org/10.5599/admet.5.1.373 https://doi.org/10.1371/journal.pone.0199348 https://doi.org/10.1016/j.chemosphere.2019.02.075 https://doi.org/10.1016/0021-9673(95)00523-4 https://doi.org/10.1016/s0002-9343(00)00623-9 https://doi.org/10.1016/s0002-9343(00)00623-9 http://creativecommons.org/licenses/by/3.0/ optimization of experimental conditions for skin-pampa measurements martí rosés et al. admet & dmpk 8(1) (2020) 16-28 admet & dmpk 8(1) (2020) 16-28 optimization of skin-pampa measurements admet & dmpk 8(1) (2020) 16-28; doi: http://dx.doi.org/10.5599/admet.761 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper optimization of experimental conditions for skin-pampa measurements sara soriano-meseguer1, elisabet fuguet1,2, adriana port3, martí rosés1,* 1 departament de química analítica i institut de biomedicina, universitat de barcelona, martí i franquès 1-11, 08028 barcelona, spain 2 serra húnter programme, generalitat de catalunya, 08002 barcelona, spain 3 esteve pharmaceuticals, drug discovery and preclinical development, parc científic de barcelona, baldiri reixac 4-8, 08028 barcelona, spain *corresponding author: e-mail: marti.roses@ub.edu; tel.: +34-934039275; fax: +34-934021233, editorial board member received: november 29, 2019; revised: february 15, 2020; available online: march 04, 2020 abstract in recent years, the parallel artificial membrane permeability assay (pampa) has been extended for prediction of skin permeation by developing an artificial membrane which mimics the stratum corneum structure, skin-pampa. in the present work, the different parameters affecting skin-pampa permeability, such as incubation time and stirring, have been studied to establish ideal assay conditions to generate quality data for a screening of active pharmaceutical ingredients (api) in early stage drug discovery. another important parameter is membrane retention, which shows dependence on lipophilicity when compounds are in their neutral form. furthermore, the stability of the membrane has been investigated at different ph values, especially at basic phs. finally, a good correlation between human skin permeability and skin-pampa permeability, with a large dataset (n = 46), has been established. the optimized assay conditions were an incubation time of 4 hours with stirring in a ph below 8. with all these considerations the thickness of the aqueous boundary layer is decreased as much as possible and the membrane stability is guaranteed. ©2020 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords skin-pampa; permeation; optimization; membrane retention; membrane stability introduction transdermal administration is considered as an alternative route to conventional oral delivery of drugs. transdermal delivery offers significant advantages over oral administration: minimal first-pass metabolism, avoidance of the adverse effects in the gastrointestinal environment and the ability to provide a controlled and prolonged drug release [1]. despite these advantages, the structure of the skin mainly composed of the stratum corneum (sc), the outermost layer which acts as the main penetration barrier, the viable epidermis and dermis, imposes a clear obstacle to the topical delivery of drugs into the systemic bloodstream. estimation of skin permeation has become crucial in the pharmaceutical and cosmetic industries. among the in vitro models used to determine the skin permeation, franz diffusion cell method applying human skin as a membrane is the most relevant [2], but this method is laborious, costly and ethically questionable. besides, it suffers low intraand inter-laboratory reproducibility. the parallel artificial membrane permeability assay (pampa) was developed by kansy and co-workers [3] for the fast determination of the permeability through passive diffusion. this technique is crucial in the early stage of drug discovery and it has many advantages like low cost and high-throughput. the first published models allowed the prediction of gastrointestinal absorption (git-pampa) [4–6]. then, other models were published for modelling the blood-brain barrier [7] and also for the estimation of skin permeation [8]. later, sinkó and co-workers developed the skin-pampa methodology consisting of a new membrane containing synthetic certramides, analogues of the ceramides found in the stratum corneum [9,10]. certramides are cheaper alternatives to natural ceramides with the potential to prolong the storage time. although certramides are structurally different from ceramides, their comparable molecular mass and hydrogen bond acceptor/donor capacity enables them to act as the lipid constituents in the pampa sandwich membrane, together with cholesterol, stearic acid and silicone oil [9,10]. skin-pampa is a model that offers high reproducibility, is more cost-effective and less laborious than other in vitro skin experiments, and has demonstrated a high prediction capability with a good correlation with the human skin penetration data [10]. git pampa assay conditions such as permeation time, assay ph, stirring, use of cosolvents and selection of detection techniques have been studied during years and optimized to generate high quality and relevant data [5]. for skin-pampa, only a few articles have been published [10,11] and most of them deal with liquid or semi-solid formulations [12,13] and transdermal patches [14]. however, permeability studies of compounds in solution are very important when the permeation properties of api are investigated. the protocol of these studies depends on the physico-chemical parameters of the api and the composition of the membrane. since the skin-pampa membrane contains different components and is more resistive compared to other pampa membranes, protocol differences as incubation time are expected. the reason for this is that pampa membranes are composed mostly of lipophilic components and hence lipophilic compounds can cross them quickly. however, skin-pampa consists of lipophilic and hydrophilic moieties, so permeable compounds are likely to cross both lipophilic and hydrophilic domains [10]. another parameter that also depends on the composition membrane and physico-chemical properties of the api is the unstirred water layer (uwl), which is formed around both sides of the lipophilic membrane. the permeation process in pampa assays may be limited by the uwl, especially for lipophilic compounds. the water layer can act as rate-limiting transport giving smaller permeability values. to solve this, the stirring decreases the uwl thickness and thus the resistance of the water layer less of a contribution to the measured permeability [15]. the use of the oblate stir disks (flippers), rotating in a horizontal axis parallel to the plane of the microtitre plates, proved to be the most efficient stirring mechanism ever reported in microtitre plate permeation assays [15]. another feature to consider is the stability of the membrane. human skin ph is normally slightly acidic, values between 4 and 6, while the body’s internal environment mainly maintains a neutral ph. fatty acids found in the lipid bilayer exist in the neutral or ionic forms depending on the ph, so they also contribute to the ionization of the membrane. at the sc surface, ph 5 causes minimal head‐group repulsion and promotes a bilayer structure. the ph 7 in the innermost sc layers produces 90 % ionization of the fatty acids leading to head‐group repulsion. an increase of the ph leads to increase head‐group repulsion, disturbing epidermal lipid lamella and thus impairing barrier function [16]. as most of the drugs are weak acids or weak bases, it is relevant to study the influence of the ph on permeability. to this end, it is necessary to study the stability of the skin-pampa membrane at different ph values, especially at basic phs. in order to establish the optimal assay conditions to generate quality data for the screening of api in early stage drug discovery, the purposes of this work are to study the different parameters affecting skin-pampa permeability, such as incubation time and stirring, and also to check the stability of the membrane with ph. materials and methods reagents acetonitrile lichrosolv grade was purchased from merck (darmstadt, germany). formic acid was obtained from scharlau (sentmenat, spain). dimethylsulphoxide was from carlo erba (milano, italy). water was purified by a milli-q deionizing system from millipore (billerica, ma, usa) with a resistivity of 18.2 mω. most solutes employed were purchased from sigma-aldrich (steinheim, germany), fluka analytical vwr (west chester, pa, usa), riedel-de haën (seelze, germany), merck (darmstadt, germany), carlo erba (milano, italy) and baker (center valley, pa, usa). some drugs were synthesized in esteve (barcelona, spain). the concentrated prisma httm solution was used to prepare the buffer solutions. this solution is a universal buffer designed by pion inc (billerica, ma, usa) and is formed by several compounds with pka values evenly spaced to produce a constant buffer capacity in the range ph 3-10. the ionic strength of the prisma httm is about 10 mm. a hydration solution from pion inc. (billerica, ma, usa) was used to rehydrate the artificial skin membrane. the skin-pampa plates, with a membrane composed by certramides, cholesterol, stearic acid, and silicone oil, were also obtained from pion inc. (billerica, ma, usa). instruments ph measurements were done with a combined crison 5202 electrode in a crison 2001 ph meter (hach lange spain, l’hospitalet de llobregat, spain). the electrode system was calibrated with the ordinary aqueous buffers of ph 4.01 and 7.00 (25 °c). permeability measurements were made with the pampa explorer permeability assay instrument from pion inc (billerica, ma, usa). this instrument is composed of the gut-boxtm and the tempplate. the gut-boxtm is a mechanical device used for the pampa assay to decrease the permeation time and reduce the unstirred water layer (uwl) thickness that is always present. the tempplate is used for the temperature control during plate incubation. chromatographic measurements were performed with a waters (milford, ma, usa) i-class uplc with diode array detector. instrument control and processing was performed by empower. the column used for the determinations was an acquity uplc beh c18 (50 x 2.1 mm, 1.7 μm). skin-pampa method before permeation assays, the top part of the skin-pampa sandwich, which contains the membrane, was hydrated overnight with the hydration solution. the samples were dissolved in diluted prisma httm buffer solution at several ph values: 25 ml of concentrated prisma httm was diluted with water to a final volume of 1 l and then, different solutions were prepared by ph adjustment between 3 and 10 with 0.5 m naoh (merck). the concentration of the sample solutions was 50 µm (containing 0.5 % v/v dmso). skin-pampa assays were carried out under gradient-ph conditions to mimic the ph change between the stratum corneum and the underlying epidermis and dermis. for this reason, the donor compartment ph was varied from 3 to 10 and the acceptor compartment ph was maintained at ph 7.4. this gradient-ph state is the first sink condition in skin-pampa. the double-sink condition that is usually used in git-pampa consists in taking advantage of chemical scavengers in the receiver compartment to make permeation of lipophilic compounds across the membrane unidirectional. this procedure simulates the situation present in the body where blood flow and serum proteins constantly shift the concentration gradient to favor absorption. however, this double-sink condition is not used in skin-pampa since this additional shift is not observed through the skin. before performing the skin-pampa sandwich, the donor compartment (or bottom plate) was prefilled with 180 µl (stirred assay) or 200 µl (unstirred assay) of sample solutions and the acceptor compartment (or top plate) was filled with 200 µl of prisma httm buffer solution at ph 7.4. the donor volume is decreased in stirred assay since the stirring bars have a volume of 20 µl that can cause overflow. as it is mentioned before, the gut-boxtm was used to stir effectively. the skin-pampa sandwich was incubated at 32 °c. after the permeation time was reached, the plates were separated and the compound concentration in acceptor, donor and reference (initial sample solution) was determined using uplc-dad. chromatographic conditions were: formic acid 0.1 % and acetonitrile as mobile phase, flow rate of 0.8 ml/min, linear gradient elution (linear gradient from 2 % to 98 % of acetonitrile in 2.5 minutes), injection volume of 5 μl and the detection by dad. 3 to 5 replicate measurements were done per compound and ph, and every well-plate contained only one compound. calculation methods the skin-pampa permeability was calculated through pampa equations. taking into account the membrane retention (mole fraction of the sample that can be lost in the membrane) under gradient-ph conditions, the equations are the following [17]: (1) (2) where pe is the effective permeability coefficient (cm/s), vd and va are the volume of solution in the donor side (180 µl stirred assay or 200 µl unstirred assay) and acceptor side (200 µl), respectively, a is the membrane area (0.3 cm2), t is the incubation time of the experiment (s), τss is the lag time (s) [τss=(54·rm+1)·60 s], εa is the apparent membrane porosity (0.76), cd(t) is the concentration in the donor side at time t, cd(0) is the initial concentration in the donor side, ca(t) is the concentration in the acceptor side at time t, rm is the membrane retention and ra is the sink asymmetry ratio (gradient-ph-induced), defined as: (3) when the ph is different in the two sides of the membrane, a gradient-ph is created and the permeation of ionizable molecules can be altered. this gradient-ph implies two different permeability coefficients, one denoted by the superscript (d→a), associated with donor to acceptor flux, and the other denoted by the superscript (a→d), corresponding to the reverse-direction flux. as equation (3) has two unknowns, pe(a→d) and pe(d→a), the following method is used to solve the equation: at least two assays are done, one with gradient-ph and the other with iso-ph, that is, the same ph at both compartments. for iso-ph, pe(a→d) = pe(d→a). therefore, pe(a→d) can be solved directly using the iso-ph equation: (4) where rv is the aqueous compartment volume ratio, defined as: (5) then, eq. (1) is iteratively solved for pe(d→a). initially, ra is assumed to be rv, but with each iteration, the ra estimation is improved by using the calculated pe(d→a). the process continues until self-consistency is reached within the precision required (0.001). the solver utility from microsoft excel was used for the iterative process. results and discussion optimization of skin-pampa assay conditions: incubation time and stirring to evaluate the different parameters that affect skin-pampa permeability and to establish general assay conditions for skin-pampa determinations in solution, the permeability of 9 drugs with different acid-base properties, lipophilicity, and in vitro franz cell human skin permeability values (kp, units in cm/s) was investigated. to assure the neutral form, acidic compounds (flurbiprofen, ibuprofen, naproxen and 5-fluorouracil) were dissolved in prisma httm buffer solution at ph 3.0. aminopyrine, a basic compound with a aqueous pka value of 5, was dissolved in the buffer solution at ph 7.0. the neutral compounds (progesterone, griseofulvin, digitoxin and hydrocortisone) were dissolved at ph 3.0. the skin-pampa sandwiches were incubated (with and without stirring) for 30 min, 4 h and 24 h. the logarithm of skin-pampa permeability values (log pe) and the corresponding membrane retention are presented in table 1. this table also shows the logarithm of the octanol-water partition coefficient (log po/w) and the logarithm of human skin permeation coefficient (log kp) for each compound. the log kp data was obtained from zhang et al. database [18]. this database provides an extensive and carefully examined data set, where experimental log kp data from literature is corrected for ionization in water and for the temperature at 37 °c. table 1. log pe values of the neutral form of the drugs, obtained from skin-pampa assays with and without stirring at different incubation times. log pe with stirring without stirring solutes log kp [18] log po/w [19] pka 30 min 4 h 24 h 30 min 4 h 24 h 5-fluorouracil -6.82 -0.89 7.86 a -5.77(±0.02) -6.16(±0.03) -6.47(±0.22) -6.07(±0.06) aminopyrine -6.55 0.80 5.00 b -5.88(±0.07) -5.65(±0.01) -5.17(±0.06) -5.79(±0.06) -5.46(±0.06) -5.22(±0.02) digitoxin -8.15 2.83 -6.42(±0.27) -6.34(±0.13) flurbiprofen -4.72 4.16 4.19 c -2.98(±0.13) -3.69(±0.02) -3.54(±0.06) -3.85(±0.03) griseofulvin -6.44 2.18 -5.61(±0.01) -5.25(±0.10) -5.12(±0.02) -5.65(±0.05) -5.33(±0.07) -5.20(±0.05) hydrocortisone -7.22 1.61 -6.13(±0.03) -6.04(±0.07) -6.30(±0.18) -6.14(±0.08) ibuprofen -4.58 3.50 4.43 d -2.71(±0.04) -3.61(±0.05) -3.59(±0.08) -3.65(±0.04) naproxen -4.97 3.34 4.28 e -3.96(±0.12) -4.19(±0.02) -3.99(±0.08) -4.24(±0.06) progesterone -4.90 3.87 -3.25(±0.25) -4.11(±0.10) -4.59(±0.15) -4.81(±0.06) -4.88(±0.22) -4.69(±0.12) a from reference [20], b from reference [19], c from reference [21], d from reference [22], e from reference [23] first, to study the effect of the incubation time and stirring in the determination of the skin permeability through the skin-pampa permeability, the log kp values were correlated to the log pe obtained from skin-pampa assays at each incubation time. the correlations obtained are presented in figure 1 a-c. each figure contains two different correlations, one for the stirred assays and another for unstirred assays. as it can be observed, the number of compounds that can be determined in 4 hours (n=8) is greater than in 30 minutes (n=6) and 24 hours (n=6). for some compounds, such as 5-fluorouracil, hydrocortisone and digitoxin, whose log kp values are quite low, an incubation time of 30 minutes is not enough to reach the steady state, therefore their skin-pampa permeability values cannot be evaluated. after 4 hours of incubation, all drugs can be determined except digitoxin which does not arrive at the steady state. digitoxin is considered very little permeable due to its very low in vitro skin permeability value (log kp = -8.15). after 24 hours of incubation, digitoxin can reach the steady state and its skin-pampa permeability values can be determined. however, compounds such as flurbiprofen, ibuprofen and naproxen (with high log kp values) cannot be determined. when a compound is highly permeable, long incubation times under gradient-ph conditions provoke that the donor and acceptor compartment concentrations achieve equilibrium values and hence the whole sample of the donor compartment moves to acceptor compartment due to sink conditions, making difficult to determine the permeability values. figure 1. effect of incubation time and stirring in the determination of the skin permeability (log kp) through the skin-pampa permeability (log pe). incubation time: 30 minutes (a); 4 hours (b); 24 hours (c); stirred assay (●); non-stirred assay (⨉). relative to stirring and non-stirring experiments, table 1 shows the results obtained at different incubation times. at 30 minutes of incubation time flurbiprofen, progesterone and ibuprofen show great log pe values in stirred assays compared to unstirred ones. this difference can be attributed to the presence of uwl due to the lipophilic character of the compounds (log po/w > 3). in this case, uwl acts as rate-limiting transport giving smaller permeability values. at 4 hours of incubation, log pe values obtained from experiments with and without stirring are almost the same except for progesterone (log pe equal to – 4.11 and -4.88, respectively) and 5-fluorouracil (log pe equal to 5.77 and -6.47, respectively). for progesterone, this difference can be justified by the presence of uwl or the high membrane retention values. the reason for 5-fluorouracil is unknown since it is a very hydrophilic compound and therefore the difference of values cannot be attributed to uwl effect. in general, it seems that with 4 hours of incubation time the aqueous boundary layer does not have much effect in most compounds. at 24 hours, apart from flurbiprofen, ibuprofen, and naproxen which couldn’t be evaluated, the results obtained with and without stirring are practically the same. the membrane retention values can sometimes be very high depending on the composition of the pampa membrane. for example, membranes made of 2 % dopc (dioleoylphosphatidylcholine) dissolved in dodecane can have rm values higher than 0.80 [4]. in the case of the skin-pampa membrane, retention is in general low and goes from 0 to 0.30 for most of the analyzed compounds (see table 2), except for progesterone whose values are very high (0.33-0.88). for some compounds this parameter can depend on the incubation time. for 5-fluorouracil, aminopyrine, and hydrocortisone this parameter is almost null; minimum retention is observed only after 24 h of incubation. note that these compounds are quite hydrophilic and are almost neutral at the ph of the determination. instead, for some other compounds such as digitoxin, griseofulvin and progesterone (all neutral compounds but with higher log po/w values) rm clearly increases with incubation time. in the case of flurbiprofen, ibuprofen and naproxen, rm decreases as the incubation time increases. these compounds are completely ionized at the ph of the acceptor compartment, and after some time they are all accumulated in there, showing negligible retention in the membrane. in general terms, this factor is independent of stirring or not the solutions as observed for all compounds except for progesterone, which retention factor increases with the stirring use. table 2. membrane retention values of the neutral form of the drugs, obtained from skin-pampa assays with and without stirring at different incubation times. rm with stirring without stirring solutes log kp [18] log po/w [19] pka 30 min 4 h 24 h 30 min 4 h 24 h 5-fluorouracil -6.82 -0.89 7.86 a 0.00 0.04(±0.04) 0.08(±0.03) 0.00 0.01(±0.03) 0.05(±0.02) aminopyrine -6.55 0.80 5.00 b 0.00 0.00 0.05(±0.02) 0.00 0.04(±0.01) 0.05(±0.02) digitoxin -8.15 2.83 0.00 0.22(±0.04) 0.29(±0.05) 0.00 0.09(±0.03) 0.25(±0.02) flurbiprofen -4.72 4.16 4.19 c 0.18(±0.03) 0.00 0.00 0.13(±0.02) 0.00 0.00 griseofulvin -6.44 2.18 0.06(±0.01) 0.19(±0.01) 0.24(±0.02) 0.03(±0.02) 0.16(±0.01) 0.20(±0.02) hydrocortisone -7.22 1.61 0.02(±0.00) 0.01(±0.01) 0.08(±0.03) 0.00 0.04(±0.01) 0.07(±0.02) ibuprofen -4.58 3.50 4.43 d 0.22(±0.03) 0.00 0.00 0.19(±0.07) 0.09(±0.03) 0.00 naproxen -4.97 3.34 4.28 e 0.13(±0.03) 0.00 0.00 0.09(±0.02) 0.04(±0.02) 0.00 progesterone -4.90 3.87 0.50(±0.04) 0.74(±0.01) 0.88(±0.01) 0.33(±0.01) 0.62(±0.04) 0.83(±0.02) a from reference [20], b from reference [19], c from reference [21], d from reference [22], e from reference [23] the membrane retention values at 24 h incubation time have been correlated to the log po/w to check if membrane retention is related to lipophilicity. flurbiprofen, ibuprofen, and naproxen are excluded from the correlation due to the reasons above mentioned. a sigmoidal relationship between rm and log po/w, which can be explained by equation 6 (see appendix), can be observed for the rest of the drugs, which are mainly in its neutral form. (6) figure 2. relationship between membrane retention (rm) and lipophilicity (log po/w) at 24 h incubation time without stirring. in this equation d is the distribution coefficient between water and membrane (ph dependent), vwater is the volume of the aqueous compartment, and vmembrane is the volume of the membrane. figure 2 shows the fit of equation 6 (using po/w of the compounds instead of log d) to the experimental data (24 h incubation time without stirring). good statistics are obtained (r2 = 0.951, sd = 0.066, f = 98), with a value of 3.19 ± 0.11 for the parameter log (vwater/vmembrane). this is a good fit given that the true equilibrium may not have been achieved, ionisation of some compounds has been neglected, and octanol is expected to be more lipophilic than the membrane. from correlations established in figure 1 a-c, it can be observed that the slopes are very close and very low data dispersion is observed. hence, the assay conditions using an incubation time of 4 hours (figure 1b) and taking advantage of stirring to decrease as much as possible the thickness of the aqueous boundary layer are considered the most appropriate. working under these conditions permits to evaluate the greatest number of compounds, despite in some cases, such as digitoxin, longer incubation might be needed. study of membrane stability the stability of the skin-pampa membrane at basic ph values was studied through measurement of the permeability of 11 compounds (3 acids, 5 bases and 3 neutrals). the skin-pampa membrane was hydrated with diluted prisma httm buffer solution at ph 7, 8, 9 and 10 during 30 min and 4 hours, to be consistent with incubation times selected in section 3.1. then, a standard permeation assay was carried out with an incubation time of 4 h and stirring at ph 5 in the donor compartment and ph 7.4 in the acceptor compartment. simultaneously, permeation assays were also performed without the previous soaking treatment at basic ph. the obtained results are presented in figure 3, which shows the skin-pampa log pe values obtained with and without the previous treatment. each subfigure corresponds to a compound and the straight line inside the figure is the log pe value in the untreated membrane. an increase of permeability values at ph 9 and 10 is shown in figure 3 for almost all the compounds, which points out the lack of stability of the membrane in this ph range. in some cases, such as aminopyrine, some distortion is already noted at ph 8. as described in the literature [16], solutions at ph 7 or higher produce a change of membrane packaging in human skin. however, in the skin-pampa membrane loose of stability is not observed up to ph 8, probably due to the different membrane composition. in most cases, the permeability values increase for both 30 minutes and 4 hours of treatment, meaning that at 30 minutes the stability of the membrane is already altered by the basic ph. griseofulvin and warfarin have a slight increase at ph 9 and 10. the values of indomethacin, 2-toluidine and sufentanyl, however, remain constant. as a general trend, it seems that this problem is more important when the compound is less permeable (log pe below -5) regardless of the nature and ionization of the compound. in conclusion, although the prisma httm universal buffer solution suggests a working ph range between 3 and 10, it is advisable not to exceed ph 8 to avoid damaging the skin-pampa membrane. in that cases where permeability must be determined at ph values higher than eight (for example to determine the log pe of the neutral forms of bases with pka values higher than 7) alternative methods should be used, such the estimation equations proposed by zhang et al. [18] or estimation through chromatographic measurements [24]. correlation with human skin permeation data the skin-pampa permeability (pe) of a large acid-base drugs dataset (n=46) obtained in the optimized assay conditions previously discussed has been correlated with literature skin permeability (kp). the log kp data were collected from zhang et. al. database [18]. table 3 shows the log kp and log pe values of the compounds used in the correlation. each compound was measured at the ph corresponding to the neutral form. figure 4 plots the log kp vs. the log pe values whose correlation is presented in eq. (7): log kp = 0.143(±0.403) + 1.156(±0.080) log pe (7) n = 46 r2 = 0.826 sd = 0.434 f = 208 a good correlation between human skin and skin-pampa data and adequate statistical parameters have been obtained. figure 3. influence of treatment time and ph in skin-pampa permeability. treatment time: 30 minutes (▲); 4 hours (∎). table 3. log kp and log pe experimental values of the neutral form of the drugs. both kp and pe in cm/s units. solutes log kp [18] log pe 2,4-dichlorophenol -4.30 -3.92± 0.00 2-isopropyl-5-methylphenol (thymol) -4.35 -4.01± 0.00 2-nitro-p-phenylenediamine -6.66 -5.25±0.02 3-methylphenol (m-cresol) -4.89 -4.33±0.01 4-amino-2-nitrophenol -5.91 -4.59±0.02 4-chlorophenol -4.52 -4.270.02± 4-ethylphenol -4.53 -4.19±0.02 4-hydroxyphenylacetamide -6.89 -6.07±0.08 4-hydroxy-methylphenylacetate -5.26 -5.07±0.09 4-nitrophenol -5.33 -4.91±0.02 5-fluorouracil -6.82 -5.77±0.02 8-methoxypsoralen -5.12 -4.30±0.03 5,5-diethylbarbituric acid (barbital) -7.29 -5.75±0.05 5-ethyl-5-phenylbarbituric acid (phenobarbital) -6.68 -6.05±0.01 benzoic acid -5.68 -4.82±0.07 4-hydroxybenzyl alcohol -6.26 -5.85±0.01 aminopyrine -6.55 -5.67±0.03 aniline -4.94 -4.55±0.00 benzyl nicotinate -4.87 -4.16±0.02 caffeine -6.85 -5.45±0.02 catechol -5.87 -5.39±0.00 cortexolone -7.20 -5.45±0.03 corticosterone -6.84 -5.59±0.01 cortisone -7.38 -6.09±0.02 dexamethasone -7.27 -6.25±0.03 diclofenac -5.30 -3.79±0.02 digitoxin -8.15 -6.38±0.15 estradiol -5.61 -4.15±0.07 fluocinonide -6.33 -5.38±0.06 flurbiprofen -4.72 -3.69±0.01 griseofulvin -6.44 -5.28±0.06 hydrocortisone -7.22 -6.17±0.05 hydroquinone -6.31 -5.87±0.06 hydroxyprogesterone -6.30 -4.70±0.05 ibuprofen -4.58 -3.61±0.06 indomethacine -5.39 -4.40±0.04 isoquinoline -5.11 -4.20±0.01 ketoprofen -5.22 -4.68±0.03 methyl 4-hydroxybenzoate -5.12 -4.88±0.04 methyl phenyl ether (anisole) -4.68 -4.34±0.04 naproxen -4.97 -4.19±0.04 o-phenylenediamine -6.70 -5.42±0.02 piroxicam -6.02 -4.67±0.05 prednisolone -7.91 -6.42±0.02 progesterone -4.90 -4.13±0.02 testosterone -5.54 -4.52±0.03 conclusions the tests performed in this work indicate that the ideal assay conditions for skin-pampa permeability measurements are 4 hours of incubation time and with the use of stirring. this incubation time allows the determination of permeability of the greatest number of compounds while stirring diminishes the thickness of the aqueous boundary layer. concerning membrane retention, this parameter is in general low (0-0.30). it has been observed that the membrane retention depends on the incubation time and also is related to the lipophilicity of compounds when they are in their neutral form. on the other hand, it has been demonstrated that the skin-pampa membrane is affected at basic ph values, so it is advisable to perform experiments below ph 8 to avoid damaging the membrane. the results shown here indicate good agreement between human skin permeability and skin-pampa permeability established under appropriate assay conditions. figure 4. correlation between human skin permeability and skin-pampa permeability. acknowledgements financial support from the generalitat de catalunya (project di-2014 modality di-esp, resolution eco/1426/2014), the ministerio de economía y competitividad from the spanish government (ctq2017-88179-p) and the catalan government (2017 sgr 1074) is acknowledged. authors also acknowledge an anonymous reviewer for the fruitful discussions regarding membrane retention and derivation of equation 6 (appendix). conflict of interest: the authors declare no conflict of interest. references j. bouwstra. structure of the skin barrier and its modulation by vesicular formulations. prog. lipid res. 42 (2003) 1–36. t.j. franz. percutaneous absorption. on the relevance of in vitro data. j. invest. dermatol. 64 (1975) 190–195. m. kansy, f. senner, k. gubernator. physicochemical high throughput screening: parallel artificial membrane permeation assay in the description of passive absorption processes. j. med. chem. 41 (1998) 1007–1010. a. avdeef. absorption and drug development: solubility, permeability, and charge state. john wiley & sons, inc, 2003. a. avdeef, s. bendels, l. di, b. faller, m. kansy, k. sugano, y. yamauchi. pampa—critical factors for better predictions of absorption. j. pharm. sci. 96 (2007) 2893–2909. a. avdeef, o. tsinman. pampa—a drug absorption in vitro model. eur. j. pharm. sci. 28 (2006) 43–50. o. tsinman, k. tsinman, n. sun, a. avdeef. physicochemical selectivity of the bbb microenvironment governing passive diffusion—matching with a porcine brain lipid extract artificial membrane permeability model. pharm. res. 28 (2011) 337–363. g. ottaviani, s. martel, p. a. carrupt. parallel artificial membrane permeability assay: a new membrane for the fast prediction of passive human skin permeability. j. med. chem. 49 no. 13, pp. 3948–3954, 2006. b. sinkó, j. kökösi, a. avdeef, k. takács-novák. a pampa study of the permeability-enhancing effect of new ceramide analogues. chem. biodivers. 6 (2009) 1867–1874. b. sinkó, t.m. garrigues, g.t. balogh, z.k. nagy, o. tsinman, a. avdeef, k. takács-novák. skin–pampa: a new method for fast prediction of skin penetration. eur. j. pharm. sci. 45 (2012) 698–707. b. sinkó, g. vizserálek, k. takács-novák. skin pampa: application in practice. admet dmpk 2 (2014) 191-198. k. tsinman, b. sinkó. a high throughput method to predict skin penetration and screen topical formulations. cosmet. toilet. 128 (2013) 192–199. l. luo, a. patel, b. sinko, m. bell, j. wibawa, j. hadgraft, m.e. lane. a comparative study of the in vitro permeation of ibuprofen in mammalian skin, the pampa model and silicone membrane. int. j. pharm. 505 (2016) 14–19. g. vizserálek, s. berko, g. toth, r. balogh, m. budai-szucs, e. csanyi, b. sinko, k. takács-novák. permeability test for transdermal and local therapeutic patches using skin pampa method. eur. j. pharm. sci. 76 (2015) 165–172. a. avdeef, p. e. nielsen, o. tsinman. pampa-a drug absorption in vitro model 11. matching the in vivo unstirred water layer thickness by individual-well stirring in microtitre plates. eur. j. pharm. sci. 22 (2004) 365–374. r. lieckfeldt, j. villalaín, j. c. gómez-fernández, g. lee. apparent pka of the fatty acids within ordered mixtures of model human stratum corneum lipids. pharm. res. 12 (1995) 1614–1617. a. avdeef. absorption and drug development: solubility, permeability, and charge state. 2nd ed. john wiley & sons, inc., 2012. k. zhang, m. h. abraham, x. liu. an equation for the prediction of human skin permeability of neutral molecules, ions and ionic species. int. j. pharm. 521 (2017) 259–266. bio-loom, 2017. biobyte corp. claremont, ca, usa. http://www.biobyte.com. date accessed: february 17, 2020. acd/percepta, www.acdlabs.com, advanced chemistry development, inc., release (build 2726, 5 359 may 2014). j. m. cabot, e. fuguet, m. rosés. determination of acidity constants of sparingly soluble drugs in aqueous solution by the internal standard capillary electrophoresis method. electrophoresis 35 (2014) 3564–3569. h. wan, a.g. holmen, y. wang, w. lindberg, m. englund, m.b. nagard, r.a. thompson. high-throughput screening of pka values of pharmaceuticals by pressure-assisted capillary electrophoresis and mass spectrometry. rapid commun. mass spectrom. 17 (2003) 2639–2648. a. port, m. bordas, r. enrech, r. pascual, m. roses, c. rafols, x. subirats, e. bosch. critical comparison of shake-flask, potentiometric and chromatographic methods for lipophilicity evaluation (log po/w) of neutral, acidic, basic, amphoteric, and zwitterionic drugs. eur. j. pharm. sci. 122 (2018) 331–340. s. soriano-meseguer, e. fuguet, a. port, m. rosés. estimation of skin permeation by liquid chromatography. admet dmpk 6 (2018) 140-152. appendix equation (6) can be derived as follows for a drug equilibrated between an aqueous compartment and a membrane: therefore: because: and (6) it should be noted that the derivation assumes that drug absorption to the filter material can be ignored. if this is not accurate then a constant can be added in order to improve the fit for less lipophilic compounds (equation 6b) (6b) ©2020 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) doi: http://dx.doi.org/10.5599/admet.761 16 22 doi: http://dx.doi.org/10.5599/admet.761 23 a concise review on lipidomics analysis in biological samples doi: http://dx.doi.org/10.5599/admet.913 1 admet & dmpk 9(1) (2021) 1-22; doi: https://doi.org/10.5599/admet.913 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index review a concise review on lipidomics analysis in biological samples ramani venkata addepalli 1 , ramesh mullangi 2 * 1 palm meadows, kompally, hyderabad-500100, india 2 laxai life sciences pvt ltd, mn park, genome valley, shamirpet, hyderabad-500 078, india *corresponding author: e-mail: ramesh.mullangi@laxai.com; tel.: +91-9611333488; fax: +91-40-66799998. received: october 09, 2020; revised: november 23, 2020; published: december 09, 2020 abstract lipids are a complex and critical heterogeneous molecular entity, playing an intricate and key role in understanding biological activities and disease processes. lipidomics aims to quantitatively define the lipid classes, including their molecular species. the analysis of the biological tissues and fluids are challenging due to the extreme sample complexity and occurrence of the molecular species as isomers or isobars. this review documents the overview of lipidomics workflow, beginning from the approaches of sample preparation, various analytical techniques and emphasizing the state-of-the-art mass spectrometry either by shotgun or coupled with liquid chromatography. we have considered the latest ion mobility spectroscopy technologies to deal with the vast number of structural isomers, different imaging techniques. all these techniques have their pitfalls and we have discussed how to circumvent them after reviewing the power of each technique with examples.. ©2021 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords lipids; lc-ms/ms; samples processing; bioanalysis introduction lipids are a complex and critical heterogeneous molecular entity, playing an intricate and key role in many biological functions such as acting as structural scaffold for the cell membrane, serving as energy storage and participating in signaling pathways. lipid classification system is spearheaded by lipid maps  lipidomics gateway (https://www.lipidmaps.org/). lipids can be divided into eight main categories: fatty acyls (fa), glycerolipids (gl), glycerophospholipids (gp), sphingolipids (sp), sterol lipids (st), prenol lipids (pr), saccharolipids (sl) and polyketides (pk) [1] (table 1). each category can be further classified into different lipid classes and subclasses, based on the number of carbon atoms and double bonds, the branching of the hydrocarbon chain, the position and orientation of double bonds, the addition of polar groups such as choline, inositol and ethanolamine; and glycosylation [2]. the lipid maps structure database currently records 45245 lipid structures (as of 12 march 2020) (https://www.lipidmaps.org/). the complete lipid profile within a cell, tissue or organism is known as a lipidome and is a branch of the “metabolome”. lipidomics is a discipline that studies the characteristics of lipid and to unravel the complex interactions of lipid metabolites in a biological system. http://dx.doi.org/10.5599/admet.913 https://doi.org/10.5599/admet.913 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:ramesh.mullangi@laxai.com http://creativecommons.org/licenses/by/4.0/ https://www.lipidmaps.org/ https://www.lipidmaps.org/ r.v. addepalli and r. mullangi admet & dmpk 9(1) (2021) 1-22 2 table 1. examples of eight categories of lipids categories structures examples typical classes: subclasses fatty acyls, fa hexadecanoic acid fatty acids: straight chain fatty acids eicosanoids fatty alcohols fatty esters fatty amides prenol lipids, pr 2e,6e-farnesol isoprenoids quinones and hydroquinones polyprenols glycerolipids, gl 1-hexadecanoyl-2-(9z-octadecenoyl)-sn-glycero monoradylglycerols: monoacyl glycerols diradylglycerols: diacyl glycerols triradylglycerols: triacyl glycerols glycerophos pholipids, gp 1-hexadecanoyl-2-(9z-octadecenoyl)-sn-glycero-3phosphocholine glycerophosphocholines glycerophosphoenolamines glycerophosphoserines glycerophospholycerols glycerophosphoglycerophosphates glycerophosphoinositols glycerophosphoglycerophosphoglycerols sphingolipids, sp n-(tetradecanoyl)-sphing-4-enine sphingoid bases ceramides phosphosphingolipids neutral glycosphingolipids acidic glycosphingolipids saccharolipids, sl udp-3-o-(3r-hydroxyl-tetradecanoyl)-αd -n-acetylglucosamine acylaminosugars acylaminosugar glycans acyltrehaloses acyltrehalose glycans sterol lipids, st cholest-5-en-3β-ol sterols cholesterol and derivatives steroids bile acids and derivatives polyketides, pk aflatoxin b1 macrolide polyketides aromatic polyketides nonribosomal peptide/ polyketides hybrids a specific alteration in the lipidome, provides potential insights into perturbed pathways, physiological processes and ultimately stages of diseases. clinical lipidomics was defined “as a new integrative admet & dmpk 9(1) (2021) 1-22 lipidomics analysis review doi: http://dx.doi.org/10.5599/admet.913 3 biomedicine to discover the correlation and regulation between a large scale of lipid elements measured and analyzed in liquid biopsies from patients with those patient phenomes and clinical phenotypes” [3]. to cite few examples, lysophosphatidic acid stimulates cell proliferation, migration and survival by acting on its cognate g-protein-coupled receptors in cancer cells [4]. alzheimer’s and parkinson’s diseases have been linked with aberrant cholesterol and abnormal glycolipid metabolism, respectively [5]. lipoprotein abnormalities are associated with type ii diabetes with increased triglycerides (tg) and very low density lipoprotein (vldl) levels, whereas decreased apolipoprotein e/vldl-tg ratio in ischemic heart disease [6]. depending on associated underlying pathways, these lipids serve as a potential biomarkers and a diagnostic tool. for instance, gl, sp, linoleic acid, cholesterol serve as a biomarker for alzheimer’s; lycophosphatidylinositol and prostaglandins (pg) for parkinson’s; ceramides, sphingomyelin for type ii diabetes mellitus and glycosphingolipid for obesity [7]. figure 1. typical work-flow of lipidomics analysis in biological samples owing to the diversification of the lipid classes with various combinations of polar head groups, fatty acyl chains, backbone structures, identification and characterization of lipids is very complicated. aggravating to this is its extensive expansion of the applications. the prime objective of lipidomics is to attain full coverage of precise structural analysis, accurate quantification and understanding its dynamics. the available analytical techniques are broadly categorized into three groups, namely global lipidomic analysis, targeted lipidomic analysis and novel lipid discovery. the global lipidomics deals with identification, quantification hundreds to thousands of cellular lipids via a high throughput basis. shotgun lipidomics-based platforms are extensively used in this category to analyze diverse pathways and networks associated with lipid metabolism, trafficking, and homeostasis. as an extension, mapping techniques have also been used to study the spatial and temporal relationships of lipids. targeted lipidomic analysis also deals with the identification but with one or a few lipid classes of interest using lc-ms (liquid chromatography coupled to mass spectrometry) and lc-ms/ms (liquid chromatography coupled to tandem mass spectrometry) based methods. while the novel lipid discovery deal with novel lipid classes and molecular species using lc coupled with ms with different enrichment technologies. in this review, we are providing an overview of the current understanding of lipid analysis taking into account the workflow, http://dx.doi.org/10.5599/admet.913 r.v. addepalli and r. mullangi admet & dmpk 9(1) (2021) 1-22 4 methodologies, technical characteristics and bottlenecks. we have also listed important papers and reviews that cover most of the aspects of lipidomics. we hope that this review will act as a bridge for biomedical and pharmacological research to generate novel approaches to disease diagnostics. a typical work flow of lipidomic analysis in biological samples is as shown figure 1. sample preparation sample processing and storage the most important and vital step in any bioanalytical work is sample processing. the biological samples can be solid in nature (e.g., tissues or cells) or comprise of highly complex biofluids (e.g., plasma, serum, urine or cyst fluid). it is laborious to handle solid samples as it involves additional disruption step. it is highly advisable to process the samples immediately, especially with the whole blood samples [8] or at least flash frozen as lipid samples exhibit substantial circadian variations. it is a well-known fact that the plasma concentration of lysophosphatidylcholine (lpc) or lysophosphatidic acid (lpa) increases when left at room temperature for a long period [9]. on the other hand, cardiolipin (cl) during the freezing process hydrolyses into monolysocardiolipin [10] and methanolic samples of lysophospholipid regioisomers isomerize at temperatures above 20 °c and at ph > 6.0 [11]. freeze-thaw is another problem with specific lipid classes, aliquoting the samples before freezing will minimize its effect, example sphingosine, polyunsaturated fatty acids and eicosanoids [12,13]. yet another problem is lipid oxidation and is a major concern with polyunsaturated fatty acid moieties, oxidized lipids and eicosanoids, special care should be exercised in handling, as these end products also contribute to both physiological and pathophysiological processes [13]. storage of plasma/serum samples and their extracts under an inert gas (e.g., argon) may also limit oxidation. antioxidants like butylhydroxytoluene (bht) have been in use, the used concentrations and the time-point of its addition vary in the literature [14]. the efficacy and protocols for the use of antioxidants should be verified. however, oxidation is not a problem in quantifying abundant lipid classes (e.g., phospholipids, sphingolipids and tgs) [13]. it is advisable to exercise caution while handling samples with each class of lipids. biofluids, such as urine, serum, plasma and whole blood are frequently used as these do not require any homogenization but while working with the lipids from a piece of tissue or ruptured cells (e.g., organelles), sample disruption nonetheless has a significant impact on the end results of a process to make it accessible to extraction solvents. widely used mechanical methods are liquid based homogenization (potter-elvehjem homogenizer, ultra-turrax), bead bearing (bead ruptor 24, elite bead mill homogenizer) and crushing of liquid-nitrogen-frozen tissue by pestle and mortar [15]. latter approach is very slow as it is performed manually. ulmer et al. (2017) have demonstrated the use of zirconium oxide or ceramic bead for softer tissues and stainless steel bead for muscle and harder tissues [16]. lipid extraction protocols there are numerous published lipid extraction methods that can also be automated for high-throughput analysis [17-19]. this process reduces the complexity of the sample by enriching the analytes of interest and getting rid of any unwanted contaminants to mass spectrometer. the extraction principles are one/two-phase liquid-liquid extraction (lle), solid-phase extraction (spe) with varying parameters of extraction like temperatures, sample/solvent ratio, re-extractions, use of sonication, vortexing, extraction under inert gasses to enhance the lipid recovery. however, this process is accountable for the artifacts in lipid identification and inconsistencies in quantification. admet & dmpk 9(1) (2021) 1-22 lipidomics analysis review doi: http://dx.doi.org/10.5599/admet.913 5 the sample preparation technique most widely used in lipidomics is lle. the folch protocol and the bligh and dyer protocol both rely on a ternary mixture of chloroform, methanol and water [17,18]. onephase extraction (ope) method was proposed by pellegrino et al. (2014) which is a mixture of methanol/chloroform/methyl tert-butyl ether (mmc) in the ratio 1.33/1/1 (v/v/v) [20]. matyash et al. (2008) and baker et al. (2001) proposed methyl-tert-butyl ether high-throughput lipidomics and acidified butanol extraction procedure for lysophosphatidic acid from biological samples [19,21]. it is well known fact that solubility can be enhanced by addition of some acid to the organic layer containing anionic lipids e.g. phosphatidic acids (pa), phosphatidylinositols (pi) or sphingosine-1-phosphate (s1p) to increase the extraction efficiency. we recommend to check the website http://cyberlipid.gerli.com/techniques-ofanalysis/ extraction-handling-of-extracts/ for guidance on extraction protocols. derivatization chemical derivatization has substantially improved the shortcomings of mass spectrometric based shotgun lipidomics, liquid chromatography and more so in gas chromatography based applications. though it is an additional step, it offers several advantages, it enhances the ionization efficiency, selectively introduce a fragment which can be used in precursor ion or neutral loss scans, masks the functional groups that contaminate the mass spectrometer (ms), often encountered with the lipids containing phosphates and most importantly, it helps in differential quantitation by selectively introducing an isotopic label. all these strategies were used by different scientists. the plasma samples for 7-oxocholesterol and 5,6epoxycholesterol were chemically derivatized with girard’s reagent p to increase the ionization efficiency of the intermediate metabolites in the patients suffering from lysosomal storage disorders [22]. wang et al. (2017) quantitatively analyzed phosphatidylglycerols and bis(monoacylglycero) phosphates by diazomethane-based methylation of phosphate group to introduces class-specific fragments into the ms/ms spectra [23]. to prevent the contamination of the ms, clark et al. (2011) methylated the phosphate groups to quantify phosphatidyl inositol phosphates [24]. last strategy was adopted by lee et al. (2017) who introduced a stable isotope-labeled methylation into one sample to enhanced detection and quantification of targeted phospholipids [25]. chemical derivatization with n-[4-(amino methyl) phenyl] pyridinium prevented the molecular masses overlapping signals between ceramides and branched fatty acid esters of hydroxy fatty acids (fahfas) to a new region thereby reducing the false results of fahfas [26]. new chemical derivatization approaches were employed on targeted lipids by zhao et al. (2020) [27]. however, there are lot of shortcomings in derivatization procedures and such procedures are expected to be rapid, high yielding, specific, in-situ, or biocompatible in order to meet the needs of studies. internal standards internal standards are required to quantify bonafide concentrations of an analyte of interest in ms analysis. both the internal standard and analyte are analyzed simultaneously to compensate for inherent variations in sample processing during the entire process of sample preparation and analysis (e.g., variations in lipid extraction efficacy, processing losses, matrix effects i.e., ionization suppression or enhancement). in an ideal scenario, both the internal standard and the analyte should be structurally similar and have a comparable ms/ms fragmentation pattern. so, a clear understanding of the types, concentration and characteristics of internal standards to be used for accurate quantification of lipid species, subclasses, and classes is of utmost importance. ideally a stable isotope of the analyte of interest, if commercially available, should be used as an internal standard for quantitative analysis, provided quantification is limited to one or a limited number of species [28,29]. http://dx.doi.org/10.5599/admet.913 http://cyberlipid.gerli.com/techniques-of-analysis/%20extraction-handling-of-extracts/ http://cyberlipid.gerli.com/techniques-of-analysis/%20extraction-handling-of-extracts/ r.v. addepalli and r. mullangi admet & dmpk 9(1) (2021) 1-22 6 internal standard mixes in lipidomic analysis approach is very common to biomarker discovery. few researchers are of the opinion that one internal standard per lipid class is enough for quantitation [29], because ionization of lipids is largely dependent on the class specific head group and not so much on the fatty acyl chains [30]. whereas, few others have contradictory report about the influence of fatty acyl chain length and unsaturation on ionization efficiency [30], it is highly advisable to extensively evaluate on case to case basis on this issue. the commercially available internal standard mixes containing isotope-labeled species include a wide range of acyl chain length and degree of saturation [31]. at least two internal standards are required to correct the effects of differential fragmentation kinetics and thermodynamics [32]. however, very few reliable internal standards with adequate chemical purity and exactly known lipid content are commercially available for a limited number of lipid subclasses and fatty acid compositions. an internal standard cocktail has been developed for the lipid maps consortium which is commercially available as splash ® lipidomix ® from avanti polar lipids for quantitative mass spectrometry analysis (https://avantilipids.com/product/330707, 2018). this is specifically designed to complement human plasma lipid analysis using lc-ms/ms platforms. this contains uncommon chain-length sphingoid bases (c17) for sphingosine (so), sphinganine (sa) and their 1-phosphates (s1p and sa1p) and c12:0 fatty acid analogs of ceramide (cer) ceramide 1-phosphate (cer1p), sphingomyelins (sm), simple monoand dihexosylceramides (hexcer and dihexcer). it also includes sulfatides, 1-deoxyand 1-(deoxymethyl)sphingoids, glycerol-phospholipids, phosphatidylinositol-bisphosphate, sterols and neutral lipids (https://avantilipids.com/product/330707, 2018). to get consistent and robust results in a high throughput clinical analysis, the use of commercially available ready-made internal standard mixes with exactly known concentrations has an added advantage. the lipidomics community should encourage the development of novel comprehensive and easily available isotope labeled internal standard mixes. analytical techniques for the study of lipids shotgun lipidomics identification and quantification of a cellular lipidomics directly from organic extracts of biological samples based on chemical and physical properties of lipid classes, subclasses, and individual molecular species are the prime aim of shotgun lipidomics [33]. this concept was based on a simple technique of a conventional loop injection, using a syringe coupled with tandem quadrupole mass spectrometry analysis of precursor masses and fragment mass, refraining from chromatographic separation [33,34]. this evolved into static nano-esi source [29] without syringe pump; resulting in higher ionization efficiencies. han et al. had used intra-source separation, favoring the ionization (i.e., positiveand negative-ion ionization) of selected lipid classes through solvent additives and subsequently using precursor ion and neutral loss scans of polar head group, resulting in fatty acid moieties [34,35]. sufficient quantitation was achieved by addition of one internal standard per lipid class [36] as ionization largely dependents on specific head group and not so much on the fatty acyl chains [37]. though some scientists contradict that fatty acyl chain length and unsaturation influence ionization efficiency [38]. another direct infusion methodology adopted by guan et al. (2006) was coupling triple quadrupole analyzer in multiple reactions monitoring (mrm) mode with syringe pump [39]. this was used to quantify the major lipid components in the lipid extract but critically lies in the knowledge of anticipated precursor and product ions. although this technique is simple, ease of management and less expensive, due to the continuous infusion cross-contamination, isobaric overlaps of the m + 2 isotope with the monoisotopic peak of the compound are major limitations. https://avantilipids.com/product/330707 https://www.powerthesaurus.org/critical/synonyms https://www.powerthesaurus.org/knowledge/synonyms admet & dmpk 9(1) (2021) 1-22 lipidomics analysis review doi: http://dx.doi.org/10.5599/admet.913 7 to overcome the contamination problem from the carryover of previous samples, multi-dimensional mass spectrometry based shotgun lipidomics (mdms-sl) was used to study the composition of lipid structures [37]. using this technique differential intra-source separation properties with various additives like li + or nh4 + or na + , resulted in unique fragments for each lipid class. nanoelectrospray ionization was integrated with chip-based nano-esi platform (advion nanomat  ) for better sensitivity and high reproducibility of sample infusion. quantitative applications of various classes of glycerophospholipids, sphingolipids and glycerolipids were studied using this technique. however, this devise is not cost effective and has limited stability of the electrospray due to clogging. moreover, samples were subjected to analysis sequentially, by charge separation (e.g., deprotonation) or adduct formation (e.g., protonation, alkaline metal adduct ion, halogen adduct ion) making the data handling difficult where specialized software packages are required, which will be discussed later in this article. other disadvantage being separation of isobaric lipids is not possible. in an unique application, differential mobility spectrometry (dms) separation was combined with ionization by electrospray ionization (esi) to separate many isobaric and isomeric lipids, which is also not fully evolved [40]. nonetheless, this was a quick and reliable technique. detection of unexpected lipid species and vulnerability for overlapping isobaric compounds is the major limitation of this method. automation was brought about by replacing syringe pump with hplc (high-performance liquid chromatography) and coupling with triple quadrupole analyzer. the hplc pump runs at micro-flow rates and the auto-sampler injects samples into the flow directly delivering into the esi (electro-spray ionization) source sans the column [41]. robustness and high automation made data acquisition simpler, as multiple precursor ions and constant neutral loss scans could be achieved by concentrated sample pulse and multiple injections. multiple standard curves were achieved in quantitation of lipids (different fatty acyl chain lengths and degrees of unsaturation) by standard addition method and one internal standard [42]. this method was further applied on various subclasses of glycerophospholipids, sphingolipids and sterols [42]. however, low resolution direct infusion technologies suffer from general dogmas of limitations, such as ion suppression, ambiguous identification of isobaric/isomeric lipid species, and ion source-generated artifacts, hindering the applications to low‐abundance lipid species, particularly in less ionizable or isomers bearing identical fragmentation patterns. one such example, diacyl and acyl-alkyl glycerophospholipids, isobaric phosphatidylcholines with odd-carbon-numbered fatty acids from plasmalogens [43]. to deal with this, shotgun lipidomics has evolved into a myriad of multi-dimensional strategies for molecular lipid characterization by coupling with high resolution mass spectrometers (hrms) like quadrupole time of flight (q-tof) or orbitrap or fourier transform ion cyclotron resonance (ft-icr) [44-46]. hrms measures exact mass, time-of-flight ms (tof-ms), orbitrap ms and ft-icr-ms deliver mass resolutions of 60,000, 240,000 and more than 1,000,000, respectively. hrms is especially useful because of their rapid acquisition of ms/ms spectra, higher mass resolution and optional ms n fragmentation, enabling fingerprint studies without prior separation and eliminating the possibilities of false-positive identification. choi et al. (2014) used untargeted lipid analysis to achieve detection of nine lipids in plasma after rosuvastatin treatment to explicate the side effects of the drug by using qtof-ms [46]. further, this facilitates data-independent acquisition (dia) ms/ms all , which has a wide mass range of spectral coverage to perform qualitative analysis. gao et al. were pragmatic in quantification of cardiolipin in mitochondrial preparations [47]. major shortcoming of ms/ms all is that it leads to eventual loss of information on precursor-fragment relationships, complicating the identification of lipids. almeida et al. (2014) utilized full fragmentation power of an orbitrap fusion tribrid and sequentially acquired higher-energy collisional induced dissociation (hcd) and collisional induced dissociation (cid) and ion trap mass spectrometry http://dx.doi.org/10.5599/admet.913 r.v. addepalli and r. mullangi admet & dmpk 9(1) (2021) 1-22 8 (itms3) to structurally characterize molecular glycerophospholipid species [44]. using this novel high confidence filtering strategy, 311 lipid species circumventing 20 lipid classes and identification of 202 distinct molecular glycerophospholipid species in mouse cerebellum and hippocampus was achieved. flaherty et al. (2019) evaluated the levels of multiple lipid species in bone marrow-derived macrophages using triple quadrupole/ion trap mass spectrometer [48]. liquid-chromatography-mass spectrometry shotgun approach is extensively used because of its simple, rapid and ease of handling the crude lipid extracts. but the major pitfall of shotgun is highly convoluted spectra due to matrix interference (ion suppression and ion enhancement effects), functional group modification, and occurrence of the molecular species as isomers or isobars throws greatest challenges to separation scientists. this requires good analytical separation platforms i.e., lc-ms to reduce the above drawbacks. further, coelution of lipids had replaced conventional hplc systems with faster run times, highly efficient separation owing to the higher backpressure allowance, ultra-high-pressure liquid chromatography (uhplc). integrating uhplc with tofms (uhplc/q-tof-ms) is powerful tool, which enables high resolution chromatographic separation coupled with structure elucidation and identification of fragmentation patterns of the comprehensive nontargeted lipid analysis [49]. the most important separation techniques used in lipidomics are rplc (reverse phased liquid chromatography), nplc (normal phase liquid chromatography) and hilic (hydrophilic interaction liquid chromatography). other separation techniques are also occasionally used include non-aqueous rplc (narp) [50], silver-ion rplc [51], chiral lc [52] and supercritical fluid chromatography (sfc) [53]. each technique was used by different group of scientists to resolve the complexity of the lipidome. firstly, nonaqueous rplc was use by lin et al. (1997) the separation of molecular species of 45 synthetic triacylglycerols and diacylglycerols, due to the advent of new column technologies this use is limited [50]. secondly, non-aqueous reversed-phase (narp) and silver-ion high-performance liquid chromatography with apci-ms and gc/fid detection were used for the characterization of fatty acids and triglycerides composition in complex samples of animal fats [51]. thirdly, complex mixtures of regioisomeric and enantiomeric eicosanoids (hydroxy and hydroperoxy fatty acids) have been resolved using chiral lipidomics approach using electron capture atmospheric pressure chemical ionization/mass spectrometry [52]. finally, sfc coupled with orbitrap mass spectrometry based lipidomics platform was used to identify diverse lipid molecular species [53]. reversed-phase lc rplc is most widely used and separations are based on lipophilicity of interacting components, e.g. shorter carbon chains and polyunsaturated analogs being more polar elute earlier as compared to longer carbon chains and saturated acyl structures, respectively. before 2004, long narrow (100-250 mm), normal bore (2-4.6 mm i.d.) columns with higher particle size (3.5-5 μm) were used. with the advent of uhplc columns with 2 μm particle size were in use with higher flow rates and are resulting in better resolution. further, to decrease the diffusional mass transfer path, operating at higher speeds and lower back pressures, “fused core” technology was used with 2.6-2.8 μm particles with a 0.35-0.5 μm porous shell fused to a solid core [54]. c18, a core-shell column, showed superior performance in case of chromatographic peak characteristics (plate number, number of detected lipid features) [55]. few applications focused on miniaturization of lipidomics analyses, employing capillary (50-650 × 0.15-0.32 mm i.d.; 1.7-5 μm) and nanobore (50-170 × 0.075 mm i.d.; 3-5 μm) columns; as these offer higher sensitivity admet & dmpk 9(1) (2021) 1-22 lipidomics analysis review doi: http://dx.doi.org/10.5599/admet.913 9 and smaller sample requirements at lower flow-rates of 0.3-10 μl/min [56]. in lipidomics, c18 or c8-modified sorbents based rplc columns of short length (50-150 mm; typically, 100 mm) microbore (1.0-2.1 mm i.d) with particle size of sub-2 μm or 2.6-2.8 μm (fused-core) are majorly used. examples are the acquity uplc beh c18, zorbax eclipse xdb-c18, acquity uplc hss t3, acquity uplc beh c8, kinetex c18 and ascentis express c8 columns. these columns can be used at a flowrate of 0.1-0.5 ml/min and 40-55°c temperatures. apart from these, c30 stationary phase are also used, though their use is limited in untargeted profiling of lipidome, but its potential has been demonstrated in separation of phospholipids [57]. the geometric and positional isomers of structurally related lipids having conjugated double bonds can be separated by polymeric c30 stationary phase; this is attributed to phase thickness [58]. a comprehensive untargeted lipidomic analysis using core–shell c30 particle column has been demonstrated by narváez-rivas and zhang [59]. these columns showed an excellent resolution of triglyceride regioisomers, which only differed in fatty acid positions (sn-1, 2 or 3) on the glycerol [60]. few examples of c30 columns are acclaim™ c30, accucore™ c30, halo c30. along these column chemistries, achieved good lc separation and detection of lipids by rplc; a mixture of water with or without different combinations of mobile phase additives of ammonium formate or acetate (5-10 mm), formic or acetic acid (0.05-0.1%) and organic solvent(s) like acetonitrile, methanol, isopropanol (ipa) or tetrahydrofuran (thf) have to be used. rplc has few limitations in separating phospholipids. normal-phase lc separation of phospholipids with rplc is not very efficient, so nplc is used even though analysis time is more because of long length (100-250 mm; typically, 150 mm) though having microbore (2 mm i.d) with 35 μm particle size and separates lipids based on their polar functional group [56]. few examples are luna 3 μm silica, lichrospher si 60, betasil silica-100, and nucleosil 100-5 oh. these columns are operated at a flow-rate of 0.1-0.5 ml/min but with higher flow-rate (1.0 ml/min) split mode should be used and maintained at temperatures of 20-35 °c. mobile phase employed over here are highly non-polar solvents with low ionization capacity like heptane, propan-1-ol, methyl ter-butyl ether, chloroform, ethanol and methanol. different proportions of these solvents can be used to get weak to strong mobile phases. at times, additives like 0.5% nh4oh; 5-15 mm ammonium acetate, ammonium formate, diethylamine, formic acid, or small amounts of water (0.5-3%) are added to get adequate separation of lipidome. hydrophilic interaction chromatography (hilic) hilic technique offers the benefits of both normal-phase and reverse-phase in lipid separations. hilic columns exhibited extraordinary separation of lipid classes based on head group composition because of its hydrophilic properties; while using the same rplc mobile phases to improve ionization efficiency and reproducibility of fatty acid chain length, degree of saturation and double bond position [61]. lipid separations under hilic conditions are usually conducted on shorter length (100-150 mm) microbore (2 mm i.d.) columns with 1.7-5 μm particle size, such as atlantis hilic silica, acquity uplc beh hilic, nucleosil 100-5 oh and spherisorb si. these columns are operated at a flow-rate of 0.1-1 ml/min and maintained at temperatures of 25-40 °c. the analysis time is typically in the range of 15-60 min. weak and strong mobile phases are used with high proportion of acetonitrile and water respectively, with methanol and ipa. additives used are 0.1-0.2 % formic acid, 5-10 mm ammonium acetate, 20 mm ammonium formate, or 10 mm nh4oh. it has been demonstrated that hilic can be used as an alternate system in a mix-mode with reverse phase liquid chromatography in the analysis of complex lipids [56,57]. extreme diversity and challenges are faced in the analysis of complex lipids (isobars, regioisomers, http://dx.doi.org/10.5599/admet.913 r.v. addepalli and r. mullangi admet & dmpk 9(1) (2021) 1-22 10 ether-linkage or head group modification). to overcome this, many scientists have combined two liquid chromatography platforms to reduce the sample complexity, analysis times and elevated linear dynamic range [61]. hilic and c30 reversed-phase chromatography (c30rp) coupled to high resolution mass spectrometry was used to analyse modified class of acylphosphatidylglycerol (acyl pg) in corn roots by hilic, and further resolution of the isomers was enhanced using c30 rp chromatography [62]. rampler et al. (2018) had used orthogonal hilic and rp separations in parallel and the effluents of both columns were combined prior to high-resolution ms detection, to achieve full separation in one analytical run [63]. long micropillar array columns (µpac) this unique technology was developed by a lithographic etching process to create a perfectly ordered separation bed on a silicon chip. freestanding nature of the pillar offers several advantages compared to conventional columns by elimination of heterogeneous flow paths in the separation bed, thereby low backpressure, high resolution and high sensitivity. sandra et al. (2017) has demonstrated inter and intra class separation and resolving isomeric lipids [64]. the blood plasma lipid differentiation of lyso‑ glycerophospholids (lyso‑ gps) and monoglycerides (mgs) from the glycerophospholipids (gps), sphingolipids (sps) and diglycerides (dgs) was studied [64]. this is commercially available with pharmafluidics (https://www.pharmafluidics.com). maldi and maldi-tof mass spectrometry maldi-tof is an ideal complementary tool for shotgun lipidomic experiments since late 1990, because of its excellent sensitivity, high tolerance against salts, sample impurities, instrument robustness and freedom from crossover sample contamination [65]. imaging lipids, mapping the distribution of various lipids in different tissues have been successfully carried out by many researchers using this technique which will be discussed later in this review. however, right choice of matrix plays a pivotal role in malditof, for example free fatty acids analysis is difficult with standard matrixes, 2,5-dihydroxy benzoic acid and alfa-cyano-4-hydroxy cinnamic acid, due to the signal and matrix overlap. instead basic matrixes like 9amino-acridine and 2-mercaptobenzothiazole have been in use. schiller et al. (1999) had comprehended an article on various conditions and matrices used in this technique [66]. so, maldi-tof provides a fast, easy, and useful tool for profiling complex lipid mixture, microbial lipids such as lipid a [67] and phosphatidylinositol mannosides [68]. maldi-ms can be very easily coupled with thin-layer chromatography (tlc) allowing the spatially resolved screening of the entire tlc plate and the detection of lipids with a higher sensitivity and nondestructively in comparison with ir lasers and uv lasers [69]. major setback of this method is in the quantitative analysis, due to lack of reproducibility, lack of universal matrix and interference of chemical background noise especially in low mass regions. a combination of highenergy collision-induced dissociation (cid) and prompt ion detection characteristic for maldi tof/tof ms/ms has a unique feature of remote fragmentation of lipids at the level of fatty acyl sn position and double bond location, making the structural analysis easy [70]. pittenauer et al. (2011) have illustrated many applications using this [71]. ion mobility spectrometry separation of isomeric and isobaric species in complex biological samples is a major challenge in hrms; either shotgun ms or coupled with liquid chromatography ms. ion mobility separates ions based on their differential mobility (size, shape, charge) through an inert gas (typically helium, argon or nitrogen) under the influence of an electric field [72,73]. the im-ms is a strong synergy between these two techniques because of their ability to ascertain complementary information about gas-phase ions. three-dimensional https://www.pharmafluidics.com/ admet & dmpk 9(1) (2021) 1-22 lipidomics analysis review doi: http://dx.doi.org/10.5599/admet.913 11 separations are achieved by a range of front-end techniques, ims and ms providing fast measurements, providing new insights into lipid biology [74]. the front-end analytical separations include gas chromatography, supercritical fluid chromatography, liquid chromatography, solid-phase extractions, capillary electrophoresis, field asymmetric ion mobility spectrometry and microfluidic devices [74]. whereas ms includes time-of-flight mass spectrometers (tof-ms), quadrupole mass spectrometers (qms), ion trap mass spectrometers (it-ms), fourier transform mass spectrometers (ftms) and magnetic sector mass spectrometer. the major ionization sources are electrospray, maldi and laser spray ionization. paglia et al. (2014, 2015) have done extensive work in this field and provide incredibly detailed protocols on various ims-coupled mass spectrometry methods [75,76]. ion mobility spectroscopy technologies there are four commercially available ims-ms technologies that have been utilized for lipidomics analysis: (i) drift time ion mobility spectrometry (dtims) here packets of ions are injected into a drift tube filled with an inert buffer gas. under the influence of a weak electric field, ions are separated by charge, size, and shape, developed by agilent technologies [77] (ii) field asymmetric ion mobility spectrometry/differential mobility spectrometry (faims/dms) an asymmetric waveform is applied to two cylindrical plates such that ions experience alternating high and low electric fields. ions traverse the region between the plates moving in a perpendicular direction to the buffer gas and with the influence of a dc potential, termed the compensation voltage (cv). only selected ions at a given cv will make it through the drift region, this approach is provided by sciex [78] (iii) travelling wave ion mobility spectrometry (twims) an alternating phase radio-frequency (rf) potential is applied to a series of stacked ring ion guides (srigs). ions are pushed through the drift region with a traveling potential wave and become mobility separated as higher mobility ions are able to ‘roll-over’ the traveling waves generated and exit the srig region, developed by waters [79] and (iv) trapped ion mobility spectrometry (tims), this uses an electric field to hold ions stationary against a moving gas, so that the drift force is compensated by the electric field and ion packages are separated based on their respective ion mobilities and process called parallel accumulation serial fragmentation or pasef, developed by bruker [80]. a detailed description of the theoretical concepts we refer the readers to reviews in this field [73,81]. improved lipid identification by im-ms identification of lipids has been done by accurate mass match with online databases such as lipid maps or lipidblast, but it provides only molecular formula. this is inconclusive as number of species belonging to different lipid classes has same molecular formula. the physicochemical characteristics of the compounds are required to allow a more accurate identification. this allows the calculation of the collision cross section (ccs), a four-dimensional orthogonal (retention time, m/z, ion mobility, intensity) physicochemical measure that can be used, together with accurate mass, fragmentation information and retention time (rt) to increase the confidence of lipid identification in milliseconds. it is a known fact that, saturated lipids bearing acyl chains are extended in the electric field and have larger ccs values as compared to the unsaturated bonds with bent structure in the acyl chain. the ccs values are more influenced by the structural characteristics of compounds than the degree of saturation [75,82-84]. this was further correlated with the ccs values of fas and pcs with both the lipid chain length and the degree of unsaturation. many scientists have determined the ccs values to cover the lipidome of complex biological matrices [83,84]. catherine et al. compiled 1856 lipid ccs values from plasma, liver and cancer cells with high quality of tims ccs values [85]. zhou et al. (2017) developed a support vector regression model using bioinformatic approaches and set of molecular descriptors, earnestly describing the subtle structural http://dx.doi.org/10.5599/admet.913 r.v. addepalli and r. mullangi admet & dmpk 9(1) (2021) 1-22 12 differences for lipids on smiles structures [86]. these in silico lipidccs values are independent data and are externally validated from the experimentally (twims and dtims) determined values. lipid ccs predictor offers (i) prediction of lipid ccs values; (ii) lipidccs database search and (iii) lipid match and identification [84]. the lipid ccs database approximately contains over 15,000 lipids with over 60,000 corresponding ccs values determined experimentally or predicted in silico. ccs evaluation an additional identifying factor to improve data interpretation and enhance lipid identification in untargeted workflows [75]. however, validation of ccs values is restricted to the limited number of commercially available lipid standards and can only be used as an in silico ccs prediction to improve identification efficiency in lipidomics. applications/isomer separation ion mobility system is exceptionally well-suited for untargeted lipidomics due to high-resolution, highthroughput and structural elucidation capabilities. notable separation occurs in ion mobility (im) before fragmentation (ms), as such product ions are mobility-aligned to corresponding precursor ions thereby improving interpretation of product ion spectra. also, identical product ions derived from different precursor ions are proportionately assigned to their precursors, enhancing the low-abundance species detection [87]. one of the major confrontation in lipid analysis is isomer separation [88], with regio[such as sn1 (16:0) or sn2 (18:1 (9z)) for gps] [89], positional (position of double bond), or geometric (cis/trans conformation of the double bond i.e. z/e) isomers [90]. typical example of glycerophospholipids is shown in figure 2. zandkarimi et al. (2019) separated, co-eluted plasmalogen phosphatidylethanolamine (pe p-) pe (p-36:1) and pe (p-38:2) lipids in mice brain tissue using lc-im-ms. as both the pe had same retention time but were separated clearly in the ion mobility region with different drift time bin number [91]. thus, structural elucidation was feasible due to ims drift time, high collisional energy in transfer region and clear fragmentation pattern [76]. technical developments in both hardware and software, empowers researchers to implement im-ms into their analytical workflows. the four major augmentations in lipidomics are firstly, im-ms crucially resolves isobaric species, thereby improves separation of lipids. secondly, im fragmentation improves the spectral clarity of product ion spectra. this is crucial in both lipid identification and structural elucidation, which are grueling task due to the isomeric nature of many lipid species. on the other hand, analysis of product ion spectra derived from untargeted fragmentation acquisitions remains challenging due to the required powerful processing tools. thirdly, im improves separation of isomeric lipids. lastly, ccs values obtained from im-ms analysis effectively increase confidence in lipid identification. finally, im-ms can be used to comprehend the conformational dynamics of a lipid system and offering a unique means of characterizing flexibility and folding mechanisms. ms lipidomics imaging and in situ lipidomics provide spatial information about the lipid composition in tissues sort of molecular microscope [92]. there are several desorption ionizations tools and imaging ms techniques but ‘lipidomics standard initiative’ (https://lipidomics-standards-initiative.org/) have recommended only secondary ion mass spectrometry (sims), desorption electrospray ionization (desi) and matrix-assisted laser desorption/ionization (maldi). out of which, maldi-imaging mass spectrometry (maldi-ims) is commonly used for lipid imaging in tissue sections. https://www.bing.com/search?q=define+empower&form=dctrqy https://lipidomics-standards-initiative.org/ admet & dmpk 9(1) (2021) 1-22 lipidomics analysis review doi: http://dx.doi.org/10.5599/admet.913 13 figure 2. isomeric structures of glycerophospholipids [polar head group indicated as ‘x’ can be glycerol, choline, ethanolamine, inositol, or serine and two fatty acyl groups at sn1 (16:0) and sn2 (18:1 (9z)) position forming the gp tail. three types of isomers are represented regioisomer presenting the fatty acyl chains in inverted positions; a positional isomer, containing the double bond in a different position (c6 instead of c9), and a geometric isomer, whose double bond is in trans (e) conformation] maldi-imaging mass spectrometry (maldi-ims) maldi-imaging mass spectrometry (maldi-ims) is a two-dimensional maldi-ms technique which gives comprehensive profiles of molecular distributions of lipids [93] with high spatial resolution without extraction, purification, separation, or labeling of biological samples. this reveals the localization and abundance of hundreds of molecular species, especially lipids on the tissue slice in a single measurement thus helping in understanding the cellular profile of the biological system. shimma et al. (2007) used this study the abnormal distribution of phospholipids in colon cancer liver metastasis [93]. this technique is widely used in brain and skin lipidomics as well [94]. goto-inoue et al. (2011) have used tlc-blot-maldi-ims to study detailed structural analysis of lipids from human brain samples [95]. this was done by a three-step process, by running tlc once, transferring to a polyvinylidene difluoride membrane and finally detection by maldi-ims. thus it was possible to separate, visualize and identify phosphatidylserine (ps) (diacyl-18:0/20:4), phosphatidylcholine (pc) (diacyl-16:0/18:1), and sphingomyelin (sm) (d18:1/c18:0) at m/z 812.5, 753.5 and 782.5 in gray matter. it was proposed that, this system would be useful in fully analyzing lipid compositions including minor components. kendall et al. applied this acyl chain head group http://dx.doi.org/10.5599/admet.913 r.v. addepalli and r. mullangi admet & dmpk 9(1) (2021) 1-22 14 technique in skin lipidomics [95] by analyzing ex vivo human skin cutaneous lipids for assessing alterations in lipid profiles linked to specific skin conditions. both maldi and desi skin imaging are used for analysis of the whole skin sections, though the analysis time was same, disadvantages being, the former had matrix selection and later had less spatial resolution. sims imaging is known for high spatial resolution but a low mass range, which can decipher the spatial distribution of multiple lipids at subcellular levels. a completely new technique was studied using infrared matrix assisted laser desorption electrospray ionization mass spectrometry (ir-maldesi), combines many benefits of maldi and esi. ir-maldesi works by combining laser desorption of neutrals with subsequent ionization by esi to increase lipid abundance with great coverage relative to commonly used parameters in negative polarity [96]. this technique can be used for quantitative analysis and ideally for drug distribution studies [96]. the use of other ambient ionization tools, including rapid evaporative ionization mass spectrometry (reims) and direct analysis in real time (dart) allow rapid, real-time screenings of lipids for predictive, preventive, and personalized medicine. reims based methods require no preparative steps or timeconsuming cell extractions which are discussed later in this review. ir-maldesi sensitivity and selectivity was augmented using silver cationization of olefinic lipids in human serum using calcifediol as the reference [97]. rapid evaporative ionization mass spectrometry (reims) reims is truly ambient analysis technique, for fast, easy molecular profiling with zero sample preparation. this is highly versatile technique as it can be used for both biological solid and liquid samples, as this provides quick determination of differences within and between samples. this is achieved by simple evaporation of the sample by joule heating or laser irradiation, the aerosol generated is introduced orthogonally to the inlet of the mass spectrometer such as high performance tof-ms. this technique has been used in intra-surgical tissue classification [98], bacterial identification [96], rapid profiling of cell lines [100] the analysis of plant material, food applications [101] and bioliquid samples. this technique has a potential to combine with other sampling techniques in providing a holistic profiling approach. few scientists have used matrix assisted version of the technique (ma-reims) to enhance the signal intensity in identifying strong phospholipid signal in the tumor which is absent in normal breast tissue [98]. in another group, in situ and real-time recognition model was employed in identifying 12 fatty acids and 37 phospholipids using “iknife” and reims in discrimination of salmon and rainbow trout without sample preparation and adulteration of minced meats [102]. despite so many advantages, the major setback of this technique is, it provides only moderate amounts of chemical information and repeatability, for further information we should still bank on lc-ms. direct analysis in real time-mass spectrometry (dart-ms) dart-ms is an ambient pressure ionization technique enabling instantaneous and sensitive analyses of gases, liquids and solids [103]. it is based on the interactions of long lived electronic or vibronic excitedstate molecule with sample and atmospheric gases at atmospheric pressure. this technique does not require laborious sample preparation, as ionization takes place directly on the sample surface, deposited or adsorbed on to surfaces or that are being desorbed into the atmosphere. the combination of this source with a high-resolution mass spectrometer (commercially available with joel as dart-accu tof) offers a rapid qualitative and quantitative measurements. it has numerous applications in the field of food science, forensics, and clinical analysis (https://www.jeol.co.jp/en/applications/pdf/ms/ms_note_en002.pdf). “noprep” analysis of lipids in cooking oils and detection of adulterated olive oil is an application using dartadmet & dmpk 9(1) (2021) 1-22 lipidomics analysis review doi: http://dx.doi.org/10.5599/admet.913 15 accu tof. despite numerous advantages, dart ionization does have several inherent limitations. firstly, fragmentation occurs at higher plasma temperatures, hindering spectra interpretation and accurate determination of the mass of intact molecules, well the decomposition fragments can also contribute to the structural information [104]. secondly, this technique subjects the analyte to oxidation artifacts, owing to the design of the instrument i.e., distance from the capillary outlet [104]. lastly, saturated hydrocarbons can undergo hydride abstraction hindering the quantitative analysis. for example, signals from aliphatic hydrocarbons and monounsaturated hydrocarbons are indistinguishable due to the same carbon length [105]. conclusions in this review, we have outlined different analytical strategies for a comprehensive identification, complete structural characterization and accurate quantitation of biogenic lipid molecules with a fewer bottle necks. it must be noted, that each technique has its own strengths and weaknesses for example high mass resolution results in accurate mass but with natural limitations. therefore, a mixture of analytical devices (chromatography, spectrometry, mass spectrometry and hyphenated methods) helps to cope up with the complexity of the lipids structure. however, lipidomics provide enormous data especially nontargeted lipids and it is critical to evaluate it using bioinformatics solutions [106]. the lipid information is available as web resources namely, cyberlipid center (http://www.cyberlipid.org/) and aocs lipid library (http://lipidlibrary.aocs. org/). tools that have been developed for analysis of ms-based lipidomic data include ms and ms/ms data by lipid consortium such as lipid maps lipidomics gateway (www.lipidmaps.org/resources/tutorials/databases.html) and national institute of standards and technology (nist) (http://chemdata.nist.gov/); and commercialized software such as lipidview™ (https://sciex.com/products/software/lipidview-software), lipidsearch™ (https://www.thermofisher.com/ hr/en/home/industrial/mass-spectrometry/liquid-chromatography-mass-spectrometry-lc-ms/lc-mssoftware.html), and simlipid (www.premierbiosoft.com) also become available. this will undoubtedly be a valuable tool for investigation of many diseases, physiological processes or in disease biomarker discovery. conflict of interest: ramesh mullangi is vice president, pre-clinical biology & dmp at laxai life sciences. references [1] e. fahy, s. subramaniam, r.c. murphy, m. nishijima, c.r.h. raetz, t. shimizu, f. spener, g. meer, m.j.o. wakelam, e.a. dennis. update of the lipid maps comprehensive classification system for lipids. journal of lipid research 50 (2009) s9-s14. doi: https://doi.org/10.1194/jlr.r800095jlr200. [2] e. fahy, d. cotter, m. sud, s. subramaniam. lipid classification, structures and tools. biochim. biophys. acta 1811 (2001) 637-647. https://doi.org/10.1016/j.bbalip.2011.06.009. [3] x. wang, d. wu, h. shen. lipidomics in health and disease: methods and application. in: x.d. wang (ed). translational bioinformatics, vol 16. berlin: springer, 2018. [4] g.b. mills, w.h. moolenaar. the emerging role of lysophosphatidic acid in cancer. nature reviews cancer 3 (2003) 582-591. https://doi.org/10.1038/nrc1143. [5] r.g. cutler, j. kelly, k. storie, w.a. pedersen, a. tammara, k. hatanppa, j.c. troncoso, m.p. mattson. involvement of oxidative stress-induced abnormalities in ceramide and cholesterol metabolism in brain aging and alzheimer’s disease. proceedings of the national academy of sciences usa 101 (2004) 2070-2075. https://doi.org/10.1073/pnas.0305799101. [6] t. murakami, t. urata, n. yamada, k. kawakubo, t. yoshida, h. kawamura, t. kashiwa, s. lino. lipoprotein abnormalities in type 2 diabetic patients. increased distribution of triglycerides (tg) in http://dx.doi.org/10.5599/admet.913 http://www.lipidmaps.org/resources/tutorials/databases.html http://chemdata.nist.gov/ https://sciex.com/products/software/lipidview-software https://www.thermofisher.com/hr/en/home/industrial/mass-spectrometry/liquid-chromatography-mass-spectrometry-lc-ms/lc-ms-software.html https://www.thermofisher.com/hr/en/home/industrial/mass-spectrometry/liquid-chromatography-mass-spectrometry-lc-ms/lc-ms-software.html https://www.thermofisher.com/hr/en/home/industrial/mass-spectrometry/liquid-chromatography-mass-spectrometry-lc-ms/lc-ms-software.html http://www.premierbiosoft.com/ https://www.ncbi.nlm.nih.gov/pubmed/?term=fahy%20e%5bauthor%5d&cauthor=true&cauthor_uid=19098281 https://www.ncbi.nlm.nih.gov/pubmed/?term=subramaniam%20s%5bauthor%5d&cauthor=true&cauthor_uid=19098281 https://www.ncbi.nlm.nih.gov/pubmed/?term=murphy%20rc%5bauthor%5d&cauthor=true&cauthor_uid=19098281 https://www.ncbi.nlm.nih.gov/pubmed/?term=nishijima%20m%5bauthor%5d&cauthor=true&cauthor_uid=19098281 https://www.ncbi.nlm.nih.gov/pubmed/?term=raetz%20cr%5bauthor%5d&cauthor=true&cauthor_uid=19098281 https://www.ncbi.nlm.nih.gov/pubmed/?term=shimizu%20t%5bauthor%5d&cauthor=true&cauthor_uid=19098281 https://www.ncbi.nlm.nih.gov/pubmed/?term=spener%20f%5bauthor%5d&cauthor=true&cauthor_uid=19098281 https://www.ncbi.nlm.nih.gov/pubmed/?term=van%20meer%20g%5bauthor%5d&cauthor=true&cauthor_uid=19098281 https://www.ncbi.nlm.nih.gov/pubmed/?term=van%20meer%20g%5bauthor%5d&cauthor=true&cauthor_uid=19098281 https://www.ncbi.nlm.nih.gov/pubmed/?term=wakelam%20mj%5bauthor%5d&cauthor=true&cauthor_uid=19098281 https://www.ncbi.nlm.nih.gov/pubmed/?term=dennis%20ea%5bauthor%5d&cauthor=true&cauthor_uid=19098281 https://doi.org/10.1194/jlr.r800095-jlr200 https://doi.org/10.1194/jlr.r800095-jlr200 https://doi.org/10.1016/j.bbalip.2011.06.009 https://doi.org/10.1038/nrc1143 https://doi.org/10.1073/pnas.0305799101 r.v. addepalli and r. mullangi admet & dmpk 9(1) (2021) 1-22 16 very low density lipoprotein (vldl) and decreased apolipoprotein e/vldl-tg ratio are risk for ischemic heart disease. artery 22 (1997) 309-327. [7] j.z. kubicek-sutherland, d.m. vu, h.m. mendez, s. jakhar, h. mukundan. detection of lipid and amphiphilic biomarkers for disease diagnostics. biosensors 7 (2017) 25. https://doi.org/10.3390/ bios7030025. [8] e.c.p. chua, g. shui, i.t. lee, p. lau, l.c. tan, s.c. yeo, b.d. lam, s. bulchand, s.c. summers, k. puvanendran, s.g. rozen, m.r. wenk, j.j. gooley. extensive diversity in circadian regulation of plasma lipids and evidence for different circadian metabolic phenotypes in humans. proceedings of the national academy of sciences usa 110 (2013) 14468-14473. https://doi.org/10.1073/ pnas.1222647110. [9] m. scherer, g. schmitz, g. liebisch. high-throughput analysis of sphingosine 1-phosphate, sphinganine 1-phosphate and lysophosphatidic acid in plasma samples by liquid chromatographytandem mass spectrometry. clinical chemistry 55 (2009) 1218-1222. https://doi.org/10.1373/clin chem.2008.113779. [10] j. kim, c.l. hoppel. comprehensive approach to the quantitative analysis of mitochondrial phospholipids by hplc-ms. journal of chromatography b 912 (2013) 105-114. https://doi.org/10. 1016/j.jchromb.2012.10.036. [11] m. okudaira, a. inoue, a. shuto, k. nakanaga, k. kano, k. makide, d. saigusa, y. tomioka, j. aoki. separation and quantification of 2-acyl-1-lysophospholipids and 1-acyl-2-lysophospholipids in biological samples by lc-ms/ms. journal of lipid research 55 (2014) 2178-2192. doi: https://doi.org /10.1194/jlr.d048439. [12] u. ceglarek, j. dittrich, c. helmschrodt, k. wagner, j.r. nofer, j. thiery, s. becker. preanalytical standardization of sphingosine-1-phosphate, sphinganine-1-phosphate and sphingosine analysis in human plasma by liquid chromatography-tandem mass spectrometry. clinica chim acta 435 (2014) 1-6. https://doi.org/10.1016/j.cca.2014.04.010. [13] j. dorow, s. becker, l. kortz, j. thiery, s. hauschildt, u. ceglarek. preanalytical investigation of polyunsaturated fatty acids and ecosanoids in human plasma by liquid chromatography-tandem mass spectrometry. biopreservation and biobanking 14 (2016) 107-113. https://doi.org/10.1089 /bio.2015.0005. [14] l.a. heiskanen, m. suoniemi, h.x. ta, k. tarasov, k. ekroos. long-term performance and stability of molecular shotgun lipidomic analysis of human plasma samples. analytical chemistry 85 (2013) 8757 8763. https://doi.org/10.1021/ac401857a. [15] k. jurowski, k. kochan, j. walczak, m. barańska, w. piekoszewski, b. buszewski. analytical techniques in lipidomics: state of the art. critical reviews in analytical chemistry 47 (2017) 418-437. https://doi.org/10.1080/10408347.2017.1310613. [16] c.z. ulmer, r.e. patterson, j.p. koelmel, t.j. garrett, r.a. yost. a robust lipidomics workflow for mammalian cells, plasma, and tissue using liquid-chromatography high-resolution tandem mass spectrometry. methods in molecular biology 1609 (2017) 91-106. doi: https:/doi.org/10.1007/978-14939-6996-8_10. [17] j. folch, m. lees, g.h. sloane stanly. a simple method for the isolation and purification of total lipids from animal tissues. journal of biological chemistry 226 (1957) 497-509. [18] j. bligh, m. lees, g.h. sloane stanley. a rapid method of total lipid extraction and purification. canadian journal of biochemistry and physiology 37 (1959) 911-917. [19] v. matyash, g. liebisch, t.v. kurzchalia, a. shevchenko, d. schwudke. lipid extraction by methyl-tertbutyl ether for high-throughput lipidomics. journal of lipid research 49 (2008) 1137-1146. doi: https://doi.org/10.1194/jlr.d700041-jlr200. [20] r.m. pellegrino, a. di veroli, a. valeri, l. goracci, g. cruciani. lc/ms lipid profiling from human serum: a new method for global lipid extraction. analytical and bioanalytical chemistry 406 (2014) 7937-7948. https://doi.org/10.1007/s00216-014-8255-0. https://doi.org/10.3390/bios7030025 https://doi.org/10.3390/bios7030025 https://doi.org/10.1073/pnas.1222647110 https://doi.org/10.1073/pnas.1222647110 https://doi.org/10.1373/clinchem.2008.113779 https://doi.org/10.1373/clinchem.2008.113779 https://doi.org/10.1016/j.jchromb.2012.10.036 https://doi.org/10.1016/j.jchromb.2012.10.036 https://doi.org/10.1194/jlr.d048439 https://doi.org/10.1194/jlr.d048439 https://doi.org/10.1016/j.cca.2014.04.010 https://doi.org/10.1089/bio.2015.0005 https://doi.org/10.1089/bio.2015.0005 https://doi.org/10.1021/ac401857a https://doi.org/10.1080/10408347.2017.1310613 https://dx.doi.org/10.1007/978-1-4939-6996-8_10 https://dx.doi.org/10.1007/978-1-4939-6996-8_10 https://doi.org/10.1194/jlr.d700041-jlr200 https://doi.org/10.1007/s00216-014-8255-0 admet & dmpk 9(1) (2021) 1-22 lipidomics analysis review doi: http://dx.doi.org/10.5599/admet.913 17 [21] d.l. baker, d.m. desiderio, d.d. miller, b. tolley, g.j. tigyi. direct quantitative analysis of lysophosphatidic acid molecular species by stable isotope dilution electrospray liquid chromatography-mass spectrometry. analytical biochemistry 292 (2001) 287-295. https://doi.org/ 10.1006/abio.2001.5063. [22] w.j. griffiths, i. gilmore, e. yutuc, j. abdel-khalik, p.j. crick, t. hearn, a. dickson, b.w. bigger, t.h.y. wu, a. goenka, a. ghosh, s.a. jones, y. wang. identification of unusual oxysterols and bile acids with 7-oxo or 3β,5α,6β-trihydroxy functions in human plasma by charge-tagging mass spectrometry with multistage fragmentation. journal of lipid research 59 (2018) 1058-1070. doi: https://doi.org/ 10.1194/jlr.d083246. [23] m. wang, j.p. palavicini, a. cseresznye, x. han. strategy for quantitative analysis of isomeric bis(monoacylglycero)phosphate and phosphatidylglycerol species by shotgun lipidomics after onestep methylation. analytical chemistry 89 (2017) 8490-8495. https://doi.org/10.1021/acs.analchem. 7b02058. [24] j. clark, k.e. anderson, v. juvin, t.s. smith, f. karpe, m.j.o. wakelam, l.r. stephens, p.t. hawkins. quantification of ptdinsp3 molecular species in cells and tissues by mass spectrometry. nature methods 8 (2011) 267-272. doi: https://doi.org/10.1038/nmeth.1564. [25] j.c. lee, s.k. byeon, m.h. moon. relative quantification of phospholipids based on isotope-labeled methylation by nano flow ultra high performance liquid chromatography-tandem mass spectrometry: enhancement in cardiolipin profiling. analytical chemistry 89 (2017) 49694977. https://doi.org/10.1021/acs.analchem.7b00297. [26] c. hu, m. wang, q. duan, x. han. sensitive analysis of fatty acid esters of hydroxy fatty acids in biological lipid extracts by shotgun lipidomics after one-step derivatization analytica chimica acta 1105 (2020) 105-111. https://doi.org/10.1016/j.aca.2020.01.026. [27] x. zhao, s. zhu, h. liu. recent progresses of derivatization approaches in the targeted lipidomics analysis by mass spectrometry. journal of separation science 43 (2020) 1838-1846. https://doi.org /10.1002/jssc.201901346. [28] r. deems, m.w. buczynski, r. bowers‐gentry, r. harkewicz, e.a. dennis. detection and quantitation of eicosanoids via high performance liquid chromatography-electrospray ionization-mass spectrometry. methods in enzymology 432 (2007) 59-82. doi: 10.1016/s0076-6879(07)32003-x. [29] x. han, k. yang, r.w. gross. multi-dimensional mass spectrometry-based shotgun lipidomics and novel strategies for lipidomic analyses. mass spectrometry reviews 31 (2011) 134-178. https://doi .org/10.1002/mas.20342. [30] m. wang, c. wang, x. han. selection of internal standards for accurate quantification of complex lipid species in biological extracts by electrospray ionization mass spectrometry what, how and why? mass spectrometry reviews 36 (2017) 693-714. doi: http://doi.org/10.1002/mas.21492. [31] t. züllig, m. trötzmüller, h.c. köfeler. lipidomics from sample preparation to data analysis: a primer. analytical bioanalytical chemistry 412 (2020) 2191-2209. https://doi.org/10.1007/s00216019-02241-y. [32] s. sales, j. graessler, s. ciucci, r. al-atrib, t. vihervaara, k. schuhmann, d. kauhanen, m. sysi-aho, s.r. bornstein, m. bickle, c.v. cannistraci, k. ekroos, a. shevchenko. gender, contraceptives and individual metabolic predisposition shape a healthy plasma lipidome. scientific reports 6 (2016) 27710. https://doi.org/10.1038/srep27710. [33] x. han, r.w. gross. shotgun lipidomics: electrospray ionization mass spectrometric analysis and quantitation of cellular lipidomes directly from crude extracts of biological samples. mass spectrometry reviews 24 (2005) 367-412. [34] x. han. characterization and direct quantitation of ceramide molecular species from lipid extracts of biological samples by electrospray ionization tandem mass spectrometry. analytical biochemistry 302 (2002) 199-212. https://doi.org/10.1006/abio.2001.5536. http://dx.doi.org/10.5599/admet.913 https://doi.org/10.1006/abio.2001.5063 https://doi.org/10.1006/abio.2001.5063 https://doi.org/10.1194/jlr.d083246 https://doi.org/10.1194/jlr.d083246 https://doi.org/10.1021/acs.analchem.7b02058 https://doi.org/10.1021/acs.analchem.7b02058 https://doi.org/10.1038/nmeth.1564 https://doi.org/10.1021/acs.analchem.7b00297 https://www.sciencedirect.com/science/journal/00032670 https://www.sciencedirect.com/science/journal/00032670/1105/supp/c https://doi.org/10.1016/j.aca.2020.01.026 https://doi.org/10.1002/jssc.201901346 https://doi.org/10.1002/jssc.201901346 http://doi.org/10.1016/s0076-6879(07)32003-x https://doi.org/10.1002/mas.20342 https://doi.org/10.1002/mas.20342 https://doi.org/10.1002/mas.21492 https://doi.org/10.1007/s00216-019-02241-y https://doi.org/10.1007/s00216-019-02241-y https://doi.org/10.1038/srep27710 https://doi.org/10.1006/abio.2001.5536 r.v. addepalli and r. mullangi admet & dmpk 9(1) (2021) 1-22 18 [35] x. han, r.w. gross. electrospray ionization mass spectroscopic analysis of human erythrocyte plasma membrane phospholipids. proceedings of the national academy of sciences usa 91 (1994) 1063510639. doi: https://doi.org/10.1073/pnas.91.22.10635. [36] k. yang, z. zhao, r.w. gross, x. han. systematic analysis of choline-containing phospholipids using multi-dimensional mass spectrometry-based shotgun lipidomics. journal of chromatography b 877 (2009) 2924-2936. https://doi.org/10.1016/j.jchromb.2009.01.016. [37] x. han, r.w. gross. shotgun lipidomics: multidimensional ms analysis of cellular lipidomes. expert review of proteomics 2 (2005) 253-264. https://doi.org/10.1586/14789450.2.2.253. [38] m. koivusalo, p. haimi, l. heikinheimo, r. kostiainen, p. somerharju. quantitative determination of phospholipid compositions by esi-ms: effects of acyl chain length, unsaturation, and lipid concentration on instrument response. journal of lipid research 42 (2001) 663-672. [39] x.l. guan, x. he, w.y. ong, w.k. yeo, g.h. shui, m.r. wenk. non-targeted profiling of lipids during kainate-induced neuronal injury. faseb journal 20 (2006) 1152-1161. https://doi.org/10.1096/fj.055362com. [40] p.r.s. baker, a.m. armando, j.l. campbell, o. quehenberger, e.a. dennis. three-dimensional enhanced lipidomics analysis combining uplc, differential ion mobility spectrometry and mass spectrometric separation strategies. journal of lipid research 55 (2004) 2432-2442. https://doi.org/ 10.1194/jlr.d051581. [41] p. wiesner, k. leidl, a. boettcher, g. schmitz, g. liebisch. lipid profiling of fplc-separated lipoprotein fractions by electrospray ionization tandem mass spectrometry. journal of lipid research 50 (2008) 574-585. doi: https://doi.org/10.1194/jlr.d800028-jlr200. [42] g. liebisch, b. lieser, j. rathenberg, w. drobnik, g. schmitz. high-throughput quantification of phosphatidylcholine and sphingomyelin by electrospray ionization tandem mass spectrometry coupled with isotope correction algorithm. biochimica biophysica acta 1686 (2004) 108-117. doi: https://doi.org/10.1016/j.bbalip.2004.09.003. [43] g. liebisch, j.a. vizcaíno, h. köfeler, m. trotzmuller, w.j. griffiths, g. schmitz, f. spener, m.j.o. wakelam. shorthand notation for lipid structures derived from mass spectrometry. journal of lipid research 50 (2013)1523-1530. doi: https://doi.org/10.1194/jlr.m033506. [44] r. almeida, j.k. pauling, e. sokol, h.k. hannibal-bach, c.s. ejsing. comprehensive lipidome analysis by shotgun lipidomics on a hybrid quadrupole-orbitrap-linear ion trap mass spectrometer. journal of the american society for mass spectrometry 26 (2014) 133-148. https://doi.org/10.1007/s13361014-1013-x. [45] m. höring, c.s. ejsing, m. hermansson, g. liebisch. quantification of cholesterol and cholesteryl ester by direct flow injection high resolution ftms utilizing species-specific response factors. analytical chemistry 91 (2019) 3459-3466. https://doi.org/10.1021/acs.analchem.8b05013. [46] j.m. choi, t.e. kim, j.y. cho, h.j. lee, b.h. jung. development of lipidomic platform and phosphatidylcholine retention time index for lipid profiling of rosuvastatin treated human plasma. journal of chromatography b 944 (2014) 157-165. doi: https://doi.org/10.1016/j.jchromb. 2013.10.029. [47] f. gao, j. mcdaniel, e.y. chen, h.e. rockwell, c. nguyen, m.d. lynes, y.h. tseng, r. sarangarajan, n.r. narain, m.a. kiebish. adapted ms/ms all shotgun lipidomics approach for analysis of cardiolipin molecular species. lipids 53 (2018) 133-142. doi: https://dx.doi.org/10.1002/lipd.12004 . [48] s.e. flaherty, a. grijalva, x. xu, e. ables, a. nomani, a.w. ferrante. a lipase-independent pathway of lipid release and immune modulation by adipocytes. science 363 (2019) 989-993. doi: https://doi.org /10.1126/science.aaw2586. [49] m. holčapek, r. jirásko, m. lísa. recent developments in liquid chromatography-mass spectrometry and related techniques. journal of chromatography a 1259 (2012) 3-15. doi: https://doi.org/10.1016 /j.chroma.2012.08.072. https://doi.org/10.1073/pnas.91.22.10635 https://doi.org/10.1016/j.jchromb.2009.01.016 https://doi.org/10.1586/14789450.2.2.253 https://doi.org/10.1096/fj.05-5362com https://doi.org/10.1096/fj.05-5362com https://doi.org/10.1194/jlr.d051581 https://doi.org/10.1194/jlr.d051581 https://doi.org/10.1194/jlr.d800028-jlr200 https://doi.org/10.1016/j.bbalip.2004.09.003 https://doi.org/10.1194/jlr.m033506 https://doi.org/10.1007/s13361-014-1013-x https://doi.org/10.1007/s13361-014-1013-x https://doi.org/10.1021/acs.analchem.8b05013 https://doi.org/10.1016/j.jchromb.2013.10.029 https://doi.org/10.1016/j.jchromb.2013.10.029 https://dx.doi.org/10.1002/lipd.12004 https://doi.org/10.1126/science.aaw2586 https://doi.org/10.1126/science.aaw2586 https://doi.org/10.1016/j.chroma.2012.08.072 https://doi.org/10.1016/j.chroma.2012.08.072 admet & dmpk 9(1) (2021) 1-22 lipidomics analysis review doi: http://dx.doi.org/10.5599/admet.913 19 [50] j.t. lin, c.l. woodruff, t.a. mckeon. non-aqueous reversed-phase high-performance liquid chromatography of synthetic triacylglycerols and diacylglycerols. journal of chromatography a 782 (1997) 41-48. https://doi.org/10.1016/s0021-9673(97)00482-2. [51] m. lísa, k. netušilová, l. franěk, h. dvořáková, v. vrkoslav, m. holčapek. characterization of fatty acid and triacylglycerol composition in animal fats using silver-ion and non-aqueous reversed-phase high-performance liquid chromatography/mass spectrometry and gas chromatography/flame ionization detection. journal of chromatography a 1218 (2011) 7499-7510. doi: https://doi.org/10 .1016/j.chroma.2011.07.032. [52] s.h. lee, i.a blair. targeted chiral lipidomics analysis of bioactive eicosanoid lipids in cellular systems. bmb reports 42 (2009) 401-410. [53] t. yamada, t. uchikata, s. sakamoto, y. yokoi, s. nishiumi, m. yoshida, e. fukusaki, t. bamba. supercritical fluid chromatography/orbitrap mass spectrometry based lipidomics platform coupled with automated lipid identification software for accurate lipid profiling. journal of chromatography a 1301 (2013) 237-242. doi: https://doi.org/10.1016/j.chroma.2013.05.057. [54] d. guillarme, j. ruta, s. rudaz, j.l. veuthey. new trends in fast and high-resolution liquid chromatography: a critical comparison of existing approaches. analytical and bioanalytical chemistry 397 (2009) 1069-1082. https://doi.org/10.1007/s00216-009-3305-8. [55] j.o. omamogho, j.d. glennon. comparison between the efficiencies of sub-2 μm c18 particles packed in narrow bore columns. analytical chemistry 83 (2011) 1547-1556. doi: https://doi.org/10 .1021/ac102139a. [56] t. cajka, o. fiehn. comprehensive analysis of lipids in biological systems by liquid chromatographymass spectrometry. trends analytical chemistry 61 (2014) 192-206. doi: https://doi.org/10.1016 /j.trac.2014.04.017. [57] h. houjou, k. yamatani, m. imagawa, t. shimizu, r. taguchi. a shotgun tandem mass spectrometric analysis of phospholipids with normal-phase and/or reverse-phase liquid chromatography/electrospray ionization mass spectrometry. rapid communications in mass spectrometry 19 (2005) 654-666. https://doi.org/10.1002/rcm.1836. [58] r. kiyonami, d. peake, d. drew, x. sun, x. liu, k. miller.large-scale lipid profiling of a human serum lipidome using a high-resolution, accurate-mass lc-ms/ms approach. thermo scientific 2016; application note 647. [59] m. narváez-rivas, q. zhang. comprehensive untargeted lipidomic analysis using core-shell c30 particle column and high field orbitrap mass spectrometer. journal of chromatography a 1440 (2016) 123-134. https://doi.org/10.1016/j.chroma.2016.02.054. [60] j.w. lee, t. nagai, n. gotoh, e. fukusaki, t. bamba. profiling of regioisomeric triacylglycerols in edible oils by supercritical fluid chromatography/tandem mass spectrometry. journal of chromatography b 966 (2014) 193-199. https://doi.org/10.1016/j.jchromb.2014.01.040. [61] m. lisa, e. cifkova, m. holcapek. lipidomic profiling of biological tissues using off-line twodimensional high-performance liquid chromatography-mass spectrometry. journal of chromatography a 1218 (2011) 5146-5156. https://doi.org/10.1016/j.chroma.2011.05.081. [62] t.h. pham, m. zaeem, t.a. fillier, m. nadeem, n.p. vidal, c. manful, s. cheema, m. cheema, r.h. thomas. targeting modified lipids during routine lipidomics analysis using hilic and c30 reverse phase liquid chromatography coupled to mass spectrometry. scientific reports 9 (2019) 5048. https://doi.org/10.1038/s41598-019-41556-9. [63] e. rampler, h. schoeny, b.m. mitic, y. el abiead, m. schwaiger, g. koellensperger. simultaneous nonpolar and polar lipid analysis by on-line combination of hilic, rp and high resolution ms. analyst 143 (2018) 1250-1258. [64] k. sandra, p. sandra, w. d. malsche, g. desmet, r. t’kindt r, j.o.d. beeck, b. claerebout, j. vandenbussche. evaluation of micro-pillar array columns (µpacᵀᴹ) combined with high resolution mass spectrometry for lipidomics. lc gc europe hplc 30 (2017) 6-13. http://dx.doi.org/10.5599/admet.913 https://doi.org/10.1016/s0021-9673(97)00482-2 https://doi.org/10.1016/j.chroma.2011.07.032 https://doi.org/10.1016/j.chroma.2011.07.032 https://doi.org/10.1016/j.chroma.2013.05.057 https://doi.org/10.1007/s00216-009-3305-8 https://doi.org/10.1021/ac102139a https://doi.org/10.1021/ac102139a https://doi.org/10.1016/j.trac.2014.04.017 https://doi.org/10.1016/j.trac.2014.04.017 https://doi.org/10.1002/rcm.1836 https://doi.org/10.1016/j.chroma.2016.02.054 https://doi.org/10.1016/j.jchromb.2014.01.040 https://doi.org/10.1016/j.chroma.2011.05.081 https://doi.org/10.1038/s41598-019-41556-9 r.v. addepalli and r. mullangi admet & dmpk 9(1) (2021) 1-22 20 [65] s. shimma, y. sugiura, t. hayasaka, y. hoshikawa, t. noda, m. setou. maldi-based imaging ass spectrometry revealed abnormal distribution of phospholipids in colon cancer liver metastasis. journal of chromatography b 855 (2007) 98-103. https://doi.org/10.1016/j.jchromb.2007.02.037. [66] j. schiller, j. arnhold, s. benard, m. müller, s. reichl, k. arnold. lipid analysis by matrix-assisted laser desorption and ionization mass spectrometry: a methodological approach. analytical biochemistry 267 (1999) 46-56. https://doi.org/10.1006/abio.1998.3001. [67] p. zhou, e. altman, m.b. perry, j. li. study of matrix additives for sensitive analysis of lipid a by matrix-assisted laser desorption ionization mass spectrometry. applied and environmental microbiology 76 (2010) 3437-3443. doi: https://doi.org/10.1128/aem.03082-09. [68] f.f. hsu, j. turk, r.m. owens, e.r. rhoades, d.g. russell. structural characterization of phosphatidylmyo-inositol mannosides from mycobacterium bovis bacillus calmette gúerin by multiple-stage quadrupole ion-trap mass spectrometry with electrospray ionization. ii. monoacyland diacyl-pims. journal of the american society for mass spectrometry 18 (2007) 479-492. https://doi.org/10 .1016/j.jasms.2006.10.020. [69] k. dreisewerd, j. müthing, a. rohlfin, i. meisen, z. vukelić, j. peter-katalinić, f. hillenkamp, s. berkenkamp. analysis of gangliosides directly from thin-layer chromatography plates by infrared matrix-assisted laser desorption/ionization orthogonal time-of-flight mass spectrometry with a glycerol matrix. analytical chemistry 77 (2005) 4098-4107. https://doi.org/10.1021/ac048373w. [70] c. cheng, e. pittenauer, m.l. gross. charge-remote fragmentations are energy-dependent processes. journal of the american society for mass spectrometry 9 (1988) 840-844. https://doi.org/10.1016/ s1044-0305(98)00053-1. [71] e. pittenauer, g. allmaier. a universal product ion nomenclature for [m−h] − , [m+h] + and [m+nna−(n−1) h] + (n=1–3) glycerophospholipid precursor ions based on high-energy cid by malditof/rtof mass spectrometry. international journal of mass spectrometry 301 (2011) 90101. https://doi.org/10.1016/j.ijms.2010.07.012. [72] c. lapthorn, f. pullen, b.z. chowdhry. ion mobility spectrometry-mass spectrometry (ims-ms) of small molecules: separating and assigning structures to ions. mass spectrometry reviews 32 (2013) 43-71. doi: https://doi.org/10.1002/mas.21349. [73] j.c. may, j.a. mclean. ion mobility-mass spectrometry: time-dispersive instrumentation. analytical chemistry 87 (2015) 1422-1436. https://doi.org/10.1021/ac504720m. [74] x. zheng, r. wojcik, x. zhang, y.m. ibrahim, k.e. burnum-johnson, d.j. orton, m.e. monroe, r.j. moore, r.d, smith, e.s baker. coupling front-end separations, ion mobility spectrometry, and mass spectrometry for enhanced multidimensional biological and environmental analyses. annual review of analytical chemistry 10 (2017) 71-92. https://doi.org/10.1146/annurev-anchem-061516-045212. [75] g. paglia, p. angel, j.p. williams, k. richardson, h.j. olivos, j. will thompson, l. menikarachchi, s. lai, c. walsh, a. moseley, r.s. plumb, d.f. grant, b.o. palsson, j. langridge, s. geromanos, g. astarita. ion mobility-derived collision cross section as an additional measure for lipid fingerprinting and identification. analytical chemistry 87 (2014) 1137-1144. https://doi.org/10.1021/ac500405x. [76] g. paglia, m. kliman, e. claude, s. geromanos, g. astarita. applications of ion-mobility mass spectrometry for lipid analysis. analytical and bioanalytical chemistry 407 (2015) 4995-5007. doi: https://doi.org/10.1007/s00216-015-8664-8. [77] y.m. ibrahim, e.s. baker, w.f. danielson, r.v. norheim, d.c. prior, g.a. anderson, m.e belov, r.d. smith. development of a new ion mobility time-of-flight mass spectrometer. international journal of mass spectrometry 377 (2015) 655-662. https://doi.org/10.1016/j.ijms.2014.07.034. [78] b.b. schneider, e.g. nazarov, f. londry, p. vouros, t.r. covey. differential mobility spectrometry/ mass spectrometry history, theory, design optimization, simulations, and applications. mass spectrometry reviews 35 (2016) 687-737. https://doi.org/10.1002/mas.21453. [79] a.a. shvartsburg, r.d. smith. fundamentals of traveling wave ion mobility spectrometry. analytical chemistry 80 (2008) 9689-9699. https://doi.org/10.1021/ac8016295. https://doi.org/10.1016/j.jchromb.2007.02.037 https://doi.org/10.1006/abio.1998.3001 https://doi.org/10.1128/aem.03082-09 https://doi.org/10.1016/j.jasms.2006.10.020 https://doi.org/10.1016/j.jasms.2006.10.020 https://doi.org/10.1021/ac048373w https://doi.org/10.1016/s1044-0305(98)00053-1 https://doi.org/10.1016/s1044-0305(98)00053-1 https://doi.org/10.1016/j.ijms.2010.07.012 https://doi.org/10.1002/mas.21349 https://doi.org/10.1021/ac504720m https://doi.org/10.1146/annurev-anchem-061516-045212 https://doi.org/10.1021/ac500405x https://doi.org/10.1007/s00216-015-8664-8 https://doi.org/10.1016/j.ijms.2014.07.034 https://doi.org/10.1002/mas.21453 https://doi.org/10.1021/ac8016295 admet & dmpk 9(1) (2021) 1-22 lipidomics analysis review doi: http://dx.doi.org/10.5599/admet.913 21 [80] f. fernandez-lima, d.a. kaplan, j. suetering, m.a park. gas-phase separation using a trapped ion mobility spectrometer. international journal of ion mobility spectrometry 14 (2011) 93-98. https://doi.org/10.1007/s12127-011-0067-8. [81] w. jiang, n.a. chung, j.c. may. in: ion mobility-mass spectrometry. encyclopedia of analytical chemistry: applications, theory and instrumentation. new york: john wiley & sons, 2013. [82] f. zhang, s. guo, m. zhang, z. zhang, y. guo. characterizing ion mobility and collision cross section of fatty acids using electrospray ion mobility mass spectrometry. journal of mass spectrometry 50 (2015) 906-913. https://doi.org/10.1002/jms.3600. [83] j.c may, c.r. goodwin, n.m. lareau, k.l. leaptrot, c.b. morris, r.t. kurulugama, a. mordehai, c. klein, w. barry, e. darland, g. overney, k. imatani, g.c. stafford, j.c. fjedsted, j.a. mclean. conformational ordering of biomolecules in the gas phase: nitrogen collision cross sections measured on a prototype high resolution drift tube ion mobility-mass spectrometer. analytical chemistry 86 (2014) 2107-2116. https://doi.org/10.1021/ac4038448. [84] k.a.z. berry, r.m. barkley, j.j. berry, j.a. hankin, e. hoyes, j.m. brown, r.c. murphy. tandem mass spectrometry in combination with product ion mobility for the identification of phospholipids. analytical chemistry 89 (2019) 916-921. https://doi.org/10.1021/acs.analchem.6b04047. [85] c.g. vasilopoulou, k. sulek, a.d. brunner, n.s. meitei, u. schweiger-hufnagel, s.w. meyer, a. barsch, m. mann, f. meier. trapped ion mobility spectrometry and pasef enable in-depth lipidomics from minimal sample amounts. nature communications 11 (2020) 331. https://doi.org/10.1038/s41467020-15613-1. [86] z. zhou, j. tu, x. xiong, x. shen, z.j. zhu. lipid ccs: prediction of collision cross-section values for lipids with high precision to support ion mobility-mass spectrometry-based lipidomics. analytical chemistry 89 (2017) 9559-9566. https://doi.org/10.1021/acs.analchem.7b02625. [87] c. hinz, s. liggi, j.l. griffin. the potential of ion mobility mass spectrometry for high-throughput and high-resolution lipidomics. current opinion in chemical biology 42 (2018) 42-50. https://doi.org/10. 1016/j.cbpa.2017.10.018. [88] m. groessl, s. graf, r. knochenmuss. high resolution ion mobility-mass spectrometry for separation and identification of isomeric lipids. analyst 140 (2015) 6904-6911. [89] m. sala, m. lisa, j.l. campbell, m. holcapek. determination of triacylglycerol regioisomers using differential mobility spectrometry. rapid communications in mass spectrometry 30 (2016) 256-264. https://doi.org/10.1002/rcm.7430. [90] a.a. shvartsburg, g. isaac, n. leveque, r.d. smith, t.o. metz. separation and classification of lipids using differential ion mobility spectrometry. journal of the american society for mass spectrometry 22 (2011) 1146-1155. doi: https://doi.org/10.1007/s13361-011-0114-z. [91] f. zandkarimi, l.m. brown. application of ion mobility mass spectrometry in lipidomics. advances in experimental medicine and biology 1140 (2019) 317-326. doi: https://doi.org/10.1007/978-3-03015950-4_18. [92] s.r. ellis, s.h. brown, m.i.h. panhuis, s.j. blanksby, t.w. mitchell. surface analysis of lipids by mass spectrometry: more than just imaging. progress in lipid research 52 (2013) 329-353. https://doi.org/ 10.1016/j.plipres.2013.04.005. [93] s. shimma, y. sugiura, t. hayasaka, y. hoshikawa, t. noda, m. setou. maldi-based imaging mass spectrometry revealed abnormal distribution of phospholipids in colon cancer liver metastasis. journal of chromatography b 855 (2007) 98-103. https://doi.org/10.1016/j.jchromb.2007.02.037. [94] a.c. kendall, m.m. koszyczarek, e.a. jones, p.j. hart, m. towers, c.e.m. griffiths, m. morris, a. nicolaou. lipidomics for translational skin research: a primer for the uninitiated. experimental dermatology 27 (2018) 721-728. https://doi.org/10.1111/exd.13558. [95] n. goto-inoue, t. hayasaka, n. zaima, m. setou. imaging mass spectrometry for lipidomics. biochimica et biophysica acta – molecular and cell biology lipids 1811 (2011) 961-969. https://doi.org/10.1016/j.bbalip.2011.03.004 http://dx.doi.org/10.5599/admet.913 https://doi.org/10.1007/s12127-011-0067-8 https://doi.org/10.1002/jms.3600 https://doi.org/10.1021/ac4038448 https://doi.org/10.1021/acs.analchem.6b04047 https://doi.org/10.1038/s41467-020-15613-1 https://doi.org/10.1038/s41467-020-15613-1 https://doi.org/10.1021/acs.analchem.7b02625 https://doi.org/10.1016/j.cbpa.2017.10.018 https://doi.org/10.1016/j.cbpa.2017.10.018 https://doi.org/10.1002/rcm.7430 https://doi.org/10.1007/s13361-011-0114-z https://doi.org/10.1007/978-3-030-15950-4_18 https://doi.org/10.1007/978-3-030-15950-4_18 https://doi.org/10.1016/j.plipres.2013.04.005 https://doi.org/10.1016/j.plipres.2013.04.005 https://doi.org/10.1016/j.jchromb.2007.02.037 https://doi.org/10.1111/exd.13558 https://doi.org/10.1016/j.bbalip.2011.03.004 r.v. addepalli and r. mullangi admet & dmpk 9(1) (2021) 1-22 22 [96] m.t. bokhart, d.c. muddiman. infrared matrix-assisted laser desorption electrospray ionization mass spectrometry imaging analysis of biospecimens. analyst 141 (2016) 5236-5245. doi: https://doi.org/10.1039/c6an01189f. [97] f. meier, k.p. garrard, d.c. muddiman. silver dopants for targeted and untargeted direct analysis of unsaturated lipids via infrared matrix-assisted laser desorption electrospray ionization (ir-maldesi). rapid communication in mass spectrometry 28 (2014) 2461-2470. https://doi.org/10.1002/ rcm.7041. [98] e.a. jones, d. simon, t. karancsi, j. balog, s.d. pringle, z. takats. matrix assisted rapid evaporation ionisation mass spectrometry. analytical chemistry 91 (2019) 9784-9791. https://doi.org/10.1021 /acs.analchem.9b01441. [99] f. bolt, s.j.s. cameron, t. karancsi, d. simon, r. schaffer, t. rickards, k. hardiman, a. burke, z. bodai, a. perdones-montero, m. rebec, j. balog, z. takats. automated high-throughput identification and characterization of clinically important bacteria and fungi using rapid evaporative ionization mass spectrometry. analytical chemistry 88 (2016) 9419-9426. https://doi.org/10.1021/acs.analchem.6b01016. [100] j. sarsby, l. mclean, v.m. harman, r.j. beynon. monitoring recombinant protein expression in bacteria by rapid evaporative ionisation mass spectrometry. rapid communications in mass spectrometry (2020) e8760. https://doi.org/10.1002/rcm.8670. [101] k. verplanken, s. stead, r. jandova, c.v. poucke, j. claereboudt, j.v. bussche, s.d. saeger, z. takats, j. wauters, l. vanhaecke. rapid evaporative ionization mass spectrometry for high-throughput screening in food analysis: the case of boar taint. talanta 169 (2017) 30-36. https://doi.org/10.1016/j.talanta.2017.03.056. [102] g. song, m. zhang, y. zhang, h. wang, s. li, z. dai, q. shen. an in-situ method for real-time discriminating salmons and rainbow trout without sample preparation using iknife and rapid evaporative ionization mass spectrometry based lipidomics. journal of agricultural and food chemistry 67 (2019) 4679-4688. https://doi.org/10.1021/acs.jafc.9b00751. [103] e.s. chernetsova, g.e. morlock, i.a revelsky. dart mass spectrometry and its applications in chemical analysis. russian chemical reviews 80 (2011) 235-255. [104] r.b. cody. observation of molecular ions and analysis of nonpolar compounds with the direct analysis in real time ion source. analytical chemistry 81 (2009) 1101-1107. https://doi.org /10.1021/ac8022108. [105] z. yang, a.b. attygalle. aliphatic hydrocarbon spectra by helium ionization mass spectrometry (hims) on a modified atmospheric-pressure source designed for electrospray ionization. journal of the american society for mass spectrometry 22 (2011) 1395-1402. https://doi.org/10.1007/s13361-0110149-1. [106] d.a. peake, r. kiyonami, y. yokoi, y. fukamachi, y. huang. processing of a complex lipid dataset for the nist inter-laboratory comparison exercise for lipidomics measurements in human serum and plasma. asms 2015 poster note 64471. ©2021 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) https://doi.org/10.1039/c6an01189f https://doi.org/10.1002/rcm.7041 https://doi.org/10.1002/rcm.7041 https://doi.org/10.1021/acs.analchem.9b01441 https://doi.org/10.1021/acs.analchem.9b01441 https://doi.org/10.1021/acs.analchem.6b01016 https://doi.org/10.1021/acs.analchem.6b01016 https://doi.org/10.1002/rcm.8670 https://doi.org/10.1016/j.talanta.2017.03.056 https://doi.org/10.1021/acs.jafc.9b00751 https://doi.org/10.1021/ac8022108 https://doi.org/10.1021/ac8022108 https://doi.org/10.1007/s13361-011-0149-1 https://doi.org/10.1007/s13361-011-0149-1 http://creativecommons.org/licenses/by/3.0/ organ-on-a-chip systems for new drugs development doi: http://doi.org/10.5599/admet.942 111 admet & dmpk 9(2) (2021) 111-141; doi: http://dx.doi.org/10.5599/admet.942 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index review organ-on-a-chip systems for new drugs development ronny vargas 1,2 *, andrea egurbide-sifre 2 and laura medina 2 1 industrial pharmacy department, faculty of pharmacy, university of costa rica 11501-2060, san josé, costa rica 2 faculty of pharmacy and food sciences, university of barcelona, av. joan xxiii, 27-1, 08028, barcelona, spain. *corresponding author: e-mail: ronny.vargas_m@ucr.ac.cr; tel.: +34-665-873-028 received: december 22, 2020; revised: march 04, 2021; published online: march 22, 2021 abstract research on alternatives to the use of animal models and cell cultures has led to the creation of organ -ona-chip systems, in which organs and their physiological reactions to the presence of external stimuli are simulated. these systems could even replace the use of human beings as subjects for the study of drugs in clinical phases and have an impact on personalized therapies. organ-on-a-chip technology present higher potential than traditional cell cultures for an appropriate prediction of functional impairments, appearance of adverse effects, the pharmacokinetic and toxicological profile and the efficacy of a drug. this potential is given by the possibility of placing different cell lines in a three-dimensional-arranged polymer piece and simulating and controlling specific conditions. thus, the normal functioning of an organ, tissue, barrier, or physiological phenomenon can be simulated, as well as the interrelation between different systems. furthermore, this alternative allows the study of physiological and pathophysiological processes. its design combines different disciplines such as materials engineering, cell cultures, microfluidics and physiology, among others. this work presents the main considerations of ooc systems, the materials, methods and cell lines used for their design, and the conditions required for their proper functioning. examples of applications and main challenges for the development of more robust systems are shown. this nonsystematic review is intended to be a reference framework that facilitates research focused on the development of new ooc systems, as well as their use as alternatives in pharmacological, pharmacokinetic and toxicological studies. ©2021 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords drug discovery; tissue engineering; cell culture; preclinical models; organoids; lab-on-a-chip devices introduction the development of new drugs is an expensive process that presents several challenges in preclinical and clinical studies [1,2], where a 59 % of the medicines that initiate a clinical study go into phase ii, just a 21 % starts phase iii, and merely a 10 % reach approval by the regulatory authority. the appearance of adverse effects is one of the main reasons for drug failure, becoming one of the biggest challenges and weaknesses presented by preclinical models as predictors of clinical performance [2–5]. however, clinical failures are not just the result of safety-related situations, but are also linked to the lack of clinical efficacy [6]. the non-detection of side effects can also become an issue for a drug that has already passed the clinical phase. the non-adequate determination of adverse effects can lead to market recalls. these http://doi.org/10.5599/admet.942 http://dx.doi.org/10.5599/admet.942 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:ronny.vargas_m@ucr.ac.cr http://creativecommons.org/licenses/by/4.0/ r. vargas, a. egurbide-sifre and l. medina admet & dmpk 9(2) (2021) 111-141 112 unwanted events are mainly liveror heart-related, due to the effect of metabolite generation [5,7]. animal models applied in preclinical phases are often unrepresentative to the human body, thus failing in the prediction of the efficacy and safety of drugs [1,3,4,8,9]. however, clinical failures might be mitigated by adding models with a superior predictor capacity in preclinical phases [10]. the use of traditional twodimensional (2d) cell cultures has made great contributions to medicine so far, and it has led to reduce the use of laboratory animals. however, these systems do not allow the study of tissue’s microenvironment or the complex interrelationships between different tissues and organs of the human body. moreover, the study of isolated cells without a systematic functional evaluation presents the risk of ignoring structural alterations that may have an impact on the organ’s function, even if those isolated cells remained viable [1,10]. in recent years, the organ-on-a-chip (ooc) technology has emerged as an alternative in drug development, as it aims to reflect the environmental, functional, and interrelationship characteristics of organs and tissues [2,11]. the concept of reproducing organic and physiological functions of the human body by using cells within a microfluidic chip was first published in 2004, when shuler et al. showed a cell culture that exemplified the interaction between the lung and the liver on a square-inch silicone chip. the term organ-on-a-chip was adopted in 2010 [2,7,12,13]. compared with traditional models, this technique provides improved information about cells mechanical properties, morphology, and differentiation, among others. ooc improve the determination of cell responses, genetic expressions, and cell function. these systems can have an impact on the candidates selection and evaluation for the development of new medicines, as they allow gaining information about metabolites secretion performed by certain cells when interacting with drugs [1]. furthermore, ooc have the potential to reduce and supplement the use of animals, cell models, and even humans in new drugs development [7]. ooc systems offer a better understanding of pathophysiology, allowing new drugs design into specific mechanisms and promote therapeutic development for pathologies such as cancer, neurological disorders, and orphan diseases [2]. choi et al. used a tridimensional (3d) cancer model for the selection of a most effective drug against lung cancer cells [14]. liu et al. studied brain metastasis with a multi-organ chip that showed functional barrier characteristics [15]. during the covid-19 pandemic caused by the sars-cov-2 in 2020, and in the middle of a global race to slow down the progression of the disease, oocs were applied in the research for treatments with fast and reliable results on the preclinical scale [16]. the development of these models brought to the emergence of an industry with more than 28 companies founded in 7 years [7,17,18]. hundreds of publications are generated each year regarding this topic, which rises the interest of companies carrying out basic research in the field of drugs and also in other areas such as cosmetics, food and medical devices [8,11,19–22]. some of the synonyms under which ooc can be found in the literature are the following ones: lab-on-achip [17], organoids-on-chip [23], organ-on-a-plate[6], tissue-on-chip [9], microphysiological system (mps)[6], and mps-based organ model [24]. this work is intended to be a reference framework that facilitates research focused on the development of new ooc systems, as well as their use as alternative in pharmacological, pharmacokinetic and toxicological studies with the aim of making them become a more powerful, accurate and used research tool in drug development. ooc basic features this field of study emerges as a combination of cellular biology, cell culture engineering, microsystem admet & dmpk 9(2) (2021) 111-141 organ-on-a-chip systems for new drugs development doi: http://doi.org/10.5599/admet.942 113 engineering, microfluidic studies, and materials engineering. this complexity of disciplines allows to create microenvironments with precise mechanical, and structural control, providing the ability to recreate concentration gradients, and nutrients and metabolites flows [2,3,7,25–31]. therefore, they are versatile and modifiable systems with the aim of showing an adequate histological composition that is suitable for live evaluation, which cannot be done by using animal models [32]. as an interdisciplinary field, a broader understanding of the ooc technology requires knowledge in each of the sciences that are part of it. since the evolution of these devices involves technology advances in each of these areas, it faces the same difficulties as they do [18], especially with the objective of mass production and sustainable implementation of ooc. some of these challenges are discussed later in this document, although the authors recommend a further study of the basic principles and state of the art of the abovementioned scientific fields. detailed reviews focused on each of these research areas can be consulted elsewhere [33–41]. the spatial configuration of each ooc depends on its design and desired function, however, most of them show the general outlines presented in figure 1; they are constituted by a physical compartment for cell confinement and several microfluidic channels, which are used for the administration and transport of substances and measurable signals. in their integrated compartment oocs allow to study, measure and control cell behavior, as well as the response of the target tissue microenvironment in front of certain stimuli [2]. it is also possible to integrate secondary tissues in the ooc, which cannot be performed in traditional cellular models [5]. the ooc can reproduce particular conditions, such as the presence of specific pathologies, age and sex differences, or metabolic modifications [3,7,8,15,42–44]. the ability of ooc systems to mimic the functionality of an organ allows to visualize the impact of a condition on the organ’s function. for example, for the study of a viral infection at kidney level, a traditional 2d culture would be able to demonstrate the virus replication but its impact on the kidney’s function would not be proved unless the 2d culture was coupled to an ooc system capable of mimicking glomerular filtration [45]. figure 1. examples of ooc devices. a) three-layer chip with a cell culture receptacle on the bottom plate, and multiple channel connectors for feed and signal control on the top plate. adapted from [46] with permission from the royal society of chemistry. copyright (2018) b) a two-channel ooc for the simulation of an epithelial tissue. reproduced from [47] under terms of the creative commons attribution license. c) ooc model for the simulation of the blood-brain barrier with gel suspended astrocytes and neurons. reproduced from [48] with permission from the royal society of chemistry. copyright (2017). http://doi.org/10.5599/admet.942 r. vargas, a. egurbide-sifre and l. medina admet & dmpk 9(2) (2021) 111-141 114 ooc systems are intended to be a useful tool to reduce the number of false positives and false negatives in the preclinical evaluation of potential drug prototypes. therefore, they might be able to raise the number of candidates that successfully complete clinical trials [6]. moreover, these systems have fewer ethical or animal welfare concerns than traditional animal models [49,50]. ooc can facilitate the design and validation of drugs that aim specific cellular and molecular targets [18,44], and could be used to establish the pharmacokinetic profile of drugs [51]. table 1 shows some of the main characteristics of existing preclinical models for a comparative benefits-drawbacks evaluation. table 1. main comparative characteristics of traditional preclinical evaluation models and ooc. modified from [5]. characteristics animal 2d ooc target tissue’s representativeness low low high ethical implications medium low low associated costs high low low viability maintenance medium high high high throughput analysis aptitude low high medium full body representativeness high low medium ooc functional characteristics the design of ooc involves reverse engineering of living organs [42,52], which requires an in-depth study of the organs and their functions to develop systems more physiologically relevant than current in vitro models. this has promoted the physio-pathological study of multi-organ interrelationships [31,42,44,49,53]. the main benefits of ooc arise from their ability to mimic three key organ aspects in their physiological function [7,54], which are: 1. interfaces that present barrier functions composed of multiple tissues and vasculature. 2. organization at a parenchymal level (i.e. the presence of different tissue substructures within the same organ). 3. the interaction between different organs. additionally, to be adaptable to hts, ooc systems require the incorporation of sensors and communication channels [7,27,28] for real-time data collection and analysis [42]. reproduction of barrier or interface properties non-homogeneous cell mixtures contain variations in shape, arrangement and interaction through interfaces of different cells, which limits the use of 2d cell cultures [7]. traditionally, epithelia used to be reproduced in vitro as several cell layers on a surface. however, the evolution of ooc production systems, and the ability to diversify 3d structures, allow a more accurate correspondence with the physiological and anatomic-functional conditions of epithelial barriers [47]. ooc microfluidic channels can incorporate polymeric membranes to model tissue interfaces, simulate barrier properties and, in some cases, they can even include programmed mechanical instruments [7,50,55]. the possibility of establishing a 3d vascular structure enables to extend the complexity and representativeness of the systems mimicked in an ooc. this is especially valuable with regard to the targeting of nutrients and exogenous substances if compared to 2d tissues, where transport occurs by simple diffusion [55]. furthermore, oocs have the ability to include dynamic factors that simulate complex biological events [7,56]. on one hand, the simulation of the physiological interface can be performed in the chip’s primary admet & dmpk 9(2) (2021) 111-141 organ-on-a-chip systems for new drugs development doi: http://doi.org/10.5599/admet.942 115 physical structure itself or by controlling the mobility of hydrogels through the channels of this polymeric structure [7]. current printing techniques allow complex 3d structures to simulate vascularized tissues and to modulate permeability through the creation of nano-porosities [7,42,57,58]. therefore, it is possible to provide mechanical stability to functional cell cultures and to generate a network designed for the transport of substances and to simulate specific desired (even pathogenic) organ-organ interaction conditions [7,26]. on the other hand, the use of hydrogels (e.g. gelatine, fibrin or collagen) within the chip channels generates hydrophilic networks that provide the system with slight selective permeability [43,59,60]. the mimicking of several epithelial structures such as blood vessels, alveoli and the blood-brain barrier have been achieved by using oocs [61]. reproduction of parenchymal tissue in a living organism, cells are specifically organized according to tissue’s function [2,7,43,62]. the arrangement of these cells is one of the main features to consider for the physiological simulation outside the native environment [8,31,42,63]. while membranes and barriers can be simulated in both 2d and 3d cultures, complex parenchymal structures cannot be mimicked in a 2d culture [1,7,8,28,43,63]. furthermore, gene expression of cells has been observed to be greater in 3d tissues compared to 2d cell cultures [64]. considering two models of cardiotoxicity assessment, an in vitro 2d model enables the evaluation of the cardiomyocytes viability in the presence of a drug but it cannot assess the functionality of the cells. nevertheless, the effect that this drug may produce in the contraction capacity of cardiomyocytes could be observed by using a 3d arrangement. an adverse reaction, such as arrhythmia, might not be detected in the 2d model and would become a potential issue on further clinical phases, but it could be noticed employing an ooc that evaluates changes in the 3d accommodation and the proper functionality of the organ [44,65]. reproduction of the interaction between organs living tissues are highly integrated, they are not isolated, in such a way that the effects in one tissue can impact other tissues [66]. intercommunication between different organs plays an important role in the safety of a drug and, thus, in the cost-benefit ratio of the clinical studies [42]. this inter-organ communication is absent in most traditional cell cultures, however, oocs allow the coupling of different organs according to their physiological disposition [42,44,51], either on a chip with different cavities or by interconnecting different chips [7,42], as shown in figure 2. the mimicking of this organ-organ interaction in ooc leads to better evaluations of absorption and metabolism [10,51]. therefore, these systems enable to observe how the organs are affected by the main compound and by the metabolites that might be generated under physiological conditions [11,67,68]. multicellular tissues and structures in living organs experience mechanical forces such as blood flow, compression, and tension, that are important for physiological processes [42,44,70,71]. blood flow is critical for cells maturation and differentiation, as well as for their proper functioning [55,70]. ooc devices enable the reproduction of some forces by modulating the flow characteristics [42,70]. this can be performed by passive manners, when talking of gravity or concentration gradient, or actively, through the integration of micro-pumps together with their respective control and measurement systems [44,54,72,73]. good simulations of cell migration from one tissue to another, as can be immune cells, or cancer cells in methasteses, are important challenges [7]. sophisticated microenvironments can also be mimicked by using multilayer technologies, parallel microchannels, cellular spheroids, turning devices, or varying the cellular arrangement pattern, among others [30]. the integration of the support and channel architectures, the 3d http://doi.org/10.5599/admet.942 r. vargas, a. egurbide-sifre and l. medina admet & dmpk 9(2) (2021) 111-141 116 cell arrangement and the flow systems enable the appropriate nutrients supply, the waste elimination and the exchange of molecular and analytical signals [13,49,50,52,58,74]. figure 2. organ interaction in ooc. a) schematic representation of miniaturized digestive system. b) heart– liver body-on-a-chip with a liver module, cardiomyocytes and a skin module on a single chip. c) body-on-achip simulation with gastrointestinal tract and liver tissue with two coupled chips. adapted respectively from [69], [67], and [68] with permission from the royal society of chemistry. copyright (2019, 2020 & 2014) measurement systems integration although several analytical methods can be coupled to ooc (see table 2 for some examples), these systems still show lower capacity to adapt to hts than 2d tissues. ooc will, however, present a higher adaptability as soon as faster and more reliable analytical systems are developed for on-line analysis [4,13,62,71,75]. table 2. main analytical techniques reported for on-line analysis of ooc. analytical technique references fluorescence [24,53,76,77][78] electrochemical detection [24,79,80] microscopy [48,80,81] chromatography [3,80] electrophoresis and magnetophoresis [3,75] high definition cameras [24] color change (ph) [28] light scattering [76] ionization spectroscopy [80] filtration [75] immuno-agglutination [76] potentiometry [24] trans-epithelial resistance [24] acoustic techniques [75] photoacoustic tomography [44] force transducers [24] flow cytometry [82] admet & dmpk 9(2) (2021) 111-141 organ-on-a-chip systems for new drugs development doi: http://doi.org/10.5599/admet.942 117 the integration of sensors in ooc systems provides an additional advantage over animal and human models as electrophysiological signals can be collected and analyzed on real-time[28,58]. this allows the quantification of cellular metabolites and study their evolution in front of external stimuli [53]. furthermore, the functionality of tissues, cells or even intracellular organelles can be tracked [80]. the main challenge of these measuring systems is to properly assess drug concentration changes and physiological responses by employing low sample quantities at low flows and with high dynamism, in a heterogeneous and complex-composition environment with the presence of salts, sugars and proteins, among others [3,28,53,80]. aside from the selection of an appropriate analytical technique, an integrated ooc system requires the development of specific computer systems dedicated to analyze and integrate the measured signals [76,83]. ooc systems production 3d printing one of the major technological contributions for the development of oocs is the 3d printing. this technique allows the construction of the desired architecture, communication channels and functional structures with cells and polymeric materials [28,44,57], which makes it possible to control specific conditions by modelling mechanical features [7,28,60]. the support structure is made of polymeric materials that must accomplish three main characteristics: biocompatibility, easy to shaping and transparency. the most commonly used polymer for this purpose is polydimethylsiloxane (pdms) [1,84]. four different techniques can be used to form ooc by 3d printing: micro-extrusion, stereolithography, inkjet bioprinting and laser bioprinting [1,28]. in micro-extrusion, a continuous flow of small drops of bio-ink (polymer and cell suspension) is generated and pressed out from a small opening whose movement is precisely controlled by a computerized system. although this technique allows the use of bio-inks with complex compositions and high viscosity, cell viability could be reduced [1,28,57]. micro-extrusion is one of the most used techniques for the formation of ooc due to its low cost, however, it shows limitations in terms of texture creation and, thus, it is not suitable for the elaboration of complex structures in heterogeneous ooc systems [58]. stereolithography creates a liquid pattern that is later polymerized using a laser or ultraviolet (uv) light. in this technique, cell viability may be affected due to the length of time of the process and the use of intense radiation [28,57,60]. however, by using stereolithography, vascularized channels can be integrated and hydrogels can be introduced in an easier way [57]. in the inkjet bioprinting technique, drops of a non-viscous bio-ink are released into a polymeric substrate obtaining a rapid formation of structures by using a low-cost process. there is, however, a lack of precision in the direction of the drop and in the control of the viscosity, which makes the construction of complex designs difficult [28,85]. finally, laser bioprinting uses a laser beam to push a gelled solution of cells onto a substrate. this technique offers high precision in tissue production and a large variety of materials can be used. however, cells viability can be affected by the high temperatures produced by the laser during the long processing time [1,58]. cell lines obtention the correct selection of biological sources for the production of cellular tissues is a key factor to obtain a http://doi.org/10.5599/admet.942 r. vargas, a. egurbide-sifre and l. medina admet & dmpk 9(2) (2021) 111-141 118 greater similarity between occs and human tissues. the choice of the source will depend on the organ and the genetic background aimed to be studied [2,26,42]. both primary tissue cells and immortalized lines can be used for these purposes, although the latter may show more difficulties performing the physiological functions of primary cells [42]. numerous authors have cited the use of induced pluripotent stem cells (ipscs) for regenerative medicine, drug discovery and disease modelling [2,3,5,42,74,86–91]. these cells show an extraordinary organizational capacity that allows them to be modelled to any physiological or pathological condition [2,62,74,86]. figure 3 shows an outline of ipscs preparation from various sources by using different reprogramming methods, including crisp-cas9 genome editing [92]. the integration of ipscs in an ooc can be a useful tool for personalized medicine. these kind of cells have a great capacity to induce genetic variability [65,90] and make it possible to execute specific pathological evaluations, which cannot be performed through current alternatives [2,5]. musah et al. established protocols to differentiate ipscs into mature-functional podocytes and built a human glomerulus in a chip with the aims of nephrotoxicity screening, therapeutic development and the study of kidney development and renal impairment [87]. smith et al. highlighted methods to produce ipsc-derived cardiomyocytes for efficacy/toxicity screening and modelling of cardiac diseases [65]. moreover, wang et al. developed liver organoids that showed hepatic functions from ipscs [89]. in another approach, vatine et al. used patient-derived ipscs from individuals with neurological diseases to simulate the disorder’s specific physiopathology [74]. figure 3. diagram of ipsc obtainment from i) fibroblasts, keratinocytes or melanocytes obtained in skin biopsy, ii) cd34+ from blood samples, iii) cd133+ from umbilical cord or iv) multipotent cell from adipose tissue; ipsc reprogrammed towards pluripotent cells through different induction mechanisms such as: v) microrna delivery, vi) viral transfection, vii) integration vectors or viii) protein transfection. adapted from [40] under terms of the creative commons attribution license. copyright (2017). applications oocs are currently used in preclinical phases as a complement to animal models and cell cultures [1,3]. 2d models cannot be used for drug evaluation in terms of metabolic activity, efficacy or systemic toxicity [31,42,43,63]. nevertheless, oocs have the potential to replace traditional experimental models, as they have shown better in vivo/in vitro extrapolation [6,51,93] and potentially represent a more favorable admet & dmpk 9(2) (2021) 111-141 organ-on-a-chip systems for new drugs development doi: http://doi.org/10.5599/admet.942 119 risk/benefit value in drug research and development [18]. these systems present outstanding features such as the chance to simulate dynamic multi-organ environment [1,2,81,94] and to replace the use of humans in some phases of clinical studies [80]. tables 3 and 4 summarize some examples of the applications of ooc technologies for organs, systems and pathophysiological conditions. the following sections point out additional features for some of the most advanced applications with the greatest impact on academic production. table 3. overview of some functionalities and applications of organs or systems on a chip. modified from [1,2]. organ or system main components of the ooc functionality preclinical application references heart contractible 3d conformation. incorporation of structural proteins. electrical stimulation. regulated mitochondrial distribution. simulation of frank-starling mechanics in cardiomyocytes. demonstration of auxotonic contractions. cardiotoxicity tests. study of physiological phenomena involved in cardiac functioning. [2][4][6][7][11][18] [21] [44][71][70] [65] [82] [84][95] [96] kidney barriers formation. permeability modulation with different types of proteins. generation of structural proteins. functional assessment under mechanical stress conditions. separation of tubular flow and interstitial fluids. selective filtration in kidney models. simulation of type ii diabetes mellitus nephropathy. evaluation of drug-induced nephrotoxicity. research on the physiological role of specific proteins. study of viral infections. [1][2][4][13] [21] [45][87][75][84] [97][98] lung simulation of the alveolus/capillary interface. modulation of cell permeability. presence of intercellular proteins. simulation of air-water interfaces with or without the presence of airflow. breathing simulation using cyclic mechanical stress. evaluation of resistance mechanisms in lung cancer. lung proteins study. metabolism of drugs. systemic toxicity studies. [1][13][18][21][29] [42][52][54][71] [99][100] [101] gut differentiation of the intestinal epithelium. reproduction of peristalsis. incorporation of extracellular proteins. reproduction of intracellular and paracellular transport. microbiota interactions studies. pathogenic bacteria models. study of viral infections. simulation of barrier function loss. drug absorption and metabolism evaluation. [1][2][4][7][8][13] [31][51][52][69] [72][102][103] [104] [105][106] liver 3d cultures with specific liver substructures. kupffer cell incorporation. incorporation of structural proteins. secretion and production of high levels of urea and albumin. enzymatic activity. production of bile and formation of gallbladder. drug metabolism studies. hepatotoxic effect of drugs and toxicity effects on other organs (including the influence of liver metabolites). glycogenesis studies (and inhibition by drugs). fatty liver drug development. [2][4][11][13][43] [51][67][68][71] [89] [99][104][107] [108] brain/central nervous system blood-brain barrier (bbb) formation in the presence of astrocytes and neurons, and with modulated cell permeability. formation of tubular vessels. structural protein incorporation. brain folding. replication of the neurovascular unit. study of neurodegenerative diseases. inflammation studies by lipopolysaccharide endotoxins or others. study of the influence of drugs in the prenatal period. determination of permeability of drugs and nanoparticles through bbb. [1][2][4][21][48] [49][64][49][74] [75][88][109] http://doi.org/10.5599/admet.942 r. vargas, a. egurbide-sifre and l. medina admet & dmpk 9(2) (2021) 111-141 120 table 3. continued… organ or system main components of the ooc functionality preclinical application references placenta formation of the placental barrier. permeability modulation for high molecular weight proteins. division of fetal and maternal chambers with fluid flow. transfer of nutrients and glucose to fetal compartments. evaluation of placental responses. simulation of drug transport across the placenta using cancer-derived cells. study of the influence of drugs in the prenatal period. [1][21][110][111] adipose tissue creation of adipose spheres that simulate adipose tissue in vivo. vascular-adipose tissue interface. fatty acid absorption. glucose uptake studies. obesity models. development of cell retention methods. [1][96][102] retina retinal pigment formation. development of the epitheliumchoroid structure. presence of structural proteins. pathologies models. reproduction of imaging processes. [1][62][112][113] muscle contractile units through chemical or electrical stimuli. evaluation of the effect and toxicity of drugs. study of degenerative diseases and muscle physiology. [2][114][115] others immune system [50,71], bone [116,117], breast tissue [118,119], pancreas [120,121], bladder, reproductive tract [104], skin [13]. heart-on-a-chip cardiotoxicity is one of the most common causes of drug failures in phase i clinical trials [13], highlighting the necessity of more accurate prediction models for a preclinical screening. an interesting number of ooc models simulating the human heart have been developed with different levels of complexity. its main application is centered on the evaluation of the contractile and rhythmic functions [6,7,13,18,44,122]. the ooc application for heart reproduction can be a powerful tool in the preclinical evaluation of a drug, as it allows to connect the heart with the lungs and the liver and to detect toxic reactions that may be absent when the drug is assessed in an isolated tissue [1,11,65]. schroer et al. published a work focused on contractile characteristics and parameter determination for personalized cardiac cell tissues [95]. the heart-on-a-chip developed by abulaiti et al, (figure 4) allowed the evaluation of heart tissue function and visualization of cardiac micro-tissue kinetics by monitoring the displacement of fluorescent particles incorporated within one of the device’s compartments through fluorescence microscopy, flow cytometry and high resolution cameras [82]. kidney-on-a-chip the kidney is the main organ for hemofiltration, it is involved in waste excretion and it is indispensable for homeostasis. therefore, the study of the renal function by using ooc can be a very interesting tool for drug development and for the determination of the systemic safety profile of drugs [2]. the greatest challenge in developing functional kidneys-on-a-chip is the simulation of the tubular structure with a sophisticated control of the physical parameters involved in the glomerular filtration [1,2,87]. the introduction of an on-line renal function detection unit in the ooc allows compounds at very low concentrations to be detected in the system [1]. admet & dmpk 9(2) (2021) 111-141 organ-on-a-chip systems for new drugs development doi: http://doi.org/10.5599/admet.942 121 figure 4. heart-on-a-chip for the evaluation of physiological parameters. a) device diagram and structure. graphs b) and c) respectively show the effect of two different isoprotenol concentrations on the heart rate and the magnitude of contractile displacement, compared to no treatment values. adapted from [82] under terms of the creative commons attribution license. copyright (2020). some examples of this technology can be found in the work of yin et al, who developed a kidney-on-achip device for drug screening and nephrotoxicity assessment (figure 5). they achieved tissue differentiation and functionality by growing the two cell cultures on opposite membranes, and evaluated the cell growth and viability using fluorescent markers [97]. on the other hand, jang et al. built a system for drug transport and nephrotoxicity assessment in preclinical phases that mimicked key functions of the human kidney proximal tubule. the chips architecture exposed cell layers to a fluid shear stress similar of that found in living kidney tubules, resulting in an enhanced epithelial cell polarization, differentiation and function compared to traditional transwell culture system [98]. wang et al. developed a kidney-on-a-chip for the study of virus pathogenic mechanisms inducing renal infections, which allowed in-line cell characterization using confocal imaging and scanning electron microscope imaging [45]. lung-on-a-chip modern ooc systems are inspired by the pioneering development of a lung-on-a-chip. the main challenges for an accurate lung function are the generation of biomechanical ventilation, reproduction of lung mucosa and the capability to overcome the possible prothrombotic interference and the rigidity of the simulated lung tissue [1,13]. stucki et al. developed a system that enables the imitation of the alveolar parenchyma and that can produce a mechanical stretching that mimics the breathing mechanism using a pneumatic micro-diaphragm. they demonstrated that the mechanical stress affects the epithelial barrier permeability [99], which id something to consider when manufacturing other lung models reproducing somehow this kind of mechanical stress. felder et al. studied wound healing recreating an alveolar epithelium on a chip also including mechanical strain with the incorporation of an ultra-thin elastic membrane (figure 6). this device used immunofluorescence microscopy and scanning electron microscopy for the evaluation of test outcomes [100]. the lung-on-a-chip developed by asad et al. incorporated on line detection of ph and trans-epithelial electrical impedance, as well as a microscope that allows visual monitoring due to the chip’s transparency [101]. http://doi.org/10.5599/admet.942 r. vargas, a. egurbide-sifre and l. medina admet & dmpk 9(2) (2021) 111-141 122 figure 5. kidney-on-a-chip device for drug screening and nephrotoxicity assessment. a) schematic representation of the chip’s two chambers. b) diagram of the complete device including: the chip and the temperature and the flow control device. adapted from [97] under terms of the creative commons attribution license. copyright (2020). gut-on-a-chip the small intestine has important functions associated with drug absorption and the immune system [2,7]. applications for ooc-gut models include pharmacodynamics, pharmacokinetics and physiological studies, and are frequently linked with the use of caco2 cells. these gut-on-a-chip models enable to investigate the way how the tissue-microbiota interrelationship can affect the nutrient absorption in the gut, and allow the study of the signaling between gastric and intestinal cells [1,54,72,103]. ooc-gut models have also been used to evaluate the mucus formation, accumulation and impact on intestinal functionality [31,72]. peters et al. highlighted the importance of in vitro testing to assess gastrointestinal safety [54]. in another approach, de haan et al. monitored enzymatic digestion of milk using a microfluidic chip and demonstrated the importance of adjusting and correlating digestion times at physiological conditions on ooc [69]. sailer et al. built a gut-on-a-chip from patient-derived intestinal subepithelial myofibroblasts, which incorporated growth factors and controlled physiological conditions that showed angiogenic properties under simulated perfusion (figure 7) [105]. the gut-on-a-chip also can be found coupled with other organs, such as the liver, which allows to study the absorption and metabolism of drugs [51]. admet & dmpk 9(2) (2021) 111-141 organ-on-a-chip systems for new drugs development doi: http://doi.org/10.5599/admet.942 123 figure 6. mimicking of breathing mechanics. schematic representation of a) the movement of the diaphragm leading to the lungs expansion during inhalation and b) the imitation of this motion by a mechanical microdiaphragm used in lung-on-a-chip devices. reproduced from [100] under terms of the creative commons attribution license. copyright (2019) figure 7. gut-on-a-chip (goc) formed by two channels separated by a thin membrane with perfusion vasculature on the lower channel. adapted from [105] under terms of the creative commons attribution license. copyright (2020). http://doi.org/10.5599/admet.942 r. vargas, a. egurbide-sifre and l. medina admet & dmpk 9(2) (2021) 111-141 124 liver-on-a-chip the liver plays a crucial role in the drug metabolism and the detoxification of the body. therefore, it is another important organ to take into consideration during the development of new drugs and their effects in the other parts of the system [2,11,21,89]. liver cells grown under 3d conditions are more viable, functional and reproducible than those cultured in 2d [2,108]. liver-on-a-chip models associated with other tissues or organs, such as the intestine, may allow to obtain first-step metabolism profiles closer to those that could be acquired through traditional cell cultures [1,3,89]. jia et al. described a 3d hepatic system for liver injury studies by controlling hepatocyte viability, function, morphology and polarity, and thus, maintaining the tissue morphology, mrna expression, bile secretion and a better cell viability compared to other types of cell cultures. they controlled the morphology encapsulating the cell culture inside a alginate hydrogel (figure 8a) [43]. freyer et al. built an ooc model for hepatotoxicity testing, highlighting the importance of the control and monitoring of chemical physiological parameters (such as lactate production and ammonia release) (figure 8b). they validated the ooc by performing a paracetamol hepatotoxicity determination, which included a histological and immunohistochemical evaluation of the damage. they also stablished the need of modifications of the ooc to include non-parenchymal cell types at physiological ratios and the creation of defined oxygen gradients in order to increase the system’s sensibility for further toxicity studies [107]. on another approach, kostrzewski et al. prepared a 3d culture on a plate from primary cells that showed positive progression on fatty liver recovery by using known anti-steatosis compounds. this model showed adipokine expression and metabolic responses similar to those seen in clinical non-alcoholic fatty liver disease. the chip is not designed for on-line quantifications and most of the analytical determinations had to be made after removing the microtissues from the chip [108]. as well as the heart or the intestine, the liver has been studied in multiorgan chips to investigate systemic effects, absorption and metabolism [51,67,68]. central nervous system-, brainor blood brain barrier (bbb)-on-a-chip it is possible to simulate physiological and pathological mechanisms of the central nervous system (cns) in an ooc [2,88]. these models can be very useful considering the complexity of the human brain, which makes it difficult to study the cns in animal models or 2d cultures [13].the blood-brain barrier plays a vital role in homeostasis and it is the main limitation for the administration of drugs in the brain [109]. the development of integrated systems reproducing the barrier properties and modelling inter-individual variation [2,15,74] makes it possible to assess whether a drug designed to treat a neuro-related disease can actually reach its therapeutic target [13,78]. adriani et al. were able to develop a chip with neurons and astrocytes that showed adequate morphology, growth and vascular function, identifying the opportunity to include other types of cells such as pericytes and microglia, which increased the model’s potential to be used for neurovascular studies and too assess drug effects on brain neural function [48]. the brain endothelium created from primary patient cells by vatine et al. reproduced complex bbb functions with the potential applications of drug screening and personalized medicine, using immunocytochemistry, confocal microscopy, calcium imaging and mass chromatography as analytical methods, among others [74]. tu et al. developed a bbb system that incorporated transendothelial electrical resistance (teer) micro-electrodes to monitor cell growth and response in the presence of drugs (figure 9) [123]. admet & dmpk 9(2) (2021) 111-141 organ-on-a-chip systems for new drugs development doi: http://doi.org/10.5599/admet.942 125 figure 8. schematic examples of liver-on-a-chip devices. a) multichannel ooc. representation of a i) microfluidic device with several channels of different compositions for the formation of encapsulated hepatic plate structures, and a ii) system of syringe pumps supplying the microfluidic channels of the ooc with solutions. adapted from [43] under terms of the creative commons attribution license. copyright (2020). b) diagram of a pressure-driven flow control system of a 3d liver bioreactor for hepatotoxicity testing under perfusion conditions. adapted from [107] under terms of the creative commons attribution license. copyright (2018). figure 9. bbb-on-a-chip with microelectrodes incorporated. a) diagram of the main channel. b) diagram of the electrodes and the chip structure. c) photograph of the integrated system. adapted from [123] under terms of the creative commons attribution license. copyright (2020). http://doi.org/10.5599/admet.942 r. vargas, a. egurbide-sifre and l. medina admet & dmpk 9(2) (2021) 111-141 126 table 4. main pathological conditions simulated in ooc systems pathology references cancer [13–15,23,42,50,63,122,124–127] inflammation [7,128–130] epilepsy [44] [64] alzheimer's disease [64] musculoskeletal abnormalities [23] arthritis [23] arrhythmias [44] diabetes [102] hematological diseases [23] pneumonia [44] parkinson [64] thrombosis [23] inflammation-on-a-chip, or immune system-on-a-chip ooc can also reproduce complex physiological events or multi-organ functions [7,128]. the development of the inflammatory system-on-a-chip enables a better understanding of how immune system cells synchronize their activities to fight a disease. the evaluation of the coordinated cellular activities related to inflammation could have interesting biological and clinical applications [128]. autoimmune diseases or immunity generated by vaccines can also be assessed by ooc systems [50,71,102]. sharifi et al. developed a system on a chip for modeling the immune response cascade of events during immune cell response to implants [130], and goyal et al. described a microfluidic chip for evaluation of vaccine response that was able to obtain antibodies and can be useful for immunotherapy candidate selection [130]. gopalakrishnan et al. worked on a real-time continuous monitoring chip (using microscopy and fluorescence) for studying cellular migration and organization during inflammatory processes in response to simulated cytokine gradients (figure 10) [131]. figure 10. microfluidic device allowing cell migration. a) simulation of an in vivo scenario with the chip construction. b) schematic of the device’s manufacture by soft lithography. reproduced from [131] with permission from the royal society of chemistry. copyright (2015). cancer-on-a-chip cancer models in mice have resulted in countless treatments to cure cancer in the animals, although they present an alarming difficulty to transfer these results to humans. ooc systems provide the possibility to study the physio-pathological mechanisms and the resistance reactions in an environment almost identical not only to human disease but to the specific individual. therefore, these systems can open the door to many treatments against cancer [13,14]. admet & dmpk 9(2) (2021) 111-141 organ-on-a-chip systems for new drugs development doi: http://doi.org/10.5599/admet.942 127 scientists have been able to modulate through ooc devices phenomena such as metastatic cells migration, tumor growth or even drug resistance [124,132]. moreover, potential therapeutic targets and possible chemical markers have been identified in preliminary research based on ooc [15]. nguyen et al. used an ooc model to demonstrate pathogenic mechanisms involved with the endothelial changes observed in pancreatic cancer, as well as to identify chemical mediators of these phenomena (figure 11) [125]. shuford et al. used primary patient tissue on a chip for the prediction of patient-specific response before clinical treatment in ovarian cancer [126]. choi et al. designed a device with the microarchitecture of breast carcinoma to be used as drug screening platform. they validated the screening capacity using it to evaluate efficacy and toxicity of paclitaxel [127]. body-on-a-chip or human-on-a-chip for adme evaluation and preclinical development oocs can assembly multiple tissues and simulate the systematic functionality of the human body. as long as each organ shows the most relevant functions associated with the pharmacokinetic phenomena to be studied, these systems have the potential to determine the pharmacokinetic profile of drugs [7,10,13,42,50,102]. the processes of absorption, distribution, excretion and metabolism (adme) can be simulated from various routes of administration and, moreover, mathematical models can be applied to make real-body predictions [10,27,51,66,132,133]. this approach has been used to evaluate the adme profile of drugs and the appearance of physiological events that cause adverse effects in devices of up to 20 different tissues [7,10,32]. furthermore, they can also be used to assess a compound effectiveness [13,133]. de melo et al. developed a heart-liver-skin device for the systemic evaluation of drugs administrated topically and for the comparison of acute and systemic toxicity (figure 2b) [67]. figure 11. pancreatic-cancer-on-a-chip. a) i: image of a pancreatic duct (seeded with pancreatic cancer cells) and a blood vessel (seeded with endothelial cells) nested within a collagen matrix that are shown on the right image; ii: representation of the cross section of the right image. b) invasion distance that cancer cells traverse within the collagen matrix as a function of time, with and without the addition of human umbilical vein endothelial cells (huvecs). c) and d) invasion of tumoral cells (green) into the blood vessel (red). e) i: invasion of the blood vessel by cancer cells, inducing to the apoptosis (marked in white in ii) of endothelial cells. modified from [125] under terms of the creative commons attribution license. copyright (2019). http://doi.org/10.5599/admet.942 r. vargas, a. egurbide-sifre and l. medina admet & dmpk 9(2) (2021) 111-141 128 the system developed by rajan et al. including tissues from the brain, liver, lung, heart and endothelium showed metabolizing capacity for prodrug transformation, as well as representativity of cardiac toxicity [10]. marin et al. studied acetaminophen absorption and metabolism with an intestine/liver microphysiological system demonstrating the potential of these kind of systems for pharmacokinetic profiling of drug substances [51]. the cardiotoxicity of the main compound and the metabolites was evaluated with an ooc device by oleaga et al., who developed a model that can be used in preclinical evaluation and for chronic studies [11]. maschmeyer et al. proposed one of the first devices to integrate intestine, liver, skin and kidney organoids for adme profiling, and repeated dose systemic toxicity testing of drug candidates, which was sustainable throughout 28 days [134]. the system of fluidically coupled vascularized chips proposed by herland et al. allowed modeling the pharmacokinetics of orally and intravenous administered drugs. their arrangement was able to couple with robotic pumps, chips that represent the gut, liver, kidney, and bone marrow, and presented an arteriovenous fluid-mixing reservoir. they obtained pharmacokinetic results similar to previously reported clinical data [135]. one of the most important challenges in the body-on-a-chip development is the proportional replication of organs in terms of size, blood flow and metabolic rates. however, in many cases, these proportions are still unknown for the human body [7]. another challenge of these ooc devices is that they represent a small number of cells, and their intercommunication substances undergo under significant dilution conditions [42]. nearly a 70 % of drug withdrawals correspond to hepato-cardio toxicity, therefore, it is believed that a system that included at least those organs would be enough for a preclinical safety screening [136]. oocs could be used in the future for clinical studies by introducing specific variables needed according to personalized groups of patients. moreover, this might allow studies of pathologies for which there are usually not enough candidates for a representative clinical trial [13]. the use of ooc systems in the development of new drugs would enable designs based on better physiological understanding and better candidate screening [86]. this would lead to an increase in drug development efficiency by reducing costs (up to 25 %) and time for the experimental phases [9,18,29], lowering the risk for patients and raising the investment in new drugs production [13]. this could increase the number of new discoveries and reduce their prices, benefiting, therefore, the health systems at various levels [9,18,20,71]. challenges and perspectives in order to exploit the full potential of ooc systems and make them become a validated alternative, several physical, physiological and regulatory challenges must be overcome. the first step to obtain sophisticated and accurate ooc with a minimum functional unit [13,133] is to have a deep knowledge of the organ’s physiology, the cellular metabolism and all pathological mechanisms that can affect the organ [31,44,50,53]. table 5 shows the main challenges or advances required in the manufacture of ooc systems for some specific organs. table 5. main challenges for the improvement of some specific organ-on-a-chip systems: modified from [1,7]. organ main challenges kidney formation of glomeruli with podocytes and entangled proximal tubules. lung generation of biomechanical ventilation without assistance. further studies with inflammatory and immunological processes and tumors. heart establishment of validated studies for reproducible cardiotoxicity with various drugs. gut deepening knowledge of mesenchymal-epithelial interactions. placenta modelling with primary cells. adipose tissue more knowledge of the functioning of brown fat in adipose tissue. brain standardization and reproducibility of brain organoids. admet & dmpk 9(2) (2021) 111-141 organ-on-a-chip systems for new drugs development doi: http://doi.org/10.5599/admet.942 129 as a field that influences a wide spectrum of disciplines [22], oocs require significant advances in the availability and the implementation of the newest trends from the scientific areas related to their design, development, application and production. overcoming the individual challenges is the key for translating ooc to cost-effective and sustainable implementation. we briefly introduce some of those relevant challenges, however, more insight into each research area can be found on specific reviews developed for this purpose, which can be found elsewhere [33–41,137,138]. challenges in fluid mechanics microfluidics are essential for the reproduction of the organ’s conditions in an ooc. the use of external systems is critical [73], as well as the 3d arrangement of the microchannels. many of the current devices present a gravity-mediated flow distribution [7,11], however, the development of better microfluidic control systems is essential to improve the reproduction and to control the microflows that generate more bio-relevant conditions within the chips [42,73]. the development of microfluidic is, perhaps, the most important driver in the progress to more complex, robust, significant and scalable oocs. for example, hts using ooc devices requires improvements in microfluidics that allow a better controlled perfusion, sampling and cell injection [18,139], while on the other hand, stem cell’s experimentation, production and analysis requires high level of microenvironmental control [140]. chips with integrated micro pumps and, with structural and mechanical considerations that allow cell circulation and passive flow control are some of the microfluidic topics with more progress in the past years [41]. challenges on cell availability physiological reproducibility of ooc mainly depends on the cells that are employed to build them. cell availability challenges in ooc are the same restrictions than those related to cellular tissues: limited access to primary cells and possible dysfunction of immortalized cells [7,42]. to have different cell culture types in the same environment brings to an extra difficulty, which is shared with regenerative medicine techniques [28,90]. obtaining reproducible tissues, suitable for application in ooc systems, requires the development of more and better techniques for the culture, maturation and differentiation of pluripotent cells, as well as the standardization of the obtention of protocols [5,55,90]. major breakthroughs in this field include relevant experimentation on novel and standardized methods to obtain and differentiate stem cells into functional organ models [141–143], and the continuous efforts for the development of a blood substitute, a representative universal cell medium for ooc [41]. challenges in materials engineering in this area it is required to develop materials that accomplish all –or most of– the application criteria in an ooc, such as plasticity, elasticity, transparency, biocompatibility or versatility [7,63]. these materials must be inert to the drugs to evaluate [44,75] and exhibit easy printing, while achieving adequate mimicry with living tissues [28]. the materials that are currently used do not meet all the requirements simultaneously. for example, pdms is one of the most widely used materials to manufacture ooc due to its versatility and biocompatibility, nevertheless, it has shown a tendency to absorb small molecules, such as many of the drugs evaluated [10,18,30,144]. therefore, the discovery of intelligent materials that can change their shape and reactivity in front of different stimuli, known as four-dimensional printing, is a promising strategy for the future ooc [58]. recent advances include the introduction of a variety of biocompatible and thermoplastic materials that are presumably more suitable for large scale manufacturing processes [41]. http://doi.org/10.5599/admet.942 r. vargas, a. egurbide-sifre and l. medina admet & dmpk 9(2) (2021) 111-141 130 challenges in measurement systems the evolution of ooc requires an improvement in the development of specific electrodes and/or biosensors [7,28] with high performance, high sensitivity and high specificity [18,53,80]. the high sensitivity is required due to the low amount of metabolic secretions [3]. moreover, it must be possible to measure the tissue response at many other levels, as well as the concentration variations of the specific substance [42]. therefore, ooc systems must have the ability to detect small non-lethal changes that affect functionality at the cellular level [7,41]. more feasible, adaptable to hts and scalable bio-assays also require microfluid regulators such as valves, pumps, mixers and other functional elements to get cell perfusion with fresh media and assay reagents [139]. these systems must allow non-invasive parallel measurements, automation and flexibility in order to fit hts programs used by big pharma companies to evaluate their compound libraries [4,18,29,41,75]. commercially available micro-sensors of thermos responsive materials and electrodes, as well as optical methods for ph and oxygen determination, have been successfully incorporated to ooc devices, [41] in addition to several prototypes of organic biotransistors [145]. in the particular case of a bbb-on-a-chip, teer has become the most popular online sense method. but other alternatives such as physical sensors, with the ability of measuring small concentrations of metabolites, and a variety of biosensors, bioreceptors that generate an electric signal after binding a biological target, have been proposed (figure 12) [146]. software-driven analyses of optical recordings have been described for several cellular functions such as the heart rate [82], which can be used to evaluate other related muscular functions. figure 12. biosensors and sensors on bbb-on-a-chip transducing signals to a pc or mobile interface. reproduced from [146] under terms of the creative commons attribution license. copyright (2019). challenges on industrial scale production there are many ooc prototypes that have been presented only as a pilot test at laboratory level. the success of an ooc system in solving real problems depends on how rigorously it fits the real model, but furthermore, on how easily it can be transferred to industrial production. in many cases, the ooc prototype admet & dmpk 9(2) (2021) 111-141 organ-on-a-chip systems for new drugs development doi: http://doi.org/10.5599/admet.942 131 is obtained through an effective 3d printing technique, but it is inefficient for large scale due to the low speed at which the structures are created [7,11,28]. low cost or reusable materials could lead to a more widespread use of ooc in early stages of drug discovery [143]. massive parallelization and automation of ooc systems are among the biggest challenges to overcome in order to meet industrial requirements of hts [18]. the work of ramadan and zourob [137] contains a comprehensive review not only of the main challenges, but of specific strategies to solve them, trends and examples of the main industrialization activities of some of the developing ooc systems. the more sophisticated an ooc becomes, the more it needs to be controlled, which increases its cost and limits its usege. to ensure reproducibility of oocs [49] specific tests must be developed in parallel for the quality control of the devices [133]. besides any limitation, the number of commercially available ooc devices, new start-up companies and patent applications have considerably increased these last years [136]. ooc validation the validation of ooc systems is particularly important for regulatory agencies to accept them as clinical and experimental alternatives, especially in toxicity trials [13,42,43]. so far, many ooc developments correspond to isolated publications with a lack of standardized protocols for operation, data collection and analyses [63]. direct comparisons require specific clinical studies and long-term gathered data. thus, many ooc system improvements are performed in collaboration between academia, industry and regulatory authorities [7,50]. regulatory agencies, academia and several companies have been conducting efforts to establish industrial standards and validation guidelines for ooc technologies [136]. the replication of already-conducted trials executed with approved products and the comparison of their results with the results obtained with approved assays is one way of achieving the acceptance of ooc systems as alternative methods for pre-clinical –or even clinical– studies [7,136]. ooc could be used to replicate cases of clinical studies in which severe adverse events were not detected pre-clinical stages and drugs had to be withdrawn from the market, therefore, it could be determined if this technology has a greater or equal predictive capacity [44]. conclusions there is growing evidence that ooc systems are a novel tool that will facilitate the search, development and evaluation of new drugs. these systems show characteristics of reproducibility, adaptability and versatility that make them a potential alternative to the use of experimental animals. oocs are an evolution of 2d cell cultures, thus maintaining some of the characteristics of cell tissue evaluations. however, these new systems overcome many of the conventional cell cultures weaknesses, such as the poor extrapolation to an entire organism. ooc simulate the real microenvironment and the key functional aspects of organs at a microscopic scale. this enables the study of drugs as well as the investigation of physiological and physio-pathological phenomena. the construction of ooc requires a deep understanding of the real functionality of the tissue/organ to be studied and mimicked. in many cases, this knowledge is limited, thus leading to a synergy between the research in physiology and the development of oocs. these systems could –scientifically and financially– promote the development of new drugs. risk reduction could revolutionize the way big pharmaceutical companies focus their development efforts, as well as the approach of how clinical and preclinical studies are conducted. http://doi.org/10.5599/admet.942 r. vargas, a. egurbide-sifre and l. medina admet & dmpk 9(2) (2021) 111-141 132 ooc's progress efforts must converge on the development of systems with enough representativeness, reproducibility, scalability and suitability for hts, which maintain at the same time reasonable production and implementations costs. although science is far from being able to mimic at 100 % the human body in a laboratory, any advance will increase the predictive potential of ooc systems and will provide innumerable benefits for new drugs development. acknowledgements: this work was possible thanks to the grant of the office of international affairs of the university of costa rica (oaice-64-2019). the authors would like to acknowledge lic. pablo vargas monge for his contribution with the elaboration of the graphical abstract. conflict of interest: the authors have no conflicts of interest to declare. references [1] r. mittal, f.w. woo, c.s. castro, m.a. cohen, j. karanxha, j. mittal, t. chhibber, v.m. jhaveri. organon-chip models: implications in drug discovery and clinical applications. journal of cellular physiology 234 (2019) 8352-8380. https://doi.org/10.1002/jcp.27729. [2] y.a. jodat, m.g. kang, k. kiaee, g.j. kim, a.f.h. martinez, a. rosenkranz, h. bae, s.r. shin. humanderived organ-on-a-chip for personalized drug development. current pharmaceutical design 24 (2019) 5471-5486. https://doi.org/10.2174/1381612825666190308150055. [3] a. lin, f. sved skottvoll, s. rayner, s. pedersen-bjergaard, g. sullivan, s. krauss, s. ray wilson, s. harrison. 3d cell culture models and organ-on-a-chip: meet separation science and mass spectrometry. electrophoresis 41 (2020) 56-64. https://doi.org/10.1002/elps.201900170. [4] j.a. kim, s. hong, w.j. rhee. microfluidic three-dimensional cell culture of stem cells for highthroughput analysis. world j stem cells 11 (2019) 803-816. https://doi.org/10.4252/wjsc.v11.i10.803. [5] c.c. miranda, t.g. fernandes, m.m. diogo, j.m.s. cabral. towards multi-organoid systems for drug screening applications. bioengineering 5 (2018) 1-17. https://doi.org/10.3390/bioengineering5030049. [6] t. meyer, m. tiburcy, w.h. zimmermann. cardiac macrotissues-on-a-plate models for phenotypic drug screens. advanced drug delivery reviews 140 (2019) 93-100. https://doi.org/10.1016/j.addr.2019.03.002. [7] b. zhang, a. korolj, b.f.l. lai, m. radisic. advances in organ-on-a-chip engineering. nature reviews materials 3 (2018) 257-278. https://doi.org/10.1038/s41578-018-0034-7. [8] l.h.m. van de burgwal, p. van dorst, h. viëtor, r. luttge, e. claassen. hybrid business models for ‘organ-on-a-chip’ technology: the best of both worlds. pharmanutrition 6 (2018) 55-63. https://doi.org/10.1016/j.phanu.2018.02.001. [9] n. franzen, w.h. van harten, v.p. retèl, p. loskill, j. van den eijnden-van raaij, m. ijzerman. impact of organ-on-a-chip technology on pharmaceutical r&d costs. drug discovery today 24 (2019) 17201724. https://doi.org/10.1016/j.drudis.2019.06.003. [10] s.a.p. rajan, j. aleman, m.m. wan, n. pourhabibi zarandi, g. nzou, s. murphy, c.e. bishop, h. sadriardekani, t. shupe, a. atala, a.r. hall, a. skardal. probing prodrug metabolism and reciprocal toxicity with an integrated and humanized multi-tissue organ-on-a-chip platform. acta biomaterialia 106 (2020) 124-135. https://doi.org/10.1016/j.actbio.2020.02.015. [11] c. oleaga, a. riu, s. rothemund, a. lavado, c.w. mcaleer, c.j. long, k. persaud, n.s. narasimhan, m. tran, j. roles, c.a. carmona-moran, t. sasserath, d.h. elbrecht, l. kumanchik, l.r. bridges, c. martin, m.t. schnepper, g. ekman, m. jackson, y.i. wang, r. note, j. langer, s. teissier, j.j. hickman. investigation of the effect of hepatic metabolism on off-target cardiotoxicity in a multi-organ humanhttps://doi.org/10.1002/jcp.27729 https://doi.org/10.2174/1381612825666190308150055 https://doi.org/10.1002/elps.201900170 https://doi.org/10.4252/wjsc.v11.i10.803 https://doi.org/10.1038/s41578-018-0034-7 https://doi.org/10.1016/j.phanu.2018.02.001 https://doi.org/10.1016/j.drudis.2019.06.003 https://doi.org/10.1016/j.actbio.2020.02.015 admet & dmpk 9(2) (2021) 111-141 organ-on-a-chip systems for new drugs development doi: http://doi.org/10.5599/admet.942 133 on-a-chip system. biomaterials 182 (2018) 176-190. https://doi.org/10.1016/j.biomaterials.2018.07.062. [12] k. viravaidya, a. sin, m.l. shuler. development of a microscale cell culture analog to probe naphthalene toxicity. biotechnology progress 20 (2004) 316-323. https://doi.org/10.1021/bp0341996. [13] k. ronaldson-bouchard, g. vunjak-novakovic. organs-on-a-chip: a fast track for engineered human tissues in drug development. cell stem cell 22 (2018) 310-324. https://doi.org/10.1016/j.stem.2018.02.011. [14] j.w. choi, s.-y. lee, d.w. lee. a cancer spheroid array chip for selecting effective drug. micromachines 10 (2019) 688. https://doi.org/10.3390/mi10100688. [15] w. liu, j. song, x. du, y. zhou, y. li, r. li, l. lyu, y. he, j. hao, j. ben, w. wang, h. shi, q. wang. akr1b10 (aldo-keto reductase family 1 b10) promotes brain metastasis of lung cancer cells in a multi-organ microfluidic chip model. acta biomaterialia 91 (2019) 195-208. https://doi.org/10.1016/j.actbio.2019.04.053. [16] v. monteil, h. kwon, p. prado, a. hagelkrüys, r.a. wimmer, m. stahl, a. leopoldi, e. garreta, c. hurtado del pozo, f. prosper, j.p. romero, g. wirnsberger, h. zhang, a.s. slutsky, r. conder, n. montserrat, a. mirazimi, j.m. penninger. inhibition of sars-cov-2 infections in engineered human tissues using clinical-grade soluble human ace2. cell 181 (2020) 905-913.e7. https://doi.org/10.1016/j.cell.2020.04.004. [17] b. zhang, m. radisic. organ-on-a-chip devices advance to market. lab on a chip 17 (2017) 23952420. https://doi.org/10.1039/c6lc01554a. [18] c. probst, s. schneider, p. loskill. high-throughput organ-on-a-chip systems: current status and remaining challenges. current opinion in biomedical engineering 6 (2018) 33-41. https://doi.org/10.1016/j.cobme.2018.02.004. [19] u. marx, t. akabane, t.b. andersson, b. elizabeth, m. beilmann, s. beken, s. brendler-schwaab, m. cirit, r. david, e.-m. dehne, i. durieux, l. ewart, s.c. fitzpatrick, o. frey, f. fuchs, l.g. griffith, g.a. hamilton, t. hartung, j. hoeng, h. hogberg, d.j. hughes, d.e. ingber, a. iskandar, t. kanamori, h. kojima, j. kuehnl, m. leist, b. li, p. loskill, d.l. mendrick, thomas neu, g. pallocca, i. rusyn, l. smirnova, t. steger-hartmann, d.a. tagle, a. tonevitsky, s. tsyb, m. trapecar, b. van de water, van den e.r. janny, p. vulto, k. watanabe, a. wolf, x. zhou, a. roth. biology-inspired microphysiological systems to advance patient benefit and animal welfare in drug development. t4 the transatlantic think tank for toxicology workshop report 37 (2020) 1-30. https://doi.org/10.14573/altex.2001241. [20] z. wang, x. he, h. qiao, p. chen. global trends of organoid and organ-on-a-chip in the past decade: a bibliometric and comparative study. tissue engineering part a 00 (2020) 1-16. https://doi.org/10.1089/ten.tea.2019.0251. [21] m. rothbauer, j.m. rosser, h. zirath, p. ertl. tomorrow today: organ-on-a-chip advances towards clinically relevant pharmaceutical and medical in vitro models. current opinion in biotechnology 55 (2019) 81-86. https://doi.org/10.1016/j.copbio.2018.08.009. [22] z. wang, x. he, h. qiao, p. chen. global trends of organoid and organ-on-a-chip in the past decade: a bibliometric and comparative study. tissue engineering part a 26 (2020) 656-671. https://doi.org/10.1089/ten.tea.2019.0251. [23] c. arrigoni, s. lopa, c. candrian, m. moretti. organs-on-a-chip as model systems for multifactorial musculoskeletal diseases. current opinion in biotechnology 63 (2020) 79-88. https://doi.org/10.1016/j.copbio.2019.12.006. [24] s.r. adam kratz, g. höll, p. schuller, p. ertl, m. rothbauer. latest trends in biosensing for microphysiological organs-on-a-chip and body-on-a-chip systems. biosensors 9 (2019). https://doi.org/10.3390/bios9030110. [25] m.m. silva. voyage from fiction to reality: organs-on-chips in the service of science and medicine. revista odontológica mexicana 20 (2016) 74-75. https://doi.org/10.1016/j.rodmex.2016.04.010. http://doi.org/10.5599/admet.942 https://doi.org/10.1016/j.biomaterials.2018.07.062 https://doi.org/10.1016/j.biomaterials.2018.07.062 https://doi.org/10.1016/j.stem.2018.02.011 https://doi.org/10.1016/j.stem.2018.02.011 https://doi.org/10.3390/mi10100688 https://doi.org/10.1016/j.cell.2020.04.004 https://doi.org/10.1016/j.cell.2020.04.004 https://doi.org/10.1039/c6lc01554a https://doi.org/10.14573/altex.2001241 https://doi.org/10.1016/j.copbio.2018.08.009 https://doi.org/10.1089/ten.tea.2019.0251 https://doi.org/10.3390/bios9030110 https://doi.org/10.1016/j.rodmex.2016.04.010 r. vargas, a. egurbide-sifre and l. medina admet & dmpk 9(2) (2021) 111-141 134 [26] f. zheng, f. fu, y. cheng, c. wang, y. zhao, z. gu. organ-on-a-chip systems: microengineering to biomimic living systems. small 12 (2016) 2253-2282. https://doi.org/10.1002/smll.201503208. [27] h. kimura, y. sakai, t. fujii. organ/body-on-a-chip based on microfluidic technology for drug discovery. drug metabolism and pharmacokinetics 33 (2018) 43-48. https://doi.org/10.1016/j.dmpk.2017.11.003. [28] j.y. park, j. jang, h.w. kang. 3d bioprinting and its application to organ-on-a-chip. microelectronic engineering 200 (2018) 1-11. https://doi.org/10.1016/j.mee.2018.08.004. [29] a. junaid, a. mashaghi, t. hankemeier, p. vulto. an end-user perspective on organ-on-a-chip: assays and usability aspects. current opinion in biomedical engineering 1 (2017) 15-22. https://doi.org/10.1016/j.cobme.2017.02.002. [30] c. tian, q. tu, w. liu, j. wang. recent advances in microfluidic technologies for organ-on-a-chip. trac trends in analytical chemistry 117 (2019) 146-156. https://doi.org/10.1016/j.trac.2019.06.005. [31] d.e. ingber. developmentally inspired human ‘organs on chips’. development (cambridge) 145 (2018) 10-13. https://doi.org/10.1242/dev.156125. [32] p. ng, c. belgur, s. barthakur, a.d. van der meer, g.a. hamilton, r. barrile. organs-on-chips: a new paradigm for safety assessment of drug-induced thrombosis. current opinion in toxicology 17 (2019) 1-8. https://doi.org/10.1016/j.cotox.2019.08.004. [33] c. tian, q. tu, w. liu, j. wang. recent advances in microfluidic technologies for organ-on-a-chip. trac trends in analytical chemistry 117 (2019) 146-156. https://doi.org/10.1016/j.trac.2019.06.005. [34] f. yu, d. choudhury. microfluidic bioprinting for organ-on-a-chip models. drug discovery today 24 (2019) 1248-1257. https://doi.org/10.1016/j.drudis.2019.03.025. [35] h. sun, y. jia, h. dong, d. dong, j. zheng. combining additive manufacturing with microfluidics: an emerging method for developing novel organs-on-chips. current opinion in chemical engineering 28 (2020) 1-9. https://doi.org/10.1016/j.coche.2019.10.006. [36] x. yu, n. wu, f. chen, j. wei, y. zhao. engineering microfluidic chip for circulating tumor cells: from enrichment, release to single cell analysis. trac trends in analytical chemistry 117 (2019) 27-38. https://doi.org/10.1016/j.trac.2019.03.027. [37] k. sakthivel, a. o’brien, k. kim, m. hoorfar. microfluidic analysis of heterotypic cellular interactions: a review of techniques and applications. trac trends in analytical chemistry 117 (2019) 166-185. https://doi.org/10.1016/j.trac.2019.03.026. [38] m.s. hall, j.t. decker, l.d. shea. towards systems tissue engineering: elucidating the dynamics, spatial coordination, and individual cells driving emergent behaviors. biomaterials 255 (2020) 120189. https://doi.org/10.1016/j.biomaterials.2020.120189. [39] d. he, h. li. biomaterials affect cell-cell interactions in vitro in tissue engineering. journal of materials science and technology 63 (2021) 62-72. https://doi.org/10.1016/j.jmst.2020.03.022. [40] m. khazaei, c.s. ahuja, m.g. fehlings. induced pluripotent stem cells for traumatic spinal cord injury. frontiers in cell and developmental biology 4 (2017) 1-9. https://doi.org/10.3389/fcell.2016.00152. [41] j.h. sung, y.i. wang, n. narasimhan sriram, m. jackson, c. long, j.j. hickman, m.l. shuler. recent advances in body-on-a-chip systems. analytical chemistry 91 (2019) 330-351. https://doi.org/10.1021/acs.analchem.8b05293. [42] s.h. lee, b.h. jun. advances in dynamic microphysiological organ-on-a-chip: design principle and its biomedical application. journal of industrial and engineering chemistry 71 (2019) 65-77. https://doi.org/10.1016/j.jiec.2018.11.041. [43] z. jia, y. cheng, x. jiang, c. zhang, g. wang. 3d culture system for liver tissue mimicking hepatic plates for improvement of human hepatocyte ( c3a ) function and polarity. biomed research international (2020). https://doi.org/10.1155/2020/6354183. https://doi.org/10.1002/smll.201503208 https://doi.org/10.1016/j.mee.2018.08.004 https://doi.org/10.1016/j.cobme.2017.02.002 https://doi.org/10.1016/j.cobme.2017.02.002 https://doi.org/10.1242/dev.156125 https://doi.org/10.1016/j.cotox.2019.08.004 https://doi.org/10.1016/j.drudis.2019.03.025 https://doi.org/10.1016/j.coche.2019.10.006 https://doi.org/10.1016/j.trac.2019.03.027 https://doi.org/10.1016/j.trac.2019.03.026 https://doi.org/10.1016/j.biomaterials.2020.120189 https://doi.org/10.1016/j.jmst.2020.03.022 https://doi.org/10.3389/fcell.2016.00152 https://doi.org/10.1016/j.jiec.2018.11.041 https://doi.org/10.1155/2020/6354183 admet & dmpk 9(2) (2021) 111-141 organ-on-a-chip systems for new drugs development doi: http://doi.org/10.5599/admet.942 135 [44] k. gold, a.k. gaharwar, a. jain. emerging trends in multiscale modeling of vascular pathophysiology: organ-on-a-chip and 3d printing. biomaterials 196 (2019) 2-17. https://doi.org/10.1016/j.biomaterials.2018.07.029. [45] j. wang, c. wang, n. xu, z.f. liu, d.w. pang, z.l. zhang. a virus-induced kidney disease model based on organ-on-a-chip: pathogenesis exploration of virus-related renal dysfunctions. biomaterials 219 (2019). https://doi.org/10.1016/j.biomaterials.2019.119367. [46] a. moya, m. ortega-ribera, x. guimerà, e. sowade, m. zea, x. illa, e. ramon, r. villa, j. graciasancho, g. gabriel. online oxygen monitoring using integrated inkjet-printed sensors in a liver-on-achip system. lab on a chip 18 (2018) 2023-2035. https://doi.org/10.1039/c8lc00456k. [47] n. torras, m. garcía-díaz, v. fernández-majada, e. martínez. mimicking epithelial tissues in threedimensional cell culture models. frontiers in bioengineering and biotechnology 6 (2018) 1-7. https://doi.org/10.3389/fbioe.2018.00197. [48] g. adriani, d. ma, a. pavesi, r.d. kamm, e.l.k. goh. a 3d neurovascular microfluidic model consisting of neurons, astrocytes and cerebral endothelial cells as a blood-brain barrier. lab on a chip 17 (2017) 448-459. https://doi.org/10.1039/c6lc00638h. [49] f. yu, w. hunziker, d. choudhury. engineering microfluidic organoid-on-a-chip platforms. micromachines 10 (2019) 1-12. https://doi.org/10.3390/mi10030165. [50] k.a. sun w, luo z, lee j, kim hj, lee k, tebon p, feng y, dokmec, m, sengupta s. organ-on-a-chip for cancer and immune organs modeling. adv healthc mater. 8 (2019) 139-148. doi: https://doi.org/10.1002/adhm.201801363. [51] t.m. marin, n. de carvalho indolfo, s.a. rocco, f.l. basei, m. de carvalho, k. de almeida gonçalves, e. pagani. acetaminophen absorption and metabolism in an intestine/liver microphysiological system. chemico-biological interactions 299 (2019) 59-76. https://doi.org/10.1016/j.cbi.2018.11.010. [52] d.e. ingber. reverse engineering human pathophysiology with organs-on-chips. cell 164 (2016) 1105-1109. https://doi.org/10.1016/j.cell.2016.02.049. [53] j. shi, l. tong, w. tong, h. chen, m. lan, x. sun, y. zhu. current progress in long-term and continuous cell metabolite detection using microfluidics. trac trends in analytical chemistry 117 (2019) 263279. https://doi.org/10.1016/j.trac.2019.05.028. [54] m. peters, a. choy, c. pin, d. leishmann, a. moisan, l.c. ewart, p. guzzie-peck, r. sura, d. keller, c.w. scott, k. kolaja. developing in vitro assays to transform gastrointestinal safety assessment: potential for microphysiological systems. lab on a chip 20 (2020) 1177-1190. https://doi.org/10.1039/c9lc01107b. [55] t. osaki, v. sivathanu, r.d. kamm. vascularized microfluidic organ-chips for drug screening, disease models and tissue engineering. current opinion in biotechnology 52 (2018) 116-123. https://doi.org/10.1016/j.copbio.2018.03.011. [56] d.t.t. phan, r.h.f. bender, j.w. andrejecsk, a. sobrino, s.j. hachey, s.c. george, c.c.w. hughes. blood–brain barrier-on-a-chip: microphysiological systems that capture the complexity of the blood– central nervous system interface. experimental biology and medicine 242 (2017) 1669-1678. https://doi.org/10.1177/1535370217694100. [57] t. grix, a. ruppelt, a. thomas, a.k. amler, b.p. noichl, r. lauster, l. kloke. bioprinting perfusionenabled liver equivalents for advanced organ-on-a-chip applications. genes 9 (2018). https://doi.org/10.3390/genes9040176. [58] f. yu, d. choudhury. microfluidic bioprinting for organ-on-a-chip models. drug discovery today 24 (2019) 1248-1257. https://doi.org/10.1016/j.drudis.2019.03.025. [59] i. maimouni, c.m. cejas, j. cossy, p. tabeling, m. russo. microfluidics mediated production of foams for biomedical applications. micromachines 11 (2020) 1-20. https://doi.org/10.3390/mi11010083. [60] e.s. bishop, s. mostafa, m. pakvasa, h.h. luu, m.j. lee, j.m. wolf, g.a. ameer, t.c. he, r.r. reid. 3-d bioprinting technologies in tissue engineering and regenerative medicine: current and future trends. genes and diseases 4 (2017) 185-195. https://doi.org/10.1016/j.gendis.2017.10.002. http://doi.org/10.5599/admet.942 https://doi.org/10.1016/j.biomaterials.2019.119367 https://doi.org/10.1039/c8lc00456k https://doi.org/10.3389/fbioe.2018.00197 https://doi.org/10.1039/c6lc00638h https://doi.org/10.3390/mi10030165 https://doi.org/10.1016/j.cbi.2018.11.010 https://doi.org/10.1016/j.cell.2016.02.049 https://doi.org/10.1016/j.trac.2019.05.028 https://doi.org/10.1039/c9lc01107b https://doi.org/10.1039/c9lc01107b https://doi.org/10.1177/1535370217694100 https://doi.org/10.3390/genes9040176 https://doi.org/10.1016/j.drudis.2019.03.025 https://doi.org/10.3390/mi11010083 https://doi.org/10.1016/j.gendis.2017.10.002 r. vargas, a. egurbide-sifre and l. medina admet & dmpk 9(2) (2021) 111-141 136 [61] d. huh, b.d. matthews, a. mammoto, m. montoya-zavala, h. yuan hsin, d.e. ingber. reconstituting organ-level lung functions on a chip. science. 328 (2010) 1662-1668. https://doi.org/10.1126/science.1188302. [62] t. takebe, b. zhang, m. radisic. synergistic engineering: organoids meet organs-on-a-chip. cell stem cell 21 (2017) 297-300. https://doi.org/10.1016/j.stem.2017.08.016. [63] p. saglam-metiner, s. gulce-iz, c. biray-avci. bioengineering-inspired three-dimensional culture systems: organoids to create tumor microenvironment. gene 686 (2019) 203-212. https://doi.org/10.1016/j.gene.2018.11.058. [64] j.-p. frimat, r. luttge. the need for physiological micro-nanofluidic systems of the brain. frontiers in bioengineering and biotechnology 7 (2019) 1-11. https://doi.org/10.3389/fbioe.2019.00100. [65] m.l. alec smith, jesse macadangdang, winnie leung, d.-h. kim. human ipsc-derived cardiomyocytes and tissue engineering strategies for disease modeling and drug screening. biotechnol adv. 176 (2017) 139-148. https://doi.org/10.1016/j.biotechadv.2016.12.002. [66] a. skardal, s. v. murphy, m. devarasetty, i. mead, h.w. kang, y.j. seol, y.s. zhang, s.r. shin, l. zhao, j. aleman, a.r. hall, t.d. shupe, a. kleensang, m.r. dokmeci, s. jin lee, j.d. jackson, j.j. yoo, t. hartung, a. khademhosseini, s. soker, c.e. bishop, a. atala. multi-tissue interactions in an integrated three-tissue organ-on-a-chip platform. scientific reports 7 (2017) 1-16. https://doi.org/10.1038/s41598-017-08879-x. [67] c.p. pires de mello, c. carmona-moran, c.w. mcaleer, j. perez, e.a. coln, c.j. long, c. oleaga, a. riu, r. note, s. teissier, j. langer, j.j. hickman. microphysiological heart-liver body-on-a-chip system with a skin mimic for evaluating topical drug delivery. lab on a chip 20 (2020) 749-759. https://doi.org/10.1039/c9lc00861f. [68] m.b. esch, g.j. mahler, t. stokol, m.l. shuler. body-on-a-chip simulation with gastrointestinal tract and liver tissues suggests that ingested nanoparticles have the potential to cause liver injury. lab on a chip 14 (2014) 3081-3092. https://doi.org/10.1039/c4lc00371c. [69] p. de haan, m.a. ianovska, k. mathwig, g.a.a. van lieshout, v. triantis, h. bouwmeester, e. verpoorte. digestion-on-a-chip: a continuous-flow modular microsystem recreating enzymatic digestion in the gastrointestinal tract. lab on a chip 19 (2019) 1599-1609. https://doi.org/10.1039/c8lc01080c. [70] h. chen, z. yu, s. bai, h. lu, d. xu, c. chen, d. liu, y. zhu. microfluidic models of physiological or pathological flow shear stress for cell biology, disease modeling and drug development. trends in analytical chemistry 117 (2019) 186-199. https://doi.org/10.1016/j.trac.2019.06.023. [71] d. kim, x. wu, a.t. young, c.l. haynes. microfluidics-based in vivo mimetic systems for the study of cellular biology. accounts of chemical research 47 (2014) 1165-1173. https://doi.org/10.1021/ar4002608. [72] a. sontheimer-phelps, d.b. chou, a. tovaglieri, t.c. ferrante, t. duckworth, c. fadel, v. frismantas, a.d. sutherland, s. jalili-firoozinezhad, m. kasendra, e. stas, j.c. weaver, c.a. richmond, o. levy, r. prantil-baun, d.t. breault, d.e. ingber. human colon-on-a-chip enables continuous in vitro analysis of colon mucus layer accumulation and physiology. cmgh 9 (2020) 507-526. https://doi.org/10.1016/j.jcmgh.2019.11.008. [73] s. mi, h. pu, s. xia, w. sun. a minimized valveless electromagnetic micropump for microfluidic actuation on organ chips. sensors and actuators, a: physical 301 (2020) 111704. https://doi.org/10.1016/j.sna.2019.111704. [74] g.d. vatine, r. barrile, m.j. workman, s. sances, b.k. barriga, m. rahnama, s. barthakur, m. kasendra, c. lucchesi, j. kerns, n. wen, w.r. spivia, z. chen, j. van eyk, c.n. svendsen. human ipscderived blood-brain barrier chips enable disease modeling and personalized medicine applications. cell stem cell 24 (2019) 995-1005. https://doi.org/10.1016/j.stem.2019.05.011. [75] k. song, g. li, x. zu, z. du, l. liu, z. hu. the fabrication and application mechanism of microfluidic systems for high throughput biomedical screening: a review. micromachines 11 (2020) 297. https://doi.org/10.3390/mi11030297. https://doi.org/10.1016/j.stem.2017.08.016 https://doi.org/10.1016/j.gene.2018.11.058 https://doi.org/10.1016/j.gene.2018.11.058 https://doi.org/10.3389/fbioe.2019.00100 https://doi.org/10.1016/j.biotechadv.2016.12.002 https://doi.org/10.1039/c9lc00861f https://doi.org/10.1039/c9lc00861f https://doi.org/10.1039/c4lc00371c https://doi.org/10.1016/j.trac.2019.06.023 https://doi.org/10.1021/ar4002608 https://doi.org/10.1021/ar4002608 https://doi.org/10.1016/j.sna.2019.111704 https://doi.org/10.1016/j.sna.2019.111704 https://doi.org/10.1016/j.stem.2019.05.011 admet & dmpk 9(2) (2021) 111-141 organ-on-a-chip systems for new drugs development doi: http://doi.org/10.5599/admet.942 137 [76] s. cho, a. islas-robles, a.m. nicolini, t.j. monks, j.y. yoon. in situ, dual-mode monitoring of organ-ona-chip with smartphone-based fluorescence microscope. biosensors and bioelectronics 86 (2016) 697-705. https://doi.org/10.1016/j.bios.2016.07.015. [77] f. sivandzade, l. cucullo. in-vitro blood–brain barrier modeling: a review of modern and fastadvancing technologies. journal of cerebral blood flow and metabolism 38 (2018) 1667-1681. https://doi.org/10.1177/0271678x18788769. [78] b. peng, z. tong, w.y. tong, p.j. pasic, a. oddo, y. dai, m. luo, j. frescene, n.g. welch, c.d. easton, h. thissen, n.h. voelcker. in situ surface modification of microfluidic blood-brain-barriers for improved screening of small molecules and nanoparticles. acs applied materials and interfaces 12 (2020) 56753-56766. https://doi.org/10.1021/acsami.0c17102. [79] l. tong, m. mozneb, e. bravo, v. ferrando, c.z. li. whole cell analysis ranging from intercellular assay to organ on a chip. trends in analytical chemistry 117 (2019) 157-165. https://doi.org/10.1016/j.trac.2019.05.021. [80] m.j.c. santbergen, m. van der zande, h. bouwmeester, m.w.f. nielen. online and in situ analysis of organs-on-a-chip. trends in analytical chemistry 115 (2019) 138-146. https://doi.org/10.1016/j.trac.2019.04.006. [81] n.f. huang, o. chaudhuri, p. cahan, a. wang, a.j. engler, y. wang, s. kumar, a. khademhosseini, s. li. multi-scale cellular engineering: from molecules to organ-on-a-chip. apl bioengineering 4 (2020) 010906. https://doi.org/10.1063/1.5129788. [82] m. abulaiti, y. yalikun, k. murata, a. sato, m.m. sami, y. sasaki, y. fujiwara, k. minatoya, y. shiba, y. tanaka, h. masumoto. establishment of a heart-on-a-chip microdevice based on human ips cells for the evaluation of human heart tissue function. scientific reports 10 (2020) 19201. https://doi.org/10.1038/s41598-020-76062-w. [83] m.c. comes, a. mencattini, d. di giuseppe, j. filippi, m.d. orazio, p. casti, f. corsi, l. ghibelli, c. di natale, e. martinelli. a camera sensors-based system to study drug cancer cells. sensors 20 (2020) 1531-1556. https://doi.org/10.3390/s20051531. [84] n.s. bhise, j. ribas, v. manoharan, y.s. zhang, a. polini, s. massa, m.r. dokmeci, a. khademhosseini. organ-on-a-chip platforms for studying drug delivery systems. journal of controlled release 190 (2014) 82-93. https://doi.org/10.1016/j.jconrel.2014.05.004. [85] m. rimann, e. bono, h. annaheim, m. bleisch, u. graf-hausner. standardized 3d bioprinting of soft tissue models with human primary cells. journal of laboratory automation 21 (2016) 496-509. https://doi.org/10.1177/2211068214567146. [86] c. liu, a. oikonomopoulos, n. sayed, j.c. wu. modeling human diseases with induced pluripotent stem cells: from 2d to 3d and beyond. development (cambridge) 145 (2018) 1-6. https://doi.org/10.1242/dev.156166. [87] c.g. musah s, dimitrakakis n, camacho dm, i. de. directed differentiation of human induced pluripotent stem cells into mature kidney podocytes and establishment of a glomerulus chip. nat protoc 176 (2016) 139-148. https://doi.org/10.1038/s41596-018-0007-8. [88] g. woodruff, s.m. reyna, m. dunlap, r. van der kant, j.a. callender, j.e. young, e.a. roberts, l.s.b. goldstein. defective transcytosis of app and lipoproteins in human ipsc-derived neurons with familial alzheimer’s disease mutations. cell reports 17 (2016) 759-773. https://doi.org/10.1016/j.celrep.2016.09.034. [89] y. wang, h. wang, p. deng, w. chen, y. guo, t. tao, j. qin. in situ differentiation and generation of functional liver organoids from human ipscs in a 3d perfusable chip system. lab on a chip 18 (2018) 3606-3616. https://doi.org/10.1039/c8lc00869h. [90] a. wnorowski, h. yang, j.c. wu. progress, obstacles, and limitations in the use of stem cells in organon-a-chip models. advanced drug delivery reviews 140 (2019) 3-11. https://doi.org/10.1016/j.addr.2018.06.001. [91] r.d. kamm, r. bashir, n. arora, r.d. dar, m.u. gillette, l.g. griffith, m.l. kemp, k. kinlaw, m. levin, a.c. martin, t.c. mcdevitt, r.m. nerem, m.j. powers, t.a. saif, j. sharpe, s. takayama, s. takeuchi, r. http://doi.org/10.5599/admet.942 https://doi.org/10.1016/j.bios.2016.07.015 https://doi.org/10.1021/acsami.0c17102 https://doi.org/10.1063/1.5129788 https://doi.org/10.1038/s41598-020-76062-w https://doi.org/10.1038/s41598-020-76062-w https://doi.org/10.3390/s20051531 https://doi.org/10.1016/j.jconrel.2014.05.004 https://doi.org/10.1242/dev.156166 https://doi.org/10.1242/dev.156166 https://doi.org/10.1038/s41596-018-0007-8 https://doi.org/10.1039/c8lc00869h r. vargas, a. egurbide-sifre and l. medina admet & dmpk 9(2) (2021) 111-141 138 weiss, k. ye, h.g. yevick, m.h. zaman. perspective: the promise of multi-cellular engineered living systems. apl bioengineering 2 (2018) 040901. https://doi.org/10.1063/1.5038337. [92] c.p.p. de mello, j. rumsey, v. slaughter, j.j. hickman. a human-on-a-chip approach to tackling rare diseases. drug discovery today 24 (2019) 2139-2151. https://doi.org/10.1016/j.drudis.2019.08.001. [93] m. haddrick, p.b. simpson. organ-on-a-chip technology: turning its potential for clinical benefit into reality. drug discovery today 24 (2019) 1217-1223. https://doi.org/10.1016/j.drudis.2019.03.011. [94] j.d. caplin, n.g. granados, m.r. james, r. montazami, n. hashemi. microfluidic organ-on-a-chip technology for advancement of drug development and toxicology. advanced healthcare materials 4 (2015) 1426-1450. https://doi.org/10.1002/adhm.201500040. [95] a.k. schroer, m.s. shotwell, v.y. sidorov, j.p. wikswo, w.d. merryman. i-wire heart-on-a-chip ii: biomechanical analysis of contractile, three-dimensional cardiomyocyte tissue constructs. acta biomaterialia 48 (2017) 79-87. https://doi.org/10.1016/j.actbio.2016.11.010. [96] p. loskill, a. mathur, b.r. conklin, a. stahl, l.p. lee, k.e. healy. organs-on-a-chip – microphysiological platforms as in vitro models of cardiac and adipose tissue. toxicology letters 258 (2016) s153. https://doi.org/10.1016/j.toxlet.2016.06.1586. [97] l. yin, g. du, b. zhang, h. zhang, r. yin, w. zhang, s.m. yang. efficient drug screening and nephrotoxicity assessment on co-culture microfluidic kidney chip. scientific reports 10 (2020) 1-11. https://doi.org/10.1038/s41598-020-63096-3. [98] k.-j. jang, a.p. mehr, g.a. hamilton, l.a. mcpartlin, s. chung, k.-y. suh, d.e. ingber. human kidney proximal tubule-on-a-chip for drug transport and nephrotoxicity assessment. integrative biology 5 (2013) 1119-1129. https://doi.org/10.1039/c3ib40049b. [99] a.o. stucki, j.d. stucki, s.r.r. hall, m. felder, y. mermoud, r.a. schmid, t. geiser, o.t. guenat. a lung-on-a-chip array with an integrated bio-inspired respiration mechanism. lab on a chip 15 (2015) 1302-1310. https://doi.org/10.1039/c4lc01252f. [100] m. felder, b. trueeb, a.o. stucki, s. borcard, j.d. stucki, b. schnyder, t. geiser, o.t. guenat. impaired wound healing of alveolar lung epithelial cells in a breathing lung-on-a-chip. frontiers in bioengineering and biotechnology 7 (2019) 3. https://doi.org/10.3389/fbioe.2019.00003. [101] m.a.u. khalid, y.s. kim, m. ali, b.g. lee, y.j. cho, k.h. choi. a lung cancer-on-chip platform with integrated biosensors for physiological monitoring and toxicity assessment. biochemical engineering journal 155 (2020) 107469. https://doi.org/10.1016/j.bej.2019.107469. [102] j. rogal, a. zbinden, k. schenke-layland, p. loskill. stem-cell based organ-on-a-chip models for diabetes research. advanced drug delivery reviews 140 (2019) 101-128. https://doi.org/10.1016/j.addr.2018.10.010. [103] a. bein, w. shin, s. jalili-firoozinezhad, m.h. park, a. sontheimer-phelps, a. tovaglieri, a. chalkiadaki, h.j. kim, d.e. ingber. microfluidic organ-on-a-chip models of human intestine. cmgh 5 (2018) 659668. https://doi.org/10.1016/j.jcmgh.2017.12.010. [104] c.a. richmond, d.t. breault. move over caco-2 cells: human-induced organoids meet gut-on-achip. cmgh 5 (2018) 634-635. https://doi.org/10.1016/j.jcmgh.2018.01.016. [105] k.m. seiler, a. bajinting, d.m. alvarado, m.a. traore, m.m. binkley, w.h. goo, w.e. lanik, j. ou, u. ismail, m. iticovici, c.r. king, k.l. vandussen, e.a. swietlicki, v. gazit, j. guo, c.j. luke, t. stappenbeck, m.a. ciorba, s.c. george, j.m. meacham, d.c. rubin, m. good, b.w. warner. patientderived small intestinal myofibroblasts direct perfused, physiologically responsive capillary development in a microfluidic gut-on-a-chip model. scientific reports 10 (2020). https://doi.org/10.1038/s41598-020-60672-5. [106] m. kasendra, a. tovaglieri, a. sontheimer-phelps, s. jalili-firoozinezhad, a. bein, a. chalkiadaki, w. scholl, c. zhang, h. rickner, c.a. richmond, h. li, d.t. breault, d.e. ingber. development of a primary human small intestine-on-a-chip using biopsy-derived organoids. scientific reports 8 (2018) 2871. https://doi.org/10.1038/s41598-018-21201-7. https://doi.org/10.1063/1.5038337 https://doi.org/10.1016/j.drudis.2019.03.011 https://doi.org/10.1002/adhm.201500040 https://doi.org/10.1016/j.actbio.2016.11.010 https://doi.org/10.1016/j.toxlet.2016.06.1586 https://doi.org/10.1039/c3ib40049b https://doi.org/10.1039/c4lc01252f https://doi.org/10.3389/fbioe.2019.00003 https://doi.org/10.1016/j.bej.2019.107469 https://doi.org/10.1016/j.addr.2018.10.010 https://doi.org/10.1016/j.addr.2018.10.010 https://doi.org/10.1016/j.jcmgh.2017.12.010 https://doi.org/10.1016/j.jcmgh.2018.01.016 https://doi.org/10.1038/s41598-018-21201-7 admet & dmpk 9(2) (2021) 111-141 organ-on-a-chip systems for new drugs development doi: http://doi.org/10.5599/admet.942 139 [107] n. freyer, s. greuel, f. knöspel, f. gerstmann, l. storch, g. damm, d. seehofer, j.f. harris, r. iyer, f. schubert, k. zeilinger. microscale 3d liver bioreactor for in vitro hepatotoxicity testing under perfusion conditions. bioengineering 5 (2018). https://doi.org/10.3390/bioengineering5010024. [108] t. kostrzewski, t. cornforth, s.a. snow, l. ouro-gnao, c. rowe, e.m. large, d.j. hughes. threedimensional perfused human in vitro model of non-alcoholic fatty liver disease. world journal of gastroenterology 23 (2017) 204-215. https://doi.org/10.3748/wjg.v23.i2.204. [109] a. bhalerao, f. sivandzade, s.r. archie, e.a. chowdhury, b. noorani, l. cucullo. in vitro modeling of the neurovascular unit: advances in the field. fluids and barriers of the cns 17 (2020) 22. https://doi.org/10.1186/s12987-020-00183-7. [110] f. yin, y. zhu, m. zhang, h. yu, w. chen, j. qin. a 3d human placenta-on-a-chip model to probe nanoparticle exposure at the placental barrier. toxicology in vitro 54 (2019) 105-113. https://doi.org/10.1016/j.tiv.2018.08.014. [111] e. knyazev, a. khristichenko, d. maltseva, t. gerasimenko, o. kindeeva, v. petrov, d. sakharov. placenta-on-a-chip model for assessing the transport and toxicity of xenobiotics in vitro. placenta 83 (2019) e59-e60. https://doi.org/10.1016/j.placenta.2019.06.195. [112] n. manafi, f. shokri, k. achberger, m. hirayama, m.h. mohammadi, f. noorizadeh, j. hong, s. liebau, t. tsuji, p.m.j. quinn, a. mashaghi. organoids and organ chips in ophthalmology. the ocular surface 19 (2021) 1-15. https://doi.org/10.1016/j.jtos.2020.11.004. [113] d.s. park, j.h. kim, h.s. kim, j.h. park, j.k. shin, m. lee. a foveated-structure cmos retina chip for edge detection with local light adaptation. en: sensors actuators, a phys., elsevier, 2003: pp. 75-80. https://doi.org/10.1016/s0924-4247(03)00292-9. [114] a. grosberg, a.p. nesmith, j.a. goss, m.d. brigham, m.l. mccain, k.k. parker. muscle on a chip: in vitro contractility assays for smooth and striated muscle. journal of pharmacological and toxicological methods 65 (2012) 126-135. https://doi.org/10.1016/j.vascn.2012.04.001. [115] b.p. jena, d.l. gatti, s. arslanturk, s. pernal, d.j. taatjes. human skeletal muscle cell atlas: unraveling cellular secrets utilizing ‘muscle-on-a-chip’, differential expansion microscopy, mass spectrometry, nanothermometry and machine learning. micron 117 (2019) 55-59. https://doi.org/10.1016/j.micron.2018.11.002. [116] q. tang, x. li, c. lai, l. li, h. wu, y. wang, x. shi. fabrication of a hydroxyapatite-pdms microfluidic chip for bone-related cell culture and drug screening. bioactive materials 6 (2021) 169-178. https://doi.org/10.1016/j.bioactmat.2020.07.016. [117] e.l. george, s.l. truesdell, s.l. york, m.m. saunders. lab-on-a-chip platforms for quantification of multicellular interactions in bone remodeling. experimental cell research 365 (2018) 106-118. https://doi.org/10.1016/j.yexcr.2018.02.027. [118] y. chen, d. gao, y. wang, s. lin, y. jiang. a novel 3d breast-cancer-on-chip platform for therapeutic evaluation of drug delivery systems. analytica chimica acta 1036 (2018) 97-106. https://doi.org/10.1016/j.aca.2018.06.038. [119] x. zhang, x. lu, w. gao, y. wang, c. jia, h. cong. a label-free microfluidic chip for the highly selective isolation of single and cluster ctcs from breast cancer patients. translational oncology 14 (2021) 100959. https://doi.org/10.1016/j.tranon.2020.100959. [120] k.s. mun, m. abu-el-haija, j.d. nathan, r.a. liddle, a.p. naren. mo1373 human pancreas-on-achip to study pressure-induced pancreatitis. gastroenterology 158 (2020) s-868. https://doi.org/10.1016/s0016-5085(20)32850-x. [121] a. essaouiba, t. okitsu, r. kinoshita, r. jellali, m. shinohara, m. danoy, c. legallais, y. sakai, e. leclerc. development of a pancreas-liver organ-on-chip coculture model for organ-to-organ interaction studies. biochemical engineering journal 164 (2020) 107783. https://doi.org/10.1016/j.bej.2020.107783. [122] j. radhakrishnan, s. varadaraj, s.k. dash, a. sharma, r.s. verma. organotypic cancer tissue models for drug screening: 3d constructs, bioprinting and microfluidic chips. drug discovery today 00 (2020). https://doi.org/10.1016/j.drudis.2020.03.002. http://doi.org/10.5599/admet.942 https://doi.org/10.3390/bioengineering5010024 https://doi.org/10.3748/wjg.v23.i2.204 https://doi.org/10.1016/j.tiv.2018.08.014 https://doi.org/10.1016/j.placenta.2019.06.195 https://doi.org/10.1016/j.jtos.2020.11.004 https://doi.org/10.1016/j.vascn.2012.04.001 https://doi.org/10.1016/j.bioactmat.2020.07.016 https://doi.org/10.1016/j.aca.2018.06.038 https://doi.org/10.1016/j.aca.2018.06.038 https://doi.org/10.1016/j.tranon.2020.100959 https://doi.org/10.1016/j.bej.2020.107783 https://doi.org/10.1016/j.bej.2020.107783 https://doi.org/10.1016/j.drudis.2020.03.002 r. vargas, a. egurbide-sifre and l. medina admet & dmpk 9(2) (2021) 111-141 140 [123] k.-h. tu, l.-s. yu, z.-h. sie, h.-y. hsu, k.t. al-jamal, j.t.-w. wang, y.-y. chiang. development of realtime transendothelial electrical resistance monitoring for an in vitro blood-brain barrier system. micromachines 12 (2020) 37. https://doi.org/10.3390/mi12010037. [124] b.a. hassell, g. goyal, e. lee, a. sontheimer-phelps, o. levy, c.s. chen, d.e. ingber. human organ chip models recapitulate orthotopic lung cancer growth, therapeutic responses, and tumor dormancy in vitro. cell reports 21 (2017) 508-516. https://doi.org/10.1016/j.celrep.2017.09.043. [125] d.h.t. nguyen, e. lee, s. alimperti, r.j. norgard, a. wong, j.j.k. lee, j. eyckmans, b.z. stanger, c.s. chen. a biomimetic pancreatic cancer on-chip reveals endothelial ablation via alk7 signaling. science advances 5 (2019) 6789-6817. https://doi.org/10.1126/sciadv.aav6789. [126] s. shuford, c. wilhelm, m. rayner, a. elrod, m. millard, c. mattingly, a. lotstein, a.m. smith, q.j. guo, l. o’donnell, j. elder, l. puls, s.j. weroha, x. hou, v. zanfagnin, a. nick, m.p. stany, g.l. maxwell, t. conrads, a.k. sood, d. orr, l.m. holmes, m. gevaert, h.e. crosswell, t.m. desrochers. prospective validation of an ex vivo, patient-derived 3d spheroid model for response predictions in newly diagnosed ovarian cancer. scientific reports 9 (2019) 1-13. https://doi.org/10.1038/s41598019-47578-7. [127] y. choi, e. hyun, j. seo, c. blundell, h.c. kim, e. lee, s.h. lee, a. moon, w.k. moon, d. huh. a microengineered pathophysiological model of early-stage breast cancer. lab on a chip 15 (2015) 3350-3357. https://doi.org/10.1039/c5lc00514k. [128] d. irimia, x. wang. inflammation-on-a-chip: probing the immune system ex vivo. trends in biotechnology 36 (2018) 923-937. https://doi.org/10.1016/j.tibtech.2018.03.011. [129] f. sharifi, s.s. htwe, m. righi, h. liu, a. pietralunga, o. yesil-celiktas, s. maharjan, b.-h. cha, s.r. shin, m.r. dokmeci, n.e. vrana, a.m. ghaemmaghami, a. khademhosseini, y.s. zhang. a foreign body response-on-a-chip platform. advanced healthcare materials 8 (2019) 1801425. https://doi.org/10.1002/adhm.201801425. [130] g. goyal, b. bausk, p. prabhala, l. xie, d. curran, j. long, l. cohen, o. levy, r. prantil-baun, d. walt, d. ingber. lymph node follicle formation and vaccination responses reconstituted in vitro in a human organ chip. biorxiv (2019) 806505. https://doi.org/10.1101/806505. [131] n. gopalakrishnan, r. hannam, g.p. casoni, d. barriet, j.m. ribe, m. haug, halaas. infection and immunity on a chip: a compartmentalised microfluidic platform to monitor immune cell behaviour in real time. lab on a chip 15 (2015) 1481-1487. https://doi.org/10.1039/c4lc01438c. [132] s. parlato, g. grisanti, g. sinibaldi, g. peruzzi, c.m. casciola, l. gabriele. tumor-on-a-chip platforms to study cancer-immune system crosstalk in the era of immunotherapy. lab on a chip 21 (2021) 234253. https://doi.org/10.1039/d0lc00799d. [133] s. ishida. organs-on-a-chip: current applications and consideration points for in vitro adme-tox studies. drug metabolism and pharmacokinetics 33 (2018) 49-54. https://doi.org/10.1016/j.dmpk.2018.01.003. [134] i. maschmeyer, a.k. lorenz, k. schimek, t. hasenberg, a.p. ramme, j. hübner, m. lindner, c. drewell, s. bauer, a. thomas, n.s. sambo, f. sonntag, r. lauster, u. marx. a four-organ-chip for interconnected long-term co-culture of human intestine, liver, skin and kidney equivalents. lab on a chip 15 (2015) 2688-2699. https://doi.org/10.1039/c5lc00392j. [135] a. herland, b.m. maoz, d. das, m.r. somayaji, r. prantil-baun, r. novak, m. cronce, t. huffstater, s.s.f. jeanty, m. ingram, a. chalkiadaki, d. benson chou, s. marquez, a. delahanty, s. jalilifiroozinezhad, y. milton, a. sontheimer-phelps, b. swenor, o. levy, k.k. parker, a. przekwas, d.e. ingber. quantitative prediction of human pharmacokinetic responses to drugs via fluidically coupled vascularized organ chips. nature biomedical engineering 4 (2020) 421-436. https://doi.org/10.1038/s41551-019-0498-9. [136] b. zhang, m. radisic. organ-on-a-chip devices advance to market. lab on a chip 17 (2017) 23952420. https://doi.org/10.1039/c6lc01554a. [137] q. ramadan, m. zourob. organ-on-a-chip engineering: toward bridging the gap between lab and industry. biomicrofluidics 14 (2020). https://doi.org/10.1063/5.0011583. https://doi.org/10.3390/mi12010037 https://doi.org/10.1016/j.celrep.2017.09.043 https://doi.org/10.1126/sciadv.aav6789 https://doi.org/10.1038/s41598-019-47578-7 https://doi.org/10.1038/s41598-019-47578-7 https://doi.org/10.1039/c5lc00514k https://doi.org/10.1016/j.tibtech.2018.03.011 https://doi.org/10.1101/806505 https://doi.org/10.1039/c4lc01438c https://doi.org/10.1039/d0lc00799d https://doi.org/10.1039/c5lc00392j https://doi.org/10.1038/s41551-019-0498-9 https://doi.org/10.1038/s41551-019-0498-9 https://doi.org/10.1039/c6lc01554a https://doi.org/10.1063/5.0011583 admet & dmpk 9(2) (2021) 111-141 organ-on-a-chip systems for new drugs development doi: http://doi.org/10.5599/admet.942 141 [138] e.m. dehne, h. erfurth, a.k. muhsmann, u. marx. automation and opportunities for industry scale-up of microphysiological systems. en: organ-on-a-chip eng. microenviron. saf. effic. test., elsevier, 2019: pp. 441-462. https://doi.org/10.1016/b978-0-12-817202-5.00014-0. [139] a.r. perestrelo, a.c.p. águas, a. rainer, g. forte. microfluidic organ/body-on-a-chip devices at the convergence of biology and microengineering. sensors (switzerland) 15 (2015) 31142-31170. https://doi.org/10.3390/s151229848. [140] j. zhang, x. wei, r. zeng, f. xu, x. li. stem cell culture and differentiation in microfluidic devices toward organ-on-a-chip. future science oa 3 (2017) fso187. https://doi.org/10.4155/fsoa-20160091. [141] i.r. mccracken, r.s. taylor, f.o. kok, f. de la cuesta, r. dobie, b.e.p. henderson, j.c. mountford, a. caudrillier, n.c. henderson, c.p. ponting, a.h. baker. transcriptional dynamics of pluripotent stemcell-derived endothelial cell differentiation revealed by single-cell rna sequencing. european heart journal 41 (2020) 1024-1036. https://doi.org/10.1093/eurheartj/ehz351. [142] s.w. tang, w.y. tong, s.w. pang, n.h. voelcker, y.w. lam. deconstructing, replicating, and engineering tissue microenvironment for stem cell differentiation. tissue engineering part b: reviews 00 (2020). https://doi.org/10.1089/ten.teb.2020.0044. [143] q. wu, j. liu, x. wang, l. feng, j. wu, x. zhu, w. wen, x. gong. organ-on-a-chip: recent breakthroughs and future prospects. biomedical engineering online 19 (2020) 1-19. https://doi.org/10.1186/s12938-020-0752-0. [144] h. hirama, t. satoh, s. sugiura, k. shin, r. onuki-nagasaki, t. kanamori, t. inoue. glass-based organon-a-chip device for restricting small molecular absorption. journal of bioscience and bioengineering 127 (2019) 641-646. https://doi.org/10.1016/j.jbiosc.2018.10.019. [145] a. spanu, l. martines, a. bonfiglio. interfacing cells with organic transistors: a review of in vitro and in vivo applications. lab on a chip (2021). https://doi.org/10.1039/d0lc01007c. [146] y. liang, j.-y. yoon. in situ sensors for blood-brain barrier (bbb) on a chip. sensors and actuators reports (2021) 100031. https://doi.org/10.1016/j.snr.2021.100031. ©2021 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license(http://creativecommons.org/licenses/by/3.0/) http://doi.org/10.5599/admet.942 https://doi.org/10.1016/b978-0-12-817202-5.00014-0 https://doi.org/10.4155/fsoa-2016-0091 https://doi.org/10.4155/fsoa-2016-0091 https://doi.org/10.1093/eurheartj/ehz351 https://doi.org/10.1089/ten.teb.2020.0044 https://doi.org/10.1039/d0lc01007c https://doi.org/10.1016/j.snr.2021.100031 http://creativecommons.org/licenses/by/3.0/ synthesis and biological activity of 2-[6-methyl-4-(thietan-3-yloxy)pyrimidin-2-ylthio]acetohydrazide derivatives doi: http://dx.doi.org/10.5599/admet.941 167 admet & dmpk 9(2) (2021) 167-176; doi: https://doi.org/10.5599/admet.941 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index original scientific paper synthesis and biological activity of 2-[6-methyl-4-(thietan-3yloxy)pyrimidin-2-ylthio]acetohydrazide derivatives svetlana meshcheryakova, alina shumadalova, ozal beylerli*, ilgiz gareev* bashkir state medical university, ufa, russia *corresponding authors: ilgiz gareev e-mail: ilgiz_gareev@mail.ru; tel.: +79374952927; ozal beylerli e-mail: obeylerli@mail.ru; tel. +79875980003 received: december 2020, 2021; revised: february 09, 2021; published online: february 18, 2021 abstract the synthesis and antimicrobial evaluation of new 2-[6-methyl-4-(thietan-3-yloxy)pyrimidin-2ylthio]acetohydrazide derivatives was investigated. according to the literature, there are a lot of antimicrobial agents among the pyrimidines and hydrazides, and therefore it seems promising to use 2-[6methyl-4-(thietan-3-yloxy)pyrimidin-2-ylthio]acetohydrazide as a base object for synthesizing new biologically active substances. 2-[6-methyl-4-(thietan-3-yloxy)pyrimidin-2-ylthio]acetohydrazide was obtained by the hydrazinolysis of ethyl thioacetate, using a 3-fold molar excess of 85 % hydrazine hydrate in ethanol, at room temperature. interaction of 2-[6-methyl-4-(thietan-3-yloxy)pyrimidin-2-ylthio]acetohydrazide with ketones during boiling in ethanol yielded n-ylidenehydrazides. the solid obtained by concentration was collected, and then purified by recrystallization. the new compounds were characterized by 1 h, 13 c nmr, ir spectroscopy and elemental analysis. the antibacterial and antifungal activities of the new compounds were analysed using agar diffusion and tenfold broth (ph 7.2 – 7.4) dilution methods, in comparison with the clinical used drugs, ceftriaxone and pimafucin. the structure– activity studies showed that, depending on the nature of the hydrazide fragment, the newly synthesized compounds exhibited varying degrees of microbial inhibition. within the same series the antimicrobial activity depends on the nature of the substituent attached to the benzene ring. the investigation of antibacterial screening data revealed that the compounds nʹ-[1-(4-aminophenyl)ethylidene]-2-[6-methyl4-(thietan-3-yloxy)pyrimidin-2-ylthio]acetohydrazide, nʹ-[1-(4-hydroxyphenyl)ethylidene]-2-[6-methyl-4(thietan-3-yloxy)pyrimidin-2-ylthio]acetohydrazide, nʹ-[1(2,5-dihydroxyphenyl) ethylidene]-2-[6-methyl4-(thietan-3-yloxy)-pyrimidin-2-ylthio]acetohydrazide were found to be more potent than the other synthesized analogues. ©2021 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords thietan; thiopyrimidine; antibacterial and antifungal activities introduction infectious diseases are still a challenging global problem and a major public health threat, and this has led to the research and development of new antibiotics for multi-drug resistant microbial pathogens. fungal and bacterial agents affect internal organs like the mucous membrane of the respiratory, gastrointestinal and urinary tracts, and cause different diseases [1,2]. increasingly, there is a growing number of individuals with a weakened immune system who can get infected. http://dx.doi.org/10.5599/admet.941 https://doi.org/10.5599/admet.941 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:ilgiz_gareev@mail.ru mailto:obeylerli@mail.ru http://creativecommons.org/licenses/by/4.0/ i. gareev et al. admet & dmpk 9(2) (2021) 167-176 168 over the last decade there has been considerable interest in the synthesis and pharmacological study of pyrimidine derivatives [3-7]. the pyrimidine fragment has been gaining special interest as it is part of many biological active compounds: vitamins, nucleotides and nucleosides. our research group has published much research in the field of the synthesis of thietan derivatives [8,9] as they possess a broad spectrum of bioactivities: antimicrobial [10,11], hypotensive [12,13], broncholytic [14] and wound healing [15,16]. compounds containing thietan and pyrimidine in their molecules are suitable candidates for further chemical modifications and may be pharmacologically active. we report the synthesis of new derivatives of 2-thiopyrimidine containing thietan ring, and investigate their antibacterial and antifungal properties. experimental materials and methods the reagents used in the experiments were all commercially available without further purification. each reaction was monitored by thin layer chromatography (tlc) on sorbfil ptsx-af-a-uv plates using ethyl acetate. visualization on tlc was achieved by uv light or iodine indicator. melting points were recorded on ptp-m (russia) melting point apparatus uncorrected. infrared spectra were recorded in kbr pellets on an infralum ft-02 (russia) spectrophotometer. 1 h and 13 c nmr spectra were recorded on bruker amx-300 300 mhz and bruker avance iii 500 mhz in cdcl3 and dmso-d6 using tetramethylsilane as internal standard (chemical shifts were expressed as δ-values, j in hertz). chns elemental analyses were performed on a hekatech euro-ea (germany) elemental analyzer. ethyl 2-[6-methyl-4-(tietan-3-yloxy)pyrimidine-2-ylthio]acetate (1) was synthesized by interaction of ethyl-2-(6-methyl-4-oxo-3,5-dihydropyrimidine-2-ilthio)acetate with 2-chloromethylthiirane [16] (figure 1). ethyl 2-[6-methyl-4-(tietan-3-yloxy)pyrimidine-2-ylthio]acetate (compound 1) light yellow crystalline powder. recrystallization from mixture etoh-water (3:1). yield: 42 %; m.p. 61 °c. ir (kbr) in cm −1 : 1630 (νс=о), 3280 (νn-h), 1530 (δn-h); 1572-1591 (с=с, с=n), 2960 (νc-h), 1443 (δc=c 6 ch3), 1167, 1050 (c-o-c), 1750 (νch2-coor). 1 h nmr (300 mhz, cdcl3, δ, ppm): 1.31 t (3н, сн3, et, j 7.1 hz); 2.34 s (3h, 6-ch3); 3.38-3.48 m [2н, s(ch)2]; 3.48-3.58 m [2н, s(ch)2]; 3.82 s (2н, sch2co); 4.22 q (2н, оch2, et, j 4,1 hz); 5.55-5.57 m (1н, och); 6.21 s (1h, 5-h). 13 c nmr (125.5 mhz, cdcl3, δ, ppm): 168.52 (c 6 ); 23.11 ( 6 c-ch3); 101.91 ( 5 ch); 166.95 ( 4 c); 69.63 (oc 3 ); 34.50 (sc 2,4 ); 169.17 ( 2 c); 32.51 (2-sch2co); 166.79 (2-sch2co), 60,61 (ch2ch3), 14,15 (ch2ch3). found, %: c 47.96; h 5.39; n 9.43; s 21.35. c12h16n2o3s2. calculated, %: c 47.98; h 5.37; n 9.33; s 21.35. (e,z)-2-[6-methyl-4-(thietan-3-yloxy)pyrimidin-2-ylthio]acetohydrazide (compound 2) light yellow crystalline powder. 85% hydrazine hydrate (0.29 g, 0.00501 mol) solution was added to a solution of compound 1 (0.5 g, 0.00167 mol) in 3 ml of ethanol. the reaction mixture was mixed at room temperature for 4 hours. the product which precipitated was filtered, dried and recrystallized from i-proh. yield: 0.32 g (67%); m.p. 119-120°c. ir (kbr) in cm −1 : 1630 (νс=о), 3280 (νn-h), 1520 (δn-h); 1572-1591 (с=с, с=n), 2960 (νc-h), 1443 (δc=c 6 -ch3), 1167, 1050 (c-o-c). 1 h nmr (300 mhz, c6d6, δ, ppm): 2.30 (3н, s, 6 c–ch3,), 3.42-3.49, (4h, m, s(ch2)2), 5.76-5.82 (1h, m, och), 6,50 (z), 6,48 (e) (1h, s, 5 c–h), 3,78 (z), 4,18 (e) s (2н, sch2со); 4,30 (z), 4,54 (e) wid. s (2н, nн2); 8,61 (e), 9,31 (z) wid. s (1н, nh). 13 c nmr (125.5 mhz, cdcl3, δ, ppm): 168.52 (c 6 ); 23.11 ( 6 c-ch3); 101.91 ( 5 ch); 166.95 ( 4 c); 69.63 (oc 3 ); 34.50 (sc 2,4 ); 169.17 ( 2 c); 32.51 (2-sch2co); 166.79 (2-sch2co). found,%: c 41.95; h 4.93; n 19.57; s 22.39. c10h14n4o2s2. calculated,%: c 41.94; h 4.93; n 19.57; s 22.39. admet & dmpk 9(2) (2021) 167-176 pyrimidine derivatives and infection doi: http://dx.doi.org/10.5599/admet.941 169 figure 1. synthetic route of the synthesized compounds 1-12 general procedure for the synthesis of compound 3-12 appropriate ketone was added to a solution of compound 2 in 5 ml of ethanol (figure 1). the reaction mixture was boiled for 1 h. the solid obtained by concentration was collected and purified by recrystallization. (e,z)-2-[6-methyl-4-(thietan-3-yloxy)pyrimidin-2-ylthio]-nʹ-(propan-2-ylidene)acetohydrazide (compound 3) white solid. recrystallization from etoh. yield: 0.21 g (61%); m.p. 141-142 °c. ir (kbr) in cm −1 : 1670 («amide i»), 1650 (c=o), 1588, 1572 (с=n, с=с), 1546 (δ nh, «amide ii»), 1393 (δс-н), 1164 (nas), 1045 (ns) (c-o-c). 1 h nmr (300 mhz, c6d6, δ, ppm): 2.10 (3н, s, 6 c–ch3,), 3.12-3.19, 3.31-3.37 (4h, m, s(ch2)2), 5.625.63 (1h, m, och), 5.82 (1h, s, 5 c–h), 4.44 (1h, s, sch2co), 9.94 (1h, wid s, conh), 1.47 s (3н, =с(сн3)); 1.65 s (3н, =с(сн3)). 1 h nmr (300 mhz, dmso-d6, δ, ppm): 2.29 (e), 2.50 (z) (3н, s, 6 c–ch3,), 3.35-3.49 (z, e) (4h, m, s(ch2)2), 5.68-5.77 (e, z) (1h, m, och), 6.45 (e), 6.48 (z) (1h, s, 5 c–h), 3.92 (z), 4.26 (e) (1h, s, sch2co), 10.29 (z), 10.39 (e) (1h, wid s, conh), 1.90 (e), 1.91 (z) s (3н, =с(сн3)); 1.92 (z), 1.94 (e) s (3н, =с(сн3)). 1 h nmr (300 mhz, cdcl3, δ, ppm): 2.35 (z), 2.41 (e) (3н, s, 6 c–ch3,), 3.33-3.37, 3.41-3.45(e), 3.493.54 (z) (4h, m, s(ch2)2), 5.78-5.86 (e, z) (1h, m, och), 6.19 (e), 6.30 (z) (1h, s, 5 c–h), 3.87 (z), 4.29 (e) (1h, s, sch2co), 8.56 (z), 9.57 (e) (1h, wid s, conh), 1.78 (e), 1.91 (z) s (3н, =с(сн3)); 2.05 (z), 2.07 (e) s (3н, =с(сн3)). 13 c nmr (125,5 mhz, cdcl3, δ, ppm): 165.07 (e, z) (c 6 ); 25.41 (e); 25.48 (z) ( 6 c-ch3); 102.13 (z); 103.30 (e) ( 5 ch); 167.45 (z); 167.65(e) ( 4 c); 69.51 (z); 70.29 (e) (oc 3 ); 35.10 (e); 35.56 (z) (sc 2,4 ); 168.30 (e); 168.49 (z) ( 2 c); 33.35 (z); 33.58 (e) (2-sch2co); 169.44 (z); 170.02 (e) (2-sch2co); 16.10 (z), 16.91 (e) (n=c(ch3)); 23.72 (z), 23.93 (e) (n=c(ch3)); 150.21 (z), 155.49 (e) (n=c). found,%: c 47.83; h 5.56; n, 17.16; s 19.64. c13h18n4o2s2. calculated,%: c 47.83; h 5.56; n, 17.16; s 19.64. (e,z)-nʹ-(butan-2-ylidene)-2-[6-methyl-4-(thietan-3-yloxy)pyrimidin-2-yl-thio]acetohydrazide (compound 4) white solid. recrystallization from etoh. yield: 0.29 g (81%); m.p. 131-132 °c. ir (kbr) in cm −1 : 1675 («amide i»), 1645 (c=o), 1586, 1571 (с=n, с=с), 1546 (δ nh, «amide ii»), 1393 (δс-н), 1168 (nas), 1050 (ns) (c-o-c). 1 h nmr (300 mhz, cdcl3, δ, ppm): 2.07 (z), 2.12 (e) (3н, s, 6 c–ch3,), 3.42-3.45(e), 3.49-3.53 (z) (4h, m, s(ch2)2), 5.73-5.86 (e, z) (1h, m, och), 5.95 (e), 6.11 (z) (1h, s, 5 c–h), 4.01 (z), 4.29 (e) (1h, s, sch2co), 9.57 (z), 10.57 (e) (1h, wid s, conh), 2.05 (e), 2.17 (z) m (2н, =ссн2); 1.92 (z), 1.99 (e) s (3н, =ссн3), 0.83 http://dx.doi.org/10.5599/admet.941 i. gareev et al. admet & dmpk 9(2) (2021) 167-176 170 (z), 1.08 (e) t (3н, =ссн2ch3). 13 c nmr (125.5 mhz, cdcl3, δ, ppm): 166.02 (e, z) (c 6 ); 23.74 (e); 23.88 (z) ( 6 c-ch3); 101.43 (z); 102.30 (e) ( 5 ch); 168.45 (z); 168.65(e) ( 4 c); 67.41 (z); 68.39 (e) (oc 3 ); 34.10 (e); 34.76 (z) (sc 2,4 ); 169.40 (e); 169.48 (z) ( 2 c); 35.36 (z); 35.78 (e) (2-sch2co); 170.42 (z); 171.01 (e) (2-sch2co); 13.20 (z), 13.81 (e) (n=c(ch3)); 11.12 (z), 12.11 (e) (ch3); 26.22 (z), 26.92 (e) (n=c(ch2)); 158.11 (z), 158.39 (e) (n=c). found,%: c 49.39; h 5.92; n, 16.46; s 18.83. c14h20n4o2s2. calculated,%: c 49.39; h 5.92; n, 16.46; s 18.83. (e,z)-nʹ-(cyclohexylidene)-2-[6-methyl-4-(thietan-3-yloxy)pyrimidin-2-yl-thio]acetohydrazide (compound 5) white solid. recrystallization from n-buoh. yield: 0.34 g (89%); m.p. 180-181 °c. ir (kbr) in cm −1 : 1661 («amide i»), 1580, 1566 (с=n, с=с), 1546 (δ nh, «amide ii»), 1449-1432 (n ch2 cyclohexane), 1164 (nas), 1047 (ns) (c-o-c), 768 (c-s). 1 h nmr (300 mhz, cdcl3, δ, ppm): 2.06 (z), 2.12 (e) (3н, s, 6 c–ch3,), 3.41-3.48(e), 3.37-3.45 (z) (4h, m, s(ch2)2), 5.73-5.86 (e, z) (1h, m, och), 5.95 (e), 6.11 (z) (1h, s, 5 c–h), 4.11 (z), 4.21 (e) (1h, s, sch2co), 9.56 (z), 10.51 (e) (1h, wid s, conh), 2.34 t (4н, =с(сн2)2); 1.65 t (6н, (сн2)3). 13 c nmr (125.5 mhz, cdcl3, δ, ppm): 166.14 (e, z) (c 6 ); 23.84 (e); 23.98 (z) ( 6 c-ch3); 101.23 (z); 102.36 (e) ( 5 ch); 168.25 (z); 168.85(e) ( 4 c); 66.34 (z); 67.58 (e) (oc 3 ); 34.12 (e); 34.78 (z) (sc 2,4 ); 169.30 (e); 169.58 (z) ( 2 c); 39.26 (z); 40.71 (e) (2-sch2co); 171.12 (z); 172.02 (e) (2-sch2co); 28.14 (e, z) (n=c(ch2); 34.12 (e, z) (n=c(ch2); 27.58 (e, z) (2ch2); 25.8 (e, z) (ch2); 167.21 (z), 167.39 (e) (n=c). found,%: c 52.44; h, 6.06; n, 15.29; s 17.51. c16h22n4o2s2. calculated,%: c 52.44; h 6.05; n, 15.29; s 17.50. (e,z)-nʹ-[1-(4-aminophenyl)ethylidene]-2-[6-methyl-4-(thietan-3-yloxy)pyrimidin-2-ylthio]acetohydrazide (compound 6) pale yellow solid. recrystallization from n-buoh. yield: 0.35 g (83%); m.p. 176-177 °c. . ir (kbr) in cm −1 : 3470 (nh2), 3187-3299 (charom), 1667 («amide i»), 1588, 1574 (с=n, с=с), 1549 (δ nh, «amide ii»), 1164 (nas), 1045 (ns) (c-o-c), 799 (c-s). 1 h nmr (300 mhz, cdcl3, δ, ppm): 2.08 (z), 2.23 (e) (3н, s, 6 c–ch3,), 3.283.34(e), 3.44-3.50 (z) (4h, m, s(ch2)2), 5.79-5.90 (e, z) (1h, m, och), 6.18 (e), 6.30 (z) (1h, s, 5 c–h), 3.87 (z), 3.92 (e) (1h, s, sch2co), 8.69 (z), 9.72 (e) (1h, wid s, conh), 2.34 (e), 2.43 (z) s (3h, n=c(ch3)); 6.18 (e, z) s (2h, nh2); 6.67 (e, z) d (2нarom, j 8.4 hz); 7.60 (e, z) d (2нarom, j 8.6 hz). 13 c nmr (125,5 mhz, cdcl3, δ, ppm): 165.97 (e, z) (c 6 ); 23.61 (e); 24.08 (z) ( 6 c-ch3); 102.29 (e); 103.54 (z) ( 5 ch); 167.46 (e, z); ( 4 c); 69.47 (e); 70.23 (z) (oc 3 ); 35.09 (z); 35.49 (e) (sc 2,4 ); 168.49 (e, z) ( 2 c); 33.36 (e); 33.80 (z) (2-sch2co); 170.78 (e, z) (2sch2co); 12.95 (e), 13.69 (z) (=c(ch3)); 126.15 (e), 126.63 (z) (2снarom); 129.51 (e), 129.69 (z) (2снarom); 137.45 (e, z) (сarom); 148.29 (e), 149.05 (z) (=c(ch3)). found,%: c 53.58; h 5.25; n, 17.36; s 15.89. c18h21n5o2s2. calculated,%: c 53.58; h 5.25; n, 17.36; s 15.89. (e,z)-2-[6-methyl-4-(thietan-3-yloxy)pyrimidin-2-ylthio]-nʹ-(1-phenylethylidene)acetohydrazide (compound 7) white solid. recrystallization from n-buoh. yield: 0.32 g (79%); m.p. 178-179 °c. ir (kbr) in cm −1 : 31873299 (charom), 1667 («amide i»), 1588, 1574 (с=n, с=с), 1549 (δ nh, «amide ii»), 1164 (nas), 1045 (ns) (c-oc), 799 (c-s). 1 h nmr (300 mhz, cdcl3, δ, ppm): 2.14 (z), 2.30 (e) (3н, s, 6 c–ch3,), 3.28-3.31(e), 3.42-3.53 (z) (4h, m, s(ch2)2), 5.77-5.87 (e, z) (1h, m, och), 6.17 (e), 6.31 (z) (1h, s, 5 c–h), 3.93 (z), 4.42 (e) (1h, s, sch2co), 9.14 (e), 9.87 (z) (1h, wid s, conh), 2.33 (e), 2.44 (z) s (3h, n=c(ch3)); 7.38-7.42 (e, z) m (3нarom); 7.74-7.79 (e, z) м (2нarom). 13 c nmr (125,5 mhz, cdcl3, δ, ppm): 165.98 (e, z) (c 6 ); 23.71 (e); 24.01 (z) ( 6 cch3); 102.19 (e); 103.34 (z) ( 5 ch); 167.46 (e, z); ( 4 c); 69.47 (e); 70.23 (z) (oc 3 ); 35.09 (z); 35.49 (e) (sc 2,4 ); 168.49 (e, z) ( 2 c); 33.36 (e); 33.80 (z) (2-sch2co); 170.78 (e, z) (2-sch2co); 12,95 (e), 13,69 (z) (=c(ch3)); 126,15 (e), 126,63 (z) (снarom); 128,33 (z), 128,47 (e) (снarom); 129,51 (e), 129,69 (z) (снarom); 137,65 (e, z) (сarom); 148,29 (e), 149,05 (z) (=c(ch3)). found,%: c 55.61; h 5.19; n, 14.42; s 16.50. c18h20n4o2s2. calculated,%: c 55.65; h 5.19; n, 14.42; s 16.50. admet & dmpk 9(2) (2021) 167-176 pyrimidine derivatives and infection doi: http://dx.doi.org/10.5599/admet.941 171 (e,z)-nʹ-[1-(4-chlorophenyl)ethylidene]-2-[6-methyl-4-(thietan-3-yloxy)pyrimidin-2-ylthio]acetohydrazide (compound 8) white solid. recrystallization from i-buoh. yield: 0.32 g (79%); m.p. 170-171 °c. ir (kbr) in cm −1 : 1630 (nс=о), 3280 (nn-h), 1520 (dn-h); 1572-1591 (с=с, с=n), 2960 (nc-h), 1443 (dc=c 6 -ch3), 1167, 1050 (c-oc), 1090 (c-cl). 1 h nmr (300 mhz, cdcl3, δ, ppm): 2.12 (z), 2.18 (e) (3н, s, 6 c–ch3,), 3.26-3.31(e), 3.41-3.54 (z) (4h, m, s(ch2)2), 5.77-5.85 (e, z) (1h, m, och), 6.19 (e), 6.31 (z) (1h, s, 5 c–h), 3.93 (z), 4.40 (e) (1h, s, sch2co), 9.02 (e), 9.90 (z) (1h, wid s, conh), 2.33 (e), 2.44 (z) s (3h, n=c(ch3)); 7.37 (e, z) d (2нarom, j 8.4 hz); 7.71 (e, z) d (2нarom, j 7.5 hz). 13 c nmr (125,5 mhz, cdcl3, δ, ppm): 163.78 (e, z) (c 6 ); 23.71 (e); 23.97 (z) ( 6 c-ch3); 102.22 (e); 103.37 (z) ( 5 ch); 167.45 (e, z); ( 4 c); 69.43 (e); 70.22 (z) (oc 3 ); 35.09 (z); 35.49 (e) (sc 2,4 ); 168.52 (e, z) ( 2 c); 33.30 (e); 33.78 (z) (2-sch2co); 170.90 (e, z) (2-sch2co); 12,87 (e), 13,48 (z) (=c(ch3)); 123,87 (e, z) (2снarom); 131,49 (z), 131,74 (e) (2снarom); 136,86 (e, z) (2сarom); 147,57 (e, z) (=c(ch3)). found,%: c 51.12; h 4.53; n 13.25; s 15.16. c18h19cln4o2s2. calculated,%: c 51.12; h 4.53; n 13.25; s 15.16. (e,z)-nʹ-[1-(4-bromophenyl)ethylidene]-2-[6-methyl-4-(thietan-3-yloxy)pyrimidin-2-ylthio]acetohydrazide (compound 9) white solid. recrystallization from n-buoh. yield: 0.40 g (81%); m.p. 178-179 °c. ir (kbr) in cm −1 : 1630 (nс=о), 3280 (nn-h), 1520 (dn-h); 1572-1591 (с=с, с=n), 2960 (nc-h), 1443 (dc=c 6 -ch3), 1167, 1050 (c-oc), 1050 (c-br). 1 h nmr (300 mhz, cdcl3, δ, ppm): 2.11 (z), 2.28 (e) (3н, s, 6 c–ch3,), 3.26-3.30(e), 3.41-3.52 (z) (4h, m, s(ch2)2), 5.75-5.87 (e, z) (1h, m, och), 6.17 (e), 6.31 (z) (1h, s, 5 c–h), 3.92 (z), 4.39 (e) (1h, s, sch2co), 9.30 (e), 9.91 (z) (1h, wid s, conh), 2.32 (e), 2.43 (z) s (3h, n=c(ch3)); 6.91 (e, z) d (2нarom, j 8,7 hz), 7.51 (e, z) d (2нarom, j 8,6 hz). 13 c nmr (125,5 mhz, cdcl3, δ, ppm): 163.78 (e, z) (c 6 ); 23.71 (e); 23.97 (z) ( 6 c-ch3); 102.22 (e); 103.37 (z) ( 5 ch); 167.45 (e, z); ( 4 c); 69.43 (e); 70.22 (z) (oc 3 ); 35.09 (z); 35.49 (e) (sc 2,4 ); 168.52 (e, z) ( 2 c); 33.30 (e); 33.78 (z) (2-sch2co); 170.90 (e, z) (2-sch2co); 12,87 (e), 13,48 (z) (=c(ch3)); 123,87 (e, z) (2снarom); 131,49 (z), 131,64 (e) (2снarom); 136,56 (e, z) (2сarom); 147,30 (e, z) (=c(ch3)). found,%: c 46.26; h 4.10; n, 11.93; s 13.72. c18h19brn4o2s2. calculated,%: c 46.26; h 4.10; n, 11.99; s 13.72. (e,z)-2-[6-methyl-4-(thietan-3-yloxy)pyrimidin-2-ylthio]-nʹ-[1-(4-nitrophenyl)ethylidene]acetohydrazide (compound 10) yellow solid. recrystallization from n-buoh. yield: 0.35 g (76%); m.p. 169-170°c. ir (kbr) in cm −1 : 1630 (nс=о), 3280 (nn-h), 1520 (dn-h); 1572-1591 (с=с, с=n), 2960 (nc-h), 1560 (nas), 1350 (ns) (no), (1443 (dc=c 6 -ch3), 1167, 1050 (c-o-c). 1 h nmr (300 mhz, cdcl3, δ, ppm): 2.39 (z), 2.43 (e) (3н, s, 6 c–ch3,), 3.263.46(e), 3.51-3.71 (z) (4h, m, s(ch2)2), 5.14-5.26 (e, z) (1h, m, och), 6.11 (e), 6.26 (z) (1h, s, 5 c–h), 3.93 (z), 4.05 (e) (1h, s, sch2co), 9.98 (e), 10.57 (z) (1h, wid s, conh), 2.39 (e), 2.43 (z) s (3h, n=c(ch3)); 8.07 (e, z) d (2нarom, j 8,7 hz), 8.34 (e, z) d (2нarom, j 8,6 hz). 1 h nmr (300 mhz, cdcl3, δ, ppm): 2.11 (z), 2.28 (e) (3н, s, 6 c–ch3,), 3.26-3.30(e), 3.41-3.52 (z) (4h, m, s(ch2)2), 5.75-5.87 (e, z) (1h, m, och), 6.17 (e), 6.31 (z) (1h, s, 5 c–h), 3.92 (z), 4.39 (e) (1h, s, sch2co), 9.30 (e), 9.91 (z) (1h, wid s, conh), 2.32 (e), 2.43 (z) s (3h, n=c(ch3)); 6.91 (e, z) d (2нarom, j 8,7 hz), 7.51 (e, z) d (2нarom, j 8,6 hz). 13 c nmr (125,5 mhz, cdcl3, δ, ppm): 163.78 (e, z) (c 6 ); 23.71 (e); 23.97 (z) ( 6 c-ch3); 102.22 (e); 103.37 (z) ( 5 ch); 167.45 (e, z); ( 4 c); 69.43 (e); 70.22 (z) (oc 3 ); 35.09 (z); 35.49 (e) (sc 2,4 ); 169.32 (e, z) ( 2 c); 40.89 (e); 41.20 (z) (2-sch2co); 170.90 (e, z) (2sch2co); 12,87 (e), 16,48 (z) (=c(ch3)); 127,77 (e, z) (4снarom), 150,26 (e, z) (сarom); 143,60 (e, z) (2сarom). found,%: c 49.87; h 4.42; n, 16.16; s 14.79. c18h19n5o4s2. calculated,%: c 49.87; h 4.42; n, 16.16; s 14.79. (e,z)-nʹ-[1-(4-hydroxyphenyl)ethylidene]-2-[6-methyl-4-(thietan-3-yloxy)pyrimidin-2-ylthio]acetohydrazide (compound 11) white solid. recrystallization from etoh. yield: 0.33 g (79%); m.p. 188-189 °c. ir (kbr) in cm −1 : 3564 (ohttp://dx.doi.org/10.5599/admet.941 i. gareev et al. admet & dmpk 9(2) (2021) 167-176 172 h), 3109, 3073 (с-нarom), 1667 («amide i»), 1591, 1569 (с=n, с=с), 1541 (δ nh, «amide ii»), 1513 (n=c), 1167 (nas), 1053 (ns) (c-o-c). 1 h nmr (300 mhz, cdcl3, δ, ppm): 2.39 (z), 2.43 (e) (3н, s, 6 c–ch3,), 3.26-3.47 (e), 3.52-3.71 (z) (4h, m, s(ch2)2), 5.14-5.26 (e, z) (1h, m, och), 6.11 (e), 6.26 (z) (1h, s, 5 c–h), 3.93 (z), 4.05 (e) (1h, s, sch2co), 9.98 (e), 10.57 (z) (1h, wid s, conh), 1.69 (e), 1.81 (z) s (3h, n=c(ch3)); 9.68 (e, z) wid s (oн), 7.74 (e, z) d (2нarom, j 8,7 hz), 6.82 (e, z) d (2нarom, j 8,6 hz). 13 c nmr (125,5 mhz, cdcl3, δ, ppm): 163.78 (e, z) (c 6 ); 23.71 (e); 23.97 (z) ( 6 c-ch3); 102.22 (e); 103.37 (z) ( 5 ch); 167.45 (e, z); ( 4 c); 69.43 (e); 70.22 (z) (oc 3 ); 35.09 (z); 35.49 (e) (sc 2,4 ); 168.52 (e, z) ( 2 c); 33.30 (e); 33.78 (z) (2-sch2co); 170.90 (e, z) (2sch2co); 12,87 (e), 13,48 (z) (=c(ch3)); 123,97 (e, z) (2снarom); 132,49 (z), 133,64 (e) (2снarom); 137,56 (e, z) (2сarom); 147,30 (e, z) (=c(ch3)). found,%: c 53.44; h 4.98; n 13.85; s 15.85. c18h20n4o3s2. calculated,%: c 53.45; h 4.98; n 13.85; s 15.85. (e,z)-nʹ-[1-(2,5-dihydroxyphenyl)ethylidene]-2-[6-methyl-4-(thietan-3-yloxy)-pyrimidin-2-ylthio]acetohydrazide (compound 12) white solid. recrystallization from i-proh. yield: 0.36 g (81%); m.p. 180-181 °c.. ir (kbr) in cm −1 : 3254 (o-н), 1667 («amide i»), 1591, 1572 (с=n, с=с), 1533 («amide ii»), 1485 (n=c), 1167 (nas), 1044 (ns) (c-o-c). 1 h nmr (300 mhz, cdcl3, δ, ppm): 2.07 (z), 2.15 (e) (3н, s, 6 c–ch3,), 3.29-3.41 (e), 3.52-3.71 (z) (4h, m, s(ch2)2), 5.14-5.26 (e, z) (1h, m, och), 6.11 (e), 6.26 (z) (1h, s, 5 c–h), 3.93 (z), 4.05 (e) (1h, s, sch2co), 9.98 (e), 10.57 (z) (1h, wid s, conh), 1.69 (e), 1.82 (z) s (3h, n=c(ch3)); 13.18 (e, z) wid s (oн), 9.46 wid s (oн),6.66 (e, z) d (2нarom, j 8,4 hz), 6.75 (e, z) d (2нarom, j 8,6 hz), 7.08 (e, z) d (2нarom, j 8,7 hz). 13 c nmr (125,5 mhz, cdcl3, δ, ppm): 163.78 (e, z) (c 6 ); 23.71 (e); 23.97 (z) ( 6 c-ch3); 102.22 (e); 103.37 (z) ( 5 ch); 167.45 (e, z); ( 4 c); 69.43 (e); 70.22 (z) (oc 3 ); 35.09 (z); 35.49 (e) (sc 2,4 ); 168.52 (e, z) ( 2 c); 33.30 (e); 33.78 (z) (2-sch2co); 170.90 (e, z) (2-sch2co); 12,87 (e), 13,48 (z) (=c(ch3)); 116,37 (e, z) (снarom); 119.62 (e, z) (снarom); 120.42 (e, z) (снarom); 120.21 (e, z) (сarom); 151.21 (e, z) (сarom); 155.51 (e, z) (сarom); 147,30 (e, z) (=c(ch3)). found,%: c 51.41; h 4.79; n, 13.32; s 15.25. c18h20n4o4s2. calculated,%: c 51.41; h 4.79; n, 13.32; s 15.25. biological studies antibacterial and antifungal activities of the novel 2-[6-methyl-4-(thietan-3-yloxy)pyrimidin-2ylthio]acetohydrazide derivatives were analyzed using the agar diffusion and the tenfold broth (ph 7.2 – 7.4) dilution methods. microbial strains of the department of microbiology and virology, bashkir state medical university deposited at l.a. tarasevich state institute of standardization and control of biomedical preparations, the ministry of health of the russian federation were used as test organisms: e. coli, c. diversus, ent. aerogenes, p. vulgaris, k. pneumoniae, ps. aeruginosa, st. aureus, str. pyogenes and lower fungi c. albicans. test compounds (10 mg) were weighed and dissolved in 1 ml dmso. these solutions were diluted in beef extract broth to achieve a final concentration of 0.1 mg/ml (stock solution). the nutrient broth, which contained logarithmic serially twofold diluted amount of test compound inoculated with 2.0∙10 6 c.f.u/ml, was used. the cultures were incubated for 24 h at 37 °c and the growth was monitored visually. the lowest concentration (highest dilution) required to arrest the growth of bacteria was regarded as minimum inhibitory concentration (mic). results and discussion chemistry this study was undertaken to synthesise acetohydrazide derivatives. synthesis of the desired compound 2 was carried out as depicted in figure 1. the structure of hydrazide 2 was confirmed by a complex of nmr spectroscopy methods. admet & dmpk 9(2) (2021) 167-176 pyrimidine derivatives and infection doi: http://dx.doi.org/10.5599/admet.941 173 e,z-isomerism was observed in acetohydrazide 2, due to the inhibited rotation around the n-c hydrazide bond, as evidenced by the doubling of the proton signals of the thioacetohydrazide fragment. the spatial form of hydrazide 2 in dmso-d6 was established using two-dimensional nmr spectroscopy methods: noesy and roesy (figure 2). according to the spectral data, the sterically more stable z-conformer prevails, as evidenced by the presence of the nuclear overhauser effect between the signal of the methylene protons of the acetamide group, and the signal of the proton bonded to the nitrogen atom of the hydrazide fragment, indicating that these protons are closer in space. figure 2. two-dimensional 1 h1 h roesy spectrum (500 mhz, dmso-d6) and structural schematic presentation of compound 2 the signals of z-conformer protons are registered in a stronger field in comparison with the same signals of the e-conformer. according to the 1 h nmr spectra, the ratio of conformers is 19 (z): 1(e). these hydrazones might exist in the form of e,z-isomers due to hindered rotation around the n-co bond and eʹ,zʹ-isomers around the c=n bond. the presence of two sets of chemical shifts in the 1 h nmr spectra indicates that these compounds exist as a mixture of two stereoisomers. acetone hydrazone might exist only as e',z'-isomers; and two sets of chemical shifts in the 1 h nmr spectra can be caused only by hindered rotation around the n-co bond. two sets of proton signals are recorded in the 1 h nmr spectra of compound 3, dissolved in polar solvents. when a solution of compound 3 in dmso-d6 is heated, coalescence occurs and the two sets of isomer signals merge into one. it confirms that the synthesized hydrazones in polar solvents exist as a mixture of two conformational e,z isomers. analysis of the 1 н nmr spectrum of hydrazone in dmso-d6 at 20.6° c showed that the signals of the protons of the сн2со, nh groups one of the methyl groups (сн3)2с=n of the e isomer are shifted to a weaker field than the corresponding signals of the z-isomer, which is consistent with the existing literature data. it is established that the sterically more stable en-co isomer predominates, according to the analysis of the integral intensities of the 1 h nmr spectrum signals of hydrazone 3. in the 1 h nmr spectrum of hydrazone 3, in non-polar solvents, i.e. benzene, one set of resonance signals was recorded, in which the values of the chemical shift indicate that hydrazone 3 in this case exists only in amide e-form. http://dx.doi.org/10.5599/admet.941 i. gareev et al. admet & dmpk 9(2) (2021) 167-176 174 antimicrobial and antifungal activity synthesized compounds were further screened for minimum inhibitory concentration against test organisms: e. coli, c. diversus, ent. aerogenes, p. vulgaris, k. pneumoniae, ps. aeruginosa, st. aureus, str. pyogenes and lower fungi c. albicans. minimum inhibitory concentration values are given in table 1. table 1. antibacterial and antifungal activities of the newly prepared compounds compound minimum inhibitory concentrations (mic), µg/ml st. aureus str. pyogenes e. coli p. vulgaris k. pneumoniae ent. aerogenes ps. aeruginosa e. сloacae c. аlbicans 1 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 2 0.5 0.05 0.5 0.05 0.05 0.05 0.5 0.05 0.5 3 50 5 0.5 5 0.05 0.05 0.05 0.05 0.05 4 5 0.5 0.5 0.5 0.05 0.05 5 0.05 0.5 5 50 5 0.05 5 5 0.5 0.05 5 0.5 6 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.5 0.05 7 0.5 0.5 0.5 0.5 0.05 0.05 0.5 0.05 0.05 8 50 0.5 0.5 5 0.5 0.5 0.5 0.5 0.05 9 50 5 5 5 0.5 5 5 0.5 5 10 50 5 0.5 5 5 0.5 0.05 5 5 11 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 12 0.05 0.05 0.5 0.05 0.05 0.05 0.05 0.05 0.05 ceftriaxone 0.5 0.5 0.5 0.5 0.05 0.05 0.05 0.5 0.05 pimafucin 0.05 the investigation of antibacterial screening data revealed that the compounds 6, 11 and 12 showed good inhibition towards all tested gram-positive and gram-negative bacteria, and lower fungi c. albicans. the derivative of 2-[6-methyl-4-(thietan-3-yloxy)pyrimidin-2-ylthio]acetohydrazide 7, prepared in reaction with acetophenone, showed good antibacterial (mic 0,05 – 0,5 µg/ml) and antifungal (mic 0.05 µg/ml) activities, but it showed weak antimicrobial activity (mic 0.5 µg/ml) against pseudomonas aeruginosa. compounds 8, 9 and 10 are less active against st. aureus, p. vulgaris. the results of antifungal testing revealed that compounds 1, 2, 4, 5, 9 and 10 are less active against c. albicans. all the other derivatives showed moderate activity against c. albicans. the structure–activity studies showed that, depending on the nature of the hydrazide fragment, the newly synthesized compounds exhibited varying degrees of microbial inhibition. it was found that hydrazide showed better activity, as compared to the ester against str. pyogenes, p. vulgaris, k. pneumonia, ent. aerogenes, e. cloacae. the presence of alkyl fragments such as methyl and ethyl radicals decrease activity against str. pyogenes, p. vulgaris. the presence of a cyclohexylidene substituent increases antimicrobial activities e. coli, ps. aeruginosa, reduces antimicrobial activities against st. aureus, str. pyogenes, p. vulgaris, k. pneumonia, ent. aerogenes, e. сloacae. within the same series the antimicrobial activity depends on the nature of the substituent attached to the benzene ring. based on the data obtained it was possible to investigate some structure-activity relationships: oh, nh2 groups in the benzene fragment in the molecule led to an increase in antimicrobial and antifungal activities; compounds with halogen, no2 groups in the benzene ring are the least active. the oh group attached to a benzene ring increased the inhibitory activity of compounds against st. aureus. str. pyogenes, p. vulgaris, ps. aeruginosa. thus, the nature of the substituent on the benzene ring attached to 2-[6-methyl-4-(thietan-3yloxy)pyrimidin-2-ylthio]acetohydrazide derivatives has a strong influence on the extent of antimicrobial activity. admet & dmpk 9(2) (2021) 167-176 pyrimidine derivatives and infection doi: http://dx.doi.org/10.5599/admet.941 175 conclusions new 2-[6-methyl-4-(thietan-3-yloxy)pyrimidin-2-ylthio]acetohydrazide derivatives have been successfully obtained and tested as potential antimicrobial and antifungal agents. n-ylidenehydrazides were synthesized by the reaction of hydrazide with alkyl and aryl ketones, without the use of acid catalysers, with good yields of 61–89%. we have characterized a hydrazone-based molecular switch and studied the isomerization mechanism of this acetohydrazide system. solvent and temperature induced switching between e,z-isomers. the synthesized compounds 1-12 have been investigated for their in vitro antibacterial and antifungal activities. it was found that the compounds nʹ-[1-(4-aminophenyl)ethylidene]2-[6-methyl-4-(thietan-3-yloxy)pyrimidin-2-ylthio]acetohydrazide (compound 6), nʹ-[1-(4-hydroxyphenyl)ethylidene]-2-[6-methyl-4-(thietan-3-yloxy)pyrimidin-2-ylthio]acetohydrazide (compound 11), nʹ-[1-(2,5dihydroxyphenyl)ethylidene]-2-[6-methyl-4-(thietan-3-yloxy)-pyrimidin-2-ylthio]acetohydrazide (compound 12) obtained from the reaction of p-aminoacetophenone, p-hydroxyacetophenone and 2,5dihydroxyacetophenone exhibit high antimicrobial and antifungal activity. highest antifungal activity was observed in compounds 3, 6, 7, 8, 11 and 12. compounds that contain thietan and pyrimidine in their molecules are suitable candidates for further chemical modifications and are pharmacological active. conflict of interest: the authors declare no conflict of interest. references [1] a. pieri, r. aschbacher, g. fasani, j. mariella, l. brusetti, e. pagani, m. sartelli, l. pagani. country income is only one of the tiles: the global journey of antimicrobial resistance among humans, animals, and environment. antibiotics (basel) 9 (8) (2020) 473. doi: https://doi.org/10.3390/antibiotics9080473. [2] m. cazzola, p. rogliani, s. aliberti, f. blasi, m.g. matera. an update on the pharmacotherapeutic management of lower respiratory tract infections. expert opin. pharmacother. 18 (10) (2017) 973988. doi: https://doi.org/10.1080/14656566.2017.1328497. [3] s.b. jadhav, p.k. vibhute, a.k. aghao. design, synthesis and microbial activity of some novel 4,6diphenyl-2-amine pyrimidine derivatives. heterocyclic letters 10 (1) (2020) 73-78. [4] s.m.h. sanad, a.a.m. ahmed, a.e.m mekky. efficient synthesis and molecular docking of novel antibacterial pyrimidines and their related fused heterocyclic derivatives. journal of heterocyclic chemistry 57 (2020) 590-605. doi: https://doi.org/10.1002/jhet.3789. [5] j. rani, s. kumar, m. saini, j. mundlia, p.k. verma. biological potential of pyrimidine derivatives in a new era. res. chem. intermed. 42 (2016) 6777–6804. doi: https://doi.org/10.1007/s11164-016-25258. [6] r.l. sawant, c.a. bansode, j.b. wadekar. in vitro anti-inflammatory potential and qsar analysis of oxazolo/thiazolo pyrimidine derivatives. med. chem. res. 22 (2013) 1884–1189. doi: https://doi.org/10.1007/s00044-012-0189-5. [7] s. meneghesso, e. vanderlinden, a. stevaert, c. mcguigan, j. balzarini, l. naesens. synthesis and biological evaluation of pyrimidine nucleoside monophosphate prodrugs targeted against influenza virus. antivira.l res. 94 (1) (2012) 35-43. doi: https://doi.org/10.1016/j.antiviral.2012.01.007. [8] v.a. kataev, s.a. meshcheryakova, v.v. lazarev, v.v. kuznetsov. synthesis of thietanyl-substituted pyrimidine-2,4(1h,3h)-dions. russ. j. org. chem. 49 (2013) 743–745. doi: https://doi.org/10.1134/s1070428013050199. [9] v. kataev, s. meshcheryakova, e. meshcheryakova, t. tyumkina, l. khalilov, v. lazarev, v. kuznetsov. direction of the reaction of 6-methylpyrimidine-2,4(1h,3h)-dione with 2-chloromethylthiirane: n 1 or n 3 -thietanyl derivative? russian journal of organic chemistry 54 (2018) 918-922. doi: https://doi.org/10.1134/s1070428018060143. [10] s.a. meshcheryakova, v.a. kataev, i.ya. fattakhova, k.v. nikolaeva, a.k. bulgakov. synthesis and antimicrobial activity of thietanylpyrmidin-2,4(1h,3h)-dione acetanilides and acetylhydrazones. http://dx.doi.org/10.5599/admet.941 https://doi.org/10.3390/antibiotics9080473 https://doi.org/10.3390/antibiotics9080473 https://doi.org/10.1080/14656566.2017.1328497 https://doi.org/10.1002/jhet.3789 https://doi.org/10.1007/s11164-016-2525-8 https://doi.org/10.1007/s11164-016-2525-8 https://doi.org/10.1007/s00044-012-0189-5 https://doi.org/10.1016/j.antiviral.2012.01.007 https://doi.org/10.1134/s1070428013050199 https://doi.org/10.1134/s1070428013050199 https://doi.org/10.1134/s1070428018060143 i. gareev et al. admet & dmpk 9(2) (2021) 167-176 176 pharmaceutical chemistry journal 49 (9) (2016) 598-601. doi: https://doi.org/10.1007/s11094-0151336-3. [11] s. meshcheryakova, v. kataev, d. munasipova, a. shumadalova, a. bulgakov. synthesis, structure, antimicrobial and antifungal activities of diacylhydrazines of thiethanylpyrimidine-2,4(1h,3h)-dione series. problems of biological, medical and pharmaceutical chemistry 20 (2017) 20-25. [12] v. kataev, v.a.i. tyurenkov, s. meshcheryakova, v. perfilova, d. munasipova, d. borodin. synthesis and hypotensive activity of pyrimidine-2,4-(1h,3h)-dione derivatives containing thietane rings with sulfur in various oxidation states. pharmaceutical chemistry journal 48 (2014) 434-438. doi: https://doi.org/10.1007/s11094-014-1126-3. [13] s. meshcheryakova, v. kataev, k. nikolaeva, v. perfilova, d. borodin, i. tyurenkov. synthesis, isomerism, and hypotensive activity of thiethane-containing hydrazones of uracilylacetic acid. russian journal of bioorganic chemistry 40 (2014) 300-307. doi: https://doi.org/10.1134/s1068162014030108. [14] s. meshcheryakova, v. kataev, r. galimova. synthesis and broncholytic activity of n-alkylpiperazinosubstituted uracil and xanthine. pharmaceutical chemistry journal 47 (2013) 247–250. doi: https://doi.org/10.1007/s11094-013-0938-x. [15] a. gabdrakhmanova, s. meshcheryakova, r. gaynutdinova, f. kil'diyarov, s. kurbanov, a. khuzhamberdiev. experimental study of the effectiveness of 6-methyl-3-(thietane-3-yl)uracilcontaining eye ointment in the treatment of corneal thermal burns. kazanskii meditsinskii zhurnal 100 (2019) 657-661. doi: https://doi.org/10.17816/kmj2019-657. [16] s. meshcheryakova, a. shumadalova. interaction of 2-[6-methyl-4-(thiethan-3-yloxy)pyrimidin-2-yl thio]acetic acid hydrazide with n-arylmaleimides. bashkir chemical journal 23 (2016) 45-48. ©2021 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) https://doi.org/10.1007/s11094-015-1336-3 https://doi.org/10.1007/s11094-015-1336-3 https://doi.org/10.1007/s11094-014-1126-3 https://doi.org/10.1134/s1068162014030108 https://doi.org/10.1134/s1068162014030108 https://doi.org/10.1007/s11094-013-0938-x https://doi.org/10.17816/kmj2019-657 http://creativecommons.org/licenses/by/3.0/ applications of biophysical techniques in drug discovery and development doi: 10.5599/admet.767 220 admet & dmpk 7(4) (2019) 220-221; doi: http://dx.doi.org/10.5599/admet.767 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index editorial applications of biophysical techniques in drug discovery and development tatjana ž. verbić university of belgrade – faculty of chemistry, studentski trg 12-16, 11000 belgrade, serbia e-mail: tatjanad@chem.bg.ac.rs; tel.: +381-11-3336-611. received: december 10, 2019; revised: december 10, 2019; published: december 11, 2019 drug discovery and development in many pharmaceutical companies and academic laboratories have benefited from the rapid advancements in biophysical experimental methods over the last two decades. at the end of the 20 th century, application of these methods was still restricted to a (relatively) small number of scientists using specialized, low throughput technologies and methods. now, automated highthroughput instruments are to be found in a growing number of laboratories [1]. techniques such as mass spectrometry, multidimensional nuclear magnetic resonance spectroscopy, image processing, x-ray diffraction, electron microscopy, atomic force microscopy, fluorescence spectroscopy, dynamic light scattering, surface plasmon resonance spectroscopy, differential scanning calorimetry and isothermal titration calorimetry are now routinely used in many key stages of the drug discovery and development. generally speaking, biophysical methods are in two ways involved in drug design: the qualitative detection of small molecule binding to a target (hit identification), and the quantitative determination of physical parameters associated to binding (hit-to-lead progression) [2]. in the first case, efforts have been made toward miniaturization, automation, and speed-up of the screening process allowing higher throughput. in the second one, sophisticated applications have been developed to derive detailed relevant information. the demand for more automated and higher-throughput versions of instrumentation continues to grow. with such advancement in instrumentation, great improvements in the speed, sensitivity and the range of possible biophysical measurements have been reached. high-resolution mechanistic, kinetic, thermodynamic and structural information on drug-target interactions are now available. in particular, biophysical measurements are useful in supporting compound progression, mechanistic understanding of the drug-receptor binding, validating potency data from biochemical and cellular assays in the discovery phases and quality control of the investigational drug, including the evaluations of drug release and stability, in the development phases. in view of the rapid advancement and the important roles of biophysical techniques in drug discovery and development, admet and dmpk devoted a special issue to this topic. among received articles, three were chosen to be published, one original scientific article and two reviews. in the following issues of the journal a couple of more articles are expected to be published. http://dx.doi.org/10.5599/admet.767 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:tatjanad@chem.bg.ac.rs admet & dmpk 7(4) (2019) 220-221 applications of biophysical techniques in drug discovery and development doi: 10.5599/admet.767 221 the manuscript written by g. holdgate, k. embrey et al. [3] provides a review about the use of biophysical methods in early phases of drug discovery, which represents an update of previously published book chapter [1]. the use of various biophysical techniques is discussed, including thorough assay development, primary screening, hit confirmation, and mechanistic characterization. each method has its own advantages and disadvantages that lead to different applications dependent upon the reagents available and the information content desired. two case studies are presented. the use of nmr spectroscopy in the various drug discovery and development processes is discussed in the manuscript written by m. zloh [4]. applications of hydrogen nuclear magnetic spectroscopy ( 1 h nmr spectroscopy), not often reported as a tool for evaluating some admet properties, are explored. the use of quantitative nmr (qnmr) in solubility, lipophilicity (log p) and acidity constants (pka) determinations is discussed. the work by i. quiroga and t. scior [5] describes the role of induced fit mechanism in binding to cytochrome p450 3a4. root means square deviations (rmsd values) of ligand bound/unbound structures of 3a4 are analyzed after the molecular dynamic (md) simulations. the study demonstrates that the smallest dynamic differences are met in pairs which share either the presence or the absence of bound substrates. substrate alone does provide an effect on the geometric changes of the enzyme which can be observed and measured in md studies. the guest editor would like to thank all the contributors for the time and efforts put in presented manuscripts. it is hoped that readers will find these contributions useful in their research. references [1] g. holdgate, s. geschwindner, a. breeze, g. davies, n. colclough, d. temesi, l. ward. biophysical methods in drug discovery from small molecule to pharmaceutical. methods in molecular biology 1008 (2013) 327-355. [2] j-p. renaud, m-a. delsuc. biophysical techniques for ligand screening and drug design. current opinion in pharmacology 9 (2009) 622–628. [3] g. holdgate, k. embrey, a. milbradt, g. davies. biophysical methods in early drug discovery. admet and dmpk 7(4) (2019) 222-241. [4] m. zloh. nmr spectroscopy in drug discovery and development: evaluation of physico-chemical properties. admet and dmpk 7(4) (2019) 242-251. [5] i. quiroga, t. scior. induced fit for cytochrome p450 3a4 based on molecular dynamics. admet and dmpk 7(4) (2019) 252-266. ©2019 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) http://creativecommons.org/licenses/by/3.0/ electrospun nanofibers: a nanotechnological approach for drug delivery and dissolution optimization in poorly water-soluble drugs doi: https://dx.doi.org/10.5599/admet.844 325 admet & dmpk 8(4) (2020) 325-353; doi: https://doi.org/10.5599/admet.844 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index review electrospun nanofibers: a nanotechnological approach for drug delivery and dissolution optimization in poorly water-soluble drugs luis castillo-henríquez 1,2, *, rolando vargas-zúñiga 1 , jorge pacheco-molina 3 and jose vega-baudrit 2,4 1 physical chemistry laboratory, faculty of pharmacy, university of costa rica, 11501-2060, san josé, costa rica 2 national laboratory of nanotechnology (lanotec), national center for high technology (cenat), 1174-1200, san josé, costa rica 3 laboratory of pharmaceutical technology, faculty of pharmacy, university of costa rica, 11501-2060, san josé, costa rica 4 laboratory of polymers (poliuna), chemistry school, national university of costa rica, 86-3000, heredia, costa rica *corresponding author: e-mail: luis.castillohenriquez@ucr.ac.cr; tel.: +506-6196-1666; fax: +506-2511-8315 received: may 01, 2020; revised: july 02, 2020; published: july 05, 2020 abstract electrospinning is a novel and sophisticated technique for the production of nanofibers with high surface area, extreme porous structure, small pore size, and surface morphologies that make them suitable for biomedical and bioengineering applications, which can provide solutions to current drug delivery issues of poorly water-soluble drugs. electrospun nanofibers can be obtained through different methods asides from the conventional one, such as coaxial, multi-jet, side by side, emulsion, and melt electrospinning. in general, the application of an electric potential to a polymer solution causes a charged liquid jet that moves downfield to an oppositely charged collector, where the nanofibers are deposited. plenty of polymers that differ in their origin, degradation character and water affinity are used during the process. physicochemical properties of the drug, polymer(s), and solvent systems need to be addressed to guarantee successful manufacturing. therefore, this review summarizes the recent progress in electrospun nanofibers for their use as a nanotechnological tool for dissolution optimization and drug delivery systems for poorly water-soluble drugs. ©2020 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords electrospinning; drug development; drug loading; drug release; nanotechnology; polymers; solubility. introduction nanofibers can be considered as nanomaterials based on their diameter, because they geometrically fall into the category of one-dimensional nanoscale elements, such as nanotubes and nanorods. however, their highly flexible nature makes them similar to two-dimensional elements and others like globular molecules as well. in addition to that, they can be viewed as nanostructured materials when filled with nanoparticles in order to form a composite. therefore, nanofibers possess characteristics such as high surface area, extreme porous structure, small pore size and surface morphologies that make them suitable for drug https://dx.doi.org/10.5599/admet.844 https://doi.org/10.5599/admet.844 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:luis.castillohenriquez@ucr.ac.cr http://creativecommons.org/licenses/by/4.0/ luis castillo et al admet & dmpk 8(4) (2020) 325-353 326 development and biomedical applications, providing sophisticated and novel solutions to current drug delivery inconveniences [1-3]. electrospun nanofibers have been the target of different applications, due to their structure and physicochemical properties. these types of nanofibers have been studied by schaub et al. based on their regenerative potential in spinal cord injuries [4,5]. according to sundara et al., their 3d architecture makes them very similar to the skin extracellular matrix structure, which allows their use as scaffolds for skin tissue engineering [6]. in addition to that, miguel et al. have worked with electrospun nanofibers loaded with bioactive molecules for improving wound healing [7]. another application, sustained release, has been evaluated by chou et al. for hydrophilic small molecule drugs [8]. recently, woon and sun have developed a charged poly(vinylidene fluoride) (pvdf) nanofiber filter to overcome novel coronavirus (sars-cov-2) capture limitations exhibited by n95 and n98 masks. this technology effectively traps virus particles at 100 nm, which reduces the risk of acquiring covid-19. their work might set new technologies for protective clothing from biological agents as well [9]. some other uses that have been given to electrospun nanofibers within the biomedical and pharmaceutical industries are water filtration, cosmetic masks, and nanosensors [10,11]. it is worth mentioning that a wide array of polymer types are being used in the creation of nanofibers, like synthetic, natural, biodegradable, non-degradable or a blend of them. thanks to the nanotechnological boom, many techniques are available and employed for the manipulation of polymers and nanofibers’ preparation [12-14]. methods can be either chemical or physical; some of these are drawing, melt blowing, template synthesis, sea-island spinning, self-assembly, phase separation and electrospinning [15-20]. among the mentioned, electrospinning is the most employed method for the production of nanofibers as an alternative delivery strategy for poorly water-soluble drugs that are not suitable for oral administration [21,22]. also, it is considered as the most efficient technique since it has been recognized as a simple, versatile and low-cost process. electrospun nanofibers are produced by an electrostatically driven jet of a solid dispersion consisting of a polymer solution. compared to other conventional solid dispersion techniques, it can produce nanofibers with enhanced dissolution and improved physicochemical properties of active pharmaceutical ingredient’ (api) particles, which can be attributed to the amorphization of the enclosed drug [23-26]. however, among the mentioned applications of electrospun nanofibers, the most relevant is their capacity to act as drug delivery systems due to their high loading capacity, faster dissolution kinetics, apis simultaneous delivery, and encapsulation. these outstanding properties make them attractive for the industry [27-30]. therefore, the present review focuses on the different electrospinning methods and materials employed for obtaining nanofibers, as well as recent and relevant investigations that have improved the delivery and dissolution performance of poorly water-soluble drugs, including smart nanofibers. historical context in the 17 th century, william gilbert was the first in documenting research with high voltage direct current applied to a solution. he observed a cone shape when a drop of water was close to electrically charged amber [31]. by the end of the 1800s, many papers had been released regarding the use of ‘electrical spinning’, using materials such as shellac, beeswax, and others, whilst in the early 1900s, the electrical charge was used mainly to spray liquids, as established by cooley and morton in their patents about electrical methods of dispersing fluids [32,33]. admet & dmpk 8(4) (2020) 325-353 electrospun nanofibers: a nanotechnological approach doi: https://dx.doi.org/10.5599/admet.844 327 nevertheless, the first electrospinning patent ‘process and apparatus for preparing artificial threads’ came in 1934 by formhals, where he described an experimental setup characterized by a collecting device, drum wheel and a plate for the production of polymer filaments using electrostatic force, which later was going to be a novel technique in nanotechnology [34]. however, it was not until 1974 when taniguchi in his publication ‘on the basic concept of nanotechnology’, that this scientific field was referenced for the first time [35]. nanotechnology is conceived in its simplest definition as the science and technology that takes place on the nanoscale (1-100 nm). precisely, due to the previous, it has been possible to obtain novel materials and structures that are integrated into larger systems improving not only chemical, physical and biological properties, but also different kinds of processes [36,37]. although electrospinning as a polymer-processing technology has been known for more than 80 years, it was not until 1994 when significant studies explored the production of nanofibers through this technique. by that year, the term ‘electrospinning’ was adopted among the scientific community thanks to the work of different researchers including doshi, reneker, and chun, who demonstrated that many organic polymers can be electrospun into nanofibers [38,39]. since then, significant effort has been done for the study and employment of this nanofiber fabrication technique. electrospinning techniques it is well known that electrospinning has its origin in ‘electrospraying’, wherein both cases the application of an electric potential to a polymer solution causes a charged liquid jet that moves downfield to an oppositely charged collector [40,41]. nevertheless, electrospraying is performed using a low viscosity liquid which due to the surface tension produces instability in the jet, resulting in the formation of polymer droplets instead of nanofibers, as occurs in electrospinning when working with polymer solutions of higher viscosity. the process can be done under room temperature unless it is necessary a heat source to keep the polymer molten [41,42]. in addition to that, nanofibers properties will depend on the type of polymer used and the method of manufacture, since there have been some modifications to the conventional electrospinning technique. the methods can be classified as needle or needleless electrospinning, where the latter is represented by roller and wire electrospinning, which offer some advantages over needle methods like the lack of clogging [43,44]. however, those advantages are not as significant as the differences within needle systems, thus this review will focus on needle techniques presented in figure 1. a single needle set-up, like solution electrospinning, is mostly used for research purposes, while multiple systems such as coaxial, multi-jet, side by side, emulsion, and melt electrospinning have been designed to improve nanofibers manufacturing [45,46]. figure 1. electrospinning techniques. https://dx.doi.org/10.5599/admet.844 luis castillo et al admet & dmpk 8(4) (2020) 325-353 328 solution electrospinning also known as co-electrospinning or blending electrospinning, it is conceived as a solid dispersion method that involves the dissolution of a drug in a polymer solution using a suitable solvent. it requires simpler equipment than other fiber manufacturing techniques, but it can create a great variety of products for different applications [47,48]. the usual set up design showed in figure 2 consists of a feeding unit with a spinneret, typically a syringe controlled by a pump that transports the polymer solution. some novel spinnerets recently used by several research groups for improving the nanofiber fabrication process are rotating disk, plucked string, plated edge, and multiple rings. there is also necessary a high-voltage supplier connected to the spinneret and the collector, charging them oppositely [49-51]. figure 2. electrospinning basic set-up design. adapted with permission from y. li et al. electrospinning in tissue engineering. in: a. haider, s. haider. electrospinning: material, techniques, and biomedical applications p.117-139. copyright (2016) intechopen [51]. regarding the process, the polymer solution is held by its surface tension at the end of the spinneret or capillary tip. once the solution is subjected to high voltage by a direct current power supply, the droplet at the tip elongates to develop a conical form known as ‘taylor cone’ [52]. when the electric field or electrostatic repelling force between the spinneret and the collector reaches a critical voltage, it overcomes the surface tension of the polymer solution, resulting in the ejection of a continuously charged jet from the tip of the taylor cone [53-55]. as it travels to the rotating or stationary collector two main events are worth highlighting. in the first place, the solvent evaporates but not completely, which represents a challenge for research groups due to solvent’s toxicity. on the other hand, the fine filament travels following a chaotic trajectory in a whip-like motion, known as bending instability. when it reaches the collector the continuous filament accumulates or deposits to form the nanofiber [56-60]. tan et al. employed this technique to fabricate vascular grafts made of poly(ɛ-caprolactone) (pcl), gelatin, and poly(vinyl alcohol) (pva). the obtained nanofibers rapidly degradated in vivo, also possessed appropriate properties for use in artery replacement, and offered scaffolds with high porosity enhancing cell proliferation and infiltration [61]. in another approach, ahmed et al. reported the creation of pcl /cellulose nanofibers for the use as biosensor strips [62]. coaxial electrospinning coaxial electrospinning presents changes regarding the method described in solution electrospinning. in this multiple needle system, two polymer solutions or a polymer and a drug solution with different properties regarding solubility and hydrophilicity, are loaded into individual feeding units to simultaneously admet & dmpk 8(4) (2020) 325-353 electrospun nanofibers: a nanotechnological approach doi: https://dx.doi.org/10.5599/admet.844 329 electrospin from coaxial spinnerets. since the last ones share an axis, one solution is injected through the inner needle into the concentric outer one at the capillary tip, resulting in a continuous core-shell nanofiber. in addition to that, the principle can be applied to triaxial systems as well, where the electrospun nanofiber will contain a core, a middle layer, and sheath [63-65]. the previously described process allows confining apis inside the core and controlling drug delivery mainly through diffusion rather than desorption of the surface [66]. moreover, coaxial electrospinning using hydrophilic and hydrophobic polymers within the composition provides a rich combination of their properties in a single engineered nanofiber while each component maintains its identity. this type of electrospun nanofiber leads to better control of drug diffusion to the medium [67-69]. wang et al. provided an approach for the design of hypromellose (hpmc)-based hydrophilic composites fabricated through coaxial electrospinning. the core fluid of the electrospun nanofibers was prepared from an electrospinnable solution containing 13 % (w/v) of hpmc, 2 % (w/v) of ferulic acid and 3 % (w/v) of poly(ethylene glycol) (peg) in a 1:1 ethanol/dichloromethane (dcm) mixture, while ethanol was used as the sheat solvent for the process. the performance of the electrospun nanofibers in the dissolution test revealed to be 10 times faster than casting hydrophilic composites. therefore, this approach can be useful as a means to obtain carriers for the delivery of poorly water-soluble drugs [70]. another research published by wu et al. proposes an alternative approach. in this case, the sheath layer consisted of a solution of poly(vinylpyrrolidone) k60 (pvp) and citric acid, while the core was composed of a peg 6000 and sodium hydroxide co-dissolving solution [71]. like in solution electrospinning, the application of a high voltage results in the formation of a taylor cone. however, in this case, it is compounded and formed by the core solution surrounded by the sheath polymer solution [72,73]. this method can solve the problem of those highly unstable substances of biomedical interest like enzymes, which can’t be electrospun due to their molecular weight or solubility properties. furthermore, when these compounds are loaded into the core the sheath structure protects them from the reactive environment [74]. multi-jet electrospinning a multiple spinnerets or nozzles set-up design is readily used for increasing the production rate of electrospun nanofibers due to its simplicity. this method can be classified into three categories: a) single spinneret with multiple jets, b) multiple spinnerets and each one with a single jet, and c) multiple spinnerets with multiple jets per each. therefore, a mass production increase of nanofibers can be achieved by increasing not only the nozzles but also the tips or holes [75-77]. multi-jet electrospinning possesses a very similar process compared to the conventional method, where a continuously charged jet trajectory is defined by the electrical field. however, in this system, the jet direction towards the collector is greatly influenced by coulomb forces caused by neighboring jets. precisely, one of the major issues of this method is a jet deviation that is exerted by the interaction of electrostatic forces between them, which provokes defects in the nanofibers [75,78,79]. in order to reduce jet repulsion, many approaches have been performed such as increasing the applied voltage, different spinnerets arrangements, and adjusting the distance between them. regarding that, liu et al. proposed placing an auxiliary grounded electrode to solve the problem, which improves the production rate as well [80]. later work from liu’s group evaluated the influence of solution properties including dielectric constant, polarity, conductivity and surface tension on multi-jet electrospinning when using an auxiliary electrode. this time it was found out that solutions with higher dielectric constant (32.2https://dx.doi.org/10.5599/admet.844 luis castillo et al admet & dmpk 8(4) (2020) 325-353 330 78.6) and larger surface tension (31.8-41.29 mn/m) are most likely to produce 2-6 jets with short stable length (1.7-6.9 mm) under the influence of low voltage (5.03-7.13 kv) [81]. side by side electrospinning sometimes it is necessary to combine blends of polymers using one solvent or different polymer solutions for nanofiber production through electrospinning. however, it is not always possible to manufacture nanofibers this way, since both polymers are required to be thermodynamically miscible when dissolving them in the same solvent. in addition, for polymer solution mixtures, polymer-polymer, polymersolvent and solvent-solvent interactions must be analyzed [82,83]. thus, side by side electrospinning introduces itself as a means for overcoming these difficulties. this method focuses on controlling viscosity and conductivity of two polymer solutions that do not get into physical contact until they reach the tip of the spinneret, where they are electrospun simultaneously. therefore, the resulting nanofiber consists of a bicomponent system with good mechanical strength and great thermal stability. in addition to that, it has different properties on both sides, based on the corresponding component placed on each [84,85]. yu et al. have employed this technique for the fabrication of janus fibers using a teflon-coated parallel spinneret, which allowed the formation of a janus taylor cone for obtaining high quality integrated janus structures. the electrospun nanofibers had one side composed of pvp k60 and ketoprofen, while the other consisted of ethyl cellulose (ec) and the drug as well. the mentioned biphasic design provided a fast dissolution on pvp k60 side and a sustained release of the remaining drug on the ec side [86]. later, they produced high quality pvp k60/shellac janus nanofibers using this electrospinning method, but with the modification of using a structured spinneret comprising two eccentric needles nested into a third one [87]. emulsion electrospinning proposed by xu et al., it has been developing recently as an alternative technique of electrospinning [88]. this novel method allows processing emulsions to encapsulate either hydrophilic or hydrophobic apis for the manufacturing of core-shell nanofibers, using a single spinneret. the process has been found to enhance encapsulation in the core and to provide better drug stability and bioavailability relevant for the development of advanced drug delivery systems. the release mechanism is mainly controlled by diffusion and enzymatic degradation of the solidified polymers [89-91]. for this purpose, either water-in-oil or oil-in-water emulsions can be electrospun [89]. during the process, the solvent molecules closer to the polymeric jet surface evaporate at a higher rate, which causes a viscosity increase in the outer layer. then, emulsion droplets are subjected to the electrical field that induces them to condense and stretch into elliptical shapes, but also several other forces have an influence on the charged jet, like coulomb’s and electrostatic forces, which provoke droplet expansion. on the other hand, viscoelastic forces and surface tension control droplet contraction to reduce the interface between the air and the polymeric jet. finally, when the jet reaches the collector, the solvent has almost completely evaporated [92,93]. hu et al. investigated the influence of span 80, sodium dodecyl sulfate (sds) and poly(ethylene oxide)poly(propylene oxide)-poly(ethylene oxide) triblock copolymer towards the morphological properties of an electrospun poly (ɛ-caprolactone) /bovine serum albumin nanofibers. it was found that a 0.4 % (w/v) sds emulsion produced the most uniform electrospun nanofibers through emulsion electrospinning [94]. in another research, moydeen et al. employed the technique for the production of core-shell pva/dextran sulfate nanofibers loaded with ciprofloxacin, which demonstrated to improve the sustained release of the drug compared to the nanofibers obtained by co-electrospinning [95]. admet & dmpk 8(4) (2020) 325-353 electrospun nanofibers: a nanotechnological approach doi: https://dx.doi.org/10.5599/admet.844 331 melt electrospinning this technique is considered one of the most relevant variations of the conventional electrospinning method because instead of using a polymer solution it employs molten polymers for achieving nanofiber production [96,97]. although molten polymer viscosity will tend to be greater compared to a polymer solution, it is possible to use some other substances as auxiliaries like plasticizers in order to reduce this parameter. when the jet arrives at the collector the solidification process will not be caused by drying or evaporation of a solvent but it will be produced by molten polymer cooling [98,99]. melt electrospinning stands out for being considered as a green technology in nanofibers manufacturing, which allows exploring new approaches regarding biomedical and bioengineering applications without the restrictions of using solvents, and related issues like incomplete elimination and their toxicity [100,101]. in addition to that, it is a promising method for industrial production and commercialization, since it not only increases the production rate of nanofibers but also gets rid of manufacturing costs represented by the use of expensive solvents [102,103]. larrondo and manley were the pioneers of this method three decades ago when they tried to produce nanofibers from molten polypropylene (pp) but ended obtaining fibers with diameters greater than 50 µm due to high melt viscosity of the polymer [104-106]. despite that, the technique hadn’t been exploited to a great extent until weimin et al. proposed in 2014 a variant of this method called ‘polymer melt differential electrospinning’ (pmdes) [107]. in pmdes scale-up, it was found out that producing multiple jets from one umbrella-shaped spinneret resulted in the obtention of nanofibers with an average diameter of 300 nm, with an efficiency 500 – 1000 times higher than solution electrospinning, and about 80 times higher compared to conventional melt electrospinning [107]. however, even the basic method of this technique is still under study, thus process parameters such as melt viscosity and processing temperatures still need special attention and evaluation [108]. polymers and solvents among the most frequently used materials in nanotechnology, polymers are actively employed in this field for the production of nanofibers by electrospinning. a wide variety of them, from natural and synthetic origins, have been electrospun under different required manufacturing conditions. one can also use polymer blends to obtain nanofibers with different structures such as core-shell, ribbon-like, porous, and aligned. therefore, polymer properties and functionalities provide nanofibers with desirable characteristics and performance to fulfill the general demands for specialized applications [109,110]. properties such as hydrophilic and hydrophobic nature of the raw material need to be addressed since they have a major effect on drug release from polymeric nanofiber matrix. hydrophobic polymers provide limited diffusion of aqueous solutions into the matrix due to their restricted contact angle allowing for longterm release, while hydrophilic polymers cause the diffusion of the api to the medium by their swelling mechanism and their dissolution [111]. furthermore, polymer crystallinity plays an important role, since amorphous regions provide better access to water molecules compared to crystalline regions. thus, highly crystalline polymers exhibit a slower release of drugs [112]. the most popular synthetic polymers that have been electrospun are hydroxypropylmethyl cellulose (hpmc), peo, pvp, sodium carboxymethyl cellulose (nacmc), pva, alginates, dextran, pcl, poly(lactic acid) (pla), poly(glycolic acid) (pga), poly(lactic acid-co-glycolic acid) (plga), and poly(lactide-co-ε-caprolactone) (plcl). regarding the natural raw materials, the use of chitosan, gelatin, collagen, chitin, fibrinogen, and silk https://dx.doi.org/10.5599/admet.844 luis castillo et al admet & dmpk 8(4) (2020) 325-353 332 fibroin have been reported [113,114]. the aforementioned polymers are extensively used in biomedical and bioengineering due to their advantages in manufacturing such as stable mechanical properties, good biocompatibility for some applications such as scaffolds, and biodegradable behavior for drug release [115]. despite the exceptional properties of polymers and the outstanding advances in nanotechnological manufacturing, not all of them are feasible for use in electrospinning. moreover, it is required a solvent with the capacity of dissolving both, drug and the polymer. solubility parameters can be useful for the selection of the most adequate solvent. therefore, non-soluble polymers like polypropylene (pp), polyethylene (pe), and poly(phenylene sulfide) (pps) can’t be electrospun satisfactorily [116,117]. polymer solubility is complex due to its molecular weight, size difference compared to the solvent, viscosity of the system and changes in terms of structure. the dissolution process is considered to occur in two stages. first, solvent molecules diffuse slowly through the polymer to produce a swollen gel, where polymer-solvent interactions are expected to break intermolecular polymeric bonds. then, when all polymer-polymer bonds are broken, a true solution is formed [118,119]. in addition to that, polymer solubility in a system can be explained by gibbs free energy (δg) [120]: ∆𝐺 = ∆𝐻 − 𝑇∆𝑆, (1) where δh is the enthalpy of the process, t stands for temperature and δs is the entropy. an adequate solvent is able to make polymer molecules to expand, while a poor one causes them to collapse [120]. also, the process temperature influences the suitability of a solvent. temperatures above flory-huggins temperature (i.e. the lowest temperature at which a polymer of infinite molecular weight is completely soluble in a determined solvent) induce polymer’s molecules to expand, which reduces gibbs free energy of the system until it reaches a negative value where the polymer is soluble [121-123]. however, in order to evaluate a solvent for a determined drug-polymer system, researchers frequently employ the triangular diagram or solubility map as the main tool for the estimation, as shown in figure 3. also, solvents can be classified according to their ability to dissolve the api and the polymer based on dielectric constant values within the range of interest [124,125]. although a polymer is soluble in a determined solvent, it does not always imply its aptitude to be electrospun for obtaining nanofibers [126]. since different solvents display different levels of electrospinnability, choosing one entails important features to be taken into consideration besides basic parameters like dielectric constant and relative permittivity. in the first place, there must be a balance in the evaporation rate, so that it is enough to allow the fiber to maintain its integrity when arriving at the collector, but not too fast for causing hardening before it reaches the nanometer range. furthermore, solvent viscosity and surface tension must not prevent the jet formation or allow the polymer solution to drawn freely from the spinneret [127,128]. a polymer solution constituted by multiple solvents may present manufacturing problems, since solvents with different solubility properties can affect several factors from the polymer and the nanofibers. polymer chain conformation, viscoelasticity, and critical minimum concentration can be influenced by the selection of the solvent [127]. on the other hand, nanofiber diameter, tensile strength, morphology and crystallinity of the loaded drug can be dramatically modified by the solvent properties [129]. kathsee et al. worked with a biodegradable polymer blend composed of pla and butylene adipate‐co‐ terephthalate (pbat) for the manufacturing of electrospun nanofibers. in this opportunity, they explored the effects of solution parameters such as types of binary solvents, solvent mixing ratio, and polymer blend concentration. they performed solubility tests of the pla/pbat blend for the selection of a suitable binary solvent system that guarantees electrospinnability, resulting in the use of a 3:2 dichloromethane (dcm)/ admet & dmpk 8(4) (2020) 325-353 electrospun nanofibers: a nanotechnological approach doi: https://dx.doi.org/10.5599/admet.844 333 dimethylformamide (dmf) mixture [130]. although the benefits of electrospinning encourage more research in this field, the non-toxic profile required for polymers and solvents has been difficult to achieve. this has represented a barrier for fast development and industrialization of the technique since more than 90 % of the electrospun polymer solutions reported in the literature at present contain toxic solvents [131]. however, current trends in research have focused on developing a new generation of electrospun nanofibers with fewer toxicity problems for safe and efficacious therapeutic applications [132]. figure 3. triangular solubility diagram for a drug-polymer system in solvent mixtures. adapted with permission from m. knopp, et al. comparative study of different methods for the prediction of drug–polymer solubility. molecular pharmaceutics 12(9): 3408–3419. copyright (2015) american chemical society [124]. drug development through electrospun nanofibers technological issues for the formulation of poorly water-soluble drugs currently, most of the drug products available do not address the growing demand for personalized therapies. many of them face several formulation challenges due to the physicochemical properties of the apis, which may prevent the development of a suitable pharmaceutical form that exerts an adequate therapeutic effect [133-135]. however, the rapid growth of nanotechnology intends to contribute greatly to medical and pharmaceutical sciences, providing innovative treatments that can improve patients’ lives. thus, an increasing number of scientists, research groups, industries, and governments are investing millions of dollars for nanotechnological development [136]. it is well known that drug molecules have to be released from the matrix of the pharmaceutical dosage form and then, dissolve in order to be absorbed and to cause the intended therapeutic effect. nevertheless, both processes are greatly dependent on the physicochemical properties of the api, and the ones from the excipients as well [137]. consequently, göke et al. state that poor water solubility is one of the major pharmaceutical challenges for drug development [138]. amidon et al. classified drug molecules as a function of their aqueous solubility and intestinal permeability in the biopharmaceutics classification system (bcs) [139]. classes ii and iv in the bcs have poor solubility due to their complete dose not being possible to dissolve when ingested with a glass of water (250 ml) [140-142]. a great majority of new drug molecules and lead compounds are within class ii and iv, and thus present limited oral absorption, slow in vivo dissolution, and low bioavailability [141,143-145]. therefore, due to the recently increasing number of poorly water-soluble drugs, many methods have https://dx.doi.org/10.5599/admet.844 luis castillo et al admet & dmpk 8(4) (2020) 325-353 334 been employed for improving dissolution [146]. electrospinning produces amorphous solid dispersions of this kind of drugs in a water-soluble matrix mainly composed of polymers than can enhance their dissolution, where that improvement is explained by the higher solubility exhibited by non-crystalline forms. in addition to that, carriers are mostly hydrophilic, and depending on the preparation method, some can cause a greater increase in the surface area than others [147,148]. nevertheless, drug-polymer systems are not exempt from presenting incompatibilities, not only chemical but also physical, such as a low degree of miscibility that may lead to an unstable amorphous phase [149]. on the other hand, the presence of polymorphism within the api can make harder the conversion to an amorphous structure, representing a disadvantage for the process [150]. according to poller et al., a major challenge for electrospinning is the scale-up process and this issue is currently being addressed by both academic and industrial researchers [151-153]. drug loading synthetic, natural and biologically active substances can be loaded into electrospun nanofibers and it has been widely reported by many research papers. electrospinning provides nanofibers with a large surface area to volume ratio, which combined with the possibility of choosing an adequate solvent for solubilizing the drug of interest, makes it a technique with high loading capacity. in addition to that, the method is considered as an economic, suitable, stable and improved medium for drug loading that provides control over drug release kinetics [154-156]. drug loading into electrospun nanofibers can be performed in different ways based on the techniques previously explained. in the first place, the drug can be dissolved directly with the polymer by the same solvent or it may require to be previously dissolved in a small amount of another one in order to be added to the polymer solution, wherein both cases the drug will end embedded in the fabricated nanofiber through co-electrospinning [154,155]. a comparative study of curcumin-loaded pcl nanofibers by melt and solution electrospinning was carried out by lian and meng. they reported that the active substance did not alter the morphology of melt electrospun nanofibers, which conserved a high crystallinity and could be loaded with a large amount of curcumin in the amorphous state. however, it did cause aggregates in solution electrospinning due to its limited solubility in the solvent system composed of 3:1 dcm/ethanol, which caused jet instability during the process [157]. in another approach made by zhu et al., coaxial electrospinning was employed to load flurbiprofen for obtaining electrospun pvp/plga core-shell nanofibers [158]. also, for drugs that are not soluble in the same solvent as the polymer, emulsion electrospinning can directly encapsulate lipophilic or hydrophilic drugs within the polymeric matrix. electrospinning also offers the possibility to load more than one api into the nanofiber thanks to the multijet variant, which is advantageous when it is needed to load lipophilic and hydrophilic drugs in the same three-dimensional structure. moreover, according to the surface adsorption method, active substances can be incorporated after the electrospinning process by immersing nanofibers in a drug solution as well [159161]. dissolution optimization nanotechnology has made the use of nanoparticles and nanostructures a trend for increasing the solubility of apis, which is known to be a limiting step for bioavailability. because of that, the dissolution test has extended its application not only to solid pharmaceutical dosage forms but to novel drug delivery systems such as electrospun nanofibers. it is well known that solubility is greatly dominated by the nature of the substance functional groups and their interactions with those of the solvent [162]. electrospinning has allowed the production of nanofibers with poorly water-soluble drugs embedded in hydrophilic admet & dmpk 8(4) (2020) 325-353 electrospun nanofibers: a nanotechnological approach doi: https://dx.doi.org/10.5599/admet.844 335 polymers, which enhances their dissolution rate [163,164]. however, it is clearly stated in the literature that the enhanced solubility and dissolution of drugs exhibited by electrospun nanofibers is mainly due to the presence of the compound in the amorphous state. this can be explained by the fact that many apis have higher kinetic energy in that arrangement than in the crystalline form [165]. aside from that, a more homogenous distribution of the drug, increased wettability, lower precipitation, highly porous and specific surface area from the nanostructure also contributes to the optimization of the dissolution rate compared to the commercial pharmaceutical product or the conventional solid dispersions. these improved properties are leading to the design and development of novel formulations including buccal, transdermal and topical dosage forms [166-168]. nazari et al. worked with solution electrospinning to fabricate different indomethacin buccal films made of ethocel, hpmc and tween 80. according to the differential scanning calorimetry (dsc) and x-ray diffraction analysis (xrd), the drug was present in the amorphous state. dissolution performance was evaluated in vitro in a medium of a buffer at ph 6.8, in which it was found that the presence of tween 80 accompanied with 5 % (w/w) of hpmc caused a fast release in the first five minutes, which increased the dissolution rate of the drug in about 62 % compared to the nanofiber produced with the water-insoluble polymer ethocel [169]. also, szabó et al. produced terbinafine hydrochloride buccal films using pva and chitosan for the electrospinning process, where the nanofibers showed fast and complete dissolution of the drug [170]. in another approach, adeli prepared several formulations of pvp k90 electrospun nanofibers loaded with irbesartan. all the evaluated prototypes showed an improved dissolution rate, but the one that exhibited the best performance had a 3:7 drug/polymer ratio. in sink conditions, almost 97 % of the drug was released in 60 minutes. aside to that, saturation solubility increased up to 245.32 ± 1.77 μg/ml, about six times that of the pure drug (40.55 ± 1.01 μg/ml) due to the electrospun nanofibers-based solid dispersion preparation [171]. drug delivery and drug release drug delivery systems are designed for improving significantly the therapeutic efficacy and safety of a drug through controlling the targeted-site of action and the release rate. the development of this type of system involves the understanding of drug release kinetics in order to select the most appropriate compounds for the intended purpose. therefore, controlled release is useful for modifying delivery kinetics, reducing toxicity and side effects, as well as for adjusting therapies according to patients’ convenience [172-174]. electrospun nanofibers have been considered as suitable structures that can control spatially and temporally drug delivery of more than one drug. the first aspect can be achieved by placing the nanofiber at the targeted site by invasive or non-invasive methods to avoid systematic exposure [175-177]. ravikumar et al. developed a tetrahydro curcumin transdermal patch through electrospinning solution, where the poorly water-soluble active was loaded at a concentration of 8.7 % (w/w of dry polymers) into a 2:1 mix of solutions of pcl 10 % (w/v) and peg 5 % (w/v) in a chloroform and acetone solvent system, obtaining nanofibers with an average diameter of 400 ± 20 nm. the cumulative % release was 95.11 % at 24 hours, exhibiting first-order kinetics that matches higuchi’s diffusion model for which this development might be feasible for once in a day transdermal drug delivery [178]. the obtained nanofibers offer different outstanding properties for hydrophobic active substances, which are explained by their nanostructured nature that improves not only physicochemical and pharmacokinetic properties but also provides protection from enzymatic or chemical degradation. this nanostructure overcomes high drug uptake limitations, exhibits efficient drug transport and has the capacity of encapsulating chemical or biological therapeutic agents during the manufacturing process [179,180]. https://dx.doi.org/10.5599/admet.844 luis castillo et al admet & dmpk 8(4) (2020) 325-353 336 paaver et al. successfully created supersaturated electrospun nanofibers for controlled release of piroxicam made of hpmc and 1,1,1,3,3,3-hexa-fluoro-2-propanol (hfip) as the solvent. the dry diameter of the nanofibers was in the range of 400-600 nm based on scanning electron microscopy studies (sem). solidstate studies of the nanofibers carried by dsc and xrd confirmed the presence of the drug in the amorphous state, but after three months they showed a slow tendency to recrystallize into form iii. physicochemical properties of the nanofibers were strongly dependent on the concentration of hpmc, exhibited short lag-time, no initial burst release and zero-order dissolution kinetics [181]. electrospun nanofibers are able to achieve controlled release at the targeted site by making use of several polymers, which can be biodegradable or non-degradable. according to that, drug release from non-degradable polymers is performed by diffusion, while release from degradable polymers occurs mainly by matrix erosion. however, the selection of the polymeric matrix is conditioned by the requirements of the application [182,183]. when using natural or biodegradable polymers it is necessary to crosslink the matrices for improving the mechanical, thermochemical and structural integrity of electrospun nanofibers, which also enhances sustained and controlled released [184]. drug release from hydrophilic polymers is possible through surface desorption, diffusion and matrix erosion due to the degradation of the nanofiber caused by enzymes. when nanofibers are constituted by this type of polymers, direct contact with an aqueous solution alters its glassy state and forms a rubbery gel, causing swelling of the polymeric matrix, which transports a drug through its network until it reaches the medium [185,186]. for high-swelling polymers, drug release depends on the diffusion through the pores of the nanofibers, while low-swelling polymer release rate is conditioned by the swelling process itself [187,188]. gordon et al. developed research for the delivery of celecoxib lipophilic nanoparticles, which were formed by loading the poorly water-soluble drug dissolved in a volatile oil-solvent system into the electrospinning aqueous polymer solution composed of a high molecular weight pva. the obtained hydrophilic electrospun nanofibers enabled fast dissolution and release of celecoxib nanoparticles, since the drug was present in the amorphous state and the average size of the nanoparticles was in the range of 21-93 nm [189]. polymer, solvent and drug compatibility need to be addressed for achieving reproducible drug release and thus, drug delivery. when the electrospun nanofiber system carries multiple drugs, their affinity to the polymer as well as the interaction between them have to be considered. if any physicochemical interaction is present it may lead to problems regarding the diffusion of the drug, which will result in different release profiles [190-192]. zhao et al. designed an implantable tissue-engineered scaffold through electrospinning made of a self-coated interfacial layer developed between an inorganic and an organic matrix for timeprogrammed multi-drug release. the scaffold’s design propitiated the absence of interactions; the insidelocated drug was ibuprofen, which experienced short-term release for 30 days, while doxorubicin was loaded outside the interfacial layer and showed a sustained long-term release during 90 days [193]. also, as stated by haider et al., the chosen electrospinning method and the process itself allow controlling the release kinetics through some critical parameters such as matrix properties, nanofiber diameter, porosity and of course their morphology. however, if the different drugs are loaded by the same electrospinning method, the release will depend on each diffusion coefficient [194]. thus, different techniques are associated with different rates of release. when drugs are loaded through physical or surface adsorption into the nanofibers this will exhibit short diffusion times, while methods like coaxial electrospinning provide controlled and sustained profiles [195]. admet & dmpk 8(4) (2020) 325-353 electrospun nanofibers: a nanotechnological approach doi: https://dx.doi.org/10.5599/admet.844 337 however, recent research is focusing on developing electrospun nanofibers that are responsive to certain stimuli or feedback factors that start the controlled release of the drug at the targeted site. these novel drug delivery systems are called smart electrospun nanofibers [196,197]. smart electrospun nanofibers although these systems have experienced considerable progress in the past years, all of them face the same challenges regarding the manufacturing process, quality control and cytotoxicity, which imply a great barrier to faster development [196]. however, it seems that electrospinning can offer a solution to many of the common problems through the selection of a suitable carrier based on the therapeutic goal, material safety profile, drug physicochemical properties, and the route of administration. as presented on table 1, several types of electrospun smart nanofibers have been developed that respond to certain stimuli such as ph, temperature, magnetic field and electric influence. in addition to that, some of them are responsive to multiple stimuli that cause a physical or chemical change in its arrangement [198,199]. table 1. electrospun smart nanofibers applications. smart system polymers drug application reference ph-responsive eudragit s100, lecithin diclofenac sodium oral-colon targeted drug delivery. [200] eudragit l100 alkaline phosphatase peroral delivery of proteins. [201] eudragit s, eudragit rs indomethacin colonic drug delivery. [202] thermoresponsive poly(nvinylcaprolactam-comethacrylic acid) ketoprofen extended drug release. [203] poly(di(ethylene glycol) methyl ether methacrylate), ethyl cellulose ketoprofen extended drug release. [204] magnetic fieldresponsive poly (d,l-lactide-coglycolide) bortezomib implantable device for endoscopic hyperthermia treatment and tumortriggered controlled drug release. [205] poly(ε-caprolactone) ketoconazole slow sustained release for fungal infections. [206] electrical fieldresponsive graphene, poly(εcaprolactone), gelatin tetracycline hydrochloride neural tissue engineering and drug delivery. [207] poly(ε-caprolactone) curcumin delivery through programmable electrical devices. [208] multi stimuliresponsive poly(nvinylcaprolactam), ethyl cellulose, eudragit l100 ketoprofen dual temperature and phresponsive delivery in lower intestinal tract. [209] poly-nisopropylacrylamide, chitosan curcumin dual temperature and phresponsive for cancer targeting. [210] ph-responsive: potential targeted sites such as tissues and organs differ in their physiological ph value, which implies a lot of interest for research in this stimulus for triggering and modulating drug release. ideally, ph-responsive electrospun nanofibers should be designed and developed in a way that makes them able to release the drug at a certain ph close to that caused by a condition or a disease in the biological surroundings [211,212]. nanofiber drug release is expected to be reduced or completely stopped whenever the condition is https://dx.doi.org/10.5599/admet.844 luis castillo et al admet & dmpk 8(4) (2020) 325-353 338 improved and the microenvironment ph shifts back to the physiological value. most polymers used for this purpose are biocompatible and biodegradable containing carboxylic acids or amine groups, which undergo protonation and deprotonation as well as changes in size, shape, and hydrophobicity causing the drug release mostly through a swelling mechanism [213]. although ph-responsive electrospun nanofibers have been found useful for many drugs, they are not suitable for protein delivery as ph variations can cause denaturation of these molecules [214]. thermo-responsive: these systems are composed of polymers that suffer solubility modifications based on the lower critical solution temperature (lcst) (i.e. the temperature at which there is a balance in the competition established by hydrophilic and hydrophobic areas of the polymer chain structure) that is usually close to normothermia (37 °c). thus, below lcst the polymer is completely miscible, but above that parameter, it is only partially due to a reversible phase transition from a hydrophilic form to a dehydrated or hydrophobic state. the aforementioned transition is an entropy-driven process caused by the release of water molecules upon heating [215]. the polymers used for the development of thermo-responsive smart electrospun nanofibers swell below their lcst, which reduces drug release at normothermia because they retain the loaded hydrophobic drug. on the other hand, a small increase in temperature (e.g. a locally heated tumor) enhances the release because of nanofibers shrinking, caused by the polymer conversion into a hydrophobic form [216-218]. magnetic field-responsive: magnetic field stimulus provides advantages compared to the other options since it is well known that it does not affect tissues and its penetration in the body is greater than heat or light, causing neither cell death nor dna damage [219,220]. the magnetic properties allow pharmaceutical nanoparticles to be delivered to a specific site due to the influence of a magnetic field [221]. for the development of these delivery systems, it is necessary to incorporate superparamagnetic iron oxide nanoparticles (spions) with high biocompatibility and low cytotoxicity like fe3o4 (magnetite) or fe2o3 (maghemite) into the polymers or polymer solution used during the electrospinning process [222,223]. magnetite and maghemite display superparamagnetism when the diameter of the nanoparticles is smaller than 20 nm, which results in a non-continuing magnetic interaction upon the removal of an external magnetic field [221]. a common developing strategy involves coating the surface of spions with polymers or encapsulating them in a biodegradable matrix to form a nanocomposite known as magnetic responsive polymer composites (mrpcs) [224]. the formation of mrpcs improves not only spions solubility and biocompatibility but also their surface properties for in vivo administration, conferring all of the superparamagnetic related properties to the entire nanocomposite [224]. magnetically active polymers react to an external magnetic field and exhibit rheological changes or experience mechanical stress. polymer’s magneto-elastic properties can be used to control deformations such as stretching, contraction, and also the movement of mrpcs [225,226]. electric field-responsive: electrospun nanofibers constituted by electric-responsive polymers experience changes in their volume and swelling mechanism when exposed to an electric field. electrospinning of iondoped conducting polymers and polymer composites or coatings has been applied for manufacturing nanofibers intended for electric field-responsive drug delivery system [227-229]. multi stimuli-responsive: this kind of system represents a new opportunity for electrospun nanofibers with enhanced performance under the influence of two or more signals. thus, in the case of multiple drugloaded nanofibers, the release of each drug can respond to a different stimulus, according to the respective polymeric matrix and the manufacturing electrospinning technique. in addition to that, drug release can be triggered at the same time or sequentially, due to the influence of combining multiple stimuli [230,231]. admet & dmpk 8(4) (2020) 325-353 electrospun nanofibers: a nanotechnological approach doi: https://dx.doi.org/10.5599/admet.844 339 concluding remarks nanotechnological development through electrospinning is giving the world the opportunity of overcoming the obstacles presented during the design, formulation, and manufacturing of pharmaceutical, biomedical and bioengineering applications. currently, electrospun nanofibers are at the forefront of nanotechnology, providing the ability to control and manipulate the properties of the obtained nanostructures, which exhibit improved in vitro dissolution performance compared to the conventional solid dispersion systems. their most important property is related to their large surface area to volume ratio, which is responsible for the optimized performance in combination with particle size reduction and amorphization of the api. therefore, electrospinning as a formulation strategy that makes use of electrostatic and mechanical forces to spin fibers, is being positioned as a novel and useful method for local and systemic drug delivery of poorly water-soluble drugs. although the potential benefits from this technique are known, there is still work to do for overcoming safety and scale-up challenges. acknowledgements: this study represents a collaborative work between lanotec and the faculty of pharmacy of the university of costa rica. the corresponding author would like to thank lanotec for the assessment and the opportunity for an internship. conflict of interest: no potential conflict of interest was reported by the authors. references [1] j. chen, t. zhang, w. hua, p. li, x. wang. 3d porous poly (lactic acid)/regenerated cellulose composite scaffolds based on electrospun nanofibers for biomineralization. colloids and surfaces a: physicochemical and engineering aspects 585 (2020). https://doi.org/10.1016/j.colsurfa.2019.124048. [2] r. ambrus, a. alshweiat, i. csóka, g. ovari, a. esmail, n. radacsi. 3d-printed electrospinning setup for the preparation of loratadine nanofibers with enhanced physicochemical properties. international journal of pharmaceutics 567 (2019) 1-8. https://doi.org/10.1016/j.ijpharm.2019.118455. [3] h. li, y. tan, l. yang, l. gao, t. wang, x. yang, d. quan. dissolution evaluation in vitro and bioavailability in vivo of self-microemulsifying drug delivery systems for ph-sensitive drug loratadine. journal of microencapsulation 32(2) (2015) 175–80. https://doi.org/10.3109/02652048.2014.985340. [4] n. schaub, c. johnson, b. cooper, r. gilbert. electrospun fibers for spinal cord injury research and regeneration. journal of neurotrauma 33(15) (2016) 1405–1415. https://doi.org/10.1089/neu.2015.4165. [5] n. schaub. electrospun fibers: a guiding scaffold for research and regeneration of the spinal cord. neural regeneration research 11(11) (2016) 1764–1765. https://doi.org/10.4103/16735374.194719. [6] d. sundaramurthi, u. krishnan, s. sethuraman. electrospun nanofibers as scaffolds for skin tissue engineering. polymer reviews. 54(2) (2014) 348–376. https://doi.org/10.1080/15583724.2014.881374. [7] s. miguel, m. ribeiro, p. coutinho, i. correia. electrospun polycaprolactone/aloe vera_chitosan nanofibrous asymmetric membranes aimed for wound healing applications. polymers 9(5) (2017) 183. https://doi.org/10.3390/polym9050183. https://dx.doi.org/10.5599/admet.844 https://www.sciencedirect.com/science/journal/09277757 https://www.sciencedirect.com/science/journal/09277757 https://doi.org/10.1016/j.colsurfa.2019.124048 https://doi.org/10.1016/j.colsurfa.2019.124048 https://doi.org/10.1016/j.ijpharm.2019.%1f118455 https://doi.org/10.3109/02652048.%1f2014.985340 https://doi.org/10.1089/neu.%1f2015.4165 https://doi.org/10.1089/neu.%1f2015.4165 https://doi.org/10.4103/1673-5374.194719 https://doi.org/10.4103/1673-5374.194719 https://doi.org/10.1080/15583724.2014.%1f881374 https://doi.org/10.1080/15583724.2014.%1f881374 https://doi.org/10.3390/polym9050183 luis castillo et al admet & dmpk 8(4) (2020) 325-353 340 [8] s. chou, d. carson, k. woodrow. current strategies for sustaining drug release from electrospun nanofibers. journal of controlled release 220 (2015) 584–591. https://doi.org/10.1016/j.jconrel.2015.09.008. [9] w. woon, q. sun. electrostatic charged nanofiber filter for filtering airborne novel coronavirus (covid-19) and nano-aerosols. separation and purification technology (2020) 116886. https://doi.org/10.1016/j.seppur.2020.116886. [10] y. ding, w. li, f. zhang, z. liu, n. ezazi, d. liu. electrospun fibrous architectures for drug delivery, tissue engineering and cancer therapy. advanced functional materials 29(2) (2019) 1802852. https://doi.org/10.1002/adfm.201802852. [11] s. miguel, d. simões, a. moreira, r. sequeira, i. correia. production and characterization of electrospun silk fibroin based asymmetric membranes for wound dressing applications. international journal of biological macromolecules 121 (2019) 524–535. https://doi.org/10.1016/j.ijbiomac.2018.10.041. [12] m.a. hassan, b.y. yeom, a. wilkie, b. pourdeyhimi, s.a. fabrication of nanofiber meltblown membranes and their filtration properties. journal of membrane science 427 (2013) 336–44. https://doi.org/10.1016/j.memsci.2012.09.050. [13] m. rolandi, r. rolandi. self-assembled chitin nanofibers and applications. advances in colloid and interface science 207 (2014) 216–222. https://doi.org/10.1016/j.cis.2014.01.019. [14] h.s. liao, j. lin, y. liu, p. huang, a. jin, x. chen. self-assembly mechanisms of nanofibers from peptide amphiphiles in solution and on substrate surfaces. nanoscale 8(31) (2016) 14814–14820. https://doi.org/10.1039/c6nr04672j. [15] k. katsogiannis, g. vladisavljević, s. georgiadou. porous electrospun polycaprolactone (pcl) fibres by phase separation. european polymer journal 69 (2015) 284–295. https://doi.org/10.1016/j.eurpolymj.2015.01.028. [16] z. zhang, w. tu, t. peijis, c. bastiaansen. fabrication and properties of poly(tetrafluoroethylene) nanofibres via sea-island spinning. polymer 109 (2017) 321–331. https://doi.org/10.1016/j.polymer.2016.12.060. [17] m. hammami, m. krifa, o. harzallah. centrifugal force spinning of pa6 nanofibers – processability and morphology of solution-spun fibers. the journal of the textile institute 105(6) (2014) 637–647. https://doi.org/10.1080/00405000.2013.842680. [18] l. yu, z. shao, l. xu, m. wang. high throughput preparation of aligned nanofibers using an improved bubble-electrospinning. polymers 9(12) (2017) 1-9. https://doi.org/10.3390/polym9120658. [19] g. rüzgar, m. birer, s. tort, f. acartürk. studies on improvement of water-solubility of curcumin with electrospun nanofibers. fabad journal of pharmaceutical sciences 38(3) (2016) 143-149. [20] o. kaljević, j. đuriš, b. čalija, z. lavrić, j. kristl, s. ibrić. application of miscibility analysis and determination of soluplus solubility map for development of carvedilol-loaded nanofibers. international journal of pharmaceutics 533(2) (2017) 445–454. https://doi.org/10.1016/j.ijpharm.2017.05.017. [21] t. potrč, s. baumgartner, r. roškar, o. planinšek, z. lavrič, j. kristl. electrospun polycaprolactone nanofibers as a potential oromucosal delivery system for poorly water-soluble drugs. european journal of pharmaceutical sciences 75 (2015) 101–113. https://doi.org/10.1016/j.ejps.2015.04.004. [22] x. wang, d. yu, x. li, s. bligh, g. williams. electrospun medicated shellac nanofibers for colontargeted drug delivery. international journal of pharmaceutics 490(1-2) (2015) 384-390. https://doi.org/10.1016/j.ijpharm.2015.05.077. [23] n. schaub, j. corey. a method to rapidly analyze the simultaneous release of multiple pharmaceuticals from electrospun fibers. international journal of pharmaceutics 574 (2020). https://doi.org/10.1016/j.ijpharm.2019.118871. [24] n. radacsi, k. giapis, g. ovari, p. szabó, r. ambrus. electrospun nanofiber-based niflumic acid capsules with superior physicochemical properties. journal of pharmaceutical and biomedical analysis 166 (2019) 371–378. https://doi.org/10.1016/j.jpba.2019.01.037. https://doi.org/10.1016/j.jconrel.%1f2015.%1f09.008 https://doi.org/10.1016/j.jconrel.%1f2015.%1f09.008 https://doi.org/10.1002/adfm.201802852 https://doi.org/10.1016/j.ijbiomac.2018.10.041 https://doi.org/10.1016/j.ijbiomac.2018.10.041 https://doi.org/10.1016/j.memsci.2012.09.050 https://doi.org/10.1016/j.cis.2014.01.019 https://doi.org/10.1039/c6nr04672j https://doi.org/10.1016/j.polymer.2016.12.060 https://doi.org/10.1080/00405000.2013.842680 https://doi.org/10.3390/polym9120658 https://doi.org/10.1016/j.ijpharm.%1f2017.05.017 https://doi.org/10.1016/j.ijpharm.%1f2017.05.017 https://doi.org/10.1016/j.ejps.2015.04.004 https://doi.org/10.1016/j.ijpharm.2015.05.077 https://doi.org/10.1016/j.ijpharm.2019.118871 https://doi.org/10.1016/j.jpba.2019.01.037 admet & dmpk 8(4) (2020) 325-353 electrospun nanofibers: a nanotechnological approach doi: https://dx.doi.org/10.5599/admet.844 341 [25] a. balogh, b. farkas, g. verreck, j. mensch, e. borbás, b. nagy b. ac and dc electrospinning of hydroxypropylmethylcellulose with polyethylene oxides as secondary polymer for improved drug dissolution. international journal of pharmaceutics 505(1-2) (2016) 159–166. https://doi.org/10.1016/j.ijpharm.2016.03.024. [26] i. sebe, p. szabó, b. kállai, r. zelkó. incorporating small molecules or biologics into nanofibers for optimized drug release: a review. international journal of pharmaceutics. 494(1) (2015) 516–530. https://doi.org/10.1016/j.ijpharm.2015.08.054. [27] x. hu, s. liu, g. zhou, y. huang, z. xie, x. jing. electrospinning of polymeric nanofibers for drug delivery applications. journal of controlled release 185 (2014) 12–21. https://doi.org/10.1016/j.jconrel.2014.04.018. [28] c. mingjun, z. youchen, l. haoyi, l. xiangnan, d. yumei, m. bubakir. an example of industrialization of melt electrospinning: polymer melt differential electrospinning. advanced industrial and engineering polymer research 2(3) (2019) 110–115. https://doi.org/10.1016/j.aiepr.2019.06.002. [29] s. miguel, r. sequeira, a. moreira, c. cabral, a. mendonça, p. ferreira. an overview of electrospun membranes loaded with bioactive molecules for improving the wound healing process. european journal of pharmaceutics and biopharmaceutics 139 (2019) 1–22. https://doi.org/10.1016/j.ejpb.2019.03.010. [30] a. jahangiri, k. adibkia. applications of electrospinning/electrospraying in drug delivery. bioimpacts 6(1) (2016) 1–2. https://doi.org/10.15171/bi.2016.08. [31] v. pillay, c. dott, y. choonara, c. tyagi, l. tomar, p. kumar. a review of the effect of processing variables on the fabrication of electrospun nanofibers for drug delivery applications. journal of nanomaterials (2013) 1-22. https://doi.org/10.1155/2013/789289. [32] j.f. cooley. electrical method of dispersing fluids. us745276a (1903). [33] w.j. morton. method of dispersing fluids. us705691a (1902). [34] a. formhals. process and apparatus for preparing artificial threads. us1975504a (1934). [35] n. taniguchi n. on the basic concept of nanotechnology(part ii). proceedings of the international conference on production engineering (1974) 18–23. [36] f. yalcinkaya. a review on advanced nanofiber technology for membrane distillation. journal of engineered fibers and fabrics 14 (2019) 1-12. https://doi.org/10.1177/1558925018824901. [37] m. nasrollahzadeh, s. mohammad, m. sajjadi, z. issaabadi. an introduction to nanotechnology. interface science and technology 28 (2019) 1-27. https://doi.org/10.1016/b978-0-12-8135860.00001-8. [38] j. doshi, d. reneker. electrospinning process and applications of electrospun fibers. journal of electrostatics 35(2-3) (1995) 151–160. https://doi.org/10.1016/0304-3886(95)00041-8. [39] d. reneker, i. chun. nanometre diameter fibres of polymer, produced by electrospinning. nanotechnology 7(3) (1996) 216–223. [40] n. bock, t. dargaville, m. woodruff. electrospraying of polymers with therapeutic molecules: state of the art. progress in polymer science 37(11) (2012) 1510–1551. https://doi.org/10.1016/j.progpolymsci.2012.03.002. [41] d. nguyen, c. clasen, g. van den mooter. pharmaceutical applications of electrospraying. journal of pharmaceutical sciences 105(9) (2016) 2601–2620. https://doi.org/10.1016/j.xphs.2016.04.024. [42] c. angammana, s. jayaram. fundamentals of electrospinning and processing technologies. particulate science and technology 34(1) (2016) 72–82. https://doi.org/10.1080/02726351.2015.1043678. [43] f. yalcinkaya. preparation of various nanofiber layers using wire electrospinning system. arabian journal of chemistry 12(8) (2019) 5162–5172. https://doi.org/10.1016/j.arabjc.2016.12.012. [44] x. wang, t. lin, x. wang. scaling up the production rate of nanofibers by needleless electrospinning from multiple ring. fibers polymer 15(5) (2014) 961–965. https://doi.org/10.1007/s12221-014-0961x. https://dx.doi.org/10.5599/admet.844 https://doi.org/10.1016/j.ijpharm.2015.08.054 https://doi.org/10.1016/%1fj.jconrel.%1f2014.04.018 https://doi.org/10.1016/%1fj.jconrel.%1f2014.04.018 https://doi.org/10.1016/j.aiepr.2019.06.002 https://doi.org/10.1016/j.ejpb.2019.03.010 https://doi.org/10.1016/j.ejpb.2019.03.010 https://doi.org/10.15171/bi.2016.08 https://doi.org/10.1155/2013/789289 https://doi.org/10.1177/1558925018824901 https://doi.org/10.1016/b978-0-12-813586-0.00001-8 https://doi.org/10.1016/b978-0-12-813586-0.00001-8 https://doi.org/10.1016/0304-3886(95)00041-8 https://doi.org/10.1016/j.%1fprogpolymsci.%1f2012.03.002 https://doi.org/10.1016/j.%1fprogpolymsci.%1f2012.03.002 https://doi.org/10.1016/j.xphs.2016.04.024 https://doi.org/10.1080/02726351.2015.%1f1043678 https://doi.org/10.1080/02726351.2015.%1f1043678 https://doi.org/10.1016/j.arabjc.2016.12.012 https://doi.org/10.1007/s12221-014-0961-x https://doi.org/10.1007/s12221-014-0961-x luis castillo et al admet & dmpk 8(4) (2020) 325-353 342 [45] s. miguel, d. figueira, d. simões, m. ribeiro, p. coutinho, p. ferreira. electrospun polymeric nanofibres as wound dressings: a review. colloids and surfaces b: biointerfaces 169 (2018) 60–71. https://doi.org/10.1016/j.colsurfb.2018.05.011. [46] a. balaji, m. vellayappan, a. john, a. subramanian, s. jaganathan, e. supriyanto. an insight on electrospun-nanofibers-inspired modern drug delivery system in the treatment of deadly cancers. rsc advances 5(71) (2015) 57984–58004. https://doi.org/10.1039/c5ra07595e. [47] s. preethi, m. haritha, c. viji, n. selvamurugan. bone tissue engineering: scaffold preparation using chitosan and other biomaterials with different design and fabrication techniques. international journal of biological macromolecules 119 (2018) 1228–1239. https://doi.org/10.1016/j.ijbiomac.2018.08.056. [48] s. tajbakhsh, f. hajiali. a comprehensive study on the fabrication and properties of biocomposites of poly(lactic acid)/ceramics for bone tissue engineering. materials science and engineering: c 70 (2017) 897–912. https://doi.org/10.1016/j.msec.2016.09.008. [49] x. chen, y. zhang, x. he, h. li, b. wei, w. yang. electrospinning on a plucked string. journal of materials science 54(1) (2019) 901–910. https://doi.org/10.1007/s10853-018-2870-7. [50] j. lagaron, a. solouk, s. castro, y. echegoyen. 3 biomedical applications of electrospinning, innovations, and products. in: t. uyar, e. kny. electrospun materials for tissue engineering and biomedical applications. woodhead publishing, cambridge, uk, 2017 p. 57–72. https://doi.org/10.1016/b978-0-08-101022-8.00010-7. [51] y. li, t. bou. electrospinning in tissue engineering. in: a. haider, s. haider. electrospinning: material, techniques, and biomedical applications. intechopen, london, uk, 2016 p.117-139. https://doi.org/10.5772/65836. [52] g. taylor, m. van dyke. electrically driven jets. proceedings of the royal society of london a mathematical and physical sciences 313(1515) (1969) 453–475. https://doi.org/10.1098/rspa.1969.0205. [53] m. morad, a. rajabi, m. razavi, s. sereshkeh. a very stable high throughput taylor cone-jet in electrohydrodynamics. scientific reports 6(1) (2016) 1–10. https://doi.org/10.1038/srep38509. [54] k. zhao, w. wang, y. yang, k. wang, d. yu. from taylor cone to solid nanofiber in tri-axial electrospinning: size relationships. results in physics 15 (2019) 102770. https://doi.org/10.1016/j.rinp.2019.102770. [55] j. he. on the height of taylor cone in electrospinning. results in physics 17 (2020) 103096. https://doi.org/10.1016/j.rinp.2020.103096. [56] y. chiu, z. zhang, k. dziemidowicz, c. nikoletopoulos, u. angkawinitwong, j. chen. the effect of solvent vapor annealing on drug-loaded electrospun polymer fibers. pharmaceutics 12(2) (2020) 139. https://doi.org/10.3390/pharmaceutics12020139. [57] j. xue, j. xie j, w. liu, y. xia. electrospun nanofibers: new concepts, materials, and applications. accounts of chemical research 50(8) (2017) 1976–1987. https://doi.org/10.1021/acs.accounts.7b00218. [58] n. ismail, f. maksoud, n. ghaddar, k. ghali, a. tehrani. simplified modeling of the electrospinning process from the stable jet region to the unstable region for predicting the final nanofiber diameter. journal of applied polymer science 133(43) (2016) 44112. https://doi.org/10.1002/app.44112. [59] m. šimko, d. lukáš. mathematical modeling of a whipping instability of an electrically charged liquid jet. applied mathematical modelling 40(21) (2016) 9565–9583. https://doi.org/10.1016/j.apm.2016.06.018. [60] m. masuelli, d. renard. advances in physicochemical properties of biopolymers (part 2). bentham science publishers, 2017, p. 497. [61] z. tan, h. wang, x. gao, t. liu, y. tan. composite vascular grafts with high cell infiltration by coelectrospinning. materials science and engineering: c 67 (2016) 369–377. https://doi.org/10.1016/j.msec.2016.05.067. https://doi.org/10.1016/j.colsurfb.2018.05.011 https://doi.org/10.1039/c5ra07595e https://doi.org/10.1016/j.ijbiomac.%1f2018.08.056 https://doi.org/10.1016/j.ijbiomac.%1f2018.08.056 https://doi.org/10.1016/j.msec.2016.09.008 https://doi.org/10.1007/s10853-018-2870-7 https://doi.org/10.1016/b978-0-08-101022-8.00010-7 https://doi.org/10.5772/65836 https://doi.org/10.1098/rspa.%1f1969.0205 https://doi.org/10.1098/rspa.%1f1969.0205 https://doi.org/10.1038/srep38509 https://doi.org/10.1016/j.%1frinp.%1f2019.102770 https://doi.org/10.1016/j.%1frinp.%1f2019.102770 https://doi.org/10.1016/j.rinp.2020.103096 https://doi.org/10.3390/pharmaceutics12020139 https://doi.org/10.1021/acs.accounts.%1f7b00218 https://doi.org/10.1021/acs.accounts.%1f7b00218 https://doi.org/10.1002/app.44112 https://doi.org/10.1016/j.%1fapm.%1f2016.%1f06.018 https://doi.org/10.1016/j.%1fapm.%1f2016.%1f06.018 https://doi.org/%1f10.1016/j.%1fmsec.2016.05.067 https://doi.org/%1f10.1016/j.%1fmsec.2016.05.067 admet & dmpk 8(4) (2020) 325-353 electrospun nanofibers: a nanotechnological approach doi: https://dx.doi.org/10.5599/admet.844 343 [62] f. ahmed, s. saleemi, z. khatri, m. abro, i. kim. co-electrospun poly(ɛ-caprolactone)/cellulose nanofibers-fabrication and characterization. carbohydrate polymers 115 (2015) 388–393. https://doi.org/10.1016/j.carbpol.2014.09.002. [63] x. qin. 3 coaxial electrospinning of nanofibers. in: m. afshari. electrospun nanofibers. woodhead publishing, cambridge, uk, 2017, p. 41–71. https://doi.org/10.1016/b978-0-08-100907-9.00003-9. [64] w. yu, q. ma, x. li, x. dong, j. wang, g. liu. one-pot coaxial electrospinning fabrication and properties of magnetic-luminescent bifunctional flexible hollow nanofibers. materials letters 120 (2014) 126–129. https://doi.org/10.1016/j.matlet.2014.01.076. [65] h. zhang, x. qin, j. wu, y. he, h. du, b. li. electrospun core–shell silicon/carbon fibers with an internal honeycomb-like conductive carbon framework as an anode for lithium ion batteries. journal of materials chemistry a 3(13) (2015) 7112–7120. https://doi.org/10.1039/c4ta06044j. [66] h. jiang, l. wang, k. zhu. coaxial electrospinning for encapsulation and controlled release of fragile water-soluble bioactive agents. journal of controlled release 193 (2014) 296–303. https://doi.org/10.1016/j.jconrel.2014.04.025. [67] s. bandyopadhyay, s. ghosh, m. sain. 19 the use of biobased nanofibres in composites. in: o. faruk, m. sain. biofiber reinforcements in composite materials. woodhead publishing, cambridge, uk, 2015, p. 571–647. https://doi.org/10.1533/9781782421276.5.571. [68] b. lee, h. yang, w. yu. fabrication of double-tubular carbon nanofibers using quadruple coaxial electrospinning. nanotechnology 25(46) (2014) 465602. https://doi.org/10.1088/09574484/25/46/465602. [69] c. gualandi, a. celli, a. zucchelli, m. focarete. nanohybrid materials by electrospinning. in: s. kalia, y. haldorai. organic-inorganic hybrid nanomaterials. springer international publishing, switzerland, 2015, p. 87–142. https://doi.org/10.1007/12_2014_281. [70] q. wang, d. yu, l. zhang, x. liu, y. deng, m. zhao. electrospun hypromellose-based hydrophilic composites for rapid dissolution of poorly water-soluble drug. carbohydrate polymers 174 (2017) 617–625. https://doi.org/10.1016/j.carbpol.2017.06.075. [71] y. wu, d. yu, h. li, x. wu, x. li. medicated structural pvp/peg composites fabricated using coaxial electrospinning. e-polymers 17(1) (2017) 39–44. https://doi.org/10.1515/epoly-2016-0244. [72] a. gañán, j. lópez, m. herrada, a. ramos, j. montanero. review on the physics of electrospray: from electrokinetics to the operating conditions of single and coaxial taylor cone-jets, and ac electrospray. journal of aerosol science 125 (2018) 32–56. https://doi.org/10.1016/j.jaerosci.2018.05.002. [73] x. wang, q. wang, f, huang, q. wei. the morphology of taylor cone influenced by different coaxial composite nozzle structures. fibers and polymers 17 (2016) 624-629. https://doi.org/10.1007/s12221-016-5730-6. [74] a. khalf, s. madihally. recent advances in multiaxial electrospinning for drug delivery. european journal of pharmaceutics and biopharmaceutics 112 (2017) 1–17. https://doi.org/10.1016/j.ejpb.2016.11.010. [75] h. salehhudin, e. mohamad, w. mahadi, a. afifi. multiple-jet electrospinning methods for nanofiber processing: a review. materials and manufacturing processes 33(5) (2018) 479–498. https://doi.org/10.1080/10426914.2017.1388523. [76] m. izzat, q. zafar, z. ahmad, k. sulaiman. humidity sensor based on electrospun meh-ppv:pvp microstructured composite. rsc advances 6(42) (2016) 35387–35393. https://doi.org/10.1039/c6ra03628g. [77] s. theron, a. yarin, e. zussman, e. kroll. multiple jets in electrospinning: experiment and modeling. polymer 46(9) (2005) 2889–2899. https://doi.org/10.1016/j.polymer.2005.01.054. [78] y. zheng, r. gong, y. zeng. multijet motion and deviation in electrospinning. rsc advances 5(60) (2015) 48533–48540. https://doi.org/10.1039/c5ra06049d. [79] y. zhang, z. cheng, z. han, s. zhao, x. zhao, l. kang. stable multi-jet electrospinning with high throughput using the bead structure nozzle. rsc advances 8(11) (2018) 6069-6074. https://doi.org/10.1039/c7ra13125a. https://dx.doi.org/10.5599/admet.844 https://doi.org/10.1016/j.carbpol.2014.09.002 https://doi.org/10.1016/b978-0-08-100907-9.00003-9 https://doi.org/10.1016/j.matlet.2014.01.076 https://doi.org/10.1039/c4ta06044j https://doi.org/%1f10.1016/j.jconrel.2014.04.025 https://doi.org/%1f10.1016/j.jconrel.2014.04.025 https://doi.org/10.1533/9781782421276.5.571 https://doi.org/10.1088/0957-4484/25/46/465602 https://doi.org/10.1088/0957-4484/25/46/465602 https://doi.org/10.1007/12_2014_281 https://doi.org/10.1016/j.carbpol.2017.06.075 https://doi.org/10.1515/epoly-2016-0244 https://doi.org/10.1016/j.jaerosci.%1f2018.05.002 https://doi.org/10.1007/s12221-016-5730-6 https://doi.org/10.1007/s12221-016-5730-6 https://doi.org/10.1016/j.ejpb.%1f2016.11.010 https://doi.org/10.1016/j.ejpb.%1f2016.11.010 https://doi.org/10.1080/10426914.2017.1388523 https://doi.org/10.1080/10426914.2017.1388523 https://doi.org/10.1039/%1fc6ra03628g https://doi.org/10.1039/%1fc6ra03628g https://doi.org/10.1016/j.polymer.2005.01.054 https://doi.org/10.1039/c5ra06049d https://doi.org/10.1039/c7ra13125a https://doi.org/10.1039/c7ra13125a luis castillo et al admet & dmpk 8(4) (2020) 325-353 344 [80] y. liu, l. zhang, x. sun, j. liu, j. fan, d. huang. multi-jet electrospinning via auxiliary electrode. materials letters 141 (2015) 153–156. https://doi.org/10.1016/j.matlet.2014.11.079. [81] y. wu, l. wang, j. fan, w. shou, b. zhou, y. liu. multi-jet electrospinning with auxiliary electrode: the influence of solution properties. polymers 10(6) (2018) 572. https://doi.org/10.3390/polym10060572. [82] p. gupta, g. wilkes. some investigations on the fiber formation by utilizing a side-by-side bicomponent electrospinning approach. polymer 44(20) (2003) 6353–6359. https://doi.org/10.1016/s0032-3861(03)00616-5. [83] k. nasouri, a. shoushtari, m. mohaddes. thermodynamic studies on polyvinylpyrrolidone solution systems used for fabrication of electrospun nanostructures: effects of the solvent. advances in polymer technology 34(3) (2015) 21495. https://doi.org/10.1002/adv.21495. [84] m. cai, h. he, x. zhang, x. yan, j. li, f. chen. efficient synthesis of pvdf/pi side-by-side bicomponent nanofiber membrane with enhanced mechanical strength and good thermal stability. nanomaterials 9(1) (2019) 39. https://doi.org/10.3390/nano9010039. [85] l. peng, s. jiang, m. seuß, a. fery, g. lang, t. scheibel. two-in-one composite fibers with side-byside arrangement of silk fibroin and poly(l-lactide) by electrospinning. macromolecular materials and engineering 301(1) (2016) 48–55. https://doi.org/10.1002/mame.201500217. [86] d. yu, c. yang, m. jin, g. williams, h. zou, x. wang. medicated janus fibers fabricated using a tefloncoated side-by-side spinneret. colloids and surfaces b: biointerfaces 138 (2016) 110–116. https://doi.org/10.1016/j.colsurfb.2015.11.055. [87] d. yu, j. li, m. zhang, r. williams. high-quality janus nanofibers prepared using three-fluid electrospinning. chemical communications 53(33) (2017) 4542–4545. https://doi.org/10.1039/c7cc01661a. [88] x. xu, l. yang, x. xu. ultrafine medicated fibers electrospun from w/o emulsions. journal of controlled release 108(1) (2005) 33-42. https://doi.org/10.1016/j.jconrel.2005.07.021. [89] c. zhang, f. feng, h. zhang. emulsion electrospinning: fundamentals, food applications and prospects. trends in food science & technology 80 (2018) 175–186. https://doi.org/10.1016/j.tifs.2018.08.005. [90] n. cai, c. han, x. luo, g. chen, q. dai, f. yu. fabrication of core/shell nanofibers with desirable mechanical and antibacterial properties by pickering emulsion electrospinning. macromolecular materials and engineering 302(3) (2016). https://doi.org/10.1002/mame.201600364. [91] x. wang, y. yuan, x. huang, t. yue. controlled release of protein from core–shell nanofibers prepared by emulsion electrospinning based on green chemical. journal of applied polymer science 132(16) (2014). https://doi.org/10.1002/app.41811. [92] a. basar, s. castro, s. torres, j. lagaron, s. turkoglu s. novel poly(ε-caprolactone)/gelatin wound dressings prepared by emulsion electrospinning with controlled release capacity of ketoprofen antiinflammatory drug. materials science and engineering c 81 (2017) 459–468. https://doi.org/10.1016/j.msec.2017.08.025. [93] p. pal, p. srivas, p. dadhich, b. das, d. maulik, s. dhara. nano-/microfibrous cotton-wool-like 3d scaffold with core–shell architecture by emulsion electrospinning for skin tissue regeneration. acs biomaterials science & engineering 3(12) (2017) 3563–3575. https://doi.org/10.1021/acsbiomaterials.7b00681. [94] j. hu, m. prabhakaran, x. ding, s. ramakrishna. emulsion electrospinning of polycaprolactone: influence of surfactant type towards the scaffold properties. journal of biomaterials science polymer edition 26(1) (2015) 57–75. https://doi.org/10.1080/09205063.2014.982241. [95] a. moydeen, p. ali, e. aboelfetoh, s. al-deyab, m. el-newehy. fabrication of electrospun poly(vinyl alcohol)/dextran nanofibers via emulsion process as drug delivery system: kinetics and in vitro release study. international journal of biological macromolecules 116 (2018) 1250–1259. https://doi.org/10.1016/j.ijbiomac.2018.05.130. https://doi.org/10.1016/j.matlet.2014.11.079 https://doi.org/10.3390/%1fpolym10060572 https://doi.org/10.3390/%1fpolym10060572 https://doi.org/10.1016/s0032-3861(03)00616-5 https://doi.org/10.1016/s0032-3861(03)00616-5 https://doi.org/10.1002/adv.21495 https://doi.org/10.3390/nano9010039 https://doi.org/10.1002/mame.201500217 https://doi.org/10.1016/j.colsurfb.2015.11.055 https://doi.org/10.1039/%1fc7cc01661a https://doi.org/10.1039/%1fc7cc01661a https://doi.org/10.1016/j.jconrel.2005.07.021 https://doi.org/10.1016/j.tifs.2018.08.005 https://doi.org/10.1016/j.tifs.2018.08.005 https://doi.org/10.1002/mame.201600364 https://doi.org/10.1002/app.41811 https://doi.org/%1f10.1016/j.msec.2017.08.025 https://doi.org/%1f10.1016/j.msec.2017.08.025 https://doi.org/10.1021/acsbiomaterials.7b00681 https://doi.org/10.1021/acsbiomaterials.7b00681 https://doi.org/10.1080/09205063.2014.982241 https://doi.org/10.1016/j.ijbiomac.2018.05.130 admet & dmpk 8(4) (2020) 325-353 electrospun nanofibers: a nanotechnological approach doi: https://dx.doi.org/10.5599/admet.844 345 [96] r. nayak, i. kyratzis, y. truong, r. padhye, l. arnold, g. peeters. fabrication and characterisation of nanofibres by meltblowing and melt electrospinning. advanced materials research 472 (2012) 1294–1299. https://doi.org/10.4028/www.scientific.net/amr.472-475.1294. [97] y. qin, l. cheng, y. zhang, x. chen, x. wang, x. he. efficient preparation of poly(lactic acid) nanofibers by melt differential electrospinning with addition of acetyl tributyl citrate. journal of applied polymer science 135(31) (2018) 46554. https://doi.org/10.1002/app.46554. [98] m. muerza, d. haylock, d. hutmacher, p. dalton. melt electrospinning and its technologization in tissue engineering. tissue engineering part b: reviews 21(2) (2014) 187–202. https://doi.org/10.1089/ten.teb.2014.0347. [99] m. yamamoto, h. yamane. melt electrospinning: electrodynamics and spinnability. polymer 132 (2017) 206–215. https://doi.org/10.1016/j.polymer.2017.11.006. [100] n. bhardway, s. kundu. electrospinning: a fascinating fiber fabrication technique. biotechnology advances 28(3) (2010) 325–347. https://doi.org/10.1016/j.biotechadv.2010.01.004. [101] d. lv, m. zhu, z. jiang, s. jiang, q. zhang, r. xiong. green electrospun nanofibers and their application in air filtration. macromolecular materials and engineering 303(12) (2018). https://doi.org/10.1002/mame.201800336. [102] t. brown, p. dalton, d. hutmacher. melt electrospinning today: an opportune time for an emerging polymer process. progress in polymer science 56 (2016) 116–166. https://doi.org/10.1016/j.progpolymsci.2016.01.001. [103] z. li, d. xiao, y. xu, y. miao, n. xin, z. yong, l. yun. recent advances in melt electrospinning. rsc advances 6(58) (2016) 53400-53414. https://doi.org/10.1039/c6ra09558e. [104] l. larrondo, r. manley. electrostatic fiber spinning from polymer melts. i. experimental observations on fiber formation and properties. journal of polymer science: polymer physics edition 19(6) (1981) 909–20. https://doi.org/10.1002/pol.1981.180190601. [105] l. larrondo, r. manley. electrostatic fiber spinning from polymer melts. ii. examination of the flow field in an electrically driven jet. journal of polymer science: polymer physics edition 19(6) (1981) 921–32. https://doi.org/10.1002/pol.1981.180190602. [106] l. larrondo, r. manley. electrostatic fiber spinning from polymer melts. iii. electrostatic deformation of a pendant drop of polymer melt. journal of polymer science: polymer physics edition 19(6) (1981) 933–40. https://doi.org/10.1002/pol.1981.180190603. [107] y. weimin, l. haoyi. principle and equipment of polymer melt differential electrospinning preparing ultrafine fiber. materials science and engineering 64 (2014) 012013. https://doi.org/10.1088/1757899x/64/1/012013. [108] h. li, h. chen, x. zhong, w. wu, y. ding, w. yang. interjet distance in needleless melt differential electrospinning with umbellate nozzles. journal of applied polymer science 131(15) (2014) 40515. https://doi.org/10.1002/app.40515. [109] a. blakney, c. ball, e. krogstad, k. woodrow. electrospun fibers for vaginal anti-hiv drug delivery. antiviral research 100 (2013) 9-16. https://doi.org/10.1016/j.antiviral.2013.09.022. [110] b. singh, n. panda, r. mund, k. pramanik. carboxymethyl cellulose enables silk fibroin nanofibrous scaffold with enhanced biomimetic potential for bone tissue engineering application. carbohydrate polymers 151 (2016) 335–347. https://doi.org/10.1016/j.carbpol.2016.05.088. [111] l. preem, m. mahmoudzadeh, m. putrinš, a. meos, i. laidmäe, t. romann. interactions between chloramphenicol, carrier polymers, and bacteria-implications for designing electrospun drug delivery systems countering wound infection. molecular pharmaceutics 14(12) (2017) 4417–4430. https://doi.org/10.1021/acs.molpharmaceut.7b00524. [112] j. yao, s. zhang, w. li, z. du, y. li. in vitro drug controlled-release behavior of an electrospun modified poly(lactic acid)/bacitracin drug delivery system. rsc advances 6(1) (2015) 515–521. https://doi.org/10.1039/c5ra22467e. [113] s. thakkar, m. misra. electrospun polymeric nanofibers: new horizons in drug delivery. european journal of pharmaceutical sciences 107 (2017) 148–167. https://doi.org/10.1016/j.ejps.2017.07.001. https://dx.doi.org/10.5599/admet.844 https://doi.org/10.4028/www.scientific.net/amr.472-475.1294 https://doi.org/10.1002/app.46554 https://doi.org/%1f10.1089%1f/ten.teb.2014.0347 https://doi.org/%1f10.1089%1f/ten.teb.2014.0347 https://doi.org/10.1016/j.polymer.2017.11.006 https://doi.org/10.1016/j.biotechadv.2010.01.004 https://doi.org/10.1002/mame.201800336 https://doi.org/10.1016/j.%1fprogpolymsci%1f.2016.01.001 https://doi.org/10.1016/j.%1fprogpolymsci%1f.2016.01.001 https://doi.org/10.1039/c6ra09558e https://doi.org/10.1002/pol.1981.180190601 https://doi.org/10.1002/pol.1981.180190602 https://doi.org/10.1088/1757-899x/64/1/012013 https://doi.org/10.1088/1757-899x/64/1/012013 https://doi.org/10.1002/app.40515 https://doi.org/10.1016/j.antiviral.2013.09.022 https://doi.org/10.1016/j.carbpol.2016.05.088 https://doi.org/10.1021/acs.molpharmaceut.7b00524 https://doi.org/10.1039/c5ra22467e https://doi.org/10.1016/j.ejps.2017.07.%1f001 luis castillo et al admet & dmpk 8(4) (2020) 325-353 346 [114] d. kai, s. liow, x. loh. biodegradable polymers for electrospinning: towards biomedical applications. materials science and engineering: c 45 (2014) 659–670. https://doi.org/10.1016/j.msec.2014.04.051. [115] j. ding, j. zhang, j. li, d. li, c. xiao, h. xiao. electrospun polymer biomaterials. progress in polymer science 90 (2019) 1–34. https://doi.org/10.1016/j.progpolymsci.2019.01.002. [116] c. luo, e. stride, m. edirisinghe. mapping the influence of solubility and dielectric constant on electrospinning polycaprolactone solutions. macromolecules 45(11) (2012) 4669-4680. https://doi.org/10.1021/ma300656u. [117] j. schiffman, c. schauer. a review: electrospinning of biopolymer nanofibers and their applications. polymer reviews 48(2) (2008) 317–352. https://doi.org/10.1080/15583720802022182. [118] m. maniruzzaman, d. morgan, a. mendham, j. pang, m. snowden, d. douroumis. drug–polymer intermolecular interactions in hot-melt extruded solid dispersions. international journal of pharmaceutics 443(1) (2013) 199–208. https://doi.org/10.1016/j.ijpharm.2012.11.048. [119] f. meng, a. trivino, d. prasad, h. chauhan. investigation and correlation of drug polymer miscibility and molecular interactions by various approaches for the preparation of amorphous solid dispersions. european journal of pharmaceutical sciences 71 (2015) 12–24. https://doi.org/10.1016/j.ejps.2015.02.003. [120] a. prudic, y. ji, g. sadowski. thermodynamic phase behavior of api/polymer solid dispersions. molecular pharmaceutics 11(7) (2014) 2294–2304. https://doi.org/10.1021/mp400729x. [121] a. sadeghi, h. nazem, m. rezakazemi, s. shirazian. predictive construction of phase diagram of ternary solutions containing polymer/solvent/nonsolvent using modified flory-huggins model. journal of molecular liquids 263 (2018) 282–287. https://doi.org/10.1016/j.molliq.2018.05.015. [122] h. lu, s. du. a phenomenological thermodynamic model for the chemo-responsive shape memory effect in polymers based on flory–huggins solution theory. polymer chemistry 5(4) (2014) 1155– 1162. https://doi.org/10.1039/c3py01256e. [123] c. donnelly, y. tian, c. potter, d. jones, g. andrews. probing the effects of experimental conditions on the character of drug-polymer phase diagrams constructed using flory-huggins theory. pharmaceutical research 32(1) (2015) 167–179. https://doi.org/10.1007/s11095-014-1453-9. [124] m. knopp, l. tajber, y. tian, n. olesen, d. jones, a. kozyra. comparative study of different methods for the prediction of drug–polymer solubility. molecular pharmaceutics 12(9) (2015) 3408–3419. https://doi.org/10.1021/acs.molpharmaceut.5b00423. [125] a. gravelle, m. davidovich, a. zetzl, s. barbut, a. marangoni. influence of solvent quality on the mechanical strength of ethylcellulose oleogels. carbohydrate polymers 135 (2016) 169–79. https://doi.org/10.1016/j.carbpol.2015.08.050. [126] c. luo, m. nangrejo, m. edirisinghe. a novel method of selecting solvents for polymer electrospinning. polymer 51(7) (2010) 1654–1662. https://doi.org/10.1016/j.polymer.2010.01.031. [127] s. mahalingam, b. raimi, d. craig, m. edirisinghe. solubility–spinnability map and model for the preparation of fibres of polyethylene (terephthalate) using gyration and pressure. chemical engineering journal 280 (2015) 344–353. https://doi.org/10.1016/j.cej.2015.05.114. [128] c. mathis, m. divandari, r. simic, v. naik, e. benetti, l. isa. atr-ir investigation of solvent interactions with surface-bound polymers. langmuir 32(30) (2016) 7588–7595. https://doi.org/10.1021/acs.langmuir.6b02086. [129] j. jacob, j. haponiuk, s. thomas, s. gopi. biopolymer based nanomaterials in drug delivery systems: a review. materials today chemistry 9 (2018) 43–55. https://doi.org/10.1016/j.mtchem.2018.05.002. [130] s. khatsee, d. daranarong, w. punyodom, p. worajittiphon. electrospinning polymer blend of pla and pbat: electrospinnability–solubility map and effect of polymer solution parameters toward application as antibiotic-carrier mats. journal of applied polymer science 135(28) (2018). https://doi.org/10.1002/app.46486. [131] m. rashid, q. zaid, s. tajuddin. chapter 11 trends in nanotechnology for practical applications. in: s. mohapatra, s. ranjan, n. dasgupta, r. kumar, s. thomas. applications of targeted nano drugs and https://doi.org/10.1016/j.%1fmsec.%1f2014.%1f04.%1f051 https://doi.org/10.1016/j.%1fmsec.%1f2014.%1f04.%1f051 https://doi.org/10.1016/j.progpolymsci.2019.01.002 https://doi.org/10.1021/ma300656u https://doi.org/10.1080/15583720802%1f022182 https://doi.org/10.1016/j.ijpharm.2012.11.048 https://doi.org/10.1016/j.ejps.2015.02.003 https://doi.org/10.1016/j.ejps.2015.02.003 https://doi.org/10.1021/mp400729x https://doi.org/10.1016/j.molliq.2018.05.015 https://doi.org/10.1039/c3py01256e https://doi.org/10.1007/s11095-014-1453-9 https://doi.org/10.1021/acs.molpharmaceut.5b00423 https://doi.org/10.1016/j.carbpol.2015.08.050 https://doi.org/10.1016/j.polymer.2010.01.031 https://doi.org/10.1016/j.cej.2015.05.114 https://doi.org/10.1021/acs.langmuir.6b02086 https://doi.org/10.1016/j.mtchem.%1f2018.05.002 https://doi.org/10.1002/app.46486 admet & dmpk 8(4) (2020) 325-353 electrospun nanofibers: a nanotechnological approach doi: https://dx.doi.org/10.5599/admet.844 347 delivery systems: nanoscience and nanotechnology in drug delivery, elsevier, 2018. p. 297-325. https://doi.org/10.1016/b978-0-12-814029-1.00011-9. [132] v. chavda. chapter 1 nanotherapeutics and nanobiotechnology. in: s. mohapatra, s. ranjan, n. dasgupta, r. kumar, s. thomas. applications of targeted nano drugs and delivery systems: nanoscience and nanotechnology in drug delivery, elsevier, 2018, p. 1-13. https://doi.org/10.1016/b978-0-12-814029-1.00001-6. [133] f. lopez, t. ernest, c. tuleu, m. gul. formulation approaches to pediatric oral drug delivery: benefits and limitations of current platforms. expert opinion on drug delivery. 12(11) (2015) 1727–1740. https://doi.org/10.1517/17425247.2015.1060218. [134] a. aleksovski, r. dreu, m. gašperlin, o. planinšek. mini-tablets: a contemporary system for oral drug delivery in targeted patient groups. expert opinion on drug delivery 12(1) (2015) 65–84. https://doi.org/10.1517/17425247.2014.951633. [135] l. castillo, r. vargas, g. carazo, g. madrigal, b. calvo, e. baltodano. development of immediate release rupatadine fumarate 10 mg tablets: a quality by design (qbd) approach. drug development and industrial pharmacy 45(10) (2019) 1674–1681. https://doi.org/10.1080/03639045.2019.1652637. [136] u. maver, t. velnar, m. gaberšček, o. planinšek, m. finšgar. recent progressive use of atomic force microscopy in biomedical applications. trends in analytical chemistry 80 (2016) 96–111. https://doi.org/10.1016/j.trac.2016.03.014. [137] b. mittal. chapter 2 pharmacokinetics and preformulation. in: b. mittal. how to develop robust solid oral dosage forms from conception to post-approval, academic press, 2017 p. 17–37. https://doi.org/10.1016/b978-0-12-804731-6.00002-9. [138] k. göke, t. lorenz, a. repanas, f. schneider, d. steiner, k. baumann. novel strategies for the formulation and processing of poorly water-soluble drugs. european journal of pharmaceutics and biopharmaceutics 126 (2018) 40–56. https://doi.org/10.1016/j.ejpb.2017.05.008. [139] g. amidon, h. lennernäs, v. shah, j. crison. a theoretical basis for a biopharmaceutic drug classification: the correlation of in vitro drug product dissolution and in vivo bioavailability. pharmaceutical research 12(3) (1995) 413–420. http://dx.doi.org/10.1023/a:1016212804288. [140] l. yu, g. amidon, j. polli, h. zhao, m. mehta, d. conner. biopharmaceutics classification system: the scientific basis for biowaiver extensions. pharmaceutical research 19(7) (2002) 921–925. http://dx.doi.org/10.1023/a:1016473601633. [141] r. kamble, s. gaikwad, a. maske, s. patil. fabrication of electrospun nanofibres of bcs ii drug for enhanced dissolution and permeation across skin. journal of advanced research 7(3) (2016) 483– 489. http://dx.doi.org/10.1016/j.jare.2016.03.009. [142] s. baghel, h. cathcart, n. o’reilly. polymeric amorphous solid dispersions: a review of amorphization, crystallization, stabilization, solid-state characterization, and aqueous solubilization of biopharmaceutical classification system class ii drugs. journal of pharmaceutical sciences 105(9) (2016) 2527–2544. https://doi.org/10.1016/j.xphs.2015.10.008. [143] s. kumar, r. kaur, r. rajput, m. singh. bio pharmaceutics classification system (bcs) class iv drug nanoparticles: quantum leap to improve their therapeutic index. advanced pharmaceutical bulletin 8(4) (2018) 617–625. https://doi.org/10.15171/apb.2018.070. [144] a. jacobsen, a. krupa, m. brandl, a. bauer. high-throughput dissolution/permeation screening-a 96well two-compartment microplate approach. pharmaceutics 11(5) (2019) 227. https://doi.org/10.3390/pharmaceutics11050227. [145] t. sou, c. bergström. automated assays for thermodynamic (equilibrium) solubility determination. drug discovery today: technologies 27 (2018) 11–19. https://doi.org/10.1016/j.ddtec.2018.04.004. [146] z. nagy, a. balogh, b. démuth, h. pataki, t. vigh, b. szabó. high speed electrospinning for scaled-up production of amorphous solid dispersion of itraconazole. international journal of pharmaceutics 480(1–2) (2015) 137–142. https://doi.org/10.1016/j.ijpharm.2015.01.025. https://dx.doi.org/10.5599/admet.844 https://doi.org/10.1016/b978-0-12-814029-1.00011-9 https://doi.org/10.1016/b978-0-12-814029-1.00001-6 https://doi.org/10.1517/17425247.2015.1060218 https://doi.org/10.1517/17425247.2014.951633 https://doi.org/10.1080/03639045.2019.1652637 https://doi.org/10.1080/03639045.2019.1652637 https://doi.org/10.1016/j.trac.2016.03.014 https://doi.org/10.1016/b978-0-12-804731-6.00002-9 https://doi.org/10.1016/j.ejpb.2017.05.008 http://dx.doi.org/10.1023/a:1016212804288 http://dx.doi.org/10.1023/a:1016473601633 http://dx.doi.org/10.1016/j.jare.2016.03.009 https://doi.org/10.1016/j.xphs.2015.10.008 https://doi.org/10.15171/apb.2018.070 https://doi.org/10.1016/j.ddtec.2018.04.004 https://doi.org/10.1016/j.ijpharm.2015.01.025 luis castillo et al admet & dmpk 8(4) (2020) 325-353 348 [147] u. paaver, i. tamm, i. laidmäe, a. lust, k. kirsimäe, p. veski. soluplus graft copolymer: potential novel carrier polymer in electrospinning of nanofibrous drug delivery systems for wound therapy. biomed research international (2014) 789765. https://doi.org/10.1155/2014/789765. [148] r. zelkó, d. lamprou, i. sebe. recent development of electrospinning for drug delivery. pharmaceutics 12(1) (2020) 5. https://doi.org/10.3390/pharmaceutics12010005. [149] l. kenry, c. lim. nanofiber technology: current status and emerging developments. progress in polymer science 70 (2017) 1–17. https://doi.org/10.1016/j.progpolymsci.2017.03.002. [150] a. bhatia, s. chopra, k. nagpal, p. kishore, m. tekade, r. tekade. chapter 2 polymorphism and its implications in pharmaceutical product development. in: r. tekade. dosage form design parameters, academic press, 2018, p. 31-65. https://doi.org/10.1016/b978-0-12-814421-3.00002-6. [151] b. poller, c. strachan, r. broadbent, g. walker. a minitablet formulation made from electrospun nanofibers. european journal of pharmaceutics and biopharmaceutics 114 (2017) 213–220. https://doi.org/10.1016/j.ejpb.2017.01.022. [152] p. vass, e. szabó, a. domokos, e. hirsch, d. galata, b. farkas. scale-up of electrospinning technology: applications in the pharmaceutical industry. wires nanomedicine and nanobiotechnology (2019) 124. https://doi.org/10.1002/wnan.1611. [153] p. vass, e. hirsch, r. kóczián, b. démuth, a. farkas, c. fehér. scaled-up production and tableting of grindable electrospun fibers containing a protein-type drug. pharmaceutics 11(7) (2019) 329. https://doi.org/10.3390/pharmaceutics11070329. [154] m. abrigo, s. mcarthur, p. kingshott. electrospun nanofibers as dressings for chronic wound care: advances, challenges, and future prospects. macromolecular bioscience 14(6) (2014) 772–792. https://doi.org/10.1002/mabi.201300561. [155] j. boateng, o. catanzano. advanced therapeutic dressings for effective wound healing--a review. journal of pharmaceutical sciences 104(11) (2015) 3653–3680. https://doi.org/10.1002/jps.24610. [156] k. ye, h. kuang, z. you, y. morsi, x. mo. electrospun nanofibers for tissue engineering with drug loading and release. pharmaceutics 11(4) (2019). https://doi.org/10.3390/pharmaceutics11040182. [157] h. lian, z. meng. melt electrospinning vs. solution electrospinning: a comparative study of drugloaded poly (ε-caprolactone) fibres. materials science and engineering: c 74 (2017) 117–123. https://doi.org/10.1016/j.msec.2017.02.024. [158] t. zhu, s. chen, w. li, j. lou, j. wang. flurbiprofen axetil loaded coaxial electrospun poly(vinyl pyrrolidone)–nanopoly(lactic-co-glycolic acid) core–shell composite nanofibers: preparation, characterization, and anti-adhesion activity. journal of applied polymer science 132(22) (2015). https://doi.org/10.1002/app.41982. [159] m. shishir, l. xie, c. sun, x. zheng, w. chen. advances in micro and nano-encapsulation of bioactive compounds using biopolymer and lipid-based transporters. trends in food science & technology 78 (2018) 34–60. https://doi.org/10.1016/j.tifs.2018.05.018. [160] p. wen, m. zong, r. linhardt, k. feng, h. wu. electrospinning: a novel nano-encapsulation approach for bioactive compounds. trends in food science & technology 70 (2017) 56–68. https://doi.org/10.1016/j.tifs.2017.10.009. [161] p. wen, y. wen, m. zong, r. linhardt, h. wu. encapsulation of bioactive compound in electrospun fibers and its potential application. journal of agricultural and food chemistry 65(42) (2017) 9161– 9179. https://doi.org/10.1021/acs.jafc.7b02956. [162] m. ruiz, m. volonté. biopharmaceutical relevance of the comparison of dissolution profiles: proposal of a combined approach. dissolution technologies (2014) 32-43. http://dx.doi.org/ 10.14227/dt210114p32. [163] j. mašek, d. lubasová, r. lukáč, p. turánek, p. kulich, j. plocková. multi-layered nanofibrous mucoadhesive films for buccal and sublingual administration of drug-delivery and vaccination nanoparticles important step towards effective mucosal vaccines. journal of controlled release 249 (2017) 183–95. https://doi.org/10.1016/j.jconrel.2016.07.036. https://doi.org/10.1155/2014/789765 https://doi.org/10.3390/pharmaceutics12010005 https://doi.org/10.1016/j.progpolymsci.2017.03.002 https://doi.org/10.1016/b978-0-12-814421-3.00002-6 https://doi.org/10.1016/j.ejpb.2017.01.022 https://doi.org/10.1002/wnan.1611 https://doi.org/10.3390/pharmaceutics11070329 https://doi.org/10.1002/mabi.201300561 https://doi.org/10.1002/jps.24610 https://doi.org/10.3390/pharmaceutics%1f11040182 https://doi.org/10.1016/j.msec.2017.02.024 https://doi.org/10.1002/app.41982 https://doi.org/10.1016/j.tifs.2018.05.018 https://doi.org/10.1016/j.tifs.2017.10.009 https://doi.org/10.1021/acs.jafc.7b02956 http://dx.doi.org/%2010.14227/dt210114p32 http://dx.doi.org/%2010.14227/dt210114p32 https://doi.org/10.1016/j.jconrel.2016.07.036 admet & dmpk 8(4) (2020) 325-353 electrospun nanofibers: a nanotechnological approach doi: https://dx.doi.org/10.5599/admet.844 349 [164] r. ricarte, z. li, l. johnson, j. ting, t. reineke, f. bates. direct observation of nanostructures during aqueous dissolution of polymer/drug particles. macromolecules 50(8) (2017) 3143–3152. https://doi.org/10.1021/acs.macromol.7b00372. [165] s. qi, d. craig. recent developments in microand nanofabrication techniques for the preparation of amorphous pharmaceutical dosage forms. advanced drug delivery reviews 100 (2016) 67–84. https://doi.org/10.1016/j.addr.2016.01.003. [166] a. mendes, c. gorzelanny, n. halter, s. schneider, i. chronakis. hybrid electrospun chitosanphospholipids nanofibers for transdermal drug delivery. international journal of pharmaceutics 510(1) (2016) 48–56. https://doi.org/10.1016/j.ijpharm.2016.06.016. [167] a. gencturk, e. kahraman, s. güngör, g. özhan, y. özsoy, a. sarac. polyurethane/hydroxypropyl cellulose electrospun nanofiber mats as potential transdermal drug delivery system: characterization studies and in vitro assays. artificial cells, nanomedicine, and biotechnology 45(3) (2017) 655–664. https://doi.org/10.3109/21691401.2016.1173047. [168] r. najafi, m. ali, r. faridi, a. amani. preparation of an ascorbic acid/pva–chitosan electrospun mat: a core/shell transdermal delivery system. rsc advances 5(62) (2015) 50462–50469. https://doi.org/10.1039/c5ra03813h. [169] k. nazari, e. kontogiannidou, r. ahmad, a. gratsani, m. rasekh, m. arshad. development and characterisation of cellulose based electrospun mats for buccal delivery of non-steroidal antiinflammatory drug (nsaid). european journal of pharmaceutical sciences 102 (2017) 147–155. https://doi.org/10.1016/j.ejps.2017.02.033. [170] p. szabó, t. daróczi, g. tóth, r. zelkó. in vitro and in silico investigation of electrospun terbinafine hydrochloride-loaded buccal nanofibrous sheets. journal of pharmaceutical and biomedical analysis 30 (2016) 156–159. https://doi.org/10.1016/j.jpba.2016.08.021. [171] e. adeli. irbesartan-loaded electrospun nanofibers-based pvp k90 for the drug dissolution improvement: fabrication, in vitro performance assessment, and in vivo evaluation. journal of applied polymer science 132(27) (2015). https://doi.org/10.1002/app.42212. [172] v. chavda. chapter 4 nanobased nano drug delivery: a comprehensive review. in: s. mohapatra, s. ranjan, n. dasgupta, r. kumar, s. thomas. applications of targeted nano drugs and delivery systems: nanoscience and nanotechnology in drug delivery, elsevier, 2018, p. 69-92. https://doi.org/10.1016/b978-0-12-814029-1.00004-1. [173] s. shahriar, j. mondal, m. hasan, v. revuri, d. lee, y. lee. electrospinning nanofibers for therapeutics delivery. nanomaterials 9(4) (2019) 532. https://doi.org/10.3390/nano9040532. [174] a. repanas, s. andriopoulou, b. glasmacher. the significance of electrospinning as a method to create fibrous scaffolds for biomedical engineering and drug delivery applications. journal of drug delivery science and technology 31 (2016) 137–46. https://doi.org/10.1016/j.jddst.2015.12.007. [175] m. saeedi, m. eslamifar, k. khezri, s. maleki. applications of nanotechnology in drug delivery to the central nervous system. biomedicine & pharmacotherapy 111 (2019) 666-675. https://doi.org/10.1016/j.biopha.2018.12.133. [176] s. maleki, f. lotfipour, m. barzegar, m. zarrintan, k. adibkia. ciprofloxacin hcl-loaded calcium carbonate nanoparticles: preparation, solid state characterization, and evaluation of antimicrobial effect against staphylococcus aureus. artificial cells, nanomedicine, and biotechnology 45(3) (2017) 535-543. https://doi.org/10.3109/21691401.2016.1161637. [177] a. akhgari, z. shakib, s. sanati. a review on electrospun nanofibers for oral drug delivery. nanomedicine journal 4(4) (2017) 197–207. [178] r. ravikumar, m. ganesh, v. senthil, y. ramesh, s. jakki, e. choi. tetrahydro curcumin loaded pclpeg electrospun transdermal nanofiber patch: preparation, characterization, and in vitro diffusion evaluations. journal of drug delivery science and technology 44 (2018) 342–348. https://doi.org/10.1016/j.jddst.2018.01.016. [179] a. ammar. polymeric nanoparticles and nanofibers for local delivery of poorly soluble drugs. recent advances in nanotechnology (2019). https://doi.org/10.11159/nddte19.116. https://dx.doi.org/10.5599/admet.844 https://doi.org/10.1021/acs.macromol.7b00372 https://doi.org/10.1016/j.addr.2016.01.003 https://doi.org/10.1016/j.ijpharm.2016.06.016 https://doi.org/10.3109/21691401.2016.1173047 https://doi.org/10.1039/c5ra03813h https://doi.org/10.1016/j.ejps.2017.02.033 https://doi.org/10.1016/j.jpba.2016.08.021 https://doi.org/10.1002/app.42212 https://doi.org/10.1016/b978-0-12-814029-1.00004-1 https://doi.org/10.3390/nano9040532 https://doi.org/10.1016/j.jddst.2015.12.007 https://doi.org/10.1016/j.biopha.2018.12.133 https://doi.org/10.1016/j.biopha.2018.12.133 https://doi.org/10.3109/21691401.2016.1161637 https://doi.org/%1f10.1016/j.jddst.2018.01.016 https://doi.org/%1f10.1016/j.jddst.2018.01.016 https://doi.org/10.11159/nddte19.116 luis castillo et al admet & dmpk 8(4) (2020) 325-353 350 [180] e. torres, g. pérez, a. serrano, d. grande, r. vera, j. cornejo. drugs loaded into electrospun polymeric nanofibers for delivery. journal of pharmacy and pharmaceutical sciences 22(1) (2019) 313–331. https://doi.org/10.18433/jpps29674. [181] u. paaver, j. heinämäki, i. laidmäe, a. lust, j. kozlova, e. sillaste. electrospun nanofibers as a potential controlled-release solid dispersion system for poorly water-soluble drugs. international journal of pharmaceutics 479(1) (2015) 252–60. https://doi.org/10.1016/j.ijpharm.2014.12.024. [182] r. rajput, j. narkhede, j. naik. nanogels as nanocarriers for drug delivery: a review. admet and dmpk 8(1) (2020) 1–15. http://dx.doi.org/10.5599/admet.724. [183] g. yazgan, a. popa, r. rossi, k. maniura, l. puigmartí, d. crespy. tunable release of hydrophilic compounds from hydrophobic nanostructured fibers prepared by emulsion electrospinning. polymer 66 (2015) 268–276. https://doi.org/10.1016/j.polymer.2015.04.045. [184] z. cui, z. zheng, l. lin, j. si, q. wang, x. peng. electrospinning and crosslinking of polyvinyl alcohol/chitosan composite nanofiber for transdermal drug delivery. advances in polymer technology 37(6) (2018) 917–928. https://doi.org/10.1002/adv.21850. [185] s. jin, a. yousaf, k. kim, d. kim, j. kim. influence of hydrophilic polymers on functional properties and wound healing efficacy of hydrocolloid based wound dressings. international journal of pharmaceutics 501(1) (2016) 160–166. https://doi.org/10.1016/j.ijpharm.2016.01.044. [186] m. ghori, b. conway. hydrophilic matrices for oral control drug delivery. american journal of pharmacological sciences 3(5) (2015) 103–109. https://doi.org/10.1016/10.12691/ajps-3-5-1. [187] m. irfan, s. rabel, q. bukhtar, m. qadir, f. jabeen, a. khan. orally disintegrating films: a modern expansion in drug delivery system. saudi pharmaceutical journal 24(5) (2016) 537–546. https://doi.org/10.1016/j.jsps.2015.02.024. [188] a. srivastava, t. yadav, s. sharma, a. nayak, a. kumari, n. mishra. polymers in drug delivery. journal of biosciences and medicines 4(1) (2015) 69–84. https://doi.org/10.4236/jbm.2016.41009. [189] v. gordon, g. marom, s. magdassi. formation of hydrophilic nanofibers from nanoemulsions through electrospinning. international journal of pharmaceutics 478(1) (2015) 172–179. https://doi.org/10.1016/j.ijpharm.2014.11.038. [190] k. ghosal, a. manakhov, l. zajíčková, s. thomas. structural and surface compatibility study of modified electrospun poly(ε-caprolactone) (pcl) composites for skin tissue engineering. aaps pharmscitech 18(1) (2017) 72–81. https://doi.org/10.1208/s12249-016-0500-8. [191] k. ghosal, a. chandra, s. roy, c. agatemor. electrospinning over solvent casting: tuning of mechanical properties of membranes. scientific reports 8(1) (2018) 1–9. https://doi.org/10.1038/s41598-018-23378-3. [192] a. aydogdu, g. sumnu, s. sahin. a novel electrospun hydroxypropyl methylcellulose/polyethylene oxide blend nanofibers: morphology and physicochemical properties. carbohydrate polymers 181 (2018) 234–246. https://doi.org/10.1016/j.carbpol.2017.10.071. [193] x. zhao, j. zhao, z. lin, g. pan, y. zhu, y. cheng. self-coated interfacial layer at organic/inorganic phase for temporally controlling dual-drug delivery from electrospun fibers. colloids and surfaces b: biointerfaces 130 (2015) 1–9. https://doi.org/10.1016/j.colsurfb.2015.03.058. [194] a. haider, s. haider, i. kang. a comprehensive review summarizing the effect of electrospinning parameters and potential applications of nanofibers in biomedical and biotechnology. arabian journal of chemistry 11(8) (2018) 1165–1188. https://doi.org/10.1016/j.arabjc.2015.11.015. [195] l. jørgensen, k. qvortrup, i. chronakis. phospholipid electrospun nanofibers: effect of solvents and co-axial processing on morphology and fiber diameter. rsc advances 5(66) (2015) 53644–53652. https://doi.org/10.1039/c5ra10498j. [196] l. weng, j. xie. smart electrospun nanofibers for controlled drug release: recent advances and new perspectives. current pharmaceutical design 21(15) (2015) 1944–1959. [197] c. alvarez, a. concheiro. smart drug delivery systems: from fundamentals to the clinic. chemical communications 50(58) (2014) 7743–7765. https://doi.org/10.1039/c4cc01429d. https://doi.org/10.18433/jpps29674 https://doi.org/10.1016/j.ijpharm.2014.12.024 http://dx.doi.org/10.5599/admet.724 https://doi.org/10.1016/j.polymer.2015.04.045 https://doi.org/10.1002/adv.21850 https://doi.org/10.1016/j.ijpharm.2016.01.044 https://doi.org/10.1016/10.12691/ajps-3-5-1 https://doi.org/10.1016/j.jsps.2015.02.024 https://doi.org/10.4236/jbm.2016.41009 https://doi.org/%1f10.1016/j.ijpharm.2014.11.038 https://doi.org/%1f10.1016/j.ijpharm.2014.11.038 https://doi.org/10.1208/s12249-016-0500-8 https://doi.org/10.1038/s41598-018-23378-3 https://doi.org/10.1038/s41598-018-23378-3 https://doi.org/10.1016/j.carbpol.2017.10.071 https://doi.org/10.1016/j.colsurfb.2015.03.058 https://doi.org/10.1016/j.arabjc.2015.11.015 https://doi.org/10.1039/c5ra10498j https://doi.org/10.1039/c4cc01429d admet & dmpk 8(4) (2020) 325-353 electrospun nanofibers: a nanotechnological approach doi: https://dx.doi.org/10.5599/admet.844 351 [198] a. goyal, g. rath, c. faujdar, b. malik. chapter 2 application and perspective of ph-responsive nano drug delivery systems. in: s. mohapatra, s. ranjan, n. dasgupta, r. kumar, s. thomas. applications of targeted nano drugs and delivery systems: nanoscience and nanotechnology in drug delivery, elsevier, 2018, p. 15-33. https://doi.org/10.1016/b978-0-12-814029-1.00002-8. [199] c. dwivedi, h. pandey, c. pandey, p. ramteke. nanofibre based smart pharmaceutical scaffolds for wound repair and regenerations. current pharmaceutical design 22(11) (2016) 1460-1471. https://doi.org/10.2174/1381612822666151215103553. [200] c. yang, d. yu, d. pan, x. liu, x. wang, s. bligh. electrospun ph-sensitive core–shell polymer nanocomposites fabricated using a tri-axial process. acta biomaterialia 35 (2016) 77–86. https://doi.org/10.1016/j.actbio.2016.02.029. [201] h. frizzell, t. ohlsen, k. woodrow. protein-loaded emulsion electrospun fibers optimized for bioactivity retention and ph-controlled release for peroral delivery of biologic therapeutics. international journal of pharmaceutics 533(1) (2017) 99–110. https://doi.org/10.1016/j.ijpharm.2017.09.043. [202] a. akhgari, z. heshmati, h. afrasiabi, f. sadeghi, a. sabbagh, b. sharif. indomethacin electrospun nanofibers for colonic drug delivery: in vitro dissolution studies. colloids and surfaces b: biointerfaces 152 (2017) 29–35. https://doi.org/10.1016/j.colsurfb.2016.12.035. [203] l. liu, s. bai, h. yang, s. li, j. quan, l. zhu. controlled release from thermo-sensitive pnvcl-co-maa electrospun nanofibers: the effects of hydrophilicity/hydrophobicity of a drug. materials science and engineering: c 67 (2016) 581–589. https://doi.org/10.1016/j.msec.2016.05.083. [204] h. li, k. liu, q. sang, g. williams, j. wu, h. wang. a thermosensitive drug delivery system prepared by blend electrospinning. colloids and surfaces b: biointerfaces 159 (2017) 277–283. https://doi.org/10.1016/j.colsurfb.2017.07.058. [205] a. sasikala, a. unnithan, y. yun, c. park, c. kim. an implantable smart magnetic nanofiber device for endoscopic hyperthermia treatment and tumor-triggered controlled drug release. acta biomaterialia 31 (2016) 122–133. https://doi.org/10.1016/j.actbio.2015.12.015. [206] b. wang, h. zheng, m. chang, z. ahmad, j. li. hollow polycaprolactone composite fibers for controlled magnetic responsive antifungal drug release. colloids and surfaces b: biointerfaces 145 (2016) 757–767. https://doi.org/10.1016/j.colsurfb.2016.05.092. [207] m. heidari, h. bahrami, m. ranjbar, p. milan. smart electrospun nanofibers containing pcl/gelatin/graphene oxide for application in nerve tissue engineering. materials science and engineering: c 103 (2019) 109768. https://doi.org/10.1016/j.msec.2019.109768. [208] a. puiggalí, a. cejudo, l. del valle, c. alemán. smart drug delivery from electrospun fibers through electroresponsive polymeric nanoparticles. acs applied bio materials 1(5) (2018) 1594–1605. https://doi.org/10.1021/acsabm.8b00459. [209] h. li, k. liu, g. williams, j. wu, h. wang. dual temperature and ph responsive nanofiber formulations prepared by electrospinning. colloids and surfaces b: biointerfaces 171 (2018) 142–149. https://doi.org/10.1016/j.colsurfb.2018.07.020. [210] t. yadavalli, s. ramasamy, g. chandrasekaran, i. michael, h. therese, r. chennakesavulu. dual responsive pnipam–chitosan targeted magnetic nanopolymers for targeted drug delivery. journal of magnetism and magnetic materials 380 (2015) 315–20. https://doi.org/10.1016/j.jmmm.2014.09.035. [211] c. tran, p. tran, t. tran. ph-independent dissolution enhancement for multiple poorly water-soluble drugs by nano-sized solid dispersions based on hydrophobic–hydrophilic conjugates. drug development and industrial pharmacy 45(3) (2019) 514–519. https://doi.org/10.1080/03639045.2018.1562466. [212] m. chen, y. li, f. besenbacher. electrospun nanofibers-mediated on-demand drug release. advanced healthcare materials 3(11) (2014) 1721–1732. https://doi.org/10.1002/adhm.201400166. [213] s. demirci, a. celebioglu, z. aytac, t. uyar. ph-responsive nanofibers with controlled drug release properties. polymer chemistry 5(6) (2014) 2050–2056. https://doi.org/10.1039/c3py01276j. https://dx.doi.org/10.5599/admet.844 https://doi.org/10.1016/b978-0-12-814029-1.00002-8 https://doi.org/10.2174/1381612822666151215103553 https://doi.org/10.1016/j.actbio.2016.02.029 https://doi.org/10.1016/j.colsurfb.2016.12.035 https://doi.org/10.1016/j.msec.2016.05.083 https://doi.org/10.1016/j.colsurfb.2017.07.058. https://doi.org/10.1016/j.colsurfb.2017.07.058. https://doi.org/10.1016/j.actbio.2015.12.015 https://doi.org/10.1016/j.colsurfb.2016.05.092 https://doi.org/10.1016/j.msec.2019.109768 https://doi.org/10.1021/acsabm.8b00459 https://doi.org/10.1016/j.colsurfb.2018.07.020 https://doi.org/10.1016/j.jmmm.2014.%1f09.035 https://doi.org/10.1016/j.jmmm.2014.%1f09.035 https://doi.org/10.1080/03639045.%1f2018.1%1f562466 https://doi.org/10.1080/03639045.%1f2018.1%1f562466 https://doi.org/10.1002/adhm.201400166 https://doi.org/10.1039/c3py01276j luis castillo et al admet & dmpk 8(4) (2020) 325-353 352 [214] l. liu, w. yao, y. rao, x. lu, j. gao. ph-responsive carriers for oral drug delivery: challenges and opportunities of current platforms. drug delivery 24(1) (2017) 569–581. https://doi.org/10.1080/10717544.2017.1279238. [215] d. ghosh, g. chakrabarti. chapter 6 thermoresponsive drug delivery systems, characterization and application. i in: s. mohapatra, s. ranjan, n. dasgupta, r. kumar, s. thomas. applications of targeted nano drugs and delivery systems: nanoscience and nanotechnology in drug delivery, elsevier, 2018, p. 133-155. https://doi.org/10.1016/b978-0-12-814029-1.00006-5. [216] l. yi, y. wang, y. fang, m. zhang, j. yao, l. wang. development of core–sheath structured smart nanofibers by coaxial electrospinning for thermo-regulated textiles. rsc advances 9(38) (2019) 21844–21851. https://doi.org/10.1039/c9ra03795k. [217] j. hu, h. li, g. williams, h. yang, l. tao, l. zhu. electrospun poly(n-isopropylacrylamide)/ethyl cellulose nanofibers as thermoresponsive drug delivery systems. journal of pharmaceutical sciences 105(3) (2016) 1104–1112. https://doi.org/10.1016/s0022-3549(15)00191-4. [218] x. lin, d. tang, z. yu, q. feng. stimuli-responsive electrospun nanofibers from poly(nisopropylacrylamide)-co-poly(acrylic acid) copolymer and polyurethane. journal of materials chemistry b 2(6) (2014) 651–658. https://doi.org/10.1039/c3tb21519a. [219] s. wu, b. wang, z. ahmad, j. huang, m. chang, j. li. surface modified electrospun porous magnetic hollow fibers using secondary downstream collection solvent contouring. materials letters 204 (2017) 73–76. https://doi.org/10.1016/j.matlet.2017.06.015. [220] j. kim, s. zhu, y. gan, k. forward. magnetic hyperthermia behavior of electrospun polyvinylpyrrolidone (pvp) nanofibers containing magnetic oxide materials. advances in research (2015) 84–91. https://doi.org/10.9734/air/2015/13547. [221] c. mortimer, c. wright. the fabrication of iron oxide nanoparticle-nanofiber composites by electrospinning and their applications in tissue engineering. biotechnol j 12(7) (2017). https://doi.org/10.1002/biot.201600693. [222] s. spadaro, m. santoro, f. barreca, a. scala, s. grimato, f. neri. peg-plga electrospun nanofibrous membranes loaded with au@fe2o3 nanoparticles for drug delivery applications. frontiers in physics 13(1) (2017) 136201. https://doi.org/10.1007/s11467-017-0703-9. [223] h. zhang, j. xia, x. pang, m. zhao, b. wang, l. yang. magnetic nanoparticle-loaded electrospun polymeric nanofibers for tissue engineering. materials science and engineering: c 73 (2017) 537–543. https://doi.org/10.1016/j.msec.2016.12.116. [224] r. meyer, j. green. biodegradable polymer iron oxide nanocomposites: the future of biocompatible magnetism. nanomedicine 10(23) (2015) 3421-34252015. https://doi.org/10.2217/nnm.15.165. [225] j. thévenot, h. oliveira, o. sandre, s. lecommandoux. magnetic responsive polymer composite materials. chem. soc. rev. 42(2013) 7099-7116. https://doi.org/ 10.1039/c3cs60058k. [226] g. filipcsei, i. csetneki, a. szilágyi, m. zrínyi. magnetic field -responsive smart polymer composites. in: oligomers polymer composites molecular imprinting. advances in polymer science. springer, berlin, (2007). https://doi.org/10.1007/12_2006_104. [227] g. chen, j. guo, j. nie, g. ma. preparation, characterization, and application of peo/ha core shell nanofibers based on electric field induced phase separation during electrospinning. polymer 83 (2016) 12–19. https://doi.org/10.1016/j.polymer.2015.12.002. [228] g. xie, y. wang, x. han, y. gong, j. wang, j. zhang. pulsed electric fields on poly-l-(lactic acid) melt electrospun fibers. industrial & engineering chemistry research 55(26) (2016) 7116–7123. https://doi.org/10.1021/acs.iecr.6b00958. [229] q. zhao, w. lu, m. wang. modulating the release of vascular endothelial growth factor by negativevoltage emulsion electrospinning for improved vascular regeneration. materials letters 193 (2017) 1– 4. https://doi.org/10.1016/j.matlet.2017.01.058. [230] r. cheng, f. meng, c. deng, h. klok, z. zhong. dual and multi-stimuli responsive polymeric nanoparticles for programmed site-specific drug delivery. biomaterials 34(14) (2013) 3647–3657. https://doi.org/10.1016/j.biomaterials.2013.01.084. https://doi.org/10.1080/%1f10717544.2017.1279238 https://doi.org/10.1080/%1f10717544.2017.1279238 https://doi.org/10.1016/b978-0-12-814029-1.00006-5 https://doi.org/10.1039/c9ra03795k https://doi.org/10.1016/s0022-3549(15)00191-4 https://doi.org/10.1039/c3tb21519a https://doi.org/10.1016/j.matlet.2017.06.015 https://doi.org/10.9734/air/2015/13547 https://doi.org/10.1002/biot.201600693 https://doi.org/10.1007/s11467-017-0703-9 https://doi.org/10.1016/j.msec.2016.12.116 https://doi.org/10.2217/nnm.15.165 https://doi.org/10.1039/c3cs60058k https://doi.org/10.1007/12_2006_104 https://doi.org/10.1016/j.polymer.2015.12.002 https://doi.org/10.1021/acs.iecr.6b00958 https://doi.org/10.1016/j.matlet.2017.01.058 https://doi.org/10.1016/j.biomaterials.2013.01.084 admet & dmpk 8(4) (2020) 325-353 electrospun nanofibers: a nanotechnological approach doi: https://dx.doi.org/10.5599/admet.844 353 [231] h. yuan, b. li, k. liang, x. lou, y. zhang. regulating drug release from phand temperatureresponsive electrospun cts-g-pnipaam/poly(ethylene oxide) hydrogel nanofibers. biomedical materials 9(5) (2014). https://doi.org/10.1088/1748-6041/9/5/055001. ©2020 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) https://dx.doi.org/10.5599/admet.844 https://doi.org/10.1088/1748-6041/9/5/055001 http://creativecommons.org/licenses/by/3.0/ current state and future of 3d bioprinted models for cardio­vascular research and drug development doi: http://dx.doi.org/10.5599/admet.951 231 admet & dmpk 9(4) (2021) 231-242; doi: https://doi.org/10.5599/admet.951 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index review current state and future of 3d bioprinted models for cardiovascular research and drug development liudmila polonchuk 1 and carmine gentile 2,3 * 1 pharmaceutical sciences, pharma research and early development, roche innovation center basel, f. hoffmann-la roche ltd., basel, switzerland 2 sydney medical school, the university of sydney, australia 3 school of biomedical engineering, university of technology sydney, australia *corresponding author: e-mail: carmine.gentile@uts.edu.au; tel.: +61 (2) 9514 4502 received: may 18, 2021; in the final form: august 16, 2021; available online: august 25, 2021 abstract in the last decade, 3d bioprinting technology has emerged as an innovative tissue engineering approach for regenerative medicine and drug development. this article aims at providing an overview about the most commonly used bioengineered tissues, focusing on 3d bioprinted cardiac cells and how they have been utilized for drug discovery and development. the review describes that, while this field is still developing, cardiovascular research may benefit from laboratory-engineered heart tissues built of specific cell types with precise 3d architecture mimicking the native cardiac microenvironment. it also describes the role played by regulatory agencies and potential commercialization pathways for direct translation from the bench to the bedside of studies using 3d bioprinted cardiac tissues. ©2021 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords bioengineered heart tissues; drug development; advanced in vitro models. introduction cardiovascular disease (cvd) is a leading cause of death worldwide, especially in the aging population. the increased incidence of cvd has been more recently associated with co-morbidity with other chronic diseases, such as kidney failure and type ii diabetes, accounting for nearly one third of all deaths [1-4]. over the past three decades extensive exploration in cardiovascular research to both prevent and treat cvd in patients has resulted in improved survival rates and quality of life [3,5]. commonly practiced therapeutic interventions include drugs and surgical procedures, which have allowed a significant reduction in the mortality of cvd patients, followed by lifestyle changes to complement these primary interventions [2,5]. however, the lack of translation from the bench to the bedside, caused by limitations of currently available in vitro and in vivo models, has prompted a need to engineer novel tools for cardiovascular research. in the past two decades, 3d bioprinting technology has emerged as a potential tool to address limitations of conventional models by generating complex multicellular systems using cells and biomaterials http://dx.doi.org/10.5599/admet.951 https://doi.org/10.5599/admet.951 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:carmine.gentile@uts.edu.au http://creativecommons.org/licenses/by/4.0/ polonshuk and gentile admet & dmpk 9(4) (2021) 231-242 232 that better recapitulate the human heart microenvironment [4,6,7]. more recently, together with the use of stem cells, advances in the field of biofabrication enabled a move toward advanced in vitro models for drug testing. this brief review aims to explore how cardiac bioprinting is best used for development of predictive in vitro tools to speed up development of new medicines and cell therapies, by covering technical and regulatory aspects of this technology. cardiac cells in 3d three-dimensional cell cultures and technologies are rapidly gaining recognition for their potential to model heart tissue pathophysiology [4,8-11]. cells can be grown in scaffolds, scaffold-free or matrices environment aiming to mimic the ecm features of the heart; for example, biomaterial scaffolds, such as collagen and fibrin, provide a 3d environment for cells to attach, interact with each other and conduct electrical signals [4,6,7,12]. cardiac myocytes cultured in 3d often employ a biomaterial such as a hydrogel or biocompatible polymer to mimic the ecm, providing a 3d architecture for cells to interact in all spatial dimensions, both with other cells and their environment. the ability to modify properties such as elasticity, stiffness, conductivity and porosity allows for fine-tuning of the microenvironment [6,12]. these are core aspects of cardiac tissue engineering as the utility of 3d culturing and bioengineering to simulate blood flow, observe contractile forces and relaxation velocity in cardiac myocytes with variable mechanical and electrical cues are the tools necessary to create a complex and accurate microenvironment [6,12,13]. the increase in parameters brings increased complexity and lack of standardized 3d culture protocols results in customized experimental design, which is not compatible with a high-throughput testing. in a recent publication by campostrini et.al. [14], a comprehensive overview of published 3d cardiac models was provided. historically, engineered heart tissues (ehts) have been used as a first approach to replicate the complex human heart microenvironment using patient-specific hipsc-cms to evaluate mutations, drug screening and individual risk of a patient such as drug-induced side effects [15,16]. given their three-dimensional nature, ehts allow the measurement of cell contractility in pathophysiological conditions together with measurements of contraction kinetics, rhythm and rate, genetic and protein analyses, and histological analyses of semi-thin, paraffin or ultrathin sections [17,18]. another popular experimental approach of culturing cardiac cells in spheroids can be applied in several ways, both with and without a scaffold to support the spheroid development in a hanging drop or in low surface adhesion plates [19,20]. a cell-sheets technique utilizes temperature-sensitive surfaces to culture monolayers of cells that can be detached as a sheet of cells and continuously stacked over each other, resulting in a thick sheet of cells [21]. more recently, 3d bioprinting of cardiac tissues emerged as improved in vitro models of human heart pathophysiology [7]. 3d bioprinting of cardiovascular tissues various processes have been applied for bioprinting cardiac constructs (figure 1). biofabrication of cardiac tissues using 3d bioprinting technology has approach-dependent advantages and disadvantages [22]. 3d bioprinting of heart tissues combines additive manufacturing, biomaterials and viable objects to produce structured biological constructs which can be assembled and oriented. this new approach enables designed anisotropy for directed response, controls 3d cell composition and alignment mimicking native myocardium [4,6]. admet & dmpk 9(4) (2021) 231-242 3d bioprinted models for cardio¬vascular research doi: http://dx.doi.org/10.5599/admet.951 233 figure 1. 3d bioprinting technology and its applications in biomedical research. extrusion-based bioprinters dispense cells embedded in hydrogels continuously in a pre-defined shape using either pneumatic or mechanical forces. this technique is the most common and the least expensive. it allows rapid printing times and can print extremely high cell densities. however, material choice is limited as viscous bioinks are required for optimal extrusion through the nozzle, a feature that can affect cell viability. for instance, kolesky et al. [23] bioprinted preformed vascular networks by lining human umbilical vein endothelial cells (huvecs) viable for at least 6 weeks. inkjet bioprinting allows the release of fluid droplets at precise locations via thermal or piezoelectric forces. this method yields high print resolutions as low as 20 µm, is compatible with a large range of bioinks and results in cell viability that can vary depending on the pressure applied. if low pressures are used, delicate cell types can be used, but this comes at the cost of lower structural integrity and, therefore, lower printed cell densities. xu et al. [24] used primary feline and h1 cardiomyocytes on a controlled porosity alginate hydrogel to bioprint viable cell populations, indicating inkjet bioprinting may be useful in engineering designed cardiac tissues. 3d bioprinting using preformed cardiac spheroid cultures has been employed in an effort to use microtissues as building blocks, by stacking each individual spheroid in a needle-based array, free from any hydrogel addition [25]. in http://dx.doi.org/10.5599/admet.951 polonshuk and gentile admet & dmpk 9(4) (2021) 231-242 234 stereolithography, unlike the previous methods, the construct is hardened via photopolymerization from a vessel of fluid containing photoactive polymers. this method is rapid and removes physical stress on cells and bioinks resulting in moderately high resolutions down to 50 µm and high cell viability. however, the number of available photoactive polymer materials is limited, and their use in cell culture is further restricted by an application of ultraviolet light harmful for the cells [26]. stereolithography bioprinting has mainly been used to generate patient-specific models to assist in surgical planning and to create vascularised tissue constructs from photoactive polymers with modifiable elasticity and tensile strength [26]. general schematics of the 3d bioprinting technology, and its application in biomedical field, are shown in figure 1. bioprinted cardiac tissues have been generated to mimic several features of the cardiac microenvironment for both in vitro and in vivo applications. these include modelling for complex diseases, drug screening and potential transplantation to replace or support the regeneration of damaged myocardium. the vascularization of the cardiac tissue, which poses a major problem for long-term survival of cells in bioengineered tissues, has progressed with the advent of bioprinting. this includes generation of vascular networks through several methods including: (i) vascular structures via simultaneous bioprinting of cells and biomaterials; (ii) addition of angiogenic factors in bioprinted constructs; and (iii) bioprinting of channel-based constructs for pre-fabricated vascular networks. previously generated spheroids have been bioprinted, used as building blocks and subsequently fused into vascular constructs with a range of cell types including human smooth muscle cells, human dermal fibroblasts and more importantly cardiac fibroblasts, endothelial cells (ecs) and ipsc-cms [4,6,7,23,25]. inkjet based bioprinters can be utilised to deposit biomaterial scaffolds and ecs simultaneously to form microvasculature scaffolds allowing ec proliferation into tubular structures with clinically relevant cell viabilities and maintained structural integrity of vasculature [27,28]. angiogenic growth factors have been explored in bioprinted constructs with some success. huvecs cultured in vegf before bioprinting with ipsc-cm were reported to integrate with host vasculature when transplanted in mice [29]. furthermore, vegf slowly released into scaffolds (demonstrated with both matrigel and alginate) promotes vessel formation and cd31 expression, which similarly have seen promising results in mice transplants [30]. more physiological constructs require a vascular network that supports flow throughout the entire structure. bioprinting uniquely offers this feature of manufacturing complex and organized networks to enable nutrient delivery for the efficient cardiac function, waste and oxygen transport, all aimed at promoting cell survival and function [4,7]. this can be accomplished by bioprinting a hydrogel containing a removable internal structure, such as sacrificial polymer or spheroids, yielding hollow networks that can be populated with ecs to mimic in vivo vasculature [31-33]. in particular, the fresh method allows integration of thermosensitive hydrogels such as gelatin and pluronics for the generation of hollow structures [34]. additionally, direct bioprinting of perfusable constructs is possible using multiple print-heads containing an outer cross-linkable hydrogel (e.g., gelma) and an inner head with the appropriate cross-linking solution [35]. despite the latest improvements in structural organization of cells within bioprinted cardiac tissues, their optimal vascular network and contractile function remain the major challenges to overcome before expanded use of clinically relevant constructs could be readily available [4]. cardiac spheroids from ipsccms already have been demonstrated to spontaneously contract but are limited by immature phenotypes [9,20,36]. therefore, multitude efforts seek to improve the microenvironment of ipsc-cms and the tissues they are grown within to increase cell contractility and contribute to overall cardiac tissue development [37]. among these strategies, the cardiac environment has been improved by the addition of conductive polymers for electrical propagation support and elastic polymers for mechanical support [38-40]. though admet & dmpk 9(4) (2021) 231-242 3d bioprinted models for cardio¬vascular research doi: http://dx.doi.org/10.5599/admet.951 235 these biomaterials have been used in cardiac tissue engineering previously, optimizing such biomaterials for bioprinting is feasible and slowly progressing, yet still requires further studies to improve the effects on biocompatibility [4]. application of 3d bioprinting in cardiac regeneration and drug testing the heart has a limited capacity to regenerate after birth, and the major focus of cardiac bioprinting has been the generation of tissues that may represent an alternative approach to regenerate infarcted heart tissue by integrating cardiac cells in presence of additional biomaterials [4,41]. the application of bioprinted cardiac tissues potentiates the creation of functional cardiac tissue to regenerate or replaced damaged tissue in the myocardium [37,41,42]. in a similar effort to the one used by other approaches aimed at bioengineering the cardiac tissue, these bioprinted tissues better mimic structural, physiological, and functional features of the native myocardium, all features that can be applied for drug testing, toxicity assays and disease modelling (e.g., myocardial infarction and heart failure [7,29,43]). this approach allows the generation of a complex 3d structure permissive for tightly-regulated molecular and cellular interactions, by supporting cells and enhancing their structural reorganisation into functional cardiac tissues [44]. optimal in vitro maturation and functional testing (including measurements of durability of tissues and their contractile function) before bioprinted tissues can be transplanted onto the defected heart is critical [4,6]. among the several tests, cell viability assays, vascular network evaluation and opticalelectrical mapping tests are required for quality control and to characterize their structural, mechanical and electrical properties [10,25,29,45]. bioprinting holds promise as a potential clinical tool to enable production of vascularized tissues with nutrient, waste and oxygen transport [46]. for example, jang et al. [47] developed a pre-vascularized 3d bioprinted myocardial tissue using bioinks containing either cardiac progenitor cells (cpcs) or human dermal tissue-derived microvascular endothelial cells (ecs) together with turbinate mesenchymal stem cells (mscs). cells were dispersed in a human decellularized extracellular matrix (decm) within the bioinks. in this study, the two bioinks were used following an alternative patterning design as dual cell threads, which facilitated the vascularization of the patch, as tested in an in vivo rat model of heart failure. however, in this study a branched endothelial cell network typical of the human myocardial tissue could not be established, remaining one of the major challenges for the 3d bioprinting of viable and functional cardiac tissues [4]. in a study by mirdamadi et al. [48] using the fresh printing technique described above, whole-size heart constructs were 3d printed by embedding soft biomaterials in a thermoreversible support bath. this study expanded the printable size range by fresh printing to a full-size model of an adult human heart from patient-derived magnetic resonance imaging (mri) data sets. alginate hydrogels were used as the printing biomaterial to mimic the elastic modulus of the human cardiac tissue. fresh-printed alginate proved to be a high print fidelity low-cost platform to create customized models for surgical training and planning. future studies including cells will be required to demonstrate the feasibility of this approach for cardiac tissue engineering purposes. even when not 3d bioprinted, cardiac tissues derived from hipsc-cms embedded with extracellular matrix proteins (e.g., fibronectin and gelatin nanofilm) present the potential to be used as a screening system for drug discovery and cardiotoxicity assay [49]. however, in addition to biological factors, a major challenge of in vitro cardiac tissues is their ability to respond to mechanical and electrical simulations in 3d [28]. in our latest study, we developed a novel approach for the biofabrication of 3d bioprinted heart tissues using bioinks containing alginate/gelatin (al/ge) hydrogels and 3d cardiac spheroids (css) [50]. http://dx.doi.org/10.5599/admet.951 polonshuk and gentile admet & dmpk 9(4) (2021) 231-242 236 cardiac spheroids were first created from human cardiac myocytes, fibroblasts and endothelial cells and then mixed within optimal al/ge hydrogels before being tested for viability and contractile function in vitro. when 3d bioprinted in a cardioexcyte 96-well microelectrode plate, the response of 3d bioprinted css to isoproterenol treatment was recorded. al/ge hydrogels enabled cs adherence to the electrode and fusion, which was further facilitated by addition of vascular endothelial growth factor (vegf). bioprinted css contracted spontaneously and under electrical stimulation, allowing us to record contractile and electrical signals on the microelectrode plates. the ability of bioprinted css to respond to drug treatments in 3d makes them suitable candidates as high-throughput assays for industrial applications. another approach to vascularize 3d tissues includes the use of microfluidic “organ-on-a-chip” devices, which provide better control over the microenvironment by mimicking various in vivo mechanical and chemical functions of the heart and can be used to monitor parameters, such as ph, nutrient supply and oxygen level [28]. microfluidics devices structure and maintain cellular morphology and cell-specific functionality of the 3d biofabricated tissue, while regulating shear stress forces through the flow rate [51]. in particular, they have also been used to promote endothelial cell network for optimal cell viability and function [52]. undoubtably, organ-on-a-chip models have advanced our way of using cardiac cells in vitro, in a way to improve their use for drug screening and to improve treatment efficacy. however, more research is required to bioengineer a device that fully recapitulates the microenvironment typical of the cardiac tissue, including combinations of 3d bioprinting and microfluidics devices. while the use of 3d cardiac bioprinting for direct clinical applications remains at an experimental stage, there are signs that the technology is slowly progressing towards more end-use applications in development of advanced cellular models for drug testing. current regulatory mandated assays for cardiac safety either do not capture the complexity of drug-induced cardiotoxicity or as with in vivo animal studies cannot be extrapolated to humans. consequently, the prediction of drug-induced cardiotoxicity remains a constant challenge for pharma-industry, and accounts for up to 33 % of drug failure due to drug withdrawal during clinical development [53]. in an effort to make cardiotoxicity testing more biologically-relevant and therefore translatable, the us food and drug administration (fda)-sponsored cardiac safety research consortium and the not-for-profit health and environmental sciences institute (hesi) have recommended incorporating hipsc-cms into the nonclinical drug cardiotoxicity assessment as an integrating model system bridging single receptor-based in vitro results together with clinical data [54]. the strategy that integrates human ips-cardiomyocytes engineered into a 3d system mimicking cardiac microenvironment offers an opportunity to develop robust, physiologically relevant in vitro functional assays, with the potential to better predict clinical outcomes, reduce animal usage, speed up decision making and lower development costs. engagement of public private partnerships for transition from the bench to the bedside in the process of generating more physiologically relevant in vitro models, regulatory agencies are preparing for ramifications of biofabricated tissue products (figure 2). however, current fda guidance primarily addresses technical aspects of devices used for 3d printing methods and products [55]. lack of regulatory standards makes it difficult to commercialize novel bioprinting technologies. public acceptance of in vitro 3d bioprinted models needs to be preceded by regulatory policies that guide technology companies within the field. ideally, they should work closely with regulatory agencies and potential customers to ensure product validation for industrial use. eventually, the purpose and the mechanistic question will determine the model to be utilized for specific applications. a recent fda white paper coauthored by academic and industrial experts outlined the general validation principles for the models to be admet & dmpk 9(4) (2021) 231-242 3d bioprinted models for cardio¬vascular research doi: http://dx.doi.org/10.5599/admet.951 237 used in the field of cardiac safety [56]. advances in the regulatory process may significantly incentivize maturation and commercialization of 3d bioprinting research in general. and government organizations could foster advancements in fundamental bioprinting technologies and potentially support early investments in this field. they can help demonstrating the utility and disseminate the use of this emerging technology into a variety of communities including academia, government, regulators, ngos, and industry. well-balanced interaction between all the stakeholders, schematically illustrated in figure 2 is essential to assure progress and foster growth on a global scale. agencies such as the us national science foundation (nsf) and department of defense (dod) awarded multiple grants to institutions dedicated to studying and developing biomaterials specifically for use in bioprinting applications over the past decade. figure 2. stakeholder’s communication scheme. schematic showing key roles and how information flow and relationships contribute to new model development. big healthcare companies followed the trend and set up collaborations with bioprinting companies. for instance, astrazeneca has established the bioventurehub, which provides emerging life science companies and academic teams access to astrazeneca’s experience and infrastructure. cellink, a global leader in bioprinting, has already established an r&d lab in bioventurehub to biofabricate kidney, heart, liver, and lung tissue models for pharmaceutical development. likewise, j&j announced the creation of a bioprinting research laboratory as part of the 3d printing center of excellence in collaboration with the advanced materials and bioengineering research (amber) institute in dublin. astellas pharma, bristol-meyers squibb, merck, novartis, procter and gamble, roche also have bioprinting programs as do some large research facilities around the world, such as the national institutes of health (nih) in the us. in australia, http://dx.doi.org/10.5599/admet.951 polonshuk and gentile admet & dmpk 9(4) (2021) 231-242 238 despite the lack of big national pharmaceutical companies, there has been a focus on emerging bioprinting technologies, with several centers supported by the government (i.e., the arc training centre in wollongong, nsw, the arc centre of excellence for electromaterial science). the emerging interest in this field is also highlighted by the recent bioprinting workshops organised by the university of technology sydney (uts) in partnerships with global industry partners [57]. abovementioned workshops and platforms for continuous education and information sharing may represent the first step in effectively bringing together future partners and catalysing public-private partnerships [57]. in the long run, innovation labs and start up incubators in commercial and public healthcare sectors would be the most effective way to develop the research idea into a commercial product. therefore, an international network composed of experts from academia, industry and regulatory bodies should facilitate the frameworks to promote fast and efficient bench-to-bedside technology transfer. conclusions and future directions emerging 3d bioprinting technologies offer therapeutic strategies to better combat heart diseases, as well as open the door for improved drug testing models in drug development. they hold much promise in personalized and regenerative medicine, particularly with the opportunity of using patient-specific cells. they also uniquely allow to recapitulate the 3d cardiac tissue architecture required for functional human myocardium. development of cardiac-specific bioinks and efficient printing methods for high throughput assays would enable the fabrication of large‐sized engineered tissue models. a faster adoption, commercialization and validation of 3d bioprinted cardiac tissues will depend on coherent strategic partnerships between academia, regulators and industry and will benefit patients and society. acknowledgements: cg is supported by a university of sydney kick-start grant, cdip grant, cardiothoracic surgery research grant, uts seed funding and catholic archdiocese of sydney grant for adult stem cell research. conflict of interest: the authors declare that the review was prepared in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. references [1] h. wang, m. naghavi, c. allen et al. global, regional, and national life expectancy, all-cause mortality, and cause-specific mortality for 249 causes of death, 1980-2015: a systematic analysis for the global burden of disease study 2015. lancet 388 (2016) 1459-1544. https://doi.org/10.1016/s01406736(16)31012-1. [2] g.a. roth, c. johnson, a. abajobir, et al. global, regional, and national burden of cardiovascular diseases for 10 causes, 1990 to 2015. j. am. coll. cardiol. 70 (2017) 1-25. https://doi.org/10.1016/j.jacc.2017.04.052. [3] h. thomas, j. diamond, a. vieco, s. chaudhuri, e. shinnar, s. cromer, p. perel, g.a. mensah, j. narula, c.o. johnson, g.a. roth, a.e. moran. global atlas of cardiovascular disease 2000-2016: the path to prevention and control. glob. heart 13 (2018) 143-163. http://doi.org/10.1016/j.gheart.2018.09.511. [4] c.d. roche, r.j.l. brereton, a.w. ashton, c. jackson, c. gentile. current challenges in threedimensional bioprinting heart tissues for cardiac surgery. european journal of cardio-thoracic surgery 58 (2020) 500-510. https://doi.org/10.1093/ejcts/ezaa093. [5] p.a. heidenreich, j.g. trogdon, o.a. khavjou, j. butler, k. dracup, m.d. ezekowitz, e.a. finkelstein, y. hong, s.c. johnston, a. khera, d.m. lloyd-jones, s.a. nelson, g. nichol, d. orenstein, p.w. wilson, y.j. woo, c. american heart association advocacy coordinating, c. stroke, r. council on cardiovascular, intervention, c. council on clinical, e. council on, prevention, a. council on, thrombosis, b. vascular, c. council on, c. critical, perioperative, resuscitation, n. council on cardiovascular, d. council on the kidney in cardiovascular, s. council on cardiovascular, anesthesia, https://doi.org/10.1016/s0140-6736(16)31012-1 https://doi.org/10.1016/s0140-6736(16)31012-1 https://doi.org/10.1016/j.jacc.2017.04.052 https://doi.org/10.1016/j.jacc.2017.04.052 http://doi.org/10.1016/j.gheart.2018.09.511 https://doi.org/10.1093/ejcts/ezaa093 admet & dmpk 9(4) (2021) 231-242 3d bioprinted models for cardio¬vascular research doi: http://dx.doi.org/10.5599/admet.951 239 c. interdisciplinary council on quality of, r. outcomes. forecasting the future of cardiovascular disease in the united states: a policy statement from the american heart association. circulation 123 (2011) 933-944. https://doi.org/10.1161/cir.0b013e31820a55f5. [6] c.d. roche, p. sharma, a.w. ashton, c. jackson, m. xue, c. gentile. printability, durability, contractility and vascular network formation in 3d bioprinted cardiac endothelial cells using alginate–gelatin hydrogels. frontiers in bioengineering and biotechnology 9 (2021) 636257. https://doi.org/10.3389/fbioe.2021.636257. [7] p. sharma, x. wang, c.l.c. ming, l. vettori, g. figtree, a. boyle, c. gentile. considerations for the bioengineering of advanced cardiac in vitro models of myocardial infarction. small (2021) e2003765. https://doi.org/10.1002/smll.202003765. [8] j. gunter, p. wolint, a. bopp, j. steiger, e. cambria, s.p. hoerstrup, m.y. emmert. microtissues in cardiovascular medicine: regenerative potential based on a 3d microenvironment. stem cells int. 2016 (2016) 9098523. https://doi.org/10.1155/2016/9098523. [9] c. gentile. filling the gaps between the in vivo and in vitro microenvironment: engineering of spheroids for stem cell technology. curr. stem cell res. ther. 11 (2016) 652-665. https://doi.org/10.2174/1574888x10666151001114848. [10] j. jang. 3d bioprinting and in vitro cardiovascular tissue modeling. bioengineering (basel) 4 (2017) 71. https://doi.org/10.3390/bioengineering4030071. [11] k. duval, h. grover, l.h. han, y. mou, a.f. pegoraro, j. fredberg, z. chen. modeling physiological events in 2d vs. 3d cell culture. physiology (bethesda) 32 (2017) 266-277. https://doi.org/10.1152/physiol.00036.2016. [12] g. vunjak novaković, t. eschenhagen, c. mummery. myocardial tissue engineering: in vitro models. cold spring harb. perspect. med. 4 (2014) a014076. https://doi.org/10.1101/cshperspect.a014076. [13] y. wang, j.a. hill. electrophysiological remodeling in heart failure. j. mol. cell cardiol. 48 (2010) 619632. https://doi.org/10.1016/j.yjmcc.2010.01.009. [14] g. campostrini, v. meraviglia, e. giacomelli, r.w.j. van helden, l. yiangou, r.p. davis, m. bellin, v.v. orlova, c.l. mummery. generation, functional analysis and applications of isogenic threedimensional self-aggregating cardiac microtissues from human pluripotent stem cells. nat. protoc. 16 (2021) 2213-2256. https://doi.org/10.1038/s41596-021-00497-2. [15] w.h. zimmermann, i. melnychenko, g. wasmeier, m. didie, h. naito, u. nixdorff, a. hess, l. budinsky, k. brune, b. michaelis, s. dhein, a. schwoerer, h. ehmke, t. eschenhagen. engineered heart tissue grafts improve systolic and diastolic function in infarcted rat hearts. nat. med. 12 (2006) 452458. https://doi.org/10.1038/nm1394. [16] i. goldfracht, y. efraim, r. shinnawi, e. kovalev, i. huber, a. gepstein, g. arbel, n. shaheen, m. tiburcy, w.h. zimmermann, m. machluf, l. gepstein. engineered heart tissue models from hipscderived cardiomyocytes and cardiac ecm for disease modeling and drug testing applications. acta biomater. 92 (2019) 145-159. https://doi.org/10.1016/j.actbio.2019.05.016. [17] m.n. hirt, n.a. sorensen, l.m. bartholdt, j. boeddinghaus, s. schaaf, a. eder, i. vollert, a. stohr, t. schulze, a. witten, m. stoll, a. hansen, t. eschenhagen. increased afterload induces pathological cardiac hypertrophy: a new in vitro model. basic res. cardiol. 107 (2012) 307. https://doi.org/10.1007/s00395-012-0307-z. [18] a. eder, i. vollert, a. hansen, t. eschenhagen. human engineered heart tissue as a model system for drug testing. adv. drug deliv. rev. 96 (2016) 214-224. https://doi.org/10.1016/j.addr.2015.05.010. [19] c. zuppinger. 3d culture for cardiac cells. biochim. biophys. acta 1863 (2016) 1873-1881. https://doi.org/10.1016/j.bbamcr.2015.11.036. [20] l. polonchuk, m. chabria, l. badi, j.c. hoflack, g. figtree, m.j. davies, c. gentile. cardiac spheroids as promising in vitro models to study the human heart microenvironment. sci. rep. 7 (2017) 7005. https://doi.org/10.1038/s41598-017-06385-8. http://dx.doi.org/10.5599/admet.951 https://doi.org/10.1161/cir.0b013e31820a55f5 https://doi.org/10.3389/fbioe.2021.636257 https://doi.org/10.1002/smll.202003765 https://doi.org/10.1155/2016/9098523 http://dx.doi.org/10.2174/1574888x10666151001114848 http://dx.doi.org/10.2174/1574888x10666151001114848 https://dx.doi.org/10.3390%2fbioengineering4030071 https://doi.org/10.1152/physiol.00036.2016 https://doi.org/10.1152/physiol.00036.2016 https://doi.org/10.1101/cshperspect.a014076 https://doi.org/10.1016/j.yjmcc.2010.01.009 https://doi.org/10.1038/s41596-021-00497-2 https://doi.org/10.1038/nm1394 https://doi.org/10.1016/j.actbio.2019.05.016 https://doi.org/10.1007/s00395-012-0307-z https://doi.org/10.1007/s00395-012-0307-z https://doi.org/10.1016/j.addr.2015.05.010 https://doi.org/10.1016/j.bbamcr.2015.11.036 https://dx.doi.org/10.1038%2fs41598-017-06385-8 polonshuk and gentile admet & dmpk 9(4) (2021) 231-242 240 [21] k. sakaguchi, t. shimizu, s. horaguchi, h. sekine, m. yamato, m. umezu, t. okano. in vitro engineering of vascularized tissue surrogates. sci. rep. 3 (2013) 1316. https://doi.org/10.1038/srep01316. [22] m. mahmoudi, m. yu, v. serpooshan, j.c. wu, r. langer, r.t. lee, j.m. karp, o.c. farokhzad. multiscale technologies for treatment of ischemic cardiomyopathy. nat. nanotechnol. 12 (2017) 845855. https://doi.org/10.1038/nnano.2017.167. [23] d.b. kolesky, k.a. homan, m.a. skylar-scott, j.a. lewis. three-dimensional bioprinting of thick vascularized tissues. proc. natl. acad. sci. u. s. a. 113 (2016) 3179-3184. https://doi.org/10.1073/pnas.1521342113 [24] t. xu, c. baicu, m. aho, m. zile, t. boland. fabrication and characterization of bio-engineered cardiac pseudo tissues. biofabrication 1 (2009) 035001. https:/doi.org/10.1088/1758-5082/1/3/035001. [25] c.s. ong, t. fukunishi, a. nashed, a. blazeski, h. zhang, s. hardy, d. disilvestre, l. vricella, j. conte, l. tung, g. tomaselli, n. hibino. creation of cardiac tissue exhibiting mechanical integration of spheroids using 3d bioprinting. j. vis. exp. (2017). https://doi.org/10.3791/55438. [26] b. zhang, m. radišić. organ-level vascularization: the mars mission of bioengineering. j. thorac. cardiovasc. surg. 159 (2020) 2003-2007. https://doi.org/10.1016/j.jtcvs.2019.08.128. [27] n. puluca, s. lee, s. doppler, a. munsterer, m. dressen, m. krane, s.m. wu. bioprinting approaches to engineering vascularized 3d cardiac tissues. curr. cardiol. rep. 21 (2019) 90. https://doi.org/10.1007/s11886-019-1179-8. [28] m. alonzo, s. anilkumar, b. roman, n. tasnim, b. joddar. 3d bioprinting of cardiac tissue and cardiac stem cell therapy. transl. res. 211 (2019) 64-83. https://doi.org/10.1016/j.trsl.2019.04.004. [29] f. maiullari, m. costantini, m. milan, v. pace, m. chirivi, s. maiullari, a. rainer, d. baci, h.e. marei, d. seliktar, c. gargioli, c. bearzi, r. rizzi. a multi-cellular 3d bioprinting approach for vascularized heart tissue engineering based on huvecs and ipsc-derived cardiomyocytes. sci. rep. 8 (2018) 13532. https://doi.org/10.1038/s41598-018-31848-x. [30] m.a. kuss, r. harms, s. wu, y. wang, j.b. untrauer, m.a. carlson, b. duan. short-term hypoxic preconditioning promotes prevascularization in 3d bioprinted bone constructs with stromal vascular fraction derived cells. rsc adv. 7 (2017) 29312-29320. https://doi.org/10.1039/c7ra04372d. [31] l.e. bertassoni, j.c. cardoso, v. manoharan, a.l. cristino, n.s. bhise, w.a. araujo, p. zorlutuna, n.e. vrana, a.m. ghaemmaghami, m.r. dokmeci, a. khademhosseini. direct-write bioprinting of cellladen methacrylated gelatin hydrogels. biofabrication 6 (2014) 024105. https://doi.org/10.1088/1758-5082/6/2/024105. [32] j. liu, h. zheng, p.s. poh, h.g. machens, a.f. schilling. hydrogels for engineering of perfusable vascular networks. int. j. mol. sci. 16 (2015) 15997-16016. https://doi.org/10.3390/ijms160715997. [33] a. robu, v. mironov, a. neagu. using sacrificial cell spheroids for the bioprinting of perfusable 3d tissue and organ constructs: a computational study. comput. math. methods med. 2019 (2019) 7853586. https://doi.org/10.1155/2019/7853586. [34] t.j. hinton, q. jallerat, r.n. palchesko, j.h. park, m.s. grodzicki, h.j. shue, m.h. ramadan, a.r. hudson, a.w. feinberg. three-dimensional printing of complex biological structures by freeform reversible embedding of suspended hydrogels. sci. adv. 1 (2015) e1500758. https://doi.org/10.1126/sciadv.1500758. [35] t. liu, w. weng, y. zhang, x. sun, h. yang. applications of gelatin methacryloyl (gelma) hydrogels in microfluidic technique-assisted tissue engineering. molecules 25 (2020) 5305. https://doi.org/10.3390/molecules25225305. [36] m.j. birket, c.l. mummery. pluripotent stem cell derived cardiovascular progenitors--a developmental perspective. dev. biol. 400 (2015) 169-179. https://doi.org/10.1016/j.ydbio.2015.01.012. [37] z. wang, s.j. lee, h.j. cheng, j.j. yoo, a. atala. 3d bioprinted functional and contractile cardiac tissue constructs. acta biomater. 70 (2018) 48-56. https://doi.org/10.1016/j.actbio.2018.02.007. https://doi.org/10.1038/srep01316 https://doi.org/10.1038/srep01316 https://doi.org/10.1038/nnano.2017.167 https://doi.org/10.1073/pnas.1521342113 https://doi.org/10.1073/pnas.1521342113 https://doi.org/10.1088/1758-5082/1/3/035001 https://dx.doi.org/10.3791/55438 https://doi.org/10.1016/j.jtcvs.2019.08.128 https://doi.org/10.1007/s11886-019-1179-8 https://doi.org/10.1007/s11886-019-1179-8 https://doi.org/10.1016/j.trsl.2019.04.004 https://doi.org/10.1038/s41598-018-31848-x https://doi.org/10.1039/c7ra04372d https://doi.org/10.1088/1758-5082/6/2/024105 https://doi.org/10.1088/1758-5082/6/2/024105 https://dx.doi.org/10.3390%2fijms160715997 https://doi.org/10.1155/2019/7853586 https://doi.org/10.1126/sciadv.1500758 https://doi.org/10.1126/sciadv.1500758 https://doi.org/10.3390/molecules25225305 https://doi.org/10.3390/molecules25225305 https://doi.org/10.1016/j.ydbio.2015.01.012 https://doi.org/10.1016/j.ydbio.2015.01.012 https://doi.org/10.1016/j.actbio.2018.02.007 admet & dmpk 9(4) (2021) 231-242 3d bioprinted models for cardio¬vascular research doi: http://dx.doi.org/10.5599/admet.951 241 [38] l. jiang, c. gentile, a. lauto, c. cui, y. song, t. romeo, s.m. silva, o. tang, p. sharma, g. figtree, j.j. gooding, d. mawad. versatile fabrication approach of conductive hydrogels via copolymerization with vinyl monomers. acs appl. mater. interfaces 9 (2017) 44124-44133. https://doi.org/10.1021/acsami.7b15019. [39] d. mawad, g. figtree, c. gentile. current technologies based on the knowledge of the stem cells microenvironments. adv. exp. med. biol. 1041 (2017) 245-262. https://doi.org/10.1007/978-3-31969194-7_13. [40] l. davenport huyer, b. zhang, a. korolj, m. montgomery, s. drecun, g. conant, y. zhao, l. reis, m. radišić. highly elastic and moldable polyester biomaterial for cardiac tissue engineering applications. acs biomater. sci. eng. 2 (2016) 780-788. https://doi.org/10.1021/acsbiomaterials.5b00525 [41] c.d. roche, c. gentile. transplantation of a 3d bioprinted patch in a murine model of myocardial infarction. j. vis. exp. (2020). https://doi.org/10.3791/61675. [42] n. noor, a. shapira, r. edri, i. gal, l. wertheim, t. dvir. 3d printing of personalized thick and perfusable cardiac patches and hearts. adv. sci. (weinh) 6 (2019) 1900344. https://doi.org/10.1002/advs.201900344. [43] h. cui, s. miao, t. esworthy, x. zhou, s.j. lee, c. liu, z.x. yu, j.p. fisher, m. mohiuddin, l.g. zhang. 3d bioprinting for cardiovascular regeneration and pharmacology. adv. drug deliv. rev. 132 (2018) 252269. https://doi.org/10.1016/j.addr.2018.07.014. [44] r.k. birla, s.k. williams. 3d bioprinting and its potential impact on cardiac failure treatment: an industry perspective. apl bioeng. 4 (2020) 010903. https://doi.org/10.1063/1.5128371. [45] a. tijore, s.a. irvine, u. sarig, p. mhaisalkar, v. baisane, s. venkatraman. contact guidance for cardiac tissue engineering using 3d bioprinted gelatin patterned hydrogel. biofabrication 10 (2018) 025003. https://doi.org/10.1088/1758-5090/aaa15d. [46] w. jia, p.s. gungor-ozkerim, y.s. zhang, k. yue, k. zhu, w. liu, q. pi, b. byambaa, m.r. dokmeci, s.r. shin, a. khademhosseini. direct 3d bioprinting of perfusable vascular constructs using a blend bioink. biomaterials 106 (2016) 58-68. https://doi.org/10.1016/j.biomaterials.2016.07.038. [47] j. jang, h.j. park, s.w. kim, h. kim, j.y. park, s.j. na, h.j. kim, m.n. park, s.h. choi, s.h. park, s.w. kim, s.m. kwon, p.j. kim, d.w. cho. 3d printed complex tissue construct using stem cell-laden decellularized extracellular matrix bioinks for cardiac repair. biomaterials 112 (2017) 264-274. [48] e. mirdamadi, j.w. tashman, d.j. shiwarski, r.n. palchesko, a.w. feinberg. fresh 3d bioprinting a full-size model of the human heart. acs biomater. sci. eng. 6 (2020) 6453-6459. https://doi.org/10.1016/j.biomaterials.2016.10.026. [49] m. takeda, s. miyagawa, s. fukushima, a. saito, e. ito, a. harada, r. matsuura, h. iseoka, n. sougawa, n. mochizuki-oda, m. matsusaki, m. akashi, y. sawa. development of in vitro druginduced cardiotoxicity assay by using three-dimensional cardiac tissues derived from human induced pluripotent stem cells. tissue eng. part c methods 24 (2018) 56-67. https://doi.org/10.1089/ten.tec.2017.0247. [50] l. polonchuk, l. surija, m.h. lee, p. sharma, c. liu chung ming, f. richter, e. ben-sefer, m. alsadat rad, h. mahmodi sheikh sarmast, w. al shamery, h.a. tran, l. vettori, f. haeusermann, e.c. filipe, j. rnjak-kovacina, t. cox, j. tipper, i. kabakova, c. gentile. towards engineering heart tissues from bioprinted cardiac spheroids. biofabrication 13 (2021) 045009. https://doi.org/10.1088/17585090/ac14ca. [51] j. paez-mayorga, g. hernandez-vargas, g.u. ruiz-esparza, h.m.n. iqbal, x. wang, y.s. zhang, r. parrasaldivar, a. khademhosseini. bioreactors for cardiac tissue engineering. adv. healthc. mater. 8 (2019) e1701504. https://doi.org/10.1002/adhm.201701504. [52] y.s. zhang, a. arneri, s. bersini, s.r. shin, k. zhu, z. goli-malekabadi, j. aleman, c. colosi, f. busignani, v. dell'erba, c. bishop, t. shupe, d. demarchi, m. moretti, m. rasponi, m.r. dokmeci, a. atala, a. khademhosseini. bioprinting 3d microfibrous scaffolds for engineering endothelialized myocardium and heart-on-a-chip. biomaterials 110 (2016) 45-59. https://doi.org/10.1016/j.biomaterials.2016.09.003. http://dx.doi.org/10.5599/admet.951 https://doi.org/10.1021/acsami.7b15019 https://doi.org/10.1021/acsami.7b15019 https://doi.org/10.1007/978-3-319-69194-7_13 https://doi.org/10.1007/978-3-319-69194-7_13 https://doi.org/10.1021/acsbiomaterials.5b00525 https://doi.org/10.1021/acsbiomaterials.5b00525 https://doi.org/10.3791/61675 https://doi.org/10.1002/advs.201900344 https://doi.org/10.1002/advs.201900344 https://doi.org/10.1016/j.addr.2018.07.014 https://doi.org/10.1063/1.5128371 https://doi.org/10.1088/1758-5090/aaa15d https://doi.org/10.1016/j.biomaterials.2016.07.038 https://doi.org/10.1016/j.biomaterials.2016.10.026 https://doi.org/10.1016/j.biomaterials.2016.10.026 https://doi.org/10.1089/ten.tec.2017.0247 https://doi.org/10.1089/ten.tec.2017.0247 https://doi.org/10.1088/1758-5090/ac14ca https://doi.org/10.1088/1758-5090/ac14ca https://doi.org/10.1002/adhm.201701504 https://doi.org/10.1016/j.biomaterials.2016.09.003 https://doi.org/10.1016/j.biomaterials.2016.09.003 polonshuk and gentile admet & dmpk 9(4) (2021) 231-242 242 [53] j.s. macdonald, r.t. robertson. toxicity testing in the 21st century: a view from the pharmaceutical industry. toxicol. sci. 110 (2009) 40-46. https://doi.org/10.1093/toxsci/kfp088. [54] k. blinova, j. stohlman, j. vicente, d. chan, l. johannesen, m.p. hortigon-vinagre, v. zamora, g. smith, w.j. crumb, l. pang, b. lyn-cook, j. ross, m. brock, s. chvatal, d. millard, l. galeotti, n. stockbridge, d.g. strauss. comprehensive translational assessment of human-induced pluripotent stem cell derived cardiomyocytes for evaluating drug-induced arrhythmias. toxicol. sci. 155 (2017) 234-247. https://doi.org/10.1093/toxsci/kfw200. [55] technical considerations for additive manufactured medical devices, guidance for industry and food and drug administration staff. available online at: https://www.fda.gov/media/97633/download. (2017). [56] z. li, g.r. mirams, t. yoshinaga, b.j. ridder, et al. general principles for the validation of proarrhythmia risk prediction models: an extension of the cipa in silico strategy. clin. pharmacol. ther. 107 (2020) 102-111. https://doi.org/10.1002/cpt.1647. [57] w. harley, h. yoshie, c. gentile. three-dimensional bioprinting for tissue engineering and regenerative medicine in down under: 2020 australian workshop summary. asaio j 67 (2021) 363369. https://doi.org/10.1097/mat.0000000000001389. ©2021 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) https://doi.org/10.1093/toxsci/kfp088 https://doi.org/10.1093/toxsci/kfw200 https://www.fda.gov/media/97633/download https://www.fda.gov/media/97633/download https://doi.org/10.1002/cpt.1647 https://doi.org/10.1097/mat.0000000000001389 http://creativecommons.org/licenses/by/3.0/ lost in modelling and simulation? doi: https:/doi.org/10.5599/admet.923 75 admet & dmpk 9(2) (2021) 75-109; doi: https://doi.org/10.5599/admet.923 open access : issn : 1848-7718 http://www.pub.iapchem.org/ojs/index.php/admet/index feature article lost in modelling and simulation? kiyohiko sugano molecular pharmaceutics lab., college of pharmaceutical sciences, ritsumeikan university, 1-1-1, noji-higashi, kusatsu, shiga 525-8577, japan e-mail: suganok@fc.ritsumei.ac.jp; tel.: +81-77-561-2773 received: november 06, 2021; revised: march 05, 2021; published: march 22, 2021 abstract over the past few decades, physiologically-based pharmacokinetic modelling (pbpk) has been anticipated to be a powerful tool to improve the productivity of drug discovery and development. however, recently, multiple systematic evaluation studies independently suggested that the predictive power of current oral absorption (oa) pbpk models needs significant improvement. there is some disagreement between the industry and regulators about the credibility of oa pbpk modelling. recently, the editorial board of amdet&dmpk has announced the policy for the articles related to pbpk modelling (modelling and simulation ethics). in this feature article, the background of this policy is explained: (1) requirements for scientific writing of pbpk modelling, (2) scientific literacy for pbpk modelling, and (3) middle-out approaches. pbpk models are a useful tool if used correctly. this article will hopefully help advance the science of oa pbpk models. ©2021 by the authors. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/4.0/). keywords physiologically-based pharmacokinetic modelling; scientific literacy; transparency; reproducibility; middleout approach; structural identifiability introduction over the past few decades, physiologically-based pharmacokinetic modelling (pbpk) has been anticipated to be a powerful tool to improve the productivity of drug discovery and development. several sophisticated software products have been commercialized since the late 1990s. plenty of case study reports have been published in peer-reviewed journals, showing nearly perfect prediction, prediction error being much smaller than the variation in the clinical plasma concentration (cp) time profile. it seems that we already have achieved a “prediction paradise”[1]. …really? recently, multiple systematic evaluation studies independently suggested that the “bottom-up” predictive power of current oral absorption (oa) pbpk models needs significant improvement [2–6]. almost all case studies had to use parameter optimization on a drug-by-drug basis to fit the simulated plasma concentration (cp) time curve to clinical data (part 3, section 3.4). expert scientists continue hard experimental works to better understand in vivo systems and improve the predictive performance of in vitro systems [7–9]. there is some disagreement between the industry and regulators about the credibility of pbpk modelling [10,11]. "publication bias" (part 2, section 2.4) and "parameter optimization" (part 3) have been identified as the main issues of the https://doi.org/10.5599/admet.923 https://doi.org/10.5599/admet.923 http://www.pub.iapchem.org/ojs/index.php/admet/index mailto:suganok@fc.ritsumei.ac.jp http://creativecommons.org/licenses/by/4.0/ kiyohiko sugano admet & dmpk 9(2) (2021) 75-109 76 case studies [11]. recently, a more realistic view about pbpk modelling has been reported by an industrial consortium [12]. are we lost in modelling and simulation? recently, the editorial board of amdet and dmpk has announced the policy for the articles related to pbpk modelling (modelling and simulation ethics). this policy is introduced to enhance the science of pbpk modelling. in this article, the background of this policy is explained. this article consists of three parts: part 1: requirements for scientific writing of pbpk modelling, part 2: scientific literacy for pbpk modelling, part 3: middle-out approach (parameter back-calculation from clinical pk data). the topics discussed in this article have been repeatedly raised before. transparency and reproducibility (calculation traceability) are critically important for scientific writing (part 1) [12–14]. from the perspective of evidence-based medicine, case studies are less reliable for assessing the predictive power of a model (part 2) [15–17]. the issue of parameter identifiability in mathematical modelling has been repeatedly warned in the literature (part 3) [18–24]. in this article, the above points will be discussed focusing on oa pbpk modelling. but this article would also be beneficial to the other pbpk models. pbpk models will be a useful tool if used correctly. this article will hopefully enhance the science of oa pbpk modelling in the future. part 1: requirements for scientific writing of pbpk modelling the policy on the scientific writing of physiologically-based pharmacokinetic modelling (pbpk) articles complements the current journal's author guidelines that cover in vivo and in vitro methods based on scientific literacy. this guideline is also in line with regulatory guidance for industry regarding pbpk modelling [12–14]. in this article, past articles that do not comply with this policy are not quoted, because it would be disadvantageous for the authors. however, readers will see that the majority of the past papers have some issues raised in this article. transparency and reproducibility are mandatory to ensure the credibility of pbpk modelling. as sciences and technology progress, model equations and physiological data can be updated in the future (section 2.1). therefore, the use of the current best estimate parameter in pbpk modelling is appropriate. however, if the physiological data or model equation changes in the future, past articles need to be re-evaluated. for example, a wide variety of small intestinal fluid volume (vsi) has been used in oral absorption (oa) pbpk modelling, ranging from less than 100 ml to over 1500 ml (the former is based on the recent mri measurements) [25–29]. despite more than a 15-fold difference, they all claimed good predictability (section 3.3). this cannot be true. if all details had been reported, we can trace the calculation and retrospectively inspect the reason for this contradiction. but if there was an undisclosed part, it is not possible to judge whether the past good prediction is just a lucky coincidence or due to other reasons. when such an inconsistency arises in regulatory submissions, it causes a more troubling situation. to ensure the credibility of pbpk modelling, authors must write the manuscript as transparent as possible to enable inspection by peer-reviewers and ensure reproducibility by independent third parties. 1.1. introduction section the purpose of pbpk modelling should be explained in the introduction section. the question of interest and the context of use (cou) for pbpk modelling should be described [10]. there is no one-size-fits-all model (section 2.8). a mathematical model that suits the purpose should be selected. the reason for selecting a pbpk model should be explained. https://pub.iapchem.org/ojs/index.php/admet/pbpk admet & dmpk 9(2) (2021) 75-109 lost in modelling and simulation? doi: https:/doi.org/10.5599/admet.923 77 1.2. method section based on the scientific ethics of transparency and reproducibility, authors are requested to disclose the model equations, physiological parameters, and drug parameters (section 2.1) that are sufficient for peerreview (or appropriate references to them). undisclosed parts cannot be peer-reviewed in the first place. in addition, according the rule of science, an experimental section should be written as detailed as possible so that to “enable” reproducing the same result. this point is critical for ensuring the credibility of science. if the word "reproducibility" is interpreted literally, even if the model equations and parameters are not disclosed (as in a black box model), the results can be reproduced by using the same input data and the same software (same version). in the context of scientific credibility, for the articles of mechanistic models (including pbpk models (section 2.1)), it should be taken as calculation traceability. a method section should be detailed enough to enable someone who wants to trace the calculation process, at least in the essential parts for cou. if a part of the model is not disclosed, that part becomes a "black box”. because a black box can mask errors in any part of the model, the credibility of the entire research is damaged by the existence of only one undisclosed part. the scientific validity of a research article is the author’s responsibility, even when using commercial software. in papers using commercial software, the model equations and physiological parameters are often described as "default”. the default information may have been disclosed in a user’s manual. however, it is not available for peer-reviewers and readers. the default model equations and physiological parameters must also be publicly disclosed. in nearly all case study reports of oa pbpk modelling, case-by-case parameter optimization (backcalculation) has been committed to fit the simulated plasma concentration (cp) time curve to the clinical observation on a drug-by-drug basis (called “local middle-out approach”, see part 3). any parameter backcalculation from clinical pk data must be explicitly stated in the method section. case-by-case parameter back-calculation could have been unconsciously committed. for example, ad hoc selection of estimated permeability values from in silico, in vitro, or in situ data is a kind of parameter back-calculation. various terms implying a subtle adjustment such as “optimize”, “fit”, “adjust”, “recover”, “refine”, and “software estimated”, have been used to refer to parameter back-calculation. however, the difference from the initial value often exceeds two-fold due to the large errors in in vitro – in vivo extrapolation (ivive) and/or in silico prediction [30,31] (section 3.10). when a parameter is calculated from a chemical structure by an in silico model, the prediction accuracy of the in silico model should be shown (or referenced) (section 3.8 (iv)). 1.3. results section when case-by-case parameter back-calculation was committed, the initial input parameters and the simulation results before parameter back-calculation must be reported in the result section (figure 1). unfortunately, the failed prediction is often undisclosed in case study reports. however, this information is important for evaluating the creditability of a back-calculated parameter. for the advancement of science, failed predictions are just as (or even more) important than successful cases (section 3.4). after case-by-case parameter back-calculation, the fitted curve must be labeled as “fitted” (not “predicted”) for the same clinical pk data used for parameter fitting. the fitted cp time curve is not a predicted curve, because the observed cp time profile has been used to back-calculate the parameter (self-referencing) (section 3.11). in oral absorption (oa) pbpk modelling, the fraction of a dose absorbed (fa) is one of the most https://doi.org/10.5599/admet.923 kiyohiko sugano admet & dmpk 9(2) (2021) 75-109 78 important outputs for understanding the oral absorption of a drug. although clinical fa data (or its surrogates (section 2.6)) is not always available for model validation, simulated fa data is important for interpreting the simulation results regarding the oral absorption of a drug, at least for biopharmaceutics and formulation scientists (section 2.2). therefore, a simulated fa value should be reported. the simulated fa value (or the fa – time profile) is available as an output in all commercial oa pbpk software products. however, unfortunately, this value has not been reported in many reports. figure 1. schematic illustration of simulated and observed cp – time profiles 1.4. discussion section when parameter back-calculation from clinical pk data is committed, the following points should be discussed (see part 3 for details) [18–24,32–34]: (i) the reason for the mismatch of the initial “bottom-up” prediction and clinical pk data, (ii) the reason for selecting a parameter as the subject of back-calculation (sections 3.5 to 3.7), (iii) parameter identifiability (section 3.2), (iv) the accuracy of the other parameters than the subject of parameter back-calculation (section 3.8), (v) the plausibility of back-calculated parameter considering physicochemical properties, in vitro data, and in vivo physiology (section 3.10), (vi) the constancy of the back-calculated parameter in the clinical and/or population conditions (cou) to be predicted (sections 3.3 and 3.13), (vii) the validity of the optimized model (section 3.11). 1.5. references the above policy applies to the references. the authors should cite reliable articles as references. part 2: scientific literacy for physiologically-based pharmacokinetic modelling in part 1, we discussed how to write a scientific article for pbpk modeling. in this part, we will discuss how to interpret pbpk papers and how to evaluate pbpk models. physiologically-based pharmacokinetic (pbpk) models are not as easy to use as a smartphone app. as mentioned in the introduction, commercial pbpk software may not be so perfect as a user might believe [2–6,11,35,36]. before using a pbpk model, we must understand the scientific literacy for mathematical modelling. in this part, the scientific literacy required for pbpk modelling is discussed. we are fully aware that there are various opinions among the modelers on this topic. the purpose of this part is to suggest several viewpoints when using a pbpk model. 2.1. the basic concept of physiologically-based pharmacokinetic models basically, a pbpk model consists of independent a priori information of drug and formulation parameters, physiological and biological parameters at the organ level, and model equations [37]. all model admet & dmpk 9(2) (2021) 75-109 lost in modelling and simulation? doi: https:/doi.org/10.5599/admet.923 79 equations and parameters have a physical and physiological basis (mechanistic basis) (figure 2). pbpk models provide a mechanistic representation of pharmacokinetics and allow a priori “bottom-up” prediction of in vivo pk profiles from in vitro data for various clinical situations. to simulate the effect of a physiological factor, the model equation must include the factor as a system parameter. as drug parameters, the parameters that are intrinsic to a drug should be used. by combining drug-intrinsic parameters with physiological factors in model equations, the effects of physiological conditions can be simulated. this model structure allows the pbpk model to handle population variability and physiological covariates. furthermore, the oral absorption (oa) pbpk model can account for the bioequivalence of various formulations considering confidence intervals. this point is one of the specific features that can be handled by pbpk modelling. figure 2. overall prediction scheme of pbpk modelling (a) and the solubility model in biorelevant media for an acidic drug (b). (a) in vitro data is first reduced to drug intrinsic parameters using the mathematical model of an in vitro system [38]. the drug intrinsic parameters are then converted to in vivo pk profiles via intermediate parameters (e.g., the permeation rate constant (kperm)). the drug intrinsic parameters are directly related to a chemical structure so that suitable for in silico prediction and drug design. (b) drug intrinsic parameters: purple, physiological parameters: green, and intermediate parameter: blue [39,40]. ka is the dissociation constant, s0 is the intrinsic solubility of a drug, and kbm,0 and kbm,are the bile micelle partition coefficients for unionized and anionic drug molecular species, respectively. [cbm] is the bile micelle concentration, [h + ] is the proton concentration (= 10 -ph ), and [h2o] is the concentration of water. an oa pbpk model consists of the model equations of solubility, dissolution rate, permeability, etc. therefore, it is important to understand these equations before using pbpk modelling (section 1.2, 2.3, and 2.7). for example, in a physiologically-based solubility model (figure 2b), the ph and the bile micelle concentration (cbm) are used as physiological parameters, and the intrinsic solubility (s0), pka, and bile micelle partition coefficients (kbm) (for each unionized and ionized drug molecules) are used as drug parameters [35,39,40]. physiologically-based dissolution [41–48] and permeation [49–55] models have already been reported in the literature and implemented in some oa pbpk models. the followings are examples of physiologically-based dissolution and permeation models in the simplest form. https://doi.org/10.5599/admet.923 kiyohiko sugano admet & dmpk 9(2) (2021) 75-109 80 dissolution model (for mono-dispersed spherical small particles) [41–48,56] 1/3 2/3eff surface dissolv 2 undissolv dissolv gip d 3 1 d undissolvx d s xx dose t s vr         where xundissolv and xdissolv are the amounts of a drug undissolved and dissolved in the gastrointestinal fluid, respectively. deff is the effective diffusion coefficient, dose is the dose strength, ρ is the true density of the drug substance, and rp is the particle radius (rp < 30 μm). ssurface and sdissolv are the solubilities of a drug at the particle surface and in the bulk fluid, respectively. vgi is the gastrointestinal fluid volume. the dissolution rate constant (kdiss) is kdiss = 3 deff ssurface/(ρrp 2 ). ssurface, sdissolv, and deff are affected by physiological factors such as ph, buffer capacity, bile micelle concentration (figure 1(b)). permeation model (for passive diffusion) [49–55] perm perm dissolv eff dissolv gi d 2 d x df k x p x t r     eff effu 0 trans0 para wc eff 1 1 pe p dve f f p p p h      where xperm is the amounts of a drug permeated the intestinal wall, kperm is the permeation rate constant, df is the degree of flatness of the small intestinal tube, rgi is the radius, peff is the effective intestinal membrane permeability, pe is the plica expansion factor, puwl is the uwl permeability, ve is the villi expansion factor, fu is the fraction of unbound drug molecule species (free fraction), f0 is the fraction of undissociated drug molecule species (calculated from pka and ph), ptrans0 is the intrinsic passive transcellular permeability, ppara is the paracellular permeability, heff is the effective thickness of the uwl, and pwc is the water conveyance. instead of using a physiologically-based mechanistic dissolution model, an experimental in vitro dissolution profile can be plugged into an oa pbpk model as a kind of intermediate parameter (figure 2a)[57]. this strategy is often used when the dissolution process is the key determinant of oral drug absorption. in this case, the dissolution profile is pre-fixed so that it is not altered by the physiological factors in the computer simulation. for example, when a dissolution profile measured at ph 1.2 is plugged into an oa pbpk model, it is not changed even if the gastric ph value is changed from ph 1.2 to ph 5.0 in the oa pbpk model. consequently, it cannot account for the effect of interor intrasubject variability of relevant physiological parameters on drug dissolution processes. such a simulation strategy may be understood as a variation of the convolution method, rather than “physiologically-based” computational modelling (in the case of using in vivo predictive dissolution testing [58–61], the in vitro dissolution profile itself is physiologically-based). 2.2. to explain or to predict? which is the position of a pbpk model? the purpose of mathematical modeling is not only to calculate the predicted values, but also to explain the observed and predicted data (understanding, interpretation) [62]. the interpretability of a model is especially important for medical applications. for an explanation-oriented model, ideally, all system parameters should be set based on independent a priori information. from the viewpoint of mathematical modelling, pbpk models are more of an explanation-oriented model when compared to more complex statistic models like deep learning [63]. in general, the interpretability of a model decreases as the model admet & dmpk 9(2) (2021) 75-109 lost in modelling and simulation? doi: https:/doi.org/10.5599/admet.923 81 becomes more complex. simple models have a great value of “interpretability” by themselves [64]. on the other hand, complex models may (or may not) show better predictive power (prediction performance, generalization performance) (section 2.8). the predictive power and description capability of a model are different. the description capability of a model is to describe existing data, whereas the predictive power is to predict unknown data. as a model becomes more complex and flexible, description capability always increases. however, predictive power does not always increase, because assumptions and errors can accumulate with the increasing complexity of a model (sections 2.8 and 3.10). when parameterization is used, a complex model is often prone to overfitting (figure 3). in mathematical modelling, generally speaking, to compensate for its lesser interpretability and its higher risk of overfitting, a complex model should show a significant advantage over a simple model with respect to predictive power for the context of use [20,64–67]. if simple and complex models show comparable predictive power with regard to the context of use, the simple model should be selected (cf. occam’s razor, the principle of parsimony) [68]. at the same time, when a complex model shows better predictive power, the complex model should be selected (section 2.8). figure 3. overfitting. in figure 3, the quintic equation perfectly describes (fits) the data. if the experimental data is error-free, it is suitable for predicting unknown y from new x data. in practice, especially for biological data, there is always error in experimental data. for objective model selection, statistical indices such as the akaike information criterion (aic) can be used [69]. 2.3. how to interpret a model there are several ways to interpret a mathematical model to understand the behavior of the system of interest. parameters the first step to understanding a pbpk model is to know the parameters that affect the physical and physiological processes. parameters that are not included in the model equation cannot be captured by the model. before using a pbpk model, we must understand the key factors that affect the pharmacokinetic processes of a drug (sections 2.1 and 2.7). intermediate parameters can be used to understand the contribution of each process. https://doi.org/10.5599/admet.923 kiyohiko sugano admet & dmpk 9(2) (2021) 75-109 82 mathematical structure a mathematical structure represents the relationship between the system parameters, the dependent variables, and the independent variables, for example, being additive, synergistic, proportional, inversely proportional, exponential, etc. this point is important for parameter identifiability (see part 3). rate-limiting step in any kinetic model, it is essential to understand the rate-limiting step (or rate-limiting parameter) [70,71]. for example, the rate-limiting step that determines fa (farls) can be diagnosed from the solubility, dissolution, and permeation numbers [53,72]. from farls, the sensitive parameters for fa of a drug can be easily identified. the rate-limiting step of intestinal membrane permeation can be diagnosed by using a physiologically-based permeability model [73,74]. the rate limiting step could differ depending on the physiology of each subject. parameter sensitivity analysis parameter sensitivity analysis (psa) has been utilized to interpret black-box models, such as deep learning. psa could also be useful in pbpk modelling. in pbpk modelling, psa has been generally used to help deciding which parameters require further consideration either for additional in vitro measurement or parameter back-calculation (section 3.5). a pbpk model includes all parameters and models, however some of these may have little or no influence on the outcome. 2.4. evidence level regarding the predictive power of pbpk models the predictive power of a mathematical model can be evaluated based on the concept of evidencebased medicine (ebm) [15–17]. according to ebm, the evidence level of a case study is low. case studies are prone to publication bias. successful cases tended to be published whereas failed cases are usually not published. historically, publication bias has caused the illusion of knowledge, especially in the medical area. a systematic study is essential to evaluate the predictive power of pbpk models. for the bottom-up prediction by oa pbpk models, several systematic evaluation studies have already been published [2–5]. however, middle-out approaches (see part 3) have been evaluated mostly by case studies (or the collection of case studies). systematic evaluation is required for middle-out approaches using a standardized procedure (more precisely, for local middle-out dynamic pbpk models) [10,75,76]. although the evidence level of a case study is limited, it is still beneficial as far as appropriately conducted and reported. however, the creditability of case studies has often been compromised by the existence of a black box in a model (see part 1) and inappropriate parameter optimization (see part 3) [11]. the middle-out approach is discussed in detail in part 3. 2.5. points to consider for systematic evaluation of predictive power dataset: in oa pbpk modelling, the biopharmaceutical classification system (bcs) can be a good starting point [6,77,78]. a test set biased to bcs class i drug (high solubility/high permeability) should be avoided, because complete absorption is easily expected for bcs class i drug. similarly, no food effect and no gastric ph effect on fa are expected for bcs class i drugs (but they could affect cmax and f). in the case of bcs class i drugs, complex oa pbpk models may add little advantage over simple models for fa prediction (depending on the purpose of the model) (section 2.8). control: a simple model or a naïve prediction (figure 4) should be used as a control to evaluate the predictive power of complex models [5,50,51,79]. admet & dmpk 9(2) (2021) 75-109 lost in modelling and simulation? doi: https:/doi.org/10.5599/admet.923 83 outcome: the absolute bioavailability (f) or fa has been used to evaluate oa pbpk models [2]. f is calculated by subtracting the effects of dose and systemic clearance (cl) from oral auc data using i.v. data (f = (aucp.o./dosep.o.)/(auci.v./dosei.v.)). fafg can then be calculated from the hepatic clearance (clh) as fafg = f/fh, fh = 1 – clh /hepatic flow [80] (fg: the fraction escaping the intestinal wall metabolism, fh: the fraction escaping the hepatic first-pass metabolism). in addition, there are several methods to estimate fa from clinical pk data (section 2.6) [5,81]. auc is less suitable for evaluating the predictive power of oa pbpk models because the main determinant of auc is dose and cl in many cases. statistics: statistics must be carefully interpreted. the percentage within a two-fold error is often used to evaluate the predictability of pbpk models. however, even when the predicted and observed f values are randomly distributed from 0 to 1, the percentage within a two-fold error becomes 50 % (figure 4a) (see also figures 3 and 4 in [6]). when predicted as average, 75 % is within a 2-fold error (figure 4b). this kind of prediction is called “naïve prediction”. this percentage increases when the data set is biased towards bcs class i. figure 4. naïve prediction. (a) random prediction, (b) prediction by average. the percentage within a twofold error is 50 % for (a) and 75 % for (b). 2.6. surrogates of fa data in oa pbpk modelling, the fraction of a dose absorbed (fa) is an important simulation output (section 1.3). however, there is no exact method to measure in vivo fa. therefore, one or a few approximations have been used to estimate fa from in vivo pk data [50,70,79,82–84]. (a) mass-balance data using a radio-labeled drug for i.v. and p.o. (b) from absolute bioavailability (f), hepatic clearance (clh), and the hypothesis of fg = 1 (section 3.8 (v)). (c) relative bioavailability of solution vs. solid formulation. (d) relative bioavailability in the fasted state vs. the fed state (when do < 1 in the fed state) (do: the dose number (dose/ (s ×vsi)), s: solubility, vsi: small intestinal fluid volume). (e) relative bioavailability between the dose strengths where do < 1 and do > 1 (auc can be corrected by elimination t1/2 for nonlinear clearance drugs). the mass-balance data would be the most reliable data to estimate fa. however, these data are rarely available. when i.v. data is available, we can estimate fafg from f and clh. when i.v. data are not available, https://doi.org/10.5599/admet.923 kiyohiko sugano admet & dmpk 9(2) (2021) 75-109 84 for high permeability drugs, the relative bioavailability to an oral formulation that eliminates the effect of solubility and dissolution ((c)-(e)) can be used as a surrogate of fa (cf. in this case, the effects of fg and fh are canceled out). the permeability category (low/high) can be reliably diagnosed by in vitro permeability assays [85]. in the case of high papp drugs (> metoprolol papp at ph 6.5, metoprolol log dph6.5 = -1.1)[86], fa will be greater than 0.8 with a very high probability when there is no solubility and dissolution limitation. based on this high reliability, caco-2 papp data has been used for regulatory biowaiver submission (note: for highly lipophilic drugs (log dph6.5 > 1.5), the caco-2 assay may underestimate the permeability of the drug due to experimental artifacts such as membrane binding [87,88], leading to a misassignment of a drug as low permeability (this can be identified by a mass balance study in caco-2). in this lipophilicity range, the diffusion through the unstirred water layer becomes the rate-limiting step of in vivo membrane permeation. therefore, a rough estimation of permeability from experimental log dph6.5 will be sufficient and could be more reliable [55,86]). the above methods showed similar fa values when applied to the same drug in most cases [50,79,82,83], confirming the validity of these methods. more than 600 clinical fa data has been compiled from the literature and has been used to evaluate oa pbpk models [5,50,73,79,83,86,89,90]. for low fa drugs the inter-subject variability is usually very high, and a clinical study may be an unrepresentative sample of the population. 2.7. often forgotten physicochemical mechanisms in oa pbpk modelling a good understanding of physicochemical processes in oral drug absorption is specifically important for oa pbpk modelling (see conclusion part). below are some examples of physicochemical mechanisms that are often forgotten in oa pbpk modelling. example 1: peff estimation from in vitro data it is well known that the effective intestinal membrane permeability (peff) is markedly affected by bile micelle binding [51,91–94], the unstirred water layer (uwl) [74,95–98], and the anatomical features (villi and fold structures) [99] (section 2.1) (cf. the peff value includes the free fraction effects by definition in most cases of commercial software). however, these factors cannot be captured by an empirical equation of peff = apapp b (log peff = a’ + b’log papp) (papp: in vitro apparent permeability). this empirical equation cannot correctly explain and predict the food effect (especially the negative food effect) [51,91,100], the formulation effect (solubility-permeability trade-off)[101,102], and species differences [52,103]. the empirical coefficients, a and b (or a’ and b’), are usually determined by the regression analysis between peff and papp using hydrophilic model drugs (log d < 1)[31]. therefore, lipophilic drugs are beyond the applicable range of the standard curve. in addition, papp is usually measured in the absence of bile micelles on the apical side and albumin on the basolateral side. the thick uwl in the in vitro systems mask the membrane permeability when not rigorously agitated [98,104]. finally, and most importantly, this equation is an empirical correlation so that it should not be referred to as “physiologically-based”. a mechanistic physiologically-based permeation model that considers these factors is available in the literature (section 2.1) [49–55] and in some commercial software products. example 2: dissolution in the stomach the gastric fluid is acidified by hydrochloric acid (hcl). because hcl is not an efficient buffer, the ph value of the gastric fluid increases when a free weak base drug is dissolved [105]. in addition, the solid surface ph is markedly increased by the dissolving free base molecules [42,43,48,61,106,107]. therefore, the solid surface solubility (ssurface) and the bulk phase solubility (sdissolv) must be differentiated in the admet & dmpk 9(2) (2021) 75-109 lost in modelling and simulation? doi: https:/doi.org/10.5599/admet.923 85 nernst-noyes-whitney equation (section 2.1) [35]. these two factors are important for predicting the effect of gastric ph on drug absorption. a mechanistic dissolution model that considers these factors is also available in the literature (section 2.1) [42,43,48,61,79,106,107] and in some commercial software products. example 3: salt dissolution the dissolution modelling of a salt form drug is not that easy as one might imagine. the solid surface solubility of a salt is significantly higher than that of the free form (> 100-fold in most cases). however, the equilibrium solubility in the ph-controlled region becomes the same regardless of the starting material is a free form or a salt form (unless the residual solid (equilibrium maker) show different solid forms) (cf. a salt form coverts to a free form in the ph-controlled region). in drug discovery and development, the solubility of a drug substance is usually measured in well-buffered media after a sufficient incubation time to achieve equilibrium [108–111]. the small intestinal ph (about ph 6.5) is in the ph-controlled region in most cases. in addition, a salt form may or may not show faster dissolution and more importantly supersaturation after dissolution, because the solid surface precipitation of a less soluble free form can inhibit the dissolution of its salt [112,113]. the mechanism of supersaturation and precipitation is not well understood at this moment. however, at least, it does not simply follow the first-order kinetics [114–119]. in addition, in silico models for the physicochemical properties of a drug are not so accurate as to be used for pbpk modelling [120]. solubility measurements are not as easy as a modeler might imagine [109,111,121]. the pka values change between 25 °c and 37 °c [122,123]. enabling formulations such as amorphous solid dispersion [124], co-crystal [125,126], nanoparticles [127–129], and self-emulsifying drug delivery system [130,131] also requires an in-depth understanding of physical chemistry for oa pbpk modelling. further investigations on these points are required in the future (see conclusion part). 2.8. a simple pbpk model or a complex pbpk model, which one to use? “everything should be made as simple as possible, but not simpler.” albert einstein as mentioned above, we are fully aware that there are various opinions among the modelers on the selection of mathematical models. a pbpk model should be selected considering the purpose of modelling and available data at each drug discovery and development stage. in the early drug discovery stages, a simple pbpk model may be sufficient. in the late drug development stages and after launch, a complex pbpk model may be required to investigate more complex clinical situations. for the use in the early drug discovery stages, simple models can show sufficient prediction performance [20,64,82,132,133]. a simple oa pbpk model has shown good prediction performance for the fraction of a dose absorbed (fa) [5,50,79,83,134] and plasma concentration (cp) time profiles [135–143]. for relative bioavailability (frel) prediction, a simple oa-pbpk model is also available for the food effect (via bile micelles) and gastric acid effect predictions [51,79,100]. by multiplying auc in the fasted state (or low gastric ph) with frel (i.e., auc ratio), auc in the fed state (or high gastric ph) can be estimated. minimal pbpk models have also been proposed to reduce the complexity of a model (see section 3.13 for metabolic ddi prediction) [21,144]. to explain and predict complex pharmacokinetics and population variation in late drug development and product life-cycle management, a pbpk model should have sufficient components. commercial software products that implements a complex dynamic oa pbpk model would be especially useful for these purposes. https://doi.org/10.5599/admet.923 kiyohiko sugano admet & dmpk 9(2) (2021) 75-109 86 part 3: middle-out approach part 2 discussed the interpretation and evaluation of the pbpk model. one of the most difficult parts of evaluating a pbpk model is the validity of the middle-out approach. therefore, middle-out approaches are discussed in this part. middle-out approaches have been widely utilized in real drug discovery and development [21]. they would be effective to improve various processes in drug development. at the same time, if they are inappropriately used, they could cause trouble. the purpose of this section is to discuss how to properly use the middle-out approach in oa pbpk modelling. 3.1. the concept of the middle-out approach in a middle-out approach, some parameters of a pbpk model are derived from in vitro data (bottom-up), while others are derived from clinical pk data (top-down) on a drug-by-drug basis (see also section 3.14). in the following sections, this prediction scheme is referred to as a “local (drug-by-drug)” middle-out approach to differentiate it from a “global” middle-out approach (section 3.15) [145]. for drug-drug interaction (ddi) prediction, the local middle-out approach using a cyp specific inhibitor or substrate has been successfully used [21] (section 3.12). however, it has been pointed out that the inappropriate use of a local middle-out approach reduces the creditability of oa pbpk modelling [17,20]. a middle-out approach brings an empirical model into pbpk modelling (figure 5, the red line and square). therefore, we should follow the good practice of empirical modelling: (i) before back-calculation, parameter identifiability must be carefully examined (sections 1.4 and 3.2) [18–23,32,33]. (ii) the degree of freedom must be enough to avoid overfitting (section 2.2, figure 3). (iii) the predictive power must be validated using clinical data that is not used to back-calculate the parameter (called “cross-validation”) (section 3.11). (iv) the optimized empirical model should be used within the parameter space limited by the data used for back-calculation (so that as interpolation) (this is a general recommendation for empirical models. for the hybrid of empirical and mechanistic models, this point (iv) needs further in-depth discussion.). 3.2. parameter identifiability: a simple explanation the issue of parameter identifiability in biological mathematical modelling (including pbpk modelling) has been repeatedly warned in the literature [18–24]. however, this issue seems to have been overlooked in many case studies using a local middle-out approach. in this section, parameter identifiability is plainly explained focusing on oa pbpk modelling. in a middle-out approach, the identifiability of a parameter(s) must be assessed before back calculating the parameter, whether the parameter(s) can be uniquely and reliably identifiable from the input-output data [18–24]. there are two types of parameter identifiability: statistical and structural. statistical identifiability is related to the experimental error of the observed data. however, even with error-free data, a model parameter could be structurally non-identifiable. to illustrate the concept of structural identifiability, let’s consider an equation of y = abx, where x is an explanatory variable, y is an object variable, and a and b are system parameters [23]. the lump quantity a·b is uniquely identifiable from x and y, while the individual parameters of a and b are non-identifiable (it is mathematically indefinite). even when multiple xy data sets are available, a and b are non-identifiable. therefore, a or b must be fixed separately. admet & dmpk 9(2) (2021) 75-109 lost in modelling and simulation? doi: https:/doi.org/10.5599/admet.923 87 figure 5. bottom-up and middle-out approaches for oa pbpk modelling. a middle-out approach brings an empirical model into a pbpk model (the red line and square). for parameter back calculation and validation, separate clinical data (fa, fa’) must be used. the statistical identifiability can be understood by considering y = 1/ (x + a). if the error of x is comparable to a, a cannot be reliably determined (even assuming zero error in y). for example, when x = 100 ± 10, we cannot determine a smaller than 10. preferably, a should be determined at x << a. another familiar example is y = 1 exp(-ax) (figure 6). in this case, ax should be < 1.5 if y has a 20 % error. the other simple examples of parameter identifiability are shown in supplemental information. figure 6. y = 1 – exp (-ax). 3.3. local middle-out approach to predict the food effect: an example to overview the process of a local middle-out approach and understand the importance of parameter identifiability, in this section, we look at a simple example of oa pbpk modelling using a local middle-out approach and then discuss the checkpoints. https://doi.org/10.5599/admet.923 kiyohiko sugano admet & dmpk 9(2) (2021) 75-109 88 step 1: problem statement the drug candidate is a poorly soluble non-ionizable compound. the formulation is a simple immediaterelease formulation. drug parameters are available from preclinical in vitro studies (table 1). the physiological data are provided as default values. table 1. drug and physiological parameters (example) parameter value comments drug parameter permeation rate constant (kperm) 0.1 h -1 from caco-2 data (peff = a papp b ) effective solubility (s) 0.1 mg/ml 0.2 mg/ml fasted state simulated intestinal fluid (fassif) fed state simulated intestinal fluid (fessif) dose strength (dose) 350 mg particle size 5 μm diffusion coefficient 6 × 10 -6 cm 2 /s physiological parameter small intestinal fluid volume (vsi) 1000 ml default (see checkpoint 4) small intestinal transit time (tsi) 3.5 h default small intestinal radius (rsi) 1.5 cm default degree of flatness (df) 1.7 default fraction escaping first pass hepatic metabolism (fh) 1 calculated from i.v. clh and hepatic flow now, the clinical pk data in the fasted state became available after the first-in-human study (both p.o. and i.v.). the bioavailability (f) was 0.30 ± 0.15 (mean ± s.d.). we are asked by the managers to predict auc in the fed state in healthy volunteers. to utilize the first-in-human clinical data, we decided to conduct a local middle-out approach. step 2: model selection the particle size is small enough so that the drug dissolution process does not become the rate-limiting step [70–72]. in addition, because the drug is non-ionizable, gastric dissolution would have little effect on f. in the case of solubility limited oral absorption, f can be calculated as [146]: g h perm si si a g h g h f f k sv tpn f f f f f f do dose    (1) where pn is the permeation number (=kperm ×tsi), and do is the dose number do = dose/ (s ×vsi). fg is the fraction escaping the intestinal metabolism. in this example, eq. 1 is used as a model equation for convenience. step 3: parameter optimization from eq. 1 and given data, kperm is optimized to be 0.3 h -1 in the fasted state. step 4: model validation using the optimized kperm value, f is calculated as, this predicted f value perfectly matches the clinical f data (= 0.30). this kperm value is additionally 𝐹 = 0.3 × 0.1 × 1000 × 3.5 350 = 0.30 admet & dmpk 9(2) (2021) 75-109 lost in modelling and simulation? doi: https:/doi.org/10.5599/admet.923 89 validated using the independent clinical f data in the fasted state. step 5: prediction the optimized kperm value is then used to predict f in the fed state. using the solubility in fessif (0.2 mg/ml), fa is calculated as therefore, auc is predicted to increase twofold in the fed state. checkpoint 1: hidden “bottom-up” prediction usually, we start a prediction study with a “bottom-up” approach at a preclinical stage (section 1.3). the input parameters are initially projected from the in vitro data. in the above case, the “bottom-up” prediction resulted in a 3-fold underestimation. kperm was increased threefold (from 0.1 to 0.3) after parameter optimization. this is not a subtle adjustment. this information alerts that there is a marked discrepancy between the kperm values estimated from the caco-2 data and back-calculated from the clinical f data (section 3.4). checkpoint 2: subject of parameter optimization the reason for selecting kperm as the subject of parameter optimization is not clear (sections 1.2 to 1.4 and 3.5 to 3.7). checkpoint 3: parameter identifiability eq. 1 suggests that the kperm value is not identifiable from the provided data (f, fh, dose, s, vsi, and tsi). fg must be separately fixed (section 3.8 (v)). in the above calculation, fg was unknowingly assumed to be 1 (section 1.2 to 1.4). checkpoint 4: hidden errors the intestinal fluid volume of 1000 ml is significantly greater than the current best estimate (most probably < 150 ml) (in the discussion below, we assume that 100 ml is realistic for the convenience of discussion) [26–29]. the optimized kperm value carries the error in the intestinal fluid volume (e.g., 1000 ml vs 100 ml). with a more realistic intestinal fluid volume of 100 ml, the kperm value is back-calculated to be 3.0 h -1 , ten-fold higher than 0.3 h -1 . this could cause a misunderstanding of the rate-limiting step for membrane permeation (section 1.4). checkpoint 5: variation of clinical data considering that cv% of f is 50 %, it is not possible to back-calculate the kperm value within a good confidence interval. checkpoint 6: constancy of optimized parameter it was implicitly assumed that the kperm values in the fasted and fed states are the same. however, it is well known that bile micelle binding reduces the free fraction, the effective diffusion coefficient, and consequently peff and kperm (section 2.7) [51,91,92,94,147,148]. 𝐹 = 0.3 × 0.2 × 1000 × 3.5 350 = 0.60 𝐹 = 0.1 × 0.1 × 1000 × 3.5 350 = 0.10 https://doi.org/10.5599/admet.923 kiyohiko sugano admet & dmpk 9(2) (2021) 75-109 90 checkpoint 7: model validation the predictive power of a model cannot be validated by the same data used for parameter backcalculation (section 3.11). checkpoint 8: prediction after parameter optimization, the predictive range of the optimized model is limited to the parameter space of the clinical data that is used for model development (limited to interpolation with some exceptions). in the above example, a simple model (eq. 1) was used for convenience to illustrate the middle-out approach. in complicated pbpk modelling, these checkpoints are not so easy to recognize. however, the mathematical principle remains the same. 3.4. when is a local middle-out approach required? a local middle-out approach is required when a bottom-up prediction is not satisfactory for the contest of use (section 1.3, figure 1). the widespread use of a local middle-out approach is consistent with the results of systematic evaluation suggesting that current bottom-up oa pbpk models need significant improvement (introduction section). discrepancies between bottom-up predictions and clinical observations may suggest opportunities to uncover unidentified mechanisms or to improve the equations and parameters of the model (sections 1.3, 3.3 checkpoint 4). before parameter back-calculation, the reason for the discrepancy should be thoroughly explored (sections 1.4 and 2.7). 3.5. how to diagnose parameter identifiability in oa pbpk modelling in the literature, peff has often been the subject of back-calculation. therefore, the identifiability of peff is discussed as an example below. simple analytical solutions are useful for diagnosing the parameter identifiability (section 2.2). the fa equation represents the relationship between the solubility, dissolution rate, and permeation of a drug to fa [5,35,149]: a 1 1 exp 1 f do dn pn             (2) where dn is the product of the dissolution rate (kdiss) and the intestinal transit time (dn = kdiss × tsi). when do < 1, set do = 1. when the oral absorption of a drug is limited by the dissolution rate (dn < pn/do), the permeation process is statistically non-identifiable from fa (cf. y = 1/ (x + a)). in other words, for peff to be identifiable, the oral absorption must be permeability or solubility-permeability limited. in this case, eq. 2 becomes: a 1 exp pn f do         (3) at pn/do > 0.7, the peff value (that is in pn) is statistically non-identifiable from clinical fa data considering the variation of clinical fa data (cf. y = 1 exp(-ax), figure 6). at pn/do < 0.7, eq. 3 can be approximated to [146]: admet & dmpk 9(2) (2021) 75-109 lost in modelling and simulation? doi: https:/doi.org/10.5599/admet.923 91 a pn f do  (4) using eq. 4 for fa, the auc value after oral administration can be expressed as: perm si g h perm si si a g h g h g h si dose k t f f k sv tdose dose dose pn dose auc f f f f f f f f cl cl cl do cl cl sv      (5) finally, g h eff si si si 2f f df p sv t auc auc clr     (6) in the absence of a unique correspondence between a parameter and a cp time profile, it is impossible to directly quantify the physiological process that involves the parameter. when the parameters are in a multiplication relationship, they are structurally non-identifiable from clinically observed data (cf. y = abx). as clearly represented in eq. 6, the individual parameters related to oral drug absorption cannot be calculated solely using auc after oral administration [19–22,33]. therefore, an oa pbpk model is essentially overparameterized. to back-calculate peff from fa, the other eight parameters (fg, fh, s, vsi, tsi, rsi, df, and cl) must be fixed separately. parameter sensitivity analysis can be used as a support to diagnose the identifiability of parameters. however, auc is sensitive to all the parameters on the right-hand side of eq. 6. being a sensitive parameter is a necessary but not sufficient condition to be identifiable [19–22]. a similar analysis of parameter identifiability can be performed for cmax and tmax: el el aa max d el k k kkfdose c v k        a max a el el 1 ln k t k k k         where ka is the absorption rate constant, kel is the elimination rate constant, and vd is the volume of distribution. ka can be approximated as 1/ka = 1/kdiss + do/kperm (section 3.14). 3.6. what kind of parameters can be candidates for back-calculation? only an intermediate parameter (e.g., peff) or an empirical scaling factor should be the candidates for the subject of parameter back-calculation from clinical pk data in a local middle-out approach (figure 2). drug-intrinsic parameters, especially physicochemical properties, should not be the subject of parameter back-calculation from clinical pk data. they should be determined by in vitro measurements. physiological parameters such as the small intestinal fluid volume (vsi) should not be the subject of drugby-drug back-calculation as well, because they do not change drug-by-drug (except for drugs with gastrointestinal effects) (a global middle-out approach has been used to estimate some physiological parameters when direct measurements are not possible (section 3-15)). post-absorptive (systemic) pk parameters such as cl and vd should be determined from i.v. cp time data [33] (unless f = 1 can be surely estimated from in vitro data, such as the case of bcs class i drug with low hepatic clearance). the same cl and vd values should be used regardless of formulations or https://doi.org/10.5599/admet.923 kiyohiko sugano admet & dmpk 9(2) (2021) 75-109 92 administration routes. oral formulation usually does not affect systemic cl and vd (however, it affects cl/f and vd /f). 3.7. which candidate parameter should be selected for back-calculation? when multiple candidate parameters can equally explain the discrepancy between prediction and observation, we cannot determine which one to be selected for back-calculation solely based on the clinical data. this situation is like a checksum process. we can validate a series of numbers by checking the sum of numbers, but we cannot tell which number has the error. the selection of a parameter depends on the clinical study design and the reliability of the estimation from in vitro data (section 3-8, (iv)). identifiability issues can be even more severe when there are regional differences in the gastrointestinal parameters in eq 6. 3.8. how to fix the other parameters there are several methods to fix the other parameter values. (i) clinical study using different administration routes and formulations to reliably estimate a parameter from clinical pk data, the interference of confounding factors must be reduced as much as possible. a specific clinical study design has been employed to eliminate the confounding factors for each parameter (table 2) (see also (3.13)). for example, systemic clearance (cl) can be obtained from i.v. data. an i.v. administration eliminates the oral absorption process (f = 1), so that cl become identifiable from auc (cf. cl = dose/auc). vd and fh can also be calculated from the i.v. data [150]. an oral solution formulation can be used to eliminate dn from eq. 2 (1/dn becomes negligible), and s and vsi from eq. 6 (mathematically, by fixing do = 1). table 2. the parameters and the clinical study design parameter clinical study design reference cl, vd i.v. administration a peff site-specific solution administration and i.v. administration [151] precipitation rate site-specific solution administration and sampling [152][153] a see any pharmacokinetic textbook it is not known whether parameter back-calculation from the oral cp time data after the administration of solid dosage forms can be accurate. theoretically, the dissolution processes can be decomposed from the permeation process by using the cp time data from solution formulations [154]. however, this deconvolution process often becomes unstable due to the variation of the data. in addition, there has been no systematic evaluation of peff back-calculation from the oral cp time data after the administration of solid dosage forms. unfortunately, in the literature, peff has been back-calculated even in the absence of i.v. data in many cases. the credibility of such peff back-calculation is at least questionable (section 3.9). (ii) experimentally measured physiological parameters experimentally measured physiological parameters can be obtained from the literature (vsi, tsi, and rsi in eq. 6). however, it should be noted that some of the physiological parameters reported in the literature have large variations between subjects and between occasions. (iii) physiological parameters estimated by a global middle-out approach the global middle-out approach has been used to back-estimate an unknown physiological parameter from clinical pk data (df in eq. 6, section 3.15). admet & dmpk 9(2) (2021) 75-109 lost in modelling and simulation? doi: https:/doi.org/10.5599/admet.923 93 (iv) preclinical in silico, in vitro, and in vivo data preclinical in silico, in vitro, and in vivo data can be used to fix a parameter if their predictability is sufficient considering the contest of use (cou) (section 1.2). the in vitro equilibrium solubility in biorelevant media such as fassif and fessif has often been assumed to be in vivo predictive. however, it should be noted that these are simplified artificial fluids, not actual intestinal fluids. fassif and fessif are very good model fluids, yet, the solubility values do not always accurately reflect the real in vivo values [155]. in addition, the solubility data could be inaccurate when inappropriately measured [109]. current in vitro-in vivo extrapolation (ivive) and allometric scaling for cl and vd is not sufficiently accurate for the cous of oa pbpk modelling in the late drug development stage [7]. in silico models for the physicochemical properties and other adme properties of a drug are not so accurate as to be used for pbpk modelling [120] (except for molecular diffusion coefficients [156, 157]). (v) hypothesis generation the peff value is still not identifiable after fixing s, vsi, tsi, rsi, df, cl, and fh. hypothesis generation is required to estimate fg [14]. hypothesis generation may include ivive, but in a less qualitative manner. hypothesis generation about the negligibility of a parameter would be most credible and useful because it can reduce the interference from the parameter for back-calculation. the credibility of a hypothesis can be improved by combining various in vitro, preclinical in vivo, and clinical observations. for example, when in vitro data suggest that a drug is not a substrate of intestinal metabolic enzymes, fg = 1 (no gut wall metabolism) can be a plausible hypothesis. low intrinsic hepatic clearance (< 100 ml/min/kg)[158], the lack of metabolites, and the lack of clinical grapefruit–drug interaction can further support the hypothesis of fg = 1. 3.9. a good example of credible parameter back-calculation sjögren et al. reported that peff can be credibly identifiable by the deconvolution of the cp time profiles after intraintestinal bolus administration as a solution, using i.v. disposition data [151]. they pointed out that the cp time profiles after an oral administration may not be suitable for peff estimation due to possible interference of confounding factors. they used a solution formulation to eliminate the uncertainty in s, vsi, and the effect of dissolution processes. intra-intestinal administration was used to eliminate the effect of gastric emptying. the i.v data was used to calculate cl and fh. hypothesis generation was used for fg. the basic concept of their approach is shown in figure 8 (they used the deconvolution method, but it was simplified to auc calculation to explain the concept). 3.10. accumulation of errors after parameter back-calculation, the simulation curve would show perfect fitting to the observed cp time data that had been used for parameter back-calculation. however, this perfect fitting does not imply the validity of all parameters and model equations. a back-calculated parameter inherits the errors of the other parameters (see checkpoint 4). in the case of eq. 6, the errors in nine separately fixed parameters (vsi, tsi, rsi, df, s, cl, fg, fh, and clinical auc) exponentially accumulate in peff (sections 2.2, 2.8). even if each parameter has only a small error of 20 % (1.2-fold), the total error can become 5.2-fold (= 1.2 9 ). to back-calculate peff within less than 20 % error, each parameter must have less than 2 % error. any experimentalist knows that this is not possible. furthermore, back-calculating one parameter hides the errors of the other parameters and model equations (section 3.4). peff itself can show high inter-subject variability [159]. https://doi.org/10.5599/admet.923 kiyohiko sugano admet & dmpk 9(2) (2021) 75-109 94 figure 8. peff calculation scheme using a set of clinical data, literature data, and hypothesis generation. each parameter is fixed step-by-step. the parameters after the evaluation are moved to the right side of the equation and shown in bold type. eq.6 is used to understand the concept of back-calculation. sjo ̈gren et al. used the deconvolution method [151], but the basic concept of parameter identification is the same (the deconvolution method can eliminate the uncertainty in tsi). 3.11. how to validate the predictive power of a pbpk model after parameter back-calculation in a local middle-out approach, the predictive power needs to be validated for each drug after parameter back-calculation. the parameter back-calculated in one clinical condition is not necessarily valid in the other clinical condition (section 3.3, checkpoint 6). the predictive power of a model cannot be validated by the same data used for parameter back-calculation, because it is self-referencing. therefore, the predictive power needs to be validated using other clinical pk studies of the context of use (cou) (“cross-validation”). the pk data under clinical conditions where some of the same pathways in the system are perturbed as in the cou can be employed for validation [10,21]. in the case of food effect prediction, the optimized model should be validated by the cp time data under a clinical condition in which the pathways of oral absorption are perturbed by the same factor for the food effect. in other words, the food effect prediction must be validated by a clinical food effect study. the clinical pk data in the same clinical condition cannot be used for validation, even if it is independently determined, because the cp time profiles are expected to be similar under the same clinical condition (this is a kind of “leakage” in the cross-validation process). multiple-dose pk data under the same clinical condition cannot be used for validation (unless the context of use is the prediction of metabolic enzyme induction or mechanism-based inhibition in the intestine after multiple-dose pk). 3.12. “confirm and refine“ strategy in drug discovery and development the quality and quantity of experimental data available for pbpk modelling increase as a research project proceeds in drug discovery and development. therefore, it would be beneficial to utilize these data for pbpk modelling (section 2.8). each module in an oa pbpk model can be confirmed by comparison with corresponding in vitro experimental data covering a wider range than in vivo conditions. for example, the solubility model should be validated by an experimental ph-solubility profile in the range of ph 1.0 to 8.0 to cover in vivo gastrointestinal conditions. the dissolution model (the nernst-noyes-whitney equation) can be confirmed admet & dmpk 9(2) (2021) 75-109 lost in modelling and simulation? doi: https:/doi.org/10.5599/admet.923 95 by comparison with dissolution test data. if the prediction result deviated from the experimental observation, an empirical correction parameter can be introduced, like the z factor for each dissolution condition [81]. similarly, in vivo animal pk data can be used to inform the confidence level of a pbpk model. however, in this case, the refinement (back-calculation) process requires the same cautions as the local middle-out approach. in addition, there may be species differences in the back-calculated parameter. first-in-human pk data can be used to inform the confidence level of a pbpk model in humans. the pk data at a low dose (dose number << 1) can be used to evaluate the oral absorption process without interference from the dissolution rate and solubility. in the case of high permeability drugs, the low dose pk data may provide information regarding systemic clearance and volume of distribution to some extent (but i.v. pk data is preferable). the confidence level of solubility and permeability values (as pn/do, eq. 2) may also be evaluated by the dose sub-proportionality of auc (so that relative bioavailability between low and high doses). however, as discussed above, it is not easy to accurately back-calculate (refine) a parameter from in vivo oral pk data. extreme caution should be exercised when using a middle-out approach. 3.13. what is the difference between metabolic ddi and food effect predictions? a local middle-out approach has been successfully used to predict metabolic ddi by pbpk modelling [12,160,161]. because it is practically impossible to clinically evaluate all ddi combinations, ddi prediction by a local middle-out approach would be of great value. in the local middle-out approach for ddi, to identify the fraction of metabolic clearance (fm) (e.g., a cyp isozyme), a specific inhibitor has been used in a clinical pk study (section 3.8 (i)). in the following, the concept of ddi prediction by the local middle-out approach is briefly explained (see [12,160,161] for details). in the case of i.v. dosing (or f = 1 for oral dosing), cl and auc in the absence of an inhibitor can be expressed as:  ma mb ma mb, 1cl cl f f f f     ma mb auc dose cl f f   where fma and fmb are the fractions of metabolic clearance pathways (a and b, respectively). the fmb can be calculated from the auc ratio in the absence (aucno-inhibition) and presence (aucinhibition, fmb = 0) of a specific strong inhibitor as: no-inhibition mb inhibition auc 1 auc f   the inhibitor should specifically inhibit only pathway b and there is no other underlying clearance mechanism (e.g. renal clearance) or uptake/efflux transporter in interplay which can be influenced by the inhibitor. in this case, the fmb value is identifiable from the auc ratio. this equation also suggests that the auc ratio is predictable without using any complex pbpk modelling. a simple prediction scheme for metabolic ddi (auc ratio) has been proposed and thoroughly validated using a large number of clinical ddi data (section 2.8) [67,162–164]. however, a similar approach cannot be simply applied to the food effect prediction, because a food intake simultaneously affects various processes of oral drug absorption (section 2.7) [165]. 3.14. what is the difference between pbpk and compartmental pk models? https://doi.org/10.5599/admet.923 kiyohiko sugano admet & dmpk 9(2) (2021) 75-109 96 in pharmacokinetics, the compartmental pk model is widely used. although both compartmental pk models and pbpk models belong to mathematical models, their applications are different. in compartmental pk models, all parameters are calculated from clinical pk data (“top-down” approach) (section 3.1). because the compartmental pk model is empirical, it is used following the good practice of empirical modelling (section 3.1). to avoid overfitting (section 2.2, figure 3), the akaike information criterion has been used to select the number of compartments appropriately [69]. for an i.v. pk model, the number of compartments is set to be one or two in most cases (2 and 4 parameters, respectively). for an oral pk model, only three parameters, that is, kel, ka, and vd /f are used in most cases (figure 7). figure 7. an oral one-compartment model with an finite absorption time of t = 3.5 h [86,166]. in many cases, an oral cp time profile can be summarized into three parameters, ka, the elimination rate constant (kel), and the lumped values of the volume of distribution (vd), fg, and fh (vd /fgfh). tmax, cmax, and auc are all described by using only one absorption parameter (ka) that represents the oral absorption of a drug. therefore, the oral absorption of a drug can be well described by fitting one of many conjugated parameters in a complex oa pbpk model (the degree of freedom is zero). from the cp -time profile after oral administration, only the composite parameter of vd/fgfh is identifiable. an i.v. data is required to fix vd (and cl) [33]. in contrast, a pbpk model consists of dozens to hundreds of parameters (section 2.1) [37]. all these parameters cannot be identified from the cp -time profile alone (section 3.8). 3.15. global middle-out approach (system parameter estimation) in pbpk modelling, it is preferable to use experimentally measured physiological parameters. however, some of the physiological parameters are not available. in such a case, a global middle-out approach has been used to back estimate a physiological parameter from the multiple pk data of multiple drugs (cf. a local middle-out approach is on a drug-by-drug basis) (section 3.1) [83,90,129,145,167]. in this approach, the following three points are usually carefully considered to ensure parameter identifiability and avoid overfitting: (i) enough number of data covering a wide range of dependent and independent variables, (ii) in vivo pk data that is sensitive to the physiological parameter (and not sensitive to the other parameters), (iii) comparison with directly or indirectly measured values. for example, the volume of the small intestinal fluid (vsi) available for drug dissolution has not been clear, because only free water can be directly measured by mri [28,29]. the vsi value available for drug dissolution was back-calculated from the clinical pk data of several low solubility drugs at various dose strengths covering do < 1 and do > 1 [83,167]. the estimated vsi values (130 ml [83] and 116 ml [167]) admet & dmpk 9(2) (2021) 75-109 lost in modelling and simulation? doi: https:/doi.org/10.5599/admet.923 97 were about 1.5 to 2-fold larger than the average free water volume directly measured by various techniques [26,28,29]. df in eq. 6 was obtained from the relationship between the clinical fa and peff values of about 20 high solubility drugs [55] (cf. fa = 1 – exp( 2df/rsi x peff x tsi). conclusion: strategy to improve oral absorption physiologically-based pharmacokinetic modelling multidisciplinary collaboration a good collaboration of pharmacokinetic, physical chemistry, formulation, and physiology experts is required to improve oral absorption (oa) physiologically-based pharmacokinetic (pbpk) modelling. there are many opportunities for both experimentalists and modellers to take advantage of collaborative works in this field [32]. the critical role of physical chemistry in oa-pbpk modelling the importance of physical chemistry in pbpk modelling has often been overlooked (section 2.7). physical chemistry plays a central role in oral drug absorption [86]. in addition, physical chemistry is also important for hepatic clearance, renal clearance, and tissue distribution (including the brain) [85,168–176]. a recent survey suggested that a poor understanding of physical chemistry is one of the reasons for the prediction failure of oa pbpk modelling [6]. a good understanding of the chemical equilibrium [177], nucleation theory [115,178], and fluid dynamics (including mass transport) [172,173,179] is required in oa pbpk modelling. fortunately, physical chemists are generally well trained in mathematics and mechanistic modelling. they are familiar with the concept of parameter identifiability. the critical role of physiological parameters in oa pbpk modelling similarly, an in-depth understanding of gastrointestinal physiology is critically important [165,180,181]. physiological parameters reported in the literature have large variations, for example, in gastric ph, intestinal ph, fluid volumes, bile concentration, and buffer capacity [27,182]. it is a combination of experimental uncertainty, true inter-subject variability and true inter-occasion variability. the buffer capacity of compendial dissolution media (phosphate buffer) is markedly higher than the real intestinal fluid (bicarbonate buffer), affecting the dissolution profiles of drugs [61,183–186]. the intestinal fluid volume (vsi) would be much smaller than originally thought [25–29] (note that a vsi value (212 ml) had already been reported as early as 1957) [26,27]. the gastric and intestinal phs of dogs are significantly different from those in humans [50,187–190]. harmonization in the future, it is desirable to harmonize physiological parameters and model equations for regulatory submission (section 1.2 and 2.1). in addition, drug intrinsic parameters such as s0, p trans0, and kbm (section 2.1), should be obtained using harmonized procedures [109]. currently, the prediction characteristics of commercial software products vary from product to product [5]. if two software products predict different outcomes, which one should be used for regulatory purposes? in the field of biopharmaceutics, the procedures and conditions for dissolution tests [191,192], in vitro – in vivo correlation [193,194], the biowaiver scheme [195], etc. have been harmonized and standardized. similarly, oa pbpk modelling should be harmonized based on proper model evaluation in the future (section 2.4). we may also need a good simulation practice when pbpk modelling is used to waive a clinical study [35,36]. https://doi.org/10.5599/admet.923 kiyohiko sugano admet & dmpk 9(2) (2021) 75-109 98 not to be lost in modelling and simulation “the greatest enemy of knowledge is not ignorance; it is the illusion of knowledge” daniel boorstin/ stephen hawking when the “bottom-up” simulation deviates from the clinical data, simply press the "optimize" button and the monitor will display a simulation curve that exactly matches the clinical plasma concentration (cp) time data. if we continue publishing this as a successful "prediction", it will eventually cause the illusion of a prediction paradise. we must exert extreme caution not to be lost in modelling and simulation. model equations, physiological parameters, and drug parameters must be disclosed to ensure proper peer-review and reproducibility. a systematic evaluation of predictive power is required to avoid publication bias. when a middle-out approach is pursued, the fitted cp time curve must be labelled as "fitted" (not "predicted"). parameter identifiability should be carefully considered. the optimized model must be validated using independent clinical pk data of the context of use. finally, and most importantly, pharmacokinetic, physical chemistry, formulation, and physiology experts should work together so that not to get lost in modelling and simulation. as mentioned in the introduction, when used correctly, oa pbpk modeling will be an excellent tool for understanding and predicting the oral absorption of a drug. this article will hopefully enhance the science of oa pbpk modelling in the future. acknowledgement i would like to thank prof. kin tam and prof. zoran mandić for giving me this opportunity to write this article. i would like to thank dr. alex avdeef, dr. zahari vinarov, and prof. shinji yamashita for their valuable and encouraging comments. this article is inspired by the discussions in the consortium of biopharmaceutical tools (cobito) project in consortium of pharmaceutical technology, research center for drug discovery and development science at ritsumeikan university. the author would like to thank all cobito members, especially mr. naoya matsumura, mr. yoshiyuki akiyama, ms. asami ono, and mr. shun hayashi. the cobito members are from asahi kasei pharma corporation, daiichi sankyo company, limited, daiichi sankyo rd novare co., ltd., japan machinery co., ltd., japan tobacco inc., mitsubishi tanabe pharma corporation, nippon boehringer ingelheim co., ltd., ono pharmaceutical co., ltd., ritsumeikan university, sawai pharmaceutical co., ltd., shionogi & co., ltd., sumitomo dainippon pharma co., ltd., and towa pharmaceutical co., ltd. (in alphabetical order). conflict of interest kiyohiko sugano is the founder of bioavailabilitydesign llc. references [1] h. van de waterbeemd, e. gifford. admet in silico modelling: towards prediction paradise? nat. rev. drug discov. 2 (2003) 192–204. [2] a.s. darwich, a. margolskee, x. pepin, l. aarons, a. galetin, a. rostami-hodjegan, s. carlert, m. hammarberg, c. hilgendorf, p. johansson, e. karlsson, d. murphy, c. tannergren, h. thörn, m. yasin, f. mazuir, o. nicolas, s. ramusovic, c. xu, s.m. pathak, t. korjamo, j. laru, j. malkki, s. pappinen, j. tuunainen, j. dressman, s. hansmann, e. kostewicz, h. he, t. heimbach, f. wu, c. hoft, y. pang, m.b. admet & dmpk 9(2) (2021) 75-109 lost in modelling and simulation? doi: https:/doi.org/10.5599/admet.923 99 bolger, e. huehn, v. lukacova, j.m. mullin, k.x. szeto, c. costales, j. lin, m. mcallister, s. modi, c. rotter, m. varma, m. wong, a. mitra, j. bevernage, j. biewenga, a. van peer, r. lloyd, c. shardlow, p. langguth, i. mishenzon, m.a. nguyen, j. brown, h. lennernäs, b. abrahamsson. imi – oral biopharmaceutics tools project – evaluation of bottom-up pbpk prediction success part 3: identifying gaps in system parameters by analysing in silico performance across different compound classes. eur. j. pharm. sci. 96 (2017) 626–642. [3] e. sjögren, h. thörn, c. tannergren. reply to ’comment on “in silico modeling of gastrointestinal drug absorption: predictive performance of three physiologically based absorption models.” mol. pharm. 14 (2017) 340–343. [4] e. sjögren, h. thorn, c. tannergren. in silico modeling of gastrointestinal drug absorption: predictive performance of three physiologically based absorption models. mol. pharm. 13 (2016) 1763–1778. [5] n. matsumura, s. hayashi, y. akiyama, a. ono, s. funaki, n. tamura, t. kimoto, m. jiko, y. haruna, a. sarashina, others. prediction characteristics of oral absorption simulation software evaluated using structurally diverse low-solubility drugs. j. pharm. sci. 109 (2020) 1403–1416. [6] a. ahmad, x. pepin, l. aarons, y. wang, a.s. darwich, j.m. wood, c. tannergren, e. karlsson, c. patterson, h. thörn, others. imi--oral biopharmaceutics tools project--evaluation of bottom-up pbpk prediction success part 4: prediction accuracy and software comparisons with improved data and modelling strategies. eur. j. pharm. biopharm. (2020). [7] f.l. wood, j.b. houston, d. hallifax. clearance prediction methodology needs fundamental improvement: trends common to rat and human hepatocytes/microsomes and implications for experimental methodology. drug metab. dispos. 45 (2017) 1178–1188. [8] f.l. wood, j.b. houston, d. hallifax. importance of the unstirred water layer and hepatocyte membrane integrity in vitro for quantification of intrinsic metabolic clearance. drug metab. dispos. 46 (2018) 268–278. [9] d.m. mudie, n. samiei, d.j. marshall, g.e. amidon, c.a.s. bergström. selection of in vivo predictive dissolution media using drug substance and physiological properties. aaps j. 22 (2020) 34. [10] c. kuemmel, y. yang, x. zhang, j. florian, h. zhu, m. tegenge, s.-m. huang, y. wang, t. morrison, i. zineh. consideration of a credibility assessment framework in model-informed drug development: potential application to physiologically-based pharmacokinetic modeling and simulation. cpt pharmacometrics syst. pharmacol. 9 (2020) 21–28. [11] m. li, p. zhao, y. pan, c. wagner. predictive performance of physiologically based pharmacokinetic models for the effect of food on oral drug absorption: current status. cpt pharmacometrics syst. pharmacol. 7 (2018) 82–89. [12] m. shebley, p. sandhu, a. emami riedmaier, m. jamei, r. narayanan, a. patel, s.a. peters, v.p. reddy, m. zheng, l. de zwart, m. beneton, f. bouzom, j. chen, y. chen, y. cleary, c. collins, g.l. dickinson, n. djebli, h.j. einolf, i. gardner, f. huth, f. kazmi, f. khalil, j. lin, a. odinecs, c. patel, h. rong, e. schuck, p. sharma, s.p. wu, y. xu, s. yamazaki, k. yoshida, m. rowland. physiologically based pharmacokinetic model qualification and reporting procedures for regulatory submissions: a consortium perspective. clin. pharmacol. ther. 104 (2018) 88–110. [13] u.s.f. and d. administration. physiologically based pharmacokinetic analyses — format and content. guidance. (2018). [14] j. dibella, j.e. sager, j. yu, i. ragueneau-majlessi, n. isoherranen, x. zhuang, c. lu, a.b. pharmaceutica sinica, european medicines agency committee for medicinal products for human use, a. nordmark, u.s. department of health and human services food and drug administration cder, e. medicines agency. guideline on the qualification and reporting of physiologically based pharmacokinetic (pbpk) modelling and simulation. acta pharm. sin. b. 44 (2016) 27–29. http://www.ema.europa.eu/docs/en_gb/document_library/scientific_guideline/2016/07/wc50021 1315.pdf%0awww.elsevier.com/locate/apsb%0awww.sciencedirect.com%0ahttp://dx.doi.org/10.10 16/j.apsb.2016.04.004. [15] d.l. sackett, w.m.c. rosenberg, j.a.m. gray, r.b. haynes, w.s. richardson. evidence based medicine: what it is and what it isn’t. (1996). https://doi.org/10.5599/admet.923 http://www.ema.europa.eu/docs/en_gb/document_library/scientific_guideline/2016/07/wc500211315.pdf%0awww.elsevier.com/locate/apsb%0awww.sciencedirect.com%0ahttp:/dx.doi.org/10.1016/j.apsb.2016.04.004 http://www.ema.europa.eu/docs/en_gb/document_library/scientific_guideline/2016/07/wc500211315.pdf%0awww.elsevier.com/locate/apsb%0awww.sciencedirect.com%0ahttp:/dx.doi.org/10.1016/j.apsb.2016.04.004 http://www.ema.europa.eu/docs/en_gb/document_library/scientific_guideline/2016/07/wc500211315.pdf%0awww.elsevier.com/locate/apsb%0awww.sciencedirect.com%0ahttp:/dx.doi.org/10.1016/j.apsb.2016.04.004 kiyohiko sugano admet & dmpk 9(2) (2021) 75-109 100 [16] i. masic, m. miokovic, b. muhamedagic. evidence based medicine--new approaches and challenges. acta inform. medica. 16 (2008) 219. [17] m. li, p. zhao, y. pan, c. wagner. predictive performance of physiologically based pharmacokinetic models for the effect of food on oral drug absorption: current status. cpt pharmacometrics syst. pharmacol. 7 (2018) 82–89. [18] k. jaqaman, g. danuser. linking data to models: data regression. nat. rev. mol. cell biol. 7 (2006) 813–819. [19] w. slob, p.h.m. janssen, j.m. den hof. structural identifiability of pbpk models: practical consequences for modeling strategies and study designs. crit. rev. toxicol. 27 (1997) 261–272. [20] s.a. peters, h. dolgos. requirements to establishing confidence in physiologically based pharmacokinetic (pbpk) models and overcoming some of the challenges to meeting them. clin. pharmacokinet. 58 (2019) 1355–1371. [21] n. tsamandouras, a. rostami-hodjegan, l. aarons. combining the “bottom up” and “top down” approaches in pharmacokinetic modelling: fitting pbpk models to observed clinical data. br. j. clin. pharmacol. 79 (2015) 48–55. [22] j.w.t. yates. structural identifiability of physiologically based pharmacokinetic models. j. pharmacokinet. pharmacodyn. 33 (2006) 421–439. [23] r. muñoz-tamayo, l. puillet, j.-b. daniel, d. sauvant, o. martin, m. taghipoor, p. blavy. to be or not to be an identifiable model. is this a relevant question in animal science modelling? animal. 12 (2018) 701–712. [24] r. bellman, k.j. åström. on structural identifiability. math. biosci. 7 (1970) 329–339. [25] d.b. turner, b. liu, n. patel, s.m. pathak, s. polak, m. jamei, j. dressman, a. rostami-hodjegan. comment on “in silico modeling of gastrointestinal drug absorption: predictive performance of three physiologically-based absorption models.” mol. pharm. 14 (2017) 336–339. [26] f. gotch, j. nadell, i.s. edelman, others. gastrointestinal water and electrolytes. iv. the equilibration of deuterium oxide (d 2 o) in gastrointestinal contents and the proportion of total body water (tbw) in the gastrointestinal tract. j. clin. invest. 36 (1957) 289–296. [27] e.l. mcconnell, h.m. fadda, a.w. basit. gut instincts: explorations in intestinal physiology and drug delivery. int. j. pharm. 364 (2008) 213–226. [28] d.m. mudie, g.l. amidon, g.e. amidon, k. murray, c.l. hoad, s.e. pritchard, p.a. gowland, m.c. garnett, r.c. spiller, l. marciani. quantification of gastrointestinal liquid volumes and distribution following a 240 ml dose of water in the fasted state. mol. pharm. 11 (2014) 3039–3047. [29] c. schiller, c.p. fröhlich, t. giessmann, w. siegmund, h. mönnikes, n. hosten, w. weitschies. intestinal fluid volumes and transit of dosage forms as assessed by magnetic resonance imaging. aliment. pharmacol. ther. 22 (2005) 971–979. [30] f.l. wood, j.b. houston, d. hallifax. clearance prediction methodology needs fundamental improvement: trends common to rat and human hepatocytes/microsomes and implications for experimental methodology. drug metab. dispos. 45 (2017) 1178–1188. [31] h. lennernäs. animal data: the contributions of the ussing chamber and perfusion systems to predicting human oral drug delivery in vivo. adv. drug deliv. rev. 59 (2007) 1103–1120. [32] m. banwarth-kuhn, s. sindi. how and why to build a mathematical model: a case study using prion aggregation. j. biol. chem. 295 (2020) 5022–5035. [33] j.w.t. yates, r.d.o. jones, m. walker, s.y.a. cheung. structural identifiability and indistinguishability of compartmental models. expert opin. drug metab. toxicol. 5 (2009) 295–302. [34] c. cobelli, j.j. distefano 3rd. parameter and structural identifiability concepts and ambiguities: a critical review and analysis. am. j. physiol. integr. comp. physiol. 239 (1980) r7--r24. [35] k. sugano. introduction to computational oral absorption simulation. expert opin. drug metab. toxicol. 5 (2009). admet & dmpk 9(2) (2021) 75-109 lost in modelling and simulation? doi: https:/doi.org/10.5599/admet.923 101 [36] k. sugano. biopharmaceutics modeling and simulations: theory, practice, methods, and applications, 2012. [37] l. kuepfer, c. niederalt, t. wendl, j.-f. schlender, s. willmann, j. lippert, m. block, t. eissing, d. teutonico. applied concepts in pbpk modeling: how to build a pbpk/pd model. cpt pharmacometrics syst. pharmacol. 5 (2016) 516–531. [38] k. sugano. emerging full mechanistic physiologically based modeling, 2014. [39] e.g. rippie, d.j. lamb, p.w. romig. solubilization of weakly acidic and basic drugs by aqueous solutions of polysorbate 80. j. pharm. sci. 53 (1964) 1346–1348. [40] j. jinno, d. oh, j.r. crison, g.l. amidon. dissolution of ionizable water-insoluble drugs: the combined effect of ph and surfactant. j. pharm. sci. 89 (2000) 268–274. [41] a.a. noyes, w.r. whitney. the rate of solution of solid substances in their own solutions. j. am. chem. soc. 19 (1897) 930–934. [42] k.g. mooney, m.a. mintun, k.j. himmelstein, v.j. stella. dissolution kinetics of carboxylic acids i: effect of ph under unbuffered conditions. j. pharm. sci. 70 (1981) 13–22. [43] k.g. mooney, m.a. mintun, k.j. himmelstein, v.j. stella. dissolution kinetics of carboxylic acids ii: effect of buffers. j. pharm. sci. 70 (1981) 22–32. [44] a. okazaki, t. mano, k. sugano. theoretical dissolution model of poly-disperse drug particles in biorelevant media. j. pharm. sci. 97 (2008). [45] r.j. hintz, k.c. johnson. the effect of particle size distribution on dissolution rate and oral absorption. int. j. pharm. 51 (1989) 9–17. [46] j. wang, d.r. flanagan. general solution for diffusion-controlled dissolution of spherical particles. 1. theory. j. pharm. sci. 88 (1999) 731–738. [47] k. sugano. theoretical comparison of hydrodynamic diffusion layer models used for dissolution simulation in drug discovery and development. int. j. pharm. 363 (2008) 73–77. [48] s.s. ozturk, b.o. palsson, j.b. dressman. dissolution of lonizable drugs in buffered and unbuffered solutions. pharm. res. 5 (1988) 272–282. [49] k. sugano. estimation of effective intestinal membrane permeability considering bile micelle solubilisation. int. j. pharm. 368 (2009) 116–122. [50] y. akiyama, t. kimoto, h. mukumoto, s. miyake, s. ito, t. taniguchi, y. nomura, n. matsumura, t. fujita, k. sugano. prediction accuracy of mechanism-based oral absorption model for dogs. j. pharm. sci. 108 (2019) 2728–2736. [51] y. akiyama, s. ito, t. fujita, k. sugano. prediction of negative food effect induced by bile micelle binding on oral absorption of hydrophilic cationic drugs. eur. j. pharm. sci. (2020) 105543. [52] k. sugano. theoretical investigation of passive intestinal membrane permeability using monte carlo method to generate drug-like molecule population. int. j. pharm. 373 (2009). [53] k. sugano, m. kataoka, c. da costa mathews, s. yamashita. prediction of food effect by bile micelles on oral drug absorption considering free fraction in intestinal fluid. eur. j. pharm. sci. 40 (2010). [54] d. pade, m. jamei, a. rostami-hodjegan, d.b. turner. application of the mechpeff model to predict passive effective intestinal permeability in the different regions of the rodent small intestine and colon. biopharm. drug dispos. 38 (2017) 94–114. [55] k. sugano. estimation of effective intestinal membrane permeability considering bile micelle solubilisation. int. j. pharm. 368 (2009). [56] a. dokoumetzidis, p. macheras. a century of dissolution research: from noyes and whitney to the biopharmaceutics classification system. int. j. pharm. 321 (2006) 1–11. [57] m. jamei, b. abrahamsson, j. brown, j. bevernage, m.b. bolger, t. heimbach, e. karlsson, e. kotzagiorgis, a. lindahl, m. mcallister, others. current status and future opportunities for incorporation of dissolution data in pbpk modeling for pharmaceutical development and regulatory applications: orbito consortium commentary. eur. j. pharm. biopharm. 155 (2020) 55–68. https://doi.org/10.5599/admet.923 kiyohiko sugano admet & dmpk 9(2) (2021) 75-109 102 [58] m. mcallister. dynamic dissolution: a step closer to predictive dissolution testing? mol. pharm. 7 (2010) 1374–1387. [59] b.j. krieg, s.m. taghavi, g.l. amidon, g.e. amidon. in vivo predictive dissolution: transport analysis of the co2, bicarbonate in vivo buffer system. j. pharm. sci. 103 (2014) 3473–3490. [60] h. lennernäs, a. lindahl, a. van peer, c. ollier, t. flanagan, r. lionberger, a. nordmark, s. yamashita, l. yu, g.l. amidon, v. fischer, e. sjögren, p. zane, m. mcallister, b. abrahamsson. in vivo predictive dissolution (ipd) and biopharmaceutical modeling and simulation: future use of modern approaches and methodologies in a regulatory context. mol. pharm. 14 (2017) 1307–1314. [61] b.j. krieg, s.m. taghavi, g.l. amidon, g.e. amidon. in vivo predictive dissolution: comparing the effect of bicarbonate and phosphate buffer on the dissolution of weak acids and weak bases. j. pharm. sci. 104 (2015) 2894–2904. [62] g. shmueli, others. to explain or to predict? stat. sci. 25 (2010) 289–310. [63] m. weisberg. simulation and similarity: using models to understand the world, oup usa, 2013. https://books.google.co.jp/books?id=rdu5e532mioc. [64] m. weisberg. simulation and similarity: using models to understand the world, oup usa, 2013. [65] d. orrell. apollo’s arrow, harpercollins, 2008. https://books.google.co.jp/books?id=snmphwaacaaj. [66] a. hisaka, y. ohno, t. yamamoto, h. suzuki. theoretical considerations on quantitative prediction of drug-drug interactions. drug metab. pharmacokinet. 25 (2010) 48–61. [67] m. tod, p.b. pierrillas, l. bourguignon, s. goutelle. comparison of the static in vivo approach to a physiologically based pharmacokinetic approach for metabolic drug--drug interactions prediction. int. j. pharmacokinet. 1 (2016) 25–34. [68] e.i. ette, p.j. williams, e.i. ette, p.j. williams. pharmacometrics : the science of quantitative pharmacology / edited by ene i. ette, anoixis corporation, paul j. williams, university of the pacific and anoixis corporation., wiley-interscience, hoboken, new jersey, 2007. [69] k. yamaoka, t. nakagawa, t. uno. application of akaike’s information criterion (aic) in the evaluation of linear pharmacokinetic equations. j. pharmacokinet. biopharm. 6 (1978) 165–175. [70] r. takano, k. furumoto, k. shiraki, n. takata, y. hayashi, y. aso, s. yamashita. rate-limiting steps of oral absorption for poorly water-soluble drugs in dogs; prediction from a miniscale dissolution test and a physiologically-based computer simulation. pharm. res. 25 (2008) 2334–2344. [71] l.x. yu. an integrated model for determining causes of poor oral drug absorption. pharm. res. 16 (1999) 1883–1887. [72] k. sugano, k. terada. rateand extent-limiting factors of oral drug absorption: theory and applications. j. pharm. sci. 104 (2015). [73] k. sugano, y. nabuchi, m. machida, y. aso. prediction of human intestinal permeability using artificial membrane permeability. int. j. pharm. 257 (2003). [74] a. avdeef, k.y. tam. how well can the caco-2/madin-darby canine kidney models predict effective human jejunal permeability? j. med. chem. 53 (2010) 3566–3584. [75] a.e. riedmaier, k. dement, j. huckle, p. bransford, c. stillhart, r. lloyd, r. alluri, s. basu, y. chen, v. dhamankar, others. use of physiologically based pharmacokinetic (pbpk) modeling for predicting drug-food interactions: an industry perspective. aaps j. 22 (2020) 1–15. [76] s. j.e., y. j., r.-m. i., i. n. physiologically based pharmacokinetic (pbpk) modeling and simulation approaches: a systematic review of published models, applications, and model verification. drug metab. dispos. 43 (2015) 1823–1837. [77] v.p. shah, g.l. amidon. g.l. amidon, h. lennernas, v.p. shah, and j.r. crison. a theoretical basis for a biopharmaceutic drug classification: the correlation of in vitro drug product dissolution and in vivo bioavailability, pharm res 12, 413–420, 1995—backstory of bcs. aaps j. 16 (2014) 894–898. [78] a. margolskee, a.s. darwich, x. pepin, l. aarons, a. galetin, a. rostami-hodjegan, s. carlert, m. hammarberg, c. hilgendorf, p. johansson, e. karlsson, d. murphy, c. tannergren, h. thörn, m. yasin, f. mazuir, o. nicolas, s. ramusovic, c. xu, s.m. pathak, t. korjamo, j. laru, j. malkki, s. pappinen, j. admet & dmpk 9(2) (2021) 75-109 lost in modelling and simulation? doi: https:/doi.org/10.5599/admet.923 103 tuunainen, j. dressman, s. hansmann, e. kostewicz, h. he, t. heimbach, f. wu, c. hoft, l. laplanche, y. pang, m.b. bolger, e. huehn, v. lukacova, j.m. mullin, k.x. szeto, c. costales, j. lin, m. mcallister, s. modi, c. rotter, m. varma, m. wong, a. mitra, j. bevernage, j. biewenga, a. van peer, r. lloyd, c. shardlow, p. langguth, i. mishenzon, m.a. nguyen, j. brown, h. lennernäs, b. abrahamsson. imi – oral biopharmaceutics tools project – evaluation of bottom-up pbpk prediction success part 2: an introduction to the simulation exercise and overview of results. eur. j. pharm. sci. 96 (2017) 610–625. [79] n. matsumura, a. ono, y. akiyama, t. fujita, k. sugano. bottom-up physiologically based oral absorption modeling of free weak base drugs. pharmaceutics. 12 (2020) 844. [80] m.v.s. varma, r.s. obach, c. rotter, h.r. miller, g. chang, s.j. steyn, a. el-kattan, m.d. troutman. physicochemical space for optimum oral bioavailability: contribution of human intestinal absorption and first-pass elimination. j. med. chem. 53 (2010) 1098–1108. [81] r. takano, k. sugano, a. higashida, y. hayashi, m. machida, y. aso, s. yamashita. oral absorption of poorly water-soluble drugs: computer simulation of fraction absorbed in humans from a miniscale dissolution test. pharm. res. 23 (2006) 1144–1156. [82] r. takano, k. sugano, a. higashida, y. hayashi, m. machida, y. aso, s. yamashita. oral absorption of poorly water-soluble drugs: computer simulation of fraction absorbed in humans from a miniscale dissolution test. pharm. res. 23 (2006). [83] k. sugano. fraction of a dose absorbed estimation for structurally diverse low solubility compounds. int. j. pharm. 405 (2011). [84] n. hisada, r. takano, n. takata, k. shiraki, t. ueto, s. tanida, m. kataoka, s. yamashita. characterizing the dissolution profiles of supersaturable salts, cocrystals, and solvates to enhance in vivo oral absorption. eur. j. pharm. biopharm. 103 (2016) 192–199. [85] k. sugano, m. kansy, p. artursson, a. avdeef, s. bendels, l. di, g.f. ecker, b. faller, h. fischer, g. gerebtzoff, h. lennernaes, f. senner. coexistence of passive and carrier-mediated processes in drug transport. nat. rev. drug discov. 9 (2010). [86] k. sugano. biopharmaceutics modeling and simulations: theory, practice, methods, and applications, wiley, n.d. [87] p. wils, a. warnery, v. phung-ba, s. legrain, d. scherman. high lipophilicity decreases drug transport across intestinal epithelial cells. j. pharmacol. exp. ther. 269 (1994) 654–658. [88] g. krishna, k.j. chen, c.c. lin, a.a. nomeir. permeability of lipophilic compounds in drug discovery using in-vitro human absorption model, caco-2. int. j. pharm. 222 (2001) 77–89. [89] r. saitoh, k. sugano, n. takata, t. tachibana, a. higashida, y. nabuchi, y. aso. correction of permeability with pore radius of tight junctions in caco-2 monolayers improves the prediction of the dose fraction of hydrophilic drugs absorbed by humans. pharm. res. 21 (2004). [90] k. sugano, n. takata, m. machida, k. saitoh, k. terada. prediction of passive intestinal absorption using bio-mimetic artificial membrane permeation assay and the paracellular pathway model. int. j. pharm. 241 (2002). [91] g.e. amidon, w.i. higuchi, n.f.h. ho. theoretical and experimental studies of transport of micellesolubilized solutes. j. pharm. sci. 71 (1982) 77–84. [92] t. yamaguchi, c. ikeda, y. sekine. intestinal absorption of $β$-adrenergic blocking agent nadolol. i.: comparison of absorption behaivor of nadolol with those of other $β$-blocking agents in rats. chem. pharm. bull. 34 (1986) 3362–3369. [93] f.g.j. poelma, r. breäs, j.j. tukker, d.j.a. crommelin. intestinal absorption of drugs. the influence of mixed micelles on on the disappearance kinetics of drugs from the small intestine of the rat. j. pharm. pharmacol. 43 (1991) 317–324. [94] f. ingels, b. beck, m. oth, p. augustijns. effect of simulated intestinal fluid on drug permeability estimation across caco-2 monolayers. int. j. pharm. 274 (2004) 221–232. [95] a.t. heikkinen, j. mönkkönen, t. korjamo. determination of permeation resistance distribution in in vitro cell monolayer permeation experiments. eur. j. pharm. sci. 40 (2010) 132–142. https://doi.org/10.5599/admet.923 kiyohiko sugano admet & dmpk 9(2) (2021) 75-109 104 [96] f.a. wilson, v.l. sallee, j.m. dietschy. unstirred water layers in intestine: rate determinant of fatty acid absorption from micellar solutions. science (80-. ). 174 (1971) 1031–1033. [97] u. fagerholm, h. lennernäs. experimental estimation of the effective unstirred water layer thickness in the human jejunum, and its importance in oral drug absorption. eur. j. pharm. sci. 3 (1995) 247– 253. [98] k.a. youdim, a. avdeef, n.j. abbott. in vitro trans-monolayer permeability calculations: often forgotten assumptions. drug discov. today. 8 (2003) 997–1003. [99] a. olivares-morales, h. lennernäs, l. aarons, a. rostami-hodjegan. translating human effective jejunal intestinal permeability to surface-dependent intrinsic permeability: a pragmatic method for a more mechanistic prediction of regional oral drug absorption. aaps j. 17 (2015) 1177–1192. [100] k. sugano, m. kataoka, c. da costa mathews, s. yamashita. prediction of food effect by bile micelles on oral drug absorption considering free fraction in intestinal fluid. eur. j. pharm. sci. 40 (2010) 118– 124. [101] t. loftsson, f. konrádsdóttir, m. másson. influence of aqueous diffusion layer on passive drug diffusion from aqueous cyclodextrin solutions through biological membranes. pharmazie. 61 (2006) 83–89. [102] a. dahan, j.m. miller, a. hoffman, g.e. amidon, g.l. amidon. the solubility--permeability interplay in using cyclodextrins as pharmaceutical solubilizers: mechanistic modeling and application to progesterone. j. pharm. sci. 99 (2010) 2739–2749. [103] r.e. oliver, a.f. jones, m. rowland. what surface of the intestinal epithelium is effectively available to permeating drugs? j. pharm. sci. 87 (1998) 634–639. [104] a. avdeef, p.e. nielsen, o. tsinman. pampa a drug absorption in vitro model: 11. matching the in vivo unstirred water layer thickness by individual-well stirring in microtitre plates. eur. j. pharm. sci. 22 (2004) 365–374. [105] m. vertzoni, e. pastelli, d. psachoulias, l. kalantzi, c. reppas. estimation of intragastric solubility of drugs: in what medium? pharm. res. 24 (2007) 909–917. [106] a.t.m. serajuddin, c.i. jarowski. effect of diffusion layer ph and solubility on the dissolution rate of pharmaceutical bases and their hydrochloride salts i: phenazopyridine. j. pharm. sci. 74 (1985) 142– 147. [107] j. al-gousous, n. salehi, g.e. amidon, r.m. ziff, p. langguth, g.l. amidon. mass transport analysis of bicarbonate buffer: effect of the co2-h2co3 hydration-dehydration kinetics in the fluid boundary layer and the apparent effective p ka controlling dissolution of acids and bases. mol. pharm. 16 (2019) 2626–2635. [108] k. sugano, a. okazaki, s. sugimoto, s. tavornvipas, a. omura, t. mano. solubility and dissolution profile assessment in drug discovery. drug metab. pharmacokinet. 22 (2007) 225–254. [109] a. ono, n. matsumura, t. kimoto, y. akiyama, s. funaki, n. tamura, s. hayashi, y. kojima, m. fushimi, h. sudaki, others. harmonizing solubility measurement to lower inter-laboratory variance--progress of consortium of biopharmaceutical tools (cobito) in japan. admet dmpk 7 (2019) 183–195. https://doi.org/10.5599/admet.704. [110] a. avdeef. suggested improvements for measurement of equilibrium solubility-ph of ionizable drugs. admet dmpk 3 (2015) 84–109. https://doi.org/10.5599/admet.3.2.193. [111] a. avdeef, e. fuguet, a. llinàs, c. ràfols, e. bosch, g. völgyi, t. verbić, e. boldyreva, k. takács-novák. equilibrium solubility measurement of ionizable drugs – consensus recommendations for improving data quality. admet dmpk 4 (2016) 117. https://doi.org/10.5599/admet.4.2.292. [112] t. uekusa, k. sugano. effect of buffer capacity on dissolution and supersaturation profiles of poorly soluble drug salt, in: apsjt annu. meet., toyama, 2019: p. 236. [113] j. oki, d. watanabe, t. uekusa, k. sugano. mechanism of supersaturation suppression in dissolution process of acidic drug salt. mol. pharm. 16 (2019) 1669–1677. [114] k. sugano. a simulation of oral absorption using classical nucleation theory. int. j. pharm. 378 (2009) 142–145. https://doi.org/10.5599/admet.704 https://doi.org/10.5599/admet.3.2.193 https://doi.org/10.5599/admet.4.2.292 admet & dmpk 9(2) (2021) 75-109 lost in modelling and simulation? doi: https:/doi.org/10.5599/admet.923 105 [115] s. ozaki. population balance model for simulation of the supersaturation–precipitation behavior of drugs in supersaturable solid forms. j. pharm. sci. 108 (2019) 260–267. [116] l. lindfors, s. forssén, j. westergren, u. olsson. nucleation and crystal growth in supersaturated solutions of a model drug. j. colloid interface sci. 325 (2008) 404–413. [117] s. ozaki, t. minamisono, t. yamashita, t. kato, i. kushida. supersaturation--nucleation behavior of poorly soluble drugs and its impact on the oral absorption of drugs in thermodynamically high-energy forms. j. pharm. sci. 101 (2012) 214–222. [118] m. warzecha, r. guo, r.m. bhardwaj, s.m. reutzel-edens, s.l. price, d.a. lamprou, a.j. florence. direct observation of templated two-step nucleation mechanism during olanzapine hydrate formation. cryst. growth des. 17 (2017) 6382–6393. [119] p. jakubiak, b. wagner, h.p. grimm, j. petrig-schaffland, f. schuler, r. alvarez-sánchez. development of a unified dissolution and precipitation model and its use for the prediction of oral drug absorption. mol. pharm. 13 (2016) 586–598. [120] a. llinas, i. oprisiu, a. avdeef. findings of the second challenge to predict aqueous solubility. j. chem. inf. model. (2020). [121] k. takács-novák, m. urac, p. horváth, g. völgyi, b.d. anderson, a. avdeef. equilibrium solubility measurement of compounds with low dissolution rate by higuchi’s facilitated dissolution method. a validation study. eur. j. pharm. sci. 106 (2017) 133–141. [122] n. sun, a. avdeef. biorelevant pk a (37°c) predicted from the 2d structure of the molecule and its pk a at 25°c. j. pharm. biomed. anal. 56 (2011) 173–182. [123] a. avdeef. solubility temperature dependence predicted from 2d structure. admet dmpk. 3 (2015) 298–344. https://doi.org/10.5599/admet.3.4.259. [124] s.s. hate, s.m. reutzel-edens, l.s. taylor. absorptive dissolution testing of supersaturating systems: impact of absorptive sink conditions on solution phase behavior and mass transport. mol. pharm. 14 (2017) 4052–4063. [125] d.j. good, r.h. naír. solubility advantage of pharmaceutical cocrystals. cryst. growth des. 9 (2009) 2252–2264. [126] m. omori, t. watanabe, t. uekusa, j. oki, d. inoue, k. sugano. effect of coformer and polymer on particle surface solution-mediated phase transformation of cocrystal in aqueous media. mol. pharm. (2020). [127] b. van eerdenbrugh, g. van den mooter, p. augustijns, j. vermant, j.a. martens, l. froyen, j. van humbeeck. solubility increases associated with crystalline drug nanoparticles: methodologies and significance. mol. pharm. 7 (2010) 1858–1870. [128] s.b. murdande, d.a. shah, r.h. dave. impact of nanosizing on solubility and dissolution rate of poorly soluble pharmaceuticals. j. pharm. sci. 104 (2015) 2094–2102. [129] k. sugano. possible reduction of effective thickness of intestinal unstirred water layer by particle drifting effect. int. j. pharm. 387 (2010). [130] j.f. cuine, c.l. mcevoy, w.n. charman, c.w. pouton, g.a. edwards, h. benameur, c.j.h. porter. evaluation of the impact of surfactant digestion on the bioavailability of danazol after oral administration of lipidic self-emulsifying formulations to dogs. j. pharm. sci. 97 (2008) 995–1012. [131] a.t. larsen, p. sassene, a. müllertz. in vitro lipolysis models as a tool for the characterization of oral lipid and surfactant based drug delivery systems. int. j. pharm. 417 (2011) 245–255. [132] d. orrell. apollo’s arrow, harpercollins, 2008. [133] s.a. peters, p.e. schroeder, n. giri, h. dolgos. evaluation of the use of static and dynamic models to predict drug-drug interaction and its associated variability: impact on drug discovery and early development. drug metab. dispos. 40 (2012) 1495–1507. [134] j.b. dressman, d. fleisher. mixing-tank model for predicting dissolution rate control of oral absorption. j. pharm. sci. 75 (1986) 109–116. https://doi.org/10.5599/admet.923 https://doi.org/10.5599/admet.3.4.259 kiyohiko sugano admet & dmpk 9(2) (2021) 75-109 106 [135] e. nicolaides, m. symillides, j.b. dressman, c. reppas. biorelevant dissolution testing to predict the plasma profile of lipophilic drugs after oral administration. pharm. res. 18 (2001) 380–388. [136] c. wagner, e. jantratid, f. kesisoglou, m. vertzoni, c. reppas, j. b. dressman. predicting the oral absorption of a poorly soluble, poorly permeable weak base using biorelevant dissolution and transfer model tests coupled with a physiologically based pharmacokinetic model. eur. j. pharm. biopharm. 82 (2012) 127–138. [137] s. chakraborty, l. yadav, d. aggarwal. prediction of in vivo drug performance using in vitro dissolution coupled with stella: a study with selected drug products. drug dev. ind. pharm. 41 (2015) 1667–1673. [138] a. kambayashi, t. yasuji, j.b. dressman. prediction of the precipitation profiles of weak base drugs in the small intestine using a simplified transfer (“dumping”) model coupled with in silico modeling and simulation approach. eur. j. pharm. biopharm. 103 (2016) 95–103. [139] a. kambayashi, j.b. dressman. forecasting gastrointestinal precipitation and oral pharmacokinetics of dantrolene in dogs using an in vitro precipitation testing coupled with in silico modeling and simulation. eur. j. pharm. biopharm. 119 (2017) 107–113. [140] a. kambayashi, j.b. dressman. an in vitro-in silico-in vivo approach to predicting the oral pharmacokinetic profile of salts of weak acids: case example dantrolene. eur. j. pharm. biopharm. 84 (2013) 200–207. [141] k. matsui, y. tsume, s. takeuchi, a. searls, g.l. amidon. utilization of gastrointestinal simulator, an in vivo predictive dissolution methodology, coupled with computational approach to forecast oral absorption of dipyridamole. mol. pharm. 14 (2017) 1181–1189. [142] m. yamane, k. matsui, m. sugihara, y. tokunaga. the provisional no-effect threshold of sugar alcohols on oral drug absorption estimated by physiologically based biopharmaceutics model. j. pharm. sci. (2020). [143] a. kambayashi, t. kiyota. a physiologically-based drug absorption modeling for orally disintegrating tablets. eur. j. pharm. biopharm. (2020). [144] y. cao, w.j. jusko. applications of minimal physiologically-based pharmacokinetic models. j. pharmacokinet. pharmacodyn. 39 (2012) 711–723. [145] r. li, h.a. barton, p.d. yates, a. ghosh, a.c. wolford, k.a. riccardi, t.s. maurer. a “middle-out” approach to human pharmacokinetic predictions for oatp substrates using physiologically-based pharmacokinetic modeling. j. pharmacokinet. pharmacodyn. 41 (2014) 197–209. [146] d.-m. oh, r.l. curl, g.l. amidon. estimating the fraction dose absorbed from suspensions of poorly soluble compounds in humans: a mathematical model. pharm. res. 10 (1993) 264–270. [147] h. lennernäs, c.g. regårdh. regional gastrointestinal absorption of the beta-blocker pafenolol in the rat and intestinal transit rate determined by movement of 14c-polyethylene glycol (peg) 4000. pharm. res. an off. j. am. assoc. pharm. sci. 10 (1993) 130–135. [148] d. riethorst, j. brouwers, j. motmans, p. augustijns. human intestinal fluid factors affecting intestinal drug permeation in vitro. eur. j. pharm. sci. 121 (2018) 338–346. [149] k. sugano. calculation of fraction of dose absorbed: comparison between analytical solution based on one compartment steady state concentration approximation and dynamic seven compartment model. chem-bio informatics j. 9 (2009) 75–93. [150] m.v.s. varma, r.s. obach, c. rotter, h.r. miller, g. chang, s.j. steyn, a. el-kattan, m.d. troutman. physicochemical space for optimum oral bioavailability: contribution of human intestinal absorption and first-pass elimination. j. med. chem. 53 (2010) 1098–1108. [151] e. sjögren, d. dahlgren, c. roos, h. lennernäs. human in vivo regional intestinal permeability: quantitation using site-specific drug absorption data. mol. pharm. 12 (2015) 2026–2039. [152] d. psachoulias, m. vertzoni, k. goumas, v. kalioras, s. beato, j. butler, c. reppas. precipitation in and supersaturation of contents of the upper small intestine after administration of two weak bases to fasted adults. pharm. res. 28 (2011) 3145–3158. admet & dmpk 9(2) (2021) 75-109 lost in modelling and simulation? doi: https:/doi.org/10.5599/admet.923 107 [153] b. hens, j. brouwers, m. corsetti, p. augustijns. supersaturation and precipitation of posaconazole upon entry in the upper small intestine in humans. j. pharm. sci. 105 (2016) 2677–2684. [154] k.k.h. chan, f. langenbucher, m. gibaldi. evaluation of in vivo drug release by numerical deconvolution using oral solution data as weighting function. j. pharm. sci. 76 (1987) 446–450. [155] s. clarysse, j. brouwers, j. tack, p. annaert, p. augustijns. intestinal drug solubility estimation based on simulated intestinal fluids: comparison with solubility in human intestinal fluids. eur. j. pharm. sci. 43 (2011) 260–269. [156] k.s. shingo takezawa, kikuo ozawa, rishichi mimura. membrane permeation mechanism of solute in dialyzer. artif. organ. 13 (1984) 1460–1467. [157] a. avdeef. leakiness and size exclusion of paracellular channels in cultured epithelial cell monolayersinterlaboratory comparison. pharm. res. 27 (2010) 480–489. [158] m. kato, k. chiba, a. hisaka, m. ishigami, m. kayama, n. mizuno, y. nagata, s. takakuwa, y. tsukamoto, k. ueda, h. kusuhara, k. ito, y. sugiyama. the intestinal first-pass metabolism of substrates of cyp3a4 and p-glycoprotei—quantitative analysis based on information from the literature. drug metab. pharmacokinet. 18 (2004) 365–372. [159] h. lennernäs. intestinal permeability and its relevance for absorption and elimination, 2007. [160] e. luzon, k. blake, s. cole, a. nordmark, c. versantvoort, e.g. berglund. physiologically based pharmacokinetic modeling in regulatory decision-making at the european medicines agency. clin. pharmacol. ther. 102 (2017) 98–105. [161] m. grimstein, y. yang, x. zhang, j. grillo, s.-m. huang, i. zineh, y. wang. physiologically based pharmacokinetic modeling in regulatory science: an update from the us food and drug administration’s office of clinical pharmacology. j. pharm. sci. 108 (2019) 21–25. [162] a. hisaka, y. ohno, t. yamamoto, h. suzuki. theoretical considerations on quantitative prediction of drug-drug interactions. drug metab. pharmacokinet. 25 (2010) 48–61. [163] y. ohno, a. hisaka, h. suzuki. general framework for the quantitative prediction of cyp3a4-mediated oral drug interactions based on the auc increase by coadministration of standard drugs. clin. pharmacokinet. 46 (2007) 681–696. [164] a. hisaka, y. ohno, t. yamamoto, h. suzuki. prediction of pharmacokinetic drug-drug interaction caused by changes in cytochrome p450 activity using in vivo information. pharmacol. ther. 125 (2010) 230–248. [165] m. koziolek, s. alcaro, p. augustijns, a.w. basit, m. grimm, b. hens, c.l. hoad, p. jedamzik, c.m. madla, m. maliepaard, others. the mechanisms of pharmacokinetic food-drug interactions--a perspective from the ungap group. eur. j. pharm. sci. 134 (2019) 31–59. [166] p. macheras, p. chryssafidis. revising pharmacokinetics of oral drug absorption: i models based on biopharmaceutical/physiological and finite absorption time concepts. pharm. res. 37 (2020) 1–13. [167] a. maharaj, n. fotaki, a. edginton. parameterization of small intestinal water volume using pbpk modeling. eur. j. pharm. sci. 67 (2015) 55–64. [168] l. di, p. artursson, a. avdeef, l.z. benet, j.b. houston, m. kansy, e.h. kerns, h. lennernäs, d.a. smith, k. sugano. the critical role of passive permeability in designing successful drugs. chemmedchem. (2020). [169] k. utsey, m.s. gastonguay, s. russell, r. freling, m.m. riggs, a. elmokadem. quantification of the impact of partition coefficient prediction methods on physiologically based pharmacokinetic model output using a standardized tissue composition. drug metab. dispos. 48 (2020) 903–916. [170] k. valko, s. rava, s. bunally, s. anderson. revisiting the application of immobilized artificial membrane (iam) chromatography to estimate in vivo distribution properties of drug discovery compounds based on the model of marketed drugs. admet dmpk. 8 (2020) 78–97. https://doi.org/10.5599/admet.757. [171] f. lombardo, j. bentzien, g. berellini, i. muegge. in silico models of human pk parameters. prediction of volume of distribution using an extensive data set and a reduced number of parameters. j. pharm. sci. (2020). https://doi.org/10.5599/admet.923 kiyohiko sugano admet & dmpk 9(2) (2021) 75-109 108 [172] d. hallifax, f.l. wood. importance of the unstirred water layer and hepatocyte membrane integrity in vitro for quantification of intrinsic metabolic clearance. drug metab. dispos. 46 (2018) 268–278. [173] k. sugano. aqueous boundary layers related to oral absorption of a drug: from dissolution of a drug to carrier mediated transport and intestinal wall metabolism. mol. pharm. 7 (2010) 1362–1373. [174] l. di, e.h. kerns, k. fan, o.j. mcconnell, g.t. carter. high throughput artificial membrane permeability assay for blood--brain barrier. eur. j. med. chem. 38 (2003) 223–232. [175] c. dagenais, a. avdeef, o. tsinman, a. dudley, r. beliveau. p-glycoprotein deficient mouse in situ blood-brain barrier permeability and its prediction using an in combo pampa model. eur. j. pharm. sci. 38 (2009) 121–137. [176] b. sinkó, t.m. garrigues, g.t. balogh, z.k. nagy, o. tsinman, a. avdeef, k. takács-novák. skinpampa: a new method for fast prediction of skin penetration. eur. j. pharm. sci. 45 (2012) 698–707. [177] a. avdeef. absorption and drug development: solubility, permeability, and charge state, second edition, john wiley & sons, inc., 2012. [178] k. sugano. a simulation of oral absorption using classical nucleation theory. int. j. pharm. 378 (2009). [179] m. higuchi, y. yoshihashi, k. tarada, k. sugano. minimum rotation speed to prevent coning phenomena in compendium paddle dissolution apparatus. eur. j. pharm. sci. 65 (2014) 74–78. [180] c. stillhart, k. vučićević, p. augustijns, a.w. basit, h. batchelor, t.r. flanagan, i. gesquiere, r. greupink, d. keszthelyi, m. koskinen, others. impact of gastrointestinal physiology on drug absorption in special populations--an ungap review. eur. j. pharm. sci. (2020) 105280. [181] m. vertzoni, p. augustijns, m. grimm, m. koziolek, g. lemmens, n. parrott, c. pentafragka, c. reppas, j. rubbens, j. van den abeele, others. impact of regional differences along the gastrointestinal tract of healthy adults on oral drug absorption: an ungap review. eur. j. pharm. sci. 134 (2019) 153–175. [182] d.m. mudie, g.l. amidon, g.e. amidon. physiological parameters for oral delivery and in vitro testing. mol. pharm. 7 (2010) 1388–1405. [183] b. hens, y. tsume, m. bermejo, p. paixao, m.j. koenigsknecht, j.r. baker, w.l. hasler, r. lionberger, j. fan, j. dickens, k. shedden, b. wen, j. wysocki, r. loebenberg, a. lee, a. frances, g. amidon, a. yu, g. benninghoff, n. salehi, a. talattof, d. sun, g.l. amidon. low buffer capacity and alternating motility along the human gastrointestinal tract: implications for in vivo dissolution and absorption of ionizable drugs. mol. pharm. 14 (2017) 4281–4294. [184] j. al-gousous, g.l. amidon, p. langguth. toward biopredictive dissolution for enteric coated dosage forms. mol. pharm. 13 (2016) 1927–1936. [185] j. al-gousous, k.x. sun, d.p. mcnamara, b. hens, n. salehi, p. langguth, m. bermejo, g.e. amidon, g.l. amidon. mass transport analysis of the enhanced buffer capacity of the bicarbonate-co2 buffer in a phase-heterogenous system: physiological and pharmaceutical significance. mol. pharm. 15 (2018) 5291–5301. [186] t. uekusa, j. oki, m. omori, d. watanabe, d. inoue, k. sugano. effect of buffer capacity on dissolution and supersaturation profiles of pioglitazone hydrochloride. j. drug deliv. sci. technol. 55 (2020) 101492. [187] h. ogata, n. aoyagi, n. kaniwa, a. ejima, t. kitaura, t. ohki, k. kitamura. evaluation of beagle dogs as an animal model for bioavailability testing of cinnarizine capsules. int. j. pharm. 29 (1986) 121–126. [188] k. sagawa, f. li, r. liese, s.c. sutton. fed and fasted gastric ph and gastric residence time in conscious beagle dogs. j. pharm. sci. 98 (2009) 2494–2500. [189] c.y. lui, g.l. amidon, r.r. berardi, d. fleisher, c. youngberg, j.b. dressman. comparison of gastrointestinal ph in dogs and humans: implications on the use of the beagle dog as a model for oral absorption in humans [published erratum appears in j pharm sci 1986 dec;75(12):1207]. j. pharm. sci. 75 (1986) 271–274. [190] n. matsumura, y. yamaura, j. katagi, s. ono, s. kim, s. yamashita, k. sugano. evaluation of using dogs to predict fraction of oral dose absorbed in humans for poorly water-soluble drugs. j. pharm. sci. 107 (2018) 2489–2496. admet & dmpk 9(2) (2021) 75-109 lost in modelling and simulation? doi: https:/doi.org/10.5599/admet.923 109 [191] r. solid, o. dosage. guidance for industry guidance for industry dissolution testing of immediate. (1997). [192] v.p. shah, l.j. lesko, j. fan, n. fleischer, j. handerson, h. malinowski, m. makary, l. ouderkirk, s. bay, p. sathe, g.j.p. singh, l. iillman, y. tsong, r.i. williams. fda guidance for industry 1 dissolution testing of immediate release solid oral dosage forms. dissolution technol. 4 (1997) 15–22. [193] fda. extended release oral dosage forms: development, evaluation, and application of in vitro/in vivo correlations. (1997) 1–24. http://www.fda.gov/cder/guidance/index.htm. [194] venkata ramana s. uppoor. regulatory perspectives on in vitro (dissolution) /in vivo (bioavailability) correlations. j. control. release. 72 (2001) 127–132. [195] redc/adf. dissolution testing and acceptance criteria for immediate-release solid oral dosage form drug products containing high solubility drug substances guidance for industry. fda. (2018) august-august. ©2021 by the authors; licensee iapc, zagreb, croatia. this article is an open-access article distributed under the terms and conditions of the creative commons attribution license (http://creativecommons.org/licenses/by/3.0/) https://doi.org/10.5599/admet.923 http://www.fda.gov/cder/guidance/index.htm http://creativecommons.org/licenses/by/3.0/